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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
83ac3e2b-f42c-4b09-b801-2a371c4fb6fb | an-evaluation-of-binary-comparative-lexical | null | null | https://aclanthology.org/2022.bea-1.24 | https://aclanthology.org/2022.bea-1.24.pdf | An Evaluation of Binary Comparative Lexical Complexity Models | Identifying complex words in texts is an important first step in text simplification (TS) systems. In this paper, we investigate the performance of binary comparative Lexical Complexity Prediction (LCP) models applied to a popular benchmark dataset — the CompLex 2.0 dataset used in SemEval-2021 Task 1. With the data from CompLex 2.0, we create a new dataset contain 1,940 sentences referred to as CompLex-BC. Using CompLex-BC, we train multiple models to differentiate which of two target words is more or less complex in the same sentence. A linear SVM model achieved the best performance in our experiments with an F1-score of 0.86. | ['Matthew Shardlow', 'Marcos Zampieri', 'Kai North'] | null | null | null | null | naacl-bea-2022-7 | ['lexical-complexity-prediction'] | ['natural-language-processing'] | [ 3.91364157e-01 2.03254186e-02 -3.33798707e-01 -2.52252340e-01
-6.38529181e-01 -3.84951413e-01 6.80979848e-01 6.07819557e-01
-7.47562110e-01 9.87348258e-01 2.76082903e-01 -4.10690963e-01
1.37392640e-01 -4.31823134e-01 -4.51991558e-01 -2.49398574e-01
4.85492706e-01 5.90315342e-01 6.27077371e-02 -5.36258221e-01
6.55427635e-01 3.55747223e-01 -1.26248193e+00 5.10449886e-01
1.25100160e+00 6.02815092e-01 3.82334530e-01 7.93834865e-01
-2.33650759e-01 4.92693573e-01 -9.92373943e-01 -8.11867177e-01
1.34791493e-01 -3.98778617e-01 -8.51326108e-01 -5.83570361e-01
6.27601504e-01 2.52123028e-01 -2.26774663e-01 1.11669195e+00
4.53951836e-01 1.29706547e-01 8.90295088e-01 -8.59962106e-01
-1.61897112e-02 1.20501482e+00 -1.59551591e-01 6.75633788e-01
3.90878916e-01 -1.26159027e-01 1.39339221e+00 -8.91336441e-01
8.46627414e-01 1.36048424e+00 6.03264213e-01 4.62075114e-01
-1.30643737e+00 -5.64728379e-01 -1.28152549e-01 5.21787107e-01
-1.27174377e+00 -7.25216269e-01 6.77731097e-01 -4.34258610e-01
1.31912649e+00 5.58829725e-01 6.07989788e-01 8.67180645e-01
5.06964982e-01 9.80839729e-01 1.05704892e+00 -7.27570474e-01
-1.75718233e-01 -4.42496780e-03 6.37704611e-01 4.86810863e-01
4.42402005e-01 -3.51616472e-01 -4.82436359e-01 -3.15612853e-02
-1.89291194e-01 -5.73140681e-01 -1.92250535e-01 3.74994189e-01
-9.94269133e-01 9.54219222e-01 -5.78351878e-02 3.41505140e-01
-1.78246185e-01 -3.82408470e-01 7.88420081e-01 6.48351669e-01
5.20947993e-01 8.10156226e-01 -6.41986251e-01 -4.35173571e-01
-9.86959040e-01 4.95192379e-01 9.36473012e-01 9.01965916e-01
1.47688121e-01 -2.59362608e-02 -4.24800903e-01 1.10639846e+00
-3.61691505e-01 4.87812161e-01 7.25001335e-01 -4.20462340e-01
9.48333740e-01 5.18703282e-01 -2.84714520e-01 -6.15905404e-01
-5.95184267e-01 -4.86798495e-01 -8.98547053e-01 -2.17427462e-01
1.52054295e-01 -1.29292145e-01 -6.53685272e-01 1.38567293e+00
-9.17308256e-02 -4.83826131e-01 6.19587824e-02 1.75913140e-01
1.00476778e+00 6.97186470e-01 9.04420018e-03 -4.11211550e-01
9.42384899e-01 -1.26830590e+00 -9.79298592e-01 -1.78606704e-01
1.14497888e+00 -1.10112870e+00 1.21540666e+00 6.94405258e-01
-1.09436810e+00 -4.31645274e-01 -1.25814509e+00 -3.33122998e-01
-4.71317679e-01 2.41524130e-01 2.09831268e-01 5.57635546e-01
-5.10303080e-01 9.49869156e-01 -4.26667809e-01 -1.78951979e-01
3.32056165e-01 1.60061076e-01 -2.05740839e-01 -1.97954588e-02
-1.18728256e+00 1.37234938e+00 4.80011493e-01 -4.86836702e-01
-1.29916802e-01 -8.41004550e-01 -7.22979665e-01 1.65831819e-01
2.05560654e-01 -4.13080752e-01 1.31219840e+00 -8.24018061e-01
-1.30430949e+00 8.57863843e-01 -3.70748430e-01 -6.78084791e-01
5.51480174e-01 -4.88875180e-01 -4.97712165e-01 -3.16515625e-01
1.95016786e-01 3.00407946e-01 8.02608132e-01 -7.94752121e-01
-8.20723891e-01 -5.28154150e-02 -1.80876255e-01 3.71791899e-01
-2.04019934e-01 3.14967215e-01 -3.13062876e-01 -9.29286122e-01
-3.41273546e-01 -9.38118994e-01 1.18318528e-01 -6.55085683e-01
-7.58894265e-01 -6.29512191e-01 4.58780885e-01 -1.06466496e+00
1.98007095e+00 -1.88895535e+00 7.44615555e-01 3.35218757e-03
2.06640348e-01 6.64746702e-01 -1.27836332e-01 3.58310789e-01
-1.98391005e-01 2.42480949e-01 -2.25537583e-01 -7.19863534e-01
-1.02658920e-01 -7.93766752e-02 3.93384974e-03 1.65043563e-01
2.28596449e-01 8.15326393e-01 -7.13051200e-01 -6.63843155e-01
2.03139976e-01 -1.43245697e-01 -5.60505748e-01 4.19168137e-02
-1.47708923e-01 -8.21822956e-02 -2.72105262e-02 5.14057696e-01
5.00736594e-01 4.34452802e-01 8.00719634e-02 -1.51824474e-01
-1.36950031e-01 7.13317692e-01 -7.12362528e-01 1.29238975e+00
-6.29716456e-01 1.13284922e+00 -3.64857435e-01 -8.31894457e-01
6.62218153e-01 6.61686342e-03 9.19679776e-02 -9.24654543e-01
3.30440342e-01 1.83959097e-01 4.64914858e-01 -3.37264150e-01
7.94743598e-01 -6.12056144e-02 -3.25382948e-01 1.99017152e-02
2.53275689e-03 -5.42191148e-01 9.30137098e-01 2.33365849e-01
1.22859609e+00 -4.79529381e-01 5.74677646e-01 -4.36751962e-01
8.57396483e-01 1.19232520e-01 5.73166788e-01 5.57030737e-01
-2.38026664e-01 5.18463433e-01 7.79369295e-01 -1.68922290e-01
-1.25951278e+00 -5.88959098e-01 -2.02311411e-01 1.07242608e+00
-3.68368089e-01 -9.84063745e-01 -7.71466017e-01 -6.24450088e-01
4.82024103e-02 1.27363610e+00 -4.39224690e-01 -2.58034706e-01
-1.08829665e+00 -5.89630783e-01 6.76911354e-01 2.05850247e-02
4.71644178e-02 -1.25187850e+00 -1.38323084e-01 1.47444740e-01
-4.51027185e-01 -1.23565471e+00 -7.37497747e-01 3.00338835e-01
-6.01134539e-01 -1.06698298e+00 -5.94625294e-01 -8.09502184e-01
8.11828598e-02 -2.94277668e-02 1.28195548e+00 3.24251428e-02
-1.65320903e-01 -6.00162208e-01 -5.71183622e-01 -7.48628080e-01
-6.90311790e-01 7.79099524e-01 1.73963353e-01 -3.98857057e-01
1.72552273e-01 1.14248097e-02 7.32615441e-02 -2.45900974e-01
-6.47063434e-01 1.72849908e-01 4.85025167e-01 8.73497605e-01
4.51142341e-01 4.80534732e-02 2.72155970e-01 -1.20922363e+00
9.72230315e-01 -3.27890575e-01 -4.38437551e-01 2.34566644e-01
-7.68040657e-01 7.64333904e-02 1.12775624e+00 -5.24097621e-01
-7.80173123e-01 -2.54112571e-01 -4.03602600e-01 -4.36541364e-02
3.07137430e-01 8.32674503e-01 -3.33753228e-01 1.75599828e-01
6.24253094e-01 1.43496722e-01 -4.07575041e-01 -7.52843499e-01
1.61806405e-01 8.09203565e-01 3.32128108e-01 -2.85277814e-01
6.69967115e-01 -3.70611668e-01 8.43675956e-02 -1.20009959e+00
-1.19028664e+00 -4.20701236e-01 -8.53850782e-01 2.10057989e-01
4.47787881e-01 -6.96732342e-01 -2.66960770e-01 5.80778539e-01
-1.26968014e+00 -3.26987505e-01 4.12311926e-02 2.29816630e-01
-1.57265574e-01 5.04463971e-01 -3.80776674e-01 -3.32188606e-01
-7.88558125e-01 -1.08216727e+00 6.55616224e-01 8.11723620e-02
-8.45390141e-01 -7.62290418e-01 1.45951182e-01 4.64558870e-01
3.54335874e-01 3.64667512e-02 1.53921103e+00 -8.72662604e-01
1.78176045e-01 -2.73534685e-01 9.87715274e-02 6.14670038e-01
-3.72846797e-02 3.68941605e-01 -4.37593967e-01 -1.72734439e-01
-3.12529355e-01 -6.53278157e-02 9.21043158e-01 2.43550345e-01
1.18870032e+00 -2.08429441e-01 -1.21846721e-01 6.25569224e-01
1.00041342e+00 1.97846303e-03 5.04954517e-01 5.29160261e-01
9.58293974e-01 4.02772158e-01 8.54199946e-01 3.26090366e-01
2.99052924e-01 8.83831382e-01 -1.68766126e-01 1.55988783e-01
-4.47572857e-01 -2.13954687e-01 4.00750518e-01 1.52132702e+00
3.12986434e-01 -3.62163305e-01 -1.02583158e+00 2.47862726e-01
-1.41527998e+00 -7.03303337e-01 -6.06926322e-01 2.03161383e+00
1.33824778e+00 7.16669500e-01 1.13576464e-02 7.01951742e-01
6.94824457e-01 1.53752401e-01 -1.93353936e-01 -9.52865601e-01
-5.45401871e-01 7.40613401e-01 4.75195348e-01 6.91131234e-01
-1.17812443e+00 1.41435683e+00 5.81445217e+00 1.36813998e+00
-9.68537927e-01 -1.02520389e-02 4.20612514e-01 -1.91744030e-01
-1.16692834e-01 -1.98996291e-01 -1.08261871e+00 7.04640687e-01
1.30153012e+00 -6.78737462e-01 4.66935307e-01 5.25232375e-01
1.97462976e-01 -1.22905925e-01 -1.13582599e+00 9.82761085e-01
2.81548142e-01 -1.19722056e+00 4.89529610e-01 -3.53088558e-01
7.64562428e-01 1.58715263e-01 -1.79697499e-01 7.15610623e-01
-1.72055721e-01 -1.16384351e+00 9.38561141e-01 4.98857856e-01
1.03853226e+00 -8.58793497e-01 8.27395439e-01 5.98497391e-01
-9.18254077e-01 1.45816937e-01 -2.80203968e-01 -6.02022447e-02
5.86130321e-02 7.54437625e-01 -6.42652094e-01 3.78226340e-01
3.50381434e-01 8.29943538e-01 -8.91912460e-01 1.07548988e+00
-2.85951644e-01 7.03595519e-01 -1.58806264e-01 -3.97513658e-01
-1.53353959e-01 -1.01693980e-01 7.67544687e-01 1.52226126e+00
-2.36278117e-01 -7.10412934e-02 -3.26572537e-01 3.11894625e-01
-3.77814800e-01 6.07487619e-01 -2.55087584e-01 -9.68776345e-02
6.81373119e-01 9.83798265e-01 -3.01320910e-01 -6.77328944e-01
-2.88911194e-01 1.02602911e+00 5.32841265e-01 -2.12496206e-01
-8.01902533e-01 -9.82868791e-01 4.85653967e-01 -1.76603436e-01
2.33055592e-01 -1.28669962e-01 -7.33739853e-01 -1.33340251e+00
8.60719159e-02 -1.24558914e+00 1.59213409e-01 -4.51764107e-01
-1.13347507e+00 7.45011568e-01 1.41262766e-02 -9.33870196e-01
1.96751028e-01 -6.29916847e-01 -6.44935250e-01 1.00332880e+00
-1.52291954e+00 -6.99268460e-01 2.39797365e-02 2.35297248e-01
1.11516345e+00 -2.79020637e-01 8.39266479e-01 4.19116139e-01
-9.72978473e-01 1.11249316e+00 5.41156292e-01 -1.69305038e-02
7.36447752e-01 -1.39125896e+00 8.32073331e-01 7.54133463e-01
1.37773128e-02 3.40525299e-01 7.93712914e-01 -7.61514008e-01
-1.03924513e+00 -1.10485327e+00 1.72886109e+00 -5.76517105e-01
5.67925334e-01 -3.12642306e-01 -8.44595075e-01 3.21852088e-01
6.18274929e-03 -6.38888597e-01 4.61415768e-01 1.39806464e-01
-1.41320631e-01 1.48561835e-01 -1.03929675e+00 6.83322132e-01
8.25626254e-01 -3.56057137e-01 -8.96012664e-01 3.26350927e-01
9.06802297e-01 -5.48956811e-01 -8.42552304e-01 2.74906546e-01
4.05740380e-01 -5.29485703e-01 6.71969235e-01 -6.88561559e-01
7.21751869e-01 3.71562958e-01 -1.92994058e-01 -1.58554649e+00
-2.34541610e-01 -6.00038290e-01 -2.84040034e-01 1.03455257e+00
8.11553538e-01 -2.99134910e-01 4.97832417e-01 1.00346938e-01
-3.10784906e-01 -8.24530244e-01 -1.01895964e+00 -6.93489134e-01
5.85800529e-01 -2.57854581e-01 4.24680084e-01 9.61145401e-01
2.45547276e-02 9.36371267e-01 -2.65905820e-02 -5.01406372e-01
4.03709799e-01 -1.65257439e-01 6.67755544e-01 -1.42029822e+00
3.68394181e-02 -9.81528223e-01 -1.19186468e-01 -8.67115319e-01
5.28528988e-01 -1.13976109e+00 -3.42601150e-01 -1.13417101e+00
2.67685235e-01 -1.93518624e-01 -3.69933881e-02 2.72082806e-01
-5.25815547e-01 -2.01880202e-01 4.46215034e-01 1.20013013e-01
-2.71632761e-01 8.10063481e-01 1.01193810e+00 -3.22642952e-01
-2.88332611e-01 2.84615278e-01 -5.32022178e-01 5.88555813e-01
1.14833057e+00 -6.70911491e-01 1.17740437e-01 -3.85133892e-01
8.08565393e-02 -9.91008803e-02 -4.54518348e-01 -7.83892989e-01
-2.43305698e-01 -7.65865520e-02 2.70479899e-02 -8.70709598e-01
2.73462832e-01 -3.42904896e-01 -2.17947841e-01 6.69020116e-01
-5.27079105e-01 3.04820418e-01 2.75870234e-01 -1.19453734e-02
-1.17524520e-01 -6.98472202e-01 8.16543519e-01 1.59763932e-01
-1.85426295e-01 -1.08306542e-01 -3.96784782e-01 3.80502582e-01
6.67818367e-01 1.12701356e-01 -3.96594197e-01 1.27347931e-01
-3.38200659e-01 3.72750871e-02 3.08734238e-01 5.79379022e-01
2.91099638e-01 -9.53135192e-01 -9.69872952e-01 3.03371740e-03
2.68673986e-01 -2.32708842e-01 -3.87358278e-01 6.71295583e-01
-8.14921856e-01 7.84329534e-01 -1.06283449e-01 -1.37621909e-01
-1.78232646e+00 2.77152598e-01 1.73459306e-01 -4.95185286e-01
-5.34892321e-01 8.42088223e-01 -5.13807893e-01 -4.98808652e-01
1.98259607e-01 -4.50026900e-01 -5.88055074e-01 6.12845644e-02
3.61715138e-01 7.04084933e-01 6.59554958e-01 -9.36902702e-01
-2.25848660e-01 3.34593683e-01 -5.18043280e-01 -1.05350681e-01
1.26035178e+00 8.12316164e-02 -2.60542095e-01 4.96397108e-01
1.38067722e+00 1.27007276e-01 -2.34012991e-01 -4.36725557e-01
4.47098553e-01 -4.26323384e-01 2.85389334e-01 -7.68930018e-01
-6.54724658e-01 6.96426451e-01 3.08217257e-01 2.81076729e-02
8.35199058e-01 -3.58936548e-01 9.67471778e-01 6.76711857e-01
-1.33555010e-01 -1.34271204e+00 -2.79500067e-01 1.24025941e+00
9.84480381e-01 -1.21063626e+00 2.93906659e-01 -5.21448374e-01
-8.26334894e-01 9.78847980e-01 4.29500431e-01 -1.91859957e-02
5.04772663e-01 -4.71643470e-02 -2.48309702e-01 -2.29235552e-02
-9.20472860e-01 -7.23323599e-02 4.55033332e-01 1.91608310e-01
6.38210475e-01 4.00090218e-01 -1.14196801e+00 8.10563266e-01
-7.87422299e-01 -4.51905847e-01 7.42968440e-01 7.58950293e-01
-4.76654500e-01 -1.22004318e+00 -1.12127125e-01 9.25926566e-01
-5.74506581e-01 -6.90434515e-01 -6.46637619e-01 6.63296878e-01
-8.38263780e-02 9.48044240e-01 -1.90498441e-01 -4.85264391e-01
6.36314511e-01 1.64309651e-01 5.43082952e-01 -7.92098343e-01
-8.04979146e-01 -2.06268698e-01 5.55180073e-01 -1.31961033e-01
1.65500760e-01 -8.58563304e-01 -1.13449371e+00 -6.51767671e-01
-3.93051386e-01 -2.93553956e-02 8.41276228e-01 1.10698175e+00
1.55461561e-02 5.74041128e-01 7.09388077e-01 -6.65409267e-01
-5.98836720e-01 -1.32772720e+00 -2.88363963e-01 5.68953097e-01
1.63824171e-01 -4.43264306e-01 -3.81412655e-01 8.14699680e-02] | [10.945839881896973, 10.402236938476562] |
9058a331-105d-4293-9238-476fe5661097 | bias-eliminated-semantic-refinement-for-any | 2202.04827 | null | https://arxiv.org/abs/2202.04827v1 | https://arxiv.org/pdf/2202.04827v1.pdf | Bias-Eliminated Semantic Refinement for Any-Shot Learning | When training samples are scarce, the semantic embedding technique, ie, describing class labels with attributes, provides a condition to generate visual features for unseen objects by transferring the knowledge from seen objects. However, semantic descriptions are usually obtained in an external paradigm, such as manual annotation, resulting in weak consistency between descriptions and visual features. In this paper, we refine the coarse-grained semantic description for any-shot learning tasks, ie, zero-shot learning (ZSL), generalized zero-shot learning (GZSL), and few-shot learning (FSL). A new model, namely, the semantic refinement Wasserstein generative adversarial network (SRWGAN) model, is designed with the proposed multihead representation and hierarchical alignment techniques. Unlike conventional methods, semantic refinement is performed with the aim of identifying a bias-eliminated condition for disjoint-class feature generation and is applicable in both inductive and transductive settings. We extensively evaluate model performance on six benchmark datasets and observe state-of-the-art results for any-shot learning; eg, we obtain 70.2% harmonic accuracy for the Caltech UCSD Birds (CUB) dataset and 82.2% harmonic accuracy for the Oxford Flowers (FLO) dataset in the standard GZSL setting. Various visualizations are also provided to show the bias-eliminated generation of SRWGAN. Our code is available. | ['Xi Li', 'Chunhui Zhao', 'Liangjun Feng'] | 2022-02-10 | null | null | null | null | ['generalized-zero-shot-learning', 'generalized-zero-shot-learning'] | ['computer-vision', 'methodology'] | [ 4.50044900e-01 3.48546863e-01 -1.49154931e-01 -4.63010907e-01
-9.78089035e-01 -5.05805671e-01 8.21127415e-01 -4.61507738e-02
-1.22647919e-01 6.80019379e-01 7.35966861e-02 2.13780493e-01
-1.78698391e-01 -9.57269430e-01 -8.07969272e-01 -1.07413518e+00
2.78490484e-01 4.07763243e-01 6.60485923e-02 -2.53966570e-01
-1.09508321e-01 1.44723251e-01 -1.79285514e+00 7.31934756e-02
6.90225065e-01 1.04017663e+00 -1.24673983e-02 4.98966426e-01
-1.40385583e-01 5.88177800e-01 -4.18489069e-01 -6.06241047e-01
2.00086057e-01 -5.11560619e-01 -7.15329826e-01 1.41961426e-01
4.01178896e-01 -1.29676059e-01 -1.39021233e-01 1.12183595e+00
6.26065373e-01 6.02119625e-01 9.99071419e-01 -1.66396010e+00
-1.17452395e+00 5.22120893e-01 -4.03393567e-01 -1.95049509e-01
1.83919787e-01 2.04931408e-01 1.17749238e+00 -1.00957799e+00
8.46346200e-01 1.28065896e+00 3.59460801e-01 1.03743052e+00
-1.43818057e+00 -6.55548334e-01 -2.50053890e-02 4.21934247e-01
-1.53749561e+00 -3.33732158e-01 8.76906753e-01 -5.20204067e-01
6.42815411e-01 1.70204923e-01 4.54017907e-01 1.50213325e+00
-1.97633371e-01 7.71628141e-01 9.18749809e-01 -4.91728216e-01
5.07954419e-01 3.02139610e-01 9.23758149e-02 5.66608489e-01
1.75416335e-01 2.63185650e-01 -5.09054065e-01 -8.20402205e-02
4.68147248e-01 2.14531675e-01 -1.74647108e-01 -7.12781250e-01
-9.60381866e-01 1.13495219e+00 8.18929672e-01 1.89239666e-01
1.59334522e-02 4.07572240e-02 3.18751544e-01 2.12287590e-01
6.10889018e-01 4.83581156e-01 -9.29500163e-02 2.05779448e-01
-6.60488129e-01 2.83395559e-01 5.13643861e-01 1.34534919e+00
8.79474282e-01 3.90913188e-01 -6.23623013e-01 8.48498344e-01
1.07068442e-01 5.02650023e-01 5.89588225e-01 -6.96312129e-01
-9.39826146e-02 3.72106522e-01 1.31226750e-02 -6.88428223e-01
-1.60217971e-01 -2.24085122e-01 -9.90830362e-01 1.20789148e-01
2.13634253e-01 3.62186320e-03 -1.34364057e+00 1.98649538e+00
5.20919979e-01 4.26701039e-01 3.63343805e-01 1.03720868e+00
1.50793111e+00 6.16468430e-01 3.49660695e-01 -1.04906806e-03
1.25322056e+00 -9.14411664e-01 -6.86479330e-01 -2.46459134e-02
4.82619703e-01 -3.71821761e-01 1.20669961e+00 -9.68465135e-02
-5.95998526e-01 -7.35486031e-01 -1.18969381e+00 -1.70985982e-01
-7.70690382e-01 -3.04761946e-01 5.22559524e-01 3.45014244e-01
-7.34441996e-01 6.15760565e-01 -6.60505056e-01 -4.79916394e-01
5.57007790e-01 7.73803666e-02 -4.53174829e-01 -1.90713957e-01
-1.43151700e+00 5.38670003e-01 5.11242986e-01 -2.90323883e-01
-1.22295380e+00 -8.27876389e-01 -1.31653941e+00 2.78990090e-01
4.37401861e-01 -6.53600156e-01 8.94150198e-01 -9.02126670e-01
-1.27119744e+00 9.85207438e-01 5.86862527e-02 -4.07558233e-01
2.37836242e-01 6.56612590e-02 -2.21482515e-01 6.78675622e-02
7.55470395e-02 1.01158774e+00 1.22568583e+00 -1.62743139e+00
-3.33718687e-01 -4.18917060e-01 7.12531656e-02 7.92168975e-02
-3.08226734e-01 -3.35627288e-01 -1.57942116e-01 -7.80562103e-01
-2.16553867e-01 -8.92605364e-01 -7.96540678e-02 5.52980900e-02
-4.71020520e-01 -3.60966921e-01 9.18194413e-01 -2.79832989e-01
7.05035985e-01 -2.42782736e+00 3.90364736e-01 3.17952782e-02
2.39072099e-01 1.71858981e-01 -2.83161610e-01 1.85339168e-01
-8.26973394e-02 2.37482637e-02 -4.14469361e-01 -3.83865476e-01
2.89380163e-01 2.95170546e-01 -4.13936734e-01 2.97346830e-01
3.94455075e-01 1.19955063e+00 -1.12509751e+00 -5.32030523e-01
3.32891792e-01 5.96343338e-01 -4.00784194e-01 4.06595916e-01
-1.11613944e-01 3.11727703e-01 -2.67699569e-01 9.08226371e-01
3.30425024e-01 -3.72642905e-01 -2.42175117e-01 -3.24173689e-01
2.99703866e-01 -4.98331517e-01 -9.74620461e-01 1.81838977e+00
-4.04783219e-01 3.55728537e-01 -5.17868578e-01 -8.32531810e-01
1.05520952e+00 3.27043921e-01 1.54734299e-01 -3.64462733e-01
2.88423508e-01 -9.57773998e-02 -2.31233239e-01 -2.79055983e-01
2.64163613e-01 -4.96738523e-01 -3.35496247e-01 1.59367874e-01
7.96086252e-01 -2.69763231e-01 1.29922614e-01 2.82698542e-01
9.01834726e-01 2.43509263e-01 6.74976587e-01 -1.14329599e-01
1.22551925e-01 -9.08239856e-02 5.95004261e-01 6.99805915e-01
-2.65292615e-01 9.64122057e-01 3.28099221e-01 -2.80336380e-01
-1.13067091e+00 -1.33567286e+00 -2.84073621e-01 1.15994203e+00
6.01277828e-01 -3.83923769e-01 -6.56611919e-01 -7.28417337e-01
2.27076784e-02 1.14794314e+00 -1.05760741e+00 -6.83459282e-01
1.98077053e-01 -7.18140543e-01 3.44781488e-01 6.92817688e-01
3.08833092e-01 -1.22742438e+00 -6.29567087e-01 1.55073386e-02
4.60100062e-02 -9.77390826e-01 -2.73866683e-01 2.29259402e-01
-3.01425487e-01 -1.04169023e+00 -8.54298890e-01 -6.60270452e-01
7.02694058e-01 5.64402714e-02 9.46372688e-01 -1.74088195e-01
-6.11532092e-01 3.84523392e-01 -8.18990827e-01 -3.94308597e-01
-8.50234032e-02 -1.85981631e-01 2.79751439e-02 3.05368304e-01
4.56190944e-01 -3.74869883e-01 -4.45683599e-01 2.06093580e-01
-8.97348881e-01 1.26066908e-01 4.89444613e-01 1.36105168e+00
7.62184560e-01 -1.68196350e-01 6.45850837e-01 -1.05407238e+00
1.91548869e-01 -5.47544360e-01 -2.63155818e-01 2.97569990e-01
-5.04915953e-01 1.67237058e-01 7.37574875e-01 -7.11915612e-01
-1.09551048e+00 7.58279860e-02 1.83959976e-02 -9.50772345e-01
-4.21949893e-01 -1.21486895e-02 -4.24882799e-01 -1.20916322e-01
7.76311874e-01 2.01066315e-01 -1.44667491e-01 -2.41598815e-01
7.04432666e-01 5.50716221e-01 5.25605500e-01 -4.28533524e-01
1.09742689e+00 4.46187735e-01 -4.55372408e-02 -7.33538628e-01
-1.13679266e+00 -4.00044709e-01 -6.27180457e-01 -7.06557259e-02
1.09680319e+00 -7.38044798e-01 -2.99062043e-01 3.36914569e-01
-8.37471128e-01 -1.96590811e-01 -7.67332196e-01 2.58261949e-01
-8.51203799e-01 -1.12354837e-01 -3.48347902e-01 -6.81489587e-01
-5.13992012e-01 -1.01982236e+00 1.34561706e+00 4.13049310e-01
-2.09485337e-01 -9.60160673e-01 4.64249142e-02 1.29873887e-01
3.08041871e-01 7.07150400e-01 1.08617425e+00 -9.96507287e-01
-2.56143481e-01 -5.24736010e-02 -2.18290016e-01 3.24757487e-01
9.22452956e-02 -5.92853166e-02 -1.34412837e+00 -3.40378791e-01
-2.81009287e-01 -7.84666240e-01 8.76091659e-01 2.66695410e-01
1.15978634e+00 -2.55620301e-01 -2.42354482e-01 8.25485528e-01
1.54031312e+00 1.84567437e-01 5.95117748e-01 2.72421837e-02
8.87127578e-01 5.89246511e-01 8.42716217e-01 6.16606891e-01
5.95320426e-02 6.77444160e-01 4.88347501e-01 -1.21574454e-01
-1.91986442e-01 -4.90218252e-01 -5.18791564e-02 3.68897200e-01
3.32516548e-03 -2.83590019e-01 -7.09240019e-01 4.94499654e-01
-1.86603332e+00 -9.44753945e-01 3.57234150e-01 2.12916946e+00
7.52981365e-01 -3.39540392e-02 -1.04807787e-01 9.65012088e-02
7.75243580e-01 3.08096319e-01 -6.78534627e-01 -1.35283306e-01
-2.06516590e-02 2.61009455e-01 3.22215706e-01 2.65355974e-01
-1.28961527e+00 1.19078040e+00 5.27883053e+00 1.13899481e+00
-7.60864019e-01 2.85982132e-01 4.65428799e-01 -1.57283783e-01
-3.13150436e-01 -9.09666941e-02 -9.02360141e-01 4.15741831e-01
5.73837578e-01 -4.38257307e-01 3.88211131e-01 1.03968525e+00
-2.98170418e-01 3.37878108e-01 -1.14456236e+00 1.11854255e+00
4.20531154e-01 -1.09746861e+00 3.21107596e-01 -2.44915679e-01
7.84981966e-01 -4.69759792e-01 1.97596669e-01 7.02941358e-01
4.64863092e-01 -1.05174482e+00 5.20748317e-01 5.17213523e-01
1.26828218e+00 -9.03493106e-01 7.17881143e-01 1.23439692e-01
-1.24452639e+00 -3.74829285e-02 -6.07682824e-01 3.72823626e-01
1.46290198e-01 1.66799828e-01 -6.12363756e-01 5.96420825e-01
8.12092602e-01 9.41985130e-01 -6.07555449e-01 6.92814529e-01
-3.92007917e-01 4.41555858e-01 5.15604541e-02 -3.00526135e-02
2.00351492e-01 2.17229854e-02 5.80906212e-01 8.20363760e-01
2.64582515e-01 1.51059344e-01 -6.86315773e-03 1.17552793e+00
-1.18497327e-01 3.68364900e-02 -7.84771800e-01 -9.35927629e-02
5.45821428e-01 1.47810328e+00 -5.54745197e-01 -4.20768589e-01
-2.41342843e-01 8.86197209e-01 3.00681233e-01 4.01969820e-01
-9.20455456e-01 -6.22557938e-01 6.80666268e-01 -1.00541346e-01
3.71258110e-01 3.78732353e-01 5.19679263e-02 -1.11000407e+00
-4.55372572e-01 -6.20022178e-01 3.88844997e-01 -9.48427379e-01
-1.60419643e+00 5.49972236e-01 2.21500143e-01 -1.50830388e+00
-3.72960091e-01 -4.47778761e-01 -6.49207532e-01 5.59671402e-01
-1.26251554e+00 -1.51797831e+00 -7.03282714e-01 6.69365048e-01
8.04055572e-01 -2.67492115e-01 1.04792607e+00 -3.61404056e-03
-3.68873864e-01 6.94362998e-01 7.02672601e-02 5.83220795e-02
8.35491359e-01 -1.44246769e+00 2.21216798e-01 6.14850223e-01
3.51350576e-01 3.17076176e-01 8.91948402e-01 -4.91564989e-01
-1.15503204e+00 -1.50953197e+00 4.50554103e-01 -3.12719584e-01
6.28615618e-01 -5.82977414e-01 -1.08648586e+00 7.15164423e-01
-2.28266437e-02 3.74120921e-01 8.18880677e-01 -6.72668368e-02
-5.45956194e-01 -1.60249532e-03 -1.27656221e+00 5.50070763e-01
1.05835724e+00 -3.36675465e-01 -7.25849390e-01 2.40214914e-01
1.13912928e+00 -2.36266762e-01 -8.87097001e-01 5.28436959e-01
4.71816540e-01 -5.40655375e-01 1.00573516e+00 -8.34896445e-01
5.82354069e-01 -1.98877633e-01 -3.71287555e-01 -1.61464787e+00
-5.32156289e-01 -2.30559915e-01 -1.49437875e-01 1.50619912e+00
1.79633513e-01 -3.49759310e-01 3.85080725e-01 4.51723576e-01
-3.20097022e-02 -6.38025522e-01 -8.11037958e-01 -9.53733385e-01
-1.77176893e-01 -3.71412560e-02 6.03330195e-01 1.06788743e+00
-3.20620656e-01 5.30852079e-01 -5.14593244e-01 -1.59610122e-01
1.02210355e+00 2.06131831e-01 7.25654542e-01 -1.18667674e+00
-3.04328680e-01 -9.38117951e-02 -7.54389167e-01 -2.77985692e-01
5.42563677e-01 -1.05996346e+00 2.45270312e-01 -1.32281911e+00
4.25469071e-01 -1.71105683e-01 -4.43972379e-01 6.67172551e-01
-2.88717240e-01 4.37207431e-01 3.33599687e-01 -8.70392993e-02
-6.52336299e-01 1.10778022e+00 1.22307563e+00 -4.32658494e-01
-1.11179743e-02 -2.43030727e-01 -6.89899683e-01 6.07798517e-01
5.77409208e-01 -4.73299354e-01 -6.08703792e-01 7.32585937e-02
-2.40229920e-01 -3.07112008e-01 6.96573198e-01 -8.88915420e-01
-8.30226094e-02 -6.55173212e-02 4.98683393e-01 -2.53978014e-01
5.68532705e-01 -7.00153947e-01 1.09389655e-01 1.93154454e-01
-5.40467083e-01 -6.78332269e-01 -1.95545107e-02 7.42512524e-01
-1.95638090e-01 -3.23523819e-01 1.09378862e+00 2.94354628e-03
-1.07848299e+00 5.15078008e-01 2.81696051e-01 3.24634224e-01
1.31583512e+00 -1.40574574e-01 -4.89664972e-01 -2.38546401e-01
-8.87373805e-01 2.09300831e-01 5.12372971e-01 6.04102194e-01
8.62964869e-01 -1.70508444e+00 -6.72478378e-01 3.55624139e-01
7.38985777e-01 1.12682007e-01 4.04670179e-01 5.41868150e-01
9.23209712e-02 5.47028147e-02 -3.70244473e-01 -5.28689981e-01
-1.12717879e+00 9.19136107e-01 4.02464124e-04 9.78349447e-02
-8.38329911e-01 9.48902607e-01 6.20534539e-01 -2.46757865e-01
1.77039638e-01 1.75280675e-01 -4.49023813e-01 1.27253488e-01
5.73927104e-01 3.23958427e-01 -2.91331351e-01 -6.91133440e-01
-2.18849167e-01 5.89509606e-01 2.95795705e-02 1.40057668e-01
1.17000067e+00 5.81436902e-02 4.34269041e-01 8.67524207e-01
1.23573291e+00 -6.33323252e-01 -1.31048930e+00 -3.27773094e-01
-3.45498532e-01 -4.57950473e-01 -2.03432515e-02 -4.86623973e-01
-1.01425552e+00 1.00345337e+00 7.21930265e-01 8.67551789e-02
9.73370552e-01 3.59277904e-01 5.94986260e-01 1.77590191e-01
4.26369011e-01 -8.09986174e-01 6.18820116e-02 1.51898205e-01
8.94160807e-01 -1.58194458e+00 -3.82102817e-01 -3.46098810e-01
-9.96189117e-01 7.27892399e-01 8.48694086e-01 -1.97836116e-01
4.87634122e-01 3.29229459e-02 -2.91209072e-01 -1.64897740e-01
-9.39608216e-01 -5.42237520e-01 4.95908976e-01 7.91441977e-01
4.15328443e-02 1.61447376e-01 9.71073508e-02 9.25038815e-01
-8.40226188e-02 -1.96428075e-01 1.82568863e-01 7.63630927e-01
-3.53932619e-01 -5.52999496e-01 -7.71020651e-02 4.64516133e-01
2.48449985e-02 -1.71042711e-01 -4.46494251e-01 8.54352415e-01
1.69078976e-01 7.91246057e-01 2.01208159e-01 -4.74639028e-01
2.57845610e-01 2.54964888e-01 4.74418879e-01 -8.61121058e-01
-2.21768960e-01 -9.50650200e-02 -2.50603735e-01 -4.44659591e-01
-3.46737355e-01 -2.67280459e-01 -1.23198986e+00 -1.29183056e-02
-3.30442607e-01 -4.63016517e-03 2.94037640e-01 8.94319296e-01
1.22606270e-01 6.30246282e-01 6.06364489e-01 -9.69154894e-01
-5.16250551e-01 -1.03472590e+00 -8.85141075e-01 8.80840778e-01
1.30416006e-01 -1.27241540e+00 -8.30033898e-01 2.27916792e-01] | [10.03982925415039, 2.484173536300659] |
51fe243b-215e-4fa9-88e5-16ec640d5776 | learning-with-noisily-labeled-class | 2211.10955 | null | https://arxiv.org/abs/2211.10955v1 | https://arxiv.org/pdf/2211.10955v1.pdf | Learning with Noisily-labeled Class-imbalanced Data | Real-world large-scale datasets are both noisily labeled and class-imbalanced. The issues seriously hurt the generalization of trained models. It is hence significant to address the simultaneous incorrect labeling and class-imbalance, i.e., the problem of learning with noisy labels on long-tailed data. Previous works develop several methods for the problem. However, they always rely on strong assumptions that are invalid or hard to be checked in practice. In this paper, to handle the problem and address the limitations of prior works, we propose a representation calibration method RCAL. Specifically, RCAL works with the representations extracted by unsupervised contrastive learning. We assume that without incorrect labeling and class imbalance, the representations of instances in each class conform to a multivariate Gaussian distribution, which is much milder and easier to be checked. Based on the assumption, we recover underlying representation distributions from polluted ones resulting from mislabeled and class-imbalanced data. Additional data points are then sampled from the recovered distributions to help generalization. Moreover, during classifier training, representation learning takes advantage of representation robustness brought by contrastive learning, which further improves the classifier performance. Experiments on multiple benchmarks justify our claims and confirm the superiority of the proposed method. | ['Weiran Huang', 'Jun Yao', 'Chun Yuan', 'Manyi Zhang'] | 2022-11-20 | null | null | null | null | ['learning-with-noisy-labels', 'learning-with-noisy-labels'] | ['computer-vision', 'natural-language-processing'] | [ 2.70791799e-01 -4.86023873e-02 -4.18505847e-01 -5.14224648e-01
-7.58516312e-01 -4.09084231e-01 1.38384044e-01 3.44604850e-01
-6.66816160e-02 9.84527588e-01 -1.58069760e-01 -3.44639309e-02
-2.42364213e-01 -9.18579519e-01 -6.51833594e-01 -9.71838176e-01
4.73433137e-01 5.17152905e-01 -7.05879927e-02 8.71679094e-03
3.12592715e-01 3.15817326e-01 -1.96342945e+00 5.09203613e-01
1.14384973e+00 1.08596778e+00 -2.67686158e-01 1.32366732e-01
-3.35961938e-01 8.98597300e-01 -9.84149456e-01 -3.25957656e-01
2.26096869e-01 -3.53555441e-01 -5.47254980e-01 3.40924978e-01
1.45775244e-01 -3.99235450e-02 3.28679830e-02 1.34270656e+00
4.76397604e-01 -6.58122525e-02 8.56656492e-01 -1.70195448e+00
-5.94103754e-01 4.76039439e-01 -9.75995958e-01 1.32909596e-01
-2.06615077e-03 -2.12868974e-01 7.80061185e-01 -7.79878438e-01
2.08283514e-01 1.16012204e+00 7.17248321e-01 3.93364131e-01
-1.09219635e+00 -1.15351522e+00 3.60258967e-01 2.26241440e-01
-1.57052600e+00 -2.50774205e-01 9.31612194e-01 -3.74787837e-01
1.17806107e-01 2.22254366e-01 2.25599706e-01 1.06037617e+00
-1.51969090e-01 9.10675883e-01 1.42656732e+00 -3.86664987e-01
2.64008462e-01 5.29618621e-01 4.19245422e-01 1.06212728e-01
8.19307923e-01 -1.15532748e-01 -2.64141530e-01 -2.42061555e-01
1.47947818e-01 3.66386622e-01 -3.68406832e-01 -3.17201227e-01
-7.68107176e-01 9.14757609e-01 3.52043092e-01 2.66299278e-01
-4.16999549e-01 -3.33631605e-01 6.06825888e-01 2.16233283e-01
7.48445570e-01 1.45844057e-01 -5.11638284e-01 2.07168981e-01
-8.55985940e-01 2.44340807e-01 4.06629264e-01 9.12761211e-01
9.19490516e-01 -4.85259965e-02 -2.63409484e-02 9.01464820e-01
1.08503766e-01 6.33624017e-01 7.34565198e-01 -4.07489836e-01
4.94183153e-01 8.24071169e-01 1.54287331e-02 -1.18424678e+00
-4.74084228e-01 -9.60702121e-01 -1.20660543e+00 2.79180836e-02
5.10612488e-01 4.23083594e-03 -7.19232082e-01 1.65020823e+00
5.23409069e-01 2.49926254e-01 2.17846766e-01 6.83050036e-01
6.21261239e-01 5.65097570e-01 1.18867293e-01 -4.50237334e-01
1.08998334e+00 -6.04266226e-01 -1.00587928e+00 -1.95850972e-02
8.47483695e-01 -6.18600547e-01 1.04318631e+00 7.58732140e-01
-6.44076705e-01 -5.78968346e-01 -1.06862187e+00 4.14367795e-01
-2.78185427e-01 9.65520665e-02 4.53470886e-01 8.57780933e-01
-2.54456699e-01 4.66197878e-01 -4.43974793e-01 1.44902959e-01
6.62461162e-01 1.57565013e-01 -3.57533455e-01 -3.06023896e-01
-1.07830191e+00 6.63585365e-01 6.22309983e-01 1.13653429e-01
-3.90792429e-01 -6.03760600e-01 -7.05871165e-01 -3.21433470e-02
3.66812527e-01 -8.55869427e-02 1.19302154e+00 -1.28899240e+00
-9.96882677e-01 6.29375756e-01 1.92690128e-03 -3.41969550e-01
5.83877087e-01 -1.13710433e-01 -6.17395461e-01 -5.27242459e-02
9.88133475e-02 4.86871898e-02 7.30654061e-01 -1.69730198e+00
-7.09063292e-01 -5.35592318e-01 -2.87607074e-01 -1.12564348e-01
-5.18569469e-01 -3.30068558e-01 7.31306151e-02 -7.22543299e-01
4.37625170e-01 -8.11437607e-01 -2.11171862e-02 -3.18181336e-01
-5.17958641e-01 -2.36554742e-01 8.81595731e-01 -4.79758978e-01
1.24115956e+00 -2.28579831e+00 -4.56236064e-01 4.26839918e-01
1.35743588e-01 4.20406073e-01 6.42038360e-02 1.16077177e-01
-4.12950754e-01 2.42267117e-01 -3.08088213e-01 -1.56359553e-01
-8.14280733e-02 4.21540588e-01 -6.17838323e-01 7.32399762e-01
1.67772979e-01 4.38374519e-01 -8.97341192e-01 -4.78427947e-01
-7.88422525e-02 2.70178050e-01 -3.31206620e-01 3.06829810e-01
-1.03600277e-02 5.34688652e-01 -3.64725351e-01 6.33471072e-01
1.30140197e+00 -2.79709190e-01 2.18731523e-01 -3.85247618e-01
3.68933767e-01 7.29688331e-02 -1.62160099e+00 9.91293073e-01
-3.92028928e-01 9.22097638e-02 -3.48513097e-01 -1.60089087e+00
1.27955806e+00 2.33316645e-01 3.10134470e-01 -7.20615804e-01
2.18413115e-01 4.42216992e-01 -3.69304866e-02 -4.77217942e-01
2.86217123e-01 -5.65817714e-01 -1.56371798e-02 4.57008690e-01
-1.29418731e-01 9.48159024e-03 -9.68512893e-03 -8.12480971e-02
5.46801865e-01 -8.22126791e-02 4.01404649e-01 -5.87730892e-02
3.97442430e-01 -2.56950051e-01 1.00314760e+00 6.26632810e-01
-1.72755167e-01 7.21233130e-01 5.80443621e-01 -3.23338985e-01
-7.69835293e-01 -8.52284610e-01 -4.76713985e-01 8.33350003e-01
1.47381261e-01 -1.98223755e-01 -5.52122176e-01 -9.89999533e-01
3.15812114e-03 7.60753393e-01 -6.41879141e-01 -4.73369509e-01
-3.71256232e-01 -1.22146487e+00 3.51133615e-01 5.61536491e-01
3.67756337e-01 -8.06358874e-01 -2.45883808e-01 1.19900376e-01
-3.56015682e-01 -8.24530900e-01 1.69947505e-01 2.87667722e-01
-8.99848521e-01 -1.21302724e+00 -4.86947685e-01 -5.87498188e-01
9.08100843e-01 4.53173339e-01 1.10955644e+00 2.65632302e-01
-4.30433601e-02 -2.33279452e-01 -5.90715706e-01 -6.02271974e-01
-4.33540046e-01 2.12612208e-02 1.56501189e-01 3.60871404e-01
4.78627354e-01 -4.99796569e-01 -4.19405341e-01 5.52147388e-01
-1.23207271e+00 -2.94704348e-01 5.45775294e-01 9.90116775e-01
7.49616444e-01 6.52530789e-01 1.16758084e+00 -1.33083892e+00
3.92413437e-01 -8.35239053e-01 -4.52768415e-01 2.41312906e-01
-6.69683278e-01 1.64224692e-02 8.37570190e-01 -5.94034910e-01
-1.05147445e+00 -1.02258295e-01 -4.29736935e-02 -3.94666702e-01
-2.38404080e-01 4.33452338e-01 -5.61133564e-01 2.60893881e-01
7.85445213e-01 1.08340271e-01 -6.74147159e-02 -4.74177390e-01
1.36821374e-01 1.03651249e+00 3.21429074e-01 -6.85840726e-01
8.29180598e-01 4.96895254e-01 -1.23342536e-01 -5.11451602e-01
-1.31650293e+00 -3.65364879e-01 -3.87894183e-01 3.94686349e-02
2.48971611e-01 -8.78035545e-01 -3.98538917e-01 5.31837642e-01
-8.31663191e-01 1.01226106e-01 -4.53158200e-01 4.55922484e-01
-2.44286463e-01 4.05397594e-01 -2.79419690e-01 -1.13679123e+00
-7.35040456e-02 -8.99987221e-01 7.98623264e-01 2.87032604e-01
-6.27956316e-02 -8.28133941e-01 -9.06557366e-02 3.45361114e-01
2.94828951e-01 4.12625104e-01 9.75298584e-01 -1.01640594e+00
-1.66247770e-01 -6.05254471e-01 -2.28688627e-01 7.36844957e-01
3.28935534e-01 5.70352599e-02 -1.37082100e+00 -4.53437567e-01
1.75322250e-01 -5.17343402e-01 6.81321561e-01 9.75821316e-02
1.81556046e+00 -2.48138413e-01 -1.25459269e-01 3.82948905e-01
1.23144746e+00 2.58487966e-02 6.33947134e-01 3.01445931e-01
4.46331143e-01 7.67840862e-01 9.71527100e-01 6.46163404e-01
2.27923647e-01 3.85707587e-01 5.21826506e-01 -9.19480175e-02
1.20228864e-01 -3.29734445e-01 -3.12723778e-02 9.20058787e-01
1.90529689e-01 -3.05112898e-01 -7.57470727e-01 4.91956919e-01
-1.77320707e+00 -7.11877227e-01 -3.32353771e-01 2.44738078e+00
1.01004326e+00 3.13848257e-01 -2.71325950e-02 8.49598408e-01
1.08568406e+00 -1.47437572e-01 -5.78025520e-01 -1.13790324e-02
-3.51224691e-01 1.17592998e-01 3.24476779e-01 -3.39420699e-03
-1.10879612e+00 3.38916570e-01 5.62459087e+00 1.05214989e+00
-1.14004409e+00 5.25895171e-02 9.94547188e-01 1.44318029e-01
-4.26467597e-01 -1.78622648e-01 -7.68925667e-01 6.19629681e-01
7.41577387e-01 -1.14173979e-01 -2.39042267e-02 9.30991828e-01
-1.37610137e-01 3.93615337e-03 -9.23080802e-01 1.08668458e+00
1.14163637e-01 -8.24526191e-01 -2.68805679e-02 -3.75150964e-02
9.51426685e-01 -4.43211913e-01 2.93249458e-01 6.23280227e-01
1.79476708e-01 -9.94176745e-01 6.48371935e-01 4.01386708e-01
6.31899238e-01 -1.03442013e+00 1.25165951e+00 6.88837647e-01
-8.52037668e-01 -3.17434400e-01 -6.12257481e-01 -1.44596979e-01
-2.66279668e-01 1.19106579e+00 -6.27522230e-01 7.63688862e-01
5.16150117e-01 6.71475708e-01 -6.05127811e-01 1.19567692e+00
-3.66873890e-01 8.21935654e-01 -9.42841992e-02 3.01920176e-01
-2.48355433e-01 2.13765297e-02 -1.29159279e-02 9.57959294e-01
3.42585325e-01 -1.64587528e-01 3.55811775e-01 4.70070362e-01
-1.82424188e-01 3.65540743e-01 -5.14665186e-01 2.60617137e-01
6.71334803e-01 1.19817185e+00 -7.11993515e-01 -5.34339666e-01
-1.80185884e-01 4.63518202e-01 3.91661197e-01 2.89544672e-01
-8.48070800e-01 -2.83788711e-01 2.45786548e-01 1.41428828e-01
1.08771451e-01 3.80294681e-01 -5.93553722e-01 -1.24084747e+00
2.99628913e-01 -1.14676845e+00 5.93377173e-01 -3.24816167e-01
-1.87021601e+00 5.03646851e-01 -1.86157487e-02 -1.59800518e+00
-6.81745484e-02 -3.94932389e-01 -3.82438689e-01 6.54509902e-01
-1.83887899e+00 -8.36988747e-01 -5.38199306e-01 4.63860095e-01
2.21737579e-01 1.17367715e-01 7.45487034e-01 5.93490720e-01
-7.25844622e-01 9.15191650e-01 1.51499227e-01 3.77897322e-02
8.95586967e-01 -1.07845008e+00 -2.06268489e-01 5.48234046e-01
-1.47583067e-01 4.13319975e-01 5.76020896e-01 -5.75802684e-01
-8.29924762e-01 -1.47958434e+00 6.08371198e-01 -1.39813229e-01
3.95335376e-01 -2.58836716e-01 -1.46608889e+00 5.11302650e-01
-2.54283041e-01 4.94801402e-01 1.00129330e+00 1.57899188e-03
-7.13022053e-01 -4.52960938e-01 -1.47194290e+00 5.94436117e-02
6.63093030e-01 -3.90886515e-01 -6.63764477e-01 5.68990290e-01
3.80283475e-01 -1.10701114e-01 -5.90921879e-01 6.18656099e-01
2.36429080e-01 -8.83494794e-01 7.63861835e-01 -6.85516834e-01
2.93684393e-01 -5.12435377e-01 -3.90450329e-01 -1.36638916e+00
-1.17567584e-01 5.24322353e-02 2.95873471e-02 1.49243808e+00
2.48331279e-01 -8.40110421e-01 6.06180429e-01 2.38363996e-01
2.76335534e-02 -6.30190074e-01 -8.00220370e-01 -9.82117414e-01
2.99273372e-01 -3.57056916e-01 8.99373412e-01 1.31269622e+00
-2.28583008e-01 1.47941709e-01 -4.57832336e-01 3.14038932e-01
7.38351226e-01 3.47157747e-01 8.37043405e-01 -1.59587550e+00
-1.10158071e-01 -1.44321069e-01 -4.08605367e-01 -6.96507096e-01
3.17274302e-01 -8.41481090e-01 2.23099858e-01 -1.21802282e+00
3.00390810e-01 -8.90329242e-01 -6.11732125e-01 4.44725215e-01
-4.68126148e-01 4.56701338e-01 -4.84641753e-02 3.29905361e-01
-6.48757637e-01 7.02863514e-01 1.12095845e+00 -1.25435501e-01
9.43632945e-02 2.29536518e-01 -9.87488508e-01 8.98390770e-01
9.80792463e-01 -8.04471076e-01 -5.18091977e-01 -1.54339718e-02
2.56296515e-01 -2.73612112e-01 1.34297520e-01 -9.56302881e-01
-1.48716852e-01 -1.11823343e-01 5.17583728e-01 -7.04515159e-01
8.81023929e-02 -9.37578976e-01 -4.48125377e-02 4.50620115e-01
-3.08099985e-01 -1.58059925e-01 -8.04800726e-03 5.71572006e-01
-3.98787946e-01 -3.74096066e-01 1.04513013e+00 -4.21704585e-03
-1.60885736e-01 1.43034890e-01 -1.07309140e-01 3.50713521e-01
1.07782638e+00 1.19449059e-02 -4.55639213e-01 -3.59920830e-01
-4.97227758e-01 3.86958681e-02 3.45825374e-01 1.17327757e-01
5.27148485e-01 -1.53991175e+00 -8.26732814e-01 4.13909227e-01
4.34171259e-01 2.89039016e-01 3.68268430e-01 7.69561529e-01
-5.08332551e-02 -9.80021358e-02 -4.36205119e-02 -7.04861939e-01
-9.46985841e-01 8.90471697e-01 1.67826414e-01 -2.13026971e-01
-3.59834462e-01 5.00407279e-01 2.79814988e-01 -7.48289704e-01
2.05986932e-01 -2.07841471e-01 -4.18975502e-01 4.24831361e-01
7.98927724e-01 3.35450798e-01 3.77678394e-01 -5.69418252e-01
-1.96136564e-01 4.77661669e-01 -2.28834108e-01 4.76084709e-01
1.27256393e+00 -2.11139675e-02 -1.39965937e-01 6.29072130e-01
1.20756149e+00 6.51677400e-02 -9.74558353e-01 -4.66931045e-01
9.99197885e-02 -7.86364734e-01 -1.60872698e-01 -6.25211895e-01
-1.25328672e+00 1.02939236e+00 7.27613091e-01 4.60541427e-01
1.22893310e+00 -3.08731198e-01 5.49798608e-01 1.85642302e-01
2.57921517e-01 -1.10806859e+00 1.32216260e-01 2.16217071e-01
6.33680761e-01 -1.28923154e+00 6.61334489e-04 -5.51795304e-01
-5.51792562e-01 1.01794326e+00 6.82734370e-01 -1.02949508e-01
6.04229927e-01 1.64357141e-01 2.71762937e-01 1.22811258e-01
-4.74523246e-01 6.00074306e-02 7.53761604e-02 7.09693849e-01
2.99906462e-01 1.45146206e-01 -2.34977901e-01 9.41184580e-01
-9.65991393e-02 -1.09065361e-01 6.48114502e-01 9.05750930e-01
-4.66155946e-01 -1.02551937e+00 -7.82752931e-01 4.99155909e-01
-3.61727595e-01 2.06711113e-01 5.92693724e-02 6.03667557e-01
4.47675735e-01 1.10711110e+00 1.03423912e-02 -3.55008841e-01
4.12461251e-01 1.54610947e-01 7.76800215e-02 -5.40857673e-01
-2.83172816e-01 -3.74545008e-02 -2.16100216e-01 -2.10890919e-01
-4.39899147e-01 -5.00902712e-01 -1.24769676e+00 -1.43221617e-01
-7.09553659e-01 3.66368294e-01 6.46793067e-01 8.75897408e-01
1.72273695e-01 6.58815324e-01 1.01917922e+00 -5.12662351e-01
-1.10290289e+00 -9.85968471e-01 -9.13587511e-01 8.01149428e-01
3.42306703e-01 -1.02983308e+00 -7.24800229e-01 -2.28507414e-01] | [9.202507019042969, 3.9029624462127686] |
8597ce8a-c5af-406c-9472-2265ef4cfb9a | semantic-segmentation-on-3d-point-clouds-with | 2307.01489 | null | https://arxiv.org/abs/2307.01489v1 | https://arxiv.org/pdf/2307.01489v1.pdf | Semantic Segmentation on 3D Point Clouds with High Density Variations | LiDAR scanning for surveying applications acquire measurements over wide areas and long distances, which produces large-scale 3D point clouds with significant local density variations. While existing 3D semantic segmentation models conduct downsampling and upsampling to build robustness against varying point densities, they are less effective under the large local density variations characteristic of point clouds from surveying applications. To alleviate this weakness, we propose a novel architecture called HDVNet that contains a nested set of encoder-decoder pathways, each handling a specific point density range. Limiting the interconnections between the feature maps enables HDVNet to gauge the reliability of each feature based on the density of a point, e.g., downweighting high density features not existing in low density objects. By effectively handling input density variations, HDVNet outperforms state-of-the-art models in segmentation accuracy on real point clouds with inconsistent density, using just over half the weights. | ['Tat-Jun Chin', 'Ian Reid', 'Simon Ratcliffe', 'Luke Haub', 'Ryan Faulkner'] | 2023-07-04 | null | null | null | null | ['3d-semantic-segmentation'] | ['computer-vision'] | [-2.07177904e-02 -3.65091935e-02 -1.87594607e-01 -6.58590972e-01
-6.92278147e-01 -4.09879386e-01 4.99290019e-01 1.93440437e-01
-2.33021945e-01 6.19918406e-01 -1.33024529e-01 -1.04449406e-01
-5.63514307e-02 -1.22089326e+00 -9.84847963e-01 -3.21998686e-01
1.58594772e-02 1.27982044e+00 7.95504630e-01 5.23135588e-02
2.45539173e-01 8.52643907e-01 -1.91783655e+00 -1.21496335e-01
1.15395343e+00 8.64017487e-01 3.58494967e-01 4.60534334e-01
-7.31892645e-01 -1.47632957e-01 -4.93569762e-01 2.53984332e-02
3.39779705e-01 6.15162067e-02 -5.30540168e-01 1.79849759e-01
1.02381599e+00 -5.08472681e-01 -9.33810174e-02 1.01249754e+00
3.05352688e-01 -9.36337039e-02 7.46792078e-01 -1.16005862e+00
-5.03948987e-01 3.30602229e-01 -7.57087409e-01 3.80094469e-01
-4.74157482e-02 2.15763375e-01 9.14574623e-01 -1.02707958e+00
4.10144508e-01 1.46594059e+00 8.71990025e-01 2.46247724e-01
-1.31681108e+00 -8.26280415e-01 4.27410990e-01 -1.68989435e-01
-1.73748326e+00 -1.97745875e-01 4.45896447e-01 -3.89853239e-01
1.21496665e+00 -7.01608881e-02 8.75989616e-01 6.97440982e-01
3.92706804e-02 5.22212625e-01 6.49671137e-01 1.78092584e-01
6.08130753e-01 3.88285238e-03 -7.88418353e-02 2.93272525e-01
6.61326766e-01 -8.34614784e-02 -6.15136206e-01 -2.58865982e-01
9.66299057e-01 -2.35344451e-02 1.12090819e-01 -7.24547029e-01
-9.62108910e-01 7.79531777e-01 6.66160762e-01 -5.24365641e-02
-2.16376692e-01 2.48853028e-01 2.02957541e-01 4.23228964e-02
7.09596694e-01 7.73399696e-02 -6.81739032e-01 -1.27732471e-01
-1.21775937e+00 5.29159606e-01 5.18737555e-01 1.59536421e+00
1.29716003e+00 1.15974611e-02 2.60041296e-01 8.20158422e-01
5.10137200e-01 9.39632237e-01 4.30318806e-03 -9.75679755e-01
7.09975362e-01 8.00682902e-01 1.16505243e-01 -4.84896898e-01
-3.42806607e-01 -3.50073427e-01 -5.26742876e-01 2.31290668e-01
2.67833799e-01 1.24273211e-01 -1.30716431e+00 1.48675656e+00
4.84561741e-01 2.82580465e-01 -2.64123797e-01 8.60580444e-01
4.57694471e-01 4.26124930e-01 1.07466899e-01 4.10934418e-01
1.11880612e+00 -2.17149809e-01 -1.83854163e-01 -7.39387453e-01
3.77241820e-01 -3.74966204e-01 1.08169210e+00 2.06741765e-02
-8.58083069e-01 -3.49833190e-01 -1.00033844e+00 -1.45437509e-01
-3.24436724e-01 -5.58346152e-01 6.22519076e-01 6.97223008e-01
-1.06434751e+00 6.41562700e-01 -8.70589316e-01 -3.60047251e-01
9.52830732e-01 3.64247292e-01 -1.37174368e-01 -2.35691667e-01
-7.79339790e-01 8.72272074e-01 3.06487270e-02 -2.04302654e-01
-5.70464253e-01 -1.28671741e+00 -1.04352307e+00 3.55737843e-02
4.44187261e-02 -7.48106182e-01 1.05620086e+00 -1.17915519e-01
-8.60141337e-01 6.92065060e-01 -2.66174257e-01 -3.68113071e-01
5.67516088e-01 -2.88505107e-01 -1.13619246e-01 -1.29732504e-01
4.60820287e-01 1.10132408e+00 6.71642065e-01 -1.36333776e+00
-7.90075600e-01 -9.08288956e-01 -4.66785699e-01 4.70646054e-01
2.61590391e-01 -6.30797565e-01 -4.32462424e-01 -6.39066026e-02
5.64723313e-01 -7.54823267e-01 -4.45925593e-01 3.06327194e-01
-1.68340623e-01 -1.16253095e-02 1.24993551e+00 8.49201679e-02
4.55155492e-01 -2.22226620e+00 -1.52170762e-01 3.45625818e-01
4.63349409e-02 -7.52671883e-02 -6.59587607e-02 1.92368001e-01
4.86704171e-01 1.46909446e-01 -6.26523316e-01 -3.39625388e-01
-4.69369851e-02 6.34312272e-01 -2.30954140e-01 6.11395359e-01
5.80607355e-01 7.05584109e-01 -1.01432395e+00 -5.67214489e-01
6.85849190e-01 7.08672583e-01 -6.42173052e-01 -1.38624147e-01
-4.31601048e-01 2.52518713e-01 -5.95977128e-01 9.72309291e-01
1.26941383e+00 -1.44604191e-01 -1.55637264e-01 4.22068872e-02
-8.14152956e-02 2.61715353e-01 -1.37608325e+00 1.81541896e+00
-4.22206044e-01 4.40744519e-01 1.51461348e-01 -3.82946879e-01
1.30788112e+00 -3.15925270e-01 5.85227907e-01 -4.87223774e-01
-6.22673854e-02 3.23037922e-01 -2.82578170e-01 -5.24028838e-02
8.73421669e-01 -3.90511662e-01 -5.96956871e-02 1.05802320e-01
1.71116963e-01 -1.04568911e+00 -1.47119716e-01 1.83017716e-01
1.11237049e+00 -2.22425722e-02 -2.96899546e-02 -2.41934970e-01
-8.10748935e-02 1.41083494e-01 3.86345565e-01 7.50122309e-01
-1.29680455e-01 1.00823236e+00 2.50493944e-01 -1.28510833e-01
-1.33177423e+00 -1.54095423e+00 -8.66571367e-01 5.17283857e-01
5.62882364e-01 -1.19726218e-01 -4.40100044e-01 -5.27654648e-01
7.29175329e-01 7.78993487e-01 -4.51939672e-01 -1.90506708e-02
-2.43655428e-01 -7.71099865e-01 4.72455084e-01 7.54756570e-01
5.27719498e-01 -6.39123678e-01 -7.16904104e-01 1.77171558e-01
2.20854163e-01 -1.28664601e+00 -9.68546420e-02 3.48507285e-01
-1.24064624e+00 -9.20725942e-01 -4.61099893e-01 -3.15660924e-01
5.69959641e-01 3.82083595e-01 1.46360838e+00 -1.08519711e-01
-2.99460381e-01 2.30393350e-01 -2.32392058e-01 -6.36606693e-01
-1.58292707e-02 4.70821023e-01 -9.99352150e-03 -6.19686127e-01
7.78323591e-01 -8.46309245e-01 -4.17129666e-01 3.65898579e-01
-7.78366089e-01 -2.88870096e-01 4.14540827e-01 4.93692368e-01
8.98316324e-01 -2.31841598e-02 4.02997553e-01 -8.28278065e-01
1.65092036e-01 -8.36670220e-01 -7.58379936e-01 -2.85144836e-01
-5.37782371e-01 1.05244834e-02 1.57179147e-01 -1.79408401e-01
-6.46695077e-01 1.73169360e-01 -1.86762646e-01 -6.69833243e-01
-3.55483919e-01 -6.55436330e-03 -2.56195486e-01 -1.72122329e-01
6.33353829e-01 -9.31116790e-02 -1.09053053e-01 -5.33352971e-01
5.43009937e-01 5.50939798e-01 4.22175527e-01 -5.57159603e-01
8.15431297e-01 6.80717528e-01 4.38408554e-02 -8.76524031e-01
-5.75881064e-01 -6.65238917e-01 -8.00560772e-01 -1.36202484e-01
5.32432258e-01 -1.15385211e+00 -1.69814974e-01 2.70794988e-01
-8.94676387e-01 -2.43372753e-01 -9.18978095e-01 3.56186658e-01
-5.53605795e-01 4.23465902e-03 -3.18374664e-01 -8.76205027e-01
1.02352034e-02 -1.05497694e+00 1.67486918e+00 1.62386626e-01
-1.92038924e-01 -6.88835680e-01 1.27781751e-02 -7.31554106e-02
2.63633728e-01 9.67227817e-02 7.87614465e-01 -3.51668119e-01
-8.34187806e-01 -2.18050912e-01 -5.45139194e-01 1.70551568e-01
9.73195806e-02 1.15834914e-01 -1.05673504e+00 -7.29783578e-03
-1.90175086e-01 -1.26575872e-01 8.25778306e-01 5.59360743e-01
1.08222079e+00 3.10530603e-01 -5.92524290e-01 7.35923409e-01
1.51230979e+00 -2.60170817e-01 6.99318886e-01 8.30257386e-02
7.03818738e-01 4.81704712e-01 5.66952229e-01 4.40588981e-01
5.52184343e-01 3.79289389e-01 8.87942851e-01 -7.16722235e-02
-1.99673027e-02 -5.10614097e-01 -2.22482547e-01 3.25403810e-01
3.59036028e-01 -1.08302645e-01 -1.08426690e+00 8.65151942e-01
-1.56550527e+00 -5.84162414e-01 -3.09293628e-01 2.21114874e+00
5.25038719e-01 2.90792137e-01 -6.58074468e-02 -8.77833441e-02
5.95176637e-01 2.81471163e-01 -9.49987292e-01 -9.27091613e-02
-2.37476960e-01 3.30202103e-01 1.07852983e+00 5.80767095e-01
-6.96766794e-01 1.09297764e+00 7.19185019e+00 6.12608194e-01
-8.40426445e-01 -1.38566941e-01 3.36081058e-01 -3.17858011e-01
-6.87596023e-01 -6.64122477e-02 -1.02592504e+00 6.03504837e-01
8.41776073e-01 1.57830045e-01 3.83462310e-02 1.05677474e+00
-6.29108176e-02 -4.42299247e-01 -8.91897738e-01 7.62842655e-01
-2.96455681e-01 -1.36600745e+00 1.72147900e-01 3.18343967e-01
5.95031977e-01 8.82642269e-01 -4.07693870e-02 2.79663820e-02
7.04163969e-01 -1.00948226e+00 7.67968178e-01 2.15849176e-01
9.69073117e-01 -7.69591391e-01 7.07261980e-01 4.91323531e-01
-1.19705415e+00 1.95261464e-01 -7.57511377e-01 2.29941290e-02
4.88933712e-01 1.08346772e+00 -8.62087011e-01 3.36581886e-01
9.67945457e-01 6.20709181e-01 -4.73350167e-01 1.06914818e+00
1.01951204e-01 3.42557937e-01 -9.52703953e-01 2.60631919e-01
1.82343632e-01 -2.12117344e-01 5.33204079e-01 1.17896247e+00
6.07791245e-01 -1.18455529e-01 1.08088544e-02 1.10563934e+00
2.29820143e-03 -3.00497204e-01 -9.27091956e-01 4.29009259e-01
1.06398129e+00 8.78815532e-01 -7.81795144e-01 -2.77182877e-01
-3.12153071e-01 4.37088698e-01 3.22667778e-01 8.15287679e-02
-6.83619738e-01 -2.02453434e-02 1.09247458e+00 5.76866627e-01
4.61384296e-01 -5.44957757e-01 -8.96248698e-01 -9.21321034e-01
-4.59055863e-02 -1.88066527e-01 1.61610413e-02 -8.64925683e-01
-1.44923389e+00 3.64177227e-01 2.70001203e-01 -1.32483673e+00
-1.41398618e-02 -3.72127295e-01 -3.13842654e-01 9.54388380e-01
-1.64922798e+00 -9.78916764e-01 -5.69500446e-01 4.40033734e-01
6.68772042e-01 2.45525301e-01 5.33111215e-01 3.16837698e-01
-4.35064435e-02 8.05340633e-02 -1.84044510e-01 -3.33128154e-01
3.36833864e-01 -1.36102855e+00 9.55913782e-01 5.84990680e-01
8.18437412e-02 3.15500766e-01 7.15102732e-01 -9.94722903e-01
-9.39953327e-01 -1.27791703e+00 5.45269430e-01 -6.72586918e-01
4.21179712e-01 -5.69191158e-01 -1.27535069e+00 5.28467476e-01
-4.75733787e-01 3.18597406e-01 3.44421804e-01 2.78945595e-01
-4.01850969e-01 2.57056281e-02 -1.66578197e+00 2.42326885e-01
1.41475809e+00 -4.72561896e-01 -3.99356365e-01 7.30735064e-02
6.16227806e-01 -6.38126254e-01 -8.74602377e-01 7.07900584e-01
3.02704722e-01 -1.06180990e+00 1.12037206e+00 -1.11205749e-01
3.20485353e-01 -4.53179926e-01 -4.63359475e-01 -1.32396436e+00
-3.18083793e-01 1.88352495e-01 -5.33079989e-02 1.08333743e+00
4.49495941e-01 -4.68309015e-01 1.16188729e+00 5.66732705e-01
-2.92059183e-01 -4.61366862e-01 -1.33917379e+00 -7.33200669e-01
1.97970852e-01 -7.62830675e-01 1.22332859e+00 7.30889142e-01
-3.71647716e-01 1.58230484e-01 1.99212849e-01 4.51907575e-01
7.60650218e-01 -1.74978733e-01 9.28519011e-01 -1.56514072e+00
4.02943134e-01 -3.32704425e-01 -6.83245599e-01 -1.20370436e+00
-1.23177711e-02 -6.18398845e-01 3.27759802e-01 -1.59942460e+00
-2.11548358e-01 -1.00169361e+00 1.94578260e-01 1.68905020e-01
-2.47982349e-02 3.74681681e-01 -6.48963749e-02 2.32231185e-01
-2.50214905e-01 6.97763264e-01 1.08567965e+00 -1.00114219e-01
-8.84391591e-02 -4.48546372e-03 -5.26453733e-01 7.30951965e-01
6.72743380e-01 -5.19830287e-01 -4.53397691e-01 -8.30756962e-01
2.15085410e-02 -3.04728091e-01 4.35819000e-01 -1.31461012e+00
5.09402752e-02 -3.92109990e-01 7.06837535e-01 -1.06421566e+00
5.97850084e-01 -1.02795851e+00 1.80303276e-01 -8.54613166e-03
1.35664180e-01 3.52185406e-02 3.69395465e-01 7.04333127e-01
-1.48888882e-02 -1.07565923e-02 9.07357752e-01 -2.87945986e-01
-7.32010484e-01 6.06415629e-01 -8.11801255e-02 8.93731341e-02
8.30421567e-01 -6.58518434e-01 -1.86512977e-01 -2.04885155e-02
-2.86268115e-01 3.74389172e-01 9.49926257e-01 4.03720617e-01
7.10184097e-01 -1.05511439e+00 -5.49465418e-01 6.10883892e-01
2.59635121e-01 1.05608392e+00 3.29816043e-01 2.55228311e-01
-5.47572970e-01 -2.53712218e-02 -1.00443818e-01 -1.27637899e+00
-7.25903690e-01 1.90344527e-02 3.93685251e-01 2.58215785e-01
-9.86639440e-01 1.04674852e+00 1.62016470e-02 -6.93095803e-01
-1.36103004e-01 -6.65157437e-01 2.96926945e-01 1.09146927e-02
1.13119096e-01 2.16989890e-01 2.68518180e-01 -6.11288309e-01
-5.68379283e-01 7.92311490e-01 7.47429207e-02 -8.25890601e-02
1.17142844e+00 -2.16053858e-01 2.70285487e-01 6.16813540e-01
8.28801274e-01 -1.69466570e-01 -1.79638922e+00 -3.14986557e-01
-1.02472408e-02 -1.04745531e+00 2.48100892e-01 -3.95497620e-01
-1.03929818e+00 8.09209645e-01 4.99956071e-01 1.18750893e-01
5.94892144e-01 4.17890757e-01 7.86265790e-01 -3.36543005e-03
8.89775753e-01 -9.92766678e-01 -4.75114882e-01 6.64806902e-01
4.65912014e-01 -1.30137563e+00 3.18944156e-01 -5.72773278e-01
-6.04503870e-01 6.71468794e-01 7.24443078e-01 -3.88795078e-01
7.52918005e-01 7.84714997e-01 -8.02490860e-02 -4.78677243e-01
-5.29487371e-01 -3.32128137e-01 -3.69341932e-02 1.08255589e+00
4.81625944e-02 2.06046283e-01 4.00556266e-01 -6.34262338e-02
-5.34370303e-01 -7.80485198e-02 2.00471848e-01 7.84538865e-01
-8.34477901e-01 -6.85153902e-01 -4.12835181e-01 8.83592725e-01
2.01728612e-01 4.02218029e-02 -2.09870771e-01 1.01661980e+00
2.19110087e-01 5.63651264e-01 9.44466352e-01 -4.21440989e-01
5.57356298e-01 -2.45051548e-01 4.10604149e-01 -8.46751809e-01
-2.19246566e-01 -2.06484795e-01 -1.06103048e-01 -5.72931290e-01
-2.08200827e-01 -9.42261875e-01 -1.42476547e+00 -4.99003351e-01
-3.77566487e-01 -7.76347741e-02 8.09028506e-01 7.75846779e-01
3.60614955e-01 4.00344491e-01 3.05248231e-01 -1.23100400e+00
-4.28906739e-01 -9.37969744e-01 -9.48580503e-01 1.96810350e-01
4.57077384e-01 -1.02150249e+00 -4.34421003e-01 -4.70744669e-01] | [8.026556968688965, -2.904064416885376] |
5aa6cd86-7517-4489-ad8c-f5c2bad284a5 | embedding-strategies-for-specialized-domains | null | null | https://aclanthology.org/P19-2041 | https://aclanthology.org/P19-2041.pdf | Embedding Strategies for Specialized Domains: Application to Clinical Entity Recognition | Using pre-trained word embeddings in conjunction with Deep Learning models has become the {``}de facto{''} approach in Natural Language Processing (NLP). While this usually yields satisfactory results, off-the-shelf word embeddings tend to perform poorly on texts from specialized domains such as clinical reports. Moreover, training specialized word representations from scratch is often either impossible or ineffective due to the lack of large enough in-domain data. In this work, we focus on the clinical domain for which we study embedding strategies that rely on general-domain resources only. We show that by combining off-the-shelf contextual embeddings (ELMo) with static word2vec embeddings trained on a small in-domain corpus built from the task data, we manage to reach and sometimes outperform representations learned from a large corpus in the medical domain. | ['Pierre Zweigenbaum', 'Olivier Ferret', 'Hicham El Boukkouri', 'Thomas Lavergne'] | 2019-07-01 | null | null | null | acl-2019-7 | ['clinical-concept-extraction'] | ['medical'] | [ 1.12922853e-02 2.61436641e-01 -7.94738308e-02 -1.72000542e-01
-9.32896852e-01 -3.59603584e-01 4.20582026e-01 9.95056212e-01
-1.10976636e+00 6.48074210e-01 5.38899004e-01 -4.85586971e-01
-8.92632678e-02 -6.39790714e-01 -1.54191405e-01 -4.38647360e-01
-5.61785288e-02 7.38880754e-01 1.25250831e-01 -3.52784574e-01
-7.49385804e-02 2.57746369e-01 -9.21174884e-01 1.88255623e-01
5.53192317e-01 6.03500247e-01 2.11547136e-01 7.42839098e-01
-6.10772550e-01 2.29142040e-01 -5.22462904e-01 -5.16980469e-01
1.23584583e-01 -1.41713873e-01 -7.36393332e-01 -1.39155835e-01
-6.06891997e-02 -8.46690387e-02 -2.95344442e-01 7.67902732e-01
8.31343651e-01 1.53872222e-01 4.43121046e-01 -4.84282613e-01
-9.12832677e-01 2.90077090e-01 -1.33455873e-01 4.41384345e-01
2.35530257e-01 2.83634737e-02 1.32129121e+00 -8.25756788e-01
1.03992748e+00 7.52733707e-01 8.55570912e-01 7.62716353e-01
-1.37612510e+00 -1.99854448e-01 1.24729156e-01 6.98118359e-02
-1.24644625e+00 -1.20014377e-01 6.17040634e-01 -3.09406102e-01
1.55771518e+00 -1.28378779e-01 3.16767782e-01 1.39297259e+00
3.39836657e-01 6.01822317e-01 6.52895153e-01 -6.33548379e-01
2.41058812e-01 3.63942474e-01 3.25648934e-01 5.08969605e-01
6.18628502e-01 -1.36557549e-01 -6.02823235e-02 -4.59933370e-01
4.84108269e-01 9.78551954e-02 -3.32595527e-01 -4.14466411e-01
-1.07913804e+00 1.25345063e+00 2.27478445e-01 9.06139612e-01
-6.28233969e-01 7.63102248e-02 8.26967657e-01 3.76401573e-01
7.38504350e-01 9.67112422e-01 -7.33383715e-01 -4.14199919e-01
-8.98917377e-01 2.24895760e-01 7.61872113e-01 5.94660997e-01
4.98124421e-01 -1.66133687e-01 7.16862306e-02 1.04119575e+00
7.64379464e-03 -1.12508526e-02 8.42260897e-01 -1.48638427e-01
3.84943813e-01 4.60155249e-01 1.40865535e-01 -8.82555246e-01
-5.70804954e-01 -5.42237163e-01 -3.68224442e-01 -1.28290236e-01
5.80454051e-01 -3.83389026e-01 -1.01638770e+00 1.67387199e+00
2.24548891e-01 5.63774481e-02 3.32958698e-01 6.63959622e-01
6.33193970e-01 5.88000476e-01 4.29368764e-01 -3.75111662e-02
1.61414111e+00 -7.61640966e-01 -9.12601829e-01 -5.09995759e-01
1.08210230e+00 -7.19160557e-01 1.06705475e+00 3.61767679e-01
-7.46745586e-01 -3.68049711e-01 -9.96788323e-01 -4.79300797e-01
-9.53147829e-01 -2.34605178e-01 4.42219973e-01 6.35195315e-01
-9.79957104e-01 5.53815246e-01 -1.03686857e+00 -7.02267230e-01
5.01970589e-01 1.32574722e-01 -7.58069038e-01 -5.82596958e-01
-1.30970955e+00 1.39314139e+00 4.87310350e-01 -2.55168438e-01
-5.38085878e-01 -9.91442025e-01 -1.02194071e+00 1.34521082e-01
3.71574968e-01 -6.90890729e-01 1.00088692e+00 -6.59295082e-01
-1.10914516e+00 9.92430210e-01 8.53925571e-02 -4.61954564e-01
3.02325279e-01 -3.72261018e-01 -4.39145714e-01 2.15221137e-01
-1.64933905e-01 3.69233519e-01 5.18228352e-01 -8.96259248e-01
-1.80335447e-01 -2.44427666e-01 4.75098342e-02 -1.97258085e-01
-7.38421321e-01 2.11913690e-01 -3.20892692e-01 -7.15920091e-01
-3.77891570e-01 -6.15974486e-01 -7.41779506e-01 1.44099206e-01
-4.71301097e-03 -4.71521467e-01 4.29943144e-01 -8.73927176e-01
1.49289238e+00 -2.14238906e+00 -5.66191971e-02 2.79752221e-02
2.22873747e-01 7.13962674e-01 -3.32628399e-01 8.38902354e-01
-3.77154559e-01 1.28080606e-01 -2.78842568e-01 -4.00303662e-01
1.41469482e-02 6.24538302e-01 6.81661218e-02 4.57421005e-01
6.09999478e-01 8.71851683e-01 -1.27885306e+00 -6.83109641e-01
1.80623680e-01 6.44556165e-01 -8.82264972e-01 2.36795217e-01
-1.43600002e-01 -1.50134768e-02 -5.67673504e-01 3.22048903e-01
2.56995857e-01 -2.53494501e-01 5.41022956e-01 1.11084633e-01
1.68966442e-01 4.89027947e-01 -7.71787345e-01 2.10139012e+00
-8.70373666e-01 3.89365673e-01 -1.09744638e-01 -1.33867705e+00
6.48497105e-01 6.16060495e-01 6.37157440e-01 -2.62653559e-01
3.34682316e-01 3.00889581e-01 3.18509609e-01 -9.14806545e-01
3.35136205e-01 -6.83151722e-01 -2.06991211e-01 4.78375167e-01
5.01674294e-01 3.70517485e-02 2.19278499e-01 -7.98778329e-03
1.67690814e+00 -1.70011133e-01 7.79404402e-01 -3.85885775e-01
3.35061491e-01 2.45456755e-01 4.52422649e-01 3.56774628e-01
-2.97697991e-01 7.63987303e-01 6.27659559e-01 -5.72624981e-01
-1.24799013e+00 -9.02437985e-01 -4.18437034e-01 9.20844018e-01
-6.28753722e-01 -7.85816729e-01 -4.74672198e-01 -7.23510325e-01
-5.22283372e-03 8.10386062e-01 -9.79800582e-01 -1.03753395e-01
-5.69065571e-01 -8.98892164e-01 3.64609778e-01 6.61871135e-01
-3.88618946e-01 -8.83873582e-01 -6.87755942e-01 7.88411379e-01
5.34157395e-01 -1.11263573e+00 -2.15198144e-01 4.61673498e-01
-7.05686510e-01 -1.01482964e+00 -9.64480400e-01 -6.90410852e-01
6.56537890e-01 -2.54211962e-01 1.32085001e+00 -7.35918656e-02
-7.15029597e-01 2.59945512e-01 -6.63998961e-01 -4.98558789e-01
-2.40829960e-01 1.86196312e-01 -9.28844362e-02 -3.16157550e-01
1.03727973e+00 -4.12378490e-01 -4.21469271e-01 -3.97173166e-01
-1.26556838e+00 -5.09719431e-01 5.77916265e-01 1.22304177e+00
4.19096917e-01 -3.26142371e-01 5.94743967e-01 -1.35831141e+00
1.01105440e+00 -6.66739285e-01 3.09064090e-02 4.78721149e-02
-6.17910445e-01 1.74380973e-01 6.86096013e-01 -6.64821267e-01
-5.82785606e-01 -3.87401909e-01 -6.09668076e-01 -4.86957133e-01
-2.20258251e-01 8.99265110e-01 2.22310811e-01 5.17877400e-01
8.54937375e-01 8.33474379e-03 1.39462724e-01 -6.82782888e-01
4.18175936e-01 7.27239847e-01 7.13444054e-02 -4.48751837e-01
5.88784814e-01 2.31957465e-01 -4.36595798e-01 -8.97454858e-01
-8.07541609e-01 -6.15030587e-01 -5.04414678e-01 5.10097861e-01
1.19685650e+00 -7.30184197e-01 3.22093465e-03 -4.50632811e-01
-1.15594816e+00 -1.66012585e-01 -5.53375185e-01 7.05680609e-01
-1.99379653e-01 3.43156219e-01 -6.53046072e-01 -5.47683716e-01
-2.61924416e-01 -9.28088725e-01 9.53940034e-01 -1.58862382e-01
-6.36520743e-01 -1.43945360e+00 5.66190600e-01 -1.31301746e-01
5.99366546e-01 2.50918061e-01 1.17589748e+00 -1.36028230e+00
1.68313831e-01 -6.91057086e-01 -1.20650433e-01 5.47288775e-01
3.58039677e-01 -2.57763743e-01 -9.53421593e-01 -1.47287205e-01
-1.76141322e-01 -2.71998942e-01 8.23412597e-01 1.32661715e-01
1.11232078e+00 7.94811696e-02 -3.02535564e-01 3.18899810e-01
1.68179214e+00 -1.49982005e-01 4.52284247e-01 3.00061971e-01
4.42835569e-01 6.77163541e-01 3.49482834e-01 4.09054130e-01
5.58565147e-02 3.96060735e-01 -5.26571535e-02 2.35312562e-02
7.48059675e-02 -1.41413614e-01 2.22497676e-02 9.38559890e-01
4.35304493e-02 -2.17756882e-01 -1.24878097e+00 1.06205130e+00
-1.51108193e+00 -6.12745881e-01 2.75352776e-01 1.84140885e+00
1.24417245e+00 2.35274017e-01 -1.81344345e-01 -5.49538247e-03
1.73373863e-01 1.35042116e-01 -1.23191141e-01 -9.19610441e-01
3.24062616e-01 9.27919269e-01 4.43341076e-01 3.22206378e-01
-9.81558621e-01 8.06143582e-01 5.90917921e+00 8.30080450e-01
-1.02238727e+00 6.52599216e-01 2.27222994e-01 -2.38345146e-01
-4.50699568e-01 -2.80969173e-01 -6.21825278e-01 3.66406411e-01
1.41492271e+00 -8.62011090e-02 -8.03460777e-02 9.49077249e-01
7.33561069e-02 2.73977757e-01 -1.33248866e+00 9.38050568e-01
2.64313668e-01 -1.28529263e+00 -2.11912543e-01 1.84135333e-01
4.12832618e-01 2.12487757e-01 -6.81593865e-02 5.51727235e-01
4.26016867e-01 -1.26852500e+00 -1.24184750e-01 3.87870036e-02
8.32712352e-01 -4.55382317e-01 1.22219563e+00 2.97188491e-01
-6.65748477e-01 -1.11725582e-02 -6.29200399e-01 2.44786978e-01
2.37119779e-01 7.46308506e-01 -1.06032920e+00 6.80667996e-01
3.19143534e-01 5.71028173e-01 -3.04394335e-01 8.07556510e-01
-1.24321818e-01 4.65538502e-01 -3.71532112e-01 -9.49272439e-02
6.96890593e-01 3.97311971e-02 1.79657713e-01 1.57598758e+00
1.92019641e-01 8.78779069e-02 1.32354513e-01 4.77698207e-01
-9.11976248e-02 4.21064854e-01 -9.29643035e-01 -5.54386795e-01
-4.83181477e-02 1.23287797e+00 -4.43295747e-01 -4.27788168e-01
-8.90790462e-01 7.19066739e-01 5.37867248e-01 1.11925378e-01
-5.47700822e-01 -4.71417874e-01 1.20529675e+00 1.81913748e-01
4.91192818e-01 -3.76410007e-01 -5.80389611e-02 -1.18099391e+00
-1.89700454e-01 -7.29415834e-01 5.50953567e-01 -2.54272014e-01
-1.74784660e+00 8.87995601e-01 -1.28293233e-02 -1.05556476e+00
-3.48066688e-01 -1.25319457e+00 -5.04625916e-01 8.95263612e-01
-1.82228601e+00 -8.01048219e-01 3.16753358e-01 2.87645847e-01
6.11046910e-01 -9.09610912e-02 1.39232779e+00 6.18987799e-01
-5.41013122e-01 5.95494390e-01 2.58075207e-01 3.48836273e-01
8.29781890e-01 -1.23810148e+00 6.52991161e-02 3.91587108e-01
1.57999501e-01 9.93195295e-01 8.27666104e-01 -3.01295012e-01
-1.22878575e+00 -8.58764291e-01 1.40512919e+00 -5.24557412e-01
1.12116539e+00 -5.69882393e-01 -1.16213977e+00 6.20257080e-01
3.74971092e-01 3.50925237e-01 1.31751859e+00 4.80425209e-01
-3.78540486e-01 2.45865241e-01 -1.08115923e+00 5.15142858e-01
8.80949318e-01 -6.91475034e-01 -1.09251368e+00 5.52975237e-01
6.36232436e-01 -9.27205607e-02 -1.12964761e+00 -1.07302181e-01
3.29694688e-01 -1.90358281e-01 1.04533195e+00 -1.34089220e+00
6.80956483e-01 2.18002409e-01 -1.61017761e-01 -1.50801814e+00
-1.17690213e-01 -4.06038433e-01 2.50863045e-01 7.94204175e-01
6.74871922e-01 -8.08421433e-01 5.45481622e-01 7.25214839e-01
-2.33312756e-01 -1.03161943e+00 -1.06185079e+00 -7.18628526e-01
6.43442571e-01 -6.18360400e-01 2.85533458e-01 1.01287019e+00
3.12051386e-01 2.57026196e-01 1.32438600e-01 -6.55113831e-02
-3.16118193e-03 -1.52309731e-01 4.47438031e-01 -1.11141574e+00
-3.65779370e-01 -2.22350568e-01 -6.51477098e-01 -5.44734478e-01
3.36741984e-01 -8.46372843e-01 7.59646520e-02 -1.60735106e+00
-6.17393740e-02 -4.51884001e-01 -7.97611058e-01 5.40417075e-01
-3.04221034e-01 9.45594832e-02 -1.23365477e-01 -3.18004847e-01
-2.43746579e-01 3.09610277e-01 8.93705130e-01 -1.74190104e-01
1.95877366e-02 -6.55005813e-01 -8.03427517e-01 5.33740222e-01
7.56897211e-01 -7.76490867e-01 -3.26356083e-01 -6.50363207e-01
2.27141082e-01 -3.19587886e-01 -1.20246261e-01 -5.65297782e-01
-8.05949420e-02 9.25053582e-02 1.21098325e-01 8.52583423e-02
3.89940768e-01 -8.62246156e-01 -4.18604404e-01 4.24254030e-01
-4.30222452e-01 1.64590061e-01 3.80629599e-01 6.96906626e-01
-3.96355420e-01 -6.27669632e-01 7.11536229e-01 -4.07566398e-01
-5.37242353e-01 2.10449949e-01 -3.18988949e-01 3.16941887e-01
9.46570158e-01 -4.50822413e-02 -7.12868944e-02 7.62958452e-02
-9.01055396e-01 7.08284602e-02 3.34378660e-01 4.12243038e-01
5.44556379e-01 -1.04292738e+00 -7.61783183e-01 1.61604032e-01
4.19970334e-01 -2.43340209e-02 1.52122006e-01 6.81705952e-01
-7.79560804e-01 6.06092155e-01 -5.74382544e-02 -5.74868731e-02
-1.05398822e+00 9.15268660e-01 -6.40904978e-02 -8.48969102e-01
-7.85490990e-01 8.94967139e-01 2.65790168e-02 -4.47278023e-01
-5.85447401e-02 -5.37734091e-01 -2.64677584e-01 3.01002055e-01
6.04213715e-01 -3.24776262e-01 2.40379438e-01 -3.03149700e-01
-5.27611852e-01 1.93418041e-01 -2.77804554e-01 -1.35128349e-01
1.88330472e+00 3.58931839e-01 2.39717424e-01 3.59546602e-01
1.55691969e+00 -2.40300242e-02 -7.21701264e-01 -4.25921082e-01
2.93508261e-01 -4.71712619e-01 2.18976825e-01 -3.75263691e-01
-7.12735176e-01 1.25339186e+00 4.16695297e-01 6.49241060e-02
7.64168262e-01 -1.08248992e-02 9.90981460e-01 4.53492373e-01
2.36030802e-01 -1.37001657e+00 -3.39783952e-02 2.42942348e-01
6.04453981e-01 -1.17651737e+00 3.43135521e-02 2.45766509e-02
-6.86304450e-01 1.19809437e+00 5.71333356e-02 -5.27793288e-01
1.10855937e+00 1.97253138e-01 1.55107141e-01 -1.70165554e-01
-9.10738766e-01 -4.06027228e-01 1.44432914e-02 7.18856215e-01
7.65619457e-01 -8.02805796e-02 -6.00771368e-01 8.13565969e-01
1.62041470e-01 2.12997481e-01 3.80737513e-01 1.18480122e+00
-1.96604460e-01 -1.72877121e+00 -6.39595911e-02 6.26646042e-01
-9.80809987e-01 -3.98031592e-01 -9.59939137e-02 9.80543911e-01
2.99667001e-01 6.73758745e-01 1.51727527e-01 3.67771909e-02
4.35857594e-01 5.83300173e-01 3.76397073e-01 -1.25522327e+00
-7.55571961e-01 -5.25155701e-02 4.20782268e-01 -3.90076220e-01
-3.17899346e-01 -5.61859131e-01 -1.00540173e+00 -1.32918917e-02
-3.44967993e-05 8.98049697e-02 5.98404646e-01 7.79779255e-01
4.13518399e-01 7.03686297e-01 -1.33704901e-01 -3.97073388e-01
-7.73880601e-01 -1.20069790e+00 -4.35953289e-01 6.23384356e-01
2.80962676e-01 -6.52101278e-01 -9.21561643e-02 -9.47345048e-02] | [8.664032936096191, 8.626298904418945] |
11de3e9e-0d4c-4f87-b5a3-fa7e6cd7b700 | dhge-dual-view-hyper-relational-knowledge | 2207.08562 | null | https://arxiv.org/abs/2207.08562v4 | https://arxiv.org/pdf/2207.08562v4.pdf | DHGE: Dual-View Hyper-Relational Knowledge Graph Embedding for Link Prediction and Entity Typing | In the field of representation learning on knowledge graphs (KGs), a hyper-relational fact consists of a main triple and several auxiliary attribute-value descriptions, which is considered more comprehensive and specific than a triple-based fact. However, currently available hyper-relational KG embedding methods in a single view are limited in application because they weaken the hierarchical structure that represents the affiliation between entities. To overcome this limitation, we propose a dual-view hyper-relational KG structure (DH-KG) that contains a hyper-relational instance view for entities and a hyper-relational ontology view for concepts that are abstracted hierarchically from the entities. This paper defines link prediction and entity typing tasks on DH-KG for the first time and constructs two DH-KG datasets, JW44K-6K, extracted from Wikidata, and HTDM based on medical data. Furthermore, we propose DHGE, a DH-KG embedding model based on GRAN encoders, HGNNs, and joint learning. DHGE outperforms baseline models on DH-KG, according to experimental results. Finally, we provide an example of how this technology can be used to treat hypertension. Our model and new datasets are publicly available. | ['Kaiyang Wan', 'Tianyu Yao', 'Gengxian Zhou', 'Ling Tan', 'Haihong E', 'Haoran Luo'] | 2022-07-18 | null | null | null | null | ['link-prediction-on-dh-kgs', 'entity-typing-on-dh-kgs', 'entity-typing'] | ['natural-language-processing', 'natural-language-processing', 'natural-language-processing'] | [-5.77031612e-01 1.01891553e+00 -8.72788787e-01 -3.61714929e-01
-3.98964316e-01 -3.53967696e-02 3.39326084e-01 5.80818355e-01
-4.03847434e-02 9.01780188e-01 7.38371968e-01 -1.12799719e-01
-6.78145766e-01 -1.49114668e+00 -6.68889165e-01 -4.11273450e-01
-5.27853310e-01 6.88837826e-01 1.08342640e-01 -2.63290673e-01
-5.24301469e-01 -3.91338132e-02 -1.22612119e+00 5.62362134e-01
6.88023031e-01 8.03776324e-01 -2.54717439e-01 1.55652925e-01
-2.49411806e-01 1.19126272e+00 -2.51042396e-01 -1.04717910e+00
-8.83226767e-02 1.25074297e-01 -1.00795710e+00 -4.69419837e-01
4.65715259e-01 -2.78764188e-01 -8.81827235e-01 7.58432746e-01
7.01948106e-01 -3.33435744e-01 5.53315580e-01 -1.45051575e+00
-1.23466420e+00 1.26972115e+00 -1.54867113e-01 -1.95030466e-01
3.08601558e-01 -4.68513221e-01 1.68172801e+00 -7.64226854e-01
1.09793270e+00 1.20953417e+00 1.01643741e+00 5.62003970e-01
-1.10337365e+00 -4.07658309e-01 6.36997148e-02 5.97532690e-01
-1.58127558e+00 -1.06735192e-02 4.12895828e-01 -3.80581945e-01
1.34641600e+00 1.56673834e-01 1.01835048e+00 1.06198013e+00
3.80350500e-02 7.72870302e-01 5.83707988e-01 -2.90288210e-01
2.58593503e-02 3.73526663e-01 4.26813364e-01 1.04845417e+00
8.84550154e-01 -4.78875414e-02 -6.17053270e-01 -3.73631001e-01
5.23921072e-01 -1.56934246e-01 -4.91158724e-01 -9.55687344e-01
-1.20267522e+00 8.86317790e-01 8.19292307e-01 1.55162051e-01
-3.86986703e-01 1.43930048e-01 6.71283901e-01 1.44239157e-01
3.48680317e-01 2.96507984e-01 -7.59809196e-01 3.24334681e-01
-3.47922981e-01 2.83126831e-01 1.16778386e+00 1.36369944e+00
5.03228486e-01 -1.97786421e-01 -1.67518839e-01 8.25215340e-01
3.30552816e-01 1.72589168e-01 3.08918178e-01 -4.98457164e-01
6.08616292e-01 1.09394443e+00 -4.47615594e-01 -1.21849251e+00
-5.94406903e-01 -5.96143365e-01 -9.21495736e-01 -4.57984746e-01
-2.27419749e-01 2.00215578e-01 -9.02670324e-01 1.64550900e+00
4.72972572e-01 4.18295890e-01 6.19832575e-01 3.98354262e-01
1.84584916e+00 2.36479878e-01 3.42209250e-01 1.91346303e-01
1.70544422e+00 -8.51248682e-01 -1.12675750e+00 4.60950941e-01
1.26981890e+00 3.84688020e-01 5.10727763e-01 -6.64699264e-03
-8.05419445e-01 -3.05272788e-01 -1.03027618e+00 -4.05053735e-01
-1.31161237e+00 1.87938511e-01 1.06482577e+00 5.49067199e-01
-9.94425595e-01 4.07370538e-01 -6.29282713e-01 -4.05257106e-01
5.42666554e-01 3.16808224e-01 -9.62180257e-01 -1.45018294e-01
-2.07024956e+00 1.01405478e+00 1.09459507e+00 4.33441922e-02
-4.56547022e-01 -1.16157675e+00 -1.33460248e+00 4.16335076e-01
6.24344110e-01 -1.24766219e+00 3.89399290e-01 4.20737207e-01
-7.67273128e-01 9.15236175e-01 4.10832345e-01 -6.11217737e-01
-9.65611190e-02 -1.83952674e-01 -1.05726326e+00 9.63971838e-02
1.53670803e-01 4.21216249e-01 1.70726478e-01 -1.40157151e+00
-6.84342444e-01 -5.91439486e-01 4.88507450e-01 1.85640678e-01
-6.18976891e-01 -6.80266559e-01 -6.56554401e-01 -3.98501426e-01
-1.51443169e-01 -6.56892836e-01 9.67254490e-03 -3.36389482e-01
-9.75731373e-01 -4.45687473e-01 5.64466119e-01 -7.54671693e-01
1.83158851e+00 -1.98591113e+00 5.23831189e-01 2.22139344e-01
8.20380807e-01 2.82756567e-01 2.41652519e-01 7.33907521e-01
-2.82925159e-01 3.44204247e-01 9.93781164e-02 2.02017482e-02
1.97665140e-01 6.89615607e-01 -1.99234784e-01 -5.40284216e-02
-5.37297018e-02 1.24654281e+00 -1.06383204e+00 -6.72097147e-01
-5.44230193e-02 5.44443071e-01 -7.32286155e-01 -3.25204358e-02
-5.38964234e-02 -3.99201244e-01 -3.84487659e-01 8.36852789e-01
4.97687817e-01 -5.71584165e-01 7.36492574e-01 -8.56775224e-01
4.09210563e-01 2.94889420e-01 -1.06642628e+00 1.55925548e+00
-2.52856493e-01 -1.01872906e-02 -4.58560646e-01 -9.25641119e-01
5.61262429e-01 5.03534734e-01 7.88630247e-01 -2.71879047e-01
-3.34310710e-01 7.69809261e-02 -2.41183266e-01 -5.95826209e-01
7.09032476e-01 2.07792982e-01 -1.02401063e-01 -1.47501647e-01
6.03576362e-01 3.49876791e-01 2.81674683e-01 8.22596312e-01
1.12741387e+00 1.50378600e-01 8.06425750e-01 6.05742913e-03
4.11363959e-01 -1.25787556e-01 6.41441345e-01 4.42479640e-01
3.15906107e-01 5.57048060e-02 9.57732916e-01 -5.76241612e-01
-6.89590156e-01 -1.20835304e+00 -5.96048176e-01 6.64627850e-01
-1.16523113e-02 -1.44608271e+00 -8.07343796e-02 -1.21754873e+00
8.59720349e-01 5.68050683e-01 -1.05768549e+00 -4.46560144e-01
-2.30707318e-01 -8.27923000e-01 7.90533245e-01 8.41545165e-01
2.89266437e-01 -7.49534965e-01 3.67764905e-02 2.11095110e-01
-7.42228776e-02 -1.33680046e+00 1.45039096e-01 2.82224387e-01
-6.56634867e-01 -1.36638832e+00 -2.07096517e-01 -6.63786352e-01
4.45027024e-01 -2.83454031e-01 1.51984704e+00 1.48733994e-02
-2.92924166e-01 5.82581818e-01 -4.43563253e-01 -3.57354581e-01
-2.44434737e-02 3.09674770e-01 6.95480108e-02 -2.51161873e-01
5.73185027e-01 -5.68091989e-01 -3.87999028e-01 -1.17892593e-01
-8.40787530e-01 7.98581839e-02 6.61425233e-01 9.95867491e-01
7.41188049e-01 2.83154368e-01 8.30507636e-01 -1.62499702e+00
3.59408259e-01 -8.90261292e-01 -7.58565217e-02 7.16859400e-01
-1.01350820e+00 1.56371012e-01 2.34452948e-01 1.16655193e-01
-7.58522928e-01 -2.84268677e-01 -5.41293882e-02 -4.85236943e-01
3.73273760e-01 1.35786021e+00 -3.41679305e-01 3.00753236e-01
4.21489060e-01 3.93664166e-02 -3.31829548e-01 -6.09673977e-01
9.64373946e-01 5.22909403e-01 4.99402463e-01 -5.80096960e-01
5.86457729e-01 3.44852835e-01 1.99237287e-01 -6.33474529e-01
-1.00931334e+00 -4.41925466e-01 -8.07321727e-01 2.26914749e-01
8.93762708e-01 -1.36570311e+00 -7.17494786e-01 -6.12792969e-02
-8.21326852e-01 6.34411871e-02 -5.13748109e-01 6.02057099e-01
-4.45869476e-01 1.39694959e-01 -8.93244505e-01 -1.31546736e-01
-3.22581887e-01 -4.56654847e-01 9.88261998e-01 -2.78743327e-01
8.96991342e-02 -1.32375884e+00 9.85160917e-02 3.28871131e-01
9.29676294e-02 3.54159236e-01 1.66960132e+00 -8.66472065e-01
-5.51787078e-01 -1.82391584e-01 -3.18356305e-01 9.46807265e-02
3.28742743e-01 -2.66082972e-01 -5.67194939e-01 -1.76474586e-01
-9.60892558e-01 -3.84995282e-01 1.04571319e+00 9.49433297e-02
1.14593303e+00 -4.39912379e-01 -8.81177425e-01 8.38909090e-01
1.70528710e+00 -1.47639737e-01 7.97747612e-01 4.52017814e-01
1.25502026e+00 4.66701269e-01 2.11881712e-01 3.81325334e-01
1.22399950e+00 9.04347658e-01 5.23512959e-01 -8.49236697e-02
-2.88479000e-01 -6.67940795e-01 -4.42403331e-02 1.19998121e+00
-3.33388716e-01 -2.03922525e-01 -7.87010789e-01 7.42202640e-01
-1.99792421e+00 -1.01576972e+00 -3.36170226e-01 1.91161191e+00
1.14017308e+00 -7.87795708e-02 6.20228015e-02 -8.59780088e-02
3.84299666e-01 2.98605189e-02 -3.02202642e-01 2.25860886e-02
-3.06178749e-01 1.10388972e-01 4.89242077e-01 2.41286904e-01
-1.14633739e+00 8.93458247e-01 5.65612698e+00 5.63060760e-01
-4.73621815e-01 4.88360375e-02 -1.06072918e-01 1.91689909e-01
-5.66544354e-01 -4.48042788e-02 -1.23583615e+00 3.16615164e-01
9.24476862e-01 -4.01076585e-01 -2.62235940e-01 9.61751759e-01
-7.68239737e-01 5.15929043e-01 -1.28501368e+00 8.61905813e-01
-1.21942256e-03 -1.69136822e+00 5.43331325e-01 3.77511114e-01
3.64109755e-01 -1.63573220e-01 -6.90875947e-02 1.02388418e+00
6.82398915e-01 -1.08128893e+00 6.13399632e-02 6.72700346e-01
9.04940486e-01 -5.63830554e-01 9.71941590e-01 -1.74700975e-01
-1.54668319e+00 4.08394001e-02 -4.64850873e-01 6.65741205e-01
-6.59957305e-02 6.48613155e-01 -8.32394242e-01 1.74235320e+00
8.82006168e-01 1.31642878e+00 -6.46175265e-01 7.30414212e-01
-3.52398843e-01 2.62428969e-01 1.76301554e-01 4.72335309e-01
-9.88582820e-02 1.11053921e-01 2.39782557e-01 1.29134107e+00
1.73575729e-01 1.66646808e-01 1.20451637e-01 7.43512750e-01
-3.45784247e-01 1.20275564e-01 -1.00122774e+00 -1.05854049e-01
5.27279198e-01 1.29660070e+00 -1.61368717e-02 -6.42903090e-01
-8.14452291e-01 5.44608951e-01 7.56901860e-01 2.92888194e-01
-7.20787227e-01 -5.66740572e-01 6.52633309e-01 6.02143854e-02
4.56220448e-01 3.03283960e-01 2.99467802e-01 -1.56297457e+00
-1.00438640e-01 -6.28139675e-01 1.20148361e+00 -6.35038018e-01
-1.60550904e+00 5.79439878e-01 2.35516578e-01 -1.16028607e+00
-1.85126007e-01 -7.03560233e-01 3.52404237e-01 6.04920924e-01
-1.71679389e+00 -1.73653162e+00 -1.85920030e-01 6.12927914e-01
-3.57726306e-01 -2.94238031e-01 1.37978363e+00 7.66565144e-01
-5.98065972e-01 7.38720238e-01 -4.99870293e-02 5.70339084e-01
6.02687240e-01 -1.77800596e+00 1.51839748e-01 4.60065762e-03
1.50290281e-01 9.86008763e-01 9.22982395e-02 -9.51624811e-01
-1.37307334e+00 -1.35783827e+00 1.16816688e+00 -8.25199008e-01
6.72540307e-01 -4.04433459e-01 -1.19466972e+00 1.29265320e+00
1.54725432e-01 4.88182157e-01 1.19906676e+00 8.05071831e-01
-7.46578515e-01 -1.75485238e-01 -1.02105772e+00 4.04134035e-01
1.36430538e+00 -5.39953887e-01 -7.88284659e-01 3.96701574e-01
9.79284763e-01 -4.87417191e-01 -1.98483205e+00 8.43863010e-01
5.87273180e-01 -5.28909624e-01 1.32526684e+00 -1.19735396e+00
4.28342551e-01 -2.91811258e-01 -3.10006678e-01 -1.49984729e+00
-6.39484107e-01 5.61686642e-02 -1.07560408e+00 1.20188105e+00
4.21402693e-01 -5.63492060e-01 7.85147905e-01 2.84140795e-01
-2.70940304e-01 -1.24442875e+00 -8.87405396e-01 -8.88783336e-01
-1.82650879e-01 6.54838886e-03 8.76275241e-01 1.63708806e+00
4.83138144e-01 5.94758749e-01 -4.77040261e-01 4.99380022e-01
6.70774519e-01 1.40640467e-01 5.73457956e-01 -1.52442276e+00
-2.03511804e-01 8.28267112e-02 -9.75945294e-01 -3.64292383e-01
2.97063559e-01 -1.49074304e+00 -8.28306675e-01 -2.13077760e+00
3.82897943e-01 -6.10403955e-01 -6.55744314e-01 9.57176685e-01
-1.51057422e-01 -2.05536947e-01 -2.03932613e-01 -7.54905790e-02
-6.08527243e-01 7.26160467e-01 9.24678802e-01 -3.45681310e-01
-3.56777050e-02 -4.72506791e-01 -7.80099988e-01 3.61739367e-01
2.51739323e-01 -3.53524745e-01 -6.49183691e-01 -2.55053580e-01
6.89927518e-01 1.25951305e-01 2.87784606e-01 -6.75278604e-01
3.88960302e-01 2.90004253e-01 2.22786158e-01 -6.47066414e-01
4.16068017e-01 -9.95844603e-01 3.30201447e-01 3.51282269e-01
-3.87232035e-01 -2.03965440e-01 -1.09349266e-02 9.78851438e-01
-4.12531435e-01 9.64973196e-02 1.12935722e-01 -1.86291337e-01
-9.65798974e-01 5.77026367e-01 4.89438534e-01 -1.73343364e-02
9.25836504e-01 2.75454335e-02 -8.25599074e-01 -6.85404837e-02
-1.23193312e+00 5.41496158e-01 -2.12396055e-01 4.76284862e-01
7.89165795e-01 -1.86445796e+00 -7.21851647e-01 3.92605439e-02
7.46398330e-01 1.01471664e-02 2.86752850e-01 8.07825685e-01
-1.20419212e-01 7.12472141e-01 -2.28095770e-01 -2.16708675e-01
-1.19494426e+00 9.47778523e-01 1.07758649e-01 -7.02081919e-01
-1.11935914e+00 6.01017833e-01 1.50090948e-01 -7.58828461e-01
3.35710913e-01 -3.44111472e-01 -7.77956903e-01 4.23291266e-01
1.90834194e-01 3.65637094e-01 2.66536653e-01 -5.27423322e-01
-3.69565189e-01 2.47291848e-01 -3.28139603e-01 4.73546386e-01
1.54650843e+00 3.42386737e-02 -5.15715182e-01 4.38423127e-01
1.08136439e+00 6.10644743e-02 -2.64253169e-01 -5.81723809e-01
2.63513476e-01 -2.84285694e-01 -1.36248227e-02 -8.58165026e-01
-1.18452406e+00 4.01181400e-01 1.51390702e-01 2.95873225e-01
7.74926066e-01 2.97124743e-01 6.44052148e-01 5.70446730e-01
4.35326755e-01 -8.33783507e-01 -1.24960206e-01 2.48774856e-01
8.77221704e-01 -8.07888448e-01 2.86642432e-01 -9.12195146e-01
-6.65676951e-01 9.69846368e-01 7.51920283e-01 3.46759290e-01
1.00421190e+00 2.24337243e-02 -2.53227741e-01 -8.71386528e-01
-1.05113769e+00 -4.82203782e-01 5.07169127e-01 7.83297181e-01
4.41181928e-01 3.16241086e-01 -1.69268146e-01 8.77715647e-01
-1.76185355e-01 1.01134911e-01 3.87086809e-01 7.38729358e-01
1.04075007e-01 -1.18948925e+00 4.99117941e-01 8.83428931e-01
-4.55931485e-01 -3.78176779e-01 -2.06829414e-01 1.22744179e+00
4.47860748e-01 2.89988458e-01 -3.08637202e-01 -7.16709852e-01
5.87765276e-01 3.14164728e-01 3.46211851e-01 -8.97342682e-01
-3.46095949e-01 -6.38613820e-01 5.99758506e-01 -4.70668375e-01
-4.29006785e-01 -2.09825218e-01 -1.21367562e+00 -2.37459302e-01
-3.50029290e-01 2.92110890e-01 2.77259022e-01 3.77399355e-01
7.16059446e-01 9.25524950e-01 1.28753886e-01 2.03507498e-01
-2.26174980e-01 -7.31269896e-01 -1.12333333e+00 5.49162328e-01
8.51790607e-02 -9.25638795e-01 -8.89478903e-03 -1.59532234e-01] | [8.70729923248291, 7.949190139770508] |
25420963-b048-40ae-9ce8-e47dfc1c19aa | a-semi-lagrangian-two-level-preconditioned | 1604.02153 | null | http://arxiv.org/abs/1604.02153v2 | http://arxiv.org/pdf/1604.02153v2.pdf | A Semi-Lagrangian two-level preconditioned Newton-Krylov solver for constrained diffeomorphic image registration | We propose an efficient numerical algorithm for the solution of diffeomorphic
image registration problems. We use a variational formulation constrained by a
partial differential equation (PDE), where the constraints are a scalar
transport equation.
We use a pseudospectral discretization in space and second-order accurate
semi-Lagrangian time stepping scheme for the transport equations. We solve for
a stationary velocity field using a preconditioned, globalized, matrix-free
Newton-Krylov scheme. We propose and test a two-level Hessian preconditioner.
We consider two strategies for inverting the preconditioner on the coarse grid:
a nested preconditioned conjugate gradient method (exact solve) and a nested
Chebyshev iterative method (inexact solve) with a fixed number of iterations.
We test the performance of our solver in different synthetic and real-world
two-dimensional application scenarios. We study grid convergence and
computational efficiency of our new scheme. We compare the performance of our
solver against our initial implementation that uses the same spatial
discretization but a standard, explicit, second-order Runge-Kutta scheme for
the numerical time integration of the transport equations and a single-level
preconditioner. Our improved scheme delivers significant speedups over our
original implementation. As a highlight, we observe a 20$\times$ speedup for a
two dimensional, real world multi-subject medical image registration problem. | ['George Biros', 'Andreas Mang'] | 2016-04-07 | null | null | null | null | ['constrained-diffeomorphic-image-registration'] | ['computer-vision'] | [-7.71614090e-02 -3.21926206e-01 4.84125674e-01 -3.28559009e-03
-9.57740724e-01 -2.61815101e-01 3.19895715e-01 -9.87019539e-02
-8.36088538e-01 9.87265944e-01 -8.99328664e-02 -3.35461050e-01
-1.10007480e-01 -5.56960642e-01 -3.33439797e-01 -9.13479507e-01
-2.66318917e-01 5.66875160e-01 2.37647280e-01 -3.83733988e-01
1.48685023e-01 6.05025828e-01 -9.99414980e-01 -2.15826500e-02
1.03094232e+00 8.26478958e-01 -1.68429390e-01 8.10942054e-01
4.27471876e-01 5.67862451e-01 2.12420076e-02 1.51473716e-01
4.14703310e-01 -7.30027795e-01 -1.28584838e+00 1.62668273e-01
1.38074130e-01 -3.79418701e-01 -7.29230866e-02 9.71498251e-01
6.29101634e-01 4.20331210e-01 5.51764250e-01 -1.19312036e+00
9.10863206e-02 -1.53813720e-01 -6.71015739e-01 2.27245137e-01
3.94267291e-01 1.99058115e-01 3.28339696e-01 -9.31204855e-01
8.27651620e-01 8.29243004e-01 1.27003002e+00 4.37197506e-01
-1.68323088e+00 -1.40555561e-01 -6.33399010e-01 -2.36227810e-01
-1.59192169e+00 -1.75515324e-01 3.51130247e-01 -7.03638852e-01
9.36590433e-01 5.40463746e-01 7.79591024e-01 3.60019803e-01
7.15092599e-01 8.25618580e-02 1.65550375e+00 -2.33133867e-01
4.06965017e-01 -3.79714817e-02 1.41279370e-01 8.65103245e-01
-7.84083828e-02 3.68322939e-01 -4.27054428e-02 -1.00493419e+00
1.11598706e+00 -2.99390435e-01 -4.93384391e-01 -2.69740134e-01
-1.37058854e+00 1.01817441e+00 3.16016853e-01 5.20335317e-01
-6.61890149e-01 3.96416895e-02 3.00331265e-01 9.65132192e-02
8.45483065e-01 3.51512104e-01 -4.13725287e-01 -3.05115189e-02
-1.22941744e+00 5.84476352e-01 1.13012290e+00 4.28036064e-01
7.31153965e-01 -1.08929060e-01 -1.52160898e-01 5.60079515e-01
2.70046949e-01 5.98166764e-01 4.21906650e-01 -1.43449903e+00
-3.78552735e-01 -3.56853940e-02 3.83591771e-01 -6.66489482e-01
-5.92132866e-01 -2.43851915e-01 -1.00278842e+00 5.35398841e-01
6.03823781e-01 -4.45903093e-01 -5.19010246e-01 1.34089863e+00
8.09232235e-01 6.79438949e-01 -1.52376488e-01 1.11590314e+00
2.26043656e-01 6.24790251e-01 -1.07481562e-01 -7.37489283e-01
1.23560941e+00 -7.12945819e-01 -6.82399809e-01 4.58533645e-01
1.21390986e+00 -8.94829214e-01 5.60340583e-01 2.23510548e-01
-1.46495450e+00 -1.69587642e-01 -6.66891038e-01 -8.50069076e-02
6.71853051e-02 5.05588017e-02 2.37056881e-01 3.87811303e-01
-1.38792729e+00 1.17151666e+00 -1.20388830e+00 -1.07689939e-01
-6.71034381e-02 5.14228940e-01 -3.78339589e-01 3.09261978e-01
-1.05970311e+00 8.59866619e-01 -1.54478908e-01 -6.01547165e-03
-4.68520820e-01 -1.21513093e+00 -8.18756700e-01 -3.39972347e-01
-2.16423675e-01 -9.02563095e-01 9.64189470e-01 -7.94714153e-01
-1.59647107e+00 1.19684875e+00 -3.55074286e-01 -2.88693190e-01
8.96144390e-01 2.68693596e-01 6.15318902e-02 8.36510360e-02
2.39741474e-01 -1.57235526e-02 4.62201536e-01 -1.19997358e+00
-1.65772699e-02 -1.08863562e-01 -3.08743298e-01 2.91790664e-01
1.43951163e-01 9.98278633e-02 2.52703186e-02 -6.49357140e-01
2.10675299e-01 -1.35259759e+00 -6.18432701e-01 1.00391708e-01
-3.64886373e-02 3.56043935e-01 4.81726944e-01 -1.13385260e+00
1.03645349e+00 -2.16103005e+00 4.35883969e-01 3.41990381e-01
1.70938075e-01 -9.57333893e-02 2.62412578e-01 3.26514244e-01
-2.14213073e-01 -2.67625868e-01 -9.82600152e-01 -4.34053242e-01
-3.79243642e-01 2.76132494e-01 7.90610239e-02 1.10478318e+00
-4.03457701e-01 5.32527328e-01 -1.05745435e+00 -5.26727855e-01
-1.17924996e-01 8.32930982e-01 -8.61756384e-01 7.14207515e-02
3.38449806e-01 1.14084685e+00 -3.94417316e-01 -7.59623051e-02
9.74085033e-01 -3.14541608e-01 1.69205979e-01 -3.91869754e-01
-7.10617244e-01 5.80088049e-03 -1.54434144e+00 1.82440972e+00
-4.98481214e-01 2.06977203e-01 9.08865631e-01 -9.21439588e-01
2.80235916e-01 5.52244067e-01 9.27050352e-01 -3.51369947e-01
2.57198870e-01 7.03540504e-01 -2.40091309e-01 -2.63598949e-01
4.05348092e-01 -6.92905605e-01 4.03110504e-01 4.98337299e-01
-2.87068009e-01 -4.86801833e-01 1.03054397e-01 2.78482795e-01
1.13474643e+00 1.04328617e-01 2.71604240e-01 -1.28682959e+00
7.93330073e-01 4.53432053e-01 4.93114233e-01 3.55674952e-01
-9.21058953e-02 5.80306470e-01 5.31627774e-01 -5.34631848e-01
-1.02580810e+00 -6.24798179e-01 -7.22364128e-01 4.44956928e-01
-3.78767475e-02 -1.60956874e-01 -9.69811141e-01 -1.12031229e-01
-2.81404657e-03 3.36786360e-01 -5.85363388e-01 2.34029517e-01
-1.00737393e+00 -1.34527767e+00 2.09050223e-01 5.43754846e-02
7.07213998e-01 -5.43856919e-01 -6.77274704e-01 2.95956165e-01
-9.47328806e-02 -1.03994000e+00 -4.74272281e-01 -2.62300432e-01
-1.18099165e+00 -9.56092656e-01 -9.36972260e-01 -7.62068212e-01
7.95763969e-01 -2.70245820e-01 1.02759695e+00 5.54227889e-01
-2.95041472e-01 4.91285652e-01 2.31494620e-01 5.87952018e-01
-5.04907489e-01 -2.85457939e-01 2.48527497e-01 -6.58892244e-02
-5.21357179e-01 -4.87905920e-01 -7.74275661e-01 2.90401340e-01
-8.00746441e-01 1.06177136e-01 -1.33839501e-02 1.09908259e+00
7.01122642e-01 -2.49369934e-01 -1.31023973e-01 -7.75214612e-01
6.69005334e-01 -4.48248476e-01 -8.73977482e-01 -2.14177862e-01
-3.80994081e-01 8.81283209e-02 3.75499368e-01 -3.10614973e-01
-9.47125971e-01 3.54776800e-01 -4.73351568e-01 -2.98861235e-01
3.65336686e-01 4.78653908e-01 5.45238972e-01 -1.00817132e+00
7.28711128e-01 2.21082404e-01 3.42721403e-01 -5.18761039e-01
1.01411544e-01 1.66570768e-01 5.12446105e-01 -9.31920469e-01
6.65107906e-01 8.43303919e-01 4.38850194e-01 -9.88331258e-01
-1.15268238e-01 -4.85091776e-01 -6.89194560e-01 -7.69447237e-02
1.12950552e+00 -5.56755066e-01 -9.36600149e-01 7.87517965e-01
-1.22282326e+00 -8.70061457e-01 -5.09286582e-01 6.72585070e-01
-8.45706224e-01 5.51661551e-01 -1.10029185e+00 -4.74611223e-01
-2.23231331e-01 -1.44816184e+00 8.85076642e-01 -3.16060901e-01
-2.03344062e-01 -1.48634255e+00 6.64317787e-01 -1.14314057e-01
7.35647440e-01 7.92282820e-01 5.21288157e-01 8.59751701e-02
-2.18026355e-01 2.80624062e-01 -1.08733080e-01 6.84530986e-03
-1.92713499e-01 -3.50588143e-01 -6.26390874e-01 -6.22202396e-01
9.08746064e-01 1.41859293e-01 3.93975079e-01 6.29220784e-01
5.47595620e-01 -4.23435301e-01 -5.22582054e-01 9.82595325e-01
1.73625219e+00 -2.17081800e-01 3.37243110e-01 3.74576934e-02
6.27176344e-01 2.39345476e-01 3.62944037e-01 4.87997413e-01
2.91697949e-01 7.00573981e-01 -2.54536308e-02 -4.47062075e-01
1.20623134e-01 4.49497312e-01 -1.26046995e-02 8.78929079e-01
-5.75398147e-01 4.94552165e-01 -1.30611336e+00 5.61167359e-01
-1.81609631e+00 -6.10391557e-01 -8.07978094e-01 2.36354399e+00
1.15444517e+00 -4.06408280e-01 1.88387170e-01 2.09588513e-01
2.88213521e-01 -4.11782295e-01 -9.62064117e-02 -3.37239027e-01
3.42462629e-01 4.35359627e-01 8.76468122e-01 1.21253514e+00
-9.64585662e-01 4.55144554e-01 6.65218592e+00 4.89557803e-01
-1.26754093e+00 7.07326651e-01 3.86642784e-01 1.83111235e-01
-3.97235435e-03 4.37142216e-02 -2.73533851e-01 3.42487425e-01
1.06579518e+00 -2.25145429e-01 5.23447633e-01 5.57024062e-01
3.83226037e-01 -4.72311348e-01 -7.68215716e-01 8.39356840e-01
-1.74960062e-01 -1.50081897e+00 -5.93281627e-01 4.34438229e-01
1.09102118e+00 1.71310991e-01 -2.56224841e-01 -2.85029292e-01
1.10285126e-01 -7.29362249e-01 4.03307825e-01 3.12612057e-01
7.53322184e-01 -4.82265115e-01 5.45598447e-01 3.35868835e-01
-1.21311247e+00 5.48551977e-01 1.68160826e-01 -3.00003320e-01
5.66158712e-01 7.19043911e-01 -6.79945275e-02 4.37007636e-01
4.94628280e-01 4.35816199e-01 1.06878921e-01 1.16027474e+00
2.62971431e-01 2.55067915e-01 -5.59764087e-01 5.06444275e-01
3.72058034e-01 -8.41018379e-01 8.53624165e-01 1.07106698e+00
3.86909068e-01 9.18527126e-01 1.83597177e-01 8.42604756e-01
3.82519811e-01 2.23136798e-01 -1.40027970e-01 7.20752239e-01
-1.67448282e-01 1.17171156e+00 -9.16764379e-01 -4.83333111e-01
-3.73118818e-01 1.15993226e+00 6.77542090e-02 4.53143567e-01
-6.01650178e-01 7.29768723e-02 7.82021821e-01 3.04505855e-01
-4.17274460e-02 -7.03762174e-01 -3.59259278e-01 -1.34509850e+00
-3.28642160e-01 -6.17413104e-01 4.51965988e-01 -4.64198858e-01
-1.09510386e+00 6.74576283e-01 2.51919001e-01 -9.20668662e-01
-2.28737369e-01 -5.93248248e-01 -6.39320731e-01 1.47720313e+00
-1.32959080e+00 -6.81180477e-01 -2.05621272e-02 9.73642290e-01
-8.11992809e-02 6.12761438e-01 9.02596831e-01 5.06324291e-01
-5.05339503e-02 2.72374302e-02 2.14581281e-01 5.78037836e-02
4.35745806e-01 -1.19198716e+00 3.41910198e-02 6.85965240e-01
-7.99251556e-01 6.65344238e-01 8.98505032e-01 -8.55757058e-01
-1.54349935e+00 -7.74011731e-01 1.06794143e+00 -2.63328850e-02
8.48733604e-01 -2.94829719e-02 -1.24329829e+00 8.86933386e-01
3.22937071e-01 6.10942662e-01 3.10719758e-01 -4.78575706e-01
5.22926271e-01 3.16795349e-01 -1.71103346e+00 6.93622977e-02
4.76377130e-01 -3.57837409e-01 -1.80627435e-01 8.09715927e-01
1.71853706e-01 -1.04666197e+00 -1.34827316e+00 2.76241601e-01
3.84557664e-01 -8.09093714e-01 8.93524706e-01 -9.50108916e-02
3.71595845e-02 -3.82920057e-01 2.52054393e-01 -1.27998400e+00
-3.64228487e-01 -1.08880448e+00 4.47584152e-01 4.98798788e-01
6.58155754e-02 -9.54119146e-01 4.40643251e-01 1.04287457e+00
-2.57424880e-02 -8.50576401e-01 -1.21648741e+00 -6.83900595e-01
4.85320657e-01 -1.75876424e-01 3.74747813e-02 1.36995602e+00
2.76632547e-01 -2.39867657e-01 -3.09090137e-01 1.59778595e-01
1.03655803e+00 -1.96053579e-01 1.62086010e-01 -1.00078785e+00
-3.33806843e-01 -8.39591324e-02 -3.10004354e-01 -7.00368643e-01
7.96636268e-02 -8.50872517e-01 5.70344329e-02 -1.27814054e+00
7.04078227e-02 -5.74934363e-01 2.41835475e-01 2.18964353e-01
9.50214937e-02 5.47431350e-01 -1.92337289e-01 2.41282701e-01
2.59736031e-01 1.02035224e-01 1.23132586e+00 2.66887486e-01
-3.56622070e-01 -3.12752992e-01 3.01658273e-01 7.00729370e-01
5.55904508e-01 -6.02887928e-01 -1.16719767e-01 -3.13204795e-01
2.38138005e-01 4.83804345e-01 6.06425285e-01 -8.87022018e-01
4.46464330e-01 -3.04079130e-02 -1.60874799e-01 -8.87736753e-02
2.40299091e-01 -4.90428418e-01 5.11958122e-01 1.03412676e+00
7.96592087e-02 2.54268825e-01 4.64842975e-01 -1.48734525e-01
3.57557535e-02 -1.53753281e-01 1.35014677e+00 -2.57437587e-01
-2.41152942e-01 3.39281976e-01 -4.91332680e-01 1.53791681e-01
7.59148955e-01 9.95809883e-02 2.38025993e-01 3.45102735e-02
-1.19329798e+00 -7.67349973e-02 7.99075663e-01 -5.71713865e-01
3.98802519e-01 -1.59261048e+00 -9.54007745e-01 4.24533635e-01
-6.26240313e-01 -3.65866005e-01 3.44141573e-01 1.62435520e+00
-1.07550347e+00 -7.93221742e-02 -1.80554748e-01 -6.99037194e-01
-1.07698596e+00 3.01646501e-01 9.79937255e-01 -4.66304541e-01
-5.83065391e-01 7.30260730e-01 4.07131054e-02 -3.73522788e-01
-4.70624030e-01 -2.97994733e-01 1.66866571e-01 -2.00775400e-01
4.26145703e-01 8.35540652e-01 1.64434403e-01 -1.05122173e+00
-4.98906165e-01 9.99927342e-01 5.48078239e-01 -5.89778721e-01
1.26221526e+00 -4.81901541e-02 -8.43634605e-01 2.15300083e-01
1.55400145e+00 -1.88380051e-02 -1.02995706e+00 -6.76332712e-02
-4.37206149e-01 -2.31410518e-01 4.78684753e-01 -3.18895638e-01
-1.08864367e+00 3.93184602e-01 5.46475530e-01 3.29767644e-01
9.36630070e-01 -4.71554935e-01 9.26650107e-01 -2.55083978e-01
3.58245045e-01 -8.74180496e-01 -7.68521786e-01 5.02313733e-01
9.01876330e-01 -1.01735032e+00 4.01382118e-01 -7.18378961e-01
-1.99406728e-01 9.42868412e-01 7.41536468e-02 -5.86912930e-01
1.02216899e+00 6.48013949e-01 6.30580410e-02 -1.70587003e-01
-2.10889980e-01 1.02756016e-01 3.89098495e-01 3.02356146e-02
7.19307303e-01 -1.33677229e-01 -1.12877226e+00 -1.74069509e-01
-1.94022916e-02 2.80282080e-01 4.54366416e-01 1.03505909e+00
1.71507344e-01 -1.12274599e+00 -5.05557179e-01 1.22035153e-01
-3.61565500e-01 -7.99287185e-02 2.83556819e-01 6.78100705e-01
8.87461007e-02 5.42242289e-01 -5.09701110e-02 2.91039705e-01
1.72464550e-01 1.25659555e-01 7.50553846e-01 -2.78626233e-01
-9.00927901e-01 2.03087077e-01 -9.03529599e-02 -9.22061026e-01
-5.95772803e-01 -1.03644788e+00 -1.67338264e+00 -4.51467037e-01
6.74074143e-02 6.07361317e-01 6.20893955e-01 1.06195664e+00
1.20316282e-01 3.31994921e-01 3.90037864e-01 -1.42478442e+00
-3.66004527e-01 -5.28418422e-01 -7.62862086e-01 3.44075322e-01
4.63971078e-01 -4.59743381e-01 -5.27294457e-01 9.88839418e-02] | [6.487897872924805, 3.296027183532715] |
4c06c2eb-5242-4624-bbf0-b7ef9ae491d2 | 3d-model-based-zero-shot-pose-estimation | 2305.17934 | null | https://arxiv.org/abs/2305.17934v1 | https://arxiv.org/pdf/2305.17934v1.pdf | 3D Model-based Zero-Shot Pose Estimation Pipeline | Most existing learning-based pose estimation methods are typically developed for non-zero-shot scenarios, where they can only estimate the poses of objects present in the training dataset. This setting restricts their applicability to unseen objects in the training phase. In this paper, we introduce a fully zero-shot pose estimation pipeline that leverages the 3D models of objects as clues. Specifically, we design a two-step pipeline consisting of 3D model-based zero-shot instance segmentation and a zero-shot pose estimator. For the first step, there is a novel way to perform zero-shot instance segmentation based on the 3D models instead of text descriptions, which can handle complex properties of unseen objects. For the second step, we utilize a hierarchical geometric structure matching mechanism to perform zero-shot pose estimation which is 10 times faster than the current render-based method. Extensive experimental results on the seven core datasets on the BOP challenge show that the proposed method outperforms the zero-shot state-of-the-art method with higher speed and lower computation cost. | ['Zhenyu He', 'Liwei Wu', 'Rui Zhao', 'Tianpeng Bao', 'Mingshan Sun', 'Jianqiu Chen'] | 2023-05-29 | null | null | null | null | ['pose-estimation'] | ['computer-vision'] | [-1.58194602e-01 -6.99993968e-02 -6.29875213e-02 -1.54395133e-01
-1.09260523e+00 -4.74330962e-01 4.74567533e-01 6.53914660e-02
-3.48602533e-01 -6.18077070e-02 -4.66578566e-02 1.75646544e-01
1.72044873e-01 -7.20812559e-01 -8.22576284e-01 -4.05839562e-01
3.71515900e-01 1.08795309e+00 1.09074485e+00 -1.16744138e-01
3.26911151e-01 3.26021492e-01 -1.66601813e+00 8.80426615e-02
5.24568260e-01 1.04415667e+00 3.10053945e-01 6.95368588e-01
-2.29324400e-01 2.23158672e-01 -2.21449569e-01 -2.84309983e-01
2.51734853e-01 -2.68878397e-02 -6.60137653e-01 4.20145720e-01
7.60576487e-01 -7.59456873e-01 -1.49685994e-01 7.32740521e-01
5.06011844e-01 2.96976089e-01 7.40365148e-01 -1.11223507e+00
6.02449141e-02 6.54911771e-02 -8.55877161e-01 -7.92279020e-02
4.23677713e-01 1.72772482e-01 1.09835315e+00 -1.32559228e+00
7.11641133e-01 1.40250242e+00 6.39248848e-01 4.75443274e-01
-1.16387928e+00 -4.75082457e-01 2.59245336e-01 1.62322789e-01
-1.58900559e+00 -3.26141864e-01 8.09230506e-01 -5.95136046e-01
8.93100202e-01 4.14138138e-02 7.44785309e-01 9.19901848e-01
-1.98436454e-01 1.07669544e+00 6.40653789e-01 -2.37093836e-01
4.66938257e-01 -2.58069545e-01 1.66938961e-01 8.21931660e-01
2.09071487e-01 -2.88764089e-01 -6.23653173e-01 -1.35423124e-01
6.38762176e-01 1.54195070e-01 8.21181908e-02 -1.10949063e+00
-1.23370123e+00 7.13860512e-01 3.61023933e-01 -1.38023183e-01
-2.18449280e-01 3.16123426e-01 2.58497268e-01 -3.40487152e-01
6.31242633e-01 2.03936160e-01 -4.68454063e-01 -9.10786092e-02
-1.12371290e+00 5.08731902e-01 9.17823374e-01 1.14600492e+00
8.96338344e-01 -4.06373143e-01 -2.85980880e-01 6.43444479e-01
5.07538319e-01 3.47532392e-01 6.39395267e-02 -7.16524065e-01
3.62496734e-01 6.18177056e-01 2.63707787e-01 -8.02713454e-01
-3.18392992e-01 -1.72383741e-01 -1.56736508e-01 -1.19252637e-01
3.30953419e-01 1.46989465e-01 -1.36085892e+00 1.24270451e+00
1.04090941e+00 6.98305845e-01 -3.33635986e-01 1.09168422e+00
1.04485440e+00 4.56299454e-01 -1.35240659e-01 1.74149901e-01
1.44608092e+00 -1.25493813e+00 -3.88238519e-01 -4.33725446e-01
7.75275648e-01 -6.70762718e-01 9.87265706e-01 1.67128280e-01
-8.18725169e-01 -3.92097086e-01 -1.15006971e+00 -2.05247581e-01
-3.33484530e-01 -5.25529981e-02 4.11696643e-01 4.24701005e-01
-4.98690486e-01 6.05572581e-01 -1.05359423e+00 -4.75954950e-01
5.40120125e-01 1.77538037e-01 -2.11730987e-01 -3.24101776e-01
-7.16945350e-01 6.52862668e-01 7.05438375e-01 -1.35998771e-01
-1.09458780e+00 -8.49029958e-01 -1.12849557e+00 4.52717096e-02
1.03397655e+00 -6.03698671e-01 1.37076938e+00 -1.58861130e-01
-1.29895914e+00 8.70486498e-01 -4.23587486e-02 -5.37407547e-02
6.46733344e-01 -6.51837885e-01 3.15860778e-01 3.27402085e-01
1.73309505e-01 7.33271837e-01 9.49450314e-01 -1.40608692e+00
-5.62835217e-01 -5.20578921e-01 1.14614122e-01 1.92958251e-01
-1.06478207e-01 -1.81722850e-01 -1.23483896e+00 -3.44518363e-01
5.63864887e-01 -1.04247737e+00 -3.14222038e-01 3.26796353e-01
-3.94603431e-01 -1.78508177e-01 1.18410528e+00 -4.11881477e-01
8.64970624e-01 -2.09392858e+00 1.46853417e-01 -7.05013350e-02
2.65100777e-01 3.16521257e-01 -2.83255372e-02 3.60477746e-01
4.11122799e-01 -1.17888518e-01 -1.00927897e-01 -8.17142129e-01
2.77214199e-01 1.02838755e-01 -1.99846134e-01 5.20121753e-01
3.77695590e-01 1.03959799e+00 -1.02447081e+00 -8.44174206e-01
5.63766003e-01 4.55789566e-01 -7.56029785e-01 2.87804306e-01
-5.64231753e-01 3.27093035e-01 -5.71973920e-01 8.43757808e-01
8.50296617e-01 -3.25560838e-01 -3.40247378e-02 -2.90980995e-01
3.29066962e-02 4.98220101e-02 -1.36933255e+00 2.11205220e+00
-2.91136771e-01 1.44525871e-01 -1.40956044e-01 -5.44597685e-01
6.36032999e-01 1.45015001e-01 5.16666114e-01 -1.61839902e-01
3.05916280e-01 8.52254257e-02 -4.02155042e-01 -3.70805472e-01
4.45233375e-01 4.51318212e-02 -1.83084220e-01 4.57507849e-01
2.96129435e-01 -4.06439573e-01 2.59167969e-01 2.91790247e-01
9.67501760e-01 7.13212252e-01 2.50364780e-01 6.17381074e-02
1.23771310e-01 8.90042484e-02 6.20096743e-01 6.60196543e-01
-1.92996681e-01 1.08695471e+00 3.05184484e-01 -3.01775485e-01
-9.16228414e-01 -9.73172903e-01 -2.18558777e-02 1.14506400e+00
6.39361858e-01 -8.15945387e-01 -8.40008199e-01 -6.78559422e-01
1.84569636e-03 5.08816242e-01 -5.55565894e-01 1.83059070e-02
-4.33125794e-01 -3.76432031e-01 2.95728929e-02 8.10584605e-01
3.82393479e-01 -5.50081849e-01 -9.63762760e-01 1.43523157e-01
-1.55851990e-01 -1.44499373e+00 -5.08370519e-01 -5.77738956e-02
-8.27407956e-01 -1.13077033e+00 -6.69313908e-01 -5.53689301e-01
6.20289862e-01 5.07163584e-01 9.05453980e-01 2.15882018e-01
-6.08336210e-01 5.08421957e-01 -6.22048736e-01 -4.59285825e-01
9.78135094e-02 1.11633427e-01 -6.92575648e-02 1.35858104e-01
3.56929690e-01 -3.37304205e-01 -8.08747530e-01 4.76573437e-01
-6.72425151e-01 1.47345304e-01 6.36709809e-01 5.04631698e-01
8.34086597e-01 -3.15139741e-01 4.82936278e-02 -7.61604607e-01
-7.42356330e-02 -3.24620426e-01 -4.80112761e-01 3.12278032e-01
-1.54210001e-01 4.94154654e-02 1.70797035e-01 -6.13667607e-01
-1.02862358e+00 6.50863171e-01 1.24661438e-02 -9.11767840e-01
-2.20758468e-01 1.22961774e-01 -3.77735585e-01 -5.80252632e-02
4.47833419e-01 -3.35826240e-02 -2.79990524e-01 -8.83303285e-01
4.92800444e-01 4.85025197e-01 4.36094016e-01 -5.72649598e-01
1.08821499e+00 6.02856576e-01 -3.71616445e-02 -8.20739508e-01
-1.18796790e+00 -1.10742545e+00 -1.20525444e+00 -2.94712663e-01
1.00661874e+00 -1.01427877e+00 -5.12729228e-01 4.96971756e-01
-1.20273411e+00 -1.82823464e-01 5.65327331e-03 4.83647436e-01
-7.56477654e-01 3.51923317e-01 -5.26310384e-01 -7.28612304e-01
-3.16857576e-01 -1.13248134e+00 1.85001802e+00 7.15459362e-02
-1.39226511e-01 -6.36253178e-01 4.32059728e-02 5.74446499e-01
-1.40136614e-01 2.74383783e-01 6.39958739e-01 -7.78743446e-01
-9.59708989e-01 -5.49763858e-01 -1.78528890e-01 -3.12837929e-01
-2.58435160e-01 -5.72397597e-02 -1.09315312e+00 -3.34896773e-01
-2.12305009e-01 -5.16479433e-01 7.83139348e-01 1.12596802e-01
1.07976139e+00 2.30610296e-01 -5.72793841e-01 7.27199554e-01
1.37719810e+00 -4.18681383e-01 4.02628332e-01 3.10807321e-02
9.98885989e-01 4.68420058e-01 1.20102656e+00 7.29682744e-01
3.36500555e-01 8.93887997e-01 6.84767306e-01 1.37120083e-01
-8.72358605e-02 -6.18142188e-01 -1.01913080e-01 6.70196056e-01
4.41123880e-02 -8.16569328e-02 -1.05618834e+00 5.67783535e-01
-2.07596755e+00 -6.48400605e-01 -1.74605280e-01 2.32204485e+00
4.55917478e-01 3.74371856e-01 1.97778672e-01 -8.33635181e-02
6.79763794e-01 3.22471082e-01 -5.52738547e-01 3.23608965e-01
6.07949257e-01 4.33030762e-02 3.00509065e-01 2.62148649e-01
-1.26079822e+00 1.28633690e+00 5.62719059e+00 8.10855389e-01
-7.76171327e-01 2.01483443e-01 1.12212360e-01 -2.63018668e-01
6.31083399e-02 2.80680954e-01 -1.09120941e+00 1.49725124e-01
4.52389628e-01 1.04916655e-01 -4.12643254e-02 1.24680042e+00
-2.41418555e-02 -3.35343838e-01 -1.33347750e+00 1.11329424e+00
2.12651432e-01 -1.17791879e+00 -1.47728011e-01 -7.19926804e-02
4.85274941e-01 -4.49806266e-02 -3.01730841e-01 5.16152799e-01
9.68441181e-03 -6.12036824e-01 8.42937171e-01 4.10357893e-01
6.59212768e-01 -7.91764677e-01 4.07598734e-01 8.20662379e-01
-1.55987573e+00 2.08924487e-01 -4.32821244e-01 1.28371626e-01
3.02582979e-01 3.53800893e-01 -9.45510626e-01 4.78199720e-01
7.34337509e-01 6.06171072e-01 -5.96961319e-01 1.27858174e+00
-2.31136635e-01 3.25379044e-01 -4.47701961e-01 1.77450366e-02
1.48775727e-01 1.44871548e-01 6.55477822e-01 9.35616016e-01
2.30170771e-01 2.03801289e-01 7.32656300e-01 8.03941429e-01
6.01974837e-02 7.79032484e-02 -3.20752591e-01 1.44560011e-02
6.06763005e-01 1.53452289e+00 -1.18345821e+00 -2.88831621e-01
-4.22950625e-01 1.13163030e+00 4.15357888e-01 -1.13036983e-01
-9.64169502e-01 -2.81249464e-01 4.14732814e-01 3.93953651e-01
7.39140809e-01 -4.35744554e-01 3.83031890e-02 -1.23837852e+00
3.80460382e-03 -4.72596467e-01 1.42757311e-01 -7.68117487e-01
-1.05234313e+00 7.24114850e-02 3.60954970e-01 -1.26156461e+00
-3.29265296e-01 -4.79015738e-01 -6.24626756e-01 4.49301362e-01
-1.13754702e+00 -1.59011245e+00 -5.47011256e-01 2.82337129e-01
9.08048153e-01 4.42916572e-01 6.89359009e-01 9.14130509e-02
-5.13622046e-01 3.00523400e-01 -2.23139599e-01 1.78223372e-01
5.54428220e-01 -1.08901954e+00 6.61632895e-01 7.72906363e-01
2.58579940e-01 4.47374046e-01 6.65734947e-01 -9.12672102e-01
-1.67639923e+00 -1.12831545e+00 4.61787492e-01 -6.41446769e-01
4.97799754e-01 -7.88219512e-01 -9.81238782e-01 6.35335684e-01
-4.42647994e-01 3.80241215e-01 4.87085283e-01 1.72041759e-01
-4.53201711e-01 2.56255865e-01 -8.16847563e-01 6.17314160e-01
1.15863466e+00 -3.15399021e-01 -6.50315583e-01 3.28080118e-01
9.13059115e-01 -7.37352669e-01 -6.56187773e-01 4.74668711e-01
5.71101010e-01 -6.47329986e-01 1.07780230e+00 -5.50611675e-01
2.59912223e-01 -4.27404016e-01 -1.84526205e-01 -8.79219830e-01
-1.40251100e-01 -4.92610157e-01 -5.36883473e-01 1.01831865e+00
1.19275384e-01 8.17488059e-02 1.01065791e+00 5.88100433e-01
-1.59337953e-01 -9.22663927e-01 -6.90545499e-01 -8.06262851e-01
-3.98623377e-01 -5.78496277e-01 3.90933096e-01 4.92041469e-01
-3.30415308e-01 5.02139807e-01 -3.96757603e-01 2.52862483e-01
8.48557293e-01 3.13157469e-01 1.22653568e+00 -1.36897182e+00
-4.20847088e-01 7.74388686e-02 -7.11984038e-01 -1.36228359e+00
1.16984501e-01 -5.58771551e-01 5.45134842e-01 -1.63062370e+00
5.03480256e-01 -9.17029828e-02 3.92568111e-02 3.64111483e-01
-2.52166241e-01 2.83803821e-01 3.26902092e-01 1.94623619e-01
-1.07269228e+00 5.46237469e-01 1.11393750e+00 7.22246096e-02
-2.20679954e-01 -8.85881856e-02 -1.67968929e-01 1.09634674e+00
2.96221346e-01 -6.23512566e-01 -4.95177358e-01 -3.07033747e-01
-1.03562929e-01 -5.33086471e-02 5.53523719e-01 -1.20095921e+00
2.85874695e-01 -1.38248757e-01 4.09634829e-01 -1.33482420e+00
8.38237226e-01 -7.48023689e-01 -1.58616111e-01 3.63296211e-01
1.10681362e-01 -4.23285127e-01 6.88290573e-04 9.51231182e-01
2.63890505e-01 -3.51451367e-01 8.57568085e-01 -2.64134258e-01
-1.02015817e+00 7.35080004e-01 1.04825310e-01 1.90118268e-01
1.41315162e+00 -3.79166931e-01 2.88271010e-02 -1.14824675e-01
-7.12071538e-01 3.86107326e-01 6.92619205e-01 5.36381006e-01
7.77520835e-01 -1.18470061e+00 -4.99287993e-01 -3.73494588e-02
4.65456307e-01 6.19423807e-01 3.38585228e-01 9.35525596e-01
-5.67435384e-01 1.71078026e-01 1.45360634e-01 -9.56695259e-01
-1.30482519e+00 8.05364847e-01 1.97179288e-01 -4.31193747e-02
-8.90462637e-01 9.37813282e-01 5.14361203e-01 -3.29621643e-01
1.90847203e-01 3.01238522e-02 1.29968837e-01 1.60190463e-01
5.32659233e-01 3.66951734e-01 1.20450914e-01 -7.74947047e-01
-4.86026227e-01 8.12965572e-01 -4.07219112e-01 -1.71165287e-01
1.37191856e+00 -2.59287655e-02 3.16669077e-01 7.12717116e-01
1.17307127e+00 -3.14020634e-01 -1.56902647e+00 -4.40004498e-01
1.45443738e-01 -8.21913779e-01 1.65758848e-01 -2.42493957e-01
-8.09905171e-01 1.04273582e+00 3.21632445e-01 -3.86614144e-01
6.26026690e-01 3.65366936e-01 1.04561543e+00 4.43583369e-01
6.25207961e-01 -1.34400082e+00 5.48591435e-01 5.19687891e-01
6.55668736e-01 -1.28911757e+00 7.56892115e-02 -8.73226881e-01
-5.10326326e-01 9.68722701e-01 8.88738275e-01 -1.50233030e-01
6.15945995e-01 1.03984229e-01 -2.71167785e-01 -3.19338918e-01
-7.24257946e-01 -5.44647098e-01 5.58836162e-01 4.21968818e-01
1.05230562e-01 -2.37794027e-01 -3.83127220e-02 5.59315205e-01
1.04935169e-01 -2.06366405e-01 1.42902941e-01 1.35030663e+00
-8.18401694e-01 -8.80172729e-01 -4.69575763e-01 2.71149397e-01
-8.45970511e-02 1.42233402e-01 -5.70198298e-01 8.37072194e-01
1.84088349e-02 7.07433105e-01 1.21435314e-01 -4.65995193e-01
3.92745912e-01 1.98223710e-01 6.57830358e-01 -1.14106941e+00
-1.19570985e-01 3.69780898e-01 -1.29983231e-01 -9.48509395e-01
-1.88107878e-01 -6.55082643e-01 -1.41318917e+00 3.41460966e-02
-6.12177491e-01 -3.52593273e-01 7.12290227e-01 1.02076840e+00
3.45152885e-01 3.36557388e-01 3.22043210e-01 -1.46433592e+00
-7.24203587e-01 -8.38489175e-01 -4.31911767e-01 5.23289561e-01
2.74070110e-02 -1.17300880e+00 -2.33684629e-01 -1.11351386e-01] | [7.562166213989258, -2.62715220451355] |
dde4efab-7112-4bfc-915b-7b05bacd4596 | nearest-neighbor-machine-translation-is-meta | 2305.13034 | null | https://arxiv.org/abs/2305.13034v1 | https://arxiv.org/pdf/2305.13034v1.pdf | Nearest Neighbor Machine Translation is Meta-Optimizer on Output Projection Layer | Nearest Neighbor Machine Translation ($k$NN-MT) has achieved great success on domain adaptation tasks by integrating pre-trained Neural Machine Translation (NMT) models with domain-specific token-level retrieval. However, the reasons underlying its success have not been thoroughly investigated. In this paper, we provide a comprehensive analysis of $k$NN-MT through theoretical and empirical studies. Initially, we offer a theoretical interpretation of the working mechanism of $k$NN-MT as an efficient technique to implicitly execute gradient descent on the output projection layer of NMT, indicating that it is a specific case of model fine-tuning. Subsequently, we conduct multi-domain experiments and word-level analysis to examine the differences in performance between $k$NN-MT and entire-model fine-tuning. Our findings suggest that: (1) Incorporating $k$NN-MT with adapters yields comparable translation performance to fine-tuning on in-domain test sets, while achieving better performance on out-of-domain test sets; (2) Fine-tuning significantly outperforms $k$NN-MT on the recall of low-frequency domain-specific words, but this gap could be bridged by optimizing the context representations with additional adapter layers. | ['Rui Wang', 'Lemao Liu', 'Yichao Du', 'Zhirui Zhang', 'Ruize Gao'] | 2023-05-22 | null | null | null | null | ['nmt'] | ['computer-code'] | [ 2.03202084e-01 -1.58157840e-01 -6.44883692e-01 -4.67289835e-01
-1.42376709e+00 -7.55663335e-01 7.77553082e-01 7.46433586e-02
-6.76298857e-01 6.47146344e-01 5.24719000e-01 -7.70473480e-01
-7.66078606e-02 -6.28762782e-01 -8.29029441e-01 -1.48775637e-01
3.38843048e-01 6.92550480e-01 -1.44957334e-01 -5.43210685e-01
3.25419605e-01 1.58140361e-01 -7.17241585e-01 6.10944033e-01
9.84185457e-01 4.62046176e-01 3.15227240e-01 3.14529061e-01
-3.10005397e-01 2.87421346e-02 -5.48019230e-01 -7.78938115e-01
2.30556726e-01 -6.03857338e-01 -8.18001211e-01 -4.39377338e-01
4.94202495e-01 -2.81044662e-01 -1.38377100e-01 7.39903808e-01
6.10496998e-01 1.55711874e-01 9.28696215e-01 -6.44271612e-01
-1.37826896e+00 6.65145218e-01 -2.90258348e-01 4.78313535e-01
1.21729255e-01 2.29777515e-01 1.05707633e+00 -1.40344334e+00
7.15481818e-01 1.13989830e+00 8.43356788e-01 5.88582814e-01
-1.53034556e+00 -7.58517206e-01 1.81879699e-02 1.61882192e-01
-1.52485621e+00 -5.28634369e-01 4.38960195e-01 -2.11888194e-01
1.81239176e+00 -1.36514217e-01 4.38331932e-01 1.23902369e+00
5.61813891e-01 6.19459808e-01 1.13667011e+00 -7.84898043e-01
2.81599313e-02 3.61584157e-01 -3.06567119e-04 2.77445585e-01
1.94921553e-01 2.39053860e-01 -7.85372078e-01 -2.73501515e-01
7.69436657e-01 -4.16396499e-01 -6.74218684e-03 2.05776971e-02
-1.19096255e+00 9.66329634e-01 2.23460913e-01 4.87549067e-01
-3.30865979e-01 -4.70609628e-02 4.73170310e-01 6.66471303e-01
7.02501953e-01 8.09263289e-01 -8.62513661e-01 -2.34468102e-01
-1.08029282e+00 1.22036859e-01 4.94633108e-01 1.01923954e+00
8.22498441e-01 -2.55662296e-02 -2.75101602e-01 1.36911798e+00
5.17927893e-02 7.16040671e-01 8.98810506e-01 -6.77148819e-01
8.07620585e-01 3.74347448e-01 1.77236572e-02 -5.09592533e-01
-1.08795036e-02 -4.96051878e-01 -4.09836888e-01 -4.07545656e-01
4.06787544e-01 -1.90203607e-01 -7.05577135e-01 2.01366186e+00
-2.04292491e-01 -2.58340716e-01 1.95693821e-01 8.65417719e-01
2.77728140e-01 7.96123028e-01 4.08762485e-01 -1.53889120e-01
1.25474536e+00 -9.85223889e-01 -5.92059970e-01 -6.98011935e-01
1.09559417e+00 -1.13811243e+00 1.61872613e+00 -1.69493571e-01
-1.03518391e+00 -5.96966326e-01 -8.88260424e-01 -2.79630721e-01
-5.83652854e-01 2.52618521e-01 4.25109953e-01 6.67096317e-01
-1.11807215e+00 5.55229783e-01 -5.64390421e-01 -8.39376986e-01
7.11120442e-02 2.93172151e-01 -3.24578643e-01 -3.60281855e-01
-1.64484084e+00 1.63422751e+00 2.13902459e-01 -3.02296400e-01
-5.33515334e-01 -6.83258176e-01 -6.90787256e-01 -1.24122947e-01
-2.55302548e-01 -1.00633323e+00 1.33230233e+00 -9.13470805e-01
-1.51852596e+00 9.80375230e-01 -6.78059757e-01 -2.07874358e-01
-1.12137258e-01 -3.11178416e-01 -5.55540025e-01 -2.13277012e-01
3.69880527e-01 7.00501442e-01 6.63714886e-01 -6.88080549e-01
-5.90379715e-01 -4.29944336e-01 -2.39158899e-01 5.27568102e-01
-4.73950505e-01 2.30859607e-01 -1.64604723e-01 -8.45313787e-01
1.12426788e-01 -1.01282668e+00 -4.94272402e-03 -5.30132115e-01
2.26660520e-01 -4.96064067e-01 3.90580982e-01 -6.42657280e-01
1.46719110e+00 -2.04223466e+00 -1.48852676e-01 -2.48440318e-02
-3.15488338e-01 5.21000862e-01 -5.90255558e-01 8.42367411e-01
-2.22160723e-02 1.78760841e-01 -2.32041106e-01 -1.52005687e-01
7.12855607e-02 1.50983974e-01 -3.95807832e-01 1.62009165e-01
4.75795239e-01 1.37152135e+00 -7.26941288e-01 -3.63926440e-01
-4.84739654e-02 3.98832470e-01 -5.88663340e-01 -1.38941035e-01
-1.98864236e-01 6.58890381e-02 -2.84948766e-01 5.44440746e-01
3.18472028e-01 -1.61681250e-01 3.14452320e-01 1.21391177e-01
-6.51921183e-02 9.76968348e-01 -4.93329972e-01 1.93034649e+00
-6.09380782e-01 7.80852437e-01 -3.11229318e-01 -9.00537252e-01
7.96252012e-01 3.35345060e-01 -5.35600446e-02 -1.19079041e+00
-2.71897435e-01 7.79516280e-01 8.52169022e-02 -2.72741795e-01
6.78200126e-01 -6.30103827e-01 -1.24458551e-01 6.93362951e-01
8.20708349e-02 -1.38224825e-01 -1.09684013e-01 -3.41676809e-02
8.93959880e-01 2.78779924e-01 2.02617511e-01 -3.63636225e-01
2.52007782e-01 5.93335807e-01 4.61135685e-01 6.61393404e-01
-1.73392057e-01 3.30220938e-01 -1.40143037e-01 -1.87269330e-01
-1.26897097e+00 -1.07153451e+00 -2.77819753e-01 1.50034332e+00
-1.92133784e-01 -3.47683877e-01 -9.15153921e-01 -6.50143921e-01
-5.42535968e-02 1.20643067e+00 -3.71606797e-01 -4.43875581e-01
-9.53764737e-01 -9.05991137e-01 8.12817395e-01 5.43613136e-01
3.54160219e-01 -9.31396127e-01 5.32174949e-03 4.40705031e-01
-6.99651003e-01 -8.49135458e-01 -9.37749386e-01 2.98716813e-01
-1.34233057e+00 -3.81924242e-01 -8.76553476e-01 -1.01550841e+00
4.36084628e-01 4.81187433e-01 1.28008342e+00 -3.49579901e-01
3.57641518e-01 2.64096081e-01 -3.07819694e-01 -2.51820832e-01
-5.27939558e-01 6.04610026e-01 2.46221140e-01 -5.29797077e-01
1.22022200e+00 -6.10235929e-01 -5.05238831e-01 5.27303636e-01
-7.44923234e-01 -3.66737515e-01 1.20061255e+00 1.08580196e+00
6.57609224e-01 -4.34188992e-01 8.23896646e-01 -7.69674599e-01
1.26308441e+00 -4.69820619e-01 -1.56503662e-01 2.80078232e-01
-1.19617486e+00 1.67713389e-01 4.38501477e-01 -7.40628958e-01
-1.02502072e+00 -4.95806932e-01 -1.30334154e-01 -1.87837690e-01
1.05795555e-01 7.85031259e-01 9.33142900e-02 1.32875144e-01
1.11560130e+00 5.56891680e-01 -2.01787546e-01 -4.83603418e-01
4.83772844e-01 8.32092583e-01 5.94903938e-02 -9.17051136e-01
7.01314151e-01 -3.62036414e-02 -6.17901444e-01 -5.85714877e-01
-7.26709008e-01 -3.29408020e-01 -5.98849595e-01 3.83884788e-01
6.89110756e-01 -1.00433028e+00 2.00983703e-01 -8.06909334e-03
-1.13181376e+00 -6.08790100e-01 -6.74583167e-02 7.85159886e-01
-6.12897635e-01 7.08263218e-02 -8.21129084e-01 -2.54313171e-01
-5.14471829e-01 -1.10021555e+00 1.05645907e+00 -2.04560861e-01
-7.42157876e-01 -1.05718827e+00 4.00586188e-01 4.10295695e-01
7.37196147e-01 -6.76503956e-01 1.47054970e+00 -8.40424001e-01
-3.11415702e-01 -3.11418809e-02 -2.45790184e-01 4.32183981e-01
2.41834179e-01 -5.58571279e-01 -7.35033393e-01 -3.65119189e-01
6.79669976e-02 -1.45283073e-01 5.70405722e-01 4.44792300e-01
3.92049104e-01 -3.62355947e-01 -4.12729770e-01 5.24208605e-01
1.39782929e+00 2.01084375e-01 4.41488832e-01 6.37407005e-01
1.77099496e-01 4.46839482e-01 7.55127192e-01 -2.17034191e-01
4.02858704e-01 9.49549198e-01 -4.11695063e-01 1.00533947e-01
-4.17475790e-01 -5.31193972e-01 6.70559466e-01 1.08152640e+00
1.89112455e-01 8.78482386e-02 -9.12713826e-01 7.93068588e-01
-1.56516385e+00 -7.78396130e-01 1.29724950e-01 2.21497512e+00
1.06148911e+00 1.13305919e-01 -7.33793480e-03 -4.67328191e-01
7.61702716e-01 -6.82741478e-02 -5.92723191e-01 -9.56723154e-01
-2.12535441e-01 5.11907697e-01 4.76490855e-01 4.50069189e-01
-6.09094024e-01 1.39171684e+00 7.10141420e+00 1.09000885e+00
-1.12255442e+00 4.72847134e-01 3.70507300e-01 -1.49070725e-01
-3.80161107e-01 6.46514371e-02 -1.04192674e+00 3.71394217e-01
1.43125844e+00 -2.92181522e-01 5.54118097e-01 6.73568547e-01
2.48624414e-01 3.00261647e-01 -1.34614301e+00 6.28334820e-01
-9.78367683e-03 -1.28848028e+00 3.18112403e-01 2.04516262e-01
9.35113251e-01 4.78886038e-01 3.01290244e-01 7.56510437e-01
2.49412015e-01 -8.51439834e-01 4.61583108e-01 -5.66507690e-02
1.13633466e+00 -6.73046410e-01 5.32067955e-01 5.04065633e-01
-6.94585562e-01 1.05075397e-01 -6.27103269e-01 -1.37752473e-01
-1.18945755e-01 3.82353872e-01 -1.18151712e+00 2.75219560e-01
4.62048531e-01 2.66035765e-01 -3.85767281e-01 4.33842927e-01
-2.04839230e-01 8.85287523e-01 -8.41554925e-02 -6.81657419e-02
5.34865916e-01 -1.14410363e-01 4.83970374e-01 1.73302197e+00
5.05048096e-01 7.36624673e-02 -3.62179577e-01 8.00852001e-01
-1.90149486e-01 4.19662416e-01 -7.13767350e-01 -2.54893452e-01
7.20477581e-01 7.14360952e-01 -1.56498492e-01 -3.78080845e-01
-6.22641861e-01 1.16573083e+00 6.94086909e-01 5.60767412e-01
-5.10011494e-01 -3.29666346e-01 1.08443713e+00 2.34464720e-01
2.68979490e-01 -4.87570405e-01 -6.22897387e-01 -1.22792244e+00
6.63876161e-02 -1.03078032e+00 3.12900335e-01 -7.69877493e-01
-1.39462829e+00 5.08887827e-01 -1.55703649e-01 -1.19881999e+00
-5.57983696e-01 -6.44717634e-01 -7.48354420e-02 1.49920094e+00
-1.43135619e+00 -1.09435165e+00 5.90441227e-01 5.86181641e-01
8.34716618e-01 -1.34784386e-01 1.28636754e+00 2.97796547e-01
-3.88579786e-01 1.16428041e+00 6.05606496e-01 2.38412216e-01
1.30650759e+00 -8.89808536e-01 8.30392957e-01 7.68849552e-01
2.07273439e-01 1.38922930e+00 4.75820094e-01 -7.54969478e-01
-1.29325438e+00 -1.14080179e+00 1.63533342e+00 -8.69417906e-01
4.68706965e-01 -2.09160194e-01 -9.57598090e-01 8.85370016e-01
3.83013278e-01 -4.59538609e-01 9.91898060e-01 4.24546152e-01
-7.05624282e-01 5.03125563e-02 -1.19491374e+00 8.29574764e-01
1.06839144e+00 -1.03159702e+00 -1.16483283e+00 2.59741515e-01
8.69987071e-01 -1.57505810e-01 -1.02540040e+00 2.50361353e-01
6.46935642e-01 -4.70371366e-01 8.98410082e-01 -7.08360255e-01
5.59766829e-01 -3.20711844e-02 -5.01703858e-01 -1.68076360e+00
-6.85398698e-01 -2.87420183e-01 2.25365192e-01 1.03942895e+00
9.41631079e-01 -8.45720589e-01 5.63000023e-01 5.22470117e-01
-3.05684507e-01 -7.24967301e-01 -1.05030668e+00 -9.28217769e-01
9.45115626e-01 -5.52662730e-01 6.17179155e-01 1.20073700e+00
2.40768477e-01 6.32827878e-01 -2.76104845e-02 -9.89934709e-03
2.76432514e-01 -1.17047749e-01 4.97032464e-01 -5.91859341e-01
-3.24946135e-01 -5.41812718e-01 1.08574063e-01 -1.26452732e+00
2.10832939e-01 -1.12257707e+00 -1.90080807e-01 -1.54816568e+00
2.12945312e-01 -3.22130919e-01 -3.55775386e-01 3.97204816e-01
-3.13361555e-01 3.10106009e-01 -8.73297155e-02 6.63548172e-01
-6.13935664e-02 3.83153707e-01 1.20187879e+00 -5.57632446e-02
-3.21069598e-01 -7.55934045e-02 -1.02454197e+00 2.29157969e-01
7.32985795e-01 -5.94753087e-01 -3.62708986e-01 -1.07185781e+00
9.03716534e-02 1.87212117e-02 -2.15049222e-01 -3.40510428e-01
1.27994284e-01 -3.45740706e-01 5.24685383e-01 -2.71033823e-01
3.42918336e-01 -4.75391924e-01 -1.24525204e-01 2.34802037e-01
-5.68237305e-01 3.79247546e-01 5.28057694e-01 2.33869448e-01
-1.30237758e-01 -1.74136162e-01 6.63499594e-01 -3.13934237e-01
-7.20255852e-01 -1.58245742e-01 -7.05760956e-01 1.58785462e-01
3.66275996e-01 -4.28254694e-01 -3.29374373e-01 -2.30743989e-01
-4.20269072e-01 -1.61232278e-01 6.03024542e-01 6.50988221e-01
4.95436102e-01 -1.54022598e+00 -9.25293565e-01 3.16515237e-01
4.59643811e-01 -5.92011333e-01 -1.26379132e-01 6.91925347e-01
2.21214201e-02 1.07032943e+00 1.33098913e-02 -4.60741758e-01
-7.96016872e-01 3.76346022e-01 3.05698156e-01 -4.25485849e-01
-2.35324576e-01 8.68436575e-01 2.09535390e-01 -8.67198467e-01
-2.55722016e-01 -1.41431764e-01 3.77889752e-01 -1.88156739e-01
3.40923756e-01 2.24287763e-01 2.83811927e-01 -5.80750704e-01
-3.93763870e-01 5.82014978e-01 -4.38572437e-01 -6.09460652e-01
1.01239312e+00 -3.68486434e-01 -5.17977849e-02 3.32458228e-01
1.28047943e+00 -1.04441501e-01 -6.85284555e-01 -6.85980201e-01
6.24328628e-02 -2.76052713e-01 -7.70237818e-02 -1.31457138e+00
-3.45511943e-01 8.52722883e-01 5.00661314e-01 -5.73259592e-01
1.02025676e+00 -9.17716250e-02 9.94007170e-01 5.36202312e-01
5.13699591e-01 -1.29116416e+00 -1.21684849e-01 8.08770537e-01
6.41789138e-01 -1.14475298e+00 -3.43174189e-01 6.44271225e-02
-5.12329102e-01 7.55331814e-01 6.43255949e-01 5.21983504e-02
3.54785770e-01 -2.79733926e-01 4.00817335e-01 6.94770962e-02
-8.56210768e-01 5.28328754e-02 3.57506692e-01 5.94383478e-01
8.86147320e-01 1.47694230e-01 -6.96514487e-01 5.84428668e-01
-4.04253721e-01 1.46124717e-02 -3.27458307e-02 7.16437936e-01
-4.36161369e-01 -1.50514936e+00 -4.27659780e-01 4.39840883e-01
-5.13158560e-01 -6.90531552e-01 -5.64867735e-01 7.80510247e-01
-1.38749167e-01 1.09803116e+00 -1.09084122e-01 -4.61180925e-01
5.04413545e-01 7.77169049e-01 5.14455914e-01 -8.35912764e-01
-8.06163073e-01 1.20906748e-01 1.73120052e-01 -3.49889070e-01
-4.93769385e-02 -7.50198781e-01 -8.97823334e-01 -3.37901175e-01
-3.52128059e-01 3.20861757e-01 8.47294450e-01 1.04132593e+00
7.21982539e-01 4.40060683e-02 3.42390627e-01 -3.97054970e-01
-8.68799627e-01 -1.36909688e+00 -1.95713297e-01 2.17925742e-01
8.09535608e-02 -2.81661063e-01 -2.79611826e-01 -1.60304576e-01] | [11.640803337097168, 10.135363578796387] |
88af0b1f-5713-48bf-923c-84d21e5cc5df | multispider-towards-benchmarking-multilingual | 2212.13492 | null | https://arxiv.org/abs/2212.13492v1 | https://arxiv.org/pdf/2212.13492v1.pdf | MultiSpider: Towards Benchmarking Multilingual Text-to-SQL Semantic Parsing | Text-to-SQL semantic parsing is an important NLP task, which greatly facilitates the interaction between users and the database and becomes the key component in many human-computer interaction systems. Much recent progress in text-to-SQL has been driven by large-scale datasets, but most of them are centered on English. In this work, we present MultiSpider, the largest multilingual text-to-SQL dataset which covers seven languages (English, German, French, Spanish, Japanese, Chinese, and Vietnamese). Upon MultiSpider, we further identify the lexical and structural challenges of text-to-SQL (caused by specific language properties and dialect sayings) and their intensity across different languages. Experimental results under three typical settings (zero-shot, monolingual and multilingual) reveal a 6.1% absolute drop in accuracy in non-English languages. Qualitative and quantitative analyses are conducted to understand the reason for the performance drop of each language. Besides the dataset, we also propose a simple schema augmentation framework SAVe (Schema-Augmentation-with-Verification), which significantly boosts the overall performance by about 1.8% and closes the 29.5% performance gap across languages. | ['Jian-Guang Lou', 'Dechen Zhan', 'Wanxiang Che', 'Dingzirui Wang', 'Mingyang Pan', 'Yan Gao', 'Longxu Dou'] | 2022-12-27 | null | null | null | null | ['text-to-sql', 'semantic-parsing'] | ['computer-code', 'natural-language-processing'] | [-7.28324950e-02 -4.69965599e-02 -3.88448775e-01 -5.71622312e-01
-1.18195796e+00 -8.15125406e-01 5.41386187e-01 3.95419747e-01
-4.87384886e-01 6.02023244e-01 4.78650957e-01 -5.97144127e-01
2.37884358e-01 -6.40069067e-01 -8.25278223e-01 9.35201421e-02
5.17238796e-01 6.63425744e-01 2.40290478e-01 -5.63205779e-01
-2.91645620e-02 6.12289719e-02 -1.23321867e+00 8.24576795e-01
1.14033175e+00 6.85451388e-01 2.05342963e-01 1.99716896e-01
-8.40072632e-01 8.20534468e-01 -4.52837646e-01 -9.13009644e-01
1.11678213e-01 -8.52819979e-02 -7.87105739e-01 -2.15611339e-01
3.78569067e-01 1.32515922e-01 1.68466181e-01 1.02123642e+00
3.76870245e-01 -5.51083922e-01 -1.01625472e-01 -1.21091568e+00
-5.92595935e-01 9.82937813e-01 -4.84432310e-01 -2.24140719e-01
7.62372077e-01 -2.89785918e-02 1.06660271e+00 -1.03086174e+00
1.07455432e+00 1.34556031e+00 8.07850778e-01 2.35784739e-01
-1.01231790e+00 -5.91339827e-01 -1.67864127e-04 -1.67173162e-01
-1.51332033e+00 -5.42380691e-01 3.75212938e-01 -4.29243743e-01
1.20249212e+00 3.39959621e-01 7.67704248e-02 1.13663924e+00
-9.61413682e-02 7.22867906e-01 1.18884778e+00 -7.13771760e-01
-1.68123305e-01 6.49433494e-01 9.29988325e-02 6.06791437e-01
2.42055878e-01 -3.66247475e-01 -9.14475501e-01 -1.04129404e-01
2.22897321e-01 -3.41934055e-01 1.52994851e-02 -4.44804817e-01
-1.31088531e+00 7.58791447e-01 -9.02634263e-02 2.66220152e-01
-1.54259473e-01 -4.82877642e-01 8.33699703e-01 3.50974113e-01
4.02626902e-01 3.26005459e-01 -1.00179470e+00 -4.10826951e-01
-5.69133520e-01 3.36124957e-01 1.06530392e+00 1.55772233e+00
7.14247763e-01 -1.87991381e-01 2.82267749e-01 1.12811053e+00
1.25002310e-01 7.09044278e-01 3.66536915e-01 -5.61152577e-01
1.24854696e+00 1.12717378e+00 9.64442734e-03 -5.45209706e-01
-4.16563630e-01 -1.42415419e-01 -4.10155505e-01 -5.13957858e-01
4.71822590e-01 -4.99146730e-02 -4.77710485e-01 1.71355653e+00
3.69288236e-01 -8.47790480e-01 3.64980996e-01 5.12165606e-01
8.78919780e-01 3.99112910e-01 2.69883007e-01 -2.53407117e-02
1.79027820e+00 -8.58933866e-01 -7.13741183e-01 -5.62831998e-01
9.80007768e-01 -1.01770675e+00 1.73914444e+00 7.09310994e-02
-8.89590621e-01 -4.88981932e-01 -7.66541600e-01 -4.03134972e-01
-8.36490810e-01 2.73751855e-01 5.25242388e-01 7.74304390e-01
-7.61906207e-01 -9.97272581e-02 -7.44027138e-01 -8.94546747e-01
3.60146314e-02 -4.11165456e-05 -5.59020519e-01 -2.86620408e-01
-1.16430998e+00 8.15957487e-01 7.45912552e-01 -4.06316072e-01
-3.55924070e-02 -9.00907099e-01 -8.72828424e-01 -3.13267350e-01
9.31764424e-01 -4.29642230e-01 1.14622331e+00 -6.09145224e-01
-1.14359343e+00 1.00368702e+00 -4.69074339e-01 -2.10947499e-01
5.67547083e-01 -4.85834330e-01 -7.89455891e-01 -4.11027849e-01
5.84735334e-01 4.50616807e-01 -5.23679592e-02 -9.38738167e-01
-7.05226719e-01 -6.33114636e-01 -1.62910864e-01 1.08820252e-01
-1.38547942e-01 7.38582313e-01 -1.05883169e+00 -7.16593027e-01
2.03462884e-01 -9.03866529e-01 1.25667408e-01 -3.76712799e-01
-5.56462824e-01 -2.09828019e-01 5.02680600e-01 -9.85828519e-01
1.45190477e+00 -2.17483187e+00 -1.59697622e-01 2.64388844e-02
-1.70473829e-01 -8.33157822e-03 3.62125188e-02 9.28685606e-01
3.29932496e-02 3.32294792e-01 -2.96021312e-01 -2.18916371e-01
-3.53086442e-02 2.14181513e-01 -3.37328911e-01 -1.56014159e-01
3.68743360e-01 1.06013572e+00 -5.06637156e-01 -6.09086156e-01
-1.11949503e-01 1.81441486e-01 -6.07409835e-01 9.20111164e-02
-3.83554667e-01 1.68741435e-01 -3.46788973e-01 1.07660782e+00
5.97754121e-01 -8.57206713e-03 6.02080405e-01 -3.13645452e-01
-5.20166636e-01 5.96807361e-01 -1.12487721e+00 1.96972215e+00
-4.59935248e-01 1.73016757e-01 7.33274668e-02 -5.45706391e-01
7.65201509e-01 1.10945471e-01 2.96409905e-01 -1.08204758e+00
-3.32127690e-01 6.73599362e-01 -2.37798616e-01 -6.73198044e-01
7.53707230e-01 6.98841810e-02 -6.27887011e-01 2.82210112e-01
-6.26033023e-02 -7.63373624e-04 4.54022735e-01 3.43171597e-01
7.01378763e-01 1.70733228e-01 4.33243603e-01 -3.52791995e-01
5.42506337e-01 3.78029406e-01 7.38598704e-01 5.54768443e-01
1.82917342e-02 3.23159486e-01 6.88939631e-01 -2.68999308e-01
-1.15568304e+00 -9.58402216e-01 -1.38384715e-01 1.28640473e+00
-3.33859660e-02 -8.54587793e-01 -7.42795169e-01 -8.28500152e-01
1.74849734e-01 9.72115457e-01 -2.92964607e-01 2.23941937e-01
-6.82412863e-01 -6.54405355e-01 9.32356417e-01 5.13031006e-01
5.29073417e-01 -9.74346876e-01 -3.00057471e-01 2.18051508e-01
-6.06336117e-01 -1.81592584e+00 -3.23958725e-01 3.19989718e-04
-6.39613867e-01 -1.02901208e+00 -4.41621207e-02 -6.58915699e-01
2.27708369e-01 -1.17746562e-01 1.23925960e+00 -3.46602768e-01
-9.32248905e-02 1.84196100e-01 -3.16587418e-01 -5.51818132e-01
-6.63488925e-01 3.90715331e-01 -1.79994866e-01 -3.13538700e-01
7.16405034e-01 1.64479956e-01 5.32730035e-02 3.21164846e-01
-8.80536973e-01 2.82468170e-01 5.49124062e-01 5.50438583e-01
6.94262624e-01 -2.16027454e-01 3.75737756e-01 -1.57361257e+00
4.96916950e-01 -6.02452457e-01 -5.60112894e-01 6.96201503e-01
-7.17415750e-01 1.69889659e-01 5.31867385e-01 -4.51298542e-02
-1.29949880e+00 -1.19222015e-01 -2.96651214e-01 2.94892937e-01
-2.03314528e-01 9.99102354e-01 -5.89077592e-01 3.89993250e-01
4.53358471e-01 9.20925885e-02 -1.40232861e-01 -7.58089781e-01
4.50428903e-01 8.33678901e-01 5.93548656e-01 -7.48194635e-01
5.43578744e-01 3.01766247e-01 -6.22308612e-01 -8.94878864e-01
-7.53298819e-01 -4.48280960e-01 -6.72113597e-01 1.73240885e-01
7.69586384e-01 -1.17826283e+00 -3.96605015e-01 4.62591141e-01
-1.15169144e+00 -1.77751213e-01 -1.26343846e-01 2.05250010e-01
-4.68688719e-02 2.35302627e-01 -6.62813008e-01 -4.80039626e-01
-2.54583061e-01 -1.14826679e+00 1.26049447e+00 -2.33098373e-01
-3.21791917e-01 -9.14195836e-01 -8.86127204e-02 7.90798783e-01
4.61876392e-01 1.44088594e-02 1.38630688e+00 -8.38816285e-01
-3.28299165e-01 5.62663190e-03 -5.49244225e-01 1.27545401e-01
-5.17315008e-02 -1.78935081e-02 -4.78071123e-01 -2.00279921e-01
-3.75221372e-02 -4.95740622e-01 2.88872331e-01 -2.22869903e-01
7.51173794e-01 -9.59984437e-02 -1.63600236e-01 4.28432643e-01
1.59842491e+00 2.77429998e-01 4.03689563e-01 5.70181906e-01
7.87751734e-01 8.86891484e-01 7.15361893e-01 2.96051025e-01
9.80001211e-01 7.83319533e-01 1.15141153e-01 -9.81959254e-02
-1.27005383e-01 -6.04995370e-01 4.53094125e-01 1.21912587e+00
5.54379642e-01 -8.22830200e-02 -1.33430052e+00 5.09950876e-01
-1.61828732e+00 -3.78209502e-01 -3.17867368e-01 2.27108788e+00
9.85941172e-01 1.49880394e-01 4.14410755e-02 -2.63477355e-01
3.84193361e-01 -1.06000170e-01 -3.58982682e-01 -3.50783378e-01
-5.58642983e-01 1.01154402e-01 4.00049418e-01 3.61307472e-01
-9.29322541e-01 1.37428474e+00 6.07964087e+00 7.01195896e-01
-1.22815144e+00 2.19215542e-01 3.20789069e-01 2.11160615e-01
-4.72395599e-01 1.43386588e-01 -1.14599931e+00 3.78627032e-01
1.04858828e+00 -2.08844021e-01 4.10270452e-01 8.81652772e-01
7.54311755e-02 -7.84482062e-02 -1.05774128e+00 8.60412776e-01
9.09179151e-02 -1.13760328e+00 1.71319157e-01 -1.36220559e-01
4.36089307e-01 5.24314344e-01 -2.93604374e-01 6.02578044e-01
1.77183703e-01 -6.21850371e-01 1.06070697e+00 8.33436549e-02
1.10108519e+00 -5.96218169e-01 8.01892042e-01 3.43206763e-01
-1.22742593e+00 1.65687911e-02 -1.44098699e-01 4.28675741e-01
2.55847394e-01 5.05887270e-01 -5.89338958e-01 7.73116112e-01
9.86222208e-01 5.12601972e-01 -8.99923921e-01 1.79602146e-01
8.59505776e-03 4.66756195e-01 -2.90650100e-01 1.69434682e-01
-7.10353851e-02 -2.94790566e-01 2.99346238e-01 1.54229736e+00
3.49420875e-01 -2.48002425e-01 3.10801685e-01 8.62120986e-01
-2.18384042e-01 6.33496046e-01 -6.63065016e-01 -3.97203892e-01
7.41228819e-01 9.21149254e-01 -4.65121150e-01 -4.72358227e-01
-9.26000178e-01 8.55828166e-01 4.33194578e-01 2.89566308e-01
-8.18710446e-01 -4.70188648e-01 4.84621882e-01 2.29839921e-01
1.45538509e-01 -1.83613300e-01 -3.92681539e-01 -1.34828436e+00
6.43223226e-01 -1.39361894e+00 4.86965656e-01 -6.96785808e-01
-1.08603299e+00 5.99417210e-01 5.64735085e-02 -7.16596186e-01
-2.71606445e-01 -4.70953017e-01 1.20151021e-01 8.86567295e-01
-1.33885860e+00 -1.39025927e+00 -1.28777191e-01 6.21383309e-01
6.44594193e-01 -3.49195987e-01 8.03793550e-01 8.14681113e-01
-7.49543667e-01 8.25187743e-01 9.43087414e-03 2.86896199e-01
8.82823467e-01 -1.11295080e+00 7.88753331e-01 9.66339350e-01
1.85354903e-01 1.07963848e+00 4.08360600e-01 -7.60031283e-01
-1.77493203e+00 -1.14424801e+00 1.78179109e+00 -7.67474771e-01
9.42743003e-01 -8.56113553e-01 -9.67289150e-01 8.54942977e-01
2.43703738e-01 -3.41865420e-01 5.38760722e-01 3.86163086e-01
-7.51112521e-01 -2.19747916e-01 -1.03954732e+00 6.62025571e-01
9.96348321e-01 -9.90070999e-01 -4.17096049e-01 2.11977243e-01
1.08227611e+00 -5.92946112e-01 -1.08175898e+00 4.42250162e-01
5.14561176e-01 -1.12238884e+00 6.62775874e-01 -6.53246403e-01
3.04744989e-01 -1.80012122e-01 -5.69394946e-01 -8.08619440e-01
3.56805086e-01 -4.21064496e-01 3.50394517e-01 1.68354046e+00
7.87041128e-01 -9.00721490e-01 6.95753098e-01 6.67778134e-01
-1.49713814e-01 -4.89817113e-01 -7.39154637e-01 -8.28387022e-01
1.44130111e-01 -8.57230604e-01 7.88018346e-01 1.17138290e+00
7.91862383e-02 5.11603236e-01 1.31935865e-01 3.61904763e-02
3.05037171e-01 9.09409001e-02 8.86627614e-01 -9.83606994e-01
-8.91311914e-02 -1.85017884e-01 7.37558380e-02 -9.76935923e-01
-3.49699035e-02 -1.05219448e+00 -2.37930253e-01 -1.31325948e+00
3.34709525e-01 -6.31626248e-01 9.53335166e-02 6.78101480e-01
-8.33945572e-02 -2.61975408e-01 2.70565510e-01 1.76919520e-01
-5.67452490e-01 1.73172861e-01 4.10181075e-01 1.50320753e-01
-1.22303218e-01 -3.36953282e-01 -9.38909292e-01 4.70549971e-01
6.27477050e-01 -5.87596714e-01 -2.35836998e-01 -9.16985154e-01
5.61070621e-01 6.98992237e-02 -1.41957954e-01 -6.35206759e-01
1.19848885e-01 -1.00374326e-01 -3.05023134e-01 -6.23717666e-01
-6.40334487e-02 -7.66488135e-01 2.73643166e-01 3.33746582e-01
-3.24598640e-01 7.20224261e-01 4.59827334e-01 1.50161654e-01
-5.23064375e-01 6.29192665e-02 3.75196695e-01 -4.30800132e-02
-7.56483197e-01 -1.28448695e-01 -7.80279860e-02 7.39457905e-01
7.39067674e-01 8.19166079e-02 -6.38986707e-01 6.49833586e-03
-2.50980794e-01 3.04487467e-01 3.96374702e-01 8.52578402e-01
1.78488418e-01 -1.14576685e+00 -7.64563322e-01 5.94819248e-01
7.69492269e-01 -2.21042916e-01 -1.08012827e-02 7.99203873e-01
-6.61337316e-01 8.80960941e-01 3.05935349e-02 -5.97150981e-01
-1.28933406e+00 3.96256983e-01 -5.33318669e-02 -4.30809766e-01
-1.84862509e-01 6.67681694e-01 -9.37453955e-02 -1.22367144e+00
1.62564710e-01 -4.60718453e-01 2.19692081e-01 -5.37210032e-02
1.62647646e-02 3.50724995e-01 4.38030392e-01 -7.52582073e-01
-5.25526702e-01 4.04754043e-01 -2.15485245e-01 -2.33462691e-01
1.13372731e+00 -1.77927896e-01 -4.85754579e-01 5.99835873e-01
1.20325768e+00 5.28087497e-01 -4.34662104e-01 -4.79308605e-01
7.23638594e-01 -1.55340448e-01 -4.34548080e-01 -1.13790190e+00
-9.54102039e-01 7.37862229e-01 1.69183284e-01 7.15136901e-02
8.72926056e-01 1.83998570e-01 9.02627826e-01 2.97659665e-01
6.97920799e-01 -1.36651587e+00 -4.12333846e-01 6.43821836e-01
8.29425097e-01 -1.47442770e+00 -1.08543523e-01 -7.11186230e-01
-9.77855861e-01 7.80779660e-01 6.54650390e-01 7.61315823e-01
4.67414171e-01 4.05140728e-01 4.47514653e-01 -3.27360272e-01
-6.84756339e-01 1.57208107e-02 5.61133996e-02 2.77207524e-01
9.04284716e-01 2.23278776e-02 -3.34145397e-01 8.88813555e-01
-4.36389714e-01 -5.44221848e-02 1.39599904e-01 1.11977434e+00
1.13677256e-01 -1.53141034e+00 -9.75905657e-02 2.94711292e-01
-6.15336418e-01 -4.36822236e-01 -5.33462882e-01 1.34196115e+00
1.66764021e-01 8.94255340e-01 -6.97347000e-02 -2.90603340e-01
7.89119959e-01 3.84700894e-01 7.94376284e-02 -6.56624138e-01
-8.50442052e-01 4.59200852e-02 4.68683451e-01 -8.07493031e-01
-1.02169141e-01 -8.95738065e-01 -1.27757335e+00 -3.73315096e-01
-6.61341846e-02 9.27162468e-02 9.58037317e-01 6.22998357e-01
6.49080455e-01 1.96418047e-01 7.00954795e-02 2.66407967e-01
-4.56477582e-01 -8.19739640e-01 -2.02915013e-01 5.70388138e-01
-2.41550282e-01 -2.77046949e-01 -1.22052997e-01 -1.28962900e-02] | [9.897665023803711, 7.926238059997559] |
e20c5535-14f4-4186-b7bb-7501bac3ab0b | eventgraph-at-case-2021-task-1-a-general | 2210.09770 | null | https://arxiv.org/abs/2210.09770v1 | https://arxiv.org/pdf/2210.09770v1.pdf | EventGraph at CASE 2021 Task 1: A General Graph-based Approach to Protest Event Extraction | This paper presents our submission to the 2022 edition of the CASE 2021 shared task 1, subtask 4. The EventGraph system adapts an end-to-end, graph-based semantic parser to the task of Protest Event Extraction and more specifically subtask 4 on event trigger and argument extraction. We experiment with various graphs, encoding the events as either "labeled-edge" or "node-centric" graphs. We show that the "node-centric" approach yields best results overall, performing well across the three languages of the task, namely English, Spanish, and Portuguese. EventGraph is ranked 3rd for English and Portuguese, and 4th for Spanish. Our code is available at: https://github.com/huiling-y/eventgraph_at_case | ['Lilja Øvrelid', 'Samia Touileb', 'David Samuel', 'Huiling You'] | 2022-10-18 | null | null | null | null | ['event-extraction'] | ['natural-language-processing'] | [-2.21424535e-01 4.68544364e-01 -3.96310180e-01 -3.49907011e-01
-9.82750893e-01 -9.42861021e-01 9.20821667e-01 4.71683174e-01
-3.52817476e-01 7.39182711e-01 8.44285548e-01 -4.18759555e-01
-1.15441792e-02 -8.05716515e-01 -5.69837987e-01 3.45991626e-02
-3.30124684e-02 5.24768770e-01 6.80553675e-01 -2.24181309e-01
-1.01323932e-01 3.41575354e-01 -1.04324377e+00 7.80861020e-01
1.63560838e-01 5.69932222e-01 -7.32209906e-02 6.35521114e-01
-3.55891377e-01 1.14293396e+00 -4.35892880e-01 -8.57878268e-01
2.94597298e-02 -4.61964726e-01 -1.35931218e+00 -5.79918802e-01
-2.53125817e-01 3.71561170e-01 -5.96887112e-01 8.81145537e-01
4.21539485e-01 1.37210377e-02 1.93297520e-01 -1.48168302e+00
-4.77224924e-02 1.30400252e+00 -2.84896016e-01 8.11840653e-01
9.30038571e-01 -3.64810452e-02 1.30554402e+00 -8.45079899e-01
1.60697567e+00 1.31033659e+00 5.20118475e-01 3.21249962e-01
-8.68773699e-01 -4.74778086e-01 2.27491707e-01 4.77686137e-01
-1.12985933e+00 -6.44687474e-01 6.16272092e-01 -4.64438796e-01
1.71186030e+00 4.15668249e-01 6.02154076e-01 1.49876344e+00
3.06323230e-01 9.03620303e-01 9.49364603e-01 -3.77196074e-01
1.78844426e-02 -3.84132564e-01 3.22243869e-01 4.93503422e-01
4.84821945e-01 1.63234279e-01 -7.66117692e-01 -4.77447331e-01
3.56779955e-02 -5.29536307e-01 -2.80855522e-02 9.60333347e-02
-1.19427013e+00 7.53968477e-01 1.12151906e-01 4.17668313e-01
-7.60167658e-01 3.14389437e-01 1.04307199e+00 2.34888494e-01
6.04997396e-01 1.22743756e-01 -8.28897238e-01 -2.81397432e-01
-6.41259551e-01 5.89222252e-01 1.20111263e+00 1.06705117e+00
1.71827093e-01 -2.80284762e-01 -1.60163850e-01 6.23403847e-01
2.43710220e-01 -1.32584972e-02 2.33796760e-01 -7.00804651e-01
8.29047501e-01 6.17504895e-01 1.94312066e-01 -7.59263158e-01
-8.10219228e-01 -2.61853188e-01 6.22750372e-02 -3.30564946e-01
4.39988285e-01 -3.82283151e-01 -7.30562925e-01 1.93191230e+00
4.77348387e-01 -1.76035881e-01 2.73524880e-01 6.08296335e-01
1.34437096e+00 5.68315804e-01 7.18861997e-01 -1.68007627e-01
1.95119309e+00 -9.11990225e-01 -9.33504105e-01 -4.24921632e-01
7.79496908e-01 -9.27585781e-01 6.35818541e-01 6.10761717e-02
-1.23937833e+00 1.14848793e-01 -8.88932645e-01 -1.54151693e-01
-5.40704250e-01 -2.50563949e-01 5.27100384e-01 7.40799010e-02
-7.06569731e-01 2.70129502e-01 -1.07013083e+00 -8.29008639e-01
4.33029532e-02 -2.81070918e-01 -4.43605810e-01 2.65912622e-01
-1.59583652e+00 9.05885994e-01 1.02408600e+00 -6.15432858e-01
-7.09372580e-01 -5.76573074e-01 -9.75545108e-01 1.07828667e-02
8.56288135e-01 -5.35461545e-01 1.73715007e+00 -4.08820421e-01
-8.19471419e-01 1.26307070e+00 -2.09657773e-01 -7.32803047e-01
4.94934946e-01 -2.75188982e-01 -1.00782621e+00 3.31633598e-01
6.50014758e-01 1.57697827e-01 3.44670713e-02 -7.32020378e-01
-8.59472215e-01 -2.67682791e-01 1.62073225e-01 4.27759700e-02
3.44913006e-01 1.06614220e+00 -3.27062011e-01 -7.01926708e-01
-1.07570902e-01 -6.21053338e-01 -7.48083815e-02 -7.22961605e-01
-7.06606984e-01 -6.75375044e-01 5.59891820e-01 -1.00648546e+00
1.42336810e+00 -2.00672603e+00 -1.58040538e-01 -8.01845044e-02
-4.72769886e-03 -3.28030258e-01 1.51273504e-01 9.80723381e-01
-5.13870001e-01 3.18731040e-01 1.03017032e-01 -5.40500367e-03
3.28068435e-01 1.58696964e-01 -9.95226055e-02 1.42326102e-01
2.25571439e-01 1.04365754e+00 -1.19636202e+00 -6.42510653e-01
-5.31113036e-02 6.54242933e-02 -1.89577565e-01 -4.86137606e-02
-4.11515743e-01 1.23958007e-01 -6.50168478e-01 5.98051727e-01
7.39002824e-02 -4.52309072e-01 7.52239227e-01 -3.61981392e-01
-2.90205479e-01 1.05906522e+00 -1.16774714e+00 1.79448164e+00
-3.85166965e-02 4.61549312e-01 1.30998358e-01 -6.89391911e-01
5.28924704e-01 7.17796326e-01 5.35408139e-01 -6.90996706e-01
-1.80108789e-02 4.14970040e-01 -1.79664880e-01 -5.33954322e-01
3.30303937e-01 8.47196504e-02 -8.44486535e-01 3.03198665e-01
2.79549271e-01 2.72557199e-01 8.85581195e-01 7.43742228e-01
1.66091239e+00 4.36918050e-01 9.95337546e-01 -3.79269034e-01
2.43935779e-01 4.94681358e-01 8.47055376e-01 4.26531583e-01
-2.12092638e-01 3.46928984e-01 8.99986744e-01 -5.15812516e-01
-8.15208495e-01 -1.09857070e+00 1.62427500e-01 1.08236027e+00
-2.17754751e-01 -9.98245120e-01 -4.82455403e-01 -1.04157066e+00
-2.35696435e-01 1.30419517e+00 -3.39390934e-01 1.77352533e-01
-9.15235877e-01 -7.21882522e-01 9.94162261e-01 5.73858857e-01
3.98295879e-01 -1.34322715e+00 -8.35332751e-01 7.79945433e-01
-7.42751241e-01 -1.48513985e+00 -2.90630668e-01 3.25522661e-01
-1.49965391e-01 -1.59876513e+00 8.21726397e-02 -4.15124595e-01
-1.83798194e-01 -5.89852452e-01 1.57069993e+00 -3.16287756e-01
-3.62516910e-01 2.64582336e-01 -7.39837170e-01 -6.59477890e-01
-4.26315397e-01 8.35832581e-02 -5.26013434e-01 -5.05491912e-01
5.83506346e-01 -2.99543858e-01 -4.00933832e-01 -9.20068398e-02
-6.33073628e-01 5.56851849e-02 -8.32305476e-02 3.33230883e-01
4.10943389e-01 -3.53703983e-02 7.79071808e-01 -1.05078137e+00
5.36945105e-01 -9.45639253e-01 -4.58169997e-01 1.20356418e-01
-3.71918321e-01 -2.79692769e-01 5.87176263e-01 3.40664148e-01
-1.25381827e+00 1.05567291e-01 -2.87576228e-01 1.94043875e-01
-3.44592601e-01 7.93801308e-01 -1.98720947e-01 7.65793741e-01
7.77099431e-01 -2.43225962e-01 -5.82425773e-01 -6.29487097e-01
3.64620447e-01 2.53162324e-01 6.48669839e-01 -7.60547459e-01
3.06064099e-01 5.29686630e-01 -2.63209760e-01 -3.10709417e-01
-8.96234214e-01 -5.15420437e-01 -1.91940621e-01 -1.33632630e-01
1.16465390e+00 -9.78888392e-01 -2.68782616e-01 2.89570242e-01
-1.41013646e+00 -4.73560810e-01 -6.80106640e-01 4.74635601e-01
-4.47090596e-01 7.22315162e-02 -8.91643047e-01 -3.39473307e-01
-4.78510499e-01 -7.29482472e-01 9.43513036e-01 4.50801067e-02
-5.93996942e-01 -1.06373465e+00 1.55771270e-01 6.37274384e-02
-8.24764520e-02 7.36604631e-01 8.45457733e-01 -1.35731602e+00
1.20546535e-01 -1.57749042e-01 -1.63964689e-01 -4.05964732e-01
-2.10135773e-01 -8.05791765e-02 -5.46432793e-01 6.82172924e-02
-2.17750758e-01 -1.50202379e-01 6.31896317e-01 2.10631415e-01
5.42774618e-01 -2.94844240e-01 -7.18299806e-01 1.58285305e-01
1.38162374e+00 3.90142828e-01 5.35035253e-01 7.99046338e-01
2.65491784e-01 5.41653395e-01 6.70012951e-01 5.53640008e-01
8.27000856e-01 6.62487030e-01 2.00370073e-01 2.55395889e-01
-5.83749890e-01 -4.36325520e-01 3.37273329e-01 2.86060661e-01
1.76184312e-01 -8.54798079e-01 -1.23166347e+00 1.08849633e+00
-2.07013202e+00 -1.10865951e+00 -6.55267715e-01 1.62065828e+00
7.11348414e-01 3.05864543e-01 3.70781094e-01 -8.75086635e-02
8.35932732e-01 4.93126601e-01 -3.50441448e-02 -5.04198134e-01
-2.40749925e-01 4.19259548e-01 3.26950252e-01 2.89878607e-01
-1.16624749e+00 1.12531006e+00 5.85103559e+00 6.64348245e-01
-6.07337534e-01 6.26402974e-01 1.46161452e-01 -1.57717109e-01
-2.98618078e-01 5.42054534e-01 -8.92186880e-01 7.05750167e-01
1.51329398e+00 -6.11639142e-01 1.20896593e-01 5.56685150e-01
2.67031610e-01 3.75059769e-02 -8.43167424e-01 4.65144187e-01
-4.06675965e-01 -1.60218990e+00 -2.56903917e-01 -2.22342595e-01
3.34527671e-01 5.72475016e-01 -8.54212701e-01 3.50260079e-01
8.01572502e-01 -6.21613443e-01 1.21127653e+00 2.16853380e-01
5.62405825e-01 -4.78843361e-01 6.21950448e-01 2.27163788e-02
-1.67193294e+00 1.18434280e-01 2.60345459e-01 3.19632828e-01
9.19746459e-01 6.40577674e-01 -6.05401158e-01 1.17630923e+00
1.00183678e+00 6.75436676e-01 -3.25316966e-01 8.42413902e-01
-7.56163836e-01 9.34005260e-01 -5.00320554e-01 1.33016780e-01
1.37566134e-01 4.39921170e-01 1.24048102e+00 1.70400822e+00
1.84990600e-01 4.72638994e-01 4.88118052e-01 4.59717751e-01
-2.34726816e-01 1.03326537e-01 -4.97063369e-01 -3.07205647e-01
7.12920368e-01 1.05556679e+00 -1.01198661e+00 -6.49993062e-01
-4.13823068e-01 5.78389108e-01 2.76031047e-01 1.22313470e-01
-9.45886850e-01 -6.54855847e-01 2.88164690e-02 1.47821516e-01
2.93412089e-01 -3.91128473e-02 6.80016130e-02 -1.02836418e+00
-5.51482849e-02 -7.20555246e-01 1.44950056e+00 -9.16006744e-01
-1.10614896e+00 7.19494164e-01 5.47732353e-01 -4.63160247e-01
-6.81965411e-01 -4.65285897e-01 -5.89356482e-01 4.28134292e-01
-1.03705728e+00 -1.16263306e+00 2.08070986e-02 6.52990580e-01
5.13955057e-01 9.23837200e-02 6.80864692e-01 6.27397895e-01
-3.69135350e-01 2.29608566e-01 -6.52908504e-01 2.29699776e-01
5.81465483e-01 -1.27405751e+00 7.82943428e-01 1.09037375e+00
7.21161887e-02 7.98069984e-02 8.63415003e-01 -1.13949144e+00
-1.12087536e+00 -1.02376151e+00 1.63407338e+00 -3.24778527e-01
1.19461823e+00 -1.18819937e-01 -5.09909391e-01 1.27015984e+00
5.61581314e-01 -8.58628079e-02 3.07792932e-01 1.52676359e-01
-5.47095299e-01 4.82596755e-01 -1.20386589e+00 3.83495718e-01
1.51300216e+00 -4.58788127e-01 -9.45635021e-01 8.63719463e-01
8.72057676e-01 -6.00715041e-01 -1.16986024e+00 4.18512404e-01
9.70891714e-02 -5.95899642e-01 7.81445503e-01 -8.74376595e-01
3.58050972e-01 -2.25307435e-01 -2.74255484e-01 -1.03718376e+00
-2.19255552e-01 -8.15345645e-01 -2.33844370e-01 1.27851784e+00
7.65144527e-01 -8.31858218e-01 3.94254386e-01 -5.43582030e-02
-5.09833157e-01 -2.92645395e-01 -1.07749915e+00 -9.13106978e-01
-1.75290093e-01 -7.16024876e-01 5.92657149e-01 1.11915839e+00
6.84470534e-02 4.42710459e-01 1.99928328e-01 1.16478950e-01
4.86270875e-01 1.86282337e-01 2.93881476e-01 -1.12422848e+00
-2.54451036e-01 -3.53937447e-01 -1.91775501e-01 -9.61135328e-02
1.83079645e-01 -1.43507504e+00 -2.77506083e-01 -1.97262812e+00
2.77749032e-01 1.22980349e-01 -1.41103715e-01 1.04757082e+00
-2.34875064e-02 -3.90750282e-02 7.50029683e-02 -5.30954171e-03
-8.06438148e-01 -3.63855064e-02 5.08043110e-01 2.25203872e-01
3.03633772e-02 -3.29318285e-01 -6.47942066e-01 8.76794517e-01
1.07566762e+00 -8.30352247e-01 -5.36881685e-02 -3.61422509e-01
4.91346806e-01 2.28296742e-01 5.29159725e-01 -5.49909770e-01
1.27260536e-01 -1.70952037e-01 -4.14024256e-02 -7.12455809e-01
-1.53553328e-02 -3.03920448e-01 6.81528807e-01 5.69406271e-01
-2.02678084e-01 4.50730741e-01 3.78143787e-01 3.77327591e-01
-1.19982004e-01 -3.02018434e-01 4.47957933e-01 -4.27300960e-01
-1.03595972e+00 -1.21556483e-01 -3.60633582e-01 7.76096523e-01
1.02445686e+00 2.54899204e-01 -7.90315092e-01 -2.36371532e-01
-1.23286664e+00 1.86061740e-01 1.46654308e-01 4.46321994e-01
1.48932263e-01 -1.15608799e+00 -1.17799127e+00 -4.56705064e-01
2.01420650e-01 -5.10472894e-01 1.71828418e-04 8.88715804e-01
-5.72398841e-01 3.61561894e-01 9.40385610e-02 -1.97015367e-02
-1.15796590e+00 4.30768222e-01 7.81374052e-02 -8.36689472e-01
-8.24914098e-01 6.49520159e-01 -2.53954440e-01 -2.74695367e-01
-2.39176169e-01 -2.92800134e-03 -9.22628418e-02 1.74772367e-01
3.62065345e-01 6.18680835e-01 2.45020926e-01 -5.13752878e-01
-9.24261749e-01 -9.34973434e-02 -4.31150906e-02 -4.04496849e-01
1.34055054e+00 1.51064724e-01 -2.06060395e-01 1.61766589e-01
8.84402514e-01 1.84651479e-01 -4.66726482e-01 7.22140912e-03
7.72736430e-01 1.51311979e-01 -2.21111268e-01 -1.25740600e+00
-7.79840171e-01 1.89281344e-01 1.85921006e-02 6.27726078e-01
8.48352671e-01 5.42510688e-01 9.41500366e-01 -3.65085565e-02
5.93911827e-01 -1.03643692e+00 -4.13130254e-01 6.03706181e-01
1.01801217e+00 -6.34589255e-01 1.18427314e-01 -7.40348518e-01
-5.51552773e-01 7.34988749e-01 2.71585077e-01 -1.82007894e-01
5.56276381e-01 5.24527192e-01 -3.27984929e-01 -8.43172908e-01
-1.07184136e+00 -3.23882729e-01 1.28730178e-01 1.71109304e-01
5.80130279e-01 3.95442396e-01 -1.14938760e+00 9.18415725e-01
-2.91388869e-01 -2.03572735e-02 2.80562520e-01 1.32662964e+00
-9.87270847e-02 -1.18189549e+00 -3.64201926e-02 3.08090150e-01
-1.12976122e+00 -2.90286779e-01 -7.20544875e-01 1.25435579e+00
-2.71693189e-02 1.12752068e+00 -5.60336262e-02 -1.00594662e-01
7.98830688e-01 5.42847693e-01 3.55780095e-01 -6.62628949e-01
-1.04966056e+00 1.36406928e-01 1.39432621e+00 -9.66403782e-01
-4.13325727e-01 -1.07732332e+00 -1.82266343e+00 -1.97482243e-01
2.15923518e-01 3.57301503e-01 5.83353400e-01 7.57445395e-01
6.22808278e-01 5.97544134e-01 3.02838422e-02 -1.05620250e-01
-1.83284596e-01 -7.80748725e-01 -1.13136053e-01 4.67821598e-01
-4.42608029e-01 -5.56388974e-01 -3.70479047e-01 3.86942290e-02] | [9.0250883102417, 9.26109504699707] |
c6b7a836-13d2-4039-be84-f9bd939bda6c | face-frontalization-based-on-robustly-fitting | 2010.13676 | null | https://arxiv.org/abs/2010.13676v2 | https://arxiv.org/pdf/2010.13676v2.pdf | Face Frontalization Based on Robustly Fitting a Deformable Shape Model to 3D Landmarks | Face frontalization consists of synthesizing a frontally-viewed face from an arbitrarily-viewed one. The main contribution of this paper is a robust face alignment method that enables pixel-to-pixel warping. The method simultaneously estimates the rigid transformation (scale, rotation, and translation) and the non-rigid deformation between two 3D point sets: a set of 3D landmarks extracted from an arbitrary-viewed face, and a set of 3D landmarks parameterized by a frontally-viewed deformable face model. An important merit of the proposed method is its ability to deal both with noise (small perturbations) and with outliers (large errors). We propose to model inliers and outliers with the generalized Student's t-probability distribution function, a heavy-tailed distribution that is immune to non-Gaussian errors in the data. We describe in detail the associated expectation-maximization (EM) algorithm that alternates between the estimation of (i) the rigid parameters, (ii) the deformation parameters, and (iii) the Student-t distribution parameters. We also propose to use the zero-mean normalized cross-correlation, between a frontalized face and the corresponding ground-truth frontally-viewed face, to evaluate the performance of frontalization. To this end, we use a dataset that contains pairs of profile-viewed and frontally-viewed faces. This evaluation, based on direct image-to-image comparison, stands in contrast with indirect evaluation, based on analyzing the effect of frontalization on face recognition. | ['Radu Horaud', 'Mostafa Sadeghi', 'Zhiqi Kang'] | 2020-10-26 | null | null | null | null | ['robust-face-alignment', 'face-alignment', 'face-model'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 3.20806563e-01 1.79825082e-01 3.13581258e-01 -4.89682585e-01
-8.57565761e-01 -6.07760668e-01 6.17520928e-01 -4.04055744e-01
-3.54512036e-01 4.26411629e-01 -6.63250610e-02 2.22273350e-01
-1.43868430e-03 -4.39419627e-01 -8.77840817e-01 -9.43954051e-01
8.45490694e-02 6.89799845e-01 -1.52947038e-01 1.00002348e-01
2.58650541e-01 1.22733712e+00 -1.74142671e+00 -2.18808204e-01
5.05097628e-01 9.29983437e-01 -2.87484109e-01 4.47077066e-01
3.28894734e-01 -1.96268842e-01 -4.50118572e-01 -5.36732435e-01
6.03580654e-01 -3.03412706e-01 -4.81460035e-01 5.51660538e-01
1.00349629e+00 -3.03989738e-01 1.24560639e-01 1.11182928e+00
3.08582813e-01 2.94898838e-01 8.19750607e-01 -1.16772199e+00
-3.02684575e-01 -2.97513843e-01 -7.55819738e-01 -1.69216126e-01
6.90837204e-01 -9.03028622e-03 3.91681999e-01 -1.18247306e+00
8.99186552e-01 1.25081098e+00 6.30189061e-01 5.41161954e-01
-1.38320363e+00 -4.43006843e-01 -1.42663851e-01 -1.49907187e-01
-1.63025045e+00 -8.35363686e-01 7.67571688e-01 -7.46069789e-01
4.85621363e-01 2.97591567e-01 5.21976054e-01 8.77648711e-01
1.90816686e-01 -9.42672044e-02 1.21343911e+00 -6.04918122e-01
2.49856174e-01 9.79706179e-03 -3.00706357e-01 7.56943762e-01
1.20793179e-01 2.26377010e-01 -1.89590037e-01 -5.17817736e-01
8.13004911e-01 -1.50699630e-01 -4.33025509e-01 -5.24586439e-01
-8.16941261e-01 4.82111186e-01 -4.30630818e-02 2.99847931e-01
-5.55376172e-01 -8.43232349e-02 -5.38521446e-02 6.50355443e-02
6.67505920e-01 3.68067212e-02 -2.70712107e-01 1.09607175e-01
-1.24404216e+00 2.42403209e-01 6.41015708e-01 7.38064051e-01
8.92724752e-01 2.34183475e-01 5.91481514e-02 5.73771179e-01
5.77482760e-01 9.78940904e-01 4.03821677e-01 -8.41526568e-01
5.74050508e-02 2.93367863e-01 4.26676795e-02 -1.06195664e+00
-2.19882265e-01 2.58818623e-02 -5.53821146e-01 5.20061374e-01
6.26929462e-01 -6.25370741e-02 -8.80363345e-01 1.80262613e+00
7.00397372e-01 2.12898508e-01 -2.50359654e-01 8.24311674e-01
3.36266100e-01 1.53499052e-01 -4.32993263e-01 -6.28508329e-01
1.36371005e+00 -4.31599021e-01 -5.73402524e-01 9.11559910e-02
-1.61881775e-01 -1.13789749e+00 8.90348732e-01 1.78349972e-01
-1.33072710e+00 -5.29119790e-01 -8.31845939e-01 -4.81428904e-03
6.00242540e-02 9.51243564e-02 -4.54288274e-01 6.91193342e-01
-1.13505900e+00 6.45178854e-01 -8.27624142e-01 -3.93865645e-01
3.72414701e-02 5.15196681e-01 -1.07665443e+00 2.54240513e-01
-4.70211536e-01 9.65794265e-01 -2.61624157e-01 1.68225393e-01
-7.14690864e-01 -5.60263336e-01 -7.35933959e-01 5.83992747e-04
1.68687508e-01 -6.05666220e-01 8.22858453e-01 -1.14636707e+00
-1.77775991e+00 1.36665547e+00 -5.53275466e-01 1.41342148e-01
7.78765798e-01 -1.22899348e-02 -3.20117861e-01 2.79911578e-01
-2.36361586e-02 3.69118989e-01 1.39639437e+00 -1.45999706e+00
1.04592264e-01 -1.00457203e+00 -5.84065020e-01 6.28545061e-02
-1.20183848e-01 1.27434120e-01 -5.70724130e-01 -6.92657173e-01
4.91344303e-01 -1.23153579e+00 1.46693990e-01 1.88564286e-01
-3.95968556e-01 2.94752836e-01 7.28784859e-01 -1.00106549e+00
6.49135351e-01 -2.35262227e+00 1.70752242e-01 7.12168694e-01
-1.36643425e-01 5.49101643e-02 -2.26155415e-01 3.20073187e-01
-6.74874842e-01 -1.74920961e-01 -2.59115726e-01 -6.60522163e-01
-2.58743972e-01 -2.82268622e-03 -1.62300512e-01 1.24440396e+00
-3.34865339e-02 3.59075963e-01 -4.85174388e-01 -2.42250621e-01
1.44281730e-01 9.00933802e-01 -3.99928391e-01 1.97877482e-01
1.23997629e-01 6.14915729e-01 -7.02833980e-02 5.52879930e-01
1.04826236e+00 3.32140595e-01 1.98159695e-01 -5.47348142e-01
-8.77070576e-02 -3.98163110e-01 -1.41417146e+00 1.34197867e+00
-5.38126528e-02 4.70024109e-01 3.24800611e-01 -4.79321182e-01
1.19714987e+00 4.72590417e-01 7.94630587e-01 -2.38480479e-01
3.97694260e-01 2.93596238e-01 -2.26658434e-01 -3.73269647e-01
2.39836603e-01 -3.27849150e-01 5.43249726e-01 4.09777999e-01
9.86265764e-02 -1.86079860e-01 -9.40236300e-02 -3.84117067e-01
5.33608735e-01 3.35379660e-01 3.63583595e-01 -4.86690521e-01
7.24097431e-01 -6.09554648e-01 3.65173548e-01 -1.66399077e-01
7.81256985e-03 1.02616107e+00 4.47604179e-01 -2.56141931e-01
-1.02732027e+00 -1.13069963e+00 -3.95876765e-01 3.27076972e-01
-2.56016225e-01 2.30226740e-02 -1.30590630e+00 -4.26412523e-01
1.96646526e-01 4.55570072e-01 -7.47090518e-01 -2.37002317e-03
-6.26127958e-01 -4.38019633e-01 3.13100636e-01 2.11891174e-01
1.17282867e-01 -5.15754879e-01 -4.31075960e-01 -2.60968596e-01
-1.75609458e-02 -1.14264858e+00 -8.18126619e-01 -3.02976340e-01
-8.51377487e-01 -1.29799891e+00 -6.71632528e-01 -5.71394503e-01
1.26706111e+00 4.89560701e-02 8.40446055e-01 -6.02122955e-02
-1.04353584e-01 6.58087492e-01 1.15603574e-01 -1.63173184e-01
-4.12765861e-01 -6.46942019e-01 5.30498385e-01 6.32719457e-01
4.90957946e-02 -5.99377215e-01 -4.86358136e-01 7.24879861e-01
-8.85790527e-01 -5.55638611e-01 -2.77548339e-02 6.71405137e-01
8.43013823e-01 -1.49875253e-01 -1.07761703e-01 -4.96536851e-01
3.05880219e-01 -3.67314182e-02 -9.67396617e-01 3.24247479e-01
-5.25325894e-01 -1.02940992e-01 3.82841021e-01 -4.38609421e-01
-9.30101573e-01 2.91544735e-01 -1.52588606e-01 -8.88047040e-01
-2.60397166e-01 -2.13963110e-02 -3.87850970e-01 -4.82186496e-01
7.22531080e-01 -3.74239758e-02 5.12179375e-01 -4.60794866e-01
3.72594893e-01 5.28510988e-01 8.15994620e-01 -6.30634129e-01
1.18363261e+00 8.66004944e-01 3.63965333e-01 -1.08848596e+00
-2.19685420e-01 -4.47207958e-01 -1.05731380e+00 -3.79756123e-01
6.98239803e-01 -6.47494674e-01 -8.05347860e-01 5.23765385e-01
-1.12326729e+00 1.81311928e-02 -4.74166572e-01 5.50011098e-01
-8.34596097e-01 6.71148598e-01 -1.65055573e-01 -7.53492057e-01
-2.88865685e-01 -1.19550514e+00 1.27907240e+00 3.76177505e-02
-2.14179695e-01 -9.73958969e-01 2.09928960e-01 4.16261852e-01
-4.05601189e-02 6.13238275e-01 5.49139142e-01 -3.75109673e-01
-1.36525005e-01 -5.46375453e-01 2.03775927e-01 6.38890624e-01
1.54360041e-01 6.48578525e-01 -1.21047795e+00 -5.67131758e-01
5.39965272e-01 2.38082692e-01 1.79785222e-01 5.50405622e-01
7.12072074e-01 -2.10797712e-01 -1.29164144e-01 7.91677117e-01
1.33675945e+00 2.06233576e-01 5.94305456e-01 3.43790911e-02
5.21748781e-01 8.21755826e-01 5.72495341e-01 4.93764520e-01
-1.34120539e-01 1.06676662e+00 4.79351342e-01 2.49278229e-02
-3.70068625e-02 -2.12563530e-01 4.67280895e-01 4.22911048e-01
-4.14533198e-01 1.40602976e-01 -7.25842535e-01 1.12126708e-01
-1.36770880e+00 -9.86216247e-01 -9.19782966e-02 2.75957251e+00
4.25236642e-01 -3.63324136e-01 2.07110330e-01 1.49689585e-01
7.85706341e-01 -2.20655173e-01 -2.73230910e-01 -2.31723115e-01
-6.52868375e-02 2.39616022e-01 3.70680183e-01 8.53040695e-01
-8.18409026e-01 4.90951151e-01 6.15620422e+00 5.56674123e-01
-1.22505426e+00 -5.59409820e-02 4.32778060e-01 4.89387028e-02
-3.05977911e-01 -9.53161567e-02 -7.59859979e-01 4.38752025e-01
7.57501900e-01 -1.88879222e-01 4.85925943e-01 8.33861828e-01
2.58624673e-01 1.27631286e-02 -1.15606403e+00 9.72672284e-01
6.08390689e-01 -9.07135785e-01 -3.67361680e-02 4.24708575e-01
6.76110625e-01 -3.68177295e-01 3.27984363e-01 -4.87809032e-01
-4.03641403e-01 -9.74220574e-01 8.16121876e-01 6.53606296e-01
1.12657547e+00 -5.85564375e-01 5.01101613e-01 4.66711484e-02
-1.00809133e+00 2.99065769e-01 -1.39332443e-01 4.68526781e-01
9.97137502e-02 5.79921365e-01 -7.93170333e-01 4.55735952e-01
4.66590494e-01 2.32252121e-01 -3.14069539e-01 7.29479134e-01
-8.63491520e-02 1.61401182e-02 -5.00768125e-01 6.71756804e-01
-2.56606311e-01 -7.88757861e-01 7.56954014e-01 8.24928284e-01
6.45483315e-01 -9.39109847e-02 -1.58299096e-02 6.34208798e-01
3.22732292e-02 3.40686291e-01 -5.38353860e-01 3.24469626e-01
3.08292896e-01 1.35880983e+00 -8.29417646e-01 -9.78305936e-02
-3.03508699e-01 9.96911705e-01 -6.08434551e-04 4.11271542e-01
-6.23722494e-01 4.59322333e-02 8.01390529e-01 5.68468690e-01
2.12302417e-01 -2.48946175e-01 -1.44682646e-01 -1.00473249e+00
2.79779643e-01 -9.15274262e-01 7.62572810e-02 -7.41544545e-01
-1.06550205e+00 8.49464238e-01 2.65783638e-01 -1.10023582e+00
-4.75362211e-01 -7.04143345e-01 -6.19983971e-01 1.18094611e+00
-1.17865396e+00 -1.11978960e+00 -3.81302744e-01 8.59710991e-01
1.16069101e-01 -1.90720454e-01 8.11040580e-01 2.75272429e-01
-4.19158250e-01 7.35900164e-01 2.00878844e-01 4.05159108e-02
9.02583361e-01 -8.45548928e-01 1.97065651e-01 7.59732902e-01
5.69816679e-02 7.36820161e-01 8.42022002e-01 -5.61153650e-01
-1.28435588e+00 -8.16570520e-01 7.86693692e-01 -4.72348809e-01
2.81597942e-01 -1.90597996e-01 -8.67001534e-01 9.65852439e-01
-2.21225038e-01 2.94461548e-01 5.91265321e-01 -1.86105981e-01
-2.64753550e-01 -1.79085165e-01 -1.68222320e+00 3.78692746e-01
6.34953201e-01 -5.05284369e-01 -4.09062535e-01 3.52517128e-01
-1.46785274e-01 -6.16925359e-01 -1.10787106e+00 3.71285975e-01
8.51681173e-01 -1.17171681e+00 9.60160673e-01 -3.61036092e-01
-5.64000532e-02 -4.03718889e-01 -2.22402841e-01 -1.27202427e+00
-2.19861418e-01 -7.51948953e-01 1.98465899e-01 1.34850144e+00
-9.62103307e-02 -6.16316974e-01 9.40202415e-01 6.81833446e-01
9.87971425e-02 -6.21938705e-01 -1.27478743e+00 -8.43384981e-01
-7.12498128e-02 -5.52761629e-02 4.19865310e-01 7.88098216e-01
-3.35871607e-01 -3.19762044e-02 -1.48890749e-01 3.84078205e-01
7.04236805e-01 4.30538580e-02 1.02178097e+00 -1.28847575e+00
-8.55556801e-02 -8.96831527e-02 -6.85855150e-01 -5.23989677e-01
4.14097190e-01 -6.11401737e-01 1.31928260e-02 -6.81624651e-01
-3.81232835e-02 7.67928809e-02 3.70419055e-01 2.08877861e-01
1.13439858e-01 2.58693248e-01 9.46531668e-02 2.85654753e-01
4.10333306e-01 5.27266026e-01 1.20076919e+00 3.80129874e-01
-9.82543230e-02 1.42835945e-01 -1.17100567e-01 1.05775213e+00
3.53255510e-01 -3.60460937e-01 -2.17316076e-01 -9.48969126e-02
-2.21961945e-01 1.35563329e-01 3.64807755e-01 -8.84170711e-01
-6.04835227e-02 -9.61820260e-02 5.27829826e-01 -2.38304704e-01
5.77536106e-01 -1.10597146e+00 6.17303550e-01 3.79103124e-01
7.71580115e-02 4.07347292e-01 4.04182710e-02 3.09676439e-01
-1.33455321e-01 -3.86379302e-01 1.29953086e+00 3.37242261e-02
3.77016403e-02 4.15158063e-01 1.12647424e-02 -1.48303092e-01
1.19017613e+00 -5.40827990e-01 4.82016765e-02 -3.47307980e-01
-8.22785378e-01 -6.00367546e-01 1.02245915e+00 8.10571909e-02
6.30547345e-01 -1.28924167e+00 -7.67254591e-01 9.52185214e-01
-1.63382217e-01 -1.76017866e-01 1.85287192e-01 9.40907061e-01
-5.68351030e-01 -5.84127754e-02 -3.08738738e-01 -5.99442601e-01
-1.76986372e+00 6.05359256e-01 4.96003777e-01 2.07587928e-01
-2.91996747e-01 6.92641914e-01 -6.15065135e-02 -3.50346893e-01
-1.00340964e-02 -1.42607510e-01 8.78965855e-02 1.78060263e-01
3.72093141e-01 5.31811297e-01 2.91159481e-01 -1.35438228e+00
-3.52510393e-01 1.24839282e+00 2.37215817e-01 -3.28180343e-01
9.27211106e-01 -6.26360252e-02 -4.00002569e-01 1.91834137e-01
1.42855716e+00 4.71473038e-01 -1.31793070e+00 3.03721614e-02
-2.07191855e-01 -8.41697514e-01 -8.73217583e-02 -2.72296041e-01
-1.28384948e+00 7.36155450e-01 8.72537613e-01 -5.98911420e-02
1.00498426e+00 -3.16899627e-01 2.87673891e-01 -9.82537493e-02
2.00050160e-01 -9.53130662e-01 -6.15534484e-02 1.28222525e-01
1.12911785e+00 -8.90424967e-01 2.49037534e-01 -6.94957256e-01
-1.73299760e-01 1.31936312e+00 4.07081008e-01 -1.88025400e-01
8.42065215e-01 3.09085339e-01 2.35093966e-01 -7.41979778e-02
-1.69517025e-01 1.50765162e-02 5.70530891e-01 6.44151390e-01
4.28663284e-01 -1.59562513e-01 -1.48263827e-01 -8.64611194e-02
-2.87616670e-01 -6.98574185e-02 2.21221477e-01 6.51038826e-01
-1.09137774e-01 -1.06201136e+00 -1.01463020e+00 1.51243191e-02
-2.50458956e-01 3.20430398e-01 -2.21910894e-01 8.35529268e-01
2.04693362e-01 6.50620222e-01 3.60347241e-01 -1.20524652e-01
6.64376020e-01 8.64135921e-02 8.80902350e-01 -3.76935691e-01
-4.18392986e-01 3.51660520e-01 -3.43920201e-01 -6.02029860e-01
-4.77974445e-01 -1.04837549e+00 -9.04762745e-01 -4.64781910e-01
-1.95753366e-01 -9.10475384e-03 9.47530985e-01 6.97503686e-01
1.86147794e-01 -8.77798349e-02 8.28896284e-01 -1.21195269e+00
-7.63741076e-01 -7.59098709e-01 -8.48095775e-01 8.03466082e-01
2.58009464e-01 -7.63352633e-01 -7.02229261e-01 2.65852541e-01] | [13.154034614562988, 0.04096107929944992] |
f1ebecae-cd12-4a42-9fb1-a1f3c8e3b123 | discern-discourse-aware-entailment-reasoning | 2010.01838 | null | https://arxiv.org/abs/2010.01838v3 | https://arxiv.org/pdf/2010.01838v3.pdf | Discern: Discourse-Aware Entailment Reasoning Network for Conversational Machine Reading | Document interpretation and dialog understanding are the two major challenges for conversational machine reading. In this work, we propose Discern, a discourse-aware entailment reasoning network to strengthen the connection and enhance the understanding for both document and dialog. Specifically, we split the document into clause-like elementary discourse units (EDU) using a pre-trained discourse segmentation model, and we train our model in a weakly-supervised manner to predict whether each EDU is entailed by the user feedback in a conversation. Based on the learned EDU and entailment representations, we either reply to the user our final decision "yes/no/irrelevant" of the initial question, or generate a follow-up question to inquiry more information. Our experiments on the ShARC benchmark (blind, held-out test set) show that Discern achieves state-of-the-art results of 78.3% macro-averaged accuracy on decision making and 64.0 BLEU1 on follow-up question generation. Code and models are released at https://github.com/Yifan-Gao/Discern. | ['Michael R. Lyu', 'Irwin King', 'Caiming Xiong', 'Steven C. H. Hoi', 'Shafiq Joty', 'Jingjing Li', 'Chien-Sheng Wu', 'Yifan Gao'] | 2020-10-05 | null | https://aclanthology.org/2020.emnlp-main.191 | https://aclanthology.org/2020.emnlp-main.191.pdf | emnlp-2020-11 | ['discourse-segmentation'] | ['natural-language-processing'] | [ 4.80752319e-01 8.64788949e-01 -2.12699190e-01 -6.75325274e-01
-1.22152793e+00 -8.33776176e-01 9.17229950e-01 1.61176428e-01
-2.04949021e-01 8.07157576e-01 8.98012638e-01 -9.42714751e-01
2.38717273e-01 -5.69063306e-01 -4.84972835e-01 -2.01989502e-01
5.16535997e-01 7.81602919e-01 5.88672869e-02 -4.94685292e-01
3.31068128e-01 -3.20657879e-01 -1.11387610e+00 1.14379859e+00
1.02253819e+00 7.28686094e-01 1.01215541e-01 1.22931635e+00
-2.74900645e-01 1.53342724e+00 -1.06384897e+00 -6.76962376e-01
-4.14584845e-01 -9.82106149e-01 -2.01414561e+00 3.62350911e-01
1.39545634e-01 -7.29155660e-01 -2.01379344e-01 8.26776981e-01
2.69752413e-01 1.95655197e-01 6.33965909e-01 -8.61904323e-01
-8.62764239e-01 1.17628837e+00 -1.77634746e-01 1.50836006e-01
8.47706020e-01 3.07340831e-01 1.34189177e+00 -6.54337108e-01
5.06462932e-01 1.45896387e+00 1.15131952e-01 1.01428497e+00
-8.02562535e-01 -1.46316946e-01 3.52716893e-01 4.49343950e-01
-6.15273774e-01 -5.81477761e-01 4.95449871e-01 -3.41415197e-01
9.99415755e-01 7.82509983e-01 1.55297026e-01 1.32616675e+00
-2.76717007e-01 1.10342216e+00 9.03051496e-01 -6.58603609e-01
2.27977306e-01 -7.50375539e-02 5.44476748e-01 6.05096519e-01
-3.03995311e-01 -5.89396238e-01 -2.56667703e-01 -6.06801473e-02
3.57144997e-02 -4.48007315e-01 -6.35670483e-01 7.59139180e-01
-9.74812865e-01 1.07018590e+00 4.08062547e-01 1.87661707e-01
-2.24707276e-01 -1.75028697e-01 2.49571055e-01 3.07045758e-01
4.42203343e-01 7.29637921e-01 -1.87885806e-01 -3.52361023e-01
-5.87384820e-01 4.15154338e-01 1.42699099e+00 7.51794875e-01
3.26875418e-01 -6.53856635e-01 -8.31823826e-01 7.92565763e-01
4.64412540e-01 3.39879066e-01 3.99932772e-01 -1.37991083e+00
7.74126530e-01 8.12934577e-01 2.80389011e-01 -5.39371669e-01
-1.85444787e-01 -1.23025134e-01 -7.25778282e-01 -1.60388827e-01
6.26365244e-01 -6.27622843e-01 -6.19010091e-01 1.46017647e+00
2.50178784e-01 -3.85586262e-01 5.04144371e-01 9.02914584e-01
1.34247077e+00 9.17815328e-01 3.66099179e-02 -2.19967619e-01
1.79595673e+00 -1.40990376e+00 -9.77638423e-01 -3.83714378e-01
5.59484720e-01 -9.63132083e-01 1.18554080e+00 9.05892998e-02
-1.22474337e+00 -2.96221882e-01 -8.47125828e-01 -4.95438248e-01
1.58540830e-01 1.25171110e-01 1.40848890e-01 2.36266255e-01
-8.68754566e-01 1.22679725e-01 -5.18508852e-01 -4.23493981e-02
1.90220132e-01 2.35292017e-02 3.05811644e-01 -1.48352534e-01
-1.37818050e+00 7.71738648e-01 1.66086689e-01 9.19618458e-02
-7.83189952e-01 -3.46802026e-01 -7.87347257e-01 1.35672107e-01
6.18353844e-01 -7.15203464e-01 2.25701094e+00 -8.76043558e-01
-1.92986226e+00 9.58352506e-01 -6.16096675e-01 -6.14367247e-01
5.49144626e-01 -3.67816776e-01 -9.18031335e-02 2.66420126e-01
1.94726765e-01 6.51003480e-01 3.37808222e-01 -1.14271998e+00
-7.48537421e-01 -7.30815381e-02 8.66778970e-01 5.71742535e-01
2.15715900e-01 3.08775064e-02 -4.74390686e-01 -2.48314440e-01
-1.33538261e-01 -8.00661802e-01 2.43891980e-02 -6.15252078e-01
-9.83093739e-01 -7.55401433e-01 6.12643898e-01 -1.15458882e+00
1.35115576e+00 -1.66248429e+00 1.34837955e-01 -2.82207906e-01
3.84271592e-01 1.68120399e-01 -1.66450366e-01 5.06587446e-01
2.81367242e-01 3.74730021e-01 -2.94420600e-01 -4.40918833e-01
1.07369527e-01 2.16491451e-03 -5.39337575e-01 -2.33068153e-01
3.79301876e-01 9.08041954e-01 -9.67934489e-01 -2.62875497e-01
-1.53387347e-02 -4.16734032e-02 -6.08397424e-01 7.57713199e-01
-1.04597819e+00 5.82062900e-01 -5.30903339e-01 3.65552127e-01
3.98168087e-01 -7.75366426e-01 3.65234524e-01 1.51953653e-01
1.49442568e-01 9.99182761e-01 -6.41123891e-01 1.31705570e+00
-4.01780277e-01 9.29835677e-01 5.28386012e-02 -6.46431088e-01
5.77228606e-01 4.74354357e-01 -3.60834509e-01 -5.92724264e-01
3.29439074e-01 1.77428871e-02 -3.96533776e-03 -7.36316621e-01
7.09758043e-01 1.81083709e-01 -2.05288574e-01 8.47922146e-01
-1.39041901e-01 -8.35347623e-02 4.59831417e-01 4.39510792e-01
1.04689753e+00 -3.41013134e-01 3.84499907e-01 7.40125775e-02
9.48093057e-01 2.12304637e-01 3.94817367e-02 8.76115561e-01
-5.25615774e-02 7.53433764e-01 1.04260266e+00 -4.96903760e-03
-5.31423807e-01 -8.04229319e-01 3.53887856e-01 1.37310731e+00
1.20847180e-01 -4.88073379e-01 -1.22014618e+00 -9.15968716e-01
-3.32504719e-01 1.29936469e+00 -5.88585913e-01 5.01570441e-02
-7.35230446e-01 -4.29317921e-01 5.90543151e-01 3.89932722e-01
7.76835442e-01 -1.19660175e+00 -3.47687781e-01 1.46635547e-01
-1.07203221e+00 -1.04660642e+00 -5.21027327e-01 -2.57579327e-01
-3.06054622e-01 -1.30794680e+00 -6.06294274e-01 -8.41614187e-01
5.99504650e-01 1.66088328e-01 1.23829031e+00 4.87680674e-01
4.08906341e-01 3.07750076e-01 -7.53408670e-01 -2.21384913e-01
-9.67985332e-01 2.56243378e-01 -6.65853739e-01 -7.75162429e-02
3.40304345e-01 1.20873831e-01 -7.69626677e-01 3.06008637e-01
-6.63310528e-01 4.11713481e-01 2.52993494e-01 8.70941639e-01
1.09566674e-01 -5.35399497e-01 7.28686035e-01 -1.05729580e+00
1.42602110e+00 -4.68520701e-01 -2.86438227e-01 5.59560001e-01
-3.27826411e-01 1.72216877e-01 4.35062796e-01 -1.54113635e-01
-1.61126339e+00 -5.18524289e-01 -5.22213995e-01 5.94797313e-01
-1.48110583e-01 5.87638319e-01 -1.32295161e-01 7.52849996e-01
7.33739793e-01 4.00914624e-02 -3.20327207e-02 -2.53593534e-01
6.32162035e-01 1.23467994e+00 7.48800099e-01 -7.05995023e-01
3.45053107e-01 -9.78221558e-03 -9.08561170e-01 -5.30530632e-01
-1.51586890e+00 -4.22963679e-01 -2.46894300e-01 -2.64414638e-01
1.35343766e+00 -7.82718837e-01 -1.04924452e+00 3.96609277e-01
-1.80001819e+00 -7.47139275e-01 -8.93653631e-02 4.85204868e-02
-3.63377422e-01 2.71186769e-01 -8.55899215e-01 -9.05612826e-01
-6.74401999e-01 -1.24787021e+00 7.97436357e-01 6.68630838e-01
-7.76326239e-01 -8.73989701e-01 -3.88894044e-02 1.29190552e+00
2.09403843e-01 -2.61073243e-02 1.01876438e+00 -9.57079947e-01
-4.50287908e-01 7.00542778e-02 -2.11159632e-01 4.25367534e-01
1.73099652e-01 -8.36082622e-02 -1.09729242e+00 8.02355856e-02
1.12820044e-01 -6.40562177e-01 6.31599247e-01 1.54784024e-01
1.17737615e+00 -8.33179951e-01 7.67277703e-02 -2.31077187e-02
6.27107501e-01 2.14290857e-01 6.46315396e-01 2.82429934e-01
3.36400688e-01 8.32838774e-01 4.52444434e-01 3.40679318e-01
9.75101531e-01 3.24147165e-01 3.23059350e-01 1.36595309e-01
-2.37912580e-01 -2.60516912e-01 3.30041617e-01 6.79377735e-01
5.26779070e-02 -8.02250445e-01 -9.74241912e-01 4.05654490e-01
-1.89609492e+00 -9.64573383e-01 -3.92063826e-01 1.59746051e+00
1.43678260e+00 1.54697672e-01 -1.04244590e-01 -7.64821917e-02
7.54037559e-01 4.04182434e-01 -4.93364364e-01 -4.89220351e-01
-7.45437369e-02 -1.36298698e-03 -2.34141648e-01 1.22757530e+00
-9.78974998e-01 8.87669086e-01 5.27355528e+00 4.51094657e-01
-8.06401074e-01 1.01619633e-02 1.22529662e+00 1.57722920e-01
-4.63186502e-01 -1.58236902e-02 -8.03670764e-01 3.05855691e-01
1.08064473e+00 -3.67988974e-01 3.26836675e-01 5.76409340e-01
1.40523583e-01 -1.39286667e-01 -1.24785650e+00 4.20952618e-01
2.01126829e-01 -1.67128778e+00 1.99994817e-01 -3.19743484e-01
6.57576978e-01 -2.87827462e-01 -9.68578532e-02 6.53668880e-01
5.60220242e-01 -1.16907072e+00 6.34751618e-01 2.48561472e-01
5.46905160e-01 -3.52739632e-01 1.06009483e+00 6.81633890e-01
-6.41711712e-01 1.34442240e-01 9.12993774e-02 -3.07190120e-01
3.11742723e-01 6.27438053e-02 -1.32567942e+00 3.79442781e-01
2.37642527e-01 8.33839253e-02 -5.61336219e-01 3.99978161e-01
-9.53928232e-01 9.19947565e-01 5.66146486e-02 -4.38258976e-01
4.51849341e-01 8.47300962e-02 4.04436499e-01 1.26915669e+00
-1.73566505e-01 6.42247260e-01 -6.22402541e-02 8.12344909e-01
-6.61187649e-01 -1.16343275e-01 1.18473366e-01 1.24970805e-02
5.65852404e-01 1.00962818e+00 -3.24867249e-01 -7.19384670e-01
-2.44902655e-01 1.12464380e+00 3.91835511e-01 4.07387435e-01
-6.40931606e-01 -2.99287498e-01 3.83415431e-01 -1.94702685e-01
3.58554311e-02 3.06581259e-01 -3.25268894e-01 -1.06128418e+00
1.61882266e-01 -1.42353213e+00 5.87112546e-01 -7.65673816e-01
-1.14787519e+00 7.94211745e-01 -2.01851502e-01 -8.14269960e-01
-7.35793352e-01 -5.80558419e-01 -1.14647484e+00 8.93626869e-01
-1.47806060e+00 -7.85628557e-01 -5.41469336e-01 2.42497697e-01
1.19497621e+00 8.65418836e-02 9.64393258e-01 -2.39873126e-01
-5.22127092e-01 4.97653067e-01 -2.18910709e-01 5.93682051e-01
6.68379784e-01 -1.50386775e+00 5.83118081e-01 7.49802828e-01
-1.58154458e-01 6.41669333e-01 7.65371382e-01 -5.09746611e-01
-9.39288199e-01 -9.41318989e-01 1.28561807e+00 -6.08120143e-01
5.47627509e-01 -1.50258586e-01 -1.13601828e+00 7.81798184e-01
1.02207077e+00 -7.86250114e-01 9.62603807e-01 1.71286806e-01
-3.06524307e-01 4.59064335e-01 -9.84887779e-01 7.45767951e-01
5.97308755e-01 -7.63338983e-01 -1.28796494e+00 6.94491208e-01
1.27947676e+00 -7.68509448e-01 -5.81071734e-01 2.08248962e-02
1.71016276e-01 -8.14213574e-01 6.42859221e-01 -8.47938836e-01
9.77633893e-01 -1.03500664e-01 -1.00039288e-01 -1.06602144e+00
7.42143616e-02 -8.03713202e-01 -2.98046529e-01 1.30777371e+00
9.48139668e-01 -3.70078146e-01 5.60403883e-01 8.52128863e-01
-2.81493038e-01 -7.58712471e-01 -6.31359577e-01 -2.35706735e-02
2.35274881e-01 -1.42013133e-01 6.29555583e-01 6.75675392e-01
5.29003084e-01 1.05677867e+00 -9.48278159e-02 3.90002169e-02
1.81646571e-01 2.68082201e-01 6.79319084e-01 -8.14654589e-01
-4.57007587e-01 -6.18626893e-01 6.63292110e-01 -1.77167678e+00
2.30434179e-01 -7.95499444e-01 3.71034354e-01 -1.98706305e+00
2.07918793e-01 1.35655925e-01 3.84364188e-01 2.84213513e-01
-6.59553647e-01 -2.86275655e-01 2.39583477e-01 2.20955893e-01
-9.44282115e-01 5.87295830e-01 1.27850223e+00 -4.45174545e-01
-2.23580137e-01 3.84712785e-01 -1.22714031e+00 7.75327504e-01
9.49896872e-01 2.80162282e-02 -3.97469819e-01 -7.89526343e-01
-4.07314785e-02 5.41688502e-01 1.74168944e-01 -4.67382014e-01
4.01114583e-01 -2.35769302e-01 5.18537126e-02 -6.31655097e-01
3.00249249e-01 2.89983060e-02 -8.14156294e-01 4.46462661e-01
-1.20799613e+00 -7.96493366e-02 -7.35567361e-02 3.82366121e-01
-4.16587472e-01 -5.58711886e-01 3.81779075e-01 -7.66032040e-02
-3.81020486e-01 -2.67640889e-01 -6.18264377e-01 5.40010035e-01
6.11925006e-01 4.57854182e-01 -1.06870997e+00 -1.09678018e+00
-6.30907953e-01 7.97410667e-01 -7.44044408e-02 4.19395238e-01
5.45867324e-01 -9.21370566e-01 -1.18049097e+00 -3.35153848e-01
5.41521981e-02 3.08361888e-01 2.13234499e-01 4.74432051e-01
-5.40201783e-01 6.48220778e-01 3.22823554e-01 -3.71415794e-01
-1.39249897e+00 -1.29127562e-01 3.11056286e-01 -6.11329257e-01
-2.61413306e-01 1.10853636e+00 1.34239599e-01 -5.38267612e-01
3.22467357e-01 -4.78061467e-01 -5.85184157e-01 9.35057327e-02
1.07514906e+00 1.59224644e-01 -9.86633152e-02 -3.14534009e-01
-1.50828257e-01 -1.00232728e-01 -3.17540228e-01 -4.30726081e-01
8.06250155e-01 -3.89092982e-01 -2.66964167e-01 1.91547424e-01
9.89115775e-01 -3.42365429e-02 -1.08934033e+00 -4.49059218e-01
1.81700960e-01 7.40637332e-02 -3.85842919e-01 -1.30026495e+00
-2.98100382e-01 8.61989617e-01 -3.26235704e-02 6.11473083e-01
8.29039812e-01 4.64413166e-01 8.34624529e-01 7.85210669e-01
-4.46650356e-01 -8.65097106e-01 4.09940779e-01 1.11333215e+00
1.42872894e+00 -1.45232904e+00 -2.60699362e-01 -3.80166650e-01
-9.64290202e-01 1.14352214e+00 8.69006097e-01 2.98960239e-01
1.01451436e-02 -5.43330703e-03 3.51889879e-01 -1.40776038e-01
-1.12172377e+00 -1.37134030e-01 4.37635034e-01 3.37749898e-01
8.64898980e-01 2.67278135e-01 -2.25345701e-01 9.42165673e-01
-6.23258293e-01 -3.69176656e-01 7.16816723e-01 7.67717838e-01
-6.51611626e-01 -9.22381878e-01 -3.09233367e-01 3.05607945e-01
-4.05831486e-01 -2.37871066e-01 -9.31495130e-01 4.88783538e-01
-5.16034544e-01 1.69072390e+00 1.81101616e-02 -3.70054245e-01
3.03650171e-01 1.93350494e-01 1.34011284e-01 -8.12383175e-01
-8.14219534e-01 -2.17362925e-01 9.29561436e-01 -1.84902057e-01
-2.30094731e-01 -4.96276706e-01 -1.63280129e+00 -4.32815969e-01
-1.15909986e-01 4.27217990e-01 2.49817729e-01 1.36821139e+00
2.51939118e-01 7.06161797e-01 5.71309388e-01 -1.92911148e-01
-9.43510354e-01 -1.42538202e+00 4.24764872e-01 5.40843070e-01
5.22391021e-01 1.35703206e-01 -4.71342891e-01 3.64691794e-01] | [11.91686725616455, 8.061984062194824] |
76ba1880-1be3-45ae-81d4-000ea6382d98 | map-disparity-estimation-using-hidden-markov | null | null | http://openaccess.thecvf.com/content_iccv_2015/html/Psota_MAP_Disparity_Estimation_ICCV_2015_paper.html | http://openaccess.thecvf.com/content_iccv_2015/papers/Psota_MAP_Disparity_Estimation_ICCV_2015_paper.pdf | MAP Disparity Estimation Using Hidden Markov Trees | A new method is introduced for stereo matching that operates on minimum spanning trees (MSTs) generated from the images. Disparity maps are represented as a collection of hidden states on MSTs, and each MST is modeled as a hidden Markov tree. An efficient recursive message-passing scheme designed to operate on hidden Markov trees, known as the upward-downward algorithm, is used to compute the maximum a posteriori (MAP) disparity estimate at each pixel. The messages processed by the upward-downward algorithm involve two types of probabilities: the probability of a pixel having a particular disparity given a set of per-pixel matching costs, and the probability of a disparity transition between a pair of connected pixels given their similarity. The distributions of these probabilities are modeled from a collection of images with ground truth disparities. Performance evaluation using the Middlebury stereo benchmark version 3 demonstrates that the proposed method ranks second and third in terms of overall accuracy when evaluated on the training and test image sets, respectively. | ['Mateusz Mittek', 'Jedrzej Kowalczuk', 'Eric T. Psota', 'Lance C. Perez'] | 2015-12-01 | null | null | null | iccv-2015-12 | ['stereo-matching'] | ['computer-vision'] | [ 7.53594637e-01 -7.49209290e-03 -3.20927799e-01 -7.10131288e-01
-9.42304015e-01 1.02099352e-01 6.50862336e-01 1.56060755e-01
-3.58778596e-01 5.63085437e-01 4.28445451e-02 -3.12633306e-01
2.83035427e-01 -9.83689368e-01 -6.81430876e-01 -7.49810457e-01
-1.62988991e-01 2.45054692e-01 7.59491622e-01 3.05160195e-01
7.10143864e-01 1.87744036e-01 -2.23101783e+00 5.18666983e-01
8.17271352e-01 1.16309428e+00 5.11523724e-01 8.88948262e-01
-1.20990127e-01 6.67134166e-01 -2.12597221e-01 -3.51923674e-01
4.88225639e-01 -3.79469335e-01 -8.48243296e-01 3.91815960e-01
8.42948496e-01 -4.98915613e-01 -2.70880371e-01 1.23870564e+00
2.31214851e-01 -2.06127092e-02 6.25053465e-01 -1.27641559e+00
2.07762614e-01 1.19078353e-01 -7.02262461e-01 1.89057156e-01
6.57671928e-01 1.97793588e-01 9.18422580e-01 -6.55287266e-01
5.29003441e-01 1.21296203e+00 5.22572339e-01 2.67070860e-01
-1.52640319e+00 -6.86919093e-01 -1.23679444e-01 3.83095682e-01
-1.31549263e+00 -4.91036803e-01 4.03333694e-01 -7.34427989e-01
1.07729173e+00 9.15907547e-02 6.87570333e-01 3.23435843e-01
6.93720102e-01 6.22173786e-01 1.43642020e+00 -6.00950718e-01
5.88047147e-01 -2.28116691e-01 2.03710929e-01 8.42680931e-01
2.16775298e-01 5.24484634e-01 -9.60896909e-01 -4.29022640e-01
6.68554187e-01 -3.37229848e-01 -1.45361573e-01 -5.65670192e-01
-1.19207871e+00 7.17621565e-01 3.17138761e-01 -3.81919891e-01
-4.34296370e-01 2.97655523e-01 -1.08597353e-02 7.60315955e-02
4.17402387e-02 -4.98159826e-01 -2.58279592e-02 1.36154830e-01
-1.00687921e+00 8.32350329e-02 6.84666872e-01 9.62597847e-01
1.65709758e+00 -4.02288854e-01 -7.04866722e-02 4.37835068e-01
6.54756188e-01 6.45786643e-01 2.39206970e-01 -1.24844694e+00
6.09530449e-01 5.41563392e-01 1.06946796e-01 -1.09669769e+00
6.22631647e-02 1.73802018e-01 -7.13442087e-01 8.90289307e-01
3.66116643e-01 1.47101328e-01 -1.43091249e+00 1.68146753e+00
3.37268919e-01 4.40191835e-01 1.94366962e-01 7.25579321e-01
5.40214121e-01 8.60677183e-01 7.90620521e-02 -1.24162279e-01
1.11163795e+00 -5.47205329e-01 -3.86776716e-01 -4.07635629e-01
3.11861485e-01 -9.94166911e-01 2.53989637e-01 1.43078357e-01
-1.13948345e+00 -7.70134866e-01 -1.21002209e+00 -6.93700016e-02
-7.23345354e-02 -3.23078841e-01 5.86399138e-01 6.85695767e-01
-1.32483888e+00 6.45842195e-01 -1.03432524e+00 -2.82708734e-01
1.39659807e-01 6.29744709e-01 -2.73335446e-03 -1.17747530e-01
-8.29206944e-01 7.69879699e-01 2.74979144e-01 9.61679071e-02
-1.00586569e+00 -8.94076824e-02 -1.08032167e+00 -1.37847126e-01
-4.43909705e-01 -9.94593799e-01 9.69915032e-01 -1.01031053e+00
-1.27390993e+00 1.42466688e+00 -8.28018367e-01 -5.60529649e-01
3.80065262e-01 2.77694911e-01 -1.78091228e-01 1.46448895e-01
3.88296306e-01 1.32302606e+00 1.15037966e+00 -1.32765234e+00
-1.49350035e+00 -3.73706937e-01 -1.20258570e-01 3.63250703e-01
3.32654655e-01 -2.50438273e-01 -5.31485498e-01 -1.92949191e-01
1.07036030e+00 -1.02584493e+00 -5.34227073e-01 1.78807136e-02
-6.42938018e-01 2.14595750e-01 4.30471748e-01 -4.98189509e-01
1.05320144e+00 -2.00526357e+00 -6.06328733e-02 5.53398788e-01
-8.07980895e-02 -4.89854723e-01 2.02531144e-01 2.31751859e-01
-1.71853062e-02 -4.51492071e-01 -4.36810195e-01 -3.67619455e-01
-3.15322489e-01 2.11208075e-01 -2.59628415e-01 5.42461336e-01
-1.40804321e-01 5.31146489e-02 -7.74814427e-01 -9.22017932e-01
5.13104439e-01 1.97624147e-01 -5.01115918e-01 1.61442816e-01
4.95099351e-02 3.36979032e-01 -2.28724584e-01 3.98006201e-01
9.89362001e-01 -4.65953164e-02 1.34362668e-01 -4.75806743e-02
-4.12997574e-01 4.33339536e-01 -1.49796414e+00 1.62692571e+00
-8.98965076e-02 6.39089346e-01 -3.98538202e-01 -3.69426578e-01
9.95362639e-01 2.57182837e-01 2.69287139e-01 -5.48547149e-01
-1.81905925e-01 2.58967370e-01 -2.76389837e-01 -2.55954772e-01
6.53640568e-01 -9.85920429e-02 4.99863774e-02 4.78614897e-01
-2.82892138e-01 -2.00207412e-01 3.16744864e-01 3.75705436e-02
9.71326530e-01 6.93570748e-02 1.53278559e-01 -3.06418508e-01
6.27217293e-01 2.96650201e-01 9.36613739e-01 8.21524084e-01
-2.73699582e-01 6.88232958e-01 1.87127575e-01 -5.26035249e-01
-9.28085327e-01 -1.34863782e+00 -3.93232286e-01 5.85965812e-01
6.47556424e-01 -4.37482670e-02 -6.72083139e-01 -9.65711661e-03
2.36412771e-02 3.25987101e-01 -3.83096546e-01 1.78045988e-01
-3.66092712e-01 -6.17715478e-01 -1.45187825e-01 2.34910980e-01
9.81136680e-01 -8.07020605e-01 -1.13887823e+00 1.28102809e-01
-4.12659079e-01 -1.06403160e+00 -2.63187259e-01 8.53779688e-02
-1.51114571e+00 -1.03667402e+00 -6.60775006e-01 -1.35244715e+00
9.55473125e-01 4.53086972e-01 9.78884041e-01 -4.86806035e-02
-1.94110885e-01 3.06613371e-02 2.79735208e-01 6.50605336e-02
-5.12447953e-01 -3.17536771e-01 -2.43149430e-01 3.30957711e-01
4.52998459e-01 -5.82071781e-01 -8.57271254e-01 4.26215053e-01
-5.03177106e-01 6.57349527e-01 5.99880040e-01 5.83894730e-01
9.78031814e-01 1.13451935e-01 -4.27583098e-01 -5.09668589e-01
-2.11735591e-01 -9.62402374e-02 -1.08072555e+00 3.34156677e-02
-4.77329344e-01 3.02012324e-01 -2.29617596e-01 5.07106781e-02
-1.30427492e+00 7.27991462e-01 2.31351838e-01 -1.74214635e-02
-2.69448191e-01 1.33248344e-01 1.95328370e-01 -7.73159564e-02
3.97193819e-01 2.49548286e-01 -1.64498061e-01 -1.32270023e-01
1.51124716e-01 6.31119788e-01 8.63492131e-01 -3.07044595e-01
6.94022894e-01 1.04655957e+00 3.72281224e-01 -7.15267301e-01
-5.11937678e-01 -6.65172279e-01 -8.49906981e-01 -4.14703578e-01
1.09266412e+00 -1.11706102e+00 -4.52106118e-01 9.53026712e-01
-1.08085454e+00 -1.55271247e-01 8.90107304e-02 6.88688219e-01
-8.37457716e-01 4.89089817e-01 -5.64704597e-01 -7.43537664e-01
-7.87810162e-02 -1.26645839e+00 1.03523612e+00 4.48271334e-01
-1.67590186e-01 -9.04787838e-01 3.25937927e-01 4.24603760e-01
-6.57307431e-02 2.01623544e-01 1.09283829e+00 3.41449469e-01
-1.31536424e+00 -1.45909473e-01 -1.71714842e-01 2.71498978e-01
6.98871091e-02 2.12904155e-01 -1.25903118e+00 -3.21166396e-01
-8.80357996e-02 1.18976079e-01 9.54949677e-01 8.47438276e-01
4.87068564e-01 1.48618311e-01 -6.09451592e-01 5.35021782e-01
1.75651217e+00 3.18642229e-01 9.71227765e-01 5.14498353e-01
2.23076910e-01 5.92832386e-01 7.02563703e-01 4.25314993e-01
5.70492089e-01 5.78742921e-01 4.44063008e-01 -2.23100007e-01
-1.59181282e-01 -3.33448350e-01 2.71595269e-01 6.42222762e-01
1.44336419e-02 3.48883241e-01 -1.07484126e+00 5.36103964e-01
-1.70751297e+00 -1.00714052e+00 -4.76804048e-01 2.85092950e+00
6.82583928e-01 5.26664019e-01 -2.51188010e-01 3.02257866e-01
1.34026885e+00 1.79355055e-01 -3.88208330e-01 -3.75376999e-01
-2.49914452e-02 6.04013056e-02 6.84675574e-01 8.51248384e-01
-9.70434368e-01 7.46470988e-01 7.26124287e+00 2.65548736e-01
-9.05287683e-01 -2.16716439e-01 7.07617223e-01 5.00703514e-01
-1.04589067e-01 6.54308975e-01 -1.05122662e+00 4.48234230e-01
6.35217667e-01 -1.84083864e-01 6.01422526e-02 6.20791197e-01
2.60830671e-01 -1.11249065e+00 -1.26775515e+00 1.04907143e+00
-1.80852905e-01 -1.24032485e+00 2.73165464e-01 1.76635236e-01
1.11454344e+00 7.11631253e-02 2.92640626e-01 -2.60439426e-01
4.28496003e-01 -5.16610622e-01 7.64315486e-01 2.90957451e-01
6.00815833e-01 -5.41804969e-01 5.06230295e-01 3.38107467e-01
-1.44465649e+00 1.02243356e-01 -4.68036205e-01 -2.91484326e-01
1.71395898e-01 6.09005988e-01 -6.03682697e-01 1.20588303e-01
8.27044129e-01 4.94587421e-01 -3.09944659e-01 1.31555200e+00
-3.05142105e-01 7.96054676e-02 -2.46201381e-01 3.63491327e-01
6.74617738e-02 -3.84656698e-01 4.42986846e-01 9.18194234e-01
1.34017348e-01 -7.49249663e-03 9.51011479e-02 6.04044735e-01
2.93363541e-01 -3.33900094e-01 -4.83383566e-01 8.03929925e-01
5.47757745e-01 7.76706338e-01 -7.54069090e-01 -7.15740383e-01
-4.24040556e-01 9.65906501e-01 -4.24749292e-02 3.93157989e-01
-4.60807800e-01 -2.73708701e-01 7.90380597e-01 -5.01124300e-02
2.67466307e-01 -8.09994936e-02 -6.65843844e-01 -7.73045421e-01
-6.05550557e-02 -4.25910681e-01 6.98017061e-01 -1.00776064e+00
-8.10262859e-01 5.19149840e-01 1.19727157e-01 -1.57906330e+00
-3.04401577e-01 -2.35979155e-01 -7.32946634e-01 1.12402654e+00
-1.94259119e+00 -4.82468039e-01 -5.48042059e-01 7.27008164e-01
5.03899336e-01 2.30930060e-01 6.93538725e-01 1.07006155e-01
-3.54654968e-01 8.61479342e-02 -1.20806910e-01 1.16691217e-02
2.36962363e-01 -1.00874662e+00 5.98355293e-01 1.12136114e+00
-1.56638607e-01 2.28615165e-01 8.26067567e-01 -6.26463652e-01
-1.08198237e+00 -8.83695781e-01 1.40631926e+00 -9.89434645e-02
1.31199241e-01 1.34464383e-01 -6.00736737e-01 8.44532311e-01
9.06105712e-02 -3.66016507e-01 3.51220816e-01 -2.24785581e-01
-6.45204633e-02 -3.01381163e-02 -1.21996689e+00 5.58670700e-01
9.30073977e-01 -5.90153039e-01 -7.37445951e-01 3.37810186e-03
4.55029979e-02 -5.44636250e-01 -6.59380555e-01 4.44829494e-01
5.61483264e-01 -1.66498923e+00 7.82337070e-01 8.80066454e-02
5.06200433e-01 -4.95668799e-01 -4.49543387e-01 -8.17903101e-01
-3.50794494e-01 -5.69004953e-01 4.12826151e-01 8.53270829e-01
4.82921988e-01 -6.11750364e-01 1.23880899e+00 6.81296110e-01
2.48070844e-02 -3.32152367e-01 -1.33146262e+00 -6.88288450e-01
-4.29760337e-01 -4.19470906e-01 4.14518297e-01 6.02078259e-01
-1.01305827e-01 1.53747842e-01 -1.57837160e-02 7.08545029e-01
1.36678028e+00 5.80430627e-01 8.71739864e-01 -1.04616952e+00
-2.34669045e-01 -1.74635693e-01 -1.00192845e+00 -1.51830482e+00
-1.79570705e-01 -5.99858880e-01 4.45978999e-01 -1.60361886e+00
5.51060200e-01 -6.46499813e-01 -1.50223272e-02 3.82277742e-02
-1.10460103e-01 1.18748516e-01 -6.10948168e-02 5.24329484e-01
-1.91562325e-01 1.97453335e-01 8.48353684e-01 -6.08528145e-02
-2.08217457e-01 2.62080818e-01 3.01378340e-01 9.85529184e-01
5.12773752e-01 -7.55252540e-01 -1.72918051e-01 -4.57906246e-01
-5.88005893e-02 4.72131878e-01 1.43566728e-01 -1.38084590e+00
5.54851890e-01 -1.45871669e-01 1.80839732e-01 -1.11425173e+00
6.17760241e-01 -6.62596881e-01 5.06024420e-01 1.00730419e+00
-1.64770812e-01 1.49684429e-01 -1.24383800e-01 6.36782944e-01
-4.28646415e-01 -5.79544790e-02 1.15520346e+00 2.57651117e-02
-1.24493837e+00 2.68119067e-01 -5.81964791e-01 -2.20393971e-01
1.06056225e+00 -9.59947169e-01 -2.09459234e-02 -3.63536000e-01
-7.01101482e-01 2.15508923e-01 7.29898512e-01 5.19851372e-02
8.84076357e-01 -1.39785242e+00 -5.72683215e-01 5.88437438e-01
2.70599991e-01 -8.53652582e-02 1.68066502e-01 5.82114220e-01
-7.60944068e-01 4.10296530e-01 -4.26160753e-01 -1.23525178e+00
-1.63066161e+00 1.17044114e-01 4.26076829e-01 -9.69098583e-02
-5.10971010e-01 8.60105634e-01 3.79946083e-01 1.10843539e-01
3.69251758e-01 -5.41090906e-01 7.60719329e-02 -3.44697386e-01
3.00540268e-01 5.56623518e-01 -1.46111503e-01 -9.40840244e-01
-2.49515608e-01 1.08018851e+00 2.09502548e-01 -6.30169809e-01
6.34177089e-01 -3.97234499e-01 -8.07062387e-02 4.79721457e-01
1.07399082e+00 -4.99635875e-01 -1.60294235e+00 -3.92423570e-01
8.13454166e-02 -1.13039219e+00 1.82893425e-01 -2.18744889e-01
-9.10770357e-01 8.57872546e-01 1.05371010e+00 -7.72898868e-02
1.05985308e+00 -2.29069695e-01 9.25389528e-01 2.27085687e-02
8.51684749e-01 -1.01266265e+00 -2.25737661e-01 3.40335369e-01
2.10427821e-01 -1.32459033e+00 -6.70366809e-02 -7.89566815e-01
-4.59732383e-01 9.42590594e-01 2.83604413e-01 -4.10800092e-02
7.51082182e-01 1.90118670e-01 2.71705300e-01 1.11407842e-02
-8.69966805e-01 -3.99863213e-01 -8.79496336e-03 7.01520085e-01
6.69758096e-02 1.80014782e-02 -2.07560826e-02 -7.04508185e-01
-1.20843552e-01 1.03304401e-01 4.39627022e-01 1.32225215e+00
-1.04100537e+00 -9.14598465e-01 -6.82921588e-01 1.86996534e-01
-1.60144120e-01 -1.42344669e-01 1.50980979e-01 3.17790061e-01
1.11103669e-01 1.10417426e+00 5.61646879e-01 -4.58306372e-01
1.62691489e-01 -2.35099167e-01 5.92335820e-01 -6.91238701e-01
-8.32271799e-02 -2.11754233e-01 6.03998452e-02 -6.68527484e-01
-5.99281967e-01 -8.49024951e-01 -1.41856909e+00 -5.11082709e-01
-1.79093778e-01 -1.76191390e-01 8.67258132e-01 7.32612193e-01
1.11257769e-01 -2.38371223e-01 9.67297137e-01 -1.00181925e+00
-2.95272678e-01 -3.15787107e-01 -4.76944596e-01 4.07772958e-01
2.69006044e-01 -3.65070850e-01 -5.49312890e-01 4.88830954e-01] | [9.068675994873047, -2.5123705863952637] |
ce4d9056-4fe4-4e0f-aa3f-7c6d747735ea | neural-structure-mapping-for-learning-1 | null | null | https://openreview.net/forum?id=l-TLGjxwajn | https://openreview.net/pdf?id=l-TLGjxwajn | Neural Structure Mapping For Learning Abstract Visual Analogies | Building conceptual abstractions from sensory information and then reasoning about them is central to human intelligence. Abstract reasoning both relies on, and is facilitated by, our ability to make analogies about concepts from known domains to novel domains. Structure Mapping Theory of human analogical reasoning posits that analogical mappings rely on (higher-order) relations and not on the sensory content of the domain. This enables humans to reason systematically about novel domains, a problem with which machine learning (ML) models tend to struggle. We introduce a two-stage neural framework, which we label Neural Structure Mapping (NSM), to learn visual analogies from Raven's Progressive Matrices, an abstract visual reasoning test of fluid intelligence. Our framework uses (1) a multi-task visual relationship encoder to extract constituent concepts from raw visual input in the source domain, and (2) a neural module net analogy inference engine to reason compositionally about the inferred relation in the target domain. Our NSM approach (a) isolates the relational structure from the source domain with high accuracy, and (b) successfully utilizes this structure for analogical reasoning in the target domain. | ['Graham W. Taylor', 'Shashank Shekhar'] | 2021-10-12 | neural-structure-mapping-for-learning | null | null | neurips-workshop-svrhm-2021-12 | ['visual-analogies'] | ['computer-vision'] | [ 3.46585989e-01 2.97885269e-01 1.32324800e-01 -4.43702698e-01
-7.73860700e-03 -7.25214124e-01 9.39775467e-01 5.29124558e-01
-3.03719610e-01 5.68830609e-01 3.51440102e-01 -7.02445745e-01
-3.80051821e-01 -9.91274893e-01 -7.83375025e-01 1.53986454e-01
1.12380907e-01 9.37383115e-01 3.10240030e-01 -4.96233344e-01
5.13185501e-01 5.38864732e-01 -1.32101190e+00 8.74504268e-01
9.14261341e-01 7.86749303e-01 4.17432934e-01 6.52986228e-01
-7.07616985e-01 1.62878633e+00 -2.33726069e-01 -7.19583690e-01
1.86454371e-01 -8.67369115e-01 -1.51312053e+00 -4.59894717e-01
4.31530088e-01 -3.40147913e-01 -2.36260429e-01 9.86355841e-01
-3.00740302e-01 3.17980438e-01 8.88387144e-01 -1.11293066e+00
-1.60851443e+00 7.75463700e-01 -1.94138408e-01 5.19284606e-01
7.41418839e-01 1.28399789e-01 1.23215985e+00 -8.51827621e-01
6.75998509e-01 1.60886145e+00 7.23581493e-01 4.55668628e-01
-1.73128390e+00 -4.03408468e-01 7.92476460e-02 7.60375440e-01
-9.69694853e-01 -1.75708622e-01 7.50058532e-01 -6.91968203e-01
1.22998273e+00 -3.37196253e-02 1.12105668e+00 9.25902009e-01
1.72402814e-01 3.91997665e-01 1.43208408e+00 -5.92869937e-01
3.65085065e-01 1.53583974e-01 2.89211750e-01 5.73405862e-01
1.31530881e-01 5.38364828e-01 -5.10619521e-01 1.93491742e-01
1.07955897e+00 2.51360573e-02 -2.92964764e-02 -3.19024861e-01
-1.28817058e+00 7.34115779e-01 1.00525057e+00 4.46243823e-01
-5.60721457e-01 1.49736330e-01 3.37530732e-01 7.51288116e-01
-3.91890734e-01 9.88490283e-01 -7.45117813e-02 4.16166604e-01
-5.09460628e-01 3.78773719e-01 6.48358047e-01 8.79294515e-01
7.55783319e-01 8.94767218e-05 1.80491921e-03 5.28506160e-01
4.46688980e-01 3.16189706e-01 6.66623056e-01 -1.03841627e+00
3.74456286e-01 7.00932860e-01 -1.86863214e-01 -9.56422627e-01
-3.36988360e-01 -8.88374969e-02 -3.06664884e-01 4.74768639e-01
4.49927658e-01 5.07311106e-01 -5.29554427e-01 2.06706190e+00
5.68453856e-02 -2.60504603e-01 3.92132998e-01 7.97715366e-01
7.78061450e-01 6.00341916e-01 3.67139250e-01 7.42503107e-02
1.50837517e+00 -5.06514907e-01 -3.37605625e-01 -6.82777226e-01
3.84493142e-01 -2.71815300e-01 1.56458557e+00 7.59944990e-02
-1.40993166e+00 -9.30278540e-01 -1.27695251e+00 -7.62785494e-01
-7.06426084e-01 -7.07969546e-01 7.23298252e-01 1.11323960e-01
-1.22162902e+00 8.19747746e-01 -4.04793829e-01 -5.81467807e-01
6.13489032e-01 2.80003939e-02 -4.11413133e-01 -1.89910695e-01
-1.23131847e+00 1.56515574e+00 8.96986127e-01 -3.14484686e-01
-4.77934390e-01 -8.94407868e-01 -1.06950760e+00 1.83942899e-01
1.62268713e-01 -1.31661224e+00 1.30230868e+00 -1.33332121e+00
-1.07099903e+00 1.05870855e+00 -3.08780894e-02 -5.27777076e-01
9.27873072e-04 -9.61760953e-02 -3.87196898e-01 3.51295978e-01
2.54048407e-01 7.63083696e-01 6.49723351e-01 -1.30362856e+00
-4.14425313e-01 -4.43511307e-01 3.52849394e-01 2.50693500e-01
1.48136824e-01 -2.69440383e-01 4.04535323e-01 -6.56480968e-01
3.40492129e-01 -5.89709282e-01 1.44573227e-01 4.42383111e-01
5.75591102e-02 -3.01378906e-01 3.04955214e-01 -9.00261998e-01
4.83286262e-01 -2.02851319e+00 5.29767156e-01 2.98882902e-01
4.64697570e-01 -1.21319033e-01 -2.70608693e-01 4.15682524e-01
-4.57499743e-01 -2.25841686e-01 -5.69886193e-02 2.93558180e-01
5.26166499e-01 3.98303956e-01 -8.66795063e-01 -2.10434079e-01
2.39520639e-01 1.55113137e+00 -1.17780769e+00 -3.08695972e-01
2.23133907e-01 -9.54601541e-02 -7.01713204e-01 3.82280201e-01
-4.12050217e-01 2.54759848e-01 1.76899403e-01 1.42142221e-01
3.29816639e-01 -3.48043084e-01 3.60741973e-01 -3.40538353e-01
2.11491093e-01 6.65661871e-01 -7.89244413e-01 1.91290855e+00
-4.91871178e-01 5.87310970e-01 -6.84878409e-01 -1.06888723e+00
8.33304346e-01 1.56311497e-01 -4.32140708e-01 -1.28291786e+00
5.59745468e-02 8.31404552e-02 4.52984542e-01 -4.90673929e-01
1.59558013e-01 -8.59687507e-01 1.02739029e-01 6.25642955e-01
2.71324217e-01 -5.09436369e-01 1.28287405e-01 4.25190598e-01
7.66588390e-01 4.95568216e-01 9.16381299e-01 -3.40750962e-01
4.69337553e-01 1.32859170e-01 1.63396150e-01 6.74227893e-01
2.27819413e-01 -8.34031329e-02 5.12782454e-01 -8.65684569e-01
-1.37114680e+00 -2.04753184e+00 1.79251686e-01 1.01912808e+00
-7.59826824e-02 -1.15810975e-01 -3.32100362e-01 -1.65329009e-01
7.12670982e-02 1.37261319e+00 -8.05838168e-01 -5.23294091e-01
-4.12720323e-01 6.81916103e-02 2.66876221e-01 1.09034669e+00
4.22676355e-01 -1.48419428e+00 -9.07344222e-01 1.09351911e-02
8.27108845e-02 -1.05614686e+00 1.05806872e-01 1.39121503e-01
-9.27993238e-01 -1.09064806e+00 -1.23505987e-01 -8.73005867e-01
5.25644422e-01 -1.16990894e-01 1.50383377e+00 8.43758732e-02
-2.75613934e-01 5.08565128e-01 -1.62313670e-01 -3.63806248e-01
-8.12233269e-01 -6.45390272e-01 -2.01533660e-02 -4.93071705e-01
7.51768649e-01 -1.03813279e+00 -1.50353774e-01 -4.47683632e-02
-8.60341966e-01 5.14868677e-01 7.71458924e-01 7.49282122e-01
3.38637233e-01 -3.35270792e-01 5.40628552e-01 -6.70235038e-01
7.82831788e-01 -4.79307681e-01 -4.87820417e-01 3.46643031e-01
-4.90862727e-01 4.88795459e-01 8.49192142e-01 -5.72328508e-01
-1.12250221e+00 -2.84805506e-01 2.58243501e-01 -4.67018753e-01
-1.88794881e-01 6.04373991e-01 -1.16446659e-01 2.66023070e-01
1.08662117e+00 3.78638983e-01 2.62769554e-02 -1.87485471e-01
9.23018515e-01 9.56071243e-02 1.11982644e+00 -1.11620724e+00
9.98504043e-01 3.28440964e-01 3.14761698e-01 -4.73585427e-01
-9.36012506e-01 2.60498494e-01 -9.85609412e-01 2.16814965e-01
1.11715889e+00 -7.57220030e-01 -1.03356838e+00 -2.63604164e-01
-1.26287603e+00 -4.01614845e-01 -7.61874974e-01 5.65000117e-01
-9.20823634e-01 1.96396187e-01 -6.80050492e-01 -4.33629304e-01
4.20206264e-02 -6.62355959e-01 3.11676115e-01 9.38542932e-02
-1.01827347e+00 -1.20375526e+00 9.71228108e-02 2.45296180e-01
2.96866864e-01 8.34325925e-02 1.89117646e+00 -7.56644428e-01
-5.57908773e-01 3.00377637e-01 -6.50716066e-01 3.74809429e-02
-2.21985560e-02 -7.58769631e-01 -8.94550562e-01 -8.81346036e-03
1.31299600e-01 -6.62511587e-01 5.44340730e-01 -1.12199478e-01
9.66744483e-01 -3.54557157e-01 -3.57440598e-02 4.80506361e-01
1.36240661e+00 3.66666675e-01 8.53333294e-01 2.73786783e-01
6.01084769e-01 7.91528165e-01 -2.62159985e-02 -7.01588616e-02
7.39838541e-01 4.47215140e-01 -1.09633848e-01 1.80069238e-01
-3.10809344e-01 -7.02643931e-01 1.15859270e-01 5.23981333e-01
-5.32885455e-02 6.40214145e-01 -1.03011441e+00 5.37318885e-01
-1.58678985e+00 -1.32795763e+00 2.25530580e-01 1.92515194e+00
1.21124768e+00 1.79942042e-01 2.83697192e-02 1.20841987e-01
3.05673957e-01 -3.62089068e-01 -7.84261823e-01 -8.84915888e-01
5.12206033e-02 5.22138953e-01 -3.73910606e-01 6.36420250e-01
-4.18560117e-01 9.42053437e-01 6.31770992e+00 1.85800016e-01
-4.12552744e-01 -1.80245385e-01 -4.87132668e-02 1.62175670e-01
-5.41000366e-01 2.36423269e-01 -3.17379124e-02 -4.52146307e-02
9.76568401e-01 -3.67760569e-01 1.05391967e+00 4.90740567e-01
-5.17985165e-01 2.73951411e-01 -2.14685035e+00 9.99221742e-01
1.37358248e-01 -1.38168681e+00 8.27678680e-01 -2.62761891e-01
2.42851421e-01 -5.58571517e-01 8.01868662e-02 4.58826780e-01
7.31869876e-01 -1.43591499e+00 8.93376887e-01 8.04806948e-01
7.49510944e-01 -5.73399782e-01 3.13471973e-01 1.80611834e-01
-1.19211197e+00 -3.49698097e-01 -5.40984988e-01 -6.56850219e-01
-1.41747370e-01 -1.48979858e-01 -7.17319608e-01 2.65162617e-01
6.55757546e-01 7.18940794e-01 -6.44303381e-01 4.80997831e-01
-7.81377375e-01 -1.42051592e-01 2.39264876e-01 5.55564426e-02
-9.66364592e-02 -1.30548671e-01 1.89242035e-01 7.14258194e-01
-2.38847695e-02 4.89904940e-01 -1.68999523e-01 1.66163266e+00
1.05463728e-01 -1.94511503e-01 -8.85502398e-01 -1.82306886e-01
6.35968804e-01 7.08703339e-01 -5.58035851e-01 -4.58641112e-01
-8.01863730e-01 1.14749551e+00 8.09740901e-01 3.37904185e-01
-4.07855242e-01 -3.12992215e-01 5.03817618e-01 2.38835197e-02
9.30461809e-02 -2.24571794e-01 -5.67192376e-01 -8.52240145e-01
-4.47501428e-02 -8.79560888e-01 5.93824387e-01 -1.54320383e+00
-1.69593191e+00 3.47186297e-01 2.50356227e-01 -8.22315216e-01
-5.41250169e-01 -1.10097218e+00 -5.86707771e-01 1.27151310e+00
-1.02366233e+00 -1.21111023e+00 -2.36634687e-01 7.61820138e-01
2.97889024e-01 -2.61300385e-01 1.06781328e+00 -4.54543531e-01
4.21290040e-01 4.35175270e-01 -5.53676248e-01 3.80132347e-01
2.32615471e-01 -1.47035396e+00 8.96417141e-01 4.41129804e-01
4.09915566e-01 1.04730654e+00 6.45111024e-01 -5.45713484e-01
-1.20733309e+00 -5.20060956e-01 8.46924365e-01 -9.09613967e-01
9.48974013e-01 -3.72796863e-01 -1.40187681e+00 1.16996050e+00
3.32873821e-01 -2.65842825e-01 1.01117802e+00 1.11621268e-01
-1.20710576e+00 1.16618231e-01 -1.06970954e+00 1.03763366e+00
1.22326660e+00 -1.11011231e+00 -2.01447725e+00 1.01212226e-01
9.74456787e-01 -1.15248837e-01 -9.68881428e-01 -1.98340341e-01
7.36450553e-01 -8.63564849e-01 1.36047792e+00 -1.34388471e+00
8.98729205e-01 -4.08258021e-01 -5.90783119e-01 -1.53627408e+00
-8.27811420e-01 -1.57262683e-02 -1.16517298e-01 5.92012644e-01
3.54892105e-01 -7.24826753e-01 2.21609890e-01 8.14108372e-01
4.37191054e-02 -2.01260373e-01 -6.88035846e-01 -8.61936688e-01
4.72895533e-01 -3.74586970e-01 7.48192787e-01 1.12254262e+00
4.93972570e-01 9.21668410e-01 2.81114459e-01 7.44579732e-02
6.60989165e-01 4.23910379e-01 4.28504497e-01 -1.76222301e+00
-5.00308454e-01 -5.91846049e-01 -6.12714171e-01 -6.67120278e-01
4.58322614e-01 -1.46141481e+00 -4.49029535e-01 -1.74830234e+00
3.95252705e-01 1.50166571e-01 -2.69773453e-01 5.51341951e-01
8.05650055e-02 1.56210773e-02 4.63783354e-01 2.52917945e-01
-1.86110482e-01 3.49846452e-01 1.24696720e+00 -6.43522665e-02
1.03193391e-02 -7.97683835e-01 -1.08606434e+00 9.08331573e-01
5.51333427e-01 -1.23454288e-01 -8.64519238e-01 -5.40683270e-01
8.63333166e-01 1.17774546e-01 1.00513840e+00 -1.09472239e+00
2.97389746e-01 -3.36575925e-01 1.22340155e+00 -1.88001096e-01
3.32754821e-01 -9.82583642e-01 -3.16436887e-02 7.04501748e-01
-6.83572292e-01 6.08734906e-01 5.81461012e-01 3.07577819e-01
5.87460436e-02 -1.95931002e-01 8.73879969e-01 -4.40532357e-01
-1.22715044e+00 -2.63692737e-01 -2.30685826e-02 4.14317578e-01
7.55639732e-01 -3.91974181e-01 -5.48797309e-01 -2.18630224e-01
-8.95935059e-01 -2.51825191e-02 6.46047175e-01 4.56289858e-01
9.47566807e-01 -1.64769268e+00 -6.53834581e-01 1.74367949e-01
3.37371171e-01 -4.04569864e-01 1.14803761e-01 4.68970090e-01
-4.73030180e-01 2.31998295e-01 -9.18736041e-01 -2.39889458e-01
-6.99917197e-01 1.31788206e+00 4.23049480e-01 3.15773696e-01
-9.47573125e-01 7.86672890e-01 5.90154707e-01 -3.56532663e-01
-4.10320371e-01 -4.74823445e-01 -2.91968703e-01 -2.29576349e-01
8.56171787e-01 1.68843076e-01 -4.26732719e-01 -4.84276623e-01
-2.04139769e-01 6.17416859e-01 -2.84102470e-01 -4.13399875e-01
1.04831743e+00 -1.23901526e-02 -5.04616439e-01 8.65631342e-01
8.56210530e-01 -4.55400020e-01 -8.19688737e-01 -6.43182278e-01
1.75978437e-01 -4.27503765e-01 -4.81649876e-01 -1.06340611e+00
-5.55245765e-02 9.82451439e-01 2.01666608e-01 -4.55320766e-03
1.03911030e+00 3.92321378e-01 4.87776905e-01 7.87506640e-01
1.38138860e-01 -8.24981272e-01 4.32636201e-01 5.35843134e-01
1.20879114e+00 -8.44347835e-01 8.05202499e-02 -4.63095419e-02
-6.45925701e-01 1.22588134e+00 6.62379086e-01 -2.39607334e-01
2.82676965e-01 -3.28232497e-02 -1.95469797e-01 -3.92551392e-01
-8.62195730e-01 2.31019184e-02 5.71957469e-01 9.75644350e-01
3.46270502e-01 -1.33625686e-01 3.14439595e-01 7.14787126e-01
-7.59432614e-01 -5.57736382e-02 2.18642727e-01 7.59521008e-01
-4.76489097e-01 -5.78946590e-01 -4.92034942e-01 3.84844244e-01
2.76166439e-01 -4.09507245e-01 -6.54375613e-01 8.30727637e-01
1.08584873e-01 5.34744143e-01 4.32065308e-01 -1.39514133e-01
4.21684206e-01 5.06676078e-01 1.10238075e+00 -5.76502621e-01
-2.12288633e-01 -8.17525804e-01 -2.79990405e-01 -6.78497374e-01
-2.99529433e-01 -4.56322879e-01 -1.58312714e+00 -4.95862991e-01
7.51382351e-01 -2.86265858e-03 1.95040137e-01 1.36001670e+00
2.07590666e-02 6.49625003e-01 -2.14068279e-01 -6.06097102e-01
-5.16744494e-01 -7.82855988e-01 -5.50293565e-01 8.69806409e-01
2.12178543e-01 -7.24360287e-01 8.75934958e-03 1.79336414e-01] | [10.595159530639648, 2.309239387512207] |
0925c820-ce2c-420e-9c43-b1abc0dda90f | sparse-local-patch-transformer-for-robust | 2203.06541 | null | https://arxiv.org/abs/2203.06541v2 | https://arxiv.org/pdf/2203.06541v2.pdf | Sparse Local Patch Transformer for Robust Face Alignment and Landmarks Inherent Relation Learning | Heatmap regression methods have dominated face alignment area in recent years while they ignore the inherent relation between different landmarks. In this paper, we propose a Sparse Local Patch Transformer (SLPT) for learning the inherent relation. The SLPT generates the representation of each single landmark from a local patch and aggregates them by an adaptive inherent relation based on the attention mechanism. The subpixel coordinate of each landmark is predicted independently based on the aggregated feature. Moreover, a coarse-to-fine framework is further introduced to incorporate with the SLPT, which enables the initial landmarks to gradually converge to the target facial landmarks using fine-grained features from dynamically resized local patches. Extensive experiments carried out on three popular benchmarks, including WFLW, 300W and COFW, demonstrate that the proposed method works at the state-of-the-art level with much less computational complexity by learning the inherent relation between facial landmarks. The code is available at the project website. | ['Min Xu', 'Xi Wang', 'JianGuo Zhang', 'Wenjian Huang', 'Weiwei qu', 'Jiahao Xia'] | 2022-03-13 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Xia_Sparse_Local_Patch_Transformer_for_Robust_Face_Alignment_and_Landmarks_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Xia_Sparse_Local_Patch_Transformer_for_Robust_Face_Alignment_and_Landmarks_CVPR_2022_paper.pdf | cvpr-2022-1 | ['robust-face-alignment', 'face-alignment'] | ['computer-vision', 'computer-vision'] | [-1.51688486e-01 6.65754378e-02 -2.79934943e-01 -5.80809057e-01
-7.71447778e-01 1.58396345e-02 6.38471365e-01 -3.35168064e-01
4.49010096e-02 4.85615194e-01 2.13264272e-01 4.79872108e-01
-9.17835012e-02 -7.11543918e-01 -7.47524619e-01 -7.52436996e-01
7.79903308e-02 3.28475118e-01 1.83912709e-01 -2.55629141e-02
2.40650788e-01 5.50451577e-01 -1.53541315e+00 7.02879354e-02
8.22401285e-01 1.26915288e+00 1.10596038e-01 -6.83585480e-02
-1.62713259e-01 3.42692077e-01 -2.08637610e-01 -2.83198118e-01
5.98438323e-01 -3.85511488e-01 -5.86933613e-01 5.58671653e-02
9.81297731e-01 -2.12446712e-02 2.61436813e-02 1.15245247e+00
4.52732027e-01 6.00982807e-04 5.05748749e-01 -1.16520000e+00
-7.13177145e-01 2.14568138e-01 -1.20405948e+00 1.91842783e-02
3.12533766e-01 -1.37659162e-01 9.78681982e-01 -1.36382008e+00
6.09343827e-01 1.38939202e+00 8.18338990e-01 3.24322373e-01
-1.16473663e+00 -9.96328473e-01 3.56315732e-01 3.85352731e-01
-1.89217746e+00 -5.55016696e-01 1.09269667e+00 -2.39587069e-01
5.69830596e-01 2.30398085e-02 5.96375346e-01 4.72033679e-01
2.35292941e-01 4.97054964e-01 1.21819723e+00 -4.76385176e-01
-4.10988666e-02 -2.19069958e-01 -1.88053310e-01 1.24914098e+00
-1.88376233e-01 -3.23725231e-02 -6.77237570e-01 -1.62457563e-02
9.33773398e-01 7.03815520e-02 -9.79252234e-02 -3.93649250e-01
-8.14144015e-01 6.65997624e-01 8.34256828e-01 2.98590034e-01
-2.99618632e-01 1.61873952e-01 -7.35874996e-02 -1.21375145e-02
7.52603650e-01 -6.55449331e-02 -2.99266845e-01 2.19003841e-01
-1.22456682e+00 -1.45426482e-01 3.53918016e-01 9.54370856e-01
1.50867629e+00 -1.36164829e-01 -2.12294310e-01 1.09544504e+00
4.86993015e-01 3.31940889e-01 4.82475251e-01 -8.62438560e-01
3.15120041e-01 8.57230842e-01 -1.09125204e-01 -1.46836197e+00
-3.29415172e-01 -3.27823192e-01 -9.06054080e-01 3.42121154e-01
1.44672737e-01 7.40102828e-02 -1.20616543e+00 1.61958039e+00
7.00841427e-01 8.24019432e-01 -5.95211267e-01 7.33796179e-01
8.02661240e-01 6.86882198e-01 7.78733268e-02 -1.42896026e-01
1.13890457e+00 -1.26228118e+00 -5.06686091e-01 -1.49596557e-01
1.38160273e-01 -7.78362215e-01 9.50498760e-01 -5.08518815e-02
-9.37353909e-01 -9.88922179e-01 -9.63586628e-01 -2.86348850e-01
-3.31491947e-01 4.62886512e-01 4.68090832e-01 4.59143847e-01
-1.37659037e+00 4.10211682e-01 -9.45212543e-01 -4.06450897e-01
8.27088416e-01 4.76317853e-01 -5.92581928e-01 -5.28973644e-04
-6.96662366e-01 6.10626400e-01 -9.43355560e-02 4.92334038e-01
-6.90574229e-01 -9.44340527e-01 -9.52696323e-01 6.72637299e-02
1.64757222e-01 -3.79201293e-01 6.74261928e-01 -1.13798690e+00
-1.66233027e+00 9.01623011e-01 -5.76207161e-01 1.37355939e-01
3.93503547e-01 -8.15100446e-02 -2.12563187e-01 2.01390572e-02
3.13353211e-01 8.90438318e-01 1.06016958e+00 -1.18400204e+00
-7.58000910e-01 -5.62863708e-01 -2.19857693e-01 1.63814202e-01
-2.69921929e-01 1.24085627e-01 -7.79163122e-01 -5.67127526e-01
2.55135536e-01 -7.11450577e-01 -1.22379206e-01 2.32669950e-01
-2.93555200e-01 -4.72380877e-01 9.29574430e-01 -4.23458874e-01
1.29567063e+00 -2.17417955e+00 1.74109757e-01 4.79592919e-01
2.06562325e-01 1.26143262e-01 -3.86203617e-01 1.49717957e-01
-1.92711383e-01 -1.14273094e-01 -1.07056022e-01 -6.02291822e-01
-1.23955078e-01 -9.01310965e-02 -2.63214968e-02 6.71475887e-01
2.91948259e-01 8.71914506e-01 -7.80934215e-01 -7.37702847e-01
1.88304856e-01 7.14778841e-01 -4.48440909e-01 1.44840181e-02
1.52263626e-01 2.83376843e-01 -5.59441686e-01 7.87760258e-01
1.00190043e+00 -7.52038732e-02 -1.54394954e-01 -4.39152449e-01
-2.05118507e-01 -2.69723535e-01 -1.11126161e+00 1.87854099e+00
-5.11138797e-01 2.54579455e-01 1.19263068e-01 -7.82673120e-01
1.23689282e+00 8.01918060e-02 6.63386464e-01 -6.74607456e-01
-4.89931405e-02 2.71341857e-02 -4.42991227e-01 -1.15805246e-01
1.92796752e-01 2.22923204e-01 3.41340214e-01 2.63845742e-01
2.26513684e-01 2.62236476e-01 -3.78252491e-02 -2.71653116e-01
6.98613942e-01 5.18550098e-01 4.82090384e-01 -5.54547548e-01
8.65055203e-01 -2.91989446e-01 9.24734771e-01 1.76901668e-01
-2.12343678e-01 8.18739593e-01 2.35636771e-01 -8.96286905e-01
-8.15114379e-01 -9.15514052e-01 -5.17922938e-01 1.13682246e+00
2.07333013e-01 -5.84749579e-01 -9.97321248e-01 -7.73623466e-01
-2.98871403e-03 -9.57404226e-02 -1.10875165e+00 1.24933153e-01
-6.75888777e-01 -6.96060240e-01 1.21322483e-01 5.33230186e-01
8.65990222e-01 -9.92794514e-01 -3.33957970e-01 4.11241017e-02
6.22497723e-02 -7.33666301e-01 -8.70214760e-01 -3.38551521e-01
-5.09292841e-01 -9.50884879e-01 -7.10507274e-01 -1.11538839e+00
1.14501607e+00 1.99298427e-01 8.43751431e-01 2.84858614e-01
-4.72943246e-01 1.83820218e-01 -1.10202767e-01 -2.05854364e-02
3.45880449e-01 1.61531165e-01 -1.98879614e-01 6.05838120e-01
5.34759462e-01 -6.88858926e-01 -7.58465171e-01 5.86000144e-01
-3.54238093e-01 5.29456995e-02 5.75395823e-01 1.04147601e+00
1.00010180e+00 -8.87568146e-02 4.42292750e-01 -8.63693893e-01
1.46207064e-01 -4.26420450e-01 -6.94420874e-01 4.25724715e-01
-4.89340395e-01 -1.24688791e-02 5.95582187e-01 -2.32710391e-01
-1.10802603e+00 4.41578060e-01 -3.61754261e-02 -4.46381241e-01
-1.33106530e-01 1.88017234e-01 -3.19944263e-01 -4.81353015e-01
2.78064102e-01 2.13123128e-01 -1.37166545e-01 -4.96004820e-01
4.62273151e-01 2.44953290e-01 5.22664547e-01 -7.68329501e-01
1.04502642e+00 5.34985542e-01 1.59418032e-01 -5.64480960e-01
-7.83801913e-01 -3.33158255e-01 -1.00467205e+00 -1.52828395e-01
7.20354319e-01 -8.69388163e-01 -4.50242281e-01 4.29904759e-01
-9.54477131e-01 -1.90779537e-01 -2.48208895e-01 3.23653631e-02
-4.17773783e-01 8.84820297e-02 -3.79683495e-01 -3.14891487e-01
-3.40608388e-01 -1.24068570e+00 1.35288572e+00 7.01691628e-01
5.00409268e-02 -9.66865361e-01 2.34654441e-01 1.52993351e-02
4.81871217e-01 3.01402539e-01 7.58622825e-01 -3.36434059e-02
-6.02849901e-01 -1.90260261e-01 -3.79516125e-01 8.70047137e-02
4.54304129e-01 3.17534775e-01 -1.02386546e+00 -2.48932168e-01
-3.19554001e-01 -1.48232102e-01 7.66726136e-01 4.01619464e-01
1.25087273e+00 -2.79466987e-01 -5.21498203e-01 1.01401234e+00
1.50982416e+00 -4.87091765e-02 7.19348073e-01 1.52912155e-01
9.02585506e-01 5.27146161e-01 5.18802345e-01 2.88364023e-01
4.64476526e-01 9.76632953e-01 1.99783996e-01 -3.38213027e-01
-5.11753976e-01 -4.85649467e-01 2.20694304e-01 6.46474183e-01
-4.85439807e-01 5.62242508e-01 -6.91735804e-01 3.67238492e-01
-1.77120423e+00 -8.86097789e-01 2.74237841e-01 2.08082032e+00
8.68338943e-01 -3.49083334e-01 -9.74687561e-02 -2.59240389e-01
7.89006054e-01 4.15225953e-01 -2.57918745e-01 -1.65379196e-01
1.31513834e-01 5.55544913e-01 1.20284781e-01 7.54205585e-01
-1.17867649e+00 1.30422306e+00 6.26446438e+00 1.14032745e+00
-1.24418044e+00 1.51880339e-01 1.05558252e+00 2.91151613e-01
-4.81613912e-02 -7.90832788e-02 -1.10836327e+00 5.69842637e-01
3.21143657e-01 -9.99396518e-02 4.15326297e-01 1.00760698e+00
-1.75765148e-04 2.70611774e-02 -8.39483619e-01 9.69735563e-01
2.75183856e-01 -1.13134646e+00 1.03330435e-02 -3.93466614e-02
1.11175680e+00 -6.05368391e-02 2.29984701e-01 6.27826154e-02
1.72718614e-01 -1.19290078e+00 4.22079146e-01 7.96142757e-01
9.34985876e-01 -8.58479083e-01 5.69011807e-01 -1.47950485e-01
-1.79958487e+00 7.86456913e-02 -7.73398817e-01 1.51642859e-01
-3.15989077e-01 4.59766567e-01 -6.75723374e-01 5.75595498e-01
9.05608475e-01 9.72562969e-01 -1.03919601e+00 9.71780717e-01
-3.18269134e-01 4.32892799e-01 -4.52675849e-01 3.72441500e-01
1.53237566e-01 -5.27476788e-01 7.91927986e-03 1.00404191e+00
4.90624934e-01 -1.12273045e-01 3.26487362e-01 8.10411811e-01
-1.33486778e-01 4.03848588e-01 -2.86359906e-01 7.26555049e-01
6.39928102e-01 1.71198094e+00 -7.70477355e-01 -3.70895192e-02
-5.13499618e-01 8.91566217e-01 6.95836842e-01 3.64153773e-01
-7.39512384e-01 -3.64120871e-01 5.30269980e-01 2.16074497e-01
3.91773373e-01 1.09649543e-02 -2.51225501e-01 -7.69640267e-01
1.64181143e-01 -5.71098745e-01 3.05434734e-01 -5.76365232e-01
-1.15945661e+00 8.77679706e-01 -1.26448229e-01 -1.16075647e+00
2.45933160e-02 -3.33770186e-01 -8.77204239e-01 9.54257250e-01
-1.71046102e+00 -1.56682873e+00 -6.57426119e-01 8.18568110e-01
4.96550024e-01 -1.36408985e-01 8.09918523e-01 1.63462207e-01
-8.05222332e-01 9.17518914e-01 -7.59283975e-02 2.87040204e-01
8.75900686e-01 -1.08513308e+00 2.69613057e-01 7.21965313e-01
3.53786558e-01 6.61635160e-01 6.33885115e-02 -5.60783982e-01
-9.25027072e-01 -1.24524927e+00 9.76140022e-01 -3.28638941e-01
3.59738350e-01 -3.33849251e-01 -7.54127443e-01 7.37506270e-01
3.01870167e-01 7.52519846e-01 4.78926033e-01 1.47962540e-01
-5.28801799e-01 -8.52261305e-01 -1.24398184e+00 4.08819020e-01
1.05720997e+00 -4.54023480e-01 -3.42842132e-01 2.39589497e-01
1.61095202e-01 -3.70538831e-01 -8.35705459e-01 3.68565738e-01
6.46621466e-01 -9.70960498e-01 9.49072063e-01 -1.69732854e-01
3.98181885e-01 -6.93194747e-01 -3.88684012e-02 -1.36701417e+00
-7.24943936e-01 -5.57161629e-01 1.41504973e-01 1.52540839e+00
2.40281284e-01 -5.36530197e-01 9.07337070e-01 4.57085490e-01
-1.22253396e-01 -1.14957404e+00 -1.31445539e+00 -4.39767540e-01
-1.40191779e-01 1.09007947e-01 9.65448737e-01 8.90142739e-01
-9.34231430e-02 9.81193408e-02 -3.49347293e-01 2.61080176e-01
6.49888813e-01 2.53833264e-01 7.65792429e-01 -1.21490860e+00
1.27949849e-01 -5.06665647e-01 -8.43402982e-01 -1.01805317e+00
3.19158256e-01 -8.38654041e-01 4.98581119e-02 -1.19699478e+00
2.08845705e-01 -6.77940011e-01 -4.88204002e-01 9.14366722e-01
-4.42394525e-01 8.48979712e-01 1.32720536e-02 1.93591461e-01
-5.75043619e-01 6.56419694e-01 1.27572763e+00 -1.51656821e-01
-3.39072943e-02 -1.52449667e-01 -6.00351155e-01 7.74243116e-01
4.98671800e-01 -2.13623837e-01 -2.67460138e-01 -4.46532995e-01
-3.45299035e-01 -5.17072678e-01 1.22501314e-01 -1.12867117e+00
4.86965835e-01 -2.50971317e-01 7.57948101e-01 -3.49894851e-01
4.29174781e-01 -8.62162232e-01 1.54103130e-01 5.36794700e-02
-8.41608122e-02 1.52947426e-01 6.17988668e-02 4.01218832e-01
-4.16537046e-01 1.48989499e-01 9.73554671e-01 1.62921965e-01
-7.88369834e-01 7.89233327e-01 4.79497403e-01 -2.38998249e-01
1.23987365e+00 -2.98441470e-01 -1.43033281e-01 -9.66437906e-02
-5.73487759e-01 1.78150460e-01 5.01874506e-01 3.56109113e-01
5.86895227e-01 -1.73957944e+00 -7.70516515e-01 7.10080087e-01
3.55348438e-02 -7.85294455e-03 2.08069712e-01 7.96647549e-01
-4.87720311e-01 2.76505411e-01 -5.24601579e-01 -6.27513766e-01
-1.18042397e+00 3.85283679e-01 4.92377311e-01 -1.07049696e-01
-6.00730658e-01 1.04516172e+00 4.81813550e-01 -2.30655506e-01
1.85184613e-01 -2.63389826e-01 -3.22477639e-01 7.06135929e-02
6.62980735e-01 2.91522831e-01 3.94368591e-03 -1.14090848e+00
-4.42468047e-01 1.49448907e+00 -2.43888661e-01 3.93979371e-01
1.36963284e+00 -1.09306946e-01 -4.41590935e-01 2.49511991e-02
1.32668614e+00 3.05748492e-01 -1.61605966e+00 -3.25934500e-01
-3.06475665e-02 -8.15787911e-01 7.26133585e-02 -5.05183935e-01
-1.65640271e+00 6.45588875e-01 6.67749584e-01 -3.60571086e-01
1.28540003e+00 -6.47052750e-02 4.70756561e-01 -1.21208400e-01
6.24849558e-01 -8.98207843e-01 -7.27136880e-02 1.33116603e-01
9.64895785e-01 -1.10820997e+00 2.01089337e-01 -6.45941436e-01
-2.77488858e-01 1.16442847e+00 8.16185117e-01 -5.24686694e-01
9.59076166e-01 1.83443531e-01 1.34651721e-01 -4.09425385e-02
-3.62710267e-01 -8.82586651e-03 4.64810789e-01 6.31356359e-01
3.92125428e-01 -3.85484733e-02 -1.86565682e-01 4.09224093e-01
-1.45534858e-01 -5.12109585e-02 -1.40834078e-01 4.66198236e-01
-3.84990603e-01 -1.30907381e+00 -3.85816395e-01 2.85456806e-01
-2.49229625e-01 -8.40547085e-02 -1.56764790e-01 6.76579595e-01
5.28927386e-01 6.02271378e-01 2.66352475e-01 -3.79439056e-01
1.28874406e-01 -1.32747563e-02 4.95820701e-01 -5.25823176e-01
-3.41867387e-01 2.31012255e-01 -5.60676992e-01 -1.01478624e+00
-5.13583779e-01 -6.51718795e-01 -1.07000947e+00 -1.65564954e-01
-8.61819759e-02 1.88658044e-01 3.36076766e-01 7.08581924e-01
6.32222176e-01 2.12307259e-01 9.74904418e-01 -1.15745044e+00
-1.16634458e-01 -1.01072526e+00 -5.83849430e-01 1.70428157e-01
2.85097539e-01 -8.80283356e-01 -1.59317315e-01 8.58787149e-02] | [13.498050689697266, 0.434471994638443] |
75b69114-d1f5-47b6-b7ab-fc6cc61b425d | spike-and-slab-generalized-additive-models | 2110.14449 | null | https://arxiv.org/abs/2110.14449v3 | https://arxiv.org/pdf/2110.14449v3.pdf | Spike-and-Slab LASSO Generalized Additive Models and Scalable Algorithms for High-Dimensional Data Analysis | There are proposals that extend the classical generalized additive models (GAMs) to accommodate high-dimensional data ($p>>n$) using group sparse regularization. However, the sparse regularization may induce excess shrinkage when estimating smooth functions, damaging predictive performance. Moreover, most of these GAMs consider an "all-in-all-out" approach for functional selection, rendering them difficult to answer if nonlinear effects are necessary. While some Bayesian models can address these shortcomings, using Markov chain Monte Carlo algorithms for model fitting creates a new challenge, scalability. Hence, we propose Bayesian hierarchical generalized additive models as a solution: we consider the smoothing penalty for proper shrinkage of curve interpolation via reparameterization. A novel two-part spike-and-slab LASSO prior for smooth functions is developed to address the sparsity of signals while providing extra flexibility to select the linear or nonlinear components of smooth functions. A scalable and deterministic algorithm, EM-Coordinate Descent, is implemented in an open-source R package BHAM. Simulation studies and metabolomics data analyses demonstrate improved predictive and computational performance against state-of-the-art models. Functional selection performance suggests trade-offs exist regarding the effect hierarchy assumption. | ['Nengjun Yi', 'D. Leann Long', 'A. K. M. Fazlur Rahman', 'Byron C. Jaeger', 'Boyi Guo'] | 2021-10-27 | null | null | null | null | ['additive-models'] | ['methodology'] | [ 8.06249753e-02 -2.31157079e-01 -2.01242894e-01 -7.17918038e-01
-8.04335475e-01 -3.33782792e-01 4.45362031e-02 1.01844184e-01
-1.17470987e-01 7.57110238e-01 2.14638516e-01 -1.99224770e-01
-3.45460862e-01 -6.08010769e-01 -7.15649128e-01 -9.46585834e-01
-1.12872377e-01 2.00553730e-01 -3.28895748e-01 1.46017075e-01
3.67680818e-01 2.33416677e-01 -1.37808156e+00 2.12444216e-01
1.29028547e+00 6.70327663e-01 2.91920573e-01 1.12124987e-01
-1.20796792e-01 3.46426457e-01 4.60890308e-02 -1.71981052e-01
4.02044237e-01 -4.60621804e-01 5.31291738e-02 -7.70311281e-02
1.72361180e-01 -2.12499693e-01 3.01146477e-01 9.57167089e-01
8.66276026e-01 2.06645101e-01 8.09908628e-01 -1.00966382e+00
-7.46408522e-01 6.05404437e-01 -8.85395408e-01 -2.23472282e-01
1.46322951e-01 4.67815101e-01 7.53551841e-01 -1.10398340e+00
3.04691404e-01 1.33477545e+00 1.01121759e+00 1.83950961e-01
-1.99095619e+00 -8.36550474e-01 2.95462549e-01 -2.72094935e-01
-1.71163571e+00 -5.66566885e-01 5.94822407e-01 -8.00166488e-01
7.85060704e-01 4.56225127e-01 5.47412455e-01 7.50873268e-01
2.98785388e-01 4.08758193e-01 1.25757861e+00 -8.27121288e-02
6.04814768e-01 1.00006960e-01 1.47778079e-01 5.11341691e-01
3.64529490e-01 -2.68087108e-02 -7.23915875e-01 -8.50539029e-01
9.70990121e-01 1.64120011e-02 -7.02187493e-02 -4.83850300e-01
-8.47972155e-01 1.09434974e+00 -5.70291653e-02 -6.12962171e-02
-6.49637759e-01 1.91085741e-01 2.62377024e-01 4.81136106e-02
7.42520094e-01 4.32281017e-01 -5.11650681e-01 1.28210649e-01
-1.22483003e+00 3.86752129e-01 4.23728168e-01 7.63621688e-01
6.20763004e-01 3.48351389e-01 -2.44312450e-01 1.09514117e+00
7.56238580e-01 2.65469134e-01 9.30984989e-02 -1.15082061e+00
-5.05592003e-02 5.38440228e-01 1.20780610e-01 -1.01809168e+00
-6.60194993e-01 -4.89127010e-01 -9.73969758e-01 1.23345606e-01
4.13125426e-01 -1.15822628e-01 -7.25867331e-01 1.83588743e+00
4.74120438e-01 3.07340175e-01 -5.63172400e-01 9.99158502e-01
6.47733569e-01 4.81880844e-01 6.80462778e-01 -7.50639379e-01
1.11313856e+00 -3.87652189e-01 -8.71460795e-01 -1.06249917e-02
6.10036671e-01 -6.75323844e-01 1.04486859e+00 6.01430237e-01
-1.39081073e+00 -2.73512542e-01 -6.28731012e-01 2.90776715e-02
1.53131792e-02 2.37302363e-01 1.02478075e+00 8.65232885e-01
-8.34471285e-01 6.27892554e-01 -1.05934834e+00 5.04872240e-02
3.08232099e-01 5.40081680e-01 1.09256901e-01 -4.12547439e-02
-8.21415544e-01 5.83381772e-01 -9.89838466e-02 1.42029330e-01
-8.67958844e-01 -1.24344647e+00 -7.70198941e-01 -3.62688564e-02
-3.16726938e-02 -1.07756710e+00 4.39355016e-01 -8.07276905e-01
-1.63564312e+00 6.88834190e-01 -2.86339104e-01 -1.87197760e-01
5.02272487e-01 -8.61724019e-02 1.38419205e-02 -1.54588535e-01
3.45039926e-02 5.22136748e-01 7.80259967e-01 -9.20445263e-01
1.58050001e-01 -5.23247778e-01 -5.67647636e-01 1.92924201e-01
-1.19059635e-02 3.20428610e-01 -1.27656356e-01 -8.66523027e-01
5.80146492e-01 -8.56058061e-01 -7.58823812e-01 1.22399367e-01
-2.13739574e-01 1.49389625e-01 9.49423611e-02 -8.17676961e-01
1.47983789e+00 -2.24820113e+00 1.21993497e-01 2.79414684e-01
1.98597517e-02 -4.05007958e-01 7.27421939e-02 1.89539000e-01
-2.10962489e-01 2.29874387e-01 -5.79424441e-01 -3.50295573e-01
8.65432024e-02 -1.50105089e-01 2.17283703e-02 9.37139213e-01
-6.28745034e-02 3.09124947e-01 -5.21581829e-01 -2.65287578e-01
1.70516685e-01 6.44467056e-01 -1.05184042e+00 2.45423131e-02
-5.93630075e-02 8.20766628e-01 -3.50354552e-01 6.36546850e-01
1.10884094e+00 -3.14984441e-01 3.40561926e-01 -1.37846917e-01
-4.61100370e-01 5.42276055e-02 -1.64911246e+00 1.63176835e+00
9.85358506e-02 -1.50278553e-01 5.62906206e-01 -1.14035630e+00
1.08846605e+00 2.41241843e-01 9.17910099e-01 -1.69142187e-01
-1.19098641e-01 4.80124414e-01 1.01269938e-01 -4.50230688e-01
-6.09925948e-03 -3.15582633e-01 3.16613913e-01 1.16041958e-01
-1.84131339e-01 -2.64324039e-01 -6.37952238e-02 -1.64482325e-01
7.47206748e-01 6.54159069e-01 2.90566117e-01 -9.14200783e-01
1.76177830e-01 -1.88166410e-01 1.09384644e+00 6.11348927e-01
1.97020583e-02 6.70073807e-01 3.83205712e-01 -2.07271636e-01
-7.91623831e-01 -9.58858669e-01 -6.01809144e-01 1.31804359e+00
-3.20477307e-01 -3.51031661e-01 -5.48286259e-01 1.71827361e-01
2.37437889e-01 9.31186974e-01 -3.74893427e-01 4.37921509e-02
-3.53874266e-01 -1.68603086e+00 2.07940876e-01 2.52669156e-01
-6.89477026e-02 -4.22382861e-01 -4.88180339e-01 5.62037170e-01
5.29944412e-02 -4.93113309e-01 -3.09141815e-01 3.33525360e-01
-1.10289991e+00 -7.51957655e-01 -7.33082175e-01 -3.30223858e-01
7.43811488e-01 3.35561857e-02 8.51541817e-01 -2.96328902e-01
-3.86650652e-01 2.50657052e-01 -9.76875424e-02 -2.79851437e-01
1.15089357e-01 -5.52895367e-01 2.93793917e-01 7.36621171e-02
3.03828984e-01 -8.14195931e-01 -9.33192968e-01 4.54663038e-01
-4.78559196e-01 -8.94803554e-03 3.31752807e-01 7.96910107e-01
1.01410186e+00 -2.75888562e-01 8.30905557e-01 -6.79443717e-01
5.85222304e-01 -7.39614367e-01 -5.16260207e-01 -2.54230611e-02
-7.56952047e-01 -5.59603646e-02 2.99178392e-01 -5.07967234e-01
-1.09293509e+00 2.08529353e-01 -9.84427109e-02 -4.83910859e-01
1.45006955e-01 9.34528589e-01 -1.42766640e-01 -2.56081577e-02
7.81106591e-01 5.70801906e-02 2.43399799e-01 -6.96635604e-01
2.82152683e-01 2.26899609e-01 -2.06400886e-01 -8.34114850e-01
-1.48806376e-02 4.16788071e-01 7.27332234e-02 -1.01874840e+00
-3.93414736e-01 -4.09452409e-01 -2.95428872e-01 9.05931517e-02
8.99403572e-01 -1.08128893e+00 -6.78470790e-01 2.52357483e-01
-6.71750963e-01 -2.77020544e-01 -1.67711049e-01 8.14915419e-01
-8.58110011e-01 4.37734872e-01 -5.56969881e-01 -1.15529966e+00
-4.02515054e-01 -1.55766761e+00 8.22275877e-01 -2.12661296e-01
-3.87692213e-01 -6.67355895e-01 6.43957332e-02 1.88421801e-01
6.35374188e-01 3.92734349e-01 9.22545135e-01 -4.99573886e-01
-4.43564653e-01 4.35272828e-02 3.14179584e-02 7.34968036e-02
-1.12955801e-01 2.51294285e-01 -7.09149837e-01 -1.55727193e-01
3.75201315e-01 -4.76464294e-02 6.70282662e-01 1.46450675e+00
1.15722167e+00 -3.50564629e-01 -3.02296430e-01 9.75330293e-01
1.24688351e+00 1.72872499e-01 6.64577782e-01 -1.40738159e-01
4.74366754e-01 6.29862845e-01 3.73832881e-01 9.13243771e-01
2.12070495e-01 5.30499279e-01 1.93244934e-01 -3.99572589e-02
3.83090794e-01 1.17111638e-01 4.33244556e-01 7.83955336e-01
-1.45574674e-01 2.90643543e-01 -7.89344847e-01 2.69506574e-01
-1.81566250e+00 -8.12255979e-01 -7.28748918e-01 2.48670483e+00
1.01446795e+00 -2.69746959e-01 1.58557400e-01 -3.44659925e-01
7.73239613e-01 -1.87605023e-01 -6.28829122e-01 -1.98849604e-01
-1.76603362e-01 4.80553284e-02 5.34858406e-01 4.64112818e-01
-8.45691800e-01 4.97607261e-01 7.05912924e+00 8.36277127e-01
-9.47602689e-01 1.91535711e-01 9.18102026e-01 -2.88565546e-01
-4.37840700e-01 2.70944357e-01 -8.46879244e-01 6.33663356e-01
1.02282619e+00 -6.62399922e-03 4.68726605e-01 7.47757435e-01
1.04059875e+00 -2.69194931e-01 -9.21900570e-01 9.76093113e-01
-2.71552593e-01 -1.09894359e+00 -3.48809600e-01 1.18968032e-01
5.71443319e-01 -2.49930192e-02 7.63508081e-02 1.13339409e-01
2.71910995e-01 -1.06499326e+00 5.57509482e-01 8.43472183e-01
7.39705861e-01 -4.47418749e-01 1.81889176e-01 3.65219116e-01
-1.08201718e+00 7.11769089e-02 -5.96555591e-01 -9.23022255e-02
3.74897897e-01 1.15037000e+00 -3.98738831e-01 2.60996372e-01
6.67513549e-01 4.57305193e-01 -3.05355102e-01 1.30788374e+00
1.95640981e-01 8.12546134e-01 -6.30464494e-01 3.66194367e-01
-2.48961523e-01 -8.44390750e-01 6.79131389e-01 1.20005512e+00
6.90623045e-01 2.99154371e-01 3.48871469e-01 1.43524599e+00
5.50956130e-01 6.17113173e-01 -4.19689894e-01 -4.67686839e-02
4.52958822e-01 9.47246075e-01 -7.52359927e-01 -2.27606058e-01
-4.77764100e-01 3.06776971e-01 -2.90023126e-02 5.18221021e-01
-7.43811548e-01 1.93866983e-01 4.81309056e-01 3.50886077e-01
2.00683959e-02 -2.98901141e-01 -8.79201412e-01 -1.05201697e+00
-3.39722633e-01 -9.26231086e-01 5.03269553e-01 -5.86399794e-01
-1.55860531e+00 -1.99976161e-01 8.67914036e-02 -1.06538832e+00
1.19346976e-01 -2.93215781e-01 -2.98127770e-01 1.11318767e+00
-8.52265537e-01 -1.08081484e+00 8.09345469e-02 3.78750443e-01
5.22875905e-01 1.53344274e-01 8.06003690e-01 3.73959243e-01
-5.65190971e-01 3.57806176e-01 3.72530460e-01 -6.25525236e-01
9.23957467e-01 -8.63313258e-01 -2.24328071e-01 5.14415920e-01
-6.42145157e-01 1.21187091e+00 1.01445961e+00 -1.04148412e+00
-1.57153511e+00 -8.56753469e-01 4.07374054e-01 -4.89667542e-02
5.02063215e-01 -2.33177602e-01 -1.13160348e+00 6.26221597e-01
-4.90655042e-02 -1.65259376e-01 1.28065908e+00 3.64556819e-01
-1.51767701e-01 1.57418013e-01 -1.31952202e+00 4.16475981e-01
8.18209469e-01 2.05736198e-02 -1.05151691e-01 3.64906937e-01
4.18852359e-01 -1.47935256e-01 -1.22479069e+00 5.88476539e-01
6.63141549e-01 -8.11222315e-01 1.16406024e+00 -3.61594945e-01
1.41979650e-01 -4.38344270e-01 -4.09613311e-01 -1.15075493e+00
-5.27653515e-01 -8.14550579e-01 -7.13303536e-02 1.14065421e+00
5.07708669e-01 -6.76388681e-01 5.43021262e-01 1.27126372e+00
-3.55825931e-01 -6.88493788e-01 -9.12683904e-01 -5.76786757e-01
1.05987974e-01 -5.15837371e-01 3.20135713e-01 1.30994773e+00
2.52466977e-01 2.88940780e-02 -5.08331537e-01 -1.14517519e-02
8.85937393e-01 -1.46622017e-01 5.21731079e-01 -1.27374315e+00
-5.45076966e-01 -4.29628104e-01 2.45104823e-02 -8.60762775e-01
-1.65612116e-01 -9.03728902e-01 -6.79093003e-02 -1.29101813e+00
2.75785387e-01 -6.13021255e-01 -1.90031797e-01 2.70854980e-01
-2.85941094e-01 -1.17066637e-01 -5.06069362e-02 2.74368256e-01
-1.21384248e-01 6.29647851e-01 1.00774896e+00 9.48710516e-02
-7.07710683e-01 -3.19574475e-02 -8.32499087e-01 9.47239578e-01
5.66260219e-01 -5.20706296e-01 -3.33895177e-01 -2.11625189e-01
4.89927918e-01 3.78760457e-01 1.81472272e-01 -5.18751502e-01
8.72507095e-02 -5.68382680e-01 2.63874948e-01 -5.90357065e-01
4.07410651e-01 -6.41681075e-01 8.38029921e-01 1.88405737e-01
-3.72409284e-01 -1.52353764e-01 2.33751938e-01 5.19947767e-01
2.38077834e-01 -4.03581858e-01 9.95675385e-01 -2.41169453e-01
5.89357540e-02 1.66306436e-01 -5.13875902e-01 -2.39865378e-01
7.18850613e-01 -1.77187935e-01 4.88869958e-02 -2.60262638e-01
-1.19580734e+00 1.33722052e-01 4.03030843e-01 -2.72380501e-01
4.26162779e-01 -1.14040220e+00 -9.25540626e-01 7.76825696e-02
-3.65883768e-01 -1.80903450e-02 6.95904970e-01 1.49574316e+00
-3.60835791e-01 6.64886683e-02 1.23464577e-02 -6.81606472e-01
-9.26638365e-01 5.64715743e-01 2.27023378e-01 -5.68674356e-02
-4.51510459e-01 8.75042975e-01 5.43544888e-01 -3.47230464e-01
4.57719862e-02 -2.05472112e-01 4.51618023e-02 2.27073476e-01
3.20535153e-01 6.88240230e-01 -1.67449385e-01 -2.76577830e-01
-4.01713282e-01 4.13494557e-01 3.87823313e-01 3.86007465e-02
1.66516614e+00 -3.45025331e-01 -4.74778533e-01 5.48277795e-01
8.16841424e-01 8.46470892e-02 -1.34130180e+00 2.90408164e-01
-1.65310919e-01 -4.36462551e-01 1.82120830e-01 -6.30371809e-01
-7.39678621e-01 6.29077852e-01 7.33043134e-01 -5.21214828e-02
9.16559398e-01 -4.03187424e-01 -1.67959165e-02 -7.46382102e-02
2.34588429e-01 -1.26579571e+00 -4.13548529e-01 1.18259430e-01
1.02066040e+00 -9.88325238e-01 5.31133473e-01 -6.70079231e-01
-6.86572850e-01 6.61490977e-01 2.94393361e-01 -2.64033049e-01
8.85537505e-01 3.04075629e-01 -3.62337202e-01 -3.68643850e-01
-6.01278484e-01 2.07326397e-01 1.98513538e-01 4.29293245e-01
1.01046062e+00 3.68106291e-02 -1.07566321e+00 9.71321225e-01
1.87133998e-01 1.75786167e-01 1.90310165e-01 7.94558585e-01
-4.71328288e-01 -8.30611765e-01 -6.76941097e-01 7.93018520e-01
-7.14217246e-01 -4.54440862e-01 1.99473098e-01 3.63012820e-01
-1.01424279e-02 9.59962606e-01 -1.01305142e-01 1.40411288e-01
3.67661715e-02 3.83507073e-01 1.76723793e-01 -6.20803595e-01
-6.11805618e-01 1.06230545e+00 7.20053464e-02 -5.01677871e-01
-4.57512945e-01 -1.12826753e+00 -1.09768748e+00 -7.90350139e-02
-5.02435863e-01 -9.87913553e-03 8.38304758e-01 6.89333022e-01
6.88839853e-01 3.46083999e-01 3.12644452e-01 -8.94442081e-01
-5.30345559e-01 -9.29213583e-01 -8.35262597e-01 1.63423523e-01
-1.34425744e-01 -8.83952975e-01 -5.32848954e-01 2.81569541e-01] | [7.417416572570801, 4.61798620223999] |
8e678dcb-0f8e-4b77-9639-5c5b9fe051e1 | autoregressive-latent-video-prediction-with | null | null | https://openreview.net/forum?id=K-hiHQXEQog | https://openreview.net/pdf?id=K-hiHQXEQog | Autoregressive Latent Video Prediction with High-Fidelity Image Generator | Video prediction is an important yet challenging problem; burdened with the tasks of generating future frames and learning environment dynamics. Recently, autoregressive latent video models have proved to be a powerful video prediction tool, by separating the video prediction into two sub-problems: pre-training an image generator model, followed by learning an autoregressive prediction model in the latent space of the image generator. However, successfully generating high-fidelity and high-resolution videos has yet to be seen. In this work, we investigate how to train an autoregressive latent video prediction model capable of predicting high-fidelity future frames with minimal modification to existing models, and produce high-resolution (256x256) videos. Specifically, we scale up prior models by employing a high-fidelity image generator (VQ-GAN) with a causal transformer model, and introduce additional techniques of top-$k$ sampling and data augmentation to further improve video prediction quality. Despite the simplicity, the proposed method achieves competitive performance to state-of-the-art approaches on standard video prediction benchmarks with fewer parameters, and enables high-resolution video prediction on complex and large-scale datasets. Videos are available at the anonymized website https://sites.google.com/view/harp-anonymous | ['Pieter Abbeel', 'Stephen James', 'Fangchen Liu', 'Kimin Lee', 'Younggyo Seo'] | 2021-09-29 | null | null | null | null | ['video-prediction'] | ['computer-vision'] | [ 2.65500337e-01 5.28801791e-02 -2.86285639e-01 -2.83386946e-01
-9.29962754e-01 -1.41205817e-01 6.46896958e-01 -7.03431487e-01
1.31623492e-01 7.87514925e-01 5.96813500e-01 -6.83595240e-02
3.22309047e-01 -6.92040205e-01 -1.13903618e+00 -7.66353011e-01
-1.14095159e-01 2.32835367e-01 1.62652448e-01 8.92226398e-02
-1.11079715e-01 2.19285145e-01 -1.60611224e+00 6.19973600e-01
6.68600976e-01 1.01297474e+00 2.67883033e-01 9.73200381e-01
4.62593675e-01 1.49669886e+00 -1.26440227e-01 -5.72574794e-01
5.25437593e-01 -6.04538918e-01 -5.60046911e-01 1.63947776e-01
3.62027913e-01 -9.33100402e-01 -7.93155253e-01 7.82472372e-01
2.48162493e-01 9.35726166e-02 5.03791034e-01 -1.45245004e+00
-8.30961347e-01 5.49466968e-01 -2.76692241e-01 1.31955137e-02
4.27532881e-01 2.55727351e-01 8.70714843e-01 -8.94065320e-01
6.89851880e-01 1.02881396e+00 5.24830163e-01 8.28552425e-01
-1.33261228e+00 -9.63112056e-01 2.25854036e-03 4.66686577e-01
-1.42649043e+00 -8.77119422e-01 7.83764005e-01 -6.81925714e-01
8.42191041e-01 1.71273142e-01 6.85024858e-01 1.70535088e+00
1.37251675e-01 7.39568174e-01 6.66455030e-01 -1.14837408e-01
7.95866102e-02 4.26469781e-02 -7.37158418e-01 6.08317077e-01
-2.88547426e-01 2.39893571e-01 -7.64980495e-01 -2.87498176e-01
1.04609716e+00 1.87136874e-01 -4.81182784e-01 -4.99679625e-01
-1.19388354e+00 8.70678425e-01 8.49705189e-02 -1.09269001e-01
-7.41504073e-01 5.35070121e-01 2.59784609e-01 1.51010320e-01
6.27256930e-01 1.09966099e-01 -6.46032467e-02 -4.90246922e-01
-1.28212237e+00 4.01277095e-01 3.97695124e-01 1.23761630e+00
5.20605564e-01 4.96196330e-01 -1.66576684e-01 5.75236917e-01
3.70386899e-01 3.17484826e-01 5.99704564e-01 -1.38895130e+00
6.66960418e-01 -5.44881560e-02 3.05233270e-01 -1.01396286e+00
3.68899912e-01 -3.49658765e-02 -1.07943010e+00 -2.48959307e-02
1.69142280e-02 -1.24807648e-01 -6.97581589e-01 1.64993334e+00
1.10963061e-01 9.19724524e-01 4.14133854e-02 8.66675377e-01
4.28233534e-01 1.24326384e+00 2.73763333e-02 -5.77656746e-01
8.29707682e-01 -1.19960535e+00 -5.75651884e-01 1.43360853e-01
3.40805560e-01 -6.52954638e-01 9.03981507e-01 3.68506312e-01
-1.33867574e+00 -8.85133028e-01 -7.36534059e-01 -9.83508378e-02
2.04020843e-01 1.58806443e-01 3.42649341e-01 3.73982698e-01
-1.27340460e+00 5.37812352e-01 -1.06280184e+00 5.11102304e-02
6.46291554e-01 2.65769958e-01 -5.06871104e-01 -2.98206836e-01
-1.18028522e+00 2.94803917e-01 2.90219039e-01 -1.12787835e-01
-1.60359275e+00 -7.85486102e-01 -8.85868192e-01 4.17547561e-02
3.67350847e-01 -9.05962586e-01 9.79536057e-01 -1.21520531e+00
-1.79173481e+00 3.46892238e-01 -3.68504316e-01 -8.75913382e-01
8.30013633e-01 -4.48630571e-01 -3.06568086e-01 2.52817988e-01
-1.80486124e-02 8.64294350e-01 1.41205966e+00 -1.22850764e+00
-6.31295383e-01 8.03221092e-02 -9.28332955e-02 1.01159617e-01
-2.95179218e-01 -2.77598333e-02 -5.56702614e-01 -9.62821901e-01
-4.35169071e-01 -9.56208706e-01 -2.93390781e-01 -1.08665451e-01
-2.05232650e-01 1.93379581e-01 9.74143326e-01 -1.09098136e+00
1.18797207e+00 -2.17936301e+00 3.02810311e-01 -1.19212903e-01
2.66009659e-01 2.20615268e-01 -1.62938863e-01 1.70470804e-01
-2.28531033e-01 1.62360355e-01 1.31153986e-01 -5.53470016e-01
-7.03260377e-02 4.58024852e-02 -8.81344199e-01 2.61303455e-01
1.35978207e-01 9.44425583e-01 -8.53023231e-01 -4.51843500e-01
2.98605621e-01 8.28570843e-01 -9.22693431e-01 6.47910953e-01
-3.31519783e-01 7.22197711e-01 -2.71080256e-01 4.15925413e-01
2.73579597e-01 -4.01322961e-01 1.84795961e-01 -1.52140766e-01
5.35464250e-02 -1.56235825e-02 -9.66935754e-01 1.46547520e+00
-2.53984004e-01 6.64860070e-01 -3.19299728e-01 -9.98197854e-01
8.74893069e-01 6.25744581e-01 6.90162659e-01 -4.44768965e-01
-1.20610908e-01 -6.80859610e-02 -4.79915410e-01 -5.54026067e-01
7.80707300e-01 -1.23953529e-01 1.58845499e-01 1.70586377e-01
8.80879983e-02 1.93283394e-01 1.35130016e-02 2.31975779e-01
8.59452844e-01 6.19546831e-01 9.96343419e-02 2.43993297e-01
3.46942604e-01 -3.39463711e-01 7.66215503e-01 4.55767691e-01
3.68872993e-02 8.52912724e-01 6.37471199e-01 -4.84874576e-01
-1.67665470e+00 -9.43424523e-01 3.60121489e-01 9.88891482e-01
-1.30623236e-01 -6.66059434e-01 -7.94016421e-01 -4.27352071e-01
-3.24654758e-01 6.55695260e-01 -5.42877436e-01 -7.09089413e-02
-6.80685520e-01 -4.85789925e-01 3.27846795e-01 6.13368571e-01
3.25672597e-01 -9.41981852e-01 -3.03920180e-01 2.40804031e-01
-4.93666977e-01 -1.33467853e+00 -4.97012943e-01 -5.08292913e-01
-6.79862022e-01 -6.70310140e-01 -8.61147165e-01 -6.05602622e-01
4.64576691e-01 1.41837066e-02 1.19299126e+00 -1.70135081e-01
-5.37161790e-02 3.41545045e-01 -5.13249576e-01 9.47001129e-02
-6.69108212e-01 -1.23305522e-01 2.39231661e-01 3.90898496e-01
-1.37338594e-01 -7.34190345e-01 -7.02283204e-01 3.40312302e-01
-8.24857891e-01 6.80763841e-01 3.02458644e-01 1.07515109e+00
1.00123036e+00 1.35799631e-01 3.50134164e-01 -8.27243865e-01
4.76576835e-02 -6.96181297e-01 -6.72069907e-01 5.29956296e-02
-3.17008913e-01 -2.24036366e-01 1.12826633e+00 -4.10700589e-01
-1.12018013e+00 3.12959045e-01 -1.47723705e-01 -1.17857206e+00
-2.12832540e-02 8.16579759e-02 -2.85753012e-01 2.45553151e-01
3.66567820e-01 5.61775386e-01 -1.37939234e-04 -2.18807310e-01
4.62739408e-01 3.95396799e-01 4.94760811e-01 -3.54937196e-01
9.05139983e-01 4.38572615e-01 -9.66956988e-02 -6.78915620e-01
-4.94121969e-01 -3.77839319e-02 -5.15916824e-01 -2.33358175e-01
9.92679894e-01 -1.49190509e+00 -3.99204642e-01 3.51557702e-01
-8.03826809e-01 -6.29230678e-01 -4.87362534e-01 4.76264030e-01
-1.20386779e+00 2.72552788e-01 -8.34397137e-01 -6.11816227e-01
-1.34219512e-01 -1.33484173e+00 1.00404620e+00 -1.52155071e-01
4.74092178e-03 -6.67964041e-01 -1.12640485e-02 4.92587000e-01
4.55612183e-01 2.49011308e-01 6.22955620e-01 -1.21176653e-01
-1.17417514e+00 -5.58156148e-02 3.36574242e-02 5.70672095e-01
-1.39087334e-01 2.13748455e-01 -7.63802946e-01 -4.03310508e-01
-6.70225620e-02 -3.92234027e-01 6.55937672e-01 4.60418105e-01
1.63823354e+00 -7.43221521e-01 -1.52495921e-01 1.05745780e+00
1.35282457e+00 1.34049386e-01 1.03500807e+00 -6.91144690e-02
9.34167206e-01 2.94837892e-01 6.59192204e-01 7.54606426e-01
4.84927326e-01 9.53323781e-01 4.23885673e-01 2.37058535e-01
-7.89673105e-02 -8.13295066e-01 6.98143363e-01 9.00219321e-01
-4.43931341e-01 -3.75110418e-01 -6.11014307e-01 3.74496400e-01
-2.08640170e+00 -1.46915627e+00 2.14540288e-01 1.97847033e+00
5.88092566e-01 -8.16770792e-02 2.80475944e-01 -7.46657625e-02
6.25041187e-01 3.98864269e-01 -5.18465281e-01 7.16192946e-02
-3.03177088e-02 -1.42020673e-01 3.74361664e-01 4.09540564e-01
-1.16198981e+00 1.05106461e+00 6.05172014e+00 9.05484855e-01
-1.06733286e+00 1.60180658e-01 1.17851973e+00 -2.23006964e-01
-3.41989726e-01 3.32113132e-02 -8.31330180e-01 8.04751039e-01
1.24080241e+00 -3.07792574e-01 7.02890694e-01 1.02778447e+00
5.65927446e-01 4.72007990e-01 -1.07922232e+00 1.17674375e+00
1.22102231e-01 -1.69786382e+00 6.28726840e-01 1.10915646e-01
8.89285326e-01 -8.07025880e-02 2.89619207e-01 2.71476209e-01
2.20102996e-01 -1.14603317e+00 8.21250498e-01 9.07568455e-01
1.13761771e+00 -6.07636213e-01 5.24450183e-01 2.80675262e-01
-1.30737698e+00 -2.54921377e-01 -5.17191410e-01 7.16087148e-02
4.61825967e-01 1.92359626e-01 -5.46982765e-01 3.61451268e-01
8.26663494e-01 1.10425067e+00 -3.44909370e-01 6.72193229e-01
-7.92627558e-02 8.11484277e-01 5.11459000e-02 5.87162912e-01
-1.39871314e-01 -1.71928793e-01 3.65305036e-01 8.65251780e-01
8.70808899e-01 2.62432158e-01 2.22562775e-01 8.28889966e-01
-3.43791991e-01 2.88792700e-02 -6.58857703e-01 -1.09895490e-01
3.61052275e-01 9.07196581e-01 -8.38429034e-02 -5.21207511e-01
-3.96420211e-01 1.05535042e+00 2.03226015e-01 3.57037306e-01
-1.26568019e+00 3.56710672e-01 6.49335623e-01 4.79707450e-01
6.44018829e-01 -1.96459293e-01 2.85924822e-01 -1.57412755e+00
-2.26692148e-02 -1.09369934e+00 1.69919088e-01 -1.08246672e+00
-9.62325275e-01 7.43398726e-01 -1.15585372e-01 -1.62147713e+00
-9.45714533e-01 -2.05822140e-01 -2.70282298e-01 6.60817623e-01
-1.38373721e+00 -1.36692560e+00 -3.61302882e-01 9.20670629e-01
7.52999604e-01 -5.71233571e-01 8.40070784e-01 3.50445718e-01
-4.33362305e-01 6.71523809e-01 2.12001652e-01 1.05668388e-01
5.04447758e-01 -9.02278006e-01 4.54679132e-01 1.14726853e+00
2.55037010e-01 2.05881417e-01 6.88540459e-01 -6.04098856e-01
-1.42039740e+00 -1.59750175e+00 5.91036081e-01 -3.35599005e-01
5.14174581e-01 -4.11379904e-01 -8.22564065e-01 1.02496302e+00
6.91002840e-03 3.83103997e-01 6.78937376e-01 -5.23906410e-01
-3.13921779e-01 -1.80055693e-01 -8.54456723e-01 6.36944592e-01
1.08025825e+00 -6.24508858e-01 2.72469060e-03 2.35578552e-01
9.17434573e-01 -4.41304713e-01 -1.13756979e+00 3.03032994e-01
4.65809852e-01 -1.10496008e+00 1.09932315e+00 -4.65810329e-01
1.00114393e+00 -2.48028636e-01 -2.90596575e-01 -8.85091662e-01
-5.49527049e-01 -1.09541631e+00 -6.20572329e-01 1.31930304e+00
1.74829885e-01 -1.81310013e-01 1.25022376e+00 7.08154678e-01
2.28160713e-02 -7.13658512e-01 -6.94886982e-01 -4.92360979e-01
-1.97057411e-01 -4.64165777e-01 6.11249208e-01 7.85365641e-01
-3.80778700e-01 8.90766159e-02 -1.42189038e+00 2.18901694e-01
6.34138882e-01 6.93002716e-02 1.12338245e+00 -4.86607879e-01
-6.96841836e-01 -1.13673098e-02 -5.23447573e-01 -1.37137508e+00
2.75730610e-01 -4.66391146e-01 -9.68176946e-02 -9.90157425e-01
3.16538334e-01 -2.17976838e-01 -9.20539424e-02 1.92018762e-01
-4.06511165e-02 4.48514551e-01 5.17395318e-01 4.76540148e-01
-7.45223820e-01 9.44038033e-01 9.68643904e-01 1.60586596e-01
-8.79281387e-03 1.86046232e-02 -3.94473016e-01 6.32003367e-01
6.45757198e-01 -2.86447346e-01 -6.55843675e-01 -4.29370373e-01
7.48247951e-02 5.94613492e-01 4.92748857e-01 -1.12460482e+00
3.76837701e-02 -1.46637902e-01 5.45695245e-01 -3.25468898e-01
7.39913464e-01 -8.32413912e-01 8.46366167e-01 2.77231127e-01
-4.57982272e-01 2.02283010e-01 -1.90828755e-01 7.20392168e-01
-4.17054772e-01 2.22528666e-01 7.63896525e-01 -1.61515370e-01
-8.58196497e-01 9.07596707e-01 -2.34804437e-01 -1.15303852e-01
1.28265893e+00 -1.90659210e-01 -1.40154168e-01 -7.89254248e-01
-7.62843668e-01 -3.57772484e-02 7.28315651e-01 5.36263645e-01
9.13622975e-01 -1.48205471e+00 -8.64101052e-01 3.07816803e-01
-5.05452119e-02 -5.37064187e-02 6.20430410e-01 4.37850714e-01
-5.63258111e-01 2.26809651e-01 -3.14572513e-01 -6.13027751e-01
-1.28023505e+00 7.99709499e-01 1.56235546e-01 -3.22663784e-01
-7.46777773e-01 7.23236501e-01 3.44485790e-01 1.48907095e-01
1.39870897e-01 1.09648019e-01 -1.87949941e-01 -2.45191023e-01
6.44628704e-01 2.82621473e-01 -6.65826440e-01 -1.08892918e+00
1.79010734e-01 3.78602684e-01 6.91920742e-02 1.74096525e-01
1.53367615e+00 -3.42561752e-01 1.71783060e-01 2.72052914e-01
1.18987513e+00 -1.90868497e-01 -1.92857659e+00 -9.71907377e-02
-4.81497496e-01 -8.35207462e-01 -2.32432708e-01 -1.93366170e-01
-1.16312850e+00 8.53786528e-01 4.82791930e-01 7.76914582e-02
1.22601998e+00 -1.79728314e-01 1.00174963e+00 -1.40856549e-01
5.14473855e-01 -7.82006443e-01 1.85032248e-01 2.13705420e-01
8.79217267e-01 -1.15448356e+00 -5.85362390e-02 -4.08517301e-01
-8.81633461e-01 8.76900375e-01 5.53405821e-01 -1.90313146e-01
5.96776545e-01 1.63018048e-01 -2.78405845e-01 3.70740563e-01
-1.16400921e+00 4.37386960e-01 1.37269124e-01 5.89995027e-01
2.36360893e-01 1.30022466e-01 2.65455514e-01 4.90942925e-01
-3.49178404e-01 4.86424595e-01 5.73100626e-01 4.64427859e-01
9.65966936e-03 -1.03479290e+00 -1.33411467e-01 5.82004070e-01
-6.58292830e-01 -1.73478797e-01 3.79509866e-01 3.02026093e-01
5.23798652e-02 5.46199441e-01 1.13609411e-01 -7.00367093e-01
-1.36220142e-01 -8.05817693e-02 1.25172600e-01 -3.08223665e-01
-2.72660777e-02 1.21707343e-01 -3.80117446e-02 -9.18626666e-01
-5.31847358e-01 -7.72443414e-01 -5.83567441e-01 -5.49544096e-01
7.36410171e-02 1.40173152e-01 2.23304167e-01 5.89600682e-01
6.77656651e-01 3.28817099e-01 7.80956149e-01 -1.16685188e+00
-3.73446852e-01 -6.60621464e-01 -4.11598891e-01 6.40411735e-01
2.12168008e-01 -2.96322614e-01 -1.22137174e-01 8.39759111e-01] | [10.703797340393066, -0.6697838306427002] |
9926b13e-94e5-4c9e-b6ba-f34798b00ca3 | dynamic-enhancement-network-for-partial-multi | 2305.15762 | null | https://arxiv.org/abs/2305.15762v1 | https://arxiv.org/pdf/2305.15762v1.pdf | Dynamic Enhancement Network for Partial Multi-modality Person Re-identification | Many existing multi-modality studies are based on the assumption of modality integrity. However, the problem of missing arbitrary modalities is very common in real life, and this problem is less studied, but actually important in the task of multi-modality person re-identification (Re-ID). To this end, we design a novel dynamic enhancement network (DENet), which allows missing arbitrary modalities while maintaining the representation ability of multiple modalities, for partial multi-modality person Re-ID. To be specific, the multi-modal representation of the RGB, near-infrared (NIR) and thermal-infrared (TIR) images is learned by three branches, in which the information of missing modalities is recovered by the feature transformation module. Since the missing state might be changeable, we design a dynamic enhancement module, which dynamically enhances modality features according to the missing state in an adaptive manner, to improve the multi-modality representation. Extensive experiments on multi-modality person Re-ID dataset RGBNT201 and vehicle Re-ID dataset RGBNT100 comparing to the state-of-the-art methods verify the effectiveness of our method in complex and changeable environments. | ['Jin Tang', 'Chenglong Li', 'Zi Wang', 'Ziling He', 'Aihua Zheng'] | 2023-05-25 | null | null | null | null | ['person-re-identification'] | ['computer-vision'] | [ 1.76890835e-01 -7.06802547e-01 6.60814345e-02 -3.44744861e-01
-2.55308867e-01 -4.10972029e-01 5.76146543e-01 -4.20171469e-01
-5.67031384e-01 5.17030656e-01 3.17303121e-01 3.33449543e-01
-1.03471421e-01 -7.57199049e-01 -3.76950324e-01 -9.52797234e-01
4.39902514e-01 1.81656569e-01 -1.61720321e-01 -4.23295826e-01
-1.01853006e-01 2.18079448e-01 -1.94267869e+00 2.13091508e-01
7.75426269e-01 7.43403852e-01 -1.18876949e-01 4.14422810e-01
5.36712185e-02 4.93254572e-01 -2.13574916e-01 -3.38921666e-01
3.65949482e-01 -4.16851789e-01 -6.96655333e-01 8.66847858e-02
3.15709293e-01 -4.01213288e-01 -7.14894533e-01 1.13053513e+00
8.50819588e-01 4.98733610e-01 4.49758053e-01 -1.58457184e+00
-8.92387092e-01 2.13545591e-01 -6.27578616e-01 3.48982178e-02
6.78619385e-01 1.57358676e-01 3.06694448e-01 -6.12826824e-01
2.83562005e-01 1.35879922e+00 7.59732902e-01 9.96905982e-01
-8.16545486e-01 -7.98004568e-01 2.68936306e-01 6.51319861e-01
-1.56560349e+00 -4.46869940e-01 8.57508898e-01 -1.84610620e-01
3.54969919e-01 3.83982658e-01 6.04403257e-01 1.12711656e+00
-3.21047872e-01 5.16038895e-01 1.36647844e+00 -2.93375045e-01
-2.65935332e-01 3.22732806e-01 1.75902843e-01 3.57069463e-01
1.25082612e-01 3.36798221e-01 -4.49108988e-01 -2.65033301e-02
5.94325662e-01 5.57509720e-01 -2.44061962e-01 -6.45658821e-02
-1.50861037e+00 2.26575077e-01 3.14683586e-01 3.45169127e-01
-2.89522350e-01 -2.51221418e-01 3.88039768e-01 5.42162240e-01
1.33986562e-01 -2.29642317e-01 -2.16580883e-01 -7.14466050e-02
-4.05540317e-01 5.66023402e-02 3.02180946e-01 7.60575712e-01
8.15566540e-01 -7.99291283e-02 -1.02669060e-01 1.00541973e+00
3.99763376e-01 9.33076918e-01 6.17397964e-01 -3.97571832e-01
4.51823890e-01 7.28456199e-01 1.52941257e-01 -8.70764911e-01
-7.21692026e-01 -2.03079522e-01 -1.36714911e+00 -6.03424124e-02
3.02216977e-01 -1.59834504e-01 -9.05534565e-01 1.88146663e+00
6.00214124e-01 3.73836786e-01 2.83643961e-01 1.22677112e+00
1.23357606e+00 5.20920098e-01 3.91924232e-01 -2.21567675e-01
1.54231048e+00 -6.86628520e-01 -7.21709549e-01 -4.84381057e-02
9.99280885e-02 -5.84633887e-01 7.03208625e-01 3.74035947e-02
-4.59962040e-01 -1.01518571e+00 -7.78522253e-01 8.47317104e-04
-5.95706165e-01 9.81624275e-02 4.84903753e-01 7.12028384e-01
-7.42780626e-01 2.05595285e-01 -2.27504075e-01 -5.40577948e-01
1.22103840e-02 3.91681284e-01 -8.78899455e-01 -4.99507278e-01
-1.59645784e+00 7.19782591e-01 4.65381205e-01 5.19813895e-01
-5.85509479e-01 -4.59094584e-01 -7.14444995e-01 -2.87501276e-01
3.14884216e-01 -8.56368363e-01 5.95480680e-01 -1.06506050e+00
-1.38290215e+00 7.12501228e-01 -1.18607044e-01 3.54674280e-01
4.80580181e-01 1.73661590e-01 -1.10381186e+00 4.87223379e-02
-1.12447828e-01 2.70610273e-01 9.37952399e-01 -1.46721029e+00
-8.01303089e-01 -8.07135940e-01 1.01677425e-01 5.64809620e-01
-5.71964383e-01 1.14156127e-01 -3.95180464e-01 -3.17270905e-01
9.43115428e-02 -1.10403025e+00 6.16298169e-02 -1.82279691e-01
-4.22877491e-01 -1.96229115e-01 8.56502295e-01 -1.02796566e+00
7.88637996e-01 -2.26595235e+00 2.65668869e-01 2.94026852e-01
2.12402493e-02 1.56274542e-01 -3.75863850e-01 1.51279464e-01
-2.32398599e-01 -1.09576985e-01 -5.12540117e-02 -4.58169311e-01
1.84031576e-02 8.77357125e-02 1.21440671e-01 6.16676867e-01
-4.19209629e-01 7.74726987e-01 -7.77846694e-01 -5.06469846e-01
5.90079010e-01 8.07895124e-01 1.35564014e-01 2.32328132e-01
4.91909564e-01 8.69944930e-01 -4.72462207e-01 1.08013391e+00
1.10507512e+00 -3.04216426e-02 -1.45002147e-02 -7.45271742e-01
-1.58269271e-01 -6.64393604e-01 -1.48322701e+00 1.67196751e+00
-1.50257483e-01 1.67269558e-01 -7.81106278e-02 -8.67756188e-01
7.58588076e-01 4.69009846e-01 8.89313519e-01 -1.03680122e+00
2.14783669e-01 1.16166694e-03 -2.68359333e-01 -6.36285722e-01
7.29976475e-01 -9.10670832e-02 -1.13812953e-01 4.35029924e-01
-2.57459670e-01 7.20058382e-01 -2.01960996e-01 -1.45544246e-01
5.60094118e-01 2.10794881e-01 -1.02745444e-01 3.04838836e-01
9.60888386e-01 -2.88462996e-01 6.58059180e-01 6.07910335e-01
-4.85506058e-01 5.95656455e-01 -3.50585669e-01 -3.85773063e-01
-1.05138338e+00 -9.41442132e-01 -1.84152368e-02 1.19186211e+00
8.04814577e-01 -7.27516785e-02 -3.63646775e-01 -4.30747360e-01
9.79887880e-03 9.89197344e-02 -6.98735297e-01 -3.32427621e-01
-4.57664341e-01 -1.14843535e+00 5.76566041e-01 3.66769642e-01
1.29498065e+00 -8.44695270e-01 -2.86395848e-01 3.97986092e-04
-7.76780009e-01 -1.26397943e+00 -4.70315158e-01 -5.36058664e-01
-4.08125758e-01 -1.13220096e+00 -1.07142079e+00 -7.60339200e-01
7.02389836e-01 7.23910749e-01 5.24050295e-01 2.40142018e-01
-1.12645634e-01 9.40752089e-01 -5.68767130e-01 -7.38841444e-02
-6.87949136e-02 -9.31195468e-02 4.44898754e-01 5.31179070e-01
6.16357803e-01 -4.53937322e-01 -5.99366188e-01 4.37713981e-01
-9.19989765e-01 1.29850330e-02 4.23475087e-01 9.66469049e-01
5.75756133e-01 3.70111108e-01 3.82313401e-01 -3.21234524e-01
3.61800671e-01 -3.47273529e-01 -8.56513754e-02 7.31622934e-01
-4.57410216e-01 -3.13537389e-01 5.67767859e-01 -9.06024039e-01
-1.39036322e+00 1.36200801e-01 -5.96374944e-02 -3.27936798e-01
-3.53166431e-01 2.30928034e-01 -5.68754554e-01 -4.27551299e-01
2.41718665e-01 6.64387286e-01 -1.55313658e-02 -5.80845892e-01
4.14793193e-01 9.49038446e-01 8.60571623e-01 -4.39218581e-01
1.16288710e+00 6.60715640e-01 8.78244787e-02 -5.65105915e-01
-3.65955412e-01 -5.06006420e-01 -6.54578686e-01 -4.23560500e-01
7.71052122e-01 -1.23698437e+00 -1.02651989e+00 1.17322981e+00
-8.49868298e-01 1.41875729e-01 -1.30915985e-01 5.62403738e-01
-1.11812070e-01 6.60542250e-01 -4.77441490e-01 -8.09202135e-01
-4.31708992e-01 -8.83227885e-01 9.73091066e-01 8.59002650e-01
5.91923952e-01 -7.53261566e-01 -1.90291703e-02 6.18969440e-01
4.39009428e-01 2.48518094e-01 5.29400170e-01 -1.64499462e-01
-2.03040540e-01 -7.66488910e-02 -5.04805446e-01 1.65505126e-01
4.13624406e-01 -4.25010145e-01 -1.04449761e+00 -4.18695599e-01
-3.50136608e-01 2.57220282e-03 8.13867152e-01 -4.90319319e-02
1.02643347e+00 -1.53954690e-02 -3.59059423e-01 7.41144419e-01
1.41944742e+00 7.64098763e-02 8.60129893e-01 6.56944156e-01
1.16168523e+00 6.28506005e-01 5.94402254e-01 5.93538105e-01
1.04813278e+00 7.11512208e-01 3.61318409e-01 -2.52493829e-01
-1.03666738e-01 -1.57368034e-01 2.40224376e-01 5.99981606e-01
-5.91412067e-01 -1.11261882e-01 -4.47688013e-01 3.68577570e-01
-1.99667358e+00 -1.41783512e+00 -2.59326193e-02 2.20488405e+00
4.40919220e-01 -6.00302756e-01 4.37844306e-01 2.31015056e-01
1.08652079e+00 4.16655093e-03 -6.72017872e-01 2.47692674e-01
-5.05397081e-01 -6.00880623e-01 4.77437288e-01 -5.78719974e-02
-1.19736826e+00 4.80950147e-01 5.61721039e+00 5.24673760e-01
-9.98034835e-01 3.05435270e-01 2.04211354e-01 2.19573915e-01
-3.30111146e-01 -1.11700736e-01 -6.16372347e-01 7.34628201e-01
4.33478832e-01 -1.25986512e-03 8.67669880e-01 4.07959759e-01
8.50801822e-04 -2.75419038e-02 -8.05888176e-01 1.76224387e+00
4.01581407e-01 -5.64226031e-01 -3.32791470e-02 5.99048175e-02
6.11944079e-01 -3.13694894e-01 1.39279127e-01 3.59035134e-01
-9.72787067e-02 -7.48557031e-01 4.75539118e-01 1.10424566e+00
1.02094185e+00 -7.61057079e-01 8.97417963e-01 1.76484600e-01
-1.65327525e+00 -4.14018035e-01 -4.41212386e-01 3.07832181e-01
2.20207676e-01 3.67714912e-01 3.56340706e-02 1.10532546e+00
1.01214254e+00 8.77612174e-01 -8.70688140e-01 8.93052101e-01
1.32511571e-01 3.12146563e-02 -2.87262291e-01 3.99828762e-01
-4.37774003e-01 -1.31147072e-01 5.38924873e-01 8.80441606e-01
3.23520124e-01 2.93148160e-01 2.78259724e-01 5.71769357e-01
8.54047835e-02 -1.02677912e-01 -2.49283329e-01 1.92035943e-01
5.44033051e-01 1.27278042e+00 -1.28083602e-01 -2.33340815e-01
-4.42698300e-01 1.30879116e+00 -1.03991680e-01 5.42773485e-01
-9.38676715e-01 -1.96473733e-01 6.07651591e-01 -3.11055392e-01
-1.30385339e-01 -6.32424429e-02 1.58144072e-01 -1.57703936e+00
1.40946195e-01 -9.01626050e-01 7.43249238e-01 -8.95232260e-01
-1.75976062e+00 5.56880236e-01 -2.17683334e-02 -1.47833228e+00
-1.13324877e-02 -3.09109688e-01 -2.02962235e-01 1.01278937e+00
-1.94125199e+00 -1.90129268e+00 -9.78536248e-01 1.36500645e+00
-6.81718215e-02 -4.27554399e-01 6.42793834e-01 8.22284758e-01
-7.29746938e-01 1.00571775e+00 1.55230835e-01 9.94865000e-02
9.26872671e-01 -7.12509036e-01 -2.45088845e-01 8.67961287e-01
-4.77004766e-01 5.63355148e-01 4.43976134e-01 -7.09046602e-01
-1.73241127e+00 -1.01264513e+00 5.31512260e-01 -1.37636095e-01
1.04351863e-01 1.04819722e-02 -8.35516274e-01 5.58116615e-01
-1.06112741e-01 1.24358580e-01 8.14826012e-01 1.45084858e-01
-5.67977309e-01 -4.20597047e-01 -1.46589482e+00 3.86543870e-01
1.23715317e+00 -7.37856746e-01 -3.90593767e-01 1.50863260e-01
4.58536237e-01 -2.60725409e-01 -1.11931813e+00 5.32316625e-01
8.59962940e-01 -6.84082508e-01 1.42311454e+00 -3.53318393e-01
-1.85239404e-01 -7.05551803e-01 -3.52521151e-01 -1.25383818e+00
-5.62465131e-01 -1.36116713e-01 -2.44203284e-02 1.63502598e+00
-3.87706459e-01 -8.47428918e-01 3.86653960e-01 8.54748547e-01
2.68502831e-01 1.04663707e-01 -1.14857149e+00 -6.43752277e-01
-3.24805409e-01 -6.31874129e-02 1.27406251e+00 1.03777111e+00
-2.40895599e-01 -1.21026233e-01 -1.09469688e+00 3.58849198e-01
8.89656901e-01 1.06651165e-01 8.03093493e-01 -1.32525718e+00
3.55953276e-02 -6.66499361e-02 -4.41947520e-01 -7.47138202e-01
1.00855507e-01 -5.19403815e-01 -1.15740448e-01 -1.42928398e+00
6.87217355e-01 -6.71401918e-01 -8.01082075e-01 6.27521813e-01
-3.86390269e-01 4.55728740e-01 3.97792578e-01 5.45433462e-01
-7.74821401e-01 8.02519441e-01 1.26134384e+00 -4.77106959e-01
-2.31014788e-01 -1.59205884e-01 -8.96917403e-01 3.39569300e-01
5.92086494e-01 -9.18375105e-02 -2.62344271e-01 -4.19403464e-01
5.14010452e-02 -8.65774006e-02 8.84653628e-01 -1.13659143e+00
4.52906996e-01 -4.14119452e-01 7.25266874e-01 -6.85894728e-01
5.35249054e-01 -1.17404962e+00 5.49282670e-01 7.26008937e-02
-5.35519160e-02 8.32382962e-02 -2.70716473e-02 4.64816809e-01
-4.96836714e-02 1.42339259e-01 6.48557365e-01 -2.71779418e-01
-1.36863661e+00 7.01382101e-01 2.50510722e-02 -4.96647507e-01
8.99536192e-01 -4.59401399e-01 -7.28956282e-01 -1.95138663e-01
-5.81974387e-01 2.51468569e-01 5.70888162e-01 7.28374422e-01
7.63131082e-01 -1.80080795e+00 -7.93182492e-01 2.37602040e-01
4.57753330e-01 -4.56230849e-01 1.15747833e+00 7.06376910e-01
1.50252089e-01 -1.29283339e-01 -5.89576006e-01 -4.83934999e-01
-1.47387791e+00 7.48344064e-01 6.40610993e-01 -1.26821594e-02
-4.94976074e-01 2.53335595e-01 -1.87620118e-01 -7.22581029e-01
-8.09940621e-02 6.64840877e-01 -7.54344523e-01 5.85921444e-02
8.78750265e-01 6.71722591e-01 -3.89402509e-01 -1.36152864e+00
-6.23225152e-01 9.90007579e-01 1.28212884e-01 2.51360182e-02
1.18334782e+00 -9.11221623e-01 -3.33904535e-01 3.39334428e-01
1.04536664e+00 -3.28242034e-01 -8.94095123e-01 -5.70369124e-01
-7.95255363e-01 -6.33284926e-01 -1.99983254e-01 -9.43540037e-01
-1.08414280e+00 5.03210008e-01 1.45376050e+00 -1.56788543e-01
1.53063738e+00 -3.30587059e-01 9.52409446e-01 3.11714262e-01
5.70608199e-01 -1.24081850e+00 -2.31573563e-02 3.42565060e-01
4.14640218e-01 -1.46325588e+00 5.12535125e-02 -3.65295947e-01
-5.39151788e-01 9.87810910e-01 8.77940834e-01 4.52867895e-01
5.72229147e-01 -2.77465880e-01 2.80805193e-02 1.87838703e-01
3.69542055e-02 -3.98595273e-01 2.62788773e-01 1.12348866e+00
-1.64138526e-01 2.22026646e-01 5.86655922e-02 5.76231182e-01
1.25813067e-01 1.49139956e-01 2.50413716e-01 7.64257073e-01
-1.09829016e-01 -1.10537660e+00 -8.90665591e-01 1.84746549e-01
4.60455343e-02 1.96971774e-01 -4.48039584e-02 5.62776208e-01
6.76137626e-01 1.40035951e+00 -2.19466537e-01 -9.62835968e-01
3.44923824e-01 -1.99617833e-01 4.48284239e-01 9.28256661e-02
-5.41321754e-01 -3.24585885e-01 -1.70179442e-01 -3.06689411e-01
-1.09033680e+00 -8.22010934e-01 -9.66442049e-01 -7.54334688e-01
-1.69386148e-01 -1.06859177e-01 6.45248592e-01 1.10502768e+00
2.31584340e-01 4.68717486e-01 9.39352393e-01 -7.06927299e-01
-3.16345543e-01 -8.04497063e-01 -7.81794310e-01 9.47547436e-01
4.67433900e-01 -8.25460672e-01 -1.88426003e-01 9.16739460e-03] | [14.644067764282227, 0.9632728695869446] |
f9f21f73-3046-477d-830d-6624285bad55 | adaptive-fine-grained-sketch-based-image | 2207.01723 | null | https://arxiv.org/abs/2207.01723v3 | https://arxiv.org/pdf/2207.01723v3.pdf | Adaptive Fine-Grained Sketch-Based Image Retrieval | The recent focus on Fine-Grained Sketch-Based Image Retrieval (FG-SBIR) has shifted towards generalising a model to new categories without any training data from them. In real-world applications, however, a trained FG-SBIR model is often applied to both new categories and different human sketchers, i.e., different drawing styles. Although this complicates the generalisation problem, fortunately, a handful of examples are typically available, enabling the model to adapt to the new category/style. In this paper, we offer a novel perspective -- instead of asking for a model that generalises, we advocate for one that quickly adapts, with just very few samples during testing (in a few-shot manner). To solve this new problem, we introduce a novel model-agnostic meta-learning (MAML) based framework with several key modifications: (1) As a retrieval task with a margin-based contrastive loss, we simplify the MAML training in the inner loop to make it more stable and tractable. (2) The margin in our contrastive loss is also meta-learned with the rest of the model. (3) Three additional regularisation losses are introduced in the outer loop, to make the meta-learned FG-SBIR model more effective for category/style adaptation. Extensive experiments on public datasets suggest a large gain over generalisation and zero-shot based approaches, and a few strong few-shot baselines. | ['Yi-Zhe Song', 'Tao Xiang', 'Pinaki Nath Chowdhury', 'Animesh Gupta', 'Parth Shah', 'Aneeshan Sain', 'Ayan Kumar Bhunia'] | 2022-07-04 | null | null | null | null | ['sketch-based-image-retrieval'] | ['computer-vision'] | [ 3.83933932e-01 -3.54092836e-01 -3.36416066e-01 -3.40354174e-01
-8.39655876e-01 -6.79624736e-01 9.18066800e-01 -2.15538993e-01
-2.85152704e-01 4.65369940e-01 1.20346181e-01 1.21260732e-01
-1.14944860e-01 -7.58845448e-01 -6.53881133e-01 -5.41587472e-01
2.71746993e-01 5.27440310e-01 3.78999472e-01 -3.76657218e-01
3.91210854e-01 4.92596686e-01 -1.59443998e+00 3.76629084e-01
6.78407550e-01 9.62968886e-01 3.47877920e-01 7.19160795e-01
-2.77964652e-01 3.68873268e-01 -5.66913068e-01 -8.28892887e-01
4.88780469e-01 -4.69871938e-01 -8.32435429e-01 2.71497160e-01
7.20159173e-01 -4.69889402e-01 -3.55062038e-01 7.29267538e-01
5.17044604e-01 6.03981912e-01 9.31854248e-01 -1.20941019e+00
-9.32600737e-01 2.00248212e-01 -4.89178538e-01 -4.87211421e-02
2.99143404e-01 2.58142114e-01 1.13081849e+00 -1.02125013e+00
8.31085324e-01 1.34661925e+00 5.43251097e-01 1.01231861e+00
-1.34045208e+00 -7.83196628e-01 5.72367132e-01 7.57029429e-02
-1.36158860e+00 -3.81213218e-01 9.66952205e-01 -8.16528201e-02
5.15700519e-01 2.47845873e-01 4.85228777e-01 1.31044269e+00
-2.41929874e-01 1.05885494e+00 8.21978569e-01 -4.52971488e-01
3.60217959e-01 3.33361447e-01 2.82149110e-02 3.76276404e-01
-3.61260585e-02 -2.09501889e-02 -4.60025430e-01 -2.39478156e-01
9.77457881e-01 3.12700868e-01 1.30252745e-02 -9.50290024e-01
-9.28467631e-01 9.08664465e-01 5.50543070e-01 2.47027293e-01
-8.74023288e-02 2.59219825e-01 4.98400867e-01 5.81650972e-01
5.07866025e-01 5.23768246e-01 -3.16521615e-01 3.46856043e-02
-1.07003319e+00 4.44278866e-01 6.68919384e-01 1.14361858e+00
9.30004954e-01 -1.87983647e-01 -4.07251894e-01 1.39897776e+00
-1.43984230e-02 2.66124547e-01 4.72538531e-01 -8.61962378e-01
3.32072705e-01 2.89125144e-01 -6.40429258e-02 -7.84200072e-01
4.03836727e-01 -2.84751803e-01 -7.64484406e-01 2.01986998e-01
1.73129365e-01 3.55327666e-01 -1.26009619e+00 1.80285788e+00
1.70492172e-01 1.54188067e-01 -2.88228512e-01 8.07697594e-01
4.91894454e-01 5.79597473e-01 2.50027269e-01 1.78553224e-01
1.10416520e+00 -1.23798525e+00 -2.24476963e-01 -3.64630461e-01
-5.40206814e-03 -6.97569370e-01 1.45770800e+00 4.26160693e-01
-1.18112922e+00 -7.36294806e-01 -1.03491557e+00 -1.05970263e-01
-6.48331463e-01 -3.14170897e-01 6.89350724e-01 5.36496103e-01
-9.49345469e-01 7.60785401e-01 -3.31393003e-01 -6.30159676e-01
5.43170989e-01 1.07581265e-01 -2.94613868e-01 -5.59920073e-01
-1.01156604e+00 7.34869301e-01 1.94522694e-01 -1.96577087e-01
-8.51221025e-01 -8.24307144e-01 -6.31323278e-01 3.50974724e-02
4.73406464e-01 -9.74815190e-01 1.23054600e+00 -1.31434917e+00
-1.64034867e+00 8.83564472e-01 -8.56835172e-02 -3.90837044e-02
7.64895618e-01 2.91918572e-02 -2.75844902e-01 1.18343368e-01
3.67588620e-03 9.52945828e-01 1.42597508e+00 -1.48168075e+00
-5.59520066e-01 -1.07484207e-01 4.07845825e-01 1.67217091e-01
-3.90375018e-01 -1.10742718e-01 -8.84190559e-01 -1.08239174e+00
-1.25078291e-01 -9.00821447e-01 -6.09606802e-02 5.02125680e-01
8.48681778e-02 -3.96903515e-01 7.07146883e-01 -3.48862320e-01
1.22059155e+00 -2.07449293e+00 3.36991131e-01 1.92239091e-01
1.40230423e-02 4.73030835e-01 -7.11457014e-01 5.62208116e-01
1.28819570e-01 1.24859467e-01 -4.26992893e-01 -5.62057555e-01
6.45248070e-02 3.94306183e-01 -4.30779248e-01 -8.92538428e-02
4.54410791e-01 9.75574553e-01 -1.16854584e+00 -2.27526590e-01
1.95019558e-01 3.95909458e-01 -7.41477549e-01 2.91167945e-01
-2.64066547e-01 2.09237993e-01 -3.54167789e-01 7.20481098e-01
6.83086097e-01 -1.22796066e-01 -6.66926056e-02 -5.13620041e-02
3.44663501e-01 -1.14963464e-01 -1.25396204e+00 2.06260943e+00
-8.45595777e-01 2.70373881e-01 -1.32219151e-01 -9.52962995e-01
8.86081219e-01 2.46382300e-02 5.87301934e-03 -6.26218557e-01
-2.23472208e-01 7.25125074e-02 -3.28597277e-01 -2.14709565e-01
5.30348778e-01 -3.70345294e-01 -1.14276521e-01 5.35628796e-01
3.21797431e-01 -3.50818574e-01 1.06124394e-01 2.91577339e-01
9.69037116e-01 3.36623996e-01 1.67683750e-01 7.05107078e-02
4.11034763e-01 -2.65073299e-01 3.58356416e-01 1.13985908e+00
-1.92891955e-01 9.42186534e-01 -6.97709098e-02 -5.27519882e-01
-1.20017827e+00 -1.20841157e+00 5.83303161e-03 1.53160679e+00
1.14391647e-01 -2.60547042e-01 -4.26371574e-01 -9.03908372e-01
1.89014852e-01 6.21767700e-01 -7.43388355e-01 -3.26350480e-01
-4.80040431e-01 -3.89541358e-01 1.75604045e-01 5.36069155e-01
4.52615172e-01 -1.14429224e+00 -3.30751896e-01 2.73809135e-01
5.97693101e-02 -5.98013341e-01 -7.95547962e-01 -5.87439165e-02
-8.69444370e-01 -7.54815698e-01 -1.18437016e+00 -6.74517870e-01
6.16298616e-01 6.40334129e-01 1.19333196e+00 3.24249059e-01
-4.05076236e-01 7.30951071e-01 -5.57940543e-01 -2.54289538e-01
-1.86539412e-01 1.48294270e-01 -3.25502515e-01 8.97222012e-02
2.69522786e-01 -5.97714782e-01 -7.49717832e-01 2.73923248e-01
-1.09543216e+00 -9.37753022e-02 7.92294025e-01 1.21692014e+00
3.17759216e-01 -2.92257160e-01 6.56126320e-01 -1.01441944e+00
6.64151013e-01 -3.31756353e-01 -2.18674868e-01 4.68110442e-01
-7.19682992e-01 1.11156017e-01 5.81821084e-01 -1.00091934e+00
-1.00501990e+00 -1.98266625e-01 1.04278112e-02 -8.65825951e-01
1.25562388e-03 2.36162886e-01 -1.72132879e-01 -2.73869306e-01
6.20993316e-01 2.25732043e-01 -6.36637630e-03 -5.80905378e-01
7.53760099e-01 6.64905429e-01 5.47163665e-01 -8.40304077e-01
1.12557423e+00 2.95294046e-01 -3.03092867e-01 -6.61097646e-01
-8.12416077e-01 -6.80641472e-01 -6.29452646e-01 -1.88321583e-02
3.39315027e-01 -7.34960437e-01 -1.91760719e-01 3.98290187e-01
-9.02547359e-01 -4.97514337e-01 -5.01194119e-01 -1.58299088e-01
-6.41810238e-01 3.82704079e-01 -4.39655691e-01 -8.28817070e-01
-3.90441984e-01 -9.51190710e-01 1.17524326e+00 1.60216480e-01
-2.12395027e-01 -1.10699785e+00 2.12221980e-01 9.19519216e-02
7.03495085e-01 -6.86374605e-02 1.02080631e+00 -6.11594975e-01
-5.35772502e-01 -3.85977477e-01 -3.87626290e-01 4.79656130e-01
1.35743126e-01 -1.41138270e-01 -1.13482225e+00 -4.95978564e-01
-4.13514763e-01 -5.64162731e-01 1.07636142e+00 -1.23853996e-01
1.39458585e+00 -3.59665811e-01 -2.18830556e-01 6.02803588e-01
1.39417636e+00 -5.04734516e-02 6.45647705e-01 3.67054999e-01
5.63863635e-01 3.51717860e-01 6.70830607e-01 2.91586488e-01
2.07466021e-01 8.95305455e-01 1.01228189e-02 2.16581021e-03
-5.18833935e-01 -4.80846018e-01 1.68205518e-02 5.64646900e-01
-1.90015305e-02 -1.45067304e-01 -4.52512920e-01 6.62722945e-01
-1.91381145e+00 -1.04873717e+00 7.58246243e-01 2.52030635e+00
9.56873417e-01 -9.45510566e-02 3.50619107e-01 -1.29834041e-01
6.76908851e-01 3.09429079e-01 -7.14269996e-01 -4.91581798e-01
8.83653313e-02 6.85888529e-01 7.02770948e-02 3.67274046e-01
-1.05742741e+00 1.14931202e+00 6.09933949e+00 1.17869413e+00
-1.14889622e+00 7.08734393e-02 5.19142807e-01 -1.47467867e-01
-4.61900890e-01 1.79422140e-01 -5.09679437e-01 3.91421258e-01
4.76761520e-01 -1.16556652e-01 9.03738022e-01 8.75892401e-01
-4.10562366e-01 1.91432476e-01 -1.23335171e+00 1.14587581e+00
3.49598765e-01 -1.16820121e+00 5.47987819e-01 -1.45116881e-01
7.44475782e-01 -2.79115587e-01 1.87736914e-01 7.08266675e-01
2.28613153e-01 -7.26769626e-01 7.99658477e-01 5.98109245e-01
1.12442112e+00 -8.33115399e-01 3.04989636e-01 1.95489794e-01
-1.14354503e+00 -2.37047613e-01 -7.26240337e-01 1.81094378e-01
-1.13354579e-01 9.56536382e-02 -4.75364506e-01 5.78730464e-01
4.92546916e-01 6.03870928e-01 -7.17105567e-01 1.04699218e+00
-9.66407284e-02 1.99787483e-01 -4.85979114e-03 8.80670547e-02
3.05790663e-01 -1.54198706e-01 5.74274302e-01 1.25612915e+00
1.37807220e-01 1.47468969e-01 1.50055155e-01 8.86699259e-01
-2.39337653e-01 -1.41057476e-01 -5.48802733e-01 5.37804402e-02
5.55159986e-01 1.37098694e+00 -4.39794451e-01 -4.60534394e-01
-4.38835680e-01 1.50365877e+00 5.86832225e-01 5.25018275e-01
-4.27614927e-01 -5.77058196e-01 5.80455720e-01 1.52725324e-01
5.64816177e-01 6.76334798e-02 2.00432930e-02 -1.24802685e+00
-8.55370313e-02 -9.19368625e-01 4.85117674e-01 -8.01482975e-01
-1.90595496e+00 3.55544776e-01 1.43231213e-01 -1.17571127e+00
-3.92196506e-01 -4.81310874e-01 -8.23738575e-01 9.90190566e-01
-1.79158783e+00 -1.62023783e+00 -1.57159239e-01 5.20936430e-01
9.01070654e-01 -1.83272481e-01 7.91718543e-01 2.28641137e-01
-3.80504936e-01 1.08826256e+00 -1.01037072e-02 3.26432921e-02
1.12525725e+00 -1.19429088e+00 5.44404685e-01 7.36792743e-01
1.48176581e-01 9.22737777e-01 5.59774339e-01 -4.44370151e-01
-1.38346016e+00 -1.23879933e+00 7.11344361e-01 -4.91880089e-01
6.01404309e-01 -4.85718369e-01 -1.14322710e+00 3.96043628e-01
-2.70201098e-02 2.12784618e-01 4.21879143e-01 2.42161751e-01
-8.08349609e-01 -1.49789304e-01 -1.24912620e+00 7.08047748e-01
1.29525828e+00 -7.00956345e-01 -6.09900057e-01 1.10830188e-01
4.11929309e-01 -2.63442323e-02 -7.94745862e-01 2.83811361e-01
9.65340555e-01 -7.01678932e-01 1.31334817e+00 -7.87374973e-01
3.70147437e-01 -1.54121578e-01 -1.62158445e-01 -1.27925634e+00
-6.75913990e-01 -5.21846533e-01 -3.44288319e-01 1.29585791e+00
8.67836624e-02 -4.95955974e-01 7.44785666e-01 8.17677200e-01
5.34728803e-02 -7.37112641e-01 -7.77805746e-01 -1.06332266e+00
2.84468114e-01 -1.79946512e-01 6.35754883e-01 8.50248814e-01
-2.43456617e-01 3.23769897e-01 -5.21028221e-01 -3.10553819e-01
5.26305258e-01 1.44511476e-01 1.08687007e+00 -1.31936800e+00
-5.21638989e-01 -6.19619489e-01 -2.97314912e-01 -1.20028520e+00
9.44280922e-02 -8.09596598e-01 5.64841479e-02 -1.34476960e+00
5.97072303e-01 -5.88418365e-01 -5.38379252e-01 4.97186393e-01
-6.00575984e-01 3.68095279e-01 5.63589931e-01 4.43253636e-01
-6.62072897e-01 6.19990706e-01 1.10367298e+00 -3.71318996e-01
-2.28692546e-01 1.64185986e-01 -7.68918335e-01 3.51383060e-01
4.46412235e-01 -3.46724242e-01 -5.64406872e-01 -2.70195633e-01
-8.38273987e-02 -3.06254148e-01 6.10226274e-01 -7.38659501e-01
1.80695280e-01 -2.43363544e-01 4.79724556e-01 -2.13050693e-01
3.99639815e-01 -5.89327931e-01 6.28282083e-03 9.95546654e-02
-4.02894378e-01 -2.94307202e-01 8.37022588e-02 7.58338511e-01
3.04507483e-02 -4.43885565e-01 8.93769205e-01 -3.29032660e-01
-9.53005135e-01 7.30088592e-01 9.94203538e-02 -4.92234007e-02
4.90570158e-01 -5.15973687e-01 -1.14588790e-01 -4.45729703e-01
-5.22528708e-01 2.02697664e-02 9.10849452e-01 7.27305770e-01
5.76418996e-01 -1.48531950e+00 -4.84093964e-01 1.63147524e-01
7.18823612e-01 -5.37937135e-02 3.19879115e-01 3.11678529e-01
5.41767105e-02 8.96002278e-02 -1.89382210e-01 -3.59467626e-01
-8.69895160e-01 9.31179106e-01 1.30640104e-01 -2.23190054e-01
-6.81108117e-01 9.01964128e-01 3.90024841e-01 -6.18290901e-01
2.98639357e-01 2.01672524e-01 1.11374028e-01 4.40605804e-02
6.25619471e-01 1.92033157e-01 -7.37438630e-03 -2.79973686e-01
-5.52998669e-02 6.66697860e-01 -5.13451159e-01 -2.84768701e-01
1.29719436e+00 -1.66967347e-01 2.12228671e-01 6.51612401e-01
1.15691829e+00 -2.50986189e-01 -1.59390497e+00 -4.92491752e-01
-5.51518910e-02 -7.94324994e-01 -9.56875980e-02 -9.86686051e-01
-7.86325634e-01 8.64666700e-01 5.56353867e-01 -1.23787686e-01
1.14796948e+00 -7.83895552e-02 7.95149803e-01 4.83292401e-01
4.82306898e-01 -1.19672847e+00 3.60559106e-01 3.74488682e-01
1.05005801e+00 -1.24168074e+00 -3.31702791e-02 2.09503230e-02
-6.58479691e-01 1.11818981e+00 4.10789043e-01 -2.48300597e-01
4.45205957e-01 -3.76609117e-01 -7.93957189e-02 1.06117718e-01
-7.09345460e-01 -1.13887928e-01 5.71175098e-01 7.20188498e-01
1.69097558e-01 -2.17117846e-01 1.42435944e-02 4.84238386e-01
2.71269232e-01 5.53927198e-02 1.37177095e-01 1.04936564e+00
-3.59213740e-01 -1.50258768e+00 -5.87992743e-02 5.59009075e-01
6.25688732e-02 -1.24917395e-01 -4.24023777e-01 7.92884469e-01
-1.12715051e-01 7.53150344e-01 1.01790279e-02 -1.12724356e-01
4.88074213e-01 3.12028646e-01 6.70924246e-01 -8.03215384e-01
-5.05094886e-01 -8.31895843e-02 -3.00011814e-01 -4.93158281e-01
-2.72515118e-01 -5.41466892e-01 -5.13072610e-01 -2.84483582e-01
-3.12466890e-01 -4.79844883e-02 5.28498471e-01 8.13084304e-01
4.22913969e-01 2.47079387e-01 8.50928843e-01 -1.18173754e+00
-7.41879523e-01 -9.25028861e-01 -5.17945945e-01 7.29147851e-01
2.89475709e-01 -8.43292832e-01 -3.61674488e-01 -1.21451169e-01] | [11.589080810546875, 0.6728093028068542] |
45e993a1-5e38-4af3-ad09-698e048258ac | marginmatch-improving-semi-supervised | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Sosea_MarginMatch_Improving_Semi-Supervised_Learning_with_Pseudo-Margins_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Sosea_MarginMatch_Improving_Semi-Supervised_Learning_with_Pseudo-Margins_CVPR_2023_paper.pdf | MarginMatch: Improving Semi-Supervised Learning with Pseudo-Margins | We introduce MarginMatch, a new SSL approach combining consistency regularization and pseudo-labeling, with its main novelty arising from the use of unlabeled data training dynamics to measure pseudo-label quality. Instead of using only the model's confidence on an unlabeled example at an arbitrary iteration to decide if the example should be masked or not, MarginMatch also analyzes the behavior of the model on the pseudo-labeled examples as the training progresses, ensuring low fluctuations in the model's predictions from one iteration to another. MarginMatch brings substantial improvements on four vision benchmarks in low data regimes and on two large-scale datasets, emphasizing the importance of enforcing high-quality pseudo-labels. Notably, we obtain an improvement in error rate over the state-of-the-art of 3.25% on CIFAR-100 with only 25 examples per class and of 4.19% on STL-10 using as few as 4 examples per class. | ['Cornelia Caragea', 'Tiberiu Sosea'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['pseudo-label'] | ['miscellaneous'] | [ 4.24869508e-02 2.76960939e-01 -4.18081403e-01 -7.27747202e-01
-1.06325495e+00 -4.90089089e-01 7.73249745e-01 7.45927989e-02
-6.03079259e-01 8.53174090e-01 -2.69669026e-01 -2.28319820e-02
1.46096110e-01 -6.22964129e-02 -9.07633126e-01 -7.82199562e-01
1.18464187e-01 6.64366782e-01 3.66494536e-01 3.18681657e-01
-5.70791438e-02 1.41775146e-01 -1.25930285e+00 1.47174373e-01
6.45444512e-01 1.06539106e+00 -2.24723816e-01 4.33164150e-01
2.20041409e-01 1.03821266e+00 -3.69908124e-01 -3.29122514e-01
5.04156172e-01 -3.83524030e-01 -6.33905888e-01 4.65714425e-01
9.90644276e-01 -1.81831375e-01 -3.11927915e-01 1.24671209e+00
7.39746243e-02 1.13242462e-01 6.95702672e-01 -1.15682733e+00
-7.61732161e-01 5.29528558e-01 -6.14876807e-01 2.58977175e-01
-3.97579730e-01 4.92422760e-01 1.02258873e+00 -9.39459503e-01
7.39061475e-01 9.40193355e-01 6.89236462e-01 7.62150466e-01
-1.71196187e+00 -6.36290073e-01 5.13796449e-01 1.91648975e-02
-1.30025554e+00 -5.84193826e-01 3.47029626e-01 -7.67720163e-01
5.41990578e-01 -2.56118864e-01 2.69512743e-01 1.13551891e+00
-1.05083603e-02 5.80440819e-01 1.51056767e+00 -3.75076562e-01
5.83842874e-01 2.92043388e-01 8.15516293e-01 7.47749686e-01
2.80003577e-01 5.03302991e-01 -5.07674038e-01 -1.46589994e-01
5.93906105e-01 -3.21039796e-01 -4.61241081e-02 -3.30579877e-01
-1.01390493e+00 6.39098942e-01 5.62380672e-01 -4.25581373e-02
-1.78725690e-01 3.02185327e-01 1.44954205e-01 2.36034974e-01
6.82020605e-01 1.87260017e-01 -5.33060372e-01 1.04084738e-01
-1.01086748e+00 1.49911806e-01 3.07205081e-01 6.93132520e-01
7.13479340e-01 1.47243440e-01 -5.22503495e-01 8.15056920e-01
4.22834337e-01 5.80762029e-01 2.61616707e-01 -1.29522932e+00
2.10620448e-01 4.27528828e-01 4.80607212e-01 -3.21022779e-01
-2.06507280e-01 -9.73292351e-01 -4.82441127e-01 4.67932135e-01
8.12632620e-01 -1.80474773e-01 -1.45576680e+00 2.11654830e+00
3.00846159e-01 5.55478990e-01 -1.79212838e-02 9.72395599e-01
5.08069396e-01 3.46855760e-01 2.41954863e-01 -3.32827508e-01
8.98829579e-01 -1.33836281e+00 -3.77286494e-01 -6.34363711e-01
5.73515475e-01 -4.31095064e-01 1.17154956e+00 4.45392698e-01
-9.00262773e-01 -7.23269939e-01 -1.05995739e+00 1.77964553e-01
-1.01086972e-02 1.95151910e-01 3.97528678e-01 2.79958874e-01
-1.00349331e+00 7.99629569e-01 -8.57830644e-01 9.36924368e-02
6.99936807e-01 1.58087552e-01 -3.38736653e-01 -1.51799187e-01
-8.72367203e-01 7.30658412e-01 1.45371258e-01 -8.97993743e-02
-1.04545820e+00 -9.70992565e-01 -4.94650275e-01 -8.09360072e-02
2.43003085e-01 -3.54789019e-01 1.46382952e+00 -1.25754690e+00
-1.08444715e+00 1.07770216e+00 -2.49325737e-01 -7.80282676e-01
8.17687094e-01 -2.50287890e-01 -5.26935339e-01 -8.84997994e-02
2.78486669e-01 1.01660383e+00 1.00277209e+00 -1.43946338e+00
-7.85583079e-01 -3.81253481e-01 -2.12391853e-01 4.89472486e-02
1.32368937e-01 -4.92325634e-01 -5.97340524e-01 -3.30919713e-01
1.69704393e-01 -1.29978120e+00 -2.61828780e-01 2.34069139e-01
-3.45824003e-01 -1.76542029e-01 5.55635452e-01 -1.88668579e-01
8.14554095e-01 -2.20072508e+00 -2.72713482e-01 -1.34845108e-01
2.63713300e-01 5.80101132e-01 -3.00600231e-01 -2.95407593e-01
-9.20724049e-02 1.07479475e-01 -3.91685992e-01 -6.60951078e-01
-1.33301646e-01 2.29171753e-01 -3.00230622e-01 6.05978549e-01
4.36376721e-01 8.80295217e-01 -1.01528966e+00 -1.74826413e-01
1.06389120e-01 3.42423409e-01 -3.70774120e-01 -5.11556603e-02
-4.21699554e-01 6.47172391e-01 -1.81854382e-01 3.18729788e-01
6.78910017e-01 -9.61108387e-01 -1.47074284e-02 -3.02679501e-02
1.98275775e-01 2.08642349e-01 -1.01106465e+00 1.40637505e+00
-1.40530452e-01 6.69258654e-01 -1.92862466e-01 -6.50110185e-01
5.75678885e-01 2.12766483e-01 2.84448773e-01 -7.75571406e-01
-4.38861214e-02 1.47824675e-01 4.01543602e-02 -1.35323882e-01
6.95594102e-02 -5.62426485e-02 3.30760628e-01 4.20666426e-01
9.52883586e-02 3.96875352e-01 2.97589988e-01 2.24003613e-01
9.80660081e-01 1.32695168e-01 -6.79336190e-02 -3.44368070e-01
2.94045001e-01 6.32037446e-02 7.65839279e-01 1.00057709e+00
-5.97159684e-01 6.52873635e-01 3.78408790e-01 -4.43700939e-01
-1.10052216e+00 -1.30181515e+00 -5.17666757e-01 8.79769742e-01
7.72920325e-02 -4.26512174e-02 -5.95113277e-01 -1.14900851e+00
2.52963245e-01 7.75908291e-01 -9.09399569e-01 -2.40807503e-01
-1.59338504e-01 -8.34642649e-01 1.44037545e-01 3.65546674e-01
5.40376127e-01 -7.97455609e-01 -2.23543659e-01 1.79632977e-01
3.12466528e-02 -1.28920662e+00 -4.13362116e-01 3.91588867e-01
-8.84731829e-01 -1.09802544e+00 -3.36518198e-01 -5.57014048e-01
1.01558030e+00 6.72090650e-02 1.32732093e+00 -4.07545380e-02
-6.02098517e-02 1.39323816e-01 -1.88148782e-01 4.31453437e-02
-5.56360543e-01 -6.31206781e-02 1.30146310e-01 1.42690524e-01
2.94672012e-01 -2.13061914e-01 -6.05461061e-01 4.13753688e-01
-5.18880606e-01 -1.15963779e-01 2.59232730e-01 1.15874815e+00
9.22127128e-01 -3.23665231e-01 7.52230823e-01 -1.38291574e+00
4.85356757e-03 -4.45377737e-01 -9.00415003e-01 3.83761436e-01
-1.36861420e+00 2.86774248e-01 5.77759027e-01 -6.51284575e-01
-1.01527655e+00 1.90175861e-01 2.53514886e-01 -7.50346899e-01
-9.23373401e-02 1.26154482e-01 4.98778850e-01 2.42766235e-02
9.65806723e-01 -6.79239631e-02 -6.89274222e-02 -4.30661470e-01
3.98213923e-01 3.71632695e-01 6.94960177e-01 -4.72428083e-01
7.11753488e-01 5.67832410e-01 -2.32285149e-02 -2.81968176e-01
-1.72106326e+00 -4.25786376e-01 -4.91252363e-01 -1.69177249e-01
6.83923066e-01 -1.19444847e+00 -4.50145453e-01 5.03201365e-01
-7.43809402e-01 -7.69708455e-01 -4.79204118e-01 4.00775254e-01
-2.66788840e-01 1.30128369e-01 -8.07726264e-01 -6.68959439e-01
-1.63812533e-01 -1.10361624e+00 9.35850918e-01 3.18452328e-01
-3.49999741e-02 -1.00374043e+00 8.13088417e-02 4.46146011e-01
1.60986662e-01 -5.24438173e-02 7.06740141e-01 -7.81739771e-01
-6.02536321e-01 -1.89872503e-01 -3.69359881e-01 6.48664713e-01
4.96377386e-02 7.77642578e-02 -1.39818668e+00 -5.49643338e-01
-9.36532468e-02 -8.53556931e-01 9.34732020e-01 4.72436756e-01
7.96687722e-01 -8.18441734e-02 -2.51367062e-01 3.78523290e-01
1.54061389e+00 -1.37419850e-01 3.42023134e-01 1.82315819e-02
4.94571954e-01 3.82949203e-01 8.28303814e-01 1.33923858e-01
8.78287926e-02 5.67218661e-01 4.53759044e-01 1.60662569e-02
-4.16002750e-01 -3.27254415e-01 8.81614089e-02 4.69644964e-01
4.41244662e-01 -1.53494984e-01 -7.00218558e-01 3.63583148e-01
-1.96247327e+00 -7.26099372e-01 -2.05887035e-01 2.51581502e+00
8.64865839e-01 5.96954525e-01 -2.02527910e-01 -2.57500976e-01
7.00293005e-01 1.37761608e-01 -1.06030798e+00 2.19527900e-01
-1.14855662e-01 -5.66696450e-02 7.02293217e-01 7.94824719e-01
-1.23760808e+00 1.14346218e+00 6.75890303e+00 7.16365814e-01
-1.30146170e+00 1.95641682e-01 1.31273687e+00 -1.61085382e-01
-2.57206857e-02 3.40078771e-02 -1.00741374e+00 5.60686529e-01
1.05398989e+00 1.79339066e-01 4.24842477e-01 9.54702020e-01
1.32584393e-01 -4.15809840e-01 -1.34431672e+00 9.32929873e-01
-1.71593815e-01 -1.34191191e+00 -2.34206751e-01 -8.08117762e-02
1.18775225e+00 7.48050213e-01 2.81552911e-01 3.14373344e-01
5.92177093e-01 -8.73889089e-01 9.70708907e-01 4.26830500e-01
7.65738785e-01 -2.91395009e-01 4.57149684e-01 4.45190281e-01
-6.74615204e-01 1.51206031e-01 -3.59161347e-01 5.98880760e-02
-1.05100006e-01 7.70874500e-01 -7.60787070e-01 5.63122192e-03
5.52678764e-01 8.30730796e-01 -7.88212955e-01 8.60315263e-01
-4.51575726e-01 1.10658765e+00 -2.52348721e-01 4.58083242e-01
3.49169195e-01 -1.74649745e-01 2.63576120e-01 9.78056729e-01
-2.23995835e-01 -1.15092948e-01 4.54088628e-01 9.39852893e-01
-2.36513421e-01 -3.22214156e-01 -2.25746676e-01 2.51657069e-01
6.77631021e-01 1.19153035e+00 -6.12989485e-01 -5.75348437e-01
-2.15329513e-01 7.81208038e-01 7.35089839e-01 5.70904851e-01
-8.34562123e-01 3.74612212e-01 4.72639143e-01 7.53354505e-02
3.76836002e-01 -4.66800220e-02 -3.57279301e-01 -1.06929457e+00
1.15820602e-01 -6.42409325e-01 1.30145714e-01 -8.92780662e-01
-1.51731348e+00 7.44501829e-01 -2.98962891e-01 -1.18135715e+00
-2.34578758e-01 -5.45337141e-01 -2.85690218e-01 8.11522782e-01
-1.69958878e+00 -7.83027947e-01 -1.61254212e-01 1.71955794e-01
5.40731370e-01 6.48754910e-02 6.62017047e-01 1.04217172e-01
-6.48680449e-01 7.21292317e-01 3.34145963e-01 1.23921759e-01
9.75358725e-01 -1.25005507e+00 5.52083671e-01 9.45908248e-01
5.47098219e-01 3.08971763e-01 6.56785071e-01 -5.15790999e-01
-6.42975748e-01 -1.22838354e+00 7.82313228e-01 -6.87344193e-01
6.19809568e-01 -2.37656876e-01 -1.13230419e+00 8.40244293e-01
-4.30197306e-02 8.61304224e-01 3.61078471e-01 1.41011383e-02
-8.51666033e-01 -1.47290975e-01 -1.28868914e+00 5.77690005e-01
1.07948697e+00 -4.84826863e-01 -2.96772569e-01 5.41145802e-01
6.31604731e-01 -4.12369251e-01 -6.47880137e-01 5.89832366e-01
3.03376615e-01 -1.21507227e+00 7.21363962e-01 -7.06553757e-01
2.56178826e-01 -4.17194158e-01 -1.34700850e-01 -1.24931931e+00
-5.88570893e-01 -3.57005686e-01 -5.40908426e-02 1.10083318e+00
8.42462361e-01 -6.77437484e-01 1.19208801e+00 7.55661905e-01
6.65621459e-02 -8.51098835e-01 -8.49561751e-01 -9.38105881e-01
2.06135046e-02 -3.93476307e-01 -3.87908774e-03 9.81028557e-01
-4.32038784e-01 4.92432266e-01 -3.11368167e-01 2.47408420e-01
1.01044834e+00 6.22184854e-03 4.05026942e-01 -1.22739339e+00
-5.68844914e-01 -1.97151780e-01 -1.13686927e-01 -9.55221295e-01
3.42053801e-01 -8.72030020e-01 1.29669696e-01 -1.01691628e+00
2.93329179e-01 -7.28139997e-01 -6.24069870e-01 6.87276900e-01
-4.65667933e-01 6.00564659e-01 1.63078532e-01 5.66841781e-01
-8.72548103e-01 3.86650354e-01 1.14217663e+00 -1.39710248e-01
7.56825581e-02 -1.11604324e-02 -4.76843268e-01 6.00832939e-01
5.92373371e-01 -7.58341134e-01 -5.98493099e-01 -4.55381393e-01
-9.04455408e-02 -1.32485792e-01 4.58271533e-01 -1.09749794e+00
1.14891957e-02 -6.55650422e-02 3.61490250e-01 -2.73020923e-01
3.40095073e-01 -7.41547525e-01 6.78710714e-02 4.63257760e-01
-8.52576375e-01 -3.12717378e-01 2.08772033e-01 9.16455805e-01
-6.52249530e-02 -2.18233392e-01 1.38892555e+00 1.32996649e-01
-6.65243506e-01 3.42336357e-01 2.43953429e-02 5.79808414e-01
9.19449985e-01 1.70084491e-01 -5.72741687e-01 -4.68903081e-03
-1.01245070e+00 3.83531541e-01 6.85457587e-01 1.80553362e-01
1.13803387e-01 -1.24080086e+00 -5.89430749e-01 1.47536486e-01
2.93290794e-01 1.01865465e-02 1.60917372e-01 5.97458780e-01
-7.50721097e-02 3.38744879e-01 6.48214221e-02 -9.95947063e-01
-9.60166097e-01 4.71893162e-01 5.90603173e-01 -4.25564498e-01
-4.32427526e-01 9.22176123e-01 1.24057174e-01 -2.05352858e-01
3.83978486e-01 -1.81267947e-01 3.15171331e-01 -3.01385403e-01
3.73052180e-01 1.82204515e-01 1.57719642e-01 -4.48113978e-01
-2.69789785e-01 5.43812573e-01 -4.68011498e-01 -1.94064349e-01
9.84195590e-01 -9.45156142e-02 2.87093729e-01 9.01536763e-01
1.08506536e+00 -3.06707740e-01 -2.25001979e+00 -5.71671605e-01
1.69869184e-01 -4.00144368e-01 2.16819689e-01 -1.26798880e+00
-1.09065151e+00 5.26187241e-01 8.32952559e-01 -1.28577784e-01
5.36901712e-01 2.40270682e-02 3.79009873e-01 3.75217408e-01
3.33266854e-01 -1.08206582e+00 1.90266788e-01 3.81347626e-01
6.08865201e-01 -1.81448698e+00 -3.74932550e-02 -2.32412204e-01
-9.11498070e-01 4.83277678e-01 7.59050548e-01 -3.55245858e-01
7.62900949e-01 8.28235224e-03 4.78190392e-01 -1.02601416e-01
-1.04640055e+00 1.07846394e-01 3.92131627e-01 2.59915352e-01
2.69248426e-01 -3.95415956e-03 1.29518822e-01 1.38475671e-01
3.46026927e-01 2.53250688e-01 3.19517791e-01 6.39386415e-01
-4.12456423e-01 -1.00133419e+00 5.72711639e-02 6.37671709e-01
-3.20371360e-01 -9.80157629e-02 -9.57346633e-02 7.05517769e-01
5.08689098e-02 9.18896198e-01 1.34355009e-01 -1.68945417e-01
7.84972757e-02 3.64965141e-01 4.15857136e-01 -8.54540586e-01
-4.65489179e-01 1.03629999e-01 -6.85781986e-02 -5.39225101e-01
-7.15151966e-01 -6.45517409e-01 -1.31804466e+00 1.44908587e-02
-3.11733037e-01 1.28308028e-01 3.70111674e-01 9.71867740e-01
5.51037848e-01 3.33108842e-01 7.23324776e-01 -4.86267000e-01
-9.72424805e-01 -9.35624659e-01 -6.05036378e-01 7.19126105e-01
5.24647832e-01 -6.01009965e-01 -6.36760116e-01 2.03484237e-01] | [9.454885482788086, 3.6236863136291504] |
e7e41f12-a110-4bd9-bb73-a6ac592dc76f | are-we-evaluating-paraphrase-generation | null | null | https://openreview.net/forum?id=YBujxHx3VmL | https://openreview.net/pdf?id=YBujxHx3VmL | Are We Evaluating Paraphrase Generation Accurately? | Paraphrase is a restatement of a text that conveys the same meaning using different expressions. The evaluation of paraphrase generation (PG) is a complex task and currently lacks a complete picture of the criteria and metrics. In this paper, we survey the automatic evaluation metrics and human evaluation criteria of PG evaluation. Base on the survey result, we propose a reference-free automatic toolkit and list clear human evaluation criteria. Moreover, we notice the paraphrases selection in downstream tasks and propose a simple but effective evaluation Filter model. It can fusion multi automatic metrics to fit the human evaluation without any references. | ['Anonymous'] | 2021-11-16 | null | null | null | acl-arr-november-2021-11 | ['paraphrase-generation', 'paraphrase-generation'] | ['computer-code', 'natural-language-processing'] | [ 4.89463881e-02 2.08153925e-03 -1.11832291e-01 -4.73155707e-01
-8.61118555e-01 -8.65835488e-01 7.91872144e-01 1.93141446e-01
-1.96514532e-01 7.93743372e-01 8.43366563e-01 -1.08488873e-01
-1.29990384e-01 -5.34086049e-01 -8.48956332e-02 -2.32692599e-01
8.86301517e-01 4.55078959e-01 -3.44034731e-02 -6.01855934e-01
9.10050869e-01 2.58763134e-01 -1.53324759e+00 9.23173368e-01
9.35356677e-01 6.08543754e-01 1.47066012e-01 7.60441244e-01
-3.07021797e-01 9.22465742e-01 -1.31166458e+00 -8.58595610e-01
-2.54070647e-02 -9.08583581e-01 -1.34679127e+00 -1.48034915e-01
3.36097598e-01 -1.08881379e-02 1.69021301e-02 1.01549602e+00
9.27268028e-01 -8.64336714e-02 7.85803735e-01 -1.49412739e+00
-9.37085986e-01 8.59396994e-01 6.80887476e-02 1.37999415e-01
1.19905889e+00 -8.51190370e-03 1.18832326e+00 -1.11540878e+00
8.61670494e-01 1.19635081e+00 5.37902951e-01 2.81301916e-01
-8.79865766e-01 -3.24114829e-01 -5.28025448e-01 6.32842600e-01
-1.11784148e+00 -3.34102899e-01 7.44811833e-01 -4.87517953e-01
1.12800074e+00 6.35648727e-01 6.86691523e-01 1.41502523e+00
-2.18638266e-03 8.67643654e-01 1.14102674e+00 -6.98131204e-01
2.25998694e-03 3.11915249e-01 7.19830453e-01 3.11153501e-01
-1.80664565e-02 -3.89333993e-01 -6.52403712e-01 -1.48510337e-01
1.52395785e-01 -3.21785837e-01 -3.92895430e-01 -8.44067428e-03
-1.30761576e+00 7.76873887e-01 -2.86476184e-02 4.98193473e-01
-1.08875306e-02 -3.61655742e-01 8.53216946e-01 8.85320842e-01
5.78672960e-02 9.76181984e-01 -1.01557903e-01 -4.00784433e-01
-1.25075316e+00 3.55044007e-01 1.01454210e+00 1.34505093e+00
4.46582586e-01 -4.77410674e-01 -9.75838900e-01 1.21003389e+00
-1.75807491e-01 4.12378609e-01 1.03759789e+00 -1.21775103e+00
3.28567654e-01 1.04483271e+00 2.62323529e-01 -9.00009155e-01
-7.74469599e-02 -3.82508993e-01 -6.12545311e-01 8.01857188e-02
1.49021531e-03 1.11591741e-01 -2.49886081e-01 1.40036416e+00
-2.58461714e-01 -7.31379271e-01 9.26419795e-02 8.64077687e-01
1.44388103e+00 5.51114380e-01 -2.31028408e-01 -2.02369645e-01
1.33304167e+00 -1.28044438e+00 -9.53411758e-01 1.38918489e-01
4.81602132e-01 -1.22163367e+00 1.51916182e+00 3.97721618e-01
-1.36691940e+00 -8.69340956e-01 -1.33026874e+00 -3.52646351e-01
-5.93217075e-01 5.53499877e-01 2.37576470e-01 4.97520477e-01
-1.11096239e+00 8.57770205e-01 1.09106138e-01 -8.25586736e-01
-3.26170743e-01 1.15616042e-02 -3.62970471e-01 3.50314826e-01
-1.52894199e+00 1.38909721e+00 4.06802297e-01 -2.54977793e-01
-2.93821156e-01 -3.46756160e-01 -6.02032721e-01 -4.98426892e-02
-1.16652176e-01 -1.21733153e+00 1.46236253e+00 -8.58005762e-01
-1.66365337e+00 1.31099284e+00 -2.76333570e-01 -2.54738480e-01
5.68728149e-01 -2.82245100e-01 -3.75342458e-01 2.50914007e-01
3.75927150e-01 3.43819141e-01 7.27989912e-01 -7.31716692e-01
-5.08717477e-01 8.87981057e-03 1.96926087e-01 6.69796109e-01
-2.36825213e-01 6.28990948e-01 -2.58061439e-01 -8.12983334e-01
-2.03145817e-01 -6.65875375e-01 1.75890505e-01 -4.36082751e-01
-3.05484414e-01 -4.82437462e-01 2.75186092e-01 -7.73863435e-01
1.90257108e+00 -1.75507748e+00 4.23880339e-01 -1.30134314e-01
1.71620354e-01 2.34844998e-01 7.47867227e-02 9.60289657e-01
-6.92732111e-02 2.45844603e-01 -1.79959983e-01 -2.21546292e-01
3.79875660e-01 -7.22396821e-02 -2.61049986e-01 -2.43665799e-01
-6.98501915e-02 1.11678052e+00 -1.19778264e+00 -8.66439104e-01
1.80737063e-01 -1.89121306e-01 5.18519506e-02 5.71599245e-01
3.91991526e-01 -1.24231227e-01 -2.40590110e-01 7.15222716e-01
3.63502502e-01 -2.75249809e-01 -2.53827780e-01 -3.80746514e-01
5.08811958e-02 3.87903720e-01 -8.78408611e-01 1.92421913e+00
-3.23691398e-01 8.04600060e-01 -3.71799916e-01 -6.85507715e-01
1.16636884e+00 2.76489556e-01 1.45943752e-02 -4.16878104e-01
4.85889502e-02 4.00018543e-01 -2.22476557e-01 -8.80591989e-01
8.57153296e-01 -1.40771940e-01 -3.54430437e-01 7.95436084e-01
2.23632112e-01 -6.56120837e-01 7.17739463e-01 5.48124671e-01
1.24213719e+00 8.21295604e-02 9.30427372e-01 -2.30146617e-01
1.05290484e+00 1.10469557e-01 9.40955654e-02 1.00351882e+00
-3.49428236e-01 9.15317357e-01 3.62790704e-01 -1.29558668e-01
-1.17539537e+00 -1.08317208e+00 5.79930060e-02 1.04394746e+00
1.63124666e-01 -1.02029622e+00 -8.85306239e-01 -7.94181347e-01
-1.81372896e-01 8.77599001e-01 -5.43533504e-01 -1.63811028e-01
-3.95889461e-01 -5.42020440e-01 8.68780136e-01 2.24039271e-01
5.49387634e-01 -1.52329803e+00 -4.92442667e-01 9.02378559e-02
-8.40823829e-01 -8.89570713e-01 -4.83044237e-01 -2.19475225e-01
-4.32894051e-01 -1.12955868e+00 -8.76835465e-01 -9.27605987e-01
2.37172753e-01 4.15573835e-01 1.59975457e+00 7.97617212e-02
6.11705333e-02 2.15894714e-01 -9.05642450e-01 -3.19796503e-01
-1.00287831e+00 1.79053441e-01 -3.44808042e-01 -5.87674558e-01
7.90746450e-01 -4.73607153e-01 -6.08070314e-01 3.71384144e-01
-5.79078674e-01 4.42149788e-01 5.42323172e-01 8.87185276e-01
1.09547347e-01 -6.43184900e-01 8.72084856e-01 -4.97592777e-01
1.86648369e+00 -3.26713145e-01 3.45615119e-01 7.80488551e-01
-7.37898171e-01 1.71095341e-01 5.91714382e-01 -4.35514376e-02
-8.70619118e-01 -4.64158267e-01 -1.17191002e-01 -9.70977023e-02
-3.94592807e-02 3.97928685e-01 1.24773607e-01 1.11147843e-01
9.66569126e-01 2.68736511e-01 -6.34112209e-02 -4.46892530e-01
5.70881426e-01 1.15290797e+00 5.47324300e-01 -3.84162396e-01
3.16079170e-01 -1.57576442e-01 -3.13153446e-01 -4.88472223e-01
-9.20939445e-01 -7.99770772e-01 -5.63078105e-01 -5.61366320e-01
4.68541056e-01 -6.51708305e-01 -2.35540897e-01 2.93099940e-01
-1.58859324e+00 1.86369166e-01 -4.00775135e-01 1.38276920e-01
-7.65146971e-01 9.39224601e-01 -7.06934810e-01 -2.83898652e-01
-1.11203909e+00 -1.07676756e+00 1.11876810e+00 8.66333172e-02
-1.12608600e+00 -6.08265400e-01 3.17815542e-01 5.68877459e-01
4.53513205e-01 8.31104517e-02 6.97102427e-01 -7.12866962e-01
2.20636666e-01 -2.60139465e-01 -3.61985654e-01 6.28568590e-01
9.13137421e-02 1.60698444e-01 -7.18578219e-01 1.39158651e-01
2.93855727e-01 -5.15720487e-01 5.33564150e-01 -3.95985730e-02
9.16456938e-01 -4.47442502e-01 4.63768747e-03 2.82976091e-01
8.62293124e-01 -9.63819921e-02 6.48816824e-01 4.89511728e-01
1.65986151e-01 8.57365012e-01 8.97137821e-01 3.33599031e-01
3.07916943e-02 6.83096409e-01 -2.84162611e-01 2.00331375e-01
-4.58292335e-01 -4.91193831e-01 5.53278923e-01 1.28554666e+00
1.91643275e-02 -3.65401179e-01 -4.88225847e-01 3.47132027e-01
-2.03836393e+00 -1.37586260e+00 -3.96609724e-01 1.86649954e+00
9.54971373e-01 5.85728465e-03 1.72478124e-01 3.83197129e-01
8.74365091e-01 3.25356089e-02 -1.50083095e-01 -1.02172041e+00
-4.95965153e-01 2.72809863e-01 -1.35926306e-01 5.54197729e-01
-5.23814082e-01 9.28944767e-01 7.62325859e+00 1.08818984e+00
-4.43591118e-01 -1.30749040e-03 1.70891494e-01 -1.21642957e-02
-3.11930388e-01 7.61917606e-03 -5.38493752e-01 6.86465323e-01
6.19570076e-01 -1.01773989e+00 1.71846628e-01 8.11886251e-01
3.59957337e-01 1.25853404e-01 -1.29440308e+00 1.37779450e+00
5.31582415e-01 -1.12440920e+00 5.67443132e-01 -5.42790592e-01
4.30332959e-01 -3.16327959e-01 -5.54956794e-01 5.94230115e-01
1.84308812e-01 -8.25444937e-01 6.43183708e-01 5.04179060e-01
6.42535269e-01 -4.66566145e-01 1.06981075e+00 2.38050014e-01
-8.49124193e-01 1.93976477e-01 -3.96416068e-01 -2.31547654e-01
4.02748615e-01 3.41234326e-01 -7.58314192e-01 7.19198465e-01
3.67671311e-01 7.43426025e-01 -1.21561182e+00 1.04634082e+00
-8.41060698e-01 2.53632128e-01 1.74955904e-01 -7.42815197e-01
-1.61976146e-03 -3.47525209e-01 7.89341688e-01 1.96575737e+00
5.73252678e-01 -1.53279543e-01 -2.65941113e-01 9.07685816e-01
7.44338408e-02 7.23147213e-01 -7.07081914e-01 2.38535166e-01
4.60477650e-01 1.41131377e+00 -2.88758904e-01 -7.18678117e-01
-6.13338165e-02 1.55208910e+00 4.35259610e-01 1.58456564e-01
-8.56670439e-01 -7.91341722e-01 1.91893101e-01 -1.97699741e-01
-5.24836540e-01 3.14144790e-01 -5.80356300e-01 -1.26429558e+00
1.85458168e-01 -1.15341508e+00 4.81646985e-01 -1.50562274e+00
-1.50207448e+00 7.02531278e-01 3.60151380e-01 -1.61845016e+00
-5.58822632e-01 -5.36185622e-01 -9.57831264e-01 9.29660320e-01
-7.95362175e-01 -9.97496843e-01 -7.09079206e-01 3.13618630e-01
8.37310016e-01 -3.37843239e-01 8.53957415e-01 9.50646102e-02
-3.93371105e-01 7.45621443e-01 -2.36683264e-01 -9.29237604e-02
1.10513377e+00 -1.38280654e+00 4.85992134e-01 7.46774316e-01
-1.26662441e-02 9.10425186e-01 1.11264622e+00 -4.52098966e-01
-5.32818258e-01 -5.43083251e-01 1.61802471e+00 -7.35184789e-01
7.64371753e-01 1.63492039e-02 -7.07593322e-01 -3.00240666e-02
8.06588769e-01 -1.05964696e+00 8.65408301e-01 3.98930721e-03
-2.74963379e-01 1.80605412e-01 -1.11253834e+00 8.26189339e-01
1.35389793e+00 -6.57721579e-01 -1.41554475e+00 5.39367139e-01
8.08328331e-01 2.88064573e-02 -7.22367704e-01 4.70950127e-01
6.81303203e-01 -1.12804389e+00 8.51100802e-01 -6.17183089e-01
8.56408000e-01 -3.03270429e-01 -6.52295873e-02 -1.44827294e+00
-5.37847042e-01 -6.95972860e-01 -4.11676243e-02 1.39610171e+00
4.93484855e-01 -2.40803614e-01 4.86030012e-01 6.07967794e-01
-3.83222729e-01 -5.74385107e-01 -5.83238542e-01 -8.76002073e-01
-1.30187586e-01 -1.17872782e-01 6.06992185e-01 8.98622334e-01
1.00526786e+00 1.00778675e+00 -2.19323993e-01 -6.42591298e-01
3.73627573e-01 4.01742041e-01 8.09010029e-01 -7.42087781e-01
-3.02613109e-01 -9.90977645e-01 -4.12807345e-01 -8.53571534e-01
1.97207004e-01 -1.32916772e+00 -4.46967959e-01 -1.89676201e+00
7.21924007e-01 4.88444388e-01 -2.92544127e-01 1.18291795e-01
-3.66664082e-01 7.10095689e-02 2.49394596e-01 4.39181834e-01
-7.90345728e-01 5.01938879e-01 1.06656682e+00 -1.68851599e-01
-8.38094056e-02 -1.36885718e-01 -7.27169335e-01 6.64646268e-01
1.23661494e+00 -4.34630990e-01 -5.80217659e-01 -2.10700288e-01
6.05778396e-01 -8.69628266e-02 2.22220033e-01 -8.26032996e-01
2.22038940e-01 -2.18044743e-01 6.17045090e-02 -7.18445301e-01
-3.48738879e-02 -1.66339576e-01 2.65800476e-01 2.58599550e-01
-7.97167599e-01 7.32015491e-01 -3.03145468e-01 -7.97682330e-02
-5.20959735e-01 -1.19415081e+00 6.61083281e-01 -4.52526093e-01
-6.71202123e-01 -3.98233384e-01 -4.26029325e-01 2.41675138e-01
6.44909084e-01 -5.53105116e-01 -3.92630756e-01 -5.44751644e-01
-5.88934422e-01 2.41134092e-01 6.73126400e-01 6.27882183e-01
6.90691411e-01 -1.52784896e+00 -1.05717063e+00 -2.29866296e-01
6.87687159e-01 -7.76339114e-01 -1.35042578e-01 3.51987213e-01
-6.94196045e-01 2.82773852e-01 -5.20700157e-01 -2.16502715e-02
-1.55742538e+00 5.11512280e-01 2.21115842e-01 -5.04090071e-01
-2.93048888e-01 5.73674262e-01 -2.26291358e-01 -2.59867787e-01
1.57310013e-02 -3.97156663e-02 -7.93018758e-01 2.16087565e-01
7.02423930e-01 7.76980460e-01 1.47409827e-01 -5.86112916e-01
-6.42066225e-02 3.40498000e-01 1.02582902e-01 -4.19152588e-01
6.72022998e-01 -3.35029662e-01 -3.71568233e-01 5.21592140e-01
1.27763474e+00 -8.30140412e-02 4.64921482e-02 1.05881892e-01
1.02526054e-01 -3.74671876e-01 -5.35746932e-01 -1.14140773e+00
-1.90398559e-01 6.99405253e-01 2.35024005e-01 4.55172181e-01
9.57864165e-01 -1.51744276e-01 7.14132369e-01 6.25992119e-01
4.68942165e-01 -1.57921898e+00 1.04528867e-01 5.13553560e-01
1.77005708e+00 -1.14953637e+00 1.84375793e-01 -3.98972273e-01
-8.75870705e-01 1.21580267e+00 5.77934146e-01 -1.48223927e-02
1.18764453e-01 -2.73823619e-01 -8.71619135e-02 -2.26144746e-01
-7.77944684e-01 -1.12980485e-01 4.51344281e-01 4.45289880e-01
8.11673641e-01 -1.15478979e-02 -1.38799322e+00 9.07506049e-01
-8.42689216e-01 1.02130629e-01 5.16791463e-01 5.71042240e-01
-5.38072288e-01 -1.22023678e+00 -9.50081646e-02 5.30116558e-01
-3.20202261e-01 -2.24000067e-01 -1.25903952e+00 4.42914814e-01
-1.43052563e-01 1.35661399e+00 -6.56730235e-01 -5.12849748e-01
7.36263037e-01 2.84293830e-01 5.41820824e-01 -6.58537328e-01
-1.11080074e+00 -6.27213597e-01 5.85486948e-01 -3.59877110e-01
-4.86069590e-01 -3.95392358e-01 -6.72360957e-01 -3.30604315e-01
-1.60286471e-01 4.88600343e-01 4.60981131e-01 6.19057596e-01
3.87645036e-01 7.88980722e-02 7.43770003e-01 -6.06578350e-01
-7.87884891e-01 -1.38596058e+00 -2.83321857e-01 1.16245079e+00
-4.17521775e-01 -2.64908552e-01 -7.63100326e-01 1.11053482e-01] | [11.513242721557617, 9.132925987243652] |
66c86086-879e-4abc-93ab-46f4745d5919 | 190807844 | 1908.07844 | null | https://arxiv.org/abs/1908.07844v1 | https://arxiv.org/pdf/1908.07844v1.pdf | Similarity Learning for Authorship Verification in Social Media | Authorship verification tries to answer the question if two documents with unknown authors were written by the same author or not. A range of successful technical approaches has been proposed for this task, many of which are based on traditional linguistic features such as n-grams. These algorithms achieve good results for certain types of written documents like books and novels. Forensic authorship verification for social media, however, is a much more challenging task since messages tend to be relatively short, with a large variety of different genres and topics. At this point, traditional methods based on features like n-grams have had limited success. In this work, we propose a new neural network topology for similarity learning that significantly improves the performance on the author verification task with such challenging data sets. | ['Dorothea Kolossa', 'Steffen Zeiler', 'Robert M. Nickel', 'Benedikt Boenninghoff'] | 2019-08-20 | null | null | null | null | ['authorship-verification'] | ['natural-language-processing'] | [-2.17764169e-01 -5.00617862e-01 -2.07550168e-01 -3.56244117e-01
-4.91471946e-01 -6.52480245e-01 8.89860868e-01 4.98031706e-01
-2.79388994e-01 6.32373810e-01 1.84496436e-02 -2.61222422e-01
1.63698979e-02 -6.31623328e-01 -1.74431428e-01 -4.54491109e-01
2.60661721e-01 5.39160728e-01 2.90460497e-01 -1.22260973e-01
8.84476185e-01 4.89753157e-01 -1.46091425e+00 5.47717586e-02
5.89750051e-01 8.10641110e-01 -1.83349028e-01 5.08211732e-01
-7.43901968e-01 6.48752570e-01 -7.70903111e-01 -7.56645858e-01
9.57712010e-02 -4.87832040e-01 -6.62249148e-01 -3.57437789e-01
6.90713465e-01 -6.60058707e-02 -3.37981582e-01 1.15607190e+00
3.60840201e-01 -1.30165279e-01 8.36607933e-01 -1.34457362e+00
-7.55535483e-01 8.86849701e-01 -6.32955611e-01 2.57654727e-01
4.31038350e-01 -3.75938267e-01 1.07316697e+00 -6.08487248e-01
4.39005673e-01 1.23215008e+00 9.66341436e-01 3.39645177e-01
-8.34098697e-01 -1.09903336e+00 -2.85794705e-01 4.52704310e-01
-1.31222999e+00 -1.73186406e-01 8.72461259e-01 -5.83560765e-01
2.87344217e-01 8.57477859e-02 2.29751706e-01 1.18959165e+00
1.87814519e-01 5.03011227e-01 1.34731829e+00 -5.02123773e-01
-4.81836498e-02 5.32931149e-01 6.15827739e-01 4.87168103e-01
5.57558775e-01 -4.28749889e-01 -4.59220082e-01 -3.20752978e-01
1.50805473e-01 1.65339217e-01 -2.10983098e-01 9.89281237e-02
-1.12857378e+00 1.05818212e+00 -1.87920898e-01 8.27633917e-01
3.67506072e-02 -2.13233739e-01 7.33476937e-01 4.74150389e-01
2.00205579e-01 7.55013585e-01 -1.14841715e-01 -1.50326774e-01
-1.32782710e+00 3.42713475e-01 1.11576772e+00 7.16759562e-01
4.79196310e-01 -3.00349504e-01 -4.72500138e-02 7.77658403e-01
2.94741452e-01 4.05819178e-01 8.13815653e-01 -6.05974078e-01
4.54853803e-01 6.89048469e-01 -1.50890006e-02 -1.74027717e+00
-2.13616148e-01 -3.45078468e-01 -9.57386971e-01 2.85185245e-03
9.11243677e-01 8.10203701e-02 -4.66235042e-01 1.23498976e+00
1.55975059e-01 -9.56732258e-02 -4.39169928e-02 7.32452869e-01
1.16120863e+00 5.79648256e-01 -1.70588523e-01 -1.10872015e-02
1.34420919e+00 -7.07959354e-01 -8.40625584e-01 2.52493143e-01
2.68415838e-01 -1.03724086e+00 6.57135308e-01 4.91720736e-01
-7.53373682e-01 -4.69041437e-01 -8.02282453e-01 -7.76211321e-02
-7.20803559e-01 -1.05468007e-02 3.98972481e-01 9.50362742e-01
-6.60698175e-01 9.06704247e-01 -1.30526289e-01 -4.11028773e-01
4.77177411e-01 1.65120900e-01 -3.09438586e-01 1.96782261e-01
-1.13663781e+00 6.89268291e-01 3.47320586e-01 -2.44978517e-01
-1.79635152e-01 -3.63561183e-01 -3.32609564e-01 1.08224392e-01
2.93896407e-01 -4.00975160e-02 9.94488418e-01 -9.60207105e-01
-1.43618715e+00 1.02273595e+00 -3.74018739e-04 -3.31571162e-01
6.55891240e-01 1.43463433e-01 -5.97860575e-01 6.81499541e-02
6.38546944e-02 -1.90808792e-02 1.04093599e+00 -1.02432156e+00
-6.52279675e-01 -5.22365868e-01 -3.18693280e-01 -4.62771595e-01
-9.52586472e-01 6.41498268e-01 1.90072004e-02 -9.29435253e-01
1.21243887e-01 -8.02698970e-01 2.00165346e-01 1.90928616e-02
-4.00349528e-01 -7.23365247e-01 8.87635708e-01 -9.31461990e-01
1.36746287e+00 -1.91840291e+00 -1.47538275e-01 3.39317024e-01
4.80408967e-01 5.42623460e-01 1.42990962e-01 6.63818538e-01
1.23094395e-01 3.70929956e-01 3.77410948e-02 -4.20386821e-01
4.45436537e-02 -1.83686465e-01 -5.54195464e-01 3.95876676e-01
-3.23197126e-01 4.69748884e-01 -6.39533043e-01 -7.81005204e-01
-5.11172593e-01 1.88807815e-01 1.84980348e-01 2.98738867e-01
8.48885179e-02 2.71967858e-01 -4.06332523e-01 7.65440464e-01
6.43935442e-01 -3.67743224e-01 1.67821348e-01 1.17077179e-01
-9.05756280e-02 1.22213759e-01 -1.18969190e+00 1.11202800e+00
-7.29758069e-02 9.75194871e-01 -1.08773567e-01 -9.74234700e-01
1.11498737e+00 2.98893720e-01 2.61600405e-01 -4.10289675e-01
5.05167663e-01 5.45061648e-01 1.63217500e-01 -6.34877801e-01
6.74211800e-01 -1.36592537e-01 -5.53253368e-02 8.60171378e-01
-2.00214893e-01 2.21999183e-01 4.46769923e-01 2.07941100e-01
1.07077408e+00 -2.57496715e-01 3.15878421e-01 4.57211435e-02
7.73218870e-01 -3.36647242e-01 5.85800529e-01 9.19016242e-01
-2.68959582e-01 8.02425683e-01 6.07925177e-01 -2.71767527e-01
-1.09778285e+00 -5.49197137e-01 -2.49671355e-01 6.36885762e-01
9.18521956e-02 -5.11097074e-01 -6.85719371e-01 -8.67828965e-01
2.96747565e-01 2.22941726e-01 -3.19999307e-01 1.72130540e-01
-4.48691070e-01 -3.78078043e-01 9.92120504e-01 3.18759114e-01
3.84745866e-01 -9.69123065e-01 -3.21083874e-01 1.15565978e-01
-4.90034372e-02 -1.15855372e+00 -4.45308715e-01 -6.28794134e-02
-7.90268064e-01 -1.28558791e+00 -9.86573458e-01 -8.42180073e-01
5.38589656e-01 2.64904052e-01 7.47779906e-01 4.07946646e-01
-1.07934430e-01 -5.75252436e-02 -6.02085352e-01 -2.83811212e-01
-6.12476289e-01 2.39434227e-01 1.76781163e-01 3.86215776e-01
5.50186038e-01 -4.69014198e-01 7.65868202e-02 2.87969261e-01
-8.26169133e-01 -5.59843540e-01 4.08254802e-01 9.64796245e-01
-8.60359594e-02 2.83599734e-01 5.67926347e-01 -9.52041626e-01
8.89672101e-01 -5.70776761e-01 -3.75532180e-01 3.38157505e-01
-8.11892390e-01 -6.23908499e-03 1.17779076e+00 -8.04447114e-01
-6.30698919e-01 -5.48313975e-01 5.07086851e-02 -3.62909496e-01
-3.93045098e-01 5.23652971e-01 4.61283214e-02 -2.94849783e-01
3.11319351e-01 2.61272430e-01 1.32112697e-01 -7.45574176e-01
-3.36358696e-01 1.33278573e+00 3.24263275e-01 -3.71428132e-01
1.03303289e+00 2.70252138e-01 -1.23310070e-02 -8.39569747e-01
-5.86879313e-01 -6.39240861e-01 -6.06346190e-01 -5.43109514e-02
5.25498152e-01 -4.00656790e-01 -9.03015435e-01 9.08832967e-01
-1.30141449e+00 2.88850039e-01 2.97932208e-01 5.16520739e-01
1.18013404e-01 1.07865536e+00 -5.80261469e-01 -8.88061106e-01
-3.54759127e-01 -8.12390387e-01 4.86063272e-01 3.80465835e-01
-5.04525006e-01 -9.87719834e-01 1.94774717e-02 4.47251886e-01
5.70790410e-01 1.27632450e-02 1.00362253e+00 -1.20614624e+00
-2.75274575e-01 -6.72701895e-01 -3.94666940e-01 3.67086858e-01
2.13151842e-01 2.81294525e-01 -6.74452841e-01 -2.72960156e-01
-1.13199905e-01 -2.48398334e-01 7.85568058e-01 -2.56593108e-01
1.32451260e+00 -3.61092120e-01 -3.54330152e-01 4.52468157e-01
1.19482195e+00 1.91123173e-01 4.66351032e-01 5.56448162e-01
7.16326892e-01 7.19734132e-01 4.65430558e-01 5.36036134e-01
2.30662510e-01 6.08817577e-01 1.47755802e-01 3.87678444e-01
1.14502512e-01 -5.60950302e-02 1.41917378e-01 8.45602274e-01
-2.45021954e-02 -6.21022224e-01 -1.11072171e+00 4.48765367e-01
-1.82060695e+00 -1.28401339e+00 -4.54133749e-01 2.12012982e+00
7.49279022e-01 1.34120911e-01 2.87288934e-01 5.78146636e-01
1.02753782e+00 2.44957983e-01 -8.29925686e-02 -4.25175548e-01
-2.96161026e-01 -6.91019073e-02 2.00516045e-01 1.75579451e-02
-9.82512414e-01 5.92137218e-01 5.88841963e+00 8.61082971e-01
-1.28847992e+00 2.94208564e-02 3.72096777e-01 3.26705009e-01
-1.01650298e-01 9.59038436e-02 -1.20558298e+00 1.19123209e+00
8.70792806e-01 -4.91435677e-02 9.93635580e-02 6.16051376e-01
-6.22304305e-02 -3.25026624e-02 -9.41986382e-01 1.29505062e+00
5.65160096e-01 -1.21218467e+00 -2.10668966e-02 7.34920353e-02
5.44667482e-01 -4.69876975e-01 -1.79472696e-02 5.25322184e-02
-1.38838023e-01 -9.82427955e-01 7.01531351e-01 5.80586791e-01
4.49454188e-01 -6.19657040e-01 1.01638210e+00 3.96182895e-01
-5.64907610e-01 -2.62757093e-01 -3.31986099e-01 -4.90552858e-02
5.79742789e-02 7.33515024e-01 -5.51597476e-01 2.21834630e-01
6.97037160e-01 8.32891524e-01 -7.32588530e-01 1.25215352e+00
-1.51737854e-01 6.99149907e-01 -2.30924398e-01 -5.62895656e-01
5.52773252e-02 -3.23681653e-01 6.09323084e-01 1.31646049e+00
4.20366466e-01 -1.99620396e-01 -1.24179386e-01 7.67902195e-01
-3.27856034e-01 4.26733255e-01 -6.25340164e-01 -5.88319361e-01
3.54649484e-01 1.34525347e+00 -9.18597519e-01 -3.50482970e-01
-4.29316700e-01 8.14175308e-01 3.02680939e-01 -6.69552684e-02
-8.13263953e-01 -7.09743977e-01 3.31434757e-01 4.47352439e-01
1.96749151e-01 -3.31099540e-01 -3.12342763e-01 -1.11678207e+00
2.05740929e-01 -1.11065030e+00 5.20757139e-01 -3.84207159e-01
-1.79109418e+00 4.59645599e-01 -4.84078854e-01 -1.34800041e+00
1.84072256e-02 -7.54529357e-01 -7.67702758e-01 8.13741505e-01
-1.42589772e+00 -9.68787491e-01 -3.55532616e-01 3.17166686e-01
2.76695490e-01 -8.27662051e-01 5.44323623e-01 4.99634802e-01
-6.33443236e-01 9.48821604e-01 6.23009205e-01 5.64945638e-01
1.16345239e+00 -1.13336289e+00 -3.24386694e-02 6.73671722e-01
1.02467887e-01 8.29772949e-01 6.36029303e-01 -7.85112977e-01
-1.33896363e+00 -5.91729760e-01 1.49168205e+00 -4.26862508e-01
8.47780764e-01 -2.08380267e-01 -1.13537514e+00 2.00936839e-01
1.85835719e-01 -1.83288783e-01 8.53799999e-01 2.82444328e-01
-6.08203232e-01 -8.97927433e-02 -1.11819732e+00 3.56470734e-01
6.28615022e-01 -5.64870059e-01 -6.27187788e-01 3.40826243e-01
-7.26348907e-02 7.87152536e-03 -8.72565866e-01 3.07927635e-02
8.13907921e-01 -7.55018294e-01 7.19402134e-01 -7.66813040e-01
8.24173987e-01 -3.01193208e-01 6.75732121e-02 -8.66306543e-01
-1.97438911e-01 -4.91035402e-01 -2.07701936e-01 1.67040813e+00
4.48237471e-02 -7.62769878e-01 7.83573806e-01 3.03753078e-01
4.25441206e-01 -3.88257414e-01 -7.73388147e-01 -1.08235586e+00
9.67150405e-02 7.00884908e-02 5.76963127e-01 1.38727832e+00
9.86738428e-02 4.15903717e-01 -4.71772850e-01 -2.53172278e-01
7.44185686e-01 4.27670270e-01 8.49698484e-01 -1.80675340e+00
-2.35521689e-01 -7.64175594e-01 -6.30877137e-01 -4.96936262e-01
4.95454133e-01 -1.00739980e+00 -2.26953819e-01 -1.06888318e+00
3.79746318e-01 -6.38232172e-01 3.66238654e-02 3.42104673e-01
-1.48704603e-01 3.29034358e-01 4.22771201e-02 5.64467132e-01
-4.24112529e-01 1.94152370e-01 7.53255665e-01 -4.27769780e-01
-1.54444622e-02 1.04761176e-01 -4.91336137e-01 6.81405962e-01
9.01663184e-01 -6.44866943e-01 2.67303556e-01 -1.17424287e-01
2.94727743e-01 -1.59901097e-01 2.49146596e-01 -9.04990375e-01
4.40757334e-01 -1.42610222e-01 1.74681857e-01 -5.95150292e-01
-2.03453237e-03 -7.91220665e-01 8.11264366e-02 3.30292344e-01
-3.48129690e-01 2.92484373e-01 -3.03560019e-01 5.40616810e-01
-5.26979029e-01 -6.80987239e-01 8.34837258e-01 -6.50662035e-02
-4.66502368e-01 2.10638091e-01 -3.82012814e-01 8.48593265e-02
8.51856172e-01 -4.43600684e-01 -3.07419330e-01 -3.13687503e-01
-3.20276290e-01 -7.72524551e-02 5.12590408e-01 4.94343221e-01
4.32569295e-01 -1.12223458e+00 -8.03449035e-01 -8.08989257e-02
5.61898462e-02 -6.66015387e-01 -2.70692080e-01 8.70532513e-01
-6.05439126e-01 3.60844970e-01 -3.83451641e-01 -2.81412840e-01
-1.75542402e+00 5.16900480e-01 4.22032662e-02 -2.17789963e-01
-3.65010709e-01 4.67209518e-01 -5.01316607e-01 -3.29767078e-01
4.13038939e-01 2.11305097e-01 -6.86046600e-01 7.40730107e-01
6.75612748e-01 6.70058012e-01 2.00987943e-02 -9.82770085e-01
-2.86584854e-01 7.39395738e-01 -2.93058634e-01 1.89804658e-01
1.31266510e+00 8.24461803e-02 -3.83605719e-01 4.49149519e-01
1.34929001e+00 3.26041520e-01 -4.63826776e-01 -3.85326624e-01
2.51969129e-01 -8.11954439e-01 -1.01255320e-01 -4.71573979e-01
-1.04718709e+00 1.05506516e+00 3.31688404e-01 5.37595809e-01
5.36217868e-01 -1.49197578e-01 1.03927720e+00 5.28297961e-01
4.30366814e-01 -1.44753218e+00 1.60226688e-01 7.11954355e-01
6.15400314e-01 -1.51740921e+00 1.14685580e-01 -1.65153384e-01
-1.84418231e-01 1.66103435e+00 4.21400070e-01 1.61588773e-01
6.34248912e-01 2.09666602e-02 9.27121043e-02 -7.70683289e-02
3.33742611e-02 1.33421332e-01 2.38162816e-01 2.15565309e-01
5.55862248e-01 -2.32736245e-01 -8.63408804e-01 6.06032193e-01
-1.68759197e-01 -1.31523326e-01 7.80780613e-01 8.12890053e-01
-2.14077368e-01 -1.40830564e+00 -6.58259273e-01 7.36219764e-01
-8.25557530e-01 3.00015453e-02 -7.66966164e-01 4.37044412e-01
8.73484686e-02 8.88006330e-01 -2.13047173e-02 -3.44341606e-01
-1.88555300e-01 1.53274253e-01 2.86592990e-01 -4.81857389e-01
-1.02397025e+00 -5.01407743e-01 -4.67785187e-02 6.58491626e-02
-4.00786281e-01 -1.11212564e+00 -1.02391040e+00 -8.17432284e-01
-3.85189325e-01 2.99208879e-01 7.10478544e-01 9.56342697e-01
6.26498386e-02 -4.45867032e-02 7.92794406e-01 -4.48731035e-01
-7.66434014e-01 -8.47443759e-01 -8.41771901e-01 6.32313728e-01
1.34537876e-01 -6.36971116e-01 -5.52627563e-01 -4.63922843e-02] | [9.557761192321777, 10.595133781433105] |
935a3ee6-3135-4aa2-819f-3ebebf8fc3c9 | efficient-semantic-segmentation-on-edge | 2212.13691 | null | https://arxiv.org/abs/2212.13691v2 | https://arxiv.org/pdf/2212.13691v2.pdf | Efficient Semantic Segmentation on Edge Devices | Semantic segmentation works on the computer vision algorithm for assigning each pixel of an image into a class. The task of semantic segmentation should be performed with both accuracy and efficiency. Most of the existing deep FCNs yield to heavy computations and these networks are very power hungry, unsuitable for real-time applications on portable devices. This project analyzes current semantic segmentation models to explore the feasibility of applying these models for emergency response during catastrophic events. We compare the performance of real-time semantic segmentation models with non-real-time counterparts constrained by aerial images under oppositional settings. Furthermore, we train several models on the Flood-Net dataset, containing UAV images captured after Hurricane Harvey, and benchmark their execution on special classes such as flooded buildings vs. non-flooded buildings or flooded roads vs. non-flooded roads. In this project, we developed a real-time UNet based model and deployed that network on Jetson AGX Xavier module. | ['Maryam Rahnemoonfar', 'Ting Zhu', 'Guanqun Song', 'Venkatesh Dasari', 'Irfan Ali', 'Farshad Safavi'] | 2022-12-28 | null | null | null | null | ['real-time-semantic-segmentation'] | ['computer-vision'] | [ 2.49538407e-01 1.84324175e-01 3.60702276e-01 -5.77473462e-01
-8.49263147e-02 -8.83406878e-01 4.38617289e-01 -1.18308648e-01
-6.52902901e-01 6.61141396e-01 -2.41764992e-01 -6.08179629e-01
-2.14715973e-01 -1.42735314e+00 -4.84026283e-01 -5.81824481e-01
-3.50593716e-01 1.03113842e+00 5.87377906e-01 -3.47209126e-01
-1.50176197e-01 8.15225482e-01 -1.60730314e+00 3.39627385e-01
4.10022408e-01 1.06391323e+00 3.29067260e-01 8.59503090e-01
-1.63436178e-02 4.84171480e-01 -7.35783815e-01 3.21529984e-01
7.78015137e-01 -1.52460963e-01 -1.05974329e+00 -2.48959865e-02
3.06374371e-01 -2.57088542e-01 -1.26582086e-01 9.55097377e-01
4.43534970e-01 2.86853194e-01 5.80094635e-01 -1.62287688e+00
6.53210655e-02 7.72470176e-01 -8.63138139e-02 4.68295217e-01
1.93755701e-02 2.61153907e-01 5.82394958e-01 -2.43420720e-01
4.97532874e-01 1.29428911e+00 9.33285236e-01 2.03187838e-01
-7.69292414e-01 -5.45928121e-01 1.04215905e-01 1.63199738e-01
-1.25074458e+00 -6.28414452e-02 1.87213004e-01 -3.57902735e-01
1.26110160e+00 3.76632363e-01 9.56015408e-01 6.06146932e-01
1.39328642e-02 4.32842910e-01 1.04536819e+00 -7.88086206e-02
7.22251773e-01 -3.98374975e-01 1.09079428e-01 6.08328104e-01
2.02059627e-01 -4.49716523e-02 2.86289986e-04 8.93908739e-02
4.49410766e-01 -2.95706354e-02 -3.32167566e-01 2.02543482e-01
-9.29193676e-01 1.02744949e+00 7.94929683e-01 3.71646047e-01
-5.35232782e-01 2.85178900e-01 3.78640205e-01 4.48646359e-02
5.01128852e-01 3.92589778e-01 -8.18985045e-01 1.98493093e-01
-1.64849246e+00 3.40672851e-01 8.95150065e-01 8.18674564e-01
7.96691418e-01 2.12996915e-01 1.73880860e-01 2.32139885e-01
2.98654258e-01 6.35043740e-01 9.10984948e-02 -1.00376773e+00
2.41213709e-01 6.07668340e-01 -4.13368829e-02 -9.07993734e-01
-1.11158884e+00 -2.12712526e-01 -7.25174665e-01 3.08432579e-01
2.46522367e-01 -6.31987274e-01 -1.44575858e+00 1.17915893e+00
2.23437428e-01 2.26174608e-01 2.92621583e-01 1.31360805e+00
7.94583976e-01 7.97533870e-01 5.96217513e-01 3.39482516e-01
1.23254704e+00 -6.23823881e-01 -2.58517504e-01 -4.20438945e-01
3.62321705e-01 -3.31370264e-01 7.15440333e-01 2.11124435e-01
-6.74312174e-01 -4.87486303e-01 -8.87718618e-01 3.31306130e-01
-1.04737532e+00 9.53837391e-03 8.50568414e-01 7.29772151e-01
-1.27113473e+00 6.29182756e-01 -9.55395818e-01 -8.57665241e-01
5.61834633e-01 5.47051430e-01 3.09172124e-02 2.28446767e-01
-1.29328144e+00 8.22348893e-01 6.65098310e-01 3.53684008e-01
-1.37717438e+00 -5.36270320e-01 -7.17262983e-01 1.58132792e-01
2.37960830e-01 -4.58000541e-01 1.02756476e+00 -1.06490159e+00
-8.74851644e-01 9.50154305e-01 5.73641598e-01 -1.02657616e+00
7.03376591e-01 -1.11175045e-01 -1.95126310e-01 5.02973199e-01
3.27798814e-01 1.28537643e+00 4.30228293e-01 -1.17807269e+00
-1.07956636e+00 -1.88635319e-01 5.06522894e-01 2.04689145e-01
1.81177735e-01 -2.52251774e-01 1.49643123e-01 -2.94178605e-01
6.19849563e-02 -9.25570428e-01 -4.16875899e-01 -8.06772113e-02
-4.28555220e-01 3.01581994e-02 1.49706578e+00 -5.65930009e-01
5.54575980e-01 -1.81117439e+00 -1.44465506e-01 2.69529343e-01
-1.87691674e-01 3.49277347e-01 3.45519511e-03 2.98072785e-01
3.17956731e-02 3.31863314e-01 -9.24164534e-01 3.13214391e-01
1.81325935e-02 6.49995804e-01 -3.33855271e-01 4.35084969e-01
-1.45565427e-03 5.11355281e-01 -7.27446496e-01 -4.87452656e-01
4.40518737e-01 3.32896531e-01 -2.74073154e-01 9.07718837e-02
-4.91040349e-01 3.02901059e-01 -5.66407681e-01 8.76860499e-01
9.27219570e-01 3.40421498e-01 1.49425030e-01 -1.79888040e-01
-3.38213474e-01 -4.84494448e-01 -9.00564015e-01 1.44634712e+00
-2.57257611e-01 9.26666975e-01 5.44250488e-01 -1.41254723e+00
8.57737005e-01 2.17914581e-01 6.08550251e-01 -5.24370372e-01
4.50752765e-01 -6.17108941e-02 -5.07393360e-01 -6.45648301e-01
6.46818340e-01 -1.26880705e-01 -2.18495369e-01 3.34858537e-01
2.59106457e-02 -7.55568504e-01 2.53907204e-01 -5.09686247e-02
1.08133352e+00 3.39761734e-01 -3.31236303e-01 -9.22973156e-01
2.02382743e-01 9.55820739e-01 3.41836780e-01 6.49729371e-01
-3.24751288e-01 7.88305998e-01 2.05432996e-01 -9.25885499e-01
-8.99069130e-01 -9.22512412e-01 -1.16190448e-01 1.02788174e+00
4.01977807e-01 1.81375101e-01 -1.37639165e+00 -6.80580974e-01
-2.53405213e-01 8.86672735e-01 -3.72963309e-01 1.50204375e-01
-4.64273393e-01 -1.37770343e+00 9.52217042e-01 3.49568725e-01
1.17846453e+00 -1.25004768e+00 -1.80845475e+00 2.47291565e-01
-1.84741601e-01 -1.36324155e+00 4.35868084e-01 4.24621701e-01
-6.87076688e-01 -1.46120429e+00 -4.34351593e-01 -6.97432637e-01
5.59627831e-01 1.71733811e-01 1.25469553e+00 1.93756610e-01
-5.97417712e-01 6.18701100e-01 -5.62813044e-01 -6.55553401e-01
1.54657722e-01 2.20869988e-01 -1.93139568e-01 -3.31693977e-01
3.48659337e-01 -3.53809625e-01 -8.19269717e-01 3.98858786e-01
-1.12514865e+00 -9.98035744e-02 1.09209836e-01 2.22152904e-01
3.96015018e-01 7.13490486e-01 2.56150901e-01 -5.00002921e-01
2.95733213e-01 -5.98556697e-01 -7.79214561e-01 1.92826778e-01
-9.40749422e-02 -4.75710422e-01 4.41529095e-01 5.77004664e-02
-9.69253302e-01 6.48100019e-01 -1.26331776e-01 -4.64781933e-02
-5.92331767e-01 4.33247417e-01 3.94066088e-02 -5.19620217e-02
5.69031060e-01 -2.74876922e-01 -5.59041679e-01 -6.02781251e-02
2.89779127e-01 6.34475052e-01 7.84027755e-01 -4.32168692e-01
6.21497750e-01 9.14100111e-01 1.23462521e-01 -1.26146901e+00
-7.72338867e-01 -3.83987665e-01 -8.94376159e-01 -5.49614608e-01
1.41203892e+00 -9.21537519e-01 -1.72839418e-01 5.48498273e-01
-9.55073416e-01 -9.56685543e-01 -3.20821911e-01 2.47410610e-01
-4.58070636e-01 1.21434256e-01 -1.27008021e-01 -6.73263013e-01
-6.10154331e-01 -1.04994822e+00 1.16536915e+00 5.36342204e-01
6.29159361e-02 -1.02544928e+00 -1.84357002e-01 2.89333999e-01
5.37716925e-01 7.20817387e-01 5.55085659e-01 -4.89047587e-01
-5.23392975e-01 -4.23900411e-02 -3.92167717e-01 1.84496656e-01
-3.94182950e-01 1.20004177e-01 -9.91495311e-01 -1.13356508e-01
-2.72226185e-01 -2.97736913e-01 9.39313531e-01 6.27512395e-01
1.00573492e+00 -6.69324026e-02 -4.57030088e-01 6.81939721e-01
1.80060661e+00 2.98545420e-01 6.27675116e-01 5.32126665e-01
3.51118982e-01 6.92171216e-01 7.54096389e-01 3.35386276e-01
2.91584045e-01 2.62680978e-01 1.02928722e+00 -3.98872286e-01
2.53806282e-02 3.24470222e-01 -1.01690562e-02 -1.45850375e-01
-7.34249037e-03 -7.97228038e-01 -1.40987468e+00 8.80976677e-01
-1.82406867e+00 -7.63529778e-01 -5.55428743e-01 1.65380776e+00
5.95361851e-02 9.06850770e-02 -1.27944395e-01 8.75281319e-02
6.75290167e-01 -8.96377396e-03 -2.22335070e-01 -3.12455386e-01
-3.41156363e-01 5.26196599e-01 9.35389280e-01 4.43181753e-01
-1.56281829e+00 1.31490302e+00 6.50230551e+00 5.54769814e-01
-1.18193066e+00 3.12532157e-01 6.39703512e-01 9.63892788e-02
1.90165326e-01 2.19571978e-01 -4.76840287e-01 2.35795394e-01
1.10630810e+00 3.94444853e-01 1.53255031e-01 8.79894853e-01
3.79851550e-01 -6.95127964e-01 -4.12193835e-01 4.12673354e-01
-3.27397734e-01 -1.19328916e+00 -1.59043148e-01 -2.46777534e-01
5.18880665e-01 5.37676692e-01 -5.90779662e-01 2.12299287e-01
6.18795455e-01 -1.03667319e+00 9.11847293e-01 4.14364815e-01
6.05062425e-01 -7.15922117e-01 1.15326369e+00 2.70516604e-01
-1.24853563e+00 -6.07607737e-02 -6.34114563e-01 -1.85965240e-01
3.52818131e-01 4.31622088e-01 -6.99648738e-01 5.75851142e-01
1.21591043e+00 3.15291852e-01 -4.12546366e-01 9.57528591e-01
-2.28871286e-01 7.08533049e-01 -8.22169542e-01 2.79305816e-01
9.51583028e-01 -2.65476495e-01 3.76456231e-01 1.45225453e+00
4.13788795e-01 4.41230267e-01 6.63715839e-01 5.88796437e-01
4.82234746e-01 -3.81541103e-01 -6.02017343e-01 2.60744035e-01
1.45965353e-01 1.45219183e+00 -1.74068761e+00 -4.73584175e-01
1.43981382e-01 8.52972388e-01 -4.00853604e-01 2.93243140e-01
-1.14835811e+00 -5.08800268e-01 4.69662905e-01 7.07935840e-02
1.36660218e-01 -2.23158285e-01 -3.26363862e-01 -5.74937463e-01
-6.88636482e-01 -2.59763420e-01 4.52622414e-01 -1.12295866e+00
-8.01962137e-01 8.64756405e-01 3.72547984e-01 -7.29131162e-01
2.10353330e-01 -5.71114421e-01 -7.69319117e-01 5.47236502e-01
-1.54464054e+00 -1.43680227e+00 -9.42849159e-01 5.86229980e-01
6.93083465e-01 -1.75236702e-01 7.45786250e-01 1.37884215e-01
-3.62228155e-01 -3.38541597e-01 -4.20878053e-01 1.81826800e-01
-1.19436018e-01 -1.07278001e+00 4.17329282e-01 9.63851988e-01
-1.95179418e-01 -5.13568699e-01 7.76860535e-01 -7.66139209e-01
-7.03549922e-01 -1.62929583e+00 4.08532768e-01 -1.21729031e-01
3.71892691e-01 -2.20862076e-01 -5.70150852e-01 6.19454026e-01
3.52926403e-01 -1.08313635e-01 1.23198532e-01 -7.90759563e-01
3.29928577e-01 1.64397880e-01 -1.64549363e+00 1.98381841e-01
9.27805960e-01 -9.54635069e-02 -4.75702524e-01 8.08126748e-01
5.75912535e-01 -2.64359653e-01 -5.41126430e-01 5.90680540e-01
1.56894416e-01 -1.18084013e+00 9.37830508e-01 -2.36879811e-01
2.42444918e-01 -5.39232254e-01 -4.36178476e-01 -1.17660773e+00
-2.59332415e-02 -2.39655867e-01 8.11651766e-01 9.03483093e-01
2.25602627e-01 -5.00187039e-01 9.96038377e-01 3.28192621e-01
-4.85555202e-01 -8.47727060e-02 -1.02185845e+00 -6.97322905e-01
-4.08664569e-02 -4.81085002e-01 7.03477383e-01 9.12822306e-01
-6.06545210e-01 -9.38409567e-02 2.63581038e-01 5.66689134e-01
5.40582538e-01 9.81962159e-02 3.29380751e-01 -1.40214717e+00
3.70232761e-01 -2.45213121e-01 -6.13301337e-01 -2.15997428e-01
2.39697188e-01 -7.37640083e-01 1.87580004e-01 -2.01372313e+00
-3.40882450e-01 -6.35329962e-01 3.58427167e-01 1.10963750e+00
5.23616552e-01 6.61067486e-01 1.70522138e-01 -5.94220683e-02
-3.46064568e-01 2.27071866e-01 4.84190851e-01 -2.79646426e-01
1.05746894e-03 2.02385578e-02 -1.76996589e-02 9.38229442e-01
1.27651191e+00 -6.51169598e-01 -4.24017102e-01 -5.76078653e-01
8.90031680e-02 -1.74894497e-01 6.55536175e-01 -1.73328424e+00
1.89431965e-01 -1.48820892e-01 2.15383485e-01 -8.36245239e-01
1.17727034e-01 -1.28481400e+00 3.80006045e-01 7.01671243e-01
1.81865051e-01 1.35686144e-01 4.41575944e-01 1.41816616e-01
-2.50180393e-01 -4.19395536e-01 9.66752589e-01 -5.92427313e-01
-1.15526283e+00 2.10739404e-01 -9.22375321e-01 6.48187995e-02
1.47739601e+00 -4.24741447e-01 -3.82564485e-01 -1.96599245e-01
-8.46606433e-01 5.91120899e-01 3.67711693e-01 2.27658555e-01
4.28654701e-01 -4.89766240e-01 -5.70044756e-01 9.38593745e-02
-4.98351336e-01 4.95086133e-01 3.31168711e-01 5.14131725e-01
-1.54209423e+00 3.93232048e-01 -6.52695596e-01 -6.96439147e-01
-1.19430757e+00 1.67104632e-01 7.72678077e-01 -3.60791311e-02
-6.73397541e-01 7.19528615e-01 2.92065553e-02 -7.01041222e-01
2.11918186e-02 -4.48751271e-01 -3.42348218e-01 3.60810220e-01
5.33680394e-02 3.75609577e-01 1.47839293e-01 -6.68985307e-01
-4.10607666e-01 7.78017223e-01 8.92613888e-01 -2.08128363e-01
1.41658282e+00 -4.54391576e-02 -1.74238428e-01 -1.66181549e-01
6.25409245e-01 -9.17315662e-01 -1.32798827e+00 5.38465738e-01
7.23295137e-02 1.23712823e-01 6.03458107e-01 -1.05683076e+00
-1.75810564e+00 9.87636387e-01 9.69023407e-01 3.09557796e-01
1.41396952e+00 -2.80153096e-01 7.51328707e-01 5.36073267e-01
5.70405245e-01 -1.48226440e+00 -4.57921594e-01 6.18614495e-01
6.62507057e-01 -9.33183014e-01 2.49508277e-01 -3.14830571e-01
-5.59583068e-01 1.28604066e+00 4.54680294e-01 -2.93284535e-01
9.60263789e-01 3.29587340e-01 2.41777658e-01 -7.15555489e-01
-1.30335927e-01 -6.13198459e-01 -3.47828269e-01 9.64328885e-01
-3.50678474e-01 4.79242563e-01 1.81272645e-02 2.38874421e-01
-4.15593863e-01 4.99047013e-03 7.27690101e-01 1.19412434e+00
-8.24152172e-01 -4.07112479e-01 -5.33469200e-01 2.76711911e-01
-4.49186474e-01 -9.85325128e-02 -5.42855918e-01 9.82442796e-01
5.81584215e-01 1.27486849e+00 3.88330162e-01 -3.14655483e-01
3.81555021e-01 -1.16722755e-01 -6.40917867e-02 -4.19533551e-01
-1.06022644e+00 -2.17617959e-01 2.74662435e-01 -5.51240325e-01
-7.73862660e-01 -6.97305441e-01 -1.65987897e+00 -2.63456136e-01
1.19972892e-01 1.91842884e-01 1.14235616e+00 9.17236865e-01
2.58049443e-02 4.99532700e-01 2.40853965e-01 -1.37448454e+00
1.33372217e-01 -7.28997648e-01 -6.41138792e-01 5.67147881e-02
-3.99060175e-02 -4.71538514e-01 -3.79095972e-01 9.33675840e-02] | [9.2211332321167, -1.328510046005249] |
6cbf1698-c7fc-4395-92a6-12c28f2e7bd2 | noisy-parallel-approximate-decoding-for | 1605.03835 | null | http://arxiv.org/abs/1605.03835v1 | http://arxiv.org/pdf/1605.03835v1.pdf | Noisy Parallel Approximate Decoding for Conditional Recurrent Language Model | Recent advances in conditional recurrent language modelling have mainly
focused on network architectures (e.g., attention mechanism), learning
algorithms (e.g., scheduled sampling and sequence-level training) and novel
applications (e.g., image/video description generation, speech recognition,
etc.) On the other hand, we notice that decoding algorithms/strategies have not
been investigated as much, and it has become standard to use greedy or beam
search. In this paper, we propose a novel decoding strategy motivated by an
earlier observation that nonlinear hidden layers of a deep neural network
stretch the data manifold. The proposed strategy is embarrassingly
parallelizable without any communication overhead, while improving an existing
decoding algorithm. We extensively evaluate it with attention-based neural
machine translation on the task of En->Cz translation. | ['Kyunghyun Cho'] | 2016-05-12 | null | null | null | null | ['video-description'] | ['computer-vision'] | [ 5.92052460e-01 1.42743543e-01 -4.00631040e-01 -1.30923226e-01
-8.40838075e-01 -1.84928313e-01 7.54699945e-01 -2.46876091e-01
-3.79682511e-01 8.26940119e-01 2.59488314e-01 -7.17741191e-01
3.08426768e-01 -4.89437759e-01 -8.71072948e-01 -8.48387182e-01
2.58894920e-01 4.87915695e-01 -1.67463213e-01 -1.46365851e-01
3.68783206e-01 4.45976436e-01 -1.24842060e+00 1.72348440e-01
6.62049234e-01 6.40209973e-01 5.97803533e-01 7.72437930e-01
-2.30139539e-01 7.74148762e-01 -4.55299020e-01 -3.90714407e-01
-1.77956074e-01 -9.72435534e-01 -7.09021032e-01 1.67976752e-01
-6.47400022e-02 -8.64165053e-02 -3.52757365e-01 1.05044436e+00
5.77849209e-01 4.36144648e-03 5.76342642e-01 -7.44793892e-01
-9.50503230e-01 8.73524129e-01 -3.15967619e-01 1.66487262e-01
1.43033549e-01 -1.65477201e-01 8.95262361e-01 -1.13895178e+00
6.67412937e-01 1.07861936e+00 2.65613467e-01 8.83979499e-01
-1.17010367e+00 -5.66978872e-01 1.38224706e-01 3.89145344e-01
-1.34215999e+00 -8.00730705e-01 8.42482388e-01 -1.21413998e-01
1.07955360e+00 3.91823381e-01 4.82401758e-01 1.54845488e+00
4.27927047e-01 1.19764733e+00 7.79921174e-01 -7.41096735e-01
1.48108706e-01 1.57707006e-01 -2.62928426e-01 4.53330904e-01
-4.46058996e-02 4.62485361e-04 -5.59193790e-01 1.03432618e-01
7.95089006e-01 -1.70905486e-01 -3.51902008e-01 -2.35914260e-01
-1.32500088e+00 8.01449239e-01 1.61290959e-01 4.33251292e-01
-3.08604330e-01 2.94848561e-01 3.64444822e-01 5.12762725e-01
5.85299194e-01 2.92750150e-01 -3.18294585e-01 -3.41101855e-01
-1.09834385e+00 9.46982428e-02 6.02489769e-01 1.21647751e+00
5.61941922e-01 3.84759128e-01 -3.32462639e-02 7.95367897e-01
2.94419587e-01 2.43614793e-01 8.16423535e-01 -5.63041627e-01
5.74738324e-01 5.51701859e-02 -5.37892245e-02 -7.73758650e-01
-3.85513991e-01 -5.68587124e-01 -1.31593287e+00 -2.78640032e-01
-1.06714986e-01 -2.74166554e-01 -8.98184538e-01 1.80148828e+00
-2.82902539e-01 2.63002515e-01 1.91486478e-01 9.12218034e-01
3.52952391e-01 8.01837862e-01 -1.89895928e-01 -5.70824802e-01
8.53474438e-01 -1.14080131e+00 -8.34353507e-01 -1.46094128e-01
7.66970694e-01 -8.28724504e-01 9.27198827e-01 3.73641461e-01
-1.45134187e+00 -3.73351157e-01 -8.68212402e-01 -9.24378261e-02
-1.10624313e-01 3.04276764e-01 5.99757016e-01 4.98759717e-01
-1.25663114e+00 5.93320072e-01 -1.08705544e+00 -4.28287655e-01
6.88355193e-02 4.37131137e-01 2.09009796e-02 7.90049657e-02
-1.14780593e+00 8.00887764e-01 4.78776634e-01 2.63110399e-01
-9.82575357e-01 -1.01612456e-01 -7.03454435e-01 2.23732337e-01
1.66097537e-01 -8.16120386e-01 1.35471916e+00 -1.42745864e+00
-2.11060405e+00 6.43305004e-01 -3.42864513e-01 -8.79894555e-01
4.48979765e-01 -1.75871179e-01 -3.55762303e-01 -8.07748884e-02
-3.54998946e-01 7.53121614e-01 1.03798056e+00 -9.98003304e-01
-4.12919849e-01 -7.60635138e-02 -1.21865354e-01 3.81242603e-01
-4.99793261e-01 3.25354367e-01 -3.10236335e-01 -9.74602580e-01
1.01703502e-01 -1.02770460e+00 -3.80931616e-01 -4.93828267e-01
-4.34480518e-01 -1.34600699e-01 6.32064164e-01 -6.12170100e-01
1.53649998e+00 -2.00572395e+00 7.33040154e-01 1.62325276e-03
2.52127182e-02 3.28497380e-01 -4.38414253e-02 6.63858473e-01
-1.49088293e-01 1.79664493e-01 -3.61160487e-01 -6.33675635e-01
-1.71388477e-01 2.14546204e-01 -5.27427793e-01 4.60171402e-01
1.10519730e-01 1.01334834e+00 -7.02943563e-01 -3.18887472e-01
3.26938182e-02 5.28721929e-01 -5.65155029e-01 1.82225242e-01
-4.44157600e-01 5.18967569e-01 -2.74875104e-01 4.84693557e-01
2.29366764e-01 -3.99597168e-01 2.65307128e-01 2.58186281e-01
-1.92499444e-01 6.45915508e-01 -6.62038863e-01 1.93018723e+00
-6.58532917e-01 9.47540820e-01 -1.60989478e-01 -1.32105267e+00
9.18263614e-01 6.82137668e-01 8.19633082e-02 -5.42982697e-01
2.45433718e-01 3.75609785e-01 1.13494433e-01 -3.33095610e-01
7.49918580e-01 -3.40629146e-02 1.61634058e-01 4.89330709e-01
-4.89131734e-02 2.84342021e-01 -1.49380207e-01 -2.25825217e-02
7.30291903e-01 2.61935592e-01 2.91073710e-01 -2.13239163e-01
3.81913424e-01 -3.25147398e-02 2.69559115e-01 6.26465142e-01
2.67141312e-01 7.90792286e-01 3.02739233e-01 -2.82678008e-02
-1.27112114e+00 -5.18983841e-01 1.36429906e-01 1.25803876e+00
-8.09364542e-02 -4.80274260e-01 -1.03459954e+00 -1.22394219e-01
-6.37731791e-01 8.11412096e-01 -4.33281273e-01 -3.87443095e-01
-9.22943294e-01 -7.39003539e-01 6.44937158e-01 4.84834135e-01
2.80310422e-01 -1.30549240e+00 -5.11716664e-01 6.06806576e-01
-1.10453181e-01 -9.51446176e-01 -4.54391927e-01 3.10913354e-01
-1.09613752e+00 -2.56766379e-01 -9.85359311e-01 -1.05636692e+00
6.29252672e-01 1.97227314e-01 1.06788051e+00 7.36832991e-02
2.05274701e-01 -4.09161747e-02 -4.07876611e-01 -4.71820496e-02
-5.90907335e-01 6.15205348e-01 8.14786255e-02 1.58000767e-01
1.38222381e-01 -5.29618204e-01 -4.83566463e-01 1.26363039e-01
-8.11452687e-01 5.46143770e-01 9.29017961e-01 1.09358096e+00
4.34917212e-01 -5.41171312e-01 6.82883739e-01 -7.35427856e-01
9.94608939e-01 -3.40627611e-01 -4.94528532e-01 4.16324884e-01
-6.25776470e-01 2.45946199e-01 7.59497762e-01 -5.21989107e-01
-8.50237906e-01 -1.96685016e-01 -4.42053199e-01 -5.19474089e-01
-4.94571887e-02 8.62204492e-01 5.54763451e-02 4.01541777e-02
3.58222067e-01 8.46714973e-01 -1.27648979e-01 -5.36949873e-01
3.10370237e-01 8.55999708e-01 4.17881787e-01 -3.88327986e-01
3.38776380e-01 5.21673188e-02 -2.18925104e-01 -7.86381185e-01
-3.87006879e-01 -6.94342256e-02 -4.08827454e-01 8.66929740e-02
8.29750359e-01 -8.32364082e-01 -4.27227587e-01 1.81017950e-01
-1.43464494e+00 -3.76476735e-01 1.23783410e-01 7.15677440e-01
-9.86116230e-01 2.59180784e-01 -8.65258634e-01 -7.69273937e-01
-4.70449090e-01 -1.29225588e+00 1.03898358e+00 -8.33724812e-02
-2.01478735e-01 -9.25308228e-01 -1.05166078e-01 -9.82908085e-02
5.44519663e-01 -1.81967810e-01 8.70560527e-01 -5.58118105e-01
-9.66978669e-01 1.08641796e-01 -2.45403331e-02 2.65278429e-01
-1.17811091e-01 -7.19962493e-02 -6.69766426e-01 -4.44912553e-01
1.28772780e-02 -2.58801967e-01 9.41844344e-01 4.04306769e-01
1.22678757e+00 -3.89826298e-01 -2.98145860e-01 6.51995301e-01
1.05109274e+00 4.02321041e-01 7.74018228e-01 1.71368331e-01
5.02478540e-01 3.21559429e-01 2.00985879e-01 2.82038271e-01
3.38278189e-02 8.11702490e-01 2.48995125e-01 6.86066076e-02
-1.46670967e-01 -2.21521750e-01 6.74073279e-01 1.32615829e+00
-2.43620723e-01 -5.91102958e-01 -8.37488949e-01 4.19867098e-01
-2.08648086e+00 -8.76184046e-01 2.55872577e-01 2.09222603e+00
7.10212886e-01 2.55616486e-01 -2.31554374e-01 -1.03582256e-02
6.72261417e-01 1.46288844e-02 -4.82024282e-01 -7.10944235e-01
-2.28287339e-01 2.77303547e-01 4.10764128e-01 4.96210337e-01
-6.83073759e-01 1.22111487e+00 5.95166397e+00 9.32185173e-01
-1.35012913e+00 2.65333742e-01 7.08948791e-01 -5.39489761e-02
-4.91978526e-01 -1.21535011e-01 -8.57199788e-01 2.84795672e-01
1.17204511e+00 -2.03396127e-01 7.02477038e-01 5.78369021e-01
2.60369599e-01 3.66442144e-01 -1.12195289e+00 1.16302764e+00
2.71229059e-01 -1.54471707e+00 2.92374611e-01 1.07450612e-01
6.97525263e-01 1.73880041e-01 1.76303282e-01 4.39067394e-01
-2.29715649e-03 -1.20463145e+00 8.07910204e-01 3.46838981e-01
8.52414787e-01 -8.11070085e-01 5.75763404e-01 6.40125871e-01
-8.37002873e-01 5.79210296e-02 -3.47224355e-01 -1.71834871e-01
3.45083296e-01 1.65361822e-01 -7.19499290e-01 4.65532422e-01
2.57635474e-01 7.69300282e-01 -1.55687004e-01 8.66980374e-01
-1.96248963e-01 8.96497428e-01 -7.94250220e-02 -3.59376162e-01
6.22709513e-01 -2.82191724e-01 5.94361663e-01 1.24664974e+00
6.51467741e-01 -2.06849519e-02 -1.07179262e-01 6.29104733e-01
-3.47474277e-01 8.81303623e-02 -5.97085714e-01 -2.35769436e-01
4.73384419e-03 7.72503197e-01 -6.03791416e-01 -4.40547168e-01
-5.59771299e-01 1.20299208e+00 4.34586495e-01 5.97877979e-01
-9.58235562e-01 -1.41164497e-01 6.48655891e-01 5.25813662e-02
3.73648196e-01 -5.36923587e-01 -3.66137207e-01 -1.40250111e+00
-4.60457616e-02 -8.05082500e-01 -1.69988304e-01 -7.75591195e-01
-6.62360549e-01 9.88380730e-01 -2.60510951e-01 -1.24963951e+00
-8.29123259e-01 -3.27477962e-01 -3.24696690e-01 8.44088078e-01
-1.43698967e+00 -7.33892977e-01 4.18145597e-01 5.78744352e-01
8.49512041e-01 -3.09507072e-01 9.36715841e-01 5.03940225e-01
-7.90345728e-01 6.91254973e-01 3.51976901e-01 3.77824232e-02
4.48808372e-01 -8.36122274e-01 4.67083305e-01 9.42888200e-01
4.05499279e-01 7.93420792e-01 6.96392119e-01 -2.93352991e-01
-1.56452131e+00 -1.11731255e+00 1.21326673e+00 2.34169811e-02
6.96232378e-01 -7.72775114e-01 -8.55591953e-01 9.51002121e-01
6.29373312e-01 -5.00693560e-01 3.66749644e-01 -4.11653742e-02
5.35672531e-02 1.21783510e-01 -4.77237761e-01 1.05434465e+00
1.15019739e+00 -5.80691159e-01 -2.99477965e-01 4.37832147e-01
8.66026461e-01 -6.20741665e-01 -3.82847190e-01 2.86978811e-01
4.66148734e-01 -6.76891744e-01 7.42213786e-01 -7.39605606e-01
4.92431641e-01 -3.80199663e-02 -2.18231708e-01 -1.23512244e+00
-2.41224036e-01 -1.01968825e+00 -3.92344654e-01 8.81827116e-01
7.15296388e-01 -4.80309814e-01 7.87377715e-01 1.45640433e-01
-5.61136425e-01 -9.83170569e-01 -9.84710872e-01 -6.25465393e-01
2.67704278e-02 -4.26633626e-01 3.20837051e-01 8.15872014e-01
-2.52810325e-02 4.61926550e-01 -9.09532309e-01 2.18437351e-02
1.72738299e-01 2.80293971e-01 4.74675179e-01 -6.79799378e-01
-5.31032622e-01 -7.31950939e-01 -1.12094417e-01 -1.68390810e+00
3.14861238e-01 -1.09184945e+00 5.00916131e-02 -1.44797230e+00
1.14113770e-01 -1.63058862e-01 -2.94318020e-01 2.62871534e-01
9.86479688e-03 5.82801737e-02 1.54598907e-01 3.84599715e-01
-4.99814123e-01 7.49497890e-01 1.10613036e+00 -6.63663670e-02
-1.68803915e-01 1.32415399e-01 -5.89358211e-01 3.73832613e-01
1.23806620e+00 -4.33187723e-01 -3.74474078e-01 -8.77835810e-01
5.12396097e-01 4.15938318e-01 2.26189762e-01 -7.83431530e-01
4.46814716e-01 9.05783400e-02 2.20861495e-03 -3.91735137e-01
3.11147183e-01 -6.37456894e-01 1.88057408e-01 5.41303992e-01
-6.66571915e-01 4.02216613e-01 -9.45687592e-02 5.28994799e-01
-4.25061315e-01 -2.28259489e-01 6.16457641e-01 -1.16593190e-01
-4.84303474e-01 2.39528611e-01 -6.52340055e-01 -3.49097610e-01
6.52229428e-01 -9.68530402e-02 1.15696527e-01 -5.21583676e-01
-9.12554264e-01 -1.34728953e-01 1.69425964e-01 5.54873705e-01
7.04260468e-01 -1.26423991e+00 -8.62932384e-01 3.68406355e-01
-3.16548869e-02 -1.52188122e-01 -2.13317752e-01 9.56375420e-01
-3.45278829e-01 7.60587037e-01 6.97838888e-02 -6.97013319e-01
-1.09912419e+00 5.40235519e-01 8.27061459e-02 -2.85263568e-01
-5.97926199e-01 7.00409055e-01 4.97682393e-02 -2.21089169e-01
2.73955792e-01 -4.92799193e-01 4.64182012e-02 -2.62434214e-01
3.97239178e-01 3.47040929e-02 1.89778537e-01 -4.51120526e-01
-1.91088647e-01 4.88290906e-01 -1.93878949e-01 -1.23678252e-01
1.13154626e+00 -2.38169745e-01 -9.78686661e-02 5.74123979e-01
1.19405460e+00 -2.97789246e-01 -8.27448070e-01 -2.62776703e-01
2.03478619e-01 4.76040989e-02 -5.33363372e-02 -4.79516536e-01
-8.63275945e-01 1.14563704e+00 3.23138237e-01 2.26746529e-01
1.23573685e+00 -4.42891568e-02 8.55897903e-01 5.57478309e-01
5.07866204e-01 -1.05192304e+00 -2.47151911e-01 7.55878747e-01
8.78086984e-01 -9.06428218e-01 -4.33426172e-01 6.46112068e-03
-4.30042863e-01 1.22381294e+00 2.57889450e-01 1.98591668e-02
5.01413763e-01 4.01830882e-01 -7.18553364e-02 3.01087379e-01
-1.22210371e+00 -9.38358083e-02 -3.54094105e-03 1.38784379e-01
7.64897525e-01 2.20984221e-01 -6.00179732e-01 2.70153821e-01
-4.98507954e-02 1.99257970e-01 4.95509684e-01 7.05876052e-01
-4.26241785e-01 -1.16447079e+00 -2.72122830e-01 2.60062426e-01
-4.84337598e-01 -6.17007732e-01 -1.63553104e-01 4.17622775e-01
-2.39749029e-01 6.83146298e-01 7.22762756e-03 -3.68355066e-01
-4.51119542e-02 1.70246646e-01 4.45563406e-01 -5.25271773e-01
-4.12827551e-01 4.05851841e-01 5.03762439e-02 -3.65828305e-01
-6.50032938e-01 -7.36975372e-01 -9.46280897e-01 -1.97007924e-01
-3.35293829e-01 2.09150821e-01 8.60236943e-01 8.45731437e-01
5.86943448e-01 4.72509086e-01 4.55165625e-01 -8.72022629e-01
-4.97135848e-01 -1.10064161e+00 -2.27186322e-01 -1.41335100e-01
2.80986130e-01 -2.39561304e-01 -7.96210915e-02 3.96304280e-02] | [10.983474731445312, 6.738154888153076] |
d70bbab7-2562-4122-8167-85f13270a802 | finding-a-needle-in-the-haystack-attention | 1811.08513 | null | https://arxiv.org/abs/1811.08513v2 | https://arxiv.org/pdf/1811.08513v2.pdf | Attention-Based Deep Neural Networks for Detection of Cancerous and Precancerous Esophagus Tissue on Histopathological Slides | Deep learning-based methods, such as the sliding window approach for cropped-image classification and heuristic aggregation for whole-slide inference, for analyzing histological patterns in high-resolution microscopy images have shown promising results. These approaches, however, require a laborious annotation process and are fragmented. This diagnostic study collected deidentified high-resolution histological images (N = 379) for training a new model composed of a convolutional neural network and a grid-based attention network, trainable without region-of-interest annotations. Histological images of patients who underwent endoscopic esophagus and gastroesophageal junction mucosal biopsy between January 1, 2016, and December 31, 2018, at Dartmouth-Hitchcock Medical Center (Lebanon, New Hampshire) were collected. The method achieved a mean accuracy of 0.83 in classifying 123 test images. These results were comparable with or better than the performance from the current state-of-the-art sliding window approach, which was trained with regions of interest. Results of this study suggest that the proposed attention-based deep neural network framework for Barrett esophagus and esophageal adenocarcinoma detection is important because it is based solely on tissue-level annotations, unlike existing methods that are based on regions of interest. This new model is expected to open avenues for applying deep learning to digital pathology. | ['Saeed Hassanpour', 'Bing Ren', 'Arief Suriawinata', 'Naofumi Tomita', 'Jason Wei', 'Behnaz Abdollahi'] | 2018-11-20 | null | null | null | null | ['medical-object-detection', 'crop-classification'] | ['computer-vision', 'miscellaneous'] | [ 2.42706522e-01 9.39764604e-02 -3.14642876e-01 -2.85892412e-02
-1.55660856e+00 -7.21505582e-01 1.52745232e-01 6.08535171e-01
-8.71222019e-01 6.70260191e-01 3.97729538e-02 -6.13124371e-01
-2.38062620e-01 -6.56220078e-01 -7.42443025e-01 -1.34132886e+00
-2.54024923e-01 2.23162845e-01 1.25443205e-01 2.57857680e-01
1.03367820e-01 4.36767578e-01 -1.08274865e+00 6.72965169e-01
6.11740410e-01 1.09975123e+00 3.12215269e-01 1.16733515e+00
3.10041070e-01 4.70289618e-01 -3.74237210e-01 -1.79879606e-01
-9.30696726e-02 -2.03488886e-01 -1.02888870e+00 -2.62968093e-01
3.83991957e-01 -3.97410542e-01 -6.47937730e-02 9.46933985e-01
4.94578242e-01 -1.90676674e-01 4.99538839e-01 -2.51930684e-01
-6.09177887e-01 3.71960908e-01 -4.41938937e-01 5.82651973e-01
-3.25608775e-02 1.74865112e-01 1.04831314e+00 -4.96164620e-01
9.22708094e-01 2.99971849e-01 1.25226820e+00 5.84506392e-01
-1.10547018e+00 -4.14093584e-01 -1.04750752e-01 2.69816548e-01
-1.23002756e+00 1.09914698e-01 1.45460904e-01 -4.03279126e-01
1.02477515e+00 4.22991872e-01 9.96104121e-01 8.76324713e-01
5.91182709e-01 5.92768371e-01 8.87144029e-01 -5.41855335e-01
1.74421057e-01 1.15645088e-01 -2.46507060e-02 8.74628961e-01
5.61367571e-01 -5.59903309e-02 1.90639913e-01 -3.38702708e-01
8.20226729e-01 2.04965785e-01 -4.42753673e-01 8.40528011e-02
-1.51182187e+00 9.10032213e-01 7.94925928e-01 9.13593292e-01
-3.33398730e-01 -1.30975947e-01 6.72593594e-01 -3.42819244e-02
5.90832531e-01 4.83030200e-01 -3.54172319e-01 3.75910640e-01
-1.20858407e+00 -2.43972197e-01 6.00923240e-01 3.32856208e-01
2.42682956e-02 -4.80790287e-01 -8.17232206e-02 7.07746029e-01
2.28830248e-01 -1.00089736e-01 9.80959892e-01 -5.86344600e-01
-3.13669145e-01 8.78718317e-01 -4.50018831e-02 -5.70444942e-01
-8.07184398e-01 -3.98461312e-01 -1.25191808e+00 3.82076263e-01
6.77173734e-01 -2.36569215e-02 -1.01693094e+00 1.11838377e+00
2.68249750e-01 1.85028967e-02 4.73871194e-02 1.10469341e+00
8.77358735e-01 2.97352046e-01 9.58225355e-02 1.24803275e-01
1.75422573e+00 -1.11384666e+00 -5.30877709e-01 5.56461811e-01
1.00018334e+00 -4.41018254e-01 8.66594851e-01 6.15434468e-01
-1.04880214e+00 -1.44364238e-01 -1.17304754e+00 -4.63169307e-01
-7.42481649e-01 4.71316934e-01 3.05241883e-01 6.40137076e-01
-1.48828006e+00 3.33357275e-01 -1.15607631e+00 -8.35663974e-01
7.27535367e-01 6.40496850e-01 -7.14104891e-01 2.19300732e-01
-7.90501535e-01 7.95971930e-01 1.86317071e-01 1.80067912e-01
-8.80465031e-01 -7.37716615e-01 -6.85478806e-01 2.18702346e-01
-1.88436538e-01 -9.49198127e-01 1.15279853e+00 -8.45232308e-01
-1.23590899e+00 1.39525318e+00 -1.20096013e-01 -6.58286273e-01
2.69637138e-01 3.28375518e-01 -5.21156602e-02 3.33852887e-01
-2.44986806e-02 5.02771914e-01 1.03097409e-01 -9.01901960e-01
-9.95429754e-01 -5.01004755e-01 -1.77531689e-02 -8.97895843e-02
-5.22917390e-01 -1.34640753e-01 -2.13972852e-01 -4.21293020e-01
-3.86014253e-01 -8.63622427e-01 -5.33384800e-01 3.85696054e-01
-3.82380426e-01 -1.83115005e-01 5.05529940e-01 -9.82993901e-01
1.13384175e+00 -2.02533960e+00 -1.77214146e-02 1.47143066e-01
2.88208127e-01 2.67848194e-01 1.63772359e-01 -1.64957670e-03
-1.28180370e-01 6.83457077e-01 -1.29255787e-01 -2.75884628e-01
-2.78473496e-01 -2.09152251e-01 3.86549085e-01 9.17793691e-01
6.51991665e-02 8.43769729e-01 -9.16074216e-01 -6.95221961e-01
3.37760389e-01 4.64799374e-01 -4.38933909e-01 1.07567310e-01
9.85379070e-02 3.42108965e-01 3.35714892e-02 1.05314064e+00
4.14192230e-01 -7.19677508e-01 2.85396010e-01 -3.70904505e-01
-4.10770416e-01 -1.28150865e-01 -3.59532624e-01 1.85552251e+00
-3.22823048e-01 6.59624457e-01 4.77019727e-01 -8.04565549e-01
2.29551241e-01 6.71673417e-01 5.70337057e-01 -2.47987747e-01
3.27775776e-01 1.76419601e-01 9.69513357e-02 -7.39860773e-01
1.41283244e-01 -1.62757754e-01 1.78689465e-01 2.02736735e-01
1.61837354e-01 2.38984764e-01 3.52113813e-01 -3.44849288e-01
1.27448165e+00 -1.91482425e-01 6.74430311e-01 -5.81162453e-01
5.86710691e-01 5.11471629e-01 3.40288728e-01 5.59965551e-01
-3.66118520e-01 7.29817927e-01 4.72825080e-01 -8.73369873e-01
-1.42818332e+00 -6.88964128e-01 -6.51542008e-01 7.79047728e-01
-1.18593290e-01 1.30047545e-01 -7.82043815e-01 -8.09087574e-01
-1.87870666e-01 2.18288437e-01 -1.24770331e+00 3.08761358e-01
-4.62624252e-01 -1.21088135e+00 8.66365075e-01 4.92977619e-01
2.58614540e-01 -8.82873893e-01 -7.36902773e-01 2.48886362e-01
-1.01236172e-01 -6.28741086e-01 -4.85580385e-01 2.10007772e-01
-8.37277710e-01 -1.45838940e+00 -1.39542305e+00 -1.19631207e+00
1.05148745e+00 -4.29319181e-02 7.86012590e-01 3.30694288e-01
-9.58295226e-01 1.13065183e-01 -2.02760145e-01 -5.30234277e-01
-4.64282155e-01 5.03362894e-01 -4.47608978e-01 -2.13767216e-01
4.02945280e-01 -6.37426376e-02 -1.25906634e+00 7.19544291e-02
-1.07831049e+00 -6.78088516e-02 1.12497663e+00 1.61000860e+00
9.83191013e-01 -4.11095545e-02 4.15994734e-01 -9.12972033e-01
4.46948886e-01 -5.39855659e-01 -4.48766172e-01 3.26090455e-01
-5.24951756e-01 -4.53401744e-01 6.72897041e-01 -3.32265884e-01
-7.93860853e-01 -3.78978066e-02 -3.93759817e-01 -2.81602681e-01
-3.16904187e-01 5.24103463e-01 7.33763993e-01 -2.03865364e-01
7.46432602e-01 9.39129591e-02 4.32015300e-01 -3.32816280e-02
-3.40754092e-01 8.02136600e-01 5.12487471e-01 -9.05076340e-02
6.25874102e-02 9.19997394e-01 -2.04590498e-03 -5.29634535e-01
-6.03670239e-01 -1.09377122e+00 -5.97209036e-01 5.62233701e-02
1.25809717e+00 -8.36809516e-01 -7.92953432e-01 4.82145607e-01
-6.67159677e-01 -5.88980913e-01 -6.74226284e-02 5.04927754e-01
-2.59766549e-01 2.72942305e-01 -1.16695762e+00 -6.03456736e-01
-9.74236131e-01 -9.83307481e-01 1.26149368e+00 2.84250200e-01
-4.77609366e-01 -1.47402740e+00 3.07144672e-01 1.44361779e-01
5.79353631e-01 5.84006906e-01 1.27752054e+00 -8.53537023e-01
-2.83465236e-01 -6.78452253e-01 -3.65441561e-01 3.28678973e-02
2.58669704e-01 3.88300389e-01 -8.52272451e-01 -6.08364761e-01
-2.58236416e-02 -1.28980130e-01 8.88843358e-01 7.73070633e-01
1.55163050e+00 -4.99592215e-01 -7.90066004e-01 7.24179447e-01
1.80649126e+00 3.52016211e-01 5.62947035e-01 7.88090885e-01
1.48735762e-01 2.18237311e-01 2.52740145e-01 2.33581886e-01
2.53541112e-01 3.01426947e-01 6.12459302e-01 -8.09538960e-01
-2.30467860e-02 2.79043347e-01 -2.66152889e-01 3.05521756e-01
-3.01974952e-01 -2.47338757e-01 -9.54002976e-01 1.07856524e+00
-1.41614830e+00 -9.43967521e-01 7.21585378e-02 2.08525157e+00
7.10892797e-01 -2.19143674e-01 3.99563685e-02 1.32660316e-02
7.41204560e-01 -2.16897532e-01 -5.44743001e-01 -3.18167001e-01
7.68522099e-02 1.44462198e-01 4.51424479e-01 1.57452479e-01
-1.21421361e+00 1.21804595e-01 6.11679983e+00 7.02874541e-01
-1.42204630e+00 1.26301438e-01 1.17437530e+00 8.15230235e-03
4.95744348e-02 -7.25828230e-01 -5.68458319e-01 3.57967108e-01
1.00201082e+00 -1.12800347e-02 1.35809213e-01 8.21214199e-01
4.37775031e-02 -1.84117004e-01 -1.01109779e+00 3.91295582e-01
4.36907560e-02 -1.67802346e+00 -1.82133973e-01 4.12604123e-01
8.17843676e-01 1.91940263e-01 2.56091356e-01 -2.21102946e-02
7.51738921e-02 -1.19350219e+00 -7.42224380e-02 6.31053090e-01
1.12243283e+00 -4.42050844e-01 1.66036785e+00 1.04057215e-01
-9.15863276e-01 -1.01294599e-01 -1.95904404e-01 4.96149153e-01
-2.97833711e-01 4.40039009e-01 -1.39038229e+00 4.12710667e-01
1.09173167e+00 3.36947471e-01 -3.27530921e-01 1.21823287e+00
4.79920685e-01 6.40899062e-01 -2.73673117e-01 -4.15579975e-01
3.71833712e-01 2.96894252e-01 2.35608056e-01 1.51677334e+00
7.45369613e-01 -1.92375168e-01 -4.16939855e-01 7.28526235e-01
-6.28603250e-02 2.69385427e-01 -3.64301801e-01 3.21286693e-02
1.49236694e-01 1.89276385e+00 -1.06445682e+00 -4.15179729e-01
-6.95422828e-01 5.67781389e-01 2.77132481e-01 1.03832364e-01
-8.18786383e-01 -3.20621431e-01 2.33781055e-01 3.00920188e-01
3.86595905e-01 4.53819692e-01 -5.26236773e-01 -8.10807884e-01
-5.35617888e-01 -8.18866372e-01 8.61742795e-01 -4.12566572e-01
-1.59678543e+00 6.65788472e-01 -4.46486324e-01 -1.16085160e+00
-9.69189033e-02 -1.04603064e+00 -7.95625865e-01 7.38092184e-01
-1.51476812e+00 -1.28659165e+00 -4.42433566e-01 3.23697776e-01
5.13091803e-01 9.65028256e-02 1.44653821e+00 7.16939420e-02
-5.32476902e-01 7.21978962e-01 5.71928799e-01 4.49104130e-01
7.71606565e-01 -1.73426259e+00 -1.96276858e-01 2.93528140e-01
-3.52903008e-01 7.98564434e-01 2.18213558e-01 -1.90875694e-01
-1.14563704e+00 -1.19007468e+00 6.64343119e-01 -2.27048650e-01
5.52228034e-01 7.47356042e-02 -8.64830673e-01 7.02909172e-01
5.51315010e-01 2.04904258e-01 1.47711301e+00 -7.59206340e-02
2.49291956e-01 3.35256495e-02 -1.69515908e+00 5.88390291e-01
2.87304461e-01 -2.34670639e-01 -2.36674905e-01 4.57636803e-01
8.36846530e-02 -7.78319478e-01 -1.49258137e+00 5.58136106e-01
8.79348338e-01 -8.05797756e-01 8.64755392e-01 -2.45567173e-01
4.05708522e-01 -3.21173482e-02 3.35469991e-01 -1.23972487e+00
-6.97754800e-01 -1.52951002e-01 3.83164108e-01 4.14558530e-01
6.98404551e-01 -6.82520330e-01 9.97542500e-01 4.33989257e-01
-5.09036422e-01 -1.25600612e+00 -1.11477256e+00 -1.40458852e-01
5.07123291e-01 4.79524076e-01 2.64571041e-01 8.17511082e-01
3.41396511e-01 -4.37270463e-01 4.77396220e-01 2.10115805e-01
4.04242128e-01 -5.65789640e-02 3.82192403e-01 -1.05862403e+00
6.73563257e-02 -8.65648270e-01 -3.87198478e-01 -2.18570888e-01
-1.82203218e-01 -8.64418268e-01 -6.86100200e-02 -1.90952706e+00
5.11423886e-01 -5.93541339e-02 -7.93886662e-01 5.14628410e-01
-2.72526860e-01 4.54474121e-01 -4.89987344e-01 2.67841309e-01
-4.67127502e-01 -3.79287779e-01 1.32454967e+00 -2.47244611e-01
7.26154447e-02 -1.98266923e-01 -7.63932824e-01 6.41956091e-01
7.68476844e-01 -2.93312043e-01 2.42537960e-01 -6.11930117e-02
-2.55615920e-01 -1.24066047e-01 5.32191336e-01 -9.67948139e-01
7.43160129e-01 3.45328182e-01 8.68698537e-01 -4.47185576e-01
6.76126555e-02 -8.53184521e-01 3.08168620e-01 8.13294947e-01
-5.04155636e-01 -2.89796386e-02 3.99256736e-01 4.78418618e-01
-3.82932901e-01 -2.07773000e-01 8.72124016e-01 -6.60332322e-01
-2.40629002e-01 2.96589524e-01 -7.84702659e-01 -5.51265478e-01
1.23891628e+00 -4.17868972e-01 -6.31505013e-01 1.34657636e-01
-8.71934414e-01 8.11474174e-02 5.46334505e-01 -3.26445609e-01
3.08544099e-01 -1.04012477e+00 -8.11270952e-01 8.18056539e-02
1.48227111e-01 5.54534018e-01 8.05621564e-01 1.44680631e+00
-9.98224795e-01 7.33915985e-01 -1.74187958e-01 -7.90877283e-01
-1.26555920e+00 6.23771489e-01 6.64798737e-01 -1.14184558e+00
-6.50388420e-01 9.45064783e-01 2.66298920e-01 -3.44136328e-01
5.22759780e-02 -8.01681280e-01 -2.87498504e-01 5.99958468e-03
4.15303886e-01 1.38655066e-01 3.16491187e-01 -1.21342897e-01
-9.66024175e-02 5.20517230e-01 -4.03877199e-01 3.42926502e-01
1.20494521e+00 3.50072533e-02 -3.07616889e-01 3.68712276e-01
1.27454352e+00 -9.35731605e-02 -7.95020342e-01 2.21684963e-01
-1.43035665e-01 -3.69565040e-02 2.97739893e-01 -1.02172601e+00
-7.21709967e-01 6.30585432e-01 7.19069898e-01 5.83446026e-01
1.24082327e+00 -4.87675816e-02 7.76903510e-01 1.20163567e-01
1.69546276e-01 -8.08488607e-01 -3.66030842e-01 -1.97573118e-02
7.25211203e-01 -1.43982852e+00 -1.32541388e-01 -2.92886347e-02
-6.95175305e-02 1.54077911e+00 5.04261851e-01 -3.07568759e-01
4.80966210e-01 5.75776517e-01 2.78504580e-01 -9.80893672e-02
-9.49446261e-01 1.12067215e-01 1.90044001e-01 4.40892190e-01
8.40905666e-01 -1.26959637e-01 -2.87058026e-01 7.76138246e-01
1.14337541e-01 4.07705098e-01 3.15962374e-01 7.00792730e-01
-2.40670592e-01 -6.05216861e-01 -2.73616314e-01 8.37764084e-01
-1.23250091e+00 -1.16951607e-01 -1.42741486e-01 1.19927120e+00
7.26504177e-02 2.92112291e-01 3.52800161e-01 1.35554969e-01
-1.58663407e-01 -6.53146803e-02 2.16016024e-01 -3.36527616e-01
-9.01681542e-01 1.32525370e-01 3.75525691e-02 -1.19071238e-01
-5.16827047e-01 -5.94019353e-01 -1.25389469e+00 -9.81370583e-02
-4.09633487e-01 1.99634761e-01 4.89113986e-01 5.35720706e-01
1.88702226e-01 7.69798756e-01 2.60976136e-01 -9.55066562e-01
-1.74669087e-01 -1.22163451e+00 -5.63807666e-01 6.81743026e-02
6.91162646e-01 -4.58078869e-02 -7.54970014e-01 2.88748145e-01] | [15.090215682983398, -2.9988787174224854] |
9f30093c-2d56-44ff-9222-f7c6c9b9d99f | siamese-sleep-transformer-for-robust-sleep | 2212.13919 | null | https://arxiv.org/abs/2212.13919v1 | https://arxiv.org/pdf/2212.13919v1.pdf | Siamese Sleep Transformer For Robust Sleep Stage Scoring With Self-knowledge Distillation and Selective Batch Sampling | In this paper, we propose a Siamese sleep transformer (SST) that effectively extracts features from single-channel raw electroencephalogram signals for robust sleep stage scoring. Despite the significant advances in sleep stage scoring in the last few years, most of them mainly focused on the increment of model performance. However, other problems still exist: the bias of labels in datasets and the instability of model performance by repetitive training. To alleviate these problems, we propose the SST, a novel sleep stage scoring model with a selective batch sampling strategy and self-knowledge distillation. To evaluate how robust the model was to the bias of labels, we used different datasets for training and testing: the sleep heart health study and the Sleep-EDF datasets. In this condition, the SST showed competitive performance in sleep stage scoring. In addition, we demonstrated the effectiveness of the selective batch sampling strategy with a reduction of the standard deviation of performance by repetitive training. These results could show that SST extracted effective learning features against the bias of labels in datasets, and the selective batch sampling strategy worked for the model robustness in training. | ['Gi-Hwan Shin', 'Young-Seok Kweon', 'Heon-Gyu Kwak'] | 2022-12-12 | null | null | null | null | ['self-knowledge-distillation'] | ['computer-vision'] | [ 5.46625182e-02 -1.37673676e-01 -6.80089667e-02 -5.80051482e-01
-4.41700906e-01 -1.76262006e-01 3.41671228e-01 -9.43936259e-02
-7.12115765e-01 1.00538135e+00 -7.28334114e-02 2.80171007e-01
-3.26987922e-01 -3.40385497e-01 -3.09217960e-01 -8.35993826e-01
-1.26329511e-01 2.35753909e-01 4.44825023e-01 -2.15287022e-02
4.05336678e-01 2.85167307e-01 -1.76388967e+00 1.46488503e-01
1.17422986e+00 1.07516873e+00 3.10304493e-01 1.53472915e-01
7.13647082e-02 6.55055106e-01 -9.20357883e-01 -5.74833862e-02
2.69188620e-02 -6.05775177e-01 -5.08009136e-01 -2.88720131e-01
4.22510616e-02 1.18230663e-01 3.39888297e-02 9.65957701e-01
8.63850832e-01 8.51657148e-03 3.78956258e-01 -1.31331396e+00
-9.30655226e-02 4.28994536e-01 -8.21168721e-02 5.80320060e-01
3.95221144e-01 3.84242646e-03 7.17199564e-01 -6.99285805e-01
2.22869590e-01 6.08192861e-01 8.32469404e-01 6.86588347e-01
-1.20215487e+00 -1.20693552e+00 -9.54707339e-02 5.15399337e-01
-1.60213709e+00 -5.64862251e-01 7.36058235e-01 -3.56535614e-01
8.48128140e-01 3.12575728e-01 1.14673984e+00 1.04438114e+00
5.21770597e-01 5.91885090e-01 1.53440344e+00 -2.45008618e-01
5.65525174e-01 6.45199299e-01 3.58159244e-01 4.88543183e-01
2.39959821e-01 1.03480354e-01 -8.38595867e-01 2.45946068e-02
3.15912068e-01 1.53689599e-03 -1.63287386e-01 -2.20753551e-01
-9.13056374e-01 6.76071048e-01 1.83551162e-01 7.41920590e-01
-5.45801930e-02 -4.90598619e-01 2.98636198e-01 3.44643891e-01
5.27434468e-01 7.68658996e-01 -4.42488849e-01 -2.32035175e-01
-1.60275364e+00 -1.39320731e-01 6.47973537e-01 5.25974572e-01
6.15726411e-01 3.14708762e-02 -3.55648607e-01 7.04268277e-01
5.80573305e-02 5.86724162e-01 1.14864969e+00 -5.93979001e-01
2.25342005e-01 9.26836193e-01 -8.51790011e-02 -8.05412471e-01
-9.87479687e-01 -7.02846348e-01 -7.85127699e-01 -8.79514888e-02
1.56601951e-01 9.33577642e-02 -6.53405786e-01 1.74958885e+00
-1.47763044e-02 1.58732668e-01 1.28931347e-02 8.11543107e-01
8.42115641e-01 1.98723361e-01 -8.08002651e-02 -4.90381151e-01
1.20400167e+00 -6.77528381e-01 -9.41920757e-01 -2.91869223e-01
5.50980210e-01 -3.49461585e-01 1.31875467e+00 6.38787866e-01
-1.04475296e+00 -7.03480244e-01 -1.18770075e+00 3.76094788e-01
-2.65774280e-01 2.99245030e-01 5.02492845e-01 8.53444576e-01
-1.03067589e+00 9.03476357e-01 -9.99829352e-01 -4.39441264e-01
4.62176472e-01 8.44752252e-01 -3.77773017e-01 4.89865184e-01
-1.27711964e+00 9.92017865e-01 1.57958582e-01 4.45299968e-02
-6.86225355e-01 -5.90789437e-01 -4.80829149e-01 2.50557542e-01
-2.50972454e-02 -3.88335139e-01 8.56082082e-01 -9.90298808e-01
-1.51483464e+00 7.36637831e-01 -1.98266983e-01 -5.17672658e-01
2.55466998e-01 -3.22663635e-01 -5.93557417e-01 4.26264666e-02
1.61329597e-01 1.71498761e-01 8.93508792e-01 -6.37353599e-01
-4.05238211e-01 -6.10638976e-01 -4.35384840e-01 -7.28435516e-02
-5.85663557e-01 -1.50409881e-02 4.47633006e-02 -3.50684226e-01
1.31219536e-01 -8.99317503e-01 6.31912500e-02 -5.18438399e-01
-3.94664444e-02 -2.52257943e-01 3.07066560e-01 -1.89549461e-01
1.51821756e+00 -2.47341919e+00 -3.74014378e-02 1.11057803e-01
8.36016387e-02 3.21964741e-01 1.90711737e-01 2.12024868e-01
-1.48079097e-01 -2.70785261e-02 -1.77601680e-01 -3.42017084e-01
-2.00508446e-01 1.37701184e-01 -1.47078544e-01 5.31575918e-01
9.69250053e-02 6.85323417e-01 -6.99175477e-01 -6.65619016e-01
2.69279373e-03 2.34814212e-01 -4.34841603e-01 2.48748630e-01
4.81865853e-01 5.35217881e-01 -2.04806447e-01 3.77722591e-01
4.12076473e-01 -2.56047666e-01 -1.92596853e-01 -2.62787133e-01
-1.56426474e-01 4.23175901e-01 -1.01732552e+00 1.67674816e+00
-2.83854574e-01 5.00234723e-01 -3.95032018e-01 -8.09388161e-01
1.00248063e+00 1.98613122e-01 5.04138112e-01 -1.31080377e+00
9.70362648e-02 3.32735479e-01 9.18735489e-02 -5.46658158e-01
2.87392855e-01 -4.60581899e-01 1.02591567e-01 4.15019065e-01
1.64631531e-01 -1.76192135e-01 1.41838878e-01 -6.70912908e-03
9.27366078e-01 -1.16029277e-01 2.72316128e-01 -4.54149842e-01
5.30854940e-01 -1.56731427e-01 7.63762474e-01 5.47401309e-01
-3.11210036e-01 5.92995524e-01 3.54449391e-01 -3.15004319e-01
-3.64101470e-01 -9.81418967e-01 -4.03690517e-01 8.59207273e-01
9.20009241e-02 -6.99230134e-01 -6.67146921e-01 -7.58095920e-01
-3.29964876e-01 7.22208619e-01 -5.33561826e-01 -7.99859941e-01
-1.78692952e-01 -1.15114117e+00 6.77439213e-01 4.61862564e-01
5.95386028e-01 -1.04716039e+00 -9.92397845e-01 1.18874565e-01
-1.73609868e-01 -8.98351312e-01 -1.46396816e-01 6.98350787e-01
-1.18405509e+00 -1.09246182e+00 -5.38082123e-01 -5.75972259e-01
5.29195964e-01 -7.26996139e-02 9.19390082e-01 -3.15436870e-02
-1.36904493e-01 4.92241094e-03 -3.02557439e-01 -4.75588769e-01
-1.56318590e-01 2.77909845e-01 4.41632301e-01 -8.17880258e-02
4.55977410e-01 -7.17131317e-01 -7.17024207e-01 6.32949114e-01
-6.64395094e-01 -2.44564444e-01 5.25550365e-01 7.10492194e-01
4.58923101e-01 2.92194813e-01 7.91093946e-01 -6.24536693e-01
6.86108768e-01 -2.77890801e-01 -3.82546246e-01 -1.09382216e-02
-1.25270927e+00 9.81081799e-02 6.40490174e-01 -4.21872646e-01
-6.72667027e-01 -3.56549211e-02 -1.86770245e-01 1.87309682e-02
7.53052458e-02 2.18176022e-01 2.35998840e-03 2.75977217e-02
6.51998758e-01 5.48398197e-01 1.13618486e-01 -5.37489772e-01
-4.04548287e-01 8.06093514e-01 7.83203989e-02 -1.08419798e-01
4.29334581e-01 4.00828630e-01 -1.31910205e-01 -6.93390667e-01
-9.44963515e-01 -4.21654880e-01 -5.04968464e-01 -1.05187573e-01
7.81606436e-01 -9.97278094e-01 -5.69243193e-01 5.85355163e-01
-3.66466254e-01 -4.76426512e-01 -4.58450824e-01 7.12024510e-01
-3.75869334e-01 3.13066214e-01 -1.21351309e-01 -8.32256794e-01
-5.65416157e-01 -1.15577137e+00 8.80644083e-01 3.18932980e-01
-4.57327068e-01 -8.51939380e-01 3.19912404e-01 1.81300670e-01
6.24684930e-01 -1.89021364e-01 6.69805050e-01 -8.55532765e-01
3.52713056e-02 -1.43713251e-01 1.03795588e-01 5.87841451e-01
2.69334197e-01 -2.71910846e-01 -1.27946055e+00 -4.17227596e-01
5.84822655e-01 -2.81873703e-01 6.88088834e-01 1.64449796e-01
8.98258209e-01 2.32985187e-02 -2.21498072e-01 6.69994831e-01
1.02045619e+00 3.35758686e-01 5.84580362e-01 5.62056303e-01
1.35761067e-01 2.40705177e-01 5.39841950e-01 3.06640714e-01
2.33384967e-01 7.41132796e-01 1.45229384e-01 -1.44900754e-01
-7.34934136e-02 -5.22792637e-02 6.42683804e-01 9.84387338e-01
-3.06323878e-02 4.45672899e-01 -7.15053558e-01 3.96224111e-01
-1.51759756e+00 -8.42624187e-01 -1.16088567e-02 2.27443409e+00
8.34964275e-01 5.29612601e-01 3.24679762e-01 5.87870002e-01
2.96164572e-01 -3.74196284e-02 -4.74133730e-01 -2.49856427e-01
-1.95773795e-01 4.59501922e-01 2.89081842e-01 -5.63229248e-02
-6.42586172e-01 5.76323509e-01 6.95908403e+00 7.54535675e-01
-1.42333984e+00 3.21274787e-01 1.85149699e-01 -5.48302293e-01
1.46755919e-01 -2.10707337e-01 -8.74775946e-01 9.99738395e-01
1.28140855e+00 -1.42127171e-01 3.64858180e-01 7.50457644e-01
4.21578288e-01 -4.14343476e-01 -9.05133426e-01 1.04716551e+00
3.18697274e-01 -8.07535529e-01 -4.01831776e-01 -1.32345766e-01
5.00462294e-01 1.93412080e-01 -1.77507401e-01 4.98571664e-01
-4.40421790e-01 -9.55260813e-01 7.04439223e-01 6.68315530e-01
6.95518374e-01 -4.65943009e-01 1.05382395e+00 4.48815495e-01
-8.52584302e-01 -3.53415728e-01 -3.00567150e-01 -3.56470197e-01
-2.61835665e-01 7.94430614e-01 -7.84315467e-01 3.31589788e-01
9.65454400e-01 7.50850379e-01 -1.24022031e+00 1.22722983e+00
-2.68768996e-01 9.24413025e-01 -4.34763402e-01 -1.74948156e-01
-2.63800025e-01 -1.80043757e-01 1.98210761e-01 1.00404274e+00
2.86806315e-01 -4.19807792e-01 -2.84449011e-01 7.56254554e-01
3.48212153e-01 -4.12856787e-02 -3.15637797e-01 1.37663662e-01
3.30280781e-01 1.25029564e+00 -8.15183103e-01 -1.40992939e-01
-1.93303123e-01 7.31383502e-01 1.52428165e-01 6.87268972e-02
-8.72056901e-01 -4.09150153e-01 1.31726488e-01 2.73446769e-01
1.51239470e-01 1.53512433e-02 -3.86006683e-01 -1.23778474e+00
-9.14211385e-03 -7.75329292e-01 2.94659644e-01 -6.60411954e-01
-1.07232499e+00 6.68922722e-01 1.52086809e-01 -1.14316666e+00
1.58032507e-01 -1.12826884e-01 -6.90153778e-01 4.34632242e-01
-1.39635789e+00 -5.16573906e-01 -3.66694093e-01 8.14574420e-01
4.42158937e-01 -2.40323395e-01 7.71382272e-01 5.33625782e-01
-8.70507836e-01 6.72702968e-01 7.92204589e-02 -2.68698007e-01
9.80392039e-01 -1.23511577e+00 -9.66739133e-02 5.46072781e-01
1.08639091e-01 8.16902459e-01 5.58529615e-01 -2.87633508e-01
-9.06561553e-01 -7.38812685e-01 9.70160306e-01 -3.40496421e-01
4.29482281e-01 -5.32345176e-01 -7.34887362e-01 2.44817600e-01
-8.76864940e-02 -2.27453753e-01 8.88346732e-01 1.00355238e-01
1.29702374e-01 -7.48197913e-01 -1.15953732e+00 2.65110224e-01
8.15744042e-01 -4.46066946e-01 -9.83129263e-01 6.55258372e-02
8.50290880e-02 -9.18502137e-02 -8.25210631e-01 5.03780782e-01
7.29124606e-01 -1.23287439e+00 5.33488214e-01 -2.24297494e-01
-6.63568825e-02 -2.14234099e-01 1.01939932e-01 -1.35180187e+00
-4.25092131e-01 -3.46246123e-01 5.88814840e-02 1.28864086e+00
3.77297491e-01 -7.27284253e-01 6.72856569e-01 5.52727759e-01
-1.10116519e-01 -7.59837329e-01 -1.17692971e+00 -8.74592721e-01
-2.00814322e-01 -1.71890870e-01 5.45623422e-01 4.42278594e-01
2.86006182e-01 5.76732576e-01 -1.65706769e-01 -2.83817917e-01
3.48374039e-01 -4.60319500e-03 5.53362012e-01 -1.51284337e+00
-3.67951579e-02 -2.29261994e-01 -4.93348867e-01 -4.25550967e-01
8.61207247e-02 -8.73241901e-01 3.62168124e-04 -1.49827206e+00
3.45961690e-01 -2.93468833e-01 -8.88023376e-01 5.32546818e-01
-3.28624636e-01 4.00369823e-01 8.69432464e-02 3.86492401e-01
-8.56096745e-01 5.37040770e-01 9.29768801e-01 -1.66712701e-02
-4.39192623e-01 4.62229609e-01 -6.83079362e-01 6.21511996e-01
1.08526611e+00 -9.40188169e-01 -7.21455216e-01 3.28303128e-02
3.85436803e-01 -3.04164946e-01 1.62833080e-01 -1.46229446e+00
2.31370941e-01 3.34614992e-01 4.80772942e-01 -3.98514092e-01
3.59190851e-01 -8.29883516e-01 2.07198754e-01 7.70937562e-01
-1.48730591e-01 9.13580433e-02 3.34161311e-01 2.05087572e-01
-1.37874514e-01 -2.19152749e-01 7.56498158e-01 -1.16459923e-02
-4.92157906e-01 -1.72400251e-01 -5.31616688e-01 1.66347295e-01
8.57459664e-01 -5.91904223e-01 -1.52872577e-01 1.03036411e-01
-7.24777222e-01 1.15159608e-01 4.99581993e-01 2.71280110e-01
3.99408340e-01 -9.68593955e-01 -2.84627974e-01 8.35898817e-01
2.63350662e-02 -2.70862073e-01 9.18942019e-02 1.49942422e+00
8.21594223e-02 4.20924902e-01 -5.23969293e-01 -7.18160748e-01
-1.23181129e+00 3.32139254e-01 3.57841909e-01 -3.49447191e-01
-5.39319813e-01 6.37040436e-01 -1.24154970e-01 -5.90619184e-02
3.98336589e-01 -5.67628920e-01 -5.35313427e-01 3.41100454e-01
4.73454356e-01 4.89503056e-01 6.17230177e-01 -2.68390834e-01
-7.55589962e-01 6.23829484e-01 6.55917376e-02 1.27614900e-01
1.41642642e+00 -2.63421625e-01 4.01339978e-02 9.32892799e-01
8.19840252e-01 1.22266136e-01 -8.70226800e-01 2.63872176e-01
6.29585236e-02 -1.74123690e-01 1.22590393e-01 -9.71589804e-01
-1.08873343e+00 7.35907197e-01 1.09529769e+00 3.47867042e-01
1.33646023e+00 -2.33472943e-01 6.15264893e-01 2.81742454e-01
5.65696001e-01 -1.13049936e+00 -3.41440365e-02 1.93346471e-01
4.59598064e-01 -1.00547981e+00 1.80693701e-01 2.21792519e-01
-7.57370532e-01 9.36755776e-01 4.50656027e-01 -1.23775214e-01
6.61993206e-01 1.30200043e-01 4.70376089e-02 -3.75073552e-01
-5.62295556e-01 -4.68550585e-02 2.34215930e-01 5.12101948e-01
2.62582183e-01 -1.98298171e-01 -5.64988554e-01 1.06963837e+00
-5.16868532e-01 5.01679182e-01 2.35166624e-01 7.16576099e-01
-4.62135732e-01 -9.32671070e-01 -1.73734263e-01 6.23607159e-01
-5.29866517e-01 -5.85385226e-02 -3.75063628e-01 7.98157036e-01
3.80944341e-01 1.14942729e+00 -1.57599851e-01 -6.35238230e-01
4.92507577e-01 4.88690197e-01 4.04715091e-01 -6.73606157e-01
-1.00416052e+00 -1.98421866e-01 -6.60035387e-02 -6.85464978e-01
-6.30772114e-01 -6.45928800e-01 -1.18654442e+00 2.83587009e-01
-7.29884505e-01 6.44346774e-01 6.16749704e-01 9.83216882e-01
4.64955598e-01 6.38785958e-01 7.10248470e-01 -5.31752110e-01
-6.04868650e-01 -1.18896818e+00 -9.33490396e-01 3.57160807e-01
2.82958388e-01 -9.04922128e-01 -8.15096796e-01 -2.79365182e-01] | [13.5126371383667, 3.527578592300415] |
262dcea8-45a6-474f-80ef-70e72bbc2aa8 | deepatom-a-framework-for-protein-ligand | 1912.00318 | null | https://arxiv.org/abs/1912.00318v1 | https://arxiv.org/pdf/1912.00318v1.pdf | DeepAtom: A Framework for Protein-Ligand Binding Affinity Prediction | The cornerstone of computational drug design is the calculation of binding affinity between two biological counterparts, especially a chemical compound, i.e., a ligand, and a protein. Predicting the strength of protein-ligand binding with reasonable accuracy is critical for drug discovery. In this paper, we propose a data-driven framework named DeepAtom to accurately predict the protein-ligand binding affinity. With 3D Convolutional Neural Network (3D-CNN) architecture, DeepAtom could automatically extract binding related atomic interaction patterns from the voxelized complex structure. Compared with the other CNN based approaches, our light-weight model design effectively improves the model representational capacity, even with the limited available training data. With validation experiments on the PDBbind v.2016 benchmark and the independent Astex Diverse Set, we demonstrate that the less feature engineering dependent DeepAtom approach consistently outperforms the other state-of-the-art scoring methods. We also compile and propose a new benchmark dataset to further improve the model performances. With the new dataset as training input, DeepAtom achieves Pearson's R=0.83 and RMSE=1.23 pK units on the PDBbind v.2016 core set. The promising results demonstrate that DeepAtom models can be potentially adopted in computational drug development protocols such as molecular docking and virtual screening. | ['Dapeng Wu', 'Xiaolin Li', 'Mohammad A. Rezaei', 'Yanjun Li', 'Chenglong Li'] | 2019-12-01 | null | null | null | null | ['molecular-docking'] | ['medical'] | [ 5.88977057e-03 -3.20844173e-01 -3.38316828e-01 -3.33957404e-01
-8.02712023e-01 -3.74121368e-01 1.65789664e-01 3.80218536e-01
-5.18977106e-01 1.31561518e+00 -2.02802643e-02 -4.37523812e-01
-7.13354647e-02 -6.56606555e-01 -1.02689302e+00 -8.99510860e-01
-9.84545052e-02 7.20916271e-01 1.15528695e-01 -2.61124730e-01
1.58918798e-01 8.73266459e-01 -8.94390404e-01 4.14573610e-01
1.01312745e+00 9.35663104e-01 1.84158832e-01 -3.56364883e-02
1.46047249e-01 2.68486470e-01 -2.69930303e-01 -3.65103900e-01
1.36354351e-02 -1.27597570e-01 -8.24925780e-01 -7.14023590e-01
2.61657983e-01 -1.76893145e-01 -1.47491008e-01 6.01591885e-01
1.10565507e+00 -2.30966136e-02 7.63088882e-01 -3.80064249e-01
-7.21153855e-01 -4.56210505e-03 -2.72220284e-01 1.18581250e-01
4.79657441e-01 4.51636225e-01 1.09860551e+00 -1.38219666e+00
6.79202020e-01 8.95870209e-01 8.09998810e-01 5.67197382e-01
-1.45397508e+00 -8.33429217e-01 -3.54177386e-01 4.97394294e-01
-1.51133370e+00 1.84302661e-03 3.49817455e-01 -5.26512444e-01
1.72239900e+00 8.56182724e-02 6.37813210e-01 1.16485786e+00
7.03680694e-01 2.67200142e-01 6.27211332e-01 5.66286258e-02
2.72551626e-01 -6.15789890e-01 6.81568459e-02 7.60340214e-01
4.42037344e-01 1.94623500e-01 -6.72796428e-01 -8.20203364e-01
4.37882870e-01 1.16992489e-01 -3.04498404e-01 -4.17633772e-01
-9.72404182e-01 8.63493979e-01 8.69070172e-01 -1.33091480e-01
-6.49735689e-01 1.80305168e-01 2.82965630e-01 -1.58493713e-01
2.28240520e-01 7.73328125e-01 -8.69651914e-01 1.60034001e-01
-3.02046567e-01 6.04958415e-01 6.11462176e-01 4.47339356e-01
5.04929841e-01 -2.90587664e-01 -1.18179932e-01 6.21235490e-01
4.51156288e-01 1.88195869e-01 1.58236787e-01 -1.65798590e-01
1.48029938e-01 6.96424663e-01 1.96871251e-01 -6.55063987e-01
-8.35598171e-01 -4.81258929e-01 -8.05351198e-01 2.39119753e-01
3.51322263e-01 3.13129127e-02 -9.08506870e-01 1.70044541e+00
3.98871601e-01 4.69008945e-02 1.28222659e-01 9.24405396e-01
1.26440680e+00 6.42923594e-01 4.31946337e-01 -2.72151113e-01
1.21815038e+00 -6.85138702e-01 -3.68552566e-01 2.70881474e-01
8.22752059e-01 -7.19497204e-01 7.64110208e-01 4.91973996e-01
-9.29022908e-01 -2.91871369e-01 -1.22294211e+00 -7.75861889e-02
-2.65643120e-01 -7.80235743e-04 9.31243718e-01 2.11497679e-01
-4.94770080e-01 8.89818132e-01 -7.80737340e-01 2.21765637e-02
7.06529319e-01 1.12349188e+00 -7.87236273e-01 -8.28883518e-03
-1.42000067e+00 1.22007227e+00 4.61615413e-01 1.56010231e-02
-9.72521484e-01 -1.07951748e+00 -3.74951959e-01 1.12741180e-01
-1.18755952e-01 -7.09500790e-01 1.01110518e+00 -2.76500553e-01
-1.36728323e+00 4.95628536e-01 -9.64789093e-02 -4.89856333e-01
7.10947737e-02 -9.36435685e-02 -1.60251241e-02 -2.28278235e-01
-5.69572039e-02 6.24087989e-01 -1.12925075e-01 -7.98575342e-01
-1.44980073e-01 -6.26003146e-01 -3.60524580e-02 2.37742037e-01
-3.52136753e-02 -8.11785236e-02 -2.06643522e-01 -1.77305534e-01
-1.32775292e-01 -8.93766224e-01 -5.73660791e-01 -2.82740854e-02
-4.21029747e-01 -3.56325895e-01 2.12529138e-01 -2.64853269e-01
9.61201429e-01 -1.62027216e+00 4.74404007e-01 5.14100254e-01
6.13381028e-01 5.91311932e-01 -8.44670013e-02 7.45935500e-01
-5.31028688e-01 -1.41307533e-01 2.36886330e-02 2.86706179e-01
-4.06494200e-01 -1.70344815e-01 2.11305305e-01 7.15627789e-01
2.82484531e-01 1.14336395e+00 -6.62030399e-01 -5.85618131e-02
-8.31629261e-02 8.95254791e-01 -8.42467785e-01 1.68449536e-01
-4.81947362e-01 6.87413692e-01 -8.28427374e-01 7.14812160e-01
8.75282824e-01 -5.51095366e-01 4.45449352e-01 -4.80743259e-01
1.32248383e-02 3.23551744e-01 -4.90989387e-01 1.78034115e+00
2.20438279e-02 8.51869117e-03 -5.43747544e-01 -5.89844584e-01
9.79237080e-01 3.41107368e-01 7.13545501e-01 -8.66912007e-01
1.02573156e-01 4.43550706e-01 6.90103114e-01 -1.99509367e-01
-3.44929665e-01 -2.08210170e-01 2.82151043e-01 4.74420264e-02
1.68003082e-01 3.30906361e-01 -1.56435981e-01 -1.00002103e-01
1.11959207e+00 1.28658757e-01 6.13217294e-01 -4.02217656e-01
5.43435037e-01 1.72619581e-01 5.67250609e-01 1.25898227e-01
1.22052506e-01 3.16895157e-01 4.45139706e-01 -1.11317134e+00
-9.86450195e-01 -6.51751995e-01 -8.23375285e-01 9.79972899e-01
-4.03131954e-02 -7.20779896e-01 -6.55190945e-01 -4.34640259e-01
3.84118825e-01 -4.56561372e-02 -7.82926798e-01 -4.63874638e-01
-3.86539191e-01 -1.34190166e+00 4.96504158e-01 3.14633638e-01
3.19037847e-02 -1.02069998e+00 -1.25708552e-02 7.47884929e-01
1.45011857e-01 -4.95855421e-01 -2.35790506e-01 8.18382680e-01
-5.22243738e-01 -1.34873271e+00 -5.12911499e-01 -7.17794240e-01
3.19108397e-01 -7.64714405e-02 9.49255288e-01 1.92539886e-01
-3.91185135e-01 -6.95815980e-01 -3.67502458e-02 -5.40550828e-01
-4.03576605e-02 6.14563525e-02 2.69976974e-01 -3.92382413e-01
7.98633993e-01 -6.52706146e-01 -1.06838083e+00 3.76761854e-01
-4.76343095e-01 1.11737348e-01 6.52927577e-01 1.11329281e+00
1.09459007e+00 -5.80362797e-01 6.55394197e-01 -7.31322348e-01
6.76566362e-01 -5.26995718e-01 -5.27837455e-01 4.82568443e-02
-6.44794583e-01 3.18919331e-01 7.86159277e-01 -3.67935002e-01
-6.09450221e-01 5.69522679e-01 -7.17606068e-01 -7.02564418e-02
1.05323665e-01 9.81686175e-01 -4.05090123e-01 -6.39150560e-01
9.88266110e-01 1.12885080e-01 -2.06151549e-02 -6.95818543e-01
-1.89522520e-01 4.05695140e-01 1.24470234e-01 -6.77739978e-01
3.10008377e-01 1.71690315e-01 4.80176926e-01 -4.51138854e-01
-6.84426010e-01 -3.82678717e-01 -6.49700284e-01 3.75921905e-01
9.69299614e-01 -9.39678669e-01 -1.45599151e+00 2.02140674e-01
-1.49291158e+00 -2.83436984e-01 4.85365659e-01 6.35749519e-01
-4.07751083e-01 3.77761841e-01 -5.63895047e-01 4.06523468e-03
-7.25289285e-01 -1.70579278e+00 1.00220942e+00 -1.92749903e-01
-3.08240145e-01 -7.03843892e-01 5.76837242e-01 3.82086694e-01
1.82108328e-01 5.75441658e-01 1.22336280e+00 -1.06361425e+00
-5.10912061e-01 -2.70888835e-01 -1.83058009e-01 -5.36180884e-02
1.55065253e-01 -9.34327170e-02 -8.33606720e-01 -1.75187528e-01
-5.02596021e-01 -5.71576416e-01 8.69590223e-01 4.95508313e-01
1.23276639e+00 -1.77938774e-01 -4.59680468e-01 8.62609625e-01
1.33313024e+00 5.61245441e-01 8.28126311e-01 3.96559447e-01
8.99873316e-01 9.82245207e-02 5.19615233e-01 5.62962592e-01
2.60297060e-02 1.16399050e+00 8.02182376e-01 -5.03480256e-01
2.61597991e-01 4.30299602e-02 -8.79070759e-02 1.86754301e-01
-5.48821032e-01 -2.57615417e-01 -1.19295263e+00 -2.07819849e-01
-1.89377904e+00 -8.72963130e-01 -4.76411790e-01 2.05783987e+00
1.40016174e+00 9.29765496e-03 4.89411969e-03 -3.27707171e-01
1.39193550e-01 -4.26861584e-01 -1.03215897e+00 -2.41441607e-01
-1.84247434e-01 8.55699122e-01 4.27204162e-01 4.62533861e-01
-9.35063958e-01 9.37335014e-01 6.25961256e+00 9.90082622e-01
-1.14081633e+00 -7.66535848e-02 7.48213172e-01 -1.49893641e-01
2.44460236e-02 -2.81507909e-01 -8.73916626e-01 3.98921013e-01
1.08894432e+00 -1.00595221e-01 1.07454650e-01 7.03089535e-01
3.99199694e-01 2.53480613e-01 -1.34232152e+00 7.96050787e-01
-4.56753343e-01 -1.99946928e+00 3.41390878e-01 6.04541361e-01
4.96449351e-01 2.86643893e-01 8.58065560e-02 -6.03629369e-03
1.96761657e-02 -1.80806899e+00 1.21365987e-01 4.02336925e-01
9.79469955e-01 -1.03287220e+00 9.17627513e-01 6.46285564e-02
-1.02252686e+00 3.28973562e-01 -6.63355112e-01 6.98585957e-02
-2.42007300e-01 3.50501209e-01 -8.07431936e-01 4.84689683e-01
4.35784966e-01 7.41413116e-01 -2.29419231e-01 1.11550975e+00
2.41468012e-01 1.97981387e-01 -6.57262504e-02 -1.39387608e-01
3.43498617e-01 -1.81987420e-01 -5.83861396e-02 1.05115771e+00
-1.93331450e-01 4.47870255e-01 3.26881498e-01 8.65305066e-01
-3.65001708e-01 4.99217957e-01 -2.81467706e-01 2.55700350e-01
2.57406443e-01 9.76445854e-01 -6.69041723e-02 3.28745171e-02
-4.29303586e-01 7.08203137e-01 5.78172743e-01 2.79134899e-01
-1.01819193e+00 -1.83521166e-01 1.25747263e+00 1.41723946e-01
9.43702906e-02 4.42691296e-02 2.81824917e-02 -7.64992774e-01
-2.09025472e-01 -1.05558789e+00 -1.94122863e-03 -5.12265861e-01
-1.24884093e+00 5.42503893e-01 -3.03663641e-01 -9.38668847e-01
3.94758105e-01 -1.17248702e+00 -5.10680974e-01 1.25991380e+00
-1.43663549e+00 -9.41602767e-01 -7.28297904e-02 4.12512422e-01
2.04837084e-01 -2.94075131e-01 1.28521109e+00 4.24029678e-01
-7.07280755e-01 6.18382692e-01 6.13752007e-01 -2.82653987e-01
8.42894673e-01 -1.07666111e+00 4.57437396e-01 -1.22282565e-01
-4.14268523e-01 9.66441154e-01 6.38140857e-01 -8.03427756e-01
-1.68657255e+00 -1.11684155e+00 5.66287696e-01 -4.29050893e-01
5.36786497e-01 -2.99682766e-01 -1.15760279e+00 1.11224748e-01
-1.43036619e-01 1.97424501e-01 1.39633536e+00 4.48364532e-03
-3.23705614e-01 2.30416089e-01 -1.07059777e+00 3.77848029e-01
1.06830204e+00 -2.48283654e-01 -7.29706064e-02 8.20414066e-01
6.21265352e-01 -7.04996467e-01 -1.27086306e+00 7.27133214e-01
8.16232383e-01 -7.56340802e-01 1.39666820e+00 -1.32211125e+00
3.36750627e-01 -2.69767135e-01 -1.80089131e-01 -1.10722220e+00
-8.97560120e-01 -2.31343716e-01 -2.09311560e-01 1.01038203e-01
8.15947473e-01 -4.79843736e-01 9.85764384e-01 4.86124873e-01
-2.95758396e-01 -1.58888841e+00 -1.18235600e+00 -5.68938911e-01
5.00461578e-01 -7.95208791e-04 6.72953963e-01 7.04637349e-01
1.10455215e-01 7.49487817e-01 -3.04386646e-01 1.42865966e-03
2.87389755e-01 -1.21140979e-01 7.33363271e-01 -1.77540684e+00
-3.30580562e-01 -2.45027766e-01 -5.70776701e-01 -8.47641349e-01
2.47642457e-01 -1.09715915e+00 -5.23670793e-01 -1.58928537e+00
6.55048728e-01 -1.66383237e-01 -6.52937353e-01 7.67128289e-01
-5.25163636e-02 1.83812082e-01 -4.82330889e-01 2.75601983e-01
-5.12737036e-01 7.92912483e-01 1.41651142e+00 -3.97286057e-01
-1.52157918e-01 -8.73033628e-02 -6.17483258e-01 2.81682134e-01
8.28424215e-01 -5.10880113e-01 -1.65266693e-01 1.38933375e-01
3.75237107e-01 -2.32966423e-01 2.17549443e-01 -6.61450386e-01
-1.32307306e-01 -4.02436286e-01 7.92202175e-01 -5.04441917e-01
4.84685063e-01 -7.90665030e-01 5.70881069e-01 8.30383778e-01
4.27986815e-04 -1.45432800e-01 4.58438158e-01 5.22244453e-01
6.74672797e-02 2.37165183e-01 8.87070537e-01 2.59213056e-02
-8.20576847e-02 8.14366758e-01 -2.17134580e-01 -3.99114311e-01
8.99321318e-01 -3.61546457e-01 -2.45012477e-01 1.98207617e-01
-7.80494153e-01 -4.84772474e-02 3.95238698e-01 -1.50266334e-01
7.65011430e-01 -1.39934516e+00 -6.38737559e-01 1.31434754e-01
5.14672279e-01 2.47944910e-02 3.24574895e-02 7.46329427e-01
-9.08777237e-01 9.23713386e-01 -4.10643458e-01 -4.48892653e-01
-1.38274097e+00 3.49452078e-01 7.86388695e-01 -2.61960357e-01
-2.30708703e-01 7.60286272e-01 5.82045734e-01 -2.34107286e-01
1.62935302e-01 -2.14648858e-01 -1.57401770e-01 -2.84510732e-01
4.91505176e-01 -9.18157920e-02 5.37153900e-01 -6.56454563e-01
-7.30489612e-01 4.20827627e-01 -5.31414568e-01 7.88339615e-01
1.81051540e+00 8.57635260e-01 -8.70350450e-02 -3.11032683e-01
1.34659564e+00 -3.63871872e-01 -1.17670596e+00 -1.78769991e-01
-5.41671738e-02 -2.55614012e-01 -2.18920573e-03 -1.05824709e+00
-6.52990639e-01 6.42457902e-01 7.97786295e-01 -6.90024078e-01
5.67938805e-01 3.63416672e-02 6.69639885e-01 9.62895751e-01
3.74438107e-01 -6.21056020e-01 1.14805333e-01 5.78788579e-01
1.05201113e+00 -1.37855661e+00 3.42467695e-01 -1.70665875e-01
-3.49418879e-01 1.12542641e+00 8.05163324e-01 1.51261315e-01
4.78924692e-01 -1.86281241e-02 -3.11904848e-01 -8.01299214e-01
-8.92405391e-01 8.53641927e-02 6.63522363e-01 5.81311345e-01
1.07895875e+00 -1.33984476e-01 -5.34970701e-01 9.12356853e-01
2.74031281e-01 -9.15972143e-02 1.74082327e-03 8.32602024e-01
-6.32379651e-01 -1.60109544e+00 -1.39451474e-01 3.14250320e-01
-6.58460915e-01 -5.18362939e-01 -1.17508686e+00 7.24679410e-01
3.72616351e-02 4.35647339e-01 -4.82051015e-01 -3.52405608e-01
5.80160797e-01 -5.01199476e-02 6.39611483e-01 -6.72372818e-01
-7.63169229e-01 3.17313224e-02 -1.29129946e-01 -5.06402373e-01
-2.19528973e-01 -3.65795307e-02 -1.73957801e+00 -5.60358405e-01
-6.67908072e-01 2.70303786e-01 5.63185930e-01 7.44775891e-01
8.21060419e-01 3.23480815e-01 3.51149738e-01 -1.06329072e+00
-3.64198864e-01 -8.87848020e-01 -3.18842858e-01 8.51206630e-02
2.45692357e-01 -9.75935221e-01 2.82939762e-01 -2.90403366e-01] | [4.907860279083252, 5.65952205657959] |
f2271078-fab2-4158-aac6-463610f47972 | convolution-free-waveform-transformers-for | 2109.15129 | null | https://arxiv.org/abs/2109.15129v1 | https://arxiv.org/pdf/2109.15129v1.pdf | Convolution-Free Waveform Transformers for Multi-Lead ECG Classification | We present our entry to the 2021 PhysioNet/CinC challenge - a waveform transformer model to detect cardiac abnormalities from ECG recordings. We compare the performance of the waveform transformer model on different ECG-lead subsets using approximately 88,000 ECG recordings from six datasets. In the official rankings, team prna ranked between 9 and 15 on 12, 6, 4, 3 and 2-lead sets respectively. Our waveform transformer model achieved an average challenge metric of 0.47 on the held-out test set across all ECG-lead subsets. Our combined performance across all leads placed us at rank 11 out of 39 officially ranking teams. | ['Jonathan Rubin', 'Corneliu Antonescu', 'Yale Chang', 'Gregory Boverman', 'Annamalai Natarajan'] | 2021-09-29 | null | null | null | null | ['ecg-classification'] | ['medical'] | [ 1.76158831e-01 -2.54352391e-01 7.59927481e-02 -2.76868284e-01
-1.66995168e+00 -1.07650602e+00 -2.91947693e-01 3.09796304e-01
-8.94917548e-02 8.97186041e-01 3.12663913e-01 -3.32286745e-01
-8.62384677e-01 -6.46342710e-02 -2.54926473e-01 -1.59954578e-01
-9.67826724e-01 5.81325889e-01 9.54899415e-02 6.70097843e-02
-1.34168267e-02 1.22545458e-01 -3.02317053e-01 7.11178064e-01
5.49421310e-01 1.19926071e+00 -8.94557953e-01 1.35489511e+00
8.21104527e-01 4.77135628e-01 -8.54422271e-01 -3.24818164e-01
2.61835784e-01 -9.29137766e-01 -5.52298248e-01 -6.65955782e-01
5.10411859e-01 1.15083531e-01 -3.62214804e-01 2.73362190e-01
1.45352721e+00 -4.73446578e-01 4.58326340e-01 -9.21794653e-01
1.23708397e-01 9.20000732e-01 -3.59829515e-01 1.09083688e+00
5.48243284e-01 1.77587256e-01 1.00052357e+00 -7.84902871e-01
6.72688544e-01 1.47906587e-01 1.59142709e+00 8.39344487e-02
-1.43288577e+00 -3.16676825e-01 -5.81860960e-01 1.08756639e-01
-1.72206318e+00 -2.52620548e-01 5.73822737e-01 -3.61987084e-01
8.50408614e-01 8.41239870e-01 8.95091355e-01 8.20493758e-01
4.75432217e-01 5.79802804e-02 1.23836386e+00 7.69521296e-02
-2.34024405e-01 -4.26447183e-01 3.97316575e-01 1.25558540e-01
5.20112336e-01 -1.03266165e-01 -9.00177419e-01 -1.06145144e+00
5.04827261e-01 -5.80890536e-01 -4.86748368e-01 2.47368813e-01
-1.63396883e+00 5.98878078e-02 -2.23802552e-01 -2.59071104e-02
-5.01963913e-01 3.41051333e-02 7.73953438e-01 6.24035001e-01
2.80714393e-01 9.23054874e-01 -7.44806468e-01 -6.29096806e-01
-1.09128225e+00 4.62447405e-01 8.62965047e-01 5.61691940e-01
-2.42014065e-01 -9.05999616e-02 -6.53198957e-01 6.90097809e-01
-2.65540723e-02 6.15166843e-01 1.33845946e-02 -1.15874040e+00
4.14143294e-01 1.34521991e-01 1.74874514e-01 -9.40229237e-01
-1.01621199e+00 -1.05979621e+00 -9.49242413e-01 -4.59622353e-01
3.47824246e-01 -4.85398412e-01 -5.68187118e-01 1.29611778e+00
-1.78826153e-01 3.30656141e-01 -2.17142150e-01 9.33305740e-01
1.11322689e+00 9.57526863e-02 -1.75546303e-01 -4.74941283e-01
1.42645526e+00 -9.83464420e-02 -4.79254067e-01 1.36615619e-01
4.02334452e-01 -6.15238190e-01 5.38897395e-01 1.15259731e+00
-1.36819375e+00 -4.31516021e-01 -1.06644213e+00 4.35949922e-01
5.60145080e-01 1.14874661e-01 2.63668112e-02 6.83089614e-01
-1.24427378e+00 1.02081227e+00 -6.04205072e-01 -1.24381222e-01
5.13317406e-01 2.84551740e-01 -2.12212086e-01 2.27187097e-01
-1.40305984e+00 7.52158403e-01 -2.29514018e-01 -6.35111183e-02
-9.85030890e-01 -1.21349776e+00 -3.69820684e-01 -3.31676930e-01
-1.76573887e-01 -5.77062666e-01 8.50342214e-01 3.68342668e-01
-7.24530339e-01 1.05026269e+00 -2.47283075e-02 -7.12649345e-01
8.19177091e-01 -2.61186272e-01 -9.47856724e-01 4.00383592e-01
2.52576977e-01 -3.02183270e-01 2.02896595e-01 -7.15111315e-01
-5.03066003e-01 -3.81427258e-01 -3.75171483e-01 6.39696093e-03
4.06348169e-01 1.29989624e-01 -2.19395459e-01 -6.62326097e-01
4.33799207e-01 -8.23835909e-01 -3.72520238e-01 -5.25489986e-01
-6.12653375e-01 2.52710789e-01 2.36731535e-03 -8.87010217e-01
1.79015934e+00 -2.23418403e+00 2.64648423e-02 6.39754832e-01
8.68023455e-01 -1.26905978e-01 9.55963805e-02 4.66885775e-01
-2.75996357e-01 4.91425216e-01 -8.39994699e-02 2.44788043e-02
-3.83579016e-01 -6.80470169e-02 -1.82247877e-01 5.96893907e-01
6.18046820e-02 1.02741146e+00 -6.34986818e-01 -3.84051085e-01
-1.93891659e-01 2.94101089e-01 -3.56177419e-01 -1.63899094e-01
6.72414839e-01 8.60748172e-01 -1.45430565e-01 3.07217717e-01
5.00376999e-01 -4.61754948e-01 4.39050347e-01 -2.68030167e-01
2.11846739e-01 5.72805822e-01 -1.07173097e+00 1.75177312e+00
4.92697299e-01 5.16994655e-01 -4.22813535e-01 -5.04814386e-01
9.83283281e-01 9.29003000e-01 8.48806202e-01 -3.21866274e-01
-4.01874810e-01 4.94683027e-01 6.69178724e-01 -4.35235023e-01
-3.95069242e-01 -1.23847634e-01 -3.14560354e-01 5.63587427e-01
-1.74911454e-01 -2.12370511e-02 6.89830184e-02 3.05287540e-01
1.70384717e+00 -2.61559546e-01 2.78374791e-01 -6.22371376e-01
1.69514686e-01 -2.47544795e-02 7.93149948e-01 1.27508879e+00
-4.20634240e-01 1.48509634e+00 1.10484469e+00 -1.11397159e+00
-5.37532806e-01 -1.48501754e+00 -5.64606428e-01 4.38703954e-01
-4.74482447e-01 -1.02239692e+00 -3.68242204e-01 -4.59151685e-01
-4.64293957e-02 1.12359583e-01 -6.64198279e-01 -4.82263491e-02
-6.14896119e-01 -1.10017908e+00 1.78087401e+00 6.31778359e-01
-9.92898941e-02 -7.72637069e-01 -1.09225535e+00 5.18200696e-01
-6.51738942e-01 -8.42378139e-01 -4.21499342e-01 2.75892466e-01
-1.03459799e+00 -1.35824680e+00 -6.94222748e-01 -3.22746336e-01
-1.75892442e-01 -7.05399752e-01 1.67047143e+00 -9.34823528e-02
-7.78867483e-01 2.34560966e-01 -2.51478791e-01 -6.70260251e-01
-9.70197259e-04 1.06396407e-01 1.83554456e-01 -2.88335253e-02
1.43593252e-01 -7.33710349e-01 -9.84537184e-01 2.18246430e-01
1.23770922e-01 -5.50085127e-01 4.32052054e-02 6.89999104e-01
4.91456807e-01 -6.46987021e-01 1.15051496e+00 -8.21992517e-01
8.72622132e-01 -3.07620317e-01 -9.19754133e-02 6.25099018e-02
-9.86544192e-01 -7.80891836e-01 2.28853956e-01 9.02863964e-03
-5.05809067e-03 -1.75715402e-01 -1.73046350e-01 -5.05024195e-02
1.49450749e-01 5.86045742e-01 5.09639561e-01 2.26625875e-01
1.16917014e+00 2.34571651e-01 -2.16802061e-01 -2.16967925e-01
-4.34601068e-01 2.93971092e-01 8.59876513e-01 -6.45659566e-01
4.98030365e-01 1.96178645e-01 2.36316472e-01 -3.14616382e-01
-7.20071316e-01 -6.18874133e-01 -4.63410109e-01 -4.90487926e-03
7.74919629e-01 -8.30253541e-01 -5.48766613e-01 2.19586223e-01
-1.20542729e+00 -3.02067418e-02 -4.85555500e-01 4.08182502e-01
-9.84368548e-02 2.42709339e-01 -7.33144462e-01 -5.59299350e-01
-9.52017069e-01 -7.64409840e-01 9.51926410e-01 -5.37868202e-01
-1.01570213e+00 -7.29474545e-01 6.04839802e-01 -1.59404695e-01
5.64647138e-01 9.73277986e-01 6.13311708e-01 -8.38054121e-01
1.65788189e-01 -3.69176865e-01 1.21626377e-01 2.02049166e-01
2.22726926e-01 -2.07296327e-01 -9.26073849e-01 -1.11123592e-01
-1.83837324e-01 -1.00098424e-01 5.15346110e-01 6.19391620e-01
1.01438868e+00 3.62134606e-01 -2.98616409e-01 6.20509326e-01
1.01226187e+00 2.23266095e-01 8.64864111e-01 -2.80624568e-01
3.92152309e-01 -5.84346205e-02 2.64891267e-01 7.63768971e-01
1.86246559e-01 2.82382160e-01 -3.23914111e-01 -3.96219343e-01
2.31513783e-01 1.08542569e-01 -2.58862019e-01 8.00720870e-01
-6.19807184e-01 8.48796740e-02 -1.47893977e+00 5.77313662e-01
-1.50854111e+00 -6.70286536e-01 -9.02124226e-01 2.22434878e+00
8.82771909e-01 4.99584407e-01 4.68366563e-01 5.59688151e-01
3.62473607e-01 -9.40488577e-02 -5.71090877e-01 -3.37854385e-01
-3.32300037e-01 6.77959561e-01 1.66123286e-01 1.60683855e-01
-1.01465201e+00 -1.93500191e-01 8.19434071e+00 9.16990936e-02
-9.21687782e-01 2.95545548e-01 9.24037933e-01 -4.08098549e-01
9.72388685e-02 -1.83511913e-01 -2.56551713e-01 3.29785377e-01
1.35953844e+00 -5.55584311e-01 3.98414135e-02 1.81848317e-01
2.06355423e-01 2.49709964e-01 -1.32555044e+00 1.46225488e+00
7.52145797e-02 -1.53737152e+00 -4.18897122e-01 -7.03130942e-03
5.90882719e-01 5.14556646e-01 -2.12480441e-01 4.72770669e-02
-5.91981351e-01 -1.42306018e+00 4.86799449e-01 7.41862357e-01
1.80548382e+00 -4.82947379e-01 9.33710277e-01 -5.07541448e-02
-1.16191387e+00 2.03197896e-01 1.09165102e-01 -1.42427713e-01
1.74822032e-01 7.36487508e-01 -6.63330555e-01 8.02066863e-01
1.03951061e+00 7.76866615e-01 -6.60537183e-01 1.34036243e+00
1.44412518e-01 1.29066896e+00 -5.19417703e-01 6.76736653e-01
-5.02909482e-01 5.08477509e-01 8.68362188e-01 1.33228374e+00
9.93987545e-03 5.65576851e-01 2.09889665e-01 3.48809898e-01
3.62363197e-02 -6.99887127e-02 -4.94648904e-01 3.63348395e-01
4.71007049e-01 1.20746613e+00 -4.86911863e-01 -5.88701308e-01
5.17757423e-02 2.42745280e-01 -2.64390767e-01 1.89364001e-01
-1.05936408e+00 -9.14760590e-01 8.80021825e-02 3.97371501e-01
-4.57211733e-01 3.54213387e-01 -9.40006554e-01 -6.70010149e-01
2.51012057e-01 -1.39582157e+00 8.35018218e-01 -4.17099327e-01
-1.45783591e+00 1.07926250e+00 -1.80857927e-01 -1.45312095e+00
-1.84704453e-01 -2.88582623e-01 -6.80512846e-01 1.29067528e+00
-7.67759800e-01 -2.40380079e-01 -3.11307818e-01 4.84082550e-01
-8.51253197e-02 7.40727261e-02 1.43528020e+00 7.58562088e-01
6.50547519e-02 9.49775159e-01 -6.29536092e-01 3.39871109e-01
1.02571821e+00 -1.53472173e+00 7.20420182e-01 3.81570756e-01
4.23438847e-02 7.43419766e-01 4.99952316e-01 -3.88569385e-01
-1.10445440e+00 -8.81832361e-01 1.14691174e+00 -1.32013690e+00
3.43328476e-01 6.60448596e-02 -6.82099581e-01 5.60141504e-01
1.20841376e-01 3.52677852e-01 1.06371176e+00 4.47308868e-01
-1.30417347e-02 -4.36014503e-01 -8.25260282e-01 -9.57515314e-02
1.09624386e+00 -6.63277924e-01 -7.02575803e-01 1.61399379e-01
4.57342118e-02 -8.32604825e-01 -1.67394483e+00 9.43948805e-01
1.29896200e+00 -7.03009665e-01 7.62621403e-01 -9.02394950e-01
2.14225009e-01 -2.35213742e-01 2.33215436e-01 -1.11933374e+00
-5.18162191e-01 -1.53243601e+00 1.09984823e-01 6.64542198e-01
8.71368170e-01 -8.46648514e-01 4.69049960e-01 8.78207833e-02
-2.65923142e-01 -1.12220335e+00 -9.04213607e-01 -4.42234278e-01
-9.64918658e-02 -5.48827767e-01 8.93659070e-02 8.50886106e-01
3.18106979e-01 4.55790132e-01 -4.47571695e-01 -7.12621659e-02
7.28951395e-01 -2.70951260e-02 2.72481948e-01 -1.57034194e+00
-6.96589768e-01 -3.34088132e-02 -4.24993306e-01 -2.71406293e-01
-1.00396717e+00 -1.08524382e+00 -3.19330961e-01 -1.62844479e+00
4.62536305e-01 -5.07132173e-01 -1.21142077e+00 4.34670031e-01
-3.14982384e-01 1.17091882e+00 -6.54903427e-02 1.42154694e-01
-7.74399519e-01 -5.35989463e-01 8.12942207e-01 1.43704996e-01
-1.58693105e-01 1.17545657e-01 -1.07742190e+00 5.97479522e-01
7.76717901e-01 -8.44148338e-01 -3.90044302e-01 -1.54803365e-01
6.49706244e-01 5.86447001e-01 3.42434138e-01 -1.30454040e+00
-2.71436144e-02 5.23528397e-01 7.63320088e-01 -8.14134121e-01
1.12926722e-01 -5.19764200e-02 6.41051054e-01 8.78805935e-01
-4.67504174e-01 6.69471920e-01 4.40395802e-01 1.65152758e-01
-6.57526553e-02 6.92426085e-01 4.81434494e-01 -8.46347511e-02
2.77486891e-01 2.90953398e-01 -2.68188119e-01 1.17213619e+00
5.56178033e-01 -1.42023936e-01 -2.97000140e-01 -1.06133126e-01
-1.51705599e+00 4.48951155e-01 -4.43903238e-01 3.31911832e-01
6.93680763e-01 -1.14321828e+00 -1.67002654e+00 -2.45374301e-03
3.80771488e-01 -2.47860461e-01 2.93675005e-01 1.70950484e+00
-7.64741421e-01 3.63817096e-01 -2.88888216e-01 -1.21262002e+00
-1.17342317e+00 -3.89079094e-01 6.98288083e-01 -5.63164711e-01
-9.29045558e-01 8.23318958e-01 -3.23858351e-01 1.17123313e-01
5.01715578e-02 -1.89046219e-01 1.80060610e-01 -4.10108641e-02
7.28116393e-01 7.73281455e-01 5.02806365e-01 7.29826540e-02
-1.16150165e+00 6.32132709e-01 2.22099885e-01 -2.93030709e-01
9.20502365e-01 2.23018169e-01 -1.83518842e-01 7.99053967e-01
8.80315125e-01 1.44154206e-01 -3.31720859e-01 2.74648219e-01
7.05579743e-02 1.04816379e-02 -4.23467875e-01 -1.32552278e+00
-5.86406410e-01 7.99600184e-01 8.76262903e-01 2.94142783e-01
1.09619880e+00 -3.08415622e-01 6.62971258e-01 8.73646960e-02
6.79572880e-01 -5.86502552e-01 -2.18339562e-01 1.87485203e-01
9.19758022e-01 -4.92878437e-01 2.19701424e-01 -9.01116654e-02
-8.54778230e-01 8.06401312e-01 -3.54074277e-02 -3.72454643e-01
8.44392657e-01 4.93444443e-01 4.76919055e-01 -5.50176084e-01
-1.02502310e+00 6.22691810e-01 5.28382540e-01 6.66666925e-01
8.53673697e-01 3.43409866e-01 -7.64118791e-01 1.03699446e+00
-7.39403516e-02 3.53807956e-01 4.92178440e-01 6.83478057e-01
7.34179392e-02 -9.52059984e-01 -4.33148630e-02 1.01018369e+00
-1.27467740e+00 -3.37724626e-01 -3.91284645e-01 3.13325584e-01
9.26934630e-02 1.15477717e+00 -3.82905662e-01 -4.66315836e-01
6.82884812e-01 4.10825968e-01 5.72903752e-01 -5.83813488e-01
-1.14539933e+00 4.33932990e-01 5.43409646e-01 -6.83627725e-01
1.55029714e-01 -9.42092478e-01 -1.13714063e+00 2.54087985e-01
1.55101925e-01 4.31134850e-01 1.02537096e-01 4.63325292e-01
8.52123439e-01 9.73963141e-01 4.60750669e-01 -5.49849905e-02
-8.44711840e-01 -1.17127395e+00 -8.39363277e-01 1.56162545e-01
5.53718865e-01 -9.30352733e-02 -3.32628459e-01 2.20145285e-01] | [14.36038875579834, 3.2723824977874756] |
192f2047-775e-476b-812c-31a3f5d0aee7 | assembled-openml-creating-efficient | 2307.00285 | null | https://arxiv.org/abs/2307.00285v1 | https://arxiv.org/pdf/2307.00285v1.pdf | Assembled-OpenML: Creating Efficient Benchmarks for Ensembles in AutoML with OpenML | Automated Machine Learning (AutoML) frameworks regularly use ensembles. Developers need to compare different ensemble techniques to select appropriate techniques for an AutoML framework from the many potential techniques. So far, the comparison of ensemble techniques is often computationally expensive, because many base models must be trained and evaluated one or multiple times. Therefore, we present Assembled-OpenML. Assembled-OpenML is a Python tool, which builds meta-datasets for ensembles using OpenML. A meta-dataset, called Metatask, consists of the data of an OpenML task, the task's dataset, and prediction data from model evaluations for the task. We can make the comparison of ensemble techniques computationally cheaper by using the predictions stored in a metatask instead of training and evaluating base models. To introduce Assembled-OpenML, we describe the first version of our tool. Moreover, we present an example of using Assembled-OpenML to compare a set of ensemble techniques. For this example comparison, we built a benchmark using Assembled-OpenML and implemented ensemble techniques expecting predictions instead of base models as input. In our example comparison, we gathered the prediction data of $1523$ base models for $31$ datasets. Obtaining the prediction data for all base models using Assembled-OpenML took ${\sim} 1$ hour in total. In comparison, obtaining the prediction data by training and evaluating just one base model on the most computationally expensive dataset took ${\sim} 37$ minutes. | ['Joeran Beel', 'Lennart Purucker'] | 2023-07-01 | null | null | null | null | ['automl'] | ['methodology'] | [-8.02801400e-02 -1.24451388e-02 4.84623134e-01 -8.41714382e-01
-1.12009108e+00 -5.92703581e-01 4.58246619e-01 3.30434054e-01
-1.72397941e-01 1.04545450e+00 -4.70986515e-01 -4.66994673e-01
-1.61874101e-01 -7.62335002e-01 -8.14919531e-01 -6.03493214e-01
-9.21574309e-02 7.32242346e-01 -1.20912097e-01 -3.24198194e-02
1.36891797e-01 -8.85522887e-02 -2.21625423e+00 8.77479434e-01
1.44915378e+00 1.09240258e+00 2.89134055e-01 7.69962013e-01
-3.71081650e-01 7.53929079e-01 -8.16449523e-01 -4.80867535e-01
1.33514076e-01 -4.30614710e-01 -5.96288204e-01 -5.05434394e-01
6.59522831e-01 2.15213105e-01 6.79975569e-01 5.64170599e-01
6.45472169e-01 1.04158662e-01 6.16926432e-01 -1.53430486e+00
3.10200423e-01 7.33272552e-01 1.84785187e-01 -2.50522763e-01
5.02679288e-01 2.95458198e-01 8.88419211e-01 -8.58593583e-01
5.20884275e-01 8.52087498e-01 8.61352146e-01 4.42324668e-01
-1.43067276e+00 -8.36498141e-01 -1.33311108e-01 1.14532188e-01
-1.20214224e+00 -5.30061305e-01 3.52006763e-01 -7.65460193e-01
1.25395644e+00 8.61674964e-01 5.78787148e-01 1.22629213e+00
3.99545223e-01 4.45982337e-01 1.40791392e+00 -3.68764907e-01
5.11589050e-01 3.53374302e-01 4.57326263e-01 6.99594021e-01
1.49564222e-01 1.68316085e-02 -4.27076966e-01 -5.33642590e-01
-5.08919396e-02 -3.78172725e-01 -4.41131964e-02 2.98094638e-02
-9.54880953e-01 3.06034476e-01 -1.14627786e-01 1.97664395e-01
-1.49218783e-01 1.06856495e-01 4.33120817e-01 4.87310559e-01
7.15754926e-01 6.06665492e-01 -1.01462245e+00 -5.30493796e-01
-9.40832198e-01 4.77287382e-01 1.24550593e+00 9.18524206e-01
1.20282888e+00 -1.17063552e-01 3.47704217e-02 9.25223589e-01
1.67173341e-01 2.93233603e-01 4.67981964e-01 -9.42007780e-01
6.52690411e-01 9.88854110e-01 2.55512092e-02 -3.72627020e-01
-2.74994105e-01 -3.49764764e-01 -6.51372313e-01 4.34706450e-01
2.43153989e-01 -4.47309047e-01 -5.70295632e-01 1.46029794e+00
3.17257851e-01 2.57755280e-01 1.36061460e-01 2.23858848e-01
9.81464207e-01 6.72280788e-01 1.21876985e-01 -4.54346359e-01
7.74405539e-01 -9.04209137e-01 -3.73365909e-01 -1.82429850e-02
1.27993023e+00 -6.81213081e-01 9.74172056e-01 6.01484418e-01
-7.55476594e-01 -7.71993458e-01 -1.11202216e+00 2.74147779e-01
-6.84647381e-01 1.34807929e-01 3.22169900e-01 6.08291805e-01
-9.20215309e-01 1.00048089e+00 -8.38703811e-01 -6.49090931e-02
6.27023876e-02 3.85770738e-01 -4.23672825e-01 8.55900645e-02
-9.39357758e-01 1.07678902e+00 5.25751889e-01 -1.95191041e-01
-9.80246603e-01 -9.36581790e-01 -7.88892210e-01 -7.85504133e-02
5.66350967e-02 -5.89679420e-01 1.24395502e+00 -9.71214652e-01
-1.21972275e+00 5.74505329e-01 -1.53777212e-01 -2.31548682e-01
5.10707378e-01 -2.83700198e-01 -5.83494186e-01 -9.75882471e-01
-1.41053706e-01 1.86201170e-01 4.53774661e-01 -1.40388060e+00
-6.32896304e-01 -3.93155754e-01 -7.50334337e-02 -1.33228630e-01
1.01242557e-01 -1.58212222e-02 6.46517575e-02 -2.54025698e-01
-3.11292738e-01 -8.90734315e-01 -3.00080568e-01 -6.99896514e-01
-3.88283670e-01 -4.46987480e-01 6.88179851e-01 -6.31364822e-01
1.73573232e+00 -1.97734511e+00 9.31724682e-02 1.67435974e-01
4.94296476e-02 1.07876107e-01 -2.25054041e-01 5.44306397e-01
-2.97969580e-01 2.67546266e-01 -6.39805019e-01 -4.97688830e-01
9.66621712e-02 2.29204863e-01 -6.71302602e-02 -1.98399112e-01
-1.09276615e-01 4.20591265e-01 -6.46267354e-01 -7.35541046e-01
2.27168694e-01 3.78211796e-01 -4.11935657e-01 3.48168790e-01
-7.90071547e-01 4.57507044e-01 -3.22233975e-01 5.00578225e-01
5.73804379e-01 6.08749650e-02 3.80163640e-01 -7.64979646e-02
-2.61660069e-01 3.88940275e-01 -1.16877818e+00 1.87525403e+00
-1.01300049e+00 5.00376046e-01 -3.92230630e-01 -6.43661320e-01
1.11007333e+00 5.15853941e-01 4.85541761e-01 -1.06904961e-01
-1.78595744e-02 7.37772703e-01 3.22025418e-01 -5.47622681e-01
-1.07367942e-02 2.82122642e-01 -2.27144748e-01 9.48979139e-01
4.26150143e-01 -5.81027925e-01 6.11636698e-01 -9.60291177e-02
1.39589369e+00 7.85996795e-01 3.82936448e-01 -2.17430994e-01
5.76527417e-01 5.88303693e-02 6.20613873e-01 6.61996841e-01
4.50760603e-01 1.92335933e-01 2.89336324e-01 -9.64455247e-01
-9.57177877e-01 -8.13959718e-01 -2.84833670e-01 1.36045468e+00
-6.13310456e-01 -1.18776679e+00 -1.03711367e+00 -1.01796198e+00
-1.78547502e-01 1.33899379e+00 -4.88346279e-01 2.63832182e-01
-4.15801853e-01 -8.37778986e-01 3.63043159e-01 1.61587194e-01
2.85913676e-01 -1.15515101e+00 -8.09777021e-01 2.91241527e-01
-2.85647303e-01 -6.47774756e-01 5.70293628e-02 4.11801666e-01
-9.44331288e-01 -1.19598448e+00 2.29741082e-01 -2.89726824e-01
5.14555752e-01 -6.72073185e-01 1.75461447e+00 1.97029710e-01
-2.30237067e-01 1.03107288e-01 -4.36485946e-01 -9.19692516e-01
-9.50016618e-01 3.03021103e-01 1.02593556e-01 -1.12231016e-01
3.06873262e-01 -9.65433836e-01 -8.09056871e-03 3.05665910e-01
-6.80723429e-01 4.93656665e-01 2.84439564e-01 7.02759087e-01
5.51875532e-01 -4.47840691e-02 3.44718039e-01 -1.11849630e+00
5.32591760e-01 -5.27538717e-01 -7.17366874e-01 7.69623220e-01
-6.61636591e-01 2.78614700e-01 7.88534582e-01 3.17605771e-02
-9.63861644e-01 2.92451918e-01 -3.04057479e-01 -1.94045171e-01
-3.26772094e-01 8.03364575e-01 -2.31144860e-01 3.47486258e-01
9.74085033e-01 1.54115692e-01 -8.44057798e-02 -6.87451303e-01
5.08099720e-02 9.52680588e-01 -6.06825352e-02 -1.01290452e+00
2.91901112e-01 -3.28348547e-01 -2.58774698e-01 -3.38189572e-01
-7.86896706e-01 2.99051225e-01 -6.78212166e-01 -3.64416182e-01
4.26659137e-01 -6.31837606e-01 -6.17421567e-01 3.03316385e-01
-1.38484490e+00 -7.42376864e-01 -3.39653671e-01 2.74894744e-01
-6.63365781e-01 -3.48449975e-01 3.88315618e-02 -9.00105298e-01
-5.75307608e-01 -1.46486795e+00 9.03240979e-01 -2.73861110e-01
-5.52620769e-01 -9.22039509e-01 6.22447789e-01 1.47820398e-01
4.71419096e-01 4.44042563e-01 1.01568830e+00 -1.13122344e+00
-3.23858082e-01 -4.52661484e-01 4.05906498e-01 6.59948707e-01
9.78024304e-02 5.80417275e-01 -1.31707907e+00 -7.90263638e-02
-1.19862847e-01 -3.32508534e-01 5.48964679e-01 -7.93463290e-02
1.74322820e+00 -2.31914297e-01 -4.71327156e-01 3.95489633e-01
1.40436924e+00 2.50585139e-01 5.11376023e-01 2.44575255e-02
4.98317540e-01 5.85877717e-01 6.21302485e-01 5.34304559e-01
3.99259537e-01 6.67612135e-01 3.71123493e-01 2.85576701e-01
3.44649434e-01 1.17910780e-01 5.97452641e-01 9.70804036e-01
-6.16502881e-01 -1.60457147e-03 -1.30295300e+00 -4.18329127e-02
-1.98724473e+00 -9.21322882e-01 -2.84148455e-01 2.59787679e+00
1.11137867e+00 -1.50031611e-01 -1.38122350e-01 1.73802003e-01
3.95999849e-01 -1.04476005e-01 -4.64150220e-01 -7.40115106e-01
1.28905654e-01 6.11355126e-01 -3.32343459e-01 6.16815925e-01
-1.02531278e+00 4.72433954e-01 5.79000139e+00 9.66989636e-01
-9.88839388e-01 2.76525229e-01 6.13441765e-01 -4.12905246e-01
-4.07849491e-01 2.05396563e-01 -8.20869803e-01 8.67266417e-01
1.46879852e+00 -3.41910630e-01 2.77781457e-01 1.12865472e+00
-6.15247749e-02 -3.36258471e-01 -1.88513970e+00 9.75621998e-01
9.45660938e-03 -1.29946744e+00 -1.66648537e-01 4.95482832e-02
8.98853660e-01 2.18975127e-01 -4.61111724e-01 6.84698522e-01
7.11516321e-01 -1.14263344e+00 5.19728780e-01 7.85598636e-01
8.14893484e-01 -5.88115335e-01 8.37567031e-01 8.08221042e-01
-9.27182376e-01 -1.60758555e-01 -2.85168663e-02 -2.40240127e-01
-2.15073243e-01 1.05584383e+00 -6.15936577e-01 9.31307077e-01
8.98675680e-01 6.80026472e-01 -7.88862467e-01 8.35478663e-01
-9.81675461e-02 5.40607154e-01 -4.24702018e-01 6.22165687e-02
-2.94270903e-01 -3.15311074e-01 3.20544034e-01 1.13831425e+00
7.19584584e-01 -2.32423499e-01 1.71158612e-01 8.11798453e-01
2.29975283e-01 2.78238893e-01 -8.64618778e-01 2.15731993e-01
4.98777598e-01 1.40028751e+00 1.87506139e-01 -5.32690167e-01
-1.51135117e-01 5.14681637e-01 3.45571727e-01 1.70477014e-02
-9.43124890e-01 -3.06069165e-01 5.70355117e-01 9.79153290e-02
-3.09306383e-01 9.06403437e-02 -4.41276640e-01 -1.01328743e+00
1.19246408e-01 -1.23888254e+00 2.76906431e-01 -8.98831129e-01
-1.20749545e+00 1.09985363e+00 2.91819815e-02 -1.43611276e+00
-5.99397540e-01 -3.88900369e-01 -7.08899319e-01 1.11467969e+00
-5.70394635e-01 -9.04897451e-01 -7.47719824e-01 2.12838948e-01
6.72913313e-01 -3.51135582e-01 1.54827547e+00 1.36328042e-01
-8.22983027e-01 3.68758082e-01 5.50151579e-02 -3.36142331e-01
8.42624664e-01 -1.26395130e+00 2.72237718e-01 3.50201994e-01
1.36966974e-01 4.60612178e-01 7.43493438e-01 -5.45802593e-01
-1.04172468e+00 -1.25876141e+00 9.73496377e-01 -1.04650438e+00
3.53402704e-01 -4.16862637e-01 -8.97488177e-01 9.45307791e-01
3.92372876e-01 3.07392292e-02 1.12943625e+00 3.15216690e-01
-2.22079605e-01 -4.97276336e-01 -9.86817002e-01 2.91692555e-01
9.56270874e-01 -1.25780150e-01 -2.70719677e-01 4.20753062e-01
4.17169631e-01 -5.79792559e-01 -1.32192230e+00 7.37324119e-01
8.32812667e-01 -1.36853731e+00 3.70618612e-01 -4.68849570e-01
7.91365802e-01 -2.54858613e-01 -4.76191968e-01 -1.79064405e+00
3.54921013e-01 -2.85030752e-01 -1.98917553e-01 1.22077322e+00
1.08108759e+00 -8.75150919e-01 2.76616454e-01 8.24592769e-01
-3.91375184e-01 -1.26204073e+00 -7.98292756e-01 -7.28089213e-01
3.05280685e-02 -1.02521467e+00 1.06967056e+00 9.48895156e-01
-8.52899551e-02 1.17145613e-01 3.55593488e-02 -8.72319564e-02
6.17086172e-01 1.95941940e-01 1.21325946e+00 -1.59300423e+00
-4.79165584e-01 -1.17245257e-01 -5.66617101e-02 -9.73636582e-02
3.35232496e-01 -1.06249642e+00 3.48458737e-02 -1.44696295e+00
1.38529062e-01 -1.12608516e+00 -3.06299835e-01 7.57146060e-01
1.92427766e-02 -2.22198337e-01 1.52304366e-01 1.40877277e-01
-4.43560660e-01 1.88143522e-01 6.80158079e-01 -1.53681964e-01
-2.79122978e-01 1.13614298e-01 -3.17357540e-01 7.85430908e-01
8.10451150e-01 -6.20084226e-01 -3.61192167e-01 -5.47130764e-01
5.39483070e-01 3.15844305e-02 1.99180037e-01 -1.54510009e+00
-1.51420133e-02 -6.54345974e-02 2.84141243e-01 -5.34059882e-01
3.75578195e-01 -7.89203644e-01 9.90702331e-01 2.14008033e-01
-1.63944200e-01 8.03249851e-02 2.48505145e-01 -4.86053377e-02
-2.25229234e-01 -3.79610270e-01 3.33380818e-01 -2.15240195e-01
-3.71996522e-01 1.23546705e-01 -6.04192354e-02 -2.62781531e-01
1.29269624e+00 -2.01401755e-01 -3.58474225e-01 1.15963794e-01
-1.18619049e+00 2.71724463e-01 5.06912291e-01 2.85860181e-01
3.06557745e-01 -9.96296406e-01 -1.06603050e+00 4.52589810e-01
1.24065928e-01 1.93907276e-01 2.91874766e-01 8.62519145e-01
-5.27092934e-01 1.27185270e-01 -1.54272392e-01 -6.76577747e-01
-1.43731725e+00 1.27949879e-01 6.24315321e-01 -6.31781816e-01
8.39725286e-02 7.72909045e-01 -2.26898342e-01 -1.18588591e+00
-1.83400095e-01 -1.55598387e-01 1.71940122e-02 3.69237773e-02
3.69305491e-01 5.55858135e-01 7.06054151e-01 -1.05972230e-01
-3.72391313e-01 4.12624538e-01 1.51011378e-01 -1.14542656e-01
1.48741341e+00 5.85063338e-01 -5.62149286e-01 1.03937447e+00
8.02512467e-01 -2.15874046e-01 -9.03046250e-01 2.42898703e-01
5.45742474e-02 -1.30395681e-01 -3.95905018e-01 -1.26548016e+00
-6.70158029e-01 9.86625314e-01 5.44545412e-01 2.80653059e-01
1.12143028e+00 -1.73039123e-01 4.41667773e-02 4.93861347e-01
1.05354536e+00 -1.02786100e+00 -4.79759395e-01 5.93795538e-01
1.28274429e+00 -1.18830466e+00 1.69422895e-01 -1.99073970e-01
-5.03483772e-01 1.07504904e+00 1.01860607e+00 1.45538554e-01
8.32898200e-01 4.92756754e-01 9.16363820e-02 -2.43524741e-03
-1.54756761e+00 2.80815572e-01 1.58579752e-01 4.00335610e-01
8.71061087e-01 3.19364697e-01 -3.31579447e-01 8.37984145e-01
-5.44804335e-01 2.53776580e-01 2.41445854e-01 9.14738834e-01
-2.03217223e-01 -1.60924268e+00 -3.59763563e-01 1.03756607e+00
-6.78297505e-02 -2.81227380e-01 -2.21389741e-01 5.66007018e-01
7.00861037e-01 8.42048645e-01 2.41901442e-01 -8.22070420e-01
3.21397007e-01 7.73403823e-01 6.06020510e-01 -8.88741314e-01
-9.99784827e-01 -5.87257743e-01 5.94673157e-01 -7.23986387e-01
-2.20159590e-01 -6.02153540e-01 -8.78252923e-01 -1.97662964e-01
-3.04018319e-01 4.06208187e-01 1.10365391e+00 1.07220352e+00
6.55875862e-01 4.78855878e-01 6.11853898e-01 -1.04876459e+00
-3.82732451e-01 -1.43169081e+00 -3.45259964e-01 8.31812993e-02
-3.51327240e-01 -6.05754912e-01 -3.62570882e-01 2.02446356e-01] | [8.762022018432617, 6.9893341064453125] |
20e8925b-9dfd-4834-8e70-efaa2e762cc1 | multi-task-transformer-with-relation | 2303.10870 | null | https://arxiv.org/abs/2303.10870v1 | https://arxiv.org/pdf/2303.10870v1.pdf | Multi-task Transformer with Relation-attention and Type-attention for Named Entity Recognition | Named entity recognition (NER) is an important research problem in natural language processing. There are three types of NER tasks, including flat, nested and discontinuous entity recognition. Most previous sequential labeling models are task-specific, while recent years have witnessed the rising of generative models due to the advantage of unifying all NER tasks into the seq2seq model framework. Although achieving promising performance, our pilot studies demonstrate that existing generative models are ineffective at detecting entity boundaries and estimating entity types. This paper proposes a multi-task Transformer, which incorporates an entity boundary detection task into the named entity recognition task. More concretely, we achieve entity boundary detection by classifying the relations between tokens within the sentence. To improve the accuracy of entity-type mapping during decoding, we adopt an external knowledge base to calculate the prior entity-type distributions and then incorporate the information into the model via the self and cross-attention mechanisms. We perform experiments on an extensive set of NER benchmarks, including two flat, three nested, and three discontinuous NER datasets. Experimental results show that our approach considerably improves the generative NER model's performance. | ['Zhoujun Li', 'Wei Wu', 'Jingang Wang', 'Zenglin Xu', 'Qifan Wang', 'Jiahao Liu', 'Hongyin Tang', 'Ying Mo'] | 2023-03-20 | null | null | null | null | ['boundary-detection'] | ['computer-vision'] | [ 3.61308712e-03 4.67531830e-02 5.17068431e-02 -6.71079636e-01
-1.09627759e+00 -6.56373501e-01 4.60236490e-01 -5.12305833e-02
-7.39448965e-01 7.68747389e-01 6.14052236e-01 -2.16528922e-01
4.75110650e-01 -8.85247052e-01 -5.74297547e-01 -4.63242322e-01
1.69385120e-01 5.13033867e-01 4.15268540e-02 -7.28124976e-02
1.07237160e-01 2.33902689e-02 -9.90800261e-01 2.80858368e-01
1.19730294e+00 5.94602168e-01 1.01079538e-01 6.03613794e-01
-6.59026206e-01 6.50562525e-01 -8.29871833e-01 -7.31749594e-01
-2.22197678e-02 -5.86348653e-01 -9.59459901e-01 -3.84494334e-01
2.26924624e-02 -5.76648675e-02 -1.54422924e-01 1.05738580e+00
8.21359515e-01 1.54800385e-01 5.31666398e-01 -1.04509413e+00
-9.39871073e-01 1.01673210e+00 -3.49639416e-01 2.87245274e-01
2.67831147e-01 -1.90684542e-01 1.22428548e+00 -9.92213070e-01
6.72533751e-01 1.00681376e+00 9.58304822e-01 8.52095008e-01
-1.00261319e+00 -5.26329935e-01 2.30535015e-01 6.05939850e-02
-1.60098767e+00 -4.28743631e-01 4.86808777e-01 -2.75778681e-01
1.15853381e+00 9.00393352e-02 8.97139683e-02 1.07590961e+00
1.56443432e-01 8.98018599e-01 9.45231855e-01 -3.55888546e-01
2.45483011e-01 -1.13313481e-01 4.12845701e-01 4.39423919e-01
1.88935608e-01 -1.96933359e-01 -3.21771502e-01 -1.69082642e-01
6.28703833e-01 -1.86257333e-01 -1.08237989e-01 1.28978819e-01
-1.14747918e+00 6.91759050e-01 4.12909538e-01 5.68837702e-01
-4.96132672e-01 -1.86394900e-01 3.97099048e-01 -5.79379797e-02
5.53166270e-01 2.00672925e-01 -5.17924130e-01 -2.91158885e-01
-9.90836680e-01 -1.29893094e-01 1.19658554e+00 1.27177012e+00
6.90243781e-01 -9.33555365e-02 -6.58850074e-01 9.64050472e-01
2.70929873e-01 3.82924676e-01 5.68403602e-01 -1.53471023e-01
7.76574612e-01 6.69296324e-01 1.03046671e-01 -5.83437681e-01
-4.71913725e-01 -3.92521352e-01 -9.42304492e-01 -4.98908728e-01
1.77340522e-01 -6.55426502e-01 -1.21598971e+00 1.76176202e+00
3.65945786e-01 1.79688558e-01 1.79970607e-01 6.58342719e-01
9.93156731e-01 7.99550831e-01 6.28204226e-01 1.43687367e-01
1.60339987e+00 -9.57134545e-01 -8.36085021e-01 -3.77202123e-01
7.93906808e-01 -6.02783203e-01 6.94120526e-01 -1.64420709e-01
-7.95628428e-01 -5.13778389e-01 -5.83187699e-01 -2.92338789e-01
-5.55672348e-01 3.87159139e-01 4.63575929e-01 8.37858677e-01
-9.31524992e-01 2.24167764e-01 -9.69338179e-01 -2.27236181e-01
3.13297421e-01 1.40116066e-01 -2.47039467e-01 5.05114421e-02
-1.54705012e+00 8.55733216e-01 6.48525655e-01 4.58447307e-01
-6.68111324e-01 -6.68003023e-01 -1.16620243e+00 3.34223598e-01
1.04560614e-01 -9.38377619e-01 1.34568369e+00 -5.86368740e-01
-1.13140762e+00 6.94388032e-01 -5.45185864e-01 -3.10689896e-01
2.50639021e-01 -3.03257585e-01 -5.90082705e-01 -4.07398194e-01
3.58453989e-01 6.91446364e-01 5.38726524e-02 -1.00109494e+00
-7.62962461e-01 -1.37749344e-01 -9.79238153e-02 3.30689430e-01
-2.06952378e-01 2.66790688e-01 -3.56581271e-01 -6.43178225e-01
-1.23768777e-01 -6.94713116e-01 -3.47730249e-01 -9.29931283e-01
-7.97752202e-01 -5.97536862e-01 1.98888838e-01 -8.85949671e-01
1.57282615e+00 -2.12568879e+00 -7.48309642e-02 -1.34705991e-01
1.33443074e-02 2.04670966e-01 -1.46992177e-01 4.41099018e-01
-1.97049584e-02 4.08092916e-01 -2.94671655e-01 -4.30989146e-01
2.77303487e-01 1.15421042e-01 -1.77237749e-01 -7.34014437e-02
5.34236252e-01 1.26109302e+00 -8.63017678e-01 -5.21333754e-01
-3.82892221e-01 6.95088863e-01 -4.71259475e-01 4.49071407e-01
-6.03282191e-02 2.33669266e-01 -4.46992904e-01 4.89015132e-01
5.77796996e-01 -1.77439943e-01 2.10098431e-01 -1.28591746e-01
-2.94076324e-01 8.89163971e-01 -1.14481711e+00 1.59162450e+00
-3.91108900e-01 2.13488698e-01 -1.79422691e-01 -6.33709490e-01
9.47653592e-01 6.07618213e-01 1.12203896e-01 -5.17283499e-01
1.21868715e-01 1.48920193e-01 -9.43135321e-02 -3.82067293e-01
8.08453739e-01 -2.81442970e-01 -5.96185684e-01 2.00062409e-01
2.64691383e-01 4.95039523e-01 4.14932042e-01 4.11064923e-02
1.18292296e+00 1.50007114e-01 2.92193770e-01 6.45406097e-02
1.08388044e-01 6.47569299e-02 1.04145634e+00 7.58910835e-01
-1.73771411e-01 5.96446693e-01 4.34468031e-01 -1.88679576e-01
-9.25095439e-01 -1.04925692e+00 -1.82708558e-02 1.23885334e+00
-8.27220529e-02 -4.28470790e-01 -9.96724904e-01 -9.85683680e-01
-4.40838754e-01 9.39556658e-01 -6.68239415e-01 -4.78415005e-03
-7.64106631e-01 -1.15408635e+00 9.50260580e-01 9.54327285e-01
7.37940192e-01 -1.36145186e+00 -3.04607432e-02 3.76607716e-01
-7.44843900e-01 -1.09054363e+00 -7.90021956e-01 3.40631664e-01
-5.82603037e-01 -6.39203250e-01 -7.91209757e-01 -1.12262821e+00
6.05627537e-01 -1.04264118e-01 1.15099239e+00 -4.02590930e-01
1.12239067e-02 1.59695476e-01 -4.01141584e-01 -2.85441458e-01
-4.49409664e-01 6.79400921e-01 -3.69088471e-01 -1.17928073e-01
6.78583622e-01 -2.20071346e-01 -2.84852713e-01 2.26404950e-01
-8.25888991e-01 -1.00955158e-01 8.99957001e-01 9.52849448e-01
5.84057987e-01 -5.45951426e-02 7.40246117e-01 -1.21648562e+00
5.49510896e-01 -7.37951696e-01 -1.46765977e-01 5.46767831e-01
-4.45733935e-01 1.57480657e-01 5.74955642e-01 -2.02934280e-01
-1.62863600e+00 7.21476413e-03 -6.90900862e-01 1.02479771e-01
-5.50198734e-01 7.42494643e-01 -7.05661118e-01 5.86115956e-01
3.76421392e-01 4.86597300e-01 -8.16697896e-01 -6.56821966e-01
3.71047467e-01 8.51496696e-01 5.71334481e-01 -3.84351939e-01
6.03914380e-01 -9.22034904e-02 -7.41135538e-01 -7.15234339e-01
-1.23063624e+00 -6.29755318e-01 -7.79851615e-01 1.77413434e-01
1.09617472e+00 -1.08218980e+00 -2.27996111e-01 5.39938211e-01
-1.41303432e+00 -1.91782027e-01 -1.48977444e-01 4.25193787e-01
-1.05059519e-01 2.52620846e-01 -8.72578740e-01 -6.50121450e-01
-6.26362085e-01 -7.12493598e-01 1.19929171e+00 5.95393598e-01
-1.17566265e-01 -1.20771503e+00 4.61882174e-01 2.96703994e-01
4.40859616e-01 -1.22558214e-01 8.49217296e-01 -1.15339994e+00
-4.31374520e-01 -1.01575807e-01 -1.89921618e-01 6.72132075e-02
8.87951553e-02 -3.16471308e-01 -8.88393641e-01 5.01466468e-02
-2.66102284e-01 5.43784164e-02 9.29515541e-01 1.26969770e-01
5.61576426e-01 -2.83279598e-01 -5.20636737e-01 5.81195533e-01
1.32468879e+00 3.34036320e-01 8.63077700e-01 3.03972334e-01
8.66410732e-01 3.06658953e-01 3.52248996e-01 2.54343450e-01
1.03913367e+00 3.85223210e-01 1.19535690e-02 -1.87250152e-01
-1.35044307e-01 -5.35563648e-01 4.87139851e-01 1.15521717e+00
7.71405697e-02 -5.78270078e-01 -1.09413695e+00 7.16280043e-01
-1.49054766e+00 -1.10416198e+00 -7.12025613e-02 1.88821816e+00
1.16330433e+00 3.21609760e-03 -3.42374258e-02 -4.54487890e-01
1.22705233e+00 3.94417122e-02 -4.52910453e-01 -2.67055690e-01
-1.04211070e-01 1.96324617e-01 3.49028170e-01 2.51112550e-01
-1.44584048e+00 1.10964370e+00 6.39585924e+00 5.80561280e-01
-8.20087314e-01 1.57630309e-01 5.50300956e-01 4.87038732e-01
-4.28715497e-01 7.70567684e-03 -1.60666835e+00 5.56608379e-01
1.08192849e+00 -2.15544000e-01 2.20399778e-02 7.71166086e-01
-1.12574212e-01 1.77778274e-01 -9.89367783e-01 5.90199471e-01
1.18510656e-01 -9.06664252e-01 2.43648514e-02 -1.66413069e-01
6.42720044e-01 2.94599354e-01 -4.38389122e-01 7.83083260e-01
6.37965262e-01 -7.42879689e-01 5.74171662e-01 5.77565432e-01
6.39335871e-01 -7.14765966e-01 9.61150944e-01 4.57286537e-01
-1.35054743e+00 2.43510246e-01 -3.37096155e-01 2.71856129e-01
5.70287704e-01 6.30190611e-01 -8.85071397e-01 7.74454474e-01
5.85608304e-01 4.21391100e-01 -4.36860979e-01 1.35959148e+00
-4.73359287e-01 9.53812897e-01 -2.80533373e-01 -1.60252854e-01
8.61757919e-02 2.99526788e-02 3.59005928e-01 1.68442333e+00
2.92411298e-01 1.40083596e-01 1.11044660e-01 9.50656474e-01
-4.19391215e-01 1.35965616e-01 -2.83094674e-01 -2.15964854e-01
7.14852691e-01 1.43399549e+00 -8.50752234e-01 -3.51209879e-01
-5.50342619e-01 1.20417619e+00 5.62907815e-01 2.08690569e-01
-1.02168906e+00 -8.86351109e-01 4.73732442e-01 -4.46575016e-01
7.60548413e-01 -2.51879901e-01 -2.84923404e-01 -1.45443535e+00
-1.23195285e-02 -4.94289786e-01 5.48446715e-01 -5.28440297e-01
-1.56266582e+00 7.37139225e-01 -2.98433572e-01 -6.90355062e-01
-1.56627193e-01 -2.91968673e-01 -7.61288345e-01 8.79014790e-01
-1.50859869e+00 -1.09463143e+00 -1.73565984e-01 1.45973220e-01
4.72580791e-01 3.20458800e-01 1.05570853e+00 6.10031903e-01
-9.78090227e-01 8.28467131e-01 5.73901311e-02 1.08034849e+00
6.83376133e-01 -1.47868526e+00 1.01953745e+00 1.02272463e+00
3.54569018e-01 7.18163133e-01 2.08540395e-01 -8.86407018e-01
-9.69385028e-01 -1.38158572e+00 1.68007278e+00 -4.33530331e-01
4.68076587e-01 -7.00573564e-01 -9.72785592e-01 9.89762545e-01
3.75676781e-01 -2.41804272e-01 1.06320024e+00 3.50668341e-01
-3.53560090e-01 3.27960610e-01 -8.16555798e-01 5.21138370e-01
1.20342147e+00 -3.51727635e-01 -9.36537802e-01 -6.96972832e-02
6.13889813e-01 -5.15950143e-01 -1.01092565e+00 1.97176024e-01
3.05976659e-01 -5.63305080e-01 6.43576920e-01 -7.14544475e-01
2.66343325e-01 -2.02037022e-01 6.57904744e-02 -1.59834230e+00
-5.69549143e-01 -3.65571916e-01 2.13030308e-01 2.07475257e+00
7.79760242e-01 -5.72158754e-01 6.11914515e-01 6.99312210e-01
-3.51540744e-01 -4.70685691e-01 -8.10834110e-01 -6.60462677e-01
1.27867520e-01 -9.63341594e-02 5.97229302e-01 1.04140186e+00
7.43792057e-02 8.88216436e-01 -1.58097282e-01 3.48470449e-01
2.65487939e-01 1.21782035e-01 3.55016679e-01 -9.63422418e-01
-4.29089516e-02 -2.32575536e-01 -2.26543725e-01 -1.33971834e+00
2.04455838e-01 -1.14493382e+00 5.16015589e-01 -1.92066932e+00
4.43385720e-01 -6.14820004e-01 -3.77507716e-01 9.58245397e-01
-6.45544708e-01 1.17373273e-01 -4.96054441e-03 5.90842888e-02
-9.19517338e-01 5.90322435e-01 7.52612054e-01 4.73489799e-02
-1.23325959e-01 -9.21589956e-02 -7.76083231e-01 4.96966869e-01
7.84550965e-01 -6.67291403e-01 9.73815396e-02 -8.28335404e-01
3.18888903e-01 -1.50633305e-01 5.33114299e-02 -8.77038777e-01
4.22966659e-01 2.15010509e-01 4.31443483e-01 -5.84926367e-01
-1.96434125e-01 -3.42604011e-01 6.66831508e-02 9.29539427e-02
-3.72657120e-01 8.18547457e-02 8.38693157e-02 4.25118625e-01
-3.49019825e-01 -2.75261879e-01 4.38334882e-01 -5.13854958e-02
-9.91011560e-01 3.09794039e-01 -3.52830619e-01 6.50064230e-01
8.45360398e-01 4.01942246e-02 -3.81967038e-01 4.14438620e-02
-8.96833897e-01 2.47955635e-01 1.78671554e-01 5.15541911e-01
3.38872045e-01 -1.31945932e+00 -8.27650011e-01 1.91011444e-01
3.69755067e-02 1.85909286e-01 3.08113277e-01 6.99410319e-01
-1.18486308e-01 4.02983934e-01 4.83990312e-02 -3.27636212e-01
-9.72001910e-01 3.23681951e-01 4.09942001e-01 -6.38280869e-01
-3.67326617e-01 9.16352272e-01 3.25536966e-01 -8.08828294e-01
9.02257636e-02 -2.09283695e-01 -4.18943197e-01 1.17158718e-01
4.52743798e-01 2.99906373e-01 1.71693191e-01 -6.63809597e-01
-4.60270464e-01 2.18602926e-01 -2.97387779e-01 1.51975378e-01
1.33148170e+00 -1.35525972e-01 1.96709596e-02 4.01100546e-01
9.86640632e-01 7.81172514e-02 -9.39443469e-01 -2.16717198e-01
4.20097858e-01 1.79670319e-01 -2.29113728e-01 -9.83072460e-01
-8.02906871e-01 8.01324129e-01 1.36713311e-01 2.86699027e-01
9.44985032e-01 8.97084773e-02 1.06369102e+00 2.42445290e-01
3.97176743e-01 -9.75853622e-01 -5.63597441e-01 1.07326090e+00
2.40100697e-01 -1.05219185e+00 -6.03952467e-01 -5.08865237e-01
-7.66684651e-01 6.93720639e-01 6.54447019e-01 1.12451099e-01
4.83379573e-01 6.79285884e-01 1.85663030e-01 6.89290538e-02
-7.32235372e-01 -4.00117159e-01 2.05344349e-01 4.09095734e-01
9.46714520e-01 7.65709803e-02 -4.29692805e-01 1.09759951e+00
-1.91500649e-01 -8.44233558e-02 2.48786047e-01 8.65083635e-01
-4.67032403e-01 -1.27362919e+00 -3.97300646e-02 3.58282596e-01
-8.28685284e-01 -6.20514870e-01 -3.50742936e-01 6.05273426e-01
9.05236676e-02 8.19280028e-01 2.42299303e-01 -2.17261583e-01
5.60657799e-01 5.61274290e-01 2.34752432e-01 -8.89688015e-01
-9.49176550e-01 -1.27629161e-01 3.42970788e-01 -1.49632215e-01
-2.77477711e-01 -7.09378064e-01 -1.49720716e+00 9.88574699e-03
-5.72823882e-01 6.30066574e-01 5.60692251e-01 1.00428414e+00
7.38950253e-01 5.57042241e-01 5.13472557e-01 -3.91665697e-01
-6.21282041e-01 -1.15768504e+00 -4.16728079e-01 2.37760127e-01
-1.36877537e-01 -4.59630042e-01 -1.80997878e-01 2.11621672e-01] | [9.689179420471191, 9.507554054260254] |
ada43423-7c24-40a9-9e9c-18fb72a211fd | on-the-limitations-of-unsupervised-bilingual | 1805.03620 | null | http://arxiv.org/abs/1805.03620v1 | http://arxiv.org/pdf/1805.03620v1.pdf | On the Limitations of Unsupervised Bilingual Dictionary Induction | Unsupervised machine translation---i.e., not assuming any cross-lingual
supervision signal, whether a dictionary, translations, or comparable
corpora---seems impossible, but nevertheless, Lample et al. (2018) recently
proposed a fully unsupervised machine translation (MT) model. The model relies
heavily on an adversarial, unsupervised alignment of word embedding spaces for
bilingual dictionary induction (Conneau et al., 2018), which we examine here.
Our results identify the limitations of current unsupervised MT: unsupervised
bilingual dictionary induction performs much worse on morphologically rich
languages that are not dependent marking, when monolingual corpora from
different domains or different embedding algorithms are used. We show that a
simple trick, exploiting a weak supervision signal from identical words,
enables more robust induction, and establish a near-perfect correlation between
unsupervised bilingual dictionary induction performance and a previously
unexplored graph similarity metric. | ['Ivan Vulić', 'Anders Søgaard', 'Sebastian Ruder'] | 2018-05-09 | on-the-limitations-of-unsupervised-bilingual-1 | https://aclanthology.org/P18-1072 | https://aclanthology.org/P18-1072.pdf | acl-2018-7 | ['graph-similarity', 'unsupervised-machine-translation'] | ['graphs', 'natural-language-processing'] | [ 2.82593906e-01 1.26496106e-01 -6.74773812e-01 -5.70277944e-02
-7.02055514e-01 -1.09836292e+00 9.43984926e-01 2.29452193e-01
-6.01694226e-01 8.21775198e-01 3.68148088e-01 -7.37965047e-01
2.12498903e-01 -6.71243608e-01 -7.34240055e-01 -6.21424496e-01
3.48756239e-02 9.35966492e-01 -2.43690893e-01 -4.45047855e-01
5.31228073e-02 3.27354044e-01 -8.15323710e-01 -2.30328664e-01
9.78660226e-01 -4.45514210e-02 -4.03509382e-03 3.46383661e-01
-5.02501279e-02 3.96333814e-01 -1.45938054e-01 -8.14428568e-01
5.54570556e-01 -6.01767838e-01 -9.30539608e-01 -1.09705649e-01
5.73142231e-01 -2.60426756e-02 -6.01392806e-01 1.18736982e+00
4.95842963e-01 -1.48988485e-01 7.38687277e-01 -7.83012748e-01
-1.14408362e+00 1.01781976e+00 -3.75288934e-01 2.39735499e-01
3.60163629e-01 1.91951543e-01 1.47473180e+00 -9.23580229e-01
1.11607671e+00 9.74555910e-01 6.97908461e-01 3.38982105e-01
-1.86053574e+00 -5.68122089e-01 -1.75355196e-01 2.80579388e-01
-1.38451076e+00 -5.18037736e-01 9.06580627e-01 -5.46321809e-01
8.81995738e-01 -1.13821879e-01 4.43489969e-01 1.47363603e+00
2.73751229e-01 4.68272835e-01 1.59017432e+00 -9.12396252e-01
-7.35194311e-02 2.22866327e-01 -4.44145799e-02 7.37398803e-01
4.79445875e-01 3.66477430e-01 -6.03271902e-01 -3.29505801e-02
6.06841385e-01 -4.42337006e-01 -1.73224241e-01 -5.82974613e-01
-1.68313634e+00 9.73821580e-01 6.02699444e-02 7.82170951e-01
9.15439241e-03 -2.55744189e-01 5.38632393e-01 8.70132864e-01
4.52887207e-01 6.84480190e-01 -3.44030291e-01 3.45124118e-02
-9.85207617e-01 -4.52445716e-01 8.70620728e-01 1.00220454e+00
1.11560249e+00 9.50626507e-02 3.07749063e-01 6.86440766e-01
-1.16109587e-02 6.49958313e-01 3.76226693e-01 -4.60572362e-01
4.78250831e-01 1.15534127e-01 -3.17937315e-01 -9.31437492e-01
-1.89179882e-01 -3.91416848e-01 -7.53869653e-01 -1.34367198e-01
5.38470149e-01 -1.80971339e-01 -5.30289948e-01 1.99188483e+00
6.88872635e-02 1.60206053e-02 5.94882220e-02 7.93021560e-01
2.30991781e-01 1.26624912e-01 2.82465927e-02 -3.55818808e-01
1.17746007e+00 -9.30239201e-01 -6.52853310e-01 -1.89693615e-01
7.67029226e-01 -9.92068172e-01 1.37889659e+00 2.40051731e-01
-8.44216049e-01 -3.34204078e-01 -1.17291427e+00 -2.69266248e-01
-5.41925311e-01 -4.01291512e-02 8.19334984e-01 7.02716768e-01
-1.25296092e+00 5.43289125e-01 -7.66642869e-01 -8.74064565e-01
4.03735489e-02 4.73679721e-01 -7.48900831e-01 -1.74331203e-01
-1.36859703e+00 1.22609401e+00 2.07832277e-01 -2.55374044e-01
-9.88461733e-01 -2.64465123e-01 -9.19332206e-01 -4.78231430e-01
1.49078608e-01 -5.35174787e-01 6.59665942e-01 -1.36854553e+00
-1.55315590e+00 1.35402465e+00 -5.68872392e-02 -4.67649370e-01
2.02405363e-01 1.15387544e-01 -4.75606531e-01 1.61664173e-01
2.37047911e-01 5.23212373e-01 8.03963780e-01 -1.16461229e+00
-5.00894226e-02 -4.17009324e-01 5.73826246e-02 1.23500198e-01
-5.52964568e-01 3.14051390e-01 -1.72106236e-01 -9.87148464e-01
-4.43725511e-02 -1.13453305e+00 -1.58380017e-01 -3.92625004e-01
-3.94313723e-01 -2.05591619e-02 2.53256530e-01 -9.04236913e-01
9.79548633e-01 -1.99273169e+00 6.39916301e-01 2.83516884e-01
1.09592482e-01 -4.80570197e-02 -6.15004897e-01 8.14639211e-01
-2.28288442e-01 1.78883448e-01 -5.86467803e-01 -2.44538009e-01
2.79907048e-01 5.91954827e-01 -3.20943147e-01 8.67616892e-01
3.69544715e-01 1.03574443e+00 -1.09954894e+00 -6.77271783e-01
-5.09777945e-03 2.56976664e-01 -4.46485609e-01 -1.75511360e-01
-4.76441495e-02 8.80922556e-01 -2.22034915e-03 8.00150990e-01
3.92885506e-01 1.67470336e-01 7.90635347e-01 -1.49706647e-01
-1.17877431e-01 6.39048874e-01 -7.75608122e-01 2.17589068e+00
-5.92963874e-01 8.76638472e-01 2.50197165e-02 -1.38361812e+00
7.18230367e-01 2.86131680e-01 4.15705621e-01 -7.23052800e-01
1.46598071e-01 5.19119442e-01 3.75089854e-01 -2.26713896e-01
2.01944143e-01 -3.53755832e-01 -1.46410167e-01 7.79351294e-01
7.21119523e-01 -1.54438749e-01 4.54107195e-01 1.10846139e-01
1.17446566e+00 2.62526333e-01 2.06537500e-01 -7.00442910e-01
3.16596925e-01 2.38685116e-01 4.87252653e-01 6.13832116e-01
-1.21818013e-01 3.09729159e-01 3.41070682e-01 -1.71770036e-01
-1.30307651e+00 -1.34164619e+00 -4.51223493e-01 1.14284086e+00
-8.87846798e-02 -4.81656224e-01 -5.87395966e-01 -1.01526487e+00
-6.30901381e-02 6.71298027e-01 -6.19857788e-01 -1.28938437e-01
-7.04786718e-01 -7.13721573e-01 8.98476660e-01 2.01538876e-01
-1.95752636e-01 -8.86626840e-01 3.38277400e-01 1.63363889e-01
-6.34468943e-02 -1.21850026e+00 -5.73844254e-01 4.98795539e-01
-1.07352805e+00 -8.35748553e-01 -4.08188939e-01 -1.16682065e+00
8.00084412e-01 -3.01648434e-02 1.46763980e+00 -1.52005494e-01
6.00425852e-03 5.70354164e-01 -2.53582507e-01 -2.82550100e-02
-6.04981661e-01 3.05033565e-01 6.37905538e-01 -1.91027019e-02
7.84601748e-01 -1.08667290e+00 -4.49453667e-02 2.31682077e-01
-8.61560762e-01 -2.58363843e-01 8.38007629e-01 1.04512942e+00
6.87408268e-01 -3.75140548e-01 3.38874400e-01 -9.81504679e-01
5.25647283e-01 -3.78866881e-01 -5.20742059e-01 2.86517650e-01
-7.74032772e-01 3.83872718e-01 8.05287659e-01 -5.79645395e-01
-3.41715336e-01 -6.87995404e-02 -3.72997299e-02 -2.92320669e-01
-9.06058922e-02 4.76029366e-01 -2.43727699e-01 -3.48366797e-01
7.36770630e-01 4.81466472e-01 -9.31403488e-02 -4.99123633e-01
7.48053968e-01 5.02766490e-01 4.55074221e-01 -9.22155321e-01
1.43904507e+00 3.84765834e-01 -1.19206756e-01 -5.69682062e-01
-4.47710812e-01 -3.81142765e-01 -1.32218218e+00 2.41081923e-01
8.54380965e-01 -1.02507889e+00 5.09563610e-02 -1.34902820e-01
-1.11380255e+00 -4.45226669e-01 -3.56303781e-01 8.70698035e-01
-5.27935624e-01 5.35508692e-01 -7.96034694e-01 -1.53055280e-01
-1.13498881e-01 -1.09679925e+00 6.89077497e-01 -4.97162968e-01
-3.73916030e-01 -1.45927060e+00 6.30812407e-01 2.38445580e-01
1.34055451e-01 6.86924085e-02 1.07729709e+00 -9.45938528e-01
-4.19224769e-01 1.79355785e-01 1.27738519e-02 4.69055653e-01
4.26777273e-01 -5.51462285e-02 -5.95047891e-01 -5.40315092e-01
-3.93743291e-02 -4.38088149e-01 7.41770089e-01 -2.55611330e-01
4.72574919e-01 -4.32876855e-01 -7.24297538e-02 6.72697425e-01
1.52938569e+00 -2.55415380e-01 4.22723651e-01 4.27721709e-01
9.10017729e-01 5.41472077e-01 1.91727042e-01 -2.48724073e-01
3.42646927e-01 5.54283202e-01 -1.67894125e-01 -3.14169616e-01
-2.65516043e-01 -4.46950912e-01 8.18023741e-01 1.91967702e+00
-2.26701438e-01 1.51843980e-01 -1.05331051e+00 9.67775285e-01
-1.33745039e+00 -8.78810823e-01 5.00100255e-02 2.27334380e+00
1.42947900e+00 1.34177163e-01 2.49827411e-02 -8.11462924e-02
5.63229322e-01 1.53133288e-01 -1.82053104e-01 -6.44647479e-01
-5.42283535e-01 1.05432498e+00 7.73083329e-01 8.08484137e-01
-9.64219511e-01 1.31574130e+00 6.63939905e+00 7.27621973e-01
-1.13914764e+00 5.82905471e-01 4.62014079e-02 2.85952508e-01
-7.46787190e-01 3.91171187e-01 -2.92585403e-01 3.52476776e-01
8.63280058e-01 -2.08091274e-01 8.38847518e-01 4.22246963e-01
-6.98885173e-02 4.46024656e-01 -1.31829643e+00 7.07272410e-01
4.16520000e-01 -9.59670365e-01 1.09021463e-01 2.51535207e-01
1.00325692e+00 5.26377022e-01 5.36834858e-02 2.29517803e-01
7.40557551e-01 -1.00051808e+00 4.67555016e-01 4.08052541e-02
1.08617353e+00 -5.73652804e-01 4.14718807e-01 9.15752649e-02
-8.66461754e-01 3.87176841e-01 -4.31476682e-01 9.31907892e-02
-4.49075736e-02 5.88093460e-01 -5.80459356e-01 6.48525000e-01
9.21566039e-02 9.08191264e-01 -6.50401890e-01 3.06840181e-01
-7.50742435e-01 9.99153137e-01 -2.82291472e-01 2.79257596e-01
2.26508185e-01 -7.78556883e-01 6.81994498e-01 1.55370331e+00
8.04001912e-02 -3.12390983e-01 2.40685880e-01 5.21893799e-01
-2.97607630e-01 5.84864199e-01 -1.06207716e+00 -4.82444674e-01
1.72837391e-01 1.15418625e+00 -6.58948064e-01 -6.77871406e-02
-7.38780141e-01 1.26988375e+00 5.66560745e-01 4.14727688e-01
-6.83180749e-01 -2.20468432e-01 5.49347162e-01 -5.80427870e-02
2.54772276e-01 -7.28782892e-01 -4.56053078e-01 -1.61838341e+00
1.59199443e-02 -1.41060245e+00 1.29786074e-01 -1.25372469e-01
-1.45304978e+00 4.29804355e-01 -4.33368355e-01 -1.19339359e+00
-7.38444179e-02 -6.92265332e-01 -4.69944030e-01 7.37573385e-01
-1.54585755e+00 -1.39540875e+00 4.18458223e-01 7.58495271e-01
1.88163549e-01 -3.65079761e-01 1.08930421e+00 5.77831030e-01
-4.69444603e-01 9.07274604e-01 3.63984853e-01 5.32850385e-01
1.03680372e+00 -1.46466351e+00 4.67889994e-01 1.01888871e+00
1.15234363e+00 9.52224135e-01 6.58735275e-01 -5.05473673e-01
-1.78736722e+00 -9.87919450e-01 1.39904499e+00 -7.88054466e-01
1.34131372e+00 -7.77980804e-01 -6.06614590e-01 9.83964682e-01
7.14704454e-01 -2.44706288e-01 1.04518235e+00 4.11640972e-01
-8.37365270e-01 -1.36348773e-02 -6.69521928e-01 7.91528463e-01
1.36212230e+00 -1.26922333e+00 -8.84602845e-01 7.05539346e-01
6.35546982e-01 2.13080257e-01 -1.16634011e+00 3.09626758e-01
3.50378841e-01 -6.00163102e-01 8.09670150e-01 -7.92019844e-01
3.93581659e-01 -2.06716701e-01 -2.45562956e-01 -1.44881654e+00
-2.24572524e-01 -8.61813009e-01 2.22406596e-01 1.20366502e+00
6.71914637e-01 -8.14631999e-01 2.18153641e-01 -2.76090145e-01
-1.59408316e-01 -3.54784399e-01 -9.65109110e-01 -1.06780028e+00
6.21837974e-01 -2.11335540e-01 2.82773465e-01 1.79061115e+00
3.07543397e-01 7.27940142e-01 -1.85652077e-01 9.33592170e-02
7.55434394e-01 -3.90517316e-03 8.27371538e-01 -1.10583436e+00
-5.31138718e-01 -5.36825299e-01 -4.36121792e-01 -8.61805797e-01
7.82487333e-01 -1.61588073e+00 -1.82804793e-01 -1.20314121e+00
1.60941586e-01 -3.96101534e-01 -5.97928166e-01 5.34872591e-01
3.42131779e-02 4.59936947e-01 -1.02148741e-01 4.73617047e-01
-3.33159238e-01 4.52265769e-01 1.08766007e+00 -4.75658327e-01
1.37095138e-01 -6.72159433e-01 -6.34069562e-01 3.91978532e-01
7.14887857e-01 -7.37997949e-01 -2.50005722e-01 -7.86437929e-01
1.83108076e-01 -3.56639385e-01 9.09440592e-02 -7.02330232e-01
9.95492488e-02 -1.09364428e-01 -1.47355184e-01 1.37855873e-01
-2.80533105e-01 -6.31564200e-01 -5.42997457e-02 3.63131672e-01
-1.66625574e-01 3.22525650e-01 1.23455703e-01 3.59676450e-01
-2.36368537e-01 -2.09877864e-01 5.68009257e-01 4.90045287e-02
-3.66272748e-01 3.35063756e-01 -2.47241735e-01 2.97055990e-01
6.09029293e-01 -7.82290772e-02 1.76256970e-02 -8.10669269e-03
-6.84480727e-01 -2.03478798e-01 8.63865972e-01 3.04654866e-01
3.37551273e-02 -1.56479454e+00 -1.10705543e+00 1.94418371e-01
2.89874107e-01 -6.52892828e-01 -6.93427980e-01 1.18985212e+00
-3.88406873e-01 2.54547834e-01 -2.18237266e-01 -4.61952478e-01
-7.11064458e-01 7.83767700e-01 -4.27805595e-02 -4.93878037e-01
-3.12902570e-01 4.77720588e-01 -3.56544815e-02 -8.87513876e-01
-2.26789281e-01 3.06647774e-02 2.16718927e-01 1.39029562e-01
-9.33374166e-02 -1.11070633e-01 8.05850849e-02 -9.00918245e-01
-3.18332911e-01 7.43175566e-01 4.96480577e-02 -4.62063104e-01
1.20880079e+00 -6.60105050e-02 -5.14657319e-01 6.11292064e-01
1.11556315e+00 7.85552382e-01 -4.72209126e-01 -5.05199194e-01
2.69472241e-01 -2.34153494e-01 -1.34680480e-01 -5.15490651e-01
-8.15143228e-01 9.00479972e-01 3.21056306e-01 3.20895761e-02
9.54743445e-01 -4.99480516e-02 7.64276564e-01 5.20294428e-01
7.37655282e-01 -1.13645089e+00 -2.38226503e-01 7.09168732e-01
6.09388649e-01 -1.26073575e+00 -7.57075399e-02 -1.90826710e-02
-3.19434851e-01 9.45793152e-01 1.69368923e-01 -1.75965175e-01
4.67235148e-01 1.59910873e-01 2.23642483e-01 -4.91583012e-02
-3.66707265e-01 -5.46103239e-01 2.59060085e-01 7.45698571e-01
4.90943134e-01 1.19313888e-01 -7.45545805e-01 1.83759391e-01
-3.44931513e-01 -4.69879061e-01 2.03021869e-01 6.12517715e-01
1.14623979e-01 -1.78433907e+00 -3.02085966e-01 1.00815453e-01
-4.25316542e-01 -7.09281445e-01 -8.08894515e-01 9.05475616e-01
3.23595762e-01 9.14955556e-01 -1.13951050e-01 -5.18191397e-01
3.15322652e-02 3.62560898e-01 8.40843141e-01 -8.31102669e-01
-6.15540743e-01 -8.64843875e-02 1.08716518e-01 -2.67407447e-01
-6.66482031e-01 -6.51129842e-01 -7.54397810e-01 -4.47823644e-01
-3.51052612e-01 3.34729761e-01 5.14719784e-01 9.18058097e-01
1.61091909e-01 -5.49110733e-02 7.68449426e-01 -4.94343072e-01
-3.78412455e-01 -1.00125170e+00 -5.01863122e-01 6.37356579e-01
9.75191891e-02 -5.33198357e-01 -5.56244552e-01 3.59762222e-01] | [11.074645042419434, 10.08076286315918] |
6c3f33a1-b588-4913-8668-c515adefbce3 | coderetriever-unimodal-and-bimodal-1 | 2201.10866 | null | https://arxiv.org/abs/2201.10866v3 | https://arxiv.org/pdf/2201.10866v3.pdf | CodeRetriever: Unimodal and Bimodal Contrastive Learning for Code Search | In this paper, we propose the CodeRetriever model, which learns the function-level code semantic representations through large-scale code-text contrastive pre-training. We adopt two contrastive learning schemes in CodeRetriever: unimodal contrastive learning and bimodal contrastive learning. For unimodal contrastive learning, we design an unsupervised learning approach to build semantic-related code pairs based on the documentation and function name. For bimodal contrastive learning, we leverage the documentation and in-line comments of code to build code-text pairs. Both contrastive objectives can fully leverage large-scale code corpus for pre-training. Extensive experimental results show that CodeRetriever achieves new state-of-the-art with significant improvement over existing code pre-trained models, on eleven domain/language-specific code search tasks with six programming languages in different code granularity (function-level, snippet-level and statement-level). These results demonstrate the effectiveness and robustness of CodeRetriever. | ['Nan Duan', 'Weizhu Chen', 'Daxin Jiang', 'Weizhen Qi', 'Bolun Yao', 'Hang Zhang', 'Xipeng Qiu', 'Yelong Shen', 'Yeyun Gong', 'Xiaonan Li'] | 2022-01-26 | null | null | null | null | ['code-search', 'code-search'] | ['computer-code', 'computer-vision'] | [-1.25113860e-01 -3.13439131e-01 -6.68943405e-01 -4.96445954e-01
-1.07417846e+00 -7.33507931e-01 5.91647863e-01 3.71414781e-01
4.72430931e-03 -2.44544689e-02 3.57515514e-01 -8.06277514e-01
9.27232504e-02 -3.89264226e-01 -6.90501451e-01 -1.91122163e-02
-6.03556037e-02 1.02608688e-01 2.53670412e-04 -3.36317122e-01
9.00443017e-01 -1.73711106e-01 -1.61043596e+00 1.00434494e+00
1.45626688e+00 6.06179416e-01 8.21251810e-01 8.33833277e-01
-8.14125001e-01 1.56197035e+00 -3.77856582e-01 -3.51255387e-01
-8.40301365e-02 -3.72330874e-01 -9.69289660e-01 -3.23682398e-01
6.06642067e-01 -4.69684787e-02 -6.51087537e-02 1.34015012e+00
3.41317733e-03 -3.45441073e-01 6.22616768e-01 -1.31335533e+00
-9.12071645e-01 1.06202567e+00 -9.29292381e-01 1.89800784e-01
4.72884536e-01 6.82165772e-02 1.38789868e+00 -1.25882995e+00
4.30330306e-01 9.25208926e-01 8.84749174e-01 5.83278596e-01
-1.19254327e+00 -9.03078079e-01 -1.90571189e-01 1.70946628e-01
-1.23016715e+00 -1.85394898e-01 9.90242362e-01 -1.16107428e+00
1.54368699e+00 9.88823175e-02 1.72644779e-01 7.44436741e-01
1.65690109e-01 8.85747075e-01 7.86781073e-01 -7.66066790e-01
-9.66758877e-02 2.64420919e-02 3.89054865e-01 1.11936474e+00
4.98532355e-02 1.21508069e-01 -1.44055843e-01 -5.18741667e-01
2.40328506e-01 4.03838485e-01 -1.82077706e-01 -5.54189384e-01
-9.78991628e-01 9.20062184e-01 6.45470619e-01 4.39066321e-01
2.82819211e-01 3.73505414e-01 9.20699954e-01 5.23017526e-01
-5.28033823e-02 6.25188589e-01 -5.74764967e-01 -3.17881376e-01
-1.27469492e+00 9.90699828e-02 7.73291409e-01 1.62811482e+00
1.39769459e+00 2.09522426e-01 -1.89637825e-01 1.28021789e+00
6.57101393e-01 4.92402911e-01 8.88740897e-01 -5.69185674e-01
1.05578589e+00 1.15562654e+00 -4.93228972e-01 -8.42575610e-01
-1.53979054e-03 -2.14881971e-01 -3.76962453e-01 1.36846155e-01
-1.89204931e-01 2.76581526e-01 -7.12181211e-01 1.35263896e+00
-6.09076619e-01 -9.15827528e-02 1.14601798e-01 3.81466538e-01
1.06999695e+00 5.02425432e-01 1.27364665e-01 3.14919472e-01
1.21888840e+00 -1.66034710e+00 -1.84901744e-01 -7.27529585e-01
1.18172479e+00 -1.05282581e+00 1.68563402e+00 4.62706499e-02
-9.16114926e-01 -7.74746656e-01 -1.20780647e+00 -2.71794766e-01
-3.10553014e-01 6.35137439e-01 5.89179099e-01 5.43967247e-01
-1.08222520e+00 2.89777040e-01 -6.71783030e-01 -5.13735190e-02
3.34371299e-01 -1.18112035e-01 -1.24628827e-01 -2.09154636e-01
-4.91021603e-01 5.23981869e-01 5.00500262e-01 -7.96740770e-01
-1.17237103e+00 -9.70281303e-01 -1.14058876e+00 5.34005761e-01
9.49922949e-02 -1.45050749e-01 1.49029839e+00 -1.41128337e+00
-9.95660603e-01 1.38373399e+00 1.97417475e-02 -8.74711201e-02
-4.48566191e-02 -2.34751835e-01 -4.58549708e-01 -1.83714449e-01
4.05865461e-01 3.89640659e-01 7.74006546e-01 -1.43749881e+00
-5.23853183e-01 1.74379930e-01 2.29379982e-01 -1.89678133e-01
-4.42824543e-01 3.05407226e-01 -4.74393189e-01 -9.54977751e-01
-3.27785075e-01 -7.50050843e-01 1.62679166e-01 -3.77886631e-02
-8.16262141e-02 -2.03146815e-01 7.22029626e-01 -7.51148403e-01
1.59782469e+00 -2.61415744e+00 -1.65803116e-02 -1.42678678e-01
4.69294578e-01 3.65448929e-02 -4.80546683e-01 4.82813209e-01
-5.41064382e-01 1.04375765e-01 -7.07695246e-01 -1.66914500e-02
1.48388162e-01 -4.77372482e-02 -5.12559772e-01 7.77468607e-02
2.00868815e-01 1.16952479e+00 -1.21561599e+00 -6.14579737e-01
-2.15758979e-01 -2.36078769e-01 -1.24112535e+00 5.40160775e-01
-3.38779867e-01 -3.17786843e-01 -3.13984841e-01 8.00504267e-01
5.55012047e-01 -5.51426768e-01 1.95051506e-01 1.63333669e-01
-2.51711786e-01 6.29932582e-01 -2.77742416e-01 2.16383338e+00
-1.09514213e+00 9.36620235e-01 -1.46830976e-01 -1.15439510e+00
1.00563514e+00 -1.37462705e-01 1.55994040e-03 -1.00397468e+00
-3.26260179e-01 5.87511480e-01 -1.24130622e-01 -6.88065886e-01
5.93805850e-01 2.88624942e-01 -5.51369667e-01 5.70264161e-01
2.47549221e-01 -3.70688051e-01 2.29584649e-01 4.01709765e-01
1.38987541e+00 2.91703165e-01 5.45522213e-01 -9.89430904e-01
7.59725332e-01 1.50756031e-01 1.43387258e-01 6.17444754e-01
5.53567661e-03 3.59699845e-01 6.58289731e-01 -1.82824329e-01
-8.60288441e-01 -1.12610555e+00 1.85233757e-01 1.61907232e+00
9.37927365e-02 -1.07005501e+00 -7.11520135e-01 -1.20415139e+00
-1.21147688e-02 7.81677783e-01 -5.20847797e-01 -2.88527548e-01
-9.27937508e-01 -7.50593990e-02 6.74359679e-01 8.03849816e-01
4.15680647e-01 -8.58696342e-01 -5.52687407e-01 -8.84941369e-02
-1.13390118e-01 -4.42534655e-01 -9.38976467e-01 5.02633631e-01
-8.16179574e-01 -1.38799155e+00 -4.90566880e-01 -1.47057211e+00
9.08492029e-01 4.51796740e-01 1.87769163e+00 6.37943447e-01
-2.76919216e-01 3.51758689e-01 -4.86471742e-01 1.68304205e-01
-9.56313729e-01 1.54711455e-02 -9.08334672e-01 -8.80457044e-01
3.60985070e-01 -3.77853513e-01 -4.42081481e-01 1.58143312e-01
-9.32862520e-01 -2.34284233e-02 6.17186248e-01 1.02140236e+00
2.79317368e-02 -4.54521120e-01 1.46990269e-01 -9.62383389e-01
7.34729528e-01 -9.52971220e-01 -7.25041807e-01 4.70205277e-01
-8.01186562e-01 4.37527180e-01 9.24024880e-01 -2.83730954e-01
-1.14043081e+00 -2.09262326e-01 1.46002159e-03 -3.28787804e-01
1.93557218e-01 9.60767746e-01 2.01292127e-01 5.52024469e-02
1.22403729e+00 4.35781389e-01 -3.62879813e-01 -4.65234041e-01
5.45113266e-01 9.82822657e-01 7.45636702e-01 -1.26217902e+00
7.05724895e-01 8.94968882e-02 -8.53649914e-01 -1.15892909e-01
-2.08737895e-01 -8.03934038e-01 -3.28588694e-01 4.12853397e-02
5.48437536e-01 -1.18516731e+00 -6.83293864e-02 2.64429957e-01
-1.17662311e+00 -5.70095479e-01 5.80770075e-02 5.67375831e-02
-6.57850862e-01 5.30866504e-01 -5.93421102e-01 -9.44236517e-02
-2.27846935e-01 -1.52283502e+00 1.07674813e+00 -7.96760321e-02
-3.60422969e-01 -9.51632142e-01 4.15308982e-01 1.78858191e-01
5.29776871e-01 -3.08655500e-01 1.63899541e+00 -6.11110151e-01
-6.00579560e-01 2.28992939e-01 -4.79381442e-01 2.77608544e-01
3.51675540e-01 1.65578753e-01 -8.24822366e-01 -4.58007842e-01
-9.21309292e-02 -7.25854099e-01 8.54008794e-01 -3.40242565e-01
1.33852959e+00 -4.24612373e-01 -3.83459955e-01 7.32344866e-01
1.90374291e+00 2.85143763e-01 6.98861718e-01 3.27119440e-01
8.40810895e-01 2.25416273e-01 4.57716852e-01 5.62298536e-01
5.13328850e-01 3.71630549e-01 4.81109291e-01 1.95306882e-01
-4.41193670e-01 -3.14982265e-01 4.57996815e-01 1.52487481e+00
3.91854078e-01 2.68677562e-01 -1.59609866e+00 9.51701105e-01
-1.68463111e+00 -8.47642481e-01 -1.25490010e-01 1.98986244e+00
1.31444275e+00 -4.46144938e-02 -1.67296618e-01 -3.76383781e-01
6.32738531e-01 2.96499338e-02 -3.05899829e-01 -4.84238565e-01
2.88237721e-01 2.63644725e-01 1.81329444e-01 2.26696447e-01
-8.95227373e-01 8.48763108e-01 5.93411064e+00 1.04595459e+00
-1.19792891e+00 3.11587185e-01 2.94414610e-01 4.92720127e-01
-8.24473083e-01 4.28171843e-01 -3.73204350e-01 7.27999926e-01
9.26195145e-01 -2.87953943e-01 6.88632011e-01 1.70855832e+00
-5.34257472e-01 6.56120554e-02 -1.48797429e+00 1.34265375e+00
3.11399281e-01 -1.56453598e+00 -1.59346253e-01 -5.95722735e-01
1.32256520e+00 3.48425776e-01 -1.06148712e-01 1.12301803e+00
6.51680112e-01 -7.44105577e-01 8.56378496e-01 2.01185659e-01
9.88349915e-01 -4.95751858e-01 2.66238064e-01 1.33437440e-01
-1.68796098e+00 -4.23553526e-01 -1.11143522e-01 5.44588044e-02
-7.23115861e-01 2.34778240e-01 -4.69079256e-01 2.47812182e-01
8.16199005e-01 1.13546097e+00 -9.73372579e-01 1.01725686e+00
-2.31657892e-01 5.76601684e-01 5.02973676e-01 -1.33541971e-01
2.68381506e-01 3.16309363e-01 -6.42100945e-02 1.91386759e+00
2.75702029e-01 -6.36150181e-01 3.09206933e-01 1.57812250e+00
-4.62836176e-01 1.20748051e-01 -4.11513329e-01 -3.33838344e-01
6.38210416e-01 1.08713067e+00 -5.94625950e-01 -6.65621579e-01
-1.11452162e+00 8.23299170e-01 7.56434739e-01 2.88269371e-01
-9.81895447e-01 -8.83724034e-01 4.90240633e-01 -1.60581112e-01
5.18348336e-01 -5.94066121e-02 -2.85834223e-01 -1.22828841e+00
2.59799689e-01 -1.27124250e+00 5.03395140e-01 -8.05765331e-01
-1.18404162e+00 4.74099755e-01 1.69284612e-01 -1.42099166e+00
-2.27205202e-01 -6.74308717e-01 -8.93646896e-01 5.97066045e-01
-1.69426274e+00 -1.00963044e+00 -5.34887195e-01 4.41412359e-01
1.01633608e+00 -7.08279610e-01 6.11136258e-01 4.50551003e-01
8.14638063e-02 7.77356982e-01 4.21669066e-01 4.37287539e-01
6.84961438e-01 -1.47570169e+00 6.54678047e-01 7.70234346e-01
5.57697862e-02 1.21262622e+00 1.40718639e-01 -7.14918852e-01
-1.43010759e+00 -9.41051364e-01 6.49641395e-01 -3.11848849e-01
1.11373103e+00 -5.20114362e-01 -8.98741424e-01 6.14133120e-01
4.70848888e-01 1.13919191e-01 6.42262340e-01 -8.98326114e-02
-1.28897071e+00 -2.04167314e-04 -6.93980217e-01 4.97306973e-01
8.61177683e-01 -1.43536997e+00 -1.08320940e+00 1.43092543e-01
7.01364040e-01 -2.89887816e-01 -7.32523203e-01 3.18278044e-01
2.46068642e-01 -1.07137942e+00 8.75278950e-01 -4.04222816e-01
1.28330147e+00 -2.17511192e-01 -3.72454405e-01 -1.12644792e+00
-3.59829962e-01 -4.86979634e-01 -6.37529120e-02 1.18027198e+00
4.07814384e-01 -2.22704068e-01 4.46946383e-01 1.14305533e-01
-6.26263678e-01 -3.70344281e-01 -2.41634607e-01 -9.69298422e-01
5.13750076e-01 -3.42179537e-01 4.98034716e-01 1.47893441e+00
7.08687663e-01 2.00726166e-01 2.69571990e-01 -3.33780825e-01
2.21572667e-01 6.14891648e-01 6.50747776e-01 -8.63435030e-01
-7.95722306e-01 -1.03150332e+00 -9.02759954e-02 -1.37534773e+00
6.69492126e-01 -1.58561838e+00 3.33572388e-01 -1.03934920e+00
9.52036679e-01 -3.97885650e-01 -2.74988890e-01 6.49293840e-01
-4.72838014e-01 -3.14487427e-01 5.23803383e-02 5.35246551e-01
-6.36861503e-01 2.80551851e-01 5.17270863e-01 -8.74886870e-01
2.31120214e-01 -5.19273758e-01 -7.30875254e-01 5.65124393e-01
4.89736378e-01 -6.26068711e-01 -7.38207936e-01 -7.89911747e-01
3.81138653e-01 9.58422758e-03 2.11010531e-01 -9.16727245e-01
2.12089092e-01 -4.42766733e-02 -3.38761419e-01 -4.01009321e-01
-6.09509349e-01 -6.96491003e-01 -3.45852524e-01 6.89877033e-01
-7.58228064e-01 4.57097828e-01 5.67061961e-01 3.45450848e-01
-4.84745651e-01 -9.44019437e-01 8.66436064e-01 -4.58917648e-01
-1.16067696e+00 -2.98465669e-01 -3.87798607e-01 7.50427961e-01
5.66319942e-01 -5.40314093e-02 -7.08625853e-01 2.97173917e-01
9.63050351e-02 -1.16298988e-01 7.38906860e-01 8.41151655e-01
6.97933912e-01 -1.31370175e+00 -5.57045877e-01 3.65283340e-01
1.04883718e+00 -5.72460890e-01 -1.03852816e-01 5.02882004e-01
-7.73199558e-01 4.22653228e-01 -2.77899593e-01 -6.68306231e-01
-1.17506969e+00 8.40347528e-01 5.51862344e-02 -2.44526595e-01
-2.20395684e-01 8.77090096e-01 4.86706585e-01 -8.78427088e-01
1.85568929e-02 -6.10384941e-01 9.39006805e-02 -4.67702061e-01
4.97942626e-01 1.65289789e-01 1.23763822e-01 -3.42492700e-01
-4.73425031e-01 5.95851183e-01 -4.22233641e-01 5.41439474e-01
1.06681418e+00 3.07238013e-01 -5.51018119e-01 1.90162644e-01
1.82438421e+00 3.87054384e-01 -1.06653357e+00 -1.94723353e-01
6.45866156e-01 -5.67042291e-01 -2.81266849e-02 -8.69990885e-01
-1.06017196e+00 9.24876392e-01 4.21303898e-01 -8.62415805e-02
9.63195384e-01 2.24536642e-01 3.22032452e-01 6.56327009e-01
3.56833488e-01 -9.41137135e-01 6.60645962e-01 7.63145208e-01
7.60673225e-01 -1.28130269e+00 -1.94129467e-01 -2.91526407e-01
-2.45676592e-01 1.30425787e+00 9.99318302e-01 -2.28369698e-01
5.55255234e-01 7.49496698e-01 4.81626876e-02 -3.37045252e-01
-7.22023845e-01 -1.19228922e-02 4.30331588e-01 4.65011477e-01
1.04916549e+00 -2.23150864e-01 -1.56772986e-01 5.50610304e-01
1.16412662e-01 -2.78911084e-01 3.47515285e-01 1.52514994e+00
-4.83836204e-01 -1.11754894e+00 -6.13192003e-03 4.71355319e-01
-2.53901798e-02 -8.65094185e-01 -4.85988289e-01 6.94316983e-01
-1.61039293e-01 6.91336632e-01 5.12569770e-02 -4.11527365e-01
4.47871760e-02 5.16173616e-03 3.95214647e-01 -9.30501819e-01
-9.55922246e-01 -2.98568785e-01 -1.57431275e-01 -5.60722709e-01
-1.63724050e-01 -1.92195430e-01 -1.54175484e+00 3.54292169e-02
-3.64286840e-01 2.16019511e-01 5.39891958e-01 7.56388247e-01
6.31853700e-01 6.81587338e-01 6.12093210e-01 -5.58339655e-01
-5.34839451e-01 -7.36082077e-01 1.51295643e-02 6.90214574e-01
4.57140803e-01 -5.22259176e-01 -3.10336292e-01 5.67065418e-01] | [7.546515941619873, 8.038717269897461] |
581f665b-7bb5-4ca4-ad7c-2a26adc25b20 | resnet18-model-with-sequential-layer-for | null | null | https://ijcrt.org/papers/IJCRT2205235.pdf | https://ijcrt.org/papers/IJCRT2205235.pdf | Resnet18 Model With Sequential Layer For Computing Accuracy On Image Classification Dataset | This residual network has been a broad domain of research in deep learning. Many complex architectures are based upon
residual networks. Residual networks are efficient due to skip connections. This paper highlights the addition of a sequential layer
to the traditional RESNET 18 model for computing the accuracy of an Image classification dataset. The classification datasets such
as Intel Scene dataset, CIFAR10 dataset, etc. These datasets consist of images belonging to various classes. In classification, we
assign the pictures to their respective categories. The addition of a sequential layer gives the accuracy in the range of 0 to 1, which
helps find the accuracy of prediction for the test set. | ['Dr. Laxman Sahoo', 'Dr. Venkateswara Rao Gurrala', 'Allena Venkata Sai Abhishek'] | 2022-05-01 | null | null | null | ijcrt-2022-5 | ['image-manipulation-detection', 'image-smoothing', 'image-matting', 'image-variation', 'image-stitching', 'image-augmentation', 'image-manipulation', 'image-morphing', 'roi-based-image-generation', 'image-cropping', 'detecting-image-manipulation'] | ['computer-vision', 'computer-vision', 'computer-vision', 'computer-vision', 'computer-vision', 'computer-vision', 'computer-vision', 'computer-vision', 'computer-vision', 'computer-vision', 'computer-vision'] | [-2.06984188e-02 -2.85756618e-01 8.65560770e-02 -6.79913044e-01
2.95978397e-01 -2.19453216e-01 3.91190529e-01 -7.76226223e-02
-6.55031979e-01 7.63061523e-01 1.46625340e-01 -2.10860178e-01
-7.85926133e-02 -1.00694311e+00 -4.53754127e-01 -5.35019219e-01
6.94944933e-02 6.28218800e-02 5.01886845e-01 -5.31370491e-02
4.41361427e-01 6.58577561e-01 -1.44324374e+00 6.25886500e-01
3.98486823e-01 1.49267733e+00 3.98112327e-01 5.40345311e-01
-1.89915702e-01 1.37755752e+00 -8.82057369e-01 -7.82013386e-02
4.94065523e-01 1.08585330e-02 -8.73425364e-01 -4.10186827e-01
3.45603198e-01 -3.36819202e-01 -6.44965172e-01 1.11417484e+00
3.69522810e-01 3.48453671e-01 5.62084556e-01 -1.30672741e+00
-7.42775500e-01 6.22490346e-01 -2.77305245e-01 6.05933964e-01
5.80113605e-02 -3.21165502e-01 6.16691530e-01 -8.95763338e-01
4.30445403e-01 1.11394048e+00 7.91267216e-01 2.21409366e-01
-7.00189352e-01 -9.02218103e-01 -8.82768109e-02 5.39106011e-01
-1.43717515e+00 -1.57080248e-01 7.23519802e-01 -4.74255353e-01
1.05632973e+00 1.71341717e-01 5.39073527e-01 5.82759202e-01
5.41766644e-01 4.48742062e-01 1.05012035e+00 -9.99821648e-02
1.16511308e-01 9.74415094e-02 9.25684333e-01 4.27943945e-01
1.98538452e-01 1.54652642e-02 5.02986684e-02 3.66774321e-01
7.91988671e-01 8.49984169e-01 -1.81366682e-01 1.94912925e-01
-9.03059363e-01 8.13125610e-01 1.12708700e+00 4.06700552e-01
-3.68109703e-01 2.19328597e-01 5.24655759e-01 6.28863692e-01
2.20865145e-01 3.96737158e-01 -6.02685630e-01 5.22904456e-01
-7.09876955e-01 -7.20863044e-02 4.60859418e-01 7.36802578e-01
8.85462284e-01 3.69403034e-01 -9.12733935e-03 1.02094007e+00
1.29802629e-01 -3.05147618e-02 1.12433755e+00 -6.84218109e-01
2.24992171e-01 1.00327599e+00 -2.49245286e-01 -1.18274498e+00
-5.35633028e-01 -7.07084715e-01 -1.37527168e+00 3.56796980e-01
-9.72518325e-02 -1.30222172e-01 -1.37746835e+00 1.19155264e+00
2.75677983e-02 6.04914486e-01 4.52198893e-01 6.86413288e-01
1.51112688e+00 8.01469386e-01 6.16807826e-02 2.50004619e-01
1.20727575e+00 -1.16738796e+00 -3.73357385e-01 -5.26667461e-02
3.77411962e-01 -8.26271117e-01 6.71915889e-01 4.47903484e-01
-4.89922255e-01 -1.18428922e+00 -1.06334698e+00 -5.73409051e-02
-8.19606423e-01 3.61002862e-01 7.69511878e-01 2.26374477e-01
-1.29126430e+00 7.92076528e-01 -3.65232825e-01 -1.23700790e-01
4.26585197e-01 6.95980906e-01 -6.35991096e-01 -1.33296221e-01
-1.25179505e+00 9.94650245e-01 8.12504113e-01 2.92724758e-01
-6.95012867e-01 -4.51485485e-01 -7.66704381e-01 6.90367371e-02
-3.32849592e-01 -2.31831193e-01 9.23451066e-01 -1.44471622e+00
-1.22579217e+00 8.20653915e-01 3.17231089e-01 -7.38964200e-01
3.31446618e-01 -6.25732616e-02 -6.09256566e-01 -9.34674889e-02
-1.28284007e-01 8.41237009e-01 6.79532826e-01 -7.19631553e-01
-7.51299024e-01 -7.65216127e-02 3.58119160e-01 1.18627384e-01
-7.07054958e-02 -2.94172671e-02 1.32220358e-01 -6.07490540e-01
2.37582400e-01 -9.03533697e-01 -5.43611050e-01 -5.18267870e-01
-3.32323581e-01 -5.41830063e-01 1.13235080e+00 -3.98370266e-01
8.48550022e-01 -2.31188393e+00 -3.37250412e-01 -1.16286881e-03
4.24119055e-01 3.83977741e-01 -1.43689826e-01 1.03668158e-03
-6.10636532e-01 2.55878389e-01 2.53198773e-01 2.69471616e-01
-3.88928026e-01 -5.56296706e-02 -2.85312265e-01 3.87043685e-01
-1.18453525e-01 6.53981686e-01 -4.88446504e-01 -2.64417440e-01
5.64022303e-01 5.34016430e-01 -2.46696591e-01 3.30533624e-01
3.35432172e-01 1.86395139e-01 -3.92336100e-01 2.17184097e-01
8.87630045e-01 -1.38189137e-01 -2.99557596e-01 -4.23799723e-01
-7.01031163e-02 1.36383340e-01 -1.09420371e+00 1.17373478e+00
-4.61629361e-01 8.06229234e-01 -3.89306962e-01 -1.14119482e+00
1.26708567e+00 1.79737836e-01 2.34889954e-01 -6.54495358e-01
3.91173273e-01 -3.88881541e-03 3.41709197e-01 -1.64811626e-01
7.59634554e-01 1.18520543e-01 6.63533881e-02 9.81099159e-02
1.14216775e-01 2.38930210e-01 1.37950301e-01 -2.19805725e-02
1.16416323e+00 -2.88376033e-01 1.43165231e-01 -3.58241618e-01
5.93661308e-01 -5.55223785e-02 5.08272886e-01 6.72027349e-01
-1.24315150e-01 5.59608698e-01 1.29233286e-01 -1.19009888e+00
-8.97151530e-01 -6.57545567e-01 -4.12045985e-01 8.53482068e-01
1.25445992e-01 -6.52534887e-02 -4.09621418e-01 -4.91863966e-01
-2.79845864e-01 3.01961750e-01 -5.80064774e-01 -2.79477865e-01
-5.61527312e-01 -6.85103476e-01 3.74152303e-01 6.63842022e-01
1.29569972e+00 -1.63322282e+00 -7.56521821e-01 -2.27936395e-02
2.53379077e-01 -9.82323587e-01 4.65109125e-02 5.06043792e-01
-9.88725603e-01 -1.29180491e+00 -5.72709560e-01 -1.39899218e+00
7.66049147e-01 3.23435366e-01 1.05121720e+00 4.32994515e-01
-1.88853860e-01 -1.36260837e-01 -4.60328072e-01 -9.46904719e-02
1.78133454e-02 1.84747100e-01 -8.79696980e-02 -2.65295625e-01
4.56247717e-01 -3.43661517e-01 -7.64719546e-01 2.65933186e-01
-7.70902216e-01 5.03868842e-03 3.81552190e-01 6.73896372e-01
4.07326579e-01 5.57467639e-01 3.86949629e-01 -1.06894648e+00
6.15194142e-01 -6.98525965e-01 -4.36710864e-01 -1.90760158e-02
-3.59126210e-01 -1.60171330e-01 1.03331387e+00 -4.40896511e-01
-6.23931289e-01 1.67490035e-01 -3.02307785e-01 -4.94609267e-01
-2.77426422e-01 5.58384895e-01 2.25668296e-01 -1.40591890e-01
4.56782579e-01 1.80267040e-02 -4.81174707e-01 -3.77430052e-01
-1.71319414e-02 6.65291965e-01 5.50995946e-01 -1.78505704e-02
3.16645473e-01 -2.46424302e-01 -9.13051516e-02 -6.76590085e-01
-6.73156321e-01 -3.19031358e-01 -6.74508154e-01 -4.02905084e-02
8.12667608e-01 -8.87031376e-01 -5.63141465e-01 7.18495846e-01
-1.08355784e+00 -4.46008086e-01 -3.15271258e-01 5.98535478e-01
2.97716316e-02 -2.45265700e-02 -8.78831267e-01 -2.06185818e-01
-6.09729409e-01 -1.43408537e+00 3.84583175e-01 5.72882116e-01
6.50210008e-02 -9.28441584e-01 -1.10892288e-01 -2.29224805e-02
5.94527960e-01 9.36289951e-02 6.18878961e-01 -9.14066732e-01
-3.31538290e-01 -5.45169473e-01 -4.71174806e-01 6.01587832e-01
9.50990543e-02 1.02820378e-02 -1.01149738e+00 -2.52460450e-01
1.60005525e-01 -3.59677166e-01 1.24316835e+00 4.11054462e-01
1.74153364e+00 -2.35824227e-01 -2.35250607e-01 7.74340034e-01
1.76121795e+00 5.82142413e-01 1.09837246e+00 4.11959678e-01
7.65524209e-01 2.06581369e-01 2.63459116e-01 2.75990497e-02
1.60161361e-01 3.18734139e-01 5.18773317e-01 -2.49928057e-01
-1.87332332e-01 1.17982812e-01 2.64089312e-02 8.84596288e-01
-3.99954654e-02 -2.69101709e-01 -9.09761190e-01 2.13701740e-01
-1.45193791e+00 -1.02594614e+00 -4.11279827e-01 1.90321493e+00
4.25740719e-01 2.44414195e-01 -4.22025889e-01 4.04801130e-01
9.45710957e-01 3.01700681e-01 -4.26857948e-01 -4.59983647e-01
7.52668455e-02 5.39485872e-01 6.77761316e-01 2.47873202e-01
-1.31429315e+00 1.19935739e+00 7.22850084e+00 7.53592610e-01
-1.63544130e+00 -5.82806207e-02 1.02119386e+00 4.71288174e-01
2.79256463e-01 -6.09317608e-02 -8.05469036e-01 5.39415598e-01
1.10934711e+00 1.92707330e-01 2.33758271e-01 1.11572862e+00
-2.04463109e-01 -1.35733262e-01 -9.54056978e-01 1.10507894e+00
1.49188926e-02 -1.31266892e+00 1.67886987e-01 -2.25255981e-01
7.30054617e-01 3.07225734e-01 1.01371929e-01 6.21234417e-01
6.27262354e-01 -1.53544807e+00 4.42842007e-01 5.87392390e-01
8.97007585e-01 -8.38609517e-01 1.34698629e+00 1.60258368e-01
-1.28218520e+00 -3.40069383e-01 -1.15845978e+00 -4.18744415e-01
-5.21455705e-01 4.14065003e-01 -7.21993506e-01 5.04648723e-02
9.68384981e-01 1.06535017e+00 -7.08910704e-01 1.19812667e+00
-1.37412250e-01 5.09897113e-01 -1.22525290e-01 1.41561642e-01
3.00688267e-01 -1.03583887e-01 -1.96298495e-01 9.91932213e-01
2.18166530e-01 3.65659446e-01 4.63626862e-01 2.75178790e-01
-5.40554464e-01 8.50995257e-02 -7.23668873e-01 4.42987621e-01
4.76986557e-01 1.37518048e+00 -9.99560952e-01 -5.63063860e-01
-1.94843292e-01 7.49249339e-01 3.93067151e-01 1.95581764e-01
-5.72100401e-01 -8.44838798e-01 2.67133772e-01 -1.85641900e-01
2.15487227e-01 -5.80164138e-04 -3.22773337e-01 -8.48741651e-01
-4.15109187e-01 -7.03633487e-01 3.62562001e-01 -9.04434562e-01
-9.56518948e-01 1.14685750e+00 -3.69843900e-01 -1.15960217e+00
2.17222214e-01 -8.49901438e-01 -9.68679845e-01 7.03892171e-01
-1.45361769e+00 -7.12089598e-01 -7.92653859e-01 9.32067752e-01
8.43057573e-01 -5.19204438e-01 7.18172908e-01 4.12701458e-01
-6.99070871e-01 4.06735539e-01 6.57325983e-02 7.24374592e-01
4.30466533e-01 -1.01602364e+00 1.69679686e-01 6.50550544e-01
-1.18541516e-01 5.48668802e-01 3.86701554e-01 -4.49317247e-01
-7.99925148e-01 -1.32580900e+00 5.36902249e-01 -9.97642800e-02
3.49825233e-01 2.14312479e-01 -7.99924731e-01 9.30904090e-01
3.67280930e-01 5.07049918e-01 4.39584106e-01 -1.58679083e-01
-1.40707836e-01 -4.73852634e-01 -1.37333691e+00 1.92535818e-01
6.94937050e-01 -3.12735021e-01 -6.59553230e-01 2.51533031e-01
7.36939192e-01 -3.48890364e-01 -8.18841398e-01 4.60412771e-01
3.61465007e-01 -8.45227540e-01 9.28721726e-01 -5.92898726e-01
6.09932721e-01 -3.29768389e-01 -3.19698930e-01 -1.30270922e+00
-6.08255506e-01 3.84425014e-01 4.93432760e-01 9.19239581e-01
1.94195285e-01 -6.58770919e-01 7.78575599e-01 4.67068821e-01
-1.44968793e-01 -5.41127264e-01 -6.87671423e-01 -2.78450400e-01
-1.36497259e-01 -2.02784449e-01 6.53922260e-01 1.04735029e+00
-5.36917031e-01 5.82014918e-01 -3.77124459e-01 -1.62341863e-01
2.08284870e-01 -1.57928206e-02 5.21165848e-01 -1.66829729e+00
3.07841510e-01 -3.70396942e-01 -1.00189281e+00 -8.35996866e-01
2.25730315e-01 -1.06812894e+00 -1.54496923e-01 -1.69715023e+00
1.43179610e-01 -1.02186453e+00 -9.09699321e-01 6.41023278e-01
1.37641713e-01 7.43931353e-01 3.90414596e-02 4.72740471e-01
-5.68455160e-01 1.92286327e-01 8.96765828e-01 -3.46928149e-01
-1.77446529e-01 1.79830596e-01 -4.01447445e-01 1.01796925e+00
1.21743894e+00 -6.07156634e-01 -5.44865906e-01 -5.35432577e-01
4.57466505e-02 -2.90256366e-02 2.23449826e-01 -1.31515408e+00
4.30292457e-01 -9.82351154e-02 9.96849477e-01 -7.31643319e-01
1.35614783e-01 -1.21909797e+00 4.37814444e-01 6.79836452e-01
-5.55640042e-01 4.26439375e-01 -2.81776339e-02 3.06683511e-01
-4.38009977e-01 -5.38301945e-01 9.87867594e-01 -5.75583160e-01
-1.33155966e+00 3.75010133e-01 -1.16996489e-01 -2.23866925e-01
1.06399214e+00 -3.68222147e-01 -4.31099415e-01 -7.63528375e-03
-5.85278153e-01 7.10996939e-03 1.22830689e-01 3.76360506e-01
9.11224544e-01 -1.27985692e+00 -5.40776312e-01 2.01804146e-01
-4.01494473e-01 8.94102529e-02 1.52695715e-01 3.63441974e-01
-1.05270839e+00 3.29768091e-01 -7.99769938e-01 -5.10615289e-01
-1.29523063e+00 5.02029836e-01 6.66478574e-01 -3.88559043e-01
-6.48656905e-01 9.17287409e-01 1.98296353e-01 -4.65550274e-01
2.15908557e-01 -3.07729512e-01 -9.82689023e-01 -1.08309373e-01
5.90553224e-01 4.40558344e-01 1.01281337e-01 -7.12648392e-01
-3.68894935e-01 5.27524948e-01 -2.56613195e-01 5.77250540e-01
1.44933140e+00 2.67422825e-01 -3.78950953e-01 4.16279286e-01
1.32631743e+00 -3.60813826e-01 -7.63527930e-01 -2.16878027e-01
-1.02757342e-01 -1.86282188e-01 2.80234486e-01 -8.08489799e-01
-1.75149786e+00 8.10831726e-01 9.95460033e-01 1.18862569e-01
1.12567782e+00 -4.72402811e-01 5.44732869e-01 5.49691737e-01
3.22346926e-01 -9.08273280e-01 3.56486836e-03 9.45030451e-01
7.54144847e-01 -1.24245095e+00 6.73245117e-02 2.07814008e-01
-5.51732242e-01 1.32714295e+00 8.38325322e-01 -9.45341527e-01
1.05726922e+00 2.01520354e-01 8.01294371e-02 -1.41488805e-01
-6.10754907e-01 8.18467215e-02 1.07734248e-01 3.51420015e-01
6.43555462e-01 1.95835903e-01 -3.20771009e-01 4.85828370e-01
-6.56892300e-01 1.84672363e-02 5.81347108e-01 6.35468245e-01
-6.73749387e-01 -7.71655798e-01 -1.65913641e-01 7.67978311e-01
-7.16540217e-01 -3.52183491e-01 -1.66575722e-02 6.60600364e-01
3.68774921e-01 9.09023583e-01 3.51480871e-01 -6.76752269e-01
1.76011994e-01 -2.06390396e-01 -2.96781790e-02 -5.46936393e-01
-8.65961969e-01 -7.26489186e-01 -3.55356157e-01 -3.84046674e-01
-2.33685374e-01 1.11374356e-01 -1.29782653e+00 -4.14700478e-01
-2.35117719e-01 3.54150355e-01 7.23124206e-01 7.93137193e-01
6.10634536e-02 7.93141246e-01 7.53890812e-01 -6.93167806e-01
-1.00025311e-01 -1.19482327e+00 -4.54241037e-01 2.09507853e-01
3.31810117e-01 -5.12565315e-01 -4.14101452e-01 -1.66830853e-01] | [9.269281387329102, 2.198275566101074] |
f21d00fb-859d-4663-95bc-17eaaf382174 | revisiting-shallow-discourse-parsing-in-the-1 | 2204.00350 | null | https://arxiv.org/abs/2204.00350v1 | https://arxiv.org/pdf/2204.00350v1.pdf | Revisiting Shallow Discourse Parsing in the PDTB-3: Handling Intra-sentential Implicits | In the PDTB-3, several thousand implicit discourse relations were newly annotated \textit{within} individual sentences, adding to the over 15,000 implicit relations annotated \textit{across} adjacent sentences in the PDTB-2. Given that the position of the arguments to these \textit{intra-sentential implicits} is no longer as well-defined as with \textit{inter-sentential implicits}, a discourse parser must identify both their location and their sense. That is the focus of the current work. The paper provides a comprehensive analysis of our results, showcasing model performance under different scenarios, pointing out limitations and noting future directions. | ['Bonnie Webber', 'Zheng Zhao'] | 2022-04-01 | revisiting-shallow-discourse-parsing-in-the | https://aclanthology.org/2021.codi-main.10 | https://aclanthology.org/2021.codi-main.10.pdf | codi-2021-11 | ['discourse-parsing', 'implicit-relations'] | ['natural-language-processing', 'natural-language-processing'] | [ 4.83824372e-01 1.35529518e+00 -4.33360845e-01 -3.96182746e-01
-4.62531507e-01 -1.02706587e+00 9.62470055e-01 7.24812329e-01
-3.55408102e-01 1.31866562e+00 9.03416216e-01 -8.41655850e-01
-4.78541136e-01 -4.65839922e-01 -3.95318717e-01 -2.23887235e-01
-1.06320314e-01 6.89249158e-01 7.02788472e-01 -5.10498106e-01
4.32579488e-01 -5.66274449e-02 -9.44637418e-01 8.22032869e-01
7.93985724e-01 4.15537804e-01 2.38703534e-01 6.11455977e-01
-3.98592710e-01 1.58804381e+00 -1.36688495e+00 -6.17080331e-01
-4.50958133e-01 -3.59654427e-01 -1.60819125e+00 -2.68823415e-01
1.59659743e-01 -2.05932379e-01 -3.94727707e-01 7.40037620e-01
8.52593407e-02 2.84590684e-02 5.42454720e-01 -1.02246499e+00
-4.84836102e-01 1.16036963e+00 -1.35880694e-01 6.79701984e-01
5.92308998e-01 -3.50539386e-01 1.43795300e+00 -5.99100590e-01
9.85215425e-01 1.45018089e+00 6.02325320e-01 4.54640061e-01
-1.11406314e+00 -1.83734894e-01 6.06619835e-01 4.08795774e-02
-8.86580229e-01 -8.26335549e-01 4.27024245e-01 -5.46947718e-01
1.48990941e+00 8.08458567e-01 3.67889434e-01 8.80205870e-01
-3.00035149e-01 6.93165481e-01 1.12362218e+00 -7.47269869e-01
-9.58853066e-02 -5.07432036e-02 8.95078838e-01 2.37305239e-01
3.07232767e-01 -3.06539148e-01 -5.88190436e-01 -2.29569867e-01
5.57982743e-01 -9.07224238e-01 -2.61743009e-01 5.20381451e-01
-1.04898465e+00 7.91288018e-01 9.34571549e-02 7.34153092e-01
1.81872264e-01 1.71666116e-01 7.08329558e-01 1.17851704e-01
6.10736966e-01 4.54866886e-01 -5.83484530e-01 -4.74295914e-01
-5.06617188e-01 3.01514536e-01 1.10472000e+00 1.23231769e+00
5.47735631e-01 -2.59855986e-01 -7.37402067e-02 7.62539804e-01
2.51331151e-01 -8.45684484e-02 6.52944148e-02 -1.49127650e+00
8.81677389e-01 7.60303915e-01 4.68026847e-01 -1.12364912e+00
-3.33425134e-01 9.45542380e-02 -4.20108646e-01 -3.40294749e-01
6.03308618e-01 -2.38436937e-01 -2.40168959e-01 1.67603743e+00
1.63911939e-01 -3.18907917e-01 1.79238513e-01 4.50546533e-01
1.03561842e+00 5.94574392e-01 4.57958668e-01 -6.97152376e-01
1.45966887e+00 -6.53005123e-01 -1.24409151e+00 -4.59423840e-01
1.10312998e+00 -8.39921653e-01 7.71111965e-01 -3.25637698e-01
-1.05216420e+00 -1.91392154e-02 -1.03235185e+00 -4.87305015e-01
-1.37526438e-01 -2.99203008e-01 5.44477522e-01 3.67738873e-01
-8.65465462e-01 1.60354063e-01 -6.51924551e-01 -1.50501296e-01
1.13865338e-01 3.25021595e-01 -1.34422243e-01 6.09284341e-01
-1.53101695e+00 1.25885832e+00 8.87548745e-01 1.57419741e-01
-8.04119259e-02 -4.29665178e-01 -9.16221201e-01 -1.77224919e-01
6.68400466e-01 -4.45351094e-01 1.47469699e+00 -9.12267566e-01
-1.25110185e+00 1.06614280e+00 -6.00686014e-01 -6.73424125e-01
3.06162477e-01 -5.20535171e-01 -3.34186941e-01 1.83974415e-01
3.73651385e-01 2.49952868e-01 7.97189325e-02 -1.23274255e+00
-7.39369750e-01 -1.04706816e-01 7.26970017e-01 4.85249668e-01
-5.06262444e-02 7.06246853e-01 9.16264951e-02 -7.55888581e-01
3.67921770e-01 -5.71954250e-01 8.87063295e-02 -6.98824286e-01
-7.12911010e-01 -8.67025256e-01 1.12374902e+00 -6.84116602e-01
1.96605623e+00 -1.88477147e+00 2.13254675e-01 -2.18960121e-01
4.07693624e-01 4.19043779e-01 3.19514573e-01 8.17164958e-01
-2.81133294e-01 6.16141856e-01 -2.99845248e-01 -2.06540421e-01
7.30752572e-02 7.69518495e-01 -6.78739905e-01 2.30761915e-01
2.72078037e-01 9.46402073e-01 -9.37213600e-01 -7.13134825e-01
4.72493023e-02 -2.78804302e-01 -1.04224220e-01 -2.60633808e-02
-6.10993326e-01 2.97790289e-01 -6.04345620e-01 1.66866407e-01
1.78533226e-01 -1.89221844e-01 8.67402792e-01 -4.39723693e-02
-4.89073336e-01 1.58219934e+00 -8.42036724e-01 1.03315866e+00
2.02119350e-01 1.01941311e+00 3.01406384e-01 -1.17385578e+00
7.04704225e-01 7.30092168e-01 2.48171583e-01 -4.40148503e-01
-1.41584594e-02 2.81728983e-01 4.53861088e-01 -6.77832961e-01
8.09250712e-01 -2.23090723e-01 -3.99735838e-01 5.41522026e-01
-4.50152516e-01 -1.32867694e-01 6.18159056e-01 5.63433766e-01
1.19946098e+00 -6.15428016e-03 7.13937819e-01 -6.48125768e-01
6.40671730e-01 4.82930154e-01 6.40544295e-01 5.65456986e-01
-1.10167079e-01 3.05553883e-01 1.23556519e+00 -3.14937741e-01
-1.00390315e+00 -8.13480556e-01 -2.59579033e-01 8.98427308e-01
1.95452511e-01 -1.00836527e+00 -6.35766745e-01 -7.78727293e-01
-3.15994412e-01 1.10077226e+00 -6.57575905e-01 6.01040184e-01
-1.47364366e+00 -5.27221024e-01 1.00383627e+00 4.26882148e-01
4.80254412e-01 -1.06201911e+00 -6.82787836e-01 3.16310763e-01
-6.88282967e-01 -1.03665400e+00 1.05454005e-01 3.17892522e-01
-5.31639457e-01 -1.37635100e+00 -2.88174767e-02 -7.00489879e-01
2.84568429e-01 -2.40907475e-01 1.33590782e+00 4.71745014e-01
3.71083140e-01 5.94506674e-02 -5.34882486e-01 -4.96764809e-01
-6.70576394e-01 1.17878400e-01 -6.33462310e-01 -8.22228372e-01
3.01250935e-01 -3.07177275e-01 -1.41775846e-01 4.21938673e-02
-5.28051078e-01 2.10776627e-01 -2.77829915e-01 8.14713597e-01
1.41029328e-01 -1.96183875e-01 7.77250469e-01 -1.39771926e+00
8.60958099e-01 -3.87370318e-01 -1.68143839e-01 2.43808225e-01
-5.46472482e-02 -4.14722860e-01 -1.06671743e-01 -1.78866774e-01
-1.34104764e+00 -8.36225569e-01 -2.34648302e-01 6.79727674e-01
-6.35257810e-02 8.86051059e-01 -1.37166619e-01 1.00099778e+00
5.36752284e-01 -5.24581552e-01 -1.48289293e-01 -2.95690686e-01
2.83711582e-01 5.01770020e-01 3.59233379e-01 -9.13513839e-01
4.49877977e-01 1.02335215e-01 -3.73967767e-01 -1.10582781e+00
-1.41658974e+00 -2.87675083e-01 -8.68564546e-01 -8.36527627e-03
1.05632269e+00 -6.47082329e-01 -6.90779448e-01 2.04737961e-01
-1.94611943e+00 -7.78582573e-01 -5.65170050e-01 1.32751688e-01
-4.79129672e-01 6.10883534e-01 -9.82962906e-01 -1.07326567e+00
-1.41217010e-02 -8.85701239e-01 7.81715751e-01 -1.41674206e-01
-1.46450424e+00 -1.27439964e+00 -6.97076172e-02 5.01524687e-01
-1.43816262e-01 2.42756084e-01 1.41706634e+00 -8.81151736e-01
-3.59494239e-01 3.40503901e-01 -3.31355125e-01 7.97894746e-02
4.60262418e-01 -7.13669807e-02 -7.56690562e-01 8.18674192e-02
1.20683588e-01 -3.01452905e-01 3.10235441e-01 2.30491087e-01
5.90197206e-01 -8.66635561e-01 -5.13719916e-01 -2.35683173e-01
7.57801175e-01 1.97991222e-01 8.77940595e-01 7.13472068e-01
2.84861386e-01 1.06592453e+00 7.57067859e-01 -5.58944233e-02
5.06462872e-01 6.36837661e-01 2.00543523e-01 2.22150996e-01
-2.92467922e-01 2.44497702e-01 3.36152494e-01 8.21713388e-01
-2.06817657e-01 -4.50412154e-01 -9.45670187e-01 6.81109309e-01
-2.00604987e+00 -1.30337691e+00 -8.57118428e-01 1.49163294e+00
1.29682744e+00 3.29546690e-01 -1.36470795e-01 4.43985850e-01
7.78502882e-01 6.91283107e-01 3.29397060e-02 -6.45501018e-01
-5.53317070e-01 3.17426354e-01 -1.17246084e-01 1.04652619e+00
-1.07505298e+00 1.14315987e+00 6.82210779e+00 5.80891430e-01
-6.54178619e-01 1.48969203e-01 5.74360371e-01 1.75301969e-01
-4.36387807e-01 4.06859905e-01 -6.63294792e-01 2.16968998e-01
7.10397005e-01 -3.02720040e-01 -1.49053723e-01 2.91432798e-01
1.93581283e-01 -6.22413337e-01 -1.10745287e+00 4.12670299e-02
-2.94503838e-01 -1.87603474e+00 -2.44769335e-01 -1.04973331e-01
4.60874170e-01 -2.10611016e-01 -2.98823982e-01 2.25203380e-01
4.81477797e-01 -1.14864540e+00 1.16404510e+00 -9.08234417e-02
7.60068715e-01 -1.99755237e-01 5.83469808e-01 5.18847525e-01
-9.89047110e-01 1.02548003e-01 2.08137725e-02 -8.05747032e-01
5.37600160e-01 3.54810715e-01 -8.61079931e-01 6.53119445e-01
6.27099454e-01 5.61597407e-01 -1.51105434e-01 1.46982059e-01
-8.26965451e-01 9.50551271e-01 -2.01317608e-01 -1.60853788e-01
1.96244672e-01 -5.35995923e-02 9.82421756e-01 1.32468569e+00
-6.72742501e-02 8.87968540e-01 5.77575155e-02 5.69556892e-01
1.09593593e-01 -2.26482481e-01 -3.54818106e-01 -1.41280532e-01
8.98074448e-01 6.53608799e-01 -8.26285422e-01 -6.03483796e-01
-7.20568001e-02 4.58862245e-01 4.91632074e-01 5.07848203e-01
-6.27870739e-01 -7.19169602e-02 4.64819998e-01 4.44360465e-01
-1.81598421e-02 -4.72748578e-01 -6.14895463e-01 -6.60470247e-01
7.57399946e-02 -8.14730406e-01 5.89163899e-01 -6.99362993e-01
-1.19763923e+00 4.88537937e-01 5.91667712e-01 -4.25372809e-01
-3.37651640e-01 -5.21425068e-01 -5.45023680e-01 1.03913081e+00
-1.09864175e+00 -1.00546813e+00 1.56566814e-01 1.46235511e-01
5.22633910e-01 2.86904842e-01 1.16097343e+00 1.54730128e-02
-3.02408338e-01 1.17935695e-01 -5.21531105e-01 5.52496016e-01
5.90493858e-01 -1.29700923e+00 2.78445959e-01 5.88516235e-01
-2.33417809e-01 1.11490953e+00 1.07610524e+00 -8.92838717e-01
-5.55524230e-01 -6.43056154e-01 1.50328958e+00 -7.68865287e-01
1.30735791e+00 -1.55032858e-01 -1.07176363e+00 1.34574819e+00
1.01428008e+00 -5.95147848e-01 7.57117450e-01 5.20997643e-01
-1.58852354e-01 5.47056258e-01 -7.28536248e-01 6.76728189e-01
1.19779313e+00 -5.89941859e-01 -1.53134608e+00 2.62040764e-01
1.25392926e+00 -8.45769405e-01 -9.01079178e-01 6.87377453e-01
1.03411553e-02 -9.82178628e-01 6.86844170e-01 -6.21178687e-01
5.30194938e-01 -1.75361902e-01 -1.91813648e-01 -6.31331980e-01
1.01509854e-01 -8.67342293e-01 -2.64252484e-01 1.67360628e+00
7.72845984e-01 -6.74026489e-01 6.42560005e-01 8.71862471e-01
-6.61005378e-01 -5.31119883e-01 -1.24223745e+00 -2.70329237e-01
4.05486315e-01 -4.12698209e-01 3.54520679e-01 1.21859109e+00
6.36363089e-01 9.45380867e-01 2.13247791e-01 -1.83272526e-01
1.65508032e-01 -2.23045964e-02 4.92480427e-01 -1.15644169e+00
-2.16058627e-01 -3.73849332e-01 1.28632471e-01 -1.46549642e+00
3.60069871e-01 -7.36276865e-01 -2.24214956e-01 -1.85155249e+00
-1.96087714e-02 -1.09752715e+00 2.78205782e-01 5.53235114e-01
-1.31067753e-01 -8.12300220e-02 1.94127500e-01 2.01386705e-01
-5.15211761e-01 4.66321826e-01 1.15309203e+00 -2.59299248e-01
-1.94335386e-01 -1.76115915e-01 -7.03169465e-01 1.11036360e+00
6.92756236e-01 -3.88469666e-01 -5.62895715e-01 -6.33885384e-01
2.23203793e-01 2.31562853e-01 1.17919564e-01 -2.48773083e-01
3.64617892e-02 -3.95883471e-01 -1.90721974e-01 -9.30059135e-01
5.43878853e-01 -2.88664728e-01 -1.54178843e-01 3.29302639e-01
-6.31569207e-01 2.24421173e-01 5.05150259e-01 2.52851576e-01
-3.54485482e-01 -6.67975605e-01 6.86694905e-02 -1.65435433e-01
-6.92571580e-01 -6.97033882e-01 -9.52349424e-01 5.77752650e-01
9.60507870e-01 -3.86830032e-01 -1.11042416e+00 -5.15630543e-02
-8.85108411e-01 6.74980581e-02 2.09876552e-01 1.09572276e-01
3.04149449e-01 -7.13531017e-01 -5.64307690e-01 -6.44294679e-01
-2.86776185e-01 5.46334088e-01 -9.14221480e-02 9.55730975e-01
-4.38954532e-01 5.45573413e-01 4.43994790e-01 -3.93273205e-01
-1.72660089e+00 1.10767290e-01 1.77281290e-01 -5.94888628e-01
-5.86715043e-01 9.01981890e-01 3.32327276e-01 -2.95108259e-01
3.20266075e-02 -1.62357852e-01 -2.85493672e-01 2.79585600e-01
2.75183201e-01 3.55234087e-01 -2.69583255e-01 -9.57315743e-01
-3.19850713e-01 -1.62460357e-01 3.32691669e-02 -3.28904122e-01
1.18793988e+00 -5.46365678e-01 -1.03964317e+00 7.92764843e-01
6.83145821e-01 3.55310798e-01 -6.25954688e-01 5.34931757e-02
9.03824091e-01 -1.88500620e-03 -6.97667122e-01 -8.73496592e-01
-4.57672589e-02 5.77495575e-01 -5.93841314e-01 7.36150682e-01
5.18870413e-01 4.05997485e-01 3.44731987e-01 4.00230229e-01
1.78852484e-01 -1.17106390e+00 2.03918606e-01 1.14183128e+00
1.20231438e+00 -6.03974044e-01 3.64286691e-01 -1.32707345e+00
-4.32461470e-01 8.98983300e-01 7.37254977e-01 1.23745658e-01
4.51718479e-01 5.85673571e-01 1.13584744e-02 -6.22378409e-01
-7.94499815e-01 2.13370129e-01 -6.61181137e-02 4.97926563e-01
1.10402179e+00 1.28112853e-01 -8.61587226e-01 4.93097037e-01
-6.08382165e-01 -5.82831502e-01 6.63690388e-01 1.17316973e+00
-4.36118513e-01 -1.55723679e+00 -2.91010380e-01 1.15197539e-01
-4.88460481e-01 -2.52568930e-01 -7.95838773e-01 1.35171473e+00
1.44634962e-01 1.48438442e+00 3.57728094e-01 1.26624495e-01
2.18439400e-01 3.54199819e-02 5.44774950e-01 -1.06037796e+00
-8.22211921e-01 9.18281749e-02 1.51411033e+00 -1.22989211e-02
-9.96172667e-01 -8.03400576e-01 -1.84584618e+00 -3.38839024e-01
-3.88524860e-01 5.42603970e-01 -1.03961647e-01 1.29531515e+00
-6.62832260e-02 6.82767570e-01 -2.98516035e-01 -3.28998208e-01
-2.93300211e-01 -1.26063812e+00 -5.87075092e-02 2.37007409e-01
1.30508795e-01 -5.81423521e-01 -1.57691285e-01 2.28289366e-01] | [10.755929946899414, 9.407858848571777] |
41732d18-2822-42bd-9b31-9c4cb8241337 | a-graph-based-approach-for-mitigating-multi | 2107.03415 | null | https://arxiv.org/abs/2107.03415v1 | https://arxiv.org/pdf/2107.03415v1.pdf | A Graph-based Approach for Mitigating Multi-sided Exposure Bias in Recommender Systems | Fairness is a critical system-level objective in recommender systems that has been the subject of extensive recent research. A specific form of fairness is supplier exposure fairness where the objective is to ensure equitable coverage of items across all suppliers in recommendations provided to users. This is especially important in multistakeholder recommendation scenarios where it may be important to optimize utilities not just for the end-user, but also for other stakeholders such as item sellers or producers who desire a fair representation of their items. This type of supplier fairness is sometimes accomplished by attempting to increasing aggregate diversity in order to mitigate popularity bias and to improve the coverage of long-tail items in recommendations. In this paper, we introduce FairMatch, a general graph-based algorithm that works as a post processing approach after recommendation generation to improve exposure fairness for items and suppliers. The algorithm iteratively adds high quality items that have low visibility or items from suppliers with low exposure to the users' final recommendation lists. A comprehensive set of experiments on two datasets and comparison with state-of-the-art baselines show that FairMatch, while significantly improves exposure fairness and aggregate diversity, maintains an acceptable level of relevance of the recommendations. | ['Robin Burke', 'Bamshad Mobasher', 'Mykola Pechenizkiy', 'Himan Abdollahpouri', 'Masoud Mansoury'] | 2021-07-07 | null | null | null | null | ['exposure-fairness'] | ['adversarial'] | [-1.50161967e-01 1.77063397e-03 -4.99001265e-01 -4.80900854e-01
-3.24815243e-01 -7.61879504e-01 2.16810301e-01 6.28603280e-01
-2.43982658e-01 5.97926259e-01 5.65109670e-01 -2.35564768e-01
-6.33046031e-01 -1.03068793e+00 -3.49973261e-01 -2.66233891e-01
-4.78857663e-04 4.18432385e-01 -2.34453250e-02 -6.80780172e-01
4.76291329e-01 4.66302007e-01 -1.23616695e+00 3.16220552e-01
1.22694397e+00 1.01109719e+00 -3.01738307e-02 2.81343102e-01
-1.38914496e-01 6.73555911e-01 -5.18660545e-01 -1.13645267e+00
8.62153590e-01 -4.39779311e-01 -3.96478295e-01 -1.13309421e-01
6.45925999e-01 -4.38371569e-01 8.58763829e-02 1.18386137e+00
4.03077394e-01 5.29603899e-01 7.16622949e-01 -1.38817620e+00
-1.00063717e+00 1.10209394e+00 -8.43338728e-01 2.60969758e-01
1.44941539e-01 -1.87042773e-01 1.78430116e+00 -7.17226207e-01
2.84046948e-01 1.13843834e+00 2.88823992e-01 8.28431025e-02
-1.22254360e+00 -7.07995057e-01 5.16453207e-01 -2.40215771e-02
-1.15000093e+00 -2.03013718e-01 3.59486163e-01 -3.58615667e-01
5.73585331e-01 8.58790874e-01 4.61546361e-01 3.06226432e-01
1.42934442e-01 4.39416528e-01 6.06547832e-01 -1.19949065e-01
3.03439647e-01 3.76707762e-01 3.86989713e-01 -7.34254047e-02
8.68210554e-01 9.32091996e-02 -4.72348094e-01 -4.29634482e-01
3.37386072e-01 3.91283900e-01 -3.31208646e-01 -2.63138235e-01
-6.41982615e-01 1.26954257e+00 7.88835287e-01 -1.58393338e-01
-9.04800236e-01 -1.57446548e-01 1.18022501e-01 6.15312994e-01
6.37119234e-01 8.59533846e-01 -1.49771571e-01 3.20059448e-01
-8.95656884e-01 4.27237868e-01 1.05243206e+00 1.15783358e+00
4.63004857e-01 -1.21132791e-01 -5.52123487e-01 7.89555252e-01
4.86202836e-01 1.87122986e-01 -2.16014564e-01 -1.04726136e+00
4.58570123e-01 6.27856135e-01 4.39850271e-01 -1.27195561e+00
-2.43757498e-02 -9.79940057e-01 -7.73987591e-01 3.65388513e-01
1.70883864e-01 -6.35055378e-02 -2.43759543e-01 1.48346734e+00
2.54037023e-01 -4.68360066e-01 -3.41783136e-01 1.46629429e+00
6.82177424e-01 7.15018034e-01 1.69551253e-01 -4.88746762e-01
1.09622371e+00 -1.13099229e+00 -5.51563203e-01 1.47577226e-01
-3.20241079e-02 -1.05110395e+00 9.14134741e-01 5.38931727e-01
-1.17082095e+00 -2.78332919e-01 -7.57042706e-01 1.35029286e-01
-2.01683462e-01 -2.96403408e-01 4.79928344e-01 8.56201589e-01
-7.00867116e-01 6.62522554e-01 2.19430804e-01 -1.96842030e-01
4.99403805e-01 4.50593799e-01 1.11898355e-01 -1.43638059e-01
-1.35749722e+00 9.80046809e-01 -2.32314840e-01 -3.54385883e-01
-3.48285198e-01 -1.02678013e+00 -3.28410953e-01 6.39143348e-01
6.05344117e-01 -9.54770625e-01 1.19349563e+00 -1.14098346e+00
-9.03117239e-01 1.01685941e-01 4.62406367e-01 -2.73322791e-01
6.40114844e-01 -1.79136634e-01 -5.44991434e-01 -4.86636519e-01
2.10655015e-02 1.90857321e-01 7.05854118e-01 -1.36128831e+00
-1.22796166e+00 -3.20484310e-01 5.38930833e-01 6.05371535e-01
-3.69163781e-01 4.83139195e-02 -1.93775207e-01 -8.70019257e-01
-3.69876027e-01 -7.67823279e-01 -4.56715405e-01 -3.04814409e-02
-2.47028396e-01 -2.63291031e-01 2.12464288e-01 -5.90550721e-01
1.58833170e+00 -1.88545966e+00 -9.05588195e-02 8.15667093e-01
4.55828935e-01 -3.39825563e-02 -2.91456968e-01 8.00479949e-01
3.52604747e-01 2.22254351e-01 3.93955886e-01 -9.78928879e-02
4.83438134e-01 -3.89064513e-02 -2.04323947e-01 5.21106124e-01
-4.11670417e-01 4.63434577e-01 -9.81076658e-01 1.61307439e-01
3.66070680e-02 7.37411231e-02 -7.67576039e-01 2.97249734e-01
-2.28385404e-01 -2.17792857e-02 -1.92021295e-01 3.92820060e-01
7.92987406e-01 -2.69546777e-01 1.23716250e-01 -3.79979670e-01
-1.23235241e-01 4.95808005e-01 -1.43111360e+00 1.03595459e+00
-3.70879292e-01 -9.27427877e-03 1.30640909e-01 -4.91899788e-01
6.99024260e-01 2.31716898e-03 6.14253402e-01 -8.22774470e-01
1.82092160e-01 5.12550771e-02 1.73981488e-01 1.42021000e-01
1.24705136e+00 -4.47935797e-02 -1.67645663e-01 7.67702222e-01
-4.72571820e-01 2.66778916e-01 4.86875176e-01 8.20260406e-01
6.76791966e-01 -4.49422181e-01 6.29040837e-01 -5.44776499e-01
1.32554367e-01 -1.62692279e-01 3.99485052e-01 7.61001587e-01
2.41127059e-01 3.36305529e-01 2.54436612e-01 -1.63077280e-01
-9.50214863e-01 -8.62808824e-01 2.76227862e-01 1.59386289e+00
4.75534290e-01 -4.62848395e-01 -1.77761286e-01 -7.76297331e-01
5.69433153e-01 1.09399152e+00 -5.58873713e-01 -1.81166083e-01
1.79437429e-01 -2.48579964e-01 -3.42654705e-01 2.37641156e-01
-1.37776673e-01 -6.62468016e-01 -2.70682544e-01 3.46784562e-01
-1.18996181e-01 -4.71843421e-01 -1.36463666e+00 -1.61792949e-01
-3.47804368e-01 -9.50527132e-01 -6.72276080e-01 -1.22892492e-01
7.32543707e-01 9.61355150e-01 1.38589811e+00 2.65775681e-01
2.67515481e-01 1.80995628e-01 -6.19533360e-01 -5.52449346e-01
-1.54143229e-01 -1.31228454e-02 1.32344961e-01 1.11386381e-01
3.46077792e-02 -3.81513625e-01 -8.93594861e-01 7.95299470e-01
-8.44074965e-01 -3.41134906e-01 2.79425830e-01 4.99120772e-01
5.83919227e-01 1.99714988e-01 9.49845135e-01 -1.34086084e+00
1.28256953e+00 -1.00452733e+00 -6.44806564e-01 3.24261993e-01
-1.23243690e+00 -3.34063679e-01 9.19090390e-01 -2.04575703e-01
-9.28247869e-01 -5.86976528e-01 -2.66414173e-02 -1.35764614e-01
4.92036104e-01 6.06817722e-01 -1.68346301e-01 -2.10818797e-02
6.16190374e-01 -5.78918040e-01 -1.09801538e-01 -4.88525271e-01
8.45005870e-01 7.27292955e-01 6.09282218e-02 -1.95852667e-01
4.61065978e-01 7.32650161e-02 -1.31056234e-01 -1.57901436e-01
-9.43633914e-01 -8.10185075e-01 8.43906254e-02 -2.73292154e-01
9.60632339e-02 -7.85605371e-01 -7.70852864e-01 -3.47341478e-01
-6.26301408e-01 7.41716474e-02 -6.93031073e-01 3.21482778e-01
-1.00470692e-01 2.31412753e-01 -2.28038371e-01 -7.45922506e-01
-7.72963822e-01 -9.05967832e-01 3.55507165e-01 4.18987751e-01
-3.72169167e-01 -7.03244746e-01 -1.31725907e-01 2.36020088e-01
7.62330711e-01 -1.70599759e-01 6.92320585e-01 -9.24068213e-01
-5.55648148e-01 -2.64650315e-01 -3.03373963e-01 2.18667984e-01
3.82913947e-01 -1.09027252e-01 -2.09471002e-01 -5.97356498e-01
-5.39608181e-01 2.10187480e-01 6.20776653e-01 5.76649547e-01
7.71317422e-01 -7.86171257e-01 4.59013227e-03 3.52277011e-02
1.36167181e+00 -6.77739158e-02 3.16253185e-01 -1.48016103e-02
3.84531051e-01 8.11732054e-01 1.15598118e+00 1.00484633e+00
6.78862691e-01 8.62896025e-01 6.78949416e-01 -2.67733425e-01
-3.02829444e-02 -7.53529444e-02 9.80870649e-02 5.77019632e-01
-1.02427997e-01 -9.39982533e-01 -5.88471396e-03 4.21810478e-01
-2.17884445e+00 -1.25629747e+00 -4.34837729e-01 2.63592625e+00
3.07256073e-01 -1.35871708e-01 7.85513937e-01 -9.72920954e-02
8.30822825e-01 -1.69534609e-01 -4.69353944e-01 -7.76570022e-01
1.29414141e-01 -2.72388190e-01 7.62355328e-01 4.76855844e-01
-5.52166820e-01 3.93828571e-01 5.84400749e+00 7.55726695e-01
-6.12808287e-01 4.24972698e-02 6.31889582e-01 -5.98105431e-01
-9.27083552e-01 -8.17713439e-02 -5.63894093e-01 7.15043783e-01
5.39530873e-01 -7.94810951e-01 8.16203475e-01 6.55074894e-01
4.37075824e-01 -9.53586865e-03 -1.01986647e+00 6.56911790e-01
-8.50870013e-02 -1.25699317e+00 2.29781605e-02 3.82015288e-01
1.07414258e+00 -1.90986052e-01 2.19925269e-01 2.17507139e-01
7.88292706e-01 -7.76575744e-01 8.31999123e-01 6.00819886e-01
3.97949010e-01 -1.21473849e+00 6.85014784e-01 3.13495129e-01
-1.24597216e+00 -2.71052778e-01 -6.79680884e-01 -1.34207055e-01
5.16098857e-01 9.48470592e-01 -4.01179433e-01 7.49303877e-01
5.29276788e-01 4.07387435e-01 -1.15760975e-01 1.43928266e+00
-2.73430217e-02 4.51295376e-01 -2.42518693e-01 -1.58355162e-01
4.80091013e-02 -4.80507046e-01 3.94913435e-01 1.09371376e+00
4.99467969e-01 4.29269433e-01 2.79170156e-01 6.08888149e-01
-6.86376274e-01 6.83795869e-01 -2.61686802e-01 1.19443834e-01
7.60537684e-01 1.58789957e+00 -3.10104817e-01 -1.77675188e-01
-4.89429355e-01 5.61171055e-01 1.86033607e-01 7.66447932e-02
-7.80702114e-01 -1.81366682e-01 9.22397733e-01 4.07333165e-01
2.47116268e-01 3.68217021e-01 -3.64972204e-01 -6.94448590e-01
-3.47229004e-01 -1.14083660e+00 8.44257236e-01 -2.94888139e-01
-1.96022964e+00 2.34481797e-01 -3.37713748e-01 -1.38865840e+00
1.83287650e-01 7.94529095e-02 -7.12311268e-01 1.13079238e+00
-1.60250604e+00 -9.28358257e-01 -2.82067835e-01 4.94189382e-01
3.68814111e-01 -2.68810153e-01 3.33438635e-01 6.49546981e-01
-4.27519903e-02 8.21281552e-01 2.37076059e-01 -7.03779042e-01
8.82372737e-01 -1.19425559e+00 2.42090628e-01 8.09064329e-01
1.90615058e-01 7.94669569e-01 8.51055205e-01 -7.90206790e-01
-1.28264570e+00 -1.22772181e+00 8.19441557e-01 -1.32845163e-01
2.75600821e-01 8.81189927e-02 -6.44872904e-01 2.80266315e-01
4.13470179e-01 -2.83159733e-01 1.14140320e+00 4.03873593e-01
-2.11384401e-01 -4.81775641e-01 -1.32100189e+00 5.46839416e-01
9.95559633e-01 7.50217512e-02 -1.28590748e-01 2.38405943e-01
4.94964808e-01 -7.22705945e-02 -1.02043319e+00 1.24794565e-01
9.21912909e-01 -9.58416581e-01 8.44232976e-01 -5.98188102e-01
1.65996239e-01 -4.03446347e-01 -2.74405926e-01 -1.77088487e+00
-1.11013305e+00 -6.79527879e-01 -1.93104550e-01 1.28237939e+00
6.09504282e-01 -4.64848340e-01 4.60936338e-01 9.81966615e-01
-3.98803540e-02 -4.99791771e-01 -2.64099985e-01 -6.80821061e-01
-2.47014299e-01 4.25733551e-02 1.05160642e+00 9.07989264e-01
6.32557273e-02 4.83458489e-01 -8.51594269e-01 1.10062204e-01
7.64003277e-01 5.85976899e-01 8.59319210e-01 -1.22119558e+00
-2.55869001e-01 -8.25747669e-01 1.91031158e-01 -1.02585423e+00
-3.15126389e-01 -9.80017900e-01 -1.13732077e-01 -2.00808477e+00
3.00962776e-01 -8.16425383e-01 -8.33113074e-01 1.88775331e-01
-3.11232805e-01 3.49278390e-01 4.38386381e-01 2.73374856e-01
-8.62569630e-01 1.81582242e-01 1.38407099e+00 -2.84927059e-02
-2.75903612e-01 6.83877707e-01 -1.77752662e+00 1.63001820e-01
6.54524982e-01 -6.19226992e-01 -7.06622124e-01 -1.33944169e-01
7.77737617e-01 -1.02271274e-01 -2.52156228e-01 -1.89479187e-01
3.40832621e-01 -6.81757510e-01 -1.65200494e-02 -5.18284976e-01
-7.22266063e-02 -1.22056341e+00 6.66242778e-01 1.84626549e-01
-6.33653343e-01 2.65549421e-01 -4.79355216e-01 6.43091500e-01
1.00187197e-01 -1.65880740e-01 7.02925265e-01 5.09500355e-02
-3.98584455e-01 6.06875479e-01 -1.38789207e-01 -1.39864504e-01
1.00598943e+00 4.69116978e-02 -5.63090026e-01 -9.36070204e-01
-4.31141615e-01 6.31659806e-01 6.55758858e-01 6.11322939e-01
4.52703893e-01 -1.34848726e+00 -1.06088448e+00 -4.11634505e-01
3.38356227e-01 -7.14836836e-01 1.47207230e-01 8.37652445e-01
-9.72378999e-02 6.10423498e-02 -2.23818794e-01 1.45439893e-01
-1.19478488e+00 6.37055099e-01 3.87461148e-02 -3.05987030e-01
-7.82977119e-02 9.63819087e-01 2.16092423e-01 -8.10567290e-02
2.92202711e-01 -2.79039331e-02 -4.62657660e-01 2.92424291e-01
5.52955031e-01 8.72492492e-01 1.13267608e-01 -6.68520153e-01
-2.97169447e-01 -3.36716287e-02 -9.94605199e-02 3.11242729e-01
1.22655296e+00 -4.88833934e-01 -5.49978688e-02 -1.95984811e-01
6.59073651e-01 7.54306853e-01 -9.89242136e-01 -1.89876422e-01
-2.32408002e-01 -1.14102113e+00 2.34648690e-01 -1.21292913e+00
-1.44271410e+00 3.59178513e-01 2.65134931e-01 8.19074392e-01
1.16269290e+00 -2.17927963e-01 6.66843593e-01 -1.98767677e-01
5.53261578e-01 -1.10622370e+00 -1.80294648e-01 -9.78212804e-03
9.39000905e-01 -1.14892185e+00 4.92721438e-01 -4.27618116e-01
-1.02290213e+00 7.37466395e-01 5.42655051e-01 -4.38708887e-02
6.32953048e-01 -1.30389586e-01 -1.68333445e-02 -3.83047611e-02
-7.51361787e-01 -3.22681814e-01 8.52961898e-01 2.58702129e-01
5.86746633e-01 4.16735977e-01 -9.39058602e-01 1.17363119e+00
1.32147595e-02 -3.06649417e-01 5.91189384e-01 5.03822565e-01
-7.09115863e-01 -1.32501328e+00 6.31127432e-02 1.17719817e+00
-6.73288643e-01 -2.03384265e-01 -3.74437124e-01 2.37320334e-01
1.92685649e-01 1.44466090e+00 -2.28130564e-01 -3.16738248e-01
7.97445118e-01 -8.70355248e-01 2.12083831e-01 -7.06104577e-01
-1.22350276e+00 1.84995860e-01 4.77090478e-01 -3.93851578e-01
-2.44253024e-01 -4.48136538e-01 -9.62566674e-01 -8.82999718e-01
-8.82486224e-01 5.47265708e-01 5.02976179e-01 5.07932007e-01
5.16633689e-01 5.90552270e-01 1.10728121e+00 -2.83247590e-01
-8.46652567e-01 -5.99358082e-01 -1.14224815e+00 7.43808150e-01
1.16343349e-01 -4.84559238e-01 -1.75979555e-01 -3.99349719e-01] | [9.706549644470215, 5.652253150939941] |
d5b99ade-cf47-4864-917a-26e4bd0b67fd | time-aware-test-case-execution-scheduling-for | 1902.04627 | null | http://arxiv.org/abs/1902.04627v1 | http://arxiv.org/pdf/1902.04627v1.pdf | Time-aware Test Case Execution Scheduling for Cyber-Physical Systems | Testing cyber-physical systems involves the execution of test cases on
target-machines equipped with the latest release of a software control system.
When testing industrial robots, it is common that the target machines need to
share some common resources, e.g., costly hardware devices, and so there is a
need to schedule test case execution on the target machines, accounting for
these shared resources. With a large number of such tests executed on a regular
basis, this scheduling becomes difficult to manage manually. In fact, with
manual test execution planning and scheduling, some robots may remain
unoccupied for long periods of time and some test cases may not be executed.
This paper introduces TC-Sched, a time-aware method for automated test case
execution scheduling. TC-Sched uses Constraint Programming to schedule tests to
run on multiple machines constrained by the tests' access to shared resources,
such as measurement or networking devices. The CP model is written in SICStus
Prolog and uses the Cumulatives global constraint. Given a set of test cases, a
set of machines, and a set of shared resources, TC-Sched produces an execution
schedule where each test is executed once with minimal time between when a
source code change is committed and the test results are reported to the
developer. Experiments reveal that TC-Sched can schedule 500 test cases over
100 machines in less than 4 minutes for 99.5% of the instances. In addition,
TC-Sched largely outperforms simpler methods based on a greedy algorithm and is
suitable for deployment on industrial robot testing. | ['Morten Mossige', 'Mats Carlsson', 'Hein Meling', 'Helge Spieker', 'Arnaud Gotlieb'] | 2019-02-12 | null | null | null | null | ['industrial-robots'] | ['robots'] | [ 1.99988291e-01 1.74930468e-01 -8.51463452e-02 -2.56033152e-01
-1.50934905e-01 -8.30900848e-01 -5.79349287e-02 -5.19451648e-02
1.71256766e-01 9.46933150e-01 -1.14023018e+00 -8.27246249e-01
-3.80689412e-01 -9.49528694e-01 -7.59137034e-01 -3.22669327e-01
-3.19029301e-01 9.61637616e-01 7.11648464e-01 1.72402546e-01
-7.51886964e-02 4.20160532e-01 -1.68483794e+00 -1.20457955e-01
7.59685457e-01 9.01629269e-01 4.89079297e-01 6.79423332e-01
1.95090249e-01 4.42874223e-01 -1.13120508e+00 2.62838572e-01
3.25377345e-01 -5.30509412e-01 -8.36816907e-01 4.12718952e-01
-4.66099471e-01 -4.22403574e-01 3.25470120e-01 1.18905556e+00
-1.65598556e-01 1.38076514e-01 -1.40256196e-01 -2.09804487e+00
3.23012203e-01 6.68929875e-01 -3.19054693e-01 -4.13977236e-01
5.84106863e-01 3.43862504e-01 3.87731075e-01 -3.25655907e-01
7.34453380e-01 7.62410760e-01 1.22693256e-01 2.74480004e-02
-1.06164122e+00 -4.90134746e-01 2.25774031e-02 -2.79647410e-02
-1.33284569e+00 1.31712466e-01 3.53720576e-01 -3.47517818e-01
1.32667649e+00 4.75437731e-01 6.81657672e-01 4.91493374e-01
1.00846291e+00 9.92717966e-03 9.55997348e-01 -5.11321068e-01
7.67383933e-01 5.71212210e-02 -1.06899634e-01 6.44268632e-01
8.45199287e-01 1.79217547e-01 2.00879723e-01 -1.99943587e-01
6.08632088e-01 9.65148211e-03 4.44464684e-02 -5.33764541e-01
-1.28963923e+00 3.62938970e-01 -5.56835175e-01 3.45152885e-01
-2.08606601e-01 6.13454990e-02 5.60167313e-01 6.42683268e-01
-1.46065488e-01 8.73111308e-01 -1.04196310e+00 -3.89001966e-01
-5.00461340e-01 2.64313430e-01 1.46552515e+00 1.48461533e+00
7.78913319e-01 -4.63053212e-03 3.54279548e-01 3.51380825e-01
1.48741111e-01 5.10646343e-01 2.36346856e-01 -8.82585227e-01
2.28199184e-01 6.80204630e-01 2.90359586e-01 -3.58887672e-01
-3.22877884e-01 4.25922014e-02 -1.42836213e-01 7.17943251e-01
8.07703212e-02 -5.22796750e-01 -7.76497543e-01 1.21180820e+00
5.29816866e-01 2.36584529e-01 7.18280897e-02 8.20825696e-01
-3.33603501e-01 4.61003870e-01 -5.34442604e-01 -4.01174903e-01
9.63956356e-01 -8.32236350e-01 -5.20767629e-01 -1.17919348e-01
8.03682327e-01 -8.96578848e-01 7.00008094e-01 6.85961843e-01
-1.07996154e+00 -3.82135183e-01 -1.55825019e+00 9.97441411e-01
-4.34802830e-01 -2.82531053e-01 6.16590381e-01 5.43583870e-01
-8.56247246e-01 9.22970235e-01 -1.15943217e+00 -2.32263401e-01
-4.10589725e-01 6.83075428e-01 -2.11967692e-01 -5.05464911e-01
-7.00814486e-01 1.05381048e+00 5.07766545e-01 9.23734456e-02
-1.14360213e+00 -1.70287386e-01 -9.89100158e-01 -7.44578838e-02
8.28065217e-01 -1.78166896e-01 1.63597238e+00 -4.39825833e-01
-1.87576854e+00 6.80483058e-02 4.88733739e-01 -5.00514219e-03
3.22226197e-01 2.52778292e-01 -8.28581274e-01 -1.75200135e-01
4.27206308e-01 1.15091026e-01 5.57942271e-01 -7.54847527e-01
-8.43502879e-01 6.48696572e-02 3.60507399e-01 -4.17154998e-01
3.59310985e-01 1.37948962e-02 -3.95774037e-01 3.09086412e-01
7.57997856e-02 -1.45572793e+00 -4.92204219e-01 -8.16865742e-01
-6.60459697e-01 1.05891012e-01 9.55741942e-01 2.66899969e-02
6.49170578e-01 -2.19744444e+00 1.05282180e-01 6.25362039e-01
-3.10220361e-01 -1.87396273e-01 -1.99725494e-01 4.89758998e-01
-1.61179662e-01 3.34108323e-02 -3.14172432e-02 5.73275685e-01
4.50855434e-01 4.54215765e-01 1.58550814e-02 4.69096720e-01
1.76922873e-01 2.27273807e-01 -8.89005661e-01 -3.48046184e-01
5.39610207e-01 -4.20795202e-01 -4.56138670e-01 1.95929244e-01
-7.74360180e-01 2.97767401e-01 -6.76425695e-01 8.71872842e-01
6.42236948e-01 -1.80838436e-01 6.80181682e-01 6.01772547e-01
-3.82153064e-01 3.50404352e-01 -1.47671354e+00 1.21632850e+00
-4.93950278e-01 1.91544354e-01 3.80825922e-02 -5.94528675e-01
7.23294318e-01 4.68968451e-01 6.23551905e-01 -5.10175526e-01
3.13970953e-01 5.37940323e-01 3.02037299e-01 -4.35012460e-01
2.24459261e-01 2.38732040e-01 -5.33412457e-01 8.44434857e-01
-1.71611443e-01 -8.67686152e-01 7.82625437e-01 -4.46828812e-01
1.86752498e+00 2.97128320e-01 3.04128081e-01 -1.91650316e-01
1.22524127e-01 6.66699171e-01 8.59258115e-01 4.95545924e-01
1.98824689e-01 -2.34264493e-01 8.02436054e-01 -2.07332790e-01
-1.07623649e+00 -6.84217572e-01 -6.39696547e-04 6.90163851e-01
3.89071196e-01 -3.92945468e-01 -5.89925051e-01 -7.40792096e-01
5.62916622e-02 1.00854492e+00 -3.79953207e-03 -3.88118476e-01
-2.67833322e-01 2.02344451e-02 5.50900167e-03 1.89120650e-01
2.58062303e-01 -1.12948382e+00 -1.27445829e+00 6.41317546e-01
6.84170663e-01 -9.73174870e-01 -2.29639247e-01 6.64035380e-01
-7.14916229e-01 -1.50858200e+00 2.37482935e-01 -6.29443526e-01
9.14961696e-01 3.77969474e-01 9.11061108e-01 3.40165138e-01
-6.06683433e-01 2.26305842e-01 -5.17709076e-01 -5.33390760e-01
-4.93647814e-01 -2.54763484e-01 -3.85472104e-02 -7.55807698e-01
-1.18620377e-02 -2.60294616e-01 2.11860508e-01 8.41356874e-01
-9.13163543e-01 -1.12008162e-01 5.80016911e-01 8.04361463e-01
7.40690172e-01 1.02329457e+00 3.34176093e-01 -7.09527433e-01
4.47534442e-01 -6.29900992e-01 -1.52726400e+00 2.80553102e-01
-7.14112341e-01 -1.58697635e-01 6.49679482e-01 -7.55775511e-01
-5.11416554e-01 1.57137632e-01 5.71780264e-01 -4.89771187e-01
-1.86090305e-01 1.05338717e+00 -4.05488938e-01 1.27809206e-02
4.93096597e-02 -3.66806030e-01 6.89195916e-02 1.19211458e-01
-2.96819896e-01 1.76532984e-01 6.23465255e-02 -7.25746751e-01
6.94745898e-01 -3.48338962e-01 2.49559894e-01 -2.72718132e-01
-1.70977842e-02 -1.05856478e-01 -1.07378446e-01 -3.45138758e-01
4.94222403e-01 -1.94370493e-01 -8.08514059e-01 6.32241219e-02
-8.18698227e-01 -9.34114158e-01 -3.78224343e-01 6.62503481e-01
-5.89651465e-01 -1.76500201e-01 -2.97133803e-01 -5.44982016e-01
4.25143570e-01 -1.75100529e+00 7.70610929e-01 5.27199470e-02
-4.95101929e-01 -4.80940789e-01 1.73714533e-01 -3.09433192e-01
1.85683399e-01 6.83905840e-01 8.12953889e-01 -6.08190477e-01
-8.72898698e-01 -8.02269816e-01 4.23656851e-01 2.91740656e-01
3.51930708e-01 3.91857982e-01 -2.36659676e-01 -5.54110467e-01
2.20254399e-02 -1.42174736e-01 -5.32125652e-01 6.81373850e-02
9.36700344e-01 -2.70203352e-02 -5.57805300e-01 -1.08222410e-01
1.37418079e+00 8.77219796e-01 5.71994185e-01 4.33663964e-01
7.94507936e-02 4.24055874e-01 1.56615007e+00 5.65100491e-01
-2.51818776e-01 6.02595925e-01 7.63067484e-01 1.61113709e-01
6.51428342e-01 3.30444157e-01 5.55234790e-01 5.59323668e-01
2.95181125e-01 -2.04700097e-01 -8.72923136e-01 3.04640532e-01
-1.75700068e+00 -5.92147350e-01 -4.44344990e-02 2.46044922e+00
5.09573519e-01 6.20338380e-01 -1.99905515e-01 3.36881906e-01
7.92820632e-01 -7.29511142e-01 -7.29829371e-01 -6.38067186e-01
7.32664168e-01 3.20751518e-01 6.40416682e-01 3.14531177e-01
-6.41273320e-01 5.17929792e-01 5.89304638e+00 6.51698485e-02
-1.10569990e+00 -1.66570947e-01 4.86143306e-02 -1.77269861e-01
1.55874163e-01 3.60034525e-01 -4.93692577e-01 5.48385203e-01
1.38453209e+00 -6.01798356e-01 5.32105505e-01 1.55832469e+00
5.87704033e-02 -6.58912718e-01 -1.53312635e+00 4.02430356e-01
-3.78128797e-01 -8.08919787e-01 -9.70934272e-01 3.68262559e-01
8.92174602e-01 9.28103030e-02 -3.87165010e-01 4.45113093e-01
7.57598639e-01 -6.77052617e-01 1.06497562e+00 -9.10553038e-02
9.18418109e-01 -8.88833642e-01 1.20213258e+00 2.21550792e-01
-9.47866619e-01 -7.23912939e-02 -3.74791265e-01 -2.46439412e-01
1.58816010e-01 8.17759812e-01 -1.33533370e+00 6.54347718e-01
6.16117299e-01 -1.74033090e-01 -6.44513369e-02 1.17147207e+00
-3.36366683e-01 2.50349939e-01 -3.44665021e-01 -3.28049242e-01
8.88634007e-03 -1.51357844e-01 2.89917976e-01 4.36088115e-01
7.89631069e-01 -1.43610790e-01 8.09254169e-01 7.11429775e-01
5.94386339e-01 -3.75249863e-01 -8.84544015e-01 -2.26010695e-01
6.29891694e-01 1.30101407e+00 -1.18125308e+00 -3.68258119e-01
-3.78169209e-01 6.17378950e-01 -2.77822703e-01 1.40472263e-01
-1.31025147e+00 -7.60261178e-01 5.40531933e-01 -1.22929320e-01
2.82632530e-01 -5.91725588e-01 -2.95470189e-02 -4.50088859e-01
2.72224247e-01 -9.81912374e-01 -1.00613879e-02 -7.90388525e-01
-7.68615246e-01 6.09133124e-01 2.11233497e-01 -1.54958534e+00
-9.09660995e-01 -6.61056340e-01 -4.93169546e-01 6.19502842e-01
-8.27272534e-01 -3.37712348e-01 -1.21320739e-01 4.57684398e-01
4.44983602e-01 -6.41286746e-02 9.92695808e-01 2.90573351e-02
-7.77704298e-01 -1.03541471e-01 -3.37984473e-01 -6.76438808e-01
5.98734736e-01 -1.12678933e+00 4.34210943e-03 8.42208564e-01
-5.01314223e-01 5.46196878e-01 9.87716794e-01 -1.06566608e+00
-2.07141995e+00 -1.15469635e+00 3.86149824e-01 -1.02768848e-02
8.06692302e-01 -2.76664913e-01 -4.50352341e-01 9.20038283e-01
1.84336275e-01 2.76439544e-02 3.95762682e-01 -1.32353243e-03
2.84523219e-01 -1.39415219e-01 -1.30029535e+00 5.59768975e-01
4.16719764e-01 -8.91223401e-02 -1.22617930e-01 7.90058851e-01
7.83867538e-01 -9.32717323e-01 -1.03987610e+00 2.36727938e-01
5.44757955e-02 -6.70500457e-01 -8.96574780e-02 -1.60000548e-01
1.14172466e-01 -7.83672929e-01 -9.64368880e-02 -1.38474429e+00
-9.02063847e-02 -7.51589715e-01 4.11366194e-01 8.93812001e-01
6.79336727e-01 -1.03661549e+00 3.79479289e-01 1.02161396e+00
-7.30025113e-01 -5.66587090e-01 -9.01644647e-01 -1.41987658e+00
-7.52035201e-01 -5.46087325e-01 1.28287661e+00 9.89510953e-01
7.79426396e-01 8.12954307e-02 3.45118076e-01 4.56785053e-01
8.99881497e-02 3.00990283e-01 7.77671576e-01 -1.11173940e+00
-5.66310585e-01 -6.72078356e-02 -4.46492970e-01 8.89147539e-03
2.81705499e-01 -4.49454904e-01 6.12079918e-01 -1.30992019e+00
-2.67970800e-01 -6.21192455e-01 -7.04478696e-02 8.90581191e-01
6.70847058e-01 -2.72128612e-01 -6.85174093e-02 -2.42249265e-01
-8.64667177e-01 -1.50746182e-01 1.11791837e+00 -1.38412431e-01
-3.60382766e-01 1.42405689e-01 4.56159934e-02 3.67732614e-01
1.01141143e+00 -5.32789826e-01 -6.94670975e-01 -1.16531894e-01
4.38570708e-01 7.07272410e-01 1.37627885e-01 -1.40782952e+00
2.38717899e-01 -7.72190452e-01 1.27806053e-01 -4.61423784e-01
-3.24758291e-02 -1.37223554e+00 1.01258850e+00 7.91396499e-01
2.05240741e-01 5.72935522e-01 3.87039870e-01 1.14158057e-01
-2.79043734e-01 -5.95669031e-01 3.05423826e-01 1.36693299e-01
-6.57190859e-01 4.40455154e-02 -7.04882801e-01 -7.46323705e-01
1.95831800e+00 -2.18815699e-01 -3.71966302e-01 3.82204890e-01
-6.60958171e-01 4.70618606e-01 8.26823711e-01 3.24847907e-01
4.00623679e-01 -9.83139515e-01 4.19022255e-02 3.33280861e-01
-1.75410160e-03 1.06632479e-01 -2.90975254e-02 7.21540034e-01
-4.92528468e-01 4.87014115e-01 -3.85354996e-01 -6.85853779e-01
-9.97764528e-01 8.56987178e-01 -3.97019237e-02 -2.39614055e-01
-5.95336735e-01 2.44809896e-01 -2.20519677e-01 -2.60541677e-01
-3.72569978e-01 -7.67921329e-01 5.21034002e-01 -6.36411190e-01
4.50169742e-01 4.56993610e-01 5.43921709e-01 3.99847239e-01
-5.10263979e-01 8.89804959e-02 1.28781453e-01 -4.41155732e-01
1.15628505e+00 3.50883782e-01 -5.45503318e-01 4.86432344e-01
5.04554808e-01 9.94674955e-03 -1.14928305e+00 6.10617816e-01
1.88465431e-01 -3.93308640e-01 -2.91839987e-01 -8.86370718e-01
-1.01406789e+00 -5.76266609e-02 1.15341559e-01 6.50597394e-01
1.04843616e+00 -1.68182075e-01 2.27743104e-01 5.18822670e-01
1.52116990e+00 -1.32624483e+00 3.35678197e-02 6.17997766e-01
6.99264526e-01 -6.04822874e-01 -2.96003893e-02 -7.13651299e-01
-3.59726638e-01 1.07196248e+00 9.57465470e-01 -1.48415208e-01
3.53879988e-01 1.04293692e+00 -3.93013328e-01 -5.24979644e-02
-1.04549861e+00 3.45909417e-01 -5.96969903e-01 4.46095586e-01
-1.05891459e-01 5.52229762e-01 -1.54893190e-01 3.66972685e-01
-1.75043389e-01 3.27647328e-01 9.14163411e-01 1.60073459e+00
-4.51607287e-01 -1.39551973e+00 -6.37239814e-01 4.46456701e-01
-9.08310562e-02 5.83670437e-01 -8.17356855e-02 1.30911350e+00
4.01398838e-01 1.26280189e+00 5.10999188e-02 -6.16494596e-01
4.58592981e-01 2.11190656e-02 5.67165375e-01 -1.15103149e+00
-2.74986207e-01 2.41475567e-01 6.29985332e-01 -1.00994778e+00
2.55882800e-01 -8.07384610e-01 -1.70059443e+00 -3.38410378e-01
-5.55983722e-01 4.03649002e-01 9.67209578e-01 8.85840595e-01
4.66305941e-01 1.19411433e+00 6.80252075e-01 -7.23518848e-01
-6.11457050e-01 -7.44075596e-01 -9.74046886e-01 -4.15468752e-01
-2.30707586e-01 -1.04475617e+00 -1.29658163e-01 -6.34001847e-03] | [5.01560115814209, 2.215137004852295] |
a8b7ca8a-a415-4ed3-822a-9211cd666985 | synthpop-a-hybrid-framework-for-generating-a | 2304.12284 | null | https://arxiv.org/abs/2304.12284v1 | https://arxiv.org/pdf/2304.12284v1.pdf | Synthpop++: A Hybrid Framework for Generating A Country-scale Synthetic Population | Population censuses are vital to public policy decision-making. They provide insight into human resources, demography, culture, and economic structure at local, regional, and national levels. However, such surveys are very expensive (especially for low and middle-income countries with high populations, such as India), time-consuming, and may also raise privacy concerns, depending upon the kinds of data collected. In light of these issues, we introduce SynthPop++, a novel hybrid framework, which can combine data from multiple real-world surveys (with different, partially overlapping sets of attributes) to produce a real-scale synthetic population of humans. Critically, our population maintains family structures comprising individuals with demographic, socioeconomic, health, and geolocation attributes: this means that our ``fake'' people live in realistic locations, have realistic families, etc. Such data can be used for a variety of purposes: we explore one such use case, Agent-based modelling of infectious disease in India. To gauge the quality of our synthetic population, we use both machine learning and statistical metrics. Our experimental results show that synthetic population can realistically simulate the population for various administrative units of India, producing real-scale, detailed data at the desired level of zoom -- from cities, to districts, to states, eventually combining to form a country-scale synthetic population. | ['Debayan Gupta', 'Kshitij Kapoor', 'Bhavesh Neekhra'] | 2023-04-24 | null | null | null | null | ['culture'] | ['speech'] | [-2.04639629e-01 1.86008308e-02 -1.60840824e-02 -5.42851686e-02
-4.12412405e-01 -3.68142307e-01 8.03115726e-01 5.54446220e-01
-5.93157768e-01 1.29356039e+00 4.49299216e-01 -4.62085247e-01
1.71089143e-01 -1.51844192e+00 -4.60129499e-01 -5.46574116e-01
-2.68416315e-01 8.25540304e-01 -6.66183606e-02 -2.23424658e-01
-2.77648717e-02 4.20642376e-01 -1.21318531e+00 -2.00314537e-01
1.09261286e+00 1.91584751e-02 -1.88244671e-01 5.03606498e-01
1.76646516e-01 1.40936419e-01 -7.78289795e-01 -7.61206031e-01
1.96208850e-01 -4.66820374e-02 -2.95897633e-01 1.29702799e-02
-8.78300294e-02 -6.71893954e-01 -7.44211897e-02 1.00475085e+00
5.37358582e-01 -3.76480371e-01 8.48767459e-01 -1.43962145e+00
-5.80970109e-01 5.07313550e-01 -8.52604270e-01 -1.53442934e-01
5.85454822e-01 2.28076085e-01 3.30319852e-01 -4.67028081e-01
6.41471982e-01 1.48415601e+00 8.02530825e-01 1.67006761e-01
-1.43922579e+00 -8.89870584e-01 -1.12098739e-01 -5.21129310e-01
-1.63671052e+00 -4.61580098e-01 2.95396060e-01 -6.22004628e-01
5.23216605e-01 4.15896863e-01 7.97104657e-01 1.20866501e+00
1.63858458e-01 3.98306996e-01 1.07275760e+00 -5.07314391e-02
3.80419344e-01 3.19555134e-01 -4.14514780e-01 5.32493830e-01
7.69469082e-01 9.65546444e-02 -6.33102953e-02 -9.14760590e-01
8.77647340e-01 3.74062300e-01 -2.28492901e-01 -4.48207647e-01
-1.32990479e+00 1.02136517e+00 2.33345672e-01 -4.28372175e-02
-5.07264912e-01 -9.44155827e-02 7.80613273e-02 2.90552396e-02
7.15771735e-01 -2.02712379e-02 -2.89870709e-01 8.86561051e-02
-7.90272772e-01 8.09784353e-01 7.47001946e-01 8.25158775e-01
6.98350132e-01 -1.41393006e-01 1.63753226e-01 6.05227828e-01
2.22821996e-01 1.18860400e+00 9.21252444e-02 -7.06295490e-01
7.42127180e-01 7.10714698e-01 4.97622281e-01 -1.28231740e+00
-5.67018151e-01 -3.36427130e-02 -1.49805820e+00 -4.44608703e-02
5.62337697e-01 -5.36840916e-01 -7.64631271e-01 1.87524807e+00
4.88546014e-01 1.62621349e-01 1.47531945e-02 5.08367121e-01
4.20338780e-01 6.64686084e-01 2.32797846e-01 -1.96883783e-01
1.33064711e+00 -1.47178009e-01 -4.39574212e-01 -3.11784130e-02
5.97243905e-01 -2.79619306e-01 7.59259820e-01 -2.44051397e-01
-1.02775347e+00 1.03012897e-01 -4.57363725e-01 7.15120196e-01
-8.23224187e-01 -2.26406977e-01 5.40547669e-01 8.89367342e-01
-1.03881884e+00 1.67061463e-01 -8.30735624e-01 -6.88432693e-01
6.12200439e-01 2.68102616e-01 -5.93424618e-01 -6.20600097e-02
-1.20229375e+00 6.11485660e-01 -7.19519779e-02 -2.41704524e-01
-3.19438279e-01 -7.02372432e-01 -9.39992905e-01 9.00382996e-02
-3.57977562e-02 -8.05338800e-01 4.47481722e-01 -3.65187943e-01
-4.74858761e-01 5.68756402e-01 1.44169899e-02 -3.62780243e-01
6.86232507e-01 4.33381677e-01 -5.20192027e-01 -3.84636104e-01
4.55797195e-01 5.96344590e-01 1.84414953e-01 -1.21242154e+00
-7.16283739e-01 -7.52903581e-01 -1.91656545e-01 4.25053872e-02
-1.10377215e-01 2.63905019e-01 -1.66225255e-01 -8.51963401e-01
-2.73536891e-01 -9.13509727e-01 -6.55870855e-01 -2.57152677e-01
-2.69814312e-01 3.63895357e-01 4.71597701e-01 -9.38060582e-01
1.21736848e+00 -1.82831264e+00 -1.55599549e-01 4.37210262e-01
2.23159537e-01 2.37584099e-01 -8.28981847e-02 7.16775775e-01
3.97572458e-01 5.04364371e-01 -7.01164544e-01 -2.77628273e-01
-2.17236862e-01 7.32850283e-02 5.61254658e-02 5.21508515e-01
3.41943502e-01 6.13466620e-01 -1.05754042e+00 -4.44289863e-01
4.31666315e-01 4.85359311e-01 -6.84259474e-01 -2.01481164e-01
8.48042965e-02 4.90027487e-01 -5.34714103e-01 7.52414763e-01
9.45913315e-01 1.39318286e-02 2.40875021e-01 6.02438331e-01
1.40282810e-01 -9.61876437e-02 -1.24755061e+00 7.80259311e-01
-2.63494730e-01 3.70475650e-01 2.35591784e-01 -5.75390041e-01
8.43880653e-01 1.63823634e-01 4.31311727e-01 -5.92361093e-01
-5.22409640e-02 4.81829382e-02 -7.87830204e-02 -1.45211741e-02
4.52145845e-01 1.58109382e-01 -3.80172193e-01 8.32979441e-01
-6.00506425e-01 8.69134441e-02 1.54589444e-01 1.41341135e-01
1.17206907e+00 -3.97103280e-01 5.00449538e-01 -2.82099962e-01
1.34338677e-01 4.89586107e-02 5.91979444e-01 6.93726480e-01
-3.36334884e-01 6.97579741e-01 5.93694687e-01 -5.94214261e-01
-1.34920156e+00 -1.24576640e+00 -2.18758523e-01 5.19137561e-01
-1.84749663e-01 7.24925548e-02 -8.24490309e-01 -5.18192470e-01
2.22629398e-01 4.83691990e-01 -5.27482152e-01 1.48594841e-01
-5.30751884e-01 -1.59829175e+00 6.97813451e-01 5.11138700e-02
6.37107372e-01 -1.11270118e+00 -4.84577835e-01 2.61427552e-01
-3.78303349e-01 -8.78355026e-01 -9.59803239e-02 -5.65204322e-01
-5.72340250e-01 -1.15848505e+00 -1.08791339e+00 -4.69207436e-01
9.36381042e-01 -3.26427519e-02 1.24186611e+00 6.77273348e-02
-1.14027835e-01 -3.99757884e-02 -1.20955229e-01 -2.99491107e-01
-6.52203500e-01 8.03738311e-02 3.26309592e-01 -5.11931069e-02
4.46218461e-01 -6.72896326e-01 -5.19947767e-01 2.92548001e-01
-9.20863748e-01 4.72190604e-02 3.47295463e-01 4.22662228e-01
1.23958169e-02 3.08873147e-01 7.68982589e-01 -8.95420790e-01
7.63939261e-01 -9.67919171e-01 -7.13949800e-01 3.07594687e-01
-1.99712396e-01 -2.13158429e-01 5.68840802e-01 -3.60300183e-01
-7.87907600e-01 -1.19395629e-01 4.04164083e-02 3.94599646e-01
-3.25474828e-01 5.31959116e-01 -3.12186033e-01 3.64646405e-01
5.02590239e-01 1.36955246e-01 1.12794779e-01 -3.06322575e-01
1.09668344e-01 1.13469470e+00 4.11110133e-01 -4.41256255e-01
9.62264776e-01 6.86285138e-01 -1.11817598e-01 -8.93761516e-01
1.27208471e-01 -1.21598147e-01 -5.35798728e-01 2.20308363e-01
3.94520611e-01 -1.11059844e+00 -6.25664353e-01 7.94822454e-01
-8.67884755e-01 -2.15147212e-01 6.10121302e-02 5.81835210e-01
-1.60700545e-01 1.11255735e-01 -2.94023365e-01 -1.00092542e+00
-1.10077992e-01 -1.10499346e+00 9.93373573e-01 3.01994551e-02
-2.35094205e-01 -1.26768684e+00 3.50576311e-01 1.19938783e-01
3.92914325e-01 8.34990323e-01 9.26643908e-01 -3.84363890e-01
-3.92988384e-01 -1.84533119e-01 -5.25821507e-01 -3.01496208e-01
5.11594474e-01 1.12849951e-01 -5.95293939e-01 -5.27423143e-01
-6.22715712e-01 3.26077007e-02 4.06128377e-01 4.85898226e-01
6.46331370e-01 -7.92644680e-01 -6.06073380e-01 3.43630522e-01
1.35767460e+00 3.57587785e-02 7.32292533e-01 1.58173263e-01
3.97324890e-01 8.18593085e-01 5.06864429e-01 8.92503619e-01
8.49518716e-01 5.95728219e-01 1.59085348e-01 -4.90693808e-01
3.20569426e-01 -3.51897269e-01 2.25927353e-01 3.05641502e-01
-1.05606288e-01 -3.89543444e-01 -1.19589615e+00 8.60891163e-01
-1.75265145e+00 -1.12957895e+00 1.11673273e-01 2.65288305e+00
7.24611402e-01 -1.81587994e-01 7.47713268e-01 -2.05576524e-01
9.12148714e-01 1.99808460e-02 -2.18052015e-01 -1.89766377e-01
-2.30402872e-01 -1.25194684e-01 1.01708019e+00 4.78247255e-01
-9.80083644e-01 5.93403518e-01 6.81107044e+00 5.90309560e-01
-8.44678462e-01 -1.94217026e-01 9.11209643e-01 4.98210303e-02
-4.95934129e-01 -1.46448568e-01 -5.26997030e-01 8.14224899e-01
1.02302504e+00 -1.63125351e-01 5.61095715e-01 2.73356408e-01
6.57126129e-01 -1.19387850e-01 -5.65486431e-01 6.10186398e-01
-3.46911728e-01 -1.37838435e+00 1.22810215e-01 6.87756538e-01
7.24528551e-01 7.24682361e-02 7.15123042e-02 2.19892412e-01
1.06054366e+00 -1.04030597e+00 3.16685081e-01 3.45675558e-01
9.25320506e-01 -9.84829903e-01 8.85678530e-01 6.82586074e-01
-1.22313297e+00 9.36422870e-02 -2.75744736e-01 -1.29686445e-01
3.54180425e-01 5.78041434e-01 -8.86911392e-01 3.08445096e-01
7.60852218e-01 1.23095557e-01 -6.56824529e-01 8.61838639e-01
4.49922919e-01 3.55737776e-01 -7.47183621e-01 1.15517750e-02
6.97855353e-02 -4.08812046e-01 2.97887206e-01 9.31134880e-01
6.56501293e-01 1.68441683e-01 3.08746733e-02 7.93674827e-01
-1.59126103e-01 -8.34589452e-02 -1.00234616e+00 5.49251027e-02
9.03800309e-01 7.83949554e-01 -5.69874525e-01 -1.31846458e-01
-3.23495299e-01 5.44993818e-01 2.01849043e-01 4.49397445e-01
-8.61993790e-01 -2.21947521e-01 9.46153522e-01 5.97069442e-01
-1.24301106e-01 -1.35242522e-01 1.39858514e-01 -1.13285613e+00
-9.26081091e-02 -1.06378543e+00 2.16556445e-01 -2.96292871e-01
-1.04810143e+00 1.59609139e-01 2.94253647e-01 -7.88785934e-01
-4.60338056e-01 -1.94126442e-01 -5.76411247e-01 8.83134246e-01
-1.14591742e+00 -1.21064425e+00 -7.21861571e-02 4.51665312e-01
9.59060329e-04 -3.72075081e-01 8.53192508e-01 4.32642788e-01
-6.71426296e-01 5.74706912e-01 5.26900947e-01 3.76594216e-01
2.93401837e-01 -1.03957891e+00 1.06841528e+00 5.85855126e-01
-2.12214589e-01 6.45068645e-01 4.56519693e-01 -9.85360742e-01
-8.64326954e-01 -1.39789808e+00 1.14688885e+00 -5.11101305e-01
4.33718234e-01 -5.37051797e-01 -5.24482131e-01 6.87305152e-01
-3.50848645e-01 -1.73558027e-01 6.20162487e-01 7.61090592e-02
-7.22905947e-03 1.02778867e-01 -1.90492010e+00 9.99197900e-01
8.05112302e-01 -1.36786684e-01 -2.78565530e-02 3.62579644e-01
5.78942597e-01 1.50707573e-01 -9.73760605e-01 2.12358490e-01
6.10448182e-01 -1.13005352e+00 1.19794941e+00 -4.20073658e-01
1.03191748e-01 -9.32460576e-02 -3.66107933e-02 -1.48832750e+00
-2.52478600e-01 -5.08495212e-01 3.94497007e-01 1.46374130e+00
4.82277244e-01 -1.12545097e+00 8.60954702e-01 9.18566585e-01
7.19304085e-01 -4.37467426e-01 -8.38925660e-01 -6.76104844e-01
2.73634553e-01 2.32106224e-02 1.61035991e+00 9.87093151e-01
-2.21983775e-01 -1.15874946e-01 -5.55123866e-01 3.75299096e-01
7.40629971e-01 -3.01896095e-01 1.13419056e+00 -1.28615844e+00
2.04185337e-01 -2.62671083e-01 -3.78250659e-01 -3.12055409e-01
-1.53693423e-01 -1.74527481e-01 -5.42868435e-01 -1.41385424e+00
3.90461117e-01 -1.06145310e+00 3.13441038e-01 3.54747772e-01
-1.62365153e-01 2.66642481e-01 -5.32326214e-02 1.13680989e-01
4.49208021e-02 3.47061664e-01 9.14453924e-01 -1.25560850e-01
-3.68051261e-01 2.23809928e-01 -7.02438354e-01 7.97013283e-01
8.43672097e-01 -4.70522612e-01 -3.26107927e-02 -2.54340172e-01
1.52165428e-01 1.65838882e-01 5.24172604e-01 -8.85985196e-01
-1.83460966e-01 -6.88317835e-01 4.05897558e-01 -6.43563092e-01
2.93985724e-01 -8.95645142e-01 7.09198296e-01 7.84598410e-01
2.35803962e-01 3.23198110e-01 -4.15196940e-02 3.16716939e-01
1.38935775e-01 2.82784879e-01 6.29639864e-01 -4.28901494e-01
1.02550387e-01 5.48758686e-01 -3.61252487e-01 3.79632227e-02
9.65641201e-01 -1.82975128e-01 -5.67057908e-01 -5.38716972e-01
-2.55613148e-01 3.20734113e-01 1.13808954e+00 4.72102575e-02
2.22087234e-01 -1.24254072e+00 -1.24060500e+00 4.54280466e-01
1.86775088e-01 -6.88890666e-02 -1.83239266e-01 3.73522371e-01
-8.89366567e-01 4.25928354e-01 -1.95070371e-01 -2.39350364e-01
-9.42305744e-01 5.28295219e-01 1.67944413e-02 -3.39093983e-01
-3.14401478e-01 7.74927586e-02 2.62468427e-01 -1.18523157e+00
-4.74720523e-02 -1.60446569e-01 -5.93577810e-02 2.68442985e-02
5.68538725e-01 5.77255130e-01 -2.96096772e-01 -1.00604737e+00
-4.08905596e-01 2.42496014e-01 3.43226284e-01 -3.12049150e-01
1.26189947e+00 -3.60599518e-01 -1.98152587e-02 -1.41634494e-01
9.82793391e-01 3.59564066e-01 -8.85243297e-01 -6.38881549e-02
-2.23726913e-01 -6.52186751e-01 -4.29239899e-01 -5.20886540e-01
-9.92300749e-01 5.32723486e-01 3.82068932e-01 3.79149944e-01
8.83859038e-01 -3.75859410e-01 7.50150084e-01 7.74411112e-02
7.20787704e-01 -5.81786573e-01 -7.82346964e-01 -1.05106264e-01
5.39333284e-01 -1.31010997e+00 1.60003528e-01 -2.40418971e-01
-3.99895906e-01 3.08379352e-01 2.62338728e-01 2.41100535e-01
6.34055734e-01 2.08594859e-01 -2.62274873e-02 1.16718434e-01
-3.89197439e-01 -1.74831301e-01 -3.25243175e-01 1.23261416e+00
-1.00557627e-02 6.17087901e-01 -3.54241729e-02 1.99090153e-01
-4.02544826e-01 -1.48538306e-01 6.88479185e-01 7.15087473e-01
-3.06009114e-01 -1.16361177e+00 -8.29764783e-01 6.52623832e-01
-3.05006981e-01 -1.69496179e-01 -4.25239593e-01 1.03800964e+00
1.87252522e-01 8.20147216e-01 5.08854866e-01 3.28848436e-02
5.12359664e-02 -5.38578093e-01 1.48575217e-01 -3.70078981e-01
-5.10199308e-01 -2.95523912e-01 2.09770381e-01 -1.29947037e-01
-1.83890730e-01 -8.79733622e-01 -7.98147559e-01 -1.39134753e+00
-5.20091429e-02 2.86455266e-02 5.12954593e-01 6.16533339e-01
3.22614014e-01 -2.87307143e-01 8.92881989e-01 -7.66087472e-01
-2.43858635e-01 -8.99793506e-01 -7.00362504e-01 4.81021881e-01
3.43826175e-01 -4.96116519e-01 -3.59725118e-01 -2.73300141e-01] | [6.311021327972412, 4.295563220977783] |
e4713e68-bcee-42a8-8fdf-b6bc18902d57 | sentiment-recognition-of-italian-elderly | 2211.07307 | null | https://arxiv.org/abs/2211.07307v1 | https://arxiv.org/pdf/2211.07307v1.pdf | Sentiment recognition of Italian elderly through domain adaptation on cross-corpus speech dataset | The aim of this work is to define a speech emotion recognition (SER) model able to recognize positive, neutral and negative emotions in natural conversations of Italian elderly people. Several datasets for SER are available in the literature. However most of them are in English or Chinese, have been recorded while actors and actresses pronounce short phrases and thus are not related to natural conversation. Moreover only few speeches among all the databases are related to elderly people. Therefore, in this work, a multi-language and multi-age corpus is considered merging a dataset in English, that includes also elderly people, with a dataset in Italian. A general model, trained on young and adult English actors and actresses is proposed, based on XGBoost. Then two strategies of domain adaptation are proposed to adapt the model either to elderly people and to Italian speakers. The results suggest that this approach increases the classification performance, underlining also that new datasets should be collected. | ['Alessandra Grossi', 'Francesca Gasparini'] | 2022-11-14 | null | null | null | null | ['cross-corpus', 'speech-emotion-recognition'] | ['computer-vision', 'speech'] | [-3.03433001e-01 4.54895765e-01 5.19280545e-02 -4.99766886e-01
-1.22669108e-01 -1.71998218e-01 1.01026309e+00 3.30469877e-01
-9.66602683e-01 1.24800074e+00 4.34561193e-01 3.33814502e-01
6.04605302e-02 -5.49742877e-01 9.23257843e-02 -5.00222266e-01
-1.26645565e-01 7.47461498e-01 1.26917893e-02 -4.75153685e-01
-5.78585174e-03 6.69214427e-01 -1.82825422e+00 4.03798878e-01
8.16991329e-01 6.00900471e-01 2.34247237e-01 6.59014463e-01
-2.15554684e-01 8.70882094e-01 -1.21187830e+00 -5.77189207e-01
-2.65629619e-01 -4.34883237e-01 -1.01371491e+00 3.35944265e-01
-2.31190994e-01 5.42639792e-02 2.07824513e-01 6.75454736e-01
9.21823323e-01 3.36375654e-01 8.77537072e-01 -1.19978178e+00
-1.97080582e-01 6.02933943e-01 1.69416159e-01 1.84206828e-01
6.52023792e-01 -2.94279814e-01 6.56341612e-01 -7.49229968e-01
7.43342102e-01 1.33037794e+00 3.65273982e-01 8.61579597e-01
-7.06979990e-01 -2.15984419e-01 2.16914669e-01 3.07936370e-01
-1.22610867e+00 -4.06261384e-01 8.91709805e-01 -3.96100253e-01
9.83594537e-01 2.87903756e-01 9.58921492e-01 1.66052961e+00
-1.44133553e-01 8.87854397e-01 1.30520475e+00 -6.84215307e-01
6.23790264e-01 9.99180436e-01 2.18254328e-01 -1.29057735e-01
-7.67129511e-02 -1.53337210e-01 -6.21097922e-01 -8.37634653e-02
-2.06401572e-01 -5.12940049e-01 -1.12312041e-01 -4.27189516e-03
-7.76345193e-01 7.48407066e-01 -3.16663474e-01 1.19890857e+00
-8.26868653e-01 -5.97918093e-01 9.18063045e-01 4.05064374e-01
8.54391873e-01 1.66628629e-01 -6.29784644e-01 -6.32591188e-01
-5.15812814e-01 3.04742783e-01 1.33347392e+00 7.08349288e-01
5.60957789e-01 -9.78698507e-02 2.72698522e-01 1.13366270e+00
2.46141702e-01 4.34448570e-01 7.75295377e-01 -1.83245137e-01
3.72496843e-01 8.42392623e-01 1.37294754e-01 -8.54486346e-01
-7.95013368e-01 -2.29540035e-01 -5.62260211e-01 1.78682074e-01
1.77387729e-01 -6.32958531e-01 -1.98783323e-01 1.54563880e+00
2.80942678e-01 -4.43870813e-01 9.35938120e-01 5.80697358e-01
9.69573200e-01 6.60910547e-01 4.56594825e-01 -7.42300689e-01
1.70506704e+00 -7.00433612e-01 -1.15028608e+00 -5.31071246e-01
4.64525044e-01 -7.66511261e-01 9.10426974e-01 6.93795562e-01
-9.94601071e-01 -5.33017516e-01 -8.70018065e-01 2.43395746e-01
-9.01837587e-01 5.55003822e-01 3.40245545e-01 1.12115061e+00
-8.98252964e-01 2.46651620e-01 -4.07809943e-01 -9.12786007e-01
-1.69182748e-01 4.59159940e-01 -4.69630271e-01 5.63687325e-01
-1.54588592e+00 1.27248430e+00 7.05390573e-01 8.11613128e-02
-1.85852766e-01 1.85977504e-01 -6.13365591e-01 -1.58077866e-01
-1.49567164e-02 -3.13417852e-01 1.14259493e+00 -1.82637715e+00
-1.92436731e+00 1.20856965e+00 -3.39928903e-02 -5.99547446e-01
7.42650986e-01 -2.24802002e-01 -8.74239922e-01 3.51744473e-01
-5.66881672e-02 5.04309475e-01 4.82392699e-01 -1.22187197e+00
-6.43275142e-01 -5.30922055e-01 -8.66810381e-02 2.56376386e-01
-6.96708858e-01 6.36619925e-01 -1.83713153e-01 -7.12619543e-01
-4.04926866e-01 -8.36722910e-01 1.62355637e-03 -7.02953696e-01
-1.45137921e-01 -5.89787900e-01 6.17800176e-01 -9.66793001e-01
1.33159173e+00 -1.97988462e+00 3.00533384e-01 2.29159117e-01
-3.39450896e-01 4.30071831e-01 3.38474363e-01 5.82267225e-01
-1.49868608e-01 -6.21000491e-02 -1.78739399e-01 -6.66373312e-01
6.50442019e-02 4.01148140e-01 3.32833081e-01 2.56966382e-01
6.31239712e-02 2.11195275e-01 -6.53818846e-01 -9.28379536e-01
1.87992901e-01 4.92204458e-01 -2.32151940e-01 3.91603023e-01
-8.85324627e-02 4.34688181e-01 -6.33004010e-01 2.91995227e-01
5.38644612e-01 6.06584132e-01 4.96671081e-01 1.64142504e-01
-4.38028365e-01 4.44674566e-02 -1.04496932e+00 1.21804178e+00
-6.68034852e-01 3.96440893e-01 8.91155750e-02 -1.11134982e+00
1.53799736e+00 8.75516534e-01 5.82539618e-01 -3.70192617e-01
5.87269783e-01 4.53313410e-01 -1.13724865e-01 -1.05889940e+00
5.84426165e-01 -2.90246010e-01 -3.58785912e-02 1.17926672e-01
4.77665849e-02 8.29824954e-02 4.99432117e-01 -4.68006641e-01
5.74162006e-01 -5.87153323e-02 5.33789992e-01 -2.79467940e-01
1.24360573e+00 -2.80097365e-01 4.30203527e-01 2.42226005e-01
-3.32925439e-01 8.22830275e-02 6.94707274e-01 -4.20303583e-01
-7.13667929e-01 -4.63025093e-01 1.08345605e-01 1.17121506e+00
-4.36437815e-01 -3.24288696e-01 -1.07705104e+00 -6.90648437e-01
-6.86465740e-01 7.58805692e-01 -4.40858036e-01 6.24740236e-02
-4.46976811e-01 -8.63937199e-01 4.79579180e-01 -1.03365700e-03
6.50234938e-01 -1.62397933e+00 -9.01531339e-01 4.53488588e-01
-5.13552964e-01 -1.19040561e+00 4.11129445e-01 1.97380170e-01
-6.51127040e-01 -6.77864671e-01 -1.05745816e+00 -8.22225749e-01
1.54405758e-01 -7.27615297e-01 1.13256073e+00 -1.93258688e-01
8.10945407e-02 4.55488861e-01 -9.82403994e-01 -8.64438891e-01
-9.31649804e-01 4.71341699e-01 7.15452582e-02 2.72002459e-01
8.71947587e-01 -4.68878627e-01 6.01286925e-02 2.60715097e-01
-6.42198503e-01 -4.34651971e-01 5.28363824e-01 5.96300125e-01
-3.15954655e-01 -2.83928901e-01 8.20418417e-01 -7.67382979e-01
7.10516036e-01 -3.83657992e-01 -1.49304653e-02 3.01473528e-01
-2.61151791e-01 -1.56047642e-01 6.29615605e-01 -5.10437489e-01
-1.40462673e+00 3.62738334e-02 -9.41734850e-01 3.48405361e-01
-7.69775093e-01 2.46485636e-01 -7.20436633e-01 3.66307944e-01
7.78123617e-01 -5.92726097e-02 -9.70025808e-02 -5.04242897e-01
-2.32802629e-01 1.25315118e+00 1.65523693e-01 -6.98117137e-01
-2.28511356e-02 7.87340477e-02 -5.65049708e-01 -1.53461075e+00
-1.73595190e-01 -4.88845915e-01 -5.59817195e-01 -6.87853575e-01
1.24006104e+00 -6.91189289e-01 -8.66681576e-01 7.90699363e-01
-1.31915474e+00 -1.02688007e-01 -2.95243323e-01 7.93705702e-01
-6.90789104e-01 3.33793074e-01 -4.79635537e-01 -1.54907191e+00
-4.44848299e-01 -9.14282143e-01 8.27104151e-01 2.31806129e-01
-5.14592648e-01 -9.11684036e-01 6.77603185e-02 2.10786417e-01
9.64867175e-02 2.75805332e-02 5.38942575e-01 -1.14649498e+00
4.47324246e-01 -2.28544161e-01 4.43863928e-01 7.24958360e-01
-1.97961573e-02 -7.96110481e-02 -1.00688231e+00 1.42067432e-01
2.00442970e-01 -4.90164042e-01 3.75465035e-01 -1.26103475e-03
3.59756619e-01 -1.70230091e-01 -1.14465460e-01 -3.39457929e-01
8.48288238e-01 6.53079629e-01 8.12689841e-01 6.47070587e-01
-1.60190836e-01 1.26065385e+00 9.18726623e-01 8.80515397e-01
3.81781310e-01 7.43965030e-01 -5.30719906e-02 -2.23079734e-02
4.13585573e-01 2.94571400e-01 7.78227031e-01 7.12452471e-01
-4.19091016e-01 -5.32919586e-01 -8.28083754e-01 5.32866478e-01
-1.75630784e+00 -9.34090793e-01 -1.22000135e-01 1.94905806e+00
6.18500292e-01 2.53832400e-01 7.06734955e-01 4.77576613e-01
8.16896915e-01 2.63933659e-01 1.35175511e-01 -7.41350770e-01
-5.76034963e-01 1.67797893e-01 -3.72320716e-03 4.64347720e-01
-1.21582961e+00 7.91865468e-01 5.13432789e+00 4.57676113e-01
-1.14147604e+00 1.72213182e-01 7.22372353e-01 2.38221243e-01
8.84897485e-02 -3.88709754e-01 -7.20342398e-01 3.96284997e-01
1.36463141e+00 -9.28135123e-03 2.52799243e-02 1.07137930e+00
4.13293004e-01 -2.24009678e-01 -6.91117823e-01 1.07061327e+00
5.39676189e-01 -4.36325103e-01 -4.71227705e-01 -1.44825369e-01
2.16035038e-01 -2.47795299e-01 -4.40888166e-01 3.73216778e-01
-3.00710320e-01 -5.07608354e-01 8.43483269e-01 8.42224598e-01
-3.61299701e-03 -9.83685255e-01 1.23130620e+00 5.02362192e-01
-8.05943251e-01 -5.85543774e-02 -1.70668721e-01 2.42693909e-02
3.55704159e-01 2.13533238e-01 -6.53921843e-01 6.00612581e-01
8.26379716e-01 5.05969107e-01 -5.78405023e-01 5.48297405e-01
-2.76090235e-01 3.23188990e-01 -2.88748741e-01 -8.82068694e-01
7.88517073e-02 -2.50480473e-01 6.65124476e-01 1.58177984e+00
3.90352875e-01 -5.43188080e-02 3.14400680e-02 3.40506673e-01
4.77163762e-01 1.00904226e+00 -5.81503451e-01 -4.32940871e-02
1.59708977e-01 1.18660474e+00 -7.00775385e-01 -5.02193332e-01
-4.29633647e-01 1.28136206e+00 -3.43630724e-02 2.47262985e-01
-4.83955860e-01 -2.65927494e-01 2.81760395e-01 -6.78083897e-02
-1.31792948e-01 1.47614375e-01 7.30152428e-02 -9.68991458e-01
8.88805315e-02 -1.14563274e+00 4.03465152e-01 -5.83377898e-01
-1.20250058e+00 1.04859066e+00 1.94863111e-01 -1.05605936e+00
-7.25794315e-01 -8.88137102e-01 -1.77916810e-01 6.84506416e-01
-9.59380329e-01 -1.06828463e+00 -1.83049574e-01 4.63749021e-01
6.42460644e-01 -6.44714355e-01 8.90405536e-01 5.54726660e-01
-4.71284062e-01 2.21839070e-01 -2.35346064e-01 -1.17411455e-02
7.01636136e-01 -1.20048630e+00 -1.63086042e-01 3.88736188e-01
-2.31385008e-01 3.12276959e-01 1.09709084e+00 -4.92499620e-01
-5.30936420e-01 -3.67386460e-01 1.74391389e+00 6.92495331e-02
4.83412027e-01 -3.92827809e-01 -7.25225270e-01 3.54236335e-01
6.83848143e-01 -5.16476035e-01 6.29568517e-01 2.39491761e-01
3.78165364e-01 -1.09630644e-01 -1.28595233e+00 5.16042829e-01
5.69282949e-01 -3.09833199e-01 -8.86542916e-01 2.55501240e-01
5.26904352e-02 -2.02729050e-02 -9.75394845e-01 1.27690881e-01
3.62422884e-01 -1.41676795e+00 7.15959489e-01 -3.97217482e-01
6.10648729e-02 5.83137795e-02 1.04129128e-01 -1.26331258e+00
5.85754812e-01 -4.89670306e-01 2.72166282e-01 1.53864706e+00
4.40742999e-01 -8.33426654e-01 6.84262395e-01 5.99987984e-01
-1.13458261e-02 -1.82535380e-01 -1.15570176e+00 -5.48737049e-01
-2.14434378e-02 -4.44674253e-01 3.42387855e-01 6.90442204e-01
2.96827883e-01 6.56126022e-01 -5.46022713e-01 -1.61515683e-01
-1.52890220e-01 -5.44613540e-01 7.47084200e-01 -1.52561700e+00
-2.19660118e-01 -3.40010017e-01 -6.40183210e-01 -4.79238212e-01
7.44360268e-01 -3.06443572e-01 -3.67705226e-02 -1.23605430e+00
-3.47740531e-01 -1.02018304e-01 2.51913190e-01 1.76330328e-01
5.23108430e-02 -5.37035167e-02 9.39545035e-02 -2.97567368e-01
-2.40173295e-01 7.32184827e-01 4.29573327e-01 -7.40557835e-02
-4.61623162e-01 5.07209241e-01 -1.55497223e-01 8.64227474e-01
9.93058681e-01 -2.26443514e-01 -2.12817773e-01 4.33447331e-01
1.69153839e-01 3.49665582e-02 1.15720697e-01 -1.22625887e+00
-4.78635207e-02 1.65454358e-01 1.78718612e-01 -5.15922904e-01
5.91071010e-01 -9.97077107e-01 5.04093878e-02 3.92204165e-01
-2.19208747e-01 2.84402758e-01 5.72360642e-02 1.51319310e-01
-4.75059152e-01 -9.56209183e-01 9.19305027e-01 -3.31382364e-01
-6.68432355e-01 -2.56579190e-01 -1.00115788e+00 -7.46217147e-02
1.26766884e+00 -1.95770100e-01 6.51719570e-02 -5.53696930e-01
-1.28854680e+00 -2.05092654e-01 2.19381735e-01 4.46859092e-01
2.00935379e-01 -9.51927006e-01 -6.92224324e-01 -2.38615870e-01
2.96101809e-01 -6.23582065e-01 5.72555959e-01 8.23306203e-01
-5.19599319e-01 2.42222726e-01 -4.53292519e-01 5.25764599e-02
-1.77155209e+00 6.79990947e-01 4.19483870e-01 -3.19125682e-01
-1.55859992e-01 4.26630914e-01 -5.00939429e-01 -4.53007519e-01
3.66443604e-01 -2.96481363e-02 -1.08501422e+00 8.72504532e-01
3.97878766e-01 4.93842930e-01 1.21996410e-01 -1.24451733e+00
-4.68492508e-01 2.01859728e-01 3.27847928e-01 -6.40691280e-01
1.35007453e+00 -2.76109457e-01 -3.72665435e-01 7.45594323e-01
1.12355411e+00 3.97826552e-01 -4.66350824e-01 1.12226278e-01
5.44952929e-01 2.18531087e-01 -3.79126042e-01 -6.48724318e-01
-6.23777688e-01 8.14932406e-01 8.77471387e-01 5.21647215e-01
1.35203338e+00 3.62313278e-02 2.09888041e-01 5.55982351e-01
2.44533762e-01 -1.90411615e+00 2.54919613e-03 5.32260060e-01
1.08828211e+00 -1.02705765e+00 -3.25081170e-01 -4.16113257e-01
-1.20255399e+00 1.45576751e+00 5.09359598e-01 2.31652096e-01
5.31574965e-01 1.91710114e-01 2.35983655e-01 -2.36152932e-02
-6.14877701e-01 -7.16534853e-01 -6.20830767e-02 7.94667542e-01
7.32035398e-01 -1.28638363e-02 -1.38158131e+00 9.71257925e-01
-2.96373427e-01 2.50394076e-01 3.56927425e-01 8.65220070e-01
-3.13307703e-01 -1.62140214e+00 -6.11564219e-01 -5.65804215e-03
-6.51649296e-01 1.42187998e-01 -8.08174729e-01 1.11365902e+00
6.14027739e-01 1.53387451e+00 -1.16074003e-01 -1.96046785e-01
7.01907396e-01 4.25000191e-01 7.67513216e-02 -2.61469483e-01
-1.13020551e+00 -2.45455410e-02 8.68385434e-01 1.65809691e-01
-1.06010032e+00 -9.75546479e-01 -8.92783284e-01 -4.21532430e-03
-2.68273950e-02 6.44549489e-01 8.68804216e-01 1.03166652e+00
-1.67644918e-01 2.91742563e-01 6.88195288e-01 -7.45362818e-01
-1.20603874e-01 -1.43983626e+00 -7.65409350e-01 3.30273747e-01
-7.37464130e-02 -5.51916897e-01 -3.67445111e-01 6.53154496e-03] | [13.162384986877441, 6.187129020690918] |
1104e36f-692d-4fff-ba0f-50f7f21ef5a9 | category-level-shape-estimation-for-densely | 2302.11983 | null | https://arxiv.org/abs/2302.11983v1 | https://arxiv.org/pdf/2302.11983v1.pdf | Category-level Shape Estimation for Densely Cluttered Objects | Accurately estimating the shape of objects in dense clutters makes important contribution to robotic packing, because the optimal object arrangement requires the robot planner to acquire shape information of all existed objects. However, the objects for packing are usually piled in dense clutters with severe occlusion, and the object shape varies significantly across different instances for the same category. They respectively cause large object segmentation errors and inaccurate shape recovery on unseen instances, which both degrade the performance of shape estimation during deployment. In this paper, we propose a category-level shape estimation method for densely cluttered objects. Our framework partitions each object in the clutter via the multi-view visual information fusion to achieve high segmentation accuracy, and the instance shape is recovered by deforming the category templates with diverse geometric transformations to obtain strengthened generalization ability. Specifically, we first collect the multi-view RGB-D images of the object clutters for point cloud reconstruction. Then we fuse the feature maps representing the visual information of multi-view RGB images and the pixel affinity learned from the clutter point cloud, where the acquired instance segmentation masks of multi-view RGB images are projected to partition the clutter point cloud. Finally, the instance geometry information is obtained from the partially observed instance point cloud and the corresponding category template, and the deformation parameters regarding the template are predicted for shape estimation. Experiments in the simulated environment and real world show that our method achieves high shape estimation accuracy for densely cluttered everyday objects with various shapes. | ['Haibin Yan', 'Jiwen Lu', 'Ziwei Wang', 'Zhenyu Wu'] | 2023-02-23 | null | null | null | null | ['point-cloud-reconstruction'] | ['computer-vision'] | [-1.95062339e-01 -1.80271506e-01 2.50453085e-01 -1.97831556e-01
-3.98229063e-01 -8.96218836e-01 -1.34376124e-01 -1.01104910e-02
8.79566669e-02 2.17452019e-01 -3.70338112e-01 2.49120414e-01
-3.16797942e-01 -7.77319312e-01 -9.22022462e-01 -9.40938115e-01
2.14745224e-01 1.21895075e+00 4.79860127e-01 -1.35482371e-01
9.66882110e-02 8.29805851e-01 -1.60558534e+00 2.51975805e-01
9.67313468e-01 1.26073408e+00 9.57561672e-01 1.94284678e-01
-2.56359398e-01 -2.30082005e-01 -4.61358666e-01 -2.11709097e-01
4.76197809e-01 6.15560710e-01 -2.45117903e-01 9.98329163e-01
-1.96352273e-01 -3.82977843e-01 -9.35819298e-02 1.22731221e+00
7.43131861e-02 4.80333455e-02 7.54130900e-01 -1.08199286e+00
-4.91722494e-01 -7.44598880e-02 -8.43376935e-01 -1.22857228e-01
1.08767837e-01 2.88955331e-01 3.25877249e-01 -1.14453149e+00
5.74702621e-01 1.41968274e+00 1.83073565e-01 7.63181224e-02
-6.41258121e-01 -5.44203639e-01 5.56244612e-01 1.16997316e-01
-1.48271906e+00 3.54803875e-02 9.51283455e-01 -4.96531367e-01
3.45480233e-01 2.49811396e-01 1.02946389e+00 6.63887382e-01
1.21854134e-01 9.27277148e-01 8.83884311e-01 2.18951866e-01
1.49603724e-01 2.20635220e-01 6.23480529e-02 4.99313146e-01
9.49048877e-01 -1.68407008e-01 3.14056516e-01 2.03374922e-02
1.05899119e+00 6.91611826e-01 -3.25929582e-01 -1.14449930e+00
-1.20253849e+00 3.27664495e-01 6.25760376e-01 -2.09149681e-02
-5.49739242e-01 -3.81165832e-01 -1.11841585e-03 -2.59695530e-01
2.81141907e-01 1.96531117e-01 -6.62850797e-01 2.49389231e-01
-1.78313047e-01 3.38220507e-01 6.27669573e-01 1.83298075e+00
8.60787034e-01 -1.27178535e-01 2.64399767e-01 6.82733297e-01
6.05702460e-01 1.14903486e+00 -9.53883529e-02 -7.84014344e-01
8.81228626e-01 9.56436336e-01 4.57960188e-01 -1.35482633e+00
-4.73841608e-01 -5.96884608e-01 -7.43904352e-01 -2.47181848e-01
3.76219183e-01 3.94469053e-01 -1.06106150e+00 8.95251513e-01
7.07394719e-01 -4.53800946e-01 4.25394885e-02 1.36742592e+00
6.41811848e-01 6.20015442e-01 -3.12358648e-01 -4.00425762e-01
1.56267869e+00 -7.06495643e-01 -5.38560331e-01 -4.78386581e-01
-1.21817030e-02 -7.71675885e-01 6.65870249e-01 6.19620383e-01
-1.03077757e+00 -5.12670040e-01 -1.01426792e+00 4.63509381e-01
-2.69481599e-01 4.03647989e-01 7.92227507e-01 2.66172022e-01
-3.32878493e-02 1.92725454e-02 -7.71230578e-01 -1.27151199e-02
5.25908411e-01 4.31966037e-01 -4.10651624e-01 -7.03698456e-01
-4.03415859e-01 5.99084377e-01 7.69337893e-01 6.58262312e-01
-8.83320391e-01 -5.27122736e-01 -6.64137661e-01 -1.96371436e-01
9.68856335e-01 -5.73652208e-01 7.63217390e-01 -5.72208703e-01
-8.47618818e-01 6.28092527e-01 8.59286636e-02 3.53419065e-01
3.49681228e-01 -4.00777385e-02 -9.68136489e-02 3.21391970e-02
4.94248681e-02 1.75653219e-01 8.44663501e-01 -2.10639405e+00
-6.89375520e-01 -1.07239676e+00 -6.01234250e-02 6.40771866e-01
3.45354259e-01 -6.19642854e-01 -8.82709503e-01 -9.70389098e-02
1.10676444e+00 -7.97565818e-01 -2.64481634e-01 -2.57132649e-01
-1.91759035e-01 8.74392390e-02 1.14746559e+00 -7.18001008e-01
4.38509554e-01 -2.34361672e+00 5.26192129e-01 3.47147405e-01
1.45933211e-01 -2.53653556e-01 2.16111034e-01 -9.40340459e-02
3.59150261e-01 -2.44755954e-01 -1.21758267e-01 2.15796426e-01
-1.57914609e-01 7.05747902e-01 -2.19136089e-01 6.52246237e-01
5.10637462e-02 6.55123532e-01 -8.32417488e-01 -6.03787124e-01
3.06106478e-01 2.30246633e-02 -5.57407200e-01 4.69752997e-01
-4.72301513e-01 5.42039037e-01 -1.00736392e+00 1.13951838e+00
1.41815579e+00 -2.51173303e-02 -1.11470390e-02 -6.56756401e-01
1.13929920e-01 -8.27257335e-01 -1.53238451e+00 1.66542995e+00
-2.15722278e-01 -5.50608337e-01 7.06690848e-01 -8.81931365e-01
1.08529842e+00 -1.41118914e-01 5.39106309e-01 -2.39204541e-01
2.23097816e-01 3.31654072e-01 8.76570120e-03 -8.29068661e-01
4.88976508e-01 1.91701010e-01 -2.55172849e-01 -2.70545930e-01
-2.22765163e-01 -8.83736551e-01 -1.11248538e-01 -1.69296235e-01
4.74047363e-01 1.61176786e-01 1.32240728e-02 -4.86996919e-02
3.81329328e-01 3.52621615e-01 6.81308806e-01 1.78501636e-01
1.18136950e-01 6.56789243e-01 -9.36497971e-02 -5.08540928e-01
-1.19915688e+00 -1.40994251e+00 -2.53881872e-01 3.22736859e-01
1.13741374e+00 4.77317274e-01 -6.08141840e-01 -4.76171583e-01
3.72729540e-01 3.47648531e-01 -9.89726186e-02 -1.20305128e-01
-6.81964695e-01 -7.66187906e-01 -4.81326908e-01 5.38662851e-01
5.76642334e-01 -7.93034732e-01 -4.52721030e-01 1.78884417e-01
-3.78875375e-01 -1.22434211e+00 -2.15497032e-01 4.58059572e-02
-1.05800998e+00 -1.32465422e+00 -5.25470972e-01 -8.81581724e-01
1.14550245e+00 6.61043167e-01 7.82577336e-01 -3.33929211e-02
-3.65101457e-01 6.50157154e-01 -4.39008743e-01 -4.33284760e-01
9.43987221e-02 -3.97367924e-01 2.85144895e-01 1.31406814e-01
2.05981597e-01 -3.92925590e-01 -4.89775032e-01 7.01086700e-01
-8.24779272e-01 7.48697147e-02 8.72848749e-01 6.89763725e-01
1.05532491e+00 3.59301001e-01 -1.18958071e-01 -3.98991436e-01
5.46369441e-02 -5.37908256e-01 -8.41000617e-01 1.69132009e-01
2.02795602e-02 -3.71934861e-01 2.50515968e-01 -5.57215273e-01
-1.01425099e+00 1.19581819e-01 4.61354166e-01 -1.15823913e+00
-4.20008183e-01 1.65304288e-01 -9.88388121e-01 1.08321995e-01
1.30322902e-02 3.36185306e-01 -5.85790798e-02 -5.46354711e-01
1.80041596e-01 7.15540290e-01 4.74698633e-01 -8.65704596e-01
8.61074746e-01 5.65705478e-01 -5.07942475e-02 -7.04715848e-01
-6.09391809e-01 -5.85715830e-01 -7.94734836e-01 -5.11316597e-01
8.83903980e-01 -9.26774144e-01 -9.90020335e-01 4.25254703e-01
-1.37287831e+00 2.38909483e-01 -2.17422053e-01 6.08833134e-01
-6.64431214e-01 3.86115640e-01 -3.15848112e-01 -9.58910465e-01
5.06607182e-02 -1.47667706e+00 1.48427522e+00 3.63205224e-01
7.93565273e-01 -3.38908076e-01 -7.74299622e-01 7.56108701e-01
-1.34119228e-01 2.48479202e-01 8.89114857e-01 -3.86774987e-01
-1.33214045e+00 -8.46909955e-02 -2.83786356e-01 -1.04040392e-01
2.82291360e-02 -2.08207697e-01 -5.00679851e-01 -3.46644491e-01
4.28902864e-01 8.90214294e-02 2.62869805e-01 5.74058294e-01
1.43764925e+00 -8.72801468e-02 -7.28466630e-01 7.89018989e-01
1.49800086e+00 4.96511132e-01 2.53223717e-01 8.80993977e-02
1.00724268e+00 6.32818878e-01 1.48557115e+00 5.00332654e-01
2.59645283e-01 7.19546795e-01 1.06184852e+00 1.45472452e-01
3.60753566e-01 -6.50823861e-02 -1.90687314e-01 1.10761297e+00
-3.37945729e-01 -1.68038800e-01 -8.81227255e-01 5.29773057e-01
-1.82829678e+00 -5.84625840e-01 -2.72292107e-01 2.00459743e+00
1.88092783e-01 3.54071483e-02 -2.21373990e-01 -2.75799125e-01
1.01584685e+00 -3.69666040e-01 -9.62361991e-01 2.72152364e-01
-2.80757830e-03 -4.64698076e-01 6.83910131e-01 1.85889497e-01
-9.76316273e-01 6.48217916e-01 4.62260771e+00 9.14900720e-01
-4.28284019e-01 7.50926509e-02 2.99480379e-01 6.94947736e-03
-2.96237528e-01 -2.15692054e-02 -8.66530299e-01 5.20215929e-01
-1.39669389e-01 2.51627594e-01 5.52350938e-01 1.22240913e+00
-1.25522807e-01 -2.56904811e-01 -7.82749951e-01 1.31740391e+00
8.22164342e-02 -7.89144516e-01 7.97761381e-02 4.15403932e-01
5.71815312e-01 -1.50757656e-01 1.32405594e-01 1.43568322e-01
1.76045522e-01 -5.57621181e-01 9.59531307e-01 8.72047603e-01
5.25964260e-01 -7.38366902e-01 8.89655173e-01 8.38759899e-01
-1.53939164e+00 -3.46313030e-01 -9.23463047e-01 2.84613699e-01
1.30859032e-01 7.40243614e-01 -6.63374126e-01 9.34546292e-01
7.77173460e-01 4.20174628e-01 -3.49192798e-01 1.02614224e+00
3.03023756e-01 -2.26212129e-01 -4.37785715e-01 5.12246192e-02
4.93915975e-02 -7.99649596e-01 8.00538659e-01 4.88802403e-01
3.48078310e-01 6.28731370e-01 8.12783003e-01 1.05258548e+00
4.78111088e-01 -1.65590137e-01 -5.93677700e-01 2.65256226e-01
5.41432917e-01 1.54673028e+00 -9.00597930e-01 -2.30314746e-01
-2.70031869e-01 6.14440620e-01 1.78677350e-01 4.05145019e-01
-9.42620218e-01 7.53510445e-02 2.99009919e-01 2.51747847e-01
6.16301358e-01 -4.70692635e-01 -2.83488750e-01 -1.27735806e+00
4.65002209e-01 -5.60881853e-01 1.44405598e-02 -1.17169368e+00
-1.34714305e+00 2.34242961e-01 3.39043856e-01 -1.56399322e+00
2.68281460e-01 -9.10933852e-01 -2.12462321e-02 9.26793039e-01
-9.42307115e-01 -1.25653505e+00 -8.01916063e-01 3.46614063e-01
8.03998649e-01 -1.80029526e-01 3.33134294e-01 -1.49332836e-01
-3.46535146e-01 -7.51273185e-02 1.05059706e-01 -2.26804227e-01
2.08552882e-01 -9.09478366e-01 -3.32607001e-01 4.10305470e-01
-4.39375877e-01 3.62899691e-01 5.70829988e-01 -1.06949854e+00
-2.17311144e+00 -1.27599013e+00 -5.01532078e-01 -3.93721402e-01
1.87871873e-01 -4.72438425e-01 -9.19969380e-01 5.14783740e-01
-5.07651508e-01 4.00715470e-01 -3.87061830e-03 -2.85170346e-01
8.61407444e-02 -1.44823208e-01 -1.50546193e+00 1.61071554e-01
1.34351933e+00 1.67260408e-01 -6.76077068e-01 4.97625291e-01
9.01795924e-01 -1.14449060e+00 -1.14183915e+00 8.18803489e-01
3.07885557e-01 -6.61881208e-01 1.11628306e+00 -4.33400333e-01
1.95695385e-01 -6.52172267e-01 -5.89911103e-01 -1.28535688e+00
-4.73533303e-01 2.29571685e-01 -8.14419612e-02 1.03264844e+00
-1.08256362e-01 -7.08608150e-01 9.12759364e-01 5.84293306e-01
-5.82699060e-01 -8.57817471e-01 -7.15390801e-01 -6.87955201e-01
-3.56381178e-01 -1.39996558e-01 9.27440226e-01 5.83579719e-01
-5.30279458e-01 1.31622292e-02 3.02992105e-01 8.93536329e-01
9.23443675e-01 7.80866504e-01 9.69172835e-01 -1.50274861e+00
-2.45851651e-01 9.56177264e-02 -5.36337793e-01 -1.12970698e+00
-9.53568220e-02 -8.35966408e-01 1.75237820e-01 -1.54106224e+00
4.97737885e-01 -9.25532103e-01 3.35605979e-01 -2.75963265e-02
-1.00850752e-02 -2.34840989e-01 2.42788255e-01 5.71126819e-01
-5.66499710e-01 6.08603716e-01 1.98435509e+00 -3.81951153e-01
-1.14997312e-01 2.21449420e-01 -3.72554213e-01 9.46668684e-01
3.92536998e-01 -1.25368070e-02 -1.88883305e-01 -7.13578463e-01
-6.42447844e-02 4.30436403e-01 2.92828649e-01 -9.70136166e-01
7.66783580e-02 -5.01796067e-01 9.16005373e-01 -1.30485690e+00
9.39878404e-01 -1.62667751e+00 3.23453039e-01 3.70577455e-01
6.10701561e-01 -1.58345178e-01 2.69585311e-01 1.04625535e+00
8.49540681e-02 -1.89585835e-01 4.83335823e-01 -5.59947550e-01
-5.35697281e-01 8.03868830e-01 1.88994974e-01 -1.43280953e-01
1.36845803e+00 -5.22846580e-01 -2.04519525e-01 2.13398322e-01
-1.02203941e+00 4.55433428e-01 8.56204152e-01 4.06889200e-01
9.29665029e-01 -1.35628378e+00 -3.88057172e-01 7.05275834e-01
2.22980410e-01 9.99065399e-01 5.42600989e-01 7.51341701e-01
-5.14156818e-01 3.20917040e-01 -2.25320458e-01 -1.26618099e+00
-9.02674496e-01 1.09899318e+00 2.35645995e-01 1.62982419e-02
-4.41707760e-01 5.58568716e-01 6.43320203e-01 -5.96263528e-01
-9.32918265e-02 -5.83185017e-01 -2.10279360e-01 -2.80481964e-01
1.32242814e-01 2.67225593e-01 1.51512802e-01 -7.68820405e-01
-2.79668689e-01 9.92062390e-01 -8.91202837e-02 2.67343909e-01
1.46274281e+00 -2.27823600e-01 -2.02333897e-01 2.65818805e-01
8.38518798e-01 -8.11222866e-02 -1.40092921e+00 -2.21950933e-01
-5.61408103e-01 -8.47965300e-01 -3.64532053e-01 -4.27468807e-01
-1.19634044e+00 7.74971187e-01 5.31397104e-01 7.17542740e-03
9.47365224e-01 3.88082057e-01 4.97334182e-01 4.64087427e-01
1.21048319e+00 -8.63210857e-01 9.56386849e-02 4.39220399e-01
1.06641817e+00 -9.04017091e-01 2.57136405e-01 -1.00934410e+00
-7.03173995e-01 1.09701002e+00 1.06324518e+00 -4.23500389e-01
4.70969141e-01 2.94742137e-01 -4.78927255e-01 -5.26092291e-01
-1.96574479e-01 -4.44988832e-02 1.10059969e-01 8.99359941e-01
-7.78127134e-01 3.35786968e-01 2.48910993e-01 1.14417946e+00
-2.15448048e-02 -6.50485039e-01 3.70288014e-01 8.69135916e-01
-8.57929170e-01 -5.56693375e-01 -1.17110729e+00 7.51511514e-01
1.40036121e-01 5.94099998e-01 1.05444100e-02 7.86027551e-01
4.30056959e-01 6.52773321e-01 4.00107384e-01 -3.25574756e-01
7.72592783e-01 -3.05179775e-01 8.79892886e-01 -7.90821850e-01
-6.54279143e-02 3.68743271e-01 -3.35802794e-01 -5.10367155e-01
-1.55732021e-01 -7.77618349e-01 -1.37975574e+00 1.32680416e-01
-6.59051716e-01 1.34351194e-01 8.17784548e-01 9.07355547e-01
1.12478040e-01 5.38224995e-01 7.62932301e-01 -1.41372967e+00
-3.93570691e-01 -7.57725894e-01 -9.86725152e-01 4.72906798e-01
-2.11429864e-01 -1.34420848e+00 -2.81970114e-01 -1.28250182e-01] | [7.447391986846924, -2.6599292755126953] |
434d8350-c2aa-4775-a03e-af1b06ec5d53 | local-water-filling-algorithm-for-shadow | null | null | https://pubmed.ncbi.nlm.nih.gov/33291572/ | https://pdfs.semanticscholar.org/cd2a/cba3283f304b232ac7a26838e747d7185d5d.pdf?_gl=1*uybhf6*_ga*MTIwOTc3MTQ0OS4xNjczNjY5MDI4*_ga_H7P4ZT52H5*MTY4MzgyMjM0NS4yNS4xLjE2ODM4MjIzNTYuMC4wLjA. | Local Water-Filling Algorithm for Shadow Detection and Removal of Document Images | Shadow detection and removal is an important task for digitized document applications. It is hard for many methods to distinguish shadow from printed text due to the high darkness similarity. In this paper, we propose a local water-filling method to remove shadows by mapping a document image into a structure of topographic surface. Firstly, we design a local water-filling approach including a flooding and effusing process to estimate the shading map, which can be used to detect umbra and penumbra. Then, the umbra is enhanced using Retinex Theory. For penumbra, we propose a binarized water-filling strategy to correct illumination distortions. Moreover, we build up a dataset called optical shadow removal (OSR dataset), which includes hundreds of shadow images. Experiments performed on OSR dataset show that our method achieves an average ErrorRatio of 0.685 with a computation time of 0.265 s to process an image size of 960×544 pixels on a desktop. The proposed method can remove the shading artifacts and outperform some state-of-the-art methods, especially for the removal of shadow boundaries. | ['C L Philip Chen', 'Bingshu Wang'] | 2020-12-04 | null | null | null | sensors-2020-12 | ['shadow-removal', 'shadow-detection-and-removal', 'shadow-detection'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 7.51955092e-01 -4.00813460e-01 6.45602703e-01 -2.66712099e-01
-2.83988148e-01 -4.97804046e-01 3.47434968e-01 -1.42651439e-01
-2.16292784e-01 5.84156215e-01 9.33330208e-02 -4.57558006e-01
2.69185722e-01 -7.81876564e-01 -3.30095589e-01 -8.24758112e-01
4.69705135e-01 9.85426083e-02 8.95572841e-01 -1.50652155e-01
6.64943516e-01 5.88345468e-01 -1.37927508e+00 7.16202706e-02
1.46790540e+00 8.13569069e-01 1.02900183e+00 6.88262403e-01
-4.46032315e-01 4.31233972e-01 -8.54831100e-01 -6.06187657e-02
9.08158571e-02 -3.48525375e-01 -2.70640165e-01 4.42585647e-01
2.03351036e-01 -7.26683855e-01 -1.67892411e-01 9.50862408e-01
6.15609407e-01 8.51194337e-02 5.19094169e-01 -7.81132162e-01
-3.49705487e-01 -2.14687437e-01 -1.11771107e+00 -5.28278612e-02
3.76278251e-01 -2.46193662e-01 2.89585680e-01 -1.14684451e+00
4.70966429e-01 1.17113781e+00 6.82106137e-01 1.26671493e-01
-7.04009354e-01 -5.74290633e-01 -2.68355086e-02 1.54170007e-01
-1.35929763e+00 -3.46855432e-01 5.98024487e-01 -2.43889466e-02
2.33757928e-01 7.61674583e-01 5.05499661e-01 2.29035437e-01
4.27562058e-01 8.56089294e-01 1.74510098e+00 -5.48075736e-01
1.42096087e-01 5.38212545e-02 2.28664249e-01 8.39403212e-01
3.91374618e-01 -4.89297003e-01 -7.03393459e-01 -2.93576092e-01
7.68720150e-01 2.40833685e-01 -8.32364380e-01 -8.09262618e-02
-7.23001659e-01 3.05813849e-01 1.00480013e-01 -1.13187887e-01
-9.17012170e-02 -2.44908720e-01 -2.56273542e-02 -3.38319331e-01
3.37850869e-01 -1.26599401e-01 9.77358501e-03 1.87978163e-01
-1.03814578e+00 4.23625186e-02 7.40122437e-01 7.94619024e-01
7.40254581e-01 -1.20602868e-01 -1.19465478e-01 9.85520482e-01
3.96351933e-01 1.02695227e+00 3.11494857e-01 -7.17431843e-01
3.11159611e-01 4.75158125e-01 4.46772784e-01 -1.11521816e+00
-1.72414958e-01 5.95891811e-02 -7.80138671e-01 5.09939849e-01
1.29572347e-01 9.00855362e-02 -1.16112375e+00 6.07558668e-01
4.72975284e-01 2.25369409e-01 1.61635041e-01 1.04149163e+00
8.51006985e-01 9.48436558e-01 -6.71415448e-01 -4.63379025e-01
1.42237318e+00 -9.93928134e-01 -1.06353235e+00 -4.14197445e-01
4.46493179e-02 -1.25598788e+00 1.28033853e+00 7.99158275e-01
-8.27604711e-01 -8.06897953e-02 -1.22721720e+00 -2.25277588e-01
-5.88349700e-02 3.59960318e-01 3.80589455e-01 7.42107034e-01
-6.78449512e-01 4.34611499e-01 -5.59810460e-01 -2.61683553e-01
1.39846444e-01 -4.43788879e-02 8.69701207e-02 -4.31541681e-01
-7.75110364e-01 7.00024843e-01 -1.04798056e-01 4.71587926e-01
-7.19457805e-01 -3.51053804e-01 -4.91407841e-01 3.77883576e-02
5.64109683e-01 1.64799154e-01 8.29919100e-01 -6.35249317e-01
-1.59456134e+00 5.51486611e-01 -6.18677974e-01 1.58183798e-01
6.65386677e-01 -4.65466619e-01 -2.62234330e-01 1.55246675e-01
-1.25495076e-01 -3.59331489e-01 9.51305449e-01 -1.83000851e+00
-3.91100138e-01 -4.57756341e-01 -4.30658430e-01 6.08341157e-01
-3.32271248e-01 3.70382331e-02 -9.37928319e-01 -5.74406683e-01
5.67718506e-01 -6.44871950e-01 -1.04299765e-02 1.89668655e-01
-5.62099874e-01 4.49221373e-01 1.24665856e+00 -1.08497858e+00
1.22449517e+00 -2.02267551e+00 -4.08680230e-01 4.19392645e-01
4.95921448e-02 4.37492698e-01 1.14144146e-01 3.61516923e-01
5.28790176e-01 -2.43315309e-01 -7.36122370e-01 -4.27569628e-01
-3.00312877e-01 1.41318470e-01 -6.54403210e-01 5.79702735e-01
-5.58441877e-01 3.70242685e-01 -5.35677791e-01 -6.63561761e-01
2.65017778e-01 6.56661749e-01 2.72601455e-01 2.43734121e-01
1.83698088e-01 -1.95791423e-02 -3.49793226e-01 8.39512229e-01
1.56486392e+00 2.04837397e-01 2.80968517e-01 3.72317545e-02
-5.49773276e-01 -9.15654898e-02 -1.43832552e+00 1.24459028e+00
-3.71060193e-01 7.96455264e-01 4.39622968e-01 4.39055040e-02
1.32385695e+00 -2.17104495e-01 -4.78768349e-02 -7.13768005e-01
4.82670702e-02 3.98881048e-01 -5.78801155e-01 -2.37741098e-01
8.17230046e-01 1.25276327e-01 4.31544960e-01 5.05293667e-01
-9.38593149e-01 -4.36691880e-01 -1.55680612e-01 2.42938727e-01
1.19335401e+00 1.55781865e-01 -3.41732427e-02 -4.28146362e-01
6.81180477e-01 -1.21602312e-01 7.65270710e-01 6.58057213e-01
8.25373977e-02 8.90525460e-01 1.30282596e-01 -2.06211463e-01
-6.97305202e-01 -9.57264841e-01 -2.27017567e-01 8.24289143e-01
9.74046111e-01 -3.29809636e-01 -9.05826926e-01 -2.14052841e-01
2.73135193e-02 6.82799101e-01 -2.66943604e-01 2.20872939e-01
-6.43397570e-01 -9.00674105e-01 1.07746914e-01 2.08758920e-01
1.01832914e+00 -1.02142656e+00 -9.04638410e-01 -1.01013184e-01
-1.67982861e-01 -9.16526258e-01 -4.96576875e-01 -8.58919322e-02
-8.16817164e-01 -1.00043845e+00 -1.01777339e+00 -7.50524223e-01
8.83810222e-01 1.15011835e+00 8.33734393e-01 5.45488358e-01
-5.07015646e-01 7.67118037e-02 -5.18901885e-01 -5.24585247e-01
1.11706093e-01 -5.37627220e-01 -3.26720715e-01 1.85248405e-01
4.95256558e-02 -1.86427310e-01 -8.06672215e-01 6.96669221e-01
-1.02604902e+00 4.18312192e-01 6.20945990e-01 5.05438924e-01
6.81521118e-01 2.18764618e-01 -2.71371037e-01 -1.02717650e+00
7.38536835e-01 5.88895008e-02 -8.57877374e-01 4.91843790e-01
-6.77542984e-01 -3.36307198e-01 4.61387604e-01 -1.15067884e-01
-1.75383210e+00 3.94122042e-02 5.05907297e-01 1.11304626e-01
2.09573939e-01 6.92870526e-04 -2.62359381e-01 -3.68854433e-01
3.20582211e-01 6.67399347e-01 -2.62042373e-01 -5.82131028e-01
-1.04005134e-03 1.05230367e+00 7.03492284e-01 -2.92125970e-01
1.01397491e+00 9.51004028e-01 -6.27193451e-02 -1.31894612e+00
-6.29160583e-01 -4.80610758e-01 -5.08133054e-01 -3.65678757e-01
6.72674000e-01 -5.57265103e-01 -6.68311596e-01 1.03900564e+00
-1.09004116e+00 -6.28069043e-01 3.81833136e-01 -3.50150722e-03
-1.43425567e-02 8.52892280e-01 -3.90473336e-01 -1.10425544e+00
-5.78136921e-01 -8.34426701e-01 1.18583381e+00 6.28541589e-01
4.22914416e-01 -5.58314919e-01 -1.00747291e-02 4.20305580e-01
3.04791301e-01 7.16038421e-02 4.77185488e-01 4.06003833e-01
-8.41091096e-01 1.28833607e-01 -7.14126706e-01 1.20428786e-01
3.08117002e-01 -1.76395737e-02 -1.12998247e+00 -2.64920980e-01
5.92964813e-02 8.90541002e-02 1.04957151e+00 2.18013883e-01
1.30254400e+00 8.39456543e-02 -4.31861371e-01 6.73209727e-01
1.56941330e+00 3.91577750e-01 1.16344059e+00 4.85458672e-01
7.65140176e-01 3.86695623e-01 1.15488398e+00 6.58702314e-01
2.58983612e-01 3.70030463e-01 3.68825585e-01 -5.07314503e-01
-2.59444565e-01 5.63137829e-02 1.41833276e-01 7.35172212e-01
-5.81745878e-02 -4.77485538e-01 -9.58139181e-01 2.40015015e-01
-1.70411730e+00 -4.12618995e-01 -7.32188284e-01 2.31113720e+00
8.38937163e-01 -1.26245692e-02 -7.59361148e-01 1.01326175e-01
9.29847181e-01 4.22975183e-01 -3.94038081e-01 -3.57533306e-01
-2.81151623e-01 3.18106301e-02 7.73617387e-01 9.29594219e-01
-6.79754019e-01 1.15121245e+00 5.47139263e+00 7.46129096e-01
-9.12478685e-01 -8.59751850e-02 3.45141083e-01 4.66204137e-01
-4.22143668e-01 1.16786338e-01 -7.33120143e-01 6.16478682e-01
2.22727150e-01 1.72362328e-01 5.61819971e-01 3.89094174e-01
3.72300088e-01 -1.02791035e+00 -1.81106672e-01 9.76048768e-01
4.98114347e-01 -7.50906110e-01 -2.98191458e-01 -1.35705173e-01
7.41551340e-01 -4.41656619e-01 -1.71204790e-01 -2.91794598e-01
1.31962463e-01 -8.54258001e-01 4.38071191e-01 7.40461111e-01
7.89906144e-01 -5.58156252e-01 6.80210054e-01 3.29562277e-01
-1.22742450e+00 2.07408652e-01 -7.10864782e-01 4.96975370e-02
1.72989011e-01 1.07550144e+00 -7.47191310e-01 3.40015322e-01
8.74282718e-01 1.03540666e-01 -5.01163840e-01 1.21609187e+00
-4.03937638e-01 3.51289779e-01 -3.66064638e-01 -1.72394440e-01
-1.30507991e-01 -6.94368064e-01 2.44791567e-01 1.18256521e+00
3.64382893e-01 5.23368180e-01 2.43174993e-02 4.31479156e-01
-5.44097312e-02 1.92999169e-01 -1.71672493e-01 4.06176716e-01
6.17291749e-01 1.41931939e+00 -1.06877136e+00 -2.69561440e-01
-1.68008909e-01 1.68897796e+00 -3.20677698e-01 4.24677193e-01
-7.63050556e-01 -1.01379371e+00 -2.90876534e-03 5.78324161e-02
6.80071712e-02 -3.52077216e-01 -5.93792319e-01 -9.71088767e-01
1.95452467e-01 -6.70976698e-01 -3.48528266e-01 -1.38229561e+00
-5.79820871e-01 3.21983278e-01 -5.40943563e-01 -9.74342108e-01
8.08043003e-01 -6.03049159e-01 -9.46247518e-01 1.10572827e+00
-1.74035668e+00 -9.22474444e-01 -1.33612156e+00 4.26864833e-01
6.68319643e-01 2.16915071e-01 6.79107904e-01 -2.89978236e-02
-4.08165425e-01 9.25316103e-03 6.96752071e-01 -2.75620550e-01
9.49618697e-01 -1.20358455e+00 3.45195502e-01 1.09753430e+00
-3.23794574e-01 4.14628804e-01 7.74517000e-01 -1.01288509e+00
-1.45560861e+00 -7.29199886e-01 6.29621685e-01 -4.78954241e-02
8.67028758e-02 -5.89433074e-01 -1.17330277e+00 2.03156963e-01
1.89117447e-01 -1.28538758e-01 3.40559632e-01 -3.38844061e-01
-6.15400495e-03 -2.96787530e-01 -1.21072197e+00 7.09013164e-01
7.96168566e-01 -3.40929061e-01 -3.92741412e-01 4.66838568e-01
2.66776055e-01 -8.05063367e-01 -1.95795789e-01 2.37264216e-01
6.85641468e-01 -1.29526782e+00 7.07903862e-01 6.45406365e-01
1.97690949e-01 -6.76649928e-01 -1.07829079e-01 -1.07262075e+00
-1.87683087e-02 -7.49250591e-01 9.33383405e-02 1.41929781e+00
-3.37630548e-02 -7.44563341e-01 8.09586227e-01 4.05438125e-01
-1.93172589e-01 -6.96255326e-01 -3.51110429e-01 -3.91789049e-01
-6.00637615e-01 1.44016460e-01 4.62028772e-01 5.39923131e-01
-6.26815975e-01 -1.01938479e-01 -3.50235254e-01 4.46501940e-01
8.64716887e-01 5.69490075e-01 8.98481011e-01 -1.17865086e+00
-5.16894087e-02 2.22164541e-01 2.04624444e-01 -1.18152058e+00
-2.63500303e-01 -1.61080047e-01 5.75635433e-01 -1.98331308e+00
5.29639542e-01 -5.18415868e-01 -7.23981718e-03 3.33856314e-01
-3.63712668e-01 4.30567086e-01 1.65874422e-01 4.05404329e-01
-3.89216036e-01 5.74577332e-01 1.34233236e+00 6.85164407e-02
-3.94082636e-01 -7.14112595e-02 -4.00413722e-01 9.99463856e-01
7.70177901e-01 -1.35407642e-01 -3.93206388e-01 -6.93213165e-01
-9.58412215e-02 -1.22039780e-01 3.08094501e-01 -8.41572165e-01
3.62309813e-02 -3.12362880e-01 6.93267286e-01 -1.15709388e+00
5.54896712e-01 -6.96490109e-01 -2.28974327e-01 4.51135010e-01
2.36757025e-01 -2.99265176e-01 3.45435068e-02 6.81537271e-01
1.32273689e-01 -5.31979322e-01 7.81490684e-01 1.13604339e-02
-7.21822262e-01 -1.43196106e-01 -3.76477838e-01 -3.05981547e-01
9.66727495e-01 -2.69315809e-01 -7.61044979e-01 -3.78521413e-01
6.62907735e-02 3.43628466e-01 7.83228397e-01 -8.43279511e-02
1.12884891e+00 -9.46056247e-01 -4.22502398e-01 1.65652156e-01
-2.76862711e-01 7.13941008e-02 2.64184505e-01 5.79841197e-01
-1.26884854e+00 4.35158387e-02 2.58901492e-02 -3.88601929e-01
-1.80091834e+00 -1.35321483e-01 3.64090651e-02 1.28142118e-01
-8.37366402e-01 7.34418929e-01 3.46827596e-01 -2.20864147e-01
1.20526724e-01 -1.88860029e-01 4.95873094e-02 -2.73985475e-01
7.08136857e-01 8.00277591e-01 1.32277310e-01 -4.09467340e-01
-4.11636710e-01 9.05988276e-01 2.72030123e-02 -2.96095669e-01
1.15053248e+00 -2.87558883e-01 -4.59500492e-01 1.58581957e-01
7.31697857e-01 5.75640917e-01 -1.47302639e+00 -1.03961915e-01
-3.40793163e-01 -1.13796580e+00 2.16244102e-01 -8.57902646e-01
-7.47003913e-01 1.01244330e+00 8.50221753e-01 1.16289407e-01
1.28630173e+00 -5.62880874e-01 1.09314811e+00 3.88443857e-01
3.10212880e-01 -1.41951513e+00 3.71904648e-03 4.96359974e-01
1.04816031e+00 -1.01611984e+00 6.02371216e-01 -7.80431807e-01
-7.01487482e-01 1.19515693e+00 4.20938402e-01 -5.88218011e-02
3.10885340e-01 5.26742220e-01 1.94800571e-01 7.31045231e-02
-2.80422028e-02 -1.20200798e-01 -8.46783351e-03 4.47965562e-01
5.17531522e-02 -5.25631802e-03 -5.92711151e-01 3.96765798e-01
-1.19084120e-02 -2.15781033e-01 1.00101435e+00 1.23764515e+00
-1.14647639e+00 -8.05894196e-01 -9.91216719e-01 3.12423170e-01
-7.32466057e-02 -1.94956258e-01 -5.74469984e-01 5.25483310e-01
-3.59921679e-02 9.26806450e-01 -1.71051305e-02 -3.72636393e-02
1.33967832e-01 -2.46525541e-01 2.98370242e-01 -4.27977234e-01
-3.54020223e-02 3.81964117e-01 -1.41209900e-01 -4.97405082e-01
-1.25013500e-01 -5.01257181e-01 -1.60080695e+00 -2.50350356e-01
-5.71870387e-01 -5.21426350e-02 8.54606748e-01 6.90747619e-01
-3.73394564e-02 4.81187254e-01 6.51939929e-01 -8.16417634e-01
-5.39899766e-02 -7.06439614e-01 -1.03149498e+00 1.52875185e-01
2.21692309e-01 -5.57189345e-01 -3.89418215e-01 7.75534585e-02] | [10.832277297973633, -3.984349250793457] |
a515e93c-169e-4752-aa2a-47967770aff8 | variable-star-classification-using-multi-view | 1911.05821 | null | https://arxiv.org/abs/1911.05821v1 | https://arxiv.org/pdf/1911.05821v1.pdf | Variable Star Classification Using Multi-View Metric Learning | Our multi-view metric learning framework enables robust characterization of star categories by directly learning to discriminate in a multi-faceted feature space, thus, eliminating the need to combine feature representations prior to fitting the machine learning model. We also demonstrate how to extend standard multi-view learning, which employs multiple vectorized views, to the matrix-variate case which allows very novel variable star signature representations. The performance of our proposed methods is evaluated on the UCR Starlight and LINEAR datasets. Both the vector and matrix-variate versions of our multi-view learning framework perform favorably --- demonstrating the ability to discriminate variable star categories. | ['S. M. Caballero-Nieves', 'R. Haber', 'K. B. Johnston', 'A. M. Peter', 'V. Petit'] | 2019-11-13 | null | null | null | null | ['multi-view-learning'] | ['computer-vision'] | [-1.20719165e-01 -5.45366824e-01 -4.02462512e-01 -4.07368660e-01
-8.01730692e-01 -1.20596528e+00 8.25283527e-01 -4.72955197e-01
1.07437238e-01 3.72257501e-01 2.80692309e-01 -9.19034705e-02
-5.79450309e-01 -6.25930429e-01 -2.24539250e-01 -8.29503477e-01
5.30523509e-02 6.66336298e-01 -2.80624628e-01 -2.17647925e-01
1.69461057e-01 6.24021411e-01 -1.68790185e+00 6.04101457e-03
4.06081527e-01 1.06948328e+00 -3.49972308e-01 6.37361825e-01
1.37973577e-01 4.89259452e-01 -2.54729241e-01 -4.71075445e-01
7.06202090e-01 -1.03499465e-01 -2.74331540e-01 5.04594147e-01
1.20096278e+00 2.40810826e-01 -1.65399671e-01 8.53073239e-01
2.22161740e-01 1.03273600e-01 9.78075683e-01 -1.30383372e+00
-9.85126853e-01 1.24945544e-01 -6.89576864e-01 2.30256841e-01
4.19437051e-01 -1.74812183e-01 1.80335963e+00 -1.37607014e+00
6.54396474e-01 1.35439205e+00 4.50587451e-01 1.35301143e-01
-1.40334570e+00 -3.68517339e-01 -2.43853033e-02 5.31746805e-01
-1.28098512e+00 -2.13416696e-01 9.54661012e-01 -8.60787332e-01
6.52234077e-01 3.98416758e-01 6.51671231e-01 9.91647124e-01
3.29014421e-01 4.67410058e-01 1.68374729e+00 -9.75997094e-03
3.55983600e-02 2.93119065e-02 2.97884613e-01 1.16839278e+00
4.14417297e-01 2.83517331e-01 -6.32287979e-01 -4.65309829e-01
5.72786748e-01 6.95421323e-02 -4.28941939e-03 -1.38381100e+00
-1.36078787e+00 1.10271263e+00 1.21282982e-02 -5.55637069e-02
-1.81324095e-01 -3.64389271e-02 3.23876590e-01 4.51244295e-01
4.33853149e-01 5.22220314e-01 -4.84321237e-01 -1.83864962e-02
-7.85832644e-01 1.46694900e-02 5.81681252e-01 1.01220286e+00
7.64347911e-01 3.20985585e-01 -1.14225455e-01 7.93398440e-01
1.80379003e-01 8.79541397e-01 3.73690993e-01 -8.65607142e-01
1.96772590e-01 7.00403214e-01 -6.81537166e-02 -8.68358135e-01
-5.03156960e-01 -6.91722393e-01 -4.47845995e-01 4.54723567e-01
4.26113635e-01 1.54382586e-01 -8.09691131e-01 1.52095425e+00
2.88489372e-01 -5.54999113e-02 1.43318966e-01 8.38685572e-01
8.77716184e-01 -9.68137830e-02 -6.37300789e-01 -2.13545814e-01
1.40763164e+00 -9.69676971e-01 -4.73715484e-01 7.58396387e-02
4.20471281e-01 -6.11940444e-01 1.03614450e+00 6.97946787e-01
-8.83191884e-01 -4.07163918e-01 -1.32188976e+00 1.23104416e-01
-4.97225165e-01 -4.78114858e-02 9.14496124e-01 7.16754913e-01
-9.25226688e-01 2.13340998e-01 -4.49401230e-01 -1.17661253e-01
2.97288954e-01 3.06349903e-01 -8.45465481e-01 -2.32823297e-01
-6.32773221e-01 9.19025600e-01 -8.09874684e-02 -3.49532217e-01
-1.00527298e+00 -5.61329246e-01 -9.84723508e-01 -6.59664646e-02
3.55127960e-01 -9.35026646e-01 7.90480196e-01 -6.19460225e-01
-1.19871390e+00 8.59356701e-01 -3.25666040e-01 1.34251684e-01
2.44083494e-01 7.71371722e-02 -6.30465686e-01 3.83873396e-02
3.54467630e-02 -9.14224684e-02 1.20764685e+00 -1.42990947e+00
-2.91741133e-01 -7.51867712e-01 2.61317521e-01 4.05949205e-01
-3.10066849e-01 -1.41584411e-01 -2.11500615e-01 -4.80428666e-01
4.76361215e-01 -1.23834193e+00 4.13714051e-02 4.22619050e-03
-3.09369043e-02 -3.74927968e-01 8.78557384e-01 -2.81304449e-01
8.78675878e-01 -1.97709405e+00 9.89517331e-01 3.49978596e-01
5.74794590e-01 -4.56299573e-01 -8.67892951e-02 2.53987521e-01
-4.50102210e-01 -4.02643472e-01 -2.07915343e-02 -8.90480056e-02
-1.64963081e-01 3.94798100e-01 2.71004904e-02 9.61044908e-01
-2.32252598e-01 7.25772500e-01 -7.55289316e-01 -4.68849659e-01
3.49552184e-01 -4.26068553e-04 -3.95049512e-01 1.98031649e-01
1.22341156e-01 -9.14354529e-03 -2.10323304e-01 9.81962919e-01
5.29797614e-01 -4.81933653e-01 3.75046991e-02 -2.34510317e-01
6.56643957e-02 -2.49001712e-01 -1.22339892e+00 1.67691731e+00
-7.26347625e-01 4.16770846e-01 -2.02558958e-03 -7.84970701e-01
9.81652856e-01 8.08659717e-02 6.98041797e-01 -1.54936090e-01
1.53582515e-02 -5.77176805e-04 4.65129539e-02 -1.84515312e-01
2.52057016e-01 -4.25245315e-01 -3.18446487e-01 3.95974517e-01
4.82592404e-01 -2.55966723e-01 4.72557358e-02 1.54302374e-01
7.12638259e-01 1.39042616e-01 5.74811876e-01 -3.88577759e-01
8.40267837e-01 -2.96595573e-01 4.86735612e-01 2.42831305e-01
-3.39255184e-02 6.44097388e-01 4.92293715e-01 -3.70182037e-01
-7.92948723e-01 -1.63189685e+00 -3.99673909e-01 1.16071069e+00
9.11068544e-03 -6.15919769e-01 -1.69468652e-02 -1.09813285e+00
3.40903699e-01 5.56270599e-01 -7.93849289e-01 -1.20089911e-02
-1.19852923e-01 -6.17766201e-01 4.87109385e-02 4.90263909e-01
-8.85691047e-02 5.68981469e-02 -2.86679804e-01 -3.57783914e-01
1.10941812e-01 -9.25181627e-01 -3.64603341e-01 2.18351200e-01
-7.41520345e-01 -1.34748590e+00 -2.54621953e-01 -4.33351696e-01
3.25127870e-01 7.50800073e-01 1.11617923e+00 -4.49078888e-01
-2.81420022e-01 1.02756023e+00 -2.64750361e-01 -5.92372678e-02
2.86420677e-02 -4.43847686e-01 5.42242050e-01 2.82054722e-01
5.58173478e-01 -6.47264898e-01 -2.77065545e-01 5.17218292e-01
-5.38055301e-01 -2.44782805e-01 3.30781996e-01 1.03754187e+00
6.25773430e-01 -6.09211028e-01 2.64953196e-01 -7.45670915e-01
2.92003840e-01 -5.57086170e-01 -7.39884496e-01 4.69985396e-01
-1.15566599e+00 3.15727323e-01 6.66093886e-01 -2.01014712e-01
-5.14295638e-01 1.07664578e-01 5.50607145e-01 -9.14012372e-01
2.58939564e-01 2.62502998e-01 -3.48820239e-01 -4.46418703e-01
6.31647229e-01 1.55362204e-01 1.42781034e-01 -4.22729373e-01
9.69650745e-01 4.47269231e-01 3.85666519e-01 -4.54057008e-01
1.44811821e+00 7.33407676e-01 6.04175568e-01 -7.26214826e-01
-7.92414367e-01 -8.24921370e-01 -1.03043997e+00 -3.01750511e-01
8.56942654e-01 -1.29618657e+00 -5.24523795e-01 6.74590692e-02
-3.68719459e-01 6.10292017e-01 -1.52205482e-01 5.28203487e-01
-1.02311015e+00 5.42980373e-01 -5.95568633e-03 -2.85269350e-01
3.47454115e-05 -1.09688509e+00 9.46463645e-01 -9.16346349e-03
-7.65891969e-02 -1.28908324e+00 4.45901543e-01 6.77043319e-01
1.41869143e-01 3.55579644e-01 9.30598259e-01 -7.83620656e-01
-3.09241831e-01 -4.34589237e-02 1.82005037e-02 1.67927712e-01
4.33847398e-01 -1.12878159e-01 -1.00043571e+00 -5.66319346e-01
-9.20589268e-03 -4.65278953e-01 7.98837066e-01 8.19005668e-02
1.02313375e+00 2.07139805e-01 -8.98137316e-03 1.05767405e+00
1.56310976e+00 -2.78281748e-01 -7.13888779e-02 2.66659111e-01
1.03126371e+00 4.65056092e-01 6.79284573e-01 6.39287531e-01
4.09753501e-01 9.66208994e-01 5.91564596e-01 -3.27800252e-02
-4.50556017e-02 -7.83786178e-02 2.63528377e-01 9.10892129e-01
-1.00962207e-01 2.22908139e-01 -5.44818461e-01 4.82949674e-01
-1.67317939e+00 -1.19498503e+00 -2.83226877e-01 2.04269981e+00
1.38589516e-01 -2.03699946e-01 3.69515568e-01 1.18583161e-02
2.00704679e-01 7.30789185e-01 -7.04773188e-01 -6.23682082e-01
-3.75907958e-01 2.88235486e-01 6.47463262e-01 5.16304910e-01
-1.27438402e+00 8.34706962e-01 7.14745808e+00 4.32714969e-01
-8.82520914e-01 1.32984743e-01 -8.31969380e-02 -5.06024659e-01
-7.16692746e-01 4.59602810e-02 -5.31608224e-01 -1.97336525e-01
8.02624524e-01 -5.19839108e-01 7.06997812e-01 1.19461286e+00
7.19089713e-03 3.35803688e-01 -1.31848824e+00 1.49846840e+00
8.65799785e-01 -1.12629461e+00 1.71087280e-01 2.16878578e-01
7.87194729e-01 4.84634638e-02 3.62941593e-01 2.68442929e-01
4.53533769e-01 -1.08352375e+00 5.93981206e-01 4.29445714e-01
1.05611026e+00 -7.33890116e-01 2.14888752e-01 -1.64875552e-01
-1.30562901e+00 -3.90435964e-01 -3.72006863e-01 -2.53043463e-03
-1.89694136e-01 3.45639646e-01 -7.66245604e-01 8.21368217e-01
4.55215275e-01 9.89395738e-01 -1.15611446e+00 6.82164192e-01
8.71902034e-02 1.74430206e-01 8.49209577e-02 3.30195010e-01
-2.97707994e-03 -7.70296156e-01 7.02807128e-01 7.90441632e-01
2.65543640e-01 -2.12464362e-01 2.77327031e-01 5.99393070e-01
3.26529652e-01 1.50391057e-01 -9.60396826e-01 1.44025996e-01
7.58341476e-02 1.71659935e+00 -2.90160835e-01 -5.68729304e-02
-1.18163693e+00 9.08495367e-01 4.75001514e-01 1.88280612e-01
-5.70423782e-01 2.04924047e-02 1.01314914e+00 -1.68158948e-01
6.65268779e-01 -6.77011132e-01 -5.67017436e-01 -1.67938912e+00
-2.18069330e-01 -1.10042822e+00 6.73396528e-01 -6.37335956e-01
-1.52062404e+00 3.40860844e-01 2.19601303e-01 -1.56913590e+00
-5.26195884e-01 -1.02077496e+00 -2.16094464e-01 8.97047460e-01
-1.16455734e+00 -1.54132819e+00 -2.61101723e-01 1.02543414e+00
4.24008608e-01 -8.60286772e-01 9.17232156e-01 -1.22189268e-01
-2.23517582e-01 8.19439352e-01 4.85165983e-01 -2.62305200e-01
9.20766294e-01 -1.69478929e+00 -2.59154402e-02 7.76293457e-01
7.35283792e-01 4.16318059e-01 7.33535051e-01 -1.80061683e-01
-2.08945966e+00 -7.55979121e-01 3.34699571e-01 -1.06345749e+00
9.02373791e-01 -5.35639763e-01 -3.54046524e-01 7.35399306e-01
3.61439213e-02 3.05881351e-01 1.36290288e+00 6.91024005e-01
-1.13849711e+00 -2.34179080e-01 -1.06378818e+00 3.33212316e-01
9.91906762e-01 -7.65390635e-01 -8.63921285e-01 2.81567335e-01
3.33642662e-01 -2.44112145e-02 -1.25238085e+00 4.06402081e-01
7.39024520e-01 -1.05750692e+00 1.14750063e+00 -8.98558974e-01
2.84927428e-01 -4.37649220e-01 -8.83214116e-01 -1.64248908e+00
-9.83123302e-01 -8.88189152e-02 -3.46084386e-01 9.33052778e-01
2.60331064e-01 -4.75740165e-01 7.24756241e-01 1.14378646e-01
-5.64594455e-02 -7.81273067e-01 -1.17878318e+00 -8.27248693e-01
1.68918371e-01 -3.25836726e-02 2.62460768e-01 1.09198940e+00
-6.04405254e-02 4.98512894e-01 -4.65135574e-01 1.56203270e-01
9.91917014e-01 7.78921247e-01 8.32283556e-01 -1.49618554e+00
-9.34866250e-01 -2.64979362e-01 -8.07361186e-01 -5.81948221e-01
4.51123416e-01 -1.27771294e+00 -5.41453779e-01 -1.17481756e+00
5.94654441e-01 -9.35180932e-02 -5.45922756e-01 2.57047694e-02
-1.52301833e-01 2.60389805e-01 2.03338414e-01 1.16226815e-01
-3.81824642e-01 5.12621462e-01 1.18052518e+00 -1.62096575e-01
2.10347325e-01 1.36364941e-02 -7.33736098e-01 5.41944921e-01
3.42798680e-01 4.91793230e-02 -6.28561020e-01 -4.53314418e-03
-1.15809729e-03 1.21345669e-01 2.73256481e-01 -7.87730932e-01
-2.75098652e-01 -5.21961033e-01 6.34489179e-01 -6.11404836e-01
6.84631646e-01 -7.84227610e-01 5.74263670e-02 5.78482822e-02
6.46982864e-02 6.37390852e-01 -1.87921375e-01 8.29927742e-01
-9.47219692e-03 1.78153664e-01 8.63031864e-01 -2.22980708e-01
-7.92633414e-01 4.32002813e-01 -1.18742427e-02 1.39809787e-01
1.16771734e+00 -2.87869543e-01 -3.27602178e-01 -3.26820731e-01
-7.74513900e-01 1.63820148e-01 7.96842873e-01 7.99353659e-01
6.50467575e-01 -1.59693742e+00 -6.59654438e-01 5.27799547e-01
8.21851075e-01 -6.56292140e-01 -8.33901241e-02 6.53516352e-01
-1.08812831e-01 3.47200543e-01 -4.19236958e-01 -8.19712698e-01
-1.72334111e+00 1.08134890e+00 3.50448400e-01 -4.42200005e-02
-4.95555043e-01 6.29713416e-01 3.07583421e-01 -4.56362844e-01
-1.33882016e-01 3.49588543e-01 -4.85479265e-01 4.89061147e-01
1.76028803e-01 4.85870421e-01 9.59272683e-02 -1.08400285e+00
-3.91428232e-01 9.77008581e-01 2.99616661e-02 -3.29281747e-01
1.29423440e+00 -6.04600310e-02 -8.71450230e-02 9.25447047e-01
1.35125685e+00 2.78970331e-01 -8.87244105e-01 -3.88896883e-01
9.80838761e-02 -8.20914865e-01 8.56545120e-02 -5.58515906e-01
-7.30443299e-01 7.07578540e-01 7.47440994e-01 -2.97968000e-01
8.68843436e-01 2.17005223e-01 6.43629208e-02 3.32608819e-01
5.06298244e-01 -8.64238203e-01 2.47343376e-01 4.98704642e-01
8.90948772e-01 -1.16007912e+00 4.56517458e-01 -2.74622917e-01
-6.81750953e-01 1.29013193e+00 6.31926298e-01 -2.78757006e-01
7.66516626e-01 -5.94174378e-02 2.58240581e-01 -5.53640902e-01
-9.42841709e-01 -2.33016729e-01 6.50569856e-01 6.33287787e-01
3.28748852e-01 3.50317568e-01 -2.81037182e-01 1.96784273e-01
-4.56845433e-01 -8.38784695e-01 7.95615017e-01 5.32592058e-01
-3.07382315e-01 -1.23426032e+00 -6.23506665e-01 7.35648096e-01
-1.56425551e-01 6.35605305e-02 -6.86174691e-01 7.08978176e-01
-2.39316300e-02 9.59456086e-01 -1.19021378e-01 -7.09972262e-01
3.36258173e-01 4.21764016e-01 6.77704930e-01 -6.48401499e-01
-4.44140673e-01 3.50878276e-02 1.90909788e-01 -6.60128236e-01
-3.69434983e-01 -1.18598223e+00 -5.17645359e-01 -1.32597983e-01
-2.98055142e-01 -7.27385432e-02 5.75638115e-01 7.72101223e-01
-2.16971394e-02 1.63729578e-01 1.26788795e+00 -5.62589467e-01
-1.17296064e+00 -7.06989527e-01 -8.31476450e-01 7.12520599e-01
3.55596572e-01 -1.33148050e+00 -6.63437188e-01 -3.64763141e-01] | [8.374481201171875, 4.535530090332031] |
9e00de3b-3e6e-4bfe-a87f-461f1f94e35d | shef-multimodal-grounding-machine-translation | null | null | https://aclanthology.org/W16-2363 | https://aclanthology.org/W16-2363.pdf | SHEF-Multimodal: Grounding Machine Translation on Images | null | ['Lucia Specia', 'Kashif Shah', 'Josiah Wang'] | 2016-08-01 | null | null | null | ws-2016-8 | ['video-description', 'multimodal-machine-translation'] | ['computer-vision', '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] | [-7.363471508026123, 3.7340803146362305] |
e75e2f22-61ce-4051-a532-a33710203b65 | transact-transformer-based-realtime-user | 2306.00248 | null | https://arxiv.org/abs/2306.00248v1 | https://arxiv.org/pdf/2306.00248v1.pdf | TransAct: Transformer-based Realtime User Action Model for Recommendation at Pinterest | Sequential models that encode user activity for next action prediction have become a popular design choice for building web-scale personalized recommendation systems. Traditional methods of sequential recommendation either utilize end-to-end learning on realtime user actions, or learn user representations separately in an offline batch-generated manner. This paper (1) presents Pinterest's ranking architecture for Homefeed, our personalized recommendation product and the largest engagement surface; (2) proposes TransAct, a sequential model that extracts users' short-term preferences from their realtime activities; (3) describes our hybrid approach to ranking, which combines end-to-end sequential modeling via TransAct with batch-generated user embeddings. The hybrid approach allows us to combine the advantages of responsiveness from learning directly on realtime user activity with the cost-effectiveness of batch user representations learned over a longer time period. We describe the results of ablation studies, the challenges we faced during productionization, and the outcome of an online A/B experiment, which validates the effectiveness of our hybrid ranking model. We further demonstrate the effectiveness of TransAct on other surfaces such as contextual recommendations and search. Our model has been deployed to production in Homefeed, Related Pins, Notifications, and Search at Pinterest. | ['Andrew Zhai', 'Zhiyuan Zhang', 'Nazanin Farahpour', 'Saurabh Vishwas Joshi', 'Neng Gu', 'Po-Wei Wang', 'Dhruvil Deven Badani', 'Nikil Pancha', 'Pong Eksombatchai', 'Xue Xia'] | 2023-05-31 | null | null | null | null | ['sequential-recommendation'] | ['miscellaneous'] | [ 5.55220433e-02 -1.31099641e-01 -8.43812346e-01 -5.20812988e-01
-1.00283659e+00 -6.49979115e-01 7.62596965e-01 7.54859000e-02
-1.12690859e-01 8.63015354e-02 1.05967653e+00 -2.62965858e-01
-4.63967979e-01 -7.16598809e-01 -5.37879229e-01 1.27298474e-01
-3.60623360e-01 3.11686039e-01 1.36278003e-01 -4.45064664e-01
2.44385675e-01 2.23659292e-01 -1.70796549e+00 9.77457821e-01
3.33044767e-01 1.23361075e+00 5.06904572e-02 9.59924817e-01
-1.37916252e-01 9.71486747e-01 -1.81896314e-01 -9.34893116e-02
3.01844478e-01 -1.70413759e-02 -7.27455378e-01 -3.47786158e-01
6.89112723e-01 -9.78576064e-01 -6.07475996e-01 -6.83775842e-02
7.45161951e-01 6.08063400e-01 2.67524332e-01 -1.12775385e+00
-8.76399696e-01 7.73733199e-01 1.34206591e-02 9.04340446e-02
8.03185225e-01 -2.91095339e-02 1.52689838e+00 -7.48776376e-01
4.90583688e-01 8.61605585e-01 7.18791127e-01 5.32931268e-01
-1.17367530e+00 -2.90375233e-01 6.35722399e-01 -9.68090352e-03
-9.17544067e-01 -6.72502279e-01 4.01397765e-01 -4.30634916e-01
1.36217439e+00 6.15607381e-01 6.33623064e-01 1.59221601e+00
-1.71148196e-01 1.26987171e+00 3.35666925e-01 2.00119917e-03
1.38116539e-01 1.52712673e-01 4.96166319e-01 4.46935326e-01
-2.03800663e-01 2.74542391e-01 -7.42432952e-01 -5.42726099e-01
7.10143805e-01 6.98175251e-01 -1.00594282e-01 -2.78994024e-01
-8.80658329e-01 5.02490699e-01 2.10939348e-01 1.97030008e-01
-5.37424028e-01 5.00676632e-01 2.70990968e-01 3.60343963e-01
5.62968194e-01 6.22960091e-01 -7.61834264e-01 -9.00315106e-01
-7.96425223e-01 2.14897707e-01 1.14880097e+00 1.15168643e+00
2.55421907e-01 -1.71289876e-01 -6.32045329e-01 1.10064161e+00
2.81039238e-01 2.83572346e-01 6.64794803e-01 -8.54222178e-01
3.45939487e-01 3.59669656e-01 6.87284648e-01 -8.61799121e-01
-2.53180087e-01 -3.33021849e-01 1.50305182e-01 -4.53536630e-01
-3.02861761e-02 -2.83977360e-01 -4.65268254e-01 1.49574137e+00
-8.71641748e-03 2.79558450e-01 -3.04953307e-01 7.79396236e-01
6.42143667e-01 6.16709411e-01 1.44957483e-01 1.47123843e-01
8.30171943e-01 -1.38483346e+00 -4.70529258e-01 -2.10797805e-02
7.78017342e-01 -5.95272064e-01 1.45414507e+00 5.16224265e-01
-9.37214553e-01 -7.32309401e-01 -9.40977037e-01 -2.14133918e-01
-3.74384850e-01 2.52432376e-01 1.03723264e+00 7.94827461e-01
-1.03739810e+00 1.03762197e+00 -9.84601021e-01 -6.44624710e-01
2.40682527e-01 6.93627894e-01 -7.53947124e-02 2.26710007e-01
-8.50099087e-01 4.43809539e-01 -4.91334766e-01 -3.40828210e-01
-5.53458273e-01 -9.15064633e-01 -2.37218946e-01 4.09368843e-01
3.65065068e-01 -3.77850831e-01 1.79455698e+00 -7.96192050e-01
-1.94374621e+00 2.01597288e-02 -1.37088120e-01 -4.28312533e-02
2.74919063e-01 -1.09933293e+00 -8.40161443e-01 -4.91384268e-01
-2.74675548e-01 1.01947464e-01 3.84568959e-01 -6.64568722e-01
-8.09468925e-01 -1.48607746e-01 3.65384221e-01 1.38543919e-01
-9.79628384e-01 -1.22317262e-01 -6.48402333e-01 -4.49234664e-01
-4.62742269e-01 -8.46465230e-01 -2.59946287e-01 -3.20304483e-01
2.41620928e-01 -4.03372347e-01 5.58179379e-01 -6.80529833e-01
1.78511238e+00 -2.22417879e+00 -2.52503157e-01 2.96201676e-01
-2.76771896e-02 7.44956136e-02 -6.97019696e-01 8.95672262e-01
6.66626543e-02 2.57320136e-01 9.42369580e-01 -3.41596127e-01
3.69116306e-01 -1.66619867e-01 -4.65026885e-01 1.44284060e-02
-3.60103101e-01 1.08885229e+00 -1.18770528e+00 2.46997654e-01
2.51748800e-01 4.66498584e-01 -1.06835473e+00 4.46877986e-01
-3.95825267e-01 -1.84227794e-01 -6.31903291e-01 4.93665516e-01
4.52312231e-02 -2.95998156e-01 5.29385149e-01 -2.56428421e-01
-1.58031717e-01 8.47321808e-01 -1.00589848e+00 1.71075904e+00
-1.07027853e+00 1.71035498e-01 -2.71088451e-01 -1.49915114e-01
5.75455725e-01 2.11135104e-01 1.04957521e+00 -1.10602283e+00
-2.26842821e-01 -7.44978860e-02 -6.52713656e-01 -5.57937443e-01
8.74568105e-01 8.06299865e-01 -2.01920137e-01 1.03749299e+00
-1.65645983e-02 6.10274136e-01 -1.02401935e-01 2.79703796e-01
1.71378338e+00 5.51112711e-01 -7.80572519e-02 1.46371350e-01
-1.71096772e-01 -3.73693168e-01 5.09186424e-02 9.20314729e-01
1.51042819e-01 4.38402206e-01 2.45538026e-01 -5.17566860e-01
-5.21319449e-01 -9.22412038e-01 5.31862915e-01 2.24033165e+00
-1.08653061e-01 -1.20995414e+00 -6.67073652e-02 -1.11608136e+00
1.73858821e-01 8.50961387e-01 -5.40470421e-01 -1.45514235e-01
-5.23101985e-01 -2.40267113e-01 1.61873952e-01 8.11100185e-01
-1.64586511e-02 -8.26041520e-01 -2.27685720e-01 5.35310447e-01
2.85498388e-02 -6.98248804e-01 -1.04305482e+00 -1.56492442e-02
-9.49744582e-01 -9.25932229e-01 -2.56041884e-01 -2.91488945e-01
2.32955724e-01 6.47455335e-01 1.08435380e+00 -6.13085516e-02
1.10436209e-01 1.10536683e+00 -6.75243497e-01 2.70332694e-01
2.86076248e-01 3.45900208e-01 2.13079140e-01 1.58637725e-02
3.62915277e-01 -8.39187324e-01 -1.03876591e+00 6.35984004e-01
-6.04612052e-01 -2.76365548e-01 4.34596449e-01 5.00976026e-01
3.56065929e-01 -3.56815487e-01 4.61907357e-01 -1.27858925e+00
9.55263019e-01 -8.51940274e-01 -1.19119652e-01 2.31292933e-01
-8.24150383e-01 -4.66196574e-02 5.70883155e-01 -7.89389610e-01
-1.00208473e+00 -1.15229420e-01 -5.02897143e-01 -6.92419484e-02
2.35017031e-01 3.68040770e-01 -4.61251400e-02 3.34880769e-01
1.03933704e+00 -1.62603185e-01 -4.13449913e-01 -9.89071071e-01
8.26195717e-01 1.14224410e+00 5.49277738e-02 -6.52724683e-01
9.01788026e-02 1.45002350e-01 -8.39243472e-01 -5.99677324e-01
-5.92857480e-01 -7.66282499e-01 -1.63456827e-01 -1.85710743e-01
1.99351460e-01 -8.02196980e-01 -8.51055086e-01 -7.71154612e-02
-6.64172769e-01 -9.52151418e-01 -3.97412479e-01 3.44169885e-01
-5.33635855e-01 4.81230877e-02 -7.68370509e-01 -7.37425566e-01
-5.27067423e-01 -6.04759991e-01 1.01503754e+00 9.32045355e-02
-8.60222042e-01 -7.65448570e-01 2.77866900e-01 5.45293093e-01
8.79655421e-01 -3.01332355e-01 6.82780027e-01 -1.06117868e+00
-4.09815252e-01 -6.49640620e-01 -1.44431582e-02 2.67861277e-01
3.50335419e-01 -1.62976965e-01 -9.19984818e-01 -2.91522771e-01
-7.76681364e-01 -1.95122123e-01 3.98481220e-01 -8.10846165e-02
1.43031621e+00 -5.80157757e-01 -4.79541928e-01 3.11865211e-01
1.16799426e+00 1.85082942e-01 5.23328125e-01 5.12996092e-02
6.21231437e-01 2.07761660e-01 8.04807842e-01 6.77459002e-01
2.49770105e-01 1.08201444e+00 6.99622184e-02 2.19648927e-01
-1.44389868e-01 -8.31761241e-01 7.93535709e-01 6.42890453e-01
-1.97342604e-01 -5.30902028e-01 -1.43406883e-01 5.10449529e-01
-2.24483776e+00 -1.01579952e+00 3.50685686e-01 2.43899608e+00
4.15391415e-01 6.50926074e-03 5.29129922e-01 -3.86892378e-01
5.03845029e-02 2.03666538e-01 -4.84645456e-01 -6.09967709e-01
6.86005592e-01 4.94680911e-01 4.99800295e-01 5.26161075e-01
-8.99656773e-01 8.60195577e-01 6.54299974e+00 5.93101978e-01
-1.04211199e+00 4.39509660e-01 2.88522840e-01 -7.76586413e-01
-5.69080830e-01 -2.65169859e-01 -7.85800576e-01 5.97658098e-01
1.28453457e+00 1.63920850e-01 7.59015083e-01 1.36343634e+00
4.24081236e-01 4.35495496e-01 -1.72986841e+00 8.01726222e-01
5.98250479e-02 -1.31997263e+00 -1.23023786e-01 3.09904844e-01
7.82954097e-01 2.55725741e-01 1.94885045e-01 9.56502020e-01
6.01642907e-01 -7.26346910e-01 3.76932710e-01 6.81196809e-01
8.54086757e-01 -4.78866875e-01 3.94420564e-01 1.45033702e-01
-1.04225588e+00 -5.37707984e-01 1.40289724e-01 -1.83460817e-01
1.87816486e-01 3.72825712e-01 -8.55428159e-01 1.38004079e-01
5.70011318e-01 9.25261378e-01 -5.35470128e-01 7.89686620e-01
1.35320827e-01 7.22455621e-01 -5.19341588e-01 -4.48831975e-01
-1.10237591e-01 1.07746003e-02 1.07005194e-01 1.25081074e+00
4.86618936e-01 3.50025594e-02 2.62709439e-01 2.45019644e-01
-2.98746794e-01 2.44685411e-01 -3.66171569e-01 -7.60568082e-01
3.78709227e-01 1.21942878e+00 -3.21950652e-02 -1.21112578e-01
-5.59029937e-01 1.22899008e+00 3.78247052e-01 3.99007261e-01
-7.95548081e-01 -2.71176577e-01 1.03890467e+00 8.44943702e-01
4.16246325e-01 -1.48236975e-01 -3.07950247e-02 -1.02577579e+00
6.62519112e-02 -8.43089461e-01 2.99526036e-01 -5.47496319e-01
-1.15337694e+00 3.59829664e-01 -2.04204306e-01 -1.04345441e+00
-4.18378800e-01 -4.01355952e-01 -6.04494989e-01 5.48279226e-01
-6.61947489e-01 -1.16065681e+00 -2.70060748e-01 2.97306806e-01
7.56114483e-01 -1.31151125e-01 1.07573676e+00 7.03112662e-01
-3.52757961e-01 9.11340594e-01 2.97237128e-01 -3.98869365e-01
7.22789943e-01 -1.13279247e+00 6.92662239e-01 2.13927299e-01
2.67028928e-01 1.02930391e+00 2.71782309e-01 -5.35188079e-01
-1.69407630e+00 -9.48587000e-01 9.88780439e-01 -1.01626480e+00
5.95280528e-01 -6.23999238e-01 -1.90725535e-01 9.80080903e-01
-1.14033714e-01 -2.50773579e-01 1.11688650e+00 1.07204819e+00
-3.18176031e-01 -5.17743230e-01 -7.66102910e-01 9.32221174e-01
1.55807269e+00 -7.52387047e-01 -1.10469334e-01 4.25584793e-01
7.84728169e-01 -1.91831272e-02 -1.09407270e+00 1.21075772e-01
1.35417438e+00 -6.77657187e-01 1.00783324e+00 -9.00726259e-01
3.13258439e-01 7.66295940e-02 -3.51254880e-01 -1.12915540e+00
-7.11195648e-01 -1.05750751e+00 -7.46810138e-01 1.13895619e+00
5.49043059e-01 -3.10115486e-01 1.07608545e+00 9.07458365e-01
3.41097787e-02 -1.03542519e+00 -2.05053046e-01 -3.32448423e-01
-7.00239956e-01 -7.60208905e-01 8.11664462e-01 5.47058225e-01
3.83065253e-01 3.85861844e-01 -7.63044834e-01 -2.17367649e-01
-2.10742876e-01 6.70712218e-02 1.07266343e+00 -1.02685130e+00
-9.58544314e-01 -2.49978453e-01 1.40924647e-01 -1.94680882e+00
-5.06133258e-01 -6.60293221e-01 -2.00071126e-01 -1.58582759e+00
-7.66421482e-02 -5.89496911e-01 -8.81981254e-01 5.59779823e-01
3.59097540e-01 -5.62513508e-02 3.51704620e-02 5.92936240e-02
-9.83759284e-01 3.32718730e-01 7.52770543e-01 -1.01153918e-01
-7.84580052e-01 5.31562984e-01 -1.29074526e+00 2.54036129e-01
6.21348977e-01 -1.09699741e-01 -8.93661618e-01 -4.48137313e-01
7.11053967e-01 -3.53650898e-02 -2.12891504e-01 -8.78052711e-01
2.67315451e-02 -2.52427936e-01 3.79321396e-01 -2.87527561e-01
5.02652228e-01 -8.81175041e-01 2.73326668e-03 -2.27674335e-01
-7.42040575e-01 -4.17404622e-02 2.06258092e-02 7.46051133e-01
4.10367876e-01 2.86689073e-01 -9.53611732e-02 2.32946202e-01
-7.43832469e-01 4.71562654e-01 -3.72786731e-01 -5.56851506e-01
6.43736184e-01 -7.85046294e-02 -1.19641989e-01 -8.17226827e-01
-9.46170509e-01 1.49175227e-01 8.84817690e-02 1.12274063e+00
4.22236651e-01 -1.37835717e+00 -1.98109122e-03 2.58136034e-01
2.24465713e-01 -9.38783586e-01 2.55023420e-01 4.75098848e-01
-1.14665516e-01 4.11268324e-01 1.28664598e-01 -1.69082824e-02
-1.16439855e+00 3.80379558e-01 9.50260460e-02 -3.90277386e-01
-3.33640099e-01 9.47232366e-01 -3.56906593e-01 -5.40180743e-01
6.36972785e-01 -2.79633284e-01 -3.04258287e-01 1.63679682e-02
6.77067935e-01 6.74193203e-01 2.38425747e-01 9.39215943e-02
-1.79053724e-01 9.15569905e-03 -4.71336454e-01 -2.70632625e-01
1.39744735e+00 -6.69899583e-02 4.46869373e-01 3.15705389e-01
1.29746592e+00 4.86895442e-01 -1.22627509e+00 -1.47091925e-01
-1.34535849e-01 -7.48674750e-01 2.14855164e-01 -1.24748135e+00
-8.13333809e-01 3.94567788e-01 9.36282694e-01 7.17021078e-02
7.64862657e-01 -1.75004810e-01 1.31431997e+00 6.45495355e-01
5.24017394e-01 -1.45768046e+00 3.76316041e-01 2.97737092e-01
7.73466647e-01 -8.36247921e-01 -4.73264642e-02 -2.94309054e-02
-5.82175672e-01 8.48639727e-01 6.98992372e-01 -1.46722212e-01
9.64348793e-01 -3.51047367e-02 -1.40386462e-01 6.29628599e-02
-1.23106253e+00 -2.33329721e-02 3.16910654e-01 3.72783035e-01
8.53954017e-01 2.20473543e-01 -2.67449141e-01 1.23050702e+00
1.60403967e-01 4.35826391e-01 3.95851582e-02 1.16211760e+00
-2.15827912e-01 -1.57234895e+00 4.64040697e-01 9.68637764e-01
-1.71996117e-01 -5.34244925e-02 -4.44145441e-01 2.71435022e-01
1.04346149e-01 1.15277672e+00 -4.90855016e-02 -1.22419262e+00
8.10598969e-01 1.37148708e-01 3.66318285e-01 -8.13955486e-01
-9.55903769e-01 -8.94479826e-02 6.64930046e-01 -1.45804214e+00
3.30849886e-01 -4.95754033e-01 -7.39830613e-01 -4.05953884e-01
-1.14915475e-01 2.18439773e-01 7.22637415e-01 5.52579463e-01
1.08447433e+00 5.67201316e-01 8.44307721e-01 -9.52643394e-01
-6.99732184e-01 -1.01579964e+00 -4.43223447e-01 5.41818380e-01
9.63537544e-02 -3.68890911e-01 -1.03968285e-01 -3.75916630e-01] | [10.202327728271484, 5.692389488220215] |
0c0150ff-af37-4095-a0d7-1e0dcc5d567a | vocal-style-factorization-for-effective | 2305.07997 | null | https://arxiv.org/abs/2305.07997v1 | https://arxiv.org/pdf/2305.07997v1.pdf | Vocal Style Factorization for Effective Speaker Recognition in Affective Scenarios | The accuracy of automated speaker recognition is negatively impacted by change in emotions in a person's speech. In this paper, we hypothesize that speaker identity is composed of various vocal style factors that may be learned from unlabeled data and re-combined using a neural network architecture to generate holistic speaker identity representations for affective scenarios. In this regard we propose the E-Vector architecture, composed of a 1-D CNN for learning speaker identity features and a vocal style factorization technique for determining vocal styles. Experiments conducted on the MSP-Podcast dataset demonstrate that the proposed architecture improves state-of-the-art speaker recognition accuracy in the affective domain over baseline ECAPA-TDNN speaker recognition models. For instance, the true match rate at a false match rate of 1% improves from 27.6% to 46.2%. | ['Arun Ross', 'Morgan Sandler'] | 2023-05-13 | null | null | null | null | ['speaker-recognition'] | ['speech'] | [ 1.04867995e-01 3.49311833e-03 7.47400522e-02 -1.00831378e+00
-5.76226771e-01 -7.31909394e-01 4.59669173e-01 -4.51439917e-01
-1.76779285e-01 3.40846568e-01 4.56524342e-01 1.87668040e-01
3.94446492e-01 -2.94187367e-01 -3.99312347e-01 -5.49215794e-01
1.59013629e-01 3.83514494e-01 -7.65273750e-01 -3.73241872e-01
-8.41136500e-02 5.86403549e-01 -1.72511578e+00 4.92652029e-01
5.17271221e-01 1.35962391e+00 -5.33634484e-01 8.26714873e-01
-4.58956391e-01 6.40284657e-01 -7.86720812e-01 -5.13579011e-01
7.14927986e-02 -3.96314591e-01 -7.92289615e-01 1.98625657e-03
6.18669927e-01 -2.46182740e-01 -3.14962491e-02 9.16277707e-01
6.83038414e-01 3.79888088e-01 7.81471312e-01 -1.23153830e+00
-7.81969011e-01 9.45919693e-01 -3.01133692e-01 4.01936658e-02
4.07738060e-01 -2.18419671e-01 9.32896256e-01 -1.30330682e+00
1.39115468e-01 1.59978032e+00 7.08531678e-01 1.08884752e+00
-1.30758202e+00 -9.90307271e-01 2.83635139e-01 3.24895680e-01
-1.53359592e+00 -9.16847527e-01 9.76138234e-01 -3.84209603e-01
8.71561229e-01 4.81638640e-01 3.75654936e-01 1.57500660e+00
-2.20094547e-01 7.47692168e-01 9.52864051e-01 -4.57419276e-01
7.11542219e-02 3.89990240e-01 2.55614996e-01 3.66731495e-01
-4.33876097e-01 2.09059238e-01 -1.08213747e+00 -1.16208501e-01
2.92194366e-01 -3.50310206e-01 1.00217603e-01 3.40963811e-01
-7.53172874e-01 7.74907887e-01 1.58124745e-01 4.23882455e-01
-3.32542002e-01 -1.67595685e-01 4.39936191e-01 4.75674719e-01
6.19912505e-01 4.80662107e-01 -2.60179818e-01 -3.55086297e-01
-9.55522180e-01 -1.16678299e-02 9.41219985e-01 5.61044693e-01
6.16761982e-01 8.66928935e-01 -1.90663815e-01 1.50270307e+00
2.21109986e-01 5.45455933e-01 8.16708446e-01 -7.36041665e-01
-1.81821764e-01 4.88189638e-01 -1.14520438e-01 -8.66793811e-01
-3.61015052e-01 -8.17551136e-01 -7.73506045e-01 -1.04647703e-01
5.09435087e-02 -3.78070027e-01 -8.35505545e-01 2.06189036e+00
1.79579526e-01 3.47856462e-01 3.58083695e-01 8.11210692e-01
1.11844313e+00 6.84573948e-01 3.23707331e-03 5.79301827e-02
1.35040390e+00 -9.34693873e-01 -8.01898360e-01 -3.73329967e-01
-1.40623629e-01 -9.32308674e-01 1.08116639e+00 4.59275395e-01
-8.95182848e-01 -9.85049367e-01 -1.01026464e+00 1.82245120e-01
-2.39394441e-01 3.25029671e-01 2.98681974e-01 1.26273179e+00
-1.14194942e+00 1.75290942e-01 -4.11109924e-01 -3.24560255e-01
1.35146931e-01 4.81822252e-01 -2.19879866e-01 4.22839671e-01
-1.11540079e+00 7.35872388e-01 -5.91466315e-02 2.49989614e-01
-1.03129601e+00 -7.92948127e-01 -4.86601532e-01 2.54519373e-01
-2.38696799e-01 -4.92224306e-01 1.54274487e+00 -1.58908057e+00
-2.08527946e+00 8.37444127e-01 -4.71844912e-01 -3.95219386e-01
1.99727625e-01 -2.55698293e-01 -1.12646949e+00 -6.75701797e-02
-2.94019997e-01 6.31463706e-01 1.01490319e+00 -1.38899827e+00
-5.49536824e-01 -6.33388460e-01 -3.33054513e-01 2.33540624e-01
-7.78198063e-01 1.83304891e-01 7.41685033e-02 -6.51978850e-01
-3.31688970e-02 -1.06049740e+00 2.51425415e-01 -3.50983500e-01
-2.29573518e-01 -2.47798383e-01 7.78305531e-01 -7.73751974e-01
9.49685276e-01 -2.18907213e+00 1.01619810e-01 3.10071856e-01
3.63007886e-03 1.61135476e-02 -3.25037897e-01 1.35939181e-01
-2.67249167e-01 -2.06034165e-02 1.70373529e-01 -6.78587496e-01
1.48757607e-01 -9.86340940e-02 -4.71939594e-01 7.76807368e-02
3.64040077e-01 5.44457674e-01 -3.76750678e-01 1.16951182e-01
-7.00511113e-02 7.53199816e-01 -4.51014578e-01 4.65510339e-01
1.80029735e-01 3.91890913e-01 1.23700984e-02 7.44678974e-01
7.29920566e-01 4.11236376e-01 8.19651261e-02 -2.02979043e-01
-6.93161264e-02 3.81655782e-01 -1.09595478e+00 1.51166284e+00
-7.11268783e-01 7.30253816e-01 2.89228022e-01 -6.54848397e-01
1.38548231e+00 4.69624460e-01 3.37033540e-01 -3.52510154e-01
4.42034543e-01 -4.02424484e-03 7.17467815e-02 -9.87429321e-02
6.86101377e-01 -5.51197827e-01 -2.17832208e-01 4.08339709e-01
4.84796941e-01 2.37588361e-01 -3.81227911e-01 -1.56186879e-01
6.36960804e-01 -6.26745582e-01 -8.29351172e-02 -2.67295063e-01
8.39335799e-01 -3.58402193e-01 6.98841453e-01 5.49221098e-01
-3.30751747e-01 4.26584423e-01 1.24480441e-01 -3.26066256e-01
-5.85858762e-01 -1.20531762e+00 -1.33149177e-01 1.68182564e+00
-2.71279663e-01 3.14166471e-02 -8.67731631e-01 -2.79102594e-01
-3.11096832e-02 9.71944928e-01 -8.76666069e-01 -4.91765529e-01
-2.80786932e-01 -4.20546889e-01 1.09875453e+00 6.00112021e-01
2.78770387e-01 -9.68331695e-01 1.41359657e-01 2.29389444e-01
-2.19724253e-01 -1.10428929e+00 -6.71014845e-01 1.30770579e-01
-3.32271367e-01 -5.16075671e-01 -4.98070627e-01 -9.43369567e-01
2.91437984e-01 -3.15461233e-02 1.05232215e+00 -4.86515492e-01
7.00127035e-02 4.76843536e-01 -3.16119671e-01 -6.57008588e-01
-7.87251592e-01 1.53442740e-01 6.73055947e-01 8.30060184e-01
5.62583327e-01 -5.52496076e-01 -4.14537549e-01 2.83979505e-01
-4.26854372e-01 -4.04085755e-01 2.37094924e-01 9.28516090e-01
2.06276059e-01 -3.57422739e-01 1.26500988e+00 -5.51341653e-01
8.04277241e-01 -3.44636202e-01 -7.04784971e-03 1.26670465e-01
-5.08862793e-01 -1.74127489e-01 5.95092475e-01 -7.63862312e-01
-1.44834173e+00 1.82381094e-01 -3.98839921e-01 -7.00737774e-01
-4.04895246e-01 2.35795915e-01 -1.24073029e-01 -8.89628530e-02
7.35200465e-01 4.35515523e-01 2.88400948e-01 -4.36134994e-01
3.73201579e-01 1.17060828e+00 6.40586078e-01 -4.85932320e-01
5.64297855e-01 2.40590811e-01 -6.79738462e-01 -9.79881704e-01
-7.19044209e-01 -4.98381317e-01 -4.57414478e-01 -5.83039820e-01
7.19287932e-01 -1.21043456e+00 -9.68240261e-01 6.52196646e-01
-1.01941097e+00 4.39822860e-02 -1.23693638e-01 4.13231581e-01
-2.54141778e-01 -7.36095756e-02 -6.56926215e-01 -1.05544066e+00
-7.61784673e-01 -9.53836918e-01 9.40403402e-01 3.64673108e-01
-6.47241831e-01 -6.69526279e-01 1.20011486e-01 5.72697043e-01
7.13412821e-01 -3.18840235e-01 7.21875250e-01 -1.08472347e+00
2.93326944e-01 -1.68673545e-02 1.89967632e-01 7.14185953e-01
4.99137551e-01 1.74737407e-03 -1.76644218e+00 -2.00423494e-01
1.48555741e-01 -2.45420486e-01 6.34318531e-01 9.10779908e-02
9.78785276e-01 -4.98258382e-01 2.01064557e-01 3.75036657e-01
7.81406939e-01 4.12543178e-01 1.26752362e-01 -1.95276335e-01
6.31251812e-01 7.36733496e-01 1.67470127e-02 5.84963918e-01
3.59642863e-01 6.02716982e-01 1.96470827e-01 1.13339975e-01
-2.44538531e-01 -6.12256080e-02 6.76683903e-01 1.04423416e+00
2.16896191e-01 -1.75752304e-02 -8.80362689e-01 4.66511756e-01
-1.25007117e+00 -1.06031537e+00 3.63249451e-01 1.84784377e+00
8.81587565e-01 -3.26027393e-01 5.55595398e-01 1.06913224e-02
8.43924165e-01 7.98133016e-02 -7.83423603e-01 -9.56291616e-01
-1.41948462e-01 2.42809996e-01 -1.71701416e-01 5.12562513e-01
-9.33793843e-01 1.05569732e+00 6.56851768e+00 4.01249915e-01
-1.75429225e+00 1.80742778e-02 7.82238185e-01 -4.42096412e-01
-1.56822428e-01 -7.32910633e-01 -8.16694975e-01 1.95949256e-01
1.27258527e+00 -3.78545880e-01 7.05007672e-01 1.12346447e+00
1.37313858e-01 7.24694014e-01 -1.10109997e+00 1.25892222e+00
6.12088263e-01 -9.28762615e-01 1.43171802e-01 -2.05060989e-01
5.78085780e-01 5.54880202e-02 5.76177835e-01 6.02819204e-01
2.85482705e-01 -1.12003291e+00 9.32679772e-01 4.24340636e-01
7.55443573e-01 -1.10884023e+00 6.08762860e-01 -4.22530621e-02
-9.81044352e-01 -1.52946964e-01 5.68069471e-03 -1.20654469e-02
-7.49051198e-02 1.94601446e-01 -1.49198484e+00 1.37239680e-01
7.40846932e-01 4.41524655e-01 -3.50484997e-01 4.20951515e-01
3.07941139e-01 9.84760463e-01 -2.79677153e-01 -1.55459449e-01
-6.47366345e-02 8.85313526e-02 4.82237190e-01 1.31327677e+00
3.81946802e-01 -2.74599344e-02 -1.40497223e-01 9.38034892e-01
-4.52394396e-01 3.31027001e-01 -2.60410845e-01 -9.16838348e-02
6.28301144e-01 1.27561307e+00 -6.17634580e-02 -2.61254311e-01
-1.54068142e-01 9.70415890e-01 2.20554382e-01 3.59458596e-01
-7.05838799e-01 -2.51376443e-02 1.27167439e+00 -3.78713310e-01
3.28056425e-01 2.21321613e-01 -4.30836797e-01 -1.11012042e+00
-2.56797731e-01 -1.27256250e+00 9.16410834e-02 -5.90242803e-01
-1.67298400e+00 1.16753840e+00 -6.01134598e-01 -1.14189732e+00
-6.35230243e-01 -5.20555973e-01 -7.30378866e-01 9.12761152e-01
-1.12997293e+00 -1.26067519e+00 -3.05034369e-01 6.23398006e-01
8.64585757e-01 -8.21989119e-01 1.27824879e+00 1.97109014e-01
-7.35255480e-01 1.22548831e+00 1.62166446e-01 3.25171024e-01
8.05424988e-01 -1.01747382e+00 5.16241670e-01 6.45419896e-01
4.18956965e-01 6.68072999e-01 6.51915669e-01 -1.28987119e-01
-1.36148810e+00 -1.09073842e+00 7.71191239e-01 -1.10827915e-01
4.41335410e-01 -4.26223189e-01 -8.66910040e-01 4.64812160e-01
4.15619999e-01 -3.67014050e-01 1.42213523e+00 6.75574958e-01
-8.41915965e-01 -5.31473339e-01 -1.27010286e+00 5.51941037e-01
7.11593509e-01 -1.00299549e+00 -6.28892362e-01 -2.24682897e-01
3.78506750e-01 3.26304697e-02 -9.10722852e-01 1.62532166e-01
1.01165211e+00 -8.11475158e-01 9.62561786e-01 -7.53676593e-01
1.17930837e-01 -7.22824875e-03 -5.75527191e-01 -1.55474377e+00
-3.92295241e-01 -2.70937920e-01 1.41892284e-01 1.60967016e+00
4.63692635e-01 -4.97382283e-01 8.00478756e-01 4.57056642e-01
-1.80519864e-01 -3.56844962e-01 -1.23584402e+00 -5.43141246e-01
1.46788910e-01 -4.75686818e-01 7.70813286e-01 1.23801637e+00
-7.34695569e-02 8.30923378e-01 -3.77813071e-01 2.48947054e-01
3.81086290e-01 7.86180198e-02 6.35188341e-01 -1.39627242e+00
-1.64114431e-01 -6.72587633e-01 -2.61356801e-01 -7.32182801e-01
6.78384662e-01 -9.68958616e-01 1.10060997e-01 -8.39847922e-01
7.61714503e-02 -2.88274698e-02 -5.79473078e-01 3.81224811e-01
2.76109576e-02 2.20149919e-01 2.53677398e-01 -9.15735513e-02
-1.31395727e-01 9.34021950e-01 6.88098609e-01 -5.16004503e-01
-2.79955775e-01 1.35006323e-01 -9.39695299e-01 6.67370021e-01
1.03031540e+00 -2.45432809e-01 -3.21994245e-01 -2.57063001e-01
-3.14829081e-01 6.37440160e-02 -1.31926015e-02 -1.03632426e+00
5.05883545e-02 1.00145444e-01 4.95136887e-01 -4.59251642e-01
9.25866306e-01 -5.87937057e-01 -3.82680967e-02 1.32230744e-01
-7.22630560e-01 -5.78930750e-02 5.00617325e-01 3.55341792e-01
-4.10512388e-01 2.48424672e-02 6.91137075e-01 2.77701139e-01
-4.86270607e-01 -1.82725210e-02 -6.81756914e-01 -2.40466192e-01
5.13369918e-01 -2.27978855e-01 -9.21154395e-02 -7.28173912e-01
-9.40898895e-01 -1.99228272e-01 -5.10859638e-02 9.98408914e-01
5.42763293e-01 -1.57633793e+00 -9.20588076e-01 4.40745085e-01
1.41427085e-01 -7.44146109e-01 5.67785442e-01 3.03701133e-01
1.37268350e-01 1.42504439e-01 -4.79277909e-01 -6.41023695e-01
-1.62416589e+00 -5.07306168e-03 6.85648799e-01 2.47617736e-01
2.45470956e-01 1.25331390e+00 2.22412959e-01 -7.69581854e-01
4.20481324e-01 -1.53494194e-01 -3.85269016e-01 4.85763073e-01
6.76079154e-01 2.71118581e-01 2.55659670e-01 -1.19324911e+00
-6.13146305e-01 2.68275410e-01 -2.80916661e-01 -3.43641430e-01
1.40848005e+00 -2.29541108e-01 -1.65171027e-02 8.52510750e-01
1.35563815e+00 1.13368697e-01 -8.34168434e-01 -2.51832962e-01
-3.35624725e-01 -9.75631475e-02 1.88746333e-01 -1.19361413e+00
-1.10378611e+00 7.92557240e-01 1.15941501e+00 9.54306424e-02
1.07304192e+00 -1.81289643e-01 6.71129823e-01 4.39661294e-01
-1.34779423e-01 -1.15347934e+00 2.11646780e-01 6.08489156e-01
1.17999840e+00 -1.31579936e+00 -6.10717654e-01 -1.05000362e-01
-9.72591162e-01 1.13610554e+00 7.42619753e-01 1.34808525e-01
7.54410088e-01 1.26365647e-01 6.72222316e-01 2.04033032e-01
-8.25562239e-01 9.46006477e-02 6.47771358e-01 4.58786368e-01
6.18400037e-01 4.20660406e-01 2.49690384e-01 1.32798445e+00
-7.87640929e-01 -4.86185342e-01 3.21733117e-01 2.63757318e-01
-2.14146748e-01 -9.99861300e-01 -4.91458267e-01 2.89866239e-01
-4.57388759e-01 -9.49370041e-02 -9.46689963e-01 6.84462339e-02
-1.05641633e-01 1.34408212e+00 3.07767242e-01 -1.03758228e+00
4.13616151e-01 7.36142218e-01 2.18852326e-01 -5.17004490e-01
-1.14516342e+00 1.11640766e-01 3.30329716e-01 -2.31026292e-01
-4.04930443e-01 -7.75189877e-01 -1.03896701e+00 -3.13686639e-01
1.56544764e-02 1.24118365e-01 9.13979292e-01 6.80531800e-01
5.58502913e-01 6.69917643e-01 9.87181306e-01 -7.23191023e-01
-7.13133454e-01 -1.29844987e+00 -5.99157810e-01 3.39936554e-01
4.93646473e-01 -5.92906237e-01 -4.29120064e-01 1.90184459e-01] | [14.060338973999023, 6.023562908172607] |
5d8dbd68-6305-4f09-bf75-d474f640fe64 | patch2pix-epipolar-guided-pixel-level | 2012.01909 | null | https://arxiv.org/abs/2012.01909v3 | https://arxiv.org/pdf/2012.01909v3.pdf | Patch2Pix: Epipolar-Guided Pixel-Level Correspondences | The classical matching pipeline used for visual localization typically involves three steps: (i) local feature detection and description, (ii) feature matching, and (iii) outlier rejection. Recently emerged correspondence networks propose to perform those steps inside a single network but suffer from low matching resolution due to the memory bottleneck. In this work, we propose a new perspective to estimate correspondences in a detect-to-refine manner, where we first predict patch-level match proposals and then refine them. We present Patch2Pix, a novel refinement network that refines match proposals by regressing pixel-level matches from the local regions defined by those proposals and jointly rejecting outlier matches with confidence scores. Patch2Pix is weakly supervised to learn correspondences that are consistent with the epipolar geometry of an input image pair. We show that our refinement network significantly improves the performance of correspondence networks on image matching, homography estimation, and localization tasks. In addition, we show that our learned refinement generalizes to fully-supervised methods without re-training, which leads us to state-of-the-art localization performance. The code is available at https://github.com/GrumpyZhou/patch2pix. | ['Laura Leal-Taixe', 'Torsten Sattler', 'Qunjie Zhou'] | 2020-12-03 | null | http://openaccess.thecvf.com//content/CVPR2021/html/Zhou_Patch2Pix_Epipolar-Guided_Pixel-Level_Correspondences_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Zhou_Patch2Pix_Epipolar-Guided_Pixel-Level_Correspondences_CVPR_2021_paper.pdf | cvpr-2021-1 | ['homography-estimation'] | ['computer-vision'] | [-3.35414074e-02 -1.93718672e-01 -1.01601884e-01 -3.27416599e-01
-9.63881552e-01 -4.84489202e-01 5.23151577e-01 3.01945150e-01
-4.32837993e-01 2.03751162e-01 3.89235727e-02 9.77731645e-02
2.02785030e-01 -6.26816690e-01 -1.16066504e+00 -3.77400815e-01
1.67671487e-01 5.41743219e-01 4.63489980e-01 8.64724964e-02
4.20367092e-01 7.56367147e-01 -1.59950256e+00 2.83532858e-01
7.36024022e-01 1.09419572e+00 -4.66027781e-02 4.74052936e-01
1.84628725e-01 3.88332307e-01 -1.03366576e-01 -1.30104870e-01
6.94230855e-01 -3.19340318e-01 -6.37270927e-01 6.80317208e-02
1.14004362e+00 -3.75409871e-01 -3.73025507e-01 1.25553501e+00
3.16800475e-01 1.31330371e-01 3.43052417e-01 -1.28720570e+00
-5.47881544e-01 1.89209804e-01 -7.43589520e-01 -6.79697096e-02
3.80567074e-01 1.24338947e-01 1.11022747e+00 -1.63679874e+00
7.24349499e-01 8.99349689e-01 1.15112495e+00 5.83896823e-02
-1.53353333e+00 -6.38022006e-01 3.91409695e-02 6.49746507e-02
-1.94051218e+00 -6.73614144e-01 7.98060358e-01 -5.06329775e-01
9.05561268e-01 3.48394252e-02 6.11568511e-01 6.50852501e-01
-1.77065320e-02 5.17068684e-01 6.61192477e-01 -4.44865078e-01
1.46184832e-01 -2.96927422e-01 -3.09857149e-02 9.26027834e-01
7.76396096e-02 3.04333389e-01 -8.01074624e-01 -3.33848327e-01
1.11747491e+00 2.34063730e-01 -2.83402354e-01 -9.34458613e-01
-1.42758679e+00 6.85311913e-01 8.31425250e-01 1.75522566e-01
-3.52782696e-01 3.37275535e-01 8.56377855e-02 2.92569220e-01
6.27445161e-01 4.10430610e-01 -2.88067698e-01 3.37794185e-01
-1.24349105e+00 1.34576738e-01 6.42982423e-01 1.12984502e+00
1.29810691e+00 -3.16995829e-01 1.51702538e-02 6.97167814e-01
2.01309964e-01 2.54234493e-01 2.26732522e-01 -1.10003865e+00
2.63494223e-01 5.58236599e-01 2.22319722e-01 -1.36576617e+00
-3.84985656e-01 -4.72122133e-01 -8.31931949e-01 3.72561127e-01
4.83650059e-01 3.25020015e-01 -1.03186536e+00 1.54148579e+00
3.42205077e-01 6.05664194e-01 -3.98944438e-01 1.07650101e+00
7.71846533e-01 2.88093179e-01 -4.09961760e-01 3.56832296e-01
1.00550902e+00 -1.12938392e+00 -1.05081216e-01 -1.67649284e-01
5.85917532e-01 -1.04972661e+00 8.18053246e-01 9.68124792e-02
-1.11784387e+00 -7.06222534e-01 -1.04679751e+00 -3.18470925e-01
-1.75760776e-01 3.79191846e-01 2.82977134e-01 3.03161070e-02
-1.37514555e+00 9.41178143e-01 -1.00280285e+00 -4.55266148e-01
5.62735021e-01 4.48202103e-01 -6.34123087e-01 -6.08028434e-02
-5.75205445e-01 6.29733443e-01 9.81328562e-02 2.08625570e-01
-6.01571858e-01 -7.61527956e-01 -1.09132159e+00 -4.81488090e-03
2.95914501e-01 -8.93397570e-01 9.91284251e-01 -8.37713361e-01
-1.17048001e+00 1.03626227e+00 -5.12879014e-01 -2.99155623e-01
7.20540524e-01 -3.07840496e-01 1.83848962e-01 1.02337621e-01
4.97583628e-01 8.74337792e-01 7.69376576e-01 -1.21257341e+00
-8.37223113e-01 -1.90205932e-01 -4.57437783e-01 6.60757944e-02
1.21574573e-01 -8.54499415e-02 -9.08729017e-01 -6.32138371e-01
9.79704440e-01 -8.54572117e-01 -3.66861999e-01 6.61988258e-01
-3.31342340e-01 -1.94376767e-01 4.05989379e-01 -5.05907357e-01
7.74160564e-01 -2.29558778e+00 -3.79532985e-02 7.10387826e-01
5.12391388e-01 -1.23897411e-01 -3.09116811e-01 2.83310384e-01
-1.62977979e-01 -1.95352837e-01 -2.19964668e-01 -9.19102550e-01
8.08045268e-02 -7.67860711e-02 -3.30879897e-01 1.04997802e+00
1.79508105e-01 8.07210147e-01 -9.53376293e-01 -2.65286356e-01
3.82247418e-01 3.30557734e-01 -7.13818491e-01 2.34233662e-01
-7.69788995e-02 5.81698418e-01 -3.72594371e-02 8.83233786e-01
9.22332168e-01 -3.48921031e-01 -2.85675347e-01 -5.87660849e-01
-3.40830773e-01 1.33011267e-01 -1.43097293e+00 2.04565763e+00
-1.37055159e-01 6.85153604e-01 -6.75242543e-02 -8.87038291e-01
1.07649779e+00 -4.69123060e-03 7.05578208e-01 -6.13045812e-01
-2.50989106e-03 4.65911329e-01 -3.00707132e-01 -8.19027517e-03
4.31143880e-01 3.95754963e-01 2.28678718e-01 3.69472057e-01
2.15753540e-01 -3.75477299e-02 -1.39262015e-02 3.73493545e-02
1.06394863e+00 4.60184187e-01 3.92227441e-01 -1.73915610e-01
3.16300243e-01 9.65314582e-02 8.97082508e-01 9.52404320e-01
-1.12350874e-01 1.22918653e+00 2.24502996e-01 -7.75945067e-01
-1.45694768e+00 -1.17995918e+00 -7.84590021e-02 7.50483155e-01
5.49428999e-01 -5.05652487e-01 -4.85603869e-01 -4.71035868e-01
2.80651093e-01 -9.76423733e-03 -3.83545518e-01 1.19490497e-01
-7.24628627e-01 -9.45692658e-02 4.74333167e-01 6.09313965e-01
5.19110560e-01 -8.58193457e-01 -2.92040259e-01 5.61169274e-02
-4.57839258e-02 -1.01865351e+00 -9.02490973e-01 5.64375371e-02
-7.25772798e-01 -1.14184403e+00 -5.89949310e-01 -9.96772170e-01
1.20773339e+00 4.24309552e-01 1.09454465e+00 5.91206789e-01
-2.03607813e-01 1.24538131e-01 -1.17548175e-01 2.29879349e-01
-5.11409296e-03 1.24808997e-01 2.65396625e-01 8.28879979e-03
4.03738678e-01 -7.38444448e-01 -7.03717530e-01 5.60968220e-01
-5.67168474e-01 6.04772009e-02 6.72633827e-01 9.33673322e-01
1.17113185e+00 -3.45409513e-01 -3.26472111e-02 -5.81649184e-01
4.39521670e-02 -1.98487625e-01 -1.00721014e+00 1.81608543e-01
-4.02843714e-01 4.91862604e-03 5.15163898e-01 -2.65592426e-01
-4.55340654e-01 7.10634053e-01 -1.71457827e-01 -8.71445179e-01
-3.09811711e-01 2.92044818e-01 9.17766169e-02 -7.08982766e-01
8.17359388e-01 9.79583785e-02 -3.46170254e-02 -5.72021961e-01
4.39290017e-01 1.92475781e-01 9.98602033e-01 -5.33599257e-01
1.32157135e+00 7.43055999e-01 2.93152258e-02 -5.52024662e-01
-8.09952080e-01 -9.24144745e-01 -1.10545766e+00 5.58921173e-02
5.31101704e-01 -1.10697937e+00 -6.62720859e-01 3.54701996e-01
-1.29912627e+00 -2.95418084e-01 -2.53258646e-01 5.48333824e-01
-6.64881766e-01 5.64114690e-01 -6.57720983e-01 -1.88224927e-01
-2.75448918e-01 -1.03854191e+00 1.16096413e+00 7.99827576e-02
-3.07729989e-01 -6.61603510e-01 2.79213578e-01 -4.20542322e-02
1.97559893e-01 1.29300788e-01 1.99466690e-01 -4.80730325e-01
-1.02700448e+00 -2.84108341e-01 -4.50159162e-01 1.11430861e-01
8.84723589e-02 4.18723524e-02 -8.24999809e-01 -4.63673383e-01
-3.13210696e-01 -1.41401917e-01 9.98071015e-01 3.69367242e-01
1.03778946e+00 -2.46015295e-01 -4.71016109e-01 1.34402025e+00
1.35899329e+00 -4.57717329e-01 5.51996827e-01 5.21485150e-01
8.85024905e-01 3.86736214e-01 6.57304049e-01 3.05626124e-01
2.64676869e-01 8.74774218e-01 4.69655216e-01 -4.90270257e-01
-1.24919392e-01 -5.15600145e-01 2.27863379e-02 5.15056610e-01
1.03340678e-01 2.12344319e-01 -8.89216185e-01 7.86785960e-01
-2.36289024e+00 -8.02404165e-01 -1.21085308e-01 2.37623978e+00
6.90369844e-01 -1.03174403e-01 8.19590166e-02 -3.68544787e-01
8.37554157e-01 8.00139830e-03 -4.03243780e-01 1.66134313e-01
-1.90366030e-01 2.37131298e-01 7.07307637e-01 8.28704715e-01
-1.30027032e+00 1.17212534e+00 5.54527378e+00 5.44487000e-01
-9.75804806e-01 -1.06949741e-02 3.39799523e-01 1.53910860e-01
-1.69911712e-01 3.62965524e-01 -6.21955395e-01 2.29849845e-01
1.08598426e-01 2.57506579e-01 3.24631125e-01 8.82169664e-01
3.14139724e-02 -7.31469467e-02 -1.36621523e+00 1.21950626e+00
2.27172509e-01 -1.63077033e+00 -5.61898500e-02 -8.48867744e-02
9.03652787e-01 4.67677206e-01 -1.37936324e-01 -2.73504525e-01
2.09680960e-01 -9.05282736e-01 7.63582170e-01 5.69548428e-01
5.93496442e-01 -6.43455982e-01 7.27349579e-01 1.76924706e-01
-1.38672674e+00 3.14754993e-01 -7.56315172e-01 7.54469037e-02
9.81042534e-02 7.04875648e-01 -7.27214515e-01 5.23318291e-01
8.95156384e-01 1.05488288e+00 -6.13822818e-01 1.52175176e+00
-3.61356825e-01 1.65076796e-02 -6.60195887e-01 5.58259785e-01
4.66016866e-03 -3.67109448e-01 5.19005001e-01 1.03407800e+00
5.49970686e-01 -1.70835361e-01 5.33113718e-01 1.26727998e+00
-1.69501290e-01 2.37937905e-02 -5.85350037e-01 5.99936187e-01
7.17455029e-01 1.22918046e+00 -8.20160389e-01 -2.40103915e-01
-4.57344919e-01 1.31034517e+00 6.71263337e-01 2.50771374e-01
-5.46683133e-01 -4.41175252e-01 7.58904815e-01 1.57677203e-01
4.26228940e-01 -3.26713353e-01 -4.15857196e-01 -1.36036682e+00
3.54569912e-01 -5.46915948e-01 1.59924626e-01 -6.42792523e-01
-1.30925035e+00 5.45373499e-01 -3.31966043e-01 -1.51564884e+00
-1.21121280e-01 -3.84316981e-01 -7.29502678e-01 9.10428286e-01
-1.38238811e+00 -1.14328146e+00 -7.49616027e-01 5.17276466e-01
2.57520825e-01 9.41964313e-02 4.86970246e-01 4.53798205e-01
-3.86400998e-01 7.53790200e-01 1.00526057e-01 4.71169949e-01
1.03909481e+00 -1.16909838e+00 8.79852712e-01 1.14730513e+00
4.96600866e-01 7.26323247e-01 5.33309340e-01 -6.63972795e-01
-9.21481848e-01 -1.13938379e+00 1.08522367e+00 -4.92843181e-01
5.26374698e-01 -4.67902452e-01 -9.64059353e-01 7.81612575e-01
-1.39683306e-01 5.53127110e-01 8.54928344e-02 -4.88522276e-02
-5.62189639e-01 -1.89069018e-01 -1.01653647e+00 6.96779609e-01
1.23961926e+00 -6.03766263e-01 -2.97048479e-01 3.10405284e-01
4.15484458e-01 -7.94603944e-01 -7.36704051e-01 3.31849366e-01
5.30160844e-01 -1.16097200e+00 1.17905402e+00 -1.29851550e-01
5.25796786e-02 -9.22320485e-01 -5.36064990e-02 -1.11114562e+00
-4.66985226e-01 -6.11477196e-01 1.22168742e-01 9.85702157e-01
3.50717068e-01 -7.17993796e-01 1.01205909e+00 3.98431063e-01
-3.09934646e-01 -5.77630639e-01 -9.94800746e-01 -7.43704259e-01
-3.60191613e-01 -2.01689646e-01 5.17780483e-01 1.00777423e+00
-1.97565272e-01 -6.17439933e-02 -3.85767639e-01 5.96299708e-01
9.99365211e-01 2.08973169e-01 1.10331249e+00 -1.13672030e+00
-2.04392746e-01 -3.91344905e-01 -7.37474918e-01 -1.53706503e+00
1.28796697e-01 -8.43547761e-01 4.70747471e-01 -1.33332801e+00
1.48118511e-01 -5.58830857e-01 -7.86425173e-03 5.99509478e-01
-2.36949753e-02 8.15026700e-01 -4.71617393e-02 7.10565269e-01
-6.30473793e-01 3.75585616e-01 6.84211552e-01 4.81881648e-02
-2.90215909e-01 3.33374850e-02 -2.23129898e-01 8.34722698e-01
7.63915837e-01 -5.87610066e-01 1.33966014e-01 -4.69677359e-01
2.28272110e-01 -3.45864356e-01 6.64396465e-01 -1.18611300e+00
8.14253986e-01 5.39412796e-02 6.67673588e-01 -8.54843616e-01
2.15618029e-01 -8.00345898e-01 2.09005550e-01 3.60625595e-01
-1.66903004e-01 4.07540917e-01 -6.27102628e-02 4.42340225e-01
-2.72142053e-01 -1.32545248e-01 7.91054845e-01 -2.59992760e-02
-7.86925495e-01 6.62026227e-01 9.18695256e-02 -2.52706617e-01
7.34244823e-01 -4.12637502e-01 -2.69795775e-01 -4.17704672e-01
-7.16387689e-01 2.40938872e-01 1.08401763e+00 1.88237399e-01
7.85907149e-01 -1.54822636e+00 -7.47173488e-01 4.92734939e-01
3.91983092e-01 4.01428312e-01 -9.18843970e-02 1.06785095e+00
-8.97294819e-01 9.47675332e-02 1.77245717e-02 -1.11738539e+00
-1.22712505e+00 3.16149980e-01 5.21479368e-01 1.64710283e-01
-9.36247230e-01 8.19931149e-01 3.31681311e-01 -6.16281986e-01
4.78128850e-01 -3.93663228e-01 3.56599182e-01 -3.95676404e-01
2.44644389e-01 2.52024621e-01 9.93202254e-02 -8.22904706e-01
-5.44071913e-01 1.06547606e+00 3.18731740e-03 4.05688658e-02
1.08251226e+00 -7.66905472e-02 -4.45397228e-01 3.53792422e-02
1.23727262e+00 9.84718576e-02 -1.39681852e+00 -6.63651288e-01
2.09679723e-01 -7.80051053e-01 -1.26674414e-01 -1.76076293e-01
-1.09702241e+00 4.93051946e-01 4.24728751e-01 -3.87926847e-01
9.00750339e-01 1.45357072e-01 6.62589550e-01 4.79564309e-01
2.81693429e-01 -1.05513597e+00 1.49051666e-01 4.71087992e-01
7.82203913e-01 -1.35235643e+00 1.44803286e-01 -4.64723974e-01
2.75860969e-02 1.15398896e+00 7.30298638e-01 -6.49034441e-01
6.21229827e-01 1.24514244e-01 2.04076961e-01 -2.54504323e-01
-2.73862451e-01 -2.99323499e-01 6.24127150e-01 5.32878578e-01
3.29261929e-01 -3.80126774e-01 4.76248227e-02 -4.66370918e-02
1.25149135e-02 -2.17773899e-01 1.85340524e-01 7.61413336e-01
-4.81545657e-01 -1.01907682e+00 -5.15522659e-01 2.21236095e-01
-7.51661733e-02 -2.89470792e-01 -4.85637635e-01 6.61180735e-01
2.16646284e-01 6.13447845e-01 3.54956925e-01 -3.32902521e-01
4.38186914e-01 -3.32279325e-01 4.07326728e-01 -7.33372748e-01
-4.81551051e-01 2.27700874e-01 -2.49552712e-01 -1.13145292e+00
-3.75261426e-01 -6.31353736e-01 -1.05179858e+00 -3.65060538e-01
-4.02005583e-01 5.32681048e-02 3.85522187e-01 7.68145800e-01
6.05933011e-01 1.69881538e-03 6.58310473e-01 -1.22731471e+00
-3.10993880e-01 -8.00485909e-01 -2.17565432e-01 4.93260622e-01
6.15834236e-01 -4.64410067e-01 -5.07917643e-01 -1.03112377e-01] | [8.074464797973633, -2.2149860858917236] |
105b84c6-81bf-4b86-80fe-473a013b2fc1 | a2-efficient-automated-attacker-for-boosting | 2210.03543 | null | https://arxiv.org/abs/2210.03543v2 | https://arxiv.org/pdf/2210.03543v2.pdf | A2: Efficient Automated Attacker for Boosting Adversarial Training | Based on the significant improvement of model robustness by AT (Adversarial Training), various variants have been proposed to further boost the performance. Well-recognized methods have focused on different components of AT (e.g., designing loss functions and leveraging additional unlabeled data). It is generally accepted that stronger perturbations yield more robust models. However, how to generate stronger perturbations efficiently is still missed. In this paper, we propose an efficient automated attacker called A2 to boost AT by generating the optimal perturbations on-the-fly during training. A2 is a parameterized automated attacker to search in the attacker space for the best attacker against the defense model and examples. Extensive experiments across different datasets demonstrate that A2 generates stronger perturbations with low extra cost and reliably improves the robustness of various AT methods against different attacks. | ['Yihua Huang', 'Ming Gu', 'Weiqiang Wang', 'ZhenZhe Ying', 'Shiwen Cui', 'Changhua Meng', 'Guanghui Zhu', 'Zhuoer Xu'] | 2022-10-07 | null | null | null | null | ['adversarial-defense'] | ['adversarial'] | [ 1.05026364e-01 1.79689787e-02 -1.41630694e-01 -2.06826240e-01
-1.14195597e+00 -1.16970813e+00 4.91816789e-01 -6.23136312e-02
-2.32221529e-01 5.46961367e-01 -2.08774939e-01 -4.78092968e-01
6.38819337e-02 -7.69550741e-01 -7.54399657e-01 -8.44280958e-01
7.28826001e-02 2.82902449e-01 3.06408852e-01 -2.52636224e-01
2.66832709e-01 7.03108251e-01 -7.82959580e-01 1.31001651e-01
7.66106546e-01 5.46041965e-01 -5.24467051e-01 6.73678100e-01
1.91922784e-01 4.20603931e-01 -9.92971063e-01 -8.87707412e-01
9.35583651e-01 -3.27022135e-01 -5.55084109e-01 -2.59118617e-01
1.91180170e-01 -5.51071107e-01 -3.54560703e-01 1.21675920e+00
6.85319781e-01 -3.56119499e-02 4.96330023e-01 -1.48412514e+00
-4.94671494e-01 1.02968419e+00 -4.94616836e-01 1.41648099e-01
1.32451534e-01 5.47297955e-01 7.33922780e-01 -7.51610100e-01
1.31139144e-01 1.44978452e+00 5.79443514e-01 1.05929017e+00
-1.21530080e+00 -1.03996193e+00 4.88421507e-02 1.53894901e-01
-1.46775901e+00 -2.19639271e-01 9.53194439e-01 -1.61568537e-01
4.05655295e-01 5.73210895e-01 -1.29474774e-01 1.70855820e+00
-1.67203188e-01 6.39348626e-01 1.09434664e+00 -4.06670213e-01
4.13537979e-01 4.57998365e-01 -3.37262154e-02 4.52219993e-01
4.59976524e-01 4.24174398e-01 -8.32592547e-02 -7.99751520e-01
5.04313171e-01 -3.68833989e-02 -3.97810847e-01 -6.44477680e-02
-5.95101357e-01 1.00128067e+00 4.08232808e-01 -7.85968080e-02
-1.13313407e-01 1.06676675e-01 4.56020474e-01 2.14821607e-01
9.82374176e-02 9.25840855e-01 -6.37422144e-01 3.66724916e-02
-4.62907910e-01 3.67711276e-01 7.34769464e-01 6.53610051e-01
5.84217548e-01 3.29181612e-01 -4.45348807e-02 6.13149464e-01
1.35355026e-01 6.40979528e-01 3.83161664e-01 -6.69616163e-01
7.73948967e-01 5.70441425e-01 1.74150407e-01 -1.06869447e+00
1.24748491e-01 -4.62855518e-01 -5.58100402e-01 1.78553849e-01
2.71227926e-01 -5.64644814e-01 -9.29281235e-01 1.87506306e+00
6.05906546e-01 3.95785928e-01 9.06476676e-02 6.01063073e-01
1.57643005e-01 4.87869740e-01 1.52588829e-01 1.58983797e-01
8.29002321e-01 -8.73332381e-01 -3.72988939e-01 -2.19050512e-01
6.09576821e-01 -5.77193737e-01 1.05162215e+00 3.88396412e-01
-8.84087980e-01 -3.69362354e-01 -1.17263651e+00 5.78218997e-01
-5.99033773e-01 -2.26584271e-01 3.61070782e-01 1.41172945e+00
-4.54903573e-01 6.68118238e-01 -8.87997866e-01 1.02972709e-01
4.35465962e-01 5.92616081e-01 -3.07387114e-01 1.21965326e-01
-1.50527501e+00 9.25748765e-01 2.64904648e-01 2.34420542e-02
-1.45222914e+00 -9.02331650e-01 -7.56191254e-01 -2.21537985e-02
4.97223645e-01 -3.21900398e-01 1.21653116e+00 -6.08630478e-01
-1.65987670e+00 2.32691199e-01 3.61336648e-01 -9.55089927e-01
8.51295888e-01 -4.83767182e-01 -3.12344402e-01 2.07258582e-01
-4.22473848e-01 4.03755128e-01 1.14077866e+00 -1.60024965e+00
-2.06783041e-01 -2.97531664e-01 3.81880254e-01 -1.35844067e-01
-9.36656833e-01 2.34807447e-01 -1.04009219e-01 -7.41613567e-01
-3.67450207e-01 -1.14705443e+00 -5.30651093e-01 -2.17408046e-01
-9.02870715e-01 2.22509980e-01 1.31567049e+00 -6.16486311e-01
1.36533403e+00 -2.02627325e+00 -6.92161471e-02 3.84897381e-01
1.72182828e-01 1.02862954e+00 -1.80918530e-01 7.10212946e-01
-1.89586192e-01 6.94195867e-01 -3.79421085e-01 -2.16760829e-01
1.44115075e-01 6.45520911e-02 -9.98306751e-01 4.24073100e-01
2.32900605e-01 8.19322348e-01 -7.50468969e-01 -1.40720189e-01
1.59529343e-01 3.24023396e-01 -6.42722905e-01 4.67616409e-01
-2.24598840e-01 4.17706341e-01 -6.66547000e-01 6.47887409e-01
8.96038413e-01 1.26195595e-01 4.48066518e-02 1.23969410e-02
3.97156596e-01 -5.47130639e-03 -1.28533185e+00 6.28122926e-01
-2.42462397e-01 2.28482679e-01 5.27539626e-02 -9.98533487e-01
9.26718891e-01 3.55265141e-01 1.00538418e-01 -8.85406584e-02
2.95605719e-01 1.39146924e-01 1.32388458e-01 -2.20146388e-01
3.73259708e-02 1.51964918e-01 -1.52423456e-01 4.30236399e-01
-3.51504564e-01 4.35053669e-02 -2.91205704e-01 6.51132464e-02
1.14286196e+00 -1.26805410e-01 2.96683908e-01 1.51588738e-01
8.77403796e-01 -2.25395769e-01 7.11393654e-01 1.00211120e+00
-3.58643949e-01 4.48639154e-01 3.09936851e-01 -3.16822499e-01
-1.00920129e+00 -9.76480007e-01 1.52470514e-01 6.48121536e-01
3.60059142e-02 -4.97432023e-01 -1.15321803e+00 -1.30700207e+00
9.19208154e-02 7.35923648e-01 -7.81344831e-01 -5.40961146e-01
-8.11777771e-01 -6.58863962e-01 1.16459000e+00 4.44159389e-01
8.57406497e-01 -7.00670838e-01 -2.33449548e-01 -2.72752699e-02
9.98043716e-02 -1.20283687e+00 -5.57467341e-01 3.10791954e-02
-6.33201599e-01 -1.10576892e+00 -4.41811979e-01 -2.53242791e-01
7.90251851e-01 2.41669059e-01 5.58455527e-01 3.06981057e-01
-2.88757324e-01 1.66873828e-01 -4.09904569e-01 -7.12147176e-01
-8.35359156e-01 7.41399676e-02 2.39010051e-01 2.61195362e-01
1.23729602e-01 -5.28359652e-01 -3.80830318e-01 6.95967674e-01
-1.02878487e+00 -5.77404678e-01 5.33416212e-01 8.72704804e-01
2.02317730e-01 2.40438849e-01 5.73797226e-01 -1.21839428e+00
6.47298157e-01 -5.52854180e-01 -7.18183458e-01 2.63927668e-01
-5.12454450e-01 1.36534959e-01 1.39323926e+00 -8.57990503e-01
-7.34675467e-01 -1.27034292e-01 -1.24157093e-01 -8.83025110e-01
-2.26409838e-01 2.82538272e-02 -5.80802083e-01 -6.59689963e-01
1.07632864e+00 2.36037463e-01 -2.34951749e-01 -3.95879835e-01
3.16292554e-01 6.03079975e-01 1.51375026e-01 -8.22866797e-01
1.91145337e+00 3.00712377e-01 2.16174126e-01 -5.24491310e-01
-1.00263119e+00 -1.39803603e-01 -2.33635649e-01 9.87672433e-02
3.48420113e-01 -4.99027222e-01 -5.92420518e-01 6.71819448e-01
-8.39346290e-01 -2.86189318e-01 6.27551377e-02 1.20593548e-01
-2.57496178e-01 6.94233179e-01 -3.37346822e-01 -9.23410058e-01
-6.04470968e-01 -1.32790387e+00 6.64696932e-01 2.13708982e-01
1.11668706e-01 -9.69491243e-01 2.07541902e-02 4.02877957e-01
5.93807995e-01 5.70004225e-01 8.73546362e-01 -1.25703478e+00
-4.70562130e-01 -6.54597998e-01 2.63438463e-01 6.99297607e-01
1.31831244e-01 2.86542892e-01 -1.08552849e+00 -5.04032373e-01
1.35003626e-01 -3.27335328e-01 3.93934339e-01 -9.48029980e-02
1.52667010e+00 -9.42250848e-01 -3.39048952e-01 8.41303170e-01
1.32831836e+00 2.81041324e-01 6.28964484e-01 4.44946468e-01
6.72581136e-01 3.07797074e-01 7.04499781e-01 4.35938179e-01
-4.41876084e-01 7.26255894e-01 6.43998146e-01 2.52511948e-02
3.32919717e-01 -4.75676417e-01 5.12956679e-01 1.55048132e-01
3.58688414e-01 -1.47906452e-01 -8.37869585e-01 2.53800303e-01
-1.50286150e+00 -1.15633881e+00 3.90808493e-01 2.24259853e+00
1.00890481e+00 4.98543054e-01 2.31513888e-01 3.57571453e-01
9.08473074e-01 9.31009576e-02 -7.47814894e-01 -3.20153505e-01
5.83867095e-02 2.23459840e-01 7.02543557e-01 6.37347162e-01
-1.12505376e+00 1.11045456e+00 6.53945637e+00 9.79666293e-01
-1.16703153e+00 -9.23556611e-02 6.11513257e-01 2.59097293e-03
-1.37527391e-01 1.75168827e-01 -1.13810670e+00 4.46302682e-01
9.93853033e-01 -4.35156614e-01 4.69648153e-01 1.23334813e+00
3.96236889e-02 7.30577588e-01 -8.78469288e-01 5.26437223e-01
6.17301976e-03 -1.12859321e+00 4.90500242e-01 1.47088021e-01
7.41225660e-01 -4.03495729e-01 3.64696324e-01 5.57227135e-01
6.63997173e-01 -9.19953406e-01 3.33208233e-01 -5.93577437e-02
4.25091654e-01 -1.15708685e+00 6.37552917e-01 5.54881513e-01
-9.16049242e-01 -5.13886333e-01 -3.82874668e-01 3.57986689e-01
-1.32917330e-01 3.26195896e-01 -1.32828510e+00 5.82371056e-01
3.21967155e-01 1.27496004e-01 -8.56064677e-01 9.76617277e-01
-5.61509371e-01 1.07045031e+00 -3.19481939e-01 -1.41439233e-02
3.01789820e-01 -7.14748027e-03 7.92152643e-01 1.05701041e+00
-9.19236392e-02 5.81174642e-02 3.80560249e-01 7.49255657e-01
-1.75115272e-01 -1.08653978e-01 -7.45461762e-01 -1.84204191e-01
1.01367021e+00 1.32148540e+00 -2.00798497e-01 -9.22952592e-03
8.33486542e-02 6.18360817e-01 2.89800256e-01 2.99628317e-01
-1.34221423e+00 -5.28371990e-01 8.32137048e-01 1.08904140e-02
1.54010236e-01 -3.00366580e-01 -2.25960344e-01 -8.41632605e-01
-9.00963992e-02 -1.51790202e+00 5.65204740e-01 -2.08988234e-01
-1.29682708e+00 7.45335519e-01 7.54387826e-02 -1.44440842e+00
3.29006538e-02 -4.97613460e-01 -8.38124871e-01 8.49848390e-01
-1.27264202e+00 -1.03178787e+00 -9.95785296e-02 7.66593337e-01
4.12451327e-01 -3.78837109e-01 8.12102318e-01 1.06331408e-02
-1.07075870e+00 1.39570177e+00 -1.90854833e-01 3.98705781e-01
6.70498908e-01 -1.25432682e+00 7.56421864e-01 1.31711388e+00
8.89845565e-02 1.09782887e+00 7.96692431e-01 -5.95979333e-01
-1.33215654e+00 -1.18569040e+00 1.03837743e-01 -5.41564763e-01
8.53049636e-01 -4.96917486e-01 -9.25442755e-01 5.24629772e-01
-4.66016568e-02 2.04748176e-02 7.58738697e-01 -4.23958540e-01
-7.72431731e-01 -4.75718677e-02 -1.65196860e+00 1.13644564e+00
6.28917933e-01 -4.64402318e-01 -1.84711054e-01 3.91944617e-01
9.63529408e-01 -3.78613949e-01 -6.87773168e-01 5.77530265e-01
-6.21205904e-02 -7.26023436e-01 1.18774903e+00 -9.47515488e-01
-1.53664546e-02 -2.58195221e-01 -1.41112059e-01 -1.19633007e+00
-4.62941192e-02 -1.38268912e+00 -4.73406315e-01 1.34335554e+00
3.23475629e-01 -9.40602899e-01 9.78386521e-01 6.61383033e-01
1.50524929e-01 -8.63907337e-01 -6.94254100e-01 -9.63727355e-01
2.24543139e-01 -9.49975252e-02 7.70684242e-01 9.38625574e-01
-4.01142597e-01 -8.35633576e-02 -5.02561927e-01 9.12594497e-01
9.29606080e-01 -3.87676030e-01 1.29400015e+00 -9.53657091e-01
-5.63136458e-01 -3.09547275e-01 -4.07587528e-01 -6.43121123e-01
2.95905679e-01 -4.86408442e-01 -1.79603413e-01 -6.18205249e-01
-7.40719810e-02 -2.71731555e-01 -1.86043933e-01 6.76069558e-01
-7.81990826e-01 -7.15183243e-02 1.68277815e-01 1.57489896e-01
-6.88494220e-02 4.67806578e-01 8.69353294e-01 -1.21100910e-01
6.19448908e-02 2.97031969e-01 -8.77262592e-01 6.56395555e-01
1.23715663e+00 -7.76107907e-01 -5.24321675e-01 -1.08601063e-01
-2.59730250e-01 -2.56193936e-01 1.81730166e-01 -1.04906964e+00
-2.22710781e-02 -4.65424478e-01 1.08571514e-01 -1.76156640e-01
2.53390908e-01 -8.71168733e-01 -1.66214034e-01 6.65696979e-01
-4.15941566e-01 1.09422177e-01 2.59258777e-01 6.55534089e-01
-2.98877470e-02 -4.77486819e-01 1.20409179e+00 9.38796904e-03
-8.91032740e-02 5.67353487e-01 1.78628162e-01 8.66375342e-02
1.23392117e+00 8.94212425e-02 -5.36838830e-01 -3.16562653e-01
-2.99764276e-01 1.20955132e-01 3.88136774e-01 2.47736171e-01
4.65046138e-01 -1.05067074e+00 -7.15344429e-01 2.43342787e-01
-1.57696471e-01 -2.87937015e-01 -1.27444090e-02 2.08644286e-01
-2.00110808e-01 2.66771406e-01 -3.99469361e-02 -1.99415177e-01
-1.57405078e+00 8.54411662e-01 6.91387355e-01 -5.13340712e-01
-3.46926838e-01 8.90329480e-01 -1.57444663e-02 -5.51418006e-01
4.09304380e-01 3.48635107e-01 -3.04944674e-03 -6.04413688e-01
7.06475079e-01 4.93985265e-01 2.84564705e-03 -1.86343163e-01
-2.50923753e-01 3.78893435e-01 -4.80688214e-01 -1.18307233e-01
1.00911057e+00 3.87088031e-01 2.35745013e-01 -2.22193524e-01
1.04492903e+00 4.44489360e-01 -1.36096823e+00 -7.80235454e-02
1.38312485e-02 -8.14083755e-01 -3.14873010e-01 -7.17259586e-01
-1.08952379e+00 8.60774577e-01 3.16543639e-01 5.42585194e-01
9.60931301e-01 -7.03245580e-01 8.64987612e-01 6.11088395e-01
3.87654543e-01 -7.48545825e-01 3.17109913e-01 2.39105716e-01
8.16556573e-01 -1.00274599e+00 -1.00694411e-01 -5.82283616e-01
-6.36838078e-01 1.03892434e+00 1.05616319e+00 -3.22157562e-01
4.64298517e-01 4.41429734e-01 1.80375725e-01 2.03938842e-01
-5.56831360e-01 4.86390471e-01 -4.95795384e-02 7.66301274e-01
-3.17666829e-01 -1.86821505e-01 -3.86156365e-02 5.38745642e-01
-1.77945867e-01 -6.26810014e-01 6.22322619e-01 8.84259284e-01
-3.37434232e-01 -1.59721255e+00 -7.24096477e-01 3.18153575e-02
-7.94259310e-01 1.64314955e-01 -7.23766565e-01 6.28969789e-01
-2.40314081e-01 1.19563019e+00 -9.36047912e-01 -7.04422295e-01
3.46526265e-01 8.55027437e-02 1.03467777e-01 -5.44145703e-01
-1.05551183e+00 -4.54708457e-01 -5.00550941e-02 -5.60779810e-01
2.77332127e-01 -3.86493355e-01 -8.95752311e-01 -2.81945974e-01
-6.57873809e-01 5.32381773e-01 6.10531509e-01 7.90596962e-01
4.51051861e-01 1.34730875e-01 1.48214424e+00 -7.15928137e-01
-1.59143782e+00 -8.79447877e-01 -1.55721366e-01 5.23817480e-01
2.10843697e-01 -6.45727038e-01 -1.06383920e+00 -2.82617152e-01] | [5.770592212677002, 7.839046478271484] |
8fd0ec76-23e2-4024-bef4-0e5f2036fb90 | whu-stereo-a-challenging-benchmark-for-stereo | 2206.02342 | null | https://arxiv.org/abs/2206.02342v1 | https://arxiv.org/pdf/2206.02342v1.pdf | WHU-Stereo: A Challenging Benchmark for Stereo Matching of High-Resolution Satellite Images | Stereo matching of high-resolution satellite images (HRSI) is still a fundamental but challenging task in the field of photogrammetry and remote sensing. Recently, deep learning (DL) methods, especially convolutional neural networks (CNNs), have demonstrated tremendous potential for stereo matching on public benchmark datasets. However, datasets for stereo matching of satellite images are scarce. To facilitate further research, this paper creates and publishes a challenging dataset, termed WHU-Stereo, for stereo matching DL network training and testing. This dataset is created by using airborne LiDAR point clouds and high-resolution stereo imageries taken from the Chinese GaoFen-7 satellite (GF-7). The WHU-Stereo dataset contains more than 1700 epipolar rectified image pairs, which cover six areas in China and includes various kinds of landscapes. We have assessed the accuracy of ground-truth disparity maps, and it is proved that our dataset achieves comparable precision compared with existing state-of-the-art stereo matching datasets. To verify its feasibility, in experiments, the hand-crafted SGM stereo matching algorithm and recent deep learning networks have been tested on the WHU-Stereo dataset. Experimental results show that deep learning networks can be well trained and achieves higher performance than hand-crafted SGM algorithm, and the dataset has great potential in remote sensing application. The WHU-Stereo dataset can serve as a challenging benchmark for stereo matching of high-resolution satellite images, and performance evaluation of deep learning models. Our dataset is available at https://github.com/Sheng029/WHU-Stereo | ['Lin Zhang', 'Wanshou Jiang', 'San Jiang', 'Sheng He', 'Shenhong Li'] | 2022-06-06 | null | null | null | null | ['stereo-matching-1'] | ['computer-vision'] | [ 1.58215612e-01 -5.18882453e-01 -1.21935541e-02 -5.45832932e-01
-7.76623547e-01 -1.70382913e-02 5.43499708e-01 -5.29134989e-01
-5.06809294e-01 6.71525240e-01 1.49062827e-01 -5.14946640e-01
-1.05375804e-01 -1.55268717e+00 -8.79364014e-01 -6.73917830e-01
1.74766071e-02 4.96710390e-01 -2.25706268e-02 -5.61609805e-01
-1.95871759e-02 5.54955661e-01 -1.82593048e+00 6.16372377e-02
1.19130576e+00 1.00355673e+00 6.08934104e-01 1.40154019e-01
1.30438153e-03 1.87622488e-01 9.89583284e-02 -1.24717876e-01
8.09928358e-01 -9.43832274e-05 -6.66232884e-01 -1.22466460e-01
1.08033776e+00 -7.59597600e-01 -7.77875006e-01 1.28994358e+00
9.36815381e-01 -1.61470652e-01 3.34968090e-01 -1.06378222e+00
-7.21545875e-01 3.21639717e-01 -7.25089014e-01 1.61146343e-01
-5.05300891e-03 1.18892089e-01 9.20453310e-01 -7.61865735e-01
3.97503048e-01 1.16706419e+00 1.10707414e+00 1.91401914e-01
-1.03137064e+00 -1.36125672e+00 -4.48908836e-01 2.19857842e-01
-1.71628010e+00 -1.08879045e-01 4.47205156e-01 -6.27665460e-01
7.96574354e-01 1.18307233e-01 8.31443489e-01 4.06830519e-01
2.06590697e-01 5.88061929e-01 1.16376424e+00 7.81048508e-03
-2.06796482e-01 -6.85164094e-01 -1.02427855e-01 4.47498322e-01
2.76494294e-01 8.46579313e-01 -1.87011302e-01 1.57312348e-01
1.20379364e+00 3.87429714e-01 -5.13415933e-01 -4.00886647e-02
-1.16524780e+00 1.01732242e+00 1.17914915e+00 1.19437680e-01
-3.50967437e-01 -7.58894384e-02 2.62330472e-01 4.31601584e-01
6.73636973e-01 -1.18556045e-01 -3.10846567e-01 4.92407084e-01
-8.70071232e-01 4.42075402e-01 8.64773020e-02 1.07131863e+00
1.28275323e+00 5.17788976e-02 2.48156756e-01 1.09417665e+00
1.01807408e-01 1.05847383e+00 3.31371754e-01 -8.66685748e-01
6.85350895e-01 7.20026672e-01 -1.72207534e-01 -1.08831167e+00
-4.19585198e-01 -8.23474377e-02 -1.54556608e+00 2.50979364e-01
9.60990880e-03 -5.72894476e-02 -7.65756905e-01 1.26795197e+00
3.02301738e-02 2.14327693e-01 5.54528050e-02 1.26855922e+00
1.43291032e+00 8.24988782e-01 -1.40686452e-01 3.60926837e-01
9.11048532e-01 -4.87821341e-01 -2.08675727e-01 -3.34388047e-01
4.15356249e-01 -6.85579121e-01 8.14545989e-01 -1.42129600e-01
-6.07819557e-01 -8.93817365e-01 -1.04249334e+00 -2.22082794e-01
-3.45651150e-01 9.37088579e-02 9.05237138e-01 6.50994526e-03
-1.17347360e+00 6.73788130e-01 -2.88134545e-01 -5.45886338e-01
6.66339695e-01 3.44997644e-01 -6.80889845e-01 -3.41302395e-01
-1.48723435e+00 7.05004156e-01 4.36070383e-01 4.40846711e-01
-5.45772910e-01 -7.65208125e-01 -1.27140939e+00 -6.13873936e-02
-2.54190564e-01 -6.68680608e-01 1.00524902e+00 -7.72765398e-01
-9.18940306e-01 1.55513573e+00 1.54014185e-01 -3.82214844e-01
4.83746797e-01 -2.15796148e-03 -4.78402942e-01 -3.24743271e-01
5.32693982e-01 9.43270922e-01 1.87275857e-01 -9.51982260e-01
-1.02263367e+00 -6.86392546e-01 -9.66308266e-02 1.95850328e-01
1.29836127e-02 -1.81043372e-01 -1.85433149e-01 -3.83762866e-01
5.46699762e-01 -7.42091835e-01 -2.28178352e-01 1.34213626e-01
-8.64079446e-02 -8.07954185e-03 7.41851747e-01 -6.22588933e-01
6.86685443e-01 -2.06093907e+00 -3.56032997e-01 6.28840476e-02
-1.12196058e-01 4.78735626e-01 -4.27897573e-01 2.52620131e-01
-2.82729506e-01 1.57994255e-02 -4.35045421e-01 1.82423115e-01
-2.61919826e-01 3.33474934e-01 -4.99947906e-01 7.49218524e-01
-1.22208305e-01 1.07993233e+00 -7.26405144e-01 -4.71749574e-01
6.72046781e-01 3.21697742e-01 -1.33396700e-01 3.45945746e-01
1.16784282e-01 5.93704522e-01 -2.92239487e-01 7.82663882e-01
1.44199264e+00 1.00802317e-01 -2.73474753e-01 -4.37206358e-01
-4.54264849e-01 -2.21476331e-02 -1.06192160e+00 1.63307512e+00
-5.00197530e-01 9.48390126e-01 -1.24236710e-01 -8.24924767e-01
1.47717428e+00 -1.68172657e-01 3.89483094e-01 -1.51868200e+00
2.50042826e-02 5.15446246e-01 -3.12162310e-01 -4.27210003e-01
5.05206227e-01 -1.42911613e-01 8.56675506e-02 -2.15768442e-02
-3.04855436e-01 -5.65057337e-01 -1.23795187e-02 -4.46351498e-01
3.50601017e-01 1.06590435e-01 1.86577365e-01 -5.01959682e-01
6.97656691e-01 3.04247737e-01 7.91876376e-01 5.24642825e-01
9.00314748e-03 1.01763058e+00 -3.31588984e-01 -1.13205719e+00
-1.45534611e+00 -8.47716928e-01 -7.84617662e-01 5.28019905e-01
5.41384518e-01 2.10935503e-01 -1.77594081e-01 1.19285107e-01
4.67990786e-01 7.81319663e-03 -3.79109442e-01 2.63135999e-01
-5.65619409e-01 -8.76442671e-01 5.76443911e-01 6.58108234e-01
1.52004576e+00 -1.05429220e+00 -2.18751907e-01 5.26002571e-02
-3.21060985e-01 -9.74682033e-01 -1.79733068e-01 -4.44714248e-01
-1.07929182e+00 -1.33225000e+00 -8.04241180e-01 -9.68533933e-01
3.58214378e-01 9.05322731e-01 1.25123549e+00 1.17828302e-01
-2.14755833e-01 -3.64818811e-01 1.66143011e-02 -1.96336284e-01
-9.96874049e-02 8.54147002e-02 -2.37581119e-01 -3.74664485e-01
7.90625811e-01 -8.07409346e-01 -6.59138262e-01 6.68373644e-01
-8.61445189e-01 4.29591209e-01 5.34417689e-01 9.83424306e-01
7.80857623e-01 1.14634931e-01 -1.30091146e-01 -5.31572640e-01
1.26120999e-01 -2.47396097e-01 -1.30572867e+00 -9.87937823e-02
-3.73112947e-01 -3.80839616e-01 3.40585709e-01 3.00872475e-01
-9.28785563e-01 4.66544516e-02 -4.84577745e-01 -2.15809986e-01
-2.74658114e-01 7.77693093e-01 -8.17569569e-02 -4.68922496e-01
5.73095262e-01 4.61558938e-01 9.72287357e-02 -4.89350945e-01
-1.79156885e-02 1.16556275e+00 9.48499382e-01 -3.53615344e-01
1.05322528e+00 7.18360126e-01 1.27066612e-01 -9.26266909e-01
-7.94046938e-01 -5.85138023e-01 -7.14319408e-01 -1.63590446e-01
6.91168785e-01 -1.75961745e+00 -6.38356686e-01 1.07738423e+00
-9.87658322e-01 -5.85257411e-01 2.17102006e-01 4.40168232e-01
-4.05888259e-01 4.41062987e-01 -4.95949149e-01 -3.11137497e-01
-6.44476831e-01 -1.02171278e+00 1.51535535e+00 4.97647166e-01
3.39688450e-01 -7.73679972e-01 1.47132456e-01 4.31253791e-01
5.67932963e-01 3.93384844e-01 5.07530153e-01 3.98695543e-02
-8.68594110e-01 -1.49036542e-01 -9.06544030e-01 3.62947792e-01
8.42821747e-02 -3.95759158e-02 -9.74363208e-01 -3.87962103e-01
-4.09331352e-01 -4.09503013e-01 1.15009284e+00 6.04811013e-01
1.24801791e+00 2.15384483e-01 -9.78253037e-02 1.43948591e+00
1.91892946e+00 5.29590473e-02 1.23111081e+00 1.06593454e+00
7.85175622e-01 3.84656578e-01 7.83755600e-01 3.54083866e-01
7.04524696e-01 5.67473054e-01 7.08837807e-01 -6.52333915e-01
1.42804548e-01 -3.33255768e-01 -1.04288831e-01 2.71325797e-01
-4.15554941e-01 4.09420669e-01 -1.23194742e+00 4.89042222e-01
-1.78589463e+00 -1.16343284e+00 -6.06048584e-01 2.23419404e+00
4.21243995e-01 -4.34236348e-01 -3.91000569e-01 1.18613385e-01
9.37035322e-01 4.90131795e-01 -4.39303368e-01 3.06014210e-01
-6.33006692e-01 2.42405310e-01 1.05957592e+00 5.04476964e-01
-1.52770722e+00 1.14253557e+00 5.36417007e+00 6.62703574e-01
-1.43275094e+00 -1.04618795e-01 3.91786397e-01 5.46097934e-01
-2.23628372e-01 -1.81881294e-01 -8.84982049e-01 4.56580281e-01
3.80952030e-01 -2.22777665e-01 2.41086498e-01 8.10358107e-01
3.16823006e-01 -8.37088153e-02 -6.24956191e-01 1.35004258e+00
-2.38730520e-01 -1.62379456e+00 1.42261013e-01 2.11955741e-01
1.00483298e+00 7.93492079e-01 -9.23637673e-02 2.74796963e-01
4.65474933e-01 -1.08536172e+00 1.76786110e-01 3.18562895e-01
1.32081831e+00 -5.87124169e-01 1.28606415e+00 3.19572777e-01
-1.52563035e+00 -7.91891664e-02 -1.12204611e+00 -3.77232105e-01
1.36894165e-02 6.78357840e-01 -1.46615028e-01 8.37299228e-01
1.17095613e+00 1.37079000e+00 -4.29838777e-01 1.22043204e+00
-2.29820296e-01 3.67206596e-02 -2.08748490e-01 5.64129174e-01
4.13825512e-01 -7.24108160e-01 1.38657792e-02 8.89838517e-01
7.14766979e-01 2.54899204e-01 1.08637691e-01 8.69975209e-01
-2.63148751e-02 7.86769092e-02 -1.12162757e+00 3.76511365e-01
5.44348538e-01 1.14034796e+00 1.61918133e-01 -1.86620176e-01
-4.56359774e-01 4.87050325e-01 4.83649410e-03 1.65796518e-01
-6.41065240e-01 -8.53107944e-02 1.00700307e+00 -5.76873869e-03
-3.51939872e-02 -2.14300871e-01 -3.40113252e-01 -1.32559693e+00
-1.57953307e-01 -8.05450082e-01 4.37291443e-01 -1.33160806e+00
-1.36408138e+00 5.22071123e-01 -1.88594759e-01 -1.76511025e+00
3.99513319e-02 -6.07006669e-01 -8.40465248e-01 1.41349435e+00
-2.24776769e+00 -1.37463152e+00 -1.36739516e+00 7.35062838e-01
2.97719538e-01 -4.34108973e-01 6.14593685e-01 5.54434001e-01
-3.82955104e-01 2.61643901e-02 4.53387946e-01 4.65555429e-01
5.02153218e-01 -7.96647787e-01 9.25437748e-01 7.39573419e-01
-3.47192228e-01 1.14821292e-01 2.77075052e-01 -4.66068506e-01
-1.17078459e+00 -1.57796037e+00 1.09816492e+00 2.02318311e-01
4.44139987e-01 1.55366704e-01 -9.25096273e-01 6.35049999e-01
-1.63358957e-01 8.99441391e-02 3.55382323e-01 -2.63487875e-01
-1.71364650e-01 -6.27818108e-01 -1.05317056e+00 3.81844372e-01
1.29641235e+00 -7.18915284e-01 -5.52830517e-01 4.42548066e-01
3.73310953e-01 -6.87178195e-01 -8.59868705e-01 1.05862606e+00
6.17764235e-01 -1.39717877e+00 1.12169755e+00 -1.45539686e-01
8.35597456e-01 -4.76257235e-01 -5.36193371e-01 -1.26412261e+00
-5.18379509e-01 1.96239755e-01 8.78000379e-01 9.14657950e-01
-8.88274238e-02 -8.15781355e-01 7.07242072e-01 1.82758361e-01
-2.46285185e-01 -3.14546436e-01 -8.80418479e-01 -8.82901251e-01
3.00703377e-01 -2.67710090e-01 1.12161493e+00 1.07101452e+00
-7.44176924e-01 2.57493816e-02 -3.74274522e-01 5.12056053e-01
7.99806356e-01 6.98944747e-01 1.27197003e+00 -1.55789292e+00
3.61143023e-01 -4.96910840e-01 -6.61517143e-01 -1.12681663e+00
3.88028473e-01 -9.66942549e-01 -1.61796823e-01 -1.65940261e+00
3.32528919e-01 -5.10192573e-01 2.68810362e-01 4.62781638e-01
1.32177006e-02 5.75153708e-01 -1.42785072e-01 5.54479301e-01
3.30997199e-01 7.81650007e-01 1.39584315e+00 -5.58933496e-01
1.71487629e-01 1.30986292e-02 -1.95665568e-01 4.96677756e-01
8.45002592e-01 8.66223499e-03 1.55847386e-01 -9.70029593e-01
1.69947490e-01 6.04584962e-02 6.45548761e-01 -1.21816146e+00
2.46602401e-01 -2.07710922e-01 3.93208474e-01 -1.19439960e+00
-4.59793657e-02 -9.23666239e-01 6.78579032e-01 5.31795621e-01
7.68838897e-02 -3.26091191e-03 2.44478151e-01 -1.44436657e-01
-7.46184111e-01 2.40268707e-01 9.91507530e-01 -2.31059790e-01
-1.51434267e+00 1.04710960e+00 2.66507655e-01 -2.38652885e-01
5.96184731e-01 -4.86942708e-01 -5.13274193e-01 -3.11304331e-01
-7.84172118e-02 6.40297294e-01 6.30483091e-01 4.01179105e-01
7.01650381e-01 -1.65557528e+00 -1.18123436e+00 6.52130842e-01
3.75682712e-01 5.94995201e-01 4.63105172e-01 4.24870133e-01
-1.18702209e+00 4.98287678e-01 -8.45590711e-01 -1.01578915e+00
-1.11324275e+00 2.48445645e-02 8.74089122e-01 2.43890420e-01
-8.17337990e-01 5.98622739e-01 2.67248869e-01 -1.04492509e+00
-1.49489298e-01 -1.05757132e-01 -2.01867640e-01 -1.62504539e-01
3.96163255e-01 1.75896883e-01 1.19952084e-02 -1.02556169e+00
-1.54210225e-01 1.18930519e+00 4.72917050e-01 4.09018487e-01
1.65719104e+00 -1.02466583e-01 -4.11007613e-01 -1.07043251e-01
1.29546976e+00 -6.97267652e-01 -1.12201786e+00 -5.89785397e-01
-4.56718802e-01 -1.08844352e+00 3.54224235e-01 -1.55100361e-01
-1.61318946e+00 1.20507884e+00 9.03578997e-01 -2.99190819e-01
1.16265142e+00 -2.65280485e-01 1.07816291e+00 4.77997720e-01
7.65850604e-01 -7.92029798e-01 -6.78817630e-01 8.69409323e-01
9.68609691e-01 -1.78869963e+00 -2.78684627e-02 -2.92879939e-01
-2.15115294e-01 1.29276466e+00 9.00151908e-01 -2.77998388e-01
5.65286994e-01 4.38226014e-02 3.75935942e-01 -2.09820986e-01
4.23556194e-02 -7.29692757e-01 1.32316858e-01 6.49603963e-01
4.38541144e-01 2.05571726e-01 -2.54481375e-01 -2.41388250e-02
-4.82941419e-01 3.68974358e-01 1.29676536e-01 5.82905173e-01
-5.55344284e-01 -8.69293511e-01 -4.54024702e-01 4.64938015e-01
2.19058335e-01 -3.98625821e-01 8.43179785e-03 1.00663853e+00
1.88892663e-01 4.51417565e-01 5.43999016e-01 -6.03012383e-01
4.34774339e-01 -6.40509248e-01 1.54802904e-01 -1.72331482e-01
-1.21297561e-01 -3.45106810e-01 -5.43969534e-02 -6.58399343e-01
-6.34131372e-01 -4.40203071e-01 -7.73713052e-01 -8.79409134e-01
-5.17074130e-02 -1.70067370e-01 4.45160002e-01 8.49743605e-01
7.64945224e-02 -1.36148334e-01 9.28316057e-01 -1.32199287e+00
-3.11906785e-01 -8.65536749e-01 -9.25739944e-01 4.75959301e-01
1.66210055e-01 -6.01428509e-01 -2.67172337e-01 -2.68957794e-01] | [8.876594543457031, -2.260822057723999] |
f3eb9c2c-21f4-495c-9c6b-74d3d205752b | dual-flow-fusion-model-for-concrete-surface | 2305.05132 | null | https://arxiv.org/abs/2305.05132v2 | https://arxiv.org/pdf/2305.05132v2.pdf | Dual flow fusion model for concrete surface crack segmentation | The existence of cracks and other damages pose a significant threat to the safe operation of transportation infrastructure. Traditional manual detection and ultrasound equipment testing consume a lot of time and resources. With the development of deep learning technology, many deep learning models have been widely applied to practical visual segmentation tasks. The detection method based on deep learning models has the advantages of high detection accuracy, fast detection speed, and simple operation. However, deep learning-based crack segmentation models are sensitive to background noise, have rough edges, and lack robustness. Therefore, this paper proposes a crack segmentation model based on the fusion of dual streams. The image is inputted simultaneously into two designed processing streams to independently extract long-distance dependence and local detail features. The adaptive prediction is achieved through the dual-headed mechanism. Meanwhile, a novel interaction fusion mechanism is proposed to guide the complementary of different feature layers to achieve crack location and recognition in complex backgrounds. Finally, an edge optimization method is proposed to improve the accuracy of segmentation. Experiments show that the F1 value of segmentation results on the DeepCrack[1] public dataset is 93.7% and the IOU value is 86.6%. The F1 value of segmentation results on the CRACK500[2] dataset is 78.1%, and the IOU value is 66.0%. | ['Yuwei Duan'] | 2023-05-09 | null | null | null | null | ['crack-segmentation'] | ['computer-vision'] | [-1.00893132e-01 -5.14067948e-01 -3.27759087e-02 -1.75796196e-01
-5.23666263e-01 6.26416579e-02 -7.88444653e-02 1.12218158e-02
-5.19252956e-01 1.72779739e-01 -4.12064791e-01 -6.78505003e-02
2.85327341e-02 -1.12776637e+00 -3.43339026e-01 -1.13133717e+00
2.45510668e-01 6.62891939e-02 8.29569936e-01 1.24638779e-02
5.53876877e-01 6.07409775e-01 -1.45539296e+00 2.15477705e-01
1.06243372e+00 1.24825156e+00 3.31449866e-01 6.10483289e-01
-1.60837650e-01 4.95766938e-01 -4.55728382e-01 6.35753646e-02
-6.21238304e-03 -1.00345962e-01 -3.84842157e-01 8.24952796e-02
-8.81378651e-02 -5.20286620e-01 -4.75323677e-01 1.08258104e+00
7.94821084e-01 2.76209787e-02 7.60501266e-01 -9.86165524e-01
-4.87480104e-01 3.69028479e-01 -1.19031513e+00 3.93417150e-01
-4.80911918e-02 2.07076892e-01 6.54632926e-01 -9.94694293e-01
1.65314823e-01 9.49531794e-01 7.44734704e-01 2.60178000e-01
-6.34824574e-01 -7.32977092e-01 -1.41991302e-01 5.47520101e-01
-1.32430828e+00 -1.11043630e-02 1.00260067e+00 -5.46840727e-01
5.61852098e-01 -4.11835760e-02 5.79144359e-01 4.84662175e-01
3.50696027e-01 8.53676498e-01 7.02468753e-01 -4.19246256e-01
-6.54275119e-02 -3.49548161e-01 4.06894624e-01 1.04863524e+00
5.27602494e-01 1.37500957e-01 1.54210236e-02 5.35205126e-01
8.42973351e-01 3.49004388e-01 -2.36455813e-01 2.30297700e-01
-7.70906448e-01 8.24899256e-01 7.17758715e-01 4.93396878e-01
-1.20764263e-01 1.79113429e-02 4.23489064e-01 -1.22621566e-01
2.08271220e-01 -6.71145618e-02 -4.88942228e-02 1.39824301e-01
-8.17177892e-01 -3.02583352e-02 2.47176617e-01 4.72966433e-01
7.25010574e-01 2.08376214e-01 3.75366546e-02 1.05161667e+00
4.91043776e-01 7.18752444e-01 2.91375846e-01 -8.15850317e-01
3.75455201e-01 7.58705318e-01 -2.36876503e-01 -1.51385427e+00
-7.70448267e-01 -4.28401858e-01 -9.57038105e-01 3.82873029e-01
4.32296634e-01 -3.46939087e-01 -1.16559899e+00 8.61554742e-01
3.19002122e-01 -7.37478361e-02 -1.79089189e-01 1.01181555e+00
1.04283798e+00 9.05095279e-01 -1.78226143e-01 -2.71829832e-02
1.15015233e+00 -9.62601066e-01 -7.72093475e-01 -2.65183210e-01
3.95125151e-01 -8.06293190e-01 9.90810812e-01 5.00278652e-01
-8.47243011e-01 -7.12248385e-01 -1.30603647e+00 -1.06756255e-01
-1.56646505e-01 4.61126804e-01 3.58325332e-01 6.85699105e-01
-4.12264764e-01 1.25266597e-01 -8.10243189e-01 2.73876358e-02
6.84622169e-01 1.78999051e-01 3.25949234e-03 -5.62927485e-01
-9.92956579e-01 6.17009878e-01 5.00055373e-01 5.81170499e-01
-5.77236354e-01 -2.51983374e-01 -7.33179808e-01 5.89662455e-02
2.71517128e-01 3.87545638e-02 8.92635167e-01 -3.04057896e-01
-1.01250494e+00 5.93104839e-01 1.02940544e-01 7.64256343e-02
4.01881158e-01 -1.40436947e-01 -3.82213354e-01 5.54598927e-01
1.93330780e-01 2.58808970e-01 7.06641495e-01 -1.33648622e+00
-1.00924170e+00 -3.87059629e-01 -2.65019029e-01 1.76347978e-02
-3.55618447e-01 -5.53071462e-02 -5.69853365e-01 -3.78648788e-01
4.66064572e-01 -4.43824708e-01 -1.94699839e-01 1.56393461e-02
-4.69781578e-01 -2.40352407e-01 1.40819347e+00 -9.09899592e-01
1.40095270e+00 -2.08933735e+00 -1.10498883e-01 4.10980940e-01
2.98459977e-01 3.47669065e-01 1.75999537e-01 8.81649405e-02
1.67523131e-01 3.56633030e-02 -6.95670724e-01 1.77335404e-02
-4.77431446e-01 9.56533551e-02 4.21944290e-01 5.87054789e-01
3.09224557e-02 5.12561321e-01 -5.92480361e-01 -8.94469798e-01
2.76420683e-01 2.49785408e-01 -3.98809493e-01 1.95359617e-01
2.82111198e-01 2.09119990e-01 -5.07276654e-01 1.10425234e+00
9.14701283e-01 2.93768942e-02 -5.45393050e-01 -4.64404941e-01
-3.59903872e-01 -7.13593900e-01 -1.31767070e+00 1.19202387e+00
-2.86139011e-01 6.45496786e-01 2.75053501e-01 -1.13316751e+00
1.18565726e+00 1.32211134e-01 5.26046097e-01 -8.75295281e-01
7.12154448e-01 4.59725171e-01 2.67878547e-02 -1.24280202e+00
1.48532569e-01 -1.41124576e-01 5.46636917e-02 3.37780684e-01
-4.39994872e-01 -1.22369207e-01 2.35878050e-01 -8.79895836e-02
7.03288972e-01 -3.59397143e-01 -4.36558276e-01 -1.16431519e-01
5.41451395e-01 -6.63824566e-03 7.02489078e-01 4.31857795e-01
-2.10180476e-01 8.85623276e-01 1.60032526e-01 -5.02613664e-01
-7.95194924e-01 -7.49733090e-01 -3.37614447e-01 4.75566417e-01
7.71412790e-01 2.24672347e-01 -8.06163549e-01 -4.01284724e-01
-1.06020153e-01 2.84097344e-01 -5.04594803e-01 -4.92876321e-01
-7.79355049e-01 -1.00902474e+00 5.71730852e-01 1.02062857e+00
1.07239163e+00 -1.03553462e+00 -8.16998541e-01 1.52319744e-01
-3.13455462e-01 -1.05192876e+00 -2.89712250e-01 -2.21833289e-01
-6.84491158e-01 -1.38942921e+00 -7.68925369e-01 -1.20223820e+00
6.19730234e-01 3.66411418e-01 4.16116685e-01 8.14534843e-01
-7.10092247e-01 2.69386973e-02 -4.25753176e-01 -2.85473734e-01
-7.82774109e-03 6.47164211e-02 -2.66481370e-01 1.60915583e-01
7.05780983e-02 -4.20097224e-02 -8.51135850e-01 2.61193126e-01
-9.55665946e-01 -3.86561155e-02 6.90548897e-01 9.94125605e-01
4.16271955e-01 6.17756069e-01 5.40548086e-01 -4.44742113e-01
5.53187132e-01 -3.11123192e-01 -4.79004711e-01 2.34550443e-02
-7.24114060e-01 -4.76867348e-01 3.25424731e-01 -2.32442215e-01
-1.32058418e+00 3.79296741e-03 -4.51719433e-01 -2.70163327e-01
-1.60024479e-01 6.83853805e-01 -2.11824417e-01 -9.55789834e-02
4.53585625e-01 1.72005057e-01 1.15335748e-01 -4.23579603e-01
-4.68126610e-02 9.44261014e-01 6.43214881e-01 -5.13200343e-01
6.69465065e-01 4.52178687e-01 -7.29670599e-02 -1.11984861e+00
-5.77807665e-01 -4.80098337e-01 -5.78426659e-01 -7.33644962e-01
1.34409201e+00 -5.38029730e-01 -7.21872330e-01 1.33415246e+00
-1.19422925e+00 -1.92089230e-01 2.86628723e-01 5.25997102e-01
2.71456037e-02 7.43389964e-01 -9.57635581e-01 -7.41058588e-01
-4.60565537e-01 -1.43842053e+00 7.80315816e-01 8.64986837e-01
5.91162503e-01 -8.62339675e-01 -3.86314154e-01 6.66428506e-01
2.13263243e-01 3.36912692e-01 7.97871768e-01 -6.79990426e-02
-4.98219967e-01 -5.86190641e-01 -7.02847540e-01 5.95565736e-01
1.11919142e-01 4.81322557e-01 -8.03389072e-01 -5.22894301e-02
-1.32549433e-02 -1.21400133e-01 1.09996629e+00 5.67823708e-01
1.26820099e+00 3.83675903e-01 -4.73810941e-01 3.73341531e-01
1.55131984e+00 7.29847550e-01 7.65353024e-01 2.42745757e-01
1.10119879e+00 5.41268349e-01 7.27317691e-01 2.00106621e-01
2.67009854e-01 2.20011488e-01 7.09500015e-01 -4.53465492e-01
-5.10104746e-02 2.26164654e-01 -6.35577664e-02 7.11360157e-01
-2.20983967e-01 -6.76383004e-02 -1.32278824e+00 6.48158967e-01
-1.61373115e+00 -1.04779184e+00 -8.58485103e-01 1.67117548e+00
4.98671800e-01 3.84153336e-01 -1.28807008e-01 6.57694101e-01
9.94402528e-01 -1.27290606e-01 -5.30309081e-01 -1.60699353e-01
-1.46807218e-02 -9.59576937e-05 4.18283850e-01 5.16706288e-01
-1.22744989e+00 6.37696147e-01 5.37203455e+00 9.75697875e-01
-1.31816769e+00 -2.57015247e-02 7.31490672e-01 3.83218050e-01
2.63995230e-02 -3.07456672e-01 -6.20781720e-01 6.74712420e-01
4.36616018e-02 4.92347479e-01 -1.31512508e-01 5.18649876e-01
1.78533867e-01 -3.84151727e-01 -3.06743741e-01 1.08693027e+00
6.57865480e-02 -1.28399050e+00 -4.81322378e-01 -1.15327440e-01
5.84995747e-01 -2.06399828e-01 4.83240858e-02 -9.41469986e-03
-2.56460667e-01 -9.75327790e-01 6.87948525e-01 5.42202950e-01
7.88208485e-01 -9.20659065e-01 1.08137453e+00 5.21812916e-01
-1.33125579e+00 -5.64390004e-01 -3.39164048e-01 -2.65033096e-02
4.24575925e-01 9.13253129e-01 -3.74288678e-01 5.39124608e-01
9.99335229e-01 7.46401131e-01 -3.21732640e-01 1.34143043e+00
-2.70984590e-01 7.21545875e-01 -4.13936794e-01 5.47728911e-02
3.01734179e-01 -3.00643384e-01 1.90174133e-01 1.39826810e+00
2.93628812e-01 1.56736180e-01 4.36039597e-01 5.24946749e-01
1.11581907e-01 1.07211798e-01 -2.92552948e-01 5.17761052e-01
5.08076787e-01 1.27935195e+00 -1.06109071e+00 -1.79379851e-01
-4.12707567e-01 5.05572081e-01 -7.60144442e-02 1.62329406e-01
-1.04335141e+00 -9.07965481e-01 9.54078734e-02 1.00258797e-01
2.61985213e-01 -2.63255626e-01 -7.90875316e-01 -7.85038531e-01
-1.19787984e-01 -4.89736766e-01 3.84839475e-01 -6.55687511e-01
-1.12709141e+00 2.72832155e-01 -1.79026768e-01 -9.94005442e-01
5.55927396e-01 -7.58499265e-01 -1.13671470e+00 6.25736833e-01
-1.44351804e+00 -1.14023924e+00 -6.72493637e-01 4.35286582e-01
7.45635509e-01 4.70446842e-03 2.55371593e-02 7.07477868e-01
-1.28491306e+00 5.31433761e-01 8.37693065e-02 6.52933180e-01
3.69725823e-01 -9.18216050e-01 -1.66459724e-01 1.30769944e+00
-4.82744247e-01 1.63625941e-01 3.37026387e-01 -6.29085541e-01
-8.98240864e-01 -9.24407542e-01 3.31284940e-01 2.18444452e-01
1.40537262e-01 1.10560805e-01 -1.30784667e+00 1.48887798e-01
1.76836457e-02 8.21145177e-02 2.02982083e-01 -3.15156400e-01
-2.35669147e-02 -2.37400115e-01 -1.25913596e+00 2.79781520e-01
4.35456216e-01 -2.09810510e-01 -3.38458717e-01 1.15670763e-01
5.15661776e-01 -6.09463871e-01 -9.34209466e-01 5.45782804e-01
7.69121528e-01 -1.10182965e+00 9.76107895e-01 2.68881395e-02
5.46125054e-01 -3.38083506e-01 -1.83769204e-02 -7.55203307e-01
-3.83393943e-01 2.34658092e-01 1.43839478e-01 1.08947909e+00
3.54250133e-01 -5.13490498e-01 8.06173623e-01 4.84643340e-01
-6.53225720e-01 -1.13378680e+00 -8.34099829e-01 -2.73799509e-01
3.18989679e-02 -5.35622239e-01 3.94945741e-01 7.59403944e-01
-3.63481551e-01 3.20561051e-01 -8.29437524e-02 3.62499356e-01
6.72119200e-01 1.60544574e-01 3.33982140e-01 -1.55242240e+00
2.95044184e-01 -5.27709186e-01 -1.48107693e-01 -9.39167678e-01
-2.32376873e-01 -3.92722905e-01 2.13905081e-01 -1.89037681e+00
4.26335372e-02 -7.14403868e-01 -2.68586934e-01 4.52587128e-01
-3.28577667e-01 3.01113188e-01 -7.92057440e-02 1.84423149e-01
-1.53862193e-01 4.34604317e-01 1.64819062e+00 -3.65145892e-01
1.45245427e-02 1.60507485e-01 -2.99485326e-01 8.26021731e-01
1.08774853e+00 -2.77524531e-01 -1.92611217e-01 -6.92329109e-01
-5.78705706e-02 1.16330832e-01 2.05090910e-01 -1.34583271e+00
5.67383170e-01 -9.58990008e-02 5.07460296e-01 -7.87564099e-01
1.91284522e-01 -8.04751694e-01 -4.62125242e-01 9.57970440e-01
2.37974524e-01 -1.17107689e-01 3.60262357e-02 6.27331197e-01
-2.58021325e-01 -5.47047317e-01 1.08348322e+00 3.65737416e-02
-9.43850994e-01 3.83934438e-01 -5.40410399e-01 -5.45106418e-02
1.15434992e+00 -6.34763896e-01 -2.14476421e-01 -5.21251336e-02
-4.68916267e-01 3.96094143e-01 2.68585563e-01 2.20177397e-01
1.13331509e+00 -1.03877962e+00 -6.79881215e-01 4.04816926e-01
-1.09326869e-01 4.31862801e-01 6.97591841e-01 8.94480526e-01
-1.07304609e+00 -2.05597863e-01 -2.29302034e-01 -8.77569139e-01
-1.23098719e+00 2.28841379e-01 5.06171644e-01 1.28181302e-03
-7.35508382e-01 9.40972090e-01 -1.41457319e-01 -7.08077401e-02
5.50482459e-02 -3.94824713e-01 -5.34875453e-01 1.82375878e-01
4.93197531e-01 9.23708320e-01 8.38333368e-02 -7.46236563e-01
-2.47986808e-01 1.09816039e+00 -4.65136915e-02 2.17246428e-01
1.16872835e+00 -5.75424545e-02 -2.14530930e-01 5.96963689e-02
1.21820045e+00 -3.06299590e-02 -1.11698115e+00 3.74185927e-02
-2.23635539e-01 -4.73791271e-01 3.78135741e-01 -6.26641333e-01
-1.84302449e+00 1.38589525e+00 8.94112408e-01 3.32664251e-01
1.28845990e+00 -2.10127130e-01 1.34187222e+00 7.77595267e-02
2.76918914e-02 -1.20869482e+00 3.61351967e-01 4.63546515e-01
4.64032918e-01 -1.42279363e+00 -4.23030406e-02 -4.80365455e-01
-4.36355948e-01 1.33374155e+00 9.99158740e-01 -8.24886560e-02
7.18354166e-01 4.17652637e-01 3.06001842e-01 -5.54172575e-01
-6.11180626e-03 -1.05040647e-01 -3.98300365e-02 5.88629901e-01
1.13529094e-01 -1.17149137e-01 -4.16584969e-01 6.63178623e-01
5.59088171e-01 -2.57652938e-01 5.62721491e-01 9.43024695e-01
-1.07718670e+00 -6.83295786e-01 -6.85576439e-01 5.94961226e-01
-5.16430497e-01 3.69354635e-01 2.40631104e-01 7.28450954e-01
4.63066876e-01 1.20905173e+00 -2.98581161e-02 -7.17558920e-01
2.84838617e-01 -4.20946658e-01 1.08874723e-01 -3.06357324e-01
-2.31519550e-01 6.58483505e-02 -2.52736151e-01 -2.17045322e-01
-2.56539255e-01 -4.39120859e-01 -1.86512899e+00 -6.88943192e-02
-6.45497739e-01 9.84240547e-02 4.73464280e-01 9.91850436e-01
-1.14462920e-01 5.93175888e-01 7.31927931e-01 -9.07532632e-01
-1.48507446e-01 -7.47632384e-01 -6.43723011e-01 1.56090036e-01
2.73068339e-01 -9.69081819e-01 -2.43178815e-01 1.01255281e-02] | [7.496574401855469, 1.5137207508087158] |
a3360027-7e7d-4e5e-85f7-b598cab1a8c4 | cutting-music-source-separation-some-slakh-a | 1909.08494 | null | https://arxiv.org/abs/1909.08494v1 | https://arxiv.org/pdf/1909.08494v1.pdf | Cutting Music Source Separation Some Slakh: A Dataset to Study the Impact of Training Data Quality and Quantity | Music source separation performance has greatly improved in recent years with the advent of approaches based on deep learning. Such methods typically require large amounts of labelled training data, which in the case of music consist of mixtures and corresponding instrument stems. However, stems are unavailable for most commercial music, and only limited datasets have so far been released to the public. It can thus be difficult to draw conclusions when comparing various source separation methods, as the difference in performance may stem as much from better data augmentation techniques or training tricks to alleviate the limited availability of training data, as from intrinsically better model architectures and objective functions. In this paper, we present the synthesized Lakh dataset (Slakh) as a new tool for music source separation research. Slakh consists of high-quality renderings of instrumental mixtures and corresponding stems generated from the Lakh MIDI dataset (LMD) using professional-grade sample-based virtual instruments. A first version, Slakh2100, focuses on 2100 songs, resulting in 145 hours of mixtures. While not fully comparable because it is purely instrumental, this dataset contains an order of magnitude more data than MUSDB18, the {\it de facto} standard dataset in the field. We show that Slakh can be used to effectively augment existing datasets for musical instrument separation, while opening the door to a wide array of data-intensive music signal analysis tasks. | ['Prem Seetharaman', 'Gordon Wichern', 'Ethan Manilow', 'Jonathan Le Roux'] | 2019-09-18 | null | null | null | null | ['music-source-separation'] | ['music'] | [ 2.74667889e-01 -3.02892745e-01 2.15393398e-02 1.54768720e-01
-1.32539070e+00 -8.56537759e-01 3.95779610e-01 -6.97488189e-02
-2.33238578e-01 5.09023190e-01 4.84240055e-01 -6.31338283e-02
-5.35108507e-01 -3.85690659e-01 -4.13978428e-01 -6.88789546e-01
-3.62604409e-02 5.59999168e-01 -3.58526438e-01 -2.63547570e-01
-5.19542955e-02 3.72862190e-01 -1.75397778e+00 2.24614069e-01
5.77922702e-01 9.61679757e-01 1.56927869e-01 7.54463494e-01
5.06360549e-03 5.79918683e-01 -8.10464680e-01 -2.35649794e-01
5.12358069e-01 -8.05276275e-01 -5.58381140e-01 -1.44799322e-01
6.11631036e-01 -3.31317298e-02 -7.19076544e-02 7.47968614e-01
9.61426020e-01 2.54184693e-01 5.52472353e-01 -1.15058577e+00
-1.72853798e-01 1.20675480e+00 -4.22964901e-01 9.31311995e-02
1.61520168e-01 -1.02493890e-01 1.36012220e+00 -6.97862327e-01
3.57247323e-01 9.43042219e-01 8.58972132e-01 2.33529851e-01
-1.47568083e+00 -8.71851921e-01 -5.82750022e-01 1.13865416e-02
-1.28859973e+00 -9.74276304e-01 1.06352687e+00 -4.47450161e-01
6.83834016e-01 5.39807916e-01 8.09102297e-01 1.26505423e+00
-6.40229225e-01 9.26929474e-01 1.04322278e+00 -5.57960093e-01
1.24179788e-01 -1.15672834e-01 -1.41281292e-01 9.27585363e-02
2.15839781e-02 2.12671682e-01 -8.62598717e-01 -3.05539519e-01
8.24203491e-01 -5.63773096e-01 -4.12958294e-01 -3.02751094e-01
-1.51511490e+00 7.14064777e-01 3.04933954e-02 3.83824259e-01
-3.12708855e-01 -4.42774184e-02 5.70155799e-01 4.90206152e-01
3.00762594e-01 8.44649434e-01 -5.07985115e-01 -5.96012115e-01
-1.47603142e+00 7.16310203e-01 8.21589410e-01 5.71969450e-01
2.07753152e-01 6.41401172e-01 -3.93475266e-03 1.29653490e+00
3.92763540e-02 5.17636776e-01 7.11188197e-01 -1.26734579e+00
4.25144672e-01 5.31956106e-02 -1.18308663e-02 -8.60643089e-01
-3.27182978e-01 -8.85805666e-01 -8.61142397e-01 3.06607246e-01
7.58858025e-01 -1.35814309e-01 -4.65910882e-01 1.83305490e+00
-8.41568261e-02 2.99931377e-01 -1.03073895e-01 8.68069589e-01
9.34582829e-01 4.66554433e-01 -6.09731674e-01 -1.41987562e-01
1.01968062e+00 -8.93061578e-01 -5.72258413e-01 -6.28097504e-02
2.72390038e-01 -1.18581390e+00 1.17452157e+00 1.05933976e+00
-1.25735056e+00 -8.23410928e-01 -1.21628881e+00 1.23180702e-01
-5.68318442e-02 1.31628409e-01 7.94524491e-01 8.35057795e-01
-7.12759972e-01 8.77948403e-01 -4.61885661e-01 -9.77766514e-03
4.22383666e-01 3.08356196e-01 -1.95771292e-01 3.23057264e-01
-1.03371716e+00 3.02928925e-01 3.04426402e-01 -9.93630439e-02
-8.22593033e-01 -9.86217380e-01 -7.13460863e-01 6.93161264e-02
2.66676247e-01 -4.29146260e-01 1.48432553e+00 -1.05278540e+00
-1.65149534e+00 8.95904541e-01 3.16682339e-01 -5.43321908e-01
4.47545141e-01 -4.16627347e-01 -7.37917185e-01 -1.18729740e-01
1.22419633e-02 3.93403083e-01 1.10957754e+00 -1.10486126e+00
-5.43187141e-01 -2.11634561e-01 -2.28445947e-01 1.33140638e-01
-3.60512435e-01 1.18763380e-01 -2.99237460e-01 -1.21254146e+00
-4.28231359e-02 -1.00789750e+00 1.18405066e-01 -6.80696666e-01
-5.13583720e-01 2.72586763e-01 4.72255319e-01 -7.83674717e-01
1.19499338e+00 -2.42893696e+00 3.04571539e-01 2.38268152e-02
6.56067580e-02 3.27187538e-01 -4.31773424e-01 4.24506605e-01
-3.94019812e-01 -1.89252660e-01 -2.97010839e-01 -5.29757917e-01
2.77874142e-01 -1.42422274e-01 -7.92330325e-01 2.93733835e-01
-1.81691095e-01 6.74124360e-01 -6.95314467e-01 -2.13138476e-01
2.13703781e-01 4.13600475e-01 -6.10484779e-01 -6.88101128e-02
-1.38516665e-01 6.27844930e-01 2.11022645e-01 5.57003558e-01
4.70285296e-01 2.78003126e-01 -3.81381400e-02 -1.51297703e-01
1.57419480e-02 6.81827962e-01 -1.51206648e+00 2.24069858e+00
-4.10951763e-01 8.96913886e-01 3.12314153e-01 -7.84482419e-01
7.67760336e-01 5.72673202e-01 7.83725083e-01 -3.54395300e-01
2.40183473e-01 5.07305622e-01 3.38337809e-01 -2.24842969e-02
6.73322082e-01 -2.48899102e-01 -8.66714716e-02 4.45716798e-01
2.39629254e-01 -5.11578798e-01 3.87963116e-01 -6.98873252e-02
9.76530433e-01 1.48262829e-01 2.40714476e-02 -4.77738939e-02
1.39104292e-01 6.73928633e-02 4.86516148e-01 7.53226459e-01
8.18329230e-02 8.84920776e-01 1.29791573e-01 1.55949518e-01
-1.02500951e+00 -1.30471385e+00 -3.04881722e-01 1.11927843e+00
-4.29274201e-01 -7.49736488e-01 -6.86309993e-01 -6.58096373e-02
-6.06583059e-02 5.60371697e-01 -1.97245821e-01 -2.00275648e-02
-4.98300165e-01 -8.31589043e-01 1.17970920e+00 4.57193375e-01
2.01946795e-01 -1.34988081e+00 -3.70662153e-01 3.24480325e-01
-2.58721143e-01 -8.13898385e-01 -3.52326632e-01 4.36771452e-01
-8.96716654e-01 -8.90180588e-01 -7.19521523e-01 -5.60168087e-01
-4.07251418e-01 1.70933858e-01 1.45958519e+00 -4.76384073e-01
-2.91601479e-01 1.76948771e-01 -3.47476423e-01 -7.81272352e-01
-5.83753467e-01 2.18285128e-01 2.91640490e-01 -2.93219127e-02
1.22127831e-01 -1.12179101e+00 -1.98591590e-01 -1.76225393e-03
-8.41578543e-01 5.07694073e-02 6.25638783e-01 6.99671626e-01
4.33738053e-01 3.59833807e-01 9.37826335e-01 -6.78569496e-01
8.36961687e-01 -3.05407345e-01 -3.83757412e-01 -4.59142536e-01
-5.00086308e-01 -2.73687840e-01 6.26831293e-01 -4.94692385e-01
-7.84531951e-01 -1.57957286e-01 -3.11952204e-01 -5.69887638e-01
-3.52111757e-01 6.68375075e-01 -2.76482999e-01 3.73221993e-01
8.04557383e-01 6.07203357e-02 -1.59998655e-01 -1.10144901e+00
4.16950494e-01 7.75986016e-01 1.07397223e+00 -6.41608536e-01
7.56398857e-01 2.06489637e-01 -1.13734201e-01 -1.10800564e+00
-8.45560670e-01 -4.66667771e-01 -4.73271072e-01 1.32178724e-01
5.13355672e-01 -9.30949390e-01 -5.06515801e-01 6.82296991e-01
-5.22735476e-01 -5.56794763e-01 -8.16854775e-01 6.65501177e-01
-6.96552932e-01 1.33453146e-01 -7.32066989e-01 -7.23137796e-01
-2.36554861e-01 -7.82515824e-01 8.76895249e-01 -1.83585271e-01
-6.37063026e-01 -6.89214110e-01 4.39816058e-01 5.36483467e-01
2.50271469e-01 1.39314845e-01 8.87941539e-01 -6.46823764e-01
-2.18751580e-01 -3.15401226e-01 3.59848857e-01 6.39306903e-01
3.24815601e-01 -3.98985371e-02 -1.52308917e+00 -2.58734882e-01
8.25256407e-02 -4.33793247e-01 8.88216197e-01 5.56446195e-01
8.94312799e-01 4.16517071e-02 2.04504743e-01 8.36884439e-01
8.58481169e-01 7.25280792e-02 5.41097045e-01 2.79120326e-01
7.21009254e-01 3.16072732e-01 2.98007965e-01 4.94112879e-01
-1.27817720e-01 9.52973485e-01 2.02821434e-01 -1.52905524e-01
-5.09743631e-01 -2.93833882e-01 3.91729474e-01 1.32556558e+00
-2.45707482e-01 -3.38532180e-02 -6.47892535e-01 5.01323879e-01
-1.50823128e+00 -1.17221439e+00 -2.17027277e-01 2.46940756e+00
1.02792406e+00 5.04017770e-02 6.40174389e-01 1.07916832e+00
3.20779145e-01 2.31513873e-01 -4.25912648e-01 -2.35781167e-03
-5.94017446e-01 6.33334339e-01 -8.71018856e-04 2.13669062e-01
-1.23676634e+00 4.61080939e-01 6.48754168e+00 1.17926025e+00
-1.05458307e+00 -5.27153052e-02 3.02410219e-03 -6.57959938e-01
-8.53886977e-02 -1.74990952e-01 -3.92040074e-01 4.54885513e-01
1.06407976e+00 -5.07282242e-02 7.99846411e-01 5.74771583e-01
1.09988891e-01 9.98301804e-02 -1.19814932e+00 1.57666457e+00
4.57056537e-02 -1.25068271e+00 -1.77707538e-01 1.49284288e-01
5.83754539e-01 1.24109462e-01 3.73468906e-01 4.81573820e-01
8.46624225e-02 -1.18557739e+00 1.10877633e+00 2.47735217e-01
7.77891278e-01 -9.96448755e-01 3.77887726e-01 4.78850871e-01
-1.04147661e+00 -5.49632348e-02 -1.45016462e-01 -2.54160970e-01
1.45519665e-02 6.39822483e-01 -6.05336905e-01 7.41390169e-01
7.27608800e-01 8.19458127e-01 -5.43380737e-01 1.11203361e+00
1.22210637e-01 1.11151147e+00 -3.58589023e-01 5.99221647e-01
-1.02387674e-01 -4.25585449e-01 9.75708246e-01 1.10450768e+00
3.98291558e-01 -4.42732424e-01 5.20273149e-02 7.67738998e-01
-1.77733436e-01 3.18655521e-01 -4.75215912e-01 -4.09397513e-01
4.12816405e-01 1.40407753e+00 -6.05696380e-01 -1.87650964e-01
-2.09326804e-01 7.53312886e-01 -4.38348912e-02 1.97270170e-01
-6.87497497e-01 -2.47145787e-01 8.16123962e-01 6.87995702e-02
1.15231186e-01 -2.18934253e-01 -3.71542245e-01 -1.09175026e+00
-2.49917388e-01 -1.59746695e+00 3.06364834e-01 -7.89870262e-01
-1.33724630e+00 6.13777936e-01 -3.01940590e-01 -1.40312803e+00
-4.74927694e-01 -4.82595325e-01 -3.12247097e-01 9.38875020e-01
-8.76457989e-01 -8.92610848e-01 -7.89901540e-02 5.21045864e-01
5.44431925e-01 -5.76886773e-01 1.18557656e+00 6.49047494e-01
-4.11766350e-01 4.70777780e-01 3.76863122e-01 1.81822315e-01
1.02529299e+00 -1.25454521e+00 2.27614000e-01 6.05043113e-01
1.20326805e+00 3.85100424e-01 6.90135717e-01 -1.37614384e-01
-1.19002926e+00 -7.37774849e-01 2.94359595e-01 -7.41501153e-01
6.33471251e-01 -4.30999041e-01 -7.53008306e-01 5.06616712e-01
1.51049212e-01 -4.28435475e-01 1.06269073e+00 4.16520149e-01
-3.55389088e-01 -1.40028417e-01 -5.82384288e-01 6.40268922e-01
9.51810539e-01 -6.85108483e-01 -5.80985248e-01 -9.07964036e-02
4.50518191e-01 -2.96247810e-01 -7.64501214e-01 5.26699066e-01
7.18834102e-01 -1.27550530e+00 1.24418283e+00 -4.82466400e-01
3.64009649e-01 -3.76647413e-01 -4.06652451e-01 -1.56458974e+00
-3.94518018e-01 -7.81052053e-01 -2.89441645e-01 1.69493365e+00
3.47520053e-01 -1.59889013e-01 6.59521520e-01 2.71578785e-02
-2.27785632e-01 -3.52731407e-01 -6.71906948e-01 -9.87757027e-01
1.38418883e-01 -1.01628971e+00 7.55200922e-01 1.19627488e+00
-2.20875263e-01 6.71779394e-01 -5.21078408e-01 -3.14834982e-01
5.80282927e-01 5.14879525e-01 1.17915833e+00 -1.54007006e+00
-1.00277007e+00 -8.82954180e-01 -1.88394651e-01 -7.46202826e-01
1.37917250e-01 -1.21523726e+00 -1.64085135e-01 -1.03458059e+00
1.04462076e-02 -6.31524205e-01 -4.47014958e-01 4.27203864e-01
3.55559811e-02 6.63369954e-01 4.05981451e-01 3.94568026e-01
-8.57466981e-02 4.89446253e-01 9.31807041e-01 -2.74537116e-01
-5.01057446e-01 3.01591009e-01 -7.87668049e-01 1.08198762e+00
7.87420332e-01 -3.25662613e-01 -5.50074756e-01 -3.10543865e-01
2.84099728e-01 5.56644909e-02 3.81496817e-01 -1.37740004e+00
-2.67801195e-01 2.85375386e-01 4.52597708e-01 -5.43972015e-01
7.70226896e-01 -5.22528827e-01 5.73519588e-01 -2.08061486e-02
-3.89695525e-01 -4.28286791e-01 4.00028169e-01 3.02262366e-01
-4.57099557e-01 -1.07122861e-01 6.09792888e-01 3.10265124e-02
-4.07298386e-01 7.07188845e-02 -2.50013322e-01 3.88572097e-01
1.62097916e-01 -1.26423612e-01 1.58783317e-01 -6.34998918e-01
-7.55872190e-01 -6.52752459e-01 3.96803707e-01 5.48916161e-01
1.39004603e-01 -1.48170257e+00 -9.33557689e-01 2.94515163e-01
1.54865962e-02 -1.42486528e-01 2.87070900e-01 8.24219346e-01
-1.76465988e-01 2.38208219e-01 -1.66773334e-01 -3.56028318e-01
-1.27066112e+00 5.59925079e-01 6.93330467e-02 -9.59204137e-02
-7.83703804e-01 9.31174815e-01 9.64337513e-02 -4.03750092e-01
3.22059035e-01 -1.59835055e-01 -1.08190812e-02 3.78563643e-01
5.14050186e-01 6.06074095e-01 1.85806304e-01 -6.95110679e-01
-1.06132835e-01 2.41702482e-01 4.80564326e-01 -6.37777030e-01
1.44370604e+00 1.68998778e-01 -8.82235914e-03 1.03762507e+00
8.12474608e-01 8.45887125e-01 -9.57394838e-01 -2.03643993e-01
-4.65501994e-02 -4.80713874e-01 7.68905953e-02 -9.21144664e-01
-1.00981033e+00 9.06237185e-01 5.28204978e-01 2.99979687e-01
1.29753554e+00 -1.33239821e-01 8.38213027e-01 2.25156739e-01
2.74417251e-01 -9.82819021e-01 -6.47359714e-02 3.90091628e-01
1.01156592e+00 -8.48979890e-01 -2.93870062e-01 -9.06430706e-02
-3.81261110e-01 7.95479715e-01 -3.29179280e-02 -9.63890702e-02
4.22295958e-01 6.38294458e-01 3.28413546e-01 6.01254106e-02
-3.64913285e-01 -2.36910984e-01 5.11812687e-01 5.45201242e-01
6.72401130e-01 1.65144056e-01 2.91981697e-01 1.17040861e+00
-1.06325054e+00 -1.62469298e-01 3.26600581e-01 4.43259090e-01
-1.38632372e-01 -1.35005081e+00 -7.80259788e-01 4.94009674e-01
-7.03964770e-01 -3.72690409e-01 -4.67029095e-01 5.67172348e-01
2.77430564e-01 1.24567032e+00 -1.73829406e-01 -3.29849184e-01
1.71603754e-01 3.70167762e-01 7.91142941e-01 -5.33273578e-01
-6.01064324e-01 5.64207017e-01 2.03789219e-01 -1.67306617e-01
-3.89519870e-01 -7.34480739e-01 -9.32531059e-01 -2.21804678e-01
-2.20837072e-01 1.54048651e-01 4.80933934e-01 6.62096977e-01
1.88024312e-01 9.03210223e-01 3.10574174e-01 -1.33371413e+00
-3.99337232e-01 -1.19085777e+00 -1.16833031e+00 5.31191051e-01
2.70960599e-01 -5.72639763e-01 -2.73016125e-01 2.40233585e-01] | [15.676801681518555, 5.4136199951171875] |
280532c2-3fe3-4ceb-875a-6687d8e05e96 | few-shot-open-set-learning-for-on-device | 2306.02161 | null | https://arxiv.org/abs/2306.02161v1 | https://arxiv.org/pdf/2306.02161v1.pdf | Few-Shot Open-Set Learning for On-Device Customization of KeyWord Spotting Systems | A personalized KeyWord Spotting (KWS) pipeline typically requires the training of a Deep Learning model on a large set of user-defined speech utterances, preventing fast customization directly applied on-device. To fill this gap, this paper investigates few-shot learning methods for open-set KWS classification by combining a deep feature encoder with a prototype-based classifier. With user-defined keywords from 10 classes of the Google Speech Command dataset, our study reports an accuracy of up to 76% in a 10-shot scenario while the false acceptance rate of unknown data is kept to 5%. In the analyzed settings, the usage of the triplet loss to train an encoder with normalized output features performs better than the prototypical networks jointly trained with a generator of dummy unknown-class prototypes. This design is also more effective than encoders trained on a classification problem and features fewer parameters than other iso-accuracy approaches. | ['Tinne Tuytelaars', 'Manuele Rusci'] | 2023-06-03 | null | null | null | null | ['open-set-learning', 'keyword-spotting'] | ['miscellaneous', 'speech'] | [ 2.91345924e-01 4.31693465e-01 -1.43346742e-01 -7.24403203e-01
-1.14606047e+00 -2.34345645e-01 4.48456198e-01 1.04320824e-01
-7.40922511e-01 4.99300748e-01 -1.07734345e-01 -5.32801151e-01
5.34299351e-02 -4.04546291e-01 -7.70650923e-01 -4.68152910e-01
-4.48445603e-02 5.74528337e-01 3.29237729e-01 -1.77530378e-01
-1.84059232e-01 7.91167989e-02 -1.96309400e+00 5.02766788e-01
6.21996045e-01 1.35280490e+00 4.14794296e-01 9.83733952e-01
-1.52605340e-01 5.88396668e-01 -1.01792014e+00 -4.55175489e-01
3.86533171e-01 -1.63517997e-01 -6.20038569e-01 -2.37452462e-01
4.10027117e-01 -6.05081081e-01 -1.83587074e-01 6.05784059e-01
9.86232102e-01 3.31209689e-01 4.09580052e-01 -1.18754637e+00
-5.77443898e-01 1.15369332e+00 3.54380310e-01 2.33444631e-01
3.08192641e-01 2.61603743e-01 1.10252297e+00 -8.03232431e-01
3.57695371e-01 7.88167357e-01 7.55847514e-01 5.60909450e-01
-1.23246980e+00 -7.32638419e-01 -3.37064087e-01 3.34272712e-01
-1.51456034e+00 -8.19655538e-01 4.81131852e-01 -3.24831307e-01
1.61372662e+00 3.84950399e-01 5.01594722e-01 1.64686072e+00
-2.68677324e-01 8.24313760e-01 5.09618700e-01 -6.35564327e-01
6.12327397e-01 8.60546947e-01 2.88742483e-01 3.52757215e-01
-9.84218195e-02 1.35616988e-01 -4.69977260e-01 -3.98187816e-01
-1.97122574e-01 -1.37346894e-01 -3.41759205e-01 -1.03729583e-01
-6.12055898e-01 9.57987487e-01 -1.86728805e-01 5.40995777e-01
-1.28214553e-01 -5.11444025e-02 8.06486368e-01 6.86208546e-01
6.15369380e-01 7.15174317e-01 -9.63550389e-01 -6.23460472e-01
-1.41423297e+00 6.85103387e-02 1.09523535e+00 1.25177240e+00
8.25820267e-01 3.46232444e-01 -4.20820862e-01 1.12293613e+00
-2.85907298e-01 2.05157608e-01 1.18739510e+00 -3.59797478e-01
3.39546412e-01 3.48832905e-01 -1.69311330e-01 -3.18365991e-02
-3.42009038e-01 -6.57323539e-01 -1.71203151e-01 -1.01507775e-01
1.37264252e-01 -2.76619703e-01 -1.04506183e+00 1.47086835e+00
1.13408707e-01 1.18718065e-01 2.76887894e-01 5.28728008e-01
7.43681729e-01 9.18061376e-01 -4.42186855e-02 -2.34664187e-01
1.19773448e+00 -9.23840225e-01 -7.35362351e-01 1.27710983e-01
9.71914768e-01 -5.14951944e-01 1.68653500e+00 2.17790321e-01
-5.35127819e-01 -4.06847745e-01 -1.20937359e+00 -9.97021049e-03
-9.00455892e-01 2.57610023e-01 3.87146622e-01 1.03511655e+00
-1.05630112e+00 7.89593220e-01 -4.08611715e-01 -4.12223786e-01
3.55427653e-01 4.19176310e-01 -1.56729341e-01 3.53101611e-01
-1.52446711e+00 8.88798773e-01 5.80821931e-01 -6.67782724e-01
-9.94826198e-01 -1.14292741e+00 -8.13713908e-01 4.18366462e-01
4.01323229e-01 -3.90690625e-01 1.65533447e+00 -1.05747867e+00
-2.04891872e+00 5.44175029e-01 2.27607608e-01 -9.79243100e-01
4.89780515e-01 -1.20164491e-01 -5.17174542e-01 -3.58127728e-02
-1.55214578e-01 3.56824130e-01 1.10535169e+00 -5.76940894e-01
-4.41961288e-01 7.16844201e-02 1.20686300e-01 -9.01636332e-02
-8.29896450e-01 9.17826518e-02 -1.17740318e-01 -4.65150535e-01
-8.04473877e-01 -5.44007838e-01 5.29845990e-02 -4.22168642e-01
-3.69153142e-01 -5.02076685e-01 9.89679396e-01 -7.45262504e-01
1.35865283e+00 -2.20869136e+00 -3.80749166e-01 -2.61692423e-02
-3.41544867e-01 6.60911560e-01 -9.24567729e-02 5.27802050e-01
-2.76853830e-01 1.99455805e-02 -5.18546104e-02 -6.91995978e-01
1.77736774e-01 1.35384262e-01 -2.48070732e-01 1.94298550e-01
2.00316265e-01 7.16967702e-01 -9.00957942e-01 -1.85914293e-01
2.27360234e-01 2.57495046e-01 -5.10826051e-01 5.39091706e-01
-3.26506168e-01 -3.01660568e-01 1.44161075e-01 4.90522563e-01
1.70587972e-01 5.39497696e-02 -2.02566311e-01 3.76472925e-03
6.70471117e-02 6.42775357e-01 -9.73405778e-01 1.68665516e+00
-9.33592021e-01 6.41923487e-01 -3.58454674e-01 -8.93987477e-01
8.71388614e-01 6.00512564e-01 2.61076838e-01 -6.37471378e-01
1.92519560e-01 3.34725201e-01 1.32802054e-02 -4.84361887e-01
4.68064696e-01 -1.06130941e-02 -1.56438336e-01 3.57143611e-01
8.69335771e-01 -1.05287265e-02 -1.48889765e-01 1.48470402e-01
1.43422902e+00 -2.12631389e-01 1.57860637e-01 -1.06223272e-02
1.98423222e-01 -2.70794421e-01 4.08189327e-01 1.08462024e+00
7.70923570e-02 5.55070400e-01 3.04176092e-01 -3.73462856e-01
-1.22250772e+00 -6.48330331e-01 -3.44179839e-01 1.32064378e+00
-4.63386446e-01 -7.93408573e-01 -8.57377946e-01 -6.94289148e-01
-2.96675637e-02 1.48065257e+00 -4.02648687e-01 -5.31533241e-01
2.81308778e-02 -3.93963367e-01 9.26481485e-01 2.01219469e-01
-1.55863360e-01 -8.09382558e-01 -8.01587403e-01 2.47674361e-01
2.06692696e-01 -1.12064433e+00 -3.41263950e-01 7.93213904e-01
-3.02473366e-01 -8.15999389e-01 -7.67076254e-01 -4.42723393e-01
2.37315863e-01 -1.64469495e-01 8.25821936e-01 -3.80954653e-01
-4.99371082e-01 3.62198323e-01 -9.57710624e-01 -6.72334194e-01
-7.94606745e-01 4.38475400e-01 3.72575730e-01 2.22826004e-01
5.04825592e-01 -5.12209833e-01 -2.02496111e-01 -7.45009854e-02
-6.64787650e-01 -1.18565589e-01 4.21397418e-01 1.11414564e+00
5.23427688e-02 -1.26531616e-01 6.44126952e-01 -7.79556096e-01
5.65203547e-01 -6.10673308e-01 -5.56817234e-01 3.38743985e-01
-8.43087912e-01 1.19092524e-01 8.20615172e-01 -9.36676860e-01
-7.38099396e-01 1.71067581e-01 -3.78473938e-01 -8.33890676e-01
-1.00653298e-01 2.18264516e-02 -2.21531287e-01 3.45544256e-02
8.79116714e-01 3.47945988e-01 -5.43057136e-02 -7.31259942e-01
3.34411532e-01 1.36665916e+00 1.49697736e-02 -9.01753679e-02
4.17502522e-01 -2.75466919e-01 -9.25953686e-01 -1.12579882e+00
-4.03651893e-01 -6.06736541e-01 -2.50527322e-01 -7.11522549e-02
3.76157194e-01 -8.50898743e-01 -2.24181712e-01 3.31490487e-01
-1.14812541e+00 -5.27640224e-01 -8.15027654e-01 5.21065235e-01
-6.58553004e-01 1.47194594e-01 -3.44832033e-01 -1.17718136e+00
-6.61705434e-01 -1.14063227e+00 1.18285978e+00 4.67515877e-03
-4.82608378e-01 -5.46549797e-01 -1.05064966e-01 6.85755089e-02
6.87923670e-01 -4.31675702e-01 7.13916004e-01 -1.68303347e+00
4.05754000e-02 -6.59254670e-01 3.59569788e-01 7.01675653e-01
1.45959020e-01 -5.81523143e-02 -1.55050457e+00 -1.49919614e-01
-2.25317031e-02 -5.83315492e-01 6.75611675e-01 2.28171609e-02
1.46958959e+00 -5.92268765e-01 -1.81259438e-01 5.56684673e-01
1.22617126e+00 4.93407220e-01 2.50028908e-01 -4.62716930e-02
2.85426587e-01 4.90684748e-01 5.27899206e-01 6.69189513e-01
2.77631599e-02 9.35570836e-01 9.91477370e-02 3.60118210e-01
4.02178019e-02 -3.61129582e-01 3.81852061e-01 5.82085788e-01
6.18723094e-01 -4.75929141e-01 -8.16274583e-01 5.51010728e-01
-1.46354413e+00 -9.95185435e-01 5.86979806e-01 2.29739332e+00
1.01160693e+00 1.89238712e-01 1.44842058e-01 3.39959711e-01
6.39988542e-01 2.19202396e-02 -5.19477487e-01 -7.11272538e-01
6.73214868e-02 4.74742651e-01 5.96689284e-01 3.68548363e-01
-1.07515037e+00 8.72024953e-01 6.08266020e+00 1.17396712e+00
-1.33832133e+00 5.66226244e-01 6.16804183e-01 -4.41998065e-01
-2.60420561e-01 -4.53050844e-02 -1.10931933e+00 9.54727829e-01
1.93789041e+00 -2.28710845e-01 3.57800245e-01 1.37509120e+00
1.35339841e-01 1.15103006e-01 -1.33454823e+00 1.15651691e+00
2.25891739e-01 -1.31042147e+00 -1.88284591e-01 -1.12970687e-01
1.48331583e-01 3.05559903e-01 -2.44718194e-02 8.12585473e-01
6.06580041e-02 -7.53611207e-01 7.86276460e-01 3.00063014e-01
1.10368550e+00 -6.23662829e-01 6.83035493e-01 6.77267730e-01
-4.44715828e-01 -3.35142463e-01 -3.58896166e-01 8.70310515e-02
6.52473271e-02 5.90968668e-01 -1.65822887e+00 2.84905404e-01
7.82574475e-01 4.54893596e-02 -2.70061195e-01 9.74825501e-01
3.45991433e-01 9.25863028e-01 -7.51971543e-01 -4.35632706e-01
2.20986694e-01 4.00268644e-01 4.83493775e-01 1.63730848e+00
2.86245137e-01 -2.35639378e-01 -1.46007478e-01 5.51696420e-01
2.42234059e-02 3.94905090e-01 -7.39261806e-01 -3.33370984e-01
5.82242608e-01 1.18239498e+00 -1.81149423e-01 -5.50534308e-01
-3.55197996e-01 1.07728052e+00 3.92278999e-01 1.57488868e-01
-7.88773060e-01 -8.54952931e-01 6.60272896e-01 1.74028859e-01
5.43843746e-01 3.23559225e-01 2.02710152e-01 -1.05749416e+00
2.46702358e-02 -7.15755641e-01 1.18414588e-01 -4.73958313e-01
-1.06747341e+00 7.31238663e-01 4.75982986e-02 -1.17862821e+00
-7.87792444e-01 -6.14144623e-01 -7.62784302e-01 8.34419727e-01
-1.10774624e+00 -8.62106919e-01 1.73821017e-01 3.09077680e-01
9.58493590e-01 -4.56890196e-01 1.16026139e+00 4.72503036e-01
-4.87876654e-01 1.25166559e+00 1.41027883e-01 2.48882119e-02
5.34668744e-01 -1.19189060e+00 4.87946481e-01 5.22795856e-01
1.69169411e-01 3.20450217e-01 8.27508986e-01 -3.75874400e-01
-1.20902634e+00 -1.04733706e+00 9.19855356e-01 -1.99200004e-01
6.13832653e-01 -9.29406524e-01 -9.01016593e-01 4.20312703e-01
2.09402576e-01 1.75197646e-01 1.01248157e+00 1.07923806e-01
-3.74159604e-01 -3.06518883e-01 -1.14525211e+00 2.24873975e-01
9.66238379e-01 -7.27236688e-01 -5.57581544e-01 6.24040425e-01
1.18531728e+00 -3.20419520e-01 -8.39636266e-01 1.16887964e-01
4.88491029e-01 -7.72155821e-01 5.89576006e-01 -8.51880610e-01
-2.57105291e-01 2.57575154e-01 -2.92617708e-01 -1.53330195e+00
8.07392299e-02 -6.86626911e-01 -3.22666675e-01 1.35804141e+00
7.78916001e-01 -5.32706439e-01 6.37438476e-01 5.54516912e-01
-4.56177086e-01 -8.65737855e-01 -1.26879025e+00 -9.95481730e-01
-4.45712626e-01 -8.00461173e-01 6.39007807e-01 8.53120029e-01
2.34956637e-01 3.73750240e-01 -4.32224602e-01 9.84527520e-04
9.74174142e-02 -4.30695742e-01 7.27238059e-01 -9.24407542e-01
-6.38535619e-01 -2.10260436e-01 -7.86889970e-01 -5.05589962e-01
4.25337493e-01 -8.00606132e-01 8.76556784e-02 -8.80103409e-01
-1.62266359e-01 -5.43577969e-01 -2.53685087e-01 7.32220352e-01
3.14751744e-01 -2.77188897e-01 -1.17234290e-01 -1.53386429e-01
-4.54510033e-01 6.83511555e-01 1.26445860e-01 -3.34913492e-01
-2.79877156e-01 3.49285632e-01 -2.71623909e-01 2.81057864e-01
7.17334688e-01 -4.96513784e-01 -7.20395148e-01 1.30161017e-01
-5.09561077e-02 -1.24840625e-01 1.25127491e-02 -1.05399573e+00
2.96702802e-01 1.96861178e-01 8.70685652e-02 -4.25704509e-01
6.67537928e-01 -7.39206374e-01 1.08291119e-01 2.97876954e-01
-4.55845535e-01 -4.94837880e-01 1.13679826e-01 5.31199157e-01
2.70708799e-02 -6.84794545e-01 6.88483477e-01 -8.27991217e-02
-8.72809052e-01 1.27675101e-01 -4.78210449e-01 -8.21529701e-02
1.10251069e+00 -3.04158777e-01 1.21330634e-01 -4.71947134e-01
-6.02151096e-01 -7.17353746e-02 1.79724365e-01 6.98705614e-01
5.04456580e-01 -9.59680498e-01 -2.78461039e-01 3.99552017e-01
5.29538393e-01 -3.20801497e-01 1.40943035e-01 5.35037339e-01
-1.04348823e-01 4.39950019e-01 3.10353249e-01 -5.39067745e-01
-1.22111177e+00 5.21248996e-01 5.67632258e-01 1.56848490e-01
-6.89515769e-01 1.07079005e+00 -3.29576612e-01 -5.29290676e-01
7.07851410e-01 -4.68584061e-01 -8.18224042e-04 2.07151204e-01
7.48955250e-01 3.30541819e-01 6.65931344e-01 -2.48522967e-01
-3.88244122e-01 -2.91790247e-01 -1.87084347e-01 -7.69391358e-02
1.24716759e+00 2.40839496e-01 7.21432567e-01 9.10326242e-01
1.61228943e+00 -2.74095923e-01 -1.11637676e+00 -3.07943553e-01
1.76726480e-03 -1.43620968e-01 8.22051615e-02 -8.50895345e-01
-5.92871547e-01 7.09587097e-01 8.08702707e-01 3.29075843e-01
7.58577704e-01 1.94294974e-01 7.49409378e-01 5.55526972e-01
3.92775774e-01 -1.54904211e+00 -1.58075824e-01 4.30155724e-01
6.53217256e-01 -1.23825359e+00 -5.19106030e-01 5.62324040e-02
-6.92538857e-01 1.03597963e+00 3.99770826e-01 2.24830583e-01
7.11024106e-01 4.43595231e-01 -1.00418754e-01 2.19576851e-01
-1.03714204e+00 -2.29306549e-01 8.65894407e-02 6.29501164e-01
1.93852752e-01 1.25632053e-02 -9.81853306e-02 1.03261101e+00
-4.60003614e-01 -6.07359782e-02 4.12756622e-01 6.82683051e-01
-3.00402939e-01 -8.86923194e-01 1.04085013e-01 8.86668682e-01
-3.93146068e-01 -4.14465606e-01 -2.01551393e-01 4.17355746e-01
1.43420443e-01 8.78543377e-01 3.25183392e-01 -7.57479489e-01
4.32301134e-01 7.40208507e-01 7.08087757e-02 -9.58530605e-01
-6.97489917e-01 -3.84902954e-01 3.72201592e-01 -5.48913300e-01
1.47403046e-01 -5.88482916e-01 -7.80568004e-01 1.91210076e-01
-7.38224506e-01 3.73922437e-01 1.02058613e+00 9.36229706e-01
5.04404783e-01 4.77435887e-01 6.92394972e-01 -7.72306502e-01
-1.22422040e+00 -1.28684735e+00 -5.86504638e-01 1.90446392e-01
3.10572654e-01 -7.35487223e-01 -7.04154313e-01 -2.87327349e-01] | [14.211549758911133, 6.474434852600098] |
f4f0b6fa-fc31-46f3-8d32-bf636ce7143f | brain-tumor-recurrence-vs-radiation-necrosis | 2306.03270 | null | https://arxiv.org/abs/2306.03270v1 | https://arxiv.org/pdf/2306.03270v1.pdf | Brain Tumor Recurrence vs. Radiation Necrosis Classification and Patient Survivability Prediction | GBM (Glioblastoma multiforme) is the most aggressive type of brain tumor in adults that has a short survival rate even after aggressive treatment with surgery and radiation therapy. The changes on magnetic resonance imaging (MRI) for patients with GBM after radiotherapy are indicative of either radiation-induced necrosis (RN) or recurrent brain tumor (rBT). Screening for rBT and RN at an early stage is crucial for facilitating faster treatment and better outcomes for the patients. Differentiating rBT from RN is challenging as both may present with similar radiological and clinical characteristics on MRI. Moreover, learning-based rBT versus RN classification using MRI may suffer from class imbalance due to lack of patient data. While synthetic data generation using generative models has shown promise to address class imbalance, the underlying data representation may be different in synthetic or augmented data. This study proposes computational modeling with statistically rigorous repeated random sub-sampling to balance the subset sample size for rBT and RN classification. The proposed pipeline includes multiresolution radiomic feature (MRF) extraction followed by feature selection with statistical significance testing (p<0.05). The five-fold cross validation results show the proposed model with MRF features classifies rBT from RN with an area under the curve (AUC) of 0.8920+-.055. Moreover, considering the dependence between survival time and censor time (where patients are not followed up until death), we demonstrate the feasibility of using MRF radiomic features as a non-invasive biomarker to identify patients who are at higher risk of recurrence or radiation necrosis. The cross-validated results show that the MRF model provides the best overall performance with an AUC of 0.770+-.032. | ['K. M. Iftekharuddin', 'A. Vossough', 'E. Lappinen', 'A. Temtam', 'W. Farzana', 'M. S. Sadique'] | 2023-06-05 | null | null | null | null | ['synthetic-data-generation', 'synthetic-data-generation'] | ['medical', 'miscellaneous'] | [ 3.32370669e-01 4.24704142e-02 -4.59759012e-02 -3.67278993e-01
-1.06600380e+00 -1.48140669e-01 5.96224308e-01 4.21075553e-01
-5.05744159e-01 1.05558372e+00 5.42865038e-01 -4.02979463e-01
-3.88676226e-01 -7.82156467e-01 -1.64466307e-01 -1.02544260e+00
-1.91527307e-01 6.28650665e-01 -6.09360524e-02 1.22579999e-01
1.49089932e-01 7.36824870e-01 -1.28407383e+00 2.83303559e-01
1.02830076e+00 6.98405325e-01 6.76460743e-01 6.47739589e-01
-5.82403969e-03 7.14115024e-01 -5.17650843e-01 2.41306424e-01
4.29060310e-02 -3.13428551e-01 -5.64915895e-01 -1.62395522e-01
1.11192510e-01 -2.41448686e-01 -3.84868741e-01 7.65009344e-01
7.34366834e-01 -2.13574320e-01 1.01984882e+00 -1.04535902e+00
-1.17346026e-01 3.75485301e-01 -5.25538146e-01 3.37313920e-01
1.66329086e-01 1.93953499e-01 3.81402314e-01 -7.88882196e-01
5.27397573e-01 5.62966049e-01 6.84953213e-01 5.60702562e-01
-1.23007405e+00 -5.25375068e-01 -3.37193042e-01 2.56589372e-02
-1.15548384e+00 -5.91064878e-02 1.10223383e-01 -8.82687449e-01
7.98500717e-01 5.90746164e-01 9.16971684e-01 8.80257547e-01
9.30144310e-01 2.96445996e-01 1.57361615e+00 -1.08617572e-02
4.30683136e-01 -3.13498378e-01 4.49274302e-01 5.81038177e-01
5.11237919e-01 2.59742528e-01 -3.36189300e-01 -2.54679620e-01
5.87119222e-01 4.54605669e-01 -5.36743164e-01 -5.33216968e-02
-1.23676240e+00 7.71921635e-01 3.03132206e-01 3.10884058e-01
-5.53458154e-01 7.67480731e-02 3.51358622e-01 1.43967003e-01
2.59473950e-01 -4.89800051e-02 -4.08226788e-01 8.23232979e-02
-1.05739975e+00 -1.54347960e-02 8.88807997e-02 5.47018111e-01
5.91770150e-02 -1.35926992e-01 -2.01270863e-01 8.84857953e-01
7.67656788e-02 4.19640630e-01 1.33915746e+00 -9.89423171e-02
1.22979522e-01 6.15712464e-01 -2.50433415e-01 -1.84147775e-01
-1.20192349e+00 -9.46916819e-01 -1.09955156e+00 3.68398458e-01
3.35252881e-01 1.34039551e-01 -1.31731796e+00 1.55400026e+00
-1.16446830e-01 -2.52000749e-01 2.05078125e-01 8.22613180e-01
1.00141549e+00 8.14840794e-02 5.86145148e-02 -4.11663383e-01
1.62638903e+00 -4.25130278e-01 -3.83149236e-01 -1.01030424e-01
1.02037227e+00 -3.53109479e-01 8.46535563e-01 1.91262588e-01
-7.66876996e-01 2.37427905e-01 -8.28950107e-01 2.60035127e-01
-4.69396301e-02 2.04574108e-01 8.03028822e-01 9.62610543e-01
-9.63451207e-01 2.66189367e-01 -1.10040867e+00 -6.40465260e-01
7.33921707e-01 5.31318188e-01 -8.29949260e-01 -2.88914025e-01
-8.63536119e-01 1.04223430e+00 1.24857634e-01 -5.84967211e-02
-9.08769310e-01 -1.10990989e+00 -4.13717121e-01 -3.03071380e-01
-4.29775000e-01 -1.00643873e+00 6.28496706e-01 -5.54186821e-01
-1.03220105e+00 8.07118773e-01 -4.19410676e-01 -5.75611234e-01
7.86422133e-01 2.52582431e-01 -2.16172546e-01 7.12222084e-02
2.32880920e-01 1.65679321e-01 3.20780784e-01 -7.82657743e-01
-7.26800501e-01 -9.82408702e-01 -4.87770230e-01 1.05608530e-01
1.08770870e-01 7.46526942e-02 4.92150396e-01 -7.01752067e-01
7.63139606e-01 -9.57429290e-01 -4.67804879e-01 -5.90725005e-01
-2.37409130e-01 2.44216308e-01 4.43545878e-01 -9.50534463e-01
7.37585902e-01 -1.53368783e+00 -2.39595041e-01 2.16709167e-01
4.94869351e-01 -2.37558454e-01 3.26982886e-01 -2.24656314e-02
-5.06238878e-01 3.02607745e-01 -2.85036683e-01 1.55565634e-01
-6.03234708e-01 -4.53194290e-01 3.01923066e-01 8.84392917e-01
-7.90138841e-02 9.92353261e-01 -7.48192430e-01 1.17589056e-01
4.56924103e-02 4.77663785e-01 -3.10409535e-02 7.22961547e-03
5.40732801e-01 7.96801984e-01 -1.65248677e-01 7.73995101e-01
5.03919899e-01 -1.88376173e-01 -1.54075563e-01 -2.88524609e-02
5.25645651e-02 6.08130060e-02 -6.11443818e-01 1.36100864e+00
-2.46567085e-01 7.18555570e-01 -2.66378015e-01 -6.24073267e-01
1.00629020e+00 2.45454997e-01 6.01905406e-01 -8.21482539e-01
1.18441582e-01 4.79073197e-01 3.46246272e-01 -3.94288659e-01
-4.06232923e-02 -5.75494289e-01 2.29108587e-01 2.69858301e-01
-3.05452764e-01 3.35710938e-03 -1.30450696e-01 -3.61479633e-02
1.77070832e+00 -3.81132722e-01 3.89754117e-01 -4.34080720e-01
4.36292976e-01 1.18181534e-01 6.03233635e-01 7.48138785e-01
-2.29601309e-01 9.16416109e-01 2.60157079e-01 -2.52445102e-01
-7.53420353e-01 -1.12330973e+00 -8.07907224e-01 1.09031983e-01
-3.52913588e-01 1.39651105e-01 -3.78553659e-01 -4.12487358e-01
-1.91552863e-01 8.65994334e-01 -7.23194659e-01 -2.07487732e-01
-1.86768457e-01 -1.73514855e+00 4.75961417e-01 4.42623168e-01
1.53656051e-01 -5.31611741e-01 -8.65866125e-01 1.67723000e-01
-2.01865539e-01 -4.82483655e-01 1.23374157e-01 4.85477030e-01
-1.26079845e+00 -1.19651961e+00 -1.09406483e+00 -5.21434903e-01
1.00260687e+00 1.45545438e-01 6.31851971e-01 -1.45041905e-02
-6.19368076e-01 2.27533475e-01 -2.26412684e-01 -2.71943241e-01
-4.21913296e-01 -9.25943106e-02 9.14343516e-04 -4.14037526e-01
2.18941614e-01 -6.17057860e-01 -8.47759664e-01 2.64992386e-01
-8.05639148e-01 3.99642676e-01 6.95637405e-01 9.17935491e-01
5.62440097e-01 6.05393574e-03 8.38256657e-01 -6.81555748e-01
4.89376247e-01 -5.61475635e-01 -3.10602248e-01 6.21150471e-02
-7.05002725e-01 -2.21294761e-01 3.32160026e-01 -1.88916653e-01
-6.71906948e-01 -1.86184779e-01 2.88216174e-01 8.94731209e-02
-3.02579135e-01 7.01033473e-01 4.08392139e-02 1.10502943e-01
6.69891417e-01 2.13436902e-01 1.95058361e-01 -1.20963855e-02
-3.41920346e-01 6.52987003e-01 3.11015457e-01 -3.41577604e-02
3.29243630e-01 6.58871233e-01 5.14333606e-01 -7.11602449e-01
-2.97061503e-01 -4.70291615e-01 -3.60808283e-01 -1.98433578e-01
6.90330207e-01 -9.23306227e-01 -4.05671835e-01 5.97540438e-01
-6.41643941e-01 -2.19785422e-01 -8.46927688e-02 1.07699406e+00
-6.34106755e-01 -3.24477293e-02 -4.21495736e-01 -7.73389637e-01
-8.31708193e-01 -1.52246904e+00 6.37713432e-01 2.67353237e-01
-4.06970292e-01 -7.28089511e-01 3.71356495e-02 6.27605677e-01
4.75589633e-01 6.49211586e-01 1.15671790e+00 -1.00453544e+00
-3.31738770e-01 -6.43150747e-01 -1.94152847e-01 -3.10635865e-01
3.77751231e-01 -2.90273190e-01 -8.70871484e-01 -5.38828313e-01
4.97894734e-02 8.75308141e-02 8.10294747e-01 8.18984091e-01
6.83473825e-01 2.58352548e-01 -4.13683087e-01 4.68775064e-01
1.64971173e+00 5.18409371e-01 7.35346735e-01 7.67416239e-01
3.92712414e-01 5.48662961e-01 2.42094010e-01 2.67042130e-01
3.11109662e-01 4.40008998e-01 5.26001275e-01 1.48294941e-01
-4.25015926e-01 1.20384105e-01 3.10539275e-01 2.87609369e-01
-1.63806230e-01 9.16610137e-02 -1.38921201e+00 6.47815526e-01
-1.42652190e+00 -8.93052340e-01 -7.65737832e-01 2.77915740e+00
4.58704233e-01 5.27576767e-02 -4.92012799e-02 1.95846826e-01
8.10036063e-01 -5.11577845e-01 -4.42727655e-01 -1.04796691e-02
-5.10837317e-01 1.56266361e-01 8.69547307e-01 3.16810578e-01
-5.42821765e-01 1.34517744e-01 5.78459883e+00 3.04726362e-01
-1.39550567e+00 1.88894615e-01 9.24484551e-01 -2.72503018e-01
-1.27084762e-01 2.08142281e-01 -5.12777925e-01 4.94489849e-01
1.09274173e+00 -3.42131913e-01 8.72002989e-02 2.43460134e-01
7.03842580e-01 -6.26464248e-01 -7.76115417e-01 8.24788809e-01
7.34026656e-02 -1.30646861e+00 -2.19745740e-01 3.76038402e-01
4.45442617e-01 4.15292382e-01 -2.13990837e-01 -1.15171941e-02
4.28922251e-02 -1.43399334e+00 6.31629348e-01 9.72346663e-01
9.57778513e-01 -6.73775434e-01 1.24561143e+00 3.14220250e-01
-6.97409809e-01 -6.62983283e-02 -1.55145964e-02 1.71357796e-01
1.12514056e-01 9.66056883e-01 -1.30876088e+00 6.36634827e-01
5.70215642e-01 2.87354410e-01 -7.76351511e-01 1.26818764e+00
4.92495932e-02 5.51582277e-01 -2.18930513e-01 7.70719647e-02
-3.73989493e-01 -7.68437460e-02 6.26575291e-01 8.34249914e-01
6.47144020e-01 1.48302495e-01 -2.76799083e-01 5.00450969e-01
5.34193873e-01 3.27180624e-01 -2.61005819e-01 2.37911433e-01
2.77211994e-01 1.10413229e+00 -1.21000195e+00 2.51487223e-03
-3.74778867e-01 5.21891296e-01 1.14885360e-01 7.56912455e-02
-5.38451254e-01 -6.53646188e-03 1.22648142e-01 5.91445684e-01
-2.72216350e-01 1.38319433e-01 -8.49416673e-01 -1.02037370e+00
-1.17374994e-01 -5.53561807e-01 3.52125376e-01 -8.31361473e-01
-1.01437795e+00 7.05930412e-01 -2.72907853e-01 -1.05358183e+00
2.61581875e-03 -3.90162706e-01 -6.99789107e-01 1.15600705e+00
-1.12282169e+00 -1.20331323e+00 -6.11942708e-01 1.88595355e-01
2.94076562e-01 -2.44493112e-01 9.86627162e-01 -2.06147403e-01
-6.00399911e-01 4.18067068e-01 1.01391971e-01 -7.47026429e-02
3.19325596e-01 -1.23066986e+00 -3.53409827e-01 6.60381377e-01
-7.10994780e-01 6.27513707e-01 7.17410743e-01 -9.95575011e-01
-1.00611126e+00 -1.18851209e+00 6.17862225e-01 -1.22334771e-01
5.54209888e-01 1.34605139e-01 -6.78076923e-01 3.84613127e-01
-3.81844640e-01 -7.63252676e-02 1.09179366e+00 -3.15415263e-01
8.26992393e-02 2.03431055e-01 -1.63292372e+00 4.74315882e-01
6.76249504e-01 -1.38918713e-01 -4.72713143e-01 3.66045892e-01
2.82517914e-02 -2.11997420e-01 -1.19799924e+00 8.07238817e-01
5.68376541e-01 -1.08516967e+00 6.41873837e-01 -5.71856022e-01
2.65686303e-01 -4.38318133e-01 -2.30270296e-01 -1.25217116e+00
-4.11794931e-01 -1.02870435e-01 4.15979266e-01 9.68015313e-01
5.01225471e-01 -9.45637524e-01 1.00247049e+00 8.99932921e-01
-1.04965538e-01 -9.77011383e-01 -1.07104039e+00 -8.32041979e-01
2.67902076e-01 -4.48058426e-01 5.01140773e-01 6.55557334e-01
-6.48318678e-02 -1.34849608e-01 3.25427324e-01 2.82168895e-01
6.05587423e-01 -1.57956347e-01 4.11265820e-01 -1.38158619e+00
2.35368177e-01 -5.86889207e-01 -9.89689112e-01 1.55768916e-01
-7.60258138e-02 -1.33362603e+00 -4.68811572e-01 -2.09491873e+00
7.64505982e-01 -5.07441998e-01 -4.80288297e-01 3.08486372e-01
8.19208473e-03 -1.25626594e-01 -1.71588376e-01 4.06320900e-01
4.10584778e-01 3.29548120e-01 9.24289942e-01 -8.95820409e-02
-2.19303101e-01 2.43618309e-01 -4.44976002e-01 5.99317551e-01
1.00698471e+00 -6.17995739e-01 -4.78938639e-01 2.27556869e-01
-1.46024868e-01 4.00055856e-01 6.38657391e-01 -1.06551290e+00
3.43879722e-02 -2.06451938e-01 7.22468376e-01 -7.62217939e-01
3.17452773e-02 -7.05225766e-01 6.48594320e-01 9.66541827e-01
-1.49909407e-01 2.66803354e-01 -7.77711570e-02 4.69331771e-01
9.07710195e-02 -4.21253830e-01 9.14841354e-01 -1.41185611e-01
-1.39534608e-01 3.41699690e-01 -8.53572249e-01 -3.09577614e-01
1.09656143e+00 -5.50574005e-01 -5.45172930e-01 -3.29615414e-01
-9.30177212e-01 -1.46470234e-01 6.49557531e-01 1.40770510e-01
6.97656810e-01 -1.16290903e+00 -8.79648507e-01 -5.32037355e-02
2.60932207e-01 1.08085588e-01 4.24867392e-01 1.49169254e+00
-6.04142666e-01 3.47638011e-01 -2.26438180e-01 -6.08827472e-01
-1.45165575e+00 -8.24566483e-02 6.50231600e-01 -4.20471728e-01
-7.93632209e-01 7.03032732e-01 1.27493903e-01 -1.76203117e-01
-8.90857354e-02 -2.41975218e-01 -2.75154322e-01 1.24462403e-01
6.18887365e-01 5.58368325e-01 6.38054132e-01 -6.93599403e-01
-4.63128954e-01 5.05494885e-02 -3.66112381e-01 -2.53178060e-01
1.59928119e+00 2.85342187e-02 -4.14248079e-01 4.62740451e-01
1.01608741e+00 -7.08732381e-02 -7.43255615e-01 2.23208517e-01
1.29283637e-01 -3.31733286e-01 4.26760375e-01 -1.07322061e+00
-9.86078858e-01 2.92662263e-01 1.20091915e+00 -2.80224025e-01
1.01468515e+00 -7.33845308e-02 3.62241805e-01 -3.16098481e-01
5.86208463e-01 -6.84316754e-01 -5.57297170e-01 2.27563575e-01
9.86627579e-01 -9.49471772e-01 1.52825475e-01 -2.74610966e-01
-2.88778037e-01 1.18553627e+00 2.40158439e-01 1.05483778e-01
5.84303379e-01 2.52724200e-01 9.47976336e-02 -2.53128052e-01
-6.70373678e-01 7.76458532e-02 -5.28927445e-02 7.70649135e-01
6.88383222e-01 5.28959870e-01 -7.77910590e-01 9.57528770e-01
-6.41596019e-01 1.14489198e-01 9.48816419e-01 1.00997758e+00
-3.79432976e-01 -8.83807361e-01 -4.88101572e-01 1.27207530e+00
-4.02161688e-01 -6.89383745e-02 -6.28391653e-02 5.55306971e-01
-1.98280618e-01 8.87867510e-01 -9.64256227e-02 -5.74913919e-02
1.65827930e-01 2.04548270e-01 4.63860929e-01 -4.28139448e-01
-5.17015576e-01 4.91070040e-02 2.32975800e-02 -9.48578268e-02
-3.54251176e-01 -9.61069584e-01 -1.59146023e+00 1.21168546e-01
-4.27657157e-01 -7.08185881e-02 1.02952230e+00 1.02302754e+00
9.83175337e-02 7.98379779e-01 4.60596442e-01 -3.82307768e-01
-3.83491404e-02 -9.90801275e-01 -7.79513597e-01 1.17767662e-01
3.12539130e-01 -7.00592458e-01 -4.46250468e-01 -2.73807794e-01] | [14.765338897705078, -2.572373151779175] |
f1495b40-d8d5-4b35-a69a-f22664625ca0 | xsim-an-improved-proxy-to-bitext-mining | 2306.12907 | null | https://arxiv.org/abs/2306.12907v1 | https://arxiv.org/pdf/2306.12907v1.pdf | xSIM++: An Improved Proxy to Bitext Mining Performance for Low-Resource Languages | We introduce a new proxy score for evaluating bitext mining based on similarity in a multilingual embedding space: xSIM++. In comparison to xSIM, this improved proxy leverages rule-based approaches to extend English sentences in any evaluation set with synthetic, hard-to-distinguish examples which more closely mirror the scenarios we encounter during large-scale mining. We validate this proxy by running a significant number of bitext mining experiments for a set of low-resource languages, and subsequently train NMT systems on the mined data. In comparison to xSIM, we show that xSIM++ is better correlated with the downstream BLEU scores of translation systems trained on mined bitexts, providing a reliable proxy of bitext mining performance without needing to run expensive bitext mining pipelines. xSIM++ also reports performance for different error types, offering more fine-grained feedback for model development. | ['Holger Schwenk', 'Alex Mourachko', 'Onur Çelebi', 'Kevin Heffernan', 'Mingda Chen'] | 2023-06-22 | null | null | null | null | ['nmt'] | ['computer-code'] | [ 2.19371051e-01 7.75127932e-02 -4.59458798e-01 -6.35242820e-01
-1.32799816e+00 -8.32586944e-01 8.46001923e-01 1.83830678e-01
-7.55355179e-01 7.21949816e-01 3.80542874e-01 -9.04738069e-01
1.19914729e-02 -7.12999105e-01 -7.49335825e-01 6.24621287e-03
7.88165182e-02 6.46879315e-01 -3.91609609e-01 -4.20692146e-01
1.59479663e-01 -7.40580410e-02 -1.20645702e+00 4.58950490e-01
9.52516973e-01 4.03426379e-01 1.22581482e-01 5.34498036e-01
-1.71066031e-01 1.51983231e-01 -6.09873116e-01 -1.26793170e+00
5.52560031e-01 -2.84652829e-01 -7.13728368e-01 -4.53370959e-01
8.41654062e-01 -7.12760389e-02 -1.50925750e-02 1.12147129e+00
4.63179320e-01 -3.78001422e-01 5.30838311e-01 -8.04550231e-01
-5.75180173e-01 1.20570207e+00 -4.80699807e-01 3.75679135e-01
4.09671158e-01 3.57123822e-01 1.54523528e+00 -9.38222587e-01
1.07953703e+00 1.08921146e+00 7.26516783e-01 3.24116647e-01
-1.39520037e+00 -8.56676221e-01 -1.10508725e-01 1.40082046e-01
-8.73928368e-01 -5.72222471e-01 3.00610840e-01 -1.85426399e-01
1.50792706e+00 4.94440228e-01 4.60435927e-01 1.61851025e+00
1.67336479e-01 9.56273735e-01 1.22413886e+00 -8.92722189e-01
-2.21539568e-02 2.85083085e-01 3.98707271e-01 6.49157465e-01
4.78100717e-01 3.65430593e-01 -5.68080544e-01 -3.94575149e-01
1.27900699e-02 -4.44468558e-01 -6.60364181e-02 1.31161317e-01
-1.33907712e+00 1.05315149e+00 -1.35137007e-01 1.25382438e-01
-2.37398654e-01 -1.35540962e-01 6.88423514e-01 8.84597361e-01
7.27256179e-01 9.03467238e-01 -7.69354403e-01 -5.63585401e-01
-1.10547245e+00 1.47548676e-01 8.37660074e-01 9.71491635e-01
7.11439073e-01 -7.34942108e-02 2.30785571e-02 1.06050980e+00
3.19297731e-01 6.05854690e-01 7.46462762e-01 -4.21879768e-01
1.18915260e+00 5.28674901e-01 -4.36971933e-01 -2.27667734e-01
4.02154699e-02 -5.14949083e-01 -6.18392117e-02 2.90775180e-01
4.16885555e-01 -3.55858684e-01 -7.51384795e-01 1.77106011e+00
2.20323503e-01 -4.18273270e-01 1.07041575e-01 5.67837715e-01
3.02575499e-01 5.46772540e-01 8.93410593e-02 -1.38686165e-01
1.47165513e+00 -8.31506371e-01 -3.26011568e-01 -4.51948553e-01
1.28524876e+00 -1.13996613e+00 1.56276059e+00 4.48998719e-01
-8.38419199e-01 -2.05415219e-01 -1.36168945e+00 1.15404546e-01
-2.10181609e-01 3.47843617e-02 8.40425611e-01 8.65842104e-01
-8.33139539e-01 9.16521609e-01 -7.49864459e-01 -5.47057211e-01
2.21509457e-01 1.97020561e-01 -4.66107428e-01 -5.30007891e-02
-1.11915386e+00 1.25052226e+00 4.12989706e-01 -4.13714945e-01
-6.12719893e-01 -8.30364943e-01 -8.94038379e-01 -3.20316881e-01
-1.68407615e-02 -6.01533890e-01 1.40124214e+00 -7.75772035e-01
-1.32502699e+00 8.68138194e-01 3.60736139e-02 -4.79433239e-01
5.86242795e-01 -5.61002433e-01 -6.47866011e-01 -4.55544531e-01
3.62477571e-01 3.28587323e-01 5.16890943e-01 -6.53249085e-01
-5.83260596e-01 -2.90820450e-01 -1.83956042e-01 1.78383246e-01
-5.27449250e-01 4.40329939e-01 -2.89520025e-01 -7.34317899e-01
-3.75263184e-01 -7.90498972e-01 -1.34483844e-01 -5.00469863e-01
-7.53437817e-01 -2.71794111e-01 5.02417922e-01 -7.07811832e-01
1.48578930e+00 -1.68484986e+00 -1.01038411e-01 3.06245744e-01
-2.41423082e-02 2.55093366e-01 -6.14207625e-01 6.19211435e-01
-1.44129112e-01 4.41668898e-01 -1.83833256e-01 -4.82634187e-01
2.56940544e-01 1.97589815e-01 -4.17761087e-01 1.02491952e-01
3.89492452e-01 9.18554723e-01 -9.96132791e-01 -7.36752510e-01
1.32689267e-01 -9.23391326e-06 -5.71463883e-01 -1.53799783e-02
-3.43399227e-01 -2.23717123e-01 -2.40235969e-01 6.38422728e-01
2.36065120e-01 3.72691303e-02 3.74768704e-01 9.09806117e-02
-3.64553370e-02 9.32825863e-01 -7.46335745e-01 1.77350163e+00
-7.21379340e-01 7.69593120e-01 -4.46741372e-01 -3.65404695e-01
7.64939368e-01 1.27550066e-01 3.82933617e-02 -8.53383899e-01
5.32855988e-02 4.10487145e-01 1.89189464e-01 -4.18441951e-01
5.30232847e-01 -1.92082599e-01 -2.19428316e-01 1.11455202e+00
3.53517115e-01 6.58302978e-02 5.65422714e-01 2.34181002e-01
1.39043272e+00 2.79809207e-01 3.38500351e-01 -3.11376661e-01
-5.44935875e-02 3.82458508e-01 5.32081544e-01 5.97620010e-01
1.12760849e-01 1.96394145e-01 3.69530842e-02 -4.48273450e-01
-1.35731578e+00 -1.14462554e+00 -5.07080078e-01 1.28142953e+00
-4.21660632e-01 -9.93617237e-01 -7.58093894e-01 -1.23958528e+00
-1.08949408e-01 1.11533558e+00 -5.13106167e-01 3.70997377e-03
-6.87309444e-01 -1.26318479e+00 9.93999541e-01 2.71581441e-01
-1.09003529e-01 -9.45458174e-01 -2.75743306e-01 3.71573716e-01
-2.13241681e-01 -6.90837324e-01 -6.22423351e-01 6.30159140e-01
-9.14186120e-01 -1.11767340e+00 -5.48469797e-02 -5.53836405e-01
1.36493340e-01 -3.26849818e-01 1.68336475e+00 -1.84385955e-01
-2.84166247e-01 -2.89307475e-01 -4.29243892e-01 -6.62993252e-01
-9.67118144e-01 4.63726223e-01 2.41303563e-01 -6.34132981e-01
1.05203938e+00 -4.90868270e-01 -2.74460316e-01 7.24543259e-02
-5.83350897e-01 -4.77762669e-02 8.82652044e-01 9.63289142e-01
5.04392862e-01 -5.28442562e-01 4.27641302e-01 -1.16603482e+00
1.00501418e+00 -5.72740674e-01 -5.11276186e-01 4.40118313e-01
-1.17320383e+00 5.21655440e-01 5.65386355e-01 -5.30566633e-01
-8.53832841e-01 -2.76194423e-01 -4.34787959e-01 5.92362396e-02
-8.48467201e-02 4.81434405e-01 -1.87765677e-02 4.14970309e-01
1.01873159e+00 1.86394546e-02 -3.61769386e-02 -7.71470904e-01
6.27103925e-01 8.29855800e-01 2.31256172e-01 -7.41802037e-01
7.99378097e-01 -1.78085372e-01 -6.61616743e-01 -3.55812490e-01
-6.73557937e-01 -2.21344754e-01 -4.78805631e-01 4.66895491e-01
5.71740925e-01 -7.84730554e-01 -2.31503889e-01 -1.50777936e-01
-1.00695682e+00 -3.28039765e-01 -2.97518373e-01 8.27176571e-01
-4.82961357e-01 4.42419767e-01 -9.43168163e-01 -3.61189812e-01
-8.95136356e-01 -1.08257675e+00 1.09287667e+00 -4.72487450e-01
-9.92444515e-01 -1.17290473e+00 8.03595841e-01 3.51452112e-01
1.31907508e-01 2.02048216e-02 1.35440326e+00 -1.01722240e+00
-1.47020919e-02 -7.41323829e-02 1.88799322e-01 3.65635008e-01
-1.11240363e-02 9.03381258e-02 -8.83299053e-01 -3.28584492e-01
-1.65426001e-01 -5.13316989e-01 7.59635687e-01 -1.52838826e-01
4.32026267e-01 -4.43293124e-01 -3.24954301e-01 7.43724704e-01
1.31707978e+00 -1.91160902e-01 3.64927143e-01 9.56856847e-01
3.70174408e-01 4.43934083e-01 7.26074040e-01 4.53341343e-02
2.08931118e-01 7.42686868e-01 1.59859061e-01 -1.70381129e-01
-5.02428897e-02 -4.73588765e-01 8.67392659e-01 1.24642611e+00
2.07618386e-01 -9.36544612e-02 -8.01439345e-01 4.23188090e-01
-1.39723587e+00 -9.71933663e-01 -1.42702639e-01 2.12065792e+00
1.40608656e+00 3.53929669e-01 1.29277259e-01 7.16057643e-02
5.43209374e-01 -1.46552712e-01 -2.92104274e-01 -9.60790336e-01
-2.19005153e-01 8.78447831e-01 6.17235839e-01 5.65789342e-01
-8.70644212e-01 1.11569560e+00 7.54310942e+00 9.57964659e-01
-1.00521255e+00 1.91771597e-01 3.01177561e-01 -5.38923502e-01
-6.80698216e-01 1.26222044e-01 -1.01135600e+00 6.05210483e-01
1.27629817e+00 -2.22050324e-01 2.35473022e-01 8.03012788e-01
-1.35908589e-01 4.36383516e-01 -1.50905585e+00 5.77641785e-01
-1.79452017e-01 -1.24819171e+00 1.23921350e-01 2.71514803e-01
4.98536974e-01 9.12324011e-01 4.24823537e-03 5.13788342e-01
1.07022750e+00 -9.39168811e-01 7.64765739e-01 -9.52752754e-02
1.13288128e+00 -6.67589664e-01 6.08091116e-01 2.66902417e-01
-6.80657446e-01 1.40850604e-01 -4.49822873e-01 1.63601920e-01
2.36057237e-01 7.85325110e-01 -1.29706681e+00 5.66078186e-01
5.24582088e-01 5.38602412e-01 -8.12812567e-01 6.06611490e-01
-4.50139850e-01 1.13519013e+00 -4.73162770e-01 -2.21560821e-01
3.18782896e-01 -2.44635612e-01 5.69111645e-01 1.86535990e+00
3.59690845e-01 -6.64116979e-01 -3.31259727e-01 7.20749915e-01
-2.62414992e-01 4.80257601e-01 -7.00837970e-01 -2.08274931e-01
7.54025638e-01 1.28622794e+00 2.96079274e-02 -4.90735322e-01
-5.14871478e-01 8.94161582e-01 7.01516092e-01 7.35961571e-02
-6.85818493e-01 -4.65366930e-01 1.15240204e+00 -9.39827561e-02
4.43275012e-02 7.63913468e-02 -2.55638957e-01 -1.35200787e+00
1.30346626e-01 -1.48470938e+00 5.13053417e-01 -2.35766008e-01
-1.51201832e+00 8.98044527e-01 -2.30093539e-01 -1.18264997e+00
-6.78810477e-01 -8.01568747e-01 -6.17702007e-01 1.01480281e+00
-1.43372667e+00 -1.10844481e+00 3.46860409e-01 3.90822500e-01
5.97868502e-01 -4.38813448e-01 1.11331642e+00 2.84110397e-01
-6.48684084e-01 1.16897511e+00 3.03132266e-01 2.41037846e-01
1.07692659e+00 -1.41480148e+00 1.24341357e+00 7.69766688e-01
8.84871304e-01 1.03150761e+00 7.29138374e-01 -6.49776757e-01
-1.35641658e+00 -1.32473791e+00 1.18091643e+00 -1.13184416e+00
8.21635067e-01 -5.53199351e-01 -5.98016322e-01 1.00113213e+00
2.66016692e-01 -5.77028275e-01 1.19687867e+00 8.66429806e-01
-8.54399979e-01 1.87870279e-01 -8.46775770e-01 6.36511862e-01
1.28695214e+00 -6.45967603e-01 -1.05159223e+00 4.34288353e-01
5.16390800e-01 -7.91054815e-02 -1.05996978e+00 3.27061445e-01
6.27493680e-01 -5.71816325e-01 7.12740183e-01 -1.13398945e+00
6.94416225e-01 4.74109128e-02 -3.43183756e-01 -1.72525454e+00
-2.48815477e-01 -4.39359605e-01 2.91661285e-02 1.18392754e+00
1.16287434e+00 -4.92914826e-01 7.91324019e-01 6.00461029e-02
-1.89343169e-01 -7.58736730e-01 -6.71988070e-01 -1.06055069e+00
4.49445426e-01 -8.73359680e-01 6.58222914e-01 1.17441130e+00
4.34642911e-01 6.92601383e-01 -1.80622846e-01 -2.08789453e-01
7.36398339e-01 8.07986110e-02 8.35703194e-01 -9.71424699e-01
-8.54678094e-01 -4.85291094e-01 -3.80985469e-01 -7.79520988e-01
1.35727793e-01 -1.51976836e+00 -2.15298936e-01 -1.19503880e+00
4.07878518e-01 -7.10251868e-01 -2.00670689e-01 6.62483275e-01
-4.63632107e-01 5.51854432e-01 4.29952005e-03 3.14987510e-01
-3.93144667e-01 2.52008468e-01 6.55151546e-01 -2.43894994e-01
5.92687353e-02 -4.00706708e-01 -9.93115008e-01 5.58332026e-01
7.08645582e-01 -7.49455512e-01 -2.87345052e-01 -9.26290870e-01
3.81042272e-01 -4.69221354e-01 -2.57893026e-01 -5.85214078e-01
-1.77110448e-01 -1.27911717e-01 3.00610125e-01 -2.95054436e-01
6.82105124e-02 -3.13394248e-01 1.29458070e-01 3.35834533e-01
-3.01764697e-01 4.86786783e-01 1.89923480e-01 9.07217264e-02
-9.44735184e-02 -3.85113984e-01 4.03785378e-01 5.13404123e-02
-5.42242169e-01 1.04864590e-01 -2.03403637e-01 2.10240677e-01
7.15320766e-01 -2.17923284e-01 -4.39816624e-01 1.23496577e-02
-4.05200511e-01 1.33215800e-01 6.49394751e-01 6.29108250e-01
2.70804912e-01 -1.28872406e+00 -1.13538122e+00 2.82806605e-01
6.50019467e-01 -6.59473240e-01 -7.10357070e-01 5.75960398e-01
-3.84122938e-01 2.23585397e-01 -2.16422193e-02 -4.28701907e-01
-1.38543177e+00 2.99037606e-01 2.46221330e-02 -6.33178949e-01
-5.11916280e-01 7.32787251e-01 -5.45584150e-02 -8.52354825e-01
-2.45804936e-01 -5.15838340e-02 3.33240092e-01 -2.01111287e-01
6.46348178e-01 1.91317573e-01 5.26395380e-01 -1.81058168e-01
-3.19381684e-01 2.60108858e-01 -5.59267461e-01 -5.16686201e-01
1.43182862e+00 1.13641143e-01 1.12367652e-01 2.54347682e-01
1.26638174e+00 4.99572963e-01 -7.26752520e-01 -1.91877261e-01
5.54474294e-01 -2.65924156e-01 -2.25889191e-01 -1.12573779e+00
-5.20272434e-01 6.77863002e-01 4.01550710e-01 -3.94636512e-01
6.52033508e-01 7.25998282e-02 8.54798615e-01 6.75355792e-01
4.38536048e-01 -1.23956025e+00 -2.51937747e-01 5.56940973e-01
5.95205486e-01 -1.28628206e+00 -9.88484696e-02 -1.21209174e-01
-4.68490958e-01 9.66876447e-01 3.52376997e-01 2.07501054e-01
3.19566607e-01 7.12266982e-01 6.18445396e-01 -2.33145431e-01
-1.09438825e+00 1.12597629e-01 6.90505803e-02 4.79555428e-01
9.87639248e-01 1.63743377e-01 -5.07180572e-01 3.96023154e-01
-1.05866277e+00 -3.98439795e-01 9.72854644e-02 7.16389656e-01
-2.69695908e-01 -1.80941892e+00 -1.11651503e-01 8.86985719e-01
-6.15161717e-01 -8.15541029e-01 -6.38441682e-01 8.87834311e-01
-9.59075391e-02 9.58965778e-01 -1.05743241e-02 -7.12465346e-01
3.62153977e-01 2.69357532e-01 5.81100523e-01 -8.45826924e-01
-8.63306344e-01 -3.66045147e-01 8.24358225e-01 -4.97573346e-01
3.07292849e-01 -8.01487148e-01 -7.08592594e-01 -5.20674646e-01
-6.63284063e-01 3.01037431e-01 8.20956051e-01 8.19253922e-01
3.08517754e-01 3.12534831e-02 4.84112144e-01 -1.82186007e-01
-1.07629085e+00 -1.47595286e+00 -1.34056106e-01 7.91456342e-01
-6.95465580e-02 -1.35966182e-01 -2.37613142e-01 -1.56647250e-01] | [11.533024787902832, 10.259288787841797] |
e6ff2e46-3fce-4c69-ac5c-a9c552340edf | plan-then-generate-controlled-data-to-text | 2108.13740 | null | https://arxiv.org/abs/2108.13740v1 | https://arxiv.org/pdf/2108.13740v1.pdf | Plan-then-Generate: Controlled Data-to-Text Generation via Planning | Recent developments in neural networks have led to the advance in data-to-text generation. However, the lack of ability of neural models to control the structure of generated output can be limiting in certain real-world applications. In this study, we propose a novel Plan-then-Generate (PlanGen) framework to improve the controllability of neural data-to-text models. Extensive experiments and analyses are conducted on two benchmark datasets, ToTTo and WebNLG. The results show that our model is able to control both the intra-sentence and inter-sentence structure of the generated output. Furthermore, empirical comparisons against previous state-of-the-art methods show that our model improves the generation quality as well as the output diversity as judged by human and automatic evaluations. | ['Nigel Collier', 'Yimai Fang', 'Sihui Wang', 'David Vandyke', 'Yixuan Su'] | 2021-08-31 | null | https://aclanthology.org/2021.findings-emnlp.76 | https://aclanthology.org/2021.findings-emnlp.76.pdf | findings-emnlp-2021-11 | ['data-to-text-generation'] | ['natural-language-processing'] | [ 4.02988523e-01 4.29714233e-01 -9.36524104e-03 -3.66925150e-01
-5.91532350e-01 -6.88309669e-01 9.86758411e-01 -5.19671850e-02
-9.31289271e-02 1.09179962e+00 5.04236102e-01 -1.62052721e-01
1.15287185e-01 -1.06895816e+00 -5.99201202e-01 -3.75711769e-01
3.70746493e-01 4.85532373e-01 -1.26744092e-01 -5.59004188e-01
2.67195106e-01 1.07375443e-01 -1.67489338e+00 8.58736277e-01
1.24098313e+00 8.42129767e-01 3.87439907e-01 6.68879986e-01
-2.23770097e-01 7.42074788e-01 -1.07029796e+00 -3.00913185e-01
2.22263206e-03 -8.06855977e-01 -5.90895176e-01 -1.13835484e-01
2.85902143e-01 -9.21588913e-02 -8.36284235e-02 8.90805006e-01
8.95024776e-01 1.71911672e-01 6.75109267e-01 -9.84489083e-01
-1.06793296e+00 1.28571153e+00 4.63435601e-04 -1.80971563e-01
3.94928157e-01 3.85038316e-01 9.13398027e-01 -7.87537038e-01
9.84799922e-01 1.32995331e+00 2.45370567e-01 8.70472014e-01
-1.34542751e+00 -4.87555325e-01 1.12018220e-01 -1.86988726e-01
-9.64688063e-01 -5.88816702e-01 6.69422030e-01 -3.32933307e-01
1.21715915e+00 2.91102618e-01 6.06286228e-01 1.33372116e+00
4.15761769e-01 8.78288925e-01 8.63341630e-01 -6.78709745e-01
2.90326685e-01 1.65933911e-02 -3.32153171e-01 5.46280861e-01
1.92918614e-01 3.43362123e-01 -8.24314177e-01 3.01930606e-01
5.27592540e-01 -7.63568044e-01 -2.90382057e-01 6.02858402e-02
-1.17747033e+00 9.67985094e-01 3.55876982e-01 4.47611511e-01
-2.83345163e-01 1.18525729e-01 3.73229980e-01 2.53330797e-01
5.40206671e-01 9.69386220e-01 -3.71024072e-01 -2.85799950e-01
-9.91257191e-01 7.08705902e-01 8.25760007e-01 1.25785494e+00
1.30182713e-01 5.05508304e-01 -1.19198155e+00 7.25283563e-01
-4.79089729e-02 4.57007438e-01 9.85145926e-01 -6.10959113e-01
1.07740676e+00 8.86574507e-01 1.90891977e-02 -8.53710771e-01
-3.32604289e-01 -4.51610744e-01 -1.03620923e+00 1.29853636e-01
1.71527520e-01 -5.36083996e-01 -1.04086483e+00 1.81171334e+00
-1.12203926e-01 -5.90405285e-01 1.76100239e-01 5.55483997e-01
8.92770052e-01 8.58831882e-01 -1.06570095e-01 -2.08742663e-01
7.80434847e-01 -1.12756383e+00 -9.14666831e-01 -2.78813154e-01
6.80537820e-01 -6.45096242e-01 1.23702741e+00 3.49575818e-01
-1.30503035e+00 -9.28533077e-01 -8.44303131e-01 1.48257330e-01
-4.90907401e-01 5.59547305e-01 3.90380234e-01 6.08237803e-01
-1.02378571e+00 6.78028941e-01 -5.42470634e-01 -1.09633334e-01
4.19124007e-01 2.84577787e-01 -9.96065214e-02 2.57313371e-01
-1.44197810e+00 7.82362163e-01 9.99219537e-01 6.22363612e-02
-8.39699209e-01 -5.26364326e-01 -7.79168069e-01 2.03221053e-01
2.50905663e-01 -9.32608247e-01 1.63606238e+00 -8.97212148e-01
-1.89298046e+00 3.56411040e-01 5.25232963e-02 -4.95743781e-01
5.84626317e-01 -1.77917019e-01 -2.61029005e-01 -1.97581962e-01
-8.91367532e-03 1.14792442e+00 4.85003144e-01 -1.16216457e+00
-6.84093773e-01 5.96061088e-02 2.56079268e-02 3.36656511e-01
-5.48724592e-01 -3.02415162e-01 -3.23866189e-01 -8.68329465e-01
-3.75137597e-01 -8.47971022e-01 -2.35037401e-01 -4.91380095e-01
-9.27529693e-01 -4.63368505e-01 3.41263294e-01 -3.62829477e-01
1.61499429e+00 -1.75306654e+00 3.03160459e-01 -2.41266280e-01
-3.04682314e-01 3.93666744e-01 -4.85126764e-01 8.40786874e-01
-4.03471403e-02 4.98561203e-01 -1.01865761e-01 -4.38541859e-01
2.12926880e-01 4.99464273e-02 -5.68286896e-01 -3.70016932e-01
4.63074505e-01 1.07275021e+00 -8.56159568e-01 -3.77383292e-01
-1.57473892e-01 1.60660192e-01 -6.22177064e-01 6.66459858e-01
-7.96895385e-01 3.03908050e-01 -5.03579378e-01 1.86453819e-01
1.44558147e-01 -4.64351028e-02 1.13493606e-01 2.64097035e-01
-1.69127569e-01 4.76354718e-01 -7.63913274e-01 1.80767119e+00
-5.27437747e-01 5.93228281e-01 -6.08120441e-01 -4.50964242e-01
9.57960486e-01 4.73281384e-01 -2.13680163e-01 -9.35197294e-01
2.05167398e-01 1.06704710e-02 2.74078310e-01 -4.89126205e-01
9.82784271e-01 1.16307691e-01 -2.57483542e-01 5.11628926e-01
6.57439902e-02 -5.72338402e-01 9.76634502e-01 1.22542560e-01
6.70237422e-01 3.79198611e-01 1.27062887e-01 -2.69480705e-01
3.31996888e-01 6.24846183e-02 3.61051172e-01 9.08215344e-01
4.63719279e-01 8.12335193e-01 6.84195578e-01 -2.05792367e-01
-1.09374726e+00 -8.08759511e-01 2.17713088e-01 1.10307813e+00
-3.99353057e-01 -3.91415507e-01 -1.37111223e+00 -7.22259581e-01
-2.24702820e-01 1.47569406e+00 -6.50226116e-01 -2.85651296e-01
-5.85344672e-01 -6.74237370e-01 8.23907912e-01 5.83784461e-01
3.31439137e-01 -1.73542619e+00 -5.58020234e-01 4.07299370e-01
-3.41456354e-01 -1.00723255e+00 -4.49608564e-01 -1.72390528e-02
-8.97718072e-01 -5.34160852e-01 -5.72046101e-01 -6.19857848e-01
6.96439028e-01 -3.81354272e-01 1.38822174e+00 -8.78035203e-02
5.04099838e-02 -3.27697515e-01 -4.72213537e-01 -6.33017361e-01
-1.01347828e+00 7.32876837e-01 -2.26111248e-01 -2.52704024e-01
-1.55984327e-01 -2.77291477e-01 -2.31887579e-01 1.18514113e-01
-1.36550283e+00 6.41116679e-01 5.98810196e-01 8.59013677e-01
3.63243312e-01 1.62724242e-01 8.55156362e-01 -9.84274685e-01
1.58612299e+00 -1.77164376e-01 -6.17467165e-01 3.06655973e-01
-9.04247105e-01 4.82176363e-01 9.63028848e-01 -4.48904932e-01
-1.29383862e+00 2.81706881e-02 -5.24546802e-02 -6.33082166e-02
-2.83834845e-01 7.82865703e-01 -3.38661581e-01 6.43173695e-01
8.56907547e-01 4.55878079e-01 -2.87136108e-01 -2.00755268e-01
4.94064718e-01 7.09005058e-01 3.08114111e-01 -6.12817585e-01
7.32168138e-01 -1.87878877e-01 -1.91431850e-01 -3.77928466e-01
-8.40820491e-01 4.95988548e-01 -4.21933651e-01 -8.69057924e-02
7.00507283e-01 -7.66430676e-01 -3.50286931e-01 5.28169036e-01
-1.56749964e+00 -7.21055686e-01 -3.53841603e-01 -1.05824389e-01
-7.37479925e-01 -1.98197603e-01 -5.67264020e-01 -7.39508331e-01
-8.13843966e-01 -1.29890120e+00 1.03606570e+00 3.24703187e-01
-5.04231334e-01 -9.58903134e-01 1.39485329e-01 1.04060665e-01
5.82483411e-01 2.80148000e-01 1.09427536e+00 -7.27673650e-01
-4.77190137e-01 -2.06241772e-01 4.77791280e-02 2.55429149e-01
-7.53318667e-02 7.77066201e-02 -7.83479452e-01 4.50953934e-03
-3.92585307e-01 -5.25652766e-01 6.60029233e-01 2.20087960e-01
1.20167899e+00 -6.14173472e-01 -3.05049837e-01 3.00163656e-01
1.14972103e+00 2.80575275e-01 8.37262690e-01 1.40640259e-01
5.22527635e-01 7.41622686e-01 6.30001605e-01 4.83593017e-01
2.34176636e-01 7.65774488e-01 3.87307763e-01 1.18121371e-01
-3.56089503e-01 -7.76058435e-01 4.57570702e-01 7.64479041e-01
-1.73817854e-02 -1.07239449e+00 -7.24452972e-01 6.12836003e-01
-1.91683435e+00 -9.76717234e-01 -2.07123294e-01 1.88940310e+00
1.28037858e+00 2.25192606e-01 -2.25709394e-01 -4.02274951e-02
6.67852461e-01 1.12021185e-01 -3.48099142e-01 -6.14605725e-01
-1.49935126e-01 1.77540407e-01 -2.48915497e-02 3.02978903e-01
-8.31797779e-01 1.26730800e+00 7.13687277e+00 1.08889949e+00
-9.73347485e-01 -3.45023274e-01 8.76337528e-01 -3.16994399e-01
-5.16833723e-01 -4.44404662e-01 -1.10209429e+00 5.24639249e-01
1.06660223e+00 -4.18432117e-01 5.49941063e-01 8.25800240e-01
5.11330247e-01 1.32850289e-01 -1.29790056e+00 5.50063014e-01
1.95301205e-01 -1.52754128e+00 6.17448151e-01 -1.80725532e-03
1.21966803e+00 -3.25735837e-01 -9.22593754e-03 5.05629539e-01
5.10310829e-01 -1.29386508e+00 1.10812294e+00 6.90454543e-01
9.33935285e-01 -9.22451556e-01 7.25916743e-01 6.68707490e-01
-7.84185588e-01 -5.54656126e-02 -2.58889347e-01 -1.52732387e-01
3.07006776e-01 7.24134326e-01 -1.11183989e+00 5.62041044e-01
2.08552822e-01 2.83984005e-01 -6.74831986e-01 5.22723854e-01
-7.41299987e-01 4.42422867e-01 1.78122222e-01 -5.76827228e-01
2.29771703e-01 7.74854422e-02 4.79191631e-01 1.23411715e+00
6.35313749e-01 -2.18997449e-01 5.32917567e-02 1.33826017e+00
-3.14108342e-01 7.14926645e-02 -8.23998332e-01 -5.29990911e-01
3.87615263e-01 1.10146224e+00 -2.49351695e-01 -3.60966057e-01
3.21097881e-01 9.34850037e-01 5.19959927e-01 3.42395723e-01
-6.08970523e-01 -6.14353716e-01 2.71786869e-01 8.96254927e-02
2.57434756e-01 -1.47696748e-01 -5.82329571e-01 -9.29526389e-01
1.62900299e-01 -1.26343334e+00 -2.36300807e-02 -9.16947663e-01
-1.07887948e+00 1.10637999e+00 4.57218587e-02 -9.83371019e-01
-9.91285682e-01 -4.68206406e-01 -6.58079028e-01 9.89215314e-01
-1.18344796e+00 -1.13217771e+00 -7.06231445e-02 1.97609574e-01
7.92682171e-01 -6.66386902e-01 1.01561582e+00 -1.60982117e-01
-5.98902404e-01 7.09515393e-01 -1.43555366e-02 1.05724260e-01
3.74354452e-01 -1.17418909e+00 8.29454482e-01 8.57921898e-01
6.73347190e-02 5.71595788e-01 7.45650053e-01 -6.63341165e-01
-1.22037983e+00 -1.23275471e+00 1.25768960e+00 -3.45721692e-01
3.81716877e-01 -5.44305742e-01 -3.57590258e-01 3.09294224e-01
8.02592456e-01 -5.62057376e-01 4.95573580e-01 -6.39357269e-02
-1.80588011e-02 7.27313235e-02 -9.06489968e-01 8.86113763e-01
1.06955290e+00 -4.98083122e-02 -4.06691015e-01 3.13639045e-01
1.03690624e+00 -7.60732234e-01 -5.26242554e-01 2.87389338e-01
3.69721830e-01 -1.03603935e+00 4.70351875e-01 -7.98859656e-01
1.29716909e+00 -1.13126308e-01 5.21339141e-02 -2.07503366e+00
-2.14553058e-01 -6.91117525e-01 -1.35260656e-01 1.50695240e+00
8.72502923e-01 -2.43418649e-01 6.52371645e-01 5.56167305e-01
-5.31014144e-01 -8.39212239e-01 -6.92549765e-01 -6.57492459e-01
2.70328134e-01 -2.65324950e-01 8.36752713e-01 4.77374136e-01
7.34054446e-02 6.42156780e-01 -2.14247838e-01 -3.92827213e-01
9.56250057e-02 1.57103911e-01 7.76715040e-01 -1.01152837e+00
-4.18290734e-01 -6.39691949e-01 2.08233565e-01 -8.85839880e-01
1.86516657e-01 -9.55775261e-01 3.74370754e-01 -1.86038268e+00
-1.09826908e-01 -1.16858400e-01 7.78121054e-02 6.04911685e-01
-2.88380861e-01 -5.43861277e-02 4.31139797e-01 -2.17905730e-01
-2.65001178e-01 9.36883330e-01 1.58495069e+00 -1.60239860e-01
-4.03368950e-01 -8.11773688e-02 -8.40897560e-01 2.59601891e-01
1.02391672e+00 -5.70659518e-01 -5.57750463e-01 -8.39882553e-01
4.71669108e-01 7.84384608e-02 -2.10437804e-01 -1.16182852e+00
-3.84651423e-02 -3.79421592e-01 3.59704316e-01 -5.36156118e-01
4.32681181e-02 -3.14958990e-01 -4.78166044e-02 4.37185168e-01
-9.72968519e-01 2.41464287e-01 3.36299032e-01 2.22811371e-01
-2.74134487e-01 -4.74448442e-01 4.51448053e-01 -1.89287871e-01
-4.46825922e-02 8.22332129e-02 -4.75727826e-01 1.50556073e-01
7.59003162e-01 1.42312899e-01 -6.34050190e-01 -6.12388670e-01
-1.70881227e-01 1.65556401e-01 6.61564618e-02 7.68197298e-01
5.38366616e-01 -1.45940590e+00 -9.12663341e-01 1.75096780e-01
2.85064995e-01 3.60576622e-02 -4.50616702e-02 8.31733495e-02
-3.67301345e-01 9.30790007e-01 -1.73006818e-01 -2.67148644e-01
-9.73826230e-01 3.03181529e-01 3.10214043e-01 -8.32510591e-01
-9.68703032e-02 6.59393251e-01 -8.12159926e-02 -7.28606701e-01
2.50340641e-01 -4.07344609e-01 -2.66822278e-01 5.07533401e-02
5.15603423e-01 1.90446034e-01 2.10307958e-03 -2.01960579e-01
3.60221803e-01 -5.28663099e-02 -1.00401305e-01 -3.82081389e-01
1.25720644e+00 3.08090270e-01 -8.57301243e-03 3.35969180e-01
6.80052459e-01 -2.78897613e-01 -9.89090085e-01 2.06029668e-01
-1.73222542e-01 -1.86178043e-01 -7.76510388e-02 -1.25409722e+00
-9.89331961e-01 9.86193597e-01 2.72521526e-01 4.99005854e-01
1.00733685e+00 -5.38396060e-01 7.22967327e-01 6.09932780e-01
2.72671461e-01 -1.35832620e+00 3.00775707e-01 8.18786621e-01
1.32348609e+00 -7.94785321e-01 -3.56379658e-01 -2.62655437e-01
-9.61864471e-01 1.11811125e+00 8.77685666e-01 1.89387769e-01
-8.12493339e-02 2.77424306e-01 1.28052965e-01 2.64306039e-01
-1.23229170e+00 4.62682135e-02 4.69730735e-01 6.55067980e-01
9.11382675e-01 4.33450006e-02 -3.65963370e-01 4.99364197e-01
-7.07001150e-01 2.52515554e-01 5.57368755e-01 6.00753129e-01
-3.64779741e-01 -1.47239923e+00 -1.60480723e-01 3.23691696e-01
-3.12566102e-01 -3.77349228e-01 -7.36330152e-01 6.70991242e-01
7.44998753e-02 1.21039832e+00 7.42283370e-03 -3.71614784e-01
4.66676265e-01 2.36636654e-01 2.85783112e-01 -9.28752124e-01
-1.05013347e+00 8.36534239e-03 5.34308255e-01 -2.35784784e-01
4.62759212e-02 -3.51385504e-01 -1.25065517e+00 -2.73205396e-02
-3.25146824e-01 5.62768430e-02 7.57309556e-01 8.41707945e-01
7.09311903e-01 9.70876396e-01 5.15276134e-01 -6.86513305e-01
-8.29645932e-01 -1.55086708e+00 -2.69961566e-01 4.84633714e-01
-1.33121386e-01 -1.48011431e-01 1.12968413e-02 2.67145991e-01] | [11.677658081054688, 9.026005744934082] |
a1ce9495-7e50-4256-851e-d41f9ff9e7c8 | fast-kinodynamic-planning-on-the-constraint | 2301.04330 | null | https://arxiv.org/abs/2301.04330v2 | https://arxiv.org/pdf/2301.04330v2.pdf | Fast Kinodynamic Planning on the Constraint Manifold with Deep Neural Networks | Motion planning is a mature area of research in robotics with many well-established methods based on optimization or sampling the state space, suitable for solving kinematic motion planning. However, when dynamic motions under constraints are needed and computation time is limited, fast kinodynamic planning on the constraint manifold is indispensable. In recent years, learning-based solutions have become alternatives to classical approaches, but they still lack comprehensive handling of complex constraints, such as planning on a lower-dimensional manifold of the task space while considering the robot's dynamics. This paper introduces a novel learning-to-plan framework that exploits the concept of constraint manifold, including dynamics, and neural planning methods. Our approach generates plans satisfying an arbitrary set of constraints and computes them in a short constant time, namely the inference time of a neural network. This allows the robot to plan and replan reactively, making our approach suitable for dynamic environments. We validate our approach on two simulated tasks and in a demanding real-world scenario, where we use a Kuka LBR Iiwa 14 robotic arm to perform the hitting movement in robotic Air Hockey. | ['Jan Peters', 'Piotr Skrzypczyński', 'Krzysztof Walas', 'Haitham Bou-Ammar', 'Davide Tateo', 'Puze Liu', 'Piotr Kicki'] | 2023-01-11 | null | null | null | null | ['motion-planning'] | ['robots'] | [ 1.49968714e-01 1.81465492e-01 -4.14975911e-01 8.12372640e-02
-1.48474351e-01 -4.18642938e-01 5.05896032e-01 -1.01660348e-01
-7.92082250e-01 7.29665041e-01 -1.53780773e-01 -2.97139108e-01
-6.06376052e-01 -5.66024840e-01 -6.18339360e-01 -7.51231670e-01
-3.33150268e-01 8.53632271e-01 3.93464714e-01 -5.31388640e-01
4.88056034e-01 7.90426075e-01 -1.40148807e+00 -3.27125371e-01
6.49040163e-01 5.03048182e-01 8.57486486e-01 5.93446851e-01
2.94465959e-01 5.31049430e-01 -5.93384616e-02 4.99310762e-01
4.78783786e-01 -3.57415944e-01 -9.87219274e-01 3.15881670e-01
-3.73433739e-01 9.89128575e-02 -4.52656746e-01 7.00598061e-01
3.24790299e-01 8.50778759e-01 6.72674000e-01 -1.10367644e+00
3.13003242e-01 4.66009974e-01 -8.44889730e-02 -1.71929643e-01
3.66478473e-01 4.39723462e-01 6.16109073e-01 -4.14943069e-01
8.81139219e-01 1.05682552e+00 4.74584877e-01 9.05554831e-01
-8.77449214e-01 -6.25417195e-03 1.59573510e-01 4.02100086e-01
-1.36444938e+00 -1.66043341e-01 6.75044239e-01 -3.96679878e-01
9.99897063e-01 -7.42741898e-02 7.18698680e-01 8.14471602e-01
8.12347829e-01 6.20205581e-01 6.92555249e-01 -3.32866997e-01
5.98874986e-01 -4.36581939e-01 -3.49963009e-01 8.32709730e-01
-3.53952199e-02 4.28793788e-01 -3.59945536e-01 6.37240931e-02
8.40287209e-01 -4.89452370e-02 -2.55522519e-01 -8.29465151e-01
-1.46646619e+00 7.62924016e-01 5.03084660e-01 2.16080174e-01
-6.50135398e-01 3.90721112e-01 1.35074064e-01 1.51157334e-01
-2.64026016e-01 7.15785086e-01 -3.89159828e-01 -1.53113648e-01
-6.29670441e-01 6.03674352e-01 1.01751232e+00 1.18061531e+00
7.03565955e-01 9.62030366e-02 9.12308395e-02 3.13293606e-01
3.28310788e-01 2.08133131e-01 4.84728694e-01 -1.27829218e+00
5.24745584e-01 4.64777797e-01 4.99733865e-01 -9.88733470e-01
-1.07872820e+00 9.46382359e-02 -5.41410863e-01 6.32034779e-01
5.65533102e-01 -5.24263799e-01 -7.73315012e-01 1.48559666e+00
5.75126886e-01 -6.06980175e-02 3.00886035e-01 1.31481314e+00
-2.33624071e-01 6.45402551e-01 -2.48820558e-01 -3.13791722e-01
9.70936358e-01 -1.11344802e+00 -6.03688180e-01 -4.97303098e-01
7.61907876e-01 -3.81823748e-01 8.90736163e-01 6.00093424e-01
-1.10125399e+00 -3.99799943e-01 -1.07926536e+00 1.55789942e-01
-4.10835557e-02 -4.00782116e-02 5.80612302e-01 -3.24918628e-02
-8.66973460e-01 9.47727084e-01 -1.24049354e+00 -7.09292591e-01
-1.02731369e-01 5.79573691e-01 -3.28049242e-01 -9.36860666e-02
-8.04861963e-01 1.53898740e+00 7.57394433e-01 4.45695370e-01
-1.03439426e+00 1.91759150e-02 -8.07978213e-01 -2.29105353e-01
7.34869003e-01 -5.87286592e-01 1.39598346e+00 -4.24342275e-01
-2.17735314e+00 1.42505661e-01 1.89064533e-01 -3.46331596e-01
5.61486185e-01 -3.07077318e-01 2.10785031e-01 2.90786829e-02
-1.48276433e-01 6.20045066e-01 6.21170342e-01 -9.98225689e-01
-3.70547295e-01 -1.33819744e-01 1.01660416e-01 6.67545259e-01
1.51067704e-01 -4.51573282e-01 -3.32129449e-01 -9.75225642e-02
5.02190053e-01 -1.64420009e+00 -9.99692798e-01 -1.34063110e-01
-2.38048345e-01 -4.23332423e-01 4.94260341e-01 -1.19369544e-01
8.64790320e-01 -1.79738379e+00 9.96674418e-01 2.57207006e-01
-4.13243055e-01 8.55363905e-02 -6.00077957e-02 7.38102853e-01
3.88427585e-01 -3.79578561e-01 -5.58443427e-01 -5.15944697e-02
-4.83869575e-02 5.50978303e-01 -1.20030716e-01 6.54976726e-01
6.82274625e-02 6.17388546e-01 -1.06903112e+00 -3.95367801e-01
1.49428204e-01 5.34249432e-02 -6.82612658e-01 1.66132838e-01
-7.79786110e-01 8.31251204e-01 -8.55458617e-01 2.05481291e-01
5.58531694e-02 3.33802700e-01 4.73778754e-01 3.98202240e-01
-5.91748893e-01 -5.28391358e-03 -1.35996890e+00 2.27169108e+00
-6.28876984e-01 4.27208394e-01 2.51515239e-01 -1.14340651e+00
9.35661852e-01 1.69830978e-01 5.49445331e-01 -2.23655432e-01
4.13020700e-01 4.04905200e-01 6.87620863e-02 -7.74124444e-01
6.20662272e-01 -8.12769756e-02 -2.63213694e-01 3.96371365e-01
-2.55340159e-01 -7.30649829e-01 1.72951475e-01 -2.59089291e-01
1.30515075e+00 7.41047144e-01 3.56642783e-01 -3.68966997e-01
5.36144495e-01 7.83140302e-01 4.41348672e-01 4.10906255e-01
-8.67279321e-02 3.54168802e-01 1.45105079e-01 -5.45601130e-01
-9.17499483e-01 -5.70869029e-01 2.47341827e-01 5.72103858e-01
5.64978659e-01 -5.71726896e-02 -7.65315711e-01 -2.78057247e-01
-2.21611455e-01 6.44936502e-01 -1.94905803e-01 -1.03292994e-01
-1.20209932e+00 -3.55974913e-01 3.90651077e-02 1.55163318e-01
2.22412318e-01 -1.60019314e+00 -1.56242394e+00 6.73686802e-01
-1.79340690e-01 -1.10122395e+00 -3.80293161e-01 2.17701003e-01
-1.11240709e+00 -1.29720402e+00 -4.70104754e-01 -8.75495374e-01
6.77975476e-01 2.84591496e-01 4.80721176e-01 2.50311553e-01
-3.18183511e-01 4.06705469e-01 -2.60336399e-01 -1.70877710e-01
-4.06398565e-01 3.10747981e-01 3.46481860e-01 -2.72562861e-01
-2.36627206e-01 -3.62392545e-01 -2.27982789e-01 5.08557916e-01
-7.40617633e-01 7.65389130e-02 7.76115477e-01 7.78436482e-01
7.36782312e-01 2.99917787e-01 4.89428461e-01 -2.67565608e-01
5.27055383e-01 -4.88347083e-01 -8.57413411e-01 -2.15059165e-02
-4.53159630e-01 2.96061575e-01 5.67471862e-01 -6.52026236e-01
-1.08099926e+00 8.73174906e-01 1.86229616e-01 -3.36568594e-01
-9.95506495e-02 8.25313210e-01 -3.38042043e-02 -1.30250916e-01
6.95661604e-01 8.36657509e-02 2.43460029e-01 -1.80887476e-01
4.20763969e-01 3.35685700e-01 5.71940958e-01 -6.21853650e-01
7.84590781e-01 4.18303251e-01 6.19425356e-01 -7.56574154e-01
-3.64955455e-01 -5.48884213e-01 -1.04522800e+00 -4.79496419e-01
9.65640724e-01 -3.51651460e-01 -9.95382011e-01 2.16400385e-01
-1.36819637e+00 -8.25993836e-01 -2.15142682e-01 8.30971360e-01
-1.37483549e+00 3.07548046e-01 -3.93284142e-01 -8.97427917e-01
5.84906116e-02 -1.26191640e+00 6.73998356e-01 1.94989517e-01
-3.82948995e-01 -8.01992178e-01 2.97444761e-01 -1.43274635e-01
3.15180004e-01 6.12773657e-01 6.78570867e-01 -1.53914586e-01
-7.52044141e-01 -2.79329121e-01 4.94630605e-01 -3.62475395e-01
-1.30403817e-01 -3.77561361e-01 -3.17360878e-01 -2.67015159e-01
1.57303318e-01 -3.24054450e-01 3.90573144e-01 3.09927344e-01
6.68383956e-01 -2.99387813e-01 -6.40327632e-01 1.86594903e-01
1.38458991e+00 2.49425441e-01 5.37465155e-01 5.87454975e-01
4.51348513e-01 1.09270656e+00 1.15637457e+00 3.32121730e-01
3.50884706e-01 9.07140970e-01 7.36951947e-01 5.08060455e-01
2.68642157e-01 -1.07352413e-01 5.71697772e-01 6.97335064e-01
-3.56185704e-01 -9.36002061e-02 -1.11037242e+00 4.96716291e-01
-2.38214302e+00 -8.67334604e-01 -1.94010675e-01 2.08906007e+00
6.52573943e-01 7.44194910e-02 -1.19517939e-02 6.08460456e-02
6.44724905e-01 -2.65217215e-01 -7.66754270e-01 -3.93952370e-01
4.45900828e-01 -8.12251568e-02 5.71728647e-01 7.90205836e-01
-8.36325705e-01 1.18863523e+00 5.95622444e+00 4.00953054e-01
-1.06236637e+00 -1.64590776e-01 -2.80130118e-01 -5.01682013e-02
2.28542417e-01 8.28946233e-02 -6.03340030e-01 1.43732443e-01
1.01143336e+00 1.17749060e-02 7.19559789e-01 1.15835238e+00
4.62430328e-01 -5.05398691e-01 -1.15351748e+00 5.17330706e-01
-1.38259560e-01 -1.20384431e+00 -5.03235698e-01 7.83993155e-02
6.57000065e-01 4.09379266e-02 -3.72024775e-01 2.80107290e-01
3.35503697e-01 -9.48688507e-01 9.18255448e-01 6.30740702e-01
2.58514017e-01 -6.65860534e-01 5.98604083e-01 1.13915360e+00
-9.59834516e-01 -3.94333899e-01 -4.94090855e-01 -3.86985540e-01
6.92196012e-01 1.09265231e-01 -9.49551046e-01 5.94691455e-01
1.97461367e-01 4.08193916e-01 1.04532987e-01 1.15959275e+00
-3.06301862e-01 -7.23238885e-02 -4.90811735e-01 -6.00233257e-01
4.58346993e-01 -4.66122925e-01 7.95800030e-01 7.64191329e-01
4.11261320e-01 3.12611133e-01 5.65468907e-01 6.18465006e-01
7.40087390e-01 -1.77341416e-01 -8.69436145e-01 2.81774729e-01
8.73451158e-02 1.24318898e+00 -9.97308373e-01 1.90810159e-01
1.88866585e-01 7.23683119e-01 2.68448055e-01 1.48890495e-01
-7.55086660e-01 -4.04478997e-01 3.51982981e-01 -1.38913766e-01
-1.01473564e-02 -1.04108655e+00 -5.73160164e-02 -8.38855326e-01
-8.02420732e-03 -4.73138928e-01 2.49319468e-02 -5.91997683e-01
-4.75925982e-01 5.27681053e-01 3.92188579e-01 -1.59420276e+00
-6.42462075e-01 -6.00969672e-01 -5.61181545e-01 5.95183671e-01
-1.50245285e+00 -6.39635324e-01 -3.00860614e-01 5.13818264e-01
9.14596856e-01 -7.83704519e-02 7.77074695e-01 -2.08025947e-01
-4.44694817e-01 -3.09751153e-01 -2.72408247e-01 -4.94566470e-01
2.53705293e-01 -8.90833557e-01 6.67684451e-02 6.74789608e-01
-2.58857757e-01 3.17510515e-01 8.65436494e-01 -6.61192060e-01
-2.00865793e+00 -7.50153542e-01 6.41654849e-01 -3.23827475e-01
5.87184846e-01 -9.32019949e-02 -7.29281187e-01 5.26293159e-01
-3.79594788e-02 -1.84166178e-01 -2.75208429e-02 -4.10339773e-01
4.64997232e-01 3.12675864e-01 -9.22455311e-01 1.01313853e+00
1.24987042e+00 2.19682604e-01 -5.31530201e-01 5.78707635e-01
5.95057368e-01 -8.70733678e-01 -4.74463105e-01 3.98106575e-01
4.19228464e-01 -4.87318099e-01 5.56736410e-01 -6.19346023e-01
3.90535444e-02 -7.05476761e-01 8.72474164e-03 -1.39094126e+00
-1.90782696e-01 -1.10263252e+00 -5.09207696e-02 3.97725970e-01
4.35939372e-01 -2.97567725e-01 9.14247751e-01 5.52759826e-01
-5.53701222e-01 -8.67970586e-01 -1.21752536e+00 -8.82596970e-01
-1.23958029e-01 -4.38492417e-01 1.72307581e-01 5.90859532e-01
4.49049205e-01 1.63001597e-01 -3.04737657e-01 3.65232021e-01
3.21452737e-01 -2.62912121e-02 8.65591943e-01 -1.13331008e+00
-2.06179678e-01 -2.74361104e-01 -2.58489460e-01 -1.00862968e+00
5.50767422e-01 -8.12694490e-01 1.01079953e+00 -1.71680450e+00
-4.40521181e-01 -7.65137374e-01 3.34344894e-01 4.04847622e-01
4.50447798e-01 -4.20263976e-01 3.44683558e-01 5.16227305e-01
-3.79342109e-01 6.22325957e-01 1.50683379e+00 1.69033960e-01
-6.02879226e-01 3.11006069e-01 1.63596421e-01 9.02387261e-01
1.05052984e+00 -4.54205424e-01 -5.99420786e-01 -5.09571552e-01
1.95155263e-01 5.57633817e-01 2.32676584e-02 -1.38507247e+00
7.95672059e-01 -8.94901931e-01 -2.55285263e-01 -3.23641449e-01
5.38487554e-01 -1.09252822e+00 2.83579767e-01 1.07307780e+00
-2.43215337e-01 2.72484571e-01 9.38887298e-02 8.33709836e-01
8.19746498e-03 -5.95916629e-01 6.44433558e-01 -2.46365473e-01
-1.03692687e+00 1.82416320e-01 -8.08889210e-01 -2.85775542e-01
1.38728452e+00 -1.99793831e-01 8.67794082e-02 -1.04083762e-01
-8.03704500e-01 4.93891954e-01 3.77198249e-01 3.35548848e-01
6.80519938e-01 -1.01115966e+00 -3.14925194e-01 -7.80577883e-02
-2.15955198e-01 5.07737458e-01 2.16215290e-03 8.46469223e-01
-8.25792611e-01 4.47681129e-01 -4.69185829e-01 -6.46144569e-01
-5.63764513e-01 5.54313838e-01 2.51354724e-01 -2.17645496e-01
-8.83314610e-01 3.66953284e-01 -2.72595465e-01 -6.14992738e-01
6.43992126e-02 -4.04228747e-01 -4.05120522e-01 -2.01959759e-01
6.73137680e-02 4.50353175e-01 -1.57318711e-01 -4.65497047e-01
-3.07113051e-01 7.74784088e-01 6.00798726e-01 -5.65852106e-01
1.21838188e+00 -9.75694507e-02 2.97162607e-02 3.99752259e-01
5.75083911e-01 -4.61068600e-01 -1.53841722e+00 1.60752442e-02
6.44693911e-01 -5.15435152e-02 -1.95668206e-01 -3.12369287e-01
-6.04456425e-01 6.94789708e-01 1.64178073e-01 -1.89709552e-02
7.30778456e-01 -2.58170903e-01 5.56967437e-01 1.02830482e+00
1.09774721e+00 -1.36619127e+00 3.72382373e-01 1.04721677e+00
1.18658829e+00 -8.85640800e-01 8.24211091e-02 -4.28280681e-01
-8.28467548e-01 1.47343755e+00 6.75899744e-01 -4.69278783e-01
4.39600825e-01 6.90152943e-02 3.89222130e-02 1.12962127e-01
-6.94973767e-01 -2.11388782e-01 9.07735154e-02 6.11452103e-01
-1.89538941e-01 -9.95763093e-02 -5.69746435e-01 2.16298953e-01
-2.38266304e-01 -5.45874499e-02 9.29245889e-01 1.37334549e+00
-8.09427381e-01 -1.03876543e+00 -3.10147196e-01 -2.93199450e-01
1.01924554e-01 5.70760965e-01 2.62064375e-02 1.14151132e+00
-1.53255165e-01 9.55503821e-01 -2.89741959e-02 -3.31057638e-01
5.41821003e-01 6.82532191e-02 5.81413090e-01 -9.23600674e-01
-2.28744969e-01 -9.05196592e-02 1.30911395e-01 -9.41956758e-01
-5.33047378e-01 -9.24634159e-01 -2.00667000e+00 6.98697194e-02
-3.30871910e-01 1.42581031e-01 1.09959888e+00 1.12627518e+00
2.21613482e-01 2.93302268e-01 6.01338506e-01 -1.58690834e+00
-8.07202399e-01 -8.18401396e-01 -3.10853362e-01 -1.49921983e-01
1.75598502e-01 -9.66372252e-01 -1.78729415e-01 5.99758215e-02] | [4.850027561187744, 1.2667981386184692] |
cf85a7c8-73c4-4ba7-8755-7e6e5bc832b9 | turning-flowchart-into-dialog-plan-based-data | 2305.01323 | null | https://arxiv.org/abs/2305.01323v2 | https://arxiv.org/pdf/2305.01323v2.pdf | Turning Flowchart into Dialog: Plan-based Data Augmentation for Low-Resource Flowchart-grounded Troubleshooting Dialogs | Flowchart-grounded troubleshooting dialogue (FTD) systems, which follow the instructions of a flowchart to diagnose users' problems in specific domains (eg., vehicle, laptop), have been gaining research interest in recent years. However, collecting sufficient dialogues that are naturally grounded on flowcharts is costly, thus FTD systems are impeded by scarce training data. To mitigate the data sparsity issue, we propose a plan-based data augmentation (PlanDA) approach that generates diverse synthetic dialog data at scale by transforming concise flowchart into dialogues. Specifically, its generative model employs a variational-base framework with a hierarchical planning strategy that includes global and local latent planning variables. Experiments on the FloDial dataset show that synthetic dialogue produced by PlanDA improves the performance of downstream tasks, including flowchart path retrieval and response generation, in particular on the Out-of-Flowchart settings. In addition, further analysis demonstrate the quality of synthetic data generated by PlanDA in paths that are covered by current sample dialogues and paths that are not covered. | ['Gholamreza Haffari', 'YuFei Wang', 'Ingrid Zukerman', 'Lizhen Qu', 'Sameen Maruf', 'Haolan Zhan'] | 2023-05-02 | null | null | null | null | ['response-generation'] | ['natural-language-processing'] | [ 3.55056971e-02 5.28207123e-01 -3.49082053e-01 -5.83678484e-01
-1.05162799e+00 -6.10632718e-01 6.62588775e-01 -1.22670792e-01
1.57733887e-01 9.64573741e-01 8.49649489e-01 -4.46239263e-01
1.44041628e-01 -6.39362574e-01 -1.34626329e-01 -1.58894539e-01
3.77094120e-01 9.47105050e-01 -2.71179467e-01 -5.93527317e-01
6.19993731e-02 -1.68081969e-01 -1.06699669e+00 4.23471391e-01
1.27546239e+00 3.50441903e-01 1.56532779e-01 6.96216881e-01
-4.65337813e-01 8.68654251e-01 -8.86268377e-01 -3.08733553e-01
2.32303385e-02 -8.35786879e-01 -1.12875259e+00 5.44844866e-01
1.57119378e-01 -7.71194041e-01 -5.03217936e-01 5.77958703e-01
3.22028130e-01 5.71959555e-01 5.03027976e-01 -1.57760417e+00
-4.99250561e-01 4.96155769e-01 -8.98902863e-02 -4.04500902e-01
1.01166463e+00 6.05741143e-01 1.35230541e+00 -9.22516882e-01
7.45711148e-01 1.78249717e+00 4.16983247e-01 9.04161870e-01
-1.49242830e+00 -3.45584452e-01 2.31613904e-01 -2.03067407e-01
-7.93543220e-01 -5.84227860e-01 7.15751648e-01 -4.04372096e-01
1.01177192e+00 3.24814349e-01 5.04511595e-01 1.32823074e+00
-1.70483232e-01 1.31083512e+00 6.28622830e-01 -1.33561119e-01
2.00419396e-01 2.58551002e-01 2.08993599e-01 9.46367741e-01
-2.55211473e-01 -3.36878359e-01 -8.54675174e-01 -2.42052138e-01
5.35378754e-01 -1.93759948e-01 -2.76211441e-01 -2.12180421e-01
-1.10804701e+00 1.28201342e+00 -1.48773611e-01 -2.39862993e-01
-2.44782701e-01 -5.01589477e-01 5.20389736e-01 4.04149294e-01
4.37321544e-01 7.86304772e-01 -4.54910398e-01 -5.50377667e-01
-8.02137852e-01 9.66039181e-01 1.34148586e+00 1.42823029e+00
6.81402266e-01 -8.67232829e-02 -7.79302657e-01 8.76872599e-01
3.29360455e-01 5.56096792e-01 3.38082135e-01 -1.24487889e+00
1.20773828e+00 9.03921247e-01 4.32631671e-01 -8.06677878e-01
-5.74716568e-01 3.75594735e-01 -6.98567867e-01 -3.60256076e-01
6.68742716e-01 -6.84142649e-01 -5.67940772e-01 1.72143638e+00
3.69890571e-01 -5.48388958e-01 3.05628449e-01 8.79035950e-01
1.12059033e+00 8.24559510e-01 -4.14293818e-02 -1.07241236e-01
1.20110607e+00 -1.43483913e+00 -1.07052636e+00 -2.68693119e-01
1.07157147e+00 -5.27999580e-01 1.67607868e+00 1.40495151e-01
-9.65806484e-01 -5.06233752e-01 -5.25567949e-01 -1.83249950e-01
4.62606624e-02 2.75418133e-01 6.80718482e-01 6.78523719e-01
-9.08852339e-01 1.37433425e-01 -6.00378871e-01 -3.34347099e-01
2.81955838e-01 -7.56457299e-02 -1.81037426e-01 -1.25004455e-01
-1.41304874e+00 7.31162786e-01 2.59376287e-01 -9.41816834e-04
-9.85911071e-01 -7.72403300e-01 -1.15306628e+00 1.79497540e-01
6.57459617e-01 -7.22503960e-01 1.81136394e+00 -2.04347491e-01
-1.93595123e+00 2.14491144e-01 -4.22312707e-01 -3.87330502e-01
6.93394840e-01 -2.27132976e-01 -1.02697918e-03 9.64296311e-02
2.42780000e-01 8.91470075e-01 7.67577291e-01 -8.68032396e-01
-5.80143452e-01 2.32516397e-02 3.63712549e-01 6.89694583e-01
-1.77862734e-01 -3.43643248e-01 -5.07169664e-01 -3.63298804e-02
-2.65702575e-01 -1.09475982e+00 -4.43890601e-01 -3.90178144e-01
-1.11443543e+00 -6.75173581e-01 9.71186042e-01 -8.37054789e-01
1.43062341e+00 -1.77848446e+00 2.31072530e-01 -9.13564190e-02
3.42226654e-01 1.79728717e-01 -1.40866756e-01 8.51499438e-01
4.15575594e-01 1.31732058e-02 -2.71976829e-01 -7.04325140e-01
3.90112430e-01 1.54328421e-01 -3.61345738e-01 -1.59854665e-01
2.51835227e-01 9.42157447e-01 -1.05032516e+00 -4.24188554e-01
2.90296614e-01 -1.57464027e-01 -9.88106430e-01 7.09819317e-01
-8.83891225e-01 8.34152699e-01 -7.16112435e-01 5.30184150e-01
3.43344659e-01 -4.13137376e-01 2.16042623e-01 2.29964122e-01
-1.39744719e-02 6.84891880e-01 -8.03038836e-01 2.03534007e+00
-7.20192313e-01 6.88004017e-01 5.77711053e-02 -3.29114676e-01
1.02870178e+00 4.12904203e-01 3.48810822e-01 -6.41693175e-01
-2.89316028e-01 -2.42980570e-01 -2.79019177e-01 -7.15311170e-01
1.09113157e+00 1.46438733e-01 -4.69429374e-01 8.38950694e-01
-1.72660854e-02 -3.88959914e-01 5.30602753e-01 8.05264056e-01
1.08759546e+00 -2.83248350e-02 7.70959780e-02 2.42042914e-01
4.24080580e-01 6.24092400e-01 3.00523251e-01 8.00357282e-01
-2.86164463e-01 4.30255532e-01 1.01698816e+00 -1.55891420e-03
-9.59867537e-01 -6.76505506e-01 2.64746934e-01 1.12942207e+00
-6.11736551e-02 -8.30308974e-01 -9.43037868e-01 -1.10727489e+00
-3.10179163e-02 1.00048912e+00 -1.90045089e-01 -3.53779085e-02
-4.62688506e-01 -5.42702556e-01 5.88272512e-01 2.69451112e-01
7.76053369e-01 -1.19714987e+00 -4.47936952e-01 2.20917836e-01
-8.28472793e-01 -1.13774002e+00 -6.61340356e-01 -3.82595628e-01
-5.80407917e-01 -1.14257944e+00 -6.66558743e-01 -5.06876707e-01
6.80214047e-01 3.32434833e-01 1.15643203e+00 -2.63034850e-01
1.85918719e-01 5.26947737e-01 -4.33261722e-01 7.30701312e-02
-8.74033451e-01 3.18266362e-01 -1.41122565e-01 -1.82024106e-01
3.64499480e-01 -1.08538326e-02 -3.44080955e-01 3.85806412e-01
-4.36871976e-01 5.84527314e-01 3.33430469e-01 1.16381073e+00
3.02538946e-02 -1.72772348e-01 9.40876305e-01 -1.33855903e+00
1.51642013e+00 -6.48538828e-01 -4.62017208e-01 4.84246947e-02
-6.40519440e-01 8.70444775e-02 6.57040060e-01 -6.84106871e-02
-1.68718803e+00 -7.57186040e-02 -2.42324486e-01 -1.45673290e-01
-3.56766433e-01 6.87136710e-01 -2.95001715e-01 5.99411428e-01
7.21890509e-01 1.29994452e-01 2.22316206e-01 -3.92662644e-01
6.12972677e-01 8.03023160e-01 2.12340847e-01 -5.46824694e-01
6.53240621e-01 -1.56111382e-02 -3.80053550e-01 -9.96269584e-01
-7.59114087e-01 -5.26090324e-01 -3.06071997e-01 -3.23077172e-01
8.52425158e-01 -8.71919572e-01 -9.34024096e-01 2.17342973e-01
-1.36049223e+00 -7.30552793e-01 -1.99979484e-01 2.19450772e-01
-6.29998386e-01 2.86120176e-01 -5.55548728e-01 -1.05117321e+00
-2.71601737e-01 -1.27287900e+00 1.12051213e+00 2.46855423e-01
-8.58128607e-01 -1.08227682e+00 2.09003866e-01 7.98635423e-01
2.25036457e-01 3.15666236e-02 1.04143810e+00 -8.57872128e-01
-6.61244869e-01 -1.02615550e-01 -1.50381625e-01 1.16676137e-01
3.03237975e-01 -2.51309484e-01 -7.85271585e-01 -1.67683914e-01
-1.10358469e-01 -7.45443881e-01 4.46714133e-01 1.75309852e-01
7.66082644e-01 -8.82562697e-01 -2.62622237e-01 -1.22962110e-02
6.11485064e-01 2.24461183e-01 1.95273608e-01 -1.42236605e-01
7.19621181e-01 1.14442456e+00 9.98732269e-01 7.26439655e-01
7.26932526e-01 6.96433485e-01 4.44958024e-02 2.51662228e-02
-4.09700833e-02 -8.00847650e-01 5.14727175e-01 6.55233026e-01
6.40971780e-01 -5.26887178e-01 -9.93088305e-01 6.27005994e-01
-2.01663423e+00 -7.83910096e-01 -2.44803295e-01 1.78642058e+00
1.05514097e+00 5.62793575e-02 4.34003115e-01 -1.13552392e-01
5.50891995e-01 4.07815933e-01 -5.98009109e-01 -3.35881174e-01
3.02880853e-01 -3.26680183e-01 -9.25400332e-02 8.94106567e-01
-7.89020419e-01 1.09401178e+00 5.84801674e+00 5.52989066e-01
-4.81125295e-01 -6.02470562e-02 5.69012523e-01 1.67812482e-01
-5.17013967e-01 -5.41244894e-02 -1.01247084e+00 5.19997001e-01
1.01926017e+00 -3.62443566e-01 2.59515792e-01 1.00499785e+00
7.31252789e-01 -9.40293744e-02 -1.34933019e+00 7.05349028e-01
-1.67571157e-01 -1.38739777e+00 5.22369705e-02 -1.26628667e-01
8.11516345e-01 -5.32606423e-01 7.54480213e-02 7.56029010e-01
7.03449130e-01 -9.53347802e-01 1.34287521e-01 2.40396217e-01
7.92761683e-01 -6.13956749e-01 5.20738304e-01 6.52763426e-01
-7.91912019e-01 3.87189537e-02 -3.59611176e-02 1.67431347e-02
5.48444271e-01 3.33904058e-01 -1.84161806e+00 4.48032558e-01
1.65101722e-01 6.10004723e-01 -2.79003650e-01 5.15962958e-01
-3.82171690e-01 7.36141682e-01 2.49453075e-02 -3.03585321e-01
4.77944970e-01 -3.27788025e-01 6.87585354e-01 1.19644117e+00
-1.26313105e-01 -2.94208969e-03 4.93169069e-01 1.20585608e+00
-1.57863989e-01 4.46025208e-02 -7.67716289e-01 -3.23775679e-01
6.00860059e-01 1.10493457e+00 -1.02928057e-01 -2.68194139e-01
-3.33412111e-01 7.98514307e-01 -3.40792321e-04 5.75614989e-01
-5.94512522e-01 -4.85594720e-01 5.59561908e-01 -1.34954900e-01
-2.38178045e-01 -1.93132967e-01 -4.19980079e-01 -1.11069262e+00
-1.11503333e-01 -1.19899321e+00 1.94499075e-01 -5.67743659e-01
-1.04267228e+00 5.32410085e-01 6.88178912e-02 -1.21760452e+00
-8.30924630e-01 -2.21059937e-02 -5.61924815e-01 1.03876114e+00
-1.15985525e+00 -8.96815479e-01 -4.10534024e-01 5.84075868e-01
1.33019078e+00 -2.91507244e-01 9.99422967e-01 1.29056454e-01
-8.05420995e-01 6.10405743e-01 -1.44685805e-01 1.69667348e-01
7.17949033e-01 -1.31458139e+00 7.14201450e-01 5.98588586e-01
-1.18308052e-01 6.74867690e-01 7.79887319e-01 -8.64943445e-01
-1.29409456e+00 -1.11411178e+00 1.11422706e+00 -5.64443767e-01
4.33299959e-01 -5.15298367e-01 -9.75772560e-01 5.64959049e-01
2.35538617e-01 -7.57055044e-01 6.50902689e-01 4.22560424e-01
1.02376588e-01 3.56813073e-01 -1.15443051e+00 8.01418543e-01
7.52893031e-01 -6.41263902e-01 -6.04283512e-01 7.45339692e-01
9.16066766e-01 -8.19831908e-01 -4.82789278e-01 -1.82511047e-01
2.70039141e-01 -6.91327155e-01 5.57606995e-01 -9.45230067e-01
7.63845801e-01 1.54152051e-01 2.62437671e-01 -1.56418777e+00
2.62900501e-01 -1.40473235e+00 -3.98572862e-01 1.37928557e+00
6.13124490e-01 -2.62318134e-01 1.20052278e+00 1.07585585e+00
-4.09913778e-01 -6.18766487e-01 -5.63860178e-01 -3.99046183e-01
-3.72366667e-01 -2.27789462e-01 5.64089358e-01 6.94577456e-01
6.05585158e-01 9.25786734e-01 -6.97989404e-01 -2.84112334e-01
3.21789742e-01 1.21857800e-01 1.34420168e+00 -9.89761472e-01
-2.08865836e-01 -1.01196043e-01 4.50821519e-01 -1.86863649e+00
4.76738572e-01 -7.55764067e-01 4.31577444e-01 -1.71835721e+00
-1.87739059e-01 -5.82878828e-01 7.02928126e-01 4.17078167e-01
-3.57859194e-01 -5.86284637e-01 2.26867989e-01 1.25726372e-01
-7.28502333e-01 9.82999146e-01 1.65182543e+00 -1.98795572e-01
-8.66689205e-01 4.32303876e-01 -7.28284538e-01 5.34360826e-01
8.06283057e-01 -2.42624730e-01 -9.25938308e-01 -1.64632112e-01
-1.10710729e-02 9.70033288e-01 -8.79651681e-03 -4.38512594e-01
2.70376831e-01 -3.51733088e-01 -2.68580049e-01 -7.70721495e-01
5.72885871e-01 -4.09589201e-01 -4.38719213e-01 2.13734373e-01
-9.82915103e-01 -3.16735953e-02 4.76729907e-02 7.95477569e-01
-2.97003478e-01 -1.92126170e-01 3.75478804e-01 -9.43263471e-02
-5.99443555e-01 2.02301219e-01 -7.94862986e-01 4.44315463e-01
6.43871427e-01 1.43603101e-01 -4.80704427e-01 -1.05552816e+00
-5.12463033e-01 8.26937854e-01 -8.01174756e-05 5.09682238e-01
7.64640152e-01 -1.16424263e+00 -6.83819234e-01 2.34868050e-01
1.38280109e-01 3.86380434e-01 4.11088020e-01 6.89736426e-01
-3.00877243e-01 1.06303227e+00 5.41641451e-02 -5.09706020e-01
-1.15638208e+00 1.09128393e-01 3.04263309e-02 -7.29115367e-01
-7.87659526e-01 8.02051663e-01 1.43695369e-01 -9.30062592e-01
3.89931232e-01 -3.45204473e-01 -3.00986975e-01 2.02361032e-01
3.53611499e-01 5.45802176e-01 -1.50726691e-01 -1.99228764e-01
9.96799991e-02 -4.68445450e-01 -3.37115049e-01 -3.96424532e-01
9.89127576e-01 -3.03708524e-01 3.19711089e-01 2.93825358e-01
8.34203362e-01 -6.25050217e-02 -1.61097896e+00 -2.37769336e-01
6.42523468e-02 -5.70442140e-01 -3.57845932e-01 -9.49043393e-01
-5.15372992e-01 8.06211233e-01 -2.01007739e-01 4.65243369e-01
5.55360317e-01 -1.78472027e-01 1.16855800e+00 8.88062894e-01
1.03039794e-01 -1.02912152e+00 5.39422095e-01 8.51337373e-01
9.81570959e-01 -1.36327434e+00 -4.53975827e-01 -3.91056001e-01
-1.33238709e+00 9.41685975e-01 9.48680580e-01 3.68788183e-01
-2.52308752e-02 -1.82934508e-01 -4.27701548e-02 -1.48308292e-01
-1.09779441e+00 4.35625389e-03 2.73398962e-02 6.40717566e-01
2.76306033e-01 -3.42764780e-02 1.76253356e-02 7.11438417e-01
-3.31009477e-01 -2.32979029e-01 8.14733267e-01 7.54359663e-01
-2.73618400e-01 -1.06749833e+00 -1.37996212e-01 5.88295281e-01
1.19830422e-01 -1.08117439e-01 -6.20931327e-01 7.36949205e-01
-5.76275706e-01 1.40420961e+00 -2.04985201e-01 -3.72898668e-01
4.71334815e-01 3.90601039e-01 -1.07265882e-01 -1.10110950e+00
-5.90843320e-01 -2.32620522e-01 7.59086311e-01 -5.43462873e-01
-9.92112681e-02 -5.90657353e-01 -1.30511272e+00 -5.03354013e-01
-2.45459631e-01 6.39621317e-01 3.71497959e-01 9.57097530e-01
3.76010686e-01 6.57292962e-01 7.87217855e-01 -5.83825588e-01
-7.76570201e-01 -1.22459471e+00 -3.29542786e-01 3.55666339e-01
5.16592264e-01 -5.44864655e-01 -1.89295962e-01 5.85960783e-03] | [12.785566329956055, 8.10977554321289] |
8dccbca6-d7dc-4f35-81e1-5704533addb6 | incremental-few-shot-object-detection-via | 2302.09779 | null | https://arxiv.org/abs/2302.09779v1 | https://arxiv.org/pdf/2302.09779v1.pdf | Incremental Few-Shot Object Detection via Simple Fine-Tuning Approach | In this paper, we explore incremental few-shot object detection (iFSD), which incrementally learns novel classes using only a few examples without revisiting base classes. Previous iFSD works achieved the desired results by applying meta-learning. However, meta-learning approaches show insufficient performance that is difficult to apply to practical problems. In this light, we propose a simple fine-tuning-based approach, the Incremental Two-stage Fine-tuning Approach (iTFA) for iFSD, which contains three steps: 1) base training using abundant base classes with the class-agnostic box regressor, 2) separation of the RoI feature extractor and classifier into the base and novel class branches for preserving base knowledge, and 3) fine-tuning the novel branch using only a few novel class examples. We evaluate our iTFA on the real-world datasets PASCAL VOC, COCO, and LVIS. iTFA achieves competitive performance in COCO and shows a 30% higher AP accuracy than meta-learning methods in the LVIS dataset. Experimental results show the effectiveness and applicability of our proposed method. | ['Jong-Hwan Kim', 'Tae-Min Choi'] | 2023-02-20 | null | null | null | null | ['few-shot-object-detection'] | ['computer-vision'] | [ 2.58975297e-01 2.84229219e-03 -3.49975526e-01 -2.31379136e-01
-9.40508485e-01 -9.26192924e-02 5.58659375e-01 6.75408691e-02
-4.56442535e-01 8.54406655e-01 -1.96884856e-01 9.20861140e-02
-1.60657331e-01 -7.95311928e-01 -5.52344143e-01 -7.51128078e-01
2.32114658e-01 2.58188725e-01 1.06979072e+00 3.66707779e-02
3.04305613e-01 4.24844593e-01 -2.10072112e+00 7.07413375e-01
8.16288471e-01 1.05770135e+00 2.61403352e-01 7.97279835e-01
-3.07765752e-01 1.01229644e+00 -6.16395533e-01 -2.23214120e-01
3.90845746e-01 -4.12108481e-01 -6.98285639e-01 2.25327220e-02
6.23490632e-01 -4.55381840e-01 -7.31820390e-02 8.37202132e-01
7.07720220e-01 3.48795921e-01 6.39031529e-01 -1.29775298e+00
-3.20779115e-01 4.87549216e-01 -7.34431565e-01 6.31455719e-01
-2.79705197e-01 2.13983148e-01 8.09740067e-01 -1.14635682e+00
6.10890865e-01 1.04786777e+00 7.37499237e-01 7.59672165e-01
-9.00927365e-01 -6.22856736e-01 1.50443286e-01 7.10367203e-01
-1.36005104e+00 -4.26143199e-01 7.74407089e-01 -4.78608757e-01
9.17858422e-01 1.08100094e-01 6.31327510e-01 7.31984794e-01
5.00457734e-03 1.09026802e+00 1.20130265e+00 -6.52937353e-01
5.29980361e-01 4.51186329e-01 5.43503225e-01 6.67497575e-01
2.52368301e-01 2.05300748e-02 -3.61510605e-01 -1.25765473e-01
2.88373262e-01 2.84733713e-01 2.01855421e-01 -4.89894152e-01
-8.79127085e-01 7.89659917e-01 3.85622025e-01 2.89524078e-01
-3.19972754e-01 1.72296185e-02 5.90331912e-01 4.59282063e-02
5.70012093e-01 1.65288135e-01 -3.86243135e-01 9.32400301e-02
-1.13164306e+00 1.41972423e-01 4.58269089e-01 1.09152281e+00
8.50679100e-01 4.93401475e-02 -6.05996251e-01 1.11760497e+00
-7.63974041e-02 8.63612071e-02 8.24347317e-01 -7.39950120e-01
3.13290536e-01 6.03252053e-01 -3.85075272e-03 -4.64910448e-01
-3.51235867e-01 -7.07050920e-01 -4.01025832e-01 3.89214456e-01
2.67564375e-02 -8.97677615e-02 -1.20549381e+00 1.23835158e+00
7.98900664e-01 5.13457358e-01 1.43396435e-02 4.18265074e-01
1.13625610e+00 6.76025331e-01 9.05565396e-02 -5.32879055e-01
1.06577981e+00 -1.43066728e+00 -5.09190440e-01 -1.40014544e-01
5.34228921e-01 -4.14387286e-01 1.03546107e+00 3.57869595e-01
-9.30572867e-01 -9.25129473e-01 -1.40333879e+00 2.25327700e-01
-4.12402004e-01 6.58542365e-02 5.55459201e-01 7.02507555e-01
-4.99507487e-01 5.57158887e-01 -7.07030296e-01 -2.92024225e-01
1.00884020e+00 1.79463699e-01 -1.83528811e-02 -2.05188707e-01
-7.53668904e-01 5.81447124e-01 7.80453503e-01 -2.08178252e-01
-1.25469053e+00 -1.13897991e+00 -5.35036623e-01 9.00188088e-02
7.99708366e-01 -5.43800831e-01 1.47516084e+00 -8.56584668e-01
-1.25945818e+00 6.58795893e-01 -2.77541690e-02 -6.24722540e-01
5.36188543e-01 -1.91460818e-01 -2.86297143e-01 1.13602959e-01
9.94066224e-02 6.28621638e-01 1.15461063e+00 -1.24038184e+00
-1.36592579e+00 -2.91527420e-01 1.63434908e-01 2.09657531e-02
-4.35516179e-01 -3.81587565e-01 -3.24165791e-01 -4.12240446e-01
-8.93712044e-02 -5.62313616e-01 -1.61322877e-01 1.48457289e-02
-1.26208335e-01 -3.51056039e-01 1.16530156e+00 -1.39804080e-01
1.47829711e+00 -2.20287919e+00 -3.25676024e-01 -2.33788103e-01
2.92593420e-01 8.84290755e-01 -1.75055161e-01 1.21746724e-02
3.93690802e-02 -2.55972981e-01 -8.13344344e-02 -8.30561072e-02
-3.25558931e-01 9.66808870e-02 -2.70294189e-01 7.13291094e-02
3.71493474e-02 8.33614409e-01 -1.09710121e+00 -8.41895521e-01
4.05749947e-01 1.01132058e-01 -5.64169765e-01 1.40580222e-01
-1.60607100e-01 -2.39806056e-01 -3.44568253e-01 9.28024292e-01
6.65747225e-01 2.07290351e-02 -2.61501878e-01 -4.46561486e-01
-2.00858966e-01 -3.27349454e-01 -1.28745198e+00 1.36488938e+00
-2.44207487e-01 4.23005074e-01 -5.12758613e-01 -1.10033226e+00
9.36153948e-01 -9.08184871e-02 4.78344500e-01 -6.84885204e-01
1.04306541e-01 1.68001398e-01 -8.60723034e-02 -6.33865476e-01
3.42451245e-01 -3.18201602e-01 2.46763960e-01 2.32465416e-01
4.98636812e-01 1.53777925e-02 4.28551972e-01 1.05765961e-01
1.07886136e+00 1.81527331e-01 8.08171809e-01 -4.23714612e-03
6.08942807e-01 2.06728101e-01 8.05799186e-01 1.09341681e+00
-7.66384423e-01 4.31420743e-01 1.18084714e-01 -5.50329089e-01
-8.98672163e-01 -1.10701144e+00 -1.24185242e-01 1.27782059e+00
1.57120839e-01 -3.56671482e-01 -8.38737071e-01 -1.12600720e+00
-1.93512999e-02 9.22389686e-01 -8.61407936e-01 -3.99486005e-01
-4.52012628e-01 -8.35179865e-01 3.06117117e-01 7.35855937e-01
8.48984241e-01 -9.99311090e-01 -9.11812842e-01 2.71155804e-01
7.62866959e-02 -7.81170666e-01 -2.53128618e-01 3.61496001e-01
-1.08757484e+00 -1.18160510e+00 -6.96370721e-01 -7.74033427e-01
5.06615698e-01 6.33804619e-01 5.95451653e-01 2.06255801e-02
-7.88523495e-01 3.66690010e-01 -4.95413452e-01 -7.66877055e-01
-1.50124863e-01 6.53752312e-02 -2.34617352e-01 1.59203559e-01
4.20094728e-01 -3.32967222e-01 -6.09293282e-01 3.23938131e-01
-6.75279498e-01 1.14034945e-02 6.68414474e-01 1.12207961e+00
8.88173759e-01 7.62897581e-02 9.06397939e-01 -1.24155021e+00
9.23674479e-02 -3.79238099e-01 -4.84191477e-01 5.16559184e-01
-8.13445508e-01 -2.26119787e-01 4.27667052e-01 -6.31486356e-01
-1.46996629e+00 1.85668141e-01 7.07465336e-02 -5.66165686e-01
-4.31338064e-02 8.61523822e-02 -6.14462979e-02 -4.77148965e-02
1.15652001e+00 2.28479415e-01 -3.11056346e-01 -4.25409615e-01
4.43397105e-01 8.93886924e-01 6.67989731e-01 -2.79774815e-01
8.17547381e-01 7.16221929e-01 -1.51886940e-01 -8.96778524e-01
-1.37422216e+00 -8.30756545e-01 -9.75095689e-01 -4.28688020e-01
5.77373385e-01 -7.36674190e-01 -1.11508846e-01 3.34556550e-01
-7.03161120e-01 -2.34068424e-01 -8.10483456e-01 4.21461999e-01
-7.10407913e-01 1.57194644e-01 -1.84779376e-01 -8.11744690e-01
-5.99347770e-01 -7.98744261e-01 7.73357272e-01 4.77067173e-01
1.18400693e-01 -5.40625870e-01 2.57382840e-01 4.39108878e-01
2.65589267e-01 2.44429424e-01 7.17025459e-01 -8.18041205e-01
-4.85023290e-01 -2.57316053e-01 -1.62557095e-01 3.67297798e-01
1.94572657e-01 -1.68575551e-02 -1.27909386e+00 -3.30665916e-01
-1.18813448e-01 -3.58290046e-01 1.19552004e+00 2.25852534e-01
1.29432225e+00 1.25162914e-01 -5.44073641e-01 6.23338282e-01
1.51087856e+00 5.39297104e-01 6.25973165e-01 5.90478718e-01
4.63700861e-01 2.20729426e-01 1.22017848e+00 5.71896553e-01
8.12291950e-02 4.87034947e-01 3.77251923e-01 1.63396463e-01
-6.08271122e-01 -1.12351917e-01 6.48122877e-02 4.42267805e-01
-1.49845317e-01 2.46741027e-01 -7.05462456e-01 7.62294888e-01
-1.99215293e+00 -1.32874811e+00 2.40214452e-01 2.10614634e+00
7.42083549e-01 4.75406021e-01 2.44627222e-01 3.46419126e-01
7.74973392e-01 6.26855344e-03 -7.84244955e-01 -2.61343658e-01
2.23009422e-01 2.76767641e-01 2.25871831e-01 5.34549095e-02
-1.22791970e+00 9.99373734e-01 5.55414820e+00 1.18915153e+00
-1.00655186e+00 6.00696445e-01 5.20305037e-01 -3.66557866e-01
2.20554844e-01 -1.92424171e-02 -1.32216656e+00 2.24367991e-01
6.91473365e-01 -2.52274990e-01 1.31503060e-01 1.22640741e+00
-2.24798277e-01 -1.78228721e-01 -9.27682400e-01 9.50255096e-01
2.91850626e-01 -1.50890660e+00 9.27026421e-02 -5.33606887e-01
9.97601092e-01 -6.57569095e-02 -2.77983963e-01 9.51045871e-01
-9.70604792e-02 -3.01869333e-01 7.35869348e-01 5.48592925e-01
7.06207693e-01 -8.04698586e-01 4.94255722e-01 5.99703252e-01
-1.19018674e+00 -6.59434974e-01 -7.88684845e-01 3.39823037e-01
-2.39195913e-01 5.32647669e-01 -1.07642305e+00 4.35200959e-01
9.03649628e-01 5.39980233e-01 -7.51320541e-01 1.54703653e+00
-3.95567454e-02 7.04032600e-01 2.37718988e-02 -8.91736224e-02
1.00672141e-01 4.31506723e-01 5.30915439e-01 1.20598519e+00
1.59426536e-02 1.97543368e-01 1.13665834e-01 5.98146379e-01
4.91718762e-02 9.94724780e-02 -2.98186034e-01 1.18681535e-01
4.69516546e-01 1.50686908e+00 -8.90564620e-01 -7.43393600e-01
-4.70025867e-01 7.45708585e-01 3.76451582e-01 1.11646175e-01
-8.78928721e-01 -7.87793279e-01 7.89852813e-02 8.85424465e-02
7.49680042e-01 3.28117698e-01 -1.22669876e-01 -9.61186111e-01
-1.76845893e-01 -6.20130181e-01 8.83273780e-01 -7.09056795e-01
-9.76726472e-01 5.68298340e-01 2.28734180e-01 -1.44092882e+00
-4.59103845e-02 -4.77761328e-01 -8.04736793e-01 2.43559137e-01
-1.59704208e+00 -1.36917877e+00 -6.71639264e-01 5.84901512e-01
1.21399057e+00 -4.02023375e-01 4.54702646e-01 3.13370287e-01
-6.19422317e-01 7.55460620e-01 9.72831398e-02 -1.74249575e-01
6.43004537e-01 -1.08151996e+00 -7.64509710e-03 6.43596470e-01
1.23646684e-01 2.07597569e-01 4.22126234e-01 -7.14195788e-01
-1.01563823e+00 -1.44684029e+00 4.53178734e-01 -2.27119237e-01
3.81020069e-01 -1.86227158e-01 -9.33282733e-01 5.61857045e-01
-1.83108717e-01 4.63610232e-01 5.97289622e-01 -1.27501860e-01
-3.71548653e-01 -5.86513698e-01 -1.39827430e+00 3.65016222e-01
1.09039032e+00 -9.01562497e-02 -9.75577056e-01 2.81170368e-01
7.71019399e-01 -2.43812695e-01 -5.34756541e-01 6.94777668e-01
6.33380473e-01 -9.32303011e-01 9.41568911e-01 -7.65913308e-01
2.26180419e-01 -4.58492339e-01 -3.13972056e-01 -1.15857446e+00
-5.32581031e-01 -2.09157571e-01 -5.42590261e-01 1.31351566e+00
3.35030288e-01 -3.67263168e-01 7.73503721e-01 8.87821168e-02
-3.07343215e-01 -1.01889551e+00 -8.99261296e-01 -9.92672741e-01
-1.74797162e-01 -4.74405557e-01 5.44394314e-01 7.46124387e-01
-2.15702131e-01 3.39922577e-01 -2.97981024e-01 -7.79355317e-02
8.30215096e-01 3.14630121e-01 9.64683175e-01 -1.16826439e+00
-4.63049352e-01 -1.81247473e-01 -5.19152343e-01 -4.41614687e-01
-2.58409977e-01 -8.08994293e-01 9.73834395e-02 -1.41794801e+00
7.21671224e-01 -3.19624990e-01 -6.15788579e-01 5.87831616e-01
-3.86124879e-01 3.43813360e-01 2.54336149e-01 1.91940755e-01
-9.64207947e-01 5.41424870e-01 1.08192885e+00 -4.39192712e-01
-4.13680226e-01 1.57735839e-01 -5.82430303e-01 8.22315753e-01
7.14145660e-01 -5.94396055e-01 -6.03761375e-01 7.57391006e-02
-3.01738262e-01 -3.93535942e-01 1.84733108e-01 -1.31476939e+00
3.27097952e-01 -2.36715838e-01 6.66206717e-01 -9.89586651e-01
2.10442036e-01 -5.98489642e-01 -1.85985327e-01 7.55534291e-01
-1.56993836e-01 -6.50550961e-01 2.39613265e-01 7.09926665e-01
2.59621385e-02 -6.27750754e-01 1.24927390e+00 -2.22611159e-01
-1.29786754e+00 3.46531063e-01 3.19047011e-02 1.00696445e-01
1.59772909e+00 -5.61889172e-01 -4.28173125e-01 3.02899092e-01
-8.91691029e-01 1.77620441e-01 -8.07511806e-02 4.53965098e-01
7.05461085e-01 -1.28904593e+00 -5.91815829e-01 2.12202147e-01
4.82919306e-01 -1.58286810e-01 4.86982256e-01 7.70754695e-01
-2.44725630e-01 1.68529704e-01 -2.38523707e-01 -8.03128898e-01
-1.49671626e+00 9.51462805e-01 4.00182396e-01 -1.03702806e-01
-6.65671468e-01 8.45790863e-01 1.91268280e-01 -2.26710215e-01
4.17313308e-01 8.09691623e-02 -4.62817609e-01 4.00579453e-01
9.59645569e-01 9.10588086e-01 1.34200215e-01 -1.88256368e-01
-2.46382713e-01 5.14250159e-01 -4.26261514e-01 2.73405254e-01
1.38489974e+00 -4.60743569e-02 2.98770100e-01 7.55988657e-01
9.53319490e-01 -2.74006933e-01 -1.32479477e+00 -5.48332632e-01
-5.99474125e-02 -6.50120318e-01 1.06705666e-01 -8.39086711e-01
-9.22541320e-01 8.51465523e-01 1.00331950e+00 -2.19242215e-01
1.25263703e+00 3.17605250e-02 6.66154027e-01 5.35011232e-01
4.63373899e-01 -1.46433282e+00 4.07375515e-01 2.67360002e-01
6.40313208e-01 -1.15951836e+00 2.38404900e-01 -3.23413163e-01
-5.06854236e-01 1.15497148e+00 1.12294400e+00 -1.27134158e-03
7.67172217e-01 3.67231928e-02 -2.41594657e-01 -1.04488775e-01
-9.85543787e-01 -2.91732579e-01 3.68814200e-01 6.22739971e-01
-2.01037392e-01 -1.45593673e-01 -3.67366105e-01 7.60623991e-01
3.08889866e-01 2.67658859e-01 3.47092152e-01 1.20933807e+00
-1.13493931e+00 -6.05045378e-01 -2.70657033e-01 8.01149666e-01
-2.28560895e-01 1.24048308e-01 -1.99176759e-01 8.84174883e-01
5.61845481e-01 6.59266889e-01 -5.06204255e-02 -4.34748322e-01
5.74243546e-01 2.45268241e-01 5.95200121e-01 -9.27233160e-01
-3.92604142e-01 1.72870066e-02 -1.96058348e-01 -5.16733110e-01
-3.39042515e-01 -6.72588408e-01 -1.16355956e+00 1.71448797e-01
-7.43424296e-01 9.03971195e-02 4.71497476e-01 1.03740454e+00
1.37954757e-01 7.29464471e-01 7.95683444e-01 -7.17354774e-01
-7.91207910e-01 -9.81188715e-01 -4.47810978e-01 -5.68444375e-03
3.21282327e-01 -9.33648944e-01 -2.78450578e-01 1.55211687e-02] | [9.664093971252441, 2.4786529541015625] |
2f13c7f5-0a35-4356-a78d-edb4810b70ee | directtracker-3d-multi-object-tracking-using | 2209.14965 | null | https://arxiv.org/abs/2209.14965v1 | https://arxiv.org/pdf/2209.14965v1.pdf | DirectTracker: 3D Multi-Object Tracking Using Direct Image Alignment and Photometric Bundle Adjustment | Direct methods have shown excellent performance in the applications of visual odometry and SLAM. In this work we propose to leverage their effectiveness for the task of 3D multi-object tracking. To this end, we propose DirectTracker, a framework that effectively combines direct image alignment for the short-term tracking and sliding-window photometric bundle adjustment for 3D object detection. Object proposals are estimated based on the sparse sliding-window pointcloud and further refined using an optimization-based cost function that carefully combines 3D and 2D cues to ensure consistency in image and world space. We propose to evaluate 3D tracking using the recently introduced higher-order tracking accuracy (HOTA) metric and the generalized intersection over union similarity measure to mitigate the limitations of the conventional use of intersection over union for the evaluation of vision-based trackers. We perform evaluation on the KITTI Tracking benchmark for the Car class and show competitive performance in tracking objects both in 2D and 3D. | ['Daniel Cremers', 'Laura Leal-Taixé', 'Aljoša Ošep', 'Nikolaus Demmel', 'Nikita Korobov', 'Mariia Gladkova'] | 2022-09-29 | null | null | null | null | ['3d-multi-object-tracking'] | ['computer-vision'] | [-3.28893065e-01 -5.72285116e-01 -1.18926845e-01 -1.20775200e-01
-7.89636552e-01 -7.33699083e-01 8.99203002e-01 2.28341103e-01
-5.79823315e-01 1.85539559e-01 -3.41047794e-01 -1.66746527e-01
4.33972813e-02 -2.53791004e-01 -6.56414211e-01 -4.99108016e-01
-1.13656573e-01 6.82135284e-01 8.32433641e-01 -1.20342098e-01
1.71217233e-01 9.31606293e-01 -1.59529424e+00 -6.33432209e-01
3.74416649e-01 9.87477183e-01 8.24886560e-02 6.86850011e-01
1.34608373e-01 2.41057605e-01 -4.82811362e-01 -2.05510825e-01
6.47210181e-01 1.33535773e-01 -1.25399768e-01 1.47891462e-01
1.07110143e+00 -2.22497255e-01 -5.23353070e-02 1.04013157e+00
6.65862203e-01 9.83993486e-02 4.06714857e-01 -1.61120212e+00
-1.11210737e-02 -2.83769995e-01 -8.22101831e-01 4.58873242e-01
5.02073407e-01 4.87824410e-01 8.07156026e-01 -1.10821342e+00
7.35360026e-01 1.39925897e+00 1.07234275e+00 2.40946963e-01
-1.27777076e+00 -8.00417125e-01 1.43034413e-01 -8.58844966e-02
-1.61721122e+00 -5.49018800e-01 4.35869932e-01 -6.83371425e-01
1.01466131e+00 7.23695084e-02 6.78765893e-01 5.47843158e-01
3.71928304e-01 4.81740057e-01 9.42943990e-01 -2.91012555e-01
-1.67703629e-01 1.27794221e-01 2.05364332e-01 8.16918135e-01
7.19099998e-01 6.51252449e-01 -5.65576911e-01 -2.22952977e-01
6.21495128e-01 3.35203521e-02 -1.88656934e-02 -1.24034584e+00
-1.38007307e+00 6.93778992e-01 5.49531221e-01 -1.33490816e-01
-2.67600089e-01 3.07963938e-01 1.53465226e-01 7.33045861e-03
5.69144368e-01 2.02478051e-01 -2.01259375e-01 -5.15991226e-02
-1.01311541e+00 5.75855672e-01 5.03169775e-01 1.30547369e+00
7.13373840e-01 -7.09245354e-02 -3.04042011e-01 3.18709433e-01
8.24548542e-01 9.79844093e-01 -1.35548800e-01 -8.70647848e-01
1.61259919e-01 4.95993018e-01 5.17351866e-01 -9.53191936e-01
-5.51633179e-01 -5.56602299e-01 -6.71531558e-02 7.49501765e-01
4.21773136e-01 2.40269661e-01 -8.47226441e-01 1.50913322e+00
1.22169602e+00 5.46890914e-01 -8.18533972e-02 1.15940082e+00
7.89272606e-01 3.32189612e-02 5.04291803e-02 -5.21222614e-02
1.42485130e+00 -8.44430566e-01 -5.72722793e-01 -2.25783020e-01
6.17685258e-01 -1.05573750e+00 4.78770792e-01 -7.52113461e-02
-1.06382489e+00 -6.08849883e-01 -1.27422559e+00 9.91281122e-02
-1.17512353e-01 2.01954860e-02 2.62110919e-01 7.08514929e-01
-9.39007521e-01 2.55210072e-01 -1.09119463e+00 -5.71508586e-01
3.20593446e-01 3.44551563e-01 -3.27785969e-01 1.11146547e-01
-5.82717121e-01 1.37295187e+00 3.02872241e-01 -8.60498995e-02
-8.68184149e-01 -9.15553272e-01 -9.87076879e-01 -3.32756042e-01
4.13583159e-01 -9.14226115e-01 1.22105014e+00 1.49253905e-01
-1.36282408e+00 1.19543624e+00 -1.89021826e-01 -6.99667037e-01
5.59645414e-01 -4.58085477e-01 -2.43689910e-01 -2.07365617e-01
2.83292532e-01 5.08203447e-01 7.64247060e-01 -1.24841309e+00
-9.55755115e-01 -6.19640172e-01 -2.16443315e-01 2.75496691e-01
2.58256346e-01 3.50336134e-02 -8.38722289e-01 -4.24388021e-01
4.63227957e-01 -1.30958366e+00 -2.06403270e-01 2.44880095e-01
-1.46748185e-01 -1.48181781e-01 1.05974281e+00 -1.76374033e-01
7.51156569e-01 -2.02619600e+00 -7.49545023e-02 7.85309374e-02
3.21405649e-01 2.21741781e-01 4.81781773e-02 -1.12749636e-01
4.23408478e-01 -5.65182030e-01 2.85829216e-01 -8.01216125e-01
1.11481763e-01 1.44070107e-02 4.54858579e-02 9.93384540e-01
2.42126420e-01 8.09468091e-01 -9.74475801e-01 -6.27597988e-01
5.27111173e-01 5.07056653e-01 -2.60127157e-01 9.30877030e-02
-1.81892626e-02 4.46725637e-01 -4.05974239e-01 7.32560933e-01
9.60668147e-01 -1.08046398e-01 -2.94011414e-01 -3.90790850e-01
-5.72789609e-01 2.30950907e-01 -1.42851937e+00 1.81006181e+00
-2.15573207e-01 4.80452448e-01 2.15069577e-01 -6.46699518e-02
9.15440857e-01 8.01326483e-02 7.45510519e-01 -4.56492901e-01
2.16481104e-01 1.79926232e-01 -1.43211573e-01 1.70561299e-01
8.59196901e-01 8.16614553e-02 -1.73727926e-02 1.52034715e-01
5.44506833e-02 -4.00769264e-01 8.60066712e-02 1.99764714e-01
1.12409091e+00 5.59446752e-01 4.71590251e-01 -2.48586580e-01
5.93927085e-01 4.62937802e-01 5.10886669e-01 7.82465398e-01
-6.07078493e-01 3.82330149e-01 -3.83772552e-01 -2.14831844e-01
-9.72095668e-01 -1.26604104e+00 -3.31642032e-01 8.53310704e-01
8.21617126e-01 -3.73841435e-01 -1.43745303e-01 -7.63884306e-01
6.90546930e-01 4.17556524e-01 -2.45918155e-01 6.92380918e-03
-6.29898608e-01 -4.89607900e-01 3.66014719e-01 3.50664884e-01
9.45006385e-02 -2.77594268e-01 -1.09407234e+00 3.73191327e-01
1.83299690e-01 -1.54263651e+00 -6.58788860e-01 7.17993081e-02
-8.65218520e-01 -1.12002540e+00 -5.10868549e-01 -3.95074725e-01
3.04542899e-01 8.69388103e-01 1.04889560e+00 -1.09092388e-02
-2.41452679e-01 8.32085252e-01 -1.30235627e-01 -6.52029634e-01
-2.68588811e-01 -1.96272448e-01 4.24739063e-01 -2.80467957e-01
4.61178929e-01 -1.31048322e-01 -4.33118820e-01 7.27168500e-01
-1.93149045e-01 -1.14428520e-01 3.37153673e-01 5.11926830e-01
7.19036102e-01 -5.67004323e-01 -1.05763786e-01 -8.81451815e-02
-3.56049575e-02 -6.93587586e-02 -1.35041547e+00 6.93034753e-02
-6.92775130e-01 -3.32960859e-02 -4.71102953e-01 -5.20146012e-01
-5.75337529e-01 4.95805532e-01 1.19903021e-01 -8.97650838e-01
1.31639853e-01 -8.01724717e-02 9.53440666e-02 -8.54664385e-01
5.28084934e-01 -7.61153772e-02 1.30773172e-01 -4.96792495e-01
3.74308616e-01 1.88450947e-01 7.08583772e-01 -2.45729834e-01
1.33811200e+00 8.59790564e-01 4.54759806e-01 -4.61671025e-01
-8.46864402e-01 -1.11882615e+00 -7.09878743e-01 -4.78272885e-01
9.50160563e-01 -1.23831022e+00 -8.96998107e-01 2.42470279e-01
-1.27975011e+00 6.03646152e-02 -1.25882655e-01 9.10281956e-01
-5.68835199e-01 4.69677001e-01 -1.67300493e-01 -9.66801405e-01
-3.34941238e-01 -1.33479369e+00 1.61112726e+00 1.44374698e-01
6.30173609e-02 -8.23433459e-01 3.90799373e-01 -1.74181443e-03
2.27509052e-01 5.00393867e-01 -1.54701501e-01 -3.93678784e-01
-1.06520951e+00 -3.50736260e-01 -2.95408666e-01 -2.97210962e-01
-9.38228667e-02 -1.75521523e-01 -8.64273667e-01 -7.47266054e-01
-1.01069778e-01 1.79682970e-01 7.08286047e-01 3.47925901e-01
-3.91499102e-02 5.20580471e-01 -8.03235173e-01 7.84627616e-01
1.53252125e+00 -1.02619022e-01 5.79533875e-02 6.90121412e-01
8.66010249e-01 1.43840656e-01 1.27950275e+00 4.49895114e-01
5.99136591e-01 1.38065016e+00 7.05160499e-01 7.89219514e-02
-4.05737787e-01 -6.33810014e-02 3.73911798e-01 4.78307456e-01
1.53931469e-01 1.84804410e-01 -9.75915790e-01 5.33480346e-01
-1.89005220e+00 -7.59100616e-01 -4.21800107e-01 2.46565413e+00
3.40148568e-01 5.47440171e-01 2.33881071e-01 -3.65229458e-01
7.40786195e-01 -2.11696178e-02 -4.15499002e-01 2.56789654e-01
2.87469551e-02 -1.75804913e-01 1.15969491e+00 5.71096063e-01
-1.34090900e+00 1.02263069e+00 5.84038687e+00 4.74331260e-01
-8.55500996e-01 4.24324751e-01 -4.92334008e-01 -1.81461141e-01
2.88179755e-01 1.47948131e-01 -1.81092680e+00 2.26231784e-01
6.68333411e-01 -1.35199040e-01 -1.13549255e-01 7.92974949e-01
7.66734034e-02 -2.14547664e-01 -1.09751689e+00 1.10637057e+00
3.30253094e-02 -1.24405813e+00 -4.55784231e-01 2.90098935e-01
6.08006179e-01 5.67097127e-01 9.67365578e-02 1.61029711e-01
3.82225901e-01 -2.33926907e-01 1.10746384e+00 2.33920276e-01
3.75221997e-01 -3.73641878e-01 3.53270203e-01 2.81909734e-01
-1.76010478e+00 2.75823355e-01 -1.38097778e-01 2.79072464e-01
4.65991706e-01 3.79629642e-01 -1.29438007e+00 6.68341517e-01
7.46250212e-01 8.84408116e-01 -6.28649950e-01 1.73450112e+00
1.66641474e-01 -3.34230959e-02 -7.17373610e-01 2.36131728e-01
3.21924478e-01 -5.10115288e-02 1.29003072e+00 1.08465946e+00
4.39043015e-01 -3.09381485e-01 6.04546010e-01 7.03325689e-01
4.32053387e-01 -2.12068245e-01 -6.66408300e-01 6.48161471e-01
5.96828043e-01 1.31342840e+00 -6.24474287e-01 -3.24438989e-01
-3.69428605e-01 4.98019546e-01 1.50696859e-02 -1.27251536e-01
-1.16302228e+00 1.48153469e-01 1.04914749e+00 9.78747569e-03
8.12787056e-01 -7.37409949e-01 -1.94638334e-02 -9.66148555e-01
4.63352017e-02 -4.20566320e-01 3.96949053e-01 -7.05976129e-01
-9.67319071e-01 5.04051387e-01 1.31472886e-01 -1.71493089e+00
-1.73096806e-01 -6.28116131e-01 -2.54063457e-01 8.66678655e-01
-1.79049563e+00 -1.37909102e+00 -5.08106232e-01 5.21812499e-01
3.08607996e-01 6.61844239e-02 3.00471842e-01 4.67296630e-01
-2.77297169e-01 3.95797014e-01 -2.45609879e-03 -1.79713145e-01
9.00063813e-01 -1.17542064e+00 6.45872116e-01 9.54128325e-01
3.01622093e-01 4.44808453e-01 1.06792414e+00 -8.57424438e-01
-1.70762670e+00 -1.10017967e+00 6.56351984e-01 -1.21283734e+00
6.17268443e-01 -3.73293012e-01 -5.99909365e-01 6.95191860e-01
-2.89540350e-01 3.76526535e-01 1.44992828e-01 -2.05903947e-02
-3.71345043e-01 -5.32694161e-02 -1.14196563e+00 2.71599025e-01
1.16278291e+00 -2.73893595e-01 -4.56941515e-01 2.40962282e-01
7.90156126e-01 -1.01621270e+00 -1.08331943e+00 6.60056293e-01
7.07830906e-01 -7.65455186e-01 1.39157951e+00 -1.41689345e-01
-9.07377720e-01 -1.09444320e+00 -1.85706794e-01 -9.92257178e-01
-3.56693357e-01 -6.24799907e-01 -2.43717581e-01 1.00266182e+00
-2.91240923e-02 -3.97401839e-01 8.09105217e-01 1.36959359e-01
-2.79398292e-01 -9.37272161e-02 -1.21799433e+00 -1.20662284e+00
-5.91335773e-01 -4.35637772e-01 5.81275821e-01 5.46471715e-01
-5.60807765e-01 1.22687176e-01 -2.95043975e-01 7.04821825e-01
1.31655371e+00 3.23297411e-01 1.35309803e+00 -1.52130437e+00
-6.11463077e-02 -3.49795580e-01 -9.59246755e-01 -1.27051258e+00
-4.65881526e-02 -8.61070454e-01 3.53659302e-01 -1.01263261e+00
6.10299297e-02 -5.41518271e-01 -2.05961272e-01 -7.77248666e-02
-1.96980551e-01 4.89144802e-01 5.86422682e-01 4.80878234e-01
-1.14648867e+00 4.79265094e-01 1.07598114e+00 9.15219821e-03
-8.72551873e-02 1.39478862e-01 2.06093322e-02 5.36569536e-01
2.55466700e-01 -7.83292472e-01 2.36953378e-01 -4.32188183e-01
-2.50242114e-01 -1.25865474e-01 8.60076904e-01 -1.28600705e+00
4.53728855e-01 1.07227951e-01 1.11921214e-01 -1.35172713e+00
6.75024509e-01 -1.12831116e+00 2.67120093e-01 6.00739598e-01
1.63330883e-01 3.43229949e-01 3.69667798e-01 6.72582507e-01
6.78141639e-02 1.51868522e-01 9.20917928e-01 2.23596901e-01
-1.06774044e+00 4.92363870e-01 1.18713908e-01 -1.54080853e-01
1.18313992e+00 -4.64496613e-01 -1.53443605e-01 3.95105667e-02
-5.36914766e-01 4.48645830e-01 9.03520226e-01 5.95334232e-01
5.45853853e-01 -1.76259565e+00 -5.61353385e-01 1.35480896e-01
5.02793670e-01 -2.21170500e-01 -2.97076285e-01 1.44451594e+00
-3.12396109e-01 5.06377459e-01 -2.04115495e-01 -1.34523690e+00
-1.68449962e+00 5.93533158e-01 4.26105112e-01 -2.59163976e-01
-5.85201859e-01 6.69448435e-01 2.79089063e-02 -3.87388766e-01
3.69878113e-01 -4.06415641e-01 9.86965895e-02 -2.65061736e-01
5.01659811e-01 4.03103650e-01 2.07369223e-01 -1.11457443e+00
-7.71812141e-01 9.67303216e-01 1.32515803e-01 -3.35018933e-02
9.66479421e-01 -4.52976704e-01 2.88221031e-01 3.04295093e-01
1.02893591e+00 -1.10018164e-01 -1.60249269e+00 -5.58703780e-01
4.81258273e-01 -7.88766265e-01 2.30031475e-01 -3.48962277e-01
-6.56008959e-01 4.95917112e-01 1.20330667e+00 6.23617433e-02
3.90010357e-01 9.75039452e-02 7.30529010e-01 1.88154489e-01
5.97696424e-01 -6.00841463e-01 -1.39639914e-01 6.72836781e-01
4.36417788e-01 -1.46886027e+00 3.99218708e-01 -4.26433176e-01
-2.31416881e-01 7.10896969e-01 5.74922681e-01 -1.77600205e-01
4.72431958e-01 4.28582728e-01 1.51432559e-01 -3.82211655e-01
-4.97590989e-01 -7.85718203e-01 5.72995365e-01 6.77012086e-01
1.47704795e-01 -2.92675108e-01 1.22378822e-02 -2.64334977e-01
1.46859214e-01 -1.28237590e-01 -9.10158679e-02 1.12467432e+00
-6.49809122e-01 -9.66206312e-01 -9.25482631e-01 -1.16621017e-01
-1.06621608e-01 3.05396736e-01 -5.25345914e-02 1.06216908e+00
1.28427997e-01 7.41506517e-01 1.56140178e-01 -2.19787478e-01
5.80589712e-01 -2.46335775e-01 8.75319123e-01 -5.15885174e-01
-6.38359368e-01 3.39930087e-01 6.69570416e-02 -8.12552452e-01
-7.46628761e-01 -1.02459824e+00 -1.02003253e+00 -5.35732880e-02
-7.94596791e-01 -1.52185932e-01 1.17540956e+00 7.93852925e-01
4.30721045e-01 2.06360608e-01 3.45051318e-01 -1.38985991e+00
-8.45306337e-01 -5.91128647e-01 -2.64601797e-01 4.06143993e-01
6.16026580e-01 -1.33295345e+00 -5.11943474e-02 -2.51228243e-01] | [6.640886306762695, -2.2243950366973877] |
7c58e776-5334-4354-abb8-74a8dd439bfa | 190910122 | 1909.10122 | null | https://arxiv.org/abs/1909.10122v1 | https://arxiv.org/pdf/1909.10122v1.pdf | Towards Best Experiment Design for Evaluating Dialogue System Output | To overcome the limitations of automated metrics (e.g. BLEU, METEOR) for evaluating dialogue systems, researchers typically use human judgments to provide convergent evidence. While it has been demonstrated that human judgments can suffer from the inconsistency of ratings, extant research has also found that the design of the evaluation task affects the consistency and quality of human judgments. We conduct a between-subjects study to understand the impact of four experiment conditions on human ratings of dialogue system output. In addition to discrete and continuous scale ratings, we also experiment with a novel application of Best-Worst scaling to dialogue evaluation. Through our systematic study with 40 crowdsourced workers in each task, we find that using continuous scales achieves more consistent ratings than Likert scale or ranking-based experiment design. Additionally, we find that factors such as time taken to complete the task and no prior experience of participating in similar studies of rating dialogue system output positively impact consistency and agreement amongst raters | ['Sashank Santhanam', 'Samira Shaikh'] | 2019-09-23 | towards-best-experiment-design-for-evaluating | https://aclanthology.org/W19-8610 | https://aclanthology.org/W19-8610.pdf | ws-2019-10 | ['dialogue-evaluation'] | ['natural-language-processing'] | [-1.57526448e-01 2.85916805e-01 6.59250021e-02 -7.89051056e-01
-7.57930875e-01 -1.06538153e+00 7.59906888e-01 4.56700683e-01
-1.06728458e+00 5.93569815e-01 6.57322943e-01 -5.01543939e-01
1.59093991e-01 -2.50582039e-01 1.16733350e-01 2.54974604e-01
4.93468672e-01 3.28867137e-01 1.57924071e-01 -4.00480479e-01
9.57095981e-01 9.57167372e-02 -1.19126153e+00 3.21737885e-01
6.20945752e-01 4.40951437e-01 -6.50384352e-02 9.88109291e-01
-3.75145487e-02 1.14293921e+00 -1.51932085e+00 -5.81822634e-01
2.94171840e-01 -5.67920566e-01 -9.55028653e-01 8.72490406e-02
3.09746206e-01 -1.00057840e+00 -7.68587813e-02 5.97900033e-01
8.38806868e-01 5.72571099e-01 6.80087924e-01 -1.26383054e+00
-9.92961287e-01 5.04441857e-01 -2.04651967e-01 1.00844875e-01
9.53131676e-01 4.47327673e-01 1.08419049e+00 -7.58518219e-01
5.17041862e-01 1.38805115e+00 6.72962964e-01 4.40722644e-01
-1.25029504e+00 -5.45106947e-01 -3.01314652e-01 -4.17420894e-01
-9.50112939e-01 -7.11458087e-01 7.04135299e-02 -8.48078251e-01
9.13319170e-01 -2.56481371e-03 1.88428849e-01 7.80276179e-01
-9.16202888e-02 -9.71581563e-02 1.84189379e+00 -3.95618349e-01
3.97404909e-01 8.55262160e-01 2.38998979e-01 1.89806700e-01
1.82154313e-01 -1.71132952e-01 -6.01767242e-01 -7.06503808e-01
7.34320283e-01 -4.29675102e-01 -3.89905348e-02 1.41271323e-01
-1.06744492e+00 8.52276623e-01 1.28723696e-01 2.56668985e-01
-2.72466183e-01 -1.92931831e-01 6.84681296e-01 5.86446345e-01
3.15566242e-01 1.32727182e+00 -3.37292880e-01 -8.25423002e-01
-5.36429584e-01 4.33400840e-01 1.40952778e+00 4.80926275e-01
2.39155054e-01 -3.31790268e-01 -5.94502747e-01 1.17868125e+00
2.22661331e-01 1.63925305e-01 5.40859461e-01 -1.38948774e+00
4.25883293e-01 4.30404902e-01 1.04679263e+00 -9.61643934e-01
-4.58591849e-01 3.63288194e-01 2.87224710e-01 6.56185865e-01
9.79213655e-01 -6.27836704e-01 -2.90889204e-01 1.53169692e+00
1.15305059e-01 -1.20392978e+00 -1.76061675e-01 1.26074731e+00
6.42534018e-01 2.58543253e-01 3.61059844e-01 -3.69630843e-01
1.45293188e+00 -7.21081793e-01 -7.93923199e-01 -2.35025659e-02
6.49813056e-01 -1.19577456e+00 1.82376635e+00 4.49382275e-01
-1.13288665e+00 -7.63906181e-01 -1.09144807e+00 -1.97355926e-01
-1.24275826e-01 -9.80374441e-02 3.50891113e-01 1.02282631e+00
-1.11479104e+00 5.58151186e-01 -2.52177119e-01 -6.69578493e-01
-5.79133749e-01 1.11425690e-01 -3.30670506e-01 4.36726153e-01
-1.07345998e+00 1.55911469e+00 -3.25901330e-01 5.22316508e-02
-3.75112057e-01 -1.54203385e-01 -4.63523537e-01 -2.27244988e-01
3.52577418e-02 -2.09012315e-01 2.07088280e+00 -9.39551592e-01
-1.67010272e+00 6.13927782e-01 5.91488034e-02 9.88504514e-02
4.55031455e-01 -2.56836861e-01 7.44157434e-02 -1.12315223e-01
2.53454059e-01 5.49416780e-01 1.46573156e-01 -9.82756615e-01
-5.19358397e-01 -2.86924750e-01 6.23782873e-01 7.14045107e-01
-4.63443607e-01 5.60018718e-01 4.31497872e-01 -2.48225093e-01
-2.83545524e-01 -9.25805449e-01 1.62004940e-02 -2.14525327e-01
3.02329391e-01 -5.82157135e-01 1.65489092e-01 -6.90322459e-01
1.26337707e+00 -1.85034502e+00 -4.11087751e-01 -1.40076324e-01
3.28563839e-01 6.95132837e-02 7.36532407e-03 9.02653515e-01
3.99843395e-01 5.21669567e-01 3.01482022e-01 -1.21223934e-01
2.02997193e-01 -3.73156339e-01 1.15529045e-01 2.86189497e-01
-1.74433798e-01 5.28743863e-01 -8.19916487e-01 -5.66429853e-01
-3.42415087e-02 1.37784228e-01 -2.31195688e-01 5.17553568e-01
2.43346930e-01 1.67080745e-01 -2.72993520e-02 1.07311644e-01
2.36405179e-01 -1.81078061e-01 2.38541916e-01 1.58135116e-01
-3.00276250e-01 7.36195862e-01 -7.36016989e-01 1.35345483e+00
-6.41373336e-01 8.07026863e-01 2.41177216e-01 1.46189388e-02
9.98599827e-01 6.81322455e-01 1.20723397e-02 -8.09985638e-01
1.93013445e-01 1.99359536e-01 4.16795105e-01 -5.17569005e-01
1.21046889e+00 -2.61891276e-01 -7.52163976e-02 1.17290926e+00
-3.01534235e-01 -5.17252207e-01 2.64603555e-01 3.35387737e-01
1.11010170e+00 -1.71882868e-01 3.40352297e-01 -2.11102515e-01
-1.61810964e-01 3.57359856e-01 2.55698621e-01 9.19545472e-01
-7.35688448e-01 3.77575696e-01 5.41517198e-01 -8.86766985e-02
-1.31733930e+00 -7.15673208e-01 -8.32858384e-02 1.46484745e+00
3.80222895e-03 -2.65158445e-01 -7.79460907e-01 -5.46151757e-01
-9.14685875e-02 9.13213015e-01 -3.98983598e-01 2.22587511e-01
-9.18605179e-03 -3.87612551e-01 7.20131516e-01 4.91352141e-01
2.72893637e-01 -1.00265753e+00 -9.28628266e-01 2.73191780e-01
-4.25786704e-01 -9.48410988e-01 -6.62546217e-01 -3.39611322e-01
-5.49916506e-01 -9.03518736e-01 -6.04377925e-01 -4.19321179e-01
3.97820681e-01 6.25775158e-01 1.21768367e+00 3.13995123e-01
-1.75334290e-02 8.18143189e-01 -6.59464359e-01 -4.11211610e-01
-6.44259155e-01 -2.30803400e-01 4.59028929e-02 -8.20250988e-01
5.33495784e-01 -2.01098882e-02 -6.72392845e-01 7.78358102e-01
-6.86867595e-01 -3.44043374e-01 5.15341163e-01 7.17592359e-01
-4.18929219e-01 -3.23979080e-01 1.00694919e+00 -8.10608983e-01
2.02732491e+00 -2.58611560e-01 -1.58952698e-01 2.31787458e-01
-1.10198879e+00 -4.25541073e-01 3.50823611e-01 -4.14492965e-01
-1.10392737e+00 -5.69793761e-01 4.47415233e-01 4.79919910e-01
-1.60297215e-01 3.75332355e-01 6.27767026e-01 -5.00614345e-02
1.30048621e+00 -7.38785744e-01 5.41405797e-01 -2.86260210e-02
6.35728016e-02 1.24760973e+00 8.38037673e-03 -6.46301508e-01
4.44317371e-01 -2.46090397e-01 -6.92918897e-01 -5.04469872e-01
-5.30557036e-01 -3.38156044e-01 -4.16222215e-01 -4.92760479e-01
1.05942333e+00 -1.06519377e+00 -9.70868468e-01 3.56694534e-02
-9.47463751e-01 -8.34048212e-01 7.32369423e-02 8.49453151e-01
-4.00032252e-01 3.71258289e-01 -1.01388514e+00 -1.26901579e+00
-1.89160377e-01 -1.19392467e+00 6.14977717e-01 4.13499981e-01
-1.29926205e+00 -8.26476574e-01 2.93658823e-01 8.55987370e-01
6.70240998e-01 -4.72269244e-02 5.53397596e-01 -9.46233332e-01
2.26770028e-01 -3.96447808e-01 -2.36604646e-01 4.21367139e-01
3.72752666e-01 5.96063212e-02 -9.22126710e-01 -2.00371012e-01
2.28538796e-01 -1.12718427e+00 -3.60477641e-02 1.45897716e-02
1.31596968e-01 -2.98343241e-01 2.37731814e-01 -6.85456693e-01
9.65261459e-01 3.64730269e-01 3.01962674e-01 4.98826593e-01
1.62468359e-01 9.48992372e-01 1.19596314e+00 4.45626646e-01
5.19063711e-01 6.34474754e-01 -3.32000881e-01 1.60799727e-01
3.41562748e-01 -1.35536939e-01 5.51662207e-01 5.39393902e-01
5.94691411e-02 -2.18319684e-01 -8.14227462e-01 2.71096170e-01
-1.54024363e+00 -6.18942916e-01 1.73230648e-01 2.53070855e+00
7.95464516e-01 2.90124178e-01 6.27951741e-01 3.97739513e-03
9.50841665e-01 1.18861265e-01 -1.31645203e-01 -1.05172467e+00
4.66373086e-01 -9.10648480e-02 3.75590324e-01 6.87894642e-01
-3.85374129e-01 4.58794892e-01 7.18106318e+00 2.51074135e-03
-7.03398526e-01 1.86152145e-01 5.93089938e-01 -2.45312020e-01
-1.93642125e-01 1.48685396e-01 -2.44622111e-01 2.15755075e-01
1.02434814e+00 -3.89418989e-01 6.44588828e-01 5.51752388e-01
5.91821492e-01 -4.40866709e-01 -1.21860063e+00 4.11694854e-01
-9.47487652e-02 -3.49278152e-01 -6.73485398e-01 1.41890153e-01
6.06002986e-01 -2.91645586e-01 -2.04279467e-01 4.65336621e-01
6.74477637e-01 -1.17902839e+00 7.52854526e-01 4.58297208e-02
6.74869895e-01 -2.91908920e-01 8.57544422e-01 3.41704667e-01
-3.70123804e-01 3.81222367e-02 -3.80852908e-01 -8.08652997e-01
1.82817489e-01 8.40628073e-02 -1.62552524e+00 -2.16510877e-01
3.83655071e-01 -3.37235063e-01 -5.49657583e-01 5.65357745e-01
-1.24553420e-01 5.26538491e-01 5.98771684e-02 -4.97694671e-01
5.05828150e-02 -1.30938292e-01 1.46702543e-01 1.00590134e+00
-9.06037390e-02 2.32730374e-01 5.95475920e-02 4.79005814e-01
-9.36879311e-03 3.48431230e-01 -7.84311354e-01 -3.39027137e-01
1.09958315e+00 1.44273031e+00 -9.88316953e-01 -2.64523089e-01
-4.58937764e-01 6.39537513e-01 3.16938639e-01 3.60536039e-01
-4.27853376e-01 -5.50695539e-01 2.43997946e-01 1.04895800e-01
-3.11196178e-01 -5.11224806e-01 -7.48755634e-01 -5.25478482e-01
3.00433576e-01 -1.36254275e+00 4.24291985e-03 -8.32498372e-01
-1.21098804e+00 5.47210693e-01 -8.59234035e-02 -9.34122801e-01
-2.06391335e-01 -3.03909093e-01 -4.54661041e-01 1.12401986e+00
-5.84677696e-01 -3.10199678e-01 -5.17561972e-01 -5.34917936e-02
6.48733139e-01 3.52011621e-02 7.79508650e-01 -1.28017038e-01
-1.03863813e-01 6.17801309e-01 -2.89118499e-01 5.83324628e-03
1.74095166e+00 -1.48299408e+00 4.11135226e-01 2.25216746e-01
-2.29836971e-01 1.21747005e+00 8.96214306e-01 -7.40965426e-01
-9.91226017e-01 -3.55430953e-02 7.22707272e-01 -8.05055797e-01
5.77412546e-01 -1.61001191e-01 -6.64195240e-01 3.08143198e-01
6.21654868e-01 -8.81473303e-01 9.22124863e-01 5.51871419e-01
-4.91285175e-01 1.49631038e-01 -1.34959555e+00 5.67583144e-01
7.33273566e-01 -1.09409988e+00 -6.07310176e-01 4.60880786e-01
6.28487051e-01 -4.30856615e-01 -1.42576706e+00 1.70973539e-01
1.07229567e+00 -9.47587252e-01 3.45107704e-01 -4.07517880e-01
6.86334610e-01 -1.71178281e-01 5.38236834e-02 -1.49803007e+00
-2.83746749e-01 -6.08603001e-01 7.11737871e-01 1.36074686e+00
6.91207349e-01 -5.90120137e-01 3.40846956e-01 1.76195383e+00
7.80108850e-03 -5.09532034e-01 -3.08586210e-01 -4.84565258e-01
1.71005607e-01 2.44598106e-01 2.74107277e-01 8.26081038e-01
7.54906774e-01 7.62754321e-01 -1.95767149e-01 -2.58686900e-01
-4.88302223e-02 -5.64296901e-01 1.24898911e+00 -1.08505964e+00
-3.62490267e-01 -3.47421020e-01 -1.14408642e-01 -7.62610316e-01
-1.93506375e-01 -2.09312513e-01 4.07733977e-01 -1.68926549e+00
2.57116169e-01 -3.08222115e-01 9.36292037e-02 2.53620762e-02
-5.27596533e-01 -1.07394373e-02 5.17342985e-01 4.72790509e-01
-7.36202657e-01 1.79868147e-01 1.10129845e+00 4.54920888e-01
-5.22166252e-01 -1.40993640e-01 -1.26257002e+00 2.41666481e-01
7.46023536e-01 -3.18036556e-01 -6.02642894e-01 -5.40563047e-01
3.50891471e-01 4.38225001e-01 1.02469005e-01 -6.81234777e-01
4.39101398e-01 -2.14028031e-01 1.75389305e-01 2.74772555e-01
1.87769040e-01 -3.76028627e-01 -2.56816387e-01 5.93988597e-02
-9.23953593e-01 8.47570837e-01 -1.03222556e-01 2.98605025e-01
-1.30508691e-01 -6.71752036e-01 4.22321260e-01 -2.73402482e-01
-5.18644415e-02 -6.65744603e-01 -8.33739817e-01 2.47111320e-01
7.37338603e-01 -3.76097023e-01 -7.20450640e-01 -1.01480150e+00
-2.02079430e-01 2.39385620e-01 9.44161654e-01 4.16496307e-01
-1.45426868e-02 -8.74369442e-01 -6.21638000e-01 -7.48470783e-01
1.42484024e-01 -3.84014904e-01 -2.65164584e-01 7.46099472e-01
-6.10216677e-01 4.01892781e-01 -2.40196794e-01 -2.10659161e-01
-1.15265727e+00 5.90034723e-02 1.00339860e-01 -7.21593350e-02
-2.80899624e-03 5.77212155e-01 -1.10637531e-01 -5.20905614e-01
3.05632412e-01 5.35595752e-02 -2.97392339e-01 2.31354684e-01
5.82844079e-01 7.19142020e-01 6.99266940e-02 -3.01231146e-01
8.55549276e-02 -1.27599284e-01 -3.11020017e-01 -1.17628181e+00
6.91220105e-01 -2.88277954e-01 6.14090376e-02 1.11944807e+00
6.76653981e-01 3.14720869e-01 -1.06873417e+00 7.48145878e-02
-4.97581735e-02 -6.99962497e-01 -4.21445280e-01 -1.05686295e+00
2.84864791e-02 4.75615621e-01 4.33471560e-01 8.62579525e-01
5.09626329e-01 -5.18740356e-01 3.25917482e-01 7.66275704e-01
6.40570879e-01 -1.68637097e+00 3.72866422e-01 3.16547483e-01
9.32189643e-01 -1.43005955e+00 1.95869520e-01 1.00706592e-01
-1.29180956e+00 8.84748042e-01 9.55727041e-01 1.72353387e-01
1.89164415e-01 -6.65331110e-02 6.49487615e-01 -1.97019249e-01
-9.44783688e-01 4.18070465e-01 -2.59963781e-01 4.24745113e-01
1.43722796e+00 2.54765004e-01 -1.24436069e+00 3.31956625e-01
-6.23340607e-01 1.03582956e-01 1.04394066e+00 1.10801291e+00
-4.91965681e-01 -1.03769469e+00 -5.30628085e-01 6.77368999e-01
-5.99606216e-01 2.12902978e-01 -1.16270876e+00 7.50973344e-01
-7.04658329e-01 1.86573100e+00 -1.57375187e-02 -6.16241038e-01
6.21647239e-01 1.69697478e-01 2.67928869e-01 -1.03110099e+00
-1.39078033e+00 -4.92255569e-01 7.49768436e-01 -2.63848633e-01
-4.18815643e-01 -6.43902719e-01 -8.63110244e-01 -5.75429678e-01
-7.57213950e-01 6.84481680e-01 7.63470173e-01 8.70664537e-01
2.06287995e-01 1.43831167e-02 7.85964251e-01 -4.28891897e-01
-1.29106379e+00 -1.83758187e+00 -5.21763802e-01 7.33027041e-01
4.34912927e-02 -7.04130113e-01 -6.00144565e-01 -3.66531938e-01] | [12.778443336486816, 8.06870174407959] |
13f72c29-3588-47cc-a4cc-39b71622ce7b | hierarchy-of-visual-words-a-learning-based | 1908.02786 | null | https://arxiv.org/abs/1908.02786v1 | https://arxiv.org/pdf/1908.02786v1.pdf | Hierarchy-of-Visual-Words: a Learning-based Approach for Trademark Image Retrieval | In this paper, we present the Hierarchy-of-Visual-Words (HoVW), a novel trademark image retrieval (TIR) method that decomposes images into simpler geometric shapes and defines a descriptor for binary trademark image representation by encoding the hierarchical arrangement of component shapes. The proposed hierarchical organization of visual data stores each component shape as a visual word. It is capable of representing the geometry of individual elements and the topology of the trademark image, making the descriptor robust against linear as well as to some level of nonlinear transformation. Experiments show that HoVW outperforms previous TIR methods on the MPEG-7 CE-1 and MPEG-7 CE-2 image databases. | ['Vítor N. Lourenço', 'Leandro A. F. Fernandes', 'Gabriela G. Silva'] | 2019-08-07 | null | null | null | null | ['trademark-retrieval'] | ['computer-vision'] | [-2.21462920e-01 -4.16628003e-01 -2.77507961e-01 -2.32400805e-01
-4.07788038e-01 -1.13062453e+00 8.76416504e-01 5.52712142e-01
-1.13897726e-01 -5.36892116e-02 7.75101110e-02 -1.91249818e-01
-3.60527277e-01 -7.47034907e-01 -4.21507359e-01 -5.00497043e-01
-2.45773867e-01 6.17454946e-01 5.98833382e-01 -2.08898127e-01
5.72206259e-01 1.15977430e+00 -1.64956355e+00 5.05626202e-01
-9.70459953e-02 1.59035420e+00 5.91279604e-02 5.22335470e-01
-2.86216438e-01 4.03991371e-01 -5.74332833e-01 -3.71565700e-01
6.55227900e-01 1.52079850e-01 -5.31985879e-01 5.52191138e-01
9.60915387e-01 -2.07918465e-01 -3.83944511e-01 1.00127637e+00
4.93728407e-02 -7.21769361e-03 1.21813643e+00 -1.38102388e+00
-1.28986657e+00 -2.05133379e-01 -5.21119297e-01 4.40502129e-02
3.84858221e-01 -3.08918595e-01 1.04895544e+00 -1.11887550e+00
9.56403673e-01 1.47579277e+00 5.29268026e-01 -1.55704752e-01
-1.11594975e+00 -3.26672316e-01 -3.25258017e-01 3.72342497e-01
-1.85270846e+00 -9.98940691e-02 7.09451377e-01 -7.31849313e-01
8.07776272e-01 4.63232309e-01 6.06214643e-01 3.32014740e-01
6.88857198e-01 2.84798831e-01 1.05295181e+00 -2.39447191e-01
1.59029365e-01 -2.87942082e-01 3.81494433e-01 9.43532765e-01
4.84791070e-01 -1.53753618e-02 -2.95594215e-01 -4.69322085e-01
1.03821135e+00 1.48786321e-01 2.10668892e-01 -9.01910126e-01
-8.70918274e-01 5.72573483e-01 4.52702343e-01 4.38624561e-01
-3.27191055e-01 2.25656524e-01 2.27522060e-01 4.78556693e-01
7.52146319e-02 1.14370815e-01 6.15000874e-02 3.88025045e-01
-7.71941125e-01 1.79847419e-01 4.99032736e-01 1.39141047e+00
7.55383968e-01 1.77097127e-01 1.35704637e-01 7.42473423e-01
6.36532307e-01 7.48223007e-01 5.77963889e-01 -9.82921898e-01
1.02178484e-01 6.46246850e-01 -9.01374444e-02 -1.57234502e+00
-1.06266268e-01 8.38071555e-02 -6.61769152e-01 4.33623314e-01
-9.65450630e-02 1.06386745e+00 -1.20226586e+00 8.60753715e-01
3.33374031e-02 -3.84258568e-01 -7.18345270e-02 4.50921476e-01
1.17298794e+00 9.19087052e-01 -5.43546230e-02 -2.51291189e-02
1.66730881e+00 -6.70526147e-01 -7.28321373e-01 2.11470321e-01
5.19812992e-03 -1.21267402e+00 6.02544427e-01 2.32886598e-01
-1.10023856e+00 -9.27953959e-01 -1.27067220e+00 -1.64097115e-01
-8.94066989e-01 5.84036633e-02 1.53944790e-01 6.60327554e-01
-1.50056458e+00 2.35293791e-01 -2.71801591e-01 -3.46109539e-01
9.15895775e-03 3.97988647e-01 -6.85229182e-01 -8.19059387e-02
-6.52032852e-01 4.53777581e-01 6.54862344e-01 -3.10639560e-01
-7.97631443e-01 -4.95301634e-01 -1.04481804e+00 1.90125927e-02
1.15184218e-01 -2.62030780e-01 6.64958358e-01 -5.77966332e-01
-8.95806968e-01 1.18371356e+00 -1.15270875e-01 -2.19499424e-01
2.17432067e-01 4.26169276e-01 -7.19404042e-01 6.75802052e-01
-9.44800116e-03 8.08096826e-01 1.28680611e+00 -1.75158560e+00
-4.78106081e-01 -5.84479570e-01 -1.65304571e-01 -8.17612186e-02
-1.22948796e-01 -1.45770550e-01 -9.61786032e-01 -1.06844890e+00
4.49148834e-01 -8.77957463e-01 2.10946172e-01 2.39867553e-01
8.83521438e-02 -3.92668188e-01 1.18596327e+00 -4.17317063e-01
1.30634117e+00 -2.61215520e+00 -2.90008038e-01 7.51110911e-01
3.22480589e-01 3.20966691e-01 -4.05937582e-01 7.52226174e-01
-2.45045781e-01 3.06883574e-01 1.11405611e-01 -5.31047955e-02
-3.32317222e-03 5.44913471e-01 -3.84115040e-01 5.41580975e-01
-2.04478383e-01 1.06449509e+00 -4.43035513e-01 -8.80705237e-01
3.46709162e-01 4.21796560e-01 -8.85350332e-02 5.03981374e-02
2.49314621e-01 -4.17407125e-01 -3.94098103e-01 8.65719736e-01
1.05855668e+00 -2.28432436e-02 9.21125561e-02 -7.19037890e-01
-2.68715888e-01 -7.53820896e-01 -1.21824455e+00 9.17644799e-01
2.25221276e-01 4.76711005e-01 1.14212679e-02 -7.18539894e-01
1.40426242e+00 5.44254594e-02 6.81451321e-01 -9.80158687e-01
-1.45295858e-02 2.20245048e-01 -6.62511528e-01 -3.74872267e-01
8.28662813e-01 1.80025011e-01 -2.65963912e-01 1.76152438e-01
5.61027080e-02 -2.77843356e-01 3.11760932e-01 1.22744769e-01
6.34681165e-01 -1.60478070e-01 4.40581113e-01 -4.55956668e-01
7.47102559e-01 5.64992130e-02 9.54801738e-02 6.24194026e-01
-1.01357251e-01 8.63974512e-01 1.85277462e-01 -5.42412579e-01
-1.44737017e+00 -1.55757010e+00 -2.80883014e-01 9.27507818e-01
4.53637630e-01 -5.49332619e-01 -6.03627980e-01 -2.16336012e-01
4.40652400e-01 1.24765508e-01 -5.37094891e-01 2.51827240e-01
-5.23390293e-01 -1.42825127e-01 4.75787073e-01 3.82793367e-01
4.55051512e-01 -7.84211874e-01 -4.60338622e-01 -1.10365422e-02
2.91148573e-01 -1.19800818e+00 -1.10773635e+00 -2.73130924e-01
-8.23819458e-01 -1.14894581e+00 -8.65878761e-01 -1.39098525e+00
9.83321249e-01 7.14837432e-01 8.76165748e-01 2.50095755e-01
-5.90426683e-01 1.02687204e+00 -5.43708563e-01 1.24020316e-01
-4.17345971e-01 -5.85370541e-01 -1.62158996e-01 3.86573225e-01
2.15633195e-02 -1.00499824e-01 -4.05877739e-01 5.07652819e-01
-1.59122372e+00 -4.77253228e-01 3.99705708e-01 4.91384923e-01
1.14606559e+00 1.91851854e-01 -1.51253864e-01 -5.06490409e-01
8.29603910e-01 -7.72535130e-02 -8.63696396e-01 7.05337167e-01
-7.99785912e-01 2.75256246e-01 4.42547977e-01 -2.05380157e-01
-5.06781340e-01 2.06202984e-01 3.46314341e-01 -7.42114902e-01
-6.01750799e-02 3.96173060e-01 -2.07060024e-01 -5.29208779e-01
1.98982716e-01 6.00366414e-01 1.54364854e-01 -6.59355164e-01
6.47599280e-01 7.29188442e-01 9.87170041e-01 -4.13191319e-01
1.05332971e+00 6.07302725e-01 4.29508179e-01 -1.32452774e+00
-9.63152424e-02 -1.00443017e+00 -1.01372540e+00 -3.35880280e-01
9.50397849e-01 -5.69780231e-01 -7.32938349e-01 4.24248308e-01
-1.06689918e+00 4.25158501e-01 -1.36924908e-01 -1.23421833e-01
-6.66686118e-01 9.46985781e-01 -4.73714083e-01 -3.79949361e-01
-2.07345665e-01 -1.10252452e+00 1.24744666e+00 -3.56663287e-01
2.58438513e-02 -9.76568937e-01 -7.37701058e-02 2.10744515e-01
8.97857398e-02 3.41515124e-01 1.21625686e+00 -5.00244141e-01
-6.15358233e-01 -4.27134693e-01 -4.28472340e-01 3.60503882e-01
4.21102904e-02 2.85804421e-01 -1.03718646e-01 -3.83186281e-01
-7.99799487e-02 -4.70746271e-02 6.54900134e-01 1.26304463e-01
9.83542621e-01 -4.48025674e-01 2.02999841e-02 6.30225778e-01
1.89225638e+00 7.56086707e-01 8.48113477e-01 3.88414383e-01
6.41932189e-01 5.53987503e-01 5.73196113e-01 3.35366189e-01
3.03422928e-01 8.09529424e-01 3.61232370e-01 -1.09450504e-01
-3.46429050e-01 -2.51349568e-01 1.69188544e-01 1.10683012e+00
-1.42824143e-01 -1.00347377e-01 -8.77179205e-01 3.59721899e-01
-1.62542570e+00 -9.03200865e-01 -4.80010271e-01 2.16899467e+00
2.20311925e-01 -1.70600727e-01 1.64083838e-01 -5.88153973e-02
7.14753330e-01 3.85738313e-01 -1.32550895e-01 -6.48746550e-01
-4.36854422e-01 9.45899561e-02 9.77781117e-01 4.70991671e-01
-1.00411224e+00 7.50996709e-01 7.45230055e+00 1.17481005e+00
-9.18259025e-01 -2.75453217e-02 3.54807317e-01 7.84171641e-01
-2.69377768e-01 -1.94191784e-01 -5.99044919e-01 3.22879732e-01
5.96070707e-01 -3.74923915e-01 6.94127306e-02 7.74262369e-01
-2.49963567e-01 9.44793001e-02 -7.22891808e-01 1.39390254e+00
5.45896947e-01 -1.31495845e+00 9.67206895e-01 3.38878781e-01
5.47424078e-01 -4.58671093e-01 4.86679912e-01 -4.03763264e-01
-1.67259961e-01 -1.01482463e+00 1.11661017e+00 4.94135082e-01
9.12485242e-01 -6.67080641e-01 5.10639071e-01 -2.92408198e-01
-1.90514338e+00 -2.18358580e-02 -5.76992512e-01 8.97116423e-01
-1.14265725e-01 -2.52820879e-01 -4.48868126e-01 6.99031472e-01
5.82094312e-01 5.70025682e-01 -1.08004320e+00 1.08606935e+00
4.74465400e-01 -5.94817922e-02 -8.04336295e-02 1.79980442e-01
5.91581702e-01 -3.52168113e-01 3.61905187e-01 1.26781201e+00
4.94987607e-01 3.60397488e-01 4.03260767e-01 4.65372175e-01
1.48525253e-01 3.56625229e-01 -7.66880870e-01 -2.59774208e-01
3.71499360e-01 9.40949380e-01 -1.07747817e+00 -5.84822416e-01
-1.81478590e-01 1.00882530e+00 -2.84521192e-01 4.21948850e-01
-4.09090579e-01 -3.63858372e-01 3.31844538e-01 5.08861244e-01
8.41270804e-01 -9.03433204e-01 -3.46456431e-02 -5.76343596e-01
-9.48913395e-02 -6.12384200e-01 5.74334979e-01 -1.04388535e+00
-1.04431474e+00 7.46549428e-01 4.97506529e-01 -1.73979223e+00
-7.96446055e-02 -1.05972207e+00 4.08787318e-02 3.90838742e-01
-1.13972986e+00 -1.39012587e+00 -1.79498121e-01 8.35270345e-01
3.70075703e-01 -5.27099788e-01 6.76691651e-01 2.11095139e-01
1.13977626e-01 5.49206853e-01 5.96044779e-01 1.28954723e-01
5.35617352e-01 -1.10540175e+00 3.07271689e-01 2.07773119e-01
3.25290471e-01 5.09060502e-01 5.06269693e-01 -7.32754350e-01
-1.60797882e+00 -9.45635021e-01 8.51130784e-01 -2.62417495e-01
6.31789565e-01 -3.93256024e-02 -8.49869013e-01 2.93442100e-01
2.54650950e-01 4.63364981e-02 6.37241960e-01 -9.39289033e-01
-9.90975440e-01 -3.89515787e-01 -1.19505000e+00 3.69987726e-01
6.96255326e-01 -5.78839481e-01 -9.67226088e-01 5.19674011e-02
5.06565332e-01 5.20951897e-02 -1.28635073e+00 1.90244332e-01
7.25074649e-01 -8.58471394e-01 1.57604349e+00 -2.02369690e-01
-6.92650005e-02 -5.99098980e-01 -4.79350209e-01 -7.93075323e-01
-6.17193639e-01 -3.29783231e-01 2.58442640e-01 9.66238737e-01
-7.11707845e-02 -5.17764330e-01 5.44719636e-01 1.47336289e-01
5.59889562e-02 -3.08998227e-01 -1.08923626e+00 -1.34960938e+00
-3.23940426e-01 -2.22863033e-01 5.53501546e-01 7.98702240e-01
-5.17279744e-01 -7.34252334e-02 -1.54860675e-01 -4.66094725e-02
1.08386123e+00 1.28439605e-01 3.87679636e-01 -1.28797007e+00
1.40465349e-01 -6.60229862e-01 -1.21051633e+00 -8.42012703e-01
-1.32997155e-01 -8.28947484e-01 -3.72645169e-01 -1.45712507e+00
-9.59235132e-02 8.69612545e-02 -2.89279342e-01 3.97328466e-01
7.43846714e-01 8.36120903e-01 5.43210149e-01 7.75771201e-01
-7.40452528e-01 2.32634395e-01 1.19106150e+00 -4.84496087e-01
2.11679012e-01 -6.60825729e-01 -7.71928057e-02 4.21874702e-01
3.63809288e-01 -1.55898824e-01 -2.67875284e-01 -2.30269358e-01
-2.68254355e-02 -4.29130904e-02 2.09182769e-01 -7.83368945e-01
3.70457500e-01 -1.12970181e-01 3.33134681e-01 -1.23165441e+00
3.86790782e-01 -1.28181016e+00 6.85623825e-01 6.34125888e-01
-2.06023499e-01 7.08223343e-01 3.60532850e-02 6.14927888e-01
-6.00359738e-01 -5.29461265e-01 7.33854175e-01 -1.06038526e-01
-1.22667754e+00 2.96280414e-01 -4.71700579e-01 -3.13679844e-01
1.22140384e+00 -8.41536343e-01 -2.75756419e-01 -3.06323439e-01
-6.78698003e-01 1.16108945e-02 6.68751359e-01 6.61931813e-01
1.16677749e+00 -1.98198843e+00 -4.21007365e-01 5.85642278e-01
5.42102337e-01 -7.88005471e-01 1.07417881e-01 1.77564532e-01
-1.26140118e+00 7.38585413e-01 -6.47844255e-01 -4.61702257e-01
-1.80930710e+00 1.03347468e+00 -6.66908845e-02 7.74877667e-02
-4.28634226e-01 6.58194050e-02 4.73736018e-01 1.78360030e-01
2.56978661e-01 -6.44652471e-02 -4.82640088e-01 4.22509223e-01
5.65157115e-01 4.91263181e-01 6.99386820e-02 -1.58908963e+00
-3.47062469e-01 1.43088603e+00 6.77968413e-02 -5.62082082e-02
1.02292168e+00 -2.23580211e-01 -5.12764692e-01 2.95831352e-01
1.83784890e+00 -1.55452564e-01 -5.69079995e-01 -2.12142393e-01
2.49141589e-01 -6.81538284e-01 -1.95915118e-01 -1.52700529e-01
-9.47478354e-01 5.45538068e-01 9.94396508e-01 4.59954351e-01
9.55600679e-01 1.70154139e-01 6.43878996e-01 4.14056480e-01
5.53462267e-01 -1.15315723e+00 4.47373569e-01 5.33787608e-01
1.33191156e+00 -5.52849114e-01 2.11146504e-01 -6.88437402e-01
-4.74913895e-01 1.53587413e+00 -1.39414161e-01 -2.81357378e-01
1.12794876e+00 1.78775951e-01 2.06175730e-01 -5.93219697e-01
-1.77880913e-01 -2.65288711e-01 9.61796880e-01 7.33000755e-01
-1.19060539e-02 1.27261177e-01 -5.05046427e-01 -7.66730532e-02
-4.47183847e-02 -4.92542773e-01 3.80913228e-01 1.05291748e+00
-7.29014516e-01 -1.23265851e+00 -9.74088371e-01 6.68294579e-02
-3.49834621e-01 1.06109023e-01 -7.92509675e-01 9.71772432e-01
5.04092760e-02 5.66066682e-01 4.30300713e-01 -3.86648804e-01
6.24617398e-01 -2.23609000e-01 4.06930000e-01 -3.12252879e-01
-4.59549546e-01 5.66466153e-01 -3.51753116e-01 -5.63018560e-01
-4.30265516e-01 -4.06286836e-01 -1.16090453e+00 -1.89226121e-01
3.70517433e-01 3.77684273e-02 5.84559917e-01 2.74425298e-01
3.51919204e-01 -8.46250653e-02 7.42223799e-01 -8.48696411e-01
-4.17271256e-01 -2.28993565e-01 -1.11513889e+00 9.85587418e-01
3.09230238e-01 -6.46950185e-01 -2.12187290e-01 4.78480011e-01] | [10.551675796508789, 0.04197406768798828] |
941f687f-892b-41aa-84cf-ed5f6262e6ca | latent-space-models-for-multiplex-networks | 2012.14409 | null | https://arxiv.org/abs/2012.14409v2 | https://arxiv.org/pdf/2012.14409v2.pdf | Latent space models for multiplex networks with shared structure | Latent space models are frequently used for modeling single-layer networks and include many popular special cases, such as the stochastic block model and the random dot product graph. However, they are not well-developed for more complex network structures, which are becoming increasingly common in practice. Here we propose a new latent space model for multiplex networks: multiple, heterogeneous networks observed on a shared node set. Multiplex networks can represent a network sample with shared node labels, a network evolving over time, or a network with multiple types of edges. The key feature of our model is that it learns from data how much of the network structure is shared between layers and pools information across layers as appropriate. We establish identifiability, develop a fitting procedure using convex optimization in combination with a nuclear norm penalty, and prove a guarantee of recovery for the latent positions as long as there is sufficient separation between the shared and the individual latent subspaces. We compare the model to competing methods in the literature on simulated networks and on a multiplex network describing the worldwide trade of agricultural products. | ['Ji Zhu', 'Elizaveta Levina', 'Peter W. MacDonald'] | 2020-12-28 | null | null | null | null | ['stochastic-block-model'] | ['graphs'] | [ 3.01101536e-01 1.58220023e-01 -4.76681262e-01 -1.14220984e-01
2.15432093e-01 -7.43926764e-01 5.76394141e-01 -3.09219003e-01
3.15228179e-02 6.67337298e-01 1.30915150e-01 -1.62990257e-01
-8.53696167e-01 -7.85476625e-01 -6.20042920e-01 -1.02983224e+00
-5.54691732e-01 7.16668963e-01 1.43183768e-01 2.47604668e-01
-2.79432714e-01 5.28399467e-01 -8.88283372e-01 4.38489541e-02
6.54175758e-01 4.38940346e-01 -1.51831478e-01 5.22117913e-01
-4.87836562e-02 6.57346129e-01 -2.85412282e-01 -2.93369889e-01
3.81850839e-01 -2.53716648e-01 -5.28852701e-01 6.35303080e-01
1.32631823e-01 1.20076582e-01 -7.93237329e-01 1.14338887e+00
4.61024279e-03 -2.65730560e-01 8.93249571e-01 -1.78945351e+00
-1.54104441e-01 1.00157619e+00 -7.93368459e-01 -1.90954387e-01
-1.51827231e-01 -2.44598895e-01 1.29821849e+00 -4.74844515e-01
8.90642464e-01 1.39116585e+00 4.51744080e-01 2.19170600e-01
-2.16217661e+00 -8.28671575e-01 4.86351669e-01 -3.42244148e-01
-1.36677849e+00 -4.89382595e-01 8.37802231e-01 -8.97480011e-01
3.08120936e-01 1.42349213e-01 7.40524411e-01 1.24239910e+00
2.16566399e-01 7.27054179e-01 9.91236031e-01 -9.02342722e-02
1.48794875e-01 5.04350998e-02 2.55679846e-01 7.56324410e-01
4.52114761e-01 -1.36567891e-01 -6.01347268e-01 -4.63597894e-01
1.12346733e+00 3.51210713e-01 -1.38711467e-01 -1.15587497e+00
-1.30302286e+00 7.84899175e-01 2.35130697e-01 2.06029832e-01
-3.62246454e-01 2.71509677e-01 3.50173516e-03 7.30592132e-01
6.41778529e-01 -8.40700045e-02 -2.23764166e-01 3.73390824e-01
-1.20788193e+00 -2.61013508e-02 1.20430899e+00 1.20588064e+00
8.19942117e-01 -1.00128122e-01 1.82705864e-01 5.82332790e-01
7.06444800e-01 3.46953064e-01 -2.95249224e-01 -9.66903806e-01
6.66996121e-01 7.40929067e-01 2.18886770e-02 -1.00071740e+00
-3.90468568e-01 -7.19453216e-01 -1.30872846e+00 3.55251990e-02
4.69020694e-01 -3.66464823e-01 -5.79640090e-01 2.13856030e+00
-9.49916840e-02 4.70881641e-01 -7.08990023e-02 3.64840806e-01
6.27382696e-02 5.66870272e-01 -3.55103642e-01 -5.30902326e-01
9.08122122e-01 -7.65364408e-01 -7.54870772e-01 -1.80473089e-01
4.29783523e-01 -4.34492707e-01 2.52236098e-01 3.81821334e-01
-1.07897055e+00 -1.53816994e-02 -8.32052112e-01 4.95441079e-01
-9.98786837e-02 -8.55383463e-03 7.21190393e-01 6.57584190e-01
-1.34208572e+00 6.92170680e-01 -1.02060246e+00 -6.58842683e-01
2.66822755e-01 7.03096449e-01 -6.33306205e-01 -2.04407588e-01
-9.11811233e-01 1.36140004e-01 5.04394025e-02 3.52896005e-01
-1.09912992e+00 -4.34773505e-01 -5.49196839e-01 2.07948282e-01
5.77419817e-01 -5.77499986e-01 4.58179682e-01 -9.35446382e-01
-1.32263303e+00 5.96539497e-01 -7.35111609e-02 -1.51851013e-01
6.90680921e-01 1.75726071e-01 -4.75446403e-01 -1.33622978e-02
4.66334075e-02 2.75377065e-01 8.29997718e-01 -1.21097839e+00
-1.40910313e-01 -2.38810942e-01 3.87376696e-02 -1.31220818e-01
-3.51412416e-01 -4.82470021e-02 -3.48399371e-01 -4.03533012e-01
7.13718414e-01 -1.33940506e+00 -2.45211273e-01 3.45443070e-01
-9.63055372e-01 1.57197699e-01 6.77069902e-01 -2.61103749e-01
1.12920928e+00 -2.26780581e+00 8.26627731e-01 8.30096304e-01
6.64535165e-01 -3.14237624e-01 -3.63388091e-01 8.54949355e-01
-4.23571587e-01 3.43011737e-01 -2.69780576e-01 -6.62799716e-01
1.78754732e-01 2.92115688e-01 -1.81117684e-01 8.68641198e-01
9.09654200e-02 6.44908547e-01 -9.58103418e-01 -2.16440707e-01
-2.57663935e-01 2.95656562e-01 -4.02696729e-01 -2.26126947e-02
-1.81131698e-02 3.13054532e-01 -3.08457702e-01 3.10077816e-01
6.05273783e-01 -7.46895730e-01 7.95856535e-01 -1.46267563e-01
-7.33581409e-02 -9.75518599e-02 -1.60975528e+00 1.50618660e+00
1.11180380e-01 6.55793786e-01 6.30232155e-01 -9.65645194e-01
6.25543058e-01 6.51311755e-01 6.59318805e-01 3.27381492e-01
-3.44780646e-02 5.18690310e-02 3.04345936e-01 -1.69916183e-01
-9.64987576e-02 -9.31652784e-02 3.71103585e-02 7.91636705e-01
1.86783686e-01 3.83939296e-01 3.73719156e-01 5.68820596e-01
1.38659847e+00 -2.19834313e-01 -3.55390534e-02 -5.28263092e-01
9.44721550e-02 -5.52375913e-01 7.17504919e-01 6.82576478e-01
9.59723145e-02 2.34007224e-01 1.41031134e+00 2.37931475e-01
-1.07846236e+00 -1.19115722e+00 9.10568237e-03 7.27235198e-01
-5.83519116e-02 -3.01169097e-01 -4.61516380e-01 -4.41212982e-01
1.07686453e-01 -1.15246184e-01 -6.53024018e-01 -4.45349962e-02
-1.69483691e-01 -8.31559718e-01 3.53731930e-01 1.53550923e-01
2.20723256e-01 -6.37428880e-01 4.10017937e-01 2.17155159e-01
-2.14300640e-02 -9.90393877e-01 -4.51960713e-01 3.12927037e-01
-1.12449062e+00 -9.95786071e-01 -7.44159818e-01 -6.82773888e-01
1.04283798e+00 2.54762381e-01 8.36975038e-01 -1.35673821e-01
-5.36090061e-02 3.88047278e-01 2.16033131e-01 3.74916017e-01
-2.85725862e-01 3.86935681e-01 3.67133647e-01 7.17483401e-01
-6.39549792e-02 -9.20112312e-01 -2.33785182e-01 5.11274040e-01
-1.02600849e+00 2.03537539e-01 6.09813631e-01 7.09257722e-01
1.75496951e-01 4.76702154e-01 1.63179129e-01 -1.24369383e+00
4.75002706e-01 -1.03092372e+00 -7.00745881e-01 4.47706640e-01
-3.70054752e-01 2.06640869e-01 8.88904631e-02 -6.53650284e-01
-6.83892906e-01 -6.19128905e-02 7.85936773e-01 -4.86259818e-01
1.32905886e-01 7.96943009e-01 -4.65055972e-01 -4.44518663e-02
8.78536850e-02 -1.14466541e-01 2.12774351e-01 -6.14582121e-01
3.02199692e-01 2.18942806e-01 -1.57693416e-01 -5.56204677e-01
1.19938946e+00 7.90423155e-01 4.92682159e-01 -9.81310010e-01
-5.21550536e-01 -4.69323277e-01 -9.73997831e-01 -1.10566087e-01
5.07937968e-01 -9.15406764e-01 -7.60901213e-01 5.11296034e-01
-9.55359697e-01 -4.40070212e-01 -1.25860825e-01 5.10603607e-01
-4.31979537e-01 2.06600204e-01 -9.27109122e-01 -9.85091448e-01
5.18059552e-01 -9.53941286e-01 5.12925208e-01 -2.40463898e-01
-1.40682071e-01 -1.55851746e+00 2.42518798e-01 -6.15786053e-02
1.90331757e-01 2.61282653e-01 1.06654751e+00 -3.40977997e-01
-1.03862548e+00 -1.75049707e-01 -3.23732674e-01 1.32441506e-01
1.78765759e-01 4.10190284e-01 -4.56474960e-01 -6.02817893e-01
-2.65727878e-01 1.42187044e-01 8.78554881e-01 4.97892529e-01
5.55089891e-01 -3.41438949e-01 -7.22494125e-01 6.47154272e-01
1.23147202e+00 -2.49127388e-01 3.04799825e-01 -3.51312488e-01
7.57977545e-01 1.13537145e+00 -2.00190708e-01 2.54127383e-01
6.96996972e-02 5.31505466e-01 4.16402817e-01 -8.62999111e-02
3.87908757e-01 -2.48365894e-01 4.43049878e-01 1.25679219e+00
5.24858087e-02 -6.13403201e-01 -8.79431427e-01 5.75512052e-01
-2.22303200e+00 -9.06637132e-01 -2.46880427e-01 2.26084852e+00
4.63572800e-01 2.73750961e-01 1.88298523e-01 7.01166317e-02
1.03972077e+00 1.93706691e-01 -5.70151627e-01 3.65882844e-01
-5.16923308e-01 -2.92479157e-01 7.18247950e-01 7.04109728e-01
-1.02188563e+00 5.90197146e-01 7.14491844e+00 4.90709513e-01
-6.00578547e-01 4.44747545e-02 4.08042699e-01 -2.79828519e-01
-4.18098629e-01 3.43913406e-01 -6.79870844e-01 3.22604805e-01
7.95166910e-01 1.86185315e-02 4.65099752e-01 2.15039983e-01
9.85527784e-02 2.23062620e-01 -1.25640082e+00 6.23842239e-01
-3.30178320e-01 -1.08580792e+00 -1.80321142e-01 7.51716852e-01
1.12810218e+00 1.75048001e-02 1.98372215e-01 -2.84041435e-01
7.63301730e-01 -8.66796851e-01 5.06633043e-01 8.05450439e-01
6.13223195e-01 -5.00757813e-01 3.91355187e-01 4.35563236e-01
-1.25204265e+00 -1.04745217e-02 -2.63286352e-01 -6.98348358e-02
2.76373804e-01 7.77175665e-01 -2.38447696e-01 4.93826061e-01
1.28089711e-01 1.22310877e+00 -2.02863216e-01 1.05798733e+00
-1.99870765e-01 7.92663574e-01 -4.88455921e-01 3.44059765e-01
1.68922678e-01 -8.16009521e-01 8.54068339e-01 7.23886490e-01
2.69758493e-01 -5.10193586e-01 2.67251730e-01 1.02883720e+00
-1.08442850e-01 -2.00420916e-01 -7.12642372e-01 -6.21605217e-01
4.21018809e-01 1.31641448e+00 -1.14131498e+00 2.11839695e-02
-4.98257577e-01 9.06202435e-01 2.85873204e-01 1.05896139e+00
-4.17049259e-01 -7.25539550e-02 8.75667572e-01 3.27703267e-01
1.86631829e-01 -4.72452760e-01 2.60260165e-01 -1.37558126e+00
-1.25256911e-01 -3.79515290e-01 3.92357893e-02 -3.75527561e-01
-1.61779535e+00 2.53544837e-01 3.27624194e-02 -1.12949681e+00
3.72837745e-02 -4.65338230e-01 -4.47857350e-01 9.48951006e-01
-1.09992206e+00 -1.12434959e+00 1.98588651e-02 4.89538252e-01
-1.80719748e-01 -3.68409127e-01 4.64750648e-01 4.21879917e-01
-9.84023333e-01 1.13516420e-01 6.65908396e-01 9.06075612e-02
4.64453042e-01 -1.19378340e+00 2.75026739e-01 7.86975265e-01
2.41781384e-01 9.27992582e-01 5.98859549e-01 -8.82165790e-01
-1.18902135e+00 -9.56192136e-01 7.14438200e-01 -1.64462507e-01
1.22730696e+00 -1.03657353e+00 -8.11311960e-01 1.19560647e+00
-5.24668545e-02 1.14690751e-01 7.48361588e-01 3.64891469e-01
-3.61819714e-01 -1.99519657e-02 -8.30308497e-01 6.94431067e-01
1.31601346e+00 -6.42799497e-01 3.26782018e-01 4.40750301e-01
6.39113426e-01 2.16491148e-01 -8.80397201e-01 2.44661234e-02
5.27306139e-01 -7.89370239e-01 7.98651934e-01 -7.16659546e-01
2.07406342e-01 -3.57057571e-01 -1.43179804e-01 -1.19481134e+00
-6.48993134e-01 -1.05586767e+00 -2.50012994e-01 1.51565766e+00
4.84596312e-01 -8.35387826e-01 1.12592196e+00 4.76363689e-01
4.93073940e-01 -3.66802037e-01 -9.16103363e-01 -9.45071459e-01
-1.32810026e-01 1.49743214e-01 4.05763835e-01 1.08684099e+00
-2.68357601e-02 3.37318569e-01 -6.28600180e-01 2.96770066e-01
1.26820183e+00 -1.91376194e-01 5.02798319e-01 -1.76924634e+00
-4.26590621e-01 -5.26249290e-01 -4.91770923e-01 -1.14248097e+00
4.85555232e-01 -9.29416895e-01 -2.80597955e-01 -1.38675284e+00
4.99024689e-01 -7.54563987e-01 -2.77753353e-01 1.36996880e-01
3.74372840e-01 1.61238983e-02 -4.03822539e-03 5.03061712e-01
-5.63542664e-01 4.23951834e-01 9.95961905e-01 -1.52504236e-01
-1.15365043e-01 3.00580770e-01 -4.96134907e-01 5.98752081e-01
4.11835462e-01 -6.46261156e-01 -6.49895608e-01 -2.21382543e-01
8.57993424e-01 3.85396034e-01 2.97658473e-01 -7.04421759e-01
4.43334222e-01 -1.98539838e-01 4.43108268e-02 -4.45583165e-01
4.85495925e-01 -1.11514115e+00 9.29047823e-01 4.10204351e-01
-4.80695933e-01 -2.52129823e-01 -3.88282776e-01 1.06958365e+00
-3.35607305e-02 -1.65435165e-01 4.85152721e-01 1.06459334e-01
-6.28909236e-03 7.51411974e-01 -3.79790813e-01 -1.69673219e-01
1.08624077e+00 -6.63269311e-02 -3.07336450e-01 -5.38619518e-01
-1.22564995e+00 4.56425607e-01 6.60585046e-01 1.84216097e-01
2.49189198e-01 -1.45395017e+00 -6.97957337e-01 3.42762113e-01
-1.09459400e-01 -2.59101540e-01 2.27244347e-01 1.06836498e+00
-2.14002937e-01 2.61290818e-01 -1.05739601e-01 -6.49113894e-01
-1.24943340e+00 3.94704878e-01 1.82395488e-01 -4.96527314e-01
-4.07282323e-01 7.34760582e-01 5.38948119e-01 -5.10161817e-01
1.78450584e-01 -6.49651629e-04 -1.38239756e-01 4.14083123e-01
1.37587398e-01 5.91779590e-01 -7.27134585e-01 -6.47304535e-01
-1.31681383e-01 4.62066293e-01 -8.59576389e-02 -3.42793465e-01
1.40394449e+00 -4.39042777e-01 -7.56404459e-01 1.16540217e+00
1.12002778e+00 -1.53019741e-01 -1.55742180e+00 -7.60127723e-01
1.72734246e-01 -2.00943798e-01 -2.36003533e-01 -2.35295184e-02
-1.28224444e+00 9.16296601e-01 1.54812574e-01 7.67287374e-01
6.52660370e-01 1.68432027e-01 -1.26334518e-01 1.01426892e-01
3.85694474e-01 -5.24252057e-01 7.72219971e-02 4.30277616e-01
6.09844685e-01 -7.27663934e-01 1.04737811e-01 -7.37799406e-01
-7.82327279e-02 9.55816865e-01 4.16998602e-02 -4.10878032e-01
1.20584154e+00 2.38389507e-01 -4.85907882e-01 -3.48702073e-01
-1.00625026e+00 9.87536162e-02 1.69941500e-01 4.17839795e-01
3.77996087e-01 2.35353321e-01 3.12143147e-01 2.22386390e-01
1.91847906e-01 -3.41658622e-01 7.45004058e-01 6.84818387e-01
-9.27985907e-02 -1.34947336e+00 -1.87915824e-02 6.86995924e-01
-2.08239555e-01 -5.38382083e-02 -2.98572868e-01 5.47190905e-01
-1.52529150e-01 7.53565252e-01 2.17726007e-01 -1.19367018e-01
1.85311623e-02 1.38388008e-01 5.09781778e-01 -9.90186691e-01
-1.46592751e-01 4.72980201e-01 -4.58338521e-02 -3.89309227e-01
-6.62872374e-01 -1.04497576e+00 -3.31240386e-01 -6.51090086e-01
-6.32425904e-01 1.11504860e-01 5.59939504e-01 8.82911623e-01
1.03390984e-01 6.28642082e-01 5.35517454e-01 -6.90355241e-01
-2.48517767e-01 -9.09822762e-01 -1.48061550e+00 1.55151069e-01
6.02508962e-01 -7.88666666e-01 -7.45062590e-01 -4.57073620e-04] | [7.0004425048828125, 5.249223232269287] |
fa6b090d-a368-4745-8428-e0cfffc564b3 | manipulating-visually-aware-federated | 2305.08183 | null | https://arxiv.org/abs/2305.08183v2 | https://arxiv.org/pdf/2305.08183v2.pdf | Manipulating Visually-aware Federated Recommender Systems and Its Countermeasures | Federated recommender systems (FedRecs) have been widely explored recently due to their ability to protect user data privacy. In FedRecs, a central server collaboratively learns recommendation models by sharing model public parameters with clients, thereby offering a privacy-preserving solution. Unfortunately, the exposure of model parameters leaves a backdoor for adversaries to manipulate FedRecs. Existing works about FedRec security already reveal that items can easily be promoted by malicious users via model poisoning attacks, but all of them mainly focus on FedRecs with only collaborative information (i.e., user-item interactions). We argue that these attacks are effective because of the data sparsity of collaborative signals. In practice, auxiliary information, such as products' visual descriptions, is used to alleviate collaborative filtering data's sparsity. Therefore, when incorporating visual information in FedRecs, all existing model poisoning attacks' effectiveness becomes questionable. In this paper, we conduct extensive experiments to verify that incorporating visual information can beat existing state-of-the-art attacks in reasonable settings. However, since visual information is usually provided by external sources, simply including it will create new security problems. Specifically, we propose a new kind of poisoning attack for visually-aware FedRecs, namely image poisoning attacks, where adversaries can gradually modify the uploaded image to manipulate item ranks during FedRecs' training process. Furthermore, we reveal that the potential collaboration between image poisoning attacks and model poisoning attacks will make visually-aware FedRecs more vulnerable to being manipulated. To safely use visual information, we employ a diffusion model in visually-aware FedRecs to purify each uploaded image and detect the adversarial images. | ['Chaoqun Yang', 'Hongzhi Yin', 'Quoc Viet Hung Nguyen', 'Shilong Yuan', 'Wei Yuan'] | 2023-05-14 | null | null | null | null | ['collaborative-filtering'] | ['miscellaneous'] | [-1.09670490e-01 -2.63010740e-01 -1.06768608e-01 -1.03370450e-01
-4.11153674e-01 -1.53959548e+00 2.83872396e-01 -5.15939593e-01
-1.83302656e-01 1.44717574e-01 1.90107614e-01 -5.57955921e-01
1.94428280e-01 -1.10941947e+00 -9.42110062e-01 -8.31876576e-01
6.81049302e-02 -2.39284575e-01 -1.18488297e-01 -1.27194643e-01
-4.11923751e-02 5.51669419e-01 -1.05084169e+00 4.93722677e-01
5.36112845e-01 8.27769816e-01 -1.01736411e-01 4.83967274e-01
3.27665955e-02 6.30236685e-01 -5.35224378e-01 -1.01511133e+00
1.00980127e+00 -2.70747066e-01 -4.52795893e-01 1.45673737e-01
2.23410189e-01 -7.82269418e-01 -9.62084293e-01 1.53633976e+00
3.52879316e-01 -1.84148923e-01 2.35687822e-01 -1.54269242e+00
-1.37871313e+00 8.21533561e-01 -5.79412222e-01 -1.53890833e-01
2.77946621e-01 6.97189033e-01 8.64298701e-01 -5.84237576e-01
5.66103458e-01 1.16231143e+00 4.29369807e-01 8.84283602e-01
-1.16591632e+00 -1.18503594e+00 5.10821104e-01 -7.67109822e-03
-1.34755158e+00 -2.60566831e-01 8.17426205e-01 -3.08157414e-01
-4.11072969e-02 9.45588887e-01 5.55061162e-01 1.30073917e+00
-3.37287664e-01 9.85942125e-01 1.16389132e+00 4.58847098e-02
3.22670639e-01 5.87547779e-01 2.75128353e-02 5.30835092e-01
4.50596333e-01 4.42334652e-01 -3.11658114e-01 -9.05327141e-01
9.02406394e-01 5.90204358e-01 -6.88852727e-01 -5.98098278e-01
-7.41411805e-01 9.23530638e-01 5.99999368e-01 -1.90907568e-01
-4.41094674e-02 4.08615097e-02 1.52620092e-01 6.92867398e-01
9.81739536e-02 2.85214275e-01 -4.42881614e-01 6.71803296e-01
-1.54571295e-01 1.90399885e-01 6.48247838e-01 9.26870584e-01
5.52756667e-01 -6.44987524e-02 -2.46616125e-01 3.84355217e-01
5.82316399e-01 7.66075313e-01 3.38774830e-01 -7.82164097e-01
3.23898256e-01 4.88193363e-01 4.57161635e-01 -1.27378476e+00
2.06389323e-01 -3.68120462e-01 -7.79724956e-01 2.29764432e-01
5.71040273e-01 -5.16587436e-01 -5.03575921e-01 1.94624507e+00
6.33599579e-01 2.43745312e-01 9.31649506e-02 1.28914857e+00
6.33907735e-01 5.16607583e-01 -1.83479730e-02 -9.36177969e-02
1.48613691e+00 -8.36396992e-01 -3.88713390e-01 2.65187770e-01
5.77191651e-01 -4.46181118e-01 1.25178599e+00 4.64254200e-01
-7.93329298e-01 -1.43820301e-01 -9.57895577e-01 3.49105775e-01
-3.49319428e-01 -1.77439973e-01 8.43510807e-01 1.11028755e+00
-6.06590569e-01 2.66356975e-01 -4.86391991e-01 -4.13663127e-02
8.53699982e-01 5.64693153e-01 -4.54753250e-01 -3.06903929e-01
-1.21916890e+00 6.32651821e-02 -3.91260982e-01 -2.72476286e-01
-1.12465012e+00 -7.37771869e-01 -5.36490738e-01 1.34334430e-01
6.25113904e-01 -8.54405642e-01 1.16532087e+00 -9.65649307e-01
-1.30064487e+00 5.56510031e-01 4.47734296e-01 -6.44572616e-01
7.60308146e-01 3.53380619e-03 -5.73652625e-01 6.84983805e-02
-4.30014610e-01 1.35082006e-01 1.13318813e+00 -1.61184573e+00
-4.00225341e-01 -6.20269299e-01 7.01332986e-01 5.86945843e-03
-9.30443645e-01 1.79591835e-01 -5.36521554e-01 -7.87116528e-01
-2.27024138e-01 -9.88882840e-01 -5.37838399e-01 5.80110371e-01
-5.13281703e-01 2.40280494e-01 1.06806695e+00 -2.38475010e-01
1.00278628e+00 -2.47518849e+00 -4.35864210e-01 4.22140837e-01
4.83832806e-01 5.00124454e-01 -3.57192963e-01 3.95350039e-01
2.18417078e-01 5.93401134e-01 2.56986290e-01 -1.85190618e-01
5.02050966e-02 6.70966730e-02 -6.93450928e-01 8.67211938e-01
-6.83471560e-01 9.71627235e-01 -6.52101696e-01 -7.16823041e-02
9.21176076e-02 5.13134956e-01 -7.55500793e-01 2.40190953e-01
-3.16352278e-01 3.90322208e-01 -7.47119069e-01 4.99448180e-01
1.24223113e+00 -4.02701735e-01 1.71995744e-01 -2.26188406e-01
2.29061276e-01 -2.11086556e-01 -1.05329192e+00 1.30255234e+00
-2.53555447e-01 -1.70450389e-01 5.52217901e-01 -4.57920820e-01
5.78065693e-01 3.77820671e-01 2.65606791e-01 -3.73279452e-01
1.99649170e-01 -3.16516608e-01 -3.47594529e-01 -2.62856752e-01
2.75518019e-02 1.53287783e-01 6.46568276e-03 9.17786479e-01
-4.02749956e-01 5.40300727e-01 -4.89805818e-01 5.56141376e-01
1.19275975e+00 -2.25821033e-01 -2.31801167e-01 5.72229773e-02
4.53295499e-01 -1.95967972e-01 5.77452779e-01 1.24539661e+00
-1.70303047e-01 5.04499137e-01 7.87376314e-02 -4.52135652e-01
-6.91606164e-01 -9.68849897e-01 1.90380931e-01 1.10037494e+00
6.89401031e-01 -5.88896334e-01 -7.02350318e-01 -1.47418988e+00
2.65628278e-01 3.81109595e-01 -6.33826494e-01 -3.68252248e-01
-1.53050125e-01 -6.48002088e-01 4.85282421e-01 1.83010131e-01
5.40718138e-01 -7.35605717e-01 -1.20463364e-01 -4.79680896e-02
1.48495391e-01 -6.60417318e-01 -9.60262716e-01 -4.35437649e-01
-4.88428503e-01 -1.40839541e+00 -6.48677468e-01 -3.59889686e-01
9.96227622e-01 1.18073750e+00 5.11541307e-01 3.79337609e-01
-1.68854762e-02 5.36390841e-01 -3.81597430e-01 -2.77566284e-01
-4.28178221e-01 -4.36280370e-01 7.72616342e-02 5.59018910e-01
3.08158100e-01 -5.92369556e-01 -1.06453109e+00 7.45476902e-01
-1.19969487e+00 -5.08155376e-02 2.70853072e-01 5.79923391e-01
3.56576025e-01 2.01653138e-01 3.63170236e-01 -1.37092352e+00
5.65982223e-01 -5.59082806e-01 -8.77849758e-01 4.14986134e-01
-4.07337815e-01 -3.89586508e-01 1.21565270e+00 -1.04721045e+00
-8.53172481e-01 2.36341909e-01 9.63891894e-02 -8.60999763e-01
-2.10058957e-01 1.33073106e-01 -6.37150943e-01 -5.25281727e-01
7.20905662e-01 2.46702313e-01 -5.90652004e-02 -8.75255048e-01
8.27186108e-01 7.89664686e-01 4.49010491e-01 -3.19080383e-01
1.36966407e+00 8.71025383e-01 -3.33517283e-01 -2.16947779e-01
-5.71776628e-01 -2.83901393e-01 2.71117985e-01 -1.20090023e-01
4.17106837e-01 -1.08055174e+00 -1.06897199e+00 5.86906254e-01
-8.61707509e-01 1.26771763e-01 -2.78201073e-01 3.39364767e-01
-6.80239648e-02 6.03925169e-01 -7.82606006e-01 -7.73777306e-01
-4.76091236e-01 -1.05581605e+00 2.68067539e-01 1.97596326e-01
4.17546868e-01 -6.52819335e-01 -1.93072230e-01 5.82121611e-01
5.23359895e-01 -6.84040785e-02 6.49343371e-01 -9.49084520e-01
-8.15477312e-01 -4.96092767e-01 -1.03192285e-01 2.53276736e-01
1.56392634e-01 -3.37239742e-01 -9.20704246e-01 -6.36778653e-01
4.26114529e-01 -1.67392582e-01 5.45635700e-01 -1.09433152e-01
1.53523183e+00 -1.19735944e+00 -3.26794297e-01 7.71781862e-01
1.34768808e+00 1.38182178e-01 5.03897488e-01 4.02763523e-02
8.32103908e-01 3.10680300e-01 3.32266659e-01 6.15870833e-01
1.79902241e-01 5.37811756e-01 8.51496518e-01 -2.25104183e-01
1.99905708e-01 -6.08885825e-01 4.33781773e-01 1.71594217e-01
2.74337709e-01 -4.57337946e-01 -2.04322115e-02 4.26490381e-02
-1.83558702e+00 -8.09466422e-01 -3.59262861e-02 2.30187154e+00
6.67376816e-01 -2.99161881e-01 2.67460644e-01 -2.40176812e-01
9.09677267e-01 -5.99956093e-03 -8.29234898e-01 2.10554432e-02
-1.53836727e-01 -2.76261151e-01 9.94281113e-01 2.72006281e-02
-1.08516645e+00 7.57176101e-01 5.42691755e+00 7.45036483e-01
-1.00868404e+00 3.79799396e-01 6.52798891e-01 -3.67932647e-01
-7.06940234e-01 1.38918638e-01 -5.73539495e-01 6.01520717e-01
3.68103981e-01 -2.59802639e-01 8.20473850e-01 1.15344644e+00
7.31551126e-02 5.51560819e-01 -8.02680135e-01 8.65210652e-01
-1.99338272e-02 -1.43290341e+00 3.20661426e-01 4.52604353e-01
8.27915788e-01 -2.35619918e-01 4.93241131e-01 1.16768353e-01
1.03529716e+00 -7.75612831e-01 5.62404692e-01 2.25755706e-01
5.72462797e-01 -8.11969757e-01 2.70114213e-01 4.20385242e-01
-8.99656415e-01 -4.91912574e-01 -7.72146165e-01 1.88739121e-01
-1.84013937e-02 2.76561588e-01 -1.33679137e-01 4.72423702e-01
7.13263571e-01 2.90261954e-01 -5.01580060e-01 1.10657048e+00
-2.40879267e-01 8.24280143e-01 -4.25466806e-01 2.02016294e-01
-1.61818452e-02 -1.72330230e-01 5.62484086e-01 6.35826588e-01
1.77899241e-01 3.54833752e-01 2.82293886e-01 1.01646745e+00
-6.03134036e-01 3.25354725e-01 -9.09419596e-01 8.57592598e-02
5.86517215e-01 1.24314547e+00 -2.13675082e-01 -1.84778981e-02
-7.56396353e-01 1.11952293e+00 5.44632114e-02 4.13891375e-01
-8.02270889e-01 7.18109831e-02 9.67279136e-01 2.00554147e-01
5.19198954e-01 9.49719995e-02 -6.90385178e-02 -1.42741966e+00
-1.66569099e-01 -1.30123103e+00 5.99616051e-01 -5.38054168e-01
-1.74400425e+00 2.44266719e-01 -4.47963625e-01 -1.26639497e+00
4.88405198e-01 -9.26520377e-02 -6.66189849e-01 5.78968108e-01
-1.02737987e+00 -1.43124700e+00 -4.26473888e-03 1.29637647e+00
5.94573803e-02 -1.86316028e-01 8.77577722e-01 2.06952959e-01
-5.17954946e-01 1.26075566e+00 3.05142432e-01 2.93485314e-01
5.23749650e-01 -7.52479196e-01 4.90731657e-01 9.18557346e-01
6.52474701e-01 1.02316701e+00 4.71704662e-01 -5.90702474e-01
-2.14121604e+00 -1.37396443e+00 -2.64882334e-02 -4.33589101e-01
5.87685168e-01 -7.01667249e-01 -8.30112040e-01 6.82749748e-01
-2.11582571e-01 5.43665111e-01 9.80324030e-01 -3.09946358e-01
-9.74196374e-01 -3.08595411e-02 -1.59484935e+00 9.57383573e-01
1.25792623e+00 -6.81254625e-01 1.35495542e-02 4.99800831e-01
9.63196516e-01 -5.83077446e-02 -5.36731422e-01 8.29502791e-02
5.92075408e-01 -8.03267241e-01 1.05512142e+00 -9.08217967e-01
-1.35531679e-01 -5.68368495e-01 -1.26716182e-01 -8.28486562e-01
-3.94366145e-01 -1.09937048e+00 -4.01743650e-01 1.33177042e+00
1.93018645e-01 -6.75151348e-01 9.49551761e-01 8.80232453e-01
5.88106334e-01 -2.54370540e-01 -4.81375068e-01 -7.99579978e-01
5.73613010e-02 -3.56307685e-01 1.14449239e+00 1.02327251e+00
-1.19051732e-01 -4.33383994e-02 -9.79252875e-01 7.09986091e-01
7.01270878e-01 1.91613391e-01 1.18613982e+00 -9.48226273e-01
-7.05047131e-01 8.64269435e-02 -1.03010952e-01 -1.33565056e+00
-8.28121528e-02 -7.76128054e-01 -5.13183832e-01 -9.58110809e-01
3.07103723e-01 -6.82556510e-01 -5.08711159e-01 6.98393703e-01
1.32614434e-01 5.83064437e-01 5.77151239e-01 5.99640906e-01
-6.64354682e-01 1.54011950e-01 1.41730475e+00 -2.72741467e-01
-1.81587487e-01 4.61365461e-01 -1.36453521e+00 5.24961412e-01
6.98763251e-01 -6.62928760e-01 -6.45161629e-01 -1.34355098e-01
2.33073056e-01 1.65146182e-03 5.38657010e-01 -5.17215788e-01
3.26980889e-01 -3.01571906e-01 1.82486415e-01 -2.09097892e-01
-3.94572038e-03 -1.25831962e+00 6.08844459e-01 4.20674205e-01
-3.43127251e-01 -4.17433292e-01 -2.61828750e-01 9.20285225e-01
3.48891884e-01 -1.27458885e-01 6.57368600e-01 -2.93221653e-01
-5.87224998e-02 7.68135071e-01 -3.80260825e-01 -4.14291143e-01
1.10150623e+00 5.41724637e-02 -4.95440453e-01 -7.20845401e-01
-6.77987099e-01 9.84428599e-02 1.01963377e+00 4.73810315e-01
4.02821422e-01 -1.28109610e+00 -4.98620629e-01 4.96240705e-01
1.10676840e-01 -3.16605061e-01 4.56037313e-01 3.82401913e-01
3.59608489e-03 -1.82765931e-01 1.77088961e-01 -2.85440292e-02
-1.62078369e+00 1.41703296e+00 1.65272981e-01 5.28191477e-02
-7.61671484e-01 8.53728235e-01 8.80704165e-01 -3.14118475e-01
5.16328752e-01 3.23911995e-01 -5.20815589e-02 -5.50235093e-01
9.58393276e-01 2.11938862e-02 -3.60221535e-01 -5.16024292e-01
-1.38831690e-01 1.84985638e-01 -5.08517087e-01 1.56645969e-01
9.82429147e-01 -3.92158389e-01 2.30539680e-01 -5.39460063e-01
1.19408000e+00 5.72740436e-01 -1.41700220e+00 -4.12883013e-01
-7.42648482e-01 -9.38917875e-01 -9.65465046e-03 -9.27619636e-01
-1.75475407e+00 4.60653096e-01 7.04850435e-01 4.28712010e-01
1.03251529e+00 -1.54347524e-01 9.41872895e-01 3.13083082e-01
7.28518426e-01 -4.10831481e-01 -1.76527634e-01 -1.86445594e-01
6.31322682e-01 -1.00458872e+00 -1.73495427e-01 -6.21261656e-01
-7.22075760e-01 5.40193021e-01 5.33912063e-01 -8.05312619e-02
8.64187837e-01 2.16183260e-01 2.20772624e-01 -7.08114430e-02
-6.18255496e-01 3.32901865e-01 -3.17543298e-02 8.53302717e-01
-4.43056792e-01 4.58888113e-02 -1.08190529e-01 1.35151422e+00
5.73539734e-02 -1.03966288e-01 4.68170702e-01 7.47296214e-01
-1.24110527e-01 -1.38061249e+00 -6.14045739e-01 4.01923805e-01
-8.30298483e-01 -4.64657582e-02 -5.42227328e-01 2.72593886e-01
8.39391276e-02 1.21327400e+00 -4.80550855e-01 -6.84816062e-01
2.25403622e-01 -4.90971982e-01 1.71894774e-01 -3.27830464e-01
-1.06286657e+00 1.69699099e-02 -3.22789222e-01 -5.87275803e-01
5.05669452e-02 -2.74158150e-01 -8.49632919e-01 -6.66450679e-01
-7.46781528e-01 3.04585129e-01 5.72574317e-01 4.61879522e-01
5.93319952e-01 -3.12102228e-01 1.43529713e+00 -1.94373101e-01
-1.05028021e+00 -2.39020362e-01 -7.75605083e-01 6.81013942e-01
4.39342633e-02 -2.40026504e-01 -7.59741545e-01 -1.11133710e-01] | [5.820383071899414, 7.1865386962890625] |
20a72d15-63bf-4298-a0dc-7ff9ba69a301 | higher-order-neural-additive-models-an | 2209.15409 | null | https://arxiv.org/abs/2209.15409v1 | https://arxiv.org/pdf/2209.15409v1.pdf | Higher-order Neural Additive Models: An Interpretable Machine Learning Model with Feature Interactions | Black-box models, such as deep neural networks, exhibit superior predictive performances, but understanding their behavior is notoriously difficult. Many explainable artificial intelligence methods have been proposed to reveal the decision-making processes of black box models. However, their applications in high-stakes domains remain limited. Recently proposed neural additive models (NAM) have achieved state-of-the-art interpretable machine learning. NAM can provide straightforward interpretations with slight performance sacrifices compared with multi-layer perceptron. However, NAM can only model 1$^{\text{st}}$-order feature interactions; thus, it cannot capture the co-relationships between input features. To overcome this problem, we propose a novel interpretable machine learning method called higher-order neural additive models (HONAM) and a feature interaction method for high interpretability. HONAM can model arbitrary orders of feature interactions. Therefore, it can provide the high predictive performance and interpretability that high-stakes domains need. In addition, we propose a novel hidden unit to effectively learn sharp-shape functions. We conducted experiments using various real-world datasets to examine the effectiveness of HONAM. Furthermore, we demonstrate that HONAM can achieve fair AI with a slight performance sacrifice. The source code for HONAM is publicly available. | ['Jinho Kim', 'Hyun-Soo Choi', 'Minkyu Kim'] | 2022-09-30 | null | null | null | null | ['additive-models'] | ['methodology'] | [ 1.35822132e-01 5.28150856e-01 -4.97813374e-01 -6.29899502e-01
-3.24544728e-01 -1.20660588e-01 2.90499538e-01 -2.32061386e-01
-1.36739880e-01 7.29205787e-01 6.55827224e-02 -5.96524596e-01
-4.73915637e-01 -7.75387645e-01 -7.72446156e-01 -5.94251752e-01
1.66507587e-01 4.89734143e-01 -2.63282299e-01 -1.86118394e-01
5.70922196e-02 -1.38680236e-02 -1.39369404e+00 6.37253821e-01
1.27108943e+00 1.31225765e+00 6.91303834e-02 3.03684980e-01
-1.05658881e-01 1.01578963e+00 -4.35957164e-01 -7.33167648e-01
2.59124547e-01 -2.36968175e-01 -7.74914861e-01 -2.31545195e-01
-5.65684587e-03 -6.62183821e-01 -3.28798383e-01 8.64090919e-01
-4.02380386e-03 -5.06284125e-02 8.10932338e-01 -1.48473668e+00
-1.33345342e+00 9.94962275e-01 -4.01900679e-01 5.65053485e-02
-6.37937114e-02 3.38444561e-01 1.56084859e+00 -8.99595261e-01
-1.04731895e-01 1.52091289e+00 5.44921398e-01 6.40917659e-01
-1.03924751e+00 -8.15977633e-01 4.33122963e-01 3.33814561e-01
-9.49364066e-01 -7.72120384e-03 8.13138545e-01 -1.36141986e-01
8.61109078e-01 5.97920954e-01 5.44845462e-01 9.99018610e-01
2.53391862e-01 1.28038323e+00 1.10261619e+00 -2.47043371e-01
1.93857163e-01 -2.12856606e-02 3.75974417e-01 9.31852281e-01
4.31808412e-01 4.79211025e-02 -5.30178308e-01 -1.93250447e-03
9.53923881e-01 5.96685946e-01 3.59048061e-02 3.91871065e-01
-1.04720211e+00 9.36986148e-01 9.55737352e-01 5.96719310e-02
-3.14917088e-01 4.39869136e-01 -4.39487286e-02 3.03613365e-01
6.59668326e-01 7.98991323e-01 -5.62960446e-01 -6.42031245e-03
-3.70360643e-01 3.14130075e-02 5.64069450e-01 9.31461036e-01
7.17554867e-01 1.71560004e-01 -2.11865917e-01 6.74747348e-01
4.38952893e-01 2.69933701e-01 4.96311665e-01 -8.73757541e-01
4.16959882e-01 1.22472823e+00 -1.39756024e-01 -1.23633707e+00
-4.15595651e-01 -4.57632542e-01 -1.34946072e+00 7.79759958e-02
1.30183741e-01 -1.53846934e-01 -9.99396801e-01 1.58079374e+00
-1.06392629e-01 -1.65273726e-01 -3.94191034e-02 9.11213756e-01
8.77483606e-01 7.39238143e-01 -6.65791109e-02 1.15289457e-01
1.48270583e+00 -1.24099803e+00 -7.75034249e-01 -5.52266777e-01
4.61979717e-01 -7.87106454e-02 1.47223544e+00 2.75560677e-01
-1.03932464e+00 -4.87320632e-01 -1.08276272e+00 -3.36395204e-01
-1.93050757e-01 2.12673098e-01 1.24146247e+00 3.56620967e-01
-6.45737171e-01 5.70108712e-01 -9.46922541e-01 3.80377710e-01
8.89570177e-01 7.43531406e-01 -2.06113368e-01 8.01429823e-02
-1.28190100e+00 7.35117853e-01 4.90065247e-01 3.58083010e-01
-7.25275517e-01 -5.68411469e-01 -7.83773363e-01 5.08772016e-01
5.74560642e-01 -8.89661312e-01 1.38646531e+00 -1.36150098e+00
-1.44453084e+00 3.98043603e-01 -3.34042072e-01 -4.28808600e-01
4.58375990e-01 -3.65864396e-01 -2.16980457e-01 -2.07212254e-01
-1.48492903e-01 7.13855147e-01 7.13941097e-01 -1.12801921e+00
-5.92311025e-01 -3.24425101e-01 5.26892602e-01 7.85033330e-02
-7.01633573e-01 -2.54148066e-01 9.25548151e-02 -6.89269066e-01
2.66319841e-01 -7.65514195e-01 -3.34160209e-01 2.87306756e-01
-4.60079640e-01 -3.62186581e-01 5.64326644e-01 -4.66937780e-01
1.36921859e+00 -2.07516098e+00 4.57542390e-02 -1.19957499e-01
8.19509685e-01 1.55259445e-01 -5.03815487e-02 -1.09803125e-01
1.39886767e-01 7.95344472e-01 -2.47625932e-01 -3.30709577e-01
3.07867110e-01 4.95828211e-01 -3.03568751e-01 -1.55856416e-01
5.19743145e-01 1.33590138e+00 -7.57546782e-01 -1.26676321e-01
-3.82738374e-02 4.31654863e-02 -6.66042566e-01 1.05680309e-01
-3.78382713e-01 1.05352193e-01 -7.29620159e-01 6.80353045e-01
4.92133975e-01 -7.54206836e-01 -1.01991102e-01 3.81720066e-03
3.82100374e-01 1.15311451e-01 -8.05384457e-01 9.52969551e-01
-3.89542192e-01 7.51496315e-01 -4.09088075e-01 -9.04224515e-01
9.96687233e-01 1.26732409e-01 -1.34342685e-01 -6.60331726e-01
8.07402208e-02 4.02470380e-01 4.16601509e-01 -3.37397099e-01
3.83839637e-01 -3.44268620e-01 -2.49302853e-02 3.66247654e-01
-2.66631603e-01 2.66800404e-01 -1.82092428e-01 -1.90238073e-01
1.06109416e+00 -2.00432405e-01 5.22676885e-01 -2.81327158e-01
1.82293221e-01 -1.83677182e-01 8.13652396e-01 8.99822116e-01
9.16852057e-02 4.93634015e-01 7.38753140e-01 -1.14089346e+00
-1.04090559e+00 -8.22833300e-01 -6.05282038e-02 1.15010011e+00
3.13691378e-01 -3.42185706e-01 -6.13723874e-01 -6.87179327e-01
1.21178880e-01 7.47824013e-01 -8.36407721e-01 -3.98272693e-01
-3.90831679e-01 -8.38746071e-01 5.50771654e-01 8.86562288e-01
9.02357161e-01 -1.05688083e+00 -5.65017402e-01 2.22246703e-02
-1.42192230e-01 -8.46500039e-01 -1.84302241e-01 2.55171359e-01
-9.28569853e-01 -7.24014580e-01 -3.65376979e-01 -5.41453362e-01
9.77799177e-01 1.62540555e-01 1.11705351e+00 2.87339419e-01
9.74678397e-02 -3.70963991e-01 -2.63093412e-01 -8.73288155e-01
-1.84146225e-01 1.62126645e-01 1.55769110e-01 2.51278188e-02
5.30975223e-01 -4.26675677e-01 -7.25466073e-01 5.75045228e-01
-1.15490401e+00 7.68006861e-01 1.00402749e+00 1.17161238e+00
3.74750614e-01 8.86268169e-02 7.80544341e-01 -9.04831052e-01
7.37094879e-01 -6.09419346e-01 -2.96215236e-01 3.47866267e-01
-8.09720516e-01 4.09886330e-01 1.12556648e+00 -4.82881784e-01
-1.07098091e+00 -3.82918328e-01 1.03332475e-01 -3.57547522e-01
1.53681636e-02 8.92181516e-01 -3.19806188e-01 3.17773998e-01
5.87392509e-01 -1.78948063e-02 -1.69772923e-01 -5.24770916e-01
1.41424894e-01 8.20946455e-01 2.52316147e-01 -6.08753324e-01
6.81753099e-01 2.99243420e-01 -1.79602094e-02 -1.99612737e-01
-1.25166619e+00 2.60650665e-01 -4.39714611e-01 1.71761855e-01
6.87711716e-01 -7.37061918e-01 -9.50896263e-01 3.98506939e-01
-1.24513197e+00 -1.93423018e-01 -1.44829556e-01 2.86023498e-01
-3.32648307e-01 -5.22416160e-02 -5.89375138e-01 -9.51716006e-01
-4.23264563e-01 -1.15363824e+00 8.39552462e-01 4.59776223e-01
-3.36811870e-01 -1.04551744e+00 -7.12885439e-01 6.81506753e-01
3.59840423e-01 2.04574645e-01 1.41430223e+00 -7.41691113e-01
-8.34654629e-01 -1.20199561e-01 -3.98207754e-01 3.56331497e-01
1.36444747e-01 -1.70493841e-01 -1.06714821e+00 1.39538914e-01
2.32460592e-02 -2.23518103e-01 1.02766931e+00 3.55621278e-01
1.81701338e+00 -1.04091942e+00 -6.17421754e-02 7.16453612e-01
9.05751646e-01 4.09320742e-01 4.57713872e-01 4.08763379e-01
8.43974948e-01 3.64326149e-01 3.86836827e-01 5.06996214e-01
6.43039703e-01 4.18632567e-01 6.87202632e-01 -3.42655301e-01
4.50175107e-01 -4.31003273e-01 3.73471737e-01 7.84999728e-01
-4.64782417e-01 -2.86356747e-01 -9.67167854e-01 2.15120509e-01
-2.32118988e+00 -6.63572013e-01 -1.09280840e-01 1.77714634e+00
7.22092628e-01 3.84441495e-01 -1.91178858e-01 1.80918798e-01
4.12975341e-01 2.20985059e-02 -1.03659606e+00 -6.04155362e-01
-1.35657579e-01 -4.86066565e-02 1.54580221e-01 2.04435810e-01
-1.00030208e+00 7.50959218e-01 5.96977901e+00 9.70878243e-01
-8.23227048e-01 6.81523904e-02 1.30097592e+00 -7.55891576e-02
-8.43422949e-01 -9.77752209e-02 -6.16196632e-01 6.28571749e-01
6.81608379e-01 -2.66151726e-01 4.26312268e-01 1.14407003e+00
1.51062712e-01 2.38402113e-01 -1.47530222e+00 9.54056561e-01
-2.16601282e-01 -1.35003138e+00 4.73739505e-01 1.80956438e-01
7.35146821e-01 -4.85817820e-01 4.07498866e-01 4.79983419e-01
3.25156003e-01 -1.71842539e+00 6.46688879e-01 4.20673579e-01
7.74637401e-01 -7.17681348e-01 1.00643623e+00 6.20276988e-01
-9.44139898e-01 -6.88203394e-01 -6.28775418e-01 -8.19434404e-01
-3.14040393e-01 6.20868623e-01 -6.67031586e-01 4.63790089e-01
5.51198661e-01 6.38665199e-01 -5.76960027e-01 6.28225148e-01
-4.30195302e-01 7.22552299e-01 -2.78893620e-01 -3.86976004e-01
4.94036645e-01 -1.63822460e-06 2.13340327e-01 6.39318109e-01
3.16058367e-01 2.85800368e-01 -6.16563149e-02 1.36496508e+00
-3.47342074e-01 -2.42354706e-01 -5.18877447e-01 -1.57769486e-01
3.44128877e-01 1.07802391e+00 -4.38359767e-01 -2.46147156e-01
-1.93060532e-01 7.22047150e-01 4.79130238e-01 2.46750280e-01
-9.23656523e-01 -1.58180550e-01 7.10689545e-01 1.10845327e-01
-1.64326295e-01 3.90766151e-02 -9.55052257e-01 -1.35357964e+00
1.67274222e-01 -1.03728986e+00 3.46705288e-01 -8.30125034e-01
-1.43358994e+00 5.15023530e-01 -3.95029821e-02 -1.12763309e+00
-1.41824901e-01 -9.27547097e-01 -6.95591629e-01 6.78885043e-01
-1.23841715e+00 -1.30895770e+00 -4.53052223e-01 2.16309682e-01
7.35924661e-01 -3.00501466e-01 7.70759225e-01 -2.98402831e-02
-7.64008462e-01 7.73874283e-01 1.49669498e-01 2.72057861e-01
7.47347921e-02 -1.28057301e+00 5.21123827e-01 4.10500556e-01
9.43129435e-02 7.74637401e-01 2.87242383e-01 -2.69322127e-01
-1.40866685e+00 -8.98658752e-01 6.56912446e-01 -5.03150940e-01
4.47258234e-01 -3.41045707e-01 -9.65967774e-01 9.40648079e-01
-1.15068726e-01 -5.72581813e-02 8.44051361e-01 4.38028157e-01
-2.30984151e-01 -2.92578548e-01 -9.81122673e-01 8.53253365e-01
1.15442395e+00 -3.43405128e-01 -7.22675920e-01 1.10963829e-01
1.05625188e+00 -2.52285123e-01 -6.26725316e-01 6.64673328e-01
6.95020616e-01 -8.86850297e-01 7.43128538e-01 -9.14496422e-01
1.25477302e+00 -6.66628703e-02 1.10316239e-02 -1.34080255e+00
-6.75499678e-01 -4.99432236e-01 -5.04740417e-01 8.84430826e-01
6.52247846e-01 -9.33826625e-01 7.03385234e-01 1.21099985e+00
-1.19428083e-01 -1.29688776e+00 -7.58414209e-01 -5.99441051e-01
-4.61497996e-03 -3.57977808e-01 9.76429343e-01 8.37153733e-01
2.52997018e-02 5.03247559e-01 -5.65753579e-01 -1.50409847e-01
4.99919325e-01 1.01253696e-01 5.45630038e-01 -1.44664371e+00
-5.39787591e-01 -7.10002780e-01 -3.87858421e-01 -1.23411751e+00
2.15273127e-01 -8.69830310e-01 -3.38912785e-01 -1.42936492e+00
5.47654152e-01 -3.99066120e-01 -5.72764516e-01 1.01016426e+00
-6.92866027e-01 -8.97703022e-02 2.88363487e-01 3.96273464e-01
-5.26184678e-01 7.25742280e-01 1.54229414e+00 -2.99733430e-01
6.07302561e-02 1.10178836e-01 -1.26956177e+00 1.11676359e+00
7.37221718e-01 -3.48883957e-01 -5.72278380e-01 -9.17363167e-01
4.94139701e-01 -1.12055190e-01 5.23686826e-01 -7.35231340e-01
1.44677371e-01 -5.12571156e-01 6.81115150e-01 -2.36084372e-01
3.26078862e-01 -8.39147806e-01 8.71166587e-02 4.42097485e-01
-6.70310140e-01 2.11091995e-01 7.49581084e-02 5.19590795e-01
-2.76506722e-01 -1.93735823e-01 4.16631222e-01 -2.00180367e-01
-5.02326727e-01 3.94940853e-01 -7.91567788e-02 -1.82217568e-01
8.95211995e-01 -2.19249696e-01 -5.20973682e-01 -5.98176718e-01
-3.00096750e-01 3.62862617e-01 7.92910084e-02 4.93643433e-01
7.44141579e-01 -1.44679236e+00 -5.87716877e-01 2.89806843e-01
-7.29649365e-02 5.81126928e-01 1.73765421e-01 5.10014176e-01
-3.93026620e-01 5.22455454e-01 -2.17765927e-01 -3.93200397e-01
-9.58508074e-01 3.26570511e-01 2.74929523e-01 -4.51569080e-01
-3.15667987e-01 8.40580821e-01 7.39528298e-01 -5.59241652e-01
-4.43223156e-02 -7.07649827e-01 5.88024855e-02 -3.01292777e-01
5.81304848e-01 3.27618480e-01 -3.51615369e-01 3.50015536e-02
-1.48143500e-01 1.24935701e-01 -1.89621761e-01 2.95278966e-01
1.51129413e+00 1.83633491e-01 -1.10137686e-01 6.01717412e-01
9.00237560e-01 -6.31965995e-01 -1.34204638e+00 -3.33687246e-01
5.35004921e-02 -5.44070959e-01 -9.11059137e-03 -8.43423784e-01
-8.92744839e-01 1.14482319e+00 1.21758737e-01 2.99744368e-01
1.18489087e+00 -2.15278134e-01 7.49054968e-01 7.88027287e-01
2.28977114e-01 -8.82700145e-01 2.09223434e-01 4.85583514e-01
8.93879294e-01 -1.56956851e+00 -4.44564484e-02 -2.76017666e-01
-7.65854120e-01 1.15981984e+00 1.01860893e+00 7.85658881e-02
3.94068182e-01 1.24231167e-01 -1.85962468e-01 -6.30046055e-02
-1.10845041e+00 1.71337843e-01 6.49193168e-01 1.13294683e-01
2.81022966e-01 4.93057936e-01 -5.58476076e-02 1.43309546e+00
-3.90299648e-01 5.22820093e-02 2.57254124e-01 4.53338981e-01
-3.52006793e-01 -7.05827534e-01 -1.88859418e-01 8.59426260e-01
-6.31166160e-01 -4.25738573e-01 -5.61669648e-01 6.50374115e-01
7.34733127e-04 8.96763563e-01 1.15954615e-01 -7.69407928e-01
7.87662156e-03 -3.87373306e-02 -2.39105765e-02 -3.69267911e-01
-5.33635616e-01 -3.29090446e-01 -3.47883701e-02 -2.64000595e-01
-7.68533535e-03 -1.43362418e-01 -1.33145642e+00 -7.07922041e-01
-2.07676202e-01 -1.57767817e-01 2.94837773e-01 1.25575650e+00
5.77608585e-01 6.59271419e-01 6.77332819e-01 -6.40550077e-01
-8.42816412e-01 -9.52994585e-01 -3.93442571e-01 4.80299979e-01
3.76195908e-01 -6.49866700e-01 -4.13768202e-01 -1.48111165e-01] | [8.928467750549316, 5.656005382537842] |
2151da96-0bbf-4b0c-bbf6-acc2312338e2 | a-survey-on-datasets-for-decision-making-of | 2306.16784 | null | https://arxiv.org/abs/2306.16784v1 | https://arxiv.org/pdf/2306.16784v1.pdf | A Survey on Datasets for Decision-making of Autonomous Vehicle | Autonomous vehicles (AV) are expected to reshape future transportation systems, and decision-making is one of the critical modules toward high-level automated driving. To overcome those complicated scenarios that rule-based methods could not cope with well, data-driven decision-making approaches have aroused more and more focus. The datasets to be used in developing data-driven methods dramatically influences the performance of decision-making, hence it is necessary to have a comprehensive insight into the existing datasets. From the aspects of collection sources, driving data can be divided into vehicle, environment, and driver related data. This study compares the state-of-the-art datasets of these three categories and summarizes their features including sensors used, annotation, and driving scenarios. Based on the characteristics of the datasets, this survey also concludes the potential applications of datasets on various aspects of AV decision-making, assisting researchers to find appropriate ones to support their own research. The future trends of AV dataset development are summarized. | ['Jianqiang Wang', 'Shaobing Xu', 'Yining Xing', 'Zeyu Han', 'Yuning Wang'] | 2023-06-29 | null | null | null | null | ['autonomous-vehicles', 'decision-making'] | ['computer-vision', 'reasoning'] | [ 4.45918404e-02 -1.19425036e-01 -7.41916955e-01 -9.09067690e-01
-2.69568950e-01 -5.21248102e-01 7.42607832e-01 3.55459809e-01
-3.65288585e-01 6.75378621e-01 8.89273807e-02 -6.49660051e-01
-3.33837628e-01 -1.08994126e+00 -3.39006573e-01 -5.79386950e-01
4.70357507e-01 3.52095485e-01 4.77165073e-01 -5.87589502e-01
4.01772320e-01 7.35277653e-01 -2.47916436e+00 -2.03132052e-02
8.48640919e-01 9.72824275e-01 3.54973525e-01 3.76270980e-01
-5.14338970e-01 6.14071667e-01 -2.33666807e-01 -1.72182605e-01
1.23264335e-01 -2.87896730e-02 -2.68469900e-01 2.06397455e-02
5.78390062e-02 -6.23721965e-02 -3.49536896e-01 7.72755742e-01
2.98002034e-01 1.93983078e-01 6.00912869e-01 -1.86983836e+00
-1.95573717e-01 3.13878983e-01 5.99861443e-02 -2.18908973e-02
-8.09315667e-02 3.77168268e-01 4.58253264e-01 -7.34648824e-01
6.17859066e-01 9.40783560e-01 2.83381492e-01 5.06380141e-01
-7.56199956e-01 -6.47076368e-01 2.34798849e-01 7.54695117e-01
-1.29684603e+00 -6.66691124e-01 9.61089015e-01 -8.67175460e-01
6.83449388e-01 2.81791985e-01 6.60587311e-01 7.51075149e-01
1.59597754e-01 6.54223800e-01 1.06807816e+00 -2.21096531e-01
3.96813154e-01 5.96079946e-01 5.32916486e-01 2.56296962e-01
8.05756688e-01 2.94446260e-01 -4.05045778e-01 2.75920898e-01
-2.04545513e-01 -1.33947536e-01 3.76936793e-01 -5.23855150e-01
-8.81309450e-01 8.83223653e-01 1.38216823e-01 2.67752290e-01
-4.43316728e-01 -1.19231075e-01 4.33655709e-01 7.24365190e-02
-1.09114649e-03 1.99445322e-01 -3.49907190e-01 -3.79075319e-01
-8.00947011e-01 6.79822445e-01 4.42076266e-01 1.14096737e+00
1.08465469e+00 1.81583494e-01 4.81802449e-02 6.01806283e-01
5.03095984e-01 8.11851025e-01 3.67072709e-02 -9.94098067e-01
3.83992851e-01 9.95263338e-01 1.46978498e-01 -1.02783728e+00
-4.81470883e-01 1.18023306e-01 -3.93045962e-01 2.00824544e-01
1.35983884e-01 -2.15309575e-01 -9.09013510e-01 1.15273035e+00
2.27541521e-01 -4.44082171e-01 2.18374372e-01 7.67714679e-01
1.15861630e+00 5.80350578e-01 1.83464438e-01 -9.05786902e-02
1.15488231e+00 -4.55385178e-01 -1.19058073e+00 -3.55312437e-01
6.86644912e-01 -4.70074624e-01 6.66931510e-01 1.50797784e-01
-5.95234156e-01 -8.51705551e-01 -1.20042872e+00 7.68240914e-02
-1.02387774e+00 -1.19108334e-01 5.24811149e-01 7.16329634e-01
-6.10572338e-01 -5.04685976e-02 -4.74171460e-01 -5.18687069e-01
3.95133585e-01 6.16569817e-02 -1.97491869e-01 -1.81272656e-01
-1.32588589e+00 1.29560304e+00 3.71461630e-01 1.43968493e-01
-9.29311395e-01 -5.04002273e-01 -8.29009414e-01 -4.95842069e-01
4.92134243e-01 -2.58403391e-01 1.12503409e+00 -1.21467680e-01
-9.35777307e-01 7.19420791e-01 -4.11499262e-01 -5.21433413e-01
4.93424147e-01 3.69407743e-01 -1.09785008e+00 -3.53326917e-01
3.46467882e-01 7.03768611e-01 4.25694108e-01 -1.36241972e+00
-1.26240504e+00 -3.22019815e-01 -2.09492639e-01 -7.27612898e-02
-1.01248436e-01 2.23745052e-02 -3.80500406e-01 5.20780534e-02
-3.15507531e-01 -1.07013357e+00 -3.53215575e-01 -1.12781912e-01
-1.38344318e-01 -6.25477731e-01 1.26112401e+00 -1.55397445e-01
1.32066298e+00 -2.15243983e+00 -4.10890430e-01 8.32218379e-02
1.93443149e-01 5.73663116e-01 1.89605087e-01 7.10438073e-01
4.06395465e-01 8.44639391e-02 -1.06924064e-01 -2.97588985e-02
1.51105165e-01 6.70471132e-01 -1.28997028e-01 2.37457216e-01
2.44172290e-01 8.04636478e-01 -6.83231890e-01 -5.02908230e-01
8.78326952e-01 2.77994543e-01 1.28333926e-01 1.78739354e-01
-2.73100257e-01 4.98196632e-01 -7.91752696e-01 7.24036872e-01
6.50926530e-01 6.05913103e-01 -1.65963396e-01 -1.12378091e-01
-9.41357970e-01 2.17780411e-01 -1.01380467e+00 1.18250263e+00
-2.58851886e-01 1.20923603e+00 -1.28525838e-01 -1.03506792e+00
1.38499928e+00 1.17375106e-01 6.09121144e-01 -9.45211649e-01
3.33762854e-01 2.89099514e-01 3.16399872e-01 -1.09488201e+00
8.50581527e-01 2.57734597e-01 -3.63364249e-01 -7.88891986e-02
-5.48915148e-01 -3.30828160e-01 5.98140478e-01 -1.17571525e-01
6.16516471e-01 -4.21418902e-03 6.08343333e-02 -2.05773503e-01
6.86911166e-01 8.69641423e-01 7.93315709e-01 3.77047271e-01
-6.88284934e-01 8.83963481e-02 2.14936912e-01 -7.03607023e-01
-9.13551807e-01 -6.44062221e-01 -5.15033782e-01 7.54544139e-01
3.87204677e-01 -2.18888536e-01 -5.14182329e-01 -3.60767215e-01
4.51620400e-01 1.23654449e+00 -5.46314836e-01 -1.46668077e-01
-2.38644779e-01 -4.10465628e-01 4.18913662e-01 5.51377237e-01
7.28612065e-01 -7.94210136e-01 -9.66438591e-01 1.48566112e-01
-1.94643244e-01 -1.16586161e+00 5.07722795e-01 -4.50434126e-02
-5.05590737e-01 -1.08562505e+00 -3.75001132e-02 -4.07278061e-01
4.15323615e-01 7.63069451e-01 7.57126391e-01 -1.34247348e-01
-1.95944193e-03 3.73880775e-03 -4.51840669e-01 -1.26725864e+00
-5.80605447e-01 1.59915507e-01 1.47700399e-01 -5.97263053e-02
1.11285841e+00 3.45224291e-02 -3.85243922e-01 7.87811458e-01
-7.11044610e-01 8.71408731e-03 4.97207940e-01 2.29178183e-02
6.10063791e-01 4.74979490e-01 7.80963004e-01 -5.06427526e-01
5.16891301e-01 -5.95837235e-01 -7.53627598e-01 3.60248201e-02
-9.63751376e-01 -7.18638897e-02 1.54541418e-01 1.13459513e-01
-1.06355250e+00 1.98055342e-01 -1.35319844e-01 9.54988878e-03
-7.78673708e-01 4.95973676e-01 -5.26713133e-01 2.98372321e-02
4.42845374e-01 4.86514233e-02 2.05015793e-01 -3.11650187e-01
3.76152396e-01 1.22863376e+00 4.01365638e-01 -1.18021145e-01
5.70165277e-01 4.02172446e-01 1.61562502e-01 -1.00456893e+00
-2.93881387e-01 -5.90572059e-01 -6.91409886e-01 -8.36931407e-01
9.29009855e-01 -6.55559003e-01 -4.52333331e-01 3.90383869e-01
-7.52395630e-01 -6.70339018e-02 -3.30416948e-01 5.22492886e-01
-4.23430473e-01 -2.41284981e-01 4.58072305e-01 -1.07094693e+00
3.00491214e-01 -1.58667147e+00 5.86906135e-01 2.87634343e-01
-2.29167521e-01 -6.51808321e-01 -7.22058341e-02 7.20695794e-01
6.22340083e-01 2.99193740e-01 7.02750146e-01 -5.11246324e-01
-4.73546833e-01 -4.82640356e-01 -8.48508105e-02 1.81528017e-01
1.37649655e-01 3.86847466e-01 -1.08460796e+00 2.44190812e-01
-3.94751191e-01 1.24027401e-01 6.77266717e-01 3.35893661e-01
9.97888803e-01 1.77729562e-01 -6.69590712e-01 5.07946499e-02
1.26119530e+00 6.14850402e-01 5.89491904e-01 6.55667901e-01
5.06137013e-01 1.24279189e+00 1.43740463e+00 2.69663781e-01
1.07505727e+00 6.42382860e-01 7.20151067e-01 1.44242980e-02
-2.10893720e-01 -3.28796297e-01 4.31768671e-02 4.03386980e-01
6.70850351e-02 -3.17551464e-01 -1.23491573e+00 8.48297656e-01
-2.02271223e+00 -9.42841053e-01 -8.13305080e-01 1.91331363e+00
1.58568934e-01 1.10049210e-01 3.82868201e-01 5.32444119e-01
5.13709009e-01 1.72010064e-01 -8.74332726e-01 -5.88923514e-01
-1.15317650e-01 -7.45693684e-01 8.92969131e-01 1.20168149e-01
-1.02092981e+00 8.41959178e-01 6.32729006e+00 3.76901388e-01
-1.17723012e+00 -1.54392987e-01 2.22271651e-01 2.66004324e-01
-3.99011642e-01 -1.14312895e-01 -1.21677315e+00 6.75838649e-01
1.28750110e+00 -3.46087962e-01 6.98518232e-02 8.86411667e-01
9.25217867e-01 -5.24290025e-01 -8.47481072e-01 8.53432298e-01
-1.83475494e-01 -1.41389024e+00 -1.69101924e-01 1.96640834e-01
7.69844055e-01 3.20920974e-01 -6.96235672e-02 2.13109583e-01
2.83771902e-01 -7.05101311e-01 7.02951372e-01 7.02498019e-01
6.43172979e-01 -7.58862615e-01 1.01431227e+00 4.23584163e-01
-1.20885372e+00 -2.73368269e-01 -3.75033110e-01 -2.58597016e-01
3.79734069e-01 5.69684982e-01 -8.31212461e-01 7.02240050e-01
7.44332135e-01 7.02757597e-01 -5.15082538e-01 9.56021190e-01
-4.46991250e-02 6.10072255e-01 -7.42513761e-02 -2.89859176e-01
1.88978493e-01 -1.98377892e-01 3.92828584e-01 1.00386882e+00
1.93978220e-01 -8.07341039e-02 9.96274352e-02 5.20981669e-01
4.59270179e-01 -5.42227477e-02 -1.16036701e+00 -1.27506644e-01
7.98600316e-01 1.16182745e+00 -3.82827848e-01 -2.90103644e-01
-7.87135005e-01 -1.83433801e-01 -2.42160708e-01 1.05275385e-01
-7.02017546e-01 -4.67883617e-01 1.36235857e+00 5.78441322e-01
4.13433164e-02 -5.37695408e-01 -7.16955006e-01 -4.14460272e-01
-9.25769731e-02 -5.02911508e-01 1.02448963e-01 -7.51434505e-01
-7.61509538e-01 4.50223237e-01 5.10359883e-01 -1.47276759e+00
-4.12197709e-01 -5.87054431e-01 -2.55724788e-01 6.57657564e-01
-1.79869449e+00 -9.48032558e-01 -7.15415061e-01 2.74427146e-01
6.48314536e-01 -4.40233350e-01 5.15272677e-01 3.75462770e-01
-7.28523850e-01 1.15678035e-01 -2.96389479e-02 1.22654820e-02
5.35692573e-01 -5.93888640e-01 3.30635607e-01 8.93192649e-01
-4.38690633e-01 5.59067726e-02 8.48460972e-01 -6.31524026e-01
-1.67590106e+00 -1.33646882e+00 1.04667866e+00 -5.15025258e-01
5.69902241e-01 -1.50252283e-01 -6.74342811e-01 4.10824537e-01
2.78689153e-02 -2.47643054e-01 5.93217313e-01 -2.34648556e-01
1.73794940e-01 -6.55529022e-01 -1.18516684e+00 4.13198382e-01
9.22537327e-01 -1.97773129e-01 -4.83556896e-01 -3.87559496e-02
2.65950382e-01 -2.74011474e-02 -5.63998520e-01 6.30776584e-01
5.03489971e-01 -7.98725665e-01 4.46409374e-01 -3.72399718e-01
-8.79588202e-02 -7.72266209e-01 -3.03362012e-01 -1.24398518e+00
-2.60495901e-01 -1.77450672e-01 2.66334891e-01 1.31639528e+00
5.47864079e-01 -8.51674199e-01 6.76823139e-01 1.04180622e+00
-4.72258925e-01 -4.43505108e-01 -6.13197803e-01 -6.38718307e-01
-1.02883339e-01 -1.12233353e+00 1.23775089e+00 6.43648624e-01
-2.10696772e-01 8.23846925e-03 -7.25158155e-02 1.33816347e-01
4.66374516e-01 2.77226213e-02 1.19291568e+00 -1.61645281e+00
9.35790479e-01 -5.03789485e-01 -9.29661870e-01 -5.66966474e-01
-3.57473828e-02 -5.78637540e-01 2.78911352e-01 -1.99710262e+00
-2.59987742e-01 -7.67928064e-01 -5.19464687e-02 4.47427660e-01
1.80133060e-01 2.98549309e-02 5.29178381e-02 1.17979281e-01
-2.00619996e-01 3.41266990e-01 1.00354123e+00 -2.81781256e-01
-1.05950937e-01 1.73500106e-01 -8.61243963e-01 6.35841012e-01
1.17726111e+00 -3.02107245e-01 -6.69109941e-01 -3.72245938e-01
1.54033273e-01 -2.67172933e-01 -3.46376076e-02 -1.06318629e+00
4.33532178e-01 -6.70197845e-01 -1.64029554e-01 -1.25541234e+00
2.07428128e-01 -1.19535494e+00 3.50653291e-01 2.92411685e-01
-2.75373280e-01 1.73190594e-01 2.71083027e-01 2.96679884e-01
-4.14790869e-01 -1.36976048e-01 4.57158387e-01 3.21592867e-01
-1.58143163e+00 2.91571736e-01 -1.05509686e+00 -3.48582923e-01
1.57877135e+00 -6.10075176e-01 -2.93615788e-01 -1.96533129e-01
-3.39149565e-01 7.13298321e-01 4.73260224e-01 9.28084493e-01
4.85037029e-01 -1.22597277e+00 -7.70902455e-01 4.68652874e-01
8.10416102e-01 2.92777494e-02 2.37038225e-01 5.86314976e-01
-3.69340450e-01 8.89512539e-01 -4.06115353e-01 -7.14645982e-01
-1.16972601e+00 4.96335208e-01 1.09659597e-01 4.50581729e-01
-1.51127145e-01 9.12541002e-02 -2.35145375e-01 -5.03147721e-01
1.11595020e-01 -1.12370253e-01 -6.43543541e-01 3.72944683e-01
4.96422261e-01 8.23520064e-01 3.86680007e-01 -1.07874417e+00
-5.35996199e-01 5.01742244e-01 3.46870869e-01 -1.33121130e-03
1.13625777e+00 -5.42317688e-01 3.28232050e-01 7.14324474e-01
7.12776005e-01 -3.24680179e-01 -1.12090397e+00 -2.66058445e-02
3.80958505e-02 -5.68875790e-01 4.42613721e-01 -7.32121110e-01
-9.27090526e-01 9.27293241e-01 6.30555868e-01 4.22846228e-01
8.43472004e-01 -1.05377045e-02 5.20726442e-01 3.04960847e-01
5.76819956e-01 -1.45525789e+00 -8.22648644e-01 2.44564548e-01
7.31564939e-01 -1.66773403e+00 -1.34490237e-01 -5.23402154e-01
-7.82602549e-01 8.87837648e-01 5.47450185e-01 3.83393854e-01
7.60208428e-01 3.62634569e-01 3.87254328e-01 -1.40615761e-01
-6.57780170e-01 -7.36660779e-01 2.50245452e-01 1.06873596e+00
1.23439588e-01 4.95824426e-01 -5.47281146e-01 2.93313384e-01
-2.67429113e-01 2.56262869e-01 4.75457609e-01 9.10592496e-01
-7.82869935e-01 -1.22060871e+00 -3.32527608e-01 6.60419762e-01
2.32816085e-01 6.44084752e-01 -4.13643658e-01 1.01411283e+00
4.99546140e-01 1.71645975e+00 1.43036395e-01 -7.83094049e-01
7.75429785e-01 -1.00718081e-01 -2.06804693e-01 -2.24883169e-01
3.90661769e-02 -6.35186672e-01 5.81901312e-01 -4.38145280e-01
-6.00318253e-01 -1.05419981e+00 -1.23945773e+00 -6.92017794e-01
-5.35947308e-02 1.47947267e-01 1.43335903e+00 9.03717935e-01
6.64944470e-01 5.41788697e-01 7.09127963e-01 -6.24388039e-01
-1.29684983e-02 -6.59922004e-01 -2.83239186e-01 6.14066124e-02
2.04210401e-01 -1.00085115e+00 1.82434730e-02 -1.79862589e-01] | [5.700277328491211, 1.1486294269561768] |
5786df5e-cd41-4302-b77a-23b7e9638cf3 | learning-stationary-markov-processes-with | 2303.05497 | null | https://arxiv.org/abs/2303.05497v2 | https://arxiv.org/pdf/2303.05497v2.pdf | Learning Stationary Markov Processes with Contrastive Adjustment | We introduce a new optimization algorithm, termed contrastive adjustment, for learning Markov transition kernels whose stationary distribution matches the data distribution. Contrastive adjustment is not restricted to a particular family of transition distributions and can be used to model data in both continuous and discrete state spaces. Inspired by recent work on noise-annealed sampling, we propose a particular transition operator, the noise kernel, that can trade mixing speed for sample fidelity. We show that contrastive adjustment is highly valuable in human-computer design processes, as the stationarity of the learned Markov chain enables local exploration of the data manifold and makes it possible to iteratively refine outputs by human feedback. We compare the performance of noise kernels trained with contrastive adjustment to current state-of-the-art generative models and demonstrate promising results on a variety of image synthesis tasks. | ['Joakim Lundeberg', 'Jens Lagergren', 'Ludvig Bergenstråhle'] | 2023-03-09 | null | null | null | null | ['image-inpainting'] | ['computer-vision'] | [ 3.26506138e-01 4.12058644e-02 -2.92292327e-01 -3.83392982e-02
-7.59744406e-01 -4.28302348e-01 1.08158219e+00 -2.91545063e-01
-7.12996125e-02 4.90963817e-01 3.26847285e-01 -2.60733455e-01
-8.76469314e-02 -6.59016609e-01 -7.22214460e-01 -7.08520055e-01
2.34369785e-01 7.51905322e-01 2.05690265e-01 3.41330953e-02
2.23836023e-02 6.97109044e-01 -1.31677866e+00 -2.69897617e-02
8.82989824e-01 5.40812910e-01 3.15925330e-01 9.12215233e-01
5.39713912e-02 6.17604852e-01 -2.29089126e-01 -1.54991552e-01
4.61370535e-02 -1.02854323e+00 -6.98226631e-01 2.90225834e-01
1.34253263e-01 -4.20372486e-02 -3.94294262e-01 9.76338267e-01
2.43122473e-01 5.06439567e-01 1.22034049e+00 -1.16255283e+00
-8.11034501e-01 6.39873564e-01 -8.31849501e-02 1.80246636e-01
9.01967734e-02 3.39423448e-01 8.59828472e-01 -1.01042974e+00
7.04254866e-01 1.29366028e+00 6.77538514e-01 7.79115140e-01
-2.08522129e+00 -5.10885119e-01 -9.68967676e-02 -1.97923481e-01
-1.13143587e+00 -6.60002768e-01 7.90137291e-01 -6.13573790e-01
7.39007056e-01 7.84041211e-02 9.28962409e-01 1.24825621e+00
6.17090940e-01 9.33100164e-01 1.25599921e+00 -5.56690693e-01
5.07058799e-01 -5.89122996e-03 -1.56537801e-01 1.01610422e+00
-1.44841015e-01 4.71636146e-01 -5.95538616e-01 -2.46534690e-01
1.40355718e+00 -1.90324754e-01 -8.38312283e-02 -9.68211651e-01
-1.26140058e+00 7.79759526e-01 1.45010769e-01 8.89301002e-02
-3.84631783e-01 7.27546155e-01 -5.90578653e-02 2.97661185e-01
3.90820205e-01 5.85959196e-01 2.34015957e-02 -3.91818851e-01
-1.07852054e+00 2.43079856e-01 7.55382836e-01 1.03618324e+00
8.51479769e-01 4.02101874e-01 -4.19100314e-01 6.98368788e-01
4.29526716e-01 6.55834377e-01 2.49569267e-01 -1.21214521e+00
-1.49732843e-01 -2.28190217e-02 2.70514429e-01 -3.05819362e-01
-1.70928493e-01 -5.15321732e-01 -1.04626334e+00 4.33653682e-01
2.35989600e-01 -6.80257380e-02 -1.20796800e+00 1.88340616e+00
2.68084496e-01 3.30102116e-01 -9.46407020e-02 4.70908105e-01
1.22214735e-01 7.11715281e-01 -1.22430036e-02 -2.70011425e-01
7.75652468e-01 -9.10340548e-01 -7.15464592e-01 -3.95702422e-02
-5.91538958e-02 -7.80989647e-01 1.25631094e+00 3.83272976e-01
-1.39599752e+00 -6.99080706e-01 -8.23683858e-01 3.66325319e-01
3.40459019e-01 -2.61916332e-02 6.39906049e-01 7.23544359e-01
-1.21854270e+00 1.06909835e+00 -1.48061609e+00 -2.30242848e-01
3.46926868e-01 1.79383606e-01 8.51303488e-02 1.89570710e-01
-7.51000166e-01 8.95331621e-01 -3.13116275e-02 -2.71410346e-01
-1.47308743e+00 -7.54450381e-01 -7.07181990e-01 -9.87596661e-02
-4.84440988e-03 -1.13510072e+00 1.57982063e+00 -7.45360732e-01
-2.20242786e+00 4.71712053e-01 -1.80591792e-01 -3.79724234e-01
5.41230321e-01 -2.32539296e-01 -2.63779432e-01 -2.26645023e-02
-1.38297856e-01 6.75060272e-01 1.58987224e+00 -1.28158939e+00
-2.30478123e-01 3.12510431e-01 -5.51171482e-01 3.35199572e-02
-9.28872004e-02 -2.80069947e-01 -2.46993363e-01 -8.40854108e-01
1.22791596e-01 -1.22746849e+00 -6.86010122e-01 -5.24289999e-03
-3.33132714e-01 -1.67307660e-01 6.03131473e-01 -2.34635800e-01
1.17290533e+00 -1.84298706e+00 6.09980524e-01 3.67084652e-01
-1.03575960e-01 -8.68180171e-02 5.58457430e-03 5.94460368e-01
2.66343243e-02 -4.83755358e-02 -4.63658273e-01 -5.57912886e-01
1.67042285e-01 2.73321718e-01 -4.45583254e-01 4.29734200e-01
4.00238127e-01 1.02066398e+00 -1.02912700e+00 -3.17102551e-01
3.68508458e-01 4.28605616e-01 -6.56940997e-01 3.71392727e-01
-2.82475114e-01 6.97285891e-01 -2.60411978e-01 1.91332415e-01
2.13319644e-01 -3.80000710e-01 -1.71194106e-01 -2.48126462e-02
1.63258418e-01 2.19163939e-01 -1.18695652e+00 1.86715639e+00
-5.93302369e-01 6.63823903e-01 -1.58268481e-01 -5.45357704e-01
8.09823036e-01 3.29130977e-01 2.75813162e-01 -4.07711864e-02
-6.89940602e-02 4.53812331e-02 -2.34378040e-01 -1.91196889e-01
4.83356714e-01 -5.40420711e-01 2.36770641e-02 7.29822040e-01
3.99203509e-01 -9.39199209e-01 6.93272501e-02 1.42851070e-01
1.01738417e+00 5.22978246e-01 1.89146966e-01 -6.97578311e-01
-1.26755729e-01 -2.36236244e-01 3.10322702e-01 9.41101372e-01
5.99075407e-02 8.29504967e-01 2.26595387e-01 -2.27282017e-01
-1.45624840e+00 -1.77440012e+00 -1.88523293e-01 8.45954716e-01
-6.98443726e-02 -3.29957545e-01 -8.78984392e-01 -4.08673555e-01
-2.83843011e-01 8.91260862e-01 -6.07596517e-01 -4.03092533e-01
-4.70158935e-01 -4.99330103e-01 3.49573284e-01 6.79201007e-01
3.20987940e-01 -1.09283113e+00 -4.47108477e-01 5.35167992e-01
7.35600516e-02 -7.41047084e-01 -9.48420823e-01 3.65160674e-01
-1.18576729e+00 -6.72908664e-01 -6.16830945e-01 -7.18542993e-01
5.35273314e-01 -1.74575254e-01 1.37260532e+00 -4.02131349e-01
-6.59067035e-02 7.61926770e-01 1.54877407e-02 -1.52313903e-01
-1.24999440e+00 9.90642682e-02 1.16333172e-01 -9.72456336e-02
-2.50327170e-01 -9.04743016e-01 -5.11900246e-01 2.32994542e-01
-9.33920026e-01 3.39945108e-01 5.15018046e-01 9.58909869e-01
5.59838951e-01 -6.52551278e-02 2.99815774e-01 -6.89924777e-01
7.58478463e-01 -1.44675747e-01 -6.09793663e-01 1.40417740e-02
-7.63967872e-01 7.91883171e-01 5.06795466e-01 -7.83388674e-01
-1.18663931e+00 3.01722169e-01 -1.40044093e-01 -5.77548623e-01
6.79542217e-03 1.41484842e-01 1.31826609e-01 1.40444875e-01
8.36499095e-01 3.53300005e-01 2.54605770e-01 -2.83513218e-01
7.19920158e-01 1.34234607e-01 6.75182581e-01 -7.30177462e-01
9.08262670e-01 4.79243755e-01 2.65649974e-01 -8.73340607e-01
-4.09540445e-01 -3.58721465e-02 -4.28406686e-01 -2.48909280e-01
8.13456833e-01 -6.99457705e-01 -4.17601287e-01 5.56257665e-01
-8.82503808e-01 -7.88034737e-01 -1.01213014e+00 5.43802679e-01
-1.28883588e+00 1.70871913e-01 -9.93218601e-01 -7.69862950e-01
-1.59945279e-01 -1.29216039e+00 1.07719517e+00 1.27246693e-01
-5.15756905e-01 -1.22955000e+00 4.95755553e-01 -3.79593760e-01
3.49126130e-01 5.87260723e-03 9.68075991e-01 1.41588766e-02
-8.53542805e-01 9.07155275e-02 4.79147226e-01 4.70536023e-01
2.76732236e-01 4.33177024e-01 -7.45358407e-01 -3.77600074e-01
-8.44396055e-02 -1.95491374e-01 1.08348823e+00 7.10613549e-01
7.67414093e-01 -9.92417932e-02 -4.72459555e-01 4.66053158e-01
1.11022449e+00 -1.00883916e-01 6.86021447e-01 -1.09165549e-01
5.70404708e-01 3.31258960e-02 1.83070034e-01 5.07095754e-01
-3.84238511e-02 5.02082348e-01 1.03699915e-01 -2.01033037e-02
-3.63854200e-01 -6.74666226e-01 4.40627247e-01 8.95724297e-01
-1.22360185e-01 5.72150387e-02 -6.65191472e-01 5.71352065e-01
-1.85575426e+00 -9.50646460e-01 2.54360557e-01 2.10243034e+00
1.13778663e+00 3.41604978e-01 2.19248608e-01 -1.11930721e-01
7.28902996e-01 6.35301098e-02 -4.72890049e-01 -3.77714574e-01
1.58404648e-01 7.60953307e-01 3.70959908e-01 7.21484721e-01
-9.83642697e-01 9.66368318e-01 8.31257057e+00 1.12438595e+00
-6.76290631e-01 9.76630151e-02 4.42659944e-01 1.40379056e-01
-7.92078555e-01 1.96527198e-01 -7.07329512e-01 2.99710423e-01
1.09901893e+00 6.61930144e-02 5.74228168e-01 8.54154825e-01
1.51701003e-01 5.08365855e-02 -1.34870589e+00 7.87523985e-01
-3.67465317e-01 -1.76458549e+00 1.02430694e-01 -1.01143606e-02
1.19799185e+00 -3.09541613e-01 4.62357968e-01 3.56171578e-02
1.09494519e+00 -1.04502141e+00 9.43219423e-01 9.52958107e-01
6.57419741e-01 -9.67932343e-01 -5.26480339e-02 3.27750921e-01
-1.15778303e+00 1.70192897e-01 -2.25942269e-01 2.94943124e-01
3.99164796e-01 5.97627938e-01 -7.97132194e-01 -2.84780413e-02
3.86791140e-01 7.04093158e-01 -2.03398779e-01 8.57450247e-01
-4.66066271e-01 8.39082122e-01 -2.47758418e-01 -3.99672948e-02
1.20512471e-01 -4.89558250e-01 8.10300231e-01 1.13409781e+00
4.30001557e-01 -1.77852601e-01 6.59787208e-02 1.23173082e+00
6.27147779e-03 -3.58896494e-01 -5.74747026e-01 -3.35006192e-02
3.98640186e-01 9.58480358e-01 -8.37144315e-01 -3.74380350e-01
6.20640889e-02 1.25514996e+00 3.36820632e-01 5.26307940e-01
-7.53234327e-01 -4.39520366e-02 7.30679095e-01 1.09886959e-01
5.91615260e-01 -6.89337552e-01 -1.79739088e-01 -1.05308855e+00
-5.28802872e-01 -8.04385424e-01 -4.19786535e-02 -6.42473817e-01
-1.14594471e+00 5.79061091e-01 8.23087916e-02 -1.19499838e+00
-6.39304161e-01 -1.58791035e-01 -4.93345320e-01 7.74049640e-01
-8.87683749e-01 -1.09094810e+00 7.37186000e-02 4.82508183e-01
6.39951527e-01 -1.38564050e-01 8.66456032e-01 -3.51513892e-01
-2.58232266e-01 3.99828643e-01 4.20599401e-01 -3.83680284e-01
4.83418733e-01 -1.46577990e+00 1.02683127e+00 9.12851453e-01
5.19030392e-01 7.17295647e-01 1.23195446e+00 -7.26200581e-01
-1.26315916e+00 -9.80649710e-01 2.35533416e-01 -6.55975223e-01
5.44035494e-01 -3.12309682e-01 -6.98338211e-01 7.94127584e-01
4.22635525e-01 -1.23630218e-01 4.43875849e-01 -1.95919424e-02
-1.91819310e-01 1.56795502e-01 -9.96661127e-01 8.93759847e-01
1.04186523e+00 -6.28983080e-01 -3.64894509e-01 9.39143002e-02
7.58619189e-01 -4.64792669e-01 -8.67769659e-01 2.13732764e-01
5.03141046e-01 -8.22874010e-01 8.98243427e-01 -7.04316080e-01
3.08326721e-01 -1.88599080e-01 7.06956461e-02 -1.77172756e+00
-8.57149720e-01 -1.26017702e+00 -4.72179323e-01 1.03181219e+00
4.45472032e-01 -3.49324107e-01 8.85024548e-01 5.34671903e-01
-3.27182353e-01 -6.11911058e-01 -8.19901705e-01 -8.80534291e-01
2.31780097e-01 -4.32094932e-01 4.89591956e-01 4.18050528e-01
-2.13224828e-01 4.03372228e-01 -5.41991472e-01 1.92751884e-02
7.93747902e-01 1.11700498e-01 6.67250693e-01 -1.05626154e+00
-7.66607761e-01 -4.62718815e-01 -1.00887045e-01 -1.51626420e+00
1.52156785e-01 -7.92920530e-01 3.39922845e-01 -1.40521276e+00
1.19759142e-01 -4.96209115e-01 -1.20295167e-01 -2.21683271e-03
-2.06854329e-01 1.53136596e-01 3.34924497e-02 3.39738011e-01
-2.03884289e-01 9.88129139e-01 1.47889614e+00 6.61322400e-02
-5.19158900e-01 3.80799443e-01 -2.92446166e-01 5.61391711e-01
4.61720914e-01 -4.22748387e-01 -5.17848313e-01 1.79893970e-02
1.22400694e-01 8.73077884e-02 2.84119576e-01 -1.22883022e+00
1.92692459e-01 -3.89078975e-01 3.75445068e-01 -4.14359391e-01
4.52219367e-01 -6.36868060e-01 6.89841509e-01 6.23186231e-01
-6.50652349e-01 7.88203776e-02 -6.24711113e-03 8.96249056e-01
2.08313391e-01 -1.98718190e-01 1.20631385e+00 -9.07448381e-02
-4.56877738e-01 3.61774385e-01 -1.00888133e+00 1.54662684e-01
6.99535668e-01 -1.52478397e-01 1.59796759e-01 -5.69319189e-01
-9.72572803e-01 -3.16421241e-01 7.78522670e-01 1.70543119e-01
7.20988393e-01 -1.68061769e+00 -5.72338939e-01 4.30592328e-01
-1.47776857e-01 -1.32071674e-01 2.32650014e-03 4.86250728e-01
-4.10290778e-01 -2.71685068e-02 -1.42503276e-01 -9.02479768e-01
-8.86615574e-01 5.33516467e-01 3.45936865e-01 -3.54058892e-01
-7.22100973e-01 9.00459588e-01 -5.59459776e-02 -2.60384619e-01
-8.65834579e-02 -6.73657119e-01 4.54253137e-01 -2.92322844e-01
2.68079996e-01 3.60412359e-01 -1.11932382e-01 -1.39082253e-01
5.34833521e-02 4.58699912e-01 9.92961600e-02 -6.59002006e-01
1.19775331e+00 -7.55403191e-02 2.25469083e-01 8.15959752e-01
8.73647928e-01 -2.49878809e-01 -1.99340034e+00 -4.37896639e-01
-2.93766975e-01 -3.37048829e-01 -1.27391061e-02 -2.60441810e-01
-6.64285958e-01 6.79548144e-01 6.10703051e-01 3.12336177e-01
8.03864241e-01 9.05387253e-02 6.59203887e-01 1.73411667e-01
3.60128671e-01 -1.18083310e+00 7.14268446e-01 4.20149356e-01
7.13689983e-01 -8.95217776e-01 -1.26595795e-01 -1.98429391e-01
-6.77652895e-01 9.23028529e-01 5.56725636e-02 -5.39510906e-01
9.07392561e-01 5.29177904e-01 -2.39281788e-01 1.19363561e-01
-8.02061200e-01 -1.50322735e-01 4.54294622e-01 8.46881032e-01
2.27723047e-01 7.44877905e-02 2.56520391e-01 -2.11133268e-02
-1.90774620e-01 -7.85782039e-02 5.57065129e-01 1.03697729e+00
-7.04914212e-01 -1.22573841e+00 -1.79202408e-01 4.68294322e-01
-1.15885645e-01 -1.56735763e-01 8.66247863e-02 5.59927046e-01
-3.97662610e-01 7.70331681e-01 1.29929334e-01 -1.83811888e-01
1.61108911e-01 1.91974387e-01 9.90076959e-01 -7.09642708e-01
-3.57606798e-01 4.01947200e-01 -8.92926678e-02 -4.43842441e-01
-3.89475554e-01 -9.00261283e-01 -1.04685271e+00 -3.22734714e-01
-4.04806077e-01 2.33733803e-01 4.80287910e-01 8.29935372e-01
2.56222218e-01 6.33149385e-01 5.28096497e-01 -9.72463071e-01
-7.81929672e-01 -8.41655552e-01 -6.34149909e-01 2.55552441e-01
5.29154122e-01 -6.60660267e-01 -1.72187984e-01 3.09687316e-01] | [11.261992454528809, -0.20386356115341187] |
bba7c842-827f-4202-8eaf-b8b6ff757e22 | juncnet-a-deep-neural-network-for-road | 1809.01011 | null | http://arxiv.org/abs/1809.01011v1 | http://arxiv.org/pdf/1809.01011v1.pdf | JuncNet: A Deep Neural Network for Road Junction Disambiguation for Autonomous Vehicles | With a great amount of research going on in the field of autonomous vehicles
or self-driving cars, there has been considerable progress in road detection
and tracking algorithms. Most of these algorithms use GPS to handle road
junctions and its subsequent decisions. However, there are places in the urban
environment where it becomes difficult to get GPS fixes which render the
junction decision handling erroneous or possibly risky. Vision-based junction
detection, however, does not have such problems. This paper proposes a novel
deep convolutional neural network architecture for disambiguation of junctions
from roads with a high degree of accuracy. This network is benchmarked against
other well known classifying network architectures like AlexNet and VGGnet.
Further, we discuss a potential road navigation methodology which uses the
proposed network model. We conclude by performing an experimental validation of
the trained network and the navigational method on the roads of the Indian
Institute of Science (IISc). | ['S. N. Omkar', 'Navaneethkrishnan B', 'Sumedh Mannar', 'Saumya Kumaar'] | 2018-08-31 | null | null | null | null | ['junction-detection'] | ['computer-vision'] | [ 2.25890335e-03 1.44003376e-01 -1.30956411e-01 -4.42314744e-01
-2.91046858e-01 -3.27390492e-01 9.68340516e-01 -5.29867932e-02
-6.62023783e-01 7.82152355e-01 -3.00559700e-01 -1.02651846e+00
-4.61305911e-03 -1.46217597e+00 -6.37944639e-01 -3.99965763e-01
5.61823323e-02 2.50603735e-01 7.16479301e-01 -5.14508069e-01
5.05192518e-01 8.29972327e-01 -1.72488844e+00 -2.43458837e-01
8.07953358e-01 9.83638883e-01 3.29905152e-02 7.40359545e-01
-3.66830021e-01 6.96931422e-01 -3.22271258e-01 -5.21182239e-01
4.14565444e-01 1.36978373e-01 -8.71593714e-01 -5.21502912e-01
8.41945767e-01 -2.34905913e-01 -5.68451405e-01 9.65700269e-01
4.13632989e-01 2.79357910e-01 6.95139408e-01 -1.29405594e+00
-3.59088868e-01 1.52146399e-01 -1.92596436e-01 5.83268285e-01
3.92111279e-02 9.73977819e-02 5.92088223e-01 -5.57870686e-01
4.49702919e-01 1.02063394e+00 1.04939961e+00 1.20370470e-01
-5.02489805e-01 -6.07971132e-01 -1.50758505e-01 4.14308071e-01
-1.40226531e+00 -3.41651797e-01 6.85308337e-01 -3.48706096e-01
1.13445961e+00 -6.83269650e-02 4.56536859e-01 7.62579799e-01
2.10587174e-01 5.44951200e-01 1.02919209e+00 -2.51735866e-01
2.21346632e-01 8.04718956e-02 1.52396858e-01 6.11398876e-01
5.16717613e-01 3.79209071e-01 1.93579972e-01 5.10796666e-01
6.94137812e-01 -2.27218255e-01 2.31327727e-01 -2.00731933e-01
-7.74903774e-01 7.98457384e-01 9.01713967e-01 3.88828158e-01
-5.37825465e-01 3.36980790e-01 5.22256196e-01 1.71209648e-01
2.23201141e-01 3.34480628e-02 -1.81284875e-01 -1.39593393e-01
-9.43193197e-01 1.28183261e-01 7.38380432e-01 9.76934731e-01
7.91752577e-01 4.00140375e-01 3.27081084e-01 6.99953258e-01
2.96821356e-01 4.23500061e-01 7.56396428e-02 -8.72238338e-01
4.23955977e-01 5.59804440e-01 5.60639240e-02 -1.35909283e+00
-5.01360476e-01 -1.47643015e-01 -7.53172338e-01 6.94687307e-01
5.48062503e-01 -2.81153917e-01 -1.33496225e+00 1.08039558e+00
2.86488801e-01 5.53345740e-01 1.95175529e-01 5.61607301e-01
1.06617522e+00 3.70070815e-01 2.76858926e-01 6.48247838e-01
1.15081859e+00 -9.88523602e-01 -4.17933315e-01 -4.21751678e-01
7.03977466e-01 -7.32556164e-01 4.16636974e-01 2.17671692e-02
-5.95278919e-01 -8.31755519e-01 -1.17500925e+00 -2.74318159e-01
-1.21569037e+00 -1.79673582e-02 8.13088298e-01 9.45518911e-01
-1.31622839e+00 4.39621806e-01 -6.46675527e-01 -8.18551779e-01
5.77892244e-01 5.28385043e-01 -4.75851864e-01 5.06481640e-02
-1.48951840e+00 1.43470228e+00 3.39015067e-01 6.00372255e-01
-5.40940583e-01 -9.74438414e-02 -8.84266436e-01 -3.25474292e-01
2.38638353e-02 -3.70903313e-01 1.04674935e+00 -7.27090716e-01
-1.26992393e+00 1.08002067e+00 -1.34790376e-01 -9.73064959e-01
7.37040997e-01 -1.34977460e-01 -8.87564600e-01 -1.90236121e-01
4.83143181e-01 1.03616428e+00 3.03901047e-01 -9.76944208e-01
-1.36892688e+00 -5.77898547e-02 3.16461176e-01 -2.30139531e-02
5.96510053e-01 -3.80103976e-01 -3.54352802e-01 -1.06015101e-01
5.68527058e-02 -1.02990496e+00 -4.44893301e-01 -7.84612298e-02
-6.92686141e-01 -3.96869451e-01 9.74316597e-01 -4.07870620e-01
9.16898966e-01 -1.85752487e+00 -8.08994830e-01 4.45524603e-01
4.07945625e-02 8.11548769e-01 1.71762943e-01 4.19165879e-01
3.34221870e-02 1.31476223e-01 -5.72948158e-02 7.21143559e-02
-2.57769912e-01 5.31147540e-01 -1.80534303e-01 3.80742908e-01
-6.59950376e-02 8.51447046e-01 -7.81492531e-01 -4.00514275e-01
7.53758371e-01 5.57648182e-01 7.83307105e-02 -3.35477144e-01
4.33190018e-01 2.99619347e-01 -5.08309603e-01 6.60462320e-01
8.80775034e-01 3.71383995e-01 -4.11109000e-01 -1.27121940e-01
-7.05496073e-01 4.57218885e-01 -1.15267992e+00 9.35942590e-01
-4.24747497e-01 1.27427959e+00 -2.44962171e-01 -1.20459259e+00
1.11494124e+00 5.75638860e-02 4.59041037e-02 -1.07143569e+00
1.34083480e-01 4.38845128e-01 4.19163592e-02 -5.05136490e-01
9.37140644e-01 2.66690016e-01 2.73038089e-01 -3.22143137e-01
-3.16065848e-01 9.47581157e-02 1.77444562e-01 -1.09414339e-01
9.64622557e-01 1.33598104e-01 -2.04262473e-02 -1.10701606e-01
7.84930944e-01 4.00100917e-01 1.94808409e-01 8.73737872e-01
-6.25378847e-01 4.69581366e-01 9.16556045e-02 -8.43883455e-01
-1.12562692e+00 -9.36359107e-01 -3.86160046e-01 5.06725192e-01
3.48938316e-01 1.63341463e-01 -6.96269035e-01 -6.22218907e-01
2.02947110e-01 6.01016104e-01 -5.30229449e-01 8.92060846e-02
-7.21403956e-01 -2.57902473e-01 9.76799607e-01 7.22578764e-01
1.31642711e+00 -1.03620005e+00 -3.82137686e-01 3.63563359e-01
1.93877161e-01 -1.23987699e+00 3.94127995e-01 6.04522079e-02
-6.67046189e-01 -1.20499504e+00 -3.98782492e-01 -1.08917141e+00
4.08453375e-01 7.76332498e-01 8.36073339e-01 1.60447702e-01
1.51003469e-02 9.93682295e-02 -1.26061449e-02 -6.05131626e-01
-3.91624689e-01 4.70500737e-01 -2.23611936e-01 -2.98332334e-01
9.92599010e-01 -3.15026224e-01 -8.64230335e-01 4.68521297e-01
-4.04565334e-01 -1.61478028e-01 5.99017680e-01 2.99817473e-01
2.73975551e-01 2.50177324e-01 8.64868164e-01 -9.08319831e-01
3.81157517e-01 -7.17031062e-01 -7.85525620e-01 -3.04484397e-01
-6.25416696e-01 -2.57807106e-01 3.57873172e-01 3.07356060e-01
-9.00437891e-01 7.62860328e-02 -8.64983022e-01 2.10598916e-01
-8.09646130e-01 1.24713227e-01 1.52454637e-02 -5.70504010e-01
7.17623711e-01 -5.02296947e-02 -5.84400780e-02 -9.18167010e-02
5.54113030e-01 9.66798365e-01 6.86785519e-01 1.50699794e-01
8.12002420e-01 7.14462638e-01 2.71260202e-01 -1.22991776e+00
-5.88485241e-01 -6.21675909e-01 -7.54392505e-01 -4.50947642e-01
1.04173481e+00 -9.59571779e-01 -5.23503244e-01 6.07181072e-01
-8.26205134e-01 -9.87789035e-02 1.48837477e-01 4.36718613e-01
-3.09855103e-01 1.87430188e-01 -2.22286910e-01 -7.87088811e-01
-2.02821687e-01 -1.01678503e+00 6.82207108e-01 7.41507411e-01
-9.84643772e-02 -1.25521183e+00 -1.41784310e-01 4.85100627e-01
8.25918496e-01 4.05595034e-01 4.22998875e-01 -6.76212490e-01
-1.00529647e+00 -6.05612457e-01 -4.30029660e-01 3.29754174e-01
4.53536548e-02 1.42116919e-01 -1.12122238e+00 3.84270966e-01
-6.11271203e-01 3.51658165e-01 1.12964988e+00 5.08005559e-01
4.86613423e-01 3.30994904e-01 -7.22207189e-01 3.33915323e-01
1.52165377e+00 6.05720580e-01 1.13398886e+00 1.11003399e+00
7.82238483e-01 4.55170006e-01 5.46302795e-01 -4.62032527e-01
8.56538355e-01 4.35131401e-01 7.12559462e-01 -3.28702539e-01
-2.08166525e-01 -2.11888716e-01 -9.15270895e-02 1.77098006e-01
-3.51273745e-01 -1.27206683e-01 -1.18662179e+00 1.02429461e+00
-1.65548456e+00 -1.04977679e+00 -8.12647045e-01 2.04548883e+00
2.02400088e-01 6.57001793e-01 3.23480330e-02 2.28007540e-01
8.15365493e-01 8.31808150e-02 -2.01925784e-01 -9.26455259e-01
-6.59935847e-02 1.52235851e-01 1.29385710e+00 5.33046067e-01
-1.70969117e+00 1.37443709e+00 6.35578251e+00 5.69173634e-01
-1.34680140e+00 -2.89911240e-01 5.47349274e-01 8.81303906e-01
1.30537644e-01 7.64646754e-02 -1.02270603e+00 4.66802567e-01
1.22648442e+00 4.22966093e-01 -1.16656646e-01 1.09307981e+00
3.41416061e-01 -6.96752429e-01 -4.95133698e-01 6.21304154e-01
-4.23408419e-01 -1.48735619e+00 -2.05358565e-01 -1.00374771e-02
5.88427663e-01 5.98744392e-01 -1.95953697e-02 4.09795910e-01
5.37916124e-01 -1.12845850e+00 2.00303763e-01 5.60806036e-01
4.39775109e-01 -8.87196958e-01 9.84965742e-01 1.56210914e-01
-1.21370542e+00 2.47403029e-02 -5.18376291e-01 -2.81423122e-01
1.62400439e-01 3.27419728e-01 -1.04905021e+00 3.96411508e-01
6.57817721e-01 7.14151382e-01 -6.33681118e-01 1.56534791e+00
-3.27254921e-01 7.28781760e-01 -4.38510895e-01 -5.00815883e-02
8.92238736e-01 -3.72862667e-01 2.79716939e-01 1.40708542e+00
1.94644168e-01 -4.45178658e-01 -1.28679112e-01 3.32551867e-01
2.39096925e-01 -1.16757236e-01 -1.38207006e+00 5.34241021e-01
5.15135109e-01 1.27701128e+00 -9.40307140e-01 -3.07648540e-01
-6.30930901e-01 4.82988685e-01 1.25983074e-01 3.73053342e-01
-7.69803703e-01 -7.01023996e-01 9.51815069e-01 2.95942366e-01
3.64553183e-01 -4.52057511e-01 -4.55917120e-01 -4.65220451e-01
-2.50376850e-01 -1.28343806e-01 -1.05290085e-01 -6.98961079e-01
-8.28541040e-01 4.66573358e-01 -2.16522902e-01 -1.09693909e+00
-1.02621034e-01 -7.15397596e-01 -9.73254025e-01 8.21868122e-01
-2.08951569e+00 -1.18262720e+00 -3.10543329e-01 1.60445914e-01
4.72023517e-01 -2.68491149e-01 4.64443177e-01 6.79170012e-01
-4.80202883e-01 3.61765444e-01 3.39384794e-01 5.61287403e-01
3.85679483e-01 -1.24779463e+00 1.11730433e+00 8.17663670e-01
-1.77734688e-01 3.65027130e-01 6.63327157e-01 -6.12229943e-01
-8.44034374e-01 -1.50413382e+00 1.09120262e+00 -4.45742935e-01
6.44323409e-01 4.86451872e-02 -8.05313110e-01 7.20767498e-01
3.18215638e-01 -1.31402209e-01 2.32006446e-01 -1.85158998e-01
-2.86287256e-03 -2.45961353e-01 -1.16849530e+00 7.63059855e-01
1.00008917e+00 -5.95002353e-01 -5.21466076e-01 2.43297413e-01
3.87887992e-02 -3.91495377e-01 -3.96854371e-01 3.14616144e-01
5.44299722e-01 -1.10152459e+00 1.15862072e+00 -2.81358033e-01
1.61544353e-01 -6.03621900e-01 1.43045917e-01 -1.17958069e+00
-1.31534651e-01 -1.01267681e-01 4.59413707e-01 1.17149627e+00
5.08135617e-01 -9.29514706e-01 1.19369781e+00 5.83543360e-01
-3.50137442e-01 -5.65126717e-01 -1.11262619e+00 -6.25753105e-01
7.60650411e-02 -8.20904434e-01 3.93352211e-01 7.29637563e-01
-6.88877583e-01 3.58884126e-01 -1.51516855e-01 3.70453477e-01
3.41949463e-01 -3.46232504e-01 1.02862382e+00 -1.39423680e+00
7.95725226e-01 -6.18442059e-01 -1.33433867e+00 -9.66782272e-01
9.54681858e-02 -7.66741931e-01 -7.68216029e-02 -2.14943051e+00
-8.69994104e-01 -6.93965375e-01 4.58981795e-03 2.38242537e-01
8.23743343e-02 3.48871768e-01 -2.77569085e-01 9.57661122e-03
-4.09102291e-01 2.94668563e-02 8.58592272e-01 -1.31279394e-01
5.62608577e-02 4.45132494e-01 -4.00138855e-01 8.19128513e-01
1.14397967e+00 -1.64957792e-01 -3.67470622e-01 -2.64330089e-01
1.92061931e-01 -4.28509057e-01 5.44553339e-01 -1.56758296e+00
5.92300594e-01 2.66952842e-01 3.78489137e-01 -1.05422723e+00
1.35345757e-01 -1.03587389e+00 2.34200675e-02 2.80118942e-01
4.77335602e-02 5.70538975e-02 2.83743024e-01 7.52219737e-01
-3.67709160e-01 -7.11148158e-02 8.01901698e-01 -6.80693462e-02
-1.53364205e+00 6.04414679e-02 -8.73833120e-01 -3.11244339e-01
1.23398066e+00 -1.01133204e+00 -4.14867014e-01 -2.74981409e-01
-6.01052225e-01 6.04906738e-01 2.54145294e-01 6.45131648e-01
5.34614503e-01 -1.17258072e+00 -3.23950142e-01 2.19807684e-01
1.28899449e-02 9.41458791e-02 6.42757565e-02 7.47409523e-01
-1.22813129e+00 7.66706526e-01 -4.35023755e-01 -5.25710225e-01
-9.54166353e-01 3.55852574e-01 6.99705124e-01 2.91224927e-01
-7.46953964e-01 6.12965584e-01 -3.17716002e-01 -5.14892817e-01
9.73404497e-02 -3.86258811e-01 -8.94239306e-01 3.08278725e-02
1.91624403e-01 5.90621233e-01 2.34194309e-01 -1.22852051e+00
-3.71890903e-01 7.29108691e-01 9.21781547e-03 2.36925691e-01
9.18112695e-01 -3.64465475e-01 3.01919669e-01 1.38390452e-01
1.02298284e+00 -3.92563760e-01 -9.75514650e-01 -2.98543298e-03
4.78563905e-01 -2.24368289e-01 2.16861054e-01 -4.75068152e-01
-1.17928982e+00 9.48245704e-01 9.32660520e-01 4.83967245e-01
5.65214276e-01 -4.67588812e-01 9.95450556e-01 7.67423034e-01
3.79088014e-01 -1.44041967e+00 -9.51218128e-01 8.99715066e-01
2.27719232e-01 -1.75024939e+00 -3.75192195e-01 -2.78874248e-01
-4.15733099e-01 1.22246945e+00 6.79972887e-01 -4.76372361e-01
1.01569664e+00 1.50328755e-01 4.55494463e-01 -3.32149416e-01
-3.61994952e-02 -7.78814971e-01 -7.38715474e-03 1.04461801e+00
3.42675716e-01 1.29157454e-01 -2.23728552e-01 -3.12766343e-01
-2.10335001e-01 1.72733307e-01 5.71549118e-01 1.01736271e+00
-8.14839780e-01 -8.15356553e-01 -1.83846697e-01 5.61406672e-01
-5.49158335e-01 -1.37941062e-01 -3.29463124e-01 1.34376061e+00
2.22510278e-01 1.12614906e+00 2.07061097e-01 -2.70259798e-01
6.61033571e-01 3.61289531e-02 -3.22978616e-01 -2.05577165e-01
-4.11169380e-01 -6.27298892e-01 6.85782433e-01 -4.36594397e-01
-5.16815722e-01 -2.45941073e-01 -1.52152765e+00 -6.48677170e-01
-2.82516390e-01 4.54710871e-02 9.94842112e-01 1.05284810e+00
3.80800098e-01 3.70848596e-01 3.89544815e-01 -7.40399897e-01
2.69894809e-01 -6.32963777e-01 -4.04521525e-01 -1.82997901e-02
3.92953485e-01 -7.97179341e-01 -7.38440305e-02 -3.44562232e-01] | [8.668585777282715, -1.186037302017212] |
2ef7c070-5573-4e05-8a37-1c1b8c38bd96 | mean-covariance-robust-risk-measurement | 2112.09959 | null | https://arxiv.org/abs/2112.09959v1 | https://arxiv.org/pdf/2112.09959v1.pdf | Mean-Covariance Robust Risk Measurement | We introduce a universal framework for mean-covariance robust risk measurement and portfolio optimization. We model uncertainty in terms of the Gelbrich distance on the mean-covariance space, along with prior structural information about the population distribution. Our approach is related to the theory of optimal transport and exhibits superior statistical and computational properties than existing models. We find that, for a large class of risk measures, mean-covariance robust portfolio optimization boils down to the Markowitz model, subject to a regularization term given in closed form. This includes the finance standards, value-at-risk and conditional value-at-risk, and can be solved highly efficiently. | ['Daniel Kuhn', 'Damir Filipović', 'Soroosh Shafieezadeh Abadeh', 'Viet Anh Nguyen'] | 2021-12-18 | null | null | null | null | ['portfolio-optimization'] | ['time-series'] | [-1.50541067e-01 1.27032906e-01 -1.80674121e-01 -4.10018623e-01
-1.11770368e+00 -6.47031367e-01 5.40491760e-01 1.24741644e-02
-4.93212610e-01 8.56024444e-01 4.09400225e-01 -5.70175111e-01
-1.10622036e+00 -8.94815326e-01 -4.60779101e-01 -7.86220491e-01
-3.54711920e-01 5.00847936e-01 -2.37743989e-01 -6.95545897e-02
4.03346717e-01 4.18507695e-01 -7.92119443e-01 -6.81451559e-01
8.00929725e-01 1.42834496e+00 -4.97321039e-01 5.46037972e-01
2.58835256e-01 6.55628204e-01 -3.08346063e-01 -8.08068752e-01
4.80986655e-01 -1.64214298e-01 -5.32080173e-01 -1.44762054e-01
-1.69409499e-01 -2.67296255e-01 -2.43416533e-01 1.26278245e+00
3.80936354e-01 2.00606450e-01 1.29715300e+00 -8.26335490e-01
-7.92103529e-01 4.54371810e-01 -3.38545710e-01 3.52356166e-01
-1.03722177e-01 -1.26466155e-01 1.20880055e+00 -6.09482527e-01
2.75942981e-01 1.18077517e+00 5.71740508e-01 4.64967251e-01
-1.24960124e+00 -1.44216105e-01 -1.17192514e-01 -5.54897070e-01
-1.30672705e+00 -1.18959725e-01 3.39574039e-01 -7.91547894e-01
5.13401270e-01 4.04541224e-01 1.95377678e-01 6.49107158e-01
6.72724783e-01 1.14231728e-01 9.12904322e-01 -2.11943313e-01
4.96166766e-01 8.26136768e-02 2.35847414e-01 5.77031672e-01
7.43352890e-01 7.19402611e-01 -1.46516919e-01 -3.36343497e-01
6.09907031e-01 -3.37212346e-02 -6.41602725e-02 -3.64952058e-01
-8.85804057e-01 1.02746320e+00 -9.37872231e-02 1.24084745e-02
-2.88849473e-01 5.55953562e-01 -1.77243084e-01 2.66264617e-01
8.56903911e-01 2.78475314e-01 -2.75433421e-01 6.20713308e-02
-6.24361634e-01 4.84578490e-01 9.03737485e-01 8.64893734e-01
3.64788294e-01 1.76748276e-01 -4.30360466e-01 3.29558551e-01
9.28771138e-01 1.30841208e+00 -4.30225641e-01 -1.47924936e+00
7.24578857e-01 -1.46103531e-01 5.24452388e-01 -6.74597383e-01
-4.75509584e-01 -5.18508494e-01 -6.06232882e-01 4.96530384e-01
8.10533106e-01 -7.50914931e-01 -4.27885413e-01 1.74916208e+00
8.34890828e-02 9.21680108e-02 2.27241173e-01 5.29366255e-01
-3.33124071e-01 5.26358306e-01 -2.30998605e-01 -6.48491025e-01
9.20389533e-01 -2.89725542e-01 -7.86758602e-01 4.30519208e-02
5.48064411e-01 -6.46319449e-01 3.96728575e-01 2.11391106e-01
-1.32978964e+00 4.91683424e-01 -8.17193449e-01 4.40088212e-01
-8.32953751e-02 -7.68125772e-01 4.45590079e-01 1.30105484e+00
-8.93792272e-01 9.43105042e-01 -4.14061785e-01 1.99073002e-01
3.55516940e-01 1.00280210e-01 1.50434986e-01 2.40924060e-01
-9.15813386e-01 8.73704910e-01 -2.46601235e-02 2.50893950e-01
-8.56469274e-01 -1.05898857e+00 -6.60548508e-01 -4.49999459e-02
4.62652206e-01 -7.57197917e-01 1.23644578e+00 -2.48135731e-01
-1.60628450e+00 3.38663012e-01 2.76778698e-01 -4.24018621e-01
7.33266771e-01 -1.83211893e-01 -1.81312665e-01 2.45193243e-02
-4.00062948e-02 -5.36796808e-01 7.90700614e-01 -8.90981078e-01
-4.02818292e-01 -2.77562141e-01 -7.06184059e-02 1.09002357e-02
5.10505065e-02 5.78305900e-01 3.20446193e-01 -7.94274986e-01
-7.82581940e-02 -8.27223539e-01 -6.35221481e-01 -1.15552433e-01
-2.86157042e-01 2.15999305e-01 -7.53800794e-02 -5.32302260e-01
1.20183086e+00 -1.66009927e+00 -7.31310109e-03 9.26749945e-01
-1.24837287e-01 -4.75909770e-01 3.03441495e-01 3.44616562e-01
1.45188540e-01 5.19914985e-01 -5.78161716e-01 -2.39225507e-01
5.86792827e-01 2.74384860e-02 -3.00149918e-01 9.87612903e-01
-1.07469838e-02 7.77221560e-01 -6.91178799e-01 -4.85773981e-02
1.70860272e-02 2.07940057e-01 -6.16232514e-01 -1.07774762e-02
2.70063225e-02 2.30809107e-01 -8.07724953e-01 6.00697696e-01
6.31477952e-01 -1.82544261e-01 -2.16113627e-01 4.09223825e-01
-2.29322210e-01 -8.63040909e-02 -1.30698323e+00 1.05108666e+00
-3.55351001e-01 1.19148970e-01 2.49932453e-01 -8.13631296e-01
5.77304184e-01 2.44983450e-01 7.55932391e-01 -2.40360647e-01
3.06446671e-01 2.59877592e-01 -2.14671522e-01 -1.34849206e-01
3.09347808e-01 -7.88182199e-01 -3.67829353e-01 8.72854292e-01
-1.91188425e-01 -8.04173946e-02 -5.68767413e-02 6.66678417e-04
7.98667073e-01 -4.59745854e-01 7.56462151e-03 -9.72065032e-01
1.94335237e-01 -4.78001297e-01 4.87841517e-01 8.97589087e-01
-1.88865811e-01 4.39161330e-01 8.04871798e-01 1.48945257e-01
-9.65749204e-01 -1.60152912e+00 -6.20200038e-01 4.96438473e-01
-1.16174169e-01 -2.70749461e-02 -8.53118777e-01 -5.98086298e-01
3.89476925e-01 8.86178017e-01 -6.91348374e-01 -1.83606163e-01
-3.35013926e-01 -1.27208138e+00 3.34737509e-01 3.61598134e-01
6.21677116e-02 -2.28547230e-01 -1.28483064e-02 1.44345939e-01
5.15074670e-01 -4.20515686e-01 -7.71141231e-01 -1.40313745e-01
-7.65022457e-01 -1.10973489e+00 -1.05764771e+00 7.69268647e-02
5.69586813e-01 -3.24604988e-01 8.70078921e-01 -3.54701370e-01
-5.42357601e-02 8.58920634e-01 1.91721991e-01 -7.20730424e-01
-1.36497945e-01 -7.93027580e-01 2.42670327e-01 2.84473062e-01
-1.59747839e-01 -8.96125361e-02 -7.01110005e-01 3.71115923e-01
-3.94120216e-01 -1.00415909e+00 1.63415208e-01 6.57224357e-01
5.86763263e-01 3.90034020e-01 5.90833604e-01 -6.44655526e-01
7.51500309e-01 -5.34534335e-01 -1.14450538e+00 4.16284114e-01
-1.07050908e+00 2.62164921e-01 -2.09583879e-01 -3.56802568e-02
-1.40972674e+00 -5.88769913e-01 3.48046944e-02 8.63074511e-02
4.90037292e-01 4.74850118e-01 -1.67396337e-01 -2.97409385e-01
2.50919640e-01 -4.59170818e-01 7.99742714e-02 -6.21571720e-01
3.98729473e-01 4.13556904e-01 1.33000776e-01 -9.18152690e-01
9.37158108e-01 6.73201561e-01 7.83308148e-01 -5.34219384e-01
-1.15560627e+00 1.01789020e-01 -4.54677284e-01 -1.94342360e-01
1.11178970e+00 -5.32005250e-01 -8.67906570e-01 3.35912704e-01
-7.16180325e-01 -2.72656113e-01 -8.25191319e-01 1.01903546e+00
-9.26808357e-01 1.47301391e-01 -6.30272806e-01 -1.53083730e+00
-2.05712393e-01 -9.82618272e-01 8.24961841e-01 -9.43398997e-02
1.60597056e-01 -1.69768441e+00 1.96322113e-01 1.28835604e-01
7.13428795e-01 4.32368755e-01 7.90636122e-01 -5.13676584e-01
-6.47317946e-01 -3.70059431e-01 -1.99044660e-01 5.72910488e-01
-2.48488441e-01 -6.02288805e-02 -6.02311015e-01 -8.85894746e-02
4.34604347e-01 3.71786416e-01 7.53893495e-01 8.52965534e-01
7.23521650e-01 -6.20195270e-01 7.45365992e-02 7.61644125e-01
1.50595069e+00 1.40829012e-01 2.31123701e-01 2.38680780e-01
3.19989175e-01 9.13779140e-01 4.24282134e-01 7.89584517e-01
3.10165405e-01 2.64980316e-01 2.52400547e-01 4.96520877e-01
7.06659317e-01 4.51138389e-04 3.73749316e-01 5.78484893e-01
-2.01353371e-01 2.87096538e-02 -9.52430964e-01 2.73430377e-01
-1.77521896e+00 -1.15055966e+00 -1.45488486e-01 2.52006793e+00
4.08478767e-01 2.34213725e-01 1.52826488e-01 -3.96074504e-01
5.42102993e-01 6.66779280e-02 -2.29205266e-01 -3.38752717e-01
-2.64302462e-01 6.85271844e-02 1.39681864e+00 1.06494045e+00
-9.48394299e-01 1.23042256e-01 8.82693481e+00 7.82169521e-01
-2.58831382e-01 2.75635988e-01 7.53297210e-01 -2.99964815e-01
-7.30010152e-01 7.75207356e-02 -8.62150371e-01 6.27573013e-01
1.28180695e+00 -8.99718285e-01 3.85877073e-01 5.77366650e-01
2.99685389e-01 1.22875519e-01 -9.44800735e-01 4.72573549e-01
-1.55051649e-01 -1.19131351e+00 -5.28110802e-01 7.32533276e-01
9.57714081e-01 -2.16735139e-01 5.30768752e-01 -4.68908325e-02
6.23449385e-01 -8.81658554e-01 8.40589404e-01 1.25781071e+00
4.86554652e-01 -1.22480702e+00 8.58640671e-01 8.69555697e-02
-8.00626159e-01 -3.03520530e-01 -5.54105699e-01 8.84005949e-02
6.52659297e-01 9.81352091e-01 2.41196856e-01 4.44730252e-01
4.58097994e-01 4.30210292e-01 1.38411000e-01 1.04752100e+00
-2.93933935e-02 3.59103531e-01 -4.92524236e-01 9.50159803e-02
1.84252203e-01 -7.54791558e-01 7.26401806e-01 8.89170945e-01
8.34678352e-01 1.14162892e-01 -1.81042999e-01 9.85218942e-01
3.57332379e-02 2.10479453e-01 -5.75871468e-01 -5.30027673e-02
9.08353403e-02 6.45260394e-01 -2.24604130e-01 -4.88517918e-02
-4.58238631e-01 2.55966961e-01 -2.93763906e-01 4.67182100e-01
-6.07513189e-01 -4.46548998e-01 8.95427167e-01 -6.63623586e-02
1.03510909e-01 -1.87470362e-01 -4.71779466e-01 -1.10751176e+00
2.47485321e-02 -1.05978511e-01 5.64193606e-01 -1.21881060e-01
-1.70194614e+00 -1.15600070e-02 4.14234936e-01 -8.00123155e-01
-3.60233784e-01 -9.87188578e-01 -8.19868147e-01 1.14857817e+00
-1.54824829e+00 -5.34312189e-01 7.30276704e-01 4.41936553e-01
-2.19989836e-01 -5.58274865e-01 4.45756167e-01 1.80711314e-01
-8.05890739e-01 4.06887114e-01 9.83673692e-01 -1.48982242e-01
2.31533423e-01 -1.64443076e+00 1.98008105e-01 8.45149577e-01
-4.77861762e-01 5.97818434e-01 8.42915475e-01 -7.61396646e-01
-1.36726117e+00 -1.06775951e+00 7.18297124e-01 -8.07880700e-01
1.33282983e+00 -2.09365174e-01 -4.80827928e-01 7.44423568e-01
6.68120459e-02 7.02394685e-03 7.49913394e-01 1.46578521e-01
-4.92630117e-02 -2.64322728e-01 -1.42926896e+00 4.15905058e-01
7.22537696e-01 -5.61512589e-01 -4.10043687e-01 7.12126911e-01
7.12602258e-01 1.13216467e-01 -1.40084779e+00 2.36254007e-01
5.67370832e-01 -6.29613161e-01 1.05372310e+00 -7.63450980e-01
-2.70538181e-01 1.89438015e-01 -6.90705776e-01 -1.07481313e+00
-2.39690334e-01 -1.24821115e+00 2.67827138e-02 1.19762897e+00
6.96911812e-01 -9.43054914e-01 5.78682542e-01 1.17671585e+00
9.68877077e-02 -6.72257423e-01 -1.40105116e+00 -1.39254820e+00
9.12264287e-01 -8.35049987e-01 6.67259932e-01 3.52257609e-01
9.34110805e-02 -2.64879256e-01 -5.83934963e-01 3.42671663e-01
1.72084892e+00 -3.01158667e-01 -4.81109768e-02 -1.29534888e+00
-4.01583463e-01 -7.56948531e-01 -2.80059292e-04 -6.94949210e-01
2.47978568e-01 -8.31902266e-01 -8.62015784e-02 -1.37152588e+00
3.81826162e-02 -4.66686666e-01 -5.45225263e-01 -3.43380332e-01
1.56636313e-01 -1.35054812e-01 5.22652268e-02 -9.84294154e-03
-4.01990503e-01 4.93725419e-01 9.97910261e-01 -1.34387019e-03
1.25287145e-01 5.19843340e-01 -7.86922514e-01 1.01552606e+00
7.88456142e-01 -6.87573493e-01 -4.13505614e-01 -8.26138109e-02
5.97562134e-01 3.79698485e-01 2.05590472e-01 -4.41301197e-01
-2.78287707e-03 -8.48071754e-01 -1.16986215e-01 -4.21113402e-01
2.11617664e-01 -7.93575287e-01 1.60900012e-01 3.56740981e-01
-4.86051083e-01 3.28443013e-02 -4.09488797e-01 8.07561874e-01
3.16353440e-01 -8.27472150e-01 8.88808191e-01 -2.89411768e-02
1.14564843e-01 4.58561450e-01 -3.63381267e-01 4.95804667e-01
1.22197855e+00 3.59526157e-01 -7.44515061e-01 -4.58157569e-01
-7.98198879e-01 3.27168167e-01 2.28426158e-01 -2.06493855e-01
5.12641013e-01 -1.54711306e+00 -8.26225042e-01 -5.94325177e-02
-2.68583000e-01 -5.48249006e-01 2.31774673e-01 8.96117210e-01
-4.40391213e-01 6.54786587e-01 4.39692467e-01 -7.14656264e-02
-3.84979457e-01 6.99684441e-01 8.38784039e-01 -1.36164561e-01
-2.99042672e-01 8.12518775e-01 2.03906611e-01 6.00652993e-02
2.04778552e-01 -8.21048841e-02 1.72267526e-01 7.02764243e-02
8.59675348e-01 9.28411722e-01 -5.13688862e-01 -7.02840269e-01
-2.52305597e-01 9.95461583e-01 3.59959394e-01 -6.83606207e-01
1.06983209e+00 -5.94915509e-01 -2.57732391e-01 5.35942018e-01
1.24960661e+00 1.90853506e-01 -1.36645544e+00 -2.31624365e-01
5.68484902e-01 -6.24949396e-01 1.66988045e-01 -5.42448640e-01
-9.91396606e-01 7.40872920e-01 4.98824954e-01 2.71581590e-01
5.08661032e-01 -1.26707237e-02 3.67217869e-01 5.00093460e-01
3.64119440e-01 -1.33451283e+00 -1.43994480e-01 4.49519843e-01
8.82956147e-01 -1.00806653e+00 1.05826572e-01 -2.39915833e-01
-6.48338675e-01 6.42472982e-01 -2.59466290e-01 -3.98934752e-01
1.80054879e+00 5.55730343e-01 -1.97903261e-01 -3.88819575e-02
-7.35342026e-01 -2.68290102e-01 6.72018766e-01 7.54079640e-01
2.28154492e-02 2.20872834e-01 -2.50018954e-01 7.77541339e-01
-6.03817105e-02 -5.68878055e-01 5.87592602e-01 7.01778710e-01
-7.80291855e-01 -1.05762243e+00 -3.75688761e-01 6.34839296e-01
-1.06972420e+00 2.99517880e-03 1.89398557e-01 5.29437721e-01
-4.01392996e-01 1.11875653e+00 1.81390002e-01 1.47721499e-01
6.12718999e-01 -1.44854039e-01 3.09034288e-01 -4.04670954e-01
-2.10571364e-01 3.83043200e-01 2.19276726e-01 -5.89838326e-01
-1.86555982e-01 -1.16829729e+00 -8.22796047e-01 -5.69647193e-01
-2.32533887e-01 3.74162823e-01 5.94932914e-01 1.04415143e+00
-1.82828665e-01 2.62115568e-01 9.62958217e-01 -4.82734263e-01
-1.40800297e+00 -4.27275836e-01 -1.52216721e+00 -7.19150379e-02
4.87157702e-01 -6.48015916e-01 -8.68276715e-01 -4.93908703e-01] | [5.00227689743042, 3.9218759536743164] |
d307f0a8-65f3-4da5-8581-5d3133904abc | privacy-preserving-deep-learning-based-record | 2211.02161 | null | https://arxiv.org/abs/2211.02161v1 | https://arxiv.org/pdf/2211.02161v1.pdf | Privacy-preserving Deep Learning based Record Linkage | Deep learning-based linkage of records across different databases is becoming increasingly useful in data integration and mining applications to discover new insights from multiple sources of data. However, due to privacy and confidentiality concerns, organisations often are not willing or allowed to share their sensitive data with any external parties, thus making it challenging to build/train deep learning models for record linkage across different organizations' databases. To overcome this limitation, we propose the first deep learning-based multi-party privacy-preserving record linkage (PPRL) protocol that can be used to link sensitive databases held by multiple different organisations. In our approach, each database owner first trains a local deep learning model, which is then uploaded to a secure environment and securely aggregated to create a global model. The global model is then used by a linkage unit to distinguish unlabelled record pairs as matches and non-matches. We utilise differential privacy to achieve provable privacy protection against re-identification attacks. We evaluate the linkage quality and scalability of our approach using several large real-world databases, showing that it can achieve high linkage quality while providing sufficient privacy protection against existing attacks. | ['Ming Ding', 'Dinusha Vatsalan', 'Thilina Ranbaduge'] | 2022-11-03 | null | null | null | null | ['data-integration', 'privacy-preserving-deep-learning', 'privacy-preserving-deep-learning'] | ['knowledge-base', 'methodology', 'natural-language-processing'] | [ 8.23380649e-02 9.74853560e-02 -2.41503850e-01 -6.67700827e-01
-9.87164199e-01 -1.09859169e+00 3.31471056e-01 1.07812846e+00
-5.58791399e-01 7.38883555e-01 1.44368196e-02 -5.17293453e-01
-3.85738075e-01 -1.22396886e+00 -9.08920825e-01 -6.36824608e-01
-2.81409681e-01 7.51014471e-01 -4.47249562e-02 1.06790043e-01
1.45777818e-02 8.53259921e-01 -9.32229578e-01 5.99185228e-01
4.73832279e-01 7.56089151e-01 -6.29523575e-01 2.95433491e-01
-1.48323849e-01 4.38319266e-01 -3.60228896e-01 -1.07521236e+00
8.34041119e-01 2.32727826e-01 -1.11389554e+00 -7.70189703e-01
3.26082915e-01 -5.00417411e-01 -1.99815571e-01 1.07917142e+00
8.20451021e-01 -4.46941435e-01 -3.98070216e-02 -1.52464128e+00
-7.00573921e-01 7.49883533e-01 -6.23747289e-01 -1.02764614e-01
4.29510087e-01 -2.18615443e-01 8.88064981e-01 4.55075987e-02
9.62800980e-01 1.10589159e+00 8.56352866e-01 5.38196981e-01
-1.64028585e+00 -1.02353418e+00 -4.37092304e-01 -1.26112878e-01
-1.08483362e+00 -4.26104546e-01 5.27516365e-01 -2.76395619e-01
5.74101090e-01 5.49937844e-01 8.12794194e-02 1.00736749e+00
2.89922297e-01 4.72530752e-01 8.60097229e-01 -1.22065037e-01
2.96069771e-01 5.05047619e-01 1.78210706e-01 8.43913555e-02
5.52320600e-01 1.88792825e-01 -3.35571557e-01 -1.10436296e+00
2.80448288e-01 1.59834445e-01 -3.13202254e-02 -8.56713831e-01
-8.95446956e-01 9.83299732e-01 7.27862656e-01 1.02050111e-01
-3.40627789e-01 -1.62743311e-02 7.89211869e-01 8.41453969e-01
2.56938070e-01 4.23451692e-01 -4.21589702e-01 4.56256688e-01
-5.32061517e-01 5.87251246e-01 1.07197893e+00 8.63057554e-01
6.87666714e-01 -9.82474983e-01 3.28305691e-01 5.13383985e-01
2.92917371e-01 -3.08206473e-02 2.35262839e-03 -8.55452597e-01
5.46030998e-01 7.37680316e-01 6.51996061e-02 -1.28181803e+00
-2.15757132e-01 -6.72818795e-02 -1.10388291e+00 1.98151454e-01
3.80191296e-01 -7.79258087e-02 -9.79357585e-02 1.69547844e+00
5.81257463e-01 -2.66686380e-01 4.44701850e-01 5.34260929e-01
5.56480706e-01 3.63383412e-01 3.07321757e-01 -5.28113218e-03
1.46507537e+00 5.54826595e-02 -6.25006735e-01 8.55193436e-02
9.35067475e-01 -3.11763078e-01 1.18893728e-01 5.60019352e-02
-8.96687388e-01 -3.29499841e-02 -9.97281969e-01 -3.22965056e-01
-7.28334963e-01 -8.12367439e-01 7.72691488e-01 8.57590318e-01
-1.09516215e+00 4.64589775e-01 -7.91373074e-01 -4.31010365e-01
1.14869714e+00 1.05060625e+00 -1.16208303e+00 -9.27531719e-02
-1.47933793e+00 5.80408394e-01 6.01964355e-01 -1.64331764e-01
-2.91119218e-01 -8.01693678e-01 -6.72423124e-01 -6.70401752e-02
-9.28011164e-03 -4.60226208e-01 8.87017727e-01 -5.58906078e-01
-5.66299915e-01 1.33946788e+00 3.16322565e-01 -9.44834054e-01
7.75899351e-01 4.88094017e-02 -4.43314582e-01 -2.50819862e-01
1.30749062e-01 3.90800446e-01 3.21702920e-02 -1.38217032e+00
-8.28669429e-01 -9.39440489e-01 -3.26808766e-02 -2.68716246e-01
-3.02833050e-01 5.37767529e-01 -4.42982651e-03 -2.31366560e-01
-1.90797839e-02 -7.96274424e-01 -2.16746092e-01 2.87804812e-01
-4.46535021e-01 -5.96148474e-03 1.06474698e+00 -8.62468958e-01
7.05907285e-01 -2.17683887e+00 -2.71622628e-01 6.64943755e-01
2.46938154e-01 5.56933343e-01 4.28476855e-02 5.54036081e-01
-2.13088989e-02 3.23041528e-01 -3.88179332e-01 -2.43389159e-01
-6.09925315e-02 1.65455863e-01 -3.37122858e-01 6.07623935e-01
1.42491341e-01 8.96203816e-01 -7.31725276e-01 -4.39268142e-01
-7.70144016e-02 5.47444284e-01 -3.88468593e-01 2.90945798e-01
4.81156744e-02 2.75260866e-01 -5.09281993e-01 6.17991626e-01
1.20146894e+00 9.81226414e-02 6.75969064e-01 8.15756172e-02
2.22250149e-01 1.07612625e-01 -1.22990572e+00 1.62605321e+00
-8.96213874e-02 3.38662833e-01 3.26798350e-01 -9.73335981e-01
1.32700872e+00 4.50674593e-01 5.21591187e-01 -6.29133999e-01
2.48471610e-02 3.69504541e-01 -3.46434265e-01 -1.41356602e-01
2.86044478e-01 1.89482458e-02 -4.35367018e-01 7.53407776e-01
-2.08766416e-01 6.45044386e-01 -5.36796153e-01 5.36544658e-02
1.11772931e+00 -3.69390756e-01 5.34760058e-02 -1.56642556e-01
5.37048459e-01 -3.13769639e-01 6.66338027e-01 7.72011101e-01
-2.53876567e-01 3.39276493e-01 6.66566193e-01 -1.08535266e+00
-1.08550203e+00 -1.14067733e+00 -2.77066439e-01 9.35480893e-01
-1.85174197e-01 1.28101841e-01 -6.45054221e-01 -8.46926272e-01
7.25470483e-01 2.31176332e-01 -4.90772158e-01 -2.89751291e-01
-6.79495931e-01 -8.27209830e-01 9.72835362e-01 3.57116550e-01
4.50790733e-01 -1.09749985e+00 -4.55373526e-01 4.56494033e-01
-7.42149279e-02 -7.01164484e-01 -1.59028620e-02 3.74097437e-01
-5.19029260e-01 -1.21119213e+00 -1.50019675e-01 -6.65973961e-01
3.18872958e-01 -2.98182338e-01 6.26887262e-01 -1.20990798e-01
-1.96482688e-01 -2.33986244e-01 1.46691144e-01 -4.63774830e-01
-8.56467485e-01 3.40135694e-01 1.38963521e-01 1.75756514e-01
9.87597048e-01 -6.17940426e-01 -3.77099127e-01 -2.58613396e-02
-1.16147828e+00 -6.75780118e-01 2.23856390e-01 3.94497722e-01
4.29760605e-01 8.22711587e-02 8.70092332e-01 -1.31416571e+00
7.13169634e-01 -6.97373986e-01 -9.29577470e-01 4.66800243e-01
-5.27994633e-01 9.09045786e-02 2.91956484e-01 -2.55842090e-01
-7.71368921e-01 1.81665063e-01 -2.50835884e-02 2.85191052e-02
-2.18045369e-01 2.52840817e-01 -7.27758050e-01 -3.29486400e-01
4.38423604e-01 -2.56217867e-01 1.79627001e-01 -6.00778997e-01
3.69678289e-01 1.14745045e+00 8.99709761e-01 -4.36096877e-01
7.78621912e-01 5.68844855e-01 1.18182898e-01 -1.37531862e-01
-1.23121150e-01 -1.19233817e-01 -7.81785488e-01 5.85941672e-01
9.01560485e-01 -8.55908453e-01 -1.19238007e+00 5.56886792e-01
-1.30688393e+00 4.85391825e-01 -5.24496436e-02 1.16997408e-02
-2.51530230e-01 2.38709509e-01 -7.12354124e-01 -6.24532819e-01
-7.39055097e-01 -8.37798953e-01 6.57649934e-01 6.26344159e-02
-4.63506281e-01 -8.49349618e-01 2.16422871e-01 4.83573675e-01
3.55297178e-01 8.00471008e-01 1.08235645e+00 -1.45050859e+00
-5.37413299e-01 -7.97333658e-01 -3.59647036e-01 9.78000760e-02
2.27103055e-01 -3.95013332e-01 -1.08015430e+00 -6.83039486e-01
6.76823631e-02 -3.14505160e-01 4.23021346e-01 -1.64829478e-01
1.08877170e+00 -6.07763231e-01 -6.08893633e-01 8.95577192e-01
1.54184508e+00 2.92295516e-01 7.23491848e-01 9.06295419e-01
6.70361936e-01 8.75782192e-01 8.44302252e-02 3.26561272e-01
4.37012881e-01 5.13672173e-01 4.62967873e-01 -8.01793858e-02
9.03619170e-01 -3.39443326e-01 -2.76970834e-01 -1.15205586e-01
4.92640287e-01 -2.46584445e-01 -9.86341238e-01 5.87622166e-01
-1.92481542e+00 -7.43694127e-01 2.70436164e-02 2.47771835e+00
9.88527536e-01 -1.78956762e-01 1.20778441e-01 -3.67090739e-02
7.90614605e-01 -1.01096161e-01 -5.78404248e-01 -7.59128392e-01
-1.90215722e-01 2.39882350e-01 9.65829611e-01 1.74375996e-01
-1.37539208e+00 4.65475708e-01 5.14184093e+00 -5.66201145e-03
-8.64116907e-01 5.99480607e-03 8.16445112e-01 -1.00113854e-01
-3.73518407e-01 7.00709075e-02 -5.96022308e-01 4.29019988e-01
1.04027498e+00 -3.59131902e-01 8.71193334e-02 6.21501505e-01
-4.45851654e-01 4.33075309e-01 -1.45646822e+00 6.96485639e-01
-6.66786075e-01 -1.55081058e+00 -3.02195847e-02 7.08784342e-01
2.85727352e-01 -2.64770314e-02 -4.07131528e-03 -2.60539830e-01
8.80818129e-01 -1.12977242e+00 1.65449306e-01 4.16382372e-01
6.72767401e-01 -1.44212627e+00 1.05235124e+00 1.74783528e-01
-6.01885676e-01 -3.40223819e-01 -4.71855521e-01 4.59687710e-01
-1.75260887e-01 2.44195074e-01 -6.43736422e-01 9.91551340e-01
9.52732086e-01 2.00796247e-01 -3.53880435e-01 7.97653675e-01
4.31528449e-01 -9.17553008e-02 -5.42514741e-01 6.27895296e-01
4.35744673e-02 -1.63319837e-02 1.23099349e-01 9.91533637e-01
2.93015808e-01 -7.88532868e-02 -3.93271968e-02 9.09227133e-01
-7.20887959e-01 -4.02354226e-02 -9.50344145e-01 5.45355398e-03
8.41307104e-01 9.66231227e-01 -1.56220824e-01 1.76523089e-01
-2.61219949e-01 8.09871852e-01 4.19260561e-01 -3.88435833e-02
-1.46993816e-01 -4.72514957e-01 1.14354551e+00 1.05691954e-01
6.34292066e-02 3.33149046e-01 -1.06118575e-01 -7.48444676e-01
2.62001455e-01 -1.07603335e+00 1.01692271e+00 -6.69992939e-02
-1.52503812e+00 4.15321946e-01 -2.31550202e-01 -7.47767329e-01
-2.82055050e-01 -1.93338301e-02 -3.21834743e-01 1.23661149e+00
-1.29558444e+00 -1.52121687e+00 2.57513791e-01 5.69830000e-01
-7.72141159e-01 -3.96658450e-01 1.16436636e+00 4.55306649e-01
-3.54845673e-01 1.15027297e+00 4.59780663e-01 6.27084076e-01
7.31487095e-01 -9.68566895e-01 9.62129593e-01 6.67822421e-01
-6.91120401e-02 8.27726245e-01 3.54323894e-01 -7.51213431e-01
-1.59084880e+00 -1.33662689e+00 1.34614646e+00 -6.66598797e-01
3.48403633e-01 -8.66531432e-01 -1.58980978e+00 9.76668656e-01
-7.74488673e-02 3.56621563e-01 1.10324419e+00 -3.36631276e-02
-8.14772666e-01 -5.08854985e-01 -1.89791965e+00 5.93716875e-02
5.70857406e-01 -9.41908658e-01 -4.15352434e-01 1.22149721e-01
9.15485024e-01 -4.66044061e-02 -1.28241956e+00 2.16148511e-01
7.34251201e-01 -9.32803512e-01 1.04526210e+00 -9.29178596e-01
3.03325746e-02 -7.13391677e-02 -2.29467034e-01 -6.28230095e-01
-1.41594440e-01 -8.23802650e-01 9.89966989e-02 1.93643367e+00
1.57035291e-01 -1.07225049e+00 9.78022814e-01 1.19207776e+00
8.05388749e-01 -1.27240047e-01 -1.23970437e+00 -5.59537768e-01
4.06041563e-01 -5.59684634e-02 1.54106212e+00 1.59459484e+00
-4.77280729e-02 -3.75860512e-01 -3.35515231e-01 7.23412335e-01
1.07020020e+00 1.54840037e-01 8.55649292e-01 -1.61369503e+00
1.38423175e-01 1.52520165e-02 -6.68670177e-01 5.09023294e-02
3.01261634e-01 -1.16061735e+00 -5.31160235e-01 -1.19927239e+00
1.60095081e-01 -8.64685655e-01 -7.78348386e-01 8.02862167e-01
2.52596021e-01 -4.29000556e-02 -1.16769202e-01 1.98295519e-01
-5.08503616e-02 -1.11346535e-01 1.52960688e-01 -9.08089578e-02
-2.23641723e-01 1.06052175e-01 -1.20690775e+00 1.32846087e-01
7.52896726e-01 -9.62393343e-01 -1.67833678e-02 -4.82604802e-01
3.24969113e-01 1.37395799e-01 5.68028331e-01 -7.88833022e-01
4.32274818e-01 1.37001857e-01 4.33999985e-01 -4.74419743e-01
-3.08772117e-01 -1.21726871e+00 8.26218367e-01 5.34414947e-01
-6.27323031e-01 -7.42442608e-02 2.93260217e-01 6.44317091e-01
-1.91364288e-02 3.05746049e-01 7.28405654e-01 3.25435065e-02
-1.72326848e-01 5.69769502e-01 3.30317169e-01 -4.46840227e-01
1.15023899e+00 -6.13106564e-02 -4.61292565e-01 3.58093269e-02
-5.00651479e-01 6.30672097e-01 1.02046263e+00 6.18854821e-01
3.21604311e-01 -1.36860180e+00 -9.26713467e-01 6.11368477e-01
3.32622766e-01 1.65639788e-01 3.48884426e-02 1.33343950e-01
-5.17797351e-01 2.30482131e-01 -6.37947261e-01 -3.09915811e-01
-1.49431241e+00 9.76714015e-01 1.93271071e-01 -1.93885729e-01
-5.38419068e-01 7.43872344e-01 -1.26233250e-01 -9.68158901e-01
3.47147346e-01 1.59945950e-01 2.66159862e-01 1.26002565e-01
7.29150474e-01 2.67067045e-01 2.94530869e-01 -6.01660550e-01
-7.51397014e-01 9.13897604e-02 -5.90856314e-01 1.18010670e-01
1.62940514e+00 2.85956766e-02 -7.29453862e-01 -1.40899643e-01
1.69189394e+00 -2.77123004e-01 -8.29217017e-01 -4.01347160e-01
5.73221624e-01 -6.21523976e-01 -3.25561345e-01 -5.48499882e-01
-1.15376854e+00 6.32324219e-01 9.83721077e-01 2.51937270e-01
8.71132493e-01 6.33412153e-02 9.56303537e-01 5.45494437e-01
5.17499804e-01 -6.22899950e-01 -1.01966941e+00 -1.57795727e-01
4.10053879e-01 -1.34105515e+00 2.44975835e-03 -1.39283976e-02
-3.27107847e-01 7.84061253e-01 2.17894301e-01 2.27596492e-01
5.00622153e-01 4.39875305e-01 3.34855944e-01 -1.08377330e-01
-5.29814601e-01 7.61209071e-01 -2.95454115e-01 1.07882738e+00
4.85232696e-02 3.99640612e-02 1.81990881e-02 5.30294120e-01
-1.75684750e-01 1.89120933e-01 3.03466648e-01 1.16812468e+00
1.64874852e-01 -1.84584641e+00 -2.19877869e-01 3.96595478e-01
-9.84892309e-01 2.87569970e-01 -6.25333667e-01 5.43369234e-01
7.35105723e-02 6.62952125e-01 2.63536304e-01 -2.62323469e-01
2.80902982e-01 1.64915249e-01 -1.74723655e-01 -1.02009408e-01
-1.17141736e+00 -4.77578580e-01 -1.99848972e-02 -5.46259284e-01
-1.91426516e-01 -7.53139257e-01 -9.27805483e-01 -7.87561119e-01
1.51744662e-02 2.50722289e-01 5.86959124e-01 3.78127307e-01
7.45281756e-01 -3.41952860e-01 7.57356882e-01 6.25049472e-02
-4.98045594e-01 -4.77891900e-02 -5.38649797e-01 7.16473579e-01
7.02069521e-01 1.05950497e-01 -1.01527311e-01 -2.92502195e-01] | [5.944058895111084, 6.740748405456543] |
2ba47d88-f1b1-4581-98e0-82bef866895b | analyzing-and-simulating-user-utterance | 2205.01763 | null | https://arxiv.org/abs/2205.01763v1 | https://arxiv.org/pdf/2205.01763v1.pdf | Analyzing and Simulating User Utterance Reformulation in Conversational Recommender Systems | User simulation has been a cost-effective technique for evaluating conversational recommender systems. However, building a human-like simulator is still an open challenge. In this work, we focus on how users reformulate their utterances when a conversational agent fails to understand them. First, we perform a user study, involving five conversational agents across different domains, to identify common reformulation types and their transition relationships. A common pattern that emerges is that persistent users would first try to rephrase, then simplify, before giving up. Next, to incorporate the observed reformulation behavior in a user simulator, we introduce the task of reformulation sequence generation: to generate a sequence of reformulated utterances with a given intent (rephrase or simplify). We develop methods by extending transformer models guided by the reformulation type and perform further filtering based on estimated reading difficulty. We demonstrate the effectiveness of our approach using both automatic and human evaluation. | ['Krisztian Balog', 'Mu-Chun Wang', 'Shuo Zhang'] | 2022-05-03 | null | null | null | null | ['user-simulation'] | ['natural-language-processing'] | [ 2.74296284e-01 3.77869964e-01 4.01534140e-01 -5.86508751e-01
-4.40141469e-01 -8.65654826e-01 1.04134810e+00 1.57662764e-01
-1.31521061e-01 6.56085134e-01 5.44210494e-01 -5.52975953e-01
4.51830029e-02 -6.04959726e-01 -3.30592722e-01 -1.23342872e-01
1.90834939e-01 9.34818089e-01 1.00234471e-01 -6.61205888e-01
3.24409336e-01 6.21711053e-02 -1.89549387e+00 6.23868763e-01
1.23858285e+00 1.22907816e-03 6.13977253e-01 1.25313938e+00
-1.32518560e-01 9.14203584e-01 -1.00998628e+00 -3.63174021e-01
-2.03483537e-01 -8.28440249e-01 -1.32015908e+00 2.70203739e-01
5.52388690e-02 -5.32337904e-01 -1.44372117e-02 6.49832249e-01
4.59924668e-01 6.19859099e-01 6.73950613e-01 -1.30173945e+00
-4.80263650e-01 1.01677823e+00 4.89442706e-01 -3.37025039e-02
1.28643227e+00 2.62817532e-01 8.20740223e-01 -4.60090399e-01
6.26187325e-01 1.29940081e+00 3.34784240e-01 8.00694585e-01
-1.18268645e+00 -2.23428175e-01 2.41695449e-01 1.17188646e-02
-9.03287530e-01 -5.45308173e-01 2.99636811e-01 -5.43151021e-01
1.20576704e+00 5.08863270e-01 5.18624067e-01 1.09359467e+00
-1.52542263e-01 6.85609698e-01 1.13267684e+00 -5.40454566e-01
2.58944243e-01 6.12801969e-01 4.89342302e-01 4.56052423e-01
-2.79981017e-01 -4.73822542e-02 -2.81542689e-01 -4.30586100e-01
1.98949352e-01 -2.00983569e-01 -5.33137560e-01 -8.54586251e-03
-8.87080252e-01 8.21867108e-01 -2.62860119e-01 2.97908545e-01
-4.13323760e-01 -5.54959595e-01 6.05206788e-02 7.30984688e-01
2.92562932e-01 1.03928316e+00 -4.79559481e-01 -9.04747844e-01
-4.69331414e-01 9.06863451e-01 1.64648449e+00 1.02439940e+00
4.75607127e-01 -5.34733772e-01 -4.30579156e-01 1.27516055e+00
2.08207846e-01 3.64156872e-01 5.84148049e-01 -1.13540506e+00
1.40093982e-01 6.41201854e-01 7.70820618e-01 -7.20024407e-01
-6.14205897e-01 -1.56167701e-01 -1.44544527e-01 -1.20232515e-02
6.36348903e-01 -4.70647693e-01 -4.27217603e-01 1.74548280e+00
9.44179073e-02 -7.97580928e-02 4.28416342e-01 7.17380941e-01
8.14264178e-01 7.07912505e-01 -2.11666673e-01 -3.65176618e-01
1.32480037e+00 -1.09432101e+00 -6.41373634e-01 8.22395459e-03
9.16728377e-01 -8.33881795e-01 1.33090293e+00 5.12102723e-01
-1.20412409e+00 -5.24660766e-01 -8.17888618e-01 3.02377015e-01
-1.39063343e-01 -1.17969796e-01 2.63444304e-01 7.92918026e-01
-1.03125167e+00 6.52848363e-01 -4.48655874e-01 -8.06092024e-01
-6.93209052e-01 1.94100171e-01 1.44024238e-01 9.50805917e-02
-1.36533892e+00 1.08502817e+00 -1.42286912e-01 -3.73052388e-01
-7.75829077e-01 -4.63679671e-01 -7.97712147e-01 1.09284565e-01
2.45490819e-01 -7.65933931e-01 2.20513725e+00 -7.61881053e-01
-2.17390966e+00 4.74172622e-01 -1.78147823e-01 -3.94425571e-01
2.88443029e-01 -8.77415687e-02 -3.47644478e-01 -3.36649179e-01
-1.66670755e-01 3.20744693e-01 3.41828734e-01 -1.31786656e+00
-9.65875566e-01 4.41082194e-02 9.98651087e-01 8.34418654e-01
4.69182208e-02 -2.48160269e-02 -2.63969421e-01 -2.13483930e-01
-3.75522703e-01 -1.17709541e+00 -2.07875594e-01 -9.75488603e-01
-2.83856988e-01 -5.82304835e-01 1.73583433e-01 -4.72176492e-01
1.39722025e+00 -1.73141253e+00 3.42134804e-01 2.04749689e-01
2.62297928e-01 3.28593314e-01 -3.43474925e-01 8.70714486e-01
2.59700447e-01 3.43592837e-03 -2.05815607e-03 -5.17420411e-01
2.89740920e-01 -2.26564556e-02 -1.06769502e-01 -1.15330450e-01
-4.23633665e-01 4.39190507e-01 -1.19307435e+00 1.10562861e-01
2.28433982e-01 7.93961734e-02 -9.45308506e-01 1.00337756e+00
-5.26499867e-01 2.97589600e-01 -2.29502171e-01 -5.18092141e-02
3.80682766e-01 -1.60291910e-01 3.52511197e-01 -4.36886307e-03
-2.30305061e-01 9.19677973e-01 -9.51514363e-01 1.36362398e+00
-9.06617522e-01 3.44547749e-01 7.35504495e-04 -5.40435612e-01
6.57102227e-01 3.00169557e-01 7.10906759e-02 -5.12372255e-01
2.05298603e-01 -2.00371981e-01 4.79233593e-01 -7.46432662e-01
8.10360134e-01 1.35546982e-01 -1.80802450e-01 1.23896253e+00
-7.32625127e-02 -3.02389711e-01 5.03876030e-01 5.07496476e-01
1.19719934e+00 -6.66424558e-02 4.35675889e-01 -7.61614293e-02
6.16577983e-01 -1.37079293e-02 -1.69515759e-01 1.23523653e+00
-4.41147760e-02 4.44079876e-01 4.49163228e-01 -3.95987183e-02
-7.25992739e-01 -7.33225405e-01 2.99398899e-01 1.42066324e+00
1.38379112e-01 -6.59473181e-01 -1.29311669e+00 -6.00976408e-01
-4.66588527e-01 1.28624570e+00 -3.73869479e-01 -1.40667766e-01
-3.50258768e-01 -4.27336335e-01 4.12079513e-01 -1.48227617e-01
2.91043699e-01 -1.27007127e+00 -6.27816081e-01 4.49162215e-01
-7.74288177e-01 -8.46034110e-01 -6.41808748e-01 -2.84472376e-01
-2.21351698e-01 -8.75986159e-01 -4.03116584e-01 -5.12679756e-01
4.88178015e-01 5.38308561e-01 1.36528027e+00 5.71304739e-01
2.80726224e-01 6.91154659e-01 -7.58902848e-01 -4.56779189e-02
-1.05721152e+00 1.50553018e-01 1.01723790e-01 -3.79359156e-01
1.76809639e-01 -4.91854578e-01 -4.60514635e-01 7.70055592e-01
-6.12450063e-01 4.63614166e-01 1.86379716e-01 7.81589568e-01
-4.01646197e-01 -9.06037688e-02 5.62521458e-01 -1.30807531e+00
1.71970332e+00 -5.75521290e-01 -1.72814101e-01 4.26887661e-01
-6.12329662e-01 -5.48353605e-02 8.74308646e-01 -4.23851430e-01
-1.31134117e+00 -2.96137512e-01 -5.52106559e-01 4.75624293e-01
-5.78605473e-01 6.36830211e-01 -2.29024198e-02 3.83655876e-01
9.49459910e-01 3.23797345e-01 1.32990450e-01 -2.59215057e-01
4.77544248e-01 1.04708338e+00 2.29081169e-01 -7.86247432e-01
4.52704132e-01 -3.63477618e-01 -1.20247936e+00 -9.65512037e-01
-6.30163789e-01 -4.87422287e-01 -8.10524672e-02 -4.46078211e-01
4.69605476e-01 -5.04642963e-01 -1.11601543e+00 3.70540291e-01
-1.19379616e+00 -8.36363912e-01 -3.45551781e-02 2.96058536e-01
-7.32824624e-01 4.51687992e-01 -6.38370097e-01 -9.82225955e-01
-2.26541281e-01 -1.42661476e+00 6.55502141e-01 4.19450790e-01
-1.00686026e+00 -8.43421459e-01 4.90038365e-01 5.91644883e-01
5.45184016e-01 -6.81207061e-01 8.82938445e-01 -1.22103548e+00
-1.11285537e-01 7.11648241e-02 8.98353532e-02 -1.09634317e-01
2.18611166e-01 -5.42479083e-02 -7.21154213e-01 -3.00024480e-01
4.15540393e-03 -2.57115006e-01 -1.86682373e-01 6.16904758e-02
6.75436854e-01 -5.91397464e-01 -1.14916496e-01 8.57846886e-02
4.88575727e-01 5.64036846e-01 5.74051678e-01 1.17785767e-01
8.03064108e-02 8.52062166e-01 6.78171694e-01 5.31066835e-01
9.40705478e-01 9.58671570e-01 -1.69186413e-01 2.86499798e-01
5.83432019e-02 -3.51227701e-01 5.31299829e-01 7.56453335e-01
-1.81017444e-01 -9.15308714e-01 -6.14882827e-01 3.17760497e-01
-1.84624171e+00 -1.09153748e+00 6.19216412e-02 2.15319371e+00
8.25315595e-01 1.79750592e-01 5.51514745e-01 -1.98796734e-01
5.30492306e-01 -2.46334195e-01 -1.92388207e-01 -7.65793681e-01
5.55848062e-01 7.30428323e-02 -2.03556791e-01 1.09283447e+00
-4.67199415e-01 9.40298855e-01 6.54916716e+00 9.62290466e-02
-7.26362407e-01 -8.95957053e-02 2.36380398e-01 3.20379406e-01
-4.53818142e-01 -1.29687116e-01 -7.67113149e-01 5.43242693e-01
1.30491805e+00 -5.12367189e-01 9.56726491e-01 6.00919306e-01
5.50597668e-01 -1.85643300e-01 -1.33052695e+00 4.29163277e-01
2.74136722e-01 -1.00044978e+00 5.83637431e-02 -2.76454329e-01
3.80117118e-01 -2.80966729e-01 -2.73108870e-01 9.56677079e-01
9.83215153e-01 -8.81259441e-01 2.28932261e-01 3.96469563e-01
1.29340410e-01 -4.61773813e-01 6.77088201e-01 8.73406589e-01
-6.99697614e-01 7.51081854e-02 -2.73059886e-02 -5.78752577e-01
3.13366503e-01 -1.13588244e-01 -1.53664792e+00 3.91274542e-01
2.84815490e-01 2.76808053e-01 -3.43918651e-01 8.14853728e-01
-1.40030012e-01 5.94808519e-01 -1.71260670e-01 -5.22016644e-01
-1.57107338e-01 -2.77743071e-01 5.99817872e-01 1.30792832e+00
3.49332213e-01 5.98230422e-01 4.57799345e-01 7.09391057e-01
1.58375278e-01 2.15105355e-01 -5.34915090e-01 -2.80319955e-02
7.42116630e-01 1.24810147e+00 -4.62599784e-01 -5.12453735e-01
-4.21535492e-01 1.15718400e+00 3.08676988e-01 3.88926238e-01
-7.00870037e-01 -2.22418845e-01 7.83945918e-01 9.61961374e-02
-1.20147094e-01 1.17453746e-02 2.17392012e-01 -1.14434636e+00
-4.31297898e-01 -1.59109688e+00 7.11076800e-03 -8.94298375e-01
-1.03385317e+00 9.69308794e-01 1.26820058e-01 -1.01082289e+00
-9.41417634e-01 1.39295319e-02 -8.16529036e-01 8.69261146e-01
-8.39611471e-01 -6.36027992e-01 -5.38425148e-01 3.98038954e-01
1.11076391e+00 -1.42612249e-01 1.09044194e+00 2.80848920e-01
-4.93005186e-01 7.27563739e-01 -3.21262211e-01 -3.52782220e-01
6.82389498e-01 -1.34048867e+00 5.73730767e-01 5.51974058e-01
5.05994521e-02 1.14396584e+00 1.48201990e+00 -5.79933822e-01
-1.04219043e+00 -7.02890098e-01 1.00288701e+00 -3.92239153e-01
4.79901344e-01 -4.05642658e-01 -9.70418751e-01 5.77896416e-01
5.89966953e-01 -1.05623126e+00 7.88422763e-01 3.75915289e-01
1.53587982e-01 4.82743055e-01 -1.07310069e+00 1.01108289e+00
1.06029212e+00 -4.51352775e-01 -8.66682649e-01 4.70744431e-01
7.11712778e-01 -5.82391202e-01 -6.00089610e-01 1.44263599e-02
5.95626533e-01 -1.11854625e+00 6.56120479e-01 -5.84640324e-01
3.66116941e-01 -2.73779064e-01 1.34055182e-01 -2.05872059e+00
-3.12331766e-01 -1.11018157e+00 1.17652565e-01 1.24722600e+00
6.63777888e-01 -5.46991825e-01 5.11371791e-01 8.51111412e-01
-1.66019067e-01 -2.47678593e-01 -9.90873501e-02 -3.04561645e-01
-2.97933429e-01 -1.92705050e-01 6.29288375e-01 5.74520051e-01
9.22554195e-01 7.96311498e-01 -5.48398852e-01 1.36002317e-01
1.77047804e-01 6.41337037e-02 1.33637726e+00 -1.05147588e+00
-6.13377512e-01 -4.24068868e-01 2.97679722e-01 -1.78862774e+00
5.51549941e-02 -6.12070560e-01 6.18633687e-01 -1.42281985e+00
1.29244417e-01 -3.84267926e-01 2.97829539e-01 5.92536740e-02
-2.21721485e-01 -2.12098390e-01 1.68205455e-01 7.20251575e-02
-6.31054401e-01 2.71166593e-01 1.02312529e+00 1.76176965e-01
-6.99787557e-01 5.67757726e-01 -8.22493792e-01 7.01544583e-01
7.50898421e-01 -1.07138604e-01 -8.63813281e-01 3.79269496e-02
1.91737235e-01 5.74196696e-01 -2.66008347e-01 -7.81096041e-01
4.02619451e-01 -1.95426896e-01 -4.92954403e-01 -4.15644616e-01
1.87436327e-01 -5.15594363e-01 2.09333971e-01 2.50818223e-01
-8.86893749e-01 3.40871602e-01 -8.93878657e-03 2.31381878e-01
1.32919043e-01 -6.30010545e-01 4.61787283e-01 3.09298187e-02
-2.97910899e-01 -1.74652755e-01 -1.29554570e+00 -9.51540247e-02
9.18905318e-01 -8.36151242e-02 -3.45726818e-01 -1.10329962e+00
-8.54628384e-01 4.03375328e-01 4.72629905e-01 6.68250978e-01
5.38568497e-01 -6.98508084e-01 -7.08177626e-01 1.66693032e-01
3.05240065e-01 -5.52503169e-01 2.08644703e-01 3.86802107e-01
-3.83311212e-01 3.43517303e-01 -3.41001619e-03 -3.62879574e-01
-1.59515190e+00 2.06555709e-01 5.65988600e-01 -3.80714774e-01
-4.32150781e-01 5.54591417e-01 2.55561352e-01 -1.02427614e+00
3.43769670e-01 -4.13458765e-01 -8.72089148e-01 -3.92957814e-02
7.78609335e-01 1.40024766e-01 1.86397061e-01 -3.10024291e-01
6.94733411e-02 -8.79371688e-02 -4.81173456e-01 -3.60863954e-01
9.13853109e-01 -5.74508190e-01 1.94332778e-01 3.53946120e-01
7.30128527e-01 8.48698989e-02 -8.05492401e-01 -1.48039132e-01
-1.41536191e-01 -4.88059193e-01 -5.34143150e-01 -1.08419240e+00
-1.99529707e-01 2.09132463e-01 2.69515961e-01 9.15203750e-01
5.95112264e-01 -2.57835686e-02 7.39002287e-01 8.76410782e-01
4.67971444e-01 -8.35712373e-01 -3.27182673e-02 8.89956832e-01
9.90609705e-01 -1.02416122e+00 -3.31910789e-01 -4.58350658e-01
-7.60516822e-01 8.08415592e-01 7.99650609e-01 1.57249197e-01
4.38086331e-01 1.18795723e-01 2.30482623e-01 -1.85051531e-01
-1.25717068e+00 -2.14466482e-01 -2.73351818e-02 6.52828872e-01
8.81044567e-01 2.12514877e-01 -6.95787311e-01 5.48116505e-01
-7.93594778e-01 2.29536518e-02 1.11549973e+00 7.05460012e-01
-4.52698797e-01 -1.33767962e+00 -6.34289160e-02 5.36602497e-01
-5.33576421e-02 -3.77149321e-02 -5.83610117e-01 3.67439300e-01
-3.68201226e-01 1.65835202e+00 -1.30916417e-01 -6.28364563e-01
8.44432771e-01 -1.75238233e-02 3.78950447e-01 -1.06209600e+00
-1.06269038e+00 -3.09285641e-01 7.82258451e-01 -2.14516789e-01
-3.23309422e-01 -8.40045035e-01 -1.05222321e+00 -4.64025408e-01
-4.23819035e-01 6.93629563e-01 4.01351035e-01 1.17866063e+00
2.34448805e-01 5.16713381e-01 7.88322747e-01 -6.68141186e-01
-6.20386839e-01 -1.30939090e+00 -6.94862604e-02 7.96089590e-01
1.76194906e-01 -5.77351511e-01 -4.96894032e-01 6.58019781e-02] | [12.600794792175293, 7.8681182861328125] |
70ba3fa4-aa67-4b2d-b974-b39dc1c6d0db | codesearchnet-challenge-evaluating-the-state | 1909.09436 | null | https://arxiv.org/abs/1909.09436v3 | https://arxiv.org/pdf/1909.09436v3.pdf | CodeSearchNet Challenge: Evaluating the State of Semantic Code Search | Semantic code search is the task of retrieving relevant code given a natural language query. While related to other information retrieval tasks, it requires bridging the gap between the language used in code (often abbreviated and highly technical) and natural language more suitable to describe vague concepts and ideas. To enable evaluation of progress on code search, we are releasing the CodeSearchNet Corpus and are presenting the CodeSearchNet Challenge, which consists of 99 natural language queries with about 4k expert relevance annotations of likely results from CodeSearchNet Corpus. The corpus contains about 6 million functions from open-source code spanning six programming languages (Go, Java, JavaScript, PHP, Python, and Ruby). The CodeSearchNet Corpus also contains automatically generated query-like natural language for 2 million functions, obtained from mechanically scraping and preprocessing associated function documentation. In this article, we describe the methodology used to obtain the corpus and expert labels, as well as a number of simple baseline solutions for the task. We hope that CodeSearchNet Challenge encourages researchers and practitioners to study this interesting task further and will host a competition and leaderboard to track the progress on the challenge. We are also keen on extending CodeSearchNet Challenge to more queries and programming languages in the future. | ['Miltiadis Allamanis', 'Ho-Hsiang Wu', 'Tiferet Gazit', 'Marc Brockschmidt', 'Hamel Husain'] | 2019-09-20 | null | null | null | null | ['code-search', 'code-search'] | ['computer-code', 'computer-vision'] | [-1.68871641e-01 1.85223326e-01 -3.09526622e-01 -3.30720156e-01
-1.04241538e+00 -1.07440710e+00 5.07758558e-01 4.97777611e-01
-2.22987548e-01 3.38674575e-01 6.12882018e-01 -6.52113616e-01
-1.43248841e-01 -2.97096372e-01 -4.37225819e-01 1.01652361e-01
-1.67915821e-01 2.31190622e-01 5.38188875e-01 -2.52185911e-01
8.68362606e-01 -1.28086209e-01 -1.58445573e+00 4.86817271e-01
1.06536317e+00 4.97038543e-01 6.85540080e-01 7.12458849e-01
-5.83233654e-01 1.25566506e+00 -5.34065783e-01 -5.22016108e-01
-2.10961789e-01 -1.77308708e-01 -1.64471412e+00 -8.00729811e-01
3.04485440e-01 2.36815572e-01 6.97914436e-02 1.18598473e+00
2.46136859e-01 -7.60602728e-02 2.47436762e-01 -1.30780053e+00
-7.91973412e-01 8.39535415e-01 -9.44262967e-02 3.15477014e-01
1.02271223e+00 -1.40877783e-01 1.39520252e+00 -9.70041215e-01
9.78325903e-01 1.03642654e+00 7.51631916e-01 6.01512492e-01
-8.43549907e-01 -3.48275691e-01 -2.05177814e-01 6.25881599e-03
-1.34997988e+00 -3.27709705e-01 1.80990323e-01 -9.33691025e-01
1.75716686e+00 5.20694613e-01 2.26750284e-01 7.86936104e-01
-1.67047139e-02 7.46738613e-01 2.37179592e-01 -6.50503933e-01
8.51499438e-02 3.02074343e-01 4.16451454e-01 9.75503027e-01
9.18486044e-02 -4.15729463e-01 -3.62018906e-02 -1.16568625e+00
1.46435991e-01 -4.69226539e-02 -4.85682696e-01 -3.24368477e-01
-1.09004819e+00 8.10830176e-01 2.50402510e-01 6.65644765e-01
1.38568282e-01 4.53123689e-01 9.43998992e-01 4.04599607e-01
5.89659847e-02 1.16387188e+00 -8.12401474e-01 -7.30948269e-01
-6.81603432e-01 3.96893859e-01 1.51031494e+00 1.70044196e+00
1.01410103e+00 -5.60695648e-01 -3.01488731e-02 1.18059993e+00
3.18463504e-01 1.75133616e-01 7.09608734e-01 -1.28857946e+00
7.34683573e-01 9.77162480e-01 1.10329874e-01 -8.70564520e-01
-8.30004215e-02 -1.13964468e-01 2.22111121e-01 -1.13936193e-01
-1.52335165e-03 2.91090250e-01 -3.06417733e-01 1.18225861e+00
-3.78385007e-01 -7.20016181e-01 -1.72319878e-02 5.33560991e-01
9.63020742e-01 5.33654153e-01 5.85862249e-02 2.42029399e-01
1.48416674e+00 -1.18057489e+00 -2.47322068e-01 -4.36984599e-01
1.02577448e+00 -1.20097864e+00 1.46726251e+00 2.08634883e-01
-9.50497389e-01 -2.20447242e-01 -6.94074571e-01 -3.23861480e-01
-6.65183663e-01 8.12835842e-02 6.80593789e-01 4.14968371e-01
-1.31368530e+00 2.91341633e-01 -6.02966309e-01 -9.56764340e-01
9.92091745e-02 -8.33864734e-02 -5.69932200e-02 -2.08280697e-01
-9.10233915e-01 8.47836733e-01 3.79685670e-01 -6.34427369e-01
-7.83315122e-01 -6.55103981e-01 -1.00741112e+00 2.83797145e-01
4.64623481e-01 -2.86727697e-01 1.74320185e+00 -7.38620102e-01
-5.00225365e-01 1.16725957e+00 -8.68353620e-02 6.56464249e-02
-7.02928603e-02 -1.12610742e-01 -4.07298952e-01 -2.25434406e-03
7.74498641e-01 3.31807435e-01 1.80217996e-01 -9.84387696e-01
-7.67837584e-01 1.35109544e-01 4.49334264e-01 -9.83812809e-02
-3.84338140e-01 8.34753931e-01 -9.34061587e-01 -6.83965325e-01
-3.37858081e-01 -8.41431379e-01 -1.30938649e-01 -9.94567946e-02
-4.95246276e-02 -7.45431185e-01 3.59577596e-01 -7.34243870e-01
1.62621140e+00 -2.34077621e+00 -1.08056469e-02 4.12533106e-03
3.33710432e-01 -8.84772763e-02 -2.18989804e-01 9.27206874e-01
-9.53239724e-02 5.24993539e-01 -3.96431357e-01 3.98858577e-01
3.95422399e-01 5.03419153e-02 -5.36663115e-01 2.70077828e-02
6.16440130e-03 9.96152878e-01 -1.35581100e+00 -6.08283401e-01
-4.60638195e-01 -2.30489345e-03 -8.14678848e-01 4.28044610e-02
-5.04778087e-01 -3.24770272e-01 -8.09184074e-01 9.41789448e-01
1.26062766e-01 -5.98957241e-01 -1.26277119e-01 4.56852585e-01
-2.32635975e-01 6.87357664e-01 -6.16552711e-01 1.97721958e+00
-7.13272870e-01 8.91534567e-01 2.40802318e-01 -5.89040697e-01
8.25787723e-01 4.17290777e-01 3.41090947e-01 -4.78312075e-01
-4.63017613e-01 5.58333337e-01 -5.58029413e-01 -1.07505238e+00
6.55277610e-01 5.05675495e-01 -6.41473651e-01 7.04007268e-01
4.13885824e-02 -4.91066962e-01 5.27009726e-01 4.99053746e-01
1.72858000e+00 2.34005049e-01 4.41641599e-01 -8.55905950e-01
7.11726129e-01 6.33980155e-01 1.91932887e-01 9.05922472e-01
7.10815541e-04 2.71355301e-01 7.01678216e-01 -5.98741293e-01
-8.03943455e-01 -5.79105020e-01 -1.14854760e-01 1.67406189e+00
-2.77216099e-02 -1.27154410e+00 -6.10002577e-01 -1.05346680e+00
3.55199464e-02 7.04551280e-01 -2.57473826e-01 8.02521333e-02
-7.42013514e-01 -1.13596745e-01 9.35288370e-01 5.00316024e-01
4.74618599e-02 -1.20566678e+00 -6.26803041e-01 2.43250445e-01
-4.56458718e-01 -7.10741222e-01 -8.85294080e-01 2.27763861e-01
-4.97712016e-01 -1.56306732e+00 -5.09295881e-01 -1.23386204e+00
6.36906385e-01 8.03767964e-02 1.89558446e+00 9.32251096e-01
-5.92581272e-01 8.49738598e-01 -7.72524893e-01 -7.02151135e-02
-8.09436142e-01 3.10824811e-01 -7.90226102e-01 -1.21278858e+00
6.97713375e-01 -1.92043021e-01 -2.64721334e-01 2.79747039e-01
-1.08974779e+00 -2.75125027e-01 2.47027874e-01 5.97154558e-01
-1.86755031e-01 -7.02413470e-02 2.36775264e-01 -9.79312539e-01
1.00091827e+00 -8.69364202e-01 -7.48050332e-01 5.84534228e-01
-8.56860757e-01 3.65041167e-01 5.56353688e-01 1.24738540e-03
-8.72241259e-01 -2.66878977e-02 -3.44887525e-01 3.19515020e-01
8.23993757e-02 9.22579944e-01 3.59922498e-01 -2.91403741e-01
1.34248817e+00 2.43941963e-01 -5.05533159e-01 -7.61593580e-01
3.03623289e-01 1.04239786e+00 4.90413636e-01 -1.15995431e+00
5.32803118e-01 6.85443729e-02 -8.70495975e-01 -4.18112844e-01
-2.43798584e-01 -8.97374690e-01 -2.65587866e-02 1.81692898e-01
5.58132291e-01 -7.53987253e-01 -4.64247286e-01 -2.53630310e-01
-1.40647757e+00 -2.08984077e-01 -4.90050167e-02 -2.23269314e-02
-4.97632891e-01 6.01919711e-01 -7.45274305e-01 -3.17336977e-01
-3.51825088e-01 -1.36404955e+00 1.24918926e+00 -1.17811628e-01
-6.97399974e-01 -9.97102261e-01 4.00443554e-01 2.11635679e-01
8.04900527e-01 -1.59255743e-01 1.36620617e+00 -7.69913256e-01
-6.00669801e-01 -3.30199182e-01 -3.15216184e-01 2.79868096e-02
-2.88994927e-02 4.10416834e-02 -6.51843607e-01 -2.73098618e-01
-1.16265208e-01 -6.48293734e-01 6.19840622e-01 -4.53689396e-01
1.22338665e+00 -5.37907958e-01 -6.62729144e-01 1.97184339e-01
1.78796911e+00 3.36617768e-01 4.05473202e-01 4.84828681e-01
4.00271952e-01 6.14057720e-01 3.81767273e-01 4.03670102e-01
6.31138682e-01 4.61537838e-01 3.49200577e-01 3.64379734e-01
1.12498343e-01 -1.71864152e-01 2.32035637e-01 8.96899819e-01
3.87854129e-01 3.83825935e-02 -1.67803097e+00 8.82873833e-01
-1.78725684e+00 -6.58231318e-01 -8.24353397e-02 1.91860008e+00
1.17288792e+00 -2.09328428e-01 -1.64645568e-01 -5.22891879e-01
5.35821080e-01 -2.22879231e-01 -3.64668339e-01 -3.27031732e-01
3.09069067e-01 5.96183091e-02 2.75869250e-01 6.05894208e-01
-8.07585776e-01 7.83569694e-01 7.04265118e+00 8.24325442e-01
-6.29927337e-01 8.90982449e-02 -7.35332724e-03 5.05538762e-01
-6.48793876e-01 5.85232019e-01 -7.36954212e-01 4.66928035e-01
8.30632091e-01 -7.54023492e-01 9.84728575e-01 1.52832115e+00
-4.46901590e-01 -1.68423250e-01 -1.31898749e+00 9.05238986e-01
1.56680003e-01 -1.42262232e+00 -3.89910519e-01 -3.93262923e-01
6.10899270e-01 6.11740410e-01 -6.34272337e-01 8.20406973e-01
7.41926491e-01 -8.48395228e-01 7.27342844e-01 3.92063946e-01
7.80465901e-01 -4.41601537e-02 5.10968745e-01 3.40861678e-01
-1.37965763e+00 -3.79375547e-01 -2.88893521e-01 -3.75131853e-02
-5.62934935e-01 1.64423510e-01 -7.78690577e-01 2.65175641e-01
1.09471810e+00 8.29403520e-01 -1.02095568e+00 1.23359704e+00
-3.74737270e-02 2.90761799e-01 7.64322877e-02 -4.64926153e-01
1.68419749e-01 4.15826648e-01 3.69261146e-01 1.80256665e+00
2.30464280e-01 -2.85926431e-01 3.08271945e-01 1.26755071e+00
-9.01717395e-02 2.24733397e-01 -7.31797338e-01 -4.72068220e-01
5.94740152e-01 1.15464258e+00 -7.99497128e-01 -5.66622972e-01
-9.62101221e-01 8.31252515e-01 2.20640421e-01 5.19985080e-01
-4.99147862e-01 -1.18517935e+00 5.40793598e-01 -1.05499178e-01
1.52573615e-01 3.59389745e-02 1.92604467e-01 -1.24363410e+00
4.12329495e-01 -1.22903860e+00 8.01149607e-01 -1.08307028e+00
-1.33404160e+00 7.92079568e-01 1.50807574e-01 -9.94576633e-01
-5.04472971e-01 -5.30814230e-01 -3.69130760e-01 9.13483858e-01
-1.25628185e+00 -3.74289632e-01 -3.87227803e-01 1.91644281e-01
6.88270211e-01 -2.40331382e-01 9.54585850e-01 5.24691224e-01
1.36191279e-01 2.33691514e-01 1.15731150e-01 3.57915372e-01
7.64151156e-01 -1.30601609e+00 8.33684444e-01 4.76926982e-01
-1.08723976e-01 1.41547906e+00 6.75240159e-01 -5.99199295e-01
-1.39181519e+00 -8.14239860e-01 1.37373579e+00 -9.67588902e-01
9.09384549e-01 -4.64743525e-01 -9.05369401e-01 6.36393964e-01
4.71070468e-01 2.98846932e-03 4.98497993e-01 -7.91046098e-02
-6.82121754e-01 2.50669688e-01 -9.30964231e-01 3.98478687e-01
9.81277049e-01 -1.11932850e+00 -9.46048498e-01 6.43263400e-01
9.88097131e-01 -3.42952639e-01 -8.40161264e-01 -1.64943617e-02
3.13738704e-01 -6.88264370e-01 7.19746411e-01 -4.07333136e-01
4.99631196e-01 -3.24060470e-01 -2.57943243e-01 -8.88409078e-01
-2.49647975e-01 -8.09518635e-01 4.21873093e-01 1.03893685e+00
6.40100598e-01 -4.49273795e-01 4.48040724e-01 9.06181097e-01
-3.58376116e-01 -4.58772033e-01 -6.19233727e-01 -8.05154562e-01
9.79426205e-02 -5.12977123e-01 5.39551556e-01 1.05592251e+00
9.36586261e-01 2.93445528e-01 3.81423026e-01 -3.27229649e-01
8.41533244e-02 2.50670195e-01 5.53583145e-01 -1.18682194e+00
-4.70229298e-01 -8.60338748e-01 -1.99184492e-01 -1.12793446e+00
2.67159253e-01 -1.49908590e+00 2.61520982e-01 -1.60223556e+00
5.59732080e-01 -3.80708814e-01 2.02974021e-01 7.89003968e-01
-1.04425408e-01 -2.86257118e-01 -3.41709852e-02 5.24611413e-01
-1.01417327e+00 3.71707343e-02 6.09986663e-01 -2.87328869e-01
1.62273124e-01 -1.73997819e-01 -1.07222950e+00 4.84684855e-01
2.95328110e-01 -7.78750241e-01 -4.76427078e-01 -7.20535994e-01
9.57653463e-01 2.23202363e-01 2.85475731e-01 -8.17440629e-01
5.69248915e-01 -6.86167255e-02 -4.44450408e-01 -1.57727614e-01
-3.63655984e-01 -9.31307316e-01 -3.25815305e-02 4.62788314e-01
-6.98957622e-01 4.86964852e-01 2.68630683e-01 2.16713995e-01
-3.23799938e-01 -1.06562448e+00 2.07070529e-01 -8.00024390e-01
-9.52414155e-01 -1.60735309e-01 -5.63372672e-01 7.85219789e-01
5.57325959e-01 2.22338941e-02 -7.28941381e-01 -2.29990676e-01
-1.82239398e-01 5.05558074e-01 9.33480620e-01 9.18510914e-01
4.29687887e-01 -1.13548744e+00 -3.83909881e-01 6.41656816e-02
8.21107090e-01 -5.67500949e-01 -6.02267921e-01 4.17125791e-01
-8.99310410e-01 8.50497484e-01 1.94718733e-01 -2.76019156e-01
-1.02681315e+00 6.39703155e-01 1.36748344e-01 -6.67272927e-03
-4.49021995e-01 7.95389652e-01 -5.77519722e-02 -8.23991120e-01
5.10457337e-01 -5.68076611e-01 -1.15268119e-01 -5.03851533e-01
6.57749772e-01 3.38287115e-01 1.97811231e-01 -1.62412673e-01
-7.32831776e-01 3.38618010e-01 1.84316278e-01 9.32021290e-02
1.18324018e+00 1.07463226e-01 -9.83441412e-01 2.56087124e-01
1.57933378e+00 3.26605588e-01 -5.63848376e-01 -2.45322496e-01
1.04710400e+00 -5.13914824e-01 -4.77895141e-01 -1.00565398e+00
-5.67685306e-01 4.67517167e-01 1.55160248e-01 5.87354779e-01
8.73255908e-01 6.51808321e-01 4.55121398e-01 8.88315618e-01
6.87423527e-01 -9.84391332e-01 4.51644361e-02 8.45365286e-01
1.13259959e+00 -9.52747881e-01 -1.02594934e-01 -2.36075997e-01
-2.47603506e-01 1.34019816e+00 4.81878906e-01 2.39391789e-01
5.86678922e-01 4.12852675e-01 1.00765720e-01 -7.61467457e-01
-8.88322055e-01 -4.04357687e-02 1.67222947e-01 3.74121457e-01
1.03750718e+00 -4.55950677e-01 -3.78181845e-01 5.90629220e-01
-2.59976368e-02 -6.03834778e-05 5.63783169e-01 1.46951866e+00
-5.41903496e-01 -1.35672665e+00 -2.06803918e-01 5.79210222e-01
-5.94318926e-01 -6.54600024e-01 -6.53493106e-01 6.38014853e-01
-1.85242772e-01 9.93961990e-01 -3.84904057e-01 -1.16713338e-01
3.86932790e-01 2.03856781e-01 2.60326248e-02 -1.21187270e+00
-9.15908694e-01 -3.49225163e-01 2.65148878e-01 -8.54958534e-01
7.56165758e-03 -4.13988918e-01 -1.45390511e+00 2.42218655e-02
-3.63019705e-01 8.98900867e-01 6.79765761e-01 5.55155516e-01
5.21206617e-01 -5.17737167e-03 1.04334019e-01 -2.11402535e-01
-4.86271828e-01 -7.38908827e-01 -3.78984027e-02 3.22653681e-01
2.37006932e-01 -2.14447349e-01 -4.10210997e-01 3.14209700e-01] | [7.542995929718018, 8.072410583496094] |
b33d31d7-7d26-4b88-a4b6-0f4e52479f5a | little-ball-of-fur-a-python-library-for-graph | 2006.04311 | null | https://arxiv.org/abs/2006.04311v2 | https://arxiv.org/pdf/2006.04311v2.pdf | Little Ball of Fur: A Python Library for Graph Sampling | Sampling graphs is an important task in data mining. In this paper, we describe Little Ball of Fur a Python library that includes more than twenty graph sampling algorithms. Our goal is to make node, edge, and exploration-based network sampling techniques accessible to a large number of professionals, researchers, and students in a single streamlined framework. We created this framework with a focus on a coherent application public interface which has a convenient design, generic input data requirements, and reasonable baseline settings of algorithms. Here we overview these design foundations of the framework in detail with illustrative code snippets. We show the practical usability of the library by estimating various global statistics of social networks and web graphs. Experiments demonstrate that Little Ball of Fur can speed up node and whole graph embedding techniques considerably with mildly deteriorating the predictive value of distilled features. | ['Oliver Kiss', 'Benedek Rozemberczki', 'Rik Sarkar'] | 2020-06-08 | null | null | null | cikm-2020-10 | ['graph-sampling'] | ['graphs'] | [-1.85108602e-01 3.44141304e-01 -5.09548962e-01 -2.58902282e-01
-1.64239734e-01 -5.48094630e-01 5.52967250e-01 3.51017565e-01
-1.21762976e-01 6.11545146e-01 1.48205683e-01 -6.99749589e-01
-3.22253555e-01 -1.23478639e+00 -2.16359437e-01 -1.62437320e-01
-6.32106304e-01 6.10432506e-01 3.81507814e-01 -1.74580440e-01
1.15225114e-01 3.96289766e-01 -1.42025518e+00 -3.51597019e-03
6.63809419e-01 5.84052205e-01 -1.06353760e-01 8.21733296e-01
-2.60419458e-01 5.30331254e-01 -2.55145490e-01 -7.02290952e-01
1.16188645e-01 -1.28207654e-01 -1.03819954e+00 -2.28780717e-01
5.20694435e-01 -2.90324867e-01 -4.76102740e-01 9.49856400e-01
6.12070143e-01 -1.46502391e-01 3.42863351e-01 -1.70277119e+00
-2.37730145e-01 1.19771242e+00 -4.06720728e-01 2.51181126e-01
6.88617826e-01 2.09810227e-01 1.31458986e+00 -7.74406552e-01
7.36460745e-01 1.31005585e+00 1.15660596e+00 2.11872816e-01
-1.59429717e+00 -4.40727830e-01 9.81475487e-02 -8.14151615e-02
-1.34557581e+00 -2.87301660e-01 7.58215964e-01 -4.19702441e-01
9.56373572e-01 6.11513495e-01 1.09609842e+00 9.92304623e-01
-2.35807255e-01 5.69654346e-01 6.64725602e-01 -3.32330793e-01
3.33308429e-01 4.21282738e-01 6.60706401e-01 9.73246217e-01
6.12638354e-01 -2.48735577e-01 -4.73769605e-01 -8.87449145e-01
5.10446250e-01 3.19689475e-02 6.43848032e-02 -8.19232345e-01
-1.03713608e+00 9.47602510e-01 4.95358586e-01 -1.74062885e-02
-9.54727177e-03 3.17940682e-01 6.46023333e-01 4.70925689e-01
5.69255888e-01 2.74943143e-01 -3.15625548e-01 -1.65322497e-01
-7.83948243e-01 3.81339461e-01 1.40147328e+00 1.26080942e+00
1.14165664e+00 -1.33638792e-02 6.66197389e-02 7.10874021e-01
1.60252213e-01 8.89821425e-02 -1.13291368e-01 -7.03857481e-01
2.40935177e-01 7.77316332e-01 -2.96190917e-01 -1.28713512e+00
-4.17928457e-01 -2.79588491e-01 -5.76296985e-01 -6.10207431e-02
4.32419658e-01 -2.68502504e-01 -4.33554083e-01 1.31157768e+00
6.95874691e-01 -1.28341988e-01 -6.70551658e-01 2.98196167e-01
1.14720833e+00 7.85860885e-03 1.21160492e-01 5.57291619e-02
1.32599366e+00 -8.29592705e-01 -3.91843885e-01 -2.41478339e-01
8.14997256e-01 -4.47943419e-01 1.51775622e+00 1.70217454e-02
-1.03473115e+00 -1.75479412e-01 -7.31763482e-01 -4.13612761e-02
-6.91578031e-01 -3.91634494e-01 1.30232966e+00 1.13721764e+00
-1.32560170e+00 1.10455322e+00 -7.87324846e-01 -8.83982599e-01
5.39717972e-01 2.46650487e-01 -3.87440711e-01 -1.03710651e-01
-8.94481242e-01 5.13635814e-01 2.88176715e-01 -5.74616551e-01
-3.87593299e-01 -1.04104269e+00 -9.98693526e-01 1.06179975e-01
5.65209627e-01 -8.65054905e-01 8.59025121e-01 -3.96891713e-01
-8.98607194e-01 7.12685108e-01 -4.71663810e-02 -3.00177753e-01
6.11627340e-01 1.09745301e-01 -1.19430795e-01 -6.55503618e-03
-3.88257504e-02 2.74477899e-01 4.52874333e-01 -7.69996881e-01
-3.01985778e-02 -4.45861548e-01 6.29699305e-02 4.87792753e-02
-9.36651349e-01 6.63602948e-02 -4.93556678e-01 -3.64008367e-01
-3.97669584e-01 -6.71001673e-01 -5.04835308e-01 2.55720824e-01
-4.72049266e-01 -3.75575393e-01 8.26252460e-01 -3.79608244e-01
1.59749794e+00 -1.82798338e+00 -2.80913085e-01 5.48377097e-01
7.97880828e-01 -1.40434191e-01 -1.21042036e-01 1.11069489e+00
-5.92802092e-03 6.82612419e-01 1.87827721e-02 -1.76341385e-01
1.07406572e-01 -9.86138824e-03 2.06941456e-01 6.05736017e-01
-6.84858784e-02 9.57049549e-01 -1.23607004e+00 -6.76922441e-01
3.87742639e-01 4.41667259e-01 -8.31443489e-01 1.93314161e-02
-1.49119034e-01 -2.09210247e-01 -4.04996216e-01 7.91088343e-01
6.77929819e-01 -5.72759271e-01 6.74978554e-01 -7.37500116e-02
2.43434310e-01 7.33939052e-01 -1.33725893e+00 1.39377844e+00
-2.12150827e-01 5.08616030e-01 2.27334604e-01 -5.65075994e-01
8.17424119e-01 -2.35902369e-01 4.38617259e-01 9.37085748e-02
1.60762726e-03 -6.72330260e-02 -2.26294413e-01 -2.28345796e-01
3.42781723e-01 6.44262731e-01 1.23695545e-01 7.96248198e-01
1.10849954e-01 -1.06275156e-01 5.07631302e-01 7.46416271e-01
1.63189447e+00 -1.39641806e-01 6.23345792e-01 -6.30013108e-01
-3.18236947e-02 5.13819866e-02 1.78493571e-03 8.09548736e-01
-2.15887189e-01 1.87860101e-01 9.35050189e-01 -7.56379843e-01
-1.16775131e+00 -9.66449142e-01 -4.31753621e-02 1.25228894e+00
-3.39295447e-01 -1.44345045e+00 -6.38524771e-01 -1.03483462e+00
4.11442369e-01 4.49229062e-01 -7.18949795e-01 1.81476817e-01
-2.16317907e-01 -7.06542432e-01 1.69168428e-01 2.03598335e-01
1.32551000e-01 -8.86256397e-01 -1.57491453e-02 -1.30434006e-01
2.39922896e-01 -9.21621978e-01 -5.57450473e-01 -1.20841600e-01
-9.51708317e-01 -1.14882088e+00 -1.40657365e-01 -7.58219719e-01
6.85495734e-01 5.20828545e-01 1.72530723e+00 3.10159206e-01
-6.54480457e-01 5.14343739e-01 -1.92620307e-01 -6.70330897e-02
-2.10745469e-01 6.70845509e-01 -2.69802988e-01 -5.75034559e-01
6.26038849e-01 -1.17588377e+00 -6.01538956e-01 1.48191869e-01
-5.78892231e-01 -9.39993188e-02 1.44724682e-01 4.85976577e-01
9.44065303e-02 -8.36614519e-02 2.81399101e-01 -1.33440006e+00
1.07364786e+00 -7.71723747e-01 -4.71182913e-01 -4.52043042e-02
-1.03983915e+00 -5.96070662e-02 3.95097643e-01 -2.08948359e-01
-1.42348647e-01 -1.55832954e-02 -9.75175500e-02 -5.20209707e-02
1.96462527e-01 7.34492064e-01 -5.38274460e-02 -3.76242399e-01
1.01620972e+00 -8.94695520e-02 4.09874231e-01 -4.98447031e-01
4.45622057e-01 6.53169811e-01 -1.51085883e-01 -5.37044048e-01
7.75424004e-01 3.22992802e-01 -1.21031180e-02 -1.06577635e+00
-1.70579895e-01 -4.83438313e-01 -5.40752292e-01 -3.02167743e-01
-1.10559747e-01 -7.15504289e-01 -9.07124400e-01 2.42497157e-02
-6.77658796e-01 -4.57314134e-01 -2.79234022e-01 -1.42557546e-02
-9.42555815e-02 6.37092173e-01 -5.06764591e-01 -7.23269880e-01
-5.24549842e-01 -8.30623150e-01 7.44539261e-01 -1.07012160e-01
-5.77586710e-01 -1.40340030e+00 2.48390362e-01 -7.28413835e-02
6.02438033e-01 5.28940618e-01 6.95724189e-01 -7.93953180e-01
-5.44178247e-01 -4.28692937e-01 -3.44237089e-01 2.43133921e-02
1.40980527e-01 5.44504046e-01 -7.39992857e-01 -5.12319028e-01
-8.10123444e-01 -2.20736250e-01 7.13198483e-01 1.07579693e-01
1.24529111e+00 -6.22067451e-01 -7.12897003e-01 7.15883195e-01
1.52957511e+00 -7.86249936e-01 4.48894203e-01 8.20478424e-02
8.50295365e-01 5.69497466e-01 1.62874714e-01 6.43422127e-01
5.70696294e-01 2.68460333e-01 3.79923701e-01 1.53056281e-02
-1.45674944e-01 -3.67506623e-01 2.78253615e-01 7.19861031e-01
1.17027648e-01 1.17426708e-01 -1.09272242e+00 8.39383602e-01
-1.62732518e+00 -8.96577179e-01 -4.24273103e-01 2.19883347e+00
9.41585124e-01 9.73631144e-02 8.98356259e-01 1.12787448e-03
7.98972189e-01 2.09290624e-01 -3.48975807e-01 -4.30734426e-01
4.62258160e-01 2.32438639e-01 5.78139246e-01 6.68534219e-01
-8.65276396e-01 7.75443971e-01 7.64646626e+00 7.57390201e-01
-5.38575053e-01 -3.10619563e-01 3.74845207e-01 7.33545935e-03
-6.94868803e-01 3.35010707e-01 -8.38816047e-01 2.90103078e-01
1.14338315e+00 -6.73256814e-01 8.23716700e-01 1.45473540e+00
9.03925225e-02 2.60683775e-01 -1.28445077e+00 7.11867273e-01
-3.36883038e-01 -1.60506415e+00 -1.95119325e-02 2.32467189e-01
2.59895712e-01 3.77480477e-01 -3.45423728e-01 3.28902662e-01
9.27597165e-01 -1.08384717e+00 1.94217879e-02 1.40719727e-01
9.68751073e-01 -6.33644700e-01 3.15320998e-01 -1.83817353e-02
-1.29529536e+00 2.02489778e-01 -3.53048533e-01 -4.22717303e-01
-2.30226532e-01 1.04910851e+00 -1.26996148e+00 4.08206701e-01
9.16494489e-01 9.44302797e-01 -8.86017025e-01 1.20531070e+00
9.87898111e-02 7.10988104e-01 -7.34381199e-01 -6.85169399e-01
-2.47364432e-01 -1.19619131e-01 5.76446533e-01 1.45276630e+00
-1.07912041e-01 -5.56431174e-01 3.92146409e-01 8.40738833e-01
-9.91956517e-02 3.66669595e-01 -1.07126272e+00 -1.45473942e-01
1.05684447e+00 1.74156642e+00 -7.77983606e-01 -1.74918368e-01
-4.39228177e-01 3.52848262e-01 7.40820765e-01 1.53653830e-01
-5.36230743e-01 -8.36863995e-01 7.99448073e-01 6.17336392e-01
2.45832801e-01 -3.16985309e-01 -3.40609938e-01 -9.94165301e-01
2.67724972e-03 -1.06600511e+00 5.48651338e-01 -2.91843057e-01
-1.46993780e+00 4.54605550e-01 2.04678297e-01 -6.20521128e-01
-1.46663696e-01 -6.48648262e-01 -8.80167723e-01 8.61739039e-01
-1.02334428e+00 -9.75174785e-01 -8.23064387e-01 4.80878353e-01
-1.49883389e-01 1.20225258e-01 1.02907860e+00 2.35490680e-01
-7.06007659e-01 7.09622860e-01 -6.42672405e-02 3.28971088e-01
4.11002427e-01 -1.53666162e+00 1.15239608e+00 3.98423851e-01
-1.38969775e-02 1.07180619e+00 7.63042331e-01 -7.39575624e-01
-1.59361768e+00 -1.10171807e+00 6.93581939e-01 -3.92744452e-01
9.36665416e-01 -8.75519693e-01 -5.41423678e-01 8.60498309e-01
2.85800785e-01 3.06824178e-01 7.81404793e-01 7.73471832e-01
-3.69531006e-01 -1.42178521e-01 -1.35922015e+00 8.72457922e-01
1.37013328e+00 -6.24069095e-01 8.03572163e-02 6.26230180e-01
5.16626358e-01 -5.36304563e-02 -1.25865340e+00 1.31803051e-01
6.20599449e-01 -1.05471754e+00 1.10958350e+00 -6.95405185e-01
-1.30288288e-01 3.79058942e-02 2.31172398e-01 -1.10708141e+00
-4.23303574e-01 -1.36960661e+00 -4.23389763e-01 1.27102268e+00
4.73406136e-01 -9.04599488e-01 1.21729326e+00 6.40347183e-01
4.34535742e-01 -8.23087156e-01 -4.03779358e-01 -5.43571711e-01
-1.36339962e-01 -2.90520906e-01 8.07334125e-01 1.04738736e+00
5.59197545e-01 6.21360898e-01 2.62732115e-02 -1.74892500e-01
8.24712396e-01 -9.12727267e-02 1.38243473e+00 -1.50209069e+00
-3.03447336e-01 -3.45494151e-01 -5.44858932e-01 -7.33926415e-01
-2.41424143e-01 -1.30160451e+00 -8.47517908e-01 -1.76240218e+00
3.63246620e-01 -5.80116332e-01 1.21892817e-01 5.74899018e-01
-2.80585438e-01 1.45428687e-01 -2.20650792e-01 -3.58308442e-02
-5.89016557e-01 3.39513630e-01 8.34131181e-01 4.45573442e-02
-2.24789485e-01 9.98431593e-02 -9.78498816e-01 4.68153805e-01
1.00683534e+00 -4.60160136e-01 -5.75899780e-01 -1.92065705e-02
5.70668936e-01 -5.22624791e-01 4.09026772e-01 -9.09674168e-01
1.27235189e-01 -4.27829213e-02 2.36189082e-01 -4.45753783e-01
1.30949721e-01 -5.50930679e-01 1.23134464e-01 4.68562365e-01
-2.84939617e-01 2.48200834e-01 1.31898507e-01 6.17678404e-01
3.86509746e-01 -1.61285214e-02 5.14184713e-01 -3.83921564e-01
-3.41966242e-01 5.57633102e-01 4.87156659e-02 3.72923434e-01
1.04492390e+00 -2.31711879e-01 -3.29865128e-01 -5.49750328e-01
-5.48455298e-01 3.09856027e-01 6.48808658e-01 1.29117131e-01
4.10777152e-01 -1.22523248e+00 -7.08997786e-01 1.92404792e-01
1.84498996e-01 -3.73844266e-01 -1.97175488e-01 6.09154642e-01
-6.72019362e-01 -1.16444297e-01 9.47069004e-02 -5.42934120e-01
-1.54378521e+00 6.83422506e-01 1.92239106e-01 -1.75468519e-01
-7.38802612e-01 7.32188404e-01 -5.55613816e-01 -8.84787202e-01
1.91662431e-01 -1.92557335e-01 9.61448997e-02 -3.81063521e-02
5.81286252e-01 6.25309646e-01 6.86796978e-02 2.56130755e-01
-5.99975884e-01 -3.85117047e-02 -1.75564930e-01 4.10556316e-01
1.59704757e+00 -9.18040276e-02 -5.03326595e-01 4.39059734e-01
1.20751715e+00 4.14803058e-01 -8.54268491e-01 -1.68060645e-01
-2.86690835e-02 -5.30738950e-01 -6.86315298e-02 -4.37087148e-01
-8.72385919e-01 4.05834317e-01 8.65896419e-02 9.61371779e-01
5.19938350e-01 1.39391974e-01 4.52872962e-01 2.90337235e-01
5.15392900e-01 -7.24131405e-01 -2.89654106e-01 6.56677186e-02
6.49976611e-01 -9.53243315e-01 6.43094659e-01 -9.43183780e-01
-3.12153041e-01 1.14496398e+00 3.61310631e-01 -3.85022014e-01
1.12011313e+00 4.04035002e-01 -4.72703844e-01 -4.55150843e-01
-9.18142259e-01 -2.02413797e-01 5.45677841e-02 9.32428598e-01
6.86544180e-01 9.37170256e-03 -2.63156533e-01 2.32276797e-01
-4.89203811e-01 -1.33137316e-01 5.34176290e-01 7.79441357e-01
-3.61088336e-01 -1.35446692e+00 1.33229986e-01 8.65565479e-01
-2.56835759e-01 -4.76378977e-01 -6.44999444e-01 1.18565249e+00
-5.13236046e-01 6.89353704e-01 -1.82677746e-01 -5.30450225e-01
1.10452347e-01 -6.90692067e-02 3.02360386e-01 -6.95336699e-01
-6.56071544e-01 -5.39392889e-01 7.76923060e-01 -6.62666321e-01
1.36722639e-01 -6.51868641e-01 -6.13609314e-01 -1.17643929e+00
-4.52826440e-01 1.50659680e-01 5.29742599e-01 1.88525438e-01
8.92411828e-01 2.61302978e-01 6.19596481e-01 -8.27448130e-01
-4.89483416e-01 -1.10247695e+00 -6.13850474e-01 1.74219295e-01
5.52773327e-02 -3.95808250e-01 -5.47434449e-01 -4.55687255e-01] | [6.97621488571167, 5.955612659454346] |
48496ba0-749e-48a8-bd38-ac75367c5d3e | is-disentanglement-enough-on-latent | 2108.01450 | null | https://arxiv.org/abs/2108.01450v1 | https://arxiv.org/pdf/2108.01450v1.pdf | Is Disentanglement enough? On Latent Representations for Controllable Music Generation | Improving controllability or the ability to manipulate one or more attributes of the generated data has become a topic of interest in the context of deep generative models of music. Recent attempts in this direction have relied on learning disentangled representations from data such that the underlying factors of variation are well separated. In this paper, we focus on the relationship between disentanglement and controllability by conducting a systematic study using different supervised disentanglement learning algorithms based on the Variational Auto-Encoder (VAE) architecture. Our experiments show that a high degree of disentanglement can be achieved by using different forms of supervision to train a strong discriminative encoder. However, in the absence of a strong generative decoder, disentanglement does not necessarily imply controllability. The structure of the latent space with respect to the VAE-decoder plays an important role in boosting the ability of a generative model to manipulate different attributes. To this end, we also propose methods and metrics to help evaluate the quality of a latent space with respect to the afforded degree of controllability. | ['Alexander Lerch', 'Ashis Pati'] | 2021-08-01 | null | null | null | null | ['music-generation', 'music-generation'] | ['audio', 'music'] | [ 1.69539496e-01 2.18192354e-01 -2.53022939e-01 -2.91735172e-01
-4.16188151e-01 -9.14760709e-01 9.72782373e-01 -1.91313669e-01
-2.07620580e-02 6.85342968e-01 5.93471766e-01 8.26309249e-02
-3.41748804e-01 -7.73724973e-01 -7.50284612e-01 -9.40710485e-01
2.52406657e-01 4.44014549e-01 -4.29584146e-01 -2.96009660e-01
-6.76449090e-02 3.68417323e-01 -1.44085991e+00 1.16672497e-02
9.23064768e-01 4.54753578e-01 1.85467757e-03 4.61964011e-01
4.97557640e-01 4.24492806e-01 -4.72339928e-01 -3.25428396e-01
2.75682449e-01 -7.67155170e-01 -2.94522762e-01 -1.96594894e-02
3.33223760e-01 -5.85804135e-02 -3.20853889e-01 1.08799648e+00
3.58923793e-01 -2.23845735e-01 9.14171875e-01 -1.25155401e+00
-5.77022433e-01 8.52297068e-01 -1.31697595e-01 5.93449697e-02
8.15308914e-02 2.42649928e-01 1.64990962e+00 -5.00700593e-01
4.27711427e-01 1.10702944e+00 1.37068138e-01 4.76892799e-01
-2.00397229e+00 -6.89186752e-01 -7.37061203e-02 -1.37051016e-01
-1.15515816e+00 -5.29527724e-01 1.06979656e+00 -8.32946777e-01
2.75793701e-01 3.42166066e-01 7.71728933e-01 1.33195364e+00
1.62327096e-01 6.05363369e-01 1.19809079e+00 -3.18608373e-01
3.33774090e-01 2.12728769e-01 -2.18209118e-01 6.38654351e-01
5.13099670e-01 3.46522689e-01 -7.46585429e-01 -3.82148810e-02
9.68259513e-01 -2.54826635e-01 -4.88983929e-01 -1.13706851e+00
-1.29489720e+00 1.11817658e+00 5.08160532e-01 3.89018416e-01
-1.35959327e-01 1.73986182e-01 1.59585655e-01 2.45115116e-01
1.71857476e-01 1.32885444e+00 -3.07216078e-01 -5.42072505e-02
-8.79511893e-01 2.55331904e-01 6.10321760e-01 5.21713257e-01
5.89971602e-01 3.20627868e-01 -3.78923506e-01 2.19461963e-01
3.53501052e-01 3.65942150e-01 2.55698591e-01 -8.24994028e-01
4.35967535e-01 6.43726170e-01 -1.84586588e-02 -9.63816106e-01
-2.22613313e-03 -7.33711541e-01 -7.75892854e-01 5.12521267e-01
4.24944013e-01 -1.51893824e-01 -6.42355382e-01 2.41335583e+00
6.82262182e-02 -7.43803009e-02 5.40428422e-02 8.81652117e-01
9.17825997e-02 3.09823722e-01 -2.12657496e-01 -6.68050051e-02
1.10815477e+00 -5.75609028e-01 -7.10756242e-01 -2.18551740e-01
1.86540842e-01 -4.94728863e-01 1.26749516e+00 2.76503652e-01
-1.08407164e+00 -4.94668722e-01 -1.40022290e+00 -5.92458844e-02
-2.91156899e-02 1.86833218e-01 5.69159210e-01 6.13537073e-01
-6.47365272e-01 6.88095808e-01 -1.12145627e+00 1.47978619e-01
3.56710643e-01 4.57809031e-01 -4.32687342e-01 2.16291666e-01
-1.25244474e+00 8.45686018e-01 2.20365286e-01 -8.73988047e-02
-1.13774920e+00 -6.74444199e-01 -7.86793888e-01 4.29785252e-01
2.76837081e-01 -1.02943337e+00 7.30303824e-01 -1.04996169e+00
-1.52659690e+00 4.64341521e-01 2.04281941e-01 -9.50837061e-02
5.16489327e-01 -2.95998603e-01 5.39694652e-02 -1.68872356e-01
1.13643132e-01 2.65312403e-01 1.17584741e+00 -1.16198206e+00
1.07672200e-01 -6.49601936e-01 3.10133129e-01 1.61211774e-01
-3.36750418e-01 -4.48848963e-01 9.91302505e-02 -7.52246082e-01
-1.07256025e-01 -1.27554667e+00 1.27891935e-02 -1.05509767e-02
-6.95938408e-01 1.16037942e-01 4.42323565e-01 -2.81788290e-01
1.04317939e+00 -2.29568648e+00 1.12007904e+00 -2.46173814e-02
5.23286104e-01 1.88295066e-01 -9.00387391e-02 5.62927127e-01
-1.63684398e-01 1.89511254e-01 -6.31504655e-02 -3.00449699e-01
2.27048606e-01 1.85375854e-01 -4.16130632e-01 3.67128968e-01
3.86729240e-01 8.36759746e-01 -7.94696152e-01 -1.75582469e-02
-8.08114558e-02 7.27375150e-01 -9.86443222e-01 4.30694222e-01
-4.48357224e-01 8.53935182e-01 -6.14516616e-01 -1.62679534e-02
1.10381298e-01 -1.97140232e-01 4.48199540e-01 -2.33655810e-01
5.84149770e-02 6.56297863e-01 -1.00468111e+00 1.79082143e+00
-4.84999478e-01 7.16844141e-01 -7.99163282e-02 -7.88858056e-01
6.82659209e-01 5.08153975e-01 2.71765292e-01 -3.73914421e-01
4.06656489e-02 1.80818796e-01 5.56979775e-01 -1.40623361e-01
2.27763519e-01 -4.68473852e-01 -2.08344385e-01 5.39705455e-01
2.04848424e-01 -7.40738735e-02 6.11775704e-02 1.78689793e-01
8.67413342e-01 2.09428549e-01 3.44590694e-01 -4.21752930e-01
6.24417402e-02 -3.54876012e-01 5.55884004e-01 4.90247399e-01
1.45740882e-01 4.60579425e-01 1.02103472e+00 4.12750915e-02
-1.24739063e+00 -1.27025783e+00 -1.75208464e-01 7.33082592e-01
-2.87696600e-01 -5.52770615e-01 -5.48636436e-01 -4.17605013e-01
-3.68640386e-02 8.31554472e-01 -8.52647603e-01 -5.84649742e-01
-3.97288024e-01 -6.52245402e-01 4.92787242e-01 4.79370654e-01
1.80338830e-01 -6.02324426e-01 -7.51099467e-01 -2.19796374e-01
-8.49923640e-02 -8.25877368e-01 -3.63645464e-01 4.04425889e-01
-7.66878366e-01 -8.23372602e-01 -3.49903166e-01 -1.99790299e-01
5.25049925e-01 -7.36589953e-02 1.13067842e+00 -5.12818754e-01
8.36967006e-02 2.13957988e-02 -1.35996640e-02 -1.78444341e-01
-4.89489824e-01 3.71526599e-01 1.58815801e-01 4.59082313e-02
-5.60786091e-02 -1.15977955e+00 -5.46620131e-01 2.15676069e-01
-1.10061407e+00 3.30450982e-01 6.00184500e-01 8.84711742e-01
1.85951307e-01 -3.34395878e-02 4.28393602e-01 -8.37143362e-01
7.01755404e-01 -3.39833707e-01 -6.29469812e-01 6.56306222e-02
-7.58187830e-01 8.73931468e-01 5.18076599e-01 -5.20318747e-01
-8.08807790e-01 -5.15456051e-02 2.21085519e-01 -5.49628913e-01
1.29091114e-01 5.24030924e-01 -7.70698786e-01 5.57040155e-01
6.29884243e-01 -2.77710846e-04 1.25283852e-01 -3.87468398e-01
8.71242344e-01 1.19441539e-01 8.01625773e-02 -6.29974782e-01
9.58373427e-01 2.69418150e-01 2.01953709e-01 -3.06300610e-01
-9.45596576e-01 1.82890996e-01 -7.56828249e-01 1.52806714e-01
1.13753271e+00 -9.90686238e-01 -4.69416529e-01 -1.82865711e-03
-7.37022758e-01 -1.67324394e-01 -4.39476699e-01 5.54976702e-01
-7.76667893e-01 -1.08813994e-01 -3.60451400e-01 -4.22395587e-01
6.02671914e-02 -1.47802675e+00 1.05333650e+00 -9.77843329e-02
-5.26628673e-01 -9.58517134e-01 4.43517804e-01 3.67059499e-01
3.36955518e-01 4.53068554e-01 1.26205981e+00 -6.94737792e-01
-8.57648790e-01 -1.70532137e-01 2.41884172e-01 2.15775624e-01
3.23215544e-01 -1.38420090e-01 -1.03393471e+00 -2.85226345e-01
1.66718423e-01 -4.43446726e-01 9.38972235e-01 2.82840073e-01
5.95509171e-01 -5.22925973e-01 4.18670252e-02 5.98453820e-01
1.20957673e+00 -1.63504228e-01 4.81522173e-01 5.31069487e-02
9.41527605e-01 4.57564890e-01 3.74178141e-02 2.71001726e-01
5.80506101e-02 1.04497218e+00 4.49096322e-01 1.28708988e-01
-7.08279088e-02 -6.51370883e-01 5.04743636e-01 7.26922631e-01
-1.56314567e-01 -1.75850555e-01 -4.97190386e-01 2.38101393e-01
-1.72166657e+00 -9.47723687e-01 5.00786789e-02 2.30590749e+00
1.05545139e+00 2.43579939e-01 1.58655003e-01 2.85639018e-01
3.05606842e-01 5.05284727e-01 -6.33475780e-01 -1.77266896e-01
-5.72959147e-02 1.49253346e-02 -3.80037166e-02 6.41984344e-01
-7.46684730e-01 7.25459635e-01 5.70070505e+00 6.06007934e-01
-1.12426662e+00 -1.26596436e-01 3.00238281e-01 -2.64165044e-01
-8.66741419e-01 2.10786536e-01 -5.88557065e-01 3.37542742e-01
5.88198006e-01 -2.06155479e-02 5.54610848e-01 5.42896509e-01
1.85824186e-01 3.04031134e-01 -1.74963236e+00 5.10304689e-01
9.03966650e-03 -1.10401750e+00 1.59880519e-01 5.19819617e-01
7.31809676e-01 -4.07071829e-01 5.11762679e-01 2.28355050e-01
2.82317072e-01 -1.22957289e+00 7.21076190e-01 6.08620763e-01
7.11730361e-01 -5.18838763e-01 1.32823840e-01 3.90343815e-01
-6.61192179e-01 1.88501745e-01 -4.75830138e-02 -1.73446208e-01
3.22251171e-02 7.22279549e-01 -6.69489622e-01 3.21962148e-01
-3.28172036e-02 5.51555276e-01 -4.44293022e-01 2.87041187e-01
-7.74557710e-01 6.43479943e-01 2.42402032e-02 6.72698347e-03
-1.69409364e-01 -3.06824863e-01 9.52281415e-01 5.86213648e-01
4.27678116e-02 -1.66134953e-01 -1.02741867e-01 1.54649889e+00
-4.38941047e-02 -2.17105687e-01 -6.86347723e-01 -6.12065494e-01
2.18332931e-01 9.81497288e-01 -2.11727604e-01 2.76711229e-02
-5.61765060e-02 9.16898429e-01 5.63453197e-01 3.56299013e-01
-7.21957743e-01 1.74790975e-02 1.03002691e+00 1.12569116e-01
3.60867798e-01 -6.07964218e-01 -4.32028025e-01 -1.66381884e+00
-1.09440081e-01 -1.06032658e+00 -8.56413096e-02 -6.29608870e-01
-1.10778689e+00 5.53227782e-01 -9.52348765e-03 -9.93146002e-01
-5.04135907e-01 -3.43929172e-01 -3.53365332e-01 1.05859637e+00
-1.01948667e+00 -1.13637805e+00 1.20743796e-01 4.29166198e-01
3.04184884e-01 -1.86115786e-01 9.67683196e-01 -8.35698545e-02
-7.74685323e-01 5.50683260e-01 1.67899579e-02 -4.16873433e-02
5.45063496e-01 -1.43338978e+00 -2.67493352e-02 8.63458514e-01
7.28703797e-01 9.29394007e-01 1.19380915e+00 -3.49459350e-01
-1.48347831e+00 -7.87729681e-01 7.37940848e-01 -8.77395749e-01
7.05028951e-01 -7.49768674e-01 -5.68733692e-01 7.85061717e-01
2.35135481e-01 -2.29214266e-01 9.26063776e-01 5.70287108e-01
-7.69421816e-01 -2.64390651e-02 -5.66167474e-01 7.31972218e-01
9.31772530e-01 -9.36845183e-01 -5.21043003e-01 -1.52650923e-01
7.83080220e-01 1.06911613e-02 -7.46370554e-01 3.36839050e-01
7.42405236e-01 -1.00073314e+00 9.36485052e-01 -8.29584122e-01
9.08497453e-01 -1.27977833e-01 -3.79286110e-01 -1.61334741e+00
-5.45826733e-01 -5.17672300e-01 -2.14238316e-01 1.24187148e+00
4.61159468e-01 -3.36631835e-01 6.15470767e-01 6.08282506e-01
1.70641392e-01 -6.29082441e-01 -6.34314299e-01 -5.66642821e-01
3.69492710e-01 -1.98197484e-01 5.25579035e-01 9.98247206e-01
8.98140520e-02 1.09518480e+00 -5.62900603e-01 1.79801762e-01
3.87914062e-01 4.59011734e-01 8.87422442e-01 -1.26305556e+00
-8.00866246e-01 -4.99272853e-01 -5.46812594e-01 -8.55409980e-01
2.04544306e-01 -1.09439957e+00 -2.04741180e-01 -1.14960289e+00
4.29108173e-01 -2.74537981e-01 -2.66358346e-01 2.71913499e-01
-2.04943776e-01 3.13642435e-02 3.34196657e-01 3.97423178e-01
-1.05154440e-01 1.00202680e+00 1.37388968e+00 1.48111368e-02
-1.61571294e-01 1.78780798e-02 -1.01141644e+00 5.22650719e-01
7.53203750e-01 -4.99176800e-01 -8.31467628e-01 -5.96906066e-01
6.55984044e-01 1.24622576e-01 4.32750463e-01 -6.70238912e-01
-2.97005981e-01 -1.41457409e-01 1.14715181e-01 2.39318922e-01
5.49269676e-01 -6.31337404e-01 2.89155722e-01 3.51367265e-01
-8.84153187e-01 1.59280524e-02 -1.22205101e-01 7.05055356e-01
-1.71618745e-01 1.06969319e-01 7.14761615e-01 1.51078597e-01
8.91036838e-02 1.12442262e-01 -3.28753591e-01 1.59433424e-01
7.55075634e-01 2.69319445e-01 -2.17295736e-02 -3.66465777e-01
-9.06748712e-01 -2.91624606e-01 5.10828912e-01 5.04636407e-01
2.60349214e-01 -1.48111749e+00 -5.95632493e-01 3.41792226e-01
9.73817706e-02 -3.74162197e-01 -1.54179275e-01 6.73587739e-01
6.41526356e-02 4.79681432e-01 -2.66313285e-01 -5.04239559e-01
-9.84336197e-01 6.31736696e-01 3.66550922e-01 -4.56065238e-01
-2.71339893e-01 6.20534182e-01 5.99407256e-01 -1.68366864e-01
-5.87531365e-02 -2.52981395e-01 -1.03197470e-01 1.17071643e-01
2.78380699e-02 4.36372459e-02 -2.77332783e-01 -4.54374164e-01
-9.42372829e-02 2.16911033e-01 1.27906337e-01 -4.54787344e-01
1.24201071e+00 1.45038694e-01 1.14524506e-01 7.57441819e-01
1.13686442e+00 4.68283266e-01 -1.46131086e+00 -2.30750013e-02
-2.55077839e-01 -5.63230932e-01 7.54746497e-02 -6.86789632e-01
-1.08498871e+00 1.11537480e+00 2.48224616e-01 2.61556715e-01
9.20468330e-01 1.22006655e-01 1.77325919e-01 1.55534437e-02
1.59354150e-01 -5.13067305e-01 4.62043107e-01 6.55281469e-02
9.99147236e-01 -1.11086977e+00 -2.14872733e-01 -2.29192391e-01
-7.68664122e-01 7.35128820e-01 3.42442960e-01 -3.70313615e-01
6.97722971e-01 1.55595198e-01 -2.98065037e-01 -4.19222295e-01
-7.91223526e-01 -1.78958371e-01 7.57454395e-01 2.72409827e-01
6.06656075e-01 3.62466693e-01 -1.93500996e-01 5.10809362e-01
-5.82487524e-01 -4.25419152e-01 4.63434517e-01 4.59855765e-01
-1.14647292e-01 -1.27451992e+00 1.07023783e-01 9.42520574e-02
-3.32136303e-01 -7.36759305e-02 -5.94660342e-01 7.61767328e-01
9.72148851e-02 7.71772921e-01 -1.61951229e-01 -3.44068259e-01
6.69798329e-02 7.68639967e-02 6.79031134e-01 -7.54588068e-01
-1.91510275e-01 2.30145618e-01 1.34660333e-01 -4.18341607e-01
-3.79383802e-01 -8.65453362e-01 -5.90592802e-01 4.76835780e-02
-4.48964685e-01 2.21540645e-01 4.06226933e-01 9.15798664e-01
4.35779870e-01 6.34288132e-01 7.03796089e-01 -5.61852992e-01
-8.33779812e-01 -8.12848508e-01 -7.05241144e-01 4.55822378e-01
5.28373957e-01 -7.35525966e-01 -4.61743325e-01 -3.07033658e-02] | [9.265371322631836, 4.876946926116943] |
3d171170-7af2-4b83-9481-c746a5109638 | genetic-analysis-of-prostate-cancer-with | 2303.15851 | null | https://arxiv.org/abs/2303.15851v2 | https://arxiv.org/pdf/2303.15851v2.pdf | Genetic Analysis of Prostate Cancer with Computer Science Methods | Metastatic prostate cancer is one of the most common cancers in men. In the advanced stages of prostate cancer, tumours can metastasise to other tissues in the body, which is fatal. In this thesis, we performed a genetic analysis of prostate cancer tumours at different metastatic sites using data science, machine learning and topological network analysis methods. We presented a general procedure for pre-processing gene expression datasets and pre-filtering significant genes by analytical methods. We then used machine learning models for further key gene filtering and secondary site tumour classification. Finally, we performed gene co-expression network analysis and community detection on samples from different prostate cancer secondary site types. In this work, 13 of the 14,379 genes were selected as the most metastatic prostate cancer related genes, achieving approximately 92% accuracy under cross-validation. In addition, we provide preliminary insights into the co-expression patterns of genes in gene co-expression networks. Project code is available at https://github.com/zcablii/Master_cancer_project. | ['Shi Zhou', 'YuXuan Li'] | 2023-03-28 | null | null | null | null | ['community-detection', 'tumour-classification'] | ['graphs', 'medical'] | [ 2.49723434e-01 7.49787763e-02 -6.38748631e-02 -1.23420656e-02
-3.87445062e-01 -4.59647417e-01 4.45988178e-01 6.72366619e-01
-1.67102814e-01 7.62045920e-01 1.63957998e-01 -4.57753778e-01
-6.56875014e-01 -1.09356999e+00 -8.65913108e-02 -9.65451062e-01
-7.43300438e-01 7.38777816e-01 -1.67718939e-02 -1.47690505e-01
2.97481149e-01 7.08795071e-01 -8.04715574e-01 1.57461315e-01
5.40727675e-01 2.58953780e-01 1.18062338e-02 1.06706154e+00
-1.11184195e-01 1.45125732e-01 -2.37297505e-01 -4.70710248e-01
-8.24795961e-02 -6.04231179e-01 -5.86253285e-01 -4.41040754e-01
-1.27621412e-01 7.22562671e-01 -3.34586293e-01 1.14032984e+00
5.57465494e-01 -3.19884658e-01 7.63830543e-01 -1.11365461e+00
1.52073875e-01 3.76102924e-01 -4.56586987e-01 1.62134185e-01
2.38439918e-01 -7.92430341e-02 8.70862603e-01 -7.45548666e-01
1.07491028e+00 1.18978357e+00 7.77231038e-01 4.91661519e-01
-1.55545390e+00 -5.93365133e-01 -7.13153660e-01 1.58432454e-01
-1.48485684e+00 -3.57965410e-01 5.26558220e-01 -6.17268026e-01
5.76988578e-01 7.13603914e-01 9.84159648e-01 8.15065861e-01
8.60636115e-01 6.09094165e-02 1.13384664e+00 -2.15946525e-01
1.83410034e-01 -2.71901548e-01 -7.51341740e-03 7.96625793e-01
5.60697019e-01 -9.50433314e-02 -3.53068113e-01 -6.33103490e-01
6.66209280e-01 3.47248405e-01 4.94551212e-02 -8.78753513e-02
-1.13475037e+00 8.10414135e-01 5.08527398e-01 6.62485778e-01
-1.83499768e-01 6.26592711e-02 5.69115758e-01 2.69391149e-01
1.10242218e-01 3.36379856e-01 -2.60509014e-01 -1.55090556e-01
-6.82279468e-01 -3.23529392e-01 8.20627749e-01 4.86698836e-01
3.09206158e-01 -8.61798525e-01 -4.04284522e-02 7.93072820e-01
7.04212308e-01 1.63202196e-01 4.83657539e-01 -5.29861212e-01
-1.19925179e-01 1.01578236e+00 -6.68327689e-01 -1.40000176e+00
-5.85115075e-01 -5.46436250e-01 -1.38833356e+00 -1.54562056e-01
5.08685708e-01 2.73302823e-01 -5.92873394e-01 1.12981606e+00
9.96836349e-02 -6.76713958e-02 -9.14327893e-03 4.63703692e-01
5.06919444e-01 2.76421737e-02 4.25504744e-01 -3.32467079e-01
1.69049406e+00 -2.62582064e-01 -5.31549752e-01 2.36108765e-01
7.46748984e-01 -6.19765341e-01 3.28018695e-01 1.29919043e-02
-5.63138604e-01 2.68345680e-02 -6.32748425e-01 2.09753782e-01
-4.85957891e-01 -7.72115737e-02 1.06639552e+00 6.92504168e-01
-9.34766054e-01 8.13703537e-01 -1.19543660e+00 -1.00129962e+00
7.01523840e-01 5.19730926e-01 -7.22208679e-01 -3.03951979e-01
-8.69043231e-01 5.74948907e-01 1.99183658e-01 2.34824140e-02
-7.57981300e-01 -6.49977207e-01 -3.04804116e-01 -1.35863855e-01
-2.42665395e-01 -8.68296325e-01 3.76320869e-01 -1.35064363e-01
-9.56753671e-01 1.01568604e+00 -5.04907668e-01 8.59161168e-02
3.07403833e-01 5.40925026e-01 -4.58656371e-01 9.13445465e-03
1.87953845e-01 2.14800552e-01 -2.78032243e-01 -7.96118379e-01
-4.95294303e-01 -7.90400624e-01 -8.62723708e-01 -1.38826529e-02
-2.46748701e-01 3.22207101e-02 -5.32912493e-01 -2.44284898e-01
4.63054687e-01 -1.12462342e+00 -8.54992330e-01 -4.59580347e-02
-5.54679334e-01 -1.82438210e-01 4.52144742e-01 -4.91679549e-01
9.64847624e-01 -2.37225389e+00 4.64137822e-01 8.00458550e-01
4.24972534e-01 -3.48949939e-01 9.85104814e-02 6.58572078e-01
-2.22467527e-01 7.35823154e-01 -4.91602793e-02 9.65124965e-02
-2.75916904e-01 1.91400290e-01 6.16498768e-01 7.74765134e-01
7.33446181e-02 8.84164989e-01 -1.12916780e+00 -6.06248319e-01
-5.71401231e-02 4.66199309e-01 -1.85884861e-03 -2.91655004e-01
3.48072529e-01 5.04930377e-01 -5.38220406e-01 1.08676541e+00
4.53268528e-01 -9.07715037e-02 6.22920156e-01 1.51154422e-03
3.54445219e-01 -1.27441928e-01 -5.84066868e-01 1.54978037e+00
1.49638727e-01 6.99779630e-01 5.13914704e-01 -7.01717675e-01
1.13617158e+00 3.19664367e-02 9.01261806e-01 -2.73192346e-01
6.90075979e-02 2.76246488e-01 4.37102914e-01 -3.68618220e-01
-4.95950831e-03 -6.24259785e-02 -1.96038380e-01 -9.59460065e-02
-2.81518817e-01 1.42519161e-01 6.14233255e-01 4.33611363e-01
1.88383222e+00 -7.66382396e-01 3.69457632e-01 -6.14064097e-01
5.93233287e-01 3.76597315e-01 7.54562616e-01 4.01612639e-01
-5.13193607e-01 1.16914496e-01 1.01630294e+00 -1.92673713e-01
-8.81740451e-01 -9.89758015e-01 -4.65596735e-01 7.22883701e-01
-2.52057493e-01 -8.87487233e-01 -1.99005287e-02 -2.74840772e-01
2.63690621e-01 1.45551711e-01 -5.68213761e-01 -1.06512345e-01
-1.69951469e-01 -1.34894657e+00 9.23345506e-01 -6.40998408e-02
2.61764735e-01 -4.09004658e-01 4.56794709e-01 5.91364466e-02
1.01857834e-01 -2.18711600e-01 1.81326672e-01 4.60738331e-01
-1.18231988e+00 -1.60115778e+00 -2.89212167e-01 -7.71779299e-01
1.03132319e+00 -2.04094827e-01 7.70295084e-01 2.28510469e-01
-9.86816764e-01 -4.65941019e-02 7.04761520e-02 -1.82783872e-01
-5.34869194e-01 -2.43064016e-02 -6.27017692e-02 -5.41741014e-01
6.68639481e-01 -6.99253798e-01 -3.42950672e-01 3.96691531e-01
-5.51043749e-01 -2.58103907e-01 9.08894718e-01 7.58115649e-01
7.30524123e-01 7.38476396e-01 1.03228450e-01 -8.41122150e-01
8.35983157e-01 -8.24026465e-01 -1.58785388e-01 -6.14294447e-02
-5.31759262e-01 -2.80532062e-01 1.51851341e-01 2.23204762e-01
-5.05515337e-01 5.83841324e-01 -1.49512336e-01 3.12256217e-01
-2.37720564e-01 8.07268262e-01 -2.80919224e-01 -2.47222129e-02
5.78456998e-01 1.83789656e-01 5.81261992e-01 -1.78982392e-01
-1.63783774e-01 4.27807122e-01 1.65756449e-01 -1.07385628e-01
4.52174872e-01 4.39599186e-01 1.00555897e+00 -1.03193843e+00
1.79138005e-01 -7.02717066e-01 -8.12122881e-01 -3.71512532e-01
7.96857834e-01 -8.28494549e-01 -9.53441024e-01 2.35125154e-01
-3.46152782e-01 -9.43710804e-02 2.26152554e-01 4.98147756e-01
-1.24267854e-01 3.47547412e-01 -9.88594711e-01 -4.44594294e-01
-1.15890928e-01 -7.54463971e-01 7.35162795e-01 2.51481444e-01
-4.21590060e-01 -1.28096628e+00 5.41055143e-01 2.65674859e-01
3.81946415e-01 6.22125685e-01 1.00662732e+00 -7.06153274e-01
-2.86386043e-01 -5.19275725e-01 8.53798464e-02 -6.22905433e-01
2.02730492e-01 5.06524682e-01 -4.20655966e-01 -4.10664290e-01
-6.52725577e-01 4.11096960e-01 9.28641856e-01 2.78035015e-01
8.59361947e-01 1.05653465e-01 -1.33974349e+00 5.32890856e-01
1.60167742e+00 2.00021073e-01 9.03061807e-01 4.63181138e-01
2.82639325e-01 6.49563730e-01 5.39705396e-01 3.94394854e-03
-1.67920575e-01 3.13078970e-01 3.59304339e-01 5.83493635e-02
1.04869843e-01 6.40295446e-02 1.53170481e-01 3.66090387e-01
-5.15117466e-01 4.73296121e-02 -1.36544538e+00 5.31932175e-01
-1.72481430e+00 -9.08921182e-01 -9.75436985e-01 1.96713495e+00
7.57334411e-01 -8.67822394e-02 -1.71656143e-02 -1.27195999e-01
8.75876427e-01 -6.20850563e-01 -1.66590914e-01 -1.59677953e-01
9.49895754e-03 4.45266999e-02 8.20168078e-01 2.15491265e-01
-1.01634932e+00 4.43983525e-01 6.22421741e+00 6.37659132e-01
-6.56182587e-01 -1.39892012e-01 7.79812753e-01 2.45323405e-03
-1.88025516e-02 1.25018597e-01 -6.29622042e-01 3.17383617e-01
1.07340837e+00 -5.29895365e-01 1.08333547e-02 7.44371831e-01
3.96005780e-01 -2.04429016e-01 -8.04192126e-01 6.04239047e-01
-5.46590507e-01 -1.30576110e+00 -6.85499907e-01 8.69034708e-01
4.70756352e-01 2.57796556e-01 -6.74757719e-01 9.73651335e-02
6.13642514e-01 -1.17281115e+00 -6.24033749e-01 1.01128769e+00
6.64604247e-01 -6.07923031e-01 1.09762394e+00 4.15864028e-02
-1.05752885e+00 -1.37417629e-01 -4.82293040e-01 -3.10167037e-02
-1.91176519e-01 1.11204112e+00 -1.14538240e+00 7.48748243e-01
7.16496348e-01 8.86704206e-01 -9.73542690e-01 1.19704962e+00
-1.23832978e-01 6.52000189e-01 -3.66706580e-01 -3.92782450e-01
-4.80023265e-01 7.72766843e-02 6.79330170e-01 1.16522491e+00
3.55370283e-01 -7.79474378e-02 1.19816117e-01 5.98634720e-01
3.66924912e-01 4.90719564e-02 -4.88322556e-01 -3.52706701e-01
6.10737443e-01 1.81173038e+00 -1.21066177e+00 -1.43982787e-02
-2.48169769e-02 8.58766854e-01 6.46943077e-02 -3.05715613e-02
-1.43296748e-01 -4.87383455e-01 9.14994001e-01 1.59903303e-01
-4.45013851e-01 -2.74700493e-01 -2.88576394e-01 -6.43247783e-01
-6.13531351e-01 -2.64631033e-01 6.72267795e-01 -9.17954743e-02
-1.68421984e+00 -6.03470840e-02 -4.13360804e-01 -8.54740560e-01
8.63752812e-02 -8.07191670e-01 -8.56870532e-01 8.58567357e-01
-7.62725711e-01 -8.32858086e-01 -3.37986320e-01 3.51029217e-01
-1.07165799e-01 -1.55855924e-01 1.31331980e+00 2.43407652e-01
-9.22521591e-01 -1.76979095e-01 4.54284370e-01 2.88574338e-01
6.92773998e-01 -1.39946532e+00 -1.24969229e-01 3.03823709e-01
-7.83549070e-01 9.30936694e-01 4.38361466e-01 -1.25906372e+00
-1.71050012e+00 -1.20902908e+00 1.06234169e+00 -1.60459459e-01
9.92922783e-01 -5.07554531e-01 -4.14554000e-01 4.58766818e-01
-6.86956123e-02 -1.51568865e-02 1.37248850e+00 2.29628280e-01
5.03849149e-01 7.30711296e-02 -1.15064180e+00 4.88414437e-01
8.33784103e-01 -1.98577225e-01 5.84528521e-02 7.17932582e-01
-1.65330216e-01 3.28957774e-02 -1.74300301e+00 9.90982801e-02
5.80911934e-01 -6.31486535e-01 1.08064842e+00 -3.68020296e-01
3.94297838e-01 -6.56501710e-01 -8.64840001e-02 -1.27579784e+00
-1.12202251e+00 -1.90654352e-01 5.08471489e-01 1.31190908e+00
5.13382018e-01 -5.00659645e-01 1.14063406e+00 4.52859998e-01
6.63845660e-03 -5.26641548e-01 -1.17998993e+00 -5.29598951e-01
2.91105546e-02 1.21723928e-01 -8.24925750e-02 1.00798285e+00
6.40057445e-01 3.30045730e-01 3.78214955e-01 -4.50645806e-03
8.59048605e-01 -2.61698455e-01 7.80426443e-01 -1.53881776e+00
7.54416212e-02 -7.78010190e-01 -1.24483001e+00 1.30683586e-01
-3.14148903e-01 -1.39543271e+00 -2.67115414e-01 -1.85799599e+00
5.21899223e-01 -4.92589027e-01 -2.38440782e-01 6.49262786e-01
-4.49357256e-02 3.05439770e-01 -2.61316091e-01 5.49388885e-01
-2.69026607e-01 -7.71235675e-02 1.15254819e+00 -2.41398349e-01
-9.63738188e-02 9.32142138e-02 -6.44023955e-01 3.10830146e-01
1.30634499e+00 -4.97119427e-01 2.38861114e-01 6.62205637e-01
3.68079484e-01 -4.29602340e-03 4.61209834e-01 -8.77015471e-01
5.15961289e-01 -2.50449449e-01 9.62024808e-01 -5.33185244e-01
2.32374936e-01 -8.83845091e-01 9.76704240e-01 1.30929506e+00
-1.98922142e-01 -2.31824055e-01 8.91686752e-02 8.08480561e-01
-1.50844008e-02 -1.13689698e-01 3.46898347e-01 -4.47009176e-01
-5.67028761e-01 -1.35997787e-01 -8.89754951e-01 -9.39930201e-01
1.32684433e+00 -2.18830764e-01 -5.43510079e-01 1.69417813e-01
-1.30030930e+00 3.58709872e-01 4.81164545e-01 5.33650815e-02
2.78792381e-01 -1.24308360e+00 -8.23656261e-01 6.72963113e-02
2.34797508e-01 -3.85517478e-01 1.03660613e-01 1.47515666e+00
-6.89418256e-01 3.89384329e-01 -3.16996366e-01 -7.85177886e-01
-1.91740870e+00 2.68283963e-01 3.36988300e-01 -5.05913496e-01
-2.18594864e-01 8.38298380e-01 -5.52903950e-01 -4.40429032e-01
-1.30178720e-01 4.09998864e-01 -3.96240950e-01 -1.85453554e-03
8.21681619e-02 5.03396630e-01 1.70134027e-02 -4.85536486e-01
-7.97764301e-01 2.86647648e-01 8.84512216e-02 3.77122432e-01
1.48783863e+00 -7.12074190e-02 -1.15541315e+00 2.50475854e-01
1.12755644e+00 3.36008608e-01 -1.41388535e-01 3.50003898e-01
3.80676121e-01 -5.42323828e-01 -1.54357687e-01 -6.80176675e-01
-8.55364859e-01 1.76660240e-01 6.63819671e-01 2.60118067e-01
1.03013861e+00 2.69109219e-01 -2.17849180e-01 3.51401091e-01
7.06195235e-02 -7.98168719e-01 -4.16274816e-01 2.58510321e-01
5.45196414e-01 -8.75085652e-01 3.64933908e-01 -9.01576161e-01
3.52751085e-04 1.50238562e+00 2.94979572e-01 -1.59818634e-01
7.96229064e-01 2.91823983e-01 -2.67725438e-01 -8.69809270e-01
-9.55281258e-01 3.20622772e-01 -2.38247350e-01 4.96854693e-01
1.02561235e+00 5.48680544e-01 -8.30129325e-01 4.01770532e-01
-4.15504992e-01 -8.24680179e-02 5.02777696e-01 9.68287170e-01
-4.11471426e-01 -1.37468565e+00 -4.41573441e-01 8.75270724e-01
-6.05864823e-01 -2.69331243e-02 -1.19420469e+00 7.85267651e-01
-8.54578689e-02 7.88174093e-01 1.34477228e-01 -4.65805113e-01
4.30822745e-02 1.63609907e-01 1.76420808e-01 -3.32773775e-01
-3.82152855e-01 3.33356649e-01 4.37702209e-01 -3.10601234e-01
-4.26000506e-01 -1.06027579e+00 -1.29802096e+00 -5.02227783e-01
-3.14820349e-01 3.61165643e-01 1.08245575e+00 2.25758106e-01
6.77145660e-01 9.22970414e-01 3.29736471e-01 -3.45572889e-01
3.33569348e-01 -1.19213092e+00 -8.99978578e-01 1.18230782e-01
-3.65375251e-01 -2.86770850e-01 -4.15893883e-01 3.12809274e-02] | [6.399196624755859, 5.49188232421875] |
d8eb2392-979e-429d-b193-5d8517208ec2 | reconnet-non-iterative-reconstruction-of-1 | null | null | http://openaccess.thecvf.com/content_cvpr_2016/html/Kulkarni_ReconNet_Non-Iterative_Reconstruction_CVPR_2016_paper.html | http://openaccess.thecvf.com/content_cvpr_2016/papers/Kulkarni_ReconNet_Non-Iterative_Reconstruction_CVPR_2016_paper.pdf | ReconNet: Non-Iterative Reconstruction of Images From Compressively Sensed Measurements | The goal of this paper is to present a non-iterative and more importantly an extremely fast algorithm to reconstruct images from compressively sensed (CS) random measurements. To this end, we propose a novel convolutional neural network (CNN) architecture which takes in CS measurements of an image as input and outputs an intermediate reconstruction. We call this network, ReconNet. The intermediate reconstruction is fed into an off-the-shelf denoiser to obtain the final reconstructed image. On a standard dataset of images we show significant improvements in reconstruction results (both in terms of PSNR and time complexity) over state-of-the-art iterative CS reconstruction algorithms at various measurement rates. Further, through qualitative experiments on real data collected using our block SPC (single pixel camera), we show that our network is highly robust to sensor noise and can recover visually better quality images than competitive algorithms at extremely low sensing rates of 0.1 and 0.04. To demonstrate that our algorithm can recover semantically informative images even at a low measurement rate of 0.01, we present a very robust proof of concept real-time visual tracking application. | ['Amit Ashok', 'Suhas Lohit', 'Pavan Turaga', 'Ronan Kerviche', 'Kuldeep Kulkarni'] | 2016-06-01 | null | null | null | cvpr-2016-6 | ['real-time-visual-tracking'] | ['computer-vision'] | [ 9.47330773e-01 -1.37882575e-01 1.75523475e-01 -1.09781161e-01
-9.25403297e-01 -5.57206750e-01 3.92852694e-01 -2.99595237e-01
-4.56742346e-01 3.41890663e-01 7.79852122e-02 -4.09026951e-01
-1.31347746e-01 -6.19111717e-01 -1.32088017e+00 -5.50590456e-01
-2.51408160e-01 -1.07076196e-02 2.03854755e-01 -1.12563446e-01
6.20728992e-02 5.69059849e-01 -1.37826955e+00 1.26296267e-01
3.65797520e-01 1.44464111e+00 3.85588109e-01 9.93322730e-01
5.92824578e-01 1.07400942e+00 -2.85630971e-01 -8.86958688e-02
5.96461296e-01 -4.24724489e-01 -4.21699911e-01 3.50769043e-01
5.36082804e-01 -8.43835056e-01 -6.15496099e-01 1.21733963e+00
4.81272787e-01 -2.31344298e-01 1.95523083e-01 -8.88341188e-01
-7.28033304e-01 7.81415761e-01 -4.14071262e-01 8.18902720e-03
4.64921832e-01 1.79807320e-01 5.68110883e-01 -8.95752728e-01
5.81755638e-01 9.48832214e-01 9.76133168e-01 3.67233753e-01
-1.22778416e+00 -6.70539200e-01 -3.86885703e-01 1.83025166e-01
-1.29284215e+00 -8.59634280e-01 8.00003886e-01 7.40948468e-02
5.37499011e-01 1.21746793e-01 5.05421937e-01 1.03236127e+00
1.01825126e-01 6.23663425e-01 1.04891992e+00 -5.48031271e-01
4.41002876e-01 -4.36283469e-01 -5.48784673e-01 6.43611908e-01
1.80312663e-01 3.80795896e-01 -5.11606812e-01 2.63335049e-01
1.06744826e+00 2.69673288e-01 -6.38583601e-01 -3.81867945e-01
-1.53962076e+00 4.86087710e-01 8.79594743e-01 2.29508638e-01
-5.89807391e-01 9.84863579e-01 -1.02764636e-01 4.59003925e-01
5.20228408e-02 -4.35257424e-03 3.17920372e-03 7.90695846e-02
-1.17170978e+00 -1.25451297e-01 6.41464293e-01 8.55545521e-01
3.84265542e-01 3.74797255e-01 2.91595459e-01 5.68755567e-01
2.56157577e-01 9.55321968e-01 1.97158724e-01 -1.66786242e+00
2.13777974e-01 -1.53445497e-01 4.23483729e-01 -1.16565156e+00
-4.21008505e-02 -7.46500909e-01 -1.09176862e+00 4.12933409e-01
3.36040296e-02 6.53277530e-05 -8.08383286e-01 1.56472790e+00
5.98946735e-02 4.58007485e-01 1.45293966e-01 1.26725090e+00
4.41753000e-01 5.38330197e-01 -3.15496624e-01 -2.48934373e-01
1.01413393e+00 -5.08195519e-01 -7.34490097e-01 -2.61871427e-01
-1.43049106e-01 -8.40387285e-01 7.25359678e-01 7.23666430e-01
-1.37645602e+00 -6.09983921e-01 -1.55804229e+00 5.64660542e-02
2.52621055e-01 2.17771024e-01 2.89176702e-01 5.28434157e-01
-1.41357183e+00 8.85737181e-01 -1.01651990e+00 -3.02133441e-01
4.33774650e-01 1.02282412e-01 -3.21826577e-01 -6.54702127e-01
-8.31892490e-01 6.71306849e-01 1.34554088e-01 5.54656982e-02
-1.34886742e+00 -6.97143972e-01 -6.37141168e-01 7.10449144e-02
3.22107762e-01 -5.67322075e-01 1.19814730e+00 -1.09635353e+00
-1.48397541e+00 5.57350397e-01 1.91795379e-01 -8.97613227e-01
5.34609795e-01 -2.35265091e-01 -4.91539598e-01 7.71022260e-01
8.27325433e-02 7.89963245e-01 1.20422363e+00 -1.55816782e+00
-4.80634838e-01 -9.87041667e-02 7.49368453e-03 -2.85183728e-01
4.01582345e-02 -7.18635172e-02 -4.66467023e-01 -5.98025918e-01
5.83455801e-01 -9.44105744e-01 -2.56979555e-01 6.76606178e-01
-3.16969872e-01 6.60453439e-01 7.62405872e-01 -7.20339835e-01
4.94546235e-01 -2.12998176e+00 1.59851174e-04 1.27947450e-01
2.82183826e-01 1.83425605e-01 -2.80922323e-01 3.23920637e-01
-2.02660203e-01 -3.30919325e-01 -5.81031203e-01 -5.34566641e-01
-2.58388132e-01 8.72754902e-02 -3.66384417e-01 1.04710865e+00
-9.71191600e-02 9.62830067e-01 -8.25010717e-01 -8.52603689e-02
6.56165242e-01 7.37452388e-01 -2.82726318e-01 4.91720326e-02
-4.19510305e-02 5.00501215e-01 -3.68344374e-02 8.80489349e-01
9.06369150e-01 -5.33187389e-01 2.13535801e-01 -5.40324867e-01
-1.06004350e-01 -2.11550027e-01 -1.21820843e+00 1.97654152e+00
-4.99239236e-01 1.03517139e+00 5.59785247e-01 -1.12934947e+00
5.52420676e-01 2.07465395e-01 6.91464722e-01 -1.02311754e+00
2.60090351e-01 3.68154407e-01 -3.93044442e-01 -3.09421718e-01
4.62834954e-01 -1.37168258e-01 9.38316658e-02 5.66699684e-01
-9.06151682e-02 -1.49039561e-02 -2.38484874e-01 3.99600297e-01
1.44366932e+00 -7.89443254e-02 1.89814880e-01 -2.33005568e-01
3.11575919e-01 -6.22490048e-02 2.32225448e-01 8.55730295e-01
-1.13793112e-01 9.90784705e-01 -1.84305489e-01 -3.94522041e-01
-1.34453869e+00 -1.33006740e+00 -9.16993171e-02 4.98190314e-01
5.06565869e-01 -1.54491842e-01 -5.72289109e-01 5.87482974e-02
-1.97218940e-01 4.58947152e-01 -4.52847064e-01 8.44568983e-02
-5.27782500e-01 -1.24655731e-01 5.81002176e-01 5.42601943e-01
9.46632504e-01 -8.39468360e-01 -1.07559323e+00 2.65688449e-01
-3.62540811e-01 -1.65838540e+00 -2.85144240e-01 -4.08253521e-02
-7.28070080e-01 -1.12174225e+00 -7.35009193e-01 -6.76625609e-01
5.95614970e-01 6.99619889e-01 1.07999015e+00 1.67153791e-01
-1.97620988e-01 7.80013680e-01 -4.58992779e-01 -7.56958872e-02
-5.59754133e-01 -7.13670254e-01 -3.71291004e-02 2.18238220e-01
-2.61067331e-01 -8.69838417e-01 -9.74973440e-01 1.38368458e-01
-1.32638741e+00 6.73252270e-02 6.49190545e-01 5.56894660e-01
6.56796336e-01 9.47995409e-02 4.19714674e-03 -3.59743923e-01
3.48333806e-01 -3.01491678e-01 -8.70982528e-01 7.84430429e-02
-5.97610116e-01 1.26909107e-01 7.49966562e-01 -4.93417591e-01
-5.91913700e-01 6.27113521e-01 -6.65967613e-02 -7.87128091e-01
6.98317811e-02 3.82930309e-01 3.58078808e-01 -5.03229320e-01
7.06869423e-01 6.17432177e-01 2.23097160e-01 -2.84710199e-01
4.56442952e-01 5.54565847e-01 1.29296958e+00 -1.21642388e-01
1.02836370e+00 1.05912244e+00 2.84357578e-01 -6.35798454e-01
-6.97060466e-01 -1.43637061e-01 -2.72978514e-01 -5.73239207e-01
6.19814456e-01 -1.27710772e+00 -9.14952576e-01 4.92119610e-01
-9.79108870e-01 -4.54699099e-01 -2.92436689e-01 5.04217207e-01
-7.44150519e-01 5.25496066e-01 -7.38247335e-01 -6.42415762e-01
-4.21934456e-01 -1.08274961e+00 1.18800104e+00 -2.89176814e-02
4.43802953e-01 -7.24304199e-01 -2.02993914e-01 2.99877256e-01
8.39475214e-01 4.34001952e-01 4.13857251e-02 2.86039859e-01
-1.16670871e+00 -1.63840011e-01 -4.31438744e-01 5.74455023e-01
-2.46008188e-01 -4.53526735e-01 -8.62929702e-01 -6.76819980e-01
2.14184180e-01 -4.65456009e-01 9.90480483e-01 5.72190166e-01
1.17036211e+00 -2.27745757e-01 -2.71779094e-02 1.10841000e+00
1.94223487e+00 -1.62088409e-01 9.91359591e-01 2.19671637e-01
4.88406181e-01 -5.48361428e-02 4.35469091e-01 3.58847171e-01
7.24713504e-02 4.15127426e-01 8.84623051e-01 -5.17659076e-02
-4.91241664e-01 -3.08615148e-01 5.19337475e-01 5.98948121e-01
1.16353810e-01 -3.31212491e-01 -4.62957293e-01 5.89984655e-01
-1.60640717e+00 -9.13208902e-01 -6.74637407e-02 2.09329796e+00
4.90569115e-01 -1.83429673e-01 -4.35658962e-01 3.97371858e-01
5.15943587e-01 2.57413954e-01 -5.76831818e-01 3.08206320e-01
-3.40835243e-01 3.54240835e-01 1.08947527e+00 4.58974868e-01
-9.33985054e-01 4.89445210e-01 6.64631748e+00 5.42909563e-01
-1.36983645e+00 2.56436348e-01 3.71617019e-01 -1.73484430e-01
-2.12942064e-01 -9.69750360e-02 6.86674239e-03 4.95346844e-01
1.04417729e+00 3.09734017e-01 9.00734007e-01 6.21464431e-01
4.77543138e-02 -1.82865500e-01 -9.96636271e-01 1.50144494e+00
3.28392744e-01 -1.89713109e+00 -1.75974831e-01 1.42299309e-01
6.72231734e-01 4.67464596e-01 1.31723940e-01 -2.94318676e-01
5.71594059e-01 -1.09337687e+00 1.12897050e+00 4.73657072e-01
1.17638886e+00 -5.10915041e-01 5.67240655e-01 3.27556759e-01
-1.06705594e+00 -6.36335909e-02 -5.09423554e-01 7.92530775e-02
3.93361568e-01 7.96063483e-01 -3.98176223e-01 4.25843477e-01
8.78767431e-01 9.22741294e-01 -3.73151422e-01 8.81708443e-01
-3.60564142e-02 6.11509144e-01 -5.48994422e-01 4.11378980e-01
1.93741933e-01 1.22854717e-01 4.49486703e-01 8.40744793e-01
7.91700780e-01 3.46008837e-01 -2.39037983e-02 9.74909425e-01
-2.59856045e-01 -7.57873833e-01 -6.24835312e-01 3.86251539e-01
4.43270922e-01 1.01930475e+00 -5.55302143e-01 -2.14031681e-01
-6.72138259e-02 1.26738477e+00 -1.48862317e-01 3.72283041e-01
-7.80769289e-01 -3.76529582e-02 7.10689351e-02 2.12014318e-01
6.56493485e-01 -3.19696754e-01 -1.02908172e-01 -1.18853700e+00
1.98296353e-01 -8.19748521e-01 -1.51702926e-01 -1.28529596e+00
-9.36110497e-01 3.59317809e-01 -3.33630800e-01 -1.55842531e+00
-2.62122780e-01 -5.39482772e-01 -3.14561337e-01 4.77399081e-01
-1.64089501e+00 -1.02162635e+00 -8.07315290e-01 7.86664963e-01
3.89293879e-01 -2.94949673e-02 5.33418119e-01 4.18727040e-01
-4.33105789e-03 3.13782156e-01 3.82465839e-01 1.73095375e-01
2.87662089e-01 -7.96236694e-01 3.63874406e-01 1.33811009e+00
1.57244086e-01 4.60217953e-01 9.77884650e-01 -3.65459591e-01
-2.06262016e+00 -1.13400948e+00 2.97672808e-01 4.88184690e-02
5.43759108e-01 -1.17493778e-01 -4.69321966e-01 7.27000594e-01
3.36449057e-01 7.00310528e-01 9.09549370e-02 -8.74486089e-01
-3.37016404e-01 -2.55695939e-01 -1.44137645e+00 3.03560823e-01
1.07459998e+00 -5.99090636e-01 -1.21368676e-01 2.61652499e-01
7.40298152e-01 -4.07076627e-01 -8.06751549e-01 3.51108402e-01
7.02024877e-01 -1.28166175e+00 1.41819358e+00 3.91868681e-01
6.25114620e-01 -5.28831303e-01 -7.49338388e-01 -1.06841457e+00
-2.08614185e-01 -6.96090519e-01 -1.81673005e-01 5.89469016e-01
-2.46336907e-02 -4.17753458e-01 7.55069196e-01 -1.23822289e-02
1.24493856e-02 -4.31853652e-01 -1.01452744e+00 -7.10101545e-01
-3.67043018e-01 -8.94472659e-01 3.85710597e-01 7.55595267e-01
-4.80166912e-01 -3.81731652e-02 -7.25178063e-01 4.71171468e-01
1.46866834e+00 1.33287355e-01 5.52891195e-01 -7.57134795e-01
-4.53354061e-01 1.12815008e-01 -4.88928080e-01 -1.52836740e+00
-2.70226300e-01 -4.75552708e-01 3.87286574e-01 -1.25280488e+00
1.69260465e-02 -4.31548357e-01 -2.39324585e-01 2.07228251e-02
3.66422266e-01 7.96715200e-01 2.80362636e-01 4.36741769e-01
-7.91372895e-01 4.67844963e-01 1.03635204e+00 -2.80837119e-01
5.05689442e-01 -3.20130438e-01 -5.29697955e-01 3.04802954e-01
5.69114506e-01 -5.21962464e-01 -2.86605269e-01 -7.28773415e-01
1.41265392e-01 6.38228595e-01 8.89451504e-01 -1.50202692e+00
5.63915670e-01 1.88675702e-01 6.69584811e-01 -6.39918387e-01
5.06898165e-01 -1.10827124e+00 6.21492803e-01 7.75044024e-01
-4.27680045e-01 -1.52021274e-01 -1.25269562e-01 8.39959025e-01
-3.96176055e-02 1.30839646e-02 9.84018803e-01 -2.39599407e-01
-5.74170470e-01 1.14459813e-01 -2.99293473e-02 -4.13464546e-01
7.43559778e-01 -2.18731895e-01 -2.81063110e-01 -8.95167768e-01
-4.02096093e-01 -1.75302118e-01 7.78337955e-01 8.10937397e-03
1.02344060e+00 -1.43239295e+00 -7.38150775e-01 3.94676387e-01
-1.13135934e-01 -1.83129430e-01 2.25406975e-01 8.16952169e-01
-1.09179640e+00 2.76375324e-01 -2.68922925e-01 -9.75370705e-01
-9.61222351e-01 5.14300525e-01 4.15243268e-01 1.89890504e-01
-9.26936686e-01 5.48053384e-01 -4.25916135e-01 -2.24205047e-01
3.50081205e-01 -4.81780827e-01 4.78628129e-01 -7.83428371e-01
7.65516043e-01 2.57155538e-01 -3.73158380e-02 -6.31172359e-01
-3.60779494e-01 6.42867982e-01 5.35969436e-01 -4.25964952e-01
1.60257375e+00 -2.59551227e-01 1.38558134e-01 -1.15627078e-02
1.32787585e+00 -4.39956114e-02 -1.80876231e+00 -3.00934792e-01
-4.43609387e-01 -6.59331203e-01 4.65547353e-01 -7.51133680e-01
-1.64700389e+00 5.19925058e-01 1.01347625e+00 1.90408468e-01
1.46504319e+00 4.64690514e-02 1.00792599e+00 3.38635057e-01
6.05633676e-01 -5.76767743e-01 1.04652643e-01 -6.90354854e-02
9.15755093e-01 -1.22889686e+00 2.64009356e-01 -1.82344347e-01
-9.73070934e-02 1.00003481e+00 -2.65336335e-01 -5.51892698e-01
5.16732037e-01 5.23892820e-01 3.62041481e-02 -2.22147629e-01
-5.23692548e-01 5.99453197e-05 -9.48025510e-02 6.94279194e-01
-9.26062763e-02 6.32058308e-02 1.74760193e-01 -2.20831469e-01
-5.66651113e-02 3.51924270e-01 8.03641379e-01 9.50155735e-01
-4.88188267e-01 -5.82342684e-01 -6.66510701e-01 7.04789981e-02
-5.45091748e-01 2.21541412e-02 -9.85902082e-03 5.52221656e-01
-1.12812273e-01 1.17720020e+00 2.78491173e-02 -5.13093412e-01
8.53704661e-02 -7.44463086e-01 6.81104898e-01 1.45098850e-01
-1.84394315e-01 4.83864453e-03 -1.18923306e-01 -1.13270009e+00
-8.15169513e-01 -5.24913073e-01 -1.07788157e+00 -6.65839732e-01
1.47528546e-02 -4.07427371e-01 1.10674548e+00 7.37585604e-01
2.97816783e-01 4.80744779e-01 7.48400390e-01 -1.11241674e+00
-6.06672823e-01 -6.85712755e-01 -5.72345674e-01 4.94881243e-01
7.83022463e-01 -1.72832951e-01 -4.18034136e-01 2.62750983e-01] | [10.977230072021484, -2.2230710983276367] |
d4e0f59e-7813-4581-9d58-29a2a1d5fe8a | transformers-are-deep-infinite-dimensional-1 | 2106.01506 | null | https://arxiv.org/abs/2106.01506v1 | https://arxiv.org/pdf/2106.01506v1.pdf | Transformers are Deep Infinite-Dimensional Non-Mercer Binary Kernel Machines | Despite their ubiquity in core AI fields like natural language processing, the mechanics of deep attention-based neural networks like the Transformer model are not fully understood. In this article, we present a new perspective towards understanding how Transformers work. In particular, we show that the "dot-product attention" that is the core of the Transformer's operation can be characterized as a kernel learning method on a pair of Banach spaces. In particular, the Transformer's kernel is characterized as having an infinite feature dimension. Along the way we consider an extension of the standard kernel learning problem to a binary setting, where data come from two input domains and a response is defined for every cross-domain pair. We prove a new representer theorem for these binary kernel machines with non-Mercer (indefinite, asymmetric) kernels (implying that the functions learned are elements of reproducing kernel Banach spaces rather than Hilbert spaces), and also prove a new universal approximation theorem showing that the Transformer calculation can learn any binary non-Mercer reproducing kernel Banach space pair. We experiment with new kernels in Transformers, and obtain results that suggest the infinite dimensionality of the standard Transformer kernel is partially responsible for its performance. This paper's results provide a new theoretical understanding of a very important but poorly understood model in modern machine~learning. | ['Joseph E. Gonzalez', 'Matthew A. Wright'] | 2021-06-02 | transformers-are-deep-infinite-dimensional | https://openreview.net/forum?id=AVKFuhH1Fo4 | https://openreview.net/pdf?id=AVKFuhH1Fo4 | null | ['deep-attention', 'deep-attention'] | ['computer-vision', 'natural-language-processing'] | [ 1.53620895e-02 2.40131676e-01 -7.60823265e-02 -5.24317741e-01
-2.40512297e-01 -5.73394537e-01 4.49088991e-01 -7.46693090e-02
-4.59527731e-01 5.39087415e-01 3.13249156e-02 -6.73327744e-01
-4.58473176e-01 -7.57151008e-01 -7.50169396e-01 -9.76383686e-01
-4.26839978e-01 4.27592903e-01 -5.67178093e-02 -3.07715029e-01
1.20795190e-01 5.66881955e-01 -1.31218660e+00 6.46481961e-02
1.04227030e+00 1.05931711e+00 -4.92499992e-02 9.01081324e-01
1.63736916e-03 7.89033234e-01 -1.97840497e-01 -3.35827202e-01
2.42110804e-01 -1.91837192e-01 -1.19648564e+00 -2.79214740e-01
2.58983582e-01 1.06083378e-01 -4.19605762e-01 1.34775460e+00
1.42154684e-02 2.36587211e-01 1.32356012e+00 -1.53489208e+00
-1.38265836e+00 6.12244904e-01 -7.05999210e-02 1.53898343e-01
4.30082381e-02 -2.15500951e-01 1.15130270e+00 -8.46616685e-01
-4.69587408e-02 1.09532392e+00 1.09520757e+00 6.25745952e-01
-1.67660356e+00 -3.07663411e-01 -5.65218866e-01 1.78048581e-01
-1.29755688e+00 -6.90404847e-02 5.88236213e-01 -6.29553378e-01
8.95821393e-01 4.41270441e-01 6.01517797e-01 7.91444242e-01
5.16845345e-01 7.44985044e-01 1.29828155e+00 -7.25001335e-01
1.42033577e-01 6.91422105e-01 3.76963884e-01 9.53083277e-01
6.37415648e-02 2.89654564e-02 -7.00554624e-02 -1.54997006e-01
9.19987977e-01 3.80144231e-02 -6.28261626e-01 -8.97311628e-01
-1.30552769e+00 1.20962775e+00 4.20984417e-01 6.85192406e-01
-7.27255344e-02 2.14805290e-01 4.46703672e-01 6.81686282e-01
4.92127687e-01 4.39678967e-01 -4.29486811e-01 9.73680094e-02
-2.50044882e-01 9.81692970e-02 1.32801533e+00 9.17280495e-01
7.23437726e-01 -9.54332203e-02 1.95007563e-01 6.46201253e-01
3.84379253e-02 4.35590148e-01 6.10800385e-01 -6.55333102e-01
7.31877908e-02 1.73466489e-01 3.53945680e-02 -5.55312991e-01
-4.88595754e-01 -3.09144020e-01 -9.23568010e-01 2.77305603e-01
6.46279037e-01 2.76223511e-01 -3.43806654e-01 2.04200745e+00
-2.63471872e-01 2.45392062e-02 3.37508798e-01 7.54316807e-01
7.29977563e-02 7.18139529e-01 -3.05651743e-02 1.34756371e-01
1.17234182e+00 -6.17351055e-01 -4.98620838e-01 2.68167287e-01
9.34867263e-01 -3.54004353e-01 1.33474803e+00 3.73070210e-01
-1.00466037e+00 -3.08651268e-01 -1.22575116e+00 -2.78228462e-01
-8.21656525e-01 2.22860393e-03 1.15029263e+00 5.88920891e-01
-1.22244918e+00 1.00206006e+00 -2.42682710e-01 -7.16034174e-01
1.42838195e-01 4.45016384e-01 -7.93232441e-01 1.74806118e-01
-1.42251587e+00 1.28394628e+00 4.30622667e-01 2.05355994e-02
-6.64195299e-01 -6.53413713e-01 -9.85583842e-01 2.68212348e-01
-2.25603089e-01 -5.14117062e-01 1.38782716e+00 -1.14358604e+00
-1.34888089e+00 1.05938613e+00 3.59748483e-01 -6.48505151e-01
2.05118611e-01 2.01173164e-02 -3.21631223e-01 1.82089454e-03
-7.36034736e-02 2.15273008e-01 9.93610322e-01 -5.65750301e-01
-3.69979560e-01 -1.82444036e-01 3.38700861e-01 -8.90315697e-02
-6.40237510e-01 -2.11692259e-01 6.73857450e-01 -5.76766551e-01
-2.03543231e-01 -7.35701680e-01 1.17191792e-01 -1.78989470e-02
-7.07278326e-02 -6.70838714e-01 6.46864891e-01 -2.11533904e-01
9.54809427e-01 -2.28479838e+00 3.69573593e-01 2.43881553e-01
4.43474382e-01 -1.57913059e-01 1.22252898e-02 5.00098884e-01
-7.12265670e-01 1.08174853e-01 -4.70832646e-01 1.44134061e-02
5.79432845e-01 1.59139857e-01 -5.99985957e-01 1.01932526e+00
1.87904105e-01 9.43559825e-01 -8.07320893e-01 -2.40670979e-01
5.05399816e-02 3.65933925e-01 -4.29691970e-01 1.32598087e-01
1.01451792e-01 -1.97152421e-01 -2.09259003e-01 2.73474902e-01
5.32140255e-01 -1.75508097e-01 -4.12165493e-01 -2.64315993e-01
-6.68354705e-02 -3.06257699e-02 -7.49707341e-01 1.55025685e+00
-4.41324890e-01 9.38182712e-01 1.43265441e-01 -1.82004035e+00
5.94787538e-01 4.04489905e-01 1.78648934e-01 -2.19358638e-01
1.22034930e-01 3.53532284e-01 1.98152289e-02 -5.15101612e-01
1.59824312e-01 -7.81664073e-01 -3.02883059e-01 4.96428102e-01
3.54373425e-01 -1.34966165e-01 -2.36531436e-01 -3.47042978e-02
1.05249095e+00 -3.10588241e-01 1.58880457e-01 -1.14452124e+00
5.84982514e-01 -8.03861246e-02 1.90150794e-02 8.64793241e-01
-2.76687175e-01 8.66742134e-02 7.54844964e-01 -4.70279306e-01
-1.33707428e+00 -1.72956216e+00 -5.27227044e-01 1.13620448e+00
-2.43676782e-01 2.85006575e-02 -7.57943213e-01 -3.14275831e-01
2.83330411e-01 7.66584396e-01 -1.14824951e+00 -8.09771001e-01
-1.77682325e-01 -5.28012931e-01 9.82609153e-01 6.80310130e-01
5.05083621e-01 -8.08991730e-01 -4.00532931e-01 -8.08083639e-02
1.73706025e-01 -6.87751591e-01 -6.76376522e-01 8.73853564e-01
-9.16629612e-01 -9.84510601e-01 -9.60835814e-01 -1.06172931e+00
3.82965446e-01 -1.92155555e-01 9.67138767e-01 -5.25272310e-01
-4.34155345e-01 7.40584016e-01 8.89110565e-02 -4.49342012e-01
-5.00097692e-01 -1.33325964e-01 5.95422506e-01 3.41412309e-03
6.36063576e-01 -6.79561615e-01 -2.01586977e-01 1.92254484e-02
-1.14704728e+00 -1.67256206e-01 5.24406254e-01 1.13904738e+00
-2.24001825e-01 1.42132014e-01 5.94648421e-01 -3.84830087e-01
9.79618192e-01 -5.55944562e-01 -6.09454811e-01 3.37507457e-01
-4.12331790e-01 6.78460598e-01 8.94468963e-01 -7.69144177e-01
-7.10768461e-01 -2.14198351e-01 4.31761503e-01 -4.84481514e-01
7.57080466e-02 5.13561189e-01 -2.88450513e-02 -2.68009782e-01
7.60935605e-01 4.46392089e-01 2.92937517e-01 -3.28817487e-01
3.84636402e-01 7.54156530e-01 6.29925132e-01 -9.02417421e-01
7.57593334e-01 4.16153729e-01 1.46484166e-01 -9.43096220e-01
-6.41062200e-01 -4.26303208e-01 -5.32243311e-01 2.47659817e-01
1.06959009e+00 -4.34075445e-01 -1.12612987e+00 4.65578437e-01
-1.04663563e+00 -4.28688884e-01 -6.90532982e-01 8.99306238e-01
-1.31548250e+00 2.49755099e-01 -9.90120471e-01 -1.06913960e+00
-1.02928169e-01 -8.66045177e-01 5.96082389e-01 -7.85264149e-02
1.40683604e-02 -1.52839923e+00 1.63734987e-01 -2.32123703e-01
6.79994881e-01 -3.90623033e-01 1.58311701e+00 -9.15187776e-01
-3.63266915e-02 -1.79083541e-01 -3.76970679e-01 8.79059792e-01
-1.53758779e-01 -5.01471996e-01 -1.08023214e+00 1.25222215e-02
4.93263364e-01 -3.26948404e-01 8.42746377e-01 2.03734264e-01
1.01315367e+00 -1.20294251e-01 5.16531952e-02 7.37771928e-01
1.44508934e+00 3.09138149e-02 4.69379932e-01 1.43805444e-01
4.55181777e-01 5.43250561e-01 1.14139616e-01 4.84758765e-02
6.12171553e-02 2.25302354e-01 2.25591570e-01 6.70055524e-02
7.30524004e-01 1.27143890e-01 5.56110740e-01 7.12783992e-01
-1.13510050e-01 3.63017768e-01 -7.28453755e-01 5.68506718e-01
-1.88069916e+00 -1.10168672e+00 1.10918581e-01 2.50403905e+00
9.68256712e-01 -8.02100673e-02 1.33719474e-01 8.03422704e-02
7.76066303e-01 -3.57780784e-01 -3.40483874e-01 -9.40827668e-01
-2.55223811e-01 7.07229853e-01 6.73810244e-01 5.62827229e-01
-1.25341427e+00 7.33044207e-01 6.80878401e+00 7.34382570e-01
-1.05445898e+00 1.23526976e-01 1.80059597e-01 4.24411982e-01
-1.61396667e-01 -1.09469876e-01 -7.85744935e-02 4.11965176e-02
1.05914319e+00 -5.10097504e-01 6.91781282e-01 1.21461821e+00
-3.73777598e-01 7.16307908e-02 -1.70409775e+00 1.17171979e+00
-6.68742657e-02 -1.14902496e+00 -2.12986946e-01 3.53402883e-01
7.79643506e-02 -1.12194210e-01 3.09360445e-01 5.81223667e-01
1.85442895e-01 -1.44474781e+00 5.45650005e-01 6.86594725e-01
7.78904021e-01 -9.74797964e-01 5.28003275e-01 3.69215786e-01
-1.11943626e+00 -2.22514093e-01 -7.53597260e-01 -3.08939904e-01
-5.45138657e-01 5.03507853e-01 -6.39811695e-01 7.07566068e-02
5.49680531e-01 6.01074517e-01 -6.00137413e-02 8.17035258e-01
3.90800029e-01 3.39326233e-01 -2.83694595e-01 -1.96691588e-01
3.92015070e-01 -5.29046357e-01 3.20375144e-01 1.42756391e+00
3.45240355e-01 6.48345947e-02 -2.80112922e-01 1.37945116e+00
2.20588613e-02 5.23138903e-02 -1.20571232e+00 -2.66299993e-01
-1.64830714e-01 9.67876971e-01 -2.89160579e-01 -3.80308181e-01
-5.30136347e-01 1.42612374e+00 5.56439400e-01 3.89215708e-01
-9.74714696e-01 -9.07849371e-01 1.10429645e+00 -1.06434792e-01
1.79095834e-01 -2.34123021e-02 -7.05083460e-02 -1.49781728e+00
-3.59292589e-02 -2.62846947e-01 1.13354750e-01 -7.78815567e-01
-1.70998609e+00 2.06762210e-01 2.46607140e-01 -9.36632574e-01
-5.64138442e-02 -1.33258355e+00 -6.33168578e-01 1.26752889e+00
-1.04697490e+00 -8.05401146e-01 3.77600700e-01 1.09964776e+00
-7.77419209e-02 -6.37176726e-03 1.19395018e+00 9.80676562e-02
-7.15845153e-02 4.07138288e-01 3.74954611e-01 3.30217868e-01
4.15703535e-01 -1.95438111e+00 2.42640302e-01 2.95249254e-01
2.29269732e-02 8.76237333e-01 6.19277358e-01 -8.61548185e-02
-1.70845056e+00 -6.59716308e-01 8.70002925e-01 -4.79389131e-01
1.20852292e+00 -5.48975646e-01 -9.21453953e-01 8.89737666e-01
8.45328942e-02 2.79806554e-01 7.44128227e-01 1.17892072e-01
-7.51445651e-01 -1.55495942e-01 -1.18903494e+00 4.27601427e-01
7.78027594e-01 -1.16122806e+00 -9.15221334e-01 4.55058843e-01
5.82872748e-01 1.34541169e-01 -1.08239555e+00 2.66421199e-01
6.84535861e-01 -9.66376841e-01 8.72152865e-01 -1.07657933e+00
4.12194207e-02 -7.16394782e-02 -3.77165049e-01 -1.35465038e+00
-5.34047723e-01 -4.62641954e-01 3.33786868e-02 4.84785527e-01
2.08841860e-01 -1.18145990e+00 2.63046831e-01 4.86975223e-01
-1.43791482e-01 -8.35433602e-01 -1.16418493e+00 -1.02817702e+00
6.69055641e-01 -6.20826960e-01 1.93785265e-01 1.25924027e+00
8.47042382e-01 3.61791879e-01 -2.66382337e-01 -5.84120862e-02
8.12280059e-01 -3.56050320e-02 1.47400752e-01 -1.55484927e+00
-4.46811646e-01 -6.86894178e-01 -9.10395145e-01 -8.12819719e-01
4.35760826e-01 -1.31156301e+00 4.00161110e-02 -8.59275043e-01
2.07927331e-01 -4.03965861e-01 -5.58296859e-01 2.38119155e-01
5.53293169e-01 -7.86358640e-02 -1.37456700e-01 1.10479845e-02
-1.41510472e-01 3.72171015e-01 7.50220001e-01 -1.22031629e-01
1.56190515e-01 4.26413156e-02 -5.85560799e-01 8.95139158e-01
8.04353774e-01 -1.55762732e-01 -5.47735333e-01 -1.68029852e-02
2.80342221e-01 -1.33220509e-01 7.42168427e-01 -1.04398155e+00
3.09108466e-01 -4.13709320e-02 3.87685299e-01 9.45492089e-02
2.55882412e-01 -9.22274053e-01 -2.53733039e-01 4.22692686e-01
-4.94608641e-01 1.55412018e-01 1.76315811e-02 4.64073777e-01
1.59273461e-01 -4.14014608e-01 8.83325815e-01 -1.76159188e-01
-4.71303672e-01 1.21859752e-01 -4.41228300e-01 1.72970667e-01
9.53111470e-01 -2.45431215e-01 -1.08831622e-01 -2.39895239e-01
-9.43424642e-01 -2.67165720e-01 4.79454100e-01 4.82939146e-02
6.04680181e-01 -1.40051091e+00 -4.46566343e-01 4.17793930e-01
1.90269873e-01 -3.70867252e-01 -7.56618530e-02 1.11236715e+00
-3.99327248e-01 6.48679376e-01 -2.02670574e-01 -4.70979184e-01
-6.27980709e-01 8.30343008e-01 7.37971187e-01 -9.02664363e-02
-5.84614396e-01 9.04709876e-01 6.38493538e-01 -4.51291502e-01
2.76224911e-01 -8.68612170e-01 2.63312370e-01 -2.06944168e-01
4.11913872e-01 1.44642293e-01 -1.85200676e-01 -4.21259820e-01
-2.24412471e-01 4.00290906e-01 -4.67770398e-02 -2.95940608e-01
1.19410896e+00 3.17258984e-01 -6.05226099e-01 1.04245198e+00
1.75400543e+00 -2.33810946e-01 -6.43065274e-01 -1.32916152e-01
-1.24721797e-02 1.40330372e-02 -3.05955052e-01 -4.09227729e-01
-5.16595244e-01 1.03974199e+00 4.54337984e-01 6.75339997e-01
9.77608263e-01 4.41492051e-01 4.72979188e-01 7.82577276e-01
2.20590785e-01 -1.04847705e+00 -2.69977421e-01 6.64784193e-01
9.77247894e-01 -1.00395942e+00 -5.12740016e-01 -4.56406772e-02
-2.74783462e-01 1.57266486e+00 1.97981119e-01 -4.42064941e-01
1.02391732e+00 2.24016950e-01 -4.26816940e-01 -2.01230459e-02
-6.27265871e-01 -1.38206884e-01 3.21207166e-01 8.52842629e-01
4.63240951e-01 1.87124446e-01 9.63720381e-02 5.93957126e-01
-2.44965702e-01 8.04154277e-02 5.02046168e-01 7.23605096e-01
-5.91539264e-01 -7.37617612e-01 -3.10106128e-01 5.00719368e-01
-3.08548570e-01 -3.90691906e-01 -1.13687240e-01 8.54007483e-01
-1.15647919e-01 4.25970584e-01 1.20876096e-01 -4.14547503e-01
8.73944238e-02 4.78469223e-01 8.31487715e-01 -5.05259752e-01
-1.35210738e-01 -5.04034162e-01 -4.17263210e-01 -3.47390383e-01
-2.17573076e-01 -2.07782716e-01 -1.15494704e+00 -6.17416024e-01
-2.57466763e-01 7.33130872e-01 6.92209125e-01 9.50814843e-01
-2.00753957e-01 7.19751343e-02 3.91723663e-01 -6.59483790e-01
-1.35460186e+00 -9.51335311e-01 -1.17208004e+00 2.46243894e-01
6.74735129e-01 -6.16273820e-01 -9.36543584e-01 -9.19554532e-02] | [7.53745698928833, 3.857351064682007] |
2e82b8dd-7386-45bd-9d9b-bff961a85b5b | inflated-3d-convolution-transformer-for | 2306.02548 | null | https://arxiv.org/abs/2306.02548v3 | https://arxiv.org/pdf/2306.02548v3.pdf | Inflated 3D Convolution-Transformer for Weakly-supervised Carotid Stenosis Grading with Ultrasound Videos | Localization of the narrowest position of the vessel and corresponding vessel and remnant vessel delineation in carotid ultrasound (US) are essential for carotid stenosis grading (CSG) in clinical practice. However, the pipeline is time-consuming and tough due to the ambiguous boundaries of plaque and temporal variation. To automatize this procedure, a large number of manual delineations are usually required, which is not only laborious but also not reliable given the annotation difficulty. In this study, we present the first video classification framework for automatic CSG. Our contribution is three-fold. First, to avoid the requirement of laborious and unreliable annotation, we propose a novel and effective video classification network for weakly-supervised CSG. Second, to ease the model training, we adopt an inflation strategy for the network, where pre-trained 2D convolution weights can be adapted into the 3D counterpart in our network for an effective warm start. Third, to enhance the feature discrimination of the video, we propose a novel attention-guided multi-dimension fusion (AMDF) transformer encoder to model and integrate global dependencies within and across spatial and temporal dimensions, where two lightweight cross-dimensional attention mechanisms are designed. Our approach is extensively validated on a large clinically collected carotid US video dataset, demonstrating state-of-the-art performance compared with strong competitors. | ['Dong Ni', 'Jie Ren', 'Jia Liu', 'Yuanji Zhang', 'Qilong Ying', 'Yuxin Zou', 'Xin Yang', 'Wufeng Xue', 'Yuhao Huang', 'Xinrui Zhou'] | 2023-06-05 | null | null | null | null | ['video-classification'] | ['computer-vision'] | [ 2.91795693e-02 -1.08138241e-01 -1.34052476e-02 -3.97867292e-01
-1.13040698e+00 -7.86000490e-01 2.40839005e-01 -1.61065817e-01
-3.44525695e-01 5.02178252e-01 1.62154779e-01 -7.72159159e-01
-9.29062963e-02 -3.49305809e-01 -4.90398198e-01 -7.28258550e-01
-2.74617940e-01 1.22516051e-01 4.26859677e-01 7.15674311e-02
-7.53208436e-03 4.76752967e-01 -1.30256331e+00 8.26769471e-02
9.35103416e-01 1.27712178e+00 3.30884069e-01 8.20826411e-01
-8.90227407e-03 4.88907605e-01 -4.28763002e-01 -4.95657206e-01
1.29802436e-01 -5.23797750e-01 -7.10748255e-01 2.31469527e-01
5.45896530e-01 -4.45886731e-01 -3.94160897e-01 1.05288494e+00
9.73768055e-01 1.84020412e-03 4.88041520e-01 -5.59103072e-01
-6.20312810e-01 3.91923457e-01 -5.12152791e-01 7.08806872e-01
-2.63685703e-01 2.43763015e-01 7.90504038e-01 -8.54539514e-01
5.83247602e-01 7.81722248e-01 5.77030838e-01 3.76155853e-01
-8.94109309e-01 -4.59286332e-01 3.76883268e-01 2.77021289e-01
-1.27709866e+00 -3.38595480e-01 7.41290569e-01 -5.21181166e-01
3.65689486e-01 2.66607136e-01 9.94131088e-01 1.02522695e+00
-3.10392622e-02 6.26089334e-01 8.96235824e-01 -1.95114106e-01
9.11864638e-02 -1.53336287e-01 1.06280677e-01 8.37149978e-01
7.16032833e-02 6.27263561e-02 1.39478862e-01 6.34633899e-02
1.10309041e+00 1.44089665e-02 -4.95300740e-01 -3.44920903e-01
-1.00984049e+00 7.78630376e-01 5.43910444e-01 5.00148475e-01
-2.83257008e-01 -3.03335730e-02 5.51429808e-01 -5.94849996e-02
4.35559303e-01 2.84483492e-01 -2.16089040e-01 -3.41413707e-01
-9.06570315e-01 -3.85404915e-01 4.61672753e-01 5.05408049e-01
1.28529936e-01 -8.71343613e-02 -2.08292484e-01 8.35831463e-01
1.08214483e-01 1.51294798e-01 8.60241175e-01 -8.52900028e-01
2.11753279e-01 3.43537182e-01 1.14010961e-03 -8.15751195e-01
-5.58725953e-01 -1.02513361e+00 -1.11033201e+00 2.84952492e-01
7.27083325e-01 -2.62063563e-01 -1.21855831e+00 1.57621408e+00
3.02990645e-01 7.54347205e-01 -1.07019968e-01 1.42407203e+00
9.81167018e-01 1.29713908e-01 2.56960720e-01 -2.12113112e-01
1.69235599e+00 -1.07277143e+00 -5.10858595e-01 -2.09868446e-01
8.39834988e-01 -6.00602150e-01 1.05541801e+00 -1.90657154e-02
-1.08124959e+00 -6.00131571e-01 -9.49547648e-01 2.90291533e-02
-8.18763599e-02 4.72713739e-01 7.17856705e-01 8.14677417e-01
-1.01030076e+00 4.22906101e-01 -1.06154346e+00 3.09879128e-02
6.01077080e-01 2.42395788e-01 -3.82091224e-01 1.37838408e-01
-1.29948974e+00 8.53678405e-01 2.97850892e-02 5.91260970e-01
-4.57895249e-01 -8.59779477e-01 -9.04533386e-01 8.58156383e-02
3.65575671e-01 -5.95025718e-01 1.19173980e+00 -7.73884237e-01
-1.63488364e+00 8.20492864e-01 -4.25381772e-02 -4.17220056e-01
6.85282290e-01 -1.00977942e-01 -4.75908697e-01 5.02305031e-01
-1.84112675e-02 2.71957695e-01 8.13647568e-01 -9.09168780e-01
-5.42220354e-01 -3.73912483e-01 1.51011780e-01 -1.70108676e-01
-1.92832857e-01 -1.40271977e-01 -8.02858412e-01 -9.53069091e-01
2.33062640e-01 -9.36518133e-01 -5.17832696e-01 1.60211161e-01
-2.73102503e-02 -9.69179720e-02 5.19614995e-01 -1.06275904e+00
1.37672150e+00 -2.42318273e+00 1.00502819e-01 3.12343508e-01
6.75093651e-01 6.87462866e-01 -9.26883072e-02 -5.92197657e-01
-1.51226088e-01 3.12173903e-01 -1.01010434e-01 -2.08103389e-01
-3.72546971e-01 1.97895127e-03 9.17494968e-02 4.61884439e-01
1.78331271e-01 9.93819118e-01 -8.61247241e-01 -6.22784019e-01
2.94726789e-01 5.87196648e-01 -5.86287141e-01 1.35976449e-01
2.97507644e-01 7.72363603e-01 -6.07532382e-01 6.54617906e-01
6.12431884e-01 -6.11148119e-01 1.88980356e-01 -6.80711091e-01
-2.25687712e-01 5.79388700e-02 -1.18120492e+00 1.88998961e+00
-7.58316934e-01 3.98287684e-01 -1.08220866e-02 -1.21140921e+00
6.49832249e-01 6.48536921e-01 6.52207911e-01 -6.32853866e-01
4.38600808e-01 5.21323979e-01 4.05292302e-01 -7.67808676e-01
-2.29550794e-01 1.35045841e-01 2.71986663e-01 3.22114795e-01
1.74096134e-02 2.35608920e-01 2.39107609e-01 -6.67008162e-02
8.91089678e-01 1.70842543e-01 2.60349005e-01 -1.79405212e-01
8.42252135e-01 -3.69815201e-01 6.77850544e-01 8.22257936e-01
-6.21800601e-01 8.61747742e-01 6.92676187e-01 -6.73411310e-01
-7.43401825e-01 -8.63609433e-01 -4.32022333e-01 6.70423329e-01
2.39035681e-01 -2.85916328e-01 -5.20143509e-01 -9.20675635e-01
-2.99648762e-01 6.94606006e-02 -7.22685158e-01 -7.44787231e-02
-1.00815308e+00 -7.61831403e-01 3.32427472e-01 9.09563422e-01
3.37845117e-01 -5.31012237e-01 -6.45063341e-01 4.80879039e-01
-2.19263211e-01 -1.25694942e+00 -7.31806099e-01 -2.06076413e-01
-8.03770244e-01 -1.12194264e+00 -1.11533022e+00 -9.56900358e-01
5.61504185e-01 1.64236501e-01 9.80532050e-01 2.05507576e-01
-5.68091393e-01 -5.59733342e-03 -3.59632492e-01 -7.55956396e-02
-1.74596056e-01 1.02086447e-01 -4.34617788e-01 1.58053920e-01
7.45947063e-02 -5.28053463e-01 -1.19929409e+00 3.70445728e-01
-4.55818057e-01 -1.97741222e-02 7.91221917e-01 1.01388943e+00
5.19435644e-01 -3.92332315e-01 7.65250385e-01 -5.67650199e-01
3.29830021e-01 -2.27095366e-01 -7.37580717e-01 3.42956722e-01
-2.08366692e-01 -1.09966472e-01 5.65697432e-01 -5.64768553e-01
-8.86986852e-01 1.31935760e-01 -4.04190987e-01 -5.10919392e-01
1.24777826e-02 5.80758333e-01 2.96305478e-01 -3.02593827e-01
5.06635606e-01 1.74247488e-01 4.67223555e-01 -4.85451043e-01
3.81961018e-01 7.99528241e-01 5.63872874e-01 -2.58195698e-01
4.04142469e-01 4.10511762e-01 -1.94692258e-02 -4.62638587e-01
-8.96741748e-01 -4.60122317e-01 -7.32053280e-01 -3.09586257e-01
1.03423750e+00 -7.31033146e-01 -6.37750685e-01 1.74089402e-01
-9.77212191e-01 -9.08804312e-03 -4.40918468e-02 7.54554808e-01
-2.04754174e-01 6.63564026e-01 -7.56538391e-01 -5.05448341e-01
-5.22027552e-01 -1.76931632e+00 1.07393265e+00 1.55765951e-01
2.45479286e-01 -7.54890203e-01 -3.28237623e-01 1.64439008e-01
7.59368598e-01 1.57929793e-01 8.00259531e-01 -5.03141701e-01
-5.42127967e-01 -3.22919935e-01 -5.34253180e-01 3.02822441e-01
2.73858815e-01 -1.14673920e-01 -7.77504623e-01 -2.28857845e-01
-2.08600163e-01 -1.55652002e-01 7.49772429e-01 8.15169930e-01
1.51274443e+00 1.79799303e-01 -2.53434986e-01 8.40711176e-01
9.73584831e-01 2.59083450e-01 4.04033422e-01 2.70022750e-01
5.35095990e-01 3.03363383e-01 3.26784372e-01 2.49189183e-01
3.44999254e-01 6.02096260e-01 2.58048534e-01 -5.06292343e-01
-4.59015518e-01 4.87692177e-01 -1.09765567e-01 6.03989244e-01
-5.15253007e-01 1.31006151e-01 -6.10319316e-01 3.31956208e-01
-1.53591037e+00 -7.70660698e-01 -3.57978456e-02 2.21817899e+00
8.62076759e-01 9.85904410e-03 1.82503223e-01 -1.39954343e-01
6.78059280e-01 4.87037562e-02 -4.11252677e-01 7.95224756e-02
5.55825513e-03 2.92348415e-01 2.90813059e-01 5.27298868e-01
-1.56235158e+00 5.93858600e-01 5.67230511e+00 5.88552356e-01
-1.61361051e+00 3.48224670e-01 9.36240017e-01 -7.47453049e-02
1.35169268e-01 -2.45288670e-01 -5.57115018e-01 6.16663694e-01
8.16627681e-01 1.11249797e-01 2.46543773e-02 6.84914589e-01
2.52774626e-01 3.60410452e-01 -6.68418109e-01 1.05973268e+00
-9.36508477e-02 -1.59596562e+00 -3.93906981e-01 -2.06940085e-01
3.17522526e-01 5.56487404e-02 -5.98113611e-02 2.68286377e-01
-2.02336162e-01 -6.11919820e-01 4.14351135e-01 2.73683846e-01
1.06147647e+00 -3.98563415e-01 9.14626539e-01 -3.03448319e-01
-1.34489489e+00 -1.14277057e-01 1.00170501e-01 5.36945224e-01
5.68207264e-01 4.84470397e-01 -3.66084248e-01 4.28806841e-01
7.10390210e-01 7.47189224e-01 -6.11744344e-01 1.29067099e+00
1.17625017e-02 5.34960866e-01 -1.95627227e-01 2.94659078e-01
3.27565372e-01 -2.30537355e-01 5.46185136e-01 1.11374438e+00
5.44044554e-01 1.54159337e-01 3.39727879e-01 5.15905023e-01
2.15543494e-01 1.78812385e-01 -1.61932725e-02 2.73381472e-01
2.18225181e-01 1.31273973e+00 -7.76331902e-01 -5.74981928e-01
-8.88801098e-01 7.59994924e-01 1.83071613e-01 4.65233684e-01
-1.05883336e+00 -2.65626162e-01 4.26381350e-01 1.86536968e-01
4.54273313e-01 -1.61854461e-01 -2.64861643e-01 -1.35337317e+00
2.54722893e-01 -6.67871296e-01 6.61265194e-01 -2.01682478e-01
-1.16657555e+00 1.00802350e+00 -3.99673432e-01 -1.58793509e+00
-2.21830085e-01 -6.34552360e-01 -4.74837482e-01 8.17141533e-01
-1.75042307e+00 -1.08788049e+00 -6.63512826e-01 3.58242780e-01
3.65151495e-01 -6.02290630e-02 7.94640064e-01 8.79275382e-01
-7.52795160e-01 7.26380348e-01 -3.03956032e-01 3.39978635e-01
7.50550270e-01 -1.20620692e+00 1.58631593e-01 9.90871131e-01
-3.31824690e-01 5.76440513e-01 2.53796846e-01 -3.36853623e-01
-9.81086612e-01 -1.10200727e+00 6.03395343e-01 -9.96814761e-03
8.26666057e-01 -6.68610558e-02 -8.79736066e-01 5.24752140e-01
-2.28752390e-01 7.21638024e-01 7.35251963e-01 -6.78626895e-02
-2.59049505e-01 -4.90824357e-02 -6.52331710e-01 5.61652303e-01
1.05230212e+00 -3.64650816e-01 -3.18534166e-01 3.54195297e-01
4.55447108e-01 -7.98799992e-01 -1.22972250e+00 5.06629109e-01
7.15056777e-01 -7.65317976e-01 1.24147677e+00 -6.49336457e-01
2.54691064e-01 -3.70980650e-01 2.59653151e-01 -1.01141334e+00
-5.49742877e-01 -6.01037621e-01 -1.23587243e-01 7.81764984e-01
3.88463765e-01 -7.07732022e-01 7.20169127e-01 3.95838231e-01
-6.31142974e-01 -8.96473229e-01 -1.20048296e+00 -5.03127277e-01
6.31023431e-03 -2.45698273e-01 3.78398895e-01 7.01226056e-01
-2.29893237e-01 1.74935777e-02 -2.96770364e-01 3.87890726e-01
5.12131155e-01 4.41462219e-01 2.67889500e-01 -1.19757450e+00
-4.49978024e-01 -8.86426866e-01 -6.02030337e-01 -1.11594844e+00
-1.56297222e-01 -8.74387741e-01 -2.72839069e-01 -1.29252493e+00
-7.36200251e-03 -4.87294704e-01 -4.57580686e-01 2.36193240e-01
-5.09196639e-01 5.19590259e-01 -9.52778235e-02 1.41807422e-01
-3.15466344e-01 3.08923215e-01 1.79037750e+00 -3.11799310e-02
-3.87863457e-01 2.54932523e-01 -6.76155388e-01 6.12870097e-01
6.23108685e-01 -2.41749659e-02 -3.44275028e-01 -3.95656437e-01
-3.70187014e-01 3.27727199e-01 6.18441880e-01 -7.48144805e-01
1.32801667e-01 2.98287779e-01 2.71840781e-01 -3.76379400e-01
1.79388419e-01 -6.91675901e-01 -3.48561466e-01 4.86569881e-01
-1.88094899e-01 -9.67648178e-02 -2.58182418e-02 4.49530154e-01
-3.79110694e-01 -1.92157015e-01 9.37652469e-01 -1.88557699e-01
-6.71868205e-01 5.97047627e-01 -3.80175978e-01 5.21271713e-02
9.14145708e-01 -5.59154637e-02 -2.24576443e-01 -2.21974149e-01
-1.29443514e+00 4.27121520e-02 -1.85312316e-01 3.94965351e-01
4.56327826e-01 -1.10957265e+00 -8.04489017e-01 3.18772584e-01
8.89388621e-02 -2.70725116e-02 6.75376534e-01 1.55963159e+00
-5.55830956e-01 2.86942542e-01 4.45279852e-02 -7.64760196e-01
-1.04465413e+00 3.95006657e-01 8.46056640e-01 -2.24829510e-01
-1.33395970e+00 6.57369196e-01 1.64032742e-01 1.26291066e-01
4.27278191e-01 -5.13192892e-01 -3.50018173e-01 -8.95057619e-02
8.17936778e-01 7.36270323e-02 2.82755524e-01 -4.95900035e-01
-2.38660276e-01 7.91870236e-01 -2.34677896e-01 3.46025854e-01
1.10149264e+00 -3.27856809e-01 1.35099605e-01 -3.11027199e-01
1.17477703e+00 -3.16396803e-01 -1.30074239e+00 -3.79981667e-01
-2.05731183e-01 -4.84967887e-01 5.50653636e-01 -6.35413885e-01
-1.58715749e+00 9.82634008e-01 1.06891215e+00 2.57315710e-02
1.15731382e+00 -5.67758046e-02 9.52995539e-01 -2.19744563e-01
-5.13905585e-02 -6.45353973e-01 -1.24842621e-01 7.13289827e-02
8.83633971e-01 -1.46004736e+00 -3.30515236e-01 -5.98187566e-01
-5.77725351e-01 1.41519761e+00 5.53432286e-01 1.21555075e-01
1.14647055e+00 2.30445817e-01 3.74251187e-01 -2.10339054e-01
-1.68472096e-01 -3.34618509e-01 5.42207539e-01 3.36243719e-01
6.48814738e-01 -6.62849620e-02 -5.11077881e-01 8.10707808e-01
1.78674296e-01 1.85863286e-01 3.60116959e-01 7.08075583e-01
-2.01333404e-01 -1.00451016e+00 8.34326446e-03 5.40516555e-01
-7.20924497e-01 -1.50118440e-01 5.04172504e-01 6.51193023e-01
2.84024817e-03 6.11352265e-01 6.15252107e-02 -1.64441168e-02
3.51061285e-01 -7.67167360e-02 2.48552233e-01 -2.15399250e-01
-3.09503764e-01 7.31351018e-01 6.81307688e-02 -6.18374765e-01
-3.46192002e-01 -5.98310947e-01 -1.13494468e+00 3.60347927e-01
-2.71404147e-01 3.50954644e-02 5.00796795e-01 1.10318100e+00
5.51485002e-01 7.86917329e-01 5.55952191e-01 -8.80419791e-01
-6.33860111e-01 -9.47627068e-01 -3.59372795e-01 5.61705709e-01
3.99453998e-01 -9.67743397e-01 -3.57355833e-01 2.11053833e-01] | [14.50469970703125, -2.457000255584717] |
5dd5f8b2-54ac-4b0d-8ec5-e156bdd518c7 | neural-koopman-pooling-control-inspired | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Wang_Neural_Koopman_Pooling_Control-Inspired_Temporal_Dynamics_Encoding_for_Skeleton-Based_Action_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Wang_Neural_Koopman_Pooling_Control-Inspired_Temporal_Dynamics_Encoding_for_Skeleton-Based_Action_CVPR_2023_paper.pdf | Neural Koopman Pooling: Control-Inspired Temporal Dynamics Encoding for Skeleton-Based Action Recognition | Skeleton-based human action recognition is becoming increasingly important in a variety of fields. Most existing works train a CNN or GCN based backbone to extract spatial-temporal features, and use temporal average/max pooling to aggregate the information. However, these pooling methods fail to capture high-order dynamics information. To address the problem, we propose a plug-and-play module called Koopman pooling, which is a parameterized high-order pooling technique based on Koopman theory. The Koopman operator linearizes a non-linear dynamics system, thus providing a way to represent the complex system through the dynamics matrix, which can be used for classification. We also propose an eigenvalue normalization method to encourage the learned dynamics to be non-decaying and stable. Besides, we also show that our Koopman pooling framework can be easily extended to one-shot action recognition when combined with Dynamic Mode Decomposition. The proposed method is evaluated on three benchmark datasets, namely NTU RGB+D 60, 120 and NW-UCLA. Our experiments clearly demonstrate that Koopman pooling significantly improves the performance under both full-dataset and one-shot settings. | ['Yadong Mu', 'Xin Xu', 'Xinghan Wang'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['skeleton-based-action-recognition', 'action-recognition-in-videos', 'action-recognition'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [-2.42083427e-02 -4.91824478e-01 -4.61538434e-01 1.08107015e-01
-3.03597420e-01 -1.18920133e-01 5.76514065e-01 -3.57275635e-01
-4.31879759e-01 4.01745826e-01 3.96420062e-01 2.68963128e-01
-2.12453067e-01 -4.95755315e-01 -4.51169163e-01 -1.03101790e+00
-2.61762202e-01 -1.68378547e-01 4.92854327e-01 -2.74918020e-01
9.90138799e-02 5.94912827e-01 -1.23374522e+00 2.20714509e-01
5.27459323e-01 1.05912519e+00 -8.83266404e-02 7.21930802e-01
1.34022251e-01 1.03283215e+00 -2.91324198e-01 -2.09663138e-02
4.59873170e-01 -4.24855322e-01 -6.32625461e-01 2.42842078e-01
1.47760406e-01 -3.25859100e-01 -8.58851790e-01 8.31029713e-01
5.43421447e-01 5.17069697e-01 4.88148600e-01 -1.30377328e+00
-3.73154193e-01 5.49508035e-02 -8.44102859e-01 5.68087518e-01
3.33511412e-01 4.49847519e-01 1.05541182e+00 -7.44471967e-01
7.64724553e-01 1.38251817e+00 5.34058273e-01 5.54147780e-01
-1.09190297e+00 -5.53352654e-01 3.31268072e-01 5.00801861e-01
-1.09915388e+00 -2.71616280e-01 8.70204687e-01 -4.26012307e-01
1.05029690e+00 4.56482843e-02 9.83287334e-01 1.05922556e+00
4.90364105e-01 1.13111866e+00 7.60950387e-01 -6.40454739e-02
5.21615185e-02 -7.80106068e-01 2.33653918e-01 7.10602224e-01
-2.97719110e-02 -1.02872819e-01 -8.22660148e-01 -1.14572696e-01
1.21695769e+00 3.01529944e-01 -3.70724469e-01 -5.04947662e-01
-1.44066107e+00 5.30030608e-01 4.74331021e-01 1.38312206e-01
-4.48634982e-01 4.99694556e-01 6.04971766e-01 1.10998973e-01
3.18691999e-01 1.49597466e-01 -1.97479308e-01 -5.17673671e-01
-6.23253405e-01 3.33254695e-01 4.00001496e-01 4.82721686e-01
5.75862050e-01 -1.07104167e-01 -3.50524902e-01 6.34127796e-01
1.51250571e-01 2.49669328e-01 7.29279518e-01 -1.04441369e+00
4.97232616e-01 8.05532396e-01 8.31605494e-02 -1.15862811e+00
-6.67483151e-01 -7.42957890e-02 -1.10661221e+00 2.19346076e-01
4.85829294e-01 2.07892861e-02 -9.97343123e-01 1.75288260e+00
2.61795133e-01 5.53865492e-01 -2.44784867e-03 1.12270069e+00
4.95833516e-01 6.81220531e-01 -8.84737670e-02 -2.27417812e-01
1.16446400e+00 -1.20156348e+00 -7.53194034e-01 4.25951788e-03
4.85203862e-01 -3.68518859e-01 1.09390819e+00 3.61088812e-01
-8.91797364e-01 -4.82387125e-01 -1.04968178e+00 -1.03024751e-01
-6.02611937e-02 2.34823421e-01 6.91077709e-01 7.11888969e-02
-6.43302798e-01 1.00498867e+00 -1.62565351e+00 -4.42267537e-01
4.09746140e-01 2.09048256e-01 -5.86712480e-01 1.12385117e-01
-9.41158175e-01 5.87700367e-01 3.28342855e-01 3.80756468e-01
-6.26925230e-01 -5.16171634e-01 -6.56551898e-01 -5.23239700e-03
5.74218750e-01 -4.66804385e-01 9.52565372e-01 -7.81130314e-01
-1.84172690e+00 3.78012717e-01 -9.41935256e-02 -3.69023204e-01
5.87378085e-01 -3.38951617e-01 -1.64934427e-01 3.78161818e-01
-1.82606071e-01 2.74003297e-01 7.60781109e-01 -5.77406108e-01
-3.78900647e-01 -4.64776903e-01 2.09351718e-01 2.50818908e-01
-5.72132647e-01 1.04428768e-01 -4.44194853e-01 -8.97962213e-01
4.48891878e-01 -9.38251078e-01 -3.62795472e-01 1.80079743e-01
-2.27934480e-01 -2.16392159e-01 8.55068266e-01 -6.15960836e-01
1.38935673e+00 -2.12740302e+00 6.06747866e-01 9.13902279e-03
2.37532914e-01 3.07051897e-01 -1.43518522e-01 5.07219315e-01
-1.95135996e-01 -8.04427564e-02 -1.70919403e-01 -2.67310828e-01
-6.32913336e-02 2.65605897e-01 -1.58987552e-01 6.07919931e-01
3.64299685e-01 1.05327165e+00 -7.07245290e-01 -3.47859889e-01
6.13061897e-02 4.35864002e-01 -5.85958242e-01 -5.44894338e-02
1.89503096e-02 5.15724361e-01 -5.90953469e-01 6.88249528e-01
4.74908382e-01 -3.31270039e-01 5.97635703e-03 -2.15001807e-01
-1.70520842e-02 -1.91708803e-02 -1.46849859e+00 1.93795478e+00
-1.31974548e-01 5.64637959e-01 -1.33082613e-01 -1.22455823e+00
6.23032272e-01 1.66923076e-01 8.87496412e-01 -4.99361873e-01
1.67738140e-01 -2.72079767e-03 9.08804759e-02 -5.87520599e-01
3.46986175e-01 -4.19165716e-02 -5.26798936e-03 4.72478390e-01
1.14728257e-01 4.39286739e-01 4.48017061e-01 9.23299119e-02
1.40201700e+00 3.51574749e-01 2.82497525e-01 -1.82810009e-01
6.60515606e-01 -3.27184051e-01 8.91938746e-01 3.85492861e-01
-5.22161961e-01 5.73479950e-01 8.56415272e-01 -5.69635570e-01
-7.63469875e-01 -8.82132113e-01 1.73384562e-01 1.00602973e+00
2.65735328e-01 -5.69065571e-01 -6.07831836e-01 -4.05117482e-01
-1.17266305e-01 -2.04256549e-01 -7.66179442e-01 -4.10231352e-01
-8.99068713e-01 -9.65877175e-01 3.12439501e-01 9.23941433e-01
8.46427500e-01 -1.14431953e+00 -9.39961731e-01 4.33810711e-01
-3.97229970e-01 -1.11373222e+00 -7.22940266e-01 -1.34692520e-01
-9.60050225e-01 -1.07380819e+00 -9.79904592e-01 -4.17726099e-01
3.81201863e-01 2.76857048e-01 4.34372187e-01 -1.52139757e-02
-4.80533630e-01 5.55584192e-01 -4.72485840e-01 9.76034701e-02
1.58964366e-01 4.58190553e-02 3.05818111e-01 5.68852663e-01
2.11826801e-01 -9.43675935e-01 -7.71803558e-01 5.86035132e-01
-8.74758124e-01 5.42495660e-02 5.05921721e-01 9.69537258e-01
6.29870534e-01 -5.31929806e-02 2.99908370e-01 9.21594575e-02
5.60929000e-01 -6.04900420e-02 -2.24714994e-01 1.69020742e-01
-3.35584395e-02 1.89002559e-01 4.96496290e-01 -9.67425346e-01
-8.91227305e-01 2.59895891e-01 2.45388895e-01 -7.70108759e-01
2.84031540e-01 5.82995772e-01 -2.04209223e-01 -1.56784043e-01
4.59392637e-01 2.81334519e-01 1.19083375e-01 -5.23486972e-01
2.87309498e-01 2.57666618e-01 5.92024624e-01 -6.40690804e-01
6.45291746e-01 9.21128929e-01 1.61745921e-01 -8.42649043e-01
-6.52142406e-01 -6.74679816e-01 -9.05035257e-01 -4.17455673e-01
1.09880686e+00 -7.36646771e-01 -1.01165926e+00 9.62411404e-01
-9.68854725e-01 -5.02650261e-01 -3.68995577e-01 6.65389419e-01
-7.71498561e-01 6.02907002e-01 -7.45374262e-01 -8.22802305e-01
-1.93792000e-01 -1.03784657e+00 9.63412046e-01 3.66204262e-01
8.88968930e-02 -5.31661153e-01 3.31721783e-01 6.88231364e-02
2.28059083e-01 4.23455924e-01 4.59161073e-01 -1.80013001e-01
-6.03739440e-01 -3.42285961e-01 -1.86305285e-01 3.85301799e-01
-9.12951156e-02 1.12242691e-01 -5.75803697e-01 -3.02568048e-01
-1.45289674e-01 -3.99291962e-01 1.21974409e+00 3.58061433e-01
1.01316702e+00 -8.69962648e-02 -1.53310180e-01 6.12985432e-01
1.03173685e+00 8.95843469e-03 9.04808819e-01 3.06215942e-01
7.82112300e-01 4.17595148e-01 4.04522777e-01 7.23472834e-01
1.11375093e-01 8.84007454e-01 2.75246471e-01 2.25242972e-01
7.56313950e-02 -2.10497707e-01 6.36633396e-01 8.10183287e-01
-7.80208290e-01 1.26186624e-01 -8.34165514e-01 2.00555995e-01
-2.33873510e+00 -1.18307829e+00 1.05194837e-01 1.90757883e+00
5.16122520e-01 2.64153659e-01 5.37213326e-01 1.21802606e-01
4.93191332e-01 4.51485157e-01 -7.43914068e-01 1.41932234e-01
-1.18982524e-01 2.68622249e-01 3.46377760e-01 6.37716874e-02
-1.20343041e+00 8.58727813e-01 5.52050686e+00 8.09818029e-01
-1.21073091e+00 1.17383577e-01 1.44931808e-01 -1.53378576e-01
4.90669072e-01 -7.12503567e-02 -4.60992992e-01 3.15456212e-01
5.64494193e-01 -2.83679903e-01 2.98792928e-01 7.58318245e-01
4.09180164e-01 -2.13203073e-01 -8.32704067e-01 1.13216519e+00
-1.92118883e-02 -1.14235568e+00 -2.52857298e-01 1.74114287e-01
5.85981250e-01 -1.57457098e-01 -2.01666772e-01 2.37948030e-01
1.82594825e-02 -8.21525514e-01 4.80661035e-01 9.38532114e-01
2.53255516e-01 -5.74993372e-01 4.86288071e-01 3.93074274e-01
-1.66962671e+00 -3.05058211e-01 -4.26943690e-01 -3.36144328e-01
4.05621618e-01 2.10949197e-01 1.99313641e-01 6.07764721e-01
8.08647454e-01 1.19657135e+00 -3.77372324e-01 1.11676180e+00
-4.36467603e-02 3.10078025e-01 -3.93965960e-01 -1.30591989e-01
1.79436490e-01 -2.07071632e-01 5.35734415e-01 8.85020316e-01
2.03604728e-01 4.07856345e-01 4.22882438e-01 5.56791902e-01
1.62882373e-01 1.31130308e-01 -1.66933924e-01 -4.33284491e-01
-3.08649629e-01 1.15850437e+00 -9.13408577e-01 -2.60251850e-01
-4.27494615e-01 1.24666607e+00 2.30732232e-01 4.91897255e-01
-7.92286396e-01 -4.42358583e-01 1.00654960e+00 -1.07507251e-01
4.38346893e-01 -7.13168979e-01 1.47567555e-01 -1.78240883e+00
3.38750958e-01 -6.66288376e-01 3.63598585e-01 -6.32609367e-01
-1.10125017e+00 9.49321836e-02 1.53775632e-01 -1.47296095e+00
3.53829674e-02 -7.65018344e-01 -7.81368911e-01 4.46410269e-01
-1.11049509e+00 -1.04584026e+00 -4.87388521e-01 7.52150297e-01
4.33779240e-01 6.00354597e-02 4.56033021e-01 2.53923506e-01
-1.10386014e+00 2.57753938e-01 -1.40426069e-01 2.85067469e-01
4.75353301e-01 -9.12165344e-01 2.18858004e-01 9.31108832e-01
-3.40886414e-03 5.72860777e-01 5.16027093e-01 -5.27364910e-01
-1.47331595e+00 -8.23637009e-01 3.55312884e-01 -2.98591524e-01
9.40359414e-01 -3.64030212e-01 -1.02845609e+00 5.96539378e-01
-1.36663780e-01 4.41178322e-01 3.15234363e-01 -2.43117884e-01
-2.73375809e-01 -1.28441289e-01 -7.00740814e-01 6.98741019e-01
1.36717308e+00 -2.72030830e-01 -4.66054797e-01 3.53035659e-01
5.13937950e-01 -4.56883878e-01 -8.90754461e-01 3.67778361e-01
9.21563923e-01 -9.35405970e-01 9.21669722e-01 -9.79154527e-01
5.93425035e-01 -2.67113060e-01 -1.14755236e-01 -1.00989044e+00
-4.50040072e-01 -9.20586288e-01 -3.67128968e-01 8.24528933e-01
4.50520311e-03 -6.87816799e-01 7.61295140e-01 3.61210406e-01
6.73085172e-03 -9.37148452e-01 -1.18607378e+00 -1.13621426e+00
-1.26232416e-03 -4.45543349e-01 2.29945317e-01 5.91194451e-01
2.50926256e-01 -6.55758902e-02 -6.50917888e-01 -9.04823020e-02
4.04409289e-01 -8.32501352e-02 1.03368521e+00 -1.00802588e+00
-4.44092393e-01 -6.39593482e-01 -9.83268023e-01 -1.31915092e+00
1.66424081e-01 -6.80255592e-01 -3.67319845e-02 -1.36933196e+00
3.30672830e-01 2.00947747e-01 -4.90853250e-01 5.79621136e-01
2.16072984e-02 2.81768382e-01 3.25974941e-01 4.66905415e-01
-5.17763436e-01 9.09275770e-01 1.48213887e+00 -2.29167985e-03
-3.79739016e-01 -6.93899766e-02 -1.01252683e-01 7.33761370e-01
5.65951467e-01 -9.76088941e-02 -2.06456602e-01 3.80987264e-02
-2.37898137e-02 -8.66177529e-02 6.95891201e-01 -1.21237278e+00
2.55153179e-01 -4.29062426e-01 1.23820812e-01 -3.92553896e-01
5.04978001e-01 -4.81118500e-01 -7.28387535e-02 4.95431691e-01
-1.50177732e-01 -2.35657804e-02 8.76910239e-02 8.02727282e-01
-2.09207088e-01 2.38782018e-01 8.37894440e-01 -4.87980917e-02
-8.26935351e-01 6.76560581e-01 -2.63894081e-01 2.29401905e-02
1.04807913e+00 -1.68022051e-01 -2.27555111e-01 -4.37146395e-01
-9.15869355e-01 3.33437562e-01 2.79837221e-01 3.94684166e-01
5.21620989e-01 -1.75196612e+00 -2.85172313e-01 2.44354889e-01
1.59731045e-01 -3.42351347e-01 5.84843338e-01 1.48497665e+00
-4.68767107e-01 2.04409778e-01 -5.52293122e-01 -6.53146207e-01
-1.05197191e+00 2.85413563e-01 4.55460727e-01 -5.60305893e-01
-1.10315704e+00 6.18479729e-01 8.00689384e-02 -7.26026893e-02
3.11505616e-01 -5.29112995e-01 -7.46209770e-02 1.90480292e-01
7.27542341e-01 6.03067636e-01 -2.79689401e-01 -7.81360209e-01
-4.45669800e-01 9.14114773e-01 1.31475583e-01 -2.48796433e-01
1.41114414e+00 7.81894401e-02 -7.46116638e-02 5.65926552e-01
1.12826180e+00 -4.30710495e-01 -1.56174302e+00 -3.19642305e-01
-1.25345260e-01 -4.45173770e-01 -2.07137510e-01 -1.05330661e-01
-1.19464219e+00 1.14136612e+00 4.36456472e-01 2.41155088e-01
1.22549534e+00 -1.21450715e-01 9.32248950e-01 5.26176512e-01
1.71682030e-01 -1.23746109e+00 5.62328458e-01 5.03410816e-01
1.18235135e+00 -9.94308114e-01 8.50471035e-02 -2.30584279e-01
-5.50673604e-01 1.39860582e+00 7.04816103e-01 -5.25790334e-01
9.23645496e-01 -2.79379841e-02 -2.34800130e-01 -1.20419331e-01
-5.80821455e-01 -3.31065714e-01 3.18137288e-01 2.15638921e-01
9.97355431e-02 -1.31542400e-01 -5.50903857e-01 7.22944379e-01
2.27044612e-01 2.53832221e-01 5.44893980e-01 1.17366445e+00
-3.55225742e-01 -9.72472608e-01 -1.28664389e-01 2.76646823e-01
-2.17628583e-01 4.37633663e-01 -3.37117583e-01 8.67012560e-01
-1.20682634e-01 4.92194831e-01 7.41960630e-02 -6.23585820e-01
5.86264551e-01 1.70201123e-01 5.46127319e-01 -3.64543766e-01
-1.72473758e-01 2.59244800e-01 -3.15417469e-01 -1.14531815e+00
-8.30397844e-01 -7.59430349e-01 -1.31109381e+00 -1.81909397e-01
-1.88611060e-01 -4.97039407e-01 2.36223325e-01 1.04778802e+00
3.46358657e-01 4.84409153e-01 3.70287478e-01 -1.15451384e+00
-6.15398347e-01 -1.00827980e+00 -7.26163805e-01 5.00659704e-01
2.76325762e-01 -1.24673486e+00 -4.67704713e-01 6.23567589e-02] | [7.862645149230957, 0.32347196340560913] |
ec2273f1-b54d-412f-a699-a5418083869e | efficient-long-sequential-user-data-modeling | 2209.12212 | null | https://arxiv.org/abs/2209.12212v1 | https://arxiv.org/pdf/2209.12212v1.pdf | Efficient Long Sequential User Data Modeling for Click-Through Rate Prediction | Recent studies on Click-Through Rate (CTR) prediction has reached new levels by modeling longer user behavior sequences. Among others, the two-stage methods stand out as the state-of-the-art (SOTA) solution for industrial applications. The two-stage methods first train a retrieval model to truncate the long behavior sequence beforehand and then use the truncated sequences to train a CTR model. However, the retrieval model and the CTR model are trained separately. So the retrieved subsequences in the CTR model is inaccurate, which degrades the final performance. In this paper, we propose an end-to-end paradigm to model long behavior sequences, which is able to achieve superior performance along with remarkable cost-efficiency compared to existing models. Our contribution is three-fold: First, we propose a hashing-based efficient target attention (TA) network named ETA-Net to enable end-to-end user behavior retrieval based on low-cost bit-wise operations. The proposed ETA-Net can reduce the complexity of standard TA by orders of magnitude for sequential data modeling. Second, we propose a general system architecture as one viable solution to deploy ETA-Net on industrial systems. Particularly, ETA-Net has been deployed on the recommender system of Taobao, and brought 1.8% lift on CTR and 3.1% lift on Gross Merchandise Value (GMV) compared to the SOTA two-stage methods. Third, we conduct extensive experiments on both offline datasets and online A/B test. The results verify that the proposed model outperforms existing CTR models considerably, in terms of both CTR prediction performance and online cost-efficiency. ETA-Net now serves the main traffic of Taobao, delivering services to hundreds of millions of users towards billions of items every day. | ['Junfeng Ge', 'Tao Zhuang', 'Shanshan Lv', 'Changhua Pei', 'Yue Xu', 'Qiwei Chen'] | 2022-09-25 | null | null | null | null | ['click-through-rate-prediction'] | ['miscellaneous'] | [-1.97301269e-01 -8.02719951e-01 -4.59701240e-01 -2.58829117e-01
-8.47964406e-01 -3.56000692e-01 1.26646757e-01 -3.17811891e-02
-1.85531780e-01 1.76153764e-01 -2.83697158e-01 -6.99379325e-01
-1.01263873e-01 -7.03081548e-01 -5.75317860e-01 -5.76860607e-01
-8.14708173e-02 4.25428092e-01 1.99711695e-01 -5.56435287e-01
4.44780886e-01 3.11471075e-01 -1.57938766e+00 4.50973004e-01
6.78617060e-01 1.52454412e+00 3.96562546e-01 4.64124441e-01
2.90256273e-02 6.99422419e-01 -4.30765390e-01 -5.24144173e-01
4.67920393e-01 -1.11835107e-01 -2.96447098e-01 -3.94218624e-01
9.70661566e-02 -7.53740311e-01 -6.59157455e-01 6.61104143e-01
5.24044633e-01 7.72984624e-02 2.36925483e-01 -1.28244579e+00
-7.83886611e-01 8.37478638e-01 -7.20151484e-01 6.58937469e-02
2.40747988e-01 1.26908019e-01 1.35418546e+00 -9.39504802e-01
1.64415568e-01 9.18790996e-01 5.06526589e-01 2.87433177e-01
-7.69208431e-01 -8.68500769e-01 2.05611333e-01 4.20692861e-01
-1.27096212e+00 -6.65234253e-02 4.72057641e-01 -6.94835037e-02
1.00946343e+00 4.03236568e-01 6.57102227e-01 8.90312374e-01
2.67092139e-01 1.24738538e+00 4.46260273e-01 -1.14863202e-01
2.27525234e-02 3.09475243e-01 3.94562811e-01 4.91840780e-01
1.30256206e-01 -6.56071166e-03 -3.39356869e-01 -2.16663733e-01
6.38945520e-01 5.07102966e-01 2.39836693e-01 -4.62930873e-02
-1.11440754e+00 7.78738439e-01 3.74035507e-01 8.20500404e-02
-4.35478210e-01 3.11769247e-01 6.20963573e-01 3.99548441e-01
3.57793897e-01 2.05117419e-01 -6.10876858e-01 -5.57879925e-01
-7.70068944e-01 3.09344918e-01 7.89084315e-01 1.38285387e+00
3.40109259e-01 6.22789562e-02 -3.18384796e-01 1.13051260e+00
3.61556172e-01 5.84718525e-01 6.57507598e-01 -4.60866183e-01
5.04395187e-01 3.58119309e-01 1.75020233e-01 -8.16858888e-01
7.94114079e-03 -6.73694313e-01 -6.17007494e-01 -4.94378090e-01
1.26875043e-01 1.86586916e-01 -7.37946391e-01 1.34148920e+00
-8.41557235e-03 1.00743864e-02 -4.15414035e-01 8.84047568e-01
1.33236259e-01 9.91459966e-01 -8.24286893e-04 -2.19611049e-01
1.18511105e+00 -1.43015361e+00 -7.51387537e-01 1.17966309e-02
7.99092829e-01 -9.08670485e-01 1.31519389e+00 7.98306346e-01
-1.20258701e+00 -8.29375625e-01 -1.05886519e+00 9.05554593e-02
-3.85929763e-01 5.38191617e-01 7.83347249e-01 8.04854035e-01
-5.00944912e-01 6.70716822e-01 -6.33604825e-01 -1.89567342e-01
2.33695835e-01 4.32674646e-01 5.01988411e-01 -1.68581694e-01
-1.27044749e+00 4.71377164e-01 3.41307209e-03 1.87226087e-01
-9.50696051e-01 -8.24042916e-01 -2.92274475e-01 4.72364455e-01
5.43597698e-01 -2.57529289e-01 1.56240809e+00 -4.01347250e-01
-1.65848207e+00 4.53401543e-02 -1.02628767e-01 -4.82735515e-01
4.43248421e-01 -5.87145984e-01 -8.23625088e-01 -3.16833794e-01
-2.41543412e-01 5.03034592e-02 5.50540030e-01 -8.86393249e-01
-9.85697865e-01 -2.67926157e-01 -1.00405589e-01 -1.62524447e-01
-7.46700644e-01 -2.55775768e-02 -9.95892882e-01 -7.14359641e-01
-2.30890960e-01 -1.11943018e+00 -3.06509107e-01 -2.21383050e-01
-2.63889402e-01 -4.07459438e-01 1.08072805e+00 -6.09423161e-01
1.99457979e+00 -2.24545741e+00 -4.06015396e-01 3.97714227e-01
-4.75485362e-02 5.02615392e-01 -2.05100760e-01 7.18770742e-01
1.33872673e-01 6.02032989e-02 3.52233231e-01 -1.64061189e-01
3.40992689e-01 -1.85866609e-01 -6.59293056e-01 2.38693669e-01
-2.46529683e-01 8.48456681e-01 -6.83768630e-01 5.46427118e-03
2.80781507e-01 2.72593200e-01 -7.25018561e-01 2.60695130e-01
-3.90352577e-01 -5.89901488e-03 -5.55623353e-01 8.65303040e-01
7.18622983e-01 -5.45242965e-01 3.53737861e-01 -2.51448452e-01
-2.45951936e-01 4.61657017e-01 -8.00519347e-01 1.45879960e+00
-7.91815698e-01 1.11042768e-01 -3.84762228e-01 -6.26224875e-01
9.71371531e-01 -4.83756773e-02 9.32830334e-01 -1.12142658e+00
3.51430893e-01 5.21977603e-01 -4.69093211e-02 -3.73176157e-01
7.67962337e-01 3.84292036e-01 -3.33856195e-01 4.65559632e-01
-3.75775129e-01 5.18942237e-01 2.95656860e-01 2.44790450e-01
1.07717526e+00 1.39075175e-01 -1.18295029e-02 -1.30021218e-02
4.99378592e-01 -1.04595870e-01 4.98776525e-01 7.66965926e-01
-2.58448143e-02 2.17755273e-01 1.03371300e-01 -4.24808085e-01
-1.29842484e+00 -7.34501123e-01 8.28016549e-02 1.43383121e+00
4.37106669e-01 -7.07963347e-01 -4.36403602e-01 -7.79137075e-01
2.36054987e-01 8.58782589e-01 -1.00383677e-01 -2.12531760e-01
-7.77488351e-01 -4.51972306e-01 3.70481074e-01 4.64547247e-01
5.87617695e-01 -7.60647953e-01 -1.39275357e-01 3.83452803e-01
-1.59451470e-01 -9.85603750e-01 -9.46383774e-01 -1.50902331e-01
-7.49913096e-01 -5.88088572e-01 -7.48365343e-01 -4.63702619e-01
4.65621278e-02 7.31229186e-01 7.91383326e-01 1.44268617e-01
-2.75074035e-01 -4.28719223e-02 -6.11986578e-01 -3.46423805e-01
-3.94803435e-02 3.54584783e-01 1.39064535e-01 1.05662242e-01
6.30304694e-01 -4.61386561e-01 -8.73957932e-01 9.27763879e-01
-6.99814022e-01 -3.74974698e-01 9.34268534e-01 6.40120625e-01
4.33498204e-01 -4.04887497e-02 7.60779440e-01 -9.76466954e-01
6.28983200e-01 -8.99625301e-01 -8.29215348e-01 3.12382340e-01
-1.18974292e+00 -2.25866109e-01 9.42809880e-01 -6.70756936e-01
-8.48473072e-01 -1.27613589e-01 -3.65060598e-01 -7.24309087e-01
5.22884965e-01 5.65251946e-01 1.20409206e-01 2.21935704e-01
1.76130936e-01 5.67750990e-01 1.33787151e-02 -7.11503565e-01
2.78688341e-01 1.11601377e+00 4.02008854e-02 -3.93900782e-01
7.78552830e-01 1.10301144e-01 -3.59577477e-01 -4.66069728e-01
-7.72189260e-01 -9.18215215e-01 -1.93422690e-01 -1.61404952e-01
2.72073507e-01 -9.60546553e-01 -1.43276346e+00 3.87502015e-01
-7.23890781e-01 -6.95198849e-02 -3.37082855e-02 5.23443878e-01
-6.07749283e-01 5.39520085e-01 -9.34900343e-01 -9.83621359e-01
-7.14166641e-01 -1.10209286e+00 9.66301203e-01 -1.39382277e-02
1.79367691e-01 -3.42341989e-01 -3.17389190e-01 5.26717186e-01
8.22995543e-01 -6.48465455e-01 1.03455365e+00 -7.98366368e-01
-8.71776342e-01 -5.93160331e-01 -3.86649698e-01 4.08100575e-01
-1.19581781e-01 -1.16984844e-01 -6.12683058e-01 -6.09820187e-01
-1.19206481e-01 -2.46867880e-01 4.77450877e-01 7.72380307e-02
1.69443548e+00 -1.69052094e-01 -3.64123166e-01 3.20939392e-01
1.36073673e+00 6.65170848e-01 8.27773690e-01 1.73248976e-01
5.15744805e-01 3.26010101e-02 1.42342770e+00 7.19441116e-01
1.23213395e-01 9.63990510e-01 3.91443074e-01 1.11115694e-01
3.23594436e-02 -4.14985090e-01 6.97534919e-01 1.52690732e+00
3.95796522e-02 -5.29737294e-01 -3.76079261e-01 4.06696141e-01
-1.97674727e+00 -8.66854072e-01 -1.17691830e-01 2.52755284e+00
4.07259554e-01 3.38105053e-01 5.34653068e-01 6.42809644e-02
5.98106325e-01 -5.27443737e-02 -7.35854745e-01 -4.11866605e-01
4.41924095e-01 -1.52867228e-01 7.73794651e-01 -1.32137492e-01
-7.00851619e-01 6.08435750e-01 5.86263657e+00 1.34204626e+00
-1.07533526e+00 1.84988603e-01 4.80191857e-01 -3.08236957e-01
-7.21655190e-02 -1.45562217e-01 -1.22421050e+00 7.52808869e-01
1.52237332e+00 -2.83483416e-01 5.43674707e-01 1.40621173e+00
3.68601143e-01 3.95727664e-01 -1.03886330e+00 1.07382512e+00
9.19124559e-02 -1.16083264e+00 2.51348406e-01 4.97311920e-01
4.12990481e-01 8.13254863e-02 3.84526551e-01 9.89974320e-01
4.77537289e-02 -5.05658448e-01 7.01667845e-01 2.96275944e-01
7.44557440e-01 -8.90109003e-01 8.20608556e-01 3.08240861e-01
-1.34339726e+00 -7.37062871e-01 -5.11569202e-01 3.29779744e-01
5.02568722e-01 6.98008180e-01 -6.99400008e-01 7.05323577e-01
8.19833279e-01 7.42013514e-01 -1.66292205e-01 1.23514664e+00
4.98148710e-01 8.49434257e-01 -3.33751202e-01 -5.39527416e-01
3.28040242e-01 -3.30288649e-01 2.50182927e-01 1.04422832e+00
6.63981616e-01 -1.17800653e-01 3.17748904e-01 5.97969651e-01
-3.67510706e-01 3.91260713e-01 -2.43985698e-01 -1.80794671e-01
4.55826789e-01 1.16433513e+00 -1.55150905e-01 -4.50933903e-01
-6.19741678e-01 8.33464921e-01 5.78388348e-02 8.59676898e-02
-1.48936665e+00 -6.42598987e-01 6.84398890e-01 3.73894274e-01
7.14916885e-01 -2.09318742e-01 -2.80417502e-03 -1.00907350e+00
1.31032959e-01 -9.85991299e-01 2.98669517e-01 -5.27840316e-01
-1.43851674e+00 4.74612206e-01 -2.82833010e-01 -1.81758296e+00
1.86400697e-01 -5.62383711e-01 -2.13343903e-01 5.53453326e-01
-1.48550677e+00 -1.08125460e+00 -1.40264288e-01 3.81800681e-01
1.04101300e+00 -1.70768887e-01 4.46519434e-01 1.13502789e+00
-7.83877730e-01 9.61931586e-01 6.27059281e-01 -1.72621891e-01
6.43258870e-01 -8.46142709e-01 4.74897832e-01 4.06723976e-01
-5.56549802e-02 1.14808452e+00 4.59527731e-01 -5.93025029e-01
-1.91935527e+00 -1.15709543e+00 7.15018868e-01 -1.55803233e-01
7.47375667e-01 -5.66835642e-01 -6.86137736e-01 5.88798702e-01
-1.07664503e-01 -1.61890551e-01 7.01337874e-01 2.46777624e-01
-5.04643083e-01 -5.60517251e-01 -8.46322656e-01 6.12846315e-01
8.44976366e-01 -4.17671531e-01 7.46088699e-02 5.31263769e-01
1.00160348e+00 -7.04460368e-02 -1.22194326e+00 2.01860383e-01
1.05339932e+00 -7.63410091e-01 8.89905155e-01 -5.47835410e-01
2.06218451e-01 -1.65353149e-01 -3.95508051e-01 -6.57850623e-01
-5.41538060e-01 -9.69357550e-01 -4.20759022e-01 1.09597325e+00
6.71562433e-01 -4.95669901e-01 9.14998233e-01 4.84215707e-01
-2.11362988e-01 -9.76745546e-01 -4.32788849e-01 -1.01644063e+00
-2.27893189e-01 -5.22974372e-01 7.20217526e-01 5.56768119e-01
9.29014087e-02 4.76881593e-01 -9.22829926e-01 -1.81317747e-01
4.62073117e-01 4.15565610e-01 8.41910660e-01 -8.60455692e-01
-5.58946609e-01 -1.64345592e-01 7.41862729e-02 -2.00653148e+00
-5.53263724e-01 -7.39309967e-01 -8.97320285e-02 -1.00493276e+00
1.96386561e-01 -7.18955755e-01 -7.36515105e-01 3.22954446e-01
2.62527019e-01 5.54786026e-02 3.15609753e-01 4.55259025e-01
-8.05101454e-01 6.78785980e-01 1.11182010e+00 -1.81491643e-01
-2.31136471e-01 5.21563351e-01 -5.67243159e-01 1.74251467e-01
8.76883030e-01 -3.90684992e-01 -5.53775847e-01 -1.67448431e-01
2.10160255e-01 1.69940099e-01 -2.18648210e-01 -8.35022986e-01
2.96021104e-01 -9.73321274e-02 4.47312044e-03 -1.08263850e+00
5.39328754e-01 -1.10654104e+00 -5.32177314e-02 5.94264388e-01
-4.10721689e-01 2.60872334e-01 3.19327042e-02 9.02364492e-01
2.77780052e-02 -8.17517266e-02 4.90087420e-01 1.87518612e-01
-5.24325967e-01 5.51278114e-01 -4.47026759e-01 -6.96351767e-01
1.11563408e+00 -9.66267511e-02 -4.34625864e-01 -4.49375480e-01
-4.34870511e-01 2.38228261e-01 -9.72187594e-02 6.77239478e-01
5.45111597e-01 -1.38258016e+00 -4.05846387e-01 1.69478804e-01
2.74519414e-01 -7.30029404e-01 4.68852818e-01 9.28805649e-01
-3.33184600e-01 9.47561860e-01 5.06532900e-02 -3.75351787e-01
-1.19590163e+00 9.31745708e-01 -2.54825920e-01 -5.93676805e-01
-4.04063404e-01 3.92840147e-01 -2.44457990e-01 -3.10536653e-01
4.11219209e-01 -3.49096864e-01 3.02899238e-02 -3.12231928e-01
5.23753047e-01 7.20110834e-01 2.00860307e-01 -1.79292828e-01
8.52158759e-03 3.85966063e-01 -7.51450062e-01 3.71547610e-01
1.19539082e+00 -3.14365834e-01 1.51372239e-01 3.31458658e-01
1.44452667e+00 7.84602463e-02 -6.43494546e-01 -2.44955927e-01
-1.74800709e-01 -8.23629200e-01 7.65506998e-02 -9.68807817e-01
-1.27282405e+00 7.79121161e-01 6.16623580e-01 4.27893519e-01
1.15554881e+00 -3.78327161e-01 1.70402646e+00 6.21660709e-01
7.15896428e-01 -1.17569137e+00 2.22762868e-01 3.54803920e-01
4.81876582e-01 -1.01461720e+00 3.42193246e-02 -5.36377013e-01
-6.54702306e-01 8.96134734e-01 6.75993025e-01 -1.69726822e-03
7.79721439e-01 -3.01328115e-02 -1.72939926e-01 -4.35525319e-04
-1.04801154e+00 3.48444283e-02 2.93577127e-02 3.35222259e-02
4.39372361e-01 -2.30235588e-02 -5.29749393e-01 7.58540690e-01
1.95495233e-01 1.84489995e-01 2.74689049e-01 9.41403747e-01
-3.92623037e-01 -1.33715987e+00 1.48042828e-01 8.13339412e-01
-6.81793988e-01 -2.26977810e-01 2.75753528e-01 6.43711030e-01
-3.98287386e-01 1.09779453e+00 -2.72329032e-01 -1.02675831e+00
5.42639911e-01 -2.87632763e-01 -3.52712199e-02 -3.15696299e-01
-9.36156213e-01 4.35733289e-01 1.97178170e-01 -8.67401481e-01
2.90360093e-01 -5.06020427e-01 -1.01272571e+00 -6.68288589e-01
-8.08152080e-01 3.22151005e-01 1.03981495e+00 3.89612973e-01
6.76282525e-01 6.08406663e-01 1.19728255e+00 -4.99536425e-01
-1.25193608e+00 -1.02892900e+00 -7.17179894e-01 3.69848788e-01
-9.15518776e-02 -4.69119608e-01 -2.53755778e-01 -3.27631384e-01] | [10.098097801208496, 5.572655200958252] |
d046b0d8-9371-4133-8396-7688eae29bd2 | estimation-of-control-area-in-badminton | 2305.04247 | null | https://arxiv.org/abs/2305.04247v1 | https://arxiv.org/pdf/2305.04247v1.pdf | Estimation of control area in badminton doubles with pose information from top and back view drone videos | The application of visual tracking to the performance analysis of sports players in dynamic competitions is vital for effective coaching. In racket sports, most previous studies have focused on analyzing and assessing singles players without occlusion in broadcast videos and discrete representations (e.g., stroke) that ignore meaningful spatial distributions. In this work, we present the first annotated drone dataset from top and back views in badminton doubles and propose a framework to estimate the control area probability map, which can be used to evaluate teamwork performance. We present an efficient framework of deep neural networks that enables the calculation of full probability surfaces, which utilizes the embedding of a Gaussian mixture map of players' positions and graph convolution of their poses. In the experiment, we verify our approach by comparing various baselines and discovering the correlations between the score and control area. Furthermore, we propose the practical application of assessing optimal positioning to provide instructions during a game. Our approach can visually and quantitatively evaluate players' movements, providing valuable insights into doubles teamwork. | ['Keisuke Fujii', 'Yingjiu Bei', 'Wenhui Jin', 'Kazuya Takeda', 'Ning Ding'] | 2023-05-07 | null | null | null | null | ['visual-tracking'] | ['computer-vision'] | [-1.93111882e-01 -2.13526055e-01 -6.30446672e-02 -2.70538926e-02
-4.97133762e-01 -7.84555376e-01 9.77096856e-02 2.42722794e-01
-6.32709801e-01 3.54709178e-01 1.28651336e-01 1.55792058e-01
-4.62775707e-01 -7.93399036e-01 -7.53826916e-01 -4.18568760e-01
-3.69855225e-01 6.08830154e-01 7.58215964e-01 -7.23271072e-01
3.59509557e-01 6.82366610e-01 -1.99170899e+00 3.72745901e-01
2.30793819e-01 8.13861847e-01 1.52637869e-01 9.76359844e-01
5.85766971e-01 7.58421421e-01 -1.11544156e+00 -5.16002059e-01
4.42029953e-01 -2.21247405e-01 -1.47589803e-01 -1.61502838e-01
7.87332773e-01 -3.93824518e-01 -5.89073420e-01 6.04883730e-01
7.56396353e-01 4.72986102e-01 4.87966657e-01 -1.31199598e+00
1.25008896e-01 4.34330195e-01 -4.92075980e-01 4.54150319e-01
4.70850378e-01 3.36237013e-01 6.96280956e-01 -2.16368586e-01
4.30835783e-01 1.04194593e+00 1.15920973e+00 1.22230247e-01
-9.17780995e-01 -6.92696154e-01 1.92511156e-02 4.32889253e-01
-1.46696365e+00 1.14777656e-02 6.23342991e-01 -6.86743796e-01
3.45339775e-01 3.86032164e-01 1.39672530e+00 9.37258899e-01
1.26220852e-01 9.09464002e-01 9.36067879e-01 -3.12212676e-01
2.47617438e-02 -3.51815403e-01 -1.00463822e-01 6.71760380e-01
2.84904659e-01 2.38230288e-01 -1.15294564e+00 1.48720086e-01
9.93143022e-01 -4.45415042e-02 2.63762637e-03 -6.10402822e-01
-8.56937468e-01 6.43693388e-01 3.79452705e-01 -1.03596069e-01
-1.95079952e-01 6.90566719e-01 3.15459341e-01 -2.52580851e-01
3.26611280e-01 2.78282523e-01 7.19168112e-02 -7.82551169e-01
-9.60710466e-01 7.80231297e-01 4.83307332e-01 5.60159981e-01
3.32695395e-01 2.26944927e-02 -3.64519745e-01 3.63004297e-01
-1.28538087e-02 2.95370013e-01 1.25123281e-02 -1.23994279e+00
5.81918716e-01 5.92732489e-01 2.72030622e-01 -1.36582983e+00
-8.56587708e-01 -4.75867122e-01 5.33816330e-02 7.26436794e-01
9.58234012e-01 -2.76327014e-01 -4.56265450e-01 1.35272479e+00
3.35651100e-01 1.30699083e-01 -4.97634917e-01 1.29807281e+00
7.11989641e-01 2.49461439e-02 -1.89813465e-01 5.33839226e-01
1.30467200e+00 -5.79906583e-01 -5.74128270e-01 -1.27063528e-01
3.78596336e-01 -4.09682781e-01 1.10248828e+00 6.68685198e-01
-1.31000328e+00 -7.86535740e-01 -9.56058145e-01 1.01714566e-01
-1.18500620e-01 6.39024496e-01 5.18636107e-01 8.68803024e-01
-7.62121201e-01 8.89052689e-01 -1.28632581e+00 -1.95730384e-02
2.56091177e-01 3.94787788e-01 -3.46368164e-01 4.06239808e-01
-1.03631675e+00 9.37887251e-01 3.50580215e-01 3.62799317e-01
-1.07309914e+00 -7.17989922e-01 -9.24983263e-01 9.92515236e-02
6.75950468e-01 -2.25189239e-01 1.08121753e+00 -5.14144301e-01
-1.49975729e+00 6.63987279e-01 5.27264178e-01 -5.93482375e-01
8.52581024e-01 -7.94508934e-01 1.39125884e-01 3.73638034e-01
5.80662936e-02 4.56122220e-01 2.63789922e-01 -1.07048237e+00
-7.19972968e-01 -4.75382984e-01 4.02052522e-01 5.54412186e-01
-1.94921389e-01 -9.36846584e-02 -6.34354413e-01 -5.45181394e-01
1.01391636e-01 -1.03572333e+00 -1.82949707e-01 2.75572896e-01
-2.82314777e-01 4.16878797e-02 3.94639730e-01 -8.69615495e-01
1.29859746e+00 -2.08973765e+00 3.81717801e-01 3.86110306e-01
3.54917377e-01 1.97118163e-01 3.63259435e-01 5.44066370e-01
2.60954887e-01 -3.68284553e-01 1.16558000e-01 -5.87023795e-02
1.11427195e-01 5.84183708e-02 2.57889144e-02 7.52784848e-01
-3.18647116e-01 7.47954428e-01 -7.39305258e-01 -5.48295856e-01
3.40504944e-01 5.00657082e-01 -5.36686480e-01 7.06868172e-02
1.68921694e-01 3.18157494e-01 -2.27847412e-01 4.63440746e-01
3.54241371e-01 5.35664439e-01 3.18468601e-01 -1.75820276e-01
-4.10317153e-01 -2.56575942e-02 -1.61671960e+00 1.92220485e+00
-1.17853247e-01 9.82868075e-01 1.46998882e-01 -8.89002979e-01
8.38510513e-01 -6.94802254e-02 5.63329577e-01 -3.83984119e-01
2.16089666e-01 -2.26557940e-01 4.83457521e-02 -4.98643726e-01
9.20823038e-01 1.08229741e-01 -4.05573159e-01 2.76268452e-01
1.24234155e-01 1.98862096e-03 3.85045975e-01 -3.89223285e-02
1.02024364e+00 8.45959187e-01 -1.09360013e-02 -4.81450371e-02
-8.17893893e-02 3.25214893e-01 2.92022109e-01 8.74281168e-01
-2.12242067e-01 8.76861036e-01 6.84846580e-01 -7.54447520e-01
-6.59296036e-01 -1.18482888e+00 4.14730668e-01 1.03676212e+00
5.13646841e-01 -5.28550625e-01 -9.76270735e-01 -2.95845717e-01
8.61400589e-02 3.24621469e-01 -8.88268709e-01 -3.10994744e-01
-8.16337168e-01 -4.82630104e-01 8.40431273e-01 9.70443785e-01
2.68197685e-01 -6.83787465e-01 -1.35240674e+00 8.63709897e-02
-3.45467746e-01 -8.85432184e-01 -2.72691607e-01 -2.48443503e-02
-5.05959868e-01 -1.45786262e+00 -7.18510091e-01 -3.29367816e-01
1.08252428e-01 1.61651149e-01 1.02064788e+00 -1.16947666e-01
-6.18260801e-01 7.64803350e-01 -2.28962123e-01 -4.04804140e-01
1.57648832e-01 -5.95748536e-02 2.61853129e-01 -1.16816655e-01
2.05954105e-01 -5.64817369e-01 -8.18555355e-01 5.50385535e-01
-3.83379012e-01 -1.26239598e-01 4.01600629e-01 2.61293620e-01
6.27450466e-01 -1.74049273e-01 -4.34728086e-01 -2.78325945e-01
6.44428432e-01 -8.30914453e-02 -7.59735823e-01 2.24671870e-01
2.53787488e-01 -3.10008585e-01 1.10942319e-01 -4.79763716e-01
-7.00796902e-01 1.99849010e-01 7.45380521e-02 -6.55698776e-01
-1.75913036e-01 1.35489300e-01 1.47699207e-01 -2.30428889e-01
9.52697694e-01 -7.88183957e-02 6.86136493e-03 -5.43360054e-01
3.12008947e-01 2.99728990e-01 9.24246609e-01 -6.73547208e-01
6.42193675e-01 8.07781577e-01 2.75125325e-01 -7.51110733e-01
-7.79171348e-01 -6.99092984e-01 -1.10047114e+00 -1.21886683e+00
1.05411732e+00 -8.38380754e-01 -1.63896155e+00 7.98096359e-02
-8.59925330e-01 -3.30057561e-01 -5.07737339e-01 9.89499331e-01
-8.54029655e-01 3.89270812e-01 -4.11999792e-01 -1.02326524e+00
3.07086021e-01 -9.47693765e-01 1.27801502e+00 3.62503409e-01
-1.12767562e-01 -6.66445494e-01 5.21401644e-01 5.65837324e-01
-1.54919997e-01 6.85686588e-01 -4.84965965e-02 -2.77357489e-01
-3.87146056e-01 -3.58589470e-01 3.18963677e-01 7.18816817e-02
-4.52302814e-01 -1.03994213e-01 -7.50563920e-01 -1.66570023e-01
-5.33762813e-01 -2.56804913e-01 5.99121392e-01 8.56430590e-01
1.19287264e+00 3.05438370e-01 -2.78888494e-01 7.52181292e-01
9.48146641e-01 -1.43468261e-01 6.10270619e-01 5.32664716e-01
7.80465245e-01 1.04564059e+00 1.00311744e+00 6.56174362e-01
4.24159110e-01 1.26970649e+00 7.06830561e-01 -1.32499918e-01
-3.19614112e-01 -5.35134077e-01 3.70805264e-01 3.06744337e-01
-1.08232307e+00 -9.16878805e-02 -9.12857056e-01 4.56337094e-01
-1.83245385e+00 -1.10324347e+00 -4.61074322e-01 2.27919841e+00
2.91982085e-01 2.25181758e-01 7.22739935e-01 1.64843127e-01
8.39723587e-01 -1.41470745e-01 -2.07938805e-01 -2.47838542e-01
2.09834993e-01 2.68456995e-01 8.36635709e-01 3.22442740e-01
-1.12176049e+00 7.25506067e-01 6.63304853e+00 1.08364236e+00
-7.09141731e-01 1.99771710e-02 -6.36062920e-02 -6.94408774e-01
2.23132163e-01 -1.90470479e-02 -8.49821091e-01 2.54275203e-01
6.51393056e-01 3.92697066e-01 7.03359023e-02 8.08420479e-01
1.72684595e-01 -3.56167078e-01 -6.43485487e-01 8.71019006e-01
2.85553813e-01 -1.28272057e+00 -6.21383905e-01 1.98243976e-01
3.32432330e-01 -3.09515715e-01 1.30525053e-01 1.87303901e-01
5.37260771e-01 -8.29281449e-01 1.26838303e+00 8.74238491e-01
6.77230179e-01 -9.08045709e-01 6.08679712e-01 4.04044330e-01
-1.24786091e+00 -2.01989368e-01 -1.97047949e-01 -3.77172053e-01
1.91801667e-01 -1.00035325e-01 -7.20873117e-01 5.54176271e-01
1.05009615e+00 4.96503323e-01 -6.08417332e-01 1.24878514e+00
-2.41225392e-01 5.09523630e-01 -6.46439612e-01 -1.33435205e-01
1.36810392e-01 -2.57948399e-01 6.50411665e-01 8.78309309e-01
3.42218399e-01 -1.05124094e-01 4.49105144e-01 5.98829150e-01
6.10668123e-01 4.53248480e-03 -3.89490306e-01 3.63294989e-01
9.90972668e-03 1.10786736e+00 -1.10235023e+00 6.14260249e-02
9.18966904e-02 5.47642410e-01 3.60450089e-01 6.37599081e-02
-1.08203709e+00 -3.94358873e-01 7.38673329e-01 8.57338369e-01
3.55671436e-01 -5.58023632e-01 -7.94294104e-02 -7.72015572e-01
1.79767057e-01 -5.66758215e-01 3.47052753e-01 -9.21072185e-01
-5.73817909e-01 3.36627662e-01 5.07631421e-01 -1.71270466e+00
-1.11656882e-01 -7.78863370e-01 -5.26900649e-01 5.38266897e-01
-8.45787823e-01 -1.09005606e+00 -5.72855532e-01 4.19367164e-01
3.18961322e-01 -7.64102936e-02 3.22949946e-01 3.69110316e-01
-4.29916084e-01 4.97337461e-01 -2.14986533e-01 2.26916954e-01
4.08929825e-01 -1.32413542e+00 5.88411577e-02 6.91827238e-01
5.06160021e-01 1.78529069e-01 9.34695005e-01 -6.53296232e-01
-1.00639069e+00 -3.00294548e-01 -3.41653787e-02 -7.65790939e-01
5.52916110e-01 -5.59999168e-01 -2.96544552e-01 6.63753211e-01
-3.34042758e-01 -1.19724989e-01 9.04407263e-01 1.24179922e-01
3.77140641e-01 -1.77425921e-01 -5.98661900e-01 6.10928357e-01
1.11871326e+00 -1.39986709e-01 -5.25526404e-01 2.85920590e-01
2.95290500e-02 -1.09263670e+00 -7.99388826e-01 1.47737235e-01
1.05495501e+00 -1.30048752e+00 1.24595213e+00 -5.98347187e-01
2.28472218e-01 -5.63405573e-01 8.21133927e-02 -1.14569604e+00
-5.90847768e-02 -4.42791015e-01 1.20221213e-01 6.44431353e-01
-1.93650648e-01 2.74673074e-01 1.20318294e+00 1.98287562e-01
-3.47507209e-01 -4.69149441e-01 -1.15155923e+00 -6.06843293e-01
-3.25268716e-01 -7.72426069e-01 1.77436113e-01 3.00153881e-01
-6.10937104e-02 -5.00097334e-01 -7.14287698e-01 2.79561251e-01
5.95555723e-01 -2.16869935e-02 1.28619564e+00 -1.39732718e+00
-5.66604078e-01 -2.13985741e-01 -1.13931620e+00 -9.92477179e-01
-2.14865401e-01 -3.96432251e-01 1.20833807e-01 -1.42900503e+00
-1.72004774e-01 -1.65993869e-01 9.55120921e-02 1.38374552e-01
7.85475001e-02 7.90799081e-01 3.14379007e-01 5.51298000e-02
-8.12636733e-01 2.58398712e-01 1.31449294e+00 2.41913244e-01
-8.12747180e-02 3.37667048e-01 -1.87328488e-01 9.37550604e-01
6.66638196e-01 -6.68403566e-01 -1.49328575e-01 -2.85850942e-01
4.88275290e-01 1.61900923e-01 8.39531004e-01 -1.58850121e+00
2.22752988e-01 -7.87288044e-03 3.27521294e-01 -8.14539850e-01
8.45828533e-01 -5.69212794e-01 2.51784742e-01 7.05096364e-01
-2.50838965e-01 1.50091082e-01 3.51685524e-01 7.54484534e-01
-1.77125975e-01 -2.75296837e-01 1.33554310e-01 -1.55660346e-01
-6.94756567e-01 -2.77065206e-03 -5.13365269e-01 8.91596675e-02
1.24930549e+00 -7.02120423e-01 -5.93834817e-02 -6.02143526e-01
-1.18277800e+00 2.24531889e-01 2.17910260e-01 2.08055377e-01
5.41835964e-01 -1.28084230e+00 -7.40477085e-01 -2.62060128e-02
1.06892921e-03 -2.62160301e-01 7.13452220e-01 1.00930810e+00
-1.35591137e+00 -4.96088676e-02 -7.51471996e-01 -8.39402556e-01
-1.43442607e+00 9.08848792e-02 5.14026940e-01 -1.80421516e-01
-7.33573675e-01 8.22208941e-01 3.39635648e-02 -2.55858868e-01
2.36912087e-01 -2.83735007e-01 -6.71046376e-01 5.12102842e-02
5.18566310e-01 8.01733017e-01 -4.90405858e-02 -8.35121393e-01
-2.51148283e-01 7.87589490e-01 5.27682304e-01 -2.75504768e-01
1.27982628e+00 1.01166092e-01 4.78694707e-01 4.50948238e-01
3.88601780e-01 3.80498797e-01 -1.92272973e+00 2.80586571e-01
-3.71081352e-01 -7.43823171e-01 -8.04962739e-02 -4.39667135e-01
-1.11781454e+00 1.31134999e+00 8.94330502e-01 2.00600028e-01
6.68296754e-01 -3.04365531e-02 2.72414446e-01 1.53868139e-01
3.41354072e-01 -1.30929935e+00 1.91054165e-01 6.94580050e-03
6.80290103e-01 -8.68372738e-01 1.61744729e-01 -3.33862334e-01
-9.54816222e-01 1.29349911e+00 6.59647822e-01 -4.07509983e-01
3.68122131e-01 4.66981769e-01 1.70295015e-01 -5.98495662e-01
-1.75220102e-01 -5.69975913e-01 6.45338655e-01 9.78171885e-01
5.69586642e-02 1.27976358e-01 -2.67082423e-01 8.17135334e-01
-6.31626248e-01 -1.96419924e-01 5.54024816e-01 1.00276983e+00
-4.68631804e-01 -7.79625535e-01 -6.75840259e-01 3.66198234e-02
-4.82175171e-01 3.86385620e-01 -4.66834843e-01 1.21336854e+00
5.44322550e-01 5.96924424e-01 2.76738554e-01 -6.88494384e-01
8.81811976e-01 -3.13139051e-01 8.31685483e-01 -5.92071652e-01
-9.72344220e-01 -6.78497329e-02 1.13860808e-01 -8.65099907e-01
-4.86333042e-01 -7.81348705e-01 -8.24784577e-01 -3.18995655e-01
-2.82717884e-01 9.73501876e-02 7.11264729e-01 7.19469547e-01
-1.12766661e-01 7.02648282e-01 1.52352424e-02 -1.33398700e+00
-1.42202973e-01 -7.87701309e-01 -8.23681235e-01 3.12637717e-01
-1.25161946e-01 -1.30439103e+00 5.59591465e-02 -8.84893760e-02] | [7.387075901031494, 0.0937662124633789] |
f2d09df0-9fdf-4b17-ac23-886cfe8f41b9 | detecting-and-grounding-multi-modal-media | 2304.02556 | null | https://arxiv.org/abs/2304.02556v1 | https://arxiv.org/pdf/2304.02556v1.pdf | Detecting and Grounding Multi-Modal Media Manipulation | Misinformation has become a pressing issue. Fake media, in both visual and textual forms, is widespread on the web. While various deepfake detection and text fake news detection methods have been proposed, they are only designed for single-modality forgery based on binary classification, let alone analyzing and reasoning subtle forgery traces across different modalities. In this paper, we highlight a new research problem for multi-modal fake media, namely Detecting and Grounding Multi-Modal Media Manipulation (DGM^4). DGM^4 aims to not only detect the authenticity of multi-modal media, but also ground the manipulated content (i.e., image bounding boxes and text tokens), which requires deeper reasoning of multi-modal media manipulation. To support a large-scale investigation, we construct the first DGM^4 dataset, where image-text pairs are manipulated by various approaches, with rich annotation of diverse manipulations. Moreover, we propose a novel HierArchical Multi-modal Manipulation rEasoning tRansformer (HAMMER) to fully capture the fine-grained interaction between different modalities. HAMMER performs 1) manipulation-aware contrastive learning between two uni-modal encoders as shallow manipulation reasoning, and 2) modality-aware cross-attention by multi-modal aggregator as deep manipulation reasoning. Dedicated manipulation detection and grounding heads are integrated from shallow to deep levels based on the interacted multi-modal information. Finally, we build an extensive benchmark and set up rigorous evaluation metrics for this new research problem. Comprehensive experiments demonstrate the superiority of our model; several valuable observations are also revealed to facilitate future research in multi-modal media manipulation. | ['Ziwei Liu', 'Tianxing Wu', 'Rui Shao'] | 2023-04-05 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Shao_Detecting_and_Grounding_Multi-Modal_Media_Manipulation_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Shao_Detecting_and_Grounding_Multi-Modal_Media_Manipulation_CVPR_2023_paper.pdf | cvpr-2023-1 | ['face-swapping', 'misinformation', 'fake-news-detection'] | ['computer-vision', 'miscellaneous', 'natural-language-processing'] | [ 2.73359835e-01 -3.16185623e-01 -3.71483475e-01 1.98460013e-01
-1.11255860e+00 -5.98636210e-01 1.00580239e+00 2.02194368e-03
8.45183805e-02 2.61475205e-01 4.99930531e-01 -5.69168404e-02
3.57742667e-01 -7.83309221e-01 -1.05360258e+00 -5.29932201e-01
2.76215345e-01 3.70239094e-02 3.16686958e-01 -5.48576593e-01
3.75540286e-01 6.11983575e-02 -1.44512558e+00 1.20749593e+00
5.96576452e-01 1.22620213e+00 -3.33117098e-01 5.76089203e-01
7.12798908e-02 1.18632507e+00 -8.04039896e-01 -1.06865752e+00
1.76464722e-01 -4.66855139e-01 -7.40329802e-01 4.01629478e-01
7.73743570e-01 -8.49841893e-01 -7.95296311e-01 1.43992960e+00
3.63984317e-01 -2.29472682e-01 6.31984890e-01 -1.75828516e+00
-1.14568985e+00 7.54375458e-01 -8.95990312e-01 2.19968602e-01
6.41951978e-01 6.77826524e-01 8.72019351e-01 -9.89362538e-01
7.50788867e-01 1.63736081e+00 6.46870017e-01 3.62800539e-01
-7.30349302e-01 -7.36775637e-01 -9.58108157e-02 4.78497833e-01
-1.05913794e+00 -4.57113117e-01 8.74211967e-01 -6.56631470e-01
5.32353163e-01 3.44736874e-01 2.80719757e-01 1.76623058e+00
4.42374088e-02 1.26774025e+00 1.11036026e+00 -2.46935144e-01
-4.91615742e-01 8.15242678e-02 2.92421151e-02 1.24470520e+00
1.97169393e-01 3.09896730e-02 -8.36439610e-01 -1.85430661e-01
5.52615643e-01 1.33232679e-02 -4.73327637e-01 2.82646101e-02
-1.62518644e+00 9.10791397e-01 4.13496226e-01 2.44663760e-01
-1.29734471e-01 4.23103184e-01 9.95755732e-01 5.17127514e-01
2.02202514e-01 3.71580213e-01 3.14127989e-02 -5.81166847e-03
-6.70407176e-01 3.70402783e-01 5.23493350e-01 1.11283183e+00
3.67924720e-01 -1.51145861e-01 -5.77811182e-01 5.57271600e-01
1.21577144e-01 7.07278132e-01 4.71157670e-01 -6.09301865e-01
1.15021193e+00 6.16218865e-01 1.41089097e-01 -1.84800518e+00
-3.42356190e-02 2.60102272e-01 -1.16664648e+00 -9.89829376e-03
4.96302068e-01 2.52514213e-01 -5.67866921e-01 1.43735206e+00
3.63458127e-01 -2.33455654e-02 -2.12362051e-01 1.13793850e+00
1.06344032e+00 3.42332155e-01 -9.47299525e-02 -7.39648640e-02
1.80189741e+00 -1.31692433e+00 -9.88478541e-01 4.35172534e-03
6.71279728e-01 -7.83971071e-01 1.32813644e+00 5.73343217e-01
-1.09211171e+00 -2.89745450e-01 -9.33251619e-01 -5.35543203e-01
-8.16321909e-01 7.27281868e-02 4.48532045e-01 5.46354830e-01
-4.13504004e-01 4.08265918e-01 -3.52246523e-01 1.83377966e-01
9.41034317e-01 -4.16688681e-01 -5.51487327e-01 -2.52473623e-01
-1.64346457e+00 9.43091273e-01 4.75569516e-01 1.30214974e-01
-1.23258138e+00 -3.81823868e-01 -9.17881906e-01 -3.40025127e-01
6.15471482e-01 -4.81333554e-01 1.00639498e+00 -9.33051050e-01
-1.14442360e+00 1.19630063e+00 1.53169408e-01 -1.67767599e-01
1.09278941e+00 -2.56744772e-01 -7.15722620e-01 5.08909047e-01
2.99804837e-01 2.41306126e-01 1.58557999e+00 -1.75655746e+00
-4.85403836e-01 -2.48672664e-01 3.61123472e-01 6.98796585e-02
-5.34807742e-01 1.78778410e-01 -4.05135155e-01 -1.09867656e+00
-8.30791797e-03 -5.16340673e-01 6.56769812e-01 1.18298404e-01
-1.02115929e+00 7.13014081e-02 1.21462691e+00 -1.26940608e+00
1.50933945e+00 -2.20930314e+00 3.68563592e-01 -2.28929669e-01
7.99116731e-01 1.16761521e-01 -1.42810717e-01 4.98272449e-01
1.12545460e-01 2.45276675e-01 -1.35053426e-01 -4.13992435e-01
4.51296329e-01 -1.62537754e-01 -6.64896846e-01 9.22905982e-01
2.07571074e-01 1.47363508e+00 -1.16076231e+00 -6.39878750e-01
2.33794793e-01 3.65975616e-03 -3.01007688e-01 1.03806891e-01
-4.95139778e-01 3.21994007e-01 -6.87811375e-02 1.48000383e+00
7.38447011e-01 -4.77020502e-01 -3.34562629e-01 -9.13385928e-01
4.07160133e-01 -1.90847993e-01 -6.45219147e-01 1.53863394e+00
-2.92906344e-01 8.07595253e-01 6.55386746e-02 -9.07496512e-01
9.53104645e-02 1.82827309e-01 1.33136511e-01 -8.37230325e-01
5.89052320e-01 2.21663088e-01 -4.22077924e-01 -1.17375112e+00
9.12948668e-01 -1.39377937e-01 -6.74312472e-01 4.88102436e-01
-1.75785404e-02 -1.20404497e-01 -1.11739881e-01 3.50868613e-01
1.12426865e+00 -8.18851516e-02 1.19708240e-01 5.13522863e-01
4.09724474e-01 8.40904471e-03 -2.51815647e-01 1.02583289e+00
-4.83785808e-01 3.95728081e-01 6.24408841e-01 -4.57955115e-02
-9.46130216e-01 -8.36943686e-01 7.27858245e-02 1.22517586e+00
8.71450007e-01 -2.59781122e-01 -4.70991462e-01 -1.11852014e+00
2.65961200e-01 4.69954371e-01 -7.68607199e-01 -4.20179278e-01
-4.88673776e-01 -6.25819087e-01 1.16110802e+00 -4.30257618e-02
1.03234327e+00 -1.09513044e+00 -2.37479687e-01 -2.53715962e-01
-9.41402853e-01 -1.46894848e+00 -4.32634324e-01 -4.93307650e-01
-2.00734437e-01 -1.30981708e+00 -5.26852608e-01 -7.07465529e-01
1.96907222e-01 5.92547894e-01 9.76849139e-01 4.27670449e-01
-1.65695161e-01 3.56504679e-01 -6.00308061e-01 5.35615757e-02
-6.76552474e-01 -4.37919080e-01 -1.69085369e-01 2.59174615e-01
9.33706760e-02 -6.51107356e-02 -3.66455585e-01 3.81737560e-01
-1.31400979e+00 1.87554002e-01 6.19801760e-01 1.05607116e+00
8.04047212e-02 4.12960500e-01 2.63758928e-01 -7.65850067e-01
7.70763636e-01 -8.99671316e-01 -1.65531084e-01 3.75427991e-01
4.60419096e-02 -4.08195883e-01 5.01021564e-01 -8.18960667e-01
-8.54276240e-01 -6.72884405e-01 2.66320676e-01 -9.02489245e-01
-1.31618492e-02 4.42823142e-01 -3.40811878e-01 -3.41054142e-01
7.18678713e-01 4.47234839e-01 -2.08632231e-01 -1.52390927e-01
7.32303858e-01 8.13938856e-01 7.72417903e-01 -4.80773747e-01
1.11626995e+00 8.12918186e-01 -2.11633876e-01 -6.00658894e-01
-9.44750488e-01 -3.51704359e-01 -2.59780228e-01 -4.22717541e-01
7.66516745e-01 -8.06231618e-01 -1.13830709e+00 1.20889628e+00
-1.50753224e+00 -2.67143905e-01 2.25556388e-01 -1.67731792e-01
-4.58968341e-01 1.05327046e+00 -1.05819035e+00 -5.51924229e-01
-4.76200208e-02 -1.24290907e+00 1.63967264e+00 -5.62326491e-01
9.24042985e-02 -8.13487232e-01 -6.20447457e-01 1.06257296e+00
3.51256877e-02 5.86576819e-01 8.64233375e-01 -4.42351937e-01
-7.93055415e-01 -4.02377039e-01 -8.38089228e-01 2.69892663e-01
-9.93854851e-02 -4.24082816e-01 -8.70878577e-01 -1.45134389e-01
7.22670034e-02 -8.81752193e-01 1.02544141e+00 -3.45402867e-01
1.43960714e+00 -6.84662342e-01 -2.08491191e-01 3.68792057e-01
1.24173033e+00 -4.49065417e-01 7.71607816e-01 4.38288659e-01
1.35525036e+00 4.49076861e-01 6.85823858e-01 5.07713079e-01
4.91044462e-01 6.50848389e-01 9.73396480e-01 2.01782018e-01
-2.81359911e-01 -3.82361233e-01 6.19399607e-01 7.64336050e-01
3.42918068e-01 -6.88493252e-01 -6.87065005e-01 3.85938615e-01
-1.91041386e+00 -1.47523081e+00 -4.20565426e-01 1.67067695e+00
7.30324268e-01 4.92365509e-02 1.24004699e-01 3.25425804e-01
9.31122124e-01 4.49981242e-01 -4.44454283e-01 1.02333434e-01
-5.12554407e-01 -4.08029974e-01 4.63536382e-01 2.61462122e-01
-1.50066233e+00 1.09456635e+00 5.33243084e+00 1.36376572e+00
-9.33579922e-01 5.82215488e-01 2.74899036e-01 1.14701375e-01
-3.10243547e-01 -4.39699471e-01 -4.38840270e-01 1.01319516e+00
1.85254484e-01 4.08429652e-01 7.39917517e-01 4.19390827e-01
5.19284829e-02 -1.43252596e-01 -9.21966434e-01 1.39883971e+00
6.14950478e-01 -1.44323862e+00 3.83047521e-01 -1.18092895e-01
8.17407310e-01 -3.89152497e-01 2.88510382e-01 4.95849341e-01
1.33753857e-02 -9.07511175e-01 1.31489563e+00 4.29327279e-01
9.76103008e-01 -3.58248949e-01 7.35411525e-01 4.88268197e-01
-8.30393612e-01 -3.03699225e-01 2.01039433e-01 1.66975543e-01
4.10020441e-01 5.26273429e-01 -2.50949919e-01 6.35218382e-01
5.57643056e-01 8.72577786e-01 -5.55984795e-01 4.58502322e-01
-2.62373865e-01 1.66714877e-01 1.53230518e-01 1.42671496e-01
2.24383280e-01 2.81380117e-01 6.61387742e-01 1.37460995e+00
2.45111436e-02 -1.92086712e-01 1.51016012e-01 8.72414231e-01
-5.22335351e-01 -3.09227347e-01 -5.19795299e-01 -5.05557895e-01
3.79599750e-01 1.05977738e+00 -3.98359597e-01 -5.78409433e-01
-4.19466615e-01 1.64821589e+00 3.57163668e-01 1.54943913e-01
-1.51950002e+00 -2.27774128e-01 4.00680453e-01 -1.23213558e-02
1.89551890e-01 7.62277842e-02 -2.20149130e-01 -1.76395726e+00
2.49047324e-01 -1.40174627e+00 4.24212039e-01 -1.02136266e+00
-1.88771176e+00 7.46230409e-02 -3.42716090e-02 -1.27072918e+00
3.63221616e-01 -7.85953760e-01 -9.94708091e-02 4.17841792e-01
-1.46847749e+00 -1.80004919e+00 -6.88666821e-01 8.26154470e-01
4.51958507e-01 -4.66138832e-02 3.80978554e-01 6.42359495e-01
-5.68442941e-01 9.41829264e-01 -2.63633072e-01 5.43569863e-01
7.02116847e-01 -8.70085835e-01 -1.73297900e-04 8.76001477e-01
-1.89250827e-01 3.39276165e-01 5.60205579e-01 -9.04526472e-01
-1.78402936e+00 -1.14133894e+00 4.65754032e-01 -7.86671460e-01
1.26577723e+00 -3.03324848e-01 -8.16326976e-01 7.09409058e-01
7.19255954e-02 2.72232536e-02 2.52519369e-01 -3.05656552e-01
-9.44466114e-01 6.25827312e-01 -1.41387022e+00 7.17002928e-01
1.19755530e+00 -1.11050093e+00 -4.99402285e-01 8.28016937e-01
8.23603809e-01 -8.51463675e-01 -7.26212323e-01 3.32010508e-01
5.21780074e-01 -1.10766900e+00 1.03736639e+00 -7.87937522e-01
1.33595562e+00 -1.61651582e-01 -3.98819983e-01 -9.01310205e-01
-9.87476185e-02 -5.81111848e-01 -8.83238852e-01 1.04388642e+00
-1.85049787e-01 -2.73361534e-01 3.82477492e-01 -1.10552333e-01
-1.91638127e-01 -5.50045669e-01 -7.83013225e-01 -6.37029707e-01
-1.59441546e-01 -7.63096213e-01 4.70092684e-01 1.50895774e+00
2.27999315e-01 1.67691752e-01 -9.59109187e-01 4.58234519e-01
6.54646933e-01 1.44487053e-01 8.57117236e-01 -4.32288587e-01
-4.19416636e-01 -7.60931671e-01 -6.29641831e-01 -1.24798143e+00
2.86396593e-01 -8.55617583e-01 -5.53451590e-02 -1.05788875e+00
5.25271356e-01 3.56220827e-02 2.27120951e-01 4.53150183e-01
-1.71878889e-01 7.89066076e-01 8.51584822e-02 3.49643141e-01
-8.97048414e-01 4.32512969e-01 1.63664091e+00 -6.36821985e-01
4.73705292e-01 -4.81357276e-01 -5.37169814e-01 7.83467889e-01
3.63810986e-01 -2.12396666e-01 -2.09663063e-02 -5.38057148e-01
4.66711789e-01 3.43949527e-01 1.18177676e+00 -6.22242808e-01
-5.53872474e-02 -2.29298756e-01 1.68941259e-01 -6.95067525e-01
4.48566496e-01 -8.07739377e-01 -2.57390916e-01 3.68269980e-01
-4.01944995e-01 -2.18028784e-01 -3.47692072e-02 9.44029033e-01
-1.73915625e-01 -1.06963061e-01 7.94952214e-01 -2.56234705e-01
-8.06696117e-01 2.60698438e-01 -1.93531647e-01 3.02973032e-01
9.86684561e-01 -1.31259084e-01 -9.95960057e-01 -5.41007996e-01
-6.87494516e-01 -1.07632525e-01 4.63407844e-01 5.86520851e-01
6.12775385e-01 -1.63273537e+00 -5.41431904e-01 -1.84843898e-01
5.96249878e-01 -2.93265134e-01 5.50597012e-01 9.98318672e-01
-5.36477566e-01 1.67322252e-02 -4.40421216e-02 -4.37945217e-01
-1.15653729e+00 1.06901407e+00 2.05376938e-01 -1.94113135e-01
-4.89366144e-01 8.76280189e-01 -5.02956435e-02 -2.07152769e-01
1.68168202e-01 -1.95161939e-01 2.51714945e-01 1.70174107e-01
6.54705048e-01 5.62902331e-01 -9.74064618e-02 -1.03704119e+00
-4.25001718e-02 7.11501837e-02 1.02428101e-01 3.05842981e-02
7.04977334e-01 -4.06427532e-01 -4.36213315e-01 3.82645816e-01
1.47489202e+00 7.43299071e-03 -7.96353877e-01 -4.85420495e-01
-1.94986299e-01 -9.65589821e-01 -3.94866876e-02 -9.85871911e-01
-1.02679574e+00 8.22155654e-01 5.87510616e-02 4.79318589e-01
1.03477716e+00 5.65095954e-02 1.26532042e+00 2.61505693e-01
3.51805478e-01 -7.67216325e-01 9.62487161e-01 4.44797099e-01
1.27123046e+00 -1.72366273e+00 -1.00432094e-02 -6.05163574e-01
-6.79991722e-01 9.14799154e-01 7.01466382e-01 8.04639086e-02
3.04661304e-01 7.73270279e-02 -2.48197660e-01 -7.05104470e-01
-1.17656536e-01 -9.71582308e-02 2.55634457e-01 3.92910689e-01
-1.89913914e-01 -5.01578161e-03 1.87934171e-02 5.18887460e-01
-7.07084732e-03 -1.05070740e-01 3.36613357e-01 8.08006644e-01
-1.11803934e-01 -4.63400662e-01 -6.81819379e-01 5.24399698e-01
-3.03480774e-01 -2.33441040e-01 -5.56508899e-01 8.05153668e-01
5.12805879e-01 9.75092947e-01 -2.86808938e-01 -6.65434182e-01
1.06328443e-01 -3.02864611e-01 6.72348320e-01 -1.58486128e-01
-6.40913785e-01 -3.23766410e-01 1.02059796e-01 -6.66447520e-01
-2.78486460e-01 -4.40712005e-01 -6.58665478e-01 -8.28721881e-01
-5.72617948e-01 -6.81065679e-01 3.25926393e-01 1.11525393e+00
1.55366972e-01 4.40524191e-01 5.80071568e-01 -1.05838203e+00
-7.18923986e-01 -9.66648996e-01 -4.39415962e-01 1.06609118e+00
7.12865472e-01 -9.94128108e-01 -6.19107068e-01 2.00266838e-01] | [8.175990104675293, 10.300833702087402] |
5f13d07c-91a2-4109-88fe-77ca12845f7a | deep-transfer-learning-for-land-use-land | 2110.02580 | null | https://arxiv.org/abs/2110.02580v3 | https://arxiv.org/pdf/2110.02580v3.pdf | Deep Transfer Learning for Land Use and Land Cover Classification: A Comparative Study | Efficiently implementing remote sensing image classification with high spatial resolution imagery can provide a significant value in Land Use and Land Cover (LULC) classification. The new advances in remote sensing and deep learning technologies have facilitated the extraction of spatiotemporal information for LULC classification. Moreover, the diverse disciplines of science, including remote sensing, have utilised tremendous improvements in image classification by Convolutional Neural Networks (CNNs) with transfer learning. In this study, instead of training CNNs from scratch, the transfer learning is applied to fine-tune pre-trained networks Visual Geometry Group (VGG16) and Wide Residual Networks (WRNs), by replacing the final layer with additional layers, for LULC classification using the red-green-blue version of the EuroSAT dataset. Moreover, the performance and computational time are compared and optimised with techniques, such as early stopping, gradient clipping, adaptive learning rates, and data augmentation. The proposed approaches have addressed the limited-data problem, and very good accuracies are achieved. The results show that the proposed method based on the WRNs performs better than the previous best-stated results in terms of the computational efficiency and accuracy from 98.57% to 99.17%. | ['Ebrahim Ghaderpour', 'Tarunpreet Kaur', 'Raoof Naushad'] | 2021-10-06 | null | null | null | null | ['remote-sensing-image-classification'] | ['miscellaneous'] | [ 2.41354436e-01 -2.53719419e-01 -1.90246001e-01 -4.45184082e-01
-5.13069570e-01 -2.23675266e-01 6.66292727e-01 -1.96077347e-01
-6.53157890e-01 1.14985323e+00 -3.57794799e-02 -6.39053345e-01
-2.65113294e-01 -1.28886175e+00 -5.37628412e-01 -9.51216698e-01
-4.58380580e-01 -8.45368430e-02 -7.98830315e-02 -4.93853658e-01
-1.20501816e-01 9.58381653e-01 -1.85000682e+00 8.52359012e-02
1.16947854e+00 1.03434098e+00 3.62229645e-01 6.15387738e-01
-1.14248106e-02 6.09191597e-01 -8.53185579e-02 2.40910470e-01
5.87008238e-01 -6.12797588e-02 -6.14766002e-01 4.07743230e-02
5.83471775e-01 -5.99425375e-01 -1.78117290e-01 8.68652225e-01
5.59592485e-01 3.27553630e-01 5.14683962e-01 -6.62636995e-01
-5.01750886e-01 3.29415590e-01 -7.87113369e-01 1.93783179e-01
-5.78486145e-01 7.36509413e-02 6.90311551e-01 -8.34992111e-01
3.78733665e-01 9.19569016e-01 9.87077177e-01 -3.06730326e-02
-1.12934935e+00 -8.87772739e-01 1.18145175e-01 6.54912591e-02
-1.82153678e+00 -3.32754850e-01 3.71020108e-01 -5.62395096e-01
1.09921718e+00 4.31993783e-01 6.41069055e-01 3.47335249e-01
5.82552515e-02 4.15457994e-01 1.02065182e+00 -4.23299372e-01
1.62299797e-02 -1.32729128e-01 -1.06661581e-01 3.84290129e-01
4.69131827e-01 4.84065950e-01 3.74283433e-01 1.45854652e-01
8.48719954e-01 3.91595006e-01 -1.19653970e-01 9.01259296e-03
-8.39644372e-01 1.03247702e+00 1.26598084e+00 5.62451303e-01
-6.45546079e-01 1.76074192e-01 1.78203061e-01 2.18142360e-01
8.91567409e-01 3.98638189e-01 -6.34052932e-01 5.35401106e-01
-1.34057486e+00 2.01547414e-01 1.96530238e-01 6.65612578e-01
1.19604003e+00 3.76677543e-01 1.14684738e-01 1.02643728e+00
3.29503536e-01 7.55498946e-01 2.27748811e-01 -4.96728510e-01
6.11444950e-01 8.48833859e-01 9.04655531e-02 -1.22570205e+00
-7.27010012e-01 -8.19642484e-01 -1.50921047e+00 5.30236304e-01
-4.84073423e-02 -2.45767474e-01 -1.31459010e+00 1.20969427e+00
3.66025716e-02 4.92689461e-02 3.13632429e-01 8.71808112e-01
8.92490208e-01 1.09867561e+00 5.37547052e-01 3.06151271e-01
9.22108412e-01 -6.81130409e-01 -4.49224085e-01 -3.89950216e-01
9.73075926e-01 -3.69927585e-01 7.79491186e-01 -1.58268005e-01
-3.71014148e-01 -9.36564684e-01 -1.17705166e+00 1.02887139e-01
-8.45934153e-01 8.32129598e-01 9.07448411e-01 5.89162886e-01
-1.15221417e+00 7.17875063e-01 -6.70910418e-01 -5.21021307e-01
8.17238271e-01 3.63346547e-01 -3.77305269e-01 -1.12157047e-01
-1.32066882e+00 6.49780393e-01 9.17333484e-01 9.29878056e-01
-6.09459341e-01 -7.05499053e-01 -8.41550946e-01 1.72430515e-01
-3.66941579e-02 -3.57189476e-02 4.62780088e-01 -1.26487815e+00
-1.00496078e+00 8.29128385e-01 3.86410743e-01 -5.47309697e-01
4.32841778e-01 -2.30121791e-01 -4.08079982e-01 -6.61712289e-02
1.81009218e-01 9.38127398e-01 4.81754959e-01 -1.17504728e+00
-1.05103505e+00 -3.19575697e-01 -6.55694157e-02 1.95412025e-01
-2.40923792e-01 -2.35928163e-01 1.50410235e-01 -6.51419997e-01
3.91467035e-01 -8.03048491e-01 -5.20239294e-01 -4.66937087e-02
1.85778782e-01 1.48052305e-01 8.78989100e-01 -8.87545645e-01
1.04442608e+00 -2.20675540e+00 -3.68605375e-01 3.90660048e-01
-4.53273803e-02 8.62059057e-01 -2.38567814e-01 2.25692242e-01
-1.85272127e-01 5.21897137e-01 -5.49657762e-01 3.24949980e-01
-4.76411194e-01 4.11892861e-01 -2.55357623e-01 6.66792810e-01
4.48041081e-01 1.11175478e+00 -6.71167791e-01 -2.87991047e-01
8.23252082e-01 5.62508047e-01 -1.74344346e-01 1.06572509e-01
-1.22564189e-01 5.05401433e-01 -4.73581374e-01 5.73897302e-01
1.25603497e+00 -6.25380501e-02 3.70344520e-02 -8.65482986e-02
-6.62089467e-01 -2.45182604e-01 -9.96361077e-01 1.14038444e+00
-5.74495614e-01 7.81023264e-01 -1.46596923e-01 -1.13121080e+00
1.32838857e+00 1.11902460e-01 3.69316995e-01 -9.33938086e-01
-1.85377210e-01 3.71703357e-01 -1.01456366e-01 -5.76136291e-01
5.23504972e-01 1.01850212e-01 3.17746907e-01 -5.22314571e-02
-2.89545923e-01 1.43620327e-01 -1.51502609e-01 -4.62579697e-01
4.10238594e-01 4.69574273e-01 2.56022841e-01 -4.45389926e-01
6.58875406e-01 4.45369422e-01 6.30940795e-02 6.88492715e-01
8.93550739e-02 5.29508710e-01 -9.53849778e-03 -1.01951826e+00
-1.23984206e+00 -4.26765442e-01 -4.24704671e-01 1.21678722e+00
-1.64618656e-01 3.91573817e-01 -2.68948764e-01 -1.90611392e-01
4.56817076e-02 4.20256943e-01 -6.93442225e-01 6.21829145e-02
-6.12103343e-01 -1.30355442e+00 7.80183554e-01 6.85470104e-01
1.40715420e+00 -1.26041186e+00 -7.23524213e-01 2.76662558e-01
2.67878100e-02 -9.97026563e-01 5.66208899e-01 1.86699852e-01
-1.19443858e+00 -6.77927554e-01 -9.40221250e-01 -6.31631672e-01
6.48134291e-01 3.94808948e-01 8.59562159e-01 3.08686823e-01
-4.04638857e-01 -4.62394208e-01 -5.50531566e-01 -3.79372425e-02
-1.16285712e-01 7.66043663e-01 -6.87601447e-01 -2.94337366e-02
2.81437814e-01 -5.92819810e-01 -6.12150729e-01 1.28158599e-01
-9.41751420e-01 3.60901028e-01 1.11990738e+00 7.71344900e-01
3.77180755e-01 8.45711529e-02 4.27958041e-01 -8.06876719e-01
-1.04103751e-01 -5.05736113e-01 -8.54529500e-01 1.47494644e-01
-6.98766887e-01 -6.18743598e-02 4.77153033e-01 -7.66819762e-03
-1.12465405e+00 3.66762847e-01 -3.64468336e-01 4.36040312e-02
-6.24886274e-01 9.90170836e-01 9.94208008e-02 -2.99377412e-01
9.50709820e-01 3.14248919e-01 -2.45832250e-01 -4.11240131e-01
1.77730575e-01 9.17673171e-01 3.87774885e-01 5.53499758e-02
8.71344209e-01 5.67682922e-01 3.68746817e-02 -1.25338376e+00
-7.31572628e-01 -6.14697814e-01 -9.62882876e-01 -1.51720628e-01
8.38656604e-01 -1.34685218e+00 -1.85455412e-01 7.58637607e-01
-6.69378221e-01 -6.71521842e-01 -7.12104663e-02 5.71939290e-01
-2.33326867e-01 -8.56125727e-02 -7.27451146e-02 -7.87396252e-01
-4.99052763e-01 -6.98240936e-01 9.67880726e-01 3.10315460e-01
4.51347560e-01 -9.08204138e-01 -2.22627655e-01 -7.95954317e-02
9.91161883e-01 7.23432302e-01 6.91519618e-01 -1.18765496e-01
-6.64591193e-01 -3.18313032e-01 -9.08895850e-01 4.52660799e-01
2.96736151e-01 -8.61347318e-02 -1.03023446e+00 -2.86238909e-01
-6.15336120e-01 -1.98669598e-01 1.25057042e+00 7.39266098e-01
1.14373195e+00 -2.44710743e-01 -2.90189832e-01 1.25173140e+00
2.01818919e+00 1.54137105e-01 1.17963099e+00 7.34707177e-01
7.54876494e-01 4.49164391e-01 5.85888803e-01 4.46223170e-01
7.76063725e-02 2.37494409e-01 9.10752058e-01 -9.01105285e-01
-3.48053910e-02 2.49631647e-02 -2.59477645e-01 -1.05974935e-01
-8.48237813e-01 5.55787608e-02 -1.17539585e+00 5.86209357e-01
-2.03237104e+00 -1.09382594e+00 -2.94994205e-01 1.93618393e+00
3.65389198e-01 -3.20423782e-01 -3.95277411e-01 2.43358582e-01
5.75823724e-01 5.22617936e-01 -3.86611730e-01 1.50439546e-01
-3.55270952e-01 3.77445728e-01 1.47951627e+00 5.23281217e-01
-1.68222570e+00 1.53976798e+00 5.64663267e+00 6.60693109e-01
-1.58212006e+00 -8.34619254e-02 6.81238711e-01 5.15239239e-01
1.96703881e-01 -2.95984507e-01 -8.80405724e-01 -7.01162964e-02
9.91635025e-01 4.35963511e-01 3.49908710e-01 8.00062835e-01
6.47630572e-01 -3.42572182e-01 -4.36288714e-02 6.41814947e-01
-3.65593433e-01 -1.49652123e+00 1.37149468e-01 1.23863958e-01
9.97990429e-01 4.37600195e-01 -6.85592517e-02 4.62736398e-01
3.80796343e-01 -1.39763010e+00 3.61122251e-01 5.91569006e-01
1.11400938e+00 -9.98754978e-01 1.19845021e+00 2.30444476e-01
-1.56734169e+00 -1.69910967e-01 -7.80274928e-01 -2.65545547e-01
-5.09221733e-01 2.71314859e-01 -6.56444073e-01 7.39543438e-01
9.58534837e-01 1.15877926e+00 -7.27636278e-01 9.19574082e-01
-2.31826842e-01 8.53132904e-01 -2.84479648e-01 5.77213280e-02
8.23676229e-01 -2.22865328e-01 -6.96635619e-02 1.23678863e+00
3.39930832e-01 2.88315952e-01 1.09406091e-01 7.24112689e-01
2.61419654e-01 1.85626790e-01 -6.54324353e-01 1.31835520e-01
2.95565110e-02 1.36478019e+00 -6.84391797e-01 -2.68319964e-01
-1.64196610e-01 4.45250958e-01 4.01871838e-02 5.97609580e-01
-6.84036970e-01 -6.19963527e-01 5.85958660e-01 2.20098719e-01
4.63746071e-01 -4.39688385e-01 -3.10042173e-01 -8.24281573e-01
-2.17156246e-01 -4.42005634e-01 1.51340455e-01 -8.53173494e-01
-5.64677238e-01 7.05445468e-01 8.62072110e-02 -1.39112723e+00
-1.78560928e-01 -7.39724696e-01 -3.28860641e-01 1.23410654e+00
-2.46098447e+00 -1.47790897e+00 -9.04760897e-01 4.28493589e-01
3.71369243e-01 -1.87983170e-01 8.66642594e-01 3.68104517e-01
-2.65423834e-01 1.38971254e-01 5.01471877e-01 4.19199854e-01
1.52851164e-01 -8.44016433e-01 3.62116009e-01 9.25699651e-01
-5.75306118e-01 -9.52276587e-03 1.70153230e-01 -4.60597694e-01
-9.26490843e-01 -2.00358891e+00 8.00358355e-01 6.35095537e-01
3.65306497e-01 -8.27412233e-02 -9.04598117e-01 6.45475268e-01
-2.43197426e-01 4.43190902e-01 2.08853781e-01 -1.40151650e-01
6.58591315e-02 -6.31827235e-01 -1.16920269e+00 3.66486996e-01
8.60014200e-01 -1.72537059e-01 -2.86129061e-02 3.23852986e-01
2.52161384e-01 -1.45168141e-01 -7.03909874e-01 7.74413228e-01
6.57771230e-01 -8.93237889e-01 9.86279070e-01 -4.42377597e-01
6.36970639e-01 -3.55646819e-01 -4.23370302e-01 -1.30551052e+00
-6.86020672e-01 2.49060929e-01 6.04557097e-01 7.59474337e-01
4.86643761e-01 -8.39432001e-01 6.85471237e-01 4.22054715e-02
-2.23760620e-01 -2.43697584e-01 -6.73904181e-01 -6.13728285e-01
1.60643771e-01 -3.34176421e-01 7.71395087e-01 8.45227599e-01
-8.96217287e-01 -1.48049191e-01 -4.09458190e-01 6.39587104e-01
3.36402804e-01 5.80855012e-02 7.69145668e-01 -1.57130778e+00
4.48125631e-01 -3.36767703e-01 -5.15656233e-01 -6.53078198e-01
5.03451983e-03 -8.39493930e-01 2.91384514e-02 -1.76641607e+00
-1.40304461e-01 -8.50212991e-01 -2.20222548e-01 1.01007259e+00
-1.33295149e-01 5.56783140e-01 -1.58790872e-01 4.99659061e-01
1.05241209e-01 5.97556472e-01 9.53364968e-01 -3.06823075e-01
-4.06321764e-01 2.15141609e-01 -3.46167445e-01 5.08012474e-01
1.11413050e+00 -3.47938299e-01 5.08857844e-03 -5.11529267e-01
3.61223072e-01 -2.36686826e-01 6.30998611e-01 -1.41790473e+00
-1.39774233e-01 -2.57040918e-01 7.71439433e-01 -1.06671596e+00
-1.25358939e-01 -1.11508274e+00 3.49679470e-01 6.77619696e-01
-2.30796769e-01 -3.59759063e-01 6.27352834e-01 7.64305890e-02
-2.65465528e-01 -1.83444783e-01 9.70335543e-01 -2.05514699e-01
-1.26872373e+00 5.22654593e-01 -5.44388294e-01 -5.96265256e-01
9.91864264e-01 -4.08223301e-01 -1.62357643e-01 -4.61968109e-02
-6.11711919e-01 2.31025130e-01 1.87534338e-03 2.70397514e-01
4.34600979e-01 -1.20394373e+00 -1.13454306e+00 6.31141365e-01
3.10611814e-01 1.97485805e-01 2.81809986e-01 5.53383052e-01
-1.15256143e+00 5.35826921e-01 -6.01940453e-01 -7.45706737e-01
-9.51209664e-01 -5.53812459e-02 7.18907416e-01 -4.60067421e-01
-6.85183346e-01 3.65568340e-01 -1.41270161e-01 -7.93670177e-01
-2.46958345e-01 -3.51652414e-01 -6.73293114e-01 1.08792298e-01
4.33579803e-01 3.21422338e-01 2.62939155e-01 -7.37565041e-01
-1.38655290e-01 6.89391792e-01 3.24749112e-01 1.67854965e-01
1.93651009e+00 -1.54141143e-01 -6.54131472e-02 1.49195999e-01
1.13630378e+00 -6.45589352e-01 -1.30436230e+00 -2.92600065e-01
-5.12254424e-03 -5.17125130e-01 6.63142681e-01 -8.04516077e-01
-1.29026020e+00 8.85255814e-01 1.09436154e+00 1.30373836e-01
1.25431728e+00 -4.17881072e-01 1.98925748e-01 6.49982154e-01
1.46585137e-01 -8.34737062e-01 -6.43960238e-01 6.93867207e-01
9.57126498e-01 -1.60100126e+00 1.55299634e-01 -8.20693374e-02
-3.08068305e-01 1.28184891e+00 4.38050777e-01 -4.42780197e-01
8.70697975e-01 1.27169132e-01 1.96166992e-01 -3.00415494e-02
-9.81470123e-02 -6.96223497e-01 2.53439158e-01 6.03727221e-01
5.29633224e-01 3.81764054e-01 -1.22209825e-01 -1.40299559e-01
-1.22376354e-02 3.62251222e-01 4.11518328e-02 8.60366106e-01
-7.96315730e-01 -4.42292720e-01 -5.44135928e-01 6.96121871e-01
-1.27977446e-01 -4.12249207e-01 1.88306198e-01 1.11364305e+00
3.00948381e-01 6.53054833e-01 2.00143993e-01 -2.18801454e-01
1.95831612e-01 -3.23054612e-01 -1.00503609e-01 -4.07723665e-01
-4.61663991e-01 -1.80242926e-01 -2.54786432e-01 -2.36225396e-01
-8.91332269e-01 -3.23808730e-01 -9.15429831e-01 -3.77170712e-01
-3.18255335e-01 -4.65699024e-02 8.95238578e-01 9.85670447e-01
2.25436255e-01 5.53464353e-01 7.85226762e-01 -1.37394512e+00
-3.46603841e-01 -1.34929490e+00 -7.34566569e-01 -1.88504964e-01
4.88983810e-01 -3.99270982e-01 -3.36506754e-01 -1.13519251e-01] | [9.52055835723877, -1.5041202306747437] |
a4c311ac-785b-4de9-b6f0-7006221ff49b | optimizing-yolov7-for-semiconductor-defect | 2302.09565 | null | https://arxiv.org/abs/2302.09565v1 | https://arxiv.org/pdf/2302.09565v1.pdf | Optimizing YOLOv7 for Semiconductor Defect Detection | The field of object detection using Deep Learning (DL) is constantly evolving with many new techniques and models being proposed. YOLOv7 is a state-of-the-art object detector based on the YOLO family of models which have become popular for industrial applications. One such possible application domain can be semiconductor defect inspection. The performance of any machine learning model depends on its hyperparameters. Furthermore, combining predictions of one or more models in different ways can also affect performance. In this research, we experiment with YOLOv7, a recently proposed, state-of-the-art object detector, by training and evaluating models with different hyperparameters to investigate which ones improve performance in terms of detection precision for semiconductor line space pattern defects. The base YOLOv7 model with default hyperparameters and Non Maximum Suppression (NMS) prediction combining outperforms all RetinaNet models from previous work in terms of mean Average Precision (mAP). We find that vertically flipping images randomly during training yields a 3% improvement in the mean AP of all defect classes. Other hyperparameter values improved AP only for certain classes compared to the default model. Combining models that achieve the best AP for different defect classes was found to be an effective ensembling strategy. Combining predictions from ensembles using Weighted Box Fusion (WBF) prediction gave the best performance. The best ensemble with WBF improved on the mAP of the default model by 10%. | ['Stefan De Gendt', 'Sandip Halder', 'Bappaditya Dey', 'Enrique Dehaerne'] | 2023-02-19 | null | null | null | null | ['defect-detection'] | ['computer-vision'] | [-1.43803313e-01 -1.86964601e-01 1.62720695e-01 1.27580658e-01
-4.38928068e-01 5.65988419e-04 3.80943716e-01 7.14320466e-02
-3.14379841e-01 4.39117342e-01 -5.79516292e-01 -1.64063275e-01
-1.18263260e-01 -6.48743391e-01 -4.90623116e-01 -8.96764517e-01
2.51672447e-01 4.07276243e-01 1.11355007e+00 -1.88552842e-01
4.27972198e-01 7.02113092e-01 -1.68490696e+00 7.73364782e-01
6.32091403e-01 1.11706734e+00 5.50205052e-01 7.07594991e-01
4.91768196e-02 5.22615373e-01 -1.10005772e+00 -2.50011474e-01
1.77770019e-01 -2.21838132e-01 -3.07064623e-01 -1.63293064e-01
4.92110491e-01 9.62522775e-02 4.22604568e-02 9.04971957e-01
8.90965164e-01 -3.50419432e-01 7.77805388e-01 -9.95975673e-01
-4.08005238e-01 2.87841678e-01 -6.68827057e-01 3.80957216e-01
1.44876331e-01 3.73556167e-01 6.05830193e-01 -9.23925042e-01
5.30486166e-01 1.12743378e+00 8.90954435e-01 4.76233065e-01
-1.31596065e+00 -6.53073609e-01 -2.40183532e-01 4.78525758e-01
-1.31230307e+00 -1.86487120e-02 5.24424136e-01 -4.71609473e-01
1.51702106e+00 -7.32466727e-02 3.25353950e-01 5.57735264e-01
5.54646611e-01 6.66202426e-01 8.66506994e-01 -8.40530217e-01
2.11876407e-01 5.13947070e-01 2.56181538e-01 9.31950748e-01
4.69314098e-01 1.79798335e-01 -4.89746839e-01 1.42887175e-01
7.05686688e-01 -4.21268612e-01 -7.97193795e-02 -3.22822809e-01
-7.86205173e-01 7.49537289e-01 5.34578621e-01 6.92341745e-01
-2.19105721e-01 6.20895214e-02 3.45788538e-01 4.46715593e-01
1.86468020e-01 8.76195908e-01 -5.42626023e-01 2.40695402e-01
-7.77319193e-01 2.55440474e-01 5.52140713e-01 6.02312148e-01
5.05268037e-01 1.87926173e-01 -3.98398161e-01 1.18402147e+00
4.36149031e-01 1.93116397e-01 5.67109942e-01 -6.32750392e-01
1.22898601e-01 9.35049236e-01 -2.16983687e-02 -5.88693559e-01
-6.88559830e-01 -5.25988102e-01 -3.91467214e-01 9.35755014e-01
6.75116301e-01 2.14121491e-02 -1.24325597e+00 9.38547194e-01
1.49915349e-02 -1.67595237e-01 -8.34057555e-02 6.89099371e-01
1.05855143e+00 5.72698057e-01 -3.03838477e-02 1.19270973e-01
1.25117266e+00 -1.01928759e+00 -4.55897093e-01 1.30310943e-02
9.52039838e-01 -9.92746115e-01 6.78106904e-01 1.01421237e+00
-8.21525455e-01 -8.54416370e-01 -1.57929087e+00 2.58199543e-01
-5.43114960e-01 7.47216582e-01 2.91646183e-01 1.11284792e+00
-1.00447583e+00 7.36255407e-01 -6.97655976e-01 -5.49655378e-01
4.40656692e-01 7.93014586e-01 -2.69574761e-01 -1.63713440e-01
-6.57141030e-01 1.09432220e+00 4.38809186e-01 -1.93533361e-01
-7.18831480e-01 -6.65652335e-01 -3.68728787e-01 -3.20368353e-03
3.13732713e-01 -4.24768597e-01 1.11020887e+00 -5.85845470e-01
-1.55040646e+00 7.77953506e-01 3.24013352e-01 -8.86231303e-01
3.84256691e-01 -1.05520725e-01 -4.99105722e-01 -8.00046921e-02
-2.88854331e-01 8.37781429e-01 7.93022454e-01 -1.09823346e+00
-6.74007118e-01 -2.50715703e-01 -6.08605631e-02 -3.76415282e-01
-4.57067378e-02 2.55885087e-02 -3.06575656e-01 -1.99001491e-01
3.93216349e-02 -7.86458075e-01 8.81812274e-02 -2.25711856e-02
-2.73081064e-01 -5.69754243e-01 1.03527057e+00 -3.88168812e-01
1.14599824e+00 -2.12291169e+00 -2.26015463e-01 -3.11579891e-02
8.74356255e-02 8.64460230e-01 -9.76088122e-02 3.82421881e-01
-1.56535089e-01 -1.73801780e-01 1.27963945e-01 -2.29689613e-01
-1.63087532e-01 -1.32032186e-01 2.86266774e-01 4.36991096e-01
2.57540643e-01 4.22173858e-01 -4.01293814e-01 -3.49104464e-01
5.79674900e-01 3.81682843e-01 -5.32092035e-01 -8.07110667e-02
-1.58002332e-01 6.89553376e-03 -5.64843491e-02 8.21523070e-01
7.00581908e-01 -1.22210555e-01 -3.09508473e-01 -3.23445529e-01
-9.12796408e-02 -1.06043763e-01 -1.24981976e+00 1.49075937e+00
-4.10715312e-01 9.59107637e-01 -2.81796128e-01 -8.96756232e-01
1.37050951e+00 2.80822456e-01 2.31902912e-01 -6.89382076e-01
2.57516801e-01 3.59390289e-01 7.23411560e-01 -4.66255337e-01
2.31939316e-01 2.58599059e-04 4.71774429e-01 -1.09211385e-01
4.90495056e-01 -2.87847519e-01 3.05928499e-01 -1.15076341e-01
1.47296131e+00 -3.95192131e-02 1.93347678e-01 -2.53093392e-01
5.88316679e-01 -4.17672209e-02 3.82602900e-01 9.21900749e-01
-2.69740611e-01 8.57326746e-01 5.81215441e-01 -4.72122401e-01
-1.01743829e+00 -9.30165887e-01 -4.56952542e-01 7.84664452e-01
4.85044904e-02 -1.78414106e-01 -7.16053188e-01 -7.09140956e-01
1.96639553e-01 8.43027174e-01 -2.08328083e-01 -1.88518479e-01
-3.18299085e-01 -1.17022300e+00 5.73103428e-01 5.35156369e-01
5.43896079e-01 -1.35785651e+00 -7.54165053e-01 1.20804317e-01
6.10411227e-01 -9.96966124e-01 5.60738981e-01 4.19106811e-01
-8.45032156e-01 -1.18064165e+00 -8.34247053e-01 -6.83518291e-01
4.85422730e-01 -9.14067775e-02 1.13204885e+00 1.44502148e-01
-8.09964836e-01 1.54380158e-01 -4.58424270e-01 -8.58689785e-01
-5.68162978e-01 -1.37587652e-01 -1.44410282e-01 -3.17684472e-01
5.79353690e-01 9.34278816e-02 -3.37814331e-01 3.38494897e-01
-6.73313081e-01 -3.17326993e-01 7.09574580e-01 9.89018440e-01
3.06788564e-01 3.76574695e-01 5.26392102e-01 -8.59816611e-01
5.05469561e-01 -1.10924214e-01 -8.67981613e-01 1.28221065e-01
-6.50176764e-01 2.66421959e-02 2.60432363e-01 -2.73609430e-01
-1.00322497e+00 5.18936999e-02 -2.42743880e-01 -4.93171781e-01
-4.10720199e-01 -5.03425188e-02 -1.77410487e-02 -4.53866273e-01
9.55723524e-01 -3.33286464e-01 -2.03929231e-01 -6.14132166e-01
-7.06341490e-02 6.44687831e-01 2.19542131e-01 -1.89959601e-01
2.51065791e-01 2.68219829e-01 1.49294689e-01 -1.08658600e+00
-4.44840312e-01 -6.66167438e-01 -7.62828410e-01 -4.34061676e-01
9.62372482e-01 -5.25419176e-01 -6.52981281e-01 9.18232083e-01
-1.30603826e+00 -1.35581881e-01 -9.75130126e-02 5.76006353e-01
-2.65355945e-01 6.35980293e-02 -4.57633346e-01 -8.74018013e-01
-1.11927785e-01 -1.30641997e+00 1.07848799e+00 3.68661344e-01
3.75853814e-02 -7.43744433e-01 -6.88144043e-02 1.65055007e-01
2.36069217e-01 1.74583998e-02 9.46515739e-01 -6.43858254e-01
-4.85006928e-01 -4.73710775e-01 -4.07875806e-01 7.08099663e-01
-1.14500105e-01 6.38401285e-02 -1.23170149e+00 -2.78284281e-01
-2.94654667e-02 -1.14627473e-01 1.43543363e+00 6.88088238e-01
9.86104131e-01 6.67521000e-01 -5.91482639e-01 1.70974225e-01
1.69195211e+00 5.95095456e-01 8.14215899e-01 6.65510595e-01
4.52103019e-01 3.12481761e-01 5.66047013e-01 2.13709742e-01
-3.68920982e-01 8.91809821e-01 6.93627000e-01 2.19637286e-02
-5.37598968e-01 3.62464756e-01 2.82295644e-01 1.94407374e-01
-1.03513010e-01 -5.47680259e-01 -1.04306650e+00 4.01080340e-01
-1.56119394e+00 -7.65380085e-01 -3.32607806e-01 2.19269300e+00
5.50938072e-03 7.33473361e-01 9.91249457e-02 6.38666928e-01
6.13529086e-01 -3.10716778e-01 -4.12977934e-01 -6.63306117e-01
-7.43863657e-02 3.41219366e-01 7.17281282e-01 4.26848941e-02
-1.01850998e+00 8.27282608e-01 6.22110605e+00 9.32310522e-01
-1.14872932e+00 2.35839918e-01 4.12688076e-01 -1.33045107e-01
5.56348085e-01 -2.90397346e-01 -1.20107961e+00 4.02588010e-01
8.67387235e-01 5.12994528e-01 -1.24363579e-01 8.38022292e-01
-2.00901300e-01 -7.18456149e-01 -1.05179477e+00 1.02823496e+00
3.95001680e-01 -1.33380473e+00 -3.45604897e-01 5.85950306e-03
8.32701504e-01 1.34179801e-01 5.49958907e-02 4.14630353e-01
7.51024783e-02 -9.10261929e-01 5.19730747e-01 6.42667294e-01
3.49447787e-01 -6.31550908e-01 1.38720930e+00 2.12641329e-01
-7.24294960e-01 -4.84472156e-01 -7.07203329e-01 1.34009212e-01
-7.85415899e-03 7.02399671e-01 -1.14699876e+00 3.52423072e-01
9.66576159e-01 4.57762122e-01 -8.77035797e-01 1.67256272e+00
-1.72641367e-01 6.53129101e-01 -3.91517133e-01 -2.40925387e-01
1.77596122e-01 2.44814873e-01 5.66970050e-01 1.18673694e+00
4.20473605e-01 -3.52296412e-01 -1.76382646e-01 8.40417981e-01
3.52033228e-01 -1.64414853e-01 -5.90850711e-01 2.40428671e-01
2.18403578e-01 9.51433480e-01 -9.46460724e-01 -1.29326925e-01
-5.58855236e-01 5.98622203e-01 -7.81705678e-02 1.34537369e-03
-7.36044705e-01 -4.49777752e-01 3.30016404e-01 4.16813999e-01
7.61579871e-01 1.50447451e-02 -4.26803559e-01 -3.44517171e-01
-2.27097064e-01 -5.19425690e-01 3.00876886e-01 -9.46097434e-01
-1.10859597e+00 6.05387509e-01 -9.75181013e-02 -1.39509785e+00
7.34456703e-02 -1.45789230e+00 -6.69489920e-01 7.28004456e-01
-1.09149337e+00 -1.00021374e+00 -2.42519692e-01 6.99598948e-03
7.05564022e-01 -7.40871131e-01 4.67857867e-01 2.62962312e-01
-6.91857994e-01 5.39650917e-01 9.43493843e-02 -2.27470756e-01
7.48972893e-01 -1.23918593e+00 4.22747992e-02 6.59711063e-01
3.41658652e-01 -7.15329871e-03 7.52964199e-01 -4.42538381e-01
-7.81027973e-01 -8.14285457e-01 5.50544024e-01 -3.35157543e-01
3.24976414e-01 5.66653796e-02 -1.01768482e+00 1.26763597e-01
2.07918450e-01 -5.97038791e-02 2.45778516e-01 -3.22996788e-02
-8.22847188e-02 -2.83337474e-01 -1.40645647e+00 1.93499029e-01
5.29590070e-01 -1.53586283e-01 -3.95798713e-01 2.97709316e-01
3.70683491e-01 -3.38889301e-01 -6.71474576e-01 6.18460894e-01
5.76860487e-01 -1.54564059e+00 8.84150147e-01 -4.00397517e-02
1.37158647e-01 -5.92984319e-01 3.33161734e-04 -1.23009336e+00
-3.93271774e-01 9.19880420e-02 1.29284263e-01 1.11548626e+00
5.22301316e-01 -5.65734744e-01 8.72751296e-01 1.54113434e-02
-3.96131366e-01 -8.66753995e-01 -9.18923974e-01 -1.06046772e+00
-1.75171658e-01 -6.40134871e-01 2.60281771e-01 1.91628307e-01
-2.98822045e-01 2.32141688e-01 -6.45928532e-02 2.13187799e-01
1.75189495e-01 -3.36608171e-01 5.72459757e-01 -1.39570916e+00
-3.97915483e-01 -7.17047513e-01 -1.23141170e+00 -4.93466556e-01
-3.42218667e-01 -7.16168523e-01 -1.10801216e-02 -1.64529002e+00
-1.51145801e-01 -4.25228506e-01 -5.13788104e-01 3.16119581e-01
2.14535251e-01 5.32284677e-01 4.04216051e-01 -4.54953946e-02
-4.12303358e-01 -4.28337008e-02 9.32738841e-01 -1.56398222e-01
-3.53775889e-01 1.95858449e-01 2.11441237e-02 7.41408646e-01
7.81960309e-01 -4.93360549e-01 -1.13357313e-01 -1.70826033e-01
3.61192785e-02 -2.99952269e-01 2.80834198e-01 -1.83753049e+00
2.61491805e-01 7.47347355e-01 8.01478505e-01 -8.39504957e-01
5.66095591e-01 -6.58906043e-01 -1.67080402e-01 9.12650704e-01
1.79081738e-01 -1.08683243e-01 6.12862647e-01 4.46625084e-01
-1.61845177e-01 -8.29034567e-01 1.25010335e+00 -5.33789806e-02
-1.01388812e+00 -3.36034328e-01 -4.37938899e-01 -6.42108560e-01
1.27380419e+00 -6.29996479e-01 -4.36686218e-01 1.51692793e-01
-8.00018311e-01 -1.31917387e-01 3.37261081e-01 3.77886176e-01
4.02623266e-01 -9.63462830e-01 -4.80380267e-01 4.48323697e-01
2.63870448e-01 -3.72577816e-01 1.78292483e-01 9.17017639e-01
-7.22834527e-01 6.55557156e-01 -5.49581587e-01 -9.37636077e-01
-1.61910892e+00 4.00565386e-01 6.82639539e-01 -3.99271429e-01
-3.15418094e-01 1.25700521e+00 -1.88361593e-02 -1.61977246e-01
2.40131557e-01 -5.38538158e-01 -5.83505809e-01 -1.15901697e-02
4.01217520e-01 8.24866056e-01 6.47350729e-01 -3.18183810e-01
-2.57616699e-01 6.75758719e-01 -2.35342070e-01 2.76577145e-01
1.11434233e+00 5.18998027e-01 1.55442208e-01 5.57060182e-01
7.84744382e-01 -2.53676087e-01 -9.02515352e-01 1.77501187e-01
1.40462682e-01 -2.67416835e-01 3.56360108e-01 -1.17707741e+00
-1.06195331e+00 1.01104963e+00 1.29733157e+00 3.27441514e-01
1.17026508e+00 9.01323557e-02 3.00669134e-01 3.42061460e-01
4.25764471e-01 -1.03099597e+00 2.39929065e-01 3.25760275e-01
7.75790751e-01 -1.34521437e+00 1.98344901e-01 -4.28263575e-01
-5.77881515e-01 1.31393826e+00 1.03525162e+00 -4.47418541e-01
7.85971820e-01 4.59771574e-01 1.24266883e-03 -3.53875399e-01
-7.38786697e-01 -2.72631794e-01 4.05035973e-01 7.91527927e-01
5.42833567e-01 -4.73855846e-02 -4.99123126e-01 3.78667653e-01
1.65637478e-01 -6.95592389e-02 3.88257146e-01 7.19245255e-01
-6.67081177e-01 -1.20138872e+00 -6.20370269e-01 8.14936221e-01
-3.31099331e-01 2.85175472e-01 -3.66266698e-01 1.05172515e+00
7.58363485e-01 9.12142932e-01 2.45685965e-01 -5.19969404e-01
5.12066185e-01 -3.60680781e-02 7.74936914e-01 -7.98002958e-01
-7.88707316e-01 -8.36137086e-02 -6.73019961e-02 -3.99401993e-01
-1.31545737e-01 -5.89480758e-01 -1.27382028e+00 1.12390555e-01
-9.50005412e-01 -2.44035318e-01 9.18831646e-01 9.81660247e-01
7.29466900e-02 9.20829833e-01 6.48625866e-02 -7.51202822e-01
-3.29310149e-01 -1.38321793e+00 -8.29243720e-01 -1.65259600e-01
9.12273303e-02 -1.14851415e+00 -2.30376825e-01 -2.28538483e-01] | [8.540609359741211, -0.5122748613357544] |
8b6d5583-657a-4aa3-95f4-a258429f9f7b | sketch-ground-and-refine-top-down-dense-video | null | null | http://openaccess.thecvf.com//content/CVPR2021/html/Deng_Sketch_Ground_and_Refine_Top-Down_Dense_Video_Captioning_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Deng_Sketch_Ground_and_Refine_Top-Down_Dense_Video_Captioning_CVPR_2021_paper.pdf | Sketch, Ground, and Refine: Top-Down Dense Video Captioning | The dense video captioning task aims to detect and describe a sequence of events in a video for detailed and coherent storytelling. Previous works mainly adopt a "detect-then-describe" framework, which firstly detects event proposals in the video and then generates descriptions for the detected events. However, the definitions of events are diverse which could be as simple as a single action or as complex as a set of events, depending on different semantic contexts. Therefore, directly detecting events based on video information is ill-defined and hurts the coherency and accuracy of generated dense captions. In this work, we reverse the predominant "detect-then-describe" fashion, proposing a top-down way to first generate paragraphs from a global view and then ground each event description to a video segment for detailed refinement. It is formulated as a Sketch, Ground, and Refine process (SGR). The sketch stage first generates a coarse-grained multi-sentence paragraph to describe the whole video, where each sentence is treated as an event and gets localised in the grounding stage. In the refining stage, we improve captioning quality via refinement-enhanced training and dual-path cross attention on both coarse-grained event captions and aligned event segments. The updated event caption can further adjust its segment boundaries. Our SGR model outperforms state-of-the-art methods on ActivityNet Captioning benchmark under traditional and story-oriented dense caption evaluations. Code will be released at github.com/bearcatt/SGR. | ['Qi Wu', 'Yuan He', 'Da Chen', 'ShiZhe Chen', 'Chaorui Deng'] | 2021-06-19 | null | null | null | cvpr-2021-1 | ['dense-video-captioning'] | ['computer-vision'] | [ 4.25895631e-01 3.73641402e-02 -2.59360999e-01 -4.28651154e-01
-1.07990134e+00 -5.19466877e-01 8.60504329e-01 2.12749869e-01
-2.50989348e-01 7.82426953e-01 8.24314654e-01 1.89329118e-01
3.43714654e-01 -7.04670370e-01 -9.93719757e-01 -4.53410268e-01
2.89132178e-01 5.10730922e-01 5.61145663e-01 -1.60125762e-01
1.56104146e-02 -4.74874582e-03 -1.35005963e+00 8.86384785e-01
4.95798618e-01 6.91620231e-01 6.71365499e-01 7.35548556e-01
-3.73987287e-01 1.06862581e+00 -7.55939901e-01 -4.15959120e-01
-2.31287509e-01 -8.29873025e-01 -7.04914212e-01 3.35632950e-01
1.40036806e-01 -4.47576404e-01 -4.43531424e-01 8.53186667e-01
3.74981791e-01 1.26526937e-01 4.18318540e-01 -1.27597427e+00
-5.84766746e-01 9.72199380e-01 -7.33966053e-01 2.82388955e-01
7.59912372e-01 1.46072686e-01 1.11050856e+00 -8.51513565e-01
8.88943434e-01 1.08113289e+00 3.16544533e-01 6.45663798e-01
-9.24032092e-01 -7.14902461e-01 6.51617885e-01 4.07814115e-01
-1.24078643e+00 -3.78834248e-01 7.46432185e-01 -4.33524221e-01
8.93732607e-01 1.81667656e-01 9.06266630e-01 1.50116956e+00
-5.21126874e-02 1.05383837e+00 3.18747371e-01 -7.68942982e-02
2.22760797e-01 -2.96132505e-01 -3.69466357e-02 4.64928746e-01
1.21351266e-02 -3.71149510e-01 -6.66096270e-01 1.54379934e-01
9.08346713e-01 6.36536181e-02 -4.64284301e-01 4.15153578e-02
-1.71678877e+00 7.52508163e-01 2.31779635e-01 2.63949305e-01
-9.92417753e-01 1.73024073e-01 5.79328537e-01 -2.50298440e-01
4.65346992e-01 3.59195381e-01 -2.04308644e-01 -3.35451752e-01
-1.36582923e+00 4.39781278e-01 5.43427289e-01 1.15599108e+00
5.80268502e-01 -2.30769768e-01 -8.38768899e-01 6.20328307e-01
2.53006279e-01 1.78901598e-01 3.10588390e-01 -7.53243446e-01
8.36616158e-01 4.63085651e-01 2.75889665e-01 -8.76408994e-01
-9.73065719e-02 -3.18144530e-01 -6.42515361e-01 -3.72445613e-01
1.19401857e-01 -2.26387948e-01 -1.11412275e+00 1.72813940e+00
3.78771484e-01 6.61904454e-01 -5.43661974e-02 1.14685404e+00
1.26861572e+00 1.39264655e+00 4.85809624e-01 -3.64877641e-01
1.72078788e+00 -1.30705082e+00 -9.60669339e-01 -4.28497404e-01
2.16849357e-01 -6.86645210e-01 8.42435420e-01 1.62308529e-01
-1.28595972e+00 -5.56096017e-01 -9.64537024e-01 -1.64087787e-01
1.56984814e-02 5.51139237e-03 1.82150826e-01 -1.36124909e-01
-7.30024576e-01 7.06246272e-02 -8.80848169e-01 -3.47785383e-01
4.10204500e-01 -2.24949047e-01 -4.07275230e-01 -1.94290072e-01
-1.30822945e+00 5.90246558e-01 8.15763235e-01 1.93316874e-03
-1.24349856e+00 -6.67626500e-01 -1.19351542e+00 2.39527911e-01
4.91371334e-01 -8.66371274e-01 1.50037932e+00 -9.01876152e-01
-1.22808766e+00 7.06814289e-01 -5.02461791e-01 -5.29596746e-01
3.72088462e-01 -2.65390813e-01 -2.64691204e-01 4.26728606e-01
4.36012954e-01 1.08033109e+00 6.32163346e-01 -1.18472219e+00
-9.04799521e-01 2.25152314e-01 2.95300484e-01 5.38771868e-01
1.44830361e-01 3.63217652e-01 -1.14241564e+00 -8.70742917e-01
-3.59193725e-03 -6.41828954e-01 -1.74642771e-01 -4.66493815e-01
-4.73216176e-01 -1.69849023e-01 6.68633819e-01 -7.77709961e-01
1.52534521e+00 -2.04723048e+00 3.02246213e-01 -4.17205513e-01
4.81538847e-02 -3.42885442e-02 -2.37196043e-01 4.66172129e-01
-2.25929931e-01 8.04045871e-02 -2.11455673e-01 -7.37152755e-01
-8.58502910e-02 1.80922717e-01 -4.27116215e-01 1.46308973e-01
5.72801590e-01 9.21932757e-01 -1.37034774e+00 -8.19176793e-01
3.10508966e-01 6.37144327e-01 -3.82947028e-01 4.46564108e-01
-5.99740624e-01 4.71069187e-01 -5.86179733e-01 5.08473277e-01
2.73029655e-01 -5.07383049e-01 -5.89303300e-02 -4.96656865e-01
-2.46234655e-01 4.25067365e-01 -1.08251584e+00 2.01632833e+00
-3.81026894e-01 7.72122979e-01 -3.59036624e-01 -8.16729188e-01
6.49031222e-01 6.62004232e-01 4.83982414e-01 -4.51129466e-01
-3.39464284e-02 -1.48993865e-01 -5.09427428e-01 -5.98218977e-01
7.27321744e-01 -1.07350968e-01 -3.92385185e-01 3.04279447e-01
1.84531912e-01 4.94775437e-02 5.82485557e-01 4.22271729e-01
1.14034605e+00 7.30687380e-01 3.62019449e-01 4.70518082e-01
4.09098089e-01 2.39759564e-01 6.46607697e-01 5.63191175e-01
-2.12224610e-02 1.34721124e+00 6.73106611e-01 -2.60986120e-01
-1.09892905e+00 -1.03197205e+00 2.28960156e-01 1.11196113e+00
4.46144551e-01 -8.41342092e-01 -7.35783637e-01 -7.43615508e-01
-6.96195602e-01 9.30847764e-01 -5.73168099e-01 1.00807190e-01
-7.42389381e-01 -5.90909421e-01 3.13171744e-01 5.76723337e-01
6.10881925e-01 -1.58306503e+00 -6.01295769e-01 6.21287286e-01
-8.44162345e-01 -1.28227258e+00 -7.51890838e-01 -1.09239928e-01
-2.69353926e-01 -8.03590834e-01 -8.93618882e-01 -8.78155291e-01
6.15682065e-01 1.41184941e-01 1.30245543e+00 -1.89345568e-01
2.69878030e-01 2.48349761e-03 -8.39584172e-01 -2.98936754e-01
-4.23196346e-01 2.74820607e-02 -4.90729570e-01 2.63957024e-01
2.51770616e-01 -3.05701554e-01 -7.16452003e-01 2.28119835e-01
-9.17850554e-01 8.49447310e-01 6.23385489e-01 5.27370751e-01
9.12620664e-01 -1.47394493e-01 4.93247926e-01 -6.61678851e-01
3.84899914e-01 -8.91397595e-01 -1.21946141e-01 2.01079771e-01
1.01911232e-01 -1.04949646e-01 5.50760806e-01 -5.45189619e-01
-1.17716300e+00 2.69960821e-01 -2.34312922e-01 -6.99929476e-01
-3.49239677e-01 5.49169302e-01 -1.91696718e-01 8.15737486e-01
3.32258642e-01 5.44320703e-01 -5.48435926e-01 -1.72353014e-01
3.62878352e-01 3.77348930e-01 8.54794085e-01 -5.02417326e-01
6.68210447e-01 2.85132915e-01 -4.63822663e-01 -4.67183173e-01
-1.09299827e+00 -5.59838772e-01 -2.47147977e-01 -5.18239677e-01
1.39551222e+00 -1.21929765e+00 -1.37050554e-01 2.76906610e-01
-1.73494554e+00 -3.22550029e-01 -2.71690696e-01 4.82660800e-01
-6.65289283e-01 1.37074575e-01 -6.99184537e-01 -4.04563516e-01
-3.12500864e-01 -1.15066004e+00 1.52798510e+00 4.97598857e-01
-4.26571339e-01 -6.10876143e-01 1.35050341e-01 3.55626583e-01
-5.88532090e-02 5.59432328e-01 2.39809826e-01 -6.01180673e-01
-5.43431044e-01 -1.90538794e-01 -2.95420259e-01 -6.11497015e-02
3.27357203e-02 9.13331136e-02 -6.73948169e-01 6.59400821e-02
-1.25454575e-01 -1.22479871e-01 7.35990226e-01 4.40325618e-01
1.00406981e+00 -4.73000526e-01 -3.66157919e-01 5.09626150e-01
1.23511207e+00 2.98021317e-01 9.02970135e-01 4.91232574e-01
8.73478353e-01 4.04029757e-01 8.29638064e-01 5.13515174e-01
6.08194530e-01 6.73500836e-01 4.06450003e-01 -2.10511103e-01
-3.15244615e-01 -6.05465949e-01 5.34631729e-01 4.63099658e-01
-1.05015770e-01 -7.55635321e-01 -6.89225614e-01 7.62584269e-01
-2.23529243e+00 -1.51142752e+00 -9.51878801e-02 1.74388647e+00
8.58167827e-01 2.22251981e-01 8.54308903e-02 -9.90898013e-02
1.34182358e+00 5.91454685e-01 -2.79501230e-01 -7.66494274e-02
-5.99441528e-02 -6.80387393e-02 3.06176413e-02 2.77244419e-01
-1.02401233e+00 1.01199210e+00 4.71941280e+00 8.61653745e-01
-9.81810093e-01 3.07774097e-01 7.43755877e-01 -2.25764796e-01
-3.12380403e-01 -2.69506183e-02 -8.71923327e-01 8.21126401e-01
8.32628667e-01 -1.45673364e-01 -6.25478551e-02 6.60781562e-01
6.20404184e-01 -3.66946235e-02 -1.13082922e+00 1.02881169e+00
1.68242097e-01 -1.71116304e+00 4.11527127e-01 -4.52075034e-01
7.19138801e-01 -1.47612318e-02 -4.59585756e-01 4.90228087e-01
1.59554724e-02 -7.67692208e-01 1.34323716e+00 4.46186393e-01
6.64804220e-01 -5.49043894e-01 6.56511426e-01 1.79885268e-01
-1.69604397e+00 2.99222797e-01 1.12177515e-02 -2.50811651e-02
9.59420383e-01 2.88680881e-01 -7.35503852e-01 6.92294478e-01
6.29819751e-01 9.12638962e-01 -1.63787812e-01 1.12744355e+00
-5.55174053e-01 7.25409329e-01 -1.82492450e-01 5.53558320e-02
5.45161307e-01 1.61035322e-02 7.23620415e-01 1.56198430e+00
4.42863792e-01 4.26403880e-01 2.63314247e-01 8.30385447e-01
-4.67156693e-02 -5.78095838e-02 -1.62180811e-01 4.07500379e-02
4.64325011e-01 1.39994586e+00 -1.00635803e+00 -6.93509281e-01
-5.47577560e-01 1.11144745e+00 5.31473048e-02 3.90873432e-01
-1.41033256e+00 -1.76447034e-01 3.14463943e-01 3.50345492e-01
6.04975641e-01 7.84792379e-02 1.58967283e-02 -1.36279142e+00
1.96497887e-01 -6.40255511e-01 4.96395350e-01 -1.34455943e+00
-8.72744977e-01 8.27409208e-01 2.84936458e-01 -1.41624665e+00
-3.28275114e-01 1.25782639e-01 -9.86375630e-01 6.13405466e-01
-1.32188237e+00 -1.16936827e+00 -5.99493980e-01 4.45180207e-01
1.37061238e+00 3.41505259e-01 4.08498079e-01 3.99359733e-01
-7.48693168e-01 1.93171844e-01 -5.66955626e-01 1.72274530e-01
5.57000160e-01 -1.00943482e+00 5.43990970e-01 1.19211578e+00
1.99554861e-01 1.23485446e-01 9.24186349e-01 -9.74751890e-01
-8.12008262e-01 -1.51110673e+00 1.04466617e+00 -3.83442879e-01
5.67802191e-01 -4.00781006e-01 -9.67306793e-01 7.59688973e-01
4.21155185e-01 -1.50421545e-01 2.34455928e-01 -3.23340058e-01
-1.34029195e-01 1.08220935e-01 -7.37878799e-01 6.71708107e-01
1.11790705e+00 -3.35195243e-01 -8.69033694e-01 3.40731263e-01
1.08208299e+00 -6.74898267e-01 -6.57634795e-01 1.83264345e-01
2.57943779e-01 -8.25634718e-01 8.14454257e-01 -3.11859608e-01
1.02511191e+00 -6.47762120e-01 1.35835052e-01 -1.03990519e+00
-3.77883375e-01 -8.64702642e-01 -2.73607522e-01 1.55787718e+00
5.25825143e-01 9.61204767e-02 6.97804153e-01 1.40089154e-01
-5.87868333e-01 -6.72923207e-01 -5.49502671e-01 -2.88663536e-01
-6.44914389e-01 -5.37866235e-01 6.50430918e-01 8.58507752e-01
6.74816817e-02 6.31868839e-01 -4.20404106e-01 3.28687608e-01
3.31853032e-01 1.09970652e-01 5.26152968e-01 -6.66626513e-01
-3.56811196e-01 -3.49008977e-01 -2.28556976e-01 -1.19139159e+00
1.17179453e-02 -6.02339625e-01 4.27308202e-01 -2.11263299e+00
5.39641857e-01 1.41780540e-01 -1.37644827e-01 4.98000860e-01
-5.15203953e-01 3.35292101e-01 2.91866779e-01 3.06804359e-01
-1.14883661e+00 4.80309486e-01 1.24881148e+00 -1.69200420e-01
-2.98229665e-01 -2.03856364e-01 -4.40872580e-01 5.74480832e-01
5.67400396e-01 -4.93768007e-01 -4.58614886e-01 -4.27193373e-01
3.06316704e-01 5.70999444e-01 4.13203657e-01 -1.22361374e+00
2.34567553e-01 -2.32724369e-01 3.72962773e-01 -8.54523838e-01
4.34422314e-01 -3.97386998e-01 4.32307363e-01 1.49822041e-01
-4.05140549e-01 1.38946906e-01 1.76341951e-01 5.80479324e-01
-4.64873344e-01 -6.37871996e-02 4.87275511e-01 -2.93920666e-01
-9.44924295e-01 5.24168849e-01 -3.43961269e-01 8.20393711e-02
1.30522168e+00 -3.73359114e-01 -3.35705638e-01 -5.55925369e-01
-7.18780339e-01 4.36431974e-01 3.65059078e-01 5.76869130e-01
6.14373207e-01 -1.31940997e+00 -1.11083424e+00 -2.83135861e-01
2.00518072e-01 5.86816251e-01 4.76092756e-01 5.29101908e-01
-5.94543755e-01 1.54662028e-01 -1.32528365e-01 -6.73912942e-01
-1.07172763e+00 6.16515517e-01 4.12419438e-02 -3.20237786e-01
-8.07864010e-01 8.31961036e-01 6.95708811e-01 4.63455975e-01
1.75048724e-01 -2.96103179e-01 -5.03091991e-01 2.21634150e-01
7.96217740e-01 -8.67935494e-02 -3.65435630e-01 -8.54712188e-01
-2.13122591e-01 4.65119272e-01 -1.68526486e-01 -2.11983263e-01
1.27554858e+00 -2.91190416e-01 1.51262209e-01 3.28171879e-01
9.03190911e-01 -3.37894529e-01 -1.60048711e+00 7.95302019e-02
-3.63019645e-01 -4.62635905e-02 -1.26057789e-01 -6.38393760e-01
-8.53019536e-01 5.27694225e-01 -6.63789958e-02 2.75441617e-01
1.21596551e+00 4.30514067e-01 1.10665917e+00 -1.42434463e-01
7.46304244e-02 -7.25198865e-01 2.62096971e-01 3.85936767e-01
1.10019958e+00 -1.00536239e+00 -2.05961853e-01 -4.93565857e-01
-9.99595046e-01 9.42978680e-01 8.07060778e-01 -5.67570375e-03
-2.77624186e-02 1.55591622e-01 -3.39454651e-01 -1.71625808e-01
-9.12559032e-01 -2.60441959e-01 3.43149513e-01 3.01900923e-01
3.85432631e-01 -1.57201737e-01 -3.30469072e-01 8.37911010e-01
8.51392187e-03 1.22636363e-01 6.35003209e-01 6.99769855e-01
-3.89739305e-01 -6.93432868e-01 -3.76790971e-01 1.80888474e-01
-4.39732283e-01 -1.39499947e-01 -1.23380072e-01 6.91655993e-01
3.57245833e-01 8.84498894e-01 3.88788342e-01 -2.51088768e-01
4.11399156e-01 -1.52999554e-02 1.87243178e-01 -9.18138623e-01
-5.46577215e-01 2.47210562e-01 3.42490137e-01 -6.14907622e-01
-5.74479282e-01 -8.20787668e-01 -1.48963189e+00 -1.34580135e-02
-5.19114099e-02 3.06712836e-01 3.87071669e-01 1.06116855e+00
3.75857979e-01 9.10440743e-01 3.26632679e-01 -1.11272073e+00
2.69357711e-01 -1.00304306e+00 4.08526659e-02 6.04948819e-01
2.11751744e-01 -4.61696059e-01 -2.05240905e-01 6.37116849e-01] | [10.424288749694824, 0.6832104921340942] |
6af22d0d-1e31-4a3b-9107-e41951ab6970 | stock-price-prediction-using-sentiment | 2204.05783 | null | https://arxiv.org/abs/2204.05783v1 | https://arxiv.org/pdf/2204.05783v1.pdf | Stock Price Prediction using Sentiment Analysis and Deep Learning for Indian Markets | Stock market prediction has been an active area of research for a considerable period. Arrival of computing, followed by Machine Learning has upgraded the speed of research as well as opened new avenues. As part of this research study, we aimed to predict the future stock movement of shares using the historical prices aided with availability of sentiment data. Two models were used as part of the exercise, LSTM was the first model with historical prices as the independent variable. Sentiment Analysis captured using Intensity Analyzer was used as the major parameter for Random Forest Model used for the second part, some macro parameters like Gold, Oil prices, USD exchange rate and Indian Govt. Securities yields were also added to the model for improved accuracy of the model. As the end product, prices of 4 stocks viz. Reliance, HDFC Bank, TCS and SBI were predicted using the aforementioned two models. The results were evaluated using RMSE metric. | ['Yogesh Agarwal', 'Usha Aiyer', 'Pranali Bhagat', 'Nutan Hindlekar', 'Milind Manjrekar', 'Himank Sharma', 'Anwesh Reddy Paduri', 'Narayana Darapaneni'] | 2022-04-07 | null | null | null | null | ['stock-market-prediction', 'stock-price-prediction'] | ['time-series', 'time-series'] | [-5.61038017e-01 -1.52765006e-01 -1.79000348e-01 -1.66953534e-01
-8.49337876e-03 -7.38403916e-01 6.96759403e-01 -1.77093148e-02
-2.79683053e-01 1.26385927e+00 2.78422207e-01 -5.89020610e-01
4.53710817e-02 -1.19169593e+00 -2.61034191e-01 -5.94963014e-01
9.14025307e-02 2.04771906e-01 1.78228870e-01 -3.90739232e-01
9.00368214e-01 5.17532229e-01 -1.02363408e+00 -1.02287896e-01
4.86865759e-01 1.48789859e+00 1.77924469e-01 2.59657025e-01
-4.76375967e-01 1.30178201e+00 -1.93001255e-01 -5.73058844e-01
7.97187209e-01 -6.42278418e-02 -3.86206746e-01 -3.10867548e-01
-7.31385469e-01 -4.53176707e-01 1.13917917e-01 1.06760800e+00
1.69068038e-01 -3.10906768e-02 4.34866935e-01 -1.08629096e+00
-5.23483694e-01 9.04587209e-01 -7.17036605e-01 5.20217240e-01
-1.81091473e-01 -8.20894167e-02 1.05099034e+00 -9.08874691e-01
3.28703701e-01 7.41077363e-01 4.28000540e-01 -4.10844117e-01
-6.73759341e-01 -1.08923411e+00 -2.36494094e-01 2.09577262e-01
-8.64072442e-01 -5.84040827e-04 9.40654755e-01 -3.94485891e-01
1.40178013e+00 -7.48324245e-02 1.08728445e+00 2.07047880e-01
7.93559611e-01 1.71484530e-01 2.03325939e+00 -5.46424091e-01
1.31122842e-01 8.90965760e-01 1.36318579e-01 1.23703390e-01
4.87632871e-01 3.03026706e-01 -3.16685200e-01 7.61636049e-02
6.07894421e-01 2.07547709e-01 3.66083235e-01 5.07563412e-01
-8.26253414e-01 1.15241611e+00 2.12201476e-01 6.26637459e-01
-8.18238735e-01 -2.42311612e-01 2.20870361e-01 5.43846071e-01
6.92333162e-01 2.43796617e-01 -9.70838964e-01 -3.23896408e-01
-1.18858743e+00 4.20171648e-01 8.62896919e-01 4.78864074e-01
5.00422597e-01 6.68115735e-01 4.79381740e-01 1.92698166e-01
4.73180413e-01 7.33918905e-01 9.62957144e-01 -3.53019089e-01
3.67252529e-01 8.32587242e-01 3.23898107e-01 -1.11181939e+00
-4.74953741e-01 -8.04860413e-01 -4.44922328e-01 2.82464355e-01
1.93402648e-01 -6.32749379e-01 -1.10303020e+00 1.08567107e+00
1.59941632e-02 1.12707712e-01 2.73340911e-01 3.42594415e-01
-1.12411447e-01 8.99545789e-01 5.82290441e-02 -5.50736785e-01
1.16276777e+00 -7.81718433e-01 -7.12610900e-01 2.48708919e-01
3.25157553e-01 -1.00645769e+00 7.41637722e-02 4.53523368e-01
-1.16992307e+00 -3.49820465e-01 -9.38331068e-01 7.66220033e-01
-8.07749748e-01 -7.38200024e-02 8.01746547e-01 5.87869287e-01
-9.56975400e-01 8.01775992e-01 -8.18732440e-01 -3.65979150e-02
2.23506108e-01 5.09624064e-01 5.93451001e-02 7.81864882e-01
-1.41135514e+00 1.50032723e+00 6.00728035e-01 7.31307417e-02
-1.40091568e-01 -4.95105088e-01 -2.69840628e-01 6.39053881e-02
-2.41059139e-01 -3.08131184e-02 1.07616317e+00 -1.14633822e+00
-1.45983553e+00 5.37253618e-01 2.72890925e-01 -9.40708280e-01
5.25224864e-01 -1.65726040e-02 -8.52875233e-01 -3.16722304e-01
1.04089558e-01 1.03682429e-01 2.42975116e-01 -5.27915597e-01
-1.17673361e+00 -5.99048972e-01 -1.86354220e-01 2.54687369e-01
1.90103240e-02 2.28153139e-01 5.89069486e-01 -9.85179603e-01
2.11150512e-01 -8.65396738e-01 -1.94779947e-01 -1.04268610e+00
-1.51161730e-01 2.42533654e-01 8.42746019e-01 -1.05698836e+00
1.20532882e+00 -1.63829362e+00 -5.57123423e-01 7.79607058e-01
-4.73794490e-01 2.28851810e-02 8.91884744e-01 7.20299125e-01
-5.15993476e-01 3.36700767e-01 3.02309729e-02 2.58098215e-01
-1.13170996e-01 -8.79037157e-02 -7.70109355e-01 1.23783685e-01
2.79260367e-01 5.52589297e-01 -1.90225199e-01 -9.40543115e-02
2.46076837e-01 -8.53167251e-02 1.19499907e-01 -1.83291491e-02
5.45843877e-02 5.44606410e-02 -4.80399758e-01 9.03218448e-01
7.75483608e-01 5.48776425e-02 -2.06865281e-01 -2.31660865e-02
-9.11839902e-01 2.05505684e-01 -1.38141418e+00 7.24605858e-01
-1.76535726e-01 3.07218075e-01 -3.56595933e-01 -1.01013577e+00
1.17582703e+00 6.15012765e-01 4.64941293e-01 -7.17140377e-01
1.71592072e-01 7.28942871e-01 3.08222383e-01 -2.16039062e-01
5.72506249e-01 -7.10199237e-01 3.52800965e-01 6.05103910e-01
-3.61564636e-01 1.80353135e-01 1.77998200e-01 -3.50648046e-01
3.20372403e-01 2.16665566e-01 7.13231266e-01 -3.23111683e-01
3.89599085e-01 4.54203576e-01 4.53705668e-01 -2.05829144e-02
-1.98846579e-01 -1.55662373e-01 3.32796514e-01 -3.29773635e-01
-1.05904555e+00 -6.13754332e-01 -2.82549411e-01 5.07482111e-01
-3.58164191e-01 7.10749209e-01 -2.13367790e-01 -7.28709102e-02
1.69771239e-01 8.77599359e-01 -4.44643855e-01 4.69383836e-01
-1.10545546e-01 -1.08123779e+00 1.59622386e-01 3.47562313e-01
8.31286669e-01 -1.16007805e+00 -8.71256948e-01 4.87166643e-01
7.06386983e-01 -6.58395648e-01 3.82239997e-01 6.99667275e-01
-1.36163592e+00 -6.92616761e-01 -8.76918793e-01 -3.77742410e-01
2.19616473e-01 -1.14634678e-01 8.68372023e-01 -2.83099711e-01
1.81161448e-01 -5.05679309e-01 -2.25266293e-01 -1.35293949e+00
-3.26594003e-02 9.92364809e-02 3.97721715e-02 -6.03256859e-02
3.98262322e-01 -7.52288163e-01 -6.93611205e-01 -2.33351186e-01
-6.36704385e-01 4.89086471e-03 6.43973351e-01 3.66316050e-01
1.99792877e-01 6.71244204e-01 1.11756837e+00 -9.94883895e-01
4.79965717e-01 -8.50494325e-01 -1.03795648e+00 -1.08792342e-01
-1.28765452e+00 5.09047136e-02 3.46211106e-01 1.06002517e-01
-1.47207725e+00 -2.19317809e-01 6.45958930e-02 1.09993584e-01
2.40374267e-01 1.22093177e+00 2.85847187e-01 2.48301223e-01
-1.48631841e-01 2.41239995e-01 -4.50323336e-03 -4.19514835e-01
-2.45584920e-01 6.65402710e-01 1.50398329e-01 1.58410102e-01
9.20620918e-01 2.18981296e-01 -8.92541429e-04 -3.79153013e-01
-6.15180373e-01 -3.37636024e-01 -6.02123499e-01 -2.74275094e-01
4.68865097e-01 -1.09165144e+00 -2.64301121e-01 8.75885189e-01
-7.80983031e-01 1.32930443e-01 6.02336042e-02 8.26618969e-01
7.51029700e-02 -5.06170928e-01 -4.67488766e-01 -1.67821169e+00
-8.17299783e-01 -8.38158965e-01 -7.90119395e-02 6.11536741e-01
-1.68329582e-01 -1.24263763e+00 1.84595451e-01 2.27505684e-01
8.56041729e-01 5.83652079e-01 9.15326059e-01 -1.11695158e+00
-4.70991701e-01 -4.71241981e-01 -2.06439361e-01 5.43447673e-01
2.97324687e-01 2.69032419e-01 -7.67838120e-01 1.13615930e-01
5.13827741e-01 -1.65186435e-01 2.90008932e-01 5.40220439e-01
2.56742209e-01 -3.37523788e-01 8.90489295e-02 2.75276154e-01
1.94108737e+00 1.15247285e+00 4.82201755e-01 9.62041557e-01
9.82944891e-02 3.65277141e-01 7.12015033e-01 6.52172029e-01
2.39456445e-01 -1.89521357e-01 1.56328157e-01 1.59551963e-01
6.57951295e-01 -1.10456079e-01 4.08374965e-01 1.19998407e+00
-4.19186354e-01 2.49036700e-01 -8.14976573e-01 3.22397560e-01
-1.33156836e+00 -1.09445357e+00 -3.08498621e-01 1.95907080e+00
5.32894731e-01 7.19665945e-01 2.67900050e-01 6.53915346e-01
2.68245995e-01 1.84662551e-01 -5.74455202e-01 -3.96733254e-01
-1.39177829e-01 5.79528749e-01 1.26967812e+00 1.34539813e-01
-6.68564916e-01 7.71445632e-01 5.79952717e+00 2.05544814e-01
-1.64889812e+00 -1.48872688e-01 8.86879504e-01 1.23751447e-01
-3.03586394e-01 2.90420622e-01 -8.63816142e-01 7.86160827e-01
1.16241336e+00 -4.00164425e-01 2.31457084e-01 8.17578673e-01
7.01622307e-01 -5.36787510e-01 -5.61597422e-02 2.50509739e-01
-4.67106581e-01 -1.46222091e+00 -1.90446660e-01 3.29985648e-01
8.64915013e-01 2.81199932e-01 9.79738086e-02 1.65452078e-01
3.64043355e-01 -7.10065007e-01 9.73009288e-01 8.07453096e-01
2.02246487e-01 -1.07527816e+00 1.30343509e+00 3.16367656e-01
-1.09955907e+00 -2.90191352e-01 -2.15264708e-01 -7.39010692e-01
1.63574383e-01 4.25288349e-01 -5.16456544e-01 4.92913306e-01
6.25033379e-01 4.65710670e-01 -1.67615071e-01 6.00890756e-01
3.18170547e-01 9.15633559e-01 -4.64174837e-01 -2.66227096e-01
6.14310741e-01 -7.26419568e-01 3.50536481e-02 7.51592159e-01
6.34215832e-01 3.09828997e-01 -3.71584207e-01 5.88263035e-01
1.02192730e-01 5.23140788e-01 -5.36133051e-01 -2.21994311e-01
6.17407620e-01 9.97408330e-01 -9.08176422e-01 -3.41356695e-01
-5.51490009e-01 2.57166803e-01 -1.53888345e-01 1.82182252e-01
-6.93956554e-01 -2.99996108e-01 1.39779374e-01 2.75636822e-01
9.39055979e-02 -1.14380196e-01 -9.16150391e-01 -8.51583123e-01
-2.22952589e-01 -6.96101308e-01 1.09850712e-01 -6.27905607e-01
-8.69896412e-01 2.71838188e-01 -5.38285896e-02 -1.01140451e+00
-2.12899178e-01 -6.53135717e-01 -9.22270060e-01 1.54526675e+00
-1.76291394e+00 -8.99153888e-01 1.72356650e-01 3.65202904e-01
4.90849733e-01 -8.34327519e-01 5.66977501e-01 -1.49706975e-02
-6.26486957e-01 -2.23753631e-01 5.92021585e-01 1.74261734e-01
-1.52287092e-02 -1.32332718e+00 1.11924447e-01 8.51092398e-01
-7.68934116e-02 4.82307822e-01 7.48929739e-01 -1.12008429e+00
-9.78846967e-01 -7.55531847e-01 8.09236944e-01 1.08368680e-01
1.36464632e+00 4.16849524e-01 -5.12249827e-01 8.67681146e-01
5.16322911e-01 -6.05089128e-01 7.14356244e-01 -5.39896965e-01
3.17825556e-01 -2.21766055e-01 -1.27431059e+00 1.44023374e-01
-2.51748055e-01 -7.61613101e-02 -6.20755970e-01 -2.07476467e-01
2.37090856e-01 -1.88276321e-01 -1.19226730e+00 5.32044172e-02
8.36464465e-01 -1.02321064e+00 4.24591303e-01 -3.70512754e-01
2.76560873e-01 -4.95763458e-02 -3.21274310e-01 -1.26735294e+00
-2.45387554e-01 -3.36652040e-01 1.50992975e-01 1.47174537e+00
9.94276762e-01 -9.75945055e-01 9.50714469e-01 9.84414160e-01
3.78970206e-01 -8.00459385e-01 -6.26090109e-01 -3.31983000e-01
1.83234319e-01 -2.55593300e-01 7.39214063e-01 8.80499184e-01
7.72284670e-03 1.61880344e-01 -2.70756662e-01 -1.05636470e-01
4.84681606e-01 1.31289139e-01 3.49923968e-01 -1.15324414e+00
-1.33131757e-01 -5.14599383e-01 -1.64155394e-01 -2.85924494e-01
-2.93573022e-01 -5.50095677e-01 -9.59300816e-01 -1.33686948e+00
-7.15039223e-02 -3.89703423e-01 -6.22277617e-01 1.76860616e-01
1.13281332e-01 -2.47204732e-02 3.94119859e-01 5.45516908e-01
9.66595590e-01 1.67857692e-01 8.92487228e-01 1.91681027e-01
-3.59006703e-01 5.61761022e-01 -7.33292699e-01 9.05531168e-01
1.19622731e+00 -2.15707526e-01 -4.59504515e-01 1.57891691e-01
6.80340230e-01 3.21873397e-01 -2.20798820e-01 -7.47749805e-01
4.79693897e-02 -5.34851313e-01 8.05561662e-01 -1.07471383e+00
1.25624269e-01 -1.12503684e+00 6.79454923e-01 7.14408934e-01
-1.22734927e-01 9.21724677e-01 8.28638971e-02 2.29143962e-01
-5.50066650e-01 -3.88757795e-01 6.61211789e-01 -4.67440069e-01
-5.53680658e-01 1.52062392e-03 -3.39970648e-01 -4.61558640e-01
1.31504667e+00 -6.40719950e-01 3.01761739e-02 -2.96014309e-01
-5.27100086e-01 2.38879934e-01 1.17756911e-01 -2.63669062e-03
2.03143388e-01 -9.36087370e-01 -5.30069232e-01 -1.11269854e-01
-6.11282766e-01 -2.86133736e-01 -1.94737256e-01 8.57738554e-01
-1.00679636e+00 7.40928113e-01 -5.52036166e-01 2.72046566e-01
-7.72061884e-01 3.04590076e-01 3.99757683e-01 -6.69800282e-01
-2.23323926e-01 4.65557158e-01 -5.75636625e-01 2.65768915e-01
-9.95227024e-02 -6.63536549e-01 -6.90241814e-01 6.32206917e-01
2.50325054e-01 5.24345458e-01 1.24656960e-01 -6.94055438e-01
-8.43422115e-02 3.45960826e-01 -7.63064623e-02 -6.09243274e-01
1.87404060e+00 -2.63249129e-01 -3.48688990e-01 1.14522350e+00
8.11979353e-01 1.19017959e-02 -9.66739178e-01 3.83291729e-02
8.00108433e-01 -2.97858536e-01 3.56240451e-01 -1.21303749e+00
-1.34838295e+00 4.11367029e-01 5.60311079e-01 4.02387112e-01
1.12885463e+00 -9.69944596e-01 8.72359455e-01 -4.24777456e-02
3.20667833e-01 -1.41596389e+00 -6.91254973e-01 3.80653799e-01
4.91050541e-01 -1.29112029e+00 1.89349562e-01 2.09531978e-01
-7.68817484e-01 1.17313111e+00 9.88735631e-02 -2.76722610e-01
1.33827603e+00 5.73012412e-01 1.75712183e-01 -8.39716420e-02
-7.26588011e-01 1.30214229e-01 -1.85290501e-01 5.97775355e-02
8.12983096e-01 2.92340130e-01 -4.77594346e-01 7.36055076e-01
-7.11583436e-01 5.34223974e-01 8.73387992e-01 1.06651533e+00
-6.58289254e-01 -6.83560908e-01 -5.04879892e-01 1.06215584e+00
-1.42226660e+00 -3.17877471e-01 -2.48315278e-02 1.08249640e+00
1.06762797e-02 6.63389444e-01 2.16710508e-01 -3.55118006e-01
2.23201185e-01 2.14374527e-01 -2.79603183e-01 -2.76045501e-01
-9.29843068e-01 7.78761739e-03 2.20811218e-02 2.21597642e-01
-4.86049145e-01 -8.75474215e-01 -1.30203092e+00 -5.20470798e-01
-6.06041491e-01 4.90462303e-01 1.09156179e+00 1.18169379e+00
-2.75012672e-01 1.18630104e-01 1.16883457e+00 -4.61620480e-01
-9.66541290e-01 -1.39666414e+00 -1.10682750e+00 -2.42792830e-01
6.73707062e-03 -4.56785470e-01 -3.99062991e-01 9.80748907e-02] | [4.499501705169678, 4.224732398986816] |
797987c9-5c79-4caa-bbb5-0557924c8921 | corelm-coreference-aware-language-model-fine-1 | 2111.02687 | null | https://arxiv.org/abs/2111.02687v1 | https://arxiv.org/pdf/2111.02687v1.pdf | CoreLM: Coreference-aware Language Model Fine-Tuning | Language Models are the underpin of all modern Natural Language Processing (NLP) tasks. The introduction of the Transformers architecture has contributed significantly into making Language Modeling very effective across many NLP task, leading to significant advancements in the field. However, Transformers come with a big computational cost, which grows quadratically with respect to the input length. This presents a challenge as to understand long texts requires a lot of context. In this paper, we propose a Fine-Tuning framework, named CoreLM, that extends the architecture of current Pretrained Language Models so that they incorporate explicit entity information. By introducing entity representations, we make available information outside the contextual space of the model, which results in a better Language Model for a fraction of the computational cost. We implement our approach using GPT2 and compare the fine-tuned model to the original. Our proposed model achieves a lower Perplexity in GUMBY and LAMBDADA datasets when compared to GPT2 and a fine-tuned version of GPT2 without any changes. We also compare the models' performance in terms of Accuracy in LAMBADA and Children's Book Test, with and without the use of model-created coreference annotations. | ['Ioannis Vlahavas', 'Nikolaos Stylianou'] | 2021-11-04 | corelm-coreference-aware-language-model-fine | https://aclanthology.org/2021.crac-1.8 | https://aclanthology.org/2021.crac-1.8.pdf | crac-acl-2021-11 | ['lambada'] | ['natural-language-processing'] | [-1.95180446e-01 3.93049687e-01 -1.14575580e-01 -5.55454552e-01
-7.50348449e-01 -5.92691541e-01 6.24642551e-01 4.12801117e-01
-8.25614929e-01 5.78550220e-01 5.21751523e-01 -2.88204074e-01
-7.08640590e-02 -8.53330076e-01 -6.21453702e-01 -4.13207650e-01
1.39450222e-01 8.88213873e-01 3.53766441e-01 -3.86283875e-01
-5.76069281e-02 1.35513812e-01 -1.02665961e+00 3.62544686e-01
9.85566318e-01 5.37822783e-01 3.57769579e-01 4.26335305e-01
-7.20046639e-01 7.90207922e-01 -2.97984272e-01 -8.39032352e-01
-1.61692668e-02 4.33447771e-02 -9.85617518e-01 -5.99108756e-01
1.61642611e-01 4.61236089e-02 -2.97753751e-01 6.57611310e-01
6.04689062e-01 2.55729973e-01 4.56340134e-01 -8.90807450e-01
-6.08392358e-01 1.24615848e+00 -4.82383102e-01 9.51270908e-02
1.90033913e-01 -3.86930645e-01 1.15587115e+00 -8.28844070e-01
4.86797869e-01 1.58939326e+00 7.88830340e-01 5.93082726e-01
-1.23716545e+00 -5.38267791e-01 3.28434825e-01 2.46710435e-01
-1.34618533e+00 -3.26400638e-01 4.83045161e-01 -1.35409385e-01
1.65730906e+00 1.20549118e-02 2.46400923e-01 1.10355937e+00
4.20707539e-02 9.99374747e-01 8.24859798e-01 -7.68947601e-01
-1.15441389e-01 2.48775706e-01 4.74768817e-01 4.96350378e-01
8.47220868e-02 -1.89130157e-01 -3.57542306e-01 -4.06692140e-02
3.96394610e-01 -2.46294945e-01 -9.28894207e-02 -1.87941417e-01
-9.79832351e-01 9.81263816e-01 3.06468964e-01 6.65921986e-01
-2.80644357e-01 9.81853679e-02 3.96861851e-01 2.19009206e-01
4.58850026e-01 4.35030431e-01 -7.81841815e-01 -3.08724284e-01
-8.86256754e-01 1.98770434e-01 1.06955600e+00 9.94901776e-01
4.77168083e-01 -3.18557799e-01 -1.19837083e-01 1.14394891e+00
3.69396031e-01 2.97624916e-01 7.65714586e-01 -5.45416415e-01
7.76881933e-01 8.69977951e-01 -1.61968231e-01 -9.13872123e-01
-6.95468605e-01 -5.68835139e-01 -6.47853673e-01 -5.42721450e-01
2.95546263e-01 -1.61889449e-01 -7.50443399e-01 1.99164331e+00
2.74954140e-01 7.76606426e-02 3.19129616e-01 3.82355630e-01
9.23976302e-01 8.88374984e-01 3.82206738e-01 -1.62551880e-01
1.46745086e+00 -1.28636062e+00 -7.65489578e-01 -4.45377856e-01
9.19705808e-01 -8.63426566e-01 1.01396215e+00 3.10832292e-01
-9.84234273e-01 -3.66601944e-01 -7.83279300e-01 -5.14482677e-01
-5.51031888e-01 -4.04789485e-02 7.32629120e-01 7.12093711e-01
-1.04803264e+00 4.27029729e-01 -7.72326171e-01 -5.73752344e-01
1.85549371e-02 4.37108994e-01 -4.52978492e-01 3.89479622e-02
-1.43661284e+00 1.16865301e+00 7.76811421e-01 -7.09439442e-02
-3.36305112e-01 -7.25741327e-01 -9.22541976e-01 2.88753450e-01
3.69180560e-01 -7.93902814e-01 1.38691115e+00 -5.52110732e-01
-1.51312792e+00 7.47500122e-01 -3.91009152e-01 -6.05259240e-01
4.36270595e-01 -5.75675070e-01 -3.37036967e-01 -1.88265458e-01
-4.50561419e-02 7.13290393e-01 2.58187503e-01 -7.30807602e-01
-5.73844671e-01 -1.68984726e-01 1.58701912e-01 2.37826854e-01
-3.48017871e-01 2.32884347e-01 -7.71990478e-01 -7.06941724e-01
-6.12591021e-02 -9.12046432e-01 -3.19835395e-01 -5.25097370e-01
-2.14084342e-01 -6.78719580e-01 4.07262981e-01 -5.48300803e-01
1.58494997e+00 -1.99155891e+00 1.57180682e-01 -1.38995070e-02
4.10660021e-02 5.33440292e-01 -2.90917665e-01 7.99656928e-01
-1.11043509e-02 3.14416438e-01 -2.09187508e-01 -6.74929023e-01
1.68548897e-01 6.27671123e-01 -3.65366012e-01 -8.74188468e-02
7.78930560e-02 9.34981227e-01 -7.66528547e-01 -4.94509310e-01
3.66824158e-02 5.92027664e-01 -7.79498279e-01 2.03755632e-01
-4.00293201e-01 1.80560216e-01 -4.23103392e-01 4.49235514e-02
5.98786354e-01 -1.70672983e-01 2.49739841e-01 1.32795095e-01
-3.07808369e-01 8.06683660e-01 -1.12073791e+00 1.83149087e+00
-6.74545348e-01 3.07819784e-01 -3.06322098e-01 -9.28128660e-01
8.69307935e-01 5.37636459e-01 1.88507974e-01 -6.74032450e-01
1.70024768e-01 2.46851474e-01 4.18136492e-02 -5.31414807e-01
5.30644417e-01 -2.24661335e-01 -1.77785292e-01 5.49809515e-01
2.34898463e-01 1.10235468e-01 3.40788692e-01 2.29434997e-01
8.91216934e-01 7.42659420e-02 5.18031061e-01 -1.93991706e-01
6.51853502e-01 -1.93423718e-01 5.44771492e-01 6.96966410e-01
1.35906890e-01 2.52065837e-01 5.13185978e-01 -4.80713457e-01
-7.80257523e-01 -6.96822464e-01 -1.57599941e-01 1.36609924e+00
-2.65549839e-01 -7.05096662e-01 -7.89953351e-01 -7.06677258e-01
-1.58367380e-01 1.00301313e+00 -4.72958952e-01 1.60876475e-02
-9.23108697e-01 -9.19828475e-01 8.39663446e-01 5.42283058e-01
4.01883304e-01 -1.12853515e+00 -3.24962705e-01 4.72307414e-01
-3.49398077e-01 -1.38550830e+00 -3.94437134e-01 2.27208748e-01
-8.55867743e-01 -8.36586058e-01 -5.22728562e-01 -8.30426574e-01
2.25268215e-01 -2.03795910e-01 1.21116769e+00 -1.22332186e-01
5.48528843e-02 1.95137456e-01 -5.00221014e-01 -5.50379038e-01
-4.28213745e-01 4.89386678e-01 2.43607163e-03 -1.74731418e-01
8.82304132e-01 -6.86116159e-01 -2.82058299e-01 -2.85178740e-02
-8.54364395e-01 1.59899324e-01 6.35287523e-01 7.37016797e-01
3.57493490e-01 -1.38425246e-01 5.92447579e-01 -1.14830148e+00
7.35711038e-01 -5.30429661e-01 -4.60226148e-01 4.40304875e-01
-6.70743167e-01 5.82676053e-01 6.15620673e-01 -3.58521730e-01
-1.26921892e+00 -2.58918852e-01 -4.89053220e-01 -4.10256647e-02
6.49856124e-03 6.17618263e-01 -9.75224152e-02 1.78283051e-01
3.96687478e-01 1.36823490e-01 -5.78413785e-01 -9.85413730e-01
5.73914647e-01 6.64339721e-01 3.11523050e-01 -8.36249769e-01
4.05065030e-01 1.59281231e-02 -3.03787917e-01 -7.41527557e-01
-1.00108278e+00 -5.44540405e-01 -7.85392642e-01 4.98204499e-01
7.05903292e-01 -7.67390549e-01 -6.48227036e-01 3.36184919e-01
-1.38886845e+00 -1.13246523e-01 -9.36837494e-02 6.62120759e-01
-2.46141374e-01 4.64619726e-01 -6.57116115e-01 -6.63333535e-01
-6.85506105e-01 -9.42920923e-01 9.26637650e-01 2.09364504e-01
-1.76250473e-01 -1.31320691e+00 4.60676372e-01 3.71730506e-01
5.58659434e-01 -5.02981655e-02 1.28749263e+00 -1.12792599e+00
-1.69606298e-01 -7.61292726e-02 -2.30960116e-01 3.82258594e-01
-2.36195058e-01 -2.57381618e-01 -1.06928051e+00 -1.22199424e-01
-8.36638659e-02 -1.61049619e-01 8.37156475e-01 9.73095223e-02
7.36314416e-01 -3.98874611e-01 -3.90785694e-01 6.44129157e-01
1.38998520e+00 3.75663601e-02 3.75902236e-01 3.44200402e-01
6.00231946e-01 6.27670825e-01 3.95284712e-01 2.19741404e-01
7.90500998e-01 6.98514938e-01 2.50539444e-02 1.56267583e-01
-3.96205336e-02 -6.97331607e-01 2.67111838e-01 1.26904368e+00
-4.17634733e-02 -3.71733725e-01 -1.29838574e+00 7.14614511e-01
-1.93522871e+00 -7.06792176e-01 -1.29041761e-01 1.90425539e+00
1.05728388e+00 5.02834171e-02 -1.80198506e-01 -1.41219318e-01
4.74671453e-01 4.42372561e-02 -1.67642519e-01 -8.01019669e-01
-1.12306088e-01 4.77812290e-01 2.78137386e-01 7.93847382e-01
-9.37989771e-01 1.35363936e+00 6.38593769e+00 8.64532351e-01
-1.10489154e+00 1.91550985e-01 1.24723345e-01 2.42179283e-03
-3.10331970e-01 1.02740407e-01 -1.11913550e+00 2.87995428e-01
1.31082404e+00 -4.02874112e-01 1.81321427e-01 7.55699933e-01
1.60779029e-01 9.56586972e-02 -1.36346233e+00 8.81102443e-01
1.25558168e-01 -9.33946371e-01 1.96587473e-01 -1.82925761e-01
4.28899944e-01 3.87724787e-01 -2.50117779e-01 8.43621075e-01
3.17239255e-01 -9.91613388e-01 4.56583589e-01 2.51357138e-01
3.38329464e-01 -6.88474476e-01 8.87008011e-01 8.04723203e-01
-1.07907379e+00 -2.55028214e-02 -5.47845304e-01 -1.02551527e-01
3.18073571e-01 5.17039955e-01 -1.12206733e+00 5.20191252e-01
5.54329038e-01 4.08016205e-01 -4.47847545e-01 9.20825005e-01
-5.20857394e-01 7.25158811e-01 -5.86342216e-01 -1.23960003e-01
4.58261997e-01 -6.24948591e-02 4.97518539e-01 1.61158192e+00
2.23599881e-01 1.30559623e-01 1.30217791e-01 6.82525218e-01
-3.72976959e-01 6.61317825e-01 -3.38223785e-01 -9.09217894e-02
6.32480085e-01 1.10471976e+00 -1.89398602e-01 -3.43979955e-01
-5.56533039e-01 7.50733912e-01 7.06039369e-01 1.09988324e-01
-5.88041008e-01 -3.85078490e-01 3.35631758e-01 -3.13633569e-02
2.20396712e-01 -1.51657119e-01 -3.42167206e-02 -1.22703576e+00
-5.26698902e-02 -8.12851727e-01 5.98206937e-01 -4.04788524e-01
-1.30771101e+00 7.99919069e-01 9.16066021e-02 -6.40639722e-01
-4.25177842e-01 -6.11573517e-01 -4.46811110e-01 9.84108090e-01
-1.77677107e+00 -1.35549879e+00 9.32649076e-02 4.89296734e-01
6.25602424e-01 -1.36272954e-02 1.15934455e+00 5.79545677e-01
-5.49408615e-01 8.23776424e-01 3.65510397e-02 3.32855999e-01
8.98391664e-01 -1.23987508e+00 5.60136795e-01 6.79665983e-01
2.55229026e-01 9.45457339e-01 7.03492343e-01 -4.17882591e-01
-1.03032899e+00 -1.01539266e+00 1.72943270e+00 -6.08856857e-01
7.73092270e-01 -4.19546902e-01 -1.06359327e+00 1.03078282e+00
3.18437219e-01 -1.65034503e-01 8.74242902e-01 5.80051661e-01
-5.31913757e-01 3.13907638e-02 -9.98694241e-01 3.68557125e-01
9.25629675e-01 -4.33971345e-01 -1.28690147e+00 1.46024987e-01
1.06477118e+00 -2.96684682e-01 -8.60633075e-01 3.78958195e-01
5.68689704e-01 -5.99092364e-01 6.94412887e-01 -6.99493289e-01
1.01209924e-01 -2.23054690e-03 -1.27926409e-01 -1.35502398e+00
-2.98228204e-01 -5.38437009e-01 -1.49535850e-01 1.37931800e+00
7.69661307e-01 -7.60102153e-01 4.42601889e-01 6.33582473e-01
-6.40041605e-02 -7.86928833e-01 -8.77720773e-01 -6.69067442e-01
5.89617431e-01 -6.22097909e-01 5.58852494e-01 9.70164239e-01
2.22620770e-01 7.11493134e-01 -2.02408761e-01 9.23363715e-02
4.28023517e-01 4.79209702e-03 5.36011517e-01 -1.20681405e+00
-4.64187533e-01 -3.23811978e-01 -1.86321393e-01 -1.27695954e+00
2.70731539e-01 -1.09004986e+00 -2.98369467e-01 -1.49628854e+00
3.19067091e-01 -6.59989059e-01 -2.57278621e-01 7.95452237e-01
-1.45555019e-01 -9.39980224e-02 2.94515491e-01 -1.13769376e-03
-4.56101209e-01 5.31348169e-01 8.31794620e-01 2.74063013e-02
-2.39408657e-01 -3.33708525e-02 -8.03877413e-01 8.90032411e-01
7.35810339e-01 -6.19310200e-01 -3.10923845e-01 -1.04273546e+00
5.20080984e-01 -1.15703471e-01 -2.63637453e-01 -7.22035706e-01
4.30659860e-01 7.77838379e-02 -6.79327641e-03 -5.73087692e-01
3.03096801e-01 -8.25672686e-01 -1.14845708e-01 2.51966059e-01
-4.35233295e-01 1.84360042e-01 3.79254222e-01 2.55208731e-01
-3.61241192e-01 -3.89061987e-01 5.78347027e-01 -8.90052021e-02
-5.90352654e-01 1.76679641e-01 -2.91694514e-02 1.54716268e-01
8.61386657e-01 2.43691310e-01 -2.98692822e-01 -4.79771495e-02
-8.09021533e-01 5.26679993e-01 8.58318433e-02 7.83124804e-01
2.24224523e-01 -9.59534585e-01 -7.75624335e-01 2.22387537e-01
-1.50297917e-02 7.43290409e-02 2.39418000e-01 9.09025609e-01
-4.01846051e-01 1.03339720e+00 2.22661808e-01 -4.12087440e-01
-1.18264472e+00 5.18103302e-01 2.81878948e-01 -8.76967192e-01
-5.74136913e-01 9.40184534e-01 3.81455392e-01 -8.95067215e-01
3.35065842e-01 -5.38461804e-01 -5.60499370e-01 1.46939740e-01
6.15829051e-01 2.06687734e-01 1.08361334e-01 -6.38975263e-01
-3.59703720e-01 6.60245001e-01 -3.95671964e-01 2.37483885e-02
1.55224359e+00 -1.01144232e-01 -2.25463554e-01 5.23906291e-01
1.04056847e+00 3.01433593e-01 -4.93333012e-01 -6.19280398e-01
4.65696216e-01 -4.35932130e-02 2.08624639e-02 -1.06354928e+00
-6.39059901e-01 1.12502468e+00 3.02691728e-01 -7.71906748e-02
7.89802909e-01 9.13353264e-02 9.21330988e-01 5.92220545e-01
4.50716972e-01 -9.39829767e-01 -3.72322351e-01 9.84156013e-01
6.64670825e-01 -1.13241041e+00 -2.88651824e-01 -3.69024098e-01
-5.48773050e-01 9.94685233e-01 4.70181644e-01 1.74666177e-02
6.25235260e-01 3.32169443e-01 1.51438802e-01 -7.69780055e-02
-1.17413437e+00 -4.60372753e-02 2.56005019e-01 3.60409439e-01
8.51143420e-01 1.32794857e-01 -6.67800903e-01 1.12497175e+00
-4.39724982e-01 3.49570103e-02 1.65814474e-01 6.90019250e-01
-5.97400427e-01 -1.62137723e+00 -6.24774098e-02 4.90039736e-02
-7.78956234e-01 -5.19564748e-01 -2.54507393e-01 9.20064330e-01
1.70471266e-01 8.03459704e-01 -1.17217980e-01 4.33718450e-02
5.25846839e-01 4.84645396e-01 3.94236058e-01 -8.69202197e-01
-7.06197619e-01 -1.10483013e-01 2.72080600e-01 -4.96023625e-01
-3.45168710e-01 -3.53927225e-01 -1.42095697e+00 -3.06567311e-01
-3.23292345e-01 4.92904842e-01 7.11951613e-01 1.10663855e+00
2.92599678e-01 3.90387595e-01 2.34425604e-01 -3.88471276e-01
-6.73380673e-01 -1.27711999e+00 -2.25648180e-01 1.66897774e-01
1.48661271e-01 -4.51790243e-01 -1.56630367e-01 -1.93577573e-01] | [10.425049781799316, 9.085943222045898] |
338597ab-a3fb-4389-99dc-09a031c6e5ad | not-just-plain-text-fuel-document-level | 2211.05343 | null | https://arxiv.org/abs/2211.05343v2 | https://arxiv.org/pdf/2211.05343v2.pdf | Not Just Plain Text! Fuel Document-Level Relation Extraction with Explicit Syntax Refinement and Subsentence Modeling | Document-level relation extraction (DocRE) aims to identify semantic labels among entities within a single document. One major challenge of DocRE is to dig decisive details regarding a specific entity pair from long text. However, in many cases, only a fraction of text carries required information, even in the manually labeled supporting evidence. To better capture and exploit instructive information, we propose a novel expLicit syntAx Refinement and Subsentence mOdeliNg based framework (LARSON). By introducing extra syntactic information, LARSON can model subsentences of arbitrary granularity and efficiently screen instructive ones. Moreover, we incorporate refined syntax into text representations which further improves the performance of LARSON. Experimental results on three benchmark datasets (DocRED, CDR, and GDA) demonstrate that LARSON significantly outperforms existing methods. | ['Jianyong Wang', 'Zhuo Wang', 'Zhenyu Li', 'Xiuxing Li', 'Zhichao Duan'] | 2022-11-10 | null | null | null | null | ['document-level-relation-extraction'] | ['natural-language-processing'] | [-2.30618613e-03 3.74775767e-01 -7.49823749e-01 -2.43181989e-01
-7.05887616e-01 -8.25255096e-01 7.61597455e-01 7.07463503e-01
-1.61995605e-01 1.15214825e+00 4.98941690e-01 -4.64331061e-01
-4.89736855e-01 -1.07904315e+00 -4.71974641e-01 -6.37990236e-02
8.54742751e-02 5.78984797e-01 4.26518291e-01 -2.83387303e-01
6.24142885e-02 3.70909452e-01 -1.29058182e+00 5.63585401e-01
1.07972288e+00 8.55178654e-01 7.58027583e-02 2.63921291e-01
-5.35721600e-01 1.29299200e+00 -4.70373541e-01 -6.58569694e-01
-2.51641959e-01 1.07478298e-01 -1.32227850e+00 1.00714058e-01
1.81427702e-01 -2.52591334e-02 -2.96224535e-01 9.22791839e-01
-2.12576240e-02 -2.17141435e-02 7.63226807e-01 -9.04524267e-01
-4.42452759e-01 1.22631717e+00 -4.73779380e-01 2.67482936e-01
5.94490349e-01 -5.97749591e-01 1.69804478e+00 -1.04801416e+00
8.71607363e-01 1.17075551e+00 5.49629986e-01 1.53458580e-01
-8.04882407e-01 -6.50180042e-01 5.56232512e-01 1.82527736e-01
-1.43730998e+00 -2.30505764e-01 6.08406901e-01 -3.00284773e-01
1.22735369e+00 4.55074549e-01 1.46468922e-01 6.97650492e-01
-1.85892314e-01 1.03043520e+00 8.12549651e-01 -5.00378132e-01
-5.31040605e-05 1.34249210e-01 7.81925321e-01 7.29504168e-01
8.28362942e-01 -4.19286907e-01 -3.57974172e-01 -2.39299744e-01
3.04872245e-01 3.27389464e-02 -1.77772358e-01 -1.32051148e-02
-8.14226687e-01 5.68433762e-01 2.31571272e-01 4.38948274e-01
-2.26247668e-01 -1.42455623e-01 5.64443588e-01 -7.47922435e-02
3.25800955e-01 6.67445183e-01 -8.38973045e-01 6.58894405e-02
-7.62290359e-01 1.57803997e-01 8.03211749e-01 1.32739091e+00
7.13478744e-01 -6.24583662e-01 -3.46997529e-01 8.93968582e-01
1.44269988e-01 1.14759326e-01 4.05222863e-01 -5.36509931e-01
9.05453503e-01 1.27265644e+00 2.13429719e-01 -9.75008070e-01
-5.53307533e-01 -6.88946128e-01 -6.37842596e-01 -6.88114703e-01
3.33584137e-02 -1.66474789e-01 -7.87982583e-01 1.20283401e+00
2.15793133e-01 -3.79951112e-02 2.83500940e-01 4.85229760e-01
1.28388476e+00 6.83635473e-01 3.12983751e-01 -3.05268854e-01
1.72907436e+00 -8.57196569e-01 -8.27729762e-01 -4.58321601e-01
1.18869042e+00 -4.68861043e-01 7.02991962e-01 2.24231765e-01
-6.58606946e-01 -1.22954309e-01 -1.01622438e+00 -3.93484831e-01
-6.57282114e-01 6.24828994e-01 1.03144264e+00 4.48600620e-01
-2.91931689e-01 4.49897945e-01 -5.74514866e-01 -1.76658668e-02
3.82385552e-01 2.10724920e-01 -4.14520204e-01 -3.53047013e-01
-1.60554612e+00 6.27834439e-01 1.10625482e+00 -1.45716444e-01
-1.79973036e-01 -6.73191965e-01 -1.16188872e+00 3.56961042e-01
1.23626423e+00 -5.08079767e-01 1.12775302e+00 2.09523439e-02
-9.40122604e-01 6.64623678e-01 -2.94271618e-01 -5.39471209e-01
-6.75058085e-03 -5.09346604e-01 -6.71478271e-01 1.08606942e-01
2.44088188e-01 2.84831017e-01 4.57106121e-02 -1.13421512e+00
-8.70451927e-01 -3.03332001e-01 4.16607559e-01 4.63642552e-02
-3.40420663e-01 2.53073037e-01 -7.11022913e-01 -7.08473384e-01
4.76349667e-02 -5.92981756e-01 -1.19971052e-01 -6.55707061e-01
-8.02864373e-01 -8.08081925e-01 5.67823708e-01 -5.81716061e-01
1.96489120e+00 -1.91565549e+00 -1.79881155e-01 2.30972737e-01
5.81187367e-01 3.32574427e-01 2.98883617e-01 6.06750786e-01
-6.99790791e-02 3.99746686e-01 -8.29894468e-02 9.81388390e-02
6.79675043e-02 4.09802854e-01 -5.95854104e-01 -2.12420806e-01
2.29458913e-01 1.12499416e+00 -1.00179696e+00 -8.95153463e-01
-9.43347290e-02 -7.45084584e-02 -5.53610086e-01 -7.79524893e-02
-3.89352739e-01 -2.51905411e-01 -8.80716562e-01 7.06166983e-01
5.28374493e-01 -6.73578799e-01 5.61465859e-01 -3.13734561e-01
6.29507378e-03 1.00997722e+00 -1.29988813e+00 1.14859331e+00
-5.64850092e-01 2.18015790e-01 -5.44339657e-01 -9.22959864e-01
8.72550130e-01 2.13140443e-01 2.84169674e-01 -3.65814745e-01
-5.46191894e-02 2.64403164e-01 -3.64473701e-01 -3.24024588e-01
8.48038256e-01 1.11034088e-01 -3.82149965e-01 1.78091317e-01
2.08555590e-02 2.91058332e-01 7.42327750e-01 6.11037910e-01
1.24467480e+00 -1.89124674e-01 1.00201380e+00 -5.91588952e-02
7.16741860e-01 1.04363017e-01 9.81502175e-01 6.56346798e-01
2.83137262e-01 6.43026233e-02 7.52943337e-01 -3.30295354e-01
-5.03205597e-01 -5.69941938e-01 -2.57737100e-01 1.01517773e+00
1.29335821e-01 -1.33192503e+00 -2.93195009e-01 -1.48001194e+00
7.67503530e-02 9.37976480e-01 -5.26558459e-01 1.55669570e-01
-6.87261760e-01 -7.52070606e-01 7.05151618e-01 8.53350580e-01
5.26980162e-01 -6.04159772e-01 -1.30338311e-01 3.42739135e-01
-2.24465802e-01 -1.56885087e+00 -3.24318618e-01 3.00256222e-01
-7.64078319e-01 -1.20397723e+00 -1.19518071e-01 -6.70847118e-01
6.26849353e-01 1.94604412e-01 1.19496548e+00 5.13719916e-02
-1.07996108e-03 -1.43898383e-01 -5.72239816e-01 -6.71592876e-02
-1.88380420e-01 3.69727701e-01 -3.20172280e-01 -5.85468829e-01
7.06884205e-01 -3.09163362e-01 -1.00123145e-01 3.89289439e-01
-7.06971645e-01 9.13824588e-02 6.61878228e-01 8.40111732e-01
5.48254669e-01 5.59159338e-01 6.68734729e-01 -1.67829204e+00
6.18061185e-01 -5.40943503e-01 -5.12439787e-01 6.90471411e-01
-1.07298028e+00 4.00969237e-01 8.27646315e-01 -5.48685975e-02
-1.30912828e+00 -1.05570264e-01 3.78105305e-02 2.48864919e-01
-9.12487805e-02 1.05910420e+00 -4.56546873e-01 6.39217913e-01
4.79834378e-01 1.27505571e-01 -9.08116281e-01 -6.53931320e-01
5.07774115e-01 7.49563694e-01 6.66430354e-01 -1.07181084e+00
7.18137681e-01 6.31219149e-02 1.45787254e-01 -4.91880238e-01
-1.42836845e+00 -7.39974737e-01 -5.70171893e-01 2.79505849e-01
3.08960676e-01 -1.00249076e+00 -5.63007474e-01 -2.05128089e-01
-1.03639138e+00 1.47189036e-01 -1.71061754e-01 3.07213783e-01
7.46866390e-02 4.53691751e-01 -6.89171433e-01 -5.35180092e-01
-3.14898163e-01 -6.25133932e-01 1.11273849e+00 2.32394233e-01
-4.11623269e-01 -1.05331206e+00 -9.98665690e-02 4.25945789e-01
-3.45797598e-01 1.91600546e-02 1.23704636e+00 -1.06949866e+00
-4.05286998e-01 -4.61992532e-01 -4.96448040e-01 -1.12550028e-01
3.69146556e-01 -3.68855894e-02 -5.33715427e-01 1.33600041e-01
-5.91073155e-01 -2.51129538e-01 9.60325062e-01 -1.94854811e-01
1.33285165e+00 -6.51978493e-01 -8.05031300e-01 1.46018162e-01
1.01377308e+00 8.85108113e-02 4.49067533e-01 2.89213508e-01
7.26059198e-01 5.93988776e-01 1.07011843e+00 4.16839868e-01
6.96300566e-01 6.53655827e-01 -8.77632946e-02 2.93076962e-01
-1.59852877e-01 -6.86273336e-01 -2.54386268e-03 7.47940421e-01
-1.49054751e-01 -3.78126770e-01 -9.19868171e-01 5.66305101e-01
-1.68739414e+00 -8.61059427e-01 -2.62460530e-01 1.59293032e+00
1.27103055e+00 4.73796606e-01 -1.80533946e-01 5.28209865e-01
7.30583787e-01 7.86223486e-02 -1.88499197e-01 1.17745176e-02
-2.60100573e-01 1.29419446e-01 5.50357461e-01 3.54226828e-01
-1.20558751e+00 1.29328728e+00 5.94971991e+00 1.21764660e+00
-5.11752844e-01 -2.13855863e-01 2.86286682e-01 3.99046063e-01
-3.75440061e-01 1.99624702e-01 -1.57375383e+00 2.34353378e-01
7.50523031e-01 -4.68048811e-01 -8.72044787e-02 1.10997796e+00
-1.17317036e-01 2.36476935e-03 -1.17057848e+00 6.34744346e-01
-2.12780133e-01 -1.57808757e+00 3.03256750e-01 -4.29836102e-02
6.47715628e-01 -4.21250105e-01 -3.67903382e-01 5.48043072e-01
7.36590743e-01 -8.81264210e-01 3.69945288e-01 1.18373685e-01
7.17058361e-01 -7.99699783e-01 9.15185273e-01 4.55959380e-01
-1.46323013e+00 -3.42562306e-03 -3.62012804e-01 1.24277279e-01
-8.63478109e-02 9.83190358e-01 -1.12349737e+00 1.06682825e+00
3.06908399e-01 9.85850930e-01 -7.47119784e-01 5.30419648e-01
-7.26576984e-01 6.71810746e-01 -9.41900387e-02 -2.72763938e-01
2.36603111e-01 3.01094264e-01 4.25375670e-01 1.65337217e+00
-3.76672558e-02 5.84729791e-01 1.79278180e-01 5.40842295e-01
-5.32212079e-01 2.70296484e-01 -2.68696338e-01 -2.71356136e-01
9.02099490e-01 1.32745886e+00 -8.18588257e-01 -5.40445626e-01
-4.73135620e-01 5.52087069e-01 6.03129923e-01 1.20618679e-01
-4.11004186e-01 -6.22314990e-01 2.66336501e-01 -8.82336348e-02
4.07097667e-01 -1.25612020e-01 -2.92116016e-01 -1.27561998e+00
1.64920852e-01 -6.82306528e-01 9.34020996e-01 -6.68116987e-01
-1.04004562e+00 4.50464129e-01 1.30591735e-01 -9.82442617e-01
-3.52869034e-01 -7.17506766e-01 -1.70374647e-01 4.78774458e-01
-1.49079764e+00 -1.12730587e+00 -1.16461121e-01 2.74361610e-01
6.37360632e-01 8.10728502e-03 7.15045869e-01 4.12521213e-01
-7.74538219e-01 6.64372265e-01 -2.00178832e-01 4.51014489e-01
4.84884858e-01 -1.45743382e+00 3.43283027e-01 8.06754589e-01
2.68325329e-01 1.05898952e+00 5.10593474e-01 -1.03085983e+00
-1.13778484e+00 -1.34824276e+00 1.47102070e+00 -5.47593594e-01
8.05969775e-01 -1.41175583e-01 -1.04552722e+00 9.80317295e-01
-1.55766368e-01 4.55095731e-02 8.03827167e-01 6.29842341e-01
-7.19003916e-01 3.84726301e-02 -8.26081157e-01 5.01127958e-01
1.15912676e+00 -5.76518953e-01 -1.15700650e+00 3.24551404e-01
8.60075772e-01 -5.81266642e-01 -9.70837176e-01 6.34968698e-01
3.26044470e-01 -3.25255841e-01 8.87546897e-01 -9.11137879e-01
6.22169495e-01 -4.64147389e-01 -1.41374588e-01 -1.06137109e+00
-2.15603396e-01 -5.51413953e-01 -7.92808235e-01 1.63836122e+00
7.80590415e-01 -3.78232449e-01 6.36282742e-01 7.46049583e-01
-9.10941288e-02 -8.52782190e-01 -4.40942734e-01 -1.03257489e+00
-2.95463372e-02 -5.25502443e-01 9.50126827e-01 1.12379169e+00
5.43542743e-01 6.13309622e-01 -1.41857237e-01 4.04653281e-01
3.40197742e-01 4.27167952e-01 4.39099818e-01 -1.47959638e+00
-1.56345695e-01 -1.82584330e-01 -2.81836987e-01 -1.33777213e+00
3.96056950e-01 -1.23584306e+00 -2.15847418e-01 -1.75688922e+00
4.29444402e-01 -7.01947808e-01 -3.28003883e-01 7.79876769e-01
-5.29072702e-01 -2.41252482e-01 -3.51087064e-01 7.84328952e-02
-9.36076581e-01 3.54594350e-01 9.84309077e-01 -1.06116362e-01
-1.15102246e-01 -1.54620339e-03 -1.14427173e+00 9.38638568e-01
4.82211083e-01 -6.68268681e-01 -3.22710752e-01 -1.09068125e-01
6.70073688e-01 -1.83620565e-02 5.50182611e-02 -4.14552093e-01
3.70319515e-01 -4.27465886e-01 1.46414563e-01 -9.68052447e-01
-2.65319645e-01 -6.69995129e-01 -1.50921389e-01 9.29605141e-02
-6.16714299e-01 -3.80125195e-01 -6.50570262e-03 5.21970272e-01
-4.34170842e-01 -4.95314181e-01 1.99951828e-01 2.06073225e-01
-8.14787328e-01 1.84079602e-01 -1.06642053e-01 4.36592370e-01
6.87892973e-01 3.46686721e-01 -6.54441178e-01 1.64317995e-01
-5.48999250e-01 5.06541610e-01 5.75658493e-03 3.63170385e-01
4.81425762e-01 -1.34713864e+00 -4.97445524e-01 -3.36186528e-01
5.62130332e-01 2.12276220e-01 -4.86114100e-02 4.09935832e-01
-1.93900645e-01 9.80986118e-01 5.20759404e-01 -5.61819188e-02
-1.38466072e+00 5.49173772e-01 -2.30749026e-02 -8.42927516e-01
-7.77384818e-01 5.59557140e-01 1.78813577e-01 -3.00833672e-01
8.07497352e-02 -7.18378544e-01 -7.26904988e-01 2.81011667e-02
7.03201711e-01 3.67230505e-01 3.21596086e-01 -3.37564349e-01
-5.53304315e-01 2.70251423e-01 -6.63559318e-01 3.54652613e-01
1.22049332e+00 -1.37233794e-01 -2.47700632e-01 1.19696654e-01
7.70828724e-01 6.49736524e-01 -7.73598611e-01 -6.54991031e-01
7.01996863e-01 -2.39485517e-01 1.81175202e-01 -1.01098275e+00
-7.99387813e-01 4.95399475e-01 -4.44328398e-01 2.11686552e-01
1.02938652e+00 4.20691311e-01 8.17851663e-01 8.69565904e-01
3.07625949e-01 -1.01781178e+00 -2.89544404e-01 7.12575912e-01
7.11038351e-01 -9.64298368e-01 3.72896075e-01 -1.26810265e+00
-5.02041042e-01 1.12916219e+00 5.49950421e-01 2.88024634e-01
4.43136603e-01 4.21997309e-01 -5.63176751e-01 -3.29622775e-01
-8.85501146e-01 -3.47589642e-01 6.08571291e-01 3.12454581e-01
7.20528305e-01 1.18827373e-01 -6.66960239e-01 1.40634286e+00
-2.49232262e-01 -6.16473917e-05 4.33756649e-01 9.92656589e-01
-5.35598397e-01 -1.24085343e+00 -3.67769203e-03 5.56850374e-01
-7.99886942e-01 -3.88558418e-01 -6.05126798e-01 7.42018461e-01
1.68958172e-01 9.56921697e-01 -3.38408291e-01 -3.78964990e-01
2.62586862e-01 1.42621398e-01 2.70703554e-01 -1.01910830e+00
-3.24871600e-01 -1.49152786e-01 8.78240705e-01 -1.52549461e-01
-3.39913547e-01 -5.63996494e-01 -1.75645292e+00 -1.57518968e-01
-4.66807127e-01 5.11384308e-01 2.12759614e-01 1.31626308e+00
3.25056523e-01 6.85968161e-01 5.80801606e-01 2.63300657e-01
-1.20172299e-01 -8.50875080e-01 -5.56862175e-01 1.23034179e-01
1.51279652e-02 -8.92854571e-01 -7.55930990e-02 -1.95862427e-02] | [9.352190017700195, 8.601242065429688] |
77e3e147-d6a1-43db-8349-38e42ef02a04 | gaze-estimation-with-eye-region-segmentation | 2112.07878 | null | https://arxiv.org/abs/2112.07878v1 | https://arxiv.org/pdf/2112.07878v1.pdf | Gaze Estimation with Eye Region Segmentation and Self-Supervised Multistream Learning | We present a novel multistream network that learns robust eye representations for gaze estimation. We first create a synthetic dataset containing eye region masks detailing the visible eyeball and iris using a simulator. We then perform eye region segmentation with a U-Net type model which we later use to generate eye region masks for real-world eye images. Next, we pretrain an eye image encoder in the real domain with self-supervised contrastive learning to learn generalized eye representations. Finally, this pretrained eye encoder, along with two additional encoders for visible eyeball region and iris, are used in parallel in our multistream framework to extract salient features for gaze estimation from real-world images. We demonstrate the performance of our method on the EYEDIAP dataset in two different evaluation settings and achieve state-of-the-art results, outperforming all the existing benchmarks on this dataset. We also conduct additional experiments to validate the robustness of our self-supervised network with respect to different amounts of labeled data used for training. | ['Ali Etemad', 'Paul Hungler', 'Zunayed Mahmud'] | 2021-12-15 | null | null | null | null | ['gaze-estimation'] | ['computer-vision'] | [ 1.55099899e-01 1.83958039e-01 -1.75692037e-01 -6.68227017e-01
-1.73272446e-01 -3.32388222e-01 4.28950101e-01 -5.89337587e-01
-4.77635711e-01 5.76266110e-01 6.86141849e-02 -1.21888839e-01
3.37249249e-01 -1.58080906e-01 -9.87603545e-01 -3.44504446e-01
3.54408994e-02 -3.17542404e-02 3.14962596e-01 -1.23448104e-01
3.08889568e-01 2.99601436e-01 -2.44965959e+00 2.94876605e-01
7.91536152e-01 1.27596855e+00 -1.30591497e-01 9.80490625e-01
3.23546797e-01 1.07977974e+00 -6.73601985e-01 -2.95037270e-01
5.77173173e-01 -5.84230542e-01 -8.88195276e-01 5.60267448e-01
1.22698808e+00 -6.09244645e-01 1.08460686e-03 9.41778243e-01
4.93836850e-01 1.65824682e-01 5.87028384e-01 -1.37323415e+00
-6.47296429e-01 -7.06630498e-02 -9.81685162e-01 5.48006475e-01
4.62860614e-01 7.29129970e-01 9.03391778e-01 -7.00749457e-01
7.54548430e-01 1.17479491e+00 1.09345235e-01 7.97547519e-01
-1.07572269e+00 -1.08955896e+00 1.85773954e-01 4.83700000e-02
-1.20558083e+00 -9.69159961e-01 3.54426861e-01 -3.06290835e-01
8.06183875e-01 -1.68924648e-02 6.28108382e-01 1.06722939e+00
1.96673438e-01 1.03798223e+00 1.46640503e+00 -4.52016473e-01
-2.36066684e-01 1.41739026e-01 2.59323835e-01 9.59438860e-01
2.31870469e-02 4.79178011e-01 -8.10883403e-01 4.01142627e-01
6.93642974e-01 -2.18279243e-01 -5.63554347e-01 -1.24903806e-01
-9.44942713e-01 5.53268194e-01 7.36701787e-01 -3.59742612e-01
-1.68303549e-01 6.06883243e-02 2.25820821e-02 4.31007564e-01
7.10325122e-01 3.70284230e-01 -3.97805214e-01 1.30257577e-01
-1.04205370e+00 2.39374071e-01 5.42634070e-01 1.10909355e+00
8.88390660e-01 -1.75704658e-01 -5.21951020e-01 5.58745980e-01
6.49487734e-01 2.46586964e-01 4.40323472e-01 -8.56480956e-01
2.05602020e-01 6.14939332e-01 1.47589564e-01 -2.10054502e-01
-5.25671780e-01 -4.60534751e-01 -3.10738027e-01 7.41351485e-01
5.85683882e-01 -5.22008121e-01 -1.42164099e+00 1.59233105e+00
3.95975471e-01 7.44064927e-01 -6.20912248e-03 1.18228281e+00
1.16437650e+00 2.41224021e-01 -1.02730125e-01 1.18939556e-01
1.38893092e+00 -1.45210397e+00 -5.59116602e-01 -3.19712669e-01
5.00800669e-01 -7.42616951e-01 1.14041460e+00 3.03843081e-01
-1.25701249e+00 -7.69379795e-01 -9.48945343e-01 -3.98023874e-01
-4.57508042e-02 3.20240229e-01 5.10213852e-01 5.85606277e-01
-1.44585347e+00 9.38918442e-02 -6.87845230e-01 -1.44070297e-01
8.97575557e-01 6.39926076e-01 -2.39248917e-01 2.57633597e-01
-6.81287169e-01 6.23732686e-01 2.53791809e-01 1.44080976e-02
-1.01063275e+00 -9.02174354e-01 -1.32600486e+00 -1.81970429e-02
1.96229756e-01 -7.66102016e-01 1.43579888e+00 -1.52177286e+00
-1.75353920e+00 1.47910845e+00 -5.87791860e-01 -5.89532197e-01
3.05966437e-01 -3.01979274e-01 -5.82062900e-01 1.89341441e-01
-1.30255207e-01 1.24332452e+00 1.58757472e+00 -1.13650858e+00
-1.04380953e+00 -2.31245428e-01 4.26427960e-01 1.17137410e-01
5.85878342e-02 6.72313690e-01 -7.21993625e-01 -3.54153901e-01
-6.50473535e-01 -8.88163030e-01 3.74348402e-01 1.59914479e-01
-6.96831465e-01 -3.94901842e-01 5.69132745e-01 -3.82458866e-01
1.09161210e+00 -2.15985060e+00 1.19287781e-01 -4.72212620e-02
8.58718872e-01 1.88415080e-01 -3.96519363e-01 -4.22201991e-01
-4.93823946e-01 -2.70608485e-01 3.38915065e-02 -1.03268802e+00
-3.46460491e-01 -2.57477969e-01 -1.90588936e-01 6.33129358e-01
7.10288167e-01 1.04264200e+00 -7.87614226e-01 -5.27730048e-01
-7.90762007e-02 5.28125346e-01 -5.41739762e-01 4.74767983e-01
-3.80339473e-01 4.16240364e-01 1.35046458e-02 6.88490510e-01
5.58710635e-01 -7.12649465e-01 -5.36770999e-01 8.81446302e-02
-1.95025057e-02 1.20391205e-01 -7.92043686e-01 1.50781393e+00
-3.91118854e-01 1.18755496e+00 -1.87690884e-01 -2.75359657e-02
5.31279981e-01 -7.13437647e-02 -1.56320333e-01 -8.14571917e-01
5.81427574e-01 -2.32065931e-01 2.10740700e-01 -4.97275501e-01
5.20402193e-01 2.98662573e-01 4.71086055e-01 8.55566025e-01
5.03416896e-01 3.24554026e-01 2.86684632e-01 -1.54671539e-02
6.04503155e-01 3.46034169e-01 1.42299104e-02 -1.82649374e-01
7.55260885e-01 -4.29791272e-01 1.81261793e-01 -1.15224766e-02
-4.70908642e-01 9.33808804e-01 7.07611740e-01 -4.10285920e-01
-5.83290219e-01 -7.54462838e-01 -2.11097494e-01 1.35271955e+00
2.76562810e-01 -2.26872399e-01 -1.08748436e+00 -9.87247586e-01
-1.00687392e-01 5.33729017e-01 -1.22286785e+00 -2.04733643e-03
3.25502530e-02 -4.47397470e-01 1.85520351e-01 3.69734079e-01
4.79235083e-01 -1.16758227e+00 -9.17448342e-01 -5.36517382e-01
3.40949982e-01 -1.25326824e+00 -7.26372898e-01 -2.44733945e-01
-3.11096400e-01 -1.59442222e+00 -8.66135657e-01 -7.99801111e-01
9.40585256e-01 2.40524739e-01 1.35850310e+00 3.35762143e-01
-4.13675517e-01 9.58016813e-02 7.33236011e-05 -7.66237557e-01
-1.85887873e-01 2.03118190e-01 1.71359889e-02 6.35729551e-01
7.66880989e-01 1.29648983e-01 -1.05852902e+00 2.83993423e-01
-6.65676832e-01 8.18994269e-02 3.25456500e-01 7.24516809e-01
3.95095468e-01 -5.01138210e-01 7.45657831e-02 -1.01689994e+00
5.68738699e-01 -3.96914661e-01 -1.05952883e+00 1.43924961e-02
-3.27300131e-01 1.15738548e-01 9.47552547e-02 -2.72619843e-01
-9.67019081e-01 -6.93588406e-02 1.57607660e-01 -7.43774652e-01
-2.87447780e-01 -1.98685989e-01 2.68667489e-01 -2.87319899e-01
9.86401796e-01 5.67902885e-02 4.68916327e-01 -2.10626110e-01
3.01642776e-01 9.69928265e-01 4.77925509e-01 -4.95714247e-02
6.51161075e-01 5.83746850e-01 -1.84514597e-01 -5.71718097e-01
-1.55167747e+00 -4.17589426e-01 -6.14321113e-01 -1.40755087e-01
9.15950418e-01 -1.20079827e+00 -1.11488748e+00 7.11896122e-01
-9.42684352e-01 -7.27507532e-01 -1.42539993e-01 3.15124810e-01
-6.45219326e-01 -9.27697420e-02 -4.32007343e-01 -5.44484377e-01
-3.24360132e-01 -1.40457213e+00 1.46109092e+00 8.35430026e-01
-8.60174969e-02 -8.23176622e-01 2.76690926e-02 4.35157955e-01
1.73641220e-01 9.99901593e-02 1.15114853e-01 -4.14863378e-01
-6.88355744e-01 2.31552184e-01 -5.88647366e-01 5.09942055e-01
1.48171246e-01 2.61042029e-01 -1.48965383e+00 -4.74110633e-01
-5.82126856e-01 -6.94648981e-01 9.48448241e-01 3.51523161e-01
1.22988343e+00 5.55047803e-02 -3.20309967e-01 1.24699414e+00
1.07104409e+00 -4.91453290e-01 6.64476871e-01 2.75839388e-01
6.24496698e-01 8.72585237e-01 5.69677532e-01 2.66767085e-01
5.62180758e-01 3.73095542e-01 6.05393350e-01 -5.39589643e-01
-4.42502886e-01 1.30899414e-01 1.80297524e-01 -2.22913235e-01
-2.00729519e-01 -3.04440528e-01 -6.96547151e-01 6.26072884e-01
-1.44201636e+00 -7.26048231e-01 2.07393929e-01 2.17651367e+00
1.16173518e+00 6.22178428e-02 6.60394013e-01 -2.43408233e-01
5.27762532e-01 7.39380941e-02 -7.73666322e-01 -1.86067864e-01
1.29972816e-01 5.90350389e-01 5.32517672e-01 3.75661939e-01
-1.30441356e+00 1.30740333e+00 6.81630754e+00 5.40438965e-02
-1.56617761e+00 -5.73555827e-02 9.16815639e-01 -7.00822771e-01
1.05006561e-01 -5.05053401e-01 -1.15557015e+00 7.25446522e-01
1.25035405e+00 1.28009945e-01 4.48253065e-01 4.82191890e-01
2.93751173e-02 -1.30209446e-01 -1.27788889e+00 1.23048663e+00
4.65146691e-01 -1.26772153e+00 -2.53510028e-01 1.65830448e-01
8.86508524e-01 5.72879970e-01 4.09568787e-01 5.53415194e-02
3.31699938e-01 -1.38784659e+00 4.77007300e-01 6.82873249e-01
1.31680191e+00 -6.41281426e-01 4.75163162e-01 -1.30288646e-01
-5.50598919e-01 8.20287466e-02 -1.87928289e-01 -9.27929953e-03
-2.35456109e-01 -2.14662388e-01 -6.95101857e-01 -5.60459048e-02
9.24189806e-01 1.02124441e+00 -1.11292040e+00 1.38379467e+00
-4.56195801e-01 4.54024702e-01 -2.87186235e-01 2.78328508e-01
1.36843011e-01 1.22277573e-01 2.68739998e-01 7.37871587e-01
-2.55285949e-01 -4.09375459e-01 -5.52129745e-01 1.14096332e+00
-5.72920799e-01 -6.96567148e-02 -3.49875808e-01 2.42204428e-01
2.61835873e-01 1.21651232e+00 -3.71329337e-01 -2.37625286e-01
-7.92759597e-01 1.04854667e+00 4.16071028e-01 8.44131529e-01
-7.33307600e-01 -5.16026855e-01 1.24841058e+00 2.62169659e-01
4.73714203e-01 2.86900729e-01 1.42045051e-01 -1.11866558e+00
-1.96532220e-01 -9.15043771e-01 1.95929065e-01 -1.06425095e+00
-9.58610415e-01 9.50611770e-01 -1.22188404e-01 -1.33965528e+00
-6.05220377e-01 -8.26624453e-01 -5.81709325e-01 1.31860590e+00
-2.35576820e+00 -1.22504210e+00 -7.58748114e-01 1.00126183e+00
5.08105755e-01 -5.11176884e-01 4.49475020e-01 -1.76370054e-01
-9.92962360e-01 1.11527073e+00 -3.32164854e-01 3.82923216e-01
1.11909950e+00 -1.32471907e+00 7.90006638e-01 9.02097523e-01
9.06097069e-02 7.13933825e-01 4.69979554e-01 -1.71111658e-01
-9.48026538e-01 -1.22420537e+00 5.22219837e-01 -8.65637422e-01
5.33779025e-01 -4.51919138e-01 -7.02408671e-01 1.22705543e+00
7.22960591e-01 7.41449118e-01 6.58409059e-01 1.31110489e-01
-3.69950682e-01 -5.26464172e-02 -9.65595424e-01 6.81696177e-01
9.17547405e-01 -5.47161043e-01 -5.91836333e-01 2.98516333e-01
7.39048958e-01 -9.87780631e-01 -6.64141834e-01 7.65586421e-02
5.15471935e-01 -1.13699389e+00 7.47304499e-01 -9.23903286e-01
6.59335792e-01 -2.78748870e-01 6.61505699e-01 -1.23019278e+00
1.95542872e-01 -1.13276517e+00 -4.25011009e-01 8.35302591e-01
4.95200843e-01 -8.54492605e-01 7.35728920e-01 2.86125392e-01
2.01331764e-01 -8.78103316e-01 -2.91318238e-01 -2.43835881e-01
-2.49698475e-01 -2.20060736e-01 8.23612273e-01 4.22238618e-01
-6.03624880e-01 5.17305315e-01 -1.96907118e-01 6.81795552e-02
9.51729178e-01 1.64186805e-01 1.21488738e+00 -1.30230832e+00
-1.63070858e-01 -4.15175647e-01 -5.99891007e-01 -1.17700410e+00
6.14734054e-01 -4.64731753e-01 -7.23428577e-02 -9.52964365e-01
-5.25204130e-02 1.31114721e-01 -4.59688723e-01 3.79493982e-01
-5.54810286e-01 8.95563364e-01 1.00925513e-01 1.89010680e-01
-7.06445456e-01 2.29675993e-01 1.59489095e+00 3.20572145e-02
-4.80572999e-01 1.31246179e-01 -9.57912385e-01 6.06263459e-01
4.37510014e-01 -4.74100523e-02 -6.11622930e-01 -4.34717327e-01
1.99999288e-01 -3.81902605e-01 3.80971313e-01 -8.19391787e-01
3.63362491e-01 1.96705937e-01 5.60944796e-01 -4.04032111e-01
1.37989491e-01 -4.65788603e-01 -8.99767458e-01 -1.45997941e-01
-3.09313923e-01 -2.28604585e-01 6.01679623e-01 4.29114342e-01
-3.00139815e-01 1.79891899e-01 1.06591177e+00 3.69805396e-01
-8.37834656e-01 6.79315329e-01 2.44163200e-01 4.30338353e-01
1.25508046e+00 -4.35533702e-01 -4.89429146e-01 -3.33144993e-01
-6.51194513e-01 6.28739953e-01 9.48484421e-01 7.42228866e-01
5.59950113e-01 -7.12264180e-01 -7.59570599e-01 9.20008838e-01
6.72406673e-01 1.54401004e-01 2.37543061e-02 9.17662740e-01
-7.65429974e-01 2.32272267e-01 -5.27145386e-01 -9.85686958e-01
-1.44842422e+00 6.37845516e-01 7.43521273e-01 4.96316373e-01
-6.24308586e-01 1.20439875e+00 5.48534989e-01 2.00687572e-01
4.61896002e-01 -6.76335692e-01 -6.03577197e-01 2.12214589e-02
1.07574952e+00 -2.67363787e-02 -1.00124501e-01 -8.56322527e-01
-1.12243123e-01 6.22599244e-01 -3.89899313e-01 2.83773065e-01
1.05846846e+00 -4.39806223e-01 -6.22616336e-03 1.93674192e-01
1.17492259e+00 1.09506864e-02 -1.81685781e+00 -3.97028416e-01
-3.49305838e-01 -5.17570853e-01 3.52759361e-01 -6.81418955e-01
-1.34829187e+00 8.17513645e-01 8.01078081e-01 -3.04188579e-01
1.42478240e+00 3.57345492e-03 6.49615407e-01 -7.39004556e-03
-5.54428026e-02 -7.84497499e-01 -9.55874175e-02 1.95941374e-01
6.87707543e-01 -1.82819319e+00 -3.10554594e-01 -3.15498322e-01
-7.52826095e-01 8.86133432e-01 8.36942673e-01 -3.75034899e-01
7.63893962e-01 1.14145465e-01 5.80426037e-01 -3.86874765e-01
-8.66813838e-01 -9.66046870e-01 9.35926855e-01 5.98393381e-01
5.39527357e-01 -4.11933184e-01 3.88999283e-01 1.81166490e-03
-5.01936197e-01 3.11733782e-01 5.75006485e-01 4.60111082e-01
-3.35659944e-02 -7.95745432e-01 -1.62625283e-01 6.15815759e-01
-6.98705316e-01 -2.52353251e-01 -4.03746665e-01 8.95985246e-01
-3.59517969e-02 8.00731063e-01 7.91487694e-01 -1.77812830e-01
1.47925541e-02 -3.88244577e-02 7.72958159e-01 -1.07837093e+00
-6.21462584e-01 -1.67855874e-01 -1.52281374e-01 -1.04506540e+00
-4.88931090e-01 -4.21481013e-01 -9.28369343e-01 -1.00157969e-01
-1.52188390e-01 -4.45169687e-01 3.49323332e-01 8.36231232e-01
5.78278244e-01 6.97889447e-01 5.43040156e-01 -9.11846638e-01
-1.36938751e-01 -1.20610738e+00 -5.92958033e-01 5.47070444e-01
1.10428393e+00 -9.23202395e-01 -3.94748658e-01 1.83133349e-01] | [14.122995376586914, 0.04828708991408348] |
2acdf16d-0c20-426a-adcf-c41c9f9cd5ab | crossner-evaluating-cross-domain-named-entity | 2012.04373 | null | https://arxiv.org/abs/2012.04373v2 | https://arxiv.org/pdf/2012.04373v2.pdf | CrossNER: Evaluating Cross-Domain Named Entity Recognition | Cross-domain named entity recognition (NER) models are able to cope with the scarcity issue of NER samples in target domains. However, most of the existing NER benchmarks lack domain-specialized entity types or do not focus on a certain domain, leading to a less effective cross-domain evaluation. To address these obstacles, we introduce a cross-domain NER dataset (CrossNER), a fully-labeled collection of NER data spanning over five diverse domains with specialized entity categories for different domains. Additionally, we also provide a domain-related corpus since using it to continue pre-training language models (domain-adaptive pre-training) is effective for the domain adaptation. We then conduct comprehensive experiments to explore the effectiveness of leveraging different levels of the domain corpus and pre-training strategies to do domain-adaptive pre-training for the cross-domain task. Results show that focusing on the fractional corpus containing domain-specialized entities and utilizing a more challenging pre-training strategy in domain-adaptive pre-training are beneficial for the NER domain adaptation, and our proposed method can consistently outperform existing cross-domain NER baselines. Nevertheless, experiments also illustrate the challenge of this cross-domain NER task. We hope that our dataset and baselines will catalyze research in the NER domain adaptation area. The code and data are available at https://github.com/zliucr/CrossNER. | ['Pascale Fung', 'Andrea Madotto', 'Samuel Cahyawijaya', 'Ziwei Ji', 'Wenliang Dai', 'Tiezheng Yu', 'Yan Xu', 'Zihan Liu'] | 2020-12-08 | null | null | null | null | ['cross-domain-named-entity-recognition'] | ['natural-language-processing'] | [-2.77971715e-01 -2.33983487e-01 -1.65956110e-01 -4.57906455e-01
-8.79915357e-01 -1.05948520e+00 6.69088840e-01 2.07910195e-01
-9.36889231e-01 8.97137344e-01 3.56312484e-01 -1.00022823e-01
1.52359381e-01 -9.03505266e-01 -4.88405704e-01 -1.76314175e-01
2.25406930e-01 5.72002828e-01 3.62588972e-01 -4.99603152e-01
-7.47409314e-02 4.25521910e-01 -1.06429768e+00 1.69126615e-01
1.18300462e+00 5.22927284e-01 2.96528220e-01 2.31978640e-01
-3.63322318e-01 1.61778480e-01 -7.52973437e-01 -6.21062398e-01
2.96044618e-01 -3.70917201e-01 -9.81031477e-01 -5.66050112e-01
1.02926292e-01 -8.14277455e-02 -2.87509084e-01 9.68678772e-01
9.06822681e-01 3.64529103e-01 6.88691616e-01 -9.04241621e-01
-9.42333817e-01 6.75450802e-01 -1.46306977e-01 3.72705936e-01
2.46882394e-01 2.70338021e-02 8.88065040e-01 -9.16499496e-01
8.75676572e-01 9.43193376e-01 9.09584403e-01 9.57936168e-01
-1.02677357e+00 -8.94693017e-01 8.15962404e-02 1.40114754e-01
-1.44664776e+00 -4.02292728e-01 7.76015282e-01 -1.17094502e-01
1.08672988e+00 -2.03666911e-01 -1.27076983e-01 1.59961820e+00
-4.02237415e-01 5.87572157e-01 9.60977554e-01 -4.15347219e-01
2.40575537e-01 1.67792678e-01 3.16654056e-01 -5.45900054e-02
3.91701758e-01 2.08166152e-01 5.42647438e-03 -2.01743916e-01
4.79246736e-01 -1.12265818e-01 -2.79669344e-01 -1.89004481e-01
-1.19585741e+00 6.23424113e-01 2.85661191e-01 7.71711648e-01
-4.31296438e-01 -5.89161396e-01 7.74828076e-01 3.48548949e-01
3.89495194e-01 8.60974610e-01 -1.10887110e+00 -1.58133149e-01
-7.69348681e-01 6.25776574e-02 1.03298652e+00 1.06112301e+00
7.37832069e-01 -9.92523730e-02 -2.76403546e-01 1.31955111e+00
-1.22335255e-01 3.64234209e-01 7.32631326e-01 -5.15861094e-01
8.84565175e-01 7.79906452e-01 2.57658750e-01 -5.11989832e-01
-4.48755860e-01 -1.97131738e-01 -7.10024595e-01 -2.27384388e-01
6.10938013e-01 -7.53653347e-01 -9.34087634e-01 2.08825493e+00
2.86300331e-01 2.62196779e-01 6.19517565e-01 6.37031615e-01
1.09727895e+00 5.43419898e-01 7.02548504e-01 3.65422815e-01
1.66244555e+00 -8.56907606e-01 -5.11478305e-01 -3.58002186e-01
1.00385845e+00 -5.72140634e-01 1.31466246e+00 -1.37438923e-01
-5.21208584e-01 -5.99457204e-01 -8.67780566e-01 -1.35972589e-01
-9.62236345e-01 3.14432859e-01 1.54120862e-01 5.48436284e-01
-5.42082667e-01 3.92633587e-01 -6.19331837e-01 -7.48251736e-01
1.81404650e-01 -8.36342853e-03 -7.02627420e-01 -3.50455165e-01
-1.83650112e+00 1.13540149e+00 9.79031384e-01 -4.56063986e-01
-7.35671997e-01 -8.99095118e-01 -9.60331261e-01 1.93480641e-01
2.00389683e-01 -5.33706903e-01 1.21882701e+00 -6.15486324e-01
-1.12125826e+00 8.88820589e-01 1.55725822e-01 -2.94129282e-01
3.21196645e-01 -3.29359800e-01 -1.06438804e+00 -1.23425521e-01
3.67523134e-01 6.19617164e-01 1.93761006e-01 -1.14680552e+00
-7.19267666e-01 -9.83475223e-02 2.55354166e-01 1.74798280e-01
-7.06201553e-01 7.95463026e-02 -2.66802847e-01 -7.21037388e-01
-5.26014566e-01 -7.48692393e-01 -2.62176394e-01 -9.16356504e-01
-2.60174096e-01 -4.18880016e-01 6.00606263e-01 -6.56762600e-01
1.47146654e+00 -2.16620779e+00 -2.65537113e-01 -3.63245457e-01
-1.48697495e-01 6.39244139e-01 -6.06240034e-01 6.61731362e-01
-3.75893205e-01 4.21465069e-01 -2.85001397e-01 -2.11580113e-01
1.10998482e-01 1.73806921e-01 -7.70713985e-02 -7.30051249e-02
4.66898561e-01 6.50937915e-01 -1.01553059e+00 -4.76421922e-01
3.70661877e-02 4.64249849e-01 -4.46904659e-01 2.76108772e-01
-4.76611927e-02 5.12654722e-01 -6.10949039e-01 5.50335407e-01
8.60924542e-01 -9.09853503e-02 3.12161803e-01 -3.72176558e-01
-7.60798231e-02 5.73069811e-01 -1.25244272e+00 1.79235768e+00
-6.09242022e-01 2.56322563e-01 -1.82736397e-01 -9.30836976e-01
9.97416377e-01 5.21541178e-01 3.01378757e-01 -8.96109998e-01
1.07287250e-01 4.63674188e-01 -1.39106378e-01 -2.32191309e-01
6.51505768e-01 -2.88234204e-01 -5.54923296e-01 1.15767553e-01
5.10663211e-01 2.96335071e-01 4.81215030e-01 -1.16910458e-01
1.29924321e+00 -1.00987842e-02 5.85859001e-01 -5.75644597e-02
5.35347879e-01 3.15220416e-01 1.01422262e+00 4.75650221e-01
-4.93058264e-01 4.74282384e-01 5.88520691e-02 -1.08176999e-01
-1.21198595e+00 -8.57201636e-01 -3.56511056e-01 1.29798532e+00
1.44137740e-01 -3.18702966e-01 -6.89522624e-01 -1.28441942e+00
-1.62978247e-01 1.05076635e+00 -3.69050384e-01 4.86453436e-03
-6.69054747e-01 -8.00916314e-01 1.22125912e+00 7.16368318e-01
7.96949029e-01 -1.20512092e+00 -1.46554142e-01 2.97354728e-01
-4.46317941e-01 -1.41552579e+00 -5.42660415e-01 4.14353162e-01
-7.79936790e-01 -9.88841057e-01 -9.40474510e-01 -1.02368248e+00
4.56523925e-01 2.75338888e-02 1.55394745e+00 -3.55987281e-01
2.73455471e-01 4.48803842e-01 -7.72947192e-01 -1.22475930e-01
-5.76673210e-01 6.69625163e-01 4.40376438e-02 -5.85177600e-01
9.57279265e-01 -6.13314927e-01 -4.03039038e-01 6.58412158e-01
-9.62373674e-01 -5.46808422e-01 6.33652210e-01 1.00486183e+00
4.99892056e-01 1.07678585e-01 9.40898895e-01 -1.07471538e+00
7.67854810e-01 -9.49896872e-01 -4.09955561e-01 4.02313501e-01
-3.50222588e-01 7.21999481e-02 1.08585966e+00 -6.78708017e-01
-1.42536998e+00 -1.22895285e-01 -3.48553270e-01 -1.19745461e-02
-6.70978189e-01 4.96930480e-01 -7.24159002e-01 3.58446151e-01
1.09053552e+00 9.29128826e-02 -8.24223936e-01 -9.09756720e-01
3.91537488e-01 7.33110189e-01 5.75473726e-01 -1.00527048e+00
7.61445224e-01 -1.04079857e-01 -6.17877662e-01 -4.44274426e-01
-6.75380528e-01 -6.55953348e-01 -7.45498657e-01 5.46747625e-01
8.36877465e-01 -1.22597623e+00 1.13738284e-01 1.87841490e-01
-1.20623207e+00 -3.51957589e-01 -2.77231157e-01 5.59406996e-01
-1.00957982e-01 4.17132467e-01 -5.58033943e-01 -3.45809639e-01
-3.86907876e-01 -7.58190215e-01 7.05912292e-01 5.52774251e-01
-1.28313214e-01 -1.34604526e+00 4.62169796e-01 4.76140492e-02
2.99581200e-01 1.40793160e-01 8.07111681e-01 -1.68547249e+00
1.98224217e-01 -8.96521881e-02 -1.82917118e-01 5.95255852e-01
2.55802542e-01 -4.25825894e-01 -7.91595459e-01 -1.80182546e-01
-3.20025951e-01 -3.88010621e-01 5.71814120e-01 -2.14058712e-01
6.93487287e-01 7.33792931e-02 -5.04675329e-01 5.19336820e-01
1.59034860e+00 1.80094182e-01 5.48996806e-01 8.51033628e-01
4.76335287e-01 4.13924724e-01 9.03953373e-01 3.92240942e-01
8.23065698e-01 5.77715755e-01 -5.34024835e-02 -2.46883854e-01
-1.88911989e-01 -4.54263568e-01 2.46374205e-01 7.09027827e-01
3.29171643e-02 -5.38275003e-01 -1.31879067e+00 1.07891822e+00
-1.41204834e+00 -8.57186258e-01 8.97375867e-02 1.96976829e+00
1.20025849e+00 -1.38680726e-01 1.49633482e-01 -2.95255631e-01
1.09100008e+00 -1.98783413e-01 -7.20950365e-01 -2.27591604e-01
-2.37097889e-01 4.42955464e-01 6.29695892e-01 2.17077620e-02
-1.46593297e+00 1.17209864e+00 5.53105164e+00 9.56656516e-01
-9.16137815e-01 3.20848554e-01 1.86118215e-01 5.16198277e-01
-1.80763349e-01 -8.38238597e-02 -1.27088583e+00 6.52288258e-01
1.25966370e+00 -3.62878084e-01 9.17973220e-02 1.13875699e+00
-2.02657774e-01 4.32799101e-01 -1.07703638e+00 6.68030977e-01
-3.68877441e-01 -7.96164572e-01 -7.30672628e-02 -5.56348078e-02
6.82968378e-01 4.65037137e-01 -3.97939920e-01 1.07008171e+00
8.76571953e-01 -6.47645533e-01 1.49144441e-01 5.84144443e-02
9.50066507e-01 -7.06658840e-01 1.03469741e+00 5.10616362e-01
-1.22567058e+00 8.48282687e-03 -6.69747114e-01 3.28976989e-01
1.98157936e-01 5.39529920e-01 -9.08751369e-01 1.03419352e+00
8.53733897e-01 5.65708280e-01 -5.07298410e-01 9.60007012e-01
-2.94660062e-01 6.44043922e-01 -4.70722288e-01 2.72039115e-01
4.64231223e-02 9.65378731e-02 4.56046313e-01 1.70854414e+00
5.77634871e-01 2.15403780e-01 9.33804139e-02 6.85824692e-01
-4.80801672e-01 2.09605366e-01 -5.64249456e-01 -2.86399782e-01
1.03761804e+00 1.33534598e+00 -4.31363225e-01 -3.41237396e-01
-7.34515727e-01 1.12359369e+00 5.39914489e-01 3.77085507e-01
-9.13830698e-01 -8.39950144e-01 1.01923752e+00 -1.54946238e-01
4.45481658e-01 -1.77435577e-01 -2.22994968e-01 -1.56559229e+00
-2.41095304e-01 -9.20393586e-01 9.18943822e-01 -3.21451932e-01
-2.00640154e+00 8.70145321e-01 5.04466519e-02 -1.37812316e+00
-2.12524354e-01 -5.84933043e-01 -5.56349933e-01 9.40396786e-01
-2.04549623e+00 -1.12053323e+00 -2.67279506e-01 7.66192377e-01
4.05860454e-01 -2.25319818e-01 1.15949500e+00 8.85291219e-01
-5.87046504e-01 1.03715909e+00 3.44004124e-01 7.41497993e-01
1.41606128e+00 -1.26494646e+00 7.05156028e-01 8.97818625e-01
-1.18188523e-01 8.02529752e-01 4.42005128e-01 -7.40418553e-01
-7.89589286e-01 -1.27721214e+00 1.10665715e+00 -7.81290054e-01
6.05761647e-01 -2.82116354e-01 -1.18511164e+00 7.92490602e-01
2.73931801e-01 5.20140566e-02 9.69412506e-01 3.57703775e-01
-4.53523576e-01 8.72276574e-02 -1.45912278e+00 5.39986134e-01
1.16503906e+00 -3.42374504e-01 -1.18623543e+00 -1.30122572e-01
7.67607629e-01 -3.57423186e-01 -1.34989166e+00 5.77716112e-01
7.72245675e-02 -5.45346618e-01 1.04164004e+00 -8.76537919e-01
3.57488006e-01 -3.12616616e-01 -1.73902035e-01 -1.81443954e+00
-3.76285225e-01 7.71748787e-03 2.77010590e-01 2.09522009e+00
5.61247289e-01 -7.00708032e-01 4.93915141e-01 5.63837409e-01
-2.65211970e-01 -2.24812478e-02 -7.91305244e-01 -1.16347539e+00
6.79619730e-01 -1.42476454e-01 8.35085690e-01 1.55932915e+00
-6.09008372e-02 5.17935455e-01 -7.52866194e-02 3.70154887e-01
1.67249560e-01 -2.65640289e-01 6.92684531e-01 -1.34066033e+00
1.10372059e-01 -1.44975111e-01 -9.95445922e-02 -9.64375138e-01
4.25069183e-01 -1.01782143e+00 1.74297746e-02 -1.52792108e+00
-1.28676863e-02 -7.24343181e-01 -6.10622644e-01 8.46509218e-01
-4.67011154e-01 -1.28546640e-01 1.80777952e-01 2.28824481e-01
-7.48932183e-01 5.93672276e-01 1.08688366e+00 1.25831068e-01
-2.72395700e-01 -2.46128440e-01 -9.97970283e-01 4.86110598e-01
8.37332845e-01 -7.79636860e-01 -1.87586173e-01 -5.50454140e-01
-2.13654533e-01 -2.58111656e-01 1.69232078e-02 -9.91331816e-01
3.15960079e-01 -6.87815323e-02 3.65644455e-01 -3.28705549e-01
-7.36321732e-02 -9.77232158e-01 -5.38608059e-02 -6.72891736e-02
-1.33095592e-01 6.38411716e-02 5.69738150e-01 3.63688201e-01
-4.90095645e-01 -4.34862554e-01 8.81308615e-01 -1.65688455e-01
-1.30084705e+00 2.03593075e-01 1.76443011e-02 8.55833173e-01
9.11636055e-01 -1.31705031e-01 -4.15910423e-01 1.45382479e-01
-6.27033532e-01 3.99027944e-01 6.30608261e-01 5.27271986e-01
-3.64687294e-02 -1.48482788e+00 -7.86977708e-01 -1.19473130e-01
4.78797257e-01 -5.55634387e-02 3.33460331e-01 2.57703185e-01
-7.75660947e-02 3.17570478e-01 -4.60743576e-01 -2.02021167e-01
-8.84820879e-01 4.93050635e-01 3.98445666e-01 -8.81679118e-01
-3.79969388e-01 6.21331036e-01 2.41525739e-01 -1.29944479e+00
-1.37089431e-01 -1.66287273e-01 -4.87850636e-01 6.48164824e-02
3.76917779e-01 3.77691448e-01 1.55969799e-01 -5.09389162e-01
-5.91818988e-01 2.37001970e-01 -6.55720904e-02 8.63811672e-02
1.51476550e+00 -1.97999239e-01 3.98263097e-01 1.84934251e-02
9.96794343e-01 2.68624425e-01 -1.09813988e+00 -5.49199164e-01
5.07954657e-01 -2.60914285e-02 -3.36715549e-01 -1.38282228e+00
-6.76773190e-01 6.46961153e-01 5.03949821e-01 -2.61588600e-02
1.30418456e+00 -6.58763051e-02 9.62797344e-01 5.12485564e-01
4.21843201e-01 -1.22337461e+00 -3.45181048e-01 1.05523467e+00
4.00961399e-01 -1.32157040e+00 -4.62263495e-01 -2.05406412e-01
-8.67838085e-01 9.83047187e-01 1.01500750e+00 -2.94513460e-02
6.04711115e-01 2.50784427e-01 3.02586764e-01 1.43974990e-01
-4.91975397e-01 -5.26956975e-01 1.45006195e-01 8.60502541e-01
7.33996868e-01 -2.06280034e-02 -5.49019396e-01 1.38956165e+00
-4.45658378e-02 9.35874358e-02 2.02781230e-01 6.84642971e-01
-1.29870117e-01 -1.74051309e+00 -2.61254668e-01 -8.45020115e-02
-6.08221233e-01 -2.97809899e-01 -4.02391136e-01 1.13051021e+00
2.94862121e-01 9.11859155e-01 -3.39940190e-01 -2.63281047e-01
9.96662140e-01 4.77565020e-01 7.91344717e-02 -8.40498388e-01
-7.63437867e-01 -5.16732931e-01 3.72360438e-01 -1.42650604e-01
-3.07976007e-01 -4.25472140e-01 -1.41740727e+00 -3.20778638e-01
-2.42003039e-01 4.69562113e-01 4.25280362e-01 7.79511988e-01
7.95208275e-01 3.20518941e-01 3.60712469e-01 -4.79146063e-01
-6.22961521e-01 -1.20022607e+00 -4.40691710e-01 7.64855206e-01
-1.14174880e-01 -8.38517904e-01 -3.03257614e-01 -5.67679526e-03] | [9.784720420837402, 9.516398429870605] |
be76be30-8515-4c29-8d2c-21809b85f19c | context-based-emotion-recognition-using | 2003.13401 | null | https://arxiv.org/abs/2003.13401v1 | https://arxiv.org/pdf/2003.13401v1.pdf | Context Based Emotion Recognition using EMOTIC Dataset | In our everyday lives and social interactions we often try to perceive the emotional states of people. There has been a lot of research in providing machines with a similar capacity of recognizing emotions. From a computer vision perspective, most of the previous efforts have been focusing in analyzing the facial expressions and, in some cases, also the body pose. Some of these methods work remarkably well in specific settings. However, their performance is limited in natural, unconstrained environments. Psychological studies show that the scene context, in addition to facial expression and body pose, provides important information to our perception of people's emotions. However, the processing of the context for automatic emotion recognition has not been explored in depth, partly due to the lack of proper data. In this paper we present EMOTIC, a dataset of images of people in a diverse set of natural situations, annotated with their apparent emotion. The EMOTIC dataset combines two different types of emotion representation: (1) a set of 26 discrete categories, and (2) the continuous dimensions Valence, Arousal, and Dominance. We also present a detailed statistical and algorithmic analysis of the dataset along with annotators' agreement analysis. Using the EMOTIC dataset we train different CNN models for emotion recognition, combining the information of the bounding box containing the person with the contextual information extracted from the scene. Our results show how scene context provides important information to automatically recognize emotional states and motivate further research in this direction. Dataset and code is open-sourced and available at: https://github.com/rkosti/emotic and link for the peer-reviewed published article: https://ieeexplore.ieee.org/document/8713881 | ['Ronak Kosti', 'Agata Lapedriza', 'Adria Recasens', 'Jose M. Alvarez'] | 2020-03-30 | null | null | null | null | ['emotion-recognition-in-context'] | ['natural-language-processing'] | [-1.78749070e-01 -2.54402697e-01 6.26382679e-02 -8.94365609e-01
6.40863925e-02 -4.62551504e-01 3.82064193e-01 1.67876825e-01
-3.58299464e-01 4.31773365e-01 4.19658691e-01 4.49619859e-01
2.45473623e-01 -4.39210802e-01 6.43480048e-02 -5.80545962e-01
2.12526321e-02 5.70636988e-02 -3.60687852e-01 -4.99623835e-01
1.05939722e-02 5.02440453e-01 -1.75339293e+00 3.72997522e-01
4.75255072e-01 1.49171102e+00 -3.89792562e-01 5.68404138e-01
-1.83261514e-01 5.15902281e-01 -5.53741693e-01 -5.43596387e-01
1.06615365e-01 -6.06998980e-01 -5.99809706e-01 2.10167617e-01
3.09322745e-01 -5.51789887e-02 -3.37327980e-02 9.58781660e-01
6.88822925e-01 1.53107509e-01 3.58614057e-01 -1.35194016e+00
-2.58957088e-01 -9.69629586e-02 -3.38133723e-01 3.76056544e-02
7.35745013e-01 5.65834939e-02 6.82645440e-01 -7.52693594e-01
5.58788776e-01 1.09367883e+00 4.92433608e-01 6.92788124e-01
-7.40428209e-01 -5.69683075e-01 6.22755252e-02 2.82249898e-01
-1.30678713e+00 -5.36389232e-01 1.12349296e+00 -4.37542826e-01
6.91345155e-01 4.13513005e-01 9.85234678e-01 1.33309472e+00
-2.33399887e-02 6.59556150e-01 1.50367999e+00 -5.44588387e-01
1.90343797e-01 6.09082580e-01 2.74227142e-01 6.45875633e-01
-1.50880307e-01 -2.00808704e-01 -4.72476035e-01 -1.91614702e-01
3.91290665e-01 7.73003558e-03 -1.88698888e-01 -2.45589674e-01
-6.34680629e-01 6.69255733e-01 3.90514761e-01 5.61646283e-01
-4.76044446e-01 -1.75301451e-02 7.75638521e-01 3.23706150e-01
7.11393058e-01 2.19060972e-01 -3.34688425e-01 -4.47145462e-01
-4.20817912e-01 7.22598145e-03 1.02687347e+00 4.47240233e-01
6.91034079e-01 -6.98684677e-02 5.68116829e-02 1.19699514e+00
1.68930277e-01 2.87622362e-01 2.93596208e-01 -7.52214015e-01
-5.51677682e-02 6.32770717e-01 1.48273826e-01 -1.47742176e+00
-6.64452255e-01 4.69251201e-02 -7.64396369e-01 2.38551810e-01
4.43254888e-01 -4.94016677e-01 -4.79136765e-01 1.86206210e+00
6.26207650e-01 -1.00639723e-01 -1.26841962e-01 1.33644450e+00
1.09917319e+00 3.94150585e-01 2.71331042e-01 -2.04647612e-02
1.73550606e+00 -4.90537494e-01 -1.08575559e+00 -2.31326923e-01
3.62785399e-01 -7.89776385e-01 9.49511051e-01 5.72373807e-01
-8.91253829e-01 -6.23839736e-01 -7.93637216e-01 4.55809496e-02
-7.87297368e-01 3.63278806e-01 1.01728225e+00 8.88408840e-01
-9.79654789e-01 3.90954316e-01 -5.42586029e-01 -7.91449904e-01
2.69182533e-01 2.86132276e-01 -6.25142813e-01 2.73195803e-01
-1.40406036e+00 1.01978016e+00 -5.58881983e-02 4.40358400e-01
-2.63762146e-01 -1.12003349e-02 -1.09340870e+00 -1.87034413e-01
2.83972859e-01 -2.85945475e-01 1.09780872e+00 -1.73006403e+00
-1.59360397e+00 1.21474600e+00 -2.05035880e-01 4.32215109e-02
3.21803719e-01 -3.52793902e-01 -6.24078512e-01 3.06359559e-01
-2.95468032e-01 6.27306461e-01 6.76477373e-01 -1.12959170e+00
-2.28168041e-01 -7.12875664e-01 5.70663475e-02 2.63515800e-01
-6.16201580e-01 5.82351327e-01 -5.10604680e-01 -3.17111224e-01
-1.95219383e-01 -9.92249608e-01 -1.22616641e-01 8.08007792e-02
-1.29752979e-01 -2.17055693e-01 7.00282335e-01 -6.14573836e-01
1.07301283e+00 -2.28513312e+00 4.18982608e-03 2.43514061e-01
6.51246086e-02 2.17477188e-01 1.59208477e-01 5.19630015e-01
-2.09931180e-01 8.11195970e-02 4.87667285e-02 -5.63135087e-01
1.57351688e-01 1.58021122e-01 3.87143977e-02 7.12203741e-01
6.63599968e-02 7.60845363e-01 -7.43300855e-01 -5.53693295e-01
3.85677844e-01 7.56475866e-01 -1.73364267e-01 2.77112067e-01
3.23795259e-01 5.75511098e-01 -4.22580034e-01 6.63306653e-01
5.21973550e-01 2.93423504e-01 2.76491433e-01 -3.18998158e-01
9.93149504e-02 -8.34751427e-02 -1.26173127e+00 1.35147405e+00
-4.38525319e-01 8.25215995e-01 4.67487484e-01 -7.98703313e-01
1.02288020e+00 6.11787617e-01 5.92910886e-01 -6.25379622e-01
7.67066360e-01 -3.68389487e-02 -5.39260954e-02 -9.92750645e-01
4.45407450e-01 -3.36914420e-01 -3.17353636e-01 2.07813442e-01
2.05600411e-02 -9.78486836e-02 2.02815294e-01 -3.30264233e-02
6.32214189e-01 1.03793137e-01 5.06133795e-01 -6.97163045e-02
4.93356347e-01 -2.68983722e-01 5.71464002e-01 2.62539834e-01
-7.48319507e-01 4.05360609e-01 5.10395646e-01 -6.98152721e-01
-7.38945365e-01 -5.57558119e-01 -1.95865586e-01 1.04103518e+00
1.60453856e-01 -4.83705878e-01 -8.14458013e-01 -2.92313784e-01
-1.53526023e-01 4.41300511e-01 -9.67655540e-01 -2.80987360e-02
-1.78523690e-01 -4.78478462e-01 4.88512456e-01 5.64401925e-01
4.85089213e-01 -1.46275246e+00 -9.64899778e-01 -3.07069063e-01
-2.97613531e-01 -1.13152766e+00 -7.32159987e-02 1.32120579e-01
-6.18114531e-01 -9.16849971e-01 -2.72468269e-01 -3.76988828e-01
6.13970995e-01 -9.04428288e-02 1.09395897e+00 1.55986294e-01
-5.45025706e-01 8.26883793e-01 -6.13168478e-01 -6.41602099e-01
-2.36737859e-02 -4.83237147e-01 2.55479097e-01 3.19062650e-01
9.19787467e-01 -2.79336870e-01 -5.93563795e-01 3.57020229e-01
-6.93038583e-01 -1.73874632e-01 1.30345553e-01 4.90624130e-01
1.73961788e-01 -1.47762880e-01 3.35183084e-01 -4.95692641e-01
7.57682383e-01 -4.61345464e-01 1.53827500e-02 -8.99312943e-02
3.27915885e-02 -6.56164229e-01 4.99315977e-01 -3.74093741e-01
-1.29728842e+00 7.02243000e-02 -2.50508726e-01 -4.15383309e-01
-1.01198041e+00 3.39610308e-01 -2.15057909e-01 9.25288275e-02
5.89978039e-01 -2.98478723e-01 1.57096848e-01 -2.15175867e-01
1.38724148e-01 8.01889181e-01 3.21104079e-01 -6.31570399e-01
1.53222397e-01 6.62762403e-01 -1.33441344e-01 -1.03097653e+00
-7.83607602e-01 -6.21389389e-01 -8.11280787e-01 -8.49761188e-01
1.03105080e+00 -7.16789901e-01 -8.28491569e-01 5.52376449e-01
-8.45216334e-01 -2.88056940e-01 -8.11780151e-03 4.91490066e-01
-4.39440906e-01 4.61802840e-01 -6.76117182e-01 -1.17544699e+00
-3.66389662e-01 -7.93887138e-01 9.61695611e-01 3.90157342e-01
-8.12961161e-01 -9.87169206e-01 -4.15987819e-02 4.70992655e-01
3.56178671e-01 6.55111969e-01 3.03169638e-01 -4.97355014e-01
4.41364855e-01 -5.17266333e-01 4.41518649e-02 4.21684980e-01
1.38157740e-01 1.01651981e-01 -1.21703029e+00 3.34208310e-02
1.43900111e-01 -7.36741960e-01 5.42070985e-01 3.58707219e-01
1.06232798e+00 7.18563423e-02 -1.26017602e-02 2.88293481e-01
1.17885947e+00 3.18178415e-01 7.07444191e-01 8.54764208e-02
4.88353252e-01 1.24222410e+00 7.43113756e-01 7.02593505e-01
1.73760563e-01 7.99220920e-01 5.05936265e-01 -2.85368681e-01
3.29840034e-01 2.10906669e-01 3.30894738e-01 3.36328506e-01
-4.21947539e-01 -1.01848736e-01 -7.85596788e-01 3.35988432e-01
-1.68064511e+00 -1.11514366e+00 -2.78705955e-01 1.86588418e+00
6.26888633e-01 -3.50722700e-01 2.94594914e-01 1.18127368e-01
6.99853301e-01 2.09998727e-01 -2.92151660e-01 -1.12371063e+00
2.23500188e-03 7.91178346e-02 -2.11349443e-01 2.30748892e-01
-1.34164286e+00 8.30249548e-01 5.35645437e+00 3.36950392e-01
-1.32337606e+00 -4.79132570e-02 8.40710223e-01 -1.80011302e-01
2.84627706e-01 -4.39164817e-01 -3.42152745e-01 3.30947340e-01
7.61099994e-01 3.15493047e-01 3.52657795e-01 1.10337555e+00
3.60780090e-01 -3.17815691e-01 -8.47530603e-01 1.34826398e+00
4.57698882e-01 -3.14831346e-01 -6.73885882e-01 -1.22603320e-01
2.53237009e-01 -3.48502189e-01 -4.00669798e-02 2.43329227e-01
-2.78577477e-01 -1.00519669e+00 5.93196332e-01 5.89112461e-01
6.62502825e-01 -6.81509674e-01 9.47048724e-01 -1.91355739e-02
-1.00291324e+00 -1.27595842e-01 -1.00907221e-01 -5.78372717e-01
1.72672987e-01 5.44842064e-01 -4.53310192e-01 2.27942079e-01
1.10122907e+00 5.63413143e-01 -4.66330826e-01 6.24063313e-01
-3.51201475e-01 4.96691287e-01 -3.44320387e-01 -3.34064692e-01
1.94943473e-02 -4.19395655e-01 4.22101349e-01 1.59676325e+00
1.44300520e-01 5.10969460e-01 6.56701922e-02 6.17060065e-01
2.38674268e-01 4.39186931e-01 -5.83962798e-01 -2.75433492e-02
2.74233408e-02 1.88944888e+00 -8.73213410e-01 -3.24551016e-01
-4.19950396e-01 1.10069394e+00 2.25532621e-01 1.54281259e-01
-8.53515804e-01 -3.37424487e-01 8.79926801e-01 -1.81411326e-01
-1.21157259e-01 -1.14187270e-01 -3.32669139e-01 -1.09538281e+00
1.02352545e-01 -8.73053670e-01 4.67027277e-01 -9.92915273e-01
-1.37862670e+00 8.39847565e-01 5.71979284e-02 -9.98651683e-01
-1.47944897e-01 -9.33365464e-01 -7.18241930e-01 7.26144075e-01
-9.03193593e-01 -9.61887538e-01 -9.11277115e-01 5.88535428e-01
3.55470479e-01 1.91040859e-01 9.76148665e-01 2.75703818e-01
-6.57082498e-01 3.48823339e-01 -4.80318666e-01 3.64916831e-01
9.68575656e-01 -1.17450929e+00 -2.79454499e-01 2.71316707e-01
-1.35025084e-01 6.39787316e-01 8.68209898e-01 -4.12595510e-01
-1.22464812e+00 -5.08312702e-01 7.15858936e-01 -4.13994312e-01
4.63103682e-01 -5.15607476e-01 -6.84546828e-01 3.62390399e-01
5.08679330e-01 -2.63360627e-02 1.17348850e+00 4.02779549e-01
-1.88507572e-01 1.63315367e-02 -1.36324918e+00 6.62973404e-01
7.81604946e-01 -4.28964704e-01 -4.82915610e-01 -6.64880797e-02
-1.76941857e-01 -4.02017832e-01 -9.07226562e-01 3.66891235e-01
8.86121154e-01 -1.28582251e+00 6.58898473e-01 -5.49969137e-01
6.59536958e-01 -9.20499936e-02 -1.75150812e-01 -1.28092766e+00
-1.67649344e-01 -7.59335831e-02 1.90775067e-01 1.24236834e+00
-8.56421664e-02 -6.72079146e-01 5.76401055e-01 9.77824450e-01
4.47687916e-02 -1.04739380e+00 -6.24116242e-01 -3.03807467e-01
-2.81346142e-01 -7.08232582e-01 2.90583789e-01 1.09456038e+00
4.54424888e-01 2.44129121e-01 -5.61388612e-01 -1.92790896e-01
8.06877017e-02 1.24086872e-01 8.69478643e-01 -1.14856160e+00
4.07899320e-02 -5.27016997e-01 -6.78227603e-01 -3.16212565e-01
2.09517673e-01 -4.78785813e-01 -6.43483549e-02 -1.31912816e+00
2.00264648e-01 -9.63190123e-02 -2.00370207e-01 6.77796125e-01
-7.83245042e-02 5.50602853e-01 2.85293609e-01 -1.39687881e-01
-7.26976156e-01 5.65714180e-01 9.73039746e-01 1.94697887e-01
-2.07862616e-01 -2.75917172e-01 -6.08896375e-01 1.02078998e+00
1.22670078e+00 5.72926812e-02 -1.85977072e-01 -1.37606645e-02
2.80508488e-01 -2.09024146e-01 4.73019332e-01 -9.48450208e-01
-7.16125444e-02 -1.82180613e-01 8.03897977e-01 -2.40374207e-01
9.63905573e-01 -9.27021205e-01 5.99470362e-02 1.88254684e-01
-1.91379443e-01 -9.47698653e-02 3.40428799e-01 1.29766643e-01
-3.90414685e-01 -2.10207775e-01 7.85680652e-01 -2.57643044e-01
-9.23452616e-01 2.12675240e-03 -5.04092455e-01 -1.95640758e-01
9.53894734e-01 -4.12319392e-01 -9.15347859e-02 -8.86170745e-01
-9.93672848e-01 1.90315381e-01 5.08418322e-01 6.43667400e-01
5.00557601e-01 -1.26632476e+00 -3.91746551e-01 4.24807854e-02
2.35415548e-01 -6.92133605e-01 5.64541996e-01 1.03828192e+00
-2.03719690e-01 1.20070092e-01 -5.87054312e-01 -4.43463862e-01
-1.71075070e+00 5.54264605e-01 4.54341859e-01 2.11538956e-01
-1.82770640e-01 6.47758186e-01 1.84978381e-01 -2.92571813e-01
8.76946077e-02 3.46308500e-02 -6.13049567e-01 4.40432191e-01
5.78666747e-01 3.08301419e-01 -1.58814475e-01 -1.29845905e+00
-4.25618052e-01 5.93489170e-01 2.24699408e-01 -1.01984078e-02
1.15507650e+00 -2.59447694e-01 -4.14906114e-01 5.20233393e-01
1.19048703e+00 1.22450374e-01 -7.44140804e-01 1.15521580e-01
-2.98111826e-01 -5.80869377e-01 -1.52947471e-01 -8.58048081e-01
-1.14683867e+00 9.67582643e-01 8.51885796e-01 2.86446810e-01
1.53580928e+00 1.58651394e-03 5.56266367e-01 2.11850882e-01
3.92188728e-01 -1.46224928e+00 1.48375437e-01 4.80320275e-01
9.81413186e-01 -1.44861817e+00 -2.81528905e-02 -5.39262712e-01
-1.17014956e+00 1.17775476e+00 8.51256013e-01 5.42762876e-02
6.99942768e-01 1.47494167e-01 5.54986894e-01 -4.95972037e-01
-4.27843809e-01 -4.82954681e-01 3.17798227e-01 4.58158523e-01
7.54855156e-01 2.57672012e-01 -3.96153510e-01 9.27941561e-01
-2.88688540e-01 -2.49206219e-02 1.63639009e-01 8.03912520e-01
-2.07041621e-01 -8.94155860e-01 -6.44209981e-01 2.77216762e-01
-6.64997458e-01 3.29794556e-01 -9.92428720e-01 6.37146235e-01
4.21526730e-01 1.28843307e+00 8.47634450e-02 -4.74449545e-01
3.59559178e-01 3.18767011e-01 4.08862203e-01 -4.09343690e-01
-7.48962641e-01 -3.44714783e-02 3.97513002e-01 -6.67667031e-01
-6.08658016e-01 -6.74351454e-01 -1.20251727e+00 -2.40033999e-01
3.33565176e-02 8.21767896e-02 8.59152615e-01 7.40541220e-01
3.12862903e-01 1.52683273e-01 5.01479149e-01 -1.17202497e+00
1.78623304e-01 -1.03417563e+00 -6.50770724e-01 9.72368181e-01
-5.19766994e-02 -8.20053041e-01 -3.68512958e-01 -9.96851996e-02] | [13.488564491271973, 2.218980550765991] |
81406312-fd7a-431e-9d0a-f6ec2b000a87 | frequency-aware-interface-dynamics-with | null | null | https://openreview.net/forum?id=uMDbGsVjCS4 | https://openreview.net/pdf?id=uMDbGsVjCS4 | Frequency-aware Interface Dynamics with Generative Adversarial Networks | We present a new method for reconstructing and refining complex surfaces based on physical simulations. Taking a roughly approximated simulation as input, our method infers corresponding spatial details while taking into account how they evolve over time. We consider this problem in terms of spatial and temporal frequencies, and leverage generative adversarial networks to learn the desired spatio-temporal signal for the surface dynamics. Furthermore, we investigate the possibility to train our network in an unsupervised manner, i.e. without predefined training pairs. We highlight the capabilities of our method with a set of synthetic wave function tests and complex 3D dynamics of elasto-plastic materials. | ['Nils Thuerey', 'Jan Bender', 'Tassilo Kugelstadt', 'Lukas Prantl'] | 2021-01-01 | null | null | null | null | ['physical-simulations'] | ['miscellaneous'] | [ 2.88404256e-01 2.36607015e-01 4.88565177e-01 1.50495037e-01
-6.46912158e-01 -8.08608234e-01 8.02562356e-01 -3.21235180e-01
-2.38470688e-01 9.18815374e-01 -8.77987370e-02 -2.11377576e-01
-1.91178054e-01 -1.23522782e+00 -1.06095719e+00 -8.20245385e-01
-4.31416482e-01 5.64755380e-01 3.21311265e-01 -4.85631734e-01
-3.86317559e-02 9.35010731e-01 -1.16630352e+00 2.98136145e-01
6.98519766e-01 7.02900708e-01 -3.48209441e-01 1.05800784e+00
3.90838206e-01 4.11955625e-01 -7.40525723e-01 -3.74870718e-01
2.61147141e-01 -4.63317275e-01 -6.06656611e-01 1.82043597e-01
2.03515291e-01 -2.21816520e-03 -4.01736051e-01 8.61366391e-01
1.60714835e-01 2.71208018e-01 1.07662189e+00 -7.82105982e-01
-4.53429073e-01 9.98954102e-02 2.49548852e-02 -5.43201752e-02
3.51667494e-01 4.74212825e-01 4.91848707e-01 -7.40831673e-01
9.57826853e-01 9.49094474e-01 1.08419394e+00 4.98359561e-01
-1.76050770e+00 -1.91013649e-01 -1.66241407e-01 -4.64154333e-01
-1.17785251e+00 -2.62212425e-01 1.13614154e+00 -6.99008048e-01
5.14017761e-01 2.83050060e-01 8.44527602e-01 1.35394502e+00
5.28939307e-01 8.96552056e-02 1.45039630e+00 -3.16892385e-01
3.71573925e-01 -1.23091675e-01 -1.69073075e-01 7.72621632e-01
4.46723923e-02 5.55833995e-01 -8.90729874e-02 -4.16258723e-01
1.33213186e+00 -2.17653036e-01 -3.08049530e-01 -2.62677252e-01
-1.17230129e+00 5.81381619e-01 1.26017988e-01 3.29660684e-01
-5.06941974e-01 3.37109894e-01 -7.90397376e-02 3.65140676e-01
5.97001791e-01 7.32806563e-01 -2.57483810e-01 1.06979869e-01
-8.09040368e-01 6.20036721e-01 1.01814497e+00 3.41161311e-01
7.26715863e-01 2.13390261e-01 -1.42825739e-02 3.70052725e-01
2.69335192e-02 6.96444333e-01 1.07543850e-02 -1.10226965e+00
-1.05677232e-01 5.40633798e-02 4.23071593e-01 -9.64187264e-01
-1.82400450e-01 -2.97194898e-01 -1.05010569e+00 8.72841954e-01
6.94055438e-01 -3.55902463e-01 -1.20651972e+00 1.66266036e+00
3.77969503e-01 7.39438355e-01 1.59132525e-01 6.70909107e-01
2.95547485e-01 6.70954466e-01 -2.25533441e-01 -3.53474498e-01
8.54671419e-01 -5.04569888e-01 -5.04961193e-01 2.93137431e-01
-5.85296825e-02 -4.64174360e-01 9.47613895e-01 3.71513039e-01
-1.52892935e+00 -4.20569569e-01 -8.82295251e-01 4.01881754e-01
-3.49577487e-01 -2.86918491e-01 3.09733510e-01 2.89891213e-01
-1.14992309e+00 1.49324715e+00 -1.32203937e+00 1.04414515e-01
2.37862080e-01 2.49607041e-01 -2.31391966e-01 4.12723541e-01
-1.21677542e+00 5.27433336e-01 -1.13794520e-01 3.83659042e-02
-9.19785440e-01 -8.24358821e-01 -5.63102305e-01 -2.18912095e-01
-5.20338975e-02 -9.03174520e-01 8.87061834e-01 -8.42478991e-01
-1.92279828e+00 7.09871948e-01 5.98129891e-02 -2.63688922e-01
8.69372189e-01 1.46693945e-01 -4.84628022e-01 3.99891764e-01
-2.04313189e-01 7.48672634e-02 1.06937778e+00 -1.67112029e+00
1.55596524e-01 1.16591036e-01 9.32707936e-02 -3.08885306e-01
1.81472316e-01 -3.68433386e-01 -1.85657278e-01 -1.13417184e+00
8.28987136e-02 -1.08886659e+00 -5.25147676e-01 1.38605520e-01
-5.77059388e-01 4.75653172e-01 5.90058982e-01 -7.33919799e-01
7.90999532e-01 -1.80016792e+00 6.48118258e-01 6.50356472e-01
2.27596521e-01 5.57766259e-02 9.81510282e-02 5.80637991e-01
1.91831104e-02 3.58878732e-01 -9.39773977e-01 -4.57856476e-01
-4.00141776e-02 2.08228171e-01 -4.30519521e-01 4.60526854e-01
4.83472019e-01 1.03045726e+00 -9.55254614e-01 -1.80364162e-01
1.47939427e-02 7.62928188e-01 -7.77658224e-01 3.15906525e-01
-5.16698778e-01 1.11790550e+00 -4.72978979e-01 4.51017261e-01
5.72882473e-01 -2.86368191e-01 8.35969970e-02 -1.24357827e-01
-9.88742262e-02 3.68859177e-03 -1.04883492e+00 1.31763709e+00
-5.36279440e-01 4.36275661e-01 2.06209555e-01 -9.81371701e-01
6.25647902e-01 4.70646530e-01 5.73202729e-01 -3.57128769e-01
-1.02814130e-01 2.91044742e-01 -1.43493712e-01 -5.36663890e-01
1.85646892e-01 -6.70726478e-01 -1.46416545e-01 5.07941365e-01
-1.88007876e-01 -2.61921823e-01 -4.00226623e-01 -2.36561760e-01
1.42665267e+00 3.26743543e-01 -1.63346574e-01 -3.10494155e-01
2.17172071e-01 7.83474967e-02 1.33906901e-01 5.87298036e-01
3.62315238e-01 9.53770936e-01 5.75203061e-01 -5.45803964e-01
-1.37357187e+00 -1.45366538e+00 -2.36353114e-01 1.44403338e-01
1.41713604e-01 1.14662677e-01 -8.14828038e-01 -4.60073173e-01
2.30253279e-01 4.62546557e-01 -1.20824134e+00 -9.95166898e-02
-1.01361930e+00 -6.17148280e-01 5.95474243e-01 1.57694563e-01
1.24695174e-01 -1.37887728e+00 -6.07698858e-01 3.19971830e-01
3.44749063e-01 -8.84505212e-01 -1.09073170e-01 -6.74241483e-02
-8.14020634e-01 -1.05278504e+00 -8.78723264e-01 -5.83197713e-01
7.50862360e-01 -4.79952574e-01 1.26069415e+00 6.11391887e-02
-2.67348140e-01 4.07158732e-01 8.21853727e-02 1.47667304e-01
-9.53649163e-01 -2.04749629e-01 -3.22473384e-02 1.98479757e-01
-7.65861690e-01 -1.35755622e+00 -7.21668363e-01 2.41497457e-01
-1.11002779e+00 1.35590089e-02 2.66685247e-01 8.39177489e-01
8.41222107e-01 2.44548634e-01 4.96899486e-01 -1.20349360e+00
6.58641875e-01 -6.06971145e-01 -5.33313036e-01 1.12588778e-01
-1.71543419e-01 2.26446673e-01 8.03623497e-01 -8.68457735e-01
-9.49328482e-01 -6.04091026e-02 -5.03834009e-01 -7.65704274e-01
-1.20354652e-01 3.53543043e-01 7.11041987e-02 -3.84993464e-01
7.86838472e-01 4.29837972e-01 5.23994863e-02 -5.24597049e-01
1.43140122e-01 -1.26676321e-01 7.81443894e-01 -9.13714349e-01
1.29304981e+00 8.24922442e-01 3.43965352e-01 -7.55245030e-01
-3.25596035e-01 2.44651347e-01 -6.63981378e-01 -2.51137525e-01
6.30698800e-01 -3.14553350e-01 -8.44424844e-01 8.43313038e-01
-1.35893691e+00 -8.98604572e-01 -7.12904036e-01 1.30746484e-01
-8.45910430e-01 1.91711597e-02 -7.43277669e-01 -6.12750471e-01
2.77548637e-02 -8.99022758e-01 1.13537633e+00 -1.33161634e-01
-1.82270050e-01 -1.50311017e+00 4.63085353e-01 -4.23417747e-01
4.88952279e-01 1.33057082e+00 9.04667139e-01 -7.29082078e-02
-7.59085655e-01 5.49875274e-02 2.74654806e-01 1.65110871e-01
2.87771404e-01 2.87291199e-01 -8.26192677e-01 -2.21593335e-01
1.81958348e-01 1.70142904e-01 8.04962099e-01 3.73147041e-01
9.77292895e-01 -6.62826478e-01 -3.60236585e-01 9.65766430e-01
1.62305081e+00 -4.74624485e-02 7.43534505e-01 -1.24517165e-01
7.22287714e-01 6.13722444e-01 -2.70302431e-03 2.01694250e-01
-1.66245550e-01 7.08785534e-01 2.91177809e-01 -1.55778602e-01
-2.13942453e-01 -2.82205820e-01 1.21861458e-01 5.21282732e-01
-5.14026701e-01 -2.40461290e-01 -9.37917292e-01 5.10295808e-01
-1.41530383e+00 -9.53814864e-01 2.15333909e-01 2.08087158e+00
9.25154328e-01 3.03119928e-01 1.04020134e-01 5.89357093e-02
4.92025107e-01 1.24465182e-01 -6.81267083e-01 -1.33534074e-01
-2.07119927e-01 9.01361883e-01 3.11697036e-01 9.17863488e-01
-8.39201689e-01 5.76701462e-01 7.06148386e+00 3.73575240e-01
-1.48209929e+00 8.40029269e-02 4.55824226e-01 1.41261339e-01
-9.95100379e-01 -2.25632057e-01 -1.26672879e-01 5.21852732e-01
9.63970840e-01 -2.67593097e-02 6.40770257e-01 1.02293685e-01
2.48999391e-02 3.69494200e-01 -9.22308862e-01 3.22649598e-01
-4.48848605e-01 -1.88483167e+00 1.45303339e-01 1.02359504e-01
9.14896548e-01 -2.30602413e-01 1.18111618e-01 -6.89210892e-02
4.50929046e-01 -1.37914705e+00 7.23024189e-01 1.17538786e+00
1.06579149e+00 -5.50403774e-01 2.03351244e-01 3.91670108e-01
-1.01778924e+00 5.48550427e-01 2.49822319e-01 1.28183350e-01
3.55433881e-01 5.13575733e-01 -4.84454691e-01 4.76894617e-01
4.19763297e-01 5.00098050e-01 -9.24059078e-02 7.25563765e-01
-6.91221207e-02 8.54817808e-01 -5.50779104e-01 2.70378649e-01
-5.28029278e-02 -4.77680922e-01 1.02338350e+00 1.04729497e+00
4.59447831e-01 4.51186508e-01 -8.82449225e-02 1.16280067e+00
-7.69649744e-02 -3.55715394e-01 -8.24263513e-01 -1.05884578e-02
2.15829998e-01 7.91683853e-01 -8.47200334e-01 3.57540362e-02
1.12329554e-02 9.25947428e-01 2.51200199e-01 7.76701987e-01
-6.95547938e-01 -1.85917988e-01 7.73816526e-01 6.21828258e-01
2.50346303e-01 -5.96295118e-01 -1.06391050e-01 -9.95310605e-01
1.05572253e-01 -5.50508261e-01 -1.53878331e-01 -6.51439011e-01
-1.24079013e+00 8.13153684e-01 -3.80085483e-02 -1.08579552e+00
-2.72329658e-01 -5.36740780e-01 -9.22310114e-01 1.02463424e+00
-1.40306842e+00 -9.90264177e-01 -1.72083035e-01 4.23933625e-01
-2.65641212e-02 1.91537172e-01 8.26783299e-01 -3.84949124e-03
-1.28945351e-01 3.02461594e-01 6.27000034e-02 4.10205945e-02
1.92338049e-01 -1.27354288e+00 9.63037729e-01 6.38147891e-01
-6.70339614e-02 4.99145985e-01 1.03117859e+00 -7.30543196e-01
-1.30927730e+00 -1.07412910e+00 3.49478900e-01 -6.33978009e-01
6.84939802e-01 -3.69446456e-01 -1.39032638e+00 6.09704852e-01
1.76578704e-02 4.39903796e-01 1.91307068e-01 -4.24147576e-01
8.44408050e-02 3.32700878e-01 -1.39873934e+00 6.14799738e-01
1.14880705e+00 -5.53094804e-01 -3.52190763e-01 3.37013721e-01
9.51692104e-01 -7.11476028e-01 -1.12138402e+00 6.51454687e-01
6.22170985e-01 -8.89337063e-01 1.15934491e+00 -6.84133112e-01
6.75944388e-01 -3.18069160e-01 1.01162836e-01 -1.44254434e+00
-5.87160029e-02 -1.16847336e+00 -2.08882794e-01 7.78816104e-01
3.02580923e-01 -8.17938924e-01 9.61206615e-01 1.53978765e-01
-7.46628717e-02 -1.06053507e+00 -1.02609193e+00 -6.40891552e-01
4.11080956e-01 -2.91766137e-01 5.23835361e-01 9.93923664e-01
-5.97027242e-01 -2.92013019e-01 -2.29361430e-01 5.88283479e-01
6.48027599e-01 1.67915434e-01 4.57101434e-01 -1.10571885e+00
-6.18875146e-01 -2.57792145e-01 -1.50933012e-01 -7.21104383e-01
3.11790615e-01 -5.64541161e-01 7.15188831e-02 -1.05400527e+00
-4.06567991e-01 -6.92520916e-01 -1.13960959e-01 -3.78448628e-02
-1.55393668e-02 5.26909232e-01 -2.63039827e-01 3.72266948e-01
1.24967970e-01 5.56411684e-01 1.66162109e+00 2.04604883e-02
-2.53696263e-01 2.94237822e-01 -8.69479254e-02 6.47570014e-01
6.97898865e-01 -4.52739298e-01 -1.90446585e-01 5.76624821e-04
3.48716438e-01 5.27151406e-01 9.59760129e-01 -1.02171791e+00
7.81211704e-02 -2.93543339e-01 2.60771692e-01 -8.65520015e-02
4.99323756e-01 -7.11250722e-01 7.96536386e-01 3.66116107e-01
-2.89178640e-01 6.44057170e-02 4.70527947e-01 6.94713771e-01
-2.05357060e-01 7.83166885e-02 9.72586155e-01 -7.58270472e-02
-2.01841053e-02 3.29543442e-01 -3.15045685e-01 1.68241009e-01
8.52865160e-01 -1.93974823e-01 1.92827713e-02 -2.77285010e-01
-1.28328109e+00 -3.01325589e-01 9.99489009e-01 -1.77209005e-01
5.11177659e-01 -1.41053927e+00 -6.65251613e-01 4.42354470e-01
-3.37926596e-01 1.28899738e-01 2.53392011e-01 3.89191151e-01
-8.99311960e-01 -1.60616860e-01 -2.69459099e-01 -5.19812524e-01
-6.70633078e-01 3.48647594e-01 8.71579230e-01 -4.51397955e-01
-7.87257016e-01 4.97952163e-01 1.61220849e-01 -4.66993064e-01
-3.85433912e-01 -3.92406732e-01 1.85774583e-02 -3.90421420e-01
-8.40171799e-02 1.13768838e-01 -5.59074096e-02 -4.33900177e-01
-1.63357928e-01 8.62362266e-01 6.54025078e-01 -4.45439309e-01
1.44263709e+00 3.22820693e-01 -1.50870241e-03 5.15752912e-01
1.13451517e+00 6.03644848e-01 -1.71335816e+00 -9.78566557e-02
-2.62982756e-01 -7.34964535e-02 -3.38329196e-01 -6.59839392e-01
-1.08974862e+00 5.54767251e-01 1.46091446e-01 5.14468431e-01
1.02753723e+00 4.73275268e-03 8.36789429e-01 7.39097372e-02
3.27419192e-01 -4.33627725e-01 1.17681950e-01 3.93812984e-01
1.05956936e+00 -7.01737463e-01 -2.01411739e-01 -6.60248697e-01
9.68626589e-02 1.21183610e+00 -9.60538723e-03 -8.70308220e-01
1.10851657e+00 7.03044116e-01 -1.11872010e-01 -3.37738335e-01
-6.23062849e-01 2.47215196e-01 4.36349183e-01 4.77722734e-01
1.25852495e-01 -2.28886213e-02 1.45525768e-01 3.36144835e-01
-1.31368428e-01 -1.15807034e-01 5.09133816e-01 7.61344969e-01
4.29366976e-02 -1.11561370e+00 -3.28551531e-01 1.34820864e-01
-5.13611913e-01 8.64210725e-02 -2.14863375e-01 1.02702487e+00
4.52280790e-02 2.07132697e-01 8.76176804e-02 -3.67834657e-01
5.05011559e-01 -7.22554559e-03 5.50572395e-01 -5.64057708e-01
-5.92181742e-01 -2.00066000e-01 7.57707581e-02 -4.89008099e-01
-4.61997747e-01 -6.93487406e-01 -1.19036984e+00 -2.62731522e-01
2.37071797e-01 2.41220087e-01 4.23594296e-01 8.48880708e-01
3.54325324e-01 8.93999636e-01 7.66297936e-01 -1.29157925e+00
-2.46704459e-01 -6.79350257e-01 -5.32311797e-01 6.90326631e-01
8.55945945e-01 -5.97598374e-01 -6.64444685e-01 3.12280953e-01] | [6.506822109222412, 3.3827428817749023] |
0f5aeb98-926a-4e61-a350-221d02120122 | zero-memory-optimization-towards-training-a | 1910.02054 | null | https://arxiv.org/abs/1910.02054v3 | https://arxiv.org/pdf/1910.02054v3.pdf | ZeRO: Memory Optimizations Toward Training Trillion Parameter Models | Large deep learning models offer significant accuracy gains, but training billions to trillions of parameters is challenging. Existing solutions such as data and model parallelisms exhibit fundamental limitations to fit these models into limited device memory, while obtaining computation, communication and development efficiency. We develop a novel solution, Zero Redundancy Optimizer (ZeRO), to optimize memory, vastly improving training speed while increasing the model size that can be efficiently trained. ZeRO eliminates memory redundancies in data- and model-parallel training while retaining low communication volume and high computational granularity, allowing us to scale the model size proportional to the number of devices with sustained high efficiency. Our analysis on memory requirements and communication volume demonstrates: ZeRO has the potential to scale beyond 1 Trillion parameters using today's hardware. We implement and evaluate ZeRO: it trains large models of over 100B parameter with super-linear speedup on 400 GPUs, achieving throughput of 15 Petaflops. This represents an 8x increase in model size and 10x increase in achievable performance over state-of-the-art. In terms of usability, ZeRO can train large models of up to 13B parameters (e.g., larger than Megatron GPT 8.3B and T5 11B) without requiring model parallelism which is harder for scientists to apply. Last but not the least, researchers have used the system breakthroughs of ZeRO to create the world's largest language model (Turing-NLG, 17B parameters) with record breaking accuracy. | ['Jeff Rasley', 'Samyam Rajbhandari', 'Olatunji Ruwase', 'Yuxiong He'] | 2019-10-04 | null | null | null | null | ['cross-lingual-document-classification'] | ['natural-language-processing'] | [-3.87014419e-01 -2.17873096e-01 -3.92509133e-01 -5.27937114e-01
-8.45405340e-01 -5.73469758e-01 1.46236554e-01 1.05715618e-01
-8.48966479e-01 4.62432474e-01 -4.11133498e-01 -9.29228425e-01
3.45116258e-02 -7.78480113e-01 -9.30696309e-01 -3.84171456e-01
1.33550197e-01 9.06985700e-01 3.01913247e-02 -2.44034216e-01
1.91373214e-01 6.62165344e-01 -1.45205641e+00 3.79033267e-01
5.46520770e-01 1.03532529e+00 2.70907253e-01 9.72119570e-01
-8.05485994e-02 6.33046627e-01 -7.51263857e-01 -1.06661402e-01
2.09856182e-01 2.06225738e-01 -9.46441650e-01 -5.49258351e-01
5.99880755e-01 -6.48412883e-01 -2.31765196e-01 3.97486269e-01
6.41419590e-01 -3.21120352e-01 7.10785016e-03 -9.80529606e-01
-2.04474270e-01 5.89674056e-01 -6.13343358e-01 3.33808094e-01
-3.05476755e-01 2.29155988e-01 7.41799653e-01 -5.08139372e-01
3.01478468e-02 9.79682624e-01 8.70103061e-01 8.29585791e-01
-1.11803699e+00 -9.44311500e-01 -4.06229079e-01 -2.43754849e-01
-1.55567062e+00 -7.55235791e-01 -2.91606307e-01 -1.68596774e-01
2.13777423e+00 4.60469723e-01 8.97776723e-01 8.29916298e-01
3.65153372e-01 4.85596418e-01 6.13452673e-01 -4.43682909e-01
7.45856240e-02 4.91857864e-02 3.63765091e-01 9.80659544e-01
4.82718647e-01 -1.39929920e-01 -7.22898185e-01 -1.04332097e-01
8.11058640e-01 -3.47245812e-01 3.04054230e-01 4.73896354e-01
-1.02641618e+00 7.11243987e-01 3.33049089e-01 2.91090697e-01
3.96799669e-02 7.33899832e-01 4.81881857e-01 2.48636484e-01
2.16248453e-01 8.99209738e-01 -7.81957030e-01 -6.91938162e-01
-1.04836905e+00 3.65198940e-01 8.60384762e-01 1.28875554e+00
6.34616435e-01 4.71041918e-01 3.84474695e-01 7.61882544e-01
1.27266601e-01 6.84093416e-01 7.83444107e-01 -7.90946841e-01
7.95058727e-01 6.09518290e-01 -1.38249189e-01 -5.14947176e-01
-8.76070440e-01 -5.40178120e-01 -9.54194009e-01 -1.20936573e-01
1.33138359e-01 -2.93647379e-01 -8.75544727e-01 1.43763995e+00
3.12205672e-01 -1.03733897e-01 -1.11016974e-01 5.89133263e-01
7.84246385e-01 9.65987027e-01 5.91208786e-02 2.44583249e-01
1.45695162e+00 -1.13080835e+00 -1.13382109e-01 -5.74104130e-01
1.59925675e+00 -8.90439510e-01 1.26379752e+00 5.07860839e-01
-1.33079398e+00 -6.46246016e-01 -1.40441144e+00 -4.72014636e-01
-1.87011197e-01 4.57280176e-03 1.17339313e+00 8.87275159e-01
-1.31187165e+00 7.81060696e-01 -1.43746197e+00 -1.00348331e-01
3.86964738e-01 1.16421044e+00 -2.45623186e-01 9.91461053e-02
-7.46637285e-01 8.07087004e-01 7.31622338e-01 -2.62631029e-01
-4.97442991e-01 -1.17232633e+00 -4.46771592e-01 2.63700277e-01
-8.24314430e-02 -9.84757423e-01 1.33622861e+00 -4.64320362e-01
-1.42020643e+00 7.71800876e-01 -1.42456353e-01 -8.02369237e-01
-1.79260410e-02 -3.56102526e-01 -4.92696732e-01 -3.93360645e-01
-4.31409627e-01 7.62277544e-01 2.93843627e-01 -4.84490961e-01
-7.19028950e-01 -4.25219417e-01 -8.50140955e-03 9.59505513e-02
-1.20556962e+00 1.41262159e-01 -8.49463880e-01 -1.74371257e-01
1.58718556e-01 -1.18539262e+00 -2.27081597e-01 -3.08389992e-01
-3.32240365e-03 1.97105762e-02 8.27207446e-01 -4.73010868e-01
1.46510351e+00 -1.96871912e+00 -2.31516257e-01 -3.38726416e-02
4.15658236e-01 6.24561131e-01 -1.02928951e-01 1.18211769e-01
4.67804760e-01 4.24031287e-01 1.88878357e-01 -4.28934962e-01
-1.74860671e-01 4.80438828e-01 -9.15004387e-02 2.46552616e-01
3.14980675e-03 1.13209391e+00 -3.37717235e-01 -3.97561759e-01
-6.60006255e-02 6.05516672e-01 -1.02573943e+00 9.07605439e-02
4.24994566e-02 2.19596326e-02 -5.68964422e-01 4.67863530e-01
5.78778565e-01 -7.39164233e-01 2.29785174e-01 -1.62964389e-01
-1.61163136e-01 5.45865238e-01 -7.79195726e-01 1.54949307e+00
-8.86592865e-01 7.99302042e-01 9.03215408e-02 -7.48280585e-01
7.59461939e-01 8.54336992e-02 3.99701923e-01 -9.12170589e-01
1.85971245e-01 4.32700545e-01 7.13219196e-02 -2.45713711e-01
7.98261821e-01 -2.31938455e-02 -1.67536706e-01 4.52948600e-01
1.60079032e-01 -4.01506364e-01 5.97394034e-02 -4.07615975e-02
1.04342282e+00 -2.71264285e-01 4.63939738e-03 -4.02242094e-01
-6.08369112e-02 1.23683579e-01 2.59876400e-01 8.59963238e-01
2.86311239e-01 3.78364362e-02 1.11640304e-01 -8.60739291e-01
-1.58811307e+00 -5.63006043e-01 -3.30252320e-01 1.48987257e+00
-1.82222933e-01 -7.98662424e-01 -9.46157515e-01 1.57412738e-01
8.46767146e-03 4.99822438e-01 2.91571859e-02 -4.92899418e-02
-1.12105703e+00 -1.12664878e+00 1.11971533e+00 7.65017629e-01
7.56247699e-01 -5.28400064e-01 -8.02690268e-01 3.14709574e-01
2.03648835e-01 -1.32296383e+00 -1.49355069e-01 4.08151746e-01
-1.38267326e+00 -2.76127309e-01 -1.21283345e-01 -7.88469434e-01
4.24494565e-01 6.15581684e-02 1.42681515e+00 4.35933083e-01
-3.39123666e-01 -2.96113372e-01 1.11881517e-01 -4.32406276e-01
-5.16360700e-01 8.15262318e-01 3.52019101e-01 -1.00594521e+00
3.87471706e-01 -5.65929472e-01 -5.54810882e-01 1.45260394e-01
-5.57248116e-01 5.88903785e-01 7.16626823e-01 8.64496231e-01
4.86010641e-01 -1.00048877e-01 3.27429086e-01 -9.11906242e-01
3.64231378e-01 -2.09527493e-01 -9.05351281e-01 2.56106853e-01
-9.56468701e-01 3.45053941e-01 8.64963412e-01 -5.05527377e-01
-6.98085308e-01 -1.06451079e-01 -4.86806005e-01 -6.59980550e-02
2.87394255e-01 3.31296057e-01 2.52178431e-01 -4.67969656e-01
9.85801041e-01 -2.66272966e-02 -1.53839678e-01 -4.29022670e-01
1.43353149e-01 8.53074789e-01 2.55986840e-01 -8.01310360e-01
2.08318189e-01 8.15092251e-02 4.29986753e-02 -8.42002869e-01
-6.11705422e-01 -2.98781902e-01 -3.21893901e-01 3.31808180e-01
5.42377770e-01 -1.14551735e+00 -1.16269016e+00 4.39486355e-01
-1.12518203e+00 -7.39205658e-01 4.34102342e-02 4.59565103e-01
-2.84508020e-01 -2.34551951e-01 -1.02355778e+00 -4.30684686e-01
-9.87928629e-01 -1.32209969e+00 1.30807722e+00 1.72316402e-01
-2.56458133e-01 -9.37009394e-01 -4.10335630e-01 6.67650461e-01
7.36083090e-01 -2.59989202e-01 1.02620935e+00 -4.14083272e-01
-6.32532358e-01 -2.71734565e-01 -2.75743544e-01 2.06062004e-01
-5.44785082e-01 -2.64883578e-01 -8.66422236e-01 -7.59297609e-01
1.10453680e-01 -5.63502908e-01 4.79808122e-01 9.87574682e-02
1.41052210e+00 -3.36783469e-01 -5.87604046e-01 1.15125930e+00
1.54155159e+00 1.30322397e-01 4.40477610e-01 2.15870485e-01
9.78510618e-01 -2.06972241e-01 2.08327815e-01 4.27634865e-01
1.11136243e-01 9.74089384e-01 2.16466524e-02 -2.13629469e-01
2.50261417e-03 -6.95532784e-02 1.06309056e-01 1.61119163e+00
-5.29487915e-02 -3.61789137e-01 -1.51779389e+00 1.67324990e-01
-1.29042840e+00 -3.51387411e-01 -2.46858194e-01 2.24676561e+00
9.92709994e-01 5.38747013e-01 -1.77634418e-01 -1.76960513e-01
1.16764456e-01 -2.72914678e-01 -6.92049921e-01 -1.01210332e+00
1.35752931e-01 5.90756595e-01 1.04445207e+00 5.88375449e-01
-6.23679161e-01 1.30809581e+00 6.99924755e+00 1.21608984e+00
-1.56572747e+00 3.11373740e-01 9.21734214e-01 -8.10825467e-01
5.47591150e-02 -2.65723169e-01 -1.60436714e+00 2.48257175e-01
1.87020326e+00 -1.27782777e-01 6.76646054e-01 1.03770888e+00
-7.62447491e-02 5.37731498e-02 -1.19986796e+00 1.28208673e+00
-1.81911424e-01 -1.64552522e+00 1.55776024e-01 4.59406227e-01
8.21472585e-01 4.95251566e-01 1.29873753e-01 6.37883723e-01
1.83853954e-01 -1.40899134e+00 5.02545655e-01 1.28205925e-01
1.07663107e+00 -9.99290407e-01 6.15457356e-01 6.75390840e-01
-9.21265066e-01 -2.66002625e-01 -6.77602947e-01 -2.99054235e-01
1.80802196e-02 4.05281484e-01 -1.10735607e+00 -4.90177497e-02
8.21388245e-01 4.99676503e-02 -6.62701964e-01 4.39052343e-01
4.31535065e-01 9.46393430e-01 -8.74316633e-01 -3.92382711e-01
3.60472828e-01 1.72297239e-01 -1.69789031e-01 1.28296983e+00
4.75813806e-01 3.39318782e-01 -1.72855020e-01 3.46570462e-01
-3.57949972e-01 -4.53967527e-02 -2.29772747e-01 -1.26369283e-01
8.40427577e-01 1.17534947e+00 -5.14183879e-01 -5.81188142e-01
-1.90319940e-01 7.94345737e-01 7.48402059e-01 -2.64974087e-01
-1.09830177e+00 -1.61937147e-01 5.99231839e-01 6.37666583e-02
-6.43622205e-02 -6.56535506e-01 -8.86427581e-01 -8.42544794e-01
-4.37005563e-03 -1.09200299e+00 -6.17419891e-02 -4.65759128e-01
-6.58143520e-01 6.97679520e-01 -2.39424154e-01 -6.54034436e-01
-2.47819275e-01 -9.05292273e-01 3.77178751e-02 9.54562426e-01
-7.68690467e-01 -1.24643230e+00 -2.32351542e-01 1.26115590e-01
4.03063029e-01 -3.43561381e-01 1.18558609e+00 6.27357960e-01
-5.68594217e-01 1.28121662e+00 4.28108901e-01 -2.75365800e-01
2.26518974e-01 -7.74791479e-01 1.15161288e+00 2.69817561e-01
9.18521956e-02 9.87901509e-01 3.73745441e-01 -2.79203862e-01
-2.01787281e+00 -9.02703464e-01 9.72932518e-01 -6.41311944e-01
5.70405483e-01 -7.24997878e-01 -7.24575222e-01 7.19201088e-01
-2.62182325e-01 6.00615218e-02 7.17063487e-01 5.17782271e-01
-2.40359321e-01 -3.35149825e-01 -8.48779082e-01 7.03648329e-01
1.06895053e+00 -4.45525110e-01 1.99588269e-01 6.75760031e-01
9.16986763e-01 -1.03921437e+00 -1.14147449e+00 4.37081546e-01
8.69643509e-01 -6.93467379e-01 9.37835872e-01 -7.11679876e-01
2.60862261e-01 3.90950799e-01 -9.10799280e-02 -7.44632959e-01
-2.12165609e-01 -5.51504195e-01 -3.62740785e-01 8.11502755e-01
6.82345748e-01 -8.15423310e-01 1.04245591e+00 9.44243371e-01
-4.02825505e-01 -1.37143600e+00 -1.03600359e+00 -7.77781427e-01
4.05235380e-01 -6.28783703e-01 9.81357634e-01 8.20142508e-01
-6.81982562e-02 5.81194460e-01 -4.19619650e-01 -8.57076794e-02
1.15235314e-01 -2.29493260e-01 9.54189241e-01 -8.16373050e-01
-8.09964597e-01 -5.42744040e-01 -4.44619894e-01 -1.53215313e+00
-1.62046015e-01 -8.85802627e-01 -4.40552652e-01 -1.15010977e+00
1.77711129e-01 -9.87527013e-01 2.48032734e-02 8.39943528e-01
3.09095383e-01 4.28559393e-01 2.00349733e-01 2.23627329e-01
-4.49107587e-01 6.56221658e-02 1.09466529e+00 8.56241435e-02
-2.31195673e-01 -2.96300888e-01 -6.49491489e-01 6.51942551e-01
8.93987060e-01 -3.89900565e-01 -4.04656291e-01 -1.38301599e+00
5.04108787e-01 4.59973514e-02 -1.14478618e-01 -1.47128522e+00
4.98240650e-01 5.74228875e-02 4.51090276e-01 -1.86967596e-01
6.57881141e-01 -5.00573456e-01 3.09498459e-01 4.43089873e-01
-1.45461261e-01 4.27909970e-01 9.58808899e-01 -2.49691725e-01
6.43575862e-02 -2.34923795e-01 7.35027254e-01 -1.78428944e-02
-4.66319561e-01 2.80806422e-01 -1.34886339e-01 -1.06059890e-02
6.84878767e-01 -1.00201294e-01 -6.00516677e-01 1.68539643e-01
-2.73509204e-01 1.51771694e-01 2.73612231e-01 3.33193272e-01
2.62229532e-01 -1.12432563e+00 -4.16593343e-01 4.28745747e-01
-2.26358011e-01 2.60930538e-01 4.24213052e-01 5.30894458e-01
-1.26844931e+00 8.64336431e-01 -1.03858851e-01 -7.93646216e-01
-1.46080506e+00 2.48654604e-01 4.59008127e-01 -5.52995145e-01
-6.27913713e-01 1.16294801e+00 -1.26940235e-01 -4.04802263e-01
-1.01639189e-01 -3.75698119e-01 4.60693866e-01 -6.05100751e-01
5.71192622e-01 3.65012705e-01 5.64228594e-01 -3.43258321e-01
-1.99771345e-01 4.26914215e-01 -4.29568440e-01 1.30383987e-02
1.30941284e+00 4.36031848e-01 -6.30777851e-02 1.45328283e-01
1.47870481e+00 -3.05424482e-01 -8.61247063e-01 1.76526025e-01
-1.25771150e-01 -9.87159014e-02 4.47164506e-01 -6.33098483e-01
-8.02954078e-01 1.11359704e+00 8.32412302e-01 1.20289195e-02
8.65366459e-01 -1.32090420e-01 1.23774266e+00 8.17876875e-01
7.64189243e-01 -1.17595899e+00 -9.39671099e-02 8.61120582e-01
4.10234839e-01 -1.08951008e+00 3.47887576e-01 -2.05941740e-02
-7.73903951e-02 9.94265497e-01 9.50734019e-01 2.35269666e-01
5.06442010e-01 9.97165382e-01 -1.95700675e-01 -1.87119424e-01
-9.05534387e-01 5.33357739e-01 1.27841324e-01 1.56674266e-01
7.61291385e-01 3.82857114e-01 -1.70890883e-01 5.78424394e-01
-8.86750042e-01 -8.28419775e-02 -7.47675949e-04 8.81255686e-01
-5.50762832e-01 -1.16489077e+00 -9.95728523e-02 7.62989581e-01
-4.31658834e-01 -4.47920561e-01 2.22870871e-01 9.20487821e-01
1.08953856e-03 8.50686789e-01 4.84958947e-01 -6.96586907e-01
1.66051224e-01 -9.15158987e-02 5.96074998e-01 -4.96512800e-01
-9.77170050e-01 -1.36429876e-01 2.16695264e-01 -5.05644619e-01
4.21984375e-01 4.07960489e-02 -1.48330224e+00 -1.21006787e+00
-4.85967457e-01 -7.49334767e-02 1.24001920e+00 8.69810700e-01
8.57470274e-01 4.47590262e-01 1.75688893e-01 -8.73165369e-01
-7.34487593e-01 -8.61588478e-01 -2.16313526e-01 -1.61957145e-01
-6.93235546e-02 -6.96029812e-02 1.09034613e-01 -1.24036461e-01] | [8.566521644592285, 3.3390743732452393] |
55e46439-a32a-4390-85bb-2584fb3078e5 | contextualized-topic-coherence-metrics | 2305.14587 | null | https://arxiv.org/abs/2305.14587v1 | https://arxiv.org/pdf/2305.14587v1.pdf | Contextualized Topic Coherence Metrics | The recent explosion in work on neural topic modeling has been criticized for optimizing automated topic evaluation metrics at the expense of actual meaningful topic identification. But human annotation remains expensive and time-consuming. We propose LLM-based methods inspired by standard human topic evaluations, in a family of metrics called Contextualized Topic Coherence (CTC). We evaluate both a fully automated version as well as a semi-automated CTC that allows human-centered evaluation of coherence while maintaining the efficiency of automated methods. We evaluate CTC relative to five other metrics on six topic models and find that it outperforms automated topic coherence methods, works well on short documents, and is not susceptible to meaningless but high-scoring topics. | ['Bernd Amann', 'Camelia Constantin', 'Hubert Naacke', 'David Mimno', 'Jacob Louis Hoover', 'Hamed Rahimi'] | 2023-05-23 | null | null | null | null | ['topic-models'] | ['natural-language-processing'] | [ 1.17166005e-01 4.88167018e-01 -2.86061168e-01 -5.58912814e-01
-1.52083790e+00 -6.27571762e-01 1.09740269e+00 6.31503403e-01
-5.01470268e-01 7.00213611e-01 5.32510817e-01 -2.31514201e-01
-3.18849176e-01 -4.96864438e-01 -6.61533624e-02 -5.21879315e-01
-1.94229990e-01 1.00363183e+00 2.69507885e-01 1.82527184e-01
5.09346902e-01 -4.02980536e-01 -1.54185295e+00 3.48131545e-02
1.14344847e+00 6.64864898e-01 1.33551389e-01 5.08460045e-01
-1.39106616e-01 3.67555439e-01 -1.11685336e+00 -3.67925704e-01
-4.24337953e-01 -3.66400093e-01 -8.73568416e-01 7.87448734e-02
7.77768731e-01 -6.59190416e-02 1.86589628e-01 8.39295566e-01
2.28557512e-01 2.38441080e-01 8.77067447e-01 -1.33506107e+00
-3.66008610e-01 8.68812203e-01 -3.86888117e-01 1.58878714e-01
1.30802676e-01 -3.12037349e-01 1.44225681e+00 -8.56639564e-01
7.39598572e-01 1.27758813e+00 9.17471111e-01 1.37860730e-01
-1.28673553e+00 -4.92437750e-01 1.21680081e-01 -2.37033620e-01
-1.24875796e+00 -2.24413574e-01 4.45057392e-01 -6.30505979e-01
1.03314912e+00 1.53669745e-01 5.54820895e-01 1.05370605e+00
2.74183124e-01 7.04008162e-01 9.23175395e-01 -2.67283946e-01
5.03031790e-01 3.59149218e-01 9.28182304e-01 3.36805731e-01
6.91955984e-01 -3.74541849e-01 -7.02453494e-01 -7.81706810e-01
3.87501568e-01 -5.73527217e-01 -1.72809169e-01 -2.40268782e-01
-1.41120851e+00 1.28727460e+00 -1.56466052e-01 3.71043444e-01
-4.32737827e-01 1.49094522e-01 4.13841635e-01 7.59433433e-02
1.23017573e+00 9.78794515e-01 -4.27712560e-01 -3.99791569e-01
-1.64587045e+00 7.78527081e-01 1.00076890e+00 8.88244927e-01
7.02866912e-01 -7.61704743e-02 -4.36452597e-01 7.99896419e-01
2.93656439e-01 2.16684997e-01 7.82687664e-01 -1.09823525e+00
2.01162890e-01 5.35016537e-01 1.82857245e-01 -9.32616770e-01
-6.33811474e-01 -5.96598268e-01 -4.66813266e-01 -2.24530488e-01
1.40123978e-01 -1.03452630e-01 -7.64006019e-01 1.61478615e+00
1.49652719e-01 -2.68313020e-01 -2.97506213e-01 4.15506363e-01
7.84103930e-01 7.45523751e-01 3.88663083e-01 -1.97428897e-01
1.39484036e+00 -8.88061643e-01 -8.75248253e-01 -1.00989453e-01
7.97100604e-01 -6.07214868e-01 1.33630455e+00 5.44979751e-01
-1.07239950e+00 -9.80338901e-02 -1.06547654e+00 -1.30641222e-01
-4.65407282e-01 1.50392443e-01 9.52009559e-01 9.99431491e-01
-1.30374479e+00 4.69450206e-01 -8.75048697e-01 -4.92350757e-01
3.80287230e-01 3.54877681e-01 6.70091761e-03 4.12561864e-01
-1.16927731e+00 8.63412082e-01 6.20231926e-01 -6.37537837e-01
-8.99609566e-01 -7.64530301e-01 -7.45444238e-01 3.02807122e-01
3.66945416e-01 -6.69164538e-01 1.67274809e+00 -4.23759997e-01
-1.24101758e+00 7.30089903e-01 -2.51077652e-01 -6.87167466e-01
2.83219129e-01 -6.03162050e-01 -8.47819373e-02 1.93070248e-01
4.63745892e-01 1.03576350e+00 5.62909782e-01 -9.59294558e-01
-7.77510822e-01 -1.37902752e-01 -3.03202420e-01 4.04274940e-01
-7.89302588e-01 1.18725993e-01 -3.78489017e-01 -7.31480718e-01
1.28607526e-01 -7.89431214e-01 -7.62243643e-02 -4.04371202e-01
-6.26385570e-01 -8.94730628e-01 8.91663432e-01 -5.22191644e-01
1.36646926e+00 -1.64829087e+00 -2.52721131e-01 1.05508961e-01
4.53339487e-01 -1.66166276e-01 1.14764594e-01 1.90923825e-01
2.19122723e-01 4.49955940e-01 -3.07468891e-01 -7.83283293e-01
2.07867622e-01 -3.42197686e-01 -5.36781311e-01 3.65980297e-01
1.46975756e-01 6.74140632e-01 -8.12918305e-01 -8.98294032e-01
4.83363587e-03 4.95013535e-01 -5.81653833e-01 -6.04237542e-02
-4.53133315e-01 -1.78319663e-01 -3.12856376e-01 5.55082977e-01
7.37320781e-02 -5.82568467e-01 8.39391202e-02 2.13571116e-01
-9.40960795e-02 8.01652312e-01 -8.87469947e-01 1.65704334e+00
-1.51581969e-02 1.27588129e+00 -2.54078507e-01 -5.88928223e-01
9.73159134e-01 6.38505399e-01 6.04295671e-01 -1.49808854e-01
2.38868054e-02 -1.56135494e-02 -4.11870569e-01 8.02979097e-02
1.23243999e+00 1.06220283e-01 -1.30690992e-01 9.37534451e-01
3.58341277e-01 -3.93065989e-01 1.69470325e-01 5.24779856e-01
9.47182238e-01 -7.95126259e-02 3.34296048e-01 -9.04887617e-01
-2.26841375e-01 2.97763854e-01 8.38636011e-02 9.72718358e-01
-9.05511826e-02 8.03202927e-01 5.08723974e-01 -1.22631222e-01
-1.19590318e+00 -8.77252936e-01 -3.28474343e-01 1.54634011e+00
-6.47832975e-02 -9.75249171e-01 -9.99816477e-01 -4.51879352e-01
-2.63472646e-01 1.11242676e+00 -7.19174266e-01 9.71395075e-02
-2.59879857e-01 -1.02460873e+00 6.15725756e-01 3.22602749e-01
1.97473839e-01 -7.43579328e-01 -7.97136605e-01 2.28027016e-01
-3.60013902e-01 -8.01129699e-01 -3.79982412e-01 3.89148921e-01
-1.02322745e+00 -5.97756028e-01 -9.19801235e-01 -3.56408328e-01
1.16026141e-01 3.65104228e-01 1.49760544e+00 -3.35109323e-01
9.67234001e-02 2.82071620e-01 -6.66020364e-02 -8.46804023e-01
-1.42057016e-01 6.03254676e-01 -6.87856823e-02 -6.14671826e-01
5.70445061e-01 -2.09350094e-01 -3.80648136e-01 3.14204693e-01
-7.88203359e-01 -1.73231542e-01 2.43188292e-01 8.18008959e-01
1.80938765e-01 1.17100351e-01 8.03658545e-01 -1.06216776e+00
1.22842252e+00 -5.46062469e-01 -3.92213345e-01 1.75339401e-01
-1.30152786e+00 -5.05154766e-02 -1.52254537e-01 -3.17187011e-01
-1.13718009e+00 -4.97384548e-01 3.68856728e-01 6.18720381e-03
-5.01955785e-02 6.00043178e-01 2.28535652e-01 5.01004875e-01
8.79411221e-01 -2.59411246e-01 -3.73548895e-01 -4.36446995e-01
4.79589581e-01 4.80354369e-01 4.65655088e-01 -5.82615376e-01
3.87950927e-01 3.85114074e-01 -5.62281966e-01 -1.03456163e+00
-9.43047225e-01 -7.74148643e-01 -2.84434527e-01 -9.57519412e-02
6.85951531e-01 -1.01427364e+00 -7.83418864e-02 1.30512297e-01
-1.32949877e+00 -1.61365867e-01 -3.00218701e-01 4.87094164e-01
-6.48119390e-01 2.45787263e-01 -4.19863254e-01 -6.72648489e-01
-6.99043572e-01 -8.60160172e-01 1.47832572e+00 8.71765986e-02
-1.03997326e+00 -1.18605316e+00 4.90786642e-01 5.31280786e-02
5.77603936e-01 -6.19950984e-03 9.82215583e-01 -1.10538304e+00
-3.36748272e-01 -3.87061089e-01 -2.08412573e-01 -1.01886071e-01
-1.47029638e-01 1.80343315e-01 -1.16370320e+00 -1.28074586e-01
4.24817204e-02 -2.86721021e-01 1.12728560e+00 8.43610883e-01
6.49633348e-01 -3.72924954e-01 -5.73788583e-01 4.36475277e-02
1.03603613e+00 -6.79672509e-03 1.97200030e-01 6.36451960e-01
1.06993400e-01 7.63190091e-01 6.30931318e-01 3.69052172e-01
5.21020174e-01 6.39580071e-01 -9.74235460e-02 -4.05257717e-02
9.41568166e-02 -6.52690604e-02 1.01180173e-01 6.97854638e-01
1.70689702e-01 -6.96956873e-01 -1.21948612e+00 8.29910338e-01
-1.89889479e+00 -7.51705647e-01 -5.87668680e-02 1.92689061e+00
1.00297260e+00 4.39389527e-01 3.19453090e-01 1.07288698e-03
7.41657317e-01 2.17399836e-01 -2.18601778e-01 -2.08453968e-01
-2.49416128e-01 8.88459012e-02 6.09477386e-02 4.84031618e-01
-1.46521318e+00 1.01600838e+00 7.98697662e+00 7.90822685e-01
-6.30281389e-01 4.49811637e-01 7.08678782e-01 -2.19168261e-01
-5.28239012e-01 8.23380239e-03 -8.76656055e-01 2.08374694e-01
1.08277762e+00 -6.05304182e-01 -4.81953532e-01 1.29507184e+00
5.77499941e-02 -3.95307690e-01 -1.05771315e+00 5.74023426e-01
2.66912788e-01 -1.29241812e+00 1.34461641e-01 1.41138181e-01
1.01715314e+00 1.01074353e-01 6.06412627e-02 4.29371119e-01
4.96620834e-01 -8.83220792e-01 6.37101948e-01 1.98535323e-01
3.53757143e-01 -5.08050919e-01 7.95132935e-01 8.35488737e-02
-8.44273508e-01 2.42313981e-01 -3.57891411e-01 2.88984627e-01
1.87615037e-01 8.75774324e-01 -1.04939878e+00 -6.16837218e-02
6.32188499e-01 3.99641156e-01 -7.99421847e-01 1.21844399e+00
5.31852804e-02 1.05408537e+00 -5.24646342e-01 -2.64925808e-01
5.99678159e-01 2.32530981e-01 9.02543426e-01 1.58558977e+00
2.06491455e-01 -3.27005744e-01 8.13340545e-02 8.53851438e-01
1.43410504e-01 1.76514849e-01 -5.84717214e-01 -1.47422507e-01
6.52482688e-01 1.20559216e+00 -1.08819675e+00 -6.89755619e-01
4.01254818e-02 3.67622823e-01 7.08951652e-02 3.38472635e-01
-5.42669356e-01 -3.80188107e-01 3.07389498e-01 -1.34729221e-01
-4.71256189e-02 -3.69367301e-01 -6.71648979e-01 -8.65044177e-01
-1.92856520e-01 -6.33666694e-01 5.24884462e-01 -6.73264503e-01
-8.78860056e-01 8.22180092e-01 5.92796922e-01 -8.00267220e-01
-6.24862373e-01 -2.56654322e-02 -8.46134186e-01 5.36323845e-01
-8.63985002e-01 -9.72221911e-01 -3.21082145e-01 5.93597479e-02
8.31120610e-01 -2.47520715e-01 8.58731091e-01 4.91998717e-02
-3.40829909e-01 4.18935716e-01 1.07600637e-01 -5.31146884e-01
9.00631607e-01 -1.59599149e+00 7.22647011e-01 4.56033409e-01
2.29477987e-01 8.32196891e-01 1.03063917e+00 -8.17945659e-01
-3.90895158e-01 -1.00960660e+00 1.22328508e+00 -6.73027694e-01
5.60521483e-01 -3.24435890e-01 -9.74539638e-01 6.36840582e-01
4.12324160e-01 -1.21512163e+00 8.31330597e-01 8.16829681e-01
-3.80345434e-01 4.89156008e-01 -6.51272476e-01 4.17325437e-01
4.01143253e-01 -4.14774179e-01 -8.72862339e-01 6.37749195e-01
1.09366834e+00 1.78405702e-01 -8.07935655e-01 2.39342555e-01
6.05952501e-01 -6.29593849e-01 7.91421771e-01 -5.27956188e-01
2.31883615e-01 1.10865310e-01 -1.68684438e-01 -1.15383840e+00
-2.02142492e-01 -6.70084715e-01 -2.82446574e-02 1.28677130e+00
5.86001575e-01 -4.07938331e-01 1.10232103e+00 7.56488383e-01
-2.62848854e-01 -7.34647870e-01 -8.08993101e-01 -7.30109751e-01
2.15968192e-01 -5.46024978e-01 3.67141783e-01 1.23839283e+00
3.44220042e-01 6.99806392e-01 -1.04624100e-01 -2.77567178e-01
7.67332911e-01 -5.85048040e-03 6.01149738e-01 -1.68316543e+00
1.46520317e-01 -9.33558643e-01 -2.19441608e-01 -6.24432266e-01
1.12706597e-03 -5.00966012e-01 1.76477045e-01 -1.58454573e+00
4.68679816e-01 -2.57337034e-01 -6.76996931e-02 3.56866717e-01
-3.26938152e-01 1.88984200e-01 -2.69375324e-01 5.83004892e-01
-1.04633105e+00 6.06733739e-01 4.33605760e-01 -1.00131340e-01
-3.08990508e-01 -1.08839869e-01 -7.65316486e-01 9.18197930e-01
7.80310452e-01 -6.02921784e-01 -5.64609110e-01 -2.50326008e-01
2.61365384e-01 -3.53398949e-01 2.33775869e-01 -1.14135456e+00
4.35936093e-01 1.46102801e-01 4.13645729e-02 -1.18527806e+00
3.40075701e-01 -1.35871902e-01 -1.99743286e-01 6.35840967e-02
-9.09013271e-01 1.17023744e-01 9.88645945e-03 5.26998460e-01
-3.31941664e-01 -2.63798207e-01 4.05205578e-01 -9.02343318e-02
-4.70171779e-01 7.66532496e-02 -5.36381245e-01 2.49973029e-01
6.29753113e-01 -9.86991152e-02 -6.50458932e-01 -7.76513398e-01
-4.02540565e-01 1.65172771e-01 2.66395181e-01 4.11236733e-01
1.44080460e-01 -9.35619056e-01 -7.22234786e-01 -4.79384601e-01
1.89742044e-01 -4.70028892e-02 -6.56650290e-02 5.73646247e-01
-1.41897485e-01 1.23067105e+00 2.44665146e-01 -8.98953915e-01
-1.13923562e+00 8.38107094e-02 -1.17222294e-01 -4.64028388e-01
-5.39367020e-01 7.52802372e-01 3.91538531e-01 -1.14416480e-01
6.70843422e-01 -3.05574536e-01 -3.18833947e-01 3.24207634e-01
3.65811706e-01 6.36437893e-01 1.60565212e-01 -1.83626890e-01
1.61213279e-02 2.68159211e-01 -2.67178059e-01 -7.10264027e-01
1.05732274e+00 -1.04414515e-01 1.34857640e-01 7.16514111e-01
1.05907524e+00 -5.92182159e-01 -7.10940540e-01 -2.22540632e-01
6.08061910e-01 5.70547357e-02 3.74038935e-01 -7.58733213e-01
-2.11550951e-01 7.75528371e-01 3.88079792e-01 6.89545572e-01
6.53700531e-01 2.16337413e-01 5.65078616e-01 8.47169161e-01
3.83389592e-02 -1.26593137e+00 1.81187391e-01 5.20302117e-01
6.93732917e-01 -1.27813697e+00 5.17484367e-01 -2.77157813e-01
-6.49941564e-01 8.10034931e-01 4.76880878e-01 5.79695366e-02
7.85433412e-01 6.84530362e-02 -1.15481742e-01 -7.52638817e-01
-1.24429417e+00 1.49936706e-01 6.11516893e-01 4.35275793e-01
7.08833873e-01 1.09911881e-01 -3.89330119e-01 4.69752401e-01
-7.24345207e-01 -3.87447834e-01 3.16310018e-01 7.57494509e-01
-8.69029462e-01 -5.37980676e-01 -1.63350433e-01 7.12701678e-01
-5.00870466e-01 -2.03129217e-01 -5.95639050e-01 1.05050814e+00
-3.55942935e-01 9.35984135e-01 3.23872656e-01 -8.93309861e-02
-2.84305722e-01 3.64819050e-01 -2.05341905e-01 -8.08784544e-01
-5.73187709e-01 4.04677302e-01 4.29468274e-01 -4.62126821e-01
-4.75431055e-01 -9.57008123e-01 -6.65825009e-01 1.51642367e-01
-8.70255828e-01 5.91627002e-01 1.07109678e+00 9.26328361e-01
2.65640974e-01 2.26313472e-01 1.74848914e-01 -7.04862297e-01
-5.28349757e-01 -1.53105128e+00 -3.51766437e-01 1.77555799e-01
1.90298378e-01 -7.73230970e-01 -4.23779488e-01 6.86839893e-02] | [10.375534057617188, 7.0937700271606445] |
c1fdb1e4-40dd-4dcb-ac13-4442179dbc22 | elastic-boundary-projection-for-3d-medical | 1812.00518 | null | https://arxiv.org/abs/1812.00518v2 | https://arxiv.org/pdf/1812.00518v2.pdf | Elastic Boundary Projection for 3D Medical Image Segmentation | We focus on an important yet challenging problem: using a 2D deep network to deal with 3D segmentation for medical image analysis. Existing approaches either applied multi-view planar (2D) networks or directly used volumetric (3D) networks for this purpose, but both of them are not ideal: 2D networks cannot capture 3D contexts effectively, and 3D networks are both memory-consuming and less stable arguably due to the lack of pre-trained models. In this paper, we bridge the gap between 2D and 3D using a novel approach named Elastic Boundary Projection (EBP). The key observation is that, although the object is a 3D volume, what we really need in segmentation is to find its boundary which is a 2D surface. Therefore, we place a number of pivot points in the 3D space, and for each pivot, we determine its distance to the object boundary along a dense set of directions. This creates an elastic shell around each pivot which is initialized as a perfect sphere. We train a 2D deep network to determine whether each ending point falls within the object, and gradually adjust the shell so that it gradually converges to the actual shape of the boundary and thus achieves the goal of segmentation. EBP allows boundary-based segmentation without cutting a 3D volume into slices or patches, which stands out from conventional 2D and 3D approaches. EBP achieves promising accuracy in abdominal organ segmentation. Our code has been open-sourced https://github.com/twni2016/Elastic-Boundary-Projection. | ['Alan L. Yuille', 'Huangjie Zheng', 'Elliot K. Fishman', 'Lingxi Xie', 'Tianwei Ni'] | 2018-12-03 | elastic-boundary-projection-for-3d-medical-1 | http://openaccess.thecvf.com/content_CVPR_2019/html/Ni_Elastic_Boundary_Projection_for_3D_Medical_Image_Segmentation_CVPR_2019_paper.html | http://openaccess.thecvf.com/content_CVPR_2019/papers/Ni_Elastic_Boundary_Projection_for_3D_Medical_Image_Segmentation_CVPR_2019_paper.pdf | cvpr-2019-6 | ['3d-medical-imaging-segmentation'] | ['medical'] | [-1.11782692e-01 2.91556954e-01 -7.00902194e-02 -6.20220117e-02
-3.76842797e-01 -6.10937655e-01 1.35697767e-01 5.79023808e-02
-2.71344870e-01 2.87960470e-01 5.04981466e-02 -3.30240130e-01
1.98750034e-01 -9.98461008e-01 -7.89122105e-01 -6.31764770e-01
-1.49954529e-02 8.67644429e-01 4.80053395e-01 -7.66562596e-02
-9.12402868e-02 7.96612442e-01 -8.88106465e-01 -9.52916220e-02
6.44855142e-01 8.93737018e-01 -1.12466821e-02 6.17285192e-01
-2.75188893e-01 -1.25316411e-01 -1.69299155e-01 -1.55645236e-01
5.75992942e-01 -3.63180339e-01 -1.02589524e+00 9.22499523e-02
-5.67377023e-02 -4.28868800e-01 2.04996783e-02 1.04598761e+00
4.88553345e-01 -2.98197623e-02 7.55806804e-01 -7.79203057e-01
-3.63891274e-01 2.42545396e-01 -9.17774141e-01 -4.97084782e-02
1.88957959e-01 -1.12616420e-01 5.05181372e-01 -9.25343931e-01
7.48376071e-01 8.49337399e-01 1.07954764e+00 8.33994091e-01
-1.02453220e+00 -2.90580630e-01 -8.82127732e-02 -6.22561514e-01
-1.22806692e+00 7.42266178e-02 1.02997184e+00 -6.08547151e-01
5.38510978e-01 3.13269556e-01 1.14687681e+00 6.24395549e-01
3.51843804e-01 8.67389441e-01 8.48000765e-01 -1.99199080e-01
1.06009118e-01 -9.36113670e-02 1.13616250e-01 7.28750646e-01
9.24301967e-02 -1.55626982e-01 1.83729231e-01 -1.59533560e-01
1.36307526e+00 2.68758625e-01 -4.68956590e-01 -8.87499630e-01
-1.03759301e+00 6.87200963e-01 5.70744753e-01 4.26326245e-01
-4.02527153e-01 -8.41001123e-02 3.17680150e-01 -2.35102192e-01
5.19822538e-01 2.96210706e-01 -3.85485023e-01 1.72120810e-01
-9.53003109e-01 1.68089896e-01 7.37695456e-01 5.11935294e-01
6.73138082e-01 -4.04011190e-01 2.69700944e-01 6.48994207e-01
3.81779283e-01 2.40833029e-01 4.15373415e-01 -9.68384564e-01
1.98861718e-01 1.03057253e+00 -7.16040507e-02 -1.02171290e+00
-8.34010422e-01 -1.54359043e-01 -1.10272312e+00 2.72429615e-01
6.69014990e-01 -1.40824914e-01 -1.16507900e+00 1.31482637e+00
9.72141743e-01 7.79435877e-03 -5.56055643e-02 1.28411281e+00
1.08490872e+00 4.17949736e-01 -4.24692541e-01 8.48229378e-02
1.33088541e+00 -9.26454544e-01 -2.32504874e-01 -1.13601066e-01
7.26533115e-01 -6.18718326e-01 8.96065235e-01 2.49702379e-01
-1.52564156e+00 -2.31335655e-01 -9.26625669e-01 -1.00042075e-01
-6.78695366e-02 -2.24974558e-01 4.16208953e-01 4.52385366e-01
-1.12498784e+00 6.95578814e-01 -1.14926672e+00 -5.16146906e-02
5.30932844e-01 3.91004890e-01 -4.54443812e-01 1.12733796e-01
-1.03358078e+00 6.76208675e-01 2.01440334e-01 2.64228284e-01
-6.01030469e-01 -7.58521199e-01 -1.00441337e+00 -9.26028788e-02
3.60730797e-01 -9.01877701e-01 1.07820702e+00 -7.78512776e-01
-1.33454084e+00 1.42367971e+00 1.60414857e-04 -2.63646066e-01
7.30589271e-01 2.47856397e-02 2.47810721e-01 3.05301696e-01
9.54338163e-03 5.58106780e-01 5.31036735e-01 -1.48769879e+00
-1.26158427e-02 -7.60793507e-01 2.72498801e-02 3.67744267e-01
3.40766981e-02 -3.48392606e-01 -7.63680339e-01 -4.02133882e-01
7.13587105e-01 -9.45468605e-01 -4.38975304e-01 2.25059107e-01
-7.68867016e-01 -1.16175868e-01 6.83664560e-01 -7.71821082e-01
9.81438458e-01 -2.15025306e+00 2.22808272e-01 2.56708354e-01
6.74406826e-01 3.27170014e-01 3.37646186e-01 -6.51409924e-02
-9.69810411e-03 3.23693961e-01 -5.67182899e-01 -3.15479279e-01
-4.29971576e-01 7.26251602e-02 2.58676678e-01 6.17797673e-01
-1.28559142e-01 1.00351679e+00 -8.40832114e-01 -6.03979647e-01
2.45718226e-01 6.67603374e-01 -6.03351712e-01 -4.78367358e-02
-1.04293987e-01 5.66185713e-01 -6.58587873e-01 4.97633666e-01
1.01186037e+00 -5.88170648e-01 1.41200006e-01 -5.70431165e-02
-1.73378721e-01 1.48041978e-01 -1.19295430e+00 1.83956575e+00
-2.75654256e-01 7.10724667e-02 3.68416995e-01 -1.11213624e+00
1.00874019e+00 3.90422314e-01 9.37485576e-01 -4.44977522e-01
3.21247607e-01 3.18385482e-01 -2.72539765e-01 -4.29962337e-01
3.20153534e-02 -4.17152584e-01 6.69777766e-02 4.11429942e-01
-2.54231155e-01 -4.23898757e-01 -1.60950541e-01 -5.73342629e-02
7.68679202e-01 2.71694183e-01 3.40365320e-01 -2.76638150e-01
4.56718266e-01 1.27617151e-01 6.26383960e-01 1.74306899e-01
-1.83462933e-01 1.09316826e+00 7.76386559e-01 -9.02453661e-01
-1.01053119e+00 -1.16484737e+00 -2.82821745e-01 3.39956917e-02
6.50093317e-01 2.55600754e-02 -9.64932621e-01 -8.89636099e-01
-5.06160818e-02 1.89794153e-01 -7.49335468e-01 1.17886193e-01
-9.01240289e-01 -6.61073208e-01 1.55674756e-01 4.64345187e-01
4.42942053e-01 -9.12523210e-01 -9.30931389e-01 2.24401146e-01
-1.32934213e-01 -7.43340969e-01 -5.70849240e-01 2.14696959e-01
-1.24162209e+00 -1.27730262e+00 -1.23649526e+00 -7.95208752e-01
9.72189665e-01 1.78432733e-01 1.14722383e+00 1.60068363e-01
-1.49654508e-01 9.08630937e-02 -8.91876146e-02 -2.47930169e-01
-4.90029842e-01 -4.42393832e-02 -1.94620922e-01 -4.20397073e-01
1.20082080e-01 -6.06060684e-01 -1.01239884e+00 4.00115401e-01
-6.59408808e-01 3.10342699e-01 2.81579137e-01 5.96271098e-01
1.04710340e+00 -1.37815207e-01 1.88289195e-01 -7.29616523e-01
2.67573446e-01 -3.05871487e-01 -5.63739777e-01 -6.80718124e-02
-2.01084271e-01 -2.45005220e-01 5.96334696e-01 -2.31375679e-01
-5.29765606e-01 1.81674808e-01 -4.84155774e-01 -7.59644866e-01
-3.10047269e-01 4.18617845e-01 2.18305096e-01 1.37091473e-01
4.61969018e-01 2.34399334e-01 4.03205663e-01 -4.62880731e-01
-7.72356018e-02 2.59683996e-01 1.30365491e-01 -2.19173491e-01
4.90409791e-01 9.03941870e-01 1.35312662e-01 -5.11031866e-01
-9.49126363e-01 -4.61964071e-01 -9.67968345e-01 -2.90829957e-01
1.15113449e+00 -4.28928077e-01 -7.41319716e-01 4.65871185e-01
-1.15521121e+00 -7.10664511e-01 -4.87763226e-01 4.86268133e-01
-4.44983661e-01 5.03294706e-01 -7.78408945e-01 -4.43903595e-01
-6.80795908e-01 -1.36178946e+00 1.07229173e+00 2.83796638e-01
-1.56032920e-01 -1.28138673e+00 2.93356389e-01 3.23088646e-01
7.76554421e-02 6.23818576e-01 8.71776342e-01 -5.22225380e-01
-2.36926407e-01 -2.81441718e-01 -1.81124732e-01 2.60140985e-01
-1.80061795e-02 -7.04133436e-02 -5.50439835e-01 -6.44145012e-02
3.04692209e-01 -2.57822294e-02 6.15407646e-01 9.21735048e-01
1.12083769e+00 -3.43711004e-02 -6.57722533e-01 9.15159822e-01
1.33881748e+00 2.01120317e-01 3.81999403e-01 2.99212396e-01
7.23384082e-01 5.12635708e-01 1.47851139e-01 1.88718453e-01
4.31326419e-01 5.91708124e-01 6.42327726e-01 -5.26238382e-01
-1.24114007e-01 -7.41083696e-02 -2.51988560e-01 8.64404619e-01
-1.06960200e-01 -5.64617962e-02 -1.25354517e+00 5.15798271e-01
-1.48563862e+00 -3.91516328e-01 -3.69865268e-01 2.27018666e+00
8.19701076e-01 2.76509255e-01 2.05509320e-01 -1.00184316e-02
6.24332249e-01 -4.51903902e-02 -6.90425813e-01 -3.16754431e-01
2.61986524e-01 -1.28217295e-01 3.21143448e-01 4.55230325e-01
-1.11846340e+00 5.07539868e-01 5.45799398e+00 5.53940237e-01
-1.43838334e+00 4.74473052e-02 8.07333946e-01 6.80453703e-02
-2.61881411e-01 -2.97689646e-01 -7.81003177e-01 3.39814067e-01
2.32813939e-01 2.43972823e-01 -2.91009713e-02 7.23940313e-01
1.92546681e-01 -1.58163458e-01 -9.02497947e-01 7.93187082e-01
-3.08477506e-02 -1.44489861e+00 -1.15889542e-01 2.49607235e-01
6.97208226e-01 1.74176916e-01 -2.48934954e-01 -3.63031775e-02
2.39548981e-02 -9.99139607e-01 4.29611862e-01 3.77183020e-01
8.14263582e-01 -5.01880944e-01 6.36655986e-01 6.53959095e-01
-1.13935697e+00 5.12289643e-01 -1.80919096e-01 1.48572668e-01
3.63047719e-01 8.38224113e-01 -9.22631025e-01 5.34223974e-01
8.14202726e-01 5.05767643e-01 5.25290929e-02 1.08304799e+00
6.29801303e-02 1.41266346e-01 -5.21813750e-01 1.62161097e-01
4.54697698e-01 -4.93159980e-01 7.29524910e-01 7.33000934e-01
3.39932829e-01 1.60816327e-01 1.54537618e-01 9.59454775e-01
-1.03627950e-01 1.62385345e-01 -7.28402853e-01 5.85312486e-01
7.63259456e-02 1.33295989e+00 -1.38127911e+00 -2.76017696e-01
-3.67723018e-01 7.97767520e-01 1.66449293e-01 -1.23469224e-02
-7.74273634e-01 -2.04965621e-02 1.33654356e-01 4.33309048e-01
2.06001490e-01 -9.27282646e-02 -4.42636758e-01 -1.02481806e+00
1.86035886e-01 -3.38651299e-01 3.79833817e-01 -7.58281350e-01
-1.01804173e+00 6.11791015e-01 -1.27231926e-01 -1.16636443e+00
9.88345668e-02 -5.21012843e-01 -5.94327092e-01 7.24754095e-01
-1.28141522e+00 -8.52287531e-01 -3.63817424e-01 3.03238809e-01
4.22960669e-01 6.88032269e-01 7.66738534e-01 2.03070313e-01
-2.64260679e-01 1.91581905e-01 -1.78335998e-02 4.98188525e-01
2.58566886e-01 -1.38039625e+00 3.78904760e-01 4.10720289e-01
-1.60571441e-01 3.94060403e-01 2.41444960e-01 -8.07639241e-01
-1.12124848e+00 -9.08204377e-01 6.62461400e-01 -4.81408656e-01
2.49167189e-01 -2.96354383e-01 -1.05020761e+00 6.01692736e-01
-1.55154526e-01 3.47370088e-01 5.52577615e-01 -2.14618146e-01
1.56782970e-01 3.82416308e-01 -1.28255951e+00 5.49517989e-01
9.06970799e-01 1.62556261e-01 -4.36580569e-01 3.56175512e-01
6.53805196e-01 -1.08667159e+00 -1.07389259e+00 6.56529486e-01
4.61116672e-01 -1.19652355e+00 1.18497455e+00 -1.59224853e-01
5.04104078e-01 -1.87747136e-01 4.38797414e-01 -1.25166857e+00
2.12455373e-02 -4.08873290e-01 -1.83723606e-02 5.31146824e-01
3.64655107e-01 -5.65319300e-01 1.24441350e+00 5.62242091e-01
-4.94602919e-01 -1.57298338e+00 -9.85046387e-01 -4.99553829e-01
4.54603821e-01 -3.31604391e-01 4.12980318e-01 1.04498672e+00
-1.62336469e-01 2.31343322e-02 2.07032979e-01 7.89644346e-02
4.62517470e-01 4.17227685e-01 5.48607290e-01 -1.25846279e+00
6.72156438e-02 -5.78046083e-01 -2.19901115e-01 -1.22382963e+00
-2.38586664e-01 -1.09217978e+00 -2.04241738e-01 -1.92757010e+00
2.43193820e-01 -6.95791721e-01 2.57943004e-01 3.37151349e-01
-7.03807175e-03 3.41578424e-01 -7.37080947e-02 3.09347183e-01
-2.84140527e-01 2.67628789e-01 2.03121400e+00 1.57425076e-01
-5.87005496e-01 3.26540619e-01 -3.67606133e-01 1.13737071e+00
8.46002579e-01 -2.90618837e-01 -1.03829764e-01 -4.26353276e-01
2.40948975e-01 5.07323146e-01 4.84109759e-01 -5.64338386e-01
1.43635586e-01 6.53262883e-02 5.47066152e-01 -1.01915276e+00
3.37910235e-01 -8.57699454e-01 9.84914526e-02 6.42402589e-01
8.92851576e-02 -1.17336467e-01 1.81200504e-01 1.80928022e-01
-2.84371655e-02 -3.83077651e-01 9.07487929e-01 -7.18533039e-01
-1.00899398e-01 5.87551236e-01 -1.03476241e-01 2.62699217e-01
1.20389199e+00 -5.99659264e-01 2.18906313e-01 6.73952997e-02
-1.10210550e+00 3.75702679e-01 7.73529530e-01 -1.55200258e-01
7.61199474e-01 -1.14501023e+00 -4.45654273e-01 2.71783173e-01
-4.91452366e-01 1.09706450e+00 4.63393450e-01 1.16537166e+00
-9.10659194e-01 2.45111465e-01 1.99432373e-02 -1.04107249e+00
-9.97938693e-01 6.29409194e-01 9.70375896e-01 -5.06008923e-01
-1.22909439e+00 9.15518761e-01 5.97941279e-01 -9.13046479e-01
1.77604631e-02 -3.49592835e-01 -3.18186492e-01 -1.00281566e-01
1.33528590e-01 1.39414743e-01 4.69910651e-02 -4.73014563e-01
-4.11501259e-01 9.44706619e-01 1.41567126e-01 2.09573045e-01
1.44621801e+00 6.71547800e-02 -1.73581213e-01 4.46641773e-01
1.36346543e+00 -1.67649746e-01 -1.35049617e+00 -5.16876318e-02
-3.66170406e-01 -1.97763801e-01 1.54305184e-02 -4.19004917e-01
-1.42317212e+00 8.63530695e-01 2.75455177e-01 4.23672765e-01
8.85946870e-01 1.83103278e-01 1.19227695e+00 -2.72065580e-01
4.40393090e-02 -7.96321571e-01 8.80144909e-02 4.91982192e-01
6.72322333e-01 -1.09263086e+00 -2.82911118e-02 -5.80081820e-01
-3.08196723e-01 1.25693405e+00 6.61354303e-01 -2.64664739e-01
9.03629839e-01 4.68185365e-01 1.56198800e-01 -7.13337839e-01
1.99396983e-02 1.77648380e-01 3.22865695e-01 2.69074917e-01
3.80106330e-01 3.54221873e-02 -1.21603943e-01 3.94314975e-01
-1.42689362e-01 1.71071533e-02 4.22827333e-01 8.43917608e-01
-1.19269162e-01 -6.88454807e-01 -4.25812423e-01 3.39785874e-01
-7.15306818e-01 2.21156344e-01 -1.34263948e-01 1.12157869e+00
2.49767955e-02 1.00088075e-01 3.29456657e-01 4.40386124e-02
4.41947490e-01 -6.75946027e-02 4.53460515e-01 -6.64004385e-01
-5.12373030e-01 4.25876141e-01 -4.30348426e-01 -4.95124280e-01
-2.84369975e-01 -6.68123722e-01 -1.74761569e+00 -6.00588322e-02
-2.64143676e-01 2.57281102e-02 6.79160595e-01 7.71779239e-01
2.10378408e-01 3.35381627e-01 6.03962183e-01 -1.16436183e+00
-2.30807349e-01 -4.32053238e-01 -4.43193674e-01 3.43829155e-01
3.48181784e-01 -5.58077335e-01 -4.11536098e-01 -1.58745259e-01] | [14.52917194366455, -2.445671319961548] |
d2b078ba-80ad-41c4-81a5-50de7eb7ba6e | twin-contrastive-learning-with-noisy-labels | 2303.06930 | null | https://arxiv.org/abs/2303.06930v1 | https://arxiv.org/pdf/2303.06930v1.pdf | Twin Contrastive Learning with Noisy Labels | Learning from noisy data is a challenging task that significantly degenerates the model performance. In this paper, we present TCL, a novel twin contrastive learning model to learn robust representations and handle noisy labels for classification. Specifically, we construct a Gaussian mixture model (GMM) over the representations by injecting the supervised model predictions into GMM to link label-free latent variables in GMM with label-noisy annotations. Then, TCL detects the examples with wrong labels as the out-of-distribution examples by another two-component GMM, taking into account the data distribution. We further propose a cross-supervision with an entropy regularization loss that bootstraps the true targets from model predictions to handle the noisy labels. As a result, TCL can learn discriminative representations aligned with estimated labels through mixup and contrastive learning. Extensive experimental results on several standard benchmarks and real-world datasets demonstrate the superior performance of TCL. In particular, TCL achieves 7.5\% improvements on CIFAR-10 with 90\% noisy label -- an extremely noisy scenario. The source code is available at \url{https://github.com/Hzzone/TCL}. | ['Hongming Shan', 'Junping Zhang', 'Zhizhong Huang'] | 2023-03-13 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Huang_Twin_Contrastive_Learning_With_Noisy_Labels_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Huang_Twin_Contrastive_Learning_With_Noisy_Labels_CVPR_2023_paper.pdf | cvpr-2023-1 | ['learning-with-noisy-labels', 'learning-with-noisy-labels'] | ['computer-vision', 'natural-language-processing'] | [-1.08213527e-02 6.95763007e-02 -2.34198481e-01 -8.84384096e-01
-1.66432834e+00 -4.08181399e-01 3.31835240e-01 -4.09485161e-01
-2.52672434e-01 6.68545067e-01 -2.32752301e-02 -1.13786608e-01
2.02556461e-01 -2.95611203e-01 -7.45561421e-01 -9.96042669e-01
1.60052374e-01 4.04146165e-01 -1.99418932e-01 3.08745772e-01
-2.28853494e-01 -9.36697498e-02 -1.35442972e+00 5.05535305e-01
7.42767930e-01 1.28364587e+00 1.09487608e-01 3.38384390e-01
-4.27049585e-02 1.05120337e+00 -8.46282184e-01 -5.20683587e-01
1.50084391e-01 -3.97859424e-01 -6.08317673e-01 1.75242245e-01
3.48902613e-01 -9.43627506e-02 -1.77632868e-01 1.42781556e+00
5.03545046e-01 1.55370101e-01 8.03348899e-01 -1.39506876e+00
-6.95869565e-01 7.77196407e-01 -6.46034300e-01 2.27878988e-01
-1.21272959e-01 6.66024312e-02 9.21421051e-01 -1.19079208e+00
1.46345675e-01 1.59365225e+00 6.69474959e-01 8.20018172e-01
-1.37700951e+00 -1.28074944e+00 4.57536995e-01 1.25337511e-01
-1.49598622e+00 -4.68123615e-01 7.39481032e-01 -4.32358205e-01
4.25929070e-01 2.33877942e-01 -5.15943617e-02 1.61222541e+00
-1.65962800e-01 1.01344693e+00 1.34891820e+00 -1.90289617e-01
2.17857361e-01 2.69730240e-01 5.35840929e-01 3.40386242e-01
-9.55864787e-02 7.55637437e-02 -3.41386616e-01 -3.54912609e-01
1.82211652e-01 1.54409828e-02 -3.71057004e-01 2.48729307e-02
-8.03188026e-01 8.66365194e-01 3.69517326e-01 -2.08180677e-02
2.74992865e-02 3.39458525e-01 4.62237805e-01 6.62472472e-02
1.06204283e+00 -1.13789655e-01 -5.78746617e-01 1.24359511e-01
-8.23971748e-01 5.94414538e-03 4.93323028e-01 1.09563911e+00
7.23298013e-01 2.61271060e-01 -2.68780947e-01 1.30542493e+00
6.53075576e-01 7.54350603e-01 7.27582872e-01 -1.00220358e+00
4.79150295e-01 1.81888312e-01 -5.88213243e-02 -5.25118709e-01
-1.13859266e-01 -9.18660045e-01 -1.01447034e+00 -1.02193981e-01
4.19605821e-02 -1.55504212e-01 -1.03533328e+00 1.81570506e+00
3.97478908e-01 7.59742081e-01 1.08352236e-01 8.05451930e-01
9.88819659e-01 7.59017706e-01 4.31345254e-01 -2.33022973e-01
1.19201636e+00 -1.00523782e+00 -9.50234950e-01 -2.81468362e-01
8.47299218e-01 -7.99642980e-01 1.03976822e+00 5.84105372e-01
-5.73745191e-01 -7.07004428e-01 -7.12953091e-01 8.41472447e-02
-1.97755978e-01 5.89068174e-01 1.11082934e-01 6.19167924e-01
-6.18514657e-01 4.01532173e-01 -7.84522414e-01 4.04802650e-01
6.30125523e-01 9.63792130e-02 -1.94199592e-01 -3.41941029e-01
-1.12034559e+00 4.38344032e-01 6.22376084e-01 1.64715216e-01
-1.16091394e+00 -4.57129270e-01 -9.92673337e-01 -9.10880119e-02
3.83263677e-01 4.74929204e-03 1.66838348e+00 -9.67553437e-01
-1.21740234e+00 8.74894977e-01 -1.94120750e-01 -4.08601969e-01
5.15227318e-01 -4.71747428e-01 -5.75079560e-01 -3.18396330e-01
4.30724353e-01 6.34435475e-01 8.21395099e-01 -1.73175168e+00
-5.69392681e-01 -2.22607583e-01 -5.05564809e-01 -7.79198855e-02
-8.11557546e-02 5.42948535e-03 -3.90174687e-01 -8.41114044e-01
3.94442439e-01 -9.74011004e-01 -8.94595608e-02 -5.35198748e-01
-7.29844689e-01 -5.84503949e-01 8.19479287e-01 -6.52387917e-01
1.08699012e+00 -2.44152784e+00 -3.57169509e-01 6.92066550e-02
1.49627745e-01 2.51118213e-01 -1.50287777e-01 -1.48620933e-01
-3.00924838e-01 2.23731577e-01 -3.36105287e-01 -9.54071283e-01
1.18982702e-01 3.39854091e-01 -7.53582478e-01 5.78302324e-01
2.49532670e-01 5.40387034e-01 -9.73901272e-01 -4.35371459e-01
1.87788326e-02 6.26774251e-01 -1.36107773e-01 4.91402596e-01
-1.59987822e-01 6.71472192e-01 -3.42403471e-01 5.76981425e-01
1.08519423e+00 -2.04744220e-01 1.18686147e-01 -1.25417620e-01
4.69490588e-01 4.03646946e-01 -1.40079737e+00 1.35072541e+00
-4.03797865e-01 1.71819940e-01 -7.97814429e-02 -1.07762599e+00
1.13333189e+00 3.81203771e-01 1.48041323e-01 -4.51111346e-01
2.48803064e-01 3.54969561e-01 -4.94475245e-01 -3.87282223e-01
1.28902510e-01 -7.73042887e-02 -3.05400252e-01 1.92131281e-01
4.87665117e-01 -9.93781462e-02 -1.21176593e-01 8.84790495e-02
5.92769444e-01 3.35651726e-01 9.06047523e-02 3.66122142e-04
2.09862158e-01 -4.60981935e-01 1.23520064e+00 8.55947912e-01
-2.85621285e-01 7.83552468e-01 4.13740218e-01 -1.76371574e-01
-5.42019367e-01 -1.12746489e+00 -3.57684702e-01 1.28022242e+00
-9.04446561e-03 -3.13592523e-01 -8.03055584e-01 -1.22203302e+00
-2.28026673e-01 1.01601923e+00 -5.05474329e-01 -3.47059995e-01
-2.87326664e-01 -1.08212876e+00 5.12947917e-01 3.91273975e-01
5.05705535e-01 -8.35299671e-01 1.81634337e-01 5.34896851e-02
-4.21383411e-01 -1.10878479e+00 -3.05176944e-01 7.05226839e-01
-6.12109423e-01 -9.56004143e-01 -4.60845321e-01 -8.50643516e-01
6.17851138e-01 2.80324638e-01 1.13818431e+00 -2.08523974e-01
9.89833847e-02 5.48693538e-02 -5.78985572e-01 -3.43619883e-01
-5.37471235e-01 -1.34622723e-01 1.32678092e-01 1.01163253e-01
5.54568946e-01 -3.48207742e-01 -3.04206043e-01 5.34039021e-01
-9.39260304e-01 -2.70719349e-01 2.92382658e-01 1.09231877e+00
1.01443267e+00 1.00055106e-01 8.35637629e-01 -1.24027288e+00
2.46018872e-01 -1.14701998e+00 -7.34897494e-01 8.67944807e-02
-4.64517117e-01 -9.97623894e-03 5.78359842e-01 -6.24525130e-01
-1.17806315e+00 2.34298602e-01 -2.94169664e-01 -7.51472950e-01
-5.06707489e-01 1.23673871e-01 -5.45866668e-01 5.70818782e-01
6.19442344e-01 2.12148115e-01 -5.14685869e-01 -1.02752650e+00
4.22532111e-01 1.11919010e+00 6.01294219e-01 -6.47091389e-01
6.11981750e-01 2.12031350e-01 -6.20699108e-01 -3.13045174e-01
-1.59775972e+00 -6.55349374e-01 -3.98563445e-01 2.66176891e-02
6.08664155e-01 -1.43871951e+00 -1.75601915e-01 4.92834240e-01
-1.05277562e+00 -3.93431544e-01 -2.68091977e-01 6.92266822e-01
-3.57974768e-01 2.57961839e-01 -6.98921084e-01 -1.11467147e+00
-1.53318346e-01 -1.39807701e+00 1.31831157e+00 1.70338660e-01
8.22598413e-02 -7.31655359e-01 7.70956129e-02 5.69307268e-01
1.66767649e-02 -1.02234095e-01 4.76972520e-01 -1.16613209e+00
-2.57576913e-01 -1.97863728e-01 -1.10760599e-01 1.08917677e+00
-3.70599739e-02 -1.84930846e-01 -1.75819612e+00 -3.47837359e-01
2.11675733e-01 -7.71260381e-01 1.19916058e+00 3.21449757e-01
1.71743619e+00 -2.28079230e-01 -3.68104577e-01 7.12347806e-01
1.17759681e+00 2.08233632e-02 3.22748423e-01 -1.63884610e-02
7.53057837e-01 3.75609100e-01 7.71215200e-01 3.59485567e-01
2.44242787e-01 5.63642204e-01 4.35401201e-01 9.25525501e-02
-2.22507536e-01 -4.91080195e-01 4.34451640e-01 1.10954809e+00
5.06676733e-01 -3.44985664e-01 -8.61579657e-01 3.41385692e-01
-1.87233484e+00 -6.23148263e-01 -2.79607534e-01 2.13556004e+00
9.92143393e-01 9.23405066e-02 -2.07698241e-01 1.24686070e-01
1.04788697e+00 4.15246263e-02 -6.02522969e-01 2.15820462e-01
-1.33795306e-01 5.15912138e-02 3.76810789e-01 5.04738033e-01
-1.72196031e+00 9.53277886e-01 5.50132895e+00 1.48758459e+00
-8.85755241e-01 7.40434706e-01 1.28700650e+00 -1.13396853e-01
-1.09281465e-01 -1.79143518e-01 -1.14096498e+00 9.63546753e-01
9.98214722e-01 3.70125622e-01 -2.63335742e-02 1.35341072e+00
1.80812374e-01 9.44064781e-02 -8.68909538e-01 1.14249563e+00
7.55346119e-02 -7.12305665e-01 -2.36202642e-01 2.68238578e-02
9.37146544e-01 3.23670387e-01 3.70696008e-01 8.97402883e-01
7.59958684e-01 -9.41950262e-01 8.76201630e-01 4.90824282e-01
7.50356019e-01 -6.93041325e-01 1.04969370e+00 5.76385140e-01
-9.33085918e-01 -1.48441777e-01 -6.19615793e-01 2.96289772e-01
-8.78294557e-02 1.16123879e+00 -7.09325552e-01 3.50270510e-01
1.02788460e+00 7.21017063e-01 -5.05990505e-01 1.01493812e+00
-5.77126205e-01 1.41290700e+00 -1.26557350e-01 4.02882785e-01
-5.91800595e-03 -2.17443198e-01 2.35555097e-01 1.47806561e+00
2.83080101e-01 -1.83918983e-01 5.65573156e-01 9.11621392e-01
-5.19677341e-01 2.20639203e-02 -4.76942569e-01 4.70772862e-01
7.69794762e-01 1.25186372e+00 -4.73640412e-01 -5.57953000e-01
-1.70933872e-01 8.22753727e-01 6.18695736e-01 5.29822648e-01
-1.11100721e+00 -5.88965677e-02 5.32895446e-01 -3.60437423e-01
1.60491869e-01 1.60318762e-01 -7.11746216e-02 -1.27612150e+00
7.58756846e-02 -8.65606308e-01 4.53857303e-01 -6.25613570e-01
-1.83328283e+00 7.62932420e-01 1.83286872e-02 -1.47901714e+00
-7.87598863e-02 -3.83046418e-01 -3.92922342e-01 7.78523564e-01
-1.68520546e+00 -9.81613219e-01 -1.25302806e-01 3.40766609e-01
7.96645343e-01 -9.02658701e-02 8.20447922e-01 4.47424412e-01
-8.53813827e-01 8.79781961e-01 5.97796082e-01 4.15439397e-01
9.20390725e-01 -1.36031330e+00 2.51767427e-01 6.78873122e-01
2.92392939e-01 2.13449731e-01 3.53159428e-01 -5.97763479e-01
-4.06793445e-01 -1.67734277e+00 6.38472974e-01 -4.73254204e-01
4.49813515e-01 -5.79176247e-01 -1.22271442e+00 9.09085453e-01
-3.05323508e-02 5.42675734e-01 9.30399597e-01 -1.43219352e-01
-8.91231239e-01 -1.28924727e-01 -1.13402593e+00 -1.11949872e-02
7.70607054e-01 -5.52403271e-01 -4.91193444e-01 7.90187240e-01
9.55451131e-01 -3.57150406e-01 -6.38682842e-01 3.87823403e-01
-5.92112429e-02 -7.74424136e-01 7.66870677e-01 -5.12687027e-01
5.37610091e-02 -3.29368323e-01 -5.60930252e-01 -1.44393730e+00
-2.60379463e-01 -2.48882681e-01 -1.41112119e-01 1.72435641e+00
4.38807338e-01 -6.11995578e-01 4.56421316e-01 2.47611508e-01
-2.32772768e-01 -7.11056709e-01 -1.14662302e+00 -9.75284874e-01
1.51891351e-01 -8.83581340e-01 4.78236079e-01 1.05638325e+00
-4.64169621e-01 2.47092053e-01 -5.62065899e-01 4.56940562e-01
8.65104616e-01 -2.74129927e-01 5.34612477e-01 -1.12280500e+00
-3.13380867e-01 -2.05795448e-02 -2.12141737e-01 -1.17480958e+00
7.56014645e-01 -1.16030490e+00 5.01566410e-01 -9.75545347e-01
3.91352862e-01 -6.70582771e-01 -6.54109180e-01 6.52207255e-01
-4.90976363e-01 4.30000603e-01 1.33539960e-01 4.24445391e-01
-1.08506000e+00 8.76910269e-01 7.58518636e-01 -2.73301184e-01
2.70202100e-01 3.03307474e-01 -6.46749794e-01 9.21614289e-01
1.15323091e+00 -1.08515608e+00 -3.98671120e-01 -3.52377117e-01
-1.64643198e-01 -4.49972957e-01 3.08957130e-01 -1.00230706e+00
-3.76915067e-01 2.16401424e-02 4.11622107e-01 -4.92110431e-01
5.38397014e-01 -6.44022048e-01 -8.28971192e-02 1.56031147e-01
-6.81694567e-01 -4.37450975e-01 -1.04884200e-01 8.17039847e-01
-3.93388391e-01 -4.53981400e-01 1.03164387e+00 -4.97942083e-02
-4.55001265e-01 1.37273476e-01 -1.29105672e-01 4.21996355e-01
8.09308708e-01 4.92558688e-01 -3.82179618e-01 -3.40352386e-01
-1.01265740e+00 3.14887017e-01 -6.24448471e-02 2.91301131e-01
5.22379279e-01 -1.55710936e+00 -8.03713441e-01 1.55273214e-01
2.49047175e-01 3.01125169e-01 3.19188774e-01 5.81153512e-01
2.00084746e-01 -2.93503031e-02 5.78551412e-01 -8.91304255e-01
-1.13615668e+00 4.43545312e-01 5.14901519e-01 -2.82385319e-01
-2.51276910e-01 1.17517519e+00 4.08339560e-01 -1.05446243e+00
5.54825246e-01 -2.75529742e-01 -1.63306609e-01 -4.45306003e-02
5.92164874e-01 2.44543955e-01 5.60527034e-02 -8.14678729e-01
-2.12959737e-01 3.11286628e-01 -1.19733244e-01 1.49279341e-01
1.18233788e+00 -2.24036962e-01 9.86682400e-02 7.70713508e-01
1.48051751e+00 -2.20127776e-01 -1.56865382e+00 -6.64310455e-01
1.70316741e-01 -4.61219430e-01 1.34217888e-01 -8.17686856e-01
-1.26415515e+00 9.35020506e-01 9.65702951e-01 6.21132962e-02
9.68728065e-01 3.49323481e-01 6.22405350e-01 2.31125683e-01
1.88392729e-01 -1.13177669e+00 1.79629400e-01 5.33861160e-01
6.59636438e-01 -1.60358787e+00 -4.32975471e-01 -5.10224819e-01
-8.57179880e-01 6.67681038e-01 7.35340059e-01 -1.72691479e-01
8.77573669e-01 1.45608842e-01 4.88897771e-01 8.25099424e-02
-6.62767291e-01 -1.88768521e-01 1.32037520e-01 6.79048657e-01
3.96534353e-01 4.08751070e-01 1.13513596e-01 1.12107027e+00
-1.39977396e-01 -2.34963447e-01 3.08017045e-01 5.55467665e-01
-4.61386323e-01 -9.44560647e-01 -5.67823172e-01 4.32173610e-01
-5.88888109e-01 -1.02376305e-01 -7.23972768e-02 3.55296254e-01
6.09320760e-01 1.35823369e+00 -7.96284005e-02 -4.35404062e-01
2.23843753e-01 4.88748908e-01 -1.38101533e-01 -8.43609869e-01
-3.37775707e-01 5.70692301e-01 -1.16947494e-01 -5.00866413e-01
-2.05256224e-01 -4.91719365e-01 -1.09535682e+00 2.64084429e-01
-5.58512151e-01 3.61416340e-01 6.14082158e-01 8.05784404e-01
3.75226915e-01 5.90578973e-01 9.65440691e-01 -8.07099819e-01
-9.71161127e-01 -1.43451643e+00 -8.08020830e-01 6.17486000e-01
3.60018015e-01 -8.35981667e-01 -8.32851529e-01 1.85854897e-01] | [9.405890464782715, 3.897252082824707] |
d86ff52f-4f19-4160-ab88-ea706c30e51f | coordx-accelerating-implicit-neural-1 | 2201.12425 | null | https://arxiv.org/abs/2201.12425v1 | https://arxiv.org/pdf/2201.12425v1.pdf | CoordX: Accelerating Implicit Neural Representation with a Split MLP Architecture | Implicit neural representations with multi-layer perceptrons (MLPs) have recently gained prominence for a wide variety of tasks such as novel view synthesis and 3D object representation and rendering. However, a significant challenge with these representations is that both training and inference with an MLP over a large number of input coordinates to learn and represent an image, video, or 3D object, require large amounts of computation and incur long processing times. In this work, we aim to accelerate inference and training of coordinate-based MLPs for implicit neural representations by proposing a new split MLP architecture, CoordX. With CoordX, the initial layers are split to learn each dimension of the input coordinates separately. The intermediate features are then fused by the last layers to generate the learned signal at the corresponding coordinate point. This significantly reduces the amount of computation required and leads to large speedups in training and inference, while achieving similar accuracy as the baseline MLP. This approach thus aims at first learning functions that are a decomposition of the original signal and then fusing them to generate the learned signal. Our proposed architecture can be generally used for many implicit neural representation tasks with no additional memory overheads. We demonstrate a speedup of up to 2.92x compared to the baseline model for image, video, and 3D shape representation and rendering tasks. | ['Nandita Vijaykumar', 'Hongyi Sun', 'Ruofan Liang'] | 2022-01-28 | coordx-accelerating-implicit-neural | https://openreview.net/forum?id=oAy7yPmdNz | https://openreview.net/pdf?id=oAy7yPmdNz | iclr-2022-4 | ['3d-shape-representation'] | ['computer-vision'] | [ 3.98259640e-01 1.27527297e-01 7.51577318e-02 -3.60554039e-01
-7.28033245e-01 -2.24551320e-01 7.01330721e-01 2.01050262e-03
-1.56686202e-01 4.87809449e-01 -3.29888165e-02 -2.25177079e-01
2.47330904e-01 -8.62810910e-01 -1.14572167e+00 -8.29464257e-01
8.37707072e-02 3.03879380e-01 -2.51625534e-02 2.20886827e-01
2.20614374e-01 9.59299862e-01 -1.75556231e+00 5.61171412e-01
3.30719560e-01 1.44289589e+00 8.30146074e-02 8.63785446e-01
-1.55151621e-01 9.44607675e-01 -7.56319523e-01 -1.10638633e-01
1.83500201e-01 1.17778271e-01 -5.07676482e-01 -2.64957640e-02
6.94759369e-01 -5.55171371e-01 -2.72437364e-01 7.36720383e-01
6.34433448e-01 2.98182636e-01 6.57673120e-01 -9.33816195e-01
-6.27004147e-01 2.36275986e-01 -7.12304115e-01 -2.40380436e-01
9.79743153e-02 -1.79795921e-02 9.18404818e-01 -1.08413684e+00
2.87284344e-01 1.47581708e+00 8.12268794e-01 5.09032428e-01
-1.52185059e+00 -5.72534740e-01 3.30255181e-01 -6.61806986e-02
-1.16521871e+00 -4.93393928e-01 7.84703434e-01 -2.95466423e-01
1.27776134e+00 2.47138694e-01 5.74806869e-01 9.15489316e-01
-9.79569182e-02 8.74012113e-01 8.03257167e-01 -3.34694684e-01
5.06167635e-02 1.79382171e-02 -4.17341106e-02 8.37161183e-01
-2.39767581e-01 -5.21984287e-02 -3.67700219e-01 -7.01873451e-02
1.32408631e+00 2.94889480e-01 -2.59521484e-01 -9.25560594e-02
-1.15604067e+00 9.26815450e-01 8.79434049e-01 -1.72677711e-01
-6.91008031e-01 8.35368276e-01 4.52206284e-01 1.53519601e-01
4.96476084e-01 2.94602096e-01 -4.18278068e-01 1.80515945e-02
-8.44559014e-01 3.05239886e-01 5.59095919e-01 8.58805060e-01
9.75720584e-01 3.50629151e-01 1.35151818e-01 9.70187128e-01
2.48107880e-01 4.51398253e-01 3.20670903e-01 -1.26394260e+00
5.24421215e-01 5.83820999e-01 -1.96017861e-01 -1.25706804e+00
-3.30176085e-01 -4.53260601e-01 -1.14535892e+00 6.56421304e-01
4.71059494e-02 -1.63719893e-01 -9.25039113e-01 1.49414480e+00
1.66104212e-01 4.80111837e-01 4.60926034e-02 7.47072279e-01
9.23584282e-01 1.26738405e+00 -2.34324589e-01 1.63364306e-01
1.19535625e+00 -8.88947487e-01 -2.33172596e-01 -1.44449383e-01
2.57351935e-01 -5.58851659e-01 7.99828708e-01 2.14863062e-01
-1.34458137e+00 -8.90842080e-01 -1.28997743e+00 -4.23500746e-01
-2.56287068e-01 4.19099301e-01 6.81544662e-01 1.37561455e-01
-1.19264901e+00 7.03317285e-01 -9.01025891e-01 3.44353139e-01
4.26541805e-01 6.88224673e-01 -1.22208856e-01 2.81629264e-01
-7.14151859e-01 5.23383617e-01 3.30321014e-01 1.29401848e-01
-4.04897064e-01 -9.48382318e-01 -9.48535264e-01 3.30376625e-01
2.31103273e-03 -8.90443742e-01 1.15606046e+00 -1.18577230e+00
-1.73860168e+00 5.85984647e-01 -2.06014842e-01 -4.67591286e-01
1.56270117e-01 -3.46848696e-01 5.75441308e-02 1.46433473e-01
-1.34969339e-01 1.10598850e+00 1.13282204e+00 -1.26378012e+00
-4.73905623e-01 -2.41554350e-01 2.56596297e-01 2.46240929e-01
9.33787599e-02 -3.26405555e-01 -6.31490529e-01 -8.17809105e-01
3.68814111e-01 -9.18860018e-01 -1.89620256e-01 3.70199561e-01
-1.08173043e-01 -2.29679793e-01 8.52128208e-01 -5.53434312e-01
5.48364997e-01 -2.37084794e+00 2.86949337e-01 1.21614948e-01
2.91384667e-01 1.04825191e-01 -1.06878616e-01 1.03741372e-02
-1.46255776e-01 -1.90675005e-01 -8.79006013e-02 -6.95961595e-01
-1.91065982e-01 4.04158443e-01 -6.97540402e-01 2.22117111e-01
6.49885058e-01 8.98465216e-01 -6.28539145e-01 -1.17894061e-01
2.67193317e-01 9.40590322e-01 -9.46040869e-01 1.86081827e-01
-4.20317560e-01 4.82792318e-01 -2.11102232e-01 3.96282017e-01
5.56489408e-01 -5.03530383e-01 5.35394177e-02 -3.71743888e-01
-5.61064743e-02 4.37490463e-01 -1.13991797e+00 1.72665310e+00
-9.28688169e-01 9.31884646e-01 -3.45097221e-02 -1.26781392e+00
1.11520827e+00 3.67688835e-01 3.35307598e-01 -5.67515075e-01
9.55057740e-02 1.27011403e-01 -3.93049061e-01 -1.56370476e-01
4.67612475e-01 -8.25947076e-02 -7.44343847e-02 4.04115438e-01
1.88133642e-01 -2.08830491e-01 -3.04197252e-01 -1.69226691e-01
7.35458195e-01 4.22619730e-01 1.27137341e-02 2.45326474e-01
5.98341763e-01 -3.39604259e-01 4.85478610e-01 4.17255282e-01
4.79824692e-01 7.27241516e-01 6.64301753e-01 -7.85412610e-01
-1.07661700e+00 -9.41608548e-01 1.11088432e-01 1.04101098e+00
-6.31338730e-02 -1.89424679e-01 -3.73693377e-01 -3.38396847e-01
1.52871042e-01 5.94181418e-01 -5.19082308e-01 6.47075474e-04
-1.11403739e+00 -4.52558964e-01 5.37087440e-01 9.31080639e-01
4.71143246e-01 -1.08378041e+00 -8.10620189e-01 2.11385339e-01
6.67616501e-02 -1.09785593e+00 -1.99119821e-01 2.36296222e-01
-1.31241786e+00 -7.34777868e-01 -7.78999746e-01 -7.84013808e-01
9.56225455e-01 9.97380614e-02 9.42678750e-01 -1.83460508e-02
-2.34576359e-01 1.35472074e-01 1.64400384e-01 -2.63101637e-01
-2.83091098e-01 -1.19850330e-01 -6.72392398e-02 1.44637078e-01
-1.20691814e-01 -9.21707869e-01 -4.72482264e-01 -1.13378540e-01
-6.78387165e-01 3.96315306e-01 8.10118675e-01 9.96934831e-01
9.65301275e-01 -4.01103407e-01 4.75846320e-01 -8.70171607e-01
4.74347711e-01 -2.05567271e-01 -7.21891999e-01 -1.15090817e-01
-7.17135817e-02 4.09224302e-01 8.75964999e-01 -5.82792699e-01
-8.91509712e-01 1.52701616e-01 -3.71465802e-01 -8.94908547e-01
-1.24037571e-01 4.27660793e-01 2.75047779e-01 -8.41679201e-02
4.96581972e-01 1.88756227e-01 1.45081386e-01 -7.69145370e-01
5.77409804e-01 3.96872848e-01 6.13112688e-01 -5.32981992e-01
5.60416162e-01 2.21135840e-01 2.35901624e-01 -9.17750716e-01
-7.02889204e-01 -1.71427429e-01 -6.50192738e-01 1.02886878e-01
6.83371723e-01 -1.07546353e+00 -1.01583803e+00 2.91624874e-01
-1.58686781e+00 -2.00549230e-01 -3.34212750e-01 4.79782850e-01
-6.50841057e-01 5.56595847e-02 -6.49816811e-01 -5.72300613e-01
-5.16315281e-01 -1.32073903e+00 1.19020545e+00 2.21568663e-02
-1.87086120e-01 -8.46961260e-01 -1.03117466e-01 5.59416264e-02
4.42989826e-01 5.29026806e-01 1.27245104e+00 -5.63360490e-02
-9.60004270e-01 -2.83317506e-01 -6.18471563e-01 5.16232550e-01
-1.46093294e-01 -5.16974144e-02 -1.15617728e+00 -1.32057920e-01
4.17909399e-02 -4.74248141e-01 8.61944973e-01 4.12007779e-01
1.53853250e+00 -5.49340069e-01 -9.61975753e-02 1.07977748e+00
1.41816306e+00 1.54136717e-01 3.34755063e-01 2.00627930e-02
7.26065457e-01 3.03189605e-01 -4.66486551e-02 3.96667063e-01
1.30845070e-01 6.18442774e-01 4.63142157e-01 -2.52589732e-01
-1.81945696e-01 -9.67018977e-02 2.37963393e-01 1.01776230e+00
-5.33319712e-01 2.86289126e-01 -4.80731219e-01 1.43033817e-01
-1.66664004e+00 -8.61214757e-01 3.32878739e-01 2.02126431e+00
6.95420027e-01 1.26822263e-01 -1.57664001e-01 2.52581865e-01
3.35684240e-01 3.85104030e-01 -6.59867167e-01 -7.10145414e-01
1.86152712e-01 5.28602540e-01 2.72871315e-01 4.30978656e-01
-1.02351892e+00 6.21328890e-01 6.12113523e+00 4.68978137e-01
-1.71792901e+00 -1.59627452e-01 6.29189134e-01 -2.52372265e-01
6.06001727e-02 -2.93159842e-01 -6.99798107e-01 9.58174467e-02
8.47865820e-01 2.41999626e-01 5.43624282e-01 9.55919266e-01
-2.57058125e-02 2.75825799e-01 -1.34029663e+00 1.39949334e+00
1.72611952e-01 -1.68365514e+00 4.30177569e-01 -1.28599405e-01
7.21165299e-01 8.59647095e-02 2.18183547e-01 3.86253387e-01
-1.31884784e-01 -1.04673040e+00 7.49803722e-01 6.09918416e-01
1.01718247e+00 -7.85602450e-01 3.99127394e-01 3.48949075e-01
-1.11165500e+00 -2.95511708e-02 -5.07329106e-01 -1.20959498e-01
-5.05779358e-03 3.30046237e-01 -7.33348012e-01 2.95883536e-01
5.52650094e-01 7.89549768e-01 -9.72950086e-02 7.08216310e-01
-2.59835780e-01 2.70443708e-01 -3.70359242e-01 -5.75311556e-02
4.37892854e-01 -8.44765306e-02 3.77574623e-01 1.03743851e+00
3.68619859e-01 1.24829136e-01 9.71548930e-02 1.02982533e+00
-4.46341664e-01 -1.82523653e-01 -5.21245539e-01 1.88471973e-01
1.85400695e-01 1.04915237e+00 -3.87415618e-01 -5.65500915e-01
-5.89511514e-01 9.17037785e-01 5.45031548e-01 5.58915973e-01
-8.80111873e-01 -5.07547319e-01 7.91229844e-01 6.53138980e-02
5.76527894e-01 -4.97015506e-01 -4.55108136e-01 -9.55001056e-01
2.42435690e-02 -5.48344016e-01 -7.47828558e-02 -7.94145286e-01
-8.32943916e-01 7.84435868e-01 -9.30070281e-02 -1.21385825e+00
-5.86451530e-01 -9.46737826e-01 -5.38091838e-01 1.02695990e+00
-1.50931692e+00 -1.07053185e+00 -3.74053031e-01 4.05998468e-01
6.46069169e-01 -7.24112466e-02 9.49713409e-01 3.20488572e-01
-2.59063333e-01 4.64659244e-01 -4.32198569e-02 2.15356424e-01
2.72826225e-01 -1.10018110e+00 6.35255814e-01 2.35939443e-01
1.95691690e-01 6.21296763e-01 2.08727777e-01 -1.68629318e-01
-1.56167722e+00 -1.20853627e+00 4.56575334e-01 -7.40813464e-02
1.79391071e-01 -4.18812335e-01 -8.66790712e-01 8.81069541e-01
-6.30316064e-02 2.53109843e-01 4.39016044e-01 6.14185780e-02
-5.32488644e-01 -3.22491258e-01 -7.86985099e-01 5.70763588e-01
7.30597436e-01 -6.88086212e-01 -4.63080108e-01 5.95052540e-02
7.57250428e-01 -6.62144959e-01 -8.74224484e-01 3.44234347e-01
7.30270565e-01 -8.74548495e-01 1.35176480e+00 -2.98825860e-01
5.75748503e-01 -1.77993029e-01 -2.03816161e-01 -1.13629508e+00
-3.78707886e-01 -3.48194838e-01 -5.04089832e-01 7.02415109e-01
4.73697007e-01 -7.35379755e-01 6.91940427e-01 3.20652634e-01
-2.42706478e-01 -1.14900827e+00 -8.01546514e-01 -2.15123206e-01
-1.01727113e-01 -4.30634171e-01 6.33760750e-01 6.71961248e-01
-5.33085883e-01 4.93458152e-01 -3.15775394e-01 3.37624699e-01
5.74404359e-01 4.69093233e-01 8.62121165e-01 -1.20611763e+00
-5.30326843e-01 -5.03700376e-01 -5.24378479e-01 -1.77907217e+00
1.16188064e-01 -9.66710627e-01 -4.75999564e-02 -1.44832611e+00
-2.64651626e-01 -7.61922896e-01 -3.62918414e-02 5.60451865e-01
4.37703058e-02 3.71500134e-01 3.46012682e-01 2.65927017e-01
-1.39547989e-01 5.95280051e-01 1.22747612e+00 -2.97844410e-01
-2.97422051e-01 1.88010752e-01 -5.04998386e-01 8.64356637e-01
5.67405581e-01 -1.71322018e-01 -4.00269061e-01 -9.03537214e-01
1.44532174e-01 3.26510668e-01 5.79536080e-01 -1.15719533e+00
1.83040679e-01 2.07025141e-01 9.27661538e-01 -9.50674236e-01
1.09533203e+00 -6.96600139e-01 1.51997298e-01 3.43339801e-01
-3.27784568e-01 1.15024731e-01 4.34279472e-01 3.88013244e-01
-3.71563494e-01 -1.43145740e-01 7.71331072e-01 -2.45665759e-01
-3.31159115e-01 3.59714508e-01 -6.76113069e-02 -3.40381563e-01
7.29702294e-01 -2.93073863e-01 3.60586606e-02 -3.49842429e-01
-8.54445815e-01 -1.34943709e-01 1.28102243e-01 2.97528356e-01
9.47052538e-01 -1.43792927e+00 -6.42505527e-01 6.53609216e-01
-4.16826665e-01 5.77656507e-01 7.91908354e-02 5.22912860e-01
-7.89895892e-01 5.06532550e-01 -2.34105781e-01 -9.32876110e-01
-1.06119573e+00 3.29973459e-01 2.60706514e-01 -1.84913978e-01
-1.03369486e+00 9.34299111e-01 4.68564451e-01 -4.20499325e-01
3.97802800e-01 -8.31366360e-01 -1.06797650e-01 -5.64241633e-02
5.66277444e-01 3.27082604e-01 -1.16349816e-01 -6.93036258e-01
5.20969853e-02 8.99372220e-01 3.82474400e-02 4.16906327e-02
1.62027514e+00 3.63616556e-01 -1.27851203e-01 5.91539383e-01
1.62003791e+00 -1.93262279e-01 -1.49960589e+00 -2.56646931e-01
-4.03118730e-01 -2.88925111e-01 1.39687046e-01 -3.80611986e-01
-1.28696859e+00 1.28066111e+00 3.25899988e-01 6.55626878e-02
1.10902214e+00 -1.44835413e-01 7.34438121e-01 6.43806517e-01
1.68863043e-01 -6.55291200e-01 2.06632122e-01 6.73186541e-01
1.18631697e+00 -7.96102881e-01 1.18395485e-01 -3.33353281e-01
-3.65206927e-01 1.35736358e+00 4.13372785e-01 -5.43906808e-01
6.05001569e-01 2.78589636e-01 -3.39129269e-02 -9.46630538e-02
-8.62044454e-01 2.86751837e-01 4.74373281e-01 3.72974247e-01
4.74026024e-01 2.66467389e-02 4.86066997e-01 2.80166388e-01
-1.27375439e-01 -1.49070978e-01 7.18606263e-02 7.86791503e-01
-2.02127367e-01 -8.85926604e-01 -3.24580908e-01 4.91742581e-01
-3.54415119e-01 -3.02472356e-04 8.61946121e-02 5.32418251e-01
4.54325825e-02 1.53886765e-01 4.80195135e-01 -1.77501470e-01
3.59744281e-01 8.69071558e-02 6.28710508e-01 -6.47963107e-01
-5.38416207e-01 1.02820039e-01 -1.32239625e-01 -5.68019867e-01
-3.62352878e-01 -3.82173091e-01 -1.46710920e+00 -1.53172314e-01
-1.21578589e-01 -1.96525231e-01 8.81201923e-01 5.65222859e-01
6.15925074e-01 7.68298566e-01 6.59649670e-01 -1.62898886e+00
-6.10762954e-01 -7.59561121e-01 -1.29574567e-01 1.79950640e-01
6.23547733e-01 -6.94791377e-01 -1.32353052e-01 1.65272042e-01] | [11.086041450500488, -1.822986364364624] |
6b1c0674-ba5b-403d-aa09-d69cbed2b2bc | deepimagespam-deep-learning-based-image-spam | 1810.03977 | null | http://arxiv.org/abs/1810.03977v1 | http://arxiv.org/pdf/1810.03977v1.pdf | DeepImageSpam: Deep Learning based Image Spam Detection | Hackers and spammers are employing innovative and novel techniques to deceive
novice and even knowledgeable internet users. Image spam is one of such
technique where the spammer varies and changes some portion of the image such
that it is indistinguishable from the original image fooling the users. This
paper proposes a deep learning based approach for image spam detection using
the convolutional neural networks which uses a dataset with 810 natural images
and 928 spam images for classification achieving an accuracy of 91.7%
outperforming the existing image processing and machine learning techniques | ['Soman Kp', 'Amara Dinesh Kumar', 'Vinayakumar R'] | 2018-10-03 | null | null | null | null | ['spam-detection'] | ['natural-language-processing'] | [ 4.47577648e-02 -2.86522537e-01 4.69265819e-01 -3.14751491e-02
3.05459172e-01 -9.02342498e-01 5.39288759e-01 1.39284924e-01
-4.00688112e-01 5.94559789e-01 -4.34391856e-01 -6.35919750e-01
3.66097033e-01 -8.77853394e-01 -6.76587701e-01 -5.15786052e-01
-2.87003126e-02 2.45982707e-01 5.44352829e-01 -2.88879097e-01
9.35420275e-01 6.89049661e-01 -1.27911186e+00 6.82311714e-01
7.71639705e-01 1.11180961e+00 -3.61977667e-01 9.76532340e-01
-2.20714331e-01 7.56067634e-01 -1.11604917e+00 -4.63384807e-01
7.10521996e-01 -2.80625135e-01 -6.83982790e-01 4.24051881e-02
1.19152033e+00 -5.44632494e-01 -7.18730271e-01 1.65402496e+00
2.81145036e-01 5.91595881e-02 4.35452998e-01 -1.27030408e+00
-1.47242761e+00 -5.68734407e-02 -5.93655169e-01 7.17536330e-01
2.92811692e-01 2.90667117e-01 -7.78391287e-02 -5.85894883e-01
4.52129722e-01 1.54372954e+00 7.38851845e-01 5.31713843e-01
-9.34466541e-01 -1.09577668e+00 -2.11729601e-01 4.05235648e-01
-1.02082717e+00 -6.75919652e-02 4.40797985e-01 -2.65820771e-01
7.55728066e-01 1.41872436e-01 5.00932157e-01 1.01000190e+00
6.58237875e-01 5.96134603e-01 1.10911226e+00 -2.67511725e-01
6.39148336e-03 3.90143186e-01 5.46906412e-01 6.86431527e-01
8.64191294e-01 3.57003152e-01 -2.38723397e-01 -4.61946100e-01
4.75290120e-01 2.38868296e-01 -4.74676192e-02 -1.30284563e-01
-5.79819322e-01 8.52906466e-01 1.07046318e+00 2.88771868e-01
-9.24979523e-02 3.79322976e-01 7.67262638e-01 7.96543300e-01
2.02965572e-01 5.96847653e-01 -2.48872742e-01 2.59065479e-01
-7.44796634e-01 2.35579625e-01 7.84565985e-01 6.31249845e-01
6.00290656e-01 5.77222526e-01 1.63142785e-01 4.66399670e-01
6.17095716e-02 6.72086298e-01 1.24149442e+00 -5.48007786e-01
4.32736948e-02 9.21578407e-01 3.37560087e-01 -1.67425895e+00
1.20106585e-01 -2.45847121e-01 -6.77188575e-01 8.69775712e-01
4.33083266e-01 -9.03990027e-03 -1.65738440e+00 7.12352931e-01
-1.20400310e-01 2.41298705e-01 -2.05761090e-01 8.59294057e-01
9.65925634e-01 6.99847400e-01 1.39460236e-01 3.89703572e-01
1.20892310e+00 -9.38705921e-01 -3.82231563e-01 -4.54423368e-01
-1.03055172e-01 -7.81054437e-01 9.77743864e-01 7.29421616e-01
-8.01459968e-01 -5.79187393e-01 -1.17412317e+00 1.58568665e-01
-1.18540001e+00 -3.46770406e-01 5.29600739e-01 1.10535872e+00
-8.35826695e-01 7.04537034e-01 -2.79283792e-01 -4.83068317e-01
1.17455995e+00 7.59885728e-01 -5.97643495e-01 1.65037453e-01
-1.07804811e+00 1.04954672e+00 5.88167429e-01 -1.41385153e-01
-7.91949034e-01 -2.84295529e-01 -4.86995310e-01 1.76501542e-01
-1.69361517e-01 -1.63131475e-01 1.16726398e+00 -2.18991041e+00
-9.41360891e-01 1.23209774e+00 1.91135302e-01 -1.09714854e+00
7.68898010e-01 -2.06298560e-01 -6.76274538e-01 4.99567181e-01
-9.80679244e-02 4.91442651e-01 1.84084034e+00 -1.31951475e+00
-3.92298549e-01 -4.51482922e-01 -4.52676237e-01 -1.58822522e-01
-3.92203122e-01 -3.26040410e-03 4.30774212e-01 -7.91014731e-01
2.14586884e-01 -7.13373423e-01 2.58626431e-01 -2.85514221e-02
-1.99806735e-01 2.23334849e-01 1.94244456e+00 -5.93199611e-01
8.50409031e-01 -1.89771998e+00 -7.11899400e-01 2.59143800e-01
3.42165351e-01 1.45688307e+00 -1.39879718e-01 3.10588181e-01
-3.93451929e-01 3.70619982e-01 1.05201022e-03 5.97633183e-01
-3.27960253e-01 -1.19709305e-01 -4.40126061e-01 7.24401653e-01
-9.14293900e-02 1.13663435e+00 -9.03729975e-01 -1.14748865e-01
4.64288384e-01 1.26210591e-02 1.45760447e-01 4.12990153e-02
1.38728470e-01 7.08269924e-02 -4.26171303e-01 8.61721992e-01
1.32845831e+00 -3.91239636e-02 -2.29839772e-01 7.57263303e-02
3.94034028e-01 -1.59444407e-01 -5.75969577e-01 5.47967315e-01
-1.01478875e-01 1.16800141e+00 -6.89552799e-02 -1.17428684e+00
8.17753255e-01 1.16648443e-01 -2.83845454e-01 -6.42059743e-01
5.12902498e-01 4.11368072e-01 1.68088421e-01 -7.49405026e-01
3.96923214e-01 -3.32453549e-01 1.21756114e-01 4.32375461e-01
-2.32529026e-02 1.57340854e-01 -2.64375001e-01 3.26916724e-01
1.40357542e+00 -5.54402709e-01 3.79296303e-01 -2.21151024e-01
1.00671983e+00 -2.22487208e-02 8.65135714e-03 1.40480673e+00
-1.06909811e+00 2.76396900e-01 1.77738607e-01 -1.15468621e+00
-1.19849527e+00 -1.53464663e+00 4.33858372e-02 8.20267677e-01
4.61058527e-01 5.28575957e-01 -8.93316925e-01 -1.30848348e+00
7.97355473e-01 4.23970699e-01 -6.84382200e-01 -5.89410603e-01
-7.14834154e-01 -4.42840040e-01 5.90063095e-01 5.37651666e-02
1.29191959e+00 -1.47696924e+00 -3.37843895e-01 7.14374557e-02
5.25768101e-01 -7.57975459e-01 -3.40781212e-01 -1.40591443e-01
-8.93426299e-01 -1.34648240e+00 -2.46567860e-01 -1.18601978e+00
1.04282761e+00 7.08871901e-01 9.62866127e-01 7.05391943e-01
-5.92591822e-01 -1.06912874e-01 -4.99300361e-01 -4.30665851e-01
-9.49942589e-01 -1.96502507e-01 -7.82797635e-02 -1.70491859e-01
9.34684932e-01 -3.27814311e-01 -8.42781544e-01 1.57313626e-02
-1.23521709e+00 -6.50235295e-01 6.68810368e-01 7.92244077e-01
-5.25992751e-01 2.98782557e-01 1.03815520e+00 -1.26445460e+00
1.08160222e+00 -3.91162694e-01 -5.05692899e-01 2.78738420e-02
-6.54659390e-01 -2.50168949e-01 1.19615090e+00 -5.98012090e-01
-7.54529774e-01 9.42694396e-02 3.26086998e-01 -9.40204710e-02
-4.44976717e-01 -3.53651106e-01 2.57750064e-01 -9.52904046e-01
1.09800816e+00 5.48811018e-01 1.74327210e-01 -3.48620772e-01
3.39125782e-01 1.26283252e+00 7.36049294e-01 2.12519228e-01
1.19789636e+00 6.57833457e-01 -3.33879679e-01 -7.54407644e-01
-4.04002219e-01 -5.51233470e-01 -5.11023402e-01 -2.08731100e-01
2.31387213e-01 -2.90072471e-01 -8.96525502e-01 1.19042909e+00
-1.07796419e+00 4.52973753e-01 2.91923165e-01 -2.05229625e-01
-3.52684222e-02 9.73323107e-01 -6.03276551e-01 -6.75605834e-01
-7.23364830e-01 -8.29857886e-01 6.34410977e-01 3.53071362e-01
-1.42430767e-01 -7.97818184e-01 -5.38685262e-01 4.21366096e-01
7.01966166e-01 2.59959757e-01 9.04742002e-01 -1.10851097e+00
-5.64590275e-01 -1.14855981e+00 -6.35873318e-01 8.75156939e-01
3.33524048e-01 -3.80710423e-01 -7.38050878e-01 -5.79341531e-01
8.97551849e-02 -2.30877504e-01 1.07985973e+00 -9.03420802e-03
1.19038606e+00 -7.06288636e-01 -3.95056605e-01 1.97215036e-01
1.71426940e+00 5.71639538e-01 9.66139555e-01 9.15949643e-01
4.71020579e-01 -9.39538479e-02 -1.01284921e-01 -1.79303423e-01
-6.06188238e-01 -1.30805373e-01 6.49232686e-01 1.94571853e-01
1.08604915e-01 -5.91300204e-02 1.81855619e-01 -8.90166014e-02
6.70113504e-01 -3.23104441e-01 -8.13760757e-01 4.02099401e-01
-1.79892802e+00 -1.34499681e+00 -3.86609256e-01 2.21577168e+00
3.81275773e-01 2.68344223e-01 3.78868431e-02 -1.91518553e-02
1.20879436e+00 1.40151277e-01 -8.68618548e-01 -1.02131951e+00
-3.28969657e-02 5.26122570e-01 1.05477273e+00 3.04586053e-01
-1.60130548e+00 1.51471746e+00 7.24280548e+00 6.09193385e-01
-1.18224251e+00 -4.55903150e-02 5.84414363e-01 1.32140115e-01
3.69006872e-01 -1.69080034e-01 -3.00865024e-01 9.07499015e-01
8.41479719e-01 -2.24328145e-01 5.99095106e-01 1.01712298e+00
4.02959883e-01 -5.00344038e-01 -3.50225747e-01 9.96976852e-01
5.06334186e-01 -1.35813820e+00 5.74401319e-01 -3.37164283e-01
9.49995577e-01 -7.31005445e-02 3.97803754e-01 2.18649879e-01
2.57649183e-01 -1.40232396e+00 3.47170621e-01 4.04705495e-01
-6.16068346e-03 -6.39309585e-01 9.41666722e-01 2.10308552e-01
-3.19457173e-01 -5.23360729e-01 -4.87499207e-01 -2.23585889e-01
-3.11626434e-01 1.68671340e-01 -8.45126092e-01 -3.71447265e-01
1.02587140e+00 3.09200197e-01 -1.23004782e+00 1.20735049e+00
-5.03135798e-03 4.45007801e-01 9.93108302e-02 -2.10066766e-01
7.93232977e-01 -8.46159682e-02 7.17583597e-01 1.19482172e+00
-6.21943511e-02 -5.11423826e-01 -2.46493638e-01 8.87428164e-01
-3.79803687e-01 -2.33097449e-01 -1.05764151e+00 -2.33824179e-01
4.16592747e-01 1.16974175e+00 -9.25965071e-01 -6.89007759e-01
-2.42447659e-01 1.50276363e+00 -3.15304957e-02 3.71054232e-01
-4.24438626e-01 -8.30536306e-01 5.16387582e-01 3.52010936e-01
4.30851638e-01 -1.36134461e-01 -3.38023454e-01 -8.55985701e-01
-1.08373985e-01 -1.18029416e+00 2.34602511e-01 -8.10770452e-01
-1.66272438e+00 4.65441167e-01 -4.80039924e-01 -1.21917868e+00
2.27888882e-01 -1.11693430e+00 -9.27169681e-01 7.96370625e-01
-1.01095498e+00 -1.20781302e+00 -5.35204589e-01 2.67935067e-01
5.10784328e-01 -8.80113184e-01 3.48555475e-01 -1.08406223e-01
-1.35478497e-01 3.35254848e-01 5.33605397e-01 3.81587178e-01
6.00151956e-01 -1.18700063e+00 7.87139773e-01 9.01855826e-01
-2.86693215e-01 8.11337590e-01 8.11331928e-01 -8.34227562e-01
-8.74867320e-01 -9.39968884e-01 4.56836224e-01 -5.11809707e-01
6.35484040e-01 -1.42689317e-01 -1.16891956e+00 5.23500025e-01
3.88283670e-01 7.16683120e-02 2.82915205e-01 -7.00733364e-01
-4.42682385e-01 -6.27306551e-02 -1.82285535e+00 4.68935162e-01
3.32737476e-01 -4.84964341e-01 -8.37675095e-01 3.45902890e-01
9.95331705e-02 -1.25175804e-01 2.03027084e-01 -1.00965776e-01
6.46640658e-01 -1.38240397e+00 1.12808847e+00 -1.06132901e+00
2.70693712e-02 -5.33997893e-01 4.29874063e-01 -1.03953993e+00
-2.68124133e-01 -4.36605155e-01 -2.36823797e-01 5.00088096e-01
-1.07075579e-01 -7.80221879e-01 1.37025332e+00 2.43807897e-01
4.42569315e-01 -2.82817986e-02 -8.43763649e-01 -9.68934774e-01
4.06419873e-01 1.26723185e-01 3.47768426e-01 8.63079727e-01
-1.92645505e-01 -7.83754289e-02 -3.44492048e-01 1.18087992e-01
5.88429511e-01 -9.03303474e-02 5.17359376e-01 -1.00139582e+00
4.04286146e-01 -3.65356266e-01 -1.30017030e+00 -5.62442720e-01
3.36500466e-01 -7.02686965e-01 -2.65125662e-01 -1.14486277e+00
1.63078055e-01 3.57357383e-01 -2.16842681e-01 2.20788911e-01
-9.52619240e-02 6.52780592e-01 -1.79034770e-02 2.93953985e-01
-5.81693769e-01 -7.57337064e-02 1.02294743e+00 -5.92143595e-01
1.68217897e-01 2.30123833e-01 -7.14810967e-01 8.98400247e-01
1.03149033e+00 -6.15429699e-01 -1.04970835e-01 -1.95999831e-01
-1.70273855e-01 -8.24845731e-01 1.01474845e+00 -1.18973029e+00
7.27535263e-02 1.25586897e-01 9.46697056e-01 -4.50629145e-01
-2.84513384e-01 -1.08142817e+00 -3.14023763e-01 1.14743555e+00
-7.62094930e-02 5.26528582e-02 3.30551624e-01 8.58833373e-01
-6.86347708e-02 -7.67880023e-01 1.49770045e+00 -9.50178683e-01
-1.10566294e+00 3.07426006e-02 -8.93171847e-01 -2.92967260e-01
1.31830657e+00 -8.39953661e-01 -9.01949525e-01 -7.06447482e-01
-5.51717043e-01 -1.88840374e-01 6.40775621e-01 7.44307637e-01
8.02931070e-01 -1.23224270e+00 -1.88621029e-01 4.25171286e-01
-1.88838709e-02 -8.90681148e-01 -1.47252157e-01 6.23618551e-02
-1.58105624e+00 4.98411030e-01 -7.21656680e-01 -6.12051748e-02
-1.49184096e+00 7.67203391e-01 7.59370685e-01 5.76024890e-01
-6.83087468e-01 7.22902477e-01 -2.15330318e-01 -2.50148684e-01
1.02498338e-01 2.77285159e-01 -3.32723372e-02 -5.82437873e-01
5.82746625e-01 6.99196815e-01 1.65992752e-02 -9.48970020e-01
-3.80351335e-01 2.69584656e-02 -6.01302266e-01 5.30288637e-01
9.07794416e-01 1.08292572e-01 -4.71282274e-01 -1.33983061e-01
1.61297250e+00 -4.14196700e-01 -6.44072652e-01 8.56441036e-02
2.53152877e-01 -1.26806128e+00 2.65497640e-02 -1.21781683e+00
-8.40909660e-01 7.09557652e-01 1.31861842e+00 7.20369101e-01
8.67570758e-01 -6.19348228e-01 1.23285306e+00 7.38474250e-01
-5.65115213e-02 -1.35563588e+00 4.58233863e-01 5.06593287e-01
7.90145159e-01 -1.72736943e+00 8.21714625e-02 -2.73129437e-02
-4.21785653e-01 1.54486740e+00 8.88059020e-01 -1.06751692e+00
6.24642015e-01 -2.03902289e-01 4.39297467e-01 -4.28094983e-01
-1.23577103e-01 1.84489772e-01 1.56345263e-01 1.05217230e+00
2.31188342e-01 -1.40948266e-01 -4.36097711e-01 -1.41842710e-02
-9.61710438e-02 6.73256069e-03 6.69791281e-01 9.47716236e-01
-1.23484957e+00 -8.89992356e-01 -7.18814671e-01 8.73042226e-01
-5.94993949e-01 -1.71610475e-01 -1.07312763e+00 6.64978802e-01
3.56327415e-01 9.06655550e-01 2.28701040e-01 -2.24298492e-01
1.03320725e-01 3.00927032e-02 7.23326206e-02 -4.67217207e-01
-1.15902233e+00 -6.69905663e-01 -4.14261997e-01 -2.27613404e-01
2.35451367e-02 -9.58140120e-02 -9.82406199e-01 -8.04861128e-01
-3.68996352e-01 -4.97389883e-02 5.41585445e-01 7.25408077e-01
2.80471772e-01 -1.47601077e-02 6.55653715e-01 -6.17459953e-01
-7.87106395e-01 -9.58578825e-01 -6.95406914e-01 9.95559275e-01
5.46980143e-01 -3.15545738e-01 -7.55039811e-01 9.43018422e-02] | [7.7259321212768555, 9.91807746887207] |
f6adaa5a-d9e0-4e96-8db3-fcff68196b14 | a-probabilistic-framework-for-visual | 2301.02086 | null | https://arxiv.org/abs/2301.02086v1 | https://arxiv.org/pdf/2301.02086v1.pdf | A Probabilistic Framework for Visual Localization in Ambiguous Scenes | Visual localization allows autonomous robots to relocalize when losing track of their pose by matching their current observation with past ones. However, ambiguous scenes pose a challenge for such systems, as repetitive structures can be viewed from many distinct, equally likely camera poses, which means it is not sufficient to produce a single best pose hypothesis. In this work, we propose a probabilistic framework that for a given image predicts the arbitrarily shaped posterior distribution of its camera pose. We do this via a novel formulation of camera pose regression using variational inference, which allows sampling from the predicted distribution. Our method outperforms existing methods on localization in ambiguous scenes. Code and data will be released at https://github.com/efreidun/vapor. | ['Patric Jensfelt', 'Alessandro Pieropan', 'Amit Dekel', 'Leonard Bruns', 'Fereidoon Zangeneh'] | 2023-01-05 | null | null | null | null | ['visual-localization'] | ['computer-vision'] | [-9.66240019e-02 -3.46092880e-02 -3.11501324e-01 -3.12716067e-01
-7.41782606e-01 -9.29116905e-01 5.33966899e-01 -2.05642775e-01
-3.66297990e-01 7.71639228e-01 6.00949861e-02 -5.69255976e-03
7.57941008e-02 -3.75305593e-01 -1.02710092e+00 -5.96593142e-01
1.84149384e-01 7.40404367e-01 2.92484522e-01 1.91781789e-01
3.08342487e-01 5.06561935e-01 -1.25404859e+00 -3.44772905e-01
2.55273014e-01 5.51481605e-01 9.63884592e-01 6.92362547e-01
3.96763802e-01 5.40709257e-01 -2.58955806e-01 -1.41853005e-01
2.74733931e-01 -4.58803587e-03 -5.31550288e-01 1.84786156e-01
6.27979457e-01 -6.12849414e-01 -4.10386115e-01 1.33476317e+00
3.84123325e-02 1.42774463e-01 7.66550064e-01 -1.35495663e+00
-3.12183648e-01 2.42544547e-01 -5.43972373e-01 -6.75968751e-02
7.16361821e-01 6.21501096e-02 1.02291322e+00 -8.32998276e-01
8.50433767e-01 1.26487255e+00 6.17650092e-01 3.60935271e-01
-1.40112567e+00 -4.71307844e-01 5.37360072e-01 1.91829517e-01
-1.73100293e+00 -4.98170316e-01 6.84066236e-01 -4.55743879e-01
6.00239694e-01 -1.27646729e-01 6.07973456e-01 1.35304308e+00
2.78896719e-01 7.42640197e-01 8.75822425e-01 -1.36148751e-01
2.71738559e-01 -1.28747717e-01 -2.95531034e-01 7.34866440e-01
3.32400471e-01 -6.76192567e-02 -5.93863070e-01 -1.23652540e-01
9.23843920e-01 2.92901635e-01 -3.41267645e-01 -9.23905015e-01
-1.41981995e+00 7.15443552e-01 4.00164962e-01 -1.65253013e-01
-3.19348186e-01 4.85368878e-01 -3.44249219e-01 -8.99462849e-02
-2.97967419e-02 4.03346926e-01 -4.01009828e-01 -1.14462294e-01
-5.88093519e-01 3.37415665e-01 6.96080267e-01 1.31584573e+00
1.05695486e+00 -3.53795290e-01 3.60365838e-01 4.46579129e-01
7.92431712e-01 9.56260204e-01 2.89556198e-02 -1.48354626e+00
1.73485056e-01 1.09318815e-01 5.66398501e-01 -1.01805127e+00
-2.32894152e-01 -3.01785972e-02 -3.77461016e-01 2.01528862e-01
3.80093068e-01 -9.00796950e-02 -8.38182509e-01 1.73359132e+00
2.46737286e-01 1.17643468e-01 1.16210423e-01 1.10750234e+00
1.70863062e-01 7.33939528e-01 -2.99447477e-01 -8.08317587e-02
9.76148784e-01 -8.56263578e-01 -5.38523734e-01 -8.98651421e-01
-4.08148728e-02 -8.67403984e-01 5.58432579e-01 5.03909111e-01
-7.50541031e-01 -2.52970904e-01 -9.09547329e-01 -6.25232756e-02
3.93183529e-02 2.48372272e-01 4.77586031e-01 3.70145813e-02
-1.16524374e+00 2.36282557e-01 -1.27076149e+00 -6.32453263e-01
2.17104815e-02 2.15093523e-01 -4.18125838e-01 -1.66279882e-01
-5.82641423e-01 1.04229164e+00 3.95977944e-01 1.94840118e-01
-1.26221550e+00 1.36215389e-01 -9.55745637e-01 -3.05953413e-01
6.08568013e-01 -7.05881834e-01 1.51981103e+00 -4.60222334e-01
-1.45961130e+00 6.30411327e-01 -7.26608396e-01 -3.96256924e-01
5.24632990e-01 -4.64155883e-01 2.87076861e-01 8.04268494e-02
3.12376976e-01 9.67433870e-01 9.84366894e-01 -1.70198619e+00
-5.43810606e-01 -4.90128785e-01 8.74243006e-02 5.37918210e-01
3.62190962e-01 -5.30973017e-01 -9.67097521e-01 -1.19999284e-02
7.27153420e-01 -1.39499664e+00 -5.23078024e-01 1.80608019e-01
-4.45833504e-01 9.87162720e-03 5.95136583e-01 -3.73447686e-01
5.32418966e-01 -1.99199414e+00 5.16036987e-01 3.29611190e-02
9.60763991e-02 -4.35618550e-01 2.21998557e-01 4.47233170e-01
4.66338336e-01 -1.28980242e-02 9.67643932e-02 -7.94829845e-01
1.47582814e-01 5.72482347e-01 -5.39368212e-01 9.21400845e-01
-2.10937679e-01 4.03294891e-01 -8.88092041e-01 -4.10063326e-01
4.86312360e-01 4.51457292e-01 -4.60043162e-01 2.23978877e-01
-3.03700477e-01 8.11621010e-01 -4.86720443e-01 6.70499563e-01
7.55971968e-01 -2.74823546e-01 4.23739016e-01 -5.02013341e-02
-1.56065136e-01 -6.86580315e-02 -1.25851917e+00 1.87283635e+00
-3.21657270e-01 6.53203130e-01 2.00865939e-01 -6.03221357e-01
7.95788229e-01 6.95812181e-02 3.27046901e-01 -1.51077092e-01
1.39943138e-01 -1.86058842e-02 -4.49516803e-01 -2.45509237e-01
7.43348122e-01 7.46984705e-02 -3.64633381e-01 -8.12751800e-03
1.28779905e-02 -3.77466351e-01 5.11118695e-02 4.33992855e-02
9.81553674e-01 5.12473524e-01 4.90241438e-01 1.85612798e-01
1.01583831e-01 1.71321243e-01 8.92257571e-01 9.01702464e-01
-3.22974265e-01 8.23010981e-01 2.80518681e-01 -2.64797240e-01
-1.16595685e+00 -1.52697694e+00 -1.79227501e-01 6.20486736e-01
9.02601123e-01 -3.19553643e-01 -3.80476385e-01 -4.10042316e-01
1.46017559e-02 5.12735069e-01 -2.44873405e-01 2.08437011e-01
-4.69882339e-01 -1.42663598e-01 1.41786141e-02 2.75774509e-01
1.95665315e-01 -6.71648383e-01 -8.06690931e-01 -1.22085035e-01
-3.99460286e-01 -1.27091932e+00 -4.20039386e-01 2.29666278e-01
-6.69278562e-01 -9.88008857e-01 -4.72987026e-01 -5.69496334e-01
9.04760182e-01 6.25337541e-01 8.56970310e-01 -2.90479392e-01
-1.55547261e-02 7.28016675e-01 -3.45076978e-01 -2.35993609e-01
-2.22642839e-01 -4.85308096e-02 4.36416328e-01 -3.55525076e-01
1.38226539e-01 -4.72414672e-01 -5.42090118e-01 5.06009817e-01
-4.24621820e-01 1.17261820e-01 7.24015832e-01 6.55306697e-01
9.92438614e-01 -1.69739693e-01 -2.96166033e-01 -3.76286805e-01
5.85663952e-02 -5.82549334e-01 -1.04477108e+00 2.73654401e-01
-3.32977742e-01 1.96698546e-01 3.66013765e-01 -3.63049597e-01
-9.62563336e-01 7.66479075e-01 2.70547301e-01 -8.45015764e-01
-4.23357904e-01 3.02397072e-01 -1.31902611e-02 1.10489160e-01
3.77225816e-01 2.20741525e-01 -3.09023093e-02 -3.73004526e-01
4.05524641e-01 3.48239303e-01 7.47201681e-01 -6.10945761e-01
9.29408729e-01 7.42312014e-01 7.78104886e-02 -7.92942584e-01
-8.31740856e-01 -6.25374377e-01 -9.44563508e-01 -3.03850889e-01
8.87293339e-01 -1.17583597e+00 -7.40680993e-01 1.87917337e-01
-1.40747559e+00 -2.51324892e-01 2.83622444e-01 7.26637900e-01
-7.17749298e-01 4.60618615e-01 -1.67367235e-01 -9.04256642e-01
3.17346007e-01 -1.41463768e+00 1.40935373e+00 4.35463816e-01
-3.40594500e-01 -8.18382144e-01 2.51094729e-01 2.07231596e-01
-1.17415614e-01 1.02072783e-01 2.41301749e-02 -1.84294015e-01
-1.22443938e+00 -1.40532091e-01 6.69212118e-02 -9.30580646e-02
-1.05506025e-01 1.32929191e-01 -7.30225682e-01 -6.32541955e-01
-1.15067631e-01 -2.50809938e-01 7.10803747e-01 5.78275144e-01
7.78198481e-01 -1.42830655e-01 -8.31799924e-01 6.84896171e-01
1.43567026e+00 1.81086898e-01 1.95961773e-01 4.09020752e-01
6.14255428e-01 3.47367674e-01 7.35896170e-01 5.14870584e-01
5.92448354e-01 7.90208578e-01 8.40865433e-01 5.08010268e-01
3.05991203e-01 -5.74383199e-01 4.27902997e-01 3.09559107e-01
1.38490722e-01 -4.47967321e-01 -1.08364010e+00 6.09630108e-01
-2.17925811e+00 -8.38844061e-01 1.24277592e-01 2.15700793e+00
3.18002224e-01 2.64085028e-02 -3.74604046e-01 -6.46150351e-01
8.74434114e-01 2.14455515e-01 -7.54064739e-01 1.79976940e-01
4.49580252e-02 -5.25928497e-01 9.51720357e-01 8.39466989e-01
-1.17101347e+00 1.04482150e+00 6.21539831e+00 3.08888435e-01
-8.95701587e-01 3.30868661e-02 1.11472420e-01 -2.58325739e-03
-1.10198781e-01 5.92032731e-01 -1.27958548e+00 3.52300465e-01
5.09863615e-01 -1.60767976e-02 6.05619192e-01 1.02184224e+00
1.64851740e-01 -6.59212589e-01 -1.19237566e+00 1.21084821e+00
1.84977785e-01 -9.89450753e-01 -1.90269396e-01 2.90991366e-01
6.14502072e-01 4.73963290e-01 1.29105017e-01 -8.82429332e-02
7.49576867e-01 -7.90892243e-01 1.10354865e+00 6.65612280e-01
3.34723413e-01 -3.85251701e-01 4.54595417e-01 6.66078210e-01
-9.72271085e-01 -5.66579066e-02 -5.94202578e-01 -1.95822775e-01
3.78555030e-01 2.22818464e-01 -1.37214696e+00 3.04195613e-01
7.89070666e-01 8.10255706e-01 -5.24316251e-01 1.13878679e+00
-5.88161111e-01 1.44893348e-01 -6.40337229e-01 -4.66992743e-02
2.36184355e-02 -2.82604694e-01 8.03932369e-01 6.48685813e-01
6.88502550e-01 -2.05635354e-01 6.38274908e-01 8.73885989e-01
1.08804129e-01 -4.75359857e-01 -9.81739402e-01 3.45185310e-01
9.82744694e-01 1.06537628e+00 -8.35318625e-01 1.23873383e-01
-3.01458329e-01 1.15495169e+00 5.01753092e-01 3.07904303e-01
-8.09642076e-01 3.73315066e-01 6.60405695e-01 -3.02756518e-01
3.49697024e-01 -7.87778735e-01 -7.19605163e-02 -1.35586572e+00
6.54367059e-02 -3.39907527e-01 1.22179605e-01 -1.18488920e+00
-1.07584071e+00 4.07685548e-01 2.71805733e-01 -1.48185742e+00
-5.39192259e-01 -6.11750841e-01 -2.21042350e-01 6.40295625e-01
-1.14595079e+00 -1.15362608e+00 -2.86021918e-01 3.05088013e-01
6.42341375e-01 1.73537076e-01 5.89559734e-01 -2.72156566e-01
-3.15637112e-01 1.80555005e-02 2.72742301e-01 -4.09199037e-02
9.09404874e-01 -1.36825430e+00 1.46754205e-01 9.78611052e-01
4.68286693e-01 7.06594050e-01 1.19929695e+00 -7.38100886e-01
-1.63767385e+00 -8.53281140e-01 4.42970693e-01 -8.42249930e-01
6.19958580e-01 -3.68829340e-01 -4.87499058e-01 1.21005428e+00
1.36770517e-01 -1.60819784e-01 1.30091697e-01 7.71850199e-02
-9.39036757e-02 -2.75831632e-02 -7.65698552e-01 8.16942692e-01
9.28573012e-01 -4.02008533e-01 -5.50751984e-01 3.43326837e-01
4.30935293e-01 -6.76101625e-01 -5.26313603e-01 2.87478507e-01
5.37517905e-01 -8.93655658e-01 9.17579710e-01 3.20360720e-01
-2.17064377e-02 -8.01664770e-01 -5.85548937e-01 -1.28504789e+00
-1.85135707e-01 -5.23703039e-01 1.44744083e-01 8.81191969e-01
3.64201218e-01 -4.87465978e-01 8.70831788e-01 3.69381338e-01
-7.61036901e-03 -2.48311952e-01 -1.08408141e+00 -7.29865789e-01
-2.74342060e-01 -4.34071064e-01 1.55204237e-01 4.89495397e-01
-3.06071967e-01 3.82971376e-01 -6.10496879e-01 1.00775909e+00
8.54225218e-01 3.13615531e-01 1.14064467e+00 -1.07964361e+00
-3.50167125e-01 -1.31612435e-01 -4.49560851e-01 -1.56565404e+00
4.13672835e-01 -5.16988277e-01 6.40369654e-01 -1.54208517e+00
3.15582424e-01 -2.52099991e-01 3.14645380e-01 3.88364762e-01
2.31556371e-01 1.96867391e-01 2.78208673e-01 6.09977901e-01
-1.02413797e+00 5.68077326e-01 1.00462151e+00 -4.03886959e-02
5.44374399e-02 9.10478830e-02 -3.10020894e-01 9.95180964e-01
8.45629394e-01 -4.29641098e-01 -3.63450170e-01 -7.35687613e-01
3.35140646e-01 4.23962474e-01 6.13451362e-01 -1.05508268e+00
5.87898612e-01 -4.29408491e-01 5.79827607e-01 -8.91678393e-01
8.30799997e-01 -9.42702234e-01 5.19341886e-01 3.26430857e-01
-2.28616223e-02 2.52469659e-01 -1.11588702e-01 1.05674148e+00
-1.51694641e-01 -4.76501971e-01 5.16829789e-01 -2.78769016e-01
-1.14794445e+00 2.42019624e-01 -5.15452445e-01 -3.89033526e-01
1.05804062e+00 -1.18161768e-01 -2.68651783e-01 -6.63830280e-01
-8.15796793e-01 4.59028304e-01 1.23397934e+00 4.93560374e-01
6.80414081e-01 -1.08602679e+00 -3.25656533e-01 4.90200287e-03
2.11640015e-01 4.99385506e-01 1.36863217e-01 7.42486537e-01
-7.48789012e-01 3.98739249e-01 -7.77454004e-02 -1.05912876e+00
-1.21219242e+00 5.60875297e-01 2.30515033e-01 1.95290104e-01
-4.35185105e-01 6.66658103e-01 2.92706639e-01 -6.20249510e-01
1.73341811e-01 -1.56008765e-01 4.43425775e-02 -3.38911235e-01
1.99103221e-01 1.61865026e-01 -4.63430405e-01 -1.02830923e+00
-4.32084680e-01 7.53677785e-01 -5.66695109e-02 -3.91509354e-01
1.11146867e+00 -7.51500249e-01 -1.20212257e-01 6.78415120e-01
9.78083551e-01 1.37914360e-01 -1.91161931e+00 -1.21868901e-01
-1.28319472e-01 -6.18989646e-01 -3.56849581e-02 -3.62518519e-01
-4.78886336e-01 5.01921058e-01 5.03580809e-01 -1.02911957e-01
6.66494846e-01 4.95477706e-01 1.88478425e-01 8.11746180e-01
8.92066598e-01 -7.97067344e-01 2.77362224e-02 7.78197944e-01
7.66691446e-01 -1.43641472e+00 1.37696519e-01 -3.32255900e-01
-5.88506222e-01 1.04462218e+00 6.40700042e-01 -2.04453543e-01
4.93773073e-01 2.34108180e-01 2.51796357e-02 -1.26211867e-02
-6.83946431e-01 -1.39041722e-01 -1.04603291e-01 6.48967505e-01
-5.38917519e-02 2.02371061e-01 2.17551783e-01 7.09184855e-02
-3.42199504e-01 -4.26034778e-01 7.31309354e-01 9.99079764e-01
-5.30196130e-01 -1.05988407e+00 -7.28510499e-01 -1.77193150e-01
-1.72736943e-01 1.94015771e-01 -3.79714966e-01 6.24881744e-01
-4.62547131e-02 8.79618585e-01 1.72019377e-01 -2.84906000e-01
-8.74706544e-04 -1.90993130e-01 6.24930859e-01 -7.04841971e-01
6.35038078e-01 3.12023968e-01 -1.94731489e-01 -6.94394946e-01
-4.66573179e-01 -1.17184699e+00 -1.11143172e+00 -1.03973374e-01
-2.29997873e-01 -2.01590136e-02 7.11397529e-01 7.95397460e-01
1.95225120e-01 -4.82836924e-02 5.99215329e-01 -1.41778159e+00
-5.17163455e-01 -7.96998620e-01 -4.78416860e-01 8.01601261e-02
6.17313683e-01 -1.04595768e+00 -4.41029429e-01 1.93865612e-01] | [7.747799396514893, -2.381242036819458] |
b0e87f41-e610-40dc-9b83-429c250ea520 | a-fair-experimental-comparison-of-neural | 2208.14822 | null | https://arxiv.org/abs/2208.14822v1 | https://arxiv.org/pdf/2208.14822v1.pdf | A Fair Experimental Comparison of Neural Network Architectures for Latent Representations of Multi-Omics for Drug Response Prediction | Recent years have seen a surge of novel neural network architectures for the integration of multi-omics data for prediction. Most of the architectures include either encoders alone or encoders and decoders, i.e., autoencoders of various sorts, to transform multi-omics data into latent representations. One important parameter is the depth of integration: the point at which the latent representations are computed or merged, which can be either early, intermediate, or late. The literature on integration methods is growing steadily, however, close to nothing is known about the relative performance of these methods under fair experimental conditions and under consideration of different use cases. We developed a comparison framework that trains and optimizes multi-omics integration methods under equal conditions. We incorporated early integration and four recently published deep learning methods: MOLI, Super.FELT, OmiEmbed, and MOMA. Further, we devised a novel method, Omics Stacking, that combines the advantages of intermediate and late integration. Experiments were conducted on a public drug response data set with multiple omics data (somatic point mutations, somatic copy number profiles and gene expression profiles) that was obtained from cell lines, patient-derived xenografts, and patient samples. Our experiments confirmed that early integration has the lowest predictive performance. Overall, architectures that integrate triplet loss achieved the best results. Statistical differences can, overall, rarely be observed, however, in terms of the average ranks of methods, Super.FELT is consistently performing best in a cross-validation setting and Omics Stacking best in an external test set setting. The source code of all experiments is available under \url{https://github.com/kramerlab/Multi-Omics_analysis} | ['Stefan Kramer', 'Tony Hauptmann'] | 2022-08-31 | null | null | null | null | ['drug-response-prediction'] | ['medical'] | [ 1.48632929e-01 -3.85616094e-01 -4.72520351e-01 -2.07271084e-01
-7.74256289e-01 -4.62610096e-01 5.37011087e-01 4.78250265e-01
-3.02488267e-01 8.91919851e-01 9.16373655e-02 -4.44114625e-01
-3.46814603e-01 -6.20999813e-01 -8.41386259e-01 -8.25480878e-01
2.08176970e-01 5.87937295e-01 -4.52600509e-01 1.77681684e-01
-1.41386958e-02 4.69772130e-01 -1.27088106e+00 6.22042477e-01
9.71012473e-01 1.04170537e+00 -8.40577483e-02 6.77429736e-01
-1.20570838e-01 4.73840564e-01 -3.09989959e-01 -3.56586128e-01
1.86861157e-01 -3.41604382e-01 -4.50809717e-01 -3.44124913e-01
1.25495359e-01 -1.58808194e-02 -1.22585855e-01 7.48608410e-01
7.34956324e-01 -4.71201777e-01 6.66448116e-01 -1.13842690e+00
-8.29502285e-01 2.36410409e-01 -1.11028843e-01 4.18848395e-02
1.15772016e-01 3.62411708e-01 8.66758347e-01 -8.82617176e-01
7.35028744e-01 9.67759907e-01 7.45055974e-01 3.70248348e-01
-1.33418417e+00 -4.99490470e-01 -8.46975967e-02 2.38531437e-02
-1.25830817e+00 -5.20861030e-01 5.80902770e-02 -7.32603073e-01
1.40482485e+00 2.83249289e-01 5.62847435e-01 1.21858609e+00
8.62858176e-01 7.71558106e-01 1.08065677e+00 -3.14587712e-01
2.43814841e-01 1.04797952e-01 2.88668454e-01 6.01106763e-01
2.08977774e-01 1.66570023e-01 -4.42288637e-01 -4.73830968e-01
3.74112666e-01 5.15451491e-01 -2.02922478e-01 -8.53897706e-02
-1.26733649e+00 7.46722639e-01 1.96708202e-01 2.68766671e-01
-4.93443489e-01 4.18080539e-02 3.94877881e-01 4.65527207e-01
6.41463161e-01 5.80397606e-01 -8.17837894e-01 1.59468681e-01
-8.30703139e-01 6.45320341e-02 6.81239426e-01 5.64678192e-01
3.76575112e-01 -1.91034406e-01 -1.22499451e-01 9.08435345e-01
1.62330911e-01 5.02756834e-02 7.86380172e-01 -5.46008587e-01
6.88074380e-02 6.74258351e-01 -8.08590427e-02 -5.21637380e-01
-5.32960057e-01 -4.74955261e-01 -8.66600215e-01 -1.39661878e-01
2.48630375e-01 -2.54548788e-01 -1.11891639e+00 1.76072049e+00
1.68770030e-01 2.67143190e-01 2.00775743e-01 6.49764597e-01
8.59279275e-01 5.93409061e-01 1.97280779e-01 -9.95881259e-02
1.47421551e+00 -7.84971118e-01 -6.40324831e-01 6.54763728e-02
7.14610040e-01 -6.34567618e-01 5.34872174e-01 4.77277011e-01
-9.71774220e-01 -2.68623203e-01 -1.03813267e+00 -7.42923692e-02
-9.03649986e-01 3.00110877e-01 8.70631099e-01 4.19298708e-01
-1.01387882e+00 9.50315356e-01 -9.47823882e-01 -3.76751840e-01
3.53132308e-01 5.58519423e-01 -4.46617365e-01 7.65781030e-02
-1.19922137e+00 8.61945927e-01 5.47850788e-01 -8.41217116e-03
-9.73021924e-01 -1.02031374e+00 -6.53133988e-01 6.20783381e-02
-1.19517349e-01 -1.21499324e+00 9.08032060e-01 -1.03030598e+00
-1.63623273e+00 7.84084618e-01 -1.41658291e-01 -4.30000186e-01
2.40092218e-01 -2.88185477e-01 -4.16179568e-01 -2.17194453e-01
-9.24056172e-02 6.66813254e-01 2.98585445e-01 -6.86728954e-01
-3.93419415e-01 -5.16824543e-01 -2.79768050e-01 5.33909500e-02
-1.97240666e-01 1.07036240e-01 -2.74521440e-01 -2.62988210e-01
-5.28482422e-02 -1.03357804e+00 -2.37134829e-01 -2.37864122e-01
-5.71631134e-01 -7.63347149e-02 2.91913629e-01 -5.93606710e-01
9.33836341e-01 -1.95975661e+00 3.82727027e-01 -1.52286708e-01
1.51934445e-01 2.71686375e-01 -2.62252450e-01 7.27900088e-01
-7.01151133e-01 3.80088091e-01 -7.82925710e-02 -3.44832152e-01
-1.00020682e-02 -9.35314596e-02 -1.60680160e-01 5.85458577e-01
2.73796380e-01 1.02318358e+00 -6.96925402e-01 -6.08676462e-04
1.40050530e-01 6.64504826e-01 -5.18332303e-01 2.24169672e-01
-4.19526070e-01 2.99996853e-01 -1.03360876e-01 1.12322557e+00
3.89590234e-01 -7.38249481e-01 2.56700873e-01 -7.24097714e-02
-1.68743953e-01 3.62718999e-01 -5.53524792e-01 1.77423286e+00
-2.25249738e-01 4.65387464e-01 -3.39127719e-01 -7.84946799e-01
5.04319727e-01 4.59020853e-01 6.29615963e-01 -5.68207264e-01
2.31479719e-01 3.68022352e-01 1.67066336e-01 -3.00234407e-01
1.13731824e-01 -2.59367794e-01 2.02489898e-01 9.91071239e-02
1.92329258e-01 4.40643430e-01 5.98418601e-02 -1.93450943e-01
1.32080412e+00 5.17542921e-02 4.80510473e-01 -4.11523506e-02
3.00174087e-01 1.14105016e-01 7.02076137e-01 5.04361391e-01
-1.00426838e-01 6.10436916e-01 6.59602582e-01 -3.59158665e-01
-1.08296061e+00 -1.05803025e+00 -5.18961370e-01 1.00182760e+00
-2.49810293e-01 -4.74646091e-01 -3.87538016e-01 -4.12848622e-01
2.45362252e-01 5.47898173e-01 -6.08058631e-01 -2.23892167e-01
-9.37270448e-02 -1.42448294e+00 6.74151182e-01 3.75346839e-01
-3.18250209e-02 -6.99343920e-01 -2.04735920e-01 2.46374831e-01
9.74365175e-02 -9.48182523e-01 -9.94144101e-03 6.14569545e-01
-9.84178782e-01 -1.04860020e+00 -7.69681811e-01 -2.87533581e-01
4.29881632e-01 -1.34281993e-01 8.89541924e-01 -7.20439255e-02
-4.12079543e-01 -1.78119272e-01 -1.51456296e-01 -7.14174807e-01
-4.06676620e-01 1.13706946e-01 1.62988946e-01 -2.00363249e-01
5.74846745e-01 -4.93603498e-01 -6.71483338e-01 9.41845477e-02
-9.61706281e-01 2.65439361e-01 8.97550106e-01 1.25573778e+00
8.08216810e-01 -3.68653864e-01 7.27365732e-01 -8.63486052e-01
5.04333556e-01 -9.76106226e-01 -4.46494907e-01 3.44843179e-01
-7.24939227e-01 -7.15088844e-02 6.50802374e-01 -4.38992456e-02
-6.60061896e-01 -3.69438976e-02 -4.85111266e-01 -6.33380175e-01
-4.32146490e-01 9.04018283e-01 -7.57572353e-02 3.06608051e-01
4.15987521e-01 2.35320941e-01 1.80706009e-01 -5.46701849e-01
8.41736346e-02 8.12057793e-01 1.90874562e-03 -3.14731985e-01
-7.34603554e-02 3.91487092e-01 -9.70187113e-02 -6.57368481e-01
-6.63647771e-01 -4.81749266e-01 -4.64963019e-01 4.04229939e-01
9.30405498e-01 -1.12747896e+00 -6.71676576e-01 6.72860324e-01
-1.05063045e+00 -3.92839879e-01 1.48800716e-01 5.55745542e-01
-4.97603893e-01 1.98443651e-01 -1.00517404e+00 -4.38496530e-01
-7.12575912e-01 -1.33214402e+00 1.15180969e+00 4.45813965e-03
-2.35772863e-01 -1.06213069e+00 3.54408503e-01 3.31113011e-01
2.82684296e-01 4.64224041e-01 1.38478243e+00 -9.96686459e-01
-4.20998573e-01 -3.09205204e-01 -2.16153771e-01 2.31610119e-01
3.34519833e-01 1.25628173e-01 -1.09690130e+00 -3.28414351e-01
-3.42815489e-01 -3.68706048e-01 9.73273396e-01 5.06324589e-01
1.32614434e+00 -1.03020482e-01 -6.97574735e-01 9.59193468e-01
1.59497440e+00 4.24841017e-01 7.29340315e-01 2.81518608e-01
3.88973653e-01 2.21085668e-01 2.07824394e-01 2.81820029e-01
1.28475845e-01 6.83167279e-01 3.32144558e-01 -2.23516300e-01
1.06884889e-01 -5.08359671e-02 5.04009426e-01 7.46767938e-01
1.58365086e-01 -6.53205931e-01 -9.38449085e-01 3.72167200e-01
-1.86037111e+00 -7.71670878e-01 1.35726795e-01 2.17420578e+00
7.98601806e-01 -2.72178352e-01 1.71514452e-02 -2.86397040e-01
4.56990421e-01 -3.05065393e-01 -8.96513641e-01 -4.87049967e-01
-2.69840151e-01 1.79782450e-01 4.12776738e-01 1.30885690e-01
-1.00205576e+00 6.04915261e-01 6.84901762e+00 6.65347219e-01
-1.28349435e+00 1.24879502e-01 8.33936393e-01 -4.41927224e-01
-2.78292388e-01 -9.24358666e-02 -9.88444686e-01 6.75752401e-01
1.55978930e+00 5.76623250e-03 1.29555166e-01 6.11315548e-01
1.17616579e-01 8.51367563e-02 -1.37022662e+00 7.75339723e-01
-1.75148904e-01 -1.71455991e+00 1.41021814e-02 2.66312391e-01
5.42072296e-01 6.35762930e-01 1.31633461e-01 2.64850378e-01
4.23747331e-01 -1.23850262e+00 3.21661353e-01 6.61381185e-01
8.50588739e-01 -4.74655241e-01 9.73687351e-01 3.84683341e-01
-5.80443025e-01 -1.37405992e-01 -2.75367707e-01 1.82236761e-01
-6.29018694e-02 9.48628187e-01 -1.05521190e+00 9.71550286e-01
4.70588058e-01 7.82910705e-01 -4.50324655e-01 9.80732799e-01
1.44951150e-01 4.58924025e-01 -2.86125213e-01 3.84686403e-02
7.93080181e-02 -1.66702136e-01 3.33886176e-01 1.35840940e+00
5.14155626e-01 -1.57281384e-01 -9.84394625e-02 8.26836765e-01
1.09993666e-01 2.08220780e-01 -5.28932393e-01 -6.06314123e-01
2.44204849e-01 1.18548465e+00 -4.55610663e-01 -3.84847999e-01
-3.36394906e-01 7.45773196e-01 3.66598487e-01 3.43306065e-01
-1.01666260e+00 -1.82008415e-01 1.00899172e+00 6.76790252e-02
7.93295279e-02 1.08635232e-01 -2.61105806e-01 -1.33526647e+00
-4.49220240e-01 -1.10582018e+00 7.76402771e-01 -7.28906691e-01
-1.51184130e+00 3.61972362e-01 -1.17227070e-01 -9.65480864e-01
-1.43375561e-01 -1.07147396e+00 -4.56463486e-01 9.50547516e-01
-1.43386269e+00 -9.50724363e-01 -1.54623052e-03 1.05230421e-01
3.52835745e-01 -3.11140299e-01 1.13131535e+00 4.84998554e-01
-1.08229601e+00 5.87328434e-01 6.22265935e-01 -1.14015639e-01
7.02957988e-01 -1.06113446e+00 6.56483173e-02 4.04726475e-01
-2.12183446e-01 1.06155527e+00 4.37866956e-01 -7.13181913e-01
-1.57245278e+00 -1.20041680e+00 7.47951806e-01 -3.68835151e-01
4.90060091e-01 -2.62370110e-01 -7.93644309e-01 8.37353885e-01
3.65688503e-01 -2.87873805e-01 1.47142363e+00 4.02551353e-01
-2.70266503e-01 8.32962170e-02 -9.09655035e-01 5.09381056e-01
6.32389963e-01 -3.92256498e-01 -2.79255986e-01 5.90097487e-01
6.42614245e-01 -4.15176868e-01 -1.25279331e+00 6.34240329e-01
7.68631637e-01 -8.81551385e-01 9.20706868e-01 -1.01756883e+00
7.96214342e-01 -2.88832188e-01 -3.84113878e-01 -1.49124241e+00
-4.86862391e-01 -1.33159205e-01 -1.45291254e-01 7.91209817e-01
6.04908347e-01 -1.03011775e+00 6.32235229e-01 3.02372843e-01
-4.08359349e-01 -1.27526605e+00 -6.95818186e-01 -6.87912822e-01
2.71887571e-01 -5.90116233e-02 6.45539224e-01 9.99048293e-01
8.06174427e-02 3.22298795e-01 -1.99847221e-01 -2.40643173e-02
2.74333686e-01 3.41391325e-01 7.00020313e-01 -9.74404931e-01
-6.17700100e-01 -5.74674308e-01 -4.41065311e-01 -5.80392063e-01
1.25702843e-01 -1.31880319e+00 -3.08145344e-01 -1.50378203e+00
4.27926064e-01 -3.02672476e-01 -7.89761186e-01 7.01803446e-01
-1.92076609e-01 -8.33486095e-02 -7.02672899e-02 2.34890953e-01
-2.77885735e-01 4.56261665e-01 8.54599237e-01 -1.00080319e-01
-1.86780736e-01 -2.73147672e-01 -8.03135931e-01 5.13044536e-01
9.29377437e-01 -3.65326822e-01 -1.46534398e-01 -3.18547457e-01
2.87656456e-01 1.51848108e-01 3.06663513e-01 -8.17418933e-01
2.05945924e-01 -1.62335679e-01 8.92395020e-01 -5.70160091e-01
4.84439224e-01 -3.21318179e-01 9.82062459e-01 6.71846449e-01
-3.32523614e-01 2.12375015e-01 5.37883937e-01 4.79862690e-01
-6.44188672e-02 3.79564650e-02 7.46719956e-01 -1.62770957e-01
-3.45828503e-01 4.35110897e-01 -3.99833173e-01 -5.24779618e-01
9.35023248e-01 -2.03503847e-01 -4.93625492e-01 3.64470482e-02
-6.74053788e-01 1.35264710e-01 4.43311810e-01 2.90372461e-01
4.46500689e-01 -1.17025638e+00 -7.39697993e-01 1.72269285e-01
2.03876838e-01 -4.04160321e-01 2.87906498e-01 1.27074659e+00
-4.89054292e-01 8.03721249e-01 -2.28966519e-01 -7.96803117e-01
-1.15415919e+00 6.06127024e-01 6.22911096e-01 -4.39230859e-01
-1.82471439e-01 6.34634554e-01 4.06831175e-01 -6.90714121e-01
-1.37399927e-01 -1.62503079e-01 1.30704880e-01 2.29507178e-01
3.85062486e-01 1.82186753e-01 3.95327955e-01 -2.71317840e-01
-3.80484343e-01 2.39995614e-01 -3.21331620e-01 3.87703359e-01
1.51067054e+00 2.13302225e-01 -3.41172487e-01 5.91276646e-01
1.42810214e+00 -4.26311940e-01 -6.47070408e-01 1.51908055e-01
-1.87412843e-01 -2.05604032e-01 7.33140260e-02 -1.10157192e+00
-8.68067980e-01 7.62122452e-01 6.28336370e-01 -9.81348902e-02
1.09298193e+00 -3.11269015e-01 6.74111784e-01 2.92195201e-01
7.98688978e-02 -7.68235087e-01 -2.40375772e-01 5.45012057e-01
6.43987894e-01 -1.06303334e+00 5.59425913e-02 -7.32247308e-02
-1.90192088e-01 9.83273506e-01 4.84798878e-01 1.90503925e-01
6.31612718e-01 3.04957360e-01 2.23757140e-02 -3.67672205e-01
-1.39760363e+00 3.81240696e-02 1.39950097e-01 8.90434980e-02
8.71553004e-01 1.32004350e-01 -1.99986786e-01 7.13661671e-01
7.53194243e-02 5.22291183e-01 1.33297473e-01 6.36604488e-01
-1.76267505e-01 -1.17438340e+00 -1.84832036e-01 6.47145987e-01
-8.17653358e-01 -2.29791418e-01 -5.22821486e-01 6.86580360e-01
1.82051897e-01 5.56629002e-01 1.76173877e-02 -4.12239522e-01
4.05492820e-02 4.84565973e-01 2.37398818e-01 -5.55245459e-01
-7.65715778e-01 1.27303511e-01 7.87952915e-02 -4.82427567e-01
-5.40596992e-02 -6.88352823e-01 -1.13437021e+00 -3.65742952e-01
-4.67069447e-01 9.58121009e-03 8.23452294e-01 7.85365999e-01
1.02194166e+00 7.26076245e-01 3.07086796e-01 -5.42948604e-01
-4.51089859e-01 -8.78765702e-01 -2.71629006e-01 1.52748212e-01
2.87276775e-01 -4.50634450e-01 -2.40043744e-01 -1.98741272e-01] | [5.972275733947754, 5.678502559661865] |
5e4173bb-d540-45cf-aee2-58baac73b8f1 | pele-scores-pelvic-x-ray-landmark-detection | 2305.04294 | null | https://arxiv.org/abs/2305.04294v2 | https://arxiv.org/pdf/2305.04294v2.pdf | PELE scores: Pelvic X-ray Landmark Detection by Pelvis Extraction and Enhancement | The pelvis, the lower part of the trunk, supports and balances the trunk. Landmark detection from a pelvic X-ray (PXR) facilitates downstream analysis and computer-assisted diagnosis and treatment of pelvic diseases. Although PXRs have the advantages of low radiation and reduced cost compared to computed tomography (CT) images, their 2D pelvis-tissue superposition of 3D structures confuses clinical decision-making. In this paper, we propose a PELvis Extraction (PELE) module that utilizes 3D prior anatomical knowledge in CT to guide and well isolate the pelvis from PXRs, thereby eliminating the influence of soft tissue. We conduct an extensive evaluation based on two public datasets and one private dataset, totaling 850 PXRs. The experimental results show that the proposed PELE module significantly improves the accuracy of PXRs landmark detection and achieves state-of-the-art performances in several benchmark metrics, thus better serving downstream tasks. | ['S. Kevin Zhou', 'Jianji Wang', 'Zhiwei Cheng', 'Huijie Hu', 'Heqin Zhu', 'Shitong Shao', 'Han Li', 'Zhen Huang'] | 2023-05-07 | null | null | null | null | ['computed-tomography-ct'] | ['methodology'] | [-6.73409319e-03 2.97367632e-01 -9.26770449e-01 1.05115017e-02
-8.37566257e-01 -3.55623543e-01 2.59819686e-01 4.57649857e-01
-3.24769616e-01 4.82092798e-01 1.57227486e-01 -7.06682086e-01
-1.82061881e-01 -9.01166141e-01 -3.62312198e-01 -5.65128684e-01
-2.55432606e-01 6.34959459e-01 3.62032026e-01 7.41239861e-02
-2.90921982e-02 8.01822603e-01 -5.17006338e-01 -1.46211132e-01
9.45840180e-01 9.86281693e-01 2.13095754e-01 3.08549643e-01
2.42377907e-01 6.69288695e-01 -1.48037791e-01 -4.56570953e-01
2.11758211e-01 -3.77698928e-01 -6.58478260e-01 -2.69344777e-01
-3.60405967e-02 -1.82525352e-01 -4.17458415e-01 7.73688376e-01
4.37685758e-01 -1.40462831e-01 6.04809523e-01 -6.83254719e-01
-9.39467102e-02 2.40785152e-01 -9.85239327e-01 1.77267686e-01
4.20148909e-01 4.24779914e-02 6.98157847e-01 -6.83088362e-01
8.58632326e-01 6.58661664e-01 1.02714658e+00 3.61476868e-01
-8.38407636e-01 -6.34240687e-01 -3.08281422e-01 -3.68374169e-01
-1.41745782e+00 -2.50206050e-02 3.29407245e-01 -1.37086466e-01
4.11017299e-01 6.30821407e-01 5.81374049e-01 7.40008175e-01
8.33117902e-01 7.30012894e-01 9.42722797e-01 -2.79602230e-01
8.91975686e-02 -4.36347395e-01 -1.22703083e-01 1.17218268e+00
3.91787380e-01 3.25146429e-02 -2.14993283e-01 -5.01824319e-01
1.10530734e+00 2.90058434e-01 -3.92703891e-01 -4.02950168e-01
-8.74659717e-01 3.10773700e-01 7.91853547e-01 1.09006748e-01
-5.66095889e-01 6.77975118e-02 6.36260092e-01 -4.10810530e-01
-3.00261024e-02 2.15657219e-01 1.42385364e-01 4.08317409e-02
-7.51569211e-01 -1.30593002e-01 4.54316169e-01 8.02606642e-01
-2.70957183e-02 -6.58552051e-01 -2.41351530e-01 7.45829940e-01
3.72685343e-01 5.74650943e-01 6.73506260e-01 -5.77988267e-01
7.02581286e-01 8.60606670e-01 -2.64559031e-01 -9.60724533e-01
-8.41583729e-01 -5.34495592e-01 -8.92969847e-01 -2.79212058e-01
4.17396963e-01 1.08250611e-01 -1.19844067e+00 1.19410431e+00
5.79473913e-01 1.25497252e-01 -3.66811514e-01 1.12559581e+00
1.15280974e+00 2.87921071e-01 2.36931622e-01 8.98663476e-02
1.86084282e+00 -9.00794744e-01 -3.91772598e-01 -4.23433542e-01
8.07737708e-01 -4.98059183e-01 7.13047981e-01 2.79593140e-01
-8.95481408e-01 1.42334048e-02 -9.93785501e-01 -1.79676995e-01
2.47235447e-01 5.89326262e-01 8.21965396e-01 6.88466549e-01
-3.37754935e-01 7.21076250e-01 -1.58152950e+00 -3.04803669e-01
6.04733527e-01 5.12819469e-01 -8.37119639e-01 -7.32447281e-02
-8.20324659e-01 1.15354276e+00 4.87814248e-02 1.81304932e-01
-5.95242918e-01 -7.17436731e-01 -1.08615577e+00 1.00463487e-01
7.56568313e-01 -7.32428193e-01 1.21579123e+00 2.28220195e-01
-1.64431965e+00 9.96110737e-01 2.98171602e-02 -2.68977612e-01
8.11611772e-01 -1.86728522e-01 -8.08786601e-02 4.96501148e-01
2.26704836e-01 1.16273649e-02 2.01804534e-01 -1.00933039e+00
-5.26149988e-01 -7.80111313e-01 -3.06944311e-01 2.97181755e-01
1.56400457e-01 -1.63702652e-01 -1.23471081e+00 -6.10210061e-01
8.66550267e-01 -1.05503309e+00 -6.76837802e-01 2.20090151e-01
-6.45233631e-01 2.92149276e-01 3.36013049e-01 -1.02695763e+00
1.32180703e+00 -1.99021006e+00 -1.03550017e-01 6.47668719e-01
4.75644171e-01 1.18823655e-01 7.43293703e-01 1.43474147e-01
9.88694876e-02 3.00437142e-03 -9.49165076e-02 -4.84522045e-01
-2.78541714e-01 3.71690899e-01 1.21527947e-01 1.02345586e+00
-9.67948586e-02 9.32540238e-01 -9.68943119e-01 -1.07758832e+00
4.05832469e-01 7.02302754e-02 -2.58416623e-01 -2.66512454e-01
4.58569676e-01 6.01818562e-01 -9.12221134e-01 1.11917245e+00
6.34856045e-01 -6.46609366e-01 5.89466453e-01 3.54961343e-02
2.28179917e-01 3.43789428e-01 -9.22242284e-01 1.88001740e+00
-4.23863977e-01 6.60815686e-02 1.86301276e-01 -4.42474246e-01
8.29775989e-01 1.98218480e-01 7.01485693e-01 -1.00138092e+00
3.53644669e-01 5.11020541e-01 1.95495151e-02 -2.98776239e-01
3.27696145e-01 -5.34166619e-02 -3.41247588e-01 1.32423431e-01
-1.89281642e-01 -6.23612255e-02 -4.01845649e-02 1.20284207e-01
1.33129835e+00 6.25894293e-02 8.67593884e-01 -3.97741020e-01
5.53052127e-01 3.54336686e-02 8.18268836e-01 5.11623740e-01
-2.22291574e-01 8.15310955e-01 6.81465387e-01 -2.07682207e-01
-3.70961905e-01 -1.30391347e+00 -5.41148007e-01 3.60683739e-01
7.16280639e-01 -5.10477543e-01 -2.55339593e-01 -8.11807752e-01
2.81467617e-01 2.22992286e-01 -7.48448193e-01 -4.49192598e-02
-9.17212069e-01 -7.46007383e-01 7.11446226e-01 6.82720602e-01
1.68093488e-01 -3.67766827e-01 -9.38238084e-01 7.26990774e-02
-3.17395210e-01 -1.18554699e+00 -1.81936488e-01 1.77723080e-01
-1.27391410e+00 -1.23816514e+00 -9.19936478e-01 -4.37436789e-01
9.06986296e-01 -1.54317841e-01 8.91633749e-01 2.22670749e-01
-4.96142387e-01 -1.80311844e-01 -1.49604782e-01 -1.02331452e-01
-1.73420608e-01 9.51532125e-02 -2.29080573e-01 -3.97459954e-01
-2.94545650e-01 -1.73578694e-01 -7.90039003e-01 4.44775075e-01
-4.45221782e-01 1.93188310e-01 7.32510865e-01 6.83233798e-01
1.01413405e+00 -3.31651449e-01 -2.69945282e-02 -1.01757467e+00
2.89844900e-01 -6.67711675e-01 -5.59414268e-01 2.25163743e-01
-3.14911097e-01 -3.39190364e-01 5.84238946e-01 1.25781476e-01
-1.05735898e+00 1.82171296e-02 -2.72871196e-01 -2.13468671e-01
2.52590060e-01 5.35879731e-01 2.00596124e-01 -1.68799251e-01
4.09708112e-01 -7.60653913e-02 -8.62947255e-02 -6.01868272e-01
-1.47849008e-01 2.69167989e-01 1.04370546e+00 -7.92071104e-01
2.77704209e-01 8.30073774e-01 6.58480346e-01 -4.83092099e-01
-4.82727736e-01 -6.05825603e-01 -4.34973508e-01 -7.97696561e-02
6.94568098e-01 -6.58978701e-01 -7.59588242e-01 -1.44570217e-01
-6.34736180e-01 2.17623599e-02 -3.84791493e-02 6.91858947e-01
-4.20124799e-01 5.50661206e-01 -1.02355170e+00 -3.27914298e-01
-6.20285213e-01 -1.45246685e+00 1.21051145e+00 6.45638108e-01
-4.42282319e-01 -7.59507060e-01 6.66041151e-02 1.94167167e-01
4.25866991e-02 8.08022857e-01 8.12990606e-01 -3.17220509e-01
-4.24170375e-01 -5.37873149e-01 -1.55431896e-01 -4.35615361e-01
1.07417345e-01 -9.00024623e-02 -6.99294865e-01 1.12000637e-01
-1.24854423e-01 9.82374549e-02 6.45324647e-01 6.01861715e-01
1.18613148e+00 3.12642455e-01 -1.06380033e+00 9.79446650e-01
1.53933620e+00 1.85551465e-01 5.95741093e-01 4.25994128e-01
5.47767282e-01 2.56508827e-01 9.16143298e-01 4.80374724e-01
4.19019222e-01 6.04587197e-01 6.56729817e-01 -3.39932770e-01
-1.68627482e-02 -3.41298908e-01 -3.10905159e-01 4.70182836e-01
-3.41868907e-01 -2.52983291e-02 -1.49995017e+00 3.93102139e-01
-1.58295369e+00 -9.78813991e-02 -4.43940759e-01 2.14207935e+00
7.94130564e-01 2.99650222e-01 -3.62541497e-01 -2.18752220e-01
5.01428962e-01 -7.06072375e-02 -2.77528912e-01 -7.67153949e-02
5.60554981e-01 5.49687624e-01 9.95830476e-01 2.84026921e-01
-1.20840120e+00 4.08522069e-01 5.84603643e+00 8.10771108e-01
-1.27436006e+00 8.90373737e-02 6.48890853e-01 8.49963129e-02
2.00824529e-01 -1.58429295e-01 -7.16927588e-01 4.56552982e-01
3.09523225e-01 -6.37648953e-03 -4.26217616e-01 8.56004477e-01
-3.93010862e-03 -7.21112967e-01 -1.18006337e+00 7.85740077e-01
-3.68057758e-01 -1.42634618e+00 -6.15572929e-01 2.37383381e-01
4.29202676e-01 1.16282493e-01 -1.38189103e-02 2.07492411e-01
8.35392475e-02 -1.12494922e+00 3.70984077e-01 4.02505189e-01
1.03750038e+00 -6.86745167e-01 9.50793445e-01 4.90080237e-01
-1.33899021e+00 2.69358397e-01 7.10789785e-02 2.27926582e-01
1.57400116e-01 4.00192827e-01 -1.01839888e+00 1.20351410e+00
4.96452510e-01 2.37792850e-01 -4.83313650e-01 1.27245474e+00
-3.94607723e-01 2.56282836e-01 -7.35161245e-01 3.49375337e-01
2.26249740e-01 -1.41105011e-01 3.34309012e-01 1.02881324e+00
2.17391804e-01 3.98772091e-01 8.13022926e-02 5.26000798e-01
-2.08424315e-01 3.11644077e-01 -3.35689545e-01 3.84594738e-01
5.04666448e-01 1.33143210e+00 -1.48107135e+00 -2.78810203e-01
-2.97205538e-01 5.57544649e-01 1.00937411e-01 -2.22857490e-01
-9.83762860e-01 -8.68540034e-02 1.56181648e-01 1.55766606e-01
-9.06721875e-02 9.07561928e-02 -7.94578969e-01 -9.56523955e-01
2.23642409e-01 -4.96743321e-01 7.14649022e-01 -2.50144392e-01
-7.67817914e-01 2.97155112e-01 -4.21308428e-02 -1.41165757e+00
-9.57179517e-02 -5.91149092e-01 -7.44416296e-01 8.30840647e-01
-1.33034682e+00 -8.52876067e-01 -5.00209033e-01 1.62957822e-05
-6.40119463e-02 5.94006181e-01 9.35814619e-01 4.09731954e-01
-6.65082514e-01 5.92664957e-01 1.07498683e-01 4.84495401e-01
5.21528304e-01 -1.21132851e+00 2.54298270e-01 6.02333128e-01
-5.94178557e-01 8.61563563e-01 1.78882226e-01 -9.18259799e-01
-1.74931192e+00 -9.08615291e-01 6.68847680e-01 -4.16166604e-01
2.46756732e-01 -6.19736798e-02 -4.97283190e-01 7.86238611e-01
-4.47894037e-01 5.27599275e-01 7.61382043e-01 -3.15204747e-02
1.76439002e-01 1.18280634e-01 -1.19804263e+00 6.29826128e-01
7.43229687e-01 -3.48883495e-02 -5.65395951e-01 7.97408223e-02
-3.03925183e-02 -1.46465933e+00 -1.26192546e+00 9.03898835e-01
7.98452079e-01 -7.53590465e-01 1.35361814e+00 -8.21664333e-02
5.58228612e-01 -1.61490738e-01 2.60214537e-01 -7.90139616e-01
1.04567700e-03 -2.10268572e-01 -1.68459877e-01 6.27802193e-01
2.25268543e-01 -4.91225123e-01 1.15962505e+00 7.02071965e-01
-3.41010571e-01 -1.21973085e+00 -1.02842581e+00 -6.26207829e-01
-1.43211365e-01 -2.79531598e-01 4.31173354e-01 1.07457113e+00
1.69148818e-01 -1.23492323e-01 1.71367630e-01 3.37086558e-01
4.80185062e-01 3.83253038e-01 6.52100086e-01 -9.83556449e-01
-4.21315402e-01 -5.67550480e-01 -4.23766315e-01 -9.47726607e-01
-5.34953952e-01 -1.13132977e+00 -1.54805675e-01 -1.79889667e+00
2.81323344e-01 -8.33229661e-01 -1.35305613e-01 4.47754800e-01
-4.33025807e-01 2.67082125e-01 1.82272792e-01 2.51885295e-01
-4.26253051e-01 3.32469851e-01 1.81488562e+00 4.37905014e-01
-6.54072091e-02 2.46130541e-01 -3.87066513e-01 1.21791112e+00
3.83264780e-01 -7.54907727e-01 -6.56449422e-02 -2.12998524e-01
-2.34694943e-01 7.01376379e-01 -2.70973127e-02 -7.40941286e-01
1.73368797e-01 -5.08175269e-02 8.19893479e-01 -8.48376751e-01
4.40953821e-01 -7.81522930e-01 7.15291649e-02 1.09997618e+00
3.66778791e-01 -1.47203237e-01 3.25089842e-01 4.00716186e-01
-1.37605295e-01 -3.32011282e-01 8.35394919e-01 -4.35278535e-01
-3.74965101e-01 2.32526556e-01 -2.02272207e-01 -1.00031063e-01
1.26171601e+00 -6.81956410e-02 -1.83136374e-01 5.85004091e-02
-8.27882349e-01 5.94544590e-01 5.01704514e-01 -3.43838297e-02
5.55452585e-01 -1.20989883e+00 -4.06002671e-01 9.12330002e-02
6.56283945e-02 7.36388505e-01 2.29291365e-01 1.33678734e+00
-1.46411753e+00 5.60942531e-01 -1.72771737e-01 -9.61139560e-01
-1.16936326e+00 9.26460326e-02 4.96402442e-01 -9.29083109e-01
-1.12521768e+00 8.93676341e-01 3.85202348e-01 -6.33881569e-01
2.14861795e-01 -7.09371805e-01 -1.27709091e-01 -2.75791794e-01
2.77187645e-01 3.81036252e-01 3.15727651e-01 -5.60712397e-01
-7.59219289e-01 4.01856542e-01 1.65017527e-02 3.33438605e-01
1.02763724e+00 -2.19538026e-02 3.45111750e-02 -2.27992237e-01
9.31729078e-01 4.08725947e-01 -7.64371216e-01 -3.37869674e-01
3.22026074e-01 -8.29676151e-01 5.57570830e-02 -6.61164403e-01
-9.30942476e-01 5.40085852e-01 2.19716355e-01 -3.59914094e-01
1.11694491e+00 1.11245364e-01 8.65682125e-01 -1.63471363e-02
5.79908550e-01 -6.88132226e-01 -2.26564586e-01 1.71653211e-01
4.75549877e-01 -1.13717008e+00 6.35115743e-01 -9.56547379e-01
-6.65308535e-01 1.13677704e+00 6.37386918e-01 -3.41297805e-01
4.79677022e-01 5.15542626e-01 -9.48191509e-02 -3.55073452e-01
-2.51780838e-01 1.61678404e-01 3.50412667e-01 1.12236761e-01
2.99109846e-01 5.18609762e-01 -5.31454027e-01 9.32108343e-01
-2.21751571e-01 3.16386558e-02 5.45958839e-02 1.45350444e+00
-9.96172205e-02 -9.42205191e-01 -5.97949147e-01 5.06773591e-01
-9.77683783e-01 1.47806540e-01 -1.12050682e-01 1.33805382e+00
-4.53339554e-02 1.81662738e-01 -1.52178317e-01 1.73288018e-01
3.78539503e-01 -3.89588892e-01 6.05389714e-01 -7.52345085e-01
-6.60817802e-01 2.92489439e-01 2.06516147e-01 -7.51257837e-01
3.28276217e-01 -6.08598053e-01 -1.87301207e+00 -1.44293383e-01
-5.73951662e-01 -4.26379442e-02 6.86602294e-01 8.92494261e-01
1.00019043e-02 7.67282367e-01 3.73634219e-01 -5.71491838e-01
-5.63512027e-01 -6.91666663e-01 -4.79225934e-01 7.75785670e-02
2.80291051e-01 -9.23008263e-01 4.74188030e-01 -2.87124664e-01] | [14.68491268157959, -2.552248477935791] |
f2f22aad-edf1-4946-83f7-7a33246808f5 | can-boosting-with-svm-as-week-learners-help | 1604.05242 | null | http://arxiv.org/abs/1604.05242v2 | http://arxiv.org/pdf/1604.05242v2.pdf | Can Boosting with SVM as Week Learners Help? | Object recognition in images involves identifying objects with partial
occlusions, viewpoint changes, varying illumination, cluttered backgrounds.
Recent work in object recognition uses machine learning techniques SVM-KNN,
Local Ensemble Kernel Learning, Multiple Kernel Learning. In this paper, we
want to utilize SVM as week learners in AdaBoost. Experiments are done with
classifiers like near- est neighbor, k-nearest neighbor, Support vector
machines, Local learning(SVM- KNN) and AdaBoost. Models use Scale-Invariant
descriptors and Pyramid his- togram of gradient descriptors. AdaBoost is
trained with set of week classifier as SVMs, each with kernel distance function
on different descriptors. Results shows AdaBoost with SVM outperform other
methods for Object Categorization dataset. | ['Dinesh Govindaraj'] | 2016-04-18 | null | null | null | null | ['object-categorization'] | ['computer-vision'] | [-5.88397622e-01 -1.07326901e+00 -5.10730088e-01 -7.45348871e-01
-3.00687969e-01 -8.04513991e-01 4.64021295e-01 6.02549672e-01
-3.27683389e-01 7.07264960e-01 3.03962708e-01 -2.07231238e-01
-4.79585648e-01 -6.60147369e-01 -1.40635222e-01 -5.75736821e-01
-3.36051524e-01 2.32191369e-01 9.23809350e-01 -1.88409165e-01
9.70815897e-01 1.09729576e+00 -1.60668421e+00 7.99834430e-01
4.19070870e-01 1.09691620e+00 -3.99284303e-01 1.33087134e+00
-2.41643339e-01 8.35368156e-01 -5.80019653e-01 2.10474014e-01
4.66346502e-01 -1.07123546e-01 -6.63099706e-01 3.94220054e-02
8.04787636e-01 -3.23221385e-01 1.42958183e-02 4.50651765e-01
4.67850536e-01 5.82103491e-01 1.31959164e+00 -1.49400759e+00
-1.08488250e+00 -4.40977931e-01 -4.71429735e-01 8.75028908e-01
5.26563883e-01 -1.83319002e-01 2.92596430e-01 -9.95001614e-01
4.94645350e-02 1.31740272e+00 1.11307454e+00 6.42488077e-02
-1.32737160e+00 -3.42567861e-01 -3.02990347e-01 8.24209630e-01
-1.13742268e+00 -2.05308925e-02 5.62887251e-01 -8.05121899e-01
1.40487337e+00 5.65920651e-01 4.22909915e-01 5.01830339e-01
6.68725431e-01 4.99718875e-01 1.72321165e+00 -9.17847693e-01
3.12008232e-01 8.82386744e-01 1.35422909e+00 7.77684987e-01
-1.83466449e-01 3.06978047e-01 -5.11664093e-01 -8.60024333e-01
2.45856151e-01 3.95722061e-01 3.49037021e-01 -3.48866999e-01
-7.84883320e-01 8.32721114e-01 3.51521611e-01 3.96009356e-01
-1.83021754e-01 -3.55726570e-01 7.08478212e-01 8.83891106e-01
3.38257402e-01 -2.38880571e-02 -8.02700520e-01 3.08924824e-01
-4.69094723e-01 3.10368061e-01 7.71212280e-01 5.36972523e-01
8.58128607e-01 1.60196912e-03 -2.11212859e-01 1.11079192e+00
2.47691274e-01 3.27021420e-01 1.07866180e+00 -9.63669568e-02
-1.84735298e-01 6.28470898e-01 4.48126942e-01 -9.00677025e-01
-4.68957961e-01 4.34792638e-01 -1.87184438e-01 1.06044996e+00
4.51651067e-01 1.85234293e-01 -1.03315914e+00 2.91192353e-01
8.56019676e-01 5.81754684e-01 4.30274010e-01 5.31392574e-01
9.27606523e-01 8.96455288e-01 1.48803368e-01 8.18384662e-02
1.38299179e+00 -1.39073658e+00 -3.08951706e-01 3.24634224e-01
6.30165040e-01 -1.32942271e+00 6.54441178e-01 5.76269448e-01
-1.41733706e-01 -9.86906946e-01 -1.02269065e+00 4.74810570e-01
-1.16467512e+00 2.29594052e-01 6.39245510e-01 8.94841671e-01
-7.21888185e-01 6.64188027e-01 -5.47637105e-01 -7.41092682e-01
2.30461881e-01 6.00169480e-01 -5.92380464e-01 9.27241147e-02
-4.43305016e-01 1.39579606e+00 7.04984963e-01 -5.15373409e-01
-3.77137721e-01 -4.04734433e-01 -6.38803720e-01 -4.84559923e-01
-4.88438070e-01 3.13854031e-02 7.78092444e-01 -1.33694553e+00
-1.53322673e+00 9.86692846e-01 -1.51578739e-01 -3.20224732e-01
1.08686008e-01 2.79763430e-01 -8.55609596e-01 -2.16330722e-01
-2.99917877e-01 -4.37109768e-02 1.02922261e+00 -8.10420215e-01
-7.20795274e-01 -7.82259703e-01 -3.08816731e-01 7.06817396e-03
-2.58731753e-01 5.02827883e-01 1.16002524e+00 -4.58915621e-01
3.13103974e-01 -8.88701677e-01 2.10366502e-01 -7.46016353e-02
9.24555510e-02 -9.96938467e-01 1.72213650e+00 -7.62097239e-01
5.79385638e-01 -2.18144894e+00 -4.47569340e-01 3.19313645e-01
-4.13060367e-01 4.11990583e-01 1.68330789e-01 5.26844800e-01
-2.94878125e-01 -5.73862910e-01 6.85753763e-01 4.29894507e-01
-2.88981646e-01 4.00816858e-01 -2.50665069e-01 8.53074253e-01
-1.71594813e-01 6.85309887e-01 -2.46622294e-01 -6.50745928e-01
6.34156287e-01 -5.78382872e-02 2.84666777e-01 2.74060875e-01
6.58917725e-01 -2.03007460e-01 -3.67897391e-01 1.00268018e+00
7.51450062e-01 2.81202912e-01 -5.70685089e-01 -3.42866749e-01
-2.02948406e-01 -7.95807183e-01 -1.50035787e+00 1.00797200e+00
-2.67509937e-01 8.67935658e-01 -6.16779685e-01 -1.64199400e+00
1.27885711e+00 2.22826213e-01 -6.04348592e-02 -8.82618278e-02
1.23551212e-01 4.95828450e-01 -7.78866485e-02 -8.22913766e-01
-2.87959874e-01 1.64288327e-01 5.25392354e-01 -1.71063505e-02
4.10339236e-01 4.49462593e-01 -2.87733912e-01 -3.77397388e-01
8.39836538e-01 -1.37792856e-01 7.75588036e-01 -5.69056034e-01
8.17153990e-01 4.07879204e-01 -1.39001727e-01 9.66852963e-01
-6.65509820e-01 1.10950999e-01 -2.64669716e-01 -1.34809303e+00
-1.10511732e+00 -1.37902629e+00 -6.70771122e-01 1.60195947e+00
1.50312051e-01 2.35696554e-01 -1.22768618e-01 -9.96165395e-01
7.02029943e-01 2.87328631e-01 -5.99516511e-01 -1.40226737e-01
-4.70039815e-01 -5.67793608e-01 1.71985731e-01 7.39999354e-01
3.78566593e-01 -7.51177967e-01 -3.77344847e-01 3.21151435e-01
9.50702488e-01 -4.53770518e-01 -1.18359245e-01 3.26315910e-01
-1.00499415e+00 -9.72114801e-01 -6.92524493e-01 -1.11049855e+00
7.38421738e-01 3.95650357e-01 4.86424476e-01 -2.68418640e-01
-1.06183696e+00 6.42156124e-01 -7.47650564e-01 -7.58925080e-01
-1.02032810e-01 -6.23886287e-01 3.81737322e-01 -1.06283441e-01
1.10046709e+00 -6.00823581e-01 -4.98879731e-01 5.38801253e-01
-2.28792399e-01 -7.27635384e-01 3.62470746e-01 1.23178267e+00
-3.97585556e-02 -1.38184875e-01 4.69509095e-01 -4.03969556e-01
5.04181027e-01 -5.20551026e-01 -7.56225765e-01 6.91272855e-01
-5.57224274e-01 -1.36538342e-01 6.50186002e-01 -8.36666405e-01
-6.19796574e-01 2.90977620e-02 5.41916728e-01 -1.98038101e-01
-7.00907171e-01 -2.70639472e-02 5.21139264e-01 -8.42309773e-01
1.67066300e+00 5.81722438e-01 -9.21025574e-02 -3.96159619e-01
3.22734453e-02 1.32297051e+00 -6.94997460e-02 -5.01227379e-01
6.83831036e-01 4.44505513e-01 -4.89390530e-02 -9.54291761e-01
-6.03952467e-01 -1.29036629e+00 -9.26204681e-01 -4.17989224e-01
8.68944585e-01 -7.05236495e-01 -5.41099489e-01 8.44724834e-01
-1.08863866e+00 1.77615032e-01 2.17232004e-01 8.11837614e-01
-1.95838898e-01 1.16264291e-01 -3.17596197e-01 -1.07829416e+00
-2.16917709e-01 -5.82557142e-01 4.95118797e-01 5.71120143e-01
-9.15429816e-02 -1.18439293e+00 1.06466621e-01 3.12505066e-01
5.12564182e-01 3.97635847e-01 7.06083596e-01 -1.30458856e+00
-7.93992728e-03 -8.11681569e-01 -2.70912856e-01 7.23248482e-01
7.20204055e-01 1.27194211e-01 -9.71059620e-01 -2.62373269e-01
-2.32423738e-01 -6.41188920e-01 7.33598769e-01 3.28099787e-01
9.45298731e-01 -4.43866342e-01 -4.21669602e-01 3.12781543e-01
1.71001148e+00 7.50919342e-01 2.29215860e-01 6.93289340e-01
3.42075944e-01 2.67937034e-01 5.57060063e-01 3.49770546e-01
-8.48841388e-03 5.64576268e-01 -3.25369567e-01 2.74407119e-01
3.29246074e-01 4.35233355e-01 2.01373562e-01 2.85995781e-01
-2.84481972e-01 2.83789307e-01 -1.11410701e+00 7.01469705e-02
-2.03831148e+00 -1.15879059e+00 -2.89611608e-01 2.02187324e+00
3.91455203e-01 6.10345323e-03 4.84773815e-01 3.58433366e-01
7.37609148e-01 -1.94458380e-01 -3.07562888e-01 -1.01953590e+00
-1.73622102e-01 1.99557915e-01 6.58073664e-01 5.13498306e-01
-1.46607351e+00 7.24794745e-01 6.24673176e+00 9.08619881e-01
-1.29916716e+00 2.86338568e-01 4.31089193e-01 4.61263806e-01
5.65273941e-01 2.84034669e-01 -1.06791031e+00 5.15472770e-01
6.92704797e-01 -1.94666088e-01 1.88275442e-01 1.47448099e+00
-6.44332170e-02 -5.20390511e-01 -8.95266294e-01 1.15015769e+00
5.22760749e-01 -1.30428183e+00 -1.75523907e-01 -5.04092872e-01
7.85695136e-01 -8.27202648e-02 -1.94424197e-01 4.00276482e-01
2.58021146e-01 -9.46179628e-01 1.25594571e-01 5.37543654e-01
-7.63389543e-02 -5.77668250e-01 7.12237656e-01 2.88738370e-01
-1.03052926e+00 -4.28257018e-01 -7.99602151e-01 -2.30965361e-01
-6.33785248e-01 2.42804974e-01 -9.50758755e-01 1.25913218e-01
9.48078990e-01 4.27364677e-01 -8.01861167e-01 1.39378154e+00
6.81793272e-01 5.13669074e-01 -4.29506153e-01 -6.40067577e-01
3.11421931e-01 -3.39876004e-02 2.51853138e-01 1.60972869e+00
9.74697247e-02 3.84342879e-01 8.02841187e-01 -1.43376112e-01
8.76320124e-01 6.90943420e-01 -6.49181247e-01 5.81868172e-01
2.63976604e-01 1.13489962e+00 -8.33948076e-01 -7.99196482e-01
-4.65480894e-01 1.26097679e+00 1.96188793e-01 2.47042745e-01
-5.92425168e-01 -8.16698551e-01 3.97246957e-01 -1.16029553e-01
1.91969439e-01 -2.75699854e-01 -1.67238846e-01 -8.10823679e-01
-2.10055951e-02 -3.86718899e-01 7.23842680e-01 -8.20982039e-01
-1.82780182e+00 3.37949187e-01 1.82778507e-01 -1.14674401e+00
2.65229791e-01 -1.41196728e+00 -1.02070916e+00 8.82341981e-01
-9.49714720e-01 -1.52079499e+00 -4.64545846e-01 9.51807797e-01
8.45642626e-01 -1.08535397e+00 9.39136147e-01 -3.84851210e-02
1.12394936e-01 4.03706193e-01 9.16563749e-01 1.69717222e-01
1.14199543e+00 -1.46003830e+00 -5.19293487e-01 9.07123759e-02
-2.06618644e-02 3.67147475e-01 6.11932933e-01 -3.13263804e-01
-1.09489369e+00 -8.04989755e-01 4.05661583e-01 -6.42670095e-01
6.04443252e-01 2.80256364e-02 -7.25693166e-01 5.13358891e-01
4.03106958e-02 1.00528383e+00 1.19574416e+00 1.51837602e-01
-7.01627314e-01 -7.16546595e-01 -1.49180889e+00 -2.55887032e-01
6.89864382e-02 -4.95589107e-01 -8.27821255e-01 1.03542721e+00
-8.08998644e-02 -4.06118482e-01 -1.00000894e+00 7.51926601e-02
1.18898082e+00 -1.17925847e+00 1.28087163e+00 -1.49060690e+00
-4.12564993e-01 -3.32147449e-01 -5.00928283e-01 -1.28797150e+00
-4.17771667e-01 2.36088023e-01 2.34252959e-01 1.11829698e+00
7.21352994e-02 -9.30543602e-01 6.91428185e-01 3.90823722e-01
1.60289392e-01 -5.85536242e-01 -1.05511975e+00 -1.43519092e+00
-4.87262271e-02 2.79676259e-01 1.69071242e-01 1.36378169e+00
-8.57903808e-02 -5.07732434e-03 -3.37598860e-01 4.57527399e-01
6.18630826e-01 2.48123974e-01 8.86303961e-01 -1.43922079e+00
-4.68913406e-01 -8.52104574e-02 -1.63848424e+00 -1.58144683e-02
4.47361767e-01 -7.75826633e-01 -5.20147085e-01 -1.29684615e+00
-5.76783791e-02 -4.14590925e-01 -5.82742274e-01 6.40416861e-01
5.36380969e-02 2.56150007e-01 -1.96325347e-01 2.17167586e-01
-2.74442852e-01 -9.52095836e-02 6.08279347e-01 -3.14312130e-01
-3.75015497e-01 1.95925400e-01 2.79319465e-01 7.45756924e-01
9.49837506e-01 -6.08690858e-01 -2.66153008e-01 1.32629508e-02
-7.25748956e-01 -1.93913490e-01 7.07308352e-01 -1.19081283e+00
3.01855028e-01 -7.88239419e-01 1.14355099e+00 -4.40317929e-01
2.66324878e-01 -7.41939545e-01 -2.69293934e-01 6.81763232e-01
1.07543739e-02 1.85749099e-01 2.14314386e-01 4.58014876e-01
-2.02026740e-01 -7.66579628e-01 1.17219079e+00 -1.36000469e-01
-1.30885744e+00 -8.81405100e-02 -6.67185605e-01 -6.78939641e-01
1.73903203e+00 -9.59883034e-01 -2.86086947e-01 3.09394032e-01
-1.16065824e+00 -2.06063479e-01 -3.69536988e-02 4.16028559e-01
5.24312258e-01 -1.59348500e+00 -7.33889401e-01 8.42322558e-02
5.97805142e-01 -9.47819769e-01 1.92930400e-01 5.13849199e-01
-9.64410841e-01 2.05015466e-01 -8.56210351e-01 -6.46698296e-01
-2.05981421e+00 5.21918416e-01 3.40759039e-01 3.47650796e-01
-1.92934453e-01 1.11672235e+00 -5.70984602e-01 -5.33486485e-01
-2.08111014e-02 -1.95740700e-01 -4.51909900e-01 -1.85569804e-02
5.41235626e-01 1.02250135e+00 1.14929929e-01 -5.00229180e-01
-5.67112923e-01 9.48936701e-01 -3.56483012e-01 3.19474995e-01
1.04502249e+00 3.97893041e-01 -4.87235814e-01 9.26704884e-01
1.65216923e+00 -3.16284418e-01 -2.63236433e-01 -1.92364946e-01
2.73798257e-01 -9.53370333e-01 -1.58300325e-01 -8.25587511e-01
6.97757378e-02 4.51336384e-01 1.88456011e+00 3.43986630e-01
6.61790252e-01 -2.93571115e-01 1.12110265e-01 1.10870397e+00
2.74644524e-01 -1.34920371e+00 1.94080517e-01 3.79867345e-01
7.40557849e-01 -1.98080266e+00 2.63702154e-01 -1.05317846e-01
-2.90245801e-01 1.86447001e+00 1.00671566e+00 -9.73841310e-01
1.57008195e+00 8.61135032e-03 4.46872532e-01 2.36522451e-01
-6.40183091e-01 1.76895931e-01 5.47416270e-01 9.11383033e-01
2.44813427e-01 1.12507388e-01 -4.58970785e-01 -1.94998756e-02
1.03071235e-01 1.32978195e-02 1.37288809e-01 1.30319417e+00
-1.07535136e+00 -9.72827613e-01 -8.35916877e-01 9.08941805e-01
-9.83272865e-02 2.51682013e-01 -1.06809273e-01 5.60556889e-01
4.37454879e-01 8.56998861e-01 9.59347636e-02 -1.91498846e-01
3.44649971e-01 3.56863260e-01 6.07898653e-01 -3.85813892e-01
-5.43198109e-01 -4.21587765e-01 -9.15307328e-02 -9.89179909e-02
-2.38547072e-01 -6.84236169e-01 -9.01808798e-01 -1.56129971e-01
-6.41289771e-01 2.88250208e-01 9.47270751e-01 1.00256932e+00
-1.76613435e-01 -2.67750978e-01 8.36690843e-01 -7.17188835e-01
-9.00493979e-01 -1.00040901e+00 -9.03889716e-01 2.86918193e-01
5.12649238e-01 -7.93209493e-01 -4.87964213e-01 2.73686290e-01] | [8.162426948547363, 3.9057109355926514] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.