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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
989e6108-a003-4129-a9a6-e4b7cae24c0d | self-supervised-sparse-representation-for | null | null | https://www.ecva.net/papers/eccv_2022/papers_ECCV/papers/136730727.pdf | https://www.ecva.net/papers/eccv_2022/papers_ECCV/papers/136730727.pdf | Self-supervised Sparse Representation for Video Anomaly Detection | Video anomaly detection (VAD) aims at localizing unexpected actions or activities in a video sequence. Existing mainstream VAD techniques are based on either the one-class formulation, which assumes all training data are normal, or weakly-supervised, which requires only video-level normal/anomaly labels. To establish a unified approach to solving the two VAD settings, we introduce a self-supervised sparse representation (S3R) framework that models the concept of anomaly at feature level by exploring the synergy between dictionary-based representation and self-supervised learning. With the learned dictionary, S3R facilitates two coupled modules, en-Normal and de-Normal, to reconstruct snippet-level features and filter out normal-event features. The self-supervised techniques also enable generating samples of pseudo normal/anomaly to train the anomaly detector. We demonstrate with extensive experiments that S3R achieves new state-of-the-art performances on popular benchmark datasets for both one-class and weakly-supervised VAD tasks. Our code is publicly available at https://github.com/louisYen/S3R. | ['Tyng-Luh Liu', 'Chiou-Shann Fuh', 'Ding-Jie Chen', 'He-Yen Hsieh*', 'Jhih-Ciang Wu*'] | 2022-10-23 | null | null | null | eccv-2022-2022-10 | ['anomaly-detection-in-surveillance-videos', 'anomaly-detection-in-surveillance-videos'] | ['computer-vision', 'methodology'] | [-6.37345575e-03 -4.66013908e-01 -3.69565129e-01 -3.18889976e-01
-5.83904147e-01 -2.93919623e-01 5.61469853e-01 9.85635146e-02
7.10393339e-02 3.32961828e-02 3.49369347e-01 -1.31651148e-01
2.71299601e-01 -5.01379967e-01 -6.36883199e-01 -5.91224730e-01
-4.06559646e-01 1.80350468e-01 9.11271274e-02 -3.55582610e-02
7.74829760e-02 3.76413614e-01 -1.78353655e+00 4.91609722e-01
6.64463699e-01 1.29435754e+00 -2.88096011e-01 6.74204111e-01
-1.28390238e-01 1.23816180e+00 -3.36875379e-01 4.47841622e-02
5.52110672e-01 -7.59634197e-01 -4.69194978e-01 4.04889464e-01
3.71192157e-01 -5.02559304e-01 -9.57398117e-01 1.21982992e+00
1.86897740e-01 9.44258422e-02 7.19212472e-01 -1.72531784e+00
-5.97500980e-01 -6.36735978e-03 -6.23423636e-01 8.63174260e-01
8.38312387e-01 4.15885061e-01 1.09039807e+00 -1.17698467e+00
4.21502411e-01 1.00685239e+00 5.36532342e-01 7.43425786e-01
-1.01915574e+00 -4.34749275e-01 3.89452308e-01 5.54191351e-01
-1.39057267e+00 -5.10488987e-01 9.23554838e-01 -7.32596695e-01
9.53062594e-01 1.68063670e-01 7.72624135e-01 1.61699545e+00
5.67005090e-02 1.35051692e+00 5.65089583e-01 -1.51047766e-01
3.65000814e-01 -3.16394567e-01 1.80841669e-01 8.88086736e-01
1.59500226e-01 9.03667137e-02 -8.44120741e-01 -4.58225697e-01
6.38009012e-01 5.39861321e-01 -3.10331792e-01 -5.94448328e-01
-1.24520195e+00 7.77725160e-01 -2.33340710e-01 2.42992982e-01
-5.34065902e-01 -1.97282329e-01 9.60245073e-01 7.39463985e-01
5.18261313e-01 5.15062772e-02 -2.35420570e-01 -3.60388815e-01
-7.98845410e-01 4.43555973e-02 4.64146137e-01 8.46667230e-01
6.19206250e-01 5.21656871e-01 -3.22090447e-01 6.37767196e-01
5.26686668e-01 4.95191872e-01 9.38943923e-01 -7.95108438e-01
2.16105402e-01 7.28036702e-01 -2.08358213e-01 -1.05966949e+00
-8.42743029e-04 -2.18493998e-01 -9.42876041e-01 2.35586405e-01
2.06464186e-01 1.66070893e-01 -1.06778753e+00 1.58340228e+00
2.58592963e-01 9.90973532e-01 2.16409713e-01 8.18643630e-01
9.01555240e-01 7.59224713e-01 -1.93658844e-01 -4.27048266e-01
8.62731874e-01 -1.09207392e+00 -8.44812930e-01 -1.83699682e-01
8.41510892e-01 -4.69481617e-01 1.21702790e+00 5.32818854e-01
-8.61513913e-01 -3.70268196e-01 -8.71359885e-01 2.06960663e-01
-4.49187867e-02 -5.85699156e-02 2.33401492e-01 2.37073094e-01
-9.13294911e-01 3.39394838e-01 -1.08548284e+00 -3.33134085e-01
7.13959336e-01 -1.41653314e-01 -5.00520706e-01 -2.28861645e-01
-8.69354904e-01 4.48146909e-01 1.22890947e-02 -7.41325468e-02
-1.57635474e+00 -4.64529425e-01 -1.28465092e+00 -2.44282469e-01
4.13815230e-01 -3.51145595e-01 1.07622600e+00 -1.14544380e+00
-1.02979088e+00 1.10145390e+00 -4.29112345e-01 -5.66464365e-01
4.01830018e-01 -4.99868035e-01 -8.30700338e-01 3.01306039e-01
3.24260116e-01 1.32198334e-01 1.27945125e+00 -1.01267755e+00
-6.00906491e-01 -3.15892249e-01 -3.94249320e-01 2.24793747e-01
-3.22389841e-01 -2.52703913e-02 -5.24229586e-01 -1.10279763e+00
2.68719912e-01 -5.62376976e-01 -4.10100035e-02 1.37575135e-01
-3.26166958e-01 -3.04070026e-01 1.25951266e+00 -5.68701327e-01
1.43597293e+00 -2.43215132e+00 1.79632783e-01 4.06793773e-01
4.78898615e-01 3.77438039e-01 -3.39111090e-01 2.16675833e-01
-4.84805882e-01 -4.75273162e-01 -2.50692278e-01 -3.87412518e-01
-2.15746254e-01 4.40414399e-01 -4.38673794e-01 7.09758937e-01
2.94058323e-01 7.20250130e-01 -1.20530796e+00 -3.17845911e-01
2.66186357e-01 1.72693640e-01 -6.97118163e-01 7.04290688e-01
-1.08283423e-01 6.06275022e-01 -3.73750031e-01 1.25162792e+00
3.50854993e-01 -1.26835778e-01 -1.58960894e-01 -4.30321768e-02
1.96108103e-01 7.15542063e-02 -1.15187871e+00 1.90722382e+00
1.62829727e-01 6.45998597e-01 -1.72433138e-01 -1.53616154e+00
7.76558697e-01 2.95067877e-01 8.29154193e-01 -5.48177123e-01
8.16105753e-02 1.46853805e-01 -3.03753436e-01 -7.54344404e-01
4.60496098e-02 4.04765368e-01 1.29482895e-01 5.18319249e-01
4.63912368e-01 5.78524351e-01 9.35984626e-02 5.13845801e-01
1.58902705e+00 1.06005058e-01 4.41599697e-01 -1.09298050e-01
8.30143332e-01 -1.91404223e-01 8.96664560e-01 7.89981365e-01
-5.70799708e-01 5.79151154e-01 7.26827681e-01 -7.03602910e-01
-8.18254352e-01 -1.28700542e+00 9.19977725e-02 9.68938947e-01
1.43152013e-01 -7.55256593e-01 -4.83610183e-01 -1.14606857e+00
-4.55199257e-02 5.23945689e-01 -6.62948787e-01 -5.18674850e-01
-3.49479228e-01 -3.91082317e-01 4.88360375e-01 4.38066244e-01
2.15038285e-01 -9.66129243e-01 -2.21473724e-01 -1.13218777e-01
-2.78972149e-01 -1.02843535e+00 -5.18202364e-01 1.11933276e-01
-8.72927666e-01 -1.14006197e+00 -3.90149236e-01 -7.61407554e-01
8.90202522e-01 3.69695544e-01 1.11389256e+00 2.23213062e-01
-2.78622806e-01 9.58118796e-01 -8.20681453e-01 -7.92502463e-02
-4.70225096e-01 -5.36434650e-01 6.68937266e-01 5.25704563e-01
9.06755090e-01 -8.13277066e-01 -5.07479846e-01 3.12202543e-01
-8.32727969e-01 -4.32179809e-01 3.42434943e-01 7.93412626e-01
9.49347794e-01 -1.63365856e-01 4.85865653e-01 -6.50059998e-01
3.01576793e-01 -1.00978720e+00 -1.80972323e-01 -9.75362286e-02
-5.07383943e-01 -9.21762958e-02 5.97192764e-01 -4.78726715e-01
-5.69737613e-01 6.11931421e-02 -3.22758764e-01 -1.26025283e+00
-4.80565220e-01 2.04994172e-01 -3.29460949e-01 1.53832644e-01
7.02120721e-01 8.10595393e-01 1.65094465e-01 -4.06595111e-01
8.74960795e-02 5.47628284e-01 7.51849949e-01 -5.23875535e-01
1.06876385e+00 6.20004475e-01 -2.99803674e-01 -8.39909732e-01
-1.02251959e+00 -8.55270445e-01 -5.52493274e-01 -4.25561458e-01
7.32936561e-01 -1.29904532e+00 5.86715201e-03 7.88323700e-01
-5.99145055e-01 -3.75153124e-01 -8.16049576e-01 4.48389918e-01
-6.40924752e-01 8.12710166e-01 -4.54659611e-01 -5.98897099e-01
-1.71759188e-01 -9.35365260e-01 1.00290704e+00 -2.02164963e-01
-1.55906215e-01 -8.55568290e-01 3.44020873e-01 9.58463475e-02
1.02214396e-01 2.96654105e-01 5.35374045e-01 -1.14572835e+00
-3.62167239e-01 -3.85455430e-01 2.56335050e-01 8.66991341e-01
3.09322983e-01 -2.45540917e-01 -8.57562482e-01 -4.71818089e-01
2.61459887e-01 -4.34288710e-01 8.86992335e-01 1.81988090e-01
1.57983768e+00 -4.00933683e-01 -1.00322003e-02 8.94074559e-01
1.03709471e+00 7.00139254e-02 7.29002893e-01 2.15585500e-01
9.78618085e-01 -3.26415077e-02 8.46279562e-01 8.91939461e-01
3.01140338e-01 4.47523147e-01 5.53635120e-01 1.07776985e-01
7.78121874e-02 -3.88639092e-01 9.12770748e-01 9.26067650e-01
-1.04954757e-01 -6.24351725e-02 -8.98339152e-01 6.42332494e-01
-2.07030869e+00 -1.26552820e+00 1.14239819e-01 2.08076406e+00
4.00319606e-01 1.73300296e-01 4.61415201e-01 4.11262691e-01
6.36963546e-01 5.21724105e-01 -7.83565760e-01 -1.42405733e-01
-2.04047531e-01 -1.35574505e-01 -7.71510378e-02 9.08067524e-02
-1.37343228e+00 7.31708527e-01 5.91729450e+00 7.73458660e-01
-8.77313793e-01 1.76056802e-01 4.24276918e-01 -2.89249271e-01
-1.43406823e-01 -3.20051491e-01 -4.01120543e-01 6.49022818e-01
7.60298669e-01 -9.25930142e-02 3.78829300e-01 1.10932970e+00
2.22933710e-01 2.01167613e-01 -1.44062054e+00 1.41506624e+00
5.09201407e-01 -1.22654414e+00 3.99694562e-01 -1.45175591e-01
6.29586101e-01 2.88697958e-01 -5.20475544e-02 4.61908013e-01
-2.15731002e-02 -7.37272263e-01 5.73879659e-01 5.46585262e-01
8.25123131e-01 -4.21090394e-01 5.66185415e-01 3.06584328e-01
-1.37277627e+00 -2.23407790e-01 -1.87619925e-01 -1.02690518e-01
5.65617755e-02 5.26442885e-01 -2.18236268e-01 3.13439786e-01
1.00197482e+00 1.61705291e+00 -5.48265755e-01 8.08788717e-01
-1.53888583e-01 9.53329563e-01 -1.14600353e-01 5.99720597e-01
1.95730656e-01 -3.02028686e-01 1.05957603e+00 1.17431211e+00
2.54758090e-01 4.38405434e-03 3.37454915e-01 4.31696266e-01
8.95195007e-02 -3.42544988e-02 -9.31808889e-01 3.09151169e-02
3.46613050e-01 9.45668161e-01 -2.75259346e-01 -3.65764856e-01
-9.00868654e-01 1.33475435e+00 1.28512457e-01 4.16284263e-01
-8.94360483e-01 -6.90331757e-02 1.18478894e+00 1.64410859e-01
3.88285965e-01 -1.93111643e-01 1.74849436e-01 -1.93509805e+00
2.06895158e-01 -1.34646201e+00 8.64316583e-01 -4.69632447e-01
-1.54594910e+00 4.70998019e-01 -2.22768396e-01 -1.91562486e+00
-2.70894587e-01 -5.41683495e-01 -9.64937985e-01 8.91604051e-02
-1.38912582e+00 -8.32767725e-01 -5.67161083e-01 1.33953643e+00
8.32590580e-01 -8.18760514e-01 8.03374588e-01 4.88495648e-01
-9.17340815e-01 8.10645103e-01 1.65131807e-01 5.07872462e-01
5.98353028e-01 -1.17817664e+00 3.48254859e-01 1.35692298e+00
3.66953194e-01 2.15232819e-01 4.91724193e-01 -6.27923310e-01
-1.54645979e+00 -1.32189381e+00 3.34017038e-01 -5.40121019e-01
8.05956244e-01 -3.58071148e-01 -1.23473620e+00 1.05772984e+00
-2.85967141e-01 7.52224386e-01 9.42956328e-01 -2.06272736e-01
-6.69272661e-01 -3.81985158e-02 -1.10573566e+00 3.75411868e-01
1.49839401e+00 -7.41691470e-01 -8.88049901e-01 4.74969000e-01
5.12007415e-01 -4.14635926e-01 -6.23672128e-01 3.86355877e-01
1.13354236e-01 -1.07052624e+00 8.43063831e-01 -9.03657854e-01
3.71861011e-01 -5.27616739e-01 -4.32556927e-01 -1.09489429e+00
-2.26014733e-01 -5.78975201e-01 -1.16500354e+00 9.52545881e-01
-3.34449625e-03 -5.06567717e-01 7.32763231e-01 -4.31978889e-02
-4.56859678e-01 -8.55779529e-01 -1.00920320e+00 -9.95775104e-01
-5.17484665e-01 -8.25419605e-01 3.28235120e-01 1.15170836e+00
-3.51329297e-02 3.51666776e-03 -7.14709818e-01 4.05851871e-01
7.55339861e-01 -2.20326483e-01 7.68585980e-01 -1.01260448e+00
-1.84581846e-01 -1.78421021e-01 -1.22960734e+00 -1.25291097e+00
4.58003312e-01 -9.22363698e-01 -1.47347182e-01 -9.96850967e-01
1.52966052e-01 2.53279246e-02 -6.77180529e-01 5.96647024e-01
-2.59218179e-02 2.67765850e-01 -4.34125096e-01 4.94470060e-01
-1.20125818e+00 8.42062712e-01 6.66293502e-01 6.20047376e-03
-2.46932730e-01 -3.23724821e-02 -5.31923592e-01 1.01182377e+00
7.42945135e-01 -3.86713833e-01 -4.77082968e-01 -3.55837405e-01
-1.54948562e-01 -2.31754884e-01 4.05578643e-01 -1.13841712e+00
6.30291551e-02 -1.38410807e-01 3.24784070e-01 -6.26303136e-01
5.44363707e-02 -6.92273498e-01 -4.84062165e-01 3.64432335e-01
-2.12623656e-01 1.41459167e-01 -1.12371534e-01 1.07233548e+00
-4.74935830e-01 9.18592960e-02 7.25134015e-01 2.22169757e-02
-1.15139723e+00 7.97107220e-01 -6.45035326e-01 3.97775352e-01
1.28506494e+00 -2.83333659e-01 -7.79423444e-03 -5.49354613e-01
-8.10166121e-01 3.59507650e-01 4.59970653e-01 5.98944426e-01
1.03796530e+00 -1.69055152e+00 -6.66420221e-01 6.67321742e-01
5.81272721e-01 1.96963139e-02 2.87744790e-01 9.24453735e-01
-4.48568523e-01 -4.48936038e-02 -2.82434195e-01 -9.61575568e-01
-1.11381900e+00 6.79788232e-01 2.93733895e-01 8.34454000e-02
-1.01987052e+00 8.02891910e-01 1.44252941e-01 -1.37810633e-01
6.02830291e-01 -5.07146306e-03 -1.38298601e-01 -1.54786557e-01
7.92140722e-01 4.01339531e-01 -1.77770793e-01 -9.60811794e-01
-5.75547636e-01 2.81359047e-01 -1.21245757e-01 4.25767064e-01
1.23586690e+00 -4.93110344e-02 -3.54443081e-02 6.74844384e-01
1.13117409e+00 -1.51856601e-01 -1.35886025e+00 -6.14967227e-01
-3.13826390e-02 -7.20282257e-01 -2.11904701e-02 -1.54749498e-01
-1.17551100e+00 4.97067571e-01 7.14640796e-01 -3.55489552e-02
1.46584320e+00 1.82127774e-01 8.80989909e-01 6.26469374e-01
1.02559380e-01 -8.82609189e-01 6.08247519e-01 5.74198604e-01
7.51045465e-01 -1.40040648e+00 -1.28578097e-01 -1.04643777e-01
-8.41628134e-01 7.79220104e-01 9.68928814e-01 -4.38843369e-01
8.04914594e-01 4.34865654e-02 9.14019942e-02 -3.65405440e-01
-7.16476381e-01 -1.92875892e-01 4.19421732e-01 6.98292375e-01
9.36067924e-02 -2.85586625e-01 1.91886485e-01 6.58413172e-01
3.67957145e-01 -1.71862677e-01 5.16053915e-01 1.15026355e+00
-3.02681595e-01 -8.52334619e-01 -1.92729428e-01 8.76436830e-01
-4.09635931e-01 9.40749720e-02 -1.59498692e-01 2.97954440e-01
9.97705385e-02 6.68849349e-01 3.05153072e-01 -6.95175529e-01
3.00606817e-01 2.56808758e-01 7.68436864e-02 -7.48011291e-01
-2.96486513e-04 -4.97019431e-03 -2.20182389e-01 -1.17920804e+00
-4.07244474e-01 -9.41061914e-01 -1.04360223e+00 -1.72065765e-01
-2.83115748e-02 1.32697150e-01 -1.04233481e-01 9.21679914e-01
5.25311470e-01 1.89092979e-01 9.62525904e-01 -6.97862923e-01
-4.32228595e-01 -6.02495193e-01 -7.39472389e-01 8.04169118e-01
7.57393360e-01 -7.52235830e-01 -7.50535309e-01 1.11914486e-01] | [7.853052139282227, 1.6024186611175537] |
73078336-be6e-4196-a928-b207f0ef2d59 | arbitrary-oriented-object-detection-with | 2003.05597 | null | https://arxiv.org/abs/2003.05597v4 | https://arxiv.org/pdf/2003.05597v4.pdf | On the Arbitrary-Oriented Object Detection: Classification based Approaches Revisited | Arbitrary-oriented object detection has been a building block for rotation sensitive tasks. We first show that the boundary problem suffered in existing dominant regression-based rotation detectors, is caused by angular periodicity or corner ordering, according to the parameterization protocol. We also show that the root cause is that the ideal predictions can be out of the defined range. Accordingly, we transform the angular prediction task from a regression problem to a classification one. For the resulting circularly distributed angle classification problem, we first devise a Circular Smooth Label technique to handle the periodicity of angle and increase the error tolerance to adjacent angles. To reduce the excessive model parameters by Circular Smooth Label, we further design a Densely Coded Labels, which greatly reduces the length of the encoding. Finally, we further develop an object heading detection module, which can be useful when the exact heading orientation information is needed e.g. for ship and plane heading detection. We release our OHD-SJTU dataset and OHDet detector for heading detection. Extensive experimental results on three large-scale public datasets for aerial images i.e. DOTA, HRSC2016, OHD-SJTU, and face dataset FDDB, as well as scene text dataset ICDAR2015 and MLT, show the effectiveness of our approach. | ['Xue Yang', 'Junchi Yan'] | 2020-03-12 | null | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/666_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123530664.pdf | eccv-2020-8 | ['object-detection-in-aerial-images'] | ['computer-vision'] | [ 1.15020059e-01 -2.35146247e-02 -1.21238448e-01 -4.86549407e-01
-3.91577482e-01 -5.57665706e-01 2.93350875e-01 -2.99757987e-01
9.87182558e-03 3.59334797e-01 1.19851520e-02 -2.50742584e-01
-1.52912617e-01 -7.46459603e-01 -5.23910582e-01 -9.91213620e-01
-1.57652184e-01 2.71113724e-01 2.80523717e-01 -3.23983788e-01
3.45367908e-01 7.32753098e-01 -1.38363159e+00 -2.24154308e-01
6.13133371e-01 1.10549986e+00 7.51373544e-03 4.25754994e-01
3.66039604e-01 4.80856389e-01 -6.48864150e-01 2.23494731e-02
4.80037361e-01 -1.55766338e-01 -6.18083775e-01 2.78268933e-01
3.95222753e-01 -2.82772750e-01 -1.03732601e-01 8.42898965e-01
6.53140485e-01 -1.39417961e-01 9.17186320e-01 -1.27350140e+00
-2.21584201e-01 2.41064966e-01 -9.30995882e-01 -1.56443074e-01
3.23234200e-01 -1.56792775e-01 6.39961183e-01 -7.40428925e-01
7.27248728e-01 1.24231195e+00 6.91323042e-01 3.35086554e-01
-7.26686716e-01 -6.34970605e-01 3.46133828e-01 1.11948885e-01
-1.57911813e+00 -3.66234541e-01 8.95433307e-01 -5.73077440e-01
6.03896320e-01 3.49023759e-01 4.28297579e-01 9.07541037e-01
1.76349729e-01 5.37243843e-01 7.85446882e-01 -3.83018553e-01
1.72096431e-01 -1.63961485e-01 5.40534891e-02 8.11848342e-01
3.00969958e-01 -9.00849923e-02 -1.37024462e-01 1.11791499e-01
6.94365263e-01 -1.87016621e-01 -4.77293044e-01 -5.48093319e-01
-1.12073958e+00 6.76508009e-01 5.78248799e-01 -1.66027904e-01
-3.10437512e-02 -2.06306547e-01 3.60247612e-01 2.91478932e-01
4.05772030e-01 6.07180476e-01 -7.20451653e-01 3.81474048e-01
-5.25514007e-01 6.86591119e-02 5.30208588e-01 1.37869048e+00
5.88411570e-01 2.13736650e-02 6.25762939e-02 8.01258802e-01
4.38228935e-01 6.68797970e-01 2.89814621e-01 -5.20572066e-01
4.23963666e-01 6.25681281e-01 1.41559854e-01 -1.37557387e+00
-9.67193604e-01 -4.14130032e-01 -9.29206192e-01 -1.18650034e-01
2.03166634e-01 -1.07090510e-01 -9.27611172e-01 1.56814432e+00
5.65494537e-01 -2.03140050e-01 1.45094723e-01 1.16284394e+00
9.14484918e-01 5.75687289e-01 -2.33451217e-01 -3.97291183e-01
1.65836418e+00 -7.03889370e-01 -6.17336392e-01 -3.77345383e-01
1.09136820e+00 -8.93275142e-01 6.86899424e-01 4.77292329e-01
-2.88569838e-01 -5.81144333e-01 -1.34614813e+00 4.90005650e-02
-2.45038673e-01 7.70825744e-01 6.87496126e-01 4.75269675e-01
-4.53917563e-01 1.88618541e-01 -5.23619294e-01 -4.31279629e-01
2.40259826e-01 3.27877134e-01 -4.09092456e-01 -2.42539048e-02
-1.09486496e+00 7.57736027e-01 2.33795375e-01 3.88214797e-01
-3.70097131e-01 -1.59890935e-01 -9.67247069e-01 -4.15706843e-01
4.68306184e-01 -2.73339629e-01 1.04606986e+00 -4.54261243e-01
-1.28940499e+00 8.82548392e-01 5.68319522e-02 -3.13504696e-01
5.29796600e-01 -1.75932482e-01 -5.68725586e-01 4.72973511e-02
1.02624170e-01 7.18670964e-01 1.05187964e+00 -1.24462366e+00
-5.27366221e-01 -4.58646625e-01 -2.57609487e-02 3.74016970e-01
-1.27875254e-01 -6.06434494e-02 -6.87394023e-01 -7.14446723e-01
7.11780727e-01 -1.33018816e+00 -1.23344213e-01 -7.35497996e-02
-6.80720925e-01 -2.09351510e-01 1.16613460e+00 -5.96145928e-01
1.15683460e+00 -2.44874239e+00 7.96828493e-02 1.56273484e-01
-7.72991776e-02 -3.93743105e-02 1.79953992e-01 6.33596554e-02
-3.37446928e-01 -1.14549287e-01 -1.34961322e-01 -9.28036124e-02
-2.37655222e-01 1.49760172e-01 -5.47505081e-01 9.59411561e-01
2.38617718e-01 2.30534047e-01 -5.46454966e-01 -5.47001421e-01
1.34043798e-01 2.38471016e-01 -4.95161355e-01 9.16991010e-02
7.90200084e-02 3.11766803e-01 -6.06960297e-01 9.55486238e-01
1.05573642e+00 -5.86794280e-02 9.20503289e-02 -7.29345381e-01
-3.70496869e-01 2.23842621e-01 -1.48953187e+00 1.28149354e+00
-2.37277418e-01 5.41579843e-01 -2.29596838e-01 -9.38847542e-01
1.36279273e+00 9.06929672e-02 4.54799592e-01 -4.31220233e-01
2.75331289e-01 1.61806382e-02 -1.77406937e-01 -4.23380494e-01
6.60367966e-01 3.22658241e-01 -3.79185557e-01 -2.66915411e-01
-2.86352932e-01 -3.11426789e-01 9.76068005e-02 -2.18111828e-01
7.41243362e-01 2.76003271e-01 5.08964717e-01 -3.44701022e-01
5.06944776e-01 2.11164132e-01 8.62066090e-01 2.35696524e-01
-8.85628685e-02 9.03182149e-01 7.47162223e-01 -6.13511801e-01
-6.22222781e-01 -4.34988141e-01 -6.90373123e-01 9.45135057e-01
5.48627675e-01 -4.28362757e-01 -5.45309126e-01 -6.55452728e-01
-1.03057697e-01 3.68944168e-01 -5.77826738e-01 -1.91605896e-01
-8.29370618e-01 -1.19822609e+00 3.69376600e-01 5.83108068e-01
7.42979050e-01 -7.19532669e-01 -8.01943064e-01 -7.99129084e-02
-2.04481646e-01 -1.37600255e+00 -2.83430278e-01 4.80780363e-01
-6.31071448e-01 -1.11172771e+00 -5.10385215e-01 -9.80148554e-01
8.76927972e-01 4.99219626e-01 7.40241349e-01 -8.56824815e-02
-4.36909884e-01 -1.44092381e-01 -5.33872247e-01 -3.36850882e-01
6.24041893e-02 2.08976775e-01 3.12002391e-01 -2.28845760e-01
8.50597024e-02 -9.93903205e-02 -5.77640295e-01 9.70229626e-01
-3.88002932e-01 3.06166917e-01 6.27910078e-01 6.67575836e-01
4.86412585e-01 -6.40182719e-02 4.69016224e-01 -6.31197691e-01
-2.76905708e-02 -2.72262812e-01 -9.34456468e-01 7.11527839e-02
-5.88211894e-01 1.33081928e-01 4.36643898e-01 -2.80327350e-01
-7.26708353e-01 4.52104598e-01 -3.91650982e-02 -2.16906846e-01
-2.39633694e-01 2.56508708e-01 -3.90323132e-01 -7.56689087e-02
6.47534370e-01 -1.79148335e-02 -4.11275148e-01 -4.11552668e-01
8.11219439e-02 9.60712314e-01 5.44161618e-01 -1.90040439e-01
9.58705604e-01 3.95080477e-01 3.57296795e-01 -1.06871223e+00
-6.26748621e-01 -5.20008922e-01 -6.40965283e-01 -1.63073733e-01
9.49615836e-01 -1.15883875e+00 -7.84829438e-01 5.21490276e-01
-1.17502761e+00 -1.10918535e-02 7.21785128e-02 4.43919152e-01
-4.09275860e-01 3.89932096e-01 -2.97575712e-01 -6.27048314e-01
-1.98434681e-01 -1.33639193e+00 1.33773029e+00 2.46434048e-01
7.01550469e-02 -3.85224462e-01 -1.16801493e-01 -1.25293368e-02
-3.71878892e-02 4.06306833e-01 6.82357550e-01 -4.11197513e-01
-5.07445097e-01 -3.08100253e-01 -3.48830491e-01 4.50779796e-02
3.80489230e-02 1.69205800e-01 -8.81857574e-01 -3.15276116e-01
-1.77965611e-01 -2.86322355e-01 8.69784951e-01 2.70059168e-01
1.21380579e+00 -7.03966618e-02 -5.78601837e-01 1.07175934e+00
1.18011165e+00 3.12676072e-01 5.39432645e-01 4.82063800e-01
7.89763927e-01 5.04800141e-01 1.33557558e+00 6.11494660e-01
3.76008153e-01 1.13149929e+00 4.98980612e-01 -2.65334696e-01
1.43001517e-02 -8.05222467e-02 2.74675727e-01 7.64048278e-01
-2.01139674e-01 -4.10846561e-01 -8.96718383e-01 2.06984758e-01
-1.64486885e+00 -5.57192206e-01 -3.62891614e-01 2.14974594e+00
5.84027469e-01 1.95425227e-01 -1.51017532e-01 3.98884207e-01
5.43484211e-01 2.61036932e-01 -2.83890307e-01 -1.85233340e-01
-1.46725386e-01 -5.34252644e-01 1.07681727e+00 1.41925439e-01
-1.73797119e+00 1.00468314e+00 5.73034525e+00 7.27836907e-01
-1.23260498e+00 -4.83601004e-01 3.80428463e-01 5.79697609e-01
4.15032864e-01 4.66894284e-02 -1.28943157e+00 2.75986254e-01
3.00074488e-01 4.67883885e-01 4.01415415e-02 1.25374305e+00
1.26406237e-01 -1.41037300e-01 -8.73413920e-01 1.00532854e+00
1.33774415e-01 -6.36242390e-01 -2.35422522e-01 8.43985677e-02
3.80664855e-01 -2.59823948e-01 -9.17845778e-03 2.99747497e-01
-8.84830430e-02 -8.62646759e-01 6.99894071e-01 1.09286614e-01
1.10153556e+00 -6.89473331e-01 8.19454908e-01 2.66099751e-01
-1.45281625e+00 -3.93298976e-02 -7.00583041e-01 2.13138118e-01
-1.07651189e-01 4.64460194e-01 -1.01128078e+00 7.92196095e-01
7.05351830e-01 8.11205029e-01 -8.44871938e-01 9.45072293e-01
-3.29137355e-01 3.42699558e-01 -5.30258536e-01 1.10225752e-01
-8.59980360e-02 -1.47086129e-01 5.62730730e-01 1.14204395e+00
3.03264350e-01 -5.63162239e-03 2.12847605e-01 4.31956723e-02
1.31519735e-01 2.39410326e-02 -6.97968841e-01 5.01138508e-01
5.11711180e-01 1.34602940e+00 -9.29579020e-01 -4.20994544e-03
-2.70647854e-01 7.74225533e-01 -1.03369094e-01 1.56649768e-01
-1.20481670e+00 -5.67058802e-01 4.18285787e-01 9.95948073e-03
5.49022079e-01 -3.27552944e-01 -1.53125122e-01 -1.03687954e+00
2.95089167e-02 -8.27526331e-01 2.85541922e-01 -5.96427083e-01
-6.45022511e-01 7.03238487e-01 2.34827653e-01 -1.85197234e+00
-1.88979596e-01 -8.30615044e-01 -3.16203356e-01 3.74453753e-01
-1.46693957e+00 -9.79614556e-01 -7.12543726e-01 3.13477993e-01
5.75498998e-01 -1.50989547e-01 7.18808651e-01 4.34992850e-01
-8.97747636e-01 5.37653387e-01 -9.09020156e-02 4.01267409e-01
1.09729481e+00 -9.75632191e-01 2.22857296e-01 6.56554222e-01
-1.87444419e-01 3.82247359e-01 9.35100555e-01 -4.58238602e-01
-1.59283161e+00 -1.21174085e+00 5.46312034e-01 -1.60139635e-01
1.23848021e-01 -5.30840576e-01 -6.45067930e-01 5.19223690e-01
-3.02991301e-01 2.23075360e-01 2.38567919e-01 -6.60475791e-02
-2.34939232e-01 -2.64885962e-01 -9.62116361e-01 4.40399110e-01
1.25834000e+00 -5.76843880e-03 -4.49104071e-01 6.78202331e-01
6.34012520e-01 -6.45029306e-01 -6.54577613e-01 9.74289298e-01
4.92672771e-01 -7.55425990e-01 9.73836303e-01 -1.43396989e-01
2.07557961e-01 -7.27729499e-01 -2.17367038e-01 -1.14149940e+00
-2.31630176e-01 -5.10541022e-01 1.07722752e-01 1.15056479e+00
2.27036938e-01 -4.43449974e-01 4.86158609e-01 -6.51763380e-02
-3.04943800e-01 -6.74864292e-01 -9.92083848e-01 -7.62948394e-01
-5.07394075e-01 -2.63017058e-01 5.43232858e-01 9.51431036e-01
-2.64845490e-01 3.64699095e-01 -5.23910999e-01 4.99618173e-01
4.11448985e-01 3.02269578e-01 1.05844879e+00 -1.16135633e+00
-1.09744415e-01 -1.74733028e-02 -6.96572483e-01 -1.59840024e+00
-2.40948230e-01 -4.35741305e-01 3.10874820e-01 -1.12002838e+00
-3.14215720e-01 -6.21644914e-01 2.65705973e-01 5.18361807e-01
1.66270927e-01 3.07725400e-01 -1.01651505e-01 3.26879352e-01
-4.47148353e-01 5.07562160e-01 1.22012448e+00 1.19791195e-01
-2.51968205e-01 6.27987757e-02 -4.83763725e-01 1.08371639e+00
7.37322807e-01 -5.54227591e-01 -2.10857332e-01 -3.34658265e-01
2.92221516e-01 7.98213258e-02 2.59165376e-01 -1.10110819e+00
1.01701722e-01 -1.42438784e-01 5.18376052e-01 -1.03772748e+00
2.67772943e-01 -8.47347140e-01 -2.34838516e-01 4.85844702e-01
3.67334597e-02 3.01280636e-02 4.76202480e-02 3.68589699e-01
-6.35171309e-02 -6.04465939e-02 9.88733828e-01 4.39675838e-01
-7.17848301e-01 1.69388726e-01 -2.00780556e-01 -1.73213184e-01
1.20715082e+00 -1.87559456e-01 -5.52172542e-01 -1.48831336e-02
-3.81432176e-01 3.75885934e-01 3.61807287e-01 5.10606885e-01
5.46177506e-01 -1.14043570e+00 -4.86442953e-01 5.00431359e-01
3.77664983e-01 3.55751544e-01 -1.17113655e-02 1.03916454e+00
-7.72852719e-01 5.35210729e-01 1.37029001e-02 -8.43021154e-01
-1.28804266e+00 5.66511512e-01 1.62891984e-01 1.06084928e-01
-6.85042620e-01 6.64666355e-01 3.55613261e-01 -3.28783959e-01
2.31416956e-01 -7.25008249e-01 -6.01085544e-01 3.06754470e-01
1.21997237e-01 1.81953758e-01 1.48219675e-01 -8.35465789e-01
-7.27646112e-01 1.17487240e+00 6.17294125e-02 3.82496089e-01
1.17343640e+00 -1.47147596e-01 2.26251557e-01 3.06029655e-02
1.07989621e+00 8.69836286e-02 -1.10517633e+00 1.72345132e-01
-1.78490132e-01 -3.32853228e-01 4.12889244e-03 -3.65920961e-01
-9.20184493e-01 5.93776941e-01 7.71838784e-01 2.53741294e-01
1.29254675e+00 -4.47197026e-03 4.45245236e-01 6.70703828e-01
3.44558954e-01 -1.06119692e+00 -3.04889679e-01 5.58126986e-01
1.02912569e+00 -1.34969831e+00 4.72737074e-01 -9.14442420e-01
-6.97025359e-01 1.27508247e+00 7.10680604e-01 -1.52607024e-01
4.72408623e-01 3.81483138e-01 9.07978881e-03 -1.08880185e-01
-4.09828782e-01 -1.33698985e-01 2.81267673e-01 4.76066947e-01
2.62872815e-01 1.02053493e-01 -4.16440219e-01 5.31145930e-01
-4.62714791e-01 -3.92993242e-01 3.99856150e-01 8.98829877e-01
-5.69038033e-01 -7.01330245e-01 -6.63659215e-01 5.99242374e-02
-8.67109820e-02 3.14237505e-01 -3.01211625e-01 1.09718192e+00
2.21003056e-01 6.86149061e-01 1.00162670e-01 -5.73659420e-01
4.73085731e-01 -2.40368426e-01 2.49239340e-01 -3.18758875e-01
6.45337701e-02 2.17845812e-01 2.01771289e-01 -3.40456694e-01
-2.86785930e-01 -5.80068767e-01 -1.26464438e+00 8.23761970e-02
-6.32562399e-01 -1.82316124e-01 7.02982485e-01 7.33003139e-01
1.89914286e-01 4.23289746e-01 1.01581609e+00 -6.21034324e-01
-5.98041952e-01 -1.03126669e+00 -3.94967318e-01 5.53925596e-02
3.39601099e-01 -1.12519443e+00 -5.49779356e-01 1.01014888e-02] | [8.686772346496582, -0.8193923234939575] |
73c3664d-c5f2-45fe-8882-4e79e95f93f4 | analyzing-large-receptive-field-convolutional | 1910.07047 | null | https://arxiv.org/abs/1910.07047v1 | https://arxiv.org/pdf/1910.07047v1.pdf | Analyzing Large Receptive Field Convolutional Networks for Distant Speech Recognition | Despite significant efforts over the last few years to build a robust automatic speech recognition (ASR) system for different acoustic settings, the performance of the current state-of-the-art technologies significantly degrades in noisy reverberant environments. Convolutional Neural Networks (CNNs) have been successfully used to achieve substantial improvements in many speech processing applications including distant speech recognition (DSR). However, standard CNN architectures were not efficient in capturing long-term speech dynamics, which are essential in the design of a robust DSR system. In the present study, we address this issue by investigating variants of large receptive field CNNs (LRF-CNNs) which include deeply recursive networks, dilated convolutional neural networks, and stacked hourglass networks. To compare the efficacy of the aforementioned architectures with the standard CNN for Wall Street Journal (WSJ) corpus, we use a hybrid DNN-HMM based speech recognition system. We extend the study to evaluate the system performances for distant speech simulated using realistic room impulse responses (RIRs). Our experiments show that with fixed number of parameters across all architectures, the large receptive field networks show consistent improvements over the standard CNNs for distant speech. Amongst the explored LRF-CNNs, stacked hourglass network has shown improvements with a 8.9% relative reduction in word error rate (WER) and 10.7% relative improvement in frame accuracy compared to the standard CNNs for distant simulated speech signals. | ['Vinay Kothapally', 'Soheil Khorram', 'John H. L. Hansen', 'Salar Jafarlou'] | 2019-10-15 | null | null | null | null | ['distant-speech-recognition'] | ['speech'] | [ 6.73955753e-02 -1.42888382e-01 6.11878276e-01 -4.17368144e-01
-7.60141194e-01 -2.67081559e-01 6.46006107e-01 -4.18613106e-01
-4.58195448e-01 4.67098683e-01 4.20582831e-01 -7.49521315e-01
-3.58968861e-02 -3.32039446e-01 -5.93044102e-01 -7.37024784e-01
2.85289790e-02 -1.78272247e-01 2.65661031e-01 -6.65229440e-01
3.35340276e-02 7.65819907e-01 -1.56987035e+00 1.56878561e-01
2.38233790e-01 1.01199365e+00 5.57680428e-01 1.23018467e+00
2.06367955e-01 8.27753544e-01 -1.12864912e+00 9.19124708e-02
2.13273037e-02 5.71468845e-03 -6.70107424e-01 -3.19255471e-01
2.77317733e-01 -1.62886441e-01 -9.50276136e-01 6.18826270e-01
1.28676045e+00 8.27301025e-01 8.94023031e-02 -4.29878384e-01
-7.39680111e-01 6.11531377e-01 2.03872547e-01 7.57317245e-01
1.09269805e-01 5.02428599e-02 5.00194371e-01 -9.49219465e-01
-1.44984704e-02 1.37395883e+00 6.51069045e-01 7.89790392e-01
-7.02879310e-01 -6.07062280e-01 -9.31310642e-04 2.08483338e-01
-1.36896920e+00 -1.14229727e+00 5.03785968e-01 1.47163585e-01
1.98942447e+00 4.57412034e-01 1.28955394e-01 1.57651043e+00
-7.94226304e-02 5.35870969e-01 8.55476320e-01 -5.09625852e-01
1.90411091e-01 -2.50479907e-01 2.91878104e-01 2.26450354e-01
-4.46448237e-01 4.37726438e-01 -4.39132094e-01 2.83266097e-01
7.57349491e-01 -2.70050377e-01 -4.22089249e-01 7.85317004e-01
-1.16540694e+00 4.53616321e-01 4.01598752e-01 7.27941692e-01
-1.38821647e-01 2.92771429e-01 5.57569742e-01 4.04116869e-01
6.10447526e-01 1.39711291e-01 -5.86116195e-01 -4.64574099e-01
-8.27600121e-01 -3.46715339e-02 7.15227842e-01 6.41675830e-01
1.72314011e-02 9.82809544e-01 -1.62317708e-01 1.65194297e+00
3.39851886e-01 4.79033321e-01 9.14738536e-01 -5.35232723e-01
6.57514691e-01 -2.88052619e-01 -2.53681153e-01 -6.91929758e-01
-3.88337553e-01 -9.79842901e-01 -1.12452877e+00 -2.03688502e-01
-7.39305690e-02 -4.46140207e-02 -1.47620690e+00 1.46618259e+00
-1.24982692e-01 3.28926980e-01 5.96611679e-01 7.80010223e-01
1.17730606e+00 1.22369981e+00 -1.07089356e-01 9.02891532e-02
1.03984344e+00 -1.20086741e+00 -9.06348944e-01 -3.28689903e-01
4.40058887e-01 -1.03087163e+00 9.22312975e-01 2.64045089e-01
-9.88020778e-01 -1.04639888e+00 -1.20844603e+00 -5.16648181e-02
-4.68252450e-01 2.99950838e-01 4.63867839e-03 8.59522402e-01
-1.55810821e+00 3.86493444e-01 -9.02606845e-01 -4.76176977e-01
-2.96200663e-02 5.02872169e-01 -2.48462439e-01 2.20887125e-01
-1.41688347e+00 7.47760415e-01 8.44825357e-02 6.74559116e-01
-1.28886616e+00 -4.87390548e-01 -6.90885842e-01 2.72519857e-01
6.43741414e-02 -2.15392172e-01 1.64637280e+00 -5.92031419e-01
-2.21679902e+00 3.17079157e-01 -2.61180699e-01 -7.88346589e-01
1.08708866e-01 -3.26714367e-01 -1.00422347e+00 -3.16412568e-01
-6.73604250e-01 4.12962228e-01 6.65721834e-01 -9.57307577e-01
-3.32728684e-01 -3.60666849e-02 -1.94299466e-03 1.37090251e-01
-1.67702615e-01 3.00673276e-01 -1.59557238e-01 -9.12162602e-01
1.10276535e-01 -1.04423165e+00 -2.17233703e-01 -8.75499666e-01
-4.24418092e-01 -3.85726035e-01 1.03252721e+00 -8.57075036e-01
1.18977177e+00 -2.37825227e+00 -2.46283114e-01 -2.03397155e-01
-2.08275676e-01 1.09770799e+00 -1.41231820e-01 3.13984185e-01
-2.63443410e-01 5.59722967e-02 8.79649451e-05 -6.52566969e-01
-1.01798698e-01 3.54231447e-01 -5.25806487e-01 3.20187718e-01
-7.31151700e-02 6.58865988e-01 -3.52354020e-01 2.98256159e-01
5.87814927e-01 1.09077549e+00 -3.98901664e-02 3.63798767e-01
2.58904934e-01 3.35609376e-01 4.82755452e-02 3.79511088e-01
4.38104719e-01 2.33390883e-01 -2.84162909e-01 8.02926272e-02
-1.97551444e-01 7.02878356e-01 -8.69779944e-01 1.35001111e+00
-9.99912679e-01 1.21517062e+00 5.75527176e-02 -1.19332421e+00
1.39134312e+00 8.47028732e-01 -3.17966461e-01 -9.92968619e-01
4.47499938e-03 5.52445054e-01 2.68869549e-01 -3.62938493e-01
4.78668839e-01 -1.00730829e-01 4.19518113e-01 -2.06441283e-01
1.50698006e-01 1.67011574e-01 -4.93638128e-01 -2.35805959e-01
1.11699033e+00 -4.05462027e-01 -5.97446710e-02 -1.64309829e-01
8.13580751e-01 -7.03296542e-01 3.18487436e-01 8.70342433e-01
-3.19980949e-01 1.05506694e+00 -2.41406098e-01 -6.08072102e-01
-1.12014818e+00 -1.04117513e+00 -2.28364274e-01 9.99530435e-01
-4.56467301e-01 -5.70305763e-03 -7.84867465e-01 4.18033311e-03
-6.41019821e-01 6.44084454e-01 -1.25321731e-01 3.93564627e-02
-1.02041519e+00 -6.81097329e-01 1.22785854e+00 7.72475958e-01
9.01180923e-01 -1.35892212e+00 -9.86883491e-02 4.23146665e-01
-1.27593517e-01 -1.56920445e+00 -2.23494589e-01 4.32106227e-01
-5.44114053e-01 -2.09144324e-01 -1.06458652e+00 -1.03109050e+00
1.17442124e-01 3.77418190e-01 7.63603389e-01 -1.01518616e-01
-7.26115033e-02 1.59234837e-01 -5.08476615e-01 -2.90131062e-01
-7.45944321e-01 2.27561399e-01 3.66342634e-01 -9.08940658e-02
-1.71109638e-03 -4.84765112e-01 -7.43085146e-01 3.57690483e-01
-7.78115451e-01 -4.60787714e-01 4.47705388e-01 7.54648566e-01
2.30396241e-01 -4.87743691e-02 8.20454478e-01 -2.42130384e-01
7.23847210e-01 -1.77335963e-01 -3.95436794e-01 9.20669138e-02
-1.24754794e-01 -1.49859518e-01 9.41708803e-01 -4.13038224e-01
-1.29305184e+00 -4.30545717e-01 -1.07364452e+00 -3.21065187e-01
-5.61938226e-01 2.70840406e-01 2.03325048e-01 7.98326284e-02
6.07491314e-01 5.00021458e-01 -3.59201550e-01 -5.83154321e-01
-1.06943987e-01 1.34776270e+00 5.10411203e-01 -1.91187069e-01
3.94476026e-01 1.19972624e-01 -3.50934088e-01 -1.62784195e+00
-5.51025152e-01 -6.19002461e-01 -1.93643317e-01 2.00251583e-02
8.84506941e-01 -1.07792401e+00 -5.88321447e-01 8.75923157e-01
-1.25327027e+00 -5.86602867e-01 1.46823019e-01 7.48695016e-01
-3.90330791e-01 1.68251649e-01 -8.04657161e-01 -1.10944533e+00
-5.31089067e-01 -1.54741621e+00 1.02834141e+00 7.50467107e-02
8.44587106e-04 -1.02612340e+00 7.88404942e-02 5.03620863e-01
1.09697354e+00 -3.91444832e-01 6.10577583e-01 -9.52365398e-01
-3.42731386e-01 -4.52522300e-02 1.53480083e-01 1.06145942e+00
-5.39397309e-03 -2.26683244e-02 -1.64291739e+00 -3.39856416e-01
9.00535211e-02 -9.42140818e-03 8.93740714e-01 7.74841964e-01
1.29285717e+00 -1.20396726e-01 7.74302930e-02 6.42949760e-01
1.06611431e+00 8.37038815e-01 1.09144616e+00 3.54696959e-01
4.77559358e-01 3.54523450e-01 -1.19313523e-01 3.20062637e-02
-4.39019203e-02 6.29579008e-01 8.86458084e-02 -1.93197608e-01
-6.09478712e-01 8.48596096e-02 6.43870294e-01 1.61256313e+00
-1.04585372e-01 -7.71752536e-01 -9.71944451e-01 6.50347054e-01
-1.41660035e+00 -8.78693581e-01 1.09278426e-01 2.05401111e+00
2.65617698e-01 6.05379492e-02 -2.86740571e-01 4.08822387e-01
7.81439424e-01 5.55375516e-01 -1.81132317e-01 -9.44521546e-01
-3.15063328e-01 5.84257543e-01 3.32408965e-01 5.99504888e-01
-8.76764834e-01 9.67019320e-01 6.37618589e+00 1.04295301e+00
-1.43089521e+00 2.91380793e-01 8.45312297e-01 1.69190578e-02
1.89380497e-01 -4.05408859e-01 -9.40318346e-01 7.92609751e-02
1.83228123e+00 5.54420650e-01 5.70062876e-01 7.21872568e-01
5.40514112e-01 4.15572643e-01 -6.34542525e-01 1.05421019e+00
-3.99661772e-02 -1.18482327e+00 -2.37751216e-01 -2.15000082e-02
6.66844189e-01 6.99803948e-01 3.67265016e-01 5.31888366e-01
1.72192872e-01 -1.45613301e+00 5.74940324e-01 2.71086633e-01
7.05774128e-01 -8.98162544e-01 1.02112579e+00 1.74039960e-01
-1.22579718e+00 -1.96906477e-01 -5.80214083e-01 -9.69384760e-02
9.51785445e-02 4.98116106e-01 -1.05714726e+00 2.29797155e-01
1.02300096e+00 1.91762000e-01 -1.70340300e-01 7.61457920e-01
2.01472089e-01 1.06098628e+00 -3.76994073e-01 -2.25861385e-01
6.02067530e-01 4.09647495e-01 5.29110730e-01 1.47589958e+00
5.68713069e-01 1.63096357e-02 -4.09952164e-01 3.66706401e-01
-1.34879202e-01 -1.24975570e-01 -5.15286028e-01 -1.23533502e-01
4.68066365e-01 9.34926331e-01 -4.88046795e-01 -2.48376489e-01
-3.31255168e-01 6.84110940e-01 1.12912223e-01 7.57327020e-01
-7.05841184e-01 -5.20534039e-01 9.01903272e-01 -1.98743269e-01
5.25492787e-01 -6.49445534e-01 -1.46060647e-03 -8.33091080e-01
2.52196006e-02 -9.58176911e-01 -2.98010826e-01 -8.99034381e-01
-9.11581337e-01 1.30623448e+00 -2.40049660e-01 -8.55935037e-01
-1.25838980e-01 -7.27865577e-01 -6.01029575e-01 1.04812312e+00
-1.66507196e+00 -9.00508821e-01 -1.04024991e-01 5.21466196e-01
1.25430596e+00 -5.56986868e-01 8.46014202e-01 3.86383593e-01
-7.28092074e-01 6.88960850e-01 3.81513178e-01 1.64031878e-01
2.91880310e-01 -9.95105624e-01 1.15270007e+00 9.67035651e-01
9.23520625e-02 8.75477433e-01 4.99227762e-01 -2.15078011e-01
-1.14365828e+00 -1.09246159e+00 7.82312512e-01 2.23372206e-02
4.70799714e-01 -5.22745609e-01 -9.42814291e-01 6.10691905e-01
3.92893136e-01 2.71638274e-01 4.44058537e-01 -6.21091202e-02
-3.29344392e-01 -3.50277722e-01 -6.87319934e-01 5.45850158e-01
1.02962041e+00 -8.71203661e-01 -4.13934857e-01 1.02133237e-01
1.17093635e+00 -4.93253827e-01 -6.78602338e-01 6.34687006e-01
4.15556282e-01 -1.01290536e+00 1.21514010e+00 -1.95995882e-01
-3.93462330e-02 -1.62809253e-01 -7.39680350e-01 -1.47621942e+00
-6.51705861e-02 -8.64113688e-01 2.43551712e-02 1.22805345e+00
4.92193878e-01 -9.07534480e-01 2.31873289e-01 7.94515535e-02
-9.36603189e-01 -5.85267603e-01 -1.32120955e+00 -9.98612821e-01
1.13344759e-01 -8.95158947e-01 5.20049214e-01 3.75361204e-01
-6.43782556e-01 2.51786858e-01 -4.11570430e-01 5.02057850e-01
-4.33431827e-02 -7.91674554e-01 4.50347364e-01 -5.86787343e-01
-2.42399916e-01 -3.05307150e-01 -5.14672518e-01 -1.33299184e+00
8.54177549e-02 -5.34040570e-01 3.74772698e-01 -1.53012884e+00
-4.77480859e-01 -3.15191507e-01 -4.21921998e-01 1.74744502e-01
1.89649388e-01 2.58047823e-02 2.99674273e-02 -6.22129887e-02
-1.85663357e-01 6.02116764e-01 9.63188231e-01 -1.10184075e-03
-1.58305705e-01 2.23088548e-01 -1.44623190e-01 4.54318404e-01
9.69870508e-01 -9.75854024e-02 -5.63566864e-01 -7.10220098e-01
-1.78316996e-01 2.46332049e-01 3.98087442e-01 -1.44623113e+00
3.64215285e-01 5.29680729e-01 1.46502271e-01 -5.21495342e-01
6.16367519e-01 -3.65861297e-01 1.10250721e-02 1.60700321e-01
-5.74384093e-01 6.94894940e-02 3.88157517e-01 4.49911594e-01
-5.62131286e-01 -2.87717953e-02 8.91425848e-01 -1.21548288e-01
-5.86988330e-01 1.64032895e-02 -6.45096898e-01 -3.07740539e-01
3.96622092e-01 -2.21480265e-01 -4.52805191e-01 -4.34762478e-01
-8.40093195e-01 -5.11934459e-01 -4.57794338e-01 7.47087836e-01
8.63241315e-01 -9.77542937e-01 -5.92268944e-01 4.14989918e-01
-3.48403811e-01 1.02463216e-01 5.97096920e-01 5.12258530e-01
-6.84076786e-01 1.08252072e+00 3.00309271e-01 -5.73219359e-01
-1.28167379e+00 1.81295760e-02 8.56693923e-01 5.35837449e-02
-6.38289988e-01 1.15781140e+00 2.97188699e-01 -7.57006288e-01
7.13540673e-01 -7.26205170e-01 -2.32925490e-01 -5.85192025e-01
5.33450723e-01 5.32885313e-01 7.93293178e-01 -7.56024122e-01
-3.36592048e-01 3.54464978e-01 -1.20904274e-01 -2.17092007e-01
1.48565948e+00 -2.41176218e-01 3.37276340e-01 5.24364412e-01
1.58368027e+00 -2.24348292e-01 -9.84395087e-01 -7.14370981e-02
-3.01432788e-01 -3.28467339e-02 5.73046267e-01 -8.56352091e-01
-1.03703952e+00 1.22103393e+00 9.57168877e-01 4.16027874e-01
1.00824249e+00 -1.07450105e-01 1.22683668e+00 6.20751739e-01
2.94806585e-02 -9.76159692e-01 6.95296153e-02 1.03667450e+00
9.88935351e-01 -1.13596070e+00 -7.96877325e-01 7.27318078e-02
-4.12291229e-01 1.39272439e+00 4.27119970e-01 -4.77037802e-02
8.37359011e-01 4.85160142e-01 3.79741162e-01 -2.29459794e-04
-7.71451533e-01 -1.65765479e-01 1.89590186e-01 5.91527045e-01
6.19770527e-01 9.74165052e-02 3.82091552e-01 2.04916060e-01
-5.78506112e-01 -4.61824983e-01 3.14213365e-01 6.99292302e-01
-5.48441589e-01 -6.19410574e-01 -4.47462082e-01 -3.03106941e-02
-8.13558578e-01 -3.34936619e-01 2.75555402e-02 5.15773356e-01
-3.44666839e-01 1.41322815e+00 1.85774416e-01 -5.46247840e-01
5.11655927e-01 5.69228409e-03 4.37152833e-02 -3.81233096e-01
-8.49960148e-01 2.74011314e-01 3.43183547e-01 -2.32194617e-01
-4.04049754e-01 -2.43464187e-01 -1.13740993e+00 -2.58418769e-01
-4.45524335e-01 -7.79084489e-02 1.22992623e+00 1.16698861e+00
3.10352266e-01 1.11704528e+00 5.64210117e-01 -6.97159469e-01
-5.57112753e-01 -1.34566903e+00 -3.36398959e-01 -1.05165981e-01
8.66789877e-01 -2.66049445e-01 -5.99058270e-01 -1.39446929e-01] | [14.722054481506348, 6.0901641845703125] |
1b11497f-ae02-4d09-8428-5914b949c160 | learning-video-salient-object-detection | 2204.02008 | null | https://arxiv.org/abs/2204.02008v1 | https://arxiv.org/pdf/2204.02008v1.pdf | Learning Video Salient Object Detection Progressively from Unlabeled Videos | Recent deep learning-based video salient object detection (VSOD) has achieved some breakthrough, but these methods rely on expensive annotated videos with pixel-wise annotations, weak annotations, or part of the pixel-wise annotations. In this paper, based on the similarities and the differences between VSOD and image salient object detection (SOD), we propose a novel VSOD method via a progressive framework that locates and segments salient objects in sequence without utilizing any video annotation. To use the knowledge learned in the SOD dataset for VSOD efficiently, we introduce dynamic saliency to compensate for the lack of motion information of SOD during the locating process but retain the same fine segmenting process. Specifically, an algorithm for generating spatiotemporal location labels, which consists of generating high-saliency location labels and tracking salient objects in adjacent frames, is proposed. Based on these location labels, a two-stream locating network that introduces an optical flow branch for video salient object locating is presented. Although our method does not require labeled video at all, the experimental results on five public benchmarks of DAVIS, FBMS, ViSal, VOS, and DAVSOD demonstrate that our proposed method is competitive with fully supervised methods and outperforms the state-of-the-art weakly and unsupervised methods. | ['Peng Chen', 'Ronghua Liang', 'Weihua Gong', 'Wentian Ni', 'Haoran Liang', 'Binwei Xu'] | 2022-04-05 | null | null | null | null | ['video-salient-object-detection'] | ['computer-vision'] | [ 1.48729265e-01 1.48817571e-02 -5.34061730e-01 -7.98362941e-02
-4.16280061e-01 -1.66253045e-01 3.04665565e-01 6.71264753e-02
-4.48104322e-01 6.76910996e-01 4.19131666e-01 2.55288184e-01
2.55653948e-01 -3.49125266e-01 -8.32997501e-01 -6.56670570e-01
5.87573610e-02 -2.91753769e-01 1.30618405e+00 -1.22594737e-01
3.63760740e-01 9.29234028e-02 -1.67144716e+00 1.91451207e-01
7.96762347e-01 1.20270836e+00 7.99046397e-01 2.85134852e-01
-1.38165772e-01 1.41777480e+00 -1.91410378e-01 9.59237963e-02
2.89703250e-01 -4.80586499e-01 -9.07979369e-01 3.12588930e-01
6.42728388e-01 -6.74056411e-01 -6.40066803e-01 1.10438883e+00
4.34661031e-01 3.11984956e-01 2.08951548e-01 -1.47795403e+00
-7.06739426e-01 7.91831195e-01 -8.03008199e-01 6.57034338e-01
2.27402449e-01 1.06663823e-01 9.17010128e-01 -1.14771211e+00
7.58972049e-01 1.13048053e+00 6.05500400e-01 5.12112379e-01
-7.24978328e-01 -3.43159407e-01 5.42556703e-01 5.14400780e-01
-1.38188720e+00 -4.06291336e-01 1.12461817e+00 -5.47894120e-01
5.63117146e-01 1.05292993e-02 7.92670727e-01 8.80996466e-01
-1.56404629e-01 1.36250257e+00 5.88126183e-01 -2.21784636e-01
2.33544186e-01 1.20921448e-01 8.69642477e-03 1.02474332e+00
2.12052032e-01 -3.21462913e-03 -7.75480270e-01 7.27413893e-02
9.64380324e-01 3.85705769e-01 -3.96447033e-01 -8.89245510e-01
-1.53148961e+00 5.14500082e-01 6.18581593e-01 1.53422296e-01
-4.10294324e-01 1.89694524e-01 6.22085273e-01 -3.25712293e-01
4.32316571e-01 1.48595283e-02 -3.16477001e-01 1.37696750e-02
-1.35616362e+00 5.77557869e-02 2.88736880e-01 1.24037707e+00
8.80338728e-01 2.10700408e-01 -7.29719222e-01 4.88322198e-01
2.13684291e-01 3.50784570e-01 5.86782873e-01 -1.01103151e+00
3.02547306e-01 7.30030298e-01 4.82757807e-01 -1.31034064e+00
-3.62737656e-01 -2.12076992e-01 -5.77929020e-01 -1.18447907e-01
1.35725230e-01 -1.38171371e-02 -9.24604535e-01 1.63277137e+00
5.18934131e-01 8.06567788e-01 -6.17335960e-02 1.56817567e+00
1.08034754e+00 6.70029581e-01 3.35223407e-01 -3.07076991e-01
9.90696907e-01 -1.57075143e+00 -7.89586902e-01 -1.35900676e-01
4.52956170e-01 -6.10075057e-01 1.02580500e+00 -1.74656898e-01
-1.29418266e+00 -8.06814849e-01 -9.60611045e-01 -3.50456864e-01
-6.17572591e-02 2.81410962e-01 5.62730968e-01 1.41917363e-01
-1.08563256e+00 3.40877652e-01 -8.33675623e-01 -3.42960566e-01
6.91897750e-01 1.53452260e-02 3.94385085e-02 1.08444169e-01
-1.17774844e+00 5.31062901e-01 6.73167706e-01 1.35901064e-01
-1.50083625e+00 -6.38885796e-01 -1.11526406e+00 1.94482088e-01
6.64549887e-01 -6.08095944e-01 1.03326023e+00 -1.36827886e+00
-1.13272762e+00 6.03528917e-01 -3.49803984e-01 -6.03740454e-01
6.61022127e-01 -3.59504431e-01 -2.15170071e-01 6.92517519e-01
4.82951909e-01 1.05254662e+00 9.90886211e-01 -1.15693760e+00
-1.10904384e+00 2.01507762e-01 1.72616601e-01 2.17362523e-01
-3.75105470e-01 2.42783681e-01 -6.97775781e-01 -7.96774268e-01
1.08323082e-01 -7.03229666e-01 -2.06405699e-01 3.84539545e-01
-5.21893144e-01 -3.08292776e-01 1.25183976e+00 -6.84161603e-01
1.28151190e+00 -2.28192115e+00 1.66894808e-01 -2.38199756e-01
3.57444018e-01 4.58581269e-01 -1.20183811e-01 5.43536898e-03
2.14921758e-01 -2.41817459e-01 -4.55232888e-01 -4.28333908e-01
-1.91251591e-01 -1.03041381e-01 -3.72126698e-01 4.92653579e-01
3.02425683e-01 9.97554064e-01 -1.49478793e+00 -1.13680661e+00
5.24816871e-01 3.05483013e-01 -5.04517436e-01 2.08867908e-01
-3.20237130e-01 2.11526200e-01 -5.51501036e-01 9.57563281e-01
6.12958789e-01 -3.86863440e-01 -2.45672837e-01 -2.81847209e-01
-4.37699497e-01 3.73646170e-02 -1.29764092e+00 1.86901879e+00
2.29492918e-01 6.04879200e-01 -1.10756755e-01 -1.05857265e+00
6.56788349e-01 9.80563536e-02 7.09879935e-01 -5.03193676e-01
-3.76061127e-02 2.72172660e-01 -5.46971858e-01 -6.89843178e-01
7.23192692e-01 3.78489226e-01 2.60574430e-01 1.82810545e-01
1.17722832e-01 4.94751602e-01 4.47843134e-01 5.63761711e-01
7.72793174e-01 4.84863997e-01 1.96588635e-01 -4.01491731e-01
8.17085922e-01 1.58189148e-01 1.08486104e+00 7.61013925e-01
-8.33109319e-01 9.49046791e-01 2.56449282e-01 -4.64943200e-01
-9.01277781e-01 -9.47229266e-01 1.84729412e-01 1.24722314e+00
1.07012749e+00 -4.38895106e-01 -7.97684014e-01 -8.70429635e-01
-1.33091271e-01 1.32846788e-01 -6.29175067e-01 -2.30418101e-01
-6.42141104e-01 -2.81544715e-01 1.28724232e-01 6.98982298e-01
8.38203132e-01 -1.40991652e+00 -8.84014189e-01 2.86346972e-01
-4.50168550e-01 -1.31519389e+00 -1.03465223e+00 -1.84386745e-01
-5.89891791e-01 -9.84759450e-01 -1.07611418e+00 -1.46691823e+00
6.77276492e-01 1.06779277e+00 8.46528172e-01 3.08661819e-01
-1.35006055e-01 1.93791211e-01 -4.48909789e-01 -2.39560768e-01
1.83603749e-01 -1.83893237e-02 6.33672774e-02 2.21616283e-01
1.32896692e-01 -1.95634246e-01 -9.81543481e-01 4.41908747e-01
-9.09852386e-01 3.63234997e-01 5.23563921e-01 7.04253137e-01
8.29169393e-01 -3.61956716e-01 6.75956309e-01 -5.90481937e-01
-1.07240379e-01 -6.16466165e-01 -4.95712757e-01 2.04050317e-01
-2.86110431e-01 -2.00551584e-01 3.74741852e-01 -3.69966418e-01
-1.07006252e+00 4.01017487e-01 3.81380260e-01 -9.19579685e-01
8.60393792e-03 2.58402884e-01 -7.81387687e-02 -1.93694845e-01
3.12400222e-01 5.83736777e-01 -2.11138442e-01 -3.75213891e-01
3.85186255e-01 5.31388342e-01 7.66731381e-01 -2.29160517e-01
7.24394739e-01 7.80926704e-01 -3.15689951e-01 -4.91929740e-01
-1.25405991e+00 -7.83717930e-01 -7.23688364e-01 -3.45305592e-01
1.03812778e+00 -1.23615766e+00 -2.92790920e-01 4.93203521e-01
-1.10702288e+00 -3.60543579e-01 -4.59073454e-01 4.68650728e-01
-5.92084050e-01 7.68671989e-01 -6.64203465e-01 -5.43845057e-01
-3.35593492e-01 -1.15844703e+00 1.27436161e+00 5.50361097e-01
1.26404002e-01 -7.40446031e-01 -3.49751681e-01 1.80460319e-01
3.09219241e-01 2.30012670e-01 2.92293727e-01 -2.19959334e-01
-1.05214620e+00 3.11595470e-01 -5.48519850e-01 2.43180275e-01
2.52307206e-01 -9.26898047e-02 -7.50785112e-01 -2.88552552e-01
-2.85077721e-01 -2.26368368e-01 1.05374241e+00 6.29032671e-01
1.36060262e+00 -2.74443328e-01 -4.48019296e-01 7.15233147e-01
1.22282958e+00 -2.57201958e-02 3.55042607e-01 5.17275512e-01
1.11111975e+00 5.29593170e-01 1.00261891e+00 5.04033983e-01
5.75926006e-01 6.48667336e-01 6.52389765e-01 -4.02359039e-01
-3.59047383e-01 -4.55873370e-01 6.71633780e-01 6.37248635e-01
-4.09166254e-02 1.11349612e-01 -4.09247279e-01 9.73126709e-01
-2.35638642e+00 -1.29537761e+00 -9.30772722e-02 1.80140281e+00
8.71724904e-01 8.40102732e-02 4.62348253e-01 -9.10681188e-02
1.14690018e+00 4.39277917e-01 -6.41018450e-01 1.14256054e-01
-4.06257540e-01 -4.89705294e-01 5.13812006e-01 1.97618753e-01
-1.45920658e+00 1.24513447e+00 5.76806164e+00 7.39722610e-01
-1.16013873e+00 2.45524064e-01 5.10792851e-01 -1.14191361e-01
-7.85027295e-02 -3.50740179e-02 -7.87029684e-01 8.48057687e-01
1.74052775e-01 -8.19314271e-02 1.64246008e-01 1.09734094e+00
5.55945873e-01 -1.56818524e-01 -7.90520012e-01 1.01367402e+00
3.25945497e-01 -1.67714131e+00 1.75465226e-01 -5.91669202e-01
1.16067517e+00 8.95292237e-02 1.25920295e-03 9.11208689e-02
-1.03784665e-01 -3.02247912e-01 1.33722448e+00 4.42497045e-01
3.03466797e-01 -5.40925264e-01 6.82065547e-01 1.98824704e-01
-1.49312544e+00 -2.07611755e-01 -5.21418989e-01 1.63574480e-02
3.69118333e-01 5.27409554e-01 -4.81806874e-01 3.98915976e-01
1.03501201e+00 1.37599754e+00 -6.48561180e-01 1.39626217e+00
-2.41851225e-01 6.51205122e-01 -5.54376468e-02 7.59981349e-02
5.19799411e-01 -4.53409590e-02 7.04869747e-01 1.27128339e+00
1.08604521e-01 -2.66115554e-02 6.01597726e-01 9.22629118e-01
2.21952572e-02 1.23897687e-01 -2.04956681e-01 2.59266675e-01
4.49410558e-01 1.32669365e+00 -9.11371946e-01 -6.36199713e-01
-6.11036181e-01 1.04458368e+00 1.66036531e-01 3.47721606e-01
-1.08341539e+00 -3.49167526e-01 3.22364122e-01 2.48090863e-01
6.81702912e-01 -8.73219296e-02 1.26469731e-02 -1.23756182e+00
1.44203141e-01 -3.75200629e-01 4.96285975e-01 -9.77484465e-01
-9.13580596e-01 2.64303654e-01 -5.76664805e-02 -1.46588349e+00
1.59650460e-01 4.24522348e-03 -6.94259226e-01 4.79011029e-01
-2.03554916e+00 -1.15972781e+00 -6.82008564e-01 8.48949671e-01
8.22275460e-01 -5.92273027e-02 -7.76227489e-02 4.08742636e-01
-7.15445578e-01 3.41711015e-01 -1.87013388e-01 3.13274026e-01
6.78953886e-01 -1.05889750e+00 1.65240899e-01 1.25707304e+00
4.26214226e-02 3.76457095e-01 4.53826874e-01 -6.86345339e-01
-1.09833527e+00 -1.57008779e+00 7.49132991e-01 -1.67495772e-01
5.87095082e-01 -2.09719047e-01 -9.53146994e-01 6.77152336e-01
2.07068726e-01 5.76588154e-01 9.98521131e-03 -6.08624220e-01
1.38546705e-01 -1.95193905e-02 -9.59975660e-01 5.72543263e-01
1.53006184e+00 -2.92814791e-01 -7.37371743e-01 3.61775905e-01
1.23274601e+00 -4.48059380e-01 -2.66156077e-01 3.83543819e-01
3.31697352e-02 -8.15072834e-01 1.12486005e+00 -4.52232301e-01
5.54263055e-01 -9.91279721e-01 1.04692899e-01 -7.70775914e-01
-3.70704561e-01 -7.84069955e-01 -4.55517888e-01 1.39649653e+00
-6.99264109e-02 -8.80595073e-02 8.02681506e-01 2.93528557e-01
-5.99791408e-01 -6.19713306e-01 -6.67902231e-01 -6.38831019e-01
-7.27256000e-01 6.45933747e-02 4.30440575e-01 1.00365412e+00
-1.26689583e-01 -4.30701524e-02 -6.77528560e-01 2.54683375e-01
7.32129455e-01 4.44429517e-01 5.76167881e-01 -9.66059983e-01
7.69910663e-02 -2.66332269e-01 -5.80152154e-01 -1.32971871e+00
1.58960968e-01 -7.64042437e-01 3.00614148e-01 -1.46298516e+00
5.95365584e-01 -3.38947862e-01 -8.44947577e-01 6.40871704e-01
-6.16127670e-01 4.03410792e-01 3.55308801e-01 3.76450628e-01
-1.52746844e+00 7.22700119e-01 1.36421502e+00 -2.47501299e-01
-3.59209627e-01 -4.20569748e-01 -6.13452852e-01 8.97837520e-01
5.31777680e-01 -3.59368056e-01 -4.39241946e-01 -3.73206198e-01
-2.66559750e-01 -2.06573159e-01 6.86146379e-01 -1.13405216e+00
3.94447178e-01 -2.90885061e-01 2.96004027e-01 -7.84043014e-01
7.28581250e-02 -5.71328819e-01 -3.82275909e-01 4.19772923e-01
-3.83179277e-01 -1.75558105e-01 5.89519879e-03 7.67314792e-01
-3.92814249e-01 -1.57067865e-01 8.76608670e-01 -1.40319904e-02
-1.64253759e+00 5.79364181e-01 -1.74040377e-01 3.00304890e-01
1.28042579e+00 -2.49399140e-01 -3.00748229e-01 -7.01169297e-02
-5.46113074e-01 5.71294665e-01 5.39210141e-01 7.18534231e-01
8.49038899e-01 -1.40045166e+00 -5.60271025e-01 -5.50127123e-03
1.86100543e-01 2.18481764e-01 4.64957029e-01 1.09137499e+00
-6.05915427e-01 2.54865795e-01 -3.81687671e-01 -8.13154101e-01
-9.99190986e-01 1.01688910e+00 2.27733880e-01 3.55799288e-01
-8.99691403e-01 8.75783503e-01 5.84455371e-01 1.54835030e-01
2.87438333e-01 -3.25750828e-01 -4.04279381e-01 1.72504038e-02
6.73091412e-01 3.29412401e-01 -4.14671034e-01 -9.39583242e-01
-3.57383370e-01 5.39169133e-01 1.64502598e-02 4.00099367e-01
1.19137049e+00 -5.64877093e-01 2.96741221e-02 3.37307155e-01
9.37866569e-01 -1.73685536e-01 -1.91102576e+00 -5.22799134e-01
1.76215172e-02 -6.91890776e-01 1.17163971e-01 -2.76228607e-01
-1.46037996e+00 6.66450918e-01 5.05726516e-01 -7.60474131e-02
1.09004438e+00 5.95005304e-02 1.20045090e+00 2.19869167e-02
3.70874435e-01 -1.31723654e+00 2.63792336e-01 2.07982436e-01
6.29098356e-01 -1.59111440e+00 -7.69343972e-02 -6.69982910e-01
-9.00624037e-01 7.31004179e-01 8.94612491e-01 -2.33988062e-01
3.69845778e-01 -1.79281041e-01 -1.25618145e-01 9.80246514e-02
-4.04033273e-01 -4.12567466e-01 2.63596714e-01 5.08414149e-01
-2.53503732e-02 -3.55671078e-01 -4.54353839e-01 5.12337804e-01
4.86480892e-01 2.32567385e-01 6.04889631e-01 1.05854297e+00
-7.69514024e-01 -3.87419969e-01 -6.55649528e-02 2.06091419e-01
-3.89403641e-01 -2.22320914e-01 -1.59076378e-01 6.18662119e-01
3.07532668e-01 7.85319448e-01 6.34449795e-02 -1.27350166e-01
7.28615522e-02 -4.64177459e-01 -7.05284439e-03 -5.22062242e-01
-2.84339547e-01 2.12536138e-02 -2.69226968e-01 -8.20006847e-01
-9.82291162e-01 -7.63553560e-01 -1.59664428e+00 7.83900693e-02
-2.23308668e-01 1.77010551e-01 7.23880157e-02 8.24903250e-01
4.51763749e-01 4.55280066e-01 6.33210599e-01 -1.22626019e+00
-1.63002193e-01 -5.77639937e-01 -6.31364346e-01 6.60723567e-01
6.27214670e-01 -1.03048563e+00 -3.19007069e-01 4.73228008e-01] | [9.663679122924805, -0.32023248076438904] |
3cffd820-9d19-4a97-98a8-0027638d8920 | automatic-parallel-corpus-creation-for-hindi | 1901.08625 | null | http://arxiv.org/abs/1901.08625v1 | http://arxiv.org/pdf/1901.08625v1.pdf | Automatic Parallel Corpus Creation for Hindi-English News Translation Task | The parallel corpus for multilingual NLP tasks, deep learning applications
like Statistical Machine Translation Systems is very important. The parallel
corpus of Hindi-English language pair available for news translation task till
date is of very limited size as per the requirement of the systems are
concerned. In this work we have developed an automatic parallel corpus
generation system prototype, which creates Hindi-English parallel corpus for
news translation task. Further to verify the quality of generated parallel
corpus we have experimented by taking various performance metrics and the
results are quite interesting. | ['Rakesh Chandra Balabantaray', 'Priyankit Acharya', 'Aditya Kumar Pathak', 'Dilpreet Kaur'] | 2019-01-24 | null | null | null | null | ['multilingual-nlp'] | ['natural-language-processing'] | [-2.19380766e-01 2.78473292e-02 1.94508266e-02 -1.22669026e-01
-1.12909615e+00 -5.36611438e-01 1.05799961e+00 -2.73399483e-02
-6.70275927e-01 1.35822463e+00 5.45956016e-01 -6.40700042e-01
4.01902169e-01 -7.45677054e-01 -7.47116864e-01 -2.71982938e-01
7.31316283e-02 1.18557847e+00 -1.16106227e-01 -6.83612943e-01
2.21547008e-01 2.00843319e-01 -8.05198252e-01 3.96273851e-01
6.20686829e-01 1.32590666e-01 4.47571218e-01 6.93750620e-01
-1.49761304e-01 3.78201514e-01 -6.03568196e-01 -4.61831629e-01
6.18606150e-01 -7.64320314e-01 -1.34936690e+00 -3.84904593e-01
-2.98564509e-02 -8.27514306e-02 -1.16700999e-01 1.02202356e+00
7.39551067e-01 -1.58542514e-01 5.34091771e-01 -8.68213534e-01
-4.73669082e-01 9.78884399e-01 -3.69824529e-01 6.60815477e-01
3.16316605e-01 -1.20657891e-01 9.84574080e-01 -8.99471164e-01
1.20718360e+00 9.11873519e-01 1.59178376e-01 3.44063729e-01
-1.01040804e+00 -4.58168089e-01 -4.59304750e-01 2.38302618e-01
-1.48222065e+00 -3.78223836e-01 6.97575867e-01 -1.46984205e-01
1.41629684e+00 2.42907647e-02 4.53226686e-01 1.33211458e+00
7.68354058e-01 7.76854038e-01 1.50684130e+00 -7.67183602e-01
-1.69789627e-01 4.36008871e-01 -1.18532129e-01 1.80209339e-01
-2.59252131e-01 -7.19917417e-02 -2.88621217e-01 6.87492266e-02
4.99545127e-01 -8.37745786e-01 9.16922167e-02 3.18724066e-01
-1.61792624e+00 9.21918392e-01 -3.50413099e-02 9.97127950e-01
-5.32764673e-01 -3.28607202e-01 1.02803516e+00 1.00709522e+00
5.35101116e-01 4.81699139e-01 -5.86842537e-01 -1.96670279e-01
-8.03997278e-01 4.07295287e-01 8.75701606e-01 1.07601273e+00
1.84498370e-01 -5.32757603e-02 1.78542778e-01 1.12366366e+00
-8.36207047e-02 6.42219305e-01 1.02651393e+00 -3.71850766e-02
9.36003864e-01 1.23632468e-01 -7.66639858e-02 -5.29485226e-01
-5.16636014e-01 -1.65061355e-01 -8.67739797e-01 -3.22889775e-01
1.80208847e-01 -4.40289736e-01 -2.33548760e-01 1.26681447e+00
3.08422863e-01 -6.22642517e-01 8.16833913e-01 8.20632398e-01
8.68416965e-01 1.30863273e+00 -1.23739704e-01 -3.02036554e-01
1.35710371e+00 -9.35355425e-01 -5.77835560e-01 9.89117771e-02
8.49221289e-01 -1.60428309e+00 1.24093032e+00 2.02161297e-01
-1.11534190e+00 -5.95512509e-01 -1.04893339e+00 -4.71161380e-02
-2.94416636e-01 1.31102622e-01 2.01162815e-01 1.98169291e-01
-1.01305103e+00 5.25295615e-01 -5.31269610e-01 -8.38265359e-01
-1.86119393e-01 3.81528020e-01 -8.10798228e-01 -3.26358825e-02
-1.47094965e+00 1.32229817e+00 6.40162289e-01 -2.23260745e-01
-4.97705877e-01 1.10092446e-01 -4.44912642e-01 -1.85522944e-01
-3.82968336e-01 -5.60012460e-01 1.26650417e+00 -1.14759505e+00
-1.60499251e+00 9.49343085e-01 1.26227677e-01 -7.51346052e-01
6.50531054e-01 -4.66890745e-02 -7.44415224e-01 -2.17074230e-01
3.45797807e-01 4.95106846e-01 2.78196167e-02 -7.14497328e-01
-8.24393928e-01 -3.68145198e-01 -2.27577388e-01 6.04432225e-01
-7.73660988e-02 6.50276124e-01 -2.77317176e-03 -5.24038374e-01
-3.31137121e-01 -1.17034316e+00 -2.66732574e-01 -1.08680928e+00
-3.17165732e-01 -2.74737179e-01 6.23521566e-01 -9.70173538e-01
6.57440126e-01 -1.55355895e+00 4.76209410e-02 -1.63774695e-02
-4.48155820e-01 3.25231761e-01 -4.22272921e-01 8.08784187e-01
-8.74080583e-02 -9.03600380e-02 2.09198575e-02 6.83510862e-03
-3.50340128e-01 2.02395886e-01 -2.69313842e-01 4.17936385e-01
1.36550084e-01 8.03491831e-01 -7.33187437e-01 -6.83273733e-01
-2.96119209e-02 2.01116681e-01 -7.91154876e-02 3.95266652e-01
-1.97884768e-01 7.03609049e-01 -3.96531850e-01 9.93411243e-02
4.25195187e-01 3.06636423e-01 1.50037289e-01 2.21309930e-01
-2.57350951e-01 8.23362112e-01 -5.12701750e-01 1.75842965e+00
-9.21402097e-01 9.29425716e-01 -3.40322435e-01 -1.05382586e+00
1.16463304e+00 9.09356534e-01 3.33922237e-01 -1.06344843e+00
2.96008945e-01 3.66829634e-01 4.48449582e-01 -5.30930996e-01
8.74554932e-01 -3.82001728e-01 -2.35177159e-01 7.13221788e-01
2.01911673e-01 1.64459869e-02 5.43607116e-01 -1.27424330e-01
7.92597651e-01 3.93645555e-01 5.58235347e-01 -5.79119086e-01
5.74139118e-01 6.87596619e-01 2.41622850e-01 3.28225344e-02
5.79119613e-03 7.32154727e-01 2.44764403e-01 -5.60891569e-01
-1.77848041e+00 -7.09789395e-01 -6.44829869e-02 8.15926433e-01
-4.35721368e-01 2.05959752e-02 -6.95522010e-01 -6.02693439e-01
-8.51809323e-01 6.55701101e-01 -3.73754203e-02 5.51816821e-01
-1.07802832e+00 -9.94266212e-01 6.81231320e-01 1.19866408e-01
2.75047600e-01 -1.48625636e+00 7.81275332e-02 6.15997314e-01
-6.44961119e-01 -1.35495329e+00 -3.88336003e-01 5.95652051e-02
-7.30528653e-01 -7.03614116e-01 -9.11567211e-01 -1.32066822e+00
2.38204733e-01 5.95334470e-02 1.24113321e+00 -4.28068429e-01
1.03644550e-01 -4.47879910e-01 -5.59525907e-01 -4.79837954e-01
-1.20389414e+00 7.13449061e-01 1.43587831e-02 -4.79763567e-01
5.14897227e-01 -8.41586351e-01 -1.39381230e-01 7.89153203e-02
-5.83234131e-01 4.14111167e-01 9.13811743e-01 7.55805850e-01
1.77402914e-01 -3.99693757e-01 8.52277994e-01 -1.01085675e+00
1.17564559e+00 -5.69258451e-01 -6.72769964e-01 6.66132495e-02
-3.28076184e-01 2.53282756e-01 1.18503392e+00 -2.39366427e-01
-1.04427743e+00 -1.39810681e-01 -6.53171241e-01 3.72677922e-01
-3.70627075e-01 6.88991487e-01 -1.85375631e-01 4.63855952e-01
1.09130096e+00 5.37507772e-01 -8.25199187e-02 -3.64416063e-01
2.17415616e-01 1.15468752e+00 2.76209325e-01 -4.74429578e-01
4.65206265e-01 -1.25914201e-01 -2.30279133e-01 -9.58075106e-01
-9.04529914e-02 -4.28485870e-01 -9.26709235e-01 2.62787547e-02
7.53507257e-01 -1.03770685e+00 -8.99045393e-02 1.02661029e-01
-1.68757761e+00 -2.61759251e-01 1.20054424e-01 7.98251867e-01
-6.59751534e-01 9.46310461e-02 -9.05095279e-01 -1.02205649e-01
-8.64728987e-01 -1.14504087e+00 7.51171470e-01 -2.88631767e-01
-4.83207315e-01 -9.49993968e-01 6.21501088e-01 2.62110025e-01
3.19463491e-01 9.03655663e-02 9.02090132e-01 -9.50593293e-01
-8.77502635e-02 -1.60758764e-01 -1.35053694e-02 1.40592873e-01
-1.49197444e-01 -1.60458952e-01 -5.05671501e-01 -1.51283428e-01
-5.99114634e-02 -3.62730950e-01 5.24646081e-02 1.66609317e-01
3.47501561e-02 -3.70413393e-01 -1.53255329e-01 -5.48494644e-02
1.37124658e+00 4.24273461e-01 6.64202929e-01 7.51940250e-01
4.27018851e-01 6.93584859e-01 6.04227304e-01 1.60209104e-01
4.51569289e-01 7.61294484e-01 -6.12221897e-01 -1.08267725e-01
-6.31947145e-02 -1.04173601e-01 5.41602194e-01 1.59829235e+00
1.85095537e-02 -3.48871380e-01 -1.07731807e+00 8.08205783e-01
-1.66648233e+00 -9.65488553e-01 -3.32864732e-01 1.82740712e+00
9.12575960e-01 1.24906413e-01 1.49212852e-01 -8.95689502e-02
7.46055365e-01 -6.07468821e-02 2.42787570e-01 -9.92099285e-01
-1.05480753e-01 3.36069047e-01 5.05872250e-01 5.60272098e-01
-7.29976237e-01 1.18027437e+00 5.72588587e+00 6.34277761e-01
-1.38632143e+00 4.43583310e-01 5.04842520e-01 1.82852745e-01
-1.16277248e-01 4.29520532e-02 -6.36511743e-01 2.12151557e-01
1.35043359e+00 -4.60234672e-01 4.14735645e-01 6.16607308e-01
4.03063744e-01 6.84974045e-02 -1.08231485e+00 8.35347235e-01
-1.65925443e-01 -1.28929925e+00 3.70669030e-02 5.91593459e-02
8.39497328e-01 7.42419660e-01 -2.47083277e-01 3.73932987e-01
5.44575214e-01 -6.88462973e-01 4.94422495e-01 -7.59929344e-02
5.51332176e-01 -9.58278239e-01 1.03420973e+00 6.06192112e-01
-6.75172150e-01 4.84310746e-01 -6.81516528e-01 -2.27728739e-01
3.80091548e-01 4.62962210e-01 -1.49284697e+00 5.91677189e-01
2.57222712e-01 1.88798651e-01 -1.17680743e-01 5.60407937e-01
-1.91145197e-01 7.59836555e-01 -3.21310818e-01 -3.43950629e-01
4.91002500e-01 -2.02030733e-01 6.25867069e-01 1.41540420e+00
4.26545888e-01 -2.28808865e-01 9.98948738e-02 5.14690280e-01
-8.93168300e-02 1.08606946e+00 -1.03474784e+00 -9.75282118e-02
1.37047380e-01 1.20932972e+00 -7.15529382e-01 -4.03649956e-01
-5.67076027e-01 1.03510106e+00 3.41037810e-01 7.88408220e-02
-4.88184690e-01 -2.99346864e-01 1.45764276e-01 2.01245502e-01
-1.91993684e-01 -4.31999773e-01 -2.94633031e-01 -1.08927023e+00
6.44533634e-02 -1.13433385e+00 5.40570766e-02 -6.21164799e-01
-1.08416426e+00 1.42087460e+00 -3.15284692e-02 -1.33133948e+00
-7.09750295e-01 -3.67805421e-01 -7.08757102e-01 1.18731201e+00
-1.04521227e+00 -1.46586835e+00 4.97170001e-01 4.22319978e-01
9.42719698e-01 -9.67401743e-01 1.04196882e+00 5.47970235e-01
-3.92902195e-01 4.45446908e-01 5.04835069e-01 3.15212280e-01
8.54603827e-01 -9.38380063e-01 1.07958066e+00 8.84333491e-01
3.82729977e-01 4.22532111e-01 9.49592710e-01 -5.78567028e-01
-1.11921895e+00 -9.43514764e-01 1.75616026e+00 -9.78529304e-02
7.50620246e-01 -3.18089306e-01 -5.11085391e-01 5.22376955e-01
1.01648796e+00 -3.81147802e-01 7.67611623e-01 1.37694776e-01
2.22148225e-01 8.58809277e-02 -8.95580411e-01 7.25848496e-01
5.04993737e-01 -2.35949829e-01 -6.41537130e-01 7.68152475e-01
5.32906532e-01 -3.29088837e-01 -9.88982260e-01 1.02381036e-02
4.13394243e-01 -4.96177047e-01 4.68289822e-01 -6.09250605e-01
1.00644469e+00 -1.83615714e-01 -5.51172458e-02 -1.70941103e+00
-3.00921053e-02 -5.14190018e-01 8.33341062e-01 1.27808058e+00
9.63839054e-01 -4.82714772e-01 6.59823358e-01 6.77103475e-02
-5.80351278e-02 -4.39519882e-01 -8.68955910e-01 -7.09045947e-01
4.80066866e-01 -1.13901272e-02 5.36888540e-01 1.13508105e+00
4.54295397e-01 1.40157032e+00 -5.47608674e-01 -2.49057278e-01
2.28459481e-02 2.18215629e-01 1.12162900e+00 -7.38101065e-01
-3.89613330e-01 -3.83281499e-01 -1.68853581e-01 -4.97871250e-01
3.37765366e-01 -1.22469270e+00 8.96974001e-03 -1.57580292e+00
1.97523296e-01 -3.53472471e-01 1.53674901e-01 1.08588345e-01
2.31775437e-02 1.72632009e-01 9.28599834e-02 4.57866788e-01
-4.65810709e-02 4.13316965e-01 1.20960402e+00 -1.97734796e-02
-1.89441741e-01 2.12653261e-02 -2.39751831e-01 1.72120675e-01
1.29399395e+00 -6.97659075e-01 -2.40289688e-01 -5.33141792e-01
2.78771501e-02 4.95960802e-01 -6.03912055e-01 -7.73802280e-01
-1.70456678e-01 -1.38998404e-01 2.25663006e-01 -3.36114854e-01
-1.76382706e-01 -5.20889044e-01 2.64749467e-01 3.71610403e-01
-3.81049514e-01 8.20951223e-01 1.36363819e-01 -2.80959636e-01
-6.91815257e-01 -1.32629022e-01 8.98573279e-01 -4.80581284e-01
-4.72693443e-01 -6.08835183e-03 -6.91524267e-01 4.38107923e-02
8.15930545e-01 2.14847818e-01 4.05163839e-02 -2.85984755e-01
-3.22441041e-01 -1.20812699e-01 2.59515136e-01 6.06265485e-01
1.72117904e-01 -1.39670789e+00 -1.43225610e+00 3.94096188e-02
1.86435953e-01 -3.38732362e-01 -1.57379240e-01 6.82184279e-01
-1.10761499e+00 7.69492388e-01 -9.52329576e-01 -1.85114831e-01
-1.34107733e+00 6.77539885e-01 -1.66042775e-01 -6.71461165e-01
-6.91296399e-01 3.23140293e-01 -2.12833390e-01 -8.65177035e-01
-4.02364939e-01 1.27789438e-01 -4.12816763e-01 -1.65506020e-01
4.21111315e-01 1.21569403e-01 3.36555064e-01 -1.14653969e+00
9.16358605e-02 1.52983181e-02 -3.47137392e-01 -6.88775957e-01
1.41134751e+00 -2.11696267e-01 -3.46976161e-01 3.25945318e-01
1.47157574e+00 -1.20630600e-01 -1.17673881e-01 -3.52974713e-01
7.24016130e-02 -9.87329632e-02 -3.80615443e-01 -4.97788370e-01
-4.81187403e-01 9.71138299e-01 5.48984408e-01 -1.03019148e-01
6.75756156e-01 -2.94360012e-01 1.20698547e+00 6.22885823e-01
4.79149282e-01 -1.41075671e+00 -5.77181995e-01 9.24978197e-01
9.53545332e-01 -1.29153919e+00 -1.96229443e-01 2.36303300e-01
-7.96702564e-01 1.20155859e+00 3.72344047e-01 -2.45895401e-01
2.39779890e-01 5.11082947e-01 5.91349304e-01 1.16743088e-01
-7.97639072e-01 3.45861465e-02 9.63644832e-02 5.34981966e-01
1.17734075e+00 1.67996883e-01 -1.39907980e+00 7.47551546e-02
-9.19983983e-01 -1.31826848e-01 5.91985404e-01 6.67474270e-01
-2.86219180e-01 -1.68842793e+00 -5.35214067e-01 6.15473241e-02
-1.02409816e+00 -3.28167319e-01 -4.72990036e-01 8.65347564e-01
-2.34476820e-01 6.58629715e-01 -9.94784236e-02 -2.87384093e-01
8.09410512e-02 5.79344220e-02 4.52266276e-01 -4.26552564e-01
-6.29934788e-01 1.48464754e-01 6.21471107e-01 6.49649138e-03
-3.80130351e-01 -6.32275999e-01 -1.12969112e+00 -4.77809727e-01
-2.10840672e-01 5.43827116e-01 1.03436863e+00 1.06115985e+00
-7.14132488e-02 2.81816348e-02 6.53675079e-01 -6.64440036e-01
-2.90060639e-01 -1.52669537e+00 -2.13121474e-01 2.94673473e-01
-2.35946521e-01 1.65409580e-01 4.75248486e-01 3.60895365e-01] | [11.485491752624512, 10.449694633483887] |
6a9cc9ea-b257-4347-91c1-de3188a4ccc6 | evolvehypergraph-group-aware-dynamic | 2208.05470 | null | https://arxiv.org/abs/2208.05470v1 | https://arxiv.org/pdf/2208.05470v1.pdf | EvolveHypergraph: Group-Aware Dynamic Relational Reasoning for Trajectory Prediction | While the modeling of pair-wise relations has been widely studied in multi-agent interacting systems, its ability to capture higher-level and larger-scale group-wise activities is limited. In this paper, we propose a group-aware relational reasoning approach (named EvolveHypergraph) with explicit inference of the underlying dynamically evolving relational structures, and we demonstrate its effectiveness for multi-agent trajectory prediction. In addition to the edges between a pair of nodes (i.e., agents), we propose to infer hyperedges that adaptively connect multiple nodes to enable group-aware relational reasoning in an unsupervised manner without fixing the number of hyperedges. The proposed approach infers the dynamically evolving relation graphs and hypergraphs over time to capture the evolution of relations, which are used by the trajectory predictor to obtain future states. Moreover, we propose to regularize the smoothness of the relation evolution and the sparsity of the inferred graphs or hypergraphs, which effectively improves training stability and enhances the explainability of inferred relations. The proposed approach is validated on both synthetic crowd simulations and multiple real-world benchmark datasets. Our approach infers explainable, reasonable group-aware relations and achieves state-of-the-art performance in long-term prediction. | ['Mykel J. Kochenderfer', 'Victoria Dax', 'Hengbo Ma', 'Jinkyoo Park', 'Chuanbo Hua', 'Jiachen Li'] | 2022-08-10 | null | null | null | null | ['relational-reasoning'] | ['natural-language-processing'] | [-2.29996353e-01 4.94457901e-01 -1.47995695e-01 -2.12605804e-01
2.34587744e-01 -3.15201402e-01 7.31677175e-01 5.32675087e-01
1.91093180e-02 8.23206604e-01 2.01580435e-01 -2.54606847e-02
-6.65071189e-01 -1.17631173e+00 -7.57421434e-01 -5.67627788e-01
-6.30148113e-01 1.10784566e+00 5.43866456e-01 -4.85692143e-01
-1.70554504e-01 3.38064820e-01 -1.44795954e+00 -9.04590040e-02
1.29658270e+00 5.22327125e-01 -1.61158904e-01 7.20863104e-01
2.50374228e-01 1.22846854e+00 -4.74721670e-01 -7.79124916e-01
1.46152928e-01 -4.05023962e-01 -7.78393328e-01 4.04536188e-01
-2.88836602e-02 -8.25158879e-02 -7.66000509e-01 7.12024391e-01
1.22956559e-01 4.70872104e-01 4.87735838e-01 -1.49490190e+00
-7.39031255e-01 9.56510246e-01 -4.19869632e-01 4.33984190e-01
2.49032035e-01 4.06947315e-01 1.06939912e+00 -1.08423904e-01
8.25988591e-01 1.29170167e+00 5.34361422e-01 1.81869879e-01
-9.91191268e-01 -3.44513357e-01 6.28406763e-01 4.58073765e-01
-1.30258250e+00 -5.84729426e-02 8.38299930e-01 -5.04537821e-01
1.08530712e+00 1.39051542e-01 1.21501589e+00 7.17163682e-01
3.33817869e-01 4.60536122e-01 4.44925159e-01 5.32817021e-02
2.73490816e-01 -1.64623916e-01 1.78413242e-01 1.21698368e+00
2.20076442e-01 -1.24177761e-01 -7.90001571e-01 -4.27827924e-01
7.02746034e-01 1.12874433e-01 -2.74872005e-01 -3.41029614e-01
-1.14716840e+00 5.48252821e-01 7.20319808e-01 7.55132586e-02
-7.93387890e-01 1.89826220e-01 1.59929633e-01 2.09640265e-01
6.82848513e-01 2.89029479e-01 -2.43690848e-01 -1.42689884e-01
-2.28674904e-01 2.59376168e-01 1.06056559e+00 9.88342226e-01
7.84851670e-01 -4.89656366e-02 -1.35208338e-01 5.24901509e-01
2.03221083e-01 2.40933433e-01 4.56169210e-02 -1.01962101e+00
5.55161476e-01 1.27459633e+00 6.34764954e-02 -1.43491805e+00
-5.68333030e-01 -4.83171880e-01 -1.08338523e+00 -4.05887097e-01
2.37411276e-01 -2.78920472e-01 -5.70668519e-01 1.77793014e+00
6.78385735e-01 8.06070447e-01 1.96211711e-01 7.40435600e-01
5.23024559e-01 6.60209656e-01 -3.09887212e-02 -4.66400832e-01
1.00259864e+00 -1.36491394e+00 -6.49673700e-01 -1.03676818e-01
4.62514162e-01 1.55973390e-01 6.01313829e-01 -5.69036566e-02
-1.25223267e+00 -2.79239833e-01 -5.74355304e-01 3.04896712e-01
-4.81606275e-02 -4.04259682e-01 1.01481521e+00 7.14551331e-03
-1.00454879e+00 5.27989507e-01 -1.21632624e+00 -4.53324616e-01
2.58083284e-01 4.34999019e-01 -1.12593323e-01 2.84543067e-01
-1.32740772e+00 7.28778780e-01 3.91788483e-01 1.57929435e-01
-8.70214939e-01 -6.54626787e-01 -6.84515417e-01 2.66740382e-01
8.35655749e-01 -9.67384458e-01 6.86058104e-01 -6.02387905e-01
-1.35802817e+00 2.38352537e-01 -1.33361727e-01 -7.07164943e-01
5.38903654e-01 2.24388704e-01 -3.42239350e-01 1.63826793e-01
-2.19584718e-01 2.88028240e-01 4.69649076e-01 -1.25621092e+00
-5.56312263e-01 -2.60273814e-01 4.27900344e-01 5.95527112e-01
-5.10459304e-01 -4.68436658e-01 -6.40866458e-01 -3.15623522e-01
2.06894148e-02 -1.28526533e+00 -3.39255691e-01 -1.81674913e-01
-3.83644342e-01 -5.21215081e-01 4.93373334e-01 -5.21140158e-01
1.35572100e+00 -1.67987299e+00 7.53255308e-01 3.37638974e-01
6.24354362e-01 1.16200127e-01 7.52652287e-02 6.80772960e-01
5.65069139e-01 -6.56324476e-02 -3.49098891e-01 -3.46008271e-01
-1.58989832e-01 6.18103921e-01 4.12311405e-02 3.44005436e-01
-1.41773179e-01 1.05372846e+00 -1.13215375e+00 -5.03855884e-01
-6.92153797e-02 4.88090098e-01 -5.74208856e-01 3.19145620e-01
-5.25511503e-01 7.06092238e-01 -7.32318342e-01 2.59009451e-01
2.84836948e-01 -8.16540658e-01 6.65975630e-01 9.03045163e-02
2.59167552e-01 -1.44411713e-01 -9.50113654e-01 1.24544489e+00
-2.61469603e-01 4.16205555e-01 -1.20267540e-01 -8.42440248e-01
1.02182078e+00 1.74748078e-01 7.04620838e-01 -2.85325706e-01
-9.85240191e-02 -2.60631114e-01 3.07648391e-01 -6.56140089e-01
4.36563104e-01 4.91153687e-01 1.33707121e-01 5.03401160e-01
-2.60866791e-01 1.68535203e-01 4.55235898e-01 6.37003005e-01
1.26760042e+00 -1.29573373e-02 2.37043828e-01 -1.34621650e-01
6.74246848e-01 -2.08202451e-02 6.83647573e-01 5.65587997e-01
6.45528957e-02 -2.15481713e-01 6.08371854e-01 -7.46077001e-01
-8.26539278e-01 -6.45111382e-01 4.58510429e-01 9.52322066e-01
7.06284583e-01 -5.29622257e-01 -5.85772932e-01 -6.27947986e-01
2.87390858e-01 4.82422858e-01 -8.17853451e-01 -3.24422240e-01
-7.72771597e-01 -7.52623081e-01 2.61177868e-01 2.71282345e-01
5.87728739e-01 -1.19616485e+00 -5.86767495e-01 3.77073348e-01
-3.37549746e-01 -1.14443219e+00 -3.60426873e-01 -2.65154898e-01
-6.44863427e-01 -1.28515613e+00 -2.47908663e-02 -6.06207073e-01
8.44069004e-01 3.62004787e-02 1.09830225e+00 6.69501781e-01
1.74540177e-01 5.14696419e-01 -3.10391933e-01 7.97448084e-02
-3.32707465e-01 2.98577040e-01 8.50969478e-02 2.99778730e-01
-1.28707036e-01 -9.14329469e-01 -4.88821983e-01 4.52158749e-01
-5.59476197e-01 2.58175671e-01 2.09035859e-01 7.16762483e-01
5.06346107e-01 4.84975219e-01 4.72708076e-01 -1.05733037e+00
8.21446300e-01 -6.72903895e-01 -6.07425928e-01 6.97890401e-01
-7.75253713e-01 2.42280707e-01 7.27781475e-01 -7.43699014e-01
-1.25954628e+00 -4.24052477e-01 4.96756166e-01 -3.55837941e-01
2.25491881e-01 7.25132346e-01 1.38746679e-01 -5.15984669e-02
4.54696923e-01 1.62874907e-01 -1.29314631e-01 1.34852165e-02
3.61009777e-01 1.32392451e-01 4.52315331e-01 -7.21232593e-01
8.40558529e-01 4.99700189e-01 3.28540534e-01 -4.88916367e-01
-6.21455669e-01 -1.34530544e-01 -5.83524048e-01 -5.62415898e-01
6.96645021e-01 -7.86905050e-01 -1.40008962e+00 4.42081839e-01
-1.06593513e+00 -7.08936155e-01 -1.13146566e-01 2.14051187e-01
-3.75888318e-01 1.74522027e-01 -9.00872588e-01 -9.65473771e-01
-3.23375672e-01 -1.01281905e+00 7.03930676e-01 4.02459592e-01
-6.18926175e-02 -1.49582446e+00 2.51452327e-01 3.83467823e-01
2.93630004e-01 4.42718506e-01 8.68042290e-01 -7.32857287e-01
-1.10969317e+00 1.49617895e-01 1.23125836e-01 -6.38668954e-01
5.63036911e-02 8.79554898e-02 -1.43739253e-01 -1.80117890e-01
-5.26489377e-01 -2.50336304e-02 6.03531539e-01 1.22923508e-01
6.69047117e-01 -7.10407734e-01 -6.92475915e-01 4.80199754e-01
8.87086034e-01 3.18700373e-01 2.31613711e-01 1.67755678e-01
9.24151540e-01 7.09986806e-01 5.65279007e-01 7.68000722e-01
1.30073190e+00 7.72416115e-01 4.77290541e-01 3.18177104e-01
1.28970236e-01 -3.26720774e-01 1.14556208e-01 1.10718501e+00
-6.24094725e-01 -7.02992737e-01 -1.07203686e+00 4.94849294e-01
-2.50459766e+00 -1.00853884e+00 -2.98344404e-01 1.64909470e+00
6.45436347e-01 -4.57170531e-02 1.46251708e-01 -3.93806845e-01
8.74402583e-01 2.09762946e-01 -9.85934436e-01 4.32040542e-02
-6.66766390e-02 -4.75196660e-01 9.56579074e-02 8.24805081e-01
-6.09542787e-01 1.11001885e+00 5.17913580e+00 2.46088535e-01
-5.56853056e-01 9.29246247e-02 5.75109243e-01 2.56414991e-03
-3.46998483e-01 5.90428412e-02 -4.66776848e-01 3.86353135e-01
8.16117108e-01 -4.26101893e-01 1.07216191e+00 6.09373569e-01
7.84022436e-02 -3.53415459e-02 -1.00852430e+00 5.63698828e-01
9.31892265e-03 -1.47370803e+00 -1.26064639e-03 1.77639708e-01
1.08174086e+00 -2.72391737e-01 -2.22084954e-01 3.06633472e-01
8.14225316e-01 -5.74256718e-01 5.42967021e-01 1.08367884e+00
-1.91743243e-02 -8.14610600e-01 6.49502575e-01 6.18633509e-01
-1.59992898e+00 -1.56358391e-01 -1.49873123e-01 -2.34511524e-01
3.60217839e-01 2.52439946e-01 -7.35109508e-01 9.04678166e-01
5.34938872e-01 1.00995111e+00 -6.11085415e-01 6.95118129e-01
-2.75969863e-01 4.30433303e-01 -3.01804632e-01 -1.66766435e-01
-2.73626484e-02 -4.92908776e-01 7.06583142e-01 5.54490983e-01
5.40294610e-02 6.22169793e-01 5.54894030e-01 6.46921575e-01
-3.35558839e-02 -3.42822522e-01 -3.37668449e-01 8.50795433e-02
8.57781589e-01 1.08168924e+00 -7.38602161e-01 -4.36108083e-01
-2.20274597e-01 7.46305823e-01 9.71473515e-01 5.19950211e-01
-9.77069497e-01 3.10308844e-01 6.72906995e-01 1.00461267e-01
1.79945081e-01 -2.97045946e-01 8.16054866e-02 -1.13834369e+00
-3.23040113e-02 -7.63347447e-01 6.62544906e-01 -4.90370333e-01
-1.25786662e+00 9.65951145e-01 1.13293350e-01 -7.11825371e-01
-3.42938572e-01 2.42368579e-01 -7.25258172e-01 3.50002259e-01
-1.24855244e+00 -1.27123809e+00 -7.46000588e-01 6.61172152e-01
2.49046326e-01 -2.63296634e-01 4.28334653e-01 -7.59508982e-02
-8.88072848e-01 2.40333825e-01 -3.07521611e-01 -1.53703183e-01
1.42750815e-01 -9.92770255e-01 4.57619786e-01 7.06269383e-01
8.52831751e-02 6.10243797e-01 7.48088598e-01 -1.13978422e+00
-1.47229528e+00 -1.18462706e+00 5.80759168e-01 -2.43608057e-01
9.35248017e-01 -1.39778763e-01 -1.11960661e+00 1.00643146e+00
2.24404901e-01 2.28902996e-01 3.31726134e-01 1.81240022e-01
-3.30062322e-02 -1.51045725e-01 -8.63101184e-01 6.78586304e-01
1.56685460e+00 -1.76862955e-01 -1.71387359e-01 5.68354845e-01
8.99905503e-01 -5.20712018e-01 -1.24816608e+00 4.07343239e-01
3.25367451e-01 -8.19871068e-01 8.91399562e-01 -6.06846452e-01
2.64258862e-01 -3.43775868e-01 2.95074552e-01 -1.45455909e+00
-4.84309584e-01 -7.03056931e-01 -8.12540531e-01 1.20523441e+00
4.28224355e-01 -9.33045924e-01 9.13914204e-01 9.58497167e-01
7.78514370e-02 -1.02970493e+00 -5.85896492e-01 -5.55531740e-01
-6.15245879e-01 2.40421563e-01 9.53060865e-01 1.06424022e+00
2.71747559e-01 4.89422262e-01 -6.75682545e-01 6.54377818e-01
7.18331695e-01 3.32135618e-01 9.29889619e-01 -1.32919407e+00
-7.19355941e-01 -3.00329775e-01 -4.24747497e-01 -9.36270118e-01
5.48977911e-01 -7.87949800e-01 -2.32016385e-01 -1.74745119e+00
3.20609063e-01 -6.95021212e-01 9.39775705e-02 3.69054765e-01
-5.67439437e-01 -2.94830441e-01 1.65055126e-01 4.16009873e-01
-1.16607499e+00 9.36119258e-01 1.42747152e+00 -1.92268789e-01
-5.55948496e-01 9.29439589e-02 -1.58346102e-01 6.15442574e-01
6.28427088e-01 -3.89523685e-01 -9.15013671e-01 -3.42133939e-01
4.66418147e-01 6.25401318e-01 3.58730346e-01 -8.32548320e-01
7.81275749e-01 -4.51229542e-01 -1.73617437e-01 -2.82733649e-01
5.10579169e-01 -8.33507061e-01 8.10695469e-01 4.27473128e-01
-6.33905411e-01 2.12229639e-01 -1.69535935e-01 1.07529974e+00
-2.92118806e-02 1.96874022e-01 1.63734600e-01 -7.09881186e-02
-3.66870403e-01 7.58322060e-01 -4.76967879e-02 1.01884872e-01
1.36283231e+00 1.83269277e-01 -6.48624003e-01 -7.57700264e-01
-8.10738564e-01 8.42548192e-01 4.88400400e-01 2.51325041e-01
4.15513754e-01 -1.08774006e+00 -6.48491979e-01 -2.30771139e-01
-3.30682546e-02 2.17808023e-01 4.53657925e-01 9.27739620e-01
-3.03561449e-01 -4.73568775e-02 -1.99173540e-02 -5.28885782e-01
-1.17128861e+00 6.33838236e-01 4.80004787e-01 -7.44730711e-01
-8.42127681e-01 7.33901739e-01 -3.89052071e-02 -3.76270622e-01
1.36183143e-01 -5.69049865e-02 -5.29696465e-01 -2.02178985e-01
1.41119972e-01 5.40818632e-01 -5.01472056e-01 -7.13593483e-01
-3.04365188e-01 3.64477545e-01 -3.56065817e-02 2.36157611e-01
1.64332104e+00 -2.60002553e-01 -3.28392744e-01 4.21243012e-01
4.42399770e-01 -1.13078922e-01 -1.65251184e+00 -4.19475168e-01
3.32191284e-03 -2.76559383e-01 -3.35981876e-01 -4.34584349e-01
-1.33018100e+00 2.66903251e-01 -2.94285506e-01 6.48780882e-01
8.81635845e-01 2.93163747e-01 7.05418527e-01 6.09873414e-01
7.31000245e-01 -6.80206060e-01 2.60275722e-01 4.24213439e-01
5.80076933e-01 -9.10729051e-01 1.66736022e-01 -7.59712279e-01
-1.07447040e+00 8.10335457e-01 9.20740664e-01 9.53416247e-03
5.36496103e-01 8.99871960e-02 -4.87812430e-01 -4.97184068e-01
-1.30404985e+00 -2.09459767e-01 1.94580629e-01 6.04536593e-01
-2.04804584e-01 1.97070643e-01 2.72136256e-02 3.43105108e-01
-1.17365055e-01 -3.13810676e-01 4.46380645e-01 5.75576425e-01
-2.62790024e-01 -8.18252146e-01 9.51611325e-02 4.23505902e-01
2.35266194e-01 2.23100469e-01 -4.55333382e-01 6.27606452e-01
4.96022776e-03 1.01386106e+00 1.40450835e-01 -5.14143705e-01
2.97435045e-01 -2.70988315e-01 3.25649172e-01 -2.77383119e-01
-8.25811565e-01 -4.68523979e-01 4.25280541e-01 -4.17182058e-01
-6.05781198e-01 -7.47900367e-01 -1.46512771e+00 -7.64178991e-01
-2.55161971e-01 3.08930963e-01 -1.02267884e-01 1.04227149e+00
7.24604487e-01 8.64332497e-01 4.25249577e-01 -5.31047940e-01
5.42991422e-02 -8.34729493e-01 -2.47895136e-01 6.57635272e-01
1.21237166e-01 -9.64766681e-01 -1.51547626e-01 3.86051014e-02] | [5.925436973571777, 0.8793867826461792] |
99f3ef8b-b4b9-4866-9d61-2c629d1cc6b7 | deep-phase-correlation-for-end-to-end | 2008.09474 | null | https://arxiv.org/abs/2008.09474v4 | https://arxiv.org/pdf/2008.09474v4.pdf | Deep Phase Correlation for End-to-End Heterogeneous Sensor Measurements Matching | The crucial step for localization is to match the current observation to the map. When the two sensor modalities are significantly different, matching becomes challenging. In this paper, we present an end-to-end deep phase correlation network (DPCN) to match heterogeneous sensor measurements. In DPCN, the primary component is a differentiable correlation-based estimator that back-propagates the pose error to learnable feature extractors, which addresses the problem that there are no direct common features for supervision. Also, it eliminates the exhaustive evaluation in some previous methods, improving efficiency. With the interpretable modeling, the network is light-weighted and promising for better generalization. We evaluate the system on both the simulation data and Aero-Ground Dataset which consists of heterogeneous sensor images and aerial images acquired by satellites or aerial robots. The results show that our method is able to match the heterogeneous sensor measurements, outperforming the comparative traditional phase correlation and other learning-based methods. Code is available at https://github.com/jessychen1016/DPCN . | ['Xuecheng Xu', 'Rong Xiong', 'Yue Wang', 'Zexi Chen'] | 2020-08-21 | null | null | null | null | ['template-matching'] | ['computer-vision'] | [ 9.34726149e-02 -6.33346289e-02 -4.45699878e-02 -4.40935612e-01
-9.68466938e-01 -5.44740736e-01 4.24191982e-01 -1.44143641e-01
-2.28082538e-01 5.82510352e-01 3.12663685e-03 1.30050123e-01
-4.67693090e-01 -6.54419720e-01 -7.23949909e-01 -8.24527502e-01
-1.83582515e-01 2.96462953e-01 2.10552007e-01 -1.52935028e-01
-1.05371036e-01 4.12045270e-01 -1.35178423e+00 -2.94191748e-01
7.08034456e-01 1.39684665e+00 5.12377739e-01 4.89904612e-01
5.89485645e-01 5.62349498e-01 -1.48161426e-01 1.70224756e-01
5.65409541e-01 -1.63371623e-01 -3.75823975e-01 -1.47629172e-01
7.66383648e-01 -1.03073828e-01 -4.23801541e-01 1.32387447e+00
6.86271369e-01 4.98003103e-02 2.92658836e-01 -1.28019214e+00
-2.11900368e-01 4.79717553e-01 -5.54883957e-01 -9.63344425e-02
3.92223537e-01 -3.81745137e-02 8.50378990e-01 -8.08149517e-01
3.78312290e-01 1.07295763e+00 1.01637316e+00 1.45783171e-01
-9.02349472e-01 -6.97382629e-01 1.15926839e-01 3.07841212e-01
-1.68650746e+00 -2.48510092e-01 8.10863316e-01 -4.50740576e-01
4.14977372e-01 1.60258338e-02 6.64198756e-01 1.03809106e+00
2.87648767e-01 4.57770437e-01 1.00134480e+00 -1.20224148e-01
1.09026968e-01 -1.74580619e-01 1.16875414e-02 8.79479647e-01
3.97717088e-01 5.45860767e-01 -8.64077985e-01 -7.66324401e-02
4.60230082e-01 4.24323380e-01 -5.97693026e-01 -6.11647010e-01
-1.37779450e+00 4.08131510e-01 1.02948332e+00 1.09029897e-01
-7.05778599e-01 1.94823578e-01 -5.49812354e-02 2.35276982e-01
2.81505078e-01 5.70640266e-01 -5.77345073e-01 1.45326361e-01
-8.04984033e-01 8.76431093e-02 6.85711324e-01 9.61381853e-01
1.06443334e+00 1.33114755e-01 2.15205252e-01 2.70576209e-01
7.03420222e-01 1.32501817e+00 3.27937573e-01 -8.65314245e-01
5.87129653e-01 5.27129650e-01 2.15721607e-01 -1.28978097e+00
-7.99908340e-01 -7.52786815e-01 -1.13596725e+00 2.08884135e-01
1.77683398e-01 -5.65251708e-01 -5.84552526e-01 1.71942675e+00
2.51518637e-01 4.39194232e-01 3.09220791e-01 1.25849247e+00
6.80263221e-01 4.60651428e-01 -3.88846785e-01 1.77069470e-01
1.09852827e+00 -8.95663559e-01 -7.55117774e-01 -6.31869733e-01
3.41159135e-01 -6.25721037e-01 5.97339630e-01 2.54668862e-01
-5.25850236e-01 -7.41115808e-01 -1.56466842e+00 4.35745895e-01
-3.15649420e-01 4.20391113e-01 4.90497530e-01 1.65067818e-02
-7.45771885e-01 8.29258382e-01 -1.06454527e+00 -4.59372044e-01
5.86031191e-02 3.60293567e-01 -4.27139342e-01 -1.78809658e-01
-1.17153466e+00 8.95600140e-01 5.48585713e-01 5.06091595e-01
-8.59678209e-01 -7.23193944e-01 -1.02717829e+00 -1.18924081e-01
3.16652536e-01 -7.11208999e-01 1.04343927e+00 -7.18873918e-01
-1.55752909e+00 1.90815091e-01 -4.29686792e-02 -4.88692582e-01
3.29669207e-01 -4.45126504e-01 -4.56778795e-01 6.90348372e-02
3.43944202e-03 6.23370886e-01 8.39436710e-01 -1.32657826e+00
-8.01111519e-01 -4.60450798e-01 -1.14273891e-01 2.02251166e-01
-5.50037064e-02 -5.75349569e-01 -3.08529407e-01 -3.14390242e-01
7.82773197e-01 -1.39658689e+00 -2.45429054e-01 -1.90211013e-02
-4.49878037e-01 2.45707765e-01 6.83756888e-01 -5.65079927e-01
8.00763547e-01 -2.21635580e+00 2.38559291e-01 4.69696254e-01
2.81837761e-01 -5.24682365e-02 -2.26876006e-01 5.53577244e-01
1.49146989e-02 -4.68939096e-01 -2.72682667e-01 -3.69331926e-01
1.62707031e-01 6.53796941e-02 -2.55545914e-01 9.58436191e-01
1.14082759e-02 7.16426253e-01 -9.54753101e-01 -1.51890695e-01
3.10777932e-01 5.07350802e-01 -1.80113569e-01 2.88169086e-01
1.33059710e-01 7.88635254e-01 -4.15467918e-01 8.73272181e-01
8.20547163e-01 -3.06584805e-01 1.36083931e-01 -8.20327282e-01
-1.48751765e-01 1.13753609e-01 -1.44284832e+00 2.13722992e+00
-4.31014806e-01 7.52704144e-01 1.52559623e-01 -7.13478565e-01
1.12183642e+00 1.42113894e-01 5.61390340e-01 -7.12033272e-01
1.03996374e-01 3.94741029e-01 -4.59875390e-02 -3.83584172e-01
3.35114866e-01 2.91524947e-01 -2.89016277e-01 -1.27923936e-01
1.21099114e-01 -4.32840019e-01 -1.21017821e-01 -1.51573196e-01
1.09688914e+00 3.54384601e-01 3.24267387e-01 -2.74859130e-01
5.43934286e-01 7.43838027e-02 9.35219765e-01 5.92987061e-01
4.65533473e-02 6.33356333e-01 -2.47125953e-01 -3.94037634e-01
-6.78009450e-01 -1.10171485e+00 -3.36075537e-02 3.99501204e-01
7.32284248e-01 -4.27484900e-01 -3.14490706e-01 -3.98420960e-01
1.07911050e-01 3.20611745e-02 -3.38716835e-01 -2.38552526e-01
-3.26509416e-01 -5.30949533e-01 3.38242233e-01 5.68772793e-01
9.00110245e-01 -4.05917555e-01 -7.14481652e-01 -4.36293818e-02
-2.46304452e-01 -1.19840086e+00 -1.11644834e-01 2.83140868e-01
-7.68522680e-01 -1.16411400e+00 -4.05926406e-01 -5.80493629e-01
7.22843766e-01 4.73857701e-01 6.55379295e-01 -1.75903782e-01
-6.74469210e-03 4.41495121e-01 -4.10403103e-01 -4.22003210e-01
1.65043026e-01 1.67339936e-01 4.58971530e-01 -1.27173886e-01
1.58343807e-01 -7.93954313e-01 -4.85354483e-01 2.59796530e-01
-4.97470021e-01 -3.02999020e-02 7.88803101e-01 8.62833321e-01
7.68599868e-01 -1.42215535e-01 2.11773440e-02 -3.28818709e-01
1.86526984e-01 -3.43138158e-01 -1.09528708e+00 1.14321308e-02
-7.45194077e-01 6.77770451e-02 5.94194233e-01 -2.42026433e-01
-6.75981462e-01 5.68591356e-01 1.12543873e-01 -6.32024348e-01
-8.53897184e-02 8.47746611e-01 -5.44752404e-02 -4.95692909e-01
7.53466427e-01 3.73952687e-02 -3.93429697e-02 -4.58612114e-01
-1.65963043e-02 4.49780822e-01 7.51509428e-01 -2.99228132e-01
1.32539105e+00 5.14333665e-01 4.12793428e-01 -8.47225130e-01
-1.10365307e+00 -5.61711729e-01 -6.80288076e-01 -1.55074373e-01
5.60697317e-01 -1.35870779e+00 -6.35294378e-01 4.09037024e-01
-1.09405184e+00 -2.19537899e-01 -2.57740706e-01 1.08854353e+00
-3.96666497e-01 2.89811790e-01 -3.09609801e-01 -5.13667643e-01
-3.54085833e-01 -9.84881401e-01 1.27917743e+00 5.50644517e-01
-4.20821179e-03 -8.93164217e-01 4.22762930e-01 5.84825575e-02
4.06699896e-01 3.48037660e-01 4.37005647e-02 -4.51842189e-01
-8.11773777e-01 -4.97215241e-01 -1.20166875e-01 2.67906040e-01
7.02454597e-02 -2.22416863e-01 -1.00665581e+00 -6.52211130e-01
-1.84526160e-01 -3.21070045e-01 7.63964295e-01 2.05333129e-01
7.12820292e-01 -5.84652498e-02 -4.37599123e-01 1.13881290e+00
1.66833615e+00 1.79953780e-03 2.81812757e-01 4.24121797e-01
7.59778976e-01 4.56669271e-01 8.83726299e-01 4.72609192e-01
5.07676661e-01 6.48779452e-01 8.74392629e-01 -2.91356295e-01
1.36622906e-01 -3.49924415e-01 4.15090978e-01 9.38967884e-01
-9.51145813e-02 -1.57988742e-01 -1.07339108e+00 2.97849834e-01
-2.16174889e+00 -6.80064559e-01 -1.20185584e-01 2.13820648e+00
1.97306648e-01 -1.46292448e-01 -5.02025545e-01 -2.24344373e-01
4.64834481e-01 3.66400361e-01 -5.32001436e-01 6.62320852e-01
-1.12538613e-01 -1.99670773e-02 9.44228411e-01 6.31247342e-01
-1.34787011e+00 7.04490006e-01 5.27115297e+00 2.40810454e-01
-1.23440492e+00 9.23112407e-02 -3.17981392e-01 1.53334573e-01
7.95733035e-02 1.19967379e-01 -8.42871189e-01 2.69936204e-01
7.61504471e-01 7.86977354e-03 3.89614642e-01 8.04977477e-01
-7.27548748e-02 -9.95678976e-02 -1.12988019e+00 1.19837201e+00
5.87876551e-02 -1.19859922e+00 -4.90708977e-01 -1.56783968e-01
6.93766356e-01 6.26911759e-01 -2.16553994e-02 1.35156408e-01
2.27283865e-01 -5.18138289e-01 7.14059591e-01 7.69655049e-01
5.01771033e-01 -4.88676101e-01 1.23365736e+00 3.61529589e-01
-1.46200144e+00 -8.85944515e-02 -4.60583568e-01 -5.49759641e-02
1.39591545e-01 6.65014386e-01 -6.92024946e-01 1.08990204e+00
8.72943461e-01 1.19450903e+00 -5.26277184e-01 1.15586829e+00
-4.74505454e-01 3.29787701e-01 -6.21060252e-01 1.86208904e-01
1.45497441e-01 -3.55343401e-01 6.42773271e-01 8.80976617e-01
8.68685842e-01 -7.21499696e-02 5.98313391e-01 7.19887853e-01
1.98644117e-01 -3.68824184e-01 -8.17430496e-01 1.98937073e-01
6.67419553e-01 1.49716270e+00 -2.24358246e-01 -1.37914225e-01
-3.82094383e-01 7.37880528e-01 1.17610814e-02 3.68054122e-01
-7.03010142e-01 -3.48276019e-01 6.53793454e-01 -3.14576626e-01
3.15823197e-01 -4.69560921e-01 1.89572304e-01 -1.15002608e+00
1.93805054e-01 -7.67911315e-01 -2.83709280e-02 -7.80800939e-01
-1.21630466e+00 7.65664101e-01 -2.86099687e-02 -1.98466301e+00
-3.20926785e-01 -6.22920692e-01 -3.46418798e-01 8.13128471e-01
-1.72556925e+00 -1.29620886e+00 -1.12881708e+00 4.86767173e-01
-1.02806166e-02 -1.42033830e-01 7.07505286e-01 4.65828896e-01
-5.17678440e-01 1.92776367e-01 5.98307736e-02 1.69555783e-01
8.55593801e-01 -1.03314435e+00 2.05726251e-01 1.21301329e+00
1.74359158e-01 3.04091662e-01 8.30311298e-01 -5.28573513e-01
-1.68297708e+00 -1.23682833e+00 6.20765686e-01 -1.82040274e-01
7.12020993e-01 -2.13409066e-01 -5.49382091e-01 6.73302889e-01
2.29974657e-01 4.31117982e-01 3.81616026e-01 -1.22745350e-01
-2.12313756e-01 -6.57981277e-01 -7.78206110e-01 1.87619254e-01
1.02053595e+00 -4.80328351e-01 -4.08212692e-01 4.30012733e-01
6.32695973e-01 -7.36043990e-01 -9.88948762e-01 8.33704889e-01
5.52852392e-01 -8.13874781e-01 8.88171613e-01 -7.72823468e-02
8.48894492e-02 -8.04103553e-01 -6.36231899e-01 -1.71120489e+00
-5.15603960e-01 -3.71252686e-01 -9.61071253e-02 9.59717155e-01
4.27191615e-01 -9.29917097e-01 6.16026640e-01 2.17190851e-02
-3.80320430e-01 -4.48165685e-01 -7.98693061e-01 -1.02110994e+00
-5.06758690e-01 -1.88844025e-01 7.47922957e-01 9.36493933e-01
-2.88785845e-01 4.26008910e-01 -4.55102444e-01 1.04498827e+00
8.83835256e-01 3.66297692e-01 8.91956151e-01 -1.41822398e+00
-4.52083558e-01 9.90537554e-02 -6.35530889e-01 -1.24750149e+00
1.46173894e-01 -8.56382549e-01 4.24919724e-01 -1.37379241e+00
-2.43179694e-01 -4.54592735e-01 -2.89293438e-01 3.92309248e-01
8.84845033e-02 5.06372452e-02 3.24921787e-01 2.49058023e-01
-6.38568819e-01 8.06873977e-01 1.02862561e+00 -3.37586790e-01
-6.28717244e-02 1.41162291e-01 -2.41631299e-01 8.48969519e-01
9.84120905e-01 -5.34857273e-01 -3.50313932e-01 -5.36423862e-01
3.16562980e-01 8.27688798e-02 6.11799300e-01 -1.66389775e+00
7.22481012e-01 1.27337337e-01 4.15541172e-01 -8.04164588e-01
4.63589966e-01 -1.35320675e+00 4.79449838e-01 5.95410526e-01
-6.26396611e-02 3.35678279e-01 1.15672462e-01 6.03161871e-01
-5.80698311e-01 -1.56657919e-02 6.93329990e-01 1.21188343e-01
-1.04182160e+00 5.48381329e-01 2.65819114e-02 -1.62015349e-01
8.89172375e-01 1.20472230e-01 -4.65962768e-01 -3.14710855e-01
-3.45038384e-01 5.15803158e-01 4.63593602e-01 2.74275333e-01
5.89900196e-01 -1.49925697e+00 -5.10594249e-01 3.33096296e-01
3.84962320e-01 3.82193267e-01 1.67194277e-01 1.23225021e+00
-4.81209308e-01 3.76119554e-01 -2.03407854e-01 -8.84302318e-01
-9.46813703e-01 3.21574330e-01 5.88780224e-01 -5.13179041e-03
-5.91177642e-01 7.04949498e-01 -1.43676609e-01 -7.74348438e-01
2.95147926e-01 -5.82764983e-01 4.63899411e-02 -1.71627104e-02
3.78342003e-01 3.65111172e-01 6.67909831e-02 -5.97702682e-01
-6.29206121e-01 9.76215661e-01 4.36037719e-01 1.43628065e-02
1.38894820e+00 -2.64168471e-01 6.44939626e-03 4.83332843e-01
1.13756144e+00 -5.81643544e-02 -1.50857604e+00 -5.46311319e-01
-2.15442963e-02 -3.59515607e-01 2.08104104e-01 -5.13754129e-01
-1.30706131e+00 7.64801621e-01 1.08424687e+00 -2.69977022e-02
1.17132103e+00 -1.61510125e-01 6.10888898e-01 9.77487981e-01
7.16308892e-01 -9.33190584e-01 -6.03961088e-02 6.99908793e-01
7.90471852e-01 -1.57512558e+00 1.29081428e-01 -3.48554671e-01
-3.65309983e-01 1.05422330e+00 6.58941269e-01 -4.30465668e-01
6.76687658e-01 3.29696238e-01 2.43317917e-01 -2.98498660e-01
-5.00164092e-01 -3.93882811e-01 3.67863238e-01 6.05350733e-01
-3.18928733e-02 4.71259803e-02 3.18349034e-01 3.23939115e-01
-2.09067971e-01 -2.16380820e-01 7.42824376e-02 9.25941408e-01
-3.27815980e-01 -8.19596469e-01 -4.27447796e-01 5.04918545e-02
1.07471131e-01 4.15108725e-02 -1.88534603e-01 8.83447826e-01
1.99569523e-01 7.68660665e-01 -1.41396765e-02 -8.03016722e-01
6.18348718e-01 -3.39758158e-01 4.24514204e-01 -3.58598292e-01
-2.76517898e-01 -1.21560343e-01 5.19037321e-02 -1.04757559e+00
-7.16977000e-01 -6.75783157e-01 -1.17728198e+00 -1.45242631e-01
-3.29421490e-01 1.01628974e-01 8.09235990e-01 8.00460517e-01
5.73674619e-01 7.06068814e-01 7.97239900e-01 -1.21362603e+00
-7.76553154e-01 -9.24515069e-01 -6.46749914e-01 1.92200448e-02
7.84950018e-01 -8.15063715e-01 -4.55082476e-01 -1.89381316e-01] | [7.527985095977783, -2.0464086532592773] |
f6e2f409-ab44-435b-9bb8-cccdc74f671c | towards-unified-human-parsing-and-pose | null | null | http://openaccess.thecvf.com/content_cvpr_2014/html/Dong_Towards_Unified_Human_2014_CVPR_paper.html | http://openaccess.thecvf.com/content_cvpr_2014/papers/Dong_Towards_Unified_Human_2014_CVPR_paper.pdf | Towards Unified Human Parsing and Pose Estimation | We study the problem of human body configuration analysis, more specifically, human parsing and human pose estimation. These two tasks, i.e. identifying the semantic regions and body joints respectively over the human body image, are intrinsically highly correlated. However, previous works generally solve these two problems separately or iteratively. In this work, we propose a unified framework for simultaneous human parsing and pose estimation based on semantic parts. By utilizing Parselets and Mixture of Joint-Group Templates as the representations for these semantic parts, we seamlessly formulate the human parsing and pose estimation problem jointly within a unified framework via a tailored And-Or graph. A novel Grid Layout Feature is then designed to effectively capture the spatial co-occurrence/occlusion information between/within the Parselets and MJGTs. Thus the mutually complementary nature of these two tasks can be harnessed to boost the performance of each other.The resultant unified model can be solved using the structure learning framework in a principled way. Comprehensive evaluations on two benchmark datasets for both tasks demonstrate the effectiveness of the proposed framework when compared with the state-of-the-art methods. | ['Xiaohui Shen', 'Qiang Chen', 'Jianchao Yang', 'Jian Dong', 'Shuicheng Yan'] | 2014-06-01 | null | null | null | cvpr-2014-6 | ['human-parsing'] | ['computer-vision'] | [ 2.40254402e-01 8.18986371e-02 -2.86712013e-02 -2.43882760e-01
-7.50782132e-01 -4.51380581e-01 2.83025950e-01 -1.76085830e-01
-1.86873987e-01 3.30489010e-01 2.02605218e-01 3.57493073e-01
-7.99051449e-02 -5.44588208e-01 -5.58356166e-01 -6.77182496e-01
1.12157010e-01 5.70558131e-01 3.16922814e-01 -1.68982267e-01
-8.56210366e-02 5.74238971e-02 -1.53769100e+00 -2.96923369e-01
8.63557637e-01 9.91373539e-01 2.92397827e-01 4.29267228e-01
1.64563537e-01 4.94880348e-01 -2.08231926e-01 -4.21317905e-01
2.58667737e-01 -3.99511725e-01 -6.58598900e-01 4.99166280e-01
4.18005794e-01 -1.72318984e-02 -1.70738831e-01 1.00651073e+00
7.02265203e-01 1.02691911e-01 5.09232163e-01 -1.00187063e+00
-1.05938062e-01 1.07933745e-01 -1.06588745e+00 -2.11464062e-01
5.73116899e-01 -7.15213045e-02 1.07638013e+00 -6.91111565e-01
6.20847881e-01 1.30152321e+00 5.81242859e-01 3.25436890e-01
-9.84856129e-01 -4.80685174e-01 3.83231580e-01 -4.26863990e-04
-1.27804482e+00 -2.59919614e-01 1.24632537e+00 -5.90054214e-01
3.69016111e-01 2.14657411e-01 8.61318588e-01 7.48726189e-01
9.65932757e-02 9.95307744e-01 9.17479336e-01 -3.95657450e-01
-3.48098315e-02 -3.35172713e-01 1.22489177e-01 1.18628311e+00
3.33369583e-01 -1.18840232e-01 -6.71876609e-01 -7.70016089e-02
9.70945776e-01 -1.04675011e-03 -3.31046730e-02 -9.68241811e-01
-1.17558074e+00 7.91876793e-01 3.60219151e-01 -3.82877775e-02
-3.80485058e-01 1.18012093e-01 4.87016857e-01 -4.64943737e-01
2.38601670e-01 7.80794844e-02 -4.08522069e-01 1.79229125e-01
-6.54317737e-01 5.96521497e-01 6.63457870e-01 1.05375743e+00
5.53033888e-01 -2.22034246e-01 -2.19666749e-01 8.91716778e-01
6.25947297e-01 4.97568816e-01 3.24680917e-02 -7.88750947e-01
7.21381843e-01 8.12546790e-01 5.55300899e-02 -1.31498909e+00
-7.93004096e-01 -6.31633222e-01 -7.34338701e-01 -2.66554624e-01
5.88303387e-01 -2.26588458e-01 -8.24641407e-01 1.92144275e+00
1.03260410e+00 1.04027102e-02 -2.04392299e-01 1.12946033e+00
8.39036882e-01 2.48452634e-01 3.20500910e-01 8.62239972e-02
2.06010938e+00 -1.18171227e+00 -6.83259130e-01 -5.85597396e-01
3.41588289e-01 -7.69297898e-01 7.25223958e-01 1.58647224e-01
-1.10892224e+00 -7.01646149e-01 -1.18013024e+00 -2.13225886e-01
1.53820012e-02 5.66176891e-01 7.20356643e-01 4.92184132e-01
-4.43315655e-01 3.39684874e-01 -1.05171120e+00 -3.01275432e-01
1.19442582e-01 4.36212689e-01 -3.77041072e-01 1.09240495e-01
-1.03049779e+00 5.39226055e-01 2.66984135e-01 5.26246905e-01
-3.33214849e-01 -1.48613766e-01 -1.16774726e+00 -2.49431103e-01
7.41367221e-01 -1.22675622e+00 1.10524297e+00 -5.09016871e-01
-1.44002640e+00 9.97620285e-01 -6.36741668e-02 -2.59335560e-04
5.22680819e-01 -7.27832913e-01 3.76746617e-02 1.91442013e-01
4.32779521e-01 4.99217659e-01 8.07964444e-01 -1.27877688e+00
-6.86929762e-01 -1.06795824e+00 -2.12133810e-01 7.15853095e-01
-5.17091118e-02 1.75312627e-02 -1.26009536e+00 -7.88984418e-01
5.97625792e-01 -1.08046508e+00 -3.97300214e-01 -4.94542681e-02
-6.70773089e-01 -2.36273661e-01 5.36123633e-01 -1.11680782e+00
1.31778657e+00 -1.71167779e+00 8.62120807e-01 3.13908368e-01
2.82464415e-01 -9.57620740e-02 1.93202049e-01 1.57954961e-01
2.77315468e-01 -3.87622893e-01 -2.06436768e-01 -5.07432520e-01
1.00795515e-01 1.75341502e-01 2.48740122e-01 7.45839775e-01
3.80010484e-03 9.73483920e-01 -8.41147125e-01 -9.24524307e-01
3.72318178e-01 4.41722721e-01 -3.84369284e-01 4.53846782e-01
5.21085784e-02 8.60068440e-01 -9.40268040e-01 8.03283393e-01
5.16354799e-01 -1.49199992e-01 5.40017068e-01 -3.64900738e-01
3.02719593e-01 -8.70039612e-02 -1.37395418e+00 2.17991567e+00
-2.13761106e-01 -2.58246243e-01 3.09720963e-01 -1.00300086e+00
8.43382716e-01 3.01811278e-01 7.42479444e-01 -4.81396914e-01
2.06483811e-01 1.71599705e-02 -2.36643344e-01 -5.57728231e-01
1.47240192e-01 -4.00602743e-02 -4.67209369e-01 1.51739269e-01
2.06289455e-01 2.57061034e-01 6.98697716e-02 -2.60661334e-01
6.55167162e-01 7.03062475e-01 7.04658389e-01 -3.21985662e-01
6.39788628e-01 -1.82739660e-01 8.40351999e-01 4.36649114e-01
-2.45368034e-01 5.57000339e-01 2.64637977e-01 -3.82429659e-01
-8.96387517e-01 -1.08418286e+00 2.05738351e-01 1.16281724e+00
5.35417020e-01 -5.27326167e-01 -1.12829280e+00 -7.07749367e-01
1.72739907e-03 -1.44516364e-01 -6.04647934e-01 1.31442487e-01
-1.04340243e+00 -6.75911367e-01 2.86797494e-01 8.80461276e-01
5.35973191e-01 -8.39268148e-01 -9.72826123e-01 2.51820952e-01
-3.80704492e-01 -1.46808624e+00 -5.19591451e-01 -3.99239874e-03
-7.27388859e-01 -1.18503606e+00 -9.27909017e-01 -9.84644234e-01
4.95469093e-01 9.30179730e-02 1.09826505e+00 1.61175936e-01
-5.46572328e-01 4.55184668e-01 -3.07820350e-01 7.66895860e-02
2.33466864e-01 1.74204901e-01 2.24947603e-03 1.01025619e-01
-1.63368315e-01 -6.52698576e-01 -1.01414216e+00 4.67913657e-01
-5.19208968e-01 4.55093592e-01 5.62279105e-01 8.19947362e-01
7.26942182e-01 2.06263247e-03 2.34696165e-01 -7.91385114e-01
1.59480378e-01 -2.05130130e-01 -3.67164046e-01 4.76968259e-01
-6.94703236e-02 1.24434926e-01 3.35144669e-01 -1.47011295e-01
-1.15963972e+00 6.35175169e-01 -1.76116586e-01 -1.37289003e-01
-2.50412703e-01 3.74821723e-01 -8.89420867e-01 -7.68365860e-02
1.86668783e-02 1.95470646e-01 -5.95731549e-02 -6.93171084e-01
5.60994446e-01 2.35170051e-01 8.63169968e-01 -9.25500751e-01
8.14196408e-01 5.15145123e-01 3.03528696e-01 -6.79578781e-01
-7.55971491e-01 -8.02696705e-01 -1.05831742e+00 -2.52464443e-01
1.30636811e+00 -1.04998910e+00 -6.76554024e-01 5.89413285e-01
-1.11153424e+00 1.57056123e-01 1.99023396e-01 2.51031876e-01
-7.49598026e-01 8.01794589e-01 -4.96023089e-01 -9.11532402e-01
-4.89960879e-01 -1.30386126e+00 1.78602946e+00 2.33149767e-01
-2.92390645e-01 -9.31945503e-01 3.24222036e-02 7.44136512e-01
-3.88433248e-01 6.92769885e-01 6.94438636e-01 -4.34069484e-01
-3.29370826e-01 -3.03296715e-01 -7.25116730e-02 -6.94086477e-02
5.28061800e-02 -3.96958530e-01 -6.80795670e-01 -2.10461825e-01
-9.78583768e-02 -1.74784586e-01 5.88668942e-01 4.28312778e-01
9.11965370e-01 -1.34649530e-01 -5.08178771e-01 6.20762169e-01
1.17057633e+00 -1.71562836e-01 2.98224390e-01 2.17316791e-01
1.16724038e+00 1.00096262e+00 9.32303727e-01 7.09060788e-01
6.62042201e-01 1.25094235e+00 3.10537130e-01 -2.37308159e-01
-7.33529851e-02 -5.20850480e-01 3.42923366e-02 8.50231051e-01
-3.88612807e-01 7.82299191e-02 -9.84558523e-01 1.08396612e-01
-2.20040011e+00 -4.72468644e-01 -7.97439441e-02 2.03784418e+00
5.42664349e-01 -8.11291784e-02 4.31243002e-01 -3.65491770e-02
8.27749848e-01 2.51550943e-01 -4.74161416e-01 4.09688711e-01
2.45093003e-01 1.55605793e-01 3.75022620e-01 1.01423867e-01
-1.59780467e+00 9.24830317e-01 5.52615356e+00 8.38612080e-01
-6.28604412e-01 -5.16073890e-02 3.12841833e-01 1.91295937e-01
1.27850875e-01 -4.93410528e-02 -8.36670756e-01 2.24467650e-01
2.12413475e-01 3.09594214e-01 1.99574605e-01 8.24829102e-01
-3.57406476e-04 -3.29250880e-02 -9.44288671e-01 1.17764986e+00
7.43981376e-02 -6.73888862e-01 -1.16525523e-01 1.36094019e-01
4.67746288e-01 -7.18137085e-01 6.43266737e-03 3.26281153e-02
-4.72399257e-02 -7.50571668e-01 1.06206787e+00 4.69275504e-01
4.68827188e-01 -6.81575835e-01 5.82287490e-01 4.42538261e-01
-1.92618918e+00 1.87891312e-02 -6.01546001e-03 3.50266173e-02
4.02159482e-01 4.27860200e-01 -1.24817140e-01 1.14697242e+00
6.36746705e-01 6.61301315e-01 -5.88026822e-01 7.90588737e-01
-4.21910435e-01 2.08485231e-01 -2.93840349e-01 3.66735786e-01
-1.61408350e-01 -3.97733480e-01 5.29968381e-01 8.80345643e-01
6.20684363e-02 1.01893462e-01 7.39028037e-01 6.69596136e-01
2.91415602e-01 3.87086481e-01 -1.84619218e-01 2.27866441e-01
2.28140727e-01 1.52835047e+00 -9.69195485e-01 -2.02447817e-01
-4.03935969e-01 1.05260098e+00 4.71522629e-01 2.18719572e-01
-1.06783283e+00 3.80837657e-02 5.58900714e-01 7.44468495e-02
3.23379427e-01 -3.21419477e-01 -4.04603988e-01 -1.27132881e+00
4.44406480e-01 -8.70038986e-01 5.15638888e-01 -3.82042617e-01
-1.15789998e+00 3.84675086e-01 9.59688798e-02 -1.13973641e+00
-2.88694710e-01 -6.20101154e-01 -3.07284355e-01 7.60184586e-01
-1.09874403e+00 -1.70324790e+00 -4.14224237e-01 6.21170580e-01
5.07475197e-01 2.98529565e-01 7.37648308e-01 1.59651205e-01
-8.57152820e-01 4.83746201e-01 -3.96889657e-01 3.49876136e-01
4.19674605e-01 -1.34592319e+00 2.83338219e-01 8.51826370e-01
1.19809933e-01 5.80504417e-01 5.94458580e-01 -9.46977258e-01
-1.53617489e+00 -9.07559395e-01 4.20289338e-01 -3.52107674e-01
2.91012257e-01 -6.51705205e-01 -5.37439287e-01 4.17696804e-01
-1.11663640e-01 -2.11010695e-01 5.52011013e-01 2.30470985e-01
-2.65094101e-01 1.14055254e-01 -7.44477630e-01 4.52968597e-01
1.36167789e+00 -2.35331371e-01 -6.09832108e-01 3.44210148e-01
5.22716165e-01 -7.08058357e-01 -9.79673088e-01 7.77875066e-01
7.84871638e-01 -9.12970841e-01 1.31893981e+00 -5.18209994e-01
3.11342388e-01 -4.69359338e-01 -2.10253105e-01 -7.29357898e-01
-2.65355140e-01 -5.64265013e-01 -2.73863018e-01 1.18218172e+00
-1.23271517e-01 -2.62121677e-01 9.00954187e-01 3.93682629e-01
3.99187580e-02 -9.15733755e-01 -8.94269884e-01 -6.01313412e-01
-2.66741246e-01 -1.27365261e-01 3.36011112e-01 4.89232332e-01
-6.47937283e-02 6.29025638e-01 -7.12494552e-01 3.33610237e-01
1.06161225e+00 3.92102510e-01 9.88057435e-01 -1.19562018e+00
-5.52928388e-01 -2.55285561e-01 -6.38820052e-01 -1.31530035e+00
1.00894414e-01 -5.41593850e-01 3.55200082e-01 -1.51596010e+00
5.02220571e-01 -2.55174011e-01 6.49505854e-02 4.62048471e-01
-6.79982305e-01 5.60324192e-02 1.58558846e-01 2.82409340e-01
-6.16019309e-01 6.28214419e-01 1.31884420e+00 1.99761540e-01
-7.81590268e-02 3.45855169e-02 -6.15316808e-01 1.02016699e+00
5.15644431e-01 -1.50890380e-01 -3.67375374e-01 -2.60556579e-01
-6.11117668e-02 3.64413947e-01 6.08649194e-01 -1.23800766e+00
1.79936036e-01 2.46036053e-02 4.34788644e-01 -6.19009018e-01
4.71042365e-01 -7.14951873e-01 2.67613143e-01 3.65668714e-01
-8.15506801e-02 8.22925754e-03 -5.76859675e-02 8.76173496e-01
-1.90118924e-01 1.87811375e-01 6.47438526e-01 -3.72045755e-01
-7.80268371e-01 4.38309819e-01 2.99444914e-01 1.29602417e-01
1.06190872e+00 -2.15210676e-01 2.43991628e-01 -2.33061358e-01
-1.07807100e+00 3.26963574e-01 3.24376166e-01 4.83935416e-01
5.39940953e-01 -1.31515014e+00 -4.52342570e-01 1.05594002e-01
1.00390211e-01 2.52358854e-01 6.15559459e-01 1.09118462e+00
-3.30168664e-01 3.57704133e-01 -2.07254723e-01 -7.51121521e-01
-1.49204373e+00 5.36792397e-01 2.77288824e-01 -7.62269139e-01
-6.80109441e-01 7.06631124e-01 7.31889367e-01 -5.06144106e-01
2.94436514e-01 -2.65571307e-02 -3.20856184e-01 -9.25449803e-02
1.56548426e-01 3.56520593e-01 -1.88369319e-01 -1.12507308e+00
-5.60236096e-01 1.21173370e+00 2.55511492e-01 2.34695934e-02
1.16349053e+00 -4.11972433e-01 -9.90743935e-03 1.18433677e-01
1.11224663e+00 -1.20447673e-01 -1.22458613e+00 -2.44088799e-01
4.61303182e-02 -2.68000662e-01 -3.62077445e-01 -4.80990469e-01
-1.17581093e+00 9.53472257e-01 4.98671204e-01 -2.41791040e-01
1.18369079e+00 3.27845246e-01 8.99407506e-01 -2.14966908e-01
4.58870739e-01 -1.20024586e+00 1.64132327e-01 4.66261171e-02
7.69519389e-01 -9.90555465e-01 2.19513074e-01 -1.22123277e+00
-6.72835886e-01 1.00966418e+00 7.11479187e-01 -1.92387477e-01
4.16858256e-01 -6.68126047e-02 -1.13116547e-01 -4.68086064e-01
-1.32871315e-01 -4.83744383e-01 9.37599301e-01 4.86917764e-01
5.23361206e-01 2.64633834e-01 -4.79226023e-01 1.08193374e+00
-1.81540146e-01 -2.62242764e-01 -4.57425416e-01 1.12648380e+00
-3.64140600e-01 -1.14753151e+00 -6.11135185e-01 5.41500114e-02
-3.42810154e-01 3.49562585e-01 -3.49753529e-01 8.03034365e-01
2.59279728e-01 7.86024213e-01 -3.78167421e-01 -5.08512974e-01
5.47169566e-01 -1.01214489e-02 7.62109816e-01 -5.30543983e-01
-3.22823375e-01 6.08049810e-01 1.78947479e-01 -9.08942580e-01
-5.73792398e-01 -7.99548447e-01 -1.28285193e+00 2.05436513e-01
-3.56662363e-01 -2.16648266e-01 3.88487965e-01 1.09264100e+00
1.77199975e-01 5.07928014e-01 2.51955539e-01 -1.03140271e+00
-7.22982466e-01 -7.14996994e-01 -8.39768529e-01 6.65185392e-01
-2.20180601e-01 -1.25787950e+00 1.22085527e-01 9.06321257e-02] | [7.564724445343018, -0.4780471622943878] |
7a09f85d-50b5-48c7-ac70-d19033d1ed90 | extractive-and-abstractive-summarization | null | null | https://aclanthology.org/2022.fnp-1.8 | https://aclanthology.org/2022.fnp-1.8.pdf | Extractive and Abstractive Summarization Methods for Financial Narrative Summarization in English, Spanish and Greek | This paper describes the three summarization systems submitted to the Financial Narrative Summarization Shared Task (FNS-2022). We developed a task-specific extractive summarization method for the reports in English. It was based on a sequence classification task whose objective was to find the sentence where the summary begins. On the other hand, since the summaries for the reports in Spanish and Greek were not extractive, we used an abstractive strategy for each of the languages. In particular, we created a new Encoder-Decoder architecture in Spanish, MariMari, based on an existing Encoding-only model; we also trained multilingual Encoder-Decoder models for this task. Finally, the summaries for the reports in Greek were obtained with a translation-summary-translation system in which the reports were translated to English and summarised, and then the summaries were translated back to Greek. | ['Álvaro Barbero Jiménez', 'David Betancur', 'Alba Segurado', 'Alejandro Vaca'] | null | null | null | null | fnp-lrec-2022-6 | ['extractive-summarization'] | ['natural-language-processing'] | [ 4.23252672e-01 6.16621017e-01 -1.53028473e-01 -8.92684385e-02
-1.49525630e+00 -6.51630938e-01 8.60231221e-01 3.58707458e-01
-3.47729772e-01 1.34042025e+00 1.24376762e+00 -1.16293430e-01
3.61422807e-01 -2.96849996e-01 -5.55358887e-01 -3.08899581e-01
2.16353759e-01 2.53838271e-01 -8.93604606e-02 -2.52606362e-01
6.75674200e-01 5.51155247e-02 -8.29840899e-01 8.74974668e-01
1.00647581e+00 4.86712456e-01 2.19552130e-01 1.20527875e+00
-7.01434910e-02 1.27404892e+00 -1.17653513e+00 -8.52609277e-01
-2.27152154e-01 -1.23313475e+00 -9.14582491e-01 2.89603770e-02
1.63116470e-01 -3.47038478e-01 -4.76921767e-01 7.93599069e-01
4.73225027e-01 1.12639684e-02 1.02878892e+00 -6.10417783e-01
-8.68245065e-01 1.18089139e+00 -2.68522292e-01 2.59961635e-01
8.71966779e-01 -1.93656921e-01 1.26471519e+00 -8.87746274e-01
9.77282524e-01 9.65364635e-01 5.07133901e-01 5.93492866e-01
-8.16548765e-01 -1.75890148e-01 -1.55408651e-01 2.54018873e-01
-8.38180006e-01 -7.97701418e-01 7.97000170e-01 -3.75983596e-01
1.60507727e+00 4.98211861e-01 7.75513053e-01 1.42882442e+00
7.52328873e-01 8.93087924e-01 5.03938377e-01 -4.18565363e-01
1.03665948e-01 1.02792002e-01 1.98138822e-02 3.65712672e-01
3.64463031e-01 -4.51918632e-01 -8.70514631e-01 -1.51317893e-02
-4.68304977e-02 -7.22187400e-01 -2.41058439e-01 6.46696091e-01
-1.43358111e+00 8.02812040e-01 -2.24926218e-01 3.67283255e-01
-6.71218157e-01 -1.60852239e-01 1.10031772e+00 4.54727620e-01
9.10516739e-01 6.63914740e-01 -2.33367741e-01 -2.43807942e-01
-1.53048790e+00 5.12062192e-01 1.33211911e+00 9.93085861e-01
2.44774356e-01 2.16574252e-01 -6.20098293e-01 6.48897171e-01
-9.13185850e-02 3.76631171e-01 7.44154334e-01 -8.02124381e-01
1.30901492e+00 3.01014215e-01 1.09811090e-01 -1.01099932e+00
-2.95369983e-01 -3.35701913e-01 -8.07736218e-01 -4.99885052e-01
-2.61258870e-01 -6.22982085e-01 -4.80390459e-01 1.15143955e+00
-4.94566441e-01 -3.03426713e-01 1.04178965e+00 5.74516356e-01
1.32850957e+00 1.10221994e+00 -2.27122605e-01 -7.70926535e-01
1.20805120e+00 -1.32247615e+00 -1.09385026e+00 -3.54443163e-01
6.88963294e-01 -8.79186451e-01 4.48631793e-01 4.78278883e-02
-1.54149961e+00 -3.60936970e-01 -1.61136663e+00 -4.43943530e-01
-1.19849466e-01 5.14291704e-01 -1.24600366e-01 4.88369912e-02
-1.11302745e+00 6.94386542e-01 -7.53385007e-01 -5.63351870e-01
7.02616945e-02 -2.33707443e-01 -3.18344206e-01 4.53544617e-01
-1.16614485e+00 1.20074260e+00 9.37143326e-01 -1.35233879e-01
-6.04255021e-01 -3.76390189e-01 -1.04492688e+00 1.45018339e-01
1.49881154e-01 -8.33955705e-01 1.59568095e+00 -8.10049593e-01
-1.82461965e+00 8.14347208e-01 -2.74648488e-01 -7.24193692e-01
4.59720165e-01 -2.29035050e-01 -4.68675017e-01 4.56071943e-01
6.73186123e-01 2.74005979e-01 5.13951123e-01 -8.47128451e-01
-6.31826818e-01 -1.71525553e-02 -2.89308667e-01 4.62400317e-01
-8.89263153e-02 4.97637659e-01 -3.71197462e-02 -9.81322825e-01
-2.45348737e-01 -3.93341333e-01 -1.08274249e-02 -1.11608756e+00
-9.11200523e-01 -8.96622464e-02 4.34881896e-01 -1.74446154e+00
1.68249750e+00 -1.92775106e+00 6.15564346e-01 -4.01567101e-01
-1.44798756e-01 1.07963584e-01 -1.52771324e-01 1.10838783e+00
-1.28597142e-02 2.89251357e-01 -6.72016859e-01 -6.52518272e-01
-9.43211988e-02 -1.68726578e-01 -5.66727161e-01 4.28534672e-02
6.35609090e-01 1.00430727e+00 -9.91014123e-01 -6.80864394e-01
-4.32455987e-01 -1.21414140e-01 -4.19306487e-01 3.89123052e-01
-4.03174669e-01 4.57394481e-01 -3.14892888e-01 2.66234398e-01
1.47707045e-01 2.37727210e-01 1.67782739e-01 1.41135409e-01
-4.68498915e-01 1.06185055e+00 -5.07247865e-01 1.84029889e+00
-3.68097037e-01 7.63374507e-01 -3.15219581e-01 -1.03058958e+00
1.06618392e+00 8.33290100e-01 3.18059415e-01 -5.69339216e-01
-7.13838860e-02 6.37047589e-01 -3.33135456e-01 -7.20945656e-01
1.29050529e+00 -2.29458675e-01 -6.56275809e-01 7.90811539e-01
2.71815091e-01 -6.20742023e-01 8.24215710e-01 5.35021245e-01
1.14183843e+00 1.98163033e-01 9.45446730e-01 1.85801461e-01
6.38397932e-01 5.78859568e-01 4.48202431e-01 4.52758193e-01
3.97607297e-01 1.10219383e+00 9.41794038e-01 -8.03408474e-02
-1.23044693e+00 -7.96949387e-01 3.52171689e-01 4.72523421e-01
-6.61084950e-01 -1.07377625e+00 -9.00860727e-01 -7.75123954e-01
-5.42445362e-01 1.42101526e+00 -3.95972341e-01 -1.17289260e-01
-7.72103727e-01 -4.47389603e-01 8.16826403e-01 3.38563442e-01
5.56823134e-01 -1.33488202e+00 -6.58545554e-01 7.36858010e-01
-7.63684988e-01 -1.22807264e+00 -7.70131707e-01 9.14768055e-02
-6.91395819e-01 -8.92781317e-01 -1.04995513e+00 -8.51103008e-01
7.71963522e-02 -3.39103758e-01 9.24717963e-01 -3.92817765e-01
3.59381467e-01 3.19485962e-02 -5.17622709e-01 -7.46532500e-01
-1.19658709e+00 6.87510729e-01 -2.60493308e-01 -3.47406894e-01
3.63462209e-03 -3.01919848e-01 2.08112761e-01 -7.44136274e-01
-7.81781375e-01 3.59127998e-01 5.36274552e-01 7.01578617e-01
1.52753815e-01 -3.66879165e-01 1.31815159e+00 -8.04518640e-01
1.24914193e+00 -4.13467765e-01 1.06563307e-01 3.82024258e-01
8.62668827e-02 4.28795479e-02 9.00998116e-01 -1.32773742e-01
-1.30092514e+00 -2.24627510e-01 -3.20801169e-01 3.20383489e-01
2.95902699e-01 1.07726336e+00 -1.25657991e-01 9.61313248e-01
5.82417727e-01 5.41532457e-01 -2.11535677e-01 -4.18665409e-01
2.71600217e-01 1.14869452e+00 7.84810960e-01 -1.28241912e-01
3.20840269e-01 -2.49503478e-01 -4.96694744e-01 -1.09939480e+00
-1.06917048e+00 -9.36717093e-02 -6.81260824e-01 -6.52824342e-02
1.14683676e+00 -1.05412149e+00 1.23216562e-01 3.82858634e-01
-1.95271719e+00 -5.39233908e-02 -6.16180182e-01 5.54831564e-01
-8.81844103e-01 5.90927005e-01 -7.68448651e-01 -5.74509382e-01
-9.02511656e-01 -7.74839997e-01 8.80079985e-01 2.34866723e-01
-7.80440509e-01 -9.77810800e-01 4.25702572e-01 4.00353707e-02
1.01573303e-01 3.55164886e-01 9.75069046e-01 -1.05373991e+00
2.49514908e-01 -8.54911506e-02 -1.17934160e-01 5.58091104e-01
1.99238598e-01 2.37716213e-02 -4.61908937e-01 -8.29492882e-03
3.54226500e-01 -2.44900703e-01 1.05660427e+00 1.20495848e-01
2.91007638e-01 -1.04088163e+00 1.28136007e-02 2.75772005e-01
9.18365121e-01 2.19707400e-01 4.78136361e-01 5.33248901e-01
4.89208341e-01 5.01574337e-01 4.32238936e-01 3.79154295e-01
6.82495236e-01 4.01068419e-01 -3.35642636e-01 3.49427164e-01
-4.85629857e-01 -4.64974880e-01 1.12764215e+00 1.73683500e+00
-4.10264470e-02 -4.96088892e-01 -5.56586504e-01 8.37511599e-01
-1.98147726e+00 -1.11298168e+00 -1.33004874e-01 1.57222986e+00
1.08910191e+00 8.16372558e-02 3.40987779e-02 -1.37862533e-01
3.93282890e-01 5.82766294e-01 -2.26486668e-01 -1.14552319e+00
-4.35499549e-01 9.39165279e-02 1.78679332e-01 3.40954125e-01
-7.82715082e-01 8.90576422e-01 6.63676310e+00 4.94948000e-01
-9.29274499e-01 1.57320723e-01 2.70828217e-01 -1.91365018e-01
-2.74165064e-01 -7.25439191e-03 -6.57797873e-01 4.62081641e-01
1.53193164e+00 -8.87792230e-01 -1.14288680e-01 4.01955783e-01
3.21035892e-01 -2.24253282e-01 -1.15620828e+00 4.36040431e-01
6.90291345e-01 -1.60598993e+00 3.61361295e-01 -2.95903027e-01
1.03135729e+00 -4.85426337e-02 -6.23240411e-01 3.57532412e-01
9.79890525e-02 -6.75144613e-01 1.19135737e+00 7.40492642e-01
8.89762104e-01 -8.24571788e-01 1.06887197e+00 5.14657736e-01
-7.68010378e-01 1.93578243e-01 -4.24448699e-01 -1.24405235e-01
5.80051541e-01 3.50718141e-01 -9.77682710e-01 1.35378838e+00
7.49978274e-02 1.22168291e+00 -5.57031870e-01 5.80037653e-01
-7.44474947e-01 7.63143599e-01 2.77102977e-01 -1.92403838e-01
3.66223246e-01 -2.94972509e-01 1.13794911e+00 1.83572876e+00
6.66669130e-01 5.73047027e-02 2.58897468e-02 5.41427314e-01
-3.32085103e-01 2.93062568e-01 -7.87238538e-01 -3.83250594e-01
6.84981346e-02 8.39841843e-01 -4.03592974e-01 -7.37621486e-01
-3.06488544e-01 1.31008804e+00 3.39242399e-01 2.98574567e-01
-6.12773359e-01 -7.94967830e-01 -1.76092282e-01 -2.16835484e-01
3.04887652e-01 -2.38042057e-01 -4.93626684e-01 -1.60149837e+00
2.40992725e-01 -7.83811629e-01 2.71200746e-01 -8.64748418e-01
-1.02500451e+00 9.01834488e-01 1.11149006e-01 -1.02573788e+00
-8.92346323e-01 9.74567607e-02 -9.28980231e-01 8.03417206e-01
-1.22160602e+00 -1.12262619e+00 4.90331829e-01 -8.42528790e-02
9.77011263e-01 -6.60679758e-01 9.04127538e-01 -3.79610918e-02
-5.94694138e-01 2.08663374e-01 2.24174604e-01 2.62778133e-01
8.20396960e-01 -1.19062817e+00 7.39727736e-01 1.14850879e+00
-2.40217559e-02 2.12691247e-01 8.73755872e-01 -1.03018332e+00
-8.09641242e-01 -1.25074434e+00 1.95161068e+00 -1.74261153e-01
8.77116919e-01 -1.99689567e-01 -7.56824851e-01 9.89508212e-01
1.19760656e+00 -1.22183335e+00 5.24181128e-01 -4.48054761e-01
2.11586729e-01 3.99007976e-01 -7.43414104e-01 7.11581647e-01
6.87907636e-01 -4.77604777e-01 -1.36873555e+00 4.72170234e-01
1.07399988e+00 -5.40503085e-01 -7.85259426e-01 -1.99467048e-01
2.08634123e-01 -3.93715769e-01 3.24595064e-01 -6.38584852e-01
1.33069313e+00 -1.77342128e-02 -6.57897666e-02 -1.85481226e+00
1.07727526e-02 -8.31845284e-01 -2.61730134e-01 1.59189224e+00
7.05516875e-01 -6.13241255e-01 1.27858534e-01 -1.46336854e-01
-7.57539213e-01 -3.91317189e-01 -1.03192532e+00 -6.47067308e-01
2.06435069e-01 -8.39962661e-02 4.25106645e-01 4.34018791e-01
5.90271115e-01 1.08257663e+00 -3.73903185e-01 -3.65110487e-01
1.99987158e-01 2.78829411e-02 4.64548796e-01 -6.32539332e-01
9.53407288e-02 -3.69180501e-01 9.40034613e-02 -7.40698576e-01
5.82132101e-01 -1.25707102e+00 4.54542786e-02 -2.25745916e+00
4.11827356e-01 6.63954079e-01 3.59858394e-01 3.61661792e-01
-2.60699272e-01 2.32601054e-02 4.79406327e-01 3.50416631e-01
-5.84366560e-01 7.60724008e-01 1.07070816e+00 -3.49043012e-01
-3.64106357e-01 1.49625391e-01 -1.11814582e+00 4.56571877e-01
8.65066528e-01 -6.95603728e-01 1.39727578e-01 -3.10948879e-01
2.67613262e-01 5.76178730e-01 -1.59342345e-02 -1.01798248e+00
3.61536354e-01 1.12232357e-01 1.27927542e-01 -1.13598454e+00
6.51843101e-02 2.65917238e-02 -1.03649236e-02 3.99623871e-01
-6.59117281e-01 2.21298262e-01 2.24537715e-01 1.36634246e-01
-5.77052712e-01 -6.86227024e-01 2.76423156e-01 -2.38502800e-01
-1.09591238e-01 -5.10640383e-01 -9.42684174e-01 2.33673871e-01
8.63812625e-01 -2.28400230e-02 -3.91478181e-01 -5.35432696e-01
-4.73576903e-01 1.17456481e-01 3.07011098e-01 1.99792042e-01
6.48250222e-01 -1.15864098e+00 -1.76357341e+00 -1.82644129e-01
-1.49332494e-01 -1.30289018e-01 -1.11873709e-02 7.95323372e-01
-7.07418740e-01 8.63253832e-01 -1.83739766e-01 4.37868871e-02
-9.96717274e-01 2.93781787e-01 -9.99162644e-02 -8.63927126e-01
-7.73938596e-01 1.32125631e-01 -2.61729360e-01 -1.89381257e-01
-2.40846887e-01 -3.41700226e-01 -5.59668183e-01 5.56035221e-01
8.25522363e-01 5.35844088e-01 3.93884391e-01 -6.99304044e-01
-1.63053051e-01 1.96863413e-01 -1.94196090e-01 -7.15334713e-01
1.67958939e+00 -2.30154559e-01 -3.04768354e-01 6.90886557e-01
1.21997058e+00 3.09166223e-01 -7.36698151e-01 3.45386155e-02
5.13662159e-01 5.18199205e-01 -3.71390700e-01 -9.14792120e-01
-4.74261969e-01 5.39108276e-01 -7.36458361e-01 3.02895039e-01
9.75425541e-01 -1.43479586e-01 9.72066104e-01 4.88775998e-01
-1.92079887e-01 -1.31611645e+00 -1.00565977e-01 1.17857707e+00
1.51952970e+00 -8.00537825e-01 3.74991447e-01 3.60445976e-02
-1.18392456e+00 1.50671732e+00 -4.69656959e-02 -1.99831054e-01
-7.56183863e-02 2.13311538e-01 -2.40797773e-01 -4.30357642e-02
-1.03536868e+00 2.28574842e-01 3.43905985e-01 4.64929879e-01
5.86239159e-01 2.79116742e-02 -8.66728723e-01 1.21967137e+00
-7.59027243e-01 3.66215594e-02 1.24190211e+00 8.33003819e-01
-2.62049884e-01 -7.85637319e-01 -3.27365160e-01 4.43522066e-01
-7.44379401e-01 -2.45309740e-01 -9.24639940e-01 5.77547848e-01
-4.57444638e-01 1.15087950e+00 -4.19797702e-03 -2.75077075e-01
6.14816964e-01 3.05813342e-01 2.15771526e-01 -1.13649142e+00
-1.15169930e+00 -2.62633245e-02 8.46533179e-01 9.10536572e-03
-3.46473008e-01 -1.04146349e+00 -1.16816151e+00 -1.13122433e-01
2.16389656e-01 5.17374277e-01 6.44170403e-01 1.00959623e+00
2.08320722e-01 1.02900934e+00 5.59592128e-01 -9.45999086e-01
-6.39474690e-01 -1.49837041e+00 -4.94198978e-01 -2.46184528e-01
3.99285436e-01 2.66006619e-01 -2.95113921e-01 4.24317569e-01] | [12.412590980529785, 9.541230201721191] |
c8aa0007-0922-4df1-ab0f-759ccdb9a6ce | ingeotec-at-semeval-2017-task-4-a-b4msa | null | null | https://aclanthology.org/S17-2130 | https://aclanthology.org/S17-2130.pdf | INGEOTEC at SemEval 2017 Task 4: A B4MSA Ensemble based on Genetic Programming for Twitter Sentiment Analysis | This paper describes the system used in SemEval-2017 Task 4 (Subtask A): Message Polarity Classification for both English and Arabic languages. Our proposed system is an ensemble of two layers, the first one uses our generic framework for multilingual polarity classification (B4MSA) and the second layer combines all the decision function values predicted by B4MSA systems using a non-linear function evolved using a Genetic Programming system, EvoDAG. With this approach, the best performances reached by our system were macro-recall 0.68 (English) and 0.477 (Arabic) which set us in sixth and fourth positions in the results table, respectively. | ["Sabino a-Jim{\\'e}nez", 'Mir', 'Eric Sadit Tellez', 'Mario Graff', 'Daniela Moctezuma'] | 2017-08-01 | null | null | null | semeval-2017-8 | ['twitter-sentiment-analysis'] | ['natural-language-processing'] | [-6.65482283e-02 4.51654434e-01 -2.08943471e-01 -3.75568956e-01
-8.46039534e-01 -7.27275908e-01 1.14569819e+00 5.12191117e-01
-4.40654755e-01 1.10492039e+00 1.81455538e-01 -4.33076859e-01
7.09979853e-04 -5.42485416e-01 -5.08141339e-01 -5.71168423e-01
5.07108271e-02 7.59970486e-01 3.00525159e-01 -9.89275515e-01
6.06556535e-01 2.31084712e-02 -1.30283451e+00 1.06774187e+00
1.09843504e+00 1.22122025e+00 -2.36411959e-01 7.50318527e-01
-1.50542527e-01 1.04843938e+00 -5.19731045e-01 -9.63629663e-01
-1.50296226e-01 -2.19604075e-01 -1.09349740e+00 -5.79730690e-01
1.74774881e-02 4.64145839e-01 3.97334278e-01 8.15653801e-01
3.46408397e-01 -4.67513651e-01 1.18290961e+00 -1.07394493e+00
-3.76578748e-01 1.04158568e+00 -2.87970543e-01 -3.34287845e-02
7.41388083e-01 -3.53037953e-01 1.11433792e+00 -8.99308741e-01
9.20408189e-01 1.36246610e+00 7.16119647e-01 1.07365929e-01
-7.76510358e-01 -3.52133274e-01 4.62530702e-02 7.44824827e-01
-9.86592412e-01 -2.93367475e-01 6.28903925e-01 -5.41712582e-01
1.23326755e+00 4.36207652e-01 8.45226228e-01 9.89174902e-01
7.09767163e-01 8.73040736e-01 1.97644842e+00 -4.68577057e-01
-5.99628873e-03 6.80698752e-01 5.55524051e-01 3.32154036e-01
1.03057839e-01 -3.87871176e-01 -5.79596043e-01 -3.45310122e-02
-6.22006297e-01 -1.01391602e+00 8.46306309e-02 1.53044164e-01
-9.91227984e-01 6.67515397e-01 2.22331151e-01 5.03792048e-01
-5.03769755e-01 -7.03743339e-01 6.33191288e-01 7.31456339e-01
5.73810399e-01 4.80556607e-01 -8.69160354e-01 7.73853511e-02
-7.07494438e-01 4.24957842e-01 1.27746415e+00 5.12331069e-01
4.09941852e-01 -2.46762365e-01 -3.21267277e-01 9.50951278e-01
5.46755075e-01 6.67883158e-01 7.68671930e-01 -2.87278652e-01
6.43862545e-01 8.38678181e-01 1.24770314e-01 -9.23273265e-01
-7.29787469e-01 -2.71573097e-01 -3.59876692e-01 -2.96870377e-02
4.19207871e-01 -5.07172287e-01 -8.84321034e-01 1.30124569e+00
-5.29149845e-02 -5.16252577e-01 9.17227507e-01 4.96890366e-01
1.30903077e+00 1.03579938e+00 2.04395786e-01 -4.20589596e-01
1.59328604e+00 -9.81518090e-01 -7.65148580e-01 -3.11107457e-01
6.03657484e-01 -1.14591348e+00 8.63782883e-01 7.81040549e-01
-1.08508885e+00 -7.77134523e-02 -1.49735761e+00 1.24402322e-01
-9.97691035e-01 2.37959549e-01 2.00419381e-01 8.12581062e-01
-1.27715206e+00 1.15640201e-01 5.31328060e-02 -2.42560893e-01
-1.19442552e-01 4.99085039e-01 -3.38476211e-01 3.65647197e-01
-1.72299409e+00 1.52534068e+00 5.81709921e-01 1.35206938e-01
-3.33968937e-01 4.72917408e-02 -7.01157868e-01 -1.04062140e-01
-3.72778205e-03 -9.97091159e-02 1.08093822e+00 -1.29741299e+00
-1.94262624e+00 1.10125172e+00 -6.78474829e-02 -7.02146471e-01
5.22486269e-01 -5.94936870e-03 -7.02099442e-01 -1.01308934e-01
-5.97051270e-02 4.69854921e-01 5.98935127e-01 -1.35049856e+00
-8.23843122e-01 -1.74572319e-01 -3.83492261e-02 4.74785835e-01
-1.23707950e-01 1.34979710e-01 -2.72266567e-01 -3.84509712e-01
-1.67042799e-02 -1.02010334e+00 2.96207279e-01 -1.25169516e+00
-2.39693075e-01 -4.34337318e-01 4.80409443e-01 -1.22369158e+00
1.32873583e+00 -1.61163294e+00 2.66213775e-01 6.78147435e-01
-3.98725748e-01 2.42481679e-01 -2.06591152e-02 3.27009678e-01
-1.97327107e-01 -7.75121376e-02 -3.93230826e-01 -4.68047112e-02
7.24554211e-02 -2.30223745e-01 -1.37221247e-01 2.09984332e-01
5.95160648e-02 6.55861318e-01 -7.35328436e-01 -5.16757488e-01
-2.36398250e-01 3.41306239e-01 -2.79287994e-01 -1.58268169e-01
-1.15749352e-01 -2.25882698e-02 -1.22335507e-02 8.55500519e-01
8.26966882e-01 2.53660649e-01 8.91807556e-01 -1.40801609e-01
-2.39810362e-01 4.82915789e-01 -1.07479393e+00 8.94505441e-01
-6.33471608e-01 5.80806434e-01 -1.06906272e-01 -7.34258533e-01
1.05189729e+00 3.95542234e-01 2.61502683e-01 -1.18396044e+00
4.41288829e-01 6.96645319e-01 2.85124362e-01 -4.99031514e-01
6.33068085e-01 9.20600444e-02 -4.86947209e-01 1.03961669e-01
1.78089708e-01 -1.78507328e-01 7.29570389e-01 1.65345341e-01
3.20039481e-01 -4.57832702e-02 4.23280835e-01 -6.48534358e-01
1.35602713e+00 3.26162577e-02 3.94950211e-01 3.23991150e-01
-7.49295950e-02 2.87652791e-01 1.33710241e+00 -2.27607638e-01
-9.19198573e-01 -5.54865956e-01 -3.98351699e-01 1.04829240e+00
-3.85101438e-02 -5.11053085e-01 -8.06636751e-01 -8.86694729e-01
-8.54552016e-02 7.19668806e-01 -7.27317810e-01 8.57317150e-02
-5.49184203e-01 -1.39861226e+00 4.35537577e-01 -7.01000914e-02
3.39027852e-01 -1.32286561e+00 4.04208712e-02 1.14719652e-01
-5.76986313e-01 -9.37151253e-01 3.69499356e-01 2.31370375e-01
-4.26452696e-01 -1.06893992e+00 -6.10978603e-01 -9.57617521e-01
2.25227267e-01 -6.21621609e-01 1.28478682e+00 -1.96506828e-01
7.21572161e-01 -5.68232909e-02 -5.80054343e-01 -7.50584245e-01
-8.62982631e-01 5.55204868e-01 -3.37031335e-01 2.53265768e-01
2.95858204e-01 1.66584104e-02 -1.83203653e-01 -2.62671169e-02
-3.65134388e-01 -1.68744445e-01 7.72435606e-01 7.66361594e-01
3.45504612e-01 -4.22094911e-01 9.00611401e-01 -1.13876545e+00
1.22076905e+00 -8.22118402e-01 -5.24787486e-01 4.89625514e-01
-9.75951016e-01 -1.97896510e-02 5.16852677e-01 8.70483834e-03
-1.03477633e+00 -6.60449117e-02 -6.89205527e-01 7.82703578e-01
3.14964563e-01 7.59880066e-01 -2.75293499e-01 1.14354035e-02
5.41747689e-01 1.45312287e-02 -1.26095653e-01 -1.55248940e-01
6.45757765e-02 1.19879818e+00 2.25608081e-01 -2.75018662e-01
-9.13263559e-02 -1.97943062e-01 -1.38324380e-01 -6.34865463e-01
-6.24190152e-01 -2.08148941e-01 -1.78643197e-01 -5.82677484e-01
8.52019072e-01 -1.01428545e+00 -9.01580930e-01 1.01891625e+00
-9.42770779e-01 -2.52236664e-01 2.89747924e-01 1.90036938e-01
-4.46695566e-01 1.30442351e-01 -6.91284597e-01 -7.99667060e-01
-7.98740029e-01 -1.31667697e+00 6.66471601e-01 -3.00294273e-02
-4.53149617e-01 -7.27376640e-01 3.35930198e-01 5.22754490e-01
1.49468333e-01 7.10917413e-02 1.13569093e+00 -8.83973300e-01
4.73204195e-01 -1.53745130e-01 -1.93136092e-02 4.20285702e-01
-4.56802458e-01 8.58623981e-02 -1.10009539e+00 3.06571331e-02
-1.63970530e-01 -5.19593060e-01 9.13452387e-01 3.06901127e-01
6.38762891e-01 -3.18270206e-01 -1.89986955e-02 4.95441742e-02
1.15424848e+00 3.83394927e-01 8.06976974e-01 9.82204497e-01
1.55508086e-01 6.85647964e-01 9.69818771e-01 -1.61828436e-02
1.10748827e+00 6.70069337e-01 3.40258777e-01 1.62784368e-01
9.44525376e-02 1.81483299e-01 6.68920517e-01 1.07385850e+00
-7.97499046e-02 -4.82804120e-01 -1.35074461e+00 4.33319420e-01
-1.80635881e+00 -6.75141275e-01 -5.94557047e-01 1.83896732e+00
9.47318017e-01 6.69712901e-01 3.63666207e-01 6.01591229e-01
2.41673231e-01 -3.49819139e-02 3.87656316e-02 -1.26216149e+00
-6.75886750e-01 1.40087545e-01 4.87126112e-01 9.15083051e-01
-1.21168613e+00 1.04024696e+00 6.36402702e+00 5.95396340e-01
-1.22413862e+00 6.03967123e-02 8.30223799e-01 4.08896357e-01
-2.89693743e-01 -7.80553296e-02 -6.47413731e-01 6.50491476e-01
1.29047143e+00 7.87947327e-02 2.46804893e-01 3.80346090e-01
-3.22263509e-01 -2.81726927e-01 -3.86784822e-01 5.60431838e-01
4.27310675e-01 -8.63519669e-01 2.44658649e-01 -3.97713661e-01
8.45136642e-01 1.45579875e-01 1.39422432e-01 7.57803023e-01
-3.51870656e-02 -7.75952101e-01 1.17796898e+00 5.57039261e-01
4.34097379e-01 -8.07485402e-01 1.39905059e+00 2.46643946e-01
-6.98970735e-01 -4.81628537e-01 7.72813186e-02 3.43899280e-02
-6.33007735e-02 4.62809086e-01 -9.74875093e-01 5.94417930e-01
6.33689523e-01 5.68975449e-01 -8.03526640e-01 7.47489333e-01
-3.42394590e-01 7.10975289e-01 -1.09165408e-01 -6.06527269e-01
3.12107563e-01 -2.76765496e-01 6.63388133e-01 1.79646122e+00
2.33224839e-01 -4.09129143e-01 -4.29574400e-01 -3.31514597e-01
-1.20191544e-01 8.47508430e-01 -1.90146118e-01 2.72198617e-01
1.86959997e-01 1.42718279e+00 -7.78493583e-01 -7.63064384e-01
-1.98153183e-02 4.68348086e-01 2.10839421e-01 3.45165171e-02
-7.89342701e-01 -4.27893281e-01 -1.55538917e-01 -1.45508900e-01
-7.31934002e-03 3.48511577e-01 -7.49384642e-01 -1.16858351e+00
-1.18631408e-01 -1.40978491e+00 6.29043102e-01 -8.03588748e-01
-9.61132407e-01 1.12579203e+00 -1.23287402e-02 -1.15808368e+00
-5.58903575e-01 -1.01094699e+00 -1.97159111e-01 8.94170463e-01
-1.59307051e+00 -1.25797057e+00 1.43955678e-01 3.00662547e-01
3.28794956e-01 -5.66636562e-01 7.53434241e-01 3.62024873e-01
-2.82929510e-01 6.69756114e-01 1.48544744e-01 -3.11712116e-01
1.09610975e+00 -1.59263003e+00 5.83834061e-03 5.03333271e-01
-1.79535881e-01 3.52315679e-02 8.02519798e-01 -6.38423681e-01
-9.05763745e-01 -4.73649502e-01 1.85385394e+00 -4.31208998e-01
6.50263369e-01 -3.52986813e-01 -5.73270023e-01 3.18277717e-01
7.82152057e-01 -9.70668197e-01 5.47099352e-01 2.35038176e-01
-1.83560863e-01 -2.94138223e-01 -1.26167226e+00 4.56913561e-01
-2.29755417e-02 -2.23656312e-01 -6.26914561e-01 3.27378958e-01
6.10244311e-02 -5.84343255e-01 -9.56952095e-01 6.66238308e-01
7.51157999e-01 -7.87640214e-01 6.23190224e-01 -1.81536004e-01
7.19945610e-01 -2.74570227e-01 -1.26227289e-01 -1.52228892e+00
1.36111304e-01 -1.31179214e-01 -4.40305471e-02 8.69313240e-01
1.36267114e+00 -8.39814484e-01 1.43897563e-01 -1.59787223e-01
-6.21906146e-02 -9.38694954e-01 -6.88295007e-01 4.34389412e-02
3.55690569e-01 -4.35761869e-01 5.27957082e-01 1.06029034e+00
3.52100939e-01 8.63838315e-01 -1.29367709e-01 -2.55660683e-01
-2.47896407e-02 7.89245516e-02 2.49595031e-01 -1.13249743e+00
1.50252998e-01 -8.52300882e-01 -1.87872782e-01 -3.58409703e-01
3.17360371e-01 -1.08914518e+00 -1.39527172e-01 -1.61710703e+00
1.74629882e-01 -3.25950235e-01 -5.22494912e-01 4.22480971e-01
-1.43913522e-01 5.04429996e-01 4.62998301e-01 -4.48013132e-04
-5.77103913e-01 1.56112671e-01 7.90593147e-01 -2.98402250e-01
-2.54027754e-01 7.62959644e-02 -1.00440192e+00 8.93761873e-01
9.22401845e-01 -3.46077353e-01 -1.63022637e-01 -5.79856597e-02
9.37651932e-01 -3.65794897e-02 -2.69170463e-01 -8.15291703e-01
-1.56261966e-01 1.09391727e-01 3.08049023e-01 -7.86632657e-01
3.59957546e-01 -5.50345600e-01 -1.24373950e-01 5.62970698e-01
-1.51106879e-01 3.73159230e-01 2.76513606e-01 -1.89561434e-02
-4.69528526e-01 -2.78203458e-01 6.95234418e-01 2.62302548e-01
-5.65213323e-01 -5.59691429e-01 -8.43716443e-01 -1.73605964e-01
9.56227720e-01 1.22709841e-01 -4.84544277e-01 -2.80936420e-01
-9.24803078e-01 4.04282540e-01 -2.99269222e-02 5.52809119e-01
2.14115933e-01 -1.13292134e+00 -1.07738900e+00 -1.49847567e-01
1.37143791e-01 -7.23875582e-01 1.24434598e-01 1.07756877e+00
-7.38483608e-01 6.49550498e-01 -4.93684202e-01 -3.60244393e-01
-1.28758442e+00 -1.12139456e-01 4.73213643e-01 -9.04726386e-01
2.98286736e-01 5.81578135e-01 -6.31709158e-01 -9.14400160e-01
-1.13201335e-01 -1.13474645e-01 -1.53024781e+00 9.46440458e-01
1.15134478e-01 4.87363994e-01 7.17823744e-01 -1.15111578e+00
-6.24046504e-01 2.63170213e-01 -6.71927407e-02 -5.67665517e-01
1.34079337e+00 2.08090499e-01 -5.88572741e-01 5.26623070e-01
9.69254673e-01 4.24340755e-01 -2.42147923e-01 3.20313454e-01
3.62888604e-01 3.77023935e-01 2.23584793e-04 -1.68465650e+00
-5.95441282e-01 5.03780067e-01 6.98802471e-01 5.81710160e-01
1.21776998e+00 -2.61477977e-01 4.36086744e-01 5.50548911e-01
-8.31680521e-02 -1.54813313e+00 -4.59192246e-01 1.27772331e+00
1.12770414e+00 -1.21504498e+00 -3.10513172e-02 -1.54261783e-01
-1.23071396e+00 1.00142133e+00 4.03659910e-01 4.15225364e-02
6.26466155e-01 1.43963054e-01 5.16753733e-01 -3.22732031e-02
-9.55275834e-01 6.84498101e-02 6.04054451e-01 1.70687929e-01
8.81378591e-01 4.08098400e-01 -1.55352616e+00 1.01762521e+00
-5.44461250e-01 -2.62095898e-01 6.03887260e-01 8.06140542e-01
-1.61824167e-01 -1.27931762e+00 -3.49365115e-01 3.92286003e-01
-8.83278131e-01 -1.19115017e-01 -9.71056163e-01 8.43704998e-01
1.72813967e-01 9.35885131e-01 -2.54686564e-01 -7.33043134e-01
5.89428425e-01 2.29964718e-01 2.10653588e-01 -9.99110378e-03
-1.09923625e+00 -2.13033217e-03 8.14808369e-01 -6.07801117e-02
-7.39808798e-01 -7.93032646e-01 -1.07772148e+00 -1.86634868e-01
-1.29020274e-01 4.58872169e-01 1.00984013e+00 9.85469222e-01
8.90858844e-02 3.05751026e-01 7.35870659e-01 -4.36913580e-01
-1.92959383e-01 -1.41581190e+00 -2.13338599e-01 1.25473231e-01
9.24505815e-02 -3.72906715e-01 -7.12973326e-02 -1.22404732e-01] | [11.168465614318848, 6.9236907958984375] |
abf43e44-f47a-486c-aaf8-39161d3f5475 | towards-more-realistic-membership-inference | 2306.12983 | null | https://arxiv.org/abs/2306.12983v1 | https://arxiv.org/pdf/2306.12983v1.pdf | Towards More Realistic Membership Inference Attacks on Large Diffusion Models | Generative diffusion models, including Stable Diffusion and Midjourney, can generate visually appealing, diverse, and high-resolution images for various applications. These models are trained on billions of internet-sourced images, raising significant concerns about the potential unauthorized use of copyright-protected images. In this paper, we examine whether it is possible to determine if a specific image was used in the training set, a problem known in the cybersecurity community and referred to as a membership inference attack. Our focus is on Stable Diffusion, and we address the challenge of designing a fair evaluation framework to answer this membership question. We propose a methodology to establish a fair evaluation setup and apply it to Stable Diffusion, enabling potential extensions to other generative models. Utilizing this evaluation setup, we execute membership attacks (both known and newly introduced). Our research reveals that previously proposed evaluation setups do not provide a full understanding of the effectiveness of membership inference attacks. We conclude that the membership inference attack remains a significant challenge for large diffusion models (often deployed as black-box systems), indicating that related privacy and copyright issues will persist in the foreseeable future. | ['Paweł Morawiecki', 'Tomasz Trzciński', 'Przemysław Rokita', 'Stanisław Pawlak', 'Antoni Kowalczuk', 'Jan Dubiński'] | 2023-06-22 | null | null | null | null | ['inference-attack', 'membership-inference-attack'] | ['adversarial', 'computer-vision'] | [ 3.67096782e-01 6.11856170e-02 2.00053789e-02 -8.57647136e-02
-6.55244350e-01 -1.22627032e+00 1.06627917e+00 -1.25183538e-01
-3.21693808e-01 4.42814112e-01 -4.66953516e-02 -7.85154939e-01
4.29294966e-02 -8.27949405e-01 -6.75946891e-01 -6.18001580e-01
-1.53664231e-01 9.76369530e-02 2.92972982e-01 -5.09216636e-02
3.96022022e-01 7.31572211e-01 -1.06300437e+00 1.96809154e-02
6.57736778e-01 8.48277569e-01 -3.12105089e-01 9.14337575e-01
4.58157748e-01 9.19108927e-01 -1.23628235e+00 -1.05637050e+00
6.78856850e-01 -3.55441123e-01 -7.13248491e-01 1.67119190e-01
7.44235218e-01 -8.93168509e-01 -2.99901754e-01 1.26816607e+00
5.18828034e-01 -1.67731598e-01 5.82804382e-01 -1.75897563e+00
-1.16668439e+00 5.99717319e-01 -6.75920844e-01 4.13603306e-01
2.36436099e-01 4.60417509e-01 7.36273706e-01 -2.82803923e-01
9.87457693e-01 1.22045052e+00 3.44529688e-01 5.86671770e-01
-1.21294415e+00 -9.73420203e-01 -2.98860180e-03 -1.57200724e-01
-1.29027569e+00 -5.93142509e-01 6.20502412e-01 -4.13636029e-01
4.03439075e-01 5.58225691e-01 2.14584753e-01 1.79794526e+00
2.15787992e-01 5.69625735e-01 1.37288892e+00 -1.79690480e-01
3.66517246e-01 5.03708005e-01 3.90193760e-02 4.29315716e-01
6.23600304e-01 1.74181163e-01 -1.89035624e-01 -7.33234107e-01
8.13210309e-01 -5.18269539e-01 -3.94634664e-01 -4.20833796e-01
-7.94023693e-01 1.12660766e+00 2.52344042e-01 1.58769824e-02
-1.73340738e-01 6.08172081e-02 2.15008616e-01 3.69258642e-01
3.47835630e-01 5.18143952e-01 1.64483562e-01 3.89101356e-02
-9.16398883e-01 2.69193858e-01 1.04347110e+00 7.71461129e-01
5.33330739e-01 4.17228714e-02 6.62613362e-02 1.02007054e-01
2.44543716e-01 5.64309120e-01 1.08929095e-03 -1.10550845e+00
2.18416229e-01 6.05398007e-02 1.80987298e-01 -1.41699302e+00
2.02303514e-01 -2.78089732e-01 -6.20932102e-01 5.97161055e-01
5.38671732e-01 -5.31383812e-01 -7.15900958e-01 1.77477682e+00
3.99337500e-01 3.30762446e-01 1.19979292e-01 6.92697823e-01
3.56196463e-01 5.54768562e-01 2.95208931e-01 -5.56894392e-02
1.21612585e+00 -4.40064251e-01 -4.10712987e-01 -4.78708297e-02
9.91060138e-02 -8.67463470e-01 8.92899990e-01 3.74838024e-01
-9.27401602e-01 -2.02747405e-01 -1.18692970e+00 1.60041139e-01
-3.82765561e-01 -3.85526985e-01 6.52477801e-01 1.48611033e+00
-1.27856719e+00 4.46475029e-01 -6.73612773e-01 -5.19974232e-01
5.60083032e-01 1.92698240e-01 -2.00781792e-01 -9.48488116e-02
-1.12471259e+00 8.34075511e-01 -9.67693850e-02 -2.77423263e-01
-1.18100679e+00 -3.39239955e-01 -6.23357296e-01 -1.62612975e-01
3.68213743e-01 -6.56740963e-01 9.14066195e-01 -8.15844595e-01
-1.09423912e+00 8.73675406e-01 2.79122084e-01 -8.17183375e-01
8.53780091e-01 6.01386540e-02 -5.99758506e-01 2.96215385e-01
-1.99701525e-02 5.26783466e-01 1.35637558e+00 -1.61458051e+00
-3.09724957e-01 -3.44470680e-01 4.40111309e-01 -7.42600560e-02
-5.79511464e-01 3.90511990e-01 -1.80133089e-01 -7.34865725e-01
-4.95109290e-01 -1.14781392e+00 -2.84963518e-01 5.76190688e-02
-7.01227665e-01 2.97623664e-01 1.36742103e+00 -6.16158962e-01
1.18813539e+00 -2.19691849e+00 -4.19714510e-01 5.15648723e-01
3.53686452e-01 4.84567523e-01 -1.61305040e-01 5.26911080e-01
2.71649778e-01 8.53458941e-01 -7.32828747e-04 -3.09974641e-01
1.48622263e-02 -2.29668990e-02 -7.60845125e-01 5.76119840e-01
-2.29671113e-02 7.38744438e-01 -4.39162016e-01 -3.03860337e-01
-5.49239516e-02 6.86609030e-01 -3.28991175e-01 9.61477682e-02
-1.27410710e-01 2.76088148e-01 -3.98692906e-01 6.54728353e-01
1.09564376e+00 -4.69903558e-01 4.40944768e-02 6.75109448e-03
1.47979334e-01 -1.65405899e-01 -1.06417561e+00 9.83578622e-01
1.07469328e-01 7.42148995e-01 2.43493840e-01 -1.90562144e-01
7.99997926e-01 5.14872707e-02 3.41351926e-02 -4.27116930e-01
2.17048779e-01 -5.31490147e-02 -1.73295327e-02 -1.80038750e-01
8.24341536e-01 1.83412760e-01 -3.22032385e-02 9.50953662e-01
-3.42046261e-01 2.93099731e-02 1.66668370e-01 7.02377319e-01
1.22139180e+00 -3.72680992e-01 -2.10792422e-01 -5.69102913e-02
1.93315953e-01 -6.48416281e-02 4.29007076e-02 1.18397379e+00
-4.87203449e-01 6.15258932e-01 6.11342549e-01 -1.88975874e-02
-1.04038310e+00 -1.26206315e+00 5.56166060e-02 7.20052063e-01
3.90900433e-01 -5.39534688e-01 -1.23836172e+00 -1.02574658e+00
-5.19469753e-02 7.87576973e-01 -5.72702408e-01 -1.79027125e-01
-2.28392929e-01 -8.71505022e-01 1.05106759e+00 6.34843558e-02
5.27248859e-01 -7.44518399e-01 -6.57714128e-01 -1.40967831e-01
1.93201769e-02 -1.15258825e+00 -5.61119795e-01 -5.90166330e-01
-3.65802199e-01 -1.21147323e+00 -6.93552554e-01 -2.41366014e-01
6.41209483e-01 4.03969735e-01 8.60752165e-01 2.42989268e-02
-7.82761425e-02 7.97744393e-01 -1.94238320e-01 -3.60546470e-01
-1.06386387e+00 2.87237298e-02 -1.74688086e-01 2.19909400e-01
2.06780672e-01 -3.81192297e-01 -9.05815959e-01 5.53512156e-01
-1.27792382e+00 -4.73244935e-01 3.91023248e-01 1.92473069e-01
4.40104958e-03 3.21842313e-01 4.53404516e-01 -1.14760137e+00
1.31813192e+00 -4.86217231e-01 -7.45850742e-01 3.52071613e-01
-8.38740349e-01 -2.78063595e-01 2.89813817e-01 -7.63138473e-01
-9.25744414e-01 -4.19101745e-01 -8.44106227e-02 -5.10436356e-01
-2.51742363e-01 3.20170462e-01 -1.97407499e-01 -5.43688238e-01
9.31862295e-01 1.48123667e-01 2.22858429e-01 -1.40324354e-01
5.00161886e-01 5.27281523e-01 4.12874877e-01 -5.43160260e-01
1.15091026e+00 7.75158167e-01 -1.73848584e-01 -8.71689260e-01
-1.68948933e-01 3.48087102e-02 6.24554753e-02 -1.54816285e-01
7.00281799e-01 -7.47345984e-01 -7.08115697e-01 6.98744357e-01
-1.03202569e+00 -1.43048331e-01 -6.86628073e-02 1.63723174e-02
-2.39101514e-01 9.06622648e-01 -7.89444387e-01 -7.90356338e-01
-4.93377477e-01 -1.30511832e+00 7.96320140e-01 1.43117219e-01
-4.58266795e-01 -8.90887737e-01 -2.74020582e-02 5.99111795e-01
7.53376126e-01 4.64247823e-01 7.22839773e-01 -7.33233273e-01
-9.99048471e-01 -3.65494639e-01 -2.81445235e-01 4.95456755e-01
-6.99858963e-02 5.07320225e-01 -9.42071021e-01 -6.14725888e-01
2.15782031e-01 -2.58678496e-01 5.58681905e-01 2.13197291e-01
7.80581415e-01 -7.33061373e-01 -7.21682683e-02 3.33458543e-01
1.20132554e+00 1.32220238e-01 9.44101632e-01 3.72683764e-01
3.85655820e-01 4.22957838e-01 1.37857065e-01 4.98795956e-01
2.67674595e-01 2.76782334e-01 4.62113053e-01 -7.92994648e-02
-2.10523773e-02 -4.54533458e-01 4.15056914e-01 -2.07537767e-02
2.30775893e-01 -7.06421852e-01 -7.75894523e-01 1.57993153e-01
-1.39688814e+00 -1.13863122e+00 6.96700662e-02 2.32016635e+00
5.92454433e-01 3.23242068e-01 2.40108222e-01 -6.78306073e-02
9.71386194e-01 4.35755968e-01 -5.53378522e-01 -3.55886191e-01
-1.51880667e-01 -4.66374122e-02 6.47569895e-01 3.67056549e-01
-1.19082904e+00 8.31465244e-01 7.26530218e+00 7.58708119e-01
-1.24532533e+00 1.47157565e-01 1.10419679e+00 2.56242961e-01
-4.98005211e-01 3.12350661e-01 -7.85768807e-01 4.65983272e-01
1.04384458e+00 -4.65358436e-01 4.42334861e-01 8.29278648e-01
-6.27874881e-02 -1.14675816e-02 -7.51975954e-01 5.87700248e-01
3.11058819e-01 -1.40862000e+00 2.30700746e-01 6.36996508e-01
6.21626198e-01 -1.51263013e-01 7.66659796e-01 -5.70717007e-02
7.54288793e-01 -9.14526105e-01 6.79186106e-01 7.51256347e-02
7.09191024e-01 -7.25445271e-01 1.85714781e-01 2.65312433e-01
-5.14017463e-01 9.72969234e-02 -3.65947932e-01 3.34709227e-01
3.50073099e-01 3.33288670e-01 -8.89402092e-01 1.72228217e-01
3.84297103e-01 2.02472717e-01 -9.05177534e-01 8.40917587e-01
-2.23485976e-01 7.36761689e-01 -3.22052389e-01 2.21674278e-01
1.04537502e-01 7.76980445e-03 6.58487499e-01 9.86910224e-01
4.17380691e-01 -3.02752942e-01 -5.11106616e-03 1.09279919e+00
-2.83517003e-01 -1.94055438e-01 -7.42887557e-01 -4.07520682e-01
4.03776437e-01 1.11686087e+00 -8.87332261e-01 -7.55230635e-02
-2.20994592e-01 1.14514792e+00 1.41820684e-02 5.38384438e-01
-9.38011885e-01 5.32155000e-02 6.80649221e-01 3.27655494e-01
1.90195084e-01 -3.51241082e-01 -9.33068916e-02 -1.24544680e+00
-2.66410470e-01 -1.29052687e+00 5.08418739e-01 -7.40123451e-01
-1.55274808e+00 6.14748895e-01 1.50790550e-02 -8.45062017e-01
-2.70689309e-01 -3.05296868e-01 -6.00493968e-01 7.62499630e-01
-1.32252860e+00 -1.25458276e+00 -7.21026286e-02 7.55525291e-01
-5.56471273e-02 -7.93403983e-02 8.22784543e-01 7.81629905e-02
-5.83886921e-01 7.73546040e-01 -7.64816552e-02 2.52869070e-01
7.87042320e-01 -9.42233384e-01 8.10133159e-01 1.36730719e+00
3.73640120e-01 1.10039854e+00 7.73196936e-01 -8.46868813e-01
-1.37493742e+00 -8.50281596e-01 3.77802849e-01 -7.71898031e-01
7.22116709e-01 -4.88551825e-01 -8.98073196e-01 7.21548855e-01
5.03592014e-01 -1.83237135e-01 7.98500597e-01 -2.46330991e-01
-6.83838665e-01 2.12370068e-01 -1.48343635e+00 8.53189468e-01
7.44814515e-01 -7.35754550e-01 -5.44476956e-02 7.51212910e-02
7.00049698e-01 -1.81605205e-01 -7.52220094e-01 1.28144786e-01
4.87344235e-01 -1.15281534e+00 1.01055670e+00 -4.10801739e-01
2.76942730e-01 1.69088296e-03 2.42352113e-02 -8.41809392e-01
-5.22843003e-02 -1.11719990e+00 -2.50656843e-01 1.54786110e+00
4.46711987e-01 -8.90579283e-01 9.17931080e-01 1.25367928e+00
7.42187858e-01 1.56580172e-02 -6.80197656e-01 -7.71470964e-01
2.39581093e-01 -2.69460171e-01 5.15495539e-01 1.18535352e+00
-4.04885501e-01 1.18486948e-01 -5.91463983e-01 5.63804567e-01
9.80183005e-01 -1.03090420e-01 1.11669064e+00 -8.27432871e-01
-3.11055601e-01 -4.21419621e-01 -3.95473301e-01 -8.17505121e-01
-2.16950867e-02 -3.80999118e-01 -7.14383602e-01 -1.12422109e+00
5.95492125e-02 -4.35190141e-01 -1.42996684e-01 9.32581499e-02
-1.39005706e-02 3.50325286e-01 4.90464509e-01 3.17249835e-01
-3.98358256e-01 1.01309270e-01 1.11479914e+00 -1.52501062e-01
2.26345330e-01 1.93223894e-01 -1.25692677e+00 4.51955169e-01
8.33568394e-01 -4.92034376e-01 -8.40535104e-01 -1.89096361e-01
1.72664747e-01 -2.45618269e-01 6.11564219e-01 -9.23977137e-01
4.38990057e-01 -2.26172581e-01 2.19430104e-01 -1.30497053e-01
2.76780874e-01 -8.47792804e-01 3.47488940e-01 2.62214929e-01
-3.75796378e-01 2.37439629e-02 1.37043044e-01 6.87923670e-01
1.36684984e-01 -1.50462940e-01 6.88280463e-01 -1.40689015e-01
-5.91259360e-01 4.65774000e-01 -4.06713545e-01 1.37324899e-01
1.34774184e+00 -2.64668584e-01 -8.50373268e-01 -9.01460469e-01
-5.39489686e-01 -2.11415127e-01 1.02609050e+00 5.41462600e-01
6.69479549e-01 -8.92131925e-01 -6.62848651e-01 3.62419933e-01
-1.16013452e-01 -6.23888373e-01 1.78075805e-01 2.78595060e-01
-5.03362179e-01 -2.75056094e-01 -1.39598429e-01 -2.51549095e-01
-1.38020599e+00 8.77264380e-01 1.13593794e-01 -1.69247895e-01
-4.40699637e-01 7.11533844e-01 2.40057275e-01 2.49205083e-02
2.14442715e-01 2.56331563e-01 1.32485703e-01 -1.06728353e-01
7.12645531e-01 3.12401026e-01 -3.41339111e-01 -7.86123514e-01
-1.04101494e-01 1.22561567e-02 -4.29873854e-01 -4.17028040e-01
8.88069510e-01 -2.83626765e-01 -2.32407488e-02 -3.51530701e-01
1.21116948e+00 3.57225627e-01 -1.30356967e+00 1.01325557e-01
-3.00013989e-01 -8.28491688e-01 -1.82004794e-01 -8.63806605e-01
-1.10305679e+00 6.71640217e-01 6.40759170e-01 7.66851425e-01
8.67008746e-01 -1.61565870e-01 9.51729298e-01 6.84254616e-02
4.77160305e-01 -7.25883007e-01 4.31521870e-02 3.69203612e-02
7.60143161e-01 -1.12596142e+00 8.10164213e-02 -5.58221519e-01
-8.06455255e-01 5.70600569e-01 4.36262369e-01 -1.73526462e-02
7.92538166e-01 5.01969993e-01 3.01759154e-01 -1.25985757e-01
-4.95023787e-01 2.51794696e-01 -9.41056311e-02 8.81430209e-01
-7.44554698e-02 -1.02113210e-01 -1.39860928e-01 1.32147223e-01
-1.62162811e-01 -1.18051775e-01 9.46117818e-01 1.11941659e+00
-9.46852639e-02 -1.34130466e+00 -5.98266363e-01 2.05318511e-01
-8.07908058e-01 -4.13023774e-03 -7.21742034e-01 6.58851862e-01
-1.64562225e-01 1.27047777e+00 -2.06931144e-01 -5.34136057e-01
7.64052868e-02 -2.95952827e-01 1.80908337e-01 -2.85907656e-01
-6.80353820e-01 -4.31576148e-02 1.29125774e-01 -3.13274026e-01
-2.76779205e-01 -6.42758191e-01 -4.85331267e-01 -9.28559780e-01
-3.80682826e-01 -7.81540945e-02 6.91075921e-01 3.44056219e-01
6.47880316e-01 -1.00781932e-01 5.60387909e-01 -3.90137404e-01
-1.04229307e+00 -4.59399045e-01 -5.15430272e-01 6.68435693e-01
1.57456309e-01 -2.39911288e-01 -6.37269020e-01 2.26264112e-02] | [12.439229965209961, 1.1276533603668213] |
0a51245e-ac3b-469f-809f-5fef495d9679 | mianet-aggregating-unbiased-instance-and-1 | 2305.13864 | null | https://arxiv.org/abs/2305.13864v1 | https://arxiv.org/pdf/2305.13864v1.pdf | MIANet: Aggregating Unbiased Instance and General Information for Few-Shot Semantic Segmentation | Existing few-shot segmentation methods are based on the meta-learning strategy and extract instance knowledge from a support set and then apply the knowledge to segment target objects in a query set. However, the extracted knowledge is insufficient to cope with the variable intra-class differences since the knowledge is obtained from a few samples in the support set. To address the problem, we propose a multi-information aggregation network (MIANet) that effectively leverages the general knowledge, i.e., semantic word embeddings, and instance information for accurate segmentation. Specifically, in MIANet, a general information module (GIM) is proposed to extract a general class prototype from word embeddings as a supplement to instance information. To this end, we design a triplet loss that treats the general class prototype as an anchor and samples positive-negative pairs from local features in the support set. The calculated triplet loss can transfer semantic similarities among language identities from a word embedding space to a visual representation space. To alleviate the model biasing towards the seen training classes and to obtain multi-scale information, we then introduce a non-parametric hierarchical prior module (HPM) to generate unbiased instance-level information via calculating the pixel-level similarity between the support and query image features. Finally, an information fusion module (IFM) combines the general and instance information to make predictions for the query image. Extensive experiments on PASCAL-5i and COCO-20i show that MIANet yields superior performance and set a new state-of-the-art. Code is available at https://github.com/Aldrich2y/MIANet. | ['Tianlin Huang', 'Yuan Feng', 'Qiong Chen', 'Yong Yang'] | 2023-05-23 | mianet-aggregating-unbiased-instance-and | http://openaccess.thecvf.com//content/CVPR2023/html/Yang_MIANet_Aggregating_Unbiased_Instance_and_General_Information_for_Few-Shot_Semantic_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Yang_MIANet_Aggregating_Unbiased_Instance_and_General_Information_for_Few-Shot_Semantic_CVPR_2023_paper.pdf | cvpr-2023-1 | ['few-shot-image-segmentation', 'general-knowledge'] | ['computer-vision', 'miscellaneous'] | [ 4.47646230e-01 2.69965902e-02 -5.25592387e-01 -6.25945687e-01
-1.17121029e+00 -2.35966101e-01 2.77791023e-01 1.41633689e-01
-5.29543817e-01 4.84237522e-01 -1.64651141e-01 2.92055994e-01
8.32407456e-03 -8.80092382e-01 -7.76773274e-01 -7.22067833e-01
3.22280526e-01 4.26946253e-01 6.13698304e-01 -2.27205157e-02
2.35196173e-01 1.02207042e-01 -1.72460127e+00 4.99520808e-01
1.29013789e+00 1.20026851e+00 4.23376799e-01 2.19014138e-01
-5.57972133e-01 3.65085512e-01 -6.21261537e-01 -3.84938955e-01
2.28702769e-01 -3.96444291e-01 -6.49713695e-01 2.71020710e-01
4.20481026e-01 -4.12991606e-02 3.82707566e-02 1.26158035e+00
4.52090532e-01 3.10827792e-01 6.92860186e-01 -1.45511138e+00
-5.51550031e-01 2.86590040e-01 -8.17773938e-01 9.89118367e-02
1.44856691e-01 1.51949570e-01 1.11939192e+00 -1.15735233e+00
6.97755992e-01 1.25090563e+00 4.14508402e-01 3.95887733e-01
-1.10703135e+00 -6.91857100e-01 3.84778410e-01 3.61671776e-01
-1.39050269e+00 -1.83269531e-01 9.06212866e-01 -2.96493590e-01
5.09686470e-01 1.81636378e-01 8.05176914e-01 9.65672672e-01
-9.40993652e-02 1.30327475e+00 9.11295712e-01 -2.84153908e-01
2.67137885e-01 4.62498963e-01 4.86592680e-01 7.72506177e-01
1.45951644e-01 -1.70482203e-01 -5.91889977e-01 1.11300454e-01
4.92337704e-01 1.02917410e-01 -2.73863196e-01 -5.48405528e-01
-9.96786296e-01 8.90757799e-01 7.83873618e-01 3.01962703e-01
-2.39458516e-01 -1.74485192e-01 3.35559607e-01 -8.78917649e-02
4.93403971e-01 2.81308323e-01 -4.02207255e-01 2.49027342e-01
-9.95856583e-01 -7.16930330e-02 5.99222660e-01 9.17410672e-01
1.33757687e+00 -3.32258731e-01 -4.91097599e-01 1.10902381e+00
4.43409473e-01 3.73116493e-01 6.04271233e-01 -8.70800972e-01
4.79079455e-01 9.41152930e-01 -1.28018260e-01 -9.07440722e-01
-9.23829805e-03 -4.89083439e-01 -3.98609102e-01 8.04749057e-02
2.16941610e-01 8.54475722e-02 -1.29323435e+00 1.69551229e+00
6.84535861e-01 3.69369924e-01 5.49321473e-02 1.03959632e+00
9.72148478e-01 7.01580524e-01 4.26749177e-02 6.56981766e-02
1.38085854e+00 -1.25626540e+00 -4.82719183e-01 -6.32124424e-01
5.17844439e-01 -5.60246468e-01 1.18256736e+00 1.09970644e-01
-8.11067164e-01 -6.81989968e-01 -1.24977016e+00 -6.92651272e-02
-6.62021279e-01 1.63238809e-01 3.10403883e-01 3.26229662e-01
-5.33793092e-01 2.91082770e-01 -7.57190645e-01 -2.74141967e-01
7.15876281e-01 1.05525270e-01 -1.62593588e-01 -3.97229850e-01
-1.31855345e+00 6.35899842e-01 7.78490186e-01 -1.25959620e-01
-7.33991444e-01 -8.51270497e-01 -1.21688020e+00 -7.22521916e-02
6.96451366e-01 -5.97500741e-01 8.32648754e-01 -1.23536813e+00
-1.17650950e+00 9.43058431e-01 -6.60785511e-02 -2.89932489e-01
2.76816219e-01 -9.51215848e-02 -8.28227848e-02 3.12457472e-01
4.72187966e-01 1.12766933e+00 9.13898528e-01 -1.55119896e+00
-7.81642854e-01 -5.07120192e-01 -7.62538239e-02 4.13232952e-01
-3.09040636e-01 -3.69867951e-01 -9.25766349e-01 -6.32471085e-01
1.95201829e-01 -7.88492978e-01 -3.00311875e-02 1.00977480e-01
-5.56125700e-01 -1.90009207e-01 8.13853800e-01 -5.50967216e-01
1.10050797e+00 -2.25776362e+00 1.27226010e-01 2.65191317e-01
-4.04349016e-03 4.17626709e-01 -3.71682554e-01 -8.55188258e-03
1.11141719e-01 -1.29855126e-01 -6.76886201e-01 -3.19048643e-01
-6.30523935e-02 4.57615405e-01 -2.83845738e-02 1.21272750e-01
4.28296655e-01 1.02302861e+00 -1.06731522e+00 -8.32004249e-01
3.59719932e-01 4.67499822e-01 -4.71613556e-01 2.10110888e-01
-2.83435196e-01 1.52683362e-01 -5.82681656e-01 7.57386446e-01
6.36649907e-01 -3.13333392e-01 -2.55496472e-01 -3.61191928e-01
7.53019303e-02 -1.46366939e-01 -1.21478570e+00 2.05986786e+00
-3.31417948e-01 9.69076753e-02 3.40756890e-03 -1.25902176e+00
9.11146581e-01 -1.24092564e-01 4.77148026e-01 -6.55979991e-01
2.43381381e-01 2.08358601e-01 -2.63719320e-01 -4.91318017e-01
2.17056826e-01 7.25042820e-02 -4.94523793e-02 1.04712613e-01
4.33901727e-01 -2.42316768e-01 2.74411321e-01 1.96782216e-01
7.25222409e-01 2.69251525e-01 2.00118944e-01 -2.77012978e-02
5.07938445e-01 1.31781861e-01 1.08890879e+00 5.40859163e-01
-3.94495040e-01 8.23117614e-01 2.79744565e-01 7.24103302e-03
-6.54886723e-01 -1.25634229e+00 -7.90399760e-02 1.09769082e+00
6.65400028e-01 -1.94772899e-01 -9.00092423e-01 -9.08388436e-01
5.21373227e-02 8.40557158e-01 -7.13787019e-01 -4.28884625e-01
-1.22916684e-01 -7.24289775e-01 8.04518759e-02 6.60686612e-01
6.90303326e-01 -9.50141072e-01 -5.87920427e-01 8.72326791e-02
-2.05364153e-01 -8.80278826e-01 -6.77516937e-01 1.07200921e-01
-6.14079833e-01 -1.03349757e+00 -7.59822428e-01 -8.71313274e-01
6.81558788e-01 2.78146595e-01 9.49583054e-01 -1.09239347e-01
-6.93856955e-01 4.19198275e-01 -3.96841526e-01 -3.63641262e-01
-4.67531979e-02 -5.02439849e-02 -3.86447817e-01 3.04276526e-01
5.42765439e-01 -1.86939135e-01 -6.98103964e-01 2.78244019e-01
-9.84746814e-01 2.51039833e-01 6.63452685e-01 1.02438188e+00
9.95502532e-01 -1.14390358e-01 6.03986561e-01 -8.38269711e-01
1.41522512e-01 -5.73263168e-01 -4.75356400e-01 4.49814767e-01
-4.11971778e-01 -8.16208944e-02 2.03156903e-01 -4.51917201e-01
-1.24089456e+00 2.21164286e-01 6.33008555e-02 -6.73909366e-01
-3.38095240e-02 4.71610069e-01 -5.58798194e-01 2.69603223e-01
4.13539022e-01 2.31116518e-01 5.16921990e-02 -2.39499152e-01
5.67024171e-01 6.53823435e-01 5.02171338e-01 -6.18790388e-01
5.82584560e-01 5.56432962e-01 -4.55730110e-01 -7.22156286e-01
-1.23872757e+00 -7.78430223e-01 -6.45364761e-01 -1.34605080e-01
1.03422940e+00 -9.17936027e-01 6.55577984e-03 4.47625756e-01
-9.76529956e-01 -1.88377693e-01 -5.79362512e-01 4.79956567e-01
-4.91667092e-01 2.09760189e-01 -4.33241040e-01 -6.57681942e-01
-3.57399225e-01 -1.29378819e+00 1.32736003e+00 6.52485967e-01
9.49935317e-02 -8.67996931e-01 -3.55295315e-02 6.67265356e-01
-1.00503430e-01 1.98285848e-01 7.46206343e-01 -8.80071223e-01
-7.51960754e-01 -3.01818281e-01 -5.17459869e-01 6.18326306e-01
1.22739516e-01 -1.03249423e-01 -1.07384837e+00 -1.25680640e-01
-6.04659393e-02 -4.43972975e-01 1.17377973e+00 4.41642493e-01
1.17724812e+00 -9.93977208e-03 -5.01219988e-01 5.65573633e-01
1.46362424e+00 1.10495739e-01 4.75522429e-01 1.28673956e-01
8.16327035e-01 6.47014201e-01 1.18054259e+00 2.78028876e-01
5.10459781e-01 5.02697408e-01 2.64329433e-01 -5.72770350e-02
-2.99732864e-01 -3.30233753e-01 1.40107095e-01 7.20773637e-01
4.33706164e-01 -3.52811590e-02 -7.69410610e-01 7.80457675e-01
-1.91689682e+00 -6.98874533e-01 3.35467100e-01 2.27870154e+00
1.00267339e+00 2.41175205e-01 -3.37913223e-02 -1.34199575e-01
9.42118347e-01 9.78090912e-02 -8.81429672e-01 2.69386563e-02
-2.38757040e-02 -6.71960339e-02 3.41962606e-01 4.14633334e-01
-1.15556252e+00 1.14894986e+00 4.47630930e+00 1.35721767e+00
-9.91028666e-01 1.97080448e-01 7.39526033e-01 -4.01835144e-02
-4.11607951e-01 -9.09248143e-02 -8.59075248e-01 5.35678446e-01
4.55751538e-01 -6.26166612e-02 1.36001796e-01 8.35243344e-01
-1.96514770e-01 -2.98460871e-01 -8.93539429e-01 1.01169813e+00
2.63599664e-01 -1.05491221e+00 9.31631774e-02 -1.73386648e-01
7.70841002e-01 7.46961385e-02 1.09373368e-01 5.68010509e-01
4.28619310e-02 -6.56195283e-01 5.18521488e-01 7.02946007e-01
5.92362940e-01 -7.15234637e-01 6.41676247e-01 4.18028504e-01
-1.26889884e+00 -8.32863003e-02 -4.25253272e-01 4.49477822e-01
5.23751378e-02 6.85296059e-01 -6.81136191e-01 6.68953598e-01
6.85438931e-01 8.48491967e-01 -5.48287690e-01 1.15933704e+00
-2.99107701e-01 3.77296031e-01 -3.44833493e-01 1.68146059e-01
3.80800337e-01 -3.82309467e-01 5.14410377e-01 1.05235040e+00
1.37714595e-01 -1.02161244e-01 4.86715764e-01 1.11115086e+00
-5.18823639e-02 2.84094244e-01 -3.25298071e-01 1.05799884e-01
5.57536185e-01 1.42547953e+00 -9.40767586e-01 -6.21288419e-01
-5.80960631e-01 1.19606709e+00 4.36189592e-01 5.19448638e-01
-8.66867065e-01 -5.77710271e-01 6.23618364e-01 -2.58346230e-01
4.49779242e-01 2.32567430e-01 -3.65334123e-01 -1.26943195e+00
1.54615223e-01 -5.05174100e-01 5.23479939e-01 -7.23938107e-01
-1.37718177e+00 3.46927047e-01 4.73502763e-02 -1.29634249e+00
4.36159372e-02 -5.06546855e-01 -7.17384994e-01 7.48782992e-01
-1.63790691e+00 -1.27871609e+00 -4.40092891e-01 3.44345033e-01
7.73335814e-01 3.41534093e-02 5.01351118e-01 1.98627800e-01
-7.55168855e-01 7.02149332e-01 -8.60090405e-02 2.37842634e-01
6.89743340e-01 -1.32819724e+00 -5.80912791e-02 7.01094270e-01
2.16478512e-01 4.45687890e-01 2.91344464e-01 -7.09235668e-01
-9.74949121e-01 -1.32925010e+00 3.48119944e-01 -3.84021848e-01
5.20987093e-01 -2.22782820e-01 -1.30710542e+00 4.76739436e-01
-1.00977391e-01 2.87197530e-01 7.61594594e-01 -2.36303717e-01
-4.85769242e-01 -2.05606624e-01 -1.25068772e+00 4.40891534e-01
8.72026265e-01 -4.98905689e-01 -7.51936972e-01 1.72891989e-01
9.34510112e-01 -3.00303429e-01 -8.90331984e-01 5.90070724e-01
3.91142458e-01 -7.54921198e-01 1.00478506e+00 -3.86599541e-01
3.58368367e-01 -5.22205591e-01 -2.90562361e-01 -1.32509077e+00
-3.75025831e-02 1.67937968e-02 5.59651367e-02 1.43042767e+00
5.03666818e-01 -5.13304830e-01 8.06233764e-01 5.67442954e-01
-1.97508618e-01 -1.11311567e+00 -1.02139413e+00 -7.70112574e-01
6.37033358e-02 -4.32709396e-01 4.57304448e-01 8.50425720e-01
-2.34040141e-01 4.03083891e-01 6.99400753e-02 1.59089878e-01
7.38547802e-01 3.94875318e-01 4.92645442e-01 -1.16661870e+00
-7.67904669e-02 -2.81996161e-01 -4.81857955e-01 -8.65520775e-01
3.70777309e-01 -1.11943722e+00 2.74906129e-01 -1.65588033e+00
4.36792403e-01 -4.30765599e-01 -6.39867783e-01 4.60735828e-01
-5.43786526e-01 3.17487776e-01 2.72260100e-01 7.01734275e-02
-8.58066499e-01 8.61681640e-01 1.32176948e+00 -3.79068404e-01
-2.71536618e-01 -9.95022207e-02 -6.12058163e-01 8.27109754e-01
7.02645063e-01 -3.85551184e-01 -5.73871017e-01 -1.83722898e-01
-2.06756070e-01 -2.06596881e-01 4.14773583e-01 -1.05577791e+00
2.19985068e-01 -8.97230729e-02 4.65970337e-01 -7.06929564e-01
5.40380061e-01 -7.60055065e-01 -3.94023031e-01 2.43963912e-01
-2.79857606e-01 -6.69584274e-01 4.25479151e-02 7.09089696e-01
-3.23604047e-01 -3.65645915e-01 9.40016150e-01 -8.97623748e-02
-1.06960797e+00 5.91758013e-01 3.08163196e-01 3.78446758e-01
1.14290392e+00 -4.58345979e-01 -2.17070520e-01 1.49754882e-01
-6.61805451e-01 6.87019289e-01 5.87877810e-01 6.11084342e-01
7.76041031e-01 -1.42543817e+00 -4.24041837e-01 2.66858339e-01
5.98657429e-01 3.50985080e-01 4.33341533e-01 8.29570889e-01
-5.06865084e-02 -1.24560725e-02 -8.47339854e-02 -8.24088871e-01
-1.14934099e+00 6.44890010e-01 2.81557828e-01 -2.29904614e-02
-4.33870256e-01 1.05247426e+00 4.70913261e-01 -5.60424984e-01
2.48183370e-01 -3.42447102e-01 -2.05989614e-01 5.29972017e-01
5.53205550e-01 1.64386660e-01 -2.34369576e-01 -7.05985308e-01
-4.22380090e-01 7.90848434e-01 -1.61080420e-01 -8.99219513e-02
1.03163016e+00 -2.04069480e-01 2.57926732e-02 7.50344336e-01
1.57946706e+00 -4.43842143e-01 -1.42724144e+00 -5.90682983e-01
6.07998902e-03 -4.90483373e-01 6.09287247e-02 -7.04921007e-01
-1.28368533e+00 9.75114465e-01 6.93907857e-01 -3.14241230e-01
1.14306605e+00 2.64823675e-01 9.63860750e-01 1.88267916e-01
1.23461112e-01 -1.49912524e+00 2.68607676e-01 2.75914937e-01
6.24343812e-01 -1.56879210e+00 -2.59888709e-01 -6.23817265e-01
-8.63894284e-01 6.97316229e-01 9.46947753e-01 -4.73702811e-02
6.62055671e-01 -8.69110748e-02 1.52216300e-01 -1.50598153e-01
-4.26895559e-01 -5.40028334e-01 5.49069464e-01 5.92677832e-01
9.76270437e-02 2.19350141e-02 -1.80139422e-01 8.52052271e-01
2.23156691e-01 2.50545621e-04 4.74080518e-02 9.36669230e-01
-7.48788774e-01 -7.61325717e-01 -2.41579816e-01 5.64732194e-01
-8.77653584e-02 -5.54003939e-02 -2.11550787e-01 5.10884464e-01
4.05021936e-01 7.95436800e-01 1.24978580e-01 -3.45366865e-01
2.38172278e-01 2.27353245e-01 2.54269928e-01 -9.14193749e-01
-2.43908733e-01 1.31544575e-01 -2.94305444e-01 -6.79508150e-01
-3.44491720e-01 -4.42007095e-01 -1.56415093e+00 4.45813835e-01
-3.64066929e-01 1.35148689e-01 5.52763581e-01 9.39490438e-01
3.04070860e-01 6.97778344e-01 5.03207028e-01 -8.16165507e-01
-3.47549886e-01 -7.69411087e-01 -5.90329528e-01 5.77009678e-01
1.55869514e-01 -8.63371491e-01 -3.75040799e-01 -1.58184648e-01] | [9.580294609069824, 1.3711696863174438] |
33467a7f-3790-466d-9666-d4287416ecf1 | experiences-of-adapting-multimodal-machine | null | null | https://aclanthology.org/2021.mmtlrl-1.7 | https://aclanthology.org/2021.mmtlrl-1.7.pdf | Experiences of Adapting Multimodal Machine Translation Techniques for Hindi | Multimodal Neural Machine Translation (MNMT) is an interesting task in natural language processing (NLP) where we use visual modalities along with a source sentence to aid the source to target translation process. Recently, there has been a lot of works in MNMT frameworks to boost the performance of standalone Machine Translation tasks. Most of the prior works in MNMT tried to perform translation between two widely known languages (e.g. English-to-German, English-to-French ). In this paper, We explore the effectiveness of different state-of-the-art MNMT methods, which use various data oriented techniques including multimodal pre-training, for low resource languages. Although the existing methods works well on high resource languages, usability of those methods on low-resource languages is unknown. In this paper, we evaluate the existing methods on Hindi and report our findings. | ['Asif Ekbal', 'Dibyanayan Bandyopadhyay', 'Baban Gain'] | null | null | null | null | mmtlrl-ranlp-2021-9 | ['multimodal-machine-translation'] | ['natural-language-processing'] | [ 2.82912225e-01 -2.69593388e-01 -3.24766874e-01 -2.49264061e-01
-1.06057858e+00 -8.28731894e-01 9.54616010e-01 -3.85788769e-01
-5.34205616e-01 1.10090566e+00 3.10613036e-01 -7.41384447e-01
5.89460552e-01 -4.68703926e-01 -7.75736749e-01 -3.59188437e-01
5.13667226e-01 8.06832135e-01 -1.27085224e-01 -5.42165756e-01
1.17111847e-01 1.03460483e-01 -8.93034101e-01 8.70006919e-01
8.39531362e-01 2.28336483e-01 2.77849734e-01 5.02211988e-01
-5.17997026e-01 3.89640719e-01 -4.35665935e-01 -8.70457768e-01
3.66663605e-01 -8.24059129e-01 -1.10178995e+00 -2.95163751e-01
3.49309415e-01 -1.00054793e-01 2.38641188e-03 9.95542049e-01
8.12914550e-01 -2.06975341e-01 6.68433607e-01 -1.21604407e+00
-1.15227354e+00 6.04048491e-01 -7.26867497e-01 7.20533878e-02
5.57944477e-01 8.65070447e-02 5.67689359e-01 -1.27085960e+00
1.04493701e+00 1.47831035e+00 1.55584246e-01 6.15015090e-01
-8.41526508e-01 -5.20432591e-01 -2.99210399e-01 1.48744851e-01
-9.98548269e-01 -4.83899802e-01 5.48711538e-01 -1.61678717e-01
1.11134768e+00 9.48488489e-02 3.30460742e-02 1.44742846e+00
3.40236396e-01 8.39197397e-01 1.71288276e+00 -9.96779859e-01
-2.98454106e-01 5.64392447e-01 -2.38174871e-01 4.98820603e-01
-1.54384747e-01 -1.13947481e-01 -8.82721305e-01 1.60158172e-01
5.84697843e-01 -3.66835833e-01 -1.96061586e-03 1.02277920e-01
-1.81078959e+00 7.71328628e-01 6.26791865e-02 5.90662181e-01
-1.96448460e-01 -3.18056464e-01 5.18472016e-01 7.77265549e-01
4.82502878e-01 2.07023099e-01 -4.56784457e-01 -3.53904516e-01
-8.86595726e-01 -1.02818251e-01 7.47525394e-01 1.10503256e+00
7.16245830e-01 -1.12899676e-01 -3.35342914e-01 1.01625931e+00
6.41084671e-01 9.33644056e-01 5.82594514e-01 -2.20262423e-01
1.41849196e+00 7.19117820e-01 -9.97466743e-02 -5.70321202e-01
-1.35137647e-01 3.92222889e-02 -7.33016729e-01 8.82310644e-02
4.32384282e-01 -4.28334087e-01 -1.05854106e+00 1.56332839e+00
2.25843415e-01 -4.82626319e-01 6.06029212e-01 1.12082112e+00
1.11681151e+00 1.06278265e+00 5.72460517e-02 -2.01436266e-01
1.32270920e+00 -1.30579543e+00 -7.22415209e-01 -2.60985225e-01
6.41064942e-01 -1.58931434e+00 1.44482064e+00 -2.66008428e-03
-1.21299469e+00 -5.21457613e-01 -7.74963021e-01 -3.37271094e-01
-6.87888682e-01 5.54531217e-01 3.87109995e-01 6.12993836e-01
-1.15145457e+00 1.38201609e-01 -6.06842279e-01 -1.07056606e+00
7.58776367e-02 4.42696691e-01 -8.03166926e-01 -2.36099765e-01
-1.34189236e+00 1.52337909e+00 3.52509886e-01 2.11077228e-01
-9.02149439e-01 1.97542444e-01 -6.27143502e-01 -5.87760627e-01
8.39620605e-02 -8.34874690e-01 1.28985357e+00 -1.41043627e+00
-1.80784416e+00 1.18175399e+00 -4.50599283e-01 -1.49784401e-01
6.25059068e-01 -2.77778804e-01 -4.01524067e-01 5.26944064e-02
3.49586532e-02 1.00260389e+00 6.36209905e-01 -1.19154012e+00
-5.41082084e-01 -3.67666423e-01 1.64680615e-01 6.89531147e-01
-1.57654926e-01 6.32854104e-01 -4.19401795e-01 -3.75220656e-01
-1.32315755e-01 -9.97409582e-01 8.58960450e-02 -3.95144701e-01
-4.87782091e-01 -1.40072390e-01 8.87941778e-01 -9.90574896e-01
9.65600491e-01 -1.82439983e+00 4.65949744e-01 -2.37080559e-01
-3.33096236e-01 2.19634145e-01 -3.64298791e-01 9.89494920e-01
2.90377766e-01 1.89324632e-01 -3.42265517e-01 -3.66076082e-01
9.60240439e-02 2.90745467e-01 -1.84528694e-01 3.71618718e-01
2.09319815e-01 1.31491613e+00 -4.64863479e-01 -7.84271598e-01
2.11749375e-02 4.41582233e-01 2.31115967e-01 2.81400412e-01
2.03042272e-02 6.14852190e-01 -8.07321519e-02 1.03696263e+00
5.96391678e-01 2.59328872e-01 -4.20737900e-02 -1.49684921e-01
-4.00715142e-01 1.05567597e-01 -5.40683389e-01 2.09011078e+00
-3.95584375e-01 9.06972289e-01 -5.82917072e-02 -7.00317085e-01
7.88015187e-01 6.29001141e-01 1.71907973e-02 -8.66383731e-01
2.51449555e-01 5.34750164e-01 2.27744967e-01 -9.02483284e-01
5.11552691e-01 -3.13390851e-01 8.89560506e-02 7.04994738e-01
1.94880739e-01 2.50233501e-01 4.45477009e-01 -7.65581205e-02
5.89051127e-01 7.75908589e-01 2.50778168e-01 1.02564737e-01
4.06549424e-01 5.50131321e-01 -5.91208227e-02 4.82175261e-01
-2.62943149e-01 7.78587162e-01 6.35286272e-02 -7.67079666e-02
-1.23463380e+00 -1.06309235e+00 2.25107729e-01 1.30098653e+00
-3.50060910e-02 -8.00071061e-02 -8.93022776e-01 -8.09282124e-01
-6.96365535e-01 4.89154428e-01 -3.72301221e-01 1.44888327e-01
-7.32885957e-01 -1.02256870e+00 9.35684025e-01 2.66999096e-01
7.17750490e-01 -1.32607651e+00 -4.51404095e-01 1.03136063e-01
-6.49884582e-01 -1.26070905e+00 -3.69789988e-01 -1.33736834e-01
-9.37751234e-01 -5.02331436e-01 -1.14993286e+00 -1.16935205e+00
5.44319808e-01 2.18750983e-01 1.13818681e+00 -4.60838288e-01
1.88418135e-01 9.61887389e-02 -5.89474082e-01 -6.31733954e-01
-6.20910645e-01 3.72503400e-01 4.24092822e-03 -9.06991661e-02
5.84431410e-01 -1.62875593e-01 -1.33285806e-01 3.21128160e-01
-8.69171679e-01 5.20924866e-01 1.20393145e+00 5.58766127e-01
2.78486222e-01 -8.11353087e-01 4.36583966e-01 -6.26958370e-01
8.90660942e-01 -3.31995457e-01 -1.77992821e-01 7.57483125e-01
-1.76674604e-01 3.18956226e-02 3.37165385e-01 -7.15681255e-01
-1.13073444e+00 3.76728289e-02 4.56540026e-02 -7.44035542e-02
-3.06950033e-01 9.37620521e-01 -1.27798155e-01 -4.14716005e-02
5.68673193e-01 3.39297950e-01 -1.79863617e-01 -4.65947777e-01
4.79504526e-01 9.81232107e-01 2.60936886e-01 -5.55281699e-01
8.07441473e-01 1.29968017e-01 -1.48428664e-01 -6.01141453e-01
-2.09414855e-01 -1.95235997e-01 -1.02615285e+00 -2.43551001e-01
1.01036620e+00 -7.93267906e-01 -7.77233317e-02 2.40696013e-01
-1.48061347e+00 -1.99400470e-01 3.50849539e-01 8.72296691e-01
-2.87174404e-01 2.76987791e-01 -6.86037183e-01 -7.14910328e-01
-7.17472196e-01 -1.31702018e+00 1.19876266e+00 9.89194512e-02
-1.39482588e-01 -9.25034344e-01 4.89977300e-01 6.17180586e-01
5.72953224e-01 -4.23948616e-02 1.02052200e+00 -5.62666237e-01
-3.64492178e-01 2.27874309e-01 -4.45280701e-01 2.21782904e-02
1.10099770e-01 7.00957850e-02 -6.51078224e-01 -1.91582888e-01
-1.88338906e-01 -5.51992655e-01 5.64989328e-01 8.36270973e-02
-2.55073071e-01 -2.35734999e-01 -1.45370662e-01 1.29779607e-01
1.36378551e+00 1.80869967e-01 8.19719970e-01 5.61589181e-01
7.80447900e-01 6.96928740e-01 6.86949968e-01 -3.02251488e-01
6.07420862e-01 6.68119371e-01 -2.60618161e-02 -4.34898943e-01
-1.58537731e-01 -1.12168476e-01 9.66815710e-01 1.18677640e+00
-3.80237192e-01 -3.76286745e-01 -1.16317558e+00 4.56803024e-01
-2.11977673e+00 -7.48722434e-01 -3.15219581e-01 1.97743881e+00
1.00369418e+00 -3.06565404e-01 1.40881732e-01 -3.19537669e-01
7.37265766e-01 -2.18558773e-01 5.03769442e-02 -8.15699995e-01
-4.49335814e-01 6.60371259e-02 1.52625173e-01 4.50157940e-01
-9.65639412e-01 1.35243845e+00 6.11135578e+00 6.85218632e-01
-1.47581327e+00 5.27144969e-01 4.62839007e-01 5.76709472e-02
-1.09591365e-01 1.17612734e-01 -7.18145669e-01 1.12477593e-01
1.08656037e+00 1.85054436e-01 5.13048410e-01 3.87866557e-01
2.36704841e-01 -2.01224342e-01 -1.14234221e+00 1.08512557e+00
4.25081819e-01 -9.51795936e-01 3.62458467e-01 -4.43754792e-02
7.79532015e-01 3.70335400e-01 5.12686111e-02 4.81623441e-01
-2.39623263e-02 -1.22146261e+00 5.57418287e-01 2.17938408e-01
8.20979536e-01 -6.47709370e-01 1.03263366e+00 4.29609865e-01
-8.82039607e-01 4.79864985e-01 -4.90082264e-01 -4.40690331e-02
3.84661973e-01 -1.79861132e-02 -1.07501853e+00 9.33918118e-01
4.83947277e-01 3.40806633e-01 -7.50881910e-01 6.75184965e-01
-2.79724598e-01 6.85014069e-01 -2.56896317e-01 -3.52154285e-01
5.69831014e-01 -3.26768816e-01 3.31642509e-01 1.59069550e+00
5.55060863e-01 -4.23306584e-01 1.44391254e-01 5.19162059e-01
-1.96784496e-01 8.83586407e-01 -9.45105195e-01 -3.34719032e-01
-1.80314034e-01 1.29413223e+00 -6.23134375e-01 -3.21572542e-01
-7.61719704e-01 1.34292197e+00 3.11658919e-01 6.54055476e-01
-8.39719117e-01 -1.88564211e-01 -1.09804735e-01 -2.06782952e-01
-2.98786402e-01 -4.35988277e-01 -3.37206274e-01 -1.38019216e+00
1.93198174e-01 -1.30996990e+00 2.81377852e-01 -1.08548117e+00
-1.30297291e+00 9.64933097e-01 1.32534847e-01 -1.41958392e+00
-2.50236422e-01 -7.25372493e-01 -3.98567379e-01 1.18896461e+00
-1.27727056e+00 -1.96435714e+00 2.90611655e-01 7.94090867e-01
8.53495538e-01 -6.11488402e-01 8.07134390e-01 4.38835233e-01
-3.39714348e-01 4.56422150e-01 7.54169375e-02 2.86674529e-01
1.28843021e+00 -9.19489503e-01 2.91334867e-01 1.02003562e+00
5.73525608e-01 7.91066110e-01 5.74168444e-01 -8.27217877e-01
-1.73270166e+00 -7.19415188e-01 1.46467280e+00 -6.62105620e-01
4.28155154e-01 -3.74735624e-01 -5.95973551e-01 7.20813096e-01
1.34857345e+00 -6.14581168e-01 6.03159368e-01 -1.08638182e-01
-1.43485516e-01 1.58692732e-01 -1.09599411e+00 9.80473042e-01
6.90126717e-01 -5.09314775e-01 -6.97514474e-01 3.95462424e-01
3.26882094e-01 -3.73123735e-01 -4.13414657e-01 4.50157374e-01
5.26733339e-01 -4.60603952e-01 7.16748536e-01 -8.00452054e-01
8.06007981e-01 -5.76956689e-01 -3.85901630e-01 -1.45087659e+00
1.44128188e-01 -5.02972305e-01 2.58255363e-01 1.28731883e+00
1.08676910e+00 -4.78623390e-01 3.49043846e-01 2.53404379e-01
-3.21507789e-02 -4.71670866e-01 -1.10008681e+00 -3.40338498e-01
2.22734109e-01 -1.71280041e-01 1.49388984e-01 1.08972597e+00
1.45774320e-01 1.08432591e+00 -7.95259297e-01 -1.19581267e-01
9.67371240e-02 1.58425048e-01 9.43559825e-01 -7.11578429e-01
-2.05337182e-01 -3.27918410e-01 -5.06594256e-02 -6.91398382e-01
-1.04568742e-01 -1.06742084e+00 -8.92759636e-02 -2.03577065e+00
5.61635792e-01 3.78206074e-01 -5.27896779e-03 7.76676297e-01
-1.48456037e-01 5.95424533e-01 3.17564875e-01 4.66737241e-01
-4.62594032e-01 3.21177930e-01 1.42529380e+00 -2.77898073e-01
-2.00155988e-01 -3.25081915e-01 -3.06733340e-01 3.63284022e-01
9.97455657e-01 -5.71177244e-01 -3.93723428e-01 -9.51320350e-01
2.91785091e-01 -1.16741866e-01 1.00157838e-02 -5.58402658e-01
1.40558392e-01 -2.32525513e-01 5.29225886e-01 -7.94315279e-01
2.34175771e-01 -7.24489093e-01 4.46343757e-02 1.61176831e-01
-2.18393087e-01 9.11240995e-01 1.56113610e-01 -1.59391567e-01
-5.46676457e-01 -2.71193236e-01 4.64160919e-01 -4.61368859e-01
-4.81445998e-01 -6.95625395e-02 -4.89667833e-01 -1.25088617e-01
8.68707895e-01 -9.29086804e-02 -5.38030148e-01 -3.91404063e-01
-3.05767864e-01 -5.01595289e-02 3.51188958e-01 7.24711895e-01
6.59837127e-01 -1.51133525e+00 -1.09596372e+00 -2.61399269e-01
1.96750432e-01 -6.93501532e-01 -2.83447385e-01 1.27188802e+00
-7.49304771e-01 6.25126183e-01 -7.65445828e-01 -5.72360635e-01
-1.52901399e+00 3.85153741e-01 3.55229084e-03 -3.35599631e-01
-2.69696731e-02 5.03270507e-01 -9.74667892e-02 -9.07215595e-01
-5.09057939e-02 1.97992194e-02 -3.09870601e-01 -2.96968292e-03
2.84586191e-01 1.34314358e-01 -3.12216505e-02 -1.14233911e+00
-3.99563909e-01 6.86141968e-01 9.64915846e-03 -1.08268356e+00
8.24139655e-01 -3.31478149e-01 -4.65275854e-01 7.96640635e-01
1.13410449e+00 2.73521394e-02 -3.40250582e-01 -2.13826299e-01
8.88592601e-02 -1.45272389e-01 -4.39755291e-01 -1.26410151e+00
-6.04754746e-01 1.33388221e+00 8.45052660e-01 -3.69441688e-01
1.04251373e+00 -1.42174467e-01 7.17785358e-01 6.05136573e-01
4.99827772e-01 -1.29738879e+00 -2.43037999e-01 8.23566198e-01
9.59807158e-01 -1.74894154e+00 -2.08413929e-01 2.85209157e-02
-1.08637583e+00 1.32279193e+00 6.54510677e-01 5.48618674e-01
4.02914546e-02 8.30058977e-02 8.26794147e-01 1.34717017e-01
-5.38979769e-01 -3.11868697e-01 6.47182703e-01 5.53655148e-01
1.02138197e+00 -7.58682145e-03 -6.99427307e-01 2.87084043e-01
-1.67453680e-02 5.60764037e-02 2.19731107e-01 1.25115466e+00
-1.78437307e-01 -1.57573771e+00 -7.16798484e-01 -8.37385058e-02
-6.79368973e-01 -5.06905377e-01 -1.04636884e+00 9.71988142e-01
5.12332283e-02 1.18633270e+00 -6.18985951e-01 -3.53363007e-01
2.66762942e-01 3.85367841e-01 7.10696697e-01 -4.79643404e-01
-8.87527645e-01 3.19135994e-01 2.76480258e-01 -7.99441859e-02
-8.83999944e-01 -5.69136143e-01 -9.62906420e-01 -2.43108109e-01
6.51483610e-02 -8.79975334e-02 9.97910261e-01 1.20675087e+00
1.64412782e-01 -4.19592634e-02 1.50300875e-01 -7.04718113e-01
-1.00162633e-01 -1.48698270e+00 1.67388827e-01 3.01078230e-01
-1.41571723e-02 -1.17804326e-01 2.52802759e-01 2.89316118e-01] | [11.485397338867188, 1.5389994382858276] |
84bbd099-57b2-4d41-b15a-5e907a53aeda | deep-convolutional-generative-adversarial-2 | 2005.04188 | null | https://arxiv.org/abs/2005.04188v1 | https://arxiv.org/pdf/2005.04188v1.pdf | Deep convolutional generative adversarial networks for traffic data imputation encoding time series as images | Sufficient high-quality traffic data are a crucial component of various Intelligent Transportation System (ITS) applications and research related to congestion prediction, speed prediction, incident detection, and other traffic operation tasks. Nonetheless, missing traffic data are a common issue in sensor data which is inevitable due to several reasons, such as malfunctioning, poor maintenance or calibration, and intermittent communications. Such missing data issues often make data analysis and decision-making complicated and challenging. In this study, we have developed a generative adversarial network (GAN) based traffic sensor data imputation framework (TSDIGAN) to efficiently reconstruct the missing data by generating realistic synthetic data. In recent years, GANs have shown impressive success in image data generation. However, generating traffic data by taking advantage of GAN based modeling is a challenging task, since traffic data have strong time dependency. To address this problem, we propose a novel time-dependent encoding method called the Gramian Angular Summation Field (GASF) that converts the problem of traffic time-series data generation into that of image generation. We have evaluated and tested our proposed model using the benchmark dataset provided by Caltrans Performance Management Systems (PeMS). This study shows that the proposed model can significantly improve the traffic data imputation accuracy in terms of Mean Absolute Error (MAE) and Root Mean Squared Error (RMSE) compared to state-of-the-art models on the benchmark dataset. Further, the model achieves reasonably high accuracy in imputation tasks even under a very high missing data rate ($>$ 50\%), which shows the robustness and efficiency of the proposed model. | ['Anuj Sharma', 'Pranamesh Chakraborty', 'Tongge Huang'] | 2020-05-05 | null | null | null | null | ['traffic-data-imputation'] | ['time-series'] | [ 5.00503838e-01 -4.27686125e-01 3.93759161e-02 -4.15011048e-01
-8.29352975e-01 1.16439521e-01 4.27509010e-01 -3.35923105e-01
-4.23329920e-02 1.27202630e+00 1.57213062e-01 -4.22397196e-01
-2.02035278e-01 -1.12538207e+00 -7.39028573e-01 -8.22102427e-01
3.09194088e-01 4.08928543e-01 -1.53136803e-02 -3.05783689e-01
-3.33070420e-02 4.02992994e-01 -1.75694847e+00 -1.01623535e-01
1.48594129e+00 1.07346666e+00 5.69392852e-02 2.76405096e-01
-2.88783073e-01 8.25629175e-01 -8.65133345e-01 -3.66839677e-01
3.71977448e-01 -3.69362980e-01 -1.73711002e-01 -6.94768131e-02
-2.15033628e-02 -3.76619607e-01 -6.72675967e-01 7.12607801e-01
3.96566719e-01 3.37198377e-01 5.44117153e-01 -1.95814621e+00
-3.90150636e-01 -3.54435071e-02 -6.51131272e-01 9.00980085e-02
-1.50720850e-01 3.91646266e-01 2.31859952e-01 -3.15200061e-01
7.37900734e-02 9.21509266e-01 7.08973527e-01 5.13805866e-01
-1.09280372e+00 -1.11989450e+00 -2.25149065e-01 3.74848753e-01
-1.39820957e+00 -5.50581634e-01 9.67066586e-01 -4.39299971e-01
5.53707242e-01 4.67188537e-01 3.70024920e-01 1.20038283e+00
2.58385092e-01 5.05955219e-01 1.11818361e+00 -1.95719004e-02
1.01197369e-01 -4.48631775e-03 -2.33002871e-01 2.08867520e-01
3.06353331e-01 2.26720273e-01 -1.18316866e-01 1.80624440e-01
5.61360240e-01 1.97531089e-01 9.45040062e-02 3.92146617e-01
-1.07016218e+00 5.75160921e-01 3.33372474e-01 -1.38724446e-01
-7.29852021e-01 3.35167438e-01 2.61026621e-01 1.84296861e-01
6.05079234e-01 -3.69700223e-01 -1.35380387e-01 -4.29079145e-01
-8.19160402e-01 3.00049812e-01 2.82489568e-01 1.03950202e+00
4.83244509e-01 7.32189655e-01 -3.61441046e-01 8.74156713e-01
6.66119158e-02 1.08280897e+00 1.24433741e-01 -9.29234743e-01
1.05184221e+00 3.09968382e-01 1.81584641e-01 -1.31032348e+00
-2.75297612e-01 -4.01040226e-01 -1.46153796e+00 7.91514516e-02
2.48822942e-01 -4.45730656e-01 -8.63289237e-01 2.01429653e+00
2.16015011e-01 5.86393952e-01 1.35698751e-01 6.70006335e-01
5.80571711e-01 9.66735244e-01 3.25525612e-01 -3.86013806e-01
8.30247700e-01 -5.02221704e-01 -1.01455402e+00 6.71853125e-02
3.31362218e-01 -7.70665884e-01 7.02748716e-01 7.68622607e-02
-7.95822620e-01 -8.05169165e-01 -6.92015648e-01 2.63059199e-01
-3.49997729e-01 7.90797696e-02 5.98305941e-01 7.21481860e-01
-4.66829300e-01 1.14426538e-01 -5.76653719e-01 -2.24672422e-01
5.97093284e-01 1.25356734e-01 -2.19025910e-01 -2.72909284e-01
-1.34345984e+00 7.22277701e-01 3.88152003e-02 3.97933096e-01
-5.60814977e-01 -8.76368105e-01 -6.53492689e-01 -1.84933603e-01
2.26251915e-01 -8.19368720e-01 7.61123717e-01 -6.46957219e-01
-1.17209113e+00 -1.30942995e-02 -3.68248403e-01 -5.87106645e-01
6.02052212e-01 1.57615572e-01 -1.13675630e+00 -3.50563943e-01
1.92503437e-01 5.52961648e-01 6.67892575e-01 -1.20341086e+00
-4.82026666e-01 -2.44532257e-01 -2.86235094e-01 -2.03878254e-01
3.30925994e-02 -3.21122438e-01 5.36067747e-02 -8.89957011e-01
-1.32148817e-01 -9.23161209e-01 -1.23841360e-01 -2.24635959e-01
-5.27738750e-01 2.70461943e-02 1.11367667e+00 -9.62784290e-01
1.30804098e+00 -1.85867012e+00 -4.62186575e-01 3.21302980e-01
-1.66852251e-01 3.74132097e-01 -3.84614915e-02 4.53416437e-01
-1.03712685e-01 1.34697989e-01 -4.33241367e-01 -3.15329045e-01
2.52303481e-02 4.56941307e-01 -3.32124054e-01 1.42286330e-01
2.15846300e-01 8.64346027e-01 -4.80404764e-01 -3.34708065e-01
7.30984271e-01 5.93699276e-01 -2.43536785e-01 2.54492611e-01
-6.44673929e-02 1.02299702e+00 -4.94218707e-01 6.22871518e-01
1.00118196e+00 2.13585988e-01 -5.05541444e-01 -3.98609608e-01
-7.79248327e-02 -1.53553471e-01 -1.20446384e+00 1.30268121e+00
-7.68113077e-01 7.02275991e-01 -3.22304428e-01 -1.10955644e+00
1.04852176e+00 3.73940885e-01 8.03149879e-01 -1.29101646e+00
5.84662482e-02 7.79996812e-02 -1.42013401e-01 -8.07687700e-01
4.78219360e-01 -1.68116152e-01 -5.75614534e-02 2.53250629e-01
-6.00962639e-01 1.56884342e-01 1.54236853e-01 -4.61032540e-02
9.89126921e-01 -7.34966472e-02 -4.38825577e-01 3.13866615e-01
5.69845378e-01 1.21218048e-01 8.22094262e-01 3.07853758e-01
-8.68114904e-02 5.69037378e-01 1.50881425e-01 -3.07628542e-01
-1.26442945e+00 -9.58525836e-01 -1.25028208e-01 2.73684114e-01
-3.45559642e-02 1.39924258e-01 -6.00089312e-01 -2.82339633e-01
3.94107923e-02 1.08154452e+00 -3.97679597e-01 -2.16518641e-01
-5.34153342e-01 -9.57099974e-01 7.19494760e-01 5.96391678e-01
1.01027930e+00 -9.13182616e-01 -8.65223482e-02 4.16972667e-01
-6.27114892e-01 -1.41457677e+00 -2.89771676e-01 -5.43303967e-01
-5.70423961e-01 -8.10813010e-01 -5.27514279e-01 -2.01223522e-01
7.61513948e-01 4.41695780e-01 7.53374636e-01 -2.18954235e-02
-1.90714389e-01 -3.40537801e-02 -2.03917131e-01 -5.20574272e-01
-4.05209988e-01 -2.59583518e-02 -6.71038870e-03 4.12229449e-01
3.21738511e-01 -7.24880338e-01 -5.93528628e-01 5.33124804e-01
-9.55642223e-01 3.26568663e-01 5.26003540e-01 6.36584759e-01
4.96346653e-01 6.14890158e-01 1.22378373e+00 -6.81658089e-01
6.55566573e-01 -8.76591265e-01 -6.26012623e-01 4.79708239e-02
-5.38399279e-01 -1.36906862e-01 8.55318248e-01 -2.42312513e-02
-1.25175822e+00 -2.75472403e-01 -4.28351462e-01 -4.09849703e-01
-3.51777554e-01 3.45342308e-01 -3.37434649e-01 1.53691500e-01
1.83645934e-01 4.23118830e-01 2.13978454e-01 -2.63902307e-01
-1.21278919e-01 9.76288438e-01 5.31679451e-01 -5.58231354e-01
1.12187648e+00 3.83968323e-01 4.91558194e-01 -9.17332172e-01
-5.04729390e-01 -8.36847499e-02 -1.55114055e-01 -5.08353829e-01
7.37127602e-01 -9.12132323e-01 -8.82071376e-01 7.31262386e-01
-1.01226544e+00 -1.82789713e-02 -1.65554449e-01 5.82008719e-01
-3.89538080e-01 1.74366951e-01 -1.24872625e-01 -1.09095180e+00
-3.45416844e-01 -1.25996208e+00 7.97680676e-01 2.23940909e-01
1.49090067e-01 -8.82173359e-01 -2.23353013e-01 8.59503686e-01
1.04427135e+00 9.83507812e-01 8.71690750e-01 -5.77648655e-02
-8.58430266e-01 -3.43258321e-01 -3.82527977e-01 5.84125638e-01
2.77173430e-01 -3.31036597e-02 -7.41253078e-01 1.15248784e-01
-3.12281579e-01 1.35586455e-01 5.63549280e-01 3.70877296e-01
1.53157878e+00 -3.67905647e-01 -6.41994029e-02 5.12202084e-01
1.38096797e+00 4.33894753e-01 1.15958011e+00 2.52119508e-02
7.72998631e-01 5.20494282e-01 6.77407980e-01 4.70237494e-01
6.60634696e-01 7.01584995e-01 5.22646725e-01 -6.87639192e-02
-2.41878867e-01 -3.95967066e-01 1.51800036e-01 9.08525646e-01
-1.13406211e-01 -6.14757001e-01 -7.50040829e-01 4.80088979e-01
-1.89260578e+00 -1.34092820e+00 -7.17615128e-01 2.38197231e+00
3.60497862e-01 1.36916742e-01 6.79661781e-02 5.02518892e-01
6.94528341e-01 2.50986945e-02 -6.82051122e-01 -3.54470223e-01
-6.91990554e-02 2.87285328e-01 7.91559994e-01 2.00625211e-01
-6.87340677e-01 4.48662698e-01 5.09494019e+00 1.08557665e+00
-9.94479716e-01 8.14732835e-02 6.44194067e-01 3.63396674e-01
-4.38670844e-01 -2.66447783e-01 -5.01694322e-01 1.23378241e+00
1.33651340e+00 -5.94774671e-02 6.36781752e-01 3.83102387e-01
8.75386477e-01 -8.83701220e-02 -3.99194658e-01 9.81417179e-01
-1.60376862e-01 -1.21855474e+00 8.03617537e-02 1.60333470e-01
7.98672497e-01 -1.39381126e-01 5.37621379e-02 3.39747429e-01
6.24339432e-02 -1.20393431e+00 1.51505515e-01 9.40926850e-01
9.24677014e-01 -1.06467569e+00 8.70664537e-01 4.57747549e-01
-1.21263468e+00 3.63910198e-02 -1.81546211e-01 5.58370017e-02
7.02345490e-01 8.63060057e-01 -4.96853471e-01 8.47748578e-01
2.43596718e-01 4.71793413e-01 -1.97046816e-01 1.13483369e+00
1.39560550e-01 8.80354047e-01 -3.57504696e-01 3.60758275e-01
-1.10046104e-01 -5.47761440e-01 4.00009155e-01 6.19877994e-01
7.75441587e-01 2.04806760e-01 -4.28166464e-02 8.18906128e-01
-1.76841766e-01 -2.15526551e-01 -7.00372219e-01 2.32201815e-01
6.77434146e-01 9.10943568e-01 -3.69504131e-02 -6.50916845e-02
-3.21735740e-01 4.68071699e-01 -3.32480907e-01 5.35042703e-01
-1.45772290e+00 -4.47860152e-01 8.54225278e-01 3.36167067e-01
-7.72110000e-02 -3.62003863e-01 -4.08652633e-01 -8.39793146e-01
3.03127080e-01 -5.23985505e-01 -6.74328022e-03 -9.34707046e-01
-1.26391637e+00 4.94596869e-01 -1.18127503e-02 -1.70707977e+00
-1.71185553e-01 1.25692110e-03 -7.08056808e-01 9.32034910e-01
-1.70537758e+00 -1.32528889e+00 -8.36302757e-01 7.49862850e-01
4.88083750e-01 -2.15908483e-01 5.71650624e-01 9.66065586e-01
-9.70247626e-01 5.34156203e-01 2.72926390e-01 2.86650565e-02
2.76882201e-01 -6.23333216e-01 3.61725360e-01 9.74854112e-01
-3.42606008e-01 9.30731669e-02 6.68947160e-01 -5.56827247e-01
-1.42241633e+00 -1.67973149e+00 8.23558748e-01 -1.08232602e-01
3.60901296e-01 -3.99469398e-02 -7.40506113e-01 4.73929852e-01
3.05317715e-02 3.87533307e-02 5.35219371e-01 -4.61110175e-01
1.42210379e-01 -8.09401214e-01 -1.55612254e+00 3.55801225e-01
9.70220745e-01 -2.15369374e-01 -1.47914395e-01 3.66069138e-01
6.08415663e-01 -1.89877227e-01 -8.80598485e-01 6.05168104e-01
3.20932180e-01 -7.46617138e-01 8.95677626e-01 -4.25316513e-01
2.15688840e-01 -6.50242090e-01 -3.38462472e-01 -1.37729585e+00
1.05352722e-01 -5.20852387e-01 5.97074023e-03 1.71486652e+00
2.49095425e-01 -8.84382546e-01 6.60303175e-01 9.60943580e-01
-2.59979934e-01 -4.57196295e-01 -1.21997821e+00 -7.43112147e-01
-2.28824332e-01 -9.09147859e-01 1.10915780e+00 7.35243082e-01
-7.58124411e-01 1.22066341e-01 -8.40812325e-01 1.51316926e-01
9.35185850e-01 -1.09166637e-01 1.18241799e+00 -1.08591115e+00
2.70595253e-01 8.00902322e-02 -5.63214064e-01 -8.02591860e-01
2.43705139e-01 -5.10024726e-01 -1.64893374e-01 -1.64532018e+00
-2.16214880e-01 -9.66316819e-01 -2.97444254e-01 3.28672737e-01
3.10539845e-02 1.59499660e-01 1.23882312e-02 3.75440996e-03
-1.21189900e-01 9.62327659e-01 1.21380258e+00 -2.25711092e-01
1.05693340e-01 4.48072642e-01 -5.12462437e-01 2.17827737e-01
1.07886577e+00 -4.41779673e-01 -6.25426054e-01 -6.05224669e-01
-1.17958024e-01 4.06759381e-01 5.30276954e-01 -1.40493321e+00
2.71716952e-01 -4.93774563e-01 2.18469411e-01 -8.61311853e-01
5.04167438e-01 -1.32829952e+00 9.20997739e-01 2.88643688e-01
2.99655553e-02 3.55366677e-01 1.68415859e-01 6.59431934e-01
-2.76921093e-01 4.87674505e-01 3.94474536e-01 3.59642148e-01
-6.95403039e-01 5.72113991e-01 -2.62583613e-01 1.01004556e-01
1.21048594e+00 -3.08218211e-01 -5.62110484e-01 -6.26413584e-01
-3.03930402e-01 3.17800522e-01 1.40447587e-01 5.33210993e-01
6.48235559e-01 -1.71077466e+00 -8.81532669e-01 4.14690197e-01
1.31593779e-01 -1.42065948e-02 7.94605434e-01 1.01133060e+00
-3.28027725e-01 3.47125739e-01 -1.99578658e-01 -5.55430353e-01
-7.49091089e-01 2.37345487e-01 -6.34348467e-02 -9.07903165e-03
-3.23353231e-01 -2.05889978e-02 -4.02274042e-01 -4.62200612e-01
-8.42992309e-03 -3.24266762e-01 6.36434108e-02 -2.79150248e-01
2.30647564e-01 8.98762226e-01 3.00217479e-01 -9.73178864e-01
-1.20092995e-01 5.38644373e-01 5.25927126e-01 2.79970348e-01
1.21094370e+00 -3.24734002e-01 2.47941345e-01 1.16089910e-01
1.11870527e+00 -1.50806859e-01 -1.12917161e+00 -4.39479994e-03
-5.76951087e-01 -6.99981570e-01 2.76217256e-02 -7.93156326e-01
-1.49867451e+00 8.83286953e-01 6.23151302e-01 1.49035618e-01
1.25412154e+00 -7.49666393e-01 1.54302323e+00 5.64303547e-02
4.95446622e-01 -9.73511100e-01 -4.23589915e-01 2.32306108e-01
6.64418399e-01 -1.42529690e+00 -3.53742480e-01 -3.88609290e-01
-7.80808091e-01 6.77423716e-01 6.05133176e-01 1.40912697e-01
6.46924675e-01 1.56307861e-01 9.67422593e-03 2.62471795e-01
-7.01945305e-01 -3.04734707e-01 2.76201144e-02 8.18397462e-01
-9.39416811e-02 2.82981962e-01 -1.44980580e-01 3.44603986e-01
-2.14711279e-01 3.43297601e-01 4.38460559e-01 7.40046799e-01
-1.20590210e-01 -1.11022151e+00 -4.28795993e-01 7.32371569e-01
-2.03194231e-01 1.78890020e-01 4.95345771e-01 7.11740375e-01
2.41668090e-01 1.45736396e+00 1.22086257e-01 -5.95411956e-01
5.12555718e-01 -1.68805450e-01 -6.59282655e-02 1.50755569e-01
-1.26019984e-01 -5.57973862e-01 2.57382542e-01 -3.99629951e-01
-5.35956621e-01 -5.55598617e-01 -1.10633278e+00 -1.03070021e+00
-1.65192842e-01 2.11988032e-01 9.46612239e-01 1.03401947e+00
6.58294976e-01 8.20091486e-01 1.11024380e+00 -3.35596502e-01
-2.89878428e-01 -8.02391171e-01 -3.18492591e-01 7.91624486e-01
2.67806500e-01 -8.73811126e-01 -2.16370508e-01 -1.47447094e-01] | [6.508443832397461, 2.040390729904175] |
5968fbf3-bcbe-4b04-b587-5fe6b265e3c5 | query-your-model-with-definitions-in-framenet | 2212.02036 | null | https://arxiv.org/abs/2212.02036v1 | https://arxiv.org/pdf/2212.02036v1.pdf | Query Your Model with Definitions in FrameNet: An Effective Method for Frame Semantic Role Labeling | Frame Semantic Role Labeling (FSRL) identifies arguments and labels them with frame semantic roles defined in FrameNet. Previous researches tend to divide FSRL into argument identification and role classification. Such methods usually model role classification as naive multi-class classification and treat arguments individually, which neglects label semantics and interactions between arguments and thus hindering performance and generalization of models. In this paper, we propose a query-based framework named ArGument Extractor with Definitions in FrameNet (AGED) to mitigate these problems. Definitions of frames and frame elements (FEs) in FrameNet can be used to query arguments in text. Encoding text-definition pairs can guide models in learning label semantics and strengthening argument interactions. Experiments show that AGED outperforms previous state-of-the-art by up to 1.3 F1-score in two FrameNet datasets and the generalization power of AGED in zero-shot and fewshot scenarios. Our code and technical appendix is available at https://github.com/PKUnlp-icler/AGED. | ['Baobao Chang', 'Yiming Wang', 'Ce Zheng'] | 2022-12-05 | null | null | null | null | ['semantic-role-labeling'] | ['natural-language-processing'] | [ 3.86110216e-01 4.08593029e-01 -6.05578125e-01 -4.84461606e-01
-6.04454815e-01 -8.87467027e-01 9.22714829e-01 5.14395475e-01
-4.69171792e-01 8.31262410e-01 6.48223877e-01 -1.38572082e-01
-1.08094789e-01 -8.67258072e-01 -5.03660560e-01 -4.21801388e-01
4.72652316e-01 4.66876149e-01 8.56655180e-01 -4.92000312e-01
1.86732098e-01 -1.80071637e-01 -1.79288054e+00 9.10158455e-01
3.92961055e-01 9.52324152e-01 2.55850732e-01 5.02674937e-01
-4.30734783e-01 1.70095384e+00 -8.03547621e-01 -7.03772545e-01
-1.92539796e-01 -3.98508608e-01 -1.41474843e+00 -3.95184755e-01
3.93510126e-02 -1.35483518e-01 -3.13347816e-01 8.54868293e-01
6.14007473e-01 2.72714764e-01 -5.52750863e-02 -1.37969506e+00
-5.47120214e-01 1.13307035e+00 -6.51865378e-02 6.43040836e-01
7.93722153e-01 -4.46034372e-01 1.44580388e+00 -7.72091210e-01
9.15001333e-01 1.80149066e+00 5.55007517e-01 7.91191518e-01
-9.24435854e-01 -4.06207561e-01 4.06666398e-01 6.14022732e-01
-7.18988180e-01 -4.95366246e-01 7.91361630e-01 -1.86776966e-01
9.01717901e-01 4.97494429e-01 5.57398915e-01 1.21729743e+00
-4.59968925e-01 1.32649648e+00 5.88837743e-01 -5.28555691e-01
1.93964317e-01 -5.75939715e-01 6.20683074e-01 5.19074857e-01
4.58295643e-02 -3.11776668e-01 -8.15796316e-01 -3.79987478e-01
5.00846565e-01 -1.55204251e-01 -1.17778182e-01 -3.20376665e-03
-1.37313521e+00 9.05423641e-01 1.90110743e-01 2.83358753e-01
-9.94323567e-02 2.77743131e-01 8.96020591e-01 2.72790879e-01
5.32096744e-01 3.29676211e-01 -3.92300427e-01 -3.91874462e-01
-1.57705694e-01 6.61607862e-01 7.16548204e-01 9.46446896e-01
2.91811913e-01 -3.76843661e-01 -5.31939745e-01 1.00103354e+00
3.19842070e-01 2.35504702e-01 4.75086838e-01 -1.02645493e+00
4.20245141e-01 1.11160100e+00 7.63212964e-02 -8.50928903e-01
-4.60097313e-01 1.69314831e-01 8.89291614e-03 -4.71200019e-01
4.55196589e-01 2.71459725e-02 -4.83741581e-01 1.67637634e+00
8.43670607e-01 4.23829079e-01 2.21964329e-01 8.55443597e-01
1.47105074e+00 3.80757004e-01 6.01959467e-01 -1.41295165e-01
1.80528653e+00 -1.08975685e+00 -9.75826144e-01 -1.27979696e-01
9.15116072e-01 -7.53469110e-01 1.25897169e+00 -2.29807839e-01
-9.83866692e-01 -5.27172804e-01 -8.13700080e-01 -4.73312020e-01
-1.87206581e-01 -9.33727100e-02 6.40036285e-01 3.26788992e-01
-4.83517081e-01 3.38525385e-01 -7.96515346e-01 -2.20342994e-01
6.11928463e-01 1.86657757e-01 -1.91053450e-01 3.23863216e-02
-1.84332454e+00 6.46969914e-01 7.33294308e-01 -3.46719265e-01
-1.16203105e+00 -8.22116077e-01 -9.98082161e-01 -1.12362005e-01
1.01990867e+00 -4.70434248e-01 1.80279350e+00 -8.43737602e-01
-1.13338614e+00 9.36817586e-01 -3.24292094e-01 -7.27440655e-01
3.09253961e-01 -4.83967483e-01 -4.70556974e-01 3.52182657e-01
5.73874176e-01 5.87674737e-01 7.16288149e-01 -9.58939314e-01
-1.09774435e+00 -2.46004716e-01 1.33536756e+00 5.25146425e-01
-1.73655152e-02 3.80982667e-01 -2.14144424e-01 -8.05431843e-01
1.24543257e-01 -6.92022860e-01 1.62895277e-01 -3.54311228e-01
-1.62042171e-01 -9.43739653e-01 1.13187730e+00 -2.26696625e-01
1.53694463e+00 -2.01372719e+00 -4.67690080e-03 -5.67315817e-01
3.06759268e-01 3.25773597e-01 1.81340247e-01 3.48185301e-01
-1.83614925e-01 1.99529082e-01 -9.80620459e-02 4.03252104e-03
-3.00298305e-03 6.40368342e-01 -4.29587513e-01 3.23282450e-01
-8.19862820e-03 9.16918635e-01 -1.50348055e+00 -6.24545991e-01
2.98129141e-01 3.23680997e-01 -3.69428337e-01 8.90968442e-02
-7.60881960e-01 3.08210671e-01 -8.32715869e-01 7.18916237e-01
2.83130974e-01 -4.15608138e-01 4.95035440e-01 -3.04527491e-01
2.68267483e-01 7.88246393e-01 -1.12793255e+00 1.70473528e+00
-2.13511869e-01 3.43926430e-01 -1.96744174e-01 -1.07308936e+00
5.26381314e-01 7.07264066e-01 4.49847996e-01 -6.12570167e-01
1.80648908e-01 1.57119423e-01 2.29376601e-03 -4.69581902e-01
3.31700861e-01 -6.81664348e-02 -2.66164362e-01 4.15102810e-01
1.15412593e-01 1.63074508e-01 6.89661801e-01 3.08894992e-01
1.08769238e+00 4.23965305e-01 4.40261692e-01 -3.56488377e-01
1.02979589e+00 1.53519258e-01 8.40517461e-01 6.30425155e-01
-1.57259077e-01 2.75205463e-01 8.62420917e-01 -8.16804230e-01
-6.44015491e-01 -6.78617179e-01 -1.36189967e-01 1.80966043e+00
6.09053075e-01 -1.01356828e+00 -7.14593589e-01 -1.19184446e+00
-1.38621837e-01 8.32673848e-01 -5.50132394e-01 8.18988234e-02
-9.14489150e-01 -5.21314859e-01 8.35411608e-01 5.67857742e-01
7.87952363e-01 -1.23984694e+00 -1.04626000e+00 3.36315334e-01
-5.39824188e-01 -1.15367067e+00 1.46518648e-01 -8.74316245e-02
-5.46200931e-01 -1.90541065e+00 -1.76573709e-01 -5.69255531e-01
3.62608522e-01 2.53170729e-01 1.40767181e+00 4.65620577e-01
2.70331830e-01 2.67078400e-01 -1.00935853e+00 -6.76049709e-01
-3.19868028e-01 -1.23952422e-02 -2.15228483e-01 -9.13119689e-02
6.04443491e-01 -1.07472926e-01 -3.96879137e-01 5.46337366e-01
-9.74492073e-01 1.16133682e-01 -3.35600704e-01 9.03794050e-01
6.39260888e-01 -1.99472886e-02 5.78359723e-01 -1.35896492e+00
6.54784143e-01 -6.37793601e-01 -4.98987794e-01 4.50410753e-01
-2.56607652e-01 -1.11089379e-01 4.48333442e-01 -2.23986983e-01
-1.47642362e+00 -1.95537642e-01 -8.09426084e-02 -8.16702694e-02
-9.45831239e-02 3.32909286e-01 -2.79889435e-01 5.34988284e-01
8.76351953e-01 -3.52185667e-01 -3.86338443e-01 -2.04878628e-01
3.37093830e-01 2.48301372e-01 2.22972065e-01 -1.12078786e+00
3.82679462e-01 6.74699426e-01 -8.37037116e-02 -5.56918621e-01
-1.64638209e+00 -4.79231477e-01 -4.49570417e-01 -3.00221980e-01
9.82334673e-01 -1.07792580e+00 -5.78246355e-01 2.54188061e-01
-1.24169171e+00 -1.58053443e-01 -3.72388870e-01 1.35829046e-01
-4.03660446e-01 2.52837360e-01 -6.09971285e-01 -6.53591752e-01
-6.14552274e-02 -1.03507459e+00 1.20762634e+00 3.02194118e-01
-3.75305980e-01 -1.09102201e+00 -2.20355839e-01 5.36634564e-01
8.13678745e-03 2.59201050e-01 7.92247236e-01 -9.49185014e-01
-4.80778515e-02 5.93750328e-02 -2.66400236e-03 -1.56487208e-02
-1.66696906e-02 -5.03205657e-01 -1.08301127e+00 7.21378103e-02
1.38018534e-01 -4.56429392e-01 7.39235580e-01 3.02191675e-01
1.09589219e+00 -1.34235442e-01 -4.70394582e-01 -1.47703616e-02
9.71880317e-01 4.49463904e-01 5.64938962e-01 5.98506868e-01
8.12646329e-01 6.16775215e-01 1.09034300e+00 3.87739718e-01
4.64705974e-01 7.18041778e-01 3.56097698e-01 9.72222984e-02
-2.05378175e-01 -3.31241637e-01 1.32046968e-01 -6.71881437e-02
-1.61788002e-01 -4.67317522e-01 -1.08408308e+00 4.35892582e-01
-2.48484421e+00 -1.30862308e+00 -4.29084092e-01 1.55288339e+00
8.77418756e-01 2.94352412e-01 2.39970475e-01 3.97084892e-01
1.10234964e+00 4.77207839e-01 -4.68971938e-01 1.66902989e-01
-2.80199856e-01 1.79423407e-01 1.52703896e-01 5.59215665e-01
-1.54697001e+00 1.24593174e+00 5.01080370e+00 1.05955648e+00
-4.61050570e-01 5.92562675e-01 6.60763979e-01 2.97922473e-02
-3.46086353e-01 3.23451102e-01 -9.54893291e-01 3.33400607e-01
7.78030455e-01 -2.76246578e-01 -2.00882684e-02 1.00275564e+00
-9.50097591e-02 -1.35346010e-01 -1.09612584e+00 8.87658775e-01
-4.82243374e-02 -1.59481204e+00 2.80434284e-02 -5.64616203e-01
3.55753094e-01 -2.46401593e-01 -6.25124693e-01 4.63198751e-01
4.38356817e-01 -7.01794267e-01 1.01383197e+00 6.17445596e-02
5.17704725e-01 -4.36320931e-01 8.06604266e-01 3.07385623e-02
-1.50598383e+00 -4.57882464e-01 -3.98818254e-01 -3.76677185e-01
3.83536488e-01 3.96097779e-01 -5.44594407e-01 6.10553443e-01
6.03226483e-01 1.01990628e+00 -3.83757174e-01 4.48921800e-01
-8.24513435e-01 8.68141711e-01 -2.25273803e-01 -2.31073592e-02
1.90054789e-01 3.94426495e-01 7.13320971e-01 8.56203198e-01
-2.25628003e-01 5.96528351e-01 4.56899434e-01 4.31716710e-01
-1.60535946e-01 6.63176738e-03 -2.68904626e-01 1.49154756e-02
9.09989715e-01 1.10626650e+00 -1.13191032e+00 -6.36350095e-01
-4.72668320e-01 5.73100805e-01 -2.63321735e-02 -7.09621832e-02
-1.20717835e+00 4.34639938e-02 6.27404988e-01 3.29288989e-01
-5.57572320e-02 2.81172067e-01 -4.26797047e-02 -1.19206357e+00
1.53157990e-02 -8.47457767e-01 1.07801819e+00 -5.72719693e-01
-1.01712561e+00 4.54648256e-01 6.42054141e-01 -1.02308667e+00
-5.65335602e-02 -7.02248037e-01 -2.29535460e-01 2.75635362e-01
-1.18902123e+00 -1.34236598e+00 -2.24010766e-01 7.12728381e-01
1.16276598e+00 -1.63428843e-01 7.79982448e-01 2.92077601e-01
-3.25067878e-01 1.15372196e-01 -7.00042427e-01 3.18297029e-01
3.48533690e-01 -1.17901063e+00 3.57236087e-01 7.10392892e-01
1.43456012e-01 5.87039888e-01 5.36969364e-01 -7.12338686e-01
-9.96034801e-01 -8.46268594e-01 1.04637563e+00 -7.52936304e-01
6.90783083e-01 -2.18978003e-01 -7.31252789e-01 6.44330621e-01
1.02162778e-01 1.95813939e-01 7.61575401e-01 1.05112784e-01
-2.94723123e-01 2.17930570e-01 -1.09473169e+00 5.39834082e-01
1.53660178e+00 -6.73962116e-01 -1.16406012e+00 5.28867602e-01
1.11613739e+00 -7.40941226e-01 -6.22873485e-01 5.32314718e-01
4.57092077e-01 -9.03145611e-01 1.13355172e+00 -7.39662111e-01
4.57183629e-01 -5.84159732e-01 -4.14292872e-01 -6.04063570e-01
-8.31070617e-02 -3.60525906e-01 -3.41974616e-01 1.30115688e+00
4.03467387e-01 -2.95701981e-01 5.81588984e-01 5.33189058e-01
-1.75993681e-01 -5.68254769e-01 -9.03466463e-01 -3.51895034e-01
-4.33763474e-01 -5.86675584e-01 7.96536088e-01 1.10520470e+00
6.27246336e-04 9.88720536e-01 6.98564723e-02 -3.07544544e-02
2.64114588e-01 3.43967266e-02 3.16096038e-01 -1.46633005e+00
-1.31317437e-01 -2.12144762e-01 -1.50838539e-01 -6.40511811e-01
7.18055964e-01 -9.30235505e-01 -4.24472779e-01 -1.64531469e+00
-1.01366118e-01 -4.28876132e-01 -2.23911315e-01 8.66088808e-01
-4.89870489e-01 -5.40347844e-02 3.24485928e-01 3.63606125e-01
-1.06258667e+00 3.47345948e-01 1.04848850e+00 -7.67789409e-02
-4.39166129e-02 -1.20194897e-01 -6.03451848e-01 9.91164207e-01
9.61123228e-01 -5.28936088e-01 -1.05214167e+00 -6.16119981e-01
3.75575900e-01 3.35802436e-02 4.29684281e-01 -9.02859688e-01
-4.22786362e-02 -1.46011218e-01 5.57550415e-02 -2.62836248e-01
1.93805665e-01 -6.01406336e-01 4.36573774e-02 2.55636007e-01
-7.05514193e-01 3.09340119e-01 6.50927201e-02 5.62798738e-01
-3.55444998e-01 -6.02776706e-01 3.23892027e-01 -5.44219255e-01
-1.25977516e+00 -7.93759823e-02 -8.63888711e-02 4.81387258e-01
1.04110932e+00 1.20888554e-01 -8.66337895e-01 -1.78438261e-01
-8.41385901e-01 3.81762832e-01 1.92583986e-02 6.54467463e-01
4.69317049e-01 -1.34439790e+00 -6.85827911e-01 -1.82234809e-01
2.88373083e-01 3.45531404e-01 2.19792217e-01 5.90220690e-01
-5.42186439e-01 1.35687545e-01 1.34056270e-01 -5.14171541e-01
-1.56749570e+00 3.05767715e-01 1.97405457e-01 -3.74276251e-01
-5.15768051e-01 1.25650942e+00 1.48750067e-01 -3.63896191e-01
8.00270066e-02 -6.26934320e-02 -6.29954040e-01 4.41187024e-01
7.82136679e-01 3.86169732e-01 -9.52396691e-02 -9.09689069e-01
-4.71955240e-01 6.97816461e-02 3.87688656e-03 -5.52782556e-04
1.18775070e+00 1.43350312e-03 -3.00191939e-01 3.20128947e-01
6.78532362e-01 -2.19242930e-01 -1.04908824e+00 -2.62184441e-01
4.74052280e-01 -5.50505221e-01 -7.98600316e-02 -6.73363268e-01
-5.50340712e-01 4.28245366e-01 4.05146092e-01 5.01842260e-01
9.10155535e-01 4.19709772e-01 6.02054656e-01 4.13513988e-01
3.67115885e-01 -1.12100089e+00 2.15451702e-01 8.46664667e-01
8.04847717e-01 -9.58487689e-01 -1.57808289e-01 -7.64204919e-01
-7.34322250e-01 9.88975763e-01 7.46581912e-01 1.56901777e-02
5.45578420e-01 1.52899206e-01 4.76274751e-02 -5.75644910e-01
-8.29388142e-01 -4.55088377e-01 9.47184637e-02 4.56447065e-01
8.72551262e-01 -1.43079804e-02 -7.13788569e-01 6.26597404e-01
-9.77810621e-02 -1.34460941e-01 1.48144737e-01 1.33387566e+00
-4.88202572e-01 -1.61323309e+00 -3.02754700e-01 1.64479136e-01
-8.50441754e-01 -1.35816604e-01 -3.16675395e-01 8.21986616e-01
4.64817822e-01 1.24852121e+00 -3.35680768e-02 -1.77582562e-01
4.65383768e-01 2.72484630e-01 3.37985605e-01 -9.48107481e-01
-6.79804564e-01 -3.82083026e-03 7.88750291e-01 -6.10561848e-01
-1.22294629e+00 -5.79010963e-01 -1.73119259e+00 -5.45065887e-02
-3.00640076e-01 3.08916122e-01 -8.73607472e-02 1.09726965e+00
1.68630406e-01 6.88312769e-01 1.85049195e-02 -4.62136447e-01
-2.74924725e-01 -8.72550368e-01 -5.37398309e-02 8.73900414e-01
-6.85133925e-03 -1.26942205e+00 -1.62361236e-03 2.54850656e-01] | [10.158230781555176, 9.199433326721191] |
4fb542be-6d9a-4392-9442-81f813af6199 | one-size-does-not-fit-all-the-case-for | 2205.02564 | null | https://arxiv.org/abs/2205.02564v1 | https://arxiv.org/pdf/2205.02564v1.pdf | One Size Does Not Fit All: The Case for Personalised Word Complexity Models | Complex Word Identification (CWI) aims to detect words within a text that a reader may find difficult to understand. It has been shown that CWI systems can improve text simplification, readability prediction and vocabulary acquisition modelling. However, the difficulty of a word is a highly idiosyncratic notion that depends on a reader's first language, proficiency and reading experience. In this paper, we show that personal models are best when predicting word complexity for individual readers. We use a novel active learning framework that allows models to be tailored to individuals and release a dataset of complexity annotations and models as a benchmark for further research. | ['Manuel Tragut', 'Sian Gooding'] | 2022-05-05 | null | https://aclanthology.org/2022.findings-naacl.27 | https://aclanthology.org/2022.findings-naacl.27.pdf | findings-naacl-2022-7 | ['complex-word-identification'] | ['natural-language-processing'] | [ 3.29001337e-01 3.64434719e-01 -3.96164954e-01 -1.66013137e-01
-5.48967838e-01 -5.76446652e-01 6.80380821e-01 1.09358156e+00
-7.13505805e-01 1.93390623e-01 5.86609483e-01 -2.94845730e-01
-4.16625738e-01 -5.41773736e-01 -1.84904635e-01 2.02695206e-01
3.20217311e-01 5.91992259e-01 -2.23968383e-02 -4.48079526e-01
5.58124602e-01 2.98900694e-01 -1.60896969e+00 8.98191035e-02
1.48388982e+00 1.99259102e-01 5.81051946e-01 1.01190996e+00
-4.61325318e-01 9.26092386e-01 -8.96611392e-01 -4.60574538e-01
-1.46152049e-01 -2.32959148e-02 -9.01952147e-01 -2.72851199e-01
7.46532083e-01 3.26952226e-02 -1.41153023e-01 7.53212154e-01
6.19112849e-01 2.40958393e-01 8.15630555e-01 -6.03894472e-01
-7.26057470e-01 1.09038115e+00 4.98856008e-01 3.13287675e-01
6.73315406e-01 2.14287072e-01 1.07420135e+00 -7.10858762e-01
5.37050188e-01 1.05953586e+00 7.70497620e-01 6.80076540e-01
-1.09778225e+00 -3.85753959e-01 5.07082641e-01 4.29454327e-01
-1.29286945e+00 -6.34133816e-01 4.48157161e-01 -6.07599258e-01
1.55214441e+00 6.59375370e-01 1.02470410e+00 1.04460359e+00
5.97783364e-02 1.06155634e+00 7.76259661e-01 -1.09436238e+00
6.37447163e-02 1.97286889e-01 1.07999384e+00 4.29742515e-01
4.58872229e-01 -1.82075456e-01 -8.42604101e-01 2.10645273e-01
8.74860119e-03 -3.38340700e-01 -3.39250714e-01 -1.25382826e-01
-1.05314231e+00 6.98311508e-01 -3.63875538e-01 2.80156225e-01
4.94265482e-02 -4.61889684e-01 1.82864338e-01 4.93270874e-01
6.47871494e-01 1.44401908e+00 -6.61005616e-01 -6.48163915e-01
-8.32535267e-01 4.85102177e-01 1.24978924e+00 1.30237019e+00
3.70244890e-01 -4.15365756e-01 -6.05189264e-01 1.05488455e+00
3.70633751e-01 5.77849329e-01 6.77667856e-01 -3.30909491e-01
3.73654753e-01 1.03873169e+00 -1.35412002e-02 -6.58952892e-01
-6.24283373e-01 -5.34860432e-01 -4.07479942e-01 2.16774270e-01
3.72768998e-01 1.62841707e-01 -6.65934026e-01 1.32124615e+00
-4.14958239e-01 -4.04144466e-01 -2.85615325e-01 1.11645557e-01
1.09098411e+00 2.75215745e-01 3.91643256e-01 -4.22330230e-01
8.79464388e-01 -9.70376134e-01 -1.09418297e+00 -5.85180879e-01
1.13678098e+00 -9.15562153e-01 1.47207463e+00 7.42841899e-01
-1.29085100e+00 -6.17008388e-01 -1.06730378e+00 -4.01721537e-01
-8.21409225e-01 -5.19327745e-02 5.26989698e-01 1.12289906e+00
-8.66995573e-01 5.11990607e-01 -5.61701834e-01 -4.04715568e-01
3.75901580e-01 3.01838338e-01 1.45261297e-02 1.93806127e-01
-1.19729996e+00 1.42427397e+00 4.36268061e-01 -5.89757800e-01
-4.63414520e-01 -1.02583408e+00 -8.48004639e-01 1.44057333e-01
9.79927257e-02 -6.99879467e-01 1.29742110e+00 -9.98819649e-01
-1.50499058e+00 9.78632689e-01 -2.80624460e-02 -3.93939197e-01
3.39536279e-01 -5.52989662e-01 -4.39874381e-01 -5.85264266e-01
-2.72972316e-01 1.80410147e-01 8.81788731e-01 -5.77532887e-01
-5.79638183e-01 -2.43873373e-01 2.25806519e-01 6.35560095e-01
-6.53726339e-01 1.47589132e-01 -3.09325248e-01 -8.64385426e-01
-4.86084282e-01 -6.56407118e-01 1.74983084e-01 -1.66157171e-01
-3.61735284e-01 -8.76614690e-01 1.33140430e-01 -8.21956456e-01
2.28238082e+00 -1.77977788e+00 4.92949069e-01 1.94674984e-01
7.59842277e-01 7.66713619e-01 -1.79252982e-01 4.66859847e-01
2.23399356e-01 6.51558459e-01 3.23737293e-01 -6.92384541e-01
1.30872011e-01 -1.01077102e-01 1.82112679e-01 -4.55395319e-02
-6.66453242e-02 1.00787818e+00 -9.54032540e-01 -6.14537001e-01
2.40167469e-01 2.24954754e-01 -5.49569130e-01 5.44692338e-01
-3.27197820e-01 9.12138894e-02 2.43714988e-01 4.10950959e-01
2.10416690e-01 -9.78044644e-02 -1.99722692e-01 3.52512658e-01
-3.76646191e-01 5.20294547e-01 -9.29641485e-01 1.52024281e+00
-9.21702504e-01 1.07402289e+00 -5.35386443e-01 -4.97770488e-01
6.52746081e-01 6.70322031e-02 -2.66088605e-01 -8.28085780e-01
2.99346223e-02 -1.15145504e-01 3.25588435e-01 -6.23887241e-01
6.93464518e-01 7.12522268e-01 9.15974975e-02 4.39243883e-01
1.43802185e-02 -2.11305201e-01 7.86108077e-01 1.69586420e-01
1.01175404e+00 -3.15915436e-01 6.50875688e-01 -5.09727716e-01
5.08770108e-01 -2.35382706e-01 -2.85373796e-02 1.23380339e+00
-3.97637300e-02 8.94300714e-02 1.10948784e-02 -1.21276908e-01
-1.11794996e+00 -9.52547550e-01 -1.06847905e-01 1.73425150e+00
-2.43688911e-01 -1.08109260e+00 -8.38211358e-01 -6.29911050e-02
-2.02865466e-01 1.33202064e+00 -4.34844524e-01 -4.19405252e-01
-4.98923331e-01 -1.74193621e-01 3.31994832e-01 4.75360453e-01
-1.23007044e-01 -8.57234597e-01 -5.07592976e-01 8.51955116e-02
-5.72846942e-02 -5.77218056e-01 -6.09597325e-01 8.29372853e-02
-5.29565930e-01 -7.59980381e-01 -5.72871625e-01 -1.03416216e+00
4.46665287e-01 4.34946604e-02 1.81428170e+00 6.00982904e-01
-5.04247427e-01 7.93825209e-01 -8.54875863e-01 -1.26011860e+00
-5.51360488e-01 8.31737459e-01 -2.70953849e-02 -7.55580723e-01
4.78495210e-01 1.25800163e-01 -3.69068265e-01 -2.21520901e-01
-5.77115178e-01 2.41721347e-01 4.06581134e-01 6.42728031e-01
2.55295455e-01 6.92694038e-02 -2.01591924e-01 -1.05466223e+00
1.30777514e+00 -1.23234913e-01 -2.39662081e-01 7.78485775e-01
-1.13889790e+00 -3.62057351e-02 4.40871507e-01 -8.00290465e-01
-9.11465108e-01 -1.40815293e-02 -9.55529436e-02 4.15275127e-01
-9.52171236e-02 5.11288881e-01 -3.02738219e-01 -1.39804155e-01
9.04437840e-01 1.04402281e-01 -3.75785790e-02 -5.87322056e-01
3.91516656e-01 7.09321797e-01 -1.24948025e-02 -2.32139245e-01
7.91772425e-01 -5.08483350e-01 -5.61910093e-01 -1.34872043e+00
-1.19603050e+00 -5.73426843e-01 -1.11294460e+00 -2.51538426e-01
5.62923789e-01 -8.93280506e-01 -5.61735749e-01 7.55534530e-01
-1.06253982e+00 -7.50410974e-01 -3.25289518e-01 2.74570733e-01
-1.96450785e-01 3.41909885e-01 8.52550939e-02 -8.73768806e-01
-6.63862646e-01 -8.65307689e-01 6.31536484e-01 3.98885906e-01
-1.37368107e+00 -1.58021033e+00 2.74047494e-01 3.60552043e-01
6.98798060e-01 -4.52952802e-01 1.33782160e+00 -1.10002840e+00
-2.29012534e-01 -2.69794047e-01 3.26844931e-01 3.88352215e-01
-3.27271484e-02 1.46766499e-01 -6.62226379e-01 -2.68785059e-01
-3.07294339e-01 -3.47431064e-01 4.65737700e-01 1.67849243e-01
1.29430902e+00 -4.26276922e-01 -1.70558751e-01 6.18855953e-01
1.05219662e+00 -6.44206405e-02 7.35717118e-01 5.67095518e-01
9.37879086e-01 3.87672961e-01 2.15279296e-01 2.97786146e-01
4.32122707e-01 5.61726153e-01 -2.41253838e-01 2.91218497e-02
-3.80218923e-01 -1.62213087e-01 1.82378873e-01 1.30953383e+00
-8.79586264e-02 -5.67774117e-01 -1.48610234e+00 3.51016670e-01
-1.17047763e+00 -7.68659294e-01 -1.95429653e-01 2.41702390e+00
1.26969230e+00 5.19734085e-01 1.15217581e-01 5.42382121e-01
2.79879570e-01 -7.79327005e-02 -3.58121336e-01 -8.29329550e-01
-2.84449190e-01 7.17254400e-01 5.62295556e-01 1.11271381e+00
-8.47481370e-01 1.06745374e+00 7.17240000e+00 1.07242811e+00
-7.25937784e-01 -5.03601506e-02 3.38182658e-01 5.03354594e-02
-3.54961038e-01 -1.43443465e-01 -1.17233145e+00 4.41204071e-01
1.03154695e+00 -8.88631046e-01 6.52202725e-01 5.98292708e-01
6.14416227e-02 -1.53224878e-02 -1.52574646e+00 1.07247400e+00
4.81990039e-01 -8.11079860e-01 4.27095503e-01 -3.09349954e-01
4.66194898e-01 -4.47245330e-01 1.79909065e-01 5.87824762e-01
3.40698585e-02 -1.50508666e+00 6.62550032e-01 1.23417127e+00
7.65324116e-01 -7.50404656e-01 2.97672987e-01 7.10651457e-01
-7.59681165e-01 -1.88039437e-01 -1.37609839e-01 -7.21261621e-01
-4.58887935e-01 1.02280691e-01 -8.75685155e-01 -3.43936712e-01
1.99292734e-01 4.34064507e-01 -1.51649368e+00 1.22892427e+00
-1.74646214e-01 7.47878134e-01 -3.06221601e-02 -8.15512478e-01
-4.04418230e-01 3.22097003e-01 5.77255368e-01 1.40955460e+00
-1.59246475e-02 -1.39498189e-02 1.15211315e-01 6.10885382e-01
1.85590535e-01 9.31095183e-01 -3.71443927e-01 -4.98578340e-01
7.95334578e-01 8.82315576e-01 -3.31328571e-01 -3.42372358e-01
-5.46405852e-01 1.09947217e+00 6.88814998e-01 -1.49540782e-01
-8.56958106e-02 -6.46116912e-01 5.17976344e-01 2.99236983e-01
-2.05758885e-01 -3.95678759e-01 -1.04898942e+00 -1.28787231e+00
-2.61153162e-01 -1.10896814e+00 -1.80738401e-02 -6.07951462e-01
-1.22064984e+00 2.15597227e-01 3.94907705e-02 -9.92164314e-01
-1.21674918e-01 -1.02597106e+00 -7.16229320e-01 1.00817502e+00
-1.10433948e+00 -8.36067796e-01 -6.68937325e-01 1.92154273e-01
1.00819457e+00 -6.33448064e-01 1.25938559e+00 -7.26637393e-02
-6.23359740e-01 1.28982842e+00 2.96068251e-01 1.21790774e-01
5.47595263e-01 -1.77154660e+00 9.52470481e-01 7.27132082e-01
2.31913373e-01 9.71533537e-01 8.52775156e-01 -7.64561057e-01
-1.30731547e+00 -7.84533858e-01 1.70971107e+00 -1.05087113e+00
6.99023724e-01 -5.52016556e-01 -8.43008041e-01 5.03175199e-01
3.43669832e-01 -9.63921249e-01 1.13053584e+00 6.23122156e-01
-1.60922199e-01 2.42086291e-01 -5.47455907e-01 8.21627736e-01
1.25298429e+00 -6.73420191e-01 -1.00404358e+00 6.37373686e-01
6.05417907e-01 -4.29829597e-01 -9.37274218e-01 -4.66615260e-02
4.47242230e-01 -6.19820774e-01 8.97898853e-01 -5.00026464e-01
-9.55570024e-04 4.97479558e-01 5.17605662e-01 -1.73984420e+00
-8.68266284e-01 -8.63768041e-01 -4.25521135e-01 1.20816028e+00
8.51843178e-01 -3.09602678e-01 3.20036650e-01 1.06194043e+00
-9.96157452e-02 -5.23141146e-01 -4.09528255e-01 -7.20889807e-01
7.14959264e-01 -6.14720404e-01 5.15132010e-01 7.65170336e-01
4.86295879e-01 7.49063313e-01 6.96274042e-02 -2.78248549e-01
4.55335826e-01 -7.35227764e-01 6.87945962e-01 -2.03714228e+00
-6.13727421e-02 -1.00882292e+00 -3.49692643e-01 -1.08760571e+00
3.28460276e-01 -9.78586197e-01 -3.53191942e-01 -1.38937092e+00
1.61342070e-01 -1.01977497e-01 6.66495040e-02 1.90391958e-01
-6.33328676e-01 -2.25330904e-01 2.43800864e-01 6.80079907e-02
-8.12523842e-01 7.17011616e-02 1.05698502e+00 -4.92247164e-01
-4.21340704e-01 2.31640399e-01 -6.98502421e-01 7.30174720e-01
8.58640909e-01 -2.24867612e-02 -7.60166585e-01 -7.68429041e-01
9.22176242e-01 -4.89829570e-01 -2.67588794e-01 -1.10596502e+00
5.09401023e-01 -8.23486317e-03 4.79389280e-01 -4.60799724e-01
-8.06706771e-02 -3.35188985e-01 -4.46306974e-01 3.21320862e-01
-9.48917270e-01 2.07417682e-01 1.04935110e-01 1.73161417e-01
3.63052547e-01 -7.89968371e-01 5.90926647e-01 -2.79288441e-01
-6.91338539e-01 2.01596677e-01 -8.31226647e-01 5.79198599e-01
6.01958871e-01 -1.79060951e-01 -2.72359271e-02 -4.46488053e-01
-7.52369940e-01 1.55090660e-01 6.46839738e-01 6.88187122e-01
5.37745178e-01 -9.26189065e-01 -9.20885384e-01 2.43177325e-01
6.38504684e-01 -1.77167237e-01 -1.06060408e-01 3.10335428e-01
-8.76281440e-01 6.05183542e-01 5.00010587e-02 -2.13704646e-01
-1.79752016e+00 4.39666241e-01 2.36732513e-01 -2.82233447e-01
-5.97854733e-01 1.07476723e+00 -5.48177600e-01 -2.25791425e-01
8.65457773e-01 -4.58501190e-01 -8.46347451e-01 3.88120055e-01
1.08364666e+00 7.19182432e-01 3.01996976e-01 -5.73221266e-01
1.02790385e-01 5.22138119e-01 -7.11133420e-01 2.94502497e-01
9.07385170e-01 -4.70130205e-01 6.54838840e-03 9.92881358e-01
9.99848962e-01 4.89448383e-02 -4.60990667e-01 -5.00076115e-01
4.54132795e-01 -3.33227217e-01 2.51546741e-01 -1.10716724e+00
-1.04034491e-01 6.51073575e-01 9.85503495e-01 2.56231576e-01
7.14621425e-01 -2.88212806e-01 4.14082289e-01 9.99369800e-01
-7.67304674e-02 -1.77683175e+00 1.57626569e-02 1.13136160e+00
8.54199350e-01 -1.25388575e+00 1.79647565e-01 -3.15963298e-01
-4.71421868e-01 1.27718365e+00 5.15153527e-01 3.05026650e-01
7.19787121e-01 1.76577792e-01 2.02407688e-01 1.06831484e-01
-8.88728082e-01 -3.80218148e-01 8.57130706e-01 9.17172372e-01
1.11702371e+00 2.06257865e-01 -6.68381751e-01 7.38025010e-01
-8.83503139e-01 -2.99359202e-01 4.61027056e-01 7.00409293e-01
-5.99355280e-01 -1.13027358e+00 -2.09915504e-01 8.52612615e-01
-8.29606205e-02 -6.59599185e-01 -8.40238631e-01 5.10196984e-01
1.97222605e-01 1.02182984e+00 2.24591538e-01 -1.92842633e-01
4.67690974e-01 5.23274958e-01 6.67522371e-01 -1.15448439e+00
-7.48381078e-01 -7.10799277e-01 2.70348549e-01 -3.86938863e-02
1.46480516e-01 -7.32580423e-01 -7.36934125e-01 -3.54500234e-01
-3.10093075e-01 -1.99771821e-01 4.28720236e-01 1.15809941e+00
-1.67870611e-01 3.44202071e-01 1.30003288e-01 -3.61567497e-01
-4.25895423e-01 -1.47841465e+00 -3.83705974e-01 6.22080147e-01
1.98296458e-01 -3.78939748e-01 -4.80015844e-01 1.41539529e-01] | [10.83163833618164, 10.377511978149414] |
11c9b13b-ec79-4264-b4e5-bf5b396b096a | anyface-free-style-text-to-face-synthesis-and | 2203.15334 | null | https://arxiv.org/abs/2203.15334v1 | https://arxiv.org/pdf/2203.15334v1.pdf | AnyFace: Free-style Text-to-Face Synthesis and Manipulation | Existing text-to-image synthesis methods generally are only applicable to words in the training dataset. However, human faces are so variable to be described with limited words. So this paper proposes the first free-style text-to-face method namely AnyFace enabling much wider open world applications such as metaverse, social media, cosmetics, forensics, etc. AnyFace has a novel two-stream framework for face image synthesis and manipulation given arbitrary descriptions of the human face. Specifically, one stream performs text-to-face generation and the other conducts face image reconstruction. Facial text and image features are extracted using the CLIP (Contrastive Language-Image Pre-training) encoders. And a collaborative Cross Modal Distillation (CMD) module is designed to align the linguistic and visual features across these two streams. Furthermore, a Diverse Triplet Loss (DT loss) is developed to model fine-grained features and improve facial diversity. Extensive experiments on Multi-modal CelebA-HQ and CelebAText-HQ demonstrate significant advantages of AnyFace over state-of-the-art methods. AnyFace can achieve high-quality, high-resolution, and high-diversity face synthesis and manipulation results without any constraints on the number and content of input captions. | ['Zhenan Sun', 'Min Ren', 'Muyi Sun', 'Qi Li', 'Qiyao Deng', 'Jianxin Sun'] | 2022-03-29 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Sun_AnyFace_Free-Style_Text-To-Face_Synthesis_and_Manipulation_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Sun_AnyFace_Free-Style_Text-To-Face_Synthesis_and_Manipulation_CVPR_2022_paper.pdf | cvpr-2022-1 | ['text-to-face-generation'] | ['computer-vision'] | [ 1.55630782e-01 -4.44106124e-02 -1.13123832e-02 -6.39544368e-01
-8.70148361e-01 -3.81417841e-01 8.05460036e-01 -8.91843617e-01
7.56386593e-02 6.35359585e-01 2.68384606e-01 3.02229166e-01
3.10835749e-01 -6.26591742e-01 -7.18904018e-01 -8.19409490e-01
4.39843148e-01 4.11079675e-01 -6.20990276e-01 -2.44899020e-01
-1.41240895e-01 7.14131415e-01 -1.97543335e+00 7.13738382e-01
4.53554273e-01 1.20026505e+00 7.05922469e-02 4.10956562e-01
-2.97320396e-01 4.10440087e-01 -3.38646680e-01 -9.85959411e-01
5.18459916e-01 -5.20301044e-01 -2.35755682e-01 4.21068221e-01
1.08044434e+00 -4.08500731e-01 -4.00672853e-01 9.41720068e-01
1.04558599e+00 -5.99953942e-02 8.01938653e-01 -1.53175628e+00
-1.04336953e+00 1.44987136e-01 -8.06939662e-01 -4.51753259e-01
4.37511772e-01 5.25565565e-01 6.49383605e-01 -1.44173872e+00
8.04871559e-01 1.94748592e+00 3.57709020e-01 1.09177482e+00
-1.15052569e+00 -1.21941864e+00 -2.70118546e-02 -6.20331615e-03
-1.54161465e+00 -1.26919770e+00 7.04717696e-01 -3.90331328e-01
4.85220224e-01 2.04548106e-01 3.62937868e-01 1.33342612e+00
-8.62048790e-02 5.25691867e-01 8.89768541e-01 -3.90440524e-01
-3.58634800e-01 2.99905360e-01 -7.52681732e-01 1.01663733e+00
-1.76963121e-01 2.17722416e-01 -9.96768475e-01 4.87314798e-02
7.87284553e-01 -1.81382000e-01 -1.68720216e-01 -8.81458521e-02
-9.96788323e-01 8.73851359e-01 1.09133907e-01 -9.26701725e-02
-1.63082376e-01 9.22855511e-02 4.91005152e-01 3.36023688e-01
5.49690902e-01 -7.14044720e-02 1.09923352e-02 2.57770419e-01
-1.05040789e+00 4.05250937e-01 4.85483557e-01 1.28701878e+00
6.56327724e-01 3.09878200e-01 -5.81485331e-01 1.24044609e+00
4.43553656e-01 1.01386797e+00 3.09598356e-01 -8.61071408e-01
4.43969905e-01 2.06613600e-01 -2.14677691e-01 -8.93558681e-01
2.08276555e-01 3.34094657e-04 -8.40656161e-01 2.17781723e-01
-1.31762773e-01 3.49503150e-03 -1.06722641e+00 1.78469670e+00
5.80762386e-01 2.81666014e-02 -2.33672351e-01 8.52240264e-01
1.16226053e+00 7.07162559e-01 6.19656267e-03 -3.44418973e-01
1.43035555e+00 -9.48374689e-01 -8.76198590e-01 -5.71751222e-03
-1.14957117e-01 -1.17305553e+00 1.02049518e+00 1.34799361e-01
-1.33084071e+00 -6.40094042e-01 -6.20263398e-01 -4.49159443e-01
-1.82625845e-01 4.50499594e-01 3.64859164e-01 7.99464524e-01
-1.02941120e+00 3.91714573e-02 -2.19810873e-01 -1.02232218e-01
9.37457860e-01 4.49414343e-01 -8.30982447e-01 -4.51261312e-01
-8.68136525e-01 7.21428871e-01 -1.22050472e-01 1.05863720e-01
-1.04123056e+00 -8.74792159e-01 -1.12359071e+00 -1.62536025e-01
1.21673107e-01 -9.47873473e-01 9.39416111e-01 -1.26337767e+00
-1.71972609e+00 1.31260908e+00 -5.54109216e-01 1.96412325e-01
6.29557550e-01 7.00739026e-02 -3.88779849e-01 1.70997977e-01
3.69937748e-01 1.21232653e+00 1.60746777e+00 -1.47814178e+00
-3.94310415e-01 -3.83931190e-01 -3.98858279e-01 1.91572338e-01
-4.42605555e-01 3.27185750e-01 -8.35216105e-01 -8.71491671e-01
-3.95214260e-01 -7.91932285e-01 4.42086250e-01 7.60007322e-01
-4.63478088e-01 -1.60031348e-01 1.04587007e+00 -6.96284473e-01
8.35603595e-01 -2.38956904e+00 2.24545285e-01 6.81411754e-03
4.71280813e-02 2.71650910e-01 -6.33222461e-01 2.51825541e-01
-2.70268857e-01 -4.88726832e-02 -8.72099251e-02 -9.23810840e-01
2.00195283e-01 5.42964833e-03 -3.72727752e-01 5.48375905e-01
4.85792756e-01 8.79516602e-01 -3.36469263e-01 -8.05005968e-01
2.05886900e-01 8.76669049e-01 -8.37390423e-01 3.23122352e-01
-2.90606499e-01 5.35342276e-01 -1.47034168e-01 9.81200278e-01
9.85576630e-01 9.74721760e-02 -1.14381582e-01 -5.22678435e-01
1.33038074e-01 -4.08442974e-01 -7.54082739e-01 1.75159538e+00
-7.73139179e-01 5.11190474e-01 2.84633666e-01 -4.94068891e-01
7.55661428e-01 3.60955000e-01 4.49281007e-01 -7.03093112e-01
3.11610967e-01 2.64566451e-01 -2.45082766e-01 -6.78655446e-01
2.24832743e-01 -4.77270722e-01 2.88449496e-01 2.52877057e-01
3.59438986e-01 -2.69159347e-01 2.12380961e-01 -4.94249091e-02
2.28188962e-01 2.74863571e-01 -1.85095116e-01 -5.21767251e-02
7.32746899e-01 -5.86790681e-01 3.96165311e-01 -1.00637510e-01
1.10180818e-01 9.77343142e-01 2.65334547e-01 -1.39105752e-01
-1.18229711e+00 -1.01218259e+00 -2.32254624e-01 1.05647802e+00
-1.47919938e-01 -3.32049996e-01 -9.09258842e-01 -4.95980799e-01
1.95016623e-01 4.67265338e-01 -6.78098917e-01 -2.68857982e-02
-4.20690864e-01 -3.35952967e-01 6.67232931e-01 3.57910022e-02
5.87205052e-01 -1.03646588e+00 5.72735108e-02 -2.59286910e-01
-2.73766339e-01 -1.25454938e+00 -1.18134248e+00 -8.17955792e-01
-1.82146639e-01 -8.60632956e-01 -9.33613002e-01 -1.08440614e+00
9.61467505e-01 2.86120195e-02 8.88586223e-01 -5.19268028e-02
-7.42442966e-01 3.30761850e-01 -1.31305605e-01 -2.93216258e-01
-3.18916053e-01 -4.55413580e-01 2.03719527e-01 7.37906337e-01
-1.31577477e-02 -3.76101583e-01 -5.90270638e-01 2.30494425e-01
-8.92812431e-01 1.20162025e-01 4.53764141e-01 1.07210994e+00
5.54028749e-01 -8.32529590e-02 4.85869169e-01 -5.84628344e-01
3.81545216e-01 -3.31798524e-01 -3.69209260e-01 3.52286011e-01
-2.37566724e-01 -6.60082921e-02 5.51758647e-01 -6.63838267e-01
-1.18914199e+00 6.25267848e-02 -2.06764311e-01 -1.04332435e+00
1.78525731e-01 -1.02635853e-01 -7.70821273e-01 -2.78177828e-01
3.71222675e-01 3.94571215e-01 4.09683675e-01 -2.93217450e-01
6.01324737e-01 8.02792788e-01 5.46537220e-01 -7.32682228e-01
9.99684751e-01 5.41014731e-01 1.01030227e-02 -9.94962454e-01
-4.23620015e-01 1.10254176e-01 -2.64756531e-01 -3.68099898e-01
7.66843319e-01 -1.15995920e+00 -9.56646085e-01 6.58121765e-01
-1.15910280e+00 -2.67442949e-02 -1.28464028e-01 9.20775160e-02
-6.77694798e-01 2.67446972e-02 -3.13462853e-01 -7.62270927e-01
-4.27515209e-01 -1.37436295e+00 1.63526165e+00 1.57460555e-01
2.16499984e-01 -6.24190152e-01 -3.91391724e-01 6.83386922e-01
4.03554410e-01 2.08199263e-01 7.37791061e-01 7.11619928e-02
-6.67206466e-01 1.04324207e-01 -3.69577497e-01 4.84586328e-01
2.71769077e-01 1.55461162e-01 -1.21137643e+00 -4.28208292e-01
-3.16160589e-01 -6.25103474e-01 7.53872037e-01 1.10163309e-01
1.31330061e+00 -6.54050827e-01 -2.61803806e-01 9.57593739e-01
1.17628658e+00 -9.89448745e-03 7.81342566e-01 -4.06090915e-01
8.64256203e-01 8.95812988e-01 4.63838667e-01 5.24663627e-01
3.14552963e-01 8.90950739e-01 2.21395805e-01 3.11198570e-02
-7.37430334e-01 -4.29287821e-01 6.77422285e-01 5.52751482e-01
2.81195998e-01 -3.05673420e-01 -3.77485663e-01 4.92544621e-01
-1.20848417e+00 -1.18433118e+00 3.63143623e-01 1.89932442e+00
1.15753269e+00 -6.24031961e-01 -6.76209927e-02 -2.27485090e-01
9.02300477e-01 4.28713001e-02 -6.69569314e-01 -1.47512376e-01
-3.82587761e-01 3.01979542e-01 2.46050465e-03 5.05631328e-01
-8.87608826e-01 1.03123581e+00 5.64525318e+00 1.43807316e+00
-1.47561550e+00 2.26895571e-01 9.20197368e-01 -6.94509745e-01
-4.88861591e-01 -4.88932222e-01 -9.54808652e-01 5.64804196e-01
5.15209198e-01 -1.08603373e-01 6.04783893e-01 5.77075124e-01
1.70242593e-01 3.28493714e-01 -1.11359024e+00 1.64527404e+00
5.34485817e-01 -1.37553394e+00 6.01020873e-01 1.15922026e-01
8.55854154e-01 -4.60044861e-01 6.24146581e-01 6.97263852e-02
-1.69822484e-01 -1.41695261e+00 1.03151548e+00 6.23234272e-01
1.87715292e+00 -9.02736664e-01 1.27940521e-01 1.17329992e-02
-1.08992839e+00 -5.47510460e-02 -1.68297306e-01 7.55692542e-01
8.80482495e-02 4.26236659e-01 -5.61729610e-01 4.99293715e-01
5.56005895e-01 5.99527121e-01 -2.61431664e-01 4.67280120e-01
2.22705394e-01 -3.06792054e-02 -2.40098998e-01 2.90557504e-01
-1.43363833e-01 -5.72571494e-02 4.95668411e-01 9.42133605e-01
6.23754323e-01 6.46273568e-02 5.54057062e-02 9.72965896e-01
-4.90781516e-01 2.86749184e-01 -6.76074862e-01 -2.19117656e-01
4.36999887e-01 1.31488204e+00 -1.08433338e-02 -1.03629537e-01
-4.71988410e-01 1.15944111e+00 5.48459664e-02 1.35207355e-01
-8.36167634e-01 -2.67443210e-01 9.94417787e-01 4.63873476e-01
3.39547902e-01 1.29353389e-01 -5.13378270e-02 -1.17249131e+00
8.64862055e-02 -1.18921351e+00 -9.31398496e-02 -9.14740860e-01
-1.62893915e+00 7.93912947e-01 5.14613539e-02 -1.14331448e+00
-2.82838553e-01 -6.86483026e-01 -3.13216537e-01 1.08351016e+00
-1.59023070e+00 -1.96195853e+00 -2.08094850e-01 1.11458004e+00
7.76148617e-01 -8.07299614e-01 6.82667673e-01 7.00707495e-01
-4.97195810e-01 1.23777056e+00 -1.53875992e-01 1.22524485e-01
1.06043899e+00 -4.66889828e-01 2.09346876e-01 5.26173890e-01
-1.63764939e-01 6.17698789e-01 4.06097382e-01 -6.37405813e-01
-1.70692396e+00 -1.45529044e+00 7.53863096e-01 -1.54383227e-01
1.73244730e-01 -6.06969893e-01 -3.60465527e-01 3.92501235e-01
3.23670924e-01 3.25501591e-01 5.82586050e-01 -5.17044902e-01
-6.84018016e-01 -3.51723731e-01 -1.55303407e+00 7.97857344e-01
1.22645271e+00 -6.67797983e-01 1.87763900e-01 4.87064421e-01
5.34286559e-01 -4.14768726e-01 -7.35244155e-01 3.23167115e-01
8.18951070e-01 -9.01615381e-01 9.81418490e-01 -3.49292636e-01
6.32926345e-01 -1.93520442e-01 -3.55485678e-01 -1.07114708e+00
3.56398267e-03 -9.68776047e-01 1.06531970e-01 1.64420509e+00
9.81522352e-02 -5.55104911e-01 5.74688196e-01 3.00055563e-01
-8.18385184e-02 -6.83723390e-01 -9.32878375e-01 -6.02596104e-01
1.34645537e-01 -1.24518931e-01 9.31042075e-01 8.34641933e-01
-4.37860489e-01 3.12921196e-01 -6.61245048e-01 -1.12302318e-01
8.55278909e-01 4.09468673e-02 8.65291059e-01 -7.63142049e-01
4.20651138e-02 -4.35615271e-01 -2.40654007e-01 -8.29972029e-01
6.84108555e-01 -1.05803573e+00 -3.76793519e-02 -8.54490101e-01
2.42958367e-01 -4.56355661e-01 5.05852461e-01 5.16203165e-01
8.55024159e-02 6.86717927e-01 3.06528121e-01 4.06374373e-02
-1.68037675e-02 1.00718784e+00 1.78739417e+00 -2.64024824e-01
2.97884047e-01 -3.82323980e-01 -7.39710271e-01 3.03515226e-01
3.30046624e-01 -1.62133500e-01 -5.33511102e-01 -5.73921800e-01
-1.61971748e-01 1.90494239e-01 4.37521756e-01 -6.60995781e-01
1.42967373e-01 -3.18292737e-01 5.52330434e-01 -3.05869699e-01
8.11045110e-01 -6.70999706e-01 2.96295911e-01 1.64286550e-02
-3.03522855e-01 -9.89186391e-02 2.17830122e-01 2.68607438e-01
-2.53760844e-01 1.60365194e-01 1.25785494e+00 -1.51613697e-01
-2.61339754e-01 1.02199841e+00 1.63291037e-01 8.39768946e-02
1.13345778e+00 -2.29104385e-01 -2.70772725e-01 -5.40242910e-01
-5.83984435e-01 6.03881739e-02 4.45873052e-01 7.32000887e-01
9.24485207e-01 -1.81418109e+00 -1.25834429e+00 8.11031580e-01
2.43058473e-01 -1.11726895e-01 4.79873300e-01 4.22104806e-01
-3.81191581e-01 2.20316157e-01 -3.86593193e-01 -4.03171808e-01
-1.54358292e+00 3.74762923e-01 3.98247957e-01 3.90152544e-01
-4.06649947e-01 1.09922206e+00 5.36104321e-01 -3.95653427e-01
2.26038605e-01 5.49562097e-01 7.41775706e-02 1.80262968e-01
7.76120543e-01 8.42060223e-02 -5.51673137e-02 -1.32732725e+00
-1.68464601e-01 8.65634799e-01 -5.96464016e-02 -3.49034220e-01
1.19060779e+00 -1.81823730e-01 -3.55345428e-01 -9.63785648e-02
1.54774964e+00 3.55852842e-02 -1.21905303e+00 -1.56286001e-01
-8.34505141e-01 -9.34878469e-01 -1.10450886e-01 -6.48232996e-01
-1.61591637e+00 1.05160844e+00 6.18292987e-01 -5.24746597e-01
1.33530426e+00 -1.85823873e-01 8.66123557e-01 -6.10799007e-02
2.77984798e-01 -8.06503773e-01 3.55421245e-01 2.35179752e-01
1.55790961e+00 -1.11947358e+00 -1.57960013e-01 -6.46685243e-01
-8.34172249e-01 1.04531920e+00 6.41750157e-01 2.12977558e-01
7.22270072e-01 3.51427644e-01 -4.30107675e-02 1.65972300e-02
-9.82932031e-01 -5.28545901e-02 4.49503660e-01 8.23178113e-01
3.04683179e-01 -2.00732887e-01 1.46818534e-01 3.15759510e-01
-4.03867096e-01 -4.32164669e-02 8.26351866e-02 4.70547467e-01
-5.75522445e-02 -1.15945172e+00 -5.75215697e-01 3.41917872e-01
-4.28239763e-01 -2.07828969e-01 -1.68591797e-01 4.25916523e-01
4.90189612e-01 8.70103121e-01 5.17303906e-02 -2.45400831e-01
1.69252589e-01 1.09909303e-01 9.47578132e-01 -6.32897794e-01
-4.34319913e-01 2.09945843e-01 -2.48492993e-02 -5.47753274e-01
-2.57011831e-01 -6.74142361e-01 -8.80349398e-01 -6.60757601e-01
-2.33581424e-01 -2.93334186e-01 8.32668304e-01 5.60378253e-01
5.62176228e-01 1.44770190e-01 9.92582321e-01 -1.19204521e+00
-3.73082370e-01 -7.96365619e-01 -5.65577209e-01 5.42378843e-01
5.04879951e-01 -9.07584727e-01 -1.39988437e-01 5.85930943e-01] | [12.60246753692627, -0.13444356620311737] |
ab431c23-9415-4d9e-9385-a3db2edd82b8 | dependency-aware-self-training-for-entity | 2211.16101 | null | https://arxiv.org/abs/2211.16101v1 | https://arxiv.org/pdf/2211.16101v1.pdf | Dependency-aware Self-training for Entity Alignment | Entity Alignment (EA), which aims to detect entity mappings (i.e. equivalent entity pairs) in different Knowledge Graphs (KGs), is critical for KG fusion. Neural EA methods dominate current EA research but still suffer from their reliance on labelled mappings. To solve this problem, a few works have explored boosting the training of EA models with self-training, which adds confidently predicted mappings into the training data iteratively. Though the effectiveness of self-training can be glimpsed in some specific settings, we still have very limited knowledge about it. One reason is the existing works concentrate on devising EA models and only treat self-training as an auxiliary tool. To fill this knowledge gap, we change the perspective to self-training to shed light on it. In addition, the existing self-training strategies have limited impact because they introduce either much False Positive noise or a low quantity of True Positive pseudo mappings. To improve self-training for EA, we propose exploiting the dependencies between entities, a particularity of EA, to suppress the noise without hurting the recall of True Positive mappings. Through extensive experiments, we show that the introduction of dependency makes the self-training strategy for EA reach a new level. The value of self-training in alleviating the reliance on annotation is actually much higher than what has been realised. Furthermore, we suggest future study on smart data annotation to break the ceiling of EA performance. | ['Guido Zuccon', 'Wen Hua', 'Tiancheng Lan', 'Bing Liu'] | 2022-11-29 | null | null | null | null | ['entity-alignment', 'entity-alignment'] | ['knowledge-base', 'natural-language-processing'] | [ 1.69720620e-01 5.61464667e-01 -4.02678877e-01 -3.08417916e-01
-3.68761897e-01 -4.44483399e-01 4.78435993e-01 3.08329612e-01
-5.64912915e-01 8.64745796e-01 1.53789833e-01 -2.42753550e-01
-1.93889260e-01 -1.07668996e+00 -9.62560713e-01 -4.34087396e-01
-1.43091129e-02 3.95572066e-01 4.49262619e-01 -4.52348232e-01
-9.19671282e-02 9.45302770e-02 -1.66004610e+00 3.35000128e-01
1.25555634e+00 6.83296800e-01 1.79514721e-01 1.15756385e-01
-3.10233086e-01 7.33755946e-01 -4.93904024e-01 -8.29797745e-01
1.76825732e-01 -3.73081714e-01 -8.70665789e-01 -4.86524522e-01
1.30738407e-01 2.09099110e-02 1.32826090e-01 1.24320483e+00
5.42089105e-01 -1.90563023e-01 4.30342197e-01 -1.47036970e+00
-5.33959627e-01 1.20425081e+00 -4.58561450e-01 6.14416376e-02
1.58294857e-01 -1.68351889e-01 1.13187253e+00 -8.11915100e-01
6.89499378e-01 7.75256813e-01 1.13168621e+00 5.49158096e-01
-9.69369829e-01 -8.18358660e-01 2.45312884e-01 3.94152105e-01
-1.48224437e+00 -4.23040718e-01 8.04692566e-01 -1.64315090e-01
1.04422164e+00 5.04239261e-01 4.99242872e-01 1.11726570e+00
-2.29120061e-01 8.18278611e-01 1.05609512e+00 -7.34405994e-01
1.67212233e-01 5.67984283e-01 2.17828289e-01 5.64010978e-01
5.85487783e-01 3.83366970e-03 -4.14357483e-01 -1.54980356e-02
5.23797393e-01 -4.94459540e-01 -4.77475971e-01 -5.64618230e-01
-1.06471121e+00 5.84190667e-01 4.61087435e-01 6.97533429e-01
-2.95657277e-01 -3.31703603e-01 4.68892306e-01 3.84249151e-01
4.88782197e-01 7.97208011e-01 -7.41478443e-01 -4.99179736e-02
-8.30653965e-01 9.53149721e-02 9.30395544e-01 9.26009774e-01
9.70911264e-01 -1.82260007e-01 -1.14890404e-01 8.09643924e-01
2.80676242e-02 1.59557730e-01 5.45550048e-01 -2.96357393e-01
5.09216189e-01 1.10365772e+00 -2.66827848e-02 -1.10764229e+00
-5.79085231e-01 -8.12046528e-01 -8.12147617e-01 -5.86852096e-02
4.53404516e-01 -3.19249451e-01 -7.86112309e-01 2.10025406e+00
3.83141428e-01 4.28790718e-01 1.18209481e-01 7.52261579e-01
7.48775125e-01 9.93190408e-02 1.96135476e-01 -2.67267615e-01
1.10773385e+00 -9.93663132e-01 -8.23857367e-01 -2.92217046e-01
1.19577622e+00 -5.10743856e-01 9.45612073e-01 1.57436326e-01
-7.94444978e-01 -5.11511862e-01 -1.17050648e+00 2.32035041e-01
-8.43206823e-01 1.07201464e-01 8.99699152e-01 8.60638916e-01
-8.31131697e-01 6.97038829e-01 -7.50899732e-01 -4.74691898e-01
2.11387902e-01 5.67614794e-01 -6.43626511e-01 2.39651635e-01
-1.66793430e+00 1.29361618e+00 9.44395304e-01 2.09029526e-01
-2.03186832e-02 -7.76931584e-01 -6.28759384e-01 1.24540150e-01
8.61556232e-01 -8.64294946e-01 8.72338533e-01 -1.19352603e+00
-1.06981695e+00 6.60692990e-01 2.00242385e-01 -6.01565778e-01
6.51009858e-01 -2.91894615e-01 -7.93605506e-01 -4.58920181e-01
5.20685874e-02 6.51903033e-01 4.25833255e-01 -1.45759344e+00
-7.41409957e-01 -2.49762967e-01 1.91807151e-01 2.34152585e-01
-7.50471413e-01 -1.84526637e-01 -5.75484216e-01 -5.77722013e-01
-1.15093246e-01 -8.70917857e-01 -1.89123631e-01 -7.33048201e-01
-2.59264678e-01 -1.77489921e-01 5.59782267e-01 -5.26000559e-01
1.54068744e+00 -1.93477798e+00 -3.50782685e-02 2.46724740e-01
1.15356326e-01 6.69627130e-01 -2.89625414e-02 4.98634100e-01
-1.91139638e-01 2.31066212e-01 -2.24809885e-01 -2.66114712e-01
-5.01144677e-02 3.70477706e-01 -1.90264344e-01 -5.88575304e-02
3.86349201e-01 1.03360963e+00 -1.00798190e+00 -4.68487591e-01
7.73501489e-03 3.72883320e-01 -5.46455383e-01 -2.01951563e-02
-1.66266873e-01 2.07876191e-01 -2.83091456e-01 4.19543922e-01
8.52572739e-01 -2.04351455e-01 3.68514240e-01 -4.49166000e-01
3.74001749e-02 2.68424004e-01 -1.60902488e+00 1.52618408e+00
-2.83772767e-01 1.86297610e-01 -9.89430472e-02 -1.07280755e+00
7.59762168e-01 1.25368491e-01 4.13915515e-01 -6.62494481e-01
1.13581069e-01 3.41066092e-01 1.26723185e-01 -3.93820405e-01
8.08755815e-01 -1.02100886e-01 1.56130731e-01 9.41911489e-02
1.98365688e-01 7.02504694e-01 2.37664133e-01 4.74834219e-02
9.85385239e-01 3.01699668e-01 5.64503908e-01 -1.06572226e-01
5.04422069e-01 2.35014156e-01 7.27283359e-01 9.03367817e-01
8.99467338e-03 3.00789326e-01 1.45518616e-01 -1.49212748e-01
-8.61265421e-01 -6.29378974e-01 -5.24392985e-02 1.16216433e+00
2.70867437e-01 -7.81047165e-01 -8.39203596e-01 -1.17529726e+00
-3.04745995e-02 8.02888274e-01 -5.80826700e-01 -5.00301182e-01
-7.33265042e-01 -1.15324521e+00 7.26517320e-01 7.10685968e-01
5.50666273e-01 -1.02548778e+00 -5.69891810e-01 2.24487469e-01
-2.18726590e-01 -1.20558226e+00 1.26409590e-01 5.30572474e-01
-8.07938337e-01 -1.07380998e+00 -3.21999043e-01 -5.96250653e-01
6.99533284e-01 8.78027454e-02 1.24444056e+00 1.05018489e-01
2.06396535e-01 1.63398981e-01 -5.68827093e-01 -6.60566390e-01
-5.70097625e-01 5.64284146e-01 1.61605656e-01 -1.94201112e-01
7.24435747e-01 -6.80308223e-01 -2.23503768e-01 5.19918919e-01
-6.36273563e-01 2.99677312e-01 9.07030344e-01 8.47436488e-01
3.28475386e-01 3.32077295e-01 9.90030169e-01 -1.38112342e+00
6.12299144e-01 -4.85196859e-01 -2.09394410e-01 5.33305883e-01
-1.09562576e+00 2.58788526e-01 5.98029673e-01 -4.72202927e-01
-1.10287297e+00 -3.93184908e-02 -2.17391744e-01 -2.41538659e-01
-6.79197237e-02 7.83353806e-01 -4.31586444e-01 -5.45166098e-02
7.80007362e-01 -3.93935554e-02 -1.80241227e-01 -4.41402942e-01
4.81944352e-01 5.50079167e-01 6.38720930e-01 -6.00215495e-01
7.26608753e-01 1.08524449e-01 -2.00163633e-01 -2.05239758e-01
-8.81690502e-01 -4.76636410e-01 -6.81468487e-01 8.81404895e-03
5.30737102e-01 -9.45970058e-01 -3.59507143e-01 1.00326017e-01
-7.57701099e-01 -1.08572356e-01 -4.08205807e-01 3.83552492e-01
-1.34706780e-01 4.46712464e-01 -1.93845972e-01 -7.82551825e-01
-2.88823843e-01 -7.29425490e-01 7.20886827e-01 1.99785426e-01
-2.25896969e-01 -9.65362847e-01 4.62219939e-02 1.72774240e-01
5.42854607e-01 -1.24731055e-02 7.61987925e-01 -1.17018962e+00
-3.97339344e-01 -2.14179263e-01 -1.64650485e-01 3.35998476e-01
2.01667503e-01 -3.44374776e-01 -1.09227765e+00 -2.08583102e-01
-4.96581532e-02 -1.73273578e-01 9.67928469e-01 -1.54799864e-01
8.86916757e-01 -1.41737863e-01 -6.39385283e-01 3.40751290e-01
1.30904734e+00 6.36057854e-02 6.16074920e-01 6.85104728e-01
7.54623115e-01 6.11663640e-01 8.15837324e-01 8.88229311e-02
6.24753118e-01 9.36228275e-01 3.67983282e-01 -3.45154881e-01
-1.57359287e-01 -4.39562112e-01 1.25989109e-01 8.20221305e-01
-3.33933264e-01 -2.46288151e-01 -9.46620524e-01 4.82450187e-01
-2.13607168e+00 -7.76081145e-01 -2.10215986e-01 2.26296520e+00
9.98002231e-01 3.77933383e-01 5.98085709e-02 4.21793967e-01
7.38490045e-01 -1.76096484e-01 -1.75536573e-01 -5.16031943e-02
-3.64279896e-01 4.80384529e-02 7.30293870e-01 2.38310501e-01
-1.25707400e+00 9.99824405e-01 5.59429359e+00 8.78207207e-01
-1.18587267e+00 1.28981143e-01 1.86498299e-01 4.50093478e-01
-2.61193782e-01 6.35354817e-02 -9.99293745e-01 4.52238351e-01
8.34377944e-01 1.44794239e-02 9.61282328e-02 9.36172128e-01
-4.17429775e-01 -7.67693222e-02 -1.10124314e+00 7.71725774e-01
8.83104652e-02 -1.17654371e+00 2.23984523e-03 -6.37670383e-02
5.06621540e-01 -1.22577891e-01 -3.50666255e-01 7.88123965e-01
3.56953919e-01 -8.33043575e-01 4.22530651e-01 3.71760011e-01
3.83952260e-01 -7.48079956e-01 1.23317778e+00 6.08179390e-01
-8.93079281e-01 -3.70825194e-02 -3.49261165e-01 -1.07808977e-01
9.35142785e-02 7.18569875e-01 -1.17905033e+00 1.27908254e+00
5.75142145e-01 3.53132635e-01 -9.50225949e-01 9.79459226e-01
-2.35823900e-01 5.94376266e-01 -4.06463563e-01 2.24444300e-01
-8.68497565e-02 9.17678047e-03 5.21562278e-01 1.40165019e+00
2.59340763e-01 -1.88633740e-01 -5.50172627e-02 6.79279983e-01
-7.00203702e-02 2.99840748e-01 -6.22312009e-01 1.27509817e-01
5.26784718e-01 1.28224075e+00 -7.39185393e-01 -3.12411606e-01
-3.92280400e-01 9.34230506e-01 6.03253424e-01 1.64394975e-02
-8.41111779e-01 -1.32131860e-01 2.30789557e-01 3.13614495e-02
3.31320643e-01 1.88817799e-01 -4.94400024e-01 -1.19833291e+00
2.23319158e-01 -7.40204573e-01 5.16013622e-01 -4.98047084e-01
-1.25288212e+00 5.60839772e-01 1.49694592e-01 -1.08810079e+00
-3.45157683e-01 -3.07817996e-01 -3.59663695e-01 4.90047991e-01
-1.49197912e+00 -1.34867537e+00 -4.09218043e-01 3.45269531e-01
2.43013613e-02 -6.06766082e-02 8.97897780e-01 7.06105888e-01
-6.35435343e-01 9.86015797e-01 -2.15113312e-01 1.02589488e-01
1.09670436e+00 -1.43214595e+00 3.11142921e-01 9.12240922e-01
3.38107526e-01 7.96215892e-01 9.27897096e-01 -8.73569667e-01
-1.09298635e+00 -1.09814322e+00 9.48355079e-01 -6.82388663e-01
4.94987518e-01 -4.09034222e-01 -1.27480555e+00 7.33944476e-01
2.27794256e-02 -6.41756132e-02 6.51452005e-01 6.94751143e-01
-2.97623396e-01 -5.82289509e-02 -9.83928025e-01 5.90828896e-01
1.37032521e+00 -2.72855878e-01 -6.69894099e-01 -1.03398196e-01
5.84289193e-01 -4.28453177e-01 -9.20728207e-01 1.00669408e+00
4.35211688e-01 -9.69094336e-01 8.85060370e-01 -5.73307812e-01
1.19372942e-01 -4.55530226e-01 1.24024153e-01 -1.54929590e+00
-2.05659494e-01 -1.33906752e-01 -2.72488773e-01 1.83607602e+00
5.43549240e-01 -6.00240588e-01 1.00655437e+00 4.76680130e-01
-3.07340622e-01 -6.42711520e-01 -6.70530200e-01 -9.71701860e-01
-1.66228728e-03 -3.80531758e-01 7.83555746e-01 1.64351523e+00
2.44865447e-01 5.31252861e-01 -4.92586762e-01 3.24628800e-01
2.14241922e-01 -1.72373578e-01 9.48835433e-01 -1.34766006e+00
-2.63918161e-01 -2.92623132e-01 -4.78531212e-01 -5.48793197e-01
2.34258190e-01 -9.99437392e-01 -1.98547632e-01 -1.44069278e+00
2.47160017e-01 -6.94458723e-01 -5.90691626e-01 8.11633050e-01
-6.04263246e-01 1.10264324e-01 4.66344692e-02 5.56723997e-02
-5.72666168e-01 3.02763462e-01 8.10702741e-01 1.47783935e-01
-3.50665599e-01 -1.55155182e-01 -9.28913891e-01 6.35722518e-01
7.35016167e-01 -5.17315507e-01 -6.42182112e-01 -2.34765336e-01
5.79665005e-01 -3.10136467e-01 1.83080107e-01 -1.01483917e+00
5.52386999e-01 1.02571853e-01 1.97587416e-01 -4.68930453e-01
-1.26602620e-01 -1.01146913e+00 5.70293367e-01 1.56949013e-01
-1.13511451e-01 -1.60757080e-02 2.17411324e-01 5.02703428e-01
-3.60842377e-01 -2.92344421e-01 3.42552423e-01 -5.40959612e-02
-9.27068770e-01 -2.65916616e-01 1.62275016e-01 -5.38077317e-02
9.48857188e-01 -4.47465837e-01 -3.63343775e-01 4.45209742e-02
-6.28198743e-01 3.46976191e-01 5.06354392e-01 5.64446032e-01
3.64085659e-02 -1.21114337e+00 -4.15306002e-01 1.80562630e-01
2.97492146e-01 6.10656999e-02 1.61628455e-01 9.14864182e-01
9.97484997e-02 3.51080358e-01 -4.20403123e-01 -3.93162340e-01
-1.22416317e+00 6.90079927e-01 2.41214007e-01 -5.95134377e-01
-6.05442643e-01 7.04190910e-01 -4.45051864e-02 -5.57944834e-01
2.90111661e-01 -1.54857293e-01 -5.68425775e-01 1.66512340e-01
2.81868100e-01 2.09762171e-01 5.28196990e-01 -4.11235869e-01
-5.26925683e-01 1.28874108e-01 -3.87470692e-01 2.99198091e-01
1.32065248e+00 -7.39344135e-02 3.24582774e-03 3.37145418e-01
6.50100589e-01 4.20055389e-01 -7.64653623e-01 -9.39168781e-02
4.75910693e-01 -2.67356008e-01 -4.76200394e-02 -1.18389833e+00
-1.00433493e+00 4.05367374e-01 5.81025004e-01 2.92153746e-01
1.16065609e+00 -5.71435839e-02 5.02708852e-01 4.09631580e-01
4.61519241e-01 -1.16728842e+00 -3.36068273e-01 3.93157274e-01
4.80833322e-01 -1.50583136e+00 3.34257968e-02 -9.00027037e-01
-6.51364326e-01 8.69763732e-01 9.20692086e-01 1.88164055e-01
3.74818861e-01 4.65831131e-01 -4.85160053e-02 -1.97013542e-01
-6.16712928e-01 -4.56253469e-01 3.51761878e-01 6.40989602e-01
5.17326593e-01 -7.47837499e-02 -6.62707984e-01 9.81614172e-01
-3.70784611e-01 2.43358314e-01 2.88862050e-01 8.19680929e-01
-1.54706702e-01 -1.42958939e+00 -5.03798239e-02 3.85020107e-01
-4.38807487e-01 -1.44044295e-01 -5.35041332e-01 1.17259479e+00
4.72983807e-01 6.07618451e-01 -2.93737113e-01 -6.56585693e-01
5.14314473e-01 2.92656362e-01 4.44097281e-01 -3.63389343e-01
-8.40069413e-01 -3.41541022e-01 5.88084936e-01 -2.11047947e-01
-6.57540977e-01 -2.40806267e-01 -1.01807976e+00 -1.04511604e-01
-8.81957293e-01 3.24698895e-01 4.40577418e-01 9.85229790e-01
5.09634554e-01 5.63106000e-01 1.79413274e-01 -1.71162933e-01
-4.66584057e-01 -1.04591596e+00 -2.74624974e-01 5.46151817e-01
-4.33776304e-02 -9.25454140e-01 -2.41542086e-01 -2.93787211e-01] | [9.242093086242676, 8.459824562072754] |
8acbedb4-00d5-4d72-b33a-a0c25da75e94 | tokenization-and-the-noiseless-channel | 2306.16842 | null | https://arxiv.org/abs/2306.16842v1 | https://arxiv.org/pdf/2306.16842v1.pdf | Tokenization and the Noiseless Channel | Subword tokenization is a key part of many NLP pipelines. However, little is known about why some tokenizer and hyperparameter combinations lead to better downstream model performance than others. We propose that good tokenizers lead to \emph{efficient} channel usage, where the channel is the means by which some input is conveyed to the model and efficiency can be quantified in information-theoretic terms as the ratio of the Shannon entropy to the maximum possible entropy of the token distribution. Yet, an optimal encoding according to Shannon entropy assigns extremely long codes to low-frequency tokens and very short codes to high-frequency tokens. Defining efficiency in terms of R\'enyi entropy, on the other hand, penalizes distributions with either very high or very low-frequency tokens. In machine translation, we find that across multiple tokenizers, the R\'enyi entropy with $\alpha = 2.5$ has a very strong correlation with \textsc{Bleu}: $0.78$ in comparison to just $-0.32$ for compressed length. | ['Ryan Cotterell', 'Mrinmaya Sachan', 'Li Du', 'Juan Luis Gastaldi', 'Clara Meister', 'Vilém Zouhar'] | 2023-06-29 | null | null | null | null | ['machine-translation'] | ['natural-language-processing'] | [ 1.08361796e-01 4.24630582e-01 -5.76992810e-01 -2.27025136e-01
-1.02011287e+00 -7.17027545e-01 4.01905119e-01 4.08525646e-01
-8.68679821e-01 6.32722735e-01 5.74783385e-01 -5.58815241e-01
-1.13112688e-01 -6.66310430e-01 -5.65399289e-01 -6.73616946e-01
-3.35829370e-02 5.50806701e-01 -2.97689080e-01 -9.67580900e-02
2.48228803e-01 -3.23987938e-02 -9.32694972e-01 1.90451711e-01
7.43602455e-01 7.66515613e-01 4.35314208e-01 6.56117558e-01
-5.68449378e-01 3.41688633e-01 -7.10433006e-01 -6.40257835e-01
7.80758932e-02 -7.00898051e-01 -9.66737151e-01 -6.81549788e-01
-2.29791686e-01 -6.92831352e-02 -2.00824678e-01 1.26345265e+00
4.84771103e-01 -2.90818047e-02 7.59564281e-01 -6.24333799e-01
-5.20143390e-01 1.66955292e+00 -3.80776465e-01 4.34408605e-01
3.64031076e-01 2.05646783e-01 1.63558054e+00 -4.41003889e-01
7.45696604e-01 1.10525489e+00 4.35778052e-01 2.94656694e-01
-1.24714947e+00 -5.07751942e-01 -3.32298934e-01 -2.05873743e-01
-1.30697143e+00 -6.20691001e-01 1.97510034e-01 -2.27190599e-01
1.27844012e+00 2.07856387e-01 5.58704257e-01 5.95964670e-01
3.81565034e-01 6.39446676e-01 6.92047000e-01 -5.71230412e-01
1.37381107e-01 2.33610347e-01 3.09144463e-02 3.11157823e-01
2.11828411e-01 -1.68314829e-01 -6.73366308e-01 6.36819899e-02
2.69128412e-01 -5.51387787e-01 -1.83039442e-01 7.74215221e-01
-1.18465698e+00 8.87075484e-01 3.01807211e-03 7.46675968e-01
-3.71380448e-01 6.04707062e-01 3.19847703e-01 3.11121821e-01
3.20721358e-01 8.47211182e-01 -6.35800004e-01 -8.32904279e-01
-1.00658822e+00 1.87391505e-01 6.95135236e-01 1.06574929e+00
7.29170978e-01 -1.25090584e-01 -3.18055809e-01 1.02666080e+00
3.41203094e-01 5.49959779e-01 5.43649733e-01 -1.07232714e+00
6.69375300e-01 2.09541917e-01 -8.14044401e-02 -3.19617242e-01
-4.12202626e-01 -5.27776599e-01 -3.82321417e-01 -6.71487689e-01
7.96225846e-01 -2.99660683e-01 -5.11473894e-01 2.05279613e+00
-3.57888550e-01 -6.34450078e-01 2.72083342e-01 5.91726780e-01
4.80023384e-01 1.07554817e+00 3.02077800e-01 -5.20798445e-01
1.56713843e+00 -3.52579564e-01 -8.48589122e-01 -3.08017671e-01
1.04763317e+00 -1.12280262e+00 1.22708559e+00 2.71653477e-02
-1.37107229e+00 -5.03017567e-02 -7.70181715e-01 -2.56470561e-01
-1.13423154e-01 -3.27187806e-01 5.84885061e-01 9.38808382e-01
-9.33899045e-01 6.68219030e-01 -5.33304036e-01 -4.31561144e-03
2.66904294e-01 2.82909751e-01 -5.57490215e-02 1.04106434e-01
-1.42473710e+00 9.16471601e-01 6.29000962e-01 -3.75294089e-01
-3.30994308e-01 -6.78098798e-01 -6.35376692e-01 5.89974165e-01
-1.55873969e-01 -3.51070672e-01 1.29661334e+00 -7.35664964e-01
-1.50205219e+00 5.84192395e-01 -2.10318804e-01 -4.19656456e-01
1.24307908e-01 6.63506910e-02 -6.49827123e-02 2.10129246e-02
-2.05365922e-02 7.61423588e-01 2.61463255e-01 -6.37182951e-01
-6.58863902e-01 -7.94951841e-02 -1.15613721e-01 3.37079972e-01
-5.30072093e-01 2.86609888e-01 -3.89892638e-01 -4.24808443e-01
1.98843658e-01 -8.61013353e-01 1.43719420e-01 -4.88648295e-01
-4.74897683e-01 -3.51415336e-01 4.18554209e-02 -6.88429534e-01
1.57051539e+00 -2.12957048e+00 -1.22580104e-01 4.84572351e-01
9.43632424e-03 7.22418576e-02 7.87891001e-02 4.67480093e-01
1.82871521e-01 6.15085423e-01 -1.51177138e-01 9.02343355e-03
3.52543861e-01 6.54822290e-02 7.25298002e-02 3.84232223e-01
-7.95695931e-02 7.41927564e-01 -1.16163623e+00 -4.68734741e-01
-6.19667619e-02 4.43327427e-01 -6.93445623e-01 -2.10774094e-01
-2.38631830e-01 -7.72268400e-02 -3.45086455e-01 3.48568290e-01
3.87447625e-01 -1.42965153e-01 3.07383358e-01 3.04677844e-01
-3.78952563e-01 7.37954259e-01 -7.48964787e-01 1.59155548e+00
-4.06408727e-01 9.08859134e-01 -2.16654629e-01 -5.32824695e-01
5.32145381e-01 4.39570099e-01 4.45515037e-01 -8.04118454e-01
2.88229167e-01 4.08642828e-01 4.78763938e-01 -2.91928113e-01
9.88659680e-01 -3.47510278e-01 -2.29066312e-01 5.98934233e-01
5.89063428e-02 -2.45816737e-01 3.91608804e-01 3.25680941e-01
1.08307648e+00 -2.16386601e-01 2.50983149e-01 -3.89356017e-01
-4.42854613e-02 -1.40583724e-01 5.63998163e-01 7.83054411e-01
-2.37706333e-01 2.85594493e-01 1.02183986e+00 3.60531807e-01
-1.46730471e+00 -9.65217531e-01 -3.50624502e-01 1.06024301e+00
-1.29081190e-01 -6.97492838e-01 -1.18780768e+00 -1.77871838e-01
-1.45554677e-01 1.23399758e+00 -1.03780270e-01 -2.19887570e-01
-4.75312024e-01 -8.96886528e-01 1.13428235e+00 2.35754490e-01
1.25272393e-01 -9.11548913e-01 -4.94414508e-01 2.26089865e-01
-5.60547054e-01 -1.01283550e+00 -7.37871885e-01 4.95054960e-01
-6.72317922e-01 -2.83106387e-01 -6.96635067e-01 -5.12115240e-01
3.75864327e-01 -7.30377585e-02 1.00601578e+00 1.01035396e-02
1.15105122e-01 -2.90362597e-01 -7.53643513e-01 -6.06993377e-01
-7.06515789e-01 3.66210371e-01 -5.97164817e-02 -6.04508162e-01
4.90120173e-01 -1.75524533e-01 -6.97137237e-01 -1.14816569e-01
-8.08518589e-01 -2.23800734e-01 6.99816167e-01 7.68356681e-01
4.83499616e-01 -8.86398368e-03 4.53008562e-01 -9.32694554e-01
6.65979743e-01 -4.71558452e-01 -2.61612743e-01 5.40548787e-02
-6.95218325e-01 4.80707675e-01 6.88435614e-01 -2.76059687e-01
-9.10601199e-01 -1.82606906e-01 -6.30980790e-01 2.25655943e-01
2.50289112e-01 6.39931738e-01 -1.05098709e-01 4.38801497e-01
6.47345066e-01 2.28655294e-01 -2.47578904e-01 -3.07381690e-01
5.90071321e-01 7.96023309e-01 7.09191263e-02 -5.53081989e-01
2.50181705e-01 5.28282039e-02 -3.74384850e-01 -1.14007604e+00
-4.52695668e-01 -3.90984952e-01 -1.84149370e-01 -3.21346894e-02
8.73780549e-01 -5.19419968e-01 -8.16680908e-01 2.05639437e-01
-1.00177109e+00 -3.16584378e-01 -6.05164051e-01 8.22851360e-01
-7.62185574e-01 2.87074596e-01 -7.25315332e-01 -8.13304722e-01
-4.70165133e-01 -1.26599979e+00 7.08816886e-01 9.73196998e-02
-6.62382901e-01 -7.83235192e-01 -1.95574403e-01 2.91949272e-01
4.50882554e-01 -3.73259306e-01 1.33729911e+00 -7.60821879e-01
-1.29592136e-01 -2.47325301e-01 -1.54048637e-01 2.55231202e-01
-2.06903964e-01 -9.02599543e-02 -8.91905010e-01 1.60350017e-02
-1.73402309e-01 3.04641724e-02 8.98289502e-01 7.41321087e-01
8.88448298e-01 -4.64717537e-01 -1.68424532e-01 4.87884998e-01
1.33805609e+00 1.25752315e-01 7.57456601e-01 1.84000388e-01
2.90197760e-01 4.35315967e-01 3.66675735e-01 6.27505481e-01
2.15783283e-01 5.08301437e-01 -2.18046214e-02 3.27780992e-01
5.14732264e-02 -4.34439749e-01 6.12450421e-01 1.03018951e+00
1.82660699e-01 -5.96569896e-01 -9.35617805e-01 4.72153276e-01
-1.13291585e+00 -1.07851243e+00 -5.73248751e-02 2.09186387e+00
1.27544487e+00 6.20911360e-01 -2.08811294e-02 2.02458799e-01
6.73001826e-01 1.46321714e-01 -3.11978702e-02 -8.40665102e-01
-1.35640413e-01 2.70696402e-01 9.87284064e-01 7.73453951e-01
-4.42974150e-01 1.23609996e+00 6.55238199e+00 1.26242578e+00
-9.46867406e-01 1.53311327e-01 6.19854152e-01 -2.97362179e-01
-7.82971501e-01 2.41440266e-01 -8.33774745e-01 1.03464127e+00
1.56017816e+00 -5.02228081e-01 4.77267414e-01 5.62465250e-01
3.11244667e-01 -3.23944747e-01 -1.01073241e+00 7.26240754e-01
-3.66403878e-01 -1.20717561e+00 -1.38281375e-01 3.29146445e-01
4.62773919e-01 8.75952914e-02 2.01720998e-01 1.75613850e-01
4.05906498e-01 -9.70414937e-01 1.06930315e+00 2.04891711e-01
1.01609635e+00 -8.41596901e-01 6.38559699e-01 6.23484969e-01
-1.02913105e+00 3.19618061e-02 -6.14954650e-01 1.70971543e-01
3.27096283e-01 9.83230710e-01 -1.02274251e+00 1.39766991e-01
3.45601290e-01 1.84013117e-02 2.56200969e-01 7.17765331e-01
-8.74471962e-02 8.36843967e-01 -6.23192787e-01 -3.61236900e-01
4.11307067e-01 -2.00439349e-01 4.25994307e-01 1.82588828e+00
4.71408546e-01 2.69031376e-01 -4.62070554e-01 7.41047144e-01
-4.00438547e-01 4.17133927e-01 -3.86001647e-01 -5.15904725e-01
9.40807939e-01 8.07185709e-01 -8.75522852e-01 -3.12280893e-01
-1.74946979e-01 6.44197583e-01 1.53420135e-01 3.49306650e-02
-5.83145022e-01 -7.22630322e-01 6.32152677e-01 -2.55364217e-02
-1.15577973e-01 -1.43779725e-01 -5.96246600e-01 -7.65764236e-01
-2.15166658e-01 -5.94721615e-01 2.85229743e-01 -2.92985469e-01
-6.93406761e-01 2.95378894e-01 -1.75877467e-01 -7.97553658e-01
-3.02334845e-01 -3.16315114e-01 -2.42806703e-01 1.10408521e+00
-1.42035520e+00 -5.42936921e-01 4.89244163e-01 3.77790295e-02
4.77638364e-01 6.06251992e-02 8.04675817e-01 4.31482881e-01
-4.45624024e-01 1.07653308e+00 6.06422901e-01 1.60315245e-01
4.43726897e-01 -1.29320526e+00 5.07810473e-01 5.58491826e-01
5.94407842e-02 8.87347937e-01 1.01041460e+00 -5.33432364e-01
-1.49313903e+00 -7.48974502e-01 1.62792158e+00 -2.19378129e-01
4.40221906e-01 -2.90998109e-02 -5.20527840e-01 5.96506596e-01
2.05540568e-01 -5.91119885e-01 9.17441666e-01 2.76528567e-01
-2.33154222e-01 1.05448559e-01 -1.21053243e+00 5.66748917e-01
7.71171510e-01 -5.34171760e-01 -3.07350963e-01 3.15823376e-01
8.49471390e-01 -1.79169208e-01 -1.19434905e+00 -1.23142593e-01
6.74345315e-01 -6.80349827e-01 5.43125629e-01 -1.91467598e-01
4.46189821e-01 2.16960773e-01 -4.06291515e-01 -1.44625843e+00
-3.62943470e-01 -7.57815003e-01 3.40378642e-01 1.16399312e+00
1.11232340e+00 -3.98192525e-01 6.61948562e-01 4.58375275e-01
-3.73260260e-01 -6.46343350e-01 -1.05270612e+00 -7.57238269e-01
5.71078598e-01 -7.78594732e-01 5.79720020e-01 7.94449091e-01
8.22281957e-01 1.65325180e-01 -2.34745350e-02 -4.21272725e-01
3.62995654e-01 -4.81246889e-01 4.76006381e-02 -7.24042356e-01
-1.87158763e-01 -8.78658831e-01 -1.73801214e-01 -1.29746222e+00
9.20657814e-02 -1.35983598e+00 2.29242250e-01 -1.37926614e+00
3.31622481e-01 -7.75938809e-01 -2.28025839e-01 2.67348945e-01
-2.39718668e-02 -2.88522579e-02 2.78414696e-01 2.19393447e-01
-2.63083667e-01 1.15602456e-01 1.01551425e+00 -1.41135445e-02
-2.39492342e-01 -1.78799674e-01 -8.50412726e-01 5.50189734e-01
8.55987608e-01 -6.14731193e-01 -1.67036220e-01 -7.14327157e-01
7.97484934e-01 2.88940430e-01 -5.44766068e-01 -5.58608949e-01
6.05527200e-02 -2.01942816e-01 1.33836284e-01 -4.01741713e-01
3.23837906e-01 -5.31148374e-01 1.06117159e-01 5.51265955e-01
-7.96881855e-01 1.06060661e-01 -1.98932096e-01 2.10239708e-01
1.38098642e-01 -7.08240569e-01 9.04297829e-01 -4.21193361e-01
-1.71167493e-01 -6.64036795e-02 -6.60941005e-01 3.81966293e-01
7.11641014e-01 -2.15548962e-01 -2.18553454e-01 -3.41452867e-01
-3.44642401e-01 5.64925410e-02 4.45654243e-01 5.77295125e-02
3.67160201e-01 -8.91794264e-01 -7.81312585e-01 6.23600222e-02
-4.13086265e-02 -4.63960856e-01 -5.55168912e-02 6.47208929e-01
-7.33272672e-01 7.54757941e-01 6.27682209e-02 -3.30585629e-01
-8.66927564e-01 -5.00758477e-02 1.67921841e-01 -1.51213735e-01
-3.96084160e-01 1.17428970e+00 -3.33458006e-01 8.51355270e-02
4.50185351e-02 -4.30020154e-01 1.46423757e-01 3.57350945e-01
4.46096897e-01 5.89934468e-01 6.23601452e-02 -7.20610023e-01
-2.66335815e-01 2.70864427e-01 -2.12448820e-01 -6.29073560e-01
9.35436487e-01 -1.08694613e-01 -1.64135307e-01 2.55821884e-01
1.42750311e+00 3.05683110e-02 -7.94518888e-01 -2.04039007e-01
1.44375816e-01 -4.66635406e-01 1.82454109e-01 -5.77356279e-01
-7.70687044e-01 7.69559443e-01 2.93160170e-01 2.79144257e-01
6.08251989e-01 1.95734799e-01 1.08293533e+00 3.08657438e-01
2.04419762e-01 -1.68554211e+00 -2.06680894e-01 8.91613960e-01
1.35810956e-01 -8.43965709e-01 -6.28347471e-02 -8.55892077e-02
-7.75309265e-01 8.05797696e-01 -4.91450876e-02 4.15794760e-01
6.07476473e-01 3.34137231e-01 -1.11525513e-01 7.17964023e-02
-6.73442185e-01 -1.92910165e-01 -2.96033055e-01 2.37024158e-01
9.57775116e-01 5.08292198e-01 -9.13898349e-01 4.56739128e-01
-9.40651000e-01 -2.85652518e-01 4.67329323e-01 5.85957050e-01
-9.21178401e-01 -1.15436387e+00 -1.25240907e-01 8.47266555e-01
-1.04431641e+00 -5.81573248e-01 -3.43638659e-02 4.00101691e-01
1.12924948e-01 1.14675081e+00 3.52216184e-01 -3.04085374e-01
-4.02417108e-02 4.41319883e-01 3.31094950e-01 -5.70105731e-01
-8.59243393e-01 2.68676162e-01 2.63308078e-01 -1.91814095e-01
1.40527889e-01 -8.57641935e-01 -1.74082410e+00 -7.85221875e-01
-5.82481563e-01 4.47916120e-01 9.94797468e-01 9.18049335e-01
4.18340228e-03 2.96056867e-01 3.34758252e-01 -3.80246550e-01
-6.19443178e-01 -1.06022966e+00 -8.71455789e-01 2.73395568e-01
1.76100537e-01 -7.80678215e-03 -3.83266509e-01 -3.95941548e-02] | [11.275471687316895, 9.323846817016602] |
a1214523-222e-439f-b742-99a01980d354 | segmentation-of-blood-vessels-optic-disc | 2207.04345 | null | https://arxiv.org/abs/2207.04345v1 | https://arxiv.org/pdf/2207.04345v1.pdf | Segmentation of Blood Vessels, Optic Disc Localization, Detection of Exudates and Diabetic Retinopathy Diagnosis from Digital Fundus Images | Diabetic Retinopathy (DR) is a complication of long-standing, unchecked diabetes and one of the leading causes of blindness in the world. This paper focuses on improved and robust methods to extract some of the features of DR, viz. Blood Vessels and Exudates. Blood vessels are segmented using multiple morphological and thresholding operations. For the segmentation of exudates, k-means clustering and contour detection on the original images are used. Extensive noise reduction is performed to remove false positives from the vessel segmentation algorithm's results. The localization of Optic Disc using k-means clustering and template matching is also performed. Lastly, this paper presents a Deep Convolutional Neural Network (DCNN) model with 14 Convolutional Layers and 2 Fully Connected Layers, for the automatic, binary diagnosis of DR. The vessel segmentation, optic disc localization and DCNN achieve accuracies of 95.93%, 98.77% and 75.73% respectively. The source code and pre-trained model are available https://github.com/Sohambasu07/DR_2021 | ['Anindya Sen', 'Ankit Bhattacharya', 'Sayantan Mukherjee', 'Soham Basu'] | 2022-07-09 | null | null | null | null | ['template-matching', 'contour-detection', 'diabetic-retinopathy-detection'] | ['computer-vision', 'computer-vision', 'medical'] | [-1.34713784e-01 -1.86121296e-02 8.29636902e-02 -4.07057494e-01
-1.56503811e-01 -3.67120355e-01 1.02682747e-02 -5.95512167e-02
-3.29940528e-01 7.03084886e-01 1.51059628e-01 -4.71656442e-01
-4.70463037e-02 -8.10872674e-01 8.97511467e-03 -8.32940280e-01
-1.07761115e-01 1.24196358e-01 1.78309858e-01 3.29110831e-01
4.99273807e-01 1.04129791e+00 -1.53559256e+00 3.32874864e-01
1.19754863e+00 1.11853313e+00 -2.69150048e-01 1.12556517e+00
-3.30109060e-01 7.91702688e-01 -4.64724153e-01 -2.55485535e-01
4.86833692e-01 -4.08720136e-01 -7.16365457e-01 3.33429277e-01
7.96221435e-01 -6.77814603e-01 -3.07739675e-01 1.07587147e+00
9.64904308e-01 -1.94643587e-01 7.89906561e-01 -4.79585797e-01
-7.58256435e-01 -5.67219332e-02 -6.91167831e-01 4.79466707e-01
-2.74650306e-01 2.64521748e-01 1.28178388e-01 -6.04357779e-01
4.58238393e-01 1.03469253e+00 5.82815468e-01 4.15134430e-01
-9.17007029e-01 -4.68798250e-01 -6.76326632e-01 2.78169692e-01
-1.18273413e+00 -4.52102154e-01 1.44522294e-01 -9.86568868e-01
5.34207642e-01 2.10533410e-01 8.78723800e-01 2.20963642e-01
1.92822069e-01 5.45874238e-01 1.53391623e+00 -4.01724786e-01
2.05212623e-01 3.50030959e-02 4.50977266e-01 8.17631781e-01
5.96283674e-01 1.10804453e-01 3.45707148e-01 9.28349644e-02
1.18989408e+00 -1.12782702e-01 -3.00938606e-01 7.49810552e-03
-6.29397690e-01 7.59413302e-01 5.28989017e-01 -2.38681678e-02
-5.23066103e-01 -1.94172189e-01 3.91349435e-01 5.56380823e-02
1.40844524e-01 2.92632341e-01 -2.79880643e-01 3.73940647e-01
-7.27614105e-01 -4.56762835e-02 4.44734305e-01 7.22429395e-01
4.60650682e-01 -5.98254465e-02 -4.19594377e-01 9.40586507e-01
2.92468607e-01 3.20314825e-01 3.60367537e-01 -1.25759435e+00
-3.44832748e-01 1.00800741e+00 6.20369762e-02 -5.82218230e-01
-4.69040334e-01 -4.66129571e-01 -9.70652759e-01 1.01220644e+00
7.91787624e-01 -6.11424327e-01 -1.56302083e+00 4.99050856e-01
3.13259393e-01 -1.00578954e-02 1.50323123e-01 1.06885397e+00
1.10385978e+00 2.17129454e-01 -3.55814560e-03 -6.93821982e-02
1.39827168e+00 -8.98690403e-01 -5.65693796e-01 1.30795926e-01
5.65016568e-01 -9.97817516e-01 5.05251408e-01 4.62782115e-01
-1.09037089e+00 -3.30751687e-01 -5.74632585e-01 -4.84956384e-01
-3.44126433e-01 8.34021449e-01 7.54887283e-01 4.60431218e-01
-1.12956393e+00 1.87802702e-01 -5.43828070e-01 -6.15741372e-01
9.74382460e-01 3.56534392e-01 -2.82839447e-01 -3.39339107e-01
-3.73493612e-01 9.70735312e-01 -3.41945626e-02 3.09776902e-01
-1.99277952e-01 -4.22728568e-01 -3.80457759e-01 -4.24213260e-01
-2.92878240e-01 -8.54938090e-01 9.82801497e-01 -7.35779345e-01
-1.23569465e+00 1.25147963e+00 -3.41090739e-01 -4.24723804e-01
6.29256070e-01 -2.98055857e-01 -4.19163644e-01 6.98614359e-01
-5.06796576e-02 4.19172674e-01 3.76169354e-01 -9.36553538e-01
-1.04914808e+00 -6.54075205e-01 -3.00498575e-01 -2.04372346e-01
3.67602229e-01 4.54171002e-01 -2.85452724e-01 -1.97219819e-01
2.42846131e-01 -5.07656157e-01 -3.60373050e-01 5.61835170e-01
-6.58627927e-01 -1.59904063e-01 4.40480292e-01 -1.19885778e+00
1.02239132e+00 -1.86874497e+00 -3.54650885e-01 2.11970106e-01
5.12579024e-01 6.78956628e-01 1.66261181e-01 -2.14678019e-01
-7.99560100e-02 1.73046663e-01 -9.28828642e-02 1.44647032e-01
-5.11159122e-01 -5.30339442e-02 3.65277171e-01 5.87986827e-01
1.82814151e-01 6.22182310e-01 -3.94724995e-01 -4.13536996e-01
4.33422595e-01 7.75346816e-01 1.49668325e-02 8.55858810e-03
3.49861294e-01 5.49493253e-01 -2.75848627e-01 1.03433335e+00
8.43140483e-01 -1.34260818e-01 -3.54128510e-01 -4.72562820e-01
-5.66576838e-01 -4.09345120e-01 -1.19159770e+00 7.96584010e-01
2.04344809e-01 8.41970086e-01 1.25431418e-01 -7.19671249e-01
1.07996857e+00 2.22041905e-01 3.37804377e-01 -6.47729516e-01
5.92331827e-01 4.42577958e-01 2.20568687e-01 -1.12760842e+00
-1.64571494e-01 2.45120972e-01 8.69924903e-01 -4.67251204e-02
-3.16332281e-01 5.28928041e-01 4.78558809e-01 -2.19110966e-01
7.92271912e-01 -1.29051328e-01 4.69981521e-01 -3.21082100e-02
5.96498549e-01 2.57033288e-01 5.90223551e-01 3.25916260e-01
-5.00246882e-01 7.83687294e-01 6.00206435e-01 -7.01871097e-01
-1.04165709e+00 -7.95943975e-01 -7.23113239e-01 9.55006108e-03
-4.46653701e-02 5.78206599e-01 -6.37682438e-01 -1.30195439e-01
2.93005884e-01 3.69670153e-01 -5.08944392e-01 3.33906800e-01
-4.60508853e-01 -9.06281769e-01 3.04220915e-01 3.79561543e-01
8.62987220e-01 -7.47560143e-01 -5.19468546e-01 9.37986374e-02
3.38640422e-01 -7.52734601e-01 1.36822715e-01 -4.38755155e-01
-1.11011064e+00 -1.58942640e+00 -1.04996729e+00 -1.05948269e+00
1.01664245e+00 1.74178347e-01 6.08554006e-01 1.63205519e-01
-1.15320909e+00 -4.69153710e-02 -1.25693023e-01 -3.75453621e-01
-1.42188609e-01 -6.59552097e-01 -3.44298154e-01 1.94885537e-01
9.66509283e-01 -4.35965568e-01 -1.10299194e+00 5.19382879e-02
-2.55166888e-01 -2.15462551e-01 1.12154567e+00 3.00845832e-01
8.05947483e-01 8.79340339e-03 2.13199973e-01 -7.07777500e-01
4.46721196e-01 -1.14335485e-01 -7.27286935e-01 -3.19327526e-02
-5.85289240e-01 -6.12355351e-01 -6.72158897e-02 -1.51073813e-01
-8.01918864e-01 5.61583228e-02 1.66814819e-01 -1.25351727e-01
-9.10202861e-01 1.13242574e-01 1.96895912e-01 -3.99201721e-01
9.01150465e-01 -8.04963186e-02 4.03802156e-01 -8.21703017e-01
2.30071709e-01 1.24208915e+00 6.16959155e-01 2.04828396e-01
2.76875734e-01 7.26462781e-01 -2.90685017e-02 -8.73682380e-01
-7.51704872e-01 -8.69949102e-01 -8.23275805e-01 -3.71236861e-01
1.37503648e+00 -8.53087842e-01 -4.44283664e-01 1.01644301e+00
-9.58997786e-01 -2.75673002e-01 1.16319150e-01 8.98636878e-01
-1.94472924e-01 2.76263386e-01 -6.98697209e-01 -8.33875299e-01
-5.93186557e-01 -9.36439097e-01 2.26353958e-01 9.77923453e-01
1.75318047e-01 -8.22336674e-01 -2.34603956e-01 7.65006185e-01
4.91498172e-01 7.90654480e-01 9.68666613e-01 -3.38076919e-01
-6.22545481e-01 -3.98953617e-01 -9.51284945e-01 7.43716776e-01
2.41616920e-01 6.48558855e-01 -8.46180320e-01 5.00235781e-02
-4.83495295e-01 2.28456885e-01 1.03176093e+00 1.17455578e+00
6.56001806e-01 -1.40986547e-01 -3.13195676e-01 7.44090617e-01
1.69642866e+00 5.97490311e-01 1.06528938e+00 5.37898302e-01
4.48860139e-01 6.89021349e-01 1.95098892e-01 3.24991673e-01
2.60375082e-01 -1.27101809e-01 4.19529468e-01 -6.47134125e-01
-6.54646814e-01 7.62175500e-01 -2.43470773e-01 1.19951293e-01
-6.86937273e-01 1.50396734e-01 -1.07807612e+00 8.23311269e-01
-1.26903439e+00 -7.79119730e-01 -1.03772247e+00 2.09223533e+00
7.21134722e-01 -8.54956135e-02 2.25839511e-01 -1.34636492e-01
1.08716869e+00 -8.06203127e-01 -6.19473755e-01 -4.89698917e-01
-2.74123877e-01 4.42777306e-01 8.98389339e-01 4.08104718e-01
-1.32695985e+00 6.72735572e-01 5.53273296e+00 -1.85430840e-01
-1.20283985e+00 -3.43600601e-01 6.97094083e-01 3.26278037e-03
6.99418664e-01 -2.29339495e-01 -8.23282897e-01 3.02684158e-01
4.79167998e-01 2.16794327e-01 2.41748318e-01 5.28672993e-01
7.17217207e-01 -5.19502819e-01 -3.97379071e-01 8.13983798e-01
-2.02087045e-01 -1.35648966e+00 -1.47888228e-01 2.48712003e-01
7.53260493e-01 2.22773910e-01 -1.39080361e-01 -2.03147322e-01
2.80656785e-01 -1.19493496e+00 -7.95946568e-02 1.29473710e+00
8.02313268e-01 -6.80656075e-01 1.22021651e+00 -3.99729908e-01
-5.88335395e-01 -3.54281217e-01 -5.56329072e-01 1.12840980e-01
6.03402443e-02 9.28632677e-01 -6.39906466e-01 2.65353150e-03
9.42508399e-01 6.21851623e-01 -7.82342553e-01 2.31234789e+00
-3.36807072e-01 6.39126539e-01 -2.52016895e-02 3.18554550e-01
-2.83627510e-02 -6.08993292e-01 6.62323475e-01 1.10890114e+00
3.52028817e-01 2.92997956e-01 -1.82867616e-01 8.33770454e-01
3.31152886e-01 3.48356575e-01 -4.86387797e-02 1.02284089e-01
2.93302953e-01 1.31471801e+00 -5.40140808e-01 -2.24506304e-01
-5.34751475e-01 4.48892534e-01 -2.19608843e-01 6.51719987e-01
-1.82965294e-01 -7.90379822e-01 6.58025026e-01 2.73535401e-01
2.06698775e-01 -2.03367129e-01 -7.15255678e-01 -6.75989091e-01
-7.68987536e-02 -4.68090236e-01 4.01545793e-01 -8.30439091e-01
-1.24469280e+00 3.00575048e-01 -6.72473788e-01 -1.03828037e+00
5.55415452e-01 -9.03370917e-01 -8.99730682e-01 1.35495448e+00
-2.00279975e+00 -1.12418652e+00 -7.29608417e-01 4.27360326e-01
1.48031548e-01 -5.04732072e-01 6.54771030e-01 3.36727917e-01
-1.03009546e+00 1.20804928e-01 2.76407391e-01 5.59580684e-01
8.26381624e-01 -1.51992965e+00 -2.21744627e-01 8.72240424e-01
-1.05455053e+00 5.11695623e-01 3.50127280e-01 -7.42570043e-01
-6.56025231e-01 -1.42297828e+00 8.41369212e-01 1.40481696e-01
4.10740107e-01 4.18390244e-01 -8.20831001e-01 4.42290068e-01
9.19255167e-02 2.27405131e-01 8.90532553e-01 -4.03538018e-01
1.80840977e-02 -6.73866570e-02 -1.42951369e+00 4.51799780e-01
4.76289034e-01 1.64875597e-01 -3.07397664e-01 4.83514011e-01
-2.45132353e-02 -4.29984987e-01 -1.24770355e+00 2.56414920e-01
6.71674967e-01 -1.16768432e+00 7.90468991e-01 -5.42362154e-01
4.23948050e-01 -6.07636631e-01 1.35353163e-01 -6.06038153e-01
-2.95629233e-01 -3.67866486e-01 3.70682292e-02 1.07498801e+00
3.21404278e-01 -8.54840994e-01 7.80611038e-01 5.35266817e-01
-3.00496876e-01 -6.48411512e-01 -4.70050037e-01 -3.39701980e-01
9.88580063e-02 3.92257154e-01 2.97960732e-02 8.37042868e-01
-7.77878642e-01 -1.35743812e-01 1.61534339e-01 6.29225969e-01
9.21784580e-01 6.17753826e-02 5.96084654e-01 -1.67645168e+00
3.60516220e-01 -6.03063643e-01 -8.51946354e-01 -4.56863910e-01
-6.14365101e-01 -6.04188085e-01 -5.22903323e-01 -2.40455842e+00
-1.92935243e-01 -3.39533448e-01 -4.54608351e-02 6.01769209e-01
2.56099761e-01 4.43153441e-01 -4.05207992e-01 3.01220536e-01
1.26577437e-01 -2.06663877e-01 1.77710021e+00 1.16773374e-01
-5.59275508e-01 3.92025977e-01 -8.38582158e-01 6.45682275e-01
1.55595541e+00 -1.15345037e-02 -5.73249571e-02 -2.82028615e-01
-2.52288640e-01 -2.36845121e-01 6.67473257e-01 -9.78228688e-01
2.83675015e-01 -1.15010299e-01 7.83763051e-01 -6.47835135e-01
-4.84885685e-02 -3.70578140e-01 -2.17676297e-01 4.62291539e-01
-1.60346389e-01 -6.88343346e-01 1.03120610e-01 2.25758940e-01
-8.08859542e-02 -1.88570872e-01 1.42276251e+00 -3.62567753e-01
-8.31511796e-01 2.34675094e-01 -6.48095727e-01 -4.15555030e-01
1.19525552e+00 -7.61439204e-01 -8.42289209e-01 2.03725025e-01
-1.18547118e+00 3.17347705e-01 3.33911180e-01 -1.02467127e-01
6.89988017e-01 -8.27589333e-01 -1.09297228e+00 1.24373876e-01
1.46092772e-02 2.63380736e-01 2.23330036e-01 1.43836045e+00
-1.18929875e+00 3.08427870e-01 -6.24302626e-01 -3.44626933e-01
-1.72635901e+00 1.06877148e-01 8.81093502e-01 6.19468153e-01
-9.38058197e-01 9.09462690e-01 -5.28357387e-01 9.38938260e-02
3.78388762e-01 -4.41546649e-01 -9.67225790e-01 7.01128468e-02
9.29909885e-01 8.92026126e-01 -1.61369637e-01 -5.43332279e-01
5.09580597e-02 9.52657402e-01 -2.79085129e-01 5.85683405e-01
1.23700023e+00 -4.00815606e-01 -7.28915036e-01 -2.94795055e-02
8.08370113e-01 -1.43259749e-01 -8.72832954e-01 -2.67218322e-01
-1.18734427e-01 -4.76514101e-01 6.27742708e-01 -1.32775545e+00
-1.25713146e+00 9.50543165e-01 1.27370059e+00 7.61554316e-02
1.19568813e+00 -1.70617878e-01 7.05347419e-01 3.59526239e-02
-2.40179047e-01 -8.93927634e-01 -6.42731786e-01 6.67796135e-02
6.89346850e-01 -1.10623157e+00 6.12556301e-02 -6.24186337e-01
-3.57693136e-01 1.62305784e+00 5.72364151e-01 -3.23506474e-01
6.02313578e-01 -1.10263631e-01 8.31107676e-01 -2.21490398e-01
-1.44720152e-01 -7.52056122e-01 3.16117465e-01 9.02902901e-01
4.26939338e-01 5.62356636e-02 -6.37170136e-01 1.27354980e-01
-6.29714206e-02 3.08481634e-01 8.90703261e-01 7.36091852e-01
-1.05222178e+00 -5.85246980e-01 -1.82606757e-01 9.41757858e-01
-6.30222619e-01 5.38609661e-02 -6.12251639e-01 7.82887340e-01
4.29229230e-01 8.99418175e-01 2.75902718e-01 2.19777942e-01
3.05939674e-01 -3.42808925e-02 1.91791609e-01 -4.49031830e-01
-2.46770501e-01 3.99945647e-01 4.16112900e-01 -4.20677543e-01
-5.93879282e-01 -5.05564988e-01 -1.50829780e+00 -1.16843291e-01
-3.52181755e-02 -2.82248646e-01 8.02978098e-01 5.68224847e-01
4.28150415e-01 4.01915401e-01 4.18070585e-01 -1.18804164e-01
-6.34426475e-02 -9.64633405e-01 -8.86544883e-01 -1.00208320e-01
6.04528189e-01 -4.56107885e-01 -4.52991247e-01 4.96778905e-01] | [15.833322525024414, -3.995621919631958] |
db5c377b-6bf4-4329-8d98-6d2309d0bbe9 | pancreas-segmentation-via-spatial-context | 1903.00832 | null | https://arxiv.org/abs/1903.00832v3 | https://arxiv.org/pdf/1903.00832v3.pdf | A Model-Driven Stack-Based Fully Convolutional Network for Pancreas Segmentation | The irregular geometry and high inter-slice variability in computerized tomography (CT) scans of the human pancreas make an accurate segmentation of this crucial organ a challenging task for existing data-driven deep learning methods. To address this problem, we present a novel model-driven stack-based fully convolutional network with a sliding window fusion algorithm for pancreas segmentation, termed MDS-Net. The MDS-Net's cost function includes a data approximation term and a prior knowledge regularization term combined with a stack scheme for capturing and fusing the two-dimensional (2D) and local three-dimensional (3D) context information. Specifically, 3D CT scans are divided into multiple stacks to capture the local spatial context feature. To highlight the importance of single slices, the inter-slice relationships in the stack data are also incorporated in the MDS-Net framework. For implementing this proposed model-driven method, we create a stack-based U-Net architecture and successfully derive its back-propagation procedure for end-to-end training. Furthermore, a sliding window fusion algorithm is utilized to improve the consistency of adjacent CT slices and intra-stack. Finally, extensive quantitative assessments on the NIH Pancreas-CT dataset demonstrated higher pancreatic segmentation accuracy and reliability of MDS-Net compared to other state-of-the-art methods. | ['Xiaohua Qian', 'Xiaozhu Lin', 'Jun Li', 'Hao Li'] | 2019-03-03 | null | null | null | null | ['pancreas-segmentation'] | ['medical'] | [-2.89028343e-02 -6.29655260e-04 -3.63983959e-02 -7.73961008e-01
-8.91468763e-01 -3.05011928e-01 1.85313478e-01 4.05843556e-01
-4.40581799e-01 1.98041558e-01 2.07467556e-01 -3.06661338e-01
-1.95847571e-01 -7.54584849e-01 -8.95082474e-01 -8.05548668e-01
-5.34362674e-01 2.98211366e-01 5.63208997e-01 1.09887354e-01
-2.13796943e-01 4.22497630e-01 -6.48419678e-01 2.75952280e-01
1.05063856e+00 1.16981602e+00 4.50405389e-01 4.96624023e-01
-3.99722725e-01 4.86823738e-01 4.44119014e-02 -8.50526690e-02
4.35360461e-01 -3.41630280e-01 -5.40458679e-01 2.77511001e-01
3.26548964e-01 -5.28514564e-01 -2.83008963e-01 1.10550928e+00
4.94160950e-01 -9.09290388e-02 4.12038773e-01 -6.40637398e-01
-3.08824897e-01 7.58126915e-01 -2.96983689e-01 4.63136733e-01
-1.18704364e-01 3.30685973e-01 3.30626160e-01 -6.62261128e-01
5.22304833e-01 7.24663019e-01 9.41055894e-01 5.79971224e-02
-1.07016265e+00 -5.18270433e-01 1.80776313e-01 -2.60732055e-01
-9.41763222e-01 3.76165837e-01 8.76364172e-01 -5.49589336e-01
6.48979962e-01 2.00760383e-02 1.23836112e+00 7.15243518e-01
6.32372081e-01 7.62285352e-01 8.75662088e-01 -1.89996406e-01
1.23297729e-01 -4.77211654e-01 2.00726435e-01 9.52280939e-01
2.06172660e-01 2.31589600e-01 4.44551930e-02 -3.33403759e-02
1.39520609e+00 3.48938227e-01 -4.32509780e-01 -8.47124994e-01
-1.43363190e+00 7.38822639e-01 1.09429073e+00 3.18107784e-01
-7.95805573e-01 -1.46570712e-01 3.49513650e-01 -2.19205528e-01
3.31718296e-01 7.87728578e-02 -3.78864020e-01 3.71575326e-01
-1.07539868e+00 1.08112790e-01 6.08898222e-01 9.03639972e-01
2.04457551e-01 -1.29909515e-01 -1.17930502e-01 5.22462368e-01
8.33604813e-01 -4.00598049e-02 7.66694069e-01 -6.67472422e-01
3.47609937e-01 1.05757141e+00 -4.39377755e-01 -3.71622056e-01
-1.00822484e+00 -7.78270781e-01 -1.07779288e+00 2.70462275e-01
4.61527050e-01 -3.50514464e-02 -1.45342767e+00 1.38893378e+00
6.82061672e-01 4.31582570e-01 -2.89843827e-01 1.42384005e+00
1.01666939e+00 1.63681269e-01 3.70356321e-01 -1.62438452e-01
1.46868384e+00 -1.17293465e+00 -4.17154908e-01 2.36517057e-01
7.66509473e-01 -5.18470228e-01 7.18486190e-01 9.64978337e-02
-1.25187755e+00 -2.20089287e-01 -1.28495693e+00 1.06409872e-02
-2.47822907e-02 -1.18075646e-01 6.15962923e-01 3.54044497e-01
-8.28396857e-01 5.91981232e-01 -1.39977920e+00 -9.43216905e-02
7.43948698e-01 4.21808153e-01 -3.36573362e-01 -1.12146720e-01
-8.54036212e-01 7.46271074e-01 4.11748052e-01 1.89489111e-01
-7.94582009e-01 -1.03482437e+00 -9.83923316e-01 6.84718937e-02
1.06965072e-01 -9.95437026e-01 1.36282074e+00 -6.15761817e-01
-1.50412452e+00 6.93513811e-01 1.91821694e-01 -4.71983582e-01
7.90669262e-01 2.44013891e-01 -1.38969332e-01 4.38491970e-01
-6.61836267e-02 4.27648276e-01 3.70618790e-01 -1.20706201e+00
-3.96632880e-01 -6.55478537e-01 -1.95542917e-01 2.14272171e-01
1.44257158e-01 -1.57173231e-01 -4.82057571e-01 -9.22780335e-01
8.74777436e-01 -8.63662124e-01 -6.33153498e-01 5.53118706e-01
-2.11285621e-01 2.17783198e-01 5.46351552e-01 -7.55506337e-01
9.63968992e-01 -2.08201528e+00 3.31821553e-02 2.49838442e-01
4.07106012e-01 3.06944758e-01 1.31049842e-01 -4.73596156e-01
-2.46350914e-01 -1.36521712e-01 -6.77545726e-01 -1.95769712e-01
-2.55651742e-01 3.22947145e-01 4.22237962e-01 6.70101464e-01
6.98829591e-02 1.11960173e+00 -8.53094459e-01 -6.69214010e-01
6.18956685e-01 6.43571913e-01 -5.80384195e-01 2.43987396e-01
-1.18692897e-01 8.49168301e-01 -5.41527271e-01 7.80276060e-01
9.04755771e-01 -5.23574531e-01 1.27791539e-01 -5.37808836e-01
-3.86460125e-01 8.52617174e-02 -8.15761447e-01 2.33442044e+00
-2.03028500e-01 -1.16552584e-01 4.85565275e-01 -6.25151515e-01
5.97961247e-01 4.66307580e-01 9.22909021e-01 -8.94394994e-01
3.88618588e-01 3.86510342e-01 2.12442756e-01 -6.61168456e-01
-5.59232123e-02 -4.36021805e-01 2.92423725e-01 3.08577955e-01
2.83111811e-01 -4.07659486e-02 -9.11991298e-03 -1.74279109e-01
7.04529524e-01 2.70578086e-01 1.40567526e-01 -5.43344259e-01
4.38833386e-01 1.79321587e-01 6.81355476e-01 3.10402960e-01
-5.66745102e-01 1.01936913e+00 2.59098202e-01 -8.75757217e-01
-9.05256927e-01 -1.15766120e+00 -5.40921628e-01 5.50187767e-01
3.80148649e-01 5.60610704e-02 -8.66438985e-01 -8.49245131e-01
8.34369957e-02 2.57093072e-01 -6.19844675e-01 7.03804344e-02
-9.30350006e-01 -1.03853667e+00 2.24729612e-01 8.01223278e-01
4.92247015e-01 -6.82558954e-01 -7.92636573e-01 6.52717590e-01
3.06162760e-02 -9.01737809e-01 -5.95947206e-01 6.24499857e-01
-1.46130681e+00 -1.27320957e+00 -1.11382008e+00 -1.14047360e+00
8.85073721e-01 1.11166425e-01 1.08047688e+00 -2.84850970e-03
-2.29377255e-01 -1.43388584e-01 -1.86012611e-01 -1.42281801e-01
-2.41801083e-01 -2.09754296e-02 -4.59074885e-01 -4.51343268e-01
6.56655878e-02 -6.98839605e-01 -1.05034971e+00 2.85787374e-01
-1.17872262e+00 2.49936864e-01 7.87366390e-01 7.94067144e-01
9.01788831e-01 -2.88371354e-01 2.05446318e-01 -5.66302598e-01
2.41303816e-01 -5.19225776e-01 -7.31284797e-01 1.27147913e-01
-3.33465695e-01 -8.73165056e-02 3.00148726e-01 -3.92383963e-01
-8.29205573e-01 3.99770498e-01 -3.30181599e-01 -5.01083255e-01
-9.03950632e-02 6.97453320e-01 1.11813419e-01 -3.91873449e-01
3.45601708e-01 2.73196906e-01 3.76753628e-01 -5.01446664e-01
3.96920592e-02 2.04554558e-01 4.99927074e-01 -3.68277878e-01
2.29354739e-01 4.51145500e-01 3.35385948e-02 -3.54323149e-01
-5.76110840e-01 -4.30262744e-01 -9.27687228e-01 -1.89563945e-01
1.12935495e+00 -9.20996308e-01 -4.52612787e-01 6.40395522e-01
-9.68625546e-01 -4.58060265e-01 -2.56472170e-01 9.61139679e-01
-4.66407657e-01 2.90887803e-01 -9.55705702e-01 -4.32488956e-02
-7.04978943e-01 -1.75563824e+00 8.78886580e-01 2.76051521e-01
1.65893003e-01 -1.09000492e+00 1.64241791e-02 1.26216561e-01
5.60940444e-01 5.71608126e-01 1.16399002e+00 -7.05846429e-01
-8.26219738e-01 -1.02275424e-01 -4.60383743e-01 2.55763173e-01
3.71202156e-02 -3.41013372e-01 -3.61994565e-01 -1.32576540e-01
2.16691643e-01 2.13932231e-01 7.11485386e-01 1.10822439e+00
1.08303738e+00 6.82265684e-02 -2.51846880e-01 1.29807663e+00
1.51388264e+00 2.20572934e-01 1.77471280e-01 5.17157435e-01
7.55296290e-01 3.28593291e-02 1.03436567e-01 3.42823923e-01
6.39711738e-01 3.31294626e-01 7.79781103e-01 -4.62828130e-01
-1.89976439e-01 5.26183620e-02 -2.69554943e-01 8.86245251e-01
1.60584509e-01 1.77249908e-01 -9.55581605e-01 5.67492902e-01
-1.67426038e+00 -2.52162784e-01 -4.12053347e-01 2.04509377e+00
6.47080362e-01 1.61895916e-01 1.28303301e-02 -3.05012524e-01
6.32808983e-01 -3.77365723e-02 -5.98159969e-01 1.63918242e-01
4.08840142e-02 -2.78001633e-02 5.78460097e-01 3.02404165e-01
-1.44604588e+00 1.81152523e-01 5.87582016e+00 2.68693954e-01
-1.52843606e+00 1.19671658e-01 5.21941423e-01 -1.02296144e-01
-1.96632594e-01 -3.51565599e-01 -2.45654911e-01 5.81227660e-01
4.54290390e-01 1.00229517e-01 -1.05447792e-01 6.58121586e-01
1.42997026e-01 -1.43653065e-01 -1.07245600e+00 6.26947224e-01
-1.75019249e-01 -1.38058698e+00 -7.71072209e-02 1.04389675e-01
6.34595692e-01 6.93627596e-01 -1.68361112e-01 1.67107657e-01
8.64466727e-02 -8.91872048e-01 7.21749187e-01 3.77944678e-01
5.35866797e-01 -5.81123650e-01 8.66273642e-01 2.93715268e-01
-1.23432863e+00 4.20981832e-02 6.78916201e-02 4.47720230e-01
4.72627133e-01 6.78662419e-01 -7.29658842e-01 7.57225811e-01
6.29594088e-01 5.21973670e-01 -2.38889500e-01 1.59084034e+00
1.24697939e-01 2.05689058e-01 -6.16522431e-01 5.08276165e-01
6.78830147e-01 -3.07508737e-01 5.95470428e-01 1.28521049e+00
5.13831437e-01 2.30968073e-01 3.24336708e-01 9.89460707e-01
-2.64345319e-03 7.79434387e-03 -1.43718487e-02 4.70931441e-01
-1.06677242e-01 1.34659135e+00 -9.54760313e-01 -3.43725055e-01
-5.43934822e-01 5.70805192e-01 4.78613377e-03 1.59976140e-01
-7.29892075e-01 1.99391842e-01 2.85738051e-01 2.27593556e-01
4.20515060e-01 -3.06996197e-01 -5.83340824e-01 -1.20623803e+00
1.54976264e-01 -3.40787560e-01 5.16338229e-01 -3.86721045e-01
-1.28658712e+00 7.25742459e-01 -1.70463279e-01 -1.42307496e+00
4.12192158e-02 -5.71295440e-01 -7.83707201e-01 9.26183701e-01
-1.94041038e+00 -1.32370090e+00 -5.39551258e-01 4.81441498e-01
3.79441291e-01 2.24696875e-01 7.35645115e-01 4.67708021e-01
-4.40544009e-01 1.50444642e-01 3.76568772e-02 3.73376220e-01
4.44321215e-01 -1.35980761e+00 6.68006167e-02 6.72851920e-01
-6.10064447e-01 5.53004503e-01 2.03453094e-01 -6.94581807e-01
-1.42575002e+00 -1.26581109e+00 4.37149853e-01 2.75703240e-02
4.59133804e-01 1.40393868e-01 -1.12082887e+00 8.85525525e-01
-5.50161302e-02 8.87652099e-01 7.97281504e-01 -5.21788597e-01
3.53244543e-02 1.58942685e-01 -1.54864740e+00 2.40800470e-01
6.28355980e-01 1.42172307e-01 -6.52686536e-01 5.39721139e-02
7.13582516e-01 -1.06696427e+00 -1.31420898e+00 7.23715603e-01
6.63595855e-01 -1.03923941e+00 1.00439405e+00 -4.66461666e-02
4.42452461e-01 -2.54838198e-01 8.21929723e-02 -1.32453072e+00
-4.06905562e-01 -1.72946557e-01 -1.35344818e-01 5.53982735e-01
1.29399627e-01 -4.35017824e-01 9.44291830e-01 6.72787011e-01
-8.52984846e-01 -1.13580370e+00 -1.12793994e+00 -3.32945257e-01
4.02785629e-01 -2.78007627e-01 6.18258178e-01 9.76629615e-01
1.71003379e-02 -3.64818841e-01 5.40759861e-01 3.87002200e-01
1.02201116e+00 1.76891059e-01 9.84967873e-02 -1.20479083e+00
1.17825553e-01 -6.89243138e-01 -3.58388424e-01 -9.85442758e-01
-3.89552981e-01 -1.24352682e+00 -2.48883460e-02 -1.70868731e+00
2.05951288e-01 -4.32631910e-01 -5.63151658e-01 9.12361145e-02
-1.89367369e-01 3.87916639e-02 1.41200036e-01 2.08808139e-01
-4.65189338e-01 4.90321100e-01 2.05913353e+00 8.34373757e-02
-2.82116801e-01 -1.84200536e-02 -2.93664187e-01 8.16210687e-01
5.04778683e-01 -3.53307903e-01 -7.47532025e-02 -6.09521627e-01
-4.97682065e-01 5.27929842e-01 5.64706624e-01 -1.01087129e+00
5.90762198e-01 1.89524353e-01 7.46401131e-01 -8.26163232e-01
-3.90418917e-02 -1.22816360e+00 -6.40148437e-03 6.27813399e-01
-2.19052747e-01 -1.53689936e-01 2.27808639e-01 2.63581395e-01
-3.68549138e-01 6.11077957e-02 1.07754660e+00 -5.77682257e-01
-2.83483505e-01 8.05635273e-01 -1.64599538e-01 -3.02609563e-01
1.12107885e+00 -2.22515941e-01 1.74598992e-01 3.57079536e-01
-8.93043935e-01 4.58543509e-01 3.02575618e-01 -2.89474837e-02
6.92499816e-01 -1.32664216e+00 -6.81683838e-01 6.20548129e-01
-3.95902991e-02 1.04076099e+00 6.57173693e-01 1.51004541e+00
-1.04434597e+00 3.68889511e-01 -3.72847885e-01 -1.06399632e+00
-7.94741333e-01 2.18076512e-01 7.51755953e-01 -5.98450780e-01
-1.15578699e+00 9.12861705e-01 3.09325993e-01 -7.22585857e-01
2.72753060e-01 -1.18078494e+00 7.60147497e-02 -2.81441033e-01
3.50676268e-01 -6.18587807e-02 2.43730038e-01 -5.20682395e-01
-3.77837837e-01 4.43164110e-01 -2.08893612e-01 3.58296871e-01
1.58634520e+00 -8.28135684e-02 9.47566412e-04 -5.35072982e-02
1.28000176e+00 -6.57712400e-01 -1.78721762e+00 -4.64589775e-01
-1.01453491e-01 -1.66182831e-01 6.44552946e-01 -1.00437474e+00
-1.46780908e+00 7.84151256e-01 8.81288946e-01 -5.55088259e-02
9.87179339e-01 -2.35023916e-01 1.28046775e+00 -4.58412707e-01
2.33429134e-01 -4.50577796e-01 -3.42803031e-01 4.31358337e-01
6.91285729e-01 -1.55311799e+00 -1.93544570e-02 -5.29946625e-01
-3.56998295e-01 1.43624902e+00 5.53568244e-01 -3.33098918e-01
8.78919661e-01 6.75234139e-01 4.55709547e-01 -2.47149348e-01
-1.98041424e-01 1.02802232e-01 3.48517835e-01 3.86795163e-01
5.13268590e-01 -4.45639802e-04 -2.06660151e-01 9.47254419e-01
1.03305869e-01 5.07573426e-01 7.30128288e-02 1.26929438e+00
-3.41928333e-01 -6.83225572e-01 -2.93113291e-01 3.21388632e-01
-5.23592770e-01 4.30063307e-02 3.95284265e-01 8.28886807e-01
8.43523592e-02 1.13837272e-01 9.04869139e-02 2.70442828e-03
3.71142387e-01 8.00396353e-02 5.80378175e-01 -3.68235737e-01
-1.12170136e+00 6.57635152e-01 -5.62222481e-01 -5.26906729e-01
-2.78257072e-01 -5.01626492e-01 -1.75343382e+00 9.41515043e-02
-9.29258391e-02 -3.16894829e-01 9.90694702e-01 8.88098121e-01
1.83864459e-01 9.57448721e-01 3.97875011e-01 -1.25004172e+00
-4.94658828e-01 -9.11741674e-01 -5.22419930e-01 3.41127396e-01
6.43901110e-01 -4.69572544e-01 -1.64214969e-01 -2.91951988e-02] | [14.491666793823242, -2.7013087272644043] |
2aae29fe-2140-459f-87cd-f28d96ec5c13 | who-should-i-engage-with-at-what-time-a | 2301.08399 | null | https://arxiv.org/abs/2301.08399v1 | https://arxiv.org/pdf/2301.08399v1.pdf | Who Should I Engage with At What Time? A Missing Event Aware Temporal Graph Neural Network | Temporal graph neural network has recently received significant attention due to its wide application scenarios, such as bioinformatics, knowledge graphs, and social networks. There are some temporal graph neural networks that achieve remarkable results. However, these works focus on future event prediction and are performed under the assumption that all historical events are observable. In real-world applications, events are not always observable, and estimating event time is as important as predicting future events. In this paper, we propose MTGN, a missing event-aware temporal graph neural network, which uniformly models evolving graph structure and timing of events to support predicting what will happen in the future and when it will happen.MTGN models the dynamic of both observed and missing events as two coupled temporal point processes, thereby incorporating the effects of missing events into the network. Experimental results on several real-world temporal graphs demonstrate that MTGN significantly outperforms existing methods with up to 89% and 112% more accurate time and link prediction. Code can be found on https://github.com/HIT-ICES/TNNLS-MTGN. | ['Zhongjie Wang', 'Xiaofei Xu', 'Zhiying Tu', 'Mingyi Liu'] | 2023-01-20 | null | null | null | null | ['point-processes'] | ['methodology'] | [-1.19906152e-03 4.37133573e-02 -4.08427656e-01 -2.55375415e-01
2.31301650e-01 -2.62424767e-01 5.34287214e-01 7.23411739e-01
-8.55609868e-03 8.21863055e-01 1.82930723e-01 -5.25418818e-01
-3.40739965e-01 -1.32636011e+00 -5.76126218e-01 -3.54360372e-01
-7.77074814e-01 5.69888711e-01 5.05530238e-01 -2.57796869e-02
-5.11825264e-01 3.73393297e-01 -9.20870125e-01 -1.45773813e-01
6.54493690e-01 6.67473912e-01 -1.30444318e-01 6.33102477e-01
-1.17984429e-01 1.19169712e+00 -5.23729205e-01 -3.88415396e-01
-1.12638168e-01 -4.10822630e-01 -4.41034943e-01 -3.81631315e-01
-3.88987601e-01 -3.14096026e-02 -1.25241637e+00 6.45263255e-01
3.20182472e-01 2.55780846e-01 3.62697095e-01 -1.55098522e+00
-6.40746772e-01 9.41343606e-01 -5.50179005e-01 6.07810378e-01
1.74597844e-01 7.81420171e-02 7.16625750e-01 -1.11700535e-01
7.96900928e-01 1.02804863e+00 6.66997015e-01 1.74179912e-01
-8.35098207e-01 -7.57417977e-01 4.42086011e-01 6.21585608e-01
-1.39780974e+00 6.75491616e-02 9.95498717e-01 -3.30739349e-01
1.01157677e+00 1.17226236e-01 9.17493224e-01 1.24679017e+00
7.37739921e-01 7.19227076e-01 4.93771404e-01 2.23779436e-02
1.04143091e-01 -6.86439395e-01 1.57656118e-01 4.57505733e-01
1.84681788e-01 2.51172453e-01 -6.55727327e-01 -1.93933502e-01
6.64318204e-01 7.24665403e-01 -2.76204169e-01 1.03754364e-01
-1.24090385e+00 5.33036888e-01 6.90240741e-01 4.40810174e-01
-7.32620955e-01 3.37993175e-01 5.88833332e-01 2.11800992e-01
9.61883903e-01 -1.96527839e-01 -4.99194384e-01 -1.02099083e-01
-6.60085320e-01 -1.28216639e-01 8.54052484e-01 8.50764215e-01
3.94720286e-01 2.97882468e-01 -2.10521877e-01 4.04487491e-01
-1.55208101e-02 2.37839356e-01 2.98466444e-01 -3.86620998e-01
2.75491953e-01 8.51407707e-01 -1.02758244e-01 -1.45754147e+00
-8.32743585e-01 -6.26709521e-01 -1.34159839e+00 -6.32899463e-01
4.00683433e-01 -3.63730341e-01 -1.01119387e+00 1.84194815e+00
5.23865461e-01 1.13272619e+00 -1.97826475e-01 4.78654653e-01
9.92227614e-01 1.04941845e+00 2.83303320e-01 -6.07191384e-01
9.62486446e-01 -4.21935588e-01 -1.00230527e+00 -9.14974138e-02
5.03021657e-01 -3.05587769e-01 2.30351597e-01 7.62513205e-02
-7.37601638e-01 -1.94899112e-01 -5.72167099e-01 3.02493036e-01
-5.93898773e-01 -2.80200452e-01 1.09975100e+00 -4.58368286e-02
-9.21279430e-01 7.45209455e-01 -1.21140075e+00 -7.10641861e-01
1.17208526e-01 1.75802320e-01 -1.97782014e-02 -4.24999855e-02
-1.72421241e+00 5.34089684e-01 9.24061120e-01 4.70101684e-01
-9.31768537e-01 -5.50546050e-01 -7.83360302e-01 2.35157803e-01
6.90129876e-01 -5.51917791e-01 9.77469802e-01 -3.35883200e-01
-8.08815181e-01 2.07886264e-01 -2.72055298e-01 -6.54959917e-01
4.72295821e-01 3.48004699e-01 -1.22138286e+00 -1.84826508e-01
-8.26766714e-02 9.89802834e-03 1.86060935e-01 -4.61582631e-01
-5.10015845e-01 -3.54296774e-01 -2.08321378e-01 -1.79441720e-02
-4.91161525e-01 -2.39573464e-01 -7.94042587e-01 -7.30060220e-01
-1.47453314e-02 -8.09511244e-01 -3.62656623e-01 -3.30610484e-01
-4.69343811e-01 -7.15166628e-01 9.84394014e-01 -7.92868614e-01
1.70466435e+00 -1.86600053e+00 -2.80292053e-02 1.19821914e-02
4.84149158e-01 2.37318985e-02 4.47247811e-02 1.16296828e+00
-7.43478462e-02 -1.30156532e-01 -5.79524636e-02 1.83354300e-02
-3.34353656e-01 4.35452670e-01 -2.86423087e-01 3.44744533e-01
-8.90820846e-02 1.03405535e+00 -1.17555833e+00 -3.46785933e-01
1.38985589e-01 6.09856427e-01 3.44257653e-01 -1.21950712e-02
-5.00286341e-01 4.26171720e-01 -5.51711321e-01 5.38894892e-01
3.04457724e-01 -8.96005213e-01 6.70720994e-01 2.41889149e-01
2.25491628e-01 -3.10279615e-02 -8.82911146e-01 1.24701738e+00
-1.84032675e-02 7.56406724e-01 -6.96829855e-01 -1.07125330e+00
1.01147282e+00 5.40329754e-01 7.76166379e-01 -8.20743203e-01
1.29541665e-01 -1.63531005e-01 6.73042610e-04 -2.98808336e-01
3.59844267e-01 1.44819796e-01 -6.24567131e-03 4.51687813e-01
-5.93147613e-02 6.11810207e-01 5.31705081e-01 6.06324196e-01
1.56279683e+00 -5.32054305e-02 3.35480005e-01 3.37075919e-01
3.41312103e-02 -1.19887188e-01 1.03197765e+00 5.82227647e-01
-3.12352508e-01 1.51366532e-01 7.70348549e-01 -8.19138646e-01
-6.83470845e-01 -1.00583792e+00 3.56680751e-01 7.34819889e-01
6.93663955e-02 -3.95395994e-01 -4.84834276e-02 -5.63232780e-01
1.14082478e-01 7.74123430e-01 -8.77307236e-01 -3.44610184e-01
-5.64291060e-01 -1.16504800e+00 3.41488987e-01 6.00969136e-01
8.60245898e-02 -1.30483365e+00 -2.14643404e-02 7.21541166e-01
-3.57317775e-01 -1.03096282e+00 -3.84304374e-01 4.70171357e-03
-8.16296041e-01 -1.28685284e+00 -4.91231620e-01 -5.61707199e-01
3.25800508e-01 5.40285930e-02 1.16099846e+00 9.22949389e-02
-3.65581453e-01 2.59100318e-01 -2.28612676e-01 -7.22248793e-01
-2.72998840e-01 -4.03669867e-04 -3.42507921e-02 1.56709000e-01
4.05590504e-01 -1.06890309e+00 -7.11242259e-01 1.34767160e-01
-9.54811573e-01 9.45398957e-02 2.27802768e-01 5.72621405e-01
6.05641663e-01 4.96685922e-01 8.79832089e-01 -8.92265499e-01
4.78103697e-01 -1.00088060e+00 -6.36300266e-01 3.91688138e-01
-8.22614253e-01 -4.72761393e-01 7.66574800e-01 -6.71869934e-01
-9.75603819e-01 -2.57163644e-01 2.18502849e-01 -4.14783567e-01
9.70791653e-02 1.12141728e+00 2.59845853e-01 5.08948982e-01
3.49420130e-01 5.11270046e-01 -4.33902770e-01 -1.06297620e-01
9.08744782e-02 1.27438739e-01 3.53383929e-01 -1.38713926e-01
4.16592330e-01 5.00697911e-01 1.93588033e-01 -5.35170794e-01
-6.39096975e-01 -5.93474627e-01 -2.58599997e-01 -6.77013636e-01
4.65638310e-01 -7.47213125e-01 -8.68010640e-01 6.62943125e-01
-1.02800417e+00 -6.00390851e-01 -2.10163653e-01 6.31179452e-01
-9.97667685e-02 2.39136502e-01 -9.56159890e-01 -7.54324675e-01
-3.89532954e-01 -3.70358378e-01 4.62320000e-01 3.89347732e-01
1.83863670e-03 -1.55283558e+00 2.35920608e-01 -2.36714467e-01
3.42672199e-01 8.65154505e-01 7.44281113e-01 -7.53686368e-01
-6.53008163e-01 -4.46454316e-01 -1.83688015e-01 -5.19876540e-01
3.17984164e-01 1.03732966e-01 -4.10827219e-01 -3.03686559e-01
-5.47383130e-01 3.20000678e-01 9.45510864e-01 7.46855557e-01
1.12295008e+00 -3.35952938e-01 -8.03007364e-01 4.98777896e-01
1.32455778e+00 5.98421276e-01 4.97588336e-01 -7.35032186e-02
8.09584498e-01 5.39191067e-01 3.56466919e-01 5.90872586e-01
7.93757319e-01 3.36000472e-01 6.94674790e-01 -5.22035956e-02
-8.00937489e-02 -5.46336055e-01 7.94600248e-02 1.02630436e+00
-1.10307083e-01 -1.00571907e+00 -1.21748340e+00 9.42914009e-01
-2.41956282e+00 -1.22854459e+00 -5.75025141e-01 2.06498694e+00
4.54850346e-01 2.22640499e-01 1.28545791e-01 -5.13997488e-02
9.52651501e-01 4.88419235e-01 -9.69777524e-01 1.19727299e-01
-2.60783911e-01 -2.11029872e-01 4.65704441e-01 1.78127870e-01
-9.58468497e-01 8.59213054e-01 5.17501926e+00 7.29504168e-01
-1.11610389e+00 3.87608334e-02 8.01447809e-01 -1.32968113e-01
9.92326252e-03 7.26689845e-02 -4.12131876e-01 6.34512007e-01
1.21858370e+00 -8.18775773e-01 2.86134064e-01 3.01333070e-01
5.26802003e-01 2.74788797e-01 -6.84901297e-01 6.21138036e-01
-3.47350061e-01 -1.36520743e+00 -4.23351862e-02 1.03768937e-01
5.52310705e-01 3.98317456e-01 -3.21740866e-01 4.60904658e-01
8.99307728e-01 -7.13118970e-01 6.32984638e-02 9.61634874e-01
5.22934020e-01 -8.47675741e-01 6.61288381e-01 5.16330719e-01
-1.62219584e+00 8.37008879e-02 -7.54513070e-02 -2.29269505e-01
5.67948341e-01 1.05483902e+00 -1.06712770e+00 1.05876303e+00
7.40749061e-01 1.39640880e+00 -3.88115883e-01 1.04212034e+00
-2.18387023e-01 1.11886358e+00 -5.03838658e-01 -2.12299660e-01
2.55903769e-02 -9.40863937e-02 6.08922541e-01 9.69011545e-01
4.84463066e-01 1.61650494e-01 4.49468315e-01 3.73755455e-01
-1.90438643e-01 -2.53868885e-02 -7.42943108e-01 -3.81411284e-01
4.42777604e-01 1.12496817e+00 -9.85443652e-01 -4.13209617e-01
-3.83990854e-01 7.72804856e-01 4.86735761e-01 7.03719318e-01
-9.94647980e-01 -3.99620503e-01 2.66921610e-01 -2.37775892e-02
2.26121858e-01 -2.91841060e-01 3.25623304e-01 -1.21743119e+00
-7.18823867e-03 -2.03710333e-01 1.04768264e+00 -8.97782564e-01
-1.41784525e+00 5.93857586e-01 -2.19206333e-01 -1.09011078e+00
-2.25789532e-01 -1.19768910e-01 -9.20799315e-01 4.97147799e-01
-1.20952725e+00 -1.38960695e+00 -2.40953282e-01 6.34734154e-01
2.16303200e-01 1.86663672e-01 4.02580917e-01 2.91979581e-01
-8.27788055e-01 7.63261542e-02 1.75695881e-01 3.70825231e-01
4.61576760e-01 -1.13383937e+00 7.11689591e-01 1.07552874e+00
8.05312172e-02 2.18528956e-01 6.86158359e-01 -1.24973261e+00
-1.32528758e+00 -1.64031887e+00 1.04247630e+00 -5.09741083e-02
1.13536370e+00 -1.45524725e-01 -1.21661520e+00 1.06323504e+00
9.99781936e-02 1.51475757e-01 4.71175700e-01 2.75948733e-01
-4.92804721e-02 -2.24098742e-01 -6.72215998e-01 6.49396062e-01
1.24133158e+00 -2.01848313e-01 8.08851942e-02 7.41931915e-01
1.00917721e+00 -4.94250238e-01 -1.07797420e+00 5.51672459e-01
2.40444407e-01 -4.65319067e-01 6.99382603e-01 -6.47460461e-01
1.78213000e-01 -3.09310734e-01 4.22979623e-01 -1.36688733e+00
-5.45318246e-01 -5.81918299e-01 -8.34478796e-01 1.08809662e+00
3.37248653e-01 -1.01706743e+00 8.02733958e-01 2.71869272e-01
1.45478234e-01 -6.47926509e-01 -1.02522624e+00 -9.23588276e-01
-5.03962219e-01 -5.25985301e-01 9.27727222e-01 1.34057617e+00
-5.04473783e-02 3.64717096e-01 -8.14884186e-01 3.89308035e-01
5.20843923e-01 1.85574472e-01 4.59459424e-01 -1.35717082e+00
-1.98379904e-01 -2.95400888e-01 -5.50258219e-01 -5.57289243e-01
7.54675362e-04 -8.77720535e-01 -3.75251174e-01 -1.96235168e+00
2.02788964e-01 -7.99888894e-02 -6.68196738e-01 7.52479732e-01
-4.19969171e-01 -2.37988308e-01 -2.36571357e-01 4.92462553e-02
-7.73008943e-01 6.99670374e-01 1.08674157e+00 -2.21962869e-01
-2.61120439e-01 1.72197208e-01 -1.74251184e-01 4.60812598e-01
1.04815626e+00 -4.37193841e-01 -5.43356538e-01 -2.57901967e-01
4.05412763e-01 8.64251733e-01 3.90304834e-01 -9.11976457e-01
6.06004715e-01 -3.63224834e-01 2.73428530e-01 -8.55224729e-01
2.78188199e-01 -7.14356005e-01 8.92691672e-01 5.30092895e-01
-1.20539837e-01 3.77055764e-01 2.03564793e-01 1.39840651e+00
-2.49691874e-01 5.58208525e-01 -6.28233030e-02 -4.68434487e-03
-8.63234401e-01 1.21386087e+00 -1.99493706e-01 -7.27481470e-02
1.18324888e+00 1.76438794e-01 -5.63547015e-01 -9.06445742e-01
-1.04367113e+00 7.46448278e-01 1.57119501e-02 5.61819613e-01
5.50204337e-01 -1.18302798e+00 -7.93743193e-01 -5.18873215e-01
1.51409790e-01 -1.66673154e-01 7.20133483e-01 8.53968561e-01
-1.64688498e-01 3.54512542e-01 4.35291380e-01 -4.66174990e-01
-1.14102340e+00 9.63048100e-01 1.88199326e-01 -6.36853933e-01
-7.76649833e-01 4.18368965e-01 1.26102110e-02 -3.23583245e-01
1.63369551e-01 5.13280928e-02 -3.28388512e-01 -2.65848357e-02
3.05242777e-01 4.50307697e-01 -1.09728023e-01 -3.36884856e-01
-2.69448221e-01 -6.68879524e-02 -8.68071690e-02 4.17782098e-01
1.59581137e+00 -5.24643846e-02 -1.82235003e-01 8.32574546e-01
7.94458687e-01 -5.06343842e-01 -9.99513388e-01 -6.93957746e-01
7.18340576e-02 1.04740843e-01 -7.20837787e-02 -7.14092672e-01
-1.33507586e+00 6.54554784e-01 2.47523442e-01 6.89269960e-01
1.34272099e+00 1.24327734e-01 1.14914548e+00 6.44937083e-02
4.38908905e-01 -6.42385423e-01 -3.32658947e-01 4.15529132e-01
6.79310858e-01 -1.02858579e+00 4.72514071e-02 -4.13089067e-01
-2.72620887e-01 9.75068927e-01 6.53080702e-01 1.30496487e-01
1.05399716e+00 -1.00287504e-01 -3.39319587e-01 -4.41131651e-01
-1.20211577e+00 -1.86757445e-01 1.99499741e-01 3.62941772e-01
3.59337598e-01 4.80931461e-01 -3.25332224e-01 1.82169303e-01
8.29483047e-02 2.93816924e-01 5.12466311e-01 7.73140728e-01
9.00088996e-02 -7.51928627e-01 1.32772595e-01 8.43679011e-01
-3.65633905e-01 -6.01032004e-02 -2.34887779e-01 7.52781570e-01
-2.41707429e-01 9.01070237e-01 5.06178848e-02 -3.67109537e-01
2.48261347e-01 -1.23660289e-01 2.19797622e-02 -3.00515175e-01
-2.94916719e-01 -1.05337642e-01 1.60040185e-01 -5.24232447e-01
-2.07225263e-01 -6.57060742e-01 -1.24745882e+00 -5.11242807e-01
-2.11692706e-01 5.53608648e-02 2.04366297e-01 6.56254888e-01
7.34434664e-01 1.06336653e+00 4.66725677e-01 -2.72117436e-01
3.55368137e-01 -1.00836444e+00 -6.27776444e-01 2.30126470e-01
1.49286106e-01 -5.59543848e-01 -2.43476346e-01 -1.01905495e-01] | [7.241791248321533, 5.781636714935303] |
43e45d2e-1072-43de-aced-ccc129acc78d | enforcenet-monocular-camera-localization-in | 1907.07160 | null | https://arxiv.org/abs/1907.07160v1 | https://arxiv.org/pdf/1907.07160v1.pdf | EnforceNet: Monocular Camera Localization in Large Scale Indoor Sparse LiDAR Point Cloud | Pose estimation is a fundamental building block for robotic applications such as autonomous vehicles, UAV, and large scale augmented reality. It is also a prohibitive factor for those applications to be in mass production, since the state-of-the-art, centimeter-level pose estimation often requires long mapping procedures and expensive localization sensors, e.g. LiDAR and high precision GPS/IMU, etc. To overcome the cost barrier, we propose a neural network based solution to localize a consumer degree RGB camera within a prior sparse LiDAR map with comparable centimeter-level precision. We achieved it by introducing a novel network module, which we call resistor module, to enforce the network generalize better, predicts more accurately, and converge faster. Such results are benchmarked by several datasets we collected in the large scale indoor parking garage scenes. We plan to open both the data and the code for the community to join the effort to advance this field. | ['Yu Chen', 'Guan Wang'] | 2019-07-16 | null | null | null | null | ['camera-localization'] | ['computer-vision'] | [-3.82992066e-03 3.39973122e-02 -2.89421737e-01 -5.64675391e-01
-4.98804122e-01 -2.82009095e-01 2.17629388e-01 -1.13007441e-01
-5.48586190e-01 9.98749077e-01 -3.84917974e-01 -3.08234394e-01
-1.75601795e-01 -1.09624755e+00 -1.03546727e+00 -1.52470589e-01
-1.67589679e-01 8.09107304e-01 2.50255942e-01 -1.79399639e-01
2.03698516e-01 8.33474278e-01 -1.55010474e+00 -4.91064817e-01
8.38543952e-01 1.22358346e+00 5.51520348e-01 2.09713370e-01
2.53491670e-01 3.41705561e-01 -1.61117893e-02 4.76995297e-02
6.02002323e-01 2.46215016e-01 -2.05207825e-01 -5.88231310e-02
4.80773211e-01 -3.62731844e-01 -2.72460669e-01 9.91195858e-01
4.13808376e-01 1.13351889e-01 1.69382021e-01 -1.35458243e+00
-3.91635120e-01 3.26986820e-01 -5.64325094e-01 -4.20318931e-01
1.45015597e-01 -8.32303539e-02 5.70393384e-01 -8.62692714e-01
6.27476990e-01 1.02676845e+00 1.13463569e+00 1.53895915e-01
-9.75397170e-01 -6.90933883e-01 -8.31919373e-04 2.21519083e-01
-1.87331104e+00 -3.48508894e-01 8.68615270e-01 -1.04288012e-01
9.53105867e-01 4.44769487e-02 6.49362862e-01 8.67375433e-01
2.65949517e-01 4.27187145e-01 1.07260871e+00 -3.90902087e-02
3.34195048e-01 1.46203250e-01 -3.41290295e-01 8.97287667e-01
4.08780247e-01 2.96769291e-01 -3.87035161e-01 1.32845238e-01
1.00942981e+00 3.38404357e-01 -4.93823402e-02 -1.01298058e+00
-1.22029364e+00 9.30391967e-01 1.04215884e+00 -3.25748846e-02
-5.25560439e-01 5.34586251e-01 -1.27730429e-01 4.74600717e-02
1.95501506e-01 5.12901366e-01 -5.97502530e-01 -2.39372492e-01
-9.20462012e-01 1.76346570e-01 7.23393679e-01 1.28044963e+00
1.26193523e+00 -1.34015277e-01 7.78929710e-01 4.93376672e-01
3.90481383e-01 5.64220846e-01 2.27969334e-01 -1.28769588e+00
5.26442528e-01 6.61896467e-01 3.36248845e-01 -1.09255648e+00
-6.21608138e-01 -3.28910619e-01 -1.05373752e+00 1.16536550e-01
-1.34276852e-01 -8.11568946e-02 -8.00302148e-01 1.38545442e+00
2.51084089e-01 2.87003815e-01 -1.28859669e-01 1.02042627e+00
5.07605135e-01 4.38287765e-01 -3.60512823e-01 1.14982001e-01
9.25421655e-01 -6.75800443e-01 -3.09688568e-01 -7.24108338e-01
5.73191404e-01 -6.58722222e-01 9.54905212e-01 3.16394597e-01
-7.61001825e-01 -5.98636389e-01 -1.44820666e+00 -3.10412031e-02
-5.13398409e-01 4.42797214e-01 9.88277793e-01 4.09898549e-01
-1.09441781e+00 9.36187029e-01 -1.11942637e+00 -6.87142134e-01
2.10667759e-01 6.88588321e-01 -5.42415321e-01 -2.04467818e-01
-9.78730679e-01 1.33748591e+00 5.31425118e-01 3.46709222e-01
-4.48345214e-01 -4.59724784e-01 -1.09650278e+00 -2.40603849e-01
4.05351132e-01 -6.46012783e-01 1.03255177e+00 -1.29985616e-01
-1.43412662e+00 6.09577596e-01 -5.16350605e-02 -7.32746363e-01
4.51788276e-01 -4.28781748e-01 -4.12340984e-02 -3.62176836e-01
2.40361124e-01 1.13925803e+00 4.86696362e-01 -1.24756086e+00
-5.12369812e-01 -6.65729582e-01 -8.18442628e-02 1.44860730e-01
-4.77434695e-02 -4.94891703e-01 -2.28069648e-01 -1.73566211e-02
7.08219111e-01 -1.38494360e+00 -6.56256855e-01 2.38611862e-01
-3.42932075e-01 -7.98032433e-02 1.00364745e+00 -4.71065998e-01
4.80453014e-01 -1.99779689e+00 -1.25961572e-01 8.63631442e-02
1.19748190e-02 -2.30644062e-01 9.54441056e-02 2.11889014e-01
2.07731143e-01 -2.37989187e-01 -1.92317098e-01 -5.71710765e-01
1.38907462e-01 4.78666186e-01 -2.29995608e-01 7.47962594e-01
1.84279978e-01 8.33515406e-01 -7.81509161e-01 -2.40720391e-01
5.72693706e-01 5.66341639e-01 -4.74550515e-01 -2.00404838e-01
-1.14148714e-01 4.39973950e-01 -4.11388248e-01 8.24342906e-01
8.26311052e-01 -3.28435570e-01 -1.17838681e-01 -3.05993915e-01
-4.51816022e-01 2.99182713e-01 -1.49990571e+00 2.19485497e+00
-7.72426784e-01 5.81389308e-01 1.45805851e-01 -8.99775922e-01
1.33096051e+00 -1.46185368e-01 5.41136742e-01 -5.88462651e-01
2.52107441e-01 3.71072590e-01 -3.03642809e-01 6.70025572e-02
1.15485418e+00 7.90391266e-02 -3.61053050e-01 -6.70484677e-02
-2.62840033e-01 -6.13733292e-01 -8.34108144e-02 -4.01350968e-02
8.46778989e-01 5.50796866e-01 4.13401693e-01 -7.55018741e-02
2.27683693e-01 2.94047266e-01 4.57782567e-01 5.48319876e-01
-1.21130705e-01 4.44679946e-01 -2.19550043e-01 -6.60563409e-01
-1.14254975e+00 -9.76048887e-01 -1.21662989e-01 5.00394702e-01
4.77579802e-01 -2.17346638e-01 -2.32641831e-01 -2.15250432e-01
3.52502018e-01 3.77036721e-01 -1.16420269e-01 5.46266250e-02
-7.50646770e-01 -2.97792882e-01 2.98912257e-01 9.15134966e-01
9.82025623e-01 -6.08805239e-01 -7.67982423e-01 3.20305526e-01
-2.19714474e-02 -1.44982767e+00 1.62077829e-01 5.06648540e-01
-9.40202415e-01 -8.21445286e-01 -1.88109830e-01 -6.22507870e-01
4.94305164e-01 3.76186311e-01 9.19022918e-01 -8.70775878e-02
-2.43312255e-01 7.65063390e-02 -8.49657506e-02 -3.08714151e-01
1.65528357e-01 2.98733830e-01 5.77121079e-01 -7.41677582e-01
2.69495279e-01 -9.18256640e-01 -4.43585277e-01 4.86106753e-01
-1.90450966e-01 2.62220018e-02 8.48034501e-01 5.31073987e-01
8.84011865e-01 9.22651216e-02 2.40777165e-01 -2.88753003e-01
2.79872507e-01 -2.67352164e-01 -1.24994171e+00 -3.51571232e-01
-6.41210973e-01 -8.43623281e-02 4.89068091e-01 -2.98517823e-01
-2.15027303e-01 9.24624085e-01 -2.43473947e-01 -3.17631125e-01
-2.06533328e-01 5.24362147e-01 -1.83103368e-01 -7.03971028e-01
5.81537783e-01 1.37541369e-01 -1.12023249e-01 -2.67317057e-01
5.70086002e-01 6.86382234e-01 8.32937360e-01 -4.47979867e-01
1.01370597e+00 4.26093161e-01 3.78783852e-01 -8.91981184e-01
-4.50946659e-01 -5.18584490e-01 -7.93275714e-01 4.40295450e-02
7.09780753e-01 -1.37418103e+00 -8.03520262e-01 -7.94199295e-03
-1.31909394e+00 -3.24905634e-01 -1.05716534e-01 6.33221090e-01
-7.22186685e-01 2.46776849e-01 -2.19060764e-01 -6.68541610e-01
-7.11961091e-02 -1.15005994e+00 1.30734825e+00 3.94952238e-01
-2.18373820e-01 -5.06683767e-01 -1.90010015e-02 1.20379917e-01
5.17280281e-01 3.82793725e-01 1.61748052e-01 -3.75731699e-02
-1.39445662e+00 -5.16401887e-01 -4.53899592e-01 6.09710030e-02
2.38587614e-02 -3.04006308e-01 -8.36411715e-01 -3.12840343e-01
-1.43471494e-01 -3.77976269e-01 4.96312827e-01 1.98623136e-01
1.04426432e+00 7.41103590e-02 -6.75140440e-01 7.60175586e-01
1.51640904e+00 -5.46838455e-02 5.83316386e-01 4.84820336e-01
6.46267831e-01 2.03368500e-01 9.24380600e-01 1.99154079e-01
7.80906856e-01 8.22337329e-01 7.91159272e-01 5.55586480e-02
2.75701851e-01 -4.98093307e-01 5.44578917e-02 8.40240419e-01
-1.39311552e-01 1.24055117e-01 -9.35295701e-01 5.22559822e-01
-1.90898883e+00 -3.92743468e-01 -1.46072274e-02 2.13233805e+00
3.32501113e-01 1.32820114e-01 -2.11375400e-01 1.10312708e-01
3.97800803e-01 7.10650459e-02 -4.40643549e-01 -2.28278022e-02
2.41417229e-01 6.43894449e-02 1.14081800e+00 4.65305626e-01
-1.10224962e+00 1.24229002e+00 5.76352644e+00 3.82685781e-01
-1.36538661e+00 4.68793027e-02 1.07626416e-01 2.39541426e-01
1.76838413e-01 7.56934211e-02 -1.01245356e+00 3.29526097e-01
9.18573380e-01 1.98304400e-01 7.09621608e-01 1.53968453e+00
1.25895068e-01 -3.93246442e-01 -1.09821832e+00 1.24776673e+00
-1.72597412e-02 -1.45326519e+00 -7.41373241e-01 3.23734373e-01
5.75032771e-01 7.43264675e-01 -1.76212147e-01 3.37889999e-01
4.84072655e-01 -9.09099996e-01 3.41067672e-01 1.78806692e-01
8.04094076e-01 -7.58256674e-01 6.87777162e-01 6.89894736e-01
-1.43592691e+00 -6.47306582e-03 -8.46750796e-01 -4.67729598e-01
3.34450871e-01 5.83949447e-01 -1.13698673e+00 4.76117164e-01
7.11820483e-01 6.19038284e-01 -4.05704558e-01 1.16601002e+00
-3.96082729e-01 -2.45581165e-01 -7.44836092e-01 -2.63899475e-01
1.49588361e-01 -2.39398614e-01 3.13951880e-01 4.93433923e-01
7.83307493e-01 -2.62529224e-01 5.97481728e-01 7.79292941e-01
-1.35145590e-01 -3.44024092e-01 -1.05746090e+00 9.62363183e-02
6.93239629e-01 1.42250049e+00 -8.28154385e-01 -7.94063434e-02
-1.38521314e-01 9.92054582e-01 2.72693306e-01 -9.82222240e-03
-7.55888104e-01 -3.97480309e-01 6.29550338e-01 1.56858891e-01
3.51448327e-01 -1.01366544e+00 -4.39960629e-01 -8.98210406e-01
1.95161924e-01 -3.50755870e-01 -5.51097631e-01 -1.02603376e+00
-8.95414829e-01 5.41641474e-01 -2.13298410e-01 -1.35523236e+00
-5.61024845e-01 -8.19904327e-01 -1.43668264e-01 8.12340736e-01
-1.73674572e+00 -1.12880838e+00 -8.02577019e-01 4.71830219e-01
1.62513644e-01 6.08825274e-02 8.66100311e-01 5.17094910e-01
-1.64223358e-01 9.78920311e-02 -2.41325945e-01 -7.75333494e-02
6.16456747e-01 -9.10661161e-01 6.39108062e-01 6.94362581e-01
9.45317596e-02 7.61193156e-01 6.70780301e-01 -8.17569971e-01
-1.89155793e+00 -1.34909320e+00 8.74516249e-01 -5.65604568e-01
7.76449561e-01 -5.93901396e-01 -3.11987221e-01 9.23764527e-01
-2.65628606e-01 2.11759299e-01 -2.23983880e-02 1.67080820e-01
1.88016251e-01 -3.02592158e-01 -1.25168204e+00 2.88675785e-01
1.18001330e+00 -5.28619826e-01 -3.69442642e-01 5.36276877e-01
8.76177609e-01 -8.07093680e-01 -8.69868457e-01 7.86523402e-01
5.41800141e-01 -8.26008439e-01 1.21248722e+00 1.41764790e-01
1.75877646e-01 -6.01946890e-01 -6.38427794e-01 -1.25763881e+00
-2.72246987e-01 -3.10087174e-01 -8.27510282e-02 9.09384370e-01
2.35469192e-01 -8.13723266e-01 1.40653384e+00 3.45308185e-01
-2.35739753e-01 -6.46228552e-01 -1.17160094e+00 -1.00377429e+00
-4.29483980e-01 -5.70464611e-01 8.16262841e-01 8.07723522e-01
-4.21469301e-01 3.77941400e-01 -6.29695535e-01 7.20113873e-01
5.73157847e-01 8.07793662e-02 1.20699334e+00 -1.58044755e+00
5.59776276e-03 9.28776637e-02 -9.38515782e-01 -1.36198628e+00
1.69840619e-01 -6.12621486e-01 4.56555575e-01 -1.60341191e+00
-4.59460586e-01 -9.12083149e-01 1.92997441e-01 2.78276622e-01
4.21872348e-01 5.15150309e-01 1.27153084e-01 4.11643833e-02
-5.92976093e-01 5.79262078e-01 9.79727209e-01 -6.99171498e-02
-1.96653858e-01 1.09001540e-01 -5.00883639e-01 9.38438356e-01
8.61932755e-01 -4.27637100e-01 -2.87815005e-01 -5.26300311e-01
3.39932561e-01 7.47725293e-02 6.18636906e-01 -1.58288443e+00
5.59042335e-01 -1.28981933e-01 5.70810854e-01 -1.20580769e+00
9.82780337e-01 -1.26069129e+00 2.61971086e-01 5.38826823e-01
4.46084738e-01 3.56034547e-01 1.02677837e-01 3.49820912e-01
-8.49503502e-02 1.16734281e-01 5.52080572e-01 -3.26332659e-01
-1.04327738e+00 5.08133888e-01 1.42311305e-01 -5.09692788e-01
1.05502188e+00 -3.57885063e-01 -2.20880255e-01 -3.86329293e-01
-3.29537421e-01 2.57280409e-01 9.11629260e-01 4.07524675e-01
6.59805715e-01 -1.51285005e+00 -1.98198080e-01 2.85864115e-01
1.37737572e-01 3.66490930e-01 -1.60411954e-01 8.13887239e-01
-7.60117650e-01 6.90820098e-01 -2.33387500e-01 -9.74357665e-01
-8.19239378e-01 4.05041456e-01 1.17041141e-01 7.04598054e-02
-5.09306252e-01 6.52654707e-01 -5.93406737e-01 -1.10360801e+00
1.90147609e-01 -6.99144006e-01 3.41527313e-01 -4.47719365e-01
1.99545771e-01 2.96349913e-01 1.45724699e-01 -5.87105632e-01
-5.43597579e-01 7.94165969e-01 4.94098157e-01 1.05643921e-01
1.38673735e+00 -2.89196879e-01 -7.98729509e-02 4.07026023e-01
9.59203541e-01 -1.04517177e-01 -1.25921643e+00 2.81179994e-02
-4.83224867e-03 -4.70981896e-01 1.12465009e-01 -4.08159137e-01
-6.43557608e-01 7.81162500e-01 8.61963272e-01 6.40982091e-02
6.70969069e-01 -1.29056439e-01 8.30846548e-01 1.09192085e+00
1.21713436e+00 -1.06035614e+00 -3.43078882e-01 7.72658885e-01
7.33158410e-01 -1.42787683e+00 3.43193054e-01 -5.16925216e-01
-1.90014347e-01 8.44624817e-01 6.39457524e-01 -6.38110340e-01
4.15777594e-01 4.60180461e-01 5.34924157e-02 2.96681225e-02
-1.63840786e-01 -6.24521598e-02 -7.94304609e-02 8.82019937e-01
1.54182553e-01 3.38027328e-01 9.07761976e-02 2.06648409e-01
-6.92527056e-01 3.00373554e-01 3.61604184e-01 1.11904275e+00
-6.04656994e-01 -1.05425894e+00 -2.94385552e-01 3.03942025e-01
3.18144530e-01 -9.60048065e-02 -2.14789137e-01 1.11497986e+00
3.54033828e-01 5.46620727e-01 9.64023247e-02 -6.65587723e-01
3.84337723e-01 -2.43774071e-01 5.60961306e-01 -4.82870549e-01
4.90765944e-02 -3.36114496e-01 9.28479582e-02 -9.73230541e-01
-2.90627182e-01 -4.13421661e-01 -1.17853272e+00 -4.29351389e-01
-3.14015627e-01 -1.01018503e-01 1.38646150e+00 8.31844926e-01
5.68271935e-01 2.48523653e-01 4.87849474e-01 -1.41598654e+00
-5.19663215e-01 -9.54533994e-01 -4.23564702e-01 -4.75110054e-01
1.67067617e-01 -8.71270299e-01 5.77573180e-02 -4.21815932e-01] | [7.580437183380127, -2.138148546218872] |
4adc754d-b7c4-4992-9eff-327ef3da2f23 | a-survey-and-an-extensive-evaluation-of | 2007.07663 | null | https://arxiv.org/abs/2007.07663v2 | https://arxiv.org/pdf/2007.07663v2.pdf | A survey and an extensive evaluation of popular audio declipping methods | Dynamic range limitations in signal processing often lead to clipping, or saturation, in signals. The task of audio declipping is estimating the original audio signal, given its clipped measurements, and has attracted much interest in recent years. Audio declipping algorithms often make assumptions about the underlying signal, such as sparsity or low-rankness, and about the measurement system. In this paper, we provide an extensive review of audio declipping algorithms proposed in the literature. For each algorithm, we present assumptions that are made about the audio signal, the modeling domain, and the optimization algorithm. Furthermore, we provide an extensive numerical evaluation of popular declipping algorithms, on real audio data. We evaluate each algorithm in terms of the Signal-to-Distortion Ratio, and also using perceptual metrics of sound quality. The article is accompanied by a repository containing the evaluated methods. | ['Pavel Záviška', 'Pavel Rajmic', 'Lucas Rencker', 'Alexey Ozerov'] | 2020-07-15 | null | null | null | null | ['audio-declipping'] | ['audio'] | [ 5.58567882e-01 -1.26397341e-01 9.56956595e-02 -8.16164538e-02
-1.17285991e+00 -7.06300974e-01 -6.79137483e-02 -4.87802885e-02
-8.57897848e-03 6.21519506e-01 5.41939020e-01 1.17841907e-01
-3.83693904e-01 3.14846113e-02 -5.32909870e-01 -4.90619987e-01
-5.10504246e-01 -3.10592204e-01 -1.13103546e-01 1.46465197e-01
1.08057939e-01 2.19981283e-01 -1.52335620e+00 2.87772745e-01
5.77115774e-01 1.36262417e+00 2.31514201e-02 1.14704466e+00
5.12026310e-01 4.60071892e-01 -7.48110354e-01 -2.43376717e-01
3.61438125e-01 -4.85662818e-01 -1.57250047e-01 2.01734141e-01
3.88113379e-01 -3.52495283e-01 -4.79816943e-01 1.27251542e+00
7.40696907e-01 4.22443151e-02 3.63246739e-01 -1.38100326e+00
8.34545791e-02 8.78459930e-01 -2.68790424e-01 1.75351471e-01
7.04576015e-01 -8.26515779e-02 9.93544042e-01 -1.44654691e+00
2.34058619e-01 9.44067955e-01 1.08193064e+00 -1.47779509e-01
-1.37635028e+00 -9.33098674e-01 -2.51893967e-01 3.02333206e-01
-1.75389755e+00 -8.93008471e-01 1.07485533e+00 -3.77343535e-01
3.34692180e-01 6.98181212e-01 5.96100807e-01 6.41939104e-01
-4.89864731e-04 8.59836638e-01 7.51014650e-01 -5.23854256e-01
2.73021311e-01 -1.75968513e-01 3.25184241e-02 3.15569937e-02
-5.01609547e-03 1.15036584e-01 -1.02092373e+00 -6.67873919e-01
3.76667887e-01 -5.46504140e-01 -7.11282969e-01 -8.95564780e-02
-9.69703078e-01 2.65267044e-01 -2.56927758e-01 2.43885908e-02
-2.11064309e-01 2.13796303e-01 4.52030569e-01 6.10846639e-01
3.62395316e-01 4.43505466e-01 -6.01954758e-02 -3.59067589e-01
-1.50172770e+00 4.33439225e-01 8.28865170e-01 9.99425292e-01
3.41012448e-01 7.28581488e-01 3.03328130e-02 1.03289676e+00
1.84708044e-01 6.32267237e-01 4.67990071e-01 -1.14136779e+00
6.38402998e-01 -4.55382198e-01 2.54395008e-01 -1.25570250e+00
-4.02394831e-01 -6.67476475e-01 -8.60178232e-01 -1.30696326e-01
2.33155683e-01 -3.60392213e-01 -2.95599997e-01 1.55851614e+00
1.70370176e-01 6.36581242e-01 -9.43529699e-03 1.05313110e+00
5.03856659e-01 8.65024447e-01 -5.85998178e-01 -6.59170985e-01
9.49921370e-01 -5.00529885e-01 -1.23263693e+00 -2.61949245e-02
-1.31809592e-01 -1.22877741e+00 8.67410719e-01 1.01569736e+00
-1.43925357e+00 -5.30466497e-01 -1.33210695e+00 2.51504987e-01
3.73655021e-01 2.81748593e-01 -9.91056561e-02 7.36953676e-01
-9.31603432e-01 4.19574350e-01 -4.47285324e-01 1.39268696e-01
-2.82752603e-01 1.76904708e-01 -1.95106640e-02 1.00555509e-01
-1.30533314e+00 1.08629026e-01 -1.09772652e-01 2.46178597e-01
-9.77233827e-01 -9.65026140e-01 -6.15531981e-01 2.22383097e-01
4.03874844e-01 -1.36797160e-01 1.57649660e+00 -7.89684772e-01
-1.58896101e+00 1.98892072e-01 -1.04834311e-01 -6.47303045e-01
4.54326928e-01 -6.59804344e-01 -9.85021174e-01 3.38511854e-01
-1.92588791e-01 -1.16418235e-01 1.56645858e+00 -9.90231752e-01
-5.53220332e-01 -3.35786119e-02 -3.81565064e-01 2.13506296e-01
-1.97923303e-01 -4.47376855e-02 -5.16535699e-01 -1.25590527e+00
5.45024514e-01 -9.56666291e-01 -1.87888652e-01 -6.11395091e-02
-7.58436978e-01 7.07204938e-01 6.91057742e-01 -7.85535872e-01
1.74355149e+00 -2.90431952e+00 4.44667041e-02 5.55222273e-01
-3.33151743e-02 -8.67987573e-02 -8.28689858e-02 7.95315325e-01
-2.97249049e-01 -2.09475353e-01 -2.48455331e-01 -3.72592807e-01
4.47952189e-02 -2.21691236e-01 -9.48286295e-01 8.92377317e-01
-2.03090027e-01 -1.34457964e-02 -5.22201777e-01 -2.42743522e-01
6.63338825e-02 4.09653068e-01 -6.94376051e-01 2.36089513e-01
3.26142848e-01 2.19796151e-01 -6.70022368e-02 5.44390082e-01
8.91885400e-01 3.05324167e-01 1.59593150e-01 -3.88393879e-01
-3.22503537e-01 2.60858208e-01 -1.81859398e+00 1.48660159e+00
-4.58974421e-01 9.64458287e-01 8.47210586e-01 -6.71218634e-01
7.76506186e-01 8.09255898e-01 6.37789667e-01 -2.54369348e-01
-7.56661743e-02 4.23705131e-01 -1.75916255e-01 -3.96521300e-01
5.84514260e-01 -5.32788783e-02 8.98944156e-04 3.58841091e-01
-2.25388147e-02 -5.61163425e-01 2.21838355e-01 7.77182356e-02
8.44975710e-01 -4.83287752e-01 3.94192636e-01 -3.39158565e-01
5.81135750e-01 -5.21050274e-01 5.44208229e-01 6.20978832e-01
-3.58304381e-02 9.20527637e-01 3.92614961e-01 1.08755961e-01
-8.95456135e-01 -1.13059270e+00 -3.16963673e-01 9.86976743e-01
-1.31019324e-01 -7.08413959e-01 -6.89527810e-01 3.93563449e-01
4.75332886e-02 4.37301636e-01 -1.39451027e-01 -7.34860003e-02
-5.25868714e-01 -3.55957776e-01 1.04319847e+00 2.69868970e-01
1.16851829e-01 -2.35359147e-01 -3.55874687e-01 5.08795083e-01
-5.45226932e-01 -1.28268790e+00 -8.09997022e-01 -2.14670021e-02
-8.99540603e-01 -6.94259048e-01 -7.23228872e-01 -4.41310287e-01
2.44924426e-01 2.66154647e-01 7.65586734e-01 -3.14109117e-01
-2.35996190e-02 5.33680499e-01 -3.62220258e-01 -4.60700959e-01
-1.25552952e-01 -2.39127368e-01 6.27927721e-01 5.34251332e-01
-4.77879345e-01 -1.05690801e+00 -3.68537664e-01 5.22154033e-01
-8.61008286e-01 -3.41479272e-01 5.05819857e-01 6.45157218e-01
5.60292542e-01 2.67400265e-01 4.77935672e-01 -3.63371849e-01
8.74204278e-01 -2.94180363e-01 -7.51209080e-01 -2.93428272e-01
-4.05457437e-01 -2.28921324e-01 7.77006805e-01 -6.83459997e-01
-7.45206773e-01 2.17703179e-01 -2.80219585e-01 -5.16258895e-01
3.41818094e-01 7.31428564e-01 -3.63522887e-01 -6.08272664e-02
5.84772527e-01 3.44549239e-01 -1.43481642e-01 -7.44265854e-01
3.11810493e-01 6.86660528e-01 8.52731049e-01 -4.82626885e-01
8.83451343e-01 3.54945511e-01 -5.78655452e-02 -1.44450903e+00
-7.03550458e-01 -5.77104330e-01 -1.83272749e-01 -3.44612569e-01
2.80791465e-02 -9.24930453e-01 -3.88274014e-01 4.27225292e-01
-9.31243539e-01 1.17123686e-01 -3.81845206e-01 9.49995875e-01
-5.57595730e-01 2.57118672e-01 -5.11961520e-01 -1.07460296e+00
-1.93487510e-01 -1.00022304e+00 7.44650245e-01 -3.27654719e-01
-3.83825272e-01 -5.19924879e-01 2.90467948e-01 -1.76862583e-01
3.86328369e-01 -1.97811006e-03 5.15488386e-01 -2.50468969e-01
-2.45282650e-01 -2.71085382e-01 1.74564511e-01 3.96393895e-01
-1.42379671e-01 2.57019531e-02 -1.27242243e+00 -4.83296126e-01
4.03177351e-01 1.30022466e-01 4.38495606e-01 7.60590315e-01
1.13358843e+00 -7.72729158e-01 -1.02342404e-01 7.49614894e-01
1.26863670e+00 2.57283337e-02 5.44895172e-01 -1.51351675e-01
4.08199057e-02 4.44603086e-01 6.74526274e-01 1.05595350e+00
-3.79230291e-01 8.40207517e-01 3.57685506e-01 2.20719442e-01
-1.32503524e-01 -2.38003150e-01 5.56295156e-01 1.40118194e+00
2.52399206e-01 -1.65464208e-01 -5.53933620e-01 4.50517505e-01
-1.37127531e+00 -8.08043718e-01 -8.61446634e-02 2.33678484e+00
7.71544337e-01 -3.26986983e-02 2.78099656e-01 1.05582559e+00
8.48334670e-01 2.53785670e-01 -5.48859239e-01 -1.70447305e-01
-2.25282058e-01 5.78236096e-02 4.38267082e-01 8.37523162e-01
-8.78612697e-01 3.53908762e-02 7.77849722e+00 6.94824278e-01
-1.21987081e+00 9.28771421e-02 -6.06777109e-02 -3.97702992e-01
1.08170375e-01 -3.49588729e-02 -5.16630948e-01 4.39356983e-01
1.06082809e+00 -6.88324630e-01 4.58354443e-01 5.78362882e-01
5.86300910e-01 1.74180076e-01 -1.04317558e+00 1.30870664e+00
9.58556905e-02 -8.53248477e-01 -1.85266733e-01 -2.59463936e-01
3.02511990e-01 -3.08473498e-01 4.85821754e-01 -1.34421632e-01
-4.47726727e-01 -7.26966262e-01 1.18062925e+00 4.83516902e-01
9.81036425e-01 -7.15324998e-01 3.03942919e-01 9.23625231e-02
-1.53816915e+00 -3.70364606e-01 -1.10942267e-01 -3.08351785e-01
3.14076364e-01 1.18719733e+00 -6.87965810e-01 4.07582015e-01
6.93335235e-01 6.07339382e-01 -1.17772385e-01 1.45058692e+00
-2.85865694e-01 1.24577212e+00 -5.36262453e-01 1.80756450e-01
-1.29901841e-01 -1.16088241e-01 1.09275901e+00 1.54768825e+00
6.60812140e-01 1.64762944e-01 2.43652031e-01 3.27035815e-01
1.23941772e-01 1.81446671e-01 -2.53593922e-01 1.48844257e-01
7.82267988e-01 7.96199024e-01 6.99845552e-02 -1.01532444e-01
-2.36558065e-01 4.41352725e-01 -6.10604525e-01 6.32565737e-01
-6.65279150e-01 -8.20621014e-01 9.01127100e-01 4.22890902e-01
5.37951142e-02 -3.13527226e-01 -1.80023178e-01 -9.26921844e-01
2.81521916e-01 -1.20939183e+00 2.70331502e-01 -7.70200849e-01
-7.89269567e-01 3.55349422e-01 -1.31961945e-02 -1.90918458e+00
-3.75420421e-01 -1.28817990e-01 -4.14881408e-01 5.95694482e-01
-8.91819954e-01 -3.21108431e-01 2.64731776e-02 6.12068951e-01
6.18280649e-01 -7.26047670e-03 4.76121217e-01 7.24993587e-01
-2.91782767e-01 6.06236815e-01 5.15852511e-01 -5.55326082e-02
6.78587437e-01 -7.82596469e-01 1.27522126e-01 1.00001216e+00
2.86296129e-01 4.87398088e-01 1.62610006e+00 -1.58570215e-01
-1.49110293e+00 -9.59896743e-01 5.46727777e-01 1.85445204e-01
8.33856404e-01 -4.68466610e-01 -8.81862640e-01 4.46594536e-01
1.26465306e-01 -1.16620056e-01 8.43063116e-01 -2.66807172e-02
-1.67748109e-01 -5.01466990e-01 -8.75384450e-01 4.00463343e-01
6.11449540e-01 -7.62588441e-01 -4.52692032e-01 7.14588687e-02
5.22868514e-01 -6.69060767e-01 -6.98934913e-01 8.37069750e-02
8.44314814e-01 -8.00342739e-01 1.01046324e+00 -3.03621709e-01
-1.33343916e-02 -4.89289910e-01 -5.92455983e-01 -1.35698771e+00
-3.67870867e-01 -1.28221953e+00 -3.60623747e-01 1.27217352e+00
5.74896574e-01 -3.46056581e-01 5.51639199e-01 1.91766635e-01
-2.44766459e-01 -4.61487621e-01 -1.14369011e+00 -1.16154242e+00
-3.75213981e-01 -1.06122124e+00 3.64108562e-01 5.30718923e-01
4.88332748e-01 2.52951682e-01 -9.84999001e-01 7.17236340e-01
9.28043246e-01 -7.47308657e-02 7.27114439e-01 -8.81012321e-01
-5.55429459e-01 -1.85756505e-01 -4.95965421e-01 -1.35424006e+00
-1.37319252e-01 -4.52772051e-01 1.44670889e-01 -8.59926581e-01
-3.97164285e-01 -1.44719839e-01 -3.35629255e-01 -1.37743860e-01
2.71228671e-01 4.80127513e-01 2.82413661e-01 1.93318143e-01
-2.25322574e-01 5.08153379e-01 6.11289084e-01 -3.22498798e-01
-4.13499206e-01 5.14384866e-01 -4.47889149e-01 9.40639257e-01
7.00906992e-01 -6.41811311e-01 -5.74939430e-01 -4.14531648e-01
4.74869162e-01 4.21623617e-01 1.26687795e-01 -1.47571063e+00
4.10284877e-01 3.05394799e-01 -9.68228802e-02 -7.13977814e-01
9.43225682e-01 -9.38847899e-01 4.45947766e-01 4.00771081e-01
-5.76380849e-01 7.42300302e-02 2.39306584e-01 6.15178704e-01
-6.82277083e-01 -2.63844162e-01 9.43647802e-01 5.44835806e-01
-9.26951841e-02 -1.27968090e-02 -7.29493201e-01 2.80290842e-01
5.58342099e-01 -6.56330436e-02 4.47205484e-01 -1.09697568e+00
-1.00370908e+00 -9.12191942e-02 -5.46898805e-02 9.33899209e-02
8.63826931e-01 -1.41887653e+00 -1.02859783e+00 2.29715422e-01
-9.95259285e-02 -5.86617291e-01 3.55897665e-01 1.01199079e+00
-1.22732669e-01 2.67842054e-01 1.74046457e-01 -6.72301471e-01
-1.52890027e+00 3.71686518e-01 2.11694136e-01 2.01718301e-01
-5.97364843e-01 6.16897404e-01 1.09936841e-01 2.71148711e-01
5.44293284e-01 -6.09870136e-01 1.21292844e-02 4.56816442e-02
9.39124465e-01 6.59052789e-01 1.78489909e-01 -5.97866714e-01
-2.14638025e-01 7.41334140e-01 4.20624584e-01 -7.24121213e-01
8.19212735e-01 -4.76219565e-01 2.32833102e-02 7.35653758e-01
1.28345895e+00 6.12122357e-01 -9.84679997e-01 -3.53078276e-01
-3.04820120e-01 -7.93553889e-01 1.10729106e-01 -3.08979422e-01
-8.85604084e-01 7.76199758e-01 6.25554502e-01 4.67956275e-01
1.47569168e+00 -5.01313806e-01 6.59976125e-01 2.77564734e-01
3.28028768e-01 -1.19974959e+00 7.66923577e-02 4.47133422e-01
1.27028322e+00 -2.95001507e-01 3.03462088e-01 -5.41125238e-01
-1.81529433e-01 9.31055248e-01 -2.05788940e-01 -2.03733131e-01
1.01449096e+00 7.99244702e-01 1.01851292e-01 4.73987609e-01
-5.77382803e-01 3.93604577e-01 3.10682923e-01 6.27281606e-01
3.45941722e-01 -7.24569410e-02 -1.59105703e-01 8.71921778e-01
-7.30516672e-01 -8.73648226e-02 7.92923927e-01 6.66629076e-01
-6.37966096e-01 -7.56284773e-01 -1.09906781e+00 3.18350077e-01
-7.25506365e-01 -2.21370459e-01 -2.45653912e-01 2.81799316e-01
-3.92646581e-01 1.33211792e+00 -1.90114900e-01 -4.73676383e-01
7.36097574e-01 -2.74970710e-01 1.91103682e-01 -1.16306469e-01
-2.31560946e-01 4.33798283e-01 1.14618376e-01 -4.92935091e-01
-4.67011556e-02 -8.97837400e-01 -8.83800268e-01 -2.39977047e-01
-5.43039203e-01 5.63315809e-01 5.81629872e-01 5.07464468e-01
2.96132267e-01 4.74856615e-01 8.86568964e-01 -9.38448071e-01
-7.42557585e-01 -8.74912679e-01 -1.16767895e+00 5.49061298e-02
8.44805419e-01 -4.64228809e-01 -6.65563226e-01 2.61436850e-01] | [15.475996971130371, 5.55682373046875] |
42fd7872-3795-4874-b96e-9941991e2294 | densely-connected-multidilated-convolutional | 2011.11844 | null | https://arxiv.org/abs/2011.11844v2 | https://arxiv.org/pdf/2011.11844v2.pdf | Densely connected multidilated convolutional networks for dense prediction tasks | Tasks that involve high-resolution dense prediction require a modeling of both local and global patterns in a large input field. Although the local and global structures often depend on each other and their simultaneous modeling is important, many convolutional neural network (CNN)-based approaches interchange representations in different resolutions only a few times. In this paper, we claim the importance of a dense simultaneous modeling of multiresolution representation and propose a novel CNN architecture called densely connected multidilated DenseNet (D3Net). D3Net involves a novel multidilated convolution that has different dilation factors in a single layer to model different resolutions simultaneously. By combining the multidilated convolution with the DenseNet architecture, D3Net incorporates multiresolution learning with an exponentially growing receptive field in almost all layers, while avoiding the aliasing problem that occurs when we naively incorporate the dilated convolution in DenseNet. Experiments on the image semantic segmentation task using Cityscapes and the audio source separation task using MUSDB18 show that the proposed method has superior performance over state-of-the-art methods. | ['Yuki Mitsufuji', 'Naoya Takahashi'] | 2020-11-21 | null | null | null | null | ['audio-source-separation', 'music-source-separation'] | ['audio', 'music'] | [-7.33610913e-02 -1.45113856e-01 1.80857927e-01 -3.66188198e-01
-5.88145971e-01 -2.31712237e-01 3.10550302e-01 -1.19791023e-01
-4.78894979e-01 4.14790452e-01 3.09447676e-01 1.63427860e-01
-6.40866384e-02 -1.00367308e+00 -7.40545928e-01 -5.27380228e-01
2.30803844e-02 7.25426152e-02 6.39658093e-01 -2.77504772e-01
-9.25238878e-02 5.46061218e-01 -1.56984186e+00 7.62917399e-01
8.74126613e-01 1.28620493e+00 5.27501941e-01 6.53872013e-01
-2.54966497e-01 1.02975190e+00 -5.65989733e-01 -6.79454207e-02
4.03817356e-01 -2.49807626e-01 -9.13491964e-01 8.93746167e-02
5.58790922e-01 -2.09306762e-01 -1.69127420e-01 1.22870862e+00
5.83798170e-01 2.32415184e-01 3.86207193e-01 -6.95382774e-01
-6.99466228e-01 8.45757604e-01 -8.92470419e-01 5.89511693e-01
-2.55584091e-01 -2.78995216e-01 9.07078683e-01 -7.16408849e-01
1.88767716e-01 1.54835093e+00 8.84446204e-01 2.06756666e-01
-1.30237472e+00 -9.35908198e-01 4.42039609e-01 1.65740222e-01
-1.68348277e+00 -1.69478476e-01 6.62577629e-01 -3.25406522e-01
9.25187171e-01 6.20452920e-03 4.53164428e-01 7.91121006e-01
-3.10548414e-02 5.66290677e-01 1.09678566e+00 -1.36013612e-01
2.46310029e-02 4.25929651e-02 -4.11032373e-03 6.15626931e-01
-1.39371380e-02 -1.59456119e-01 -1.87510297e-01 2.06473559e-01
1.47264934e+00 1.39743894e-01 -2.92235196e-01 1.53120428e-01
-9.98964787e-01 8.88207018e-01 8.73122811e-01 7.73760855e-01
-4.02230203e-01 5.89124411e-02 3.22427213e-01 2.34586358e-01
6.37805402e-01 3.42233449e-01 -5.03327787e-01 4.04867977e-01
-1.19446075e+00 1.17800035e-01 3.68186891e-01 8.71278465e-01
8.60610187e-01 2.22253382e-01 -1.14237584e-01 1.27865899e+00
-2.65069120e-02 4.18144725e-02 6.69254303e-01 -8.61191154e-01
3.81467104e-01 6.69601083e-01 -1.87681764e-01 -1.11801636e+00
-5.48020303e-01 -8.03955913e-01 -1.49479580e+00 1.51019856e-01
2.73585200e-01 -1.17219523e-01 -1.12932670e+00 1.68623078e+00
1.38598636e-01 4.75485623e-01 1.15447953e-01 1.01640773e+00
1.01248538e+00 9.36346352e-01 1.84303030e-01 1.23200297e-01
1.30979002e+00 -1.27907372e+00 -6.13951802e-01 -1.53283060e-01
1.30154461e-01 -9.03465211e-01 8.16005766e-01 3.15075338e-01
-1.18978894e+00 -1.18769336e+00 -1.12753272e+00 -2.81966686e-01
-4.52296257e-01 2.38673076e-01 5.01590610e-01 2.04244763e-01
-1.10092592e+00 6.77208841e-01 -6.44152105e-01 -2.94957340e-01
4.62141573e-01 3.66787493e-01 -2.61175126e-01 1.77970249e-02
-1.29658091e+00 7.31476307e-01 5.41435182e-01 1.33486643e-01
-6.89945400e-01 -6.41009450e-01 -8.12626302e-01 2.71322966e-01
7.56177604e-02 -4.94856089e-01 1.21537375e+00 -1.30237031e+00
-1.26340926e+00 4.28649038e-01 2.95496034e-03 -6.53774202e-01
3.27709615e-01 -1.99396715e-01 -5.15863299e-01 4.04056817e-01
3.13227661e-02 1.02258241e+00 9.89201844e-01 -9.44276035e-01
-8.94486666e-01 -1.42025813e-01 2.43073255e-01 3.46358299e-01
-1.83132157e-01 -6.27011135e-02 -4.61489737e-01 -9.29726064e-01
3.86309355e-01 -3.87484223e-01 -5.21944046e-01 -3.59807819e-01
-1.22533077e-02 -9.65588540e-02 8.46018493e-01 -6.19145811e-01
1.19810653e+00 -2.32262850e+00 9.02450904e-02 1.62667483e-02
2.85081774e-01 2.82091677e-01 -3.98853809e-01 -2.52509266e-01
-1.55390516e-01 2.07951143e-01 -1.45424947e-01 -3.34206522e-01
-2.96588689e-01 1.00449204e-01 -2.29148433e-01 1.75871700e-01
3.84073019e-01 6.51523590e-01 -5.80775857e-01 -5.32449543e-01
2.60209680e-01 9.00622964e-01 -6.21554971e-01 1.25635102e-01
5.29145449e-02 5.51221013e-01 -3.06877524e-01 4.36988622e-01
9.97965693e-01 -3.72961015e-01 -1.66249368e-02 -4.38622206e-01
-3.01718742e-01 2.03738943e-01 -1.47150397e+00 1.86607814e+00
-6.11689925e-01 4.08458322e-01 3.52585614e-01 -1.09623241e+00
1.11790001e+00 5.17745137e-01 3.48076105e-01 -9.21775639e-01
1.22169495e-01 3.13161701e-01 8.73306207e-03 -1.70002893e-01
5.45490563e-01 -2.95785517e-01 5.45127457e-03 1.04786821e-01
4.40452904e-01 1.63753062e-01 -4.56155241e-02 -1.81227148e-01
8.23080361e-01 -2.13090837e-01 1.27371013e-01 -3.35557669e-01
5.90141356e-01 -3.17671031e-01 8.25425267e-01 7.65663385e-01
-5.13519496e-02 1.02094865e+00 2.63153642e-01 -7.17177689e-01
-1.07750058e+00 -6.52563810e-01 -2.58933753e-01 1.02410197e+00
2.24053159e-01 -1.09096758e-01 -9.21174586e-01 -4.36211169e-01
-2.43417859e-01 1.43499905e-02 -6.20805383e-01 2.20786482e-01
-6.94831192e-01 -8.39907348e-01 7.04213560e-01 7.63579905e-01
1.18108487e+00 -1.10074019e+00 -5.47997117e-01 4.46902901e-01
-3.33578378e-01 -1.45359373e+00 -3.02702039e-01 4.75174516e-01
-9.13755715e-01 -7.68039882e-01 -1.06255162e+00 -1.13585246e+00
3.78417611e-01 2.77506620e-01 1.02335775e+00 -2.62973070e-01
-1.83901697e-01 -1.93457872e-01 -3.78259927e-01 -1.08812116e-01
-1.21303901e-01 4.35542673e-01 -3.03555518e-01 2.49371231e-01
2.87615001e-01 -8.59090924e-01 -5.24825990e-01 3.29726666e-01
-9.13088799e-01 2.85730898e-01 6.51433766e-01 6.30256355e-01
9.70369697e-01 5.62016606e-01 6.75941408e-01 -8.09194565e-01
4.15008008e-01 -3.71338397e-01 -5.85573435e-01 2.00447217e-02
-9.62235406e-02 -5.58640845e-02 7.23295391e-01 -6.24242604e-01
-1.18912721e+00 -3.21087278e-02 -4.11818832e-01 -5.96195817e-01
-5.31983197e-01 2.69871175e-01 -1.38784975e-01 6.86124936e-02
5.00674069e-01 -7.53313676e-02 -3.86903971e-01 -9.18518662e-01
2.64418483e-01 5.90670705e-01 5.29265702e-01 -3.66171062e-01
5.21671951e-01 3.99357736e-01 -3.19113582e-01 -8.77509117e-01
-1.03027320e+00 -4.14361954e-01 -8.46598268e-01 1.15882352e-01
1.25491512e+00 -1.45719266e+00 -3.57736409e-01 6.30644381e-01
-1.32329214e+00 -2.62456685e-01 -4.59457338e-01 5.33167541e-01
-2.60712832e-01 -9.04152095e-02 -7.80763447e-01 -4.10344601e-01
-2.03157082e-01 -1.18256330e+00 9.48077381e-01 5.58900952e-01
-2.98572369e-02 -8.93998563e-01 -2.14992598e-01 1.80197820e-01
7.12188661e-01 1.36172280e-01 7.51981497e-01 -5.97051263e-01
-6.58354700e-01 1.06446959e-01 -6.60609782e-01 6.93648398e-01
2.14401171e-01 -4.04013216e-01 -1.22763753e+00 8.78491811e-03
1.85348019e-01 -2.56067693e-01 1.36477566e+00 7.35764682e-01
1.46425974e+00 -1.89704031e-01 1.88688770e-01 9.04330134e-01
1.61434352e+00 1.05888724e-01 6.72166169e-01 2.92707384e-01
7.99533248e-01 4.35258746e-01 3.28380793e-01 2.52373457e-01
3.31491649e-01 3.69235098e-01 4.71663445e-01 -7.54438221e-01
-3.53193998e-01 -6.27697706e-02 -4.13146093e-02 9.30080533e-01
-2.43750885e-01 6.95875660e-02 -5.40245295e-01 6.37323856e-01
-1.85610580e+00 -9.80442405e-01 1.08632237e-01 1.78974211e+00
9.50867414e-01 1.00223385e-01 5.11788353e-02 2.42749766e-01
9.91324186e-01 4.31968302e-01 -2.60502070e-01 -5.32565713e-01
-3.65903586e-01 6.45896971e-01 4.67077464e-01 5.79617858e-01
-1.30780077e+00 1.02901125e+00 6.13705301e+00 1.11054671e+00
-1.34760857e+00 3.42861086e-01 8.20791543e-01 4.18296754e-02
-7.66124353e-02 -4.47871923e-01 -9.77077663e-01 1.09981403e-01
8.64675999e-01 3.91208589e-01 4.02947724e-01 6.85738921e-01
1.00007206e-01 -7.10628703e-02 -6.50514185e-01 1.04085970e+00
-2.13283807e-01 -1.35715330e+00 1.40514120e-01 -1.54425263e-01
8.68175566e-01 9.61360782e-02 1.80514187e-01 2.46997714e-01
1.44262210e-01 -1.28199232e+00 7.54947543e-01 3.38301063e-01
6.06638074e-01 -1.07945907e+00 8.74232650e-01 2.68957376e-01
-1.67345250e+00 -2.50686288e-01 -7.74725199e-01 8.67793523e-03
-1.42204061e-01 5.15713990e-01 -3.02387565e-01 4.95729446e-01
1.17660129e+00 8.15480053e-01 -4.59020138e-01 9.84629333e-01
-1.07055962e-01 4.60952908e-01 -2.14346841e-01 5.58859944e-01
5.06343186e-01 7.14520812e-02 3.40145826e-01 1.40049708e+00
3.27402115e-01 1.40260383e-01 1.29418373e-01 9.83367562e-01
-2.32685298e-01 1.60243793e-03 -2.33933955e-01 1.50051922e-01
2.15631083e-01 1.34653854e+00 -8.58915806e-01 -3.08227539e-01
-5.75901270e-01 8.12526405e-01 2.55355179e-01 4.97349739e-01
-8.35433543e-01 -4.23348218e-01 7.36099482e-01 1.68587938e-01
8.52007985e-01 -1.57665238e-01 -3.81962270e-01 -1.00105619e+00
-2.64637887e-01 -8.69369864e-01 3.33291292e-01 -6.41104043e-01
-1.19405091e+00 1.16957104e+00 -9.00786668e-02 -1.32430065e+00
1.00171193e-01 -3.74629945e-01 -3.61151755e-01 1.23626471e+00
-2.02673745e+00 -1.07462585e+00 -3.66414636e-01 9.93798852e-01
8.89582992e-01 -6.29203171e-02 8.65676463e-01 6.52357101e-01
-4.64643866e-01 4.14632708e-01 -2.63250023e-01 2.43000939e-01
5.75607359e-01 -1.07007849e+00 2.13102371e-01 7.09206402e-01
3.10940593e-02 4.12264228e-01 1.82326600e-01 -3.42916459e-01
-3.18412989e-01 -1.39853632e+00 7.48383701e-01 3.86082172e-01
4.26237464e-01 -2.20296144e-01 -1.26666093e+00 5.63284099e-01
3.49888116e-01 3.34709615e-01 5.62367380e-01 5.91428839e-02
-5.64837396e-01 -3.35670590e-01 -1.13925540e+00 2.02083677e-01
7.55911708e-01 -5.67265093e-01 -4.83984649e-01 -2.01526016e-01
1.02256167e+00 -3.52723420e-01 -8.95417511e-01 5.96363127e-01
3.22647035e-01 -1.11497188e+00 9.96873319e-01 -1.38003215e-01
4.54446137e-01 -4.54846382e-01 -3.98980469e-01 -1.04763710e+00
-6.93784475e-01 -2.55625248e-01 3.61639820e-02 1.22926390e+00
3.34824026e-01 -4.74109709e-01 3.48908961e-01 3.83967385e-02
-1.35180905e-01 -5.87966144e-01 -8.13364685e-01 -5.81065595e-01
1.97148204e-01 -3.76720488e-01 7.80868888e-01 9.95748103e-01
-5.38335621e-01 4.66226518e-01 -3.90543252e-01 5.61859608e-01
3.91144753e-01 1.67602807e-01 2.65350014e-01 -1.44774914e+00
-2.22754657e-01 -4.02238280e-01 -3.30650359e-01 -1.28695798e+00
-3.95712517e-02 -6.27647221e-01 -8.99557099e-02 -1.57480884e+00
1.61728859e-01 -4.28898096e-01 -7.20368087e-01 4.95181412e-01
5.71700335e-02 4.88396466e-01 6.31124750e-02 2.24014536e-01
-5.74029088e-01 4.82268095e-01 1.50018251e+00 -2.04881236e-01
-5.19812942e-01 -1.18680611e-01 -8.25638056e-01 1.12574530e+00
8.16086173e-01 -3.05616945e-01 -4.08330768e-01 -7.53395259e-01
-1.58354148e-01 -1.96494207e-01 3.61056149e-01 -1.37120330e+00
2.48977229e-01 8.33004266e-02 7.23075986e-01 -7.34288692e-01
3.46917838e-01 -9.52661395e-01 5.99770620e-02 -1.97672192e-02
-5.54006517e-01 -1.18923001e-01 5.77261508e-01 3.77969205e-01
-8.32524002e-01 -1.09121818e-02 1.27625418e+00 -4.75374877e-01
-9.45595264e-01 4.15032744e-01 -2.61122078e-01 -1.64893359e-01
5.57343185e-01 -2.31528521e-01 -7.33447298e-02 -1.38633341e-01
-1.11124110e+00 -3.33669409e-02 7.26918206e-02 5.86677134e-01
5.87931395e-01 -1.30719590e+00 -5.82146287e-01 3.03660959e-01
-4.69802618e-01 4.85306323e-01 5.17450213e-01 7.38264441e-01
-4.66759771e-01 2.49081120e-01 -3.93353969e-01 -5.38502991e-01
-1.02705038e+00 2.18833894e-01 7.24274933e-01 -5.23593605e-01
-7.13165283e-01 1.05614471e+00 6.94987476e-01 -3.54103476e-01
4.05281246e-01 -5.34332216e-01 -6.74134433e-01 2.18123794e-01
7.39115655e-01 2.43687436e-01 1.16330180e-02 -7.88275421e-01
-4.57491316e-02 7.42995441e-01 -1.98022023e-01 -1.60577744e-02
1.51689780e+00 -2.75101960e-01 -2.07789287e-01 4.49771404e-01
1.24642467e+00 -2.19474971e-01 -1.44542873e+00 -5.50261378e-01
-3.80864620e-01 -2.71223426e-01 3.80006999e-01 -6.85901165e-01
-1.53748441e+00 1.21391308e+00 7.05269337e-01 2.98096567e-01
1.56057048e+00 -1.41672879e-01 9.19400573e-01 -5.06930947e-02
1.46382511e-01 -1.18129325e+00 1.33465245e-01 8.36368859e-01
8.98986161e-01 -1.05668128e+00 -1.77368939e-01 -4.12260026e-01
-5.23988426e-01 1.16362321e+00 7.72465050e-01 -3.70392442e-01
8.78127158e-01 3.40480149e-01 3.32555138e-02 1.04361929e-01
-5.04034400e-01 -5.15323102e-01 3.22523177e-01 4.74236965e-01
4.61003363e-01 -8.52064267e-02 1.68175235e-01 9.32879686e-01
-1.54932529e-01 -1.58540502e-01 1.58495814e-01 7.04494059e-01
-6.43963337e-01 -8.78063262e-01 -3.91297311e-01 -2.29545012e-02
-7.22116292e-01 -2.36158580e-01 7.83281401e-02 7.41306305e-01
7.07143188e-01 8.34128678e-01 5.74417233e-01 -4.40844268e-01
2.80006856e-01 -1.15230329e-01 2.80019581e-01 -5.95676720e-01
-8.63726616e-01 7.09725022e-01 -3.44891757e-01 -5.34171462e-01
-7.39688337e-01 -1.05220094e-01 -1.37554395e+00 -2.10760996e-01
-1.40824273e-01 2.27032863e-02 4.50633556e-01 8.45479548e-01
2.78747439e-01 1.07655346e+00 4.33064520e-01 -9.34029222e-01
-7.37108514e-02 -1.11784542e+00 -9.81439352e-01 -1.70394429e-03
5.92237711e-01 -4.09463286e-01 -2.43293077e-01 1.21493936e-01] | [15.47749137878418, 5.375308513641357] |
7f3620dc-d6db-4eb7-b19a-13c3e4808c0d | modeling-exemplification-in-long-form | null | null | https://openreview.net/forum?id=QXSIhKXBaFP | https://openreview.net/pdf?id=QXSIhKXBaFP | Modeling Exemplification in Long-form Question Answering via Retrieval | Exemplification is a process by which writers explain or clarify a concept by providing an example. While common in all forms of writing, exemplification is particularly useful in the task of long-form question answering (LFQA), where a complicated answer can be made more understandable through simple examples. In this paper, we provide the first computational study of exemplification in QA, performing a fine-grained annotation of different types of examples (e.g., hypotheticals, anecdotes) in three corpora. We show that not only do state-of-the-art LFQA models struggle to generate relevant examples, but also that standard evaluation metrics such as ROUGE are insufficient to judge exemplification quality. We propose to treat exemplification as a \emph{retrieval} problem in which a partially-written answer is used to query a large set of human-written examples extracted from a corpus. Our approach allows a reliable ranking-type automatic metrics that correlates well with human evaluation. Human evaluation shows that examples retrieved from our retriever are more relevant than examples generated from state-of-the-art LFQA model. | ['Anonymous'] | 2022-01-16 | null | null | null | acl-arr-january-2022-1 | ['long-form-question-answering'] | ['natural-language-processing'] | [ 9.29335803e-02 6.23549402e-01 1.23096913e-01 -5.02512753e-01
-1.35935128e+00 -9.40417290e-01 1.21730590e+00 4.54713374e-01
-2.69265026e-01 1.18200135e+00 3.64928961e-01 -4.71839666e-01
-3.41863573e-01 -6.26734197e-01 -7.18566716e-01 6.35262281e-02
5.25418699e-01 1.30236232e+00 2.46009350e-01 -8.66774142e-01
6.30683720e-01 2.47926965e-01 -1.60769928e+00 7.35684872e-01
1.34742987e+00 5.62414825e-01 1.57439053e-01 4.30118203e-01
-7.74032950e-01 8.35650265e-01 -1.49157023e+00 -1.10290432e+00
-2.33966544e-01 -7.38275409e-01 -1.42051780e+00 -3.14445645e-02
8.76162708e-01 -3.07747163e-03 2.96870440e-01 7.72667348e-01
2.73460448e-01 -5.78107722e-02 9.09778953e-01 -1.00839984e+00
-9.97866035e-01 6.22107744e-01 1.34482786e-01 1.19220801e-01
1.05903661e+00 1.53668597e-01 1.28168821e+00 -1.02789533e+00
9.27583277e-01 1.44851470e+00 4.24733132e-01 8.50684285e-01
-1.06457865e+00 -1.56559244e-01 -3.58276099e-01 3.23108226e-01
-9.95069087e-01 -2.24854976e-01 5.27278125e-01 -2.31219500e-01
9.57745790e-01 7.53010988e-01 5.06379306e-01 7.89161921e-01
-2.42428064e-01 8.33388388e-01 1.43799496e+00 -9.76905406e-01
1.44890696e-01 5.08610725e-01 4.89800721e-01 4.73821610e-01
2.27390707e-01 -3.09285671e-01 -4.41558450e-01 -3.91944617e-01
4.64717537e-01 -4.08259600e-01 -6.21289790e-01 5.44251651e-02
-1.08026254e+00 9.93670404e-01 3.57369870e-01 4.75536168e-01
-4.94201332e-01 2.13944465e-01 1.90520763e-01 3.42912167e-01
1.27515838e-01 1.53333688e+00 -5.04670143e-01 -1.33896723e-01
-9.81769562e-01 8.42973888e-01 1.10626173e+00 8.82047176e-01
6.66325331e-01 -4.75933969e-01 -4.05806571e-01 8.50272655e-01
2.01310769e-01 5.58866084e-01 5.94939709e-01 -1.40997922e+00
2.16624349e-01 7.53217399e-01 6.47466600e-01 -5.60830474e-01
7.20728189e-02 -5.57756603e-01 -6.30695000e-03 -1.51168793e-01
6.93407238e-01 3.13196748e-01 -4.67610002e-01 1.44744205e+00
8.26605111e-02 -4.34600204e-01 4.21251923e-01 9.16550457e-01
1.34676623e+00 7.49377012e-01 -2.76638810e-02 -5.29150903e-01
1.53421116e+00 -1.05785954e+00 -1.03879523e+00 -2.35286728e-01
6.89751029e-01 -8.56981635e-01 1.91434085e+00 4.73973513e-01
-1.19230735e+00 -5.43476760e-01 -8.69541466e-01 -2.50963777e-01
-4.77762520e-01 8.18072855e-02 4.79218543e-01 6.76516891e-01
-7.31425047e-01 3.23115647e-01 1.62476584e-01 -3.31069052e-01
1.67624429e-01 -7.17772469e-02 -2.02137738e-01 -2.89835453e-01
-1.38081872e+00 1.36471784e+00 2.49898195e-01 -2.04068840e-01
-7.44837046e-01 -6.65365756e-01 -7.07471967e-01 5.24495877e-02
7.01762080e-01 -9.59969044e-01 1.70593214e+00 -8.98762643e-01
-9.95302141e-01 1.03432250e+00 -4.13584173e-01 -6.05061173e-01
1.86672062e-01 -5.91419160e-01 -6.33619845e-01 3.85049611e-01
5.24874628e-01 6.28850698e-01 8.57619226e-01 -1.81255865e+00
-2.19203606e-01 -5.68289198e-02 6.50808454e-01 1.95581034e-01
6.62147924e-02 2.27250278e-01 4.41158749e-02 -4.58245158e-01
-1.65107965e-01 -4.48279262e-01 -4.18731607e-02 -2.64963984e-01
-3.02612007e-01 -7.85391748e-01 5.65433025e-01 -5.72306395e-01
1.52629066e+00 -1.55800629e+00 -1.88392207e-01 -8.20924714e-02
1.68509349e-01 4.31168407e-01 -6.25343947e-03 8.32884967e-01
2.52355248e-01 5.05656123e-01 -2.70675480e-01 1.36759028e-01
4.37871486e-01 4.79039818e-01 -7.90470839e-01 -3.98323864e-01
4.44103122e-01 1.28689539e+00 -1.29645121e+00 -7.48978198e-01
-2.49107867e-01 -7.21414983e-02 -1.98062316e-01 2.94656157e-01
-7.91753232e-01 9.50182602e-02 -5.15521169e-01 5.16896904e-01
8.85216170e-04 -6.28492475e-01 -3.35745335e-01 -9.56059396e-02
3.78102988e-01 7.84899294e-01 -7.16949463e-01 1.61937547e+00
-5.49609601e-01 5.92158973e-01 -7.03097641e-01 -4.65815723e-01
9.31935132e-01 4.58346456e-01 -6.72557175e-01 -4.55863982e-01
-2.05526307e-01 8.88070583e-01 1.02081351e-01 -8.45062256e-01
8.62133563e-01 -4.95319158e-01 -3.15844446e-01 7.88383126e-01
5.36658429e-02 -1.04830384e+00 7.70458996e-01 5.59703648e-01
9.79338944e-01 2.92913139e-01 6.50464594e-01 -3.09089303e-01
7.94141471e-01 7.22966611e-01 -1.66058809e-01 1.14083481e+00
2.21773103e-01 5.03459275e-01 3.33766818e-01 -4.50466037e-01
-1.04872346e+00 -9.69229758e-01 -1.37293056e-01 7.70520747e-01
-5.26895896e-02 -7.66353250e-01 -8.44030738e-01 -9.55274701e-01
-2.71994770e-01 1.47950256e+00 -5.47384202e-01 1.40033588e-01
-7.37939417e-01 -2.92960674e-01 5.95981777e-01 4.35925163e-02
4.19713616e-01 -1.45233929e+00 -1.00304902e+00 3.75374407e-01
-4.68084753e-01 -5.51130474e-01 1.52286347e-02 -9.98348743e-02
-8.02031755e-01 -9.63742852e-01 -7.16122270e-01 -4.52539325e-01
5.00735700e-01 -8.68132487e-02 2.07535791e+00 5.87927818e-01
-6.04880452e-02 6.47156894e-01 -7.59406745e-01 -7.04135418e-01
-8.38940442e-01 -2.98542798e-01 -3.34728718e-01 -7.10046113e-01
6.40787959e-01 -1.70704305e-01 -3.03254098e-01 1.98148265e-01
-1.14879763e+00 -1.91688791e-01 5.38423359e-01 1.13620245e+00
5.67904055e-01 -4.43435401e-01 1.06964469e+00 -1.28919899e+00
1.52258301e+00 -3.22966903e-01 6.30866662e-02 9.24191594e-01
-5.12696862e-01 1.99210048e-01 6.57273352e-01 -4.08861130e-01
-1.08468986e+00 -5.26393116e-01 -4.90436703e-02 1.62256494e-01
-2.34385654e-01 8.27473581e-01 8.89851078e-02 2.71674156e-01
1.49552274e+00 1.88483387e-01 -1.06717072e-01 -1.59489259e-01
8.53438735e-01 5.99563003e-01 5.53265214e-01 -1.00511909e+00
7.72041619e-01 -2.98349205e-02 -1.05164118e-01 -6.78621531e-01
-1.42622983e+00 -3.28777790e-01 -3.97579968e-01 -3.33083361e-01
3.69853020e-01 -3.81389439e-01 -2.84594297e-01 -5.66213965e-01
-1.50234258e+00 -1.36332512e-01 -8.36364210e-01 -2.85377931e-02
-7.50130713e-01 3.11760157e-01 -2.79488385e-01 -9.75923836e-01
-4.04386640e-01 -7.52234459e-01 1.24378073e+00 3.70588690e-01
-9.01831388e-01 -7.29186833e-01 -2.92861797e-02 8.04693162e-01
3.26764047e-01 7.83204138e-02 1.27424145e+00 -1.12795067e+00
-4.74240422e-01 -2.49875233e-01 1.59180686e-01 1.12728231e-01
-1.56174898e-01 -1.53634086e-01 -8.73650610e-01 2.66827166e-01
2.18893528e-01 -8.97605956e-01 4.74314302e-01 -8.74676257e-02
8.40163529e-01 -7.34425306e-01 4.55894694e-02 -4.33981448e-01
1.23220336e+00 8.99072438e-02 8.32203865e-01 4.21957463e-01
2.08575148e-02 1.07999015e+00 1.29948676e+00 -2.04798877e-02
4.93986234e-02 5.77367127e-01 1.88871980e-01 3.61678600e-01
-2.68583655e-01 -1.88181117e-01 -1.05318397e-01 5.20299196e-01
-3.36729251e-02 -3.58787298e-01 -1.05250430e+00 6.48842454e-01
-1.67216504e+00 -1.19156933e+00 -4.77206171e-01 1.93658948e+00
1.19313192e+00 1.86280414e-01 -9.39890668e-02 2.37684920e-01
3.73024672e-01 -7.47704208e-02 -6.74313828e-02 -6.16310179e-01
-4.74209130e-01 6.68892980e-01 -6.66491628e-01 6.65263593e-01
-5.08801103e-01 1.04441583e+00 6.20583963e+00 1.00082994e+00
-2.58204162e-01 6.88783675e-02 4.98610914e-01 2.93508857e-01
-8.45874786e-01 1.98295027e-01 -5.22041678e-01 2.05162726e-02
8.12879443e-01 -4.30998057e-01 2.10814863e-01 7.49165356e-01
-1.62638187e-01 -1.37647718e-01 -1.37598968e+00 5.67407906e-01
5.19410551e-01 -1.55587184e+00 7.48269618e-01 -4.40360725e-01
6.09115481e-01 -8.57296824e-01 -1.46463355e-02 5.67161858e-01
1.05640441e-02 -1.43470407e+00 6.76848292e-01 6.87278271e-01
6.08059525e-01 -4.21705067e-01 9.30868447e-01 5.66075563e-01
-3.78551632e-01 6.75480627e-03 -4.23339248e-01 -1.51285931e-01
4.70883757e-01 2.64407754e-01 -1.08183992e+00 5.51861823e-01
4.11544055e-01 2.49851728e-03 -8.78977835e-01 1.02552330e+00
-9.05064404e-01 5.94070673e-01 -2.39225794e-02 -9.01879609e-01
3.68744999e-01 7.81720579e-02 5.35870790e-01 1.13896477e+00
3.16121370e-01 5.78491867e-01 -4.12453949e-01 1.25592697e+00
-2.48054676e-02 4.41707373e-01 -4.94175285e-01 -8.82054642e-02
6.77718103e-01 1.10615480e+00 -2.43318990e-01 -7.05272913e-01
-1.11009754e-01 8.96991849e-01 3.97317797e-01 2.51525491e-01
-3.01674157e-01 -4.26459074e-01 -2.97834784e-01 1.02676012e-01
-4.04306017e-02 1.97631136e-01 -3.81071754e-02 -8.82366836e-01
2.70346642e-01 -1.35496151e+00 3.03770483e-01 -1.64186037e+00
-1.71459985e+00 7.90731311e-01 2.62483686e-01 -1.14428043e+00
-8.67947757e-01 -6.38927877e-01 -6.63052917e-01 1.05143225e+00
-1.27671957e+00 -1.02370715e+00 -2.45989785e-01 2.59788007e-01
7.96124697e-01 -1.05310857e-01 1.19866145e+00 -1.35334700e-01
4.39279824e-01 2.12622225e-01 -4.85204160e-01 -2.97769845e-01
8.51313591e-01 -1.68663514e+00 1.20201476e-01 6.09259307e-01
8.40566456e-01 1.28248692e+00 1.37105489e+00 -5.71034372e-01
-9.04888570e-01 -4.54512745e-01 1.52581036e+00 -1.08911610e+00
6.64892733e-01 1.75198555e-01 -1.27350545e+00 3.29652607e-01
4.30047393e-01 -3.58179152e-01 9.11039650e-01 1.94772948e-02
-5.29315948e-01 2.00112328e-01 -1.11719775e+00 7.91029871e-01
7.02312648e-01 -6.27577484e-01 -1.79815722e+00 9.04672325e-01
8.40142012e-01 -2.26085678e-01 -7.25161910e-01 2.14863732e-01
1.90927476e-01 -8.57129216e-01 8.62024009e-01 -1.07443619e+00
1.00376773e+00 -6.21019602e-01 -1.54441357e-01 -1.42640364e+00
2.55545884e-01 -6.85604930e-01 -3.39127481e-01 1.09326661e+00
9.29751575e-01 -2.73303509e-01 4.31922793e-01 6.83140159e-01
-4.45792884e-01 -8.78069222e-01 -8.82559776e-01 -8.22143137e-01
1.63673639e-01 -1.94942221e-01 8.47402513e-01 8.34472895e-01
4.45725024e-01 8.54364097e-01 1.02484584e-01 -2.99001783e-01
2.96889156e-01 3.86171907e-01 9.24261212e-01 -1.26148605e+00
-3.96553069e-01 -3.39383602e-01 5.51150702e-02 -1.26163292e+00
7.83939287e-02 -6.32452309e-01 1.44595103e-02 -1.94664037e+00
1.84711680e-01 -3.45975906e-01 3.00178438e-01 1.74780324e-01
-5.32546997e-01 2.15660557e-01 1.28741249e-01 4.17166978e-01
-8.56834471e-01 4.58727896e-01 1.61565804e+00 -9.14572701e-02
1.22547098e-01 -2.19398946e-01 -8.01903844e-01 6.14800751e-01
5.72291851e-01 -3.26915801e-01 -4.50715989e-01 -3.81728530e-01
6.89521730e-01 3.66019845e-01 5.33044100e-01 -6.57006919e-01
5.95420487e-02 2.42656022e-02 1.14004403e-01 -3.90607685e-01
5.28235555e-01 -6.85077906e-01 -3.48727554e-02 4.68987077e-01
-6.72476530e-01 2.99413770e-01 -9.45560858e-02 3.49267155e-01
-6.03565872e-01 -1.22182739e+00 1.33234978e-01 -5.48124731e-01
-6.23386860e-01 -4.00719821e-01 -1.62215233e-01 5.43144286e-01
4.97538179e-01 -1.40016392e-01 -1.08607912e+00 -8.39083850e-01
-5.37209630e-01 1.01985097e-01 3.91117334e-01 7.76773468e-02
7.42741287e-01 -1.34401345e+00 -1.00347936e+00 -6.93448007e-01
5.78109264e-01 -1.70589328e-01 -1.38108820e-01 3.58109444e-01
-7.76902914e-01 6.36780441e-01 2.01163329e-02 -3.51985991e-01
-9.94903326e-01 5.71125269e-01 3.06114227e-01 -3.39747399e-01
-8.74348581e-02 9.36773717e-01 -5.72517514e-02 -4.57023978e-01
-8.19224119e-02 -2.33765654e-02 -6.37214601e-01 -1.00085817e-01
6.43218875e-01 6.61912784e-02 -1.36223108e-01 -6.00163221e-01
-1.90178137e-02 2.77650744e-01 -1.23248018e-01 -5.76139748e-01
9.11069155e-01 8.01655948e-02 -4.00420010e-01 7.40444124e-01
4.52608943e-01 1.70532614e-01 -4.22973603e-01 -2.89824039e-01
3.55753660e-01 -5.37244916e-01 -7.11087823e-01 -1.54628503e+00
8.05669799e-02 9.92494524e-01 -4.21337895e-02 6.91492498e-01
9.31047857e-01 5.30845463e-01 6.68124795e-01 1.05005801e+00
5.28885365e-01 -9.06798244e-01 6.15482926e-01 4.81543899e-01
1.66957140e+00 -1.18234730e+00 3.30573916e-02 -5.48313379e-01
-8.71383607e-01 1.23736274e+00 6.69814527e-01 5.13101853e-02
-1.46463916e-01 -4.47769821e-01 2.76092499e-01 -7.97789514e-01
-7.10158527e-01 -2.07989868e-02 6.55615509e-01 4.99124259e-01
7.25530386e-01 -1.03993028e-01 -9.88361001e-01 9.51457441e-01
-7.81605124e-01 -3.52439106e-01 7.86191642e-01 8.26070547e-01
-7.13370264e-01 -1.12042773e+00 -3.50228190e-01 7.62341380e-01
-5.15922487e-01 -3.70379597e-01 -1.03133547e+00 1.10494864e+00
-9.06173140e-02 1.17297101e+00 -3.91744345e-01 2.08725780e-01
4.37679440e-01 3.77065361e-01 7.42435515e-01 -1.09447539e+00
-1.09686863e+00 -3.76075357e-01 8.07805896e-01 -1.33699164e-01
-7.52252758e-01 -4.63010222e-01 -1.21740508e+00 3.04886967e-01
-6.31200254e-01 9.31837797e-01 5.36155403e-01 1.24646211e+00
-1.10096037e-01 1.58243060e-01 -5.11831492e-02 -1.50570601e-01
-8.20864677e-01 -1.27672565e+00 -3.27840626e-01 8.29043984e-01
7.30367377e-02 -5.05876720e-01 -4.47278619e-01 4.68776515e-03] | [11.37888240814209, 8.087306022644043] |
4d98f2a6-d50b-4097-b7b3-f647027a6b83 | categorizing-items-with-short-and-noisy | 2110.11431 | null | https://arxiv.org/abs/2110.11431v1 | https://arxiv.org/pdf/2110.11431v1.pdf | Categorizing Items with Short and Noisy Descriptions using Ensembled Transferred Embeddings | Item categorization is a machine learning task which aims at classifying e-commerce items, typically represented by textual attributes, to their most suitable category from a predefined set of categories. An accurate item categorization system is essential for improving both the user experience and the operational processes of the company. In this work, we focus on item categorization settings in which the textual attributes representing items are noisy and short, and labels (i.e., accurate classification of items into categories) are not available. In order to cope with such settings, we propose a novel learning framework, Ensembled Transferred Embeddings (ETE), which relies on two key ideas: 1) labeling a relatively small sample of the target dataset, in a semi-automatic process, and 2) leveraging other datasets from related domains or related tasks that are large-scale and labeled, to extract "transferable embeddings". Evaluation of ETE on a large-scale real-world dataset provided to us by PayPal, shows that it significantly outperforms traditional as well as state-of-the-art item categorization methods. | ['Erez Shmueli', 'Yonatan Hadar'] | 2021-10-21 | null | null | null | null | ['embeddings-evaluation'] | ['natural-language-processing'] | [-7.13539496e-02 -4.54918057e-01 -4.34444070e-01 -6.35817230e-01
-6.55406535e-01 -9.08159733e-01 2.70641923e-01 5.90676725e-01
-4.96320009e-01 4.47066277e-01 1.72085181e-01 -2.71394402e-01
-1.84850350e-01 -8.60108316e-01 -4.01322633e-01 -3.46340120e-01
1.59616411e-01 7.29992867e-01 -1.58317342e-01 -1.01195760e-01
4.83145803e-01 2.12497771e-01 -1.63517070e+00 4.03934002e-01
7.06717193e-01 1.46487582e+00 -4.11254130e-02 2.40086675e-01
-6.10611856e-01 3.96538645e-01 -3.80407929e-01 -5.55162430e-01
2.61972696e-01 -2.41457880e-01 -9.16844785e-01 2.23617941e-01
1.96931273e-01 -2.49488816e-01 3.65425169e-01 1.13170433e+00
3.08853865e-01 1.65157109e-01 1.03192151e+00 -1.22523165e+00
-1.45669687e+00 6.15177631e-01 -2.36397728e-01 8.39454904e-02
1.45488486e-01 -3.27354670e-01 1.31865048e+00 -1.19289112e+00
4.63942975e-01 8.75706255e-01 8.64861071e-01 5.33515334e-01
-1.25530553e+00 -6.74242973e-01 2.79173285e-01 1.78449228e-01
-1.30522227e+00 -7.51037225e-02 7.27255821e-01 -7.25443184e-01
6.32302284e-01 -3.53911780e-02 2.57268459e-01 1.20086014e+00
1.97957441e-01 4.36381668e-01 9.84831750e-01 -5.18912017e-01
5.96960366e-01 1.11174297e+00 7.78756440e-01 1.38031527e-01
4.36429143e-01 -3.52642924e-01 -1.38597399e-01 -3.05497974e-01
4.25287098e-01 4.77185726e-01 5.41171208e-02 -5.24398923e-01
-1.07863462e+00 1.24913216e+00 3.18511456e-01 5.32340765e-01
-6.24995768e-01 -3.32119286e-01 6.68355227e-01 7.31674433e-01
7.87103653e-01 6.27025247e-01 -8.68396461e-01 8.20941553e-02
-4.34625536e-01 -1.31224722e-01 7.87109435e-01 1.08801758e+00
6.66321993e-01 -2.94162989e-01 -5.39787412e-02 1.11558783e+00
3.55681121e-01 1.61728367e-01 9.60463464e-01 -1.79854646e-01
4.09230113e-01 8.46107185e-01 5.29409766e-01 -8.01708817e-01
-2.15467006e-01 -4.21872973e-01 -6.43592238e-01 -4.59820367e-02
2.96861768e-01 -2.31736060e-02 -7.47520506e-01 1.57098734e+00
4.14664626e-01 -2.20174447e-01 9.29854065e-02 9.42204475e-01
7.97395051e-01 6.12264156e-01 3.94454539e-01 -2.83096097e-02
1.63534248e+00 -8.84421289e-01 -7.59614587e-01 -1.61865074e-02
8.40804517e-01 -7.60648370e-01 1.39315820e+00 5.38971782e-01
-6.38940215e-01 -9.50580060e-01 -1.08794451e+00 -5.08439504e-02
-1.09330153e+00 2.75342762e-01 6.76795065e-01 9.12721395e-01
-4.71691877e-01 4.46054131e-01 -1.74543738e-01 -4.67832208e-01
3.33656907e-01 4.08320129e-01 -6.20314479e-01 -2.00278640e-01
-1.33485186e+00 7.54261374e-01 4.67599273e-01 -2.23329365e-01
-3.05950731e-01 -7.24566102e-01 -7.71729350e-01 1.52350768e-01
1.88547835e-01 -4.88420457e-01 1.12195456e+00 -1.46227157e+00
-1.10933483e+00 8.67507041e-01 1.69345915e-01 -3.59038383e-01
1.30787000e-01 -3.29433739e-01 -8.29523683e-01 -1.82371393e-01
1.33248314e-01 3.36748958e-01 7.89367616e-01 -1.28769267e+00
-8.64764392e-01 -7.04013944e-01 1.57390032e-02 7.23607987e-02
-1.17000842e+00 1.84054375e-01 -8.77833664e-02 -6.20301604e-01
-1.11747250e-01 -9.75554883e-01 -6.79437146e-02 -1.48157865e-01
-1.08227350e-01 -6.70820653e-01 8.07413459e-01 -6.26915455e-01
1.26740730e+00 -2.18863368e+00 -6.28571734e-02 1.54073536e-01
5.03710322e-02 3.57719958e-01 -8.19526538e-02 3.51943642e-01
-1.81204379e-01 2.61701792e-01 1.71936780e-01 -2.99673408e-01
3.99852931e-01 4.85327095e-02 -3.63067895e-01 1.81472912e-01
1.39609858e-01 7.78769493e-01 -7.73464918e-01 -3.69943082e-01
2.60108620e-01 9.15559158e-02 -3.88818562e-01 3.49487931e-01
8.08210224e-02 1.56872377e-01 -5.83629608e-01 7.35903621e-01
6.27163351e-01 -1.35077193e-01 9.07348245e-02 -2.54748046e-01
3.14889997e-02 1.92000061e-01 -1.41518700e+00 1.42849946e+00
-7.50079811e-01 3.38886753e-02 -3.48787904e-01 -1.22465146e+00
1.08230960e+00 3.43392283e-01 4.67579216e-01 -5.37247837e-01
3.55138123e-01 3.06833506e-01 -6.66291192e-02 -3.76432061e-01
7.06870437e-01 -9.27416161e-02 -4.19961244e-01 7.10064709e-01
3.54001701e-01 3.14390600e-01 6.54639751e-02 -5.38772903e-02
7.05134869e-01 -2.37498522e-01 5.45647264e-01 -4.04129475e-01
3.82109046e-01 5.91567047e-02 3.10931921e-01 6.69627011e-01
-3.02196771e-01 4.46366936e-01 5.46979941e-02 -6.82665408e-01
-1.16848421e+00 -8.15387368e-01 -4.87782419e-01 1.47716892e+00
1.35975212e-01 -1.94641471e-01 -6.12946272e-01 -1.23232555e+00
3.41957718e-01 7.96457589e-01 -9.57844496e-01 -3.40345472e-01
-9.72914696e-02 -5.21413088e-01 -9.62028354e-02 9.45331633e-01
1.43555149e-01 -1.00293994e+00 -1.82941660e-01 6.68486416e-01
3.82373594e-02 -8.86017859e-01 -6.88062310e-01 4.56701815e-01
-6.21758580e-01 -8.41105819e-01 -5.94682693e-01 -1.00213313e+00
6.53513908e-01 4.05384958e-01 1.29692721e+00 -1.25846773e-01
-1.59942955e-01 2.26945654e-01 -9.45841193e-01 -7.65072465e-01
-1.45131618e-01 8.49680230e-02 4.34659392e-01 4.09825265e-01
1.21588838e+00 -8.41951221e-02 -4.24823105e-01 5.73172867e-01
-9.49969947e-01 -4.44710195e-01 7.08307862e-01 1.02547681e+00
6.82182491e-01 2.22891748e-01 1.11162257e+00 -1.37805593e+00
7.90496945e-01 -9.34480488e-01 -1.31906539e-01 2.92402238e-01
-9.22235727e-01 -1.44447342e-01 1.06319273e+00 -9.58769202e-01
-8.31424415e-01 -3.37544173e-01 -1.80281356e-01 -2.44308501e-01
-3.46092194e-01 4.73114729e-01 -2.34609962e-01 2.06660077e-01
5.34502983e-01 6.15576981e-04 -2.54670352e-01 -7.53734112e-01
5.03456235e-01 1.47892153e+00 1.09606504e-01 -4.18623060e-01
5.29733062e-01 1.28782585e-01 -5.87353826e-01 -2.97298104e-01
-1.01468027e+00 -1.05613673e+00 -1.10941374e+00 1.32461861e-01
6.01256192e-01 -6.60343409e-01 -7.76477829e-02 6.22251295e-02
-8.05667818e-01 2.07470909e-01 -4.76849824e-01 8.03159654e-01
-2.92140245e-01 -1.75466016e-02 -5.43337226e-01 -6.28803730e-01
-5.76416671e-01 -7.21195757e-01 9.45001900e-01 5.17976433e-02
-3.82996738e-01 -1.13324237e+00 -1.12482591e-03 2.66144842e-01
5.31554282e-01 -2.30215281e-01 1.21216571e+00 -1.57698727e+00
2.14585692e-01 -5.64023435e-01 -2.47787803e-01 8.61027718e-01
5.71652174e-01 -2.54234731e-01 -7.91787684e-01 -4.57695961e-01
5.05355261e-02 -6.16865039e-01 5.98612964e-01 7.86455721e-02
1.24891317e+00 -3.21750492e-01 -1.78899243e-01 2.09922820e-01
1.57687426e+00 4.42273259e-01 2.37224191e-01 4.58865106e-01
4.87517059e-01 7.15905130e-01 1.07015002e+00 5.09748757e-01
2.13459596e-01 7.35560000e-01 5.21049909e-02 -1.55600369e-01
1.32401824e-01 -2.55071402e-01 8.14018846e-02 9.82933939e-01
3.02970022e-01 -2.91315645e-01 -4.41334188e-01 6.13665044e-01
-1.74440360e+00 -6.63018703e-01 -3.20313424e-02 2.45610070e+00
7.18713582e-01 1.29842326e-01 2.44619876e-01 3.51757944e-01
7.65162170e-01 -4.54482377e-01 -6.54965043e-01 -7.85704553e-01
3.00433397e-01 1.18583150e-01 4.63098556e-01 -1.55771211e-01
-1.44874454e+00 6.08416975e-01 5.59928131e+00 8.29223096e-01
-9.57242370e-01 3.21413159e-01 6.11778975e-01 4.34825987e-01
-2.39074200e-01 -4.59114671e-01 -8.77038002e-01 8.44962835e-01
1.04969215e+00 -2.24400073e-01 2.17285641e-02 1.25467706e+00
-1.73977941e-01 4.70781446e-01 -1.39258742e+00 8.42710853e-01
3.24621171e-01 -9.27745223e-01 1.53571457e-01 1.30324110e-01
5.92866957e-01 -5.16182303e-01 3.68291080e-01 8.06807280e-01
3.00569504e-01 -8.69486392e-01 6.01138353e-01 1.72534525e-01
8.33736002e-01 -8.29099119e-01 1.39787745e+00 2.76641637e-01
-1.21174026e+00 -4.24275845e-01 -7.87379086e-01 2.39165977e-01
-7.27320388e-02 4.80161369e-01 -3.82839233e-01 6.23726845e-01
8.23911846e-01 7.66504288e-01 -5.22438586e-01 9.20137167e-01
2.24156961e-01 7.38453448e-01 8.72335285e-02 -1.53945982e-01
3.27725410e-01 -4.94031012e-01 -1.92758381e-01 1.29141712e+00
4.77722019e-01 1.26366168e-01 2.12084174e-01 6.43254995e-01
-3.88105661e-01 6.33434534e-01 -7.10333467e-01 -1.29805729e-01
5.70149839e-01 1.55353105e+00 -5.91579735e-01 -4.50696260e-01
-7.31946945e-01 9.82368648e-01 3.54695082e-01 -2.05625761e-02
-7.75296986e-01 -6.95890903e-01 6.41732335e-01 -6.33588582e-02
4.19876665e-01 8.22678208e-02 -2.97042638e-01 -1.09806395e+00
1.98544487e-02 -7.26079702e-01 6.18232012e-01 -2.74382144e-01
-2.06310153e+00 6.72716260e-01 -2.33778641e-01 -1.81707966e+00
-5.82772978e-02 -9.13857639e-01 -3.66252095e-01 7.31258810e-01
-1.53771508e+00 -1.26010966e+00 -2.36503288e-01 5.02800345e-01
7.71405339e-01 -2.94198126e-01 1.09169376e+00 6.78194523e-01
-5.06087303e-01 9.02140141e-01 6.23887599e-01 1.74624041e-01
9.83931303e-01 -1.47675645e+00 1.88817352e-01 3.16818088e-01
2.37247795e-01 8.04758251e-01 3.44031453e-01 -1.21672474e-01
-1.33655798e+00 -1.45743179e+00 9.95216310e-01 -6.65744483e-01
7.64723122e-01 -5.98256528e-01 -1.16561103e+00 7.34791279e-01
-6.11883961e-02 6.58902004e-02 1.38113260e+00 2.88384616e-01
-6.09259903e-01 -1.89309567e-01 -1.50049114e+00 6.22404702e-02
8.00836802e-01 -4.39745754e-01 -9.65795279e-01 5.40833235e-01
6.13206923e-01 1.72807753e-01 -1.15825009e+00 5.88575564e-02
6.71499550e-01 -5.59249043e-01 9.85136390e-01 -1.20905900e+00
2.10358709e-01 -1.80583932e-02 -4.19898450e-01 -1.69292367e+00
-7.35241890e-01 1.52966619e-01 3.13888192e-02 1.60959470e+00
3.68957341e-01 -8.00103068e-01 5.40208340e-01 7.20187187e-01
-3.35834138e-02 -6.54983163e-01 -5.44886827e-01 -1.05790651e+00
1.78058878e-01 -2.85790592e-01 8.99575233e-01 1.20688951e+00
-5.21017797e-02 4.14837718e-01 -2.53105640e-01 3.77406552e-02
5.24476588e-01 4.17630792e-01 6.76151097e-01 -1.59549701e+00
5.07838151e-04 -1.67305246e-01 -5.31622291e-01 -8.14098299e-01
2.77262479e-01 -7.18686998e-01 -1.48321569e-01 -1.40574551e+00
3.21872413e-01 -8.76192451e-01 -8.89035583e-01 3.48682731e-01
-7.98924640e-02 4.64124471e-01 -1.42972305e-01 3.82148921e-01
-5.82913876e-01 3.03089082e-01 6.17585242e-01 -2.53976166e-01
-1.47031412e-01 1.79085165e-01 -1.08250523e+00 6.54148996e-01
8.35244060e-01 -5.38371325e-01 -3.61041099e-01 -2.81064868e-01
-1.52769417e-01 -5.16226888e-01 -2.05467284e-01 -6.44607902e-01
1.73925608e-02 -1.48963735e-01 5.17444730e-01 -4.59055901e-01
1.60483912e-01 -1.19340587e+00 -5.51609807e-02 1.58381239e-01
-6.62381768e-01 1.18170291e-01 -1.40573591e-01 5.10297775e-01
-3.54360700e-01 -7.92917252e-01 8.12919676e-01 5.04780710e-02
-9.47109759e-01 4.10330325e-01 -9.28466469e-02 2.72120629e-03
1.28411388e+00 -3.23305905e-01 -1.56265855e-01 -5.08582592e-02
-6.08010352e-01 -5.66448979e-02 4.23431844e-01 7.21664429e-01
3.89865577e-01 -1.63826215e+00 -6.95067167e-01 2.88640589e-01
8.38202834e-01 -6.50255799e-01 6.89406022e-02 3.37828070e-01
-9.43011194e-02 5.33970237e-01 -2.88851798e-01 -2.05218464e-01
-1.19131517e+00 1.12442577e+00 -1.67710021e-01 -2.34071478e-01
-3.30705166e-01 6.79685831e-01 5.43926954e-02 -7.37119198e-01
2.36814439e-01 -3.90637934e-01 -7.02795923e-01 4.62101519e-01
7.27289438e-01 5.80437519e-02 4.62468028e-01 -7.53274024e-01
-2.44212240e-01 6.18113577e-01 -3.17751497e-01 3.52297246e-01
1.26803827e+00 -3.31422150e-01 1.88997909e-01 7.19807625e-01
1.53501606e+00 -7.64000714e-02 -6.31778419e-01 -6.45063818e-01
9.90250781e-02 -6.64780319e-01 4.62972410e-02 -7.19036698e-01
-7.79389858e-01 8.58573079e-01 9.02194798e-01 6.26748562e-01
9.89340782e-01 -1.27665564e-01 9.25584793e-01 3.90599489e-01
7.19625711e-01 -1.38332105e+00 2.10856184e-01 1.65516332e-01
5.63025951e-01 -1.58019054e+00 -2.97275871e-01 -2.58157939e-01
-9.64118123e-01 1.10308576e+00 6.32213175e-01 1.49563877e-02
7.88460314e-01 -1.77887946e-01 1.92327634e-01 7.91852698e-02
-5.86606503e-01 -1.38745412e-01 3.16197872e-01 6.54067695e-01
4.46845263e-01 2.09037319e-01 -3.32056284e-01 1.36547649e+00
1.40451953e-01 3.54906470e-02 3.10294241e-01 8.15079391e-01
-4.63029295e-01 -1.21136928e+00 -2.53093362e-01 8.51327837e-01
-5.42303205e-01 -9.66681819e-03 -3.45611513e-01 6.95182085e-01
3.87743473e-01 1.06386256e+00 1.33277521e-01 -4.19403434e-01
6.31058097e-01 1.33577585e-01 3.27909924e-02 -9.30151463e-01
-6.89481556e-01 -4.07055050e-01 -1.90795250e-02 -1.08126700e-01
-1.96320355e-01 -3.81371409e-01 -8.02440584e-01 -2.07136095e-01
-6.49080455e-01 5.55445254e-01 8.14663231e-01 7.35656202e-01
4.36716497e-01 1.92334026e-01 1.26608717e+00 -7.35749006e-01
-1.08101070e+00 -1.14580739e+00 -1.09253478e+00 1.27019227e+00
-1.85174309e-02 -8.57117414e-01 -4.80919629e-01 7.97919556e-02] | [10.017016410827637, 6.223222255706787] |
b495142e-67f1-4db3-bfa3-c68e41a385ec | feature-extraction-of-hyperspectral-images | null | null | https://doi.org/10.1109/TGRS.2013.2275613 | https://doi.org/10.1109/TGRS.2013.2275613 | Feature Extraction of Hyperspectral Images With Image Fusion and Recursive Filtering | Feature extraction is known to be an effective way in both reducing computational complexity and increasing accuracy of hyperspectral image classification. In this paper, a simple yet quite powerful feature extraction method based on image fusion and recursive filtering (IFRF) is proposed. First, the hyperspectral image is partitioned into multiple subsets of adjacent hyperspectral bands. Then, the bands in each subset are fused together by averaging, which is one of the simplest image fusion methods. Finally, the fused bands are processed with transform domain recursive filtering to get the resulting features for classification. Experiments are performed on different hyperspectral images, with the support vector machines (SVMs) serving as the classifier. By using the proposed method, the accuracy of the SVM classifier can be improved significantly. Furthermore, compared with other hyperspectral classification methods, the proposed IFRF method shows outstanding performance in terms of classification accuracy and computational efficiency. | ['Jón Atli Benediktsson', 'Shutao Li', 'Xudong Kang'] | 2013-09-16 | null | null | null | ieee-transactions-on-geoscience-and-remote-17 | ['few-shot-image-classification'] | ['computer-vision'] | [ 8.89470041e-01 -8.06171894e-01 1.26709923e-01 -1.99444398e-01
-2.68081784e-01 -4.11487013e-01 2.90819734e-01 2.43709758e-02
-2.60535300e-01 7.30999291e-01 -2.09913194e-01 -9.02196690e-02
-6.15464449e-01 -1.07030499e+00 1.05089411e-01 -1.27443075e+00
3.91087145e-01 -3.11488599e-01 2.43209712e-02 -1.11952228e-02
2.43454382e-01 7.05216050e-01 -1.90294695e+00 5.27908020e-02
1.58111405e+00 1.48715901e+00 4.05691594e-01 1.69473544e-01
-1.36792749e-01 2.74845898e-01 -3.27425450e-01 5.49825281e-02
3.93360168e-01 -6.07153118e-01 -5.01575112e-01 5.54017842e-01
-3.40759219e-03 -1.24620974e-01 -1.87492818e-01 1.49620557e+00
3.02987933e-01 5.15774012e-01 5.62473536e-01 -8.83398175e-01
-3.53573263e-01 3.68788511e-01 -8.70972633e-01 -1.35592129e-02
-9.07529667e-02 -2.98753917e-01 6.67104840e-01 -7.61685550e-01
-1.04652964e-01 8.01579535e-01 2.58356422e-01 -4.12365347e-02
-1.14732158e+00 -8.64399970e-01 -4.65365946e-02 3.86562616e-01
-1.43266523e+00 -1.51944667e-01 7.16345847e-01 -4.04123545e-01
4.18509960e-01 5.74046075e-01 7.86785483e-01 -2.22917363e-01
2.91210264e-02 4.32001382e-01 1.42365003e+00 -4.41044033e-01
1.45827636e-01 -3.79373506e-02 5.64520419e-01 5.29119909e-01
5.55523038e-01 4.11265865e-02 -1.00576013e-01 -2.04651326e-01
3.57997179e-01 5.96797585e-01 -7.77042270e-01 5.13973553e-03
-1.00463080e+00 7.15999365e-01 7.72202134e-01 5.38424730e-01
-6.24176860e-01 -7.23970592e-01 1.40728861e-01 1.65908962e-01
6.83416903e-01 1.82951018e-01 -2.83703536e-01 6.15679562e-01
-9.49349880e-01 -1.43591613e-01 3.28766793e-01 1.25743195e-01
1.06752956e+00 6.02868982e-02 2.08192080e-01 1.15439653e+00
2.12060735e-01 5.88381946e-01 5.93881488e-01 -4.56985831e-01
1.43229634e-01 8.54426205e-01 8.81461650e-02 -1.12869620e+00
-3.87072742e-01 -4.43075687e-01 -1.03023481e+00 2.58977324e-01
-6.48520980e-03 -2.14358389e-01 -8.79338622e-01 1.15106571e+00
2.03217924e-01 2.78295487e-01 5.06865859e-01 8.97927761e-01
7.08462477e-01 1.06648517e+00 7.76509047e-02 -6.57899797e-01
1.20829701e+00 -6.46349609e-01 -6.41398132e-01 3.91453691e-02
1.68440878e-01 -7.76297748e-01 5.57178259e-01 6.04377091e-01
-4.73389357e-01 -6.18807912e-01 -1.22830272e+00 4.98647928e-01
-3.40081990e-01 5.72705448e-01 8.57153058e-01 6.32660329e-01
-3.52759957e-01 4.18323010e-01 -6.38169706e-01 -1.32215917e-01
3.90888005e-01 4.20410186e-01 -3.94052058e-01 -2.45962292e-01
-8.96762133e-01 4.87116605e-01 9.55883265e-01 1.47414774e-01
3.34867798e-02 -3.86507064e-01 -8.17410231e-01 1.88958675e-01
1.67766824e-01 -1.39775217e-01 9.36778188e-01 -1.12483799e+00
-1.34140754e+00 3.25465411e-01 -4.35490847e-01 -1.10137813e-01
-1.65414661e-01 1.40899457e-02 -7.43030190e-01 3.32658947e-01
-2.21973564e-02 5.05198091e-02 8.56016695e-01 -1.05399907e+00
-1.10260260e+00 -7.38252461e-01 -2.50879377e-01 3.75541598e-01
-6.56586409e-01 -1.28830537e-01 3.82853657e-01 -3.46064448e-01
9.22014534e-01 -7.19299555e-01 -1.02168947e-01 -3.83249134e-01
-2.60482073e-01 -1.35062933e-01 1.33121383e+00 -7.85435736e-01
1.09354401e+00 -2.51385593e+00 5.36657833e-02 4.19702590e-01
-7.47376233e-02 5.58260798e-01 6.99443072e-02 1.13945477e-01
-2.98028588e-01 -1.12818502e-01 -4.54603076e-01 3.76600236e-01
-5.23241699e-01 -8.01290944e-02 -3.30262810e-01 4.54393625e-01
4.77572083e-02 3.36471379e-01 -6.09099746e-01 -1.40572086e-01
5.79929590e-01 3.62770289e-01 1.00049432e-02 1.00476101e-01
1.42012656e-01 4.56100374e-01 -5.16550064e-01 6.64311707e-01
1.03292060e+00 -9.88200754e-02 -2.24375352e-02 -5.94734192e-01
-5.00148118e-01 -2.48798966e-01 -1.23355484e+00 9.88259196e-01
-1.41199023e-01 1.75646573e-01 8.67531523e-02 -1.22792828e+00
1.18818223e+00 3.33493292e-01 6.48767948e-01 -2.84909248e-01
1.55825570e-01 3.31143498e-01 1.97806861e-02 -5.01406848e-01
4.24163401e-01 -2.85231978e-01 4.37446862e-01 1.20617703e-01
-2.36262381e-01 -3.76869768e-01 2.92737931e-01 -4.64290142e-01
2.57957041e-01 -2.88121819e-01 7.19378114e-01 -1.61194131e-01
9.28921044e-01 1.61349297e-01 4.72926021e-01 3.41653749e-02
4.48427955e-03 3.45967971e-02 -1.83534220e-01 -3.54175389e-01
-6.29840732e-01 -5.69323957e-01 -5.01776993e-01 5.82828879e-01
5.11442542e-01 1.90084159e-01 -4.38060790e-01 -2.89459229e-01
-1.12866154e-02 5.29992163e-01 -1.25800699e-01 -2.95104552e-02
1.50445998e-01 -1.23948169e+00 1.56798344e-02 1.39724106e-01
1.22345293e+00 -6.87553346e-01 -3.65228027e-01 2.44874939e-01
-2.57500082e-01 -6.92754090e-01 6.42903745e-02 5.04636765e-02
-1.05474973e+00 -1.06352377e+00 -6.28025413e-01 -7.79696167e-01
6.22973561e-01 1.23593390e+00 -1.90867539e-02 9.32014883e-02
-2.13633701e-01 -2.80918092e-01 -6.20678961e-01 -5.26003361e-01
-7.87770525e-02 -1.61676183e-01 -1.19468071e-01 4.55541193e-01
3.09377939e-01 -4.78475511e-01 -4.75581706e-01 2.31536731e-01
-1.11760366e+00 7.43068829e-02 5.73426247e-01 9.62253094e-01
5.83312273e-01 1.25910008e+00 4.82594401e-01 -5.22471070e-01
4.41892982e-01 -2.45504528e-01 -8.70754302e-01 1.50413066e-01
-4.78420854e-01 -5.07751048e-01 7.81200111e-01 -9.11413878e-02
-1.43791437e+00 2.89116353e-01 1.81691825e-01 4.81010303e-02
-3.54704171e-01 9.30853367e-01 -2.05961034e-01 -5.65976202e-01
5.16409516e-01 6.62151575e-01 2.10728675e-01 -4.02869999e-01
-6.06933348e-02 1.07830632e+00 5.00301898e-01 -3.53294313e-02
8.09325635e-01 3.95693958e-01 1.29801482e-01 -1.09704661e+00
-9.38305736e-01 -6.02471650e-01 -4.90733296e-01 -9.81649086e-02
8.09613526e-01 -9.05434966e-01 -6.14443302e-01 8.61069262e-01
-6.33382380e-01 3.91971111e-01 2.91659772e-01 9.13024068e-01
9.74816177e-03 6.40680730e-01 -4.14388090e-01 -1.08115482e+00
-4.37893331e-01 -1.02989459e+00 6.36197150e-01 8.21462035e-01
4.55059767e-01 -6.62118256e-01 -3.91603440e-01 4.81119037e-01
3.17195088e-01 2.88786083e-01 1.02208781e+00 -5.72310269e-01
-1.74115688e-01 -3.71615291e-01 -6.56740725e-01 6.15696192e-01
8.51305008e-01 1.58433586e-01 -8.76602769e-01 -3.32532018e-01
2.71454424e-01 -5.30979000e-02 9.99454260e-01 4.87779289e-01
1.30970883e+00 -6.30709976e-02 -4.14695263e-01 6.30606592e-01
1.80594158e+00 8.76692712e-01 5.91212749e-01 3.31580102e-01
4.89692330e-01 5.26535988e-01 8.38495731e-01 4.69974011e-01
-1.70383751e-01 4.47448753e-02 3.80707741e-01 -4.17819291e-01
5.07369757e-01 3.70744377e-01 1.68011505e-02 5.62031806e-01
-4.05405372e-01 4.02214937e-03 -5.26800513e-01 1.90939307e-02
-1.74670243e+00 -1.23854423e+00 -3.61222118e-01 2.16042566e+00
5.33569336e-01 -4.78487283e-01 -2.37279192e-01 8.73589694e-01
1.02621973e+00 1.33443832e-01 -4.66687232e-01 2.03290924e-01
-3.04594666e-01 4.52189237e-01 3.00969124e-01 2.98095644e-01
-1.48945189e+00 4.53036100e-01 5.48162794e+00 7.30227411e-01
-1.58764291e+00 -1.69456452e-01 4.10777748e-01 4.43592191e-01
1.77017659e-01 -7.48723224e-02 -4.39212143e-01 4.59578097e-01
3.34209323e-01 -3.59832495e-01 5.74770868e-01 5.14164984e-01
2.91503400e-01 -5.93791485e-01 -1.67959109e-01 1.11769247e+00
-4.60021161e-02 -9.45221484e-01 1.40473709e-01 -8.64966437e-02
7.68283486e-01 -4.16688889e-01 -4.40202653e-02 -2.06822947e-01
-1.65039450e-02 -7.01762915e-01 2.17643693e-01 5.16695201e-01
6.11379147e-01 -1.02853251e+00 9.85354662e-01 4.42023456e-01
-1.39030468e+00 -5.16326308e-01 -4.52386498e-01 -1.92466214e-01
-1.41534477e-01 1.03232408e+00 -2.80326724e-01 1.17984533e+00
5.92301250e-01 8.32275569e-01 -2.25703076e-01 1.31578529e+00
-2.27574017e-02 4.68051612e-01 -2.94280827e-01 9.70396176e-02
8.79147798e-02 -8.57522190e-01 2.27830440e-01 6.70797765e-01
7.90168703e-01 8.53325725e-01 5.08066416e-01 3.85969281e-01
2.77144581e-01 4.58301842e-01 -4.29211080e-01 -2.94169307e-01
3.37829113e-01 1.46648967e+00 -7.87625253e-01 -4.10157114e-01
-5.08477271e-01 7.75188804e-01 -2.59920239e-01 3.06332886e-01
-5.07497191e-01 -7.74687707e-01 4.18975830e-01 -4.41255242e-01
2.72015959e-01 -9.68510807e-02 -2.46816739e-01 -1.22895110e+00
-2.31081754e-01 -6.76000297e-01 5.68184674e-01 -8.16804588e-01
-1.11853337e+00 5.71153104e-01 -2.15781778e-01 -1.48851919e+00
4.05455083e-01 -7.05509603e-01 -3.42174411e-01 1.13419676e+00
-1.78523803e+00 -9.22555745e-01 -9.07084346e-01 6.52117372e-01
2.51199007e-01 -1.40283450e-01 9.02563393e-01 1.01017021e-01
-8.04031551e-01 -5.28023280e-02 4.90330338e-01 -1.20684743e-01
3.57007623e-01 -6.96816683e-01 -7.17173636e-01 1.03618848e+00
-2.53969818e-01 2.95605212e-01 2.77718544e-01 -5.32746077e-01
-9.86277699e-01 -1.18275499e+00 3.99706721e-01 8.11804295e-01
4.26432997e-01 4.55317408e-01 -1.02783895e+00 1.64887920e-01
5.72765665e-03 -2.01446354e-01 1.08460534e+00 -4.03636664e-01
-1.17514104e-01 -4.06713933e-01 -1.32265103e+00 3.60467881e-01
3.83509219e-01 -2.94404119e-01 -4.90604877e-01 5.04374027e-01
3.75552863e-01 -1.98045909e-01 -1.09993327e+00 7.14392006e-01
4.31599855e-01 -8.90132844e-01 7.03176975e-01 -5.75322956e-02
3.26039314e-01 -7.54423320e-01 -3.81570756e-01 -1.55012929e+00
-6.25886559e-01 -5.35121076e-02 6.02464676e-01 9.59820032e-01
2.28611678e-01 -1.24017966e+00 4.27871346e-01 3.23699981e-01
6.01731837e-02 -3.06591451e-01 -5.52287519e-01 -6.58105552e-01
-6.38719261e-01 1.15823790e-01 8.64999473e-01 1.04455519e+00
1.46051541e-01 2.22522184e-01 -8.45011473e-02 4.97721374e-01
6.61698341e-01 6.82116270e-01 3.24902147e-01 -1.73736715e+00
-4.86628301e-02 -4.26696777e-01 -3.64343286e-01 -5.19699097e-01
1.21367462e-01 -8.43277037e-01 -1.64653793e-01 -1.61111057e+00
2.91184932e-01 -3.02806646e-01 -4.50416535e-01 7.56115079e-01
-4.26864445e-01 4.49907839e-01 1.61836565e-01 4.76777673e-01
3.18489611e-01 4.51864779e-01 1.29437721e+00 -2.68430173e-01
-4.73168075e-01 3.30587029e-01 -8.16269636e-01 5.69375873e-01
9.85686898e-01 -7.44910538e-02 -3.93401712e-01 -2.32762061e-02
-5.50562203e-01 2.82022238e-01 7.79471695e-02 -1.17255151e+00
2.75382400e-02 -4.45064247e-01 7.06446707e-01 -6.39510214e-01
3.39113355e-01 -9.75111187e-01 4.80312318e-01 6.92255259e-01
5.08027300e-02 -6.05082333e-01 3.13243061e-01 2.65774846e-01
-5.44491351e-01 -3.55991662e-01 9.92052853e-01 2.40625311e-02
-9.35151696e-01 3.50662500e-01 -3.29551876e-01 -1.06537735e+00
1.23142779e+00 -5.19850552e-01 -3.33044291e-01 -1.11041345e-01
-7.04269826e-01 -8.96153152e-02 2.47811139e-01 -1.08117595e-01
5.88370085e-01 -1.14866340e+00 -7.09996104e-01 3.58380616e-01
1.92188010e-01 -3.11723143e-01 4.13711905e-01 8.69563162e-01
-4.65571582e-01 2.74830282e-01 -3.10256690e-01 -4.71792638e-01
-1.64425278e+00 3.89077187e-01 3.02990764e-01 2.37457324e-02
-2.50040799e-01 4.83359933e-01 -2.40490455e-02 -8.06281045e-02
-3.50776255e-01 -8.04694891e-02 -5.88197351e-01 2.84083575e-01
7.05569029e-01 3.85815173e-01 3.23643893e-01 -8.02891850e-01
-1.00013778e-01 7.80036330e-01 7.15407133e-02 1.48181975e-01
1.40731573e+00 1.70459449e-02 -7.87710309e-01 1.35249598e-02
1.01679063e+00 -6.36029094e-02 -7.40032852e-01 -3.61866862e-01
-4.56554115e-01 -6.99731529e-01 3.85356784e-01 -7.17342079e-01
-1.12027586e+00 6.71247602e-01 7.27584124e-01 5.47721624e-01
1.83509505e+00 -6.21222734e-01 5.23865104e-01 3.63816410e-01
4.94487047e-01 -6.75337136e-01 -6.26626253e-01 2.80660361e-01
5.92223585e-01 -1.17237425e+00 2.12810308e-01 -8.99532676e-01
-2.61177570e-01 1.47762918e+00 4.44222212e-01 9.46667269e-02
7.18295157e-01 -6.53191879e-02 -3.17023136e-02 1.26493990e-01
-1.90818694e-03 -4.00465637e-01 2.63811082e-01 3.53806853e-01
4.00208980e-01 2.67818093e-01 -7.22791493e-01 4.77451682e-01
-1.43393636e-01 9.43254530e-02 1.78927064e-01 8.41571271e-01
-1.02179766e+00 -7.92670250e-01 -7.36144245e-01 8.84189725e-01
-2.75211334e-01 6.44786134e-02 -9.87729356e-02 2.90900677e-01
1.44133940e-01 1.37155986e+00 -1.47809207e-01 -5.76290488e-01
2.27817774e-01 1.29134312e-01 3.04576516e-01 -4.68375117e-01
-2.40925729e-01 2.51250118e-01 -3.05152148e-01 -7.27805346e-02
-7.52611220e-01 -4.85975116e-01 -1.30384409e+00 -1.52566284e-01
-9.89704430e-01 4.77870464e-01 6.12437248e-01 1.09674096e+00
9.74030793e-02 3.12006593e-01 1.12147677e+00 -9.75382268e-01
-5.12716472e-01 -8.12418640e-01 -1.14631009e+00 2.99254954e-01
2.31330007e-01 -6.31183207e-01 -5.75225234e-01 1.26036197e-01] | [9.816518783569336, -1.7810112237930298] |
9396a6cd-ca36-442c-81d0-e48431da68f6 | edge-augmented-graph-transformers-global-self | 2108.03348 | null | https://arxiv.org/abs/2108.03348v3 | https://arxiv.org/pdf/2108.03348v3.pdf | Global Self-Attention as a Replacement for Graph Convolution | We propose an extension to the transformer neural network architecture for general-purpose graph learning by adding a dedicated pathway for pairwise structural information, called edge channels. The resultant framework - which we call Edge-augmented Graph Transformer (EGT) - can directly accept, process and output structural information of arbitrary form, which is important for effective learning on graph-structured data. Our model exclusively uses global self-attention as an aggregation mechanism rather than static localized convolutional aggregation. This allows for unconstrained long-range dynamic interactions between nodes. Moreover, the edge channels allow the structural information to evolve from layer to layer, and prediction tasks on edges/links can be performed directly from the output embeddings of these channels. We verify the performance of EGT in a wide range of graph-learning experiments on benchmark datasets, in which it outperforms Convolutional/Message-Passing Graph Neural Networks. EGT sets a new state-of-the-art for the quantum-chemical regression task on the OGB-LSC PCQM4Mv2 dataset containing 3.8 million molecular graphs. Our findings indicate that global self-attention based aggregation can serve as a flexible, adaptive and effective replacement of graph convolution for general-purpose graph learning. Therefore, convolutional local neighborhood aggregation is not an essential inductive bias. | ['Dharmashankar Subramanian', 'Mohammed J. Zaki', 'Md Shamim Hussain'] | 2021-08-07 | null | null | null | null | ['graph-property-prediction', 'graph-regression'] | ['graphs', 'graphs'] | [ 1.82943001e-01 2.61809796e-01 -2.19229698e-01 -1.82902217e-01
-7.90720657e-02 -4.02513921e-01 5.06208837e-01 5.51841557e-01
-3.73238266e-01 8.63632739e-01 -4.74302471e-03 -7.86669254e-01
7.10915998e-02 -1.24633503e+00 -1.28356886e+00 -1.06474113e+00
-4.91713196e-01 5.34400582e-01 1.66624069e-01 -5.21950364e-01
1.45114660e-02 5.45820475e-01 -9.59223390e-01 5.27641714e-01
9.42114353e-01 9.26414371e-01 2.11283207e-01 6.55066371e-01
-4.04093564e-01 7.02673554e-01 -2.54177302e-01 -6.18356287e-01
-6.86315000e-02 -2.59353995e-01 -6.72733665e-01 -5.13634205e-01
4.24411178e-01 3.06557894e-01 -6.14635348e-01 9.12646294e-01
6.44414306e-01 8.75073373e-02 6.12488270e-01 -9.64022994e-01
-1.29578686e+00 8.61290455e-01 -1.32467896e-01 1.73077941e-01
1.89241186e-01 6.83668852e-01 1.48375368e+00 -6.47950292e-01
8.69113684e-01 1.38411939e+00 7.39894986e-01 5.22474527e-01
-1.54724562e+00 -7.72020221e-01 4.27771568e-01 2.43620038e-01
-1.01240516e+00 9.39505175e-03 8.11631799e-01 -3.48256975e-01
1.69690132e+00 6.39285296e-02 9.74711418e-01 1.30455661e+00
7.69191325e-01 4.46393311e-01 7.69263268e-01 1.04196161e-01
7.25791529e-02 -5.69948077e-01 4.26256567e-01 1.13802838e+00
4.00994927e-01 8.57252553e-02 -5.06501079e-01 -2.13153735e-01
6.35744870e-01 1.32037833e-01 -2.04852045e-01 -3.09920311e-01
-1.13908458e+00 8.57933640e-01 1.46867776e+00 -2.19658427e-02
-2.78613210e-01 6.65722251e-01 5.02493382e-01 6.43423736e-01
5.44772208e-01 7.27998853e-01 -4.30003792e-01 4.12784100e-01
1.86197888e-02 -3.11435573e-03 6.20751977e-01 6.89453542e-01
9.88337159e-01 2.13137418e-01 -3.28925967e-01 3.44687164e-01
2.97451079e-01 3.09867740e-01 1.76031202e-01 -1.00138530e-01
4.54020381e-01 1.08435488e+00 -5.16878366e-01 -8.18951964e-01
-6.62432313e-01 -7.57379711e-01 -1.23531163e+00 -2.38988027e-02
1.24821380e-01 3.94248366e-02 -1.18574321e+00 1.93531466e+00
2.39623964e-01 1.57628134e-01 4.98493202e-02 5.45493722e-01
1.45558679e+00 6.19181812e-01 3.84623826e-01 1.20424733e-01
1.05393517e+00 -9.93482530e-01 -5.66489816e-01 -1.92254677e-01
9.68916416e-01 -1.17392629e-01 1.05332029e+00 6.07265458e-02
-8.75431895e-01 -4.72098976e-01 -1.10132384e+00 -3.36446673e-01
-9.65051055e-01 -3.80251288e-01 1.40358627e+00 1.87543869e-01
-1.21268928e+00 9.91455257e-01 -6.95253670e-01 -1.21371731e-01
5.61810553e-01 8.17442775e-01 -5.99616766e-01 -9.78272781e-02
-1.45387554e+00 5.36814272e-01 4.65613157e-01 1.73160687e-01
-7.93319941e-01 -7.52449811e-01 -1.02340353e+00 2.37838104e-01
2.24953234e-01 -9.29269373e-01 5.54986656e-01 -7.22555220e-01
-1.33035874e+00 6.24423862e-01 4.00854461e-02 -7.63272107e-01
-6.73703775e-02 2.05275908e-01 -2.96528250e-01 -6.79970682e-02
-2.12154627e-01 6.60885215e-01 6.71952546e-01 -7.51552165e-01
2.09196731e-01 -4.58829969e-01 1.83258981e-01 3.62071469e-02
-1.86190218e-01 -4.05893445e-01 -2.81840891e-01 -5.54260850e-01
-1.42482877e-01 -1.04189897e+00 -4.75153923e-01 -3.35733801e-01
-5.39219797e-01 -4.27418411e-01 6.60680175e-01 -2.48373583e-01
1.24839044e+00 -1.82969499e+00 5.70166230e-01 1.68560892e-01
9.27976787e-01 1.64087355e-01 -4.49683815e-01 6.40567064e-01
-5.14178157e-01 3.22187662e-01 -2.78773546e-01 -1.38870999e-01
-1.47762001e-01 1.15790613e-01 -3.61402109e-02 2.17134833e-01
5.76960325e-01 1.81672120e+00 -1.01758027e+00 1.71543267e-02
4.57118861e-02 7.10213065e-01 -8.44617069e-01 1.69073135e-01
-7.04304695e-01 4.13788497e-01 -4.83681023e-01 5.04984975e-01
5.38095176e-01 -9.09193993e-01 3.13342333e-01 -3.99631441e-01
2.11088225e-01 5.28557897e-01 -3.48231018e-01 1.75029683e+00
-1.52763233e-01 3.40123743e-01 -1.85382694e-01 -1.11437917e+00
8.08562279e-01 -1.16352320e-01 3.40535343e-01 -8.62518907e-01
4.01883312e-02 5.09296358e-02 5.96605539e-01 -3.89764830e-02
9.86991823e-02 1.52713090e-01 -5.63461706e-02 3.30127269e-01
4.70991969e-01 1.67799652e-01 1.33133292e-01 5.88294625e-01
1.41928029e+00 -6.79534823e-02 8.61855969e-02 -4.79668677e-01
3.31250548e-01 -2.93701231e-01 2.60987043e-01 6.72968388e-01
2.90318191e-01 6.04351461e-02 8.89152110e-01 -7.70408213e-01
-8.45124662e-01 -7.54611850e-01 7.66121969e-02 1.34196186e+00
-4.95982356e-03 -9.88525033e-01 -4.89203274e-01 -8.14224601e-01
2.87620366e-01 1.91143021e-01 -8.52458596e-01 -7.01272190e-01
-4.23128843e-01 -1.20795536e+00 3.10687810e-01 4.93025362e-01
3.33822638e-01 -1.23658001e+00 3.04471999e-01 4.88903522e-01
4.47101712e-01 -8.99823785e-01 -7.82976091e-01 6.99050188e-01
-7.32284427e-01 -1.19102168e+00 -2.06857964e-01 -6.00747764e-01
6.99418366e-01 1.92977056e-01 1.36731970e+00 4.65463966e-01
-3.37308466e-01 7.28973374e-02 -9.31617394e-02 -7.44689107e-02
-4.06767666e-01 3.58734250e-01 -2.06240073e-01 1.70072526e-01
1.10641874e-01 -8.67576480e-01 -7.54578352e-01 -2.46235691e-02
-7.16325462e-01 2.33053252e-01 6.15855157e-01 1.19210935e+00
6.17380261e-01 -3.29543233e-01 7.05017507e-01 -1.37903285e+00
8.42282593e-01 -4.69667494e-01 -6.00507855e-01 2.39829376e-01
-6.62269831e-01 7.28003383e-01 9.81910169e-01 -2.92062104e-01
-4.75599587e-01 -5.77523112e-02 -2.51371026e-01 -3.70653331e-01
4.86203074e-01 8.59004796e-01 -3.57066214e-01 -6.09345675e-01
6.52002692e-01 1.58326000e-01 8.84992480e-02 -1.78840607e-01
6.06876493e-01 -1.58383697e-02 1.86621934e-01 -7.23792791e-01
4.58812892e-01 6.91166669e-02 5.73039591e-01 -8.34360957e-01
-4.58141863e-01 -7.94580504e-02 -6.17832482e-01 1.54898554e-01
1.04829884e+00 -8.88650358e-01 -1.14559388e+00 7.25670099e-01
-1.11657953e+00 -8.51583898e-01 8.69261324e-02 1.14503652e-02
-7.53722563e-02 3.34400922e-01 -9.72110987e-01 -4.38879192e-01
-6.66624188e-01 -1.49430335e+00 1.14080429e+00 -1.16912693e-01
3.48583311e-01 -1.44494462e+00 -1.92293227e-01 1.54774077e-02
4.86717314e-01 4.23556447e-01 1.33867788e+00 -7.26882875e-01
-1.01809025e+00 -6.47868216e-02 -3.76805484e-01 1.26626790e-01
2.10070191e-03 -3.10548488e-02 -7.47054636e-01 -5.69519043e-01
-8.52147341e-01 -3.96294475e-01 1.56382251e+00 2.05870211e-01
1.49930477e+00 -3.42404127e-01 -5.16336083e-01 1.07372940e+00
1.20256066e+00 -1.86058551e-01 6.51444972e-01 -2.38913685e-01
1.47369075e+00 -5.33092543e-02 -2.24898398e-01 -8.92213210e-02
4.97704744e-01 4.69842732e-01 7.54327297e-01 -4.38286245e-01
-1.97462693e-01 -4.03155446e-01 5.26121676e-01 9.21471357e-01
-2.93488264e-01 -5.95391512e-01 -8.61590028e-01 -1.65033177e-01
-1.91511035e+00 -7.33012795e-01 -3.44682813e-01 2.02602220e+00
7.98964798e-01 3.30615044e-01 -1.01586469e-01 -3.53508651e-01
5.01790524e-01 3.67067069e-01 -8.56075048e-01 -4.34868425e-01
-3.40013623e-01 7.11754858e-01 7.81717062e-01 6.09631777e-01
-1.04045272e+00 1.20383430e+00 6.02865601e+00 7.53923476e-01
-1.35175419e+00 1.54374260e-02 7.38136113e-01 2.67944694e-01
-6.15885854e-01 -3.28805819e-02 -6.60577476e-01 5.77223480e-01
1.12890315e+00 9.77587551e-02 7.09632576e-01 5.51677585e-01
-3.56238604e-01 6.22244656e-01 -1.11056554e+00 8.43523920e-01
-2.79973328e-01 -2.01596808e+00 5.03614187e-01 3.72735918e-01
5.15811563e-01 5.15190482e-01 1.08874813e-01 4.80769455e-01
6.68248773e-01 -1.52284336e+00 1.34657115e-01 4.66639429e-01
9.92267966e-01 -5.84353268e-01 5.32996535e-01 -1.47986442e-01
-1.50611508e+00 4.28146571e-02 -4.01502252e-01 3.06478217e-02
-2.79157996e-01 5.46015143e-01 -7.09084272e-01 6.54941320e-01
3.59884918e-01 1.13027489e+00 -7.87598729e-01 6.05221927e-01
-2.50779331e-01 4.63401794e-01 -1.75575674e-01 -1.89933300e-01
5.00707626e-01 -4.53163207e-01 3.83727789e-01 1.26748633e+00
-8.75876099e-02 -8.96559060e-02 3.05404693e-01 1.01066041e+00
-7.46806502e-01 1.20543838e-02 -9.80421782e-01 -5.95082045e-01
2.20879629e-01 1.24483705e+00 -5.64633131e-01 -1.87631935e-01
-3.70762646e-01 9.65431929e-01 1.00758338e+00 6.01988435e-01
-8.08277905e-01 -2.78417230e-01 8.11238170e-01 1.51564449e-01
2.89379746e-01 -3.64845127e-01 1.86702266e-01 -1.13905621e+00
-3.15103710e-01 -7.96682954e-01 4.86014843e-01 -6.35447145e-01
-1.47754788e+00 6.95690155e-01 -5.91430068e-01 -4.26869214e-01
4.09604400e-01 -1.20162821e+00 -8.72778177e-01 8.33829284e-01
-1.48107409e+00 -1.38142323e+00 -2.44265676e-01 5.29852748e-01
-8.21049958e-02 -2.25682810e-01 8.83168936e-01 1.47463679e-01
-7.84784794e-01 7.89243639e-01 1.36271026e-02 1.13960765e-01
4.48852241e-01 -1.64451683e+00 1.06040788e+00 4.30128664e-01
1.54656649e-01 9.67283368e-01 2.03373700e-01 -9.11381364e-01
-2.19401550e+00 -1.52973199e+00 3.20862293e-01 -6.75340176e-01
9.41328168e-01 -9.38544273e-01 -1.20219374e+00 7.54607201e-01
1.79863781e-01 4.11059350e-01 6.70331240e-01 3.77302647e-01
-6.93387866e-01 -5.95183447e-02 -6.53198481e-01 5.49956620e-01
1.61085320e+00 -7.06204891e-01 6.54159635e-02 6.34779096e-01
1.50934815e+00 -3.41639727e-01 -1.15534937e+00 4.59344506e-01
1.64580941e-01 -5.62244892e-01 1.17642736e+00 -1.22753572e+00
2.96405822e-01 -6.42444566e-02 1.16829619e-01 -1.52969861e+00
-7.07005680e-01 -1.03660464e+00 -3.25408995e-01 5.06923318e-01
8.29241633e-01 -1.17374682e+00 6.18803263e-01 1.91804036e-01
-6.12310171e-01 -9.91538525e-01 -8.46341908e-01 -4.83025938e-01
2.12878585e-01 -1.28088385e-01 8.01072240e-01 8.54415596e-01
-1.68232232e-01 9.62186933e-01 -6.33687228e-02 -9.72655937e-02
6.58283353e-01 4.36620116e-02 6.61737442e-01 -1.40306115e+00
-4.32603240e-01 -7.02566624e-01 -5.72812676e-01 -9.44099486e-01
4.19175863e-01 -1.72445261e+00 -5.35782754e-01 -1.52576804e+00
8.75315517e-02 -3.78771365e-01 -7.29593933e-01 5.65112829e-01
-4.23652202e-01 2.16738228e-02 -1.21657349e-01 -1.67897299e-01
-7.57232070e-01 8.71914506e-01 1.70703924e+00 -5.43005228e-01
-7.44449720e-02 -3.33012819e-01 -8.54601264e-01 -1.50479209e-02
5.74898243e-01 -1.21316969e-01 -4.71681774e-01 -3.14284563e-01
6.98253155e-01 -2.37889469e-01 4.15191442e-01 -6.39705598e-01
1.77379996e-01 -2.63956040e-02 2.84957618e-01 -2.21790791e-01
2.59425759e-01 -4.47961509e-01 1.79083377e-01 5.30184150e-01
-3.03070784e-01 2.09358573e-01 2.44164079e-01 9.87823248e-01
2.23060399e-01 5.41115642e-01 5.85942745e-01 -3.60948622e-01
-4.24571216e-01 1.11705375e+00 6.26135990e-02 -8.14748406e-02
6.53353751e-01 1.07016340e-01 -7.06482768e-01 -7.80319124e-02
-8.14297080e-01 3.36553752e-01 5.34841835e-01 1.19699091e-01
5.49035549e-01 -1.33037031e+00 -5.06059647e-01 4.69353110e-01
3.24160308e-01 5.23661301e-02 5.87155707e-02 7.65799701e-01
-4.61231947e-01 5.49555659e-01 -9.54798609e-02 -5.19135356e-01
-8.68667781e-01 9.15069222e-01 5.18089771e-01 -4.77358699e-01
-7.07938671e-01 1.03643680e+00 6.74369812e-01 -5.91796517e-01
-1.41607106e-01 -6.47664785e-01 6.42130896e-02 -1.67748839e-01
9.73554179e-02 -3.48392427e-01 2.52059907e-01 -3.41846108e-01
-2.34134793e-01 3.39585662e-01 -2.18459725e-01 7.97779143e-01
1.40289259e+00 4.46737945e-01 -5.46886206e-01 2.40690485e-01
1.28392112e+00 -3.38095754e-01 -1.05851805e+00 -2.69821465e-01
-2.44328335e-01 2.00226367e-01 1.33039579e-01 -6.30679309e-01
-1.21118498e+00 1.09353185e+00 1.90969408e-01 2.66389966e-01
6.80664062e-01 1.19582795e-01 6.87810063e-01 8.50145876e-01
3.54029775e-01 -5.24980366e-01 3.17173302e-01 7.16205418e-01
9.86233056e-01 -1.25805783e+00 -1.38002068e-01 -3.00117314e-01
-3.82858664e-01 9.92136896e-01 7.79125869e-01 -1.47971064e-01
6.37096822e-01 1.52572673e-02 -6.90812290e-01 -7.55020857e-01
-1.21206248e+00 -2.63886988e-01 5.14973462e-01 5.71772218e-01
6.22600436e-01 3.73959780e-01 -1.73183158e-02 5.62724829e-01
1.43977720e-02 -5.94138980e-01 1.23965427e-01 4.18620139e-01
-3.36878896e-01 -1.23629522e+00 4.68956828e-01 7.36419201e-01
-2.03515917e-01 -7.09498763e-01 -7.35136628e-01 6.96492910e-01
-1.81294471e-01 2.84737885e-01 -1.51453406e-01 -5.63089490e-01
1.31658278e-02 2.07628131e-01 5.34526169e-01 -7.86851883e-01
-7.19601870e-01 -2.20111698e-01 9.34391022e-02 -6.62868977e-01
-8.03995579e-02 -5.76834194e-02 -1.31240416e+00 -4.98043627e-01
-2.98819453e-01 2.50468105e-01 2.70472825e-01 5.83333850e-01
8.62362266e-01 1.07494676e+00 3.38433802e-01 -7.72918224e-01
-1.40548199e-01 -8.64795327e-01 -2.32589900e-01 4.48142201e-01
4.53379989e-01 -6.98197663e-01 -2.80896038e-01 -4.89201725e-01] | [6.796009063720703, 6.217291831970215] |
d0fc8ada-6690-4b8a-a039-09c4cda223e2 | assessing-pattern-recognition-performance-of | 2012.10355 | null | https://arxiv.org/abs/2012.10355v2 | https://arxiv.org/pdf/2012.10355v2.pdf | Assessing Pattern Recognition Performance of Neuronal Cultures through Accurate Simulation | Previous work has shown that it is possible to train neuronal cultures on Multi-Electrode Arrays (MEAs), to recognize very simple patterns. However, this work was mainly focused to demonstrate that it is possible to induce plasticity in cultures, rather than performing a rigorous assessment of their pattern recognition performance. In this paper, we address this gap by developing a methodology that allows us to assess the performance of neuronal cultures on a learning task. Specifically, we propose a digital model of the real cultured neuronal networks; we identify biologically plausible simulation parameters that allow us to reliably reproduce the behavior of real cultures; we use the simulated culture to perform handwritten digit recognition and rigorously evaluate its performance; we also show that it is possible to find improved simulation parameters for the specific task, which can guide the creation of real cultures. | ['Giuseppe Amato', 'Federico Cremisi', 'Tommaso Pizzorusso', 'Guido Marco Cicchini', 'Claudio Gennaro', 'Fabrizio Falchi', 'Raffaele Mazziotti', 'Gabriele Lagani'] | 2020-12-18 | null | null | null | null | ['handwritten-digit-recognition'] | ['computer-vision'] | [ 3.22666019e-01 -3.16312201e-02 7.20508277e-01 2.30499029e-01
-1.49281114e-01 -5.60009420e-01 6.70124412e-01 -1.74258664e-01
-4.93602037e-01 1.02520525e+00 -2.89394379e-01 -2.64073461e-01
3.30445357e-02 -6.74134612e-01 -1.04730701e+00 -9.19828415e-01
-5.94431311e-02 2.55123526e-01 4.50117409e-01 1.70040265e-01
2.58683473e-01 9.37409401e-01 -1.76763010e+00 1.10491611e-01
4.98584688e-01 5.36994517e-01 4.09633875e-01 5.65736830e-01
3.85070533e-01 4.34318870e-01 -6.91042304e-01 2.23743603e-01
1.55275986e-01 -7.74862170e-01 -4.41610366e-01 1.96999520e-01
-1.24725295e-04 1.14584319e-01 1.11328065e-01 8.11043441e-01
3.75706106e-01 -2.31364921e-01 1.08268845e+00 -1.04278362e+00
-3.09441566e-01 1.29300937e-01 2.45206907e-01 2.86903322e-01
1.77816957e-01 2.80551203e-02 7.43464753e-02 -5.99263132e-01
7.38897622e-01 7.07579076e-01 4.74770546e-01 5.90650856e-01
-1.32586837e+00 -7.76966512e-01 -2.87782490e-01 -2.29176939e-01
-1.47189128e+00 -4.96931612e-01 5.35376191e-01 -4.23258603e-01
8.22898746e-01 -6.05500527e-02 1.08444011e+00 1.17650664e+00
7.00496435e-01 3.75590950e-01 1.45449507e+00 -7.28348792e-01
7.31307805e-01 1.44965991e-01 -1.66877672e-01 5.69724917e-01
5.15943408e-01 1.59428671e-01 -1.28745988e-01 2.27658689e-01
1.26642978e+00 -4.85507578e-01 -4.42167491e-01 -2.78295457e-01
-1.17093587e+00 3.02512795e-01 -1.43323198e-01 1.02686942e+00
-5.52421749e-01 1.60732150e-01 -8.65398496e-02 2.26028278e-01
-7.87422955e-02 7.18554199e-01 3.51395682e-02 -1.36055350e-01
-7.16378927e-01 1.96848050e-01 1.18956864e+00 5.54510891e-01
4.99783069e-01 2.09453672e-01 1.62842780e-01 6.08779848e-01
1.64526910e-01 3.30025285e-01 4.70594525e-01 -1.21528423e+00
-2.51491159e-01 5.14202893e-01 1.23458147e-01 -7.11735487e-01
-3.14556599e-01 -4.36199695e-01 -6.60544574e-01 7.79869616e-01
6.36905849e-01 -4.72028464e-01 -6.79450631e-01 1.45943463e+00
-3.27029109e-01 5.77893317e-01 3.59346032e-01 6.35159254e-01
1.64928019e-01 6.91862226e-01 -1.61761075e-01 -5.29175580e-01
6.96049690e-01 -3.15042883e-01 -4.29294825e-01 1.19661875e-01
7.01597691e-01 -3.11347008e-01 9.42999363e-01 6.04872048e-01
-1.14171350e+00 -2.56339639e-01 -1.46278083e+00 8.33993971e-01
-3.75502229e-01 -6.41289726e-02 3.09223592e-01 6.83424056e-01
-1.28345835e+00 7.90098727e-01 -7.87594736e-01 -7.14955926e-01
1.22954182e-01 5.96688330e-01 -3.56394917e-01 8.74083638e-02
-7.96108603e-01 1.26310956e+00 2.53381252e-01 2.03978140e-02
-9.69677031e-01 -2.65165955e-01 -3.83043230e-01 7.24641159e-02
-3.07578206e-01 -7.15881228e-01 9.01403368e-01 -9.62671280e-01
-1.79377294e+00 6.75152600e-01 -5.16261570e-02 -3.87342006e-01
1.06540658e-01 5.17753899e-01 -2.60742396e-01 1.87149912e-01
-4.66662645e-01 6.55540645e-01 9.58847851e-02 -1.77385271e+00
-2.42867827e-01 -2.50598490e-01 -2.03803048e-01 -1.15918033e-01
-4.03279334e-01 6.52124956e-02 -1.90622866e-01 -3.13700259e-01
-6.13548718e-02 -9.82825220e-01 -6.37852922e-02 -3.10671657e-01
1.13776617e-01 1.45529941e-01 4.64665115e-01 -1.42196655e-01
6.35318518e-01 -2.07983994e+00 1.91944659e-01 4.92493391e-01
-8.58137310e-02 4.82877284e-01 -3.09568673e-01 6.27635241e-01
2.14344665e-01 1.77327588e-01 -3.47287327e-01 -1.29935578e-01
-4.19503570e-01 5.37830949e-01 -1.52680248e-01 3.24664772e-01
2.41101176e-01 7.52572119e-01 -3.49340260e-01 -3.72176111e-01
-1.31011438e-02 7.04984009e-01 -2.33200684e-01 -1.48127019e-01
-5.02786711e-02 5.90529263e-01 -3.42479944e-01 3.26098740e-01
8.45726281e-02 2.33252365e-02 2.05817327e-01 2.38610983e-01
-4.80824560e-01 -2.28527829e-01 -1.10492551e+00 9.77478802e-01
-3.64396214e-01 6.81912243e-01 2.22320110e-03 -1.04074657e+00
1.10526776e+00 5.55636644e-01 3.65410179e-01 -6.97714329e-01
5.87806523e-01 4.86252487e-01 2.15882197e-01 -1.98730871e-01
-3.66775274e-01 -3.18925738e-01 2.57347047e-01 5.38281798e-01
2.42927726e-02 -3.45515162e-01 3.23138863e-01 -4.35607523e-01
1.20188498e+00 -2.03607157e-01 -1.38160348e-01 -4.49335724e-01
4.82005924e-01 -1.17360413e-01 1.84525967e-01 6.16586626e-01
1.72455952e-01 6.90685391e-01 5.10343671e-01 -4.86707408e-03
-1.21804428e+00 -1.07216537e+00 -2.55346417e-01 1.40749171e-01
1.16149172e-01 2.60100037e-01 -1.03519821e+00 6.62128329e-02
-2.21153289e-01 6.38736844e-01 -6.89694881e-01 -2.99143083e-02
-4.58057970e-01 -9.19563949e-01 7.28788733e-01 3.13900292e-01
5.05368948e-01 -1.36782837e+00 -7.05547154e-01 4.86865014e-01
5.36954641e-01 -9.44687009e-01 3.80854726e-01 3.90874654e-01
-8.01109791e-01 -8.64617646e-01 -9.84886467e-01 -1.13777649e+00
8.47654760e-01 -8.02233070e-02 4.19249415e-01 2.35679954e-01
-7.55702853e-02 5.42658925e-01 -2.29986027e-01 -4.49908316e-01
-7.11490035e-01 -1.82566255e-01 2.20536992e-01 -9.74072814e-02
1.64733127e-01 -1.19040298e+00 -3.35431755e-01 5.10035396e-01
-1.15455401e+00 -9.70231462e-03 7.87869215e-01 4.44720715e-01
6.79520905e-01 -1.28722545e-02 8.89944196e-01 -6.06133819e-01
8.93483877e-01 -1.37527376e-01 -6.87494934e-01 3.38164270e-01
-6.51848793e-01 5.27876616e-02 7.79071689e-01 -4.91821051e-01
-6.63384974e-01 1.84029460e-01 -2.53771335e-01 -1.50924131e-01
-5.16445041e-01 5.63896775e-01 -5.64677492e-02 -6.74425960e-01
5.63898563e-01 7.76477277e-01 3.16080481e-01 -2.09919870e-01
-4.55831915e-01 6.06207728e-01 5.05986631e-01 -6.16024792e-01
4.41873521e-01 2.56328821e-01 2.31220856e-01 -9.18066442e-01
1.13316968e-01 2.53927261e-02 -5.60764074e-01 -5.99143922e-01
5.69987416e-01 -5.00467420e-01 -7.42317677e-01 7.14288294e-01
-9.23507690e-01 -8.72824609e-01 8.25758651e-02 5.80107987e-01
-9.23909783e-01 1.17857559e-02 -6.20029867e-01 -7.75611103e-01
4.66833115e-02 -9.75052178e-01 7.03053772e-01 2.68011034e-01
-1.24195307e-01 -9.92249012e-01 4.83238697e-01 -3.01955193e-01
3.34066182e-01 2.73610771e-01 1.11736584e+00 -4.39238280e-01
-6.86401486e-01 -2.67164204e-02 1.27550900e-01 3.61985475e-01
-6.72738850e-02 3.55002314e-01 -9.94814098e-01 -2.21045196e-01
1.52385473e-01 -5.73172495e-02 6.47885621e-01 5.04644156e-01
8.72508109e-01 1.44428179e-01 -6.55672014e-01 3.15519840e-01
1.53068423e+00 7.61566877e-01 1.22368193e+00 3.74246061e-01
-8.27383716e-03 5.49846530e-01 -9.08295065e-02 -1.58483788e-01
-1.02019347e-01 5.08339763e-01 3.00174579e-02 9.92443562e-02
-6.08874001e-02 6.85595302e-03 2.37625644e-01 1.03176856e+00
-4.47585553e-01 -4.32343602e-01 -1.05714393e+00 4.40249592e-01
-1.61601818e+00 -6.73845768e-01 -1.44922705e-02 2.14046383e+00
5.12369633e-01 2.72052109e-01 1.56698361e-01 5.03242373e-01
5.53420722e-01 -7.14921057e-01 -1.38725147e-01 -3.57116014e-01
-3.14651072e-01 3.86582762e-01 1.83254331e-01 4.45020974e-01
-4.60279971e-01 5.09972394e-01 7.90432024e+00 1.88054681e-01
-1.34721959e+00 -3.33421409e-01 5.12451351e-01 -8.83153901e-02
-1.09557353e-01 -2.45348290e-01 -7.00470150e-01 3.76768470e-01
1.18357372e+00 -2.66918808e-01 4.27053750e-01 2.46268287e-01
3.46564263e-01 -3.06220919e-01 -1.16329408e+00 6.35052919e-01
6.15364350e-02 -1.51557279e+00 1.72748238e-01 3.56076092e-01
6.89651847e-01 -2.91473627e-01 -1.14525616e-01 1.62901282e-01
6.17921911e-02 -1.18222642e+00 3.68138105e-01 8.53484154e-01
2.72881538e-01 -6.78423345e-01 8.02246630e-01 6.38261139e-01
-4.92041081e-01 3.61389597e-04 -3.16224456e-01 -3.30519646e-01
-3.91955562e-02 3.61742228e-01 -8.90053630e-01 1.71412118e-02
2.66941875e-01 3.65400940e-01 -5.60299814e-01 1.68464267e+00
3.08676690e-01 6.79891467e-01 -3.56329769e-01 -6.70237422e-01
3.75144579e-03 -2.41820320e-01 3.11304301e-01 9.76146579e-01
8.32301974e-01 1.67142764e-01 -3.51506650e-01 9.13129628e-01
2.27773950e-01 1.23368546e-01 -9.49957848e-01 -1.90911517e-01
3.22610140e-01 9.21042919e-01 -1.10444784e+00 -1.61023308e-02
-2.25428492e-01 6.59818947e-01 3.68050843e-01 5.96510172e-01
-3.73812020e-01 -2.44309932e-01 3.21871817e-01 8.57975632e-02
3.75584871e-01 -5.15290558e-01 -4.79689151e-01 -8.85728598e-01
-2.58713037e-01 -7.40109980e-01 -3.36100668e-01 -9.82908726e-01
-8.34254503e-01 5.16085327e-01 -1.21768810e-01 -9.02908683e-01
-4.06622440e-01 -8.55416298e-01 -7.26467133e-01 7.54351974e-01
-9.77760017e-01 -5.99069834e-01 5.19119669e-04 3.11997980e-01
1.92857683e-02 -3.28894347e-01 1.06225395e+00 1.46013469e-01
-6.52022302e-01 2.56852031e-01 2.78063208e-01 -1.46816909e-01
1.71066850e-01 -7.98747301e-01 -8.61417130e-02 7.05835521e-01
2.48806894e-01 5.50242364e-01 9.52127516e-01 -5.10869384e-01
-1.28040135e+00 -7.31579185e-01 6.47243619e-01 -3.02161098e-01
3.67857367e-01 -3.79932433e-01 -1.07084036e+00 4.70784038e-01
1.92676499e-01 -3.95053774e-01 4.70300466e-01 -3.30982834e-01
2.95142144e-01 -1.02630518e-01 -1.00648832e+00 7.61259198e-01
9.20934618e-01 -1.92331970e-01 -5.53380489e-01 6.52630534e-03
-3.19770463e-02 2.01848775e-01 -7.47584045e-01 3.30665499e-01
7.48108983e-01 -7.41296947e-01 4.93938863e-01 -1.58055648e-01
2.80022651e-01 -4.00116026e-01 7.73532093e-02 -1.49652946e+00
-2.25448698e-01 -1.46828547e-01 1.80683061e-01 1.21344793e+00
7.36365914e-01 -9.69784498e-01 7.17076421e-01 4.94624287e-01
-1.30404726e-01 -7.01206446e-01 -9.81858253e-01 -9.98631060e-01
5.39510548e-01 -8.55497085e-03 1.80982456e-01 5.88339150e-01
3.15618634e-01 6.41325712e-02 1.81758866e-01 2.00053886e-01
3.01681221e-01 -1.15353785e-01 4.63779122e-01 -1.28022826e+00
-2.01297164e-01 -5.83220184e-01 -7.32001066e-01 -5.27636588e-01
2.98196197e-01 -5.31534553e-01 3.72269154e-01 -1.64191091e+00
4.13490236e-02 -2.65980572e-01 -3.55936438e-01 1.49121448e-01
4.86130327e-01 4.13553536e-01 -1.20428167e-02 1.16316371e-01
-2.14010552e-01 8.56454745e-02 1.28161919e+00 2.54123420e-01
-5.74501902e-02 -8.57328102e-02 -4.63211328e-01 4.20722872e-01
8.95305574e-01 -4.69694346e-01 -3.54660660e-01 -2.45425731e-01
1.52145788e-01 -5.71340173e-02 5.07616878e-01 -1.79258525e+00
3.58538359e-01 1.66674718e-01 6.51868820e-01 -1.31714731e-01
7.37189800e-02 -7.88045466e-01 5.35844684e-01 5.89953125e-01
-3.29438955e-01 -2.77782440e-01 4.57537800e-01 3.49167079e-01
-2.24875554e-01 -6.23192608e-01 6.37221992e-01 -7.01908320e-02
-4.44170058e-01 -2.67689705e-01 -1.23218095e+00 -5.59637249e-01
1.15930927e+00 -4.95419830e-01 -1.54047102e-01 -2.29736358e-01
-7.61303484e-01 -6.25846386e-02 9.42198217e-01 -8.17389563e-02
4.64402467e-01 -1.08880067e+00 -4.53653604e-01 3.07574332e-01
-2.24415019e-01 -5.48603415e-01 -3.00739348e-01 6.59064174e-01
-8.28160048e-01 6.42283201e-01 -7.55299687e-01 -4.38186586e-01
-1.07814980e+00 5.11693120e-01 7.23060071e-01 2.53805906e-01
-2.82038897e-01 3.60607415e-01 3.83322015e-02 -2.41067683e-05
8.60935226e-02 -4.70737517e-01 -3.71443868e-01 -4.41744864e-01
4.50544417e-01 -1.55470341e-01 5.48794195e-02 -1.48306981e-01
-1.36796251e-01 7.06712365e-01 3.37907523e-01 -4.69343126e-01
1.49936962e+00 1.16084933e-01 8.90165344e-02 6.42805457e-01
8.64789784e-01 -2.59994209e-01 -1.53222764e+00 4.11408931e-01
-4.53407206e-02 -4.59223054e-02 -2.60244645e-02 -6.94827914e-01
-6.55439138e-01 6.41864896e-01 5.32693624e-01 4.48089331e-01
1.17171037e+00 -1.19699746e-01 3.68780822e-01 7.47123241e-01
8.49242628e-01 -8.53678763e-01 -5.51505610e-02 5.67789435e-01
8.99646163e-01 -5.39153934e-01 -4.65675086e-01 -4.36052024e-01
-4.32638228e-02 1.27723670e+00 4.15673822e-01 -5.50311208e-01
6.25307798e-01 8.00750554e-01 -2.99204662e-02 -1.57105729e-01
-9.13410962e-01 -5.28345741e-02 -1.30318180e-01 8.24338198e-01
2.85390794e-01 -4.10796493e-01 -4.91648167e-01 3.06485444e-01
-2.52873041e-02 8.02329898e-01 8.71455252e-01 9.58746612e-01
-7.94287682e-01 -1.12778401e+00 -2.56469965e-01 3.83779734e-01
-2.53452212e-01 2.91243374e-01 -7.21500933e-01 8.45493078e-01
-6.99715540e-02 6.76414251e-01 2.14249462e-01 -2.75287122e-01
3.92224431e-01 4.75528002e-01 1.12806892e+00 -5.50590992e-01
-3.98583412e-01 -3.46763954e-02 1.73434243e-01 1.35155857e-01
-5.12653589e-01 -7.45597959e-01 -1.26039767e+00 -2.41528936e-02
1.96709973e-03 1.97823063e-01 6.77782834e-01 1.20058119e+00
2.42926374e-01 4.95527565e-01 3.02177370e-01 -9.54604983e-01
1.57658517e-01 -8.01404953e-01 -9.00685549e-01 1.08611941e-01
3.17009576e-02 -7.63982177e-01 -5.14501810e-01 3.26721251e-01] | [8.033114433288574, 2.890683889389038] |
b4cfc41f-2c01-4746-b4e4-77197369123b | adjacent-level-feature-cross-fusion-with-3d | 2302.05109 | null | https://arxiv.org/abs/2302.05109v1 | https://arxiv.org/pdf/2302.05109v1.pdf | Adjacent-level Feature Cross-Fusion with 3D CNN for Remote Sensing Image Change Detection | Deep learning-based change detection using remote sensing images has received increasing attention in recent years. However, how to effectively extract and fuse the deep features of bi-temporal images to improve the accuracy of change detection is still a challenge. To address that, a novel adjacent-level feature fusion network with 3D convolution (named AFCF3D-Net) is proposed in this article. First, through the inner fusion property of 3D convolution, we design a new feature fusion way that can simultaneously extract and fuse the feature information from bi-temporal images. Then, in order to bridge the semantic gap between low-level features and high-level features, we propose an adjacent-level feature cross-fusion (AFCF) module to aggregate complementary feature information between the adjacent-levels. Furthermore, the densely skip connection strategy is introduced to improve the capability of pixel-wise prediction and compactness of changed objects in the results. Finally, the proposed AFCF3D-Net has been validated on the three challenging remote sensing change detection datasets: Wuhan building dataset (WHU-CD), LEVIR building dataset (LEVIR-CD), and Sun Yat-Sen University (SYSU-CD). The results of quantitative analysis and qualitative comparison demonstrate that the proposed AFCF3D-Net achieves better performance compared to the other state-of-the-art change detection methods. | ['Yao Qin', 'Jianwei Fan', 'Guangyang Lei', 'Liang Zhou', 'Mengmeng Wang', 'Yuanxin Ye'] | 2023-02-10 | null | null | null | null | ['change-detection'] | ['computer-vision'] | [ 4.56182212e-02 -8.64788473e-01 5.68574190e-01 -4.61696148e-01
-2.98651636e-01 -8.35240856e-02 5.83037019e-01 1.44429401e-01
-4.51468915e-01 5.09674728e-01 2.25270092e-01 -1.73963636e-01
-4.52432007e-01 -1.18596005e+00 -3.57129514e-01 -9.62464452e-01
-3.19888622e-01 -5.49437940e-01 3.34252238e-01 -4.47579443e-01
-1.59668103e-01 7.36908197e-01 -1.69544911e+00 7.82572702e-02
1.07603765e+00 1.35791087e+00 2.94924855e-01 2.70635217e-01
1.95066229e-01 3.01380694e-01 -2.90460717e-02 3.05115402e-01
6.52566671e-01 -3.16642225e-01 -5.49316704e-01 9.42922533e-02
4.74370867e-01 -6.07935309e-01 -5.06655812e-01 1.24034667e+00
8.01260948e-01 2.82342076e-01 3.01575243e-01 -8.11563969e-01
-7.52780020e-01 3.59517545e-01 -8.30673814e-01 8.57741594e-01
-2.03499198e-01 2.42548257e-01 7.24448204e-01 -1.05039752e+00
3.36093783e-01 1.30478954e+00 8.55630219e-01 -2.67735898e-01
-8.63634348e-01 -6.76574111e-01 5.33284605e-01 5.17403185e-01
-1.57102549e+00 -1.33621901e-01 8.18100274e-01 -5.70622861e-01
8.72872591e-01 2.07450181e-01 9.41018999e-01 1.06058903e-01
3.17385644e-01 6.81720853e-01 1.19845402e+00 -1.51214436e-01
-5.02644926e-02 -5.34931779e-01 1.37898460e-01 7.52601922e-01
3.01543117e-01 2.18298599e-01 -7.95041695e-02 2.91894853e-01
8.12721133e-01 6.22491002e-01 -5.84892213e-01 6.09727092e-02
-1.29759812e+00 7.43028283e-01 1.16199291e+00 6.93181753e-01
-5.55544138e-01 -2.35349685e-02 2.27829367e-01 3.17195714e-01
8.54187787e-01 -1.24517784e-01 -4.96401221e-01 3.23450983e-01
-9.51931536e-01 1.98006347e-01 -3.92875038e-02 4.15337235e-01
9.99984920e-01 4.10413230e-03 -2.94544697e-01 7.47365355e-01
4.23512459e-01 6.88386619e-01 4.76560533e-01 -3.83114845e-01
3.66450727e-01 8.84262025e-01 -6.09226078e-02 -1.42785835e+00
-5.85608482e-01 -6.91263258e-01 -1.28699446e+00 2.68587112e-01
-2.76872575e-01 -9.81635004e-02 -9.76972818e-01 1.37103951e+00
6.06255472e-01 1.66411281e-01 -6.45160750e-02 1.10331345e+00
1.08525848e+00 9.50735927e-01 -1.90480769e-01 -2.33143330e-01
1.20809865e+00 -7.54391193e-01 -6.25502110e-01 1.34083247e-02
2.54347891e-01 -4.34475213e-01 9.38216984e-01 -2.54176140e-01
-5.99539161e-01 -8.70230377e-01 -1.14311802e+00 1.28387762e-02
-5.85973740e-01 3.19328278e-01 5.68212926e-01 9.69443023e-02
-8.32547903e-01 4.05396491e-01 -9.00953710e-01 -4.39453512e-01
7.61744082e-01 1.70137614e-01 -4.46108520e-01 -3.33999902e-01
-1.30885363e+00 7.75641859e-01 5.05299747e-01 7.23909020e-01
-6.17869735e-01 -7.95381188e-01 -8.39057684e-01 1.68845892e-01
-6.16993606e-02 -4.45482254e-01 7.56927907e-01 -9.52542663e-01
-1.00041878e+00 5.06611049e-01 1.21861033e-01 -9.67710763e-02
3.67565662e-01 -9.06602591e-02 -6.79479778e-01 1.13859735e-01
1.71433408e-02 6.46850049e-01 6.13876522e-01 -1.02573895e+00
-1.00361919e+00 -5.81596851e-01 -1.20842585e-03 2.86850929e-01
-4.50051814e-01 -2.08436161e-01 -4.12783511e-02 -9.01137590e-01
5.75157821e-01 -4.88230109e-01 -1.70838341e-01 2.72104710e-01
1.03542529e-01 -1.65798992e-01 1.38089299e+00 -9.26811457e-01
1.37373829e+00 -2.34740329e+00 -1.72595888e-01 1.54826134e-01
2.04645753e-01 6.49128199e-01 -2.64184862e-01 1.14890620e-01
-1.44918010e-01 1.37467027e-01 -6.69249594e-01 1.59931570e-01
-1.08925350e-01 5.68836108e-02 -9.80163366e-03 7.28779554e-01
4.00670558e-01 9.02086914e-01 -7.64862239e-01 -3.29260170e-01
5.94539225e-01 5.66452324e-01 -3.54532599e-01 -2.04282761e-01
1.76739350e-01 4.29428250e-01 -5.22384286e-01 7.98181057e-01
1.20175231e+00 4.51077037e-02 -4.93059754e-01 -5.22613764e-01
-5.86095393e-01 -1.87637359e-01 -1.23896873e+00 1.39847624e+00
-2.89722919e-01 5.94414234e-01 -1.01933755e-01 -9.10630524e-01
1.04893601e+00 -1.49145916e-01 6.40874684e-01 -9.54482317e-01
1.71563163e-01 2.19285190e-01 2.57357769e-02 -5.80497205e-01
3.08071047e-01 -2.18828902e-01 4.98161577e-02 1.42380595e-01
-3.45429122e-01 1.46846533e-01 5.57305291e-02 -2.89767295e-01
9.19534922e-01 7.48198479e-02 3.06228638e-01 -4.16985542e-01
8.68526816e-01 -4.77252528e-02 8.99097979e-01 3.92045587e-01
-5.62282145e-01 3.63935888e-01 -3.44420612e-01 -8.47056627e-01
-6.57249808e-01 -7.68545508e-01 -3.76051217e-01 6.07149541e-01
4.95125055e-01 1.78376287e-01 -3.55991900e-01 -6.89902663e-01
2.48987868e-01 1.99302495e-01 -6.33201182e-01 -1.51965097e-01
-4.99312103e-01 -9.77466166e-01 2.50015318e-01 4.62546468e-01
1.52424896e+00 -8.99979770e-01 -6.47582710e-01 3.52722645e-01
-2.56977946e-01 -7.88123488e-01 -5.05171716e-01 -9.45931748e-02
-7.85563052e-01 -8.91263485e-01 -5.36489129e-01 -9.14600313e-01
5.71913421e-01 8.54269743e-01 5.96106291e-01 7.78806284e-02
-3.74877810e-01 -2.48846486e-02 -5.51746011e-01 -2.31310189e-01
2.22628623e-01 -8.50759521e-02 -1.96568415e-01 1.32057592e-01
3.75577122e-01 -7.75881588e-01 -9.38708782e-01 1.93660185e-01
-1.17695594e+00 7.32427612e-02 7.14587569e-01 6.95244014e-01
6.22067273e-01 7.15837538e-01 5.21518230e-01 -9.54662338e-02
2.63212204e-01 -3.78864437e-01 -6.28823400e-01 1.91064283e-01
-6.05300188e-01 -3.62874091e-01 3.37477475e-01 -1.49312526e-01
-1.25386477e+00 8.36263373e-02 -1.66991353e-01 -2.49444082e-01
-7.11234957e-02 9.27164018e-01 -3.69446903e-01 -1.27403751e-01
5.10860562e-01 4.25696194e-01 -2.82555908e-01 -6.95794821e-01
2.43629485e-01 8.87804270e-01 5.80081940e-01 6.59920126e-02
9.04115617e-01 6.45528674e-01 -1.13607973e-01 -6.52776837e-01
-8.60853434e-01 -5.07006347e-01 -8.54913473e-01 -3.02102894e-01
9.24217165e-01 -1.38102686e+00 -3.75252277e-01 1.18664289e+00
-8.40173244e-01 -1.09159783e-01 -3.38163614e-01 6.73092246e-01
-6.04356378e-02 2.19997540e-01 -3.37774307e-01 -4.79814470e-01
-7.55485117e-01 -9.00938988e-01 1.02527821e+00 5.64640343e-01
5.86539268e-01 -8.23983133e-01 2.34559830e-02 5.48597686e-02
7.49543667e-01 6.76237285e-01 7.24999666e-01 -1.23654336e-01
-4.93369222e-01 -8.15008115e-03 -6.05995059e-01 5.10137916e-01
6.91872895e-01 2.74973409e-03 -6.47820115e-01 -4.06898588e-01
-2.37864256e-02 1.44229308e-01 1.24567950e+00 4.75915670e-01
9.30313051e-01 -2.88092971e-01 -3.09217870e-01 7.54421055e-01
1.76629698e+00 1.76253840e-01 6.16579771e-01 5.96592069e-01
8.44592571e-01 1.08688101e-01 6.20061100e-01 6.28920615e-01
5.68293095e-01 3.98924470e-01 4.69707191e-01 -4.72672105e-01
-2.45582446e-01 1.88095272e-02 3.45087379e-01 7.11663604e-01
9.89142526e-03 1.48673117e-01 -7.94814408e-01 6.94747448e-01
-1.80010772e+00 -1.19089723e+00 -4.19212461e-01 1.62771010e+00
8.10785532e-01 -2.36727074e-01 -2.34644011e-01 2.91321605e-01
9.76119161e-01 4.11573350e-01 -7.57444322e-01 3.51627558e-01
-4.60802287e-01 1.28206551e-01 4.27236766e-01 2.80677944e-01
-1.64584792e+00 5.77803016e-01 5.00831032e+00 7.33002782e-01
-1.52636540e+00 3.39019299e-01 3.94540459e-01 4.00411934e-02
-1.88199013e-01 -1.97032705e-01 -7.17969954e-01 5.42468369e-01
2.21753210e-01 -3.41120660e-02 1.95315704e-01 4.05781269e-01
6.74738228e-01 -2.22412512e-01 -5.29917657e-01 9.22816038e-01
-1.23936292e-02 -1.22004116e+00 6.69289306e-02 -4.92108427e-02
1.12141907e+00 3.35751504e-01 -1.96664229e-01 2.80240089e-01
1.63802192e-01 -6.37648344e-01 6.51899576e-01 9.47470009e-01
5.95913768e-01 -7.37977803e-01 1.02460480e+00 3.19191009e-01
-1.67412901e+00 -4.23144639e-01 -3.71682167e-01 -1.46519393e-01
-3.16073224e-02 1.17125833e+00 -9.08196792e-02 1.03572834e+00
1.12106299e+00 1.27149785e+00 -8.09253812e-01 1.25707686e+00
5.61129339e-02 5.06583631e-01 -4.62224334e-01 3.78729999e-01
4.96204257e-01 -2.15489626e-01 3.93813372e-01 1.23138750e+00
6.86141074e-01 5.15618324e-01 3.00350636e-01 7.62278080e-01
7.23413751e-02 -2.31315326e-02 -3.08885723e-01 2.41188616e-01
3.33855093e-01 1.44180226e+00 -5.31179547e-01 -2.95387030e-01
-2.96407312e-01 7.53445148e-01 -1.41730262e-02 1.52154133e-01
-7.27196455e-01 -6.00721478e-01 8.47271681e-01 -1.09139197e-01
7.82772839e-01 -3.67015690e-01 -2.08874464e-01 -1.29755569e+00
2.05115661e-01 -6.31471813e-01 3.45178068e-01 -7.43111849e-01
-1.27005911e+00 4.27503109e-01 -3.29403169e-02 -1.51305413e+00
5.55227578e-01 -2.15889946e-01 -7.80748785e-01 1.10473084e+00
-2.08860517e+00 -1.38036048e+00 -9.57051039e-01 8.54421496e-01
3.54275972e-01 1.07326575e-01 3.75773042e-01 6.67111397e-01
-7.61395037e-01 1.63010851e-01 4.15673643e-01 2.78535843e-01
2.32826591e-01 -7.74437428e-01 2.86583245e-01 1.36991501e+00
-3.54292393e-01 1.58570871e-01 1.84200108e-01 -5.26290238e-01
-9.50702250e-01 -1.70710182e+00 6.38189852e-01 4.45911914e-01
3.17415923e-01 1.33614302e-01 -1.12991750e+00 3.31470460e-01
-3.44291329e-02 5.93150854e-01 3.36742401e-01 -5.20714462e-01
-1.65086269e-01 -7.63007581e-01 -1.40196598e+00 1.45111308e-01
1.13192022e+00 -3.49016488e-01 -5.42506516e-01 1.23636134e-01
8.01271021e-01 -1.71333343e-01 -9.78371859e-01 7.82667935e-01
3.91039282e-01 -9.22482848e-01 7.25945234e-01 -1.56597570e-02
3.58588725e-01 -1.00120664e+00 -4.95307028e-01 -1.41963387e+00
-8.74958932e-01 1.12668753e-01 2.39458084e-01 1.32034361e+00
-2.29906887e-01 -7.06171989e-01 -7.44059011e-02 -1.36575848e-01
-3.95355225e-01 -6.73086524e-01 -1.00017667e+00 -6.04456246e-01
6.39208173e-03 -1.86028317e-01 9.62520957e-01 1.11701071e+00
-5.37411451e-01 1.06881395e-01 2.11240561e-03 4.74690408e-01
2.64093518e-01 4.10341591e-01 4.02187347e-01 -1.50676417e+00
2.37873465e-01 -6.33103311e-01 -5.92402637e-01 -9.33894992e-01
-1.13265201e-01 -1.01533616e+00 1.35795146e-01 -2.02040553e+00
1.37746051e-01 -5.69364309e-01 -5.07637262e-01 8.73188615e-01
-4.21057761e-01 2.50133991e-01 1.44884720e-01 3.71847987e-01
-1.31657749e-01 1.15774655e+00 1.59722686e+00 -4.71956313e-01
-2.01366529e-01 -2.29943156e-01 -6.60610080e-01 3.92830044e-01
6.52357996e-01 -2.37551510e-01 -1.24420285e-01 -5.09828269e-01
-1.09547801e-01 -4.42307502e-01 6.48242950e-01 -1.31973803e+00
1.32295907e-01 -2.13433489e-01 7.09363163e-01 -1.08677852e+00
-1.94351584e-01 -9.30206299e-01 2.26891816e-01 6.51506007e-01
7.54749849e-02 2.09865207e-03 2.96048969e-01 4.72299099e-01
-4.78766382e-01 2.22501829e-01 9.80890334e-01 -9.06100050e-02
-1.10662675e+00 6.64023101e-01 -2.55525500e-01 -4.61678982e-01
1.10930049e+00 -1.25198528e-01 -4.34347391e-01 2.75170635e-02
-5.02750754e-01 4.28701401e-01 1.42067343e-01 3.87600869e-01
7.00670362e-01 -1.77592254e+00 -1.02325940e+00 3.79671782e-01
1.99171022e-01 2.69785762e-01 7.07827508e-01 9.81028080e-01
-5.75511217e-01 6.69608414e-02 -4.25734282e-01 -7.17434645e-01
-1.11992121e+00 3.04749608e-01 7.46439993e-01 -1.25753403e-01
-7.33721435e-01 6.08106375e-01 5.93726002e-02 -5.58672786e-01
-2.73341089e-01 -5.99114776e-01 -2.39512816e-01 2.23955899e-01
5.65540195e-01 4.03582662e-01 3.46469492e-01 -8.05430830e-01
-5.66594183e-01 6.58845663e-01 2.21622318e-01 2.63792694e-01
1.82272875e+00 -4.72919285e-01 -3.98523092e-01 2.14751109e-01
1.29862654e+00 -4.00565028e-01 -1.46615267e+00 -7.63739347e-01
-5.00582874e-01 -6.56261325e-01 6.09808445e-01 -8.48538637e-01
-1.45547211e+00 9.33675289e-01 1.43849635e+00 7.13654310e-02
1.65955925e+00 -3.33134085e-01 9.56855297e-01 3.76756072e-01
4.07895856e-02 -8.59215975e-01 -1.67451069e-01 6.22440875e-01
9.96096253e-01 -1.35512590e+00 2.31585458e-01 -9.98137668e-02
-8.84717181e-02 1.05271721e+00 6.16021037e-01 -1.82792097e-02
1.13448751e+00 -7.33679682e-02 6.24097558e-03 -3.30127925e-01
-2.17405528e-01 -6.22937500e-01 3.95148546e-01 3.25299084e-01
7.68721625e-02 1.81905672e-01 -3.46115202e-01 4.72288728e-01
1.02177873e-01 6.70661554e-02 1.00901455e-01 1.09572077e+00
-8.64755869e-01 -5.40086865e-01 -4.91490275e-01 6.00952446e-01
-1.08382314e-01 -2.43013486e-01 4.71455418e-02 6.58046544e-01
8.26890171e-01 9.41145599e-01 3.20722699e-01 -6.18914843e-01
3.76864940e-01 -3.93779129e-01 1.64906934e-01 -2.24093452e-01
-6.63503706e-01 6.10313006e-02 -4.24446315e-01 -1.78453475e-01
-1.03421509e+00 -8.30377638e-01 -1.30534935e+00 -2.70180047e-01
-4.46013033e-01 -1.08482942e-01 4.21118468e-01 1.06135249e+00
5.20187080e-01 5.59206605e-01 1.05328870e+00 -9.84104156e-01
-2.94061303e-01 -1.25292587e+00 -7.34109461e-01 1.10971034e-01
5.82661748e-01 -7.51338959e-01 -2.94991285e-01 1.64312452e-01] | [9.75543212890625, -1.323255181312561] |
af957e55-c45e-43e0-845a-7c596f8bfef5 | automatic-evaluation-of-turn-taking-cues-in | 2305.17971 | null | https://arxiv.org/abs/2305.17971v1 | https://arxiv.org/pdf/2305.17971v1.pdf | Automatic Evaluation of Turn-taking Cues in Conversational Speech Synthesis | Turn-taking is a fundamental aspect of human communication where speakers convey their intention to either hold, or yield, their turn through prosodic cues. Using the recently proposed Voice Activity Projection model, we propose an automatic evaluation approach to measure these aspects for conversational speech synthesis. We investigate the ability of three commercial, and two open-source, Text-To-Speech (TTS) systems ability to generate turn-taking cues over simulated turns. By varying the stimuli, or controlling the prosody, we analyze the models performances. We show that while commercial TTS largely provide appropriate cues, they often produce ambiguous signals, and that further improvements are possible. TTS, trained on read or spontaneous speech, produce strong turn-hold but weak turn-yield cues. We argue that this approach, that focus on functional aspects of interaction, provides a useful addition to other important speech metrics, such as intelligibility and naturalness. | ['Gabriel Skantze', 'Joakim Gustafson', 'Éva Székely', 'Siyang Wang', 'Erik Ekstedt'] | 2023-05-29 | null | null | null | null | ['speech-synthesis'] | ['speech'] | [ 1.94893375e-01 4.31890517e-01 2.65854765e-02 -4.91279781e-01
-9.03823674e-01 -7.43365645e-01 8.20439816e-01 -2.09740713e-01
1.51165977e-01 7.77954578e-01 8.15161586e-01 -4.76113647e-01
1.70614854e-01 -3.85151863e-01 -2.11067259e-01 -4.43394393e-01
2.85364985e-01 2.64429361e-01 2.77195662e-01 -7.21952856e-01
1.28308907e-01 5.35920680e-01 -1.75768185e+00 6.29736841e-01
6.03319764e-01 7.71062374e-01 3.05857539e-01 1.13349128e+00
-2.04653725e-01 7.83085644e-01 -1.08833981e+00 -9.96933803e-02
-1.49180651e-01 -6.93433225e-01 -8.18466365e-01 -6.83749095e-02
3.25884484e-02 -8.49562660e-02 1.38012141e-01 5.95240295e-01
7.38399208e-01 8.01792890e-02 5.05511165e-01 -1.08806908e+00
-2.09995642e-01 1.10438073e+00 3.67111713e-01 -4.49663438e-02
1.21331620e+00 6.07238352e-01 1.09363985e+00 -5.35114527e-01
6.10134244e-01 1.53249896e+00 3.59571934e-01 8.50466192e-01
-1.44389045e+00 -3.70929271e-01 -1.29259095e-01 -4.32565540e-01
-8.68171334e-01 -1.14838791e+00 7.84548998e-01 -4.59999979e-01
1.00140607e+00 1.02184856e+00 5.33023238e-01 1.50591254e+00
-1.57640785e-01 8.66544962e-01 1.18213332e+00 -7.02878058e-01
3.75939049e-02 4.65079606e-01 1.03721723e-01 1.22080140e-01
-6.81750119e-01 3.87697041e-01 -7.92869151e-01 -6.19309805e-02
3.59360486e-01 -9.13756490e-01 -7.61130869e-01 2.54383028e-01
-1.26913714e+00 4.87695187e-01 -3.32071111e-02 6.39627457e-01
-4.25198913e-01 -2.40591586e-01 4.57740217e-01 5.59437335e-01
2.03913718e-01 7.52622783e-01 -3.54196131e-01 -9.03245986e-01
-8.07654023e-01 5.05680680e-01 1.26826143e+00 1.13634324e+00
2.29089424e-01 1.68743968e-01 -7.97064900e-01 1.02940798e+00
1.93239525e-01 6.23623371e-01 3.84419143e-01 -1.02150273e+00
5.21606624e-01 1.35342121e-01 3.48453850e-01 -8.41781974e-01
-4.54813808e-01 -3.77972028e-03 -1.39125586e-01 1.43632069e-01
5.77620804e-01 -4.18470800e-01 -4.80646700e-01 1.90248942e+00
1.16717592e-01 -7.29736924e-01 3.55613887e-01 7.43519545e-01
9.07356501e-01 8.37804019e-01 -2.26888716e-01 -7.35641301e-01
1.31316257e+00 -9.08127010e-01 -1.34849513e+00 5.15388139e-03
4.58959311e-01 -1.24501836e+00 1.78414321e+00 5.56722224e-01
-1.36378074e+00 -6.12018228e-01 -9.82588351e-01 1.81902543e-01
-1.04461096e-01 9.26645547e-02 2.61918604e-01 1.02494299e+00
-1.11881196e+00 4.62285936e-01 -3.67126137e-01 -1.76288188e-01
-6.82428420e-01 9.91838947e-02 -2.03827515e-01 7.64651656e-01
-1.15718162e+00 7.97209680e-01 -1.41191110e-01 -1.51155353e-01
-4.32339042e-01 -3.33409339e-01 -7.74147391e-01 6.03939500e-03
2.09249437e-01 -1.48043290e-01 1.98834920e+00 -8.45026731e-01
-2.50272107e+00 6.15696788e-01 -1.79062590e-01 -3.01702529e-01
5.67713916e-01 -3.26314628e-01 -6.36358798e-01 3.02422255e-01
-2.00283378e-01 8.19122970e-01 7.19176412e-01 -1.37925017e+00
-3.08925450e-01 2.13528916e-01 -7.43359253e-02 1.22665562e-01
-1.03251241e-01 4.60197330e-01 -1.24481149e-01 -4.86712962e-01
-1.13483138e-01 -1.14271617e+00 2.43395194e-01 -3.80463302e-01
-9.94361460e-01 -1.01244308e-01 8.13147128e-01 -6.48740530e-01
1.55088496e+00 -2.09627533e+00 2.24929936e-02 8.22374299e-02
-3.46977562e-01 3.76462996e-01 3.95315932e-03 7.96219707e-01
3.95758748e-02 2.05032766e-01 -1.62654743e-01 -4.09339994e-01
2.64693052e-01 -1.05159087e-02 -4.72445101e-01 -1.75015569e-01
1.02389477e-01 6.98529005e-01 -6.35772228e-01 -1.99335515e-01
4.41868126e-01 3.80718768e-01 -5.83518267e-01 5.41790962e-01
-3.04975927e-01 7.43224084e-01 -5.91976047e-02 3.63942176e-01
2.47972384e-01 5.49967468e-01 2.00995579e-01 4.42040190e-02
-5.58256090e-01 1.03638494e+00 -8.89679015e-01 1.23499167e+00
-7.97003388e-01 9.93365705e-01 2.97899425e-01 -1.90722123e-01
1.00237417e+00 7.78151333e-01 -3.00335158e-02 -6.03746831e-01
9.99222621e-02 3.74213517e-01 5.42607069e-01 -7.00697839e-01
6.29138768e-01 -1.62302002e-01 -2.82001793e-01 3.44182938e-01
-1.91248000e-01 -9.32998002e-01 1.00538999e-01 -1.64724126e-01
7.24449396e-01 -4.81417887e-02 3.38081121e-01 -3.35441232e-01
6.46630228e-01 -3.89021933e-01 -2.82935202e-02 5.22955060e-01
-2.28121892e-01 8.69250298e-01 7.83547521e-01 1.40489906e-01
-7.86951065e-01 -8.97428691e-01 9.55091789e-02 1.12287295e+00
-3.12739551e-01 -4.63620216e-01 -1.28416431e+00 -2.16292888e-01
-4.83114004e-01 1.36280704e+00 -1.75532311e-01 1.60946127e-03
-6.93745136e-01 -6.85359463e-02 7.73265243e-01 1.45921662e-01
-3.29653546e-02 -1.83724701e+00 -5.58180451e-01 4.83878195e-01
-6.46420419e-01 -1.14257169e+00 -8.11084569e-01 2.14082435e-01
-3.19257855e-01 -6.25674248e-01 -6.93932652e-01 -4.09580261e-01
-5.49579374e-02 -3.43412086e-02 1.08995676e+00 -3.68838236e-02
2.89708495e-01 3.29349875e-01 -6.00551009e-01 -3.40900719e-01
-1.33479881e+00 -4.92494181e-02 1.40374422e-01 -4.33065407e-02
-6.77104443e-02 -6.23493314e-01 -3.09973747e-01 7.21153438e-01
-6.01711333e-01 1.00182258e-01 2.25599632e-01 5.51182508e-01
-1.50071904e-01 -6.63631618e-01 7.06875861e-01 -4.92362887e-01
1.36827874e+00 2.49409467e-01 -6.49939030e-02 -3.17770801e-02
-1.54077739e-01 2.80669630e-01 5.78102052e-01 -4.35926914e-01
-1.17502463e+00 -1.80816099e-01 -8.14373553e-01 2.57849898e-02
-5.47697842e-01 1.93914428e-01 -6.11677587e-01 1.60061046e-01
8.79742801e-01 1.63050339e-01 2.96543479e-01 -2.62632877e-01
4.11901176e-01 1.23445261e+00 4.20435369e-01 -6.40415728e-01
2.83432573e-01 -1.15933985e-01 -5.72785914e-01 -1.39696705e+00
-3.95641834e-01 -3.23888808e-01 -2.39998609e-01 -3.25646728e-01
6.97642326e-01 -4.12543803e-01 -8.34632456e-01 3.54596674e-01
-1.47799325e+00 -6.27568305e-01 -2.86456764e-01 4.18690741e-01
-9.76560354e-01 2.01631173e-01 -6.33014083e-01 -1.28369308e+00
-3.07287306e-01 -1.50763726e+00 1.28981483e+00 9.69131105e-03
-1.03937912e+00 -5.62871277e-01 4.70806770e-02 4.77792263e-01
4.67405140e-01 -2.17693783e-02 7.10040212e-01 -7.51569211e-01
7.95258358e-02 6.24374710e-02 2.99983740e-01 5.65633357e-01
3.96165252e-01 4.21493560e-01 -1.26560247e+00 1.82655081e-01
-5.59397601e-02 -4.03781205e-01 2.00977325e-01 4.42116618e-01
4.70468670e-01 -6.38013959e-01 -1.67664699e-02 -2.74501201e-02
4.55797136e-01 6.61412060e-01 8.58658493e-01 -3.35751951e-01
2.30897248e-01 1.16941357e+00 6.62699699e-01 3.32910717e-01
3.45403925e-02 1.20016658e+00 5.74750789e-02 1.57843202e-01
-4.61512327e-01 -3.01150858e-01 8.01533699e-01 1.23497474e+00
-8.15138668e-02 -6.64170325e-01 -6.24427378e-01 3.09451222e-01
-1.34256160e+00 -1.08486974e+00 -4.15526927e-01 2.12926555e+00
1.17100036e+00 5.30760765e-01 4.26903307e-01 6.18634641e-01
7.16524124e-01 2.61199117e-01 1.13505527e-01 -1.26183224e+00
-1.37742385e-01 1.25186637e-01 -1.53764188e-01 1.10575998e+00
-5.71591675e-01 1.09976923e+00 7.25143671e+00 7.83242285e-01
-1.48531270e+00 -1.92634746e-01 3.80557120e-01 7.40286633e-02
-4.85452563e-01 -4.16602194e-01 -6.68794513e-01 4.43287760e-01
1.29978383e+00 5.63838184e-02 4.67587382e-01 6.30799592e-01
9.70372438e-01 -6.16453215e-02 -1.32881320e+00 7.07603693e-01
-1.15169823e-01 -8.95022154e-01 -1.35639995e-01 -2.54543982e-02
2.64532804e-01 -6.20079875e-01 1.70498770e-02 3.18783045e-01
-9.83982906e-02 -9.36315894e-01 1.13576543e+00 5.16282558e-01
7.54866540e-01 -5.05798817e-01 3.66133571e-01 3.28563720e-01
-9.90182996e-01 2.14204788e-01 4.21154261e-01 -3.31156075e-01
5.25134206e-01 3.02617192e-01 -1.22661400e+00 2.26178318e-01
1.21602476e-01 -1.47843346e-01 -3.34595963e-02 5.89009285e-01
-5.61880469e-01 9.88941014e-01 -3.32561463e-01 -6.18923366e-01
1.22743122e-01 5.57255633e-02 1.10677397e+00 1.47425675e+00
4.09613043e-01 -8.95711556e-02 -4.12976965e-02 8.13599348e-01
3.54718596e-01 3.67888331e-01 -7.13920176e-01 -1.75452739e-01
6.42177403e-01 9.45173860e-01 -4.08058494e-01 -1.59601718e-01
5.17081246e-02 9.65409338e-01 -3.22996736e-01 2.70150512e-01
-6.19218230e-01 -3.40321511e-01 7.00197935e-01 1.05676755e-01
-7.82122687e-02 -1.42044678e-01 -1.51227087e-01 -4.88139898e-01
2.81540900e-01 -1.29133844e+00 -4.88148808e-01 -9.86350954e-01
-5.99432230e-01 1.07199252e+00 9.53384116e-02 -1.27687478e+00
-9.92277741e-01 -5.46044052e-01 -8.09160709e-01 7.83576131e-01
-8.29859555e-01 -8.74419749e-01 3.57721783e-02 1.31594405e-01
9.62170243e-01 -3.10769752e-02 9.08282399e-01 -3.18601243e-02
-1.71273112e-01 7.01798797e-01 -4.02783543e-01 -2.15905666e-01
5.86390495e-01 -1.33241785e+00 4.86148864e-01 4.01723146e-01
2.34369099e-01 5.56818902e-01 1.45627427e+00 -1.02234520e-01
-8.05615962e-01 -3.59852701e-01 1.09728575e+00 -3.29194605e-01
4.64919031e-01 -6.63422287e-01 -6.88623726e-01 2.48217672e-01
5.84752619e-01 -5.16869605e-01 6.63154662e-01 1.39800847e-01
1.13488883e-01 1.56919330e-01 -1.05673063e+00 8.61372113e-01
9.50088859e-01 -6.29797101e-01 -8.77044499e-01 3.29563841e-02
1.16264892e+00 -3.96102995e-01 -4.75586593e-01 1.36204869e-01
7.03439295e-01 -1.51470780e+00 6.78462744e-01 -3.55590522e-01
3.61981452e-01 1.12905033e-01 -1.96716100e-01 -1.76008987e+00
2.35722408e-01 -1.49441457e+00 2.66080558e-01 1.48209238e+00
9.42472696e-01 -5.73176801e-01 3.16855967e-01 4.20968175e-01
-5.96213222e-01 -3.02358568e-01 -9.08809125e-01 -7.52553165e-01
-5.54721057e-03 -6.97162271e-01 6.38884425e-01 4.58862960e-01
6.78122342e-01 6.47732258e-01 -5.41858315e-01 -4.70889919e-02
-2.77378321e-01 -2.16168359e-01 9.54089105e-01 -9.62846816e-01
-2.33162954e-01 -6.46826565e-01 -2.89474763e-02 -1.21821165e+00
-2.94790603e-02 -2.29321271e-01 6.56635523e-01 -1.02119935e+00
-7.29684472e-01 -1.79247081e-01 2.70405799e-01 2.00566977e-01
1.62564777e-02 -2.72052020e-01 4.67620760e-01 -1.46746308e-01
5.70661239e-02 6.34421706e-01 1.44543290e+00 5.58821782e-02
-7.81061411e-01 4.60243434e-01 -4.75267559e-01 6.95546627e-01
7.63952911e-01 -9.46337059e-02 -4.50389117e-01 1.51174501e-01
-1.24326415e-01 7.01966465e-01 1.09811105e-01 -1.17874503e+00
2.34931875e-02 -1.79702759e-01 -4.61859375e-01 -4.38786179e-01
6.98727787e-01 -5.22056520e-01 -2.07196157e-02 3.47842991e-01
-7.87758768e-01 -1.58725455e-01 2.96025097e-01 1.80425961e-02
-3.48405570e-01 -2.24872857e-01 5.74306548e-01 6.06458932e-02
4.95852493e-02 -5.29411614e-01 -1.01134825e+00 -6.37881458e-02
6.61334872e-01 -2.87669897e-01 -9.99060087e-03 -9.26338136e-01
-9.04135942e-01 -2.13824078e-01 -2.31290404e-02 6.51891887e-01
4.00823176e-01 -1.13538563e+00 -5.62813401e-01 1.81951791e-01
1.45376056e-01 -5.81373096e-01 -1.21426277e-01 8.08632135e-01
-4.66964900e-01 7.61293650e-01 -3.04020122e-02 -7.22453594e-01
-1.42246079e+00 1.98544651e-01 3.26102704e-01 1.98191125e-02
-4.02309090e-01 6.68475926e-01 6.55444805e-03 -6.10265911e-01
2.61032432e-01 -8.74305487e-01 -3.01824033e-01 1.45440936e-01
5.22977531e-01 2.33510226e-01 1.36274517e-01 -6.87279999e-01
-2.01439157e-01 7.80084282e-02 3.67680460e-01 -9.66336310e-01
5.33486009e-01 -2.32269019e-01 2.98332363e-01 8.23720396e-01
9.03504014e-01 8.51623535e-01 -8.93990934e-01 3.39577049e-01
2.91014388e-02 -1.78280473e-01 -2.48276144e-01 -1.01239002e+00
-3.72485936e-01 9.56454873e-01 1.50056884e-01 9.95287359e-01
7.97903657e-01 -7.36128837e-02 8.01413894e-01 4.91633266e-01
2.47955322e-01 -1.16012645e+00 4.10425551e-02 5.98160505e-01
1.52380335e+00 -1.03209054e+00 -7.34706044e-01 -9.11572456e-01
-8.74660194e-01 1.06742549e+00 4.98426944e-01 4.76245522e-01
2.72377253e-01 6.01392925e-01 4.20387834e-01 1.46333084e-01
-9.99231994e-01 -2.31339902e-01 2.08588570e-01 7.64531851e-01
9.15646970e-01 4.54270363e-01 -6.03362381e-01 5.31278193e-01
-1.08664799e+00 -3.29191804e-01 4.98256147e-01 5.34675181e-01
-6.04946494e-01 -1.18521082e+00 -6.07896149e-01 -2.75880797e-03
-3.23576003e-01 -2.22319722e-01 -9.83567297e-01 7.43445277e-01
-2.73019582e-01 1.59552538e+00 -1.53325528e-01 -5.11712015e-01
7.12001085e-01 3.47873658e-01 1.83301151e-01 -6.85981333e-01
-1.00683379e+00 2.65642136e-01 8.84455502e-01 -3.74649227e-01
-2.68090278e-01 -6.66060150e-01 -1.11357450e+00 -3.44282389e-02
-4.32545215e-01 4.50094134e-01 6.03957415e-01 1.00878835e+00
1.72430053e-01 7.26598501e-01 8.99565101e-01 -9.16757703e-01
-5.26905775e-01 -1.55288291e+00 -2.68410355e-01 1.52335539e-01
3.96431088e-01 -4.36969787e-01 -7.12875187e-01 -2.12834850e-02] | [14.748151779174805, 6.6167778968811035] |
0a9d8279-e42c-4c62-a4af-68f477ca43df | range-nullspace-video-frame-interpolation | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Yu_Range-Nullspace_Video_Frame_Interpolation_With_Focalized_Motion_Estimation_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Yu_Range-Nullspace_Video_Frame_Interpolation_With_Focalized_Motion_Estimation_CVPR_2023_paper.pdf | Range-Nullspace Video Frame Interpolation With Focalized Motion Estimation | Continuous-time video frame interpolation is a fundamental technique in computer vision for its flexibility in synthesizing motion trajectories and novel video frames at arbitrary intermediate time steps. Yet, how to infer accurate intermediate motion and synthesize high-quality video frames are two critical challenges. In this paper, we present a novel VFI framework with improved treatment for these challenges. To address the former, we propose focalized trajectory fitting, which performs confidence-aware motion trajectory estimation by learning to pay focus to reliable optical flow candidates while suppressing the outliers. The second is range-nullspace synthesis, a novel frame renderer cast as solving an ill-posed problem addressed by learning decoupled components in orthogonal subspaces. The proposed framework sets new records on 7 of 10 public VFI benchmarks. | ['Shunqing Ren', 'Jimmy S. Ren', 'Xijun Chen', 'Dongqing Zou', 'Yu Zhang', 'ZHIYANG YU'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['video-frame-interpolation'] | ['computer-vision'] | [ 2.84181535e-01 -4.52698141e-01 -2.67871141e-01 -1.48553913e-02
-8.13509226e-01 -5.99703968e-01 4.74794894e-01 -7.52590001e-01
-2.43479177e-01 9.50975120e-01 3.05119276e-01 -1.77982047e-01
-2.21910805e-01 -2.00509921e-01 -8.93380404e-01 -7.44199872e-01
-5.20667844e-02 8.13438371e-02 1.75004035e-01 2.11470798e-01
5.38766384e-01 6.84726000e-01 -1.48945892e+00 1.68513492e-01
9.88151371e-01 8.32283914e-01 1.01222001e-01 1.12650144e+00
1.44167006e-01 9.84022856e-01 -3.19970191e-01 -1.30510256e-01
5.03498256e-01 -3.97607297e-01 -5.50288379e-01 3.71198118e-01
1.05704951e+00 -7.34374225e-01 -4.28615659e-01 9.89490509e-01
1.62884191e-01 5.91132998e-01 5.55771172e-01 -1.52074087e+00
-3.40254873e-01 1.60090595e-01 -5.32414079e-01 3.29934567e-01
6.73648119e-01 4.61067408e-01 6.91722870e-01 -1.20542622e+00
9.74872589e-01 1.28318357e+00 4.34954554e-01 5.65689266e-01
-1.17085898e+00 -4.73748773e-01 1.87947392e-01 4.25058514e-01
-1.19792318e+00 -7.72734582e-01 7.17819929e-01 -8.42636704e-01
6.51375830e-01 4.78347689e-01 5.70298016e-01 1.07064879e+00
1.95269659e-01 8.22691202e-01 6.02657795e-01 -2.04750989e-02
1.04209997e-01 -3.87177080e-01 -7.13352188e-02 5.72320223e-01
1.73101947e-01 3.15724462e-01 -6.61862731e-01 -2.10555077e-01
1.02876508e+00 -2.35639978e-02 -7.54846871e-01 -3.34777236e-01
-1.76595223e+00 5.41688085e-01 -7.05556348e-02 -2.88289577e-01
-3.49375635e-01 2.36961514e-01 1.65492252e-01 1.11160658e-01
2.67995656e-01 3.36007565e-01 -1.80021107e-01 -2.81131387e-01
-1.32229185e+00 6.95504665e-01 5.15856206e-01 1.15211475e+00
7.19205737e-01 5.60404837e-01 -4.58547443e-01 1.70017421e-01
3.50381076e-01 6.18451536e-01 9.80611593e-02 -1.68274343e+00
5.86613297e-01 1.26357907e-02 5.40103734e-01 -1.12505627e+00
-1.73780799e-01 -1.55888364e-01 -7.54705071e-01 3.50646704e-01
6.68064296e-01 -1.24584161e-01 -7.45402336e-01 1.47447753e+00
5.21816850e-01 1.28962088e+00 -9.68503356e-02 1.20516205e+00
5.68026364e-01 9.00932491e-01 -1.65255591e-01 -6.01871908e-01
8.52775395e-01 -1.13193941e+00 -8.22402060e-01 5.81236370e-02
9.57129821e-02 -1.09390640e+00 6.47582293e-01 5.48139691e-01
-1.27461743e+00 -7.50580430e-01 -9.39035892e-01 -3.42639416e-01
3.90579224e-01 1.50969386e-01 3.31266463e-01 2.83533007e-01
-1.01117837e+00 5.91800213e-01 -7.82006800e-01 1.87251166e-01
3.54122579e-01 2.48060271e-01 -3.62979203e-01 -1.21101432e-01
-7.42237747e-01 3.72895837e-01 4.18873876e-01 2.39019945e-01
-8.98526251e-01 -9.92370486e-01 -7.89671361e-01 -2.46668532e-01
5.27234554e-01 -1.08437026e+00 9.53048825e-01 -9.27650869e-01
-1.70673490e+00 3.72761339e-01 -6.96051121e-01 -4.25810069e-01
9.40435886e-01 -5.04539371e-01 -2.68046379e-01 1.97359517e-01
1.65525943e-01 6.64602101e-01 1.53606510e+00 -1.19731963e+00
-9.74894762e-01 1.60555672e-02 -1.23810470e-01 2.42557570e-01
2.90999025e-01 -1.62771896e-01 -5.69270194e-01 -8.80719662e-01
2.15184689e-01 -1.13601589e+00 -3.26500952e-01 2.58134872e-01
-2.59455144e-01 8.14097151e-02 9.79352653e-01 -8.51584375e-01
1.17778444e+00 -2.02387547e+00 5.78376472e-01 1.53014725e-02
3.46059531e-01 4.19891447e-01 2.88394149e-02 -1.46735087e-01
8.11296627e-02 -3.13272417e-01 -1.55071199e-01 -3.04663450e-01
-4.45355356e-01 7.81465396e-02 -6.69505298e-01 5.62085271e-01
2.57017583e-01 6.93972468e-01 -1.22485769e+00 -4.57142621e-01
6.12528026e-01 5.71681917e-01 -8.14363480e-01 3.28451842e-01
-2.17969969e-01 1.13847208e+00 -3.86455864e-01 5.70805073e-01
8.89719427e-01 2.24953201e-02 -2.92919546e-01 -4.98294622e-01
-3.60740811e-01 -2.24019602e-01 -1.60923159e+00 1.95278132e+00
-4.49427143e-02 8.78710270e-01 -4.16650325e-02 -6.10420525e-01
5.77002347e-01 2.97880262e-01 8.83509159e-01 -7.36454502e-02
-2.13803109e-02 1.49458006e-01 -3.11645061e-01 -7.02389181e-01
7.27414787e-01 1.98391885e-01 5.06034255e-01 -7.94476941e-02
-6.06496632e-02 9.47015509e-02 2.71806121e-01 1.14125416e-01
7.46152043e-01 7.23076224e-01 2.54113555e-01 -1.46843240e-01
9.56917405e-01 5.56865036e-02 1.05966032e+00 5.71825802e-01
-4.06783819e-01 1.07866859e+00 2.46258557e-01 -6.79625034e-01
-1.11532497e+00 -1.17679334e+00 1.86406389e-01 4.34738398e-01
2.29646087e-01 -3.58507395e-01 -5.22682905e-01 -3.86021674e-01
-1.43940508e-01 5.88533103e-01 -2.33703509e-01 1.48596242e-01
-1.21151090e+00 -2.60268569e-01 1.26512587e-01 3.71739596e-01
2.88844168e-01 -7.10081935e-01 -6.58374190e-01 3.59478116e-01
-6.02835834e-01 -1.38724935e+00 -9.16153848e-01 -5.97708642e-01
-8.01829696e-01 -1.10761321e+00 -8.81215632e-01 -5.92072308e-01
5.36853373e-01 6.63285255e-01 1.11731744e+00 4.17753268e-04
-2.19767064e-01 2.35206485e-01 -1.17180586e-01 1.07731543e-01
-3.86329234e-01 -3.45840096e-01 2.37565771e-01 4.35404211e-01
-5.88221923e-02 -3.78004134e-01 -7.65918553e-01 4.33455229e-01
-8.04344058e-01 3.04243267e-01 -6.53608469e-04 7.91826427e-01
8.35935116e-01 -4.13074732e-01 1.92742854e-01 -5.49441397e-01
2.35605821e-01 -5.09060144e-01 -9.02702212e-01 2.07899481e-01
-1.62380561e-01 3.09089795e-02 8.15129936e-01 -4.53062147e-01
-1.22019374e+00 3.78927946e-01 -3.57805714e-02 -1.06506383e+00
3.06727551e-03 -9.52283740e-02 -1.35703430e-01 -2.00854838e-01
4.96173948e-01 2.13681266e-01 -1.47916496e-01 -1.43857062e-01
5.51399231e-01 1.55507952e-01 1.04856312e+00 -7.97915280e-01
1.01025867e+00 8.12879980e-01 2.72610486e-01 -1.00306618e+00
-5.79479694e-01 -7.55391777e-01 -7.61654794e-01 -4.85152811e-01
9.72780347e-01 -9.85402167e-01 -7.47672617e-01 3.82798642e-01
-1.33304298e+00 -7.86508098e-02 -1.79181099e-01 6.54712379e-01
-9.73016441e-01 7.71657407e-01 -3.55834574e-01 -5.50201178e-01
-2.13014349e-01 -1.55834103e+00 1.01506341e+00 2.61064142e-01
-1.41704291e-01 -8.04970920e-01 2.93879747e-01 2.78604746e-01
-7.91795272e-03 6.15750670e-01 2.34028786e-01 2.58532226e-01
-1.21693122e+00 2.62140483e-01 -1.38774022e-01 2.36238882e-01
1.81629911e-01 5.71281612e-01 -8.20383191e-01 -4.55230713e-01
7.49174431e-02 2.48447925e-01 7.98938572e-01 7.12185383e-01
9.88290489e-01 -3.04930478e-01 -2.24544972e-01 1.39508486e+00
1.43635094e+00 2.76440352e-01 6.36992633e-01 1.76317140e-01
1.20130670e+00 3.48486334e-01 9.19326365e-01 5.92820525e-01
2.32839078e-01 7.95851886e-01 2.68776804e-01 3.70587587e-01
-2.54565895e-01 -2.47201473e-01 4.39603806e-01 6.38289094e-01
-5.07495999e-01 -2.03550145e-01 -7.10071087e-01 4.70520169e-01
-2.22966051e+00 -1.32554305e+00 -4.33310419e-01 2.19563198e+00
3.88245016e-01 -1.09608449e-01 1.31826952e-01 8.10071826e-02
5.64521611e-01 2.23921269e-01 -6.38481438e-01 1.25074953e-01
-1.99742362e-01 -1.17309771e-01 4.66901720e-01 9.08569455e-01
-1.10834086e+00 9.35032308e-01 5.69903088e+00 6.88475072e-01
-1.16020334e+00 -8.49299356e-02 4.34686273e-01 -2.38210171e-01
-3.30580652e-01 2.12941021e-01 -9.11324024e-01 5.11045277e-01
6.94683671e-01 -2.36270100e-01 6.30148947e-01 4.77398813e-01
7.08703518e-01 6.22968748e-03 -1.16899645e+00 1.36728156e+00
1.19252823e-01 -1.86407554e+00 2.71166086e-01 -2.23810375e-01
1.09809649e+00 -3.18041146e-01 7.83698708e-02 -1.86124772e-01
-1.03759415e-01 -8.64633858e-01 9.60791945e-01 8.62363100e-01
7.84263611e-01 -6.35663748e-01 6.39438257e-02 2.00887620e-01
-1.31816328e+00 2.97774523e-02 -2.99379855e-01 1.50973233e-03
6.10485435e-01 3.97858560e-01 -5.30408621e-01 7.08850980e-01
4.90299374e-01 1.14333916e+00 -2.14572549e-01 1.35869098e+00
3.67573649e-02 5.08234322e-01 -1.58108696e-01 6.71006739e-01
2.38208964e-01 -6.65897846e-01 9.93691266e-01 1.08539867e+00
6.30644798e-01 1.87629029e-01 3.80696476e-01 8.52958262e-01
3.14927399e-01 -1.98887527e-01 -6.10033393e-01 4.06479031e-01
3.35798681e-01 1.06164944e+00 -5.67407250e-01 -3.53947163e-01
-5.91382265e-01 1.06651914e+00 -6.02910388e-03 7.64642298e-01
-1.02833426e+00 7.01416805e-02 1.01239598e+00 1.37430549e-01
8.58367234e-02 -6.13276005e-01 -8.00769851e-02 -1.71899319e+00
1.90796673e-01 -8.84016514e-01 1.83495000e-01 -7.10266054e-01
-8.42878222e-01 4.24383730e-01 -1.62887406e-02 -1.79609275e+00
-5.96281528e-01 -5.70024788e-01 -5.37305117e-01 7.78273106e-01
-1.70187700e+00 -9.69553113e-01 -5.43758512e-01 9.90834415e-01
1.05536973e+00 -1.21755652e-01 1.87174648e-01 5.10781765e-01
-4.16404635e-01 2.41140768e-01 2.16121241e-01 -1.03139892e-01
8.31840932e-01 -9.70756531e-01 4.91444290e-01 1.30570006e+00
1.91947356e-01 4.72561091e-01 8.80280018e-01 -6.55405223e-01
-1.57213330e+00 -1.24965954e+00 5.07894397e-01 -4.52206761e-01
3.95684749e-01 1.06376022e-01 -1.03425825e+00 7.79966891e-01
-5.22725470e-02 3.38367343e-01 1.10759132e-01 -8.15542519e-01
-1.03550712e-02 9.47257876e-02 -8.57521057e-01 7.31843472e-01
1.22348666e+00 -1.96683213e-01 -2.71136314e-01 2.39350140e-01
8.18310797e-01 -7.78680921e-01 -6.86197579e-01 3.01130205e-01
4.38782126e-01 -1.14156258e+00 1.48069978e+00 -6.42747343e-01
3.87358814e-01 -7.45033979e-01 -8.19362774e-02 -1.10669899e+00
-2.84764558e-01 -1.36970639e+00 -5.93014479e-01 1.03751111e+00
-2.13034511e-01 -2.53502190e-01 8.86431456e-01 5.32354236e-01
-1.07527003e-01 -4.12770361e-01 -7.83878565e-01 -6.75210476e-01
-1.90786943e-01 -4.53994691e-01 4.18533593e-01 7.95837581e-01
-6.28974855e-01 -1.11363746e-01 -8.76924217e-01 2.68060625e-01
9.86370504e-01 1.34997005e-02 1.05384195e+00 -8.42449069e-01
-2.67205805e-01 -2.80107558e-01 -4.35588777e-01 -1.42385828e+00
3.47420096e-01 -5.06700039e-01 7.86564201e-02 -1.14111650e+00
-1.82625592e-01 -2.04314023e-01 9.38914344e-02 -4.46148366e-01
-5.16691267e-01 3.36323567e-02 4.20953751e-01 3.88026834e-01
-4.38124001e-01 4.69790518e-01 1.36223042e+00 -1.18957192e-01
-2.42340997e-01 2.53301710e-01 -2.41827276e-02 9.14891005e-01
5.77157497e-01 -1.33357495e-01 -7.94426739e-01 -6.51332676e-01
1.38843432e-01 7.18117297e-01 5.55932999e-01 -1.26471734e+00
3.97840649e-01 -6.73756421e-01 3.23768824e-01 -8.88070941e-01
3.76954138e-01 -7.97117949e-01 4.65706587e-01 1.81375012e-01
-1.37469769e-01 3.09820563e-01 -6.61895126e-02 7.91876316e-01
-1.88100606e-01 -1.09990574e-01 7.75307775e-01 -3.32042724e-02
-1.10321569e+00 7.40166485e-01 -2.47665659e-01 3.15986156e-01
1.07992327e+00 -4.22510207e-01 3.12721245e-02 -3.16129565e-01
-7.52780437e-01 1.05550230e-01 4.65101540e-01 6.31219685e-01
1.03030300e+00 -1.41504586e+00 -9.74627137e-01 4.05393392e-01
-1.94515586e-01 1.47158340e-01 3.31010818e-01 7.20753253e-01
-9.01627302e-01 2.54091114e-01 -3.34515303e-01 -9.66961801e-01
-1.13293922e+00 6.51305377e-01 1.89108282e-01 1.37067497e-01
-7.77117074e-01 8.46241772e-01 2.42127523e-01 1.41103983e-01
1.35440409e-01 -4.09838229e-01 -1.67211920e-01 -1.91727236e-01
7.38685727e-01 8.73296618e-01 -2.11245388e-01 -9.89656329e-01
-1.21354677e-01 9.32658672e-01 2.12881789e-01 -7.80702010e-02
8.28673959e-01 -4.95134652e-01 6.09668791e-02 2.58677870e-01
1.31642687e+00 -1.03360405e-02 -2.03599572e+00 -1.09673880e-01
8.90991837e-02 -1.09201467e+00 -6.04566820e-02 -3.42507474e-02
-1.07007229e+00 6.62066460e-01 4.01455879e-01 -4.74606216e-01
1.02591586e+00 -7.87387490e-01 1.09106481e+00 2.09503576e-01
3.99663419e-01 -8.24096978e-01 -6.53218031e-02 5.18653870e-01
7.14896560e-01 -1.29469717e+00 1.60745010e-02 -6.73544884e-01
-4.87889796e-01 1.35728753e+00 5.84023654e-01 -4.22939807e-01
4.78473008e-01 6.54121339e-02 -1.89621747e-02 3.22741419e-01
-6.52600646e-01 1.49237856e-01 6.34837329e-01 5.57500541e-01
2.76064336e-01 -1.76098794e-01 -1.87627003e-01 -1.52066350e-01
1.05757676e-01 3.55874360e-01 6.43590629e-01 6.00418270e-01
-2.14665413e-01 -7.66449511e-01 -6.48345947e-01 1.56581327e-01
-3.34401846e-01 1.09345183e-01 3.72663647e-01 4.52464551e-01
1.47052348e-01 8.77540886e-01 -4.89322096e-02 -8.96619856e-02
1.53200328e-01 -2.14273557e-01 4.53407854e-01 -1.08334171e-02
-1.96682110e-01 2.85442740e-01 -7.63241127e-02 -1.03135717e+00
-5.99123299e-01 -8.28778625e-01 -1.19272065e+00 -3.19651961e-01
5.42126782e-02 -9.63504463e-02 2.47225001e-01 8.24345946e-01
3.59971195e-01 4.74560231e-01 5.65117359e-01 -1.22319555e+00
-4.12530094e-01 -1.65693700e-01 -1.94534272e-01 6.74947858e-01
8.94598126e-01 -6.45238400e-01 -4.19435114e-01 5.68561554e-01] | [10.646709442138672, -1.4642970561981201] |
85a3915b-4fd4-44bd-a349-640a7294bfa2 | neural-shuffle-exchange-networks-sequence-1 | null | null | http://papers.nips.cc/paper/8889-neural-shuffle-exchange-networks-sequence-processing-in-on-log-n-time | http://papers.nips.cc/paper/8889-neural-shuffle-exchange-networks-sequence-processing-in-on-log-n-time.pdf | Neural Shuffle-Exchange Networks - Sequence Processing in O(n log n) Time | A key requirement in sequence to sequence processing is the modeling of long range dependencies. To this end, a vast majority of the state-of-the-art models use attention mechanism which is of O(n^2) complexity that leads to slow execution for long sequences.
We introduce a new Shuffle-Exchange neural network model for sequence to sequence tasks which have O(log n) depth and O(n log n) total complexity. We show that this model is powerful enough to infer efficient algorithms for common algorithmic benchmarks including sorting, addition and multiplication. We evaluate our architecture on the challenging LAMBADA question answering dataset and compare it with the state-of-the-art models which use attention. Our model achieves competitive accuracy and scales to sequences with more than a hundred thousand of elements.
We are confident that the proposed model has the potential for building more efficient architectures for processing large interrelated data in language modeling, music generation and other application domains. | ['Agris Šostaks', 'Emīls Ozoliņš', 'Karlis Freivalds'] | 2019-12-01 | null | null | null | neurips-2019-12 | ['lambada'] | ['natural-language-processing'] | [ 5.40522039e-01 -3.19419414e-01 -5.38741378e-03 -4.18166548e-01
-8.18235695e-01 -8.19215596e-01 2.63451010e-01 3.94873708e-01
-7.87672579e-01 5.92358470e-01 3.16662490e-01 -7.75308311e-01
5.79424761e-02 -7.57206321e-01 -1.01072574e+00 -3.40978324e-01
-2.71861374e-01 7.53169179e-01 3.70790005e-01 -7.10845053e-01
5.49305797e-01 1.88309446e-01 -1.72509003e+00 9.29878056e-01
5.47167718e-01 9.46702898e-01 4.88363802e-01 1.49486923e+00
-6.64761901e-01 1.25840104e+00 -7.42746532e-01 -4.12631363e-01
1.60924256e-01 -2.39728525e-01 -1.29558742e+00 -6.69763148e-01
5.50359488e-01 -2.21773610e-01 -9.95562449e-02 8.13428938e-01
8.40210259e-01 1.61610901e-01 2.41918966e-01 -9.48856831e-01
-8.18372667e-01 1.29154289e+00 -5.49185932e-01 7.08746552e-01
6.10141993e-01 -6.22616820e-02 1.40488589e+00 -5.19975483e-01
4.06932294e-01 1.28372228e+00 7.98249483e-01 4.61302996e-01
-9.31144178e-01 -6.30445898e-01 1.95057034e-01 6.41272843e-01
-9.89919424e-01 -4.44154322e-01 3.19747806e-01 -2.93102622e-01
1.69474196e+00 5.26059628e-01 3.68306011e-01 7.45745420e-01
-7.70461038e-02 9.49323833e-01 4.82506663e-01 -5.92981935e-01
-9.30398107e-02 -5.28904080e-01 5.46394527e-01 6.45875871e-01
-8.94844532e-02 -2.72576094e-01 -9.23398674e-01 -1.88089758e-01
3.45496684e-01 -5.40822744e-01 -1.07908554e-01 3.30694586e-01
-1.29380393e+00 7.12060511e-01 1.93386689e-01 4.05911297e-01
-9.49560553e-02 7.64487207e-01 8.99260461e-01 6.39040709e-01
1.13051170e-02 4.86561209e-01 -7.61284709e-01 -6.64153576e-01
-9.00043368e-01 3.98673266e-01 1.01284206e+00 1.09310150e+00
1.75229207e-01 5.16132601e-02 -1.39868259e-01 8.77522349e-01
2.55898060e-03 5.49756410e-03 7.76010633e-01 -8.46025825e-01
6.40885890e-01 2.70062149e-01 -9.66807976e-02 -6.48699880e-01
-4.22756284e-01 -3.75164598e-01 -8.70985270e-01 -1.52672991e-01
4.16576833e-01 1.93143412e-01 -7.01143563e-01 1.79371476e+00
-2.57225577e-02 7.22076371e-02 -3.69866431e-01 6.45783365e-01
7.52321362e-01 8.56999159e-01 -1.20688286e-02 -1.35372847e-01
1.29170203e+00 -1.21002376e+00 -6.00436687e-01 -3.81921232e-01
8.38800132e-01 -9.10724223e-01 1.42715907e+00 4.86162245e-01
-1.56044555e+00 -8.44477534e-01 -1.19977498e+00 -6.00050271e-01
-2.18078971e-01 -3.15588087e-01 1.03384244e+00 6.47482455e-01
-1.09577537e+00 8.29138100e-01 -6.65487349e-01 -3.77933711e-01
1.69990957e-01 7.38825798e-01 1.65780056e-02 1.45165220e-01
-1.13090122e+00 4.83275563e-01 6.20679915e-01 1.16052210e-01
-5.21320343e-01 -8.22903991e-01 -5.14896631e-01 3.28557730e-01
2.54383802e-01 -9.65913594e-01 1.54639530e+00 -9.52939749e-01
-1.46427703e+00 7.52121687e-01 -3.42442721e-01 -9.55746412e-01
2.07923725e-01 -5.28466821e-01 -1.67870000e-01 -4.43059281e-02
-2.74719268e-01 6.06816649e-01 3.96151930e-01 -5.47827721e-01
-6.25258207e-01 -3.60103458e-01 -3.37102311e-03 7.87842050e-02
-2.48601377e-01 4.00917172e-01 -4.33043599e-01 -7.08646417e-01
-2.44737521e-01 -9.13457990e-01 -2.05757573e-01 -2.13956371e-01
-6.42552078e-02 -3.45599681e-01 3.01733255e-01 -6.21917903e-01
1.61243129e+00 -1.83576703e+00 1.61427528e-01 -1.74228415e-01
-3.89054120e-02 2.84054190e-01 -4.56559807e-01 6.67041361e-01
-2.01415971e-01 3.21443141e-01 -2.69881278e-01 -4.50957492e-02
1.55719355e-01 1.78769797e-01 -4.99668747e-01 1.23614326e-01
-8.88768956e-02 1.22225571e+00 -6.87642574e-01 -1.18451193e-01
-1.39162689e-01 7.02302810e-03 -9.18006003e-01 1.85323015e-01
-6.69575095e-01 -4.76245992e-02 3.86683047e-02 3.12662750e-01
4.05134767e-01 -5.28201282e-01 3.31189394e-01 1.71925083e-01
5.52452281e-02 7.32634008e-01 -9.81334388e-01 2.26060581e+00
-5.45195818e-01 7.93644011e-01 -4.49166782e-02 -9.62892354e-01
6.32935226e-01 1.02165401e-01 2.03392506e-01 -5.17764807e-01
6.62902519e-02 2.32897818e-01 6.72069907e-01 -5.93975723e-01
7.44839370e-01 -3.36662680e-02 -4.52841222e-02 6.06655121e-01
-3.37063260e-02 1.04623184e-01 5.76956153e-01 3.46829474e-01
1.35991752e+00 1.03271522e-01 3.55396926e-01 -3.13321322e-01
7.87727654e-01 -2.61038482e-01 2.09319875e-01 1.22658241e+00
2.45167175e-04 2.67513663e-01 5.00352383e-01 -6.79434419e-01
-1.34711039e+00 -7.36366570e-01 4.34308529e-01 2.22108960e+00
-3.89941275e-01 -6.07077718e-01 -6.03026807e-01 -8.87797028e-02
-2.15504766e-01 6.73621118e-01 -5.78928411e-01 1.27728820e-01
-1.06692636e+00 -6.84951067e-01 9.75541174e-01 8.82410228e-01
2.19485030e-01 -1.45130253e+00 -6.30963027e-01 7.41020024e-01
-2.95768768e-01 -9.46887612e-01 -6.58711076e-01 3.41895223e-01
-8.50828588e-01 -7.28731632e-01 -4.10687536e-01 -1.09231579e+00
-6.23671710e-02 -1.00950412e-01 1.74549460e+00 2.77591139e-01
-4.06441957e-01 -7.09808692e-02 -4.61055845e-01 -6.35370016e-01
-5.17521441e-01 6.19844913e-01 -3.06574553e-01 -5.06930232e-01
5.75001121e-01 -8.87458920e-01 -3.88126820e-01 -1.49736315e-01
-1.10018539e+00 -4.86829095e-02 5.22902191e-01 9.71300960e-01
3.25181454e-01 -4.88869190e-01 7.94280052e-01 -1.07450271e+00
8.76139104e-01 -2.91197985e-01 -6.45117879e-01 2.61242300e-01
-2.34933570e-01 4.72023010e-01 9.45279241e-01 -6.21546626e-01
-6.14355624e-01 6.47483394e-02 -6.30672872e-01 -4.26595435e-02
-6.15948997e-02 7.01938689e-01 7.74469748e-02 2.31776595e-01
6.31866038e-01 4.58323359e-01 -1.90870613e-01 -5.29158592e-01
4.36031044e-01 6.75747871e-01 7.73764193e-01 -7.24786758e-01
3.09467524e-01 3.47720921e-01 5.65419868e-02 -7.63301611e-01
-7.83596873e-01 -6.78069890e-01 -5.08901238e-01 4.55244511e-01
7.05282509e-01 -5.20699441e-01 -1.40373838e+00 4.84757096e-01
-1.45627737e+00 -4.39752221e-01 -1.54377120e-02 1.67448252e-01
-8.09960544e-01 6.64872348e-01 -1.08654690e+00 -9.20996606e-01
-8.50547493e-01 -8.40313792e-01 1.03606200e+00 -1.31420746e-01
-4.75564361e-01 -7.25008309e-01 9.03750584e-02 4.67229366e-01
4.44915444e-01 -3.63445073e-01 1.49681079e+00 -8.40006232e-01
-7.67773807e-01 9.70951542e-02 -9.33057368e-02 1.00488670e-01
-5.45071244e-01 -2.78471202e-01 -7.61127949e-01 -1.15079850e-01
-1.71759605e-01 -5.80402970e-01 1.21090007e+00 2.46085837e-01
1.72549069e+00 -2.31193498e-01 1.49739817e-01 7.16772676e-01
1.27451372e+00 3.32585692e-01 8.45921814e-01 4.30769205e-01
5.06981075e-01 3.55565786e-01 2.40592286e-01 5.25842369e-01
3.20520073e-01 3.80495131e-01 3.51115018e-01 3.33456039e-01
-8.76317322e-02 -2.57209152e-01 2.56942838e-01 1.40237236e+00
1.26463220e-01 -6.64656401e-01 -1.05501497e+00 8.23163986e-01
-1.91931558e+00 -1.27316034e+00 -2.17690974e-01 2.07600069e+00
8.23523402e-01 2.27210954e-01 6.98264316e-02 5.39655447e-01
4.25477028e-01 2.51664519e-01 -4.10412133e-01 -1.13691521e+00
-5.77785708e-02 7.30967700e-01 4.80716676e-01 5.97368419e-01
-9.35872197e-01 1.04954720e+00 7.25779295e+00 9.29510236e-01
-6.95147276e-01 -1.36959821e-01 5.12695312e-01 -2.04912290e-01
-2.01422140e-01 -3.43737781e-01 -8.13838482e-01 3.80177200e-01
1.35253584e+00 -2.32148245e-01 7.78267980e-01 5.58204055e-01
-2.19685733e-01 9.45949480e-02 -1.40155065e+00 1.11763060e+00
1.94782302e-01 -1.66011655e+00 2.94276386e-01 -2.51910537e-01
4.74774152e-01 2.46524706e-01 -2.63379753e-01 3.29742074e-01
3.40179890e-01 -1.52087247e+00 5.20421565e-01 2.95172453e-01
6.69716716e-01 -9.47245836e-01 7.10619509e-01 5.47757626e-01
-1.19100988e+00 -3.48991841e-01 -5.44351995e-01 -4.68753874e-01
4.20103401e-01 1.43654466e-01 -9.17251766e-01 4.23209012e-01
7.34922707e-01 4.51718032e-01 -4.15182650e-01 9.31809664e-01
2.55499125e-01 7.87378252e-01 -5.08239686e-01 -4.56826985e-01
3.12846750e-01 1.34456024e-01 2.51018941e-01 1.49280107e+00
3.16688091e-01 8.78141820e-02 4.60774917e-03 4.21084315e-01
-3.38463813e-01 3.67238998e-01 -3.25152695e-01 -3.58239353e-01
3.74388546e-01 7.34965086e-01 -4.98681933e-01 -5.71268022e-01
-4.25063699e-01 9.45723772e-01 6.22863054e-01 7.93530699e-03
-8.40198338e-01 -5.72449744e-01 5.48427582e-01 -9.59193781e-02
7.41576672e-01 -3.79312277e-01 -6.69288278e-01 -9.71316576e-01
-4.13617082e-02 -1.33976722e+00 7.54693866e-01 -7.92095900e-01
-1.36752021e+00 7.26492822e-01 -3.98308933e-01 -5.75299859e-01
-5.68902552e-01 -8.83086026e-01 -3.40889484e-01 9.12859499e-01
-1.31718612e+00 -8.11035454e-01 -6.65737987e-02 3.35391045e-01
7.77973711e-01 -2.12094724e-01 1.00661349e+00 5.34302175e-01
2.90540833e-04 8.37810159e-01 8.29343721e-02 2.13450238e-01
4.06956255e-01 -1.20797396e+00 1.09481311e+00 5.16838968e-01
6.10961735e-01 8.06399941e-01 6.27352118e-01 -2.25741535e-01
-1.67403650e+00 -7.05104053e-01 1.23195839e+00 -4.43738073e-01
9.00918067e-01 -5.96564770e-01 -9.69843090e-01 8.09062421e-01
3.10590684e-01 -3.35368544e-01 1.15054703e+00 3.29801053e-01
-5.86173594e-01 -1.97443888e-02 -4.06318784e-01 5.36683321e-01
1.40948522e+00 -6.12287641e-01 -6.04994953e-01 2.95529068e-01
9.40871656e-01 -4.84742701e-01 -6.25530839e-01 2.16475368e-01
8.14008117e-01 -1.00146282e+00 9.61970091e-01 -1.12738025e+00
6.16009116e-01 -1.41342655e-01 -3.16030025e-01 -7.78511643e-01
-3.17408234e-01 -9.38543856e-01 -2.71862358e-01 8.45958948e-01
5.95746517e-01 -2.03673333e-01 9.25571918e-01 -7.13723898e-02
-1.36376083e-01 -7.01332211e-01 -7.63408422e-01 -7.23200440e-01
3.39481026e-01 -8.85463119e-01 7.36968338e-01 7.14657664e-01
-1.32393897e-01 7.57268727e-01 -5.55095911e-01 -1.31270453e-01
2.38949612e-01 4.62920219e-01 7.23671138e-01 -9.83561754e-01
-8.85667384e-01 -5.57080746e-01 -3.44986886e-01 -1.55137479e+00
2.35967368e-01 -1.17843664e+00 -6.49821013e-02 -1.50420606e+00
2.03349590e-01 -2.80717295e-02 -2.02488631e-01 1.14125110e-01
-9.99057442e-02 2.21679643e-01 4.51808184e-01 -2.40170375e-01
-5.76340973e-01 2.58289188e-01 8.84633601e-01 -1.25826359e-01
4.94932942e-02 -2.88465232e-01 -7.31588900e-01 6.43238604e-01
8.32589090e-01 -4.66841638e-01 -2.53952503e-01 -1.06064963e+00
7.89145529e-01 -4.54126894e-02 -9.53161195e-02 -9.22130227e-01
4.31430995e-01 8.68273005e-02 1.99234307e-01 -8.92358899e-01
3.12325209e-01 -4.30792689e-01 -1.75807774e-01 4.35507983e-01
-9.63178873e-01 6.16615534e-01 3.96032095e-01 4.19414669e-01
-2.05137104e-01 -4.28682327e-01 4.62378621e-01 -3.76838952e-01
-6.83527946e-01 1.94670379e-01 -3.42089087e-01 3.95854890e-01
5.48884690e-01 2.56011519e-03 -2.74140507e-01 -7.14461625e-01
-5.55083334e-01 1.60933331e-01 -1.07318580e-01 5.12080014e-01
2.83426195e-01 -1.02207041e+00 -9.22047019e-01 6.31220192e-02
1.44328913e-02 -1.94466367e-01 5.04482761e-02 4.09856200e-01
-1.10498154e+00 8.07851195e-01 -8.84558558e-02 -4.40370977e-01
-1.63215840e+00 8.21458101e-01 -1.86802335e-02 -6.91794574e-01
-1.47030324e-01 1.25252175e+00 -7.35350922e-02 -5.46285927e-01
4.83475685e-01 -4.79764700e-01 9.22648683e-02 -8.42218921e-02
8.98164213e-01 3.94807577e-01 1.21398807e-01 -2.65549004e-01
-4.00257736e-01 3.76369774e-01 -3.03962409e-01 -9.78569984e-02
1.31356001e+00 1.28704265e-01 -5.16067207e-01 6.53729856e-01
1.19051433e+00 -4.54352088e-02 -4.83955115e-01 8.28799829e-02
4.51084524e-01 -2.14818209e-01 -4.06668812e-01 -7.82657623e-01
-4.77452010e-01 1.27013910e+00 3.97752464e-01 8.52347538e-02
1.12736058e+00 -1.36579886e-01 1.11112142e+00 9.21134412e-01
2.34848768e-01 -9.87776816e-01 2.70334817e-02 1.16662002e+00
8.67164016e-01 -8.78563106e-01 -2.38655522e-01 -2.32866004e-01
-3.77348751e-01 1.23219013e+00 5.26464164e-01 -1.22180782e-01
2.95232356e-01 7.84184813e-01 3.27256694e-02 2.51913875e-01
-1.44754326e+00 -1.30897239e-01 1.69663027e-01 4.19598520e-01
1.03328872e+00 -2.14220554e-01 -5.18246114e-01 6.02764130e-01
-6.61897004e-01 2.52186030e-01 5.21260321e-01 8.98844659e-01
-4.90465045e-01 -1.53339946e+00 -1.85519800e-01 5.38707614e-01
-9.25006747e-01 -7.67438352e-01 -3.51563752e-01 4.73023474e-01
-5.91682866e-02 8.25986743e-01 2.12329686e-01 -1.72088161e-01
3.41692455e-02 4.66297537e-01 6.17439806e-01 -4.78678048e-01
-1.17163455e+00 6.63809187e-04 3.26420516e-01 -4.94889855e-01
-2.53053695e-01 -3.37658376e-01 -1.26223576e+00 -5.27860284e-01
-6.40187413e-02 -5.80083132e-02 6.10343516e-01 5.71452141e-01
5.51852882e-01 6.56808555e-01 -4.55877520e-02 -5.07543981e-01
-7.58046925e-01 -1.05775273e+00 -2.93700993e-01 4.37156647e-01
2.54642189e-01 1.86205432e-01 1.42053485e-01 2.06538916e-01] | [10.841506004333496, 7.24830436706543] |
ea5eb57a-dfa1-4bfd-9d48-33de977c9ea6 | programmable-spectral-filter-arrays-for | 2109.14450 | null | https://arxiv.org/abs/2109.14450v2 | https://arxiv.org/pdf/2109.14450v2.pdf | Programmable Spectral Filter Arrays using Phase Spatial Light Modulator | Spatially varying spectral modulation can be implemented using a liquid crystal spatial light modulator (SLM) since it provides an array of liquid crystal cells, each of which can be purposed to act as a programmable spectral filter array. However, such an optical setup suffers from strong optical aberrations due to the unintended phase modulation, precluding spectral modulation at high spatial resolutions. In this work, we propose a novel computational approach for the practical implementation of phase SLMs for implementing spatially varying spectral filters. We provide a careful and systematic analysis of the aberrations arising out of phase SLMs for the purposes of spatially varying spectral modulation. The analysis naturally leads us to a set of "good patterns" that minimize the optical aberrations. We then train a deep network that overcomes any residual aberrations, thereby achieving ideal spectral modulation at high spatial resolution. We show a number of unique operating points with our prototype including dynamic spectral filtering, material classification, and single- and multi-image hyperspectral imaging. | ['Tuo Zhuang', 'Ryuichi Tadano', 'Vijay Rengarajan', 'Vishwanath Saragadam', 'Jun Murayama', 'Hideki Oyaizu', 'Aswin C. Sankaranarayanan'] | 2021-09-29 | null | null | null | null | ['material-classification'] | ['computer-vision'] | [ 1.15386021e+00 -3.78972322e-01 2.15675503e-01 -2.48425757e-03
-2.53055930e-01 -4.99289334e-01 4.09332454e-01 -5.75983346e-01
-3.26448590e-01 7.32747018e-01 -1.93802521e-01 -3.31389695e-01
-5.45266807e-01 -8.68020475e-01 -6.57915294e-01 -1.18170536e+00
1.79760069e-01 2.57536471e-01 2.12609768e-01 -2.87846141e-02
2.96369761e-01 5.82336485e-01 -1.79116130e+00 2.03669742e-01
1.12634969e+00 1.04053771e+00 7.79045165e-01 6.45598888e-01
1.46260977e-01 4.19014901e-01 -7.41795897e-01 5.09948254e-01
4.81576145e-01 -5.31869233e-01 -2.82203674e-01 3.06556970e-01
8.98298323e-01 -2.20843837e-01 -2.19570294e-01 1.40203798e+00
4.37275678e-01 2.22390652e-01 6.77363813e-01 -6.76902294e-01
-7.88243473e-01 2.90490717e-01 -5.49200416e-01 -1.92683384e-01
3.10845613e-01 7.59270549e-01 7.53179729e-01 -6.26945555e-01
5.70223093e-01 8.89325142e-01 5.46607792e-01 5.24923384e-01
-1.52156699e+00 -6.82276309e-01 -2.72484899e-01 1.66306779e-01
-1.32996261e+00 -6.20480776e-01 6.69601083e-01 -5.91285408e-01
1.08158791e+00 3.94354343e-01 8.30114663e-01 4.36624318e-01
2.85868675e-01 -1.44013055e-02 1.42151248e+00 -4.93981272e-01
2.10177794e-01 -2.03901216e-01 3.73168848e-02 4.92856830e-01
2.85734057e-01 5.74203491e-01 -5.33523738e-01 2.00928792e-01
7.00254798e-01 -1.55794293e-01 -1.09905422e+00 -1.38361648e-01
-6.23136044e-01 2.30435908e-01 7.62489140e-01 3.51903886e-01
-2.21478790e-01 -9.09230933e-02 -5.19309580e-01 3.03054273e-01
2.89694875e-01 1.28049278e+00 -7.93980807e-02 5.25134087e-01
-1.04048061e+00 -5.33260703e-02 4.19951856e-01 6.00446582e-01
9.11170363e-01 2.38153115e-02 -3.56617302e-01 6.50473356e-01
4.39502627e-01 8.57973993e-01 2.54195601e-01 -1.06686270e+00
-1.42546907e-01 6.10540569e-01 3.79902571e-01 -6.03559256e-01
-6.77233398e-01 -3.74660969e-01 -9.86193657e-01 8.33908260e-01
4.15903002e-01 5.21807000e-02 -1.03718805e+00 1.52868724e+00
-6.42339885e-02 3.40939671e-01 8.36684648e-03 1.03506505e+00
4.61976171e-01 7.59673238e-01 -3.01542282e-01 -6.96814358e-01
1.01810277e+00 -6.60308063e-01 -9.53884900e-01 -1.59322262e-01
2.58699358e-01 -8.00517917e-01 1.09433389e+00 4.70030665e-01
-1.06612170e+00 -1.71594769e-01 -1.42760801e+00 8.40729624e-02
-3.09821457e-01 -8.82577822e-02 4.44069207e-01 6.57842338e-01
-1.02393305e+00 9.05281484e-01 -6.89747393e-01 2.09377035e-01
6.02093399e-01 5.35684586e-01 6.66446760e-02 1.09014735e-01
-5.98033130e-01 8.47738206e-01 2.42788032e-01 1.36239186e-01
-1.83571175e-01 -1.17710745e+00 -3.58596414e-01 7.99397677e-02
2.68363804e-01 -6.86187565e-01 9.23370063e-01 -7.80570686e-01
-1.91767323e+00 9.42802012e-01 -5.34405470e-01 -3.80999506e-01
5.94180524e-02 1.67284291e-02 -3.75149369e-01 1.66789144e-01
-3.10117602e-01 1.13786995e-01 9.64490175e-01 -9.82546866e-01
-5.49007952e-01 -3.66106987e-01 -1.35633290e-01 1.68736264e-01
-2.90207773e-01 -1.50476709e-01 1.10929206e-01 2.18035821e-02
2.39875674e-01 -8.11325610e-01 -2.04897001e-01 2.96395575e-03
-3.84037584e-01 2.46530846e-01 1.01284611e+00 -1.26768649e-01
1.16293561e+00 -2.29674911e+00 1.70263380e-01 -1.16226718e-01
2.26642013e-01 6.36669934e-01 -2.21261814e-01 4.88498136e-02
-1.01864181e-01 -1.20073132e-01 -4.98131931e-01 -7.19187409e-02
-2.99695134e-01 -3.21456820e-01 -3.70755413e-04 4.72264260e-01
2.27411613e-01 7.47856259e-01 -6.80759192e-01 3.56231213e-01
4.57124203e-01 4.52454925e-01 -6.67628467e-01 2.63169378e-01
-6.83536112e-01 7.55966365e-01 5.03081605e-02 4.83125985e-01
1.14662898e+00 -5.18760085e-01 2.45193437e-01 -5.12094200e-01
-8.10546160e-01 3.74672234e-01 -1.23943973e+00 1.37960887e+00
-3.58359158e-01 9.70900774e-01 5.88598728e-01 -5.17100155e-01
5.25422752e-01 9.34135914e-02 4.95496154e-01 -8.54466021e-01
-7.23502636e-02 4.79956627e-01 7.51706213e-02 -4.80375350e-01
3.56375575e-01 -3.18395674e-01 6.95496559e-01 2.93176144e-01
-1.34901062e-01 -5.18991292e-01 2.16477603e-01 -6.40236735e-01
8.15630436e-01 -1.10142633e-01 1.41513199e-01 -5.66688895e-01
6.06238842e-01 -2.01855615e-01 2.05706879e-01 6.80548966e-01
8.48335773e-02 5.31237364e-01 -3.63775492e-02 -2.91142911e-01
-9.48087573e-01 -8.32703352e-01 -4.77971345e-01 4.51567918e-01
3.65333706e-01 1.43841433e-03 -4.37813729e-01 4.28357184e-01
5.01457825e-02 3.01017076e-01 7.26184323e-02 -7.79655799e-02
-4.92257744e-01 -9.97092903e-01 -1.05134889e-01 -4.41180468e-02
7.01298118e-01 -8.98957074e-01 -6.64417624e-01 2.21652195e-01
1.84807584e-01 -9.74982858e-01 -9.14802253e-02 1.88253388e-01
-4.37048852e-01 -1.53550959e+00 4.47496437e-02 -5.84866166e-01
3.51635218e-01 5.41912973e-01 8.72816205e-01 -1.90378487e-01
-5.85123420e-01 -4.84152958e-02 2.46699616e-01 -3.98032576e-01
-3.70480210e-01 -2.41196185e-01 1.50322855e-01 1.32441774e-01
4.94640052e-01 -7.93493927e-01 -9.12623584e-01 1.26598075e-01
-7.99088240e-01 2.77924895e-01 4.64433551e-01 8.10098588e-01
6.06711984e-01 4.17312711e-01 7.03617111e-02 -9.05527830e-01
4.54955190e-01 2.96848565e-01 -1.33943760e+00 1.07230194e-01
-7.07103193e-01 -6.03401847e-03 6.44174397e-01 -4.08949107e-01
-1.01502514e+00 -8.23969990e-02 -1.55126154e-01 1.58059373e-02
-2.88879097e-01 2.80131042e-01 -1.64494619e-01 -7.24483728e-01
1.08008707e+00 1.68284714e-01 5.84924668e-02 -3.80904615e-01
1.11743577e-01 8.85626793e-01 4.83198822e-01 -1.67535543e-01
1.00891948e+00 6.63330436e-01 5.45097768e-01 -1.57614946e+00
-9.89386261e-01 -4.61131483e-01 -3.14313203e-01 -1.99942529e-01
7.96891093e-01 -8.55622828e-01 -1.22776103e+00 9.42204058e-01
-9.99590576e-01 -6.15126967e-01 -3.21949124e-01 3.10935676e-01
-4.11566764e-01 5.69254793e-02 -6.37902141e-01 -8.08578253e-01
-3.01959813e-01 -1.16434193e+00 9.57232952e-01 5.22581637e-01
4.31194790e-02 -7.58278966e-01 -2.55880743e-01 1.82228163e-01
8.05969775e-01 1.48498520e-01 9.28530633e-01 4.59338129e-01
-1.44881415e+00 2.12410510e-01 -2.87809372e-01 1.85507759e-01
4.60258126e-01 1.85397834e-01 -1.30597234e+00 -4.60360914e-01
5.02732813e-01 -2.90653892e-02 1.14089465e+00 1.00393558e+00
1.24959302e+00 2.45625712e-02 -4.06840920e-01 1.41587222e+00
1.83898056e+00 4.52139854e-01 7.11534858e-01 3.47465605e-01
6.66466594e-01 1.62486240e-01 1.90038726e-01 1.38137117e-01
-5.34938037e-01 5.54961085e-01 4.44123089e-01 -2.04672143e-01
-3.17884058e-01 3.33009750e-01 -8.20954815e-02 2.02479795e-01
-4.62478995e-02 -4.11584109e-01 -5.03850043e-01 -5.62798753e-02
-1.39297903e+00 -1.06438208e+00 -4.35180515e-01 2.17193985e+00
9.39396918e-01 -2.66866654e-01 -4.90802199e-01 7.93455243e-02
6.17683828e-01 3.68866742e-01 -5.72069466e-01 3.88997257e-01
-4.58916366e-01 7.40319431e-01 7.94937670e-01 9.68923032e-01
-1.21998060e+00 7.88986444e-01 6.71544170e+00 4.52351421e-01
-1.50048733e+00 -2.88875729e-01 9.20797349e-04 -1.59200668e-01
-2.61096418e-01 -3.01350981e-01 -9.23083067e-01 6.20465159e-01
6.41506076e-01 4.64266762e-02 9.69008386e-01 1.04035191e-01
4.66925383e-01 -3.54978234e-01 -9.34420764e-01 1.01184022e+00
-4.17289227e-01 -1.73445213e+00 9.65510979e-02 -7.80439824e-02
8.75436544e-01 9.79246851e-03 4.25313115e-01 -6.32471204e-01
-9.94238183e-02 -1.09620285e+00 2.75987446e-01 7.10170686e-01
1.26685405e+00 -3.82347107e-01 1.29232826e-02 4.60993797e-01
-8.35220575e-01 -3.21504414e-01 -5.09269536e-01 -3.67967546e-01
-7.71586299e-02 7.91938126e-01 -5.24218500e-01 1.45795509e-01
3.14475596e-01 6.62062228e-01 -1.51850015e-01 1.05920386e+00
-2.73532301e-01 2.91727066e-01 -4.60680664e-01 4.70612757e-03
-1.64405152e-01 -8.84195685e-01 6.51218414e-01 7.88166881e-01
5.72793245e-01 3.71976525e-01 -1.36614889e-02 1.31247020e+00
-8.50373656e-02 -4.29978848e-01 -6.70045197e-01 -2.26481676e-01
5.01107872e-01 1.17859697e+00 -2.18123928e-01 4.68741450e-03
-5.66594720e-01 7.01469481e-01 3.48265399e-04 6.91409707e-01
-2.02100053e-01 -2.39977613e-01 1.27558732e+00 3.38568568e-01
-9.31405753e-04 -3.26271594e-01 -3.86835933e-01 -1.21494246e+00
-1.14850134e-01 -6.05708241e-01 -7.63614699e-02 -9.84033823e-01
-1.23790419e+00 1.62050948e-01 -6.70835853e-01 -1.04648423e+00
3.53599668e-01 -1.29960799e+00 -3.19518000e-01 1.39209366e+00
-2.04642153e+00 -6.80034220e-01 -5.00419974e-01 4.57820147e-01
6.07700720e-02 2.23128032e-02 9.61211562e-01 3.30495030e-01
-2.88246393e-01 1.20577961e-02 3.77334833e-01 -3.80129963e-01
8.42691541e-01 -1.15325046e+00 -7.68470839e-02 1.04026556e+00
-2.69752860e-01 4.92870152e-01 6.58940494e-01 -4.22566593e-01
-1.50043535e+00 -1.16733730e+00 5.41838646e-01 -2.42286250e-01
5.63015878e-01 -1.62422448e-01 -9.18438435e-01 3.45190108e-01
3.48290771e-01 -3.69758755e-01 8.07412863e-01 -2.80997455e-01
-1.32573629e-02 -1.94528118e-01 -1.05397093e+00 6.47499144e-01
1.25996614e+00 -5.57065070e-01 -1.08745240e-01 9.39140797e-01
6.99698806e-01 -5.81322312e-01 -4.92518932e-01 5.68684459e-01
3.95282179e-01 -1.03609765e+00 9.05290961e-01 -1.80917069e-01
5.13174891e-01 -6.69932425e-01 1.52449831e-01 -1.52298427e+00
-7.86277831e-01 -9.22563434e-01 -7.64536019e-03 5.34879267e-01
2.53765136e-01 -1.09106541e+00 7.19880879e-01 2.77898788e-01
-4.09571081e-01 -9.80106220e-02 -6.40195906e-01 -6.95870936e-01
-1.91108912e-01 -5.13156317e-02 7.05241919e-01 9.32242870e-01
-2.94000059e-01 3.02995920e-01 -5.12567498e-02 8.00642192e-01
9.03566480e-01 4.73919541e-01 1.56343222e-01 -1.34004247e+00
-2.77622193e-01 -8.80822062e-01 -1.06048368e-01 -1.35305727e+00
6.14693724e-02 -9.48074281e-01 5.34829274e-02 -1.24034929e+00
-8.37584957e-02 -2.65730321e-01 -3.05256825e-02 -1.51318908e-01
-1.36860937e-01 3.55398685e-01 3.52574810e-02 1.16871789e-01
-1.42736477e-04 1.59230903e-01 1.64658201e+00 -3.88818145e-01
-6.26032352e-01 8.97610113e-02 -4.94709373e-01 5.02898037e-01
6.13667309e-01 7.02628866e-02 -2.08658785e-01 -5.14502108e-01
3.93188477e-01 -1.47102341e-01 4.08701003e-01 -1.36438155e+00
5.77889919e-01 -4.56762940e-01 5.18233895e-01 -3.38758826e-01
4.06993806e-01 -8.98480415e-01 4.20028716e-01 5.61557353e-01
-4.17956524e-02 -1.03571928e+00 -1.23099433e-02 2.18415543e-01
-1.76374257e-01 -1.99949741e-01 1.39960301e+00 -3.23884666e-01
-7.09355831e-01 3.44131857e-01 -5.29151499e-01 -3.87240022e-01
7.74601817e-01 -2.49139085e-01 -7.87710130e-01 1.46773055e-01
-6.01331711e-01 2.03954354e-01 8.81739199e-01 -2.50837117e-01
3.75470459e-01 -1.11597812e+00 -2.07813084e-01 8.82747889e-01
-1.00332133e-01 5.49174063e-02 3.62443179e-01 4.96944368e-01
-9.01120722e-01 4.61157322e-01 -1.22853592e-01 -7.45910406e-01
-1.16972232e+00 3.38128000e-01 1.13952911e+00 5.67024387e-02
-5.52749872e-01 1.01127064e+00 3.41693908e-02 -4.90688905e-02
-4.20822576e-02 -3.08883429e-01 -1.32155895e-01 -7.52717257e-02
7.00125575e-01 1.81250945e-01 1.91628233e-01 -1.78421304e-01
-4.21306156e-02 9.23555851e-01 6.79252326e-01 3.17600012e-01
1.49728894e+00 -6.91626221e-02 -5.74279070e-01 1.64010629e-01
8.28474343e-01 4.22981866e-02 -1.39267802e+00 -2.89458662e-01
-3.71973813e-01 -7.74516165e-01 8.75228569e-02 -7.31731415e-01
-7.02123523e-01 9.08348918e-01 6.94777906e-01 3.63950431e-01
1.57972622e+00 -4.99075413e-01 4.80846882e-01 5.26249051e-01
2.95869231e-01 -1.11094165e+00 -2.03783602e-01 7.00997472e-01
4.12672251e-01 -1.07404661e+00 -4.35356312e-02 -7.75013864e-01
3.81291628e-01 1.20046186e+00 4.21804428e-01 8.31321552e-02
6.24795139e-01 5.50428689e-01 3.29667330e-02 -3.68111938e-01
-5.00064492e-01 -2.83868670e-01 3.39262843e-01 9.05997634e-01
4.00768757e-01 3.31647307e-01 -2.84681171e-01 -2.58112401e-01
-1.65972486e-01 -1.72278322e-02 5.99702060e-01 5.04406333e-01
-9.03240085e-01 -8.73103321e-01 -3.12227666e-01 5.56320012e-01
-4.34575453e-02 -9.80806276e-02 -1.30543321e-01 2.35776365e-01
1.37109742e-01 8.51034939e-01 2.91106671e-01 -4.74279560e-02
4.96587485e-01 2.95764893e-01 7.71434903e-01 -7.46975660e-01
-1.14554219e-01 1.64558023e-01 -2.79720902e-01 -5.82265913e-01
-7.67914295e-01 -3.81081820e-01 -9.84412253e-01 -1.53279856e-01
-3.68098587e-01 -3.88102889e-01 3.15082133e-01 8.07482660e-01
3.37575763e-01 6.41081214e-01 6.26478553e-01 -8.49767923e-01
-3.91200334e-01 -9.23048139e-01 -9.23391759e-01 4.36929762e-01
7.45592296e-01 -4.67122048e-01 -6.15087628e-01 -2.85573881e-02] | [10.256566047668457, -2.477782964706421] |
f26c1110-df91-471c-bab7-7ee2015be881 | evaluating-the-robustness-of-machine-reading | 2304.03145 | null | https://arxiv.org/abs/2304.03145v1 | https://arxiv.org/pdf/2304.03145v1.pdf | Evaluating the Robustness of Machine Reading Comprehension Models to Low Resource Entity Renaming | Question answering (QA) models have shown compelling results in the task of Machine Reading Comprehension (MRC). Recently these systems have proved to perform better than humans on held-out test sets of datasets e.g. SQuAD, but their robustness is not guaranteed. The QA model's brittleness is exposed when evaluated on adversarial generated examples by a performance drop. In this study, we explore the robustness of MRC models to entity renaming, with entities from low-resource regions such as Africa. We propose EntSwap, a method for test-time perturbations, to create a test set whose entities have been renamed. In particular, we rename entities of type: country, person, nationality, location, organization, and city, to create AfriSQuAD2. Using the perturbed test set, we evaluate the robustness of three popular MRC models. We find that compared to base models, large models perform well comparatively on novel entities. Furthermore, our analysis indicates that entity type person highly challenges the MRC models' performance. | ['Tunde Oluwaseyi Ajayi', 'Clemencia Siro'] | 2023-04-06 | null | null | null | null | ['reading-comprehension', 'machine-reading-comprehension'] | ['natural-language-processing', 'natural-language-processing'] | [ 1.75524086e-01 5.37032485e-01 4.96349692e-01 -3.70247781e-01
-9.68652368e-01 -9.63442564e-01 6.78902149e-01 3.03591251e-01
-5.43020546e-01 1.21742094e+00 3.35101783e-01 -4.28838551e-01
-4.92743142e-02 -8.49190652e-01 -1.12909949e+00 -1.63030505e-01
-3.67593467e-02 5.48713744e-01 3.75152439e-01 -6.39973402e-01
2.44856551e-01 1.42553911e-01 -1.36483574e+00 3.76720637e-01
1.45848596e+00 4.19005722e-01 -5.65183401e-01 8.29723001e-01
4.88416523e-01 1.04789436e+00 -1.28353477e+00 -1.20678985e+00
8.09890702e-02 -2.33697712e-01 -1.31664741e+00 -6.55085564e-01
8.05154443e-01 -1.94859728e-01 -2.73850799e-01 9.70790625e-01
8.20309579e-01 2.27401614e-01 7.81604469e-01 -1.37214768e+00
-9.86472368e-01 8.59948933e-01 1.02444313e-01 3.98893803e-01
1.11030769e+00 1.67911276e-01 9.48309004e-01 -7.85045385e-01
8.73185158e-01 1.18390954e+00 7.97450006e-01 7.03178585e-01
-1.10349691e+00 -5.24371207e-01 -3.98522755e-03 4.62049574e-01
-1.50841105e+00 -3.85870397e-01 2.90643185e-01 -2.10499600e-01
8.37905645e-01 6.54773891e-01 -2.60856420e-01 1.64991856e+00
1.62697241e-01 5.15978098e-01 1.02103102e+00 -3.49486619e-01
3.12356919e-01 1.92275569e-01 1.61865920e-01 5.01875699e-01
2.68534392e-01 -1.45506680e-01 -3.41781408e-01 -4.43169206e-01
-1.50385529e-01 -8.01440775e-01 -8.07002127e-01 -4.01978455e-02
-1.31315398e+00 6.36249244e-01 5.74723721e-01 1.49624303e-01
-2.01264039e-01 -1.56113625e-01 1.91170022e-01 6.40566885e-01
3.95584673e-01 1.30146444e+00 -8.08748484e-01 -5.20681329e-02
-4.54143167e-01 9.03735280e-01 1.11424053e+00 1.07206976e+00
2.68365651e-01 -4.45596993e-01 -5.81027567e-01 5.28601527e-01
-1.95610836e-01 5.75810790e-01 4.59112823e-01 -7.19899714e-01
1.03847504e+00 6.23906314e-01 4.57568973e-01 -1.11021912e+00
-5.87176979e-01 -3.87245506e-01 -8.19460273e-01 -3.09470594e-01
6.26796484e-01 -2.73922920e-01 -7.55113482e-01 2.03143287e+00
2.52298355e-01 1.65495411e-01 5.96600354e-01 5.44517934e-01
1.21387470e+00 3.85435522e-01 1.64591968e-01 3.72542173e-01
1.43082833e+00 -9.15431261e-01 -5.12355983e-01 -4.87562791e-02
7.33402610e-01 -3.00441861e-01 1.27429080e+00 1.48363426e-01
-1.18228364e+00 -4.14481938e-01 -8.11014175e-01 4.96957079e-02
-9.49190557e-01 -4.66957062e-01 -5.22081740e-02 7.93739259e-01
-7.39606023e-01 4.12457913e-01 -1.73216939e-01 -2.58230656e-01
2.91156888e-01 7.94766322e-02 -7.70184278e-01 -3.04262996e-01
-1.98598099e+00 1.46530628e+00 4.18286949e-01 -1.27630621e-01
-7.92217672e-01 -7.75822043e-01 -1.02759767e+00 1.49430484e-01
3.27570409e-01 -9.74965096e-01 1.34306073e+00 -4.62763011e-01
-1.06015825e+00 7.21545756e-01 -4.11733277e-02 -6.38884366e-01
8.02492321e-01 -4.52164918e-01 -8.42538893e-01 2.54066080e-01
1.98633641e-01 3.98525834e-01 5.99462807e-01 -1.13632619e+00
-3.01904708e-01 -7.16145858e-02 5.30595005e-01 2.18663841e-01
2.01684721e-02 1.74508736e-01 1.09103702e-01 -6.48848414e-01
-2.48698339e-01 -6.40639603e-01 -9.11814943e-02 -7.13440537e-01
-8.87827218e-01 -1.79329574e-01 1.95900470e-01 -1.04919219e+00
1.47558737e+00 -1.85204458e+00 1.97690889e-01 3.02796483e-01
1.39889374e-01 1.94267526e-01 -1.34536654e-01 3.83880436e-01
-2.69866496e-01 7.19888031e-01 -3.71550143e-01 2.03471944e-01
1.60577714e-01 -7.90213123e-02 -3.06128263e-01 8.90273526e-02
5.73360324e-01 1.05499530e+00 -9.42709565e-01 -3.00112545e-01
-5.66997409e-01 -1.42106891e-01 -6.91729784e-01 2.08601743e-01
-4.24955100e-01 1.74205899e-01 -4.13548410e-01 7.32214034e-01
5.73042810e-01 -1.96613476e-01 -3.05056363e-01 6.63943440e-02
3.91073495e-01 3.46589893e-01 -8.66882026e-01 1.14332759e+00
1.08481199e-02 4.12441134e-01 -3.92207175e-01 -3.42023760e-01
7.05169618e-01 4.54725534e-01 -3.36250663e-01 -6.68104291e-01
-1.33994654e-01 1.32269114e-01 1.92182645e-01 -9.00343776e-01
7.48563230e-01 6.75897747e-02 -4.44568902e-01 2.86640137e-01
-7.81457126e-02 -2.57604957e-01 1.98086366e-01 4.05130029e-01
1.58381832e+00 -1.71493262e-01 2.02831894e-01 -6.97083697e-02
6.71562672e-01 2.04722837e-01 2.51803398e-01 1.14580834e+00
-2.62237012e-01 8.58134031e-01 4.91908997e-01 -8.88899490e-02
-1.07460284e+00 -1.34068751e+00 -2.60771364e-01 9.91843045e-01
1.01811230e-01 -2.57860839e-01 -1.09348845e+00 -1.15465546e+00
1.02411933e-01 1.20350432e+00 -8.70736003e-01 -5.20512998e-01
-6.90193295e-01 -7.04485714e-01 1.33923113e+00 3.19640666e-01
7.72802711e-01 -1.19995785e+00 -4.96617407e-01 1.83303833e-01
-7.15408564e-01 -1.00819099e+00 -1.67219356e-01 -3.66754323e-01
-1.25114292e-01 -1.25710166e+00 -8.30661118e-01 -6.96486473e-01
6.36980951e-01 -4.94296253e-01 1.71294439e+00 2.64476743e-02
1.32387042e-01 3.76153350e-01 -6.66287959e-01 -5.19424677e-01
-7.20720768e-01 4.25233811e-01 1.76328629e-01 -2.54964858e-01
2.42555588e-01 -2.22638935e-01 -3.31801981e-01 7.46519327e-01
-1.10750461e+00 -3.06164265e-01 1.78199798e-01 7.93037295e-01
5.20168478e-03 -3.42688821e-02 1.21606123e+00 -1.20462871e+00
1.07498801e+00 -9.82069194e-01 -1.10661183e-02 8.42407882e-01
-3.66333574e-01 -8.08103755e-02 7.27211416e-01 -4.74175692e-01
-9.82955873e-01 -8.09189618e-01 -1.54846102e-01 1.61686435e-01
-3.54720950e-01 6.92483306e-01 -4.02548522e-01 2.37204731e-02
1.39742041e+00 -8.53647813e-02 -4.25580412e-01 -3.44207764e-01
4.50271815e-01 5.55988967e-01 8.49042058e-01 -7.03467190e-01
1.21113849e+00 -1.65607959e-01 -5.73945224e-01 -4.04010892e-01
-8.25724304e-01 -2.07363628e-02 -3.14388692e-01 -4.65665609e-02
7.89348066e-01 -8.21708441e-01 -4.14962828e-01 6.17757738e-01
-1.08649611e+00 -1.71894804e-01 -1.48519963e-01 1.97738975e-01
-4.47154909e-01 2.90342271e-01 -3.30903471e-01 -3.38508934e-01
-2.85144269e-01 -6.42411351e-01 5.99334657e-01 3.24659526e-01
-4.86135989e-01 -9.98548746e-01 2.70200908e-01 6.30214751e-01
5.44703186e-01 7.91910052e-01 1.24423957e+00 -1.24313450e+00
-2.38396734e-01 -1.92373738e-01 1.36775851e-01 3.62740755e-01
-1.22060932e-01 -2.03325134e-02 -8.37869644e-01 -3.30615580e-01
-2.34870151e-01 -4.96631920e-01 4.20630157e-01 -3.89220178e-01
1.02998805e+00 -5.71878314e-01 -1.26330197e-01 3.28359425e-01
9.53034759e-01 -1.45826668e-01 9.72323477e-01 7.50176013e-01
4.29120660e-01 8.23249876e-01 7.67412663e-01 -9.55540221e-03
6.84636712e-01 3.75807852e-01 3.27730358e-01 1.08180963e-01
2.73226827e-01 -3.05628985e-01 4.32988167e-01 4.83554989e-01
-1.79876033e-02 -7.97303557e-01 -1.29604828e+00 6.40483260e-01
-1.53624475e+00 -1.10232162e+00 -1.98455881e-02 1.97696602e+00
1.00343239e+00 2.38446638e-01 -2.05961928e-01 -6.96255937e-02
5.59716403e-01 -1.46926403e-01 -5.42924225e-01 -3.74355912e-01
-6.89024389e-01 3.55633765e-01 1.88800052e-01 3.52619618e-01
-9.86197531e-01 8.41780603e-01 6.67149973e+00 3.89697313e-01
-2.84139305e-01 -2.66344361e-02 5.68512559e-01 1.85819581e-01
-6.18538439e-01 -3.10074717e-01 -4.09023523e-01 5.66345274e-01
1.08051026e+00 -5.21323979e-01 1.83390692e-01 5.42905331e-01
-2.64649302e-01 6.29220903e-02 -1.22981894e+00 3.01524639e-01
3.66537601e-01 -9.36194897e-01 2.26307169e-01 -4.00828719e-01
8.80642831e-01 -1.35470361e-01 1.55822024e-01 8.26995254e-01
5.42432249e-01 -1.49672461e+00 6.88301563e-01 7.01403260e-01
6.74803317e-01 -6.84527934e-01 1.13767040e+00 4.49602872e-01
-3.48907620e-01 -2.08158925e-01 -3.46092552e-01 2.23010778e-01
1.09121099e-01 7.61991218e-02 -1.04980457e+00 8.43520761e-01
1.01312232e+00 -5.57038598e-02 -1.36718369e+00 9.49106932e-01
-4.78627145e-01 7.62964427e-01 -2.19504192e-01 1.13236997e-02
1.47057781e-02 3.13302130e-01 6.45956039e-01 9.56710756e-01
5.91460802e-02 2.35720515e-01 -4.45734382e-01 8.39873612e-01
-4.91119832e-01 5.30046411e-02 -6.78035259e-01 1.77012727e-01
6.37627363e-01 6.02201879e-01 2.43980512e-01 -3.41621637e-01
-1.21464036e-01 8.60863566e-01 6.14260137e-01 6.01565301e-01
-9.64478135e-01 -6.41196728e-01 3.86234820e-01 1.00470521e-01
7.96871558e-02 3.02959144e-01 1.53022990e-01 -1.35397458e+00
4.19548750e-01 -1.63849950e+00 5.54568887e-01 -1.07850182e+00
-1.50857556e+00 7.95782208e-01 1.95381075e-01 -8.66911054e-01
-3.81312877e-01 -3.75245959e-01 -6.53486073e-01 9.71448421e-01
-1.26624286e+00 -9.17983413e-01 -4.01331812e-01 7.10799694e-01
-9.73175559e-03 -3.85441720e-01 9.96545672e-01 2.30390862e-01
-5.87855577e-01 1.33100533e+00 1.17280357e-01 3.12257081e-01
1.00542188e+00 -1.48449624e+00 9.92569745e-01 9.77918208e-01
-1.64895952e-01 7.74902225e-01 9.02669966e-01 -6.02112770e-01
-6.24227703e-01 -1.07387769e+00 1.20029449e+00 -1.28562140e+00
6.31533384e-01 -3.39537352e-01 -1.29542434e+00 8.34040105e-01
1.63876683e-01 -4.16023046e-01 6.99677825e-01 8.39627087e-02
-6.35852516e-01 3.57085377e-01 -1.62452328e+00 9.27022576e-01
9.54617321e-01 -5.74967682e-01 -1.28924417e+00 4.41121250e-01
1.08538973e+00 -5.63435197e-01 -1.16915309e+00 7.30907857e-01
2.14752600e-01 -8.09229553e-01 9.08404946e-01 -1.43201578e+00
5.84812999e-01 -3.54092985e-01 -1.32524341e-01 -1.60282540e+00
-2.92165577e-01 -5.06640017e-01 -8.75831097e-02 1.37279201e+00
1.04210806e+00 -1.01401699e+00 5.02404034e-01 9.90449965e-01
3.32419574e-02 -4.72208768e-01 -1.03156376e+00 -7.31822014e-01
4.72411454e-01 7.30250105e-02 9.35940623e-01 1.23669255e+00
3.87785658e-02 3.62894714e-01 -5.89971384e-03 6.00357413e-01
1.53394222e-01 -4.12872881e-01 9.39463735e-01 -9.95824873e-01
-2.46221006e-01 -1.57975540e-01 -4.57170486e-01 -4.70195323e-01
2.38098219e-01 -5.79751492e-01 -1.29607484e-01 -1.36247075e+00
-5.26908599e-02 -4.25470024e-01 -2.43663162e-01 3.08600903e-01
-9.43524301e-01 -5.68181425e-02 1.15977190e-01 -1.46497712e-01
-5.68025768e-01 4.79791462e-01 1.02586198e+00 -2.13451117e-01
2.17526913e-01 4.88677695e-02 -7.92507648e-01 4.81108308e-01
1.08261240e+00 -4.76833791e-01 -1.89781174e-01 -3.99595678e-01
6.71989620e-01 -2.73587316e-01 5.47572732e-01 -1.05754411e+00
1.06123984e-01 2.36337453e-01 1.62161335e-01 -2.68611461e-01
6.61525577e-02 -4.87595350e-01 3.50787230e-02 3.10754716e-01
-6.29406989e-01 5.69083214e-01 1.92392215e-01 4.16740447e-01
-3.50103140e-01 -4.95698035e-01 3.57086062e-01 5.69731109e-02
-4.31954980e-01 -1.13503382e-01 -1.86194688e-01 8.10306370e-01
9.56359565e-01 -8.36326033e-02 -1.07716453e+00 -5.68133056e-01
-7.13498294e-01 6.10755324e-01 4.71931517e-01 6.32311285e-01
4.41654950e-01 -1.04595959e+00 -1.34868431e+00 -9.47262272e-02
4.22379851e-01 1.55159563e-01 2.98531622e-01 4.48235691e-01
-7.15064883e-01 1.02103591e-01 -1.36898473e-01 -1.34245276e-01
-1.13293326e+00 6.34760976e-01 5.23808360e-01 -3.52414042e-01
-1.35805756e-01 1.08441281e+00 -7.05919787e-02 -1.07101536e+00
9.81651396e-02 -2.13494092e-01 -5.39072335e-01 -1.29024997e-01
4.83508646e-01 5.00878751e-01 9.41931754e-02 -4.66409117e-01
-3.23482096e-01 1.30479291e-01 -3.19682628e-01 -2.54041590e-02
8.24478567e-01 1.97208315e-01 3.24526764e-02 4.26778123e-02
7.93889582e-01 4.12640959e-01 -5.58427334e-01 -1.78411916e-01
2.02091098e-01 -1.09783880e-01 -8.34174931e-01 -1.40381551e+00
-2.62575686e-01 3.35754335e-01 2.61844724e-01 3.91516089e-01
8.47110868e-01 -6.55557364e-02 5.13598561e-01 7.01514244e-01
3.74188185e-01 -8.31088126e-01 -1.57558352e-01 7.45209038e-01
1.16883314e+00 -1.15655613e+00 -1.44603923e-01 -2.86428541e-01
-6.65946186e-01 4.10615742e-01 1.00717342e+00 2.43096635e-01
2.44460583e-01 -3.36497217e-01 1.05479851e-01 -1.15656845e-01
-8.42166245e-01 1.87586546e-01 3.02366018e-01 7.88199186e-01
-1.08058061e-02 1.54260829e-01 3.81978564e-02 6.48913205e-01
-9.51687336e-01 -6.10234559e-01 7.81991839e-01 7.86430061e-01
-1.30523562e-01 -6.70617044e-01 -5.09727299e-01 4.37811106e-01
-6.35119200e-01 -3.34081352e-01 -7.45353520e-01 1.14939761e+00
4.73402292e-02 1.06529808e+00 -2.27985218e-01 -4.70800549e-01
8.91407967e-01 3.56601238e-01 2.70053476e-01 -4.78972524e-01
-1.12400377e+00 -1.35923100e+00 6.70033693e-01 -3.17578107e-01
-5.45083545e-02 -5.36372542e-01 -7.44450629e-01 -4.64283615e-01
-3.76594722e-01 4.90031451e-01 2.12314621e-01 7.46237814e-01
4.73116755e-01 4.39266682e-01 4.97587085e-01 1.08067773e-01
-9.47396636e-01 -1.22814751e+00 -6.13598153e-02 9.60922241e-01
2.29765549e-01 -5.23591697e-01 -6.76910698e-01 -9.58311185e-02] | [11.022895812988281, 8.043771743774414] |
200e90c0-6f10-40c7-aafc-e3fd38b24b7e | instant-visual-odometry-initialization-for | 2107.14659 | null | https://arxiv.org/abs/2107.14659v1 | https://arxiv.org/pdf/2107.14659v1.pdf | Instant Visual Odometry Initialization for Mobile AR | Mobile AR applications benefit from fast initialization to display world-locked effects instantly. However, standard visual odometry or SLAM algorithms require motion parallax to initialize (see Figure 1) and, therefore, suffer from delayed initialization. In this paper, we present a 6-DoF monocular visual odometry that initializes instantly and without motion parallax. Our main contribution is a pose estimator that decouples estimating the 5-DoF relative rotation and translation direction from the 1-DoF translation magnitude. While scale is not observable in a monocular vision-only setting, it is still paramount to estimate a consistent scale over the whole trajectory (even if not physically accurate) to avoid AR effects moving erroneously along depth. In our approach, we leverage the fact that depth errors are not perceivable to the user during rotation-only motion. However, as the user starts translating the device, depth becomes perceivable and so does the capability to estimate consistent scale. Our proposed algorithm naturally transitions between these two modes. We perform extensive validations of our contributions with both a publicly available dataset and synthetic data. We show that the proposed pose estimator outperforms the classical approaches for 6-DoF pose estimation used in the literature in low-parallax configurations. We release a dataset for the relative pose problem using real data to facilitate the comparison with future solutions for the relative pose problem. Our solution is either used as a full odometry or as a preSLAM component of any supported SLAM system (ARKit, ARCore) in world-locked AR effects on platforms such as Instagram and Facebook. | ['Luc Oth', 'Christian Forster', 'Jesús Briales', 'Michael Burri', 'Alejo Concha'] | 2021-07-30 | null | null | null | null | ['monocular-visual-odometry'] | ['robots'] | [-5.59281670e-02 6.23068772e-02 -1.30750373e-01 -3.42533179e-02
-4.25541192e-01 -9.32561576e-01 7.64515936e-01 -5.48538603e-02
-5.16727149e-01 6.93722546e-01 -1.83482319e-01 -3.25339079e-01
7.09894150e-02 -5.16062021e-01 -7.97237754e-01 -4.33929175e-01
-1.19739719e-01 7.01167047e-01 1.58060506e-01 -2.99722970e-01
3.34689289e-01 9.12168145e-01 -1.60580683e+00 -7.67266870e-01
8.39668930e-01 7.43103445e-01 1.61766529e-01 1.10114920e+00
5.50935149e-01 6.30886376e-01 -5.45677841e-01 2.07255244e-01
5.30220389e-01 1.42091170e-01 -1.86025918e-01 9.86616127e-03
8.28680336e-01 -5.98124087e-01 -2.09550589e-01 9.27628756e-01
7.00125575e-01 7.17214495e-02 3.22539121e-01 -1.38368893e+00
8.78795534e-02 -2.98101872e-01 -4.64254856e-01 -1.81884751e-01
8.94132912e-01 3.63589972e-01 5.23018599e-01 -8.62606406e-01
1.04956317e+00 9.94905651e-01 5.75924158e-01 1.01795666e-01
-1.00793529e+00 -3.37528974e-01 -4.13767174e-02 -2.44251892e-01
-1.64336562e+00 -7.18698621e-01 5.86707830e-01 -5.42435467e-01
8.28898549e-01 3.32068533e-01 5.78473926e-01 1.13359380e+00
3.92883569e-01 2.11632714e-01 1.02404141e+00 -2.36263245e-01
4.18551594e-01 1.29321918e-01 -1.92033425e-01 6.62688017e-01
7.73998499e-01 2.21156821e-01 -7.17478395e-01 -2.18954822e-03
1.02357113e+00 -5.03205620e-02 -3.31233472e-01 -1.24403167e+00
-1.50889564e+00 3.81653428e-01 4.90993083e-01 -2.52773046e-01
-2.86107570e-01 2.80773818e-01 -1.41716659e-01 4.16214794e-01
1.56779200e-01 3.62181336e-01 -2.55757540e-01 -7.56953418e-01
-7.27275789e-01 2.24296272e-01 5.74919224e-01 1.32366455e+00
8.18040490e-01 7.71628991e-02 4.82673705e-01 4.93614711e-02
3.88788104e-01 1.09716880e+00 3.34155262e-01 -1.10497189e+00
4.86956894e-01 4.13930088e-01 7.91862071e-01 -1.15454888e+00
-6.26960039e-01 -3.98209691e-01 -4.13371027e-01 6.28746152e-01
4.70706403e-01 -2.20571965e-01 -8.57738674e-01 1.56414223e+00
7.16678441e-01 1.75815538e-01 8.42372105e-02 1.13213563e+00
4.41751987e-01 -2.23722346e-02 -5.53287446e-01 -5.47535755e-02
1.09820366e+00 -4.93262947e-01 -8.40586960e-01 -5.39375424e-01
7.55912960e-01 -8.84896517e-01 1.03296101e+00 5.86627066e-01
-7.83502400e-01 -4.36844021e-01 -1.58982682e+00 -2.79108375e-01
-9.58097726e-02 2.79661059e-01 3.40792328e-01 7.04263151e-01
-1.07650650e+00 2.55602807e-01 -1.08590972e+00 -6.73907399e-01
-4.62841809e-01 5.42130888e-01 -4.41234559e-01 -7.68443942e-02
-9.40722644e-01 1.24739778e+00 -8.18627849e-02 6.08769208e-02
-5.66253006e-01 -1.94815591e-01 -1.01776767e+00 -4.93290544e-01
3.74072880e-01 -8.51887584e-01 1.23702395e+00 -4.37325150e-01
-1.80777311e+00 8.74408603e-01 -3.57650161e-01 -4.12590951e-01
1.05398679e+00 -6.47731364e-01 -6.11956418e-02 2.24471893e-02
-1.07998643e-02 5.78789532e-01 7.87450314e-01 -1.63615155e+00
-3.50680560e-01 -5.02321601e-01 3.14453542e-01 7.71568894e-01
3.91665787e-01 -9.27636325e-01 -5.58860600e-01 -2.39416063e-02
7.45810151e-01 -1.33778644e+00 -1.96284503e-01 1.27626002e-01
-4.37132001e-01 8.36554885e-01 7.61360705e-01 -2.52071291e-01
7.62367308e-01 -2.07653880e+00 2.41450995e-01 1.69339672e-01
1.43474430e-01 -2.70960659e-01 3.33534956e-01 4.46664810e-01
3.40957910e-01 -2.91179806e-01 3.07367474e-01 -8.12708259e-01
-1.00717612e-01 3.04432005e-01 -3.06577533e-01 1.08180046e+00
-4.83024985e-01 5.89022994e-01 -9.74682987e-01 -4.97920476e-02
6.69263899e-01 6.32669091e-01 -3.14192265e-01 3.20315659e-02
1.01447076e-01 8.32827985e-01 -5.37032373e-02 8.04806471e-01
8.94371986e-01 5.72724789e-02 2.70475566e-01 -2.72522331e-03
-4.64360029e-01 5.04904032e-01 -1.57001328e+00 2.05317593e+00
-7.57222831e-01 6.65053248e-01 1.93210602e-01 -1.66118458e-01
7.90644884e-01 3.81483920e-02 2.82722443e-01 -8.02370787e-01
2.82954276e-01 4.06035423e-01 -1.97150499e-01 -6.63204938e-02
1.07216966e+00 1.37738183e-01 -9.26833507e-03 6.84721768e-02
-4.57783580e-01 -4.21659082e-01 -2.81167984e-01 1.05490148e-01
1.03713584e+00 4.75249201e-01 7.27361083e-01 -6.85842633e-02
2.08704606e-01 1.54240340e-01 3.11550707e-01 6.32091761e-01
-1.31134316e-01 7.77783573e-01 6.38604909e-02 -1.60292074e-01
-9.81816947e-01 -1.24472225e+00 -9.79257897e-02 2.33470410e-01
8.52149010e-01 -3.87173504e-01 -2.60203421e-01 -2.21757412e-01
1.93745792e-01 4.44523782e-01 -2.97139972e-01 1.20411664e-02
-5.28505862e-01 -1.87288627e-01 3.52029800e-01 1.46614820e-01
3.05960208e-01 -2.71090567e-01 -1.21445501e+00 -3.59892212e-02
-9.32913199e-02 -1.25548303e+00 -1.73484117e-01 -6.18106797e-02
-9.67919052e-01 -9.39249575e-01 -5.14479816e-01 -7.45490491e-02
6.58720315e-01 7.51604497e-01 8.37017417e-01 -2.26540387e-01
1.44154519e-01 6.75766051e-01 -1.41516358e-01 -2.58321017e-01
-3.90271433e-02 -5.90280443e-02 5.07282555e-01 -2.59910077e-01
-1.68095857e-01 -5.91360867e-01 -7.33691752e-01 4.96146053e-01
-3.43045712e-01 1.89240903e-01 3.16791415e-01 2.47924238e-01
5.38555086e-01 -4.30295020e-01 -2.18756661e-01 -4.36768144e-01
1.39396101e-01 -3.17402661e-01 -9.45974708e-01 -5.57960868e-01
-5.92043221e-01 -6.17194027e-02 1.35884866e-01 -4.82651085e-01
-7.76800513e-01 3.00207436e-01 9.37325731e-02 -4.76747334e-01
-7.21100196e-02 4.00447577e-01 -1.08980924e-01 -3.04722905e-01
8.60502183e-01 -1.13337614e-01 5.77619337e-02 -2.69084662e-01
4.81412441e-01 5.87260246e-01 5.56428313e-01 -3.17466229e-01
1.17974877e+00 1.08154500e+00 4.45264697e-01 -1.06768715e+00
-1.11162357e-01 -5.79743028e-01 -8.08359385e-01 -1.62600636e-01
5.72871268e-01 -1.32852852e+00 -1.00631714e+00 3.10617030e-01
-1.08592415e+00 -2.64900118e-01 -1.88330248e-01 5.83096564e-01
-6.91821873e-01 4.22997296e-01 -1.37405142e-01 -9.92186368e-01
1.22494906e-01 -1.34138572e+00 1.51972115e+00 1.89121261e-01
-3.53324920e-01 -8.59632671e-01 3.88022870e-01 -5.25934771e-02
1.41435996e-01 5.19676328e-01 -8.27906132e-02 1.97623074e-01
-9.45082366e-01 -3.05361360e-01 1.18705161e-01 -5.42251647e-01
2.18397737e-01 -4.24574912e-02 -9.98546779e-01 -6.59638703e-01
-7.77828172e-02 -6.67906180e-02 2.95621663e-01 1.70392543e-01
-4.99614477e-02 -5.19036278e-02 -3.33644658e-01 7.99024165e-01
1.49345124e+00 1.03827633e-01 7.03808010e-01 8.57388079e-01
1.01586080e+00 2.88365245e-01 1.14246058e+00 4.63402629e-01
7.83082724e-01 1.32321560e+00 9.59610045e-01 -1.02797948e-01
7.90263116e-02 -4.13160503e-01 6.07212961e-01 5.03434658e-01
-2.77024567e-01 -1.21662237e-01 -8.42917681e-01 2.22086698e-01
-1.76589811e+00 -4.01562840e-01 -4.39033389e-01 2.94106293e+00
3.51562768e-01 2.02470139e-01 -1.00821652e-01 6.14296310e-02
5.16825505e-02 1.72020763e-01 -4.86606151e-01 -2.65539348e-01
1.28796026e-01 -2.08619550e-01 1.10848546e+00 1.21616173e+00
-7.48694897e-01 9.16566074e-01 5.09458399e+00 -3.45273986e-02
-1.48727190e+00 8.95535946e-02 -3.42134118e-01 -2.32896373e-01
-1.01449840e-01 3.99627864e-01 -9.07477379e-01 1.31651014e-01
9.28069949e-01 1.90655217e-01 2.47238636e-01 1.03033316e+00
4.23959434e-01 -6.54894531e-01 -1.02694285e+00 1.21788418e+00
-2.14873590e-02 -8.17111135e-01 -5.73644459e-01 4.48810458e-01
5.65984905e-01 2.56224990e-01 1.28592581e-01 5.31564914e-02
1.48475468e-01 -7.54691660e-01 9.90964174e-01 3.46575469e-01
1.12278879e+00 -5.66411674e-01 5.15342176e-01 6.50188625e-01
-1.20915174e+00 3.34782809e-01 -1.38362095e-01 -5.49861848e-01
3.58973861e-01 4.92846221e-01 -1.14765179e+00 8.14856708e-01
2.26705506e-01 6.25299335e-01 -6.19729578e-01 8.75694036e-01
-3.49375516e-01 -1.16883323e-01 -5.86501300e-01 4.08988416e-01
-7.84654841e-02 -1.93878666e-01 8.81508231e-01 6.22846186e-01
6.14387453e-01 -2.99190313e-01 1.76811919e-01 2.91422129e-01
3.17634195e-01 -2.54576176e-01 -1.06011045e+00 4.57446247e-01
5.43491125e-01 9.67597306e-01 -5.84200084e-01 -9.95620266e-02
-1.49300322e-01 1.27712512e+00 3.42537686e-02 4.92772073e-01
-8.18960905e-01 -2.75231954e-02 1.08059943e+00 3.24387372e-01
-4.51938733e-02 -1.12003660e+00 -5.07686198e-01 -1.53846073e+00
1.77110270e-01 -6.93311155e-01 -1.68754101e-01 -1.01730371e+00
-2.46400356e-01 4.40530121e-01 4.07195650e-02 -1.87170744e+00
-6.75280690e-01 -5.00837803e-01 1.36477100e-02 8.56412053e-01
-1.54677331e+00 -1.04519296e+00 -7.87258685e-01 3.00014764e-01
3.41213018e-01 4.80001509e-01 6.80187404e-01 2.15224728e-01
2.37351693e-02 2.27870852e-01 -1.06839083e-01 -5.23254693e-01
1.03874457e+00 -1.26743102e+00 6.05925739e-01 9.70913827e-01
2.84602493e-01 9.56570029e-01 1.21888876e+00 -8.26006413e-01
-1.88477516e+00 -4.86689270e-01 7.53183246e-01 -9.81190860e-01
4.53214854e-01 -7.07149088e-01 -3.72872144e-01 1.01161432e+00
-2.94389099e-01 1.13300852e-01 -2.08589852e-01 -8.70286375e-02
-4.71369214e-02 -1.14568789e-02 -1.06071734e+00 8.20963025e-01
1.10821605e+00 -4.95652974e-01 -1.86264172e-01 5.73379584e-02
6.17355764e-01 -1.43573260e+00 -4.06636715e-01 4.02449012e-01
8.68663490e-01 -1.16334987e+00 1.11232901e+00 1.73598439e-01
-4.24136966e-01 -8.50199819e-01 -3.41370106e-01 -1.10679090e+00
2.52650619e-01 -9.26421821e-01 -4.04687017e-01 7.31284738e-01
-6.24638908e-02 -7.54819214e-01 8.25503767e-01 2.62073129e-01
1.70107111e-01 -1.83683947e-01 -1.23697186e+00 -8.79016697e-01
-5.58860719e-01 -6.67438388e-01 2.54091173e-01 8.31138253e-01
-2.66419172e-01 3.88796151e-01 -7.00733662e-01 7.92232096e-01
6.38286769e-01 -7.59630501e-02 1.66793346e+00 -1.18058014e+00
-3.35107982e-01 2.22295374e-01 -7.84725726e-01 -1.67260671e+00
-3.46002400e-01 -2.85135508e-01 9.98789147e-02 -1.25845242e+00
-5.14231145e-01 -4.53899115e-01 4.85488057e-01 -8.29168335e-02
2.21082404e-01 2.74257779e-01 1.54620811e-01 3.99380952e-01
-2.95626372e-01 3.86640996e-01 9.33586776e-01 3.75310540e-01
-5.50794899e-01 1.04864150e-01 -1.19670302e-01 8.11554253e-01
4.97154742e-01 -3.85511488e-01 -8.07583034e-01 -4.10575330e-01
7.02144563e-01 3.24173361e-01 6.04425728e-01 -1.21262109e+00
2.22565070e-01 -1.04251683e-01 2.36707225e-01 -7.23127484e-01
7.45675862e-01 -1.01351559e+00 5.30528605e-01 5.34684777e-01
4.82860863e-01 3.59377861e-01 1.66242957e-01 4.08033162e-01
1.31845772e-01 1.20847829e-01 3.96950781e-01 7.68528283e-02
-6.17686272e-01 7.91295543e-02 -2.14119360e-01 -3.57591450e-01
8.36879909e-01 -6.15577996e-01 -4.66985762e-01 -8.58428836e-01
-5.45522451e-01 -6.04446158e-02 1.46882057e+00 3.83982420e-01
5.17136931e-01 -1.20861423e+00 -2.69750003e-02 2.86899209e-01
3.85703444e-01 2.99753487e-01 1.18304446e-01 1.19307518e+00
-9.47872877e-01 4.25184429e-01 -1.80627648e-02 -1.12869167e+00
-1.23164344e+00 2.67628878e-01 2.01877564e-01 -3.72824557e-02
-5.16073108e-01 2.75689393e-01 -3.30474600e-02 -5.06886184e-01
1.18155286e-01 -5.40785015e-01 2.90956181e-02 -1.79309204e-01
2.87398130e-01 3.58557880e-01 2.42902696e-01 -9.30643499e-01
-5.11059105e-01 9.01529372e-01 3.40394288e-01 -6.86843991e-01
7.73405075e-01 -8.65135789e-01 4.48791653e-01 7.70058870e-01
9.51281190e-01 8.16545010e-01 -1.47037029e+00 1.50070593e-01
-2.08211854e-01 -6.18386626e-01 -1.38394699e-01 -4.42418784e-01
-2.68516332e-01 6.94179237e-01 9.32570219e-01 -2.57262707e-01
7.45820642e-01 -4.07132328e-01 2.28580177e-01 5.36880136e-01
9.97018754e-01 -6.97249115e-01 -9.75230709e-02 7.32797801e-01
7.87218571e-01 -1.38205004e+00 3.83666694e-01 -5.53179383e-01
-4.78076547e-01 8.90688360e-01 4.89881545e-01 -8.16375092e-02
1.10879168e-01 4.77689862e-01 4.07555401e-01 -6.34828061e-02
-3.09680820e-01 -2.92804837e-01 1.28561348e-01 8.42388451e-01
2.10202783e-01 -2.43663345e-03 -1.57427177e-01 -2.58924246e-01
-4.91217226e-01 -2.30390295e-01 9.14359152e-01 1.30988097e+00
-3.73779982e-01 -1.03663552e+00 -7.15559721e-01 -4.64968920e-01
1.48026735e-01 2.09234476e-01 -3.62504244e-01 1.31458032e+00
-4.00689505e-02 7.60006189e-01 1.25864558e-02 -5.05215287e-01
4.00293618e-01 -2.84877568e-01 6.31540895e-01 -5.35033047e-01
-1.73625425e-02 1.64743438e-01 3.01289618e-01 -1.08676314e+00
-2.51762569e-01 -6.08746886e-01 -1.48495650e+00 -3.81032526e-01
-8.74579623e-02 -5.07459223e-01 1.15373373e+00 6.73799038e-01
5.33193648e-01 2.45380491e-01 3.19263160e-01 -1.34370553e+00
-2.98949242e-01 -7.89073229e-01 -4.29968923e-01 1.30665869e-01
9.24190700e-01 -9.99287009e-01 -4.43043411e-01 -3.27178895e-01] | [7.424172878265381, -2.1401848793029785] |
345472c8-72cd-4ba3-9c65-dbc3cbc335a5 | non-convex-approaches-for-low-rank-tensor | 2303.12721 | null | https://arxiv.org/abs/2303.12721v1 | https://arxiv.org/pdf/2303.12721v1.pdf | Non-convex approaches for low-rank tensor completion under tubal sampling | Tensor completion is an important problem in modern data analysis. In this work, we investigate a specific sampling strategy, referred to as tubal sampling. We propose two novel non-convex tensor completion frameworks that are easy to implement, named tensor $L_1$-$L_2$ (TL12) and tensor completion via CUR (TCCUR). We test the efficiency of both methods on synthetic data and a color image inpainting problem. Empirical results reveal a trade-off between the accuracy and time efficiency of these two methods in a low sampling ratio. Each of them outperforms some classical completion methods in at least one aspect. | ['Yifei Lou', 'HanQin Cai', 'Longxiu Huang', 'Zheng Tan'] | 2023-03-17 | null | null | null | null | ['image-inpainting'] | ['computer-vision'] | [-1.71486780e-01 -3.88516754e-01 3.01467180e-02 -6.67069852e-02
-9.04791415e-01 -4.86332357e-01 3.64597291e-01 -2.63660401e-01
-3.38465065e-01 7.23409355e-01 1.67701736e-01 -2.01306850e-01
-2.47038156e-01 -4.56306458e-01 -6.58519089e-01 -5.50480306e-01
-2.48876780e-01 1.63373098e-01 -8.00050721e-02 -2.88701326e-01
2.56092817e-01 2.55026668e-01 -1.03056848e+00 7.79761225e-02
1.06634653e+00 1.09978271e+00 -6.85574114e-02 5.00754714e-01
-5.92925437e-02 6.92906976e-01 -1.41082495e-01 -7.93781400e-01
6.04523957e-01 -9.24985185e-02 -9.56789613e-01 3.71194214e-01
2.32970923e-01 -3.16197097e-01 -2.65939265e-01 1.03165054e+00
1.29585028e-01 2.52317995e-01 4.29424822e-01 -1.11389065e+00
-5.36152959e-01 3.59580487e-01 -9.87178385e-01 1.16468564e-01
4.46458489e-01 -8.89887437e-02 1.05874276e+00 -1.34992790e+00
6.54585540e-01 1.32412624e+00 8.95887375e-01 -2.50423830e-02
-1.41550612e+00 -3.52041215e-01 -3.23118828e-02 1.67917863e-01
-1.48785877e+00 -3.06327313e-01 8.90436172e-01 -4.99983758e-01
1.77652448e-01 3.63379449e-01 3.42669040e-01 5.96045494e-01
1.83069199e-01 9.32444274e-01 1.60541570e+00 -1.42873824e-01
2.65609682e-01 -2.35583708e-01 5.63748218e-02 7.41082549e-01
3.27663511e-01 -7.45035857e-02 -7.08553135e-01 -4.07086432e-01
8.21054459e-01 8.73751044e-02 -1.81811348e-01 -2.46756896e-01
-1.62538862e+00 7.82132924e-01 3.37768137e-01 1.14893094e-01
-3.66468996e-01 3.91286552e-01 6.13286614e-01 3.86947095e-01
8.10023129e-01 5.45087904e-02 -1.21541083e-01 -2.82353312e-01
-8.46232712e-01 3.73700351e-01 6.28112912e-01 1.19106567e+00
9.48931754e-01 2.63088256e-01 -2.84018796e-02 9.39543366e-01
6.06029071e-02 4.50878620e-01 1.84464410e-01 -1.10187507e+00
7.26851523e-01 3.17666888e-01 3.37914616e-01 -1.26082945e+00
-7.11746514e-02 -3.06660473e-01 -1.01681662e+00 -9.75081846e-02
3.41322184e-01 3.60177048e-02 -4.85664427e-01 1.26354754e+00
5.86731732e-01 3.67861390e-01 -1.64256439e-01 9.75661457e-01
3.22237968e-01 4.65858370e-01 -2.61271477e-01 -2.25549743e-01
1.30391467e+00 -9.94099796e-01 -9.16419506e-01 1.70040652e-01
5.34996569e-01 -1.20678294e+00 1.03851938e+00 6.79802477e-01
-1.27216744e+00 -3.61031413e-01 -8.33189368e-01 -3.08728099e-01
-1.19022913e-01 3.94194156e-01 1.09371281e+00 6.29876256e-01
-9.48932946e-01 8.85743737e-01 -9.40130174e-01 -1.28780618e-01
1.91469982e-01 8.62769932e-02 -4.26822960e-01 -6.48289740e-01
-6.32864118e-01 4.80309576e-01 -8.12365338e-02 4.12231684e-01
-1.03231657e+00 -6.60028458e-01 -5.28524101e-01 -3.45791578e-01
4.18363303e-01 -5.13334274e-01 9.86593246e-01 -5.23333549e-01
-1.47793412e+00 7.20169425e-01 -1.92110211e-01 -1.68347344e-01
7.94428885e-01 -4.48837996e-01 -1.10071927e-01 7.45135844e-02
2.71228075e-01 1.56074598e-01 1.21712577e+00 -1.14681280e+00
-2.96718508e-01 -4.08221096e-01 1.10812165e-01 1.75496578e-01
-2.07796395e-01 4.89711910e-02 -5.25224209e-01 -1.07618761e+00
3.67501736e-01 -9.67161894e-01 -3.74311000e-01 1.63775548e-01
-5.81010818e-01 -2.70253748e-01 7.17460752e-01 -9.45367694e-01
1.01431155e+00 -2.09945440e+00 4.88368720e-01 9.25589427e-02
4.32047248e-01 6.79705814e-02 -1.26232982e-01 7.13700116e-01
-5.23906350e-02 6.52723163e-02 -4.70829993e-01 -7.72299528e-01
-3.47573496e-02 4.05838013e-01 -3.51039082e-01 8.36985409e-01
8.41856599e-02 4.40484434e-01 -9.95887637e-01 -5.44298351e-01
4.77104485e-02 4.45726663e-01 -6.30820274e-01 2.64057398e-01
-1.86911207e-02 4.74970013e-01 -5.56394696e-01 9.52170253e-01
1.04059708e+00 4.70062532e-02 1.06422186e-01 -4.81303126e-01
-2.99289405e-01 -1.68920937e-03 -1.54272556e+00 1.98821998e+00
-2.77538002e-01 3.57152522e-01 4.56988871e-01 -1.12874210e+00
8.93479407e-01 2.73389548e-01 6.14541531e-01 -1.84836790e-01
5.51881306e-02 4.80588108e-01 -4.81631398e-01 -4.90054458e-01
8.18629026e-01 -1.23422913e-01 2.48047397e-01 7.26757407e-01
-2.39926711e-01 -1.95068941e-01 5.38802087e-01 3.02201629e-01
1.05991256e+00 1.17103271e-01 -1.04142882e-01 -5.51183164e-01
2.86674798e-01 -1.70677572e-01 7.31684983e-01 4.02588964e-01
-2.00398386e-01 7.90344000e-01 6.93147838e-01 -3.76358926e-01
-1.20929372e+00 -9.66085553e-01 -4.51528616e-02 9.81956482e-01
6.64604008e-02 -6.53699100e-01 -5.59679627e-01 -4.49729264e-01
-2.11386289e-02 2.24835768e-01 -7.15588927e-01 3.07048887e-01
-6.11221194e-01 -7.39586055e-01 3.68534297e-01 3.39551598e-01
5.83930016e-01 -3.76072496e-01 -1.30364165e-01 -2.03594510e-02
-4.13769722e-01 -1.28405011e+00 -8.36746454e-01 -3.52179170e-01
-1.19147301e+00 -1.21050489e+00 -9.05983806e-01 -3.64100397e-01
9.76011515e-01 7.58711338e-01 1.05304003e+00 2.26370916e-02
-1.74254522e-01 5.12070715e-01 -6.38770938e-01 9.86164659e-02
7.74162486e-02 -3.83254327e-02 1.56720534e-01 6.13690495e-01
-1.96374834e-01 -6.16173983e-01 -8.39482307e-01 5.08562744e-01
-1.26583028e+00 2.31139478e-03 5.24326921e-01 8.42079103e-01
6.35463536e-01 -1.19896285e-01 -1.91339374e-01 -7.52983630e-01
1.01699519e+00 -5.11269927e-01 -5.73994339e-01 1.19511768e-01
-4.40422267e-01 2.86311269e-01 6.57880664e-01 -3.83172005e-01
-8.00133705e-01 2.23614909e-02 1.83861777e-01 -8.69072378e-01
6.77240133e-01 9.18962598e-01 2.42258742e-01 -2.53217846e-01
5.43939412e-01 2.71407396e-01 3.94078530e-02 -8.82462502e-01
3.88292700e-01 1.12242304e-01 3.13752621e-01 -1.15490103e+00
1.04452157e+00 9.44713831e-01 2.18663350e-01 -9.34541225e-01
-6.49876177e-01 -4.11128700e-01 -4.86043334e-01 -2.07512662e-01
3.12396169e-01 -6.95776105e-01 -8.18344474e-01 3.62271786e-01
-1.06966650e+00 -9.01404023e-03 -2.35798612e-01 6.78960919e-01
-5.19634128e-01 8.35899234e-01 -7.23419309e-01 -7.03354239e-01
-2.68123895e-01 -1.38691151e+00 1.03647113e+00 -4.12194699e-01
4.23396111e-01 -9.10438061e-01 2.75654048e-01 2.86546916e-01
4.66408432e-01 5.61000884e-01 3.33427370e-01 1.69708177e-01
-6.60678327e-01 -1.91961989e-01 -6.79231822e-01 6.20284855e-01
-1.97879389e-01 2.26300612e-01 -5.58384955e-01 -5.96181870e-01
2.45723482e-02 -3.68594825e-01 7.46903539e-01 1.88728482e-01
1.29175818e+00 -3.80827665e-01 2.58098822e-02 8.79802465e-01
1.60442102e+00 -2.24249855e-01 8.52025092e-01 -4.11697254e-02
7.79190302e-01 5.53054810e-01 6.20431006e-01 7.51446426e-01
2.83122659e-01 6.96149945e-01 4.08263177e-01 -1.29733430e-02
1.50174228e-02 -1.03779249e-01 2.85216630e-01 1.38941908e+00
-6.42599702e-01 5.62452555e-01 -8.38452399e-01 6.42212689e-01
-2.03755260e+00 -6.25444233e-01 -3.33214283e-01 2.31120229e+00
8.67734313e-01 -1.88660190e-01 2.61530817e-01 2.74591774e-01
6.11206949e-01 1.64186180e-01 -1.74820811e-01 -4.81244415e-01
4.16487753e-02 3.98996681e-01 7.00492024e-01 5.08559763e-01
-9.45214987e-01 7.05330074e-01 7.12092113e+00 9.46839094e-01
-7.68111348e-01 2.53685981e-01 5.20595074e-01 2.44258180e-01
-5.17880261e-01 3.97349894e-01 -4.56949443e-01 3.42306763e-01
5.48249125e-01 -1.94850504e-01 9.00478661e-01 9.60016727e-01
2.61121124e-01 -7.01431930e-02 -9.51542437e-01 1.23344815e+00
2.42608897e-02 -1.41350281e+00 5.86631000e-02 9.00993347e-02
6.97815299e-01 -1.38580218e-01 1.98264495e-02 1.18435316e-01
4.18285340e-01 -9.13377523e-01 5.87384224e-01 4.84730542e-01
8.87231708e-01 -6.23202980e-01 4.24334019e-01 -1.64869785e-01
-1.58751357e+00 3.08773726e-01 -4.88501102e-01 -9.49926376e-02
6.25451952e-02 1.04958117e+00 -4.12961781e-01 9.32771921e-01
7.61380851e-01 9.14278150e-01 -3.71448755e-01 1.03588533e+00
-1.16260573e-01 5.33925235e-01 -4.19758022e-01 1.95594475e-01
3.26915622e-01 -8.83535504e-01 7.53050625e-01 1.04899061e+00
2.84796894e-01 1.07058302e-01 2.59863108e-01 8.43491197e-01
-1.57890067e-01 1.08545274e-01 -6.21396840e-01 4.17403728e-02
3.06011140e-01 1.19331896e+00 -5.86565614e-01 -9.42711234e-02
-3.81583899e-01 8.96817327e-01 3.01658571e-01 4.84995961e-01
-7.95257926e-01 -4.81339842e-01 7.33697534e-01 -3.88538949e-02
2.24052429e-01 -9.25513744e-01 -3.35409105e-01 -1.50319922e+00
3.68012816e-01 -6.75026059e-01 3.19509387e-01 -5.96463859e-01
-1.38573289e+00 4.83611941e-01 2.11176366e-01 -1.54093599e+00
1.78372905e-01 -7.83812404e-01 -3.33172619e-01 6.72812223e-01
-1.55604470e+00 -1.08492458e+00 -4.21901107e-01 1.00960815e+00
3.87029320e-01 1.78537533e-01 5.41159153e-01 5.85879982e-01
-9.03481841e-01 5.64538240e-01 5.79777777e-01 -1.70884412e-02
4.05255705e-01 -1.23397493e+00 1.77882209e-01 9.95316505e-01
-2.19120502e-01 7.33537078e-01 7.42060006e-01 -3.71810049e-01
-2.16891503e+00 -9.93135631e-01 5.34711123e-01 -1.22309297e-01
7.73216903e-01 -1.67252705e-01 -6.36502981e-01 8.01393569e-01
1.17122240e-01 2.35285103e-01 5.79014242e-01 2.18161285e-01
-6.47759795e-01 -4.66707796e-01 -1.10513926e+00 6.79187596e-01
8.11823785e-01 -4.87580955e-01 -2.42511868e-01 5.66397250e-01
5.26806712e-01 -3.05148333e-01 -1.27668476e+00 2.30901599e-01
3.77638191e-01 -1.03537583e+00 1.11168492e+00 -4.22350705e-01
7.33830035e-01 -3.72196287e-01 -5.57659924e-01 -1.18854856e+00
-1.37626603e-01 -1.15911531e+00 -5.25902249e-02 9.45922136e-01
-8.44433531e-02 -6.96352184e-01 6.41326070e-01 6.13086581e-01
-2.29672775e-01 -6.97702825e-01 -1.11998379e+00 -1.03667152e+00
-8.05174038e-02 -7.07254052e-01 4.31480825e-01 1.11874795e+00
-4.45514079e-03 8.24303702e-02 -5.84975243e-01 -2.01546982e-01
1.06307137e+00 1.17107935e-01 1.00450695e+00 -7.91465402e-01
-2.31223345e-01 -1.62197053e-01 -1.06224924e-01 -1.13833785e+00
-3.35540295e-01 -7.20552206e-01 -3.26910347e-01 -1.15869892e+00
1.71955317e-01 -7.67963886e-01 -9.47157219e-02 2.84699708e-01
-2.90697217e-02 4.05221611e-01 1.17949903e-01 5.94105542e-01
-5.06520927e-01 8.77997994e-01 1.55514598e+00 -5.08085042e-02
4.02186215e-02 -2.22172260e-01 -3.40287119e-01 5.90386391e-01
4.99155015e-01 -3.36592704e-01 -1.76093146e-01 -8.12655985e-01
3.60103011e-01 3.20614845e-01 1.76719666e-01 -7.41160572e-01
-4.89968210e-02 -3.85805279e-01 -2.45201126e-01 -6.68942153e-01
4.62117851e-01 -6.67829394e-01 1.73889831e-01 2.89943099e-01
-1.20768778e-01 4.76692080e-01 -1.16653912e-01 7.00625658e-01
-2.89534926e-01 -1.59652039e-01 7.19888330e-01 -1.21484563e-01
-2.71179736e-01 5.79343259e-01 -3.85981649e-02 1.28476515e-01
8.51889133e-01 -1.40839204e-01 -1.60910904e-01 -2.90655524e-01
-5.46750546e-01 3.20393592e-02 2.09873080e-01 7.47711584e-02
8.21942508e-01 -1.63165903e+00 -9.13845718e-01 -3.88585515e-02
-1.47785209e-02 6.71728253e-02 1.58789009e-01 1.36341715e+00
-1.02505326e+00 3.24953616e-01 -4.47919704e-02 -6.54666901e-01
-7.43848920e-01 6.80620253e-01 2.61597335e-02 -5.68242490e-01
-4.95874971e-01 7.17758894e-01 -5.22838719e-02 -3.48356187e-01
1.58394009e-01 -4.36661601e-01 2.80658156e-01 -2.41996676e-01
4.73788589e-01 9.63832438e-01 4.35470007e-02 -4.95335728e-01
-1.84166655e-01 5.57133138e-01 -9.75559428e-02 -3.85460034e-02
1.35557747e+00 -2.18284931e-02 -5.61626017e-01 2.96365559e-01
1.34443307e+00 -6.57427758e-02 -1.17196000e+00 -4.36555833e-01
-1.92316428e-01 -1.10990608e+00 -2.36151386e-02 -1.09134026e-01
-1.42542863e+00 6.08591735e-01 3.17303836e-01 4.82106507e-01
1.10918939e+00 -4.78727102e-01 9.32517767e-01 1.25736415e-01
4.35575098e-01 -1.09226263e+00 3.00076425e-01 3.81551892e-01
1.26070487e+00 -1.05449247e+00 4.12307173e-01 -8.27297747e-01
-4.49179798e-01 1.04015517e+00 2.10029751e-01 -6.20332956e-01
8.05437505e-01 -1.86441839e-01 -4.74349678e-01 -2.43809447e-01
-5.83326459e-01 -7.95850903e-02 8.96533728e-02 3.78701478e-01
5.49510717e-01 2.40347758e-01 -9.15678501e-01 -5.12736067e-02
-7.86869898e-02 2.40935180e-02 5.56496501e-01 1.10554433e+00
1.93281651e-01 -1.25335598e+00 -7.74096668e-01 3.12347442e-01
-4.42433298e-01 -7.04780454e-04 -1.00350708e-01 6.72153413e-01
-2.95299619e-01 8.73695374e-01 -3.21907997e-01 -5.30408800e-01
2.10813507e-01 -4.05528992e-01 3.76354009e-01 -1.22956172e-01
-4.17286158e-01 2.13198408e-01 8.80258009e-02 -7.28453219e-01
-6.40375674e-01 -7.49726593e-01 -5.71823657e-01 -9.59684849e-01
-2.10124850e-01 1.80285886e-01 7.31166065e-01 8.07574213e-01
3.44426304e-01 1.77365333e-01 1.05744469e+00 -8.93509626e-01
-6.88452899e-01 -8.92141700e-01 -9.25468862e-01 6.72663510e-01
1.15090773e-01 -8.18518221e-01 -2.59371430e-01 1.36855558e-01] | [7.369930267333984, 4.514855861663818] |
5730d1b9-a348-45a6-b022-9066e2d78021 | openicl-an-open-source-framework-for-in | 2303.02913 | null | https://arxiv.org/abs/2303.02913v1 | https://arxiv.org/pdf/2303.02913v1.pdf | OpenICL: An Open-Source Framework for In-context Learning | In recent years, In-context Learning (ICL) has gained increasing attention and emerged as the new paradigm for large language model (LLM) evaluation. Unlike traditional fine-tuning methods, ICL instead adapts the pre-trained models to unseen tasks without any parameter updates. However, the implementation of ICL is sophisticated due to the diverse retrieval and inference methods involved, as well as the varying pre-processing requirements for different models, datasets, and tasks. A unified and flexible framework for ICL is urgently needed to ease the implementation of the aforementioned components. To facilitate ICL research, we introduce OpenICL, an open-source toolkit for ICL and LLM evaluation. OpenICL is research-friendly with a highly flexible architecture that users can easily combine different components to suit their needs. It also provides various state-of-the-art retrieval and inference methods to streamline the process of adapting ICL to cutting-edge research. The effectiveness of OpenICL has been validated on a wide range of NLP tasks, including classification, QA, machine translation, and semantic parsing. As a side-product, we found OpenICL to be an efficient yet robust tool for LLMs evaluation. OpenICL is released at https://github.com/Shark-NLP/OpenICL | ['Zhiyong Wu', 'Yu Qiao', 'Jingjing Xu', 'Jiangtao Feng', 'Jiacheng Ye', 'Yaoxiang Wang', 'Zhenyu Wu'] | 2023-03-06 | null | null | null | null | ['semantic-parsing'] | ['natural-language-processing'] | [-8.23930800e-02 -4.20264274e-01 -3.92771661e-01 -5.79552531e-01
-1.37131476e+00 -9.68309224e-01 5.47180057e-01 1.60185784e-01
-5.19792855e-01 5.58231592e-01 1.18683912e-01 -5.13817012e-01
3.45631763e-02 -3.69938821e-01 -6.44867122e-01 -2.26065442e-01
5.53051889e-01 8.18247855e-01 1.70003757e-01 -5.98222613e-02
1.98006630e-01 2.90455073e-01 -1.30647755e+00 7.44184971e-01
9.01737630e-01 6.82004869e-01 7.18057871e-01 5.48442185e-01
-5.16184330e-01 2.59181589e-01 -5.38974345e-01 -6.29836440e-01
-2.23966256e-01 -5.89633323e-02 -9.82876003e-01 -6.15401030e-01
5.15694439e-01 1.26147613e-01 1.65395677e-01 8.28722119e-01
8.78528476e-01 2.30834946e-01 3.23289424e-01 -9.83669996e-01
-6.71329260e-01 7.00145125e-01 1.64597252e-04 1.52087659e-01
4.39402848e-01 4.06977236e-02 1.02961147e+00 -1.28887343e+00
6.31505013e-01 1.52702188e+00 4.87669885e-01 6.61191642e-01
-1.02643931e+00 -6.48411870e-01 2.39438564e-01 4.45304871e-01
-1.13811898e+00 -5.24067342e-01 4.63677049e-01 -1.25837103e-01
1.29696739e+00 2.90407062e-01 1.55463979e-01 1.44730246e+00
1.81708381e-01 1.13227654e+00 9.99430954e-01 -7.98427403e-01
2.31843069e-01 2.02837333e-01 1.23541065e-01 4.96277958e-01
8.16330388e-02 -2.38941684e-01 -4.43139881e-01 -3.28055650e-01
2.86399990e-01 -1.73014283e-01 3.30777392e-02 -7.17153400e-02
-1.22005665e+00 8.26308370e-01 -1.50259817e-02 5.10203421e-01
-4.33038659e-02 7.89983124e-02 6.32809937e-01 2.82565415e-01
4.86762583e-01 6.35365427e-01 -9.51252222e-01 -3.66730928e-01
-8.34671021e-01 1.38660654e-01 8.23524237e-01 8.97822857e-01
5.44792235e-01 -4.01177019e-01 -3.50366443e-01 1.31027234e+00
3.86296809e-01 5.51412046e-01 7.11889982e-01 -1.07737875e+00
6.98910654e-01 5.41958690e-01 -2.12613657e-01 -4.30487365e-01
-2.73262501e-01 -5.54177701e-01 -5.73287666e-01 -4.65069503e-01
1.22916341e-01 4.10430133e-02 -7.76712775e-01 1.81971753e+00
2.82582134e-01 -8.03960785e-02 8.71906988e-03 5.54980993e-01
9.96215284e-01 8.58861864e-01 3.63096476e-01 2.60650963e-02
1.36681974e+00 -1.20743966e+00 -7.12000549e-01 -6.59321308e-01
1.00363457e+00 -1.25274253e+00 1.63839376e+00 1.38266996e-01
-1.08030343e+00 -4.30267721e-01 -6.89316869e-01 -3.36949617e-01
-7.64037967e-01 2.88739383e-01 7.09989786e-01 4.64110732e-01
-1.05600226e+00 1.87404871e-01 -7.44533002e-01 -4.61780220e-01
1.03359900e-01 1.46675393e-01 -1.75263062e-01 -5.43487847e-01
-1.57049894e+00 1.16345251e+00 5.23103952e-01 1.40776038e-01
-7.56256044e-01 -7.68480301e-01 -9.05286610e-01 -4.59856279e-02
4.62560982e-01 -8.75147998e-01 1.66855383e+00 -5.18475235e-01
-1.54086053e+00 9.04916644e-01 -3.82872999e-01 -4.10431530e-03
2.22337663e-01 -3.65474403e-01 -4.76369530e-01 -2.49789700e-01
2.50345558e-01 5.79757452e-01 5.58555007e-01 -8.42216253e-01
-3.02015096e-01 -2.27778390e-01 -2.95152552e-02 3.95560235e-01
-1.76126078e-01 4.27583039e-01 -9.96095181e-01 -5.03392756e-01
-1.90142319e-01 -8.74068797e-01 -1.82123721e-01 -4.53723490e-01
-9.01951045e-02 -5.03143072e-01 7.07872510e-01 -4.95155543e-01
1.54535067e+00 -2.05375743e+00 3.07848845e-02 -2.65494287e-01
-2.59027779e-01 8.40193629e-01 -5.73633730e-01 7.51147568e-01
3.70318621e-01 2.15239763e-01 -1.91145629e-01 -5.69228947e-01
2.31077209e-01 3.00610721e-01 -1.34520993e-01 -9.61462855e-02
1.45570040e-01 1.18424392e+00 -9.74766612e-01 -5.90125322e-01
3.64666373e-01 5.22903800e-01 -5.10665357e-01 2.54059851e-01
-6.54562771e-01 4.70157593e-01 -6.08653307e-01 8.32786977e-01
4.66678590e-01 -4.98352826e-01 1.95821673e-01 5.42437658e-02
-6.55802041e-02 6.56542361e-01 -9.84170377e-01 2.07110882e+00
-9.78296101e-01 4.26651061e-01 6.20693974e-02 -7.79531777e-01
6.36725724e-01 3.99181366e-01 -1.15102112e-01 -7.39852905e-01
-9.86260325e-02 4.87967819e-01 -3.14044863e-01 -4.13863480e-01
4.95426565e-01 2.76812166e-01 -2.76289493e-01 3.64054888e-01
2.58486301e-01 -1.31129384e-01 4.26667899e-01 1.89212918e-01
9.21762049e-01 2.65778601e-01 4.05787468e-01 -6.21950515e-02
6.59203112e-01 -7.56578594e-02 5.46971738e-01 8.54735136e-01
-5.94998710e-02 1.94423035e-01 -3.26286927e-02 -2.09522083e-01
-5.92260540e-01 -1.02412534e+00 -3.38090271e-01 1.66223347e+00
-2.63007373e-01 -5.99092007e-01 -7.86895335e-01 -5.90174973e-01
-1.65605798e-01 8.37459981e-01 -1.00910753e-01 1.39182638e-02
-6.10561907e-01 -8.10143054e-01 6.76755071e-01 2.61391371e-01
5.19609511e-01 -1.45233667e+00 -6.29347470e-03 3.51501673e-01
-6.08008862e-01 -1.38062072e+00 -5.75788021e-01 7.98899829e-02
-9.25987661e-01 -8.48146915e-01 -3.93202990e-01 -1.08771682e+00
3.98024112e-01 2.84284167e-02 1.55577743e+00 -1.17444135e-01
-1.94325835e-01 3.52324694e-01 -4.17259693e-01 -3.17576557e-01
-6.34384751e-01 6.28185272e-01 -1.78770646e-01 -4.75226283e-01
5.25491416e-01 -1.45769835e-01 -4.16802347e-01 3.03147912e-01
-9.13755655e-01 -1.45953614e-03 7.36088395e-01 7.05773413e-01
7.87662685e-01 -2.79654473e-01 6.71294212e-01 -1.05994260e+00
9.28577900e-01 -3.39366853e-01 -6.16608441e-01 8.48465145e-01
-7.63155580e-01 1.63006425e-01 6.12759471e-01 -3.32786441e-01
-1.21802890e+00 -3.13658744e-01 -4.98492211e-01 -8.57458115e-02
-1.66169927e-01 8.39976251e-01 -4.74771947e-01 1.84979141e-01
4.93655562e-01 5.92852151e-03 -4.33154672e-01 -9.08688307e-01
7.54445553e-01 8.79698157e-01 2.98948914e-01 -8.46223056e-01
3.57013315e-01 -1.41382173e-01 -4.62504953e-01 -5.76183558e-01
-1.08549535e+00 -4.78675872e-01 -4.48756397e-01 1.19189359e-01
6.20518923e-01 -8.66418004e-01 -3.59381109e-01 4.15025085e-01
-1.21243346e+00 -5.36273956e-01 7.37145618e-02 4.26825076e-01
-1.78176254e-01 3.58826131e-01 -8.67956221e-01 -4.40144986e-01
-7.70234883e-01 -1.47593057e+00 1.21851051e+00 1.50812551e-01
-2.63322800e-01 -1.37555623e+00 2.51902699e-01 7.06923783e-01
5.90086281e-01 -4.25022930e-01 1.18091285e+00 -6.87554002e-01
-4.52748299e-01 -3.06194603e-01 -1.08791344e-01 4.34741557e-01
-8.30594152e-02 -2.98968609e-02 -1.03229880e+00 -3.71962309e-01
-2.02987939e-01 -6.18047774e-01 6.34556234e-01 3.45897943e-01
1.16576886e+00 -2.27730989e-01 -3.99881095e-01 6.48862958e-01
1.24513626e+00 -1.04359388e-02 4.72378582e-01 6.04157329e-01
5.22979736e-01 3.92739981e-01 8.43841970e-01 -1.77703500e-02
4.30549741e-01 7.59717047e-01 7.31457919e-02 2.98305787e-02
-8.80758464e-02 -3.32304657e-01 7.03361571e-01 1.28329730e+00
3.95237744e-01 -2.95463771e-01 -1.04162574e+00 2.80916184e-01
-1.78965855e+00 -6.43896759e-01 2.08651736e-01 2.12384105e+00
1.16195309e+00 7.87779631e-04 -5.13063014e-01 -5.62404990e-01
5.65767825e-01 1.79447472e-01 -5.32153785e-01 -6.90667033e-01
-1.72641292e-01 3.88099074e-01 2.07999628e-02 6.95776939e-01
-8.77871990e-01 1.49154055e+00 6.01932621e+00 1.32514048e+00
-1.06824315e+00 5.31233847e-01 4.62619454e-01 -5.27401716e-02
-4.28311706e-01 1.83791861e-01 -1.26779819e+00 4.74010885e-01
1.27784848e+00 -2.19667852e-02 4.72691029e-01 9.53686953e-01
1.22210488e-01 5.07443026e-02 -1.03662980e+00 8.48994076e-01
-3.39645520e-02 -1.37563813e+00 2.73808986e-01 -3.71455938e-01
5.52184820e-01 6.18969440e-01 4.08424251e-02 8.10175896e-01
1.99822545e-01 -8.17988038e-01 2.92548835e-01 1.11710273e-01
9.76464152e-01 -4.47084457e-01 8.04634869e-01 4.73038435e-01
-1.08174980e+00 8.89632851e-02 -4.50791717e-01 8.90359879e-02
2.01191857e-01 5.94324946e-01 -7.94954479e-01 4.42227900e-01
6.91170514e-01 4.15752709e-01 -7.31999516e-01 7.92048454e-01
-4.16397214e-01 7.83929884e-01 -1.87069669e-01 1.65113271e-03
2.59216309e-01 -1.10001720e-01 3.35028082e-01 1.64073503e+00
1.02743514e-01 -4.46884364e-01 3.47138137e-01 5.82145631e-01
-2.25034922e-01 4.42385823e-01 -4.32149619e-01 -1.74409464e-01
8.36860359e-01 1.33689475e+00 -2.69473046e-01 -3.97948325e-01
-5.58398783e-01 8.23443592e-01 6.10847414e-01 3.35415363e-01
-6.76772177e-01 -2.29058623e-01 5.97897947e-01 -1.95097551e-01
-2.28987094e-02 -1.58987507e-01 4.53289002e-02 -1.34113097e+00
1.29406542e-01 -1.09501708e+00 6.34320140e-01 -7.96683133e-01
-1.27799785e+00 7.96637893e-01 1.02186747e-01 -8.27149332e-01
-4.55876410e-01 -7.06419349e-01 -3.35494727e-01 8.88567567e-01
-1.69486725e+00 -1.27147019e+00 1.03090219e-02 4.55343068e-01
9.82998729e-01 -2.02325821e-01 1.14152849e+00 5.19737244e-01
-7.39712000e-01 7.36901104e-01 3.56657833e-01 -9.77964401e-02
1.03117621e+00 -9.82409358e-01 6.26951098e-01 7.49657273e-01
2.14436144e-01 8.78068388e-01 2.86042184e-01 -5.63208520e-01
-1.34981620e+00 -1.21270394e+00 1.37884283e+00 -7.62170732e-01
7.64825523e-01 -5.52144051e-01 -8.78093541e-01 7.71855175e-01
1.38086796e-01 -1.41735464e-01 8.82341623e-01 3.97995442e-01
-4.36041534e-01 -9.57608894e-02 -8.07935059e-01 5.98892808e-01
8.03849518e-01 -8.28443646e-01 -6.60110295e-01 6.01545036e-01
9.17771578e-01 -4.76448804e-01 -8.12025130e-01 5.64835668e-01
4.32822019e-01 -5.70968688e-01 9.97282505e-01 -5.02339125e-01
1.51255503e-01 -2.01153085e-02 -4.73118007e-01 -1.20319819e+00
-3.13846290e-01 -6.52392328e-01 -1.65178955e-01 1.33877277e+00
8.49309206e-01 -8.07459593e-01 2.86740720e-01 5.61119378e-01
-3.54102790e-01 -9.85943913e-01 -7.97195852e-01 -6.95402622e-01
2.06527188e-01 -8.89415681e-01 4.77238983e-01 8.12826455e-01
-1.17622212e-01 7.45262265e-01 -6.45890087e-02 9.66203362e-02
4.32245821e-01 2.07774475e-01 5.96727490e-01 -9.69174683e-01
-4.95964348e-01 -4.04715627e-01 1.98395222e-01 -1.12082386e+00
2.87799895e-01 -1.36225545e+00 -1.90430820e-01 -1.61367655e+00
3.44968826e-01 -6.46297276e-01 -4.66338009e-01 6.47768140e-01
-3.96643639e-01 9.20211375e-02 2.76934475e-01 3.76715362e-01
-9.31998312e-01 4.14969862e-01 1.16652071e+00 1.32782105e-02
-2.12576792e-01 1.49921790e-01 -6.63807511e-01 6.43626809e-01
9.93414223e-01 -4.84862506e-01 -4.61531162e-01 -9.18859959e-01
2.73568839e-01 -2.02596664e-01 -2.77008433e-02 -6.62285924e-01
1.42668217e-01 -7.49156028e-02 1.79490387e-01 -6.27615511e-01
2.82364130e-01 -4.13927436e-01 -1.97883951e-03 8.42210874e-02
-3.88637930e-01 3.08463424e-01 4.92434740e-01 1.18695661e-01
-2.92135328e-01 -4.63101208e-01 5.64125180e-01 -4.34175014e-01
-9.14578795e-01 2.06179872e-01 -9.72796381e-02 5.78494310e-01
4.27297026e-01 2.85716653e-01 -4.87673014e-01 -1.09126560e-01
-5.42304397e-01 4.20428157e-01 3.33178908e-01 6.93927765e-01
4.14752752e-01 -1.04494619e+00 -6.21135414e-01 1.59239367e-01
3.26992214e-01 4.57687601e-02 2.60370970e-01 6.58681870e-01
-3.59274924e-01 9.18040574e-01 3.19808364e-01 -5.34851789e-01
-1.17105281e+00 4.09568310e-01 4.84446771e-02 -8.39428961e-01
-3.44235569e-01 6.95228696e-01 1.67906865e-01 -1.16933179e+00
3.92880678e-01 -2.65999258e-01 6.18824288e-02 -1.81012988e-01
4.91800159e-01 1.01466715e-01 4.31319237e-01 -2.95000434e-01
-4.63546485e-01 4.32082832e-01 -3.80613416e-01 -7.39439651e-02
1.10760367e+00 -1.92560494e-01 -2.41385281e-01 5.60498714e-01
1.18433750e+00 -6.89393058e-02 -7.73377061e-01 -4.34691310e-01
3.21859121e-01 -1.09398335e-01 5.21484502e-02 -1.34427214e+00
-5.35807312e-01 8.43958735e-01 3.12328219e-01 -3.92549604e-01
1.05142760e+00 2.60364324e-01 1.00406492e+00 6.98717237e-01
5.77927053e-01 -1.17851722e+00 -1.07469052e-01 9.03018773e-01
8.89282107e-01 -1.31740475e+00 -1.04871869e-01 -2.27420285e-01
-5.77192008e-01 7.58575439e-01 5.11868715e-01 5.81137955e-01
4.45001811e-01 1.50572225e-01 4.59725201e-01 1.35928940e-03
-9.89561439e-01 -9.28043667e-03 4.42009479e-01 3.46633583e-01
8.36546481e-01 9.62532163e-02 -3.51137787e-01 5.53345919e-01
-8.82942826e-02 7.06995800e-02 -7.01895580e-02 8.86920035e-01
-2.44001582e-01 -1.84907687e+00 -2.58446544e-01 2.15142384e-01
-5.88232219e-01 -5.54884255e-01 -2.61857957e-01 7.70647585e-01
-1.02122538e-01 9.52058434e-01 -3.70641798e-01 -2.81255543e-02
2.62563795e-01 5.06208241e-01 3.63001734e-01 -8.81782532e-01
-5.95951557e-01 -5.03905416e-02 1.58663586e-01 -6.98253572e-01
-1.10390402e-01 -4.65494841e-01 -1.20130992e+00 -5.39495796e-02
-3.46548855e-01 3.36086154e-01 1.00652122e+00 8.88555586e-01
7.84140646e-01 1.84363991e-01 2.38461643e-01 -6.23873055e-01
-6.08872175e-01 -1.13789225e+00 -3.29619576e-03 2.25308105e-01
-1.20782875e-01 -5.26646554e-01 -2.36553401e-01 -2.65819699e-01] | [11.031598091125488, 8.687082290649414] |
3f60001c-45ac-498e-ae63-02c3a2c69038 | multi-turn-response-selection-for-chatbots | null | null | https://aclanthology.org/P18-1103 | https://aclanthology.org/P18-1103.pdf | Multi-Turn Response Selection for Chatbots with Deep Attention Matching Network | Human generates responses relying on semantic and functional dependencies, including coreference relation, among dialogue elements and their context. In this paper, we investigate matching a response with its multi-turn context using dependency information based entirely on attention. Our solution is inspired by the recently proposed Transformer in machine translation (Vaswani et al., 2017) and we extend the attention mechanism in two ways. First, we construct representations of text segments at different granularities solely with stacked self-attention. Second, we try to extract the truly matched segment pairs with attention across the context and response. We jointly introduce those two kinds of attention in one uniform neural network. Experiments on two large-scale multi-turn response selection tasks show that our proposed model significantly outperforms the state-of-the-art models. | ['dianhai yu', 'daxiang dong', 'Ying Chen', 'Yi Liu', 'Hua Wu', 'Xiangyang Zhou', 'Lu Li', 'Wayne Xin Zhao'] | 2018-07-01 | null | null | null | acl-2018-7 | ['conversational-response-selection'] | ['natural-language-processing'] | [ 4.67607111e-01 2.31246784e-01 -3.39427859e-01 -5.67099869e-01
-1.20596051e+00 -6.97235227e-01 6.82195425e-01 1.00010909e-01
-5.41146398e-01 8.31817448e-01 7.54513681e-01 -3.45455825e-01
1.26346529e-01 -5.90808749e-01 -6.32927060e-01 -1.58053800e-01
6.38193309e-01 7.78020263e-01 3.03378403e-01 -6.76313519e-01
5.01896203e-01 1.05465457e-01 -9.08626378e-01 8.12167048e-01
9.35639501e-01 7.62942433e-01 2.64789969e-01 6.91390991e-01
-3.56172472e-01 8.29682529e-01 -5.25573730e-01 -7.11844027e-01
-6.87476620e-02 -9.11177635e-01 -1.55127859e+00 -1.52900130e-01
2.75584042e-01 -1.61591649e-01 -3.66523087e-01 6.66613698e-01
6.18667006e-01 2.70020068e-01 5.28893471e-01 -5.89884400e-01
-1.00270867e+00 1.00336909e+00 -5.67536831e-01 3.84862334e-01
8.39616656e-01 -4.11083922e-02 1.26637506e+00 -9.05938447e-01
6.73995674e-01 1.34408915e+00 4.32685077e-01 8.30122769e-01
-1.32944942e+00 -3.90589178e-01 4.39233869e-01 4.56094652e-01
-7.90354073e-01 -4.90625679e-01 8.36162090e-01 -1.15145348e-01
1.25341356e+00 4.99056965e-01 2.91424036e-01 1.42187154e+00
3.06002256e-02 1.12198377e+00 1.01576495e+00 -7.09185779e-01
-2.89366424e-01 -7.05914050e-02 5.21972060e-01 2.92310834e-01
-4.88536179e-01 -2.55272537e-01 -3.95484179e-01 -2.38802820e-01
3.62992436e-01 -4.14677382e-01 -4.54115212e-01 1.16644785e-01
-1.39467525e+00 1.06773543e+00 3.66164774e-01 4.13796812e-01
-4.27112430e-01 -1.17159665e-01 5.05368590e-01 5.14146686e-01
4.82459635e-01 6.04297042e-01 -5.52779675e-01 1.52409179e-02
-4.76709127e-01 2.44277194e-01 1.00860751e+00 1.13462281e+00
6.16927505e-01 -4.18056011e-01 -7.37404168e-01 1.17032027e+00
-3.60559300e-02 1.79204375e-01 6.72504663e-01 -7.06398368e-01
9.82756436e-01 4.18140501e-01 1.54460743e-01 -8.75917435e-01
-4.38909829e-01 -1.62917972e-01 -5.53201020e-01 -6.34187996e-01
4.53584850e-01 -4.01493639e-01 -5.91679633e-01 1.87378025e+00
3.09584677e-01 -1.49704248e-01 8.36910754e-02 1.16054606e+00
1.07761073e+00 4.62539256e-01 9.49620605e-02 -3.20756495e-01
1.32322335e+00 -1.47052598e+00 -7.67117202e-01 -4.03371632e-01
5.62706172e-01 -9.90263343e-01 1.08425498e+00 -3.26975025e-02
-1.28947377e+00 -6.27257943e-01 -7.53140450e-01 -3.61475348e-01
7.42943734e-02 2.00716168e-01 5.14717698e-01 -1.56666376e-02
-8.60018432e-01 5.25317907e-01 -2.76558161e-01 -5.88982642e-01
-7.41048232e-02 3.15043032e-01 -3.45173866e-01 1.70484573e-01
-1.64710593e+00 1.21050084e+00 -3.71755846e-02 2.74453104e-01
-3.02469403e-01 -7.74345547e-02 -6.63482308e-01 5.44160232e-02
4.88862336e-01 -8.97780061e-01 1.73162234e+00 -1.28121889e+00
-1.99341846e+00 9.57373261e-01 -4.42583025e-01 -3.42032820e-01
3.93033117e-01 -3.60455126e-01 -2.19472691e-01 3.00439924e-01
1.34756923e-01 4.99825776e-01 5.45245171e-01 -9.17976379e-01
-5.09400904e-01 -2.78536409e-01 3.91824573e-01 3.75805348e-01
-2.75421131e-04 4.85419571e-01 -4.18618292e-01 -5.35621464e-01
1.69776127e-01 -1.09961665e+00 -1.75993159e-01 -7.97090352e-01
-5.22210538e-01 -6.65298820e-01 2.57471412e-01 -7.15125084e-01
1.26452959e+00 -1.51936054e+00 5.85112751e-01 -3.11913311e-01
-3.46404351e-02 2.05016062e-01 -4.15622085e-01 8.76812279e-01
-1.05856761e-01 1.34215236e-01 -2.70178974e-01 -2.63329148e-01
-2.16268674e-02 -2.83527728e-02 -5.36741614e-01 2.60927051e-01
4.75680649e-01 1.17566967e+00 -8.61125946e-01 -4.18857902e-01
-3.22631687e-01 4.30385843e-02 -5.46362579e-01 5.16081452e-01
-1.69773787e-01 7.00367928e-01 -6.52017355e-01 2.77379811e-01
3.76055986e-01 -4.23417017e-02 5.85802197e-01 -8.99705067e-02
-9.15495604e-02 1.08014750e+00 -5.16424894e-01 1.89796412e+00
-6.72381222e-01 4.59735394e-01 1.91980749e-02 -1.14808810e+00
8.42736065e-01 4.65622991e-01 9.72445086e-02 -7.89788008e-01
2.74771929e-01 2.16175109e-01 2.82073915e-01 -7.07715392e-01
6.49279952e-01 -4.25793864e-02 -5.25133312e-01 8.81038368e-01
2.48356014e-01 1.38718814e-01 1.65451601e-01 2.76934445e-01
7.86900043e-01 3.34063560e-01 5.45540869e-01 -2.70466238e-01
7.93785691e-01 -1.10475004e-01 6.16473556e-01 8.05996060e-01
-2.12318853e-01 7.05156982e-01 7.18401074e-01 -3.52316648e-01
-7.14846969e-01 -5.59097350e-01 7.43234977e-02 1.56694770e+00
1.54018328e-01 -1.89422682e-01 -8.56186688e-01 -1.05115187e+00
-2.60392576e-01 6.59994066e-01 -8.60638797e-01 1.99965984e-02
-1.16466928e+00 -4.18917835e-01 3.79279077e-01 7.07624912e-01
2.37921387e-01 -1.49789357e+00 -4.91999716e-01 5.69383681e-01
-8.31208706e-01 -1.06602359e+00 -8.03491592e-01 1.27023652e-01
-5.73508263e-01 -1.01590955e+00 -6.30613685e-01 -8.85643780e-01
3.11695129e-01 3.70297253e-01 1.39086545e+00 1.14872634e-01
3.16385150e-01 1.09274492e-01 -6.69061303e-01 -5.77610843e-02
-3.22220236e-01 5.64859688e-01 -3.39231640e-01 1.70449819e-02
5.29128194e-01 -4.15468067e-01 -4.64635998e-01 2.45890900e-01
-4.64006037e-01 2.05061868e-01 5.57206035e-01 1.16271460e+00
3.78860891e-01 -1.21404350e+00 9.68070030e-01 -1.20954704e+00
1.03327441e+00 -7.64342785e-01 2.34353393e-02 5.05066514e-01
-2.81421274e-01 1.43291622e-01 6.68034673e-01 -5.05740285e-01
-1.27108026e+00 -1.90487370e-01 -2.78529018e-01 7.27884397e-02
-2.43301153e-01 3.81731570e-01 -1.68471530e-01 2.50811547e-01
4.47881848e-01 1.75980777e-01 -4.71876681e-01 -5.20347476e-01
5.66573620e-01 9.28093731e-01 3.29965532e-01 -9.47984099e-01
1.22038469e-01 4.48797122e-02 -4.92586434e-01 -2.52287507e-01
-9.46077526e-01 -4.37241167e-01 -7.71816790e-01 4.26989328e-03
8.91479194e-01 -4.15098339e-01 -6.27076983e-01 2.17955306e-01
-1.72080874e+00 -2.78035313e-01 2.26501748e-01 5.49728990e-01
-7.50907719e-01 4.01305854e-01 -9.94254410e-01 -6.22214198e-01
-5.61865211e-01 -1.13588810e+00 8.98134410e-01 2.27290362e-01
-5.31304836e-01 -6.90131366e-01 3.32198769e-01 4.65043217e-01
3.24515790e-01 -9.24068838e-02 7.80574739e-01 -1.09360290e+00
-2.04190180e-01 4.56731096e-02 -2.92258292e-01 -9.95205417e-02
-7.07116723e-02 -2.33616680e-01 -7.58785248e-01 -9.28092226e-02
1.76476672e-01 -6.38144076e-01 1.09926200e+00 8.20152685e-02
9.00204957e-01 -5.33217549e-01 -3.74383301e-01 3.37651283e-01
9.64282811e-01 1.07975923e-01 4.42698866e-01 2.76250780e-01
5.95480442e-01 1.09968448e+00 5.58034837e-01 6.80606365e-02
6.39962077e-01 8.40019047e-01 2.05598637e-01 3.59596536e-02
8.80913362e-02 -2.11493298e-01 1.19566098e-01 1.04626834e+00
7.19332229e-03 -5.29944539e-01 -6.00902617e-01 6.47110462e-01
-2.29509854e+00 -1.03842258e+00 -4.14621890e-01 1.95620728e+00
1.07593977e+00 -2.23296490e-02 1.31754339e-01 -4.73151773e-01
9.19345379e-01 1.42736375e-01 -2.42686078e-01 -1.13920784e+00
-1.40467137e-01 3.92897904e-01 8.37461203e-02 6.75932527e-01
-9.69565034e-01 1.14931655e+00 5.98612309e+00 3.86048734e-01
-9.47750211e-01 2.10007757e-01 4.89055634e-01 -5.22525236e-02
-4.44997400e-01 2.02558950e-01 -8.00842881e-01 3.76015514e-01
8.81843686e-01 -4.66156751e-02 3.26188087e-01 3.51460427e-01
-1.27126381e-01 1.43517181e-01 -1.26787055e+00 3.61743629e-01
4.03534502e-01 -1.11687446e+00 -9.77524966e-02 -3.46005350e-01
6.06539369e-01 -9.88521427e-02 -2.71941483e-01 5.95608413e-01
3.88565958e-01 -8.16767335e-01 6.40307724e-01 5.18712044e-01
6.24336362e-01 -6.36596203e-01 8.25363636e-01 3.25958580e-01
-9.42913055e-01 -1.20646276e-01 -2.26058051e-01 -2.02481121e-01
3.95433336e-01 -1.03433788e-01 -5.90222001e-01 7.26404190e-01
3.22215259e-01 3.22130203e-01 -1.84724018e-01 6.07612193e-01
-6.57287419e-01 6.83831990e-01 4.38812338e-02 -1.47900403e-01
4.46164995e-01 -2.46474773e-01 5.33171177e-01 1.49730361e+00
-1.56696346e-02 4.77906406e-01 2.10804299e-01 7.52026737e-01
-2.66022056e-01 4.22439039e-01 -3.10452789e-01 2.15312049e-01
5.93498647e-01 1.30110765e+00 -3.09167087e-01 -4.93177235e-01
-6.26826108e-01 1.13667130e+00 9.86955464e-01 3.53080034e-01
-8.68241489e-01 -3.78800243e-01 4.97470945e-01 -2.40744978e-01
4.41314042e-01 1.81887299e-01 -2.05483004e-01 -1.21721053e+00
4.62699309e-02 -9.97507095e-01 4.27798092e-01 -4.91061389e-01
-1.36173904e+00 1.07399428e+00 -2.16181636e-01 -1.00579524e+00
-5.79850018e-01 -2.34622985e-01 -7.66587555e-01 1.25893760e+00
-1.50052893e+00 -1.39404559e+00 1.37354225e-01 5.15008867e-01
1.00292420e+00 1.96162730e-01 1.02959025e+00 9.24422517e-02
-6.30723357e-01 8.60541701e-01 -2.81244546e-01 3.53975147e-01
1.10115290e+00 -1.05041850e+00 6.52206957e-01 6.93183422e-01
1.39042780e-01 9.37428772e-01 4.35156345e-01 -4.16239887e-01
-1.08258247e+00 -6.56312764e-01 1.56158781e+00 -4.74041134e-01
6.27920806e-01 -2.21032053e-01 -1.08515775e+00 8.93822312e-01
1.00318265e+00 -4.19315159e-01 8.15643013e-01 3.82693082e-01
-4.12278354e-01 2.92231530e-01 -7.78496265e-01 5.75494409e-01
1.03536296e+00 -6.74999237e-01 -1.24848104e+00 2.86616683e-01
9.16527152e-01 -6.29264712e-01 -5.73555648e-01 4.37631309e-01
5.60773075e-01 -9.48448181e-01 6.96507454e-01 -1.19332814e+00
7.32543886e-01 2.16666698e-01 -1.04908898e-01 -1.41267371e+00
-7.10416436e-01 -8.38299096e-01 -4.21792455e-02 1.34092581e+00
5.65235376e-01 -5.25595188e-01 1.24294594e-01 5.56901157e-01
-4.65285599e-01 -9.32936490e-01 -7.69120216e-01 -2.01494738e-01
4.15249527e-01 1.64422449e-02 6.90705419e-01 1.02045417e+00
5.59286475e-01 1.17583966e+00 -4.42044079e-01 -4.24797863e-01
6.28607050e-02 7.18202889e-01 6.26267791e-01 -1.04728258e+00
-6.07668161e-01 -6.59360230e-01 4.00967985e-01 -1.47578859e+00
6.33997619e-01 -8.47786129e-01 1.66720361e-01 -1.51283562e+00
4.07848835e-01 -1.60346046e-01 -3.45699251e-01 2.99849033e-01
-6.91586792e-01 8.42074528e-02 3.90293509e-01 3.31114590e-01
-6.13877773e-01 4.35735136e-01 1.20228517e+00 -9.98996496e-02
-1.26566455e-01 1.60089005e-02 -1.03275812e+00 4.70978439e-01
9.08198953e-01 -4.76045370e-01 -1.15682676e-01 -8.84952068e-01
4.49232385e-02 6.45526111e-01 -7.16258734e-02 -2.86204278e-01
2.87811100e-01 -3.08313131e-01 1.98082663e-02 -4.16952521e-01
1.24425031e-01 -2.72661537e-01 -4.40714926e-01 8.80049467e-02
-1.09269691e+00 2.06137866e-01 1.09478282e-02 5.88283658e-01
-2.33304530e-01 -3.64649922e-01 5.80135286e-01 -3.62097234e-01
-4.11768615e-01 1.14723565e-02 -3.68918508e-01 3.30533892e-01
5.75879574e-01 9.65068117e-02 -3.42139155e-01 -3.58443409e-01
-6.37734056e-01 3.96318346e-01 1.29667476e-01 6.25472605e-01
4.66625512e-01 -1.27276778e+00 -1.01408958e+00 2.48779394e-02
2.18248993e-01 -2.81094134e-01 2.15187669e-01 1.07085919e+00
9.12537202e-02 7.26245761e-01 -1.85329646e-01 -2.06858948e-01
-1.16781199e+00 5.56462586e-01 1.20407380e-01 -5.10580301e-01
-3.48460048e-01 1.09421444e+00 3.24912310e-01 -6.35294378e-01
-1.64536938e-01 -1.86847299e-01 -5.54896533e-01 2.01286644e-01
5.09320915e-01 -5.61910346e-02 3.14586237e-02 -8.68855000e-01
-4.03090060e-01 4.22478288e-01 -4.72231299e-01 -2.23857924e-01
1.06929696e+00 -3.78258944e-01 -2.16062501e-01 3.93907696e-01
1.08389127e+00 1.32868096e-01 -1.01004112e+00 -5.34036815e-01
1.06015779e-01 -2.96079814e-01 -3.99033219e-01 -8.88490081e-01
-7.65547872e-01 9.13896859e-01 -3.14242423e-01 2.52421498e-01
8.91200840e-01 7.39892945e-02 1.12846172e+00 3.28556150e-01
2.17712969e-01 -1.18678153e+00 -3.37589495e-02 8.75101089e-01
1.13812339e+00 -1.28922582e+00 -3.73569638e-01 -1.80557281e-01
-9.88858521e-01 1.16561270e+00 9.70413804e-01 -2.78649181e-01
3.40404399e-02 -4.11933176e-02 1.52653009e-01 8.14801604e-02
-1.24949360e+00 -3.57776582e-01 3.65977913e-01 3.62870961e-01
1.07619452e+00 -2.06894052e-04 -1.11990809e+00 9.60954487e-01
6.78993240e-02 -2.57222235e-01 5.29539227e-01 6.99144483e-01
-3.08591545e-01 -1.40379143e+00 -1.21640943e-01 2.14088112e-01
-5.78770041e-01 -3.75084609e-01 -1.06356895e+00 5.36755919e-01
-3.26161414e-01 1.24406934e+00 -1.23031273e-01 -4.67163175e-01
5.71220875e-01 2.65436739e-01 5.78678429e-01 -7.75351644e-01
-1.12668514e+00 1.01491764e-01 5.24910748e-01 -5.45678318e-01
-3.86032820e-01 -6.02321625e-01 -1.16380000e+00 -6.32977933e-02
-5.25385499e-01 2.72829682e-01 2.84853995e-01 1.07823133e+00
4.67322409e-01 5.14836609e-01 8.73260856e-01 -7.21483827e-01
-9.94827092e-01 -1.45904040e+00 5.24761193e-02 5.64887345e-01
2.35599726e-01 -5.05819023e-01 -1.30939186e-01 -2.72731602e-01] | [12.406606674194336, 7.940566062927246] |
78428275-bfe0-465c-ad90-20f1f0acdd47 | fednilm-applying-federated-learning-to-nilm | 2106.07751 | null | https://arxiv.org/abs/2106.07751v1 | https://arxiv.org/pdf/2106.07751v1.pdf | FedNILM: Applying Federated Learning to NILM Applications at the Edge | Non-intrusive load monitoring (NILM) helps disaggregate the household's main electricity consumption to energy usages of individual appliances, thus greatly cutting down the cost in fine-grained household load monitoring. To address the arisen privacy concern in NILM applications, federated learning (FL) could be leveraged for NILM model training and sharing. When applying the FL paradigm in real-world NILM applications, however, we are faced with the challenges of edge resource restriction, edge model personalization and edge training data scarcity. In this paper we present FedNILM, a practical FL paradigm for NILM applications at the edge client. Specifically, FedNILM is designed to deliver privacy-preserving and personalized NILM services to large-scale edge clients, by leveraging i) secure data aggregation through federated learning, ii) efficient cloud model compression via filter pruning and multi-task learning, and iii) personalized edge model building with unsupervised transfer learning. Our experiments on real-world energy data show that, FedNILM is able to achieve personalized energy disaggregation with the state-of-the-art accuracy, while ensuring privacy preserving at the edge client. | ['Jiadong Lou', 'Xudong Wang', 'Yi Wang', 'Qianyi Huang', 'Guoming Tang', 'Yu Zhang'] | 2021-06-07 | null | null | null | null | ['non-intrusive-load-monitoring', 'non-intrusive-load-monitoring', 'non-intrusive-load-monitoring'] | ['knowledge-base', 'miscellaneous', 'time-series'] | [ 6.62787259e-02 -1.34491697e-02 -4.34079647e-01 -3.50984126e-01
-9.00656521e-01 -4.18448657e-01 2.50991076e-01 1.90748610e-02
1.63805112e-01 5.88497698e-01 2.05178931e-01 -4.23707485e-01
-1.38135195e-01 -1.06344950e+00 -6.18017614e-01 -6.85856760e-01
-1.14683464e-01 4.59236085e-01 -5.67456603e-01 4.06735867e-01
-2.61663854e-01 2.63673753e-01 -1.38066578e+00 5.42345107e-01
9.90427554e-01 1.75150979e+00 -2.43255332e-01 3.45591843e-01
7.47770816e-02 8.07091236e-01 -9.25710425e-02 -8.05535495e-01
6.52112663e-01 1.86091408e-01 -7.52629697e-01 -1.69196025e-01
7.73593492e-05 -6.54612839e-01 -2.31459990e-01 1.07534921e+00
6.24874949e-01 -1.05972022e-01 4.46003497e-01 -2.13427615e+00
-5.88042855e-01 8.85692179e-01 -4.59174037e-01 -3.91966611e-01
7.85953104e-02 3.64459194e-02 8.50056648e-01 -2.47765169e-01
1.06300622e-01 6.20595634e-01 9.65305030e-01 3.34061742e-01
-1.30379558e+00 -1.00074756e+00 -1.25326766e-02 6.39771163e-01
-1.39513290e+00 -6.27065957e-01 7.84461379e-01 1.96311343e-02
1.38783610e+00 8.61209989e-01 4.91897106e-01 7.61878490e-01
2.68817246e-01 9.51193810e-01 1.16416645e+00 -1.95685387e-01
5.89075983e-01 3.37345392e-01 -2.09220037e-01 3.66818875e-01
2.81541646e-01 1.96154892e-01 -7.45852530e-01 -6.95677698e-01
-2.57977724e-01 3.56808513e-01 -1.91639319e-01 -3.80057096e-01
-4.06598300e-01 6.09936655e-01 2.39354149e-01 -1.32525906e-01
-6.03659451e-01 5.54567315e-02 9.99632180e-01 3.25526834e-01
6.95780158e-01 -3.18437308e-01 -1.13860190e+00 -4.18958440e-02
-1.06084967e+00 -2.07427725e-01 1.13781667e+00 1.37044156e+00
9.79608178e-01 -1.41838685e-01 -1.21360786e-01 1.87959552e-01
5.88735789e-02 5.27443945e-01 4.94517118e-01 -9.30886269e-01
4.98166233e-01 6.33370697e-01 -1.53209791e-01 -5.25361955e-01
-3.56447726e-01 9.71961170e-02 -1.25937569e+00 5.61156459e-02
-2.41644278e-01 -4.34286177e-01 -1.46187872e-01 1.65994525e+00
5.33673465e-01 3.52791756e-01 6.17057569e-02 2.46333122e-01
2.55789310e-01 5.28223634e-01 2.62793899e-01 -3.71310681e-01
1.57571375e+00 -7.79191971e-01 -6.99735343e-01 1.32336050e-01
8.28700125e-01 -3.61236453e-01 6.63690507e-01 2.57776082e-01
-8.96599889e-01 1.04993701e-01 -6.54722333e-01 2.67376434e-02
-8.52645576e-01 -9.71038565e-02 8.52667093e-01 9.78763223e-01
-7.92998672e-01 5.11483908e-01 -7.58193254e-01 -3.49554121e-01
1.08538043e+00 7.14170039e-01 -4.43783134e-01 -6.75242692e-02
-9.24990356e-01 6.53471053e-01 4.91635501e-01 -2.48369157e-01
-4.18274045e-01 -1.11616588e+00 -7.48771071e-01 4.43084776e-01
2.78921872e-01 -1.03748858e+00 1.12281513e+00 -2.60227531e-01
-1.33514869e+00 6.41136110e-01 1.07199475e-01 -9.15164113e-01
6.72902107e-01 -6.71392977e-02 -7.01646388e-01 -1.81193024e-01
-1.59279734e-01 5.43206781e-02 9.50913370e-01 -9.73127425e-01
-1.05793464e+00 -7.46615171e-01 -3.61458123e-01 -1.17694952e-01
-8.58075917e-01 -3.88160914e-01 1.89886197e-01 -3.76496404e-01
-5.65679431e-01 -6.67560577e-01 -5.19539136e-03 -1.59646899e-01
-3.98769081e-01 -3.22908372e-01 1.66900027e+00 -1.03630626e+00
1.24362767e+00 -2.09532642e+00 -5.31250238e-01 4.20551926e-01
2.49690339e-01 3.48483801e-01 2.13379934e-01 3.66729170e-01
1.01230033e-01 -5.21686152e-02 -5.98871075e-02 -8.88408959e-01
6.34436548e-01 2.51573831e-01 -1.07046306e-01 4.83186096e-01
-5.33797920e-01 1.45526874e+00 -7.21983314e-01 -4.31134462e-01
6.44922256e-01 4.15795237e-01 -4.18691576e-01 3.58192742e-01
-2.57749617e-01 1.31603137e-01 -1.48032501e-01 9.38736379e-01
1.01965511e+00 -3.21986824e-01 3.77606690e-01 -5.60321271e-01
2.66080558e-01 3.14709209e-02 -1.16130269e+00 1.48041511e+00
-1.07645357e+00 1.81997195e-01 4.97380912e-01 -8.99642110e-01
6.40559435e-01 5.45072436e-01 1.06805146e+00 -6.69527352e-01
3.17976385e-01 7.46457055e-02 -9.39917922e-01 -3.11673760e-01
2.70134121e-01 3.58933747e-01 -2.20715731e-01 8.38913202e-01
-8.84863883e-02 3.46938521e-01 -4.52176392e-01 -1.25366643e-01
9.83854949e-01 -2.78682411e-01 4.28499728e-01 -3.63302529e-01
4.32439476e-01 -6.25616848e-01 6.37388408e-01 3.55384231e-01
-2.88045377e-01 -2.34237656e-01 -1.67253062e-01 -5.72077692e-01
-6.40867949e-01 -8.84781063e-01 -1.50629925e-02 1.21060371e+00
-3.79870296e-01 -6.18893802e-01 -8.74234259e-01 -9.41114187e-01
5.54383218e-01 1.19469309e+00 -3.15454662e-01 -4.08202410e-01
-2.25006267e-01 -9.25558329e-01 2.77788252e-01 4.17674989e-01
8.41704965e-01 -8.47159266e-01 -7.97053456e-01 1.92001700e-01
-2.66599655e-01 -1.09777606e+00 -9.56825733e-01 5.08225143e-01
-4.83913839e-01 -1.12551117e+00 4.79358658e-02 -5.65462828e-01
5.36981821e-01 2.13812381e-01 1.22975659e+00 -4.86700475e-01
-3.72656316e-01 6.20586276e-01 -3.48449610e-02 -5.54858863e-01
-2.50339895e-01 4.72014517e-01 1.14929453e-01 2.77027965e-01
1.00566125e+00 -1.22942936e+00 -7.10461736e-01 1.17534004e-01
-7.29154348e-01 -2.67716572e-02 3.07511151e-01 6.22687876e-01
6.22802496e-01 7.58728087e-01 7.66705871e-01 -1.27831125e+00
3.83153915e-01 -6.21294856e-01 -9.05868709e-01 5.42599082e-01
-1.70767498e+00 -4.36789274e-01 1.06575358e+00 -3.11867088e-01
-1.01529586e+00 3.31041485e-01 9.88187343e-02 -5.18898249e-01
-1.58355281e-01 -2.71885246e-01 -8.98885787e-01 -3.41848850e-01
7.12672472e-02 4.32219684e-01 -2.26177007e-01 -5.08121908e-01
7.53811479e-01 1.17675459e+00 7.36249328e-01 -4.81677622e-01
6.11913621e-01 4.97762442e-01 1.27477109e-01 -2.82781959e-01
-2.33894885e-01 -4.11336511e-01 -2.51186103e-01 2.68396795e-01
6.34503484e-01 -1.14482820e+00 -1.51332533e+00 6.89710915e-01
-6.88936353e-01 -3.24985713e-01 -8.01749885e-01 4.59716702e-03
-7.79582620e-01 1.76803783e-01 -4.81454730e-01 -9.06012118e-01
-1.39522231e+00 -8.04030061e-01 9.40932393e-01 7.57509023e-02
-1.61257252e-01 -9.61995840e-01 -4.79555398e-01 5.13175845e-01
7.36609936e-01 1.75920874e-01 1.10842347e+00 -8.00288856e-01
-6.66620970e-01 -5.94676197e-01 -1.84128404e-01 3.88686299e-01
5.34000754e-01 -9.15226400e-01 -1.33075058e+00 -7.65301049e-01
3.07989210e-01 -1.65823191e-01 2.71787256e-01 3.07902902e-01
1.75934601e+00 -1.08086360e+00 -3.31830591e-01 1.26549053e+00
1.54026675e+00 -1.05306774e-01 4.64592010e-01 3.71237565e-03
7.55965233e-01 3.30020159e-01 -6.96656294e-03 1.00653183e+00
9.16156948e-01 1.83270246e-01 5.72332561e-01 7.51124397e-02
2.89946258e-01 -5.56900203e-01 1.73991024e-01 6.15306497e-01
5.80220997e-01 -2.03038767e-01 -1.68756142e-01 4.50481325e-01
-2.16188669e+00 -7.71005154e-01 4.25438523e-01 2.16758776e+00
4.88260001e-01 -2.25042373e-01 2.99872816e-01 1.85239241e-01
5.19772410e-01 9.99776348e-02 -1.15452003e+00 -5.53122282e-01
-9.16478038e-02 1.07431747e-02 1.20935547e+00 8.16889331e-02
-1.13863778e+00 3.53425592e-01 5.27491379e+00 9.78438735e-01
-8.88517737e-01 6.01586819e-01 8.29807103e-01 -2.12437555e-01
-2.02131793e-01 -1.98470458e-01 -6.27786636e-01 8.95915508e-01
1.07369411e+00 -7.34107494e-01 1.18281710e+00 1.29561496e+00
-2.91500743e-02 3.07107180e-01 -1.29407835e+00 1.49319208e+00
-2.61170715e-01 -1.41692758e+00 -4.08614665e-01 3.97263587e-01
9.83377934e-01 2.26713419e-01 -1.13054588e-02 2.14172274e-01
5.44874012e-01 -8.10120702e-01 2.32887372e-01 4.06345457e-01
9.03743207e-01 -1.10412753e+00 5.97055256e-01 4.92801696e-01
-1.63596833e+00 -6.14518702e-01 -1.01746231e-01 1.81971848e-01
1.30200788e-01 8.89271736e-01 -7.04940975e-01 8.66520107e-01
1.22257686e+00 3.28166485e-01 -1.98362216e-01 4.94737953e-01
3.75143856e-01 3.80848259e-01 -7.80004263e-01 3.27261776e-01
-4.64254797e-01 -3.06196600e-01 -6.93027629e-03 1.14182019e+00
4.08033937e-01 -6.92249760e-02 2.80852467e-01 5.90333164e-01
-6.86346591e-01 1.73610687e-01 -7.98122406e-01 2.15412423e-01
8.22173834e-01 1.66777539e+00 -5.42528462e-03 -1.71801031e-01
-6.48031592e-01 1.33522761e+00 9.84497964e-02 3.33560519e-02
-5.34555852e-01 -1.19698547e-01 1.08600593e+00 5.58021106e-02
4.41896796e-01 4.73674744e-01 -5.48533142e-01 -1.06890202e+00
5.95062673e-02 -6.79024041e-01 9.44989741e-01 -1.78330407e-01
-2.02206540e+00 1.71696737e-01 -2.94770271e-01 -7.68730760e-01
-4.58060563e-01 -1.19035996e-01 -8.67007434e-01 7.71415293e-01
-1.65803921e+00 -1.97891712e+00 -8.93695652e-02 9.96645451e-01
3.68399289e-03 -2.10942522e-01 1.30901730e+00 7.18104303e-01
-4.44860369e-01 1.16791153e+00 5.80793798e-01 -1.27960473e-01
2.89574325e-01 -1.25695431e+00 5.51637530e-01 4.97920722e-01
-5.39467819e-02 -1.64489880e-01 1.11959510e-01 -4.39619511e-01
-1.79001510e+00 -1.83537269e+00 8.43255997e-01 -2.26513207e-01
3.12853247e-01 -7.82071531e-01 -5.61218917e-01 9.34210420e-01
2.22180963e-01 2.92693943e-01 1.14746320e+00 -1.54822648e-01
-2.43519232e-01 -7.71474004e-01 -2.13272929e+00 4.88891825e-02
8.95921290e-01 -8.93582284e-01 1.50250137e-01 6.77249789e-01
7.17464328e-01 -6.68120235e-02 -1.36893809e+00 3.06748539e-01
3.96730483e-01 -8.76285255e-01 8.23549569e-01 -5.17347217e-01
-5.73514342e-01 -1.11146353e-01 -6.37293696e-01 -1.13998103e+00
-4.50047642e-01 -1.33002985e+00 -1.27579558e+00 1.98398733e+00
-1.46939933e-01 -1.04926169e+00 1.09599113e+00 1.26066315e+00
3.35924357e-01 -8.15520525e-01 -1.24708772e+00 -7.40997553e-01
-8.27905983e-02 -6.35469913e-01 1.84675252e+00 8.31638336e-01
-7.63535500e-02 -1.46096870e-01 -6.28786981e-01 5.43190181e-01
1.10227275e+00 4.21936989e-01 6.40730143e-01 -9.90286589e-01
-1.78128004e-01 -1.54887378e-01 -1.60602540e-01 -4.60436165e-01
5.03346086e-01 -1.16814089e+00 -4.39002901e-01 -1.19088042e+00
6.11083582e-02 -3.93820345e-01 -5.76373756e-01 7.86752462e-01
4.42643374e-01 1.76564485e-01 3.64573866e-01 -9.40513313e-02
-6.32252038e-01 5.34903705e-01 2.38200232e-01 -3.72881711e-01
-4.45014648e-02 4.18852836e-01 -9.36594605e-01 6.05924726e-01
9.12750661e-01 -3.30387086e-01 -4.45941746e-01 -3.12416911e-01
-2.19590321e-01 -2.06228286e-01 2.41863042e-01 -8.21671247e-01
5.46041906e-01 -1.31314963e-01 4.10993874e-01 -7.48256207e-01
6.70723543e-02 -1.69693589e+00 7.15926290e-01 5.27947605e-01
3.12386096e-01 -5.04802577e-02 4.24893834e-02 3.14476460e-01
4.96927023e-01 2.51471072e-01 5.79750359e-01 -5.98259456e-02
-6.08202994e-01 9.24095869e-01 1.71585813e-01 -2.17096910e-01
1.34867799e+00 2.41069615e-01 -2.69368708e-01 -2.78424352e-01
-4.71261114e-01 8.42179954e-01 6.63257957e-01 1.61329895e-01
5.56785390e-02 -1.48856318e+00 -4.81582165e-01 6.36772335e-01
9.23326537e-02 5.13884984e-02 3.42162460e-01 3.76459837e-01
3.29713911e-01 3.39272290e-01 2.71000594e-01 -3.23288143e-01
-1.10279560e+00 1.09673953e+00 2.10928440e-01 -4.81879741e-01
-7.25912571e-01 3.63201559e-01 -1.53434485e-01 -5.43741703e-01
5.72798848e-01 -1.30719081e-01 6.75158441e-01 8.70763436e-02
4.87694621e-01 1.06709969e+00 5.62629163e-01 -3.10416520e-01
-3.39883447e-01 1.79586932e-01 2.07016230e-01 7.75868356e-01
1.41504264e+00 -6.67884588e-01 -3.25642794e-01 -7.54239261e-02
1.70435870e+00 -2.50691950e-01 -1.20340228e+00 -2.75628895e-01
2.75556110e-02 -4.71130908e-01 2.56284744e-01 -9.58873034e-01
-1.48945916e+00 3.96369815e-01 1.06361628e+00 2.36732006e-01
1.96901762e+00 -1.88601822e-01 1.83818603e+00 1.70718074e-01
8.33754897e-01 -1.22956204e+00 -7.98887908e-01 -2.29729116e-01
3.28653492e-02 -1.28873813e+00 -1.17546350e-01 -7.33608603e-02
-3.05685252e-01 5.68194151e-01 2.91797608e-01 3.78385097e-01
1.24211597e+00 6.79787755e-01 -1.98839173e-01 1.90481439e-01
-7.53900051e-01 3.17931950e-01 -1.64834023e-01 9.52289522e-01
-5.97526133e-01 6.19508862e-01 3.01762849e-01 1.09720099e+00
-2.54047185e-01 4.03414071e-01 -2.49419361e-01 8.79087746e-01
3.50781888e-01 -1.20065475e+00 -2.69678861e-01 1.09948480e+00
-4.87279028e-01 2.28057727e-02 1.92005455e-01 -1.04753319e-02
5.79239845e-01 9.38009322e-01 5.64183369e-02 -4.89628553e-01
2.82746643e-01 4.07708406e-01 2.19414413e-01 2.03547254e-01
-5.77661991e-01 -4.46294814e-01 -3.36081266e-01 -6.98614419e-01
2.10602537e-01 -7.52284586e-01 -9.63845313e-01 -9.71151292e-01
-2.57672489e-01 1.74458902e-02 6.67429388e-01 6.77693486e-01
9.38888967e-01 1.83712691e-01 1.34795380e+00 -6.55560732e-01
-9.22360897e-01 -5.32454729e-01 -9.60936725e-01 5.78263402e-01
2.46019214e-01 4.25981246e-02 -2.95537978e-01 -6.51441440e-02] | [5.876384735107422, 2.81178879737854] |
88ca41a2-661a-427c-9ba0-0e4f52d76985 | reconstruction-of-fragmented-trajectories-of | 2110.10428 | null | https://arxiv.org/abs/2110.10428v1 | https://arxiv.org/pdf/2110.10428v1.pdf | Reconstruction of Fragmented Trajectories of Collective Motion using Hadamard Deep Autoencoders | Learning dynamics of collectively moving agents such as fish or humans is an active field in research. Due to natural phenomena such as occlusion and change of illumination, the multi-object methods tracking such dynamics might lose track of the agents where that might result fragmentation in the constructed trajectories. Here, we present an extended deep autoencoder (DA) that we train only on fully observed segments of the trajectories by defining its loss function as the Hadamard product of a binary indicator matrix with the absolute difference between the outputs and the labels. The trajectories of the agents practicing collective motion is low-rank due to mutual interactions and dependencies between the agents that we utilize as the underlying pattern that our Hadamard deep autoencoder (HDA) codes during its training. The performance of our HDA is compared with that of a low-rank matrix completion scheme in the context of fragmented trajectory reconstruction. | ['Anura P. Jayasumana', 'Randy Paffenroth', 'Yonggi Park', 'Kelum Gajamannage'] | 2021-10-20 | null | null | null | null | ['low-rank-matrix-completion'] | ['methodology'] | [-3.87061507e-01 -1.55533418e-01 5.11073291e-01 -1.01016350e-02
-4.11723256e-02 -5.47036469e-01 8.30222785e-01 -6.03229143e-02
-6.33408546e-01 4.94242668e-01 5.49858034e-01 1.43718734e-01
-4.25897509e-01 -7.47109354e-01 -1.10245061e+00 -1.11419237e+00
-6.81067348e-01 8.39183092e-01 3.17690223e-02 1.88361201e-02
-3.62776339e-01 5.51344931e-01 -1.43930399e+00 8.24002326e-02
3.01247537e-01 3.65934908e-01 2.27090418e-02 9.85434413e-01
3.26581955e-01 1.12397754e+00 -5.05394876e-01 -3.62114340e-01
5.23174286e-01 -2.02277064e-01 -4.28499311e-01 9.97597277e-02
3.19711804e-01 -7.17818677e-01 -1.36455071e+00 1.04344428e+00
1.18793352e-02 5.15955985e-01 9.01020646e-01 -1.43143308e+00
-7.67076731e-01 6.64979696e-01 -2.58656859e-01 9.62344036e-02
-8.85918662e-02 3.41742933e-01 8.09092581e-01 -6.67899549e-01
7.28036225e-01 1.39180124e+00 9.41544712e-01 5.28948605e-01
-1.08708024e+00 -3.64172816e-01 4.78183292e-03 2.81504929e-01
-1.33699167e+00 -1.77260324e-01 3.86973679e-01 -1.02890837e+00
6.68014944e-01 -1.40666589e-01 8.01116288e-01 1.14331138e+00
2.49235570e-01 6.99343681e-01 4.70971137e-01 9.78135765e-02
2.18757734e-01 -3.60158652e-01 1.10051572e-01 1.02605033e+00
2.78041422e-01 5.55947304e-01 -1.91793084e-01 -4.08477455e-01
9.09959316e-01 5.30100882e-01 -2.63137341e-01 -1.99621469e-01
-1.42759001e+00 8.54310453e-01 5.46125472e-01 1.47667632e-01
-6.41426742e-01 4.70966607e-01 3.63605320e-02 3.59580278e-01
-1.99686196e-02 1.69426098e-01 -1.88133404e-01 2.21764715e-03
-6.38558686e-01 5.10132611e-01 9.67118919e-01 6.17602468e-01
1.02877283e+00 1.84552327e-01 1.03233859e-01 1.90700427e-01
3.23362380e-01 7.32632458e-01 3.20939988e-01 -1.20119298e+00
-6.03525490e-02 7.07167804e-01 3.21293116e-01 -1.16093671e+00
-4.81626928e-01 -3.12864512e-01 -1.06078506e+00 4.52586174e-01
6.91714287e-01 -5.89094043e-01 -6.03516161e-01 2.00454688e+00
4.30696756e-01 6.02015257e-01 3.44099671e-01 1.09256876e+00
5.95595181e-01 9.01699483e-01 -1.33416623e-01 8.53265226e-02
8.52486551e-01 -9.47390914e-01 -6.78900659e-01 2.62010485e-01
3.52725804e-01 -1.17135391e-01 4.20719773e-01 -3.78438495e-02
-6.86698198e-01 -4.05101031e-01 -8.40442359e-01 5.50503023e-02
-2.79788047e-01 1.29498571e-01 7.25508094e-01 -5.09123877e-02
-1.04033148e+00 8.65003884e-01 -1.45456922e+00 -3.08872104e-01
1.35906667e-01 4.08206284e-01 -6.22814000e-01 2.27855787e-01
-8.65396082e-01 7.70069242e-01 5.69962114e-02 2.44908318e-01
-1.69405770e+00 -5.70746124e-01 -7.13609338e-01 -2.12783329e-02
-8.13352466e-02 -7.49208331e-01 8.97876740e-01 -8.06537151e-01
-1.34212232e+00 4.26829219e-01 3.41969788e-01 -5.49661040e-01
6.07003570e-01 -4.00130212e-01 -2.65646130e-01 1.11748613e-01
5.98225631e-02 7.56845653e-01 8.04226398e-01 -1.09472537e+00
-7.13495493e-01 -3.95011038e-01 1.32121712e-01 2.39384651e-01
-2.59446055e-01 -5.57990372e-01 -8.86810049e-02 -5.00262737e-01
-2.87439197e-01 -1.38164139e+00 -3.35492939e-01 2.28928506e-01
-1.64433405e-01 -1.76261276e-01 9.76042867e-01 -7.32426286e-01
8.86378467e-01 -2.27782512e+00 6.91943586e-01 -1.77096665e-01
3.39163959e-01 -3.96881513e-02 -4.95488077e-01 8.17366600e-01
4.47211057e-01 -4.07423854e-01 -3.54137301e-01 -4.25124854e-01
5.75793721e-02 5.57118893e-01 -3.37334514e-01 1.05409908e+00
-2.92936135e-02 4.96261507e-01 -1.08267403e+00 4.54761572e-02
-4.42358106e-02 7.00624108e-01 -4.17955458e-01 3.93935680e-01
-2.27991968e-01 6.68750763e-01 -4.23772544e-01 2.02049643e-01
4.57226127e-01 -3.19039524e-01 -6.97707906e-02 -1.85634270e-02
-3.70955139e-01 -2.44335681e-01 -1.04752815e+00 1.60601914e+00
1.83710515e-01 9.36769128e-01 2.25457132e-01 -5.73951662e-01
4.08901393e-01 3.97630662e-01 8.90610695e-01 2.32412275e-02
6.64283410e-02 -1.21354342e-01 2.19533548e-01 -6.25123143e-01
3.51883352e-01 9.25934985e-02 1.85806379e-01 5.95080197e-01
7.22971335e-02 5.09394944e-01 1.37297481e-01 3.55628014e-01
1.55017567e+00 1.27866685e-01 -1.76272452e-01 -6.97020888e-02
2.05070212e-01 3.27602744e-01 6.28560305e-01 7.90831745e-01
-2.73761690e-01 3.12380284e-01 2.12225884e-01 -1.02817202e+00
-1.49398804e+00 -9.72906768e-01 1.89776421e-01 9.59712207e-01
1.02712333e-01 3.33887935e-02 -6.81452632e-01 -2.87602514e-01
2.05405354e-01 3.21719646e-01 -1.00692713e+00 -1.00499786e-01
-7.51620650e-01 -7.85225391e-01 9.45011735e-01 2.69411087e-01
3.59508574e-01 -1.14431703e+00 -9.42010820e-01 4.17006075e-01
2.38136053e-02 -1.01144481e+00 -3.60142797e-01 -1.38917536e-01
-6.98451519e-01 -1.30919290e+00 -6.61065459e-01 -7.21935093e-01
6.42095625e-01 9.85896066e-02 7.18927622e-01 -1.95437111e-02
-1.14678793e-01 5.38018882e-01 -2.47721076e-01 -1.24280579e-01
-5.14031291e-01 -2.80034363e-01 4.25567031e-01 3.25774878e-01
1.55570790e-01 -8.43273461e-01 -7.58359790e-01 4.64925766e-02
-9.95904684e-01 -1.73485115e-01 5.91779470e-01 7.25474954e-01
1.51123181e-01 2.29085937e-01 1.39151350e-01 -3.37092340e-01
4.95386213e-01 -6.59264982e-01 -7.68902659e-01 -1.90163462e-03
1.82342991e-01 1.07175060e-01 4.50820029e-01 -8.60290527e-01
-6.65422022e-01 3.70684117e-01 2.21724436e-01 -1.05334175e+00
-2.04584256e-01 1.79843992e-01 4.01619077e-01 4.62936163e-02
5.33870101e-01 2.29898766e-01 2.46049076e-01 -4.05912489e-01
4.70592052e-01 2.35609487e-01 6.63963318e-01 -2.83450752e-01
8.52427900e-01 9.62197006e-01 1.67322010e-01 -9.64941442e-01
-3.30973446e-01 -2.71689683e-01 -5.46989501e-01 -4.81455535e-01
9.98734772e-01 -1.07526422e+00 -1.30546951e+00 7.43446589e-01
-1.38603389e+00 -5.39096236e-01 -2.50478566e-01 1.02433383e+00
-3.53537887e-01 2.03430384e-01 -1.02606452e+00 -7.70089626e-01
-3.68741304e-02 -9.42760170e-01 8.09193373e-01 2.34132260e-01
1.59256935e-01 -1.09295249e+00 7.37450600e-01 -2.72906363e-01
1.24800183e-01 5.59343755e-01 7.87300825e-01 -6.79246724e-01
-8.93346846e-01 -1.51801556e-01 2.41553798e-01 9.15574357e-02
-1.06977731e-01 1.49067611e-01 -6.17487729e-01 -5.65863967e-01
-2.14885529e-02 -1.96720675e-01 9.75871444e-01 5.83441496e-01
2.21442327e-01 -7.77649581e-01 -5.97518265e-01 5.76596081e-01
1.33108997e+00 1.68309957e-01 1.52241752e-01 1.94636941e-01
9.29375827e-01 7.53768265e-01 -1.26350865e-01 7.01387525e-01
5.88776648e-01 1.96444422e-01 6.66136324e-01 2.71860778e-01
1.04345754e-02 -2.35673621e-01 6.83208942e-01 1.07921052e+00
-1.12561516e-01 -3.89189214e-01 -9.28712666e-01 8.09731245e-01
-2.26590014e+00 -1.19096661e+00 -1.56525418e-01 1.80907953e+00
1.91932812e-01 -5.64402044e-01 1.44950330e-01 -3.20511580e-01
7.68253088e-01 1.03751414e-01 -7.78829575e-01 2.21716091e-01
-2.31133789e-01 -5.35665214e-01 5.28586268e-01 6.18197322e-01
-1.27592361e+00 7.18468666e-01 6.28804493e+00 1.37733713e-01
-7.84104645e-01 6.03796765e-02 -1.80608094e-01 1.15290824e-02
2.11449787e-02 -6.14459291e-02 -5.78374445e-01 3.88376147e-01
8.07693422e-01 2.05255583e-01 8.01374614e-01 5.34950972e-01
1.84196115e-01 2.83475429e-01 -1.28669369e+00 6.42518342e-01
-1.16671555e-01 -1.19368684e+00 1.77088842e-01 3.70966107e-01
1.00562370e+00 4.83827591e-01 9.87905115e-02 2.83313334e-01
1.11409187e+00 -8.67402136e-01 7.83715248e-01 9.65049863e-01
1.24037802e-01 -4.73693877e-01 4.84316230e-01 6.25780046e-01
-9.64938700e-01 -4.37960565e-01 -4.81476039e-01 -4.22401547e-01
2.52442062e-01 3.11821014e-01 -7.71554887e-01 1.19339988e-01
5.18319905e-01 8.26469183e-01 -1.92585379e-01 1.13027668e+00
2.73411810e-01 5.10445654e-01 -5.60263276e-01 2.48571858e-02
7.08223641e-01 -7.16809332e-01 1.17391264e+00 9.82542932e-01
3.81134957e-01 3.44469637e-01 3.59933704e-01 8.31733406e-01
-1.11957677e-01 -4.74242866e-01 -8.59639108e-01 -2.92741835e-01
3.01015168e-01 9.59695995e-01 -4.34326172e-01 -3.00860435e-01
-5.14067352e-01 8.21009994e-01 5.93114078e-01 5.07114887e-01
-5.50690055e-01 1.66361153e-01 1.23077083e+00 -2.73287952e-01
4.46252316e-01 -4.34479445e-01 3.87961030e-01 -1.14668906e+00
-4.35177505e-01 -5.84242404e-01 3.78784090e-01 -3.72362971e-01
-1.55806279e+00 6.41374052e-01 -1.93172947e-01 -1.26094377e+00
-1.57462671e-01 -5.44887841e-01 -5.17624438e-01 2.00435534e-01
-9.73855793e-01 -1.15146542e+00 -7.40680397e-02 6.50190353e-01
2.77110338e-01 -5.20911694e-01 7.29203045e-01 5.03805161e-01
-5.49416006e-01 -1.24657005e-02 6.72053397e-01 5.65891862e-01
8.61111954e-02 -1.11674559e+00 1.52099907e-01 7.71515310e-01
3.58878016e-01 5.23591280e-01 9.57359612e-01 -8.31478596e-01
-1.54167342e+00 -1.24759018e+00 3.47760975e-01 -4.23456877e-01
9.84005153e-01 -1.33875972e-02 -8.92291546e-01 1.20718193e+00
2.13185981e-01 4.22658212e-02 5.43644547e-01 -3.26240957e-01
-1.01657748e-01 1.99998870e-01 -7.06940889e-01 4.98560011e-01
9.80889320e-01 -4.91744995e-01 -4.47394252e-01 3.48262846e-01
6.23806596e-01 -2.53558278e-01 -6.83097005e-01 2.86954224e-01
7.73102343e-01 -7.54241586e-01 7.71142542e-01 -1.01707721e+00
3.46995056e-01 -6.74771547e-01 -2.32543543e-01 -1.30124652e+00
-5.96744537e-01 -5.35564423e-01 -2.74909377e-01 6.53776288e-01
1.48553941e-02 -1.58874735e-01 6.67811036e-01 3.19502711e-01
-2.04905227e-01 -2.95389414e-01 -8.45619917e-01 -5.44706643e-01
1.23130329e-01 1.95819557e-01 4.64569956e-01 9.30664241e-01
-4.24695402e-01 1.15706712e-01 -8.04733455e-01 8.36520314e-01
9.47705328e-01 1.21727310e-01 9.21501815e-01 -1.32499278e+00
-2.87449300e-01 -1.87234104e-01 -5.47148645e-01 -1.10637820e+00
3.75077903e-01 -7.40767777e-01 2.41720319e-01 -1.20959818e+00
2.82382518e-01 -2.66465209e-02 -1.10542148e-01 1.77564770e-01
2.72721201e-01 -1.61352396e-01 2.13078842e-01 5.22826552e-01
-6.88678086e-01 8.51332009e-01 1.18507075e+00 -4.04471338e-01
-2.13719442e-01 -9.24568474e-02 9.40511897e-02 9.11708295e-01
1.64612517e-01 -8.77163947e-01 -6.87500387e-02 -9.05610859e-01
1.99514508e-01 3.95290852e-01 5.98393321e-01 -1.21835005e+00
7.62313247e-01 3.35296965e-03 1.75290421e-01 -6.67610466e-01
4.69200015e-01 -1.08656561e+00 8.40172112e-01 6.91293359e-01
-4.17719126e-01 3.58449578e-01 -1.03930958e-01 1.14758098e+00
-1.48217246e-01 -4.30888273e-02 5.04004478e-01 -3.55028450e-01
-5.03094554e-01 5.80781996e-01 -8.29265594e-01 -1.17468916e-01
8.80424976e-01 1.96382284e-01 -2.77884811e-01 -6.06209338e-01
-8.50822747e-01 4.86178815e-01 4.90418166e-01 1.81947395e-01
5.85352778e-01 -1.33872283e+00 -9.16388988e-01 -9.06496346e-02
-3.60304266e-01 -2.31917184e-02 4.27591413e-01 6.54620051e-01
-7.00334370e-01 2.03022987e-01 -5.60437441e-01 -4.20528471e-01
-7.04874754e-01 5.23530006e-01 6.07500911e-01 -1.55291632e-01
-9.88643706e-01 5.82210243e-01 5.26022792e-01 -4.16300595e-01
4.22254562e-01 -1.04639642e-01 -3.68407041e-01 -3.26992534e-02
5.53550661e-01 7.93128908e-01 -4.99137580e-01 -1.02964711e+00
-1.11994699e-01 3.96439880e-01 1.59921959e-01 -3.38822514e-01
1.69462991e+00 -1.26247525e-01 -1.48272783e-01 5.33983171e-01
1.17020655e+00 -3.78243983e-01 -1.99915743e+00 -1.30863428e-01
-2.52996594e-01 -8.50502476e-02 -1.24047607e-01 -2.97371417e-01
-9.60578799e-01 7.36392975e-01 5.55217564e-01 2.31940076e-01
5.07113159e-01 -1.02625459e-01 8.04139733e-01 8.70322645e-01
6.40543103e-02 -6.17560029e-01 1.55459210e-01 7.90785134e-01
8.18565488e-01 -1.02633429e+00 -3.25720012e-01 3.65285516e-01
-5.75731814e-01 1.09054399e+00 3.78398538e-01 -6.02889776e-01
8.21555018e-01 2.65316039e-01 1.49952173e-01 -4.67114657e-01
-9.53741431e-01 -3.63811642e-01 -6.86722100e-02 6.62552476e-01
-3.99883837e-01 1.42671227e-01 1.50771499e-01 2.39413768e-01
-8.95597786e-02 -2.98995852e-01 6.09403670e-01 6.31763935e-01
-3.77153963e-01 -4.42320913e-01 -2.50975162e-01 1.61804110e-01
-2.42960930e-01 3.80742073e-01 -2.12975845e-01 7.72409081e-01
1.97926953e-01 6.93778396e-01 4.91242647e-01 -4.67443556e-01
1.78595588e-01 -1.82939768e-01 1.91462070e-01 -1.87548637e-01
-4.57361013e-01 -1.78889632e-01 -3.28407884e-01 -2.42382586e-01
-7.50387192e-01 -9.23507929e-01 -1.27719402e+00 -4.53580379e-01
-1.98908169e-02 1.55977517e-01 4.21543479e-01 8.11078370e-01
2.64046848e-01 3.13456059e-01 4.15260106e-01 -1.02635384e+00
-5.44527113e-01 -9.82089937e-01 -5.56418419e-01 8.70503843e-01
1.06295729e+00 -7.87743151e-01 -5.61625063e-01 3.24693978e-01] | [5.935495376586914, 0.7603923678398132] |
0831caac-aa84-4bea-8afe-532039912453 | prototypical-model-with-novel-information | 2112.03134 | null | https://arxiv.org/abs/2112.03134v1 | https://arxiv.org/pdf/2112.03134v1.pdf | Prototypical Model with Novel Information-theoretic Loss Function for Generalized Zero Shot Learning | Generalized zero shot learning (GZSL) is still a technical challenge of deep learning as it has to recognize both source and target classes without data from target classes. To preserve the semantic relation between source and target classes when only trained with data from source classes, we address the quantification of the knowledge transfer and semantic relation from an information-theoretic viewpoint. To this end, we follow the prototypical model and format the variables of concern as a probability vector. Leveraging on the proposed probability vector representation, the information measurement such as mutual information and entropy, can be effectively evaluated with simple closed forms. We discuss the choice of common embedding space and distance function when using the prototypical model. Then We propose three information-theoretic loss functions for deterministic GZSL model: a mutual information loss to bridge seen data and target classes; an uncertainty-aware entropy constraint loss to prevent overfitting when using seen data to learn the embedding of target classes; a semantic preserving cross entropy loss to preserve the semantic relation when mapping the semantic representations to the common space. Simulation shows that, as a deterministic model, our proposed method obtains state of the art results on GZSL benchmark datasets. We achieve 21%-64% improvements over the baseline model -- deep calibration network (DCN) and for the first time demonstrate a deterministic model can perform as well as generative ones. Moreover, our proposed model is compatible with generative models. Simulation studies show that by incorporating with f-CLSWGAN, we obtain comparable results compared with advanced generative models. | ['Huiwen Yang', 'Meiying Zhang', 'Feng Chen', 'Zhan Xiong', 'Hanchu Shen', 'Chunlin Ji'] | 2021-12-06 | null | null | null | null | ['generalized-zero-shot-learning', 'generalized-zero-shot-learning'] | ['computer-vision', 'methodology'] | [ 2.28858158e-01 5.22967637e-01 -1.73499525e-01 -3.01158130e-01
-8.10513496e-01 -3.92756999e-01 8.51025581e-01 -1.31591529e-01
-1.73092753e-01 7.37726152e-01 7.17486292e-02 1.38763815e-01
-6.01854503e-01 -1.24125957e+00 -9.74518001e-01 -9.97067750e-01
1.38694137e-01 5.93045354e-01 2.13836916e-02 -4.09590565e-02
-1.93232238e-01 1.94686696e-01 -1.65710878e+00 -1.08530946e-01
8.02524030e-01 1.14170587e+00 5.49272113e-02 2.63930649e-01
-1.83613345e-01 7.01200604e-01 -4.51352805e-01 -6.30812824e-01
4.78582650e-01 -5.18661559e-01 -8.00782979e-01 -2.33584300e-01
1.42032161e-01 -9.05061364e-02 -3.76868129e-01 1.27599740e+00
5.36654353e-01 1.57489285e-01 9.26507950e-01 -1.63806987e+00
-7.90739000e-01 8.63335073e-01 -6.89358860e-02 -2.06658781e-01
-1.14614807e-01 -1.22949220e-01 1.19826543e+00 -6.93088830e-01
6.62186146e-01 1.30184627e+00 5.72450578e-01 6.21638834e-01
-1.26548743e+00 -7.37882137e-01 -5.92247210e-02 3.39954495e-01
-1.35372841e+00 -2.32941419e-01 8.92085314e-01 -4.61741209e-01
6.60992205e-01 6.85528815e-02 4.34665352e-01 1.33796549e+00
1.96138799e-01 5.89381456e-01 8.85503531e-01 -5.31485856e-01
4.51435655e-01 4.80240554e-01 3.45541716e-01 4.54790086e-01
3.64111125e-01 3.17671597e-01 -4.81389284e-01 1.63266480e-01
3.27215374e-01 8.08730274e-02 -4.41990077e-01 -9.51490521e-01
-9.49638724e-01 1.12311041e+00 5.62200367e-01 3.74794543e-01
8.87988880e-03 2.59755850e-01 2.06657231e-01 3.80698949e-01
3.40742439e-01 1.49919614e-01 -2.54600972e-01 2.03002110e-01
-6.73734486e-01 -5.04800081e-02 9.20290232e-01 1.33104289e+00
1.04974389e+00 1.34775519e-01 -3.54591280e-01 6.98085666e-01
5.28469861e-01 7.73509324e-01 4.91177559e-01 -6.50450766e-01
3.46427858e-01 6.11216724e-01 -3.01761806e-01 -9.64654028e-01
2.77916691e-03 -7.36436605e-01 -9.30757701e-01 1.66220531e-01
1.63786799e-01 1.05676346e-01 -8.89691472e-01 2.31362987e+00
1.43448532e-01 4.10316199e-01 6.35833263e-01 4.55201775e-01
8.20470631e-01 6.63544118e-01 -6.27232864e-02 -1.36663616e-01
1.14442015e+00 -6.20442867e-01 -8.23801279e-01 -1.31856188e-01
5.19889116e-01 -4.27758008e-01 9.17387903e-01 1.12959377e-01
-7.27716446e-01 -4.40399855e-01 -1.37968719e+00 -7.77417198e-02
-7.51311183e-01 -7.45091066e-02 3.67560506e-01 9.32688177e-01
-9.43200588e-01 8.17498803e-01 -7.72555470e-01 -3.50311309e-01
4.30238098e-01 3.32302123e-01 -3.29700261e-01 -5.08508123e-02
-1.62476730e+00 9.44907606e-01 7.27791846e-01 -1.80039883e-01
-1.01356852e+00 -9.46071088e-01 -1.00780153e+00 4.44953293e-01
1.66754767e-01 -7.08624542e-01 7.98711538e-01 -7.23068058e-01
-1.44853246e+00 6.90147340e-01 2.98184663e-01 -7.27682590e-01
4.88589168e-01 -1.86725169e-01 -2.79809296e-01 -1.02205835e-01
3.88471670e-02 5.18178463e-01 6.35510027e-01 -1.40956318e+00
-3.66867930e-01 -2.72390515e-01 -1.64457243e-02 2.18548387e-01
-4.89491314e-01 -6.01649106e-01 -2.93919027e-01 -5.18514216e-01
3.17227356e-02 -8.68146837e-01 2.99268872e-01 -8.97266790e-02
-4.31679130e-01 -8.23015496e-02 9.32111859e-01 -3.55404019e-01
9.68586385e-01 -2.15250182e+00 3.11867386e-01 2.03220904e-01
1.04493089e-01 2.19698548e-01 1.18180048e-02 2.81691134e-01
-1.33334473e-01 -6.74538985e-02 -5.07386744e-01 -4.37124342e-01
4.06503111e-01 3.50291401e-01 -5.71206748e-01 4.75579500e-01
3.85564677e-02 8.79808187e-01 -9.19860423e-01 -2.12167904e-01
5.61386228e-01 9.80720282e-01 -4.56172973e-01 2.62698412e-01
2.96216235e-02 -1.74009711e-01 -1.51102096e-01 1.69485018e-01
7.85591662e-01 -1.63301036e-01 1.22745477e-01 -4.97730076e-01
4.44290251e-01 -8.90730973e-03 -1.19861746e+00 1.82680213e+00
-6.68895721e-01 5.46252072e-01 -3.74822915e-01 -1.12487578e+00
1.23612595e+00 1.69069439e-01 3.34624618e-01 -4.16528165e-01
4.12144214e-01 8.33353996e-02 -1.57212049e-01 4.36953716e-02
-2.62888595e-02 -4.04426277e-01 -2.08217502e-01 1.80472091e-01
9.33678508e-01 2.70868726e-02 -2.36005530e-01 1.76696464e-01
7.79088080e-01 1.35757893e-01 3.26222509e-01 -4.38577175e-01
3.89050007e-01 -4.19316202e-01 5.66954315e-01 6.93361282e-01
-1.41834877e-02 6.46028996e-01 4.73755389e-01 -1.14479221e-01
-9.46782053e-01 -1.38771713e+00 -3.15280497e-01 5.90771317e-01
4.49171245e-01 -1.88477591e-01 -8.35137665e-01 -7.67915905e-01
5.32480851e-02 1.25759685e+00 -8.55166554e-01 -8.56012583e-01
-6.95870221e-02 -8.28509986e-01 4.72832263e-01 2.94027716e-01
7.62283027e-01 -5.93692303e-01 -4.18012351e-01 -3.49874119e-03
1.27067780e-02 -9.95928943e-01 -2.84379926e-02 5.05798817e-01
-4.93159741e-01 -1.07389462e+00 -5.49397528e-01 -6.19077325e-01
3.20042282e-01 -1.22618891e-01 1.02203155e+00 -5.05971968e-01
-1.33936405e-01 5.34277320e-01 -2.96714783e-01 -3.63519877e-01
-4.89593595e-01 3.18504311e-02 -1.47425368e-01 2.82302350e-01
5.57063699e-01 -7.58610129e-01 -4.85638827e-01 1.07367076e-01
-9.13180232e-01 1.86213888e-02 4.54653591e-01 9.44381833e-01
4.49611336e-01 1.51917562e-01 4.76866245e-01 -9.53291655e-01
2.07054213e-01 -7.00571358e-01 -5.13060510e-01 6.23013794e-01
-1.07483935e+00 4.77937073e-01 5.11953235e-01 -2.83473223e-01
-1.16593814e+00 -9.34329703e-02 1.55970082e-02 -8.18999231e-01
1.75985992e-01 5.16000874e-02 -6.76675618e-01 8.27832893e-02
5.83013475e-01 2.91357487e-01 1.14305817e-01 -3.41976792e-01
5.69801092e-01 4.85189229e-01 4.15113240e-01 -5.95255017e-01
9.05199230e-01 4.71596122e-01 1.27091810e-01 -4.36977178e-01
-9.18069363e-01 -2.37651303e-01 -7.03491569e-01 4.09837924e-02
8.68082523e-01 -8.82475257e-01 -5.56634247e-01 3.89560878e-01
-1.05893481e+00 2.03280132e-02 -7.53441453e-01 5.37683189e-01
-6.56057358e-01 8.66937190e-02 -2.25386485e-01 -7.65237153e-01
-4.16203409e-01 -1.11311257e+00 1.01088929e+00 1.02232575e-01
2.81636596e-01 -1.43722236e+00 2.06186071e-01 -1.28669709e-01
4.05352235e-01 4.43218350e-01 9.34843183e-01 -9.87516940e-01
-6.36120498e-01 -2.21731275e-01 -1.00426875e-01 7.33340561e-01
1.51754767e-01 -2.50088304e-01 -1.22546077e+00 -3.27270359e-01
2.79030055e-01 -2.34304592e-01 1.19947791e+00 2.94951648e-01
9.00793850e-01 -2.94286966e-01 -2.69282430e-01 9.32510555e-01
1.76963007e+00 3.12533230e-01 7.86341190e-01 1.08336858e-01
6.27347708e-01 5.52478194e-01 2.53530771e-01 2.79343605e-01
4.10627395e-01 5.44376552e-01 4.53896105e-01 2.27608860e-01
-3.85186166e-01 -5.72379291e-01 3.75432521e-01 7.82494664e-01
3.52295071e-01 -4.66161281e-01 -7.80752838e-01 4.17315573e-01
-1.81216013e+00 -9.91494060e-01 4.28486347e-01 2.45774174e+00
7.01394498e-01 3.40152122e-02 -3.87144506e-01 2.90481955e-01
8.38178098e-01 1.54403508e-01 -4.85220641e-01 -1.00067714e-02
-2.76324600e-01 2.20421091e-01 6.67863190e-01 7.36357629e-01
-1.15073061e+00 7.69343793e-01 5.54453993e+00 1.21915948e+00
-1.04925001e+00 4.26509440e-01 5.45110047e-01 -1.31302020e-02
-6.16023719e-01 1.27334982e-01 -1.07309711e+00 5.88386595e-01
9.56246674e-01 -6.35252237e-01 3.00201297e-01 9.88406479e-01
-4.99525696e-01 3.63999933e-01 -1.40259027e+00 1.08438206e+00
2.06732571e-01 -1.25910413e+00 2.33481377e-01 9.14416686e-02
6.15751803e-01 -1.36749744e-01 1.84573486e-01 6.37144148e-01
4.66025084e-01 -9.75972414e-01 6.73930585e-01 6.25015616e-01
7.00196922e-01 -8.46327066e-01 7.98016071e-01 1.65605858e-01
-1.07957351e+00 1.65088773e-02 -5.72040260e-01 2.88019210e-01
-1.12130076e-01 7.11007774e-01 -6.22553468e-01 8.22480500e-01
3.44207793e-01 8.30735326e-01 -3.83456677e-01 7.24489033e-01
-2.09249824e-01 4.69216585e-01 -2.98171163e-01 8.88148174e-02
4.09973823e-02 -2.31748745e-01 5.23159325e-01 9.85379517e-01
6.81209207e-01 -2.77130902e-01 -2.40627587e-01 1.38007641e+00
-9.61196125e-02 -9.45601612e-02 -8.60747576e-01 9.50448737e-02
6.74235463e-01 9.69111502e-01 -4.96100456e-01 -2.53719956e-01
-1.37417644e-01 8.76968861e-01 2.84476101e-01 3.72030258e-01
-1.04990053e+00 -4.59109455e-01 6.47578657e-01 -2.83386528e-01
1.12430647e-01 3.08933884e-01 -2.22718164e-01 -1.23905563e+00
-1.54116321e-02 -3.34592640e-01 3.52147847e-01 -4.31584120e-01
-1.33678508e+00 6.90877318e-01 3.20362061e-01 -1.35097003e+00
-2.14549705e-01 -5.51144183e-01 -4.71180677e-01 8.53880286e-01
-1.53837669e+00 -1.19846964e+00 -3.37025553e-01 7.42276609e-01
1.48355886e-01 -4.88092005e-01 9.84643579e-01 3.00912440e-01
-4.87105936e-01 8.85854006e-01 4.76710111e-01 2.87349299e-02
4.06628162e-01 -1.32191229e+00 9.97457653e-02 7.59684801e-01
2.46311381e-01 5.23453474e-01 6.38552666e-01 -4.88978088e-01
-1.26745296e+00 -1.29015481e+00 6.46750271e-01 -3.07560682e-01
4.43770528e-01 -5.92691839e-01 -8.62101674e-01 7.66261160e-01
1.06330700e-01 1.83669969e-01 6.44216895e-01 5.00266179e-02
-6.43138647e-01 -4.89740103e-01 -1.34125936e+00 1.98365360e-01
1.02539456e+00 -6.53098464e-01 -7.21971810e-01 2.94859648e-01
1.08535767e+00 -1.62586704e-01 -9.35675740e-01 5.39624691e-01
4.16425198e-01 -1.00773180e+00 1.07622123e+00 -2.96747267e-01
3.91109586e-01 -1.15149096e-01 -5.91044366e-01 -1.45367634e+00
-1.67676091e-01 -3.53652060e-01 -1.84184358e-01 1.62300932e+00
2.79387355e-01 -7.80443549e-01 5.71848094e-01 6.15906775e-01
-9.08347219e-03 -5.38609385e-01 -1.30316293e+00 -1.18265021e+00
4.13905591e-01 -3.89021248e-01 6.51191175e-01 9.89496708e-01
-2.51027793e-01 3.08705181e-01 -4.68393475e-01 1.85732707e-01
1.11977267e+00 -1.32875675e-02 4.15275633e-01 -1.42546451e+00
-2.46923581e-01 -4.46589768e-01 -7.56315887e-01 -5.48344672e-01
4.44646001e-01 -1.23830080e+00 -1.54527158e-01 -1.41282213e+00
2.59843349e-01 -4.54463929e-01 -6.19222522e-01 3.65960538e-01
1.03807181e-01 -1.78939641e-01 3.28695834e-01 3.18929590e-02
-1.87122241e-01 1.00321198e+00 7.83824265e-01 -2.02944100e-01
7.64322728e-02 -1.63956940e-01 -6.63084090e-01 4.76570278e-01
6.58195913e-01 -5.46774387e-01 -9.59627032e-01 -1.18619204e-01
9.95785370e-02 -1.75126135e-01 5.70128560e-01 -1.18000197e+00
2.29060605e-01 8.45616887e-06 1.16231389e-01 -2.62099028e-01
5.37982583e-01 -1.05593657e+00 4.72670078e-01 4.71134156e-01
-5.23255765e-01 -7.80571282e-01 -1.01971045e-01 7.54264176e-01
-2.36785710e-01 -2.68798053e-01 1.09992898e+00 2.23047093e-01
-5.65588892e-01 3.37405175e-01 4.19495523e-01 1.51161924e-01
1.34078121e+00 -3.20895910e-02 -4.07447636e-01 -2.80936807e-01
-7.72586524e-01 1.63518358e-02 3.39173824e-01 5.76383352e-01
5.56466103e-01 -1.74306619e+00 -4.91136998e-01 3.74148965e-01
1.32518828e-01 -1.49841204e-01 2.94045687e-01 4.06548202e-01
-1.58191487e-01 4.60595608e-01 -2.20692098e-01 -5.18659234e-01
-7.65172839e-01 6.77574873e-01 4.97201651e-01 -2.05004543e-01
-5.74330866e-01 8.90927196e-01 5.80146670e-01 -5.95017612e-01
4.56794202e-01 -1.90929636e-01 -3.24711278e-02 9.72332954e-02
2.86670744e-01 3.77503067e-01 7.12045804e-02 -6.01598024e-01
-3.91373038e-01 6.00367010e-01 1.61931396e-01 1.60208512e-02
1.22454834e+00 -1.88752860e-01 2.19822749e-01 6.16237104e-01
1.73537791e+00 -5.03564656e-01 -1.27283084e+00 -5.07162690e-01
-2.52085358e-01 -3.32909226e-01 2.11537644e-01 -6.77853763e-01
-1.20853364e+00 1.21813095e+00 9.09210205e-01 1.54007420e-01
9.34012532e-01 9.97530892e-02 4.02924329e-01 2.21547306e-01
4.29665238e-01 -8.98756385e-01 -1.22250944e-01 3.46332967e-01
8.75473022e-01 -1.16005933e+00 -3.65336865e-01 -3.99954975e-01
-5.88918507e-01 1.09353185e+00 4.24473822e-01 -2.31281698e-01
9.86884236e-01 1.83654606e-01 -4.00629342e-01 -1.22757658e-01
-7.22293854e-01 -3.32117796e-01 5.15729129e-01 6.96210623e-01
9.40001309e-02 1.11956760e-01 -5.85243851e-02 7.97558546e-01
-3.53882879e-01 -1.55593410e-01 3.04623276e-01 3.76213580e-01
-3.84840339e-01 -9.92623508e-01 -3.44880787e-03 2.30038717e-01
-6.51536137e-02 -1.51610151e-01 -8.60933140e-02 9.23898220e-01
2.24751428e-01 6.93240583e-01 1.37866020e-01 -5.42034864e-01
1.94343075e-01 3.07282537e-01 4.40636605e-01 -5.49917281e-01
-1.14858910e-01 -1.88892990e-01 -3.04848224e-01 -3.81536663e-01
-3.52687627e-01 -4.43203598e-01 -1.05373514e+00 -2.41604522e-01
-3.23688686e-01 1.07422724e-01 6.35297656e-01 9.56624866e-01
4.02191430e-01 5.96155941e-01 6.23064995e-01 -4.58918035e-01
-8.77795637e-01 -7.41450727e-01 -8.75809014e-01 5.33037424e-01
2.00961933e-01 -1.06726444e+00 -7.94434488e-01 -1.57345086e-01] | [9.941956520080566, 2.6592588424682617] |
0ddc4746-94d4-4c25-9ca4-776e0907aa63 | is-translation-helpful-an-empirical-analysis | 2305.12480 | null | https://arxiv.org/abs/2305.12480v1 | https://arxiv.org/pdf/2305.12480v1.pdf | Is Translation Helpful? An Empirical Analysis of Cross-Lingual Transfer in Low-Resource Dialog Generation | Cross-lingual transfer is important for developing high-quality chatbots in multiple languages due to the strongly imbalanced distribution of language resources. A typical approach is to leverage off-the-shelf machine translation (MT) systems to utilize either the training corpus or developed models from high-resource languages. In this work, we investigate whether it is helpful to utilize MT at all in this task. To do so, we simulate a low-resource scenario assuming access to limited Chinese dialog data in the movie domain and large amounts of English dialog data from multiple domains. Experiments show that leveraging English dialog corpora can indeed improve the naturalness, relevance and cross-domain transferability in Chinese. However, directly using English dialog corpora in its original form, surprisingly, is better than using its translated version. As the topics and wording habits in daily conversations are strongly culture-dependent, MT can reinforce the bias from high-resource languages, yielding unnatural generations in the target language. Considering the cost of translating large amounts of text and the strong effects of the translation quality, we suggest future research should rather focus on utilizing the original English data for cross-lingual transfer in dialog generation. We perform extensive human evaluations and ablation studies. The analysis results, together with the collected dataset, are presented to draw attention towards this area and benefit future research. | ['Xiaoyu Shen', 'Shuai Yu', 'Lei Shen'] | 2023-05-21 | null | null | null | null | ['cross-lingual-transfer'] | ['natural-language-processing'] | [-4.00037691e-02 1.51109353e-01 8.19590464e-02 -2.94958800e-01
-1.05788791e+00 -8.55729580e-01 7.05519080e-01 -4.63520974e-01
-6.19016528e-01 1.14446259e+00 4.71732020e-01 -3.99044633e-01
5.64883530e-01 -4.69374150e-01 -3.80080789e-01 -3.96614999e-01
5.17320037e-01 8.50446045e-01 8.81707668e-03 -7.48850703e-01
3.28457579e-02 1.28871396e-01 -9.04283047e-01 4.74266768e-01
1.24956286e+00 8.88971537e-02 5.48449636e-01 4.17841434e-01
-3.47760469e-01 7.54064143e-01 -1.05953264e+00 -8.10153604e-01
1.71487048e-01 -8.95669937e-01 -9.91470873e-01 6.11844063e-02
9.80739072e-02 -4.39089686e-01 8.22424367e-02 7.90417433e-01
5.84137738e-01 9.47504013e-04 4.24696863e-01 -8.91660035e-01
-9.15904522e-01 8.13158751e-01 -1.71574369e-01 1.43024931e-03
3.43889356e-01 4.66286391e-01 9.07757044e-01 -9.80944753e-01
8.18130612e-01 1.34164751e+00 4.36619103e-01 9.35325503e-01
-1.22791243e+00 -5.83498597e-01 3.94942984e-02 -2.67844081e-01
-1.09369886e+00 -6.35955334e-01 6.48814261e-01 -3.46968502e-01
1.01752055e+00 1.18286878e-01 2.93605745e-01 1.65824759e+00
-2.70088352e-02 6.01447463e-01 1.34311199e+00 -7.76158512e-01
-1.14012919e-01 8.39480817e-01 -3.12185138e-01 3.32892269e-01
-5.64117581e-02 -3.80027667e-02 -7.60222912e-01 -1.04640089e-01
4.92831409e-01 -5.54372489e-01 -2.67560989e-01 2.32035637e-01
-1.17020071e+00 9.96109307e-01 1.08527184e-01 7.34120548e-01
-2.41255537e-01 -4.00202900e-01 4.26291645e-01 7.45831013e-01
7.48577595e-01 1.03556097e+00 -6.00764453e-01 -6.35611773e-01
-6.17328584e-01 2.08274052e-01 1.20946813e+00 1.28462112e+00
7.28625536e-01 -6.94689974e-02 -2.30510756e-01 1.29040420e+00
-8.09305310e-02 5.94371498e-01 7.56724119e-01 -8.54569018e-01
8.95849645e-01 5.20636857e-01 2.28114948e-01 -5.02718747e-01
-1.01011798e-01 9.64403991e-03 -2.99189419e-01 -6.02437630e-02
8.48286211e-01 -7.73310900e-01 -3.60799253e-01 2.11640549e+00
1.90760747e-01 -6.60704553e-01 3.95190924e-01 8.33660483e-01
4.22381818e-01 5.52717686e-01 4.73829508e-02 -2.34166518e-01
1.31777298e+00 -1.15929937e+00 -7.37660646e-01 -5.72744310e-01
9.22415316e-01 -1.24319208e+00 1.85975587e+00 -9.91804227e-02
-9.95547116e-01 -4.02450562e-01 -7.56164432e-01 -7.99111947e-02
-3.08831573e-01 2.51486331e-01 2.54999459e-01 6.98799729e-01
-9.45236742e-01 3.63633662e-01 -6.00279152e-01 -8.59222531e-01
-2.11949229e-01 2.19666213e-02 -1.15023203e-01 -1.57048479e-01
-1.51058316e+00 1.41851044e+00 7.62116862e-03 -7.05525000e-03
-5.46908379e-01 -4.41610515e-01 -6.18198872e-01 -1.66115984e-01
2.87203014e-01 -3.98197353e-01 1.49770308e+00 -1.23026383e+00
-1.85272849e+00 7.88076043e-01 4.77892756e-02 -5.68071119e-02
7.31996715e-01 -2.90997058e-01 -1.56822339e-01 -1.31104842e-01
1.70749620e-01 6.97873592e-01 4.30557698e-01 -1.08397245e+00
-4.77961421e-01 -1.62534893e-01 1.87020347e-01 5.12786031e-01
-6.89641416e-01 2.85335988e-01 -2.56514311e-01 -5.35816729e-01
-6.13352478e-01 -1.29669070e+00 -8.42741504e-02 -4.95819032e-01
-2.07036033e-01 -1.53841645e-01 4.34876531e-01 -9.38998461e-01
9.64239359e-01 -1.85103321e+00 2.21054897e-01 -3.40374172e-01
-3.09792846e-01 1.67967930e-01 -2.88674086e-01 9.02024925e-01
5.47333479e-01 2.84421563e-01 -8.67718607e-02 -4.62816417e-01
-1.15557499e-01 2.42537975e-01 -1.85993433e-01 -2.71259602e-02
4.88766611e-01 9.05189633e-01 -7.75693655e-01 -4.80812788e-01
-7.26236030e-02 2.29620874e-01 -4.97228652e-01 5.80916941e-01
-3.07643980e-01 8.27145636e-01 -3.78145367e-01 3.87144625e-01
2.68966496e-01 -2.87093781e-02 4.76954043e-01 2.41311416e-01
-1.67078048e-01 8.28241229e-01 -4.28957909e-01 1.81418228e+00
-1.01134944e+00 6.46812499e-01 1.36697039e-01 -3.57987702e-01
8.49855840e-01 5.66027284e-01 1.08276747e-01 -7.76041389e-01
8.59504044e-02 5.03861010e-01 5.57116508e-01 -5.33647716e-01
7.13909388e-01 -4.97808933e-01 -2.74623364e-01 7.42398858e-01
1.53921157e-01 -4.10428137e-01 3.15727890e-01 1.60505027e-01
9.04323876e-01 1.10212684e-01 4.65388000e-02 -3.98287386e-01
2.93059856e-01 4.29598600e-01 4.49986070e-01 4.09826636e-01
-2.17510179e-01 4.66147751e-01 4.47061777e-01 1.76819190e-01
-1.08261549e+00 -7.87187219e-01 6.64837807e-02 1.30876112e+00
-2.01324210e-01 -3.57582390e-01 -9.84593570e-01 -8.34982455e-01
-4.01316404e-01 9.94677126e-01 -2.59886354e-01 -1.75044313e-01
-8.22231829e-01 -5.79832911e-01 8.67019594e-01 2.27534547e-01
4.32901919e-01 -1.20386779e+00 -2.62363136e-01 3.88185382e-01
-7.26371646e-01 -1.22281134e+00 -6.79695189e-01 -1.06563695e-01
-6.23033762e-01 -5.99915564e-01 -1.01677096e+00 -8.15515041e-01
4.29693967e-01 2.45974332e-01 1.37561738e+00 2.82623563e-02
2.72972554e-01 1.38596669e-01 -6.00494385e-01 -3.82376701e-01
-1.18940187e+00 5.90862393e-01 1.51451141e-01 -3.63134295e-01
5.69561601e-01 -4.83271420e-01 -5.75484112e-02 5.18772125e-01
-6.34413123e-01 1.84326366e-01 6.84226871e-01 1.11231148e+00
-2.39758313e-01 -4.64682043e-01 9.87922251e-01 -1.09239459e+00
1.05802989e+00 -4.11107540e-01 -4.03115511e-01 4.63356167e-01
-4.07725990e-01 -5.82471350e-03 9.40810621e-01 -6.91338003e-01
-1.50904644e+00 -3.48771036e-01 -1.74073622e-01 2.23201583e-03
-8.18966255e-02 5.34963489e-01 -2.58007228e-01 3.53866965e-01
9.45530713e-01 2.21614316e-02 1.20611548e-01 -5.67808688e-01
4.52328265e-01 1.08372068e+00 2.71928981e-02 -9.64352071e-01
5.99294066e-01 -5.81190512e-02 -8.92657399e-01 -8.48127425e-01
-5.53351700e-01 -1.69032902e-01 -6.76247120e-01 -2.28912905e-01
9.23143744e-01 -9.37179625e-01 -1.96991742e-01 3.91563714e-01
-1.46976423e+00 -7.99922824e-01 -2.04165541e-02 4.92885917e-01
-3.01997125e-01 -1.33715812e-02 -9.18769360e-01 -7.61445999e-01
-2.32660413e-01 -1.42976165e+00 8.93965065e-01 7.57925957e-02
-6.34485841e-01 -1.16369438e+00 3.32057178e-01 5.99899411e-01
6.41377032e-01 -4.24679518e-01 9.30284441e-01 -1.00398076e+00
-3.18944603e-01 -1.04048196e-02 6.50104061e-02 5.19171596e-01
6.07312560e-01 -2.72993147e-01 -9.48813140e-01 -3.47451031e-01
2.02527493e-01 -8.93052459e-01 1.81170981e-02 -2.88700134e-01
9.11875516e-02 -4.36495781e-01 1.10695012e-01 -1.68315202e-01
1.03017485e+00 4.64470796e-02 2.70981342e-01 1.25681698e-01
6.30425334e-01 1.28296649e+00 7.97327936e-01 1.48244932e-01
5.59673607e-01 8.75151575e-01 -4.24159914e-01 -2.07046464e-01
-1.81851059e-01 -3.60139847e-01 8.67688835e-01 1.51705956e+00
2.10649654e-01 -3.15894693e-01 -9.94095027e-01 6.25693321e-01
-1.56050611e+00 -7.16512263e-01 1.34347469e-01 2.07616878e+00
1.51667988e+00 -4.61924709e-02 2.70559967e-01 -7.54012227e-01
6.62099242e-01 -1.36593729e-01 -2.57681131e-01 -5.50042391e-01
-1.55584306e-01 -4.54962701e-02 2.15452075e-01 6.27543986e-01
-4.68241155e-01 1.41723502e+00 5.82612705e+00 7.38671541e-01
-1.31101882e+00 4.17834729e-01 6.77674472e-01 -9.19068828e-02
-3.19161564e-01 1.21150605e-01 -7.45136917e-01 5.84830165e-01
1.23029053e+00 -3.57732654e-01 6.44224584e-01 6.41254902e-01
3.43866378e-01 -4.40080427e-02 -1.17178154e+00 4.34991270e-01
-7.13678682e-03 -8.84340405e-01 -1.75908089e-01 1.31533325e-01
6.61252081e-01 1.67087406e-01 -1.08138837e-01 6.38933897e-01
7.68353224e-01 -8.46997738e-01 5.24222851e-01 -2.72242188e-01
8.15463006e-01 -4.46529895e-01 7.42853224e-01 6.75964236e-01
-7.31885612e-01 4.20804799e-01 -4.16419685e-01 -1.98552594e-01
3.77124697e-01 7.06594959e-02 -1.28710258e+00 4.38095033e-01
4.19686347e-01 3.05745661e-01 -5.91570377e-01 1.11881696e-01
-2.68513471e-01 7.99250007e-01 -3.18336934e-01 -2.96842784e-01
2.52305806e-01 -4.63706434e-01 3.88963997e-01 1.32606018e+00
2.91039973e-01 -1.48411006e-01 2.07934946e-01 9.04981911e-01
-2.20819727e-01 4.74842578e-01 -9.85797405e-01 -2.58888304e-01
6.57018483e-01 1.21587527e+00 -3.39589983e-01 -2.83695787e-01
-7.62163579e-01 1.19363999e+00 6.61122441e-01 3.27408910e-01
-5.85782886e-01 -3.84304486e-02 6.81851387e-01 -1.53295472e-01
-1.20899215e-01 -2.96126634e-01 -5.06360471e-01 -1.18297625e+00
1.59487024e-01 -1.50842667e+00 1.68954939e-01 -4.76748198e-01
-1.70784593e+00 1.04257643e+00 -1.51454821e-01 -1.29135418e+00
-6.85102820e-01 -4.06987995e-01 -4.27450866e-01 1.18971753e+00
-1.30323851e+00 -1.33859694e+00 1.13631785e-01 4.49894160e-01
1.12452900e+00 -3.06892604e-01 8.72977734e-01 4.39846188e-01
-5.48686564e-01 9.12135839e-01 -2.52588782e-02 1.65956110e-01
1.35947585e+00 -1.07895446e+00 4.40376461e-01 8.56693506e-01
1.78521752e-01 8.58926892e-01 7.24595904e-01 -7.93570936e-01
-1.15939999e+00 -8.35330009e-01 1.09313810e+00 -7.85179317e-01
8.37749183e-01 -6.79270625e-01 -1.05944121e+00 6.43925369e-01
9.02882695e-01 -8.49023998e-01 7.43084967e-01 3.69710207e-01
1.89314410e-02 3.14181983e-01 -1.01320446e+00 1.07649267e+00
8.39073718e-01 -6.62127018e-01 -6.33248985e-01 3.35609287e-01
9.71055925e-01 -3.15228850e-01 -9.24093962e-01 -1.87791325e-02
3.62438589e-01 -5.80924511e-01 4.02458459e-01 -6.91377401e-01
6.34786487e-01 2.12380230e-01 -4.78021316e-02 -1.67733264e+00
4.97999936e-02 -8.37167501e-01 6.49164319e-01 1.79087687e+00
9.63020265e-01 -6.01086378e-01 5.74714780e-01 1.05244708e+00
-7.09987655e-02 -4.25635189e-01 -8.52269828e-01 -8.88544202e-01
7.59597838e-01 -6.27709106e-02 4.38575447e-01 1.10328376e+00
4.15337175e-01 1.12491333e+00 -6.53719783e-01 -2.13060349e-01
-2.28690878e-02 3.39705609e-02 1.15040982e+00 -6.78179741e-01
-4.13323849e-01 -2.27803931e-01 4.25798595e-01 -1.09142828e+00
4.84419823e-01 -7.88918495e-01 4.88011539e-01 -1.04005885e+00
-5.49624078e-02 -7.93201983e-01 2.97633946e-01 2.47448578e-01
-5.50530314e-01 9.67642814e-02 4.74847883e-01 3.91873419e-01
-1.62259057e-01 7.83645093e-01 1.37195158e+00 1.59315377e-01
-5.50377548e-01 2.58704219e-02 -8.58641863e-01 4.66955215e-01
9.13011968e-01 -5.54826736e-01 -4.85612661e-01 -8.16875875e-01
-2.75498331e-01 8.63519385e-02 -3.64151835e-01 -5.41785181e-01
4.03014533e-02 -3.02233636e-01 -2.50756919e-01 6.50168732e-02
5.84374905e-01 -6.88894808e-01 -2.34439716e-01 2.15161905e-01
-5.21532476e-01 4.00221497e-01 2.35084593e-01 3.85324955e-02
-2.67186731e-01 -3.24763685e-01 7.07175255e-01 -4.04203534e-01
-2.16712058e-01 -2.36646980e-01 -7.43629456e-01 4.33720946e-01
6.81188166e-01 8.74219164e-02 -4.88628775e-01 -6.91678226e-01
-2.11904272e-01 1.83323562e-01 8.51874948e-01 6.86946094e-01
3.17003652e-02 -1.27705586e+00 -9.68174994e-01 -4.57705110e-02
2.58338958e-01 -3.47758383e-01 -4.81435955e-02 7.42051899e-01
-4.19333130e-01 4.85908210e-01 -3.44813555e-01 -3.65830004e-01
-1.02480352e+00 1.87529877e-01 2.35939011e-01 -5.67298889e-01
-1.74899921e-01 6.67412937e-01 3.05810302e-01 -8.57108116e-01
-4.95350510e-02 7.06034303e-02 1.48614869e-02 4.81674299e-02
2.82440245e-01 1.68957159e-01 1.19399831e-01 -5.42420208e-01
-1.80730611e-01 -5.82996234e-02 -2.86099523e-01 -9.12711799e-01
1.04015601e+00 -3.65641177e-01 -3.80168259e-02 6.35782897e-01
1.01097548e+00 6.03419483e-01 -1.15343034e+00 -3.42494905e-01
2.30576932e-01 -4.69430745e-01 -6.42266333e-01 -1.01073134e+00
-7.05738902e-01 8.27124715e-01 1.18988097e-01 1.22169890e-01
8.02940011e-01 1.01510152e-01 8.90427351e-01 5.01174688e-01
5.36296606e-01 -1.32452238e+00 2.34355286e-01 7.74940550e-01
1.00997961e+00 -1.44557357e+00 -4.57687557e-01 -3.50702137e-01
-1.30796945e+00 7.64007807e-01 1.13262022e+00 2.54637331e-01
2.15341210e-01 3.14876050e-01 7.96055973e-01 4.75095771e-02
-9.98915553e-01 -1.66748777e-01 -9.96712670e-02 5.71213484e-01
9.44798529e-01 -1.55054806e-02 -5.65901756e-01 5.67819357e-01
-5.80254018e-01 -3.30937594e-01 7.13585556e-01 8.21374655e-01
4.69231457e-02 -1.74008036e+00 -3.26036125e-01 1.74231023e-01
-5.29776514e-01 -4.66618389e-01 -9.98483121e-01 8.99294555e-01
-2.90423632e-01 1.44672394e+00 -1.47578448e-01 -4.43928570e-01
3.91224831e-01 3.48830163e-01 2.54658371e-01 -9.78145480e-01
-1.00614512e+00 1.54732957e-01 6.04788840e-01 -7.78255910e-02
-2.63255775e-01 -5.88687658e-01 -8.11527729e-01 -4.05067742e-01
-4.93342221e-01 2.92676479e-01 5.39970517e-01 1.03701723e+00
3.79235268e-01 2.36129597e-01 6.21428788e-01 -6.07797921e-01
-7.63843536e-01 -1.53753102e+00 -1.26032218e-01 4.98903841e-01
-1.19195521e-01 -4.23968256e-01 -2.26002797e-01 5.18938750e-02] | [12.507572174072266, 8.43808650970459] |
0617ad2e-d2f0-41d9-8677-75498e2f472a | weakly-supervised-spatial-context-networks | 1704.02998 | null | http://arxiv.org/abs/1704.02998v2 | http://arxiv.org/pdf/1704.02998v2.pdf | Weakly-Supervised Spatial Context Networks | We explore the power of spatial context as a self-supervisory signal for
learning visual representations. In particular, we propose spatial context
networks that learn to predict a representation of one image patch from another
image patch, within the same image, conditioned on their real-valued relative
spatial offset. Unlike auto-encoders, that aim to encode and reconstruct
original image patches, our network aims to encode and reconstruct intermediate
representations of the spatially offset patches. As such, the network learns a
spatially conditioned contextual representation. By testing performance with
various patch selection mechanisms we show that focusing on object-centric
patches is important, and that using object proposal as a patch selection
mechanism leads to the highest improvement in performance. Further, unlike
auto-encoders, context encoders [21], or other forms of unsupervised feature
learning, we illustrate that contextual supervision (with pre-trained model
initialization) can improve on existing pre-trained model performance. We build
our spatial context networks on top of standard VGG_19 and CNN_M architectures
and, among other things, show that we can achieve improvements (with no
additional explicit supervision) over the original ImageNet pre-trained VGG_19
and CNN_M models in object categorization and detection on VOC2007. | ['Larry S. Davis', 'Zuxuan Wu', 'Leonid Sigal'] | 2017-04-10 | null | null | null | null | ['object-categorization'] | ['computer-vision'] | [ 4.27120298e-01 2.50712246e-01 -3.20392370e-01 -4.98772979e-01
-4.98901010e-01 -4.15625423e-01 8.81953776e-01 3.05247545e-01
-3.90374541e-01 5.68632007e-01 3.43347490e-01 -3.04942638e-01
-5.61236516e-02 -8.28644216e-01 -1.20883811e+00 -6.00293875e-01
-2.49680057e-01 9.34771821e-03 3.31556231e-01 -2.72208489e-02
4.15562451e-01 4.93930131e-01 -1.73428905e+00 7.66150534e-01
3.18803698e-01 1.19880939e+00 5.54535985e-01 8.36647987e-01
2.08954401e-02 9.64206040e-01 -4.65198576e-01 2.70097911e-01
1.17980279e-01 -3.89685988e-01 -1.15297782e+00 2.04869375e-01
1.08641589e+00 -1.47684723e-01 -4.43259060e-01 7.52092004e-01
9.29998979e-02 8.89471471e-02 8.28734815e-01 -9.65516210e-01
-1.23408532e+00 4.25299883e-01 -2.94509858e-01 3.06721091e-01
2.21080240e-03 3.40131313e-01 1.13553095e+00 -1.21586251e+00
7.59075105e-01 1.04809809e+00 6.32506609e-01 5.03686845e-01
-1.69103086e+00 -2.43567005e-01 5.89863539e-01 3.31049353e-01
-1.52380025e+00 -3.13008100e-01 6.42233968e-01 -3.80193681e-01
1.30562723e+00 -2.07223818e-02 5.32425046e-01 1.22125900e+00
1.93040356e-01 8.38513851e-01 9.03441191e-01 -5.71302295e-01
2.75323659e-01 9.32338983e-02 8.70806202e-02 4.65296179e-01
-2.32598092e-02 3.78968298e-01 -5.36832929e-01 2.06938326e-01
9.61906612e-01 1.62337288e-01 -2.22494885e-01 -5.09202242e-01
-1.29524767e+00 8.06626976e-01 1.24635983e+00 3.09185356e-01
-2.44502321e-01 6.62706256e-01 5.88253848e-02 2.83645689e-01
2.31044069e-01 5.31009436e-01 -4.12928313e-01 4.00518924e-01
-1.01447582e+00 6.26005158e-02 2.87076414e-01 1.24287999e+00
1.17026460e+00 7.16765001e-02 -3.09883893e-01 4.55713242e-01
4.25613672e-02 2.30188742e-01 4.33270574e-01 -9.71454799e-01
2.36638322e-01 5.48864067e-01 -7.09323063e-02 -8.65159452e-01
-2.89167851e-01 -8.06855440e-01 -6.95679009e-01 4.17656720e-01
2.16264337e-01 1.64053068e-01 -1.19192922e+00 1.88048792e+00
-1.42320335e-01 6.49788618e-01 3.05838161e-03 8.76247942e-01
6.58591330e-01 6.27238929e-01 2.71750987e-01 3.45891446e-01
1.03852832e+00 -1.02933633e+00 -2.03046530e-01 -5.85299373e-01
6.35866582e-01 -4.21221346e-01 1.05760896e+00 2.17651248e-01
-7.55392730e-01 -1.00398910e+00 -1.24887538e+00 -1.96246609e-01
-6.01564288e-01 1.88664958e-01 6.03503048e-01 4.33820151e-02
-1.42777574e+00 7.90121853e-01 -6.11735404e-01 -4.22475785e-01
6.27356410e-01 2.79799432e-01 -6.59413815e-01 -2.60798689e-02
-7.63441026e-01 8.89832318e-01 4.32386130e-01 -1.04678050e-01
-1.35311222e+00 -7.17221081e-01 -9.49095309e-01 3.19658108e-02
1.63450181e-01 -5.65926075e-01 9.82032061e-01 -1.31527352e+00
-8.76949012e-01 7.55157769e-01 -1.78548485e-01 -7.96382308e-01
-1.56751230e-01 -8.36742520e-02 -2.98009604e-01 3.41934115e-01
1.67787164e-01 1.55707443e+00 1.01473224e+00 -1.50852478e+00
-7.57438540e-01 -1.49376959e-01 2.09662259e-01 6.44006282e-02
-1.19169457e-02 -3.36182237e-01 -3.36780310e-01 -6.36467695e-01
7.98293501e-02 -7.65714228e-01 -3.67494076e-01 2.61688888e-01
-2.87498116e-01 -5.22422120e-02 1.11499047e+00 -3.13436419e-01
7.39699364e-01 -2.44929123e+00 9.18692052e-02 1.91981941e-01
8.11417401e-03 1.49273038e-01 -5.41035473e-01 2.19569951e-01
-3.41224492e-01 1.58277676e-01 -4.57053035e-01 -3.74207228e-01
-2.29682431e-01 5.38044274e-01 -5.59072971e-01 4.01627272e-01
7.79793680e-01 1.04694951e+00 -9.41380858e-01 -3.04332197e-01
3.93300742e-01 6.93895936e-01 -8.02418113e-01 2.14157440e-02
-4.58567917e-01 4.38703984e-01 4.14864086e-02 3.17103326e-01
3.53573591e-01 -4.93832529e-01 1.89516753e-01 -2.01821715e-01
1.24621382e-02 4.86727148e-01 -1.03159332e+00 1.99087334e+00
-7.55351841e-01 1.03139198e+00 -8.63978639e-02 -1.23063958e+00
7.88597703e-01 2.17687428e-01 4.61898861e-04 -9.44878459e-01
-1.72117621e-01 -7.21548423e-02 -1.24647200e-01 -1.34031206e-01
7.09562540e-01 1.76346615e-01 8.51734877e-02 3.02460611e-01
6.02586150e-01 1.70136392e-01 -1.91305429e-01 3.24182302e-01
1.00957763e+00 2.85218805e-01 5.63188642e-02 -3.96871060e-01
1.91565588e-01 1.09162316e-01 2.32857391e-01 8.55989754e-01
-1.11499075e-02 1.01538515e+00 4.18985635e-01 -4.88316566e-01
-1.02793789e+00 -1.09537685e+00 -3.25079709e-01 1.33904445e+00
2.38610759e-01 -4.93892401e-01 -4.48170066e-01 -7.91238427e-01
1.67580098e-01 7.00012386e-01 -1.23910201e+00 -1.85616851e-01
-6.67865098e-01 -1.27329782e-01 2.18340188e-01 1.11103320e+00
3.90860140e-01 -1.39447260e+00 -7.65774190e-01 1.85246781e-01
1.27824903e-01 -8.04356873e-01 -3.18503916e-01 8.69680882e-01
-9.42984760e-01 -1.03545725e+00 -3.54804873e-01 -1.09352839e+00
8.78811181e-01 3.75195444e-01 1.26722229e+00 2.95401067e-01
-4.51048225e-01 5.53150475e-01 -3.48801076e-01 -8.32607448e-02
1.07071185e-02 2.02336356e-01 -3.91444981e-01 -1.43530041e-01
1.75085247e-01 -6.40535951e-01 -9.35715199e-01 3.07587087e-01
-9.90029335e-01 5.00047989e-02 6.36332750e-01 1.12964678e+00
9.31446016e-01 -2.61983544e-01 5.86254418e-01 -8.71723473e-01
1.54749323e-02 -5.40384769e-01 -3.79548043e-01 1.32888734e-01
-4.67406720e-01 3.49072218e-01 5.36137521e-01 -3.35693479e-01
-6.86903238e-01 2.91355729e-01 6.36543408e-02 -5.21584511e-01
-5.11341393e-01 4.08566833e-01 5.29714078e-02 -6.39456660e-02
1.01978707e+00 2.86931425e-01 -9.65754762e-02 -4.53059345e-01
8.34103942e-01 2.41562873e-01 8.60406697e-01 -5.81623077e-01
5.10877430e-01 5.13375282e-01 -4.91045304e-02 -5.28977692e-01
-8.19103301e-01 -3.60009581e-01 -9.08655047e-01 2.22456276e-01
1.08303380e+00 -9.96085405e-01 -3.73504668e-01 -1.89095989e-01
-1.07100308e+00 -7.53507733e-01 -5.87782204e-01 2.36149788e-01
-7.19107866e-01 -1.60214797e-01 -3.82202923e-01 -4.59853202e-01
1.99470907e-01 -1.04652941e+00 1.17038238e+00 -7.71154417e-03
-2.28417143e-01 -1.12364745e+00 -3.03660840e-01 -3.37816626e-01
5.45417845e-01 1.43563315e-01 8.44304025e-01 -3.33371788e-01
-1.00921977e+00 2.12535039e-01 -4.23002273e-01 3.51028800e-01
1.30396217e-01 -1.98514566e-01 -1.09463382e+00 -3.38452786e-01
-3.98785740e-01 -4.27067369e-01 1.29328763e+00 3.95454317e-01
1.38112926e+00 -3.50375473e-01 -5.10208845e-01 7.66987383e-01
1.68713915e+00 4.64811549e-02 6.57068670e-01 3.75669092e-01
6.75449669e-01 5.05231142e-01 4.45029616e-01 1.03756592e-01
1.27912000e-01 6.31404817e-01 9.04978096e-01 -1.84421033e-01
-5.43574870e-01 -4.84830767e-01 3.59902345e-02 2.14042626e-02
1.63773373e-01 2.60902308e-02 -7.66199887e-01 9.70389366e-01
-1.76124406e+00 -9.45479393e-01 2.24929124e-01 2.11345005e+00
7.37950087e-01 -2.95895413e-02 -3.65802906e-02 -8.00414905e-02
5.69414437e-01 3.00895840e-01 -5.78713477e-01 -3.72030467e-01
-2.18652502e-01 5.84212124e-01 6.19512200e-01 6.34573936e-01
-1.27221060e+00 1.09331179e+00 7.15270662e+00 4.39464420e-01
-1.39579415e+00 7.17282519e-02 6.38808906e-01 -1.98545456e-01
-4.10227060e-01 1.51794493e-01 -6.66144431e-01 2.42336571e-01
8.24365437e-01 3.89110535e-01 3.64888757e-01 1.10421968e+00
-3.80326986e-01 -9.22418833e-02 -1.63164830e+00 7.96097577e-01
6.31898195e-02 -1.69447911e+00 2.32280493e-01 1.98827863e-01
8.58469248e-01 9.45878699e-02 4.57233071e-01 3.55372965e-01
2.76491761e-01 -1.52310598e+00 9.50976551e-01 3.85958165e-01
8.55828285e-01 -6.61173403e-01 4.69266981e-01 7.27416277e-02
-1.16009665e+00 -2.80270278e-01 -7.64342904e-01 -6.45866320e-02
-2.40744650e-01 2.39052966e-01 -9.25439894e-01 3.27988565e-01
7.40352869e-01 9.82809424e-01 -9.04228210e-01 9.45325732e-01
-5.45283496e-01 7.14186907e-01 -6.48110285e-02 3.33444566e-01
5.44017851e-01 2.96779275e-01 2.81526685e-01 1.46983743e+00
5.63956685e-02 -3.88106741e-02 1.27257511e-01 1.08944154e+00
-7.78075308e-02 -2.41098613e-01 -8.30532670e-01 4.42363560e-01
4.54348415e-01 9.50787604e-01 -4.88051534e-01 -5.89491367e-01
-4.67536956e-01 1.01540840e+00 6.10712945e-01 6.12525702e-01
-6.75802350e-01 -3.68556678e-01 7.72256196e-01 3.69958133e-01
9.74993765e-01 -3.27793479e-01 -4.73762900e-01 -8.17400396e-01
-2.21718654e-01 -5.46476722e-01 3.31197411e-01 -9.39697921e-01
-1.01156569e+00 5.75888515e-01 -3.82992625e-02 -1.32676065e+00
-4.04188544e-01 -9.42304850e-01 -7.16831625e-01 9.90010321e-01
-1.64628553e+00 -1.24837697e+00 -2.94711828e-01 8.09675694e-01
3.84745777e-01 -3.59183624e-02 9.10484552e-01 -5.50805889e-02
2.82836091e-02 6.42927527e-01 -1.14305601e-01 2.85511851e-01
6.60067618e-01 -1.47460365e+00 4.20657396e-01 8.02728236e-01
6.44725978e-01 8.46219003e-01 4.14102942e-01 -4.09469604e-01
-1.17164028e+00 -1.59118903e+00 5.25050879e-01 -5.81813693e-01
3.81057113e-01 -4.48598325e-01 -9.23326015e-01 1.13151240e+00
6.11643493e-01 6.84396684e-01 3.79035711e-01 3.49700987e-01
-7.99752891e-01 -1.60397321e-01 -8.66227150e-01 6.03106081e-01
1.31797349e+00 -8.72811437e-01 -7.79929161e-01 1.11485392e-01
8.36239398e-01 -2.43747681e-01 -6.23920500e-01 3.04777652e-01
3.03958237e-01 -1.09071159e+00 1.17378807e+00 -7.21403658e-01
5.24790108e-01 -5.21646082e-01 -4.70744401e-01 -1.43307424e+00
-6.17188692e-01 6.06108941e-02 1.12513512e-01 9.04182196e-01
5.93172252e-01 -4.77157712e-01 8.98095489e-01 2.30820090e-01
-4.23184544e-01 -6.94387376e-01 -7.88392663e-01 -8.12315106e-01
3.40324283e-01 -5.70572734e-01 5.99067807e-01 7.75387168e-01
-2.63575077e-01 2.95258671e-01 -1.87075213e-01 3.64214599e-01
4.78913218e-01 3.09885830e-01 5.98254085e-01 -7.78432429e-01
-5.04708111e-01 -4.99065548e-01 -8.15810502e-01 -1.30266070e+00
5.31794354e-02 -1.02811754e+00 1.95118845e-01 -1.62434423e+00
-3.09990346e-02 -5.11581898e-01 -7.34538972e-01 9.37630534e-01
-5.86721227e-02 5.89864194e-01 1.90101519e-01 1.27779886e-01
-7.38429070e-01 3.86547059e-01 1.17827082e+00 -3.56327713e-01
-7.38490596e-02 -4.43649709e-01 -9.07709002e-01 3.28683883e-01
6.38981879e-01 -3.07876766e-01 -6.48703396e-01 -6.98751032e-01
-3.34634632e-02 -2.64954805e-01 9.14521337e-01 -1.19963920e+00
2.78474659e-01 3.21860448e-03 9.29297090e-01 -6.03258431e-01
3.36131126e-01 -8.75563979e-01 -5.61021045e-02 1.51564464e-01
-6.38974667e-01 -1.29021272e-01 3.00445616e-01 6.50882006e-01
-4.86969352e-01 -1.10891618e-01 6.58102810e-01 -2.23890513e-01
-1.18399048e+00 3.07650119e-01 -2.71380782e-01 -1.30071744e-01
7.56763816e-01 -4.49059218e-01 -4.63127255e-01 -2.36381486e-01
-9.88758564e-01 -2.45906394e-02 5.55617332e-01 4.13835108e-01
7.52163231e-01 -1.53040576e+00 -4.21990246e-01 2.85772115e-01
5.13229966e-01 2.68032908e-01 1.22250803e-01 4.77969229e-01
-2.36557394e-01 5.07028759e-01 -3.20756614e-01 -1.08822143e+00
-8.61910045e-01 8.68662894e-01 2.06473604e-01 1.58820197e-01
-7.14307606e-01 8.77388120e-01 6.92906618e-01 -2.21414179e-01
3.67130816e-01 -7.13670254e-01 -8.04459304e-02 -3.19464535e-01
5.79814970e-01 -3.05024117e-01 5.77107221e-02 -5.86058319e-01
-3.69112462e-01 5.93815267e-01 -5.25155663e-02 -8.49388987e-02
1.33563864e+00 -4.48683323e-03 2.11144120e-01 2.92795509e-01
1.44420302e+00 -2.62397170e-01 -1.76505411e+00 -3.69826287e-01
-7.22735282e-03 -4.44904327e-01 3.83275114e-02 -9.09983337e-01
-1.07096791e+00 1.12847829e+00 5.69147229e-01 -6.61913157e-02
1.09320676e+00 3.23725373e-01 8.28948691e-02 3.92036378e-01
4.76624876e-01 -8.89025033e-01 3.90072107e-01 6.29018426e-01
1.16459298e+00 -1.21843874e+00 -1.64967075e-01 -1.93644404e-01
-5.76732934e-01 9.20483589e-01 8.66081595e-01 -5.13237417e-01
8.28288734e-01 5.70943095e-02 6.82481681e-04 -2.31611356e-01
-1.05712473e+00 -4.31143135e-01 4.34067458e-01 9.07135546e-01
5.20083487e-01 1.77756995e-02 4.84602183e-01 2.09367052e-01
-6.00880794e-02 -2.79081494e-01 1.61503762e-01 1.04125977e+00
-5.62505484e-01 -1.00956857e+00 -2.54667461e-01 2.80414701e-01
-4.76295315e-02 -3.35284144e-01 -2.56799489e-01 9.56256568e-01
4.51049328e-01 7.05655873e-01 5.23255169e-01 -3.86466801e-01
1.80132359e-01 1.09170012e-01 5.39488733e-01 -9.57921088e-01
-4.78514642e-01 -2.36977473e-01 -8.20237100e-02 -6.40213311e-01
-3.54432642e-01 -3.61417681e-01 -1.22868609e+00 1.25770316e-01
-9.33856964e-02 -1.40874730e-02 6.64808571e-01 8.50419343e-01
6.49265349e-01 5.06681919e-01 5.43765426e-01 -1.09905064e+00
-1.08273946e-01 -9.76301849e-01 -5.23462594e-01 5.56572020e-01
6.23135030e-01 -6.61782146e-01 -2.22346410e-01 2.41890892e-01] | [9.875931739807129, 1.5740197896957397] |
c56ac986-e7f7-4747-85e1-761f4338ff56 | bioimageloader-easy-handling-of-bioimage | 2303.02158 | null | https://arxiv.org/abs/2303.02158v1 | https://arxiv.org/pdf/2303.02158v1.pdf | BioImageLoader: Easy Handling of Bioimage Datasets for Machine Learning | BioImageLoader (BIL) is a python library that handles bioimage datasets for machine learning applications, easing simple workflows and enabling complex ones. BIL attempts to wrap the numerous and varied bioimages datasets in unified interfaces, to easily concatenate, perform image augmentation, and batch-load them. By acting at a per experimental dataset level, it enables both a high level of customization and a comparison across experiments. Here we present the library and show some application it enables, including retraining published deep learning architectures and evaluating their versatility in a leave-one-dataset-out fashion. | ['Anatole Chessel', 'Emmanuel Beaurepaire', 'Xingjian Zhang', 'Seongbin Lim'] | 2023-03-02 | null | null | null | null | ['image-augmentation'] | ['computer-vision'] | [ 3.28518420e-01 -2.48300254e-01 8.47717673e-02 -5.96867681e-01
-5.48107862e-01 -7.41044819e-01 7.18045473e-01 3.36108774e-01
-8.48423421e-01 6.95841730e-01 -8.14768448e-02 -5.71450651e-01
1.72355995e-01 -3.76974016e-01 -7.25599706e-01 -8.60971391e-01
-2.73874223e-01 6.70510769e-01 2.84142733e-01 2.88206991e-02
-3.05374842e-02 9.98437583e-01 -1.55805027e+00 6.40616775e-01
1.06276900e-01 7.15007722e-01 1.41796067e-01 9.46917415e-01
4.21863757e-02 4.59095597e-01 -6.99329674e-01 -4.29557323e-01
2.47027412e-01 -1.77069623e-02 -8.22176099e-01 -1.67843148e-01
3.62798631e-01 -1.52741387e-01 -1.98122159e-01 4.95203555e-01
9.79459345e-01 -4.29399908e-01 1.26780927e-01 -1.20636380e+00
-2.63027132e-01 2.98591673e-01 -1.88465253e-01 6.88774526e-01
1.15607142e-01 6.20017111e-01 3.16118836e-01 -5.47891080e-01
9.23005998e-01 1.03317571e+00 8.13055575e-01 5.44275224e-01
-1.51817024e+00 -5.73147416e-01 -1.64222673e-01 1.39322437e-04
-9.49474871e-01 -7.03060091e-01 1.40089869e-01 -5.50253391e-01
1.33657086e+00 4.57654148e-01 5.65789878e-01 1.30402708e+00
2.86747187e-01 6.10364020e-01 1.32159257e+00 -1.42009526e-01
7.96057284e-02 -2.65391380e-01 1.99127600e-01 4.40371275e-01
1.31757990e-01 -1.84140429e-01 -3.05272996e-01 -2.65622199e-01
6.22044742e-01 1.29749984e-01 -1.93041742e-01 -5.20153046e-02
-1.47180212e+00 3.26224595e-01 3.16962808e-01 4.36919481e-01
-3.17333072e-01 -5.23982123e-02 1.00521445e+00 2.77852952e-01
2.05421671e-01 4.71741289e-01 -7.30829716e-01 8.51023942e-02
-7.93482184e-01 3.73520404e-01 6.75228536e-01 7.98257291e-01
6.69935763e-01 -2.41489753e-01 -1.42484635e-01 5.41735590e-01
-2.60730505e-01 -5.89186810e-02 1.15982032e+00 -6.42623603e-01
-2.25887179e-01 4.46056932e-01 -2.14789659e-01 -4.55661148e-01
-8.88643622e-01 -3.28161389e-01 -7.90633380e-01 3.75497878e-01
3.88227373e-01 3.57119180e-03 -1.16096652e+00 1.53363860e+00
3.86506557e-01 9.55718085e-02 -2.32556984e-01 6.62315488e-01
1.11140168e+00 2.88712643e-02 5.94004095e-01 2.36432459e-02
1.69900477e+00 -7.28979647e-01 -3.51020902e-01 -1.32870778e-01
7.72642255e-01 -7.39703834e-01 1.30572236e+00 4.76866603e-01
-1.06857502e+00 -3.56060416e-01 -1.19858766e+00 -5.07061660e-01
-1.01449955e+00 6.54297620e-02 6.67744458e-01 4.94031817e-01
-1.36339676e+00 7.94242382e-01 -1.06602359e+00 -5.46941459e-01
5.78822672e-01 5.05344927e-01 -9.69123960e-01 1.74021050e-01
-7.29466736e-01 8.57533336e-01 5.93718231e-01 -2.20747590e-01
-9.01438057e-01 -9.62978721e-01 -6.01745963e-01 -1.32046729e-01
1.26931340e-01 -9.65109408e-01 1.20376992e+00 -4.83541429e-01
-1.31340551e+00 1.59920061e+00 1.21739432e-01 -7.33793020e-01
6.22910440e-01 1.33917958e-01 -2.85728008e-01 3.08867358e-02
1.20939529e-02 8.42203259e-01 3.72463524e-01 -5.74355662e-01
-2.03317583e-01 -6.43567204e-01 -1.64317131e-01 -1.67942658e-01
-7.94255584e-02 3.22186917e-01 -7.71936238e-01 -6.42758250e-01
-3.00730467e-01 -7.22646236e-01 -3.22000265e-01 -1.38467923e-01
-3.41851085e-01 1.18178464e-01 6.47800505e-01 -5.24507821e-01
6.58572853e-01 -2.07875514e+00 -1.71716258e-01 -1.85544446e-01
5.21156788e-01 5.44451177e-01 -2.49006987e-01 4.74037349e-01
-5.46997309e-01 1.58623710e-01 -1.20216198e-01 -4.68215555e-01
-2.20345989e-01 2.02477500e-01 2.11751312e-01 5.37236094e-01
2.46869385e-01 1.04629195e+00 -6.84725165e-01 -3.84506404e-01
2.30343729e-01 5.67740381e-01 -3.40056151e-01 2.22292855e-01
2.30249623e-03 7.56781697e-01 -4.49856147e-02 6.78146780e-01
7.47023106e-01 -4.34444308e-01 3.68388116e-01 -4.54550445e-01
-1.67210713e-01 1.21171296e-01 -7.32100964e-01 1.73480225e+00
-1.64425462e-01 5.88225842e-01 2.42769465e-01 -9.42441702e-01
6.69821143e-01 1.49765298e-01 5.18174767e-01 -5.22492290e-01
3.58908474e-01 -3.19678858e-02 -1.07004933e-01 -7.59403467e-01
1.38109416e-01 5.73515287e-03 2.07891718e-01 4.29040253e-01
5.84341943e-01 3.39542329e-01 3.90851378e-01 7.78981894e-02
1.41775870e+00 2.16607362e-01 5.12157023e-01 -3.95167023e-01
4.93020356e-01 -2.59244833e-02 2.54002333e-01 7.49254107e-01
-4.03629422e-01 6.62915170e-01 5.05552649e-01 -9.75925386e-01
-1.25806475e+00 -1.00700736e+00 -3.48848432e-01 1.38064396e+00
-5.77109039e-01 -2.89649963e-01 -9.50548470e-01 -5.66705942e-01
-3.20186578e-02 1.34387821e-01 -9.32515860e-01 1.54547140e-01
-3.33584070e-01 -1.18766761e+00 8.78418863e-01 2.81773806e-01
4.70982939e-01 -1.12279403e+00 -6.72813058e-01 1.47129700e-01
1.80426285e-01 -1.29582322e+00 1.93953812e-02 5.04054487e-01
-6.49261594e-01 -1.27605426e+00 -6.27888024e-01 -6.30459905e-01
4.39829051e-01 -8.56423751e-03 1.30769002e+00 2.79107124e-01
-1.00740957e+00 -1.95144322e-02 -1.53812155e-01 -7.67810702e-01
-4.48143274e-01 2.54140139e-01 -6.34292513e-02 -2.93494821e-01
4.20461625e-01 -8.15184951e-01 -8.53145540e-01 2.24419519e-01
-1.38902485e+00 -3.95698124e-04 6.78067803e-01 7.19470799e-01
6.45960033e-01 -5.74223578e-01 4.75753546e-01 -1.11063850e+00
6.54270768e-01 -3.92765552e-01 -5.78494132e-01 2.76064962e-01
-5.22944927e-01 -1.10344358e-01 5.27202308e-01 -2.06924394e-01
-4.26221251e-01 3.58964771e-01 -5.41368783e-01 -2.20612019e-01
-6.76094234e-01 3.55761230e-01 -1.91498742e-01 -1.89016446e-01
8.18291426e-01 2.61002541e-01 4.64038730e-01 -8.09965968e-01
4.44585025e-01 5.71383238e-01 9.95611131e-01 -3.00060302e-01
5.70703037e-02 5.06407261e-01 -3.91216241e-02 -7.07585096e-01
-4.17698205e-01 -4.78803724e-01 -8.37226748e-01 -2.87952833e-03
8.53161931e-01 -7.09778249e-01 -9.69636619e-01 6.94999397e-01
-6.90594435e-01 -4.92002875e-01 4.15597893e-02 1.25694573e-01
-3.81448209e-01 1.46067888e-01 -6.14268005e-01 4.94956672e-02
-5.75545430e-01 -1.50510287e+00 9.64090526e-01 4.22143452e-02
-2.26224035e-01 -8.36149275e-01 1.76760003e-01 1.63583130e-01
5.31406939e-01 5.51737428e-01 7.02063739e-01 -1.21878552e+00
-1.42806649e-01 -5.40519416e-01 -3.36092979e-01 1.03862308e-01
-3.24641876e-02 4.98199239e-02 -1.33055210e+00 -4.09703285e-01
-2.53193408e-01 -4.67396230e-01 8.10505688e-01 2.26024091e-01
1.59442389e+00 -2.67901242e-01 -5.31838715e-01 8.63421381e-01
1.19483697e+00 -1.06502004e-01 9.40864205e-01 6.87332153e-01
2.49906465e-01 3.28110188e-01 -1.51763886e-01 1.35123894e-01
1.65462121e-01 4.90830779e-01 3.89040977e-01 -5.16085207e-01
-2.25938279e-02 3.76797378e-01 -2.44171694e-01 2.12745607e-01
-6.25286549e-02 -4.06070128e-02 -8.95216405e-01 2.83271790e-01
-1.58184969e+00 -7.31395900e-01 -4.14979979e-02 2.11326146e+00
1.01593602e+00 -1.24121234e-01 5.81295073e-01 -3.28678153e-02
5.61578929e-01 -2.64382035e-01 -5.65431952e-01 -3.38586152e-01
-2.14360625e-01 4.78529692e-01 5.23284614e-01 1.39584124e-01
-1.23097742e+00 8.67478907e-01 8.25971603e+00 5.72187364e-01
-1.53411746e+00 1.86362475e-01 8.90849590e-01 -1.70517296e-01
3.90289694e-01 -2.72292703e-01 -6.26910448e-01 3.70148778e-01
1.13304913e+00 6.61474913e-02 3.24217260e-01 7.49432564e-01
1.06662937e-01 -4.16877754e-02 -1.10116696e+00 8.58028710e-01
-1.41132325e-01 -1.74446726e+00 1.60131037e-01 1.56945363e-02
-5.86630180e-02 7.68584847e-01 -1.73093766e-01 1.27953902e-01
1.90621734e-01 -1.10491669e+00 2.70354420e-01 4.79052842e-01
7.14781165e-01 -2.99961269e-01 8.46270144e-01 5.91291897e-02
-5.48430622e-01 2.42620870e-01 -5.07253110e-02 7.65410205e-03
-5.38046695e-02 6.20022118e-01 -1.13515675e+00 5.00644982e-01
1.08237743e+00 2.12432817e-01 -9.38525856e-01 1.33236349e+00
3.32138717e-01 1.38582915e-01 -2.29962915e-01 2.01726720e-01
1.93390213e-02 9.37984809e-02 3.79899591e-01 1.80832613e+00
-6.57214001e-02 -6.99205697e-02 1.90059602e-01 5.09740889e-01
-8.81224275e-02 2.90220797e-01 -4.44140583e-01 -1.55289888e-01
2.95827657e-01 1.70263338e+00 -1.18184721e+00 -5.84774554e-01
-2.14155808e-01 9.95556355e-01 3.87816101e-01 9.63755101e-02
-6.05555415e-01 -6.14372849e-01 7.53803313e-01 7.72647560e-02
1.22842565e-01 -2.78161019e-01 -2.84399062e-01 -9.33483601e-01
-2.11537659e-01 -1.02729607e+00 5.51863492e-01 -9.26199317e-01
-1.20659912e+00 7.62717724e-01 9.72808748e-02 -6.99607611e-01
4.11652550e-02 -1.03956306e+00 -5.02902746e-01 7.28052974e-01
-1.27579379e+00 -1.25478506e+00 -4.79080975e-01 5.49029291e-01
6.62779529e-03 -1.12434290e-01 1.08809876e+00 5.06073356e-01
-9.26307917e-01 6.84307873e-01 -6.33466914e-02 8.99584442e-02
8.84930730e-01 -1.19518948e+00 5.66571057e-01 4.70133603e-01
-2.12184444e-01 1.04263699e+00 7.64525950e-01 -2.52839059e-01
-1.23612821e+00 -9.47029114e-01 6.22742236e-01 -3.32758546e-01
7.81677186e-01 -4.16871846e-01 -8.80860507e-01 8.39897752e-01
3.73935878e-01 1.79861948e-01 9.75202322e-01 -1.39227957e-01
-4.31036532e-01 -7.96750858e-02 -1.39058805e+00 5.89708567e-01
8.04618061e-01 -2.73887694e-01 -2.53947556e-01 5.39261758e-01
4.08918738e-01 -6.14118993e-01 -1.11036861e+00 3.61678600e-01
7.42600262e-01 -1.24693441e+00 1.01829052e+00 -9.55692410e-01
1.39854744e-01 -3.21275324e-01 2.44031340e-01 -1.09538245e+00
-2.80299574e-01 -8.36530864e-01 3.12953025e-01 1.06439912e+00
4.96254474e-01 -8.77108037e-01 5.63055694e-01 2.90688843e-01
-3.04467052e-01 -8.44809532e-01 -8.98205996e-01 -6.41796410e-01
3.11673339e-02 -2.43273631e-01 1.05199790e+00 9.10814762e-01
1.08251527e-01 1.20269172e-01 6.90303296e-02 -7.87164345e-02
2.30101272e-01 -3.99517000e-01 1.08811224e+00 -1.01695585e+00
-2.95751363e-01 -7.63906777e-01 -8.88974667e-01 -4.52299744e-01
-3.50643158e-01 -1.07706368e+00 -1.57974660e-01 -1.21894300e+00
3.03476691e-01 -3.02174240e-01 -4.05567199e-01 1.05257475e+00
-1.31103143e-01 7.86731184e-01 -8.56718346e-02 2.95876801e-01
-5.33803344e-01 -2.36026138e-01 7.18825817e-01 3.75125185e-02
-6.81931600e-02 -3.17547470e-01 -6.96039617e-01 5.49962163e-01
8.79947364e-01 -3.42529595e-01 -1.97565872e-02 -4.48553920e-01
-1.18446022e-01 -5.84630609e-01 6.18287146e-01 -1.32286513e+00
2.89224219e-02 1.90206558e-01 8.20346415e-01 -3.71013701e-01
7.11253732e-02 -4.46434587e-01 3.28933269e-01 4.54255283e-01
-2.59084493e-01 4.11045313e-01 6.32174492e-01 1.87716126e-01
1.90033361e-01 -2.34719068e-02 9.36396003e-01 -4.56789881e-01
-4.55882341e-01 3.67219865e-01 -4.72678065e-01 -1.68808952e-01
1.13804555e+00 -1.02715492e-01 -5.72912395e-01 7.46043324e-02
-1.17528725e+00 1.10421307e-01 6.75277591e-01 1.58485264e-01
-7.75988959e-03 -8.97892416e-01 -4.97999132e-01 5.21159232e-01
4.32013094e-01 -3.15596253e-01 2.73125261e-01 1.11944759e+00
-8.77855837e-01 2.98279494e-01 -7.15873361e-01 -8.51032436e-01
-1.51719439e+00 7.13156641e-01 5.45736492e-01 -3.61724287e-01
-7.06287086e-01 8.26895416e-01 -8.07666779e-03 -5.90169549e-01
1.05466500e-01 -2.42494926e-01 -2.49615997e-01 -8.67304504e-02
9.91784871e-01 1.33282810e-01 8.04085255e-01 -3.08533102e-01
-4.69563276e-01 2.23918110e-02 -2.48431474e-01 1.26598388e-01
1.56959796e+00 1.37348836e-02 -2.79191762e-01 3.08312714e-01
1.23285949e+00 -3.53624076e-01 -1.05326819e+00 1.86475098e-01
-7.74988579e-03 -1.10008180e-01 -6.04327060e-02 -1.04455721e+00
-8.08875680e-01 7.36417413e-01 7.98827887e-01 2.93365508e-01
1.16279244e+00 3.30823846e-03 5.40077150e-01 3.29072386e-01
1.74603283e-01 -6.20113134e-01 -3.78935367e-01 3.45738590e-01
8.60489368e-01 -9.96816933e-01 1.99839726e-01 -1.17547862e-01
-3.77632350e-01 1.23466134e+00 3.24629724e-01 1.07334077e-01
6.49221718e-01 6.94173336e-01 4.37694788e-01 -3.69718969e-01
-6.69691384e-01 -2.64892131e-01 -3.79928388e-02 8.97488058e-01
8.24009240e-01 -1.29618958e-01 -3.51791769e-01 4.52958465e-01
-2.25622118e-01 5.11342406e-01 3.31587046e-01 1.20249033e+00
-8.61497000e-02 -1.27981639e+00 -1.79623932e-01 4.57545221e-01
-8.58772039e-01 -6.67504668e-02 -5.38390338e-01 9.55318987e-01
1.71013206e-01 3.41905326e-01 1.57719046e-01 -3.26601744e-01
2.31944010e-01 3.72601569e-01 5.40640771e-01 -2.93534875e-01
-8.92714560e-01 1.09371312e-01 2.19592586e-01 -6.88458860e-01
-3.22599024e-01 -4.22955066e-01 -1.06631172e+00 -3.10371220e-01
6.96115494e-02 -2.65856147e-01 1.00868654e+00 1.01940346e+00
1.02417052e+00 5.19058704e-01 1.81133345e-01 -1.32271647e+00
-1.26682773e-01 -1.17800534e+00 -3.21122974e-01 4.64223355e-01
2.49765307e-01 -3.80604178e-01 -6.12597615e-02 3.26116830e-01] | [14.776898384094238, -2.997745990753174] |
1afe5642-f563-42f6-8a15-c524af8cabf0 | ivt-an-end-to-end-instance-guided-video | 2208.03431 | null | https://arxiv.org/abs/2208.03431v1 | https://arxiv.org/pdf/2208.03431v1.pdf | IVT: An End-to-End Instance-guided Video Transformer for 3D Pose Estimation | Video 3D human pose estimation aims to localize the 3D coordinates of human joints from videos. Recent transformer-based approaches focus on capturing the spatiotemporal information from sequential 2D poses, which cannot model the contextual depth feature effectively since the visual depth features are lost in the step of 2D pose estimation. In this paper, we simplify the paradigm into an end-to-end framework, Instance-guided Video Transformer (IVT), which enables learning spatiotemporal contextual depth information from visual features effectively and predicts 3D poses directly from video frames. In particular, we firstly formulate video frames as a series of instance-guided tokens and each token is in charge of predicting the 3D pose of a human instance. These tokens contain body structure information since they are extracted by the guidance of joint offsets from the human center to the corresponding body joints. Then, these tokens are sent into IVT for learning spatiotemporal contextual depth. In addition, we propose a cross-scale instance-guided attention mechanism to handle the variational scales among multiple persons. Finally, the 3D poses of each person are decoded from instance-guided tokens by coordinate regression. Experiments on three widely-used 3D pose estimation benchmarks show that the proposed IVT achieves state-of-the-art performances. | ['Dongmei Fu', 'Jian Wang', 'Qiansheng Yang', 'Zhongwei Qiu'] | 2022-08-06 | null | null | null | null | ['3d-pose-estimation', '3d-multi-person-pose-estimation'] | ['computer-vision', 'computer-vision'] | [-2.87209481e-01 -1.36254475e-01 -2.85124123e-01 -2.00211719e-01
-8.42785835e-01 -1.33105367e-01 4.12560523e-01 -3.58436942e-01
-3.53549302e-01 2.87099034e-01 4.48547214e-01 5.02969861e-01
2.17372805e-01 -3.49817097e-01 -9.47284579e-01 -6.43185675e-01
-1.16017900e-01 6.07777476e-01 2.33967483e-01 -5.95494732e-02
6.50896281e-02 1.61197007e-01 -1.31838846e+00 1.32792622e-01
4.23454225e-01 1.16340506e+00 1.84654161e-01 5.57406962e-01
1.03995286e-01 7.27612972e-01 -5.07509589e-01 -1.69360429e-01
2.73412198e-01 -3.85028929e-01 -4.59534079e-01 5.72094440e-01
6.61936462e-01 -7.35503137e-01 -6.94439828e-01 8.68607759e-01
6.51052773e-01 2.60163188e-01 5.44238806e-01 -1.36107099e+00
-1.94515646e-01 -7.28102028e-02 -8.05323660e-01 1.75036699e-01
8.68356049e-01 2.11628228e-01 8.22506309e-01 -1.17403328e+00
8.43912840e-01 1.69058776e+00 3.72723907e-01 5.80386698e-01
-6.23756051e-01 -4.26177323e-01 6.71390295e-01 5.61291456e-01
-1.41217136e+00 -1.61276937e-01 1.13580346e+00 -5.61934114e-01
9.23088372e-01 -3.09420303e-02 1.28556812e+00 1.07926357e+00
2.41162717e-01 1.36582720e+00 4.42665219e-01 -1.54182732e-01
-1.78978235e-01 -5.41988134e-01 -1.74013361e-01 1.02614248e+00
-2.26535887e-01 3.54561093e-03 -8.49120259e-01 1.20198309e-01
1.18142605e+00 4.24576163e-01 -2.57955611e-01 -6.92895472e-01
-1.57440257e+00 5.72300017e-01 6.27728105e-01 -1.66272476e-01
-6.33095145e-01 4.40408975e-01 4.48560268e-01 -4.53657322e-02
5.55675566e-01 -3.23816627e-01 -3.75125945e-01 -9.05344039e-02
-5.33500373e-01 6.20840371e-01 2.25057378e-01 1.16243637e+00
5.56331098e-01 -1.01489395e-01 -5.12010992e-01 4.98325467e-01
5.53236425e-01 6.53975606e-01 2.33114436e-01 -1.20592034e+00
8.47114205e-01 6.99405313e-01 2.79173732e-01 -1.07395303e+00
-2.91082174e-01 -2.80894071e-01 -5.20861030e-01 -2.75887817e-01
3.87802958e-01 1.18959323e-03 -9.39770997e-01 1.55873358e+00
9.10180867e-01 3.64343852e-01 -4.37669575e-01 1.42382526e+00
7.46855378e-01 6.96675003e-01 -3.13833542e-02 -8.89062509e-02
1.57209325e+00 -1.17982614e+00 -6.35178685e-01 -2.98785180e-01
2.12697133e-01 -3.84155124e-01 7.35906303e-01 2.72996306e-01
-1.30020189e+00 -8.83695900e-01 -6.66640162e-01 -3.79225492e-01
6.51822314e-02 1.63915113e-01 2.14588016e-01 1.36457821e-02
-5.89292169e-01 8.76870826e-02 -1.18650329e+00 -1.13240667e-01
2.00362295e-01 2.80422360e-01 -5.73029935e-01 -1.15561150e-01
-1.16420281e+00 7.05456138e-01 5.11116497e-02 6.00711703e-01
-1.03778923e+00 -3.49456549e-01 -1.29063320e+00 -2.60692924e-01
7.23325491e-01 -1.06731677e+00 1.13684416e+00 -4.94876742e-01
-1.24180472e+00 8.77099633e-01 -5.54537177e-01 -6.35550767e-02
9.04798508e-01 -6.84835494e-01 1.99583396e-01 6.19437158e-01
3.98842156e-01 7.10831404e-01 1.13104057e+00 -1.03090179e+00
-7.58893311e-01 -8.42894197e-01 3.82006802e-02 6.45024359e-01
-2.68096849e-02 -2.17916802e-01 -1.28695309e+00 -6.36281788e-01
4.31841582e-01 -8.68042529e-01 -1.63949534e-01 3.30627710e-01
-5.36527812e-01 -5.89111090e-01 5.97529233e-01 -8.98303509e-01
8.09570670e-01 -2.02650785e+00 9.86194670e-01 -2.39118896e-02
3.47261697e-01 -1.56508744e-01 7.32999370e-02 1.03647955e-01
3.03174585e-01 -3.80611867e-01 2.37409607e-01 -6.69900715e-01
1.58001736e-01 1.97924808e-01 8.66536498e-02 6.33805513e-01
2.99895763e-01 1.03719759e+00 -9.49072123e-01 -8.37858915e-01
6.07250571e-01 7.39509284e-01 -7.43846238e-01 5.01442373e-01
-2.06233650e-01 6.14754140e-01 -8.64959717e-01 7.78712451e-01
3.37812513e-01 -1.29398823e-01 -6.52605444e-02 -5.45967042e-01
3.96422036e-02 -1.07687870e-02 -1.00706887e+00 2.21408772e+00
-6.17569836e-04 6.96063191e-02 -7.56340995e-02 -9.81145203e-01
5.14767528e-01 3.98481041e-01 7.39236355e-01 -5.98656654e-01
2.03461662e-01 -1.81260020e-01 -6.43429697e-01 -7.18082726e-01
2.15489835e-01 1.33732498e-01 -3.75412107e-01 -1.41941637e-01
1.38981700e-01 1.79296702e-01 1.82212479e-02 9.18565094e-02
6.96076095e-01 6.66019738e-01 1.82685271e-01 3.93321693e-01
5.70107400e-01 -1.52221829e-01 8.53591204e-01 2.89322466e-01
-5.69080889e-01 6.48227274e-01 4.65204686e-01 -5.78693807e-01
-1.07020485e+00 -1.38016236e+00 5.38581192e-01 7.72399127e-01
5.18359661e-01 -5.55199921e-01 -6.99474037e-01 -6.91618800e-01
1.37664331e-02 -1.35935441e-01 -7.79720902e-01 -1.13154478e-01
-9.33673680e-01 2.47261152e-02 1.12027720e-01 7.00131416e-01
5.97044110e-01 -9.07882869e-01 -8.76609147e-01 2.37200275e-01
-7.53187299e-01 -1.31811953e+00 -9.67100859e-01 -3.58029991e-01
-7.68029153e-01 -1.17173457e+00 -1.20480001e+00 -8.74312282e-01
5.88066638e-01 3.35980564e-01 7.73613155e-01 -1.91185344e-02
-3.59310448e-01 5.32148480e-01 -3.03377658e-01 4.66818139e-02
4.35532331e-01 -2.22034723e-01 2.22281322e-01 1.64539129e-01
3.70262325e-01 -2.62634754e-01 -8.91641200e-01 3.76969516e-01
-2.39002645e-01 1.09214649e-01 5.17699957e-01 8.56476307e-01
8.62569869e-01 -2.04907835e-01 6.90890327e-02 -2.16635972e-01
-8.87188315e-02 -1.63086936e-01 -1.89978406e-01 2.52783507e-01
4.00518954e-01 -1.75428420e-01 2.18367070e-01 -5.01331389e-01
-8.50684166e-01 3.80387038e-01 -1.03756234e-01 -9.81541276e-01
-6.17679097e-02 2.15126649e-01 -3.76657069e-01 3.40047807e-01
-2.31619217e-02 4.22524154e-01 5.67017645e-02 -4.22920644e-01
1.48918286e-01 2.35543415e-01 7.09005296e-01 -6.28241479e-01
8.40258121e-01 3.71916085e-01 -4.33936045e-02 -6.98173702e-01
-9.98278916e-01 -5.65793872e-01 -8.54226470e-01 -7.32545972e-01
1.25222611e+00 -1.36001575e+00 -8.81720543e-01 6.76570952e-01
-1.15753162e+00 -1.21776588e-01 8.18694606e-02 6.70776963e-01
-7.75058925e-01 5.88739693e-01 -7.50372589e-01 -6.38105989e-01
-1.98277533e-01 -1.31994534e+00 1.89030230e+00 1.00127026e-01
-7.47574791e-02 -7.29161322e-01 -2.45565340e-01 5.87896049e-01
-4.53337252e-01 5.31800747e-01 5.06426930e-01 1.09165393e-01
-6.76883817e-01 -3.64509046e-01 7.89969042e-02 5.33697158e-02
-9.66554061e-02 -3.12568963e-01 -4.32441592e-01 -3.78101379e-01
-1.06342018e-01 -3.16811204e-01 6.32215917e-01 7.06540704e-01
8.10887992e-01 -2.16647729e-01 -3.49075645e-01 6.38042748e-01
9.52136517e-01 -1.47679925e-01 3.93055588e-01 2.15853631e-01
1.17287409e+00 6.67799473e-01 1.11305654e+00 5.53731382e-01
7.12307453e-01 1.10549498e+00 5.33381045e-01 1.54027745e-01
3.59603316e-02 -6.77465975e-01 5.68022311e-01 6.55870020e-01
-4.22526151e-01 1.33818924e-01 -4.35701340e-01 4.03533846e-01
-2.04781795e+00 -9.29252923e-01 1.44638032e-01 2.01383281e+00
4.85478491e-01 2.20804915e-01 5.49280524e-01 8.59586447e-02
8.07339907e-01 2.82183796e-01 -6.31278217e-01 5.34596503e-01
3.29516709e-01 -4.42257494e-01 1.84114948e-02 3.92331153e-01
-1.04522848e+00 9.84990597e-01 4.60035515e+00 6.12624645e-01
-7.39615917e-01 -1.22487787e-02 1.56465903e-01 -4.79239821e-01
6.87700957e-02 -2.53025085e-01 -9.01371479e-01 4.74997759e-01
8.34841803e-02 2.14432985e-01 -3.53602879e-02 6.90356135e-01
3.67061645e-01 8.60209614e-02 -1.19741988e+00 1.25985599e+00
1.57734603e-01 -8.76673043e-01 2.27730274e-01 1.28802940e-01
3.99929613e-01 -3.20750594e-01 4.02314328e-02 2.69622087e-01
-3.88353944e-01 -5.11579216e-01 1.16875970e+00 7.37710059e-01
5.65198660e-01 -9.51868713e-01 4.67333704e-01 3.93405676e-01
-1.62789035e+00 -9.23530385e-02 -4.16880578e-01 -1.38707310e-01
5.12028337e-01 2.16062188e-01 -4.02053356e-01 6.02932572e-01
9.91765022e-01 1.13594341e+00 -3.45651388e-01 9.42692339e-01
-4.06435102e-01 -5.66766970e-03 -2.49239415e-01 1.02504775e-01
2.15243563e-01 1.31740481e-01 6.64174378e-01 6.90400362e-01
3.82148921e-01 2.52046764e-01 6.93380058e-01 5.25102079e-01
2.44287238e-01 -1.94744498e-01 -4.19861734e-01 3.69091839e-01
2.98364520e-01 8.84341717e-01 -4.77184236e-01 -3.68253678e-01
-3.44886094e-01 1.37777245e+00 3.08418751e-01 4.11265522e-01
-9.66241896e-01 -1.77596305e-02 7.44788826e-01 8.45028758e-02
5.30314147e-01 -4.50596660e-01 3.82731050e-01 -1.46945810e+00
4.88586962e-01 -7.60329902e-01 4.43155795e-01 -9.82767642e-01
-1.13056040e+00 3.14155936e-01 1.81878135e-01 -1.48317301e+00
-5.67497313e-01 -5.46261370e-01 -2.52130151e-01 5.32632411e-01
-1.06529808e+00 -1.31077957e+00 -5.63524246e-01 1.02769983e+00
9.22141135e-01 2.20134929e-01 3.49616289e-01 1.40503898e-01
-4.48349714e-01 4.81282890e-01 -5.72455943e-01 5.46187758e-01
6.64516926e-01 -1.11063790e+00 5.68768382e-01 5.91650248e-01
-1.12947514e-02 4.87486064e-01 5.53654730e-01 -7.43269086e-01
-1.71004272e+00 -9.76626098e-01 9.60151732e-01 -6.29495919e-01
1.09862879e-01 -4.90534455e-01 -3.43579262e-01 7.19610512e-01
-3.94397259e-01 2.18346655e-01 1.41934544e-01 -1.61111131e-01
-2.75788665e-01 -5.46995439e-02 -8.44572604e-01 5.98151267e-01
1.43466353e+00 -5.72468698e-01 -8.12850356e-01 2.53993541e-01
7.70908833e-01 -8.21471870e-01 -8.89632702e-01 2.61582583e-01
8.35219383e-01 -7.41047919e-01 1.43453550e+00 -5.15462220e-01
4.86265600e-01 -4.91391033e-01 -1.28992096e-01 -9.49259937e-01
-1.57420516e-01 -2.95747340e-01 -4.42013472e-01 8.63539934e-01
-3.08033615e-01 2.39508506e-02 9.74733472e-01 4.63310868e-01
-6.22230470e-02 -7.95311987e-01 -1.10338140e+00 -4.92053062e-01
-3.21436316e-01 -3.47598881e-01 2.82275617e-01 4.12039042e-01
-2.28227284e-02 2.67392009e-01 -8.96076977e-01 1.93757758e-01
1.04852998e+00 1.68349296e-01 1.01160169e+00 -9.25407112e-01
-2.07608163e-01 -1.49975242e-02 -8.63503873e-01 -1.72682571e+00
1.33062273e-01 -3.65786791e-01 1.66320071e-01 -1.42363286e+00
3.51002187e-01 2.79480010e-01 -8.83771852e-02 2.99385220e-01
-4.97005790e-01 3.06827158e-01 4.73595291e-01 1.59817815e-01
-9.00043666e-01 7.89417565e-01 1.70195019e+00 -3.09393227e-01
-1.48750162e-02 4.09812406e-02 -3.12745757e-02 7.94315517e-01
1.94659427e-01 -2.92300075e-01 -3.97756130e-01 -5.61604321e-01
-1.57740220e-01 6.99092090e-01 1.04028130e+00 -9.90446210e-01
2.36944169e-01 -1.02454074e-01 9.24673140e-01 -1.16294444e+00
8.52935195e-01 -7.23014057e-01 -1.27276778e-01 5.20606875e-01
-2.35445023e-01 1.21419966e-01 -4.74775404e-01 8.65781188e-01
-2.01488078e-01 3.92943203e-01 2.85221130e-01 -4.56614047e-01
-9.80572522e-01 9.08450961e-01 -8.60589147e-02 1.61649331e-01
9.78772938e-01 -3.58676761e-01 2.57698357e-01 -3.67435783e-01
-1.06279695e+00 5.84632456e-01 3.97393733e-01 6.78000212e-01
1.06255794e+00 -1.56835973e+00 -6.33794069e-01 2.28702515e-01
1.04109220e-01 4.44642186e-01 7.27813125e-01 8.50089550e-01
-2.73498654e-01 3.85985434e-01 -2.78037012e-01 -1.21900856e+00
-1.30044198e+00 5.19081891e-01 4.14012820e-01 -8.18893015e-02
-8.78476381e-01 9.38439608e-01 4.64797437e-01 -2.10320085e-01
5.58767557e-01 -3.37153345e-01 -1.29502580e-01 5.94983920e-02
6.12873852e-01 2.62269467e-01 -4.37387437e-01 -1.03342450e+00
-5.43308854e-01 1.10953581e+00 4.02077623e-02 -2.06695706e-01
1.17761230e+00 -4.09400582e-01 1.54669538e-01 4.96967465e-01
1.50734055e+00 -3.67994040e-01 -1.95880377e+00 -3.21894258e-01
-4.83511299e-01 -7.65472591e-01 -3.90537500e-01 -2.11349845e-01
-1.27035475e+00 1.05714142e+00 4.96725738e-01 -5.55998504e-01
9.25253034e-01 2.02258810e-01 1.06555319e+00 3.01814824e-01
6.44515991e-01 -1.12737679e+00 5.60573220e-01 3.63711357e-01
7.93371737e-01 -1.05975413e+00 -1.18358359e-01 -4.35575932e-01
-6.71977580e-01 8.55397463e-01 8.81678998e-01 -3.55901450e-01
5.04523695e-01 -2.55994320e-01 -6.79929703e-02 -1.86495051e-01
-6.77799881e-01 -1.58092946e-01 6.05491579e-01 7.55301237e-01
2.14636192e-01 -1.08481713e-01 -8.39635283e-02 5.96188307e-01
1.07165225e-01 -4.70086820e-02 -1.02452733e-01 8.93803358e-01
-4.53099102e-01 -7.87501991e-01 -4.89305079e-01 -1.30525408e-02
-3.19421053e-01 3.36788803e-01 -2.34221980e-01 8.88481200e-01
1.73814923e-01 6.56608284e-01 3.31282150e-03 -5.33254147e-01
6.45820081e-01 -7.72918202e-03 8.48817289e-01 -4.09429967e-01
-2.79523820e-01 5.45355737e-01 -1.36992007e-01 -9.19970870e-01
-5.80365896e-01 -8.07718992e-01 -1.39723468e+00 -8.08779746e-02
5.71349226e-02 1.30715340e-01 1.92287117e-01 1.08828664e+00
4.18524025e-03 4.47992653e-01 4.84764040e-01 -1.59700215e+00
-4.23419237e-01 -7.37000823e-01 -4.78284597e-01 5.83883822e-01
5.74397087e-01 -1.14492118e+00 -1.22875951e-01 4.88801114e-02] | [7.162216663360596, -0.7350761890411377] |
a35f1f34-3da8-43ec-87a5-55459134d96e | mixgen-a-new-multi-modal-data-augmentation | 2206.08358 | null | https://arxiv.org/abs/2206.08358v3 | https://arxiv.org/pdf/2206.08358v3.pdf | MixGen: A New Multi-Modal Data Augmentation | Data augmentation is a necessity to enhance data efficiency in deep learning. For vision-language pre-training, data is only augmented either for images or for text in previous works. In this paper, we present MixGen: a joint data augmentation for vision-language representation learning to further improve data efficiency. It generates new image-text pairs with semantic relationships preserved by interpolating images and concatenating text. It's simple, and can be plug-and-played into existing pipelines. We evaluate MixGen on four architectures, including CLIP, ViLT, ALBEF and TCL, across five downstream vision-language tasks to show its versatility and effectiveness. For example, adding MixGen in ALBEF pre-training leads to absolute performance improvements on downstream tasks: image-text retrieval (+6.2% on COCO fine-tuned and +5.3% on Flicker30K zero-shot), visual grounding (+0.9% on RefCOCO+), visual reasoning (+$0.9% on NLVR2), visual question answering (+0.3% on VQA2.0), and visual entailment (+0.4% on SNLI-VE). | ['Mu Li', 'Bo Li', 'Wanqian Zhang', 'Aston Zhang', 'Srikar Appalaraju', 'Yi Zhu', 'Xiaoshuai Hao'] | 2022-06-16 | null | null | null | null | ['visual-reasoning', 'visual-reasoning', 'visual-entailment'] | ['computer-vision', 'reasoning', 'reasoning'] | [ 3.35816815e-02 8.65343437e-02 1.61496729e-01 -2.96516269e-01
-8.09694707e-01 -8.09569538e-01 1.09149289e+00 1.14761837e-01
-7.96823621e-01 2.49073878e-01 2.42927432e-01 -5.91665268e-01
4.92009521e-01 -5.92438161e-01 -9.81369197e-01 -5.93728982e-02
4.78160679e-01 4.65548366e-01 1.92080408e-01 -5.39332867e-01
5.40110208e-02 2.04408154e-01 -1.38927460e+00 1.09807086e+00
5.97647488e-01 1.02595317e+00 2.93813169e-01 1.01537323e+00
-3.24456364e-01 1.25479090e+00 -5.13041496e-01 -8.55819523e-01
2.65798718e-01 -2.92048544e-01 -1.05025327e+00 -1.10481396e-01
1.15731680e+00 -6.10543013e-01 -5.47826231e-01 6.98111296e-01
4.13271338e-01 -9.23141390e-02 6.96982741e-01 -1.46563065e+00
-1.36520612e+00 4.08525079e-01 -5.54454923e-01 1.34263456e-01
1.16739646e-01 7.70553052e-01 1.18923628e+00 -1.29020810e+00
7.26065278e-01 1.47421706e+00 6.04317546e-01 4.58276093e-01
-1.13550198e+00 -4.21121627e-01 -4.48945425e-02 3.83443952e-01
-1.19250703e+00 -6.68253303e-01 1.13266453e-01 -4.98754591e-01
1.55631995e+00 2.96101660e-01 5.65110207e-01 1.16833925e+00
-8.00576061e-02 1.16806209e+00 9.21880305e-01 -4.29198384e-01
-1.17548391e-01 3.80615406e-02 2.89373826e-02 1.01826692e+00
2.26795580e-02 -1.00829005e-01 -4.65154201e-01 4.56377655e-01
6.22877777e-01 -2.05896854e-01 -2.15158209e-01 1.16048090e-01
-1.23688340e+00 6.99594378e-01 8.86495769e-01 -3.44610326e-02
-2.80466944e-01 7.25641310e-01 5.55280983e-01 3.37859422e-01
2.76450306e-01 5.66159844e-01 -3.17932904e-01 1.14424266e-01
-9.43700016e-01 3.63846242e-01 3.92396182e-01 1.11548650e+00
6.10626042e-01 2.36413285e-01 -8.88547540e-01 9.04996574e-01
4.55637842e-01 9.06006098e-01 4.81887311e-01 -1.14568353e+00
6.48838699e-01 6.25506401e-01 1.40863508e-02 -6.08179986e-01
-2.38389492e-01 -3.55982274e-01 -8.08089733e-01 1.57350838e-01
4.61662829e-01 1.30613565e-01 -1.46765745e+00 1.42604148e+00
-1.90502077e-01 -6.86924607e-02 2.83162475e-01 8.77433658e-01
1.65226972e+00 9.92755651e-01 2.38612577e-01 2.19810009e-01
1.55748606e+00 -1.57450461e+00 -6.36043191e-01 -5.26761591e-01
7.71316648e-01 -1.04759288e+00 1.76858628e+00 4.05958623e-01
-1.36565053e+00 -9.46454585e-01 -8.12264085e-01 -1.02226484e+00
-4.87523079e-01 4.01856542e-01 3.00793380e-01 1.91515431e-01
-1.53838956e+00 1.37258664e-01 -3.94996881e-01 -3.26003462e-01
5.76333284e-01 4.83451188e-02 -3.43887269e-01 -5.34954429e-01
-1.05144393e+00 9.41441894e-01 3.39637637e-01 6.45483471e-03
-1.12488222e+00 -8.30481112e-01 -8.58616412e-01 -3.80709320e-02
3.04829776e-01 -1.14564359e+00 1.59340286e+00 -9.89138186e-01
-1.16197395e+00 1.29502690e+00 -5.90945445e-02 -9.21008348e-01
7.40751743e-01 -5.00432014e-01 -2.34965339e-01 3.68993342e-01
1.95852682e-01 1.52946281e+00 9.28417504e-01 -1.14979351e+00
-5.13711751e-01 -1.45343229e-01 2.89187551e-01 3.37532073e-01
-5.57382517e-02 -1.07041396e-01 -1.19024169e+00 -6.43417716e-01
-5.32322168e-01 -8.06246996e-01 -2.07694285e-02 3.42791617e-01
-3.83840263e-01 -3.35605919e-01 7.33010590e-01 -1.09321737e+00
6.95767879e-01 -1.99026740e+00 -1.50267978e-03 -3.12523454e-01
3.54697496e-01 6.78200245e-01 -7.43299067e-01 3.24206799e-01
6.63611740e-02 1.84989303e-01 -1.71049580e-01 -7.90931880e-01
-4.82175983e-02 1.99929968e-01 -4.17271733e-01 1.08734876e-01
5.84230125e-01 1.85338974e+00 -7.35650718e-01 -4.72585678e-01
5.02862573e-01 4.93803471e-01 -6.34135544e-01 1.02487646e-01
-7.11105645e-01 -1.59190997e-01 2.36364469e-01 5.49098551e-01
4.42141086e-01 -4.71813470e-01 -4.08728868e-01 -5.38887799e-01
-8.20951164e-02 -2.34960089e-03 -5.95262647e-01 1.82276797e+00
-6.71763420e-01 1.15819466e+00 -1.21705405e-01 -6.35549009e-01
5.22381067e-01 7.03028440e-02 -1.31572187e-01 -1.38206816e+00
9.74964723e-02 -2.34445661e-01 -2.29520708e-01 -7.87440836e-01
9.48879659e-01 2.95974463e-01 1.28728837e-01 3.08039010e-01
3.64560038e-01 -4.52359647e-01 3.98824334e-01 6.31923199e-01
7.46567249e-01 4.11365107e-02 2.17476860e-02 5.69738969e-02
3.38076025e-01 4.31628555e-01 -2.98317879e-01 9.21793520e-01
-3.70929241e-02 7.65335858e-01 4.82016832e-01 -1.67594522e-01
-1.26893532e+00 -1.11502779e+00 2.83150107e-01 1.35701859e+00
-7.41117354e-03 -7.56240964e-01 -5.99004030e-01 -6.30299032e-01
2.67266203e-02 9.90515649e-01 -5.57645679e-01 -3.15674126e-01
-2.82473594e-01 -3.33877444e-01 9.93947744e-01 7.83380330e-01
8.53655696e-01 -1.28255534e+00 -2.91004092e-01 -2.70330548e-01
-3.42192948e-01 -1.38334632e+00 -4.42472667e-01 -2.08842129e-01
-6.03470385e-01 -8.92792344e-01 -7.95654297e-01 -7.74638057e-01
4.29070026e-01 6.24018133e-01 1.51251793e+00 1.46978393e-01
-4.23644662e-01 6.20618582e-01 -3.82376939e-01 -4.41907376e-01
-4.77694362e-01 -2.55965590e-01 -4.60165948e-01 -1.40551955e-01
5.93069792e-02 2.11460680e-01 -8.08546484e-01 -1.12371728e-01
-9.94970024e-01 4.74791735e-01 7.83634841e-01 9.21331704e-01
6.15363061e-01 -5.98375678e-01 1.73847556e-01 -6.55868411e-01
6.44410789e-01 -1.72052741e-01 -6.12120271e-01 4.21709150e-01
-5.25212824e-01 9.50190723e-02 5.20940125e-01 -2.45830342e-01
-9.47830975e-01 1.45306420e-02 -1.64755955e-01 -8.17796648e-01
-4.31379229e-02 4.57826197e-01 1.38467595e-01 2.45431840e-01
1.00913334e+00 2.10192785e-01 8.44915360e-02 -2.84987807e-01
1.34696245e+00 6.31344020e-01 1.05184984e+00 -1.88103288e-01
6.19826794e-01 5.61935127e-01 -3.06559563e-01 -9.10833299e-01
-9.01897430e-01 -4.06900823e-01 -2.52496332e-01 -2.27879155e-02
1.15319538e+00 -1.29962945e+00 -7.54723310e-01 4.22733605e-01
-1.30615330e+00 -7.43158340e-01 -3.03157777e-01 9.65499580e-02
-3.50168735e-01 4.27294850e-01 -6.10848427e-01 -3.97817373e-01
-9.13537681e-01 -1.12321055e+00 1.32531834e+00 1.30428290e-02
-1.58106089e-01 -6.54703856e-01 -2.61583805e-01 8.62994254e-01
3.70054007e-01 -2.04650521e-01 6.81497931e-01 -3.53319407e-01
-5.77221513e-01 1.20853111e-01 -1.00861204e+00 6.81675136e-01
-4.14008796e-01 2.63318032e-01 -1.06062257e+00 -3.07848006e-01
-5.77722132e-01 -7.96495497e-01 1.34903550e+00 3.14686298e-01
1.29845130e+00 -3.69628727e-01 4.95812781e-02 6.24021769e-01
1.28661752e+00 -1.38818339e-01 7.89383352e-01 2.22409308e-01
1.11777318e+00 5.58944881e-01 4.38662052e-01 8.32573697e-02
6.03364468e-01 6.11410260e-01 7.61745632e-01 -3.92456621e-01
-8.46300662e-01 -2.07598448e-01 4.37020510e-01 3.69583905e-01
9.50388536e-02 -4.04624432e-01 -1.21196091e+00 7.48031318e-01
-1.81616044e+00 -8.14559221e-01 -6.87173963e-01 1.68960440e+00
8.56163085e-01 -6.34223074e-02 2.66846400e-02 -2.92905182e-01
3.29477966e-01 -3.80175672e-02 -4.64672178e-01 -3.46748680e-01
-3.08024347e-01 4.36030984e-01 3.05431575e-01 6.53530240e-01
-1.01447070e+00 1.55420732e+00 5.54330635e+00 6.74085140e-01
-1.09632599e+00 1.03195891e-01 6.91637814e-01 -4.44778830e-01
-2.12511748e-01 -2.00882107e-01 -5.62069058e-01 1.87725171e-01
7.29346573e-01 1.46248296e-01 3.50613385e-01 7.75738955e-01
9.55907777e-02 -4.60885763e-02 -1.12051523e+00 1.39250946e+00
2.48502538e-01 -1.87629890e+00 5.61709762e-01 -3.82830173e-01
6.73074782e-01 5.56203485e-01 2.88019955e-01 6.34357810e-01
4.40991789e-01 -1.41000497e+00 9.70559955e-01 4.31745142e-01
1.11497462e+00 -5.93148589e-01 7.18851686e-01 2.36807503e-02
-8.66361737e-01 1.91509932e-01 -2.87239581e-01 6.18647151e-02
8.13495442e-02 3.39564323e-01 -1.17007077e+00 4.66207564e-01
9.85751152e-01 7.35144913e-01 -1.14090407e+00 7.41700768e-01
-4.04912829e-01 5.24636388e-01 -2.27011040e-01 7.18524978e-02
5.44262648e-01 2.14308575e-01 7.48716593e-02 1.55741739e+00
-5.14400378e-02 -2.29711294e-01 -1.04353972e-01 8.58780324e-01
-5.20957649e-01 2.92058364e-02 -5.56938231e-01 -1.42370388e-01
2.54122227e-01 1.26293778e+00 -2.75736809e-01 -7.20987141e-01
-5.53468883e-01 1.49234772e+00 3.09488088e-01 6.12399340e-01
-8.38893116e-01 -3.46200407e-01 4.24594134e-01 -7.04219788e-02
2.58019328e-01 -1.15247659e-01 -2.34234154e-01 -1.09918773e+00
-8.10940564e-02 -9.82572734e-01 4.38589841e-01 -1.53542507e+00
-1.19488907e+00 5.57378948e-01 -1.08575225e-01 -7.63155818e-01
-1.99374750e-01 -8.89093578e-01 -1.89341798e-01 8.99517298e-01
-1.48518705e+00 -1.61410344e+00 -7.28757799e-01 9.10868108e-01
8.39403987e-01 -2.09385872e-01 5.45049012e-01 2.42364272e-01
-2.99849659e-01 6.80706143e-01 -1.31959036e-01 2.98353225e-01
7.86015689e-01 -1.23571074e+00 9.69001532e-01 8.62779438e-01
6.44575179e-01 3.29611540e-01 3.34394753e-01 -4.29049045e-01
-1.42022574e+00 -1.48538172e+00 8.25911820e-01 -7.77417541e-01
6.94781959e-01 -3.30349743e-01 -9.37746406e-01 8.13688576e-01
7.80985475e-01 2.23033547e-01 2.59084225e-01 -2.23844275e-01
-8.76190364e-01 -2.25550625e-02 -8.44129145e-01 1.11283302e+00
1.00974953e+00 -8.47125471e-01 -8.12849402e-01 5.18797636e-01
1.10986888e+00 -4.29760367e-01 -5.53976715e-01 2.39699453e-01
3.60847861e-01 -8.54318261e-01 1.28513813e+00 -7.76806533e-01
8.13259065e-01 -4.11609322e-01 -2.03869358e-01 -9.96373832e-01
-3.18964124e-01 -3.20702165e-01 -8.35603401e-02 1.11812210e+00
3.26161593e-01 -1.24426439e-01 4.78813022e-01 4.67423126e-02
-2.78124303e-01 -3.88193965e-01 -6.14475906e-01 -3.97913307e-01
6.57176450e-02 -7.99104095e-01 7.14530051e-02 9.84347105e-01
-5.29153705e-01 8.55050385e-01 -2.33569399e-01 -1.50584895e-02
2.62871772e-01 -2.03331962e-01 1.04095340e+00 -7.20715582e-01
-3.73232216e-01 -6.88854337e-01 -2.42872015e-02 -1.20614707e+00
-2.66888104e-02 -1.16845846e+00 -1.86769128e-01 -2.14833426e+00
9.09446850e-02 8.39787349e-02 -3.21103819e-02 1.01716006e+00
-1.97052211e-01 6.30432546e-01 6.39190316e-01 2.31601879e-01
-6.74912274e-01 5.09689867e-01 1.32911742e+00 -4.57693905e-01
-3.09209675e-02 -5.92274666e-01 -6.95145488e-01 5.69354892e-01
7.81120420e-01 7.40572363e-02 -6.10593915e-01 -1.00840616e+00
4.52726334e-01 -2.64044911e-01 9.60504591e-01 -7.95324743e-01
3.23812552e-02 1.91536039e-01 4.73933369e-01 -8.77497852e-01
4.55243260e-01 -5.06170213e-01 -2.85467267e-01 4.66662526e-01
-4.77956355e-01 1.96039468e-01 6.73942089e-01 3.64825398e-01
-1.50569797e-01 -6.55708611e-02 6.84168637e-01 -3.19076419e-01
-1.21488929e+00 6.98844269e-02 -1.78187758e-01 3.68018866e-01
6.40165448e-01 5.58490679e-02 -9.15926456e-01 -4.92648542e-01
-4.24339265e-01 5.98742485e-01 2.58647323e-01 6.96780503e-01
9.57804799e-01 -1.17904484e+00 -9.31401551e-01 -6.34296089e-02
5.04843414e-01 2.42715001e-01 3.10304701e-01 6.94849610e-01
-8.90591979e-01 6.49891734e-01 -2.67398030e-01 -8.70992839e-01
-1.44455040e+00 5.56399345e-01 3.90192986e-01 -7.97475576e-02
-4.79429603e-01 1.24663115e+00 4.13743824e-01 -2.36331433e-01
2.24317893e-01 -7.56088674e-01 -2.24589109e-02 1.23862974e-01
5.26295841e-01 1.52916282e-01 7.62321204e-02 -4.36216086e-01
-2.45879859e-01 4.12618518e-01 -3.35293442e-01 -3.86605978e-01
9.42523718e-01 -1.72489733e-02 -2.21768901e-01 1.88689381e-01
1.31067717e+00 -3.17737758e-01 -1.22982705e+00 -3.10134232e-01
-2.56984919e-01 -5.14540859e-02 1.76167548e-01 -1.21100819e+00
-1.05851364e+00 1.18988526e+00 5.75499296e-01 -8.75860006e-02
1.07731044e+00 2.52002001e-01 5.80493093e-01 7.67015457e-01
-3.68893236e-01 -9.22869682e-01 7.05989897e-01 7.06693232e-01
1.44573176e+00 -1.49706042e+00 -1.33971214e-01 -1.07036650e-01
-1.00025213e+00 7.83931077e-01 7.73828924e-01 4.30227742e-02
2.70528849e-02 -1.63924694e-01 1.33012936e-01 -2.32974306e-01
-8.59486341e-01 -6.30047083e-01 7.76103377e-01 5.79314709e-01
4.05834407e-01 -1.76242217e-01 1.39062971e-01 1.60925940e-01
-2.71756053e-01 -1.53127313e-01 4.09084529e-01 5.87747097e-01
-2.41430551e-01 -6.97973728e-01 -2.67347038e-01 3.66411418e-01
-2.52787173e-01 -6.91112757e-01 -6.85271084e-01 1.01198947e+00
7.31701180e-02 9.10016119e-01 3.41617644e-01 -8.24780092e-02
4.75428641e-01 3.50273177e-02 5.71715474e-01 -6.03687108e-01
-6.68233097e-01 3.84602919e-02 4.41280514e-01 -6.95146680e-01
-1.90815225e-01 -1.16515912e-01 -1.45183635e+00 -1.31789342e-01
9.43128690e-02 -4.00164038e-01 6.82344496e-01 7.42885947e-01
6.00028217e-01 7.71646619e-01 -2.25904942e-01 -5.49056709e-01
-1.82516083e-01 -8.72961462e-01 2.55960047e-01 5.10228872e-01
3.41345310e-01 -1.13192216e-01 -1.83955073e-01 4.44304019e-01] | [10.908561706542969, 1.562545895576477] |
ac076472-a1eb-40d9-9f20-4d76a4fc63a3 | a-multi-task-mean-teacher-for-semi-supervised | null | null | http://openaccess.thecvf.com/content_CVPR_2020/html/Chen_A_Multi-Task_Mean_Teacher_for_Semi-Supervised_Shadow_Detection_CVPR_2020_paper.html | http://openaccess.thecvf.com/content_CVPR_2020/papers/Chen_A_Multi-Task_Mean_Teacher_for_Semi-Supervised_Shadow_Detection_CVPR_2020_paper.pdf | A Multi-Task Mean Teacher for Semi-Supervised Shadow Detection | Existing shadow detection methods suffer from an intrinsic limitation in relying on limited labeled datasets, and they may produce poor results in some complicated situations. To boost the shadow detection performance, this paper presents a multi-task mean teacher model for semi-supervised shadow detection by leveraging unlabeled data and exploring the learning of multiple information of shadows simultaneously. To be specific, we first build a multi-task baseline model to simultaneously detect shadow regions, shadow edges, and shadow count by leveraging their complementary information and assign this baseline model to the student and teacher network. After that, we encourage the predictions of the three tasks from the student and teacher networks to be consistent for computing a consistency loss on unlabeled data, which is then added to the supervised loss on the labeled data from the predictions of the multi-task baseline model. Experimental results on three widely-used benchmark datasets show that our method consistently outperforms all the compared state-of- the-art methods, which verifies that the proposed network can effectively leverage additional unlabeled data to boost the shadow detection performance.
| [' Pheng-Ann Heng', ' Wei Feng', ' Song Wang', ' Liang Wan', ' Lei Zhu', 'Zhihao Chen'] | 2020-06-01 | null | null | null | cvpr-2020-6 | ['shadow-detection'] | ['computer-vision'] | [ 3.12726557e-01 3.72021556e-01 -2.77557343e-01 -6.89797103e-01
-8.30524981e-01 -1.40779942e-01 3.42053920e-01 -5.10114670e-01
-2.03818291e-01 9.37516689e-01 -1.18998280e-02 -5.13537765e-01
5.27601719e-01 -2.41703764e-01 -6.63129151e-01 -1.18135417e+00
3.71693760e-01 2.08847523e-01 1.09065831e+00 4.77581978e-01
-2.56258920e-02 4.33079004e-02 -1.26709557e+00 -6.55477792e-02
1.22634041e+00 1.25264215e+00 4.60911691e-01 4.35782075e-01
4.97301072e-02 1.08292878e+00 -6.10368133e-01 5.41791432e-02
2.25277115e-02 -2.74863303e-01 -1.02222785e-01 9.23782140e-02
6.10903502e-01 -7.38615513e-01 -2.79343486e-01 6.12969518e-01
6.63952827e-01 1.50955558e-01 6.94689631e-01 -1.58167207e+00
-3.53709221e-01 1.45483151e-01 -9.61768389e-01 1.07514784e-01
-1.47278652e-01 -4.16406803e-02 6.93513751e-01 -1.07647789e+00
7.55838379e-02 1.17110097e+00 7.40444362e-01 9.47660282e-02
-9.46742177e-01 -9.78741884e-01 5.48574984e-01 1.91783130e-01
-1.21824574e+00 -2.91935623e-01 8.18737090e-01 -8.02267343e-02
2.90833265e-01 -4.16052490e-02 3.41600031e-01 1.19548190e+00
2.24134505e-01 1.26759398e+00 1.66432059e+00 -3.47426414e-01
5.38997203e-02 4.01260972e-01 2.12445572e-01 1.09963799e+00
9.38525796e-02 7.98048675e-02 -7.89671123e-01 -1.36743441e-01
3.30450475e-01 9.06195119e-02 -3.42238098e-01 -5.51499128e-01
-8.26758087e-01 6.60548627e-01 4.03401911e-01 -2.32188240e-01
-1.25617785e-02 1.18356049e-01 4.37231958e-02 -2.58781433e-01
8.44721198e-01 -5.08526802e-01 -4.64778781e-01 5.38139284e-01
-9.27902460e-01 -1.77168116e-01 7.19077408e-01 9.85085845e-01
9.69585478e-01 6.87082335e-02 -5.60323298e-01 5.32224357e-01
6.00757658e-01 1.25963891e+00 1.81439176e-01 -6.56028628e-01
4.24783677e-01 2.90772855e-01 1.10115066e-01 -5.41265011e-01
-2.95946032e-01 -3.56970519e-01 -4.79642451e-01 1.94675267e-01
3.29291999e-01 -3.40301514e-01 -1.29021621e+00 1.59374249e+00
6.79871917e-01 7.31037378e-01 -4.88683470e-02 8.70716572e-01
9.83066499e-01 5.81770778e-01 5.02898581e-02 -2.83839226e-01
1.07277787e+00 -1.54539132e+00 -6.72395110e-01 -4.41041678e-01
3.57173771e-01 -8.03949535e-01 1.09296858e+00 1.68098271e-01
-3.80150020e-01 -5.51100671e-01 -1.27050900e+00 -2.63443124e-02
-2.27872327e-01 4.45179254e-01 5.88565111e-01 7.50337005e-01
-6.72795475e-01 1.59825251e-01 -7.57748127e-01 -1.05508350e-01
5.74688196e-01 4.70270142e-02 4.57275897e-01 -8.72050449e-02
-8.97107005e-01 7.85657108e-01 1.24590829e-01 4.11969498e-02
-1.40187287e+00 -7.78572142e-01 -5.66197991e-01 1.37727354e-02
9.31807160e-01 -1.21808425e-01 1.37704957e+00 -6.71450853e-01
-1.39264870e+00 4.03322905e-01 -3.28475237e-01 -2.02263236e-01
7.16238260e-01 -4.81038988e-01 -8.59136134e-03 -8.56265724e-02
1.38212129e-01 6.45099223e-01 1.00835145e+00 -1.75703776e+00
-9.54048514e-01 -2.18990877e-01 -1.50646031e-01 4.48484391e-01
-4.89349872e-01 -3.35601360e-01 -5.77094078e-01 -4.38079685e-01
-5.16658276e-02 -1.12086332e+00 4.46220068e-03 2.31390446e-01
-7.38965929e-01 -2.97736466e-01 1.62471986e+00 -4.91934717e-01
9.56073225e-01 -2.08558798e+00 -5.52610934e-01 -3.81459780e-02
2.82866120e-01 1.73273906e-01 3.16990204e-02 -7.36582130e-02
5.35124540e-01 -4.94373798e-01 -2.93014854e-01 -7.81676829e-01
-1.77730218e-01 4.08004463e-01 -6.89492226e-01 6.16458297e-01
-2.24155784e-01 9.03664410e-01 -9.79540646e-01 -1.16898799e+00
2.28357926e-01 5.60796618e-01 3.67652029e-01 6.85784459e-01
-2.96419054e-01 4.25335974e-01 -7.01225042e-01 8.28178465e-01
8.02747905e-01 -3.97430807e-01 1.61935702e-01 -1.42634898e-01
1.22263551e-01 9.56516638e-02 -8.87081921e-01 1.14628196e+00
-5.97356975e-01 9.25908923e-01 7.43113682e-02 -6.73798323e-01
7.09224641e-01 2.17938766e-01 4.03789401e-01 -6.03646874e-01
5.71486503e-02 1.42013514e-02 -3.79530549e-01 -2.26381242e-01
1.08161174e-01 1.71529669e-02 7.08918050e-02 8.81346345e-01
-1.93171769e-01 -1.62451565e-01 -5.98475635e-01 3.43438148e-01
8.35460842e-01 4.37952012e-01 -5.20144515e-02 -1.27513796e-01
2.91746527e-01 -3.96283507e-01 7.13580787e-01 9.04920936e-01
-5.24905443e-01 4.41708893e-01 1.24584287e-01 -1.29095867e-01
-5.00251710e-01 -1.34234357e+00 -7.15000182e-02 1.68310285e+00
5.04384458e-01 2.17642054e-01 -7.02140152e-01 -1.40093219e+00
1.47295043e-01 7.27349401e-01 -5.99264860e-01 9.65805203e-02
-3.85165900e-01 -9.14637327e-01 5.61084509e-01 7.30062604e-01
6.41784549e-01 -8.90358925e-01 -9.32676017e-01 -1.85703889e-01
-3.11937302e-01 -1.52251542e+00 -5.39725959e-01 7.21127331e-01
-5.08977175e-01 -1.07764518e+00 -7.17355609e-01 -6.08706892e-01
8.00258040e-01 1.10430443e+00 8.72238874e-01 2.52794981e-01
-2.61865318e-01 2.73904055e-01 -1.62446707e-01 -1.05487061e+00
-4.23767269e-02 -2.87472247e-03 5.02129272e-02 -3.43963644e-03
-5.03228046e-03 -3.76353115e-01 -6.53316796e-01 5.01873791e-01
-6.38423026e-01 2.48653233e-01 7.68780351e-01 6.55068517e-01
3.79900187e-01 -1.83856085e-01 5.52645266e-01 -1.21624589e+00
5.11642871e-03 -3.06381017e-01 -7.44598389e-01 6.21901810e-01
-1.26824749e+00 7.45840967e-02 5.72255552e-01 -4.43803906e-01
-1.82464135e+00 4.33097005e-01 5.02978981e-01 -5.14502764e-01
-3.10757291e-02 -2.30236083e-01 -2.67919779e-01 -4.04037774e-01
4.47173178e-01 2.14614525e-01 -3.82386208e-01 -2.37977564e-01
3.62817526e-01 5.96016705e-01 5.44330776e-01 -5.82103550e-01
1.22733057e+00 8.72193396e-01 1.52308255e-01 -6.08888447e-01
-1.84958172e+00 -6.08100176e-01 -7.94238925e-01 -4.70011950e-01
5.76952517e-01 -1.02344072e+00 -3.49852294e-01 7.54907012e-01
-9.43399072e-01 -8.48895133e-01 1.09428845e-01 9.82641876e-02
-1.69443801e-01 6.73767209e-01 -2.07142174e-01 -1.22160351e+00
-3.03848326e-01 -9.79646325e-01 1.33337867e+00 4.65524912e-01
6.59653962e-01 -1.12305665e+00 -4.36621159e-02 4.50567991e-01
2.02337608e-01 7.78263658e-02 6.24390304e-01 -5.29355645e-01
-8.61673713e-01 2.00119168e-01 -6.75059497e-01 3.78223687e-01
1.04909569e-01 -1.53857604e-01 -1.62960267e+00 -1.98586926e-01
-1.92846823e-02 -9.20798540e-01 1.37944055e+00 3.93175334e-01
1.41964829e+00 -2.56032180e-02 -6.46118462e-01 5.72276711e-01
1.25863469e+00 -1.95490658e-01 1.42919600e-01 3.05437669e-02
1.07291698e+00 5.92102289e-01 1.10837233e+00 4.44165468e-01
6.96412444e-01 4.07467097e-01 5.55158973e-01 -6.20046496e-01
-2.73560166e-01 -3.04566503e-01 4.95470852e-01 4.36995089e-01
1.64046660e-01 -2.90482968e-01 -6.30151212e-01 2.96604931e-01
-2.07880282e+00 -4.91894156e-01 -2.52792239e-01 2.05656791e+00
7.24556863e-01 2.27807730e-01 -2.27717906e-02 -1.60758361e-01
5.19599020e-01 7.55419314e-01 -7.75931776e-01 3.49484175e-01
-3.10454462e-02 -1.03295051e-01 9.75573003e-01 3.98472786e-01
-1.24433768e+00 1.16654181e+00 6.52672625e+00 9.50532615e-01
-1.09275687e+00 4.32109624e-01 6.92052364e-01 7.10956706e-03
-2.86890298e-01 1.32120416e-01 -1.06787121e+00 4.70909178e-01
6.90611362e-01 3.14957619e-01 2.96900153e-01 9.90271866e-01
2.65563548e-01 -7.31618762e-01 -8.42472792e-01 3.92215788e-01
4.31706756e-01 -6.88879848e-01 -5.42708278e-01 8.82342532e-02
1.10740149e+00 2.00969517e-01 2.04438731e-01 4.92841870e-01
7.59099305e-01 -6.28898084e-01 6.60071969e-01 2.61091471e-01
7.23540008e-01 -3.51637006e-01 7.21839368e-01 4.64275360e-01
-1.41389477e+00 -7.38320465e-04 -2.69141257e-01 2.41319388e-01
8.57576802e-02 6.69783235e-01 -1.28065097e+00 3.07092577e-01
6.15826845e-01 2.95398891e-01 -7.09011257e-01 7.77180195e-01
-7.94370651e-01 9.14125860e-01 -4.14531082e-01 -1.18335254e-01
1.53905287e-01 2.14396939e-02 -7.28450865e-02 1.18870544e+00
-3.64393368e-02 7.31822401e-02 7.97202528e-01 5.73537350e-01
-1.24392845e-01 -2.18557864e-01 -3.17095518e-01 4.97233212e-01
7.95163274e-01 1.59730244e+00 -9.37382460e-01 -4.92819071e-01
-6.40633166e-01 9.44626749e-01 3.59741598e-01 5.18614054e-01
-1.49522650e+00 2.67310683e-02 -1.01235643e-01 -1.88275352e-01
1.81844875e-01 5.90864643e-02 -4.35201406e-01 -7.82693684e-01
-8.25891569e-02 -2.69144177e-01 2.65948743e-01 -1.00262928e+00
-9.64804947e-01 1.23598382e-01 1.35703012e-01 -8.89855206e-01
1.09946333e-01 -3.41630191e-01 -1.10017157e+00 5.90100110e-01
-2.08920026e+00 -1.63522017e+00 -9.36700463e-01 5.69846451e-01
5.75409174e-01 6.13554604e-02 3.57029945e-01 1.92800630e-02
-7.04762757e-01 7.66259134e-01 2.27597222e-01 -4.24049981e-02
1.22739053e+00 -1.43430471e+00 1.62779629e-01 8.59208345e-01
8.87027755e-02 -2.24744305e-01 4.95384663e-01 -6.73287451e-01
-1.03990638e+00 -1.30013633e+00 4.87224251e-01 -5.28964221e-01
5.84943056e-01 -3.61670226e-01 -8.34705770e-01 5.71848571e-01
-1.49518456e-02 2.89938003e-01 5.03687024e-01 3.78409214e-02
-4.69373643e-01 -3.22450429e-01 -8.43404889e-01 2.15442955e-01
8.31068277e-01 -4.96211827e-01 -2.31543690e-01 8.02899301e-01
8.33055794e-01 -5.64139128e-01 -2.74236053e-01 5.49305499e-01
6.26560450e-01 -1.08523369e+00 8.23339105e-01 4.02764902e-02
1.72401294e-01 -2.28918836e-01 -1.69928689e-02 -1.12280464e+00
1.55205503e-01 -3.11069310e-01 -4.98602808e-01 1.32454097e+00
3.51822197e-01 -5.05195022e-01 1.10400498e+00 3.01511139e-01
-3.71829212e-01 -1.03658199e+00 -7.51524448e-01 -7.73571253e-01
-3.34744036e-01 -2.64005393e-01 2.79569685e-01 4.51747715e-01
-6.39055133e-01 4.89876598e-01 -8.09785306e-01 5.97894847e-01
1.06403840e+00 4.97616917e-01 1.14204752e+00 -1.22159719e+00
-2.35977203e-01 1.79261267e-01 5.86765051e-01 -1.34981513e+00
5.41135669e-01 -3.27705324e-01 9.25377905e-01 -1.40045023e+00
8.85721982e-01 -6.56118333e-01 -3.63529176e-01 7.20739603e-01
-6.46402657e-01 3.04729581e-01 1.01436414e-01 4.07785118e-01
-1.07004905e+00 9.08048093e-01 1.32313228e+00 6.51122956e-03
-1.83789641e-01 2.47514457e-01 -3.73856187e-01 9.34008062e-01
4.54528958e-01 -6.49630070e-01 -6.51443779e-01 -2.89732456e-01
-5.96615911e-01 -1.03974551e-01 4.98017311e-01 -9.32455778e-01
2.97850966e-01 -5.69796979e-01 5.84747612e-01 -9.15751159e-01
5.54738641e-01 -8.53804171e-01 -6.09699667e-01 3.12999338e-01
-6.34442866e-02 -8.09394538e-01 -2.44640097e-01 9.87265885e-01
3.93597364e-01 1.19133085e-01 9.78634000e-01 1.96517944e-01
-6.59368038e-01 4.21056449e-01 -1.01056732e-02 8.28159004e-02
1.17567801e+00 -3.61277573e-02 -7.45897174e-01 -5.14653087e-01
6.98801801e-02 6.59232795e-01 3.24051440e-01 2.44794995e-01
5.76691687e-01 -9.45331514e-01 -3.57109010e-01 -8.07841197e-02
1.59056157e-01 2.89543033e-01 3.02481148e-02 8.91682744e-01
8.08858052e-02 3.56631041e-01 1.09659567e-01 -7.52388597e-01
-1.42093384e+00 3.82596493e-01 2.42254615e-01 -3.30260217e-01
-4.38481897e-01 7.74219632e-01 6.35086715e-01 -3.23787212e-01
8.47936451e-01 -2.16138422e-01 1.54396713e-01 -2.14889750e-01
1.64352894e-01 3.83580238e-01 -2.00311691e-01 -4.41056162e-01
-3.53228509e-01 2.77376890e-01 6.36105686e-02 -3.25996615e-02
1.06902039e+00 -2.15783268e-01 3.19987029e-01 5.46134651e-01
7.99443007e-01 1.42671093e-01 -2.03389525e+00 -5.85662723e-01
-2.26374060e-01 -4.61882383e-01 2.08786950e-01 -8.55327189e-01
-1.14246917e+00 9.60336745e-01 7.64593959e-01 -1.41778514e-01
1.14661407e+00 9.56121534e-02 1.02974069e+00 5.63407004e-01
3.57827753e-01 -1.29864562e+00 6.35055184e-01 4.86297816e-01
2.03149945e-01 -1.79566646e+00 4.27729428e-01 -6.00486457e-01
-7.73180723e-01 7.00631559e-01 1.06218374e+00 1.35471895e-01
7.52619743e-01 5.61033428e-01 4.10447359e-01 6.46658093e-02
-6.44502819e-01 -2.61920750e-01 4.47515249e-01 7.86213994e-01
1.08086489e-01 1.54012918e-01 8.46946165e-02 3.94236326e-01
3.67167294e-01 -2.87120342e-01 2.61419892e-01 8.97978365e-01
-1.01962709e+00 -8.00220728e-01 -5.70538223e-01 4.86421287e-01
-1.33120328e-01 -5.01002856e-02 -7.48325646e-01 6.23314142e-01
8.26782212e-02 1.04968715e+00 -4.76700991e-01 -3.72101724e-01
-1.09816037e-01 -6.44314736e-02 2.34412745e-01 -6.17996097e-01
-3.88572961e-02 2.01420486e-01 -1.70846835e-01 -4.44435149e-01
-4.01288271e-01 -4.86365288e-01 -1.34202766e+00 -6.68899268e-02
-7.05192327e-01 -4.18200046e-02 4.75461543e-01 1.29861534e+00
-7.65066594e-02 5.67682683e-01 8.85074854e-01 -1.04576361e+00
-8.10781121e-01 -1.05728590e+00 -6.16649389e-01 -4.74303104e-02
4.69844937e-01 -1.00369930e+00 -5.04714727e-01 4.51560169e-02] | [10.850373268127441, -4.1140055656433105] |
0a261c5a-14ca-4b0d-a75e-ce3af4fd18f4 | bridge-prompt-towards-ordinal-action | 2203.14104 | null | https://arxiv.org/abs/2203.14104v1 | https://arxiv.org/pdf/2203.14104v1.pdf | Bridge-Prompt: Towards Ordinal Action Understanding in Instructional Videos | Action recognition models have shown a promising capability to classify human actions in short video clips. In a real scenario, multiple correlated human actions commonly occur in particular orders, forming semantically meaningful human activities. Conventional action recognition approaches focus on analyzing single actions. However, they fail to fully reason about the contextual relations between adjacent actions, which provide potential temporal logic for understanding long videos. In this paper, we propose a prompt-based framework, Bridge-Prompt (Br-Prompt), to model the semantics across adjacent actions, so that it simultaneously exploits both out-of-context and contextual information from a series of ordinal actions in instructional videos. More specifically, we reformulate the individual action labels as integrated text prompts for supervision, which bridge the gap between individual action semantics. The generated text prompts are paired with corresponding video clips, and together co-train the text encoder and the video encoder via a contrastive approach. The learned vision encoder has a stronger capability for ordinal-action-related downstream tasks, e.g. action segmentation and human activity recognition. We evaluate the performances of our approach on several video datasets: Georgia Tech Egocentric Activities (GTEA), 50Salads, and the Breakfast dataset. Br-Prompt achieves state-of-the-art on multiple benchmarks. Code is available at https://github.com/ttlmh/Bridge-Prompt | ['Jiwen Lu', 'Jie zhou', 'Jianjiang Feng', 'Zhilan Hu', 'Yueqi Duan', 'Lei Chen', 'Muheng Li'] | 2022-03-26 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Li_Bridge-Prompt_Towards_Ordinal_Action_Understanding_in_Instructional_Videos_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Li_Bridge-Prompt_Towards_Ordinal_Action_Understanding_in_Instructional_Videos_CVPR_2022_paper.pdf | cvpr-2022-1 | ['action-understanding'] | ['computer-vision'] | [ 3.25909495e-01 -1.80108875e-01 -5.55988491e-01 -5.09122491e-01
-5.98618031e-01 -5.09007692e-01 7.21286297e-01 -1.95096787e-02
-2.77608037e-01 4.10164267e-01 7.32997060e-01 2.37878710e-02
-1.13285162e-01 -2.92827785e-01 -8.07523310e-01 -6.33692503e-01
6.38135746e-02 -4.37302068e-02 4.09705967e-01 7.27116913e-02
2.61020541e-01 1.96806833e-01 -1.68829405e+00 8.43954682e-01
6.41085625e-01 9.99195039e-01 2.61893153e-01 7.67645419e-01
-1.29355490e-03 1.58865178e+00 -3.59660238e-01 -1.54133722e-01
4.29299176e-02 -7.91483164e-01 -1.00483167e+00 4.44326609e-01
4.67504919e-01 -6.80421650e-01 -7.09454179e-01 8.74312758e-01
2.18885317e-02 5.81711709e-01 3.34495246e-01 -1.45647740e+00
-2.82064587e-01 6.58965886e-01 -4.68326449e-01 5.36229908e-01
8.20017636e-01 4.11781609e-01 1.08449745e+00 -6.97866499e-01
7.49212861e-01 1.08587468e+00 1.83898240e-01 4.80532378e-01
-6.75930262e-01 -4.71060634e-01 5.78067780e-01 7.79589355e-01
-9.24257517e-01 -3.83916914e-01 7.27497518e-01 -6.68567836e-01
9.71863210e-01 1.48536354e-01 9.88720000e-01 1.49711204e+00
1.53051734e-01 1.51943338e+00 7.23525643e-01 -2.11619273e-01
1.48092002e-01 -6.49276674e-01 2.51327127e-01 6.27651095e-01
-2.58097202e-01 -1.47174880e-01 -9.53156948e-01 4.45530027e-01
9.04998362e-01 5.48700094e-01 -3.72791588e-01 -2.49870002e-01
-1.59675658e+00 4.37363416e-01 1.85119659e-01 4.04908895e-01
-5.82442343e-01 2.38518655e-01 5.09274483e-01 1.52364582e-01
1.85713559e-01 2.47969717e-01 -3.09530348e-01 -8.27274323e-01
-6.64248466e-01 1.25135615e-01 4.86074477e-01 1.04987967e+00
4.43716884e-01 -2.06595957e-01 -5.46266913e-01 4.84934300e-01
1.23544060e-01 1.02549627e-01 5.63077033e-01 -1.36442852e+00
6.29906356e-01 6.43812299e-01 9.96224582e-02 -7.32227504e-01
-2.59494156e-01 1.08922601e-01 -3.46585363e-01 -2.86206216e-01
5.65706968e-01 4.70976382e-02 -9.04002666e-01 1.65389550e+00
4.44838941e-01 7.89914250e-01 8.08797851e-02 9.72300351e-01
8.27310026e-01 7.14481533e-01 3.25028658e-01 -2.45486841e-01
1.36469197e+00 -1.53353393e+00 -8.56943905e-01 -2.54094839e-01
9.19979990e-01 -4.78438348e-01 1.19311643e+00 4.53210294e-01
-1.01143527e+00 -7.17123091e-01 -6.85051858e-01 -1.26739174e-01
-1.65940344e-01 2.03074679e-01 4.87692028e-01 -1.00941069e-01
-6.41475677e-01 5.32714069e-01 -1.24069130e+00 -3.34707499e-01
4.84607846e-01 2.81247329e-02 -4.25729334e-01 -1.37102395e-01
-9.91096258e-01 3.72143984e-01 4.41023856e-01 4.80986908e-02
-1.18402338e+00 -5.40039122e-01 -1.06073272e+00 -1.27430931e-02
9.28396225e-01 -2.91644901e-01 1.69557202e+00 -1.25110948e+00
-1.58051169e+00 6.57157183e-01 -1.41897753e-01 -4.61067230e-01
4.89272594e-01 -6.60824358e-01 -4.25880849e-01 5.55488944e-01
1.59910455e-01 7.70610154e-01 6.34526253e-01 -7.01385081e-01
-8.42086792e-01 -2.04754189e-01 4.58958477e-01 4.99084473e-01
-2.51966596e-01 1.61868006e-01 -8.15914035e-01 -7.11487353e-01
-1.14910394e-01 -9.20230508e-01 -5.25526777e-02 -2.49390870e-01
-2.01193497e-01 -4.81589198e-01 7.92898595e-01 -6.04831219e-01
1.39055371e+00 -2.38206840e+00 2.81012267e-01 -3.39532971e-01
-5.87518513e-02 1.16163440e-01 -2.96986848e-01 4.47252274e-01
-2.34187856e-01 -3.69821161e-01 2.85463259e-02 -2.27784067e-01
-2.86173839e-02 4.13851708e-01 -3.51976663e-01 3.16168547e-01
2.20328420e-02 9.78958488e-01 -1.26502204e+00 -6.11178637e-01
4.37738448e-01 2.42846444e-01 -5.92042983e-01 4.32841331e-01
-4.46156681e-01 7.24855959e-01 -6.38377666e-01 7.92799234e-01
-5.09033874e-02 -3.30038697e-01 1.85117424e-01 -8.95883739e-02
-4.39271517e-02 4.37503099e-01 -9.91689682e-01 2.18532014e+00
-1.64049387e-01 7.37379789e-01 -4.13787544e-01 -1.08528149e+00
3.98441851e-01 3.86820942e-01 9.86239314e-01 -8.42409313e-01
5.82229644e-02 -9.70745832e-02 -1.69301793e-01 -9.07512546e-01
4.02984351e-01 3.44211429e-01 -1.88362688e-01 3.34979624e-01
3.66791576e-01 3.18820029e-01 6.59313023e-01 3.17095608e-01
1.31061125e+00 7.37215102e-01 4.79041159e-01 1.92555949e-01
5.85087419e-01 1.10425474e-02 6.20424211e-01 6.03619874e-01
-5.11161923e-01 4.99393761e-01 8.21765482e-01 -4.02770668e-01
-6.13613605e-01 -8.82090271e-01 3.81178081e-01 1.41413212e+00
3.66353363e-01 -7.43519068e-01 -8.45176697e-01 -1.03726816e+00
-3.54350924e-01 6.12163603e-01 -6.25140131e-01 -1.87452599e-01
-8.02502751e-01 -7.75345862e-02 3.35353553e-01 9.26414967e-01
5.55412591e-01 -1.45029497e+00 -8.85691941e-01 1.28843531e-01
-5.94770789e-01 -1.59267485e+00 -7.66027927e-01 2.33592968e-02
-7.82423019e-01 -1.39270449e+00 -4.48029280e-01 -6.03493750e-01
5.16950786e-01 5.38042307e-01 9.93412852e-01 -9.18171406e-02
-4.82316650e-02 8.57093811e-01 -7.16242254e-01 -4.03156467e-02
-6.98867142e-02 -2.53984183e-01 -3.36017199e-02 2.74589837e-01
6.31208658e-01 -3.29261988e-01 -7.30926394e-01 4.64902729e-01
-9.14087474e-01 4.03600723e-01 5.57122529e-01 7.61896431e-01
7.55596876e-01 -1.76845416e-01 1.92640230e-01 -6.14789724e-01
1.12139113e-01 -4.50377733e-01 -2.72271305e-01 3.19034606e-01
3.22936103e-02 -1.84637487e-01 5.80427825e-01 -5.43804228e-01
-1.13691115e+00 3.04868340e-01 1.16249964e-01 -7.87064850e-01
-5.54215968e-01 4.68670130e-01 -3.34034503e-01 5.18414140e-01
2.70646781e-01 3.11342478e-01 -3.58844697e-01 -3.34558249e-01
2.33175039e-01 5.28394163e-01 7.51886129e-01 -6.52949333e-01
2.61566639e-01 3.78026754e-01 -1.59613892e-01 -7.12915838e-01
-1.15104067e+00 -8.28806460e-01 -8.80429208e-01 -7.03602076e-01
1.29583788e+00 -1.14441466e+00 -6.97429180e-01 6.92473233e-01
-9.56192732e-01 -8.08143616e-01 -4.12939578e-01 7.83865273e-01
-9.49942291e-01 3.68760228e-01 -8.15160036e-01 -4.98631150e-01
1.64459765e-01 -1.18345213e+00 1.15860212e+00 4.19090152e-01
-3.28528613e-01 -7.83785701e-01 -1.13118425e-01 8.81545663e-01
-3.73537689e-01 1.76197410e-01 3.36207688e-01 -7.65140295e-01
-6.34166956e-01 -5.64632826e-02 3.60970274e-02 2.57462353e-01
1.83662549e-01 2.89168651e-03 -7.08975136e-01 -4.31420207e-02
-1.69090629e-01 -5.28805852e-01 7.15434074e-01 3.41399163e-01
1.53260052e+00 -3.74887317e-01 -1.73237756e-01 4.14020538e-01
8.43871176e-01 5.16084731e-01 6.80642366e-01 1.83848158e-01
9.53306556e-01 6.73482895e-01 1.18864596e+00 4.94236410e-01
3.06323141e-01 7.41611004e-01 2.98206747e-01 3.10806990e-01
-8.10968652e-02 -4.80000228e-01 8.53930056e-01 8.86358202e-01
-2.79643387e-01 -2.84752637e-01 -8.70123029e-01 4.68575358e-01
-2.32270217e+00 -1.33355534e+00 -9.46834460e-02 1.78592575e+00
7.89103806e-01 1.82769239e-01 2.27536812e-01 -1.04103491e-01
4.98483121e-01 5.81196725e-01 -5.67890406e-01 -5.02855964e-02
3.53133082e-01 -1.81419924e-01 1.77996412e-01 2.26102933e-01
-1.32883286e+00 9.64917183e-01 5.29983807e+00 7.04645753e-01
-8.87616277e-01 9.75997522e-02 4.80417490e-01 -4.13423926e-01
9.81868356e-02 -1.79117043e-02 -6.66490078e-01 6.45422041e-01
9.10540462e-01 -1.13221025e-02 1.82474181e-01 9.64518726e-01
4.71574545e-01 -3.19720775e-01 -1.42682445e+00 9.63293016e-01
2.78896898e-01 -1.13654172e+00 5.42373545e-02 -1.63340807e-01
7.65520096e-01 -1.44539744e-01 -1.98329240e-01 4.74419534e-01
1.90431789e-01 -7.13746309e-01 8.48018944e-01 6.26562238e-01
4.88322288e-01 -4.16855127e-01 4.60125268e-01 3.06417257e-01
-1.35581696e+00 -3.10912073e-01 1.94284618e-01 -1.92757308e-01
2.17209563e-01 1.20548137e-01 -4.77955401e-01 5.97183585e-01
7.85790145e-01 1.54501510e+00 -5.06003797e-01 8.43722224e-01
-6.38414025e-01 5.86315095e-01 3.35321650e-02 1.92267343e-01
5.64535558e-01 -2.59076387e-01 4.12431479e-01 1.15725923e+00
2.35793740e-01 5.09328306e-01 5.04359901e-01 3.26844573e-01
1.67519581e-02 -1.54832840e-01 -5.72570503e-01 -4.04285192e-01
2.28800878e-01 1.01090252e+00 -8.26659858e-01 -6.09714210e-01
-8.30752909e-01 1.09882557e+00 7.35897794e-02 4.54469055e-01
-1.17583036e+00 4.61998805e-02 8.31587732e-01 -4.27119173e-02
3.53652149e-01 -2.84842223e-01 1.36347786e-01 -1.45333230e+00
1.62927240e-01 -1.09410036e+00 7.75208533e-01 -1.01179373e+00
-7.23776817e-01 6.65673167e-02 2.97607571e-01 -1.65664005e+00
-3.54875237e-01 -5.71906626e-01 -5.94734430e-01 5.17451726e-02
-1.09160137e+00 -9.01717842e-01 -6.66996479e-01 8.74498129e-01
1.27247310e+00 1.37445331e-01 3.52368176e-01 3.53154421e-01
-8.41628671e-01 2.21000656e-01 -3.82581770e-01 3.47678125e-01
6.94685638e-01 -1.13481522e+00 9.30864960e-02 1.16809106e+00
5.33288896e-01 3.45414042e-01 5.37483096e-01 -7.11832404e-01
-1.28483522e+00 -1.01981997e+00 6.41231358e-01 -5.19855917e-01
7.76025057e-01 -1.55734345e-01 -8.07720065e-01 1.06527221e+00
2.30345324e-01 9.53093395e-02 7.65395761e-01 -1.46474302e-01
-2.24653378e-01 -2.25389767e-02 -3.90547127e-01 7.54510522e-01
1.49254847e+00 -7.33573198e-01 -8.92659843e-01 3.53366315e-01
7.96339273e-01 -6.35171890e-01 -7.17243671e-01 3.73130023e-01
5.67159832e-01 -9.88446951e-01 9.47380066e-01 -8.55274916e-01
1.01712501e+00 -3.20689917e-01 -1.32258654e-01 -9.92421865e-01
-1.17210336e-01 -5.67950904e-01 -5.27432501e-01 1.02497888e+00
-7.50918314e-02 -1.97349656e-02 7.33365893e-01 4.64218199e-01
-5.22132516e-01 -8.52584362e-01 -7.14008272e-01 -8.06976199e-01
-5.98705053e-01 -4.98005211e-01 4.05572623e-01 1.03694642e+00
3.02945524e-01 1.27149478e-01 -4.06331837e-01 -5.66722974e-02
2.93779433e-01 2.47143507e-01 8.30813706e-01 -7.42483854e-01
-3.81004244e-01 -4.78717744e-01 -4.87634957e-01 -1.49557507e+00
3.08736175e-01 -5.62614441e-01 1.94926560e-01 -1.53985000e+00
2.78954774e-01 2.50989735e-01 -5.51605105e-01 7.56046951e-01
-2.35284477e-01 -3.27476824e-04 2.16516539e-01 2.18286097e-01
-1.35773420e+00 5.25547683e-01 1.42008197e+00 -1.81312174e-01
-2.92770714e-01 -1.28526956e-01 -2.39609793e-01 8.71727109e-01
6.36343420e-01 -2.44968638e-01 -6.83009326e-01 -3.88106853e-01
-1.57087713e-01 3.80057126e-01 4.82306421e-01 -1.28513849e+00
3.61910254e-01 -7.33490527e-01 2.53103793e-01 -5.64304650e-01
2.68936455e-01 -7.53905714e-01 -2.35739183e-02 3.42393667e-01
-6.37608111e-01 -1.28991127e-01 -1.53584212e-01 6.38596833e-01
-6.14096642e-01 -8.11119527e-02 3.19182038e-01 -9.19769853e-02
-1.38387883e+00 4.04103279e-01 -3.95748138e-01 1.74235612e-01
1.41507757e+00 -3.84834468e-01 -3.25846732e-01 -3.52741480e-01
-8.33704650e-01 3.89672905e-01 2.66665310e-01 8.41600716e-01
5.22588253e-01 -1.32078230e+00 -2.74755418e-01 7.43947998e-02
3.87245983e-01 4.53870818e-02 6.38175428e-01 1.21640038e+00
-3.37767899e-01 4.12079334e-01 -4.38304871e-01 -6.86154425e-01
-1.45404708e+00 5.76746225e-01 1.85971200e-01 -2.37143412e-01
-9.37163413e-01 7.23366082e-01 4.23446804e-01 1.40011877e-01
4.53034699e-01 -6.40684605e-01 -4.58793133e-01 2.88073301e-01
6.48187816e-01 2.89622992e-01 -3.46369147e-01 -5.58417499e-01
-3.59713227e-01 3.43266904e-01 5.12419492e-02 -1.91360004e-02
1.23662746e+00 -7.71824718e-02 2.55953252e-01 6.98960721e-01
1.01606011e+00 -3.52171212e-01 -1.88995945e+00 -2.96829432e-01
1.48634106e-01 -6.34796143e-01 -2.23681122e-01 -5.26812851e-01
-1.10242498e+00 7.98292577e-01 1.16512187e-01 -6.44025579e-02
1.30337274e+00 1.36474952e-01 8.96902144e-01 3.40887278e-01
3.08977038e-01 -1.32189393e+00 8.14344704e-01 5.52656054e-01
7.03881085e-01 -1.25325429e+00 -8.07540789e-02 -3.97358686e-01
-1.05036294e+00 1.08173335e+00 1.02655399e+00 1.22395620e-01
2.33490750e-01 -2.09581237e-02 -1.45070791e-01 -2.42845371e-01
-9.58367527e-01 -3.08983594e-01 3.21064532e-01 1.92930102e-01
4.01298702e-01 1.01822488e-01 -2.07978711e-01 7.33998120e-01
2.02673301e-01 3.07509601e-01 6.12880290e-01 1.15984154e+00
-3.38538229e-01 -8.44334424e-01 -9.98353213e-02 3.33311856e-01
-4.16292071e-01 2.46332020e-01 -2.59583145e-01 5.98203182e-01
1.10345684e-01 8.37545097e-01 3.53620946e-01 -4.98616606e-01
4.07071978e-01 3.04914415e-01 4.42628145e-01 -6.13263011e-01
-3.36668432e-01 7.39816576e-02 1.51138440e-01 -1.28145730e+00
-7.98982084e-01 -1.08657670e+00 -1.45948136e+00 -1.75658558e-02
2.35446259e-01 1.95537172e-02 7.20647648e-02 1.15004504e+00
2.51335829e-01 8.18061709e-01 4.94564056e-01 -6.81666374e-01
-2.65363306e-01 -8.25428605e-01 -3.75106186e-01 8.06431592e-01
2.38662496e-01 -7.57236183e-01 -2.17248619e-01 5.09237885e-01] | [8.54498291015625, 0.6338307857513428] |
97c8ea4b-726d-4743-aa42-32e51bd5ddff | faceless-person-recognition-privacy | 1607.08438 | null | http://arxiv.org/abs/1607.08438v1 | http://arxiv.org/pdf/1607.08438v1.pdf | Faceless Person Recognition; Privacy Implications in Social Media | As we shift more of our lives into the virtual domain, the volume of data
shared on the web keeps increasing and presents a threat to our privacy. This
works contributes to the understanding of privacy implications of such data
sharing by analysing how well people are recognisable in social media data. To
facilitate a systematic study we define a number of scenarios considering
factors such as how many heads of a person are tagged and if those heads are
obfuscated or not. We propose a robust person recognition system that can
handle large variations in pose and clothing, and can be trained with few
training samples. Our results indicate that a handful of images is enough to
threaten users' privacy, even in the presence of obfuscation. We show detailed
experimental results, and discuss their implications. | ['Rodrigo Benenson', 'Seong Joon Oh', 'Mario Fritz', 'Bernt Schiele'] | 2016-07-28 | null | null | null | null | ['person-recognition'] | ['computer-vision'] | [ 1.90891981e-01 2.18137249e-01 7.37253278e-02 -5.66505015e-01
-1.21286847e-01 -8.95378053e-01 4.22604173e-01 1.67914152e-01
-6.59014404e-01 6.68919802e-01 2.84626365e-01 5.71981929e-02
3.20790648e-01 -7.78781533e-01 -6.59031212e-01 -3.46599430e-01
-2.32727200e-01 1.19341187e-01 1.29240632e-01 1.92099307e-02
-2.60820776e-01 4.17492092e-01 -1.46255851e+00 2.39389986e-01
3.59515458e-01 9.08740163e-01 -3.83636534e-01 5.02078891e-01
3.40026855e-01 1.05441384e-01 -7.43595839e-01 -1.51682222e+00
6.48249805e-01 1.09237626e-01 -5.24947166e-01 1.64091676e-01
9.09191310e-01 -6.87974274e-01 -3.92553896e-01 1.07614732e+00
6.14129126e-01 -1.99336141e-01 1.57739341e-01 -1.18459833e+00
-4.62061137e-01 2.07585365e-01 -3.26801479e-01 3.12365647e-02
7.43933499e-01 3.40027846e-02 5.31835616e-01 -2.74153113e-01
6.29035532e-01 1.01927590e+00 9.56870437e-01 6.45402372e-01
-1.24416351e+00 -6.36952698e-01 2.55356468e-02 -1.75575286e-01
-1.72713768e+00 -7.05880284e-01 5.65810025e-01 -3.26325148e-01
1.41599521e-01 8.83815289e-01 7.30003715e-01 1.47886038e+00
-4.79178950e-02 5.15533924e-01 9.57772136e-01 -3.70679080e-01
1.05975695e-01 8.74633133e-01 1.58792585e-01 5.03276229e-01
9.22837377e-01 4.71269637e-02 -5.38787603e-01 -6.81019783e-01
4.96823728e-01 1.80559292e-01 -1.62698299e-01 -6.91767931e-01
-7.39366710e-01 6.94478869e-01 1.91795111e-01 5.78194074e-02
-1.37318214e-02 -3.13791424e-01 4.51782614e-01 7.25665465e-02
3.93357038e-01 3.33289325e-01 -2.78467178e-01 -5.28216828e-03
-7.49799311e-01 4.70291823e-01 9.50807750e-01 1.08927310e+00
4.77331400e-01 -5.91889322e-01 1.11353800e-01 7.79672623e-01
1.32387146e-01 6.69580638e-01 3.02317291e-01 -7.41516054e-01
5.01926184e-01 5.88869750e-01 4.17422265e-01 -1.53477764e+00
-8.55989531e-02 -6.54013157e-02 -4.31891441e-01 -2.90039051e-02
8.08485806e-01 -5.45072138e-01 -3.56051236e-01 1.47335744e+00
3.86359990e-01 -3.54515880e-01 -4.29699302e-01 9.33309138e-01
4.91224706e-01 1.05175758e-02 1.74897075e-01 -4.91313078e-02
1.63022316e+00 -2.10974589e-01 -8.65505278e-01 -4.54598665e-01
4.06800121e-01 -5.64890563e-01 7.01386750e-01 2.70186454e-01
-9.67855692e-01 -3.15817595e-01 -8.85780931e-01 5.94063662e-02
-7.74327695e-01 -2.34502062e-01 8.08576167e-01 1.71789384e+00
-8.66020739e-01 5.97349524e-01 -4.48529571e-01 -8.59877408e-01
5.70780694e-01 5.47758758e-01 -6.67637885e-01 -4.54580970e-02
-1.16721296e+00 7.28294492e-01 3.46426759e-03 4.45592701e-02
-1.82378829e-01 -2.58223206e-01 -8.81755412e-01 -2.21473500e-01
3.40509295e-01 -5.22056937e-01 1.03408694e+00 -1.10402429e+00
-7.34793663e-01 1.29988146e+00 -4.70402092e-02 -4.09290016e-01
1.14084673e+00 -4.07899857e-01 -7.79112160e-01 8.07855278e-03
-1.17146544e-01 4.38438892e-01 9.11680579e-01 -1.37697375e+00
-3.55047584e-01 -1.03941703e+00 2.02865917e-02 2.26734020e-02
-5.94116032e-01 3.81220400e-01 -4.25079286e-01 -5.82077384e-01
-2.23088101e-01 -1.08373022e+00 -4.53209542e-02 4.83617336e-01
-4.45893735e-01 6.03727281e-01 9.80785251e-01 -1.11431170e+00
1.12716305e+00 -2.14148998e+00 -3.09702069e-01 3.41593713e-01
1.69894829e-01 5.67603528e-01 4.46718723e-01 3.67523015e-01
2.72499800e-01 5.92266560e-01 1.02424406e-01 -6.23624384e-01
7.87971243e-02 1.75746381e-01 -7.21019134e-02 7.74940550e-01
-4.52643067e-01 7.23196983e-01 -6.40277863e-01 -4.16811764e-01
9.32467654e-02 6.01641357e-01 -2.61237621e-01 2.16817290e-01
4.16691571e-01 2.09168553e-01 -2.97069520e-01 7.34435320e-01
1.06438148e+00 2.35572577e-01 3.93206388e-01 2.23711152e-02
1.22585684e-01 1.84071437e-03 -1.18006098e+00 9.12066102e-01
9.36358497e-02 5.00494897e-01 4.71566975e-01 -1.62125036e-01
6.55032992e-01 1.68628722e-01 1.31329462e-01 -4.00422096e-01
3.89866710e-01 -1.26724720e-01 -3.69318873e-01 -5.18962741e-01
6.76822603e-01 1.18299425e-01 -1.96756348e-01 2.47362390e-01
-3.47528338e-01 5.73441505e-01 -1.93014517e-01 4.28386107e-02
8.67846489e-01 -3.59328985e-01 3.84175420e-01 -9.75804105e-02
1.83350384e-01 -5.62070131e-01 3.37326437e-01 7.39047527e-01
-8.38831306e-01 5.93574405e-01 9.16968286e-02 -4.92422611e-01
-1.05182552e+00 -1.10692406e+00 -1.65003359e-01 1.07566762e+00
3.15908998e-01 -3.99650067e-01 -9.77592587e-01 -9.74781752e-01
5.20986676e-01 4.94144857e-01 -6.14517033e-01 -1.61819860e-01
-6.32771999e-02 -6.63976252e-01 8.41803014e-01 1.28795594e-01
6.76344097e-01 -5.60680091e-01 -5.55187941e-01 -3.51502508e-01
-1.24897890e-01 -1.21359134e+00 -5.54759502e-01 -5.23455739e-01
-4.28943932e-01 -9.82471228e-01 -8.41820657e-01 -3.25414836e-01
8.62793922e-01 4.77565467e-01 7.59099185e-01 2.65720308e-01
-3.75575811e-01 7.93824017e-01 -3.41925174e-01 -5.74880719e-01
-1.50255457e-01 -2.26767793e-01 2.52026021e-01 5.31257272e-01
6.26971722e-01 -3.89948428e-01 -5.10595918e-01 5.40081680e-01
-6.73052132e-01 -3.47048938e-01 7.15676509e-03 4.14179228e-02
-1.46247596e-01 2.03396976e-01 -2.23508030e-02 -1.19087112e+00
5.58990598e-01 -4.28294867e-01 -2.03187570e-01 1.34383515e-01
-2.76299953e-01 -5.25947094e-01 2.55770892e-01 -4.96941894e-01
-8.64767313e-01 1.71714008e-01 1.78695410e-01 -3.26315984e-02
-5.07994771e-01 -3.72578412e-01 -5.23388445e-01 -5.34186959e-01
4.79144990e-01 1.98344141e-02 1.49865851e-01 -5.15389383e-01
1.59417331e-01 9.14591193e-01 1.43230155e-01 -1.53320938e-01
9.26162362e-01 8.68519425e-01 -4.81433183e-01 -1.18510926e+00
-4.69898045e-01 -6.03498757e-01 -6.80668235e-01 -5.02082229e-01
7.54121304e-01 -6.81585968e-01 -1.02904069e+00 5.96547067e-01
-9.02931333e-01 2.46882752e-01 1.24398433e-01 1.57020941e-01
3.14729251e-02 7.33814240e-01 -4.74053711e-01 -1.23216808e+00
1.20090423e-02 -6.34388447e-01 8.64304960e-01 1.80051342e-01
-7.36505210e-01 -8.87429416e-01 -1.68143660e-01 7.71588206e-01
2.82331407e-01 4.32140619e-01 -5.98602407e-02 -6.72486007e-01
-3.81203771e-01 -7.82998979e-01 -1.01286307e-01 1.69183701e-01
2.82076389e-01 -3.55133533e-01 -1.09403718e+00 -5.37362039e-01
-7.47892708e-02 -6.15025237e-02 3.33721340e-01 -1.44947201e-01
1.13104713e+00 -9.42135990e-01 -5.50717354e-01 5.43937325e-01
1.11034513e+00 -1.96265757e-01 8.19783926e-01 3.03227723e-01
7.77782083e-01 1.01482487e+00 1.59471542e-01 6.11361027e-01
4.80384529e-01 7.41440177e-01 3.95885497e-01 6.19230345e-02
3.73034149e-01 -4.44620967e-01 2.08482087e-01 -7.32215792e-02
-2.14022368e-01 -2.53380269e-01 -6.96964025e-01 3.77251148e-01
-1.52406168e+00 -9.80892897e-01 1.26547907e-02 2.64902663e+00
5.13331115e-01 -1.20817581e-02 6.83401346e-01 7.65290037e-02
1.11116588e+00 2.50801504e-01 -2.42795974e-01 -4.32144463e-01
8.57102498e-02 -3.41847152e-01 1.23074269e+00 4.08085555e-01
-1.23296893e+00 5.37494302e-01 6.93703175e+00 -1.95066761e-02
-9.17791665e-01 -1.35448739e-01 7.67858446e-01 -2.45310053e-01
-4.31160927e-02 -4.19689357e-01 -8.76241803e-01 8.29903245e-01
8.00562620e-01 1.27450109e-01 5.70521533e-01 8.43454301e-01
-5.48988804e-02 -1.12098143e-01 -1.03257835e+00 1.05645776e+00
4.94975448e-01 -6.88131273e-01 -1.94083884e-01 5.85058451e-01
1.93694711e-01 -4.40750033e-01 1.85134307e-01 -9.81217176e-02
1.53228074e-01 -8.10208082e-01 6.78459704e-01 5.13079882e-01
4.65477526e-01 -6.64320648e-01 7.31078804e-01 3.83071989e-01
-7.60907292e-01 -7.71737620e-02 -2.52897620e-01 -1.50496095e-01
2.20018640e-01 2.42609322e-01 -8.06853473e-01 1.70743123e-01
8.26354802e-01 9.73109007e-02 -9.60255384e-01 1.00225163e+00
4.64928001e-02 3.84221911e-01 -5.26815772e-01 -2.39939466e-01
-5.41063488e-01 5.92348874e-02 5.83238780e-01 1.17697859e+00
1.73113018e-01 9.44124535e-02 -1.82194367e-01 5.58690071e-01
-5.12734056e-02 3.73786427e-02 -8.19793701e-01 -3.16230301e-03
4.75301206e-01 1.09374523e+00 -4.30171758e-01 5.95228449e-02
-4.71841961e-01 1.60880232e+00 2.72946149e-01 2.49876365e-01
-5.55987477e-01 1.68287533e-03 7.89023936e-01 7.04717577e-01
1.95880190e-01 -2.36759707e-01 -3.13210636e-01 -1.20718670e+00
4.09792811e-01 -8.32344592e-01 5.63232541e-01 -5.02809227e-01
-1.24541974e+00 6.32602572e-02 -1.46911651e-01 -9.24094141e-01
-3.90541442e-02 -3.88106316e-01 -9.85741913e-02 5.17583489e-01
-8.63604069e-01 -1.26679015e+00 -2.46522442e-01 4.84123945e-01
-2.98821509e-01 -2.14756001e-02 9.14994061e-01 4.55078185e-01
-4.99864787e-01 1.15200686e+00 -6.43498525e-02 4.25835848e-01
6.60848200e-01 -8.01776707e-01 4.89350051e-01 7.20842302e-01
-2.24418342e-02 1.00346041e+00 7.78878629e-01 -8.28962982e-01
-1.51649928e+00 -1.04087186e+00 1.15981472e+00 -8.43557656e-01
2.36225918e-01 -9.58688617e-01 -7.22792268e-01 9.55491722e-01
-2.64566660e-01 8.94496217e-02 1.08088160e+00 3.72510135e-01
-5.58373988e-01 -2.35766664e-01 -1.88693202e+00 6.41188323e-01
1.16632557e+00 -6.88974082e-01 -4.50468808e-01 2.73064524e-02
1.75509065e-01 -2.20006362e-01 -7.44271159e-01 8.58955532e-02
1.06450582e+00 -1.28647697e+00 1.09893942e+00 -4.00096387e-01
-1.78120092e-01 1.02668785e-01 -1.89891547e-01 -8.43972266e-01
-2.03633681e-01 -6.31862819e-01 -6.76544607e-02 1.58887959e+00
9.88523588e-02 -7.80024827e-01 1.16054296e+00 1.72266388e+00
8.11488092e-01 -7.85543025e-02 -8.83345306e-01 -8.02663147e-01
-4.31060106e-01 -2.92109221e-01 6.18998468e-01 1.18280065e+00
1.52162269e-01 -2.18686953e-01 -8.34107041e-01 4.39615011e-01
9.26386297e-01 -5.45334339e-01 1.00722754e+00 -1.12243390e+00
-2.32101396e-01 2.01725960e-01 -8.40839326e-01 -5.25763273e-01
-3.25807452e-01 -4.84262347e-01 -5.93441725e-01 -8.36636901e-01
3.19005430e-01 -3.21118683e-01 1.13213211e-01 3.04615825e-01
4.65669557e-02 4.66702074e-01 3.99464220e-01 -3.06995828e-02
-6.24659419e-01 -8.80271643e-02 7.26986766e-01 2.10479256e-02
7.79000074e-02 3.56772959e-01 -1.04324937e+00 6.82947457e-01
7.38964558e-01 -2.17946231e-01 -6.02407940e-02 -1.29952073e-01
2.26361096e-01 -2.82442927e-01 7.56433785e-01 -9.75148916e-01
-1.87318828e-02 1.12205923e-01 8.63714755e-01 -1.69853359e-01
6.41431630e-01 -1.35065162e+00 2.64703631e-01 3.12157989e-01
-2.44429395e-01 -3.52244318e-01 2.70417541e-01 5.65122306e-01
4.32182848e-01 -2.10532665e-01 6.86386704e-01 -3.43233258e-01
-4.01807457e-01 3.38637590e-01 -2.20923871e-01 -2.81074554e-01
1.25860739e+00 -5.51811695e-01 -2.26547033e-01 -5.39553106e-01
-8.47372949e-01 3.65243070e-02 1.00879371e+00 4.99524474e-01
1.99758232e-01 -1.19784594e+00 -2.63219714e-01 5.20458937e-01
1.94267556e-01 -6.40833557e-01 2.20487222e-01 6.71816766e-02
-3.24589908e-01 1.07116066e-01 -2.94667602e-01 -1.80340603e-01
-1.91225982e+00 7.17861414e-01 2.60372102e-01 3.81100744e-01
-4.67934668e-01 8.19433272e-01 -1.19583368e-01 -4.51344103e-01
5.02677798e-01 1.36377856e-01 -4.57657408e-03 7.62560889e-02
8.94293368e-01 5.66444039e-01 -9.09118280e-02 -1.13431287e+00
-5.64927757e-01 1.02632858e-01 -2.74909228e-01 1.63426186e-04
9.19573605e-01 -4.51518327e-01 8.22950006e-02 1.86455339e-01
1.32021487e+00 5.06358385e-01 -1.14971256e+00 4.45230864e-02
-2.65423656e-01 -1.14816117e+00 -4.35648412e-01 -6.09954298e-01
-7.50480115e-01 3.38191569e-01 8.19932401e-01 7.50054121e-01
7.26084590e-01 -1.64784845e-02 8.18386078e-01 1.66938111e-01
8.56173754e-01 -1.12864959e+00 -5.30477822e-01 -2.69004889e-02
6.75628006e-01 -1.36729980e+00 3.31120342e-01 -8.93010855e-01
-6.32163405e-01 6.35969520e-01 3.03840935e-01 1.05459251e-01
6.13176882e-01 9.81320441e-02 8.70831236e-02 -2.90280767e-02
3.84298302e-02 1.65071905e-01 7.13059381e-02 1.08116388e+00
1.31429955e-01 4.50614780e-01 -2.72541583e-01 7.32719064e-01
-7.52653480e-01 -1.57487333e-01 4.18258727e-01 1.08455861e+00
-2.78004646e-01 -1.13686621e+00 -7.57036328e-01 4.15018320e-01
-6.84698761e-01 3.02520603e-01 -1.02710235e+00 6.19261265e-01
3.40782583e-01 8.79925489e-01 -1.52076343e-02 -5.02882063e-01
4.68633115e-01 2.88860574e-02 4.96345460e-01 -3.60106617e-01
-6.79331005e-01 -5.29645503e-01 5.63853741e-01 -4.82327193e-01
-1.77934244e-01 -1.06882977e+00 -4.85628128e-01 -8.51463258e-01
-1.30623087e-01 -2.64434487e-01 6.22179627e-01 6.50999308e-01
3.47652167e-01 -2.74486125e-01 4.49083537e-01 -6.05750620e-01
-3.05749893e-01 -4.57044303e-01 -7.98027456e-01 8.33815753e-01
3.55568647e-01 -4.78334725e-01 -4.16120708e-01 9.62220207e-02] | [12.794586181640625, 0.7971318364143372] |
b3c7c92d-1f82-4470-9f75-288b7332c70b | kg-sp-knowledge-guided-simple-primitives-for | 2205.06784 | null | https://arxiv.org/abs/2205.06784v1 | https://arxiv.org/pdf/2205.06784v1.pdf | KG-SP: Knowledge Guided Simple Primitives for Open World Compositional Zero-Shot Learning | The goal of open-world compositional zero-shot learning (OW-CZSL) is to recognize compositions of state and objects in images, given only a subset of them during training and no prior on the unseen compositions. In this setting, models operate on a huge output space, containing all possible state-object compositions. While previous works tackle the problem by learning embeddings for the compositions jointly, here we revisit a simple CZSL baseline and predict the primitives, i.e. states and objects, independently. To ensure that the model develops primitive-specific features, we equip the state and object classifiers with separate, non-linear feature extractors. Moreover, we estimate the feasibility of each composition through external knowledge, using this prior to remove unfeasible compositions from the output space. Finally, we propose a new setting, i.e. CZSL under partial supervision (pCZSL), where either only objects or state labels are available during training, and we can use our prior to estimate the missing labels. Our model, Knowledge-Guided Simple Primitives (KG-SP), achieves state of the art in both OW-CZSL and pCZSL, surpassing most recent competitors even when coupled with semi-supervised learning techniques. Code available at: https://github.com/ExplainableML/KG-SP. | ['Zeynep Akata', 'Massimiliano Mancini', 'Shyamgopal Karthik'] | 2022-05-13 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Karthik_KG-SP_Knowledge_Guided_Simple_Primitives_for_Open_World_Compositional_Zero-Shot_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Karthik_KG-SP_Knowledge_Guided_Simple_Primitives_for_Open_World_Compositional_Zero-Shot_CVPR_2022_paper.pdf | cvpr-2022-1 | ['compositional-zero-shot-learning'] | ['computer-vision'] | [ 3.05860877e-01 3.91425818e-01 -3.87556255e-01 -8.30274522e-02
-6.25830054e-01 -5.54529488e-01 8.83905768e-01 -9.18281153e-02
-2.01158613e-01 3.43676865e-01 1.58548519e-01 7.09717572e-02
1.77295759e-01 -6.91059530e-01 -1.01720536e+00 -7.79968262e-01
2.32103541e-01 8.71143341e-01 4.25276995e-01 1.42012849e-01
-1.09032743e-01 9.32078287e-02 -1.89114082e+00 3.84966642e-01
5.66102087e-01 1.15050364e+00 2.77590483e-01 7.79760957e-01
-1.30768910e-01 1.19776118e+00 -7.29086995e-02 -5.12089372e-01
3.36430341e-01 -2.77982205e-01 -7.80215740e-01 4.54861224e-01
4.28570420e-01 -3.72471601e-01 -4.89841968e-01 9.08848166e-01
1.08785175e-01 1.16190381e-01 7.21635282e-01 -1.58485901e+00
-8.71388972e-01 7.17100143e-01 9.05474871e-02 -2.15502173e-01
1.46969542e-01 6.20602906e-01 1.39438999e+00 -9.09295857e-01
8.46294045e-01 1.03678930e+00 1.65949449e-01 9.81689334e-01
-1.59936225e+00 -5.20958543e-01 5.97871065e-01 5.01233816e-01
-1.31560934e+00 -7.78482676e-01 7.41927624e-01 -7.06377864e-01
1.04309750e+00 1.03244193e-01 6.40037537e-01 1.42009592e+00
-3.43330950e-01 1.18896639e+00 9.32037711e-01 -3.03661913e-01
5.39390981e-01 1.13322720e-01 2.99833328e-01 1.03009188e+00
7.06151798e-02 -3.29838134e-02 -7.78323829e-01 -1.60062283e-01
5.32910526e-01 1.32575631e-01 -2.01740533e-01 -9.69190896e-01
-1.35049808e+00 4.56345558e-01 1.86014816e-01 -1.46240890e-02
-2.36685351e-01 3.95609260e-01 4.46804762e-02 2.25222662e-01
2.27317005e-01 2.43390620e-01 -7.40797520e-01 5.24577536e-02
-6.41579926e-01 1.72978714e-01 9.94352162e-01 1.11749566e+00
1.13233840e+00 -1.61716700e-01 -2.35983193e-01 3.76991630e-01
2.57119894e-01 3.59401613e-01 4.44537401e-01 -8.68880451e-01
2.94168979e-01 7.31817603e-01 -5.97127574e-03 -4.74815853e-02
5.03008580e-03 -1.96222201e-01 -5.17239988e-01 1.72298089e-01
2.17760697e-01 1.05215594e-01 -1.39571202e+00 2.03437996e+00
3.96997482e-01 8.49765837e-01 1.17355719e-01 6.87670588e-01
6.29688263e-01 7.41999328e-01 4.39172350e-02 -5.32627851e-03
1.19213939e+00 -1.42662930e+00 -3.26863259e-01 -3.49756956e-01
3.60522002e-01 -1.58590585e-01 1.15654683e+00 4.53586489e-01
-9.00078475e-01 -4.83147085e-01 -9.14244413e-01 -1.66565001e-01
-3.22186321e-01 2.19400272e-01 5.62257707e-01 3.24086368e-01
-9.09711123e-01 7.08097696e-01 -1.37100780e+00 1.03911767e-02
6.47442758e-01 5.28054237e-01 -4.35122401e-01 -2.52385557e-01
-6.92147195e-01 7.72358775e-01 4.46768612e-01 -2.90179104e-01
-1.77795470e+00 -7.16198981e-01 -1.21798646e+00 2.10582316e-01
9.04989600e-01 -7.87894487e-01 1.45314503e+00 -8.52602005e-01
-1.61958790e+00 6.59835100e-01 -3.55111986e-01 -4.56757247e-01
3.87288421e-01 -1.89417645e-01 -3.95001993e-02 9.68585908e-02
-7.20151514e-02 8.02958906e-01 1.13809586e+00 -1.42955673e+00
-5.49478292e-01 -1.71055928e-01 1.76519021e-01 9.71568003e-02
-7.48213530e-02 -3.01599026e-01 -7.46104062e-01 -1.45416006e-01
9.13175419e-02 -9.89176154e-01 -2.16913894e-01 4.14918214e-01
-6.10329807e-01 -3.68755102e-01 6.46314979e-01 -3.76914442e-01
7.08407104e-01 -2.00568414e+00 5.90084732e-01 -8.71363655e-02
4.26647216e-01 1.74915195e-01 -3.55317056e-01 3.02537471e-01
1.40300065e-01 -1.04021914e-01 -5.69929719e-01 -9.28364694e-01
4.28941339e-01 6.63218319e-01 -3.81756186e-01 5.23208320e-01
6.32410228e-01 1.09053171e+00 -9.64615345e-01 -4.55780447e-01
5.50209999e-01 3.15790564e-01 -6.60676897e-01 3.38193446e-01
-8.91454995e-01 4.62803960e-01 -3.33996028e-01 6.36412442e-01
3.63666028e-01 -5.32991052e-01 1.64311931e-01 -2.04568580e-02
8.52349550e-02 2.82044172e-01 -1.29670727e+00 1.86694109e+00
-6.66439235e-01 2.05040589e-01 -1.75707087e-01 -9.67462957e-01
4.08503503e-01 4.46160764e-01 4.56162602e-01 -1.47343606e-01
5.64600416e-02 1.86941959e-02 2.73481272e-02 -5.57753801e-01
-6.09457344e-02 -1.99353084e-01 -6.11889735e-02 6.51678205e-01
9.28470612e-01 -1.14909537e-01 2.79577971e-01 3.50714564e-01
1.11478317e+00 6.67370856e-01 3.70282620e-01 1.07797131e-01
4.00691479e-01 -3.00592065e-01 8.35098624e-01 7.44296372e-01
-1.03261292e-01 6.45055473e-01 5.80745101e-01 -3.19541007e-01
-1.13280988e+00 -1.18577504e+00 2.56905973e-01 1.14397264e+00
3.45455468e-01 -6.92152739e-01 -5.51684797e-01 -9.29201245e-01
-1.63350217e-02 8.31721365e-01 -6.42767668e-01 -2.85201699e-01
-4.34304059e-01 -1.27522141e-01 2.12017342e-01 4.02853727e-01
-2.53121816e-02 -1.12608981e+00 -7.26971686e-01 6.44920766e-02
2.56018694e-02 -1.27902949e+00 -2.94060349e-01 3.84171426e-01
-4.88184690e-01 -1.11268628e+00 -2.73839951e-01 -6.88583255e-01
7.69330144e-01 -1.76615104e-01 9.43063736e-01 3.16573940e-02
-2.86521316e-01 4.12325531e-01 -2.35219389e-01 -1.58559188e-01
-3.35427135e-01 -5.47479987e-02 3.62860709e-02 4.02636379e-01
5.51704057e-02 -6.44001424e-01 -3.24852198e-01 2.05526706e-02
-8.20921838e-01 5.88420272e-01 6.98382676e-01 7.75306284e-01
7.48952091e-01 -1.04663290e-01 7.52823055e-02 -8.66018176e-01
-1.45489156e-01 -4.30015653e-01 -5.61966002e-01 4.85143721e-01
-4.35965180e-01 6.14751399e-01 8.46628666e-01 -7.98478007e-01
-1.03040969e+00 4.39271450e-01 7.73466378e-02 -1.07546353e+00
-4.81454760e-01 3.23735662e-02 -5.34172118e-01 2.32996911e-01
4.92057651e-01 4.45986450e-01 -1.19137548e-01 -7.22193241e-01
6.45313263e-01 3.31780225e-01 6.89755440e-01 -8.11801732e-01
9.85745966e-01 6.90158904e-01 -1.32700384e-01 -5.56294262e-01
-1.03145063e+00 -6.24738932e-01 -7.11490393e-01 1.50366932e-01
6.92400277e-01 -9.91769135e-01 -6.58698261e-01 5.25674403e-01
-1.03159249e+00 -8.62346053e-01 -7.23753989e-01 3.38267922e-01
-8.21029782e-01 3.17135036e-01 -6.97326660e-01 -9.63986456e-01
-1.57774910e-01 -1.30866337e+00 1.29578507e+00 -8.27211961e-02
-6.55344725e-02 -7.26814032e-01 9.15078148e-02 3.04037094e-01
-4.61152941e-02 -3.07804942e-02 9.47340846e-01 -8.54066849e-01
-8.83011639e-01 -9.96862426e-02 1.06205031e-01 4.22305256e-01
-1.00551685e-02 7.66602904e-02 -1.13328838e+00 -4.38178740e-02
-1.12862319e-01 -7.56323934e-01 1.11590910e+00 2.72056852e-02
1.20575523e+00 -5.31779647e-01 -3.70114982e-01 5.73055029e-01
1.27859533e+00 -2.37694874e-01 3.07476580e-01 -1.10487618e-01
7.90744245e-01 5.30436277e-01 2.30788156e-01 5.11023045e-01
5.76312244e-01 5.70774853e-01 5.91388702e-01 4.41578984e-01
-4.79816020e-01 -7.23194599e-01 5.20753860e-01 6.24896288e-01
4.34418656e-02 -3.47466916e-01 -8.42293561e-01 5.67914486e-01
-2.07089496e+00 -7.35537469e-01 3.91122937e-01 2.09839249e+00
9.14085925e-01 8.54250267e-02 -7.94247389e-02 -2.26613656e-02
5.32970190e-01 2.76278168e-01 -9.84370768e-01 2.23747224e-01
2.82193869e-02 4.71891522e-01 2.72938073e-01 6.21604621e-01
-1.21301842e+00 1.23421824e+00 4.95991135e+00 6.53564811e-01
-1.02105141e+00 3.89090300e-01 2.53126591e-01 -4.44396496e-01
-4.72456157e-01 5.54486334e-01 -8.27840328e-01 4.15667444e-01
8.74240637e-01 5.20897331e-04 7.68506229e-01 9.44674313e-01
-3.14870924e-01 2.19384488e-02 -1.49979687e+00 8.16751599e-01
3.08003843e-01 -1.28889585e+00 1.00843824e-01 -8.08756128e-02
8.63573670e-01 1.79239944e-01 3.79717047e-03 6.01347983e-01
7.16011822e-01 -6.23854697e-01 1.26549268e+00 4.56613451e-01
7.60424137e-01 -1.53159589e-01 2.72411436e-01 8.06521177e-01
-1.24166811e+00 -3.21898013e-01 -2.32153147e-01 -1.82979956e-01
2.67138898e-01 1.87572107e-01 -5.41541040e-01 4.54482764e-01
1.94521904e-01 8.93743753e-01 -4.24731821e-01 8.80029321e-01
-8.60277712e-01 6.47624433e-01 -4.51724201e-01 2.53529131e-01
9.02659744e-02 -2.70419549e-02 5.32303274e-01 7.25547016e-01
9.89483446e-02 1.79857705e-02 4.37523484e-01 1.21672249e+00
-2.38348275e-01 -3.75235170e-01 -2.46073052e-01 -2.13238776e-01
2.75304079e-01 1.15452588e+00 -5.38215876e-01 -7.93840110e-01
-4.54059601e-01 1.12083173e+00 6.35870457e-01 3.16443712e-01
-7.95903444e-01 3.10002059e-01 7.94947863e-01 -2.05107465e-01
5.43463171e-01 -1.15743212e-01 6.44629747e-02 -1.86341047e+00
4.10084575e-02 -7.14486301e-01 3.85758162e-01 -5.63537836e-01
-1.12683463e+00 3.23362470e-01 4.95856926e-02 -1.02798939e+00
-1.18148416e-01 -6.35776222e-01 -6.12891197e-01 5.14014125e-01
-1.56555903e+00 -1.62009811e+00 -9.96665433e-02 6.30175650e-01
8.53637815e-01 8.66806731e-02 8.02132607e-01 -7.89245442e-02
-7.90944159e-01 2.66474515e-01 -3.06383342e-01 -1.17141515e-01
4.52778369e-01 -1.30099165e+00 5.76831698e-01 8.69616151e-01
7.49513507e-01 4.53988969e-01 6.28561318e-01 -6.83415651e-01
-1.67357731e+00 -1.19962871e+00 7.61714101e-01 -6.99849606e-01
7.54904449e-01 -9.06120241e-01 -7.77815163e-01 1.06660450e+00
-1.84748009e-01 6.32754922e-01 5.03716469e-01 -1.31667461e-02
-7.89032280e-01 2.81057447e-01 -7.56954670e-01 6.74925148e-01
1.37467802e+00 -7.44061351e-01 -6.54448986e-01 3.31323355e-01
1.04055417e+00 -3.60585481e-01 -4.84669179e-01 2.04636753e-01
5.65927386e-01 -5.96773624e-01 7.69683182e-01 -8.31831694e-01
4.41298425e-01 -4.66510117e-01 -1.62050024e-01 -1.04781485e+00
-1.88123330e-01 -5.97950101e-01 -9.93272126e-01 8.74283314e-01
5.08238852e-01 -4.62984294e-01 9.35047746e-01 8.68664026e-01
-3.26176465e-01 -8.56112003e-01 -8.27843547e-01 -8.69321406e-01
-1.93520233e-01 -6.10001564e-01 5.86144269e-01 6.24891698e-01
-4.34397385e-02 4.61385250e-01 -5.24678707e-01 4.09219950e-01
7.11955905e-01 4.01574492e-01 9.08707917e-01 -1.09089696e+00
-9.03710842e-01 -2.61280179e-01 -4.21799242e-01 -1.04133010e+00
7.24263728e-01 -1.01724362e+00 2.22337082e-01 -1.53474975e+00
5.41449904e-01 -3.75258058e-01 -4.21691507e-01 9.40426350e-01
-1.67131513e-01 2.29869154e-03 5.26695192e-01 3.39323342e-01
-9.61301029e-01 8.24554443e-01 1.07279277e+00 -3.85704190e-01
-1.00093357e-01 2.66766176e-02 -4.23070788e-01 7.27279961e-01
5.15902042e-01 -4.76216882e-01 -5.24249017e-01 -4.58513558e-01
-1.38624042e-01 -1.32068694e-01 6.75772727e-01 -1.04528236e+00
3.10874075e-01 -2.59635895e-01 -2.23670043e-02 -2.91851819e-01
7.98814356e-01 -8.14906657e-01 3.23320299e-01 4.69911009e-01
-5.26645601e-01 -7.13588953e-01 -3.40609342e-01 8.92199039e-01
1.34058326e-01 -2.77049839e-01 5.47956347e-01 -1.89060718e-01
-8.96129370e-01 8.37653339e-01 -6.68663681e-02 -1.52608320e-01
1.24606872e+00 2.68861447e-02 -4.33904409e-01 -1.82124659e-01
-1.01951718e+00 2.69393533e-01 7.42626607e-01 3.21963191e-01
5.10787904e-01 -9.91079092e-01 -2.64833212e-01 3.19232345e-01
4.88226771e-01 3.91062886e-01 3.92692387e-01 7.93633997e-01
-7.76955336e-02 2.63110995e-01 -4.85596731e-02 -3.76264036e-01
-9.85599339e-01 9.64559853e-01 1.75093517e-01 -2.18365625e-01
-9.36535656e-01 1.01586974e+00 5.64718008e-01 -5.99737704e-01
3.47361386e-01 -4.99327064e-01 1.86303243e-01 -2.03242406e-01
4.15822059e-01 9.74206775e-02 -2.63975799e-01 -6.69384420e-01
-2.13491783e-01 3.46211851e-01 -4.39304337e-02 -2.70526856e-01
1.33710206e+00 1.47358641e-01 5.46051450e-02 7.79563010e-01
1.09358084e+00 -3.26898515e-01 -1.89145172e+00 -6.92686677e-01
-6.08689198e-03 -2.45590761e-01 -1.63370654e-01 -5.61545074e-01
-9.08280373e-01 1.06792080e+00 1.29905447e-01 -2.86339402e-01
6.66094184e-01 5.85117936e-01 6.21650159e-01 5.62331676e-01
5.38391352e-01 -6.83286250e-01 2.17019156e-01 3.57083738e-01
4.45431739e-01 -1.20594609e+00 -1.72731712e-01 -2.92468429e-01
-8.17242444e-01 7.88491905e-01 6.99849606e-01 -2.62129009e-01
4.77906942e-01 3.00763398e-01 -4.20179188e-01 -1.40320450e-01
-1.47711051e+00 -6.39238894e-01 1.69259846e-01 4.84706044e-01
-2.25461379e-01 1.86169371e-01 4.88607109e-01 7.61825144e-01
-1.64319575e-02 1.67087168e-02 2.49828517e-01 1.09680724e+00
-4.43006456e-01 -1.18616450e+00 8.71300548e-02 6.16926253e-01
-9.14874673e-02 -1.22447327e-01 -3.86051416e-01 3.31366420e-01
4.86100852e-01 6.16864681e-01 6.43877462e-02 -3.08337122e-01
5.47434315e-02 4.29753155e-01 6.26412749e-01 -1.17047560e+00
-1.51580676e-01 -6.49220273e-02 8.95239934e-02 -7.26686776e-01
-2.09543243e-01 -5.06076813e-01 -1.17019618e+00 1.22289889e-01
-4.86205995e-01 -9.58590731e-02 5.08567333e-01 1.14893246e+00
2.70164639e-01 3.42194498e-01 3.10953081e-01 -1.08541250e+00
-7.04482436e-01 -8.19649875e-01 -4.21199352e-01 5.13534248e-01
4.98633713e-01 -8.67165089e-01 -3.52134764e-01 4.09377784e-01] | [10.264893531799316, 2.2346768379211426] |
79b4f440-7134-452c-86cc-36be417b22ed | improving-complex-knowledge-base-question | 2212.13036 | null | https://arxiv.org/abs/2212.13036v1 | https://arxiv.org/pdf/2212.13036v1.pdf | Improving Complex Knowledge Base Question Answering via Question-to-Action and Question-to-Question Alignment | Complex knowledge base question answering can be achieved by converting questions into sequences of predefined actions. However, there is a significant semantic and structural gap between natural language and action sequences, which makes this conversion difficult. In this paper, we introduce an alignment-enhanced complex question answering framework, called ALCQA, which mitigates this gap through question-to-action alignment and question-to-question alignment. We train a question rewriting model to align the question and each action, and utilize a pretrained language model to implicitly align the question and KG artifacts. Moreover, considering that similar questions correspond to similar action sequences, we retrieve top-k similar question-answer pairs at the inference stage through question-to-question alignment and propose a novel reward-guided action sequence selection strategy to select from candidate action sequences. We conduct experiments on CQA and WQSP datasets, and the results show that our approach outperforms state-of-the-art methods and obtains a 9.88\% improvements in the F1 metric on CQA dataset. Our source code is available at https://github.com/TTTTTTTTy/ALCQA. | ['Weiming Lu', 'Xiaoxia Cheng', 'Yechun Tang'] | 2022-12-26 | null | null | null | null | ['knowledge-base-question-answering', 'question-rewriting'] | ['natural-language-processing', 'natural-language-processing'] | [ 3.53074253e-01 1.81782544e-02 -6.60614595e-02 -5.47651470e-01
-1.36169434e+00 -8.06064725e-01 5.10862231e-01 -8.16241875e-02
-3.61041397e-01 5.95579743e-01 5.64512372e-01 -3.73831362e-01
-1.87258676e-01 -7.78957665e-01 -6.33019865e-01 -1.64086118e-01
6.54505789e-01 3.73793691e-01 6.14709020e-01 -5.17423630e-01
4.86585766e-01 -1.34725407e-01 -1.44633698e+00 6.17541671e-01
1.34779441e+00 7.58956969e-01 2.59085506e-01 7.57655084e-01
-6.85166121e-01 1.36341488e+00 -5.95077813e-01 -7.28801966e-01
1.13236338e-01 -1.21734536e+00 -1.52980506e+00 -2.67391689e-02
5.10688484e-01 -3.41004461e-01 -3.44980985e-01 8.87771904e-01
3.88737619e-01 3.25998515e-01 1.44854486e-01 -1.15308619e+00
-9.31201279e-01 2.45493844e-01 -2.30051056e-02 2.26284266e-01
1.04199541e+00 4.93737161e-01 1.52735555e+00 -5.98213911e-01
4.85462755e-01 1.26185131e+00 8.18072632e-02 7.43156314e-01
-6.86197519e-01 -3.35722536e-01 2.65672505e-01 7.77436376e-01
-1.09865212e+00 -3.87659758e-01 5.59001982e-01 -1.05975598e-01
1.08786952e+00 4.90804702e-01 4.58065063e-01 7.15148389e-01
9.50226001e-03 1.06270647e+00 9.86580372e-01 -4.25842673e-01
1.23936139e-01 -2.62074322e-01 2.34240323e-01 8.73630762e-01
-3.40861320e-01 -4.82350707e-01 -5.90191782e-01 -2.84089446e-01
2.93250471e-01 -1.08347736e-01 -2.22089157e-01 -7.37619102e-02
-1.08339632e+00 8.49204123e-01 1.02198236e-01 2.45928600e-01
-3.95222306e-01 2.15192825e-01 2.90363252e-01 5.28382838e-01
-1.78771645e-01 7.64396071e-01 -4.31663364e-01 -4.35348302e-01
-2.61526316e-01 6.20205224e-01 9.27121878e-01 9.90269482e-01
7.83879220e-01 -5.20049393e-01 -8.21999669e-01 8.92463148e-01
3.37639660e-01 5.54492176e-01 4.03164297e-01 -1.55310798e+00
7.02564359e-01 9.83365059e-01 4.12928045e-01 -9.02004659e-01
-5.61412498e-02 1.05116293e-01 -2.61290640e-01 -5.44864595e-01
4.44357097e-01 1.24914862e-01 -6.00872934e-01 1.63128448e+00
6.19427979e-01 1.10371307e-01 2.42217362e-01 8.07376444e-01
8.29815924e-01 8.00996959e-01 1.81454048e-01 -2.34275162e-01
1.71346796e+00 -1.50468767e+00 -9.10633266e-01 -2.77583808e-01
7.71547139e-01 -6.66206717e-01 1.66523528e+00 3.28529291e-02
-1.22351229e+00 -4.32806373e-01 -4.77177441e-01 -4.60504025e-01
-3.37105505e-02 3.99817433e-03 3.61606032e-02 2.75643706e-01
-7.06916034e-01 1.13733947e-01 -4.21952009e-01 -3.14886004e-01
2.27346987e-01 -8.58786553e-02 6.09882325e-02 -5.21342278e-01
-1.62873936e+00 8.19680631e-01 2.05995500e-01 -4.32297103e-02
-9.96320367e-01 -5.94353497e-01 -7.04949319e-01 3.11805829e-02
1.01388586e+00 -8.84165287e-01 1.83060455e+00 -7.48929918e-01
-1.80541873e+00 5.85164487e-01 -5.17807901e-01 -2.58378863e-01
2.56592155e-01 -4.75079715e-01 -3.98702860e-01 4.16789562e-01
3.95818889e-01 3.91117066e-01 5.06616771e-01 -7.93724120e-01
-9.01096880e-01 -1.60475835e-01 7.75403976e-01 2.88597256e-01
6.94985688e-02 2.93750226e-01 -7.39872932e-01 -3.43229085e-01
-6.41282499e-02 -8.45909953e-01 -3.98008019e-01 -1.01632260e-01
-6.85725510e-02 -7.55169392e-01 2.26811484e-01 -1.07831836e+00
1.55366409e+00 -1.69537067e+00 3.58353034e-02 -8.71100798e-02
-1.36064947e-01 3.96308273e-01 -6.44024789e-01 8.87838721e-01
3.15836281e-01 3.91407311e-02 -5.20080745e-01 1.68239653e-01
1.63052604e-01 2.91272610e-01 -4.68981415e-01 -1.16313279e-01
2.96693206e-01 1.10538304e+00 -1.24211133e+00 -6.38738275e-01
-1.80265144e-01 -2.10906446e-01 -7.38740981e-01 7.38951445e-01
-9.83264327e-01 4.47921127e-01 -9.61668134e-01 5.79882979e-01
3.74349952e-01 -4.62589681e-01 2.38262817e-01 5.91983236e-02
2.19486728e-01 7.87604451e-01 -8.28188777e-01 2.06131005e+00
-4.86393213e-01 7.95396119e-02 -4.52890992e-01 -9.64357436e-01
9.58415568e-01 3.36295515e-01 2.09760472e-01 -1.28380203e+00
-1.97959468e-01 1.14789858e-01 -5.21031320e-02 -1.21674550e+00
4.83813554e-01 -3.71783189e-02 -2.38360718e-01 5.89001179e-01
-1.05403163e-01 -3.32226664e-01 5.50704658e-01 3.73427153e-01
1.39383388e+00 4.09197301e-01 3.73241991e-01 9.96932834e-02
1.11602104e+00 2.67325670e-01 5.37883461e-01 7.54793644e-01
-4.01035160e-01 3.61575544e-01 6.73577964e-01 -2.14175180e-01
-6.37081563e-01 -9.10986900e-01 5.70661545e-01 1.30545890e+00
1.98108479e-01 -5.50266922e-01 -8.05141866e-01 -9.10778821e-01
-2.32779101e-01 1.06034648e+00 -4.13278937e-01 -3.03712338e-01
-8.53421450e-01 -3.60030979e-02 6.00437045e-01 2.87669688e-01
6.53975844e-01 -1.35351026e+00 -3.26559424e-01 4.37482506e-01
-9.56815362e-01 -1.25670755e+00 -7.85451472e-01 -6.92118347e-01
-6.10147893e-01 -1.45690954e+00 -4.33667541e-01 -7.77697921e-01
3.40309501e-01 3.96569759e-01 1.25998139e+00 3.43344927e-01
1.11702122e-01 6.87859416e-01 -8.02675486e-01 -8.07521120e-02
-5.65106392e-01 5.39506488e-02 -5.33946395e-01 4.35958020e-02
4.98121142e-01 -3.39879155e-01 -9.13642168e-01 4.37327415e-01
-1.00442803e+00 -2.94667366e-03 5.10685146e-01 5.01639068e-01
6.29126310e-01 -5.64780056e-01 9.26619351e-01 -7.17220068e-01
9.26869690e-01 -4.44801331e-01 -5.18040359e-01 9.07273948e-01
-4.67583358e-01 2.05412343e-01 7.88882494e-01 -4.58944067e-02
-1.31002676e+00 -1.24522418e-01 -3.62152100e-01 -4.26609591e-02
-1.40573606e-01 4.72100258e-01 -4.21769261e-01 3.47600847e-01
5.39605975e-01 5.61718702e-01 -1.36697233e-01 -3.63754302e-01
7.30785310e-01 6.07122481e-01 4.31675553e-01 -6.66002810e-01
7.02334285e-01 3.32291722e-01 -3.78142834e-01 -3.03820550e-01
-1.27144897e+00 -6.62521899e-01 -2.79156923e-01 -2.52274156e-01
1.06848121e+00 -6.70530796e-01 -8.60952675e-01 2.44943485e-01
-1.30482447e+00 -2.74450809e-01 -2.79287219e-01 1.44072011e-01
-8.48984063e-01 6.19135857e-01 -4.46237773e-01 -6.93364263e-01
-5.32690883e-01 -9.42972600e-01 6.66306615e-01 4.76666570e-01
-1.45686164e-01 -4.99508649e-01 3.73640597e-01 1.27504647e+00
2.04999223e-01 -1.90107882e-01 9.83225703e-01 -8.72488976e-01
-9.36757743e-01 9.48581398e-02 -5.36469296e-02 4.71543938e-01
3.10997486e-01 -2.76506484e-01 -3.59830856e-01 -9.39852651e-03
9.85019468e-03 -6.39715493e-01 4.93167341e-01 -2.71542221e-01
1.14279151e+00 -5.13143182e-01 2.07723647e-01 -5.28227091e-02
1.13337755e+00 4.17944074e-01 9.27789450e-01 3.84739697e-01
4.79169309e-01 7.12934434e-01 1.19905865e+00 2.75767654e-01
8.74127686e-01 5.87852061e-01 2.29695961e-01 5.12693465e-01
-3.85184772e-02 -5.87186873e-01 4.55529720e-01 1.01365757e+00
2.79665589e-01 -3.38165611e-01 -8.61842573e-01 9.15259480e-01
-1.95383883e+00 -1.01258874e+00 -1.50512680e-01 1.82003665e+00
1.20598686e+00 -2.38676146e-01 3.31852250e-02 -2.56840438e-01
4.83258516e-01 3.64221931e-01 -6.97966576e-01 -3.53525251e-01
2.34124035e-01 3.44805092e-01 -1.45192847e-01 6.11629784e-01
-6.81728125e-01 1.11035311e+00 5.19099236e+00 9.78711963e-01
-3.80608946e-01 1.43022925e-01 1.77403033e-01 7.60724128e-04
-6.21819913e-01 3.19285959e-01 -5.49857259e-01 4.01676089e-01
9.78796065e-01 -4.64607716e-01 4.96115655e-01 4.70843852e-01
3.89216959e-01 -1.29168108e-01 -9.11284745e-01 6.19369090e-01
1.99581757e-01 -1.34349144e+00 5.14555514e-01 -5.24255276e-01
6.25977874e-01 -3.72225523e-01 -3.81495476e-01 6.38248980e-01
5.53245604e-01 -6.82104766e-01 3.19106013e-01 7.84489214e-01
2.06435680e-01 -6.53818071e-01 4.43504035e-01 4.05537099e-01
-1.17434108e+00 -1.84706107e-01 -3.28454673e-01 8.38305205e-02
2.80099839e-01 2.01596562e-02 -5.06533504e-01 1.02540410e+00
6.52248919e-01 3.84690315e-01 -5.19838035e-01 8.28237832e-01
-7.60035038e-01 8.16519797e-01 1.60769597e-01 -3.26892167e-01
3.94226134e-01 -5.05035639e-01 2.23540157e-01 7.62435853e-01
1.87937364e-01 7.97921002e-01 1.14538975e-01 6.81153893e-01
-2.96297580e-01 4.60437626e-01 -5.81989586e-02 -1.76902056e-01
4.86598670e-01 8.54066432e-01 -8.80090520e-02 -4.90629107e-01
-7.86715269e-01 1.14109218e+00 4.09620732e-01 3.83673012e-01
-1.01404560e+00 -6.60446107e-01 7.22631156e-01 -1.08411945e-01
3.89990687e-01 1.08239397e-01 3.53962988e-01 -1.20451832e+00
4.45701569e-01 -1.56796229e+00 7.97065496e-01 -9.73203838e-01
-1.07582915e+00 2.29664490e-01 -2.01641783e-01 -1.15309036e+00
-1.64368212e-01 -1.46291237e-02 -5.88869333e-01 6.46017790e-01
-1.69419706e+00 -8.66778731e-01 -3.12969357e-01 6.72439635e-01
7.30669141e-01 2.42555931e-01 5.56854486e-01 2.29049221e-01
-3.18952948e-01 5.07496834e-01 -3.09414640e-02 2.21324816e-01
6.87569261e-01 -9.09010708e-01 6.96448147e-01 7.67470062e-01
1.67204887e-01 5.31563699e-01 4.52389777e-01 -4.43757504e-01
-1.59902477e+00 -1.14361143e+00 1.21300697e+00 -6.41703606e-01
6.06957376e-01 1.76654626e-02 -1.25387132e+00 6.39264941e-01
4.42264944e-01 -4.38808322e-01 9.39306319e-01 -3.30971330e-01
-4.17763561e-01 -1.09691687e-01 -9.37877536e-01 8.02774191e-01
1.21459746e+00 -8.96193326e-01 -1.10912704e+00 4.69322681e-01
1.29582667e+00 -2.35941783e-01 -8.39634717e-01 2.78385371e-01
3.01354766e-01 -8.49412620e-01 7.97584116e-01 -1.06086457e+00
5.77195704e-01 -5.95221102e-01 -1.85662046e-01 -7.99749076e-01
-2.00168833e-01 -5.97247601e-01 -1.23635225e-01 1.19966400e+00
5.40872514e-01 -5.48963845e-01 6.61652923e-01 6.39397502e-01
-1.55999884e-01 -9.66637433e-01 -9.66333985e-01 -8.45361054e-01
9.15817767e-02 -1.28357500e-01 7.43810952e-01 8.02294791e-01
7.21984659e-04 5.57624757e-01 -3.17608565e-01 1.91502005e-01
1.13817677e-01 6.27981424e-01 8.73577535e-01 -4.75118428e-01
-4.62808281e-01 -1.78970411e-01 2.64960378e-01 -1.70372069e+00
9.12628397e-02 -7.80353189e-01 2.63125092e-01 -2.06502724e+00
1.90638453e-01 6.70767874e-02 -1.59256116e-01 4.10242707e-01
-6.90012932e-01 -3.26528043e-01 3.00550133e-01 6.94968924e-02
-1.16239583e+00 9.98818576e-01 1.65249681e+00 -2.91402228e-02
-7.58019611e-02 -4.77872230e-02 -7.25067258e-01 3.89984727e-01
9.43572283e-01 -5.01505971e-01 -6.77891731e-01 -6.23967409e-01
3.98055404e-01 4.85494524e-01 1.48651391e-01 -7.60556340e-01
2.49533847e-01 -6.62796438e-01 -4.81667101e-01 -5.28834403e-01
1.75112337e-01 -5.21314263e-01 -3.63503218e-01 5.98941863e-01
-7.90646374e-01 1.19826272e-01 -2.28432752e-02 6.74653471e-01
-3.29990119e-01 -5.39953232e-01 4.92356867e-01 -2.31798872e-01
-7.38714397e-01 3.24503690e-01 -3.39345634e-01 6.96637332e-01
1.02963531e+00 1.08587503e-01 -5.07470012e-01 -6.97753251e-01
-3.79470944e-01 8.96351755e-01 1.01008594e-01 6.37021959e-01
5.91449320e-01 -1.27942991e+00 -8.96644950e-01 -4.67830479e-01
4.34928894e-01 -5.20949923e-02 5.61207652e-01 6.10853016e-01
-3.67577732e-01 6.50485873e-01 9.93637592e-02 -1.69964731e-01
-1.32274270e+00 5.25972247e-01 5.05988300e-01 -5.16053736e-01
-1.18761331e-01 7.66022444e-01 3.01804580e-02 -7.99031317e-01
1.51309306e-02 -3.28315586e-01 -3.62369120e-01 -2.21046492e-01
6.30926669e-01 2.79264718e-01 -2.61467129e-01 -2.86845982e-01
-4.15776789e-01 5.72009265e-01 -2.08328426e-01 -9.42205638e-02
9.11701620e-01 -2.55550176e-01 -1.82048440e-01 6.83179051e-02
1.03092062e+00 -2.79735066e-02 -8.46018136e-01 -6.76774859e-01
3.07043254e-01 -6.54792547e-01 -7.65931427e-01 -9.25700724e-01
-5.97466588e-01 7.58239031e-01 2.38690138e-01 1.19560190e-01
1.12519932e+00 3.55854541e-01 1.19782269e+00 9.66950536e-01
1.21450335e-01 -1.12889600e+00 7.93648660e-01 6.52959108e-01
1.21437490e+00 -9.66744602e-01 -2.06148297e-01 -3.69076461e-01
-8.09828222e-01 6.17582083e-01 1.02033460e+00 1.09911198e-02
-6.19633347e-02 -7.03274310e-01 2.85716623e-01 -2.06940159e-01
-1.07237387e+00 -4.35303956e-01 4.09355424e-02 2.57565379e-01
2.81528324e-01 -2.37631664e-01 -7.16774464e-01 5.46777904e-01
-9.76050869e-02 1.69528916e-01 4.65109050e-01 1.23670268e+00
-7.08132386e-01 -1.46654403e+00 -1.34070396e-01 3.02508384e-01
-3.54155958e-01 -1.56792253e-01 -4.62697029e-01 4.69370544e-01
-3.39187026e-01 1.44431496e+00 -2.09370226e-01 -2.84420133e-01
8.30895066e-01 4.65099156e-01 3.51846129e-01 -6.25125885e-01
-7.27691650e-01 -5.16443372e-01 3.99627835e-01 -9.13366616e-01
-4.76665705e-01 -6.48251653e-01 -1.54719555e+00 -6.74592629e-02
-2.20135391e-01 5.23093760e-01 1.27806336e-01 1.17443168e+00
6.36061132e-01 3.66600275e-01 8.82043123e-01 4.82662082e-01
-9.41961050e-01 -7.99868405e-01 5.95173985e-02 5.93374431e-01
2.00642973e-01 -8.02714229e-02 -1.11845128e-01 1.44358486e-01] | [11.455384254455566, 7.9579997062683105] |
11d80b97-6df2-4716-890e-a09ab4febbef | beyond-semantic-to-instance-segmentation | 2109.09477 | null | https://arxiv.org/abs/2109.09477v3 | https://arxiv.org/pdf/2109.09477v3.pdf | Beyond Semantic to Instance Segmentation: Weakly-Supervised Instance Segmentation via Semantic Knowledge Transfer and Self-Refinement | Weakly-supervised instance segmentation (WSIS) has been considered as a more challenging task than weakly-supervised semantic segmentation (WSSS). Compared to WSSS, WSIS requires instance-wise localization, which is difficult to extract from image-level labels. To tackle the problem, most WSIS approaches use off-the-shelf proposal techniques that require pre-training with instance or object level labels, deviating the fundamental definition of the fully-image-level supervised setting. In this paper, we propose a novel approach including two innovative components. First, we propose a semantic knowledge transfer to obtain pseudo instance labels by transferring the knowledge of WSSS to WSIS while eliminating the need for the off-the-shelf proposals. Second, we propose a self-refinement method to refine the pseudo instance labels in a self-supervised scheme and to use the refined labels for training in an online manner. Here, we discover an erroneous phenomenon, semantic drift, that occurred by the missing instances in pseudo instance labels categorized as background class. This semantic drift occurs confusion between background and instance in training and consequently degrades the segmentation performance. We term this problem as semantic drift problem and show that our proposed self-refinement method eliminates the semantic drift problem. The extensive experiments on PASCAL VOC 2012 and MS COCO demonstrate the effectiveness of our approach, and we achieve a considerable performance without off-the-shelf proposal techniques. The code is available at https://github.com/clovaai/BESTIE. | ['Junmo Kim', 'Chaeeun Rhee', 'Youngjoon Yoo', 'Beomyoung Kim'] | 2021-09-20 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Kim_Beyond_Semantic_to_Instance_Segmentation_Weakly-Supervised_Instance_Segmentation_via_Semantic_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Kim_Beyond_Semantic_to_Instance_Segmentation_Weakly-Supervised_Instance_Segmentation_via_Semantic_CVPR_2022_paper.pdf | cvpr-2022-1 | ['weakly-supervised-instance-segmentation', 'point-supervised-instance-segmentation', 'image-level-supervised-instance-segmentation'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 5.40967286e-01 2.82261431e-01 -3.17207217e-01 -5.68623781e-01
-7.99668670e-01 -5.01839936e-01 3.05592179e-01 1.28382882e-02
-5.99146783e-01 8.44210863e-01 -5.17369509e-01 -1.86970979e-02
-3.40172346e-03 -4.93846864e-01 -9.98089373e-01 -9.50428128e-01
3.74311090e-01 5.29821873e-01 8.07948291e-01 1.99353695e-02
1.52140811e-01 -2.46685985e-02 -1.70150530e+00 2.53786594e-01
1.10092449e+00 1.04354095e+00 3.65849167e-01 4.03765261e-01
-6.05756342e-01 4.36389953e-01 -7.17687726e-01 -1.95433676e-01
2.53798634e-01 -4.97873664e-01 -9.45832014e-01 3.79383266e-01
3.93002838e-01 2.16310143e-01 2.52280533e-01 1.39287651e+00
2.59631604e-01 -5.08396104e-02 2.85714418e-01 -1.33287942e+00
-1.86167821e-01 2.76457191e-01 -7.54477561e-01 1.21200860e-01
-1.48534849e-01 5.03892265e-03 5.41198671e-01 -7.12548137e-01
6.59912229e-01 8.53518248e-01 7.31032670e-01 8.52079988e-01
-8.96631062e-01 -7.13950813e-01 5.52453041e-01 2.44985864e-01
-1.53309906e+00 -4.00628924e-01 9.86589849e-01 -3.33800405e-01
2.50469476e-01 2.58812517e-01 5.12775779e-01 8.89250278e-01
-5.17118931e-01 1.23855782e+00 1.36256897e+00 -4.19053525e-01
3.91795009e-01 5.09867966e-01 4.40573782e-01 6.41295195e-01
3.37708920e-01 -2.47042194e-01 -4.20268089e-01 2.12426484e-01
5.05331159e-01 9.19018760e-02 -4.07788754e-01 -4.22433674e-01
-1.09619164e+00 3.29241395e-01 4.11534846e-01 5.62660038e-01
-2.60387715e-02 -8.25141668e-02 2.30294451e-01 1.58139206e-02
8.93696964e-01 3.74395493e-03 -7.88254082e-01 7.85566121e-02
-1.24864793e+00 -4.59946133e-02 5.81767082e-01 1.22496986e+00
9.86335814e-01 -1.69800952e-01 2.62169540e-02 9.54336286e-01
2.19567671e-01 4.74174738e-01 6.29472971e-01 -6.11891329e-01
3.31809819e-01 7.37015486e-01 2.02917576e-01 -4.95380849e-01
-3.18316936e-01 -7.90577769e-01 -5.89562118e-01 -1.41293079e-01
4.55966741e-01 -7.94036593e-03 -1.31155503e+00 1.59494650e+00
8.19849849e-01 7.03689218e-01 7.08475485e-02 9.30492938e-01
8.15872967e-01 6.34226441e-01 1.63872376e-01 -2.63840377e-01
1.15507329e+00 -1.18355620e+00 -8.10219407e-01 -3.56803507e-01
5.95608234e-01 -5.22480845e-01 1.09745216e+00 3.28117132e-01
-8.16638112e-01 -7.44771361e-01 -1.10433757e+00 3.38982940e-01
-5.08609056e-01 2.29346678e-01 3.89692634e-01 5.42218804e-01
-8.85186672e-01 5.98395467e-01 -8.53427172e-01 -3.89333367e-01
6.89668477e-01 4.09894675e-01 -1.66060805e-01 1.01331264e-01
-9.18344378e-01 4.69338149e-01 7.63260543e-01 2.57570356e-01
-7.92349994e-01 -6.43424034e-01 -7.32126534e-01 -3.22419822e-01
9.34880674e-01 -3.23395789e-01 1.26017845e+00 -1.48564553e+00
-1.43015957e+00 1.03034294e+00 -4.43665743e-01 -2.53292859e-01
6.98684871e-01 -1.44768834e-01 -3.81012172e-01 1.33083031e-01
3.46087158e-01 5.95667124e-01 1.12869573e+00 -1.77257502e+00
-9.84433055e-01 -5.30429661e-01 -1.23879708e-01 1.79947630e-01
-3.10753196e-01 -3.61166269e-01 -7.74765670e-01 -5.31813741e-01
4.10417497e-01 -1.01176775e+00 -1.20511465e-01 -2.78512329e-01
-4.05120283e-01 -3.80401999e-01 9.78179038e-01 -4.46491688e-01
1.06421196e+00 -2.34368396e+00 -1.15814835e-01 1.95534248e-02
5.05177416e-02 6.69236243e-01 -1.17575182e-02 -3.05683970e-01
3.09195034e-02 6.34931326e-02 -8.41213524e-01 -5.46079338e-01
-2.67380297e-01 4.64409739e-01 3.65861058e-02 3.33332986e-01
3.98572594e-01 7.78411269e-01 -1.18690503e+00 -7.27421403e-01
1.72374994e-01 1.32075176e-01 -2.15223998e-01 2.39876613e-01
-4.32289511e-01 7.83616722e-01 -4.82779622e-01 8.33499968e-01
9.58778441e-01 -2.54430950e-01 -1.25893861e-01 -2.16218546e-01
-3.47441882e-02 -9.78365541e-02 -1.33495581e+00 1.99957180e+00
-2.45588511e-01 7.33472556e-02 3.36139239e-02 -1.38414121e+00
8.11209381e-01 1.40265748e-01 4.00797367e-01 -5.69924951e-01
4.44946587e-02 5.37299752e-01 -2.93684423e-01 -5.15473127e-01
2.29456872e-01 -2.26515368e-01 9.81193185e-02 -1.44717861e-02
3.23522687e-01 -1.29045159e-01 2.89479226e-01 1.19532175e-01
7.30836749e-01 7.30211496e-01 -5.32151840e-04 -3.37176979e-01
8.38633895e-01 3.56795222e-01 9.92607951e-01 6.57158852e-01
-3.94208074e-01 6.87443197e-01 1.35520771e-01 -1.58145785e-01
-6.04578137e-01 -9.33021605e-01 -2.54854888e-01 8.66618574e-01
6.51819468e-01 -2.13923022e-01 -1.24355388e+00 -1.20731437e+00
-2.06994861e-01 4.63276058e-01 -4.04013425e-01 -1.01855151e-01
-4.40629512e-01 -9.83097255e-01 2.86852777e-01 3.89113784e-01
9.26056802e-01 -9.77456808e-01 -3.34309816e-01 1.94762647e-01
-3.13367188e-01 -1.24913990e+00 -3.83896381e-01 3.32265079e-01
-9.35680151e-01 -1.06865525e+00 -8.93520892e-01 -9.03216064e-01
1.07162488e+00 3.18863541e-01 8.08982670e-01 1.65267825e-01
-3.26922864e-01 3.14765006e-01 -5.06841063e-01 -5.16796350e-01
-2.53865391e-01 1.77690804e-01 -1.57898143e-01 3.15759540e-01
3.89415205e-01 -1.79192156e-01 -7.55432427e-01 4.89940047e-01
-9.64831471e-01 1.72867492e-01 5.03979266e-01 7.94760704e-01
9.89212215e-01 1.76635236e-01 7.63945520e-01 -1.39630413e+00
-5.85758165e-02 -4.39182132e-01 -5.64146519e-01 2.51107544e-01
-6.81210816e-01 -3.87192331e-02 3.85623515e-01 -3.41529489e-01
-1.37350059e+00 4.00487334e-01 -1.61123425e-01 -2.67036170e-01
-5.70998967e-01 1.71590284e-01 -3.00366670e-01 -1.53670445e-01
4.67037708e-01 3.45001191e-01 -2.19237477e-01 -5.32858610e-01
2.40875274e-01 8.76891911e-01 5.27533770e-01 -6.49957716e-01
7.61270881e-01 6.84304535e-01 -4.44259405e-01 -7.65250683e-01
-1.36667681e+00 -9.21684861e-01 -7.63725996e-01 -2.65759230e-01
8.00084531e-01 -9.45956290e-01 -2.78353095e-01 6.72795951e-01
-1.05528903e+00 -4.47061658e-01 -5.04478931e-01 1.85327277e-01
-4.80695873e-01 3.57687682e-01 -3.06636930e-01 -8.62163901e-01
-2.19693616e-01 -1.04509938e+00 1.18316138e+00 4.87046838e-01
1.34506240e-01 -7.95327008e-01 -2.08417952e-01 8.04010153e-01
1.15843661e-01 1.53568238e-01 2.79492438e-01 -7.30895102e-01
-6.29880428e-01 -8.39031264e-02 -4.42221761e-01 6.68859184e-01
1.89424545e-01 -3.35706115e-01 -1.19381511e+00 -2.10008815e-01
1.92335308e-01 -1.62033439e-01 9.39515948e-01 2.70878494e-01
1.23006296e+00 -3.18110622e-02 -4.05655533e-01 6.85275912e-01
1.50841904e+00 2.94865519e-01 2.61989057e-01 4.58244801e-01
9.80290949e-01 6.57440722e-01 1.03686261e+00 1.72577694e-01
4.17727798e-01 4.38330084e-01 2.21237555e-01 -3.86605322e-01
-2.78807431e-01 -1.12703055e-01 1.06379740e-01 8.99655998e-01
-8.26088041e-02 -6.76240306e-03 -8.35910857e-01 6.96639180e-01
-2.21036315e+00 -4.27097410e-01 -2.85175949e-01 2.23307300e+00
1.14043415e+00 3.70454699e-01 -4.19114381e-02 2.37782508e-01
9.05312359e-01 -6.10792749e-02 -7.32951462e-01 1.17536351e-01
1.06587820e-01 2.48157740e-01 7.02857554e-01 4.39195335e-01
-1.22701037e+00 1.35275483e+00 4.55180979e+00 1.00702250e+00
-9.89048898e-01 6.63604796e-01 7.22007930e-01 9.68787745e-02
5.12885265e-02 -7.83279166e-03 -1.03637755e+00 7.00646818e-01
5.93735099e-01 2.49406308e-01 7.73743168e-02 8.61083388e-01
7.11327344e-02 -2.81023502e-01 -9.03822601e-01 9.61006522e-01
1.87262416e-01 -1.00258839e+00 -1.97096005e-01 -3.71970892e-01
8.96131575e-01 -1.26930818e-01 -2.92121708e-01 2.74565965e-01
-7.54056545e-03 -3.92606527e-01 8.68030369e-01 3.05055112e-01
6.28943801e-01 -5.11943698e-01 9.02126253e-01 4.96045530e-01
-1.09163201e+00 9.03805867e-02 -2.42521346e-01 1.61307231e-01
-2.82648932e-02 8.14323425e-01 -6.32695019e-01 7.66519845e-01
8.44488621e-01 7.94487596e-01 -5.89762330e-01 1.06734610e+00
-4.85345662e-01 8.19286585e-01 -2.66818494e-01 1.98176622e-01
3.28871965e-01 -1.71030700e-01 3.48522246e-01 1.23972023e+00
6.46726340e-02 -1.15610668e-02 3.48184347e-01 7.50780046e-01
3.01831346e-02 1.46840895e-02 -1.64537355e-01 2.61112630e-01
2.10114419e-01 1.25564444e+00 -1.16521287e+00 -6.77296340e-01
-1.69686735e-01 1.35517120e+00 3.05453129e-02 5.31887293e-01
-9.67194498e-01 -2.82318830e-01 1.50778517e-01 1.56218976e-01
2.63183266e-01 1.38489930e-02 -3.21917713e-01 -1.31550753e+00
4.06328231e-01 -5.00976980e-01 3.48656178e-01 -5.48293889e-01
-1.05737829e+00 5.68724692e-01 2.64049601e-02 -1.16823411e+00
1.85749620e-01 -4.10125196e-01 -3.66702467e-01 5.44577599e-01
-2.10806108e+00 -1.15152383e+00 -7.03968346e-01 6.47763312e-01
8.76965344e-01 2.64838040e-01 2.95402706e-01 5.22637844e-01
-7.25874364e-01 4.90725011e-01 -3.00557557e-02 -3.53929168e-03
8.32067072e-01 -1.34337354e+00 2.72190180e-02 8.70261192e-01
1.61281556e-01 2.26597920e-01 4.58614767e-01 -6.84617043e-01
-7.52706230e-01 -1.33798146e+00 5.74865520e-01 -2.76764899e-01
4.50165987e-01 -4.21157926e-01 -1.07146668e+00 4.55661148e-01
-1.41608760e-01 3.08766216e-01 4.94896472e-01 -1.35177121e-01
5.92266656e-02 -4.32041407e-01 -1.22832704e+00 2.57866293e-01
1.40448940e+00 -2.24635735e-01 -5.72412610e-01 5.40082097e-01
8.72749329e-01 -4.45942461e-01 -5.53924143e-01 7.57201970e-01
4.12641242e-02 -7.25322843e-01 8.14963758e-01 -7.63507411e-02
7.31398165e-02 -8.42590749e-01 2.77648449e-01 -1.12949550e+00
9.72326547e-02 -3.71638060e-01 5.85760660e-02 1.58186293e+00
4.21586990e-01 -7.03954577e-01 1.10239124e+00 3.43250811e-01
-3.06362033e-01 -5.87488651e-01 -8.84792328e-01 -9.96428370e-01
-2.71841705e-01 -3.74380887e-01 4.91356760e-01 1.09193695e+00
-3.89722645e-01 4.58977073e-02 -3.02167274e-02 3.28804404e-01
8.05411696e-01 4.50862087e-02 6.93996012e-01 -1.14396226e+00
-1.14686899e-01 -1.39804780e-01 -4.38318580e-01 -9.56824124e-01
1.84176490e-01 -8.35056007e-01 5.70402265e-01 -1.40058756e+00
1.01864457e-01 -8.74619365e-01 -6.56697631e-01 6.11518860e-01
-4.88302767e-01 4.21082586e-01 1.23848513e-01 3.82788450e-01
-1.04255235e+00 4.99627680e-01 1.25622118e+00 -3.09501767e-01
-2.64123887e-01 1.10151447e-01 -5.96774518e-01 9.60923314e-01
8.84582698e-01 -7.02902555e-01 -5.76477230e-01 -3.13312620e-01
-2.08401009e-01 -3.44358623e-01 2.89289385e-01 -1.16752803e+00
2.40124315e-01 4.05714773e-02 1.99263513e-01 -5.87286949e-01
1.89082623e-02 -8.64430726e-01 -1.14516102e-01 3.54694068e-01
-8.74369219e-03 -7.18286872e-01 2.54049569e-01 7.07467318e-01
-4.61690187e-01 -6.44091249e-01 9.41717207e-01 -3.24715793e-01
-1.11567819e+00 3.29802036e-01 1.63712651e-01 1.71096221e-01
1.28277528e+00 -3.19906205e-01 -6.44425154e-02 3.10939491e-01
-8.25342596e-01 3.82996410e-01 3.71225178e-01 3.98815215e-01
3.45341831e-01 -8.84028375e-01 -4.38236833e-01 1.61871925e-01
3.68729979e-01 7.83442855e-01 2.99982280e-01 8.00143003e-01
-3.44166189e-01 2.07927953e-02 2.88899571e-01 -1.03774583e+00
-1.17409599e+00 4.42716748e-01 1.13599032e-01 -1.28382400e-01
-4.00148481e-01 1.00844145e+00 3.29517841e-01 -5.72221279e-01
2.79505998e-01 -3.37345809e-01 -9.01459455e-02 3.43086361e-03
3.36064607e-01 9.60909203e-02 1.51743248e-01 -4.77769583e-01
-4.24764067e-01 7.07796216e-01 -9.21281353e-02 1.03926711e-01
1.16667128e+00 -3.18978548e-01 -7.63057023e-02 6.80470645e-01
9.75207746e-01 -3.71964395e-01 -1.50091672e+00 -4.31519359e-01
4.07359660e-01 -3.91948700e-01 -2.99243294e-02 -8.80289853e-01
-1.14840364e+00 7.77094781e-01 8.99197757e-01 -1.26321977e-02
1.16325140e+00 1.30382180e-01 9.82499123e-01 9.46666896e-02
4.66812491e-01 -1.47864997e+00 -8.16452429e-02 1.97790578e-01
3.88246953e-01 -1.61481869e+00 -2.32977167e-01 -8.34713042e-01
-5.81768632e-01 7.09648371e-01 8.56660008e-01 4.34035175e-02
6.78026319e-01 1.56155616e-01 2.14017749e-01 -3.39373830e-03
-1.75597474e-01 -5.30099452e-01 2.02582955e-01 4.69169497e-01
1.85189098e-01 4.05532792e-02 -4.81953114e-01 6.84016287e-01
2.72642314e-01 2.73570627e-01 2.00010031e-01 1.10816932e+00
-4.78494048e-01 -1.14979768e+00 -3.55614156e-01 1.57248169e-01
-3.27932745e-01 -1.24772340e-02 -1.66125417e-01 5.82425296e-01
7.28966355e-01 9.39309061e-01 -1.55770071e-02 -1.87058300e-01
4.05359387e-01 3.20025891e-01 3.52083504e-01 -7.38578856e-01
-4.08302397e-01 2.06561640e-01 -1.37421831e-01 -5.07612526e-01
-9.73282814e-01 -5.12392163e-01 -1.54357445e+00 3.59156191e-01
-6.47534668e-01 3.36551189e-01 9.94100392e-01 1.06473124e+00
2.99117446e-01 7.15561211e-01 5.81860423e-01 -7.75377214e-01
-1.41030177e-01 -8.00450444e-01 -4.49551582e-01 7.85522938e-01
3.40615630e-01 -7.09141493e-01 -4.89235014e-01 3.93916041e-01] | [9.522485733032227, 0.9647348523139954] |
3cae823c-4d3d-4f46-9e2b-4b59861cdd7a | artificial-error-generation-with-machine | 1707.05236 | null | http://arxiv.org/abs/1707.05236v1 | http://arxiv.org/pdf/1707.05236v1.pdf | Artificial Error Generation with Machine Translation and Syntactic Patterns | Shortage of available training data is holding back progress in the area of
automated error detection. This paper investigates two alternative methods for
artificially generating writing errors, in order to create additional
resources. We propose treating error generation as a machine translation task,
where grammatically correct text is translated to contain errors. In addition,
we explore a system for extracting textual patterns from an annotated corpus,
which can then be used to insert errors into grammatically correct sentences.
Our experiments show that the inclusion of artificially generated errors
significantly improves error detection accuracy on both FCE and CoNLL 2014
datasets. | ['Zheng Yuan', 'Ted Briscoe', 'Mariano Felice', 'Marek Rei'] | 2017-07-17 | artificial-error-generation-with-machine-1 | https://aclanthology.org/W17-5032 | https://aclanthology.org/W17-5032.pdf | ws-2017-9 | ['grammatical-error-detection'] | ['natural-language-processing'] | [ 6.38300776e-01 3.77451748e-01 1.69864327e-01 -4.19394881e-01
-7.32474089e-01 -4.66770738e-01 4.72682178e-01 3.31797928e-01
-5.59264898e-01 1.16875267e+00 2.45378092e-01 -7.05648601e-01
3.93261522e-01 -6.33908033e-01 -9.53377962e-01 2.99006313e-01
5.33882022e-01 5.64461708e-01 -1.18904293e-01 -3.35059255e-01
8.12389195e-01 -6.25247881e-02 -1.36076081e+00 5.08149803e-01
1.66122329e+00 2.49007046e-01 2.71080703e-01 7.47652531e-01
-4.68703032e-01 8.35238039e-01 -1.11222506e+00 -8.69121492e-01
-7.91291296e-02 -9.40030456e-01 -1.24123573e+00 5.98287918e-02
2.48570681e-01 -8.58844146e-02 3.18710923e-01 1.15188444e+00
3.64850521e-01 -2.15964794e-01 6.18867218e-01 -9.25101399e-01
-9.43310678e-01 8.79753828e-01 2.44307742e-01 3.41132402e-01
7.36192524e-01 -8.58614519e-02 6.69461966e-01 -1.25050557e+00
9.63816226e-01 8.28806758e-01 6.01492465e-01 8.71895432e-01
-9.45841610e-01 -3.77002537e-01 -5.40719151e-01 2.70055830e-01
-1.06083226e+00 -5.12331963e-01 5.27772129e-01 -3.19205314e-01
1.73746181e+00 5.01700938e-01 3.34281564e-01 1.23873425e+00
2.61718959e-01 6.82804167e-01 1.08622944e+00 -1.47808158e+00
1.92920901e-02 2.73888052e-01 9.61994827e-02 8.15986097e-01
5.34240246e-01 4.76989374e-02 -5.37054896e-01 -2.88791023e-03
1.37951016e-01 -5.06266654e-01 -4.49017406e-01 7.89764822e-01
-9.47879672e-01 6.30079865e-01 -1.78704828e-01 5.91633797e-01
-1.59617707e-01 -2.27433771e-01 4.60350066e-01 8.16315353e-01
8.01855564e-01 1.23185396e+00 -7.32518613e-01 -5.23984551e-01
-9.96992826e-01 2.75479317e-01 9.60961759e-01 1.23456907e+00
6.00347877e-01 -1.74918056e-01 -1.85945094e-01 8.85473371e-01
1.74793914e-01 4.43653464e-01 9.57095981e-01 -6.16702139e-01
9.99510586e-01 8.68742645e-01 3.07988971e-01 -8.97230625e-01
-1.31419227e-01 -1.89223513e-01 -3.48615766e-01 -2.47908190e-01
2.87921846e-01 -2.23976925e-01 -6.64770365e-01 1.25428629e+00
-1.00685269e-01 -4.93974715e-01 1.06676631e-01 4.65426505e-01
4.98865932e-01 5.18767297e-01 -1.78482190e-01 -2.65027940e-01
8.64515066e-01 -1.06952834e+00 -1.08267581e+00 -3.90255153e-01
1.44952309e+00 -1.10191953e+00 1.28706133e+00 4.09827352e-01
-1.17364717e+00 -3.25175703e-01 -8.49377036e-01 -1.37635857e-01
-4.91787016e-01 5.99468172e-01 1.17297113e-01 4.75643337e-01
-7.47764647e-01 8.32608819e-01 -4.96730983e-01 -2.06588849e-01
1.84192479e-01 -1.99449748e-01 -2.55315572e-01 -8.72026160e-02
-1.24465489e+00 1.34834456e+00 5.37656426e-01 -2.66324934e-02
-2.73079798e-02 -5.30352890e-01 -1.03637528e+00 -3.94023478e-01
7.55562857e-02 -5.97777009e-01 1.48198056e+00 -1.33848345e+00
-1.23893356e+00 9.40197647e-01 -5.10583222e-01 -4.26350057e-01
6.04863286e-01 -2.97365606e-01 -6.79898620e-01 -3.98186445e-01
1.84879974e-01 2.81830043e-01 4.65776712e-01 -8.75107467e-01
-8.03937674e-01 -3.16556156e-01 -5.13344228e-01 6.10726662e-02
-3.06234449e-01 6.18897736e-01 1.03317900e-02 -1.02737570e+00
-5.54886609e-02 -7.27241158e-01 2.19247639e-02 -4.87520903e-01
-4.59385216e-01 -5.70322812e-01 1.95429757e-01 -1.33237970e+00
1.87147856e+00 -1.66714609e+00 2.49307796e-01 1.41675413e-01
-2.72946477e-01 2.59386927e-01 -6.24036863e-02 3.79819483e-01
1.41689435e-01 8.29245508e-01 -3.83111984e-01 -5.01963258e-01
-3.64573225e-02 3.91625047e-01 -2.51392663e-01 -2.79076755e-01
6.13768995e-01 9.67083633e-01 -9.91298497e-01 -3.06571424e-01
-2.96711534e-01 -1.78184107e-01 -4.89211351e-01 2.87230223e-01
-2.58833975e-01 2.95987248e-01 -3.30844447e-02 5.84555864e-01
2.11945549e-01 1.07145295e-01 -4.57337089e-02 5.37975192e-01
-1.35377795e-01 1.20645630e+00 -7.45498776e-01 1.42582536e+00
-6.39125347e-01 8.91942978e-01 -6.88044965e-01 -6.07414603e-01
9.82519805e-01 2.34615311e-01 -4.14846599e-01 -7.94124663e-01
1.61700472e-01 7.67896056e-01 -4.40223105e-02 -7.48953760e-01
9.57565069e-01 8.60199239e-03 -1.52246594e-01 4.13751543e-01
2.13302806e-01 -1.85353413e-01 6.87324047e-01 -3.03116255e-03
1.26494694e+00 1.41671196e-01 2.14716166e-01 6.48610964e-02
3.76251489e-01 5.04348040e-01 4.21146572e-01 7.91792035e-01
1.49357468e-01 4.84213799e-01 1.16330925e-02 -2.78280169e-01
-1.45292473e+00 -2.99946934e-01 -5.97422197e-02 7.20597267e-01
-3.97734612e-01 -7.28362560e-01 -9.53197598e-01 -9.47427213e-01
-1.86651766e-01 1.16739452e+00 -4.19135839e-01 -2.38088429e-01
-7.17103899e-01 -4.02134269e-01 9.10074294e-01 4.86093879e-01
2.29461432e-01 -1.43869448e+00 -5.18327713e-01 4.78499204e-01
-7.81455815e-01 -8.27835619e-01 -4.16488707e-01 1.49549037e-01
-8.11497450e-01 -1.07909429e+00 -2.04833791e-01 -9.01420414e-01
9.53749299e-01 -3.53244871e-01 1.38215327e+00 7.67293453e-01
-3.46603453e-01 -1.60479337e-01 -9.88394976e-01 -7.85332143e-01
-1.27535176e+00 1.06187209e-01 -5.25977761e-02 -7.39410818e-01
7.91695058e-01 1.26333013e-02 2.64173329e-01 4.29524481e-02
-8.02715659e-01 1.55966118e-01 4.29050356e-01 1.04260957e+00
2.46198982e-01 -1.39433146e-01 5.25226355e-01 -1.19523776e+00
1.05609822e+00 -4.09670949e-01 -2.27657765e-01 4.55218554e-01
-9.48025227e-01 2.59958029e-01 6.93072617e-01 -1.54251739e-01
-1.05458999e+00 -3.62421811e-01 -4.61622089e-01 1.25618830e-01
-2.30642527e-01 7.18912423e-01 2.37731546e-01 6.63620383e-02
9.95506465e-01 3.45063955e-01 -3.41973692e-01 -5.68903267e-01
7.12652430e-02 1.19882905e+00 2.80749500e-01 -4.73734409e-01
2.14575812e-01 -5.04843950e-01 -4.89152998e-01 -5.07274270e-01
-5.73445618e-01 -1.08218513e-01 -5.65766454e-01 -9.20157284e-02
3.06745321e-01 -4.17167306e-01 3.47372383e-01 4.00540888e-01
-1.68588543e+00 -3.36005658e-01 -4.22495544e-01 2.60439157e-01
-4.28678155e-01 4.86049145e-01 -6.81959212e-01 -6.21381104e-01
-4.83751982e-01 -7.72197664e-01 9.01736498e-01 -3.00034713e-02
-9.04101491e-01 -9.09667969e-01 1.68665260e-01 4.58580703e-01
2.88147181e-01 -2.48455346e-01 9.19239581e-01 -8.20672095e-01
1.21949827e-02 -2.01546311e-01 2.33960927e-01 6.03613019e-01
1.99003462e-02 1.15788005e-01 -2.56869346e-01 1.01573326e-01
-1.42840907e-01 -4.05828416e-01 4.04986233e-01 -3.80603373e-01
1.10280097e+00 -8.23517799e-01 -1.79186210e-01 1.45576969e-01
1.06073940e+00 2.48386040e-01 8.57912302e-01 4.84159917e-01
3.54026079e-01 7.03263700e-01 7.82172143e-01 4.11169261e-01
2.72894919e-01 6.16890013e-01 -1.48503676e-01 2.94481486e-01
-1.75339237e-01 -5.58158994e-01 3.55081320e-01 1.27584338e+00
5.21807298e-02 -6.89114094e-01 -1.16430950e+00 8.02583754e-01
-1.56097901e+00 -8.60480070e-01 -6.75087810e-01 1.80519962e+00
1.41435695e+00 7.64874965e-02 -4.25728559e-01 4.78349119e-01
6.11749411e-01 -7.15585947e-01 7.90586248e-02 -9.54299927e-01
-7.58057162e-02 7.23845184e-01 3.70707035e-01 5.76265335e-01
-6.81403935e-01 1.01342630e+00 6.93572521e+00 5.57122409e-01
-7.39386559e-01 3.18530738e-01 2.98359483e-01 2.31019944e-01
-5.82314134e-01 4.65614907e-02 -8.49957407e-01 1.02896047e+00
1.06981277e+00 -1.12838686e-01 3.45714778e-01 4.79695827e-01
1.77513883e-01 -1.20452732e-01 -1.03163588e+00 5.89941621e-01
4.81600821e-01 -1.25416017e+00 1.62077710e-01 -3.17634970e-01
7.77029395e-01 -2.76039630e-01 -4.96018946e-01 4.87877458e-01
4.23848003e-01 -1.00862527e+00 7.78807223e-01 7.36400545e-01
6.34392262e-01 -6.23555899e-01 1.04175913e+00 7.80231595e-01
-2.94196516e-01 1.72977611e-01 -5.16179681e-01 -5.27760029e-01
-5.97518403e-04 7.00013816e-01 -9.56862152e-01 5.72876751e-01
6.15438521e-01 4.93598878e-01 -1.00805712e+00 8.76635730e-01
-8.26241553e-01 9.17311013e-01 -1.75146893e-01 -4.05988663e-01
-1.84202462e-01 7.00338511e-03 4.22975183e-01 1.49723434e+00
7.15215743e-01 -7.03955293e-02 -5.95955253e-02 1.03004003e+00
-3.05992424e-01 5.10347605e-01 -6.56642199e-01 -2.98606098e-01
6.52898610e-01 6.82058156e-01 -9.27981362e-02 -5.10474086e-01
-1.98986456e-01 1.63703632e+00 6.55604839e-01 4.65548970e-02
-4.75134462e-01 -7.70661056e-01 1.31251216e-01 -3.75534408e-02
5.61337778e-03 -8.58506933e-02 -5.97828329e-01 -1.39573467e+00
6.11246347e-01 -1.23653221e+00 -3.79848629e-02 -1.04630268e+00
-1.14249396e+00 6.06516719e-01 -6.10231102e-01 -9.82495129e-01
-4.06112641e-01 -6.13452017e-01 -5.47691524e-01 1.16068959e+00
-1.25494528e+00 -7.07280576e-01 -2.16476575e-01 -5.49915135e-02
1.05193830e+00 -1.73539564e-01 8.62540722e-01 5.16646206e-01
-5.39136350e-01 9.46662843e-01 9.82861035e-03 4.03631538e-01
8.94735694e-01 -1.25250721e+00 6.75147593e-01 1.24841845e+00
2.22765177e-01 6.44284666e-01 8.06476295e-01 -1.16604340e+00
-7.95212567e-01 -1.37383258e+00 2.14845920e+00 -8.38313878e-01
5.11510909e-01 -1.43441603e-01 -1.01024675e+00 6.48759782e-01
2.27123290e-01 -3.14229518e-01 6.63346291e-01 -1.55105829e-01
-9.23767239e-02 7.17235982e-01 -1.12253201e+00 4.66675013e-01
1.20643926e+00 -4.95390862e-01 -1.16960037e+00 4.37024534e-01
6.51263356e-01 -6.52089238e-01 -6.68403685e-01 3.17454308e-01
1.33470356e-01 -4.38171715e-01 5.59648089e-02 -1.06399310e+00
1.23868704e+00 -1.23945393e-01 7.94371590e-02 -1.91662097e+00
-2.32874248e-02 -4.94087547e-01 -1.42285705e-01 1.38793600e+00
1.00000381e+00 -3.92184883e-01 2.66083568e-01 5.31785131e-01
-5.26422024e-01 -6.49450302e-01 -7.30934799e-01 -8.49239171e-01
2.40863994e-01 -3.91510993e-01 4.92401004e-01 1.15635347e+00
5.17502546e-01 2.26195872e-01 -2.37009093e-01 -3.65548968e-01
-2.36390978e-02 -4.31406230e-01 6.49870038e-01 -8.33194911e-01
-3.37398738e-01 -2.78113008e-01 -3.44412357e-01 -8.55600119e-01
2.27566987e-01 -1.08903873e+00 4.66061592e-01 -1.41278076e+00
-3.92077006e-02 -4.86188650e-01 4.27214980e-01 5.70837498e-01
-5.23563862e-01 2.30689570e-01 -1.18155172e-02 1.90338776e-01
-2.01714754e-01 3.13561112e-01 7.96060741e-01 -2.69760303e-02
-7.52102211e-02 -1.96904391e-01 -5.21600425e-01 5.55378139e-01
1.08302748e+00 -8.64172101e-01 1.02745801e-01 -6.91294789e-01
7.52551854e-01 -2.35416085e-01 -2.82532908e-02 -9.39769626e-01
2.05402300e-02 -4.89734001e-02 2.29684576e-01 -2.69172192e-01
-6.66846514e-01 -6.36179924e-01 7.97729939e-02 3.86149585e-01
-7.72806048e-01 6.46364570e-01 1.86669275e-01 2.45902017e-01
-3.89077783e-01 -1.07645285e+00 3.48399580e-01 -2.88098603e-01
-3.88131946e-01 -5.28182328e-01 -6.95414424e-01 2.52563268e-01
7.10185707e-01 -1.89487487e-01 -3.95960271e-01 4.82063666e-02
-4.92576659e-01 1.20199490e-02 6.70552313e-01 5.66310823e-01
4.88301277e-01 -1.14495718e+00 -9.32248831e-01 4.05883849e-01
4.04662967e-01 -3.22031438e-01 -4.26134557e-01 5.98363638e-01
-7.63984859e-01 3.47341448e-01 -7.19953403e-02 -2.08717093e-01
-1.41202688e+00 2.21178666e-01 3.61665606e-01 -4.15559709e-01
-1.91994339e-01 8.95236790e-01 -1.13648283e+00 -8.89293492e-01
2.50269249e-02 -4.32881773e-01 -2.25395747e-02 -4.17711884e-01
5.49078226e-01 4.89409029e-01 7.28636444e-01 -2.55512059e-01
-1.58103947e-02 -7.79454187e-02 -1.14208505e-01 -1.81743041e-01
1.02135742e+00 -1.80208355e-01 -3.81525397e-01 4.22382057e-01
7.91944504e-01 2.58299530e-01 -2.52788126e-01 -1.31019995e-01
6.92535818e-01 -6.33711815e-01 -2.16514781e-01 -1.36882746e+00
-2.40971699e-01 6.38223529e-01 2.54288137e-01 1.56914681e-01
8.69865596e-01 -1.54804125e-01 7.78442502e-01 6.67365611e-01
5.48698902e-01 -1.63926578e+00 -1.48218125e-01 1.05419815e+00
9.45393443e-01 -1.33563840e+00 -4.89547044e-01 -5.28716922e-01
-3.23929816e-01 1.28615832e+00 8.01898301e-01 2.79133134e-02
3.44624370e-02 2.12025046e-01 -1.00883968e-01 -7.70648643e-02
-6.96379066e-01 1.70890167e-01 2.44912192e-01 3.63714159e-01
8.97964835e-01 -4.48500328e-02 -1.29525518e+00 5.88500500e-01
-5.83011270e-01 2.83422619e-01 9.19357896e-01 1.26701581e+00
-4.98329490e-01 -1.52194262e+00 -1.96195588e-01 7.73486257e-01
-4.94411886e-01 -5.08137882e-01 -9.61334109e-01 4.67018843e-01
4.63036716e-01 1.08754671e+00 2.05285341e-01 -2.34888926e-01
5.26242137e-01 6.98597729e-01 7.43069112e-01 -1.03686023e+00
-8.15106452e-01 -6.89402044e-01 6.28012598e-01 -2.66689569e-01
-1.07415415e-01 -5.93859375e-01 -1.14568591e+00 -2.21670672e-01
-7.54443347e-01 3.02357882e-01 8.35280836e-01 1.25192368e+00
5.67941010e-01 4.80105132e-01 3.69221300e-01 -7.78877512e-02
-7.01405406e-01 -1.46913218e+00 2.82288138e-02 7.80913115e-01
-7.54960179e-02 -2.76501894e-01 -4.93558347e-01 3.14327329e-01] | [11.147491455078125, 10.567237854003906] |
b745ce22-0b0a-4de7-9ea1-686b8f1368cc | learning-to-rank-with-bert-in-tf-ranking | 2004.08476 | null | https://arxiv.org/abs/2004.08476v3 | https://arxiv.org/pdf/2004.08476v3.pdf | Learning-to-Rank with BERT in TF-Ranking | This paper describes a machine learning algorithm for document (re)ranking, in which queries and documents are firstly encoded using BERT [1], and on top of that a learning-to-rank (LTR) model constructed with TF-Ranking (TFR) [2] is applied to further optimize the ranking performance. This approach is proved to be effective in a public MS MARCO benchmark [3]. Our first two submissions achieve the best performance for the passage re-ranking task [4], and the second best performance for the passage full-ranking task as of April 10, 2020 [5]. To leverage the lately development of pre-trained language models, we recently integrate RoBERTa [6] and ELECTRA [7]. Our latest submissions improve our previously state-of-the-art re-ranking performance by 4.3% [8], and achieve the third best performance for the full-ranking task [9] as of June 8, 2020. Both of them demonstrate the effectiveness of combining ranking losses with BERT representations for document ranking. | ['Xuanhui Wang', 'Marc Najork', 'Shuguang Han', 'Mike Bendersky'] | 2020-04-17 | null | null | null | null | ['passage-re-ranking'] | ['natural-language-processing'] | [-2.32025638e-01 -1.66222766e-01 -4.14521575e-01 -4.61402088e-01
-1.82234514e+00 -7.14135408e-01 1.31538868e+00 6.00424290e-01
-5.69607854e-01 8.91052246e-01 5.84236503e-01 -5.49381152e-02
-5.88079214e-01 -5.90466976e-01 -6.51871562e-01 -1.33011907e-01
-6.89632177e-01 9.67600524e-01 4.39936042e-01 -7.80496001e-01
7.66085267e-01 3.57728839e-01 -1.73082113e+00 9.39665914e-01
8.15433741e-01 9.90422785e-01 -2.63271090e-02 1.14377022e+00
-2.70498246e-02 8.61665487e-01 -8.09297025e-01 -4.87273961e-01
-3.12847346e-02 2.76514024e-01 -1.29355156e+00 -9.54837263e-01
6.48458898e-01 -4.22983825e-01 -5.55013597e-01 5.10145664e-01
7.35778093e-01 5.01502216e-01 9.67487574e-01 -7.27242053e-01
-5.58667541e-01 8.80258977e-01 -1.76444933e-01 2.32605904e-01
4.18650240e-01 -7.12343514e-01 1.77316976e+00 -1.05977488e+00
4.05432791e-01 1.12938118e+00 3.51538926e-01 4.46629763e-01
-7.46352077e-01 -2.92602301e-01 -1.72853902e-01 4.10763144e-01
-1.26702189e+00 -5.02815366e-01 3.34153175e-01 -1.68447003e-01
1.21436870e+00 5.86230516e-01 2.07287371e-01 7.70583928e-01
5.49089424e-02 9.76515412e-01 1.01860297e+00 -7.99831271e-01
-2.04502478e-01 1.02429725e-01 3.31159949e-01 4.67193335e-01
-1.47740230e-01 6.63199648e-02 -6.51860833e-01 -5.39064705e-01
-2.01607347e-01 -2.72472978e-01 1.22852884e-01 1.47875354e-01
-9.75577116e-01 6.99387014e-01 3.33732545e-01 4.51200068e-01
-3.60908248e-02 2.56474346e-01 7.41271436e-01 5.56600511e-01
8.99840057e-01 7.59826958e-01 -4.98470008e-01 -3.10688257e-01
-1.36994159e+00 5.88370681e-01 6.86219573e-01 5.33324480e-01
3.76480818e-01 -6.18967414e-01 -1.01438808e+00 1.55614686e+00
2.13426679e-01 3.19214612e-01 6.38558209e-01 -1.03676236e+00
5.06226420e-01 1.38200194e-01 3.63993108e-01 -6.15171552e-01
-2.75118858e-01 -8.50061893e-01 -5.10219038e-01 -1.48050740e-01
7.13805929e-02 5.23923695e-01 -6.86873436e-01 1.31159282e+00
-4.55829650e-01 -1.37950867e-01 3.27865742e-02 4.73476797e-01
8.56849194e-01 9.19024110e-01 -1.71828538e-01 -3.15273292e-02
1.11500585e+00 -1.38379884e+00 -2.18127578e-01 1.19630262e-01
7.99609780e-01 -9.84659910e-01 1.07424772e+00 5.60964882e-01
-1.21362650e+00 -6.85742140e-01 -1.05595684e+00 -2.68200725e-01
-5.50408602e-01 4.05730933e-01 5.79962611e-01 5.59244275e-01
-1.57471979e+00 1.00295329e+00 -3.12857270e-01 -3.46886873e-01
-6.90614432e-02 4.28333580e-01 -4.88515496e-02 -2.43286669e-01
-1.77026546e+00 8.99224102e-01 2.74925828e-01 -1.33892015e-01
-1.00380301e+00 -6.32376552e-01 -5.01767546e-02 1.65452555e-01
-6.73100799e-02 -6.31799102e-01 1.43719280e+00 -1.66650817e-01
-1.51860404e+00 1.21989679e+00 -1.08550511e-01 -4.63451207e-01
6.43310010e-01 -8.01213026e-01 -7.12350667e-01 1.49277955e-01
2.23441362e-01 4.40343946e-01 2.63254285e-01 -1.17446601e+00
-8.97805095e-01 -1.19555481e-01 3.19318891e-01 2.66577274e-01
-7.27889419e-01 1.74364656e-01 -7.29817450e-01 -5.80061495e-01
-2.40348607e-01 -5.82435548e-01 4.58925441e-02 -9.40455675e-01
-2.89605439e-01 -8.07729781e-01 2.40629449e-01 -8.42707634e-01
1.66088128e+00 -1.75944221e+00 -1.23953089e-01 2.75957733e-01
-7.50243140e-04 3.74396116e-01 -5.17814040e-01 8.94111633e-01
4.38205600e-02 3.71627897e-01 4.23511267e-01 -5.28007150e-01
1.60586521e-01 -3.86534452e-01 -7.52697349e-01 1.40361533e-01
-1.20181814e-01 9.43165898e-01 -1.06057632e+00 -5.21879017e-01
-1.56747997e-01 3.54310244e-01 -4.95768815e-01 3.28673154e-01
-2.32131958e-01 1.05657190e-01 -5.17083526e-01 3.48251820e-01
2.34217778e-01 -2.84865886e-01 3.65870446e-02 -2.22859140e-02
-2.48533815e-01 9.28494453e-01 -4.61164743e-01 1.61876178e+00
-6.36414051e-01 6.31973863e-01 -5.05286634e-01 -8.13605964e-01
8.84769261e-01 1.18929118e-01 6.57933831e-01 -9.91023004e-01
-4.34162945e-01 6.64293051e-01 -5.59390366e-01 -8.24458599e-02
1.18211794e+00 4.34067219e-01 -1.04483955e-01 3.85640383e-01
-2.29528174e-01 -3.16846311e-01 7.97057331e-01 7.50529408e-01
1.26949990e+00 1.14226058e-01 -2.05520809e-01 -3.15508634e-01
7.95999885e-01 -6.16247430e-02 -2.66965777e-01 1.34098470e+00
2.74372935e-01 8.63311470e-01 2.91466445e-01 -2.38666087e-01
-1.08653915e+00 -9.30618405e-01 -1.94160163e-01 1.88130331e+00
-3.43156785e-01 -6.86674416e-01 -3.95608366e-01 -8.62699986e-01
2.74413358e-02 8.60236168e-01 -3.57630402e-01 -1.98247284e-01
-6.69033468e-01 -7.99676597e-01 7.74345577e-01 2.56099582e-01
1.74582213e-01 -9.17134583e-01 1.76775575e-01 2.47541204e-01
-5.10970294e-01 -4.62212205e-01 -4.39547330e-01 2.29392663e-01
-8.99737358e-01 -6.76100314e-01 -1.34675872e+00 -5.87583065e-01
1.50788143e-01 1.15146026e-01 1.65228081e+00 3.71468276e-01
-1.22512437e-01 6.18414640e-01 -5.96049428e-01 -5.70539981e-02
-3.38020086e-01 5.41712880e-01 1.17431641e-01 -4.50624168e-01
-2.13212464e-02 -1.04209848e-01 -6.86733365e-01 1.11036986e-01
-8.14980567e-01 -6.10235453e-01 4.82443929e-01 1.00844336e+00
4.62144196e-01 -6.07854985e-02 8.37897182e-01 -9.68010306e-01
9.96169984e-01 -2.26093367e-01 -5.07163703e-01 7.01459944e-01
-1.25096834e+00 2.94448882e-01 6.02652967e-01 8.09852704e-02
-8.44608366e-01 -7.48645425e-01 -2.34373972e-01 1.91130266e-01
5.95208704e-01 6.82683110e-01 5.40402174e-01 2.70468593e-01
8.50494921e-01 6.59682453e-02 -6.72694325e-01 -8.01182806e-01
5.24971783e-01 1.11984539e+00 2.84828603e-01 -9.07340884e-01
6.71715558e-01 3.04701805e-01 -4.89228293e-02 -1.41614616e-01
-1.36866856e+00 -1.04388690e+00 -3.84111881e-01 -2.35969961e-01
2.42590040e-01 -1.07319748e+00 -4.24776822e-01 2.55925030e-01
-1.07848740e+00 -9.43312123e-02 -1.76051646e-01 4.89056945e-01
-5.43238401e-01 5.20416975e-01 -7.58467436e-01 -6.51038945e-01
-6.72264457e-01 -6.84056163e-01 1.30479884e+00 -1.41967088e-02
1.09388679e-02 -7.94008136e-01 6.14828408e-01 6.24427021e-01
5.55464089e-01 -4.00144160e-01 1.25198877e+00 -7.22469747e-01
-2.69794643e-01 -5.89849710e-01 -5.12432814e-01 6.61024928e-01
-3.19622427e-01 -1.72593500e-02 -9.91709948e-01 -6.95718169e-01
-8.84928644e-01 -6.18303239e-01 1.53986549e+00 2.74047673e-01
1.21010327e+00 1.79735329e-02 -3.49316508e-01 7.92899504e-02
1.10240221e+00 -7.87855983e-02 8.50468874e-01 6.74831748e-01
3.06279570e-01 7.24647582e-01 8.79682183e-01 4.78552729e-01
3.93736899e-01 1.04129124e+00 1.82803050e-01 2.19987988e-01
-2.26458579e-01 -2.65868723e-01 6.42345607e-01 1.26685524e+00
-4.62146997e-01 -5.57843566e-01 -7.42700636e-01 3.94551009e-01
-1.83591056e+00 -1.08719420e+00 -1.60062924e-01 2.66352034e+00
8.03310633e-01 1.94114000e-01 1.19548522e-01 -6.66712523e-02
5.84371090e-01 4.94224072e-01 -2.06929520e-02 -5.74172020e-01
-2.16503009e-01 2.98752338e-01 4.88324016e-01 6.44902408e-01
-1.10908592e+00 8.56792390e-01 7.10945320e+00 1.26088428e+00
-6.43534899e-01 9.44040865e-02 8.67289364e-01 -2.33293042e-01
-4.49566543e-01 2.16554608e-02 -1.05573893e+00 3.72324973e-01
1.35667217e+00 -4.90789294e-01 6.09900117e-01 7.02006698e-01
-7.31803775e-02 1.99389875e-01 -1.05183244e+00 7.56994843e-01
3.16589624e-01 -1.32152700e+00 2.28727326e-01 -1.35206088e-01
8.57004046e-01 4.03923422e-01 2.22707346e-01 1.13964772e+00
2.96643287e-01 -1.25432968e+00 6.77314401e-01 9.33527470e-01
1.05629361e+00 -8.88519406e-01 8.47005486e-01 1.48496807e-01
-1.07186627e+00 -1.47156477e-01 -6.34345829e-01 3.39879483e-01
-6.74972162e-02 1.02904880e+00 -8.41624796e-01 9.53658342e-01
8.47501397e-01 6.67021036e-01 -8.78462911e-01 1.20357502e+00
-3.16913351e-02 7.68093467e-01 -9.19298753e-02 -2.39189595e-01
1.88810721e-01 1.79141730e-01 6.22847557e-01 1.53448427e+00
3.96070063e-01 -3.31886262e-01 4.05770913e-02 1.28234893e-01
-6.69675171e-01 4.25531805e-01 -2.56169707e-01 -2.30855092e-01
2.03399315e-01 1.11440420e+00 -8.46117362e-02 -5.42012453e-01
-1.20973894e-02 1.04627478e+00 5.95759511e-01 3.74035329e-01
-5.05199611e-01 -7.49808669e-01 -1.14023790e-01 9.69789401e-02
-7.64499083e-02 3.68096808e-04 3.25174749e-01 -9.85904753e-01
1.47530019e-01 -6.50131702e-01 7.40774632e-01 -7.16856539e-01
-1.47031486e+00 8.02190959e-01 1.73180640e-01 -1.39357197e+00
-5.43575466e-01 -5.32292426e-01 -1.72294319e-01 9.60328221e-01
-1.79170883e+00 -9.36772406e-01 3.27997178e-01 8.75564069e-02
3.93764436e-01 -6.50442183e-01 1.01896548e+00 7.85761535e-01
1.22607134e-01 6.67909980e-01 1.04598236e+00 -1.99086130e-01
1.12054634e+00 -1.49640906e+00 2.35611975e-01 5.50792143e-02
3.55761975e-01 8.14753234e-01 2.62166947e-01 -3.65924329e-01
-9.72853124e-01 -1.04231644e+00 1.81891489e+00 -6.88235044e-01
4.98611003e-01 -2.82403767e-01 -4.75780696e-01 2.47745961e-01
-2.36864761e-02 -1.65326998e-01 3.93133461e-01 7.65983880e-01
-6.41281426e-01 -5.02495408e-01 -6.60775721e-01 5.31945169e-01
7.54660964e-01 -8.97883117e-01 -7.44389713e-01 1.02118301e+00
7.10858226e-01 -2.90434629e-01 -7.85192549e-01 6.20204866e-01
8.47482741e-01 -9.09779549e-01 1.38349986e+00 -7.60094881e-01
5.65894008e-01 -6.09206855e-02 -3.21805388e-01 -1.26303720e+00
-5.53776801e-01 -6.15463793e-01 -2.12293670e-01 1.20380354e+00
6.11837983e-01 -2.72222310e-01 6.28655732e-01 -1.83659032e-01
-2.01904833e-01 -7.75656462e-01 -9.48881567e-01 -1.02363431e+00
3.43327403e-01 -2.77335018e-01 3.85693222e-01 3.69584978e-01
-1.16207570e-01 4.12855029e-01 -4.87339884e-01 -3.39202821e-01
2.59488702e-01 -2.71418542e-02 4.35939819e-01 -1.46834195e+00
-3.79663587e-01 -7.30843663e-01 6.71340339e-03 -1.23746431e+00
2.71249592e-01 -1.42732942e+00 6.48748577e-02 -1.87470460e+00
5.74844420e-01 -4.99775857e-01 -1.20784235e+00 1.56502396e-01
-2.16964617e-01 4.63934690e-01 1.97327885e-04 5.86932361e-01
-1.35968411e+00 5.96897960e-01 7.36080885e-01 -4.80281651e-01
6.27807677e-02 1.90713391e-01 -6.79788172e-01 6.82528019e-02
3.27237040e-01 -6.06842995e-01 -1.75556362e-01 -4.55356747e-01
7.19047964e-01 -8.18636119e-02 1.81811050e-01 -8.65163565e-01
1.35409072e-01 3.42002809e-01 2.85998702e-01 -8.80953789e-01
2.35165134e-01 -6.32717460e-02 -4.40610498e-01 3.32792938e-01
-1.00254524e+00 2.44554669e-01 4.57985327e-02 5.49321175e-01
-5.32075584e-01 -6.16004169e-01 4.88701224e-01 -1.11633874e-02
-5.05614936e-01 2.16839075e-01 -1.07857719e-01 2.41819575e-01
1.80662856e-01 4.75068331e-01 -5.92341781e-01 -6.76715732e-01
-5.00569224e-01 2.89596021e-01 -1.65324762e-01 5.33315718e-01
4.14154202e-01 -1.24938357e+00 -1.34812987e+00 -4.74242210e-01
3.77058357e-01 -6.81382298e-01 1.79425940e-01 4.29062337e-01
-4.91261601e-01 1.05849516e+00 2.66726732e-01 -2.36786544e-01
-1.16970623e+00 1.50938081e-02 6.25960827e-02 -1.47892463e+00
-3.07842284e-01 9.27383363e-01 -2.73923397e-01 -6.29881859e-01
3.64437342e-01 3.43777984e-01 -6.58080459e-01 1.74227223e-01
7.31973886e-01 6.71533048e-01 6.42837942e-01 -2.01113418e-01
-3.59920979e-01 6.24107838e-01 -6.73380136e-01 -6.48753405e-01
1.26245952e+00 -7.56856799e-02 -3.74456972e-01 4.23253715e-01
1.62119043e+00 4.53250974e-01 -4.01388943e-01 -2.41472021e-01
5.34012914e-01 -2.20793232e-01 2.23918006e-01 -1.29429531e+00
-4.92917955e-01 8.18600297e-01 5.67297637e-01 3.19373727e-01
9.46796834e-01 -1.11664683e-01 6.79570258e-01 8.95245731e-01
4.60441858e-01 -1.43783784e+00 3.95236164e-01 1.10076058e+00
1.20554590e+00 -9.88781750e-01 2.86100328e-01 3.44996095e-01
-3.46384406e-01 1.23634243e+00 -4.08474766e-02 -3.72345606e-03
4.17357236e-01 -4.83651906e-01 -1.75579056e-01 -1.01258159e-01
-1.01398885e+00 -9.21476409e-02 9.67543602e-01 1.83255062e-01
9.10839736e-01 8.28444138e-02 -6.49924874e-01 3.81820679e-01
-1.06274679e-01 -1.82516314e-02 1.22570088e-02 5.23033321e-01
-5.12974381e-01 -1.61224365e+00 -6.33248612e-02 7.23787904e-01
-8.82282436e-01 -6.40497267e-01 -3.54245305e-01 3.44214916e-01
-3.78573090e-01 1.21281719e+00 -2.90031970e-01 -4.57098871e-01
4.31627214e-01 7.19832852e-02 3.43401670e-01 -8.34224939e-01
-1.01287961e+00 -1.42463699e-01 6.04882479e-01 -5.00503004e-01
-5.93691990e-02 -3.32480490e-01 -1.00829852e+00 -1.17631949e-01
-1.93413705e-01 8.37500930e-01 7.39149690e-01 4.87542927e-01
3.96804333e-01 5.56263924e-01 9.41351950e-01 -4.59447235e-01
-1.04782963e+00 -1.12533998e+00 -7.54159808e-01 3.61207306e-01
2.01474443e-01 -3.45728874e-01 -5.34980595e-01 -3.20286393e-01] | [11.502025604248047, 7.611375331878662] |
c34040ad-c2d5-41df-930d-c33697f36fa9 | a-multilingual-wikified-data-set-of | null | null | https://aclanthology.org/L18-1073 | https://aclanthology.org/L18-1073.pdf | A Multilingual Wikified Data Set of Educational Material | null | ['Antal Van den Bosch', 'Maria Stasimioti', 'Valia Kordoni', 'Thanasis Naskos', 'Maja Popovic', 'Katia Lida Kermanidis', 'Hugo de Vos', 'Eirini Takoulidou', 'Markus Egg', 'Panayota Georgakopoulou', 'Menno van Zaanen', 'Iris Hendrickx', 'Vilelmini Sosoni'] | 2018-05-01 | a-multilingual-wikified-data-set-of-1 | https://aclanthology.org/L18-1073 | https://aclanthology.org/L18-1073.pdf | lrec-2018-5 | ['cross-lingual-semantic-textual-similarity'] | ['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] | [-7.465922832489014, 3.6727030277252197] |
197e3e61-a5ca-4355-9a0b-e20ddd1e36cf | scope-structural-continuity-preservation-for | 2304.14572 | null | https://arxiv.org/abs/2304.14572v1 | https://arxiv.org/pdf/2304.14572v1.pdf | SCOPE: Structural Continuity Preservation for Medical Image Segmentation | Although the preservation of shape continuity and physiological anatomy is a natural assumption in the segmentation of medical images, it is often neglected by deep learning methods that mostly aim for the statistical modeling of input data as pixels rather than interconnected structures. In biological structures, however, organs are not separate entities; for example, in reality, a severed vessel is an indication of an underlying problem, but traditional segmentation models are not designed to strictly enforce the continuity of anatomy, potentially leading to inaccurate medical diagnoses. To address this issue, we propose a graph-based approach that enforces the continuity and connectivity of anatomical topology in medical images. Our method encodes the continuity of shapes as a graph constraint, ensuring that the network's predictions maintain this continuity. We evaluate our method on two public benchmarks on retinal vessel segmentation, showing significant improvements in connectivity metrics compared to traditional methods while getting better or on-par performance on segmentation metrics. | ['Nassir Navab', 'Ehsan Adeli', 'Yongjian Tang', 'Rui Xiao', 'Amr Abu-zer', 'Goktug Guevercin', 'Azade Farshad', 'Yousef Yeganeh'] | 2023-04-28 | null | null | null | null | ['anatomy'] | ['miscellaneous'] | [ 3.90691273e-02 4.90805507e-01 -1.64764315e-01 -4.81144845e-01
1.33717537e-01 -6.57545507e-01 3.17505419e-01 6.07781231e-01
-1.09536916e-01 5.74427843e-01 1.23280272e-01 -5.37498951e-01
-7.51710311e-02 -9.17913079e-01 -6.16960168e-01 -4.73887950e-01
-4.20496948e-02 5.42060435e-01 4.45606738e-01 7.35579953e-02
2.79747532e-03 8.09884965e-01 -8.52432847e-01 -4.28510481e-04
1.04424715e+00 9.17046964e-01 -2.92491823e-01 5.29421628e-01
-3.81445050e-01 6.57541215e-01 -2.10263550e-01 -5.92536449e-01
4.59029078e-01 -6.83111906e-01 -1.02845478e+00 2.62784958e-01
7.54991710e-01 1.71671547e-02 -2.80446798e-01 1.47157729e+00
2.00223848e-01 -1.96610540e-01 7.07598746e-01 -9.04875040e-01
-4.29563463e-01 5.97760975e-01 -5.65034270e-01 1.65243194e-01
-1.31651223e-01 2.28761390e-01 1.21675706e+00 -8.20356607e-02
7.60672987e-01 8.01407754e-01 8.23694766e-01 3.35090846e-01
-1.83496594e+00 -1.12636350e-01 2.03107491e-01 -2.96198100e-01
-1.26782799e+00 -2.62603402e-01 5.45727730e-01 -7.40808606e-01
4.24609482e-01 3.84475589e-01 1.08022022e+00 4.90594447e-01
2.41829216e-01 3.66775692e-01 9.09138024e-01 2.72435211e-02
3.06878358e-01 3.82684022e-02 2.61238366e-01 8.30624998e-01
5.34394741e-01 -1.98946595e-01 -1.56468853e-01 -9.58720595e-03
1.06642902e+00 -1.85832158e-01 -3.14525217e-01 -5.63892841e-01
-1.30032098e+00 6.63797796e-01 8.35357308e-01 4.41416353e-01
-3.84429932e-01 9.10207108e-02 2.39397600e-01 1.51546836e-01
3.08846653e-01 5.71543515e-01 -1.00225993e-01 2.58440286e-01
-1.22106946e+00 1.06926672e-01 8.81667912e-01 7.27152348e-01
5.26558757e-01 -1.60115451e-01 -9.77095068e-02 4.58021820e-01
3.78689378e-01 -7.73281679e-02 1.86772659e-01 -1.00035834e+00
-1.33878589e-01 1.13745046e+00 -2.30892166e-01 -1.12552953e+00
-6.64824247e-01 -8.22376549e-01 -1.18147671e+00 3.46244961e-01
9.99514997e-01 3.95786017e-02 -1.11273980e+00 1.64207149e+00
3.90552253e-01 1.55310273e-01 -3.12382460e-01 9.76566911e-01
8.50925803e-01 1.40774190e-01 2.09922209e-01 -4.64126281e-02
1.07076454e+00 -6.92999244e-01 -4.62496400e-01 -3.51934023e-02
5.59911847e-01 -4.32702363e-01 7.55837381e-01 3.40418845e-01
-1.14449835e+00 -1.93587258e-01 -7.10101545e-01 -9.43275392e-02
-7.88437948e-02 -2.75195569e-01 8.91195476e-01 6.21673286e-01
-1.22187483e+00 7.76451290e-01 -1.00576079e+00 -3.32010865e-01
9.67823982e-01 4.71483827e-01 -4.78328317e-01 2.37597063e-01
-5.55040240e-01 7.58425951e-01 9.96067971e-02 3.31119895e-02
-3.21504146e-01 -1.12881792e+00 -5.62400460e-01 1.86475262e-01
8.82494450e-02 -9.97947395e-01 6.64560795e-01 -1.08511531e+00
-1.03870928e+00 1.26560211e+00 -1.94576918e-04 -7.59607136e-01
1.05690730e+00 2.52795994e-01 -5.15255108e-02 2.04421952e-01
-3.32412899e-01 8.69294047e-01 4.74943638e-01 -1.27982187e+00
-4.33242202e-01 -3.64922196e-01 2.45512471e-01 -7.73447379e-03
-1.19245835e-01 -3.67652446e-01 -4.16258603e-01 -4.31193411e-01
6.09166980e-01 -6.56840563e-01 -6.91935539e-01 6.50540888e-01
-9.35222447e-01 1.41371414e-01 3.42170328e-01 -5.86017191e-01
1.08578861e+00 -2.01074171e+00 2.05557004e-01 5.61501265e-01
8.31995368e-01 2.30557080e-02 8.27975124e-02 -2.04721466e-01
4.84172814e-03 4.05542761e-01 -6.35005713e-01 -2.43023023e-01
-1.67834312e-01 2.31140718e-01 -2.40328533e-04 7.19986439e-01
1.69037253e-01 1.03451729e+00 -7.20658958e-01 -7.50306845e-01
1.27825111e-01 4.77219880e-01 -8.05189967e-01 -1.93395898e-01
-2.80149460e-01 9.05364573e-01 -3.88491929e-01 5.13477206e-01
4.92896438e-01 -7.12637544e-01 4.40466702e-01 -3.45954865e-01
9.80269015e-02 2.17680335e-01 -9.53093171e-01 1.57824540e+00
5.57813011e-02 5.36645174e-01 5.09164520e-02 -1.07221913e+00
7.07661629e-01 2.26929545e-01 1.11751640e+00 -4.49241847e-01
3.65157068e-01 1.88064888e-01 6.35350943e-01 -2.09240392e-01
1.03465281e-02 -1.27299830e-01 4.69883323e-01 1.66110411e-01
-2.86394626e-01 -2.49499381e-01 2.11242884e-01 1.75908610e-01
1.06466091e+00 4.71636914e-02 1.21710852e-01 -5.69139600e-01
2.24876568e-01 2.81472325e-01 7.54086375e-01 4.27755564e-01
-2.25355223e-01 8.66020977e-01 1.07637703e+00 -5.43266594e-01
-1.16268790e+00 -1.14572990e+00 -5.48308730e-01 9.83347893e-02
2.02118173e-01 -2.08555549e-01 -1.01423848e+00 -8.34611595e-01
1.47928432e-01 2.96594173e-01 -7.22357154e-01 2.35379059e-02
-4.41107243e-01 -8.84016037e-01 5.35534322e-01 3.13393950e-01
3.01154882e-01 -7.69320548e-01 -6.67735338e-01 2.35319257e-01
-1.08280312e-02 -1.23005652e+00 -3.83109599e-01 -8.63612145e-02
-1.03404272e+00 -1.30994987e+00 -5.97821593e-01 -7.18303680e-01
1.30383015e+00 -1.99795082e-01 1.28486133e+00 7.39378035e-01
-5.78713536e-01 -1.10690556e-01 5.69058806e-02 -8.93123224e-02
-4.74252492e-01 -5.67890108e-02 -4.38146383e-01 1.57756716e-01
4.58448790e-02 -8.14291298e-01 -8.27462852e-01 1.91733778e-01
-8.26623321e-01 1.84816346e-01 4.58753496e-01 6.57033741e-01
8.22358191e-01 9.97491553e-02 2.31862381e-01 -1.17972708e+00
2.03068927e-01 -3.02502751e-01 -7.92878807e-01 1.95172131e-01
-6.66732609e-01 -1.85626715e-01 4.35493588e-01 -3.73396367e-01
-4.92953360e-01 3.77243131e-01 -6.15656599e-02 -2.36137897e-01
-3.48286271e-01 6.25450075e-01 -8.86802897e-02 -4.46174413e-01
5.73419511e-01 6.82858229e-02 2.45711178e-01 -3.72234523e-01
1.89776316e-01 -1.76412910e-02 4.76451367e-01 -4.01243210e-01
6.14585519e-01 8.23434770e-01 4.85488147e-01 -7.61590004e-01
-7.51914859e-01 -2.31668919e-01 -9.28221345e-01 -1.37418449e-01
1.03753626e+00 -3.33856612e-01 -6.30002201e-01 1.40038624e-01
-9.64082837e-01 -4.94035780e-01 -4.42697495e-01 3.67805898e-01
-3.73416007e-01 5.56848347e-01 -5.19721210e-01 -3.54651958e-01
-7.95182735e-02 -1.23581314e+00 5.95868409e-01 2.05464512e-01
-4.43192035e-01 -1.39733851e+00 -1.89875457e-02 1.52586654e-01
4.58293825e-01 7.56625891e-01 1.39446461e+00 -4.72499907e-01
-6.34710014e-01 -1.72689110e-01 -3.43154639e-01 2.26611570e-01
3.90355170e-01 3.33533973e-01 -7.01331615e-01 4.20085862e-02
-3.50203395e-01 1.82686433e-01 8.62155437e-01 7.48070240e-01
1.17091596e+00 -2.02844754e-01 -3.62034440e-01 7.91345119e-01
1.54343915e+00 -2.17181340e-01 7.06551075e-01 -2.15557501e-01
8.00911665e-01 1.02243578e+00 -1.78558439e-01 1.93198204e-01
2.74714023e-01 4.04943079e-01 7.72272587e-01 -7.34329641e-01
-3.78851473e-01 -5.71534261e-02 -2.72359759e-01 4.82847273e-01
-2.14719430e-01 -1.04275882e-01 -1.06550527e+00 6.73436880e-01
-1.62242854e+00 -6.15306318e-01 -8.23900759e-01 2.36585021e+00
1.01854348e+00 3.10923219e-01 2.44241685e-01 -9.24494714e-02
6.03748620e-01 -2.42489174e-01 -7.05778956e-01 -6.09078109e-02
-2.23132819e-01 1.66065127e-01 6.20280445e-01 5.95803857e-01
-1.04637432e+00 9.31731820e-01 6.80224228e+00 1.79765299e-01
-1.42937469e+00 -1.35548070e-01 9.63217318e-01 6.90824585e-03
-4.06240821e-01 1.43207356e-01 -3.83348733e-01 3.96115154e-01
3.07085514e-01 -5.15306853e-02 2.12683052e-01 4.64997172e-01
3.49834949e-01 3.31059583e-02 -1.22009826e+00 5.88181555e-01
-5.21080494e-01 -1.44094300e+00 1.64197251e-01 4.81934637e-01
7.51285315e-01 -3.98097634e-02 -3.36254463e-02 -4.55806643e-01
3.86914432e-01 -1.41540480e+00 6.35545433e-01 6.70628905e-01
5.42565167e-01 -3.91369879e-01 5.17923355e-01 1.35544911e-01
-8.30258727e-01 4.70106602e-01 -3.49024951e-01 2.88633078e-01
2.43274838e-01 9.84606743e-01 -5.96944809e-01 4.01056916e-01
4.62738156e-01 8.18402708e-01 -5.22147834e-01 1.69101453e+00
-1.40222073e-01 7.71630049e-01 -3.24041545e-01 4.42383826e-01
1.75217211e-01 -4.76647228e-01 6.29163265e-01 1.00269485e+00
-5.42782173e-02 -1.17339380e-01 1.81787163e-01 1.41597927e+00
-1.69757277e-01 2.80687511e-01 -4.73469049e-01 -2.25253373e-01
1.15933008e-01 1.39185798e+00 -1.24060476e+00 -1.87767282e-01
-4.62325633e-01 5.53603292e-01 2.43582681e-01 3.18151832e-01
-8.20946038e-01 1.73080921e-01 7.42527604e-01 5.79461575e-01
6.63740560e-02 -2.69562155e-01 -9.49167371e-01 -9.19468701e-01
1.79505683e-02 -7.20194817e-01 2.09670231e-01 -1.96972579e-01
-1.36748409e+00 5.08651316e-01 -6.49691880e-01 -1.05758965e+00
4.41297144e-01 -4.77759689e-01 -7.00202167e-01 7.75542200e-01
-1.68547678e+00 -1.19445765e+00 -3.99685383e-01 4.49611008e-01
-7.27510527e-02 2.92788655e-01 6.56004250e-01 4.11700726e-01
-3.91102970e-01 3.11870188e-01 -3.42099994e-01 2.94538647e-01
4.75562543e-01 -1.61855078e+00 4.03445393e-01 7.65190423e-01
3.52970064e-01 8.18570316e-01 6.61573827e-01 -8.04635942e-01
-1.01522124e+00 -1.19331753e+00 7.61962116e-01 -2.09756196e-01
7.29949355e-01 -1.18005544e-01 -1.29195547e+00 6.78263664e-01
1.06425554e-01 5.69741070e-01 6.10057473e-01 -4.60744724e-02
-2.86406040e-01 3.21841873e-02 -1.14469349e+00 5.95503092e-01
1.04280102e+00 -8.27114359e-02 -1.98732659e-01 5.39667845e-01
4.91031140e-01 -4.65909034e-01 -1.19222498e+00 3.10423791e-01
4.79740709e-01 -9.85700488e-01 9.18347478e-01 -9.09895957e-01
4.89540964e-01 -4.19702470e-01 1.42253339e-01 -1.09362113e+00
-3.61799717e-01 -5.31487346e-01 1.77988246e-01 8.89567912e-01
6.84300125e-01 -5.82592130e-01 1.11241162e+00 7.96757698e-01
-2.22346634e-01 -7.55322337e-01 -9.01793778e-01 -7.13525653e-01
4.12041426e-01 -6.47408292e-02 4.65651453e-01 1.32991600e+00
-2.58819610e-01 -1.63221993e-02 1.39266819e-01 2.57685065e-01
6.71229899e-01 -2.88282195e-03 5.24048746e-01 -1.69463444e+00
-2.64784306e-01 -1.08903289e+00 -6.99046969e-01 -7.21686184e-01
1.19871877e-01 -1.17259896e+00 -1.24358155e-01 -1.90884447e+00
-4.08154689e-02 -8.34107935e-01 -2.08066180e-01 5.19498289e-01
5.05001619e-02 4.52343464e-01 -2.61178523e-01 1.57732680e-01
-2.45981455e-01 1.14189245e-01 1.63767099e+00 -1.49869800e-01
-1.77067235e-01 -3.02898139e-02 -7.50400722e-01 8.80960941e-01
6.65748060e-01 -2.70268410e-01 -3.47893119e-01 -3.96854967e-01
3.40004891e-01 -2.01084986e-01 8.05243313e-01 -9.97515321e-01
3.27728808e-01 -1.44940615e-01 2.65524060e-01 -1.00789346e-01
-3.15476730e-02 -1.00266171e+00 2.05923572e-01 6.87089264e-01
-4.48229939e-01 -2.86765069e-01 6.60597607e-02 3.61649483e-01
-1.86042204e-01 6.56223074e-02 1.15776587e+00 -2.09400579e-01
-1.63096145e-01 5.10773778e-01 4.48060930e-02 2.02019736e-01
8.93098950e-01 -3.41656119e-01 -2.35426143e-01 -1.49433285e-01
-9.60127115e-01 1.80038497e-01 8.28368127e-01 -5.66439517e-02
4.15959835e-01 -8.67603898e-01 -7.95608044e-01 2.69497419e-03
-1.69797987e-01 3.50520402e-01 7.49619901e-02 1.44694698e+00
-1.00260019e+00 9.34903100e-02 -2.19111040e-01 -8.30058753e-01
-1.17983663e+00 2.48839721e-01 9.35006738e-01 -1.41735032e-01
-1.03570187e+00 6.95904076e-01 3.50033045e-01 -2.05848381e-01
2.42398858e-01 -6.65566623e-01 -2.63228923e-01 -2.21035615e-01
-1.04606450e-02 -1.10245019e-01 6.66851103e-02 -5.68984389e-01
-2.15836316e-01 5.60174167e-01 8.34300146e-02 2.70285904e-01
1.24543977e+00 -9.05138440e-03 -5.54533958e-01 2.17723548e-01
6.27855539e-01 1.86414376e-01 -1.21953928e+00 -2.30975375e-01
1.29228994e-01 -3.00497532e-01 1.76806927e-01 -9.51506376e-01
-1.53256941e+00 9.35535550e-01 2.79932171e-01 5.91262221e-01
7.54016340e-01 -2.34870017e-02 7.66933262e-01 -6.55196384e-02
2.30732366e-01 -6.69276714e-01 -4.09067512e-01 1.14872284e-01
6.69357717e-01 -1.05480206e+00 5.98327145e-02 -9.42437172e-01
-4.01183963e-01 9.97323096e-01 5.08928537e-01 -2.58052647e-01
7.77461231e-01 2.27398321e-01 5.38707525e-02 -4.59913045e-01
-3.39914322e-01 -1.79325372e-01 6.18384719e-01 4.05943185e-01
6.29400611e-01 1.61585566e-02 -2.50792325e-01 1.65962309e-01
-3.79393041e-01 -1.42798394e-01 3.75765681e-01 4.98152196e-01
-3.16911459e-01 -9.63933587e-01 1.43461451e-01 6.04600370e-01
-7.05746174e-01 -2.55807936e-01 -7.79846787e-01 8.65723014e-01
1.58725724e-01 5.70279300e-01 2.78553307e-01 9.36925486e-02
2.27311969e-01 -2.01446727e-01 5.63940585e-01 -6.93545163e-01
-7.59341955e-01 1.50987059e-01 -1.30720854e-01 -5.71671069e-01
-3.00353199e-01 -5.64683318e-01 -1.67022967e+00 -3.05263758e-01
-9.60559547e-02 -4.40542772e-02 6.17683232e-01 9.20442939e-01
2.62376785e-01 7.54934669e-01 2.01213852e-01 -3.94613706e-02
-1.77718326e-01 -5.96478343e-01 -6.99548960e-01 6.95123672e-01
3.49841088e-01 -4.15614814e-01 -2.32832566e-01 3.99963588e-01] | [14.335618019104004, -2.8060457706451416] |
640cbb74-30a3-4141-b60f-b84557cbb296 | physics-informed-machine-learning-for-the | 2008.08162 | null | https://arxiv.org/abs/2008.08162v1 | https://arxiv.org/pdf/2008.08162v1.pdf | Physics-informed machine learning for the COVID-19 pandemic: Adherence to social distancing and short-term predictions for eight countries | The spread of COVID-19 during the initial phase of the first half of 2020 was curtailed to a larger or lesser extent through measures of social distancing imposed by most countries. In this work, we link directly, through machine learning techniques, infection data at a country level to a single number that signifies social distancing effectiveness. We assume that the standard SIR model gives a reasonable description of the dynamics of spreading, and thus the social distancing aspect can be modeled through time-dependent infection rates that are imposed externally. We use an exponential ansatz to analyze the SIR model, find an exact solution for the time-independent infection rate, and derive a simple first-order differential equation for the time-dependent infection rate as a function of the infected population. Using infected number data from the "first wave" of the infection from eight countries, and through physics-informed machine learning, we extract the degree of linear dependence in social distancing that led to the specific infections. We find that in the two extremes are Greece, with the highest decay slope on one side, and the US on the other with a practically flat "decay". The hierarchy of slopes is compatible with the effectiveness of the pandemic containment in each country. Finally, we train our network with data after the end of the analyzed period, and we make week-long predictions for the current phase of the infection that appear to be very close to the actual infection values. | ['G. P. Tsironis', 'G. D. Barmparis'] | 2020-08-18 | null | null | null | null | ['physics-informed-machine-learning'] | ['graphs'] | [-3.67309421e-01 9.77478325e-02 -1.72746718e-01 2.75031328e-01
1.42132610e-01 -3.66564304e-01 1.10888565e+00 2.97603697e-01
-5.55823267e-01 9.23115551e-01 2.27153137e-01 -5.62502086e-01
-4.45355326e-01 -1.10115600e+00 -3.70193362e-01 -1.03729773e+00
-4.19304818e-01 1.06441593e+00 9.37341377e-02 -7.35902548e-01
8.33290964e-02 5.14542222e-01 -7.51324236e-01 -2.46148393e-01
8.72560143e-01 4.39933121e-01 -1.81781113e-01 5.90270758e-01
9.72672552e-03 2.47256070e-01 -8.03526759e-01 -7.07526058e-02
1.71306953e-01 -4.01694417e-01 -6.06958687e-01 -3.77342820e-01
-5.23598492e-01 -2.21909508e-01 -4.44962025e-01 5.03804922e-01
2.57321119e-01 -3.81900042e-01 1.05810511e+00 -1.10202289e+00
-4.91542846e-01 4.98251319e-01 -6.54917836e-01 3.65610063e-01
2.51608491e-01 2.27480829e-01 4.31293875e-01 -2.37391472e-01
7.62459040e-01 1.25353038e+00 9.08048332e-01 2.63814598e-01
-1.32037127e+00 -4.55482602e-01 -2.71219403e-01 -2.81988144e-01
-1.35031283e+00 2.15994254e-01 4.89583760e-01 -9.29752409e-01
7.50559986e-01 1.99865387e-03 1.04413831e+00 1.05244935e+00
6.31223440e-01 -2.88834155e-01 1.02887869e+00 -3.71425420e-01
2.46276692e-01 2.18504041e-01 -2.10093111e-01 3.77728164e-01
5.20189583e-01 3.35989296e-01 1.21873654e-01 -5.41984558e-01
8.41706455e-01 1.65405180e-02 -3.40183705e-01 -2.52732988e-02
-8.61436248e-01 1.09778738e+00 6.01940513e-01 7.65889704e-01
-3.71429563e-01 -1.49853528e-01 2.42039159e-01 4.62699354e-01
7.66705513e-01 3.73445332e-01 -8.69085789e-01 2.78455585e-01
-7.87424386e-01 2.43765920e-01 1.01753783e+00 -6.04134127e-02
7.15569615e-01 -4.09229062e-02 2.06708059e-01 5.08736789e-01
2.11590737e-01 8.50376189e-01 7.34275132e-02 -7.25431502e-01
9.37345177e-02 3.84172410e-01 3.20726365e-01 -9.28361058e-01
-7.43944287e-01 -6.37057424e-01 -1.13597417e+00 1.36093125e-02
9.15206969e-01 -6.22445345e-01 -5.25895774e-01 1.89400017e+00
1.43283054e-01 1.34058714e-01 -1.88222677e-01 4.51384664e-01
1.12075999e-01 9.35244560e-01 1.52420485e-02 -7.60937274e-01
1.04016221e+00 -2.53013432e-01 -3.27749461e-01 1.96634889e-01
7.28272498e-01 -3.26396197e-01 4.75575149e-01 1.70437470e-01
-7.66990602e-01 -1.32980132e-02 -6.24798357e-01 9.29752707e-01
-4.94022578e-01 -6.37044787e-01 4.25060719e-01 4.67601866e-01
-1.24299455e+00 8.63012016e-01 -5.86273849e-01 -7.70610571e-01
-7.36045539e-02 2.07403675e-01 -4.91056591e-02 4.58312035e-01
-1.58442402e+00 1.01341951e+00 1.53796956e-01 -1.78964764e-01
-6.33470774e-01 -9.78128135e-01 -1.71575963e-01 -4.56912816e-02
-1.75002646e-02 -8.95570099e-01 5.82467735e-01 -8.89376402e-01
-1.09748495e+00 7.22216010e-01 2.55938649e-01 -4.85639393e-01
6.65632904e-01 2.81432599e-01 -4.88114029e-01 8.05079937e-02
-5.25719225e-02 2.59552270e-01 5.85176766e-01 -1.32974958e+00
-1.07699722e-01 -5.43468773e-01 -5.91438338e-02 -1.24840073e-01
-2.90067971e-01 2.76588917e-01 -3.05961762e-02 -4.58521903e-01
-4.04311150e-01 -1.05781639e+00 -3.78270209e-01 -4.72941697e-01
-2.11194471e-01 -7.63407946e-02 5.23794174e-01 -8.42362761e-01
1.10679448e+00 -1.82108021e+00 3.46413910e-01 4.36153412e-01
3.06246996e-01 1.56203046e-01 4.69878167e-02 9.16880786e-01
-1.16318956e-01 4.16385621e-01 -3.92391890e-01 1.32915705e-01
-3.53978038e-01 1.90772250e-01 -6.35293126e-02 6.87028587e-01
1.90754682e-01 6.58233941e-01 -8.82594168e-01 -1.78536490e-01
-8.30325857e-02 7.16678381e-01 -6.45713687e-01 -1.97129510e-03
-6.80066273e-02 8.34733248e-01 -3.21315020e-01 -6.10283129e-02
6.86893106e-01 -2.17154041e-01 2.40210041e-01 2.80557722e-01
-3.01079303e-01 -1.62131667e-01 -4.90004122e-01 7.95209944e-01
-3.37497056e-01 4.39811051e-01 1.80973828e-01 -1.08616984e+00
7.98956811e-01 3.65975380e-01 9.34110582e-01 -5.51866472e-01
2.39698738e-01 2.60708064e-01 3.27990383e-01 -2.43581712e-01
2.48372275e-02 -4.79724228e-01 -3.23780775e-02 6.18834913e-01
-1.25275552e-01 -1.59809127e-01 5.78715801e-02 2.43312776e-01
1.07960582e+00 -3.65633518e-01 2.88115829e-01 -7.12229848e-01
3.29354554e-01 -1.12941466e-01 1.72256887e-01 5.26138067e-01
-9.93156806e-02 5.29416092e-02 9.67701495e-01 -4.39011097e-01
-1.32025838e+00 -1.22796488e+00 -3.46051306e-01 5.65428376e-01
-8.51390734e-02 -2.33663991e-02 -9.24436390e-01 -2.71526098e-01
1.97444096e-01 5.45730472e-01 -8.15925598e-01 -2.90204287e-01
-7.13919818e-01 -1.24661684e+00 4.84295189e-01 -1.83058098e-01
3.49207819e-01 -8.17417741e-01 -1.99232504e-01 -4.15649265e-02
-5.59054576e-02 -7.28632390e-01 1.32108303e-02 1.08209088e-01
-8.04953396e-01 -9.65417922e-01 -9.55792189e-01 -3.42676431e-01
4.99714553e-01 -3.99574071e-01 1.01176918e+00 3.38387907e-01
-8.76474828e-02 1.27781510e-01 1.41990215e-01 -3.48474354e-01
-8.07096303e-01 6.53619394e-02 3.44196975e-01 -3.06759536e-01
-8.93068835e-02 -6.59903228e-01 -7.01356053e-01 2.87400723e-01
-7.32294321e-01 -5.15982091e-01 7.94042051e-02 2.97756851e-01
-2.27416992e-01 2.48412088e-01 8.06210339e-01 -4.03423101e-01
6.60535514e-01 -9.04935122e-01 -6.11751080e-01 -8.42778385e-03
-6.49044156e-01 -8.94746259e-02 6.76532030e-01 -5.06640971e-01
-9.58945334e-01 -5.66153109e-01 -2.83764713e-02 1.59368008e-01
-3.57544683e-02 5.77216208e-01 4.11299288e-01 1.70794591e-01
6.50151074e-01 -2.07403880e-02 1.54923409e-01 -5.07134318e-01
2.29975298e-01 6.42351925e-01 2.30615363e-01 -4.96647567e-01
1.06062305e+00 5.86412251e-01 1.39568793e-02 -1.18570316e+00
-4.72377509e-01 -5.20684198e-02 -5.13130903e-01 -2.95406640e-01
7.87458956e-01 -7.52743542e-01 -8.77068222e-01 7.64199793e-01
-1.16394067e+00 -5.77762246e-01 -3.47411782e-01 4.73929971e-01
-3.99547994e-01 -7.18879700e-03 -8.55167150e-01 -1.02170396e+00
-1.78315952e-01 -4.89652902e-01 5.58661699e-01 -3.04676630e-02
-2.42993757e-01 -1.60658944e+00 7.85341561e-01 -7.28820190e-02
8.51195395e-01 4.28742379e-01 1.27295280e+00 -4.99589801e-01
-9.93205309e-02 -1.97198436e-01 -1.86903849e-01 -5.67823723e-02
2.27728963e-01 2.95977980e-01 -4.29823160e-01 -4.33400780e-01
2.26511538e-01 2.81339049e-01 8.19545805e-01 8.19067180e-01
2.22675070e-01 -4.77299601e-01 -5.09854555e-01 3.06044698e-01
1.60114729e+00 2.32137904e-01 5.89115620e-01 2.12624192e-01
2.59999335e-01 6.98225915e-01 3.05719972e-02 5.35077512e-01
2.07023218e-01 7.42576778e-01 2.80565619e-01 -2.60570168e-01
2.58275449e-01 -8.95826221e-02 4.02536124e-01 9.24429238e-01
-5.84579527e-01 -1.63641587e-01 -1.13533700e+00 3.63351464e-01
-1.38801157e+00 -1.07568550e+00 -2.99442500e-01 2.52352881e+00
7.68135726e-01 3.24705660e-01 6.60849214e-01 -2.15729132e-01
6.49224222e-01 1.29165233e-03 3.05181090e-02 -5.60819507e-01
-1.04794927e-01 2.68753916e-02 7.62744546e-01 9.92673278e-01
-6.08694732e-01 4.95557845e-01 8.02839470e+00 4.45678502e-01
-1.27803826e+00 1.19593397e-01 7.05149531e-01 9.07048061e-02
-3.72589886e-01 6.93875402e-02 -4.30179298e-01 7.39679933e-01
1.46854067e+00 -4.88887608e-01 5.45957386e-01 -2.04344876e-02
6.67777479e-01 -1.59886405e-01 -5.93483865e-01 2.62458235e-01
-4.70741928e-01 -1.12026119e+00 -1.04048833e-01 5.24428427e-01
8.44887733e-01 3.29862028e-01 -2.33527169e-01 9.12990570e-02
3.38492334e-01 -9.07027125e-01 2.04648778e-01 7.59161472e-01
7.47900605e-01 -7.02708662e-01 7.88084388e-01 7.07722425e-01
-1.03714812e+00 -7.72507787e-02 -6.04768135e-02 -3.75570834e-01
3.97169322e-01 8.53311241e-01 -6.87777042e-01 4.04795229e-01
4.02879536e-01 4.20283109e-01 -2.70284921e-01 8.21344197e-01
1.35089442e-01 7.14749873e-01 -6.36869133e-01 1.04265064e-01
7.96585754e-02 -6.60726428e-01 6.92139804e-01 1.02218974e+00
3.11580002e-01 -8.69034752e-02 -3.22225839e-01 7.26643026e-01
2.76908994e-01 -1.57087848e-01 -8.10725152e-01 -1.12861343e-01
1.27264485e-01 8.53957951e-01 -6.43195808e-01 -1.40062734e-01
-1.11450441e-01 5.11705995e-01 1.05587929e-01 6.85511827e-01
-9.01912928e-01 5.13076782e-04 5.17003715e-01 6.30309641e-01
1.13928795e-01 -3.92466903e-01 5.59150577e-02 -9.31337655e-01
-4.23177958e-01 -4.61120069e-01 1.55798003e-01 -3.77665073e-01
-1.11732030e+00 3.42658699e-01 2.64285713e-01 -5.90032339e-01
-5.15248120e-01 -4.26047415e-01 -8.91712606e-01 9.80604470e-01
-1.06817353e+00 -5.69068611e-01 3.16356510e-01 4.09172863e-01
-1.51017308e-01 -8.29139352e-03 8.00665021e-01 3.68329138e-01
-4.79800731e-01 4.98787165e-02 4.04959887e-01 -7.10333064e-02
2.73558706e-01 -8.94538045e-01 2.30156124e-01 2.22364292e-01
-4.36399281e-01 7.83646047e-01 1.10245073e+00 -1.08036935e+00
-8.92877400e-01 -6.26439869e-01 1.13482475e+00 -3.95992637e-01
1.23181546e+00 -2.02072322e-01 -7.44709671e-01 4.91456658e-01
3.66532415e-01 -3.89242113e-01 2.13291705e-01 3.94429127e-03
-2.08434775e-01 -1.57639000e-03 -1.21133363e+00 5.22831440e-01
7.34234810e-01 -3.92190218e-01 -3.21823537e-01 5.17456532e-01
7.74882972e-01 5.43050587e-01 -8.18269610e-01 4.01372373e-01
5.75280070e-01 -8.33064735e-01 1.06041849e+00 -7.57825196e-01
3.80950600e-01 1.88786119e-01 5.20231500e-02 -1.61789560e+00
-3.83187830e-01 -6.10279381e-01 1.35134399e-01 9.30275738e-01
5.21933556e-01 -1.13477433e+00 4.73877996e-01 3.07142120e-02
7.09167123e-01 -9.30768907e-01 -1.14621890e+00 -9.39059734e-01
8.60312521e-01 1.77660078e-01 2.75352091e-01 1.06172931e+00
-1.54919118e-01 2.60101467e-01 -3.65173906e-01 2.03535222e-02
6.06305122e-01 -4.09451067e-01 4.84901965e-01 -1.45798552e+00
-4.25218970e-01 -5.48626781e-01 -1.91039890e-01 -4.16851133e-01
-1.94516256e-01 -6.02232695e-01 -4.06503737e-01 -1.38260972e+00
3.59918118e-01 -5.21874249e-01 -1.25184208e-01 -1.90799177e-01
1.83437303e-01 -9.45905689e-03 1.86894797e-02 3.75576794e-01
1.51535004e-01 2.81213284e-01 9.49713349e-01 8.76199082e-02
-3.86882573e-01 1.86865870e-02 -4.75104898e-01 1.04619539e+00
9.96896565e-01 -5.13693988e-01 -1.62190378e-01 -1.14922626e-02
5.61975300e-01 3.37906003e-01 4.24367875e-01 -7.10290313e-01
-2.05174297e-01 -4.56697971e-01 1.41903356e-01 -2.80113101e-01
2.26013005e-01 -8.71249497e-01 5.24649978e-01 1.15066195e+00
3.15273441e-02 2.90054344e-02 6.89143389e-02 3.73752028e-01
3.91759932e-01 -9.00477245e-02 9.14244115e-01 1.09608382e-01
4.45486993e-01 1.95698723e-01 -6.51523352e-01 1.69478655e-01
9.83536661e-01 2.20227301e-01 -5.01495719e-01 -6.09347224e-01
-5.95632374e-01 6.76558986e-02 6.27711892e-01 1.35484919e-01
1.80183202e-02 -1.04565203e+00 -1.02278852e+00 -2.58957185e-02
-5.59244215e-01 -8.79123688e-01 1.07424356e-01 1.09522283e+00
-6.96824372e-01 3.47765148e-01 -1.90890104e-01 -4.47406739e-01
-7.19160497e-01 6.96567595e-01 7.50054717e-01 -6.96080029e-01
-1.89651877e-01 2.84350514e-01 2.83833027e-01 -4.91892070e-01
-2.13167846e-01 2.28369743e-01 -2.71611065e-01 3.71665716e-01
2.85121232e-01 4.17641610e-01 -3.69937360e-01 -7.48275042e-01
-4.45510447e-01 8.97583425e-01 2.11193040e-01 -2.37604845e-02
1.32822657e+00 -1.00459829e-01 -3.54346484e-01 6.26773596e-01
1.13363695e+00 1.31953448e-01 -8.93270612e-01 1.05616093e-01
-4.03197967e-02 -3.12883779e-02 -4.02560800e-01 -7.76073337e-01
-1.00726485e+00 5.47908604e-01 3.88520122e-01 9.91369903e-01
7.71510959e-01 3.42900753e-01 5.32034695e-01 1.04949512e-01
4.07786220e-01 -7.27063298e-01 -1.08624607e-01 4.84387219e-01
1.00838149e+00 -8.91420424e-01 1.35712624e-01 -3.00816447e-01
-8.67506564e-02 6.21526539e-01 5.64195886e-02 -3.24161947e-01
1.03995454e+00 4.03688461e-01 -2.22838983e-01 -1.70282200e-01
-6.32408142e-01 2.33364016e-01 -7.92169720e-02 6.69346333e-01
2.28254050e-01 3.88461202e-01 -8.00612211e-01 8.21141079e-02
-5.89782037e-02 -6.37876913e-02 3.85270625e-01 3.46283078e-01
-6.18977249e-01 -9.50963378e-01 -4.54942524e-01 4.04072195e-01
-4.10527498e-01 -1.76967308e-01 -5.94623029e-01 1.04478455e+00
1.55411899e-01 7.83173382e-01 3.39582413e-01 -7.84046017e-03
3.45997252e-02 1.55815308e-03 4.21875298e-01 -1.20077863e-01
-5.76338291e-01 -3.98925729e-02 -4.15982381e-02 -5.88866100e-02
-2.63865978e-01 -5.08477092e-01 -1.25493383e+00 -1.02643490e+00
-1.61459088e-01 2.97368199e-01 6.05798244e-01 1.04917800e+00
-2.56772321e-02 2.07694322e-01 8.43753099e-01 -9.60706592e-01
-5.53708673e-01 -9.17433918e-01 -7.27113187e-01 2.19750062e-01
4.03585792e-01 -3.80772024e-01 -1.02386558e+00 -4.18273687e-01] | [5.992849826812744, 4.382608890533447] |
58893bca-fa17-4f8d-8701-01695d9876d2 | the-limited-integrator-model-regulator-and | 2304.13161 | null | https://arxiv.org/abs/2304.13161v1 | https://arxiv.org/pdf/2304.13161v1.pdf | The Limited Integrator Model Regulator And its Use in Vehicle Steering Control | Unexpected yaw disturbances like braking on unilaterally icy road, side wind forces and tire rupture are very difficult to handle by the driver of a road vehicle, due to his/her large panic reaction period ranging between 0.5 to 2 seconds. Automatic driver assist systems provide counteracting yaw moments during this driver panic reaction period to maintain the stability of the yaw dynamics of the vehicle. An active steering based driver assist system that uses the model regulator control architecture is introduced and used here for yaw dynamics stabilization in such situations. The model regulator which is a special form of a two degree of freedom control architecture is introduced and explained in detail in a tutorial fashion whereby its integral action capability, among others, is also shown. An auxiliary steering actuation system is assumed and a limited integrator version of the model regulator based steering controller is developed in order not to saturate the auxiliary steering actuator. This low frequency limited integrator implementation also allows the driver to take care of low frequency steering and disturbance rejection tasks. Linear simulation results are used to demonstrate the effectiveness of the proposed method. | ['Levent Guvenc', 'Bilin Aksun-Guvenc'] | 2023-04-25 | null | null | null | null | ['steering-control'] | ['computer-vision'] | [-4.51544635e-02 8.43625784e-01 -2.71688044e-01 -1.17413513e-01
2.17164576e-01 -6.16524756e-01 6.92647219e-01 -5.38654745e-01
-1.29800305e-01 6.88395143e-01 -3.55487391e-02 -7.77940214e-01
-1.91984758e-01 -1.50561318e-01 -9.72387344e-02 -8.19553852e-01
6.55919015e-01 1.06772907e-01 3.29215825e-01 -9.22048986e-01
4.01075751e-01 7.45783746e-01 -1.82059741e+00 -6.76983356e-01
7.60475934e-01 4.65997905e-01 3.29961479e-01 6.88620746e-01
2.25060076e-01 2.32078373e-01 -8.53665322e-02 4.97230321e-01
1.85465202e-01 -3.63883190e-02 -8.05848911e-02 3.83767068e-01
-4.34247404e-03 -4.85708326e-01 -5.85223921e-03 5.19280314e-01
5.02607703e-01 3.46865654e-01 7.77962863e-01 -1.53902936e+00
5.21957099e-01 -3.44792485e-01 -2.52395809e-01 -1.35426745e-01
2.91578304e-02 4.24611062e-01 1.73011988e-01 -8.20058405e-01
5.75572014e-01 1.18708658e+00 2.04858392e-01 7.50313640e-01
-1.15762377e+00 -7.10120440e-01 -4.25153933e-02 9.99170318e-02
-9.55458164e-01 -7.31534600e-01 7.49643266e-01 -4.37178731e-01
7.25570977e-01 3.74374747e-01 3.64302069e-01 6.03819370e-01
7.47404039e-01 2.03267753e-01 1.05309319e+00 -2.99401134e-02
-7.00750053e-02 6.90778852e-01 6.26497805e-01 2.98865914e-01
4.45898980e-01 1.73739210e-01 -3.10857985e-02 5.93180694e-02
4.56440121e-01 -3.18799555e-01 -2.76222825e-01 -5.96122384e-01
-7.48495817e-01 6.35359168e-01 7.80023858e-02 -4.26988661e-01
-6.05231166e-01 -8.19896832e-02 6.68157995e-01 4.19982851e-01
4.60821576e-02 7.51185715e-02 -8.87418091e-02 -3.29653174e-01
-1.57895893e-01 5.70313513e-01 8.98866057e-01 1.03239548e+00
3.37226689e-01 7.36379862e-01 5.20352125e-01 3.52867514e-01
4.22731638e-01 8.82081926e-01 -6.29089773e-02 -7.31463790e-01
2.32005879e-01 4.67239976e-01 7.57753253e-01 -6.59156561e-01
-6.41566575e-01 1.44531101e-01 -5.82471311e-01 7.44113266e-01
1.91465437e-01 -4.57107544e-01 -8.12367082e-01 1.35921156e+00
6.67593241e-01 -6.21469676e-01 3.09735507e-01 7.76056588e-01
4.18312788e-01 8.85581493e-01 -2.79748708e-01 -6.20508492e-01
1.39026546e+00 -3.62407565e-01 -1.43545592e+00 -2.49193609e-01
1.72250986e-01 -9.51179504e-01 8.90147984e-01 4.40573931e-01
-9.05219078e-01 -4.89911020e-01 -1.16822171e+00 1.97452173e-01
-4.74318355e-01 4.19560432e-01 -1.07276559e-01 6.54314399e-01
-7.91650057e-01 -1.59871966e-01 -6.78183138e-01 -3.91124547e-01
-6.72299743e-01 4.59719956e-01 -3.54062229e-01 6.48583472e-01
-1.42851269e+00 1.76370764e+00 -1.01306476e-01 6.23989224e-01
-3.03358763e-01 -6.12986326e-01 -8.73752832e-01 -4.90038067e-01
3.59071821e-01 -3.81440371e-01 1.15864265e+00 -3.74049604e-01
-2.07219338e+00 4.70136762e-01 -6.37478888e-01 -1.35044530e-01
5.76170146e-01 -4.14758116e-01 -5.89193821e-01 -6.50483277e-03
-2.39430964e-01 2.04905570e-01 1.34870994e+00 -9.72629845e-01
-4.22680676e-01 -1.37227923e-02 -3.37081879e-01 3.06008250e-01
3.73861969e-01 -2.18450829e-01 4.01826710e-01 7.60822818e-02
-1.17059402e-01 -1.35458541e+00 -1.34285808e-01 -2.97733158e-01
-3.68195921e-01 -2.13474840e-01 1.72198355e+00 -1.59310028e-02
1.14585900e+00 -2.07354140e+00 -2.07099766e-01 3.96495849e-01
-9.97050107e-02 5.13668537e-01 3.96510214e-01 8.85949194e-01
-1.83460563e-01 -5.11460006e-01 3.17128897e-01 3.91904175e-01
-1.45627066e-01 2.69255728e-01 -7.87708461e-01 6.59270167e-01
5.60797989e-01 1.07440613e-01 -2.07122833e-01 3.64582211e-01
4.19267416e-01 5.96798658e-01 2.69475561e-02 2.24364251e-01
6.48696303e-01 2.45461226e-01 -5.12645423e-01 4.28683907e-01
8.25912178e-01 6.11852050e-01 -3.70317847e-02 -4.10719216e-01
-9.71855938e-01 3.17222565e-01 -1.51244700e+00 3.53648663e-02
-4.44042742e-01 9.12554443e-01 8.79924595e-01 -8.15895855e-01
1.29248250e+00 5.56032121e-01 -6.58147782e-02 -5.16847134e-01
3.03115785e-01 5.18778682e-01 3.10887903e-01 -7.48791575e-01
6.29961789e-01 -2.77201682e-01 2.02943701e-02 3.49137932e-02
-6.01831734e-01 -6.74000025e-01 -5.43205775e-02 2.36190241e-02
3.10079992e-01 -1.00534000e-01 4.66649711e-01 -7.44432986e-01
1.14743698e+00 1.32946998e-01 6.50288522e-01 -1.00793503e-01
-4.52905893e-01 -2.48422071e-01 6.34216011e-01 -3.45113009e-01
-1.26701486e+00 -5.89141726e-01 -2.62082666e-01 6.24302387e-01
4.47964877e-01 1.04038134e-01 -3.54975492e-01 4.47022825e-01
3.23660880e-01 8.49821329e-01 -2.78665572e-01 -5.35061002e-01
-8.52827013e-01 -6.29671961e-02 1.46603957e-01 4.83524173e-01
1.50349498e-01 -4.11198080e-01 -1.17301941e+00 2.52877057e-01
3.40884477e-01 -8.12377334e-01 -2.86775619e-01 1.09788820e-01
-5.47947168e-01 -1.11044300e+00 -1.34462461e-01 -4.35996652e-01
7.77732193e-01 6.11713529e-01 1.93694100e-01 -2.38293067e-01
-1.58310145e-01 3.47362906e-01 1.82290196e-01 -9.48348999e-01
-5.07348895e-01 -2.33751744e-01 6.83972239e-01 -1.37092501e-01
2.22901538e-01 9.53852013e-02 -3.83218914e-01 9.74359751e-01
-4.31100160e-01 1.63361028e-01 2.09683552e-01 8.66923630e-01
-1.10188544e-01 -2.81103045e-01 1.31136382e+00 -4.02066857e-01
9.16466236e-01 -6.67823255e-02 -1.22590590e+00 -4.98488635e-01
-8.56106460e-01 -7.75003508e-02 5.74747741e-01 -1.58106729e-01
-1.32398927e+00 1.88357100e-01 5.73096201e-02 5.18966652e-02
-1.53373241e-01 -1.83764659e-02 -2.51844943e-01 -1.64816141e-01
3.05619985e-01 -1.81844812e-02 1.35131240e+00 -1.66499600e-01
2.89626896e-01 8.91042173e-01 5.14826179e-01 2.66559780e-01
1.08985984e+00 2.04267055e-01 6.77951694e-01 -1.20837021e+00
1.31044343e-01 -5.70859909e-01 -4.35502172e-01 -6.58791065e-01
6.58197999e-01 -8.97803783e-01 -1.31765234e+00 4.40991044e-01
-8.87622178e-01 -6.94370270e-02 1.84021726e-01 5.37763000e-01
-7.64219761e-01 -1.63333356e-01 -5.93613200e-02 -1.51385927e+00
-3.17833126e-01 -1.35457146e+00 5.96255422e-01 5.12728870e-01
-4.53661621e-01 -8.15825462e-01 -3.12761396e-01 1.06123634e-01
9.17018712e-01 3.88871521e-01 6.81034386e-01 -2.70332024e-02
-1.74883947e-01 -6.07265234e-01 3.07445407e-01 2.21335068e-01
4.16483842e-02 4.26359862e-01 -8.38434756e-01 -4.22637999e-01
5.88907897e-02 -2.51210690e-01 2.48237118e-01 3.64577830e-01
-1.40038282e-01 -2.15750575e-01 -5.16853273e-01 -1.04215607e-01
1.24154890e+00 5.55691361e-01 4.60376084e-01 3.44528675e-01
3.64338845e-01 1.21383786e+00 1.44238126e+00 1.13814451e-01
2.27041528e-01 5.98411798e-01 6.93105102e-01 -2.36398697e-01
5.57032287e-01 3.66149783e-01 6.22195601e-01 3.27677250e-01
-9.59709734e-02 1.82577983e-01 -6.28667653e-01 5.05856454e-01
-1.58879244e+00 -6.82402432e-01 -9.18388844e-01 2.37581325e+00
4.84986633e-01 3.43517452e-01 1.50137976e-01 4.66050059e-01
7.88361728e-01 -1.20389044e-01 -3.69666547e-01 -1.48410666e+00
4.38338161e-01 -4.57758099e-01 1.17531359e+00 1.09423161e+00
-6.93881452e-01 5.27995586e-01 5.74812126e+00 3.60400617e-01
-1.41697156e+00 -5.74725628e-01 -2.08771363e-01 1.16949096e-01
6.17462061e-02 2.62723356e-01 -1.35534537e+00 3.23228151e-01
1.15715921e+00 -6.68773055e-01 -6.60594404e-02 1.09412324e+00
1.32681870e+00 -6.10642254e-01 -5.09185195e-01 3.52681458e-01
-2.59834260e-01 -6.12379491e-01 -5.09044230e-01 1.03337362e-01
1.40545323e-01 -5.48897684e-01 7.12237135e-02 2.57724643e-01
-5.61797202e-01 -6.31986439e-01 3.38946968e-01 5.92490196e-01
5.65080225e-01 -1.32042384e+00 7.31387138e-01 6.77001238e-01
-1.14550591e+00 -1.95911437e-01 -1.90474391e-01 -3.85769218e-01
9.49130058e-01 2.52698094e-01 -1.28476608e+00 2.85298824e-01
-2.91432679e-01 1.30316556e-01 -1.69520937e-02 5.24582684e-01
-5.40277548e-02 3.96294802e-01 -4.81154621e-01 -3.14949185e-01
4.76321012e-01 -3.81696254e-01 1.12661350e+00 8.46329451e-01
-3.13503593e-02 1.50814444e-01 -1.69360772e-01 4.57873255e-01
9.69800293e-01 -3.49963099e-01 -1.13405824e+00 3.23353976e-01
3.35523754e-01 1.51143897e+00 1.77056730e-01 -2.16356471e-01
-8.35713930e-03 8.35775733e-02 -8.93821239e-01 3.10784817e-01
-8.44233155e-01 -1.15897810e+00 1.23718321e+00 8.56043994e-01
-3.04189567e-02 -4.40214217e-01 -6.70893639e-02 -2.63668031e-01
-8.53869170e-02 -2.87979186e-01 -6.26303971e-01 -6.44786179e-01
-3.88711005e-01 2.19655171e-01 3.74269158e-01 -1.68660128e+00
-1.04042006e+00 -6.59472167e-01 -1.05508363e+00 1.45595813e+00
-1.59786701e+00 -7.40711272e-01 -3.76712441e-01 3.72212917e-01
7.30367661e-01 -2.99693018e-01 4.53756392e-01 1.62062775e-02
-5.33832848e-01 2.95882784e-02 -8.42892192e-03 -7.82538772e-01
8.86189699e-01 -9.56781387e-01 -7.36444667e-02 7.77453065e-01
-2.00429964e+00 9.22596872e-01 1.63843763e+00 -5.37361801e-01
-2.10071230e+00 -8.48999500e-01 1.13080812e+00 -1.11589767e-01
6.62551045e-01 -2.99821883e-01 -9.15683925e-01 2.67270982e-01
4.62790966e-01 -3.10983092e-01 1.26550524e-02 -6.90804601e-01
4.82732564e-01 -4.04823244e-01 -9.34698641e-01 6.49546981e-01
-9.80032980e-02 -2.08669469e-01 -6.97439551e-01 -8.89740586e-02
3.23773891e-01 -5.86236060e-01 -4.08927739e-01 4.20977712e-01
6.55640662e-01 -3.69577855e-01 4.28946316e-01 -2.73986965e-01
-4.43907201e-01 -7.64403462e-01 5.35727620e-01 -1.22895765e+00
-4.51388620e-02 -1.23461413e+00 1.92058817e-01 9.29611981e-01
2.89293468e-01 -1.13862801e+00 2.55355299e-01 7.04081893e-01
-1.97144568e-01 -6.13918543e-01 -8.79462481e-01 -7.17076838e-01
-3.74459088e-01 2.20824659e-01 -3.94115567e-01 2.68168986e-01
4.29256111e-01 6.86503887e-01 -5.58683813e-01 4.18709368e-01
3.44424665e-01 -4.79981184e-01 1.33694291e+00 -1.14940917e+00
5.27595580e-01 -3.80652770e-02 -4.48006511e-01 -6.60883248e-01
8.33691210e-02 1.41108736e-01 3.15830082e-01 -1.18422592e+00
-7.89710999e-01 2.19826117e-01 4.95447665e-01 7.71415755e-02
1.95414461e-02 -3.01969022e-01 -1.91813216e-01 -1.60268366e-01
6.54714167e-01 3.52900058e-01 1.10103393e+00 2.75107294e-01
-5.85080266e-01 7.98325300e-01 -3.99898104e-02 5.97778797e-01
8.35083485e-01 -1.59489382e-02 -1.21424890e+00 3.89647871e-01
5.55417649e-02 3.34901750e-01 3.79734218e-01 -6.46549821e-01
1.88047573e-01 -4.03386772e-01 -3.53428572e-01 -1.13094902e+00
4.02023971e-01 -9.94263530e-01 3.21536422e-01 9.15666342e-01
3.07642743e-02 4.24509525e-01 3.29826534e-01 4.88199651e-01
-3.23813617e-01 2.21950978e-01 1.22642171e+00 5.46580136e-01
-5.29081166e-01 -5.47158718e-01 -1.21351755e+00 -8.01125884e-01
1.27157450e+00 -6.67951047e-01 -5.05494475e-01 -7.27876306e-01
-6.46473467e-01 6.16179466e-01 1.29425243e-01 6.09762013e-01
4.84167695e-01 -9.23519969e-01 -2.24578470e-01 3.79736990e-01
-6.83482438e-02 -4.26421642e-01 2.71679789e-01 1.55178046e+00
-3.62769097e-01 1.14288712e+00 -7.01887727e-01 -5.01499116e-01
-1.60806251e+00 2.18184978e-01 3.82870883e-01 5.91417730e-01
-5.76811314e-01 4.72279042e-02 8.98397639e-02 -1.26565337e-01
-1.85913190e-01 -3.63904625e-01 -4.96470720e-01 3.02760750e-01
3.98791671e-01 6.53661013e-01 2.08187595e-01 -6.79752052e-01
-4.90888149e-01 7.00663030e-01 6.18137140e-03 -2.18411922e-01
8.70822668e-01 -5.75961947e-01 3.10820550e-01 5.94744325e-01
8.72547805e-01 -3.75483818e-02 -1.41893244e+00 5.55514216e-01
-3.50076556e-01 -1.32065728e-01 2.07279101e-02 -3.02739799e-01
-4.02407914e-01 6.44264698e-01 5.26864171e-01 3.55272233e-01
9.09990251e-01 -8.44080567e-01 4.48139369e-01 4.13153887e-01
-1.56519309e-01 -1.63493490e+00 -6.26328349e-01 7.19511390e-01
1.05651903e+00 -9.95840609e-01 9.16581750e-02 -6.41411483e-01
-1.02150118e+00 1.35717773e+00 7.81942964e-01 -5.31978488e-01
6.86767161e-01 5.33160329e-01 3.33911717e-01 2.41465464e-01
-1.31150889e+00 -1.80188656e-01 2.60366440e-01 6.66634083e-01
4.77732927e-01 -6.86621591e-02 -1.20327330e+00 1.55357486e-02
2.29524449e-01 -3.43995579e-02 1.16076446e+00 1.04504657e+00
-1.05197215e+00 -8.49361837e-01 -6.22881591e-01 -3.29702273e-02
-1.92927439e-02 6.30636692e-01 -2.00930938e-01 1.52806616e+00
-3.56330156e-01 9.40292299e-01 3.46417576e-01 -2.65533365e-02
1.02507007e+00 2.03075483e-01 -4.60982382e-01 -1.21694736e-01
-3.24286580e-01 4.25458252e-01 6.12580419e-01 -7.17695594e-01
2.28954069e-02 -3.92583042e-01 -1.45375907e+00 -3.12348723e-01
-4.49983686e-01 6.58737272e-02 1.13663173e+00 6.06584609e-01
2.52891511e-01 2.72559851e-01 9.20828938e-01 -6.94023550e-01
-9.03269827e-01 -9.77387071e-01 -5.59423864e-01 -3.09652627e-01
9.34353113e-01 -1.32410097e+00 -6.24054015e-01 -1.14379570e-01] | [5.469031810760498, 1.919940710067749] |
af6d10e1-84dd-4b9f-9ef1-024127c5872f | bayesian-learning-of-feature-spaces-for | 2209.03028 | null | https://arxiv.org/abs/2209.03028v1 | https://arxiv.org/pdf/2209.03028v1.pdf | Bayesian learning of feature spaces for multitasks problems | This paper presents a Bayesian framework to construct non-linear, parsimonious, shallow models for multitask regression. The proposed framework relies on the fact that Random Fourier Features (RFFs) enables the approximation of an RBF kernel by an extreme learning machine whose hidden layer is formed by RFFs. The main idea is to combine both dual views of a same model under a single Bayesian formulation that extends the Sparse Bayesian Extreme Learning Machines to multitask problems. From the kernel methods point of view, the proposed formulation facilitates the introduction of prior domain knowledge through the RBF kernel parameter. From the extreme learning machines perspective, the new formulation helps control overfitting and enables a parsimonious overall model (the models that serve each task share a same set of RFFs selected within the joint Bayesian optimisation). The experimental results show that combining advantages from kernel methods and extreme learning machines within the same framework can lead to significant improvements in the performance achieved by each of these two paradigms independently. | ['Emilio Parrado-Hernández', 'Vanessa Gómez-Verdejo', 'Ascensión Gallardo-Antolín', 'Carlos Sevilla-Salcedo'] | 2022-09-07 | null | null | null | null | ['bayesian-optimisation'] | ['methodology'] | [-1.33768380e-01 2.60542810e-01 -1.29492775e-01 -4.18897182e-01
-5.28947353e-01 -5.75974211e-02 7.13734031e-01 -4.83046323e-01
-1.13581613e-01 6.33041561e-01 3.09531391e-02 5.00670224e-02
-8.90779614e-01 -4.38363612e-01 -4.02773917e-01 -1.12006676e+00
6.31800964e-02 4.19139534e-01 -2.23209970e-02 1.03306167e-01
4.48192805e-01 4.61453676e-01 -1.47169971e+00 2.53379345e-01
1.16180658e+00 1.05179870e+00 4.52073425e-01 2.92195886e-01
-2.50530779e-01 7.33907640e-01 -3.34397592e-02 -3.60123962e-01
1.20549031e-01 3.74783725e-01 -3.03365409e-01 -1.50502518e-01
2.58135766e-01 -1.86628345e-02 -3.47553678e-02 9.22844231e-01
3.32803041e-01 2.39731967e-01 1.35585284e+00 -1.17031550e+00
-5.81798077e-01 4.86386538e-01 -6.15903318e-01 -1.34968251e-01
1.85544103e-01 -1.63035169e-01 9.05903339e-01 -1.19119847e+00
-5.04816212e-02 1.14165747e+00 9.19978380e-01 -8.83404762e-02
-1.62208116e+00 -2.63773710e-01 2.16816828e-01 2.88400978e-01
-1.57943833e+00 -3.91316295e-01 9.37013209e-01 -6.26450002e-01
7.34042466e-01 9.26196873e-02 4.56869900e-01 1.09054518e+00
4.18852657e-01 6.10468090e-01 1.46570361e+00 -6.09558046e-01
7.04436719e-01 3.71907800e-01 4.45189476e-01 6.13809109e-01
2.88997144e-01 1.34218380e-01 -6.71951056e-01 -6.70019984e-01
5.13148367e-01 2.16485560e-01 -1.65090397e-01 -8.85326505e-01
-8.12014997e-01 1.30558133e+00 7.73088336e-02 3.95408213e-01
-5.43347657e-01 2.58973300e-01 1.89314976e-01 6.83010519e-02
6.01378739e-01 1.74879462e-01 -4.36729401e-01 3.65634337e-02
-1.21566069e+00 7.85741135e-02 1.06488526e+00 5.72046280e-01
8.01220834e-01 2.45074034e-01 -1.21547803e-01 8.32902968e-01
8.91349733e-01 2.93905705e-01 2.70058542e-01 -9.23538268e-01
1.02411404e-01 8.93083438e-02 3.13397616e-01 -9.47315872e-01
-3.50880861e-01 -8.22229922e-01 -6.40768051e-01 4.10480201e-01
3.68639678e-01 -1.09958768e-01 -7.00521946e-01 1.62028933e+00
4.55135316e-01 3.33251864e-01 -8.98721740e-02 5.81134677e-01
2.76634455e-01 6.85096085e-01 1.71234205e-01 -3.63500655e-01
1.39419639e+00 -6.19726896e-01 -8.10029984e-01 -1.24003500e-01
3.54069561e-01 -7.45959163e-01 7.71484375e-01 9.15165424e-01
-8.56717050e-01 -7.00872362e-01 -9.89222884e-01 3.77513170e-01
-1.73590913e-01 2.04756126e-01 8.43971610e-01 8.04198146e-01
-8.99611115e-01 4.97278601e-01 -7.12533653e-01 1.69085890e-01
1.83484793e-01 3.48207474e-01 -4.15520892e-02 -5.37589528e-02
-1.07670057e+00 1.20787513e+00 6.09691262e-01 3.76636446e-01
-5.18593132e-01 -6.46262825e-01 -5.83345830e-01 2.66744375e-01
3.57802987e-01 -8.87692392e-01 8.74519229e-01 -9.00013566e-01
-1.82104957e+00 2.88436532e-01 -7.16114342e-02 -3.72946084e-01
5.37325561e-01 -5.49849272e-01 1.24978404e-02 2.68643707e-01
-2.83686280e-01 1.76477104e-01 1.79354012e+00 -1.17068243e+00
-2.09052488e-01 -4.42489237e-01 -4.16068524e-01 -3.15729193e-02
-5.04867971e-01 -1.32103562e-01 1.07051887e-01 -6.37354255e-01
2.84447074e-01 -8.64057720e-01 -2.63313264e-01 -3.04804325e-01
-1.18661046e-01 -5.07773876e-01 6.22836471e-01 -6.43047690e-01
1.20425725e+00 -2.14343643e+00 7.21879601e-01 4.24511254e-01
3.34290951e-01 -2.28519440e-02 3.77012908e-01 3.90092522e-01
-2.03362331e-01 -3.05124521e-01 -1.53334245e-01 -4.43642706e-01
3.59098077e-01 3.95009406e-02 -5.34459889e-01 6.32664680e-01
5.76326298e-03 6.34990335e-01 -5.61739624e-01 -5.79387486e-01
4.48234051e-01 8.68584573e-01 -3.98647130e-01 2.87171364e-01
-1.44053221e-01 1.29567623e-01 -8.23027492e-01 3.39021355e-01
7.88580894e-01 6.18754253e-02 -3.45878713e-02 -2.87748456e-01
-5.09525537e-02 -4.39923495e-01 -1.61561692e+00 1.68889582e+00
-6.55209601e-01 5.79609051e-02 6.57024905e-02 -1.28624630e+00
1.18031538e+00 5.14758885e-01 3.79646331e-01 1.08325526e-01
-4.21167612e-02 3.25430959e-01 -3.89639467e-01 -6.25208974e-01
1.49870142e-01 -4.69550759e-01 -2.24432573e-02 4.00455296e-01
4.68341380e-01 -8.72851629e-03 -3.51663172e-01 -1.62286833e-01
4.61004913e-01 6.29824042e-01 5.59574723e-01 -3.85176718e-01
3.37551147e-01 -5.05870342e-01 4.18308675e-01 1.02950311e+00
-2.49076379e-03 1.96357116e-01 3.77404243e-01 -6.29595399e-01
-8.81167054e-01 -1.41849613e+00 -6.85140073e-01 1.18308723e+00
-6.62555575e-01 -1.59944385e-01 -5.64276278e-01 -3.47503006e-01
2.34375000e-01 9.91294980e-01 -6.97438240e-01 -2.58751512e-01
-8.46577212e-02 -1.14884138e+00 2.09425256e-01 2.15675786e-01
1.90533921e-01 -4.58253354e-01 -5.80216527e-01 7.08117038e-02
-2.75085252e-02 -7.25092709e-01 1.64936736e-01 7.59205341e-01
-1.24559128e+00 -3.76573056e-01 -9.16523337e-01 -3.33152175e-01
1.30071223e-01 -5.53558357e-02 7.87488878e-01 -4.82561409e-01
-2.35752985e-01 4.48451936e-01 -3.24857205e-01 -2.79511541e-01
-2.22822100e-01 -2.65935995e-02 2.90759832e-01 5.22096872e-01
5.18058836e-01 -8.87114525e-01 -2.44157448e-01 2.24963315e-02
-5.32484949e-01 -1.87316999e-01 7.57018864e-01 9.01032746e-01
2.69404382e-01 1.24312989e-01 8.27460170e-01 -5.92420280e-01
4.30828154e-01 -8.18951130e-01 -5.57753086e-01 3.44741046e-01
-7.64275491e-01 4.62719500e-01 4.49074179e-01 -7.46177793e-01
-1.45939600e+00 1.48598984e-01 2.60829002e-01 -4.61430162e-01
-3.62830520e-01 5.22450745e-01 4.44550328e-02 -2.27282465e-01
6.25718832e-01 3.41961741e-01 -1.22406101e-02 -6.45290852e-01
5.96966505e-01 8.87492359e-01 8.00095573e-02 -6.85644388e-01
5.95888495e-01 3.26273888e-01 3.59194875e-01 -9.79570746e-01
-8.79474878e-01 -2.74073392e-01 -9.47012246e-01 -2.57008582e-01
8.23185384e-01 -8.70487630e-01 -7.83984244e-01 3.91596794e-01
-8.91953707e-01 1.99221447e-01 -1.70903981e-01 9.21856403e-01
-1.17815983e+00 6.27667010e-01 -2.37380281e-01 -1.47114503e+00
-1.11751653e-01 -9.46463585e-01 8.93937945e-01 7.42377043e-02
-1.81619853e-01 -1.17680800e+00 2.10956052e-01 4.35624272e-01
3.58451694e-01 6.80470169e-02 1.13386750e+00 -7.81248689e-01
-3.63134682e-01 -3.54401022e-01 3.16920161e-01 3.81778896e-01
-1.56291798e-01 9.23020467e-02 -1.23141265e+00 -2.86874980e-01
7.61395574e-01 -3.98471177e-01 1.13299048e+00 8.31812322e-01
6.16867304e-01 1.19027928e-01 -2.63560474e-01 8.19976032e-01
1.51710677e+00 -2.78379440e-01 4.83111918e-01 1.22334741e-01
4.34694260e-01 7.27830946e-01 4.39048737e-01 6.76748097e-01
-3.00767161e-02 8.17431152e-01 3.79662156e-01 1.45834133e-01
3.30455571e-01 1.89476803e-01 4.61342186e-01 7.64325798e-01
-2.10686848e-01 5.34186006e-01 -7.91022658e-01 1.18035868e-01
-2.05615449e+00 -1.13828397e+00 -5.35172671e-02 2.40346837e+00
6.72583640e-01 1.01915427e-01 6.72316179e-02 3.72903794e-02
8.05291891e-01 6.71533942e-02 -2.26120487e-01 -5.84535599e-01
3.05972785e-01 -2.21584216e-02 -4.10733446e-02 5.22675097e-01
-9.88984227e-01 5.70418715e-01 6.43791056e+00 1.14886165e+00
-6.90940201e-01 9.41625386e-02 3.24786097e-01 -2.11636797e-01
-6.79103732e-02 2.50244051e-01 -9.82763588e-01 5.20107925e-01
1.20215034e+00 2.00429007e-01 4.85762805e-01 1.07573450e+00
3.21909964e-01 -3.37183982e-01 -1.02640581e+00 7.86892295e-01
8.05426314e-02 -8.79676223e-01 3.26792002e-02 2.63771057e-01
4.55632240e-01 -2.40457699e-01 5.41067481e-01 4.17866677e-01
2.09297493e-01 -1.12855601e+00 6.53740823e-01 1.31692052e+00
2.97438860e-01 -9.29291725e-01 5.78952193e-01 6.83019102e-01
-9.30555165e-01 -6.00552797e-01 -7.00173855e-01 -2.69029569e-02
7.69452676e-02 1.06597888e+00 -6.89940631e-01 6.94110751e-01
4.92022485e-01 3.20923567e-01 -2.96893477e-01 9.71541405e-01
-1.75296585e-03 5.15461504e-01 -3.57320815e-01 5.39224818e-02
1.79831043e-01 -6.13466620e-01 6.70405567e-01 1.11807191e+00
3.23245287e-01 -3.70067656e-01 2.91846633e-01 1.31097102e+00
9.47045088e-01 1.07280254e-01 -7.42130280e-01 3.12460393e-01
3.59124243e-01 1.42880857e+00 -6.57467246e-01 -1.15026720e-01
-6.07006669e-01 6.50956511e-01 5.78517437e-01 6.01772368e-01
-8.69430304e-01 -2.56378770e-01 3.09834808e-01 -1.86726496e-01
8.21176291e-01 -2.68760949e-01 -3.85917634e-01 -1.20271599e+00
-1.58100843e-01 -5.51448405e-01 2.30433434e-01 -8.81893277e-01
-1.39403105e+00 3.52708697e-01 4.90458876e-01 -8.65284741e-01
-6.28905535e-01 -6.76396012e-01 -5.86313188e-01 1.21082485e+00
-1.51711953e+00 -1.29859650e+00 1.60301626e-01 5.77316284e-01
2.71303982e-01 -2.90690452e-01 1.07606566e+00 -3.36961746e-01
-4.25568998e-01 2.30268210e-01 5.89929998e-01 -5.57052195e-01
5.65116465e-01 -1.44744670e+00 -3.61647397e-01 3.32720399e-01
1.75435722e-01 9.27733362e-01 8.17398787e-01 -6.83625102e-01
-1.44179308e+00 -5.68893969e-01 5.99392593e-01 -3.45423311e-01
7.18657196e-01 -1.54490158e-01 -7.62825966e-01 4.68732625e-01
-1.24992788e-01 -8.73283297e-02 1.05969405e+00 6.12622201e-01
-3.39286208e-01 2.45287381e-02 -1.03711081e+00 3.27928603e-01
2.42741197e-01 -5.94307899e-01 -1.15268612e+00 1.69462115e-01
1.69633821e-01 4.45395410e-01 -1.24518859e+00 2.19274178e-01
7.07236588e-01 -9.97998118e-01 1.18989944e+00 -7.76544511e-01
-4.37408797e-02 -1.34578511e-01 -4.89248931e-01 -1.29970372e+00
-6.70836985e-01 -6.24177814e-01 -8.20640862e-01 9.70278144e-01
-4.39499766e-02 -6.66374505e-01 5.65633416e-01 4.75509912e-01
-3.70629579e-02 -8.64357591e-01 -1.29130268e+00 -8.69317770e-01
2.81189978e-01 -5.69006979e-01 4.81755380e-03 6.76544666e-01
1.80235133e-02 2.15043843e-01 -6.54513299e-01 2.08200365e-01
1.17503595e+00 2.62189776e-01 4.59730715e-01 -1.84668767e+00
-8.30167651e-01 -1.25509158e-01 -3.43460798e-01 -6.70064926e-01
4.27185178e-01 -9.07123029e-01 -1.83742687e-01 -1.09737468e+00
3.25679749e-01 -3.08102697e-01 -4.70130652e-01 1.03088103e-01
-1.22850761e-01 -6.37407526e-02 1.44428492e-01 3.12132895e-01
-3.01100910e-01 7.75326431e-01 5.70325315e-01 2.18122751e-01
-1.92514747e-01 4.33925450e-01 -5.75934947e-01 1.18656933e+00
4.95755166e-01 -7.36511230e-01 -4.05450404e-01 -8.01054314e-02
2.60049373e-01 6.53562471e-02 4.59855348e-01 -8.44002903e-01
2.16665864e-01 -1.97910488e-01 7.50052869e-01 -2.46441528e-01
6.94986641e-01 -9.58077371e-01 2.13726372e-01 4.94239450e-01
-7.89801255e-02 -5.39488196e-01 -3.19786668e-01 9.62696254e-01
9.66605544e-02 -8.50220203e-01 8.46978784e-01 -2.02492177e-01
-2.63768196e-01 -2.69905686e-01 -4.58142549e-01 -2.78237313e-01
1.16587973e+00 -5.81636190e-01 2.93260396e-01 -2.61207461e-01
-1.37966526e+00 -2.08000302e-01 3.58075090e-02 -1.50229320e-01
5.60987055e-01 -1.15500438e+00 -7.53891706e-01 1.38735026e-01
-1.97268561e-01 -4.45477068e-01 2.60702819e-01 1.10879302e+00
2.11227521e-01 4.29454267e-01 -1.40520647e-01 -8.21168423e-01
-8.63112450e-01 6.29560351e-01 3.98946285e-01 -2.64198512e-01
-5.89079261e-01 6.50982141e-01 1.00042917e-01 -3.08373779e-01
3.84960771e-01 -1.74995176e-02 -2.87708431e-01 3.20089132e-01
5.18885911e-01 7.82145321e-01 -8.70655924e-02 -4.76523817e-01
-9.84238535e-02 7.54077077e-01 -1.56586394e-01 -1.88969716e-01
1.69333899e+00 -1.68581810e-02 -7.13008568e-02 8.92325699e-01
8.87864947e-01 -1.06742799e-01 -1.52888954e+00 -2.90873438e-01
2.72274554e-01 -3.26282769e-01 5.00196934e-01 -6.87923968e-01
-4.53003556e-01 8.77249122e-01 5.21923900e-01 1.81540579e-01
8.08991194e-01 -1.29325399e-02 4.81962711e-02 6.86537206e-01
2.42939517e-01 -1.23743367e+00 1.29990190e-01 3.47921044e-01
8.32216918e-01 -8.46877635e-01 2.32143849e-01 -2.90539742e-01
-4.12044078e-01 1.53628540e+00 1.12487497e-02 -2.71156043e-01
8.82285893e-01 2.28244111e-01 -6.06382608e-01 -1.72532141e-01
-7.07110226e-01 1.14001200e-01 3.47670764e-01 7.81466365e-01
2.68634021e-01 1.71473380e-02 -1.78535163e-01 8.73149693e-01
2.75795937e-01 1.14936478e-01 3.13806772e-01 7.64024794e-01
-8.70332539e-01 -8.21719646e-01 -7.37730503e-01 4.02885318e-01
-3.96488994e-01 -2.20094070e-01 2.34745845e-01 4.41137850e-01
-9.29169655e-02 7.50292182e-01 -2.84972250e-01 -7.40218014e-02
-1.80487290e-01 8.45954597e-01 6.41124547e-01 -5.97942650e-01
-2.79501528e-01 3.46082211e-01 -1.83348820e-01 -2.60548741e-01
-1.42878473e-01 -7.79761136e-01 -7.19977438e-01 4.85679246e-02
-6.57248497e-01 3.78803164e-01 9.77756441e-01 1.08276582e+00
2.50299573e-01 1.47248060e-01 7.89802134e-01 -1.07882738e+00
-1.56465459e+00 -1.15777934e+00 -1.02172303e+00 -1.71307296e-01
1.36070833e-01 -1.16594160e+00 -6.99540198e-01 -1.15885809e-01] | [7.485313415527344, 4.03417444229126] |
899c2201-64a1-439e-a3df-e72000c06c68 | mirror-mirror-on-the-wall-whos-got-the | 1805.11589 | null | http://arxiv.org/abs/1805.11589v2 | http://arxiv.org/pdf/1805.11589v2.pdf | Mirror, Mirror, on the Wall, Who's Got the Clearest Image of Them All? - A Tailored Approach to Single Image Reflection Removal | Removing reflection artefacts from a single image is a problem of both
theoretical and practical interest, which still presents challenges because of
the massively ill-posed nature of the problem. In this work, we propose a
technique based on a novel optimisation problem. Firstly, we introduce a simple
user interaction scheme, which helps minimise information loss in
reflection-free regions. Secondly, we introduce an $H^2$ fidelity term, which
preserves fine detail while enforcing global colour similarity. We show that
this combination allows us to mitigate some major drawbacks of the existing
methods for reflection removal. We demonstrate, through numerical and visual
experiments, that our method is able to outperform the state-of-the-art methods
and compete with recent deep-learning approaches. | ['Dong-Dong Chen', 'Sabine Süsstrunk', 'Carola-Bibiane Schönlieb', 'Angelica I. Aviles-Rivero', 'Qingnan Fan', 'Georg Maierhofer', 'Daniel Heydecker'] | 2018-05-29 | null | null | null | null | ['reflection-removal'] | ['computer-vision'] | [ 5.85272133e-01 4.67961282e-02 6.45419717e-01 1.18536189e-01
-8.12679946e-01 -9.07938927e-02 6.12714648e-01 7.97340274e-02
-5.06031692e-01 6.53175831e-01 1.99689697e-02 -1.75500046e-02
-3.57443631e-01 -6.42957211e-01 -4.08099473e-01 -9.07138169e-01
-3.34210023e-02 -2.71059632e-01 3.48230124e-01 -3.02143693e-01
3.93074065e-01 6.87159777e-01 -1.61768115e+00 1.33139372e-01
9.01350319e-01 1.16009760e+00 1.43272489e-01 4.21088696e-01
1.41977608e-01 5.68171203e-01 -3.58370930e-01 -2.79268593e-01
4.06509131e-01 -6.08930886e-01 -6.85691655e-01 2.16101155e-01
4.88987058e-01 -2.58302744e-02 -1.02658145e-01 8.84638488e-01
7.22832441e-01 4.37913418e-01 6.06600046e-01 -6.21373117e-01
-2.44986236e-01 -3.74260306e-01 -8.06005657e-01 9.81233343e-02
3.46074760e-01 -1.11673482e-01 7.70746112e-01 -8.37113917e-01
4.63886529e-01 8.85301709e-01 8.77200127e-01 3.18573594e-01
-1.38270533e+00 -2.90261775e-01 -3.87881114e-03 1.68545157e-01
-1.40236330e+00 -6.37943804e-01 1.17645717e+00 -2.94433475e-01
8.03704441e-01 4.36768591e-01 5.43904722e-01 6.20952785e-01
2.90745556e-01 4.89764869e-01 1.36042321e+00 -8.94455075e-01
2.40370736e-01 6.04857691e-02 -3.97626340e-01 7.61249959e-01
-4.42809016e-02 2.95653880e-01 -5.08590817e-01 -2.19183154e-02
9.10915017e-01 -1.94181740e-01 -6.67399466e-01 -7.50938118e-01
-9.12844360e-01 8.63044620e-01 4.82747704e-01 2.88073003e-01
-3.18586946e-01 1.19663544e-01 1.15701277e-02 2.14346156e-01
7.40649641e-01 4.01195884e-01 -1.49270803e-01 8.74903947e-02
-1.02921915e+00 1.16050608e-01 5.94549894e-01 3.34417880e-01
7.73834705e-01 -5.97199388e-02 1.51033834e-01 1.00753260e+00
5.77863812e-01 2.51834959e-01 7.85908103e-03 -9.32229400e-01
-6.68882430e-02 4.25182909e-01 1.40845969e-01 -1.13150609e+00
-5.13327658e-01 -4.82920140e-01 -1.13086009e+00 9.07140732e-01
2.90809482e-01 1.61698759e-01 -8.06728125e-01 1.35445988e+00
3.89375895e-01 1.45631447e-01 -1.57344081e-02 9.09453034e-01
6.85810149e-01 4.06726509e-01 -3.60753536e-01 -4.75125581e-01
1.11059654e+00 -9.28512573e-01 -6.37019575e-01 -1.64437637e-01
1.13552652e-01 -1.01592028e+00 6.61465943e-01 8.23277891e-01
-1.38821268e+00 -3.83137196e-01 -1.04247880e+00 -1.66082252e-02
-3.21746916e-02 -8.01013336e-02 4.49213862e-01 8.45755458e-01
-1.14799607e+00 7.91859210e-01 -5.81258059e-01 -1.60479322e-01
3.30387026e-01 2.64772564e-01 -3.54968548e-01 -1.61480501e-01
-8.22543621e-01 9.41545725e-01 -6.25431836e-02 4.56426501e-01
-3.00330102e-01 -6.60767317e-01 -8.90007555e-01 -6.66123405e-02
4.62992340e-01 -6.32041633e-01 9.80752170e-01 -7.73895621e-01
-1.85403073e+00 8.70546043e-01 1.35429855e-03 -1.24122478e-01
8.22501481e-01 -3.90178502e-01 -2.48519391e-01 4.60330814e-01
-4.55858976e-01 9.59097669e-02 1.14633918e+00 -1.72694850e+00
-4.31430757e-01 -1.30955324e-01 4.04154174e-02 2.22364902e-01
-1.76811889e-01 5.21553084e-02 -5.29939771e-01 -7.53821611e-01
1.37841657e-01 -7.60544598e-01 -6.47804558e-01 3.62912655e-01
-2.11392298e-01 3.81880075e-01 5.75337589e-01 -5.63358247e-01
1.06304967e+00 -2.07356787e+00 2.74938703e-01 4.12513971e-01
3.62293661e-01 5.89981258e-01 6.90441355e-02 6.52856052e-01
-9.06220376e-02 -2.60845214e-01 -7.68313229e-01 -6.01639926e-01
-1.57417804e-01 -2.73234621e-02 2.41703279e-02 8.26238215e-01
2.44095460e-01 6.00181162e-01 -7.12435305e-01 -3.53689671e-01
6.37096345e-01 1.14975381e+00 -4.12384838e-01 1.26938581e-01
1.32484794e-01 6.85461283e-01 -1.81003734e-01 2.24441022e-01
9.94712949e-01 -2.45162528e-02 1.35208175e-01 -1.91818684e-01
-4.35653239e-01 -5.26744314e-02 -1.47966290e+00 1.56878686e+00
-7.60643840e-01 4.92584825e-01 4.15764123e-01 -1.35196340e+00
8.50697935e-01 1.39566705e-01 5.99631965e-01 -1.12589371e+00
1.56227633e-01 2.72754163e-01 -2.11490974e-01 -3.19410592e-01
2.14093432e-01 -5.47706962e-01 2.37968221e-01 4.40459639e-01
-2.76362240e-01 -2.48693034e-01 -1.97235405e-01 -1.64119169e-01
1.17022479e+00 1.20462991e-01 4.30801481e-01 -3.58335882e-01
8.94903779e-01 -5.38820624e-01 2.86917716e-01 7.90846348e-01
-6.38868511e-02 1.21855521e+00 2.40341261e-01 -4.25029308e-01
-7.32801020e-01 -9.23873603e-01 -2.78933406e-01 5.20945430e-01
2.91236967e-01 -1.13890231e-01 -6.10846460e-01 -5.10360599e-01
-3.24989647e-01 2.70833075e-01 -6.09962285e-01 -1.83351770e-01
-8.53785813e-01 -8.45357180e-01 -8.91960636e-02 3.03884685e-01
5.36537886e-01 -1.08884382e+00 -1.03429508e+00 2.94250995e-01
-1.39884412e-01 -1.18679380e+00 2.91783269e-02 7.46467337e-02
-8.72729003e-01 -1.02941382e+00 -1.08815479e+00 -5.56550205e-01
5.90974808e-01 5.51692843e-01 1.33197331e+00 5.93119740e-01
-6.72237635e-01 4.62229759e-01 -4.41540331e-01 -1.55363977e-01
-2.27852955e-01 -3.12837780e-01 -5.05746067e-01 1.07165731e-01
-1.15104474e-01 -6.53034866e-01 -9.40264285e-01 2.94998080e-01
-1.14014125e+00 1.42019302e-01 7.94080317e-01 8.64335716e-01
5.32750845e-01 3.48666996e-01 1.34703249e-01 -7.34119177e-01
3.60915720e-01 3.89009938e-02 -5.71914136e-01 -1.44137479e-02
-4.90140229e-01 5.96146472e-02 6.04030013e-01 6.61138892e-02
-1.19311953e+00 2.21776247e-01 -3.59800667e-01 -3.35041247e-02
-1.73007280e-01 2.45349586e-01 -3.96017581e-02 -8.11942875e-01
3.56420606e-01 1.94176450e-01 2.31397096e-02 -7.58051634e-01
2.60448605e-01 2.49807939e-01 3.93299997e-01 -3.07538033e-01
9.20765579e-01 1.04772711e+00 4.87956852e-01 -1.27487445e+00
-7.94797599e-01 -7.41112649e-01 -7.85805762e-01 -3.16759586e-01
5.25776803e-01 -5.12800515e-01 -7.18367815e-01 6.18944824e-01
-1.07983685e+00 -3.87455344e-01 -3.17202061e-01 2.56339610e-01
-8.40776801e-01 7.53943861e-01 -3.92546952e-01 -1.05807936e+00
-2.58676440e-01 -9.56338108e-01 1.10212004e+00 4.99712341e-02
2.76397645e-01 -1.21740198e+00 3.24059337e-01 3.34247530e-01
5.76595247e-01 5.41182339e-01 4.94314164e-01 2.35815868e-01
-7.04729259e-01 7.67329857e-02 -5.73628724e-01 5.30118227e-01
1.52187701e-02 -2.74702519e-01 -1.10090935e+00 -4.73252267e-01
3.47222507e-01 -1.25250146e-01 1.25154400e+00 3.05535942e-01
9.77494478e-01 -2.71403566e-02 -1.83430329e-01 7.67904580e-01
1.74429584e+00 -1.64161697e-01 1.12386453e+00 5.34381866e-01
5.80973506e-01 7.25364983e-01 4.81619924e-01 4.06137466e-01
4.27802093e-02 9.59560931e-01 5.75988233e-01 -7.53559887e-01
-3.60744834e-01 4.15731400e-01 -1.12208962e-01 5.92033982e-01
-4.79736745e-01 -2.24319667e-01 -5.60601711e-01 4.49069887e-01
-1.94941735e+00 -7.29027987e-01 -4.18622792e-01 2.50271440e+00
6.05486155e-01 4.31146957e-02 -3.97562981e-02 6.82547629e-01
2.20520630e-01 2.82315433e-01 2.74309311e-02 -4.36179847e-01
3.34928855e-02 6.87967122e-01 3.23380291e-01 7.45474160e-01
-1.21092772e+00 3.74085039e-01 6.71851730e+00 6.86662972e-01
-9.00425196e-01 6.91841124e-03 3.36209625e-01 1.57532275e-01
-2.10500151e-01 -2.21960217e-01 -3.13527554e-01 3.08206171e-01
5.46450794e-01 3.99130940e-01 2.52543628e-01 9.47802067e-02
1.96205378e-01 -3.96494806e-01 -7.10162044e-01 8.79590988e-01
4.62064922e-01 -1.00132680e+00 -4.10286099e-01 1.39556676e-01
6.67858660e-01 -2.43156999e-01 3.68436836e-02 -2.71188170e-01
-8.97100195e-02 -1.14064479e+00 3.76271248e-01 7.36219406e-01
6.05553985e-01 -9.54110384e-01 6.38127029e-01 1.60897121e-01
-1.11172259e+00 1.14291929e-01 -2.09699020e-01 -9.39155445e-02
3.81152332e-01 9.97397959e-01 -1.08622156e-01 9.11965549e-01
7.89401531e-01 4.55128133e-01 -3.32112908e-01 1.45404565e+00
-2.82279879e-01 6.03685714e-02 -4.11131084e-01 2.87350982e-01
2.25720465e-01 -3.96383166e-01 5.46405256e-01 1.38305128e+00
3.20758581e-01 1.40829608e-01 -5.47036417e-02 6.08339071e-01
2.57250488e-01 9.69542563e-02 -4.43435967e-01 5.86823940e-01
-4.18442070e-01 1.37511492e+00 -9.14853513e-01 8.02976117e-02
-5.86413145e-01 1.27495039e+00 3.01999211e-01 3.45472217e-01
-4.04755443e-01 -4.46431488e-01 4.79232222e-01 2.74279773e-01
6.53249502e-01 -3.81519675e-01 -1.78486988e-01 -1.04872394e+00
3.95728290e-01 -6.45465314e-01 2.59395331e-01 -5.54807305e-01
-1.17781794e+00 4.97191370e-01 -1.90551296e-01 -1.23493600e+00
9.93242115e-02 -7.62075841e-01 -5.42426646e-01 8.21009755e-01
-2.33723068e+00 -1.05869818e+00 -5.07832229e-01 5.00920653e-01
2.05104262e-01 3.54826570e-01 8.42643142e-01 5.23420215e-01
-1.60918728e-01 3.50913793e-01 4.35096651e-01 -1.88447103e-01
4.68098640e-01 -1.24742711e+00 3.41571361e-01 8.95824492e-01
9.94531289e-02 4.08845395e-01 9.14180517e-01 -1.94683954e-01
-1.23223042e+00 -7.02856004e-01 8.13929141e-01 -1.30922928e-01
3.97521824e-01 -3.27381313e-01 -1.06089938e+00 1.79062963e-01
3.02011520e-01 2.53735185e-01 6.79524124e-01 -1.40895680e-01
-2.41633028e-01 7.57708400e-02 -1.26986456e+00 3.81992996e-01
8.63343894e-01 -2.63279617e-01 -3.87294650e-01 2.44135767e-01
9.54836681e-02 -2.90351957e-01 -6.83039546e-01 4.95328575e-01
6.69838548e-01 -1.65783441e+00 1.42892730e+00 8.89181644e-02
3.26692849e-01 -2.03976497e-01 1.29498690e-01 -1.47924781e+00
-2.99510777e-01 -9.57839429e-01 -2.75143117e-01 8.73555720e-01
5.01114912e-02 -6.67488873e-01 6.17332757e-01 3.50988507e-01
-4.75297421e-02 -6.99174464e-01 -1.00991499e+00 -7.94856787e-01
2.90946458e-02 -2.78555870e-01 -6.79295957e-02 6.81831419e-01
-2.41110563e-01 6.55360594e-02 -5.78042626e-01 6.45517483e-02
1.06000614e+00 1.07987329e-01 4.97426957e-01 -1.41991413e+00
-4.33479965e-01 -5.23978531e-01 -4.45908248e-01 -9.88415539e-01
-2.33101472e-01 -2.46544212e-01 3.57526869e-01 -1.68051243e+00
1.36607558e-01 -5.41488469e-01 -5.25866270e-01 8.50972086e-02
-2.36335963e-01 9.65094626e-01 1.07233860e-01 -8.83423388e-02
-5.55213630e-01 7.24185646e-01 1.22839272e+00 2.34447837e-01
-1.90714970e-01 3.15611064e-02 -6.40584290e-01 8.83829772e-01
6.05020225e-01 -2.94217616e-01 -1.17349714e-01 -3.27607185e-01
3.36375922e-01 -2.51135051e-01 6.69471741e-01 -1.11738896e+00
-3.26215435e-04 1.12826847e-01 3.44153225e-01 -3.71674538e-01
7.31867254e-01 -1.02121210e+00 -1.28866225e-01 3.94030631e-01
5.18797487e-02 -4.28179681e-01 2.67427057e-01 6.11673176e-01
-4.98372018e-02 -3.38577300e-01 1.35559022e+00 -1.83432922e-01
-3.65538180e-01 -1.02758303e-01 -3.38077128e-01 -1.26796886e-01
8.09406459e-01 -2.45419264e-01 -1.77084096e-02 -3.79087687e-01
-4.47632968e-01 -1.59104586e-01 5.71023583e-01 8.17778930e-02
8.30352604e-01 -1.11008692e+00 -6.75679028e-01 3.86894733e-01
7.32622743e-02 -2.88917478e-02 4.76959944e-01 1.17070186e+00
-6.32412732e-01 2.73233056e-02 -8.98038561e-04 -3.94346595e-01
-1.38818395e+00 4.77898657e-01 3.62913907e-01 -3.15522194e-01
-1.21517849e+00 8.11241031e-01 2.94909775e-01 -2.92754114e-01
1.53840855e-01 -5.72466152e-03 -2.95211434e-01 -1.59115404e-01
7.46951878e-01 4.58412737e-01 4.70235348e-01 -7.68339157e-01
-3.74591976e-01 1.11769867e+00 6.22616522e-02 -1.73687339e-01
1.58062518e+00 -3.51081848e-01 3.98339611e-03 8.47626776e-02
1.24775970e+00 1.07118212e-01 -1.39657331e+00 -3.05441856e-01
-2.09961221e-01 -9.36078608e-01 4.80148733e-01 -7.24254668e-01
-1.19781041e+00 1.05460405e+00 6.49017453e-01 5.46253443e-01
1.46227419e+00 -2.49242917e-01 8.97157907e-01 9.07253101e-02
8.99613723e-02 -1.02888596e+00 6.30624518e-02 2.29100138e-01
8.92202377e-01 -1.41714025e+00 4.65815276e-01 -6.53913140e-01
-1.45487830e-01 1.11759806e+00 8.95777419e-02 -2.38270104e-01
7.66479433e-01 3.04607540e-01 2.78699487e-01 -2.96313792e-01
-3.56488913e-01 -4.43618506e-01 5.28955042e-01 7.78184295e-01
4.46745872e-01 -4.41707462e-01 -4.71982002e-01 -1.46302208e-01
2.60545999e-01 -3.23402956e-02 4.28204536e-01 1.07703161e+00
-4.45801497e-01 -1.16171014e+00 -5.57043195e-01 7.55243003e-02
-5.69236040e-01 -7.15511590e-02 -3.00246358e-01 7.38597155e-01
-5.32907099e-02 9.96020377e-01 -1.94129154e-01 1.62083849e-01
4.45660204e-01 -4.45285499e-01 7.36261547e-01 -3.11025023e-01
-4.52390581e-01 4.90032256e-01 -3.40984017e-02 -7.76431978e-01
-8.37037563e-01 -5.90349913e-01 -9.08206582e-01 7.62590393e-02
-4.06419545e-01 -1.09705009e-01 6.79195881e-01 9.40331817e-01
1.28726855e-01 6.55218303e-01 8.22382152e-01 -1.28313553e+00
-2.22367555e-01 -5.99022746e-01 -6.30486012e-01 4.71734017e-01
6.74402297e-01 -8.85714829e-01 -6.75777078e-01 -9.75499898e-02] | [10.813300132751465, -2.650067090988159] |
2ded49e2-193d-4d3a-9209-5691c271687e | cross-lingual-training-of-dense-retrievers | null | null | https://aclanthology.org/2021.mrl-1.24 | https://aclanthology.org/2021.mrl-1.24.pdf | Cross-Lingual Training of Dense Retrievers for Document Retrieval | Dense retrieval has shown great success for passage ranking in English. However, its effectiveness for non-English languages remains unexplored due to limitation in training resources. In this work, we explore different transfer techniques for document ranking from English annotations to non-English languages. Our experiments reveal that zero-shot model-based transfer using mBERT improves search quality. We find that weakly-supervised target language transfer is competitive compared to generation-based target language transfer, which requires translation models. | ['Jimmy Lin', 'He Bai', 'Rui Zhang', 'Peng Shi'] | null | null | null | null | emnlp-mrl-2021-11 | ['document-ranking', 'passage-ranking'] | ['natural-language-processing', 'natural-language-processing'] | [-2.07501173e-01 -2.72195309e-01 -6.61344707e-01 -5.62924445e-02
-1.83806992e+00 -6.87083423e-01 9.38722968e-01 2.17112582e-02
-9.61249590e-01 1.20691335e+00 7.45426834e-01 -4.16035056e-01
3.04192510e-02 -7.58663356e-01 -7.01006055e-01 -6.03117887e-03
1.11377306e-01 8.68426204e-01 5.24759352e-01 -7.51138270e-01
2.98562944e-01 -8.53674188e-02 -7.70994842e-01 6.88075125e-01
9.15920317e-01 6.20283820e-02 2.67390043e-01 8.34465086e-01
-2.87184775e-01 7.79393375e-01 -5.41053295e-01 -3.05720448e-01
-1.62542447e-01 -6.54831588e-01 -1.26288199e+00 -9.24438417e-01
3.27756792e-01 -3.51917297e-01 -5.37953317e-01 9.35847342e-01
9.35289383e-01 5.10236382e-01 9.38775003e-01 -4.64308977e-01
-1.30126023e+00 6.76895797e-01 -2.44327530e-01 4.87771928e-01
6.70267880e-01 -5.93414128e-01 1.20164227e+00 -1.26749599e+00
1.03368068e+00 1.29639316e+00 5.50084591e-01 9.18110967e-01
-8.44308615e-01 -3.90560746e-01 -2.46736363e-01 2.56459534e-01
-1.31509733e+00 -3.97615343e-01 3.43867660e-01 -7.41471127e-02
1.52233315e+00 2.92558640e-01 2.13667721e-01 1.30159545e+00
1.00753129e-01 9.22870219e-01 1.01425719e+00 -1.06928933e+00
-2.37051502e-01 3.76086056e-01 9.44976285e-02 5.73972940e-01
1.63401276e-01 -2.92880763e-03 -7.58317471e-01 -1.74904734e-01
4.41785872e-01 -2.69970298e-01 -4.27023023e-01 -3.48696932e-02
-1.03585601e+00 9.64716196e-01 4.46465254e-01 4.60921913e-01
-8.33113026e-03 5.56259267e-02 6.83580160e-01 7.75499821e-01
1.00968242e+00 8.25579166e-01 -7.70634592e-01 -3.39867353e-01
-8.98748159e-01 -4.22267132e-02 8.58675003e-01 1.20094657e+00
5.88301063e-01 -2.46887803e-01 -6.97189569e-01 1.35449994e+00
1.37848288e-01 6.51950479e-01 8.76895547e-01 -3.45883161e-01
5.86694539e-01 1.07913502e-01 2.55007505e-01 -4.05017972e-01
7.99973533e-02 -3.93614382e-01 -5.11692107e-01 -5.51326156e-01
-1.08905874e-01 -6.21061660e-02 -9.97583032e-01 1.55323100e+00
-2.05543622e-01 -2.89427698e-01 5.57268083e-01 8.02758932e-01
1.04427862e+00 1.05326414e+00 2.11468920e-01 -1.88263692e-02
1.10728443e+00 -1.45605171e+00 -5.38579047e-01 -7.67414570e-02
1.00055110e+00 -1.07531679e+00 1.49044168e+00 -2.76220381e-01
-1.04312348e+00 -4.52773660e-01 -5.15510857e-01 -5.21321177e-01
-6.48987114e-01 3.21836889e-01 4.48099583e-01 4.56536442e-01
-1.28546274e+00 3.10140550e-01 -6.76771641e-01 -1.01007819e+00
-1.18309027e-02 1.41292796e-01 -1.90395057e-01 -5.69016457e-01
-1.79629672e+00 1.23892355e+00 4.65248108e-01 -3.36308032e-01
-1.01908588e+00 -5.14785588e-01 -8.41593325e-01 1.56633444e-02
-4.68473397e-02 -8.51797521e-01 1.56626058e+00 -3.68249297e-01
-1.56966853e+00 8.45156431e-01 -2.46656954e-01 -4.20944750e-01
2.80789405e-01 -4.59782243e-01 -3.84518117e-01 8.84039775e-02
3.37660491e-01 7.10820615e-01 3.86757642e-01 -8.95810783e-01
-5.88322759e-01 4.40978110e-02 2.22170055e-01 6.16878927e-01
-9.69028175e-01 4.87970948e-01 -8.69903564e-01 -5.86842358e-01
-4.71364528e-01 -7.27285147e-01 -5.65674417e-02 -6.14465773e-01
-2.97420584e-02 -6.22313797e-01 5.61237335e-01 -7.94128358e-01
1.45103741e+00 -1.56841004e+00 2.31141746e-02 -1.62126243e-01
-4.08134311e-01 4.59276527e-01 -4.69062269e-01 9.61496353e-01
4.74164695e-01 3.87255758e-01 1.78845122e-01 -1.17318265e-01
-1.16780840e-01 -2.77978610e-02 -5.15913904e-01 -6.62899390e-02
5.03102876e-02 1.28893840e+00 -1.44712007e+00 -8.16281736e-01
-1.28142238e-01 4.15918797e-01 -4.70245630e-01 1.53937340e-01
2.85751354e-02 2.62871273e-02 -8.11073005e-01 7.63244152e-01
8.04875568e-02 -9.99742001e-02 -9.41906497e-02 2.02317253e-01
3.15371528e-02 8.30573857e-01 -3.36218141e-02 2.06244278e+00
-8.73710573e-01 6.23682618e-01 -5.18661320e-01 -4.75753307e-01
6.72079980e-01 5.69767594e-01 -1.33408774e-02 -9.31534111e-01
-2.46711895e-01 6.45814359e-01 -1.62099823e-01 -1.93830624e-01
9.95121300e-01 -1.83046013e-01 -2.84237564e-01 4.02277380e-01
3.53121966e-01 -2.76021749e-01 5.11467159e-01 5.53309739e-01
1.27321243e+00 4.60882753e-01 1.39891341e-01 -2.31525809e-01
2.61417747e-01 4.05439705e-01 3.23039363e-03 1.04296601e+00
8.54348242e-02 4.99640882e-01 -3.34087282e-01 1.00394577e-01
-9.84431326e-01 -9.80701566e-01 -8.54172036e-02 1.56985772e+00
9.24466103e-02 -5.17023325e-01 -5.44405580e-01 -7.73751199e-01
-2.33987793e-01 8.63963723e-01 -4.92345899e-01 -5.65880418e-01
-5.53564906e-01 -4.95367378e-01 7.66998112e-01 5.38914084e-01
2.19141260e-01 -1.32839239e+00 -1.69708803e-01 3.36770535e-01
-4.84595239e-01 -9.26027894e-01 -9.31454241e-01 6.11167699e-02
-9.73770142e-01 -5.39460003e-01 -1.53631830e+00 -1.15178275e+00
5.87312818e-01 4.84940171e-01 1.37909627e+00 4.46389988e-02
-7.67384395e-02 8.94904077e-01 -8.36151540e-01 -3.92363876e-01
-3.53081226e-01 7.74812222e-01 -1.88534278e-02 -8.96735430e-01
7.50304103e-01 1.34313479e-01 -4.54542965e-01 8.41882527e-02
-7.39878237e-01 -3.63569170e-01 5.88253319e-01 1.05307889e+00
3.53747070e-01 -5.89359879e-01 6.45574212e-01 -7.50048161e-01
1.43982530e+00 -3.88817698e-01 -7.35387504e-02 8.05099189e-01
-6.70215547e-01 4.86358553e-02 3.78207088e-01 -4.41000015e-01
-1.23239112e+00 -7.11910427e-01 3.64114121e-02 -1.09931670e-01
3.26281428e-01 9.67192590e-01 5.11281669e-01 4.76687476e-02
9.51036751e-01 1.47523090e-01 -5.77351272e-01 -6.35666847e-01
4.41030651e-01 8.49013805e-01 1.89952515e-02 -8.87964964e-01
4.90037501e-01 7.56935030e-02 -6.19743168e-01 -8.02289367e-01
-1.01119733e+00 -8.69257927e-01 -4.64288920e-01 -6.16970509e-02
7.68406808e-01 -1.15489066e+00 3.82971853e-01 -1.74970686e-01
-1.47473145e+00 -4.15839672e-01 -4.85727519e-01 8.37994933e-01
-2.39056736e-01 1.75780013e-01 -1.34011209e+00 -4.73435700e-01
-9.32584763e-01 -7.32422411e-01 1.19396317e+00 -3.01762372e-02
-1.61742792e-01 -1.17112231e+00 9.28379953e-01 2.10858852e-01
5.03082514e-01 -7.56837666e-01 7.62837231e-01 -6.64939940e-01
-3.14810872e-01 -6.35419488e-01 -2.07463413e-01 1.63676351e-01
1.01361603e-01 -3.19035828e-01 -5.76796174e-01 -6.25572383e-01
-6.25575542e-01 -1.00202382e+00 1.16825724e+00 2.43062750e-01
5.83577096e-01 3.25837694e-02 -5.04916072e-01 5.19416481e-03
1.28449523e+00 1.16994912e-02 6.33773625e-01 6.07280552e-01
4.92678821e-01 3.52501243e-01 8.82690728e-01 -1.69612497e-01
3.19162279e-01 7.17734754e-01 -6.16152167e-01 -3.19253504e-01
-5.59967101e-01 -6.90398872e-01 5.76883733e-01 1.53788900e+00
-2.88767010e-01 -6.26719475e-01 -9.18004870e-01 8.15534115e-01
-1.72956264e+00 -9.39557731e-01 2.70523518e-01 2.32268214e+00
1.14133275e+00 -3.76157500e-02 -1.85484201e-01 -8.39222848e-01
5.70688963e-01 -1.46227255e-01 -1.93064183e-01 -4.66983676e-01
-1.64635301e-01 4.82287347e-01 3.22704881e-01 7.21252501e-01
-8.79003227e-01 1.67184842e+00 6.85632896e+00 1.11921191e+00
-7.79149532e-01 6.76624119e-01 2.74341673e-01 1.01086274e-01
-5.10416865e-01 1.01994872e-01 -1.05304909e+00 -4.01010364e-02
9.64152992e-01 -7.55066574e-01 1.41898133e-02 7.52190173e-01
-1.57666102e-01 3.12968314e-01 -9.26948607e-01 4.95094091e-01
4.02279973e-01 -9.81014907e-01 4.31578636e-01 -1.08828433e-01
1.04934084e+00 5.29497385e-01 -1.15960643e-01 1.15740740e+00
6.12807214e-01 -7.62710512e-01 2.87725002e-01 4.23624337e-01
1.07529628e+00 -6.37945354e-01 7.87977695e-01 3.26212078e-01
-9.83183742e-01 3.91540676e-01 -9.19304192e-01 4.39403579e-02
1.66752592e-01 8.67351331e-03 -9.05303240e-01 5.45474350e-01
6.49683595e-01 6.45867169e-01 -3.94348651e-01 1.05417049e+00
-3.79786223e-01 8.13627601e-01 -4.80325036e-02 -5.40758729e-01
4.47585106e-01 -8.62459987e-02 5.01316488e-01 1.75774646e+00
6.14550948e-01 -1.07910827e-01 3.22898090e-01 4.85046506e-01
-5.24197698e-01 6.50915623e-01 -1.00656223e+00 -2.23995805e-01
1.57644391e-01 1.02195358e+00 -3.75092417e-01 -7.12047279e-01
-4.11467284e-01 1.51999581e+00 7.76216388e-01 6.54472053e-01
-3.77082676e-01 -5.44350505e-01 9.36540589e-03 -1.94161713e-01
7.37148672e-02 -1.16677999e-01 3.67601991e-01 -1.54024351e+00
-1.51331946e-01 -6.89567029e-01 6.16778910e-01 -7.64996231e-01
-1.34530950e+00 7.66877174e-01 8.83936808e-02 -1.39194942e+00
-4.92402613e-01 -3.57702821e-01 -3.47534269e-01 1.03185117e+00
-1.74495506e+00 -1.28518867e+00 3.57824653e-01 4.47185487e-01
1.15408361e+00 -4.07018751e-01 1.32264304e+00 4.60010052e-01
-1.59453720e-01 7.31750548e-01 6.05137229e-01 2.16466337e-01
1.22233391e+00 -1.17977309e+00 2.95927763e-01 6.71591818e-01
5.33005416e-01 1.00938463e+00 3.61293823e-01 -7.77772784e-01
-1.21415365e+00 -9.12428856e-01 1.55288506e+00 -6.20746195e-01
7.63637424e-01 -1.46599188e-01 -7.41914392e-01 5.40552616e-01
6.97000742e-01 -1.70661539e-01 7.96202540e-01 5.71343720e-01
-3.83627504e-01 1.63909718e-01 -6.14472806e-01 7.57300913e-01
8.87883604e-01 -1.07932329e+00 -9.09574211e-01 6.83769643e-01
9.21390414e-01 -2.41764858e-01 -6.83532417e-01 4.61332351e-01
4.29658324e-01 7.15719461e-02 1.09456611e+00 -1.23851788e+00
7.13191688e-01 1.45757152e-02 -1.51831612e-01 -1.67819273e+00
-4.11230236e-01 -2.74350226e-01 -1.24350362e-01 1.16119564e+00
8.92229676e-01 -3.34069282e-01 4.00623858e-01 2.37974331e-01
-2.32782587e-01 -4.42834914e-01 -8.42687368e-01 -9.76359963e-01
6.80792332e-01 -1.40452776e-02 -3.09514180e-02 1.03637004e+00
3.01736772e-01 1.00781000e+00 -4.80794489e-01 -3.78681153e-01
1.81618601e-01 7.01198354e-02 5.19565880e-01 -7.96344697e-01
-2.75271595e-01 -3.42794985e-01 8.53762180e-02 -1.16604400e+00
4.59635586e-01 -1.28297329e+00 3.06115389e-01 -2.06318402e+00
7.35981047e-01 -1.15185902e-01 -7.73381650e-01 3.90033126e-01
-4.68319386e-01 5.57680488e-01 -7.46440217e-02 4.06845003e-01
-1.08303463e+00 8.39226663e-01 1.19751024e+00 -3.88577640e-01
-1.28581896e-01 -2.34418064e-01 -3.93011808e-01 2.63761371e-01
6.90505028e-01 -6.12532139e-01 -4.51553106e-01 -9.83747959e-01
2.45131791e-01 1.10065937e-02 -2.76712626e-01 -6.57662928e-01
1.56876206e-01 2.13411823e-01 1.40461788e-01 -2.99106508e-01
2.02263594e-01 -2.43063897e-01 -7.12075710e-01 3.92575920e-01
-8.65800142e-01 1.67561635e-01 2.97936261e-01 5.14475942e-01
-5.72392344e-01 -5.50682843e-01 8.67946669e-02 -4.29423690e-01
-8.07046950e-01 2.09662363e-01 -5.01123607e-01 4.91587222e-01
2.94050664e-01 1.58660814e-01 -2.43901968e-01 -6.66899145e-01
-2.76253402e-01 1.53948560e-01 2.25866988e-01 7.56545007e-01
5.79663634e-01 -1.45246494e+00 -1.14392734e+00 -5.73551118e-01
5.38199604e-01 -6.62638307e-01 -8.03648755e-02 4.78685558e-01
-3.34502429e-01 1.12364495e+00 1.53087601e-01 -1.27145812e-01
-1.33175945e+00 2.83706695e-01 -4.59494032e-02 -9.02484536e-01
-3.77949357e-01 9.14415479e-01 -1.08718529e-01 -8.99509728e-01
4.38672751e-02 1.80449530e-01 -3.45225185e-01 -1.69616282e-01
5.04655421e-01 2.69224644e-01 2.39590317e-01 -3.59826088e-01
-1.66339055e-01 5.06611705e-01 -6.26277745e-01 -7.12632656e-01
9.88512337e-01 -1.51309997e-01 -1.79452911e-01 6.36659682e-01
1.39739358e+00 1.66127667e-01 -2.44847327e-01 -3.99312496e-01
3.26335430e-01 -2.82517880e-01 2.50183016e-01 -9.47170317e-01
-2.82627136e-01 1.00720668e+00 4.78728771e-01 -2.70401150e-01
8.82544518e-01 1.20744362e-01 9.55439627e-01 1.05109429e+00
8.95189285e-01 -1.30852032e+00 9.00145099e-02 1.08367229e+00
9.43039894e-01 -1.34407449e+00 -1.07609876e-01 3.26658823e-02
-5.35846353e-01 8.49265873e-01 7.00287044e-01 1.72895208e-01
4.54248011e-01 -3.70175034e-01 1.54299572e-01 -1.16521508e-01
-8.23089898e-01 -4.52446997e-01 8.76802742e-01 3.67791653e-01
1.20207119e+00 1.87636341e-03 -8.80766153e-01 9.98953432e-02
-9.37714893e-03 2.43492052e-01 5.81333600e-02 1.15156102e+00
-3.73550028e-01 -1.55851066e+00 -6.94223195e-02 5.08189619e-01
-5.46894252e-01 -9.30172145e-01 -5.67836285e-01 6.66933537e-01
-5.60425997e-01 8.42809200e-01 -1.71970963e-01 -1.16996370e-01
3.47262084e-01 1.96251601e-01 6.07085168e-01 -1.14499772e+00
-6.05712235e-01 3.22907954e-01 4.14878994e-01 -1.09453730e-01
-3.60604048e-01 -3.40152234e-01 -8.35492134e-01 2.39255533e-01
-6.96822226e-01 9.09613252e-01 5.18601239e-01 6.32381618e-01
3.49318922e-01 3.91319811e-01 3.21587145e-01 -4.20448154e-01
-6.84974253e-01 -1.57182765e+00 -2.62543261e-01 2.61722118e-01
-6.23582453e-02 -2.72856861e-01 -1.24396294e-01 -1.33125082e-01] | [11.399682998657227, 9.728362083435059] |
ee57be7e-36a4-448e-ae98-913dd4b2b3e5 | emotion-recognition-by-fusing-time | 2010.14102 | null | https://arxiv.org/abs/2010.14102v2 | https://arxiv.org/pdf/2010.14102v2.pdf | Emotion recognition by fusing time synchronous and time asynchronous representations | In this paper, a novel two-branch neural network model structure is proposed for multimodal emotion recognition, which consists of a time synchronous branch (TSB) and a time asynchronous branch (TAB). To capture correlations between each word and its acoustic realisation, the TSB combines speech and text modalities at each input window frame and then does pooling across time to form a single embedding vector. The TAB, by contrast, provides cross-utterance information by integrating sentence text embeddings from a number of context utterances into another embedding vector. The final emotion classification uses both the TSB and the TAB embeddings. Experimental results on the IEMOCAP dataset demonstrate that the two-branch structure achieves state-of-the-art results in 4-way classification with all common test setups. When using automatic speech recognition (ASR) output instead of manually transcribed reference text, it is shown that the cross-utterance information considerably improves the robustness against ASR errors. Furthermore, by incorporating an extra class for all the other emotions, the final 5-way classification system with ASR hypotheses can be viewed as a prototype for more realistic emotion recognition systems. | ['Philip C. Woodland', 'Chao Zhang', 'Wen Wu'] | 2020-10-27 | null | null | null | null | ['multimodal-emotion-recognition', 'multimodal-emotion-recognition'] | ['computer-vision', 'speech'] | [ 2.05103800e-01 -3.84992622e-02 1.77569330e-01 -7.24308491e-01
-1.02157891e+00 -4.22362417e-01 6.88977420e-01 9.47897285e-02
-6.05940044e-01 1.04998879e-01 1.83936879e-01 -1.39361188e-01
2.34765008e-01 -1.49950460e-01 -2.96587557e-01 -7.63657510e-01
-6.05739616e-02 1.44495562e-01 -3.84206064e-02 -2.26716682e-01
-3.40007722e-01 3.20517242e-01 -1.52062857e+00 5.58796883e-01
3.44120204e-01 1.58310640e+00 -7.90509060e-02 1.04682457e+00
-2.42820919e-01 8.22068155e-01 -8.42056274e-01 -5.11763871e-01
-7.61147067e-02 -2.82749593e-01 -5.92285931e-01 2.94571240e-02
5.52891158e-02 -9.83004868e-02 -4.76251602e-01 6.17133141e-01
7.64400840e-01 4.89379942e-01 3.61855060e-01 -1.34140432e+00
-1.85282052e-01 5.47238827e-01 -7.45798573e-02 -3.66570391e-02
4.53721017e-01 -2.97528476e-01 9.95976985e-01 -1.19700837e+00
2.56264001e-01 1.14750624e+00 6.43325090e-01 6.75546825e-01
-1.14725912e+00 -4.58417416e-01 2.68853128e-01 3.01626891e-01
-1.19268346e+00 -7.16144145e-01 9.05241728e-01 -1.13512181e-01
1.68488955e+00 5.54237187e-01 4.74654675e-01 1.61817122e+00
1.60123333e-01 1.04126108e+00 8.98059666e-01 -5.76463044e-01
3.21792811e-01 4.31888372e-01 3.92768860e-01 2.87616223e-01
-8.80730033e-01 -3.16486061e-02 -8.10861409e-01 -1.77520722e-01
2.56934240e-02 -2.49335572e-01 -3.42815310e-01 -5.99500723e-03
-9.27220821e-01 6.56802118e-01 -1.51204085e-02 5.17901480e-01
-4.36217606e-01 9.65216160e-02 9.77809072e-01 6.28947854e-01
6.50512636e-01 8.48680884e-02 -7.60109842e-01 -6.80181623e-01
-9.36508954e-01 -3.92087936e-01 9.25903440e-01 6.13985956e-01
3.22650224e-01 4.48957473e-01 -7.84347728e-02 1.37945604e+00
3.44561726e-01 4.64342535e-01 9.16600525e-01 -5.90280354e-01
4.81753647e-01 2.05184042e-01 -1.67533159e-01 -8.61636937e-01
-5.96025586e-01 -1.55119732e-01 -6.82448208e-01 -2.09221393e-01
8.83359089e-02 -3.59893054e-01 -8.97949636e-01 1.88405049e+00
1.01701155e-01 1.89695895e-01 5.38370848e-01 7.38823295e-01
9.95605350e-01 1.23083448e+00 -7.70195425e-02 -2.48762712e-01
1.55446517e+00 -1.13173461e+00 -1.17110479e+00 -2.05670089e-01
6.56114161e-01 -7.43847251e-01 7.91261494e-01 6.25109434e-01
-9.77388859e-01 -7.19002366e-01 -1.16222548e+00 -1.54804841e-01
-9.35120285e-01 3.91226768e-01 1.88713908e-01 7.82262146e-01
-1.17430985e+00 1.55726478e-01 -9.29249287e-01 -4.11812454e-01
-4.17661577e-01 3.08132708e-01 -6.02615952e-01 3.19743097e-01
-1.66310251e+00 1.03924453e+00 1.78401262e-01 4.97721344e-01
-4.65875357e-01 -2.66046643e-01 -1.33705461e+00 1.57058656e-01
1.09524131e-02 -3.91406603e-02 1.43522358e+00 -1.05767751e+00
-2.15883851e+00 5.63965321e-01 -5.38811088e-01 -3.40063632e-01
1.28677264e-02 -1.43506244e-01 -1.06226802e+00 3.11559975e-01
-5.48185706e-01 5.93399644e-01 1.02533913e+00 -1.11520374e+00
-4.37713295e-01 -4.11891967e-01 -4.23662931e-01 2.75319278e-01
-5.98779500e-01 2.90515214e-01 -5.44708848e-01 -5.96026957e-01
-7.77462423e-02 -8.49658132e-01 1.60537705e-01 -5.23589492e-01
-1.81318671e-01 -3.12603623e-01 1.11319065e+00 -7.02468753e-01
1.59954798e+00 -2.41754341e+00 3.25355738e-01 1.87431768e-01
-1.97486967e-01 1.39662400e-01 -4.35618192e-01 6.58152997e-01
-5.60281813e-01 -3.28843519e-02 4.04181071e-02 -1.11867929e+00
3.83615732e-01 4.47006702e-01 -4.66835350e-01 1.88751876e-01
4.95932102e-01 7.51297057e-01 -4.68149483e-01 -1.27831295e-01
3.19442362e-01 8.78211141e-01 -1.12165786e-01 3.32597166e-01
2.85477847e-01 -6.84239417e-02 7.60218874e-02 3.26434672e-01
3.08389515e-01 2.73420870e-01 1.70523569e-01 -3.45653862e-01
-5.17970845e-02 4.25146371e-01 -1.09218705e+00 1.57477057e+00
-7.58469522e-01 1.10120344e+00 2.41280794e-01 -1.05075479e+00
1.10140204e+00 1.09945726e+00 2.97996283e-01 -7.62675881e-01
4.28951055e-01 5.10655828e-02 -1.27597839e-01 -5.08994281e-01
7.79593408e-01 -1.73693866e-01 -5.98473966e-01 3.01203758e-01
6.74376190e-01 -1.33885473e-01 -8.06664377e-02 7.36942701e-03
9.06616986e-01 -2.98243701e-01 1.00053228e-01 3.60051185e-01
5.41803122e-01 -6.64239347e-01 4.48148906e-01 5.50290763e-01
-3.97250235e-01 6.88757420e-01 4.66823459e-01 -2.73849517e-01
-8.14669430e-01 -9.92083311e-01 -3.46472114e-02 1.17749631e+00
-2.11572155e-01 -6.17315173e-01 -5.73317766e-01 -5.24555206e-01
-3.20235729e-01 8.97072494e-01 -7.44700193e-01 -1.91508263e-01
-2.75116414e-01 -5.14897585e-01 8.78516316e-01 7.92248309e-01
1.02720350e-01 -9.61915910e-01 -6.00520790e-01 3.91861618e-01
-4.83463466e-01 -1.45792794e+00 -4.77461845e-01 8.25778008e-01
-4.08953726e-01 -3.72400671e-01 -5.37605226e-01 -6.07179582e-01
5.48702180e-02 -7.87272155e-02 6.71629250e-01 -4.13168699e-01
-5.26591912e-02 8.67361248e-01 -7.94797957e-01 -4.34146136e-01
-4.00540918e-01 -3.18849653e-01 1.40344650e-01 6.11895859e-01
5.69294870e-01 -3.93177688e-01 -3.79793979e-02 3.32844675e-01
-9.94567454e-01 -3.71203691e-01 1.57822475e-01 1.21123815e+00
2.49364629e-01 1.15507483e-01 6.38730764e-01 -1.02170430e-01
7.55392015e-01 -2.64490008e-01 -1.20879233e-01 1.67973995e-01
-1.25863433e-01 -1.65928721e-01 5.44067919e-01 -8.07693720e-01
-1.15061879e+00 -6.81992993e-02 -4.47050989e-01 -5.79065859e-01
-3.52776229e-01 5.66473067e-01 -2.82724053e-01 4.37407196e-01
1.54997841e-01 4.09329832e-01 5.78560643e-02 -2.85514474e-01
4.18599427e-01 1.26351964e+00 4.50516224e-01 -3.73966187e-01
1.77059337e-01 3.82881090e-02 -6.71771705e-01 -1.34980154e+00
-3.27027828e-01 -7.37845182e-01 -4.81671363e-01 -5.42651594e-01
1.02977860e+00 -9.12735879e-01 -5.16702414e-01 6.10861957e-01
-1.46401477e+00 -3.19096386e-01 -2.52609938e-01 7.77909994e-01
-5.51942647e-01 2.55011976e-01 -9.37863171e-01 -1.21710920e+00
-1.97562158e-01 -1.40384841e+00 1.49287057e+00 1.45872489e-01
-5.69666207e-01 -1.14677489e+00 2.05763225e-02 1.83553666e-01
4.00099218e-01 -1.74675807e-01 6.56564355e-01 -1.14560390e+00
3.13160449e-01 -6.16035104e-01 2.13844851e-01 8.80414665e-01
-1.13661319e-01 6.51428178e-02 -1.56991720e+00 -1.71390310e-01
4.00314540e-01 -4.92354274e-01 6.40039563e-01 1.65159866e-01
7.88076162e-01 4.09543402e-02 6.97929636e-02 7.63115212e-02
1.03199840e+00 7.19709516e-01 6.88088238e-01 -8.77430812e-02
2.76191592e-01 6.40873432e-01 3.92903984e-01 4.05244291e-01
3.39209884e-01 7.84849226e-01 9.98229906e-02 -2.59452343e-01
3.45676392e-01 2.50177085e-01 9.44866717e-01 1.60672641e+00
4.37870115e-01 -5.57676852e-01 -7.84423470e-01 5.37384450e-01
-1.68662596e+00 -9.91653502e-01 1.95560411e-01 1.99077654e+00
6.71511233e-01 -1.82592717e-03 -1.25327259e-01 4.52761173e-01
4.80567575e-01 4.13391024e-01 -2.17490464e-01 -1.16580129e+00
-2.41772115e-01 2.77306587e-01 -1.41795561e-01 5.67690253e-01
-1.08908653e+00 7.14541256e-01 6.16718149e+00 8.22271943e-01
-1.34174085e+00 3.44422847e-01 4.85694587e-01 -1.68581098e-01
7.64975399e-02 -4.55751777e-01 -5.90130031e-01 3.24907333e-01
1.60013974e+00 2.59710044e-01 3.22392702e-01 5.59059441e-01
1.80731654e-01 -5.89009412e-02 -1.30722904e+00 1.24634242e+00
4.45283443e-01 -5.75826764e-01 -2.38172844e-01 -3.37951720e-01
1.46247238e-01 2.50699501e-02 1.45110071e-01 7.31094301e-01
-3.05117637e-01 -9.04538155e-01 8.67479324e-01 4.32462215e-01
7.30334461e-01 -7.85536587e-01 1.00849998e+00 9.84554961e-02
-1.40106404e+00 -1.76458091e-01 9.84803215e-02 1.08005494e-01
4.01356071e-01 2.86163807e-01 -6.16471291e-01 7.96160460e-01
6.34279072e-01 4.80034173e-01 -3.79616559e-01 3.99129748e-01
1.02536045e-01 6.17534280e-01 -3.75103027e-01 -1.68849468e-01
3.44382674e-01 2.92908065e-02 6.21882796e-01 1.78594279e+00
2.42184222e-01 -2.16333736e-02 -8.60961154e-03 2.94872314e-01
7.51744956e-02 8.01298097e-02 -4.79253441e-01 -3.69678944e-01
4.70221788e-01 1.27830911e+00 -4.70798999e-01 -4.04856533e-01
-4.14271563e-01 1.34385478e+00 2.00275376e-01 7.04183161e-01
-8.67358625e-01 -8.70875061e-01 8.08044374e-01 -8.91318798e-01
4.08259481e-01 -2.25186795e-01 4.43132594e-02 -1.10127783e+00
1.13976792e-01 -8.16249669e-01 2.71282524e-01 -1.12583399e+00
-1.26268518e+00 1.06902301e+00 -2.55038559e-01 -1.01035523e+00
-3.14913809e-01 -9.05539036e-01 -4.81111914e-01 8.13792527e-01
-1.15902436e+00 -8.76519740e-01 1.32215858e-01 4.35108423e-01
8.63542736e-01 -1.86702713e-01 1.34863198e+00 3.25169265e-01
-8.36745560e-01 8.85047257e-01 1.11672744e-01 1.57945290e-01
8.32864165e-01 -1.35436749e+00 1.11149643e-02 5.52747488e-01
2.44623229e-01 4.16568667e-01 6.53653085e-01 -4.18296047e-02
-1.39938593e+00 -8.08509588e-01 1.04317522e+00 -4.74250883e-01
7.82407939e-01 -7.85752475e-01 -9.03182626e-01 6.96225047e-01
6.40605092e-01 -2.80963987e-01 1.09612310e+00 2.63799399e-01
-6.09299481e-01 -3.25427473e-01 -7.46713161e-01 5.08526742e-01
1.36138663e-01 -1.18197024e+00 -8.96744311e-01 -1.61108702e-01
1.05384290e+00 -3.06994975e-01 -1.13448322e+00 4.12913173e-01
7.62099504e-01 -7.71420956e-01 6.54451489e-01 -4.62267011e-01
2.26202846e-01 2.01994160e-04 -7.38512993e-01 -1.56173122e+00
2.74420917e-01 -6.82342529e-01 -7.19379857e-02 1.40921354e+00
6.09713078e-01 -6.51520848e-01 1.09575480e-01 5.39282560e-01
-2.08550096e-01 -8.47678900e-01 -1.38081729e+00 -7.43462205e-01
-2.19672397e-01 -1.14876437e+00 2.41341472e-01 9.57098961e-01
5.92292726e-01 6.54520333e-01 -4.36318368e-01 3.01969677e-01
-6.95806667e-02 -3.54312152e-01 5.07077098e-01 -7.15332210e-01
-2.20522925e-01 -5.40488839e-01 -6.05994999e-01 -1.21152854e+00
3.38573664e-01 -6.44773185e-01 3.75289947e-01 -9.98846412e-01
-2.83555746e-01 1.34371519e-01 -5.18567026e-01 4.35195386e-01
9.50462297e-02 8.04391503e-02 2.16767445e-01 -5.15566111e-01
-5.53152978e-01 1.01989603e+00 4.34881479e-01 -8.49190950e-02
-2.05000639e-01 -3.54109854e-01 -6.41989931e-02 4.82034832e-01
5.29078722e-01 -2.92657584e-01 -3.03427726e-01 -2.35816896e-01
-1.51099607e-01 3.69016409e-01 1.12298690e-01 -8.12336445e-01
4.03845787e-01 4.47818041e-01 2.14873880e-01 -7.56467044e-01
1.25546384e+00 -1.10702550e+00 -3.20926383e-02 -9.35909674e-02
-4.65557724e-01 1.13252506e-01 6.78072155e-01 5.66469133e-01
-7.05544531e-01 -3.58476155e-02 5.87327838e-01 5.59516728e-01
-4.31972116e-01 -1.72927976e-01 -9.85978723e-01 -4.19316560e-01
6.37895882e-01 -2.64336199e-01 -7.98024014e-02 -5.93367279e-01
-1.16601157e+00 3.11409950e-01 -3.10390383e-01 8.04893374e-01
7.68818676e-01 -1.48691452e+00 -3.37746978e-01 3.47355813e-01
2.19638392e-01 -5.43074846e-01 4.46952999e-01 9.54077303e-01
3.47675800e-01 4.74686474e-01 2.50737846e-01 -7.62539387e-01
-1.54287684e+00 1.55802518e-01 5.00310004e-01 -2.36492828e-01
-2.41703257e-01 8.60289991e-01 6.16729036e-02 -8.03222418e-01
7.44430602e-01 -3.39641482e-01 -2.23069474e-01 5.79604149e-01
6.78752840e-01 7.22194985e-02 3.32109809e-01 -1.03365862e+00
-3.37674350e-01 4.18139070e-01 2.17670598e-03 -7.51538694e-01
1.17991269e+00 -4.41587478e-01 -1.48567716e-02 1.43124163e+00
1.66068888e+00 -1.49463832e-01 -6.99087322e-01 -2.17904925e-01
-2.98348032e-02 1.50585130e-01 1.90460443e-01 -9.32621837e-01
-7.54962087e-01 1.19845998e+00 6.82079136e-01 5.81819773e-01
1.48956430e+00 -1.97771162e-01 6.99954629e-01 4.63233382e-01
-6.72915503e-02 -1.38355303e+00 1.97562948e-01 9.25573766e-01
1.06907177e+00 -9.85196352e-01 -8.67187917e-01 6.51877373e-02
-9.09513593e-01 1.45116127e+00 2.73949802e-01 3.08972061e-01
7.62286901e-01 4.56694901e-01 4.54830468e-01 -7.65234903e-02
-1.19934666e+00 5.25965840e-02 3.31951797e-01 2.57206261e-01
3.63146633e-01 1.91349480e-02 9.18040872e-02 1.08310318e+00
-9.16781873e-02 -5.58153808e-01 4.25540239e-01 8.10039401e-01
-1.67233840e-01 -1.05130994e+00 -3.79836768e-01 5.87515421e-02
-4.05363262e-01 -4.74441461e-02 -4.30083901e-01 7.50630140e-01
-2.42307931e-01 1.36727333e+00 3.27879936e-01 -8.57242703e-01
5.89677155e-01 9.64431107e-01 1.84430823e-01 -3.26177329e-01
-7.92370498e-01 4.31810409e-01 4.32228416e-01 -7.46229589e-01
-3.19211721e-01 -4.71113116e-01 -1.14531982e+00 3.50079000e-01
-5.66662669e-01 3.46513510e-01 1.12503040e+00 9.07430589e-01
4.38215703e-01 8.62011731e-01 8.04411888e-01 -1.01660991e+00
-4.97103989e-01 -1.27642858e+00 -6.48159683e-01 2.36177966e-01
6.24831438e-01 -4.59843546e-01 -5.20447016e-01 4.82035987e-02] | [13.574153900146484, 5.720552444458008] |
095764c7-abce-4bf5-926e-bb65ede6d503 | robust-one-class-classification-with-signed | 2303.01978 | null | https://arxiv.org/abs/2303.01978v1 | https://arxiv.org/pdf/2303.01978v1.pdf | Robust One-Class Classification with Signed Distance Function using 1-Lipschitz Neural Networks | We propose a new method, dubbed One Class Signed Distance Function (OCSDF), to perform One Class Classification (OCC) by provably learning the Signed Distance Function (SDF) to the boundary of the support of any distribution. The distance to the support can be interpreted as a normality score, and its approximation using 1-Lipschitz neural networks provides robustness bounds against l2 adversarial attacks, an under-explored weakness of deep learning-based OCC algorithms. As a result, OCSDF comes with a new metric, certified AUROC, that can be computed at the same cost as any classical AUROC. We show that OCSDF is competitive against concurrent methods on tabular and image data while being way more robust to adversarial attacks, illustrating its theoretical properties. Finally, as exploratory research perspectives, we theoretically and empirically show how OCSDF connects OCC with image generation and implicit neural surface parametrization. Our code is available at https://github.com/Algue-Rythme/OneClassMetricLearning | ['Andres Troya-Galvis', 'Quentin Vincenot', 'Mathieu Serrurier', 'Guillaume Coiffier', 'Thibaut Boissin', 'Paul Novello', 'Louis Bethune'] | 2023-01-26 | null | null | null | null | ['one-class-classification'] | ['miscellaneous'] | [ 2.61947691e-01 4.66446996e-01 -2.65244961e-01 -4.51231688e-01
-1.27784753e+00 -1.13737297e+00 7.56787598e-01 -2.03332044e-02
-3.05459917e-01 8.15560699e-01 -1.74270704e-01 -4.25897956e-01
-2.47451812e-01 -7.84842730e-01 -1.33263469e+00 -7.09230542e-01
-4.61342961e-01 3.54250461e-01 7.73288831e-02 3.79651897e-02
2.95699924e-01 8.32021058e-01 -1.24170876e+00 1.83591053e-01
6.99480951e-01 1.31770647e+00 -7.90072381e-01 7.48930752e-01
2.60394573e-01 4.78847444e-01 -6.11664176e-01 -6.58024073e-01
7.43539333e-01 -9.66578573e-02 -8.36222231e-01 -4.78200734e-01
1.03962064e+00 -1.83835447e-01 -2.82335520e-01 1.31087792e+00
4.47363079e-01 -1.94854885e-01 8.36095870e-01 -1.66591907e+00
-6.93007588e-01 4.86158580e-01 -4.05080110e-01 -1.68883264e-01
2.01310337e-01 8.28159600e-02 1.12150979e+00 -7.87222326e-01
6.19211853e-01 1.20121467e+00 8.50763977e-01 5.39536178e-01
-1.16276872e+00 -5.92475593e-01 -1.59958825e-01 4.14808504e-02
-1.23206306e+00 -2.45873168e-01 5.83883703e-01 -3.14969987e-01
1.20783053e-01 5.43768227e-01 2.03122973e-01 1.26836491e+00
2.36901805e-01 9.52475846e-01 1.34563923e+00 -2.84598053e-01
3.52531016e-01 1.85194895e-01 -8.59939754e-02 6.10115170e-01
3.99295568e-01 3.42646718e-01 -2.37292022e-01 -3.64116311e-01
4.98570502e-01 -3.82199734e-01 -3.80174607e-01 -7.40637302e-01
-7.84391105e-01 9.02915537e-01 4.03008491e-01 7.72909522e-02
3.57532442e-01 4.18232620e-01 5.13113976e-01 6.11676037e-01
5.35418689e-01 3.59812677e-01 -5.34648895e-01 3.73687893e-02
-5.67984760e-01 3.08870941e-01 9.31551218e-01 7.21671402e-01
5.43272793e-01 1.39677241e-01 -1.64373983e-02 4.57656115e-01
5.95752075e-02 8.91994238e-01 1.32577151e-01 -1.29917490e+00
1.15015678e-01 1.34097084e-01 -4.11485992e-02 -1.10896945e+00
-2.33525589e-01 -3.55370551e-01 -6.81780398e-01 7.28578091e-01
7.43585169e-01 4.18932214e-02 -5.49064279e-01 2.04914403e+00
3.86622012e-01 1.79968029e-01 -7.16181174e-02 7.24031985e-01
3.10887903e-01 4.06394452e-01 -5.42156935e-01 4.95056696e-02
8.77944827e-01 -5.86683750e-01 -2.27889091e-01 7.34260604e-02
6.76189482e-01 -5.23246586e-01 1.18916476e+00 8.37255001e-01
-1.01142800e+00 -3.48771475e-02 -1.42710209e+00 8.01776797e-02
-5.87147057e-01 -4.00094599e-01 4.49003726e-01 1.09772599e+00
-1.10464227e+00 8.55407834e-01 -8.11698675e-01 1.14317842e-01
7.11360633e-01 3.08761537e-01 -6.53983057e-01 1.32584408e-01
-1.16724169e+00 6.26132011e-01 1.79002315e-01 -1.57827258e-01
-1.04328477e+00 -7.43105054e-01 -7.99623609e-01 -2.15861529e-01
4.23653871e-01 -1.77532405e-01 1.08088434e+00 -1.20589674e+00
-1.36786997e+00 1.10250807e+00 2.95430660e-01 -7.75480151e-01
1.09975290e+00 -1.16878211e-01 -1.97778434e-01 2.06934407e-01
-1.63405538e-01 5.54563224e-01 1.21939921e+00 -1.49070883e+00
-2.48852968e-01 -4.58443314e-01 1.14504293e-01 -2.27836668e-01
-5.07433593e-01 -1.46060541e-01 -4.58138995e-02 -6.99706316e-01
-2.29788899e-01 -8.83508921e-01 1.20832831e-01 7.55536377e-01
-5.23873985e-01 -2.09025621e-01 8.29459548e-01 -4.11380976e-01
8.65669489e-01 -2.35658789e+00 -1.25239998e-01 5.84780157e-01
2.58784354e-01 2.91330040e-01 -4.24823195e-01 1.55315727e-01
-1.49248764e-01 3.05754006e-01 -6.91793501e-01 -2.33580619e-01
3.14345390e-01 7.13472962e-02 -6.11017048e-01 1.09622252e+00
2.99392547e-02 9.23450112e-01 -8.46441567e-01 -1.97545826e-01
-3.98154199e-01 4.74705815e-01 -3.65918934e-01 -1.49349958e-01
-2.01195300e-01 4.94967513e-02 -1.12446733e-01 6.51632130e-01
1.10832012e+00 -4.78878692e-02 2.24270858e-02 1.45349026e-01
2.44297028e-01 -8.18424150e-02 -1.13056099e+00 1.45954001e+00
-2.38409191e-01 8.01399827e-01 1.34169117e-01 -1.06825864e+00
1.07508457e+00 -6.29868656e-02 2.91651338e-01 -5.89861333e-01
2.63580024e-01 5.28366327e-01 -4.25202936e-01 2.36706603e-02
1.32187933e-01 1.56598583e-01 -1.44664273e-01 6.73883975e-01
-8.26080814e-02 -1.15140989e-01 -1.85054272e-01 2.17644110e-01
9.78438020e-01 1.16993636e-01 8.51667002e-02 -4.43299711e-01
5.90666652e-01 -3.71575058e-01 4.12070513e-01 8.72864187e-01
-3.92273813e-01 7.40725100e-01 9.05588806e-01 -4.88154113e-01
-1.07458544e+00 -1.52522945e+00 -5.08110523e-01 7.60213912e-01
8.90114084e-02 -8.52585658e-02 -1.13586307e+00 -1.07981086e+00
3.51735443e-01 3.96703899e-01 -9.83950853e-01 -2.83593535e-01
-4.09549922e-01 -2.63755649e-01 1.07814181e+00 4.41956848e-01
2.53217578e-01 -7.64306843e-01 -1.95906609e-01 -3.89695644e-01
2.34714508e-01 -7.82815993e-01 -5.67812264e-01 9.51295868e-02
-5.94290435e-01 -1.32765567e+00 -6.10938311e-01 -6.39457226e-01
4.38336968e-01 -8.88365805e-02 8.81615043e-01 -9.86711606e-02
-2.16348037e-01 4.77109939e-01 -7.62929618e-02 -5.11916399e-01
-6.34486794e-01 -8.11292045e-03 2.75804907e-01 2.49008968e-01
2.72785351e-02 -7.19722509e-01 -6.95635796e-01 4.09676820e-01
-1.16599071e+00 -5.24734080e-01 1.59177840e-01 6.56058550e-01
7.43011355e-01 -2.02531710e-01 6.15549743e-01 -9.39569056e-01
4.72700655e-01 -3.79548967e-01 -1.04068017e+00 2.59558648e-01
-7.56993592e-01 1.61065131e-01 8.32596302e-01 -3.96202147e-01
-3.76324356e-01 -2.16666579e-01 4.37480733e-02 -6.21847749e-01
-1.58531040e-01 1.74744621e-01 -1.79214582e-01 -4.79798704e-01
9.89234328e-01 -3.02966707e-03 2.95264393e-01 -2.51279682e-01
5.78529239e-01 5.23199499e-01 8.40429485e-01 -7.41624057e-01
1.17551100e+00 9.20527697e-01 2.08373994e-01 -4.76491243e-01
-1.00769997e+00 3.67433354e-02 -5.40134490e-01 -1.88747063e-01
6.19709074e-01 -4.75433707e-01 -9.19477880e-01 6.99158311e-01
-1.05209947e+00 -5.48128545e-01 -4.53675985e-01 3.32847461e-02
-7.52746105e-01 4.73013967e-01 -4.62939143e-01 -7.74664104e-01
-4.85971123e-01 -7.80947685e-01 9.87560391e-01 -2.13770941e-01
5.95260374e-02 -1.09403241e+00 2.25406244e-01 1.97978690e-01
3.00721586e-01 9.34727192e-01 8.29855859e-01 -1.09563613e+00
-5.03301740e-01 -5.18194735e-01 -2.13354200e-01 8.11256707e-01
-2.73197979e-01 9.66678634e-02 -1.26495588e+00 -6.88917398e-01
5.43556251e-02 -6.43408954e-01 6.10214829e-01 3.06642532e-01
1.58829415e+00 -7.25242019e-01 1.32537901e-01 1.05948353e+00
1.41317272e+00 -1.71058714e-01 8.10346305e-01 6.23377085e-01
5.34541607e-01 4.37293380e-01 4.74214017e-01 4.00357306e-01
4.42998447e-02 3.70666355e-01 7.38613904e-01 -5.30352965e-02
5.02710156e-02 -1.48386776e-01 6.50046706e-01 4.07387435e-01
2.69568831e-01 -1.93489075e-01 -8.11877966e-01 1.64435625e-01
-1.48627353e+00 -8.69446397e-01 -2.54849941e-02 2.46319985e+00
9.21419263e-01 4.72460419e-01 2.37489268e-01 3.56006414e-01
6.60995066e-01 2.22589254e-01 -7.78621674e-01 -6.63135231e-01
-4.66729879e-01 3.65842432e-01 7.06214190e-01 6.88561618e-01
-1.24788499e+00 5.21853089e-01 6.07794523e+00 1.32045591e+00
-1.17592216e+00 3.72386128e-02 9.84173477e-01 -7.33059719e-02
-4.31858063e-01 -1.85696945e-01 -3.40127081e-01 4.65779364e-01
8.64458919e-01 -2.86083013e-01 5.89150429e-01 1.00536048e+00
-2.08124578e-01 2.93556869e-01 -1.19046569e+00 7.01618493e-01
2.19981611e-01 -1.26514649e+00 -1.37506165e-02 2.57732928e-01
8.67557168e-01 9.18151364e-02 6.58582926e-01 -9.62671191e-02
2.17409804e-01 -1.19948602e+00 8.70472550e-01 2.82209337e-01
1.35317445e+00 -1.05814230e+00 5.72665393e-01 7.94564337e-02
-8.09118807e-01 2.53520049e-02 -2.18155444e-01 4.40593541e-01
-3.57311100e-01 3.20822120e-01 -5.74695110e-01 4.58483368e-01
5.91585815e-01 5.68058968e-01 -5.57511151e-01 9.66807842e-01
-2.52033859e-01 5.86635172e-01 -4.02429551e-01 2.64241904e-01
2.34266698e-01 1.52753210e-02 7.68356085e-01 1.00395811e+00
2.41003588e-01 -4.21817154e-01 -9.63036641e-02 8.89923215e-01
-3.95843357e-01 -4.62616161e-02 -8.28173339e-01 5.09070083e-02
4.69271809e-01 1.14204907e+00 -5.96636653e-01 1.37791093e-02
-7.81992171e-03 8.68682563e-01 1.63056478e-01 1.60316825e-01
-9.29025114e-01 -7.46264696e-01 6.17813110e-01 1.74710721e-01
1.72344148e-01 5.35256378e-02 -5.09962142e-01 -9.38714087e-01
2.93942958e-01 -1.21961820e+00 4.97208357e-01 -5.39511859e-01
-1.59365273e+00 5.14452934e-01 -1.61407977e-01 -1.43296325e+00
9.37443301e-02 -1.06893313e+00 -5.50066471e-01 3.49255413e-01
-1.38506937e+00 -9.41679060e-01 -1.20964654e-01 6.92611396e-01
-9.47479233e-02 -2.34283805e-01 8.13226581e-01 1.12817466e-01
-3.49714339e-01 1.33829451e+00 4.50035155e-01 2.49945223e-01
8.45549107e-01 -1.57774949e+00 4.57672268e-01 7.61226296e-01
1.95532531e-01 2.15837330e-01 5.83520710e-01 -3.00898165e-01
-1.34325540e+00 -9.73766208e-01 2.89655298e-01 -6.42494798e-01
1.02988756e+00 -5.37377179e-01 -1.07905841e+00 5.00733912e-01
-7.72129074e-02 3.09699148e-01 4.96685416e-01 -2.32968763e-01
-1.07054651e+00 -2.67484128e-01 -1.43613219e+00 5.51560402e-01
9.40273285e-01 -7.37342417e-01 -1.69664636e-01 5.41372597e-01
9.01107013e-01 -3.88497949e-01 -8.42167020e-01 4.69884366e-01
7.18622029e-01 -9.99162853e-01 9.47347701e-01 -5.49490571e-01
3.33861440e-01 -1.67242691e-01 -3.21152300e-01 -1.02445018e+00
2.60136992e-01 -1.16232157e+00 -3.11264068e-01 8.67176712e-01
5.34974873e-01 -1.05832267e+00 7.54083872e-01 2.31386647e-01
2.63021346e-02 -1.02564120e+00 -1.27580702e+00 -1.38232648e+00
7.67903090e-01 -6.14211738e-01 7.18397021e-01 1.15244961e+00
-3.27937663e-01 -4.48280901e-01 -2.46598303e-01 3.07726741e-01
9.24032271e-01 -1.99275807e-01 8.38019550e-01 -1.26750958e+00
-4.50505167e-01 -6.87066078e-01 -6.81527197e-01 -5.73130250e-01
3.53767246e-01 -1.29091501e+00 -2.22204313e-01 -8.89194429e-01
-1.00739874e-01 -3.57212782e-01 -4.80336756e-01 5.78195632e-01
3.92156392e-01 5.29562235e-01 6.72355220e-02 -2.02767216e-02
-4.37254578e-01 4.23999578e-01 1.22679329e+00 -2.47058898e-01
1.15125172e-01 2.51926929e-02 -4.46830750e-01 7.84653544e-01
9.12955761e-01 -6.71268821e-01 -1.72463000e-01 -2.37239346e-01
4.23034608e-01 -1.58413529e-01 5.60856402e-01 -8.89425874e-01
-1.29650719e-02 -6.01199456e-02 1.17476992e-01 -2.78613061e-01
1.60673290e-01 -6.69834495e-01 -2.88136601e-01 5.35748005e-01
-5.39560616e-01 -5.15356213e-02 1.71896055e-01 5.59500992e-01
6.55666590e-02 -2.34869719e-01 9.64094222e-01 4.17832404e-01
-1.36032075e-01 4.88833159e-01 1.58559699e-02 4.78311956e-01
1.03395867e+00 -2.92051017e-01 -6.86245739e-01 -4.98784274e-01
-4.34155524e-01 1.12087347e-01 6.00927472e-01 3.18025202e-01
6.49736345e-01 -1.47273695e+00 -7.60610640e-01 2.99286187e-01
1.03306018e-01 -8.49444941e-02 -9.73668694e-02 7.22771525e-01
-7.88119137e-01 -4.41702791e-02 -6.28271922e-02 -5.44073701e-01
-1.08524883e+00 4.61722970e-01 6.28876030e-01 -2.73341481e-02
-4.65519667e-01 9.25385594e-01 6.87462762e-02 -6.89212918e-01
6.00765228e-01 6.40384434e-03 4.25627172e-01 -1.28237620e-01
6.00212753e-01 4.89205271e-01 8.05184692e-02 -3.75070870e-01
-3.47263485e-01 3.99548948e-01 8.06786418e-02 -3.24181706e-01
1.25878572e+00 3.40057433e-01 -2.11432189e-01 2.86561430e-01
1.63788509e+00 2.44663566e-01 -1.42921853e+00 1.01242354e-02
7.87478536e-02 -4.59398657e-01 -2.94595689e-01 -8.21464539e-01
-1.10914242e+00 8.14971149e-01 8.04514408e-01 3.54386389e-01
9.05124724e-01 -1.35840401e-01 7.65290797e-01 5.92682362e-01
2.65953600e-01 -9.47699904e-01 2.40859821e-01 4.79337722e-01
1.07309854e+00 -1.03808022e+00 -1.17779940e-01 -1.61489457e-01
-4.23140943e-01 1.08293629e+00 3.36667269e-01 -4.95298624e-01
7.77984142e-01 5.67496777e-01 1.25685975e-01 7.80578032e-02
-4.17681634e-01 3.90480399e-01 5.49212098e-01 6.80839837e-01
-3.96084599e-02 6.11246340e-02 2.98989750e-02 2.83821225e-01
-4.56948757e-01 -3.32119823e-01 5.12648821e-01 7.11843729e-01
-3.75928074e-01 -1.10528064e+00 -3.93976420e-01 2.30472103e-01
-5.97795784e-01 1.28044873e-01 -5.42579830e-01 8.76459658e-01
-1.11560374e-01 5.53440154e-01 -6.56282250e-03 -3.98767799e-01
8.47615302e-02 -1.56937227e-01 4.14092690e-01 -6.60225600e-02
-2.47628376e-01 -3.74380797e-01 -2.04639867e-01 -7.94712365e-01
4.08805571e-02 -5.02760231e-01 -1.14940178e+00 -5.66629648e-01
-1.70964852e-01 -2.39457544e-02 6.68257535e-01 7.56818712e-01
1.69881925e-01 -1.68311700e-01 1.08316600e+00 -6.67941809e-01
-1.06607699e+00 -4.15736526e-01 -6.00088716e-01 4.27824676e-01
5.81709027e-01 -4.23010975e-01 -9.02585566e-01 -2.31392279e-01] | [5.672607898712158, 7.799917221069336] |
09eb8c11-6d0f-4640-b187-62985e3f449b | large-car-following-data-based-on-lyft-level | 2305.18921 | null | https://arxiv.org/abs/2305.18921v1 | https://arxiv.org/pdf/2305.18921v1.pdf | Large Car-following Data Based on Lyft level-5 Open Dataset: Following Autonomous Vehicles vs. Human-driven Vehicles | Car-Following (CF), as a fundamental driving behaviour, has significant influences on the safety and efficiency of traffic flow. Investigating how human drivers react differently when following autonomous vs. human-driven vehicles (HV) is thus critical for mixed traffic flow. Research in this field can be expedited with trajectory datasets collected by Autonomous Vehicles (AVs). However, trajectories collected by AVs are noisy and not readily applicable for studying CF behaviour. This paper extracts and enhances two categories of CF data, HV-following-AV (H-A) and HV-following-HV (H-H), from the open Lyft level-5 dataset. First, CF pairs are selected based on specific rules. Next, the quality of raw data is assessed by anomaly analysis. Then, the raw CF data is corrected and enhanced via motion planning, Kalman filtering, and wavelet denoising. As a result, 29k+ H-A and 42k+ H-H car-following segments are obtained, with a total driving distance of 150k+ km. A diversity assessment shows that the processed data cover complete CF regimes for calibrating CF models. This open and ready-to-use dataset provides the opportunity to investigate the CF behaviours of following AVs vs. HVs from real-world data. It can further facilitate studies on exploring the impact of AVs on mixed urban traffic. | ['J. W. C. van Lint', 'Simeon C. Calvert', 'Victor L. Knoop', 'Yiru Jiao', 'Guopeng Li'] | 2023-05-30 | null | null | null | null | ['autonomous-vehicles', 'motion-planning'] | ['computer-vision', 'robots'] | [-2.45499909e-01 -2.57252991e-01 -3.32377106e-01 -4.34572041e-01
-6.98931396e-01 -4.70117509e-01 8.73955429e-01 2.46630996e-01
-5.43734550e-01 5.03801346e-01 1.51839688e-01 -8.22149754e-01
-3.38831514e-01 -9.59821641e-01 -6.76072776e-01 -8.66199195e-01
-8.80661309e-02 2.09389791e-01 5.61211586e-01 -5.62242866e-01
2.80791461e-01 8.49984705e-01 -1.93402290e+00 -5.94108291e-02
8.59892428e-01 7.67260969e-01 3.49339336e-01 6.14729881e-01
-3.32566677e-04 4.15872902e-01 -7.35688023e-03 -1.84187472e-01
3.37703943e-01 -1.35300726e-01 -3.68635803e-01 -2.04341918e-01
1.50590986e-01 -2.15567037e-01 -5.98100901e-01 5.31410992e-01
1.78042799e-01 6.13064766e-01 7.05292046e-01 -1.78876209e+00
-1.25778439e-02 -9.05332342e-02 7.83353969e-02 3.32739741e-01
9.49146375e-02 7.99783528e-01 5.37547052e-01 -5.31630695e-01
6.55988693e-01 1.17240882e+00 4.97035980e-01 1.63057789e-01
-1.20675719e+00 -7.04708993e-01 -1.39454409e-01 6.25664830e-01
-1.41922104e+00 -5.86669981e-01 6.58361912e-01 -8.75921190e-01
7.98609138e-01 3.25208217e-01 9.43448067e-01 1.17294669e+00
3.86498511e-01 5.72342634e-01 8.66133630e-01 1.24932967e-01
2.62562126e-01 1.90319449e-01 5.03624916e-01 3.70163321e-02
3.97870481e-01 8.61998618e-01 -2.47049630e-01 2.02084556e-01
-1.69092149e-01 -1.76058277e-01 7.54433796e-02 -1.15281776e-01
-9.47552145e-01 9.63070452e-01 2.98704356e-01 -8.67943093e-02
-5.62779784e-01 -2.92544123e-02 4.59077895e-01 2.89208978e-01
2.17797503e-01 -9.89753455e-02 -1.94883510e-01 -4.88635361e-01
-1.07214379e+00 9.87536490e-01 3.02120626e-01 1.11792445e+00
1.05225551e+00 -2.58902609e-02 -1.26543641e-01 3.99074346e-01
3.48301411e-01 1.04417908e+00 -1.51023969e-01 -1.22386086e+00
4.72720116e-01 4.12929952e-01 9.56044570e-02 -9.98665750e-01
-5.48763037e-01 -2.19785217e-02 -4.97710973e-01 2.91014761e-01
6.45665884e-01 -1.40463933e-01 -6.39198840e-01 1.36210442e+00
3.15672547e-01 7.68541917e-02 -2.39263568e-02 6.85280919e-01
6.14026964e-01 7.48912811e-01 2.53519803e-01 -7.54036158e-02
1.24183726e+00 -5.06722271e-01 -8.19002867e-01 -1.05467588e-01
1.02956653e+00 -4.66729075e-01 8.13493371e-01 -7.16428384e-02
-8.34360957e-01 -8.29394996e-01 -6.39208794e-01 1.99984387e-01
-7.73432374e-01 -3.17997426e-01 7.41225556e-02 6.23333514e-01
-7.49354601e-01 3.23466152e-01 -6.85695410e-01 -4.10343021e-01
4.30362582e-01 -4.44824658e-02 -3.52583021e-01 -9.34115201e-02
-1.44778681e+00 1.01895654e+00 1.55883402e-01 2.45328516e-01
-9.90578592e-01 -1.10716963e+00 -1.04125416e+00 -4.06659961e-01
4.03872997e-01 -2.98886031e-01 1.13726497e+00 -2.67184973e-02
-8.60698342e-01 3.95400912e-01 -4.91375297e-01 -7.64788568e-01
7.27509320e-01 4.09282278e-03 -1.05240571e+00 -4.69671786e-02
3.98245573e-01 4.31851208e-01 5.96643686e-01 -1.32542288e+00
-1.15767539e+00 -2.92690068e-01 -3.99969041e-01 -1.12244345e-01
4.77499396e-01 -4.18526530e-02 -2.35208482e-01 -1.88189864e-01
-3.49596113e-01 -1.35402548e+00 -1.00981466e-01 -3.89761776e-01
-3.23616564e-01 -3.48239481e-01 1.08115911e+00 -6.74015999e-01
1.50242269e+00 -2.42628741e+00 -4.94740576e-01 5.11792302e-01
4.78676111e-02 5.71540892e-01 1.49354795e-02 7.68155575e-01
3.76345336e-01 4.47413139e-02 -3.13206643e-01 -5.97426062e-03
-2.33392809e-02 4.65015948e-01 -1.95844680e-01 6.17703080e-01
2.26892740e-01 8.03827405e-01 -7.70733774e-01 -3.52052838e-01
9.38274682e-01 2.60946721e-01 -4.05525714e-01 -2.06735041e-02
2.28856355e-01 6.61480665e-01 -3.97175997e-01 5.61779320e-01
9.36153412e-01 8.84262860e-01 -6.07948601e-01 7.98008777e-03
-9.12998676e-01 -4.94784229e-02 -8.44681025e-01 7.51957834e-01
-3.74877870e-01 1.30288148e+00 1.28726065e-01 -7.85905898e-01
9.40164208e-01 2.43716743e-02 4.36701357e-01 -1.25184155e+00
1.34186238e-01 2.86658466e-01 2.46373534e-01 -9.20358300e-01
7.73633480e-01 9.21946764e-02 -1.47714719e-01 -3.72653194e-02
-4.43257362e-01 -1.46781847e-01 6.19316936e-01 1.29318267e-01
8.92516136e-01 -2.05685675e-01 -3.18388104e-01 -4.73679572e-01
7.38692343e-01 5.09994805e-01 4.15081114e-01 4.95391399e-01
-7.26347625e-01 2.63144583e-01 4.75862890e-01 -1.97625875e-01
-1.24044752e+00 -1.19887400e+00 -6.90052152e-01 5.40589690e-01
3.16577435e-01 -2.43410990e-01 -6.47267342e-01 -2.10344285e-01
5.13424873e-01 1.45469987e+00 -5.58446407e-01 -5.47843933e-01
-7.43655205e-01 -3.20119202e-01 6.22584522e-01 5.48495114e-01
4.77408588e-01 -9.70142901e-01 -7.02320039e-01 2.69124508e-01
-2.36022294e-01 -1.18554628e+00 -1.35549381e-01 -2.92760253e-01
-3.04903507e-01 -1.16515219e+00 -1.81771681e-01 -2.61177927e-01
3.23805392e-01 7.55063653e-01 7.03111887e-01 1.56550705e-02
6.54546097e-02 2.98726708e-01 -2.47955099e-01 -5.72081506e-01
-6.48124278e-01 -1.07997963e-02 1.22989938e-01 2.79925972e-01
7.03592598e-01 -1.23396531e-01 -5.86311638e-01 6.72041774e-01
-5.70269644e-01 -2.86694676e-01 4.27256197e-01 2.41455302e-01
4.46287692e-01 3.10257286e-01 6.12553596e-01 -4.36440319e-01
4.51009244e-01 -8.50713432e-01 -5.62243640e-01 -3.94930035e-01
-7.50897884e-01 -2.35950202e-01 5.22603452e-01 -6.61517382e-02
-1.15689611e+00 -4.68607619e-02 -3.90254378e-01 -1.70665443e-01
-9.03432488e-01 3.16010863e-01 -7.21023455e-02 2.35272706e-01
4.78203684e-01 -9.75707173e-03 3.04522693e-01 -1.71459422e-01
3.60669762e-01 9.35017288e-01 7.01601923e-01 -6.60102889e-02
1.03674269e+00 6.61956787e-01 1.60429627e-01 -1.62621117e+00
-7.58160204e-02 -8.91948521e-01 -6.82092965e-01 -7.92168736e-01
1.15058744e+00 -8.52587461e-01 -7.85289466e-01 5.33696532e-01
-5.86203992e-01 -5.01801729e-01 -1.34435326e-01 8.30131590e-01
-5.87059319e-01 -2.29954273e-02 -5.45977950e-02 -1.14681745e+00
4.33253795e-01 -1.32704139e+00 9.90930498e-01 1.39413867e-02
-4.73818749e-01 -9.45554674e-01 4.92870845e-02 5.66087365e-01
3.82385164e-01 2.76494980e-01 6.87218726e-01 -2.74633557e-01
-5.24747550e-01 -3.70121926e-01 -6.99849725e-02 2.10503042e-01
-1.96243316e-01 4.36438203e-01 -7.86478519e-01 -1.13615356e-01
-4.34872121e-01 3.63007396e-01 8.57442856e-01 5.83684623e-01
5.69974601e-01 6.56607002e-03 -4.39859450e-01 4.11796302e-01
1.14767969e+00 5.31391263e-01 9.59710479e-01 6.50958478e-01
5.42682230e-01 9.90168571e-01 1.21106911e+00 3.70722562e-02
5.80224037e-01 7.30874002e-01 4.80217963e-01 9.50730890e-02
-1.41709849e-01 -2.57598966e-01 3.84501755e-01 3.46074462e-01
-8.20729509e-02 -1.47032633e-01 -1.16081476e+00 1.03589594e+00
-1.73596251e+00 -1.35777414e+00 -1.24656940e+00 2.22238326e+00
-1.20098643e-01 3.79530072e-01 4.70099270e-01 4.60807025e-01
7.13833392e-01 9.58290920e-02 -2.70193309e-01 -4.88723278e-01
-3.24221626e-02 -2.61760443e-01 1.10536361e+00 5.88754654e-01
-7.54010260e-01 6.88456297e-01 5.59898138e+00 8.85287404e-01
-8.37061644e-01 -1.58410624e-01 1.44386679e-01 -3.57764624e-02
-3.50573450e-01 2.11580053e-01 -1.09819162e+00 9.02719676e-01
1.73926127e+00 -2.81828403e-01 3.00944299e-01 5.05419731e-01
1.18560541e+00 -5.40991843e-01 -5.53414941e-01 6.37338996e-01
-5.39066553e-01 -1.16191089e+00 -3.20140094e-01 3.52874041e-01
4.06432986e-01 1.14045031e-01 -3.85916010e-02 5.71401417e-01
-6.23490214e-02 -7.51112938e-01 1.07810605e+00 8.08866501e-01
4.68344927e-01 -1.10677028e+00 8.49867582e-01 4.59261984e-01
-1.53405404e+00 -3.06042373e-01 -2.06147268e-01 6.67732721e-03
8.39515448e-01 5.01939833e-01 -4.93074805e-01 7.80165434e-01
7.66770065e-01 7.74253070e-01 -8.17578077e-01 9.75244999e-01
2.14411721e-01 8.57508004e-01 -2.13878646e-01 1.14483751e-01
5.64968050e-01 -5.73902845e-01 8.52682471e-01 1.17212808e+00
4.23875600e-01 -1.49994731e-01 -1.09051421e-01 6.57569528e-01
6.76516175e-01 -2.00437486e-01 -1.12457800e+00 1.03901744e-01
6.51848853e-01 1.08545899e+00 -4.31390643e-01 -2.43093222e-01
-4.95271415e-01 7.92190805e-02 -3.03016633e-01 4.67333972e-01
-1.03888047e+00 -3.17299247e-01 1.23479962e+00 8.02092433e-01
1.47736982e-01 -5.62140763e-01 -1.65725023e-01 -1.99292541e-01
-1.84835657e-01 -2.95867920e-01 1.31129930e-02 -6.22849822e-01
-8.56358230e-01 2.50877172e-01 5.12559175e-01 -1.39376771e+00
-2.96463549e-01 -4.60220456e-01 -7.02330470e-01 7.31360972e-01
-1.55537236e+00 -7.50149667e-01 -5.05791128e-01 5.80039561e-01
4.29563254e-01 -8.04895163e-02 -3.01765651e-02 5.76751053e-01
-7.50103652e-01 2.41794750e-01 2.90968090e-01 -2.73313373e-01
3.91467154e-01 -7.47260690e-01 6.31353199e-01 8.26872826e-01
-5.29617310e-01 1.60873935e-01 8.40542257e-01 -8.85033309e-01
-1.30595040e+00 -1.63744676e+00 1.06162941e+00 -5.77871621e-01
6.22184515e-01 -1.58265471e-01 -1.04853797e+00 4.33224767e-01
2.44783796e-02 -1.33800194e-01 2.81353295e-01 -4.89099979e-01
4.70254749e-01 -2.17083290e-01 -1.02998579e+00 6.93388283e-01
1.05199087e+00 -6.07938111e-01 -2.84913689e-01 -5.30040935e-02
4.11432803e-01 1.13676094e-01 -8.34295154e-01 4.49630469e-01
5.06094158e-01 -1.13704967e+00 8.71347129e-01 -4.84447211e-01
-5.92886508e-02 -4.23587263e-01 -1.27429917e-01 -1.25709379e+00
-4.71057415e-01 -3.25994968e-01 2.38877222e-01 1.19739342e+00
3.27354610e-01 -8.70203733e-01 4.10443842e-01 6.99119866e-01
-7.47173309e-01 -3.94811630e-01 -1.04970312e+00 -1.11439955e+00
2.72011191e-01 -1.25246191e+00 8.18950951e-01 4.96553570e-01
-5.33263385e-01 -1.76831290e-01 -3.87137942e-02 2.72447079e-01
5.87322831e-01 -4.40359503e-01 1.16295874e+00 -1.32234633e+00
7.15128064e-01 -6.38455868e-01 -5.68650424e-01 -5.08331060e-01
3.37963551e-01 -8.60530436e-01 -4.73062284e-02 -1.46319211e+00
-4.47940618e-01 -4.05501693e-01 3.30854475e-01 -1.71723112e-01
1.19235978e-01 4.28607985e-02 1.59559801e-01 1.91160485e-01
7.50230551e-02 6.14008427e-01 1.08771968e+00 1.94240168e-01
-2.70612329e-01 2.22757623e-01 -9.41210240e-02 5.51098824e-01
9.88741398e-01 -3.47424269e-01 -4.51681018e-01 2.65765250e-01
-1.74684986e-01 1.11682087e-01 7.56099164e-01 -1.08563304e+00
1.90083116e-01 -4.25039142e-01 -1.90847039e-01 -1.37775755e+00
1.65177062e-01 -1.00442922e+00 4.97230411e-01 5.07645607e-01
-9.68568772e-02 1.90226629e-01 3.06970596e-01 6.24682486e-01
-1.89554408e-01 -4.97027934e-02 6.44697726e-01 2.75009274e-01
-1.07889819e+00 3.57616514e-01 -1.22945106e+00 2.36301459e-02
1.62731743e+00 -6.14953697e-01 -2.85435200e-01 -3.99968982e-01
-5.01968265e-01 6.72132194e-01 3.71490926e-01 6.37028575e-01
3.91678184e-01 -1.57962060e+00 -7.49957681e-01 6.55800760e-01
3.15511227e-01 -1.69349119e-01 7.38572776e-01 1.16020703e+00
-4.34666663e-01 7.82420039e-01 -2.08811596e-01 -7.77195871e-01
-9.67741847e-01 7.14549720e-01 1.18296050e-01 4.54964221e-01
-7.61848271e-01 1.05910096e-02 1.23721699e-03 -4.25335407e-01
-2.52648741e-01 -2.27679491e-01 -2.56403118e-01 4.71708357e-01
3.23519498e-01 1.21723080e+00 1.41456887e-01 -1.56290019e+00
-2.73764759e-01 5.25672615e-01 4.62321371e-01 -1.85666159e-01
8.05894494e-01 -6.71564758e-01 4.44178432e-01 6.51509345e-01
1.41242564e+00 -4.13794145e-02 -1.35787392e+00 3.69696110e-01
5.89813367e-02 -6.29995167e-01 1.60845160e-01 -4.54322584e-02
-8.86126161e-01 8.06908548e-01 5.98512471e-01 3.87839079e-01
7.12078393e-01 -3.91378924e-02 1.07031691e+00 -2.75656730e-02
1.73702508e-01 -1.28146613e+00 -7.73593605e-01 3.68666232e-01
7.51552284e-01 -1.19333434e+00 -5.21615803e-01 -3.70855361e-01
-7.42531955e-01 7.54868150e-01 3.04401398e-01 3.21734585e-02
9.29069102e-01 4.37237732e-02 1.32210329e-01 -2.16985896e-01
-6.76175296e-01 -6.77036047e-01 3.09361547e-01 7.93764591e-01
-2.23592728e-01 3.43084335e-01 -2.97427595e-01 1.99631259e-01
-6.01354241e-01 -2.49922290e-01 6.11877859e-01 7.80078590e-01
-4.09450680e-01 -8.27219963e-01 -4.50775206e-01 7.81107306e-01
3.06979030e-01 4.06754851e-01 -4.84980829e-02 1.23288751e+00
2.04323247e-01 1.44466066e+00 2.87969381e-01 -5.63435555e-01
7.98054099e-01 1.48746789e-01 -2.24481195e-01 8.29487666e-02
-2.81450748e-01 -4.07873213e-01 3.90056163e-01 -7.60532320e-01
-2.17770770e-01 -1.33283174e+00 -1.33755791e+00 -1.00526428e+00
-1.45772874e-01 4.67475578e-02 7.41421700e-01 1.04762661e+00
3.19090515e-01 5.58224380e-01 7.92198360e-01 -8.71948481e-01
-1.17981978e-01 -6.15015209e-01 -6.31912768e-01 5.89713097e-01
6.53174579e-01 -1.04645193e+00 -7.25371659e-01 -2.78347880e-01] | [5.838973045349121, 1.1605075597763062] |
8b727d84-a3dd-49d4-9ce2-69c10d59c3d3 | spatio-temporal-relation-modeling-for-few | 2112.05132 | null | https://arxiv.org/abs/2112.05132v2 | https://arxiv.org/pdf/2112.05132v2.pdf | Spatio-temporal Relation Modeling for Few-shot Action Recognition | We propose a novel few-shot action recognition framework, STRM, which enhances class-specific feature discriminability while simultaneously learning higher-order temporal representations. The focus of our approach is a novel spatio-temporal enrichment module that aggregates spatial and temporal contexts with dedicated local patch-level and global frame-level feature enrichment sub-modules. Local patch-level enrichment captures the appearance-based characteristics of actions. On the other hand, global frame-level enrichment explicitly encodes the broad temporal context, thereby capturing the relevant object features over time. The resulting spatio-temporally enriched representations are then utilized to learn the relational matching between query and support action sub-sequences. We further introduce a query-class similarity classifier on the patch-level enriched features to enhance class-specific feature discriminability by reinforcing the feature learning at different stages in the proposed framework. Experiments are performed on four few-shot action recognition benchmarks: Kinetics, SSv2, HMDB51 and UCF101. Our extensive ablation study reveals the benefits of the proposed contributions. Furthermore, our approach sets a new state-of-the-art on all four benchmarks. On the challenging SSv2 benchmark, our approach achieves an absolute gain of $3.5\%$ in classification accuracy, as compared to the best existing method in the literature. Our code and models are available at https://github.com/Anirudh257/strm. | ['Bernard Ghanem', 'Fahad Shahbaz Khan', 'Rao Muhammad Anwer', 'Salman Khan', 'Sanath Narayan', 'Anirudh Thatipelli'] | 2021-12-09 | spatio-temporal-relation-modeling-for-few-1 | http://openaccess.thecvf.com//content/CVPR2022/html/Thatipelli_Spatio-Temporal_Relation_Modeling_for_Few-Shot_Action_Recognition_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Thatipelli_Spatio-Temporal_Relation_Modeling_for_Few-Shot_Action_Recognition_CVPR_2022_paper.pdf | cvpr-2022-1 | ['few-shot-action-recognition'] | ['computer-vision'] | [ 4.27381158e-01 -5.10725975e-01 -6.52513683e-01 -3.12363237e-01
-1.00771368e+00 -1.78076670e-01 6.92368865e-01 2.64148116e-01
-3.12431753e-01 4.18381304e-01 4.19512779e-01 4.40966755e-01
-3.27111125e-01 -5.73119879e-01 -4.77010161e-01 -6.77391648e-01
-2.71371156e-01 -9.51273888e-02 8.21310639e-01 -1.70660526e-01
3.19657624e-01 4.21281278e-01 -1.86931932e+00 8.52280200e-01
4.72314090e-01 1.38254571e+00 7.88149089e-02 5.12942970e-01
1.17422827e-01 1.25050151e+00 -2.49401405e-01 1.34643897e-01
2.93231905e-01 -4.37166303e-01 -8.66943419e-01 2.40489885e-01
4.96963918e-01 -3.14399153e-01 -6.91651225e-01 5.95993519e-01
3.94619226e-01 6.17799580e-01 4.53013867e-01 -1.26501417e+00
-4.53437924e-01 4.60319929e-02 -7.24810362e-01 7.43103445e-01
5.71585476e-01 5.98372161e-01 1.17558467e+00 -9.33556080e-01
7.22160697e-01 1.05946088e+00 4.60490584e-01 3.98651183e-01
-9.97586191e-01 -3.97352695e-01 4.06857401e-01 6.72688663e-01
-1.28091037e+00 -4.86902535e-01 7.66752064e-01 -5.22992074e-01
1.19424260e+00 1.00776814e-01 6.04723692e-01 1.13222337e+00
2.36364901e-01 1.19481277e+00 1.00926769e+00 -1.45651400e-01
2.74913609e-01 -5.91765702e-01 3.58733028e-01 7.00181842e-01
-3.66461545e-01 2.15608820e-01 -9.75925624e-01 -8.93787146e-02
7.38009512e-01 6.62605286e-01 -1.10354222e-01 -4.79540586e-01
-1.20288622e+00 5.87930799e-01 5.11650324e-01 5.21246970e-01
-5.77198029e-01 4.52010840e-01 6.71033502e-01 5.53979948e-02
3.78514111e-01 3.40625197e-02 -3.07260782e-01 -6.24484479e-01
-8.48643422e-01 9.93127078e-02 1.00868143e-01 8.70157540e-01
7.69588590e-01 -1.23842224e-01 -8.75385106e-01 9.22422290e-01
1.21315923e-02 6.36523515e-02 6.13459170e-01 -9.51856732e-01
4.05349255e-01 8.83735001e-01 -3.80357727e-02 -8.98774326e-01
-2.26035863e-01 -3.01909387e-01 -5.25839567e-01 2.48220898e-02
1.81207344e-01 4.75971311e-01 -8.93985271e-01 1.70742738e+00
4.12106335e-01 7.25207388e-01 -3.73129435e-02 8.12782109e-01
7.84192741e-01 5.69123805e-01 2.51451313e-01 -1.36484921e-01
1.46482146e+00 -1.26424062e+00 -4.55991119e-01 -6.60934672e-02
6.89457417e-01 -4.01869684e-01 1.19974017e+00 -3.80613841e-02
-9.56646621e-01 -7.58510709e-01 -9.02075410e-01 1.44270524e-01
-3.54303122e-01 7.89547861e-02 5.80579877e-01 2.34341784e-03
-6.17837131e-01 7.60463715e-01 -1.00209057e+00 -5.61405122e-01
8.21527541e-01 1.05839327e-01 -3.81126970e-01 -1.66354433e-01
-1.08055508e+00 4.63955909e-01 3.20668191e-01 -3.62116903e-01
-1.19134116e+00 -8.34975243e-01 -9.00534153e-01 -7.29741007e-02
6.63744092e-01 -2.75373131e-01 1.21227443e+00 -8.77240539e-01
-1.30502093e+00 7.42143452e-01 -3.04345399e-01 -4.81638104e-01
3.07829946e-01 -1.86220109e-01 -4.47070897e-01 5.83610177e-01
2.87853241e-01 5.68130732e-01 6.92319870e-01 -7.79050708e-01
-8.97076607e-01 -3.59945744e-01 3.75693262e-01 1.09662168e-01
-3.42789590e-01 3.68884094e-02 -6.68268740e-01 -9.14605618e-01
-5.73248975e-02 -7.32742190e-01 -8.53519067e-02 1.74875960e-01
1.41923338e-01 -3.35056752e-01 8.63562942e-01 -2.64600396e-01
1.40204883e+00 -2.40397954e+00 6.60641864e-02 -2.01378658e-01
-7.60408193e-02 2.51609981e-01 -4.38754439e-01 6.83742642e-01
-1.35840133e-01 -2.84261882e-01 -2.51744092e-01 -1.54659376e-01
-3.89286131e-02 6.13475256e-02 -4.80386727e-02 5.17467797e-01
4.31802392e-01 1.14927208e+00 -1.13459289e+00 -4.99155074e-01
4.31639791e-01 4.86641467e-01 -4.34002131e-01 2.09854618e-02
-1.56058922e-01 4.03638005e-01 -5.86786032e-01 9.66316164e-01
1.22579716e-01 -3.79087299e-01 -6.37293682e-02 -2.21232712e-01
2.79278196e-02 1.32549822e-01 -1.06144035e+00 2.17217255e+00
-3.89282435e-01 3.52156997e-01 -4.90093946e-01 -1.14014709e+00
7.02130377e-01 2.42003888e-01 1.01182985e+00 -9.97723401e-01
9.51862708e-02 -7.21979216e-02 -2.19533518e-01 -5.15837133e-01
4.27528709e-01 1.48140773e-01 -1.95890769e-01 3.80927920e-01
2.63088018e-01 5.69389284e-01 2.87178397e-01 3.85684282e-01
1.66268861e+00 5.31695664e-01 5.35143077e-01 -1.68475032e-01
5.96900761e-01 -1.14146106e-01 7.53984332e-01 5.80661833e-01
-7.37372518e-01 4.51624602e-01 2.51240045e-01 -4.98046041e-01
-5.51904917e-01 -8.87384951e-01 1.14872426e-01 1.44540250e+00
3.32145214e-01 -7.97461450e-01 -3.67428541e-01 -8.65753412e-01
7.01846555e-02 4.30419147e-01 -1.08387434e+00 -4.38435584e-01
-4.74846244e-01 -3.26298058e-01 3.75501484e-01 9.94722247e-01
6.02287412e-01 -1.31916273e+00 -1.08329487e+00 1.95134625e-01
-1.52560189e-01 -1.13681424e+00 -6.39795244e-01 -3.49852704e-02
-9.36245203e-01 -1.11809838e+00 -6.26281857e-01 -4.14643079e-01
3.32426071e-01 6.08294904e-01 9.26892519e-01 2.20550373e-02
-7.71048248e-01 8.39841664e-01 -8.16759408e-01 8.04192796e-02
1.17478207e-01 -1.45713419e-01 -7.66781941e-02 4.87488210e-01
4.18189943e-01 -7.88113534e-01 -8.37865353e-01 5.00623524e-01
-8.77407014e-01 -1.99764252e-01 5.79081059e-01 8.29513550e-01
8.84276092e-01 -2.31070951e-01 5.57002366e-01 -4.81836259e-01
-3.50900628e-02 -4.57898408e-01 -1.95977330e-01 2.26226911e-01
-1.44199342e-01 -1.10665642e-01 2.62378514e-01 -5.29587924e-01
-9.46180046e-01 1.74593061e-01 3.40162218e-01 -7.10183084e-01
-2.52938569e-01 3.76830161e-01 -1.63342729e-01 1.60160944e-01
5.79514205e-01 5.87143481e-01 -2.24198684e-01 -5.04786849e-01
2.73244858e-01 3.24991256e-01 5.35923660e-01 -7.01640904e-01
4.73792493e-01 8.12033296e-01 -3.88731062e-02 -7.30749369e-01
-8.60197604e-01 -9.32857990e-01 -8.70522022e-01 -3.67221922e-01
8.70516360e-01 -9.26742315e-01 -5.78406155e-01 4.67506886e-01
-6.80809379e-01 -4.82711375e-01 -7.12067783e-01 3.69879931e-01
-9.40077245e-01 4.17903155e-01 -4.80727315e-01 -7.62398422e-01
-1.95404336e-01 -1.02000368e+00 1.41403353e+00 1.27578616e-01
-2.31845692e-01 -6.33579254e-01 2.80604154e-01 4.98031288e-01
1.34203136e-01 4.85790491e-01 4.93122280e-01 -5.71188748e-01
-5.67677677e-01 -2.07990587e-01 -1.38395324e-01 1.08290397e-01
3.15335184e-01 -1.23819262e-01 -8.10788095e-01 -3.12083572e-01
-5.07147849e-01 -6.06350422e-01 1.23654151e+00 2.66592711e-01
1.30027568e+00 9.50461850e-02 -3.19388121e-01 4.07001555e-01
1.44009614e+00 3.56050789e-01 7.61831462e-01 2.40849271e-01
5.84316730e-01 3.45425010e-01 1.24744701e+00 9.06108975e-01
2.24322885e-01 1.10564470e+00 3.29447240e-01 2.68045485e-01
-2.63162076e-01 -1.90945849e-01 6.34993970e-01 3.11391890e-01
-3.52995157e-01 2.02561900e-01 -7.73954093e-01 7.11225569e-01
-2.28617311e+00 -1.43373203e+00 2.04164177e-01 1.97819531e+00
5.83273411e-01 7.23179877e-02 5.33764839e-01 2.05500171e-01
4.86107707e-01 6.35220349e-01 -6.37808979e-01 5.88220656e-02
4.44334894e-02 3.23693275e-01 1.54834136e-01 1.69313774e-01
-1.53586483e+00 9.43956792e-01 5.04010534e+00 1.15080762e+00
-8.17869663e-01 2.78227806e-01 2.98033834e-01 -4.83746022e-01
3.11550468e-01 -2.43599433e-02 -6.31176233e-01 3.89032722e-01
7.38408983e-01 -9.65465009e-02 1.19882450e-01 8.28954339e-01
3.33970189e-01 -2.10395977e-01 -1.20920050e+00 1.03829062e+00
2.04643339e-01 -1.45000529e+00 -1.01044774e-02 -1.18446266e-02
6.45170093e-01 -2.03934871e-02 -8.27449635e-02 5.30284166e-01
1.55486353e-02 -7.07376361e-01 6.75364196e-01 7.15153813e-01
7.21737623e-01 -7.81409144e-01 3.11904490e-01 1.07950903e-01
-1.90482938e+00 -3.96748096e-01 -2.25764886e-01 -9.21400040e-02
2.11419594e-02 1.31074920e-01 -1.88985452e-01 8.70398700e-01
8.54386866e-01 1.53712249e+00 -6.79830611e-01 8.18816960e-01
4.32351232e-02 4.67728615e-01 1.00009233e-01 2.05981925e-01
5.15421152e-01 1.61321729e-01 4.59188223e-01 1.30431056e+00
3.56332436e-02 5.56931257e-01 4.77141291e-01 4.22945529e-01
1.44768238e-01 1.20112501e-01 -4.92638856e-01 -6.20160066e-02
2.43098944e-01 1.26933777e+00 -7.24723399e-01 -5.10257304e-01
-5.98117352e-01 1.12549627e+00 3.32923949e-01 1.47135004e-01
-1.07064605e+00 -1.30444258e-01 8.96938324e-01 5.01290783e-02
6.48090422e-01 9.73459613e-03 1.89177290e-01 -1.28983724e+00
1.87407658e-01 -9.23913240e-01 8.83103132e-01 -5.04630506e-01
-1.25131953e+00 3.39706093e-01 1.50584787e-01 -1.71561217e+00
2.09627971e-02 -3.21288198e-01 -6.79739416e-01 3.22250098e-01
-1.28946090e+00 -1.35180891e+00 -5.17702460e-01 9.21818376e-01
1.06357443e+00 -3.19824904e-01 6.75024390e-01 3.31514329e-01
-6.90914631e-01 6.28498852e-01 -1.13810137e-01 1.32529512e-01
5.09237468e-01 -7.64371693e-01 1.49358615e-01 7.49346733e-01
3.10128242e-01 3.95713985e-01 1.30753517e-01 -6.10523224e-01
-1.45928693e+00 -1.31346786e+00 5.07782459e-01 -5.18839478e-01
7.30074406e-01 -2.34527826e-01 -8.09262693e-01 5.23142040e-01
-1.30372956e-01 6.02276325e-01 7.55148292e-01 -1.12128899e-01
-7.14751899e-01 -2.74156898e-01 -9.18061256e-01 3.56893331e-01
1.52269626e+00 -6.82570159e-01 -6.05634689e-01 2.30967611e-01
5.73109508e-01 7.77014531e-03 -1.00798440e+00 5.11477530e-01
8.56881201e-01 -1.03892136e+00 1.11312318e+00 -9.63749945e-01
6.36794388e-01 -3.90391886e-01 -6.37991130e-01 -7.90368140e-01
-6.28579795e-01 -4.80127484e-01 -3.91698748e-01 1.13136959e+00
-3.78941819e-02 -2.51451164e-01 7.14620173e-01 1.03047408e-01
-2.13659689e-01 -1.25995839e+00 -1.05389047e+00 -1.22988379e+00
-2.38108411e-01 -6.30752027e-01 3.63716573e-01 8.23658407e-01
1.26288608e-01 7.92736039e-02 -4.34199959e-01 -1.52107507e-01
4.64042813e-01 4.00899887e-01 6.76062405e-01 -8.79306555e-01
-5.22160470e-01 -4.68754113e-01 -8.96992207e-01 -9.57388699e-01
1.13449529e-01 -9.18950260e-01 -4.50096512e-03 -1.44659948e+00
5.74329436e-01 -4.57809418e-02 -9.30013418e-01 8.27382565e-01
-2.48641998e-01 4.43147004e-01 3.99797529e-01 2.33193919e-01
-1.22332811e+00 8.05699050e-01 1.11311388e+00 -3.41573834e-01
-1.67048991e-01 -1.79475322e-01 -2.61564493e-01 5.96889377e-01
6.04927838e-01 -3.92193705e-01 -6.50049090e-01 -6.84404150e-02
-4.77882862e-01 -8.68576169e-02 7.03590631e-01 -1.33603156e+00
4.86137085e-02 -4.77677703e-01 2.85254568e-01 -6.35199308e-01
7.02653885e-01 -6.80706918e-01 -4.30313051e-02 5.58426619e-01
-6.53864384e-01 -2.75767148e-01 1.57904714e-01 8.86414409e-01
-3.18902373e-01 1.74807668e-01 1.04123116e+00 -1.29057258e-01
-1.29660094e+00 7.37939477e-01 -1.04539789e-01 2.26855785e-01
1.41964734e+00 -2.71315187e-01 -1.47110984e-01 -2.53260702e-01
-8.04550111e-01 1.15196608e-01 4.49219525e-01 6.09903157e-01
8.16257894e-01 -1.68800342e+00 -6.51384890e-01 -2.43563857e-02
7.86035419e-01 -6.94958448e-01 7.03495026e-01 1.14843261e+00
7.61229694e-02 2.72438169e-01 -3.95985126e-01 -7.42870927e-01
-1.41029942e+00 7.11825788e-01 2.24774987e-01 -4.15216535e-01
-9.22161639e-01 8.06838870e-01 3.97999316e-01 1.48767248e-01
3.01340640e-01 -2.82410502e-01 -1.83586955e-01 1.10619843e-01
7.67957151e-01 6.33616865e-01 -2.08268628e-01 -9.66022670e-01
-8.37577879e-01 6.21130586e-01 6.61529377e-02 6.67543616e-03
1.45951569e+00 1.32968560e-01 2.72620201e-01 6.79702461e-01
1.27951157e+00 -4.99191910e-01 -1.49772024e+00 -5.79436183e-01
9.52521637e-02 -7.81594753e-01 -2.58371294e-01 -7.12907970e-01
-1.20954585e+00 8.63527834e-01 7.04933941e-01 -3.22198361e-01
1.42072821e+00 2.09232867e-01 7.81385779e-01 1.77565366e-01
5.40528774e-01 -1.03333008e+00 7.31012642e-01 4.28770840e-01
9.24304187e-01 -1.21025181e+00 6.00159094e-02 -2.58242458e-01
-8.02555263e-01 7.91495383e-01 6.72440588e-01 -3.40003639e-01
5.77489853e-01 -1.52762178e-02 -4.05810982e-01 -4.33322877e-01
-1.00033772e+00 -4.80882198e-01 4.72049147e-01 4.58559871e-01
3.27710122e-01 -6.87322840e-02 -3.45766217e-01 7.59504318e-01
6.93427563e-01 2.34499782e-01 -2.61841044e-02 1.45774889e+00
-3.96438718e-01 -1.05078232e+00 4.48106751e-02 4.11922902e-01
-2.42682725e-01 9.59743634e-02 -3.16903621e-01 6.65938556e-01
1.62449092e-01 8.02644968e-01 1.89763736e-02 -6.91665471e-01
5.43423176e-01 1.64088175e-01 5.46388924e-01 -6.12545013e-01
-6.47883832e-01 4.21055816e-02 -2.23746076e-02 -1.30947745e+00
-7.43383348e-01 -9.44911122e-01 -1.39024282e+00 1.59129843e-01
1.54060528e-01 -1.74914658e-01 -1.70595437e-01 8.46375287e-01
7.44945943e-01 6.50588155e-01 6.59471512e-01 -8.70802283e-01
-3.93406719e-01 -8.58555555e-01 -5.63896239e-01 7.73064673e-01
1.39219120e-01 -1.17656934e+00 -3.93568762e-02 1.65261209e-01] | [8.436676979064941, 0.789359986782074] |
801e9b00-f915-4a43-85a3-b520d83d33ce | the-role-of-packaging-sites-in-efficient-and | 1502.05029 | null | http://arxiv.org/abs/1502.05029v1 | http://arxiv.org/pdf/1502.05029v1.pdf | The role of packaging sites in efficient and specific virus assembly | During the lifecycle of many single-stranded RNA viruses, including many
human pathogens, a protein shell called the capsid spontaneously assembles
around the viral genome. Understanding the mechanisms by which capsid proteins
selectively assemble around the viral RNA amidst diverse host RNAs is a key
question in virology. In one proposed mechanism, sequence elements (packaging
sites) within the genomic RNA promote rapid and efficient assembly through
specific interactions with the capsid proteins. In this work we develop a
coarse-grained particle-based computational model for capsid proteins and RNA
which represents protein-RNA interactions arising both from non-specific
electrostatics and specific packaging sites interactions. Using Brownian
dynamics simulations, we explore how the efficiency and specificity of assembly
depend on solution conditions (which control protein-protein and nonspecific
protein-RNA interactions) as well as the strength and number of packaging
sites. We identify distinct regions in parameter space in which packaging sites
lead to highly specific assembly via different mechanisms, and others in which
packaging sites lead to kinetic traps. We relate these computational
predictions to in vitro assays for specificity in which cognate viral RNAs are
compete against non-cognate RNAs for assembly by capsid proteins. | [] | 2015-02-17 | null | null | null | null | ['genome-understanding', 'virology'] | ['medical', 'miscellaneous'] | [ 2.45096400e-01 -3.64126354e-01 -4.64159325e-02 -1.39327357e-02
-2.80670226e-01 -1.52118945e+00 6.19353175e-01 1.63981155e-01
-4.16191906e-01 9.36837494e-01 3.70912135e-01 -6.87166154e-01
2.69537836e-01 -6.40649021e-01 -6.87569618e-01 -1.15933740e+00
-2.04535022e-01 1.02064013e+00 2.13142529e-01 3.91599312e-02
3.80584598e-01 7.86529124e-01 -8.71465027e-01 1.98632151e-01
8.37455332e-01 -5.43209948e-02 6.38411343e-01 9.40144002e-01
-4.92386073e-01 -1.66884422e-01 -5.43477274e-02 9.07133445e-02
9.17346030e-02 -3.89243454e-01 -9.56804678e-02 -9.24493372e-02
-4.46461141e-01 4.42471690e-02 4.78042245e-01 4.57449526e-01
4.72732961e-01 -3.89377534e-01 1.17970943e+00 -6.29713476e-01
-6.76511049e-01 -3.64684463e-01 -3.51857185e-01 -8.60697869e-03
2.92269558e-01 4.18464839e-01 9.80423033e-01 -1.09050393e+00
1.37448609e+00 1.10727048e+00 5.55483937e-01 6.74179673e-01
-1.63688576e+00 1.76245391e-01 -2.86198229e-01 -5.62556088e-01
-9.41022098e-01 -3.67614061e-01 4.02714252e-01 -8.46086383e-01
1.36136985e+00 5.22735357e-01 7.98807502e-01 6.57112360e-01
1.03635097e+00 1.02290966e-01 1.07808137e+00 5.52743375e-02
5.53181171e-01 -2.35998198e-01 4.06814218e-01 4.37846601e-01
6.42323017e-01 -3.36389244e-02 2.34251812e-01 -1.16861165e+00
5.29750764e-01 -8.84722918e-03 -2.75674164e-01 -8.87383282e-01
-8.20052564e-01 9.07833338e-01 -3.82123649e-01 6.40899390e-02
-6.26645267e-01 -3.41516733e-01 3.27195227e-01 -4.80295211e-01
-1.08266965e-01 3.44878554e-01 -7.58156061e-01 -1.58011436e-01
-1.93890423e-01 1.41971707e-01 8.77084494e-01 1.37465194e-01
6.59921348e-01 -6.70079350e-01 1.21388502e-01 5.75853467e-01
8.06655943e-01 7.66069531e-01 -2.93928534e-01 -1.22315660e-01
7.38909841e-02 3.35641652e-01 6.43666208e-01 -4.63242620e-01
-7.23733723e-01 2.17698477e-02 -4.71519053e-01 9.81438207e-04
5.85619271e-01 -3.92117321e-01 -8.68909538e-01 1.78606367e+00
3.77701789e-01 -3.56020600e-01 2.86644161e-01 7.08157480e-01
3.76067847e-01 1.03904414e+00 3.22839171e-01 -1.15778756e+00
1.73164904e+00 -3.02039415e-01 -4.79450375e-01 -1.95450634e-01
1.98312372e-01 -8.50103140e-01 5.71738839e-01 -3.48846108e-01
-8.81807387e-01 1.98450878e-01 -7.08120346e-01 2.88021475e-01
-7.27050677e-02 -3.22827429e-01 3.36083889e-01 6.28046215e-01
-1.00766563e+00 2.42812842e-01 -1.06847429e+00 -9.99514639e-01
-3.16580206e-01 4.06200916e-01 -2.02660426e-01 2.91525304e-01
-6.66537881e-01 1.10998058e+00 1.15691043e-01 9.51700434e-02
-5.84416687e-01 -1.43302813e-01 -3.70921552e-01 -5.38929738e-02
-4.17537928e-01 -8.45797479e-01 7.58312345e-01 -3.07779253e-01
-1.18584883e+00 9.05579209e-01 -4.51114029e-01 2.88310857e-03
-1.29161313e-01 1.80803716e-01 1.59311280e-01 1.58712551e-01
-2.36842334e-02 2.55770653e-01 2.10490882e-01 -1.57954383e+00
2.19465539e-01 -5.20346284e-01 -6.26232505e-01 2.71776676e-01
8.93513441e-01 5.21628022e-01 -5.77419251e-02 -1.90059558e-01
2.74281919e-01 -1.53150809e+00 -4.89396393e-01 -3.71323854e-01
-1.77305803e-01 -1.32050455e-01 7.41645038e-01 -2.66245186e-01
4.40474153e-01 -1.76358736e+00 4.60382432e-01 4.76546109e-01
6.03596978e-02 5.55268764e-01 -1.12723909e-01 1.14332247e+00
1.63439617e-01 1.56802684e-01 -9.04790498e-03 5.87825239e-01
-2.85604507e-01 1.13418221e-01 -2.85504162e-01 7.25575626e-01
2.64277607e-01 9.88713920e-01 -7.56119132e-01 5.49837612e-02
9.98829082e-02 1.02426803e+00 -4.77796346e-01 2.69836098e-01
-5.58574915e-01 7.70400167e-01 -8.31988573e-01 3.47585887e-01
1.06154394e+00 -5.03759086e-01 1.28337145e+00 -1.07222550e-01
-5.54469407e-01 6.14864469e-01 -4.56265002e-01 7.61007309e-01
3.18332195e-01 2.81931221e-01 4.51850593e-01 9.82553735e-02
6.69093549e-01 3.40298176e-01 9.59997028e-02 -3.73273827e-02
1.45609453e-01 1.07954629e-01 3.17259163e-01 1.32273315e-02
1.76921070e-01 -6.07383311e-01 2.72215486e-01 6.10495985e-01
-4.86278504e-01 1.13357969e-01 -9.84982625e-02 2.47856379e-01
9.66543794e-01 1.00670755e-01 5.20345509e-01 -7.24909008e-01
7.98807219e-02 8.70745331e-02 8.23684096e-01 4.77597922e-01
-2.09793851e-01 5.85013270e-01 6.90775812e-01 -3.71225327e-01
-1.66063225e+00 -1.29824901e+00 -2.76456743e-01 9.86120701e-01
3.61749619e-01 -4.33983237e-01 -8.43521893e-01 1.83871582e-01
-2.48934552e-02 4.21031058e-01 -2.92042673e-01 1.34620979e-01
-6.29221559e-01 -1.48904800e+00 1.84175447e-02 1.28580764e-01
-5.70332170e-01 -1.19585443e+00 -5.68451941e-01 6.26662672e-01
-1.06355712e-01 -5.68736494e-01 -5.62650144e-01 3.08925271e-01
-7.78710425e-01 -1.18558729e+00 -2.61617959e-01 -4.75850910e-01
8.40650618e-01 5.40427089e-01 7.31476903e-01 2.77966529e-01
-2.52168983e-01 1.93157136e-01 8.75909813e-03 -1.56823099e-01
-7.87849844e-01 -2.53241211e-01 3.66184592e-01 -5.31343162e-01
3.81973594e-01 -5.32781780e-01 -8.31771374e-01 2.11199611e-01
-6.13550186e-01 2.07491294e-01 4.57851022e-01 5.78625143e-01
4.56095755e-01 -8.54956627e-01 3.16615939e-01 -6.14609778e-01
6.98312283e-01 -6.24545038e-01 -7.08099961e-01 2.14214414e-01
-3.57102334e-01 2.07090929e-01 4.54673976e-01 1.02630451e-01
-1.24765170e+00 1.47375807e-01 -3.89400274e-01 6.86481833e-01
1.08109869e-01 8.43601003e-02 -4.63551551e-01 1.07410438e-01
3.47343057e-01 8.11780810e-01 4.92370069e-01 -1.69538707e-01
9.02694538e-02 3.25090289e-01 -3.56134921e-01 -6.74606383e-01
8.13792706e-01 6.58239067e-01 -5.25637269e-02 -9.31378365e-01
-4.28241305e-02 -2.35794768e-01 -4.10074532e-01 9.03154723e-03
1.35876620e+00 -5.98277211e-01 -1.10287654e+00 4.80985492e-01
-1.57027614e+00 -1.10636286e-01 4.34607834e-01 4.46948767e-01
-6.01707339e-01 4.46264535e-01 -1.11334968e+00 -9.84228551e-01
-2.50616401e-01 -1.50114381e+00 9.65974391e-01 2.69556165e-01
-2.44977161e-01 -7.64346182e-01 1.01155341e+00 5.70450425e-02
3.93552721e-01 -3.99508551e-02 1.30907249e+00 -6.84822500e-01
-1.07397318e+00 3.94838691e-01 -1.96682528e-01 -3.49597394e-01
1.70997649e-01 5.67372561e-01 -2.78020620e-01 -4.16046262e-01
-1.94417313e-01 1.62419379e-01 5.00896037e-01 7.76892483e-01
-3.50256324e-01 -3.26596797e-01 -1.01510835e+00 2.55981386e-01
1.55036330e+00 9.45387006e-01 7.84694135e-01 3.77547354e-01
4.92044091e-01 4.18287009e-01 3.90564293e-01 3.40200424e-01
-2.74385344e-02 4.37993377e-01 1.68769926e-01 2.76956141e-01
6.22170150e-01 9.63299572e-02 2.18763173e-01 7.40600586e-01
-3.05265516e-01 -5.85809112e-01 -1.15342331e+00 2.11926028e-01
-1.46179986e+00 -8.86413991e-01 -3.94785762e-01 2.13977814e+00
1.07205415e+00 6.51430711e-02 6.72867447e-02 -8.51544261e-01
7.86273420e-01 -1.28177647e-02 -4.30833101e-01 -6.58182144e-01
-3.39482039e-01 -1.89269453e-01 5.59498847e-01 1.01001108e+00
-7.49331117e-01 6.09881520e-01 7.45497036e+00 1.97626978e-01
-9.00489986e-01 -1.70276724e-02 3.07351977e-01 1.55976728e-01
-7.69235611e-01 5.12637436e-01 -7.87351906e-01 4.29298013e-01
6.57061517e-01 2.32482463e-01 2.34126121e-01 5.22984028e-01
4.67521280e-01 -3.56221125e-02 -5.80580950e-01 2.70553648e-01
-4.65606600e-01 -1.91314614e+00 -9.56271067e-02 5.26922643e-01
7.96737671e-01 3.65284801e-01 -3.50897014e-01 -6.18951857e-01
1.72598168e-01 -7.31846154e-01 -6.60322681e-02 4.11371678e-01
3.69376719e-01 -5.11346936e-01 4.67518181e-01 4.32269901e-01
-1.10571969e+00 3.20616275e-01 -2.63418615e-01 -2.43237894e-02
5.58365703e-01 4.81587648e-01 -1.23646617e+00 -4.58133519e-01
3.08581870e-02 -3.57622951e-01 3.81114185e-01 5.62810183e-01
1.14157125e-01 3.25185359e-01 -3.76583606e-01 -3.50148261e-01
2.04007208e-01 -9.64562953e-01 9.59912062e-01 1.31431735e+00
-1.64172381e-01 7.11192727e-01 -7.02506080e-02 6.28229439e-01
2.86847264e-01 1.24510117e-01 -7.46232569e-01 -2.64783144e-01
4.21328515e-01 1.31246996e+00 -1.23462892e+00 -1.08490750e-01
-3.77299070e-01 7.99328744e-01 2.96009600e-01 4.65823621e-01
-4.33663577e-01 2.28777274e-01 1.42541718e+00 3.11174661e-01
5.29092550e-01 -6.85935557e-01 1.77026466e-01 -7.36874938e-01
-2.82505244e-01 -4.61711943e-01 -2.52901316e-01 -6.46122396e-01
-1.06139660e+00 2.88853228e-01 -6.93301737e-01 -4.93187755e-01
-7.62538835e-02 -4.67316061e-01 -6.08136475e-01 8.93594980e-01
-7.39295065e-01 -8.57241929e-01 5.03976345e-01 -3.58064324e-01
1.07456334e-01 1.89210907e-01 9.72194910e-01 -5.04130065e-01
-3.20100933e-01 -4.30663943e-01 9.96774077e-01 -1.11999214e-01
4.48535681e-01 -1.01909029e+00 8.40195835e-01 6.16101027e-01
-4.69091713e-01 1.46057868e+00 1.30042744e+00 -1.32664037e+00
-1.93499827e+00 -9.28883314e-01 9.78738666e-01 -9.19618428e-01
3.48800778e-01 -8.54201138e-01 -7.67746449e-01 8.36714327e-01
3.72012049e-01 -4.85943645e-01 1.25276530e+00 -6.88253939e-02
-3.24165970e-01 7.87085831e-01 -1.19705057e+00 8.52459013e-01
6.63443148e-01 -4.82861072e-01 -5.65817952e-01 5.38647056e-01
5.69192469e-01 -2.09816620e-01 -3.98806095e-01 3.28574121e-01
7.51071095e-01 -1.03955305e+00 7.98185050e-01 -9.37759459e-01
-2.76097059e-01 -7.63460100e-01 2.91990153e-02 -1.04246557e+00
-7.14517891e-01 -7.15451419e-01 3.61379713e-01 6.11084044e-01
4.47917461e-01 -6.07136965e-01 4.91017401e-01 8.12213182e-01
2.25070745e-01 -6.43946469e-01 -9.21858788e-01 -3.93194169e-01
-1.90576434e-01 3.74928981e-01 2.09980354e-01 5.77945411e-01
1.64055660e-01 7.93001115e-01 -2.27616236e-01 1.17855921e-01
5.09673178e-01 3.55508357e-01 3.55209678e-01 -1.10654247e+00
4.41601202e-02 5.23597598e-02 -5.40215969e-02 -8.81878436e-01
-1.04732282e-01 -5.17075181e-01 1.99076667e-01 -1.80259967e+00
9.70789909e-01 -6.27798617e-01 2.56445378e-01 -6.71580732e-02
-1.17043473e-01 -1.53000832e-01 2.93949544e-01 3.31577241e-01
-4.83029187e-01 3.68440479e-01 8.33831012e-01 5.07511556e-01
-5.02518952e-01 -1.75085604e-01 -4.57135618e-01 5.12299657e-01
7.95058846e-01 -4.87453908e-01 -3.21724325e-01 6.42687157e-02
6.93987012e-01 2.57365674e-01 -6.96123391e-02 -7.77174905e-02
-3.41252118e-01 -4.63896126e-01 2.10016161e-01 -6.37112439e-01
4.14633006e-01 -6.91116929e-01 8.57634008e-01 8.30863416e-01
3.28443319e-01 1.94243014e-01 2.92664677e-01 6.29245102e-01
7.69390285e-01 3.01330332e-02 9.45520461e-01 -1.21243201e-01
1.44384250e-01 -4.21990715e-02 -1.53233886e+00 -3.78721729e-02
1.28473723e+00 -2.43892789e-01 -7.80375719e-01 7.74799613e-03
-8.45954776e-01 -3.88298959e-01 1.23710728e+00 1.93636253e-01
2.70664394e-01 -9.54035401e-01 -5.19324720e-01 7.80793056e-02
-1.69974059e-01 -7.04385817e-01 3.42598468e-01 5.66127896e-01
-1.08354652e+00 1.04425955e+00 -2.36250743e-01 -8.81428123e-01
-1.38523781e+00 8.90244544e-01 3.20600837e-01 -2.18733530e-02
-2.02250764e-01 6.70186341e-01 5.93812346e-01 -7.03643024e-01
-6.58095002e-01 9.87826735e-02 1.01203963e-01 -1.96534261e-01
6.73516572e-01 2.61403508e-02 -2.00169250e-01 -7.93241441e-01
-8.77580941e-01 6.79900348e-01 -4.51939374e-01 2.99242556e-01
1.14656365e+00 -8.92262831e-02 -8.25919747e-01 1.62175357e-01
8.00964236e-01 4.32081163e-01 -1.22531974e+00 2.38599107e-01
-8.66584480e-02 -2.20820740e-01 -8.48230660e-01 -6.09591007e-01
-2.69413114e-01 2.22666964e-01 3.74951273e-01 1.11620612e-01
4.93273526e-01 5.33978760e-01 6.12938583e-01 1.07754074e-01
5.99303007e-01 -7.09170580e-01 -4.09541011e-01 4.74307686e-01
5.25623381e-01 -8.71622860e-01 -2.38907680e-01 -4.95021611e-01
-4.90907997e-01 6.83072090e-01 4.68492150e-01 5.14704362e-02
3.47702295e-01 5.45402765e-01 2.56749958e-01 -2.68511117e-01
-1.41628373e+00 7.09920302e-02 -3.15874010e-01 6.89669073e-01
7.15038717e-01 3.94769460e-01 -1.17036498e+00 2.90411294e-01
4.81077045e-01 -3.71830165e-01 6.16266549e-01 1.02857983e+00
-8.79259825e-01 -1.53298604e+00 -5.44622958e-01 3.16477388e-01
-3.09801638e-01 -3.56570303e-01 -7.17528343e-01 3.50582004e-01
-2.13791654e-01 6.68044209e-01 2.69475877e-01 3.21508378e-01
-2.40400299e-01 3.89944494e-01 4.25525337e-01 -7.04845071e-01
-3.81423563e-01 5.69655120e-01 1.41892761e-01 -1.35820746e-01
-1.30009815e-01 -8.95866573e-01 -1.89625418e+00 -3.38481277e-01
-3.24571878e-01 6.60125554e-01 8.07067633e-01 8.31349134e-01
1.02259207e+00 -2.08482489e-01 2.77366042e-01 -6.92087948e-01
-2.26441845e-01 -4.44076151e-01 -5.97517133e-01 -6.41231015e-02
3.68192941e-01 -3.36173654e-01 -2.99693972e-01 -4.28177208e-01] | [4.78074836730957, 5.22767448425293] |
2567b5f5-bc41-4c6a-9cac-e056a4446fd3 | visual-goal-step-inference-using-wikihow | 2104.05845 | null | https://arxiv.org/abs/2104.05845v2 | https://arxiv.org/pdf/2104.05845v2.pdf | Visual Goal-Step Inference using wikiHow | Understanding what sequence of steps are needed to complete a goal can help artificial intelligence systems reason about human activities. Past work in NLP has examined the task of goal-step inference for text. We introduce the visual analogue. We propose the Visual Goal-Step Inference (VGSI) task, where a model is given a textual goal and must choose which of four images represents a plausible step towards that goal. With a new dataset harvested from wikiHow consisting of 772,277 images representing human actions, we show that our task is challenging for state-of-the-art multimodal models. Moreover, the multimodal representation learned from our data can be effectively transferred to other datasets like HowTo100m, increasing the VGSI accuracy by 15 - 20%. Our task will facilitate multimodal reasoning about procedural events. | ['Chris Callison-Burch', 'Mark Yatskar', 'Li Zhang', 'Qing Lyu', 'Artemis Panagopoulou', 'Yue Yang'] | 2021-04-12 | null | https://aclanthology.org/2021.emnlp-main.165 | https://aclanthology.org/2021.emnlp-main.165.pdf | emnlp-2021-11 | ['vgsi'] | ['computer-vision'] | [ 4.35266435e-01 6.23757422e-01 -1.23520002e-01 -3.24057519e-01
-8.46459270e-01 -6.05146945e-01 1.13030183e+00 1.77375183e-01
-3.99917871e-01 7.74259031e-01 7.10886359e-01 -4.10861403e-01
1.10082507e-01 -6.35347486e-01 -8.67984295e-01 -2.26881325e-01
1.20361038e-02 8.88976574e-01 3.45268212e-02 -2.55284786e-01
3.03172559e-01 1.26135454e-01 -1.47907043e+00 1.05032420e+00
5.05953431e-01 8.04672062e-01 2.60655463e-01 9.45768416e-01
-3.62431735e-01 1.53800559e+00 -4.64351922e-01 -6.11532271e-01
-1.49758726e-01 -7.10479021e-01 -1.45313966e+00 5.26035354e-02
6.26509428e-01 -4.82195407e-01 -3.63649637e-01 8.81587863e-01
1.79440290e-01 5.57093263e-01 9.85417247e-01 -1.53068173e+00
-5.60540318e-01 8.30825269e-01 -3.00938636e-01 -2.98079848e-01
9.68877077e-01 5.00595391e-01 9.66334641e-01 -7.50993192e-01
1.06348717e+00 1.76418102e+00 7.67319724e-02 7.42290735e-01
-1.14427912e+00 -1.53584629e-01 3.51459622e-01 5.48473895e-01
-9.84228671e-01 -5.55544913e-01 6.06882751e-01 -4.25216109e-01
1.25618637e+00 2.70792067e-01 6.26269996e-01 1.28789377e+00
-8.23585242e-02 1.10645092e+00 7.96637297e-01 -4.10737574e-01
1.06121525e-01 -3.83063257e-01 -2.06955001e-02 1.19171572e+00
-6.48239776e-02 -2.24241167e-01 -1.06551731e+00 7.93997347e-02
7.52925813e-01 -2.53690034e-01 -9.20775309e-02 -2.44136468e-01
-1.66500294e+00 6.65241122e-01 4.67127919e-01 -1.43363401e-02
-3.74907404e-01 5.59267282e-01 2.89716393e-01 1.47746548e-01
-1.18430950e-01 5.46913445e-01 1.36644363e-01 -4.67031062e-01
-6.97026789e-01 5.18660665e-01 7.28366733e-01 1.03227127e+00
6.58915937e-01 -1.96490869e-01 -5.09719849e-01 4.26117092e-01
4.78535801e-01 7.93921113e-01 -1.10824354e-01 -1.65290594e+00
1.03109539e+00 7.88908064e-01 5.10389745e-01 -9.25505280e-01
-5.05801499e-01 5.10097265e-01 -5.19878566e-01 5.59516907e-01
1.01172864e+00 -1.69603229e-01 -1.09950411e+00 1.50031805e+00
2.42999896e-01 -2.73593068e-01 1.16581790e-01 9.29556310e-01
1.09476185e+00 8.81996334e-01 5.69485128e-01 2.10863352e-01
1.54276896e+00 -1.05930424e+00 -8.15257072e-01 -7.75498271e-01
7.43192375e-01 -3.20891380e-01 1.33327508e+00 4.98150527e-01
-1.30207551e+00 -4.17406678e-01 -6.93325043e-01 -4.72073704e-01
-5.02636611e-01 2.11960629e-01 7.09606409e-01 1.67867448e-02
-8.78460288e-01 2.60543823e-01 -7.79346347e-01 -5.31323731e-01
6.30928516e-01 -1.47924244e-01 -7.21780241e-01 -3.98261964e-01
-9.25646424e-01 1.08929467e+00 6.03619814e-01 -3.11707407e-02
-1.31476271e+00 -4.05969024e-01 -1.26625502e+00 -6.53672777e-03
7.66692936e-01 -1.00053978e+00 1.28050971e+00 -8.57124150e-01
-1.10469854e+00 1.18681729e+00 -5.14828026e-01 -4.53918725e-01
7.01414406e-01 -1.27359927e-01 -3.49737331e-02 6.66899502e-01
2.24701762e-01 1.22205925e+00 7.31900275e-01 -1.40987921e+00
-6.54630363e-01 -4.38060760e-01 5.48266232e-01 6.39892399e-01
2.76537180e-01 4.46131453e-02 -8.35971355e-01 -2.01064363e-01
-2.49497041e-01 -8.97825122e-01 -1.80675641e-01 4.45743293e-01
-5.15351892e-01 -2.40188777e-01 4.29762304e-01 -9.44553673e-01
1.03586900e+00 -1.80371988e+00 6.93662286e-01 -3.31990831e-02
1.62634015e-01 -2.14860216e-01 -2.13915735e-01 6.50836110e-01
2.16780901e-01 9.53574330e-02 -1.63003683e-01 -5.05089819e-01
3.98096025e-01 2.45084718e-01 -2.41035089e-01 1.10070087e-01
1.39885455e-01 1.30778265e+00 -9.52293277e-01 -9.51096535e-01
4.70941901e-01 2.19360307e-01 -1.77427590e-01 4.57777530e-01
-7.00923741e-01 3.85164320e-01 -2.72104323e-01 6.44059598e-01
2.56720483e-01 -5.20744681e-01 4.37330306e-01 -2.36627236e-01
2.31969461e-01 3.77287269e-02 -9.16129529e-01 2.17721033e+00
-2.28962034e-01 7.35496700e-01 -1.79951012e-01 -5.70562780e-01
4.82815802e-01 1.91461816e-02 -6.53572306e-02 -8.17633092e-01
-3.58008482e-02 -4.10173833e-01 -3.02506089e-01 -7.65601754e-01
4.30388510e-01 -1.25548914e-01 -5.02777219e-01 4.56258774e-01
5.65722436e-02 -3.54122013e-01 6.17860496e-01 5.62164843e-01
8.38891387e-01 6.11312330e-01 3.94553572e-01 2.01541871e-01
3.54757905e-01 6.68033421e-01 -1.12961210e-01 1.19653285e+00
-1.01975605e-01 4.77132618e-01 7.48147845e-01 -4.87736255e-01
-7.67186165e-01 -1.19351089e+00 7.06401467e-01 1.46444869e+00
2.48359814e-01 -4.94767666e-01 -6.29432082e-01 -7.91502595e-01
-1.35754958e-01 1.17202437e+00 -8.14775407e-01 1.25044569e-01
-6.85049593e-01 -3.19570266e-02 8.04844558e-01 7.98555970e-01
6.98899209e-01 -1.37443221e+00 -1.00142562e+00 -1.67385384e-01
-7.88250625e-01 -1.32815492e+00 -1.61611721e-01 -1.86256394e-01
-4.98971254e-01 -1.41865301e+00 -5.34844100e-01 -6.47255123e-01
8.36609423e-01 -1.27425995e-02 1.50833988e+00 1.44090205e-01
-2.89686024e-01 8.80263329e-01 -3.90719265e-01 -4.61879224e-01
-4.27260876e-01 -4.25368190e-01 -6.57116234e-01 -2.72313148e-01
2.39923805e-01 8.08408409e-02 -3.67667556e-01 6.28167111e-03
-6.11035049e-01 9.27601993e-01 3.13290685e-01 4.90541786e-01
4.43269312e-01 -2.83712208e-01 -7.16067851e-02 -7.57683456e-01
5.95231116e-01 -3.50392014e-01 -2.28167459e-01 6.32008374e-01
6.85485750e-02 1.93712637e-01 4.55641359e-01 -2.04756767e-01
-1.67605400e+00 1.67691037e-01 1.98654786e-01 -2.12847546e-01
-4.90484387e-01 6.49067283e-01 -6.23876639e-02 3.57736498e-01
6.64109886e-01 1.39928117e-01 -6.00558333e-02 -2.43307615e-04
9.90681171e-01 7.98492059e-02 7.60226905e-01 -9.46869254e-01
4.79539216e-01 8.53318632e-01 3.12968194e-01 -8.35124493e-01
-8.32604825e-01 -1.01379909e-01 -5.35495043e-01 -7.26682544e-01
1.33904850e+00 -9.27932918e-01 -1.22026753e+00 1.65099829e-01
-1.30763042e+00 -9.10241127e-01 6.00478575e-02 1.33035317e-01
-1.05371189e+00 3.77074808e-01 -4.50907826e-01 -1.00672877e+00
-1.95796892e-01 -8.54319572e-01 1.21182394e+00 2.34052479e-01
-5.56444883e-01 -1.01865363e+00 -3.01922888e-01 7.36919045e-01
-3.98670760e-04 4.09838706e-01 1.07856131e+00 -3.08263123e-01
-6.37955368e-01 1.62251681e-01 -4.16348606e-01 -4.36476558e-01
-2.93228835e-01 -1.63560539e-01 -4.99176502e-01 -1.82571262e-02
-5.97594202e-01 -8.91615331e-01 8.33771586e-01 1.92552403e-01
1.07080853e+00 -2.84394890e-01 -4.71932620e-01 1.94199920e-01
1.07752764e+00 2.42249250e-01 9.87327754e-01 3.39036644e-01
8.28836203e-01 7.86198854e-01 1.26349974e+00 4.71396297e-01
8.11165154e-01 4.56514508e-01 7.01421857e-01 -9.07953978e-02
-3.54436100e-01 -5.92539370e-01 2.86864728e-01 -3.57574731e-01
-5.20021021e-01 -5.68732679e-01 -1.41407847e+00 6.92876518e-01
-2.32989073e+00 -1.23578024e+00 -4.40211184e-02 1.70667505e+00
6.79422259e-01 -4.01902944e-02 1.12370864e-01 -3.18103820e-01
3.31483245e-01 3.23483825e-01 -3.55719835e-01 -2.62328953e-01
7.28557408e-02 -2.23968789e-01 9.81894583e-02 6.57357931e-01
-1.11660326e+00 1.24426508e+00 6.53965998e+00 6.14489555e-01
-4.36390400e-01 -2.81375051e-01 4.87806171e-01 -1.16327778e-01
-2.77496606e-01 6.55469447e-02 -7.23275542e-01 -1.72464997e-02
6.84567630e-01 -7.67467692e-02 7.37466753e-01 5.25981486e-01
2.65638679e-01 -6.73802018e-01 -1.46462023e+00 1.22986817e+00
5.84344149e-01 -1.39351761e+00 5.67407489e-01 -3.35399359e-01
3.38825196e-01 -5.05656719e-01 -1.64729327e-01 4.74864542e-01
8.17780733e-01 -1.60851562e+00 8.53222430e-01 1.04858541e+00
8.30739677e-01 -6.90227807e-01 2.11284965e-01 5.86641908e-01
-1.17087960e+00 -4.39660586e-02 -2.84300949e-02 -1.83505774e-01
5.47094584e-01 -2.32321158e-01 -1.11918664e+00 5.46194553e-01
6.17178261e-01 7.06961095e-01 -8.37582350e-01 5.24534345e-01
-6.57691240e-01 1.62393078e-01 -8.07317197e-02 -1.87211871e-01
3.15380663e-01 -2.02690214e-01 4.76165503e-01 1.25736034e+00
1.90223545e-01 9.91616026e-02 1.68918699e-01 9.22501087e-01
1.18676750e-02 -3.24745685e-01 -8.77510548e-01 -3.59706879e-01
2.44692087e-01 9.20610189e-01 -5.61140180e-01 -6.62139237e-01
-2.53701419e-01 1.28529298e+00 7.20661461e-01 6.24911427e-01
-1.03931999e+00 -2.58908540e-01 5.34567535e-01 -4.66637425e-02
1.65238544e-01 -4.46376324e-01 -2.78390974e-01 -9.91256714e-01
-1.75914913e-01 -1.00663793e+00 7.39286005e-01 -1.75925446e+00
-8.84099066e-01 1.73757806e-01 4.14365858e-01 -8.49010587e-01
-5.75249374e-01 -7.94155419e-01 -4.42668498e-01 6.04004025e-01
-1.09199739e+00 -1.80101311e+00 -6.41129196e-01 7.68270910e-01
8.02011192e-01 1.24299049e-01 8.33114445e-01 -2.72011787e-01
9.98298824e-03 -2.68801358e-02 -7.01475024e-01 2.79655278e-01
6.51757658e-01 -1.37569702e+00 3.55164051e-01 7.16722250e-01
2.53023744e-01 5.26132345e-01 7.95799077e-01 -8.43343854e-01
-1.54786515e+00 -9.44302619e-01 8.92121375e-01 -8.57775450e-01
5.76845944e-01 -2.09352955e-01 -5.44285357e-01 1.23890185e+00
5.86607158e-01 -5.13428271e-01 4.28030163e-01 -4.37546447e-02
-4.12901342e-01 5.06235659e-01 -8.54653358e-01 1.33766043e+00
1.54360104e+00 -6.13421261e-01 -9.89986777e-01 2.98099220e-01
4.19810176e-01 -6.08720005e-01 -5.83074033e-01 8.90607983e-02
4.47275579e-01 -7.85986483e-01 1.24685693e+00 -1.33875465e+00
9.70562696e-01 -4.38764811e-01 -1.75630808e-01 -1.24263787e+00
-8.98730606e-02 -2.64678717e-01 -2.85322994e-01 8.20590973e-01
4.20047998e-01 1.11056969e-01 6.42087162e-01 9.84450936e-01
9.13138166e-02 -2.82051504e-01 -7.63501048e-01 -3.24315995e-01
-1.21323213e-01 -6.68337464e-01 3.22235227e-01 6.59228384e-01
3.88693422e-01 4.14415866e-01 -6.55069947e-01 1.63360313e-02
6.06634855e-01 1.35820359e-01 1.24501860e+00 -8.94625008e-01
-1.18657418e-01 -2.61262506e-01 -6.93850666e-02 -1.28458834e+00
3.01116616e-01 -8.24738204e-01 8.75933692e-02 -2.36116958e+00
4.17613238e-01 4.02596921e-01 3.68395597e-01 9.72481728e-01
-1.15784600e-01 7.13686319e-03 6.48849010e-01 -1.11559592e-01
-1.30853105e+00 3.24645877e-01 1.51412237e+00 -4.76777881e-01
6.97427690e-02 -7.43425488e-01 -6.77259982e-01 8.68468583e-01
5.58554113e-01 -1.06204659e-01 -4.27868724e-01 -5.91252029e-01
7.01910019e-01 5.63428104e-01 7.03783929e-01 -7.48016894e-01
4.20908034e-01 -6.39949322e-01 5.82731903e-01 -7.23064423e-01
8.58485520e-01 -7.97477543e-01 2.93593705e-01 2.83390790e-01
-6.37506247e-01 -1.19294479e-01 2.67760903e-01 6.48581088e-01
-1.05940774e-01 -1.62115037e-01 2.53155470e-01 -5.16111374e-01
-1.27888024e+00 -2.12987155e-01 -6.86570346e-01 2.36870795e-01
1.15111113e+00 -5.00640273e-02 -8.18823814e-01 -8.66569519e-01
-1.12069738e+00 6.31238639e-01 4.23297226e-01 4.02843475e-01
9.37363803e-01 -1.34432364e+00 -6.62744284e-01 -4.55593467e-01
4.76910055e-01 -1.45684153e-01 4.07014787e-01 6.03589416e-01
-6.87481046e-01 2.87482411e-01 -3.99397403e-01 -4.80860442e-01
-1.62994242e+00 4.67071891e-01 1.39080659e-01 -2.81394303e-01
-5.08594632e-01 7.46567905e-01 2.49993667e-01 -2.71195859e-01
2.20447421e-01 -9.73565876e-03 -4.10644174e-01 2.19590194e-03
6.40361488e-01 3.50166023e-01 -4.85931963e-01 -6.40909851e-01
-4.50526118e-01 4.27117705e-01 1.78484708e-01 -5.12256742e-01
1.04336643e+00 -2.04146311e-01 -8.66670813e-03 5.48084557e-01
7.72469938e-01 -5.07345915e-01 -1.24251544e+00 8.04965869e-02
-1.66521847e-01 -4.67669964e-01 -4.61472631e-01 -1.20648873e+00
-3.45722914e-01 9.93266582e-01 1.15003355e-01 -1.71343982e-01
7.86694407e-01 3.71392637e-01 2.05262095e-01 9.62729454e-01
5.21459162e-01 -1.20212579e+00 5.25180101e-01 6.64153099e-01
1.38073266e+00 -1.61923659e+00 -3.64841893e-02 -2.62714624e-01
-1.45602691e+00 1.19292390e+00 6.67098999e-01 3.71823788e-01
-1.17722571e-01 -9.86212492e-02 4.46236245e-02 -5.87292910e-01
-8.74927998e-01 -5.03364623e-01 4.21081811e-01 7.77193010e-01
2.52041966e-01 2.01066792e-01 7.19095767e-02 4.73526925e-01
6.93820864e-02 3.10381651e-02 3.80513668e-01 1.09670389e+00
-4.90318030e-01 -6.07003331e-01 -5.13625801e-01 2.35176742e-01
6.37706071e-02 -5.91497421e-02 -8.10668111e-01 8.92827034e-01
-3.11224639e-01 1.17202795e+00 2.23594308e-02 6.65471330e-02
3.36937457e-01 3.82467747e-01 8.70768785e-01 -4.96788383e-01
-3.30880940e-01 -3.64646822e-01 8.42555523e-01 -1.03264666e+00
-6.57891512e-01 -5.46947300e-01 -1.79974389e+00 -2.54502773e-01
6.00596547e-01 -3.36394429e-01 4.26324755e-01 9.51499820e-01
2.03022078e-01 4.98859584e-01 -4.30554092e-01 -1.01212287e+00
-2.73386352e-02 -6.67381287e-01 -1.20441668e-01 9.85014915e-01
-1.33364171e-01 -6.11006260e-01 -1.54029891e-01 5.94313204e-01] | [10.670262336730957, 1.662308931350708] |
3b10d11d-d048-442d-9417-33b03d3c8e51 | s4rl-surprisingly-simple-self-supervision-for | 2103.06326 | null | https://arxiv.org/abs/2103.06326v2 | https://arxiv.org/pdf/2103.06326v2.pdf | S4RL: Surprisingly Simple Self-Supervision for Offline Reinforcement Learning | Offline reinforcement learning proposes to learn policies from large collected datasets without interacting with the physical environment. These algorithms have made it possible to learn useful skills from data that can then be deployed in the environment in real-world settings where interactions may be costly or dangerous, such as autonomous driving or factories. However, current algorithms overfit to the dataset they are trained on and exhibit poor out-of-distribution generalization to the environment when deployed. In this paper, we study the effectiveness of performing data augmentations on the state space, and study 7 different augmentation schemes and how they behave with existing offline RL algorithms. We then combine the best data performing augmentation scheme with a state-of-the-art Q-learning technique, and improve the function approximation of the Q-networks by smoothening out the learned state-action space. We experimentally show that using this Surprisingly Simple Self-Supervision technique in RL (S4RL), we significantly improve over the current state-of-the-art algorithms on offline robot learning environments such as MetaWorld [1] and RoboSuite [2,3], and benchmark datasets such as D4RL [4]. | ['Animesh Garg', 'Ajay Mandlekar', 'Samarth Sinha'] | 2021-03-10 | null | null | null | null | ['d4rl'] | ['robots'] | [-4.09630202e-02 2.85119653e-01 -3.66793513e-01 -1.82549968e-01
-4.86817628e-01 -6.64992750e-01 5.73263049e-01 2.16491669e-01
-8.17026019e-01 1.10803735e+00 -1.64823115e-01 -5.72598755e-01
-3.46257836e-01 -6.67348266e-01 -1.34912062e+00 -5.70332110e-01
-5.90684354e-01 8.10122013e-01 4.24197555e-01 -7.01063216e-01
9.64598730e-02 5.23852706e-01 -1.87956989e+00 -3.54592532e-01
7.69855618e-01 8.19886804e-01 1.65276602e-01 9.25727844e-01
3.07737082e-01 1.20427811e+00 -6.24156475e-01 2.58919805e-01
6.93225682e-01 -3.06244791e-01 -6.89292908e-01 -1.46614499e-02
1.39446855e-01 -6.86422110e-01 -7.37290621e-01 6.43922567e-01
2.71199971e-01 7.85886347e-01 2.07460731e-01 -1.48267007e+00
-3.40871006e-01 4.46684241e-01 -2.06865475e-01 4.54963036e-02
2.41715088e-01 7.19162941e-01 6.41161621e-01 -8.67562518e-02
7.11290836e-01 1.32827473e+00 4.37708020e-01 8.11511278e-01
-1.25669193e+00 -6.74738765e-01 3.40790957e-01 1.56693697e-01
-5.61108947e-01 -3.59635323e-01 4.96693671e-01 -1.55055597e-01
1.21067166e+00 -3.39971602e-01 7.65398443e-01 1.44301999e+00
1.51319116e-01 1.08196604e+00 1.44603074e+00 -3.68585438e-01
6.60319448e-01 -1.13757007e-01 -2.03673229e-01 7.22802520e-01
4.11371849e-02 8.22087884e-01 -5.87828934e-01 -7.91626945e-02
8.78499091e-01 5.95772080e-03 2.42372081e-01 -1.02489746e+00
-1.03733170e+00 7.49069035e-01 5.34677565e-01 -1.93952724e-01
-3.28253031e-01 5.37932277e-01 4.69576567e-01 7.23517835e-01
2.81616360e-01 7.41501272e-01 -7.49382019e-01 -6.10164821e-01
-3.45737576e-01 6.78505540e-01 9.05284584e-01 1.04517853e+00
9.38220501e-01 2.31737643e-02 1.38535619e-01 5.86474121e-01
-1.48993179e-01 4.82148141e-01 6.03548348e-01 -1.26533651e+00
5.34717500e-01 3.06930482e-01 7.19998002e-01 -2.28100285e-01
-5.94721138e-01 -2.79526085e-01 -7.83056319e-02 6.19856894e-01
4.72953767e-01 -5.53042650e-01 -1.06234944e+00 1.80736387e+00
6.28264070e-01 4.09408957e-01 3.61745358e-01 7.64225900e-01
4.29074429e-02 4.17485446e-01 5.24950176e-02 1.22303851e-01
6.35697603e-01 -1.32076716e+00 -5.13368905e-01 -7.04087615e-01
8.55864167e-01 -6.05467223e-02 1.36178195e+00 6.64349973e-01
-9.86006141e-01 -7.12958217e-01 -1.27634239e+00 8.02555457e-02
-4.37092155e-01 -1.89602315e-01 8.69238853e-01 2.64201581e-01
-1.08020389e+00 1.11647463e+00 -1.47084594e+00 -3.61963123e-01
4.50438738e-01 6.98222280e-01 -4.45436567e-01 -6.50950223e-02
-1.12391996e+00 1.33352101e+00 5.01150250e-01 -3.72839659e-01
-1.73567152e+00 -6.08488441e-01 -8.30801964e-01 -4.19478923e-01
8.81173193e-01 -3.18060189e-01 1.76643968e+00 -9.16148126e-01
-1.95488369e+00 4.21177804e-01 3.55400860e-01 -8.82142782e-01
5.19156635e-01 -6.80661917e-01 -1.76015168e-01 1.05125783e-02
4.34741005e-02 8.04706395e-01 8.13399255e-01 -1.26598144e+00
-9.12624180e-01 -3.75320941e-01 5.87279081e-01 5.19709945e-01
-2.00224027e-01 -4.98314947e-01 -4.05989811e-02 -5.45468852e-02
-4.63160574e-01 -1.28249907e+00 -7.23619342e-01 -1.47852525e-01
1.30042374e-01 -2.87342012e-01 9.32418287e-01 -3.63554478e-01
5.90471327e-01 -2.00568128e+00 1.58445686e-01 2.91184802e-02
-5.04497439e-02 1.65130213e-01 -3.76983732e-01 6.92094564e-01
2.31359318e-01 -3.91535729e-01 1.03012484e-03 -5.10748506e-01
2.01653615e-01 8.43244672e-01 -2.36220732e-01 5.73451757e-01
7.26403072e-02 8.44702840e-01 -1.54060078e+00 -1.08852968e-01
2.69486547e-01 -3.13001759e-02 -5.81192017e-01 6.11872792e-01
-6.90525353e-01 8.07526648e-01 -5.33495128e-01 2.57005274e-01
1.28201798e-01 1.18129745e-01 1.35126561e-01 6.12185597e-01
-3.34534049e-02 4.59753335e-01 -9.24098909e-01 2.12196016e+00
-7.87351727e-01 4.75702703e-01 1.78562775e-01 -1.24180186e+00
8.35301638e-01 3.51230800e-02 7.04351902e-01 -8.95026982e-01
5.74351214e-02 7.32248724e-02 2.56015807e-01 -6.46022856e-01
5.55500269e-01 -1.71233580e-01 -1.41265318e-01 3.87973249e-01
3.26391429e-01 -3.61696750e-01 3.31859171e-01 2.09626667e-02
1.49885321e+00 7.62406945e-01 1.48776412e-01 -1.03989370e-01
-2.03590766e-02 3.74068230e-01 2.81703115e-01 1.06709790e+00
-4.57326263e-01 -1.50762588e-01 4.17957306e-01 -4.36574280e-01
-1.15916181e+00 -8.81109059e-01 2.79389113e-01 1.50534678e+00
2.47256219e-01 -1.11658745e-01 -7.05139399e-01 -9.71804559e-01
4.08964038e-01 8.92273009e-01 -8.11567664e-01 -5.98097980e-01
-5.80731750e-01 -3.00533980e-01 5.33841252e-01 8.00688803e-01
3.47667933e-01 -1.32308698e+00 -1.27967525e+00 4.70180511e-01
3.50868613e-01 -1.03480017e+00 3.58506329e-02 9.83777463e-01
-8.77010226e-01 -1.00076473e+00 -9.23791900e-02 -4.77977753e-01
5.11115789e-01 -3.12251523e-02 1.16924310e+00 -8.30142200e-02
-1.88279450e-01 8.27749431e-01 -5.21983385e-01 -7.18543351e-01
-5.01520514e-01 9.47453827e-02 5.91484070e-01 -4.74979550e-01
2.56841213e-01 -6.53624773e-01 -5.07035851e-01 1.56807050e-01
-6.73905313e-01 -3.58254880e-01 6.43985212e-01 1.05994272e+00
3.99732232e-01 2.32032955e-01 5.58748484e-01 -8.85443330e-01
6.24612331e-01 -5.67320466e-01 -9.00868654e-01 -8.06889161e-02
-6.33566141e-01 4.02376711e-01 9.44110930e-01 -7.92494059e-01
-8.64367366e-01 2.02399135e-01 -2.81679234e-03 -6.20603323e-01
-2.86736518e-01 1.16795830e-01 1.90731302e-01 -1.90115303e-01
8.68434489e-01 7.37804640e-03 3.77508640e-01 -3.41942072e-01
5.74002266e-01 4.67597067e-01 6.26770675e-01 -1.07981527e+00
8.68055046e-01 3.72249812e-01 1.38739124e-01 -4.91780370e-01
-9.91931796e-01 -4.58147973e-01 -6.34955704e-01 -1.10791884e-01
5.15783906e-01 -8.08292866e-01 -9.73222792e-01 2.98771739e-01
-4.46307540e-01 -1.36938012e+00 -7.63685584e-01 4.17290449e-01
-1.27371919e+00 -6.46669343e-02 -3.73573214e-01 -1.02119160e+00
2.19714299e-01 -1.14139438e+00 9.75066721e-01 2.91215360e-01
1.85049027e-01 -1.02615511e+00 4.27091986e-01 1.29557386e-01
4.00837600e-01 2.34474525e-01 5.07243574e-01 -7.13788152e-01
-3.19338590e-01 1.10820010e-01 3.50385606e-01 4.51735407e-01
1.06993943e-01 -5.58953226e-01 -8.92821252e-01 -7.10688531e-01
-2.65408963e-01 -1.17289627e+00 5.40176511e-01 9.06935483e-02
1.19269371e+00 -2.65553951e-01 -3.36598992e-01 3.63311410e-01
1.23143888e+00 3.21574301e-01 4.02899265e-01 7.68052101e-01
4.12814587e-01 6.00283146e-01 1.19042301e+00 5.33460140e-01
3.95325094e-01 4.30011004e-01 8.28000486e-01 7.35491738e-02
2.72718728e-01 -8.96223366e-01 6.91308618e-01 1.33229285e-01
1.51913494e-01 -4.42717457e-03 -7.75177121e-01 6.61786616e-01
-2.27838850e+00 -5.54098785e-01 5.94725132e-01 2.14614606e+00
9.71013367e-01 3.88205588e-01 4.79608417e-01 -6.17934130e-02
-7.28039816e-02 -9.06578377e-02 -1.29273808e+00 -3.73699576e-01
2.85949051e-01 4.79012430e-01 1.08364356e+00 4.75124121e-01
-1.15330529e+00 1.12090015e+00 6.54932356e+00 5.51206410e-01
-8.86987329e-01 1.52447656e-01 4.02990639e-01 -3.06138247e-01
2.72836387e-01 1.06372513e-01 -6.51125133e-01 2.09728613e-01
1.39554322e+00 2.23950490e-01 1.07127297e+00 1.23951209e+00
1.87650546e-01 -3.79635543e-01 -1.53292727e+00 6.26275659e-01
-1.82957083e-01 -8.70987713e-01 -6.80093884e-01 1.30594790e-01
8.31673443e-01 4.53503311e-01 1.75205871e-01 1.01618576e+00
1.11336482e+00 -1.09906089e+00 7.37641156e-01 3.17983687e-01
4.81656790e-01 -8.53879869e-01 4.99434024e-01 8.63861859e-01
-6.51441216e-01 -7.62547791e-01 -3.31784904e-01 -3.58700007e-01
-1.15948975e-01 -9.06981751e-02 -9.71711755e-01 2.60518163e-01
8.20142567e-01 6.94276452e-01 -4.46718365e-01 7.01838374e-01
-3.93014461e-01 6.11875474e-01 -4.61264521e-01 -3.89751136e-01
6.98578954e-01 -1.83912829e-01 3.54992896e-01 5.26018739e-01
-6.23499490e-02 -7.36446446e-03 7.48952627e-01 5.28054476e-01
2.13545412e-02 -2.87333220e-01 -1.03958964e+00 -1.08058073e-01
2.89110422e-01 9.57522511e-01 -3.85846317e-01 -3.03263396e-01
-2.55244285e-01 7.33280659e-01 8.46273363e-01 2.68994391e-01
-6.68106258e-01 -1.89429015e-01 8.32616866e-01 -3.76250669e-02
2.02349916e-01 -5.43352425e-01 1.82289183e-01 -8.07671905e-01
8.78565852e-03 -1.17443871e+00 1.34876043e-01 -5.52492023e-01
-1.06703961e+00 1.50415853e-01 2.51998663e-01 -1.16888118e+00
-4.93530601e-01 -7.77380586e-01 -2.42500484e-01 3.08517396e-01
-1.64704299e+00 -8.29586089e-01 -6.61245063e-02 6.29260242e-01
5.24015784e-01 -2.58398414e-01 6.72819793e-01 -9.83387381e-02
-3.79731447e-01 3.87197196e-01 4.13631111e-01 -1.84074789e-01
6.72468781e-01 -1.60684097e+00 4.22031462e-01 4.35230851e-01
1.48125384e-02 4.36467320e-01 8.33707154e-01 -5.99522650e-01
-1.89973533e+00 -9.15592253e-01 -2.75556594e-01 -7.45869339e-01
1.09528387e+00 -6.03390217e-01 -5.64277768e-01 9.73191917e-01
1.18088514e-01 3.70757461e-01 6.25313520e-02 3.26679260e-01
-5.67413643e-02 -2.27116913e-01 -1.21778226e+00 7.56757975e-01
1.34128869e+00 -3.58313918e-01 -6.00101709e-01 4.41985011e-01
9.37296987e-01 -7.51499236e-01 -7.24285305e-01 3.05200070e-01
4.09760982e-01 -8.02876830e-01 8.64621401e-01 -1.26097441e+00
2.85723418e-01 -8.17070007e-02 -1.77397802e-02 -1.89531291e+00
7.81414956e-02 -1.00945580e+00 -5.35810769e-01 6.34148419e-01
2.55621701e-01 -7.12411761e-01 1.01449263e+00 3.73505950e-01
-2.37014472e-01 -8.54813695e-01 -8.98241341e-01 -1.13488281e+00
2.70811409e-01 -3.23888630e-01 5.45278311e-01 4.03866976e-01
2.76524425e-01 1.48974597e-01 -4.91334170e-01 2.75228508e-02
6.13530457e-01 -2.23158330e-01 1.29914856e+00 -9.55591619e-01
-6.26272917e-01 -1.04607530e-02 -1.81247115e-01 -1.33560598e+00
6.56073034e-01 -4.24117446e-01 5.33105135e-01 -1.25149214e+00
-3.06977361e-01 -9.42671299e-01 -5.41083872e-01 7.31660485e-01
-1.00532830e-01 -4.37082738e-01 6.48853928e-02 -1.03344321e-01
-9.33690190e-01 9.24964905e-01 1.46536493e+00 -5.87520748e-02
-4.82689410e-01 -3.50687541e-02 -3.66702646e-01 4.93313044e-01
9.34651971e-01 -3.66332114e-01 -9.60578382e-01 -3.15721542e-01
9.79303122e-02 1.54157817e-01 2.90335178e-01 -1.25719655e+00
1.78794608e-01 -4.65754360e-01 2.07056940e-01 -2.07881674e-01
5.69254100e-01 -9.38149154e-01 -4.67972845e-01 6.84141397e-01
-5.75247049e-01 5.06185852e-02 4.65606362e-01 9.00510550e-01
1.62005082e-01 -1.57351539e-01 7.01690912e-01 -2.59769082e-01
-7.97525644e-01 2.46474087e-01 -3.50489348e-01 3.25098336e-01
1.29378211e+00 6.14042580e-02 -3.67014825e-01 -4.91796732e-01
-5.44661820e-01 7.71045387e-01 5.49587965e-01 5.78746855e-01
3.36439192e-01 -9.95967865e-01 -3.00889432e-01 1.85921788e-01
1.66655958e-01 1.97884291e-01 -6.45010546e-03 6.11861527e-01
-3.58901531e-01 1.38995871e-01 -5.97641110e-01 -4.43708718e-01
-6.47159576e-01 8.77053738e-01 3.81305933e-01 -4.86195058e-01
-7.64021099e-01 6.20597541e-01 -1.13313064e-01 -9.00680125e-01
4.68883961e-01 -3.89441729e-01 1.48146246e-02 -4.11818296e-01
1.38589397e-01 2.73776561e-01 -3.98918837e-02 2.00631544e-02
-4.25824635e-02 1.97034497e-02 -7.58704543e-02 -2.87991554e-01
1.42620957e+00 1.30234137e-01 4.58674252e-01 6.42967820e-01
7.55016863e-01 -5.02200484e-01 -2.09590340e+00 -1.57466531e-01
8.26328397e-02 -2.97338635e-01 -5.09632491e-02 -8.01445901e-01
-6.79131985e-01 6.61361873e-01 6.72154665e-01 1.57877609e-01
8.83121789e-01 -1.61935210e-01 7.89013863e-01 9.81163919e-01
1.00837398e+00 -1.79969144e+00 4.61040229e-01 6.34543836e-01
5.14374375e-01 -1.46781898e+00 -5.70867360e-02 3.43254447e-01
-8.19743812e-01 6.71315193e-01 9.96944368e-01 -6.10852718e-01
5.43677330e-01 3.80878240e-01 2.54803877e-02 5.76542504e-02
-1.15412176e+00 -4.14106101e-01 -2.44326577e-01 1.02703285e+00
-2.16925934e-01 1.76567152e-01 2.20246747e-01 3.22550654e-01
-3.57686311e-01 -3.36534977e-02 5.85732579e-01 1.50498950e+00
-5.09180307e-01 -1.23937130e+00 -1.77445784e-01 4.87565070e-01
-5.71727343e-02 3.94925624e-01 -3.16179276e-01 1.07591534e+00
4.52442020e-02 1.04625845e+00 -1.44883201e-01 -3.60806048e-01
5.41023731e-01 7.36674294e-03 8.12210619e-01 -7.14221418e-01
-5.86745262e-01 -3.97895426e-01 2.72713184e-01 -1.09500790e+00
-2.73129106e-01 -7.47522235e-01 -1.46978498e+00 6.26737177e-02
-7.40894526e-02 1.22284621e-01 7.21634924e-01 1.07573581e+00
2.32859090e-01 6.27375960e-01 7.70592690e-01 -1.09365761e+00
-1.25015652e+00 -9.28922296e-01 -5.61034203e-01 4.34634835e-01
8.33947241e-01 -1.23735750e+00 -2.42130086e-01 -3.37771863e-01] | [4.21461820602417, 1.5904086828231812] |
23cee43a-d586-4c2b-a65b-12885a29fdc0 | stable-training-of-autoencoders-for | 2109.13748 | null | https://arxiv.org/abs/2109.13748v3 | https://arxiv.org/pdf/2109.13748v3.pdf | Improving Autoencoder Training Performance for Hyperspectral Unmixing with Network Reinitialisation | Neural networks, in particular autoencoders, are one of the most promising solutions for unmixing hyperspectral data, i.e. reconstructing the spectra of observed substances (endmembers) and their relative mixing fractions (abundances), which is needed for effective hyperspectral analysis and classification. However, as we show in this paper, the training of autoencoders for unmixing is highly dependent on weights initialisation; some sets of weights lead to degenerate or low-performance solutions, introducing negative bias in the expected performance. In this work, we experimentally investigate autoencoders stability as well as network reinitialisation methods based on coefficients of neurons' dead activations. We demonstrate that the proposed techniques have a positive effect on autoencoder training in terms of reconstruction, abundances and endmembers errors. | ['Krisztián Búza', 'Bartosz Grabowski', 'Michał Cholewa', 'Michał Romaszewski', 'Przemysław Głomb', 'Kamil Książek'] | 2021-09-28 | null | null | null | null | ['hyperspectral-unmixing'] | ['computer-vision'] | [ 3.90390247e-01 -3.94632995e-01 2.86528051e-01 -9.42491218e-02
2.40492254e-01 -5.86625457e-01 5.59891284e-01 7.61258006e-02
-4.98527527e-01 8.98424804e-01 3.05080991e-02 -2.54366606e-01
-2.82898694e-01 -8.85986328e-01 -6.99301124e-01 -1.27933323e+00
1.55941263e-01 3.52028042e-01 -4.40552622e-01 -2.12352797e-01
-1.48756951e-01 8.06925774e-01 -2.02023530e+00 1.61239505e-01
1.11021078e+00 9.37500715e-01 1.10256933e-01 4.79312241e-01
-2.36406907e-01 6.42991841e-01 -5.65619230e-01 -1.07470132e-01
3.51929486e-01 -3.76917839e-01 -2.58688658e-01 3.60675365e-01
2.75274575e-01 -1.12088792e-01 -8.33259597e-02 1.43810010e+00
2.73472160e-01 3.91987890e-01 1.15314066e+00 -9.66830313e-01
-4.16784406e-01 9.32754874e-01 -3.36388171e-01 -8.54541361e-02
-3.84562790e-01 1.19725630e-01 6.62142634e-01 -6.13566101e-01
5.85783981e-02 8.51339042e-01 8.00054371e-01 2.60670155e-01
-1.31157839e+00 -7.08623886e-01 -1.20765731e-01 1.10925101e-01
-1.40838790e+00 -5.13134897e-01 8.17043245e-01 -6.26606107e-01
6.87835872e-01 4.70133960e-01 7.91087866e-01 8.55909050e-01
-2.86533181e-02 2.69071132e-01 1.20067072e+00 -5.95868230e-01
3.76490891e-01 4.82655823e-01 1.37653604e-01 1.72297522e-01
6.96142852e-01 2.19029352e-01 1.07336596e-01 1.61763318e-02
6.35314584e-01 8.30024779e-02 -5.47619164e-01 -3.30897540e-01
-9.68359053e-01 7.19099224e-01 6.84269011e-01 6.45413637e-01
-9.32496965e-01 -2.64769465e-01 -1.08623587e-01 3.54493469e-01
5.51351547e-01 6.93193197e-01 -2.50849277e-01 5.49724340e-01
-8.63539994e-01 -1.47847623e-01 4.65549976e-01 2.09006682e-01
1.10297406e+00 5.10580063e-01 1.56446800e-01 1.05786097e+00
1.42847702e-01 5.69938183e-01 9.12608862e-01 -5.10189772e-01
1.03499457e-01 8.06064308e-01 2.27208927e-01 -9.06337857e-01
-5.21543324e-01 -8.97301257e-01 -1.28206754e+00 5.72033226e-01
2.65608579e-01 -4.62869287e-01 -1.00979674e+00 1.56603420e+00
2.03977108e-01 2.01083720e-01 5.98788738e-01 1.02043462e+00
6.26673698e-01 1.02708745e+00 2.59097219e-01 -4.10884708e-01
8.65188777e-01 -7.06206441e-01 -9.29807007e-01 -3.08636278e-01
1.40282556e-01 -5.99195540e-01 5.64215302e-01 3.98475736e-01
-8.47110450e-01 -6.37549341e-01 -1.24696290e+00 5.23210406e-01
-6.53069854e-01 4.51875329e-01 6.81452692e-01 6.66244030e-01
-7.53837645e-01 1.17907560e+00 -7.47711658e-01 -1.42790928e-01
2.75336634e-02 4.91868317e-01 -3.79927546e-01 3.59683067e-01
-1.09282434e+00 9.08708155e-01 1.11738074e+00 4.54155236e-01
-6.57574952e-01 -5.24182439e-01 -4.90515500e-01 2.99085826e-01
-1.08678930e-01 -4.48680460e-01 6.91288352e-01 -1.63330007e+00
-1.52251494e+00 5.20577252e-01 3.36570412e-01 -5.23013055e-01
2.57507473e-01 1.36922970e-02 -7.47751951e-01 1.11446530e-02
-5.32471538e-01 6.13708317e-01 1.21305549e+00 -1.43635905e+00
-4.32688981e-01 -3.87714237e-01 -3.03043544e-01 2.70050883e-01
-1.00254798e+00 -4.50004727e-01 4.56035435e-01 -5.81873715e-01
3.33563238e-01 -8.06361735e-01 -2.53810823e-01 -3.31621826e-01
-9.83705968e-02 2.49982104e-01 4.90849584e-01 -7.68789411e-01
9.25837874e-01 -2.23308539e+00 3.15765321e-01 5.12578309e-01
3.57758589e-02 5.06564975e-01 -1.23268574e-01 1.49237037e-01
-6.64480627e-01 -1.95778623e-01 -5.46681166e-01 -1.30890071e-01
-1.78432494e-01 2.44396716e-01 -1.93433478e-01 6.34682238e-01
3.35967153e-01 3.89332384e-01 -7.08307266e-01 -4.21736874e-02
6.61300838e-01 8.58163476e-01 -2.68618077e-01 3.02699476e-01
-1.90460876e-01 3.20226133e-01 1.52334496e-01 3.31551105e-01
8.66146386e-01 -1.79335460e-01 3.23139518e-01 -6.36833906e-01
-3.18926513e-01 -1.13339111e-01 -1.32523310e+00 1.07248831e+00
-4.76068974e-01 4.28279132e-01 2.25578442e-01 -1.20442319e+00
8.06644499e-01 4.62021321e-01 5.65444648e-01 -1.04474783e-01
4.53403980e-01 3.60576540e-01 3.42553735e-01 -1.09663583e-01
4.80460554e-01 -5.74654639e-01 8.49870265e-01 3.03236067e-01
3.05236042e-01 7.12904558e-02 2.38067672e-01 -5.28256178e-01
2.41620854e-01 -2.52726734e-01 3.02194446e-01 -4.32645917e-01
7.63293445e-01 -1.70117877e-02 -1.92001630e-02 3.78246456e-01
3.51552844e-01 3.88189554e-01 -3.52544105e-03 -2.21318975e-01
-1.28350770e+00 -7.07004607e-01 -4.02889401e-01 6.53207242e-01
-1.73226610e-01 4.32996899e-01 -8.23212802e-01 8.50058720e-02
-1.72236800e-01 7.55115986e-01 -4.34236556e-01 -3.67286801e-01
-6.37031347e-02 -1.37383938e+00 2.62683511e-01 2.72041500e-01
2.54016429e-01 -1.11837387e+00 -4.26714003e-01 3.52844089e-01
-6.17747530e-02 -8.37426066e-01 2.92083174e-01 7.73880363e-01
-1.26601470e+00 -8.86734486e-01 -1.00835443e+00 -4.61906075e-01
8.29003394e-01 3.07403117e-01 7.46479511e-01 -6.70218840e-02
1.37946429e-02 -1.87559128e-01 -3.17587972e-01 -5.52897811e-01
-8.84551466e-01 -4.43229042e-02 2.47016743e-01 5.45301914e-01
1.89171717e-01 -1.02449417e+00 -4.38712239e-01 3.77965048e-02
-1.37694883e+00 -4.57643010e-02 7.82439291e-01 7.84044206e-01
3.83995742e-01 6.49611712e-01 2.15742454e-01 -8.19292068e-01
7.68378377e-01 -4.23279256e-01 -8.13467383e-01 1.57469630e-01
-8.13765347e-01 2.96195924e-01 9.07095492e-01 -5.76732934e-01
-1.17648888e+00 1.92037821e-01 -2.33401448e-01 -7.26881325e-01
-4.85987008e-01 6.38657331e-01 -2.18551885e-02 -2.61917263e-01
1.17525387e+00 3.84381324e-01 1.98462203e-01 -4.65730041e-01
2.23541811e-01 7.54788995e-01 6.49524212e-01 -1.69973701e-01
8.09484065e-01 4.38537240e-01 -4.71457057e-02 -1.13005447e+00
-4.50844735e-01 -6.23683214e-01 -4.92735893e-01 -2.93626249e-01
6.28426731e-01 -1.08660412e+00 -3.92400801e-01 6.26889467e-01
-9.21239436e-01 -1.41019970e-01 -4.45111573e-01 6.76461458e-01
-9.31786671e-02 3.96678418e-01 -4.44945097e-01 -1.02054656e+00
-3.71672451e-01 -9.93426323e-01 2.91456163e-01 3.24657708e-01
1.88599169e-01 -9.91789877e-01 1.56301502e-02 8.97222906e-02
4.33603138e-01 2.57704675e-01 9.63686109e-01 -4.86896425e-01
9.42766741e-02 -2.20543131e-01 -8.63027647e-02 1.00313115e+00
2.60955244e-01 1.43254369e-01 -1.27895093e+00 -2.02720255e-01
2.23109469e-01 -2.19220314e-02 9.91423011e-01 5.78209579e-01
9.72311020e-01 -4.44325417e-01 1.11515582e-01 6.34326637e-01
1.61353087e+00 2.69075632e-01 7.00419962e-01 1.87479138e-01
5.80949426e-01 7.06588566e-01 -1.11746728e-01 5.39135516e-01
-6.03868544e-01 2.23933198e-02 7.16043472e-01 -2.64546007e-01
4.71780263e-02 6.85802922e-02 2.86128044e-01 8.47181380e-01
-5.61381817e-01 -3.52867156e-01 -6.33697152e-01 3.96240324e-01
-1.67373180e+00 -1.03222084e+00 -4.93302673e-01 2.30787158e+00
7.80864120e-01 -3.05259883e-01 -3.36593315e-02 6.69659913e-01
9.59009349e-01 2.38401085e-01 -4.31802392e-01 -2.59174556e-02
-4.53852087e-01 4.29076910e-01 6.24017894e-01 5.96536636e-01
-1.20434391e+00 6.19789660e-01 6.36231947e+00 3.90130639e-01
-1.48685658e+00 9.70709249e-02 3.88440371e-01 2.04279840e-01
-2.95994908e-01 -2.82522887e-01 -2.15060338e-01 3.37742239e-01
8.89424860e-01 2.96970516e-01 8.69463801e-01 6.80581987e-01
-1.83728449e-02 8.51994157e-02 -4.96894509e-01 1.10577571e+00
3.57945971e-02 -1.13533533e+00 2.74810940e-01 -2.95113027e-02
1.06501734e+00 -3.37042622e-02 -4.54952419e-02 -8.08938406e-03
1.86055109e-01 -9.64400709e-01 6.37426198e-01 4.95088518e-01
4.32577014e-01 -8.33596110e-01 9.38749909e-01 3.09456140e-01
-5.43732047e-01 -3.64792377e-01 -7.11344242e-01 -2.30580255e-01
-8.17060843e-02 1.05262065e+00 -6.21158183e-01 3.96105468e-01
3.31168771e-01 4.50622231e-01 -3.64690930e-01 1.02375174e+00
-2.52783597e-01 5.95457077e-01 -5.03017128e-01 -1.29703477e-01
1.95048735e-01 -8.63731503e-01 4.29705232e-01 1.04062879e+00
4.51578885e-01 1.87569335e-01 -3.44468951e-01 1.08623600e+00
3.86432782e-02 2.22750381e-01 -3.19052458e-01 -5.84083080e-01
1.55609891e-01 1.51204419e+00 -6.88819408e-01 -4.51055467e-01
6.84441254e-02 7.63931274e-01 -5.66555671e-02 5.68209469e-01
-5.59570730e-01 -1.21354423e-01 8.51659596e-01 8.27551037e-02
1.88060418e-01 4.43132222e-02 -1.96908399e-01 -1.30783439e+00
-2.42500082e-01 -9.40854728e-01 2.29287595e-01 -8.66700053e-01
-1.32084990e+00 6.83003247e-01 -2.15687871e-01 -1.16875827e+00
-2.43187636e-01 -1.13850594e+00 -4.47813928e-01 9.22230363e-01
-1.48866773e+00 -7.39953101e-01 -6.53999686e-01 5.72243631e-01
-5.47550619e-02 -2.89737016e-01 8.76325130e-01 4.43622172e-01
-6.90850794e-01 -6.19503378e-04 5.82352340e-01 -1.68939114e-01
3.23688924e-01 -1.17235076e+00 -3.45851243e-01 9.06310916e-01
3.21999937e-01 3.93404573e-01 1.11322296e+00 -3.58176738e-01
-9.83265340e-01 -1.15071881e+00 4.12327170e-01 2.26551563e-01
7.18143880e-01 1.52638674e-01 -9.64427471e-01 4.11502182e-01
1.93847612e-01 -2.08746970e-01 8.96127522e-01 6.83387369e-02
1.73620060e-02 -2.59862244e-01 -9.47477579e-01 5.69648743e-01
5.29684126e-01 -4.00150061e-01 -2.83480942e-01 4.10035282e-01
1.74942091e-01 -7.91202933e-02 -9.10904050e-01 4.95870948e-01
5.26500046e-01 -1.17728961e+00 1.10842943e+00 -6.31582439e-01
5.61177015e-01 -4.62699026e-01 1.53738039e-03 -1.66865516e+00
-6.40725374e-01 3.19043137e-02 -1.25847459e-01 8.82382154e-01
5.60379148e-01 -6.41321957e-01 6.99300110e-01 1.05133012e-01
-1.13841668e-01 -2.82445680e-02 -4.72836971e-01 -5.89930654e-01
-1.23783112e-01 -2.22639635e-01 6.65274441e-01 1.28658390e+00
-3.22043091e-01 2.65885353e-01 -4.55196142e-01 5.30433774e-01
5.33012807e-01 4.05438729e-02 3.80935341e-01 -1.51262844e+00
-4.14927185e-01 -7.21694231e-01 -2.12253526e-01 -3.61568034e-01
3.04435194e-01 -8.34223449e-01 4.78932410e-02 -1.21859455e+00
-1.21450864e-01 -2.77704418e-01 -6.01345599e-01 5.06914318e-01
-1.25661016e-01 3.29258651e-01 -1.59169644e-01 3.08775842e-01
2.91443139e-01 5.86437941e-01 8.06910038e-01 -4.74559456e-01
-4.45037931e-01 3.54722105e-02 -4.93934304e-01 6.63939416e-01
9.99788940e-01 -3.10763985e-01 -3.65791380e-01 -3.72853935e-01
2.27242023e-01 -2.01716363e-01 2.94025302e-01 -1.45785582e+00
9.77415591e-04 -3.60273510e-01 6.60335839e-01 -2.39173219e-01
4.15681183e-01 -1.19294453e+00 8.25514436e-01 5.10324478e-01
-9.94708389e-02 -3.59658808e-01 4.01582599e-01 2.55481660e-01
-4.19719100e-01 -8.24064851e-01 7.92058527e-01 -2.81189442e-01
-5.52730918e-01 1.58287063e-01 -5.64662457e-01 -8.57836068e-01
6.24793351e-01 -3.06516945e-01 3.97121459e-02 -2.68205643e-01
-7.00497985e-01 -2.97698408e-01 3.74869257e-01 8.66730511e-03
2.75387734e-01 -1.09609306e+00 -7.05837429e-01 1.78895071e-01
-3.93639356e-02 -1.68141559e-01 3.89188528e-01 6.72479093e-01
-8.23644638e-01 2.55462289e-01 -6.92836821e-01 -3.30487788e-01
-9.75805283e-01 7.04401612e-01 7.07853556e-01 -4.29814756e-02
-1.22174680e-01 6.55980170e-01 9.67897773e-02 -5.08879602e-01
1.51990339e-01 -3.30817908e-01 -5.07768512e-01 2.86039561e-01
4.24269259e-01 2.68925875e-01 3.19684833e-01 -5.62276006e-01
1.00614838e-01 2.05613717e-01 4.85260278e-01 2.66938433e-02
1.63521695e+00 2.52931118e-01 -4.64499861e-01 4.51710790e-01
8.15893352e-01 -2.30552673e-01 -1.13220644e+00 9.38280374e-02
-4.15916771e-01 -1.21869706e-01 5.12486696e-01 -7.23994434e-01
-1.11043227e+00 1.04463983e+00 8.61363232e-01 4.41058248e-01
1.38204336e+00 -5.44900060e-01 2.29191929e-01 5.63626170e-01
-1.23054117e-01 -9.58988249e-01 -3.38592529e-01 1.47042543e-01
8.39617431e-01 -1.13443410e+00 4.58783060e-02 -2.29640111e-01
-3.03997636e-01 1.32245719e+00 3.79881144e-01 -6.94765300e-02
5.96144497e-01 6.25726432e-02 1.93899229e-01 -4.96244021e-02
-7.80291185e-02 -2.65436113e-01 2.70773292e-01 4.47296590e-01
5.47666252e-01 3.99339050e-01 -2.42246285e-01 3.17697018e-01
-2.33739883e-01 -3.33820492e-01 2.88326919e-01 4.13221598e-01
-6.04725957e-01 -8.70449007e-01 -8.99380624e-01 4.70491260e-01
-1.60406947e-01 -1.69412404e-01 -4.13257927e-01 3.67650598e-01
3.94942671e-01 7.50808418e-01 2.35646069e-01 -4.32359338e-01
1.30429640e-01 3.64128858e-01 3.18149209e-01 -1.94300175e-01
-5.58804214e-01 1.26108542e-01 -1.15789354e-01 2.48567373e-01
-7.74029076e-01 -3.62688631e-01 -7.83406377e-01 -3.02961349e-01
-8.37134957e-01 3.41255844e-01 9.32346582e-01 8.96288753e-01
-1.73028428e-02 6.38286710e-01 5.27712822e-01 -9.63980079e-01
-6.87725008e-01 -1.42945683e+00 -9.79365170e-01 5.62425196e-01
2.91954875e-01 -4.23369914e-01 -6.34011090e-01 1.23977549e-01] | [10.076043128967285, -2.0221869945526123] |
829d391d-4331-4d28-9a0f-3a5d56593a2e | motion-estimation-for-fisheye-video-sequences | 2212.01164 | null | https://arxiv.org/abs/2212.01164v1 | https://arxiv.org/pdf/2212.01164v1.pdf | Motion estimation for fisheye video sequences combining perspective projection with camera calibration information | Fisheye cameras prove a convenient means in surveillance and automotive applications as they provide a very wide field of view for capturing their surroundings. Contrary to typical rectilinear imagery, however, fisheye video sequences follow a different mapping from the world coordinates to the image plane which is not considered in standard video processing techniques. In this paper, we present a motion estimation method for real-world fisheye videos by combining perspective projection with knowledge about the underlying fisheye projection. The latter is obtained by camera calibration since actual lenses rarely follow exact models. Furthermore, we introduce a re-mapping for ultra-wide angles which would otherwise lead to wrong motion compensation results for the fisheye boundary. Both concepts extend an existing hybrid motion estimation method for equisolid fisheye video sequences that decides between traditional and fisheye block matching in a block-based manner. Compared to that method, the proposed calibration and re-mapping extensions yield gains of up to 0.58 dB in luminance PSNR for real-world fisheye video sequences. Overall gains amount to up to 3.32 dB compared to traditional block matching. | ['André Kaup', 'Michel Bätz', 'Andrea Eichenseer'] | 2022-12-02 | null | null | null | null | ['camera-calibration', 'motion-compensation'] | ['computer-vision', 'computer-vision'] | [ 2.31632710e-01 -3.23323160e-01 6.56705424e-02 -3.57616395e-01
-1.66905612e-01 -6.77721560e-01 6.16430402e-01 -6.64703369e-01
-4.43282664e-01 5.27160943e-01 -9.08976570e-02 -4.06260431e-01
1.65851578e-01 -6.63020849e-01 -8.13353121e-01 -7.75639415e-01
1.50674373e-01 -2.47795343e-01 8.40183079e-01 -2.16561258e-01
6.18957877e-01 5.16148627e-01 -1.31327343e+00 1.41970679e-01
2.66421199e-01 7.45293438e-01 4.14149582e-01 1.03836274e+00
3.84849608e-01 6.78460777e-01 -2.74988562e-01 -5.45467854e-01
6.16868436e-01 -1.53165534e-01 -9.76987332e-02 3.53133529e-01
9.20776308e-01 -1.08212340e+00 -5.44747353e-01 1.20410585e+00
-7.29971053e-03 1.04540452e-01 2.87791520e-01 -1.18414748e+00
-1.41311616e-01 -1.45915583e-01 -7.54024088e-01 2.42710844e-01
6.68965757e-01 -5.13407551e-02 4.46163148e-01 -8.71920943e-01
8.52119744e-01 9.64736044e-01 7.51518846e-01 2.74456799e-01
-4.75133508e-01 -5.74614704e-01 -2.93979719e-02 6.55519128e-01
-1.24418759e+00 -7.16541886e-01 7.11063445e-01 -3.00713807e-01
7.19530284e-01 1.97055981e-01 7.33919740e-01 5.90169787e-01
7.86752284e-01 2.68843591e-01 8.47467959e-01 -4.63362068e-01
-1.66059583e-01 2.48804569e-01 -5.01345880e-02 6.16184950e-01
4.35579628e-01 3.69730502e-01 -3.26420397e-01 2.14347333e-01
9.90763426e-01 3.51918191e-01 -8.48625422e-01 -9.85580325e-01
-1.20153248e+00 5.20560205e-01 -4.85468507e-02 -1.29929096e-01
-6.29695132e-02 9.97214839e-02 7.43720531e-02 3.03970814e-01
1.74759045e-01 9.28239673e-02 -3.77724953e-02 -1.84957281e-01
-1.03924966e+00 1.55844793e-01 7.69165039e-01 1.46709061e+00
4.98322010e-01 1.61031440e-01 6.86461687e-01 4.08219904e-01
3.97518754e-01 8.32995832e-01 -1.37640432e-01 -1.30670667e+00
6.39252961e-01 3.02505493e-02 5.82170308e-01 -1.23210335e+00
-2.00052544e-01 -8.47125277e-02 -4.53239888e-01 8.54288340e-01
7.19605684e-01 -9.23654288e-02 -2.99133092e-01 1.12097931e+00
3.07122856e-01 2.97126591e-01 2.21956283e-01 1.09707570e+00
5.26710331e-01 8.74571145e-01 -8.15649450e-01 -3.87047887e-01
1.20449340e+00 -9.79714990e-01 -8.59998584e-01 -1.41159922e-01
3.98277849e-01 -1.05137324e+00 5.47512174e-01 8.92465889e-01
-1.22159111e+00 -4.21519876e-01 -1.35830617e+00 5.72632551e-02
4.88542281e-02 -1.93920564e-02 -1.40681058e-01 1.08369434e+00
-1.15897262e+00 8.15778524e-02 -6.03124559e-01 -3.23653191e-01
-3.22530597e-01 2.42846623e-01 -6.19341731e-01 -3.89129579e-01
-9.36993659e-01 1.17398167e+00 3.35760005e-02 2.20882505e-01
-5.05414367e-01 -5.36714137e-01 -8.51283252e-01 1.01316810e-01
7.85054088e-01 -4.03117269e-01 1.05199444e+00 -7.79574335e-01
-1.76731110e+00 5.12661338e-01 -2.37634704e-01 -4.58382607e-01
5.79211235e-01 -5.79996109e-01 -3.84323597e-01 5.07727921e-01
-3.37713927e-01 5.26850641e-01 9.44455028e-01 -1.14450407e+00
-9.55968440e-01 -8.72994289e-02 4.19798791e-01 4.65960950e-01
2.78162286e-02 -8.83054361e-03 -5.45705914e-01 -3.60110670e-01
6.44719005e-02 -1.01276624e+00 1.11067697e-01 4.74822015e-01
7.89422616e-02 5.86450577e-01 1.40550721e+00 -6.33597314e-01
1.04008973e+00 -1.93487966e+00 -4.64179628e-02 -4.51849289e-02
2.12988511e-01 3.38403910e-01 3.75454247e-01 5.10476887e-01
9.04954001e-02 -6.23001754e-01 1.82520717e-01 1.26083151e-01
-4.52518612e-01 -9.11250338e-02 -1.84019566e-01 9.30450320e-01
-3.88996422e-01 3.16567242e-01 -8.79763603e-01 -2.58339256e-01
7.03540027e-01 4.61562306e-01 -5.65140069e-01 8.96053016e-02
4.50128704e-01 9.66157299e-03 1.64449751e-01 5.92669904e-01
1.18370271e+00 3.46781999e-01 2.34258205e-01 -3.00775111e-01
-5.76777637e-01 -2.31225580e-01 -1.12841845e+00 1.22449672e+00
-5.17998338e-01 1.28040075e+00 3.30941767e-01 -5.76574028e-01
1.02706635e+00 1.94124818e-01 2.51684397e-01 -3.18146795e-01
-2.46062037e-02 -4.24104333e-02 -1.99757963e-01 -4.67522562e-01
8.58192205e-01 -2.22852141e-01 6.68649822e-02 6.45907298e-02
-2.32545793e-01 -2.34906510e-01 1.10890023e-01 -1.23687096e-01
8.60076547e-01 2.70626128e-01 7.50954986e-01 -3.51762682e-01
8.93512547e-01 -8.19949508e-02 6.17049694e-01 4.38578635e-01
-1.32506564e-01 7.15037823e-01 3.06515068e-01 -6.58314466e-01
-1.34801757e+00 -8.78566444e-01 -3.22511196e-02 3.34776938e-01
8.04124296e-01 -3.77697945e-01 -7.85303771e-01 -3.60990733e-01
-5.04880667e-01 5.43205202e-01 -7.28476346e-02 -3.33153568e-02
-8.63749444e-01 -1.54463366e-01 1.92954719e-01 2.82467633e-01
9.16763246e-01 -3.71574879e-01 -1.41669559e+00 -1.42687801e-02
-5.69064878e-02 -1.54587269e+00 -7.62053609e-01 -4.24028605e-01
-8.54087651e-01 -1.39325368e+00 -9.46672261e-01 -4.34869885e-01
4.70942765e-01 1.23991358e+00 9.55403507e-01 -1.72088698e-01
2.46088982e-01 6.30303383e-01 -4.76933479e-01 -7.13561773e-02
-4.15516883e-01 -5.96497715e-01 2.09502224e-02 -3.90887558e-02
4.70905393e-01 -2.17057779e-01 -8.75091434e-01 8.03432107e-01
-7.16437221e-01 1.59098759e-01 5.44153631e-01 7.77115226e-01
-1.15220845e-02 9.05491188e-02 -1.97137564e-01 -6.81897163e-01
-1.25156522e-01 -3.13044310e-01 -1.37090826e+00 1.81051418e-01
-5.50043643e-01 -3.92488807e-01 7.51845837e-01 -1.88386932e-01
-1.20075953e+00 1.96958259e-01 1.25819877e-01 -5.81957579e-01
-1.48521528e-01 -2.11656377e-01 -3.01034480e-01 -4.13945228e-01
6.48468919e-03 2.01847628e-01 8.81508589e-02 -4.80779633e-02
1.92424893e-01 5.39699197e-01 6.42370343e-01 3.36695075e-01
8.05099428e-01 8.03032935e-01 1.33788988e-01 -1.28468168e+00
-3.36070135e-02 -6.69087887e-01 -8.71699572e-01 -6.41524136e-01
9.49976385e-01 -7.96543360e-01 -7.15941250e-01 6.35523379e-01
-1.17621243e+00 7.22741708e-02 4.12973076e-01 8.52428675e-01
-5.80380738e-01 9.11356509e-01 -1.66249380e-01 -6.93596840e-01
7.60130435e-02 -1.39902329e+00 9.62496042e-01 1.25277951e-01
1.06927656e-01 -1.07883084e+00 -1.38554439e-01 2.75319934e-01
1.64054781e-01 7.40209781e-03 2.83949494e-01 2.06684038e-01
-1.03126514e+00 -1.96108475e-01 -1.81958973e-01 2.86492914e-01
-8.95423517e-02 1.23222560e-01 -5.51613629e-01 -3.61458719e-01
4.46198463e-01 4.40051764e-01 5.38316965e-01 6.15880251e-01
2.00902581e-01 -1.20075226e-01 -3.19416195e-01 9.95873630e-01
1.61075771e+00 8.55948091e-01 1.05338871e+00 6.45739973e-01
4.26143348e-01 5.52203298e-01 1.07740712e+00 4.17049289e-01
1.38669521e-01 1.23304164e+00 6.58300102e-01 1.05629839e-01
1.36225626e-01 1.40246019e-01 7.59884238e-01 5.39197028e-01
-1.92845553e-01 -4.66198772e-01 -5.78648090e-01 2.04522118e-01
-1.36855316e+00 -1.34219801e+00 -4.97214824e-01 2.63722372e+00
1.39364332e-01 1.57712609e-01 -5.18843606e-02 2.50986278e-01
7.70173907e-01 2.51167983e-01 -1.00888401e-01 -6.06205285e-01
9.01754424e-02 -5.27196884e-01 1.19464922e+00 9.08316255e-01
-1.01822734e+00 6.43729031e-01 5.94965458e+00 3.93504083e-01
-1.26230526e+00 2.40974128e-02 -9.33974832e-02 -5.64139560e-02
-3.79461050e-02 2.43085638e-01 -1.13780761e+00 5.09865999e-01
8.38776052e-01 1.05167545e-01 4.39627543e-02 7.12871134e-01
3.70204955e-01 -7.55634487e-01 -8.22804630e-01 1.25041974e+00
4.54734206e-01 -1.12841082e+00 -2.55731732e-01 3.28636587e-01
5.48480511e-01 -5.80433726e-01 -1.97852198e-02 -4.37730819e-01
-2.28262544e-01 -4.33365852e-01 7.32342958e-01 5.57916343e-01
6.93233490e-01 -8.02296460e-01 8.56627524e-01 3.90496999e-01
-1.11524737e+00 -7.44081885e-02 -7.60886252e-01 -1.45591926e-02
6.97698593e-01 2.57079870e-01 -8.69455695e-01 4.73712206e-01
7.26092279e-01 7.38936186e-01 -2.48763651e-01 1.08746362e+00
1.27281815e-01 2.21994683e-01 -9.90121141e-02 2.36661479e-01
4.33984160e-01 -5.98636389e-01 9.17980611e-01 1.11781454e+00
8.65404963e-01 1.90402448e-01 -5.36431015e-01 2.74969786e-01
5.24754047e-01 -3.03269923e-01 -1.15991354e+00 7.34012961e-01
4.08019751e-01 8.39033842e-01 -6.49220169e-01 -2.51040727e-01
-1.01455581e+00 1.07554781e+00 -5.84009767e-01 2.92469680e-01
-1.05175149e+00 -5.62745035e-01 4.87700045e-01 5.42079270e-01
6.42755926e-01 -5.58787942e-01 2.29348019e-01 -1.19722629e+00
-1.07561015e-01 -6.18303120e-01 -2.08811298e-01 -1.08001840e+00
-3.26970816e-01 5.80123007e-01 5.01998246e-01 -2.07742786e+00
-6.22985601e-01 -8.48942995e-01 -3.15982789e-01 5.46476662e-01
-1.53216445e+00 -8.85628223e-01 -6.29537940e-01 5.31027913e-01
1.04258859e+00 -2.67603725e-01 2.67861694e-01 2.65987545e-01
-7.90789947e-02 2.51949072e-01 5.89161634e-01 -2.51811832e-01
9.87151980e-01 -9.26076949e-01 2.13299036e-01 1.26596153e+00
-1.48265928e-01 5.59569657e-01 9.97461438e-01 -3.81209761e-01
-1.58023977e+00 -6.20767534e-01 5.98312736e-01 -4.60313350e-01
3.01945090e-01 -1.60935700e-01 -6.63779318e-01 6.06047869e-01
5.89372635e-01 -4.46577780e-02 2.68908381e-01 -8.14390719e-01
-1.53558748e-02 -3.68025333e-01 -9.15276289e-01 4.60732371e-01
8.86152029e-01 -1.51309565e-01 -3.67718905e-01 -3.07320267e-01
3.21745276e-02 -5.38378179e-01 -5.90994060e-01 1.70420349e-01
1.13118196e+00 -1.67831409e+00 1.14122272e+00 9.19151902e-02
3.60482544e-01 -5.55835485e-01 -3.21309417e-01 -1.04167879e+00
-6.62561208e-02 -7.98335195e-01 2.09628761e-01 8.03100467e-01
-1.57360241e-01 -5.06621838e-01 7.84760475e-01 2.58357555e-01
-3.14425111e-01 -3.43805820e-01 -8.47029567e-01 -7.08045423e-01
-4.13240522e-01 -4.43274602e-02 5.53753302e-02 6.04021490e-01
-5.78624289e-03 -4.56888368e-03 -1.15204465e+00 3.97296369e-01
9.35073316e-01 2.16202121e-02 1.09087086e+00 -8.81717503e-01
-2.11815506e-01 -5.27504086e-02 -9.05542791e-01 -1.67155504e+00
-1.08402558e-01 -1.96566917e-02 8.12749937e-02 -1.00267339e+00
6.56388141e-03 -1.17129110e-01 3.35196257e-01 -6.37483597e-01
8.36060718e-02 4.97935712e-01 3.06326389e-01 1.40445054e-01
-2.85952419e-01 5.48152104e-02 9.90285814e-01 2.70883918e-01
-3.49133536e-02 3.87865096e-01 -1.02698080e-01 9.81352925e-01
5.95386386e-01 -1.23590499e-01 -8.65096569e-01 -6.43230677e-01
1.58463359e-01 7.01610684e-01 4.31537092e-01 -1.10555637e+00
5.62855124e-01 -9.19718146e-02 1.60324544e-01 -9.76550817e-01
7.42345929e-01 -1.40651417e+00 4.37240928e-01 3.10026437e-01
1.85450032e-01 5.12135983e-01 -7.88869187e-02 6.31976902e-01
-3.10810536e-01 -6.50228143e-01 9.19885099e-01 -1.34486750e-01
-1.06617105e+00 -2.38485172e-01 -8.35783005e-01 -4.55637485e-01
1.41509199e+00 -1.09561288e+00 -4.10887092e-01 -9.16961432e-01
-3.06071132e-01 -2.06439361e-01 1.03401721e+00 1.34409234e-01
1.14641666e+00 -9.56477642e-01 -3.32923770e-01 3.86322170e-01
-2.83112526e-01 -3.21353585e-01 3.10402989e-01 1.12628496e+00
-1.24631631e+00 9.20350432e-01 -6.63923144e-01 -7.55167902e-01
-1.95891893e+00 6.28580451e-01 3.20867807e-01 2.14440838e-01
-8.76192689e-01 3.92782867e-01 5.42213142e-01 2.24419788e-01
-1.03013985e-01 -3.05093467e-01 -2.59652168e-01 -3.71648252e-01
7.23213077e-01 7.76636958e-01 -6.98656440e-02 -9.43248034e-01
-3.17023873e-01 1.13086605e+00 9.10983086e-02 -2.33595654e-01
1.06439197e+00 -7.64165103e-01 2.94326156e-01 1.56539828e-01
1.01198339e+00 4.93354976e-01 -1.62984633e+00 1.49199933e-01
-1.98933616e-01 -1.17454135e+00 8.35392773e-02 -4.69499715e-02
-9.53839004e-01 1.12223220e+00 6.19237900e-01 1.37506612e-02
1.27405906e+00 -4.27730680e-01 7.92847216e-01 6.56068444e-01
4.28556532e-01 -7.18349934e-01 -3.81802827e-01 2.62664735e-01
6.77303076e-01 -1.01601303e+00 3.41942608e-01 -7.84440160e-01
-6.55440748e-01 1.75649714e+00 4.57656115e-01 -2.64326811e-01
4.59693104e-01 4.68601763e-01 1.52989596e-01 1.86936140e-01
-7.46425450e-01 2.47367755e-01 1.62943397e-02 8.35045218e-01
1.48066401e-01 -5.96976757e-01 -2.20732719e-01 -1.86867967e-01
8.73534828e-02 -2.30822433e-02 1.33552372e+00 9.42809761e-01
-7.58788109e-01 -9.03440475e-01 -9.17272449e-01 -1.25056505e-01
-3.53098541e-01 3.27451453e-02 4.10596766e-02 1.27904201e+00
-2.07278371e-01 9.49948251e-01 2.31052697e-01 -3.92221749e-01
3.64690453e-01 -3.26432019e-01 5.87666810e-01 -1.00275986e-01
5.33620752e-02 3.39111269e-01 2.09374651e-01 -7.53517210e-01
-7.08099484e-01 -6.66360080e-01 -7.03043759e-01 -4.21891779e-01
-4.27443057e-01 4.20811698e-02 5.17547965e-01 5.85638106e-01
-1.41184777e-01 -3.99011467e-03 7.12570667e-01 -1.11100149e+00
-3.38479996e-01 -4.82058376e-01 -5.45582652e-01 -9.86117497e-02
5.56555748e-01 -8.37742150e-01 -3.30225557e-01 4.81664240e-01] | [8.865074157714844, -2.4264702796936035] |
561449cf-c89d-44ca-a64c-b7297f3b35e6 | improving-breast-cancer-detection-using | 1808.08273 | null | http://arxiv.org/abs/1808.08273v1 | http://arxiv.org/pdf/1808.08273v1.pdf | Improving Breast Cancer Detection using Symmetry Information with Deep Learning | Convolutional Neural Networks (CNN) have had a huge success in many areas of
computer vision and medical image analysis. However, there is still an immense
potential for performance improvement in mammogram breast cancer detection
Computer-Aided Detection (CAD) systems by integrating all the information that
the radiologist utilizes, such as symmetry and temporal data. In this work, we
proposed a patch based multi-input CNN that learns symmetrical difference to
detect breast masses. The network was trained on a large-scale dataset of 28294
mammogram images. The performance was compared to a baseline architecture
without symmetry context using Area Under the ROC Curve (AUC) and Competition
Performance Metric (CPM). At candidate level, AUC value of 0.933 with 95%
confidence interval of [0.920, 0.954] was obtained when symmetry information is
incorporated in comparison with baseline architecture which yielded AUC value
of 0.929 with [0.919, 0.947] confidence interval. By incorporating symmetrical
information, although there was no a significant candidate level performance
again (p = 0.111), we have found a compelling result at exam level with CPM
value of 0.733 (p = 0.001). We believe that including temporal data, and adding
benign class to the dataset could improve the detection performance. | ['Yeman Brhane Hagos', 'Albert Gubern Merida', 'Jonas Teuwen'] | 2018-08-17 | null | null | null | null | ['breast-cancer-detection', 'breast-cancer-detection'] | ['knowledge-base', 'medical'] | [ 2.39804953e-01 2.77867168e-01 -2.97818840e-01 -4.00594324e-01
-8.91576171e-01 -1.43417254e-01 3.18618298e-01 6.58459663e-01
-6.78295255e-01 4.38233137e-01 1.37487769e-01 -8.29775095e-01
-2.50447363e-01 -8.32580447e-01 -7.31606901e-01 -6.31547630e-01
-3.11533839e-01 3.04622091e-02 4.22857136e-01 -8.87245163e-02
-7.12438226e-02 4.62133914e-01 -1.02640605e+00 6.16778791e-01
4.39853489e-01 1.15337372e+00 -4.84103151e-02 8.31911802e-01
3.85422409e-01 6.26428783e-01 -4.99339134e-01 -3.98031324e-01
4.17214245e-01 -4.27838713e-01 -5.64793587e-01 -1.69919685e-01
5.48164845e-01 -3.91243130e-01 -2.02866286e-01 8.44003201e-01
6.23818278e-01 -4.07703668e-01 5.65121114e-01 -5.66900611e-01
-4.01280075e-01 4.81704593e-01 -8.84494483e-01 5.82855821e-01
-3.33411366e-01 5.26739955e-02 5.18217742e-01 -5.05463839e-01
5.26523054e-01 6.00589216e-01 1.08829749e+00 2.59931654e-01
-8.46816897e-01 -6.51755869e-01 -4.01745409e-01 6.00168034e-02
-1.17191076e+00 1.01827294e-01 2.66136289e-01 -2.16852412e-01
8.07436883e-01 4.71153080e-01 8.80425036e-01 5.85319579e-01
8.75928164e-01 4.42267716e-01 8.80567431e-01 -6.21935606e-01
-1.47172287e-01 2.72796929e-01 2.00129580e-02 8.53682160e-01
4.18911427e-01 2.96498448e-01 5.61711006e-02 -2.05607682e-01
8.92724335e-01 1.17638126e-01 2.41941884e-01 -1.29968107e-01
-1.27892971e+00 9.23786879e-01 1.00333881e+00 6.63603127e-01
-3.15578699e-01 -4.61150473e-03 7.12324440e-01 2.21434936e-01
9.57921147e-02 2.94734180e-01 -1.05206400e-01 3.72292370e-01
-1.13087273e+00 1.45396024e-01 3.76670867e-01 3.44813079e-01
5.01304120e-02 -1.64373249e-01 -2.96490937e-01 7.44627237e-01
7.76778460e-02 2.85609215e-01 7.98828244e-01 -5.18355966e-01
-4.40863613e-03 6.88137472e-01 -1.64613530e-01 -1.13066590e+00
-9.34514403e-01 -8.20465446e-01 -1.22927678e+00 3.10311783e-02
5.14358282e-01 -1.35194406e-01 -1.27184069e+00 1.48818421e+00
8.72727111e-02 -1.24993272e-01 5.19506894e-02 8.86575401e-01
9.86024082e-01 1.44614309e-01 1.17126398e-01 6.24049716e-02
1.75273168e+00 -5.94338894e-01 -4.58254039e-01 -1.12447836e-01
9.77174461e-01 -6.48821235e-01 6.01719260e-01 1.22472040e-01
-7.86886513e-01 -4.57599342e-01 -1.51901996e+00 1.41419038e-01
-8.68939534e-02 4.36187148e-01 6.02518141e-01 8.96312833e-01
-1.03355753e+00 2.77947217e-01 -1.14443290e+00 -6.50330484e-01
7.12966383e-01 4.83926266e-01 -5.00776589e-01 -2.60588199e-01
-9.21333134e-01 9.12642598e-01 3.28233570e-01 -1.71445943e-02
-8.06058168e-01 -6.72222257e-01 -6.29693449e-01 -1.66420192e-01
1.34636462e-01 -5.31259179e-01 1.19580650e+00 -9.95541692e-01
-6.19022310e-01 1.01501620e+00 7.34950379e-02 -9.63529468e-01
5.95836580e-01 2.39609390e-01 -4.99402463e-01 2.03064531e-01
2.79869467e-01 1.01847625e+00 1.66279435e-01 -7.09807336e-01
-8.99628758e-01 -4.88649726e-01 -1.25264004e-01 -6.50330856e-02
-9.60593224e-02 9.04284418e-02 -4.01277870e-01 -5.39721549e-01
4.21815932e-01 -1.07749093e+00 -4.02716368e-01 2.02560887e-01
-9.34746191e-02 3.05864327e-02 5.47448933e-01 -9.40949500e-01
9.19850528e-01 -2.06778860e+00 -7.64208555e-01 2.98773140e-01
2.72033155e-01 4.57104266e-01 2.25298643e-01 -1.37226507e-01
-1.86694652e-01 2.51664698e-01 -2.54732490e-01 3.91974211e-01
-6.64182305e-01 -2.06179246e-01 6.10992193e-01 6.08668625e-01
5.37389278e-01 1.18356121e+00 -6.27089500e-01 -4.46048141e-01
2.10163325e-01 4.66760039e-01 -5.34213960e-01 -3.89969528e-01
3.10776561e-01 1.84685037e-01 -3.05948853e-01 7.92259812e-01
5.95136464e-01 -5.28425634e-01 1.42300799e-01 -4.70016658e-01
8.44693929e-02 -2.08936021e-01 -8.67632627e-01 1.51163745e+00
-1.02484591e-01 9.61315215e-01 -1.17852546e-01 -1.09547389e+00
1.02027035e+00 4.02298599e-01 6.65799260e-01 -9.49696422e-01
2.38148108e-01 2.85506725e-01 9.76869166e-01 -3.12963456e-01
2.49405026e-01 -2.59070098e-01 1.97770804e-01 3.64331342e-02
-7.85500407e-02 2.29038209e-01 -1.36290696e-02 1.55634731e-01
1.62077868e+00 -3.85370940e-01 4.76367414e-01 -5.28959632e-01
3.11721891e-01 4.38032150e-01 2.49398500e-01 8.29665661e-01
-4.74210501e-01 7.21293151e-01 5.43476939e-01 -6.13414168e-01
-9.90306854e-01 -8.09712231e-01 -7.89856076e-01 2.64572352e-01
-8.05300474e-02 -1.55234337e-01 -3.32127869e-01 -5.58272958e-01
-3.20870802e-02 1.40647754e-01 -1.02814484e+00 -1.73195764e-01
-6.74715698e-01 -1.06074309e+00 6.96283221e-01 8.60421896e-01
7.64809132e-01 -8.36518347e-01 -1.11505675e+00 2.24533632e-01
8.71536732e-02 -9.29363847e-01 -4.57726000e-03 4.55026120e-01
-1.06594646e+00 -1.01302028e+00 -1.05107439e+00 -7.16905355e-01
6.33350968e-01 2.24850386e-01 8.88015628e-01 1.50083870e-01
-8.48436058e-01 2.56356280e-02 -3.48779202e-01 -6.90674543e-01
-4.25758332e-01 -1.53222010e-01 -4.82324779e-01 -4.54803348e-01
5.23727953e-01 2.46062446e-02 -8.76534939e-01 3.09516579e-01
-9.73466337e-01 1.38844857e-02 1.19850290e+00 1.03221810e+00
6.12608850e-01 -1.27800658e-01 6.24486268e-01 -6.71083570e-01
2.50826716e-01 -5.49985290e-01 -1.87184051e-01 -2.09162850e-03
-4.67567831e-01 -4.27726567e-01 8.41739699e-02 -4.65278655e-01
-6.20435476e-01 -5.80978394e-02 -7.22549036e-02 -9.63675156e-02
-1.46786645e-01 7.11279273e-01 5.27997375e-01 -6.33729547e-02
1.00370085e+00 -2.02692568e-01 1.82546869e-01 -1.74281187e-02
-3.94167572e-01 4.03916210e-01 6.81770265e-01 -9.38952435e-03
2.02215731e-01 4.66313928e-01 2.94720292e-01 -5.24359345e-01
-7.06424713e-01 -6.07147336e-01 -4.31104243e-01 7.57863820e-02
9.20184374e-01 -8.69153261e-01 -4.29598361e-01 1.80911526e-01
-6.70503914e-01 -6.62782937e-02 -2.05265954e-01 8.88407946e-01
-8.44587609e-02 8.37795809e-02 -5.60327828e-01 -5.68889260e-01
-4.83631670e-01 -1.14002991e+00 6.99189723e-01 3.40462893e-01
-3.75167727e-01 -7.39226818e-01 -2.87640482e-01 3.09888214e-01
6.93742633e-01 8.67954612e-01 8.47234845e-01 -9.88635480e-01
-1.62763387e-01 -9.28384542e-01 -7.23213792e-01 2.96736062e-01
3.78728032e-01 -9.65525284e-02 -7.22105920e-01 -3.05522949e-01
8.01435858e-03 -5.52071556e-02 9.46451068e-01 1.00400436e+00
1.18559575e+00 1.56021744e-01 -6.24739945e-01 4.17091191e-01
1.51164556e+00 7.97677398e-01 6.89620614e-01 6.11285090e-01
2.89229244e-01 3.07823002e-01 4.64876801e-01 1.35038659e-01
1.36746466e-01 4.35947984e-01 5.37956655e-01 -4.93227363e-01
-2.64507025e-01 8.14684406e-02 -2.11984649e-01 -1.04993857e-01
3.32907960e-02 -1.07016459e-01 -1.31349015e+00 7.86628127e-01
-1.46293604e+00 -6.81184471e-01 -1.49360597e-01 2.05397105e+00
4.65457052e-01 5.44560194e-01 3.58645208e-02 3.28118503e-01
7.36539125e-01 -3.48802507e-01 -5.31295776e-01 -5.12612820e-01
-1.40896603e-01 3.21471512e-01 7.48635411e-01 -7.72645548e-02
-1.20807123e+00 1.70381591e-01 5.45718431e+00 5.66279471e-01
-1.41800833e+00 -8.84632487e-03 1.14642775e+00 -3.51690985e-02
4.50056702e-01 -2.04932421e-01 -5.74817598e-01 2.44205356e-01
1.13651621e+00 -1.22191086e-01 -6.21643603e-01 5.30820012e-01
-9.55684334e-02 -7.07472622e-01 -9.87875938e-01 6.73168302e-01
2.10464284e-01 -1.42567205e+00 -2.19901651e-01 2.69108027e-01
6.97141528e-01 1.30173236e-01 1.79503679e-01 2.93616295e-01
-2.41926014e-01 -1.38166869e+00 1.07236266e-01 2.38916010e-01
7.89845884e-01 -5.75977683e-01 1.46618998e+00 2.82243878e-01
-7.60397911e-01 1.25583619e-01 -1.29307002e-01 8.82027298e-02
-3.01291913e-01 5.64560115e-01 -1.48712969e+00 4.92425025e-01
1.06356835e+00 3.62977356e-01 -8.13008368e-01 1.39772058e+00
3.91252279e-01 6.99611843e-01 -4.05161530e-01 -1.12414353e-01
4.84807611e-01 4.66440558e-01 1.79195911e-01 1.30239940e+00
4.86147791e-01 2.15160862e-01 -1.33750021e-01 4.04845864e-01
2.69988775e-02 3.33746165e-01 -4.74754393e-01 1.92047834e-01
6.82362467e-02 1.29172778e+00 -1.10579896e+00 -2.41768479e-01
-5.89837372e-01 3.42342019e-01 -3.85906041e-01 -3.08332324e-01
-7.69477904e-01 -3.68793756e-01 -3.65843117e-01 5.02536237e-01
1.39449224e-01 3.44397247e-01 -4.81260687e-01 -4.23782110e-01
7.50465840e-02 -7.63242960e-01 6.43240035e-01 -4.88958448e-01
-8.92551839e-01 4.54419464e-01 3.58108319e-02 -1.13456488e+00
-5.44678271e-02 -7.13599741e-01 -7.37128675e-01 8.62459540e-01
-1.07016504e+00 -1.16534889e+00 -3.83473039e-01 2.15523571e-01
2.81423122e-01 -1.47557437e-01 7.92612612e-01 2.17666909e-01
-1.55398577e-01 1.03220022e+00 -2.84762271e-02 5.61796308e-01
8.09317112e-01 -9.76003826e-01 -1.23415403e-01 5.53404808e-01
-2.48859748e-01 5.13261378e-01 2.69972354e-01 -6.51539087e-01
-1.02059531e+00 -1.14244044e+00 5.63761592e-01 -1.27880007e-01
3.21546972e-01 4.01201487e-01 -8.68291080e-01 5.76619565e-01
1.59335993e-02 3.25416118e-01 9.12535131e-01 -1.29440218e-01
-1.08995624e-01 -4.49198782e-02 -1.58269179e+00 2.55501896e-01
3.30720246e-01 2.29223534e-01 -2.61325419e-01 1.69414014e-01
3.58875126e-01 -6.39662206e-01 -1.17377114e+00 1.27769268e+00
8.75723839e-01 -9.72418547e-01 7.75234222e-01 -2.31986806e-01
5.76632082e-01 -3.00532933e-02 -2.69386381e-01 -7.92269349e-01
-4.04426515e-01 2.45124534e-01 2.82247245e-01 5.55790186e-01
8.17597568e-01 -6.26665592e-01 1.00828898e+00 3.75844479e-01
-3.75073701e-01 -1.25392640e+00 -9.96668637e-01 -5.16738713e-01
4.81108934e-01 -2.95477748e-01 2.16126308e-01 7.86535442e-01
-1.20430104e-01 -5.68766929e-02 9.65151563e-02 5.70852607e-02
1.79584071e-01 -1.67940989e-01 3.36686045e-01 -9.82461333e-01
-1.68805867e-01 -4.91898596e-01 -8.79039764e-01 -2.92732835e-01
-6.64451599e-01 -1.04540741e+00 -4.27680701e-01 -1.70445597e+00
7.52840579e-01 -4.11874175e-01 -7.95423627e-01 5.41971266e-01
-5.96944876e-02 4.62786674e-01 -6.80542588e-02 1.18039742e-01
-1.93258300e-01 -2.47407362e-01 1.37754536e+00 -2.79409409e-01
-5.47409616e-02 9.75914747e-02 -9.10237670e-01 5.57941437e-01
1.20160949e+00 -4.45406884e-01 -6.03634790e-02 -3.55233520e-01
-1.27728358e-01 4.69898403e-01 5.93898058e-01 -1.37401986e+00
1.86516687e-01 2.29980186e-01 1.18273628e+00 -9.00216043e-01
1.87304184e-01 -4.86862987e-01 1.55072957e-01 1.21807897e+00
-2.63708293e-01 3.40033583e-02 7.51432657e-01 4.20756549e-01
-2.52219439e-01 -1.17522560e-01 9.05543268e-01 -1.95706114e-01
-4.20202315e-01 1.03465214e-01 -2.18299836e-01 -4.13456649e-01
1.16335464e+00 -5.73500574e-01 -1.89146385e-01 -1.32226180e-02
-7.42537022e-01 -2.86885165e-02 1.30051970e-01 4.41080719e-01
4.36081260e-01 -1.28966272e+00 -9.81341124e-01 1.81921676e-01
2.80514598e-01 -5.61882630e-02 1.64617211e-01 1.11004555e+00
-6.13889217e-01 7.35990584e-01 -3.42155337e-01 -1.01009989e+00
-1.42074442e+00 6.18151128e-02 5.93166292e-01 -4.94999349e-01
-5.88757932e-01 8.81105542e-01 7.24402368e-02 -1.37747273e-01
9.67570990e-02 -6.00116432e-01 -5.99853881e-02 -2.15897903e-01
4.53048885e-01 1.93543464e-01 5.56314349e-01 -3.50962549e-01
-5.35745800e-01 3.74533296e-01 -5.16909182e-01 -6.64236397e-02
1.18219101e+00 5.11169195e-01 2.40250140e-01 9.67760980e-02
1.21633351e+00 -2.05778077e-01 -6.86452210e-01 -1.92097932e-01
-3.06633282e-02 -2.08547294e-01 1.90349877e-01 -1.25684798e+00
-1.02691495e+00 7.19882607e-01 1.44455314e+00 4.47065160e-02
1.06317902e+00 1.02164030e-01 4.00876850e-01 3.09534550e-01
7.64814168e-02 -5.70336759e-01 4.97679487e-02 2.62612045e-01
7.26826966e-01 -1.65957797e+00 1.79689214e-01 9.77135915e-03
-6.67581677e-01 1.11693680e+00 8.21114898e-01 -3.24160933e-01
9.27707076e-01 2.75754750e-01 1.23880513e-01 -2.72062451e-01
-5.62172651e-01 1.90661550e-01 3.98183644e-01 4.65489626e-01
9.42018569e-01 2.92101413e-01 -7.12042391e-01 6.99592352e-01
-1.95563570e-01 8.70268494e-02 4.01479095e-01 1.17323411e+00
-5.08938551e-01 -6.92236245e-01 -3.87584716e-01 1.11184156e+00
-1.05600381e+00 7.19058588e-02 -1.23475760e-01 1.20709300e+00
3.38027894e-01 8.29304457e-01 2.31919020e-01 -1.97479472e-01
2.26377994e-01 -1.82445735e-01 3.12599629e-01 -6.13922358e-01
-7.40381598e-01 2.31510863e-01 -7.45635927e-02 -2.05464914e-01
-3.50338161e-01 -6.36572838e-01 -1.31763279e+00 4.52280918e-04
-6.46474898e-01 5.29853813e-02 8.12126756e-01 5.71601570e-01
9.59953144e-02 9.32246804e-01 1.89107612e-01 -3.75657707e-01
-3.11520696e-01 -1.18088865e+00 -2.97800541e-01 6.56619444e-02
3.69627595e-01 -3.98320615e-01 2.12011635e-02 -1.50328666e-01] | [15.222908020019531, -2.5156972408294678] |
59f7853e-513e-433e-a0d2-6064ab21a956 | 190600389 | 1906.00389 | null | https://arxiv.org/abs/1906.00389v4 | https://arxiv.org/pdf/1906.00389v4.pdf | Disparate Vulnerability to Membership Inference Attacks | A membership inference attack (MIA) against a machine-learning model enables an attacker to determine whether a given data record was part of the model's training data or not. In this paper, we provide an in-depth study of the phenomenon of disparate vulnerability against MIAs: unequal success rate of MIAs against different population subgroups. We first establish necessary and sufficient conditions for MIAs to be prevented, both on average and for population subgroups, using a notion of distributional generalization. Second, we derive connections of disparate vulnerability to algorithmic fairness and to differential privacy. We show that fairness can only prevent disparate vulnerability against limited classes of adversaries. Differential privacy bounds disparate vulnerability but can significantly reduce the accuracy of the model. We show that estimating disparate vulnerability to MIAs by na\"ively applying existing attacks can lead to overestimation. We then establish which attacks are suitable for estimating disparate vulnerability, and provide a statistical framework for doing so reliably. We conduct experiments on synthetic and real-world data finding statistically significant evidence of disparate vulnerability in realistic settings. The code is available at https://github.com/spring-epfl/disparate-vulnerability | ['Mohammad Yaghini', 'Bogdan Kulynych', 'Carmela Troncoso', 'Michael Veale', 'Giovanni Cherubin'] | 2019-06-02 | null | null | null | null | ['membership-inference-attack'] | ['computer-vision'] | [-5.21663837e-02 -8.89799185e-03 -2.65440285e-01 -3.84230435e-01
-1.03044724e+00 -1.19669068e+00 3.86722118e-01 2.88077325e-01
-3.33446145e-01 8.01249623e-01 -1.24080509e-01 -9.00883555e-01
-2.08076119e-01 -1.09910536e+00 -7.37159193e-01 -5.06136954e-01
-3.77897948e-01 2.77225703e-01 -1.61199212e-01 1.80995136e-01
3.36276174e-01 6.77004814e-01 -9.33805346e-01 2.62785703e-01
8.26723754e-01 4.35413569e-01 -9.72139716e-01 7.61076629e-01
3.37483495e-01 5.03269494e-01 -8.94849241e-01 -9.01995361e-01
8.93591106e-01 -3.96209210e-01 -6.57138109e-01 -7.64012814e-01
7.90371895e-01 -4.87943411e-01 -3.92005026e-01 1.35349536e+00
5.07146180e-01 -2.57390320e-01 1.00412881e+00 -1.75658977e+00
-7.23474801e-01 9.09239113e-01 -4.55573410e-01 4.60044861e-01
3.72557551e-01 2.04055771e-01 8.96885633e-01 -2.24446088e-01
1.46897122e-01 1.11865723e+00 9.23600733e-01 7.91722059e-01
-1.28097618e+00 -1.22568977e+00 -1.64922580e-01 -3.62053573e-01
-1.70719874e+00 -6.38602674e-01 3.48924011e-01 -4.11265880e-01
3.61058712e-01 9.26337659e-01 3.65289263e-02 9.27586913e-01
1.93451539e-01 2.51366764e-01 1.06782413e+00 -1.84931353e-01
5.19027293e-01 1.99837893e-01 4.80881453e-01 5.22210658e-01
1.00059509e+00 3.90496075e-01 -6.44758865e-02 -1.35133088e+00
3.72552663e-01 -7.14119673e-02 -1.41835824e-01 -1.28513202e-01
-4.77871656e-01 1.02029181e+00 9.46890488e-02 -7.20825093e-03
1.12973794e-01 2.39591166e-01 3.48594159e-01 5.68555415e-01
2.65327126e-01 3.48349929e-01 -3.80781412e-01 2.78101712e-01
-6.69508100e-01 6.90117776e-01 1.18714237e+00 7.66717672e-01
2.39812210e-01 -8.45862627e-02 -7.50528947e-02 2.11313859e-01
1.30670369e-01 8.55323672e-01 1.64885014e-01 -1.14870906e+00
5.63579261e-01 3.24001536e-02 3.23700935e-01 -1.14692211e+00
-2.03313529e-02 -4.75975946e-02 -7.56860018e-01 3.81622761e-01
9.44504023e-01 -4.76995468e-01 -4.90099758e-01 2.31851435e+00
2.71481216e-01 2.60237455e-01 1.39540672e-01 3.99439484e-01
2.36632138e-01 3.80132318e-01 4.20835912e-01 -1.22075565e-01
1.02230847e+00 2.01315224e-01 -2.43342251e-01 1.89856723e-01
7.30615437e-01 -3.11909556e-01 7.61255145e-01 -8.75330046e-02
-1.07857478e+00 8.56148601e-02 -8.07656109e-01 3.01178992e-01
-3.70466650e-01 -6.96859300e-01 5.36211491e-01 1.52706504e+00
-8.67012918e-01 4.20762658e-01 -5.10525346e-01 -9.70349237e-02
9.22079027e-01 5.20471990e-01 -4.20069069e-01 2.54598141e-01
-1.47771454e+00 6.05915368e-01 2.34858375e-02 -4.16872501e-01
-8.93275738e-01 -1.00079286e+00 -7.88132966e-01 2.06108868e-01
1.45447090e-01 -7.74352849e-01 8.71863782e-01 -7.31990635e-01
-5.12305140e-01 8.51856351e-01 -3.26700732e-02 -8.08702886e-01
7.92901397e-01 5.98911718e-02 -4.08808500e-01 3.11418101e-02
1.19587995e-01 -7.09428191e-02 5.85529208e-01 -1.39578247e+00
-4.31807727e-01 -6.25436962e-01 2.54699498e-01 -3.15503925e-01
-4.23974544e-01 3.33956897e-01 4.94937509e-01 -7.04417229e-01
-4.90492940e-01 -8.17686975e-01 -5.11367857e-01 1.93990767e-02
-7.78800547e-01 1.97405547e-01 6.32088125e-01 -5.81156611e-01
1.48186862e+00 -1.97080481e+00 -6.59065902e-01 8.96652758e-01
4.49335009e-01 4.65885878e-01 -3.77512595e-04 2.50594884e-01
7.75808692e-02 8.72808695e-01 -6.47109985e-01 -5.14645986e-02
2.58328676e-01 -1.58335626e-01 -7.43684411e-01 1.06216156e+00
-3.13035280e-01 5.02486706e-01 -4.49259639e-01 -2.79023081e-01
-1.05513535e-01 1.47045061e-01 -8.63342702e-01 1.69558570e-01
1.05459727e-01 1.03941716e-01 -4.77594405e-01 6.80154204e-01
1.06331336e+00 7.58253187e-02 1.51968256e-01 3.47353458e-01
2.05076382e-01 1.36154428e-01 -1.01263297e+00 5.21959126e-01
-1.57487944e-01 2.85051882e-01 1.23736203e-01 -8.49656880e-01
6.18618071e-01 2.05554977e-01 9.09494460e-02 -1.99178666e-01
4.85039890e-01 1.86613664e-01 1.08538114e-01 -2.18919232e-01
1.27442345e-01 -3.61798227e-01 -6.91374004e-01 9.17364001e-01
-5.16923308e-01 4.26121652e-01 -4.52424288e-01 2.55934566e-01
1.15635943e+00 -7.18258083e-01 3.92069489e-01 -6.95050538e-01
4.03578430e-01 -1.98689476e-01 7.97366381e-01 1.27525723e+00
-7.94066370e-01 2.95799345e-01 7.36891091e-01 -3.32859814e-01
-1.17873847e+00 -1.56557429e+00 -3.01799923e-01 7.75941432e-01
2.53257394e-01 -1.06045440e-01 -9.81469870e-01 -1.00184083e+00
6.04624331e-01 8.20686817e-01 -8.09000731e-01 -3.53106737e-01
-2.58418828e-01 -9.85557675e-01 1.41163480e+00 2.95089066e-01
4.52735901e-01 -4.46126670e-01 -2.73140550e-01 -5.15652001e-01
4.27130014e-02 -7.03356624e-01 -5.45909107e-01 -2.95991182e-01
-5.08380115e-01 -1.23068964e+00 -2.33505115e-01 -3.46231937e-01
6.71262026e-01 -1.21535230e-02 8.01633000e-01 2.96030641e-01
-3.59475255e-01 3.41471732e-01 1.99389860e-01 -5.24852574e-01
-8.01503062e-01 -1.09309047e-01 3.80271137e-01 -1.77461207e-01
7.26570249e-01 -7.65452385e-01 -5.55874169e-01 2.11646453e-01
-7.95846879e-01 -7.81004548e-01 1.34104386e-03 2.85116166e-01
1.57665089e-01 -5.96309118e-02 9.48763728e-01 -1.41009784e+00
8.13049614e-01 -1.06021786e+00 -6.80158854e-01 3.47889394e-01
-4.82179254e-01 -2.05885768e-01 8.88109744e-01 -4.73961592e-01
-8.34409773e-01 -2.42108911e-01 -6.22313768e-02 -2.21318364e-01
-3.49803984e-01 6.82762563e-02 -4.94341552e-01 -2.77006060e-01
9.33409870e-01 -1.45924717e-01 -3.50594856e-02 -2.22406402e-01
4.08455104e-01 8.39405954e-01 5.03688872e-01 -1.00966799e+00
1.09625089e+00 6.55084312e-01 6.52829334e-02 -4.93398249e-01
-6.04431570e-01 -2.02039047e-03 -2.71971583e-01 2.53436774e-01
3.60256076e-01 -6.51266813e-01 -9.93134558e-01 3.61122787e-01
-7.19048321e-01 -4.66551989e-01 -3.27252001e-02 2.13727340e-01
-3.27624768e-01 6.27213061e-01 -5.94006956e-01 -1.02108777e+00
-3.44913572e-01 -6.73065841e-01 2.20720798e-01 1.59840703e-01
-5.54705739e-01 -1.23618972e+00 2.05485478e-01 4.18258280e-01
2.80118912e-01 4.47275192e-01 9.13710535e-01 -1.39821875e+00
-1.22083649e-01 -6.12079918e-01 -1.21736296e-01 2.07598254e-01
3.25713120e-02 2.71037787e-01 -1.25393355e+00 -3.82680416e-01
-4.78480477e-03 -1.07223317e-01 6.04857087e-01 3.46876293e-01
1.35616422e+00 -9.99751747e-01 -3.40068132e-01 7.40251780e-01
1.40022409e+00 3.36410441e-02 6.19816780e-01 1.12721473e-01
4.90431845e-01 5.09624004e-01 1.42318740e-01 7.23761201e-01
3.47592980e-01 3.12078744e-01 1.65776029e-01 7.67236501e-02
5.07497668e-01 -3.26016098e-01 3.35181653e-01 -1.89743564e-01
1.21781170e-01 -2.25811556e-01 -8.80154669e-01 5.65725803e-01
-1.45672321e+00 -1.47810197e+00 -5.82617149e-02 2.76727343e+00
8.80321264e-01 -1.65708400e-02 6.61896944e-01 -5.95871396e-02
9.58297729e-01 -1.39961109e-01 -5.93860686e-01 -9.16138589e-01
-5.33208735e-02 1.04168683e-01 1.05709064e+00 8.46020460e-01
-1.10735154e+00 6.27714097e-01 6.95004082e+00 7.90478051e-01
-6.65843487e-01 9.99930054e-02 1.14367545e+00 -2.50114799e-01
-7.56444454e-01 8.93908832e-03 -5.47167063e-01 6.14253283e-01
1.20474946e+00 -9.96601164e-01 1.78881302e-01 8.55461776e-01
-1.94976777e-01 3.37583691e-01 -1.10066617e+00 4.14309621e-01
-1.05124421e-01 -1.19256616e+00 6.28811121e-02 5.04767418e-01
7.11913645e-01 -3.32149655e-01 4.56999898e-01 2.34367371e-01
9.62001622e-01 -1.25168645e+00 5.25012672e-01 2.80886620e-01
8.18832636e-01 -1.26696253e+00 5.80001652e-01 3.98502856e-01
-8.48597109e-01 -3.82601917e-01 -5.36905944e-01 -6.50701821e-02
-2.88737148e-01 4.68315214e-01 -3.97439152e-01 3.38384151e-01
4.73152906e-01 -1.58044070e-01 -3.73942047e-01 7.90192306e-01
1.71961859e-01 7.10317552e-01 -6.15246058e-01 2.62575507e-01
-7.49184266e-02 8.61094072e-02 5.53828895e-01 1.22377431e+00
2.03134164e-01 3.78545791e-01 1.90060548e-02 1.10811639e+00
-3.73962164e-01 -8.24673399e-02 -1.04058969e+00 9.38445926e-02
1.20403874e+00 7.23747134e-01 -1.53917283e-01 -2.70599276e-01
-7.44059905e-02 6.86371863e-01 1.32112056e-01 3.19374502e-01
-8.17391753e-01 -4.25522059e-01 1.19598281e+00 8.80660936e-02
-2.05935314e-01 1.80696249e-01 -7.71360159e-01 -1.20488489e+00
-1.65923923e-01 -9.17608023e-01 9.51064169e-01 4.76930439e-02
-1.75742936e+00 1.44077703e-01 7.62826055e-02 -9.70275640e-01
-1.55160606e-01 -2.07144141e-01 -1.13476384e+00 1.30122268e+00
-9.73032594e-01 -1.02371979e+00 3.57401401e-01 7.08063483e-01
-4.04469967e-01 -2.70461887e-01 9.56967950e-01 1.65136635e-01
-7.08450794e-01 1.54191124e+00 3.99161652e-02 5.09096146e-01
5.82837641e-01 -1.03635001e+00 7.29543567e-01 1.14265311e+00
-1.37329489e-01 1.01582348e+00 6.86363339e-01 -8.22436035e-01
-1.06742942e+00 -1.08815587e+00 9.45182979e-01 -9.44907725e-01
6.70592606e-01 -3.67237806e-01 -9.44180131e-01 9.57009673e-01
-3.57764602e-01 1.38029397e-01 1.31034398e+00 8.97378102e-02
-1.04993844e+00 1.07041828e-01 -2.03680849e+00 7.20500112e-01
9.03751612e-01 -7.36134171e-01 -4.08584654e-01 -8.73982683e-02
4.62539673e-01 3.91306877e-02 -8.40846419e-01 2.45815024e-01
8.06598186e-01 -1.10652053e+00 1.19055140e+00 -1.06253088e+00
6.65884698e-03 -9.39731300e-02 -4.18096840e-01 -7.49361694e-01
-1.19289368e-01 -7.43891180e-01 -7.70733133e-02 1.41219544e+00
5.70888579e-01 -1.18893600e+00 6.53079033e-01 1.22399485e+00
6.77022636e-01 -3.75700861e-01 -1.01117086e+00 -8.46147716e-01
1.03248262e+00 -3.15089434e-01 9.13612127e-01 1.43054879e+00
2.21794188e-01 -5.63844085e-01 -3.05061281e-01 8.24799776e-01
1.23730755e+00 5.14816120e-02 7.98673570e-01 -1.05534708e+00
-1.77008867e-01 -5.85357666e-01 -5.22343874e-01 -1.43522784e-01
4.49118972e-01 -8.36356819e-01 -3.63107264e-01 -6.78296566e-01
4.98134851e-01 -9.43436503e-01 -2.43208617e-01 4.02393132e-01
-4.39136386e-01 3.53274584e-01 2.32920468e-01 2.42656514e-01
-9.03312415e-02 -5.03473915e-02 3.18784475e-01 6.24937452e-02
6.32285848e-02 4.34589773e-01 -1.29322159e+00 7.94296384e-01
1.44197631e+00 -6.46033227e-01 -3.95842522e-01 2.44615138e-01
1.99283272e-01 5.06398343e-02 6.80466652e-01 -8.23256791e-01
4.28796336e-02 -4.15491641e-01 3.24260920e-01 2.01284856e-01
-2.24530548e-01 -6.44540846e-01 2.20935822e-01 8.25742662e-01
-7.18182147e-01 -1.55237094e-01 2.46349975e-01 5.94039917e-01
4.86905813e-01 -1.30255371e-01 1.01623881e+00 7.20445216e-02
-3.68184745e-02 5.48404396e-01 -3.35640937e-01 4.31932509e-01
1.24743032e+00 8.22102875e-02 -8.38421881e-01 -4.89030927e-01
-5.84083438e-01 2.15063527e-01 9.82829213e-01 -1.85432136e-01
3.55508000e-01 -1.02104223e+00 -9.39917982e-01 2.31540784e-01
-2.76968116e-03 -6.61893249e-01 2.39764780e-01 3.61827731e-01
-4.24389213e-01 -4.42877635e-02 -2.29814664e-01 1.31229386e-01
-1.55769944e+00 7.70810425e-01 6.52013361e-01 1.63285688e-01
-3.55715454e-01 7.85982192e-01 3.23019683e-01 -3.83126974e-01
2.58889169e-01 1.74356326e-01 3.86193305e-01 -4.88458604e-01
8.82220745e-01 4.72621024e-01 -4.95687574e-01 -6.49245918e-01
-5.17651200e-01 2.57135272e-01 -2.60843579e-02 -2.58122683e-01
6.32363558e-01 -1.46539986e-01 -3.29322070e-02 -2.21044049e-02
1.31210196e+00 5.37629604e-01 -9.05150294e-01 3.08300722e-02
-2.13210851e-01 -8.04646134e-01 -4.40605104e-01 -7.38875031e-01
-8.61025035e-01 6.81589425e-01 3.96047443e-01 5.54799795e-01
7.89088964e-01 -1.99677679e-03 6.59438074e-01 2.38132030e-01
4.81785417e-01 -5.13212323e-01 -8.12384784e-01 -1.67055711e-01
4.33546305e-01 -1.08386993e+00 1.25424847e-01 -4.12832558e-01
-4.98409569e-01 5.96172631e-01 3.86618853e-01 -3.36119175e-01
9.09488738e-01 6.19264245e-01 1.43892363e-01 8.06214437e-02
-5.04597008e-01 4.04468477e-01 -1.11031972e-01 8.42217207e-01
1.70144979e-02 5.00842810e-01 -3.01887602e-01 9.56101179e-01
-3.11991036e-01 -2.48424396e-01 7.30760098e-01 8.25305641e-01
-4.12406564e-01 -9.78608489e-01 -4.74123418e-01 5.13076544e-01
-1.10077012e+00 -9.37122405e-02 -6.80019975e-01 6.12068653e-01
-6.28824905e-02 8.81763577e-01 8.79633650e-02 -3.87246370e-01
4.40618582e-03 -1.05927631e-01 2.54789114e-01 -4.20291185e-01
-8.09502840e-01 -6.46267176e-01 2.51884788e-01 -3.96844566e-01
6.04781955e-02 -6.50965989e-01 -1.01619852e+00 -1.42701864e+00
-1.04939630e-02 3.12380672e-01 2.13702857e-01 5.81774592e-01
4.93185133e-01 -3.15572500e-01 1.01364076e+00 -5.92517853e-02
-1.05682635e+00 -4.03438270e-01 -7.04538882e-01 4.09788162e-01
3.88693929e-01 -3.72805037e-02 -9.86601830e-01 -1.29565179e-01] | [5.975590705871582, 7.0305328369140625] |
b28b0074-4a23-4b9c-8fc6-2199abc77e8b | xiaomingbot-a-multilingual-robot-news-1 | 2007.08005 | null | https://arxiv.org/abs/2007.08005v1 | https://arxiv.org/pdf/2007.08005v1.pdf | Xiaomingbot: A Multilingual Robot News Reporter | This paper proposes the building of Xiaomingbot, an intelligent, multilingual and multimodal software robot equipped with four integral capabilities: news generation, news translation, news reading and avatar animation. Its system summarizes Chinese news that it automatically generates from data tables. Next, it translates the summary or the full article into multiple languages, and reads the multilingual rendition through synthesized speech. Notably, Xiaomingbot utilizes a voice cloning technology to synthesize the speech trained from a real person's voice data in one input language. The proposed system enjoys several merits: it has an animated avatar, and is able to generate and read multilingual news. Since it was put into practice, Xiaomingbot has written over 600,000 articles, and gained over 150,000 followers on social media platforms. | ['Yu-Ping Wang', 'Lei LI', 'Xijin Zhang', 'Songcheng Jiang', 'Runxin Xu', 'Yuxuan Wang', 'Li Chen', 'Jun Cao', 'Hao Zhou', 'Xiang Yin', 'Jiaze Chen', 'Ying Zeng', 'Mingxuan Wang'] | 2020-07-12 | xiaomingbot-a-multilingual-robot-news | https://aclanthology.org/2020.acl-demos.1 | https://aclanthology.org/2020.acl-demos.1.pdf | acl-2020-6 | ['news-generation', 'voice-cloning'] | ['natural-language-processing', 'speech'] | [-3.21921617e-01 7.30246305e-01 -4.76646990e-01 1.97000541e-02
-8.00632358e-01 -7.42061079e-01 1.19906294e+00 -2.09030703e-01
-1.83055624e-01 1.09793162e+00 7.37965167e-01 -2.79861182e-01
6.23483598e-01 -6.87101662e-01 -8.85063350e-01 -1.31266296e-01
3.65758002e-01 8.40595961e-01 -1.37579679e-01 -7.86898553e-01
2.17164587e-02 6.87229410e-02 -1.24548113e+00 7.02221245e-02
9.59019065e-01 6.44692183e-01 6.09819829e-01 8.17091584e-01
-4.54278439e-01 1.39797950e+00 -1.06042624e+00 -7.94847727e-01
-2.38925979e-01 -5.73502898e-01 -9.71078277e-01 -2.47435525e-01
-1.69418484e-01 -3.51161629e-01 -3.82784396e-01 1.11188138e+00
3.82419825e-01 -2.28509188e-01 3.39633137e-01 -1.51894999e+00
-1.25981629e+00 1.51676488e+00 -4.23372537e-03 -5.04172564e-01
1.14047432e+00 3.16138193e-02 9.41077828e-01 -5.92703462e-01
1.25525868e+00 1.40865552e+00 4.65318829e-01 5.62076271e-01
-6.82184517e-01 -2.37222642e-01 -4.15805757e-01 -1.88513592e-01
-1.04144430e+00 -5.49070835e-01 4.61026549e-01 -3.37008685e-01
8.13971937e-01 3.02173883e-01 8.67840827e-01 1.49998260e+00
5.97302914e-01 6.64010465e-01 9.61404502e-01 -4.71616685e-01
-2.14229703e-01 7.20743895e-01 -2.79417008e-01 8.43999326e-01
-3.04023594e-01 -2.20882073e-01 -8.33332717e-01 -1.57283694e-01
7.29473889e-01 -5.69939315e-01 -1.39058456e-01 2.90028095e-01
-2.17343593e+00 6.74148679e-01 2.59692338e-03 1.52511016e-01
-3.63951206e-01 -2.19632089e-02 5.45416951e-01 4.32118475e-01
2.02598080e-01 6.44563735e-01 -1.58150747e-01 -8.24136496e-01
-2.01440945e-01 3.67524847e-02 1.24445605e+00 1.70989585e+00
2.31908098e-01 7.58514479e-02 1.17567152e-01 9.02854383e-01
5.45363426e-01 1.28525639e+00 7.04925895e-01 -1.24734318e+00
5.76146483e-01 4.27892685e-01 4.44110781e-01 -1.10337257e+00
-4.17533696e-01 2.33046278e-01 -5.13121426e-01 -2.02847555e-01
1.35779798e-01 -7.09306598e-01 -1.32430807e-01 1.30906641e+00
2.04942942e-01 -5.79843402e-01 7.76429296e-01 7.45679617e-01
1.42390418e+00 1.29099572e+00 -2.29798764e-01 -3.33768368e-01
1.58786201e+00 -1.43960130e+00 -1.15924835e+00 2.06474159e-02
5.79048060e-02 -1.09891701e+00 1.32045424e+00 3.87575567e-01
-1.06437314e+00 -5.72670102e-01 -9.67181504e-01 -2.97272682e-01
-3.55374485e-01 5.55923581e-01 3.29396307e-01 2.76474088e-01
-8.06185663e-01 8.23301300e-02 -4.67293620e-01 -7.76768565e-01
-5.62320590e-01 -1.00919351e-01 -4.23251927e-01 6.58319354e-01
-1.34650385e+00 1.03275394e+00 3.74863654e-01 -3.82001936e-01
-7.00282037e-01 -9.97560769e-02 -8.36864471e-01 -3.63535255e-01
3.55273545e-01 -7.96322167e-01 1.82301521e+00 -7.56303132e-01
-2.61538696e+00 7.02824950e-01 1.11151218e-01 -2.66318172e-01
8.21746826e-01 -3.07119220e-01 -9.83732998e-01 1.67969570e-01
4.91315901e-01 5.22793710e-01 5.30644596e-01 -1.00361204e+00
-8.89538944e-01 1.19379044e-01 6.96595758e-02 2.35746354e-01
1.08723722e-01 4.73189712e-01 -6.51201010e-01 -7.42367387e-01
-5.99186897e-01 -8.52430224e-01 8.22886229e-02 -4.07817483e-01
-8.98786426e-01 -5.28143384e-02 4.35055524e-01 -7.53722429e-01
8.99851561e-01 -1.86219525e+00 2.78037757e-01 -8.02600756e-02
7.45379403e-02 -7.15996325e-02 -2.56499667e-02 9.41610277e-01
8.55999112e-01 -7.73243681e-02 1.96231425e-01 -4.54247072e-02
2.50957042e-01 1.40680656e-01 -4.86739367e-01 7.11263940e-02
-3.35777551e-01 9.26186740e-01 -9.57767963e-01 -6.06735945e-01
-8.28122944e-02 3.55718464e-01 -1.93173826e-01 2.85062373e-01
-5.19457817e-01 4.35493946e-01 -2.94976383e-01 5.00694692e-01
1.43587485e-01 -1.85528100e-01 4.09859866e-01 2.03729078e-01
-7.08182514e-01 2.38478214e-01 -8.19524884e-01 1.60683048e+00
-6.44720137e-01 8.32506239e-01 2.10160077e-01 -1.10828653e-01
1.34099722e+00 5.88883281e-01 1.78373605e-01 -5.72671056e-01
5.17220974e-01 4.03665900e-01 -6.28882527e-01 -1.05376351e+00
1.08631694e+00 2.90048957e-01 -7.41818070e-01 7.70410538e-01
1.27431318e-01 -1.79541320e-01 2.53314853e-01 2.60739714e-01
4.85471219e-01 2.38391206e-01 6.30404949e-01 -5.76535938e-03
5.55992007e-01 5.65465629e-01 1.78293437e-01 6.27718389e-01
1.01034120e-01 2.63748288e-01 5.75070381e-01 -1.11490525e-01
-1.12083399e+00 -1.11769724e+00 3.01020980e-01 1.32053208e+00
5.55809259e-01 -5.87023854e-01 -9.13182139e-01 -4.36787128e-01
-2.75200546e-01 1.02172518e+00 -1.48244396e-01 2.19789222e-01
-3.83768022e-01 -1.91677690e-01 1.06260574e+00 9.76308957e-02
7.01479793e-01 -1.46623421e+00 -5.43823302e-01 4.46218580e-01
-8.18338513e-01 -1.31250226e+00 -6.59111440e-01 -3.42660427e-01
-7.70651922e-02 -8.85006070e-01 -8.46164823e-01 -1.18649697e+00
1.69559300e-01 1.02468878e-02 8.03283334e-01 -4.48928088e-01
3.67610008e-01 3.76732618e-01 -5.65461338e-01 -4.93578434e-01
-1.45674539e+00 3.26667041e-01 3.17002267e-01 -1.81368947e-01
-1.71039671e-01 -1.54543415e-01 3.70984882e-01 2.17829123e-01
-3.66229206e-01 6.78887546e-01 6.06677711e-01 7.12588727e-01
9.11499485e-02 -6.49894357e-01 7.65003204e-01 -9.48211849e-01
1.12932026e+00 -4.93865281e-01 -4.20137346e-01 6.94575489e-01
-5.38787305e-01 3.26765403e-02 9.55005288e-01 -3.80231172e-01
-1.29021692e+00 -1.46160662e-01 -1.79728433e-01 3.49628896e-01
-1.21575892e-01 7.97462642e-01 -9.89163071e-02 4.40178931e-01
7.71475017e-01 5.95737934e-01 5.25511324e-01 -4.93995667e-01
9.18975830e-01 1.36033309e+00 1.08085895e+00 -5.37962019e-01
4.74180073e-01 7.41571486e-02 -8.46695900e-01 -1.04974687e+00
-3.24820564e-03 1.64379418e-01 -3.49527389e-01 -7.04395950e-01
9.59166944e-01 -9.39513862e-01 -1.36049628e+00 6.83717072e-01
-1.59381104e+00 -3.85877252e-01 -6.05701245e-02 6.14519417e-01
-8.18935275e-01 8.71722028e-02 -1.06935084e+00 -5.46792507e-01
-5.69090724e-01 -1.06773877e+00 7.50427067e-01 4.13115442e-01
-5.25933981e-01 -7.47938693e-01 2.29672596e-01 2.39287257e-01
4.07313943e-01 1.03802942e-02 7.21528590e-01 -6.32969558e-01
-1.25695646e-01 -6.63267672e-02 -4.32800837e-02 -1.07519068e-01
1.26692489e-01 6.88477218e-01 -2.68614322e-01 4.42292064e-01
-6.02538943e-01 -3.17359567e-01 -9.76190791e-02 -8.83124098e-02
-1.72823831e-01 -9.79102254e-01 -1.34128332e-01 2.73224562e-01
6.70276165e-01 6.79278195e-01 3.07243407e-01 7.02846050e-01
5.92517197e-01 5.99782050e-01 7.66069114e-01 4.54269588e-01
9.88808393e-01 6.21255398e-01 3.23955789e-02 3.81090403e-01
1.07659616e-01 -8.01954865e-01 9.42389727e-01 1.45457494e+00
-2.71252066e-01 -3.08223516e-01 -7.82902956e-01 2.90097259e-02
-1.85952330e+00 -1.12373126e+00 -1.73282087e-01 1.38917148e+00
9.02848661e-01 -7.78419450e-02 4.10533398e-01 -5.97174644e-01
8.89751136e-01 6.25784248e-02 -2.92777389e-01 -5.78138947e-01
-3.87431681e-01 -5.87928891e-01 3.91142249e-01 5.63593268e-01
-6.47917211e-01 1.25918090e+00 6.96839142e+00 3.76642883e-01
-1.40363431e+00 5.31696752e-02 1.65550604e-01 5.00519633e-01
-3.10880005e-01 -2.77059227e-01 -6.38719976e-01 3.40535492e-01
9.78677690e-01 -1.05301094e+00 7.81725049e-01 1.09171927e+00
3.04575831e-01 1.76601503e-02 -7.36260831e-01 1.02890205e+00
3.26555550e-01 -1.84329081e+00 1.70768276e-01 -4.14461732e-01
5.30537188e-01 1.65709049e-01 -7.34249353e-02 3.71502072e-01
6.73806548e-01 -6.11248314e-01 1.69562030e+00 8.04792643e-01
8.59600544e-01 -6.57126248e-01 6.16731703e-01 7.01085329e-01
-9.13948238e-01 1.01073183e-01 -1.32077664e-01 -9.22371075e-03
5.56488574e-01 -2.29515865e-01 -1.11740947e+00 6.91446424e-01
4.05837774e-01 7.49903917e-01 -3.49429160e-01 4.48733717e-01
-4.15820599e-01 3.49540234e-01 -2.08798479e-02 -7.88030326e-01
-5.88597506e-02 -5.30926108e-01 8.31505179e-01 1.22580683e+00
7.64264822e-01 -2.27874130e-01 1.00165740e-01 5.49126506e-01
-2.58044124e-01 5.05661964e-01 -6.11655235e-01 -5.42202353e-01
6.09808028e-01 1.23088109e+00 -4.49890733e-01 -7.09826887e-01
-2.88703024e-01 1.03660667e+00 -1.61211729e-01 2.60583222e-01
-1.04090393e+00 -6.78107381e-01 1.08415373e-01 -3.51480246e-01
-3.60683948e-01 -1.56094179e-01 -3.34904492e-01 -1.16366005e+00
-3.77355009e-01 -1.34523129e+00 2.85636019e-02 -1.41854882e+00
-9.55608189e-01 1.47256851e+00 -1.96603030e-01 -1.41053581e+00
-6.95943296e-01 -2.39389509e-01 -4.23238695e-01 3.71009737e-01
-5.73244810e-01 -1.31377304e+00 -4.49164882e-02 3.91642511e-01
7.12884128e-01 -8.90934110e-01 9.02820289e-01 2.77619749e-01
-4.72830296e-01 3.87163013e-01 1.15490049e-01 2.32888177e-01
8.72209311e-01 -9.79058027e-01 3.50670457e-01 4.57670718e-01
1.86091632e-01 6.93961143e-01 1.12144911e+00 -8.35042417e-01
-1.63953674e+00 -8.49252582e-01 1.40577173e+00 -1.86183333e-01
9.93689477e-01 -2.15909168e-01 -2.19456419e-01 9.93474424e-01
1.03375649e+00 -8.72625947e-01 4.36253607e-01 -5.41695833e-01
-5.60379960e-02 -8.64991620e-02 -1.01334321e+00 1.01648653e+00
7.55384088e-01 -6.36056781e-01 -8.59748542e-01 3.62979531e-01
1.13075662e+00 -6.67020082e-01 -1.00837255e+00 -3.10901105e-01
7.63026237e-01 -6.24655128e-01 4.14228678e-01 -3.05857271e-01
8.00959170e-01 -4.16515291e-01 -8.33605081e-02 -1.59958053e+00
1.42174467e-01 -1.30419326e+00 2.70547718e-01 1.61480439e+00
1.03510606e+00 -8.57375920e-01 6.70643002e-02 1.79629862e-01
-4.95290935e-01 -6.11633919e-02 -5.79444587e-01 -5.37380040e-01
-5.13375700e-01 -2.90764421e-01 8.93617749e-01 1.05749321e+00
8.83526325e-01 7.74483263e-01 -1.00318420e+00 5.41545637e-02
3.20229195e-02 -2.17581745e-02 1.31512976e+00 -1.02372837e+00
-2.74766445e-01 -2.44552284e-01 8.20061490e-02 -1.30839443e+00
1.04204364e-01 -9.69726562e-01 2.27450281e-01 -1.71780884e+00
-1.70333534e-01 4.02748697e-02 8.40448260e-01 3.63560855e-01
4.49328065e-01 -3.37567963e-02 3.84879380e-01 4.83953893e-01
-6.53681040e-01 6.10778451e-01 1.49031174e+00 -2.55242825e-01
-3.45071554e-01 -2.99837012e-02 -6.83665931e-01 7.10327685e-01
6.66289449e-01 -2.81320840e-01 8.04522168e-03 -1.61576122e-01
4.90282416e-01 6.65708721e-01 8.00048783e-02 -7.47500360e-01
5.16155601e-01 -3.02411199e-01 -1.27671823e-01 -4.57817227e-01
1.21464506e-01 -6.40618980e-01 6.94734693e-01 5.14708161e-01
-6.03006005e-01 6.79440141e-01 -2.91356146e-01 2.83355713e-01
-4.12533969e-01 -6.63234852e-03 3.45461726e-01 -9.50545073e-02
-8.03143203e-01 -3.61954153e-01 -1.18872404e+00 -1.01323210e-01
1.08189428e+00 2.24641144e-01 -9.02541578e-01 -7.13089705e-01
-6.43767476e-01 1.28186569e-01 5.08809805e-01 7.86578596e-01
4.59794670e-01 -1.49451005e+00 -5.89101493e-01 4.29539643e-02
1.73393995e-01 -5.53294301e-01 -2.66169488e-01 6.95531547e-01
-1.34867465e+00 3.55102807e-01 -5.05568445e-01 -3.01200658e-01
-1.05453908e+00 4.27472830e-01 -8.12440142e-02 2.56238818e-01
-7.54184723e-01 2.45710418e-01 -7.14102983e-01 -9.06684160e-01
2.71262795e-01 -1.85002074e-01 -4.92778450e-01 1.92853957e-01
6.76190555e-01 6.57331526e-01 -5.10373354e-01 -1.11361694e+00
-3.69092412e-02 2.44353965e-01 3.21469158e-01 -8.74006152e-01
1.06544197e+00 -6.68517888e-01 -4.22796249e-01 1.06044209e+00
8.20354283e-01 7.44462311e-01 -5.64850152e-01 3.17363441e-02
3.71127296e-03 1.30058959e-01 -8.20408881e-01 -9.47299242e-01
-5.12092054e-01 1.37166977e-01 -1.80530757e-01 5.38940966e-01
4.10449147e-01 3.04856271e-01 1.21597552e+00 7.69366920e-01
7.44751930e-01 -1.33998382e+00 -8.13498870e-02 9.24142599e-01
1.18679571e+00 -1.08665991e+00 -4.70678329e-01 -2.16726393e-01
-1.45600533e+00 1.47319961e+00 4.31695789e-01 2.24826559e-01
2.96640962e-01 4.83149379e-01 1.03274894e+00 -1.02924533e-01
-8.36923718e-01 3.47272456e-01 -6.43709227e-02 7.91634142e-01
2.53744930e-01 4.16990101e-01 -4.08118188e-01 8.50243568e-01
-1.06690252e+00 1.09139699e-02 1.00903893e+00 5.68923295e-01
-5.40577710e-01 -7.63357222e-01 -5.98018169e-01 -9.39638019e-02
-3.34678173e-01 2.17700079e-01 -6.29126370e-01 9.70044971e-01
-2.42687404e-01 1.20391035e+00 -1.67604387e-01 -8.00642312e-01
4.95076984e-01 -1.86858978e-02 -1.75045475e-01 -3.46289635e-01
-6.32528603e-01 5.23892872e-04 6.63157225e-01 -2.07097352e-01
-4.54059243e-01 -4.33142304e-01 -1.61131513e+00 -6.18419588e-01
2.47620463e-01 6.63737953e-01 7.87630975e-01 9.00264025e-01
3.09888333e-01 3.06327552e-01 6.47627771e-01 -7.34535217e-01
-2.57935256e-01 -1.00446558e+00 -4.79100198e-01 -1.51809618e-01
1.95921645e-01 -1.52967900e-01 -2.54342202e-02 3.68595600e-01] | [12.401538848876953, 8.490050315856934] |
99c139d8-c373-4bc2-a8ac-1164a30457c1 | superglue-learning-feature-matching-with | 1911.11763 | null | https://arxiv.org/abs/1911.11763v2 | https://arxiv.org/pdf/1911.11763v2.pdf | SuperGlue: Learning Feature Matching with Graph Neural Networks | This paper introduces SuperGlue, a neural network that matches two sets of local features by jointly finding correspondences and rejecting non-matchable points. Assignments are estimated by solving a differentiable optimal transport problem, whose costs are predicted by a graph neural network. We introduce a flexible context aggregation mechanism based on attention, enabling SuperGlue to reason about the underlying 3D scene and feature assignments jointly. Compared to traditional, hand-designed heuristics, our technique learns priors over geometric transformations and regularities of the 3D world through end-to-end training from image pairs. SuperGlue outperforms other learned approaches and achieves state-of-the-art results on the task of pose estimation in challenging real-world indoor and outdoor environments. The proposed method performs matching in real-time on a modern GPU and can be readily integrated into modern SfM or SLAM systems. The code and trained weights are publicly available at https://github.com/magicleap/SuperGluePretrainedNetwork. | ['Daniel DeTone', 'Paul-Edouard Sarlin', 'Tomasz Malisiewicz', 'Andrew Rabinovich'] | 2019-11-26 | superglue-learning-feature-matching-with-1 | http://openaccess.thecvf.com/content_CVPR_2020/html/Sarlin_SuperGlue_Learning_Feature_Matching_With_Graph_Neural_Networks_CVPR_2020_paper.html | http://openaccess.thecvf.com/content_CVPR_2020/papers/Sarlin_SuperGlue_Learning_Feature_Matching_With_Graph_Neural_Networks_CVPR_2020_paper.pdf | cvpr-2020-6 | ['image-matching'] | ['computer-vision'] | [-1.16961934e-01 1.67668372e-01 -1.01806454e-01 -6.73089385e-01
-6.87760830e-01 -5.95969498e-01 5.17728090e-01 2.29436964e-01
-5.32025337e-01 5.03671288e-01 1.29762217e-01 -1.96231768e-01
-7.53572732e-02 -8.25916350e-01 -1.18686759e+00 -2.35394970e-01
-3.21054041e-01 9.78098035e-01 1.69721887e-01 -7.28477761e-02
4.01928753e-01 7.20064044e-01 -1.45759141e+00 -5.69952764e-02
5.81847131e-01 1.16917038e+00 4.99774367e-01 6.62873149e-01
1.07504167e-01 3.67764115e-01 -3.15425470e-02 -2.93640763e-01
5.48495173e-01 2.88677216e-01 -8.23822021e-01 -4.07355428e-02
1.11207747e+00 -1.72194391e-01 -6.02293730e-01 9.86879170e-01
5.36057830e-01 5.23096502e-01 2.22509742e-01 -1.31849730e+00
-6.45563245e-01 1.03080116e-01 -4.47135121e-01 -5.13642877e-02
5.44072926e-01 4.13217135e-02 1.08493483e+00 -1.21603811e+00
6.91545665e-01 1.26870012e+00 9.96354938e-01 8.15380812e-02
-1.16617477e+00 -3.95994544e-01 4.49454606e-01 1.23895846e-01
-1.56072938e+00 -3.49646747e-01 6.29911482e-01 -2.89616227e-01
1.38425660e+00 1.61829621e-01 6.37450218e-01 8.27356160e-01
3.16906273e-01 6.17581725e-01 5.42820215e-01 -1.02466591e-01
2.09528446e-01 -4.04829830e-01 -2.37465799e-01 1.19878948e+00
1.15617383e-02 1.67940378e-01 -8.14903557e-01 -3.45034689e-01
9.11781907e-01 1.42445946e-02 -2.74850905e-01 -9.73550677e-01
-1.72258067e+00 5.81175327e-01 1.07569277e+00 -3.25380504e-01
-3.75438273e-01 6.78050160e-01 1.30548552e-01 1.63827002e-01
5.53601444e-01 4.38039958e-01 -4.92089838e-01 8.31940845e-02
-5.73980451e-01 3.43525857e-01 6.87870085e-01 1.43252182e+00
1.07004547e+00 -3.80436122e-01 1.60960540e-01 5.56066513e-01
3.88927579e-01 6.78260148e-01 -1.82759881e-01 -1.28771353e+00
6.63762927e-01 3.43510538e-01 3.09007764e-01 -1.25619066e+00
-6.15645409e-01 -5.19237578e-01 -6.33811414e-01 1.65523201e-01
1.86546877e-01 -2.77271192e-03 -1.00388408e+00 1.58969140e+00
6.84988141e-01 4.65804756e-01 -2.95193374e-01 1.38519526e+00
7.13246822e-01 5.16259253e-01 -3.00703734e-01 6.41317129e-01
9.82397377e-01 -1.31577992e+00 -6.71881139e-02 -5.78421831e-01
3.82079780e-01 -8.93503666e-01 7.06177652e-01 8.00085589e-02
-9.36214089e-01 -5.10931492e-01 -9.56500828e-01 -5.52895665e-01
-4.04430836e-01 6.00720458e-02 8.88163030e-01 7.36596361e-02
-1.27243996e+00 8.41328442e-01 -1.13906074e+00 -5.40231705e-01
4.00289029e-01 7.42202640e-01 -5.46157300e-01 -2.01013476e-01
-6.41227782e-01 8.44295979e-01 3.32823068e-01 5.61888397e-01
-7.84639478e-01 -6.72271848e-01 -1.18170059e+00 -2.61888742e-01
3.93989235e-01 -1.18969405e+00 1.22129524e+00 -4.20859694e-01
-1.24177897e+00 1.01123023e+00 -1.98528185e-01 -3.94067496e-01
5.27170658e-01 -6.44556224e-01 6.48645386e-02 -1.24120824e-01
3.72808963e-01 9.40302849e-01 5.05486012e-01 -1.17474926e+00
-7.11835623e-01 -4.39058781e-01 3.06501277e-02 4.36692446e-01
5.43281913e-01 -4.03715312e-01 -8.50840271e-01 -3.01825613e-01
7.19178259e-01 -1.14873338e+00 -6.42535508e-01 5.24317384e-01
-5.31634688e-01 -4.55281697e-02 5.98697305e-01 -3.99957716e-01
2.22770154e-01 -1.83156753e+00 2.87282050e-01 4.75546986e-01
2.10645512e-01 -2.16361612e-01 -3.29661518e-01 4.56118762e-01
3.29752296e-01 -3.80111039e-01 -2.69806325e-01 -8.13427985e-01
3.84711921e-01 3.74928236e-01 -1.79394171e-01 9.91765320e-01
6.49473630e-03 8.52592707e-01 -1.29250991e+00 -1.00170985e-01
5.97046554e-01 6.41582847e-01 -6.22294009e-01 1.80139557e-01
-3.26599270e-01 5.39593220e-01 -4.63640034e-01 6.89606369e-01
9.54392195e-01 -2.70446002e-01 -3.75007093e-02 -1.34650037e-01
-1.66150838e-01 5.26778936e-01 -1.27007008e+00 2.62909174e+00
-6.80104256e-01 6.70130312e-01 1.72627032e-01 -7.70556986e-01
1.07561111e+00 -3.08444232e-01 3.19956243e-01 -4.67505217e-01
1.31053209e-01 3.23419601e-01 -6.79905117e-01 -1.41202867e-01
9.47418213e-01 4.95662242e-01 -1.64812773e-01 1.66290119e-01
2.36181349e-01 -4.02641773e-01 -2.66642541e-01 7.38513246e-02
1.11487222e+00 8.06927502e-01 1.87326118e-01 -4.46065009e-01
3.02215457e-01 2.19484091e-01 6.56923413e-01 9.06641245e-01
1.84053406e-01 6.48937643e-01 -1.92128614e-01 -1.04549944e+00
-9.49873030e-01 -1.29886901e+00 5.84399924e-02 1.03627098e+00
8.85551095e-01 -4.08888668e-01 -2.03917459e-01 -5.48927367e-01
5.14006257e-01 4.59127605e-01 -4.87177014e-01 2.55624712e-01
-6.68431401e-01 -2.73753107e-01 1.48883119e-01 5.65860689e-01
5.38240254e-01 -7.86744893e-01 -6.87809646e-01 2.87436694e-01
-1.38483256e-01 -1.22394741e+00 -5.89680731e-01 3.67857099e-01
-6.06456220e-01 -1.04027152e+00 -1.45492375e-01 -9.15460408e-01
9.68126297e-01 5.75167000e-01 1.40157712e+00 2.07494423e-01
-2.89575487e-01 4.05142248e-01 -4.27694134e-02 -9.31125507e-02
2.00359017e-01 3.66445929e-01 1.51454747e-01 -1.55420393e-01
2.47309178e-01 -6.88698411e-01 -7.56640851e-01 4.98966545e-01
-6.49603531e-02 1.21430844e-01 3.42105865e-01 6.04558170e-01
1.14446723e+00 -5.91633856e-01 -2.68773943e-01 -5.75410366e-01
1.52880922e-01 -3.59035552e-01 -1.16961229e+00 4.64610979e-02
-2.26939484e-01 2.73487270e-01 3.87820274e-01 -8.01363364e-02
-7.64478683e-01 4.50803876e-01 -1.07499048e-01 -5.79439938e-01
-2.79810429e-01 2.05840826e-01 1.16853416e-02 -6.85047030e-01
3.52530211e-01 -8.89121294e-02 -2.49721959e-01 -4.86477852e-01
4.61370379e-01 1.65983737e-01 8.94315541e-01 -9.81207371e-01
9.54698741e-01 7.02154934e-01 3.64012122e-01 -4.57531601e-01
-9.32431102e-01 -6.45415306e-01 -8.62901092e-01 -1.66887805e-01
7.08426297e-01 -9.88188565e-01 -9.95127618e-01 3.66780698e-01
-1.25550890e+00 -6.84000850e-01 -9.58863571e-02 6.77360833e-01
-8.96335483e-01 2.75438577e-02 -3.45993251e-01 -4.24393505e-01
-1.64464653e-01 -1.09667039e+00 1.67915535e+00 9.94916819e-03
-1.10712513e-01 -1.01793182e+00 2.28897318e-01 8.78570303e-02
2.75017112e-01 4.62661177e-01 3.21546078e-01 -1.41023427e-01
-1.22355187e+00 -8.79845917e-02 -3.24205577e-01 -2.54218042e-01
-1.07867047e-01 -2.11118996e-01 -8.31113636e-01 -4.41693932e-01
-4.98431087e-01 -2.84613520e-01 9.18367267e-01 3.23113710e-01
1.26741016e+00 -2.08041623e-01 -5.75043976e-01 1.29809272e+00
1.61362851e+00 -3.56854141e-01 2.36477569e-01 5.17209828e-01
8.64203453e-01 4.68945831e-01 7.17308104e-01 3.67837042e-01
7.91856289e-01 7.06858158e-01 1.13874829e+00 -1.26195878e-01
-2.06003599e-02 -5.96135616e-01 -7.87533522e-02 5.51505685e-01
-1.17550313e-01 -3.24683785e-01 -9.79018450e-01 6.52785897e-01
-2.30062127e+00 -6.47176564e-01 -8.70297998e-02 2.28370285e+00
9.79831666e-02 -2.23409571e-02 -3.93619508e-01 -7.14005709e-01
8.11757088e-01 3.08841556e-01 -6.15597308e-01 -1.70226574e-01
1.70132853e-02 1.66908335e-02 1.05751157e+00 1.01097536e+00
-1.40196371e+00 1.05932248e+00 5.28885031e+00 3.72926205e-01
-9.55167770e-01 5.70874661e-02 3.01711291e-01 -2.29604885e-01
-1.86407700e-01 1.99364901e-01 -7.49115884e-01 4.76139277e-04
4.00963426e-01 2.76547372e-01 6.65028870e-01 9.32316482e-01
-6.13081120e-02 -2.37030257e-02 -1.25410283e+00 1.04164720e+00
-1.06531903e-01 -1.72506845e+00 -3.64879131e-01 1.25602186e-01
6.92819536e-01 9.27662373e-01 -1.03510402e-01 5.18374853e-02
6.67893648e-01 -6.42158866e-01 1.04772139e+00 5.42643666e-01
5.61121821e-01 -7.13374734e-01 6.92176938e-01 2.26606444e-01
-1.40322137e+00 1.36672944e-01 -6.13367856e-01 -2.05768362e-01
2.25518793e-01 3.95065457e-01 -9.31640565e-01 6.91648185e-01
8.80610108e-01 8.38894606e-01 -4.56938416e-01 1.39305401e+00
-3.56900424e-01 -8.73471126e-02 -7.36003637e-01 -4.02361825e-02
4.86447632e-01 -9.92016494e-02 5.91651320e-01 1.10856497e+00
5.04588008e-01 -1.47630125e-01 6.02295756e-01 8.69091570e-01
-1.55982375e-01 -3.62037063e-01 -6.22544944e-01 6.11684620e-01
6.46291316e-01 1.21205997e+00 -7.59532750e-01 -5.61144799e-02
-3.09141546e-01 1.17980111e+00 6.89752996e-01 2.45926529e-01
-7.75401235e-01 -3.75642866e-01 9.23940241e-01 -1.81228146e-02
4.09379154e-01 -5.83729565e-01 -2.48896793e-01 -1.22621489e+00
1.83073133e-01 -1.69526622e-01 1.89382404e-01 -9.11500931e-01
-1.24854791e+00 6.47036254e-01 -9.15122926e-02 -1.22256601e+00
3.98820937e-02 -6.97303772e-01 -4.67152417e-01 7.04283655e-01
-1.54589045e+00 -1.31972349e+00 -8.01874399e-01 5.40152967e-01
3.11060518e-01 1.72604591e-01 8.01344037e-01 8.40789303e-02
-1.61983937e-01 1.92065224e-01 -7.66475499e-02 1.20957561e-01
5.61981320e-01 -1.32184160e+00 1.25358033e+00 8.36303294e-01
3.04658920e-01 5.38157105e-01 6.28163755e-01 -6.29966736e-01
-1.63617003e+00 -1.33568120e+00 9.93502498e-01 -5.00463605e-01
6.16697788e-01 -7.74130642e-01 -4.76040900e-01 9.35705185e-01
2.05856025e-01 5.44983029e-01 1.80638984e-01 2.59283185e-01
-4.52963859e-01 -1.70674920e-02 -1.17916358e+00 5.20181060e-01
1.79449570e+00 -4.51603055e-01 -4.43281591e-01 7.62118459e-01
8.95787656e-01 -1.27915204e+00 -7.48599946e-01 3.16444367e-01
3.80239815e-01 -9.58949983e-01 1.17292404e+00 -3.12515914e-01
-7.94297159e-02 -5.44177055e-01 -5.97838521e-01 -1.22844255e+00
-4.29257214e-01 -9.03217256e-01 -6.11691512e-02 5.94452739e-01
4.12255138e-01 -7.86358535e-01 1.04112673e+00 4.76936311e-01
-6.34347856e-01 -8.03930342e-01 -1.14063895e+00 -7.77642310e-01
-5.58124363e-01 -5.29223740e-01 9.45947647e-01 8.16493571e-01
-4.96090561e-01 1.66499969e-02 -3.65103245e-01 8.74018371e-01
1.02155375e+00 4.78682578e-01 1.03737450e+00 -1.07493448e+00
-2.70708025e-01 -1.10478655e-01 -8.18187773e-01 -1.42730796e+00
4.31680799e-01 -1.04435325e+00 4.15704519e-01 -1.79639351e+00
-3.28906745e-01 -7.31245458e-01 -7.54831284e-02 6.97403312e-01
3.19966942e-01 4.83495831e-01 7.09136650e-02 2.01779343e-02
-9.06525850e-01 6.63436770e-01 1.01983142e+00 -1.91514328e-01
-2.08798975e-01 -4.19943258e-02 -1.22545548e-01 7.77219474e-01
6.43639445e-01 -4.84005004e-01 -1.70349330e-01 -1.14293933e+00
3.73488516e-01 -2.75459588e-02 7.70029902e-01 -1.17924261e+00
5.71198821e-01 -2.45843723e-01 4.46088284e-01 -9.53176558e-01
7.15894043e-01 -9.08911407e-01 1.28226101e-01 3.01114351e-01
-2.06126794e-01 4.34814930e-01 2.28944317e-01 6.11131191e-01
3.35411318e-02 1.81986183e-01 3.28812271e-01 -2.28070810e-01
-1.17450309e+00 8.12347174e-01 2.27985531e-01 7.27942865e-03
6.77223325e-01 -1.45370260e-01 -4.65956897e-01 -2.03582212e-01
-6.81584895e-01 5.35723686e-01 8.08605075e-01 5.86624146e-01
9.19932961e-01 -1.48554325e+00 -5.12901366e-01 4.02353019e-01
2.25361988e-01 6.04558706e-01 6.81543499e-02 7.55714059e-01
-1.00286222e+00 2.90018409e-01 -1.07392423e-01 -9.25329566e-01
-8.12226832e-01 3.73809516e-01 4.59409684e-01 1.81457773e-01
-8.40684891e-01 1.17325330e+00 -1.60556342e-02 -1.10807669e+00
4.34599102e-01 -4.85376030e-01 6.95338130e-01 -5.75945318e-01
1.80585429e-01 2.43479490e-01 1.95378944e-01 -7.69799113e-01
-7.82622874e-01 7.64776587e-01 1.88102096e-01 7.71833360e-02
1.55521429e+00 -1.95771664e-01 -3.20604235e-01 3.07711381e-02
1.27756476e+00 -1.29860463e-02 -1.60116339e+00 -3.83998811e-01
-1.56384960e-01 -8.40401232e-01 1.96460947e-01 -5.15082896e-01
-8.58639836e-01 5.21972716e-01 4.42769915e-01 -4.46830750e-01
7.09352493e-01 6.23106062e-02 7.37135291e-01 9.40668344e-01
9.59235728e-01 -8.53365183e-01 -1.43359184e-01 6.99991405e-01
9.30989146e-01 -1.32927287e+00 6.47148490e-02 -4.79722351e-01
-2.87627667e-01 1.03968120e+00 6.26701236e-01 -7.01177657e-01
5.01383245e-01 2.47143149e-01 -1.24131385e-02 -2.47019514e-01
-4.58994299e-01 -1.21416479e-01 4.40498561e-01 7.50575066e-01
2.83847610e-03 4.91035469e-02 4.44624960e-01 -2.55666703e-01
-4.20133501e-01 -2.78795600e-01 1.38053283e-01 9.29202378e-01
-3.97932619e-01 -9.29695129e-01 -4.59445745e-01 1.52580976e-01
9.52416137e-02 -5.75687811e-02 -3.01957965e-01 8.13967288e-01
2.60090619e-01 6.03549778e-01 4.14968252e-01 -4.73012835e-01
3.07606518e-01 -4.25974578e-01 7.78563559e-01 -6.08707011e-01
-4.50107247e-01 -2.32800916e-01 2.52254695e-01 -1.26204479e+00
-4.44941044e-01 -6.06153488e-01 -1.27618062e+00 -3.76612753e-01
-8.69326144e-02 3.33322138e-02 8.44423771e-01 7.43043482e-01
8.08284402e-01 1.87013328e-01 3.68115902e-01 -1.63347971e+00
-1.52238324e-01 -4.35373247e-01 -1.50933355e-01 1.66110128e-01
6.37092233e-01 -7.79898882e-01 -6.39242753e-02 -4.22995120e-01] | [7.628195762634277, -2.1943531036376953] |
b9c64ccc-93c5-4bcd-8349-1890c9b6f733 | enhanced-masked-image-modeling-for-analysis | 2306.10623 | null | https://arxiv.org/abs/2306.10623v1 | https://arxiv.org/pdf/2306.10623v1.pdf | Enhanced Masked Image Modeling for Analysis of Dental Panoramic Radiographs | The computer-assisted radiologic informative report has received increasing research attention to facilitate diagnosis and treatment planning for dental care providers. However, manual interpretation of dental images is limited, expensive, and time-consuming. Another barrier in dental imaging is the limited number of available images for training, which is a challenge in the era of deep learning. This study proposes a novel self-distillation (SD) enhanced self-supervised learning on top of the masked image modeling (SimMIM) Transformer, called SD-SimMIM, to improve the outcome with a limited number of dental radiographs. In addition to the prediction loss on masked patches, SD-SimMIM computes the self-distillation loss on the visible patches. We apply SD-SimMIM on dental panoramic X-rays for teeth numbering, detection of dental restorations and orthodontic appliances, and instance segmentation tasks. Our results show that SD-SimMIM outperforms other self-supervised learning methods. Furthermore, we augment and improve the annotation of an existing dataset of panoramic X-rays. | ['Longin Jan Latecki', 'Amani Almalki'] | 2023-06-18 | null | null | null | null | ['self-supervised-learning'] | ['computer-vision'] | [ 4.78504241e-01 8.04771483e-01 -6.15628719e-01 -7.76002407e-01
-1.46220684e+00 3.28856289e-01 1.08100593e-01 2.30568960e-01
-4.03615981e-01 3.21095288e-01 1.18596099e-01 -2.15375468e-01
2.95454320e-02 -7.59843290e-01 -7.51732349e-01 -7.48706043e-01
1.24799788e-01 6.49496913e-01 3.31992358e-01 2.02088639e-01
-9.20310915e-02 4.04455721e-01 -1.68127346e+00 5.19203305e-01
7.35657990e-01 8.98375332e-01 7.31692016e-01 3.12247187e-01
-1.87229030e-02 6.76335931e-01 -2.97283888e-01 -3.99886519e-01
1.03772745e-01 -2.72097021e-01 -6.05968714e-01 5.51448762e-01
8.03508103e-01 -7.29421198e-01 -1.47380725e-01 9.44131136e-01
6.88887775e-01 -2.85193533e-01 5.42826056e-01 -8.02096069e-01
-6.02292836e-01 7.19923913e-01 -8.18793714e-01 3.93280953e-01
-1.77910060e-01 1.42953947e-01 7.53379524e-01 -7.84345984e-01
7.16144443e-01 1.27662241e+00 3.82841527e-01 5.74789166e-01
-1.15350628e+00 -5.93215108e-01 -5.31764515e-02 3.70205522e-01
-9.13178205e-01 -2.53551692e-01 6.95621789e-01 -3.19403231e-01
4.40431982e-01 1.33870035e-01 5.59004664e-01 7.36781657e-01
1.72608852e-01 1.30668008e+00 1.36058092e+00 -7.15368867e-01
4.87318486e-01 -4.36381213e-02 3.07600975e-01 7.47428954e-01
-5.20918481e-02 2.08955884e-01 -1.48574397e-01 -6.43471554e-02
7.82541513e-01 1.79479375e-01 -2.43213940e-02 1.16978064e-01
-5.71994424e-01 9.45786417e-01 6.83856368e-01 1.78715557e-01
-5.14820814e-01 -3.46251667e-01 1.23858735e-01 -5.90967575e-05
8.64458919e-01 1.35216964e-02 -3.20240825e-01 5.30862868e-01
-1.10637355e+00 -1.27700046e-01 2.30495900e-01 2.19386503e-01
7.01703012e-01 -1.72322661e-01 1.71817228e-01 1.48322845e+00
4.66015190e-01 5.65743327e-01 8.91119242e-01 -8.02730322e-01
-9.93098784e-03 7.63727665e-01 -9.06933308e-01 -2.40252763e-01
-5.89829683e-01 -3.79160404e-01 -7.02056825e-01 5.58910966e-01
2.14904249e-01 2.94804007e-01 -1.58975995e+00 1.34534132e+00
8.68730128e-01 4.21538427e-02 -3.00703108e-01 8.61911178e-01
9.85048950e-01 3.19424689e-01 2.45303854e-01 -2.69963086e-01
1.61283767e+00 -1.08010209e+00 -6.53818130e-01 -5.58821261e-01
6.28020644e-01 -9.13610816e-01 1.24320006e+00 2.95878470e-01
-1.14347970e+00 -3.54808569e-01 -9.90319967e-01 -3.37286383e-01
2.58020926e-02 3.86170745e-01 6.96927965e-01 5.22045970e-01
-8.25342357e-01 4.78195965e-01 -1.28778648e+00 1.42056450e-01
1.00795674e+00 5.19565642e-01 -4.44774568e-01 -4.59690601e-01
-4.69083965e-01 7.41933286e-01 -1.19201712e-01 -4.80792895e-02
-4.92005587e-01 -8.72481883e-01 -9.58312273e-01 -2.45475888e-01
3.00038040e-01 -3.36774945e-01 1.55968392e+00 -6.88539743e-01
-1.58857620e+00 1.40956295e+00 1.86436966e-01 -3.64913553e-01
5.60249209e-01 -3.31883699e-01 -2.70605475e-01 4.27319646e-01
5.74312091e-01 9.75577414e-01 7.07216084e-01 -1.34316647e+00
-5.21002471e-01 -1.07555938e+00 -3.34655225e-01 -3.29767130e-02
6.96352199e-02 -1.22691676e-01 -7.14066327e-01 -8.74828041e-01
7.17851222e-01 -1.11293721e+00 -5.33338189e-01 4.03661758e-01
-8.16676497e-01 -3.05941552e-01 6.00350320e-01 -8.50971043e-01
7.48273134e-01 -2.35186863e+00 -4.51589406e-01 1.56279206e-01
1.88795086e-02 9.36856121e-02 1.41731668e-02 -2.88854957e-01
-2.71451682e-01 -4.00418937e-01 -6.29862309e-01 -7.88435876e-01
-4.17685002e-01 5.96048236e-01 7.30865672e-02 4.82423574e-01
5.78250457e-03 5.85062683e-01 -5.13298512e-01 -1.01635540e+00
5.40660739e-01 6.32017672e-01 -6.70996249e-01 9.83760580e-02
-3.46994966e-01 6.01752639e-01 -3.84258479e-01 1.07821536e+00
9.22583580e-01 -6.44613147e-01 5.92945442e-02 -3.76311004e-01
2.34262362e-01 3.71224046e-01 -9.87681270e-01 1.69213092e+00
-8.74539137e-01 2.83351243e-01 1.76203832e-01 -8.62704098e-01
3.54257196e-01 1.47677436e-01 7.73003519e-01 -1.17734802e+00
9.92488340e-02 3.96887571e-01 -1.70803219e-01 -7.09012687e-01
-2.16637760e-01 -2.76434213e-01 4.57443386e-01 8.33603978e-01
2.37331484e-02 -1.75535157e-01 -2.09689257e-03 4.08772305e-02
7.50318348e-01 7.88067374e-03 3.39009136e-01 1.41917542e-01
2.39949778e-01 -1.26275063e-01 5.11644185e-01 3.17694873e-01
4.61129919e-02 1.02087975e+00 -6.73788860e-02 -2.77073413e-01
-8.73916626e-01 -1.14545977e+00 -7.54024684e-01 9.42259192e-01
-1.34820521e-01 1.51142776e-01 -9.82085049e-01 -7.36405134e-01
-6.82723820e-02 5.33411622e-01 -6.00566864e-01 3.85291874e-02
-9.96762574e-01 -1.20108831e+00 -7.44673982e-02 5.52817643e-01
3.19384277e-01 -1.15942991e+00 -4.14811313e-01 2.98231900e-01
3.74786966e-02 -1.12073350e+00 -3.87914896e-01 3.04749340e-01
-1.29941344e+00 -1.21348464e+00 -9.33940649e-01 -1.17301691e+00
8.82004142e-01 5.47527596e-02 8.99383962e-01 1.34084299e-01
-8.92017961e-01 1.73731044e-01 3.98946702e-02 -4.28614765e-01
-6.24600589e-01 -6.15449809e-02 -4.91702825e-01 -1.59557462e-01
3.58823895e-01 -6.24830604e-01 -1.12708306e+00 2.24898338e-01
-1.16367042e+00 4.44725975e-02 1.01999724e+00 8.35759103e-01
1.14074957e+00 -2.12467447e-01 3.80968809e-01 -1.23063910e+00
8.74695703e-02 -2.95799017e-01 -4.06071842e-01 2.17133865e-01
-8.06664586e-01 1.88287601e-01 -8.74483511e-02 -3.77820849e-01
-1.25212181e+00 4.86977488e-01 -6.79655552e-01 1.03257045e-01
-1.81837186e-01 1.74623504e-01 -3.64400819e-02 2.46575773e-02
5.02434611e-01 -1.22852176e-01 2.48854280e-01 -8.57776821e-01
3.92723262e-01 8.08845043e-01 7.53216863e-01 -1.70855507e-01
2.78254181e-01 9.53917623e-01 -6.48625121e-02 -9.85194266e-01
-1.32377982e+00 -4.65498954e-01 -4.54508364e-01 -1.38438344e-01
1.04192233e+00 -9.94594812e-01 -3.00363183e-01 5.27796388e-01
-7.35134661e-01 -1.72950640e-01 -4.89336103e-01 8.53745997e-01
-3.70211929e-01 5.15172958e-01 -1.06544018e+00 -5.64640105e-01
-5.23887873e-01 -1.67519295e+00 1.59951556e+00 1.40363097e-01
4.14427221e-02 -5.76068103e-01 1.11194901e-01 1.16334808e+00
-1.24952689e-01 -5.95756993e-02 1.15638757e+00 -4.03748751e-01
-6.99579358e-01 -1.60925716e-01 -1.25348866e-01 7.69398689e-01
2.51048625e-01 -4.25262630e-01 -1.20553827e+00 1.41002640e-01
4.02393818e-01 -3.78766567e-01 1.09290826e+00 1.07772028e+00
1.38145208e+00 -1.58497885e-01 -3.76195788e-01 5.07601917e-01
1.17709994e+00 3.05434823e-01 5.94014883e-01 6.46628022e-01
5.65016091e-01 7.86097229e-01 5.19657075e-01 4.63917851e-02
6.01399064e-01 6.84299469e-01 5.11803567e-01 -5.96120596e-01
-6.67457342e-01 -4.05546129e-02 -3.57781611e-02 8.22390974e-01
2.11554959e-01 2.46533915e-01 -7.33984470e-01 5.12824476e-01
-1.33332658e+00 -3.82220000e-01 -1.70812421e-02 1.83675611e+00
1.16439986e+00 -1.19107448e-01 -3.45745273e-02 4.35195297e-01
8.24823022e-01 -2.50569880e-01 -9.24301386e-01 3.37803274e-01
-1.72987089e-01 8.63198340e-01 4.31584299e-01 5.63848257e-01
-1.11572731e+00 5.63082337e-01 5.82138538e+00 8.41598570e-01
-1.30150604e+00 5.86377323e-01 1.10927689e+00 -9.95879397e-02
5.15503995e-03 -3.56533706e-01 -7.48625875e-01 5.17406642e-01
4.82320756e-01 6.90169454e-01 -3.62931430e-01 1.35678351e+00
3.88212167e-02 -4.65374440e-01 -1.13287508e+00 9.14585650e-01
1.53899476e-01 -1.27922237e+00 -1.86243072e-01 2.69195944e-01
7.00392425e-01 2.10497618e-01 3.21074516e-01 -5.28261699e-02
3.26067358e-01 -7.43339956e-01 2.64675289e-01 -1.36994034e-01
5.63681781e-01 -3.27152699e-01 5.78032136e-01 -2.72858869e-02
-5.21368921e-01 4.69410196e-02 -3.94574180e-02 6.32701516e-01
4.79523659e-01 1.09057713e+00 -1.07242870e+00 1.32191572e-02
6.83389068e-01 1.68786213e-01 -4.03189003e-01 1.08452380e+00
-3.73932451e-01 6.29547238e-01 -6.00319386e-01 6.60624862e-01
1.62216499e-02 2.27617379e-02 9.86179709e-02 5.81289470e-01
9.84570384e-02 2.15205133e-01 1.85278580e-01 4.49333817e-01
-1.12117767e-01 3.84809375e-01 1.23489611e-01 6.67996407e-01
1.90460876e-01 9.00833964e-01 -9.26762640e-01 -4.22962338e-01
-4.19193506e-01 7.40626156e-01 -1.61654446e-02 -9.11620036e-02
-2.92445511e-01 6.71592832e-01 1.28881171e-01 5.49417853e-01
1.18018791e-01 3.32326561e-01 -4.51820910e-01 -8.70051205e-01
1.34705007e-01 -9.02034879e-01 6.33113980e-01 -6.22840762e-01
-1.38103199e+00 6.34299636e-01 -4.60141748e-02 -1.03911781e+00
-2.89563417e-01 -5.38531244e-01 -3.54278505e-01 1.39653012e-01
-1.73119462e+00 -1.38823128e+00 -1.42353371e-01 4.35409486e-01
7.87811995e-01 -2.31908597e-02 6.03309393e-01 7.79703140e-01
-6.63717866e-01 4.70024616e-01 2.40996871e-02 -9.72126275e-02
6.70627356e-01 -1.22125793e+00 -3.29144150e-02 2.93609262e-01
2.04565480e-01 2.25408792e-01 4.35858220e-01 -6.43750131e-01
-7.66364396e-01 -1.02331364e+00 5.89368820e-01 1.03149556e-01
1.29964590e-01 1.08097002e-01 -9.29560363e-01 6.92840815e-01
-1.13017492e-01 2.00002402e-01 1.00499392e+00 -2.73486435e-01
-1.75722644e-01 -1.19061835e-01 -1.35741079e+00 3.78965229e-01
4.45599079e-01 -4.62432086e-01 -5.40142298e-01 9.59836602e-01
5.39273262e-01 -4.47643191e-01 -8.89617264e-01 6.13779068e-01
4.07178402e-01 -8.95150959e-01 1.32247829e+00 4.03188467e-02
6.00785732e-01 1.67396292e-01 -2.59056240e-01 -8.43738854e-01
4.20190722e-01 6.99369907e-02 1.61162138e-01 8.68662000e-01
3.97593290e-01 -4.35705423e-01 1.33450294e+00 4.37393248e-01
-4.74831223e-01 -9.65295434e-01 -1.21052384e+00 -4.04947042e-01
1.82779327e-01 -5.65694690e-01 3.00765634e-01 7.76060283e-01
-4.98674005e-01 2.69062281e-01 6.15383312e-02 2.80056715e-01
1.05139387e+00 2.92560935e-01 3.41003418e-01 -1.19220030e+00
-6.01520598e-01 -2.44353250e-01 -3.26177597e-01 -9.92100358e-01
-6.79295287e-02 -9.37302411e-01 -8.33257809e-02 -1.76870596e+00
2.43796095e-01 -5.80209553e-01 -1.18157968e-01 5.62951207e-01
-2.80391634e-01 5.95600307e-01 -2.20340818e-01 4.54986513e-01
7.86364675e-02 5.04633665e-01 1.65694726e+00 -4.20165479e-01
-2.01452702e-01 4.46746111e-01 -2.84689069e-01 1.09540260e+00
4.73087639e-01 -7.56948113e-01 -3.21054786e-01 -4.16985363e-01
-2.98639417e-01 1.10211775e-01 4.19174254e-01 -8.27216804e-01
1.25711843e-01 2.61285484e-01 2.01089695e-01 -1.06880426e+00
5.70996463e-01 -5.88318169e-01 -2.67465234e-01 8.10349226e-01
-3.05389702e-01 -2.71604002e-01 -8.52881595e-02 3.69931161e-01
-5.34746170e-01 -1.27909303e-01 1.06021404e+00 -4.96769816e-01
-3.81586432e-01 2.86743015e-01 5.23669682e-02 -2.31249422e-01
9.24242318e-01 -1.75370291e-01 -1.51247039e-01 4.15571071e-02
-1.22476518e+00 -1.00753821e-01 4.39127572e-02 1.37809873e-01
6.78338766e-01 -8.32262635e-01 -6.03020668e-01 2.68889248e-01
8.50802585e-02 6.51641428e-01 7.68645644e-01 7.62287498e-01
-6.31240368e-01 8.81803185e-02 1.67115498e-02 -1.05152166e+00
-1.48470783e+00 1.99785665e-01 1.02413103e-01 -3.92756999e-01
-1.13656330e+00 8.42951596e-01 5.98814487e-01 -4.53725427e-01
3.33136380e-01 -6.39071703e-01 -1.69517666e-01 -7.56694600e-02
5.57366908e-01 2.71932065e-01 6.00662708e-01 -6.01563632e-01
-1.91355079e-01 8.07812512e-01 -6.43486619e-01 -9.67775732e-02
1.79866958e+00 -5.76452650e-02 -1.19611628e-01 -2.60863453e-02
1.35668933e+00 -2.79013276e-01 -9.79285121e-01 -6.06861651e-01
-9.44599323e-03 -1.67421490e-01 7.37330496e-01 -7.18314230e-01
-1.67323995e+00 8.75802398e-01 1.37318146e+00 -3.77307683e-01
1.14686310e+00 5.89424253e-01 1.08724177e+00 -7.04098912e-03
9.38259289e-02 -1.01081085e+00 3.31450075e-01 -5.21822631e-01
6.87958479e-01 -1.75268555e+00 3.68982047e-01 -9.14711177e-01
-4.46378410e-01 9.12404060e-01 6.12344384e-01 4.42836322e-02
1.06699181e+00 3.88006389e-01 5.21592200e-01 -5.82149684e-01
-2.64259279e-01 -4.78839129e-02 2.88749248e-01 4.70891178e-01
2.92710483e-01 -4.52598324e-03 -2.84010947e-01 1.02765493e-01
-3.68286192e-01 1.13592772e-02 1.48711726e-01 1.09383428e+00
-2.74676085e-01 -1.33222473e+00 -5.94614208e-01 4.76647884e-01
-7.94553459e-01 -1.29531369e-01 1.33318588e-01 7.95986533e-01
2.18867913e-01 8.26771021e-01 1.71300560e-01 1.55930459e-01
1.40979230e-01 -3.93550396e-01 4.71947968e-01 -9.37213540e-01
-3.34427685e-01 5.09281814e-01 -2.29629889e-01 -4.49722975e-01
-6.47990167e-01 -5.96100509e-01 -1.38544691e+00 4.16142374e-01
-5.55857122e-01 -3.85645956e-01 1.16009200e+00 1.07837737e+00
6.51041344e-02 6.70258224e-01 7.19200253e-01 -5.49304247e-01
-5.80769897e-01 -9.82551038e-01 -4.91773605e-01 3.78515691e-01
5.08922756e-01 -6.53399646e-01 -4.14104134e-01 2.13743418e-01] | [13.761332511901855, -2.2351796627044678] |
8e0227a2-cd51-4887-9298-12b9c64755b2 | investigating-the-challenges-of-temporal | null | null | https://aclanthology.org/W18-5607 | https://aclanthology.org/W18-5607.pdf | Investigating the Challenges of Temporal Relation Extraction from Clinical Text | Temporal reasoning remains as an unsolved task for Natural Language Processing (NLP), particularly demonstrated in the clinical domain. The complexity of temporal representation in language is evident as results of the 2016 Clinical TempEval challenge indicate: the current state-of-the-art systems perform well in solving mention-identification tasks of event and time expressions but poorly in temporal relation extraction, showing a gap of around 0.25 point below human performance. We explore to adapt the tree-based LSTM-RNN model proposed by Miwa and Bansal (2016) to temporal relation extraction from clinical text, obtaining a five point improvement over the best 2016 Clinical TempEval system and two points over the state-of-the-art. We deliver a deep analysis of the results and discuss the next step towards human-like temporal reasoning. | ['Diana Galvan', 'Naoaki Okazaki', 'Kentaro Inui', 'Koji Matsuda'] | 2018-10-01 | null | null | null | ws-2018-10 | ['temporal-relation-extraction', 'temporal-information-extraction'] | ['natural-language-processing', 'natural-language-processing'] | [ 1.31583855e-01 6.95154607e-01 -5.59787333e-01 -2.77173609e-01
-8.86029303e-01 -4.11439031e-01 6.95046544e-01 7.80973196e-01
-7.24085152e-01 7.29561388e-01 7.30734229e-01 -6.93427980e-01
-6.13533556e-01 -3.23549569e-01 -1.83710769e-01 -2.88370252e-01
-7.44105756e-01 7.02388227e-01 1.35149032e-01 -4.28699195e-01
-3.68717134e-01 3.81097108e-01 -5.88463426e-01 1.00231028e+00
2.25370288e-01 9.49385166e-01 -8.93587708e-01 7.21107006e-01
7.40093291e-02 1.69823384e+00 -2.95708597e-01 -3.75622302e-01
-2.02140898e-01 -2.71714985e-01 -1.46696186e+00 -7.32285082e-01
-2.93545216e-01 -1.23982787e-01 -8.37137997e-01 4.43689376e-01
3.65298063e-01 1.70917436e-01 2.31953710e-01 -8.41005147e-01
-2.52256960e-01 9.80616808e-01 -1.59179613e-01 8.81821334e-01
5.59485257e-01 2.85192549e-01 7.89698839e-01 -6.50622472e-02
1.11768508e+00 8.98989141e-01 9.91057992e-01 6.22076690e-01
-1.12930620e+00 -4.36168164e-01 2.33896151e-01 4.42090988e-01
-1.17950344e+00 -6.03861809e-01 9.22732279e-02 -4.36228245e-01
2.02349305e+00 1.47847645e-02 6.70843244e-01 1.33400321e+00
7.47330546e-01 9.00470793e-01 1.09587526e+00 -1.86940163e-01
6.02760762e-02 -6.44861698e-01 4.29307014e-01 7.19393849e-01
-4.33689922e-01 3.09578925e-01 -7.76237547e-01 -2.59479821e-01
5.11356711e-01 -2.85016388e-01 -7.90487006e-02 3.03559422e-01
-1.71369314e+00 5.58589339e-01 5.96906066e-01 8.53753746e-01
-7.80705035e-01 3.33841622e-01 1.17757821e+00 2.74512529e-01
4.84449118e-01 3.25970292e-01 -8.76422405e-01 -5.11473954e-01
-1.26302898e+00 2.45640904e-01 7.60594308e-01 5.53708196e-01
-6.53148293e-01 -2.83290714e-01 -6.61872745e-01 4.49366182e-01
1.35848284e-01 -2.76117176e-01 7.15532184e-01 -6.14652753e-01
4.77423936e-01 4.83313620e-01 -2.43543953e-01 -7.14428782e-01
-1.18810296e+00 -2.87857980e-01 -8.89084697e-01 -3.15257728e-01
6.42722666e-01 -6.40606955e-02 -1.07210469e+00 1.81080544e+00
9.57376063e-02 1.44192934e-01 2.48379111e-01 3.98530394e-01
9.49910164e-01 5.20448267e-01 6.43039227e-01 -7.68454850e-01
1.82824504e+00 -6.09192073e-01 -1.25552607e+00 -2.70299226e-01
1.14452028e+00 -3.47297996e-01 6.30903095e-02 5.58400273e-01
-1.27662694e+00 2.39863530e-01 -9.61593330e-01 -2.82526851e-01
-4.90737498e-01 -2.92475194e-01 1.10804129e+00 -5.32679856e-02
-1.08691633e+00 8.90134752e-01 -1.39143884e+00 -5.83379447e-01
4.88945425e-01 5.55905223e-01 -5.93297124e-01 3.35152626e-01
-1.82792389e+00 1.62032843e+00 5.83152711e-01 5.27205527e-01
-6.26858592e-01 -1.10573685e+00 -9.13720131e-01 -3.19046676e-01
3.58382195e-01 -7.97728956e-01 1.82908189e+00 6.23243898e-02
-1.19504023e+00 1.12369335e+00 -2.53802150e-01 -1.33881879e+00
8.74061525e-01 -1.09124675e-01 -9.14573550e-01 3.38944077e-01
-6.46969955e-03 4.04125869e-01 -6.94589466e-02 -1.50817469e-01
-4.69310701e-01 -1.97742984e-01 -1.74660742e-01 -3.11781228e-01
2.71353781e-01 5.51650286e-01 -2.58501053e-01 -5.99255085e-01
3.48581895e-02 -7.52290845e-01 -6.69775724e-01 4.94829565e-02
-3.80924076e-01 -6.29131973e-01 2.53259480e-01 -1.02908266e+00
1.74725556e+00 -1.83968556e+00 -1.32337734e-01 -1.48742676e-01
4.65367049e-01 1.66147441e-01 1.16748266e-01 7.19299853e-01
-8.26651692e-01 2.68053357e-02 -3.21481735e-01 -1.95891634e-01
-1.21835798e-01 2.47056946e-01 -3.27596813e-01 7.09398627e-01
4.07644451e-01 1.32019603e+00 -1.30695748e+00 -7.29412735e-01
-2.23737732e-02 4.35819089e-01 -3.41488309e-02 -2.24416569e-01
-3.49313468e-01 4.15148139e-01 -3.43560815e-01 5.44755995e-01
1.33970324e-02 -3.44218433e-01 4.67859268e-01 -1.89262658e-01
-1.07781142e-01 1.15856302e+00 -4.72756654e-01 1.98086226e+00
-2.33065814e-01 6.37387037e-01 -4.10517544e-01 -6.91941023e-01
3.25043797e-01 1.21864307e+00 1.07644260e+00 -8.92080843e-01
2.16701478e-01 1.12858452e-01 4.43395317e-01 -7.14786053e-01
1.33710027e-01 -7.43370473e-01 -1.98488310e-01 6.33782566e-01
-1.45689517e-01 3.21838111e-01 2.63053149e-01 3.54088247e-01
1.67874718e+00 1.80860624e-01 8.46399963e-01 -1.63698062e-01
4.01674271e-01 3.01032066e-01 7.18940258e-01 3.87126714e-01
-6.38813078e-01 1.12076417e-01 1.00568676e+00 -8.18420231e-01
-6.10789478e-01 -8.21270347e-01 -2.53430963e-01 6.77739322e-01
-8.45230222e-01 -6.78528309e-01 -1.65928423e-01 -9.49057221e-01
-2.95185924e-01 8.42102945e-01 -1.20482421e+00 -1.42607227e-01
-1.01029468e+00 -8.78114223e-01 1.17603588e+00 8.31339002e-01
8.73001441e-02 -1.26568472e+00 -8.97778749e-01 6.19766414e-01
-3.60214442e-01 -1.64271140e+00 -1.86520189e-01 5.17125309e-01
-1.08203661e+00 -1.07256377e+00 -4.29787725e-01 -4.51875985e-01
5.22621209e-03 -7.10377693e-01 1.20355928e+00 -3.43715996e-01
-2.78113216e-01 3.92850824e-02 -1.63460359e-01 -4.42210227e-01
-5.58230639e-01 -1.70527920e-02 9.29374695e-02 -6.51854277e-01
4.64545786e-01 -5.27652025e-01 -6.77314878e-01 -2.25622624e-01
-6.91351056e-01 6.58206493e-02 2.03422636e-01 7.72428691e-01
3.51260066e-01 -2.19425187e-02 4.91168767e-01 -8.47426355e-01
5.56629717e-01 -5.69597125e-01 -1.20956421e-01 2.07991838e-01
-7.65176833e-01 3.05677027e-01 3.21149975e-01 -3.27824324e-01
-8.92961144e-01 -3.71021926e-02 -1.00186594e-01 -1.54045597e-01
1.62289679e-01 9.24558580e-01 8.50176811e-01 4.58699554e-01
6.67556286e-01 -7.89302960e-02 -1.45488773e-02 -3.87423709e-02
4.00993019e-01 -4.82925251e-02 8.01281095e-01 -3.32045645e-01
1.64977804e-01 6.48485243e-01 2.12669820e-01 -1.19233586e-01
-1.12936592e+00 -3.86124074e-01 -7.23539829e-01 5.22350259e-02
1.14097583e+00 -7.99842894e-01 -1.25221205e+00 -1.96289942e-02
-1.43244910e+00 -6.77537382e-01 -4.63301063e-01 4.90677059e-01
-6.31921947e-01 8.50341320e-02 -1.24281335e+00 -7.07780302e-01
-6.69422150e-01 -7.67253280e-01 9.64535892e-01 -3.73460621e-01
-9.75710332e-01 -1.23425019e+00 4.00772095e-01 9.02409106e-02
2.04992458e-01 6.57416880e-01 1.18158233e+00 -7.99246907e-01
-1.03327505e-01 -2.10392967e-01 -1.73872858e-01 -6.44415021e-01
4.79053222e-02 -9.03768316e-02 -7.67486095e-01 1.34750500e-01
1.33101016e-01 -1.74968198e-01 9.50061500e-01 5.28326690e-01
5.29182196e-01 -3.38144690e-01 -6.97699487e-01 3.06070089e-01
8.66370857e-01 4.34676319e-01 6.29766762e-01 5.08356035e-01
2.70455569e-01 6.74277186e-01 6.17790699e-01 3.87723714e-01
5.54199874e-01 4.94919538e-01 1.73354730e-01 7.78907016e-02
2.47112345e-02 1.31548280e-02 1.07609950e-01 6.19888604e-01
-1.05344288e-01 1.48679852e-01 -1.58781874e+00 8.90266180e-01
-2.21427560e+00 -1.00262058e+00 -2.59758443e-01 1.65881741e+00
1.28745532e+00 3.60893518e-01 2.12434560e-01 2.88410485e-01
3.61911394e-02 2.51301497e-01 -3.59070152e-01 -7.08872080e-01
1.18821293e-01 4.17727143e-01 4.86906707e-01 2.61399567e-01
-1.25062811e+00 1.10929370e+00 6.95949411e+00 4.84522760e-01
-1.12460601e+00 2.57055089e-02 5.67528784e-01 -3.88268113e-01
2.62175769e-01 -1.26278177e-01 -2.81697989e-01 -1.02938771e-01
1.76755750e+00 -2.95462668e-01 1.91484034e-01 7.01368377e-02
7.08658695e-01 -3.01032551e-02 -1.62866676e+00 1.04373240e+00
-2.85988629e-01 -1.54265273e+00 -4.84694719e-01 -1.49454223e-02
2.93554008e-01 3.44984442e-01 -8.89803991e-02 4.02190983e-01
5.66034257e-01 -1.44999230e+00 4.27815139e-01 6.85529828e-01
4.85816717e-01 -3.54256928e-01 6.95169508e-01 1.65705815e-01
-1.14931643e+00 8.58954433e-03 4.46921259e-01 3.75105329e-02
7.78953850e-01 4.75658715e-01 -1.25400519e+00 8.58491957e-01
6.43124342e-01 1.07213151e+00 -3.30027103e-01 6.99113309e-01
-3.23428273e-01 5.62915683e-01 -3.20063800e-01 2.68667787e-01
6.17756307e-01 4.29637223e-01 4.48869228e-01 1.61496603e+00
-7.61001259e-02 5.98039806e-01 -1.77634075e-01 5.51206112e-01
-6.00552410e-02 -8.71255845e-02 -6.19169712e-01 -5.16649127e-01
-2.49795131e-02 1.01456833e+00 -8.42656076e-01 -5.71603656e-01
4.90364321e-02 7.15996444e-01 2.23062053e-01 2.96252575e-02
-9.67203736e-01 2.01642260e-01 4.18990910e-01 -2.98514992e-01
6.97802901e-02 -2.65024453e-01 -4.53580648e-01 -7.59544671e-01
-1.47268726e-02 -9.95747566e-01 1.37694824e+00 -6.74063265e-01
-1.22258723e+00 8.49779487e-01 7.31200129e-02 -1.20639169e+00
-6.70468688e-01 -5.83124161e-01 -2.50908464e-01 5.22946060e-01
-1.08624578e+00 -1.26511467e+00 3.70353401e-01 4.69337612e-01
3.40643585e-01 2.72864640e-01 1.06052756e+00 3.99835169e-01
-5.29107392e-01 4.34244812e-01 -5.85418582e-01 3.87540370e-01
7.08750188e-01 -1.06526911e+00 5.90774834e-01 6.48350537e-01
3.36606055e-03 9.87008452e-01 8.20150137e-01 -6.22685969e-01
-1.32816875e+00 -9.94231820e-01 1.69074106e+00 -8.12812030e-01
1.38892794e+00 9.90982503e-02 -8.01865458e-01 1.18816841e+00
5.01304388e-01 -1.31681543e-02 6.60007775e-01 2.74963439e-01
-5.22018433e-01 3.08363616e-01 -1.04102409e+00 7.25022972e-01
9.52552557e-01 -8.64621043e-01 -1.05635750e+00 7.11707473e-01
9.82950091e-01 -7.82134831e-01 -1.45441437e+00 7.01219559e-01
6.37072682e-01 -5.80332458e-01 8.31631184e-01 -1.24538302e+00
6.84180737e-01 -2.23929569e-01 2.20507249e-01 -8.35960090e-01
-1.33512124e-01 -1.15470314e+00 -3.90718818e-01 8.25530887e-01
6.32898211e-01 -4.47923094e-01 5.86002529e-01 7.86295176e-01
-1.22351788e-01 -1.10937011e+00 -1.27862906e+00 -5.87892771e-01
2.33008057e-01 -8.71175647e-01 1.97454225e-02 1.40050185e+00
9.74170208e-01 5.01649320e-01 -6.92556202e-02 -3.53739262e-02
3.32943462e-02 -3.30850899e-01 -1.51389822e-01 -8.21582675e-01
-1.98905781e-01 -6.54091537e-01 -3.10158670e-01 -3.34639341e-01
2.74800718e-01 -1.12443006e+00 -1.83873013e-01 -2.05360889e+00
5.45043834e-02 5.53780161e-02 -7.16053545e-01 1.26732814e+00
2.84276396e-01 -6.49914099e-03 5.19119166e-02 -1.90967973e-02
-5.64514279e-01 3.09409201e-02 9.95864034e-01 -3.97073597e-01
-3.71164113e-01 -2.48300269e-01 -4.80618715e-01 5.65144300e-01
4.19820875e-01 -6.68720484e-01 -1.78176105e-01 -1.96338221e-01
5.66836357e-01 7.88389683e-01 1.44948021e-01 -5.96149027e-01
6.93437099e-01 -1.14161307e-02 9.28143114e-02 -8.10232043e-01
3.14049363e-01 -5.22438407e-01 9.18617100e-02 1.08766687e+00
-7.82465935e-01 3.86254907e-01 9.04568851e-01 3.91150534e-01
-2.15663835e-01 3.27918112e-01 5.46737313e-01 -2.52711356e-01
-3.84986818e-01 2.32334346e-01 -7.49074399e-01 5.97758479e-02
8.43451262e-01 1.83749795e-01 -6.12865150e-01 -1.76126719e-01
-1.28718281e+00 5.03600419e-01 -4.89039958e-01 2.23679960e-01
3.21734965e-01 -1.00549603e+00 -7.28114545e-01 -6.10277176e-01
6.19907938e-02 -1.40197456e-01 3.99465978e-01 1.67921388e+00
-4.38180655e-01 1.08303344e+00 1.35983661e-01 -3.69377881e-01
-1.21846807e+00 9.44992721e-01 5.16443670e-01 -1.21431768e+00
-1.10899591e+00 6.92981541e-01 -1.91654503e-01 -1.81519017e-02
1.38179541e-01 -9.61531043e-01 -2.63778687e-01 3.51164699e-01
4.81366038e-01 1.00613348e-02 3.96466315e-01 -1.57006815e-01
-9.85754609e-01 6.95051923e-02 -4.13481027e-01 -4.34973180e-01
1.60436845e+00 4.86570537e-01 -4.93733138e-01 6.72300756e-01
1.01859355e+00 -4.95230675e-01 -3.79485756e-01 -3.08327049e-01
7.61663496e-01 6.31410420e-01 -8.06591436e-02 -1.28720033e+00
-5.70917010e-01 5.92335641e-01 3.67524415e-01 8.67292751e-03
9.32219386e-01 1.27482548e-01 9.10861015e-01 4.52104926e-01
2.14102462e-01 -8.94179642e-01 -4.43955272e-01 8.79293025e-01
8.16972256e-01 -8.92312706e-01 3.30804884e-01 -9.01583433e-02
-7.20137000e-01 1.16759205e+00 4.13437523e-02 9.67579782e-02
7.78107047e-01 4.18907315e-01 2.24984497e-01 -6.68728173e-01
-1.47135162e+00 -7.68109560e-02 3.38205069e-01 3.91882733e-02
1.00655401e+00 -3.00666993e-03 -5.41394293e-01 6.05948865e-01
-1.86514586e-01 6.35057271e-01 3.12298477e-01 1.05524409e+00
5.45922279e-01 -9.86302674e-01 3.18236314e-02 3.26022297e-01
-8.80932510e-01 -4.98100907e-01 -2.40785182e-01 9.08927441e-01
-2.63923049e-01 9.66100276e-01 -7.04940781e-02 -6.52790219e-02
5.06436765e-01 4.84960258e-01 5.07180095e-01 -3.70113611e-01
-1.26619184e+00 2.51462534e-02 8.41587305e-01 -1.17439640e+00
-5.20811796e-01 -8.12098384e-01 -1.83341706e+00 -1.00374054e-02
3.61208498e-01 3.77984233e-02 2.80323684e-01 1.23606706e+00
1.69448286e-01 1.28603780e+00 -3.06494355e-01 -1.15021609e-01
-2.32931688e-01 -8.79792690e-01 -4.84650247e-02 5.30568399e-02
6.47238076e-01 -3.35818827e-01 8.97830445e-03 4.32493575e-02] | [8.499953269958496, 9.015286445617676] |
a49eec92-3206-446c-9e82-97adda90db09 | integrating-prior-knowledge-in-post-hoc | 2204.11634 | null | https://arxiv.org/abs/2204.11634v1 | https://arxiv.org/pdf/2204.11634v1.pdf | Integrating Prior Knowledge in Post-hoc Explanations | In the field of eXplainable Artificial Intelligence (XAI), post-hoc interpretability methods aim at explaining to a user the predictions of a trained decision model. Integrating prior knowledge into such interpretability methods aims at improving the explanation understandability and allowing for personalised explanations adapted to each user. In this paper, we propose to define a cost function that explicitly integrates prior knowledge into the interpretability objectives: we present a general framework for the optimization problem of post-hoc interpretability methods, and show that user knowledge can thus be integrated to any method by adding a compatibility term in the cost function. We instantiate the proposed formalization in the case of counterfactual explanations and propose a new interpretability method called Knowledge Integration in Counterfactual Explanation (KICE) to optimize it. The paper performs an experimental study on several benchmark data sets to characterize the counterfactual instances generated by KICE, as compared to reference methods. | ['Marcin Detyniecki', 'Christophe Marsala', 'Marie-Jeanne Lesot', 'Thibault Laugel', 'Adulam Jeyasothy'] | 2022-04-25 | null | null | null | null | ['counterfactual-explanation'] | ['miscellaneous'] | [ 5.95682681e-01 1.15657806e+00 -1.67453393e-01 -6.98358297e-01
-1.51625663e-01 -4.00152206e-01 8.06047857e-01 2.76024908e-01
-2.03447819e-01 9.91283596e-01 2.81824857e-01 -6.01215720e-01
-8.47590625e-01 -5.81123650e-01 -7.15175748e-01 -2.69198507e-01
5.10274880e-02 1.02241981e+00 -3.83413374e-01 7.46595114e-02
4.42317128e-01 2.75950909e-01 -1.78433645e+00 3.90703231e-01
1.54206729e+00 7.45517254e-01 1.33377435e-02 4.94793773e-01
2.39288993e-02 4.97715980e-01 -6.54600084e-01 -5.26038826e-01
2.25839972e-01 -6.81827426e-01 -1.42245495e+00 2.18190134e-01
1.54992357e-01 8.27714801e-02 5.75223267e-01 7.41283357e-01
-1.46288440e-01 2.79087037e-01 9.70548213e-01 -1.75366700e+00
-7.69370854e-01 9.30744886e-01 1.74901605e-01 -7.96237960e-02
4.39362943e-01 2.35061143e-02 1.18534315e+00 -4.29950118e-01
5.81070960e-01 1.45838213e+00 2.89817542e-01 7.54093170e-01
-1.62844586e+00 -7.54326731e-02 4.90868628e-01 6.24716163e-01
-6.59805298e-01 -3.65986079e-02 6.39376223e-01 -4.36816663e-01
8.91399741e-01 8.62446368e-01 6.54432118e-01 9.60614026e-01
-8.90996903e-02 7.69384325e-01 1.31946480e+00 -7.87269294e-01
6.73302770e-01 3.87174666e-01 6.26948476e-01 4.73423183e-01
8.79087269e-01 3.61541957e-01 -2.04552844e-01 -1.94385201e-01
3.35990638e-01 1.30857364e-03 -5.67719460e-01 -5.74990749e-01
-1.02757907e+00 1.09320831e+00 4.93064910e-01 2.80251712e-01
-4.82615650e-01 1.24169827e-01 1.81673571e-01 3.47958446e-01
4.94340658e-01 1.24118280e+00 -7.68052042e-01 2.33918965e-01
-2.67236620e-01 4.17483866e-01 1.11031103e+00 6.74660444e-01
8.31245720e-01 -2.61214912e-01 -3.44940782e-01 2.76556462e-01
4.70764488e-01 1.79582953e-01 5.91287494e-01 -1.08031547e+00
3.97792906e-01 1.02525067e+00 3.93614203e-01 -4.98751730e-01
-3.67746115e-01 -4.98106897e-01 -5.08867323e-01 6.05253696e-01
2.40600541e-01 -1.79669157e-01 -8.30891907e-01 1.68338180e+00
3.62841338e-01 3.74364406e-02 3.10090780e-01 9.77731884e-01
3.97674441e-01 4.05204982e-01 2.30999053e-01 -4.16907907e-01
1.17346752e+00 -9.93805408e-01 -9.12028015e-01 -1.41696066e-01
5.99443316e-01 -2.26768896e-01 1.33289778e+00 3.62622648e-01
-8.46533537e-01 -4.11959887e-01 -1.26281691e+00 8.46884325e-02
-4.17471111e-01 -1.90520026e-02 1.01333058e+00 7.37503767e-01
-7.01848924e-01 7.41080165e-01 -5.13061702e-01 -1.77949652e-01
5.65028563e-02 6.58181429e-01 -1.89574391e-01 3.66382986e-01
-1.22353685e+00 1.17000151e+00 8.04124117e-01 3.58488671e-02
-5.98464012e-01 -6.40746176e-01 -8.71235371e-01 5.57211339e-01
6.82085693e-01 -1.36786568e+00 1.29809380e+00 -1.75131238e+00
-1.72797906e+00 5.11353970e-01 -1.81880683e-01 -8.25616241e-01
6.72284782e-01 -4.07975376e-01 -1.45900428e-01 -2.96254486e-01
1.48641080e-01 5.12706161e-01 5.50524771e-01 -1.67450821e+00
-5.43041885e-01 -4.48621511e-01 6.97619021e-01 3.16700399e-01
1.25688374e-01 -6.31935894e-01 1.91848755e-01 -4.99668986e-01
-1.46905884e-01 -9.43327248e-01 -3.45123529e-01 -4.72779781e-01
-5.50589204e-01 -2.47806132e-01 6.62146211e-01 -3.38753760e-01
1.14691544e+00 -1.69182169e+00 4.46256012e-01 4.60018337e-01
1.70589581e-01 2.57576890e-02 7.38622015e-03 2.79983468e-02
-5.17565310e-01 4.80205387e-01 -4.37968493e-01 -2.46243015e-01
3.84005666e-01 5.13111830e-01 -1.63665131e-01 -3.64784175e-03
5.46741337e-02 7.87209153e-01 -7.95034587e-01 -1.55296372e-02
3.18083942e-01 -2.26443373e-02 -8.11136723e-01 3.37903917e-01
-6.17138565e-01 2.32585400e-01 -5.50547600e-01 -2.38326434e-02
5.44910789e-01 -3.25233787e-01 4.62660760e-01 6.34466037e-02
-5.33871688e-02 2.36772165e-01 -1.28898704e+00 1.34563923e+00
-7.73130655e-01 4.04366821e-01 -4.43716139e-01 -1.08445990e+00
5.69915950e-01 4.15719986e-01 -9.45093259e-02 -1.61113799e-01
2.02994123e-01 1.94078147e-01 3.19494694e-01 -5.29613674e-01
2.21621484e-01 -3.51853788e-01 1.19085878e-01 4.79164839e-01
2.87890788e-02 2.53584795e-02 1.45297244e-01 -2.00426474e-01
6.99103594e-01 6.60976395e-02 1.10356903e+00 -5.75842440e-01
1.00838292e+00 7.26969093e-02 3.47430557e-01 9.55515206e-01
3.45084012e-01 4.18465227e-01 5.71173012e-01 -8.50823760e-01
-9.12548065e-01 -7.33989418e-01 -4.13287012e-03 5.75968444e-01
1.38046026e-01 -1.32835776e-01 -1.08768868e+00 -1.19508171e+00
8.34743530e-02 1.44322598e+00 -9.28616166e-01 -1.09999850e-01
-2.48081550e-01 -6.79718852e-01 -5.07008880e-02 3.23442519e-01
5.68333864e-01 -9.70133483e-01 -9.06595707e-01 8.68779025e-04
-2.77598798e-01 -5.45712411e-01 -6.64888620e-02 3.26817989e-01
-1.04502833e+00 -1.18618762e+00 -3.32883187e-02 2.09877477e-03
7.60880888e-01 -2.01136157e-01 1.23019433e+00 5.40164173e-01
4.17080894e-02 4.42707151e-01 -3.88068318e-01 -8.41698527e-01
-6.47146344e-01 1.21041432e-01 -1.17893301e-01 1.06793478e-01
7.76296481e-02 -4.70977068e-01 -4.21845138e-01 1.92856759e-01
-1.16806936e+00 3.52269322e-01 5.28163671e-01 9.94512022e-01
3.88112456e-01 -1.82828754e-01 5.56130648e-01 -1.48718202e+00
7.85257995e-01 -3.95000875e-01 -7.81569898e-01 5.82401395e-01
-1.37163496e+00 1.01944828e+00 7.83616662e-01 -1.62867457e-01
-1.61658859e+00 -5.21043651e-02 1.76406175e-01 1.91789135e-01
-4.03039843e-01 5.16625404e-01 -7.99788535e-01 1.14268214e-01
7.56859660e-01 -4.67520326e-01 -2.24070922e-01 -4.14542556e-01
6.47260904e-01 3.62285793e-01 2.98500925e-01 -6.85143471e-01
7.98529804e-01 2.80895084e-01 2.38118798e-01 -9.06614885e-02
-9.04891431e-01 3.03278770e-02 -5.37835360e-01 2.55818982e-02
7.98661470e-01 -1.67192668e-01 -1.03419673e+00 -4.21064585e-01
-1.31275249e+00 -3.24232206e-02 -5.72860301e-01 4.80744928e-01
-9.85774100e-01 1.35926485e-01 4.10386920e-01 -7.64791667e-01
-2.31787905e-01 -1.26266980e+00 7.98150778e-01 2.10906118e-01
-7.44151592e-01 -1.43368816e+00 1.06905982e-01 5.08404553e-01
5.05132526e-02 4.23462510e-01 1.30038691e+00 -1.15774703e+00
-6.78588033e-01 1.32598430e-01 4.69337357e-03 2.42494106e-01
5.10118268e-02 -2.46961385e-01 -1.11588442e+00 3.41176689e-01
2.29121298e-01 3.31170917e-01 6.57387316e-01 4.18528885e-01
1.21747565e+00 -1.08939338e+00 -3.51818383e-01 3.94466609e-01
1.58216166e+00 3.99854869e-01 5.67665696e-01 6.77492976e-01
3.77946615e-01 6.43421888e-01 8.22537482e-01 2.19653577e-01
1.07499175e-01 8.71593475e-01 6.25405550e-01 -3.33836265e-02
3.24641615e-01 -8.05647075e-02 -7.60076940e-02 -3.00572425e-01
-7.17783272e-01 -4.11402881e-01 -7.52820849e-01 3.10442746e-01
-2.40756059e+00 -9.92683828e-01 -3.21130037e-01 2.28073955e+00
3.63498032e-01 4.71101888e-02 -1.66519180e-01 3.88479918e-01
5.59764683e-01 -4.15211499e-01 -3.80645603e-01 -8.97854745e-01
1.70680240e-01 1.44578040e-01 2.78333902e-01 1.26289201e+00
-8.32163632e-01 4.23669875e-01 6.11982965e+00 4.49231088e-01
-6.87357605e-01 1.33513644e-01 6.67070150e-01 1.95700631e-01
-1.01597559e+00 4.91804123e-01 -2.50537157e-01 2.23031953e-01
9.66699839e-01 -5.45436502e-01 5.25354803e-01 1.02168584e+00
4.90317285e-01 -1.18267082e-01 -1.68878114e+00 3.62032771e-01
-7.60605559e-02 -1.38491404e+00 4.80956793e-01 8.84267390e-02
7.92414129e-01 -9.15117145e-01 1.69283394e-02 3.78351919e-02
1.61046371e-01 -1.08070707e+00 7.18712389e-01 6.15375161e-01
9.52545181e-02 -7.90744185e-01 1.06484497e+00 3.77023369e-01
-4.79178220e-01 -3.13389331e-01 1.01187243e-03 -4.68475938e-01
-3.71333882e-02 3.08770448e-01 -1.26951635e+00 9.78961706e-01
9.11241695e-02 2.14433819e-01 -4.23324585e-01 8.54893446e-01
-6.60090506e-01 3.38364989e-01 1.42266636e-03 -1.56420559e-01
2.20892102e-01 -1.57278448e-01 7.32981682e-01 8.65247846e-01
2.31794283e-01 3.25120360e-01 -3.42645258e-01 1.19556952e+00
1.47767052e-01 1.52483173e-02 -5.66694200e-01 4.14201766e-01
5.82715608e-02 8.40220451e-01 -5.78467548e-01 -6.61010265e-01
8.22017528e-03 1.19882286e+00 3.03388596e-01 4.09586579e-01
-8.83744657e-01 -3.74140292e-02 7.69681334e-01 -7.30563281e-03
-1.03120811e-01 5.58244109e-01 -8.32095027e-01 -1.23488092e+00
5.76224960e-02 -9.50224400e-01 6.60621583e-01 -8.64723027e-01
-9.86649692e-01 6.52101278e-01 6.17900729e-01 -9.56954300e-01
-6.31268442e-01 -8.55003238e-01 -6.60818100e-01 7.56715655e-01
-1.35830605e+00 -8.80941808e-01 -2.95605242e-01 1.31209686e-01
6.59266829e-01 8.32232460e-02 8.76910329e-01 -2.42689148e-01
-2.89833009e-01 2.94322073e-01 -1.53137386e-01 -7.23330438e-01
1.47024423e-01 -1.81033885e+00 1.07168630e-01 8.71993184e-01
1.32776305e-01 8.64422321e-01 1.31761515e+00 -3.61293316e-01
-8.09083104e-01 -8.76984715e-01 1.14659226e+00 -6.17425442e-01
1.26839638e-01 1.29526436e-01 -8.16208065e-01 1.02974820e+00
4.26689118e-01 -5.93575478e-01 6.51742578e-01 4.60788816e-01
-8.08873326e-02 2.35426370e-02 -1.34968531e+00 6.54422402e-01
1.04137635e+00 -1.15535548e-02 -1.12898731e+00 1.83420435e-01
8.52138281e-01 -1.40384927e-01 -5.19869924e-01 4.84204799e-01
3.93487036e-01 -1.10806739e+00 8.33887815e-01 -9.36114132e-01
3.97392631e-01 -6.16259694e-01 3.16080779e-01 -1.62251139e+00
-3.26357126e-01 -6.31482422e-01 -4.42356840e-02 9.66641366e-01
7.74529696e-01 -8.58972132e-01 4.84672070e-01 1.29271078e+00
-1.53371155e-01 -3.72667134e-01 -8.34360301e-01 -6.55510306e-01
-4.47153896e-01 -2.51093149e-01 1.09013510e+00 7.79880524e-01
3.45876515e-01 5.07792532e-01 -2.17283115e-01 4.90556031e-01
5.21310747e-01 3.56045991e-01 7.01430380e-01 -1.46695137e+00
-5.94105124e-01 -4.81486320e-01 -4.27509457e-01 -5.65117896e-01
4.58751947e-01 -8.91877592e-01 -1.39010057e-01 -1.48682523e+00
2.90021837e-01 -7.82001857e-03 -5.68370409e-02 6.05695665e-01
-6.27084076e-01 -4.59344447e-01 3.15619409e-01 -1.30327210e-01
-3.51127416e-01 5.89252770e-01 1.03401196e+00 -1.45872608e-01
-1.87987179e-01 2.75856495e-01 -1.00647271e+00 9.28140163e-01
6.44920707e-01 -4.83967215e-01 -8.95067334e-01 -2.65881419e-01
1.24955930e-01 1.12445597e-02 5.98348320e-01 -8.26001048e-01
-2.80358136e-01 -5.16754568e-01 1.57724202e-01 4.33184355e-02
-1.20336041e-01 -1.22529745e+00 7.08580852e-01 5.70999324e-01
-6.48873925e-01 2.02923324e-02 5.62353879e-02 5.33068776e-01
-2.54931062e-01 -6.34124458e-01 4.05060470e-01 -3.46471556e-02
-6.52457297e-01 -2.19903812e-01 -3.42277855e-01 -2.30591282e-01
1.06276405e+00 -2.24651560e-01 -1.70322150e-01 -3.23034853e-01
-9.69393611e-01 1.67246401e-01 4.23768938e-01 2.44180501e-01
4.46192950e-01 -1.09717321e+00 -4.40528303e-01 -8.18658024e-02
2.80842572e-01 -2.95892388e-01 7.26659968e-03 3.99329215e-01
-4.09692913e-01 7.08402455e-01 -2.09834188e-01 -9.48486328e-02
-1.28266144e+00 6.64611518e-01 5.81741154e-01 -5.89638531e-01
-1.09420724e-01 2.57480890e-01 4.97520208e-01 -4.79250789e-01
-1.41461849e-01 -7.91192174e-01 -4.25307244e-01 -4.64354575e-01
2.35746309e-01 4.67298150e-01 -6.28450960e-02 3.84032764e-02
-2.16949925e-01 9.93414968e-02 2.01136187e-01 -3.62575471e-01
1.17245603e+00 -1.50045305e-01 -2.29866914e-02 3.12786281e-01
7.63529480e-01 -3.08091432e-01 -8.52493227e-01 3.36034656e-01
5.25790393e-01 -5.12606680e-01 -3.25052649e-01 -1.28308046e+00
-5.14688849e-01 4.77003485e-01 3.99467200e-01 4.35496420e-01
1.23662055e+00 -8.15731138e-02 -1.38630986e-01 7.20033705e-01
1.38945043e-01 -6.99957490e-01 -3.91216934e-01 2.05234110e-01
1.29799914e+00 -1.23528326e+00 2.16402784e-02 -6.01286113e-01
-8.03464532e-01 1.21935940e+00 6.42595470e-01 5.10228872e-02
2.09127158e-01 -2.07903743e-01 -1.00138158e-01 -1.63606972e-01
-8.38252306e-01 -1.37237847e-01 7.91602910e-01 5.05534649e-01
4.53068316e-01 4.17874604e-01 -1.01451004e+00 9.36154187e-01
-2.61983275e-01 5.64120561e-02 5.13803482e-01 3.84083003e-01
-3.10136050e-01 -1.40465140e+00 -4.44244802e-01 1.42037645e-01
-1.28277406e-01 -1.73351467e-02 -7.20900118e-01 1.08770490e+00
3.63306612e-01 9.97842133e-01 -2.78369188e-01 -3.15503106e-02
4.51434106e-01 2.25725695e-01 2.96048313e-01 -5.81329882e-01
-7.27362871e-01 -4.91369247e-01 4.26467329e-01 -4.25525814e-01
-5.90567648e-01 -5.58920622e-01 -1.26355040e+00 1.17726754e-02
-4.03357983e-01 7.03795731e-01 5.29115140e-01 1.39144361e+00
3.49807829e-01 6.23899937e-01 5.19550800e-01 -5.19003510e-01
-5.42307138e-01 -8.24191272e-01 -2.03452557e-01 6.94911480e-01
3.61803234e-01 -7.63617933e-01 -6.58166528e-01 9.74589288e-02] | [8.754082679748535, 5.710731029510498] |
c06d7a15-dc9d-4975-b90c-7c3c2e9b65f4 | randomized-quantization-is-all-you-need-for | 2306.11913 | null | https://arxiv.org/abs/2306.11913v1 | https://arxiv.org/pdf/2306.11913v1.pdf | Randomized Quantization is All You Need for Differential Privacy in Federated Learning | Federated learning (FL) is a common and practical framework for learning a machine model in a decentralized fashion. A primary motivation behind this decentralized approach is data privacy, ensuring that the learner never sees the data of each local source itself. Federated learning then comes with two majors challenges: one is handling potentially complex model updates between a server and a large number of data sources; the other is that de-centralization may, in fact, be insufficient for privacy, as the local updates themselves can reveal information about the sources' data. To address these issues, we consider an approach to federated learning that combines quantization and differential privacy. Absent privacy, Federated Learning often relies on quantization to reduce communication complexity. We build upon this approach and develop a new algorithm called the \textbf{R}andomized \textbf{Q}uantization \textbf{M}echanism (RQM), which obtains privacy through a two-levels of randomization. More precisely, we randomly sub-sample feasible quantization levels, then employ a randomized rounding procedure using these sub-sampled discrete levels. We are able to establish that our results preserve ``Renyi differential privacy'' (Renyi DP). We empirically study the performance of our algorithm and demonstrate that compared to previous work it yields improved privacy-accuracy trade-offs for DP federated learning. To the best of our knowledge, this is the first study that solely relies on randomized quantization without incorporating explicit discrete noise to achieve Renyi DP guarantees in Federated Learning systems. | ['Jacob Abernethy', 'Juba Ziani', 'Zihao Hu', 'Yeojoon Youn'] | 2023-06-20 | null | null | null | null | ['quantization'] | ['methodology'] | [-3.95690426e-02 5.71397990e-02 -3.36804062e-01 -2.58573830e-01
-1.34024787e+00 -1.06400383e+00 5.18922448e-01 2.99026221e-01
-5.25779724e-01 8.35019827e-01 9.05191228e-02 -4.41038400e-01
-1.63275987e-01 -8.98048818e-01 -9.65392768e-01 -9.42374647e-01
-1.88461915e-01 2.91738272e-01 -1.35835648e-01 1.80204019e-01
-1.19935393e-01 5.36605954e-01 -1.41660500e+00 1.65771186e-01
5.39817214e-01 1.06197262e+00 -5.07300377e-01 7.55153954e-01
-1.13185287e-01 9.20583069e-01 -4.35006678e-01 -7.62604415e-01
8.05268407e-01 -4.39600915e-01 -8.89406323e-01 -3.16868812e-01
4.86428201e-01 -6.80118322e-01 -3.16798270e-01 1.27260280e+00
3.78065854e-01 -1.19026475e-01 3.06589901e-01 -1.44847262e+00
-2.89887935e-01 8.02107453e-01 -3.63998801e-01 -5.28850630e-02
1.82525665e-01 1.49831116e-01 9.87611830e-01 -4.52630997e-01
6.66459262e-01 8.65836382e-01 6.95624769e-01 5.48626721e-01
-1.48319602e+00 -8.88251901e-01 -1.54282779e-01 -1.06048509e-01
-1.32621467e+00 -6.45832062e-01 4.47490573e-01 -2.13196039e-01
2.87957728e-01 7.77684152e-01 1.90878928e-01 7.71232069e-01
8.67290646e-02 7.79005885e-01 1.38910043e+00 -4.59944278e-01
6.96102679e-01 4.96374071e-01 7.98374340e-02 5.16493082e-01
5.00001848e-01 3.45037282e-01 -7.52871037e-01 -8.76302063e-01
4.07759249e-01 2.62432486e-01 -4.63166654e-01 -8.04643869e-01
-7.76810288e-01 8.82138789e-01 1.54991761e-01 1.13044173e-01
-2.15272456e-01 3.49155694e-01 5.09487689e-01 9.43773925e-01
5.08278906e-01 -9.07308236e-02 -6.19663596e-01 -5.25575224e-03
-8.85567784e-01 3.00172985e-01 1.36350429e+00 9.37507749e-01
1.14479184e+00 -3.42559844e-01 9.54392180e-02 1.85388997e-01
1.88242599e-01 3.43331516e-01 3.26221049e-01 -1.34081137e+00
5.52249789e-01 1.42272294e-01 2.75460124e-01 -7.26252496e-01
3.21396828e-01 -4.76863571e-02 -7.90094495e-01 4.01490867e-01
6.91132486e-01 -5.34632504e-01 -1.01759680e-01 1.98774064e+00
6.58686161e-01 -1.85702667e-01 3.89324665e-01 7.03400970e-01
3.61581407e-02 3.72583061e-01 -1.46698490e-01 -3.48365039e-01
1.11255229e+00 -5.67026496e-01 -6.26867890e-01 4.93546605e-01
8.09867084e-01 -3.07218075e-01 5.89953601e-01 5.02188146e-01
-1.08274066e+00 3.54130745e-01 -9.59714830e-01 -6.38499334e-02
-4.27153558e-01 -4.78715211e-01 7.97746539e-01 1.20166516e+00
-1.19714296e+00 7.35254765e-01 -1.01710868e+00 -2.15273395e-01
7.18905568e-01 5.96735418e-01 -7.15119600e-01 -6.78943843e-02
-1.00089228e+00 3.00752610e-01 -1.01291174e-02 -5.50679147e-01
-9.26121473e-01 -7.59868205e-01 -6.96687520e-01 1.90305069e-01
3.90358239e-01 -8.19320261e-01 1.54527855e+00 -9.26804006e-01
-1.31478190e+00 7.71698058e-01 -1.42025918e-01 -9.38593864e-01
1.04278803e+00 2.30668619e-01 -5.99316545e-02 1.49267793e-01
-1.38127118e-01 -1.00527719e-01 7.20592976e-01 -1.33455873e+00
-8.79392207e-01 -9.19753969e-01 -8.65383167e-03 6.11324683e-02
-6.99203432e-01 3.67491581e-02 -4.58068550e-02 -4.91230696e-01
-1.16112262e-01 -5.69657624e-01 -3.21756452e-01 4.90286410e-01
-9.90698487e-02 1.13095613e-02 9.86949980e-01 -2.94341743e-01
9.84555364e-01 -2.27840042e+00 -4.35282499e-01 4.23238993e-01
5.99938452e-01 6.02689832e-02 2.36265108e-01 7.01456428e-01
3.63217354e-01 3.42361808e-01 -2.78466552e-01 -7.75695860e-01
3.33860904e-01 2.95771599e-01 -5.07806242e-01 9.90217626e-01
-6.16939366e-01 7.30369568e-01 -7.51477540e-01 -3.95827413e-01
-5.34259565e-02 3.87555361e-01 -7.57340670e-01 7.37129599e-02
-3.36544216e-01 3.17073703e-01 -5.16769648e-01 5.11923552e-01
8.67094100e-01 -2.39522681e-01 3.62249732e-01 2.45085359e-01
-1.85157388e-01 1.93353340e-01 -1.43132401e+00 1.61666274e+00
-3.32391560e-01 2.19086081e-01 8.94929945e-01 -6.80178702e-01
4.98665571e-01 7.44833946e-01 7.80202329e-01 -2.76745915e-01
1.86883301e-01 3.31625611e-01 -7.61748075e-01 -1.25492036e-01
1.68435410e-01 -9.47820246e-02 -1.47479847e-01 1.12603629e+00
-3.59540842e-02 3.05274636e-01 -4.99434888e-01 2.09098980e-01
1.31697750e+00 -2.84462392e-01 2.89496571e-01 -2.52328902e-01
4.21185315e-01 -1.55187160e-01 6.71299160e-01 9.79000926e-01
-4.77249771e-01 2.73719341e-01 4.79733795e-01 -4.15754467e-01
-7.90019393e-01 -9.41249967e-01 1.13763280e-01 1.20097506e+00
-1.16182625e-01 -5.42420268e-01 -8.40798795e-01 -9.29214895e-01
4.70242560e-01 4.99194324e-01 -4.66263354e-01 1.28783152e-01
-1.89626902e-01 -4.01930004e-01 8.51636708e-01 6.05699345e-02
4.22197819e-01 -4.49475586e-01 -6.34056866e-01 9.85628366e-02
1.69719756e-01 -5.69819748e-01 -6.65664315e-01 5.60152650e-01
-8.12103152e-01 -1.10595918e+00 -2.80317336e-01 -3.59913349e-01
6.46034479e-01 2.49285311e-01 7.10290730e-01 -2.25032523e-01
2.71292794e-02 6.66120172e-01 4.62261550e-02 -3.95310074e-01
-4.42407608e-01 7.55018415e-03 1.17471084e-01 3.09053153e-01
5.61380625e-01 -7.76885986e-01 -6.19183004e-01 -8.90246406e-02
-1.27523255e+00 -4.93504077e-01 9.43555906e-02 6.99270189e-01
6.00205302e-01 1.96003556e-01 2.87174314e-01 -1.40544200e+00
4.99769509e-01 -6.09586418e-01 -8.88577521e-01 3.15245956e-01
-9.65836287e-01 2.01056272e-01 1.15758932e+00 -1.10999599e-01
-7.62918830e-01 3.11427683e-01 1.68764725e-01 -6.97082639e-01
-1.98999733e-01 -1.19724656e-02 -4.60287035e-01 -5.05642474e-01
6.69927657e-01 2.54127085e-01 2.92719901e-01 -6.58520103e-01
6.95327401e-01 9.74974394e-01 5.71961641e-01 -9.20393288e-01
8.56017470e-01 8.07945073e-01 4.22923565e-02 -2.08843097e-01
-2.35463530e-01 -2.77606457e-01 -2.06949607e-01 3.08054209e-01
3.33266616e-01 -9.15265203e-01 -1.25209713e+00 2.75548488e-01
-8.44052732e-01 -2.05726966e-01 -7.25893319e-01 2.77922362e-01
-7.39441395e-01 5.60285687e-01 -6.36699378e-01 -1.03545737e+00
-5.70593715e-01 -8.64307046e-01 6.34794831e-01 -7.88287148e-02
1.06645552e-02 -9.81990159e-01 2.43941903e-01 3.82970363e-01
5.24187386e-01 4.14256930e-01 6.09246433e-01 -1.03894329e+00
-7.64648736e-01 -3.06645155e-01 1.76775947e-01 1.88648045e-01
2.33827382e-01 -4.85328227e-01 -1.31727493e+00 -8.15996528e-01
4.52254236e-01 -4.59764123e-01 4.01248991e-01 -2.22165644e-01
1.27241504e+00 -1.23033190e+00 -1.44658759e-02 9.46011782e-01
1.78489578e+00 -3.84128571e-01 -2.51887590e-02 2.48737693e-01
3.37773591e-01 3.59845251e-01 1.33022666e-01 1.06206620e+00
5.84538817e-01 1.67144552e-01 4.17357564e-01 1.22017175e-01
3.82625908e-01 -5.43579400e-01 3.23606163e-01 3.06687593e-01
4.60608542e-01 9.57881287e-02 -4.98386145e-01 5.49865603e-01
-1.75460434e+00 -9.53674495e-01 2.36596584e-01 2.52943492e+00
1.18036008e+00 -5.04133344e-01 2.03984290e-01 8.98149461e-02
4.61470455e-01 8.07450563e-02 -8.01345587e-01 -5.73643923e-01
-3.03779934e-02 3.31163019e-01 1.28126717e+00 4.90168095e-01
-1.00839782e+00 5.39973795e-01 5.69862223e+00 5.33298373e-01
-9.89705026e-01 4.76817697e-01 6.41772151e-01 -3.90644968e-01
-6.17118418e-01 1.83927640e-01 -6.16195500e-01 4.76972073e-01
1.12148261e+00 -7.30897307e-01 1.00592768e+00 1.03123963e+00
-1.23072363e-01 2.46517211e-01 -1.46936059e+00 1.06711364e+00
-3.04679334e-01 -1.28739631e+00 -2.41108045e-01 5.44858396e-01
7.49579370e-01 1.57970518e-01 1.43987656e-01 7.26584764e-03
1.14210141e+00 -8.60221744e-01 5.11699200e-01 3.96890044e-01
8.66580665e-01 -1.03794849e+00 2.50189513e-01 7.17555821e-01
-9.74597275e-01 -3.18236470e-01 -2.23825708e-01 1.33433208e-01
-4.67571914e-01 4.23576921e-01 -3.97803515e-01 7.75176764e-01
7.75861382e-01 1.10897705e-01 -2.15766668e-01 7.30759442e-01
2.56561100e-01 7.23127842e-01 -6.87320888e-01 2.31847465e-01
2.50367261e-02 -2.47315198e-01 4.24542904e-01 9.55028713e-01
1.20653816e-01 1.00453764e-01 8.94161537e-02 7.48378873e-01
-5.88720202e-01 2.07222253e-01 -8.27600002e-01 1.51991978e-01
8.49893630e-01 9.62230504e-01 8.54519159e-02 -1.65804431e-01
-4.85564411e-01 9.97982264e-01 4.15978521e-01 3.11151177e-01
-1.95015684e-01 -2.07359970e-01 9.81358826e-01 -7.25918934e-02
3.90463442e-01 -1.04716029e-02 -3.99709404e-01 -1.35666895e+00
1.14084408e-01 -1.28476036e+00 1.07048249e+00 1.79517031e-01
-1.64970100e+00 7.17806816e-02 -2.97546536e-01 -9.38450813e-01
-4.13137972e-01 9.01921988e-02 -2.70512849e-01 8.91912937e-01
-1.60866916e+00 -1.01313007e+00 2.50095576e-01 1.20853114e+00
-2.92482227e-01 -3.77269723e-02 1.17284465e+00 1.49142981e-01
-3.50484461e-01 1.20451164e+00 8.28880072e-01 7.46075017e-03
6.84570193e-01 -1.35096943e+00 1.18235394e-01 9.65400040e-01
2.10550889e-01 8.00518990e-01 5.18966436e-01 -3.38770717e-01
-2.06024814e+00 -1.22347569e+00 9.79723334e-01 -4.43497151e-01
5.43868124e-01 -6.31462753e-01 -7.16726542e-01 1.03239357e+00
2.65348461e-02 5.02631783e-01 1.05015290e+00 -2.69157469e-01
-6.97338104e-01 -5.75174809e-01 -1.95246267e+00 3.75328004e-01
5.41392803e-01 -8.51551473e-01 -4.58023362e-02 3.74102026e-01
8.12897801e-01 -2.01329663e-01 -1.09834135e+00 -1.10603116e-01
3.58505100e-01 -1.23122191e+00 5.70650935e-01 -5.58437169e-01
-3.77368927e-01 -2.53924876e-01 -5.22807598e-01 -7.52510190e-01
1.53065473e-01 -1.41338897e+00 -6.68434918e-01 1.46867442e+00
-6.18573464e-02 -1.16461277e+00 1.22771204e+00 1.22949553e+00
7.59058952e-01 -4.29671168e-01 -1.36082637e+00 -7.30241477e-01
4.76545095e-01 -3.66067171e-01 1.13277864e+00 1.14851344e+00
6.95448890e-02 -6.52259767e-01 -2.67959684e-01 3.95762265e-01
1.20392966e+00 1.38690770e-01 1.01284397e+00 -9.87812161e-01
-5.05539298e-01 -1.24445431e-01 -1.78719029e-01 -8.05946767e-01
1.09399125e-01 -9.30642366e-01 -3.24069351e-01 -7.97949135e-01
-7.88360648e-03 -6.78879440e-01 -5.05914152e-01 8.32608104e-01
3.32869023e-01 -3.13103497e-02 3.02050263e-01 4.13782209e-01
-5.82331300e-01 4.68755245e-01 6.64366901e-01 4.19184119e-02
-7.58459568e-02 3.06839317e-01 -1.04691768e+00 1.53097898e-01
7.71743655e-01 -6.87974215e-01 -4.01096344e-01 -3.54851425e-01
-1.40583009e-01 2.87032455e-01 4.45015609e-01 -7.84111261e-01
8.04155529e-01 -1.58736408e-01 9.89568830e-02 -2.12181523e-01
-1.08742997e-01 -1.29144323e+00 4.93054599e-01 5.36577582e-01
-4.56133157e-01 -1.03178330e-01 -2.09136844e-01 7.60092199e-01
-9.81722865e-03 4.10860591e-02 7.58450389e-01 -2.74778545e-01
-5.42727113e-03 6.80019438e-01 -3.06435466e-01 -2.10638996e-02
1.17867649e+00 9.62231681e-02 -1.56538799e-01 -5.21328628e-01
-3.78594249e-01 3.63250345e-01 1.12378609e+00 -1.67343065e-01
1.56246603e-01 -1.07987547e+00 -4.31360871e-01 4.37564611e-01
-1.30855173e-01 1.29624605e-01 -1.85864359e-01 5.24792492e-01
-2.68782318e-01 4.24403757e-01 3.26755106e-01 -2.20517457e-01
-1.27716708e+00 7.43923724e-01 4.97157902e-01 -1.21160068e-01
-7.19656408e-01 4.99189675e-01 -1.64080441e-01 -6.63569868e-01
7.36187577e-01 3.50754219e-03 6.76033020e-01 -1.57415837e-01
6.32386804e-01 4.36680675e-01 2.20602706e-01 -3.50397259e-01
-2.40139768e-01 -7.50917494e-02 -1.56032458e-01 -3.89555395e-01
1.32711804e+00 -3.27264845e-01 -3.42018723e-01 2.10080639e-01
1.58090508e+00 2.37529695e-01 -1.45736110e+00 -6.64612174e-01
8.10860172e-02 -8.06883574e-01 2.92957034e-02 -5.45042157e-01
-1.16210711e+00 4.58023518e-01 7.20775306e-01 1.33127242e-01
1.20208406e+00 -2.71941543e-01 9.38131273e-01 4.90463644e-01
8.27912629e-01 -8.25123012e-01 -7.41683245e-01 3.15208063e-02
1.48641005e-01 -9.03454781e-01 -1.02386259e-01 5.05697764e-02
-3.32979292e-01 9.40401614e-01 1.60993278e-01 1.48786545e-01
6.97658002e-01 4.85053509e-01 3.38498682e-01 1.27765745e-01
-1.05464554e+00 2.99197137e-01 -4.51082349e-01 5.73953331e-01
-1.07285470e-01 2.20027938e-02 -1.84739709e-01 7.00075090e-01
-1.97850302e-01 2.81950027e-01 2.88262695e-01 1.41635609e+00
-1.19614787e-01 -1.49350834e+00 -6.19998872e-01 2.19978482e-01
-9.83896911e-01 1.91658944e-01 -4.23421413e-01 3.24034184e-01
1.93606783e-03 9.72088933e-01 -2.52889931e-01 -9.81933251e-02
-9.94658470e-02 2.02316940e-01 9.80071202e-02 -2.09092259e-01
-1.05867529e+00 -1.91412240e-01 -3.87785822e-01 -7.22408593e-01
-1.39782757e-01 -7.85142303e-01 -1.23639655e+00 -8.59005868e-01
-1.41387075e-01 5.32888651e-01 8.85510445e-01 5.12704432e-01
5.89760423e-01 -3.83112341e-01 1.38845801e+00 -1.81603014e-01
-1.47728598e+00 -1.07927024e-01 -9.74204242e-01 4.07433122e-01
7.92528033e-01 2.46529013e-01 -6.97640955e-01 -1.48010358e-01] | [5.8773674964904785, 6.6325860023498535] |
e581edb3-bba9-45f1-825a-03be135f7421 | goal-oriented-next-best-activity | 2205.03219 | null | https://arxiv.org/abs/2205.03219v1 | https://arxiv.org/pdf/2205.03219v1.pdf | Goal-Oriented Next Best Activity Recommendation using Reinforcement Learning | Recommending a sequence of activities for an ongoing case requires that the recommendations conform to the underlying business process and meet the performance goal of either completion time or process outcome. Existing work on next activity prediction can predict the future activity but cannot provide guarantees of the prediction being conformant or meeting the goal. Hence, we propose a goal-oriented next best activity recommendation. Our proposed framework uses a deep learning model to predict the next best activity and an estimated value of a goal given the activity. A reinforcement learning method explores the sequence of activities based on the estimates likely to meet one or more goals. We further address a real-world problem of multiple goals by introducing an additional reward function to balance the outcome of a recommended activity and satisfy the goal. We demonstrate the effectiveness of the proposed method on four real-world datasets with different characteristics. The results show that the recommendations from our proposed approach outperform in goal satisfaction and conformance compared to the existing state-of-the-art next best activity recommendation techniques. | ['Sampath Dechu', 'Renuka Sindhgatta', 'Avani Gupta', 'Prerna Agarwal'] | 2022-05-06 | null | null | null | null | ['activity-prediction', 'activity-prediction'] | ['computer-vision', 'time-series'] | [ 3.06664318e-01 1.78282350e-01 -7.10297525e-01 -5.57328820e-01
-2.46864617e-01 -2.25428149e-01 7.01092720e-01 2.04737633e-02
-7.01615065e-02 7.47508883e-01 8.47208679e-01 -2.12874845e-01
-8.40456665e-01 -9.27548409e-01 -5.59458733e-01 -3.29617411e-01
-3.05211335e-01 5.26073575e-01 1.11378888e-02 -1.75067633e-01
8.12926888e-01 1.81200862e-01 -1.59327304e+00 3.48084539e-01
8.62851083e-01 1.11257184e+00 3.79198194e-01 5.58428705e-01
-1.35727003e-02 1.21298254e+00 -3.98321092e-01 -1.97333187e-01
5.47518015e-01 -6.43128216e-01 -9.12353277e-01 4.14599776e-01
5.96667528e-02 -5.42281210e-01 -3.27472419e-01 2.85094172e-01
-4.80309278e-02 7.64929235e-01 5.39927959e-01 -1.59988463e+00
-6.18383229e-01 7.31717706e-01 -1.22984894e-01 3.56330991e-01
6.04576468e-01 4.19546179e-02 1.15331507e+00 -2.32970983e-01
4.08300847e-01 5.34842551e-01 3.04407299e-01 4.65595543e-01
-7.70634115e-01 -3.53628844e-01 7.15833664e-01 3.41395646e-01
-7.25073457e-01 -2.32661858e-01 5.80477297e-01 -3.46094102e-01
1.41337121e+00 -5.95227368e-02 9.48108613e-01 9.04979467e-01
5.64900994e-01 5.55610776e-01 6.73381627e-01 -1.40647158e-01
5.20879030e-01 -4.12313402e-01 3.60603374e-03 2.76591599e-01
4.42980975e-02 1.88275710e-01 -7.07129657e-01 -8.22297782e-02
5.92851460e-01 6.75107419e-01 -3.93689498e-02 -3.01135451e-01
-1.26393759e+00 5.11848271e-01 -4.74400958e-03 4.89977479e-01
-1.06054854e+00 3.83847594e-01 -1.12578683e-01 2.17728525e-01
5.81518142e-03 6.43566251e-01 -4.85992849e-01 -5.31064630e-01
-1.08142531e+00 4.19543654e-01 1.17576194e+00 8.51720512e-01
5.86173058e-01 -1.55863501e-02 -7.05891848e-01 2.07994670e-01
4.81614947e-01 -6.69895932e-02 4.29103047e-01 -1.49356496e+00
5.02084434e-01 5.82566619e-01 9.38631654e-01 -7.44135320e-01
-1.74308062e-01 -6.28998458e-01 -1.82447910e-01 1.45951226e-01
5.99871874e-01 -1.07710898e-01 -4.56045628e-01 1.43797481e+00
-6.07035384e-02 8.10128629e-01 1.16224892e-01 8.03623915e-01
2.73687914e-02 7.67240167e-01 1.71017498e-01 -6.79525793e-01
6.44500494e-01 -1.25472188e+00 -8.10179055e-01 -4.42691684e-01
4.72615182e-01 -1.94924414e-01 9.16722894e-01 6.24834239e-01
-1.13641655e+00 -6.06411040e-01 -1.08386910e+00 7.10436404e-01
2.36645535e-01 -2.75240451e-01 8.03109944e-01 5.13498902e-01
-7.94940174e-01 1.17655396e+00 -9.58838224e-01 -3.03341687e-01
2.05514744e-01 3.42553735e-01 2.38582748e-03 2.27609668e-02
-9.04757559e-01 6.22383237e-01 3.16433132e-01 -1.40567720e-01
-1.23255253e+00 -5.11783540e-01 -4.54135329e-01 5.32317638e-01
7.28964984e-01 -5.54300606e-01 1.40386665e+00 -1.13116097e+00
-1.56989264e+00 2.25105852e-01 1.09730691e-01 -7.09175050e-01
4.70652491e-01 -5.41831195e-01 -6.53211415e-01 -4.63104844e-01
-1.18679829e-01 1.56657454e-02 5.54616988e-01 -6.75347507e-01
-1.34040368e+00 -1.53808758e-01 2.75283575e-01 6.83777034e-01
-6.17171824e-01 -2.58111060e-01 -3.07633519e-01 -1.05456904e-01
-1.71738192e-01 -5.36085546e-01 -4.88087744e-01 -5.92440963e-01
2.88879484e-01 -5.97350001e-01 2.24012360e-01 -5.69196999e-01
1.59466767e+00 -1.73761666e+00 -1.01745740e-01 2.45208472e-01
-4.58540432e-02 -3.11425269e-01 -1.02081731e-01 8.00270915e-01
3.69711608e-01 -6.22887947e-02 3.16449761e-01 1.21360324e-01
1.30966306e-01 3.11425805e-01 -1.95453733e-01 3.55944544e-01
-3.07471365e-01 6.09277248e-01 -1.06055737e+00 -2.61891298e-02
2.27207631e-01 6.82806894e-02 -6.17788911e-01 8.64710689e-01
-4.84590292e-01 5.16860366e-01 -6.45349205e-01 2.96593517e-01
-1.64791159e-02 -5.21027803e-01 6.46550536e-01 4.15668219e-01
-7.70123824e-02 7.71473825e-01 -1.37881923e+00 1.59349918e+00
-3.94367248e-01 -2.10153490e-01 -3.80551338e-01 -1.17576838e+00
1.26708245e+00 2.00902537e-01 1.12884402e+00 -1.06684923e+00
-2.87698895e-01 5.91311678e-02 1.76508680e-01 -4.53796536e-01
5.29625893e-01 4.77114804e-02 4.06870444e-04 9.72461283e-01
-8.35635886e-02 4.73721027e-01 4.40464199e-01 -5.80828302e-02
1.44228542e+00 8.90425801e-01 5.69543660e-01 6.78414758e-03
4.69318718e-01 -1.65180251e-01 8.91177177e-01 1.17483866e+00
-4.11641032e-01 1.65536329e-02 3.83426398e-01 -6.04102194e-01
-7.62101531e-01 -6.41319513e-01 7.95199871e-01 1.51574099e+00
1.58753797e-01 -2.52120316e-01 -2.67197043e-01 -1.08112407e+00
-1.11720733e-01 1.13305163e+00 -4.57032621e-01 -4.09082137e-02
-6.77491426e-01 -1.68504193e-01 -2.45277688e-01 6.17931843e-01
3.92731309e-01 -1.43019450e+00 -8.53813469e-01 6.43714428e-01
-2.07756594e-01 -7.68983424e-01 -7.77622581e-01 -7.48381345e-03
-8.73720825e-01 -1.09523880e+00 -3.23060095e-01 -5.02674699e-01
2.56667227e-01 8.30019414e-02 1.28460658e+00 1.67869687e-01
7.42564201e-01 5.12619495e-01 -3.68608147e-01 -1.48818111e-02
-1.99157313e-01 -4.15731966e-03 1.42343074e-01 1.38096109e-01
6.08246505e-01 -7.57597387e-01 -9.33096230e-01 4.81076837e-01
-3.92934054e-01 6.15588501e-02 3.64419013e-01 4.28013563e-01
6.79857910e-01 2.71849930e-01 1.07560778e+00 -5.07967114e-01
9.16290224e-01 -7.93556273e-01 -4.36488390e-01 3.32693160e-01
-1.48137224e+00 5.49828932e-02 6.80192113e-01 -4.57483619e-01
-1.36557007e+00 -3.11397091e-02 -1.39784575e-01 -2.16091592e-02
-2.43192464e-01 5.88526547e-01 -1.30499125e-01 8.17402482e-01
4.98695463e-01 3.58020663e-01 -4.61091876e-01 -2.23861352e-01
2.09745809e-01 5.69906354e-01 3.51422071e-01 -3.98892045e-01
7.81256035e-02 3.27039421e-01 -1.73813075e-01 2.92734116e-01
-1.11680245e+00 -7.09148169e-01 -2.63433635e-01 -6.73099339e-01
3.86373848e-01 -4.67350662e-01 -1.12137711e+00 1.04192626e-02
-5.92294395e-01 -6.23773396e-01 -3.52220744e-01 7.10819066e-01
-1.28870988e+00 2.33862832e-01 -3.06147575e-01 -1.15281796e+00
-3.64015490e-01 -6.39936090e-01 4.01012421e-01 4.21620250e-01
-4.10607934e-01 -8.13534379e-01 2.60547936e-01 5.50739110e-01
4.11302030e-01 7.67319202e-02 2.52146244e-01 -9.42320883e-01
-7.21387863e-01 -1.39583156e-01 3.25861037e-01 -6.37290860e-03
3.47669929e-01 -4.97433573e-01 -1.69127300e-01 -3.10048968e-01
2.81934198e-02 -1.30307404e-02 3.95982772e-01 4.48956579e-01
1.11279702e+00 -7.04628527e-01 -2.38083571e-01 3.43635827e-01
1.52078485e+00 7.69112170e-01 9.24046934e-01 7.47140586e-01
2.07028583e-01 6.37847543e-01 1.43669641e+00 1.02140462e+00
3.20977151e-01 5.73769808e-01 7.68194497e-01 6.18336439e-01
3.22380215e-01 -5.26030719e-01 5.71690500e-01 2.49742568e-01
-3.67588758e-01 -6.14663363e-01 -7.96776950e-01 7.70938635e-01
-2.55910468e+00 -1.50923741e+00 1.55310154e-01 2.38303757e+00
2.16984510e-01 4.50819880e-01 2.86167979e-01 1.01865873e-01
4.43938673e-01 3.57902348e-02 -9.34126139e-01 -5.23217380e-01
5.96814573e-01 -1.74498260e-01 3.71666849e-01 3.13842744e-01
-7.62581766e-01 6.06075048e-01 6.75518179e+00 2.84162253e-01
-5.85651159e-01 -1.00804411e-01 6.03160739e-01 -3.64001364e-01
-3.37044924e-01 4.21841741e-02 -9.10543740e-01 3.95957500e-01
1.23781574e+00 -3.49575758e-01 7.72720337e-01 1.02495670e+00
7.58390546e-01 -3.66549268e-02 -1.45295942e+00 5.68132401e-01
-8.83349311e-03 -1.48376274e+00 -2.31778547e-01 3.29668641e-01
9.45563436e-01 -1.89562082e-01 -3.65252197e-01 6.36552811e-01
6.59043014e-01 -1.00207067e+00 6.09744966e-01 1.16517007e+00
1.00249641e-01 -9.50427532e-01 6.64866805e-01 8.12604070e-01
-9.49990273e-01 -6.85686648e-01 1.33418720e-02 -7.61294246e-01
3.20432603e-01 2.01526836e-01 -9.67134893e-01 4.09798801e-01
5.88682652e-01 9.79643822e-01 -5.59830712e-03 8.16852689e-01
-7.04416782e-02 8.50087583e-01 3.84927280e-02 -2.74890214e-01
2.47162089e-01 -4.69998211e-01 7.50211924e-02 7.32754529e-01
7.53074288e-01 -1.65492207e-01 6.61250651e-01 6.29536271e-01
9.71679688e-02 2.17834771e-01 -3.13178629e-01 -2.94782311e-01
3.61600101e-01 1.01801300e+00 -4.94058430e-01 -3.69811296e-01
-6.69823229e-01 8.69005919e-01 4.66536254e-01 2.37431705e-01
-7.83498049e-01 7.64423981e-02 6.71885073e-01 3.85581881e-01
4.87595588e-01 -5.30640446e-02 -3.64816278e-01 -7.19150662e-01
1.02489643e-01 -8.04225683e-01 7.93814123e-01 -7.57434070e-01
-1.13201213e+00 1.00626305e-01 -3.47610027e-01 -1.33400881e+00
-6.11651003e-01 -6.49130791e-02 -7.72297680e-01 7.45384157e-01
-1.52033997e+00 -8.80311549e-01 -3.65878195e-01 4.61194694e-01
6.31430507e-01 -4.17905211e-01 5.83897591e-01 -8.61890148e-03
-1.31691486e-01 8.68154783e-03 1.07571386e-01 -4.69614625e-01
5.63017309e-01 -1.11473858e+00 1.44794032e-01 9.42648649e-01
5.73778301e-02 4.47641075e-01 7.29178309e-01 -9.78411674e-01
-1.17810667e+00 -8.04624796e-01 1.09743655e+00 -3.82245362e-01
5.87985039e-01 2.55203396e-01 -6.74340785e-01 9.60746408e-01
2.97835082e-01 -3.06760758e-01 9.06535387e-01 1.99345812e-01
2.79892951e-01 -3.59501213e-01 -1.16008043e+00 4.73225594e-01
1.46475971e+00 2.68775791e-01 -7.96361327e-01 1.66586205e-01
5.43712974e-01 -9.25661400e-02 -1.14831614e+00 1.96554050e-01
9.10358608e-01 -1.20382917e+00 6.40337050e-01 -1.05867732e+00
6.26587391e-01 -9.15264860e-02 -1.41468242e-01 -1.23053026e+00
-8.14697444e-01 -7.07263291e-01 -9.41523910e-01 9.53195393e-01
1.83772027e-01 -1.20377839e-02 1.30982292e+00 8.42806995e-01
-3.70407552e-01 -8.94837499e-01 -5.69825888e-01 -6.84657454e-01
-4.79237616e-01 -4.85154241e-01 1.26310802e+00 8.04089367e-01
3.96093607e-01 7.28735179e-02 -9.32994127e-01 -2.24760529e-02
4.31817621e-01 5.11615515e-01 7.97355294e-01 -1.09774661e+00
-7.52912760e-01 -2.55847245e-01 8.44045803e-02 -1.36737549e+00
-9.51275751e-02 -6.40577137e-01 1.86693575e-02 -2.34416389e+00
1.24609277e-01 -1.40891537e-01 -1.09559190e+00 5.06954372e-01
6.14361204e-02 -4.52864766e-01 7.18957707e-02 2.24751458e-01
-1.13923204e+00 3.43399286e-01 1.36185193e+00 1.04052603e-01
-7.62629092e-01 5.83230257e-01 -1.17253184e+00 4.62378651e-01
1.00596225e+00 -4.03401673e-01 -5.82300961e-01 7.98516199e-02
6.15536690e-01 7.16789067e-01 -3.67847353e-01 -1.14681387e+00
1.98649466e-01 -1.14621067e+00 1.86300755e-01 -5.73486090e-01
1.35956593e-02 -1.01520467e+00 3.84053499e-01 6.22465670e-01
-8.66233885e-01 -1.60855263e-01 -6.19897604e-01 1.04234815e+00
2.50238717e-01 -1.95891678e-01 2.52851278e-01 -7.70210102e-02
-9.89963293e-01 4.34295505e-01 -5.51795185e-01 -2.91118711e-01
1.38298023e+00 -6.07887924e-01 3.48859914e-02 -8.13826442e-01
-8.86523426e-01 5.62645793e-01 1.08943716e-01 7.03837037e-01
6.02517486e-01 -1.25952613e+00 -6.91446006e-01 -1.19705558e-01
2.16701552e-01 -7.42892146e-01 4.23987582e-02 5.27868092e-01
7.99219012e-02 3.80100518e-01 -6.00961447e-01 -1.03120901e-01
-8.78787339e-01 6.80995822e-01 4.95589525e-01 -1.01608562e+00
-5.56030214e-01 1.38578132e-01 -2.62829006e-01 -2.46177331e-01
5.38311183e-01 -7.87226260e-02 -7.58074522e-01 -4.60399598e-01
6.02606654e-01 8.10631156e-01 -1.28559679e-01 -9.46284011e-02
-9.56300870e-02 -2.29545295e-01 9.12613124e-02 -7.56752864e-02
1.50681329e+00 -4.25069779e-01 4.50933397e-01 2.97878087e-01
2.21739396e-01 -2.79451460e-01 -1.85341227e+00 -1.74929321e-01
2.80244857e-01 -7.59407640e-01 7.03079700e-02 -1.11542404e+00
-8.41995835e-01 2.18607366e-01 5.46395123e-01 3.35257590e-01
1.04592848e+00 -2.02168152e-01 8.46237957e-01 3.81389052e-01
4.74101305e-01 -1.60736096e+00 4.67240602e-01 2.78496861e-01
7.09364355e-01 -9.93245661e-01 3.07074469e-02 1.09776899e-01
-9.16295826e-01 8.62932265e-01 1.13014722e+00 -2.51385152e-01
6.65386260e-01 -6.17255867e-02 -3.30836236e-01 -6.93530962e-02
-9.52482104e-01 -6.95826560e-02 1.85125068e-01 7.26763189e-01
6.21201038e-01 1.08975835e-01 -7.54219234e-01 7.48050809e-01
1.08987056e-01 5.69698393e-01 6.21501505e-01 8.32862198e-01
-7.08241224e-01 -1.13554442e+00 -1.39036193e-01 1.12999952e+00
-5.92847228e-01 4.40388262e-01 -1.76241130e-01 1.54037625e-01
2.09891517e-02 1.33574748e+00 3.03225905e-01 -1.97713211e-01
4.65158761e-01 1.65879697e-01 3.76435935e-01 -6.48682535e-01
-7.81475008e-01 1.11325622e-01 3.75268549e-01 -8.87844265e-01
-6.33176088e-01 -7.89066136e-01 -1.39820540e+00 -3.24907005e-01
-9.30494964e-02 2.06095949e-01 2.90553778e-01 1.35383892e+00
4.35305446e-01 8.12772453e-01 6.98051095e-01 -4.04021353e-01
-7.54824698e-01 -8.26833129e-01 -7.74149060e-01 8.14216673e-01
-3.31132300e-02 -5.46819806e-01 3.27812359e-02 -3.74156539e-03] | [7.922903537750244, 0.7013313174247742] |
2403bba3-5b0d-4af2-8218-93af990e7817 | impulsive-noise-removal-from-color-images | 1707.03126 | null | http://arxiv.org/abs/1707.03126v1 | http://arxiv.org/pdf/1707.03126v1.pdf | Impulsive noise removal from color images with morphological filtering | This paper deals with impulse noise removal from color images. The proposed
noise removal algorithm employs a novel approach with morphological filtering
for color image denoising; that is, detection of corrupted pixels and removal
of the detected noise by means of morphological filtering. With the help of
computer simulation we show that the proposed algorithm can effectively remove
impulse noise. The performance of the proposed algorithm is compared in terms
of image restoration metrics and processing speed with that of common
successful algorithms. | ['Alexey Ruchay', 'Vitaly Kober'] | 2017-07-11 | null | null | null | null | ['color-image-denoising'] | ['computer-vision'] | [ 6.04074478e-01 -8.24233234e-01 8.56613994e-01 9.82378200e-02
-1.35629445e-01 -3.58596414e-01 1.51680782e-01 -3.44567448e-02
-8.97486687e-01 5.66454947e-01 3.33419405e-02 -3.71368706e-01
-9.46381614e-02 -8.44557226e-01 2.88266148e-02 -1.02533698e+00
2.91777045e-01 -4.54599112e-01 2.52408713e-01 -3.09614748e-01
4.76145566e-01 8.31331372e-01 -1.39024711e+00 -1.95101753e-01
7.40368843e-01 8.89536619e-01 2.20462248e-01 1.23988676e+00
-1.22830048e-01 4.85958129e-01 -8.13253701e-01 8.54850039e-02
6.19566262e-01 -7.81444490e-01 -3.32601577e-01 6.73554182e-01
-1.31595150e-01 -2.63544828e-01 -3.27108532e-01 1.46721017e+00
6.28509045e-01 5.64883411e-01 4.91177231e-01 -5.16734719e-01
-6.48597360e-01 -8.25530738e-02 -8.30938220e-01 8.11098695e-01
1.53369322e-01 1.19596861e-01 -3.63253020e-02 -7.86823630e-01
6.29124343e-01 1.19477391e+00 6.77534103e-01 2.31567279e-01
-1.23745096e+00 -3.46705705e-01 -3.52188617e-01 3.08424294e-01
-1.39581645e+00 -1.33681223e-01 8.14894736e-01 1.69164930e-02
6.16379499e-01 4.64598179e-01 3.73752147e-01 6.18963651e-02
5.84426701e-01 2.94462770e-01 1.53665459e+00 -9.42574322e-01
2.41690338e-01 -4.23642285e-02 4.52953935e-01 4.68942672e-01
6.55347168e-01 6.89618513e-02 7.82804713e-02 -1.76860597e-02
6.28584623e-01 -5.86218899e-04 -3.54242057e-01 3.57792526e-01
-7.31903017e-01 4.77340698e-01 3.89460474e-02 7.13959932e-01
-6.53053343e-01 1.00646429e-01 4.63262528e-01 5.33033073e-01
4.36474085e-01 -7.54735842e-02 -4.87551726e-02 1.46563128e-01
-9.20829952e-01 -2.12685764e-01 7.40077496e-01 7.40786433e-01
2.91490763e-01 6.45106256e-01 1.07661813e-01 8.34703743e-01
3.05001944e-01 8.60347211e-01 2.19332993e-01 -9.95467901e-01
-2.63621390e-01 1.31119400e-01 1.52983889e-01 -1.09640074e+00
-2.28587374e-01 -1.32147670e-01 -1.10139096e+00 1.00091922e+00
3.04056704e-01 -3.16628963e-01 -1.33966017e+00 8.78656507e-01
1.13135345e-01 1.65688068e-01 3.84772539e-01 5.69630444e-01
6.90020442e-01 1.00623059e+00 1.53203100e-01 -7.88021147e-01
1.02737892e+00 -4.84828919e-01 -1.44725108e+00 1.09842256e-01
-4.97199059e-01 -1.43289626e+00 3.03884387e-01 9.68106449e-01
-1.16832268e+00 -7.54498005e-01 -9.40782130e-01 1.52221516e-01
-3.66519690e-01 1.64994776e-01 3.39286894e-01 8.91588807e-01
-9.14500654e-01 4.25540626e-01 -6.29048109e-01 -5.45970261e-01
8.13587978e-02 3.58525336e-01 -3.29287589e-01 -1.87297866e-01
-4.48689640e-01 8.08681428e-01 3.40084285e-01 6.44383550e-01
-3.37618053e-01 3.90556194e-02 -5.80329955e-01 -1.87866092e-01
1.53090835e-01 -5.27517855e-01 7.75016010e-01 -1.11291778e+00
-1.27394235e+00 5.90069532e-01 -3.37076008e-01 -2.31711164e-01
6.64982617e-01 -2.96253413e-01 -6.51518941e-01 4.65405464e-01
-1.79422021e-01 -2.18494937e-01 1.17668664e+00 -1.48241913e+00
-7.73898363e-01 -2.12695003e-01 -7.04088986e-01 -7.48219574e-03
1.38226479e-01 1.65915906e-01 -4.16229248e-01 -9.20089245e-01
6.64795101e-01 -3.64985615e-01 -4.57963616e-01 -8.64230469e-03
-3.21778953e-01 2.36026108e-01 1.32551980e+00 -9.51466143e-01
1.00591409e+00 -2.30900025e+00 -3.13456714e-01 6.83622360e-01
-2.10583821e-01 5.44866443e-01 -1.27989694e-01 5.63098133e-01
-1.06604934e-01 1.04505777e-01 -4.01612908e-01 -1.05581149e-01
-4.88327444e-01 2.55449146e-01 3.35307866e-02 8.09801042e-01
-1.48663461e-01 -8.34393650e-02 -3.74090999e-01 -4.37452137e-01
6.38897300e-01 9.10964310e-01 1.90513670e-01 -6.14969656e-02
6.55991375e-01 2.50479549e-01 -1.68324977e-01 7.42373049e-01
1.19424284e+00 7.02147543e-01 -2.51704305e-02 -5.74752808e-01
-4.07662690e-01 -5.63513517e-01 -1.78690851e+00 6.33191884e-01
-1.16733633e-01 7.04104483e-01 7.90259957e-01 -8.12012553e-01
1.12071466e+00 4.27892148e-01 3.42521578e-01 -4.97863442e-01
6.17977142e-01 3.78070116e-01 3.77004705e-02 -7.98699677e-01
2.39315912e-01 -2.74629951e-01 7.08706975e-01 1.28636584e-01
-1.44852355e-01 -2.11238772e-01 5.67562878e-01 -1.25375792e-01
1.02975249e+00 -1.48177072e-01 3.63878995e-01 -4.96531337e-01
1.04507351e+00 1.36097252e-01 5.49141705e-01 1.02479446e+00
-4.94466424e-01 2.82476217e-01 -1.56009391e-01 -1.74247593e-01
-1.05432427e+00 -9.62516308e-01 -1.46923900e-01 5.25424123e-01
4.57050204e-01 3.95495445e-01 -7.72150397e-01 1.80049807e-01
-2.32234895e-02 3.02226841e-01 -3.53868455e-01 1.95092961e-01
-9.22725141e-01 -1.02155316e+00 1.88667506e-01 6.97309971e-02
8.80110860e-01 -1.19570637e+00 -5.81173599e-01 3.63176227e-01
2.36526608e-01 -9.39564884e-01 8.00042972e-03 1.83724135e-01
-1.18697536e+00 -1.17460585e+00 -4.00969774e-01 -1.21740365e+00
1.02247369e+00 7.98737228e-01 6.52025521e-01 8.64602149e-01
-8.40386689e-01 5.84554374e-01 -4.85105485e-01 -3.26694936e-01
-5.55223465e-01 -9.67082918e-01 -1.14594713e-01 7.34065771e-02
3.78021538e-01 -6.67885661e-01 -7.94042647e-01 1.33874053e-02
-1.32261419e+00 -6.68098390e-01 7.69425094e-01 6.37074888e-01
6.21456802e-01 1.40906966e+00 1.22121699e-01 -8.30341220e-01
1.01293337e+00 9.82590839e-02 -7.51310229e-01 -7.21136630e-02
-5.19569457e-01 -5.69807172e-01 6.75487757e-01 -1.89725012e-01
-1.28930092e+00 3.43676955e-02 -9.36592296e-02 2.13732123e-01
-4.21948105e-01 2.12753952e-01 1.35533139e-01 -6.71672523e-01
3.39979082e-01 6.93111658e-01 3.24990332e-01 -1.05661309e+00
1.24484554e-01 6.00287020e-01 1.16605496e+00 6.49863034e-02
1.13177848e+00 8.05128634e-01 3.53424996e-01 -1.31177819e+00
3.37422729e-01 -7.10455239e-01 -4.51534897e-01 -2.79374808e-01
7.59629965e-01 -5.27550280e-01 -7.18389392e-01 1.18625414e+00
-8.25756848e-01 2.70021111e-01 1.36060134e-01 5.44485390e-01
-3.48947383e-02 7.55536914e-01 -1.17953062e+00 -1.24500215e+00
-5.05156457e-01 -8.80944073e-01 1.07794479e-01 4.65175599e-01
3.60369712e-01 -1.11846519e+00 -3.11853260e-01 -5.93904257e-02
6.98495209e-01 3.28689814e-01 8.54628086e-01 -9.26345438e-02
-3.08674961e-01 -6.17753148e-01 -2.23703191e-01 7.97333121e-01
4.65280592e-01 3.50948572e-01 -6.60635233e-01 -3.49645585e-01
7.17951179e-01 4.71819490e-01 1.18106937e+00 8.56082559e-01
5.32328367e-01 7.28590861e-02 -1.68539062e-01 4.77544248e-01
2.45496488e+00 7.79558897e-01 1.19604397e+00 7.46765375e-01
1.54590517e-01 1.53122395e-01 4.54852223e-01 2.92893976e-01
-6.89514995e-01 -1.95380151e-01 2.32749179e-01 -6.46664381e-01
-3.48856777e-01 6.77322447e-01 -7.31593370e-02 5.96206129e-01
-3.46591532e-01 -4.07211214e-01 -4.60572004e-01 4.94655401e-01
-1.30517852e+00 -9.99379277e-01 -5.65282226e-01 1.57719433e+00
4.35705930e-01 1.73481464e-01 -2.74138391e-01 8.59383404e-01
9.79147971e-01 -1.88234493e-01 -6.82788044e-02 -7.06412435e-01
-3.82408828e-01 5.10369182e-01 8.54692757e-01 7.75247991e-01
-1.22048616e+00 3.68640929e-01 7.06917715e+00 6.99765205e-01
-1.00096762e+00 -8.83166865e-02 2.79444784e-01 4.82263505e-01
2.75447398e-01 -1.74758345e-01 -1.62827119e-01 3.43774199e-01
1.89458624e-01 2.75231618e-02 3.03082407e-01 2.18741447e-01
8.30029666e-01 -9.81822073e-01 3.46413217e-02 9.47926939e-01
1.66970268e-01 -7.38885224e-01 1.17708929e-02 -3.29134762e-01
5.63187242e-01 -6.64652705e-01 -1.47732571e-01 -3.20069492e-01
-5.30991443e-02 -7.21937060e-01 4.42921847e-01 8.73129904e-01
-1.63950622e-01 -9.24192667e-01 1.18112385e+00 7.02745840e-02
-9.12029326e-01 -2.20305040e-01 -4.74585593e-01 -2.39201933e-01
2.32949749e-01 7.40828454e-01 -1.21188030e-01 3.61648798e-01
6.32833064e-01 3.02998483e-01 -3.88820678e-01 1.85138917e+00
-1.30350858e-01 7.95730233e-01 -2.73044646e-01 3.42842698e-01
1.42632186e-01 -8.40664566e-01 7.39265263e-01 1.54086435e+00
4.50879693e-01 6.26696587e-01 2.79470794e-02 4.41595405e-01
3.25942904e-01 3.35193872e-01 -3.16385299e-01 3.24137837e-01
3.82376164e-01 1.04326773e+00 -1.27429819e+00 -3.85873318e-01
-4.23881203e-01 1.01411879e+00 -8.78753662e-01 6.29740834e-01
-4.88239639e-02 -9.69353974e-01 4.15980071e-01 -2.40524456e-01
6.09548271e-01 -4.20980603e-01 -5.77431500e-01 -3.38009328e-01
-1.94882944e-01 -6.54493213e-01 2.33787164e-01 -7.09314466e-01
-1.09708762e+00 5.90133369e-01 -2.00748399e-01 -1.05528271e+00
5.81535399e-01 -7.81663656e-01 -9.48190689e-01 1.07244563e+00
-1.28832221e+00 -5.86263299e-01 -3.48501444e-01 6.79537654e-01
7.42239833e-01 -2.35131204e-01 3.71534318e-01 2.52327591e-01
-6.19592011e-01 1.65062938e-02 6.00318432e-01 2.28260253e-02
3.85375351e-01 -1.16110206e+00 -5.60121536e-02 1.65033376e+00
-6.48975551e-01 5.81562936e-01 1.33259165e+00 -8.54773045e-01
-1.34098804e+00 -5.24979711e-01 6.06745183e-01 5.29093862e-01
2.26634324e-01 1.74623862e-01 -9.70431089e-01 1.46559969e-01
6.67562127e-01 1.80558600e-02 3.15953672e-01 -7.10328162e-01
1.63661033e-01 -3.34354550e-01 -1.64932823e+00 2.76063204e-01
2.69589067e-01 -5.59650138e-02 -4.08628523e-01 6.02866672e-02
1.02290384e-01 1.26536367e-02 -5.35803020e-01 2.39028782e-01
1.90098494e-01 -9.31199968e-01 9.82317746e-01 2.48352084e-02
-3.55951488e-01 -8.68867457e-01 -2.60791127e-02 -8.72507095e-01
-6.25954628e-01 -6.81195438e-01 7.01487482e-01 1.17278528e+00
-4.10195254e-02 -3.56921822e-01 3.11343849e-01 1.91940099e-01
6.45162016e-02 1.38117820e-02 -7.03127086e-01 -6.09525561e-01
-5.30634344e-01 -2.97818601e-01 -2.41634369e-01 4.25432652e-01
-4.74203855e-01 -2.10377559e-01 -4.46740985e-01 3.12309057e-01
9.79195237e-01 -3.54008675e-01 1.65188640e-01 -9.82181787e-01
1.82728529e-01 -4.09076095e-01 -5.92695832e-01 -1.99353307e-01
-4.76539582e-01 5.77774942e-02 2.65301496e-01 -2.04228520e+00
-8.18502009e-02 3.32060754e-01 -5.25699139e-01 1.00999258e-01
-3.07380289e-01 9.22905803e-01 1.40617892e-01 9.30469036e-02
-6.97127134e-02 -9.88900736e-02 1.15177107e+00 -6.18199520e-02
-2.28282005e-01 4.73762862e-02 -3.57294053e-01 7.86429942e-01
6.82307124e-01 -3.77568424e-01 -1.34229749e-01 -1.64696574e-01
-3.27382475e-01 -1.54782355e-01 4.29257065e-01 -1.22890615e+00
1.68326542e-01 1.51423095e-02 6.91455483e-01 -5.48727393e-01
1.20002672e-01 -1.19820797e+00 2.05640703e-01 8.52835357e-01
1.22292697e-01 1.71715423e-01 3.85692775e-01 7.82752335e-01
-2.65504152e-01 -5.02925634e-01 1.15724933e+00 -2.56695181e-01
-1.22553515e+00 -7.00278997e-01 -1.29968309e+00 -7.13215828e-01
9.46172833e-01 -8.16349089e-01 -4.55626011e-01 -3.97665918e-01
-1.17631149e+00 -4.20425743e-01 3.25150073e-01 -2.77149796e-01
7.74806082e-01 -8.01437020e-01 -7.20961154e-01 2.30170995e-01
-7.90805161e-01 -7.73829103e-01 2.81729698e-01 9.28487897e-01
-1.31463861e+00 -2.42116630e-01 -5.56154907e-01 6.04856154e-03
-1.88533831e+00 4.64714140e-01 4.65798765e-01 2.77100056e-01
-1.01192141e+00 7.24726439e-01 -5.86940229e-01 5.45262635e-01
3.06896359e-01 -2.05972902e-02 -3.64662260e-01 -3.85607541e-01
7.62766719e-01 9.77648377e-01 -4.87080123e-03 -6.36610389e-01
-1.12861007e-01 8.59656036e-01 1.42383456e-01 -1.29158407e-01
1.25230277e+00 -6.55743301e-01 -9.26369548e-01 2.83045396e-02
9.73694026e-01 3.14251870e-01 -8.47271442e-01 -1.80802464e-01
-1.28607273e-01 -8.50445092e-01 3.78142208e-01 -9.95795786e-01
-1.10405540e+00 3.06695491e-01 1.35569632e+00 4.14864004e-01
1.97813320e+00 -8.55301797e-01 7.78784454e-01 3.89304191e-01
-1.66492805e-01 -1.45622122e+00 -2.96206415e-01 1.99253544e-01
4.85142291e-01 -1.19985688e+00 4.06452298e-01 -8.44910443e-01
1.06505342e-01 1.57304752e+00 3.11581105e-01 -5.52869797e-01
8.22464406e-01 4.00663972e-01 6.50145769e-01 -2.26686578e-02
-2.80367315e-01 -5.45029461e-01 -9.46756378e-02 8.29469800e-01
2.05759063e-01 -2.50612736e-01 -1.24562395e+00 7.94419721e-02
4.62758571e-01 5.16055301e-02 9.97302651e-01 1.36299372e+00
-1.18835497e+00 -9.18209612e-01 -1.27309740e+00 2.41466150e-01
-9.10113335e-01 -6.62285537e-02 -7.41524100e-02 8.69651377e-01
3.49717200e-01 1.64122260e+00 -2.52775520e-01 -3.41968685e-02
5.84922850e-01 -2.05649778e-01 2.91121304e-01 4.00108285e-02
-6.60426021e-01 6.95575833e-01 -1.60593852e-01 -2.27698591e-02
-6.95403159e-01 -4.34281409e-01 -1.08921778e+00 -3.90481383e-01
-1.49174988e-01 3.84865493e-01 9.70274031e-01 8.01301241e-01
-3.24183732e-01 5.77700436e-01 5.07423460e-01 -5.44932425e-01
-2.49946594e-01 -9.23338652e-01 -1.12684381e+00 6.72082424e-01
5.15455723e-01 -2.19244987e-01 -6.35286212e-01 4.30886626e-01] | [11.226486206054688, -2.5073060989379883] |
2d849643-ff2c-4f7d-bc8f-3bf6a87ac542 | otfs-a-mathematical-foundation-for | 2302.08696 | null | https://arxiv.org/abs/2302.08696v1 | https://arxiv.org/pdf/2302.08696v1.pdf | OTFS -- A Mathematical Foundation for Communication and Radar Sensing in the Delay-Doppler Domain | Orthogonal time frequency space (OTFS) is a framework for communication and active sensing that processes signals in the delay-Doppler (DD) domain. This paper explores three key features of the OTFS framework, and explains their value to applications. The first feature is a compact and sparse DD domain parameterization of the wireless channel, where the parameters map directly to physical attributes of the reflectors that comprise the scattering environment, and as a consequence these parameters evolve predictably. The second feature is a novel waveform / modulation technique, matched to the DD channel model, that embeds information symbols in the DD domain. The relation between channel inputs and outputs is localized, non-fading and predictable, even in the presence of significant delay and Doppler spread, and as a consequence the channel can be efficiently acquired and equalized. By avoiding fading, the post equalization SNR remains constant across all information symbols in a packet, so that bit error performance is superior to contemporary multi-carrier waveforms. Further, the OTFS carrier waveform is a localized pulse in the DD domain, making it possible to separate reflectors along both delay and Doppler simultaneously, and to achieve a high-resolution delay-Doppler radar image of the environment. In other words, the DD parameterization provides a common mathematical framework for communication and radar. This is the third feature of the OTFS framework, and it is ideally suited to intelligent transportation systems involving self-driving cars and unmanned ground/aerial vehicles which are self/network controlled. The OTFS waveform is able to support stable and superior performance over a wide range of user speeds. | ['Robert Calderbank', 'Ananthanarayanan Chockalingam', 'Ronny Hadani', 'Saif Khan Mohammed'] | 2023-02-17 | null | null | null | null | ['self-driving-cars'] | ['computer-vision'] | [ 5.72751164e-01 -1.03188969e-01 -3.33963215e-01 2.10894734e-01
-1.15676090e-01 -5.10433853e-01 6.84225380e-01 -2.73003995e-01
-4.26276147e-01 8.12818646e-01 -1.91832989e-01 -3.13479334e-01
-4.50708240e-01 -9.47923303e-01 -1.51747793e-01 -1.22566175e+00
-8.09760690e-01 1.21283438e-02 2.30884835e-01 -3.20480287e-01
-2.24498473e-02 9.88068759e-01 -1.69629455e+00 -5.66561401e-01
6.70998573e-01 1.42811286e+00 2.26175159e-01 7.82429397e-01
1.92641050e-01 1.33848205e-01 -6.91007853e-01 6.87355623e-02
4.15483654e-01 -2.41800025e-01 2.08740920e-01 -2.74730057e-01
-2.98418701e-02 -1.35142654e-01 -7.56670237e-01 7.70533919e-01
4.15064424e-01 -1.53193265e-01 8.50282311e-01 -1.15602994e+00
-7.21178353e-02 -6.29817545e-02 -3.37784469e-01 1.45165652e-01
3.18035275e-01 -1.48359627e-01 5.96014619e-01 -5.75771332e-01
3.63376528e-01 8.76003623e-01 6.38877988e-01 1.08689710e-01
-9.12532806e-01 -7.70063818e-01 -3.58982950e-01 2.33331531e-01
-1.54515040e+00 -4.69751447e-01 4.57261741e-01 -3.17105830e-01
4.73398775e-01 3.97493541e-01 7.82164752e-01 7.38607109e-01
7.75230110e-01 1.34882122e-01 8.62076104e-01 -6.77391052e-01
3.53606373e-01 -7.08992481e-02 -1.23458967e-01 4.76457894e-01
7.80917943e-01 9.47062373e-01 -5.05609512e-01 -1.75517276e-01
5.58605492e-01 -2.35324189e-01 -7.20116258e-01 -3.70297492e-01
-1.07007778e+00 7.10075557e-01 1.09184697e-01 4.79094118e-01
-6.80027246e-01 2.64598757e-01 -1.00885257e-01 6.14208817e-01
-3.45618539e-02 1.83599338e-01 1.20785877e-01 -1.02437355e-01
-1.04739642e+00 2.32355207e-01 9.84522164e-01 9.04967487e-01
7.79915929e-01 5.38817525e-01 -3.89089361e-02 2.88879097e-01
5.80837727e-01 1.73080504e+00 -3.52864452e-02 -7.19611526e-01
7.32035190e-02 -4.18967187e-01 2.38931805e-01 -1.09545362e+00
-6.44077182e-01 -1.01901531e+00 -9.37566757e-01 3.82454693e-01
1.75824195e-01 -3.86882842e-01 -6.26118958e-01 1.53887010e+00
2.33886808e-01 1.64787561e-01 5.46287894e-01 6.50084674e-01
1.68899685e-01 6.95788145e-01 -3.66206497e-01 -6.26678050e-01
1.34924555e+00 -9.47520062e-02 -9.74470258e-01 -4.09713179e-01
1.38507426e-01 -7.62461603e-01 5.67729548e-02 5.85607946e-01
-6.61491990e-01 -4.23059851e-01 -1.55469990e+00 9.20005977e-01
-1.99492827e-01 -7.53042921e-02 2.69482344e-01 1.18129539e+00
-7.34353542e-01 1.13782071e-01 -4.16058868e-01 1.32780477e-01
-7.75186643e-02 6.19469173e-02 4.04185951e-02 -4.36809361e-01
-1.71504045e+00 8.52903724e-01 3.27784270e-02 3.04481268e-01
-4.16377068e-01 -6.24768078e-01 -8.38020027e-01 -6.44632727e-02
-9.33716446e-02 -3.41036379e-01 9.85198617e-01 -5.27866960e-01
-1.45089364e+00 8.99164826e-02 -2.99018443e-01 -7.53524959e-01
7.24316910e-02 4.77120802e-02 -1.29609036e+00 5.31540096e-01
-1.74493358e-01 -1.36468440e-01 1.29258561e+00 -1.19235480e+00
-6.77652776e-01 -2.02664331e-01 -4.36988026e-01 -7.47685228e-03
2.70081107e-02 -4.66850996e-01 5.45872822e-02 -4.63441223e-01
3.16234350e-01 -1.12262905e+00 -3.07832837e-01 -7.16768876e-02
7.53152519e-02 6.12792552e-01 1.13150120e+00 -9.22304317e-02
1.34557128e+00 -2.36183929e+00 -3.61815900e-01 7.71203101e-01
5.75572923e-02 3.72312784e-01 1.55306254e-02 9.32054520e-01
1.68168202e-01 -5.12495875e-01 -2.14332342e-01 2.47462839e-01
-7.02302679e-02 2.54385769e-01 -5.93978226e-01 9.34633076e-01
-3.41514945e-02 2.83647776e-01 -6.88827515e-01 1.25848830e-01
1.89551705e-04 6.32038474e-01 -6.00978918e-02 -1.07230380e-01
2.80022442e-01 3.40192527e-01 -6.05252385e-01 5.69301903e-01
1.12228227e+00 4.65233624e-01 -9.25752446e-02 -1.78571180e-01
-7.47869492e-01 -2.10382611e-01 -1.30218589e+00 9.96961474e-01
-5.20425797e-01 1.03314149e+00 6.04775548e-01 -8.91397953e-01
1.37551999e+00 4.43168014e-01 6.62787259e-01 -1.23530507e+00
2.83073597e-02 4.49431121e-01 6.72303420e-03 -4.94574070e-01
6.29233718e-01 -3.83167028e-01 -9.70803872e-02 4.30283725e-01
-2.44911790e-01 -1.16575122e-01 8.79285783e-02 5.90943033e-03
1.15610099e+00 -6.42468333e-01 3.63145322e-01 -2.50849783e-01
5.74215412e-01 -2.93005079e-01 4.60365206e-01 7.34345138e-01
5.90245239e-02 -3.34603973e-02 -1.46953459e-03 2.45378107e-01
-5.26326537e-01 -1.12139904e+00 -4.48322207e-01 2.04358622e-01
5.90134203e-01 1.00989178e-01 -3.10408827e-02 1.91599131e-01
3.28909725e-01 6.30784869e-01 -1.97873354e-01 -3.11415315e-01
-4.00225252e-01 -4.09189135e-01 6.87457204e-01 -1.17177054e-01
5.23391128e-01 -2.38015980e-01 -9.14878666e-01 6.34212673e-01
1.62208125e-01 -1.37728941e+00 1.48061156e-01 2.66680151e-01
-5.20379901e-01 -1.02390897e+00 -6.45269215e-01 -2.03735158e-01
2.18588397e-01 8.37748528e-01 6.13004088e-01 -1.77495673e-01
-3.35745633e-01 8.85658622e-01 -5.78631401e-01 -6.03491604e-01
-1.56104058e-01 -5.36842346e-01 1.30885631e-01 4.06697899e-01
2.69397255e-02 -5.19195259e-01 -5.52380025e-01 2.97536850e-01
-7.64795184e-01 -5.56576371e-01 7.29306817e-01 5.41029334e-01
1.99433416e-01 5.89997590e-01 7.04664826e-01 -3.47392857e-01
4.19249833e-01 -6.32996917e-01 -8.59982014e-01 -2.15194821e-01
-5.50994039e-01 -2.15340614e-01 4.61957604e-01 -1.22218750e-01
-9.08321738e-01 -5.82175627e-02 3.07471910e-03 1.47048205e-01
2.97339708e-02 6.03255987e-01 -1.03308775e-01 -6.05162024e-01
6.62682354e-01 3.48373055e-01 3.55961144e-01 1.36254400e-01
1.72142118e-01 7.50962138e-01 4.42938507e-01 -8.03598017e-02
1.50267410e+00 8.68322015e-01 6.45847321e-01 -1.75939882e+00
-2.28220463e-01 -4.21697229e-01 -3.28027785e-01 -5.31857550e-01
3.65534604e-01 -8.70081425e-01 -5.80158472e-01 4.84580517e-01
-9.05351937e-01 -9.48443562e-02 -2.77044088e-01 1.27421451e+00
-4.40101922e-01 3.88618052e-01 -6.31337464e-02 -1.39678311e+00
1.28934290e-02 -6.24842763e-01 5.29622614e-01 3.19843352e-01
-1.08327650e-01 -1.07233202e+00 6.97509572e-03 -3.92910540e-01
9.05504048e-01 4.85127628e-01 6.93596661e-01 -2.89154891e-02
-8.01449418e-01 -5.05746305e-01 8.62623900e-02 4.59011570e-02
-5.24668321e-02 -7.30324611e-02 -7.23485410e-01 -5.08473039e-01
2.17668533e-01 4.54309076e-01 7.88216889e-01 6.79724276e-01
2.72943556e-01 -1.68629978e-02 -6.50016963e-01 5.53031147e-01
1.44399798e+00 5.84428787e-01 8.71740162e-01 -1.55160949e-02
-1.79772347e-01 3.23771656e-01 1.08059573e+00 5.83872318e-01
8.55579004e-02 7.10232079e-01 4.81911182e-01 -2.48022795e-01
-8.86306074e-03 3.54946554e-01 4.16830838e-01 3.12617958e-01
3.45497169e-02 -4.70158666e-01 -3.81232798e-01 1.49339765e-01
-1.13532758e+00 -1.20079350e+00 -4.17141527e-01 2.56565166e+00
2.06865147e-01 1.06207155e-01 -2.00702041e-01 4.77726042e-01
6.73941016e-01 2.78879553e-01 -2.70038188e-01 -3.59205425e-01
-2.70929247e-01 4.87566888e-01 1.26401830e+00 5.11420012e-01
-6.80106997e-01 -3.16342227e-02 6.30475140e+00 7.30222225e-01
-1.46913064e+00 -2.26964518e-01 -4.92512047e-01 3.01735699e-01
-5.33621609e-01 -1.31661758e-01 -8.33576083e-01 3.86372507e-01
1.15761471e+00 -4.04608071e-01 1.50083005e-01 1.91542715e-01
4.31022942e-01 -4.67604637e-01 -5.35830498e-01 7.89016366e-01
-5.52172884e-02 -1.14540541e+00 -4.75116491e-01 5.15099645e-01
2.79943347e-01 -1.47476584e-01 2.64890015e-01 -3.14940922e-02
-3.52993011e-01 -8.25801849e-01 7.71192849e-01 6.65047407e-01
1.07572353e+00 -7.81062722e-01 5.58503091e-01 3.80654961e-01
-1.43454313e+00 -4.25394684e-01 -1.45610228e-01 4.00357358e-02
6.43460095e-01 1.22283912e+00 -6.46839619e-01 8.95664036e-01
1.54852822e-01 2.93968916e-01 3.17739904e-01 1.30762041e+00
-3.97077240e-02 5.82574666e-01 -4.09541905e-01 -1.43215254e-01
2.08440185e-01 -2.05332473e-01 9.84941304e-01 1.31048357e+00
1.07453978e+00 2.91634083e-01 6.29654899e-02 2.78857559e-01
5.46931565e-01 -5.19068837e-01 -8.46695065e-01 -9.81459469e-02
1.08022058e+00 9.23609138e-01 -3.72916371e-01 1.75854802e-01
-3.34830344e-01 1.59412354e-01 -8.29395413e-01 7.45989799e-01
-7.43296742e-01 -9.33030427e-01 9.17630017e-01 2.60329813e-01
6.59000099e-01 -8.78232181e-01 9.09616575e-02 -4.22211945e-01
-3.11140507e-01 -5.45295537e-01 -5.61172739e-02 -1.37571797e-01
-8.80807102e-01 4.63679314e-01 -1.09572746e-01 -1.85694909e+00
-1.92973778e-01 -4.94054645e-01 -3.33202451e-01 7.83693194e-01
-2.14464641e+00 -6.56003535e-01 -3.82395506e-01 7.00627685e-01
-3.01615536e-01 -3.71651351e-01 1.05468249e+00 5.65563560e-01
4.66556475e-02 2.33823851e-01 5.05065322e-01 -3.31438780e-01
4.44610447e-01 -6.71550214e-01 -1.14291860e-02 8.62568080e-01
-2.19457030e-01 4.98143524e-01 1.03036654e+00 -5.85037112e-01
-1.90431380e+00 -1.07693291e+00 5.99510849e-01 4.95466053e-01
4.52326715e-01 -2.09535837e-01 -4.35129106e-01 7.99644925e-03
6.55913800e-02 -9.81315374e-02 9.12618935e-01 -3.30311924e-01
-2.65078396e-01 -4.02164161e-01 -1.15813982e+00 1.67783827e-01
5.02739668e-01 -2.48418421e-01 -2.36128241e-01 8.27953368e-02
1.98581100e-01 -3.34347099e-01 -5.79381287e-01 4.25318420e-01
8.36932957e-01 -9.66555119e-01 1.07387042e+00 4.47218984e-01
-4.35789257e-01 -5.36816120e-01 -4.97262567e-01 -1.36504114e+00
-4.26350296e-01 -9.91623700e-01 -1.96174100e-01 8.51683736e-01
3.13885957e-01 -1.39809918e+00 4.36933815e-01 -2.79484332e-01
-1.86656505e-01 -3.99547279e-01 -1.35166490e+00 -1.22864795e+00
-4.33072478e-01 -5.15001893e-01 4.52170312e-01 4.15866286e-01
-1.93361118e-01 3.91395018e-02 -3.76303732e-01 7.74638832e-01
1.10440099e+00 1.08418599e-01 5.60584664e-01 -1.60010111e+00
-2.41214424e-01 -1.22974947e-01 -5.21811545e-01 -1.40700591e+00
-3.95447686e-02 -5.44224381e-01 2.02457353e-01 -1.17598510e+00
-9.77414727e-01 -1.05437219e+00 1.57748610e-01 -4.47938740e-01
7.42538095e-01 2.59372175e-01 8.93536396e-03 2.38051414e-02
3.27297598e-01 6.52277172e-01 1.14446092e+00 2.46080965e-01
-5.20663671e-02 7.14748740e-01 -3.29294056e-01 3.43882829e-01
6.55635715e-01 -2.88292170e-01 -5.12002826e-01 -1.02769136e-02
3.98413800e-02 4.16865349e-01 3.89168739e-01 -1.43384433e+00
5.32295525e-01 -1.86384678e-01 2.11413220e-01 -6.16543591e-01
8.69177043e-01 -1.33607292e+00 4.61643070e-01 9.61048484e-01
3.08564365e-01 -4.79064941e-01 -4.89425138e-02 1.04324234e+00
-3.32277000e-01 -8.09216797e-02 1.04367757e+00 3.53448153e-01
-8.29090476e-01 2.50102609e-01 -9.58343863e-01 -1.79268643e-01
1.24543881e+00 -7.66809165e-01 -5.19160092e-01 -7.17062354e-01
-2.09347963e-01 -2.18035467e-02 2.08831839e-02 -4.49798889e-02
5.13448119e-01 -1.15945888e+00 -6.31719351e-01 7.12262034e-01
-7.99654871e-02 -6.49382055e-01 4.21590507e-01 9.98112798e-01
-3.20124060e-01 8.70737553e-01 -2.37003565e-01 -7.01948166e-01
-1.01894736e+00 -1.61392838e-01 2.43982702e-01 2.56419927e-01
-5.45918703e-01 5.15544593e-01 -2.22166270e-01 2.82585621e-01
5.03396094e-02 1.16389230e-01 -2.80908138e-01 2.35741854e-01
9.58806396e-01 3.14226389e-01 2.53083169e-01 -7.26176083e-01
-2.75859296e-01 8.35015893e-01 4.88840193e-01 -2.55228639e-01
1.01773763e+00 -4.34714854e-01 6.95346221e-02 8.37367401e-02
9.26397979e-01 5.73586285e-01 -9.84331787e-01 -4.06660855e-01
-3.21612418e-01 -6.31187439e-01 3.59642148e-01 -3.78733933e-01
-8.54332149e-01 4.98004615e-01 5.44495404e-01 8.25043082e-01
1.22305048e+00 -4.33102310e-01 7.65107632e-01 2.20269203e-01
6.85247958e-01 -9.12074268e-01 -1.83037192e-01 8.57336640e-01
5.79911530e-01 -1.60819054e-01 -6.77726883e-03 -6.02001190e-01
-1.59526557e-01 1.45550120e+00 -3.09672296e-01 4.59359288e-02
9.11859393e-01 8.14653754e-01 6.32972866e-02 -1.79552913e-01
-3.89807642e-01 -4.46716785e-01 -3.05111073e-02 1.22234344e+00
5.11783883e-02 1.66084155e-01 -3.91419321e-01 1.38134703e-01
-3.25942755e-01 -2.55885452e-01 7.88870931e-01 9.56031501e-01
-1.22336757e+00 -1.11150861e+00 -8.63715291e-01 3.08713108e-01
2.82835122e-02 2.10040361e-01 3.69830310e-01 9.32954013e-01
-9.20868386e-03 1.20645261e+00 1.23670064e-01 -2.98201740e-01
5.46434224e-01 -3.17054570e-01 3.60084385e-01 -1.89494997e-01
3.10736269e-01 1.05149057e-02 3.79133582e-01 -4.36743498e-01
-3.15192193e-01 -7.20774949e-01 -1.36589217e+00 -3.18346322e-01
-3.54233831e-01 5.66771686e-01 8.79028678e-01 8.25968862e-01
2.14008451e-01 4.94220644e-01 1.21256268e+00 -7.38578618e-01
-5.75803638e-01 -1.74500600e-01 -1.32135189e+00 -6.24385655e-01
8.54296386e-01 -9.98960674e-01 -5.94687641e-01 -3.80916625e-01] | [6.4268479347229, 1.246323823928833] |
6077e26f-858d-4e29-b984-386c08be6aa3 | elucidation-of-molecular-targets-of-bioactive | 1506.02008 | null | http://arxiv.org/abs/1506.02008v1 | http://arxiv.org/pdf/1506.02008v1.pdf | Elucidation of molecular targets of bioactive principles of black cumin relevant to its anti-tumour functionality - An Insilico target fishing approach | Black cumin (Nigella sativa) is a spice having medicinal properties with
pungent and bitter odour. It is used since thousands of years to treat various
ailments, including cancer mainly in South Asia and Middle Eastern regions.
Substantial evidence in multiple research studies emphasizes about the
therapeutic importance of bioactive principles of N. sativa in cancer
bioassays; however, the exact mechanism of their anti-tumour action is still to
be fully comprehended. The current study makes an attempt in this direction by
exploiting the advancements in the Insilico reverse screening technology. In
this study, three different Insilico Reverse Screening approaches have been
employed for identifying the putative molecular targets of the bioactive
principles in Black cumin (thymoquinone, alpha-hederin, dithymoquinone and
thymohydroquinone) relevant to its anti-tumour functionality. The identified
set of putative targets is further compared with the existing set of
experimentally validated targets, so as to estimate the performance of insilico
platforms. Subsequently, molecular docking simulations studies were performed
to elucidate the molecular interactions between the bioactive compounds & their
respective identified targets. The molecular interactions of one such target
identified i.e. VEGF2 along with thymoquinone depicted one H-bond formed at the
catalytic site. The molecular targets identified in this study need further
confirmatory tests on cancer bioassays, in order to justify the research
findings from Insilico platforms. This study has brought to light the
effectiveness of usage of Insilico Reverse Screening protocols to characterise
the un-identified target-ome of poly pharmacological bioactive agents in
spices. | [] | 2015-02-20 | null | null | null | null | ['molecular-docking'] | ['medical'] | [ 4.91355807e-01 4.32469733e-02 -4.75407988e-01 3.43707889e-01
-2.33722433e-01 -8.92418146e-01 5.97552896e-01 8.25613022e-01
-2.78466105e-01 1.28889894e+00 4.89393435e-02 -3.83815646e-01
-7.49297142e-02 -6.01765692e-01 -2.33157098e-01 -1.11024261e+00
-4.93518114e-01 1.86865926e-01 7.95904454e-03 -4.81382549e-01
6.67372942e-01 9.71575975e-01 -9.87661242e-01 -1.16071455e-01
1.33164263e+00 5.06560087e-01 4.23903048e-01 3.63753468e-01
8.87743896e-04 3.56554449e-01 -5.11071920e-01 -3.77392471e-01
-3.16597596e-02 -4.04227048e-01 -4.87739712e-01 -4.53918636e-01
-2.26458773e-01 -5.14280535e-02 4.49611545e-01 9.08436775e-01
5.62547624e-01 -2.01504938e-02 8.41714203e-01 -7.64193177e-01
-4.19512838e-01 1.35579601e-01 -1.44494250e-01 1.21752501e-01
6.08884871e-01 -2.46252716e-02 6.28440738e-01 -9.90410566e-01
5.84601820e-01 8.85927737e-01 7.33490959e-02 4.44296896e-01
-8.44465375e-01 -7.72083640e-01 -5.30001760e-01 2.32553080e-01
-1.28956497e+00 -1.33925527e-01 4.04903382e-01 -2.80723482e-01
7.34540403e-01 5.94861746e-01 8.35601926e-01 6.27260566e-01
9.16242540e-01 3.64353985e-01 1.56533587e+00 -1.78093940e-01
4.39339399e-01 4.97951627e-01 -1.00171015e-01 4.72079426e-01
6.29301548e-01 2.64308482e-01 -6.73247516e-01 -1.97377637e-01
3.13631743e-02 -4.60181618e-03 -4.18565065e-01 -7.94268623e-02
-3.64935040e-01 8.65341663e-01 6.74858153e-01 6.15389466e-01
-3.65727723e-01 -2.48538449e-01 5.06116629e-01 -3.41351688e-01
3.94228131e-01 5.29389143e-01 -6.59553766e-01 8.02397206e-02
-2.21646413e-01 -3.16571862e-01 8.66734326e-01 2.22300097e-01
4.10784811e-01 -9.48641077e-03 5.01141310e-01 3.91158819e-01
5.26027560e-01 3.23967755e-01 4.16253269e-01 9.48918797e-03
-2.43890196e-01 5.95408261e-01 6.01375662e-02 -7.39667058e-01
-6.22634947e-01 -2.61384130e-01 -4.11750734e-01 1.32794827e-01
3.04240555e-01 -2.15790778e-01 -4.51187044e-01 1.35560036e+00
6.62453592e-01 -4.09341529e-02 2.50777245e-01 7.80759215e-01
8.37761641e-01 6.29054904e-01 8.94771636e-01 -4.36868548e-01
1.64005280e+00 -4.58677709e-01 -4.20275599e-01 5.82180858e-01
5.83850205e-01 -9.88651395e-01 4.31823641e-01 6.36372030e-01
-7.09518671e-01 -9.88692567e-02 -1.00003481e+00 4.31178898e-01
-7.96990514e-01 1.18375728e-02 9.06045914e-01 1.25059128e+00
-3.30829084e-01 7.98234999e-01 -8.63810837e-01 -4.94642466e-01
2.04120278e-01 7.00992584e-01 -5.16419232e-01 1.15033798e-01
-1.33125257e+00 1.09369743e+00 6.94853187e-01 2.52911776e-01
-1.03253043e+00 -7.55336761e-01 -5.42849481e-01 -3.85828130e-02
1.96562544e-01 -4.13668782e-01 5.48624694e-01 -8.07671487e-01
-1.77242982e+00 4.92395312e-01 -4.24591005e-02 -4.16544080e-01
-4.14743535e-02 -2.37360433e-01 -6.28724575e-01 4.71720606e-01
8.94744843e-02 2.64725715e-01 -1.28926054e-01 -1.14908588e+00
-4.51024294e-01 -6.14787400e-01 -6.25843331e-02 1.85390592e-01
-2.93602198e-01 1.00387998e-01 4.05773699e-01 -5.55688441e-01
-1.53298378e-01 -8.94935846e-01 -3.93026441e-01 -2.94999540e-01
-7.27545083e-01 -1.03997111e-01 6.39239371e-01 -5.07626057e-01
6.69912815e-01 -1.66052699e+00 -5.96035793e-02 4.12250042e-01
5.16987741e-02 4.26959038e-01 1.35002792e-01 1.32593000e+00
-1.87808275e-01 6.14538550e-01 1.86790928e-01 8.70997131e-01
-5.41958272e-01 -3.50377649e-01 -1.91858448e-02 8.79689574e-01
2.71895379e-01 4.26910639e-01 -1.12736189e+00 -2.95223624e-01
1.82701245e-01 9.43404377e-01 3.42900045e-02 -4.95785207e-01
-1.55236363e-01 5.81490815e-01 -1.10563612e+00 1.17589533e+00
8.37960243e-01 3.50272864e-01 5.04279315e-01 -2.89521843e-01
-4.49604154e-01 2.46757314e-01 -5.13028920e-01 9.68404174e-01
1.01746153e-03 1.16727069e-01 -4.75502163e-01 -4.73199844e-01
9.07489598e-01 5.24069309e-01 5.97108424e-01 -9.23009574e-01
2.20614329e-01 5.18049002e-01 1.81581572e-01 -2.73224264e-01
2.68399745e-01 -5.86632371e-01 3.97834510e-01 -4.79269028e-01
-1.35521501e-01 4.03523356e-01 4.68982100e-01 9.36683118e-02
6.38148665e-01 3.90931964e-01 6.79951251e-01 -8.45621467e-01
9.76746380e-01 2.55970895e-01 2.58230090e-01 -1.60158500e-01
2.53060814e-02 -2.92946756e-01 5.56731641e-01 -2.22577751e-01
-4.80247825e-01 -8.73933196e-01 -6.95052385e-01 6.41912282e-01
2.87755102e-01 -4.71602559e-01 -6.28724992e-01 -2.67023861e-01
-3.91103059e-01 7.17033565e-01 -7.74303138e-01 -1.99505717e-01
-4.17239405e-02 -8.37246478e-01 6.53814197e-01 -6.73787147e-02
4.72872496e-01 -6.23712301e-01 -5.34304202e-01 4.46577668e-01
5.94311178e-01 -4.09184366e-01 1.28601894e-01 5.02374649e-01
-9.21167672e-01 -1.65722167e+00 -7.71116555e-01 -2.81041473e-01
3.18263233e-01 2.64704466e-01 4.29335982e-01 1.65101122e-02
-2.40525439e-01 -3.16144168e-01 -6.28582835e-01 -1.08514202e+00
-5.36920130e-01 -2.27627993e-01 -9.46722850e-02 -4.19171095e-01
3.19363505e-01 -3.59692574e-01 -8.58685076e-01 3.58537227e-01
-7.94317961e-01 -4.38427001e-01 7.91282892e-01 5.79663515e-01
5.05079031e-01 6.34486601e-03 8.34659874e-01 -1.27018714e+00
5.51623404e-01 -9.03293610e-01 -4.67339963e-01 4.08627577e-02
-4.70482677e-01 -1.60688370e-01 7.51874864e-01 -3.99868786e-01
-9.19867873e-01 -1.03926182e-01 -2.11008996e-01 4.77109522e-01
-6.68875799e-02 1.14044559e+00 -5.40279806e-01 -3.46077859e-01
5.36843538e-01 6.33812919e-02 1.22258030e-01 -3.80884826e-01
-2.13873833e-01 1.85602009e-01 -5.68592697e-02 -5.47586262e-01
6.57103956e-01 5.82990050e-01 7.03927696e-01 -1.18611169e+00
-4.62046653e-01 -6.53147578e-01 -2.01160848e-01 -1.05501398e-01
1.04415309e+00 -8.26085567e-01 -1.14534974e+00 1.16257191e-01
-8.62850249e-01 3.65991257e-02 4.21525538e-01 7.05189168e-01
-3.96902747e-02 4.43663418e-01 -4.05547023e-01 -7.19510674e-01
-8.00623059e-01 -9.79603469e-01 4.03726310e-01 5.81234694e-01
-9.18744579e-02 -1.09508491e+00 4.66666579e-01 3.73563081e-01
7.43768811e-02 1.01767778e+00 1.19289279e+00 -8.58461440e-01
-3.62261444e-01 -3.02247375e-01 2.03190416e-01 -2.19319865e-01
2.04999089e-01 3.01052868e-01 -6.99357152e-01 -1.01000302e-01
-2.11608037e-01 -2.10812807e-01 3.70971441e-01 2.60273010e-01
4.77728486e-01 -3.73713151e-02 -5.92437148e-01 2.61740655e-01
1.74607801e+00 7.69672155e-01 9.51297104e-01 7.04174042e-01
1.88282475e-01 5.21277964e-01 1.09405446e+00 4.65675354e-01
-3.74392450e-01 4.13918614e-01 7.89470077e-01 -3.02013338e-01
2.47932434e-01 -3.69295701e-02 4.85562742e-01 -1.23033129e-01
-6.27571344e-01 -2.69568324e-01 -6.10597193e-01 1.43925130e-01
-8.89088631e-01 -9.08088386e-01 -7.91243374e-01 2.46751046e+00
8.58927071e-01 -4.70042899e-02 1.90817222e-01 -1.78305522e-01
3.84395868e-01 -4.58070099e-01 -4.15969759e-01 -8.03721786e-01
-2.72850186e-01 9.01459098e-01 6.86408699e-01 3.48619252e-01
-7.03204572e-01 6.55343056e-01 5.39767790e+00 7.60500789e-01
-1.31583631e+00 -4.64120656e-01 4.77481961e-01 2.36681223e-01
-1.90218627e-01 7.02035367e-01 -7.18865335e-01 3.96915555e-01
1.04994380e+00 -4.01996136e-01 -1.87621400e-01 5.37129939e-01
7.41077065e-01 -8.20555508e-01 -7.49975920e-01 1.15128152e-01
-2.62413681e-01 -1.42819464e+00 -9.56699252e-02 6.31301045e-01
4.68178689e-01 -1.13593891e-01 -2.41950572e-01 -1.32471800e-01
-2.18407854e-01 -1.04019368e+00 4.65198249e-01 3.90230834e-01
5.24460852e-01 -1.28327489e+00 6.86633289e-01 1.90262616e-01
-1.02489257e+00 2.69803908e-02 -3.08198810e-01 4.04400080e-02
-7.02538863e-02 5.44286966e-01 -1.17434847e+00 7.94417799e-01
2.16272816e-01 3.30174059e-01 -7.73270309e-01 1.20482635e+00
-4.13480431e-01 4.18529123e-01 -1.60078689e-01 -5.75385928e-01
2.26362616e-01 -5.74859321e-01 3.38102937e-01 6.96205854e-01
5.84014952e-02 1.04107864e-01 6.59964532e-02 5.44091344e-01
2.32828006e-01 8.11002970e-01 -5.15127599e-01 -2.76587307e-01
2.82806009e-01 1.30896592e+00 -1.13264680e+00 -1.99895233e-01
-2.80024558e-01 5.29219747e-01 -2.37832844e-01 2.02691555e-01
-5.14518321e-01 -2.72174567e-01 5.49738467e-01 2.46579036e-01
-1.10288575e-01 9.14670229e-02 7.01384172e-02 -5.50156236e-01
-7.27497280e-01 -7.51543522e-01 4.99003023e-01 -2.99761593e-01
-6.72253251e-01 6.87532723e-02 1.89201072e-01 -9.97611403e-01
5.30908406e-01 -5.31705260e-01 -8.45066547e-01 1.13549769e+00
-1.47170341e+00 -1.19706190e+00 6.81803674e-02 7.88244903e-02
3.02182943e-01 2.90756315e-01 1.24380755e+00 -2.81138659e-01
-7.25583851e-01 -1.72542050e-01 5.10312498e-01 -5.85799396e-01
7.19586730e-01 -1.01146924e+00 -5.54899812e-01 3.31798136e-01
-4.41772312e-01 9.82560992e-01 9.88722563e-01 -9.42724109e-01
-1.39074111e+00 -6.85904145e-01 6.41495049e-01 -3.93198915e-02
6.90149784e-01 7.17920959e-02 -3.71845394e-01 1.13389827e-01
4.02711987e-01 -6.89322650e-01 1.66775870e+00 -7.11710528e-02
6.58864528e-02 4.78456497e-01 -1.19075525e+00 6.92972183e-01
7.65471980e-02 -1.15647778e-01 8.70219171e-02 5.22959769e-01
3.15569378e-02 -1.50590017e-01 -1.41883028e+00 8.34198594e-02
7.06586897e-01 -8.76942992e-01 9.62983966e-01 -6.21554255e-01
6.24635339e-01 -5.62847078e-01 2.58855999e-01 -9.55866158e-01
6.46355376e-02 -5.50130844e-01 1.76914662e-01 8.64219368e-01
3.63206536e-01 -7.52973676e-01 7.51125515e-01 4.02313650e-01
-9.05009359e-02 -9.70211864e-01 -8.15995097e-01 -4.62368637e-01
2.71639228e-01 6.00215256e-01 1.19785182e-01 8.12011421e-01
2.61051387e-01 6.40552342e-01 -4.43375856e-02 4.61194701e-02
3.75378788e-01 -1.21915944e-01 6.35302842e-01 -1.37082267e+00
-1.86358523e-02 -3.92364599e-02 -3.65892500e-01 1.46061078e-01
-1.16222918e-01 -7.27654934e-01 -9.91305888e-01 -1.14438188e+00
2.69301832e-01 -8.38518292e-02 -8.23729262e-02 1.31965011e-01
-1.45400107e-01 2.05696389e-01 1.51912749e-01 7.55692124e-02
1.20141350e-01 1.60945967e-01 1.16510010e+00 -6.51488081e-02
-6.04674935e-01 -8.10430199e-02 -1.19896150e+00 5.95466554e-01
1.21241820e+00 -4.85707462e-01 -4.99983341e-01 7.10611820e-01
3.42971563e-01 8.26251283e-02 8.98618773e-02 -5.06982088e-01
-1.90548837e-01 -6.03901684e-01 5.65561116e-01 -4.79997039e-01
2.46425882e-01 -8.63238931e-01 6.12543702e-01 1.27329385e+00
3.07297617e-01 -6.11699760e-01 3.62437516e-01 5.18280566e-01
5.52266240e-02 -6.06690109e-01 8.07879090e-01 -1.97000325e-01
-5.75985014e-01 -1.04600802e-01 -9.43383336e-01 -6.16920531e-01
1.29743516e+00 -8.67416263e-01 -3.38235408e-01 1.91476226e-01
-7.06901670e-01 -1.26664206e-01 5.61734140e-01 -2.15111837e-01
4.19565588e-01 -7.85266340e-01 -4.49392319e-01 -3.85741353e-01
3.38928610e-01 -6.12843633e-01 3.96740675e-01 1.36459470e+00
-1.03380406e+00 6.24271154e-01 -6.82633877e-01 -1.41143024e-01
-1.63126791e+00 9.13547993e-01 3.09221268e-01 -2.68204987e-01
-1.62524328e-01 3.20430398e-01 3.08246315e-01 2.82250851e-01
-2.85498857e-01 1.59509674e-01 -7.30816007e-01 1.69564679e-01
4.60137069e-01 6.32233918e-01 2.69584388e-01 -6.06779873e-01
-4.97316331e-01 2.77579963e-01 -1.74479082e-01 5.82040787e-01
1.34257627e+00 2.34386414e-01 -2.83809036e-01 -3.28354305e-04
1.12802410e+00 4.92029190e-01 -6.81373179e-01 6.63931370e-01
6.40854314e-02 -3.55775326e-01 -7.45142400e-02 -1.07465589e+00
-4.24987376e-01 4.67342257e-01 7.04501927e-01 -9.74835921e-03
9.89384770e-01 -1.39768317e-01 3.17206740e-01 1.50662005e-01
2.73916721e-01 -8.71269286e-01 3.01665086e-02 -1.64904892e-01
9.70804811e-01 -7.89600015e-01 2.58789331e-01 -1.03506529e+00
-3.22940946e-01 1.43537962e+00 1.79178417e-01 4.87031735e-04
4.32333529e-01 -1.40791073e-01 -1.23950280e-01 -4.67267513e-01
-5.66658914e-01 -3.78447808e-02 1.51389524e-01 6.66117668e-01
1.17280972e+00 1.46816447e-01 -1.61043465e+00 1.33427233e-01
6.03680983e-02 -1.69602349e-01 7.51569927e-01 1.04679620e+00
-5.26302993e-01 -1.70391810e+00 -5.92043817e-01 1.53770775e-01
-9.03999746e-01 -3.18882674e-01 -9.41188812e-01 1.23805296e+00
8.01614858e-03 9.60733056e-01 -8.28779519e-01 8.85358453e-02
1.13932580e-01 -1.15527324e-01 4.08116639e-01 -3.07620406e-01
-8.68609190e-01 7.95520365e-01 5.25778651e-01 -1.18279278e-01
-7.80428529e-01 -4.79617298e-01 -1.43835473e+00 -2.26849228e-01
-9.79178786e-01 6.39467895e-01 1.02590835e+00 8.87714982e-01
-8.53945389e-02 2.05579206e-01 8.24313760e-01 -5.73494673e-01
-4.09440510e-02 -6.75739884e-01 -8.59377444e-01 1.69085175e-01
-3.63685101e-01 -5.93245268e-01 -8.22724029e-02 1.52350470e-01] | [4.660658836364746, 5.111394882202148] |
af2b6750-104c-49f9-87e8-afa87a847262 | sdd-fiqa-unsupervised-face-image-quality | 2103.05977 | null | https://arxiv.org/abs/2103.05977v1 | https://arxiv.org/pdf/2103.05977v1.pdf | SDD-FIQA: Unsupervised Face Image Quality Assessment with Similarity Distribution Distance | In recent years, Face Image Quality Assessment (FIQA) has become an indispensable part of the face recognition system to guarantee the stability and reliability of recognition performance in an unconstrained scenario. For this purpose, the FIQA method should consider both the intrinsic property and the recognizability of the face image. Most previous works aim to estimate the sample-wise embedding uncertainty or pair-wise similarity as the quality score, which only considers the information from partial intra-class. However, these methods ignore the valuable information from the inter-class, which is for estimating to the recognizability of face image. In this work, we argue that a high-quality face image should be similar to its intra-class samples and dissimilar to its inter-class samples. Thus, we propose a novel unsupervised FIQA method that incorporates Similarity Distribution Distance for Face Image Quality Assessment (SDD-FIQA). Our method generates quality pseudo-labels by calculating the Wasserstein Distance (WD) between the intra-class similarity distributions and inter-class similarity distributions. With these quality pseudo-labels, we are capable of training a regression network for quality prediction. Extensive experiments on benchmark datasets demonstrate that the proposed SDD-FIQA surpasses the state-of-the-arts by an impressive margin. Meanwhile, our method shows good generalization across different recognition systems. | ['Yuan-Gen Wang', 'Liujuan Cao', 'Yong Li', 'Jilin Li', 'Shaoxin Li', 'Yuge Huang', 'Ruixin Zhang', 'Xingyu Chen', 'Fu-Zhao Ou'] | 2021-03-10 | null | http://openaccess.thecvf.com//content/CVPR2021/html/Ou_SDD-FIQA_Unsupervised_Face_Image_Quality_Assessment_With_Similarity_Distribution_Distance_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Ou_SDD-FIQA_Unsupervised_Face_Image_Quality_Assessment_With_Similarity_Distribution_Distance_CVPR_2021_paper.pdf | cvpr-2021-1 | ['face-image-quality', 'face-image-quality-assessment'] | ['computer-vision', 'computer-vision'] | [ 1.02492087e-01 -2.35103458e-01 3.15285325e-02 -8.88393164e-01
-7.37251341e-01 -1.26089454e-01 4.41806346e-01 -2.72708952e-01
1.24344155e-02 3.41913819e-01 -1.90904364e-03 1.59819931e-01
-4.67242718e-01 -8.19691420e-01 -3.23944896e-01 -9.63685215e-01
1.65357247e-01 2.75887370e-01 -2.19776213e-01 6.63771182e-02
1.93757355e-01 6.22806191e-01 -1.78859270e+00 1.98462263e-01
1.03972912e+00 1.67717421e+00 -2.13441640e-01 1.88176185e-01
-5.35187908e-02 4.50297266e-01 -5.76160669e-01 -6.45034432e-01
3.93062294e-01 -6.41627192e-01 -4.56957698e-01 2.90988564e-01
6.53967202e-01 -3.67976069e-01 -2.86954910e-01 1.38308966e+00
4.65829760e-01 -1.22602932e-01 8.99828434e-01 -1.62850845e+00
-8.17104936e-01 -4.60451134e-02 -4.21910673e-01 7.37865642e-02
3.62553924e-01 2.30526030e-01 8.56151223e-01 -1.07597601e+00
2.05766290e-01 1.36636126e+00 4.01966870e-01 6.12063050e-01
-1.10052431e+00 -7.99897015e-01 -2.06282750e-01 4.80100572e-01
-1.54400539e+00 -6.85959101e-01 9.32370722e-01 -4.90931034e-01
3.35543036e-01 1.54651731e-01 3.43445390e-01 6.74635708e-01
1.60624608e-01 5.52714765e-01 1.34619713e+00 -1.92664310e-01
3.52146238e-01 1.20620787e-01 -1.18517920e-01 8.16516876e-01
5.38785905e-02 3.13908279e-01 -5.63621342e-01 -6.81659877e-02
5.84229589e-01 -2.07862332e-02 -2.89470017e-01 -5.04423916e-01
-8.26966703e-01 5.31117201e-01 3.52578282e-01 3.03400606e-01
-2.65702248e-01 -3.27281892e-01 1.58658013e-01 5.21900535e-01
4.54861164e-01 -8.18319172e-02 -2.62781531e-02 2.28471011e-02
-1.03081810e+00 -2.45221093e-01 5.19326389e-01 5.93532622e-01
8.04742515e-01 6.34881184e-02 -3.60850006e-01 9.24654961e-01
6.07185066e-01 7.85365403e-01 4.50022608e-01 -9.56529975e-01
2.57158697e-01 7.01364458e-01 2.97295209e-02 -1.47155178e+00
-2.14321986e-02 -4.04968381e-01 -1.03636158e+00 4.99152064e-01
5.41788995e-01 3.13842237e-01 -7.72178650e-01 1.67419279e+00
4.60387915e-01 3.36869717e-01 5.92887178e-02 9.10044432e-01
7.68506408e-01 6.32979751e-01 -7.59667009e-02 -6.49438143e-01
1.06410909e+00 -5.99179506e-01 -8.16642165e-01 1.46955937e-01
1.19735762e-01 -7.61161268e-01 1.00536978e+00 6.11156523e-01
-9.55304384e-01 -9.17787015e-01 -1.09453583e+00 4.69004750e-01
1.04455620e-01 3.29015195e-01 2.12531343e-01 1.01127076e+00
-1.03787887e+00 6.21306360e-01 -5.26679695e-01 1.21497400e-02
6.21627331e-01 4.23110783e-01 -7.81770051e-01 -2.75335312e-01
-9.55573499e-01 6.76390946e-01 6.71267733e-02 2.27723658e-01
-1.03105080e+00 -6.02420151e-01 -6.81662023e-01 6.36926517e-02
1.89900488e-01 -2.17769817e-01 7.48954237e-01 -1.21570325e+00
-1.63112915e+00 7.15747714e-01 -2.33066693e-01 1.23016581e-01
3.24045360e-01 3.16093326e-01 -8.91137123e-01 2.92649597e-01
-7.34903589e-02 3.96097660e-01 1.27044046e+00 -1.32084620e+00
-4.65371728e-01 -7.34088063e-01 -2.12401986e-01 2.22332347e-02
-5.81810355e-01 -8.97863433e-02 -3.62057000e-01 -4.39520776e-01
3.63977760e-01 -4.55242187e-01 4.91874754e-01 4.67119098e-01
-1.48639992e-01 -5.14652073e-01 7.54093885e-01 -5.52714050e-01
1.07947385e+00 -2.37134194e+00 8.54645893e-02 4.16780770e-01
1.26501083e-01 4.55532342e-01 -5.05274236e-01 6.00967631e-02
-7.82593042e-02 -9.78456587e-02 -3.58564407e-01 -1.75323859e-01
8.33241493e-02 2.22342219e-02 -3.73177715e-02 6.42194510e-01
6.15929604e-01 5.38621128e-01 -8.19501400e-01 -6.64289117e-01
9.21861157e-02 6.57463610e-01 -3.87108862e-01 5.50824642e-01
2.50163525e-01 3.91192734e-01 -3.47084999e-01 8.34052205e-01
1.16146171e+00 -1.19568529e-02 8.33984613e-02 -6.11960649e-01
4.18574661e-01 -1.97808847e-01 -1.31209123e+00 1.31146085e+00
-2.88565516e-01 2.53529936e-01 -7.82246366e-02 -1.01332223e+00
1.22393966e+00 2.29450390e-01 5.22825122e-01 -9.50810611e-01
1.37365013e-01 3.45820934e-01 1.25311106e-01 -4.72515702e-01
-9.61436108e-02 -3.60234648e-01 4.49585468e-01 4.81012315e-01
2.91806608e-01 8.27412084e-02 4.10679206e-02 -6.80000186e-02
7.73308694e-01 -1.49120361e-01 1.48913234e-01 -2.94626594e-01
1.01078784e+00 -9.06762362e-01 6.87742293e-01 1.29609674e-01
-7.33513474e-01 6.68953657e-01 5.63164949e-01 -2.59999812e-01
-8.45485210e-01 -1.29587269e+00 -4.75603521e-01 4.84366983e-01
3.23912233e-01 -1.34544969e-01 -8.90028119e-01 -1.07392418e+00
1.46826744e-01 1.52406588e-01 -6.76439643e-01 -5.53886294e-01
-9.78996009e-02 -4.54844117e-01 4.68473107e-01 2.71556854e-01
6.63299203e-01 -9.52176213e-01 8.39895606e-02 -1.27003938e-01
-5.91155775e-02 -7.88814723e-01 -5.70863366e-01 -5.03178418e-01
-5.85807323e-01 -1.08229625e+00 -6.94099724e-01 -6.69637501e-01
8.45358670e-01 2.04039276e-01 8.72893155e-01 2.69135803e-01
-9.58630964e-02 3.16983968e-01 -2.86240309e-01 1.46471381e-01
-3.31278712e-01 -5.74282527e-01 4.02940691e-01 7.24098742e-01
5.02951860e-01 -4.65301514e-01 -7.36039817e-01 8.26156616e-01
-9.63420272e-01 -5.40986478e-01 6.43868268e-01 9.33637321e-01
7.29995191e-01 3.85808438e-01 8.10376287e-01 -4.12882000e-01
5.62374651e-01 -3.88636440e-01 -3.04392219e-01 5.83506703e-01
-8.94900680e-01 1.24102555e-01 5.64698398e-01 -3.17081034e-01
-1.06519556e+00 -1.67808056e-01 -2.36373365e-01 -3.61251503e-01
-1.27360880e-01 3.25630695e-01 -7.69626498e-01 -3.70835423e-01
4.58548933e-01 3.29849839e-01 3.78898054e-01 -2.38251463e-01
1.51032493e-01 9.69668031e-01 4.40515816e-01 -5.13394773e-01
9.39108849e-01 4.27837670e-01 1.94975615e-01 -6.89986289e-01
-6.73614502e-01 -3.41625869e-01 -5.61828911e-01 -4.83514696e-01
4.63633895e-01 -7.24249184e-01 -9.44547534e-01 6.40822887e-01
-8.11593413e-01 9.64186341e-02 -1.42912209e-01 4.01387006e-01
-3.79010975e-01 7.90088773e-01 -4.11788166e-01 -1.01254797e+00
-2.44404629e-01 -1.37117016e+00 1.14197850e+00 3.07481825e-01
3.21741521e-01 -7.58207738e-01 -1.29066050e-01 4.97769624e-01
3.00679326e-01 3.68163027e-02 7.62307405e-01 -3.85245621e-01
-3.40330601e-01 -2.25963280e-01 -5.16699374e-01 9.29652333e-01
5.34203649e-01 8.55480433e-02 -1.05833244e+00 -4.16065484e-01
2.02470362e-01 -5.47639072e-01 6.80271089e-01 8.70809332e-02
1.37139785e+00 -3.09341431e-01 2.47818932e-01 5.56421101e-01
1.21363771e+00 2.33946800e-01 9.95802462e-01 -2.08129883e-01
4.17003781e-01 8.16260338e-01 9.16630745e-01 4.43446577e-01
1.21113077e-01 7.73786128e-01 5.53103507e-01 3.21558893e-01
-1.24906637e-01 -2.33466402e-01 6.34184778e-01 9.45933104e-01
1.33469477e-01 -1.95145622e-01 -5.92513084e-01 1.39173657e-01
-1.42904925e+00 -1.05536973e+00 2.41101444e-01 2.35031247e+00
7.65234292e-01 -8.65366310e-02 9.36646089e-02 5.26930392e-01
7.18543887e-01 7.39861205e-02 -6.81180239e-01 -1.40771151e-01
-5.05951084e-02 1.50262490e-01 -2.67006636e-01 4.10200715e-01
-8.74490023e-01 4.88560766e-01 5.73130989e+00 1.28217733e+00
-1.06249940e+00 -1.37070522e-01 9.48732078e-01 2.67760187e-01
-3.68933618e-01 -3.81402761e-01 -6.74584508e-01 6.45800471e-01
8.06262016e-01 -4.59803194e-02 4.12096679e-01 8.02405000e-01
4.09144871e-02 7.13995472e-02 -1.18890226e+00 1.24129295e+00
5.58409512e-01 -5.72081506e-01 2.75048107e-01 7.24535137e-02
6.50351524e-01 -6.71284676e-01 4.81807113e-01 1.27506033e-01
-3.08272183e-01 -1.16806388e+00 4.53515232e-01 8.57733905e-01
1.12669611e+00 -8.74234796e-01 8.93141747e-01 1.76975474e-01
-1.16490722e+00 2.87303422e-03 -6.80859804e-01 2.56253511e-01
-2.17376396e-01 9.12902772e-01 -4.37168926e-01 6.70019448e-01
6.20455325e-01 7.95812190e-01 -6.74164951e-01 9.41185057e-01
-8.08886215e-02 4.09777224e-01 -3.92619986e-03 2.35098869e-01
-1.65232271e-01 -5.58636248e-01 2.67398924e-01 6.52159452e-01
6.02875292e-01 1.84709817e-01 -1.52189389e-01 7.47004628e-01
-2.51323223e-01 2.77772725e-01 -3.54849964e-01 -1.41651407e-01
4.49672312e-01 1.24223244e+00 -3.83786231e-01 -1.82552874e-01
-1.54379249e-01 9.93442237e-01 2.18963489e-01 1.40295461e-01
-6.03325963e-01 -4.12542045e-01 7.98118055e-01 -1.06594093e-01
9.71018448e-02 8.65586102e-02 -1.35452181e-01 -1.08255661e+00
4.02043611e-01 -1.09822845e+00 2.66315043e-01 -5.60130954e-01
-1.73332906e+00 9.17794466e-01 -4.17788386e-01 -1.49849570e+00
7.06414282e-02 -7.10276246e-01 -4.29580122e-01 7.57900357e-01
-1.62895525e+00 -9.68906224e-01 -4.97097224e-01 6.49658620e-01
2.59632379e-01 -4.77555722e-01 5.62799394e-01 4.99140263e-01
-5.55900455e-01 1.08141994e+00 2.04035208e-01 1.46355346e-01
8.23904395e-01 -9.10887063e-01 -2.00353548e-01 6.56609356e-01
7.96874613e-02 4.73065466e-01 3.73057127e-01 -4.42805737e-01
-1.31679749e+00 -9.97320354e-01 7.40341723e-01 -3.69310677e-01
3.49744052e-01 3.59667949e-02 -1.04698777e+00 2.11662427e-02
-7.28031620e-02 3.93268675e-01 8.40434194e-01 -1.46580547e-01
-7.50817478e-01 -8.74629021e-01 -1.59678543e+00 1.64800271e-01
1.02540576e+00 -8.17522764e-01 -4.18893397e-01 8.55927318e-02
1.93287656e-01 2.29418576e-01 -1.16292131e+00 6.70814812e-01
8.40114057e-01 -1.29813170e+00 9.24469531e-01 -3.07401866e-01
3.97957861e-01 -6.19820654e-01 -3.76683831e-01 -1.23222017e+00
-3.47017139e-01 4.29954529e-02 -1.48975998e-01 1.50491834e+00
7.15750977e-02 -5.28864503e-01 7.22831011e-01 4.62656409e-01
1.26875877e-01 -7.27214217e-01 -1.23572755e+00 -1.01405323e+00
-7.64073357e-02 -2.01609343e-01 8.95238340e-01 7.81977415e-01
-3.83003540e-02 -1.10748254e-01 -4.66370463e-01 3.05706948e-01
9.13429558e-01 4.11273614e-02 4.38352674e-01 -1.37406874e+00
-2.05733240e-01 -4.66041178e-01 -9.56178129e-01 -7.94553101e-01
3.32014233e-01 -8.89991939e-01 8.94805789e-02 -1.14520431e+00
4.70867634e-01 -5.22339284e-01 -6.81165099e-01 2.64003217e-01
-1.96565330e-01 6.09000564e-01 2.60867048e-02 3.39966834e-01
-6.55577719e-01 1.06628096e+00 1.39967179e+00 -4.79024202e-01
1.69211343e-01 -4.90037613e-02 -4.84847337e-01 3.57694656e-01
6.39997780e-01 -3.19047838e-01 -4.83151913e-01 -2.14491889e-01
-1.47563880e-02 2.13173423e-02 1.49183482e-01 -1.27034795e+00
8.53368789e-02 -3.20814341e-01 4.89663482e-01 -1.84377596e-01
3.00526381e-01 -1.00173008e+00 8.55224431e-02 3.82272005e-01
-2.66851187e-01 -3.58252496e-01 -1.49929121e-01 6.34353638e-01
-5.79330206e-01 -1.40056983e-01 1.04838872e+00 3.48342091e-01
-4.53470796e-01 6.95717514e-01 1.22795358e-01 -1.51458606e-01
1.04035318e+00 -2.74008006e-01 -2.76564151e-01 -3.81949306e-01
-4.30774033e-01 -3.55209783e-02 3.81558031e-01 3.66212815e-01
1.10477161e+00 -1.83692074e+00 -8.17162097e-01 6.48201704e-01
4.24299657e-01 -5.57384431e-01 4.26031947e-01 6.93915009e-01
-2.77962416e-01 1.39609113e-01 -5.31929851e-01 -7.14125693e-01
-1.38451767e+00 5.38865328e-01 3.53930831e-01 6.85424209e-02
1.13562465e-01 7.79741168e-01 2.57743388e-01 -4.44236726e-01
2.10451335e-01 7.73748010e-02 -3.35217267e-01 4.70736809e-02
8.62001836e-01 3.73408645e-01 2.01574370e-01 -1.10204673e+00
-3.85259628e-01 8.75804424e-01 5.66980653e-02 1.71025768e-01
1.02608275e+00 -1.62241593e-01 -3.10911119e-01 2.30429515e-01
1.50042772e+00 -1.63795248e-01 -1.33285534e+00 -3.23134035e-01
-5.93354739e-02 -9.96544600e-01 1.16262406e-01 -7.94984341e-01
-1.47391570e+00 1.11792421e+00 1.09397936e+00 1.41257579e-02
1.50431871e+00 -2.03009948e-01 4.77267355e-01 1.86606213e-01
5.03171325e-01 -9.67173696e-01 6.39530063e-01 -7.78791234e-02
1.04258454e+00 -1.44412220e+00 -1.42342016e-01 -4.86171067e-01
-6.24392748e-01 1.13766587e+00 7.02610970e-01 1.15884446e-01
9.39280391e-01 -3.45913202e-01 7.25018978e-02 6.24342682e-03
-4.61688429e-01 1.18745267e-02 7.58067310e-01 7.33950198e-01
2.32885405e-01 1.04025297e-01 -2.00170919e-01 5.08964360e-01
1.22474670e-01 -8.17227550e-03 5.91777749e-02 4.74942476e-01
-5.37855268e-01 -1.27069223e+00 -3.03653002e-01 5.54410875e-01
-1.74776718e-01 1.88732728e-01 -2.42017224e-01 2.15330109e-01
3.35857540e-01 1.27614570e+00 -8.27955157e-02 -8.13302517e-01
2.55894423e-01 1.53095108e-02 5.17729461e-01 -3.49793226e-01
-1.08995467e-01 -1.74682617e-01 -3.99836600e-01 -6.27766788e-01
-4.67363358e-01 -5.15565515e-01 -9.17319059e-01 -4.36574936e-01
-3.85913432e-01 3.02503169e-01 5.63990533e-01 8.76053751e-01
3.35560411e-01 7.73739815e-02 1.32566524e+00 -4.37513232e-01
-8.58703375e-01 -8.82180750e-01 -9.27159905e-01 6.13325775e-01
3.72611314e-01 -8.11353147e-01 -6.23494327e-01 -8.58061016e-02] | [13.089505195617676, 0.6631797552108765] |
7c8a8a5d-d40c-4f79-a74c-31ba435f57c7 | prototype-based-domain-generalization | 2204.07358 | null | https://arxiv.org/abs/2204.07358v1 | https://arxiv.org/pdf/2204.07358v1.pdf | Prototype-based Domain Generalization Framework for Subject-Independent Brain-Computer Interfaces | Brain-computer interface (BCI) is challenging to use in practice due to the inter/intra-subject variability of electroencephalography (EEG). The BCI system, in general, necessitates a calibration technique to obtain subject/session-specific data in order to tune the model each time the system is utilized. This issue is acknowledged as a key hindrance to BCI, and a new strategy based on domain generalization has recently evolved to address it. In light of this, we've concentrated on developing an EEG classification framework that can be applied directly to data from unknown domains (i.e. subjects), using only data acquired from separate subjects previously. For this purpose, in this paper, we proposed a framework that employs the open-set recognition technique as an auxiliary task to learn subject-specific style features from the source dataset while helping the shared feature extractor with mapping the features of the unseen target dataset as a new unseen domain. Our aim is to impose cross-instance style in-variance in the same domain and reduce the open space risk on the potential unseen subject in order to improve the generalization ability of the shared feature extractor. Our experiments showed that using the domain information as an auxiliary network increases the generalization performance. | ['Seong-Whan Lee', 'Ji-Hoon Jeong', 'Dong-Kyun Han', 'Serkan Musellim'] | 2022-04-15 | null | null | null | null | ['open-set-learning'] | ['miscellaneous'] | [ 4.98414040e-01 -8.99852663e-02 5.78799069e-01 -5.14524579e-01
-4.85482335e-01 -4.98025596e-01 2.77323902e-01 -2.57553041e-01
-5.12162685e-01 1.16916978e+00 -2.85796821e-01 1.93763852e-01
-6.20549440e-01 -5.16172767e-01 -6.80851400e-01 -8.82388413e-01
-3.87286842e-02 5.76403499e-01 8.26024041e-02 -1.80238456e-01
3.14009905e-01 5.89609087e-01 -1.71961939e+00 2.67102748e-01
1.17419255e+00 9.34676528e-01 4.39001739e-01 8.63255113e-02
-1.14632333e-02 -1.23323435e-02 -1.06480956e+00 -2.05516648e-02
3.80950123e-01 -5.39259672e-01 -6.37650967e-01 -9.84726753e-03
1.11811087e-01 1.91083491e-01 -9.64708552e-02 1.02803099e+00
7.19506443e-01 2.85830826e-01 7.75869310e-01 -1.38297474e+00
-2.92885423e-01 3.32112402e-01 -2.28141099e-01 5.35883844e-01
2.24919543e-01 -1.41974688e-01 2.88148910e-01 -6.80286944e-01
4.38002557e-01 6.15260661e-01 2.63870299e-01 8.46355617e-01
-1.45317090e+00 -1.18642843e+00 1.01136483e-01 5.87560236e-01
-1.74613631e+00 -4.07183141e-01 1.08206892e+00 -4.48210776e-01
6.40012562e-01 1.72554255e-01 6.39710367e-01 1.56673932e+00
2.69842803e-01 4.70216304e-01 1.34066093e+00 -4.13277715e-01
5.30532300e-01 8.43220413e-01 2.50091732e-01 -2.66378433e-01
2.72781551e-01 1.62462130e-01 -5.74178874e-01 -6.80270791e-02
5.75297892e-01 -1.47679940e-01 -7.34298825e-01 -6.75382257e-01
-7.11955488e-01 5.52713573e-01 3.99501592e-01 8.09960663e-01
-3.86706203e-01 -6.23973191e-01 3.45687985e-01 5.74864089e-01
4.58066851e-01 7.95649827e-01 -6.70544922e-01 -2.19769552e-01
-9.55567181e-01 4.95256446e-02 9.20854449e-01 9.03033257e-01
7.67067373e-01 -1.87668383e-01 8.79596695e-02 9.47923958e-01
-3.36374134e-01 2.25990176e-01 9.63911653e-01 4.60469611e-02
5.32012939e-01 7.60578811e-01 -1.81006327e-01 -8.79770160e-01
-4.83535767e-01 -7.17597187e-01 -8.09855878e-01 2.88215399e-01
3.90067697e-01 -2.00999469e-01 -8.86597395e-01 1.84848535e+00
1.57384828e-01 4.08031464e-01 6.78721592e-02 8.02007496e-01
4.38538343e-01 3.98035973e-01 -1.76862881e-01 -3.19496095e-01
9.74060714e-01 -2.19483286e-01 -6.27693474e-01 -2.31341943e-01
4.47865427e-01 -3.19468766e-01 9.52058911e-01 8.12162757e-01
-5.96880257e-01 -6.25335932e-01 -1.24042392e+00 6.07081473e-01
-7.55334139e-01 5.67118786e-02 7.68819824e-02 6.75653517e-01
-7.43484557e-01 5.53304136e-01 -5.08870840e-01 -5.03502488e-01
3.88392091e-01 7.30077207e-01 -6.48953676e-01 1.77388087e-01
-1.36446333e+00 1.14460003e+00 8.82046580e-01 1.51394814e-01
-7.50938594e-01 -7.87091076e-01 -3.07671338e-01 1.27870575e-01
5.33063650e-01 -2.90984452e-01 5.00581741e-01 -1.27711272e+00
-1.57127225e+00 5.43216586e-01 9.78900269e-02 -3.68118972e-01
4.29098666e-01 9.25646815e-03 -7.54373848e-01 -1.42933220e-01
-8.45041201e-02 1.78834870e-01 1.16339374e+00 -1.18663907e+00
-5.96126497e-01 -7.76698768e-01 -2.83427835e-01 4.91519012e-02
-6.18631303e-01 -2.80532748e-01 6.65102005e-02 -6.08114302e-01
1.76918536e-01 -9.28858638e-01 4.07237381e-01 -6.33163810e-01
-1.46808237e-01 -2.06920132e-01 9.28959668e-01 -6.33723259e-01
1.17893195e+00 -2.36124492e+00 4.25835341e-01 7.92900145e-01
8.12400356e-02 2.93844193e-01 -6.22711703e-02 5.22068217e-02
-7.39629686e-01 -4.13129866e-01 -4.35784340e-01 1.57393113e-01
-4.49051499e-01 2.54147917e-01 5.69793768e-03 4.40783501e-01
2.66723216e-01 3.75981122e-01 -5.73697686e-01 -1.22517943e-01
1.06817357e-01 4.46186095e-01 -3.96511465e-01 4.37200814e-01
2.45408729e-01 9.25857425e-01 -2.28053719e-01 5.95110022e-02
7.02555954e-01 1.12857513e-01 -3.79338339e-02 -2.90222228e-01
1.42421603e-01 3.84285152e-02 -1.55729449e+00 1.80909741e+00
-5.36322176e-01 3.33991081e-01 -1.04849778e-01 -1.52268875e+00
1.25292242e+00 4.95693266e-01 5.60190380e-01 -6.22062445e-01
3.22959900e-01 3.72354776e-01 4.53364938e-01 -4.39029545e-01
-1.52793825e-01 5.11130802e-02 1.44829571e-01 2.02801391e-01
4.39885795e-01 9.31857824e-02 -1.23454474e-01 -1.72383815e-01
8.67426574e-01 1.01961397e-01 3.25976819e-01 -6.29949272e-01
7.98554897e-01 -3.15231383e-01 4.61945474e-01 5.43301463e-01
5.56527078e-02 4.91074443e-01 2.45299935e-01 -3.91489804e-01
-7.56880999e-01 -8.79995406e-01 -5.92267990e-01 5.46268642e-01
-5.64939082e-02 9.77756307e-02 -8.79544675e-01 -7.19983220e-01
-6.60623387e-02 8.19917560e-01 -7.84492910e-01 -6.08945251e-01
-4.50446099e-01 -7.04963028e-01 1.63709223e-01 4.39584166e-01
4.61909920e-01 -9.16028261e-01 -6.33294284e-01 4.52748030e-01
-5.33217900e-02 -6.70441508e-01 -7.77811557e-02 9.08257842e-01
-8.27421069e-01 -1.06407726e+00 -8.86106372e-01 -5.58781445e-01
6.06356919e-01 -9.61133242e-02 6.08244538e-01 -4.19406444e-01
-3.96976441e-01 2.51701325e-01 -4.07398999e-01 -5.10588944e-01
-3.11250329e-01 3.47771645e-01 3.81484419e-01 4.26280111e-01
7.07041800e-01 -9.45807278e-01 -4.15946662e-01 5.91113567e-01
-9.20556605e-01 -2.59968311e-01 5.06249607e-01 1.10810602e+00
2.66112864e-01 2.91191697e-01 9.62589025e-01 -7.09092915e-01
7.55061924e-01 -5.16615391e-01 -5.89234829e-01 4.22542483e-01
-6.04094744e-01 1.23621918e-01 6.91136062e-01 -7.86722720e-01
-1.02725828e+00 -2.84112860e-02 1.61639154e-01 -2.69446611e-01
-3.85837972e-01 3.94782960e-01 -7.21554160e-01 -3.29962641e-01
9.70217645e-01 3.58122230e-01 1.12327784e-01 -6.07769966e-01
-2.16534346e-01 1.10339522e+00 3.74236226e-01 -4.54822391e-01
6.50938511e-01 1.29015855e-02 -1.64810732e-01 -8.05445254e-01
-4.91770357e-01 -4.79793847e-01 -9.89254773e-01 -5.03722057e-02
5.47702193e-01 -7.84209490e-01 -3.72020096e-01 4.44212973e-01
-1.08216357e+00 -9.07658413e-03 -1.65606141e-01 6.30905211e-01
-4.46471632e-01 1.04110628e-01 3.84101331e-01 -5.57912409e-01
-1.73422650e-01 -1.30977750e+00 5.15176892e-01 1.47719398e-01
-2.95788854e-01 -7.50014424e-01 -3.52500044e-02 -4.89944592e-02
4.57050323e-01 1.73532572e-02 7.80661643e-01 -1.35636806e+00
-1.05613716e-01 -2.13958770e-01 -6.55278265e-02 8.98046672e-01
4.87912923e-01 -7.45232880e-01 -1.21469915e+00 -5.11457920e-01
6.52547657e-01 1.18978750e-02 3.51753175e-01 1.40043020e-01
1.33155596e+00 4.43737991e-02 -4.09397334e-01 5.36850691e-01
1.29558885e+00 6.22082174e-01 7.20046639e-01 4.34684068e-01
4.34616208e-01 4.64577854e-01 3.46199423e-01 4.79673475e-01
-2.49260232e-01 8.14033329e-01 -9.86309275e-02 2.18507081e-01
1.77404240e-01 1.68599293e-01 3.84367853e-02 4.08215195e-01
-7.08004907e-02 2.29605529e-02 -6.96073771e-01 4.53293264e-01
-1.48056829e+00 -6.93766773e-01 2.70703346e-01 2.57336235e+00
7.54277647e-01 1.03041552e-01 9.17964056e-02 5.22690058e-01
7.12996900e-01 -6.51327670e-01 -6.93526208e-01 -1.80900097e-01
-9.30541903e-02 6.37525380e-01 2.13057026e-01 1.32382363e-01
-8.13404381e-01 4.25972581e-01 5.10161686e+00 8.02317739e-01
-1.33774054e+00 1.08959503e-01 1.56690702e-01 -1.05019800e-01
1.56705126e-01 -2.62365252e-01 -7.52937436e-01 7.63201296e-01
1.04203713e+00 -3.78544033e-01 6.61161125e-01 7.75645912e-01
-6.03076816e-02 -1.58391401e-01 -1.57695353e+00 1.32716286e+00
3.78959090e-01 -6.68617964e-01 -4.30271178e-02 1.90796614e-01
4.14400727e-01 -3.70091736e-01 -1.07997097e-01 3.52232307e-01
-7.23547995e-01 -8.96459937e-01 2.42908552e-01 6.22095704e-01
7.65265346e-01 -8.17574918e-01 8.07212770e-01 4.73602593e-01
-6.74197614e-01 -2.31273174e-01 -3.92945468e-01 2.59480067e-02
-4.08445209e-01 3.84333163e-01 -1.00645268e+00 6.37080610e-01
8.02427649e-01 5.16359210e-01 -6.95496440e-01 1.17232406e+00
4.50586565e-02 3.90770227e-01 -4.42006081e-01 7.22626075e-02
-1.18247561e-01 -3.31460476e-01 6.66803420e-01 7.70360410e-01
5.65070093e-01 2.69965436e-02 -8.13693255e-02 9.62510586e-01
3.26082915e-01 1.30505815e-01 -8.16415012e-01 3.40251863e-01
2.67512530e-01 9.34237123e-01 -4.19334710e-01 -1.89638630e-01
-3.00833881e-01 1.32551956e+00 2.70268470e-01 3.99657160e-01
-8.13214302e-01 -7.07327664e-01 4.51467425e-01 2.29324289e-02
1.57675251e-01 1.98540092e-01 -2.75689989e-01 -1.31669211e+00
2.57087290e-01 -9.92340088e-01 4.34743136e-01 -5.13578892e-01
-1.44169450e+00 9.47218657e-01 2.87940055e-01 -1.45733225e+00
-3.30214739e-01 -7.19172895e-01 -2.70816505e-01 1.32224011e+00
-1.27462697e+00 -6.38554573e-01 -4.57889080e-01 1.16252732e+00
3.66701186e-01 -4.81400281e-01 1.03114426e+00 5.35144687e-01
-4.87245351e-01 8.46574724e-01 2.36617818e-01 -1.49222717e-01
9.97716188e-01 -8.99688482e-01 -2.91432291e-01 6.32737756e-01
4.78930883e-02 7.09234595e-01 7.30291963e-01 -4.68530744e-01
-1.01361060e+00 -7.18312263e-01 4.82103378e-01 -3.80057454e-01
4.49134827e-01 -6.57927513e-01 -1.28321755e+00 4.57475185e-01
-9.13889408e-02 -1.62017792e-01 7.10683167e-01 2.14872614e-01
4.19901684e-02 -6.50308490e-01 -1.28608584e+00 3.63629431e-01
9.18980241e-01 -4.16189879e-01 -1.01569068e+00 1.30096143e-02
1.07011229e-01 -1.69505045e-01 -9.49665964e-01 2.86776215e-01
4.81120259e-01 -6.39976323e-01 6.13492250e-01 -4.88976240e-01
-2.28952035e-01 -1.82907268e-01 -1.17907673e-02 -1.88838553e+00
-2.15963945e-01 -4.16751534e-01 3.22080463e-01 1.18047476e+00
4.68583882e-01 -1.03605163e+00 6.51561856e-01 1.01453841e+00
-2.19688024e-02 -4.27405238e-01 -1.30018735e+00 -1.10109019e+00
-2.62236577e-02 -3.89657110e-01 8.66036296e-01 6.98962569e-01
2.57818341e-01 2.96821058e-01 -3.25720310e-01 2.06900343e-01
3.64976466e-01 -1.23574384e-01 7.74379313e-01 -1.67190170e+00
-2.21647993e-01 5.77693395e-02 -8.34308803e-01 -6.30935371e-01
1.99377671e-01 -9.53855038e-01 -4.94184941e-02 -8.26929152e-01
4.04657684e-02 -4.67031509e-01 -7.67219007e-01 3.53353739e-01
-2.03998079e-05 -7.34324232e-02 5.83457313e-02 -9.72020533e-03
5.98421320e-02 5.33346713e-01 1.01143873e+00 -1.08239919e-01
-6.26679122e-01 3.21411133e-01 -7.28273094e-01 4.07919139e-01
8.53768170e-01 -6.58256948e-01 -7.40952432e-01 -1.69060394e-01
-1.72456399e-01 -1.46648616e-01 3.38048190e-01 -1.55923617e+00
2.35221863e-01 2.28265569e-01 7.76771963e-01 -3.25471401e-01
3.70591491e-01 -1.46062493e+00 3.31208289e-01 1.52149051e-01
-2.67277986e-01 -3.29933643e-01 4.98157352e-01 4.92856443e-01
-1.83890700e-01 -4.63150620e-01 7.18515456e-01 2.94692572e-02
-6.85271382e-01 1.04539424e-01 -1.19896352e-01 -6.62249103e-02
1.26477945e+00 -7.49037743e-01 1.92631662e-01 5.03791571e-02
-9.94369209e-01 -9.39163864e-02 1.68807000e-01 5.44318795e-01
6.21761501e-01 -9.86803234e-01 -5.90239882e-01 7.36795843e-01
4.36582655e-01 -2.57474065e-01 3.88328135e-01 7.70641267e-01
2.63973534e-01 3.07322592e-01 -7.08294868e-01 -6.70396328e-01
-1.15475357e+00 7.22639561e-01 3.92207146e-01 5.73042557e-02
-6.51166320e-01 7.63407052e-01 2.27203265e-01 -1.81359366e-01
4.52106595e-01 -2.79405057e-01 -3.93917620e-01 1.70471817e-01
5.95976591e-01 2.77109772e-01 5.45591235e-01 -2.68132240e-01
-5.26915073e-01 3.55039984e-01 -2.78600067e-01 -2.00138963e-03
1.47285879e+00 -6.52325973e-02 2.76323874e-02 5.72788775e-01
1.37919009e+00 -3.68763477e-01 -9.06907201e-01 -1.66807964e-01
6.17445260e-02 -6.22892320e-01 -1.06520765e-02 -1.14514804e+00
-9.02557015e-01 8.11989605e-01 1.20876348e+00 4.17089872e-02
1.44501221e+00 -2.16457501e-01 2.71312565e-01 3.91710073e-01
8.06222379e-01 -1.26353323e+00 -3.41117978e-01 1.60178661e-01
1.10809875e+00 -1.13822842e+00 -1.92697808e-01 -4.77406271e-02
-7.08998501e-01 1.19689286e+00 6.75226390e-01 -1.70688346e-01
8.37249160e-01 1.43223405e-01 -1.85980693e-01 -6.44148141e-02
-3.01451385e-01 1.51459470e-01 5.03451824e-01 7.97633886e-01
1.12265505e-01 -7.67202973e-02 -4.27647918e-01 9.77878451e-01
-6.21316992e-02 4.63463426e-01 4.09532458e-01 6.79444551e-01
-1.10463105e-01 -1.23096395e+00 -5.06608248e-01 6.88035548e-01
-6.35009781e-02 1.05765447e-01 -1.34885833e-01 8.58149469e-01
3.51662517e-01 9.20976639e-01 -1.04232922e-01 -4.90420669e-01
6.55328929e-01 4.91777211e-01 5.90232551e-01 -6.05830014e-01
-5.89882433e-01 -2.08295062e-01 -3.68368387e-01 -5.31998813e-01
-2.19304278e-01 -6.31069779e-01 -7.60124981e-01 3.48042041e-01
-4.65584964e-01 3.67044151e-01 9.08225358e-01 9.74159241e-01
5.33299148e-01 6.16520226e-01 7.79988825e-01 -7.57760704e-01
-7.86707222e-01 -1.15646720e+00 -9.37656462e-01 6.82955861e-01
1.79323509e-01 -1.10716975e+00 -5.84580541e-01 -1.06675416e-01] | [13.124414443969727, 3.456477642059326] |
952558be-792c-456b-a874-a4a32a880749 | finxabsa-explainable-finance-through-aspect | 2303.02563 | null | https://arxiv.org/abs/2303.02563v3 | https://arxiv.org/pdf/2303.02563v3.pdf | FinXABSA: Explainable Finance through Aspect-Based Sentiment Analysis | This paper presents a novel approach for explainability in financial analysis by utilizing the Pearson correlation coefficient to establish a relationship between aspect-based sentiment analysis and stock prices. The proposed methodology involves constructing an aspect list from financial news articles and analyzing sentiment intensity scores for each aspect. These scores are then compared to the stock prices for the relevant companies using the Pearson coefficient to determine any significant correlations. The results indicate that the proposed approach provides a more detailed and accurate understanding of the relationship between sentiment analysis and stock prices, which can be useful for investors and financial analysts in making informed decisions. Additionally, this methodology offers a transparent and interpretable way to explain the sentiment analysis results and their impact on stock prices. Overall, the findings of this paper demonstrate the importance of explainability in financial analysis and highlight the potential benefits of utilizing the Pearson coefficient for analyzing aspect-based sentiment analysis and stock prices. The proposed approach offers a valuable tool for understanding the complex relationships between financial news sentiment and stock prices, providing a new perspective on the financial market and aiding in making informed investment decisions. | ['Erik Cambria', 'Gianmarco Mengaldo', 'Ranjan Satapathy', 'Wihan van der Heever', 'Keane Ong'] | 2023-03-05 | null | null | null | null | ['aspect-based-sentiment-analysis'] | ['natural-language-processing'] | [-5.59576452e-01 -1.67801574e-01 -5.75278223e-01 -4.00002092e-01
-1.31220862e-01 -1.04258931e+00 4.96433645e-01 6.05685413e-01
-1.61348179e-01 5.18253505e-01 4.87047255e-01 -5.74717820e-01
-2.02146366e-01 -1.10879052e+00 -2.20003560e-01 -5.02294660e-01
6.87570572e-02 7.33208731e-02 -2.33285740e-01 -5.73049366e-01
1.11791992e+00 5.28735578e-01 -1.44115126e+00 7.10420758e-02
3.47420335e-01 1.32685459e+00 3.55788246e-02 1.99318111e-01
-6.74898565e-01 1.01422644e+00 -8.03035140e-01 -7.77160406e-01
3.60382795e-01 -1.90111771e-01 -2.26657152e-01 1.66491896e-01
-4.74420875e-01 -2.21571058e-01 7.63101339e-01 1.06997991e+00
-2.72938550e-01 -2.73061156e-01 4.47685391e-01 -1.06992865e+00
-7.46015012e-01 8.12632918e-01 -7.91303813e-01 6.53038979e-01
5.15013099e-01 -2.08730504e-01 1.56597161e+00 -1.06231165e+00
3.45186323e-01 8.14869523e-01 4.15927291e-01 -5.63006878e-01
-5.28215170e-01 -8.46900344e-01 3.46317530e-01 -1.70299932e-02
-6.64726615e-01 2.16153890e-01 8.88726592e-01 -6.33337438e-01
1.00709438e+00 4.29748684e-01 1.38448560e+00 -5.58032990e-02
6.88737392e-01 3.22194010e-01 1.41100645e+00 -4.52034563e-01
2.29511082e-01 9.27973926e-01 4.23116744e-01 -1.21288765e-02
1.04740858e+00 -1.74393699e-01 -7.16035664e-01 1.74644478e-02
2.77333498e-01 4.67224836e-01 1.14763796e-01 3.22296560e-01
-1.07905972e+00 1.20243919e+00 3.01224977e-01 5.80736697e-01
-8.33009660e-01 -3.50348651e-01 4.47668463e-01 5.13937175e-01
7.80584753e-01 8.11075151e-01 -6.58077300e-01 -3.45995605e-01
-8.84519160e-01 3.15612614e-01 9.43338275e-01 5.79096556e-01
4.94798064e-01 3.14449608e-01 4.35548544e-01 4.14383598e-02
6.67409599e-01 7.30292559e-01 7.29312479e-01 -3.39348972e-01
4.08291608e-01 1.03316879e+00 3.76469880e-01 -1.52003694e+00
-2.94665366e-01 -6.87103748e-01 8.08039010e-02 4.57406133e-01
-1.67966142e-01 -1.25228941e-01 -2.20122725e-01 8.70920718e-01
1.02745406e-02 -4.46937501e-01 4.44876790e-01 7.84163415e-01
6.92196786e-01 6.93397045e-01 -1.68716479e-02 -1.38750538e-01
1.85193872e+00 -5.41773200e-01 -1.06814265e+00 -8.18096027e-02
3.09687287e-01 -1.18644321e+00 8.64639580e-01 5.12888670e-01
-1.01274204e+00 -3.59642982e-01 -1.28024375e+00 4.84471887e-01
-6.87861741e-01 4.50573443e-03 9.37514544e-01 6.82765424e-01
-7.10514367e-01 4.20235813e-01 -7.06102669e-01 -1.94219090e-02
-1.79595530e-01 2.46046126e-01 3.65141854e-02 5.57225585e-01
-1.06266844e+00 1.09690297e+00 4.79764700e-01 1.07141636e-01
3.62564117e-01 -4.04095024e-01 -6.82281017e-01 1.47990376e-01
9.76740289e-03 -1.80947959e-01 1.09743834e+00 -9.59862411e-01
-1.03055012e+00 5.17105937e-01 -1.47773162e-01 -5.93191087e-01
1.01095818e-01 -2.21811309e-01 -6.26022100e-01 1.12005308e-01
3.54638934e-01 -3.71991038e-01 2.13227063e-01 -8.73935997e-01
-1.09093606e+00 -3.69073540e-01 1.52021676e-01 2.16402784e-01
-2.94255763e-01 3.87126237e-01 9.50738639e-02 -9.14648950e-01
5.50897777e-01 -5.02457619e-01 3.86545956e-02 -8.99059236e-01
-1.04418606e-01 1.33827269e-01 5.48369646e-01 -6.50072038e-01
1.30886400e+00 -1.79231429e+00 -5.65793812e-01 7.23985672e-01
5.96779250e-02 -2.99238831e-01 4.97039974e-01 8.75499845e-01
-1.74644887e-01 4.64820921e-01 2.62620479e-01 2.13303223e-01
4.19002101e-02 -7.76253045e-02 -6.33328438e-01 1.86085492e-01
4.16621625e-01 7.22346008e-01 -5.05380034e-01 9.04920101e-02
2.65577585e-01 2.75742888e-01 -1.42449051e-01 4.71743802e-03
8.61911923e-02 1.14959255e-02 -7.32866108e-01 9.23416197e-01
4.61249709e-01 -1.86861053e-01 1.07637115e-01 -2.09949195e-01
-6.97643936e-01 6.64016306e-01 -1.13430798e+00 3.73739213e-01
-3.24492842e-01 8.00714791e-01 -4.65240777e-01 -6.96442485e-01
1.60350323e+00 2.28640348e-01 3.01828414e-01 -6.40300870e-01
4.80562866e-01 4.99987662e-01 8.16158876e-02 -1.46813348e-01
1.01384282e+00 -7.84097552e-01 -1.18211173e-01 8.91783237e-01
-4.62278634e-01 -2.26069584e-01 6.21157348e-01 -3.10417470e-02
2.88779467e-01 -3.72942299e-01 8.30442250e-01 -3.92419338e-01
6.64590478e-01 1.11352280e-01 3.30500156e-01 -1.34753007e-02
1.57266915e-01 -2.97809672e-02 8.00366104e-01 -6.36443377e-01
-8.59860003e-01 -4.80018407e-01 -2.18389928e-01 3.50567818e-01
-1.60322770e-01 -1.97741538e-01 -3.21797371e-01 -8.99117738e-02
2.16511771e-01 8.92154574e-01 -8.82550418e-01 3.51370513e-01
2.65455931e-01 -9.55336511e-01 -4.07396823e-01 5.35331666e-01
4.86262172e-01 -8.67743134e-01 -9.57159877e-01 2.01377407e-01
1.53037205e-01 -8.72699797e-01 2.48140663e-01 1.69769749e-01
-1.14411664e+00 -1.14621830e+00 -4.81861711e-01 -2.33808398e-01
7.99072385e-01 4.18718994e-01 1.07855856e+00 1.48564771e-01
3.21223170e-01 1.29538536e-01 -7.23224103e-01 -1.17856824e+00
-2.48347446e-01 -9.32194665e-02 -8.94884095e-02 -1.56253994e-01
8.73019218e-01 -1.19247168e-01 -5.89100242e-01 1.78928673e-01
-9.75387454e-01 -3.31313610e-01 5.38278103e-01 3.00750762e-01
4.60267186e-01 8.34048212e-01 8.34043443e-01 -1.06637728e+00
1.14325154e+00 -8.13663244e-01 -1.02205741e+00 3.54474061e-03
-1.31611300e+00 -1.15706570e-01 3.62890452e-01 3.70770097e-01
-1.20262015e+00 -7.89591134e-01 1.87558785e-01 4.47266608e-01
3.09226483e-01 1.44378448e+00 4.04200435e-01 4.49020714e-02
2.99669523e-02 4.04957309e-02 2.65522264e-02 -2.14317605e-01
-9.85103752e-03 3.35721940e-01 7.66095370e-02 3.76357734e-02
8.32749248e-01 5.42930007e-01 -4.23198985e-03 -2.35692143e-01
-9.83352304e-01 -7.48911798e-01 -6.32180452e-01 -3.86838049e-01
7.27338552e-01 -1.02342331e+00 -7.10006416e-01 1.33237720e-01
-6.83381259e-01 7.43034244e-01 -1.99666321e-01 9.24145639e-01
-1.29469901e-01 -1.03200153e-01 -4.52322930e-01 -1.19732666e+00
-3.52692813e-01 -1.13447809e+00 3.26939732e-01 3.96682948e-01
-8.10257077e-01 -1.50606406e+00 8.12193900e-02 5.85518599e-01
5.41829884e-01 5.59946477e-01 7.79101133e-01 -1.18943143e+00
-4.85834897e-01 -9.63354588e-01 -1.39721334e-01 2.73270488e-01
4.80923444e-01 4.36310112e-01 -4.94118780e-01 1.70742288e-01
5.42761564e-01 1.82353050e-01 3.46948922e-01 3.11576426e-01
-4.09314595e-02 -5.08383393e-01 3.63782376e-01 1.01859473e-01
1.80401540e+00 6.83769464e-01 2.81281769e-01 1.46760654e+00
1.61216587e-01 9.78396893e-01 1.28302062e+00 9.86397266e-01
4.21249241e-01 6.67691156e-02 2.33125106e-01 3.33925411e-02
6.85364485e-01 5.11326566e-02 3.65848392e-01 1.23797667e+00
-2.71420956e-01 3.01865578e-01 -7.87372649e-01 4.15456980e-01
-1.49775052e+00 -1.01306641e+00 -2.68088341e-01 1.76859975e+00
2.58062989e-01 5.84197521e-01 2.85456181e-01 6.69656992e-01
4.10931885e-01 2.68244445e-01 -1.76794052e-01 -9.71088648e-01
-3.59783649e-01 1.75030127e-01 6.75450027e-01 2.72689402e-01
-6.30787849e-01 5.77487886e-01 6.86827374e+00 -6.02175556e-02
-1.32344210e+00 -5.23238003e-01 7.84838080e-01 9.24928114e-02
-8.52636755e-01 9.28703770e-02 -7.78270841e-01 4.18604851e-01
8.44453812e-01 -1.02677131e+00 -2.29129881e-01 1.07348871e+00
7.31660903e-01 -5.26045918e-01 -4.67991382e-01 1.55047655e-01
-4.12783660e-02 -1.53959954e+00 2.50172704e-01 2.94823617e-01
8.64345849e-01 -3.58021319e-01 4.19403434e-01 -4.08269942e-01
1.58564359e-01 -4.73598957e-01 1.07082748e+00 5.76045692e-01
-3.02443892e-01 -1.25875366e+00 1.61563373e+00 -1.67649984e-01
-1.29505193e+00 -8.75414833e-02 -2.82534868e-01 -7.99123347e-01
1.90230057e-01 6.63699865e-01 -1.06586325e+00 5.96563041e-01
7.46735573e-01 7.98021913e-01 -2.85180986e-01 5.49662888e-01
-2.61311918e-01 4.80544895e-01 1.78396463e-01 -4.88153249e-01
3.11408848e-01 -7.43719637e-01 1.28434658e-01 9.47509646e-01
6.79836810e-01 1.77726641e-01 -5.99840224e-01 7.77764320e-01
3.59231383e-01 6.67745173e-01 -5.45796633e-01 -4.63255912e-01
2.94681847e-01 1.03814089e+00 -1.60661733e+00 -4.38876867e-01
-9.52241361e-01 9.88007486e-02 -4.76268888e-01 1.16852805e-01
-3.48176688e-01 -3.00727993e-01 5.81070602e-01 2.31725037e-01
5.23410082e-01 1.75774656e-03 -1.20637012e+00 -9.71774936e-01
8.12191442e-02 -7.30131626e-01 2.07515642e-01 -8.05382133e-01
-9.46840167e-01 5.19871950e-01 -8.44859146e-03 -1.41665852e+00
-2.56015718e-01 -7.31362760e-01 -9.49351370e-01 9.67378557e-01
-1.78783441e+00 -4.72492188e-01 1.15814634e-01 4.16937061e-02
2.83804357e-01 -5.97556710e-01 6.41553640e-01 -2.84551501e-01
-3.46158743e-01 -1.79973960e-01 1.30220860e-01 -5.56496233e-02
4.45815027e-02 -1.32041144e+00 4.58209395e-01 1.00245929e+00
2.33543236e-02 9.30685401e-01 1.07597911e+00 -9.45144653e-01
-1.05227351e+00 -2.92691410e-01 1.18140221e+00 -2.91134626e-01
1.28903282e+00 3.98069650e-01 -6.17256224e-01 4.76191014e-01
4.14529175e-01 -7.78663158e-01 1.50497198e+00 1.56326294e-01
-1.10463977e-01 -2.54848987e-01 -1.10990965e+00 3.06877077e-01
-4.54996407e-01 -2.47164533e-01 -1.01169086e+00 -1.15759425e-01
5.07138491e-01 8.47680047e-02 -1.16507781e+00 -1.73648801e-02
7.60938585e-01 -1.43258691e+00 5.14726102e-01 -2.88078606e-01
7.25312829e-01 -2.49519244e-01 -1.53002441e-01 -1.14932823e+00
-2.22966895e-01 -1.57748852e-02 4.73385751e-01 1.23766947e+00
1.02877891e+00 -1.04773140e+00 7.58126140e-01 1.09770989e+00
3.61129045e-01 -8.83175671e-01 -3.43782932e-01 -4.29989338e-01
-1.45791784e-01 -6.57742202e-01 1.10481966e+00 9.65755701e-01
3.68813664e-01 -3.05907905e-01 5.77090159e-02 2.23721992e-02
3.76415730e-01 4.91815984e-01 6.69302940e-01 -1.09669614e+00
-7.97672644e-02 -7.13384867e-01 -5.14810979e-01 -5.22388667e-02
-2.19313830e-01 -3.71058941e-01 -8.30194414e-01 -1.50622392e+00
4.89113517e-02 5.48050068e-02 -6.86610878e-01 1.29485351e-03
-9.75732729e-02 2.06308424e-01 6.09696984e-01 6.32083476e-01
1.48764908e-01 -3.82619090e-02 1.12180793e+00 2.39465833e-01
-2.05007285e-01 2.93617636e-01 -1.49011350e+00 9.80993748e-01
9.47513044e-01 -4.52091902e-01 -3.87320042e-01 2.08410531e-01
1.10868740e+00 -2.15706155e-01 -1.54765099e-01 -4.46609646e-01
-2.82336082e-02 -3.88422698e-01 3.55538338e-01 -9.09367621e-01
-1.77908614e-01 -1.07349598e+00 2.00488985e-01 7.58837402e-01
-1.16123147e-01 1.02501369e+00 3.28082860e-01 1.78237900e-01
-9.66102540e-01 -4.75814492e-01 6.79508448e-02 -2.28727996e-01
-7.86740541e-01 -3.55694562e-01 -5.44223726e-01 -4.47084159e-01
1.21230948e+00 -6.42630398e-01 -2.06473738e-01 -6.89173877e-01
-5.33721089e-01 -3.52873094e-02 3.52675825e-01 2.53166258e-01
6.35536849e-01 -1.15787911e+00 -4.52435970e-01 1.78539172e-01
-2.04048138e-02 -5.30231595e-01 -1.65305987e-01 7.59972990e-01
-1.13200498e+00 8.42186749e-01 -4.64625835e-01 4.80206646e-02
-1.00690186e+00 1.77824453e-01 -2.37355426e-01 -3.93631697e-01
-2.15900794e-01 6.76230609e-01 3.70642506e-02 3.59624594e-01
-2.78277367e-01 -7.72258699e-01 -8.81920338e-01 8.79824042e-01
6.44544959e-01 3.46806794e-01 3.00394502e-02 -9.61359382e-01
-4.10029620e-01 8.70021701e-01 -1.50532305e-01 -5.51638782e-01
1.67350984e+00 -4.55230057e-01 -3.85502607e-01 9.63839829e-01
7.83942401e-01 5.66785574e-01 -7.25422919e-01 1.29496723e-01
6.16955459e-01 -6.62579954e-01 1.00453585e-01 -8.06355417e-01
-1.33161795e+00 2.94028133e-01 -3.38917039e-02 8.18925142e-01
1.24982297e+00 -2.24972516e-01 5.74366331e-01 2.09968716e-01
3.95503081e-02 -1.30179691e+00 -1.68221861e-01 2.51381807e-02
8.68825912e-01 -1.27587831e+00 5.21284580e-01 -3.72284919e-01
-1.03922915e+00 1.62525749e+00 -1.20606519e-01 -2.64540792e-01
1.11736608e+00 2.40586624e-01 6.94946349e-01 -5.13841629e-01
-5.34769773e-01 1.44910440e-01 3.99891496e-01 3.70513685e-02
1.10305381e+00 1.68622881e-01 -8.08635890e-01 9.55723286e-01
-1.10016227e+00 -2.29342565e-01 9.62200582e-01 1.11731303e+00
-6.19659364e-01 -1.10896111e+00 -6.06852531e-01 5.10329425e-01
-1.22831202e+00 -2.35046998e-01 -6.29159808e-01 1.25400233e+00
-3.59002978e-01 9.26698327e-01 2.08021462e-01 -3.70820582e-01
5.84126770e-01 -1.70286238e-01 -3.57848138e-01 -5.69100380e-01
-8.71491671e-01 3.82026672e-01 1.78756967e-01 -1.43830061e-01
-9.51368928e-01 -1.03007615e+00 -1.17828321e+00 -5.01824260e-01
-4.08454239e-01 6.20110989e-01 1.14021742e+00 1.12303686e+00
1.80234723e-02 6.14818156e-01 1.00630283e+00 -3.26428771e-01
-2.41173938e-01 -6.34667397e-01 -1.00280464e+00 1.29512861e-01
2.70451277e-01 -5.64771891e-01 -6.40361488e-01 -3.06441057e-02] | [4.4908366203308105, 4.335035800933838] |
56911c8f-6e95-47b0-b377-cd2927d974a4 | semantically-aware-mask-cyclegan-for | 2306.06577 | null | https://arxiv.org/abs/2306.06577v1 | https://arxiv.org/pdf/2306.06577v1.pdf | Semantically-aware Mask CycleGAN for Translating Artistic Portraits to Photo-realistic Visualizations | Image-to-image translation (I2I) is defined as a computer vision task where the aim is to transfer images in a source domain to a target domain with minimal loss or alteration of the content representations. Major progress has been made since I2I was proposed with the invention of a variety of revolutionary generative models. Among them, GAN-based models perform exceptionally well as they are mostly tailor-made for specific domains or tasks. However, few works proposed a tailor-made method for the artistic domain. In this project, I propose the Semantic-aware Mask CycleGAN (SMCycleGAN) architecture which can translate artistic portraits to photo-realistic visualizations. This model can generate realistic human portraits by feeding the discriminators semantically masked fake samples, thus enforcing them to make discriminative decisions with partial information so that the generators can be optimized to synthesize more realistic human portraits instead of increasing the similarity of other irrelevant components, such as the background. Experiments have shown that the SMCycleGAN generate images with significantly increased realism and minimal loss of content representations. | ['Zhuohao Yin'] | 2023-06-11 | null | null | null | null | ['image-to-image-translation', 'image-to-image-translation'] | ['computer-vision', 'miscellaneous'] | [ 7.20883906e-01 3.86274099e-01 1.88328400e-01 -2.98557252e-01
-2.47904792e-01 -6.06646359e-01 8.12767029e-01 -8.16933632e-01
9.23542604e-02 8.99789989e-01 3.64790857e-02 1.44137189e-01
3.61379683e-01 -9.03798342e-01 -6.27956808e-01 -7.32371390e-01
5.27428269e-01 5.47180831e-01 9.34268627e-03 -3.46267909e-01
1.29483685e-01 5.16670108e-01 -1.48174810e+00 2.28975281e-01
7.71514833e-01 5.61363101e-01 2.47688174e-01 6.46285534e-01
4.04836843e-03 6.22812450e-01 -9.30366695e-01 -5.88385344e-01
4.83519018e-01 -1.28402984e+00 -3.83247554e-01 3.50163847e-01
5.51494539e-01 -8.52192938e-02 -3.12581778e-01 1.21257818e+00
4.00548726e-01 -1.09703742e-01 9.02805984e-01 -1.62450254e+00
-1.09662509e+00 2.43044838e-01 -5.55423737e-01 -2.18785167e-01
1.73618391e-01 3.83620918e-01 4.96557653e-01 -5.41519523e-01
9.32253301e-01 1.48050117e+00 2.11678773e-01 8.72359395e-01
-1.42613339e+00 -8.56796384e-01 -1.69903010e-01 1.77044898e-01
-1.29469156e+00 -3.43195528e-01 1.09760189e+00 -2.93894380e-01
2.53188968e-01 5.58210731e-01 6.26086056e-01 1.39232600e+00
1.53287843e-01 6.11035764e-01 1.35171700e+00 -5.52039206e-01
1.95928648e-01 6.63892746e-01 -7.51829445e-01 5.69102585e-01
2.69490391e-01 1.15398288e-01 -2.79260486e-01 2.29723155e-01
9.08229530e-01 -2.24467367e-01 -3.28294724e-01 -1.66341826e-01
-1.04780710e+00 8.79817069e-01 4.38047707e-01 4.68133837e-01
-2.66724050e-01 1.33954331e-01 -7.91837648e-02 2.90906787e-01
3.18311036e-01 6.86368704e-01 2.41371572e-01 3.17550153e-01
-1.08099866e+00 2.34653518e-01 5.25179565e-01 9.94833767e-01
8.45711768e-01 6.38497353e-01 -4.53112721e-01 5.32160401e-01
9.57874507e-02 7.41840303e-01 6.17466748e-01 -9.73483324e-01
1.91953108e-01 5.12932897e-01 6.02482818e-03 -1.14489937e+00
2.52079844e-01 -3.39296579e-01 -1.08806324e+00 7.20210433e-01
5.00071459e-02 -1.62228987e-01 -1.13915265e+00 1.78531611e+00
1.77288786e-01 -6.13230579e-02 -1.61155239e-02 1.05108440e+00
6.28437757e-01 8.55365634e-01 -1.25000536e-01 1.85138166e-01
1.04063118e+00 -9.08332467e-01 -7.49592602e-01 -4.20552552e-01
-4.35476415e-02 -9.34442520e-01 1.23502886e+00 1.20684355e-01
-1.13472712e+00 -7.15455472e-01 -1.10488260e+00 8.85568783e-02
-3.08943093e-01 -3.55113372e-02 1.83165953e-01 1.07267332e+00
-1.06529367e+00 3.82939577e-01 -2.46210441e-01 -1.15083911e-01
4.62042570e-01 3.05377822e-02 -2.85463512e-01 -9.98641253e-02
-9.68403220e-01 8.99155021e-01 3.60624880e-01 -1.39215752e-01
-1.22391069e+00 -5.41764021e-01 -6.44719541e-01 -1.82866380e-02
7.83946812e-02 -8.62535596e-01 6.70756400e-01 -2.00611234e+00
-1.85882831e+00 1.07507551e+00 1.01309828e-01 -3.01545948e-01
8.64440024e-01 3.84251207e-01 -3.70519787e-01 9.45469067e-02
6.90580010e-02 1.04591382e+00 1.34047925e+00 -1.52532172e+00
-3.73913258e-01 -9.87013131e-02 -1.67002603e-01 1.07504629e-01
-2.52230823e-01 -1.41804248e-01 -3.11394334e-01 -1.06349003e+00
-3.48196328e-01 -1.22072172e+00 -3.07932831e-02 7.53468648e-02
-6.59127355e-01 4.24246997e-01 1.33136868e+00 -5.72976172e-01
5.38478553e-01 -2.16540265e+00 2.58685201e-01 1.04676254e-01
3.47038656e-02 4.86396730e-01 -4.78476614e-01 4.69054371e-01
-1.49665967e-01 2.93496758e-01 -3.91968280e-01 -3.60765666e-01
-1.22612558e-01 2.92748928e-01 -3.13917249e-01 2.47198462e-01
4.56148058e-01 1.03659892e+00 -7.30792403e-01 -4.72203135e-01
3.42830747e-01 5.94768167e-01 -5.00452101e-01 3.38079005e-01
-3.92571837e-01 9.34269786e-01 -1.80576250e-01 3.20812315e-01
8.31164896e-01 -8.35109502e-03 2.19698716e-02 -4.19932418e-02
2.89238900e-01 -2.87585765e-01 -6.95131838e-01 1.35979331e+00
-4.31728035e-01 1.08106744e+00 -4.12837639e-02 -8.77265334e-01
1.06141114e+00 1.57299072e-01 1.70953527e-01 -6.52307570e-01
1.69793442e-01 -6.11539464e-04 8.51743072e-02 -2.99682796e-01
4.75385338e-01 -4.00234997e-01 -1.20503262e-01 1.87014863e-01
-1.64826021e-01 -6.37388706e-01 -4.13696021e-02 1.59075763e-02
6.89892411e-01 1.97027460e-01 9.87043902e-02 -3.17105055e-02
5.32097876e-01 7.92751461e-02 4.68923122e-01 5.04175603e-01
1.23318695e-01 1.05473018e+00 3.06743056e-01 -3.03399205e-01
-1.43902051e+00 -1.04847956e+00 4.25046206e-01 4.41656172e-01
2.39140928e-01 9.66433436e-02 -1.10937858e+00 -5.78329682e-01
-3.44201416e-01 1.12793875e+00 -7.40105271e-01 -4.60940748e-01
-4.88051414e-01 -3.32684606e-01 5.48076451e-01 4.77202190e-03
1.02055264e+00 -1.29912817e+00 -6.21282101e-01 -1.00463256e-01
-1.74998358e-01 -1.06388676e+00 -6.65694296e-01 -4.18261558e-01
-4.91619200e-01 -7.95268893e-01 -1.03066087e+00 -8.34432006e-01
9.94031787e-01 2.82720715e-01 9.34845567e-01 -2.09507257e-01
-5.04166067e-01 9.17613134e-02 -4.08739418e-01 -4.22273338e-01
-9.48036611e-01 -2.10988641e-01 -3.42562109e-01 4.82696056e-01
-2.05678552e-01 -6.03646934e-01 -7.38067091e-01 5.02900779e-01
-1.36255145e+00 5.61881483e-01 6.40345812e-01 7.44626284e-01
4.59237754e-01 7.93567300e-02 3.15784663e-01 -9.50900674e-01
4.65072691e-01 -7.96549618e-02 -4.01408106e-01 2.35684559e-01
-6.06091261e-01 9.50829461e-02 9.71244991e-01 -6.96688771e-01
-1.22895205e+00 -3.65988687e-02 1.83924794e-01 -7.40023732e-01
-1.05037868e-01 -1.21420786e-01 -6.51811659e-01 -2.41308257e-01
6.74025536e-01 5.05655289e-01 3.46960947e-02 -1.53002381e-01
6.21244192e-01 5.31398892e-01 6.46614134e-01 -2.72973746e-01
1.13565242e+00 5.95006168e-01 1.31689474e-01 -8.41675282e-01
-3.25991541e-01 3.49549204e-01 -3.23579967e-01 -3.15504670e-01
1.00244451e+00 -5.49276829e-01 -1.83497369e-01 4.68536109e-01
-1.00735807e+00 -4.30894941e-01 -6.53621018e-01 -3.92084382e-02
-5.99817693e-01 2.50884652e-01 6.48901612e-03 -5.42158544e-01
-2.68908620e-01 -1.20409322e+00 1.02944922e+00 4.53030407e-01
-3.12835336e-01 -8.05851638e-01 1.37865162e-02 3.17737877e-01
5.93679845e-01 7.09758878e-01 9.93837297e-01 -1.37115598e-01
-6.84101999e-01 -2.39698201e-01 -2.13876203e-01 6.19358301e-01
3.72873276e-01 -4.48381938e-02 -9.84187841e-01 -3.16321909e-01
-7.31862858e-02 1.63528360e-02 5.69160402e-01 1.64093897e-01
1.11374950e+00 -7.38791585e-01 -1.89355671e-01 8.72451365e-01
1.41752958e+00 4.51429009e-01 1.13646758e+00 1.07555710e-01
7.78100967e-01 5.18294454e-01 3.30616653e-01 -1.73478015e-02
-3.58048677e-02 9.46436644e-01 4.96475428e-01 -3.14830691e-01
-8.16142917e-01 -6.87566936e-01 3.36916715e-01 2.32449263e-01
-8.71789902e-02 -6.90942168e-01 -3.90869945e-01 4.32961315e-01
-1.52656436e+00 -1.09655285e+00 7.51216486e-02 2.02887511e+00
6.53936327e-01 -1.56228647e-01 -6.88360035e-02 -2.91794892e-02
9.15811777e-01 1.20008744e-01 -5.85662663e-01 -5.50087333e-01
-3.58101845e-01 3.03833067e-01 4.00222033e-01 3.51474583e-01
-7.40973949e-01 1.20072162e+00 6.09530449e+00 9.74827290e-01
-1.45021021e+00 1.58552587e-01 9.53502297e-01 -3.77758108e-02
-7.03946650e-01 1.18779972e-01 -3.37605208e-01 7.95882225e-01
6.21664286e-01 -2.24579707e-01 4.71170336e-01 7.86641002e-01
2.14394227e-01 5.29298112e-02 -7.25816667e-01 1.09854627e+00
3.40358078e-01 -1.34337473e+00 4.65088695e-01 1.18028276e-01
1.14642537e+00 -7.25549936e-01 5.17461777e-01 -7.87478164e-02
2.32930332e-01 -1.18875933e+00 8.99816513e-01 6.95406258e-01
1.24260592e+00 -7.50511348e-01 4.44925010e-01 1.66701481e-01
-5.12911201e-01 1.90076604e-01 -2.67737687e-01 2.12949067e-01
1.90932199e-01 3.06472093e-01 -9.77744937e-01 3.19842190e-01
3.66136879e-01 3.04642469e-01 -5.43627143e-01 7.09811330e-01
-4.97073084e-01 4.98724282e-01 -5.68809034e-03 1.50659718e-02
2.29349732e-01 -5.34151852e-01 8.14687550e-01 1.00907910e+00
6.17467046e-01 -1.96503490e-01 -8.87863040e-02 1.35481441e+00
-1.22936644e-01 -1.34740129e-01 -8.42795432e-01 -7.34145343e-02
1.91589907e-01 1.21013784e+00 -6.49970174e-01 -3.83426219e-01
1.06483795e-01 1.69752395e+00 -2.59574294e-01 3.87958467e-01
-9.73720193e-01 -3.67399901e-01 5.79341948e-01 4.28630263e-01
2.72912532e-01 5.89609146e-02 -3.02528352e-01 -9.85587716e-01
-1.88643336e-01 -1.10309541e+00 -1.09503992e-01 -1.09951711e+00
-1.07409811e+00 8.55924070e-01 -5.23097105e-02 -1.22902763e+00
-4.37695742e-01 -3.49222302e-01 -7.39906132e-01 9.76804197e-01
-1.23558164e+00 -1.54569268e+00 -5.40886879e-01 6.34915769e-01
6.25124574e-01 -4.36182171e-01 7.35519469e-01 6.61694407e-02
-2.83575982e-01 6.37200534e-01 6.22686036e-02 3.55795585e-02
7.20814526e-01 -9.86832082e-01 3.97521704e-01 1.02237070e+00
2.32582301e-01 6.32406846e-02 9.52415466e-01 -6.85972273e-01
-1.01603854e+00 -1.36346829e+00 5.79036891e-01 -7.94754401e-02
3.82503681e-03 -3.53919119e-01 -5.27149677e-01 4.22285110e-01
5.83444118e-01 -2.65302509e-01 3.29595327e-01 -9.17752743e-01
-2.64861524e-01 -2.07339376e-01 -1.44518316e+00 8.83059204e-01
9.19471800e-01 -3.11150074e-01 -1.06500648e-01 1.20242260e-01
5.99619389e-01 -9.89946425e-02 -3.10747057e-01 5.05846553e-02
4.30949509e-01 -1.07953370e+00 9.80123580e-01 -3.21506590e-01
6.84383512e-01 -4.70543116e-01 3.08419601e-03 -1.53418589e+00
-4.06899959e-01 -9.47210371e-01 3.61627162e-01 1.27432430e+00
1.73840314e-01 -5.84585667e-01 7.72311091e-01 4.82817024e-01
2.78317146e-02 1.59680948e-03 -7.50228167e-01 -9.54827368e-01
-1.33989319e-01 1.23071849e-01 7.63854027e-01 8.43440950e-01
-7.66936064e-01 3.12277824e-01 -9.21195626e-01 -1.26797944e-01
5.80717623e-01 2.76699811e-01 1.09806311e+00 -7.90554345e-01
-2.79230535e-01 -4.28186476e-01 -6.84830904e-01 -4.27716136e-01
5.22148348e-02 -8.76070857e-01 -7.56690726e-02 -1.16236639e+00
3.65309604e-02 -2.76708573e-01 2.42463157e-01 2.93264061e-01
-1.07743055e-01 8.69933009e-01 5.87727308e-01 2.83387065e-01
5.85936606e-02 4.88847971e-01 1.77835071e+00 -4.26132470e-01
-1.69139192e-01 6.22090921e-02 -8.76629710e-01 4.92436916e-01
8.78211498e-01 -5.19611537e-01 -7.32804537e-01 -3.54214430e-01
5.82849025e-04 -1.24181740e-01 6.90973103e-01 -1.11245656e+00
-3.14349800e-01 -2.52473921e-01 7.30402172e-01 -1.71219796e-01
4.53012168e-01 -8.42833281e-01 8.42227757e-01 5.18671811e-01
-3.68199587e-01 -1.23199649e-01 -4.48687896e-02 6.17110908e-01
-2.46253729e-01 -1.70037434e-01 1.23366714e+00 -3.28172743e-01
-6.08817101e-01 2.11283356e-01 -3.54310185e-01 8.54322389e-02
1.32273316e+00 -4.19602692e-01 -1.72532573e-01 -8.24634790e-01
-4.43645120e-01 -4.92410690e-01 8.25040698e-01 4.43225145e-01
7.59522557e-01 -1.37665904e+00 -9.18428540e-01 3.15661162e-01
-3.74501781e-03 -3.26715648e-01 2.47253880e-01 1.35407016e-01
-6.67005002e-01 3.14898252e-01 -7.78080225e-01 -3.74493688e-01
-1.33479428e+00 6.68682694e-01 2.16974348e-01 -9.28977355e-02
-6.94255710e-01 6.17592812e-01 6.38936937e-01 -7.79352635e-02
-2.17984885e-01 4.05061208e-02 9.44171697e-02 -2.61194021e-01
4.90369141e-01 1.02899499e-01 -2.72453696e-01 -7.46645749e-01
-7.90574923e-02 4.59544569e-01 2.79915899e-01 -2.64839441e-01
1.19920170e+00 -2.97802556e-02 1.89558826e-02 -3.37751620e-02
1.17576969e+00 1.86714400e-02 -1.42072427e+00 7.86355585e-02
-5.90752542e-01 -7.72975147e-01 -5.40294610e-02 -9.53937113e-01
-1.38086569e+00 8.18842530e-01 7.25776553e-01 2.00699344e-02
1.37235332e+00 -2.31411561e-01 6.52908027e-01 -2.39481062e-01
3.27303767e-01 -8.21315646e-01 3.39908868e-01 -1.01290934e-01
1.21656597e+00 -9.03093874e-01 -1.90795287e-01 -2.81565905e-01
-1.06633472e+00 8.57680142e-01 5.11905372e-01 -4.19207454e-01
1.55545965e-01 5.00656329e-02 2.96451431e-02 4.68871407e-02
-2.52853245e-01 -1.47791788e-01 5.03357351e-01 1.14170647e+00
-3.59146148e-02 1.85768232e-01 -5.79808354e-02 7.73620978e-02
-4.38110590e-01 -1.25681192e-01 5.31339109e-01 3.71121407e-01
-1.04277804e-01 -1.27875304e+00 -6.86442554e-01 4.42676358e-02
-1.77711248e-01 8.44376460e-02 -7.64472127e-01 9.21803832e-01
3.88828695e-01 9.23084617e-01 1.14272043e-01 -2.81800985e-01
6.28371462e-02 -9.69664827e-02 6.70997679e-01 -3.98852766e-01
-4.47530776e-01 -7.51941279e-02 -1.36863291e-01 -3.77369791e-01
-3.24530095e-01 -3.79937559e-01 -6.22404814e-01 -3.84693325e-01
1.16434924e-01 -4.97281402e-02 8.21566284e-01 6.19881392e-01
5.09323776e-01 6.85546994e-01 7.35955417e-01 -9.73710716e-01
-4.48435731e-02 -7.68284321e-01 -4.46975142e-01 6.95509195e-01
5.36118709e-02 -3.12276483e-01 -1.98866382e-01 3.96528482e-01] | [11.83109188079834, -0.3908349871635437] |
93c1dcc0-5f34-4a6c-bb09-5665e8038a15 | global-stock-market-prediction-based-on-stock | 1902.10948 | null | http://arxiv.org/abs/1902.10948v1 | http://arxiv.org/pdf/1902.10948v1.pdf | Global Stock Market Prediction Based on Stock Chart Images Using Deep Q-Network | We applied Deep Q-Network with a Convolutional Neural Network function
approximator, which takes stock chart images as input, for making global stock
market predictions. Our model not only yields profit in the stock market of the
country where it was trained but generally yields profit in global stock
markets. We trained our model only in the US market and tested it in 31
different countries over 12 years. The portfolios constructed based on our
model's output generally yield about 0.1 to 1.0 percent return per transaction
prior to transaction costs in 31 countries. The results show that there are
some patterns on stock chart image, that tend to predict the same future stock
price movements across global stock markets. Moreover, the results show that
future stock prices can be predicted even if the training and testing
procedures are done in different countries. Training procedure could be done in
relatively large and liquid markets (e.g., USA) and tested in small markets.
This result demonstrates that artificial intelligence based stock price
forecasting models can be used in relatively small markets (emerging countries)
even though they do not have a sufficient amount of data for training. | ['Yookyung Koh', 'Raehyun Kim', 'Jaewoo Kang', 'Jinho Lee'] | 2019-02-28 | null | null | null | null | ['stock-market-prediction'] | ['time-series'] | [-9.96260047e-01 -8.19110200e-02 -3.30599248e-01 1.77967444e-01
-1.21153116e-01 -8.97405505e-01 5.82780004e-01 -3.51246625e-01
-3.28964174e-01 9.94576454e-01 8.47953781e-02 -8.93764317e-01
3.29355687e-01 -1.49704361e+00 -7.25502968e-01 -3.36276203e-01
-2.61476815e-01 4.28636819e-01 2.10218187e-02 -6.26033783e-01
4.45622444e-01 4.80090052e-01 -1.07193303e+00 5.48278317e-02
5.07594049e-01 1.38729608e+00 -6.68474883e-02 4.60995615e-01
-4.59054232e-01 1.35507798e+00 -1.01662242e+00 -8.00925672e-01
1.12903059e+00 -2.99429327e-01 -2.32349843e-01 -4.19393748e-01
1.67542204e-01 -9.14279938e-01 -1.43501624e-01 1.03982699e+00
-4.03100887e-04 -3.42725217e-01 6.22360766e-01 -1.11630976e+00
-1.03567958e+00 8.01885545e-01 -4.79644477e-01 4.52432960e-01
-2.01059222e-01 3.11763465e-01 1.27332306e+00 -6.84636891e-01
5.22035003e-01 9.21562493e-01 7.42434919e-01 6.79929703e-02
-8.58001113e-01 -1.18704391e+00 -2.70663589e-01 -4.17093784e-01
-5.18069565e-01 2.18284100e-01 5.45180619e-01 -5.67511618e-01
1.21591723e+00 3.51824798e-02 1.12353647e+00 5.00997424e-01
1.06582868e+00 5.16461492e-01 9.85154629e-01 -3.94578651e-02
2.34292746e-01 1.00730591e-01 -4.14360583e-01 6.63604364e-02
5.59980094e-01 6.42889023e-01 -1.73473835e-01 -7.06812963e-02
1.26781523e+00 1.65964022e-01 2.11054772e-01 2.29408532e-01
-1.34892213e+00 1.20972383e+00 7.52832413e-01 5.82916379e-01
-6.96493030e-01 1.82489485e-01 4.81914580e-01 1.07198203e+00
5.59082508e-01 6.88393950e-01 -8.20607722e-01 -1.98137417e-01
-1.19708645e+00 5.52817941e-01 1.02070892e+00 4.66562837e-01
5.36266804e-01 8.68287444e-01 2.79970735e-01 2.04370841e-01
-1.49261868e-02 7.80648768e-01 9.31690633e-01 -7.32647777e-01
6.00333750e-01 5.81558645e-01 5.98481715e-01 -1.04161358e+00
-4.66596186e-01 -6.04062498e-01 -5.75020373e-01 7.96060145e-01
7.63553500e-01 -6.85232818e-01 -7.58510113e-01 1.05681908e+00
-4.94481027e-01 2.38535970e-01 5.09549499e-01 5.91731846e-01
3.10343802e-01 1.01125813e+00 -1.60190314e-01 1.55218225e-02
9.82176781e-01 -7.74434924e-01 -4.88208055e-01 4.16268669e-02
4.47153300e-01 -5.01525998e-01 4.89972383e-01 3.07335466e-01
-1.34260941e+00 -5.42324364e-01 -1.02532399e+00 6.84594214e-01
-6.53502762e-01 -2.01112300e-01 7.52126634e-01 6.44988239e-01
-1.22689307e+00 1.00573337e+00 -6.06922805e-01 4.04080361e-01
2.41536543e-01 3.67681026e-01 -7.12406961e-03 9.48802412e-01
-1.51731837e+00 1.23840272e+00 6.15542352e-01 2.48559527e-02
-5.28942823e-01 -5.19625604e-01 -5.90771377e-01 5.07654130e-01
-2.18739271e-01 -1.60180077e-01 1.49187279e+00 -1.67359006e+00
-1.35992706e+00 3.38401139e-01 7.48844564e-01 -1.19010544e+00
6.09451354e-01 1.33667752e-01 -6.44950986e-01 -1.56159639e-01
8.28329399e-02 4.91638929e-01 5.87011755e-01 -5.80617428e-01
-1.07799530e+00 -2.16461629e-01 1.14046052e-01 -9.95262712e-03
-3.77433077e-02 2.37874389e-01 4.73658353e-01 -1.06801379e+00
-8.33860785e-02 -5.74918926e-01 -6.04899451e-02 -5.95942914e-01
3.50348145e-01 -1.03838459e-01 3.54239136e-01 -9.45649922e-01
9.60542858e-01 -1.57942355e+00 -7.18281269e-01 4.29707021e-01
-2.47085959e-01 -8.92699137e-02 6.13221712e-02 6.00726128e-01
-3.49115342e-01 4.38305050e-01 1.13073878e-01 5.22906840e-01
2.55816519e-01 -1.44828901e-01 -8.34232330e-01 2.01334819e-01
3.01770061e-01 1.43957686e+00 -5.32942593e-01 1.99613065e-01
-3.49657461e-02 -3.16696942e-01 -2.64559209e-01 -1.16899602e-01
-3.75790805e-01 -1.00908555e-01 -1.94885761e-01 7.78473675e-01
8.22011530e-01 -2.36189172e-01 2.16263738e-02 4.16637421e-01
-5.58765948e-01 2.40616903e-01 -1.10540485e+00 7.55339563e-01
-2.75992095e-01 6.96273386e-01 -4.85331982e-01 -8.55626941e-01
1.27070653e+00 3.93135935e-01 4.85492647e-01 -1.30264127e+00
6.13233857e-02 9.27002788e-01 5.20856559e-01 1.05879381e-01
6.94778979e-01 -8.09080720e-01 -5.45288995e-02 8.12028766e-01
-7.84209296e-02 -9.59500670e-02 3.77603054e-01 -3.77260417e-01
7.24551380e-01 -2.05264566e-03 1.94139972e-01 -3.75740618e-01
1.66060448e-01 1.93877324e-01 6.06221259e-01 4.58950311e-01
-2.83322215e-01 2.26417035e-01 9.32522833e-01 -1.25288010e+00
-1.51359499e+00 -1.03836274e+00 -1.13622680e-01 8.47260118e-01
-4.05506670e-01 4.38218921e-01 -4.54964906e-01 -4.32470798e-01
4.52231467e-01 6.43616676e-01 -5.33333600e-01 4.46305335e-01
-3.48199695e-01 -7.15709984e-01 3.55866641e-01 6.68003082e-01
6.94020569e-01 -1.72655737e+00 -9.28473055e-01 5.37386835e-01
7.57375181e-01 -4.32280630e-01 -8.61690789e-02 1.94329768e-01
-1.15259540e+00 -8.72643769e-01 -1.29019284e+00 -6.33573055e-01
2.12500677e-01 -4.92208749e-01 1.51368797e+00 -8.64581540e-02
7.23137379e-01 -1.44395724e-01 7.33234733e-02 -1.10712922e+00
-3.97415787e-01 1.04128765e-02 -6.85697049e-02 -3.19850653e-01
5.45302987e-01 -2.22632781e-01 -6.96167409e-01 2.55630791e-01
-8.01163852e-01 -3.05036306e-01 4.14592594e-01 9.79986072e-01
-1.36415428e-02 3.79375994e-01 1.22099435e+00 -6.20596349e-01
8.15304399e-01 -6.86045349e-01 -1.43675268e+00 1.14773273e-01
-9.78491426e-01 -7.34435841e-02 7.37712979e-01 -1.45768672e-01
-8.98371041e-01 -3.31760406e-01 2.56526265e-02 -2.16439396e-01
4.01897103e-01 9.60833728e-01 6.79245353e-01 4.35796916e-01
4.09411252e-01 2.46629030e-01 1.42829731e-01 -1.50178403e-01
-1.65127307e-01 3.83110732e-01 3.35514069e-01 -4.76833954e-02
9.68438983e-01 1.14091799e-01 -2.99361229e-01 4.33502533e-02
-1.53387925e-02 9.22665074e-02 -4.80524331e-01 -1.67022154e-01
5.13580322e-01 -1.15920854e+00 -7.06601679e-01 8.29288185e-01
-8.51682782e-01 -3.09612215e-01 -4.61463720e-01 5.75106561e-01
-4.06708866e-01 -4.03807014e-01 -8.97973239e-01 -1.14166522e+00
-6.05869412e-01 -7.46649206e-01 3.39700073e-01 4.97681469e-01
-1.06954813e-01 -1.35796225e+00 3.33401233e-01 -1.97096005e-01
7.52917707e-01 3.72262895e-01 5.47219574e-01 -9.48870599e-01
-6.07762992e-01 -3.11707556e-01 -1.76099673e-01 6.46413386e-01
2.94391692e-01 6.73422888e-02 -6.32199287e-01 -2.71489978e-01
3.15277069e-03 -2.73576558e-01 7.99564362e-01 6.79290712e-01
1.18539385e-01 -3.98996204e-01 4.49956447e-01 2.58228689e-01
1.60102105e+00 1.09526968e+00 7.87189901e-01 1.15906501e+00
-1.21580578e-01 4.82953161e-01 3.69482845e-01 2.78026313e-01
1.73966214e-01 -1.60156161e-01 3.12197536e-01 -6.37755319e-02
6.97355688e-01 -3.67459238e-01 6.68715954e-01 5.67696333e-01
-3.12021732e-01 2.41162613e-01 -1.10400867e+00 5.19495845e-01
-1.54109406e+00 -1.35331154e+00 2.87740946e-01 1.97995126e+00
5.96742511e-01 6.99011922e-01 5.87447464e-01 -8.53073522e-02
3.40007365e-01 1.86857864e-01 -9.85638559e-01 -6.89966857e-01
-3.81682843e-01 5.34267545e-01 1.03771663e+00 3.62380012e-03
-8.76147270e-01 8.70771110e-01 7.34370899e+00 2.65175104e-01
-1.65407193e+00 -4.33479965e-01 1.33452928e+00 6.71295449e-03
-6.28315628e-01 -9.78114679e-02 -4.01903212e-01 7.59984374e-01
1.37389350e+00 -7.11851656e-01 3.30777854e-01 9.56954956e-01
1.80631086e-01 -3.26830633e-02 -5.35565317e-01 8.28238368e-01
-7.39585578e-01 -1.98597407e+00 -1.34427533e-01 1.66262269e-01
1.00738442e+00 2.08101004e-01 3.72074395e-01 5.04800737e-01
5.45156002e-01 -1.11176431e+00 1.16044140e+00 7.50451505e-01
3.94492567e-01 -1.16362011e+00 1.24254954e+00 3.66456240e-01
-1.12045097e+00 -3.93128932e-01 -6.03582680e-01 -6.12889230e-01
-3.02158773e-01 1.94318458e-01 -6.06730103e-01 3.23693126e-01
9.09192383e-01 6.70381427e-01 -2.67010778e-01 5.60749471e-01
2.13420361e-01 5.55687010e-01 -3.06105524e-01 -3.58161986e-01
7.38668978e-01 -7.72611320e-01 -3.36212575e-01 6.43318594e-01
8.91784370e-01 -1.38767377e-01 -4.72471952e-01 1.11129665e+00
-2.00358182e-01 2.32596472e-01 -8.47730517e-01 -2.93313980e-01
8.39202404e-02 5.93333721e-01 -7.05706537e-01 -3.09689581e-01
-9.82521117e-01 6.22612298e-01 -1.14357792e-01 3.20906073e-01
-5.81812680e-01 -3.58963698e-01 4.76885259e-01 2.04507142e-01
4.55767214e-01 -3.20815668e-02 -6.36894107e-01 -1.23854125e+00
4.05554436e-02 -7.29075730e-01 3.23388040e-01 -8.16764951e-01
-1.29608095e+00 3.29562247e-01 -3.27614725e-01 -1.42216980e+00
-8.92468452e-01 -1.12621081e+00 -1.16576874e+00 1.29936051e+00
-1.70027232e+00 -5.66396892e-01 7.34059036e-01 2.91940451e-01
2.72454053e-01 -1.08332121e+00 4.86860216e-01 -2.78375059e-01
-2.57453948e-01 1.40454799e-01 7.14035451e-01 7.50507474e-01
-4.18789126e-02 -1.55381989e+00 1.15006983e+00 6.40885532e-01
1.42197371e-01 4.52539593e-01 2.04216942e-01 -8.85989010e-01
-7.60976493e-01 -5.71774006e-01 8.81066620e-01 -2.48705968e-01
1.33355129e+00 2.41564199e-01 -8.08382988e-01 9.52451348e-01
8.98159444e-01 -3.74742687e-01 4.57572639e-01 -3.09056699e-01
-1.30026951e-01 -3.38724136e-01 -1.15200388e+00 3.98960024e-01
8.74449387e-02 -2.25256175e-01 -8.24366510e-01 -1.35588795e-01
4.23519820e-01 -2.43039876e-01 -1.13536274e+00 -3.30611430e-02
8.85668933e-01 -1.29638684e+00 6.43456757e-01 -6.85021937e-01
5.70913911e-01 1.04913525e-01 1.90366775e-01 -1.53808475e+00
-2.05531478e-01 -5.16852438e-01 9.34464261e-02 8.55983675e-01
9.66017365e-01 -1.41725624e+00 1.07492924e+00 1.12254500e+00
3.97493631e-01 -6.09648645e-01 -1.06419826e+00 -9.43917930e-01
1.01681972e+00 -2.35461369e-01 1.45004821e+00 8.50947022e-01
-5.66513371e-03 -2.34308884e-01 -2.18529716e-01 -2.61933833e-01
2.27076396e-01 5.41531444e-01 6.45573616e-01 -1.22813594e+00
1.39698580e-01 -8.48139405e-01 -3.38061333e-01 -3.90112728e-01
2.45158225e-01 -6.57037079e-01 -7.17075348e-01 -1.22924161e+00
-3.35073680e-01 -2.77135044e-01 -7.97729611e-01 4.69680816e-01
2.92930752e-01 -1.44735545e-01 5.77578068e-01 4.77269411e-01
4.84842628e-01 2.63939053e-01 1.47482049e+00 -3.08581829e-01
-8.10804218e-02 2.45752856e-01 -7.05484509e-01 5.95709264e-01
1.19142914e+00 -8.64735618e-02 -1.41565949e-01 8.70136619e-02
8.10318291e-01 3.02273661e-01 1.45520028e-02 -9.73658383e-01
-4.47810516e-02 -4.64360982e-01 1.05464053e+00 -9.21679556e-01
-3.55929047e-01 -1.00505745e+00 5.41528940e-01 9.02777255e-01
9.73182097e-02 7.66297340e-01 4.27341789e-01 -2.28386223e-02
-6.50685787e-01 -3.77953798e-01 6.28062248e-01 -6.01168752e-01
-7.73611188e-01 2.57418811e-01 -3.52963388e-01 -2.90998995e-01
1.16345215e+00 -4.93628353e-01 -2.81250983e-01 -6.74244583e-01
-4.28902686e-01 2.57540077e-01 5.01697898e-01 4.17937219e-01
4.35445130e-01 -1.60819304e+00 -8.63990247e-01 3.99393708e-01
-3.86563450e-01 -4.43624049e-01 -3.66131157e-01 7.95153603e-02
-1.30687392e+00 6.43179595e-01 -7.87935913e-01 -4.08099480e-02
-6.99199811e-02 4.31263924e-01 9.41617131e-01 -5.56178212e-01
-4.96844023e-01 3.50620329e-01 -2.19023228e-03 -3.85069668e-01
-2.65821069e-01 -1.07757890e+00 -3.60181004e-01 6.97088361e-01
5.89832306e-01 1.50089398e-01 -3.08859535e-03 -5.95180154e-01
1.51339084e-01 4.45992887e-01 1.98825166e-01 -3.81830156e-01
1.80061245e+00 4.09825236e-01 1.27717899e-02 6.54964030e-01
9.19213891e-01 -1.48079947e-01 -1.51494420e+00 2.04350755e-01
5.07769704e-01 -4.22179490e-01 2.27367748e-02 -8.29014480e-01
-1.68463898e+00 4.09597933e-01 5.76412737e-01 8.12940717e-01
1.07748747e+00 -6.20997429e-01 1.14615810e+00 2.99552500e-01
4.64204222e-01 -1.35387254e+00 -4.41252083e-01 6.06339812e-01
9.79925275e-01 -1.19364357e+00 -2.25940719e-01 7.76726186e-01
-7.66498268e-01 1.56443918e+00 2.48233438e-01 -7.45322704e-01
1.00662982e+00 3.04551750e-01 4.51507121e-01 -1.25762418e-01
-7.59519041e-01 1.67999223e-01 4.18188386e-02 3.22345704e-01
4.58249152e-01 3.55180174e-01 -9.88800377e-02 6.39362991e-01
-8.06645036e-01 1.57125577e-01 7.73753464e-01 7.54895508e-01
-4.46204454e-01 -7.71369338e-01 -6.70114517e-01 8.19918275e-01
-8.85442793e-01 -2.97928363e-01 -2.96822965e-01 1.21279514e+00
-1.99543074e-01 3.00479710e-01 8.09050620e-01 -1.55568838e-01
4.39417630e-01 2.85785854e-01 -1.21423759e-01 -2.52121747e-01
-1.13248801e+00 1.52900115e-01 -2.15510637e-01 -1.95614681e-01
-2.77042091e-01 -8.61644626e-01 -1.37019324e+00 -7.00017810e-01
-2.70017665e-02 1.08647786e-01 4.10348266e-01 4.56089437e-01
-1.22463532e-01 6.06817119e-02 1.00880373e+00 -6.78227842e-01
-9.20182765e-01 -1.09522545e+00 -1.46188629e+00 2.12441117e-01
5.36089420e-01 -2.39422336e-01 -4.86765653e-01 -4.04413454e-02] | [4.456366062164307, 4.217419147491455] |
992ff40f-432d-4e8d-b000-411d48776f8f | dali-dynamically-adjusted-label-importance | 2301.12077 | null | https://arxiv.org/abs/2301.12077v2 | https://arxiv.org/pdf/2301.12077v2.pdf | ALIM: Adjusting Label Importance Mechanism for Noisy Partial Label Learning | Noisy partial label learning (noisy PLL) is an important branch of weakly supervised learning. Unlike PLL where the ground-truth label must conceal in the candidate label set, noisy PLL relaxes this constraint and allows the ground-truth label may not be in the candidate label set. To address this challenging problem, most of the existing works attempt to detect noisy samples and estimate the ground-truth label for each noisy sample. However, detection errors are unavoidable. These errors can accumulate during training and continuously affect model optimization. To this end, we propose a novel framework for noisy PLL with theoretical guarantees, called ``Adjusting Label Importance Mechanism (ALIM)''. It aims to reduce the negative impact of detection errors by trading off the initial candidate set and model outputs. ALIM is a plug-in strategy that can be integrated with existing PLL approaches. Experimental results on benchmark datasets demonstrate that our method can achieve state-of-the-art performance on noisy PLL. \textcolor[rgb]{0.93,0.0,0.47}{Our code can be found in Supplementary Material}. | ['JianHua Tao', 'Bin Liu', 'Lei Feng', 'Zheng Lian', 'Mingyu Xu'] | 2023-01-28 | null | null | null | null | ['partial-label-learning'] | ['methodology'] | [ 3.31913650e-01 2.77222842e-01 -5.64482033e-01 -4.30791527e-01
-1.38344228e+00 -7.37851560e-01 1.25575900e-01 2.23818749e-01
-4.21092182e-01 9.78647530e-01 -3.60782713e-01 -3.42402346e-02
8.69056582e-02 -5.37832737e-01 -9.39100444e-01 -1.00158727e+00
1.51971981e-01 2.95705467e-01 3.06431800e-01 4.18217689e-01
-1.28534976e-02 1.28247157e-01 -1.48649013e+00 3.06549102e-01
6.87789083e-01 1.26668990e+00 1.82053775e-01 2.90828407e-01
-2.03195468e-01 1.04292893e+00 -7.65412569e-01 -2.20479682e-01
2.66167432e-01 -4.59697276e-01 -5.79138279e-01 1.48728430e-01
3.65087807e-01 2.76193372e-03 1.50063545e-01 1.39499092e+00
6.45826161e-01 -1.10255875e-01 5.19368768e-01 -1.43896568e+00
-3.14624935e-01 7.29301810e-01 -6.26766264e-01 -1.13464892e-01
5.60166277e-02 1.69765130e-01 1.15033102e+00 -9.44452465e-01
4.96076405e-01 1.05746472e+00 7.89875209e-01 6.22392774e-01
-1.25041902e+00 -7.53840685e-01 4.62882131e-01 -1.48761243e-01
-1.48587763e+00 -3.22297961e-01 7.50174761e-01 -1.76304877e-01
1.98625922e-01 3.72527987e-01 5.43107033e-01 1.03642988e+00
-3.83620560e-02 1.09177017e+00 1.65879047e+00 -4.79093194e-01
4.46360499e-01 2.82528579e-01 2.98767298e-01 9.13337350e-01
2.44791448e-01 1.18463345e-01 -9.29962575e-01 -1.74586505e-01
4.39968288e-01 -2.95635492e-01 -4.16516930e-01 -2.14177758e-01
-1.14509571e+00 4.91567552e-01 4.57975745e-01 -3.55966054e-02
7.61077479e-02 4.93766934e-01 2.53525823e-01 2.42101818e-01
6.76318347e-01 3.73231739e-01 -6.13352835e-01 -8.95306189e-03
-9.96546805e-01 9.62141994e-03 6.62379563e-01 9.75284219e-01
9.25057471e-01 -1.95038140e-01 -5.46428621e-01 7.51341403e-01
4.80749309e-01 4.35499221e-01 2.62229502e-01 -1.17259634e+00
1.43890262e-01 5.14838576e-01 2.03072742e-01 -5.19934893e-01
-4.03925091e-01 -8.65733325e-01 -7.07052231e-01 1.81511834e-01
6.09717309e-01 -1.76249921e-01 -9.22513962e-01 1.72068059e+00
6.92983031e-01 5.47440112e-01 -1.71051666e-01 9.34737742e-01
9.23997104e-01 4.19178098e-01 3.70521806e-02 -4.58532214e-01
9.81325626e-01 -1.08527541e+00 -8.79683375e-01 -2.81526148e-01
7.70769656e-01 -8.64274561e-01 1.45195782e+00 6.74629331e-01
-7.67646611e-01 -3.74508828e-01 -9.68052089e-01 4.28073816e-02
-8.43709037e-02 3.90353382e-01 5.44629991e-01 8.77779245e-01
-7.65486181e-01 5.61603844e-01 -6.95683181e-01 1.86110020e-01
5.68466067e-01 2.99768895e-01 3.38969491e-02 -2.18785554e-01
-9.06063855e-01 4.08135414e-01 3.67508531e-01 2.61510372e-01
-9.02248621e-01 -6.74606740e-01 -6.29638970e-01 -3.64509583e-01
7.55377650e-01 -2.80115038e-01 1.47957289e+00 -1.08250618e+00
-1.38267589e+00 9.00452077e-01 -1.95571035e-01 -1.32765144e-01
8.20851266e-01 -1.51542306e-01 2.93850359e-02 -2.71312386e-01
2.04094574e-01 7.18128860e-01 8.21374536e-01 -1.76782548e+00
-8.10481787e-01 -1.85188707e-02 -1.60592154e-01 -2.43673325e-02
-2.35509962e-01 -3.41816038e-01 -7.05230653e-01 -8.59894454e-01
3.78386706e-01 -9.65043545e-01 -2.75210977e-01 2.84715682e-01
-6.24006569e-01 -3.84801656e-01 7.43863523e-01 -2.88261827e-02
1.27702701e+00 -2.03450847e+00 -3.37112665e-01 3.43191236e-01
2.22280145e-01 5.62531166e-02 1.46943200e-02 -5.41845113e-02
2.62458354e-01 2.67612964e-01 -3.04030508e-01 -7.74518847e-01
1.96220264e-01 3.51204574e-01 -8.64769891e-02 7.41695166e-01
9.58850011e-02 8.04457426e-01 -1.26782489e+00 -5.53908110e-01
1.08645968e-01 2.22774744e-01 -2.71588027e-01 1.22690670e-01
-5.55519521e-01 5.82182765e-01 -3.24034989e-01 1.05290282e+00
7.01410472e-01 -4.19711053e-01 3.30828764e-02 -2.16267347e-01
1.19048730e-01 -1.75892767e-02 -1.54369116e+00 1.49942112e+00
-2.07550138e-01 3.11231583e-01 7.17074871e-02 -6.25274956e-01
8.06620359e-01 1.59265980e-01 5.43824077e-01 -5.81520677e-01
2.89118290e-01 4.17542726e-01 -4.23331290e-01 -3.58064622e-01
3.00739378e-01 -1.28401935e-01 -1.94323108e-01 4.61217016e-01
-1.52691260e-01 -1.27310291e-01 1.13375179e-01 -9.99124944e-02
9.09271121e-01 3.95467997e-01 9.15994272e-02 -2.19173267e-01
3.99245888e-01 -7.80993104e-02 9.51351047e-01 1.16885591e+00
-4.54210967e-01 7.51927912e-01 6.13172233e-01 -1.59286097e-01
-6.78399861e-01 -8.49987447e-01 -2.66990930e-01 1.02672946e+00
3.59064758e-01 -4.24403071e-01 -8.77823591e-01 -1.06190538e+00
-1.25418454e-01 4.38012868e-01 -5.47723353e-01 -1.51954859e-01
-2.91707397e-01 -9.43717241e-01 4.92586344e-01 3.98348510e-01
3.86664867e-01 -9.09307837e-01 -1.82624891e-01 1.41405731e-01
-3.06280017e-01 -1.02965736e+00 -4.88847315e-01 5.95881522e-01
-5.29563963e-01 -1.07339883e+00 -3.76330048e-01 -8.10220361e-01
1.01988411e+00 1.54202089e-01 1.18949962e+00 3.55368346e-01
4.28133458e-02 2.70960838e-01 -5.20309925e-01 -4.91052330e-01
-4.42169040e-01 1.43124625e-01 -1.25376418e-01 6.83911517e-02
2.47978851e-01 -3.88043672e-02 -4.59766626e-01 4.69188541e-01
-9.77647960e-01 -3.50726815e-03 4.02769059e-01 9.59446311e-01
1.36255050e+00 2.06403360e-01 6.31199002e-01 -1.14823854e+00
3.17761838e-01 -4.23455566e-01 -8.42375517e-01 3.38659108e-01
-8.29163134e-01 1.19819298e-01 5.15445828e-01 -5.80964744e-01
-7.26146638e-01 4.97303724e-01 -1.81123346e-01 -3.57517064e-01
2.67440230e-02 3.84968847e-01 -3.05494994e-01 -1.22862458e-01
4.76461172e-01 -1.15414381e-01 -4.16430324e-01 -5.11770666e-01
2.34559998e-01 6.43678129e-01 5.30994415e-01 -7.63387680e-01
6.25285923e-01 6.23763204e-01 8.81024152e-02 -2.87898302e-01
-1.55509889e+00 -6.46878660e-01 -5.27367234e-01 -5.37936330e-01
2.83780903e-01 -1.09608793e+00 -6.42329276e-01 5.45951426e-01
-7.94150174e-01 -6.83080733e-01 -4.94249046e-01 1.82370111e-01
-2.84148425e-01 2.56551653e-01 -4.59193826e-01 -8.79953623e-01
-1.36101946e-01 -1.32901669e+00 1.34617412e+00 2.00608313e-01
-8.99183676e-02 -7.69991577e-01 -2.49329671e-01 3.43637079e-01
-5.48895784e-02 3.93539876e-01 4.07658130e-01 -2.59538084e-01
-7.41220236e-01 -2.56253779e-01 -1.92092791e-01 4.37439203e-01
1.32738739e-01 -1.10788852e-01 -1.32703078e+00 -4.39674169e-01
-3.74847837e-02 -7.16709375e-01 1.15289521e+00 3.43034267e-01
1.34638023e+00 -1.62482560e-01 -2.42630675e-01 5.00330985e-01
1.51688051e+00 -3.23047191e-01 2.56920129e-01 3.14859778e-01
8.89723003e-01 4.98130649e-01 1.08507276e+00 5.83993137e-01
3.10503602e-01 5.40061474e-01 5.57885051e-01 -2.13725924e-01
-3.83491099e-01 -3.20912153e-01 4.29541856e-01 6.63931668e-01
4.99412268e-01 -4.72701341e-01 -7.89663851e-01 3.28889012e-01
-2.03104162e+00 -3.56197506e-01 -4.91945148e-01 2.14643073e+00
1.23436117e+00 4.94410783e-01 -8.76774266e-02 5.23469985e-01
7.07617044e-01 2.81155687e-02 -6.16073608e-01 1.60213068e-01
-2.28251234e-01 -5.05793579e-02 7.27527440e-01 7.14085460e-01
-1.30543292e+00 1.10659015e+00 5.44480705e+00 1.23023450e+00
-1.12957168e+00 4.40850258e-01 9.32123661e-01 -2.04426289e-01
-3.06236863e-01 -1.13434665e-01 -1.10478747e+00 7.04056621e-01
6.13660395e-01 3.21084797e-01 2.38802075e-01 8.12799931e-01
3.86356592e-01 -4.48338568e-01 -1.11161792e+00 8.67918432e-01
-6.90777227e-02 -1.03601062e+00 -3.97876561e-01 -1.43549323e-01
1.08343613e+00 -4.40470725e-02 1.28376886e-01 3.19768697e-01
4.17287797e-01 -8.48827124e-01 1.19558907e+00 5.94634831e-01
9.34599400e-01 -8.11965942e-01 8.31540585e-01 5.98496854e-01
-1.13249862e+00 4.76471037e-02 -3.14570457e-01 1.27570197e-01
-1.71175301e-02 1.28027236e+00 -6.71343684e-01 1.89278722e-01
8.29962373e-01 6.49022043e-01 -6.49818361e-01 1.11425388e+00
-6.83588624e-01 9.68656123e-01 -4.12081003e-01 -4.22870442e-02
2.09154159e-01 9.38309878e-02 2.97278136e-01 1.06276011e+00
2.18912903e-02 -2.18658745e-01 6.34689867e-01 6.60802901e-01
-4.32127446e-01 1.04208574e-01 -2.00534284e-01 3.40576887e-01
5.69578469e-01 1.19820249e+00 -1.23353517e+00 -2.67233849e-01
-1.40001446e-01 8.00215960e-01 3.27627808e-01 2.18289465e-01
-9.18928146e-01 1.14902578e-01 1.34042472e-01 -4.46848646e-02
5.10990545e-02 1.45090312e-01 -5.57411492e-01 -9.88141477e-01
2.48362809e-01 -9.43573833e-01 2.50260442e-01 -5.51974356e-01
-1.44759893e+00 2.92837739e-01 -3.91659886e-01 -1.35352647e+00
3.25798899e-01 -3.87724549e-01 -1.22688543e-02 4.80412632e-01
-1.84426880e+00 -1.17901373e+00 -4.36375856e-01 2.88272291e-01
5.84141672e-01 3.45033020e-01 5.54896116e-01 3.16633761e-01
-6.81776345e-01 8.48188818e-01 3.31875652e-01 -1.03618354e-01
9.11126316e-01 -1.39640844e+00 5.02044521e-03 7.27466285e-01
1.21506527e-01 1.29391342e-01 5.69778442e-01 -7.19221950e-01
-1.06343269e+00 -1.56821358e+00 7.59943962e-01 -4.74410862e-01
5.38575530e-01 -5.35711288e-01 -9.30498123e-01 3.93883169e-01
-2.95976758e-01 5.58742046e-01 7.51990795e-01 -2.00776219e-01
-2.80254692e-01 -2.72575021e-01 -1.26742446e+00 4.18451011e-01
9.80137289e-01 -3.42470348e-01 2.45803192e-01 7.25895882e-01
7.71365702e-01 -7.64854789e-01 -7.78120875e-01 5.61912239e-01
4.10432667e-01 -9.05196488e-01 7.34785140e-01 -2.63990946e-02
1.20077953e-02 -7.00839818e-01 -1.54915571e-01 -1.00492477e+00
-7.78841823e-02 -7.58336902e-01 -3.22839200e-01 1.44975185e+00
4.66761976e-01 -3.53246033e-01 1.09219360e+00 3.14541847e-01
-2.06712216e-01 -1.03284669e+00 -1.11408556e+00 -7.72470117e-01
-2.38449529e-01 -8.74517143e-01 4.70552862e-01 8.32789302e-01
-4.48846132e-01 3.84375565e-02 -5.86798191e-01 4.24564779e-01
9.02678907e-01 -9.73730683e-02 4.74683702e-01 -1.15377367e+00
-3.63985926e-01 -2.58890510e-01 9.91396084e-02 -1.10901320e+00
2.35089839e-01 -8.99673939e-01 5.70214331e-01 -1.31409824e+00
7.07673207e-02 -1.05052352e+00 -6.34227991e-01 8.12926233e-01
-5.51209390e-01 7.89246261e-01 6.96199834e-02 3.55623484e-01
-1.03075600e+00 4.96725112e-01 1.43131840e+00 -1.75997645e-01
-4.83076312e-02 1.00193039e-01 -5.90266526e-01 8.52578521e-01
8.65324378e-01 -9.89235461e-01 -3.55603009e-01 -1.80378035e-01
4.73262548e-01 -3.52426082e-01 2.70572394e-01 -9.30990994e-01
2.01353699e-01 -1.51095062e-01 2.59114176e-01 -6.12645626e-01
2.07202300e-01 -8.93096566e-01 -7.46131539e-02 4.15425718e-01
-4.33775067e-01 -3.65935117e-01 -5.97551316e-02 8.28643918e-01
-4.63531464e-02 -4.98072118e-01 9.60391939e-01 -1.54361770e-01
-5.57949126e-01 1.92552820e-01 -1.31444886e-01 1.35285556e-01
9.76002932e-01 -6.48917034e-02 -2.30937734e-01 -7.47860223e-02
-5.56552112e-01 5.44616878e-01 4.52516407e-01 9.50243548e-02
3.71305346e-01 -1.32897663e+00 -4.92422342e-01 -4.88658473e-02
2.84605235e-01 4.26445484e-01 -2.33517602e-01 7.74429440e-01
-2.09244743e-01 3.10853273e-02 7.16973484e-01 -8.31427693e-01
-1.19570518e+00 3.24743927e-01 4.10820484e-01 -3.04005504e-01
-2.34638512e-01 1.14800239e+00 -2.76562989e-01 -6.06676817e-01
8.43914926e-01 -4.56922412e-01 5.93070760e-02 8.42973068e-02
3.36291224e-01 3.13674331e-01 1.10834904e-01 -5.06800115e-01
-1.79535389e-01 4.88510877e-01 1.96091861e-01 9.25847515e-02
1.12946939e+00 -2.16518939e-01 -2.39488050e-01 8.19743872e-01
1.01033604e+00 1.62015915e-01 -1.55653274e+00 -5.24528503e-01
2.78739631e-01 -4.05525506e-01 3.46420020e-01 -9.16431665e-01
-1.18119287e+00 4.21847939e-01 7.74640620e-01 1.22852810e-01
1.13792622e+00 1.04823157e-01 7.58180320e-01 3.75070363e-01
4.71640468e-01 -1.28310418e+00 2.88675070e-01 2.92754054e-01
4.98796493e-01 -1.51002562e+00 1.99332740e-02 -6.33348465e-01
-3.53625894e-01 7.69179404e-01 6.94532394e-01 1.26812562e-01
6.00165844e-01 4.87227529e-01 3.56729418e-01 -1.09986536e-01
-6.37225986e-01 -2.70327359e-01 1.48026481e-01 3.69357407e-01
4.04585630e-01 -2.95484047e-02 -3.65927607e-01 6.43164635e-01
1.74174178e-02 3.11552901e-02 3.77830923e-01 1.14902270e+00
-6.01991355e-01 -1.39236474e+00 -7.26538181e-01 4.68591332e-01
-5.53486705e-01 -1.92179959e-02 -3.42505068e-01 4.12167698e-01
8.83106291e-01 1.06532383e+00 -3.15028995e-01 -2.12999240e-01
2.61148602e-01 -7.86805004e-02 3.92060488e-01 -7.24834859e-01
-5.58010936e-01 5.50406456e-01 -2.40330361e-02 -4.37592775e-01
-7.14999378e-01 -5.76819420e-01 -1.45794857e+00 1.23241186e-01
-7.41132081e-01 1.31981373e-01 4.47087169e-01 9.41025317e-01
-9.76597592e-02 6.38257921e-01 6.35261178e-01 -5.82281470e-01
-4.60940361e-01 -7.63343751e-01 -5.96271455e-01 1.78944424e-01
5.34861743e-01 -6.00326240e-01 -4.96357411e-01 2.30565131e-01] | [9.436103820800781, 3.944321632385254] |
2237aaeb-0f05-43e8-9b1b-973c412644de | transformers-in-3d-point-clouds-a-survey | 2205.07417 | null | https://arxiv.org/abs/2205.07417v2 | https://arxiv.org/pdf/2205.07417v2.pdf | Transformers in 3D Point Clouds: A Survey | Transformers have been at the heart of the Natural Language Processing (NLP) and Computer Vision (CV) revolutions. The significant success in NLP and CV inspired exploring the use of Transformers in point cloud processing. However, how do Transformers cope with the irregularity and unordered nature of point clouds? How suitable are Transformers for different 3D representations (e.g., point- or voxel-based)? How competent are Transformers for various 3D processing tasks? As of now, there is still no systematic survey of the research on these issues. For the first time, we provided a comprehensive overview of increasingly popular Transformers for 3D point cloud analysis. We start by introducing the theory of the Transformer architecture and reviewing its applications in 2D/3D fields. Then, we present three different taxonomies (i.e., implementation-, data representation-, and task-based), which can classify current Transformer-based methods from multiple perspectives. Furthermore, we present the results of an investigation of the variants and improvements of the self-attention mechanism in 3D. To demonstrate the superiority of Transformers in point cloud analysis, we present comprehensive comparisons of various Transformer-based methods for classification, segmentation, and object detection. Finally, we suggest three potential research directions, providing benefit references for the development of 3D Transformers. | ['Linlin Xu', 'Kyle Gao', 'Jonathan Li', 'Mingqiang Wei', 'Qian Xie', 'Dening Lu'] | 2022-05-16 | null | null | null | null | ['point-cloud-classification'] | ['computer-vision'] | [-1.34233953e-02 -4.08169687e-01 2.01127559e-01 -2.13743210e-01
-5.15245855e-01 -6.31064355e-01 5.02969444e-01 1.58375889e-01
-1.01428293e-01 1.48068285e-02 -3.98564696e-01 -5.78525662e-01
-3.01671743e-01 -1.04874182e+00 -5.04598439e-01 -4.90937859e-01
-4.59533595e-02 8.05151880e-01 3.76714617e-01 -2.30644539e-01
6.04916632e-01 1.37722230e+00 -2.11873531e+00 2.02419460e-01
9.13473666e-01 1.38298857e+00 5.18313587e-01 2.28906780e-01
-8.74009848e-01 1.48873016e-01 -5.67301929e-01 -2.53686488e-01
3.43627304e-01 4.70409840e-01 -7.45152295e-01 2.24396706e-01
4.97263670e-01 1.02409914e-01 -1.91875592e-01 1.10718179e+00
4.51113552e-01 -5.10145873e-02 8.02547336e-01 -1.41913509e+00
-8.26605976e-01 -8.42429250e-02 -4.92073834e-01 4.30952579e-01
2.60693431e-01 -1.82131643e-03 7.85973072e-01 -1.36532331e+00
3.55436563e-01 1.35144973e+00 7.59373665e-01 7.47060329e-02
-6.70192480e-01 -5.17305255e-01 2.25066409e-01 3.88282090e-01
-1.50175047e+00 -4.26651314e-02 9.99479175e-01 -7.24523127e-01
1.32057738e+00 4.12429810e-01 9.82011080e-01 4.99009758e-01
1.49278387e-01 9.77443337e-01 1.07307053e+00 -4.56063479e-01
3.94127071e-02 8.61722082e-02 2.72742420e-01 5.25632501e-01
1.76145971e-01 -1.20145142e-01 -2.44530424e-01 -2.19505072e-01
1.18903887e+00 5.35829226e-03 -6.80094287e-02 -5.70983648e-01
-1.18498003e+00 8.35481644e-01 5.40823698e-01 4.29375499e-01
-5.06093085e-01 -2.82558441e-01 3.66468132e-01 8.85654911e-02
8.83603930e-01 4.30330426e-01 -3.51625085e-01 1.72279924e-01
-7.70314038e-01 3.22272867e-01 2.99805969e-01 1.40948236e+00
5.76847374e-01 1.28733188e-01 -1.52206384e-02 8.42545211e-01
3.92814636e-01 7.41287529e-01 3.37144285e-01 -7.04090178e-01
3.86198431e-01 8.28019202e-01 -1.45608455e-01 -1.11644614e+00
-4.58233356e-01 -2.98877597e-01 -7.82413602e-01 4.31619674e-01
-4.08736244e-02 6.25246644e-01 -1.00412786e+00 9.65174854e-01
4.64882016e-01 9.90830585e-02 -3.58743399e-01 8.06398273e-01
1.20471179e+00 6.59709752e-01 1.32229971e-02 -7.18312413e-02
1.54790628e+00 -5.72278917e-01 -4.86167431e-01 -3.54860201e-02
1.20642744e-01 -8.83983314e-01 1.17490876e+00 3.95087212e-01
-1.10922122e+00 -8.50938618e-01 -8.02029908e-01 -2.99399346e-01
-8.07236910e-01 -3.61392982e-02 6.80104792e-01 4.89740193e-01
-1.08150923e+00 5.60832143e-01 -8.76663387e-01 -6.53547227e-01
5.66670895e-01 3.51144671e-01 -3.73253822e-02 2.14141354e-01
-7.84005225e-01 1.08493364e+00 5.29379845e-02 1.48960069e-01
-3.37206244e-01 -7.47119308e-01 -5.35324633e-01 -5.81429750e-02
6.54167905e-02 -9.94729459e-01 1.28134799e+00 -4.52231348e-01
-1.26000512e+00 1.26217270e+00 -2.62550652e-01 -2.67844021e-01
2.43476480e-01 -2.10714847e-01 -7.44298100e-02 1.90407813e-01
2.57038295e-01 5.58418810e-01 8.91795635e-01 -1.22194028e+00
-8.35368097e-01 -6.96362793e-01 9.15077925e-02 5.56428671e-01
-3.04336473e-02 2.82061636e-01 -7.04295218e-01 -4.12910044e-01
5.31047165e-01 -6.49465084e-01 -1.81578740e-01 3.50753486e-01
-1.29537866e-01 -7.99385548e-01 9.41556931e-01 -2.23762676e-01
8.37122381e-01 -2.27069068e+00 -7.11449385e-02 1.67613238e-01
5.11852503e-01 4.13743407e-01 4.28193659e-02 3.52150023e-01
-2.84545541e-01 3.95718753e-01 -2.92438298e-01 -3.89124811e-01
3.94367687e-02 2.08819062e-01 -5.87201238e-01 3.63863558e-01
1.96915060e-01 9.31483209e-01 -7.49036610e-01 -6.56711638e-01
8.49761188e-01 5.58079958e-01 -2.47371599e-01 -7.76319392e-03
-2.26868484e-02 2.36089647e-01 -7.89074600e-01 1.03672493e+00
1.00751472e+00 -2.00633705e-01 -5.54807365e-01 -4.84083563e-01
-6.80280983e-01 2.03716710e-01 -8.75990927e-01 1.37120438e+00
-3.29352409e-01 6.31684124e-01 1.38846591e-01 -1.01980627e+00
1.06457198e+00 1.92921370e-01 7.79876709e-01 -5.60515344e-01
6.82886988e-02 2.43034303e-01 -3.86399269e-01 -5.50738871e-01
6.33909583e-01 -3.16766202e-01 1.82776004e-01 2.89733638e-03
-1.54910088e-01 -8.95160615e-01 -4.06491719e-02 -1.72569945e-01
7.32959270e-01 -1.00869164e-02 3.88680041e-01 -1.60250261e-01
5.20301402e-01 4.34875548e-01 1.95560157e-01 7.40860403e-01
-3.18661362e-01 8.55478644e-01 1.18770078e-01 -6.31010532e-01
-8.87646973e-01 -1.21353197e+00 -4.60683763e-01 6.27843738e-01
2.73060560e-01 -2.97863811e-01 -3.29852670e-01 -4.02471662e-01
2.91894436e-01 5.58051169e-01 -2.74289697e-01 3.41489941e-01
-4.17275578e-01 -7.32801020e-01 3.37913483e-01 5.37233353e-01
5.60051978e-01 -1.05075824e+00 -6.37632370e-01 5.11503257e-02
-2.34640703e-01 -1.27757204e+00 1.57168254e-01 1.28503621e-01
-1.41723013e+00 -9.97137308e-01 -5.53810358e-01 -7.74867356e-01
5.02682388e-01 7.59598851e-01 1.36821628e+00 2.82220356e-02
-9.75971743e-02 7.48922229e-01 -3.73197019e-01 -9.48358059e-01
2.20658734e-01 -1.24786995e-01 1.62224993e-01 -3.39878172e-01
6.21225238e-01 -5.86296797e-01 -1.65102541e-01 2.01462418e-01
-6.28533304e-01 -1.34090036e-01 4.97978985e-01 2.80595839e-01
1.00935185e+00 2.66200602e-01 -1.19183980e-01 -5.63236356e-01
7.43498385e-01 -1.12628385e-01 -6.58541203e-01 1.13644198e-01
-2.13808864e-01 -3.81843477e-01 2.71394432e-01 -1.78089693e-01
-7.01543272e-01 6.12299815e-02 -5.59794843e-01 -1.05785429e+00
-4.24673498e-01 4.14637715e-01 -1.15478262e-01 -4.52798158e-01
4.39751834e-01 3.38342249e-01 -7.24289119e-02 -7.00676262e-01
3.35196972e-01 7.28280306e-01 2.19966456e-01 -6.85749829e-01
8.26977730e-01 7.14138687e-01 -7.10539967e-02 -1.40980983e+00
-6.26190126e-01 -7.73662746e-01 -8.18214893e-01 -2.27251768e-01
9.05020714e-01 -6.28018022e-01 -6.93426371e-01 5.57785213e-01
-1.70484197e+00 2.45450214e-01 -5.91119289e-01 2.06741393e-01
-5.89933217e-01 5.08470178e-01 -3.28850210e-01 -9.40259159e-01
-4.09238964e-01 -1.40621364e+00 1.43604326e+00 -6.22464269e-02
1.83776930e-01 -7.51788259e-01 -1.16308317e-01 3.07718039e-01
3.29487413e-01 1.73915908e-01 1.17804313e+00 -3.87894511e-01
-7.96264172e-01 -1.18715659e-01 -5.02072632e-01 2.24512801e-01
-4.44541723e-02 2.20705658e-01 -1.06831479e+00 1.10031858e-01
4.01557535e-01 3.39141279e-01 3.54313254e-01 5.51586986e-01
1.47155523e+00 1.52447820e-01 -5.78116000e-01 6.93714559e-01
1.34842074e+00 2.74235040e-01 5.44888318e-01 2.95880973e-01
6.38633013e-01 5.55112064e-01 5.69535792e-01 1.34415984e-01
4.03167725e-01 6.84910595e-01 6.98201656e-01 -1.51207164e-01
-9.92179736e-02 -2.22958773e-02 -1.67116165e-01 1.09021378e+00
-5.23674548e-01 -1.33290663e-01 -1.27711296e+00 4.31394607e-01
-1.55341864e+00 -7.36542523e-01 -4.63516176e-01 2.09774733e+00
3.47888246e-02 6.64252117e-02 -2.75788102e-02 3.23129117e-01
7.79900908e-01 1.18023790e-02 -4.75070626e-01 -3.47059309e-01
-4.10792492e-02 4.02519375e-01 3.82958859e-01 1.69990897e-01
-1.35783672e+00 1.01259089e+00 6.74879313e+00 8.69454503e-01
-1.24861419e+00 4.61408943e-02 1.12813547e-01 2.01470807e-01
-5.34909740e-02 -3.39509487e-01 -6.94233179e-01 2.63084054e-01
3.70297015e-01 -2.45471939e-01 1.75785273e-01 1.09096360e+00
1.33125424e-01 1.70127273e-01 -9.70988154e-01 1.57432890e+00
2.34218556e-02 -1.49754643e+00 3.89190406e-01 5.93806766e-02
8.85544270e-02 6.39794409e-01 1.04464218e-01 1.86524332e-01
-2.24712282e-01 -7.54496336e-01 9.54568923e-01 3.30295175e-01
7.54056454e-01 -4.39920545e-01 5.92936695e-01 4.44676250e-01
-1.49134290e+00 1.21986851e-01 -6.71796620e-01 -7.12487847e-02
2.08655611e-01 7.72670627e-01 -5.75955689e-01 8.84691358e-01
1.16894305e+00 8.18880081e-01 -4.60225195e-01 1.40150142e+00
5.78645244e-03 1.25152871e-01 -7.53319800e-01 -5.83688244e-02
3.05119246e-01 -3.90587538e-01 9.09009755e-01 1.06950819e+00
5.53421140e-01 3.07721645e-01 4.33173962e-02 9.52217579e-01
2.87658870e-01 8.31004903e-02 -9.93517220e-01 7.17449933e-02
5.56934893e-01 1.07958603e+00 -9.78725970e-01 -2.60554641e-01
-5.23696959e-01 5.12239814e-01 2.22394347e-01 2.67216623e-01
-6.16688848e-01 -3.07996243e-01 9.18250024e-01 3.97411674e-01
2.99833119e-01 -7.06147492e-01 -7.76859283e-01 -9.67791677e-01
4.35334034e-02 -5.29849768e-01 2.35386521e-01 -1.20355725e+00
-1.67175233e+00 8.57880175e-01 4.28216636e-01 -1.53896475e+00
4.09045994e-01 -9.90861595e-01 -5.54427087e-01 7.14366376e-01
-1.68379462e+00 -9.86103356e-01 -2.81318188e-01 5.38400888e-01
6.25957966e-01 -4.29516174e-02 6.18537068e-01 4.90990460e-01
-2.37209454e-01 -1.07131027e-01 -6.25422448e-02 -2.62812022e-02
1.03121415e-01 -1.14322817e+00 6.84348702e-01 4.29793596e-01
2.48205885e-01 5.94688594e-01 3.88103873e-01 -4.97343630e-01
-1.50823009e+00 -9.51353848e-01 9.76808190e-01 -7.80164778e-01
4.71459985e-01 -3.35347980e-01 -1.04338109e+00 5.72084069e-01
-2.92029500e-01 -6.57873228e-02 2.88703740e-01 5.87155260e-02
-5.97929163e-03 1.42406253e-02 -1.30579877e+00 4.25637186e-01
1.31893611e+00 -4.60054934e-01 -9.57005441e-01 4.91786033e-01
6.43345654e-01 -6.71883881e-01 -7.40913033e-01 6.33765221e-01
1.46378130e-01 -1.02045238e+00 1.49877298e+00 -1.63016170e-01
-6.20161369e-02 -4.41300184e-01 -1.48578778e-01 -9.46248293e-01
-6.96933448e-01 -1.42680615e-01 1.86675817e-01 8.82086396e-01
-2.03114580e-02 -8.14789176e-01 8.19655418e-01 3.51519197e-01
-6.69308960e-01 -7.70505726e-01 -1.26254010e+00 -7.45700836e-01
2.25017935e-01 -1.10060978e+00 8.71902943e-01 7.95776844e-01
-3.60288739e-01 2.13798776e-01 3.70504051e-01 3.51823032e-01
6.92515850e-01 4.03853178e-01 5.66894591e-01 -1.72437203e+00
5.13143301e-01 -7.49162138e-01 -7.63161838e-01 -1.35203838e+00
-4.21368554e-02 -1.12944496e+00 -2.47422725e-01 -1.84839988e+00
-2.43418306e-01 -6.95977807e-01 6.16875403e-02 3.12143147e-01
2.86086291e-01 2.47843042e-01 4.58102405e-01 6.24024153e-01
-2.65038818e-01 5.65057397e-01 1.34655201e+00 -2.05762848e-01
-1.10084280e-01 1.71640396e-01 -4.99996215e-01 9.90504205e-01
6.37051523e-01 -3.34383994e-01 -3.28448772e-01 -8.21003199e-01
1.56653225e-01 -2.28394181e-01 4.50172067e-01 -1.05229759e+00
3.42329830e-01 -1.16282687e-01 2.13388875e-01 -1.42921519e+00
6.13697112e-01 -1.12670743e+00 -6.47826791e-02 2.30585281e-02
4.47656780e-01 4.20985401e-01 4.43708837e-01 2.97503322e-01
-2.98526466e-01 -1.04511827e-01 7.77573943e-01 -4.87869889e-01
-9.06983674e-01 5.96336544e-01 -4.19097155e-01 -1.81426823e-01
9.24593627e-01 -7.73914695e-01 3.70182656e-02 6.19717985e-02
-5.03951073e-01 2.63375252e-01 3.53737652e-01 5.43018699e-01
1.05156946e+00 -1.22396684e+00 -5.52114785e-01 4.70685929e-01
1.33361802e-01 4.37910378e-01 1.50820792e-01 7.10434258e-01
-6.91897154e-01 7.52094507e-01 -2.28554383e-01 -1.15154588e+00
-1.18049276e+00 7.37600982e-01 3.05403918e-01 -7.02733994e-02
-8.21105957e-01 7.54502654e-01 3.79647106e-01 -6.94106102e-01
1.46983773e-01 -7.51733959e-01 -3.05350274e-01 -1.01877283e-02
8.58438760e-02 3.54123175e-01 5.04318774e-01 -7.30140865e-01
-7.27407336e-01 1.26854300e+00 2.68788904e-01 1.85599923e-01
1.32510364e+00 8.21413100e-02 -2.90424347e-01 5.31948864e-01
7.69296229e-01 -3.50225538e-01 -6.15308285e-01 -2.04025388e-01
-8.07854608e-02 -5.57563484e-01 1.88212454e-01 -4.11127299e-01
-9.87832248e-01 1.27915156e+00 6.02981627e-01 5.87808073e-01
1.09104168e+00 3.02568585e-01 5.21413147e-01 2.10265145e-01
6.64447725e-01 -7.05052733e-01 -4.57503825e-01 8.77258778e-01
1.24456251e+00 -1.00930679e+00 1.01025581e-01 -8.24631453e-01
-3.64355773e-01 1.11306369e+00 4.13233399e-01 -1.10631607e-01
9.60217357e-01 2.02967107e-01 -6.34971038e-02 -7.46940255e-01
-2.99145520e-01 -4.00806189e-01 4.83456880e-01 9.99257684e-01
3.40158820e-01 -6.14418350e-02 7.24628344e-02 2.68834412e-01
-4.16863412e-01 4.01729159e-02 -3.02788112e-02 9.25271094e-01
-5.77159643e-01 -8.36967826e-01 -8.42484117e-01 5.21111727e-01
1.93477795e-02 -6.20314712e-03 -4.31754351e-01 9.68156338e-01
2.79516488e-01 7.27767646e-01 3.73722970e-01 -2.30277598e-01
7.84894407e-01 2.58827247e-02 6.34999633e-01 -7.17535973e-01
-6.53728664e-01 -6.90564811e-02 -3.22358638e-01 -5.35454392e-01
-5.06283700e-01 -7.50522137e-01 -1.02358496e+00 -3.03785503e-01
-3.68655652e-01 1.81939721e-01 8.79382849e-01 7.60603249e-01
4.82915461e-01 3.71786922e-01 3.04003149e-01 -1.31781971e+00
-2.34619901e-01 -6.62607908e-01 -4.45028722e-01 5.09699807e-02
1.37575552e-01 -9.77752328e-01 -3.51302236e-01 -2.71753401e-01] | [7.976195812225342, -3.4697670936584473] |
d0456c99-ca51-4034-bdb5-ed23dea44dbb | sleepeegnet-automated-sleep-stage-scoring | 1903.02108 | null | http://arxiv.org/abs/1903.02108v1 | http://arxiv.org/pdf/1903.02108v1.pdf | SleepEEGNet: Automated Sleep Stage Scoring with Sequence to Sequence Deep Learning Approach | Electroencephalogram (EEG) is a common base signal used to monitor brain
activity and diagnose sleep disorders. Manual sleep stage scoring is a
time-consuming task for sleep experts and is limited by inter-rater
reliability. In this paper, we propose an automatic sleep stage annotation
method called SleepEEGNet using a single-channel EEG signal. The SleepEEGNet is
composed of deep convolutional neural networks (CNNs) to extract time-invariant
features, frequency information, and a sequence to sequence model to capture
the complex and long short-term context dependencies between sleep epochs and
scores. In addition, to reduce the effect of the class imbalance problem
presented in the available sleep datasets, we applied novel loss functions to
have an equal misclassified error for each sleep stage while training the
network. We evaluated the proposed method on different single-EEG channels
(i.e., Fpz-Cz and Pz-Oz EEG channels) from the Physionet Sleep-EDF datasets
published in 2013 and 2018. The evaluation results demonstrate that the
proposed method achieved the best annotation performance compared to current
literature, with an overall accuracy of 84.26%, a macro F1-score of 79.66% and
Cohen's Kappa coefficient = 0.79. Our developed model is ready to test with
more sleep EEG signals and aid the sleep specialists to arrive at an accurate
diagnosis. The source code is available at
https://github.com/SajadMo/SleepEEGNet. | ['U. Rajendra Acharya', 'Fatemeh Afghah', 'Sajad Mousavi'] | 2019-03-05 | null | null | null | null | ['sleep-stage-detection'] | ['medical'] | [ 6.20366223e-02 -3.26586366e-01 1.69968963e-01 -5.77576637e-01
-5.93066454e-01 -2.32247025e-01 -1.48577064e-01 7.95680732e-02
-7.08355904e-01 9.40486014e-01 -6.06110580e-02 2.85166744e-02
-1.18348628e-01 -1.85991243e-01 -2.41336808e-01 -6.73572183e-01
-2.46506438e-01 7.90986046e-02 5.77310063e-02 7.70376697e-02
2.75577158e-01 1.98574197e-02 -1.56425440e+00 1.84099630e-01
1.13646245e+00 1.26332808e+00 4.42162782e-01 4.97749239e-01
2.32928187e-01 2.32369825e-01 -9.32191253e-01 -1.70049474e-01
-9.38114673e-02 -5.69056988e-01 -4.98899460e-01 -3.04717720e-01
-1.08707741e-01 1.42349929e-01 -8.51681381e-02 1.15916920e+00
9.31954861e-01 1.18371490e-02 2.45794699e-01 -1.20570719e+00
-2.90102601e-01 2.87616491e-01 -1.94617957e-01 1.00797713e+00
2.25440279e-01 2.04124019e-01 6.48468196e-01 -6.47236824e-01
-8.92163664e-02 1.63547084e-01 7.97117710e-01 6.59151971e-01
-9.69275951e-01 -1.30569839e+00 -3.67040962e-01 8.25732112e-01
-1.71325421e+00 -5.49339294e-01 5.45586348e-01 -3.87662590e-01
1.09274399e+00 1.70761347e-01 1.22517264e+00 1.17735732e+00
8.32304001e-01 2.01500013e-01 1.26391447e+00 4.73674498e-02
3.55661958e-01 -1.38364108e-02 3.42305690e-01 4.84420031e-01
7.20052570e-02 -1.57645494e-01 -8.62655163e-01 2.13734448e-01
3.17669064e-01 2.50524700e-01 -4.49472815e-01 6.05708659e-01
-8.58128726e-01 4.16900665e-01 3.75950873e-01 3.73520344e-01
-5.28108776e-01 -2.26497769e-01 4.97338921e-01 2.76031315e-01
6.44774795e-01 4.16558444e-01 -5.32577574e-01 -7.21379757e-01
-1.44411492e+00 -1.71303958e-01 5.79806507e-01 6.30687773e-01
4.85283285e-01 -6.39552027e-02 -2.65967041e-01 9.34851706e-01
7.98582956e-02 3.34935993e-01 9.94146466e-01 -5.22126555e-01
1.73767075e-01 6.48685694e-01 -1.13012232e-01 -4.85223651e-01
-9.93169010e-01 -7.04402149e-01 -9.65223074e-01 -2.41062373e-01
-5.90902381e-02 -5.67107014e-02 -7.13499129e-01 1.59111607e+00
-3.15734267e-01 3.75225008e-01 -4.32381600e-01 9.57626879e-01
9.14824545e-01 3.08864117e-01 4.20856588e-02 -3.49774212e-01
1.67789316e+00 -6.79358661e-01 -9.36831951e-01 -4.81025815e-01
2.01410830e-01 -5.63818276e-01 1.21165240e+00 6.53362155e-01
-1.10930085e+00 -5.18403411e-01 -1.22064936e+00 9.77818072e-02
-1.90094650e-01 4.52890843e-01 2.30602324e-01 4.99753088e-01
-1.23663831e+00 4.47415859e-01 -1.33915019e+00 -3.07368875e-01
7.25474060e-01 8.55479836e-01 -3.27867955e-01 3.30042928e-01
-1.20358491e+00 9.32538211e-01 2.33653247e-01 2.93994844e-01
-8.83830547e-01 -6.50852919e-01 -4.11961228e-01 1.97980806e-01
-1.07493713e-01 -6.58256114e-01 1.13082945e+00 -8.27560186e-01
-1.30142379e+00 8.51392508e-01 -4.88455594e-01 -4.32408899e-01
-1.00728519e-01 -2.72850186e-01 -7.57874966e-01 9.50203277e-03
2.09844530e-01 2.66369313e-01 5.29760182e-01 -2.63803571e-01
-5.25556624e-01 -6.58997178e-01 -3.24303061e-01 2.58737635e-02
-2.98632175e-01 2.05153421e-01 -2.72167206e-01 -4.43041414e-01
-1.06649332e-01 -8.95120859e-01 2.98202962e-01 -3.85949075e-01
-2.72674412e-01 -3.05213213e-01 1.03789330e-01 -8.41608703e-01
1.55456567e+00 -2.24824762e+00 -2.82912832e-02 -6.56612888e-02
3.84297371e-01 2.45489240e-01 2.76075333e-01 1.00059181e-01
-3.12009692e-01 -1.57787532e-01 -4.10899699e-01 -5.88567674e-01
-2.43150517e-01 2.90538259e-02 3.32185596e-01 6.65387869e-01
6.51643947e-02 6.49649918e-01 -7.17077255e-01 -8.21956843e-02
1.03470966e-01 5.38158774e-01 -2.55297512e-01 4.02069926e-01
4.76679295e-01 5.34267724e-01 1.29047334e-01 5.24202108e-01
3.88202071e-01 -2.78581262e-01 -2.84253865e-01 -1.45062417e-01
-2.88524151e-01 8.40503573e-01 -6.25323117e-01 1.91678667e+00
-6.03990912e-01 9.09425914e-01 -2.00654194e-01 -7.54120469e-01
8.53357494e-01 4.66267347e-01 4.22032595e-01 -8.51801097e-01
4.66421485e-01 2.22511068e-01 3.38943362e-01 -6.25753284e-01
1.26819881e-02 -2.71029443e-01 2.09526941e-01 2.77243435e-01
2.97175080e-01 1.48276836e-01 2.44290903e-01 -2.67176867e-01
1.39468467e+00 -4.25527036e-01 3.72729361e-01 -4.49976593e-01
3.95646185e-01 -5.14291286e-01 8.57393086e-01 3.96678299e-01
-3.31334352e-01 6.26666367e-01 1.95541933e-01 -2.76270092e-01
-6.18609309e-01 -9.71403122e-01 -4.26070273e-01 6.08937204e-01
-1.00330651e-01 -8.05185735e-01 -7.81798184e-01 -2.43366450e-01
-5.17489076e-01 6.54551387e-01 -6.40657723e-01 -5.50871611e-01
-1.03983544e-01 -9.82301593e-01 6.11858130e-01 5.25951624e-01
5.88518023e-01 -1.23376560e+00 -8.61437678e-01 1.81746408e-01
-3.58753204e-01 -8.59671712e-01 -5.01944244e-01 6.30607665e-01
-6.43773675e-01 -1.11263824e+00 -5.64200282e-01 -6.13791585e-01
4.26060647e-01 -1.95808262e-01 9.13819373e-01 -5.72134331e-02
-4.56582397e-01 -1.37425601e-01 -3.93104821e-01 -6.83719456e-01
2.21366405e-01 2.96801418e-01 2.41988197e-01 -6.24103844e-02
8.79623353e-01 -1.04539251e+00 -1.14982474e+00 1.76804677e-01
-5.70189238e-01 -1.44281732e-02 5.85773408e-01 8.07092547e-01
6.27944410e-01 1.22223251e-01 6.92639828e-01 -2.50316918e-01
7.78501213e-01 -5.83694518e-01 -5.60581505e-01 -1.48670718e-01
-1.01153803e+00 -3.93130928e-01 6.66844249e-01 -1.96160749e-01
-5.69552302e-01 -1.64893895e-01 -5.77192366e-01 -1.94615528e-01
-3.13053638e-01 3.17067653e-01 2.31829043e-02 1.47068262e-01
4.73808020e-01 5.69291532e-01 -1.68584466e-01 -4.13609445e-01
-5.84165990e-01 9.75662291e-01 5.52399576e-01 1.18817851e-01
3.96455601e-02 1.62736088e-01 -3.16168904e-01 -8.40455592e-01
-9.39937174e-01 -6.73845947e-01 -4.22083050e-01 -1.36671737e-01
1.07667494e+00 -1.12827849e+00 -8.60729337e-01 5.62863410e-01
-9.38028574e-01 -4.72221136e-01 1.00666480e-02 7.44333267e-01
-2.93246686e-01 1.99155838e-04 -4.99629259e-01 -6.69018507e-01
-1.01899981e+00 -1.20354474e+00 8.76794279e-01 4.11463439e-01
-5.99547744e-01 -6.69838011e-01 1.15038842e-01 1.87078878e-01
4.71635789e-01 -2.03246459e-01 5.03177106e-01 -7.07338870e-01
5.11062369e-02 -9.31349322e-02 1.11047320e-01 5.24021745e-01
3.34569097e-01 -4.18110877e-01 -1.13512313e+00 -2.42611572e-01
3.82119805e-01 -3.29916924e-02 5.50678134e-01 5.47180116e-01
1.40344667e+00 -1.21877529e-01 -2.65074726e-02 8.69775772e-01
1.26003408e+00 5.58420181e-01 7.78761268e-01 3.91710877e-01
2.60480523e-01 2.57352609e-02 1.72150791e-01 5.86407959e-01
4.57221717e-01 5.27123272e-01 1.83400467e-01 1.98054954e-01
-4.19315845e-02 3.91123235e-01 3.74408394e-01 1.13689733e+00
-7.74740428e-02 -1.23463795e-01 -9.46218967e-01 5.88168323e-01
-1.42259741e+00 -7.08944142e-01 -1.53298229e-01 2.18410254e+00
8.58366191e-01 3.14829677e-01 7.63722956e-02 3.23971987e-01
4.32248145e-01 -3.40547085e-01 -5.86964488e-01 -3.92368644e-01
7.02298358e-02 8.33870471e-01 2.25683138e-01 3.07534076e-03
-7.20898092e-01 4.98813957e-01 5.40451527e+00 6.01935685e-01
-1.37300563e+00 6.17010236e-01 1.97112754e-01 -8.31719041e-01
4.41089302e-01 -4.35118169e-01 -6.70586765e-01 1.24351311e+00
1.51957119e+00 -1.93112090e-01 8.41661811e-01 4.20163214e-01
7.74899364e-01 -3.03854674e-01 -8.79941404e-01 1.44594705e+00
1.05601788e-01 -1.00648141e+00 -8.40164721e-01 -1.17022172e-01
2.95530349e-01 6.06683791e-01 -1.91495925e-01 1.61851481e-01
-4.86961931e-01 -1.04626310e+00 6.94799364e-01 7.69491315e-01
1.05036628e+00 -7.37171531e-01 1.16152728e+00 2.15268835e-01
-1.20431578e+00 -1.77562147e-01 -1.72893256e-01 -2.39952773e-01
2.77561657e-02 7.10171402e-01 -5.82345366e-01 2.57198423e-01
1.18660057e+00 9.93325233e-01 -7.52703786e-01 1.57041931e+00
-5.23014545e-01 1.00404751e+00 -1.64915279e-01 -4.66379859e-02
-2.71608353e-01 -2.83551831e-02 3.16146374e-01 1.07572985e+00
6.11463368e-01 1.65149048e-01 -4.17471260e-01 9.33870912e-01
-1.35878399e-02 -2.86759228e-01 -1.23731859e-01 1.40136972e-01
4.92590070e-01 1.40569425e+00 -8.33161235e-01 1.81878246e-02
-3.37010205e-01 1.06300092e+00 2.16088682e-01 1.63334295e-01
-9.25566494e-01 -7.15030313e-01 5.93255520e-01 -3.97532620e-02
-1.07221745e-01 -8.86218715e-03 -5.57710946e-01 -1.11925101e+00
1.99660778e-01 -6.61500335e-01 2.07518727e-01 -9.93296802e-01
-1.23400128e+00 9.67644632e-01 -1.62781537e-01 -1.11774492e+00
1.76085547e-01 -3.30492377e-01 -9.41137135e-01 9.61768031e-01
-1.26585305e+00 -5.13796031e-01 -5.96394539e-01 6.95703208e-01
6.84014738e-01 -4.27157395e-02 9.21253383e-01 7.68977582e-01
-9.45188642e-01 7.02320576e-01 -5.25702126e-02 -1.38573632e-01
7.90155351e-01 -1.21413851e+00 4.54885885e-02 7.23246038e-01
-2.03144357e-01 8.10711026e-01 5.04832327e-01 -3.67787361e-01
-9.05441463e-01 -1.07037473e+00 9.73080873e-01 -2.40548030e-01
5.29189587e-01 -5.47674417e-01 -7.81743288e-01 4.58191067e-01
2.60357231e-01 -3.19445550e-01 1.10289073e+00 3.70936692e-02
3.55881840e-01 -5.59834719e-01 -9.90170121e-01 1.87947437e-01
8.77511621e-01 -6.39165103e-01 -6.89599216e-01 1.47965387e-01
2.39961296e-01 -2.77271122e-01 -9.83231723e-01 1.77009255e-01
6.64164245e-01 -1.09843183e+00 3.63049001e-01 1.19105175e-01
1.62539795e-01 -1.66884094e-01 2.13373214e-01 -1.49764383e+00
-3.00568104e-01 -4.77896214e-01 5.96117377e-02 9.82972980e-01
3.54118109e-01 -6.94434464e-01 4.93236840e-01 4.83558387e-01
-7.88143754e-01 -1.06310415e+00 -1.11349440e+00 -6.32362306e-01
-4.99674022e-01 -6.39708519e-01 6.01238787e-01 3.41825515e-01
2.01454580e-01 5.26720226e-01 -1.13443278e-01 1.78554982e-01
2.01521739e-01 -2.42617875e-01 6.72386736e-02 -1.15646720e+00
1.02112442e-01 -3.91477108e-01 -4.92357016e-01 -3.89813870e-01
1.73267260e-01 -1.04769933e+00 2.15044826e-01 -1.56723177e+00
3.44414413e-01 -8.88618454e-02 -8.81931007e-01 7.24310458e-01
-1.12953767e-01 5.52991331e-01 -2.32183024e-01 3.93031389e-02
-6.31882370e-01 6.01787806e-01 6.94157064e-01 3.42637300e-02
-3.13089371e-01 2.58907169e-01 -6.19686604e-01 6.28742874e-01
1.29960823e+00 -8.01694453e-01 -5.04824579e-01 -2.73359835e-01
2.19505385e-01 -1.49650127e-01 2.74217755e-01 -1.51599717e+00
3.48576069e-01 4.15684968e-01 5.67247093e-01 -4.28008825e-01
4.63136971e-01 -5.85985839e-01 3.23098570e-01 4.84258175e-01
-1.56283692e-01 3.18844527e-01 3.83465648e-01 1.56787172e-01
-1.26499161e-01 -3.89474258e-02 6.31123960e-01 1.70240663e-02
-4.47795093e-01 4.18268800e-01 -6.50904715e-01 -5.68812899e-03
6.12525642e-01 -2.37372413e-01 -3.43850523e-01 -3.42008881e-02
-7.85355449e-01 2.42763937e-01 2.27035969e-01 3.47009808e-01
7.63682127e-01 -1.00816274e+00 -3.73681992e-01 6.62870288e-01
3.39836538e-01 -2.17180714e-01 5.53462982e-01 1.48165929e+00
-2.32203320e-01 4.47512776e-01 -5.80902338e-01 -5.27280092e-01
-1.28663635e+00 2.23297011e-02 3.57148051e-01 6.25853613e-02
-6.49064064e-01 1.06629443e+00 -1.22738451e-01 2.06770346e-01
2.99020112e-01 -5.58358848e-01 -3.79301548e-01 8.52619577e-03
6.18948460e-01 3.34605753e-01 7.72037685e-01 -4.15456921e-01
-7.62645781e-01 2.57272065e-01 1.34959653e-01 1.77741468e-01
1.47746038e+00 -1.63239107e-01 -2.98141271e-01 6.73522949e-01
1.09770298e+00 -3.23480397e-01 -9.40459728e-01 4.77941185e-01
-1.53783828e-01 -9.76232216e-02 1.97652012e-01 -1.07566702e+00
-1.06120169e+00 1.07743704e+00 1.30637300e+00 4.89825085e-02
1.60030401e+00 -1.61367208e-01 1.05901861e+00 9.06947330e-02
3.30712318e-01 -9.30555999e-01 -2.35462025e-01 2.91330665e-01
7.54746675e-01 -1.02580261e+00 -7.49320611e-02 3.06956470e-01
-4.46157515e-01 9.89632487e-01 6.22805655e-01 -9.44616646e-02
8.00304234e-01 6.83567896e-02 -7.88378939e-02 -4.80507970e-01
-7.06407666e-01 -3.74201424e-02 4.73014593e-01 3.35878223e-01
5.57903111e-01 1.98061645e-01 -6.32569611e-01 1.39278495e+00
-6.28040791e-01 4.38623756e-01 3.41117114e-01 6.46255314e-01
-2.18135774e-01 -7.73716688e-01 5.28252870e-02 1.02739167e+00
-8.31958950e-01 -3.21176827e-01 -2.25845054e-02 4.23381537e-01
5.05889535e-01 1.24120700e+00 2.19509900e-01 -8.38231742e-01
2.56957650e-01 2.90493041e-01 3.39844912e-01 -8.70260000e-01
-7.61929631e-01 -2.06225049e-02 -8.84449556e-02 -6.64368510e-01
-4.48272705e-01 -4.10082370e-01 -1.26919520e+00 -1.65451895e-02
-2.77390927e-01 3.59659135e-01 7.82653213e-01 1.07171524e+00
6.07504129e-01 9.39541340e-01 3.71816814e-01 -6.96603835e-01
5.36858961e-02 -1.32605553e+00 -7.81441271e-01 2.85218172e-02
5.06976843e-01 -8.13074172e-01 -3.88551980e-01 4.45159115e-02] | [13.503805160522461, 3.5023579597473145] |
3111c885-361e-4bc5-9662-b48bdcb92bbd | unsupervised-cnn-based-co-saliency-detection | null | null | http://openaccess.thecvf.com/content_ECCV_2018/html/Kuang-Jui_Hsu_Unsupervised_CNN-based_co-saliency_ECCV_2018_paper.html | http://openaccess.thecvf.com/content_ECCV_2018/papers/Kuang-Jui_Hsu_Unsupervised_CNN-based_co-saliency_ECCV_2018_paper.pdf | Unsupervised CNN-based Co-Saliency Detection with Graphical Optimization | In this paper, we address co-saliency detection in a set of images jointly covering objects of a specific class by an unsupervised convolutional neural network (CNN). Our method does not require any additional training data in the form of object masks. We decompose co-saliency detection into two sub-tasks, single-image saliency detection and cross-image co-occurrence region discovery corresponding to two novel unsupervised losses, the single-image saliency (SIS) loss and the co-occurrence (COOC) loss. The two losses are modeled on a graphical model where the former and the latter act as the unary and pairwise terms, respectively. These two tasks can be jointly optimized for generating co-saliency maps of high quality. Furthermore, the quality of the generated co-saliency maps can be enhanced via two extensions: map sharpening by self-paced learning and boundary preserving by fully connected conditional random fields. Experiments show that our method achieves superior results, even outperforming many supervised methods. | ['Yung-Yu Chuang', 'Yen-Yu Lin', 'Chung-Chi Tsai', 'Kuang-Jui Hsu', 'Xiaoning Qian'] | 2018-09-01 | null | null | null | eccv-2018-9 | ['co-saliency-detection'] | ['computer-vision'] | [ 8.17348301e-01 5.03036976e-01 -2.05233052e-01 -3.78167152e-01
-7.35726953e-01 -2.01825172e-01 6.15722477e-01 1.65619925e-01
-2.31672212e-01 7.67731965e-01 -1.86368264e-02 2.19183356e-01
-4.22595738e-04 -5.67994952e-01 -1.08714712e+00 -6.11937106e-01
1.62853897e-01 -3.92829515e-02 8.33817661e-01 -9.69359502e-02
3.42127979e-01 2.85140753e-01 -1.79857600e+00 1.28200039e-01
1.37305331e+00 1.08573687e+00 6.50341690e-01 2.36759648e-01
-2.00633816e-02 6.53088570e-01 -2.56787360e-01 -1.69609755e-01
1.82029277e-01 -4.15642887e-01 -6.93809569e-01 4.03923064e-01
3.66856456e-01 2.05268919e-01 5.30394949e-02 1.36782980e+00
1.76903918e-01 2.02665180e-02 4.81733173e-01 -1.39285660e+00
-5.09485006e-01 4.87750679e-01 -8.59880388e-01 6.13588914e-02
-7.30910450e-02 3.76635492e-02 1.00038159e+00 -9.19574380e-01
4.33726639e-01 1.23285198e+00 4.43978310e-01 2.12560147e-01
-1.34842730e+00 -6.15482748e-01 2.30897114e-01 2.39015464e-02
-1.34252441e+00 -2.38044947e-01 1.09732151e+00 -3.57260317e-01
4.22764599e-01 1.37079522e-01 4.06090617e-01 4.09841835e-01
6.71362877e-03 1.06951714e+00 1.13020682e+00 -5.37567616e-01
2.70132154e-01 3.56802970e-01 -2.46066853e-01 8.09137404e-01
2.71614075e-01 1.09105572e-01 -6.13885224e-01 1.02685921e-01
1.07921350e+00 2.29230300e-01 -3.40409487e-01 -7.45610058e-01
-1.33381021e+00 7.42269695e-01 9.43222106e-01 2.64100134e-01
-2.81469703e-01 3.94197889e-02 6.42984931e-04 -1.63303703e-01
6.07283294e-01 5.95645666e-01 -2.78448135e-01 6.35289192e-01
-1.24372721e+00 2.64394641e-01 1.15921855e-01 1.13591444e+00
1.23572135e+00 -1.03240505e-01 -2.84276068e-01 7.23018885e-01
1.32311448e-01 5.36927402e-01 3.41002256e-01 -5.80230772e-01
3.19480985e-01 6.37589633e-01 2.35191464e-01 -1.07833290e+00
-1.68332890e-01 -8.12079668e-01 -7.05772698e-01 2.65041262e-01
1.27775714e-01 -1.17223077e-01 -1.16599417e+00 1.81603515e+00
3.21056336e-01 3.21090549e-01 -2.07923695e-01 9.01262343e-01
6.68189108e-01 3.70298654e-01 4.24384214e-02 -4.77619655e-02
1.15430391e+00 -1.50811613e+00 -8.38144362e-01 -4.48444158e-01
1.07982904e-01 -7.81795800e-01 8.30117404e-01 -1.24841601e-01
-1.46830583e+00 -5.41663587e-01 -1.18730772e+00 -1.07732363e-01
-4.55787152e-01 2.38233328e-01 6.51634395e-01 7.33846277e-02
-1.08535397e+00 4.41398442e-01 -6.20241284e-01 1.98208526e-01
8.67440701e-01 3.97700012e-01 -1.80012453e-02 1.52936071e-01
-1.06391990e+00 7.45959997e-01 4.41162795e-01 -2.15411246e-01
-9.36846793e-01 -7.17881560e-01 -1.14769197e+00 2.13726223e-01
5.09555101e-01 -5.18552899e-01 9.59424376e-01 -1.41802454e+00
-1.13110352e+00 1.10778928e+00 -2.92491436e-01 -5.78061938e-01
4.00491923e-01 -1.91228375e-01 -3.16446513e-01 3.04738253e-01
6.89233601e-01 1.06718695e+00 1.15150833e+00 -1.49530506e+00
-9.56184089e-01 -1.19691372e-01 -2.16907412e-01 4.15858656e-01
-1.52432453e-02 7.24468902e-02 -5.65509558e-01 -9.98881102e-01
1.61299378e-01 -6.63611650e-01 -3.24785501e-01 1.52689144e-01
-7.66657233e-01 -2.44193915e-02 9.21841800e-01 -5.29787898e-01
9.09298241e-01 -2.24797058e+00 1.73467621e-01 1.01197883e-01
3.46702963e-01 1.58904001e-01 -6.31601512e-02 -2.76089668e-01
-1.90115660e-01 6.34416342e-02 -8.23078930e-01 -5.37978232e-01
-3.72110784e-01 -2.38231927e-01 -3.87894899e-01 3.61328900e-01
8.30602586e-01 1.15952134e+00 -1.16236603e+00 -7.68620491e-01
1.93386018e-01 3.20021123e-01 -2.91489840e-01 1.19469732e-01
-2.35934079e-01 4.35421765e-01 -4.39716846e-01 5.03996193e-01
8.59052241e-01 -5.21951556e-01 -2.74671137e-01 -4.29753698e-02
-2.27545306e-01 1.30463496e-01 -1.02272391e+00 1.75547910e+00
-2.91450471e-01 4.10882145e-01 6.40150979e-02 -8.89989793e-01
9.23314512e-01 -1.18904456e-01 2.91992217e-01 -5.75472713e-01
-1.31362468e-01 2.97394633e-01 -3.72061491e-01 -3.10379565e-02
5.29326379e-01 -6.14335537e-02 7.51248077e-02 2.53154337e-01
1.79776862e-01 -1.23374879e-01 1.15960382e-01 2.55141914e-01
4.98026937e-01 -1.25311211e-01 5.28798848e-02 -6.20279193e-01
4.80676383e-01 -7.47888759e-02 6.29306912e-01 6.38594747e-01
-7.87877664e-02 1.01899362e+00 4.23251659e-01 7.03946799e-02
-1.01086915e+00 -1.15877163e+00 -2.06083909e-01 7.44415045e-01
9.49729979e-01 1.23823509e-01 -7.85071135e-01 -8.88807774e-01
-6.62251413e-02 4.82097596e-01 -7.23272264e-01 -3.51174504e-01
-4.96057779e-01 -4.69494164e-01 -3.20777930e-02 5.27946115e-01
8.09242427e-01 -1.30979669e+00 -7.58259475e-01 1.26259457e-02
-2.64569819e-01 -1.10930204e+00 -9.67902541e-01 2.67482519e-01
-8.53253722e-01 -9.23632264e-01 -1.25335121e+00 -1.38452828e+00
9.92534101e-01 6.63740277e-01 9.97586668e-01 1.05719425e-01
-2.67808527e-01 -1.50477543e-01 -1.34561241e-01 -5.22290349e-01
1.11410335e-01 3.97709087e-02 -2.21578747e-01 5.54485202e-01
-1.01014465e-01 -5.85559607e-01 -5.60872674e-01 4.06798422e-01
-1.00353301e+00 5.08933306e-01 8.83007288e-01 9.40162241e-01
9.14143682e-01 1.43926576e-01 5.73230386e-01 -7.39942908e-01
1.27922043e-01 -3.60851437e-01 -6.56386852e-01 2.40592346e-01
-5.54440320e-01 2.10120708e-01 1.13693677e-01 -2.25802749e-01
-1.14688456e+00 3.16264778e-01 4.16707128e-01 -5.02187610e-01
-4.59124595e-02 2.94667453e-01 -3.47877890e-01 -1.92920551e-01
2.66993403e-01 4.88792926e-01 -4.89107892e-02 -4.40215766e-01
3.16726267e-01 1.89514786e-01 8.17369461e-01 -2.80590337e-02
9.21671987e-01 6.82854712e-01 -1.09599389e-01 -6.74874783e-01
-1.14987361e+00 -6.19760692e-01 -6.25433445e-01 -5.96080460e-02
9.71085608e-01 -1.06540465e+00 -7.39155104e-03 6.47735596e-01
-1.02262688e+00 -2.13257954e-01 -4.79821831e-01 9.74762440e-02
-6.24530494e-01 2.78071046e-01 -7.36870393e-02 -7.25701869e-01
-2.40205482e-01 -1.17876768e+00 1.28357196e+00 6.01191401e-01
3.52588296e-01 -8.52473736e-01 -2.88962632e-01 7.22770244e-02
4.26580995e-01 4.31643844e-01 6.35359585e-01 -3.54570925e-01
-9.62290168e-01 3.00587099e-02 -6.72087312e-01 3.62665445e-01
1.83945715e-01 -3.81070495e-01 -9.77303743e-01 -1.75872684e-01
-1.94801658e-01 -2.81699598e-01 1.19796014e+00 7.11001575e-01
1.26687241e+00 -1.87108383e-01 -5.31557918e-01 6.28799379e-01
1.45142245e+00 -2.65692715e-02 7.16174841e-01 1.29733652e-01
7.08379567e-01 6.39559448e-01 7.08011270e-01 9.35531333e-02
2.27633744e-01 8.39201987e-01 6.68942273e-01 -7.11297274e-01
-4.02458817e-01 -5.58867574e-01 6.39375448e-02 2.45509431e-01
1.95587158e-01 1.81800395e-01 -6.24281943e-01 1.12067258e+00
-1.98474598e+00 -7.64109731e-01 -1.40262932e-01 2.27464724e+00
8.87250721e-01 2.77765304e-01 2.48092800e-01 -1.12446409e-03
1.30749774e+00 2.07526147e-01 -6.69456422e-01 2.36400127e-01
-5.78264296e-01 1.68777049e-01 6.15418375e-01 5.84353030e-01
-1.46968889e+00 1.05634928e+00 5.79192448e+00 9.98614132e-01
-1.07878745e+00 2.80250102e-01 9.21591043e-01 1.88592330e-01
-4.95954305e-01 2.03370482e-01 -6.14261866e-01 7.28101254e-01
1.56497151e-01 -3.44472677e-02 3.46383825e-02 8.84711266e-01
-1.73710316e-01 -4.82163548e-01 -5.70845008e-01 8.61510694e-01
1.71645246e-02 -1.57652092e+00 -7.52297863e-02 -1.57119989e-01
1.27153790e+00 -8.88591632e-02 3.91107619e-01 -1.47317454e-01
2.53506750e-01 -9.00720954e-01 1.12057078e+00 5.56554437e-01
1.01174390e+00 -7.52438486e-01 5.58452547e-01 1.53999880e-01
-1.20583653e+00 3.90324630e-02 -1.97059497e-01 2.90330738e-01
2.90218800e-01 9.75172520e-01 -3.31627071e-01 4.46960360e-01
8.32861364e-01 9.10634696e-01 -7.10684538e-01 1.43305826e+00
-3.05898041e-01 2.94236124e-01 -2.44223148e-01 7.47521147e-02
2.59804219e-01 4.49411348e-02 7.46187449e-01 1.00986433e+00
-1.78795606e-01 -3.33164006e-01 1.84828758e-01 1.36551726e+00
-1.37269031e-02 -1.75312907e-01 -2.72838622e-01 3.67819667e-01
3.27891558e-01 1.19184375e+00 -9.84972000e-01 -4.18276757e-01
-5.68115525e-02 1.21480525e+00 2.73356557e-01 4.56840426e-01
-7.83493757e-01 -8.29482496e-01 3.90460104e-01 2.40657926e-01
6.12773240e-01 2.36472055e-01 -7.39946842e-01 -1.16609669e+00
2.36568853e-01 -2.32876077e-01 -2.42925584e-02 -7.42578566e-01
-1.05988395e+00 4.09592122e-01 1.07684098e-01 -1.19082391e+00
1.88383356e-01 -1.93227232e-02 -9.29923832e-01 1.09093046e+00
-2.15769482e+00 -1.21132290e+00 -2.86958814e-01 5.64795732e-01
5.12365997e-01 3.76591645e-02 1.01210877e-01 1.70088559e-01
-3.28074723e-01 5.79822004e-01 -2.75787804e-02 -1.26324177e-01
5.36240399e-01 -1.33462620e+00 2.87475675e-01 1.04407644e+00
-1.71215773e-01 1.62557900e-01 5.85325897e-01 -6.91197813e-01
-2.89018273e-01 -1.40793419e+00 1.14419675e+00 -1.20124683e-01
2.49180958e-01 -4.39824164e-01 -1.03417861e+00 3.23164254e-01
8.36477652e-02 1.93846852e-01 -1.35182112e-01 -3.49776655e-01
-1.26183197e-01 1.20346226e-01 -1.18977010e+00 5.12162268e-01
8.95397604e-01 -3.50841880e-01 -4.05516207e-01 2.67765909e-01
1.20663655e+00 -2.01280445e-01 -2.53438532e-01 5.35098135e-01
-1.56730984e-03 -8.97132933e-01 1.09966314e+00 -3.46309632e-01
7.96097338e-01 -6.13812566e-01 2.88616598e-01 -1.13642216e+00
-3.78718615e-01 -5.72924435e-01 9.04982984e-02 1.21735680e+00
5.49229443e-01 -4.19565856e-01 6.20445430e-01 1.38243422e-01
-2.68738300e-01 -7.69588470e-01 -5.54494262e-01 -6.55947328e-01
-2.36963198e-01 4.88831475e-02 7.16710687e-01 7.81567752e-01
-1.85794085e-01 3.31589967e-01 -3.69900584e-01 2.73551106e-01
9.22402442e-01 3.61468136e-01 3.52436960e-01 -1.15324497e+00
-6.74799085e-02 -6.18354738e-01 -3.48119885e-01 -1.04904962e+00
-7.02889860e-02 -9.40655112e-01 3.80485356e-01 -1.31944358e+00
4.34908539e-01 -6.02803409e-01 -4.79422092e-01 5.47164321e-01
-5.18444955e-01 2.20388383e-01 6.01419657e-02 2.89404601e-01
-9.03672516e-01 9.17927802e-01 1.40464044e+00 -2.29149252e-01
-3.24646771e-01 4.41232920e-02 -8.12461257e-01 7.63604701e-01
7.29067862e-01 -4.83557701e-01 -3.51330936e-01 -2.23875418e-01
-2.52672225e-01 -1.07016325e-01 7.22812414e-01 -1.15425313e+00
2.40308404e-01 -2.15238288e-01 2.01665238e-01 -6.22491419e-01
6.11580312e-02 -3.82013589e-01 -2.50460327e-01 2.55017370e-01
-5.66331029e-01 -3.22574407e-01 1.25676200e-01 7.98037291e-01
-4.32042658e-01 -8.57924819e-02 1.23423409e+00 -5.81232505e-03
-6.87742174e-01 2.68469244e-01 9.20029283e-02 6.19638190e-02
1.07144713e+00 -1.67897195e-01 -3.26534957e-02 -4.17706043e-01
-6.57953143e-01 3.63493413e-01 5.38885772e-01 4.96781617e-01
8.58884394e-01 -1.48830211e+00 -5.47707617e-01 3.71221006e-01
3.14038813e-01 5.26298821e-01 3.62439871e-01 9.74211752e-01
-1.42326757e-01 4.78484869e-01 -2.91422218e-01 -6.91243827e-01
-9.18758690e-01 6.76890910e-01 1.20164305e-01 -4.85359311e-01
-5.37891686e-01 1.15935731e+00 7.95305669e-01 -3.13990176e-01
3.45333189e-01 -1.26283795e-01 -1.81412354e-01 -4.54546928e-01
4.96805012e-01 9.56274420e-02 -1.70374498e-01 -7.20557570e-01
-3.41922998e-01 5.42200625e-01 -2.26871908e-01 -1.44471331e-02
1.23997951e+00 -1.67875186e-01 -1.55562371e-01 1.67259961e-01
1.17473221e+00 -2.41310433e-01 -1.68871486e+00 -5.44408977e-01
3.34995389e-02 -5.19268334e-01 2.56073207e-01 -8.97443175e-01
-1.27057910e+00 7.55347729e-01 5.62896132e-01 -2.72012260e-02
1.26459455e+00 3.78974915e-01 8.18161845e-01 -2.52724022e-01
2.67053217e-01 -1.03105915e+00 3.52502137e-01 2.84328848e-01
9.48645532e-01 -1.43085945e+00 -1.49842069e-01 -8.43045473e-01
-7.26593912e-01 4.61751342e-01 5.76433599e-01 -4.14731205e-01
8.22254479e-01 -5.54333553e-02 -3.63068700e-01 -2.28190914e-01
-3.56670946e-01 -6.90615237e-01 7.49513566e-01 6.57372177e-01
-6.34360909e-02 1.63945869e-01 -2.04932585e-01 5.21580577e-01
9.60996896e-02 -8.14478621e-02 3.06674659e-01 8.40178847e-01
-6.28901482e-01 -5.83714604e-01 -3.33045870e-01 5.04708350e-01
-2.53876120e-01 -2.27107644e-01 -2.43571296e-01 3.29129308e-01
2.44015828e-01 8.63462985e-01 2.63454348e-01 -2.53696144e-01
9.99038145e-02 -4.07042086e-01 1.81797117e-01 -7.57536471e-01
-1.44181922e-01 1.85903236e-01 -4.12356615e-01 -3.94477040e-01
-5.05688071e-01 -5.51965773e-01 -1.14667857e+00 3.39944810e-01
-7.09839940e-01 1.97949889e-03 4.26930249e-01 9.45453107e-01
5.12628317e-01 5.43677270e-01 7.97572851e-01 -1.12822509e+00
-1.45670593e-01 -7.11986542e-01 -6.97930753e-01 4.12100047e-01
6.26019359e-01 -1.01273596e+00 -4.16494399e-01 2.85874844e-01] | [9.807169914245605, -0.30722007155418396] |
ff92b164-bf80-44ab-85a4-4c2b302c156f | simulspeech-end-to-end-simultaneous-speech-to | null | null | https://aclanthology.org/2020.acl-main.350 | https://aclanthology.org/2020.acl-main.350.pdf | SimulSpeech: End-to-End Simultaneous Speech to Text Translation | In this work, we develop SimulSpeech, an end-to-end simultaneous speech to text translation system which translates speech in source language to text in target language concurrently. SimulSpeech consists of a speech encoder, a speech segmenter and a text decoder, where 1) the segmenter builds upon the encoder and leverages a connectionist temporal classification (CTC) loss to split the input streaming speech in real time, 2) the encoder-decoder attention adopts a wait-$k$ strategy for simultaneous translation. SimulSpeech is more challenging than previous cascaded systems (with simultaneous automatic speech recognition (ASR) and simultaneous neural machine translation (NMT)). We introduce two novel knowledge distillation methods to ensure the performance: 1) Attention-level knowledge distillation transfers the knowledge from the multiplication of the attention matrices of simultaneous NMT and ASR models to help the training of the attention mechanism in SimulSpeech; 2) Data-level knowledge distillation transfers the knowledge from the full-sentence NMT model and also reduces the complexity of data distribution to help on the optimization of SimulSpeech. Experiments on MuST-C English-Spanish and English-German spoken language translation datasets show that SimulSpeech achieves reasonable BLEU scores and lower delay compared to full-sentence end-to-end speech to text translation (without simultaneous translation), and better performance than the two-stage cascaded simultaneous translation model in terms of BLEU scores and translation delay. | ['Tie-Yan Liu', 'Tao Qin', 'Yi Ren', 'Jinglin Liu', 'Chen Zhang', 'Zhou Zhao', 'Xu Tan'] | 2020-07-01 | null | null | null | acl-2020-6 | ['speech-to-text-translation', 'simultaneous-speech-to-text-translation'] | ['natural-language-processing', 'natural-language-processing'] | [ 2.66933054e-01 1.81079403e-01 -3.01179260e-01 -2.47750074e-01
-1.37665439e+00 -6.55281305e-01 4.31580007e-01 -3.63828152e-01
-5.64524651e-01 3.80746752e-01 1.34828612e-01 -1.23675799e+00
5.26199579e-01 -2.27317661e-01 -1.05279505e+00 -5.17497599e-01
3.99072737e-01 7.22487390e-01 3.58317867e-02 -1.18698478e-01
-4.15978640e-01 -4.51303236e-02 -6.71933055e-01 5.86007595e-01
9.66169953e-01 8.86868656e-01 6.31645024e-01 1.07804656e+00
-2.03877985e-01 8.22553992e-01 -5.47873080e-01 -6.62086785e-01
3.13479751e-01 -9.10873830e-01 -9.97764170e-01 -1.86097473e-01
-3.67784873e-02 -4.03639048e-01 -5.68621874e-01 7.34196186e-01
7.69345582e-01 7.72311985e-02 3.56271595e-01 -8.55024457e-01
-8.15662801e-01 1.12771988e+00 -3.34431976e-01 3.76907259e-01
2.39277035e-02 4.27311808e-01 9.93479133e-01 -1.26517034e+00
1.40756994e-01 1.41831899e+00 2.96385854e-01 6.01856649e-01
-1.17846990e+00 -7.01718926e-01 2.54925102e-01 1.53124765e-01
-1.31795251e+00 -1.10126877e+00 1.91348806e-01 -7.08095431e-02
1.91609108e+00 2.09983662e-01 5.11982679e-01 1.18917811e+00
2.56738394e-01 1.09106815e+00 5.40808976e-01 -5.37096798e-01
8.54674652e-02 2.13234052e-01 -2.02765614e-01 5.52969575e-01
-7.18520522e-01 2.79232115e-01 -7.43324816e-01 2.08815798e-01
6.05008602e-01 -3.65232378e-01 -2.61824310e-01 3.72604579e-01
-1.36796260e+00 5.88528693e-01 1.93119645e-01 1.18496262e-01
-3.97953957e-01 1.31348759e-01 6.37341142e-01 8.72532010e-01
5.85122764e-01 1.38498634e-01 -7.56434798e-01 -6.10965550e-01
-1.05103874e+00 -4.56645042e-01 6.90359473e-01 1.37171006e+00
4.45886552e-01 5.18505692e-01 -3.15021992e-01 8.38103950e-01
1.78115413e-01 1.09177852e+00 7.91211545e-01 -4.25244004e-01
1.31670976e+00 1.69927955e-01 -2.75022358e-01 -1.14921350e-02
1.80421710e-01 -5.22410810e-01 -7.93756962e-01 -3.64969522e-01
-5.72244488e-02 -5.06482303e-01 -1.01979601e+00 1.71043921e+00
3.32331322e-02 1.96510360e-01 5.14263153e-01 8.84300351e-01
4.07310426e-01 1.24903297e+00 -2.18878120e-01 -4.34444219e-01
1.16927230e+00 -1.44153023e+00 -9.19396877e-01 -5.24724185e-01
8.42353940e-01 -9.76630330e-01 1.25352693e+00 -1.00312978e-01
-1.68292356e+00 -7.06130624e-01 -9.03205276e-01 -3.10304165e-01
-9.74419490e-02 4.53111708e-01 -8.69374126e-02 3.85596454e-01
-1.29302418e+00 2.24399999e-01 -1.14328063e+00 -1.08162716e-01
5.49395196e-03 6.35404587e-01 9.12791956e-03 1.48488179e-01
-1.45453370e+00 1.02624547e+00 2.95057476e-01 2.33522177e-01
-9.40115631e-01 -5.80786228e-01 -8.34255219e-01 2.83028901e-01
2.67417729e-01 -8.81753743e-01 1.70482183e+00 -1.39626670e+00
-2.33869195e+00 4.21863556e-01 -5.63517988e-01 -6.61086023e-01
4.29476082e-01 -2.61241466e-01 -3.35812956e-01 6.53822646e-02
-1.57507390e-01 6.77485824e-01 9.88256931e-01 -5.39504051e-01
-7.45803833e-01 1.07512452e-01 -4.39340740e-01 6.21941686e-01
-1.33020446e-01 3.50292593e-01 -6.34192467e-01 -7.11241663e-01
-1.62254184e-01 -9.52437758e-01 1.38327852e-01 -5.69190085e-01
-4.15835291e-01 -4.98017110e-02 7.77099371e-01 -1.13941348e+00
1.43289840e+00 -2.25087047e+00 5.26466548e-01 -1.52437419e-01
-3.98351818e-01 5.54695964e-01 -4.04212892e-01 4.79664892e-01
-1.91002131e-01 -1.83042716e-02 -1.45166263e-01 -9.72985566e-01
-7.19294176e-02 2.95729965e-01 -7.66317427e-01 1.89651832e-01
3.79236460e-01 1.36280060e+00 -6.50008202e-01 -3.03904295e-01
1.53982788e-01 3.54804277e-01 -4.18428093e-01 4.50839847e-01
-2.83356130e-01 3.81442815e-01 -7.97355399e-02 3.24504435e-01
1.40646175e-01 2.79720891e-02 1.37558043e-01 2.32964948e-01
-1.09466590e-01 1.12829161e+00 -4.53761399e-01 1.74550653e+00
-7.54468262e-01 8.01451862e-01 2.21430674e-01 -7.53719032e-01
5.92836320e-01 8.58900428e-01 -3.69241163e-02 -9.09576595e-01
2.84445465e-01 3.51960629e-01 1.10383332e-01 -1.75817847e-01
3.53957772e-01 -1.29931182e-01 -3.40514109e-02 5.91088653e-01
1.32281154e-01 -2.17250615e-01 -2.39358589e-01 1.40259311e-01
9.10688221e-01 -4.90684733e-02 -1.96388155e-01 1.29186232e-02
3.83753121e-01 -1.45772602e-02 3.98398459e-01 4.68282133e-01
1.81106515e-02 4.02054191e-01 2.11432800e-01 8.99356157e-02
-1.26177335e+00 -1.21629214e+00 5.22184134e-01 1.34852099e+00
-2.36804605e-01 -1.51834935e-01 -8.97728980e-01 -6.96602404e-01
-3.78345579e-01 1.00400889e+00 -9.02893767e-02 -4.31914568e-01
-9.28439140e-01 -2.63230234e-01 8.60566497e-01 6.45260692e-01
4.85220671e-01 -9.20876682e-01 8.73600245e-02 3.67056727e-01
-5.69124043e-01 -1.38947737e+00 -1.43312013e+00 4.97961968e-01
-7.10168362e-01 -3.15664351e-01 -7.71607280e-01 -1.06439793e+00
4.69940096e-01 2.75694907e-01 8.38009298e-01 -4.16808307e-01
4.29161847e-01 8.48616473e-03 -4.00244355e-01 -1.68356687e-01
-9.97775614e-01 4.90109682e-01 3.31025004e-01 7.59641156e-02
2.04794213e-01 -3.20844859e-01 -3.42262834e-01 4.23100322e-01
-5.64039528e-01 4.57070589e-01 9.53798056e-01 1.05996561e+00
5.49138486e-01 -2.61678398e-01 5.18145919e-01 -1.89782992e-01
6.42105281e-01 -2.61965990e-01 -6.30603075e-01 3.77486169e-01
-6.39037669e-01 1.31004870e-01 9.51796472e-01 -7.52274156e-01
-9.04640794e-01 -3.20223272e-02 -4.45188999e-01 -8.45664680e-01
3.92651230e-01 5.57541132e-01 -2.40687966e-01 4.44757730e-01
3.43806565e-01 9.45522845e-01 1.30635083e-01 -3.20853144e-01
5.29348433e-01 1.18498170e+00 5.25173843e-01 -2.65444815e-01
6.00370884e-01 -2.93681622e-01 -8.39605153e-01 -6.20569527e-01
-4.27703381e-01 -2.73356557e-01 -5.22182167e-01 2.14844540e-01
9.48621690e-01 -1.21440291e+00 -6.53634131e-01 4.55411047e-01
-1.50133097e+00 -9.91979778e-01 -3.67723823e-01 7.44487107e-01
-6.98786438e-01 7.74122551e-02 -1.07959080e+00 -6.60794675e-01
-8.29840422e-01 -1.47448015e+00 1.19097090e+00 -2.84623921e-01
-3.93123552e-02 -8.70882988e-01 -2.59856611e-01 4.83184189e-01
6.48177087e-01 -9.52174842e-01 9.88099813e-01 -6.74285710e-01
-6.15020931e-01 1.16927095e-01 -1.23748183e-01 7.46031284e-01
7.90747069e-03 -2.89773524e-01 -7.61573613e-01 -6.58041000e-01
1.72646455e-02 -3.03606868e-01 7.18224168e-01 3.55503023e-01
4.85383987e-01 -7.35376358e-01 -6.82559460e-02 5.93489170e-01
9.37756896e-01 6.35328531e-01 4.57485825e-01 -3.33917469e-01
6.29257679e-01 2.70173818e-01 3.23546588e-01 -1.26560286e-01
5.28895199e-01 8.37104082e-01 -1.03449211e-01 -3.46393019e-01
-4.76587355e-01 -3.88411671e-01 1.41014040e+00 1.94597518e+00
5.17598510e-01 -5.74369788e-01 -9.17329311e-01 4.75104779e-01
-1.91409135e+00 -7.86062360e-01 2.22582862e-01 2.22866821e+00
1.15584886e+00 1.38915911e-01 6.35732431e-03 -1.42455742e-01
6.63259327e-01 -1.05743058e-01 -6.97219014e-01 -8.53050947e-01
-1.15550552e-02 1.88575789e-01 5.54082274e-01 1.05785513e+00
-6.31713986e-01 1.51231015e+00 5.93071461e+00 1.20136297e+00
-1.48006916e+00 6.39590621e-01 6.19942248e-01 -3.89909267e-01
-1.39843196e-01 -4.54091541e-02 -8.79046082e-01 4.96589363e-01
1.70394707e+00 -1.08388446e-01 9.02900875e-01 4.85271513e-01
4.42657828e-01 3.29770535e-01 -1.33710742e+00 9.17832017e-01
-2.19194874e-01 -1.15426040e+00 1.84383780e-01 -1.57130867e-01
4.71513391e-01 5.68454146e-01 1.82027861e-01 7.22301304e-01
5.17065287e-01 -8.44525337e-01 1.01605201e+00 7.11593553e-02
1.28956938e+00 -8.89783084e-01 6.02430761e-01 6.52529657e-01
-1.26032519e+00 3.31412069e-02 4.38293442e-02 5.49042318e-03
3.55583608e-01 1.37758791e-01 -1.35120583e+00 5.84775746e-01
1.77719414e-01 3.75717521e-01 5.52580804e-02 2.39582241e-01
-3.73790920e-01 1.18759179e+00 -2.52365947e-01 -1.15414105e-01
5.06011844e-01 -1.13137372e-01 4.66040224e-01 1.54163849e+00
3.42158228e-01 -1.70915946e-01 1.44789606e-01 5.94082594e-01
-1.96251228e-01 7.98247159e-02 -1.72744647e-01 -4.08490896e-01
5.79229593e-01 5.22267342e-01 -2.11123928e-01 -5.37217438e-01
-4.81155813e-01 1.66612756e+00 3.35566968e-01 7.62156606e-01
-9.52407002e-01 -3.87186319e-01 7.58319736e-01 -1.19662799e-01
4.54612434e-01 -4.95008200e-01 -2.96268314e-01 -1.22584653e+00
1.28469080e-01 -1.03259158e+00 -1.16716385e-01 -7.81098843e-01
-8.63507092e-01 1.01943350e+00 -4.50823247e-01 -9.99735773e-01
-5.17357171e-01 -3.16905200e-01 -3.71436596e-01 1.50148106e+00
-1.52341211e+00 -1.22153258e+00 5.57832956e-01 7.13366508e-01
1.13295412e+00 -3.13832045e-01 8.66726637e-01 3.98543268e-01
-8.72985661e-01 9.95873213e-01 2.21430704e-01 4.20064837e-01
7.02644289e-01 -9.71351504e-01 1.11965370e+00 1.14981067e+00
1.10362038e-01 4.29102302e-01 2.24333838e-01 -5.95924020e-01
-1.68743098e+00 -1.33326161e+00 1.27127707e+00 -2.95988172e-01
6.16585255e-01 -7.66349792e-01 -7.06111133e-01 9.04770315e-01
7.19675183e-01 -4.00697857e-01 5.57023823e-01 -1.88552156e-01
-2.06822053e-01 -1.36217579e-01 -4.82915610e-01 8.15475523e-01
8.49868536e-01 -1.01547873e+00 -5.08465052e-01 3.04559886e-01
1.43806028e+00 -5.90656102e-01 -5.69377720e-01 3.14926170e-02
3.73124659e-01 -2.00349599e-01 5.97871959e-01 -5.45481384e-01
3.01527888e-01 -1.34872153e-01 -3.12400043e-01 -1.57981467e+00
-1.81255549e-01 -1.19564450e+00 -2.17813671e-01 1.04665148e+00
1.09235740e+00 -6.76475108e-01 3.42005014e-01 2.43953660e-01
-6.63318634e-01 -8.43338966e-01 -1.38103461e+00 -9.45375383e-01
3.18987727e-01 -5.42638123e-01 4.92025107e-01 6.71629429e-01
2.02235803e-01 9.58102167e-01 -4.02990311e-01 2.39662096e-01
-9.21063125e-02 -1.45361438e-01 5.41752756e-01 -1.64864659e-01
-7.12574899e-01 -6.04229510e-01 3.21089149e-01 -1.93114281e+00
2.30166405e-01 -1.29302168e+00 4.10604954e-01 -1.41374457e+00
-2.33642161e-02 -1.74507633e-01 -4.39311378e-02 7.23177195e-01
-2.24704042e-01 -4.14163321e-01 1.52626887e-01 2.79251009e-01
-3.49290282e-01 9.57775116e-01 1.21215594e+00 -1.70410261e-01
-4.16392028e-01 1.33875236e-01 -3.47247511e-01 5.29458858e-02
5.53177893e-01 -4.55068946e-01 -5.77162266e-01 -1.06632590e+00
-1.24261573e-01 6.54324055e-01 -1.51495308e-01 -5.50798655e-01
5.91035187e-01 -3.77326086e-02 -1.42266169e-01 -5.06794989e-01
4.05683666e-01 -6.53105199e-01 -1.79982156e-01 4.56622571e-01
-4.33174223e-01 3.67746681e-01 3.37950021e-01 2.95512289e-01
-2.67435551e-01 2.57632226e-01 8.41805756e-01 1.17206626e-01
-8.79117772e-02 3.39014351e-01 -7.15984762e-01 4.95832153e-02
6.21969044e-01 -3.76017801e-02 -1.99282184e-01 -5.29438198e-01
-6.95986807e-01 3.95135373e-01 1.35617061e-02 7.16926813e-01
6.60339296e-01 -1.31567967e+00 -9.16583836e-01 4.89352107e-01
-3.03047091e-01 -3.71560790e-02 -4.73807706e-03 1.22504592e+00
-2.63012975e-01 6.95970654e-01 5.15062809e-01 -6.23500764e-01
-1.25239968e+00 5.05463660e-01 5.26533961e-01 -2.35685140e-01
-4.08575624e-01 1.09200466e+00 1.83274031e-01 -5.87643385e-01
4.23459381e-01 -6.56928122e-01 7.47221589e-01 -3.30761790e-01
3.73538643e-01 1.61131427e-01 3.04154307e-01 -5.17095566e-01
-3.83011580e-01 1.35443389e-01 -2.76554972e-01 -6.54327095e-01
8.48027229e-01 -5.35854757e-01 1.93012521e-01 4.33636844e-01
1.26182187e+00 -1.43399790e-01 -1.21371114e+00 -5.25184572e-01
-2.97761470e-01 9.98582765e-02 1.77703157e-01 -1.14505637e+00
-8.61763418e-01 1.25305092e+00 1.85765877e-01 -2.21158907e-01
1.18395197e+00 -1.50192305e-01 1.35253656e+00 4.25915152e-01
1.21317454e-01 -1.23740041e+00 6.87199607e-02 1.12650824e+00
9.20863390e-01 -1.03801179e+00 -8.14506292e-01 -1.67235136e-01
-8.76755714e-01 1.00621307e+00 3.31137836e-01 5.08375347e-01
5.32747746e-01 7.59200037e-01 4.41050351e-01 4.47487950e-01
-1.33248162e+00 -1.17889844e-01 2.06970781e-01 1.85387731e-01
2.72131473e-01 1.88278124e-01 1.90752044e-01 7.14523613e-01
-3.83154809e-01 -2.06726670e-01 -2.60966644e-02 6.42782927e-01
-2.66812980e-01 -1.01539409e+00 -1.10864900e-01 -1.81413945e-02
-3.42497170e-01 -7.61941791e-01 -4.67056423e-01 1.94864422e-01
-2.05793276e-01 1.26584518e+00 1.85744390e-01 -6.71207070e-01
4.35363114e-01 4.11958188e-01 1.72876582e-01 -6.43868804e-01
-1.04615259e+00 5.19555628e-01 1.95328668e-01 -4.31644887e-01
3.37464631e-01 -4.96518791e-01 -1.40905023e+00 -3.81988972e-01
-6.72834039e-01 3.51115525e-01 8.03322792e-01 1.10654473e+00
8.13163579e-01 7.22858548e-01 9.01929915e-01 -5.89716911e-01
-8.12924087e-01 -1.14788032e+00 -1.05937392e-01 -2.19042525e-01
6.70536637e-01 1.79047465e-01 -2.22579852e-01 2.41141394e-01] | [14.487174987792969, 7.1487812995910645] |
5795f6be-a96b-4e7c-9dc5-c6df72bb78a8 | can-large-language-models-be-an-alternative | 2305.01937 | null | https://arxiv.org/abs/2305.01937v1 | https://arxiv.org/pdf/2305.01937v1.pdf | Can Large Language Models Be an Alternative to Human Evaluations? | Human evaluation is indispensable and inevitable for assessing the quality of texts generated by machine learning models or written by humans. However, human evaluation is very difficult to reproduce and its quality is notoriously unstable, hindering fair comparisons among different natural language processing (NLP) models and algorithms. Recently, large language models (LLMs) have demonstrated exceptional performance on unseen tasks when only the task instructions are provided. In this paper, we explore if such an ability of the LLMs can be used as an alternative to human evaluation. We present the LLMs with the exact same instructions, samples to be evaluated, and questions used to conduct human evaluation, and then ask the LLMs to generate responses to those questions; we dub this LLM evaluation. We use human evaluation and LLM evaluation to evaluate the texts in two NLP tasks: open-ended story generation and adversarial attacks. We show that the result of LLM evaluation is consistent with the results obtained by expert human evaluation: the texts rated higher by human experts are also rated higher by the LLMs. We also find that the results of LLM evaluation are stable over different formatting of the task instructions and the sampling algorithm used to generate the answer. We are the first to show the potential of using LLMs to assess the quality of texts and discuss the limitations and ethical considerations of LLM evaluation. | ['Hung-Yi Lee', 'Cheng-Han Chiang'] | 2023-05-03 | null | null | null | null | ['story-generation'] | ['natural-language-processing'] | [ 3.24090689e-01 4.34160084e-01 1.93203270e-01 -3.63630831e-01
-1.13534260e+00 -9.95085001e-01 9.64841187e-01 2.77673900e-01
-7.71757185e-01 8.96157801e-01 4.43374753e-01 -3.09584200e-01
6.29799888e-02 -5.99444270e-01 -6.87726676e-01 -1.07526846e-01
4.72293139e-01 8.20138752e-01 1.15616377e-02 -3.19145620e-01
5.27521908e-01 4.44646999e-02 -1.33967364e+00 7.31868207e-01
9.91254926e-01 5.19381344e-01 6.24638703e-03 1.02831852e+00
-4.27129343e-02 9.54540730e-01 -1.34097219e+00 -1.02903402e+00
6.01623245e-02 -6.21566236e-01 -1.20617759e+00 -9.49992836e-02
3.33327979e-01 -4.48027134e-01 5.00237495e-02 8.46880853e-01
8.11079502e-01 2.69382894e-01 9.07888114e-01 -1.24406815e+00
-7.62042761e-01 9.08521473e-01 1.41752914e-01 7.20922053e-02
1.21560466e+00 4.35886383e-01 9.64412630e-01 -7.68694222e-01
7.81030059e-01 1.54598773e+00 5.39683521e-01 8.36704254e-01
-1.12424660e+00 -4.19218749e-01 -2.74333745e-01 1.28543362e-01
-1.02061236e+00 -4.58306283e-01 4.27687436e-01 -4.68808711e-01
7.07680285e-01 2.63457149e-01 1.08051397e-01 1.69072723e+00
2.97003865e-01 7.39282370e-01 1.11313224e+00 -7.00463116e-01
3.15899819e-01 6.38912559e-01 6.95093274e-02 3.25531602e-01
1.37733072e-01 1.70455113e-01 -5.41131198e-01 -4.05517399e-01
1.13571510e-01 -9.03913975e-01 -3.48148346e-01 3.22907865e-01
-1.37897027e+00 1.22552049e+00 -9.60112438e-02 3.90421301e-01
-3.08679342e-01 -1.49714544e-01 5.22928119e-01 4.63532358e-01
4.17753965e-01 1.13407421e+00 -2.68932462e-01 -2.89086133e-01
-8.82806599e-01 5.95830262e-01 1.14841723e+00 5.55991471e-01
-3.48139405e-02 3.59758623e-02 -5.67873240e-01 7.25106895e-01
2.19132513e-01 5.88752687e-01 8.73303950e-01 -1.20133650e+00
6.37064636e-01 9.38228592e-02 5.16720176e-01 -1.27368140e+00
-2.84032553e-01 9.44479648e-03 -4.54152703e-01 2.43368670e-01
6.53090775e-01 -4.52657521e-01 -2.70725429e-01 1.83050787e+00
-1.70096308e-02 -4.67065454e-01 3.03205341e-01 6.83289111e-01
9.17092323e-01 8.90406847e-01 1.44313976e-01 -2.78228372e-01
1.19573009e+00 -6.27725065e-01 -8.51681113e-01 -8.60329941e-02
8.36158216e-01 -9.53523934e-01 1.63359940e+00 6.04868352e-01
-1.14460981e+00 -7.39649475e-01 -1.07661521e+00 2.61459231e-01
-1.32304043e-01 -1.07354797e-01 8.76473263e-02 8.67629886e-01
-9.67569113e-01 7.07868338e-01 -1.56729981e-01 -1.56301364e-01
-1.82705652e-02 6.74089566e-02 -2.80718446e-01 4.45504524e-02
-1.86068439e+00 1.25538099e+00 3.76592398e-01 -2.34033838e-01
-8.85813355e-01 -2.56970108e-01 -9.33475971e-01 -1.92404836e-01
1.35976955e-01 -4.56735194e-01 1.67742145e+00 -9.45278764e-01
-1.42841303e+00 9.07453120e-01 -9.07365084e-02 -6.12771809e-01
7.55012929e-01 -2.92019397e-01 -4.69147772e-01 2.99661398e-01
2.85145372e-01 6.02817357e-01 7.70653307e-01 -1.22118950e+00
-1.89669758e-01 2.32070789e-01 1.92783009e-02 1.68537036e-01
-1.27416521e-01 4.58931148e-01 3.59614879e-01 -7.74271548e-01
-5.64812481e-01 -7.51429677e-01 -1.27933800e-01 -4.73860264e-01
-4.84279126e-01 -2.85417646e-01 2.47567087e-01 -7.21073806e-01
1.33032846e+00 -1.93352759e+00 -2.37074479e-01 6.00411184e-02
-6.72897995e-02 5.32149374e-01 -5.23171484e-01 6.26092434e-01
1.82952192e-02 7.59839833e-01 -3.98739316e-02 -3.19469392e-01
4.64573950e-01 -1.10438079e-01 -5.50235152e-01 7.05299973e-02
1.97960481e-01 1.08751225e+00 -8.78240645e-01 -7.17259288e-01
-1.06442384e-01 1.66505620e-01 -4.56628144e-01 3.85191709e-01
-2.43709594e-01 2.46412188e-01 -2.35431805e-01 1.51577279e-01
1.66040827e-02 -3.94602716e-02 -3.07954043e-01 2.11958021e-01
2.95672208e-01 4.07170296e-01 -1.01273274e+00 1.05618894e+00
-3.54854703e-01 8.91220689e-01 -2.32008815e-01 -4.13195074e-01
9.77095902e-01 8.26849043e-01 -2.72488922e-01 -6.03834152e-01
2.08170861e-01 2.43170813e-01 1.33148104e-01 -8.09625089e-01
6.07965469e-01 -3.88675034e-01 -4.05392557e-01 1.08299387e+00
2.26739999e-02 -7.36186624e-01 4.78635311e-01 5.07608175e-01
9.59552407e-01 -2.07456231e-01 3.19357455e-01 -4.89870422e-02
5.59125900e-01 4.84606512e-02 1.85830399e-01 1.22892129e+00
-2.12554678e-01 6.13720775e-01 5.21714449e-01 -2.26532146e-01
-1.13182998e+00 -9.16718900e-01 2.34968625e-02 1.09967256e+00
-3.24039698e-01 -3.62255663e-01 -1.15803444e+00 -6.54655993e-01
-4.10527706e-01 1.53252888e+00 -5.56198955e-01 -2.38230750e-01
-3.95645469e-01 -5.69030643e-01 9.59938705e-01 2.22848177e-01
7.05186874e-02 -1.70573008e+00 -6.55464590e-01 2.42538452e-01
-6.61338151e-01 -1.15565753e+00 -3.85435730e-01 -2.46367022e-01
-3.71246666e-01 -8.58803451e-01 -4.34095711e-01 -5.77603102e-01
4.44385141e-01 -3.53037298e-01 1.27444220e+00 4.87003177e-02
2.55688936e-01 4.19465780e-01 -6.84591174e-01 -8.03815126e-01
-1.40285122e+00 -4.41926718e-02 1.84806153e-01 -1.93561271e-01
2.59216130e-01 -2.56398052e-01 -9.45906192e-02 4.68986243e-01
-1.31542218e+00 -9.33869556e-02 5.10462761e-01 7.62991309e-01
-3.49276513e-02 1.73699528e-01 8.58588934e-01 -1.07076097e+00
1.51874137e+00 -2.63926297e-01 -7.61435702e-02 3.37277979e-01
-5.02857208e-01 2.64436305e-01 8.50403547e-01 -8.14884782e-01
-9.40904021e-01 -4.30845559e-01 -3.64215016e-01 9.52072144e-02
-3.99700493e-01 4.93156165e-01 -2.40629226e-01 1.91268682e-01
1.17619336e+00 -1.87797770e-01 -1.16022706e-01 -1.64312795e-01
3.21392119e-01 6.89965010e-01 4.50268745e-01 -7.42853343e-01
9.76358175e-01 -7.82193169e-02 -5.52981257e-01 -5.79111874e-01
-9.15679574e-01 2.63456076e-01 -2.71997273e-01 -4.71202910e-01
8.54560733e-01 -3.16506654e-01 -5.94407678e-01 2.59290338e-01
-1.57117522e+00 -3.96245867e-01 -4.11022753e-01 5.39185524e-01
-6.87451780e-01 4.54523385e-01 -8.22548091e-01 -9.00899887e-01
-4.77353811e-01 -1.20657730e+00 9.33316767e-01 -1.03329733e-01
-1.24685073e+00 -9.84540164e-01 1.85691640e-01 7.16398239e-01
3.04005235e-01 3.70449096e-01 1.22370780e+00 -1.17987108e+00
2.51768142e-01 -6.43947303e-01 3.28840107e-01 4.67599332e-01
-2.41392434e-01 8.49501714e-02 -9.73191619e-01 -4.51658927e-02
4.44059610e-01 -9.88179207e-01 2.64511913e-01 1.77968174e-01
5.41345596e-01 -7.36442327e-01 2.84734368e-01 -1.17778726e-01
9.03084159e-01 6.11189418e-02 6.89257205e-01 3.35908562e-01
2.22544953e-01 1.17766953e+00 6.92064464e-01 2.15486914e-01
-1.28139518e-02 3.63981813e-01 -6.21797927e-02 2.59352297e-01
3.17936122e-01 -6.92469478e-01 8.26552749e-01 8.99306774e-01
3.98454905e-01 -7.12302744e-01 -8.17061961e-01 2.93973953e-01
-1.43066955e+00 -1.19208443e+00 -1.55216470e-01 2.09146190e+00
1.09688580e+00 5.09578645e-01 -6.97429478e-02 3.91314358e-01
7.59230971e-01 1.19512975e-01 -7.20180422e-02 -9.12446260e-01
-3.82672548e-01 3.11827242e-01 -1.60683036e-01 7.97606587e-01
-8.35796237e-01 5.90086401e-01 7.41741467e+00 9.41673636e-01
-5.84225833e-01 6.25066459e-02 7.71480739e-01 -7.36031011e-02
-5.15516758e-01 -1.20172724e-01 -5.10018826e-01 4.63635772e-01
1.42181396e+00 -5.86138308e-01 2.54680187e-01 5.56286693e-01
4.85853821e-01 -9.02726650e-02 -1.32005858e+00 6.54147804e-01
3.04042697e-01 -8.58064651e-01 3.72726321e-01 -2.66610533e-01
7.03671515e-01 -5.14110804e-01 -4.25292440e-02 4.50635940e-01
4.33127195e-01 -1.34174991e+00 9.53916132e-01 2.94774830e-01
5.84138870e-01 -6.09715998e-01 1.05011868e+00 8.86795282e-01
-4.19767737e-01 8.59590154e-03 -2.59866357e-01 -2.36577049e-01
4.87439424e-01 2.59375930e-01 -1.09450078e+00 1.78255901e-01
1.71453565e-01 -2.24696323e-01 -7.82243729e-01 5.18605649e-01
-6.14117026e-01 7.54485309e-01 5.24024069e-02 -4.91663933e-01
1.98725164e-01 2.19514370e-01 6.45940304e-01 1.21413159e+00
1.08872361e-01 -3.24273482e-02 -1.84875056e-02 8.78582001e-01
-1.04180351e-01 5.31121492e-01 -7.87675500e-01 -4.00383502e-01
4.61929262e-01 9.48249936e-01 -3.62092882e-01 -3.75440061e-01
-1.41440034e-01 7.96862841e-01 9.28204432e-02 3.09923947e-01
-6.70243502e-01 -4.38131154e-01 -1.01857394e-01 2.49181166e-01
-4.61363316e-01 1.71732515e-01 -2.71080375e-01 -9.27245677e-01
-1.05302734e-02 -1.55077350e+00 3.07294041e-01 -1.17070758e+00
-1.45659673e+00 9.09880936e-01 1.55413672e-01 -1.12054777e+00
-8.81984949e-01 -6.41063571e-01 -7.04534411e-01 8.51846337e-01
-7.01183259e-01 -4.76281971e-01 9.89802852e-02 2.09373668e-01
6.13780677e-01 -1.28852829e-01 9.15121078e-01 -6.26774505e-02
-2.41460830e-01 7.78972626e-01 -3.21183175e-01 3.39654744e-01
9.55910087e-01 -1.04687345e+00 4.24393088e-01 7.76510358e-01
3.09130222e-01 6.18773878e-01 1.12206495e+00 -6.55945122e-01
-5.79180300e-01 -7.52314866e-01 1.34670198e+00 -9.35913086e-01
6.52565777e-01 -2.86826998e-01 -8.97415817e-01 3.06781948e-01
4.60252792e-01 -9.23054457e-01 8.47110152e-01 -4.03988183e-01
-1.42154023e-01 6.21912241e-01 -1.27304399e+00 7.80652344e-01
4.21945602e-01 -5.95355272e-01 -1.15672433e+00 5.01196742e-01
1.09205401e+00 -1.12503551e-01 -6.45818889e-01 2.95757115e-01
4.54038799e-01 -6.58144891e-01 6.20623291e-01 -8.72621775e-01
8.56244326e-01 -7.97357038e-03 -2.30511516e-01 -1.39819980e+00
-9.36635956e-02 -8.09380293e-01 3.21003467e-01 1.43359435e+00
9.98858154e-01 -5.86754858e-01 4.00352299e-01 1.02011454e+00
2.30594769e-01 -5.39828062e-01 -5.14897525e-01 -6.08453333e-01
3.27244222e-01 -6.91437125e-01 3.61069828e-01 8.31758916e-01
1.75027937e-01 6.99270248e-01 -3.95211488e-01 -2.25121185e-01
3.51733506e-01 -3.32196653e-01 8.25889111e-01 -1.05965054e+00
-5.44343829e-01 -4.78525698e-01 -7.03229010e-02 -4.67455000e-01
4.26629007e-01 -7.57486403e-01 2.06598133e-01 -1.30720222e+00
1.66808337e-01 2.12298647e-01 2.62118250e-01 -2.63608973e-02
-4.65707421e-01 6.02571741e-02 4.40683633e-01 3.31914797e-02
-5.03585577e-01 4.34300303e-01 1.13592970e+00 6.12822827e-03
-1.42339513e-01 -1.91208180e-02 -9.00648415e-01 8.32439244e-01
9.79586601e-01 -6.33076668e-01 -4.43055242e-01 -3.89063478e-01
4.73263621e-01 6.44548684e-02 2.96644032e-01 -9.41092253e-01
1.10501826e-01 -2.65159905e-02 4.06193942e-01 -2.10044503e-01
5.62988035e-02 -2.58957475e-01 2.04390436e-02 3.90489757e-01
-1.04325461e+00 3.05200219e-01 1.41067758e-01 1.80482447e-01
-3.08157742e-01 -1.01592684e+00 5.52409828e-01 -3.33282918e-01
-2.18726918e-01 -1.88649997e-01 -8.21253955e-01 5.58819175e-01
6.68603599e-01 -1.02147542e-01 -3.31838489e-01 -1.06299162e+00
-5.21231592e-01 1.75937936e-01 4.82258409e-01 3.71725202e-01
6.51769340e-01 -1.27684689e+00 -1.24958336e+00 -2.06474751e-01
-3.94447185e-02 -4.16482568e-01 1.91315394e-02 4.36566204e-01
-5.82128823e-01 4.28809673e-01 6.36412278e-02 -1.07872486e-01
-1.09242737e+00 6.39641345e-01 4.73802071e-03 -6.58123970e-01
-1.19650476e-02 7.02400208e-01 1.09256811e-01 -5.41056991e-01
2.73645043e-01 1.16391696e-01 -3.42681229e-01 1.08811639e-01
8.59192014e-01 2.24462122e-01 -1.44859895e-01 -6.49447322e-01
6.59153461e-02 8.74887556e-02 4.40558642e-02 -9.36592638e-01
8.19111347e-01 9.14198384e-02 1.55802682e-01 6.25918388e-01
9.53543842e-01 2.37862572e-01 -5.68369865e-01 -6.69175163e-02
1.30605370e-01 -2.16828197e-01 -6.29069388e-01 -8.70728850e-01
-2.43749812e-01 9.20883894e-01 -2.60481276e-02 3.65057468e-01
7.64344156e-01 -1.67064339e-01 7.92969763e-01 5.50260723e-01
4.53891635e-01 -1.23968720e+00 4.71076399e-01 4.53438580e-01
1.26486492e+00 -1.10991132e+00 -2.02930167e-01 -9.41272974e-02
-1.05285406e+00 8.84219170e-01 5.73958158e-01 1.40018329e-01
1.15506127e-01 1.17430478e-01 3.74814034e-01 1.27076074e-01
-1.00064218e+00 3.63361984e-01 2.43800893e-01 6.71947002e-01
5.59077621e-01 -6.11986825e-03 -6.44304156e-01 8.00927520e-01
-8.63853276e-01 -2.12312594e-01 6.97407842e-01 7.41783321e-01
-4.20430213e-01 -1.20215857e+00 -8.22001517e-01 4.52929020e-01
-7.71606445e-01 -1.67715147e-01 -9.82271373e-01 6.64388359e-01
-2.23842040e-01 1.55366421e+00 -5.06806552e-01 -5.70528626e-01
4.43830788e-01 4.80036974e-01 1.55324876e-01 -7.17802525e-01
-8.55515718e-01 -4.96633351e-01 5.51977277e-01 -3.23247343e-01
-1.00554414e-01 -5.43331265e-01 -9.78536189e-01 -3.43343168e-01
-2.81027466e-01 6.51158214e-01 2.86058187e-01 1.04556942e+00
-8.71752575e-02 5.53581789e-02 6.39837801e-01 -6.82132661e-01
-9.90770102e-01 -1.27105534e+00 -1.61912724e-01 9.55079019e-01
-6.02940880e-02 -2.37885788e-01 -7.58228660e-01 1.44275054e-01] | [11.742654800415039, 8.829221725463867] |
ace2454e-d1c0-42d8-875d-5b5d0e8df6cc | down-sampled-epsilon-lexicase-selection-for | 2302.04301 | null | https://arxiv.org/abs/2302.04301v1 | https://arxiv.org/pdf/2302.04301v1.pdf | Down-Sampled Epsilon-Lexicase Selection for Real-World Symbolic Regression Problems | Epsilon-lexicase selection is a parent selection method in genetic programming that has been successfully applied to symbolic regression problems. Recently, the combination of random subsampling with lexicase selection significantly improved performance in other genetic programming domains such as program synthesis. However, the influence of subsampling on the solution quality of real-world symbolic regression problems has not yet been studied. In this paper, we propose down-sampled epsilon-lexicase selection which combines epsilon-lexicase selection with random subsampling to improve the performance in the domain of symbolic regression. Therefore, we compare down-sampled epsilon-lexicase with traditional selection methods on common real-world symbolic regression problems and analyze its influence on the properties of the population over a genetic programming run. We find that the diversity is reduced by using down-sampled epsilon-lexicase selection compared to standard epsilon-lexicase selection. This comes along with high hyperselection rates we observe for down-sampled epsilon-lexicase selection. Further, we find that down-sampled epsilon-lexicase selection outperforms the traditional selection methods on all studied problems. Overall, with down-sampled epsilon-lexicase selection we observe an improvement of the solution quality of up to 85% in comparison to standard epsilon-lexicase selection. | ['Franz Rothlauf', 'Dominik Sobania', 'Alina Geiger'] | 2023-02-08 | null | null | null | null | ['program-synthesis'] | ['computer-code'] | [ 5.04937053e-01 7.95805454e-02 -2.56116390e-01 -7.22743049e-02
-3.15123647e-01 -3.06644827e-01 2.68184572e-01 3.96033585e-01
-2.32128412e-01 1.21256840e+00 -2.73654431e-01 -2.35377327e-01
-6.33345783e-01 -1.11183739e+00 -6.34563088e-01 -7.10037410e-01
-1.84213102e-01 6.77950442e-01 3.76322776e-01 -5.71421504e-01
6.94631398e-01 3.56220782e-01 -2.21447301e+00 3.51054817e-01
1.11439061e+00 4.21244830e-01 7.73607418e-02 3.93442929e-01
1.52156115e-01 5.98067604e-02 -6.29278243e-01 2.03212813e-01
2.82744676e-01 -8.07081103e-01 -5.45006394e-01 -3.98247004e-01
5.29677160e-02 5.39753795e-01 4.81046677e-01 1.07169092e+00
3.83640438e-01 1.15553223e-01 4.21268165e-01 -1.44309282e+00
7.84693006e-03 1.08023441e+00 -8.80931556e-01 -6.64068386e-02
5.70755482e-01 -6.43149167e-02 8.41666341e-01 -5.23562372e-01
7.54673958e-01 1.33858824e+00 7.86405802e-01 1.04313694e-01
-1.64460719e+00 -6.81103289e-01 9.98508781e-02 -2.80488163e-01
-1.46563148e+00 6.56630024e-02 4.01914448e-01 -2.66389608e-01
1.16837382e+00 6.05862260e-01 9.78148043e-01 3.24069619e-01
6.69256985e-01 3.98911238e-01 1.06221437e+00 -7.44644880e-01
5.53917170e-01 4.90788333e-02 1.17865510e-01 4.80289102e-01
7.31380343e-01 3.62602592e-01 -4.47600067e-01 -5.12833655e-01
5.66992104e-01 -3.98532599e-01 -1.14868671e-01 -5.92897117e-01
-1.01293814e+00 1.02791262e+00 -4.34634909e-02 3.82786095e-01
-2.28363827e-01 4.02541786e-01 4.70253736e-01 5.32646894e-01
4.20266926e-01 1.43054152e+00 -6.39593303e-01 -3.92642945e-01
-1.03303385e+00 5.72429717e-01 6.96900070e-01 1.00409985e+00
6.31264806e-01 3.46785784e-01 -2.19963625e-01 9.98177350e-01
-1.46845102e-01 3.54767174e-01 2.68628925e-01 -7.96990812e-01
3.02082479e-01 9.34047282e-01 -1.35603860e-01 -8.56798291e-01
-4.62352902e-01 -4.42979008e-01 -5.36708772e-01 6.10654831e-01
4.11042631e-01 -4.98223603e-01 -6.87224030e-01 1.54780757e+00
3.93051386e-01 -2.59024441e-01 -1.24651283e-01 4.59243149e-01
5.34503311e-02 8.11935246e-01 -3.47744048e-01 -5.66380560e-01
8.70375216e-01 -8.16270530e-01 -1.88328922e-01 1.96222544e-01
9.44403708e-01 -4.75233197e-01 8.87210846e-01 8.23640347e-01
-1.01694286e+00 -3.92067194e-01 -1.01890171e+00 6.46313071e-01
-2.31028169e-01 2.07800806e-01 8.09202909e-01 1.06730974e+00
-7.15347469e-01 9.12312388e-01 -6.49783909e-01 -1.80112496e-01
-6.95137903e-02 7.91453242e-01 -2.02217191e-01 1.78100076e-02
-8.50664973e-01 6.62728190e-01 7.97614813e-01 -2.67242759e-01
-1.59642234e-01 -1.05344081e+00 -8.23453844e-01 1.03504308e-01
8.86357367e-01 -2.84304947e-01 6.94890320e-01 -7.22098410e-01
-1.61314487e+00 2.33169436e-01 -1.75712481e-01 -3.40593636e-01
5.65949857e-01 3.79970759e-01 -3.11039090e-02 -2.45349392e-01
-2.80553754e-02 3.55029374e-01 2.70858049e-01 -9.65089262e-01
-7.38446891e-01 -2.35375732e-01 -1.36986509e-01 1.87510531e-02
1.77622661e-02 6.37560934e-02 5.87884188e-02 -7.14516580e-01
2.59756953e-01 -1.42647433e+00 -3.95371228e-01 -6.73935115e-01
-3.27500671e-01 -1.19897023e-01 5.99849403e-01 -1.49231672e-01
1.35576904e+00 -2.12775135e+00 5.58892250e-01 5.31407833e-01
-2.22101554e-01 2.69614965e-01 -3.07867173e-02 7.07181513e-01
-5.05558491e-01 1.29854485e-01 -3.49840194e-01 3.20687145e-01
-1.02962300e-01 -5.10348231e-02 -1.88825130e-02 3.29047441e-01
3.69440764e-01 3.47515970e-01 -1.11089575e+00 -1.57704398e-01
1.62151400e-02 4.89954725e-02 -1.02701843e+00 -4.32092786e-01
-4.31832820e-01 1.16425693e-01 -4.13574457e-01 8.11178148e-01
5.03930151e-01 3.57750267e-01 3.53162795e-01 4.76451784e-01
-4.32487160e-01 1.40839860e-01 -1.29859865e+00 1.09995615e+00
-3.82597834e-01 5.36394954e-01 -5.86107194e-01 -9.14897859e-01
1.43965662e+00 8.44927579e-02 5.63329101e-01 -5.68387628e-01
1.76990524e-01 5.55122554e-01 3.12436551e-01 3.63536999e-02
6.50288582e-01 -1.62374392e-01 -3.14805120e-01 5.38718760e-01
-5.35252750e-01 -4.82912123e-01 1.08127451e+00 -2.95671552e-01
9.77107584e-01 3.28987122e-01 6.37561977e-01 -6.62742853e-01
6.43661559e-01 2.53044724e-01 8.16087663e-01 7.07247019e-01
4.38999146e-01 5.07791579e-01 1.38546324e+00 -3.45679745e-02
-9.85874474e-01 -1.64114416e-01 -3.62003416e-01 9.87481296e-01
-2.40010326e-03 -5.38490534e-01 -4.29168314e-01 -6.79440081e-01
3.21503967e-01 1.20737422e+00 -6.73785746e-01 -4.94792104e-01
-8.10379982e-01 -1.16844356e+00 6.90223753e-01 3.39837819e-01
1.48431793e-01 -1.06437361e+00 -9.47027504e-01 1.93581268e-01
6.49761558e-01 -2.94714183e-01 9.18588880e-03 7.05225766e-01
-1.07527947e+00 -7.72912204e-01 -7.03071773e-01 -5.09399652e-01
7.74677455e-01 -1.98607147e-01 9.97269750e-01 3.98517340e-01
-1.87990770e-01 -3.46568644e-01 -6.50265515e-01 -2.98723638e-01
-6.76353693e-01 2.43881837e-01 -3.96601200e-01 -4.72630024e-01
2.57323086e-01 -4.97837961e-01 3.28514129e-01 4.50026453e-01
-5.15423000e-01 -6.48221225e-02 2.16361925e-01 1.16035259e+00
5.92805624e-01 5.00656724e-01 6.61581635e-01 -1.10826349e+00
6.19711399e-01 -4.11425859e-01 -1.08455408e+00 2.07035020e-01
-7.62970269e-01 5.53623140e-01 8.52826059e-01 -7.41565287e-01
-5.18832445e-01 1.64956376e-01 2.59635746e-01 -3.47090214e-02
1.30090162e-01 6.34524345e-01 1.19832717e-01 -3.18040848e-01
4.64137346e-01 -6.91030920e-02 -1.58294246e-01 -1.57320350e-01
-3.33986044e-01 1.79928258e-01 -2.13222370e-01 -7.47964919e-01
2.88748890e-01 -2.07395554e-01 5.94977438e-01 -8.85193884e-01
2.56686717e-01 1.72357097e-01 -1.70764059e-01 6.60755187e-02
3.25575829e-01 -1.54237747e-01 -6.56315625e-01 3.68420422e-01
-6.00552499e-01 -4.59884763e-01 -5.22987783e-01 5.01136184e-01
-6.95788682e-01 -4.49708290e-02 4.14766818e-02 -9.19946969e-01
2.18915090e-01 -1.51136112e+00 9.33233559e-01 9.40693319e-02
-8.42564344e-01 -6.36623144e-01 6.66843206e-02 -1.61801875e-01
1.31584421e-01 6.78335547e-01 1.32035959e+00 -6.74162269e-01
-1.75133303e-01 -1.10980153e-01 2.36410841e-01 -2.08469093e-01
-2.32239634e-01 3.74339044e-01 1.31074395e-02 -3.91430974e-01
-3.70121568e-01 -1.46271959e-01 6.72360361e-01 4.29296732e-01
6.99442744e-01 9.70836133e-02 -5.49770594e-01 6.23827159e-01
1.42785513e+00 8.56117249e-01 6.57755733e-01 7.41240680e-01
1.71183988e-01 5.51396608e-01 1.31228840e+00 6.09653175e-01
-4.31199253e-01 8.07873309e-01 1.52945086e-01 1.20922551e-01
9.69518051e-02 -1.59723330e-02 1.18117683e-01 -1.23381928e-01
-3.26897711e-01 -3.97866368e-01 -1.03286529e+00 4.75584775e-01
-1.57863450e+00 -7.34290957e-01 -2.92100310e-01 2.16658616e+00
9.89757836e-01 1.92393288e-01 4.57561046e-01 7.30343699e-01
8.54089618e-01 -1.54044241e-01 -4.58810031e-01 -9.47655737e-01
7.97006339e-02 4.10760134e-01 6.79853737e-01 5.03732204e-01
-5.00867724e-01 7.56476939e-01 5.80031538e+00 7.43370473e-01
-1.31998420e+00 -4.99310434e-01 3.31625134e-01 -3.27564389e-01
-4.38088715e-01 2.05284148e-01 -1.06638157e+00 5.22901356e-01
5.85148692e-01 -5.33373713e-01 4.55973178e-01 9.92685616e-01
-4.58753593e-02 -6.13825083e-01 -1.13863873e+00 3.41215521e-01
-7.40736350e-02 -9.88326609e-01 -3.00085932e-01 1.91891313e-01
1.10760212e+00 -7.01620519e-01 -1.59801826e-01 4.24971521e-01
2.59689420e-01 -9.63128924e-01 7.36090899e-01 -1.02290660e-01
6.56110406e-01 -1.49465621e+00 7.63952911e-01 3.03673804e-01
-8.99150610e-01 -1.81106567e-01 -1.50108039e-01 -3.24536525e-02
9.85968858e-02 7.85103798e-01 -9.58346725e-01 6.63537383e-01
4.57335949e-01 4.65864867e-01 -4.41817760e-01 1.23441637e+00
-1.40538782e-01 3.11432421e-01 -3.95683885e-01 -4.48518306e-01
1.18924184e-02 -4.09281880e-01 6.92613065e-01 9.67346549e-01
7.54102111e-01 -1.66240349e-01 -4.19518054e-02 8.89618397e-01
6.12981915e-01 -7.34770764e-03 -4.30096745e-01 -2.29479387e-01
5.38850904e-01 2.64514774e-01 -1.01060665e+00 -9.74686295e-02
2.38843873e-01 2.57225692e-01 -2.66597033e-01 1.84933946e-01
-9.48404372e-01 -9.63045239e-01 2.09050104e-01 4.29312974e-01
5.84557533e-01 -2.65900157e-02 -5.27234077e-01 -5.52947104e-01
-3.03349704e-01 -1.46072590e+00 9.11034644e-02 -4.88792002e-01
-4.61906344e-01 8.22827518e-01 6.97817206e-01 -7.99843669e-01
-6.55206025e-01 -2.64939576e-01 -4.22743171e-01 8.33558559e-01
-7.78164923e-01 -4.14682925e-01 5.04324138e-02 -7.46401474e-02
5.50184846e-01 -4.65512902e-01 6.46812379e-01 -4.75747734e-02
-6.71603084e-01 6.01569295e-01 1.40150025e-01 -7.38111317e-01
3.16130072e-01 -8.80867422e-01 2.21915722e-01 7.56174743e-01
-6.05001688e-01 8.10085177e-01 1.07187986e+00 -1.14370096e+00
-1.38895357e+00 -8.22417200e-01 8.28656614e-01 3.33661199e-01
3.68506014e-01 -2.25965947e-01 -7.02570260e-01 4.37541425e-01
-1.93527713e-02 -4.33033913e-01 1.75874263e-01 2.21295416e-01
1.08977057e-01 -8.15493613e-02 -1.41963053e+00 6.80286169e-01
8.73683393e-01 2.21303269e-01 -3.24911445e-01 -1.48464122e-03
4.00269687e-01 -4.51497346e-01 -7.59431124e-01 7.60260046e-01
4.68318731e-01 -1.21350551e+00 7.36010373e-01 -9.94970948e-02
6.30449414e-01 -3.92816544e-01 -1.23487413e-02 -1.53283620e+00
-3.92455816e-01 -7.28732526e-01 5.58416188e-01 1.07190013e+00
7.09201813e-01 -1.15737343e+00 9.03489888e-01 2.07231760e-01
-5.76097667e-02 -1.03968048e+00 -6.78285360e-01 -1.06253707e+00
2.13639960e-01 2.18604177e-01 7.87443638e-01 7.86995232e-01
2.51035452e-01 -2.52229515e-02 -2.60872543e-01 -2.69388676e-01
3.22076946e-01 3.25770289e-01 7.75584340e-01 -1.17774761e+00
-7.70547271e-01 -5.45888305e-01 -4.27776128e-01 -1.87523812e-01
6.50127903e-02 -7.84428418e-01 1.07702473e-02 -1.09488857e+00
5.96641973e-02 -7.95367718e-01 1.91320062e-01 4.14182097e-01
-1.58506572e-01 1.40830427e-01 1.90785691e-01 -4.95271891e-01
2.28563339e-01 2.89620131e-01 1.04699636e+00 1.16048865e-01
-8.05364788e-01 -7.57612241e-03 -5.18824637e-01 5.87710321e-01
8.13642442e-01 -1.03737485e+00 -7.02881753e-01 8.37372467e-02
7.99584091e-01 3.58190864e-01 -2.92715400e-01 -9.47625816e-01
-2.28485048e-01 -6.44320846e-01 -6.34485409e-02 -6.86513484e-01
1.65513754e-01 -4.19340581e-01 7.19523609e-01 9.30619180e-01
-5.25652885e-01 2.31371239e-01 5.29031277e-01 1.26454443e-01
-7.83289671e-02 -9.04836714e-01 6.96933150e-01 2.75022853e-02
-3.41062486e-01 -6.33493364e-01 -5.97254813e-01 -4.53888737e-02
1.50486982e+00 -6.19082808e-01 -7.52343088e-02 2.24367276e-01
-4.83652532e-01 2.23598346e-01 8.74818683e-01 6.61989450e-02
3.11644614e-01 -9.08264756e-01 -7.22347558e-01 7.04358757e-01
-2.75718898e-01 -2.80302882e-01 -3.91244352e-01 9.05195236e-01
-6.52492464e-01 5.32819569e-01 -4.89115298e-01 -6.55603230e-01
-1.81934607e+00 3.66615236e-01 -3.78412046e-02 -3.35653633e-01
-3.94212157e-01 9.73677993e-01 -4.32347834e-01 -5.08325458e-01
1.26657486e-01 -5.63022852e-01 -5.74515276e-02 1.59046143e-01
7.26132542e-02 7.93676913e-01 7.78543949e-02 -7.23085925e-02
-5.70034027e-01 8.01077485e-01 1.26078531e-01 -1.64892912e-01
1.62138915e+00 5.41377723e-01 -5.95781565e-01 3.53855282e-01
7.87480772e-01 3.18744659e-01 -5.54478586e-01 7.20110238e-01
-5.93508482e-02 -8.01994324e-01 -1.23127095e-01 -5.51768601e-01
-8.95633459e-01 2.47458383e-01 -7.69475922e-02 1.93573803e-01
1.22290468e+00 -4.20667738e-01 2.77824640e-01 3.26660156e-01
7.38281250e-01 -8.74231279e-01 -2.39977628e-01 6.90107942e-01
9.97291625e-01 -4.85703140e-01 4.53870207e-01 -9.22191739e-01
-6.50064766e-01 1.16029918e+00 6.04303062e-01 -5.18040061e-01
2.69208014e-01 5.52005768e-01 -8.18953514e-01 1.17960749e-02
-8.79551470e-01 1.21979155e-01 -2.46097464e-02 3.33901852e-01
6.32913470e-01 -5.51863164e-02 -9.67536449e-01 2.14410752e-01
-5.72618663e-01 2.15236738e-01 7.39895105e-01 1.43242276e+00
-4.26625580e-01 -1.50450408e+00 -7.14854598e-01 4.20285612e-01
1.00958504e-01 -1.71434805e-02 -6.53875887e-01 1.04605496e+00
4.33117419e-01 7.63203800e-01 6.84634447e-02 -3.27650458e-01
3.24882090e-01 -6.07489012e-02 7.84541905e-01 -7.55439222e-01
-1.20176363e+00 2.12421283e-01 6.33424401e-01 -2.29617953e-01
4.34287786e-02 -9.05311465e-01 -1.46693885e+00 -3.88898730e-01
-6.59663916e-01 4.86765593e-01 6.37888908e-01 6.77503645e-01
3.24641526e-01 9.16028559e-01 4.68457520e-01 -6.48781598e-01
-4.37176704e-01 -4.39231724e-01 -6.07929409e-01 -3.08905303e-01
2.03917585e-02 -7.88826883e-01 -4.32634234e-01 -6.00615799e-01] | [8.026570320129395, 7.204242706298828] |
4e7d81a3-e648-455a-a6e3-4c84044dc13b | wisdom-of-the-contexts-active-ensemble | 2101.11560 | null | https://arxiv.org/abs/2101.11560v4 | https://arxiv.org/pdf/2101.11560v4.pdf | Wisdom of the Contexts: Active Ensemble Learning for Contextual Anomaly Detection | In contextual anomaly detection, an object is only considered anomalous within a specific context. Most existing methods for CAD use a single context based on a set of user-specified contextual features. However, identifying the right context can be very challenging in practice, especially in datasets, with a large number of attributes. Furthermore, in real-world systems, there might be multiple anomalies that occur in different contexts and, therefore, require a combination of several "useful" contexts to unveil them. In this work, we leverage active learning and ensembles to effectively detect complex contextual anomalies in situations where the true contextual and behavioral attributes are unknown. We propose a novel approach, called WisCon (Wisdom of the Contexts), that automatically creates contexts from the feature set. Our method constructs an ensemble of multiple contexts, with varying importance scores, based on the assumption that not all useful contexts are equally so. Experiments show that WisCon significantly outperforms existing baselines in different categories (i.e., active classifiers, unsupervised contextual and non-contextual anomaly detectors, and supervised classifiers) on seven datasets. Furthermore, the results support our initial hypothesis that there is no single perfect context that successfully uncovers all kinds of contextual anomalies, and leveraging the "wisdom" of multiple contexts is necessary. | ['Onur Dikmen', 'Mohamed-Rafik Bouguelia', 'Slawomir Nowaczyk', 'Ece Calikus'] | 2021-01-27 | null | null | null | null | ['contextual-anomaly-detection'] | ['miscellaneous'] | [ 3.20799142e-01 -3.54842544e-01 -1.01462193e-01 -5.94369590e-01
-6.91047549e-01 -4.69250351e-01 5.24904072e-01 6.79184437e-01
-3.77803296e-03 3.94565850e-01 1.32068262e-01 -4.58236665e-01
-2.46558398e-01 -7.77773857e-01 -4.60160792e-01 -6.92467690e-01
-4.94927727e-03 2.02233464e-01 6.72367036e-01 5.40309474e-02
5.38968503e-01 3.53741884e-01 -1.87003303e+00 4.36983973e-01
1.08291245e+00 1.01859164e+00 -1.96601510e-01 4.95288610e-01
-4.00544673e-01 1.04520929e+00 -7.92034924e-01 -2.06688285e-01
5.58737926e-02 -4.39950138e-01 -5.11888087e-01 -8.59067738e-02
6.00890458e-01 -1.87499225e-01 2.52974302e-01 7.85057962e-01
9.48249474e-02 1.66129157e-01 5.24791002e-01 -1.44322610e+00
-4.73702908e-01 3.88498962e-01 -4.34947520e-01 6.23210609e-01
4.51457739e-01 1.96704417e-01 1.25398624e+00 -9.96767998e-01
2.22447619e-01 9.27072167e-01 7.16611266e-01 2.97367752e-01
-1.17701435e+00 -6.06856406e-01 8.19947481e-01 4.94588524e-01
-1.16920328e+00 -3.62999231e-01 8.62353206e-01 -4.72553730e-01
8.83787632e-01 4.94952828e-01 5.49958229e-01 1.36741555e+00
2.75223851e-01 6.05719566e-01 8.06518555e-01 -5.43534636e-01
7.11940169e-01 -3.09468582e-02 4.71027136e-01 9.88091528e-02
6.72705054e-01 -5.00099324e-02 -6.09957397e-01 -9.32472587e-01
1.02130279e-01 4.81684566e-01 1.79224446e-01 -3.92154485e-01
-8.94591153e-01 6.92304552e-01 -1.80508733e-01 3.79401922e-01
-3.64899993e-01 -2.29322731e-01 3.51972699e-01 2.40969300e-01
4.46856558e-01 6.78989768e-01 -5.62276721e-01 -1.29348338e-01
-4.28685844e-01 2.25558415e-01 6.18576944e-01 6.65819168e-01
7.86013007e-01 -8.70290771e-02 -7.70278350e-02 7.57373810e-01
1.78429767e-01 2.18214273e-01 3.20896029e-01 -5.74202299e-01
1.96796507e-01 1.21342301e+00 8.30533430e-02 -1.11974621e+00
-1.50877863e-01 -1.42906889e-01 -5.71382761e-01 5.27691096e-04
2.33896166e-01 -6.03955947e-02 -7.82058895e-01 1.56276548e+00
5.17378032e-01 8.97415340e-01 -1.39836311e-01 5.05119443e-01
4.76569414e-01 1.15141675e-01 3.26431870e-01 -1.13433048e-01
7.85030007e-01 -3.64729792e-01 -6.65647388e-01 -5.18718719e-01
6.03418529e-01 -6.21733308e-01 1.17113554e+00 5.20339787e-01
-3.29431683e-01 -2.16292039e-01 -1.09046161e+00 6.03985608e-01
-5.64469695e-01 -2.85784841e-01 7.43308544e-01 5.72210670e-01
-5.46826482e-01 3.18715870e-01 -8.58712435e-01 -2.84238040e-01
3.66348982e-01 1.20961228e-02 -1.94698438e-01 -3.18293944e-02
-8.94216061e-01 6.27039075e-01 3.88623297e-01 6.46106079e-02
-7.47850180e-01 -5.35896599e-01 -7.16273665e-01 -7.42412060e-02
9.60242331e-01 -1.80466771e-01 9.91951585e-01 -9.75132108e-01
-5.03517509e-01 3.61573100e-01 -4.20871586e-01 -2.94849753e-01
1.82592824e-01 -6.69712603e-01 -9.47470427e-01 -6.01342656e-02
2.04821855e-01 -3.04772168e-01 7.98769593e-01 -1.34607244e+00
-9.91547942e-01 -5.02327800e-01 -1.74264535e-02 -2.37841159e-01
-3.32918078e-01 2.01851353e-01 -2.23716214e-01 -6.34587109e-01
3.63549531e-01 -9.15320218e-01 -4.63362098e-01 -4.38453376e-01
-5.77001810e-01 -2.77002454e-01 1.31215239e+00 -2.78868139e-01
1.75284410e+00 -2.33594322e+00 -5.47101319e-01 6.39718056e-01
4.12695825e-01 1.19575739e-01 3.23707700e-01 4.01832044e-01
-1.99943632e-01 2.84142703e-01 -4.82595205e-01 -3.05472333e-02
-3.08349550e-01 6.11404061e-01 -7.59006500e-01 5.43686822e-02
4.67096925e-01 4.66337740e-01 -1.17837620e+00 -4.06223416e-01
1.32047921e-01 -8.41313154e-02 -4.98094857e-01 2.39752531e-01
-2.30714202e-01 5.10718942e-01 -5.25320590e-01 9.10004616e-01
3.09967339e-01 -1.74728051e-01 2.97723114e-01 5.45543194e-01
1.39319360e-01 1.16527438e-01 -1.52967620e+00 9.79739308e-01
-1.60512328e-01 4.64104474e-01 -6.19244456e-01 -1.19159842e+00
8.67301643e-01 3.09479713e-01 7.77163506e-01 -4.82382655e-01
-4.46863264e-01 3.39493424e-01 -7.00362772e-02 -6.86825633e-01
3.27828318e-01 4.96691048e-01 -2.41052926e-01 5.97502053e-01
-1.86973423e-01 5.37437260e-01 1.62047207e-01 2.34806970e-01
1.77314448e+00 -1.86102316e-01 6.06141865e-01 6.79073483e-02
5.19127965e-01 -1.72108993e-01 1.21039605e+00 1.17979956e+00
-3.54505926e-01 7.05153942e-01 9.64058518e-01 -5.99110782e-01
-7.77020693e-01 -1.10276163e+00 -9.66588184e-02 1.16601276e+00
1.04372293e-01 -8.15725625e-01 -2.18263090e-01 -1.27807093e+00
8.57467800e-02 8.60885501e-01 -6.83292925e-01 -3.24924409e-01
-7.40574479e-01 -9.40991342e-01 3.48427087e-01 6.60898805e-01
1.70710385e-01 -1.02171457e+00 -7.43744373e-01 2.31431663e-01
-1.26523122e-01 -1.08297229e+00 -6.52622581e-02 2.01609701e-01
-8.68514359e-01 -1.57232261e+00 4.74506289e-01 -1.90223455e-01
7.16472328e-01 2.32295603e-01 1.41664672e+00 4.61442173e-01
-2.68552780e-01 5.38114667e-01 -7.53875852e-01 -9.04102206e-01
-3.17073226e-01 -2.56002456e-01 4.79110144e-02 2.76997447e-01
9.21480298e-01 -7.27606475e-01 -4.39622760e-01 4.90857154e-01
-9.08144116e-01 -6.63649321e-01 5.30950665e-01 8.71487379e-01
6.93896115e-01 1.95460051e-01 9.77219999e-01 -1.38556075e+00
5.25760293e-01 -8.87931108e-01 -2.59903252e-01 4.57469553e-01
-8.64323258e-01 -1.73991129e-01 5.94603956e-01 -6.53629065e-01
-9.87893701e-01 6.83865175e-02 2.11409211e-01 -3.47277105e-01
-6.29655004e-01 3.05754840e-01 -3.82770419e-01 3.49625677e-01
1.00627708e+00 -7.17278272e-02 -5.34902096e-01 -3.51470470e-01
1.53940823e-03 4.81701612e-01 3.04944873e-01 -8.07595789e-01
5.95219374e-01 3.06562901e-01 -1.62261412e-01 -7.08135188e-01
-9.49024677e-01 -7.67349064e-01 -6.66909695e-01 -1.98800057e-01
4.15220201e-01 -6.08930171e-01 -6.96213990e-02 4.30601060e-01
-7.48350739e-01 1.85059756e-03 -4.03853059e-01 3.21991235e-01
-6.21293299e-02 3.93997699e-01 -1.93914194e-02 -1.02767336e+00
-8.46045986e-02 -9.63618159e-01 8.83036613e-01 7.27446601e-02
-7.33642757e-01 -8.19607973e-01 1.49329036e-01 -1.55125288e-02
3.14079225e-01 6.58747196e-01 9.71132874e-01 -1.26104569e+00
-4.71614063e-01 -4.07342046e-01 2.05875337e-01 2.00080961e-01
4.78297025e-01 3.81838292e-01 -1.14909494e+00 5.22368550e-02
-4.90397476e-02 1.07798927e-01 7.91728139e-01 1.06002018e-02
1.47981024e+00 -3.70729536e-01 -4.03158337e-01 -3.43249403e-02
1.23786187e+00 7.12445974e-01 5.05985379e-01 1.89376324e-01
7.05848575e-01 3.99395049e-01 6.88773334e-01 4.87654179e-01
1.86216474e-01 3.85418147e-01 4.33805585e-01 1.53308183e-01
5.14674067e-01 -2.61416268e-02 2.61788875e-01 3.84230316e-01
-1.64364800e-02 4.44026031e-02 -1.37018108e+00 7.53907323e-01
-2.15572548e+00 -1.07824004e+00 -1.40360475e-01 2.51438236e+00
6.68432653e-01 6.05342329e-01 7.00541958e-02 4.25338894e-01
7.38462806e-01 -1.23336412e-01 -8.65767956e-01 -3.81217957e-01
-9.42094773e-02 2.17035800e-01 -4.66765687e-02 1.00731201e-01
-1.33996379e+00 4.15985137e-01 6.25667191e+00 2.82292008e-01
-9.20891464e-01 -1.11182481e-02 5.40606737e-01 -1.17382720e-01
-2.31227487e-01 4.04219329e-01 -6.72722161e-01 7.46534109e-01
8.51223648e-01 1.30767059e-02 -2.50475466e-01 1.15799928e+00
-1.63395524e-01 -3.84745985e-01 -1.34432435e+00 7.25991249e-01
8.91961679e-02 -9.24982727e-01 -8.48829001e-02 1.84234425e-01
7.69504428e-01 -2.68588185e-01 -1.47148818e-01 3.53041232e-01
3.41769814e-01 -6.11920357e-01 3.77471864e-01 6.96531475e-01
1.98438197e-01 -7.70269632e-01 7.77764916e-01 3.74834418e-01
-1.03043675e+00 -6.80565536e-01 3.85845676e-02 1.98919587e-02
-3.54306340e-01 1.08792448e+00 -8.04133296e-01 5.07201672e-01
8.89175653e-01 7.06556082e-01 -1.06687200e+00 1.17364359e+00
-2.26684570e-01 1.10774350e+00 -3.43037248e-01 2.34431311e-01
-1.17942438e-01 5.83413132e-02 7.27403879e-01 9.50518727e-01
2.59524405e-01 9.13837329e-02 3.51623029e-01 5.30959189e-01
4.12921786e-01 1.56107888e-01 -8.53338540e-01 2.72447556e-01
8.34283352e-01 9.71536398e-01 -7.32445657e-01 -3.36376518e-01
-6.94904685e-01 3.80197287e-01 6.63709715e-02 3.04402262e-01
-6.04394674e-01 -2.28858531e-01 5.59282184e-01 -5.30050974e-03
9.32123661e-02 6.42125495e-04 -5.40095091e-01 -1.29397750e+00
4.56745028e-01 -8.74787569e-01 8.69820952e-01 -3.09716344e-01
-1.61503243e+00 2.57855296e-01 1.26098230e-01 -1.50234628e+00
-4.45837915e-01 -3.21123719e-01 -1.24295926e+00 3.31933320e-01
-9.79114354e-01 -8.76592278e-01 -4.49499011e-01 6.34503841e-01
3.76006663e-01 -2.19536424e-01 8.99222970e-01 2.29809806e-01
-8.43498409e-01 5.71830869e-01 -1.08804658e-01 2.61291444e-01
9.39893365e-01 -1.54677749e+00 4.24040586e-01 1.36731422e+00
3.46870691e-01 8.24961305e-01 6.14109039e-01 -8.66184115e-01
-1.01325238e+00 -1.05638802e+00 7.94038057e-01 -8.76547992e-01
6.58061266e-01 -3.48572314e-01 -1.39686716e+00 8.66376936e-01
-2.01858863e-01 2.09795803e-01 1.16536236e+00 7.05533683e-01
-4.52555239e-01 -2.42470562e-01 -1.07114649e+00 5.38444161e-01
1.14858019e+00 -3.88240546e-01 -7.83268631e-01 6.70142919e-02
5.55944204e-01 -1.50865853e-01 -5.11457622e-01 7.45580435e-01
4.20252502e-01 -1.11206937e+00 8.32139075e-01 -8.26615453e-01
2.04765499e-01 -2.40907028e-01 -3.72371256e-01 -1.11460531e+00
1.93191711e-02 -2.27455094e-01 -6.88295126e-01 1.37396824e+00
5.04265666e-01 -9.73780811e-01 5.98928928e-01 9.27586079e-01
-1.56696245e-01 -8.77677560e-01 -9.09058809e-01 -6.05840325e-01
-3.92449290e-01 -9.93732750e-01 1.01563764e+00 1.39762688e+00
-1.02246553e-01 2.00129732e-01 -1.11845382e-01 4.41192716e-01
4.74927753e-01 9.82086435e-02 8.07159066e-01 -1.80013633e+00
-8.18035677e-02 -1.76849455e-01 -6.76648855e-01 -1.31659061e-01
4.50623147e-02 -3.24028820e-01 -6.82924613e-02 -1.00611854e+00
2.85324082e-02 -8.84375930e-01 -7.53408611e-01 7.97256410e-01
-7.62230396e-01 -2.88143277e-01 -3.53888452e-01 2.54463255e-01
-8.00454021e-01 -3.75523716e-02 4.25818980e-01 6.58342242e-02
-4.94604409e-01 2.87467837e-01 -9.19400752e-01 1.01439381e+00
8.69898081e-01 -4.35642689e-01 -5.41362405e-01 4.32424843e-02
4.63089108e-01 -4.41036582e-01 5.32115459e-01 -1.09828973e+00
2.87531435e-01 -5.55161655e-01 5.87842822e-01 -7.62574136e-01
3.20646837e-02 -9.91312683e-01 -2.14347839e-02 -1.50872558e-01
-3.10981631e-01 2.43947864e-01 -4.75360546e-04 9.18867707e-01
-2.69185066e-01 -9.24601331e-02 4.10950482e-01 -1.17864043e-01
-1.09957111e+00 5.05641252e-02 -2.37057611e-01 1.85156018e-01
1.25224710e+00 -2.02833205e-01 -3.66797626e-01 -1.20555170e-01
-8.23499441e-01 2.35775843e-01 2.64562547e-01 6.22835696e-01
6.96551144e-01 -1.39502609e+00 -5.09361148e-01 3.84803176e-01
6.50427103e-01 2.45979026e-01 2.23093837e-01 6.20474935e-01
8.47662985e-02 2.97252685e-02 2.26167470e-01 -9.40702915e-01
-1.32088172e+00 5.88161230e-01 1.54627264e-01 -9.93768871e-02
-7.87759662e-01 3.43979985e-01 4.58079055e-02 -2.88995266e-01
2.20949173e-01 -2.98874319e-01 -3.18538994e-01 2.31888011e-01
7.66719580e-01 3.35344672e-01 4.74108636e-01 -4.88813758e-01
-4.28285271e-01 2.49268323e-01 -3.08497548e-01 2.89725333e-01
1.18923211e+00 9.15084854e-02 -4.72441763e-02 8.81724238e-01
5.28381884e-01 2.87673175e-01 -8.87050331e-01 -6.76135242e-01
6.31594241e-01 -7.76445866e-01 -3.78257900e-01 -7.04433858e-01
-8.33697140e-01 5.13374567e-01 5.29214501e-01 4.61933285e-01
1.45219588e+00 1.13561824e-02 4.29758430e-01 5.48681021e-01
2.84204513e-01 -1.28634000e+00 2.75075972e-01 4.15079474e-01
5.64496577e-01 -1.46429133e+00 7.12743998e-02 -3.23321968e-01
-7.82409251e-01 8.34364474e-01 1.14307487e+00 -2.68495679e-02
7.88594723e-01 2.28419363e-01 -2.00135261e-03 -1.33447260e-01
-1.12473643e+00 -9.60344300e-02 3.49985749e-01 4.67167407e-01
2.22950563e-01 1.58506051e-01 -7.29170591e-02 6.18996263e-01
1.57077722e-02 -4.76914793e-01 5.36055505e-01 1.42357707e+00
-4.24008489e-01 -1.04564166e+00 -6.14902258e-01 9.76968408e-01
-6.98070407e-01 2.20231295e-01 -6.94050014e-01 5.81843376e-01
5.25629640e-01 1.28901589e+00 2.70667762e-01 -5.50402403e-01
5.74803352e-01 4.79160547e-01 -2.28356898e-01 -8.54761362e-01
-4.74901497e-01 -2.00970009e-01 5.44752926e-02 -7.37195492e-01
-2.33832404e-01 -8.50663245e-01 -1.02207160e+00 7.15917870e-02
-4.12632048e-01 -7.77702853e-02 2.10863307e-01 1.23952866e+00
4.76538688e-01 3.76180857e-01 9.00017262e-01 -9.34521928e-02
-1.83773816e-01 -7.62535989e-01 -2.29551464e-01 7.11414456e-01
4.83698398e-01 -8.82224739e-01 -5.30790925e-01 -7.18957037e-02] | [7.614894390106201, 2.5221176147460938] |
832ef3a9-77c3-448d-8285-f240344066ff | pasts-progress-aware-spatio-temporal | 2305.11918 | null | https://arxiv.org/abs/2305.11918v1 | https://arxiv.org/pdf/2305.11918v1.pdf | PASTS: Progress-Aware Spatio-Temporal Transformer Speaker For Vision-and-Language Navigation | Vision-and-language navigation (VLN) is a crucial but challenging cross-modal navigation task. One powerful technique to enhance the generalization performance in VLN is the use of an independent speaker model to provide pseudo instructions for data augmentation. However, current speaker models based on Long-Short Term Memory (LSTM) lack the ability to attend to features relevant at different locations and time steps. To address this, we propose a novel progress-aware spatio-temporal transformer speaker (PASTS) model that uses the transformer as the core of the network. PASTS uses a spatio-temporal encoder to fuse panoramic representations and encode intermediate connections through steps. Besides, to avoid the misalignment problem that could result in incorrect supervision, a speaker progress monitor (SPM) is proposed to enable the model to estimate the progress of instruction generation and facilitate more fine-grained caption results. Additionally, a multifeature dropout (MFD) strategy is introduced to alleviate overfitting. The proposed PASTS is flexible to be combined with existing VLN models. The experimental results demonstrate that PASTS outperforms all existing speaker models and successfully improves the performance of previous VLN models, achieving state-of-the-art performance on the standard Room-to-Room (R2R) dataset. | ['Qijun Chen', 'Huiyi Chen', 'Qingqing Yan', 'Shu Li', 'Zongtao He', 'Chengju Liu', 'Liuyi Wang'] | 2023-05-19 | null | null | null | null | ['vision-and-language-navigation'] | ['robots'] | [ 2.26844206e-01 -2.94319242e-01 -1.12694524e-01 -7.27479517e-01
-8.29803348e-01 -9.44130719e-02 6.99641168e-01 -2.64575452e-01
-5.63137949e-01 3.74370456e-01 5.53042948e-01 -5.67008972e-01
1.83378980e-01 -4.67761010e-01 -1.03332782e+00 -5.86196363e-01
2.90875942e-01 -2.08156332e-02 2.64213175e-01 -3.84781420e-01
2.36220956e-01 1.01419777e-01 -1.62509239e+00 3.60266417e-01
1.14935231e+00 7.85848379e-01 8.83347750e-01 4.28762853e-01
-6.12671912e-01 8.34668636e-01 -5.61217487e-01 -9.47210193e-03
-1.36166796e-01 -3.54334503e-01 -4.44614619e-01 -2.43175849e-01
7.46855199e-01 -3.45312595e-01 -5.57037711e-01 8.63175154e-01
6.38567805e-01 6.49716556e-01 3.90109301e-01 -1.08169889e+00
-9.87962484e-01 6.92432046e-01 -5.72896361e-01 2.19788566e-01
2.08308950e-01 1.50652170e-01 6.36457980e-01 -1.16973996e+00
1.65202795e-03 1.54184389e+00 4.56746370e-01 8.68438065e-01
-1.02378547e+00 -8.92807662e-01 8.07841480e-01 4.42842305e-01
-1.21374249e+00 -7.60126412e-01 7.77772307e-01 -1.12278365e-01
8.27137351e-01 5.57896076e-03 2.24570349e-01 1.63534343e+00
2.22597167e-01 1.05395198e+00 8.64523768e-01 -3.53989542e-01
6.19846657e-02 -1.74459554e-02 1.74073368e-01 8.30287874e-01
-4.40898031e-01 6.25607818e-02 -1.09375393e+00 3.52510273e-01
8.03433120e-01 1.43133894e-01 -4.85125124e-01 -3.08273762e-01
-1.35623181e+00 6.67321563e-01 7.98026204e-01 1.69165194e-01
-2.24913612e-01 1.25204651e-02 4.56114113e-01 -4.58394475e-02
1.17352374e-01 6.49227425e-02 -3.34151447e-01 -1.39910147e-01
-9.40709710e-01 -1.99938789e-01 2.14352220e-01 9.63271439e-01
4.95368809e-01 5.13971090e-01 -5.62048554e-01 9.69606519e-01
6.27862036e-01 4.61587012e-01 9.39099371e-01 -8.07385504e-01
9.47262526e-01 3.96393538e-01 -2.56165504e-01 -5.50192714e-01
-2.24977598e-01 -5.32548368e-01 -8.58914793e-01 1.19261116e-01
1.52465001e-01 3.10203671e-01 -1.37538910e+00 1.96159601e+00
7.44741410e-02 4.45734322e-01 1.33314565e-01 9.12924647e-01
1.12014544e+00 8.89845252e-01 1.50926605e-01 4.46278378e-02
1.24753249e+00 -1.55638802e+00 -1.05846918e+00 -7.30796754e-01
4.37972218e-01 -5.72562218e-01 1.65186524e+00 2.13267654e-01
-7.48964250e-01 -1.02295506e+00 -1.17268133e+00 -3.16654712e-01
-3.69594991e-01 2.00865358e-01 4.05311853e-01 4.88214701e-01
-1.15352750e+00 1.26243457e-01 -1.02518392e+00 -5.94804175e-02
3.01281095e-01 2.36487240e-01 -3.05420309e-01 -3.54728490e-01
-1.24142325e+00 8.10955167e-01 1.01610571e-01 6.15416288e-01
-1.08046091e+00 -5.75338483e-01 -1.20484793e+00 -6.42085671e-02
2.22645700e-01 -5.66016197e-01 1.21498597e+00 -6.29139483e-01
-1.85814190e+00 3.24416518e-01 -8.01039696e-01 -2.95107514e-01
1.22849710e-01 -3.60987514e-01 -5.14462531e-01 -3.43137801e-01
8.30665007e-02 8.97725284e-01 7.75532544e-01 -1.20279062e+00
-4.34568971e-01 -5.41181087e-01 -1.84106842e-01 4.96744663e-01
-4.99543518e-01 -2.14034453e-01 -6.35236740e-01 -6.54847026e-01
2.98747480e-01 -7.42601633e-01 -2.03900024e-01 -3.00549030e-01
-4.29303855e-01 -2.59419948e-01 8.35253179e-01 -8.95706952e-01
1.33655047e+00 -2.13207197e+00 3.16603273e-01 -2.12087452e-01
-1.32030010e-01 2.74692804e-01 -5.08708119e-01 2.72542797e-02
2.61636972e-02 -2.77675956e-01 -1.72473118e-01 -8.69570673e-01
-1.16978645e-01 2.00092271e-01 -3.44048649e-01 1.48801759e-01
-3.93474214e-02 8.32884967e-01 -5.86679280e-01 -3.11175287e-01
3.70340258e-01 7.64070094e-01 -3.99109483e-01 2.50369519e-01
-1.75770730e-01 8.18986237e-01 -2.50857532e-01 4.96196538e-01
5.16804874e-01 -4.52449508e-02 -4.67470437e-01 -9.21480954e-02
-3.10046732e-01 6.87075019e-01 -9.85382915e-01 2.42585731e+00
-9.46629941e-01 6.27132714e-01 1.23835437e-01 -7.64254093e-01
1.04982221e+00 2.99839109e-01 -1.54081658e-01 -1.07937813e+00
-6.81764558e-02 7.72263482e-02 -1.64523855e-01 -4.35623825e-01
6.43840075e-01 2.23002121e-01 -9.55293700e-03 1.51766494e-01
1.39018614e-02 4.27315652e-01 -3.52462977e-01 5.01349084e-02
8.23937058e-01 3.15300345e-01 -2.46075690e-01 3.31968553e-02
8.72282207e-01 -4.35319275e-01 6.16301954e-01 7.26531625e-01
-3.96794617e-01 4.93775636e-01 -2.17013925e-01 -1.46338567e-01
-4.12202567e-01 -1.09401023e+00 1.05684511e-01 1.59423733e+00
6.96893036e-02 -1.72725692e-01 -6.59797013e-01 -7.24546075e-01
-3.91817480e-01 1.21084714e+00 -6.11978590e-01 -4.17664856e-01
-6.85696542e-01 -4.29518551e-01 4.04680163e-01 8.23899925e-01
6.56195104e-01 -1.16536319e+00 -2.39255309e-01 2.35757127e-01
-7.33004510e-01 -1.08242011e+00 -9.28658247e-01 4.05130237e-01
-8.41933489e-01 -5.40336907e-01 -7.33803809e-01 -9.78362203e-01
5.37204266e-01 4.73569691e-01 7.11198807e-01 -5.08526266e-01
1.82192609e-01 2.02277407e-01 -1.65403903e-01 -1.86005443e-01
-2.83076793e-01 1.91522121e-01 1.48933142e-01 2.02949578e-03
5.93825638e-01 -3.58825594e-01 -6.10438585e-01 3.34409326e-01
-6.07866526e-01 2.11529583e-01 7.18092322e-01 1.03512692e+00
6.05658948e-01 -2.90070146e-01 6.85148180e-01 -3.12365651e-01
5.58413029e-01 -1.97465047e-01 -5.09497225e-01 3.60518575e-01
-5.40695965e-01 4.12457645e-01 6.85773849e-01 -5.90987742e-01
-1.40234983e+00 7.20536783e-02 -3.18114012e-01 -6.75484717e-01
-2.74895281e-01 3.80196244e-01 -5.46158373e-01 7.24214539e-02
3.32365513e-01 7.84837723e-01 -5.77325234e-03 -6.42063320e-01
4.50607896e-01 7.40567327e-01 7.50049114e-01 -2.86441505e-01
4.26099002e-01 1.20455876e-01 -2.98999071e-01 -7.63047934e-01
-8.23258996e-01 -3.86486709e-01 -5.15452206e-01 -1.48051113e-01
9.93827283e-01 -1.22052693e+00 -6.22056067e-01 4.96459484e-01
-1.18235505e+00 -5.71616709e-01 1.22347571e-01 7.18141854e-01
-4.24275100e-01 3.70147787e-02 -6.09068215e-01 -7.43475676e-01
-2.43736401e-01 -1.62305295e+00 9.09078956e-01 4.34779406e-01
3.71514522e-02 -8.52939963e-01 -1.04572602e-01 7.73377657e-01
7.35291243e-01 -4.54916596e-01 6.70597494e-01 -4.52974170e-01
-6.60267711e-01 1.05807699e-01 -7.77078327e-03 3.16340476e-01
2.80209720e-01 -3.98752302e-01 -1.16682971e+00 -2.25903839e-01
2.63180554e-01 -2.90726453e-01 1.26484370e+00 6.08560264e-01
1.34083772e+00 -1.56529307e-01 -2.64532536e-01 7.66949236e-01
1.03050065e+00 1.74042478e-01 5.78378260e-01 5.78055978e-01
1.24021506e+00 6.21885538e-01 5.43640256e-01 -5.25830984e-02
8.55094850e-01 7.36801207e-01 5.56152463e-01 1.02380298e-01
-2.92361438e-01 -5.07091880e-01 9.17807639e-01 1.11226046e+00
2.45347768e-01 -1.31873056e-01 -7.70888984e-01 4.45516139e-01
-1.68877661e+00 -7.01442420e-01 -1.20411828e-01 2.36323929e+00
8.51499438e-01 1.65675491e-01 -3.82902712e-01 -1.70081034e-01
6.06599867e-01 4.15647864e-01 -6.70686901e-01 -3.78022462e-01
6.69815019e-02 -1.13158941e-01 3.44784826e-01 6.81456327e-01
-1.04128194e+00 1.07687652e+00 5.35625553e+00 8.01576257e-01
-1.22725105e+00 2.50807047e-01 2.95943052e-01 -1.22707091e-01
-2.32231840e-01 -4.48034823e-01 -1.28201425e+00 4.79529828e-01
1.24391103e+00 3.47263962e-01 4.98573214e-01 7.54781067e-01
4.80635047e-01 2.64446214e-02 -1.14002824e+00 9.81831372e-01
4.21054482e-01 -9.14001584e-01 1.59290746e-01 -1.21946320e-01
5.32035172e-01 1.11365728e-01 4.71693248e-01 7.54940093e-01
1.79337978e-01 -1.07442272e+00 6.48048043e-01 5.29634893e-01
5.91290355e-01 -8.34888220e-01 4.19859231e-01 4.70046401e-01
-1.49267793e+00 -3.05842847e-01 -1.92184791e-01 2.85809636e-01
4.69494641e-01 6.62651286e-02 -6.07425451e-01 3.87923539e-01
7.67727792e-01 5.81258595e-01 -7.67179787e-01 9.37881052e-01
-3.17666918e-01 4.83502030e-01 -6.93783313e-02 1.24262944e-01
3.66162300e-01 -4.03195024e-02 4.39209670e-01 9.85558152e-01
4.56075370e-01 -3.35951060e-01 9.71840024e-02 7.51915276e-01
-4.50762622e-02 -5.23083843e-02 -5.20743310e-01 3.26294363e-01
5.59446275e-01 8.31624806e-01 -1.90630451e-01 -8.27861503e-02
-7.27802634e-01 1.26959896e+00 4.04525131e-01 7.38791049e-01
-8.85758519e-01 -2.80195236e-01 6.30724251e-01 -5.54094128e-02
4.89151776e-01 -2.96876967e-01 -1.79537356e-01 -9.92913604e-01
1.16609365e-01 -8.47354710e-01 1.39883339e-01 -8.53325605e-01
-1.01807034e+00 8.11134994e-01 -2.74976343e-01 -1.23367321e+00
-1.49171099e-01 -4.25738990e-01 -6.63672447e-01 1.15733504e+00
-1.89429724e+00 -1.28077531e+00 -4.08959806e-01 8.52384150e-01
1.16006172e+00 -2.51081318e-01 8.80296707e-01 3.75757664e-01
-9.44292128e-01 9.97639596e-01 -2.58872509e-02 -7.03267381e-02
1.09321046e+00 -1.07283485e+00 3.98896337e-01 1.01606905e+00
8.16732943e-02 8.47486019e-01 6.49627090e-01 -4.93898720e-01
-1.37142026e+00 -1.32796454e+00 7.84133375e-01 -5.10226190e-01
3.88094604e-01 -4.67661887e-01 -1.10382569e+00 8.61246288e-01
3.13865393e-01 2.69143078e-02 7.01980591e-01 1.17174350e-01
-5.38371384e-01 -2.52861917e-01 -5.89414179e-01 7.48424590e-01
9.43823457e-01 -7.40919888e-01 -7.20012009e-01 -5.42051978e-02
9.97795224e-01 -5.16816258e-01 -2.92566866e-01 3.56191784e-01
3.20204288e-01 -9.52058673e-01 1.00028181e+00 -1.49908543e-01
2.70059735e-01 -3.82683367e-01 -2.96236426e-01 -1.37928069e+00
-2.38794953e-01 -1.94961339e-01 -3.39671463e-01 1.38960803e+00
4.14912343e-01 -4.46456194e-01 6.42617404e-01 4.39864993e-01
-6.59997404e-01 -6.82402492e-01 -1.16170883e+00 -7.63640344e-01
-1.10930406e-01 -4.41302359e-01 5.42193055e-01 7.35739112e-01
-3.58894438e-01 6.92313135e-01 -4.51770097e-01 3.73686194e-01
6.57651246e-01 -1.74784601e-01 5.71331143e-01 -1.00347257e+00
-9.85470787e-02 -5.32556772e-01 1.01423867e-01 -1.80206561e+00
4.71443623e-01 -8.20279419e-01 5.64970136e-01 -1.75964141e+00
-6.42706081e-03 -2.16479391e-01 -6.07266009e-01 3.89128596e-01
-4.19407517e-01 -2.39043713e-01 -6.79123700e-02 -3.91890435e-03
-7.16218472e-01 1.20033753e+00 1.29624975e+00 -1.65519834e-01
-5.76854944e-01 2.27379397e-01 -5.33769965e-01 5.40066183e-01
6.48306549e-01 -2.13502169e-01 -6.32357538e-01 -8.23755503e-01
-3.48273903e-01 1.09510854e-01 2.56107777e-01 -1.08596146e+00
6.25546515e-01 9.47103798e-02 3.73537481e-01 -1.11685097e+00
6.13727689e-01 -6.98947847e-01 -4.78943408e-01 1.73685536e-01
-5.98742366e-01 1.32710442e-01 4.46026802e-01 9.08271313e-01
-3.63957554e-01 -3.92097188e-03 4.95442480e-01 5.69538996e-02
-9.42178786e-01 2.63292104e-01 -3.69095951e-01 -2.63542384e-01
6.46235466e-01 -1.50034562e-01 -4.11157906e-01 -2.73167044e-01
-4.70172465e-01 6.13155544e-01 1.96093440e-01 1.02942717e+00
9.35612500e-01 -1.53229070e+00 -4.06385452e-01 3.75118911e-01
3.31369370e-01 2.46743619e-01 6.57067060e-01 8.21926296e-01
-8.13567638e-03 7.54389226e-01 -2.71390826e-02 -7.62654245e-01
-1.10767555e+00 4.86443698e-01 3.48734647e-01 -8.53428021e-02
-6.12372100e-01 1.25574207e+00 3.72062832e-01 -8.66914272e-01
9.16537941e-01 -3.23887318e-01 -5.11276186e-01 -1.10939682e-01
8.20857823e-01 9.67485607e-02 1.65859144e-02 -7.92760491e-01
-3.72706205e-01 2.91270256e-01 -3.46411943e-01 -2.57757187e-01
1.12216592e+00 -5.91217816e-01 1.79925099e-01 9.83296216e-01
1.00962269e+00 -1.90744251e-01 -1.70715976e+00 -4.62422878e-01
-2.64121324e-01 -9.12543684e-02 4.45401311e-01 -7.43957639e-01
-1.00808704e+00 1.31782436e+00 9.10801113e-01 -5.11059344e-01
9.40293729e-01 -2.26183251e-01 8.92842710e-01 4.34426516e-01
1.26626968e-01 -9.56839263e-01 2.34130576e-01 7.85616338e-01
9.27419424e-01 -1.56759679e+00 -6.36585355e-01 3.49796340e-02
-7.19352722e-01 9.89420831e-01 1.19982207e+00 5.93301713e-01
4.52554345e-01 -1.52512277e-02 2.53510505e-01 2.00394899e-01
-6.64176762e-01 4.58971672e-02 3.76750559e-01 5.44085026e-01
6.19908154e-01 -2.06304878e-01 2.43306383e-01 8.26527357e-01
-8.36935192e-02 -2.07236782e-01 2.92175770e-01 8.76336098e-01
-3.96798849e-01 -1.01804578e+00 -4.70285267e-01 1.49759397e-01
-8.24906379e-02 -3.90191436e-01 2.30853215e-01 3.61964613e-01
-1.28740102e-01 1.01789105e+00 1.20554596e-01 -4.78859842e-01
4.22858685e-01 4.07560825e-01 2.41149634e-01 -5.52353919e-01
-3.66575480e-01 1.32419020e-01 -3.02889138e-01 -8.28433096e-01
-3.14923465e-01 -4.06784087e-01 -1.32434487e+00 2.49599367e-02
-3.80751222e-01 -3.35631631e-02 8.73994768e-01 9.54704463e-01
4.15787995e-01 1.27404737e+00 4.47194606e-01 -9.49885607e-01
-5.74467182e-01 -1.16480756e+00 -2.12073073e-01 -5.21618500e-02
6.84289813e-01 -6.91319525e-01 -3.55394751e-01 -5.86657450e-02] | [14.238630294799805, 5.0958685874938965] |
e927c27b-bcfe-4bf7-a64c-50b7aad33a5b | local-shape-spectrum-analysis-for-3d-facial | 1705.06900 | null | http://arxiv.org/abs/1705.06900v1 | http://arxiv.org/pdf/1705.06900v1.pdf | Local Shape Spectrum Analysis for 3D Facial Expression Recognition | We investigate the problem of facial expression recognition using 3D data.
Building from one of the most successful frameworks for facial analysis using
exclusively 3D geometry, we extend the analysis from a curve-based
representation into a spectral representation, which allows a complete
description of the underlying surface that can be further tuned to the desired
level of detail. Spectral representations are based on the decomposition of the
geometry in its spatial frequency components, much like a Fourier transform,
which are related to intrinsic characteristics of the surface. In this work, we
propose the use of Graph Laplacian Features (GLF), which results from the
projection of local surface patches into a common basis obtained from the Graph
Laplacian eigenspace. We test the proposed approach in the BU-3DFE database in
terms of expressions and Action Units recognition. Our results confirm that the
proposed GLF produces consistently higher recognition rates than the
curves-based approach, thanks to a more complete description of the surface,
while requiring a lower computational complexity. We also show that the GLF
outperform the most popular alternative approach for spectral representation,
Shape- DNA, which is based on the Laplace Beltrami Operator and cannot provide
a stable basis that guarantee that the extracted signatures for the different
patches are directly comparable. | ['Federico M. Sukno', 'Dmytro Derkach'] | 2017-05-19 | null | null | null | null | ['3d-facial-expression-recognition'] | ['computer-vision'] | [ 2.94425428e-01 3.97321992e-02 9.41672027e-02 -3.93980682e-01
-4.00634110e-01 -5.49558163e-01 6.48667097e-01 3.52607667e-02
-2.63462663e-01 3.32631379e-01 -7.12077990e-02 1.24224953e-01
-2.30388522e-01 -7.71146953e-01 -3.95319611e-01 -1.06877398e+00
-1.00196406e-01 1.29590511e-01 7.74030760e-02 -4.17045683e-01
8.93529058e-02 1.11379409e+00 -1.95005894e+00 2.17803255e-01
4.34331715e-01 1.20788503e+00 -3.51474404e-01 2.39031285e-01
-2.79382259e-01 1.09961659e-01 -4.83157814e-01 -2.78224140e-01
1.51796132e-01 -3.78276438e-01 -3.12736630e-01 4.38031793e-01
4.91840273e-01 3.98419261e-01 2.21564636e-01 1.09673810e+00
2.92504430e-01 -1.33038640e-01 1.09051073e+00 -1.09270477e+00
-1.11944593e-01 -2.83384889e-01 -5.54451466e-01 -2.70035952e-01
6.62137806e-01 -4.88495141e-01 8.29782903e-01 -1.09686172e+00
8.35107923e-01 1.41883850e+00 6.61001027e-01 2.15171710e-01
-1.71488976e+00 -4.03945357e-01 -2.09367171e-01 2.14096278e-01
-1.73920119e+00 -5.41013360e-01 1.16653669e+00 -5.55782676e-01
4.93556291e-01 4.44629848e-01 6.36844754e-01 9.12711680e-01
-1.85791869e-02 4.28295195e-01 1.37492025e+00 -7.48101175e-01
2.32487828e-01 1.75699234e-01 9.37283188e-02 7.86241114e-01
5.39648868e-02 -2.69113362e-01 -3.32024723e-01 -4.62134749e-01
7.10972369e-01 -2.57391542e-01 -4.50706065e-01 -8.12943935e-01
-6.48182094e-01 8.23430896e-01 6.22839816e-02 7.64561832e-01
-5.40536880e-01 -1.82668313e-01 3.22411209e-01 2.96041518e-01
4.58373040e-01 1.73294116e-02 6.68767765e-02 2.99209766e-02
-9.23242867e-01 2.18513235e-01 8.86883020e-01 4.81726557e-01
1.01726449e+00 4.45406213e-02 -8.53767022e-02 7.04899669e-01
4.18267906e-01 6.93511248e-01 2.73655623e-01 -6.68601274e-01
-1.29740894e-01 8.48463476e-01 -7.50415996e-02 -1.49002445e+00
-4.97169137e-01 -2.13532820e-01 -7.38244534e-01 6.06573284e-01
4.86298025e-01 2.04320699e-01 -4.44146305e-01 1.79196858e+00
4.51710612e-01 3.87484618e-02 -5.12064509e-02 6.86624050e-01
5.32495916e-01 3.43288124e-01 -1.83926210e-01 -2.93597788e-01
1.24866533e+00 -1.40465572e-02 -7.01238871e-01 6.36029005e-01
5.73797047e-01 -6.86468780e-01 7.47383535e-01 6.42089486e-01
-8.10391963e-01 -5.05577743e-01 -9.50663686e-01 3.01112294e-01
-4.29129690e-01 3.72830719e-01 2.25490883e-01 9.36102033e-01
-1.30491626e+00 4.79288548e-01 -5.96723855e-01 -5.38931310e-01
6.06309809e-02 2.86207557e-01 -8.79241705e-01 2.55514055e-01
-8.26892138e-01 7.45127082e-01 -1.64380148e-02 2.60916445e-02
-1.77668333e-02 -3.40836465e-01 -9.06274974e-01 4.03408147e-02
-1.22461610e-01 -1.49250799e-03 5.10986209e-01 -1.28360808e+00
-1.70794594e+00 1.20850968e+00 -3.22580099e-01 -2.69533582e-02
4.74863172e-01 4.22509164e-01 -3.34418535e-01 4.99899805e-01
-3.30500871e-01 8.35463628e-02 1.30034542e+00 -1.21786702e+00
1.65891960e-01 -7.24869251e-01 -2.59277046e-01 -1.73192590e-01
-4.64019358e-01 -9.21974778e-02 -1.57010898e-01 -4.94458288e-01
2.78615862e-01 -1.03280687e+00 2.96096414e-01 1.67451754e-01
-7.09322244e-02 -1.66773424e-01 9.39043462e-01 -5.60696959e-01
1.15140498e+00 -2.35527515e+00 5.84565043e-01 7.88552165e-01
-1.10159643e-01 2.76862085e-01 -1.27551958e-01 6.26123488e-01
-4.53423262e-01 -4.32963520e-02 -3.20264876e-01 -2.28850290e-01
-1.25306901e-02 4.59050909e-02 -5.80431037e-02 9.55089986e-01
4.43688422e-01 4.53899682e-01 -4.22063082e-01 -3.80599141e-01
1.69445470e-01 9.05374706e-01 -3.94970864e-01 -2.16156006e-01
1.80255845e-01 4.88252521e-01 -5.50807476e-01 3.61453086e-01
8.78816247e-01 2.53918469e-01 1.92321941e-01 -3.86647999e-01
-1.02367766e-01 -3.62229586e-01 -1.41495895e+00 1.60344791e+00
-5.22573948e-01 4.85945851e-01 4.49952513e-01 -1.46119714e+00
1.68064749e+00 3.99434417e-01 8.13159108e-01 -3.30693275e-01
1.92487746e-01 3.71594965e-01 -2.68616050e-01 -3.33254486e-01
-1.41128615e-01 -1.85407713e-01 8.35346582e-04 2.55733550e-01
2.35506088e-01 -4.57446277e-02 2.53727883e-02 -4.36466455e-01
8.06375265e-01 3.04145694e-01 3.74924988e-01 -6.83304071e-01
1.37966621e+00 -5.86176991e-01 8.11619461e-02 9.03782547e-02
1.76919535e-01 4.65715110e-01 7.64241755e-01 -3.23696434e-01
-6.84991479e-01 -1.01109338e+00 -5.79679489e-01 6.22881353e-01
-2.04032406e-01 -3.74511778e-01 -1.07582569e+00 -5.06677210e-01
1.73971236e-01 3.94818574e-01 -7.91372001e-01 -1.99190259e-01
-3.79092067e-01 -5.42306483e-01 5.31213582e-01 -1.84076667e-01
3.01444352e-01 -6.23832583e-01 -7.94110060e-01 -1.23395309e-01
8.28390718e-02 -9.65623260e-01 -1.20456800e-01 -2.30308756e-01
-8.89926434e-01 -9.91401196e-01 -9.24504697e-01 -3.77121031e-01
6.46086752e-01 -8.45544711e-02 7.05386937e-01 -8.30123425e-02
-4.41439927e-01 9.03114200e-01 -5.49692750e-01 -2.32874751e-01
-5.29873908e-01 -2.82913625e-01 3.83068211e-02 1.10874772e+00
4.31066155e-01 -9.51254487e-01 -1.81487262e-01 3.31727922e-01
-1.01152956e+00 -3.65849823e-01 5.18357575e-01 4.82612222e-01
5.59306979e-01 -2.25591511e-01 4.34199870e-01 -6.84012294e-01
5.74225962e-01 -2.28790686e-01 -5.11085331e-01 2.76754469e-01
-3.11749220e-01 4.07739192e-01 5.04065394e-01 -3.58927757e-01
-8.23606074e-01 4.04037535e-01 -1.82507336e-01 -6.47654891e-01
-3.53309393e-01 3.26272607e-01 -1.29745916e-01 -7.19550610e-01
8.56387973e-01 4.79024023e-01 4.28154379e-01 -7.04882026e-01
1.43913180e-01 5.45074105e-01 1.34047732e-01 -5.05095601e-01
7.71309018e-01 7.23452926e-01 6.26551032e-01 -1.55343676e+00
-1.86690703e-01 -5.68713546e-01 -8.73513103e-01 -5.26012778e-01
8.05357218e-01 -3.90212804e-01 -8.73998225e-01 3.02276939e-01
-1.27478516e+00 2.99746275e-01 -2.75362849e-01 3.12743366e-01
-8.03359270e-01 7.08409786e-01 -1.38920605e-01 -1.26025116e+00
-7.26502482e-03 -9.04737294e-01 1.51462317e+00 -1.59990609e-01
-1.15644380e-01 -1.05009758e+00 2.70594269e-01 -1.31956741e-01
3.09984773e-01 7.65230358e-01 9.74883139e-01 -3.03849190e-01
-3.60845476e-02 -5.53050876e-01 4.69050817e-02 4.56079662e-01
1.96030095e-01 1.83882311e-01 -1.06767964e+00 -2.04931453e-01
1.91229448e-01 -7.53045529e-02 7.50903904e-01 1.40014783e-01
8.70179653e-01 1.86478277e-03 -8.29269588e-02 6.04523301e-01
1.49816656e+00 -4.34913766e-03 6.81781530e-01 3.06998529e-02
3.45002264e-01 8.95972669e-01 3.25951725e-01 6.87302172e-01
-2.28035122e-01 1.20171893e+00 3.08491826e-01 -1.43991083e-01
-4.16706093e-02 -3.85832861e-02 5.31575739e-01 4.19610411e-01
-5.70101678e-01 3.15757066e-01 -6.47972167e-01 -1.20420912e-02
-1.68360901e+00 -8.72777283e-01 -9.43417102e-02 2.55570745e+00
3.73083472e-01 -2.71738619e-01 3.71686429e-01 5.36078572e-01
5.94655812e-01 1.46721408e-01 9.01416782e-03 -7.42836177e-01
-3.01972836e-01 5.92663944e-01 2.38936454e-01 6.16163909e-01
-8.47981870e-01 4.77163523e-01 5.87484837e+00 9.03084278e-01
-1.36205912e+00 -7.50036612e-02 2.12913468e-01 4.30357784e-01
-2.53162533e-01 -3.11257869e-01 -6.08875990e-01 1.98842898e-01
7.16314971e-01 -2.26473138e-01 1.38210759e-01 7.10715771e-01
2.72117257e-01 1.85796749e-02 -9.77292359e-01 1.27586782e+00
3.10357153e-01 -8.99494052e-01 2.96924531e-01 3.48344415e-01
3.50051254e-01 -6.70137882e-01 1.59199134e-01 -1.69597104e-01
-7.59779692e-01 -9.97612417e-01 7.35266864e-01 9.25833881e-01
8.71266782e-01 -6.47365689e-01 5.77201724e-01 2.98395336e-01
-1.19806075e+00 1.18046030e-01 -2.62013018e-01 1.36329561e-01
-1.07135169e-01 4.28421915e-01 -6.09318137e-01 6.09440207e-01
3.14405739e-01 6.21838689e-01 -4.10528332e-01 6.02562547e-01
9.45348516e-02 3.46544623e-01 -3.93604398e-01 -4.58866395e-02
3.65461595e-02 -8.31165195e-01 7.71628499e-01 1.41684330e+00
6.24688685e-01 -3.81775424e-02 -9.88577902e-02 1.00093782e+00
4.00138736e-01 7.62827098e-01 -8.33436072e-01 1.53947843e-03
-2.42016420e-01 1.41088629e+00 -6.25888944e-01 1.67901888e-01
-5.58907092e-01 9.31179285e-01 1.31776348e-01 3.20701540e-01
-4.01741326e-01 -2.26785079e-01 6.77220881e-01 4.77496386e-01
3.65416229e-01 -4.26397026e-01 3.60943764e-01 -9.56548154e-01
2.63233870e-01 -6.95522249e-01 1.65163472e-01 -5.12768507e-01
-1.03756273e+00 7.45703936e-01 1.49114266e-01 -1.29552698e+00
-4.17678475e-01 -1.00781357e+00 -4.25355673e-01 1.14487016e+00
-1.33628821e+00 -1.13813686e+00 -3.48964423e-01 9.58589256e-01
-9.18244347e-02 -1.82684585e-02 1.33853853e+00 1.66942492e-01
1.38255730e-02 5.10438859e-01 -3.77967134e-02 -5.71174473e-02
3.92711014e-01 -9.38191175e-01 -2.21421182e-01 2.37909570e-01
3.49853754e-01 2.93863595e-01 7.05399692e-01 -2.30667710e-01
-1.48124135e+00 -5.24302065e-01 7.28844106e-01 -2.28221342e-01
3.66771668e-01 -4.99650627e-01 -9.18937683e-01 1.93970338e-01
-3.93463552e-01 2.26866324e-02 7.21376777e-01 -6.57518357e-02
-4.54215229e-01 -2.85816103e-01 -1.15858996e+00 1.39495939e-01
8.30143571e-01 -5.54309011e-01 -4.01845604e-01 6.65061697e-02
-2.23609045e-01 1.58388600e-01 -1.07237566e+00 3.70267481e-01
6.46829307e-01 -1.38909149e+00 7.79600739e-01 -4.11386043e-01
-2.22630829e-01 -2.25837067e-01 -2.98631996e-01 -1.00410306e+00
-2.73628652e-01 -5.20969748e-01 3.40342432e-01 9.74456251e-01
2.01081663e-01 -6.30900502e-01 8.80562544e-01 5.18001504e-02
3.44768286e-01 -6.88451231e-01 -1.34916282e+00 -8.27188313e-01
-1.71357185e-01 -3.79780442e-01 1.34478480e-01 8.44681978e-01
1.38414100e-01 3.68097201e-02 -3.66776176e-02 -3.14767249e-02
6.68167233e-01 2.53412068e-01 6.54229343e-01 -1.59447050e+00
-3.20139349e-01 -5.58287442e-01 -1.24244833e+00 -5.12384295e-01
6.03720367e-01 -1.07848799e+00 -3.62110376e-01 -8.87858331e-01
-1.87497050e-01 -2.43389130e-01 -1.04880027e-01 1.78786024e-01
5.36484122e-01 4.43922967e-01 1.37426019e-01 5.18187396e-02
2.78422944e-02 6.73814952e-01 1.04105127e+00 9.31644738e-02
-3.33063364e-01 3.69101726e-02 -1.90182757e-02 7.78264225e-01
2.88826913e-01 6.98175058e-02 -2.13517100e-01 1.32586330e-01
2.22786590e-02 4.08900082e-02 3.18768740e-01 -1.01514316e+00
-1.59821659e-01 1.66685626e-01 1.39661938e-01 3.73744890e-02
7.26855099e-01 -1.00199854e+00 4.12080258e-01 3.36012930e-01
-5.33165857e-02 -2.95001715e-01 2.14267671e-01 4.28501815e-01
-4.75475192e-01 -1.74338549e-01 9.34580326e-01 1.79022089e-01
-5.00024974e-01 1.85378224e-01 -3.42170894e-01 -5.89936435e-01
1.04978156e+00 -6.17904782e-01 3.42516273e-01 -6.45249546e-01
-1.06032932e+00 -6.15488708e-01 6.83372259e-01 1.19371526e-01
5.58198452e-01 -1.52621949e+00 -7.78504908e-01 6.30070984e-01
2.57266372e-01 -7.54088521e-01 1.65565684e-01 1.10824049e+00
-4.19075578e-01 3.30577344e-01 -5.66648424e-01 -9.01700556e-01
-1.52625680e+00 5.46768427e-01 5.26304901e-01 1.14501104e-01
-4.87715751e-01 3.38959038e-01 3.64010841e-01 -1.79849163e-01
9.02254134e-03 -9.21406746e-02 -4.99733776e-01 3.89790297e-01
3.67969930e-01 2.67429650e-01 2.11590976e-01 -1.42276073e+00
-4.73398387e-01 1.34011471e+00 7.39088476e-01 -3.48712206e-01
1.21897042e+00 1.94130570e-01 -3.37516487e-01 6.17985785e-01
1.57155514e+00 3.67790729e-01 -8.02852869e-01 -9.94403437e-02
1.15899645e-01 -3.74537855e-01 3.78798274e-03 -2.07707688e-01
-8.55433583e-01 9.59558308e-01 6.46783173e-01 4.42386895e-01
1.35875475e+00 -1.40595855e-02 -4.88446690e-02 9.79303867e-02
4.92294073e-01 -7.06360996e-01 -2.24804521e-01 -3.33344489e-02
1.35135162e+00 -4.84523714e-01 -1.87904477e-01 -8.70980144e-01
-2.02398360e-01 1.67187333e+00 -1.96899295e-01 -3.91403288e-01
9.23611164e-01 4.48847637e-02 -2.35779770e-02 -2.17567593e-01
-1.92060754e-01 -3.86269420e-01 6.17511809e-01 5.98478615e-01
5.18922985e-01 -2.01802272e-02 -5.49878180e-01 4.33528066e-01
7.05447840e-03 -1.04713999e-01 2.45763019e-01 4.43087518e-01
-3.37976038e-01 -1.29806137e+00 -6.40621305e-01 -1.34635549e-02
-2.83223093e-01 5.00744402e-01 -6.40479743e-01 8.48081350e-01
1.42034099e-01 6.32102251e-01 5.39099164e-02 -2.96214223e-01
7.37294078e-01 6.17008269e-01 9.89495873e-01 -4.95813072e-01
2.16407739e-02 1.91340104e-01 2.08205190e-02 -8.27968836e-01
-7.74437666e-01 -9.03416812e-01 -9.74844635e-01 5.73934764e-02
-1.20219335e-01 2.03009918e-01 9.68947351e-01 5.95298409e-01
2.96086848e-01 -1.36782736e-01 8.48192036e-01 -1.15430427e+00
-5.68818450e-01 -7.68049419e-01 -1.17666972e+00 8.18829775e-01
2.78313607e-01 -1.08239996e+00 -5.08893013e-01 4.89978604e-02] | [12.864557266235352, 0.7841561436653137] |
12063b47-29fb-4c97-9bec-cc1de1a11ec0 | cross-document-event-coreference-search-task | 2210.12654 | null | https://arxiv.org/abs/2210.12654v1 | https://arxiv.org/pdf/2210.12654v1.pdf | Cross-document Event Coreference Search: Task, Dataset and Modeling | The task of Cross-document Coreference Resolution has been traditionally formulated as requiring to identify all coreference links across a given set of documents. We propose an appealing, and often more applicable, complementary set up for the task - Cross-document Coreference Search, focusing in this paper on event coreference. Concretely, given a mention in context of an event of interest, considered as a query, the task is to find all coreferring mentions for the query event in a large document collection. To support research on this task, we create a corresponding dataset, which is derived from Wikipedia while leveraging annotations in the available Wikipedia Event Coreference dataset (WEC-Eng). Observing that the coreference search setup is largely analogous to the setting of Open Domain Question Answering, we adapt the prominent Deep Passage Retrieval (DPR) model to our setting, as an appealing baseline. Finally, we present a novel model that integrates a powerful coreference scoring scheme into the DPR architecture, yielding improved performance. | ['Ido Dagan', 'Avi Caciularu', 'Alon Eirew'] | 2022-10-23 | null | null | null | null | ['cross-document-coreference-resolution', 'passage-retrieval', 'coreference-resolution', 'open-domain-question-answering'] | ['natural-language-processing', 'natural-language-processing', 'natural-language-processing', 'natural-language-processing'] | [ 1.04403824e-01 4.08051640e-01 -2.86163300e-01 -1.94774255e-01
-1.57201493e+00 -9.51116025e-01 9.55962241e-01 5.60583055e-01
-7.49636889e-01 7.34382272e-01 1.09073162e+00 -2.20762119e-02
-5.35877943e-01 -5.53785324e-01 -7.62152493e-01 -3.82859290e-01
1.11675575e-01 1.06287944e+00 4.08906072e-01 -4.54635262e-01
3.51044357e-01 3.98596942e-01 -1.34106624e+00 5.28732657e-01
6.58351839e-01 4.37814474e-01 3.19750369e-01 5.02262354e-01
-5.11044189e-02 6.61270678e-01 -4.16506588e-01 -9.37044680e-01
-3.13335598e-01 -1.50986597e-01 -1.94093108e+00 -9.29265738e-01
7.65306890e-01 2.17353925e-01 -7.00949192e-01 8.73171270e-01
6.28528357e-01 6.22818887e-01 5.06499469e-01 -9.75972116e-01
-3.06441933e-01 1.13770199e+00 -1.73859403e-01 6.72892630e-01
8.96122992e-01 -6.91337824e-01 1.88724470e+00 -5.61219871e-01
1.13841867e+00 1.38299477e+00 4.80316430e-01 5.46881318e-01
-1.30497873e+00 -3.50692093e-01 -7.32119083e-02 5.30103743e-01
-1.09137881e+00 -5.92141509e-01 5.53250849e-01 -4.49138358e-02
1.11416805e+00 5.36887705e-01 -8.94137844e-02 1.48804414e+00
-2.98007578e-01 8.85083079e-01 1.69860616e-01 -6.40139401e-01
-2.34854683e-01 -1.88728958e-01 7.99713552e-01 -2.06758194e-02
6.45267889e-02 1.25150263e-01 -5.57401240e-01 -5.53651869e-01
3.22376452e-02 -4.49548960e-01 -8.14960003e-01 -4.83930111e-01
-1.06866169e+00 8.73607278e-01 5.79772294e-01 6.69162452e-01
-3.14470768e-01 -2.51811117e-01 6.18290424e-01 3.51238042e-01
-3.18916552e-02 7.41778553e-01 -4.68523771e-01 -9.54423845e-02
-6.62557662e-01 6.88242376e-01 1.35787666e+00 8.68391216e-01
4.14082497e-01 -1.13421130e+00 -4.75008368e-01 1.00654709e+00
2.31298506e-01 2.55137593e-01 1.99558511e-01 -1.33426905e+00
8.19607019e-01 2.99762875e-01 3.93520772e-01 -9.15171206e-01
-5.43382525e-01 -2.06050724e-01 -2.95314461e-01 -6.23934329e-01
4.74545747e-01 2.27517694e-01 -8.39992091e-02 2.32689357e+00
3.91674697e-01 1.14667602e-01 5.16778052e-01 9.26176071e-01
1.19324291e+00 5.73293507e-01 2.11484700e-01 -3.21786135e-01
1.80107546e+00 -8.55045378e-01 -6.23198330e-01 -1.06014602e-01
6.51290059e-01 -7.49315381e-01 6.77664042e-01 -5.94461784e-02
-1.12357879e+00 -3.16462629e-02 -7.76441336e-01 -5.56543052e-01
-2.77791917e-01 -3.39687735e-01 3.33710164e-01 -2.50121891e-01
-8.20985794e-01 6.73025310e-01 -4.91474271e-01 -1.12043560e+00
-3.35322917e-01 3.11847366e-02 -4.35537875e-01 -7.36148357e-02
-1.81950831e+00 1.17169857e+00 6.64704859e-01 -2.78589278e-01
-6.79888308e-01 -8.19996834e-01 -7.11650431e-01 3.45123947e-01
5.66821218e-01 -9.14698184e-01 1.83234191e+00 -4.70988810e-01
-8.03049207e-01 1.33701146e+00 -2.36532182e-01 -6.62781894e-01
3.05790119e-02 -5.61825871e-01 -6.62591457e-01 4.49343026e-01
3.86330813e-01 4.29440260e-01 9.18540582e-02 -1.11542344e+00
-7.04094291e-01 -6.14899635e-01 4.59365368e-01 3.91104460e-01
5.56834973e-02 6.29696429e-01 -6.60639644e-01 -3.63431931e-01
-2.77066618e-01 -9.08560097e-01 1.41904518e-01 -6.40154898e-01
-4.57678586e-01 -7.78601944e-01 5.58615625e-01 -5.47112226e-01
1.52795947e+00 -2.06862235e+00 3.94331694e-01 -1.19379111e-01
4.88311052e-02 2.52922386e-01 -4.34882343e-01 1.00558782e+00
-4.53097075e-01 -1.28210694e-01 -2.33776346e-01 -1.95734605e-01
2.92282760e-01 3.12357157e-01 -8.76812398e-01 1.61010161e-01
-6.73904568e-02 8.34569633e-01 -1.22703648e+00 -7.63648927e-01
-4.09001321e-01 2.03117311e-01 -5.88634789e-01 2.51883596e-01
-4.25826401e-01 1.07619502e-01 -7.51922548e-01 7.86956027e-02
3.57347637e-01 -1.22080103e-01 3.35071921e-01 -5.56670487e-01
-1.91683963e-01 9.84228373e-01 -1.05155230e+00 2.21835279e+00
-2.65034735e-01 6.15853071e-01 2.27096483e-01 -1.00411427e+00
5.64962268e-01 8.06565762e-01 3.95269305e-01 -7.79543459e-01
-2.18961522e-01 2.37103343e-01 -2.10948408e-01 -6.68558240e-01
1.01504052e+00 7.15472400e-02 -4.48796868e-01 7.17685282e-01
5.38445354e-01 3.93058807e-01 4.35190350e-01 8.94090772e-01
1.16801810e+00 4.07359470e-03 2.59081185e-01 -4.04927164e-01
6.64514482e-01 2.68735230e-01 4.30681169e-01 9.13954735e-01
-1.15382612e-01 4.62592214e-01 4.37025219e-01 -1.77160073e-02
-7.43705273e-01 -1.21118808e+00 -4.70510274e-01 1.35580039e+00
9.75930393e-02 -5.97083688e-01 -6.03403687e-01 -7.04196692e-01
-1.28366664e-01 8.16926658e-01 -5.99223495e-01 6.49764389e-02
-9.99449372e-01 -3.35717857e-01 9.22218740e-01 2.97038794e-01
2.21813247e-02 -1.23821533e+00 -5.57669818e-01 2.83516496e-01
-1.06458437e+00 -9.89294350e-01 -7.28845298e-01 1.60930574e-01
-3.88832211e-01 -1.53023505e+00 -5.66811919e-01 -1.08594394e+00
-3.58353585e-01 8.73366967e-02 1.95451069e+00 -7.65493140e-02
-6.41224533e-02 8.62848938e-01 -5.25145829e-01 -1.61325067e-01
-2.66221046e-01 4.05622005e-01 -1.49671316e-01 -3.96692574e-01
7.31532693e-01 -5.50281584e-01 -4.88232434e-01 -2.40482576e-03
-9.01447415e-01 -6.23649836e-01 1.18989013e-01 7.45172858e-01
3.28317791e-01 -7.22009122e-01 6.25304282e-01 -8.07148695e-01
8.03248882e-01 -7.64992595e-01 -2.68357456e-01 6.57214820e-01
-1.58417150e-01 3.59054267e-01 2.02216119e-01 -5.47637306e-02
-1.43698049e+00 -5.29890358e-01 -2.79885560e-01 2.01744167e-03
-1.53490275e-01 6.96009815e-01 -2.81668991e-01 4.46644485e-01
6.61444962e-01 -1.90663651e-01 -5.19126356e-01 -8.41078281e-01
6.17120266e-01 6.58577085e-01 1.34500289e+00 -1.04012525e+00
2.97054678e-01 3.04762185e-01 -2.98794597e-01 -5.06219268e-01
-1.39208281e+00 -1.06866252e+00 -6.46224737e-01 3.26179713e-01
8.46237123e-01 -8.00122201e-01 -1.04388475e+00 -2.76637316e-01
-1.73648810e+00 1.31172732e-01 -1.91488251e-01 4.22276199e-01
-6.47646666e-01 6.02066696e-01 -6.05656564e-01 -3.87108415e-01
-5.43915212e-01 -6.38649702e-01 1.10024238e+00 1.88789904e-01
-5.61665416e-01 -1.07606220e+00 1.03201902e+00 3.66868287e-01
-1.92515813e-02 6.08548895e-02 1.09254992e+00 -1.39242721e+00
-2.23943919e-01 5.32044992e-02 -2.14915916e-01 -4.60574329e-01
-4.09526885e-01 -2.35018477e-01 -8.21304381e-01 -2.18121082e-01
-1.71665624e-01 -5.55538893e-01 1.06682312e+00 8.82728100e-02
3.07759851e-01 -1.70255721e-01 -8.36750507e-01 2.66385436e-01
1.25958920e+00 -7.03350082e-02 5.09873867e-01 8.97657514e-01
1.33424699e-01 8.48803878e-01 7.14114726e-01 2.58859456e-01
7.51991689e-01 9.65993226e-01 1.71227649e-01 4.29870307e-01
-1.41290873e-01 -4.37487125e-01 -2.39554849e-02 5.93195260e-01
1.50968716e-01 -3.57496232e-01 -9.68906403e-01 9.63872612e-01
-2.13114882e+00 -1.61683249e+00 -2.36510381e-01 2.07696033e+00
1.13708687e+00 -3.34570318e-01 -1.34618338e-02 -3.48346442e-01
8.05739522e-01 2.18633398e-01 -2.41013914e-01 -1.49896875e-01
-3.42271447e-01 2.74330974e-01 -9.78119671e-02 7.52570748e-01
-1.23006797e+00 6.43517077e-01 5.87897348e+00 3.41757357e-01
-3.84475410e-01 1.02407403e-01 -3.24786901e-01 -6.20521829e-02
-4.16429073e-01 1.89854562e-01 -1.00439072e+00 3.61446291e-02
1.20713115e+00 -5.51394641e-01 1.77311257e-01 2.20726043e-01
-2.01181844e-01 -6.33277968e-02 -1.68865407e+00 7.63024926e-01
1.84667185e-01 -1.27822447e+00 1.23360164e-01 -3.95360500e-01
2.07661629e-01 3.38858813e-01 -3.61823440e-01 5.46464264e-01
7.10988343e-01 -6.12171173e-01 4.60665524e-01 4.84082103e-01
3.03577662e-01 -6.51020169e-01 6.55715942e-01 1.71924561e-01
-1.14564109e+00 -9.14155841e-02 -2.07099482e-01 4.43940729e-01
5.54031372e-01 -2.95681171e-02 -3.26933175e-01 1.07865965e+00
9.61718619e-01 6.38233781e-01 -1.49720341e-01 1.21099424e+00
-1.79959297e-01 3.77805680e-01 -2.77521431e-01 3.11369807e-01
2.62965769e-01 1.52976990e-01 1.04378843e+00 1.63011658e+00
5.36318980e-02 5.32353699e-01 -6.18685409e-02 7.26161480e-01
-3.94052505e-01 1.16215590e-02 -4.58783507e-01 2.46007904e-01
1.07097113e+00 1.15934503e+00 1.72253534e-01 -2.11526692e-01
-4.59750921e-01 8.50249648e-01 8.69620562e-01 3.28825563e-01
-6.86988413e-01 -6.40939772e-01 6.79548323e-01 -5.85293293e-01
2.68218368e-01 4.50151324e-01 6.47129953e-01 -1.21936560e+00
1.66785736e-02 -9.29241121e-01 1.39799476e+00 -7.90011883e-01
-1.38243198e+00 5.89886785e-01 1.43203884e-01 -7.89747953e-01
-4.12137896e-01 -1.34412289e-01 -7.40958214e-01 8.53488326e-01
-1.77165437e+00 -9.31079686e-01 -1.15202696e-04 8.08903515e-01
3.03774983e-01 2.95769036e-01 1.06814933e+00 5.44722557e-01
-5.67547679e-01 5.52684963e-01 2.38833278e-01 5.23974240e-01
1.25263166e+00 -1.19858766e+00 1.54313207e-01 7.17206180e-01
4.63746607e-01 9.55733597e-01 1.02967191e+00 -3.30253750e-01
-1.30656803e+00 -8.35594952e-01 1.69393551e+00 -8.38032722e-01
1.01276004e+00 2.56380923e-02 -1.16222823e+00 1.14729321e+00
6.90640986e-01 -4.69031066e-01 7.37080336e-01 5.68249941e-01
-6.71008289e-01 2.68359810e-01 -8.11524451e-01 4.54715699e-01
1.19498777e+00 -8.31132710e-01 -1.81578100e+00 1.99334264e-01
1.00449109e+00 -4.89687949e-01 -1.19292533e+00 3.56312811e-01
2.74212301e-01 -5.59333265e-01 1.29266298e+00 -1.24234498e+00
2.61795998e-01 1.02676950e-01 -4.71219569e-01 -1.26201808e+00
-5.62969089e-01 -7.57326245e-01 -2.50624299e-01 1.81030679e+00
4.90859181e-01 -2.58350104e-01 2.18477905e-01 6.63345993e-01
-3.75054419e-01 8.73029232e-02 -1.14863479e+00 -4.63848591e-01
3.26659501e-01 -1.63935319e-01 4.38013852e-01 1.06048608e+00
7.21204460e-01 9.85234678e-01 1.18460342e-01 4.76956040e-01
7.46339500e-01 6.49228752e-01 3.21609139e-01 -1.70303023e+00
-2.56060839e-01 -3.47968459e-01 2.92689562e-01 -1.15430522e+00
6.84632063e-01 -1.22056425e+00 1.54991761e-01 -1.46752024e+00
7.22355545e-01 -2.66124010e-01 -4.75749701e-01 1.36682346e-01
-2.66901135e-01 -2.29668841e-01 2.07664803e-01 4.18447882e-01
-1.13330305e+00 3.33805352e-01 7.52011955e-01 -2.03778639e-01
1.24896131e-01 -2.48902068e-01 -9.14439142e-01 4.78824675e-01
8.32084939e-02 -5.58671236e-01 -4.64445213e-03 -6.68216169e-01
3.55710745e-01 6.52822733e-01 4.26964909e-01 -3.27820301e-01
7.24933982e-01 1.31805509e-01 -4.78908747e-01 -5.74018955e-01
1.63859561e-01 -6.47826493e-01 -5.33201694e-02 7.81506971e-02
-9.40286279e-01 6.75376281e-02 2.35972747e-01 6.37873411e-01
-5.38017809e-01 -6.61674380e-01 3.92161816e-01 -1.34515777e-01
-8.64422619e-01 1.04795232e-01 3.74891460e-02 9.33976769e-01
5.15059829e-01 4.90059823e-01 -7.72960603e-01 -1.26725227e-01
-9.62140262e-01 6.27143383e-01 2.30682299e-01 6.43373072e-01
1.16506159e-01 -1.30936205e+00 -9.35690224e-01 -6.98067665e-01
4.79823560e-01 -3.39105159e-01 1.84038892e-01 8.19934547e-01
3.27584982e-01 1.24065852e+00 2.39615262e-01 -1.81038529e-01
-1.35255063e+00 8.71656299e-01 2.57695615e-01 -7.31556416e-01
-5.96505105e-01 6.72092497e-01 2.48866946e-01 -6.00962043e-01
4.24160987e-01 8.81838575e-02 -7.26962328e-01 4.10078138e-01
8.25023592e-01 2.98561692e-01 2.05144554e-01 -7.07330346e-01
-5.65810680e-01 3.53349268e-01 -2.79126018e-01 -1.67771891e-01
1.42804253e+00 -4.75888431e-01 -3.22249740e-01 2.19178841e-01
1.34770155e+00 4.80917701e-03 -5.67272961e-01 -6.81250453e-01
7.24225402e-01 7.91448355e-02 -2.36370713e-01 -9.17375803e-01
-4.81066227e-01 5.42534947e-01 1.19841523e-01 2.80349642e-01
7.20634043e-01 7.10636914e-01 7.87021816e-01 8.94270599e-01
2.62436688e-01 -7.55476475e-01 -2.74321049e-01 9.52837765e-01
1.07139218e+00 -9.82061863e-01 -4.81977910e-01 6.36440963e-02
-3.59393090e-01 9.82110023e-01 3.60256463e-01 1.08261645e-01
3.48653316e-01 1.05779842e-01 -1.34267509e-01 -5.86024642e-01
-1.02632463e+00 -3.23629618e-01 2.20364720e-01 4.54262972e-01
6.70311987e-01 -3.32746297e-01 -5.10378003e-01 8.76449645e-01
-5.73906489e-02 -1.16059549e-01 3.45584661e-01 5.96339166e-01
-3.76391232e-01 -1.12534404e+00 -3.11990589e-01 1.47937089e-02
-7.10115075e-01 -4.21847910e-01 -5.19283831e-01 7.35585451e-01
-4.32255000e-01 1.03748584e+00 4.34079796e-01 2.52947092e-01
7.33642340e-01 2.71888465e-01 4.90525246e-01 -4.10652280e-01
-6.06341004e-01 -5.33106089e-01 6.34238660e-01 -6.93104506e-01
-8.92523468e-01 -9.67709601e-01 -1.24660432e+00 -8.16533566e-02
-3.63204092e-01 9.52460587e-01 -8.03368464e-02 1.05504668e+00
4.61730212e-01 1.34084880e-01 2.76852220e-01 -4.61132139e-01
-6.09061956e-01 -9.20658767e-01 -4.37481016e-01 1.02985263e+00
4.68999505e-01 -7.12493956e-01 -2.30139911e-01 -8.01039413e-02] | [9.278921127319336, 9.542388916015625] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.