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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
77b078ed-75d7-4965-9457-9ff8e905d9af | streaming-algorithms-for-diversity | 2208.00194 | null | https://arxiv.org/abs/2208.00194v1 | https://arxiv.org/pdf/2208.00194v1.pdf | Streaming Algorithms for Diversity Maximization with Fairness Constraints | Diversity maximization is a fundamental problem with wide applications in data summarization, web search, and recommender systems. Given a set $X$ of $n$ elements, it asks to select a subset $S$ of $k \ll n$ elements with maximum \emph{diversity}, as quantified by the dissimilarities among the elements in $S$. In this paper, we focus on the diversity maximization problem with fairness constraints in the streaming setting. Specifically, we consider the max-min diversity objective, which selects a subset $S$ that maximizes the minimum distance (dissimilarity) between any pair of distinct elements within it. Assuming that the set $X$ is partitioned into $m$ disjoint groups by some sensitive attribute, e.g., sex or race, ensuring \emph{fairness} requires that the selected subset $S$ contains $k_i$ elements from each group $i \in [1,m]$. A streaming algorithm should process $X$ sequentially in one pass and return a subset with maximum \emph{diversity} while guaranteeing the fairness constraint. Although diversity maximization has been extensively studied, the only known algorithms that can work with the max-min diversity objective and fairness constraints are very inefficient for data streams. Since diversity maximization is NP-hard in general, we propose two approximation algorithms for fair diversity maximization in data streams, the first of which is $\frac{1-\varepsilon}{4}$-approximate and specific for $m=2$, where $\varepsilon \in (0,1)$, and the second of which achieves a $\frac{1-\varepsilon}{3m+2}$-approximation for an arbitrary $m$. Experimental results on real-world and synthetic datasets show that both algorithms provide solutions of comparable quality to the state-of-the-art algorithms while running several orders of magnitude faster in the streaming setting. | ['Michael Mathioudakis', 'Francesco Fabbri', 'Yanhao Wang'] | 2022-07-30 | null | null | null | null | ['data-summarization'] | ['miscellaneous'] | [ 1.58539847e-01 -2.54679602e-02 -5.06817758e-01 -4.43630546e-01
-7.49524772e-01 -5.74048042e-01 -3.93514842e-01 6.26826584e-01
-6.59468293e-01 8.18947256e-01 -7.11022168e-02 -2.43655130e-01
-8.08826685e-01 -1.15122032e+00 -4.52319652e-01 -8.07458520e-01
-8.13042462e-01 6.41727328e-01 -1.29587740e-01 -3.34200650e-01
2.90682584e-01 5.42980321e-02 -1.85574186e+00 -8.57647732e-02
1.09612083e+00 1.50897586e+00 -9.12471637e-02 4.89886910e-01
-3.87404680e-01 3.79312724e-01 -9.10496175e-01 -4.84750867e-01
5.95119059e-01 -7.46034026e-01 -8.58571529e-01 -2.70073279e-03
2.26441007e-02 -2.17754781e-01 -1.83545500e-01 1.08016610e+00
5.22050321e-01 3.16263348e-01 4.21288669e-01 -1.47979665e+00
-2.12539174e-02 1.24922752e+00 -9.84301150e-01 3.92342538e-01
3.56781006e-01 -2.13055164e-01 1.22622693e+00 -1.61221713e-01
3.87473494e-01 8.95187140e-01 2.47904956e-01 3.99393111e-01
-1.08917117e+00 -1.05335343e+00 3.41537386e-01 -1.70625001e-01
-1.57754791e+00 -3.40877920e-01 4.45083231e-01 8.41310769e-02
4.97948438e-01 9.67733383e-01 4.54275489e-01 -1.42166272e-01
-4.80146736e-01 8.49917889e-01 4.76552695e-01 -2.61598229e-01
6.14031911e-01 -3.43052559e-02 9.20216739e-02 2.80983776e-01
4.61887479e-01 -4.91307378e-01 -6.96873963e-01 -6.63411915e-01
1.20073572e-01 -3.61028649e-02 -4.03777957e-01 -4.92664464e-02
-8.44259083e-01 1.06732750e+00 1.56325568e-02 7.55390301e-02
-3.78441960e-01 2.35339820e-01 3.47960413e-01 4.45735306e-01
3.73524934e-01 2.84407914e-01 -3.01085889e-01 -2.44974047e-01
-1.51023507e+00 5.02127588e-01 9.92373943e-01 1.29337585e+00
6.59210980e-01 7.37154623e-03 -1.14094028e-02 8.90659451e-01
-1.75597873e-02 7.64339447e-01 7.12818503e-02 -1.46742427e+00
7.55173087e-01 5.90250850e-01 5.17272279e-02 -8.55588138e-01
-1.63005084e-01 -3.38211149e-01 -1.15270710e+00 -1.45516336e-01
4.77287710e-01 -2.94856697e-01 -2.48197600e-01 2.15875244e+00
3.41452301e-01 -1.13926806e-01 -1.49764895e-01 8.39622378e-01
5.22906959e-01 8.88845503e-01 -1.53793171e-01 -1.31723917e+00
1.05573523e+00 -1.85672626e-01 -3.57046932e-01 2.23582968e-01
3.12425762e-01 -5.91694176e-01 6.49330080e-01 5.25352597e-01
-1.80435026e+00 7.34659657e-02 -9.05823648e-01 5.83132029e-01
2.01331168e-01 -6.47818327e-01 5.32881439e-01 1.13669336e+00
-8.65804315e-01 2.54767388e-01 -2.56117553e-01 -1.93523914e-01
5.04313588e-01 5.07002771e-01 1.76187575e-01 -4.73873354e-02
-1.07305992e+00 -2.80692857e-02 1.68797210e-01 -5.73593318e-01
-4.74080086e-01 -7.71102846e-01 -4.20725495e-01 2.64516652e-01
7.54009128e-01 -5.88060200e-01 8.49703729e-01 -7.90191174e-01
-9.16098297e-01 8.26150179e-01 -1.90075815e-01 -5.23878336e-01
5.44439614e-01 4.79524791e-01 -1.73082501e-01 2.77921647e-01
3.09095904e-02 3.05671990e-01 3.50530446e-01 -1.20136559e+00
-1.25563753e+00 -4.84609544e-01 2.68087119e-01 2.75505990e-01
-5.40032208e-01 2.15577900e-01 -4.79427040e-01 -4.78809565e-01
2.10372657e-01 -5.15500188e-01 -5.10629535e-01 -1.31718501e-01
-3.50815058e-01 -2.00912908e-01 2.93758810e-01 -2.08135128e-01
2.03172183e+00 -2.02879786e+00 -1.02960160e-02 8.19535911e-01
2.30402783e-01 2.49288917e-01 6.97232410e-02 6.29176378e-01
4.70891386e-01 4.02571976e-01 -5.18052638e-01 -2.93027967e-01
1.26497120e-01 2.69271076e-01 -1.85995340e-01 5.43228507e-01
-7.68200636e-01 4.59624603e-02 -7.55603671e-01 -2.78014839e-01
-2.48470053e-01 -9.46802571e-02 -7.91171849e-01 2.58777857e-01
-3.91536921e-01 -1.60708264e-01 -3.80792260e-01 5.43974102e-01
1.08541882e+00 -1.06748119e-01 4.35422212e-01 2.82694876e-01
-2.48268187e-01 -5.65319397e-02 -1.81352270e+00 1.49649978e+00
-1.18436880e-01 -5.95186763e-02 5.28497338e-01 -1.39098787e+00
8.16466093e-01 7.75942132e-02 1.20949721e+00 -5.74563861e-01
1.78619340e-01 4.41194981e-01 -1.34909496e-01 -1.36233181e-01
8.57816637e-01 -2.00961739e-01 -5.85984051e-01 1.03561890e+00
-4.32856232e-01 1.03844583e-01 6.93404555e-01 6.06978834e-01
1.19069552e+00 -8.68300259e-01 2.51336128e-01 -4.58438784e-01
3.75709414e-01 -3.61817330e-01 9.11942959e-01 1.01504469e+00
-3.10939223e-01 5.09244084e-01 6.99957073e-01 -1.00215815e-01
-8.30198586e-01 -1.00949764e+00 9.39914808e-02 1.27620804e+00
5.72524488e-01 -4.91563708e-01 -8.08652818e-01 -4.57022071e-01
1.46665588e-01 7.90538907e-01 -3.82360667e-01 1.15547523e-01
-4.82760906e-01 -8.81658018e-01 5.89805663e-01 -9.16357636e-02
5.22180200e-01 -6.72219753e-01 -1.01893342e+00 3.41395259e-01
-4.96859044e-01 -5.81087649e-01 -8.40260088e-01 1.07707523e-01
-7.15867221e-01 -9.06738460e-01 -6.00449383e-01 -4.17817742e-01
5.64979792e-01 4.36782271e-01 1.04845965e+00 1.34661913e-01
-2.18043879e-01 3.05427790e-01 -4.03903633e-01 -3.85769278e-01
-1.57276038e-02 8.54957923e-02 -1.94609631e-02 -7.71910474e-02
2.79344052e-01 -7.16976404e-01 -9.30796027e-01 3.36509198e-01
-1.28563464e+00 -6.29921973e-01 -1.24610066e-01 3.31321836e-01
9.28319097e-01 6.18592501e-01 9.97195423e-01 -1.00001729e+00
6.31859601e-01 -8.32452357e-01 -3.15639257e-01 2.32371241e-01
-7.01350272e-01 -1.22030511e-01 9.17339861e-01 -1.38788313e-01
-4.52878505e-01 -3.33134979e-01 -6.05078824e-02 -1.09033003e-01
2.85067767e-01 5.40158868e-01 -3.83163214e-01 5.93539536e-01
4.46252704e-01 4.07426745e-01 1.60707578e-01 -2.26774186e-01
4.23423111e-01 6.81032240e-01 4.82864946e-01 -5.83081961e-01
2.25383669e-01 5.46278000e-01 -5.35772257e-02 -7.71984518e-01
-3.99679661e-01 -5.36189795e-01 2.46041298e-01 -1.09641925e-01
4.57210094e-03 -7.93850601e-01 -1.31240416e+00 3.43304873e-02
-3.04707408e-01 -1.23356603e-01 -8.28424394e-01 1.22537300e-01
-6.93970621e-01 6.33136332e-01 -7.23457262e-02 -1.32160246e+00
-7.72066176e-01 -7.15729535e-01 4.35672373e-01 4.16173816e-01
-1.67697757e-01 -3.62039715e-01 -3.02790463e-01 1.44649386e-01
4.77676839e-01 3.04004192e-01 9.90813613e-01 -6.66878343e-01
-4.92200017e-01 -2.25087658e-01 1.18759330e-02 1.59208968e-01
-3.35857319e-03 -1.49980098e-01 -1.70655698e-01 -5.87469220e-01
-1.68818742e-01 -4.21506492e-03 7.11643934e-01 6.05951905e-01
1.42801106e+00 -7.19894052e-01 -3.08562994e-01 5.50441384e-01
1.40403628e+00 4.29940969e-01 5.44242144e-01 -7.89644346e-02
-7.86693022e-02 5.36579251e-01 7.81316042e-01 1.57851601e+00
5.99857569e-01 4.27296817e-01 5.83263636e-01 3.93623918e-01
4.37335849e-01 2.68014938e-01 1.29115224e-01 5.64314902e-01
1.18932888e-01 -9.59837556e-01 -3.94497424e-01 7.98373103e-01
-1.85775983e+00 -1.21019530e+00 3.44824232e-02 2.93526745e+00
1.02032244e+00 6.96482435e-02 7.93807924e-01 5.86101472e-01
8.49177361e-01 1.41284674e-01 -7.12323487e-01 -7.77517438e-01
-3.92121732e-01 5.47050834e-01 6.56670392e-01 2.80536711e-01
-6.44634485e-01 2.30806544e-01 4.86494446e+00 1.23031044e+00
-8.35243583e-01 -2.98714578e-01 8.70553970e-01 -1.01379633e+00
-8.35195541e-01 -1.59225404e-01 -6.19860053e-01 1.08933902e+00
1.06594813e+00 -7.88187206e-01 6.10219479e-01 4.54895675e-01
3.38277400e-01 -6.64763331e-01 -9.86531973e-01 1.12998354e+00
1.03415497e-01 -1.15397739e+00 -5.85936084e-02 2.22676724e-01
8.53692055e-01 -4.44993049e-01 2.54027665e-01 -2.82174256e-02
3.42149943e-01 -7.30215669e-01 6.59598708e-01 2.39642739e-01
1.06430161e+00 -1.53081930e+00 2.89817840e-01 6.75091565e-01
-1.20755684e+00 -3.27254951e-01 -3.14352512e-01 1.27101108e-01
5.75522482e-01 1.05640388e+00 -1.24990739e-01 6.62423193e-01
1.00230157e+00 -8.13527554e-02 4.47563499e-01 9.98679876e-01
5.22564054e-01 5.19680262e-01 -8.64579380e-01 -1.13578416e-01
3.46122347e-02 -1.79540262e-01 6.35459363e-01 1.11906004e+00
7.41837859e-01 7.58252263e-01 3.99157763e-01 2.41792306e-01
-4.82066154e-01 5.82495332e-01 -1.61562189e-01 2.02695206e-01
1.20360994e+00 7.86589980e-01 -5.77321768e-01 -4.08523589e-01
5.15780561e-02 4.92669851e-01 -1.28280178e-01 4.90043871e-02
-8.34952295e-01 -8.46529067e-01 1.09555149e+00 4.67633635e-01
4.43197370e-01 -4.98237088e-02 -5.59522927e-01 -7.71266818e-01
7.56577328e-02 -8.47699881e-01 1.16166329e+00 9.87434387e-02
-1.23750627e+00 4.03223336e-01 7.57543221e-02 -1.11808991e+00
-1.61258563e-01 2.71235257e-01 -4.22370642e-01 8.82470131e-01
-1.18924892e+00 -2.72747010e-01 -1.30398840e-01 7.40080357e-01
2.41786137e-01 -4.70148623e-02 6.21040940e-01 5.34093797e-01
-4.00981307e-01 1.18722868e+00 4.27266449e-01 -5.32967508e-01
3.37079704e-01 -8.64111364e-01 -3.31112832e-01 7.53398180e-01
-4.09236938e-01 4.62963760e-01 8.09428513e-01 -3.07619959e-01
-1.26556611e+00 -9.19147372e-01 9.94503498e-01 3.08099568e-01
-1.30287096e-01 1.90978982e-02 -6.69314265e-01 4.33489569e-02
6.69996887e-02 -6.07144944e-02 1.18217957e+00 -2.43582800e-01
-2.07396716e-01 -5.51540911e-01 -1.72191143e+00 3.82970631e-01
1.34012270e+00 1.31131247e-01 1.50436148e-01 -8.23953450e-02
4.54793572e-01 -2.46169776e-01 -1.09958577e+00 4.80026245e-01
5.59216380e-01 -1.14813292e+00 7.97204792e-01 -3.60613197e-01
7.65885860e-02 -2.09554493e-01 -6.42612755e-01 -9.21969891e-01
-1.34505089e-02 -1.04961014e+00 -1.65737942e-01 1.31728399e+00
5.70184410e-01 -3.87535930e-01 1.03673637e+00 7.99514949e-01
1.82967633e-01 -8.60048532e-01 -1.23331320e+00 -6.80290103e-01
2.82937080e-01 -3.90620440e-01 9.49948430e-01 8.00436616e-01
4.33708012e-01 -1.53874621e-01 -4.53901917e-01 -8.63512158e-02
8.00684214e-01 5.69459856e-01 7.57959902e-01 -1.05392444e+00
-3.42920035e-01 -7.00469375e-01 1.07536279e-01 -1.18275905e+00
-2.29889840e-01 -8.51215899e-01 6.73671961e-02 -1.35891008e+00
3.55520099e-01 -1.12110543e+00 -3.23985845e-01 1.23731300e-01
-1.20045707e-01 -1.19503371e-01 2.70980120e-01 6.62060753e-02
-1.08056927e+00 3.64503175e-01 6.58553541e-01 -1.74077321e-02
-5.16983569e-01 3.71801496e-01 -1.11613655e+00 2.26438567e-01
7.83230126e-01 -4.06705022e-01 -6.12314641e-01 -2.96471059e-01
4.57035869e-01 5.94171226e-01 -6.29447818e-01 -6.43647730e-01
2.93587714e-01 -5.87458491e-01 -3.35962176e-01 -6.91812992e-01
1.55398637e-01 -6.42958403e-01 4.40173328e-01 4.94565248e-01
-5.14694870e-01 -5.27736321e-02 -1.48324907e-01 4.46554989e-01
-1.15557753e-01 -2.06694111e-01 8.45639527e-01 -8.72823820e-02
-1.41419768e-01 7.54708409e-01 -3.10662270e-01 6.05528414e-01
1.16466379e+00 -3.53548825e-01 -3.86959106e-01 -7.85921931e-01
-4.17888284e-01 8.16977561e-01 3.95268083e-01 -5.36615364e-02
4.86966342e-01 -1.10399461e+00 -1.02692294e+00 -1.33877955e-02
1.01689138e-01 2.10837230e-01 6.09157860e-01 5.86450577e-01
-3.12133759e-01 -3.74919176e-02 8.90847221e-02 -4.67388779e-01
-1.24834716e+00 5.19598186e-01 1.92443170e-02 -3.73732537e-01
6.56613708e-02 1.23632836e+00 -2.90431082e-01 5.66242523e-02
4.98439729e-01 -3.11311297e-02 1.37291983e-01 2.95832366e-01
6.26518846e-01 9.07884002e-01 -1.75233662e-01 -5.62169790e-01
-4.36282188e-01 1.66060656e-01 6.86277747e-02 -3.66709024e-01
1.18204308e+00 -5.79459548e-01 -2.82047063e-01 -4.30080928e-02
1.21131158e+00 1.90005213e-01 -8.72174323e-01 -5.35212040e-01
-3.46489787e-01 -7.61776626e-01 -5.10093212e-01 -5.39236128e-01
-1.38553178e+00 4.70198303e-01 2.86790073e-01 7.77726233e-01
1.68295157e+00 -6.78722113e-02 1.07974982e+00 -1.68401986e-01
7.67460465e-01 -1.28183091e+00 -1.81486011e-01 2.99216479e-01
3.98848802e-01 -7.19618976e-01 1.43206730e-01 -1.98695973e-01
-6.31819665e-01 7.63884127e-01 4.36020821e-01 6.92212358e-02
8.32097590e-01 1.83870047e-01 -2.76727140e-01 4.55421619e-02
-8.23160827e-01 -2.41829827e-01 -2.32331499e-01 1.89249858e-01
3.26459616e-01 4.72998410e-01 -1.20196104e+00 1.04843152e+00
-3.10722977e-01 -1.62597358e-01 5.87819576e-01 1.13808739e+00
-9.49731708e-01 -1.34627533e+00 -2.80232549e-01 9.42091286e-01
-7.78478265e-01 9.77835581e-02 8.58106390e-02 1.35505900e-01
4.39299971e-01 1.49130750e+00 2.79285729e-01 -1.86084241e-01
4.42731589e-01 -2.24562019e-01 2.61772007e-01 -2.82972217e-01
-8.84325147e-01 1.46535367e-01 7.74076059e-02 -4.28598285e-01
-3.23283136e-01 -9.18798506e-01 -1.65742755e+00 -1.03247201e+00
-1.90654635e-01 6.75745845e-01 4.89413351e-01 5.31463087e-01
3.78466934e-01 1.03469223e-01 1.22996676e+00 -1.19782882e-02
-6.43571377e-01 -5.10096133e-01 -1.14021647e+00 2.77256668e-01
-1.31753713e-01 -7.81551674e-02 -3.60675156e-01 -1.85821310e-01] | [6.553548336029053, 4.917885780334473] |
ab62e1a9-4e57-4a92-a801-7d559c612d23 | attribute-guided-encryption-with-facial | 2305.13548 | null | https://arxiv.org/abs/2305.13548v1 | https://arxiv.org/pdf/2305.13548v1.pdf | Attribute-Guided Encryption with Facial Texture Masking | The increasingly pervasive facial recognition (FR) systems raise serious concerns about personal privacy, especially for billions of users who have publicly shared their photos on social media. Several attempts have been made to protect individuals from unauthorized FR systems utilizing adversarial attacks to generate encrypted face images to protect users from being identified by FR systems. However, existing methods suffer from poor visual quality or low attack success rates, which limit their usability in practice. In this paper, we propose Attribute Guided Encryption with Facial Texture Masking (AGE-FTM) that performs a dual manifold adversarial attack on FR systems to achieve both good visual quality and high black box attack success rates. In particular, AGE-FTM utilizes a high fidelity generative adversarial network (GAN) to generate natural on-manifold adversarial samples by modifying facial attributes, and performs the facial texture masking attack to generate imperceptible off-manifold adversarial samples. Extensive experiments on the CelebA-HQ dataset demonstrate that our proposed method produces more natural-looking encrypted images than state-of-the-art methods while achieving competitive attack performance. We further evaluate the effectiveness of AGE-FTM in the real world using a commercial FR API and validate its usefulness in practice through an user study. | ['Rama Chellappa', 'Jiang Liu', 'Chun Pong Lau'] | 2023-05-22 | null | null | null | null | ['adversarial-attack'] | ['adversarial'] | [ 4.04400736e-01 9.01111066e-02 4.79385823e-01 -6.11884356e-01
-8.51519227e-01 -9.04442549e-01 6.17710471e-01 -5.44297934e-01
-9.53863338e-02 5.04072726e-01 -9.68561843e-02 -1.13266803e-01
2.44926378e-01 -7.72018552e-01 -6.69331074e-01 -6.16431475e-01
-1.32145450e-01 -4.14430439e-01 -4.32880044e-01 -2.20813572e-01
-1.82983652e-01 6.37255907e-01 -1.49103761e+00 4.46965814e-01
8.88289630e-01 1.21836793e+00 -6.58914804e-01 6.59559608e-01
4.41394955e-01 5.39821565e-01 -8.78942072e-01 -1.25392175e+00
9.21027243e-01 -3.00596237e-01 -5.43702066e-01 -7.95283467e-02
1.05257928e+00 -8.02030385e-01 -5.69354236e-01 1.27634609e+00
6.86043203e-01 -2.35093355e-01 4.07426715e-01 -1.73662031e+00
-1.18331623e+00 -3.04519683e-02 -4.96693194e-01 -3.18037540e-01
7.23457098e-01 4.25328434e-01 3.38575542e-01 -8.46928358e-01
1.46942005e-01 1.57598877e+00 7.41791487e-01 1.10554135e+00
-1.31949008e+00 -1.23186171e+00 -2.43959889e-01 -1.19579308e-01
-1.82081366e+00 -8.29288721e-01 5.37879467e-01 -1.90253988e-01
3.97694141e-01 6.20285451e-01 3.58327746e-01 1.23681259e+00
2.50065088e-01 3.94995630e-01 1.40477765e+00 -4.41235341e-02
-6.28957525e-02 4.33437020e-01 -3.51870716e-01 7.14533269e-01
-8.89818743e-02 4.13482726e-01 -4.80162263e-01 -7.58225679e-01
8.13346207e-01 -6.81497604e-02 -4.83619601e-01 1.27456650e-01
-6.50378287e-01 8.03088248e-01 4.23884362e-01 -3.37614924e-01
-1.95271134e-01 1.16077669e-01 1.19920433e-01 4.82422322e-01
5.46853900e-01 2.00108394e-01 1.56259481e-02 2.53894717e-01
-6.52644932e-01 3.30452532e-01 5.66922843e-01 9.01194334e-01
6.89580619e-01 7.56621286e-02 -1.59556434e-01 8.63470852e-01
3.05887610e-01 9.33319986e-01 -2.95679308e-02 -9.58526134e-01
2.84552336e-01 3.32885981e-01 8.85884538e-02 -1.38183165e+00
3.41104865e-01 2.02011690e-01 -9.44427013e-01 4.98386681e-01
2.38436699e-01 -3.88049990e-01 -8.78607273e-01 1.90130186e+00
3.34531099e-01 3.09689969e-01 3.13743114e-01 7.47228980e-01
6.72404289e-01 6.06216311e-01 2.87003398e-01 2.07267702e-01
1.37926912e+00 -3.12198043e-01 -6.90925479e-01 4.06419113e-02
1.66128114e-01 -1.01002491e+00 1.14827228e+00 4.42397416e-01
-9.84195232e-01 -7.08046854e-01 -1.02612448e+00 2.12512210e-01
-2.87162125e-01 1.06707789e-01 6.17993832e-01 1.60502911e+00
-1.16990852e+00 4.53800827e-01 -4.23683703e-01 -1.10210896e-01
9.38886225e-01 7.69706249e-01 -9.11597192e-01 -1.05232365e-01
-1.36778867e+00 3.30330729e-01 -2.98615605e-01 2.03842282e-01
-9.70973313e-01 -8.98895979e-01 -9.54987466e-01 -1.15519442e-01
-1.69879291e-02 -5.10567069e-01 9.31743443e-01 -1.32260716e+00
-1.64253807e+00 9.17087913e-01 1.62055537e-01 -4.37759101e-01
6.52925372e-01 -4.18040484e-01 -9.94902849e-01 3.30515951e-01
-1.75804749e-01 8.77031446e-01 1.59230494e+00 -1.12270331e+00
-2.61172354e-01 -5.19616544e-01 2.95614183e-01 -5.99513426e-02
-8.16114366e-01 2.02917144e-01 3.98763977e-02 -9.23611462e-01
-5.27732313e-01 -1.00929666e+00 1.28144205e-01 3.79764825e-01
-3.00750762e-01 2.22909346e-01 1.33504033e+00 -6.90084815e-01
8.78834844e-01 -2.46891832e+00 -2.81232566e-01 3.72107923e-01
1.73447490e-01 6.29048586e-01 -2.06082001e-01 2.51725972e-01
-2.27772549e-01 4.60206896e-01 -8.45912844e-02 -3.24545681e-01
-1.70150772e-02 -1.22093976e-01 -5.92606366e-01 6.60596609e-01
4.79832828e-01 8.03506076e-01 -5.45605123e-01 -2.03872606e-01
7.46472329e-02 1.09788442e+00 -6.68458223e-01 4.25322264e-01
1.92281604e-01 3.67313743e-01 -3.47532481e-01 7.64777780e-01
1.23458135e+00 2.78386176e-01 -1.10903054e-01 -2.40689233e-01
5.27824461e-01 -4.81699228e-01 -1.10877419e+00 9.29358363e-01
-2.32895389e-01 4.63882357e-01 2.83276647e-01 5.75806247e-03
8.40853095e-01 4.09660637e-01 7.27229416e-02 -4.40203995e-01
1.88867867e-01 -2.54749060e-02 -2.01181382e-01 -1.84376791e-01
3.83489728e-01 -1.10284917e-01 -7.38648474e-02 2.96884596e-01
-1.49927482e-01 2.06827790e-01 -5.90539098e-01 2.19121337e-01
9.87025976e-01 -7.80453086e-02 -1.70518890e-01 -3.11031848e-01
9.25654113e-01 -7.58367419e-01 2.57336348e-01 4.58204210e-01
-3.12966108e-01 7.40518153e-01 2.08722964e-01 -4.59954888e-01
-8.40090215e-01 -1.30393839e+00 -1.42357677e-01 5.93897045e-01
1.62939936e-01 -5.10861218e-01 -1.32616544e+00 -1.10499358e+00
2.46820152e-01 2.89584756e-01 -7.32215166e-01 -3.84209633e-01
-3.52484047e-01 -5.00483453e-01 1.27469766e+00 2.33808532e-01
1.13858986e+00 -9.37591553e-01 -9.80616286e-02 -2.15302870e-01
-3.97885218e-02 -1.22303069e+00 -8.90097558e-01 -1.05937719e+00
-1.05335802e-01 -1.13701582e+00 -7.90708363e-01 -5.37733674e-01
9.78441536e-01 1.92516029e-01 5.82036436e-01 2.71894366e-01
-5.69833696e-01 6.71643853e-01 -5.06255552e-02 -4.50430125e-01
-6.49418116e-01 -3.63331407e-01 4.51706171e-01 9.26933348e-01
3.59272510e-01 -2.32629105e-01 -9.41112339e-01 6.85635686e-01
-1.08886647e+00 -4.98096168e-01 2.35166296e-01 6.36134744e-01
9.10538957e-02 2.17571884e-01 5.76567948e-01 -9.97683644e-01
7.44307041e-01 -1.04224771e-01 -6.12445772e-01 1.94195867e-01
-3.01329315e-01 -3.82964522e-01 8.25550437e-01 -6.08226061e-01
-1.08681178e+00 -1.85602363e-02 -4.02250767e-01 -5.59356332e-01
-1.44673705e-01 -3.02688330e-01 -5.71906686e-01 -8.09682965e-01
7.39989758e-01 2.61910141e-01 4.00267690e-01 -7.86821619e-02
2.91127771e-01 9.91366506e-01 5.70507586e-01 -4.81159389e-01
1.35466933e+00 6.11706078e-01 -1.04329750e-01 -9.55879986e-01
-3.65837902e-01 2.38123015e-01 -1.54851153e-01 -3.57904673e-01
7.80652404e-01 -1.12511802e+00 -1.21610856e+00 1.13305867e+00
-7.65449166e-01 -1.26147158e-02 2.14954451e-01 -5.43671623e-02
-4.07171577e-01 5.61092794e-01 -7.20520914e-01 -8.92327070e-01
-6.41362965e-01 -1.13629568e+00 1.14635432e+00 2.38152593e-01
-1.31569564e-01 -6.37781322e-01 -4.62916940e-01 4.92765605e-01
6.12808287e-01 6.59198344e-01 3.23607773e-01 -8.69556144e-02
-6.55019403e-01 -3.82362157e-01 -1.16109163e-01 8.42448890e-01
5.47238708e-01 -1.71726290e-02 -1.27005744e+00 -6.90943062e-01
-8.14984203e-04 -3.59831214e-01 3.15711856e-01 -2.52153665e-01
1.42619443e+00 -8.10067713e-01 -4.20777313e-02 9.30452049e-01
1.22657430e+00 -5.42431176e-02 1.23227429e+00 -1.03170358e-01
6.23249233e-01 5.52417219e-01 6.34421349e-01 4.02435482e-01
-1.53539795e-02 5.55958450e-01 3.46189797e-01 -2.52947450e-01
1.01422884e-01 -4.15818006e-01 6.66447222e-01 -1.77456409e-01
7.03808218e-02 -1.47496402e-01 -4.14210707e-01 7.06672668e-02
-1.24827969e+00 -1.04490519e+00 2.94083267e-01 2.30719948e+00
7.02924728e-01 -1.73598781e-01 2.27692798e-01 2.80362405e-02
8.59764993e-01 6.98028281e-02 -2.39020407e-01 -3.49302202e-01
-1.20991401e-01 4.25451934e-01 4.96208966e-01 4.33570564e-01
-1.30813742e+00 1.01019251e+00 5.87391186e+00 7.80670524e-01
-1.19562602e+00 1.03870898e-01 9.94666815e-01 7.08743744e-03
-1.11895390e-01 -4.04804796e-01 -9.15864885e-01 4.81264204e-01
9.08278644e-01 -1.20369822e-01 6.66993141e-01 8.32392931e-01
-1.07005537e-01 5.05475461e-01 -9.50067163e-01 1.21083641e+00
2.27700353e-01 -1.23453212e+00 3.61064523e-01 4.86315936e-01
5.49177170e-01 -6.76396906e-01 7.09660470e-01 -2.65575573e-02
1.65942118e-01 -1.39626372e+00 5.12294829e-01 3.72576952e-01
1.42679608e+00 -1.10988975e+00 4.80557263e-01 -1.99497208e-01
-1.15840447e+00 -5.78014441e-02 -4.15633172e-01 3.08783472e-01
-1.41246319e-01 1.21879302e-01 -6.49121761e-01 4.28919554e-01
8.19746852e-01 2.99580306e-01 -6.27776623e-01 2.77450323e-01
-1.38367504e-01 4.09871519e-01 -3.83085191e-01 2.71221936e-01
2.76017711e-02 -5.52328154e-02 4.77623940e-01 8.11507821e-01
2.89513648e-01 3.30296606e-01 -3.06707710e-01 8.58070672e-01
-4.33414787e-01 4.75149415e-02 -9.03473437e-01 -7.67985657e-02
4.56983626e-01 1.37892628e+00 1.70820449e-02 5.73225878e-02
-2.38900319e-01 1.47796202e+00 1.05778407e-03 2.92250276e-01
-1.04254639e+00 -3.24325204e-01 1.35008061e+00 4.20215666e-01
-6.03614748e-02 -4.99627031e-02 3.61556381e-01 -1.04630077e+00
9.91369709e-02 -1.45204175e+00 2.84366071e-01 -5.76331973e-01
-1.50940812e+00 1.05528951e+00 -3.95004988e-01 -1.24069667e+00
-7.96718337e-03 -4.21341330e-01 -3.18223178e-01 1.16615784e+00
-1.17491639e+00 -1.47954214e+00 -3.69041175e-01 1.24479818e+00
4.49430794e-02 -5.95992148e-01 1.08990753e+00 3.84765804e-01
-3.87801200e-01 1.47958386e+00 -2.86430031e-01 4.44654286e-01
8.25316370e-01 -9.05235350e-01 8.17522764e-01 8.35168362e-01
-7.65929148e-02 9.28927958e-01 3.00526917e-01 -6.79673612e-01
-1.63629150e+00 -1.48904085e+00 2.15576559e-01 -5.71418822e-01
4.86889072e-02 -7.80542552e-01 -8.03547084e-01 6.51408613e-01
2.94637799e-01 3.26268524e-01 9.54515755e-01 -6.31073236e-01
-7.37349391e-01 -4.00794744e-01 -2.08637047e+00 7.75773108e-01
9.19467986e-01 -8.77692342e-01 -2.19111834e-02 1.85646057e-01
4.77479488e-01 -1.54216632e-01 -1.14882648e+00 5.13246357e-01
7.89165318e-01 -9.88840878e-01 1.09927702e+00 -4.09805268e-01
1.18887946e-01 -4.09749568e-01 -2.94018954e-01 -1.04764223e+00
4.69553657e-02 -1.47367561e+00 9.43969041e-02 1.67374110e+00
3.83869372e-02 -1.09367275e+00 8.67978632e-01 9.43337262e-01
6.06651902e-01 -3.99931133e-01 -7.68879056e-01 -7.33540535e-01
-1.35952681e-01 -1.21099412e-01 1.05219340e+00 9.61977124e-01
-4.80912328e-01 -2.27107242e-01 -6.75556958e-01 7.13078082e-01
9.62926865e-01 -5.77976704e-01 1.17817223e+00 -7.41714180e-01
-1.30972475e-01 1.88554719e-01 -7.22607374e-01 -6.39221847e-01
2.51723588e-01 -5.07410347e-01 -4.06690449e-01 -3.80845547e-01
-1.28812417e-01 -4.31987137e-01 1.50089888e-02 5.17149627e-01
-2.17674568e-01 1.06753123e+00 3.25881124e-01 -3.13986354e-02
4.44114543e-02 5.93059361e-01 9.24796045e-01 -1.56635091e-01
1.45698711e-01 -6.25949576e-02 -9.17886853e-01 6.45342469e-01
6.28809512e-01 -2.64821231e-01 -5.66891909e-01 -1.31042227e-01
-6.05456457e-02 -2.27990508e-01 7.00916529e-01 -1.14436436e+00
-2.07978055e-01 1.44685939e-01 7.48893857e-01 2.61533111e-02
5.80777764e-01 -9.93404031e-01 4.69732583e-01 3.23988795e-01
-8.83958712e-02 1.55375928e-01 3.92775834e-01 3.76438648e-01
-5.54360412e-02 4.29617584e-01 1.14218748e+00 8.13459605e-02
-3.41370225e-01 7.79342890e-01 -6.29951525e-03 -2.98084557e-01
1.32191968e+00 -2.61946887e-01 -5.01084566e-01 -7.58685887e-01
-6.23329461e-01 -3.49119939e-02 7.07601964e-01 5.29871047e-01
1.06404674e+00 -1.49081898e+00 -9.15465832e-01 9.97500002e-01
6.32198229e-02 -6.47711992e-01 3.87531251e-01 8.41438174e-02
-6.39635921e-01 -1.15369640e-01 -3.97589803e-01 -4.94378239e-01
-1.88474011e+00 6.16030574e-01 5.73782980e-01 3.81130308e-01
-4.49543595e-01 7.99728036e-01 4.59850550e-01 -3.48131806e-01
3.27904895e-02 5.44797659e-01 2.21685469e-01 -3.95705104e-01
1.05797446e+00 1.36325940e-01 -8.74549598e-02 -9.76915121e-01
-3.04772824e-01 5.57980359e-01 -2.12553740e-01 -3.49191912e-02
9.82212305e-01 -1.20693207e-01 -9.17985737e-02 -6.47156894e-01
1.45474911e+00 3.58499259e-01 -1.29628217e+00 7.73099214e-02
-7.49618351e-01 -1.23856199e+00 -2.66902208e-01 -5.71210682e-01
-1.26716924e+00 5.91251016e-01 9.73301947e-01 3.45892727e-01
1.43025529e+00 -4.42399025e-01 9.60476995e-01 3.58003862e-02
5.11247158e-01 -4.18788105e-01 9.68834907e-02 -1.56247884e-01
1.09963930e+00 -1.05366647e+00 -1.54012412e-01 -9.96917486e-01
-5.52725494e-01 8.67381573e-01 6.51169598e-01 -2.51067877e-01
8.03084910e-01 2.88041860e-01 4.52680558e-01 8.54738578e-02
-2.85975128e-01 4.71406341e-01 2.77657300e-01 8.62441361e-01
-2.16904115e-02 -2.00506542e-02 1.65419593e-01 4.31543142e-01
-5.03605068e-01 -1.88512713e-01 3.88633192e-01 7.47488916e-01
3.50918055e-01 -1.42016673e+00 -6.97461307e-01 2.61844158e-01
-1.06439435e+00 -3.60693485e-02 -6.68024957e-01 7.56597698e-01
2.65114326e-02 1.03154612e+00 -1.34104311e-01 -6.73561990e-01
2.05314040e-01 -1.20214656e-01 6.27956390e-01 -2.44953603e-01
-9.37278807e-01 -2.41827637e-01 -6.57071620e-02 -7.46726573e-01
-2.22104579e-01 -5.68320692e-01 -6.52989984e-01 -9.11181033e-01
-7.77459145e-02 -8.40678997e-03 4.26357627e-01 5.17307401e-01
6.94742680e-01 -1.36628032e-01 1.23253870e+00 -5.24546981e-01
-6.55953825e-01 -5.83854556e-01 -4.76947010e-01 7.53062129e-01
4.52667058e-01 -3.19017738e-01 -3.41287464e-01 1.50280073e-01] | [12.87615966796875, 0.8965027928352356] |
4ce5595c-c379-4782-95dd-e547f54d480b | correcting-prepositional-phrase-attachments | null | null | https://aclanthology.org/W17-6311 | https://aclanthology.org/W17-6311.pdf | Correcting prepositional phrase attachments using multimodal corpora | PP-attachments are an important source of errors in parsing natural language. We propose in this article to use data coming from a multimodal corpus, combining textual, visual and conceptual information, as well as a correction strategy, to propose alternative attachments in the output of a parser. | ['Sebastien Delecraz', 'Frederic Bechet', 'Alexis Nasr', 'Benoit Favre'] | 2017-09-01 | null | null | null | ws-2017-9 | ['prepositional-phrase-attachment'] | ['natural-language-processing'] | [ 1.42020341e-02 4.26780105e-01 3.89513552e-01 -4.68919814e-01
-5.24231553e-01 -7.60977507e-01 4.97165799e-01 9.49338794e-01
-5.19992769e-01 6.51547015e-01 5.68164825e-01 -3.14504653e-01
1.35747656e-01 -5.99837840e-01 -3.23088259e-01 1.40671343e-01
3.41984451e-01 5.80869853e-01 6.98724926e-01 -4.72312510e-01
5.19472003e-01 6.04502916e-01 -1.20871687e+00 9.53791797e-01
6.97938740e-01 1.36112317e-01 4.81066316e-01 4.32532698e-01
-1.03159976e+00 5.32670081e-01 -8.03277791e-01 -1.11075115e+00
-5.16822994e-01 -3.79547238e-01 -1.14776230e+00 -1.02257863e-01
1.20724611e-01 3.12551290e-01 3.94533604e-01 1.04167688e+00
3.37102652e-01 -1.96979612e-01 6.24749541e-01 -6.80343091e-01
-3.78467470e-01 1.06061280e+00 -3.33917230e-01 4.10683185e-01
1.07842159e+00 -1.82247370e-01 9.69738364e-01 -1.17569423e+00
1.32093024e+00 1.67776358e+00 5.41874170e-01 6.05995238e-01
-1.24364245e+00 2.24282622e-01 -5.49963415e-02 1.60985082e-01
-9.63844061e-01 -5.33851743e-01 7.51578808e-01 -4.45244551e-01
1.45055676e+00 2.39203259e-01 5.67423344e-01 1.21993971e+00
2.59823233e-01 6.95324957e-01 9.90527928e-01 -1.36864436e+00
-1.21181337e-02 4.26786393e-01 1.46844268e-01 6.59408033e-01
2.43360654e-01 -2.36669078e-01 -7.25033522e-01 -3.19559187e-01
2.53340483e-01 -5.65759897e-01 -1.57827344e-02 1.20403603e-01
-7.67616212e-01 7.34709859e-01 2.25555554e-01 8.84959936e-01
-4.12856430e-01 -2.42636114e-01 3.80544335e-01 1.81200758e-01
4.70194906e-01 5.73418081e-01 -2.38902077e-01 -2.86897004e-01
-2.28788108e-01 1.32519603e-01 8.18747401e-01 7.04207182e-01
5.33357680e-01 -4.14059311e-01 -6.10546991e-02 7.95663714e-01
8.20507705e-01 2.20592722e-01 4.83739793e-01 -6.46441877e-01
1.01904023e+00 8.98261368e-01 1.21785171e-01 -1.28813064e+00
-7.18095362e-01 2.25004509e-01 -1.59312673e-02 -3.93043421e-02
3.93354148e-01 -7.54557326e-02 -4.39400077e-01 1.53364456e+00
4.28564996e-01 -7.32383668e-01 4.08445686e-01 4.17765975e-01
9.84543383e-01 3.77626538e-01 5.34016013e-01 -4.54311132e-01
1.50057983e+00 -4.27068204e-01 -9.39082205e-01 -4.46117669e-01
8.24083686e-01 -1.30868638e+00 1.14382565e+00 2.56197393e-01
-1.32428396e+00 -3.14245969e-01 -8.41138840e-01 -1.56234562e-01
-4.68374908e-01 2.29173824e-01 1.15449205e-01 3.92644018e-01
-8.37573111e-01 6.43684745e-01 -5.69847047e-01 -7.59042561e-01
-8.01482499e-02 -7.48890452e-03 -5.05010605e-01 1.12622954e-01
-8.78161848e-01 1.24166894e+00 5.02148569e-01 2.08560795e-01
4.22941208e-01 3.41855213e-02 -7.25865185e-01 -1.29909322e-01
1.58512652e-01 -6.63822770e-01 1.28017664e+00 -9.62912440e-01
-1.33789051e+00 1.23140359e+00 -2.11810380e-01 7.67170489e-02
2.69497037e-01 -1.55481130e-01 -4.66104180e-01 2.97852874e-01
9.31596458e-02 4.99691874e-01 5.72341681e-01 -1.27934015e+00
-6.60665512e-01 -8.48997653e-01 1.00222871e-01 5.00873774e-02
1.01762302e-01 7.68045604e-01 -1.94864392e-01 -6.58345640e-01
5.06171286e-01 -5.70373178e-01 -3.96909229e-02 -6.39071643e-01
-3.62558395e-01 -6.31505370e-01 3.85312110e-01 -1.10627580e+00
1.11325622e+00 -2.19620800e+00 7.06690013e-01 2.86627352e-01
1.19845346e-01 1.14659861e-01 -1.24419145e-01 7.54621446e-01
-6.90100715e-02 5.12170196e-01 -3.03538114e-01 -7.72066832e-01
-1.46674970e-02 6.20735228e-01 3.33293267e-02 -1.45178452e-01
5.30004740e-01 7.03652322e-01 -8.14721584e-01 -8.10798645e-01
1.37810037e-02 2.51984358e-01 -1.78791881e-01 2.02611968e-01
-1.99672565e-01 2.32879773e-01 -4.11178321e-01 5.84210753e-01
3.88926595e-01 5.33013046e-01 4.94921416e-01 -2.04112738e-01
-3.97213250e-01 5.57184696e-01 -7.46276736e-01 1.63717055e+00
-2.65795022e-01 5.94728529e-01 -1.82587788e-01 -6.86232090e-01
9.23761725e-01 3.64883244e-01 -3.57174098e-01 -7.87113845e-01
2.22793281e-01 2.62351125e-01 -5.78347147e-02 -9.44158494e-01
6.67479396e-01 -1.33731961e-01 -1.64864257e-01 1.63913429e-01
3.46698105e-01 9.76701826e-02 6.55158937e-01 3.12769562e-01
7.86779284e-01 2.47920498e-01 4.94538873e-01 8.24550632e-04
7.18865037e-01 1.04817003e-01 3.94758463e-01 4.29158747e-01
2.09581509e-01 4.22749341e-01 1.17636836e+00 -5.39951742e-01
-1.05460262e+00 -6.68760061e-01 1.98796839e-01 1.06494927e+00
-2.85126597e-01 -7.28648484e-01 -9.65564132e-01 -7.59388983e-01
-6.37432158e-01 9.01507437e-01 -3.51292312e-01 1.81931630e-01
-9.08192277e-01 -5.79129934e-01 4.98614222e-01 5.05505502e-01
-2.88969994e-01 -1.30404317e+00 -7.29895890e-01 4.91555810e-01
-3.68593872e-01 -1.18462002e+00 2.44672194e-01 1.19069904e-01
-9.55586672e-01 -9.87424016e-01 -1.32339329e-01 -5.82885742e-01
7.17224240e-01 -4.22478378e-01 1.35135400e+00 6.47201478e-01
-3.06110941e-02 1.63110077e-01 -9.80784178e-01 -3.27382326e-01
-1.11534965e+00 -4.31088582e-02 -2.74369895e-01 -2.37744927e-01
3.36728424e-01 -2.41566062e-01 4.09415215e-01 -2.52204865e-01
-7.53999531e-01 1.33482620e-01 5.54398715e-01 5.29824317e-01
9.81162414e-02 -8.01325560e-01 -2.76323147e-02 -1.04903960e+00
8.16825211e-01 -2.65247285e-01 -3.68981063e-01 5.97823203e-01
-3.12268168e-01 1.98811516e-01 3.57293636e-01 -7.84873143e-02
-1.45098639e+00 2.02115644e-02 -7.93612003e-01 2.93533921e-01
-4.83329833e-01 6.93383455e-01 -4.16795135e-01 3.24653052e-02
7.34265089e-01 -2.74980336e-01 -3.11537445e-01 -9.33766127e-01
3.00269097e-01 3.82779211e-01 3.84111375e-01 -6.00540221e-01
2.17756763e-01 -1.77648872e-01 3.27807367e-02 -9.11079586e-01
-3.78727287e-01 -1.75665170e-01 -9.57141757e-01 -1.98288023e-01
6.86647832e-01 -2.15638608e-01 5.08517511e-02 9.77275968e-02
-2.04182696e+00 1.98566183e-01 -2.59924442e-01 4.52677786e-01
-2.05981448e-01 5.18833578e-01 -7.72726476e-01 -6.81658030e-01
-1.56944573e-01 -9.70026374e-01 6.72693193e-01 1.47954360e-01
-4.79845256e-01 -9.70230639e-01 2.47363001e-01 6.73109740e-02
-4.51786770e-03 -2.03242376e-02 1.26709390e+00 -7.39959955e-01
5.85749671e-02 8.95733163e-02 -6.98513091e-02 -1.69925302e-01
-1.92105949e-01 6.09305859e-01 -5.91325998e-01 4.05441612e-01
-1.51993424e-01 1.80772250e-03 6.27301395e-01 -9.75156948e-02
2.26401359e-01 -3.96605521e-01 -4.88921791e-01 4.64780554e-02
1.32327485e+00 2.92166352e-01 5.46363354e-01 1.92110419e-01
5.70550084e-01 1.26496112e+00 5.55207551e-01 3.50182742e-01
3.47214758e-01 9.53501523e-01 3.59741867e-01 2.78521240e-01
-7.51302168e-02 -2.00742930e-01 3.73584092e-01 7.98714757e-01
-1.77166998e-01 -4.84554350e-01 -1.26751161e+00 3.65555257e-01
-1.75486219e+00 -7.48773634e-01 -9.57813263e-01 1.62089431e+00
6.02201998e-01 2.67111003e-01 1.04940750e-01 1.67388827e-01
7.16745198e-01 -1.62404895e-01 6.33474827e-01 -9.70461249e-01
-2.27220640e-01 5.88174984e-02 -8.57280008e-03 8.86444211e-01
-8.25214565e-01 1.10923302e+00 6.78577185e+00 4.04083252e-01
-8.79235804e-01 2.66345590e-01 -5.32658100e-02 6.28696322e-01
-8.63928497e-01 1.63627286e-02 -6.98095918e-01 3.34824264e-01
1.21274090e+00 3.63373458e-01 -3.75091820e-03 3.62975836e-01
-3.63376811e-02 -1.94832534e-01 -9.69141185e-01 5.75621009e-01
4.01689917e-01 -1.26747632e+00 1.81567922e-01 -5.73707819e-01
-3.27485174e-01 -1.52372882e-01 -4.39152926e-01 -1.02326654e-01
-2.12317124e-01 -5.98256707e-01 1.09189653e+00 6.62881911e-01
2.10727230e-01 -4.11269367e-01 1.10059118e+00 1.82660639e-01
-9.60876048e-01 2.46536106e-01 -1.64256468e-01 -3.20025086e-01
3.17248106e-01 -1.67657845e-02 -5.00465453e-01 6.41400099e-01
6.11299038e-01 4.73704249e-01 -1.12434101e+00 8.70069981e-01
-7.64792264e-01 3.17741275e-01 -3.26196074e-01 -4.89912152e-01
-6.51066899e-02 -4.28518429e-02 7.97698498e-01 1.59289408e+00
2.41280660e-01 2.97829092e-01 -1.52195960e-01 6.00953341e-01
3.90632659e-01 6.91890955e-01 -8.26659918e-01 -1.48780331e-01
3.80144238e-01 1.19242907e+00 -1.20806265e+00 -2.46759266e-01
-5.81606925e-01 9.79144633e-01 6.20569468e-01 9.89385322e-02
-1.63465291e-01 -2.79278994e-01 9.81203988e-02 -1.50216490e-01
6.27062917e-02 -2.71563083e-01 -7.05751078e-03 -1.07786632e+00
2.48455063e-01 -4.21502888e-01 6.39310718e-01 -1.16564775e+00
-1.10140729e+00 9.58898664e-01 2.14810789e-01 -8.73801529e-01
-4.86706704e-01 -9.24905002e-01 -5.11543393e-01 8.66033375e-01
-9.76673484e-01 -1.21989191e+00 2.49158412e-01 2.55308062e-01
4.96912926e-01 -1.91870376e-01 1.03816903e+00 1.39252767e-01
-6.33983076e-01 1.82291001e-01 -4.69811320e-01 5.53224422e-02
4.40371811e-01 -1.47210693e+00 4.53473151e-01 1.13023448e+00
5.30663669e-01 3.76086771e-01 1.04845166e+00 -4.44192529e-01
-8.87651622e-01 -2.59782940e-01 1.83706617e+00 -4.99944478e-01
5.59451938e-01 1.47273662e-02 -1.11848032e+00 6.28969073e-01
6.24401093e-01 -2.67199904e-01 6.65256858e-01 1.25854135e-01
-1.78233370e-01 1.38317257e-01 -1.19825840e+00 3.94431949e-01
8.80825639e-01 -3.34833741e-01 -1.58003497e+00 1.43215759e-02
6.55944407e-01 -7.87295222e-01 -4.67810035e-01 -3.28264348e-02
2.46183842e-01 -8.63211513e-01 7.05893159e-01 -6.89466715e-01
5.10981739e-01 -1.16417240e-02 -9.25047845e-02 -1.40642309e+00
-1.43921375e-01 -5.14333487e-01 1.66769058e-01 1.62190294e+00
8.20132792e-01 -2.76038527e-01 1.66111976e-01 4.64333624e-01
-4.77691889e-01 -9.58520398e-02 -1.25200880e+00 -2.05397736e-02
-1.50778010e-01 -5.13927639e-01 2.09251195e-01 4.79862273e-01
5.98776639e-01 7.07964420e-01 -2.61869431e-02 1.02852844e-01
1.07903540e-01 -3.48036140e-01 4.25352961e-01 -1.60778046e+00
-2.90480405e-02 -4.05074120e-01 -2.55626321e-01 -2.88759559e-01
2.47333631e-01 -5.31435728e-01 1.40093982e-01 -1.58769906e+00
-1.64379910e-01 -1.47051141e-01 9.54852030e-02 6.82553411e-01
1.45226985e-01 2.47120544e-01 6.15605116e-01 2.32354626e-01
-4.87520605e-01 2.05006704e-01 5.47903180e-01 1.34679854e-01
-2.10156977e-01 -4.12707567e-01 -5.14447570e-01 1.45974588e+00
6.97608054e-01 -8.83236289e-01 2.29081735e-01 -7.85069942e-01
9.88820434e-01 3.95055681e-01 2.08137706e-01 -4.53315943e-01
2.43928153e-02 -1.30410835e-01 2.87150770e-01 -4.48412210e-01
2.62415081e-01 -9.20569718e-01 -5.88027015e-02 5.74626267e-01
-1.58979371e-01 8.14172089e-01 3.36297303e-01 2.40464270e-01
-1.94910690e-01 -1.12634528e+00 2.60436893e-01 -3.83212090e-01
-8.67361426e-01 -6.27868831e-01 -8.01184297e-01 9.43112522e-02
7.11187840e-01 -5.66000864e-02 -4.23263967e-01 3.47086608e-01
-1.10760760e+00 -3.16969961e-01 4.54004705e-01 3.71979415e-01
7.56991506e-01 -1.07451713e+00 -5.67768514e-01 1.87192142e-01
3.21056068e-01 -6.12619996e-01 -3.12244929e-02 8.16342056e-01
-9.10029888e-01 1.51025504e-01 -5.08325160e-01 -3.34635973e-01
-1.59329093e+00 3.31689060e-01 5.83325773e-02 7.27187544e-02
-5.66479623e-01 8.31101537e-01 -6.16278350e-01 -2.34001428e-01
-1.53474852e-01 -5.17175436e-01 -9.35498118e-01 3.34728688e-01
4.70209837e-01 -3.23160924e-02 2.51478791e-01 -1.15222776e+00
-7.27950573e-01 5.75091243e-01 2.64211863e-01 -5.49033344e-01
1.02293789e+00 -6.34066701e-01 -5.64979136e-01 7.83158898e-01
5.56325674e-01 6.40795887e-01 -2.82387197e-01 -7.03046564e-03
7.06505299e-01 -4.71283197e-01 -4.69334692e-01 -8.63903940e-01
-4.74495918e-01 9.91665781e-01 3.11384499e-01 5.97909570e-01
6.81022048e-01 5.53118348e-01 2.82656074e-01 3.88969064e-01
4.82279919e-02 -1.08568788e+00 -1.15212388e-01 6.43495739e-01
1.17784524e+00 -1.08692038e+00 -2.29325280e-01 -7.79728830e-01
-5.73354363e-01 1.76448083e+00 5.54794550e-01 2.79385298e-01
1.75428256e-01 1.92095295e-01 2.02311680e-01 -2.13703051e-01
-8.95868242e-01 -2.41819382e-01 1.64879590e-01 7.08323777e-01
8.64928186e-01 -1.22232355e-01 -1.30525172e+00 8.56381714e-01
5.96437789e-02 -2.67452329e-01 9.02077258e-01 8.82846594e-01
-2.79486835e-01 -1.70708585e+00 -4.58207250e-01 -1.67731136e-01
-6.86471283e-01 -1.27693698e-01 -8.12691092e-01 7.03550458e-01
5.84247589e-01 9.16826725e-01 3.19374800e-01 -1.02506392e-01
5.79101443e-01 2.82733649e-01 8.03035498e-01 -7.86991656e-01
-7.95749903e-01 -5.77746611e-03 6.45724535e-01 -5.09508550e-01
-9.15256679e-01 -8.50095391e-01 -1.35155284e+00 -9.75826010e-02
8.10983703e-02 1.67420030e-01 6.74949467e-01 1.43895340e+00
1.57657951e-01 3.30735832e-01 -9.83741209e-02 -7.19634354e-01
7.97148794e-03 -1.11684191e+00 -1.06407344e-01 4.28296059e-01
-2.07320541e-01 -5.70105612e-01 -9.28644836e-02 5.09106703e-02] | [10.480956077575684, 9.87310791015625] |
3e770424-d043-4f99-b470-54a9e8d8f9f3 | best-buddies-registration-for-point-clouds | 2010.01912 | null | https://arxiv.org/abs/2010.01912v1 | https://arxiv.org/pdf/2010.01912v1.pdf | Best Buddies Registration for Point Clouds | We propose new, and robust, loss functions for the point cloud registration problem. Our loss functions are inspired by the Best Buddies Similarity (BBS) measure that counts the number of mutual nearest neighbors between two point sets. This measure has been shown to be robust to outliers and missing data in the case of template matching for images. We present several algorithms, collectively named Best Buddy Registration (BBR), where each algorithm consists of optimizing one of these loss functions with Adam gradient descent. The loss functions differ in several ways, including the distance function used (point-to-point vs. point-to-plane), and how the BBS measure is combined with the actual distances between pairs of points. Experiments on various data sets, both synthetic and real, demonstrate the effectiveness of the BBR algorithms, showing that they are quite robust to noise, outliers, and distractors, and cope well with extremely sparse point clouds. One variant, BBR-F, achieves state-of-the-art accuracy in the registration of automotive lidar scans taken up to several seconds apart, from the KITTI and Apollo-Southbay datasets. | ['Raja Giryes', 'Shai Avidan', 'Tal Shomer', 'Amnon Drory'] | 2020-10-05 | null | null | null | null | ['template-matching'] | ['computer-vision'] | [-2.46247903e-01 -3.88075203e-01 1.31298229e-01 -5.71046889e-01
-1.37224960e+00 -4.09149379e-01 7.10756063e-01 3.43641311e-01
-5.71082711e-01 4.35856998e-01 -2.71786332e-01 6.71986840e-04
-4.79539812e-01 -5.64300418e-01 -9.37189460e-01 -6.15600646e-01
-3.64590198e-01 9.90647197e-01 3.61848086e-01 -6.04066372e-01
3.65542680e-01 1.16150510e+00 -1.60325933e+00 -3.46508265e-01
5.94622970e-01 1.07190001e+00 -1.55613706e-01 2.02119976e-01
8.86339396e-02 6.73615187e-02 -4.14959133e-01 -5.19590735e-01
7.16455102e-01 2.17607602e-01 -6.10031307e-01 -1.46011099e-01
8.58013213e-01 1.15666926e-01 -2.02306539e-01 8.29868555e-01
5.20749569e-01 2.63662487e-01 6.81489170e-01 -1.46741915e+00
-3.38812619e-01 -8.80063996e-02 -7.18799591e-01 3.22211608e-02
2.98851252e-01 1.31675407e-01 9.76323664e-01 -1.36594403e+00
5.98249197e-01 1.28313088e+00 1.19095039e+00 1.19148806e-01
-1.33581865e+00 -6.14814639e-01 -2.20195159e-01 1.30979523e-01
-1.70891726e+00 -4.43354428e-01 6.86310828e-01 -7.13508010e-01
7.50902176e-01 2.97457278e-01 4.16038424e-01 5.70636988e-01
2.67344475e-01 2.42921352e-01 8.75385940e-01 -1.79332718e-01
1.87758639e-01 -2.11873516e-01 1.07918404e-01 4.34851646e-01
3.34169149e-01 4.56014156e-01 -4.34344769e-01 -5.74260414e-01
5.54142952e-01 -9.74010229e-02 -9.62929130e-02 -8.76712561e-01
-1.22493613e+00 9.33091462e-01 5.81457376e-01 4.84304577e-02
-3.52470964e-01 1.19990297e-01 2.63125211e-01 3.91976237e-01
6.00108445e-01 3.28414202e-01 -3.31986725e-01 -1.36007331e-02
-7.90080905e-01 4.81089979e-01 6.75411999e-01 9.94945705e-01
9.71849859e-01 -1.55324057e-01 3.56560826e-01 9.61204886e-01
3.81779969e-01 6.26909554e-01 3.13232213e-01 -1.03557026e+00
5.90903819e-01 1.65988147e-01 2.12457791e-01 -1.16284418e+00
-2.99295276e-01 -8.48320406e-03 -7.21059322e-01 6.27557576e-01
3.83688539e-01 2.66096801e-01 -6.57691360e-01 1.34454894e+00
5.36021352e-01 3.21496397e-01 -1.12190604e-01 9.30440843e-01
7.42087364e-01 2.76212007e-01 -4.56986845e-01 1.30387228e-02
8.40762198e-01 -3.81473273e-01 -1.27777964e-01 -2.86348641e-01
3.63610774e-01 -9.75728154e-01 9.58084702e-01 1.30793452e-01
-1.20342958e+00 -5.76855361e-01 -1.10543942e+00 5.76728582e-02
-3.95869344e-01 -4.14866507e-01 1.03578687e-01 2.47004017e-01
-9.79414403e-01 1.10539401e+00 -7.07624793e-01 -2.71140337e-01
3.60773623e-01 5.56027532e-01 -7.20195949e-01 1.54555626e-02
-7.28688478e-01 1.02506101e+00 -2.11804718e-01 1.49637848e-01
-3.44224066e-01 -8.69995117e-01 -8.14201951e-01 -3.21239740e-01
1.57159388e-01 -5.26621103e-01 1.01826680e+00 -3.60340655e-01
-1.16220295e+00 1.29925644e+00 -8.55287760e-02 -5.83763182e-01
7.70875990e-01 -2.75318891e-01 -2.54621387e-01 -4.16747838e-01
1.92862153e-01 6.28395975e-01 6.62826240e-01 -1.40631747e+00
-2.82343566e-01 -7.05995739e-01 -5.84673166e-01 -8.29080939e-02
4.40132499e-01 3.48211527e-02 -2.10932210e-01 -4.17095780e-01
6.21858418e-01 -1.07326269e+00 -3.00986499e-01 2.88573563e-01
-2.56674677e-01 -2.57164270e-01 8.21252584e-01 -3.88569832e-01
3.81911039e-01 -2.37080598e+00 -1.95172742e-01 5.76766849e-01
-4.69960161e-02 7.26331249e-02 -2.42610618e-01 5.00729680e-01
-2.37459779e-01 -1.19174667e-01 -4.49503332e-01 -4.65406567e-01
2.55067814e-02 3.21198016e-01 -4.36741084e-01 9.97918308e-01
1.31113529e-01 6.17803812e-01 -7.80490816e-01 -1.42533198e-01
3.08886081e-01 5.73427379e-01 -5.34818582e-02 6.37804046e-02
2.82430768e-01 4.37711805e-01 -6.68628141e-02 3.90204310e-01
8.99614692e-01 2.09768966e-01 -5.92023611e-01 -2.82013297e-01
-1.88235357e-01 1.61175564e-01 -1.24346352e+00 1.33494461e+00
-2.41772681e-01 7.11825669e-01 7.87129477e-02 -6.70456409e-01
1.46754146e+00 4.14303038e-03 9.00549650e-01 -6.54436052e-01
-1.05893925e-01 5.63955724e-01 -6.66173175e-02 -1.46830484e-01
4.89117622e-01 -1.64868981e-01 9.17719901e-02 8.56515169e-02
-2.57275254e-01 -7.15567231e-01 3.04125925e-03 -2.16139957e-01
9.80443180e-01 -8.65249336e-02 2.67099917e-01 -2.11329877e-01
5.13718963e-01 1.12700194e-01 5.72955072e-01 5.69528282e-01
-1.15452453e-01 1.07753611e+00 5.54739758e-02 -5.97477734e-01
-1.20175052e+00 -1.35438991e+00 -3.55334997e-01 5.27156889e-01
4.71939951e-01 -1.13516763e-01 -3.07827294e-01 -3.23252946e-01
6.84838355e-01 5.21669209e-01 -3.01206619e-01 -1.52984232e-01
-8.51080596e-01 -3.56587499e-01 3.91014457e-01 5.83725035e-01
1.36668041e-01 -6.43113732e-01 -5.01920342e-01 9.23524052e-02
1.16276726e-01 -1.22979295e+00 -5.47529101e-01 5.13534546e-02
-1.13711548e+00 -1.22798729e+00 -4.29348946e-01 -5.01272857e-01
5.02448976e-01 2.96531767e-01 1.43981290e+00 -8.19364469e-03
-2.97009587e-01 4.99635190e-01 -7.18458965e-02 -4.80365157e-01
-3.13995272e-01 -3.55077326e-01 1.79971576e-01 1.52220428e-02
3.96448284e-01 -7.07485318e-01 -3.73335302e-01 7.24942982e-01
-4.92590904e-01 -7.60105491e-01 2.97805041e-01 7.45398998e-01
1.11336982e+00 -4.52011675e-01 -2.37202421e-02 -5.25705934e-01
5.40180326e-01 -3.40143353e-01 -9.33069170e-01 -1.30791470e-01
-5.93268871e-01 -1.20370150e-01 1.62466988e-01 -1.29560620e-01
-1.30567506e-01 2.63725549e-01 -1.75107166e-01 -7.03914523e-01
-1.85815826e-01 7.32194781e-02 1.66344821e-01 -7.85823286e-01
6.44734502e-01 -4.78840843e-02 1.45239532e-01 -5.22453010e-01
4.25702870e-01 4.76233333e-01 8.67124498e-01 -6.49240911e-01
1.05223966e+00 8.54746103e-01 2.97266155e-01 -9.95657682e-01
-4.51604128e-01 -8.51205051e-01 -1.02560472e+00 -9.61023644e-02
5.73320866e-01 -7.18097627e-01 -7.00214386e-01 4.77957249e-01
-1.24753153e+00 7.50929536e-03 -7.27839708e-01 4.81806725e-01
-8.36875737e-01 3.69643331e-01 -1.81896448e-01 -5.40937185e-01
-3.45184058e-01 -1.19012725e+00 1.44325650e+00 -2.42791474e-01
-9.43329781e-02 -7.96773970e-01 4.57188606e-01 7.77647644e-02
1.67375445e-01 7.59926140e-01 6.72023177e-01 -7.43688822e-01
-5.52098811e-01 -6.07855678e-01 -9.96336564e-02 3.97752941e-01
5.06285727e-02 2.18648672e-01 -7.62123168e-01 -4.88935918e-01
2.27624387e-01 -9.46125463e-02 5.80712736e-01 2.93968618e-01
9.03508723e-01 -1.29775763e-01 -3.61158460e-01 8.52005839e-01
1.44455349e+00 -9.10280570e-02 7.13667274e-01 6.39320910e-01
5.37902832e-01 4.46682572e-01 9.09176052e-01 2.26503581e-01
3.83952051e-01 1.09601164e+00 8.22541475e-01 -1.35052338e-01
7.35384971e-02 1.54344052e-01 1.12479322e-01 7.48212576e-01
-1.24627925e-01 2.86766082e-01 -1.17127979e+00 4.98673499e-01
-1.79823959e+00 -7.74077356e-01 -4.54184324e-01 2.67322087e+00
2.80487895e-01 9.57030952e-02 2.92581797e-01 3.19703668e-02
6.55205071e-01 1.53232753e-01 -5.30551374e-01 -2.47375071e-01
-2.26762399e-01 2.38363758e-01 8.75573337e-01 5.66024184e-01
-9.77899790e-01 6.01795077e-01 6.45580959e+00 4.39010024e-01
-9.90504384e-01 1.03555575e-01 2.53272861e-01 -1.71819050e-02
1.18103944e-01 -6.55544698e-02 -6.04566216e-01 4.96703595e-01
7.80966401e-01 -9.73223373e-02 3.05769652e-01 8.80107224e-01
9.30414572e-02 5.78431301e-02 -1.19045794e+00 1.25293362e+00
4.53158356e-02 -1.16561794e+00 -2.90913880e-01 1.38903618e-01
5.91046691e-01 7.14202106e-01 -1.40351951e-02 -1.17374614e-01
3.81471217e-01 -9.67589915e-01 9.05422330e-01 5.94485581e-01
4.80940014e-01 -8.40700388e-01 7.95743883e-01 2.03584716e-01
-1.21578813e+00 2.32808739e-01 -5.75221539e-01 4.26707417e-01
2.16293722e-01 7.82095909e-01 -7.12041855e-01 6.16693676e-01
9.07186389e-01 5.10802627e-01 -5.11115134e-01 1.43184447e+00
1.77143961e-01 2.38630883e-02 -7.55566776e-01 4.11438197e-01
2.68254638e-01 -5.79860330e-01 8.39957654e-01 9.16207194e-01
4.31021333e-01 -1.05470225e-01 2.07977116e-01 7.15844691e-01
3.11866794e-02 3.67125496e-02 -8.39274108e-01 6.96056664e-01
7.98401892e-01 1.06950259e+00 -3.14524561e-01 1.20050907e-01
-1.69699356e-01 4.91907179e-01 1.56621054e-01 3.01911253e-02
-6.53305769e-01 -4.24457967e-01 1.13885319e+00 4.34250921e-01
3.59947324e-01 -6.49323165e-01 -4.71213490e-01 -5.27744591e-01
4.14956957e-01 -7.36706793e-01 2.26747602e-01 -8.55232775e-01
-1.76345122e+00 7.33049691e-01 1.48671851e-01 -1.52824080e+00
-1.52776554e-01 -6.48193479e-01 -8.02127004e-01 8.14671934e-01
-1.40274131e+00 -8.14532101e-01 -5.15446663e-01 5.95586658e-01
2.60096371e-01 -2.01150775e-01 4.27214086e-01 4.01367068e-01
-6.42658994e-02 4.36403930e-01 3.67232472e-01 4.02962901e-02
7.67102599e-01 -1.13376701e+00 9.82406855e-01 4.82281566e-01
3.22721452e-01 4.41830248e-01 7.87288606e-01 -5.56317925e-01
-1.31167257e+00 -1.04337323e+00 6.97440505e-01 -6.37823045e-01
6.15130007e-01 -1.25541747e-01 -1.13023138e+00 6.57609642e-01
-2.13120341e-01 3.59540373e-01 2.70254672e-01 1.70439563e-03
-4.11099702e-01 -5.23547471e-01 -1.49829471e+00 3.70449692e-01
9.69446182e-01 -3.53537261e-01 -6.11664832e-01 6.69403315e-01
5.06976962e-01 -7.03392923e-01 -1.00608659e+00 7.91569412e-01
2.83895105e-01 -1.19660556e+00 1.45672035e+00 -4.68190670e-01
-1.66796073e-01 -4.01317328e-01 -4.21654820e-01 -1.25765169e+00
-1.70105308e-01 -6.69233859e-01 1.54068679e-01 1.06856215e+00
2.27568150e-01 -9.21431124e-01 6.89276397e-01 4.36853409e-01
-2.36875236e-01 -9.27044630e-01 -1.38522899e+00 -1.44841492e+00
1.65937677e-01 -5.39140105e-01 9.63775218e-01 7.80893147e-01
-6.32916152e-01 2.06919685e-01 7.93159846e-03 2.45049298e-01
8.96984994e-01 7.81825930e-02 1.23603070e+00 -1.66758192e+00
1.28740624e-01 -4.96879935e-01 -1.02699077e+00 -7.70127833e-01
1.28579676e-01 -7.69261658e-01 1.69203669e-01 -1.23109984e+00
-5.49026251e-01 -7.53788829e-01 1.75210729e-01 2.09687516e-01
2.10343212e-01 2.93862343e-01 3.25435817e-01 5.82148790e-01
-1.23687290e-01 6.44332707e-01 8.05262923e-01 -1.69658795e-01
-8.31134617e-02 3.03278893e-01 -3.61659527e-02 9.56590652e-01
5.16265810e-01 -7.40829170e-01 2.14552268e-01 -4.91116792e-01
4.90089543e-02 -1.55288041e-01 6.17487013e-01 -1.29628110e+00
4.57841039e-01 -7.56892860e-02 8.49732384e-02 -9.73055243e-01
8.10877085e-01 -1.06301975e+00 4.26481366e-01 2.65058041e-01
2.16541756e-02 7.67347336e-01 8.79058540e-02 5.49222767e-01
-4.01898026e-01 -1.97988957e-01 1.14739823e+00 5.61486036e-02
-4.60690558e-01 5.17789364e-01 4.55957741e-01 1.94675252e-01
1.01885760e+00 -5.53390563e-01 -1.44139037e-01 -3.88522863e-01
-4.86811787e-01 2.02511668e-01 6.81591332e-01 4.28653210e-01
7.46865630e-01 -1.67466104e+00 -9.84643817e-01 3.12249422e-01
2.82837898e-01 2.49292597e-01 -2.50355780e-01 1.06748474e+00
-6.00975931e-01 9.07297730e-02 2.88043451e-02 -1.13942909e+00
-1.45390749e+00 3.29934776e-01 4.70951229e-01 1.02598898e-01
-6.92411900e-01 8.61537695e-01 -3.51476043e-01 -7.72974253e-01
1.64573863e-01 -3.99120390e-01 3.50694090e-01 -1.24591924e-01
7.37630650e-02 7.41876662e-01 5.96472442e-01 -1.12700665e+00
-6.18054867e-01 1.18943048e+00 4.82059307e-02 -3.84056829e-02
1.45028949e+00 2.07463846e-01 -1.90231100e-01 5.33280432e-01
1.33645225e+00 4.33248188e-03 -1.06355858e+00 -3.28583747e-01
2.71268278e-01 -8.25493753e-01 -2.75598198e-01 -2.61696100e-01
-1.11338937e+00 7.46312499e-01 9.16490495e-01 2.66063690e-01
6.06238425e-01 1.66361090e-02 9.97436583e-01 4.77447629e-01
6.06649935e-01 -9.10983920e-01 -2.17529595e-01 7.62217104e-01
1.14610159e+00 -1.37420833e+00 3.19056094e-01 -3.08833152e-01
-4.36418384e-01 1.07012188e+00 3.25621545e-01 -6.54369831e-01
8.01982880e-01 1.37509897e-01 2.17843086e-01 -4.52319920e-01
-2.15716109e-01 -2.14368224e-01 5.12747526e-01 6.53470576e-01
1.05205290e-02 -1.23055346e-01 -8.12722668e-02 -5.78214414e-02
-6.39768600e-01 -4.73458409e-01 4.80783172e-02 7.93127954e-01
-4.13742065e-01 -9.93187964e-01 -7.19149947e-01 3.18469524e-01
4.72609252e-02 2.83561289e-01 -6.64407551e-01 1.17377484e+00
-1.70236360e-02 6.77894652e-01 3.34575564e-01 -3.88844401e-01
9.18174624e-01 -1.71453357e-01 5.38438261e-01 -3.58134091e-01
-6.24248624e-01 -1.86331376e-01 -2.11076349e-01 -9.51496422e-01
-3.51659834e-01 -9.05801713e-01 -1.08204913e+00 -5.00318527e-01
-4.43867862e-01 1.73000678e-01 9.20625091e-01 8.25787246e-01
4.21363592e-01 -8.24666321e-02 8.79060805e-01 -1.23754919e+00
-9.66498554e-01 -5.43989658e-01 -5.53188801e-01 6.84242666e-01
5.16543984e-01 -7.20776498e-01 -4.18105096e-01 -3.72249037e-01] | [7.738664627075195, -2.832556962966919] |
8844f494-db7b-45b7-ae86-b3ebcd724fa0 | a-little-bit-attention-is-all-you-need-for | 2302.14574 | null | https://arxiv.org/abs/2302.14574v1 | https://arxiv.org/pdf/2302.14574v1.pdf | A Little Bit Attention Is All You Need for Person Re-Identification | Person re-identification plays a key role in applications where a mobile robot needs to track its users over a long period of time, even if they are partially unobserved for some time, in order to follow them or be available on demand. In this context, deep-learning based real-time feature extraction on a mobile robot is often performed on special-purpose devices whose computational resources are shared for multiple tasks. Therefore, the inference speed has to be taken into account. In contrast, person re-identification is often improved by architectural changes that come at the cost of significantly slowing down inference. Attention blocks are one such example. We will show that some well-performing attention blocks used in the state of the art are subject to inference costs that are far too high to justify their use for mobile robotic applications. As a consequence, we propose an attention block that only slightly affects the inference speed while keeping up with much deeper networks or more complex attention blocks in terms of re-identification accuracy. We perform extensive neural architecture search to derive rules at which locations this attention block should be integrated into the architecture in order to achieve the best trade-off between speed and accuracy. Finally, we confirm that the best performing configuration on a re-identification benchmark also performs well on an indoor robotic dataset. | ['Horst-Michael Gross', 'Dustin Aganian', 'Jannik Lübberstedt', 'Markus Eisenbach'] | 2023-02-28 | null | null | null | null | ['person-re-identification'] | ['computer-vision'] | [-6.85472637e-02 1.00805545e-02 7.90687725e-02 -4.77763832e-01
-2.14809462e-01 -3.03299487e-01 5.88155210e-01 3.55256081e-01
-9.55910206e-01 5.76395154e-01 -3.58920068e-01 -1.25917196e-01
-3.17760706e-01 -6.47665322e-01 -7.18661547e-01 -4.73484427e-01
1.11897886e-01 9.01922703e-01 2.69315958e-01 -8.95229280e-02
-7.05904737e-02 6.99336171e-01 -1.91922510e+00 -2.56054878e-01
7.44719088e-01 8.03208411e-01 3.39137167e-01 5.05783021e-01
6.40194044e-02 3.60738009e-01 -4.91912782e-01 -7.53631413e-01
2.45011419e-01 -9.22305211e-02 -8.98027420e-01 -9.29695442e-02
2.86911190e-01 -4.67162818e-01 -9.17126834e-02 1.12085915e+00
5.44236302e-01 3.47589880e-01 6.12996995e-01 -1.14587867e+00
-1.62751466e-01 6.18689477e-01 -2.71159351e-01 1.84804127e-01
2.42186993e-01 1.71066448e-02 7.91868925e-01 -6.11204863e-01
4.19876993e-01 1.13690388e+00 7.89367199e-01 3.68999451e-01
-1.24307406e+00 -4.38182235e-01 3.72391909e-01 5.37580550e-01
-1.44596553e+00 -5.87357759e-01 5.45905828e-01 -2.67428070e-01
1.07077217e+00 2.22159073e-01 5.00949621e-01 1.06747401e+00
-3.07406448e-02 5.56572139e-01 4.65572447e-01 -3.37059706e-01
2.16727957e-01 3.42837960e-01 3.19498152e-01 4.90320593e-01
6.52956724e-01 -3.64969641e-01 -1.46422893e-01 7.85319433e-02
3.30801755e-01 1.39939085e-01 -1.26261130e-01 -3.89584750e-01
-9.48540390e-01 5.54750085e-01 6.09340012e-01 6.36015415e-01
-5.34725785e-01 1.77286357e-01 4.45119530e-01 3.54636639e-01
1.01791009e-01 6.53620005e-01 -5.39690077e-01 -3.05447280e-01
-6.48622751e-01 2.12385133e-01 6.42190516e-01 8.10605526e-01
8.67009997e-01 -3.90186310e-01 -4.65973429e-02 8.77285480e-01
8.36998150e-02 2.89338708e-01 6.63765669e-01 -7.12673128e-01
2.19130263e-01 7.52444029e-01 4.98432159e-01 -1.03327429e+00
-7.38675117e-01 -7.14398026e-01 -9.00945842e-01 1.59402728e-01
7.88547516e-01 -8.42958316e-02 -5.47411323e-01 1.71521473e+00
1.84873149e-01 1.46829337e-01 -7.88305253e-02 9.88316774e-01
2.48608842e-01 2.02287182e-01 6.47417232e-02 -7.64245633e-03
1.75411332e+00 -9.14534986e-01 -4.39617008e-01 -4.51233834e-01
6.28490269e-01 -4.47936177e-01 1.02000654e+00 4.24806774e-01
-8.09772372e-01 -8.04800630e-01 -1.09257531e+00 -1.14245377e-01
-5.76849818e-01 4.49708700e-01 5.41199446e-01 8.09175730e-01
-7.77877152e-01 8.73579562e-01 -8.71423602e-01 -7.73828983e-01
1.33135870e-01 7.57549942e-01 -3.92475575e-01 7.77456611e-02
-1.07596552e+00 1.08360922e+00 2.26378322e-01 5.18569410e-01
-3.15336466e-01 -3.19607228e-01 -5.98765910e-01 5.04963398e-01
4.24906224e-01 -7.42975354e-01 1.22640026e+00 -1.04458010e+00
-1.62423885e+00 6.50043011e-01 -2.34329045e-01 -7.57940710e-01
8.05002928e-01 -3.93482149e-01 -3.41092050e-01 -2.44818747e-01
-3.65667860e-03 6.16698921e-01 8.71646047e-01 -7.07921624e-01
-7.94950724e-01 -5.94441295e-01 4.95195478e-01 1.37018636e-01
-5.67940414e-01 -2.04575405e-01 -5.89517355e-01 -2.89179713e-01
-1.36123568e-01 -1.44204032e+00 -1.15187511e-01 -3.40369552e-01
-2.75078923e-01 -3.92015696e-01 6.12370610e-01 -5.46084166e-01
8.96847129e-01 -1.83453059e+00 3.00904810e-01 1.96982890e-01
-1.55435175e-01 4.44413960e-01 1.57850042e-01 -9.94986370e-02
4.28434424e-02 3.89959110e-04 4.39535454e-02 -7.57551968e-01
3.59467752e-02 -6.79993108e-02 1.42528847e-01 4.45407450e-01
1.91618744e-02 7.35621572e-01 -6.87137961e-01 -1.10814422e-01
3.03610593e-01 4.06028807e-01 -6.50036573e-01 -7.99586102e-02
2.54005045e-02 4.80538845e-01 -2.44440630e-01 1.89963967e-01
4.23989147e-01 -2.52709150e-01 2.27453068e-01 -2.58056641e-01
5.17236593e-04 2.16912955e-01 -1.23618829e+00 1.70607460e+00
-8.14072371e-01 6.47232890e-01 1.11878039e-02 -1.05396366e+00
6.70896232e-01 1.58688292e-01 2.50081927e-01 -6.53655767e-01
3.81373316e-01 3.11616719e-01 3.23693246e-01 -3.31108183e-01
7.84835577e-01 2.84219593e-01 -8.28922912e-02 3.80152315e-01
-3.01952600e-01 6.02668405e-01 2.28958517e-01 -2.14506775e-01
9.60539579e-01 -6.31101280e-02 2.56903708e-01 -4.18963552e-01
8.53235781e-01 -1.06365621e-01 4.96742934e-01 8.38377476e-01
-1.07813127e-01 3.10788810e-01 1.75340489e-01 -4.64630485e-01
-1.03438854e+00 -2.56765306e-01 -1.09457336e-02 1.07662523e+00
2.13413313e-01 -2.27578521e-01 -8.85075808e-01 -6.62950635e-01
-6.44845590e-02 5.94768584e-01 -5.22053540e-01 -2.05895007e-01
-5.56724727e-01 -6.89877272e-01 5.02337039e-01 4.81440812e-01
6.48679793e-01 -9.46595252e-01 -1.05550897e+00 3.22780520e-01
-1.52539298e-01 -1.15396261e+00 -1.54338390e-01 2.38062382e-01
-6.22462630e-01 -1.02240944e+00 -1.02668917e+00 -4.49980527e-01
6.72436416e-01 3.29630762e-01 7.48720169e-01 2.67236799e-01
-3.04943994e-02 3.20772767e-01 -3.93678248e-01 -3.76297951e-01
-1.93751350e-01 7.22455323e-01 4.56894815e-01 2.19173953e-01
4.92089540e-01 -5.35498857e-01 -5.60621321e-01 4.87095535e-01
-4.02928352e-01 -1.76798537e-01 5.21080494e-01 8.53837967e-01
9.02894065e-02 3.93956780e-01 5.73812604e-01 -7.09569931e-01
3.47517371e-01 -2.19707355e-01 -8.18012655e-01 2.67972440e-01
-6.09884560e-01 4.41592813e-01 7.29164720e-01 -5.91900170e-01
-8.59015048e-01 2.49310926e-01 -2.54418641e-01 -2.22862795e-01
-2.10222736e-01 3.67233336e-01 -4.00796086e-01 1.25203699e-01
5.66427827e-01 -4.92598303e-02 5.21370098e-02 -6.70220494e-01
-6.61007613e-02 8.73523474e-01 2.21674576e-01 -1.35285005e-01
4.96006161e-01 4.37885791e-01 -1.20165251e-01 -8.14210594e-01
-5.05469441e-01 -4.03451979e-01 -7.63172328e-01 8.35592225e-02
5.53315222e-01 -8.17990065e-01 -1.15022099e+00 5.77286601e-01
-1.12697780e+00 -4.00980353e-01 -2.61364412e-02 4.25837725e-01
-4.08970147e-01 3.22979212e-01 -2.56509542e-01 -8.24017107e-01
-2.05273747e-01 -1.42046452e+00 8.12340856e-01 2.83330560e-01
-4.78489995e-01 -6.39102280e-01 -3.85356396e-01 1.64375991e-01
4.86270875e-01 -3.13576251e-01 7.02228010e-01 -7.37128377e-01
-3.89723986e-01 -2.86032706e-01 -2.84430921e-01 3.92956510e-02
2.16571033e-01 -4.09693867e-01 -1.02761436e+00 -4.58231360e-01
-3.41524750e-01 2.50992894e-01 7.12293923e-01 1.73478380e-01
1.18741679e+00 -1.35963395e-01 -4.79235768e-01 3.23184103e-01
1.01672626e+00 1.10488035e-01 4.74016935e-01 5.20660996e-01
6.55370116e-01 7.26933718e-01 5.54162502e-01 2.44210854e-01
4.74937171e-01 1.33012807e+00 3.87129068e-01 1.29185095e-01
5.02001084e-02 2.57935598e-02 1.93099529e-01 1.28996208e-01
-2.13779643e-01 -2.34375328e-01 -9.49250579e-01 5.19509137e-01
-2.16050768e+00 -8.33458304e-01 5.12082642e-03 2.57707334e+00
2.49881729e-01 3.59204888e-01 5.54048181e-01 5.39925277e-01
7.43169785e-01 -3.85574073e-01 -6.43184006e-01 -2.43767709e-01
2.87218869e-01 -2.85751373e-01 7.63730526e-01 4.73215669e-01
-1.05300987e+00 7.34224916e-01 4.80066681e+00 6.25354707e-01
-1.20522809e+00 1.63573742e-01 5.07083356e-01 4.59798835e-02
1.74589843e-01 -1.95472375e-01 -1.12871385e+00 7.43581235e-01
1.18258846e+00 7.71921948e-02 5.42996168e-01 1.07010472e+00
9.79526564e-02 -2.63568163e-01 -1.43390214e+00 1.38180292e+00
-1.65434629e-01 -9.10090804e-01 -3.25086206e-01 2.72922754e-01
1.63938791e-01 9.62301902e-03 -1.80700257e-01 4.60437953e-01
-6.53933808e-02 -8.37359488e-01 7.81891048e-01 4.95836675e-01
4.91574705e-01 -8.80457640e-01 1.05951476e+00 4.85690534e-01
-1.07578290e+00 -3.54618251e-01 -4.69329655e-01 -1.44878492e-01
1.65422454e-01 5.32879889e-01 -9.47101116e-01 3.94076228e-01
8.79901052e-01 4.98313934e-01 -6.94170177e-01 1.09223235e+00
-5.29063307e-02 6.17089309e-02 -4.36178029e-01 -2.52367795e-01
1.09860569e-01 1.22093871e-01 5.05069137e-01 1.20310974e+00
5.90151608e-01 -2.65376776e-01 -2.17569053e-01 7.14941502e-01
-4.79476601e-02 -9.83779356e-02 -5.39082706e-01 3.51078361e-01
4.31882620e-01 1.26832223e+00 -8.46486270e-01 -3.32790613e-01
-2.13844791e-01 1.39912558e+00 4.85516727e-01 1.24439843e-01
-8.95472348e-01 -3.64023417e-01 8.58909845e-01 3.26893836e-01
3.63889396e-01 -1.89438283e-01 -6.13753460e-02 -1.03445113e+00
-5.35785221e-02 -4.99938667e-01 1.22254029e-01 -4.70031649e-01
-1.06297219e+00 6.67217910e-01 -2.48938724e-01 -1.13883889e+00
-6.21504128e-01 -7.21992552e-01 -8.57605711e-02 7.48239160e-01
-1.44974887e+00 -7.88780212e-01 -4.76969481e-01 4.91849899e-01
5.69228113e-01 -1.57959968e-01 8.22126329e-01 7.43550777e-01
-8.65086555e-01 8.60294938e-01 -7.40402937e-02 5.87102808e-02
5.46682119e-01 -1.10456026e+00 4.91575569e-01 7.93390274e-01
1.95036709e-01 7.34842300e-01 7.31864929e-01 -3.57591122e-01
-1.34643817e+00 -9.17257667e-01 1.01431417e+00 -4.38489556e-01
1.86744124e-01 -5.26765466e-01 -9.13097799e-01 6.73046470e-01
-2.40257382e-01 -1.33868471e-01 2.86380351e-01 6.02220714e-01
-2.44924389e-02 -3.62469554e-01 -1.07351291e+00 6.89276636e-01
1.07579851e+00 -3.80292296e-01 -2.45709136e-01 -3.12137045e-02
3.45988840e-01 -2.17506021e-01 -5.67016780e-01 2.84962744e-01
7.79782951e-01 -9.56059575e-01 1.06007648e+00 -3.37236524e-01
-1.84126079e-01 -4.07592118e-01 1.56123742e-01 -1.22022390e+00
-4.95432049e-01 -4.34787214e-01 -1.52443666e-02 1.22934115e+00
4.80811119e-01 -7.55001009e-01 9.45911705e-01 1.07989168e+00
2.85886109e-01 -3.71571928e-01 -1.00859618e+00 -9.64143097e-01
-3.90688002e-01 -3.64245176e-01 8.14369500e-01 7.46075809e-01
-1.07439421e-01 6.13496542e-01 -5.19453704e-01 2.51226008e-01
2.75964111e-01 -1.19084358e-01 1.05548549e+00 -1.60983145e+00
-2.28218645e-01 -4.70202684e-01 -4.96981025e-01 -1.13845301e+00
3.50698978e-01 -6.75760746e-01 4.87104356e-02 -1.20981586e+00
9.45855677e-02 -8.34185064e-01 -2.50723213e-01 4.31692868e-01
-2.12623850e-02 2.33702004e-01 2.92544872e-01 2.14108557e-01
-6.25900924e-01 4.24320281e-01 3.93516630e-01 -2.29355112e-01
-5.18572509e-01 4.29568052e-01 -6.57972634e-01 8.31622064e-01
8.60697925e-01 -3.58279169e-01 -3.26496929e-01 -6.35063767e-01
4.17728513e-01 -2.49050871e-01 5.18075407e-01 -1.35292685e+00
5.16328394e-01 3.17674041e-01 2.72120506e-01 -3.56029958e-01
6.52036786e-01 -1.36193061e+00 2.59084195e-01 5.94212174e-01
-6.19621053e-02 1.15334049e-01 1.23829395e-01 5.64545870e-01
7.74220377e-02 -5.39376318e-01 7.95143962e-01 1.57037769e-02
-7.31469333e-01 8.76489878e-02 -5.00803113e-01 -6.49786949e-01
8.14289749e-01 -3.51480275e-01 1.29903570e-01 -3.90793920e-01
-7.73357213e-01 9.66966972e-02 5.58032095e-01 6.32386088e-01
1.26712129e-01 -1.08366811e+00 -3.64994854e-01 1.46762982e-01
6.95262924e-02 -8.91938433e-02 3.17183554e-01 8.95869970e-01
-3.37475657e-01 6.22304261e-01 -2.54294634e-01 -5.75615466e-01
-1.32592034e+00 6.78753912e-01 4.50579703e-01 -3.51885259e-01
-6.82613015e-01 5.86722553e-01 -1.93572819e-01 -3.08935225e-01
3.89001936e-01 -4.99373198e-01 -3.96933526e-01 2.78216183e-01
6.61915004e-01 6.47932589e-01 5.40276051e-01 -6.34820938e-01
-5.74211121e-01 5.29331088e-01 -1.09079175e-01 7.54315406e-02
1.09103525e+00 -4.46917415e-01 2.60625839e-01 2.45248824e-01
9.37429965e-01 -7.17934594e-02 -1.18020451e+00 -2.15388164e-01
2.09080160e-01 -2.73466289e-01 -2.96055190e-02 -6.33090258e-01
-9.62521136e-01 7.68056929e-01 8.13081503e-01 1.86428502e-01
9.34614718e-01 -2.45261014e-01 6.07748568e-01 8.84663284e-01
7.32105434e-01 -1.18113697e+00 -2.71093965e-01 5.42578697e-01
6.77790642e-01 -1.44878435e+00 3.42294667e-03 -1.97664902e-01
-3.97718251e-01 8.82581055e-01 5.32368004e-01 -6.31928369e-02
5.11449695e-01 -5.68316467e-02 -4.80634987e-01 1.38103217e-01
-2.13742808e-01 -3.73025537e-01 9.17432681e-02 4.40123767e-01
6.90777972e-02 3.32698822e-02 -6.49519414e-02 7.26656318e-01
-2.96206951e-01 -9.20647830e-02 3.40281278e-01 5.16609192e-01
-2.78461516e-01 -1.14143431e+00 -3.00218582e-01 2.75828212e-01
-4.60248351e-01 2.35307559e-01 -1.77090779e-01 7.03281879e-01
2.84292072e-01 9.30072129e-01 2.48845339e-01 -3.23375046e-01
4.31743592e-01 -2.56995056e-02 4.01612371e-01 -3.99036109e-01
-7.28953838e-01 -3.30495149e-01 2.17744216e-01 -5.49508154e-01
-3.67793530e-01 -6.82113826e-01 -1.00113213e+00 -3.92540246e-01
-2.76878685e-01 4.44965288e-02 7.61398494e-01 9.64478016e-01
6.46957576e-01 6.23056233e-01 1.53180674e-01 -1.14856708e+00
-2.66143978e-01 -1.09009480e+00 -3.30168366e-01 2.52821416e-01
1.50374621e-01 -9.84260738e-01 5.06758839e-02 -2.00677872e-01] | [14.378056526184082, 0.7874850630760193] |
9368890d-5608-4824-89ae-b67125feb25f | emns-imz-corpus-an-emotive-single-speaker | 2305.13137 | null | https://arxiv.org/abs/2305.13137v2 | https://arxiv.org/pdf/2305.13137v2.pdf | EMNS /Imz/ Corpus: An emotive single-speaker dataset for narrative storytelling in games, television and graphic novels | The increasing adoption of text-to-speech technologies has led to a growing demand for natural and emotive voices that adapt to a conversation's context and emotional tone. The Emotive Narrative Storytelling (EMNS) corpus is a unique speech dataset created to enhance conversations' expressiveness and emotive quality in interactive narrative-driven systems. The corpus consists of a 2.3-hour recording featuring a female speaker delivering labelled utterances. It encompasses eight acted emotional states, evenly distributed with a variance of 0.68%, along with expressiveness levels and natural language descriptions with word emphasis labels. The evaluation of audio samples from different datasets revealed that the EMNS corpus achieved the highest average scores in accurately conveying emotions and demonstrating expressiveness. It outperformed other datasets in conveying shared emotions and achieved comparable levels of genuineness. A classification task confirmed the accurate representation of intended emotions in the corpus, with participants recognising the recordings as genuine and expressive. Additionally, the availability of the dataset collection tool under the Apache 2.0 License simplifies remote speech data collection for researchers. | ['Jian Jun Zhang', 'Xiaosong Yang', 'Kari Ali Noriy'] | 2023-05-22 | null | null | null | null | ['expressive-speech-synthesis', 'speech-synthesis'] | ['speech', 'speech'] | [-1.45154810e-02 2.67179042e-01 1.78479448e-01 -6.13008976e-01
-6.73082054e-01 -6.70154452e-01 9.79254067e-01 -2.13373937e-02
7.82186091e-02 6.62073135e-01 9.82630551e-01 3.60406607e-01
1.58105016e-01 -1.49824649e-01 1.18592277e-01 -3.22359473e-01
-1.79016665e-01 1.27842724e-01 -1.99257597e-01 -6.83436751e-01
1.46588415e-01 2.33772099e-01 -1.91150987e+00 9.34219778e-01
2.74325073e-01 1.20854592e+00 -1.07814018e-02 1.37959242e+00
-4.57438767e-01 1.57653201e+00 -1.24403107e+00 -5.90129018e-01
-4.98559058e-01 -5.43393373e-01 -9.24272358e-01 1.58048823e-01
-1.69775441e-01 -1.10633269e-01 -2.42659763e-01 2.05761805e-01
8.60878170e-01 4.02366251e-01 2.62625873e-01 -1.44096994e+00
-4.38414216e-01 5.94879508e-01 2.63763458e-01 5.12960441e-02
1.07325804e+00 -7.74967819e-02 1.05838859e+00 -7.66831756e-01
1.05855107e+00 1.07417965e+00 5.46080291e-01 8.28254461e-01
-1.03752089e+00 -6.20330274e-01 -3.43160480e-01 -5.32184914e-02
-1.10233998e+00 -1.09555006e+00 9.21730101e-01 -5.22273242e-01
1.30365384e+00 8.83050382e-01 7.78388321e-01 1.73574758e+00
-4.24165606e-01 6.65367067e-01 1.07331288e+00 -3.62863123e-01
2.97882497e-01 8.08819354e-01 -2.29635373e-01 3.27287391e-02
-9.90040839e-01 -3.69872689e-01 -1.10583651e+00 -4.50435251e-01
3.07324603e-02 -4.99340773e-01 -2.74114817e-01 1.79062381e-01
-1.08735597e+00 7.32358396e-01 -3.61710161e-01 6.25374317e-01
-5.82581520e-01 -3.71768624e-02 1.41556633e+00 4.34897810e-01
7.85681605e-01 5.58653772e-01 -2.11131930e-01 -1.44487858e+00
-5.40131271e-01 1.95093378e-01 1.31091404e+00 9.66022670e-01
-1.06194586e-01 2.53942490e-01 -1.99896947e-01 1.56490767e+00
-1.99476808e-01 3.15876782e-01 5.34107804e-01 -1.01418519e+00
8.72432664e-02 5.02157509e-01 1.63047254e-01 -1.29133785e+00
-3.76650959e-01 1.82282329e-01 -3.48274976e-01 -1.74659342e-01
-5.95836639e-02 -3.89677435e-01 6.35909662e-02 1.72905266e+00
1.42193034e-01 -4.19056803e-01 5.84182024e-01 7.01900959e-01
1.41862917e+00 9.00682330e-01 3.29663396e-01 -3.14633936e-01
1.52399373e+00 -4.51517433e-01 -1.49331808e+00 -8.52540955e-02
5.36842704e-01 -9.71842885e-01 1.46944726e+00 4.90572810e-01
-1.19787431e+00 -2.63598859e-01 -9.36168194e-01 -1.55499550e-02
-4.54919755e-01 4.07921709e-02 6.63062036e-01 8.75314236e-01
-9.41141248e-01 2.46169835e-01 -1.38306290e-01 -4.59326327e-01
1.91297039e-01 -1.19269438e-01 -6.42555416e-01 7.73471773e-01
-1.30572021e+00 8.12716246e-01 -2.20843479e-02 -3.63206893e-01
-4.42758560e-01 -5.15217662e-01 -8.48108828e-01 -2.41392374e-01
-3.90819646e-02 3.62351656e-01 1.69979227e+00 -1.35550940e+00
-2.16496038e+00 9.93765891e-01 4.70013805e-02 -1.98922232e-01
4.20125663e-01 -9.26393494e-02 -1.00094700e+00 6.90305829e-01
-6.15267269e-02 5.98017454e-01 4.94010121e-01 -1.09905791e+00
-3.38424832e-01 2.16773868e-01 2.41624303e-02 1.51638165e-01
-7.40761399e-01 8.32240641e-01 7.69479126e-02 -5.93867064e-01
-4.80749577e-01 -5.39738357e-01 4.64462757e-01 -3.95637929e-01
-3.45788926e-01 -2.81319201e-01 9.94163871e-01 -8.21768403e-01
1.37984157e+00 -2.47461915e+00 -1.82316750e-01 -1.75568566e-01
2.50468969e-01 -1.28882304e-01 2.25200847e-01 1.16519380e+00
-7.85848871e-02 5.46573475e-02 1.18582152e-01 -5.99370956e-01
4.85610098e-01 1.54413611e-01 -3.58194530e-01 2.13547036e-01
2.06138611e-01 5.63366175e-01 -1.03939760e+00 -5.68112910e-01
2.79453516e-01 8.76079798e-01 -2.07901806e-01 5.46843112e-01
7.20169172e-02 2.80597717e-01 -1.32991344e-01 3.77659708e-01
7.60762915e-02 2.62666404e-01 1.11216716e-01 4.51341495e-02
-3.51780653e-01 4.38799948e-01 -8.01291108e-01 1.25653017e+00
-7.47877955e-01 1.41237640e+00 2.96213031e-01 -3.85339528e-01
1.50067377e+00 1.06055856e+00 5.46838343e-01 -5.83411634e-01
3.62096608e-01 9.14266482e-02 -3.55532736e-01 -1.22696185e+00
8.31326008e-01 -4.91803974e-01 -6.96785212e-01 3.92047346e-01
1.32534252e-02 -6.39070392e-01 -6.56781495e-02 3.57352734e-01
1.07714379e+00 -3.16646218e-01 3.32801402e-01 -2.05326881e-02
3.78729761e-01 -2.07074676e-02 -4.54850588e-03 2.03628719e-01
-5.08667290e-01 4.98476565e-01 6.37532651e-01 -3.12740713e-01
-9.14161861e-01 -5.30438483e-01 -8.19733664e-02 1.57850003e+00
-4.10509139e-01 -7.78845012e-01 -7.53267586e-01 6.36983961e-02
-5.31813800e-01 9.90771294e-01 -5.86754203e-01 -9.70368758e-02
2.48428062e-02 5.96426353e-02 1.14400673e+00 2.31077075e-01
4.23442394e-01 -1.68414247e+00 -9.37399089e-01 3.82836848e-01
-8.07060897e-01 -1.40551627e+00 -2.42247730e-01 9.59023759e-02
4.88467403e-02 -6.46314859e-01 -3.87777448e-01 -5.31783581e-01
-2.21898388e-02 -2.08214670e-01 1.19280326e+00 -3.24408114e-01
-2.18647853e-01 8.39205086e-01 -9.36110556e-01 -5.27444839e-01
-9.99390781e-01 -1.94047898e-01 -4.76365024e-03 1.31630972e-01
4.83843595e-01 -5.28446555e-01 -6.33493736e-02 1.61307350e-01
-7.22975910e-01 3.38882864e-01 -2.98239380e-01 6.12584651e-01
-2.62572795e-01 -3.50892216e-01 1.06962478e+00 -3.95666540e-01
1.38832796e+00 -6.42015874e-01 6.97853804e-01 -2.74332196e-01
1.77209243e-01 -7.10631788e-01 7.29769886e-01 -6.99847102e-01
-1.35316372e+00 -2.12443903e-01 -2.89872378e-01 5.29171014e-03
-5.77403605e-01 1.27785757e-01 -7.39891455e-03 3.27339560e-01
6.82963669e-01 -1.02621736e-02 1.22826912e-01 1.02406985e-03
3.16050202e-01 1.55164385e+00 8.51291478e-01 -4.27138805e-01
-5.50738834e-02 2.10013345e-01 -7.53371716e-01 -1.34996760e+00
-3.92031729e-01 -3.27494293e-01 -5.98991066e-02 -1.06570828e+00
7.48610795e-01 -7.87259758e-01 -1.13606298e+00 3.66527230e-01
-1.12651443e+00 -3.82235736e-01 -5.70258081e-01 3.42646688e-01
-8.25962305e-01 6.12709634e-02 -8.06890607e-01 -1.44410384e+00
-5.14495313e-01 -7.05897629e-01 9.67699587e-01 9.40622911e-02
-1.42995632e+00 -7.42766082e-01 1.23402337e-02 4.85398293e-01
5.76114953e-01 6.50888681e-01 3.55875343e-01 -9.32352066e-01
9.70536768e-01 -5.52322149e-01 1.53024584e-01 1.41252816e-01
1.52865574e-01 2.76255608e-01 -1.42233360e+00 2.66677231e-01
7.35115185e-02 -1.01625121e+00 -2.40937188e-01 -6.46770746e-02
7.12698162e-01 -6.12556696e-01 3.65818590e-01 -1.19239300e-01
6.76037729e-01 4.36095357e-01 6.28590345e-01 2.59985775e-01
9.80147123e-02 1.33056390e+00 4.64253545e-01 1.15527308e+00
5.21287084e-01 6.23078942e-01 1.57812148e-01 1.85809061e-01
1.55477583e-01 -1.94270507e-01 5.49642384e-01 9.52342093e-01
1.90592587e-01 -2.75132775e-01 -7.62720644e-01 5.10343730e-01
-1.54952788e+00 -1.40963924e+00 -2.04873458e-01 1.49024165e+00
1.06555688e+00 -1.41841456e-01 3.86936456e-01 4.88593310e-01
6.30356908e-01 4.70699966e-01 -2.57364780e-01 -1.53122795e+00
-3.31170619e-01 -5.24724536e-02 -3.38442236e-01 4.56480086e-01
-7.41121888e-01 6.41295552e-01 6.51297998e+00 6.87395215e-01
-1.11779654e+00 6.83937520e-02 6.89815998e-01 -5.71603119e-01
-2.00543180e-01 -6.33299112e-01 -8.60773176e-02 4.88487691e-01
1.46333981e+00 -3.19283903e-01 4.73572969e-01 8.19645762e-01
6.95782304e-01 3.28813791e-02 -7.99021125e-01 1.06493676e+00
3.55347574e-01 -1.21047258e+00 -5.13257325e-01 -4.56967473e-01
3.22012067e-01 -3.72809201e-01 -1.88467085e-01 5.06376266e-01
-5.46353385e-02 -9.29043710e-01 1.08262312e+00 5.89474380e-01
1.08420265e+00 -8.10154974e-01 6.08574688e-01 2.35789284e-01
-9.62908030e-01 1.89032033e-02 3.65976453e-01 -6.37979448e-01
4.45766300e-01 -1.45907123e-02 -9.66067255e-01 -1.49789870e-01
8.38297427e-01 4.78578597e-01 6.41528983e-03 5.53557202e-02
2.01854974e-01 6.43221080e-01 -1.44569322e-01 -7.35760212e-01
8.32162797e-02 9.52014476e-02 7.93906868e-01 2.00581646e+00
1.35949031e-02 4.85252768e-01 -1.76621795e-01 4.79623497e-01
3.57324220e-02 4.99107897e-01 -6.89738810e-01 -6.66861057e-01
7.83277273e-01 1.49171925e+00 -5.92700899e-01 -4.05282140e-01
-1.11958787e-01 1.04918957e+00 -2.18891371e-02 1.89062729e-02
-7.66529381e-01 -6.62153125e-01 7.19013035e-01 -1.97740361e-01
-1.74835280e-01 4.25517671e-02 -1.54844001e-01 -4.82234567e-01
5.16171902e-02 -1.12805390e+00 2.28592027e-02 -1.21685481e+00
-1.28132272e+00 1.24255288e+00 -9.06252563e-02 -1.04199862e+00
-5.33820510e-01 -8.46166909e-02 -6.43477440e-01 5.21916568e-01
-6.52505815e-01 -8.68094623e-01 -6.22625291e-01 2.97228813e-01
8.39589238e-01 -2.08768457e-01 1.35683274e+00 3.72928321e-01
-4.05677557e-01 4.16033804e-01 -3.77451569e-01 -6.80641010e-02
8.25034499e-01 -1.10085535e+00 -9.89613310e-02 8.82454030e-03
-4.12721306e-01 2.65327007e-01 1.23778653e+00 -1.63181365e-01
-1.31952596e+00 -4.65499401e-01 1.26127410e+00 -2.31734291e-01
1.02066350e+00 -5.80305219e-01 -7.46699452e-01 2.05299616e-01
7.24292934e-01 -3.76352668e-01 1.42528093e+00 -1.96948364e-01
-2.15771347e-01 3.53443176e-01 -1.55064428e+00 6.17048562e-01
7.43372738e-01 -9.25895929e-01 -6.09559476e-01 2.54006177e-01
6.43965185e-01 -3.63721162e-01 -1.21605945e+00 -1.94658637e-01
7.55999565e-01 -1.09577084e+00 5.60550034e-01 -6.95609331e-01
9.45367038e-01 3.48662257e-01 -4.24134344e-01 -1.13013661e+00
1.83930174e-01 -1.23555958e+00 -9.24661476e-03 1.99347961e+00
1.29709050e-01 -4.71522689e-01 3.38263243e-01 1.18555212e+00
-2.55295426e-01 -6.25138283e-01 -9.11504507e-01 -2.62289941e-01
-4.47616011e-01 -9.24317777e-01 5.55661321e-01 1.14668477e+00
1.17253041e+00 5.19549906e-01 -5.97045600e-01 -4.22069609e-01
-3.27626258e-01 -3.38629812e-01 9.34051633e-01 -9.86403763e-01
1.14136390e-01 -8.00592244e-01 -5.44264972e-01 -4.08948272e-01
2.45519981e-01 -7.26760626e-01 1.04762860e-01 -1.13529611e+00
5.59774926e-03 -3.65535989e-02 5.07068872e-01 3.45153689e-01
2.36979216e-01 3.52657646e-01 2.45088309e-01 -8.97682756e-02
-5.60078919e-01 6.45571411e-01 7.75432169e-01 1.20039552e-01
-4.23026711e-01 -3.60798448e-01 -7.58812904e-01 6.10257030e-01
8.82532001e-01 -4.67835516e-02 -3.62887353e-01 3.58225644e-01
3.01970869e-01 1.94919854e-01 1.69750541e-01 -8.49384725e-01
-3.71316895e-02 1.73465256e-02 -4.93811853e-02 -3.28032643e-01
9.98558700e-01 -6.21592224e-01 3.56085300e-01 -8.19549188e-02
-9.31673110e-01 -6.16620891e-02 4.75839823e-01 -2.19532903e-02
-3.06610882e-01 -1.45056143e-01 5.96067369e-01 2.95208007e-01
-5.94172239e-01 -3.79195064e-01 -1.02535582e+00 2.44215235e-01
1.02916110e+00 -6.72393739e-01 -2.58695304e-01 -1.27322853e+00
-7.32494473e-01 -6.37070462e-02 2.23804817e-01 7.54499257e-01
5.62293231e-01 -1.44815552e+00 -8.05461287e-01 -2.15209350e-02
4.70815122e-01 -6.93354845e-01 3.95187169e-01 5.84188998e-01
-2.82211363e-01 6.74886107e-02 -1.76190794e-01 -2.57240355e-01
-1.56211042e+00 -1.41287044e-01 3.18425924e-01 3.68386269e-01
-8.27175200e-01 8.39942157e-01 -4.84422714e-01 -3.37799430e-01
2.78005004e-01 1.82120010e-01 -5.51217616e-01 5.99403262e-01
9.00433958e-01 6.29039466e-01 -2.26801485e-01 -1.17482698e+00
-1.91747963e-01 -1.63149834e-01 3.59959364e-01 -6.62066638e-01
1.38186157e+00 -4.00275797e-01 4.95371744e-02 1.14842021e+00
1.43386328e+00 1.74790874e-01 -8.75490427e-01 1.27578139e-01
5.47881946e-02 -3.89222920e-01 -8.10320899e-02 -9.03349757e-01
-4.77748156e-01 5.93672216e-01 2.12897360e-01 9.25474703e-01
8.84231210e-01 1.31689876e-01 9.26916897e-01 1.38037845e-01
8.88588838e-03 -1.46358263e+00 2.82627583e-01 5.76245070e-01
1.41801858e+00 -9.94781613e-01 -4.75695074e-01 -8.76585767e-02
-1.55394220e+00 1.29578280e+00 3.06336135e-01 3.74704331e-01
2.79029578e-01 6.06806576e-01 4.72969711e-01 -5.15186489e-01
-1.06555629e+00 3.07743639e-01 -1.01170167e-01 6.93568587e-01
7.97940373e-01 4.87644523e-01 -2.61492133e-01 1.22723055e+00
-9.47482049e-01 -1.27013296e-01 6.03402674e-01 7.88838565e-01
-1.76073849e-01 -4.52286631e-01 -3.50794584e-01 1.87993854e-01
-5.81023216e-01 2.74684817e-01 -1.13256669e+00 5.85169554e-01
-4.09639329e-01 1.61679900e+00 1.26637712e-01 -5.51996827e-01
6.26277864e-01 4.95305300e-01 -1.19777896e-01 -3.45531970e-01
-1.20415223e+00 -2.31232181e-01 1.01993632e+00 -4.80994850e-01
-4.61216778e-01 -7.48925447e-01 -1.47279298e+00 -5.91555536e-01
-4.09676284e-02 4.43130255e-01 9.61214304e-01 7.58696437e-01
5.12962878e-01 6.67698145e-01 1.01839828e+00 -1.00244200e+00
-1.59187347e-01 -1.31507671e+00 -6.01531148e-01 6.36700094e-01
1.60837397e-01 -3.85858655e-01 -6.05879545e-01 2.53504336e-01] | [13.090340614318848, 6.142116546630859] |
28714e73-9bb3-4155-9930-be228f51587a | a-multi-cell-experimental-design-to-recover | 2302.13857 | null | https://arxiv.org/abs/2302.13857v2 | https://arxiv.org/pdf/2302.13857v2.pdf | A multi-cell experimental design to recover policy relevant treatment effects, with an application to online advertising | Experiments are an important tool to measure the impacts of interventions. However, in experimental settings with one-sided noncompliance, extant empirical approaches may not produce the estimands a decision-maker needs to solve their problem. For example, these experimental designs are common in digital advertising settings, but typical methods do not yield effects that inform the intensive margin -- how much should be spent or how many consumers should be reached with a campaign. We propose a solution that combines a novel multi-cell experimental design with modern estimation techniques that enables decision-makers to recover enough information to solve problems with an intensive margin. Our design is straightforward to implement. Using data from advertising experiments at Facebook, we demonstrate our approach outperforms standard techniques in recovering treatment effect parameters. Through a simple advertising reach decision problem, we show that our approach generates better decisions relative to standard techniques. | ['Brett R. Gordon', 'Caio Waisman'] | 2023-02-27 | null | null | null | null | ['experimental-design'] | ['methodology'] | [ 3.66988033e-01 1.17733411e-01 -1.17583680e+00 -5.68522155e-01
-9.45494294e-01 -6.60368085e-01 3.54264379e-01 1.89300030e-01
-7.19920754e-01 9.73983824e-01 3.90253663e-01 -9.88605857e-01
-1.54404685e-01 -7.36181200e-01 -9.73786175e-01 -2.77928114e-01
9.26544219e-02 4.86581624e-01 -2.49983624e-01 2.73009986e-01
3.61807674e-01 6.95581734e-02 -1.24518108e+00 2.05923975e-01
7.85393596e-01 6.63208187e-01 -2.58361667e-01 3.45326781e-01
2.76136577e-01 4.61043775e-01 -3.60584795e-01 -4.04685259e-01
6.57895386e-01 -3.34633976e-01 -1.57334417e-01 3.55101138e-01
5.04893720e-01 -8.95236313e-01 -1.63757652e-01 9.91669834e-01
4.72581595e-01 -2.60470748e-01 6.64809942e-01 -1.08187973e+00
-4.75273877e-01 1.06473553e+00 -9.46118653e-01 1.76865414e-01
2.18509346e-01 3.41732830e-01 1.35369849e+00 1.43262699e-01
7.30670571e-01 1.65335238e+00 3.79573315e-01 1.02108210e-01
-2.09066296e+00 -9.69707251e-01 2.42079183e-01 -2.88841218e-01
-9.42519188e-01 -7.08349049e-01 4.22166854e-01 -4.16728377e-01
7.42316060e-03 3.50516558e-01 5.28471410e-01 1.48626959e+00
1.95685789e-01 8.07274640e-01 1.97904289e+00 -2.86765784e-01
5.70265651e-01 2.89515495e-01 4.79900479e-01 8.50533620e-02
8.30218732e-01 6.78057969e-01 -3.66334133e-02 -5.70393443e-01
7.66587615e-01 8.46397653e-02 -3.37080359e-02 -2.67703623e-01
-7.17373788e-01 1.39877474e+00 1.22107796e-01 6.80379495e-02
-5.46737194e-01 2.71815300e-01 9.06004459e-02 5.60308814e-01
4.20411140e-01 5.88669717e-01 -5.03703356e-01 6.87609911e-02
-8.87277246e-01 9.37433839e-01 9.03993964e-01 7.26395667e-01
5.38298786e-01 -4.94340807e-01 -6.75669432e-01 3.20142567e-01
2.08043665e-01 5.54976225e-01 1.49610236e-01 -1.26938748e+00
4.80306685e-01 2.95858026e-01 8.23792398e-01 -6.01189315e-01
-3.41464132e-01 -1.79282188e-01 -3.53519738e-01 -1.24185113e-02
9.65103388e-01 -5.20754218e-01 -6.69468164e-01 1.85587943e+00
3.95307600e-01 -1.74735896e-02 -4.68441725e-01 6.90117776e-01
-1.99960217e-01 3.28197777e-01 3.83537948e-01 -6.68091178e-01
1.60774684e+00 -2.30700329e-01 -7.62294352e-01 -5.50300062e-01
8.77069354e-01 -6.55391872e-01 1.19915712e+00 3.02700937e-01
-1.13695800e+00 7.88486153e-02 -6.57479286e-01 3.90817285e-01
6.34293780e-02 -3.14185530e-01 1.13207483e+00 1.06778276e+00
-5.43954432e-01 5.01686513e-01 -1.72948301e-01 -2.13325039e-01
6.66921735e-01 4.01274920e-01 1.93149984e-01 -3.16443026e-01
-8.47478509e-01 5.41863263e-01 -2.51302987e-01 -4.08372283e-01
-8.67175698e-01 -1.17402065e+00 -3.96653861e-01 3.47949237e-01
9.75753009e-01 -6.51906908e-01 1.82850564e+00 -9.20433998e-01
-1.31488729e+00 4.07791704e-01 1.00688562e-01 -6.57973647e-01
6.02594435e-01 2.78074592e-01 3.92315723e-02 -4.61616993e-01
4.65366513e-01 3.55466247e-01 6.38749957e-01 -1.05284834e+00
-6.65232778e-01 -7.47181952e-01 1.05754547e-01 -1.98329613e-01
-3.21585126e-02 2.67135501e-01 3.35489959e-01 -4.61241454e-01
-2.57225007e-01 -1.07539761e+00 -7.35843301e-01 -1.48122162e-01
-7.00147271e-01 -6.04775595e-03 2.69139916e-01 -1.92713916e-01
1.22815228e+00 -1.87526631e+00 -4.72180039e-01 3.86141747e-01
2.16327563e-01 -1.67177886e-01 -2.03371793e-01 3.48094821e-01
2.93081641e-01 4.66984212e-01 3.41833271e-02 -9.92136821e-02
6.88620031e-01 -3.57032451e-03 -2.82890826e-01 5.94002545e-01
-3.53971303e-01 7.59351790e-01 -3.75614136e-01 -3.22243631e-01
-1.90125838e-01 -8.32774118e-02 -1.01438105e+00 -1.70546651e-01
-5.38329422e-01 2.33002856e-01 -8.28606725e-01 4.58907455e-01
7.85674512e-01 -2.49089897e-01 8.11628580e-01 2.50535786e-01
-2.06405818e-01 4.11900699e-01 -1.19776630e+00 8.75159264e-01
-4.24167097e-01 3.54021639e-01 5.45138896e-01 -1.21449375e+00
1.45399109e-01 -1.08226597e-01 3.20734411e-01 -9.58495557e-01
4.77192700e-01 1.36846319e-01 3.51134807e-01 -3.69554788e-01
1.20520815e-01 -3.48392427e-01 -3.78234386e-01 7.80841827e-01
-5.93128800e-01 6.15731440e-02 2.72153556e-01 4.42845784e-02
1.34412670e+00 -5.43508887e-01 3.89445454e-01 -4.65939373e-01
-3.76542270e-01 8.36378112e-02 8.28978956e-01 1.19109893e+00
-2.22315028e-01 -1.65760681e-01 1.09221780e+00 -6.55252859e-02
-1.09407520e+00 -6.93151712e-01 -3.48173916e-01 1.03308105e+00
-3.65182996e-01 3.43918838e-02 -7.95796931e-01 -5.48417568e-01
5.98767161e-01 1.08347476e+00 -9.19570327e-01 2.95363605e-01
-7.48614967e-03 -1.00177443e+00 1.25940191e-02 2.32944116e-01
3.35520566e-01 -3.66336286e-01 -1.21976800e-01 2.43852869e-01
1.56593427e-01 -1.04339612e+00 -7.31345057e-01 -2.75047213e-01
-9.73305762e-01 -1.13573527e+00 -4.24909025e-01 -2.81316012e-01
5.59664488e-01 2.98749477e-01 9.07413900e-01 -2.99255490e-01
-1.22547351e-01 1.64577261e-01 1.52013391e-01 -5.05780637e-01
-4.70502526e-01 -1.80086210e-01 2.01780312e-02 1.03773281e-01
8.97723436e-01 -6.08727455e-01 -9.15739298e-01 3.97529304e-01
-6.91011965e-01 -4.80706096e-01 1.03367972e+00 5.66249430e-01
3.65939498e-01 -1.05352193e-01 8.39608431e-01 -1.43204331e+00
1.03737855e+00 -6.63849115e-01 -1.28391016e+00 -5.69620589e-03
-1.05134881e+00 2.03087777e-01 2.88778663e-01 -1.02042937e+00
-7.67556071e-01 -1.28084928e-01 4.03017521e-01 2.81239063e-01
-1.71278194e-01 4.20987904e-01 -9.19281840e-02 8.98467898e-02
8.39057088e-01 -5.06247401e-01 1.87399313e-01 -5.95811903e-01
3.48208308e-01 9.97563362e-01 -3.54961514e-01 -7.44155169e-01
4.08860564e-01 2.50285119e-01 -4.12783660e-02 -4.71249700e-01
-8.97619069e-01 -9.28130597e-02 2.46727765e-01 2.71534026e-01
5.04481494e-01 -1.08857167e+00 -1.31265175e+00 -2.67525017e-01
-4.92039770e-01 -6.63802862e-01 -4.00029123e-01 8.92234325e-01
-5.46026528e-01 -2.25865226e-02 -6.11757517e-01 -1.00582802e+00
2.54481316e-01 -1.35761404e+00 9.20251369e-01 9.14155915e-02
-2.71208256e-01 -7.93947458e-01 3.27206850e-02 5.95172822e-01
4.54865694e-01 -1.61868986e-03 8.69994521e-01 -6.65084422e-01
-8.53236496e-01 -3.97187680e-01 -2.80205160e-01 -3.11657041e-01
-2.11497903e-01 -2.91901290e-01 -6.42886281e-01 -3.32042307e-01
-1.32832676e-01 -2.91487217e-01 5.43483734e-01 1.41636777e+00
1.19691455e+00 -8.96092057e-01 -5.64441681e-01 -5.46907596e-02
1.08697891e+00 5.81424236e-02 6.39014006e-01 1.78091407e-01
-2.06119552e-01 5.76129615e-01 5.46928167e-01 7.80198276e-01
5.07527351e-01 9.96181250e-01 -6.39112368e-02 2.78694145e-02
2.78665751e-01 -5.76833189e-01 4.58698392e-01 -9.81700420e-02
4.75127190e-01 -2.06196606e-01 8.16417951e-03 3.74834001e-01
-1.71660423e+00 -8.55908871e-01 -1.50603443e-01 2.67029047e+00
1.08791494e+00 2.99165159e-01 8.11253905e-01 -1.10721752e-01
5.10102272e-01 -1.54511198e-01 -5.56234300e-01 -4.86185849e-01
1.59867287e-01 1.76311970e-01 1.12893057e+00 6.69421315e-01
-5.97750843e-01 6.03574455e-01 7.53580141e+00 7.90869236e-01
-5.11617243e-01 2.86322355e-01 8.78225505e-01 -3.17333162e-01
-5.26167750e-01 5.09019613e-01 -1.05530822e+00 6.12935245e-01
1.40037584e+00 -3.17996949e-01 3.93878400e-01 5.67437291e-01
8.46569419e-01 -5.59487164e-01 -1.35377407e+00 5.23652196e-01
-5.19319892e-01 -1.12214136e+00 -5.57338893e-01 9.18855906e-01
7.61508048e-01 -6.00410759e-01 3.17928880e-01 5.05270660e-01
1.26303661e+00 -8.77053142e-01 3.30582768e-01 1.91448674e-01
5.55770516e-01 -3.24816287e-01 3.20323497e-01 5.10491669e-01
-3.08168083e-01 -7.53261209e-01 -3.99252474e-01 -4.83240545e-01
-1.94904543e-02 1.10385609e+00 -9.33709264e-01 -2.06627205e-01
1.40842497e-01 4.83795069e-02 -2.81654060e-01 1.05239534e+00
-1.45561576e-01 9.83230591e-01 -4.48622674e-01 -3.74767721e-01
1.12535112e-01 -5.01266301e-01 1.32634073e-01 5.59874594e-01
4.72205818e-01 1.69102564e-01 2.29661226e-01 1.09247494e+00
-4.06951904e-01 3.71525496e-01 -6.89227343e-01 -3.64826679e-01
6.17729306e-01 9.11945999e-01 -4.27780539e-01 -2.28678823e-01
-5.24445772e-01 2.46586949e-01 1.14726471e-02 4.83762532e-01
-6.00460768e-01 2.60336399e-01 7.88349926e-01 6.61915720e-01
1.18586041e-01 4.27422151e-02 -2.85403401e-01 -8.91571045e-01
-2.44688064e-01 -1.21893799e+00 3.96823764e-01 -4.24403161e-01
-1.26944578e+00 -6.99310899e-01 3.06884468e-01 -5.69276690e-01
-1.54063866e-01 -5.60140371e-01 -3.23852837e-01 4.62419361e-01
-1.18782294e+00 -6.01801157e-01 4.05347228e-01 4.01250906e-02
4.45284098e-01 3.27915758e-01 3.56449604e-01 4.15531963e-01
-7.85183489e-01 6.33185744e-01 2.87644535e-01 -3.69313836e-01
6.76588476e-01 -1.09572041e+00 -1.78581685e-01 4.48593140e-01
-2.54667699e-01 4.90827978e-01 9.62427676e-01 -6.94664955e-01
-1.58712292e+00 -5.86834550e-01 6.29020631e-01 -4.33674529e-02
1.04459023e+00 -5.47758162e-01 -1.47981495e-01 1.16842878e+00
5.61403930e-02 -5.14356554e-01 7.00573742e-01 7.23268390e-01
-2.76580185e-01 -3.85419726e-01 -1.28015828e+00 1.01191652e+00
9.79627311e-01 -1.43820226e-01 -6.52782694e-02 7.82077491e-01
5.83071887e-01 1.06508553e-01 -8.95694077e-01 -1.11736134e-01
5.83040893e-01 -7.22123504e-01 6.29346669e-01 -8.33529234e-01
1.82538718e-01 2.93293864e-01 -3.73831689e-01 -1.27418864e+00
-3.11547518e-01 -9.57487404e-01 2.63132095e-01 1.06908417e+00
8.08381081e-01 -8.72886896e-01 7.68270493e-01 1.01236951e+00
5.97309232e-01 -1.58478662e-01 -6.80653691e-01 -6.20885313e-01
5.57716966e-01 -1.32337049e-01 7.95658529e-01 8.76158953e-01
-1.98240072e-01 5.24002135e-01 -3.22360128e-01 -5.06347939e-02
1.00906408e+00 3.30910861e-01 1.01238310e+00 -1.48723733e+00
-5.62514544e-01 -4.52420145e-01 -1.69126280e-02 -1.18176317e+00
3.75409961e-01 -3.66984338e-01 -2.55288869e-01 -1.08185244e+00
7.23423004e-01 -7.18441546e-01 1.71235800e-01 2.47085124e-01
-9.86994058e-03 -2.31860504e-01 -1.33713782e-01 -2.58613378e-01
-2.61257201e-01 1.06041990e-01 1.22614431e+00 -9.22341198e-02
-4.50233579e-01 4.45196986e-01 -1.46283996e+00 3.59131575e-01
7.43519843e-01 -7.65596271e-01 -4.15848523e-01 1.67061910e-01
3.33987802e-01 4.35029209e-01 3.62442344e-01 -2.93620259e-01
-4.03581485e-02 -7.44775772e-01 2.62333304e-01 -1.23699278e-01
-8.55512694e-02 -8.55001986e-01 4.70406681e-01 4.64409530e-01
-7.86661148e-01 -2.41806075e-01 -2.10777260e-02 7.02265382e-01
5.27819335e-01 -3.47232014e-01 5.45633852e-01 -6.54304698e-02
2.56748289e-01 2.39148438e-01 -5.44282317e-01 1.77838221e-01
1.01656032e+00 3.90961230e-01 -6.25049055e-01 -7.02800512e-01
-6.42205894e-01 2.80876964e-01 5.90079904e-01 -5.63205220e-02
-1.23242043e-01 -1.20042157e+00 -8.46315086e-01 -1.27371639e-01
-1.11386955e-01 -9.58516717e-01 5.41424155e-02 1.10152316e+00
1.89247310e-01 3.64138603e-01 2.41555229e-01 -1.38599843e-01
-9.86868799e-01 9.00111735e-01 1.54242873e-01 -4.79814917e-01
-2.64352202e-01 2.53242940e-01 5.28327584e-01 -1.51304845e-02
5.58570251e-02 -4.80627835e-01 2.89148450e-01 2.71654814e-01
6.50448322e-01 1.73286721e-01 -3.29711825e-01 1.75874084e-02
1.86770231e-01 5.77159822e-02 -2.86169708e-01 -4.51293737e-01
1.16559076e+00 -3.22140604e-01 3.40442777e-01 2.86137611e-01
1.01207435e+00 3.14786732e-01 -1.32021058e+00 -3.18827003e-01
3.40548567e-02 -8.31568837e-01 4.76987809e-01 -7.89332747e-01
-9.45134342e-01 3.58645767e-01 5.96180797e-01 6.23870254e-01
7.45702505e-01 -2.63706893e-01 3.46409470e-01 2.46895403e-01
3.23509932e-01 -1.16031909e+00 -2.04793110e-01 -6.36289001e-01
3.48788947e-01 -1.35895455e+00 3.66079390e-01 -4.26120549e-01
-3.55715990e-01 1.84848905e-01 1.33716181e-01 5.22576384e-02
6.48405910e-01 3.39313835e-01 -3.23223889e-01 -2.53735989e-01
-8.60113800e-01 -3.40771586e-01 -2.33670339e-01 3.69046003e-01
1.84958562e-01 5.19770741e-01 -9.93579209e-01 6.48751676e-01
-4.74519357e-02 1.69656485e-01 1.00551915e+00 8.65863800e-01
-4.57041740e-01 -1.40753388e+00 -6.05617285e-01 1.13346028e+00
-8.42189372e-01 -4.55864519e-02 -3.62415761e-01 1.05980408e+00
-4.73847836e-01 1.26146019e+00 1.52553961e-01 1.56451073e-02
5.55906951e-01 -1.61893055e-01 4.72053826e-01 -4.43095922e-01
-2.68977314e-01 4.86092985e-01 4.49824691e-01 -8.00227463e-01
-3.94842505e-01 -9.37420428e-01 -3.39381278e-01 -8.83851409e-01
-6.59524083e-01 2.07917973e-01 5.56768060e-01 7.77813315e-01
5.07664382e-01 1.33488283e-01 1.09630013e+00 -4.30757284e-01
-1.26612723e+00 -1.09217989e+00 -1.08015645e+00 2.89184958e-01
1.13602035e-01 -8.70029688e-01 -6.89042926e-01 -4.59104866e-01] | [8.095216751098633, 5.3411126136779785] |
abb0456c-b1a1-42ee-ad2d-44103134203b | sonnet-a-self-guided-ordinal-regression | null | null | https://ieeexplore.ieee.org/document/9709151 | https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=9709151 | SONNET: A Self-Guided Ordinal Regression Neural Network for Segmentation and Classification of Nuclei in Large-Scale Multi-Tissue Histology Images | Automated nuclei segmentation and classification are the keys to analyze and understand the cellular characteristics and functionality, supporting computer-aided digital pathology in disease diagnosis. However, the task still remains challenging due to the intrinsic variations in size, intensity, and morphology of different types of nuclei. Herein, we propose a self-guided ordinal regression neural network for simultaneous nuclear segmentation and classification that can exploit the intrinsic characteristics of nuclei and focus on highly uncertain areas during training. The proposed network formulates nuclei segmentation as an ordinal regression learning by introducing a distance decreasing discretization strategy, which stratifies nuclei in a way that inner regions forming a regular shape of nuclei are separated from outer regions forming an irregular shape. It also adopts a self-guided training strategy to adaptively adjust the weights associated with nuclear pixels, depending on the difficulty of the pixels that is assessed by the network itself. To evaluate the performance of the proposed network, we employ large-scale multi-tissue datasets with 276349 exhaustively annotated nuclei. We show that the proposed network achieves the state-of-the-art performance in both nuclei segmentation and classification in comparison to several methods that are recently developed for segmentation and/or classification. | ['Jin T. Kwak', 'Kyungeun Kim', 'Trinh T. L. Vuong', 'Boram Song', 'Tan N. N. Doan'] | 2022-02-09 | null | null | null | ieee-journal-of-biomedical-and-health-2 | ['multi-tissue-nucleus-segmentation', 'nuclear-segmentation', 'nuclei-classification'] | ['medical', 'medical', 'medical'] | [ 3.49066645e-01 1.02208316e-01 -2.10253984e-01 -3.43794852e-01
-7.78407574e-01 -4.65277791e-01 1.14116341e-01 7.62464404e-01
-8.80880833e-01 7.37381160e-01 -2.16093838e-01 -3.05568557e-02
-1.55905321e-01 -7.84423292e-01 -3.02589446e-01 -1.31945276e+00
-1.85690001e-02 7.75845289e-01 3.30947012e-01 1.85155541e-01
1.46820650e-01 8.45389307e-01 -1.34401023e+00 6.61259666e-02
9.85504210e-01 1.03709996e+00 7.33451694e-02 5.24658501e-01
-1.01997353e-01 3.04813802e-01 -2.49103338e-01 1.77495763e-01
-4.34647612e-02 -1.96970448e-01 -6.65530324e-01 3.21354747e-01
2.26878986e-01 3.16724367e-02 -2.82347854e-02 1.20033443e+00
3.92358720e-01 4.25712653e-02 1.20497608e+00 -6.08111382e-01
-3.09089541e-01 4.03249115e-01 -5.26063919e-01 3.57380509e-01
-3.94884199e-01 7.99548551e-02 9.46175158e-01 -6.81755185e-01
6.03860021e-01 6.24065340e-01 6.58965170e-01 4.51615214e-01
-1.34005380e+00 -4.21322525e-01 -1.00103021e-01 -1.27842098e-01
-1.78341305e+00 -2.19430670e-01 5.96957505e-01 -7.19830632e-01
4.20323193e-01 8.00062120e-02 6.43259883e-01 2.44115084e-01
9.59383100e-02 7.21424103e-01 1.03056145e+00 -2.36806124e-01
3.23383778e-01 2.71474496e-02 2.97682226e-01 8.09242189e-01
2.34181270e-01 -5.37024319e-01 2.21855626e-01 -4.86682216e-03
7.98516870e-01 5.27564716e-03 -1.73971385e-01 -2.83031493e-01
-1.14715397e+00 6.79798603e-01 5.71336389e-01 7.27698386e-01
-4.09789056e-01 -8.33950117e-02 4.50772732e-01 -2.86417872e-01
4.51133728e-01 3.44616532e-01 -3.85685563e-01 4.17854458e-01
-1.33772635e+00 -3.41030583e-02 2.93629080e-01 2.71406651e-01
7.86272347e-01 -6.24747127e-02 -2.83058196e-01 9.57612395e-01
3.05215001e-01 -2.65513398e-02 6.41057312e-01 -9.11711097e-01
-8.98744166e-02 1.06862569e+00 -6.38341680e-02 -9.61533070e-01
-8.47855031e-01 -6.35445178e-01 -1.38808322e+00 1.69591799e-01
9.00995195e-01 -5.92133030e-03 -1.00273716e+00 1.43453205e+00
6.12149775e-01 -2.21808292e-02 -1.06922820e-01 6.74542844e-01
8.89088154e-01 5.49940467e-01 6.97459206e-02 -5.24733901e-01
1.22470903e+00 -6.92817211e-01 -6.00102961e-01 2.43890360e-01
7.57567346e-01 -1.51649401e-01 7.72873461e-01 3.07774723e-01
-9.24858928e-01 -4.08033013e-01 -9.38012779e-01 2.83651520e-03
-4.85813200e-01 4.81123656e-01 3.58601332e-01 4.36701894e-01
-9.40898776e-01 5.79593480e-01 -8.48766387e-01 -2.39565134e-01
7.78917849e-01 7.72802889e-01 -3.56427073e-01 3.73864979e-01
-8.62932086e-01 3.74199569e-01 7.02978253e-01 4.26162750e-01
-2.52576590e-01 -6.09138489e-01 -7.03366697e-01 -7.84025353e-04
1.97809622e-01 -3.08633268e-01 6.64786637e-01 -7.14521527e-01
-1.23592019e+00 1.15794837e+00 3.37421708e-02 -2.76940972e-01
3.75960022e-01 6.04081810e-01 -8.02926626e-03 3.43714088e-01
-2.47201063e-02 1.11358249e+00 4.04872537e-01 -1.09359717e+00
-7.06155598e-01 -5.55957317e-01 -2.92152256e-01 2.19808772e-01
-5.33184826e-01 -4.46534753e-01 -6.23967946e-01 -5.84704280e-01
4.29942161e-01 -6.75392330e-01 -4.95581448e-01 3.23190451e-01
-3.49497497e-01 -3.36690128e-01 6.17585897e-01 -5.67978799e-01
1.09220874e+00 -2.11432171e+00 3.41272384e-01 5.75935602e-01
3.99299890e-01 5.46174124e-02 2.36168757e-01 -4.40309376e-01
2.58985043e-01 2.57194161e-01 -5.17668247e-01 -1.00714877e-01
-3.02237034e-01 2.20688045e-01 4.31652486e-01 8.20428193e-01
2.14862987e-01 5.88904023e-01 -8.20126534e-01 -1.01289582e+00
3.75953056e-02 3.99578691e-01 -4.68642473e-01 1.59060493e-01
-1.48237869e-01 4.39341515e-01 -4.47010100e-01 1.07347155e+00
4.71771389e-01 -4.62252885e-01 3.51565659e-01 -4.20958489e-01
-1.12609535e-01 -4.82087940e-01 -1.10287893e+00 1.06619740e+00
5.75528070e-02 4.33757097e-01 2.95450240e-01 -1.10982800e+00
8.61979842e-01 2.42580801e-01 7.14438081e-01 -1.72112972e-01
5.68591297e-01 3.26306462e-01 1.05819300e-01 -2.03869119e-01
3.46083075e-01 -2.90318161e-01 4.04548012e-02 9.99403000e-02
-6.13273308e-03 -1.18648402e-01 5.45033991e-01 -2.62150824e-01
8.91751289e-01 -3.03097188e-01 4.13748145e-01 -4.08291131e-01
8.68721485e-01 -1.47502705e-01 8.66228819e-01 4.92178828e-01
-5.10751009e-01 6.96019173e-01 7.33469427e-01 -3.12359691e-01
-9.53256905e-01 -9.13626075e-01 -4.09350902e-01 8.28646600e-01
7.13607967e-02 2.96810031e-01 -9.09534276e-01 -8.13792109e-01
-5.95984049e-02 5.52465618e-02 -9.95890498e-01 1.86023936e-01
-4.69109088e-01 -1.22125268e+00 6.36983812e-01 5.93583226e-01
2.96538800e-01 -1.13078773e+00 -3.21817189e-01 1.36135682e-01
-1.57372162e-01 -1.04485273e+00 -3.57107520e-01 4.97162342e-01
-1.11421824e+00 -1.05327976e+00 -8.28935921e-01 -1.00518012e+00
1.33153749e+00 -2.66947806e-01 7.56334484e-01 4.20759887e-01
-3.90229255e-01 -9.78757292e-02 -5.27780168e-02 1.26798227e-01
-3.12828660e-01 4.16651905e-01 -1.45009011e-01 7.55838156e-02
-1.21997654e-01 -3.77351731e-01 -6.39427423e-01 3.18605810e-01
-1.08564496e+00 -2.12220728e-01 6.93592310e-01 1.07418954e+00
1.25644624e+00 4.75806206e-01 4.54643756e-01 -1.18858790e+00
1.97924048e-01 -4.58456248e-01 -6.62386239e-01 2.19320297e-01
-5.23526967e-01 -1.08855642e-01 6.72613084e-01 -4.50186133e-01
-6.65682197e-01 4.20829326e-01 -2.14897886e-01 -1.33131400e-01
-1.76525041e-01 5.81328571e-01 -3.83339673e-02 -3.55179638e-01
3.92077684e-01 2.12848499e-01 1.99761633e-02 -4.76240506e-03
7.73924813e-02 6.20500207e-01 6.46123052e-01 -3.02424252e-01
5.43913364e-01 6.54397070e-01 2.43019655e-01 -7.01020956e-01
-6.37533784e-01 -6.50724888e-01 -8.70864928e-01 -3.47379774e-01
9.82070148e-01 -4.93608832e-01 -7.97398210e-01 6.25989437e-01
-6.36178792e-01 -3.92512143e-01 -3.24143052e-01 2.19613910e-01
-4.07013059e-01 3.66155684e-01 -9.02673244e-01 -6.38217330e-01
-3.58092457e-01 -1.09510291e+00 8.36089671e-01 4.69662309e-01
-1.38756692e-01 -9.71669734e-01 6.30024029e-03 4.29306120e-01
9.11808386e-03 4.14706290e-01 1.29279387e+00 -7.48353779e-01
-3.05524737e-01 -2.92468637e-01 -3.46972644e-01 3.24701339e-01
3.39914739e-01 3.13844889e-01 -5.72621465e-01 -2.27826834e-01
-2.79872805e-01 -4.81646180e-01 1.00558162e+00 7.47202516e-01
1.40546572e+00 -1.09648481e-01 -4.46065485e-01 6.97661161e-01
1.54255879e+00 1.83042169e-01 5.44918001e-01 3.06647658e-01
5.11294782e-01 6.97515965e-01 5.58370709e-01 2.74934798e-01
2.33875886e-01 1.89194888e-01 3.66102189e-01 -5.09080589e-01
2.78152466e-01 3.35254282e-01 -3.38413328e-01 6.33766592e-01
-1.02408439e-01 -1.91934451e-01 -8.89827549e-01 8.01289260e-01
-1.58382690e+00 -5.79978168e-01 5.18944152e-02 1.78304374e+00
1.08862746e+00 3.84241611e-01 9.29486528e-02 3.94936651e-01
8.17323446e-01 -5.43400198e-02 -6.50446236e-01 -4.37708385e-02
-1.41897827e-01 1.23812564e-01 5.35961211e-01 3.02495629e-01
-1.26248145e+00 6.78918481e-01 6.18685150e+00 1.09780991e+00
-1.13524377e+00 -2.41051555e-01 1.37454355e+00 2.84696877e-01
-4.76849191e-02 -6.11866474e-01 -8.21753800e-01 4.14390653e-01
3.19251031e-01 3.53228390e-01 1.36929182e-02 4.58210111e-01
2.38171011e-01 -2.39135325e-01 -7.99577653e-01 4.55846161e-01
-1.82361364e-01 -1.41003001e+00 5.63186295e-02 -2.51898244e-02
8.12053680e-01 -2.62432694e-01 1.24733120e-01 1.01665080e-01
-1.13896556e-01 -1.03272259e+00 2.76953012e-01 7.29721487e-01
7.54453599e-01 -7.12492108e-01 1.19359457e+00 4.11003947e-01
-1.04699349e+00 -1.24564178e-01 -1.76035479e-01 3.44480485e-01
-2.44523570e-01 6.58534050e-01 -1.00248742e+00 2.38648623e-01
3.75413537e-01 3.66701216e-01 -5.44344842e-01 1.10285735e+00
3.31493497e-01 6.55610144e-01 -4.41345870e-01 -4.54125069e-02
2.74610370e-01 -4.35115367e-01 1.04226880e-01 1.28414369e+00
2.10324049e-01 2.82133758e-01 1.91351131e-01 7.93752372e-01
-2.24835530e-01 4.42422956e-01 8.67048502e-02 -1.84625372e-01
2.67822534e-01 1.63466346e+00 -1.67497671e+00 -2.10404709e-01
-6.76788986e-02 3.38358551e-01 5.15157759e-01 1.96519434e-01
-6.23738706e-01 -3.07612717e-01 1.31073445e-01 3.79824221e-01
3.22693974e-01 -9.28019881e-02 -5.51655233e-01 -5.28381407e-01
-3.97584051e-01 -4.94445950e-01 6.48045778e-01 -3.03173006e-01
-1.21383917e+00 5.89849710e-01 -3.21375817e-01 -1.01240158e+00
1.43995047e-01 -5.96041799e-01 -3.32689583e-01 3.35927725e-01
-1.29985392e+00 -9.93067443e-01 -3.39154959e-01 1.07811056e-01
3.26486886e-01 5.39330877e-02 5.62768638e-01 2.68559426e-01
-7.28810430e-01 5.60570657e-01 3.78764808e-01 3.20317209e-01
3.86761189e-01 -1.41989315e+00 -4.27078038e-01 3.40315074e-01
-2.97082156e-01 3.94273341e-01 3.90776306e-01 -5.65705121e-01
-6.01758480e-01 -1.20471847e+00 6.14235640e-01 2.35568747e-01
4.88786370e-01 -1.65630475e-01 -1.03056002e+00 2.85904378e-01
-2.76004583e-01 5.53028584e-01 8.56434524e-01 -8.91635343e-02
2.47029945e-01 -2.98616946e-01 -1.55201256e+00 4.64202762e-01
3.64002824e-01 -2.46866912e-01 -1.37617007e-01 2.27667436e-01
2.62572348e-01 -3.80672306e-01 -1.06890953e+00 7.14182913e-01
5.02523243e-01 -7.47941077e-01 7.89305449e-01 -1.89592615e-01
9.05605927e-02 -5.96896410e-01 4.12698351e-02 -8.39732528e-01
-4.91479844e-01 -9.18492861e-03 1.41928539e-01 1.12704718e+00
4.39936161e-01 -3.99416000e-01 1.30729544e+00 1.96592748e-01
-9.62275490e-02 -1.35940230e+00 -1.03514028e+00 -2.78037041e-01
1.08405001e-01 1.14059992e-01 1.48025319e-01 7.73650110e-01
-2.70428300e-01 -3.85741703e-02 2.42133081e-01 1.55231342e-01
6.65265679e-01 -6.01360463e-02 1.22123249e-01 -1.34993935e+00
6.61563501e-02 -6.49694383e-01 -6.04446530e-01 -7.24566042e-01
2.79783219e-01 -8.51872504e-01 3.11244369e-01 -1.40621221e+00
3.23587656e-01 -6.01337910e-01 -6.74024522e-01 5.48103094e-01
-1.64038703e-01 8.70950997e-01 -3.84930491e-01 1.56688511e-01
-6.96468830e-01 3.70015800e-01 1.23399687e+00 -4.32825863e-01
-2.94111401e-01 8.15187395e-02 -6.34143114e-01 1.04753304e+00
5.79397440e-01 -3.64335030e-01 -7.04129860e-02 2.27055237e-01
2.68282276e-02 9.08339098e-02 1.33394986e-01 -1.12885439e+00
4.70703393e-01 -1.48449376e-01 7.71473408e-01 -9.14108694e-01
1.29064068e-01 -6.92987621e-01 -9.12221819e-02 5.28745055e-01
-4.59031522e-01 -4.59363461e-01 1.09570762e-02 5.77615261e-01
-1.51811332e-01 -3.21580440e-01 1.12022483e+00 -1.50024712e-01
-1.02902330e-01 3.42012912e-01 -7.48657227e-01 1.07241916e-02
9.84762669e-01 -6.71018898e-01 8.26312006e-02 1.61185130e-01
-9.47746277e-01 4.05918419e-01 3.92989725e-01 -4.37947005e-01
4.64703590e-01 -1.12885296e+00 -5.53380311e-01 -4.47728764e-03
4.12267223e-02 3.81836802e-01 4.39899921e-01 9.61385071e-01
-6.71202719e-01 1.20676801e-01 -2.52597749e-01 -8.62586915e-01
-1.35179114e+00 1.11624010e-01 5.96492052e-01 -6.99622631e-01
-1.59464091e-01 8.79583716e-01 1.72052220e-01 -5.58263540e-01
3.26321691e-01 -7.06123769e-01 -8.07335436e-01 2.39594147e-01
4.10199165e-02 5.26259840e-01 1.03468150e-01 -6.68123126e-01
-2.70771861e-01 6.13425732e-01 -2.03070983e-01 3.77471328e-01
1.19469905e+00 6.07649498e-02 -4.23471481e-01 6.07371330e-01
9.54954743e-01 -2.45638743e-01 -1.30352163e+00 -1.54979482e-01
-7.24268705e-02 5.86110540e-02 2.07305104e-01 -6.47762239e-01
-1.10572171e+00 4.54197735e-01 7.19725847e-01 1.14720695e-01
1.23238850e+00 -5.65436520e-02 7.07006752e-01 2.78667331e-01
-1.05676249e-01 -1.29030395e+00 -1.53345580e-03 3.66043478e-01
2.97932267e-01 -1.07707810e+00 2.31598094e-01 -5.65935194e-01
-1.94385290e-01 1.20922697e+00 5.90857625e-01 -8.98667891e-03
5.41536152e-01 3.82205755e-01 -6.64733499e-02 -1.44690573e-01
-4.72311497e-01 -1.03534050e-01 3.34780216e-01 3.36350799e-01
3.03724766e-01 2.01931316e-02 -4.73738790e-01 7.51705647e-01
2.92801350e-01 1.38929049e-02 3.25087130e-01 5.95039546e-01
-6.50819838e-01 -7.55844355e-01 -3.57953161e-01 8.18954051e-01
-5.62840104e-01 9.84379053e-02 -2.81586826e-01 7.53028214e-01
5.55144727e-01 3.89887005e-01 3.14243078e-01 -1.58127937e-02
-4.45362180e-02 -2.99362093e-01 4.88727450e-01 -4.70871359e-01
-6.50153041e-01 2.46080548e-01 -3.55921596e-01 2.02732608e-01
-4.60088193e-01 -7.20590532e-01 -1.90098202e+00 2.30908781e-01
-4.70627725e-01 2.21767560e-01 3.06369781e-01 1.07295251e+00
-2.13188514e-01 6.21014595e-01 6.36419237e-01 -6.99685633e-01
-3.91712636e-01 -8.15136552e-01 -8.77921581e-01 3.35067250e-02
3.56928378e-01 -6.12587094e-01 -3.95317972e-01 1.60517961e-01] | [14.892904281616211, -3.1278457641601562] |
6bdb1bd8-edba-45f6-9851-383a2c05d868 | definition-modeling-to-model-definitions | 2306.08433 | null | https://arxiv.org/abs/2306.08433v1 | https://arxiv.org/pdf/2306.08433v1.pdf | "Definition Modeling: To model definitions." Generating Definitions With Little to No Semantics | Definition Modeling, the task of generating definitions, was first proposed as a means to evaluate the semantic quality of word embeddings-a coherent lexical semantic representations of a word in context should contain all the information necessary to generate its definition. The relative novelty of this task entails that we do not know which factors are actually relied upon by a Definition Modeling system. In this paper, we present evidence that the task may not involve as much semantics as one might expect: we show how an earlier model from the literature is both rather insensitive to semantic aspects such as explicit polysemy, as well as reliant on formal similarities between headwords and words occurring in its glosses, casting doubt on the validity of the task as a means to evaluate embeddings. | ['Timothee Mickus', 'Vincent Segonne'] | 2023-06-14 | null | null | null | null | ['word-embeddings'] | ['methodology'] | [ 2.21869245e-01 2.61105746e-01 -8.90942588e-02 -5.17122030e-01
-2.79471010e-01 -7.08658993e-01 1.10467267e+00 6.85374439e-01
-8.35072279e-01 6.40229285e-01 6.49302244e-01 -6.93829417e-01
-3.06642890e-01 -8.94141912e-01 -2.47391716e-01 -4.33237165e-01
3.23813736e-01 3.14897329e-01 2.31307238e-01 -4.49513227e-01
4.39689070e-01 2.70083815e-01 -1.49259555e+00 2.70088911e-02
5.16501725e-01 5.44575453e-01 2.69654751e-01 2.32778773e-01
-4.68381703e-01 4.03813124e-01 -8.20428014e-01 -6.96214616e-01
-2.26133659e-01 -4.03423578e-01 -9.13274765e-01 -4.81508262e-02
2.81829268e-01 6.63895682e-02 1.16340332e-02 1.15218508e+00
1.58511505e-01 7.07669035e-02 9.32970285e-01 -8.53000343e-01
-1.01962054e+00 6.59879446e-01 2.85879105e-01 4.26781565e-01
4.16399539e-01 1.00809976e-01 1.48933566e+00 -8.97630215e-01
7.46441722e-01 1.06899703e+00 6.87075138e-01 4.96393383e-01
-1.23832560e+00 -1.91062018e-01 3.13221991e-01 -1.28722992e-02
-1.32191336e+00 -4.56679672e-01 5.31737685e-01 -4.90508616e-01
1.10922718e+00 1.76530078e-01 7.06028044e-01 1.18840170e+00
2.09093392e-01 8.65560398e-02 1.08981776e+00 -7.10908592e-01
3.01245749e-01 3.52525622e-01 5.74434996e-01 2.09323749e-01
9.68611777e-01 2.18534306e-01 -3.98949176e-01 -1.50448591e-01
5.09381115e-01 -3.78549576e-01 -2.32329145e-01 -3.19584869e-02
-1.12459588e+00 1.06883299e+00 1.98623911e-01 8.80347967e-01
-7.61526078e-02 2.25075275e-01 6.21730328e-01 9.64884460e-02
4.24148530e-01 9.42462683e-01 -4.16874766e-01 -1.97910562e-01
-6.06675267e-01 3.82199317e-01 6.87366784e-01 7.84388721e-01
5.18948019e-01 1.28527001e-01 3.06762606e-01 5.85618436e-01
3.44758749e-01 3.36243451e-01 8.11204731e-01 -6.63247705e-01
-5.27434163e-02 4.65519428e-01 1.56717017e-01 -9.75866973e-01
-2.08181515e-01 -2.02769265e-01 8.39953423e-02 -8.38956758e-02
1.78748034e-02 2.54424047e-02 -6.45808697e-01 2.22627974e+00
-3.75166312e-02 -2.19820470e-01 7.54658729e-02 6.57229900e-01
7.40805328e-01 2.62786061e-01 5.61152220e-01 -2.42284581e-01
1.45000803e+00 -2.06534743e-01 -6.60363317e-01 -4.74317878e-01
8.47233355e-01 -6.93075418e-01 1.46854043e+00 4.05693501e-02
-9.07258093e-01 -5.02398312e-01 -1.34884191e+00 -2.88664550e-02
-8.94279540e-01 -6.23588741e-01 9.70994174e-01 8.40759456e-01
-9.84571218e-01 5.15164375e-01 -5.72899103e-01 -6.98125124e-01
-2.45965078e-01 -1.13519013e-01 -4.36020553e-01 2.52202988e-01
-1.45606017e+00 1.59773326e+00 6.86259747e-01 -7.44301230e-02
-4.61351633e-01 -2.93921083e-01 -1.18231571e+00 -2.86475886e-02
1.77304000e-01 -7.51368046e-01 1.12100506e+00 -1.25022280e+00
-6.80612147e-01 9.76602316e-01 -2.85884768e-01 -2.76001245e-01
-3.62082809e-01 -1.03836723e-01 -8.81991684e-01 -1.68543920e-01
3.74479949e-01 1.45965010e-01 5.43677211e-01 -1.34609973e+00
-3.64345491e-01 -4.49870795e-01 3.20458800e-01 2.59687543e-01
-1.29610866e-01 2.09389120e-01 2.16340438e-01 -7.13349402e-01
1.59079820e-01 -5.56425750e-01 3.36330337e-03 -1.97974101e-01
1.00962952e-01 -4.77547467e-01 1.74997643e-01 -3.10999781e-01
1.43630159e+00 -2.19534993e+00 -1.93758875e-01 1.27890110e-01
1.84714228e-01 3.40636015e-01 8.52438509e-02 8.30982983e-01
-3.75710309e-01 7.33935475e-01 -2.86765397e-01 -1.09037422e-01
6.28972352e-01 5.65938652e-01 -5.31713128e-01 4.00326818e-01
3.35149020e-01 9.51750696e-01 -1.07602096e+00 -2.58323044e-01
2.95229048e-01 5.39440989e-01 -3.33630860e-01 -1.43812343e-01
-2.34090358e-01 -5.97727537e-01 -2.20056936e-01 1.07821532e-01
1.90655008e-01 -4.45117950e-02 6.38450980e-01 -1.96623594e-01
-2.07595751e-01 1.18870473e+00 -1.05878770e+00 1.43204510e+00
-4.91870195e-01 6.60466433e-01 -3.90781134e-01 -9.26097870e-01
6.78357601e-01 3.16186368e-01 -9.44604799e-02 -6.22393489e-01
7.79539943e-02 4.72428381e-01 2.30587795e-01 -4.88883704e-01
7.30098307e-01 -8.39803994e-01 -2.55787849e-01 5.00641227e-01
1.54085636e-01 -3.48132700e-01 2.72861540e-01 1.68310225e-01
8.36088955e-01 1.54014930e-01 4.62331265e-01 -6.89953864e-01
1.72312722e-01 7.41279647e-02 5.57284176e-01 4.67244208e-01
3.48899141e-02 3.51970315e-01 3.45919162e-01 -3.03372383e-01
-1.11091387e+00 -1.39848220e+00 -7.20954537e-01 7.06666291e-01
1.88993394e-01 -1.06933308e+00 -5.98368526e-01 -5.31548560e-01
-1.20924547e-01 1.42311907e+00 -7.84760833e-01 -1.56901091e-01
-2.34607100e-01 -6.99548960e-01 4.50032175e-01 6.26286209e-01
-3.94291967e-01 -9.95475411e-01 -9.33589220e-01 3.09970707e-01
2.29311585e-01 -8.60133827e-01 -2.26976946e-01 3.01276892e-01
-5.99258244e-01 -1.05841315e+00 -4.70097438e-02 -6.35060847e-01
5.46620011e-01 1.54473959e-02 1.44680643e+00 2.79263437e-01
3.59388068e-02 6.13772571e-01 -5.12725353e-01 -4.57418561e-01
-4.86814350e-01 -2.71919042e-01 4.04702723e-02 -5.37703276e-01
1.17455816e+00 -5.76294363e-01 -3.41230184e-01 -2.35589966e-01
-1.47380841e+00 -5.05891919e-01 3.30515504e-01 6.82275116e-01
1.03592977e-01 -1.06484465e-01 6.93505287e-01 -1.08775032e+00
1.02957559e+00 -3.88776839e-01 -4.95774113e-02 2.45858163e-01
-9.45644021e-01 2.33114749e-01 1.90660790e-01 -1.56260237e-01
-6.69031501e-01 -6.69652104e-01 -3.49192500e-01 3.80187988e-01
-2.85058320e-01 7.24288166e-01 -1.25234753e-01 3.87735218e-01
7.00582623e-01 2.23346785e-01 -5.51720671e-02 -2.75938451e-01
6.41695201e-01 3.96846920e-01 1.41031697e-01 -8.33850503e-01
6.66576207e-01 1.02994375e-01 -2.12679535e-01 -1.17087817e+00
-8.75957906e-01 -3.89351398e-01 -5.68000555e-01 2.21386239e-01
1.09313059e+00 -7.58852839e-01 -1.37933642e-01 -2.79000342e-01
-1.22518528e+00 5.93019798e-02 -5.74945748e-01 5.61901331e-01
-4.22087133e-01 5.21628141e-01 -3.21294487e-01 -6.17863774e-01
-6.25159517e-02 -9.23119664e-01 7.48052239e-01 -2.75200099e-01
-9.69440639e-01 -1.54461622e+00 7.15102628e-02 -1.89882368e-01
5.46521604e-01 1.92406952e-01 1.37997270e+00 -1.05163932e+00
5.94293550e-02 -1.76801980e-01 -2.53998190e-02 5.69581211e-01
3.67312700e-01 7.30463937e-02 -8.60335171e-01 -4.44906345e-03
3.69280934e-01 -2.29752436e-01 6.78378880e-01 -1.06596246e-01
4.96098161e-01 -3.38667363e-01 -1.53040634e-02 1.24606013e-01
1.89364946e+00 1.11246608e-01 5.73527575e-01 2.18134165e-01
2.51652122e-01 7.56060600e-01 3.46201181e-01 8.29288587e-02
3.78918111e-01 5.44256210e-01 4.69949888e-03 1.79125056e-01
-7.34916627e-02 -5.27024329e-01 2.08848551e-01 9.39914584e-01
4.56589043e-01 -1.61844134e-01 -1.05071044e+00 9.65990543e-01
-1.45485401e+00 -9.84029591e-01 -7.43353441e-02 2.41833568e+00
8.60755384e-01 4.54413503e-01 -3.26341540e-02 1.24403104e-01
4.69906688e-01 4.73410904e-01 2.22469866e-01 -7.98117220e-01
-5.09105660e-02 5.76808453e-01 1.47712305e-01 7.91493773e-01
-5.93402684e-01 9.78203714e-01 7.27628994e+00 4.40855861e-01
-8.38985860e-01 1.01564482e-01 -4.82961871e-02 3.97062987e-01
-1.05288899e+00 3.92372578e-01 -5.17085433e-01 5.91555417e-01
1.03275049e+00 -3.82502139e-01 -6.32578582e-02 5.59046209e-01
-5.04642688e-02 -1.25731111e-01 -1.35849142e+00 4.94752198e-01
4.01207626e-01 -1.01077259e+00 6.13692224e-01 -6.82637095e-02
9.73693281e-02 -2.70073444e-01 -1.00440569e-01 1.61683515e-01
2.22085819e-01 -1.25729883e+00 8.77593994e-01 1.75366238e-01
6.23588741e-01 -7.12494016e-01 8.92054617e-01 8.38487893e-02
-8.99842143e-01 3.75843585e-01 -4.96276498e-01 -4.00038570e-01
4.05576259e-01 3.89071316e-01 -6.38187289e-01 2.44905382e-01
-8.48228037e-02 3.33731353e-01 -7.26698875e-01 5.62279642e-01
-5.78662157e-01 5.78506589e-01 -2.32141793e-01 -3.31629097e-01
3.91468525e-01 -7.14559928e-02 4.56427515e-01 1.51438498e+00
7.40229851e-03 2.66560502e-02 -6.74901009e-02 9.43859100e-01
4.59750950e-01 3.12162012e-01 -9.96438444e-01 -4.52714145e-01
6.24068856e-01 7.22700775e-01 -6.87152505e-01 -2.23948836e-01
-7.23225713e-01 6.84547067e-01 3.49234462e-01 2.46982887e-01
-6.34212911e-01 -3.10057044e-01 8.98129642e-01 3.47375721e-01
1.72692776e-01 -5.68908334e-01 -3.64328623e-01 -1.13405383e+00
1.14075758e-01 -4.80314434e-01 9.88853797e-02 -9.18239295e-01
-1.34051287e+00 5.50120592e-01 6.52370304e-02 -5.42495131e-01
-2.85134792e-01 -1.02598941e+00 -5.22801578e-01 1.09907126e+00
-1.24334323e+00 -8.84054899e-01 2.99321979e-01 1.16180763e-01
3.70180070e-01 2.35756099e-01 1.20897853e+00 -2.40637213e-02
-1.04640849e-01 3.06546867e-01 -3.10691506e-01 4.57336791e-02
5.59300482e-01 -1.48568773e+00 5.00139713e-01 8.85085940e-01
5.76637030e-01 1.49356484e+00 1.11796963e+00 -4.97188330e-01
-1.04459059e+00 -5.61568916e-01 1.62706339e+00 -9.23360705e-01
8.77641380e-01 -2.15850875e-01 -9.02715921e-01 7.83201933e-01
3.75123858e-01 -3.90191466e-01 1.04438233e+00 6.38091087e-01
-6.89930260e-01 2.14343637e-01 -8.21467996e-01 6.92948401e-01
9.57458675e-01 -9.75683451e-01 -1.71416211e+00 9.44333747e-02
1.06214964e+00 2.49767229e-01 -6.72991335e-01 3.49787474e-01
3.57779741e-01 -8.60252559e-01 7.85063446e-01 -8.45101595e-01
3.52758110e-01 -2.53558099e-01 -6.28674269e-01 -1.21548712e+00
-4.33163315e-01 -8.39161277e-02 5.27784944e-01 1.24675798e+00
8.18070173e-01 -7.76349783e-01 2.33443856e-01 7.02478111e-01
-2.44407430e-01 -4.63620603e-01 -9.64637697e-01 -8.05881321e-01
5.09163678e-01 -8.88474822e-01 6.71541333e-01 1.15306163e+00
3.25686425e-01 9.51000035e-01 4.34880704e-01 -1.36587188e-01
3.64074975e-01 -2.16262966e-01 2.49655291e-01 -1.14388740e+00
-2.12675646e-01 -6.46646082e-01 -6.81650281e-01 -5.96897840e-01
4.40682650e-01 -1.05625081e+00 -1.53535500e-01 -1.70666993e+00
-2.88591031e-02 -3.70683283e-01 -3.84629667e-01 5.28149717e-02
-2.18115330e-01 -1.86987314e-02 1.91725671e-01 -8.35847631e-02
-3.63227844e-01 4.56647009e-01 6.15256727e-01 3.59446794e-01
3.44992042e-01 -4.71386790e-01 -1.16031599e+00 8.16140831e-01
5.89922905e-01 -4.10780162e-01 -7.07629800e-01 -6.40094876e-01
7.15315700e-01 -5.14032245e-01 4.50863332e-01 -5.63220203e-01
-2.58336961e-01 -2.62022108e-01 5.96123450e-02 5.06364964e-02
2.49115452e-01 -1.01661801e+00 2.76628166e-01 2.93906599e-01
-4.76921022e-01 5.72000146e-01 1.59720093e-01 3.02587390e-01
-2.02735126e-01 -7.79030383e-01 4.40608889e-01 -3.85044903e-01
-1.06221855e+00 -2.16022521e-01 -4.19253349e-01 4.38779593e-01
7.45418787e-01 -4.30380344e-01 -2.52784580e-01 -4.28461935e-03
-6.25900269e-01 -3.32066983e-01 9.50645685e-01 4.96130824e-01
5.59710681e-01 -1.31475496e+00 -4.98971224e-01 1.27840742e-01
5.28988123e-01 -5.72655976e-01 -3.59052658e-01 1.83001518e-01
-3.75211060e-01 5.18692553e-01 1.68180496e-01 -9.08485353e-02
-8.21658134e-01 1.01396322e+00 2.50957727e-01 3.47017050e-01
-6.54119015e-01 6.16342723e-01 3.25196356e-01 -1.89867750e-01
-8.14783722e-02 -4.17250872e-01 3.19283493e-02 1.44305348e-01
5.42092621e-01 -1.58424810e-01 -6.53721541e-02 -8.37204695e-01
-3.93243998e-01 3.59667689e-01 1.04959421e-01 -4.47180361e-01
1.10521483e+00 -2.05437496e-01 -1.59842715e-01 1.10930085e+00
1.17538714e+00 9.21652615e-02 -2.91090727e-01 -1.22890472e-01
4.24519360e-01 -5.57412446e-01 -5.59061356e-02 -8.07624400e-01
-1.79482281e-01 8.03509116e-01 1.35919690e-01 3.55463237e-01
6.14916682e-01 7.30882138e-02 6.30061686e-01 1.14179827e-01
5.66411018e-01 -1.09318495e+00 -2.00895682e-01 5.12540340e-01
9.11870956e-01 -8.98997247e-01 3.80761698e-02 -2.65036851e-01
-5.36883771e-01 1.01002681e+00 3.69876862e-01 -2.16660053e-01
6.94239557e-01 5.47062904e-02 -3.79374027e-02 -5.64439476e-01
-9.25154805e-01 -6.00357056e-01 1.23846732e-01 7.53870428e-01
8.88324320e-01 1.65010855e-01 -1.07341850e+00 6.03369713e-01
-4.28326458e-01 -4.45907593e-01 5.31764865e-01 8.73438478e-01
-6.01424336e-01 -1.14954281e+00 8.37160945e-02 3.22792768e-01
-5.89328825e-01 -5.46039224e-01 -5.98953128e-01 1.09174573e+00
2.96006978e-01 7.42136478e-01 1.18056178e-01 -2.48685151e-01
2.97808558e-01 4.85233188e-01 5.06865203e-01 -1.29020143e+00
-5.99904835e-01 -2.90715903e-01 5.08634627e-01 -3.17930460e-01
-4.25370455e-01 -4.46782351e-01 -1.09140730e+00 -2.08728760e-01
-3.55083138e-01 5.50146341e-01 6.48688197e-01 1.17034829e+00
2.13774759e-02 1.66838199e-01 7.32294098e-02 -8.71258304e-02
-6.99204028e-01 -9.16891217e-01 -7.27129579e-01 7.45200455e-01
1.58208251e-01 -7.40536451e-01 -7.80484498e-01 -2.23186269e-01] | [10.327476501464844, 8.957219123840332] |
6169d4f3-83c1-499b-8fd8-8cb68ce031d8 | m-cner-a-corpus-for-chinese-named-entity | null | null | https://aclanthology.org/L18-1706 | https://aclanthology.org/L18-1706.pdf | M-CNER: A Corpus for Chinese Named Entity Recognition in Multi-Domains | null | ['Qi Lu', 'Zhenghua Li', 'Wenliang Chen', 'Min Zhang', 'YaoSheng Yang'] | 2018-05-01 | m-cner-a-corpus-for-chinese-named-entity-1 | https://aclanthology.org/L18-1706 | https://aclanthology.org/L18-1706.pdf | lrec-2018-5 | ['chinese-named-entity-recognition'] | ['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] | [-7.2136759757995605, 3.7917046546936035] |
59858a93-2693-416c-831c-dd919a244027 | contextual-dependencies-in-time-continuous | null | null | https://aclanthology.org/L18-1196 | https://aclanthology.org/L18-1196.pdf | Contextual Dependencies in Time-Continuous Multidimensional Affect Recognition | null | ['Maxim Sidorov', 'Wolfgang Minker', 'Dmitrii Fedotov', 'Denis Ivanko'] | 2018-05-01 | contextual-dependencies-in-time-continuous-1 | https://aclanthology.org/L18-1196 | https://aclanthology.org/L18-1196.pdf | lrec-2018-5 | ['multimodal-emotion-recognition', 'multimodal-emotion-recognition'] | ['computer-vision', 'speech'] | [-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] | [-7.2173261642456055, 3.5678350925445557] |
fa7d3208-04dc-462e-9918-9b1e737cbe58 | quantum-machine-learning-for-material | 2208.08273 | null | https://arxiv.org/abs/2208.08273v1 | https://arxiv.org/pdf/2208.08273v1.pdf | Quantum Machine Learning for Material Synthesis and Hardware Security | Using quantum computing, this paper addresses two scientifically pressing and day-to-day relevant problems, namely, chemical retrosynthesis which is an important step in drug/material discovery and security of the semiconductor supply chain. We show that Quantum Long Short-Term Memory (QLSTM) is a viable tool for retrosynthesis. We achieve 65% training accuracy with QLSTM, whereas classical LSTM can achieve 100%. However, in testing, we achieve 80% accuracy with the QLSTM while classical LSTM peaks at only 70% accuracy! We also demonstrate an application of Quantum Neural Network (QNN) in the hardware security domain, specifically in Hardware Trojan (HT) detection using a set of power and area Trojan features. The QNN model achieves detection accuracy as high as 97.27%. | ['Swaroop Ghosh', 'Rasit Onur Topaloglu', 'Satwik Kundu', 'Collin Beaudoin'] | 2022-08-16 | null | null | null | null | ['retrosynthesis'] | ['medical'] | [ 6.45711839e-01 -2.71683007e-01 -5.56416273e-01 4.10021514e-01
-1.05667841e+00 -6.05267942e-01 3.93207788e-01 4.38547939e-01
-5.99495471e-01 8.19410920e-01 -7.43141532e-01 -1.15280354e+00
3.53628516e-01 -8.78666937e-01 -8.91293406e-01 -9.36559558e-01
1.37833536e-01 -1.83332101e-01 5.18907234e-02 -5.19653440e-01
7.28415310e-01 4.92124557e-01 -7.98499346e-01 4.46124941e-01
4.77844715e-01 1.11654401e+00 -2.12696448e-01 1.01595545e+00
3.80546123e-01 8.90231431e-01 -9.70460951e-01 -5.53690970e-01
-1.10629991e-01 -2.01956779e-01 -9.00072694e-01 -1.02622688e+00
1.81844682e-01 -1.73313096e-01 -3.32133442e-01 1.55548143e+00
5.90163708e-01 -2.50109315e-01 4.25259113e-01 -7.66462088e-01
-4.95267808e-01 7.05565095e-01 -2.61125833e-01 4.98752922e-01
3.66458654e-01 3.95631373e-01 1.01188636e+00 -5.31373858e-01
2.84556478e-01 7.30954051e-01 7.32715309e-01 6.23870730e-01
-1.12549448e+00 -1.12271023e+00 -9.08982933e-01 3.38663369e-01
-1.34639382e+00 -2.95807570e-01 2.69788802e-01 -1.07675463e-01
1.78741062e+00 -1.63984761e-01 2.35488191e-01 1.22820556e+00
1.44718742e+00 2.10758492e-01 9.58121538e-01 -4.37782347e-01
4.30982441e-01 -2.50073552e-01 2.21508637e-01 8.44779313e-01
1.64001822e-01 7.50168443e-01 -6.28106833e-01 -2.74542212e-01
6.38613284e-01 -8.24058428e-02 3.71687263e-01 6.43001258e-01
-1.06970477e+00 9.03694808e-01 6.60248697e-01 7.52049863e-01
-1.29256874e-01 1.06565595e+00 6.72943830e-01 4.42102373e-01
7.58640096e-02 8.86073947e-01 -4.41963285e-01 -3.16072464e-01
-8.07790816e-01 -1.74297065e-01 4.55715597e-01 6.68225944e-01
5.21998823e-01 5.39771080e-01 3.72279808e-02 -2.15855569e-01
-6.15205355e-02 5.13977766e-01 5.30050397e-01 -5.83431602e-01
3.85567434e-02 8.51917546e-04 -3.11361223e-01 -3.06390822e-01
-5.20329237e-01 -4.49444652e-01 -8.96619678e-01 4.60216440e-02
2.00179145e-01 -6.74445778e-02 -9.65455413e-01 1.53965330e+00
-1.37522072e-01 2.10587204e-01 2.35504389e-01 1.31665885e-01
6.38605714e-01 9.31949317e-01 3.69720191e-01 -2.33676732e-01
1.54876339e+00 -4.16410029e-01 -7.48598695e-01 1.79887682e-01
1.06663513e+00 -7.82155037e-01 3.83734554e-01 7.31810868e-01
-8.59856963e-01 -3.54396343e-01 -1.73457825e+00 -1.30538031e-01
-7.86529183e-01 -1.57529861e-01 7.83945620e-01 1.40733433e+00
-8.39144528e-01 1.33553672e+00 -8.54292035e-01 -1.15319462e-02
1.98067695e-01 1.07131338e+00 -1.15110673e-01 3.38243395e-01
-1.57400703e+00 9.12122667e-01 5.15297055e-01 -1.29718915e-01
-1.29658771e+00 -6.46410525e-01 -4.87634599e-01 4.65062773e-03
3.18465203e-01 -5.73310196e-01 1.65887678e+00 -3.27432215e-01
-1.73319113e+00 4.21999246e-01 1.07595630e-01 -1.01875734e+00
-3.37760746e-01 2.40347281e-01 -6.11970365e-01 4.63791713e-02
-1.98746517e-01 4.42449749e-01 7.85926700e-01 -2.64759988e-01
-4.76383895e-01 -3.57281655e-01 -9.66483876e-02 -6.77908480e-01
-3.05863738e-01 1.08821809e-01 5.13705075e-01 -1.54752538e-01
-3.57535899e-01 -1.11959565e+00 -1.89418495e-01 -9.02953446e-01
-6.49677217e-01 -2.83404320e-01 4.74371105e-01 -5.08095562e-01
1.17689323e+00 -1.74355507e+00 -3.75262082e-01 2.18310282e-01
2.28246942e-01 4.94197100e-01 1.94583554e-02 4.27980930e-01
-4.08307612e-01 3.10965896e-01 1.10581465e-01 3.95251811e-01
-1.49280280e-01 -2.70214587e-01 -3.70210141e-01 5.00671625e-01
2.68436998e-01 1.36142671e+00 -9.64662135e-01 -3.01272087e-02
9.23916921e-02 2.98795015e-01 -2.04446927e-01 -6.24486625e-01
-3.61444533e-01 3.37820798e-01 -3.43966424e-01 1.02155435e+00
2.05282286e-01 -5.12564242e-01 2.00960293e-01 -2.33147204e-01
-1.23735182e-01 5.19579053e-01 -3.11415225e-01 1.79154241e+00
-3.06033790e-01 5.89554191e-01 -6.58902347e-01 -4.90946472e-01
4.87519413e-01 5.00298083e-01 2.85909861e-01 -1.30757678e+00
5.80755174e-01 6.24858499e-01 5.24528146e-01 1.72765598e-01
7.76091993e-01 -6.16882384e-01 -4.32303071e-01 3.09931278e-01
3.07483554e-01 -9.07053128e-02 7.22618923e-02 1.38013080e-01
1.47373605e+00 -1.63776100e-01 2.75667548e-01 -3.60835940e-01
4.12487507e-01 -9.06922203e-03 2.39277914e-01 6.80454910e-01
-4.97568071e-01 -1.04681611e-01 4.18926179e-01 -5.11231005e-01
-1.34774983e+00 -8.79362345e-01 -1.99183673e-01 1.05050468e+00
-2.51299739e-01 -4.19689596e-01 -9.83078718e-01 -3.31534594e-01
-4.55982298e-01 6.42889321e-01 -3.59123707e-01 -6.25036120e-01
-5.19969225e-01 -1.02891266e+00 1.16843438e+00 3.77354652e-01
3.43866616e-01 -9.55919027e-01 -4.14478719e-01 5.77914059e-01
4.34115052e-01 -1.30981624e+00 -2.50629902e-01 8.78940880e-01
-8.39652359e-01 -9.73667026e-01 -2.50384599e-01 -5.47261715e-01
1.20642573e-01 -2.15594217e-01 8.04558754e-01 -5.39841652e-02
-7.03633130e-01 -4.01018143e-01 1.79765731e-01 -3.73743087e-01
-1.13163424e+00 2.65246361e-01 2.61253446e-01 -7.65041173e-01
4.26262558e-01 -3.73819947e-01 -4.83449996e-01 -2.29514241e-01
-7.59197891e-01 -4.30683821e-01 7.66652226e-01 9.18731153e-01
4.26199138e-01 2.60763198e-01 6.07712328e-01 -7.41157413e-01
3.67879808e-01 -7.71910250e-02 -8.75820160e-01 2.56959975e-01
-8.78107607e-01 4.62774962e-01 1.02793133e+00 -3.12703818e-01
-2.58296907e-01 2.15307381e-02 -5.76129436e-01 -1.68560535e-01
3.30998778e-01 4.63769346e-01 2.52478302e-01 -7.92552233e-01
9.38616693e-01 2.06617434e-02 -2.84606457e-01 1.97536558e-01
7.97891170e-02 2.23874539e-01 3.72495353e-01 -3.86943698e-01
6.41107500e-01 8.74927193e-02 8.69105279e-01 -1.03751922e+00
-6.25553191e-01 -6.07274026e-02 -3.26251477e-01 3.38817418e-01
8.42719853e-01 -6.70756638e-01 -1.70898473e+00 3.71899664e-01
-1.32207525e+00 -2.06685930e-01 4.18434793e-04 3.39640558e-01
-3.10573250e-01 3.52701873e-01 -1.40152717e+00 -9.33154106e-01
-9.69447732e-01 -1.48528528e+00 1.18289745e+00 9.82595980e-02
-1.52674019e-01 -9.79575932e-01 -1.25053138e-01 5.30808158e-02
1.74241483e-01 3.95025313e-01 1.35210371e+00 -6.18830264e-01
-6.32233322e-01 -4.27026898e-01 -7.27501884e-02 1.86975181e-01
-1.10970244e-01 -6.21208437e-02 -1.37219346e+00 -4.29015547e-01
2.12079845e-02 -4.77526218e-01 9.19781208e-01 4.11401570e-01
9.60845590e-01 -6.88685430e-03 -2.27848560e-01 2.38018081e-01
1.50183928e+00 6.67569518e-01 8.46562147e-01 7.10725188e-02
6.54944181e-01 -1.23411618e-01 2.85343230e-01 8.18670690e-02
-2.40734398e-01 4.24194068e-01 4.81556684e-01 2.60480940e-01
2.11384803e-01 -3.60499978e-01 8.55359316e-01 7.52286017e-01
4.77670766e-02 2.46645883e-02 -1.06834519e+00 -2.80160636e-01
-1.17240632e+00 -9.68490779e-01 -2.58335292e-01 2.38401389e+00
9.27749634e-01 6.65135324e-01 8.73646736e-02 5.05202770e-01
6.46402240e-01 -2.99177766e-01 -6.88497961e-01 -1.01725888e+00
9.27033797e-02 1.22982466e+00 1.35873425e+00 3.35240632e-01
-1.14412796e+00 1.23131585e+00 7.01444387e+00 1.74369669e+00
-1.38444412e+00 3.28256011e-01 9.27349746e-01 3.92628372e-01
1.03937322e-02 7.99124315e-02 -9.32917237e-01 2.08964884e-01
2.24722934e+00 -8.15259963e-02 3.65339041e-01 4.92821962e-01
-2.06300080e-01 -1.02529176e-01 -1.27005935e+00 9.75955665e-01
-3.80472362e-01 -1.83957541e+00 -5.46515919e-02 5.08127749e-01
7.13920116e-01 6.24121465e-02 6.60557747e-01 3.48783433e-01
-3.23679298e-02 -1.50091231e+00 4.33662683e-01 -1.38383925e-01
1.29441500e+00 -1.28232646e+00 7.45971203e-01 1.81178093e-01
-1.09943271e+00 -2.42918268e-01 -3.92362952e-01 -9.30401683e-02
-3.77258480e-01 4.18443173e-01 -1.09455240e+00 5.43242276e-01
2.81880975e-01 3.18810582e-01 -4.50123757e-01 5.43469429e-01
-8.17123652e-02 4.07121241e-01 -1.47507653e-01 -6.62766635e-01
5.62437952e-01 -3.44551331e-03 -2.19534663e-03 9.04074907e-01
3.77615720e-01 -2.05056146e-01 -2.14598984e-01 8.39455366e-01
-4.47822481e-01 -3.91685069e-01 -6.38873994e-01 -7.97848046e-01
2.66981125e-01 1.07165670e+00 -8.33646715e-01 -2.04283550e-01
1.14823155e-01 8.89992058e-01 -3.76898438e-01 -1.22506104e-01
-1.00302851e+00 -6.21645808e-01 2.05378458e-01 -3.88031155e-01
-6.21082149e-02 -2.72146404e-01 -3.92421633e-01 -6.43920600e-01
-6.48067117e-01 -8.38094592e-01 9.49328691e-02 -2.11504489e-01
-6.47455752e-01 3.61160696e-01 -7.20139503e-01 -9.69404280e-01
-3.25732082e-01 -1.21488714e+00 -3.71527523e-01 7.27941632e-01
-1.42691946e+00 -1.08525121e+00 6.65918648e-01 1.37341782e-01
1.60518706e-01 -1.34731367e-01 1.15551090e+00 4.35729623e-01
-7.73973405e-01 8.22901845e-01 4.52722043e-01 9.36511997e-03
3.03492814e-01 -1.19936943e+00 9.37173426e-01 5.68082273e-01
2.50522971e-01 9.47234154e-01 8.71745586e-01 -6.11430645e-01
-2.17827320e+00 -9.95171905e-01 7.06819773e-01 -5.32462239e-01
1.18296421e+00 -4.34559792e-01 -3.19070488e-01 3.12903970e-01
7.64648318e-02 -2.13904887e-01 9.95967031e-01 -1.99105427e-01
-4.38843250e-01 1.83703184e-01 -1.19527912e+00 5.05591571e-01
3.96108955e-01 -1.25023973e+00 5.09051643e-02 7.83246100e-01
8.73494565e-01 -2.73901790e-01 -1.10212409e+00 3.53515208e-01
7.50141025e-01 -6.44066513e-01 1.00436604e+00 -4.48864996e-01
4.25627857e-01 -4.05705050e-02 -7.99532160e-02 -7.36541867e-01
-2.13489421e-02 -1.11386764e+00 -9.49485227e-02 3.75522137e-01
7.32641637e-01 -5.92169046e-01 1.16717541e+00 1.67273387e-01
-1.61092892e-01 -5.39236605e-01 -1.45186174e+00 -1.09783888e+00
6.31478012e-01 -6.35921776e-01 2.12584093e-01 5.84461093e-01
3.05824935e-01 7.10946918e-01 -2.27863491e-01 3.55951600e-02
5.10669172e-01 -2.42753237e-01 3.53749394e-02 -6.69966936e-01
-3.24085832e-01 -4.48270470e-01 -5.58654964e-01 -5.94596505e-01
2.23829716e-01 -9.79234159e-01 -2.00660259e-01 -5.91301143e-01
3.85742366e-01 1.16415091e-01 -7.54564285e-01 2.34045267e-01
3.00237000e-01 6.13411188e-01 -1.42512634e-01 -1.80833772e-01
-5.24883807e-01 2.33107314e-01 9.38481212e-01 -5.68648279e-01
9.90656391e-02 -2.39804819e-01 -3.52417052e-01 3.10618192e-01
6.97966814e-01 -5.71366072e-01 1.32111639e-01 3.01212907e-01
6.20137572e-01 3.86421770e-01 2.20243007e-01 -1.21196043e+00
4.04955804e-01 1.33705318e-01 3.89169604e-01 -5.49116433e-01
4.68294054e-01 -3.56296182e-01 -4.99856882e-02 1.53711820e+00
-5.65751009e-02 1.43970400e-01 4.36387330e-01 3.95434320e-01
1.33181676e-01 -3.78128409e-01 8.01284611e-01 -5.38122989e-02
-4.07790631e-01 3.99357587e-01 -5.65356016e-01 -5.48914552e-01
8.70930016e-01 -2.23279282e-01 -8.15743804e-01 4.75347638e-02
-3.62298697e-01 -1.63935736e-01 2.05728024e-01 -4.97086681e-02
7.40354538e-01 -1.06781340e+00 -1.36132330e-01 5.47612719e-02
3.27630118e-02 -8.32621872e-01 2.40708888e-01 7.45638490e-01
-7.46918857e-01 1.24175787e+00 -8.26842114e-02 -4.68098521e-01
-1.17047119e+00 8.66677046e-01 4.26550508e-01 -3.93665791e-01
7.15003163e-02 8.78230751e-01 -1.15395010e-01 -3.33984159e-02
-1.34102419e-01 -5.02207994e-01 1.66024491e-01 -2.55994707e-01
4.56464350e-01 3.78536165e-01 6.77406132e-01 -4.32981163e-01
-4.94064838e-01 4.01928037e-01 -1.31423041e-01 -9.99437571e-02
8.47290218e-01 7.59855568e-01 -1.79625511e-01 5.56549788e-01
1.47914696e+00 -4.69479173e-01 -3.05934906e-01 1.14297703e-01
5.55846810e-01 3.73568058e-01 4.16098714e-01 -6.66531503e-01
-3.61419350e-01 1.33077133e+00 1.02028680e+00 2.21721232e-01
6.75274730e-01 -5.56423187e-01 1.51514804e+00 9.71391261e-01
7.10540593e-01 -1.17939866e+00 2.95332611e-01 7.92632878e-01
-8.93267840e-02 -9.61228609e-01 1.85567766e-01 4.78925221e-02
1.27375320e-01 1.61800122e+00 -2.81770490e-02 -1.78099066e-01
3.67182791e-01 1.98708743e-01 -6.27586365e-01 -2.72009343e-01
-8.85937214e-01 3.75738344e-03 -9.99782886e-03 1.30448788e-01
7.03585684e-01 2.75560319e-01 -7.43311793e-02 1.17778219e-01
-3.44236970e-01 -5.23812473e-02 7.42129505e-01 1.13137114e+00
-4.91199166e-01 -1.46664989e+00 -3.05322796e-01 3.23524475e-01
-1.16142857e+00 -6.01889372e-01 -4.16388899e-01 3.94726634e-01
1.31010935e-02 1.05102956e+00 -3.04199487e-01 -9.02467191e-01
-2.51172632e-01 1.25092953e-01 9.37087536e-01 -5.96454799e-01
-1.12431824e+00 1.08346939e-01 5.33005781e-02 -4.65913117e-01
2.81334221e-02 -1.19112290e-01 -1.66821218e+00 -8.18590701e-01
-8.83934379e-01 1.43234223e-01 1.20128560e+00 1.07655346e+00
3.73462960e-02 8.96340907e-01 6.72839522e-01 -5.17992556e-01
-7.76435018e-01 -7.00078368e-01 -6.05745673e-01 -4.71016109e-01
5.51072776e-01 -1.21150017e-01 -4.55668271e-02 -3.79686832e-01] | [5.528236389160156, 5.107929229736328] |
4b4b3f9d-5051-4401-905c-ecb190eb0962 | a-multiscale-graph-convolutional-network | 2006.12542 | null | https://arxiv.org/abs/2006.12542v1 | https://arxiv.org/pdf/2006.12542v1.pdf | A Multiscale Graph Convolutional Network Using Hierarchical Clustering | The information contained in hierarchical topology, intrinsic to many networks, is currently underutilised. A novel architecture is explored which exploits this information through a multiscale decomposition. A dendrogram is produced by a Girvan-Newman hierarchical clustering algorithm. It is segmented and fed through graph convolutional layers, allowing the architecture to learn multiple scale latent space representations of the network, from fine to coarse grained. The architecture is tested on a benchmark citation network, demonstrating competitive performance. Given the abundance of hierarchical networks, possible applications include quantum molecular property prediction, protein interface prediction and multiscale computational substrates for partial differential equations. | ['Pietro Liò', 'Alex Lipov'] | 2020-06-22 | null | null | null | null | ['protein-interface-prediction'] | ['miscellaneous'] | [ 1.29920870e-01 5.05016148e-01 -1.26366958e-01 -1.07403837e-01
1.08962491e-01 -5.94526947e-01 6.22784793e-01 3.39158773e-01
4.80856113e-02 6.80941284e-01 1.72583759e-01 -3.63184601e-01
-5.73739052e-01 -9.43715930e-01 -2.17590287e-01 -8.10949862e-01
-8.69744301e-01 7.81788707e-01 2.56411910e-01 -1.90788180e-01
3.80541921e-01 8.46202254e-01 -1.10060298e+00 4.92563993e-01
7.67863035e-01 7.86113620e-01 -3.32559347e-02 8.64242971e-01
-4.50228080e-02 9.44947004e-01 -2.29140982e-01 2.90632043e-02
1.30771071e-01 -2.62831122e-01 -9.94892955e-01 9.23677459e-02
2.57826775e-01 2.07311809e-01 -4.81421769e-01 1.04160285e+00
-1.80282135e-04 1.41635258e-02 1.07226634e+00 -1.08933926e+00
-6.56186759e-01 6.16999984e-01 -2.10453153e-01 2.99045652e-01
3.00817907e-01 9.62974802e-02 1.55376077e+00 -6.66675925e-01
1.01146340e+00 1.65287423e+00 7.40046501e-01 -6.89365566e-02
-2.14215994e+00 -3.48294765e-01 -2.40346834e-01 1.21539898e-01
-1.40404820e+00 2.97476947e-02 7.89784253e-01 -7.64011025e-01
1.53203630e+00 -1.90403104e-01 7.18675554e-01 1.04591858e+00
9.38174009e-01 1.08077750e-01 1.17735481e+00 -4.98565808e-02
4.54408973e-01 -4.25126463e-01 3.05081397e-01 1.14475226e+00
2.74003655e-01 -1.87754810e-01 -5.43079853e-01 -6.41468465e-01
1.03945029e+00 -4.74656560e-02 1.04328729e-01 -9.49350715e-01
-1.09032536e+00 1.02342033e+00 8.90376687e-01 3.71949077e-01
-5.19067943e-01 1.39621750e-01 2.27767199e-01 5.83025694e-01
3.86479080e-01 6.25847816e-01 -1.58930570e-01 3.44180286e-01
-1.17076075e+00 7.49576315e-02 1.30269456e+00 6.66362107e-01
1.01836491e+00 -7.47331530e-02 4.24528211e-01 3.09654623e-01
6.15331292e-01 -2.51559168e-01 1.63739011e-01 -1.35046351e+00
-7.39279240e-02 1.08310950e+00 -2.99158692e-01 -1.32473981e+00
-1.05637562e+00 -3.00456345e-01 -1.47274029e+00 3.75021935e-01
1.70505077e-01 1.86053604e-01 -8.73960853e-01 1.30627489e+00
1.58865556e-01 -7.99916983e-02 -1.52673125e-01 5.45090020e-01
5.67932904e-01 5.56865215e-01 1.37426332e-01 -1.26457095e-01
1.20303690e+00 -6.55352354e-01 -4.13334936e-01 4.03273225e-01
5.17814755e-01 -3.40234101e-01 2.57802010e-01 3.19167763e-01
-1.07905221e+00 -4.03670281e-01 -1.17905188e+00 -1.38746321e-01
-8.05362523e-01 -3.82587552e-01 9.46925402e-01 1.95584282e-01
-1.77907872e+00 1.31455874e+00 -1.08889723e+00 -7.56760061e-01
4.43259895e-01 9.58803952e-01 -4.73592162e-01 1.64479345e-01
-1.33220077e+00 7.18898356e-01 6.96413934e-01 -2.05056593e-01
-7.36691415e-01 -4.61371273e-01 -6.67971373e-01 4.50658053e-01
-4.44143176e-01 -1.20698988e+00 3.64182770e-01 -5.24571240e-01
-1.56741548e+00 6.03358924e-01 -2.41830088e-02 -7.85962760e-01
1.82353884e-01 9.27747488e-01 -1.86271027e-01 6.80147409e-01
1.39159143e-01 9.20741022e-01 7.89368510e-01 -6.93148255e-01
8.75351653e-02 -3.19793522e-01 -8.29568878e-02 7.01505989e-02
-2.13117704e-01 -1.10285737e-01 1.70836195e-01 -3.75353843e-01
5.80592453e-01 -1.00384021e+00 -7.32779145e-01 -2.07700253e-01
-7.18334496e-01 -5.17155111e-01 7.50159025e-01 -3.43394309e-01
9.84390378e-01 -1.54100883e+00 7.22131848e-01 7.78267920e-01
1.12095737e+00 -5.20908177e-01 -7.00039230e-03 9.39494431e-01
-3.10409665e-01 4.77990329e-01 -3.34893793e-01 1.67184055e-01
-7.54776895e-02 2.35377140e-02 2.53180802e-01 4.39602435e-01
4.20307338e-01 9.59459424e-01 -6.18402421e-01 -4.45623904e-01
1.19509041e-01 7.06308246e-01 -4.61915851e-01 7.03231618e-02
-2.17919141e-01 3.80599648e-01 -4.52579409e-01 4.45916653e-01
5.52573323e-01 -1.17436850e+00 7.27237105e-01 -2.30868384e-01
-5.33810109e-02 2.39541829e-01 -9.22010839e-01 1.59540570e+00
3.10298949e-01 5.81746042e-01 3.74149799e-01 -1.08131850e+00
8.52958739e-01 2.80223906e-01 7.10106552e-01 -2.35831246e-01
-4.55953628e-02 -6.49382249e-02 3.53269726e-01 7.36125633e-02
1.48354426e-01 -3.26304585e-02 7.06150755e-02 5.10628819e-01
4.53603685e-01 1.14400581e-01 3.94102424e-01 8.47616673e-01
1.72695577e+00 -1.03548229e-01 4.45196986e-01 -9.10011590e-01
3.98195565e-01 2.63039947e-01 9.17076245e-02 6.32300854e-01
-2.03633711e-01 1.91890091e-01 1.09538078e+00 -6.29388869e-01
-1.48782957e+00 -1.27329051e+00 -2.34853104e-01 9.83266711e-01
-1.61996633e-01 -8.23349893e-01 -7.99481273e-01 -1.40566513e-01
6.80644661e-02 -3.11221302e-01 -7.94233441e-01 -2.38751583e-02
-2.65429676e-01 -8.34199786e-01 3.39865416e-01 1.66832969e-01
1.31113976e-01 -1.04687989e+00 -2.26009399e-01 5.60896635e-01
4.01679337e-01 -9.92587090e-01 1.20606229e-01 4.80965048e-01
-1.31871128e+00 -1.15490627e+00 -1.64971158e-01 -6.82618856e-01
4.56675947e-01 1.30460737e-02 1.28034234e+00 1.82848245e-01
-6.54681444e-01 1.82814136e-01 1.49785101e-01 3.64631593e-01
-5.89869320e-01 4.74726200e-01 3.61953706e-01 -1.42211065e-01
3.03165138e-01 -1.21735859e+00 -7.15579808e-01 -1.77096017e-02
-6.73448920e-01 1.60304740e-01 8.11818898e-01 7.27162838e-01
3.60676229e-01 2.60621250e-01 2.53924459e-01 -8.11252952e-01
8.98686707e-01 -6.83114529e-01 -4.75072592e-01 -7.10434839e-02
-6.55135334e-01 3.17376018e-01 6.86026335e-01 -5.04459739e-02
-3.87465924e-01 1.34190679e-01 4.18157250e-01 -1.95889935e-01
-3.57525349e-01 5.62280476e-01 3.11274882e-02 -4.86655235e-01
5.40755332e-01 -1.65106356e-02 4.32687104e-01 -4.12097275e-01
5.24599314e-01 2.92280704e-01 7.66492486e-02 -4.42709178e-01
5.51877499e-01 4.45845515e-01 6.54118419e-01 -1.31929636e+00
-2.03585193e-01 -4.68400270e-01 -1.34260929e+00 -1.75997410e-02
1.17752123e+00 -7.39677489e-01 -1.03454757e+00 1.27128452e-01
-1.14084697e+00 -1.04741953e-01 2.30989978e-01 1.10222638e-01
-7.25251257e-01 4.48543966e-01 -1.30379546e+00 -3.39598745e-01
-1.88633755e-01 -9.78287935e-01 6.79359436e-01 -7.89164677e-02
-4.56715614e-01 -1.40877807e+00 2.56312013e-01 7.46969208e-02
4.74072337e-01 6.29314959e-01 1.43280387e+00 -5.97628534e-01
-9.75205958e-01 -5.66422418e-02 -4.74511623e-01 -2.25782350e-01
-7.87158161e-02 4.13486123e-01 -4.99460071e-01 -5.37068009e-01
-4.13101494e-01 -2.63469100e-01 9.90104258e-01 4.74666953e-01
6.93246782e-01 -3.61812562e-01 -6.21365964e-01 4.91334677e-01
1.29659867e+00 -4.22849953e-01 3.39162171e-01 1.92126408e-01
9.89370406e-01 7.68873870e-01 -6.03354752e-01 2.02755287e-01
4.35806900e-01 3.05287004e-01 4.45450217e-01 -7.27642477e-02
2.83027124e-02 8.23711455e-02 1.84854522e-01 1.21624625e+00
-2.35989183e-01 2.21245810e-01 -1.29500306e+00 2.62938589e-02
-1.80740392e+00 -9.95524228e-01 -3.51376325e-01 1.29506636e+00
5.25297523e-01 4.53157902e-01 1.42717540e-01 -2.43512332e-01
8.44358563e-01 1.60487339e-01 -9.33087289e-01 -5.40582180e-01
-3.81886184e-01 1.87966719e-01 3.16265494e-01 5.92372417e-01
-1.12716317e+00 9.81627107e-01 7.94117022e+00 4.35046792e-01
-5.30708432e-01 -2.35701188e-01 5.97366929e-01 3.65458697e-01
-1.91252157e-02 2.24490926e-01 -2.88219362e-01 8.43673870e-02
1.52456272e+00 -1.92184910e-01 6.32702291e-01 4.69082266e-01
3.21180731e-01 2.39076346e-01 -9.58885968e-01 5.57988703e-01
-5.35826266e-01 -1.92855561e+00 3.32240969e-01 5.62008440e-01
7.32583940e-01 5.51639259e-01 -9.35112834e-02 -6.97759688e-02
9.23986912e-01 -1.31196666e+00 4.19287719e-02 7.20393658e-01
4.57234472e-01 -7.78062403e-01 2.18248829e-01 1.55467987e-01
-1.49781680e+00 2.75552627e-02 -5.95753431e-01 -4.62452233e-01
-2.71107465e-01 4.63454962e-01 -7.59373128e-01 4.53553528e-01
5.16923368e-01 1.19132745e+00 -8.13789964e-01 8.13916028e-01
7.64662549e-02 4.66443092e-01 -3.29006165e-01 1.39015406e-01
5.14897108e-01 -6.80797875e-01 4.82214004e-01 1.15120983e+00
-1.52832374e-01 3.49178538e-02 4.59167480e-01 1.21762264e+00
-2.81320840e-01 3.95160019e-02 -9.03783619e-01 -4.89845872e-01
2.71881729e-01 1.49629021e+00 -1.32956302e+00 -3.79708588e-01
-2.23411769e-01 9.18652952e-01 7.49623835e-01 6.02593958e-01
-2.82560170e-01 -3.67049217e-01 6.26339972e-01 2.25721538e-01
2.53712803e-01 -6.87852323e-01 -2.07183212e-01 -1.04010975e+00
-8.40827882e-01 -6.14804626e-01 4.48005885e-01 -7.50021935e-01
-1.61509907e+00 5.63718081e-01 -4.12023842e-01 -6.76104903e-01
-2.39938777e-02 -8.59213948e-01 -7.10553885e-01 8.22516203e-01
-8.76780927e-01 -9.59453881e-01 1.25866950e-01 3.87255222e-01
-5.66128194e-02 -2.32857376e-01 1.11922204e+00 -2.25818247e-01
-5.86157382e-01 -8.59245881e-02 4.77346241e-01 4.38591838e-02
1.07831582e-01 -1.66172290e+00 5.95699608e-01 2.79727042e-01
4.55931015e-02 1.00971937e+00 6.56445086e-01 -7.60097086e-01
-1.28451598e+00 -9.60812271e-01 6.19757354e-01 -4.82703835e-01
1.47063053e+00 -5.59588373e-01 -1.23103166e+00 4.61255223e-01
5.80563366e-01 -1.58931091e-01 8.79840910e-01 1.86504960e-01
-5.43077052e-01 4.46173489e-01 -1.05535817e+00 4.30003464e-01
9.27799642e-01 -8.67653310e-01 -3.60596716e-01 3.52688760e-01
7.65284359e-01 3.20214868e-01 -1.64600217e+00 1.94160759e-01
3.97009045e-01 -1.02543306e+00 9.27719116e-01 -9.80471194e-01
3.09105575e-01 -1.65955946e-01 1.01552702e-01 -1.09288287e+00
-1.08510900e+00 -9.32682872e-01 -5.47306180e-01 5.25638878e-01
4.61128473e-01 -4.24602717e-01 9.99539852e-01 3.56633723e-01
2.43630156e-01 -6.57391965e-01 -8.97303283e-01 -5.29793262e-01
3.16196293e-01 3.11640114e-01 2.05046877e-01 1.08884346e+00
3.78986865e-01 9.90143478e-01 1.11632422e-01 -1.62837822e-02
1.13083982e+00 2.94616014e-01 1.89650208e-01 -2.03783035e+00
-6.28619939e-02 -8.96093130e-01 -9.02882993e-01 -6.74475670e-01
3.19314003e-01 -1.33363211e+00 -6.92410469e-01 -1.65968180e+00
4.07478243e-01 1.64037630e-01 -5.34998238e-01 2.30597064e-01
3.37773383e-01 4.41751719e-01 -2.91981623e-02 7.00087070e-01
-1.03225720e+00 4.20695633e-01 9.58927095e-01 -1.19765550e-01
1.15771778e-02 -4.63246375e-01 -2.50695586e-01 6.41676843e-01
9.31917965e-01 -4.06272322e-01 -5.32863773e-02 2.60337800e-01
3.34764570e-01 1.76177859e-01 2.66433895e-01 -1.20691502e+00
5.15711308e-01 2.17462853e-01 6.61119461e-01 -5.69417655e-01
1.73627272e-01 -4.89647537e-01 2.35971749e-01 7.49482989e-01
-4.88911539e-01 2.78100878e-01 -1.65202305e-01 1.01108813e+00
-1.49367794e-01 4.90519494e-01 9.10426855e-01 -4.02841061e-01
-2.55753696e-01 4.09357697e-01 -9.17787135e-01 -3.25435400e-01
6.82473302e-01 -1.94961399e-01 -2.37205133e-01 -1.03295080e-01
-1.36634290e+00 2.13578790e-01 7.79671133e-01 8.89641866e-02
3.92453641e-01 -1.29152107e+00 -4.61620778e-01 1.23600475e-01
-1.65316761e-01 -4.48940188e-01 9.29890275e-02 8.54797721e-01
-7.23201036e-01 6.71295643e-01 -6.46415770e-01 -7.26412833e-01
-1.00438857e+00 7.35144258e-01 3.31299007e-01 -5.46795905e-01
-7.28489816e-01 3.03279012e-01 -6.91284239e-03 -3.53136688e-01
-1.48967192e-01 -1.04854360e-01 -3.51805300e-01 2.69414842e-01
1.55811667e-01 4.35768127e-01 -2.57880121e-01 -7.72731304e-01
-2.07263023e-01 4.29406613e-01 -6.19765818e-02 2.72146881e-01
1.47699392e+00 -2.92568326e-01 -1.08478975e+00 5.16249061e-01
1.40181422e+00 -8.43575835e-01 -1.21401167e+00 -1.12327188e-01
7.36543536e-01 5.27802527e-01 -9.95265543e-02 -4.37053144e-01
-5.04452169e-01 1.04787946e+00 3.39971244e-01 1.07171941e+00
4.94425565e-01 3.70293818e-02 3.89999092e-01 8.13253582e-01
2.61536419e-01 -8.56519103e-01 1.06479675e-01 8.10870469e-01
4.92209792e-01 -9.88486469e-01 2.36956954e-01 -1.57771870e-01
-8.25971365e-02 1.66157079e+00 1.73671156e-01 -6.32832944e-01
9.45204973e-01 9.31658968e-02 -4.50708985e-01 -8.84144545e-01
-1.04089344e+00 -1.06989807e-02 3.76058131e-01 4.58238304e-01
5.75805485e-01 2.62225926e-01 5.53606041e-02 2.18068976e-02
-1.00790873e-01 -4.88246143e-01 7.33170211e-01 6.19225383e-01
-7.44976938e-01 -1.12693119e+00 1.56172207e-02 6.96366787e-01
-3.21315050e-01 -2.52922148e-01 -9.88095343e-01 6.05981886e-01
-3.29550624e-01 6.36144817e-01 -4.86042835e-02 -1.91799194e-01
-1.88989654e-01 3.22008729e-01 2.89501280e-01 -7.14371622e-01
-5.19036233e-01 1.32357314e-01 -9.48067531e-02 -7.87739575e-01
-5.46990275e-01 -2.64506787e-01 -1.46844625e+00 -6.00126386e-01
1.59601226e-01 3.95792603e-01 4.57395941e-01 6.00372910e-01
7.34458745e-01 5.59593201e-01 4.88265008e-01 -1.38517880e+00
-3.45403329e-02 -9.12192881e-01 -8.65991175e-01 1.52573049e-01
2.88848519e-01 -6.21735990e-01 -4.45704430e-01 -8.28609839e-02] | [6.900984764099121, 5.797737121582031] |
1e755a76-5003-4864-97c2-dce49d2045b4 | programs-as-black-box-explanations | 1611.07579 | null | http://arxiv.org/abs/1611.07579v1 | http://arxiv.org/pdf/1611.07579v1.pdf | Programs as Black-Box Explanations | Recent work in model-agnostic explanations of black-box machine learning has
demonstrated that interpretability of complex models does not have to come at
the cost of accuracy or model flexibility. However, it is not clear what kind
of explanations, such as linear models, decision trees, and rule lists, are the
appropriate family to consider, and different tasks and models may benefit from
different kinds of explanations. Instead of picking a single family of
representations, in this work we propose to use "programs" as model-agnostic
explanations. We show that small programs can be expressive yet intuitive as
explanations, and generalize over a number of existing interpretable families.
We propose a prototype program induction method based on simulated annealing
that approximates the local behavior of black-box classifiers around a specific
prediction using random perturbations. Finally, we present preliminary
application on small datasets and show that the generated explanations are
intuitive and accurate for a number of classifiers. | ['Sameer Singh', 'Carlos Guestrin', 'Marco Tulio Ribeiro'] | 2016-11-22 | null | null | null | null | ['program-induction'] | ['computer-code'] | [ 3.22879761e-01 9.02184725e-01 -5.68962276e-01 -8.13013494e-01
-2.95834839e-01 -4.97063726e-01 5.90093732e-01 2.65069216e-01
4.05251741e-01 7.73789525e-01 5.80907166e-02 -9.70482588e-01
-6.34478033e-01 -5.73987901e-01 -8.55189562e-01 -4.22139853e-01
2.86114067e-02 8.11156571e-01 -3.14342119e-02 -5.77261858e-02
4.84611690e-01 4.87561047e-01 -1.83407474e+00 6.25787020e-01
1.19328141e+00 5.48165977e-01 -1.47779331e-01 8.24588656e-01
-3.49097878e-01 8.67304444e-01 -4.02946383e-01 -4.80186820e-01
-3.59916873e-02 -4.71657991e-01 -1.04615152e+00 2.20097169e-01
2.20240831e-01 -1.07339315e-01 3.51957858e-01 6.33917451e-01
-2.68430054e-01 1.89974178e-02 7.48962224e-01 -1.76031268e+00
-7.67143786e-01 1.07201540e+00 -1.37786835e-01 -3.26899588e-01
3.35348517e-01 2.17995375e-01 1.07714593e+00 -5.12188017e-01
3.69454592e-01 1.50877619e+00 7.98829436e-01 1.06937301e+00
-1.78252220e+00 -3.75636071e-01 6.08263552e-01 6.21104054e-02
-5.52947223e-01 -1.09362483e-01 4.76672977e-01 -5.43668032e-01
1.09787130e+00 8.71433973e-01 6.90200448e-01 9.46650505e-01
2.01031908e-01 6.28899515e-01 1.27512455e+00 -7.71813810e-01
5.17654836e-01 5.22388756e-01 8.25734735e-01 9.44787204e-01
6.78627014e-01 3.32811624e-01 -4.10775870e-01 -7.08256721e-01
3.58212203e-01 1.68030471e-01 -2.89914280e-01 -9.16228116e-01
-7.52978683e-01 1.22079754e+00 3.67586911e-01 -1.39831766e-01
-3.82442474e-02 4.75837618e-01 2.85559595e-01 2.29568914e-01
2.26404384e-01 8.81215692e-01 -7.92337775e-01 4.23807986e-02
-2.94631869e-01 3.68621379e-01 1.06082380e+00 9.36224937e-01
7.70746589e-01 1.36598602e-01 1.96195453e-01 3.51690084e-01
4.15879875e-01 6.87548071e-02 4.12095189e-01 -1.03543091e+00
1.27659425e-01 9.49564040e-01 1.11157708e-01 -5.64185202e-01
-4.33383763e-01 -3.59644860e-01 -6.09268367e-01 5.93223035e-01
3.58554542e-01 -1.47509903e-01 -9.12943900e-01 1.62377679e+00
6.83598369e-02 -4.64186102e-01 1.51707172e-01 5.57899117e-01
5.65531909e-01 4.40675795e-01 2.07537889e-01 -1.33616254e-01
8.08648348e-01 -1.22898364e+00 -2.38538250e-01 -7.04488397e-01
9.53245759e-01 9.12450906e-03 1.29973507e+00 4.58785266e-01
-8.84214103e-01 -5.20973921e-01 -1.08212316e+00 1.73078239e-01
-2.98408061e-01 -1.53579801e-01 1.21473026e+00 8.73027563e-01
-7.67747879e-01 9.04729426e-01 -8.23011816e-01 -3.51257414e-01
2.64921308e-01 6.01358056e-01 -2.04810083e-01 1.28146410e-01
-4.63816285e-01 1.06537712e+00 5.22436082e-01 -3.92648220e-01
-6.43022060e-01 -6.67808712e-01 -8.49455953e-01 4.22515512e-01
2.24527240e-01 -1.07495761e+00 1.43687689e+00 -1.73677850e+00
-1.12962449e+00 4.73392636e-01 -4.88565385e-01 -7.08161294e-01
3.98621261e-01 -1.36321396e-01 -6.09045476e-02 -3.57238144e-01
-2.86488116e-01 7.51749754e-01 6.17820859e-01 -1.58203626e+00
-3.81457359e-01 -1.98263958e-01 5.03297329e-01 -1.58423081e-01
-1.31797791e-02 -6.81015775e-02 3.67797852e-01 -2.15686977e-01
2.43556395e-01 -1.13056779e+00 -7.51702368e-01 7.48495981e-02
-6.32971287e-01 -5.52396514e-02 4.76923347e-01 -1.44574299e-01
1.25784051e+00 -1.73140049e+00 -2.77452730e-02 4.16271716e-01
2.59367287e-01 -1.72587067e-01 8.78763273e-02 2.54890025e-01
-5.90914965e-01 7.73240566e-01 -3.81426454e-01 5.86260408e-02
3.21813703e-01 5.97836614e-01 -6.72286212e-01 -5.81549257e-02
3.69425148e-01 8.11558902e-01 -7.05464363e-01 -2.93209016e-01
7.47289881e-02 -1.14588991e-01 -8.85019839e-01 9.13149619e-04
-7.26387620e-01 2.46989116e-01 -6.01138949e-01 2.06593052e-01
3.72898161e-01 -4.69155461e-01 3.82039279e-01 6.33091390e-01
1.70709223e-01 4.68658000e-01 -9.86714125e-01 9.29534078e-01
-4.31002110e-01 6.15482569e-01 -5.50967515e-01 -1.03740931e+00
9.60585415e-01 1.02942407e-01 -2.95779943e-01 1.69816583e-01
-1.77654356e-01 2.25566685e-01 3.44270706e-01 -5.87078929e-01
1.14806049e-01 -2.44540900e-01 -3.67103219e-02 8.60611618e-01
-3.55236113e-01 -2.75628954e-01 5.69096133e-02 -7.29592517e-02
1.09211302e+00 2.77463436e-01 8.74896586e-01 -3.17823410e-01
2.22969607e-01 6.04905427e-01 5.19672215e-01 1.11662531e+00
4.01012659e-01 6.56071484e-01 8.07619929e-01 -1.01711679e+00
-9.78991687e-01 -7.36929297e-01 -7.49712959e-02 1.02174222e+00
-1.53542951e-01 -4.60535824e-01 -6.58202827e-01 -9.67675865e-01
2.03531951e-01 1.31983662e+00 -7.12996364e-01 -2.55062819e-01
-3.47370654e-01 -6.16426647e-01 9.90803987e-02 6.37230814e-01
-4.11877148e-02 -9.04374421e-01 -9.41181600e-01 -4.55303304e-02
2.01795712e-01 -4.15936947e-01 2.13934891e-02 8.87677431e-01
-1.46889818e+00 -1.29298472e+00 1.27423689e-01 -3.36821377e-01
1.10484433e+00 2.24025831e-01 1.41978061e+00 7.00187862e-01
1.36635408e-01 2.55872995e-01 -2.66558707e-01 -6.86551630e-01
-1.03195941e+00 -8.79394065e-04 -3.68051976e-02 -4.30485040e-01
2.40893140e-01 -5.32433629e-01 -1.05668291e-01 1.17445923e-01
-6.86462998e-01 4.37998414e-01 5.35389960e-01 1.13978994e+00
3.04320514e-01 -1.10788479e-01 3.04964602e-01 -1.54331219e+00
6.89599991e-01 -3.96639347e-01 -3.65672082e-01 5.99621773e-01
-1.05125809e+00 8.43717694e-01 9.59285736e-01 -6.05205059e-01
-1.11283040e+00 4.36515093e-01 1.40272707e-01 -1.37432411e-01
-4.65655714e-01 4.10247475e-01 -2.01400101e-01 -2.72940397e-02
1.23364258e+00 -8.47536623e-02 -1.18368410e-01 -2.76933551e-01
5.52669108e-01 3.81317556e-01 6.15684167e-02 -7.63340890e-01
8.59221935e-01 1.34902477e-01 5.99341430e-02 -2.44766682e-01
-6.89093590e-01 2.66244709e-01 -4.23133075e-01 2.55240947e-01
3.41748118e-01 -3.54678422e-01 -5.61389327e-01 -4.53005582e-01
-1.26763868e+00 -3.83030415e-01 -4.93181795e-01 1.52952120e-01
-8.67974579e-01 2.21875519e-01 -1.62493866e-02 -9.84117031e-01
-1.35009483e-01 -1.20928764e+00 5.92085540e-01 2.96231359e-01
-1.19721437e+00 -9.93878424e-01 -1.60703644e-01 2.02444568e-01
1.75409019e-01 2.79492110e-01 1.78786516e+00 -1.20484328e+00
-7.65584171e-01 -2.53957272e-01 2.48001516e-01 1.19249366e-01
-8.26471820e-02 4.36529696e-01 -1.15976071e+00 2.49890402e-01
-1.26745939e-01 -1.98611990e-01 5.84547997e-01 5.31682909e-01
1.57118285e+00 -1.03686929e+00 -6.46555722e-01 4.87723351e-01
1.05828929e+00 3.14112395e-01 4.42213804e-01 4.79696602e-01
3.11295956e-01 7.81695366e-01 4.68795329e-01 2.34175891e-01
1.46090493e-01 4.81889784e-01 5.94831228e-01 5.10224588e-02
4.54016477e-01 -4.64901000e-01 2.07614198e-01 -7.77380690e-02
3.39994580e-02 -4.66324855e-03 -1.01432645e+00 3.31118494e-01
-2.17113948e+00 -1.14411676e+00 -4.20216233e-01 2.11949158e+00
6.08160317e-01 3.91975164e-01 1.37179270e-01 3.15302759e-01
4.62120354e-01 -2.78691351e-01 -7.70115793e-01 -1.24105251e+00
-1.98205514e-03 2.08053872e-01 1.59030229e-01 6.66999638e-01
-7.34463513e-01 5.51448584e-01 7.07336378e+00 3.14785689e-01
-7.93432355e-01 -1.40343338e-01 9.17784810e-01 2.02632304e-02
-9.15983200e-01 6.27006888e-01 -5.48832774e-01 4.98883128e-02
9.96159256e-01 -3.36138844e-01 3.14373136e-01 1.57690167e+00
1.46883473e-01 8.87401924e-02 -1.71974683e+00 3.91955346e-01
-3.79282773e-01 -1.49336708e+00 1.98633641e-01 -1.12786919e-01
7.71716952e-01 -5.16710162e-01 8.74208752e-04 2.57634550e-01
8.27508390e-01 -1.40023339e+00 9.31025922e-01 3.46078664e-01
1.07500702e-01 -6.56216562e-01 4.34291750e-01 7.34192252e-01
-6.16836548e-01 -6.68233991e-01 -3.79938871e-01 -4.72550750e-01
-4.84984905e-01 2.78172851e-01 -9.40493524e-01 7.75865689e-02
5.29025793e-01 3.51405114e-01 -7.65095890e-01 9.58640933e-01
-5.90292633e-01 6.01622105e-01 -6.18528537e-02 -4.08385724e-01
-1.85977034e-02 1.44228816e-01 2.67505139e-01 1.05059433e+00
3.16323966e-01 2.01288790e-01 -1.28678754e-01 1.24213159e+00
4.86749023e-01 -2.46408910e-01 -9.18053389e-01 1.72328964e-01
4.23221946e-01 1.05318916e+00 -4.93606538e-01 -4.14372802e-01
-1.79178685e-01 4.02760595e-01 3.92964035e-01 2.97795266e-01
-7.56876349e-01 6.17692471e-02 7.60072112e-01 2.08367467e-01
-7.29464460e-03 2.25055054e-01 -9.73514676e-01 -1.12176216e+00
-1.34789467e-01 -1.35844600e+00 5.06369829e-01 -1.17540157e+00
-9.68748212e-01 5.51096678e-01 2.08957687e-01 -1.04290140e+00
-7.32653320e-01 -7.25072265e-01 -9.57888484e-01 6.29526496e-01
-8.25535476e-01 -9.61639762e-01 -3.17062587e-01 -4.07111049e-02
5.82065463e-01 7.82034397e-02 1.12023973e+00 -8.57848465e-01
-3.77437443e-01 3.81265998e-01 -1.52826980e-02 -4.56911772e-01
3.82217243e-02 -1.49554777e+00 5.49326122e-01 5.82194328e-01
1.94162443e-01 1.01530337e+00 1.27814555e+00 -3.68360370e-01
-1.31874967e+00 -9.10262823e-01 8.05480063e-01 -6.92757010e-01
3.64003390e-01 -3.15701887e-02 -1.01468647e+00 1.17691362e+00
2.73732487e-02 -3.11914325e-01 5.41530311e-01 5.02326369e-01
-5.24659455e-01 -1.87320225e-02 -1.16759038e+00 9.34493601e-01
1.09903061e+00 -1.76925033e-01 -6.62392795e-01 3.91261756e-01
7.56621003e-01 -2.01654837e-01 -3.08079809e-01 3.61067563e-01
5.77760518e-01 -1.26613140e+00 8.85293663e-01 -1.32982981e+00
7.57733881e-01 -2.17804939e-01 4.74025635e-03 -1.49723446e+00
-3.12162369e-01 -5.96560836e-01 -1.29105538e-01 1.04274118e+00
8.89876723e-01 -8.71348262e-01 9.83569145e-01 1.45222723e+00
-1.57182440e-01 -9.20130134e-01 -5.14546633e-01 -6.62607074e-01
1.45720109e-01 -5.24641633e-01 1.07202840e+00 5.47521174e-01
5.62187552e-01 4.23821449e-01 4.64206636e-02 1.95857987e-01
5.15038013e-01 7.01162994e-01 1.05961239e+00 -1.45135701e+00
-7.98470140e-01 -6.14663184e-01 -1.42717868e-01 -8.66787612e-01
4.65433717e-01 -9.32392359e-01 -2.36167386e-02 -1.38949728e+00
5.23500800e-01 -6.10579550e-01 2.55671740e-01 7.87596107e-01
-1.99793354e-01 -6.29586756e-01 3.76515724e-02 6.36153817e-02
-1.38258651e-01 2.29845539e-01 7.74484158e-01 -2.96692461e-01
-1.77502856e-01 4.17403609e-01 -1.25713134e+00 1.12808800e+00
9.72660363e-01 -6.82227790e-01 -6.52239978e-01 -2.73077488e-01
3.57994527e-01 1.67474672e-01 5.38818538e-01 -6.62801147e-01
-1.53153181e-01 -7.00511217e-01 3.43238443e-01 2.36726403e-02
1.76664650e-01 -8.53374600e-01 6.66725755e-01 8.79834950e-01
-8.97474468e-01 1.92905322e-01 2.47393265e-01 5.24024308e-01
1.83890477e-01 -7.86675096e-01 6.70114398e-01 -3.14296901e-01
-4.94314194e-01 -2.16171175e-01 -4.55548018e-01 -2.90816516e-01
9.55687344e-01 -4.54581439e-01 -4.94994938e-01 -6.31441891e-01
-6.92915320e-01 6.77591339e-02 6.11405194e-01 2.81944633e-01
5.58600783e-01 -9.06084538e-01 -2.38465920e-01 1.63054615e-01
3.25023383e-01 -2.09120914e-01 -1.76283121e-01 3.65050763e-01
-4.47500110e-01 5.68011820e-01 -1.32626174e-02 -5.92052698e-01
-1.50452483e+00 7.52976060e-01 5.76534808e-01 -2.55080551e-01
-3.52290630e-01 6.21804476e-01 3.96985650e-01 -8.16918433e-01
1.43507108e-01 -7.68654823e-01 2.48777922e-02 -6.47137225e-01
2.61779040e-01 7.93514252e-02 -2.71228284e-01 4.59921770e-02
-3.15907538e-01 4.30532008e-01 1.65628940e-01 2.17137098e-01
1.43541288e+00 2.08624303e-01 5.47834998e-03 4.39940512e-01
4.38218564e-01 -4.15065169e-01 -1.20845783e+00 2.51772434e-01
4.32904631e-01 -3.82830560e-01 -5.89684427e-01 -8.73932064e-01
-6.23640656e-01 1.19957221e+00 1.55100182e-01 5.68461120e-01
1.02507687e+00 1.21392801e-01 -1.14927888e-01 7.38917172e-01
2.95282394e-01 -4.34014797e-01 -1.88297987e-01 2.01068118e-01
1.02980530e+00 -1.16889524e+00 1.08694538e-01 -4.73731220e-01
-8.17467093e-01 1.43939888e+00 7.88747609e-01 3.51493061e-02
2.06182078e-01 2.41605893e-01 -1.80444673e-01 -9.00240988e-03
-1.45630789e+00 7.40907937e-02 2.08381280e-01 8.83740902e-01
3.92556131e-01 2.91471362e-01 1.48800723e-02 9.12505627e-01
-5.95868826e-01 -5.99924363e-02 7.32483923e-01 8.67046058e-01
-7.44652510e-01 -1.13200986e+00 -5.84534466e-01 6.09678388e-01
-7.04192221e-02 -5.46853766e-02 -7.42308259e-01 1.07066834e+00
-1.22630261e-02 1.09901786e+00 -2.78971791e-01 -4.91470098e-01
7.30343834e-02 3.82441431e-01 4.79421020e-01 -8.26699913e-01
-6.72651768e-01 -6.86993897e-01 4.88800615e-01 -2.92900503e-01
-8.21771845e-02 -5.59562087e-01 -1.38120496e+00 -6.21991158e-01
-3.54174942e-01 3.03813636e-01 3.92780334e-01 1.06680381e+00
2.92398125e-01 2.74375707e-01 3.96369576e-01 -6.15041137e-01
-9.47580397e-01 -5.42719960e-01 -1.82728037e-01 1.81816012e-01
3.46593976e-01 -3.98110211e-01 -5.61420083e-01 6.81876540e-02] | [8.782017707824707, 5.808671951293945] |
32d58f15-2e2b-4979-a6e3-0a0a4355aa7c | unsupervised-multi-document-opinion | 1911.02247 | null | https://arxiv.org/abs/1911.02247v2 | https://arxiv.org/pdf/1911.02247v2.pdf | Unsupervised Opinion Summarization as Copycat-Review Generation | Opinion summarization is the task of automatically creating summaries that reflect subjective information expressed in multiple documents, such as product reviews. While the majority of previous work has focused on the extractive setting, i.e., selecting fragments from input reviews to produce a summary, we let the model generate novel sentences and hence produce abstractive summaries. Recent progress in summarization has seen the development of supervised models which rely on large quantities of document-summary pairs. Since such training data is expensive to acquire, we instead consider the unsupervised setting, in other words, we do not use any summaries in training. We define a generative model for a review collection which capitalizes on the intuition that when generating a new review given a set of other reviews of a product, we should be able to control the "amount of novelty" going into the new review or, equivalently, vary the extent to which it deviates from the input. At test time, when generating summaries, we force the novelty to be minimal, and produce a text reflecting consensus opinions. We capture this intuition by defining a hierarchical variational autoencoder model. Both individual reviews and the products they correspond to are associated with stochastic latent codes, and the review generator ("decoder") has direct access to the text of input reviews through the pointer-generator mechanism. Experiments on Amazon and Yelp datasets, show that setting at test time the review's latent code to its mean, allows the model to produce fluent and coherent summaries reflecting common opinions. | ['Arthur Bražinskas', 'Ivan Titov', 'Mirella Lapata'] | 2019-11-06 | unsupervised-opinion-summarization-as-copycat | https://aclanthology.org/2020.acl-main.461 | https://aclanthology.org/2020.acl-main.461.pdf | acl-2020-6 | ['review-generation', 'unsupervised-opinion-summarization'] | ['natural-language-processing', 'natural-language-processing'] | [ 4.65536356e-01 4.72286403e-01 -2.80393660e-01 -3.99247527e-01
-7.95380235e-01 -7.99633265e-01 6.95355356e-01 3.63748819e-01
-4.47865017e-03 6.79693997e-01 6.24517679e-01 -1.70626909e-01
4.40546215e-01 -8.28594685e-01 -7.35266685e-01 -6.69014812e-01
5.56924164e-01 5.70000172e-01 -3.12788844e-01 -1.06207877e-01
4.64150846e-01 -1.84286252e-01 -1.51862597e+00 3.42568189e-01
1.17795634e+00 5.47270417e-01 1.57579288e-01 1.04441762e+00
-1.17472529e-01 6.94916248e-01 -1.22983205e+00 -8.02929699e-01
1.63710624e-01 -8.87868941e-01 -4.50572997e-01 6.10046804e-01
1.88116863e-01 -3.66689622e-01 8.81119147e-02 1.14005160e+00
3.67550403e-01 3.11722666e-01 1.08398461e+00 -9.69226956e-01
-1.15930474e+00 1.06847024e+00 -5.31377017e-01 1.75300892e-03
4.03597832e-01 1.42320588e-01 1.56305993e+00 -7.93565273e-01
7.36944020e-01 9.49345171e-01 1.60107598e-01 7.00881362e-01
-1.39197016e+00 -1.33717060e-01 5.06143034e-01 -4.70809370e-01
-7.19020545e-01 -5.71924627e-01 7.84920573e-01 -3.90224576e-01
9.04869139e-01 1.30870879e-01 6.32606328e-01 1.15076125e+00
4.99802053e-01 1.10056448e+00 5.38135469e-01 -4.31742996e-01
7.41736233e-01 5.42647719e-01 2.53167778e-01 1.90002337e-01
5.23549378e-01 -5.17285228e-01 -5.99995911e-01 -3.22480381e-01
1.61646888e-01 -5.38788475e-02 -2.97576457e-01 -2.54263848e-01
-7.57583618e-01 1.24410892e+00 -1.16743781e-01 1.54088020e-01
-8.23387980e-01 2.76050158e-02 4.42187279e-01 3.77721280e-01
8.84517252e-01 5.40185213e-01 -3.50154400e-01 -1.55514283e-02
-1.14543414e+00 5.29861152e-01 1.30487514e+00 8.49650979e-01
7.77966440e-01 1.50425909e-02 -3.60903949e-01 7.06527889e-01
1.88699588e-01 5.77232778e-01 6.97853744e-01 -9.83123362e-01
4.61625665e-01 5.68300843e-01 2.52098680e-01 -9.26163852e-01
3.17309529e-01 -3.76099139e-01 -6.18114114e-01 -9.17892829e-02
-3.01359683e-01 -5.48296690e-01 -1.06579614e+00 1.63064384e+00
-2.15615220e-02 -3.24463576e-01 5.25129139e-01 6.26262486e-01
6.90822780e-01 1.04998922e+00 -1.88096806e-01 -5.71432650e-01
7.46437132e-01 -1.06042993e+00 -7.01043963e-01 -3.66128236e-01
5.28490782e-01 -6.13501728e-01 1.00083470e+00 5.25373876e-01
-1.36083651e+00 -3.86626840e-01 -1.25404620e+00 1.12501673e-01
-2.32315600e-01 1.23306654e-01 3.14466864e-01 3.79876107e-01
-1.12673593e+00 6.98172033e-01 -6.49362803e-01 -2.20956430e-01
2.35261455e-01 1.49236798e-01 -7.61253238e-02 4.01901901e-02
-9.91414547e-01 7.09300995e-01 -2.50659809e-02 1.27593977e-02
-7.10352540e-01 -4.34411973e-01 -1.00716436e+00 2.46086329e-01
3.69668394e-01 -1.06199515e+00 1.66716135e+00 -1.46851718e+00
-1.55867839e+00 3.67268533e-01 -5.33146560e-01 -3.92226398e-01
2.26753786e-01 -1.48625985e-01 -1.39970809e-01 3.34532037e-02
4.82013524e-01 7.05660462e-01 1.05393672e+00 -1.42971897e+00
-5.85593700e-01 -1.45061746e-01 1.02706268e-01 3.90952557e-01
-2.97265232e-01 -1.93513408e-01 -3.30562055e-01 -3.69371980e-01
-2.80486971e-01 -9.94203508e-01 -4.82464164e-01 -6.64054036e-01
-6.78495884e-01 -2.71872848e-01 4.49559540e-01 -5.66228509e-01
1.30051470e+00 -1.89950585e+00 3.14878285e-01 2.68656760e-01
2.47152463e-01 4.60070930e-02 -1.74695268e-01 5.97144365e-01
2.01821327e-01 4.26965386e-01 -4.86808926e-01 -8.36027622e-01
1.71966389e-01 2.70457596e-01 -6.71418250e-01 5.78852259e-02
3.38466555e-01 9.21312034e-01 -1.08344138e+00 -2.81457722e-01
-2.27668643e-01 1.75248340e-01 -9.03015316e-01 1.57112122e-01
-6.54721618e-01 1.82964629e-03 -6.22728765e-01 2.11350292e-01
4.41699088e-01 -3.46912950e-01 -3.54261347e-03 2.00838909e-01
-2.11215243e-02 4.13547128e-01 -1.04823387e+00 1.30160010e+00
-5.92100859e-01 5.63911855e-01 -2.42488533e-01 -7.77716339e-01
8.15325618e-01 3.21971059e-01 9.03046951e-02 -1.30016536e-01
1.12914398e-01 2.88773179e-02 -1.59617141e-01 -3.79261762e-01
1.04610431e+00 -2.10854366e-01 -1.68062583e-01 9.29098368e-01
-3.29433419e-02 -4.44351912e-01 7.21619725e-01 6.78402960e-01
1.09061038e+00 -1.66448280e-01 3.13828170e-01 1.61916867e-01
3.06046247e-01 1.03883408e-01 3.61888766e-01 1.08552182e+00
2.06157535e-01 8.78815711e-01 9.88667488e-01 8.04346725e-02
-1.07599378e+00 -1.05227971e+00 4.15889978e-01 8.65079343e-01
-1.69524953e-01 -6.65589213e-01 -7.35655665e-01 -7.38433123e-01
-2.31052432e-02 1.30981433e+00 -8.07491958e-01 -2.20124811e-01
-1.26578048e-01 -4.68557447e-01 -1.00096419e-01 3.44900817e-01
6.65515214e-02 -1.32793295e+00 -5.49692810e-01 4.46236581e-01
-1.45373970e-01 -6.08404100e-01 -8.52145791e-01 1.53216064e-01
-7.40672112e-01 -6.72296166e-01 -7.46177077e-01 -5.62991560e-01
8.98795009e-01 8.18723738e-02 1.38887930e+00 -2.23699614e-01
2.93462396e-01 5.09100735e-01 -4.57162112e-01 -6.13415360e-01
-8.40250909e-01 2.68438071e-01 -6.89563528e-02 1.63732648e-01
3.29668134e-01 -4.79471475e-01 -4.75440234e-01 -3.26939464e-01
-1.22974825e+00 -8.23073387e-02 7.04380691e-01 8.14207494e-01
5.60176909e-01 7.35643506e-02 8.91590834e-01 -1.20568097e+00
1.47000706e+00 -6.13977790e-01 -1.48481950e-01 2.58735240e-01
-7.64635503e-01 2.83856422e-01 7.75254309e-01 -6.09135389e-01
-1.03141546e+00 -3.02189797e-01 1.63434520e-01 -2.74006724e-01
1.58039302e-01 9.74317551e-01 7.05525875e-02 9.76305425e-01
5.89213431e-01 3.53936195e-01 8.00881186e-04 7.89523572e-02
5.76726973e-01 7.79135108e-01 2.88485408e-01 -2.86070168e-01
4.07027602e-01 1.24023855e-01 -7.46159971e-01 -6.61363125e-01
-8.21342826e-01 -3.23965371e-01 -2.82763243e-01 -4.09576036e-02
7.25646973e-01 -7.39851654e-01 -1.67401060e-01 9.97482687e-02
-1.29169047e+00 -1.14588805e-01 -7.53099978e-01 1.52580306e-01
-3.69801372e-01 1.20821767e-01 -3.28145266e-01 -9.36037779e-01
-6.79994106e-01 -1.11032379e+00 1.16510916e+00 4.49049979e-01
-7.88405776e-01 -1.02575099e+00 4.37604576e-01 1.22652657e-03
3.24533403e-01 8.70625749e-02 9.07790422e-01 -9.99673903e-01
-3.22006732e-01 -5.14467239e-01 3.61269176e-01 9.45086360e-01
2.76268154e-01 4.52437103e-01 -6.54087126e-01 -3.90026160e-02
1.99236080e-01 -2.90615916e-01 9.57446277e-01 5.91574669e-01
4.78057861e-01 -7.45784104e-01 -1.68027937e-01 -2.17727453e-01
1.00072658e+00 1.22636303e-01 2.56312251e-01 -2.17888236e-01
4.04284000e-01 8.20204556e-01 2.57035375e-01 4.94559675e-01
4.34194505e-01 6.35530725e-02 -9.49102715e-02 3.09292138e-01
3.82049382e-01 -4.22539890e-01 6.31064236e-01 1.11381888e+00
4.12604541e-01 -7.69584835e-01 -3.74769688e-01 8.12765896e-01
-1.90008664e+00 -1.03762174e+00 3.84710968e-01 2.10007143e+00
9.63998735e-01 3.45237285e-01 -1.31900415e-01 -1.79479405e-01
6.65286481e-01 4.45689619e-01 -8.22473109e-01 -1.02013004e+00
2.02162191e-02 2.40447119e-01 1.24590687e-01 6.74352407e-01
-5.90743601e-01 7.46339381e-01 5.83500957e+00 3.34202796e-01
-1.09719312e+00 -2.16603115e-01 7.74646759e-01 -4.59042072e-01
-1.09446025e+00 1.26296073e-01 -7.55727768e-01 5.36913931e-01
1.12076235e+00 -7.20595181e-01 1.93097174e-01 8.47069144e-01
3.94553185e-01 -4.01095957e-01 -1.34006870e+00 4.38683361e-01
4.65335757e-01 -1.19577026e+00 4.94385481e-01 5.76149002e-02
1.28941703e+00 -1.60397515e-01 1.99190184e-01 4.40246999e-01
6.54866576e-01 -8.48421931e-01 6.94225490e-01 5.03655016e-01
3.18617165e-01 -9.04228985e-01 8.77041459e-01 4.50911969e-01
-6.68576360e-01 4.71270457e-02 -4.88137960e-01 -1.19828649e-01
3.51322711e-01 8.94751132e-01 -8.98980141e-01 4.14172351e-01
1.49280220e-01 8.53900850e-01 -4.16783243e-01 4.61284727e-01
-4.26571965e-01 6.90326512e-01 -4.87060435e-02 -4.39670950e-01
2.56127775e-01 -4.27010834e-01 5.10542095e-01 1.00293994e+00
3.39143395e-01 -9.80031714e-02 2.54839718e-01 1.05032265e+00
-3.58082235e-01 2.35458657e-01 -8.07516158e-01 -4.24398094e-01
3.14063907e-01 1.13508487e+00 -5.18811285e-01 -6.67847633e-01
-2.86870867e-01 1.07993364e+00 1.05720714e-01 5.31941533e-01
-3.17362577e-01 -3.87443244e-01 3.79708290e-01 -2.73179054e-01
7.44275331e-01 1.86204955e-01 -5.18974960e-01 -1.20535243e+00
3.40695143e-01 -9.90377665e-01 -3.47816832e-02 -9.60670114e-01
-1.14570153e+00 5.83871603e-01 -1.06447384e-01 -8.69837046e-01
-8.26730907e-01 -2.85740290e-03 -1.20145679e+00 9.60113943e-01
-1.09527504e+00 -6.06192410e-01 2.45779887e-01 -7.83313364e-02
9.50020313e-01 -7.03303609e-03 6.30712926e-01 -4.06835020e-01
-5.42538047e-01 3.65866810e-01 1.36889130e-01 -2.87381738e-01
6.62131131e-01 -1.48874867e+00 5.59298158e-01 8.50968838e-01
2.19433665e-01 1.01728284e+00 1.25446796e+00 -8.98291588e-01
-1.06077588e+00 -9.76227820e-01 1.19435751e+00 -5.96533179e-01
6.71132803e-01 -2.03637540e-01 -8.14204633e-01 6.81927204e-01
6.21355653e-01 -6.69349313e-01 7.36955285e-01 -9.96549707e-03
-4.04406413e-02 1.92511547e-02 -8.31519961e-01 9.12907898e-01
3.66033435e-01 -5.00119328e-01 -8.09439361e-01 1.68499276e-01
9.06371236e-01 2.17591152e-02 -3.67249489e-01 8.66807997e-02
4.75305051e-01 -9.13501680e-01 2.43102908e-01 -4.09062207e-01
1.15162361e+00 -1.60329610e-01 1.02517888e-01 -1.99452603e+00
-4.52688982e-04 -5.86639225e-01 -2.85027534e-01 1.37520349e+00
8.61071050e-01 -4.32368249e-01 8.26140225e-01 7.95763135e-01
-1.76799506e-01 -1.02353954e+00 -3.91425222e-01 -1.89903691e-01
1.10348225e-01 -1.32467583e-01 4.20183599e-01 2.60456055e-01
1.00276992e-01 7.75174558e-01 -1.57030493e-01 -5.70677826e-03
1.68147832e-01 2.29449272e-01 9.35397744e-01 -9.99329209e-01
-5.00112593e-01 -4.70177203e-01 2.62302309e-01 -1.26362336e+00
3.13678056e-01 -6.63242936e-01 4.65968549e-01 -1.83228493e+00
3.23433220e-01 2.10864425e-01 -7.08739320e-03 3.05010881e-02
-5.17499626e-01 -1.87223449e-01 7.48129785e-02 2.32884213e-01
-6.56140983e-01 6.37990355e-01 1.12321317e+00 -2.74710655e-01
-6.21793628e-01 3.09926480e-01 -1.42115843e+00 4.98471379e-01
6.63593411e-01 -4.64853615e-01 -7.99596190e-01 -2.13015378e-01
5.50264895e-01 1.84324682e-01 -1.47699341e-01 -3.76337528e-01
3.67671460e-01 -2.54169456e-03 1.97229207e-01 -7.43052602e-01
-4.99781407e-03 -2.03252167e-01 -9.49317068e-02 1.03814729e-01
-8.99408996e-01 3.16296041e-01 -1.67380854e-01 7.17110813e-01
-3.09009135e-01 -6.61365569e-01 3.05503935e-01 -3.21035564e-01
7.55404681e-02 5.23621589e-02 -7.63570547e-01 4.81667370e-02
5.35743535e-01 -1.82557657e-01 -1.85602725e-01 -8.88437450e-01
-5.11768460e-01 3.45392197e-01 6.86366141e-01 3.11715245e-01
5.69411039e-01 -9.26811397e-01 -8.79909277e-01 1.67036086e-01
1.79572657e-01 2.07941175e-01 3.12317088e-02 4.52101499e-01
2.22041886e-02 3.06647360e-01 4.41660941e-01 -3.29117119e-01
-8.97937596e-01 3.56886744e-01 -5.52821644e-02 -5.78700960e-01
-3.32504988e-01 6.46936834e-01 2.29337737e-01 -2.46229947e-01
6.57599559e-03 -4.31490660e-01 -3.54342610e-01 4.19894755e-01
5.41958392e-01 -1.64205041e-02 8.27716812e-02 -2.44595960e-01
1.60595641e-01 1.53075635e-01 -6.39165044e-01 -5.69250882e-01
1.23046064e+00 -1.50571421e-01 -8.21906775e-02 8.10740829e-01
1.12000608e+00 3.96994531e-01 -1.20498180e+00 -2.92408634e-02
-1.16854988e-01 -4.39773947e-02 -1.20996676e-01 -6.20195568e-01
-8.11091661e-01 4.94244397e-01 -2.05543056e-01 5.88311434e-01
7.36331642e-01 1.51828498e-01 6.52829111e-01 6.60642326e-01
-1.66564390e-01 -1.17650902e+00 3.00179601e-01 5.08008897e-01
8.63139093e-01 -1.21037543e+00 5.85811809e-02 1.66066542e-01
-1.14469504e+00 9.05973732e-01 3.58215094e-01 -4.23423201e-01
3.14223915e-01 9.52927917e-02 1.08494841e-01 -1.85923725e-01
-1.26408219e+00 1.69830441e-01 1.75102964e-01 4.30954337e-01
4.18106198e-01 9.73558202e-02 -2.03095347e-01 7.24082112e-01
-5.56583345e-01 -1.08863905e-01 1.14889562e+00 9.55104113e-01
-6.45479321e-01 -9.87817824e-01 -5.52474149e-02 9.67491508e-01
-3.53054076e-01 -3.20153594e-01 -6.55545890e-01 3.23307723e-01
-3.34530354e-01 1.25790894e+00 1.68936491e-01 -1.49788380e-01
2.97128379e-01 1.33288428e-01 -6.11988362e-03 -1.20626128e+00
-5.64952135e-01 5.36887050e-02 1.41892150e-01 4.68316972e-02
-2.64112920e-01 -9.27114546e-01 -1.00945210e+00 -2.41567045e-01
-4.35053885e-01 6.98619068e-01 6.70588613e-01 1.06824827e+00
3.90340358e-01 5.89539170e-01 8.32142830e-01 -8.38249028e-01
-9.08220708e-01 -1.02459466e+00 -4.69359249e-01 3.67688119e-01
5.49501657e-01 -1.30865663e-01 -7.60744810e-01 3.08420122e-01] | [12.404030799865723, 9.283031463623047] |
658e8732-c23f-46b3-8d9c-b111c9286c2c | extreme-classification-for-answer-type | 2304.12395 | null | https://arxiv.org/abs/2304.12395v2 | https://arxiv.org/pdf/2304.12395v2.pdf | Extreme Classification for Answer Type Prediction in Question Answering | Semantic answer type prediction (SMART) is known to be a useful step towards effective question answering (QA) systems. The SMART task involves predicting the top-$k$ knowledge graph (KG) types for a given natural language question. This is challenging due to the large number of types in KGs. In this paper, we propose use of extreme multi-label classification using Transformer models (XBERT) by clustering KG types using structural and semantic features based on question text. We specifically improve the clustering stage of the XBERT pipeline using textual and structural features derived from KGs. We show that these features can improve end-to-end performance for the SMART task, and yield state-of-the-art results. | ['Vinay Setty'] | 2023-04-24 | null | null | null | null | ['type-prediction', 'extreme-multi-label-classification', 'type'] | ['computer-code', 'methodology', 'speech'] | [-5.40922768e-03 3.08884442e-01 -1.94632590e-01 -4.70973462e-01
-1.05461931e+00 -7.07640171e-01 2.83867508e-01 5.04014909e-01
-1.60072207e-01 3.13727438e-01 2.59680748e-01 -5.89516103e-01
-4.17440295e-01 -9.84015822e-01 -7.07053900e-01 -1.40594989e-01
3.76145810e-01 8.20026398e-01 6.95551753e-01 -4.67879564e-01
5.33620834e-01 -3.45473699e-02 -1.71810710e+00 1.16160464e+00
1.15481341e+00 1.47065508e+00 -3.15618888e-02 9.18959022e-01
-7.85578012e-01 1.56602550e+00 -1.73370287e-01 -8.30311358e-01
-2.21624672e-01 -2.32217386e-01 -1.88163662e+00 -5.89550436e-01
9.48030651e-01 2.54650861e-01 -8.56100395e-02 7.81814933e-01
3.02037776e-01 9.66635123e-02 6.55327857e-01 -1.41666412e+00
-3.49167913e-01 5.04710197e-01 -1.63870007e-01 9.58578736e-02
5.84016442e-01 -3.56150776e-01 1.82756734e+00 -6.73866093e-01
8.88506532e-01 1.33365917e+00 8.49990368e-01 4.05302852e-01
-9.83033359e-01 -1.35543272e-01 -9.33588818e-02 1.13866329e+00
-1.17703533e+00 -2.83345491e-01 8.49800527e-01 -3.63375127e-01
1.20950115e+00 3.51500750e-01 1.59950078e-01 2.63886184e-01
-1.31895214e-01 9.74428535e-01 1.30813456e+00 -7.41278410e-01
2.12045625e-01 8.64033997e-02 8.50102246e-01 1.42941475e+00
-1.91548795e-01 -8.37002277e-01 -6.96783900e-01 -5.58462620e-01
-3.68675739e-01 -5.30741513e-01 -6.98604956e-02 -4.02802080e-01
-6.30600393e-01 1.05545509e+00 6.16432369e-01 1.76057324e-01
-1.21217430e-01 4.13132071e-01 2.90979773e-01 5.71000099e-01
4.81261432e-01 7.37674296e-01 -1.01252592e+00 -2.05788538e-01
-5.81371367e-01 4.60886955e-01 1.20916581e+00 8.55647206e-01
1.05539060e+00 -7.59276628e-01 -2.23706514e-01 1.08378637e+00
2.07642540e-01 2.38475412e-01 1.95664123e-01 -1.05560517e+00
5.40423393e-01 1.28733063e+00 -1.83434337e-01 -7.47360706e-01
-7.00452745e-01 -3.10461402e-01 -1.05716668e-01 -4.53559518e-01
6.18404865e-01 1.76442042e-01 -7.50405133e-01 1.35656023e+00
7.28106260e-01 -1.08158134e-01 2.00851653e-02 4.49139446e-01
1.26888263e+00 3.54555935e-01 1.85437232e-01 1.80102110e-01
1.77515566e+00 -9.75822270e-01 -5.09324908e-01 -2.69021392e-01
1.51518464e+00 -5.18199205e-01 1.17118049e+00 1.66619003e-01
-6.09355986e-01 -1.55615240e-01 -4.46817368e-01 -5.56713820e-01
-7.20019460e-01 -1.25034466e-01 5.40053248e-01 8.09325159e-01
-1.13194883e+00 1.74884021e-01 -2.62741506e-01 -2.45675847e-01
5.19780815e-01 2.44602859e-01 -6.37668669e-02 -7.17833340e-01
-1.25563359e+00 8.84061098e-01 2.69498229e-01 -4.61290121e-01
-3.50977689e-01 -9.39528584e-01 -6.51989281e-01 1.04882650e-01
1.00958610e+00 -1.04653895e+00 1.31078064e+00 -3.47308576e-01
-1.14261591e+00 8.26209545e-01 -6.09250069e-01 -3.24439913e-01
-1.03050746e-01 9.95137468e-02 -9.02728885e-02 8.79457414e-01
-4.55615297e-02 3.94093424e-01 7.18582153e-01 -1.13714504e+00
-8.37853909e-01 -6.93592131e-01 2.19330534e-01 1.42494902e-01
-2.78096735e-01 1.89673424e-01 -2.78134227e-01 -5.16558811e-02
2.02190876e-01 -8.80111814e-01 1.20576739e-01 -5.02930701e-01
-2.91831076e-01 -1.07671249e+00 8.27894807e-01 -1.06756043e+00
1.21623671e+00 -1.46438479e+00 1.45167103e-02 3.92555296e-01
6.37916148e-01 7.40804225e-02 1.10972047e-01 4.06352550e-01
3.49566340e-01 1.96192890e-01 1.87696964e-01 -2.96546593e-02
2.63647676e-01 6.05065599e-02 -1.74194336e-01 -1.51687577e-01
-1.17871992e-01 1.44538689e+00 -8.18860948e-01 -7.85883248e-01
-1.89884380e-01 -2.97125369e-01 -8.44767272e-01 1.54864386e-01
-1.11573374e+00 -1.65401325e-01 -6.10195696e-01 7.85832226e-01
3.10769945e-01 -6.54894948e-01 1.06418744e-01 -3.81085008e-01
5.09784281e-01 7.46452272e-01 -8.24976325e-01 1.37032461e+00
-4.58654433e-01 1.48663327e-01 -4.17322852e-03 -1.21826208e+00
5.67060947e-01 1.62536576e-01 3.07961911e-01 -7.05319881e-01
-6.19553998e-02 3.88738751e-01 -1.50242656e-01 -7.57328629e-01
5.42169392e-01 -2.08014384e-01 -2.68008143e-01 3.05780500e-01
3.26094419e-01 -3.46177042e-01 2.53031522e-01 7.50005007e-01
1.61835945e+00 -2.31744409e-01 -1.18191347e-01 -2.64311135e-01
4.50087875e-01 5.70917010e-01 2.05971122e-01 7.56849468e-01
9.32201445e-02 5.81989177e-02 9.02284563e-01 -1.35478199e-01
-5.44269562e-01 -6.80342913e-01 2.82300174e-01 1.45683885e+00
-1.33194536e-01 -7.50600517e-01 -6.98044121e-01 -1.35680306e+00
2.47026712e-01 8.71669412e-01 -4.64985639e-01 -2.88772225e-01
-6.34541333e-01 -5.53995967e-01 7.21580684e-01 3.53906721e-01
4.86052752e-01 -7.89229512e-01 -2.38380015e-01 1.97194695e-01
-8.53781164e-01 -1.47754669e+00 -3.23217995e-02 1.43949026e-02
-5.29289842e-01 -1.53009987e+00 -1.56277329e-01 -8.61896694e-01
3.37983340e-01 2.15547189e-01 1.43078828e+00 3.09942931e-01
-2.25130305e-01 8.85918736e-01 -8.87512326e-01 -1.46950215e-01
-4.28944379e-01 4.59542513e-01 -4.95053768e-01 -6.40442362e-03
3.42027456e-01 -6.70688972e-03 -5.46912134e-01 2.01332420e-01
-6.53063118e-01 -1.07047088e-01 2.00371832e-01 6.08073950e-01
4.18744534e-01 2.21260324e-01 8.31486166e-01 -1.22057700e+00
6.21862292e-01 -4.06527102e-01 -3.56345117e-01 8.70381057e-01
-1.00279534e+00 3.80657345e-01 7.93541193e-01 2.07346320e-01
-1.06695962e+00 -3.93412679e-01 -4.16456014e-01 3.03685432e-03
4.12954092e-02 9.26189482e-01 -1.70183226e-01 -5.56211233e-01
5.94501317e-01 -1.06368452e-01 -2.10207865e-01 -5.87492764e-01
8.57764900e-01 9.07934666e-01 2.03542560e-01 -7.05763638e-01
4.17818874e-01 2.52700895e-01 2.51005143e-01 -6.41931176e-01
-1.44484234e+00 -1.12829137e+00 -3.43703032e-01 -3.64162832e-01
8.19595158e-01 -5.33677697e-01 -1.17152643e+00 3.91106158e-01
-7.37795770e-01 -3.75320882e-01 -1.07225589e-02 -2.08356440e-01
-5.04024565e-01 5.36224544e-01 -8.24560046e-01 -6.29929125e-01
-4.79427040e-01 -7.96267271e-01 9.73006070e-01 -1.15685739e-01
-5.45093976e-02 -1.08209753e+00 -2.79265456e-02 1.51950884e+00
2.51023799e-01 -2.33499095e-01 1.89264274e+00 -8.46375883e-01
-6.44531786e-01 -1.18353680e-01 -3.75235707e-01 1.31869256e-01
-4.39702034e-01 -6.89648390e-01 -7.97846615e-01 -5.19950874e-03
-1.43558681e-01 -8.23892176e-01 1.14850783e+00 -4.27493788e-02
1.19870853e+00 -5.13853848e-01 -2.88054973e-01 4.21082415e-02
1.58916855e+00 -3.65205705e-01 4.33589488e-01 2.15703994e-01
1.15463424e+00 1.02048409e+00 5.27088165e-01 -1.08510673e-01
1.25040257e+00 7.87500083e-01 2.93456942e-01 4.88943338e-01
-3.99783671e-01 -3.48473936e-01 8.45012888e-02 7.47823000e-01
5.16617537e-01 -2.12702602e-01 -1.45624304e+00 7.68634379e-01
-1.74488485e+00 -9.15594399e-01 -7.80626178e-01 1.72713852e+00
9.60279524e-01 -1.22640312e-01 1.59453258e-01 2.75160819e-01
3.64657462e-01 -2.73023266e-02 -3.27863932e-01 -3.69094640e-01
-3.81127074e-02 6.22146070e-01 5.26003897e-01 6.22158051e-01
-7.42096245e-01 1.31242919e+00 5.40482092e+00 1.29403448e+00
-4.05146956e-01 3.16146761e-01 4.18829978e-01 3.75777870e-01
-5.06235540e-01 2.69020647e-01 -1.00067341e+00 1.55884251e-01
1.23711884e+00 1.07382208e-01 5.95609605e-01 6.24692857e-01
-4.77325439e-01 -4.23940867e-01 -7.59830296e-01 6.39129817e-01
3.40383559e-01 -1.52020311e+00 1.08623959e-01 -2.45584220e-01
5.13204157e-01 6.08702414e-02 -1.97980449e-01 5.28128922e-01
6.54816151e-01 -7.28017092e-01 3.66469234e-01 3.54696661e-01
5.55119753e-01 -6.14455402e-01 5.87904632e-01 5.95474899e-01
-1.14592493e+00 -3.63575011e-01 -7.86843970e-02 3.19951057e-01
-1.66122243e-01 7.09973395e-01 -1.11872292e+00 7.88027167e-01
7.04524577e-01 2.65815675e-01 -1.02085745e+00 7.23815501e-01
-2.20389068e-01 1.00052404e+00 -3.49353552e-01 -4.51842010e-01
1.29315406e-01 1.18579760e-01 1.41147569e-01 9.71731186e-01
-1.08614437e-01 3.84445131e-01 -3.60989459e-02 5.30131876e-01
-4.74713087e-01 3.00718546e-01 -1.35192975e-01 -5.17782532e-02
3.03061604e-01 1.34864044e+00 -5.04318655e-01 -6.15367770e-01
-3.20090294e-01 7.73513615e-01 8.83798718e-01 2.63546593e-02
-3.15242112e-01 -5.56638777e-01 9.35907215e-02 8.06589276e-02
5.26585758e-01 3.34904622e-03 -3.48067522e-01 -1.23133159e+00
1.30350396e-01 -8.81664574e-01 1.14357567e+00 -9.07372832e-01
-1.36620951e+00 4.64386567e-02 -1.40636489e-01 -3.36551845e-01
-2.21958473e-01 -7.68916011e-01 -8.68956149e-02 5.82703352e-01
-1.82832265e+00 -1.43509269e+00 -3.07050496e-01 7.97808290e-01
1.29403427e-01 1.07625142e-01 8.31068218e-01 1.08564094e-01
-2.01141104e-01 7.95249343e-01 1.09750748e-01 8.12023729e-02
5.30929565e-01 -1.75675416e+00 4.64226976e-02 4.34725881e-01
3.26583922e-01 3.00403625e-01 5.37083447e-01 -6.85368121e-01
-1.75173640e+00 -1.16796577e+00 1.31750858e+00 -8.53904903e-01
7.46306002e-01 -2.56251574e-01 -8.80088151e-01 4.09175128e-01
-1.04587086e-01 -1.21381395e-01 8.18178952e-01 7.46746540e-01
-9.47349727e-01 2.41858903e-02 -1.24545729e+00 2.06855446e-01
9.85556960e-01 -9.34894502e-01 -8.11948597e-01 4.62329417e-01
8.23188007e-01 -1.82874158e-01 -1.14850390e+00 4.65200424e-01
1.83104619e-01 -7.55653143e-01 8.04869592e-01 -6.48491561e-01
4.17406708e-01 -1.46809533e-01 -9.42220017e-02 -1.36626339e+00
-1.52949944e-01 -2.07386538e-01 -4.78089064e-01 1.04997087e+00
6.19163275e-01 -6.26093447e-01 1.15578926e+00 8.24480832e-01
-2.64178276e-01 -8.66296232e-01 -1.08915293e+00 -5.02114296e-01
6.17488623e-02 -4.63993102e-01 3.25982094e-01 1.11748183e+00
5.27577281e-01 8.77642572e-01 2.46938869e-01 1.08072355e-01
6.28729641e-01 6.28964484e-01 5.26106656e-01 -1.35163391e+00
-3.43284309e-01 -3.76449883e-01 -3.27274472e-01 -1.04341137e+00
4.09932703e-01 -1.46851516e+00 -2.28751093e-01 -1.99985242e+00
2.46839359e-01 -7.36779869e-01 -5.34467511e-02 7.09994733e-01
-5.84768414e-01 5.61884344e-02 1.10931098e-01 1.24927625e-01
-1.33501279e+00 4.92316991e-01 8.22899580e-01 -1.72957152e-01
3.47170502e-01 -2.27386683e-01 -7.55299151e-01 4.73835528e-01
7.90710449e-01 -4.48050767e-01 -3.16230416e-01 -3.00457895e-01
7.87255764e-01 2.72496283e-01 3.43051344e-01 -5.70629358e-01
3.91022503e-01 1.48033097e-01 -2.92487800e-01 -6.91047192e-01
3.74074787e-01 -5.28841376e-01 -5.55264950e-01 3.72165382e-01
-5.06184042e-01 -1.44556597e-01 -1.71647578e-01 6.46601975e-01
-1.38751313e-01 -4.13157046e-01 4.58391190e-01 -1.74812317e-01
-7.85306454e-01 9.56597850e-02 -5.72615266e-02 7.27931499e-01
5.67957640e-01 2.28955150e-01 -8.77479911e-01 -3.80353034e-01
-4.67964858e-01 5.62977791e-01 1.89492315e-01 1.89874128e-01
6.02891803e-01 -9.92641628e-01 -5.43720245e-01 -3.75792623e-01
5.69275141e-01 -9.91746783e-02 3.70214462e-01 8.26537430e-01
-4.81373489e-01 5.90134621e-01 4.87323880e-01 -2.32003868e-01
-1.57435036e+00 5.99911869e-01 3.67985308e-01 -7.30576217e-01
-2.29768559e-01 9.98328984e-01 -3.76979917e-01 -8.02070558e-01
-1.22866184e-01 -1.07921422e-01 -4.46061015e-01 1.97418764e-01
2.79850036e-01 5.99389911e-01 5.69407344e-01 -3.92456204e-01
-2.81004161e-01 4.02995944e-01 -2.39335254e-01 -2.80215144e-02
1.09414399e+00 -1.20995920e-02 -5.59476376e-01 1.56313494e-01
1.48643601e+00 -7.30777383e-02 -3.24287772e-01 -6.42026424e-01
6.50179923e-01 -2.56489992e-01 1.67352766e-01 -1.20682085e+00
-6.85331404e-01 7.29031146e-01 1.60972953e-01 3.84140640e-01
9.15856004e-01 6.20028675e-01 1.18293130e+00 7.60241747e-01
5.42508364e-01 -1.24709129e+00 2.28718251e-01 7.14053154e-01
4.21735913e-01 -1.03391433e+00 -2.73679733e-01 -7.59628952e-01
-4.14530694e-01 6.60550773e-01 4.16001916e-01 3.94510061e-01
5.98702252e-01 -3.78483534e-01 -1.01646027e-02 -8.69840980e-01
-1.09182382e+00 -5.54920554e-01 5.57715058e-01 3.05162489e-01
1.27796933e-01 1.90180242e-02 -2.19716653e-01 4.19754922e-01
-3.59679610e-01 -1.82220146e-01 2.60667205e-01 9.38818872e-01
-7.78690338e-01 -1.25221956e+00 -1.86673045e-01 1.23402476e+00
-6.46883309e-01 -3.54980975e-01 -6.35724962e-01 1.25183642e-01
-4.83694545e-04 1.70554030e+00 -6.07562006e-01 -5.22732913e-01
2.51063764e-01 6.95097566e-01 6.55491948e-01 -6.00674748e-01
-8.08479488e-01 -6.45347059e-01 7.50809014e-01 -5.33187151e-01
-2.11016342e-01 -6.19587719e-01 -1.21485651e+00 -3.29903573e-01
-5.10525227e-01 5.17713070e-01 4.98892188e-01 1.21884322e+00
7.13200390e-01 2.50667751e-01 5.08212924e-01 4.49617356e-01
-6.04670703e-01 -9.41983998e-01 -3.19470018e-01 7.21386731e-01
-2.42053047e-02 -4.75980997e-01 -2.84376234e-01 -4.92218025e-02] | [10.507360458374023, 7.906509876251221] |
88290d66-1881-47c1-a62d-16caecc208a2 | diffsurv-differentiable-sorting-for-censored | 2304.13594 | null | https://arxiv.org/abs/2304.13594v1 | https://arxiv.org/pdf/2304.13594v1.pdf | Diffsurv: Differentiable sorting for censored time-to-event data | Survival analysis is a crucial semi-supervised task in machine learning with numerous real-world applications, particularly in healthcare. Currently, the most common approach to survival analysis is based on Cox's partial likelihood, which can be interpreted as a ranking model optimized on a lower bound of the concordance index. This relation between ranking models and Cox's partial likelihood considers only pairwise comparisons. Recent work has developed differentiable sorting methods which relax this pairwise independence assumption, enabling the ranking of sets of samples. However, current differentiable sorting methods cannot account for censoring, a key factor in many real-world datasets. To address this limitation, we propose a novel method called Diffsurv. We extend differentiable sorting methods to handle censored tasks by predicting matrices of possible permutations that take into account the label uncertainty introduced by censored samples. We contrast this approach with methods derived from partial likelihood and ranking losses. Our experiments show that Diffsurv outperforms established baselines in various simulated and real-world risk prediction scenarios. Additionally, we demonstrate the benefits of the algorithmic supervision enabled by Diffsurv by presenting a novel method for top-k risk prediction that outperforms current methods. | ['Spiros Denaxas', 'Roland Eils', 'Aylin Cakiroglu', 'Benjamin Wild', 'Andre Vauvelle'] | 2023-04-26 | null | null | null | null | ['survival-analysis'] | ['miscellaneous'] | [ 3.53903264e-01 7.04695135e-02 -6.87224150e-01 -9.58749831e-01
-1.14792764e+00 -3.45611393e-01 4.15352285e-01 5.70699871e-01
-4.50205833e-01 1.13183701e+00 3.91173869e-01 -4.65147465e-01
-7.24479139e-01 -6.47601008e-01 -5.56164980e-01 -6.59622729e-01
-4.79854614e-01 7.12710917e-01 -5.26461937e-02 4.92380746e-02
5.92959598e-02 3.85746241e-01 -1.29282928e+00 2.33928323e-01
9.78067279e-01 6.91058695e-01 -3.40970218e-01 2.62678862e-01
2.66539007e-01 6.24656916e-01 -2.78620750e-01 -5.13508618e-01
3.28842551e-02 -4.63289618e-01 -6.43215001e-01 -6.02899611e-01
7.41139948e-02 -3.48751724e-01 -3.28240432e-02 8.57292712e-01
4.31524783e-01 -1.15480898e-02 1.09317541e+00 -1.48805285e+00
-8.66209436e-03 7.82403708e-01 -2.59354234e-01 -1.02435328e-01
2.10874751e-01 -5.44188619e-01 1.04425442e+00 -4.99208003e-01
3.96000862e-01 1.15203595e+00 1.01011479e+00 4.12469000e-01
-1.45298159e+00 -5.33711910e-01 -1.45639926e-01 1.61497161e-01
-1.28952563e+00 -1.97861701e-01 5.32398880e-01 -5.49537420e-01
4.19565141e-01 6.97019994e-01 2.68602401e-01 8.66639495e-01
5.48413754e-01 7.72696137e-01 1.29475856e+00 -3.47484142e-01
4.79859442e-01 4.88785058e-02 6.43377364e-01 5.53577900e-01
3.81247610e-01 5.55785716e-01 -4.14336443e-01 -8.42651367e-01
3.72298479e-01 4.05878991e-01 -3.37943166e-01 -7.07586646e-01
-1.30540729e+00 9.57668066e-01 1.24289036e-01 -3.58738869e-01
-2.92226523e-02 -2.41202004e-02 5.05878389e-01 2.79559225e-01
4.63715732e-01 -5.85413463e-02 -6.58319592e-01 1.16649583e-01
-1.01304698e+00 3.00659128e-02 8.87268841e-01 7.00352490e-01
2.22526938e-01 -7.27655172e-01 -5.60104787e-01 6.95407093e-01
5.12265086e-01 1.81569472e-01 1.97045520e-01 -8.29903364e-01
2.02735052e-01 6.65573120e-01 1.19718783e-01 -3.16451937e-01
-7.13378727e-01 -6.99071050e-01 -1.20010638e+00 3.25416028e-01
7.02871382e-01 1.37185127e-01 -6.47385657e-01 1.92516363e+00
3.58059138e-01 2.99327105e-01 9.96102765e-02 5.67796588e-01
4.17756408e-01 1.48313686e-01 8.77646431e-02 -6.58189595e-01
1.27329981e+00 -6.46367311e-01 -6.54739618e-01 2.29569376e-01
1.00275946e+00 -1.85850456e-01 8.02692056e-01 6.07173979e-01
-6.51122928e-01 1.32664993e-01 -9.41028655e-01 4.28818353e-02
-1.03145286e-01 1.53751656e-01 7.68237412e-01 8.07648897e-01
-7.75214374e-01 1.00339007e+00 -8.89056802e-01 -2.06991330e-01
4.89904791e-01 5.07482767e-01 -4.20341372e-01 -9.55127403e-02
-1.29671109e+00 5.92680514e-01 1.95979878e-01 6.55513257e-02
-7.46517658e-01 -7.33461857e-01 -6.50123715e-01 1.19698666e-01
2.68965751e-01 -9.88067687e-01 9.90812480e-01 -5.44912696e-01
-1.12889886e+00 7.69823790e-01 -1.84448123e-01 -6.31991267e-01
9.22271311e-01 -2.13488355e-01 -1.21147521e-01 -2.46526971e-01
3.99296023e-02 1.13899931e-02 3.12936991e-01 -9.50231791e-01
-5.75575054e-01 -5.39684057e-01 -2.99904734e-01 -2.13826122e-03
-3.46224219e-01 5.06137386e-02 -1.25869468e-01 -7.66639948e-01
1.13275893e-01 -8.40537965e-01 -4.24068153e-01 -4.06744052e-03
-6.14647806e-01 -3.35155934e-01 1.34210046e-02 -5.81554294e-01
1.41333687e+00 -2.00705242e+00 3.70180756e-02 4.15302336e-01
-4.26834598e-02 -2.76010841e-01 2.16964364e-01 5.15667796e-01
-1.17154911e-01 -2.47441810e-02 -6.88356996e-01 -6.06070757e-01
1.54429367e-02 1.18230000e-01 -3.22302163e-01 5.75490236e-01
-3.28645915e-01 5.86544633e-01 -8.45327318e-01 -6.40157580e-01
-6.87107444e-02 4.04056847e-01 -4.79915023e-01 1.49235189e-01
2.10991368e-01 4.66112405e-01 -2.37186447e-01 5.66111982e-01
5.72158158e-01 -3.79661441e-01 4.97053593e-01 1.64868206e-01
2.24729732e-01 7.93259069e-02 -9.38207626e-01 1.33496714e+00
-1.19566739e-01 1.64462719e-02 -4.26904082e-01 -1.13130188e+00
7.83955216e-01 1.49560869e-01 4.17757899e-01 2.46077199e-02
1.03988171e-01 3.82136434e-01 -3.46854538e-01 -8.85995552e-02
-1.23679750e-01 -5.09004831e-01 -3.00718904e-01 2.77759373e-01
-3.07755560e-01 4.02663976e-01 -1.80060908e-01 1.94288999e-01
1.24175298e+00 3.22023337e-03 7.44702876e-01 -3.89411867e-01
5.03737152e-01 -1.72182217e-01 9.11256969e-01 1.03771782e+00
-2.62301356e-01 8.22714567e-01 9.45330381e-01 -3.61169934e-01
-4.14265603e-01 -1.31883609e+00 -8.70779097e-01 7.54707813e-01
-3.24964821e-02 -1.84296921e-01 -5.05270600e-01 -1.18190134e+00
3.11928868e-01 8.59837770e-01 -8.74395370e-01 -3.17178816e-01
-1.32980540e-01 -1.42671275e+00 5.13747692e-01 4.66509044e-01
-9.56304669e-02 -5.73687971e-01 -9.39197987e-02 6.52141273e-02
-1.34629741e-01 -4.40819770e-01 -4.83453214e-01 4.08560306e-01
-1.35534358e+00 -1.35965490e+00 -7.24310279e-01 -5.83253384e-01
7.19083071e-01 -2.51553714e-01 1.10397923e+00 -1.15826707e-02
-7.15976506e-02 1.61269121e-02 -2.40216583e-01 -1.71143070e-01
-4.76525903e-01 -1.08086197e-02 1.10091619e-01 1.85266733e-02
3.21978539e-01 -5.00404000e-01 -6.92738414e-01 5.19052684e-01
-7.86736548e-01 2.10857391e-02 5.94618380e-01 1.21243811e+00
9.06823933e-01 -3.23556773e-02 1.04152286e+00 -1.34409618e+00
1.53878793e-01 -7.44238138e-01 -6.16861284e-01 5.75160503e-01
-1.03014648e+00 2.67081469e-01 4.58101243e-01 -1.54141262e-01
-8.52736294e-01 2.13000491e-01 4.46290895e-02 2.79093701e-02
-1.09928435e-04 7.49902725e-01 -2.69496351e-01 3.15688401e-01
3.04236948e-01 2.94684004e-02 8.52800012e-02 -5.99059880e-01
-7.28455558e-02 7.31303513e-01 3.73222113e-01 -3.12518179e-01
4.27300751e-01 7.71643817e-01 6.09438658e-01 -2.86197394e-01
-1.01378489e+00 -4.29940939e-01 -6.00614011e-01 1.10742167e-01
6.01057410e-01 -7.28509426e-01 -9.51965094e-01 4.87850904e-01
-6.82219923e-01 -3.12390566e-01 -1.86907113e-01 8.29446971e-01
-6.63705111e-01 5.46130419e-01 -5.56025684e-01 -1.01367176e+00
-2.83872336e-01 -8.71231914e-01 1.07543182e+00 -2.66231567e-01
-1.10489301e-01 -1.24473917e+00 2.39116520e-01 3.97579849e-01
2.83361841e-02 5.47920823e-01 1.23038900e+00 -1.07926893e+00
-2.60844231e-01 -5.52830875e-01 -1.02885723e-01 2.27889642e-01
1.33859649e-01 -4.75628287e-01 -7.38482594e-01 -4.53750402e-01
-1.57800049e-01 -9.44880918e-02 1.24122965e+00 6.82238519e-01
1.29910254e+00 -1.40828639e-01 -6.38503373e-01 7.36394107e-01
1.24953628e+00 4.69416231e-02 4.82929826e-01 2.84786135e-01
4.01080161e-01 9.59328473e-01 8.97179306e-01 5.90548038e-01
5.35472631e-01 7.61377394e-01 4.57934439e-01 7.30146933e-03
2.19133213e-01 -4.19752896e-01 2.50580817e-01 4.96636987e-01
2.97202826e-01 -1.68090388e-01 -9.01340663e-01 3.35509717e-01
-2.07306314e+00 -6.91588104e-01 -4.55443025e-01 2.94387650e+00
8.92796397e-01 2.22327441e-01 9.42103639e-02 3.01728427e-01
6.80916846e-01 -4.32107091e-01 -5.85689723e-01 3.38221304e-02
-1.49998605e-01 -1.67246193e-01 6.63371384e-01 5.52300155e-01
-1.09311569e+00 9.76091921e-02 6.73153877e+00 7.19887316e-01
-3.82937461e-01 -5.32246456e-02 9.51666117e-01 -8.64363685e-02
-4.01732713e-01 2.37602621e-01 -7.66946733e-01 5.16566098e-01
7.78110266e-01 -1.27599999e-01 -2.89967246e-02 6.15843832e-01
2.51587808e-01 -4.05484959e-02 -1.61955619e+00 6.12689912e-01
-1.71895735e-02 -8.04792225e-01 -9.51306596e-02 2.57185161e-01
5.43987215e-01 -2.14145631e-01 5.25934063e-03 2.73158073e-01
4.12089974e-01 -1.07278621e+00 4.16963845e-01 8.95369053e-01
9.69258904e-01 -7.60967970e-01 1.14496160e+00 4.80700076e-01
-6.40179038e-01 -1.26184121e-01 -3.74937691e-02 1.32719830e-01
4.11773652e-01 1.23006320e+00 -7.47267783e-01 7.84863114e-01
5.05063355e-01 7.99110413e-01 -4.95005906e-01 1.18333697e+00
-9.36097577e-02 9.04757261e-01 -2.67154366e-01 2.13117540e-01
-5.12162149e-01 -2.99020708e-01 4.37740892e-01 8.84307086e-01
5.41762888e-01 -2.40829075e-03 -9.50082392e-02 2.78052509e-01
2.90354826e-02 1.71781555e-01 -3.13278168e-01 4.73406911e-01
3.56154203e-01 7.50464022e-01 -6.96039677e-01 -2.37636656e-01
-3.28645259e-01 8.04359436e-01 1.48957849e-01 6.84596375e-02
-7.35622644e-01 -2.37595856e-01 4.37536061e-01 1.40804246e-01
-1.19165964e-01 2.19067454e-01 -4.49735433e-01 -1.16878891e+00
-6.36398047e-02 -5.55441380e-01 1.12196004e+00 -1.94826812e-01
-1.66759419e+00 3.10360879e-01 2.47597903e-01 -1.41160893e+00
-1.50773525e-01 -4.99210000e-01 -3.77503008e-01 6.57062769e-01
-1.56264639e+00 -9.89329040e-01 -6.27012597e-03 3.48414838e-01
1.54587731e-01 2.92237047e-02 8.74223948e-01 1.93849728e-01
-7.62580633e-01 8.74471843e-01 6.96001649e-01 -2.82618433e-01
1.16500485e+00 -1.55330122e+00 4.58841911e-03 3.91068369e-01
-2.10293606e-01 5.14743686e-01 5.65958261e-01 -8.35332155e-01
-9.94327009e-01 -1.13681054e+00 1.29493558e+00 -6.41497552e-01
3.36200953e-01 -3.16509813e-01 -9.70582485e-01 6.45954907e-01
-6.93817616e-01 1.30101085e-01 1.14666784e+00 5.09591162e-01
-2.75392979e-01 -3.20778400e-01 -1.24558210e+00 2.82966495e-01
1.16456974e+00 -1.67471781e-01 -5.08426607e-01 4.17175591e-01
4.24642175e-01 -6.10620901e-02 -9.66377497e-01 6.92358255e-01
7.92652249e-01 -1.08179224e+00 1.01870358e+00 -7.92137384e-01
3.74884307e-01 -2.34308422e-01 -1.31095916e-01 -1.02567494e+00
-2.12880567e-01 -4.27664518e-01 -1.69031173e-02 1.14525485e+00
3.86606723e-01 -9.57238019e-01 8.44747484e-01 6.23109281e-01
3.26392762e-02 -8.20969939e-01 -1.27965283e+00 -1.03388882e+00
2.17751041e-01 -3.46047163e-01 6.18986189e-01 9.38900828e-01
2.27575079e-01 -1.69567868e-01 -5.74152350e-01 2.72258013e-01
1.22864270e+00 5.54248914e-02 4.19761211e-01 -1.74274755e+00
-5.09637475e-01 -5.86581863e-02 -4.14728940e-01 -5.72377145e-01
2.67600715e-01 -1.17449033e+00 -1.11296348e-01 -1.45422149e+00
8.33434641e-01 -6.32972479e-01 -8.57765973e-01 4.54059869e-01
-2.87080109e-01 9.38884094e-02 -3.56626123e-01 4.59880382e-01
-5.01958251e-01 5.24918854e-01 5.85360706e-01 1.58600539e-01
-1.79174230e-01 5.00902951e-01 -6.35576665e-01 7.29591429e-01
7.58184254e-01 -8.01560581e-01 -3.99091065e-01 1.50096461e-01
4.69911128e-01 5.48320353e-01 4.64605778e-01 -7.42376506e-01
1.14433177e-01 -2.22463384e-01 1.47525772e-01 -6.06266916e-01
-1.02151781e-01 -8.18060160e-01 5.28523386e-01 7.53396511e-01
-6.23226821e-01 -2.80114263e-01 -2.61346310e-01 1.01920998e+00
-1.73998863e-01 -1.45232812e-01 5.75994253e-01 5.20589948e-01
2.68825293e-02 1.73645005e-01 -3.81985269e-02 -4.54340018e-02
1.09942532e+00 -3.00216470e-02 -2.13518247e-01 -3.07929665e-01
-8.14277232e-01 5.85534871e-01 4.30589348e-01 8.57118666e-02
5.89925587e-01 -1.42077374e+00 -9.45049703e-01 1.68859139e-01
3.04482281e-01 -9.93712693e-02 2.97450572e-01 1.26943302e+00
-1.48805022e-01 3.29633325e-01 2.32546240e-01 -3.85268092e-01
-1.49487352e+00 6.89277232e-01 2.60253638e-01 -7.47294724e-01
-4.15757149e-01 4.52221215e-01 4.99973893e-01 -4.82537359e-01
3.46842736e-01 -1.74986184e-01 -3.84804904e-01 9.62989926e-02
4.21711475e-01 7.29168534e-01 1.12994686e-01 -3.58153045e-01
-6.29754543e-01 3.46887976e-01 -7.60642588e-02 -1.35484770e-01
1.25586867e+00 -1.59056604e-01 -1.91807553e-01 6.48081243e-01
1.27584493e+00 -7.01087788e-02 -1.19835448e+00 -1.41327351e-01
3.39358658e-01 -3.20716739e-01 -5.94790317e-02 -1.09748709e+00
-4.99309838e-01 6.12975001e-01 6.76214099e-01 -9.38866585e-02
1.23210359e+00 1.93852428e-02 4.76835966e-01 1.26355797e-01
4.96849835e-01 -6.50620580e-01 -4.69874293e-01 5.61616085e-02
7.31177509e-01 -1.45649195e+00 1.24418456e-02 -5.91093123e-01
-5.04270494e-01 8.15026641e-01 -9.84508246e-02 2.31899843e-01
9.28715646e-01 -5.54349693e-03 1.00520477e-01 6.41451180e-02
-7.50513673e-01 2.44233578e-01 4.89665359e-01 4.41321135e-01
4.64406073e-01 2.56223321e-01 -8.00314367e-01 1.09825230e+00
3.62059567e-03 3.21526945e-01 3.28053832e-01 6.75306857e-01
-1.12466171e-01 -1.42049062e+00 -4.49709922e-01 1.00000179e+00
-7.38382518e-01 -1.06232464e-01 -9.85850543e-02 3.99891675e-01
-3.17571640e-01 8.73138607e-01 -1.29602447e-01 -6.91562369e-02
3.15974891e-01 9.91438702e-02 8.13761875e-02 -4.19131249e-01
-3.42061698e-01 -2.32810751e-01 5.38372435e-02 -3.53510290e-01
-2.58560956e-01 -1.12818456e+00 -1.18947983e+00 5.34851942e-03
-5.39956987e-01 3.50212187e-01 4.12038356e-01 8.21529627e-01
2.07145855e-01 3.63725245e-01 8.31967473e-01 -2.69123018e-01
-1.19618237e+00 -5.49724698e-01 -9.10327911e-01 4.04259026e-01
4.89791930e-01 -7.42441833e-01 -8.34360957e-01 -9.77895111e-02] | [7.773879051208496, 5.463418483734131] |
8b581b38-1a58-466f-a438-73b18fc8157f | a-close-look-at-deep-learning-with-small-data | 2003.12843 | null | https://arxiv.org/abs/2003.12843v3 | https://arxiv.org/pdf/2003.12843v3.pdf | A Close Look at Deep Learning with Small Data | In this work, we perform a wide variety of experiments with different deep learning architectures on datasets of limited size. According to our study, we show that model complexity is a critical factor when only a few samples per class are available. Differently from the literature, we show that in some configurations, the state of the art can be improved using low complexity models. For instance, in problems with scarce training samples and without data augmentation, low-complexity convolutional neural networks perform comparably well or better than state-of-the-art architectures. Moreover, we show that even standard data augmentation can boost recognition performance by large margins. This result suggests the development of more complex data generation/augmentation pipelines for cases when data is limited. Finally, we show that dropout, a widely used regularization technique, maintains its role as a good regularizer even when data is scarce. Our findings are empirically validated on the sub-sampled versions of popular CIFAR-10, Fashion-MNIST and, SVHN benchmarks. | ['L. Iocchi', 'L. Brigato'] | 2020-03-28 | null | null | null | null | ['small-data'] | ['computer-vision'] | [ 1.32106438e-01 -8.17630067e-02 -4.75620598e-01 -4.23916042e-01
-6.44881129e-01 -2.98609078e-01 6.00092232e-01 1.98979110e-01
-9.49267387e-01 9.46542799e-01 -6.45185113e-02 -1.94315180e-01
-5.68253733e-02 -5.79847038e-01 -1.05119145e+00 -5.93136907e-01
-5.51817343e-02 4.76837248e-01 -2.50296351e-02 -2.77611911e-01
1.22241853e-02 6.21990860e-01 -1.62497914e+00 2.34591872e-01
5.40756047e-01 1.16765964e+00 -1.81193084e-01 6.06378853e-01
2.06582472e-02 8.91656399e-01 -5.47276139e-01 -3.88512820e-01
5.50688863e-01 -1.21449955e-01 -8.19142103e-01 1.66862085e-01
7.08605111e-01 -2.79128283e-01 -3.36311668e-01 7.25026667e-01
4.99096900e-01 9.68916938e-02 5.22156239e-01 -1.04352701e+00
-6.53534591e-01 8.87654841e-01 -2.84810275e-01 3.49714071e-01
-3.68783176e-01 3.46055955e-01 1.01494765e+00 -9.13388252e-01
5.67050159e-01 9.51740026e-01 8.38203192e-01 6.06363118e-01
-1.65326226e+00 -7.02378392e-01 4.72054243e-01 7.40003660e-02
-1.28560793e+00 -4.75686222e-01 8.09205830e-01 -2.70066977e-01
9.59466636e-01 2.09798157e-01 3.34353983e-01 1.38781512e+00
-2.64567077e-01 9.38602090e-01 1.16297209e+00 -5.87019742e-01
2.19636977e-01 3.50710124e-01 4.89583910e-01 4.77176338e-01
5.02592266e-01 -1.42733548e-02 -2.94517964e-01 3.33476737e-02
5.90716481e-01 -9.35143605e-02 -2.09270805e-01 -3.04897130e-01
-9.87922907e-01 9.69563782e-01 5.26475787e-01 6.63000584e-01
-4.03950751e-01 1.96344048e-01 6.40309989e-01 7.42104888e-01
6.13861620e-01 7.24933565e-01 -6.82773829e-01 8.26867018e-03
-1.08105457e+00 2.83134282e-01 9.44641411e-01 7.35065341e-01
7.00486124e-01 3.92601907e-01 -2.52342701e-01 8.88580501e-01
-1.94840208e-01 2.53947992e-02 6.21306419e-01 -8.42118204e-01
4.99033183e-01 5.80454767e-01 5.22846403e-03 -5.70465028e-01
-4.50895935e-01 -1.17313170e+00 -1.24251568e+00 2.21432716e-01
8.79570246e-01 -1.69620290e-01 -9.08838272e-01 1.85816765e+00
-1.51937172e-01 1.98184922e-01 -1.13381976e-02 7.91714311e-01
5.91129661e-01 3.32517177e-01 5.14826365e-02 -9.55805648e-03
1.07552648e+00 -1.07424963e+00 -4.97980595e-01 -5.39204359e-01
7.81386256e-01 -5.87038755e-01 1.51161075e+00 7.38202095e-01
-1.02471507e+00 -7.45789945e-01 -1.06636012e+00 -6.38899282e-02
-5.26451051e-01 4.83161271e-01 9.47377503e-01 7.59796441e-01
-1.00371861e+00 8.56100559e-01 -7.72777617e-01 -2.63155848e-01
8.10145617e-01 4.27843958e-01 -3.79567862e-01 -3.15648735e-01
-1.04402328e+00 7.74711072e-01 3.37300628e-01 2.43326306e-01
-1.04284513e+00 -8.48025024e-01 -5.17209172e-01 1.60764173e-01
2.86270440e-01 -4.08658862e-01 1.30646372e+00 -1.28636432e+00
-1.35881054e+00 9.89579737e-01 1.85816780e-01 -1.13272417e+00
7.67352581e-01 -4.38617706e-01 -1.44914970e-01 -2.27191865e-01
-4.61798102e-01 6.43461943e-01 6.05171442e-01 -1.16267550e+00
-4.52812016e-01 -2.00338423e-01 3.61394763e-01 -2.43340269e-01
-5.60719311e-01 -3.37176174e-02 -2.17773318e-01 -5.81965268e-01
-4.34765548e-01 -8.82057846e-01 -4.59722042e-01 -7.13322461e-02
-3.82502675e-01 -1.69167861e-01 4.98081774e-01 -3.54617268e-01
1.15130460e+00 -2.00268507e+00 2.92822085e-02 1.46237597e-01
1.11152366e-01 5.63726425e-01 -3.27857226e-01 1.83324948e-01
-2.45354980e-01 3.09289575e-01 -2.24753320e-01 -6.38062298e-01
2.72418894e-02 4.37174499e-01 -6.52842224e-02 4.97758389e-01
3.91316742e-01 8.04339588e-01 -4.56831723e-01 1.24743916e-01
1.84117287e-01 4.38139886e-01 -6.63632691e-01 6.09256625e-02
-2.16504991e-01 2.92214543e-01 -7.94158578e-02 3.83945793e-01
5.19394279e-01 -5.58685005e-01 -1.77731663e-02 -5.30842207e-02
7.46560320e-02 2.79522032e-01 -1.21260524e+00 1.69960821e+00
-6.37205362e-01 9.16285396e-01 9.78218466e-02 -1.48908448e+00
8.25113356e-01 1.09638557e-01 5.25949858e-02 -7.59323418e-01
2.71627456e-01 3.12337309e-01 4.07471687e-01 -1.61577180e-01
3.53909940e-01 1.35825917e-01 3.43915820e-01 2.71067739e-01
2.26199031e-01 3.20060313e-01 6.22309387e-01 -1.39900282e-01
1.09985864e+00 -2.85113126e-01 1.09356381e-01 -3.49758238e-01
4.58937228e-01 -5.41192889e-02 3.48020434e-01 1.03028357e+00
3.92120518e-02 6.18873894e-01 7.31973171e-01 -6.51387632e-01
-1.42604840e+00 -5.40519238e-01 -3.30525696e-01 1.35680890e+00
-6.80824041e-01 -1.85867012e-01 -8.78725767e-01 -4.94824409e-01
-1.07769802e-01 5.51519632e-01 -8.38145494e-01 -1.65768772e-01
-7.81684637e-01 -1.30238497e+00 5.79571545e-01 6.63170397e-01
5.33072770e-01 -1.02174819e+00 -3.45826417e-01 2.14708343e-01
3.20088267e-01 -1.28844190e+00 1.22563712e-01 5.62835336e-01
-1.08218348e+00 -1.02291369e+00 -8.32449973e-01 -7.10690200e-01
6.86436117e-01 -3.15216370e-02 1.48407531e+00 3.09813976e-01
-1.71773687e-01 2.11644415e-02 -4.04905260e-01 -3.28120291e-01
-3.84620816e-01 7.25671589e-01 -1.19520657e-01 -3.62083428e-02
2.90143490e-01 -6.20518625e-01 -5.74164689e-01 1.45510748e-01
-9.76718128e-01 2.11128369e-02 7.94407964e-01 1.07501376e+00
4.60624307e-01 -9.13287178e-02 7.21512914e-01 -1.28870356e+00
6.66247606e-01 -4.78429198e-01 -6.20381236e-01 -8.36764798e-02
-7.11397707e-01 1.94426477e-01 1.07865524e+00 -8.01596403e-01
-6.15691185e-01 -6.17891401e-02 -2.34877884e-01 -4.04017448e-01
-4.86509830e-01 5.58998585e-01 9.20379758e-02 -6.06636330e-02
1.04488409e+00 1.03865594e-01 -1.42365946e-02 -7.69397259e-01
2.87945598e-01 3.81870747e-01 7.39121735e-02 -5.40623784e-01
5.19664943e-01 5.12747288e-01 -1.44552529e-01 -7.75839448e-01
-9.16792691e-01 -8.29753652e-02 -6.38397217e-01 3.03513050e-01
4.41823214e-01 -9.66118038e-01 -4.86854404e-01 6.17699742e-01
-8.40429842e-01 -8.36987555e-01 -5.97496688e-01 5.22661328e-01
-2.71223634e-01 -3.70293222e-02 -8.89954567e-01 -7.52997279e-01
-2.45471269e-01 -1.16891491e+00 5.37484765e-01 1.51844114e-01
1.70171782e-02 -9.49910700e-01 -1.52105823e-01 1.31838784e-01
5.62192917e-01 1.64855912e-01 7.90176570e-01 -1.15750468e+00
-5.00126660e-01 -3.48895073e-01 -1.79063424e-01 8.70857239e-01
-6.27390444e-02 -1.09407276e-01 -1.29547286e+00 -4.85326827e-01
-2.05886617e-01 -6.96836412e-01 1.18271732e+00 4.11125332e-01
1.55337858e+00 -2.58250624e-01 1.32225469e-01 7.53974378e-01
1.40053582e+00 -1.72965929e-01 5.73299766e-01 5.36706686e-01
6.71080768e-01 5.45928419e-01 2.18836352e-01 2.88360357e-01
1.20949633e-01 4.86339867e-01 3.99354458e-01 -4.49175835e-01
-1.16095431e-01 1.35586590e-01 1.05394363e-01 6.58718348e-01
-1.77748501e-01 -2.12430522e-01 -9.33384657e-01 4.76301163e-01
-1.73083162e+00 -6.90311551e-01 -1.51383445e-01 2.14822245e+00
8.80495965e-01 4.34276849e-01 2.54029155e-01 5.15705705e-01
3.10630888e-01 1.50909767e-01 -4.85229611e-01 -2.44370639e-01
-4.02774543e-01 4.98745292e-01 7.40422845e-01 3.06376159e-01
-1.13778853e+00 8.43128145e-01 6.58134842e+00 9.46948349e-01
-1.24240494e+00 1.90008894e-01 1.17288733e+00 -3.34306687e-01
3.77822667e-02 -3.74253780e-01 -7.96637177e-01 4.08039570e-01
1.26236916e+00 2.27840349e-01 4.68230754e-01 1.07133424e+00
9.20464769e-02 6.15893789e-02 -1.47644258e+00 9.79508221e-01
6.66977884e-03 -1.50532091e+00 -2.35061988e-01 4.00257967e-02
7.38334000e-01 4.08483535e-01 2.08165124e-01 5.98306835e-01
4.20918763e-01 -1.29894686e+00 5.49973011e-01 2.96952099e-01
4.91305470e-01 -7.95937061e-01 9.52761054e-01 3.97275567e-01
-7.20061064e-01 -2.04966530e-01 -5.92742443e-01 -2.31648609e-01
-2.39232540e-01 8.72267783e-01 -5.67921579e-01 3.71357411e-01
6.05264962e-01 5.80058992e-01 -9.25401807e-01 1.06187928e+00
-1.04209892e-01 1.00271368e+00 -4.58907396e-01 8.52477849e-02
4.84225750e-01 6.09463081e-02 -5.79506047e-02 1.47869241e+00
-2.39227936e-02 -2.84903616e-01 6.11087792e-02 6.92732334e-01
-5.12534738e-01 1.28738165e-01 -5.22278786e-01 6.68697730e-02
1.68784380e-01 1.21816790e+00 -6.26786172e-01 -4.64409024e-01
-5.18580019e-01 5.90348899e-01 6.92589045e-01 5.01558602e-01
-6.19343519e-01 -9.95343924e-02 5.58279872e-01 1.96310848e-01
3.38330299e-01 -1.57965049e-01 -4.26085770e-01 -1.22590292e+00
5.77707589e-02 -1.21819258e+00 2.95410722e-01 -2.67456651e-01
-1.35884190e+00 7.65545845e-01 -2.32198477e-01 -9.70017314e-01
-2.72275448e-01 -8.89777601e-01 -4.19074863e-01 8.05401742e-01
-1.83380759e+00 -9.37296629e-01 -1.44876420e-01 4.56434220e-01
6.20518625e-01 -3.56828094e-01 7.92094290e-01 6.92105949e-01
-8.32697332e-01 8.14078748e-01 3.04269135e-01 3.91561478e-01
3.70692641e-01 -1.09240782e+00 4.75313544e-01 7.73910820e-01
3.20376337e-01 5.23383617e-01 6.78413212e-01 -1.28805429e-01
-1.17790294e+00 -1.03010726e+00 5.58138549e-01 -1.94921449e-01
7.13993430e-01 -7.85187840e-01 -1.28620291e+00 7.22919941e-01
2.43101746e-01 4.58157063e-01 6.06098533e-01 3.99476349e-01
-4.22168911e-01 -2.62369901e-01 -9.84887302e-01 4.45703059e-01
1.00607228e+00 -3.02740693e-01 -2.75186777e-01 3.35487902e-01
4.53863919e-01 -3.44127923e-01 -7.14442492e-01 3.50265622e-01
4.55322802e-01 -9.76752102e-01 9.56149697e-01 -1.00112164e+00
4.90676910e-01 3.06326717e-01 -2.28235438e-01 -1.33189452e+00
-2.11994514e-01 -3.88146162e-01 -1.81593299e-01 1.27394295e+00
6.27628744e-01 -5.34192145e-01 1.12828827e+00 6.37348950e-01
-2.15772595e-02 -8.00778210e-01 -8.32611442e-01 -9.95637417e-01
4.65188116e-01 -6.24826431e-01 3.18300128e-01 1.04007304e+00
-7.31880128e-01 2.58461595e-01 -4.60167170e-01 -2.76405275e-01
4.32484806e-01 -2.48729646e-01 8.20748031e-01 -1.42551398e+00
-3.17026109e-01 -5.85857153e-01 -1.86254412e-01 -8.86165500e-01
4.47190762e-01 -9.20390666e-01 -3.43338102e-01 -1.08906579e+00
1.26267709e-02 -7.26027906e-01 -6.00469172e-01 5.69339275e-01
8.92452225e-02 3.69318604e-01 2.72329926e-01 1.06007777e-01
-3.20633918e-01 3.05408329e-01 1.03217268e+00 -1.20159879e-01
-1.97500095e-01 7.86185116e-02 -6.87721908e-01 7.63052166e-01
1.06160378e+00 -5.06643891e-01 -3.74997795e-01 -5.23103476e-01
2.57553250e-01 -5.91458976e-01 3.72021258e-01 -1.14494026e+00
3.82029545e-03 8.63310918e-02 4.05812174e-01 -1.91361189e-01
3.11454952e-01 -9.59453344e-01 -1.47269726e-01 5.49992383e-01
-7.68753350e-01 -4.64904942e-02 3.65660399e-01 2.48792037e-01
-1.99168354e-01 -4.03067797e-01 8.86847138e-01 -1.81010768e-01
-4.38641310e-01 2.41003096e-01 -3.08613688e-01 2.78281391e-01
6.56770468e-01 -1.03731658e-02 -2.91437685e-01 -1.70505300e-01
-6.72747195e-01 5.14310263e-02 2.02658847e-01 3.71960491e-01
2.26407722e-01 -1.31490397e+00 -8.63404989e-01 2.96012998e-01
-3.61858532e-02 3.63388918e-02 1.69354275e-01 8.99720848e-01
-4.54350621e-01 4.55253512e-01 -1.24410316e-01 -5.94880641e-01
-1.04316831e+00 4.67865258e-01 3.14396918e-01 -5.60929418e-01
-5.60593605e-01 7.79166400e-01 5.04397601e-02 -3.88448328e-01
5.61133146e-01 -8.53300035e-01 -1.45553678e-01 9.43478122e-02
5.76125324e-01 3.01199406e-01 4.38515604e-01 -2.08250254e-01
-9.50353295e-02 1.60591915e-01 -3.69817615e-01 2.45629340e-01
1.69614041e+00 2.98429519e-01 2.70129710e-01 5.79730630e-01
1.16864359e+00 -2.86849439e-01 -1.44562995e+00 -3.87071669e-01
5.09868562e-02 -2.89124340e-01 1.66575179e-01 -7.14402974e-01
-1.36015379e+00 1.05069327e+00 6.72631264e-01 3.30287367e-01
1.12148118e+00 -2.40095079e-01 3.35327476e-01 7.28875339e-01
1.41363472e-01 -1.15960824e+00 -6.05631433e-02 6.16759598e-01
8.50798190e-01 -1.45168042e+00 -9.48455334e-02 -5.06887995e-02
-5.07895291e-01 8.37042332e-01 6.60942018e-01 -5.13558388e-01
5.61877668e-01 3.89923990e-01 -9.37784016e-02 1.15318649e-01
-1.00317585e+00 -2.13511512e-01 2.37340659e-01 5.57915509e-01
6.47882581e-01 -2.09209174e-01 -2.08160624e-01 6.04010284e-01
-2.03408867e-01 8.31075832e-02 5.32651246e-01 6.94277227e-01
-3.13228995e-01 -1.14063966e+00 -1.62465394e-01 6.55071735e-01
-8.72483373e-01 -3.12001705e-01 -2.33058065e-01 1.35332394e+00
-1.43718896e-02 6.57701075e-01 1.68022349e-01 -7.39440098e-02
4.12471592e-01 1.59168437e-01 4.67090875e-01 -5.83289564e-01
-1.09972417e+00 -1.85009405e-01 1.91991001e-01 -5.23226321e-01
-5.66447318e-01 -2.07912385e-01 -7.29290605e-01 -4.01919216e-01
-1.24790415e-01 -5.85284568e-02 6.72350109e-01 9.09137249e-01
2.00134769e-01 6.82579696e-01 3.03975612e-01 -8.01525235e-01
-7.09126234e-01 -1.15600765e+00 -4.83688951e-01 5.54467082e-01
5.42112470e-01 -4.43789333e-01 -6.25003397e-01 5.27884299e-03] | [8.978752136230469, 3.2660391330718994] |
0594c29e-c25a-4459-a047-96051d1a4ab3 | gaitgs-temporal-feature-learning-in | 2305.19700 | null | https://arxiv.org/abs/2305.19700v2 | https://arxiv.org/pdf/2305.19700v2.pdf | GaitGS: Temporal Feature Learning in Granularity and Span Dimension for Gait Recognition | Gait recognition is an emerging biological recognition technology that identifies and verifies individuals based on their walking patterns. However, many current methods are limited in their use of temporal information. In order to fully harness the potential of gait recognition, it is crucial to consider temporal features at various granularities and spans. Hence, in this paper, we propose a novel framework named GaitGS, which aggregates temporal features in the granularity dimension and span dimension simultaneously. Specifically, Multi-Granularity Feature Extractor (MGFE) is proposed to focus on capturing the micro-motion and macro-motion information at the frame level and unit level respectively. Moreover, we present Multi-Span Feature Learning (MSFL) module to generate global and local temporal representations. On three popular gait datasets, extensive experiments demonstrate the state-of-the-art performance of our method. Our method achieves the Rank-1 accuracies of 92.9% (+0.5%), 52.0% (+1.4%), and 97.5% (+0.8%) on CASIA-B, GREW, and OU-MVLP respectively. The source code will be released soon. | ['Bin Feng', 'Wenyu Liu', 'Xinggang Wang', 'Xiaohu Huang', 'Yunze Deng', 'Haijun Xiong'] | 2023-05-31 | null | null | null | null | ['gait-recognition'] | ['computer-vision'] | [ 4.88202535e-02 -7.47191489e-01 -2.79712915e-01 -1.44768670e-01
-6.26036108e-01 -1.85589522e-01 4.65261310e-01 2.82718763e-02
-1.77656591e-01 9.65381920e-01 2.16289982e-01 3.96359652e-01
-2.02831700e-01 -7.87318707e-01 -7.60862157e-02 -8.01786005e-01
-5.56832016e-01 -1.84317634e-01 4.58266884e-01 -6.91047609e-02
3.13528031e-01 3.30138713e-01 -1.69683194e+00 3.06588918e-01
6.82970583e-01 1.09079635e+00 -1.21560045e-01 8.05222988e-01
2.70498067e-01 4.58749622e-01 -4.52411413e-01 -1.70528546e-01
-9.14357379e-02 -4.23815966e-01 -4.24472153e-01 6.75492883e-02
6.32424504e-02 -2.50408411e-01 -4.72929627e-01 6.11169696e-01
5.60678065e-01 1.08419232e-01 6.36470497e-01 -1.31481266e+00
-3.67955357e-01 -9.36061610e-03 -8.32515538e-01 5.85167825e-01
6.12542212e-01 4.55548376e-01 8.52531493e-01 -7.45947778e-01
3.95476848e-01 1.03138876e+00 7.41074800e-01 3.38249713e-01
-9.00598228e-01 -5.40357471e-01 6.60374239e-02 5.12527168e-01
-1.29989648e+00 -4.60776180e-01 7.27397442e-01 -5.53246319e-01
9.47915435e-01 2.02675417e-01 7.92364061e-01 1.16082776e+00
5.37896872e-01 7.73070633e-01 1.22579181e+00 -1.64910674e-01
2.83471290e-02 -8.84834468e-01 2.41777435e-01 7.81252384e-01
4.15925175e-01 1.97429046e-01 -9.58835065e-01 -9.87203047e-03
8.10098350e-01 1.81955025e-02 -1.99911445e-01 8.61814544e-02
-1.40113747e+00 3.86620790e-01 1.35642126e-01 3.16926479e-01
-4.77826923e-01 1.94269806e-01 5.93553603e-01 5.09290434e-02
2.52828300e-01 -1.64592847e-01 -1.59928203e-01 -6.44724786e-01
-1.02406061e+00 3.21888983e-01 3.93625587e-01 5.50129473e-01
4.48270112e-01 1.41335055e-01 -3.49429727e-01 7.42394567e-01
2.59223640e-01 6.24333382e-01 8.49584818e-01 -6.42339885e-01
3.52119833e-01 6.44389510e-01 1.00060575e-01 -1.38197970e+00
-5.97191811e-01 -4.61937070e-01 -1.06993556e+00 -2.94080198e-01
1.84289515e-01 7.90532157e-02 -7.82212019e-01 1.65364182e+00
3.94989461e-01 8.44122648e-01 -4.41540312e-03 8.81759524e-01
6.88474298e-01 5.14556706e-01 1.59629375e-01 -4.05426919e-01
1.46861756e+00 -5.74349701e-01 -5.57134092e-01 -1.33333728e-01
5.38272619e-01 -5.53944707e-01 7.84873426e-01 2.42903456e-01
-8.92595589e-01 -7.01299489e-01 -1.24177158e+00 3.56516331e-01
-2.16070384e-01 3.92044574e-01 5.25316954e-01 5.13880253e-01
-6.21874690e-01 7.68450439e-01 -1.27524006e+00 -5.88023543e-01
2.02249512e-01 3.65363240e-01 -4.93776679e-01 -2.42365319e-02
-1.19269109e+00 4.24722761e-01 2.67587572e-01 1.96834952e-01
-5.47596216e-01 -3.31769615e-01 -7.30237544e-01 -1.46298394e-01
-5.71769802e-03 -4.77478385e-01 6.39802635e-01 -2.50139460e-02
-1.21805513e+00 5.80853343e-01 -2.80276179e-01 -4.78580415e-01
4.42738086e-01 -2.66862184e-01 -7.01554239e-01 3.53245378e-01
3.44609737e-01 2.03440726e-01 6.24203026e-01 -6.53235376e-01
-7.95874059e-01 -6.00561023e-01 -4.99424875e-01 -9.05085728e-02
-3.41237605e-01 -1.62191689e-01 -4.45425928e-01 -9.71472919e-01
1.23170979e-01 -8.51379275e-01 1.77360669e-01 -2.85243899e-01
-3.17614406e-01 -1.55198365e-01 8.42552125e-01 -9.72915411e-01
1.58647275e+00 -2.04253268e+00 4.90552410e-02 3.32892239e-02
1.29327163e-01 4.39229131e-01 4.23292294e-02 5.50305486e-01
2.33415604e-01 -7.95698985e-02 -1.89964682e-01 -1.25777990e-01
-1.02917301e-02 3.59921157e-02 1.65891834e-02 6.58470929e-01
2.98419625e-01 8.33563864e-01 -9.00766850e-01 -5.16654074e-01
2.98425317e-01 4.06071514e-01 -3.16365957e-01 -1.09168869e-02
4.59171027e-01 3.79011869e-01 -6.49143159e-01 1.08894753e+00
4.99675155e-01 -1.38051242e-01 7.25944638e-02 -2.83966392e-01
-1.42877474e-01 -1.21546187e-01 -1.26851404e+00 1.53918898e+00
6.81835786e-02 3.61569554e-01 -5.19889295e-01 -1.06810582e+00
1.10327697e+00 3.55087876e-01 8.82294595e-01 -7.18353152e-01
-4.47766110e-02 2.17152804e-01 -1.19185902e-01 -6.69399858e-01
3.23255837e-01 1.14222430e-01 -3.34553212e-01 2.25345224e-01
2.24255938e-02 8.01319599e-01 5.22625387e-01 -3.23546588e-01
1.48200750e+00 1.54341504e-01 5.41319072e-01 -9.60691050e-02
7.57280290e-01 -3.39026421e-01 1.07893288e+00 3.45944285e-01
-8.39704156e-01 4.51715380e-01 3.92551810e-01 -5.26584923e-01
-7.35363901e-01 -1.08495891e+00 8.45765546e-02 7.65911758e-01
1.07341386e-01 -6.05561614e-01 -4.74653006e-01 -4.26005870e-01
1.29648939e-01 -9.12096053e-02 -7.06033051e-01 -4.22668844e-01
-8.03809583e-01 -9.25920069e-01 7.97290564e-01 7.95146108e-01
1.04993451e+00 -9.39307392e-01 -1.00450623e+00 4.72738951e-01
-3.87685388e-01 -1.17192578e+00 -5.56959391e-01 -3.96398127e-01
-8.71542215e-01 -1.06186259e+00 -7.12095499e-01 -5.23241699e-01
1.31024718e-01 1.78573236e-01 6.17740333e-01 1.82997495e-01
-5.45389354e-01 2.53047384e-02 -5.78441143e-01 3.31744045e-01
4.20100361e-01 1.58091009e-01 3.14145505e-01 3.53561312e-01
3.50250781e-01 -8.26180816e-01 -8.31120551e-01 4.95087177e-01
-4.88076001e-01 4.84443754e-02 6.28117323e-01 1.02332306e+00
4.73075479e-01 -1.45197213e-01 8.46445024e-01 -1.33900151e-01
5.18674374e-01 -4.75597322e-01 -1.98674902e-01 2.03669816e-01
-3.19029957e-01 -8.34887400e-02 4.29461062e-01 -4.30661649e-01
-6.37124777e-01 4.68123378e-03 5.84793128e-02 -3.65069121e-01
-2.18055144e-01 7.12019861e-01 -6.98311031e-02 1.22378640e-01
3.88894945e-01 7.34335065e-01 -2.86030084e-01 -4.94865447e-01
-9.45392475e-02 6.19802237e-01 9.06960070e-01 -6.85807943e-01
6.31737947e-01 4.55247849e-01 2.25635007e-01 -9.13160205e-01
-2.37813726e-01 -5.67654490e-01 -7.20225692e-01 -5.02496123e-01
8.14819753e-01 -8.52165878e-01 -7.46756077e-01 8.01111817e-01
-6.75162673e-01 5.48759140e-02 1.93841636e-01 4.94880766e-01
-6.08577967e-01 5.90730429e-01 -6.90322220e-01 -8.54942083e-01
-4.81283933e-01 -9.09481287e-01 1.21881270e+00 3.87598336e-01
-3.66286159e-01 -4.98744577e-01 9.67059210e-02 3.25399727e-01
3.33121896e-01 9.68616784e-01 4.63302165e-01 -2.30306849e-01
-2.76352435e-01 -2.88822114e-01 -2.17984408e-01 -1.76598534e-01
4.62250829e-01 1.50438413e-01 -7.44926274e-01 -3.52612942e-01
-4.48803872e-01 -1.63077429e-01 8.73061061e-01 3.00683588e-01
7.85876989e-01 -1.04812734e-01 -4.67976332e-01 4.67113763e-01
1.24835646e+00 3.83511037e-01 6.83490515e-01 3.81760627e-01
5.44317961e-01 4.33076322e-01 7.81270623e-01 9.23428416e-01
3.52735102e-01 9.39573526e-01 -3.99974361e-02 5.32926261e-01
-1.52686745e-01 -1.56625062e-01 5.15302300e-01 8.44774842e-01
-2.96956629e-01 -4.55013327e-02 -1.04272723e+00 7.29344130e-01
-2.14166808e+00 -1.27040446e+00 -6.75689727e-02 2.10195732e+00
4.11438257e-01 2.50603735e-01 5.54497838e-01 6.41498506e-01
8.72747004e-01 2.49705926e-01 -4.74920690e-01 3.16263959e-02
-4.14862148e-02 3.75036895e-02 2.67000109e-01 2.93196607e-02
-1.41136527e+00 8.19467187e-01 5.23085070e+00 6.91389263e-01
-1.05347061e+00 -3.00673455e-01 5.07126093e-01 1.10860951e-02
5.64699829e-01 -4.31134641e-01 -7.75771141e-01 9.96769011e-01
9.76946414e-01 -2.54615366e-01 1.16807200e-01 4.94673848e-01
4.66846913e-01 -6.00143410e-02 -7.02754319e-01 1.24671209e+00
-1.59552366e-01 -1.11740708e+00 -1.05557166e-01 1.70302406e-01
4.47885394e-01 -4.85433936e-01 1.93991121e-02 2.03899056e-01
-1.60019279e-01 -8.89491916e-01 4.67923671e-01 8.23311627e-01
8.06291640e-01 -8.44873726e-01 7.63192713e-01 1.30239502e-01
-2.08336997e+00 1.32239927e-02 -3.17697003e-02 -1.49840787e-01
2.95680761e-01 6.81315124e-01 -3.04292411e-01 8.10216069e-01
8.51169646e-01 1.12795448e+00 -6.12351298e-01 1.25376225e+00
6.49996400e-02 7.19315290e-01 -3.00616324e-01 -2.87430966e-03
3.47826779e-02 1.10909902e-01 3.16471905e-01 1.36951041e+00
5.91519356e-01 1.67277634e-01 4.12204415e-01 2.51630098e-01
2.53210664e-01 -8.17847028e-02 -3.75459760e-01 -2.61974782e-01
5.76065004e-01 9.90895033e-01 -8.06080937e-01 -1.38713256e-01
-3.67698580e-01 1.01468146e+00 1.13852352e-01 6.43581226e-02
-8.73581469e-01 -5.99920630e-01 8.44036222e-01 -1.42068788e-01
3.71338218e-01 -4.78281051e-01 -1.77815974e-01 -1.34602225e+00
2.39561796e-01 -7.26282656e-01 6.47829115e-01 -3.89176965e-01
-1.41660643e+00 3.12082916e-01 -8.87481421e-02 -1.54739773e+00
-4.09866273e-01 -4.44597721e-01 -6.33898854e-01 6.80208743e-01
-1.12525427e+00 -1.34918213e+00 -5.92073202e-01 6.54807031e-01
4.58108604e-01 -1.84491277e-01 5.97294807e-01 4.40803140e-01
-8.87373269e-01 6.96837485e-01 -8.44243243e-02 4.17179346e-01
5.03377199e-01 -7.40175009e-01 4.71997887e-01 1.05582464e+00
-2.07693160e-01 5.77148438e-01 5.96310616e-01 -8.04863811e-01
-1.60710382e+00 -1.09369266e+00 9.50159907e-01 -2.93168537e-02
6.37178123e-01 -3.88670936e-02 -8.65111709e-01 3.12207907e-01
-4.64284271e-01 2.75693655e-01 7.23029196e-01 -1.56310916e-01
-1.90578327e-01 -2.63204813e-01 -1.20893002e+00 5.63787818e-01
1.24041283e+00 -3.24944705e-01 -5.59254348e-01 -2.50399560e-01
1.69751704e-01 -2.28813335e-01 -1.19497967e+00 7.16290116e-01
1.00474000e+00 -8.86878371e-01 1.04274035e+00 -4.58693027e-01
2.49089926e-01 -6.62427545e-01 -5.30236483e-01 -8.81845534e-01
-5.41244566e-01 -3.68034720e-01 -6.69830441e-01 1.29751587e+00
-2.58444101e-01 -5.84807634e-01 9.33867872e-01 4.61720787e-02
1.50657457e-03 -9.21689868e-01 -1.23202801e+00 -9.98611987e-01
-4.12796170e-01 -3.72581720e-01 6.19939148e-01 8.53790164e-01
2.46102408e-01 -1.67004038e-02 -6.74994767e-01 1.40464976e-01
6.75103247e-01 2.66784191e-01 7.03649223e-01 -1.23581421e+00
-2.77483851e-01 -3.48030090e-01 -1.09421873e+00 -6.72119796e-01
-1.74344569e-01 -4.89040226e-01 -1.38913751e-01 -1.32791936e+00
1.27877653e-01 4.14090790e-02 -6.97018087e-01 4.59533274e-01
-2.96325892e-01 4.19987738e-01 1.21722408e-01 3.45862061e-01
-5.61205566e-01 6.99286103e-01 9.76019144e-01 -7.05351606e-02
-9.02134180e-02 -2.15968266e-01 -3.14896733e-01 4.57689255e-01
1.16153240e+00 -8.01056325e-02 -2.44701192e-01 -3.19576301e-02
-4.66799647e-01 1.77294999e-01 4.41791922e-01 -1.67913187e+00
1.36227578e-01 -3.25681657e-01 7.22127080e-01 -7.17776716e-01
5.23594201e-01 -9.92030799e-02 2.36123294e-01 8.60739768e-01
3.51873897e-02 4.19803917e-01 1.23950168e-01 7.63460338e-01
-2.44230658e-01 4.77520466e-01 6.88625813e-01 9.71470028e-02
-1.04786253e+00 4.20986146e-01 -3.82874310e-01 -8.54822770e-02
1.06972396e+00 -4.34426576e-01 -4.46192563e-01 -4.53009047e-02
-7.25662470e-01 2.36450523e-01 2.46073842e-01 4.44998085e-01
8.83685589e-01 -1.75218880e+00 -6.17286265e-01 1.84197024e-01
3.54803950e-01 -7.06208289e-01 5.02998769e-01 9.19341803e-01
-4.21524376e-01 5.04917741e-01 -6.04782224e-01 -6.58086121e-01
-1.27839017e+00 1.66866884e-01 9.17069986e-02 -5.55721164e-01
-6.28708005e-01 4.18277979e-01 -3.03412974e-01 1.88289851e-01
-1.53354377e-01 -1.94391832e-01 -3.02134037e-01 7.24976510e-02
6.87799335e-01 8.56007338e-01 -6.50681108e-02 -9.60597992e-01
-7.79413760e-01 8.35514605e-01 9.61547494e-02 -8.63576680e-02
1.16897070e+00 -2.04090145e-03 1.39425695e-01 4.51991886e-01
1.06929123e+00 -3.80826086e-01 -1.35514796e+00 5.33195101e-02
3.60299766e-01 -5.79426825e-01 -3.99435282e-01 -5.45796514e-01
-8.93950999e-01 8.54450107e-01 8.48136902e-01 2.04035416e-02
1.31863713e+00 -5.00718117e-01 1.15453315e+00 1.55429598e-02
7.12023735e-01 -9.40335035e-01 3.07813346e-01 4.69819635e-01
6.20935857e-01 -8.28754663e-01 -9.58937593e-03 -3.59645307e-01
-3.09686631e-01 1.02525771e+00 6.40500486e-01 -2.96768546e-01
4.02053922e-01 5.35999872e-02 -3.72442603e-01 3.49758342e-02
-8.76684546e-01 -2.81440079e-01 2.73558766e-01 6.88775003e-01
5.83918691e-01 2.86741942e-01 -6.51290357e-01 7.05002129e-01
-8.59537274e-02 3.69578689e-01 1.68357491e-01 1.27764428e+00
-6.44862294e-01 -1.04141212e+00 -3.33205730e-01 5.78324378e-01
-4.29760724e-01 3.58460099e-01 -5.86357526e-02 6.29422784e-01
2.30673030e-01 1.11048806e+00 -2.76858300e-01 -9.82990682e-01
1.59793720e-01 1.10664554e-01 3.47086102e-01 -1.32952854e-01
-1.47351637e-01 3.41048725e-02 1.37551039e-01 -8.12854588e-01
-3.57802331e-01 -8.84012282e-01 -1.24572003e+00 -5.14009595e-01
2.05534115e-01 -2.05291770e-02 1.29073799e-01 8.88342857e-01
5.65807402e-01 4.79050279e-01 4.50739026e-01 -8.26667249e-01
-1.88978821e-01 -9.27461147e-01 -7.16652453e-01 4.15708482e-01
2.26821572e-01 -1.10386217e+00 -4.82765473e-02 2.33940378e-01] | [14.272857666015625, 1.4439592361450195] |
82034608-0c07-48d1-85db-e100655b8056 | correcting-underrepresentation-and | 2306.11112 | null | https://arxiv.org/abs/2306.11112v1 | https://arxiv.org/pdf/2306.11112v1.pdf | Correcting Underrepresentation and Intersectional Bias for Fair Classification | We consider the problem of learning from data corrupted by underrepresentation bias, where positive examples are filtered from the data at different, unknown rates for a fixed number of sensitive groups. We show that with a small amount of unbiased data, we can efficiently estimate the group-wise drop-out parameters, even in settings where intersectional group membership makes learning each intersectional rate computationally infeasible. Using this estimate for the group-wise drop-out rate, we construct a re-weighting scheme that allows us to approximate the loss of any hypothesis on the true distribution, even if we only observe the empirical error on a biased sample. Finally, we present an algorithm encapsulating this learning and re-weighting process, and we provide strong PAC-style guarantees that, with high probability, our estimate of the risk of the hypothesis over the true distribution will be arbitrarily close to the true risk. | ['Emily Diana', 'Alexander Williams Tolbert'] | 2023-06-19 | null | null | null | null | ['classification-1'] | ['methodology'] | [ 2.10157841e-01 6.29527092e-01 -5.73394060e-01 -5.12758732e-01
-1.51980329e+00 -6.45332575e-01 3.29248130e-01 5.04874349e-01
-8.54202509e-01 1.08927739e+00 6.91357702e-02 -2.38133103e-01
-2.68322885e-01 -9.43453491e-01 -1.33472395e+00 -8.78434598e-01
-4.40016627e-01 8.09485495e-01 9.35419127e-02 5.77743709e-01
2.24444956e-01 6.95789233e-02 -1.25514269e+00 1.00596324e-01
8.16356242e-01 6.60084188e-01 -5.05946457e-01 7.60501266e-01
5.84629476e-01 3.77000093e-01 -6.20501220e-01 -6.32529616e-01
5.16490638e-01 -3.17478925e-01 -5.28864801e-01 -5.96848018e-02
7.41801023e-01 -6.39635086e-01 2.60366430e-03 1.52227128e+00
4.01265115e-01 4.51229773e-02 8.97806108e-01 -1.24170518e+00
-2.68424720e-01 1.09684181e+00 -8.19955409e-01 3.92781757e-03
1.53940991e-01 5.56134842e-02 1.35084331e+00 -4.69782293e-01
5.46255052e-01 1.30573618e+00 7.34938741e-01 5.67456067e-01
-1.62467206e+00 -1.17009163e+00 2.20083833e-01 -4.65933233e-01
-1.25084448e+00 -3.32157612e-01 1.47958666e-01 -4.41111505e-01
1.51412398e-01 9.97824371e-02 3.46746057e-01 9.37060356e-01
-8.84185955e-02 5.41741192e-01 9.48408723e-01 -3.21935356e-01
6.27608836e-01 4.39071238e-01 3.87614191e-01 5.60211241e-01
1.01873004e+00 2.42678061e-01 -5.11400342e-01 -8.65032911e-01
2.85575122e-01 1.03582583e-01 -4.81077075e-01 -5.05057216e-01
-6.42821491e-01 9.85398173e-01 1.89006567e-01 -1.59508049e-01
1.90747920e-02 3.33528936e-01 1.72246456e-01 3.62154692e-01
6.20674372e-01 3.34220767e-01 -5.79291940e-01 3.00176978e-01
-7.86425292e-01 6.64745152e-01 9.38094735e-01 8.09753776e-01
6.57619834e-01 -5.60541749e-01 -2.65463799e-01 4.76484865e-01
6.99703097e-02 7.67065525e-01 1.11651123e-01 -8.27302754e-01
7.84715235e-01 2.01313928e-01 7.61908412e-01 -6.92902923e-01
-1.38966322e-01 -3.95682156e-01 -4.85828370e-01 1.86433613e-01
1.00427103e+00 -5.89881778e-01 -6.89916253e-01 2.43829274e+00
3.58319074e-01 7.39193037e-02 -2.51508296e-01 5.44325292e-01
-3.20467860e-01 4.59972709e-01 2.53374249e-01 -5.24427474e-01
1.05066359e+00 -8.34555104e-02 -4.30061311e-01 -7.27320015e-02
8.05881321e-01 -1.70136526e-01 7.88344562e-01 3.76269788e-01
-1.19950724e+00 7.57172704e-02 -1.05177462e+00 1.72288179e-01
1.09962076e-01 -3.62846971e-01 4.97022718e-01 8.98654401e-01
-5.11068761e-01 7.19980121e-01 -6.10344172e-01 -4.00694758e-02
9.49167371e-01 4.32132095e-01 -2.69785851e-01 -1.81816280e-01
-1.01467586e+00 4.96035904e-01 4.49767023e-01 -4.34254706e-01
-1.18436122e+00 -1.39460588e+00 -4.20090288e-01 1.64154470e-01
6.21476710e-01 -6.13813698e-01 1.16425169e+00 -8.30452442e-01
-5.96329749e-01 8.68031979e-01 -1.61615878e-01 -8.76275301e-01
1.19205749e+00 -2.00376779e-01 1.26808554e-01 -1.67900056e-01
2.62175888e-01 1.12728015e-01 7.39556372e-01 -1.34122396e+00
-1.16051269e+00 -7.58380592e-01 -2.17614081e-02 1.49493009e-01
-3.22195530e-01 -3.28672409e-01 1.72303215e-01 -3.40546221e-01
-8.46912339e-02 -6.82501614e-01 -3.80258083e-01 4.06210512e-01
-4.36810672e-01 -1.64592728e-01 4.44150776e-01 -4.11129713e-01
1.13342822e+00 -2.23715186e+00 -3.74167413e-01 5.47286749e-01
3.23096246e-01 -5.58030717e-02 1.27517834e-01 8.20225552e-02
1.02032848e-01 4.85017598e-01 -5.49697280e-01 -3.66459042e-01
-3.61961462e-02 -5.04076108e-02 -5.08731902e-01 1.05864060e+00
-1.78184491e-02 9.37168002e-02 -1.07134902e+00 -4.72715832e-02
-3.61012399e-01 -1.33711576e-01 -9.95986700e-01 1.72074050e-01
2.96447636e-03 -1.65381044e-01 -3.35633874e-01 7.94614330e-02
9.29969072e-01 -1.00980952e-01 2.48602703e-01 2.04159841e-01
4.08579469e-01 1.65966600e-01 -1.45358431e+00 6.74872637e-01
-3.77299339e-01 6.67938590e-02 2.70776749e-01 -9.69038129e-01
4.98460680e-01 1.74155563e-01 2.02108711e-01 1.44546688e-01
1.57633707e-01 4.92391437e-01 -2.00712562e-01 -9.98671129e-02
1.59696370e-01 -8.82638693e-01 -2.57896394e-01 7.36327112e-01
7.78650492e-02 2.17042789e-01 -2.48572707e-01 1.78871453e-01
8.97995353e-01 -7.00595498e-01 3.28964621e-01 -2.61079907e-01
1.04192961e-02 -4.36787665e-01 7.39107311e-01 1.46705234e+00
-1.86137751e-01 5.24783015e-01 9.98263240e-01 -1.25443369e-01
-1.27194095e+00 -1.24916518e+00 -4.61904824e-01 8.61134648e-01
-7.65789449e-02 1.07591078e-01 -5.77602744e-01 -1.09099448e+00
6.74347341e-01 6.62085950e-01 -9.31745946e-01 -3.67636293e-01
-4.71873358e-02 -1.25047255e+00 2.52417922e-01 1.87340081e-01
1.55117720e-01 -1.67363927e-01 -4.74152304e-02 -2.15018228e-01
1.42965302e-01 -7.25093782e-01 -5.98486543e-01 9.28295776e-02
-8.71206522e-01 -1.38090277e+00 -5.79966903e-01 -4.46047544e-01
1.04619706e+00 -2.87570059e-02 8.39378417e-01 5.35157211e-02
1.94397181e-01 -1.61617342e-02 1.87601686e-01 -6.80522442e-01
-4.15107131e-01 -8.11364204e-02 1.49337783e-01 8.82587582e-02
4.40633088e-01 -3.62007171e-01 -5.04199266e-01 3.62439118e-02
-6.53473675e-01 -6.66887105e-01 1.14298522e-01 8.91155243e-01
4.78870541e-01 2.29582921e-01 9.18876529e-01 -1.56097412e+00
5.19325972e-01 -8.15722346e-01 -1.12886930e+00 2.36595929e-01
-5.19306481e-01 5.65634295e-02 5.17217815e-01 -5.64699531e-01
-8.15491438e-01 -2.16478705e-01 3.79143655e-01 -1.33816257e-01
1.78911448e-01 4.27124128e-02 -4.37293530e-01 9.50345621e-02
8.71726096e-01 -2.80336350e-01 1.23497257e-02 -4.13437665e-01
2.90371776e-01 7.19099641e-01 5.25622964e-01 -7.86163867e-01
7.81828284e-01 6.68776095e-01 1.04556479e-01 -3.35203975e-01
-1.36045933e+00 -1.60271615e-01 -1.52643219e-01 1.82354033e-01
3.71079952e-01 -1.01600170e+00 -8.98818612e-01 3.20101500e-01
-6.07786536e-01 -3.69449109e-01 -6.89131498e-01 6.00353003e-01
-5.00282526e-01 1.55385047e-01 -1.00070007e-01 -1.30783403e+00
5.92425056e-02 -6.11767530e-01 6.37863398e-01 -1.05215363e-01
-2.25186691e-01 -9.76348341e-01 -1.87935561e-01 2.42110372e-01
-1.36951938e-01 3.49619418e-01 1.06538200e+00 -1.19524348e+00
-4.67372686e-01 -7.81357229e-01 -1.59890443e-01 4.95936066e-01
-4.36833799e-02 -1.76609650e-01 -9.48999107e-01 -5.81921816e-01
3.05422526e-02 -4.86343682e-01 1.04212821e+00 6.31066024e-01
1.53270829e+00 -1.02957392e+00 -4.02707815e-01 3.53792757e-01
1.44716454e+00 -1.83738187e-01 1.52843222e-01 7.05250911e-03
3.21172476e-01 5.32167315e-01 6.85139298e-01 8.51845086e-01
5.12218624e-02 1.56843260e-01 1.68181643e-01 4.01886731e-01
4.00749475e-01 -6.38762474e-01 2.07007065e-01 -2.47696057e-01
4.19925928e-01 -2.60699958e-01 -4.68518436e-01 9.33455110e-01
-1.64370620e+00 -9.97538745e-01 2.33942494e-01 3.32521605e+00
1.35956657e+00 4.37783629e-01 5.19957483e-01 1.32334203e-01
9.41624463e-01 -6.11017942e-02 -7.61936069e-01 -1.38411641e-01
-2.68966444e-02 6.51875287e-02 1.26695931e+00 9.85395789e-01
-1.01303995e+00 3.87036771e-01 7.59906387e+00 8.20267498e-01
-5.68571389e-01 2.32578702e-02 1.07884526e+00 -4.11287069e-01
-6.54232621e-01 1.39917493e-01 -9.68898475e-01 6.45685136e-01
9.92699504e-01 -6.59474015e-01 1.60366729e-01 9.38473523e-01
-1.08856484e-01 -2.38248140e-01 -1.49548340e+00 3.62169594e-01
-2.40804598e-01 -1.01612639e+00 4.78505120e-02 3.70404184e-01
9.83527422e-01 -3.00795436e-01 1.03897028e-01 1.02376111e-01
1.02352607e+00 -7.59780109e-01 6.59528375e-01 2.35033199e-01
6.88601851e-01 -1.24107409e+00 8.17001700e-01 5.47031164e-01
-1.75725952e-01 -5.54177821e-01 -5.03293931e-01 -1.76162980e-02
-1.51209280e-01 1.23328149e+00 -9.27294195e-01 -9.38256010e-02
3.66614550e-01 3.29039723e-01 5.67336529e-02 1.20113051e+00
-7.64158741e-02 8.83834779e-01 -6.40177667e-01 4.32524979e-02
-2.97396302e-01 9.07089338e-02 5.95865488e-01 9.04974222e-01
3.74756068e-01 8.84328708e-02 1.55179143e-01 7.12833583e-01
-8.43627751e-01 -2.13425025e-01 -5.82214415e-01 2.01863259e-01
1.02150023e+00 7.06385136e-01 -1.22325324e-01 -3.73776704e-01
4.84474376e-02 2.41775915e-01 3.96694601e-01 2.94534296e-01
-5.80200255e-01 -3.17454398e-01 7.89925277e-01 1.44078895e-01
2.56507605e-01 4.10112500e-01 -4.12328780e-01 -8.77089441e-01
9.60055441e-02 -6.81376338e-01 9.25307989e-01 -2.30024442e-01
-1.73780918e+00 -4.31700706e-01 1.83661029e-01 -8.24791014e-01
-7.93503821e-02 -4.88948703e-01 -4.54207331e-01 8.54545474e-01
-1.08733833e+00 -2.84953654e-01 4.84714478e-01 1.58486828e-01
-6.51540980e-02 1.55799195e-01 4.35548514e-01 4.99695763e-02
-3.58169347e-01 1.15678489e+00 1.80052400e-01 7.84743279e-02
8.46964359e-01 -1.41763425e+00 -1.75686225e-01 6.94361389e-01
-4.12575930e-01 5.32724977e-01 9.90863442e-01 -6.64342642e-01
-9.29531991e-01 -1.15361226e+00 9.33397353e-01 -6.74123108e-01
7.87874818e-01 -4.36835676e-01 -9.73752797e-01 1.03571248e+00
-4.86315399e-01 3.09765607e-01 7.66520441e-01 5.32081246e-01
-6.83256269e-01 -4.90825325e-01 -1.92836916e+00 4.75467563e-01
8.12739730e-01 -3.37412983e-01 -4.69386309e-01 4.78966892e-01
4.86416489e-01 -2.05377981e-01 -6.88534737e-01 2.25189075e-01
5.83291888e-01 -7.21102118e-01 7.97574997e-01 -1.02210593e+00
2.68863410e-01 -1.38004512e-01 -2.72740453e-01 -1.20335627e+00
-8.34401250e-02 -6.13058150e-01 -1.55455321e-02 1.25587857e+00
7.11002946e-01 -8.45337510e-01 8.81977558e-01 8.71542871e-01
6.59697890e-01 -4.12961811e-01 -1.27646923e+00 -6.34157956e-01
4.50839788e-01 -2.63275653e-01 4.66409028e-01 7.63296843e-01
-5.13556264e-02 -4.19024974e-02 -5.95319211e-01 6.52120769e-01
1.20753765e+00 -1.49566159e-01 6.23302221e-01 -1.31470418e+00
-4.42891210e-01 -2.11369917e-01 -3.03196073e-01 -5.49012959e-01
1.24429651e-01 -7.28754997e-01 1.71505034e-01 -7.96889842e-01
9.55901384e-01 -5.73068380e-01 -3.73251915e-01 2.65874922e-01
-4.89289969e-01 -6.43110573e-02 -6.68532476e-02 -1.05465025e-01
-5.44216096e-01 2.40823850e-01 9.52170312e-01 9.21368971e-02
-3.26757878e-02 2.95385838e-01 -1.34204865e+00 8.34473014e-01
6.14655495e-01 -1.04258001e+00 -2.91341186e-01 6.51580989e-02
3.53461474e-01 -5.10218404e-02 3.56142372e-01 -7.92308390e-01
-1.22332415e-02 -2.81058311e-01 3.47221017e-01 -3.21159661e-01
-1.27702400e-01 -6.88084900e-01 -2.46959060e-01 6.24353290e-01
-1.27543819e+00 -6.66228235e-01 -2.59338349e-01 1.10473299e+00
4.47386831e-01 -3.25504124e-01 1.14337516e+00 1.77647322e-02
4.70350564e-01 3.55506301e-01 -2.36076340e-01 6.58229530e-01
9.87852275e-01 4.41870034e-01 -3.41013670e-01 -4.97662336e-01
-4.67659563e-01 5.47931194e-01 4.01864856e-01 -1.94805130e-01
1.47764966e-01 -1.18127203e+00 -9.44057763e-01 1.90931678e-01
2.17515767e-01 7.82406554e-02 5.91237247e-02 4.39369410e-01
2.43537650e-01 -1.63934141e-01 3.97524178e-01 -2.88570344e-01
-1.18593121e+00 5.52794218e-01 5.06378531e-01 -1.12894632e-01
-1.73686609e-01 1.03874624e+00 4.53840345e-01 -3.89911592e-01
4.65689361e-01 -1.52600691e-01 4.60165292e-01 1.26119852e-01
9.46006715e-01 5.38064241e-01 -2.26992682e-01 -2.74397079e-02
-1.12214215e-01 1.27384007e-01 -4.17261928e-01 -3.19435090e-01
1.11410630e+00 -7.98965991e-02 -4.49656844e-02 5.54540873e-01
1.27359581e+00 3.11831325e-01 -1.53457320e+00 -3.20089400e-01
2.92707589e-02 -1.06641448e+00 -1.80184111e-01 -8.24950039e-01
-8.48660409e-01 7.04626501e-01 4.53569919e-01 1.51340932e-01
6.88044548e-01 7.68400654e-02 2.50749707e-01 3.83798629e-01
4.94179159e-01 -1.08652151e+00 -3.71130556e-01 -9.90043432e-02
6.03309691e-01 -1.22343278e+00 2.65172482e-01 -2.86587983e-01
-3.14042032e-01 4.92023259e-01 2.56819636e-01 -5.23726463e-01
8.52624595e-01 2.33350337e-01 -3.85687739e-01 3.15028012e-01
-9.55000639e-01 2.63860881e-01 -3.02276146e-02 8.16156089e-01
1.38038397e-01 4.21946585e-01 -4.05651897e-01 7.91859329e-01
-2.18282208e-01 7.19069391e-02 8.91880214e-01 5.28936267e-01
-8.04751873e-01 -7.66972840e-01 -3.80182654e-01 8.78146946e-01
-1.06866872e+00 1.76160738e-01 -1.80561215e-01 7.09500134e-01
-7.95033108e-03 6.65206850e-01 2.58675218e-01 -2.20981650e-02
2.81301111e-01 -3.61733139e-02 4.47113723e-01 -6.25714600e-01
-1.80839926e-01 -1.87309861e-01 1.48516804e-01 -3.78184676e-01
-8.92189443e-02 -8.38806331e-01 -9.43895936e-01 -5.73379576e-01
-2.57001340e-01 3.22330356e-01 5.75603694e-02 9.04595852e-01
-3.63867288e-03 -2.52120316e-01 1.06995797e+00 -1.87387764e-01
-1.46255481e+00 -5.82284629e-01 -1.01171112e+00 5.48475862e-01
8.38127553e-01 -3.19053739e-01 -1.20002687e+00 -3.17137182e-01] | [8.315094947814941, 4.8316874504089355] |
809b35ed-6bfd-43b7-9865-5fec75b0535c | improved-decipherment-of-homophonic-ciphers | null | null | https://aclanthology.org/D14-1184 | https://aclanthology.org/D14-1184.pdf | Improved Decipherment of Homophonic Ciphers | null | ['Hermann Ney', 'Julian Schamper', 'Malte Nuhn'] | 2014-10-01 | null | null | null | emnlp-2014-10 | ['decipherment'] | ['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] | [-7.42834997177124, 3.658884048461914] |
eec70365-d2a9-4763-87b3-561d1eec40c8 | em-driven-unsupervised-learning-for-efficient | 2201.02074 | null | https://arxiv.org/abs/2201.02074v3 | https://arxiv.org/pdf/2201.02074v3.pdf | EM-driven unsupervised learning for efficient motion segmentation | In this paper, we present a CNN-based fully unsupervised method for motion segmentation from optical flow. We assume that the input optical flow can be represented as a piecewise set of parametric motion models, typically, affine or quadratic motion models. The core idea of our work is to leverage the Expectation-Maximization (EM) framework in order to design in a well-founded manner a loss function and a training procedure of our motion segmentation neural network that does not require either ground-truth or manual annotation. However, in contrast to the classical iterative EM, once the network is trained, we can provide a segmentation for any unseen optical flow field in a single inference step and without estimating any motion models. We investigate different loss functions including robust ones and propose a novel efficient data augmentation technique on the optical flow field, applicable to any network taking optical flow as input. In addition, our method is able by design to segment multiple motions. Our motion segmentation network was tested on four benchmarks, DAVIS2016, SegTrackV2, FBMS59, and MoCA, and performed very well, while being fast at test time. | ['Patrick Bouthemy', 'Anaïs Badoual', 'Etienne Meunier'] | 2022-01-06 | null | null | null | null | ['unsupervised-object-segmentation', 'motion-segmentation'] | ['computer-vision', 'computer-vision'] | [-1.13876648e-02 1.87139641e-02 -1.70561820e-01 -1.40697315e-01
-1.73828974e-01 -6.41653836e-01 4.89602447e-01 -1.93270326e-01
-7.81501293e-01 6.65163994e-01 -1.27128139e-01 -3.31569493e-01
1.14331640e-01 -6.55004501e-01 -9.16379809e-01 -6.23776555e-01
2.14541078e-01 5.47420919e-01 4.65828419e-01 9.30471644e-02
2.17979938e-01 5.94127357e-01 -1.15936053e+00 -4.50325161e-01
9.00959790e-01 7.49624789e-01 1.90286458e-01 1.04801452e+00
6.50141621e-03 1.19107163e+00 -2.63661742e-01 -3.08539391e-01
3.25161695e-01 -5.72176933e-01 -1.38257670e+00 5.60497522e-01
6.81381345e-01 -4.51328278e-01 -2.36850753e-01 8.09439063e-01
2.24835083e-01 5.75804591e-01 6.56593323e-01 -1.14360142e+00
-5.06582558e-02 3.21120888e-01 -2.88497239e-01 1.15510359e-01
1.92907795e-01 4.30243105e-01 7.57089198e-01 -7.08714366e-01
1.14433253e+00 1.08472419e+00 4.71231699e-01 7.67125547e-01
-1.47516656e+00 -2.12524980e-02 1.60588890e-01 1.97820336e-01
-9.79471684e-01 -4.01260704e-01 6.71438813e-01 -8.26114058e-01
6.97745979e-01 -1.96796432e-02 6.42123580e-01 1.01604795e+00
-1.73965201e-01 9.59748268e-01 5.84326506e-01 -3.81904811e-01
4.14237559e-01 -8.33593011e-02 1.20670438e-01 8.97168338e-01
-2.32345108e-02 -2.28757694e-01 2.96129994e-02 8.54623616e-02
9.00422275e-01 -1.38396844e-01 -5.33351123e-01 -6.29366159e-01
-1.18180537e+00 7.75724411e-01 4.69219029e-01 2.50578731e-01
-2.89825857e-01 4.58679944e-01 2.80403078e-01 -2.92227436e-02
3.56326759e-01 2.81395674e-01 -3.04376841e-01 -9.92217436e-02
-1.29288304e+00 2.77268380e-01 9.90548193e-01 5.89449763e-01
1.25828218e+00 1.83684006e-02 -2.14105517e-01 3.23256463e-01
4.12945300e-01 3.10500562e-01 4.11806911e-01 -1.44757557e+00
4.20659214e-01 2.76825666e-01 3.72865707e-01 -1.13663399e+00
-4.82897878e-01 -2.14914545e-01 -7.89741635e-01 2.96002477e-01
9.44659472e-01 -4.39925671e-01 -9.90863383e-01 1.76776505e+00
4.81691271e-01 6.88446999e-01 -8.85314122e-02 1.16342974e+00
3.72465998e-01 4.82610643e-01 -9.10667926e-02 -8.72997493e-02
7.97762692e-01 -1.39055777e+00 -6.71293437e-01 -1.48582086e-01
1.00187218e+00 -6.35991395e-01 7.35652149e-01 4.17724371e-01
-1.30968630e+00 -6.50478244e-01 -8.66619825e-01 -3.19220841e-01
1.78045873e-02 1.60779998e-01 4.80758458e-01 5.16104043e-01
-1.25333762e+00 9.99394536e-01 -1.28684580e+00 -3.16548258e-01
3.80366981e-01 3.43443453e-01 -3.88826549e-01 1.47362530e-01
-7.39000857e-01 6.50933504e-01 5.34492314e-01 4.95662123e-01
-9.41731870e-01 -4.20453489e-01 -1.08479512e+00 -1.17220968e-01
4.62572008e-01 -1.27888787e+00 1.08082974e+00 -1.04522443e+00
-1.81578767e+00 5.82086861e-01 -3.54101658e-01 -6.71077967e-01
1.05700850e+00 -2.50645161e-01 2.66041279e-01 5.32286942e-01
-1.46468177e-01 1.01664495e+00 9.08506393e-01 -1.13455498e+00
-3.85513574e-01 6.07011504e-02 2.05870092e-01 -2.21168190e-01
-1.39836729e-01 -2.84996599e-01 -5.31193614e-01 -4.96870399e-01
-1.48785010e-01 -1.10109055e+00 -6.13357544e-01 3.60061340e-02
-5.30471981e-01 4.72425036e-02 8.08419585e-01 -6.33900106e-01
1.11416924e+00 -1.80122244e+00 5.96708059e-01 1.10125050e-01
2.16864496e-01 5.82224071e-01 -5.93451411e-02 -4.87667993e-02
1.91100731e-01 1.12897769e-01 -9.00211751e-01 -9.28929269e-01
-8.98618624e-02 4.65624481e-01 -7.10799769e-02 5.54185033e-01
5.52738309e-01 1.01362431e+00 -1.05869448e+00 -6.29249990e-01
5.46791196e-01 5.45218229e-01 -9.33996201e-01 3.16652983e-01
-5.60010672e-01 9.98693168e-01 -2.48364970e-01 9.57377478e-02
7.41651058e-01 -3.60965401e-01 -2.14182548e-02 5.93634807e-02
-1.56043097e-01 3.50634078e-03 -1.51791322e+00 2.13201261e+00
-4.86532778e-01 7.37460613e-01 5.70183732e-02 -1.13613617e+00
6.56528175e-01 1.48091644e-01 7.34416306e-01 5.99018410e-02
3.56324792e-01 2.03167707e-01 -2.31350482e-01 -7.04289198e-01
4.12751794e-01 1.02323517e-01 3.79294813e-01 4.01588768e-01
3.74148965e-01 1.21095851e-01 5.77248693e-01 -3.87984626e-02
9.57592368e-01 5.47738791e-01 -1.47438645e-01 -1.82098746e-01
9.87344503e-01 1.19402759e-01 6.60715818e-01 5.91108561e-01
-2.18636662e-01 8.59133065e-01 4.09892440e-01 -5.24859965e-01
-8.69535625e-01 -7.59430051e-01 5.19474335e-02 4.82614815e-01
1.50887921e-01 -1.88876092e-01 -9.30311799e-01 -9.10213768e-01
-2.97952682e-01 2.56366521e-01 -4.78167176e-01 2.06238851e-01
-8.68913352e-01 -6.25838578e-01 3.54511172e-01 4.85794812e-01
5.73108435e-01 -1.07807839e+00 -6.91624343e-01 4.01264042e-01
-3.62035513e-01 -1.60575998e+00 -7.26069987e-01 -2.42323712e-01
-1.01774073e+00 -1.15142500e+00 -9.15443420e-01 -7.76499927e-01
6.75863504e-01 -7.32143298e-02 1.04883504e+00 1.11487620e-01
-1.94926351e-01 4.42459464e-01 -7.03998134e-02 2.12565124e-01
-3.22513849e-01 2.05703050e-01 -3.80401015e-01 5.59602916e-01
-7.54179657e-02 -5.70284009e-01 -8.83983493e-01 2.16434240e-01
-1.00479269e+00 -1.81421697e-01 4.34711695e-01 6.06011510e-01
5.17127275e-01 -4.46636170e-01 4.56214994e-02 -8.66038978e-01
-1.82049293e-02 -3.59595925e-01 -7.68269360e-01 -1.05876112e-02
-3.38703662e-01 2.38106817e-01 7.28815019e-01 -3.12785864e-01
-8.95989656e-01 5.87053120e-01 -1.52323157e-01 -8.30524266e-01
-4.51170862e-01 2.59188384e-01 -1.90180000e-02 -1.98955104e-01
5.48973978e-01 1.04056105e-01 8.66505504e-02 -4.78269756e-01
6.05879903e-01 1.10242903e-01 8.04560423e-01 -4.14826721e-01
8.74588072e-01 9.51433003e-01 4.02420580e-01 -9.67605948e-01
-4.85525757e-01 -7.41536379e-01 -1.12678421e+00 -2.47786537e-01
1.29642510e+00 -5.78123510e-01 -7.95838475e-01 6.70376122e-01
-1.48850310e+00 -7.82155275e-01 -1.73630640e-01 6.71565771e-01
-8.09806406e-01 8.02431703e-01 -7.05198586e-01 -6.95721984e-01
-1.62736103e-01 -1.30439210e+00 8.30919504e-01 2.30090797e-01
-6.21216111e-02 -1.63482094e+00 2.53839016e-01 3.63819033e-01
2.56405443e-01 6.24997079e-01 2.91511744e-01 -2.47281000e-01
-9.86627221e-01 1.36261031e-01 -2.61529125e-02 6.75024331e-01
1.42243160e-02 4.10400331e-01 -8.86359692e-01 -1.87432066e-01
-3.17148082e-02 -2.35791504e-01 1.24780595e+00 5.88082314e-01
9.65065360e-01 -3.18471193e-01 -2.04648033e-01 1.03457046e+00
1.47116935e+00 -2.26700604e-01 6.01781785e-01 1.96879268e-01
1.04054785e+00 6.13012671e-01 3.02397609e-01 2.67597497e-01
4.54258472e-01 7.08044469e-01 5.67939878e-01 -4.64822724e-02
-2.66320407e-02 2.91625876e-02 3.53142947e-01 6.23297095e-01
-2.97253817e-01 -2.90375024e-01 -7.65609622e-01 7.68710256e-01
-2.06394601e+00 -9.31698263e-01 -5.70218980e-01 2.22434878e+00
5.55100560e-01 7.67914280e-02 3.56958836e-01 6.77610934e-02
5.07672966e-01 1.63996875e-01 -2.91638672e-01 -3.23529512e-01
-4.24600057e-02 2.96523869e-01 5.14255106e-01 9.75469530e-01
-1.50902092e+00 1.04290664e+00 5.28538752e+00 2.90597081e-01
-1.33144033e+00 -3.74856591e-02 5.13373554e-01 8.36111084e-02
-5.93102872e-02 2.33154565e-01 -6.64924502e-01 4.33055639e-01
7.95818210e-01 2.30815142e-01 2.74920255e-01 5.62743723e-01
5.52902162e-01 -1.66736141e-01 -1.05545533e+00 8.00124645e-01
-1.61812261e-01 -1.45595646e+00 -3.05750277e-02 -6.12569898e-02
8.75903487e-01 -1.27741508e-03 -1.95625037e-01 -7.15890005e-02
5.86927123e-02 -8.88715804e-01 5.38374722e-01 7.18585253e-01
1.55974597e-01 -5.60532928e-01 6.22968435e-01 3.94834548e-01
-1.00401151e+00 2.60151565e-01 -1.32640600e-01 -4.19970751e-02
5.22172451e-01 5.82978904e-01 -6.02144539e-01 7.76848972e-01
3.02162975e-01 9.19086158e-01 -4.17292178e-01 1.32319915e+00
-2.01448798e-01 6.78407848e-01 -2.55436808e-01 1.73768654e-01
6.19417787e-01 -4.12326992e-01 6.70910716e-01 1.43703103e+00
1.03423283e-01 -4.22459513e-01 4.80563760e-01 1.07832706e+00
2.82467362e-02 1.11534677e-01 -2.54946560e-01 1.95805073e-01
-1.20802589e-01 1.39731455e+00 -8.58537495e-01 -2.62988746e-01
-2.94192046e-01 1.12355781e+00 2.91746825e-01 5.30282795e-01
-7.44929194e-01 -2.69951016e-01 5.88363230e-01 3.29201408e-02
4.89706725e-01 -6.08197451e-01 8.11924972e-03 -1.40060103e+00
1.48696125e-01 -1.98808149e-01 4.32931930e-02 -4.27838802e-01
-9.20913994e-01 4.52791274e-01 -1.02753818e-01 -1.13218343e+00
-6.71965361e-01 -7.51022756e-01 -8.00639868e-01 9.08700705e-01
-1.78955936e+00 -8.15385282e-01 -3.98091733e-01 6.00986063e-01
3.43427747e-01 3.68362099e-01 3.18473607e-01 4.26974058e-01
-7.97267556e-01 1.68311700e-01 -1.62172094e-01 5.51373363e-01
6.54673815e-01 -1.53452480e+00 4.03303951e-01 1.12107468e+00
1.79139465e-01 3.61472547e-01 6.64647043e-01 -2.62387961e-01
-1.12575722e+00 -1.26009190e+00 7.67087281e-01 -3.82151842e-01
7.39460886e-01 -1.66176006e-01 -1.05620492e+00 7.63816476e-01
2.11965233e-01 4.47247535e-01 1.31554767e-01 -5.33862412e-01
2.73993701e-01 2.64544874e-01 -8.31014812e-01 4.65061486e-01
9.94553983e-01 -1.46015376e-01 -3.21479887e-01 2.34805956e-01
6.13950610e-01 -5.61165094e-01 -7.51065075e-01 3.65533590e-01
1.71913937e-01 -9.67590868e-01 9.52719212e-01 -7.31489003e-01
4.92158502e-01 -5.62515378e-01 3.98025334e-01 -1.10555029e+00
6.07256182e-02 -1.12630284e+00 -3.88371885e-01 1.10992324e+00
3.51573110e-01 -4.25555408e-01 1.05133808e+00 5.55912912e-01
-1.57986224e-01 -4.91085857e-01 -7.12837815e-01 -7.61804521e-01
8.31237882e-02 -5.18728971e-01 3.56988609e-02 1.00739717e+00
-6.07550979e-01 1.44264221e-01 -3.71634096e-01 1.42447293e-01
7.72933960e-01 -2.62291096e-02 1.08931518e+00 -1.00314188e+00
-5.41588724e-01 -6.54594839e-01 -6.25124931e-01 -1.46865845e+00
5.34316421e-01 -9.43155706e-01 1.24201342e-01 -1.47456861e+00
-2.65179634e-01 -9.24487412e-02 -2.53080111e-02 2.69013435e-01
-2.42925361e-01 1.94605231e-01 2.68438727e-01 1.92736223e-01
-5.21620631e-01 5.36733866e-01 1.55326104e+00 -9.52035859e-02
-3.88128340e-01 2.29191765e-01 7.09920526e-02 8.52724075e-01
5.89209437e-01 -3.19999695e-01 -4.64673668e-01 -5.31212509e-01
-4.40929793e-02 7.77759552e-02 8.11317801e-01 -1.06824601e+00
4.98228222e-01 2.29952857e-02 6.04621880e-02 -3.75579089e-01
4.45365869e-02 -6.29672289e-01 -1.19661734e-01 4.88800645e-01
-1.27655149e-01 -2.50060380e-01 4.11719419e-02 4.80637014e-01
-3.17831576e-01 -4.76278722e-01 7.01398075e-01 -8.46737474e-02
-8.37753296e-01 4.87821966e-01 -3.11581582e-01 2.93369740e-01
7.23033786e-01 -1.09842151e-01 2.82373447e-02 -2.28437811e-01
-1.13019931e+00 3.48844290e-01 5.07563889e-01 1.59218445e-01
4.25071120e-01 -1.10036933e+00 -6.68253601e-01 -6.41230196e-02
-2.11535096e-01 4.53332961e-01 1.43613458e-01 1.22308862e+00
-1.08285367e+00 3.58965665e-01 -1.82409972e-01 -9.85447884e-01
-6.96479678e-01 6.87026799e-01 6.08079195e-01 -2.75000960e-01
-5.69074214e-01 5.77797890e-01 2.63112038e-02 -2.72589713e-01
1.55690163e-01 -6.64515316e-01 -2.27696374e-01 -1.34685144e-01
3.34645599e-01 5.22688746e-01 -7.90517330e-02 -6.60572827e-01
-1.95212334e-01 7.73128569e-01 2.52824545e-01 -3.14995825e-01
9.59235549e-01 -2.21949011e-01 -9.89885479e-02 3.69036883e-01
1.46382284e+00 -2.39527032e-01 -1.72199535e+00 -1.96697302e-02
3.24256599e-01 -3.46295595e-01 -4.31514159e-02 -1.63868085e-01
-1.44567084e+00 9.34185088e-01 2.58987784e-01 1.14599019e-01
9.89203870e-01 -3.48485887e-01 9.85673487e-01 3.84907633e-01
4.02164273e-02 -8.66175950e-01 3.14688422e-02 5.02911687e-01
3.18473637e-01 -1.29553628e+00 -4.18159932e-01 -3.81259561e-01
-3.23839575e-01 1.38451576e+00 4.06834215e-01 -5.22156239e-01
6.34643853e-01 9.75077450e-02 6.45289347e-02 1.85225770e-01
-4.91564244e-01 -4.74732399e-01 4.64645535e-01 3.56720656e-01
3.32868218e-01 -3.70387197e-01 -3.42801303e-01 -1.33069694e-01
3.34968301e-03 3.56548548e-01 7.24337697e-01 7.89502025e-01
-2.38120914e-01 -1.20972574e+00 -9.29168761e-02 -2.19595984e-01
-5.18837035e-01 2.16190234e-01 -3.28504652e-01 8.61917019e-01
6.83559403e-02 9.60011303e-01 1.10544473e-01 -1.30037546e-01
2.15821475e-01 2.02288665e-03 4.72043604e-01 -4.10650462e-01
-3.90266925e-01 5.35656475e-02 -1.74632952e-01 -7.93476880e-01
-9.23063278e-01 -6.90815330e-01 -1.39955401e+00 -1.20791480e-01
-1.76568508e-01 -1.10021969e-02 5.71655869e-01 1.23774552e+00
6.19874820e-02 3.76260132e-01 6.39338374e-01 -1.09626675e+00
-9.22379121e-02 -7.98332930e-01 -8.39297399e-02 6.87147021e-01
7.24075615e-01 -4.47421759e-01 -4.72421288e-01 5.06730199e-01] | [9.084859848022461, -0.4034520387649536] |
779e0422-22b4-44b3-930e-83b89e221ca9 | ordinal-time-series-analysis-with-the-r | 2304.12251 | null | https://arxiv.org/abs/2304.12251v1 | https://arxiv.org/pdf/2304.12251v1.pdf | Ordinal time series analysis with the R package otsfeatures | The 21st century has witnessed a growing interest in the analysis of time series data. Whereas most of the literature on the topic deals with real-valued time series, ordinal time series have typically received much less attention. However, the development of specific analytical tools for the latter objects has substantially increased in recent years. The R package otsfeatures attempts to provide a set of simple functions for analyzing ordinal time series. In particular, several commands allowing the extraction of well-known statistical features and the execution of inferential tasks are available for the user. The output of several functions can be employed to perform traditional machine learning tasks including clustering, classification or outlier detection. otsfeatures also incorporates two datasets of financial time series which were used in the literature for clustering purposes, as well as three interesting synthetic databases. The main properties of the package are described and its use is illustrated through several examples. Researchers from a broad variety of disciplines could benefit from the powerful tools provided by otsfeatures. | ['José Antonio Vilar Fernández', 'Ángel López Oriona'] | 2023-04-24 | null | null | null | null | ['outlier-detection'] | ['methodology'] | [-2.05223784e-01 -3.92484307e-01 4.79034707e-02 -5.89934409e-01
-4.53683227e-01 -6.80860996e-01 6.91759527e-01 7.78011620e-01
-3.22616756e-01 5.08991957e-01 -2.53503054e-01 -4.97530162e-01
-5.73946595e-01 -6.65846109e-01 -6.30326718e-02 -7.22829640e-01
-7.62163281e-01 4.25891906e-01 5.58863319e-02 -1.27370775e-01
5.25346160e-01 6.78434312e-01 -1.99613047e+00 -6.26114830e-02
8.04023266e-01 1.27272928e+00 -1.20897338e-01 3.33966315e-01
-1.35092080e-01 3.70888144e-01 -8.02245498e-01 -1.11268230e-01
2.24299043e-01 -2.01115891e-01 -3.03550452e-01 -5.51900230e-02
-4.90094781e-01 1.01755075e-01 1.41610458e-01 7.43968129e-01
2.50099719e-01 3.20296288e-01 6.16322637e-01 -1.50638938e+00
-8.11956227e-02 4.51985568e-01 -3.55148703e-01 2.28366241e-01
4.19746906e-01 -1.56653840e-02 8.76811028e-01 -6.20699465e-01
1.80579081e-01 9.47461963e-01 6.10922515e-01 -4.64100540e-01
-1.23684263e+00 -5.48040509e-01 -4.90549207e-01 1.80481866e-01
-1.23771226e+00 -7.06688985e-02 8.25293720e-01 -4.91577685e-01
9.06052709e-01 3.97927493e-01 6.73555315e-01 6.61978662e-01
1.67342767e-01 4.35032815e-01 1.46880114e+00 -3.72644365e-01
5.01506567e-01 2.84946598e-02 2.38238573e-01 1.27264574e-01
1.81835234e-01 2.34381512e-01 -1.62910402e-01 -5.35133600e-01
7.58691430e-01 2.30219766e-01 2.87604272e-01 -1.56569511e-01
-1.32739711e+00 8.11649024e-01 -7.05986246e-02 8.07477176e-01
-4.23501223e-01 -2.42241308e-01 7.72511065e-01 7.57907212e-01
6.84870601e-01 7.29373753e-01 -5.34901321e-01 -6.33427501e-01
-9.54945683e-01 4.26540166e-01 8.79707277e-01 5.45784593e-01
4.34723914e-01 6.09716699e-02 3.28359485e-01 7.36581683e-01
4.52459650e-03 1.83767125e-01 5.17028868e-01 -9.93126869e-01
1.19138427e-01 6.98339283e-01 1.95810005e-01 -1.09850025e+00
-5.85396051e-01 -1.92409739e-01 -6.80094421e-01 2.00123325e-01
8.25977564e-01 -2.48818826e-02 -4.30400401e-01 1.17090392e+00
2.14996427e-01 -2.00760424e-01 -1.72361419e-01 4.65561867e-01
3.96113306e-01 6.33619249e-01 -1.52099863e-01 -6.33854747e-01
1.02588379e+00 -1.03720384e-04 -6.88576221e-01 4.80754882e-01
4.80697870e-01 -8.20403457e-01 9.29569244e-01 7.36029625e-01
-9.65446532e-01 -3.36667746e-01 -6.50635660e-01 2.51027584e-01
-6.93796635e-01 -3.73360544e-01 8.60915363e-01 5.18087804e-01
-8.64819527e-01 9.23356533e-01 -1.01976264e+00 -4.05122072e-01
1.46691412e-01 3.95925343e-01 -2.86000043e-01 2.77437627e-01
-1.04235470e+00 7.87721097e-01 4.24414098e-01 -7.20472038e-02
-1.41097203e-01 -4.34840947e-01 -7.45635390e-01 -1.12810731e-01
2.97295839e-01 -2.16807157e-01 1.05124652e+00 -9.47940350e-01
-1.22022355e+00 7.12192833e-01 1.83930889e-01 -2.14692459e-01
6.18190885e-01 1.02207184e-01 -7.16960073e-01 3.89738679e-02
2.33445063e-01 -2.13130787e-01 8.00969601e-01 -3.68030459e-01
-4.94910538e-01 -4.50790644e-01 -2.76220322e-01 -3.06628376e-01
-3.49210858e-01 5.66451073e-01 9.47975963e-02 -1.02060807e+00
5.76464040e-03 -6.06685281e-01 -3.01334023e-01 -2.82652527e-01
-2.13266537e-01 -5.58422804e-01 6.05070829e-01 -3.12756628e-01
1.41481864e+00 -2.47318816e+00 2.03043930e-02 5.04958332e-01
-2.18463987e-01 -3.78245026e-01 3.70761782e-01 9.60921228e-01
-6.37197614e-01 1.63994744e-01 -3.65387827e-01 -8.72502029e-02
3.18945938e-04 2.64986426e-01 -3.41891468e-01 7.60336816e-01
8.88319090e-02 3.15996200e-01 -1.07290697e+00 -2.98327953e-01
5.18901229e-01 -9.18654725e-02 -1.50220290e-01 2.24850550e-01
-6.73157498e-02 3.51332396e-01 -3.64588737e-01 7.29962707e-01
1.95546240e-01 -1.95906460e-01 4.76318933e-02 4.10574228e-01
-7.75129199e-01 2.02537447e-01 -1.05002916e+00 1.40438557e+00
9.75829810e-02 4.81641233e-01 -3.13090652e-01 -1.15086496e+00
1.13256073e+00 4.49926883e-01 1.02076256e+00 -3.90044838e-01
2.91245162e-01 4.38977987e-01 1.97028890e-01 -5.44556141e-01
4.47577327e-01 -2.43608639e-01 -3.62727612e-01 6.94879472e-01
-3.32040519e-01 -3.91530454e-01 5.24884403e-01 -2.25454643e-01
1.16691101e+00 6.79124668e-02 7.22696483e-01 -4.44633812e-01
3.71881157e-01 2.73676813e-01 3.14581215e-01 2.94714481e-01
-1.20330870e-01 2.18190774e-01 6.15491092e-01 -5.53422153e-01
-1.14819515e+00 -1.04988897e+00 -6.00982428e-01 9.72610533e-01
-4.18279856e-01 -4.54765648e-01 -2.11659044e-01 7.33884517e-03
2.98235208e-01 7.81139553e-01 -6.68940425e-01 2.64194041e-01
-2.16834903e-01 -8.28519523e-01 3.20324957e-01 3.00223947e-01
-2.11924706e-02 -1.37078476e+00 -7.71554649e-01 1.69854924e-01
2.07758218e-01 -6.89502001e-01 1.88943252e-01 3.19073886e-01
-1.35514462e+00 -9.52921927e-01 -4.11462218e-01 -2.92091817e-01
4.76935416e-01 8.73364881e-02 1.02411592e+00 -1.65670753e-01
-5.01353562e-01 3.88455749e-01 -5.89287817e-01 -7.53062665e-01
-3.54171365e-01 7.74284899e-02 2.04018563e-01 -1.19801991e-01
5.24760604e-01 -9.33587074e-01 -2.04123199e-01 3.85295540e-01
-1.08903933e+00 -4.33962077e-01 1.03702679e-01 6.03591084e-01
4.60621327e-01 4.99853492e-01 9.46339786e-01 -3.75447214e-01
7.72641182e-01 -1.02068424e+00 -8.23514044e-01 -2.06069693e-01
-4.47152823e-01 -1.65481195e-01 9.77410078e-01 -2.51959413e-01
-3.42364043e-01 -2.29291335e-01 6.08944595e-02 -4.04012889e-01
-4.31674480e-01 9.52426434e-01 2.60721743e-01 4.67730999e-01
4.41700101e-01 8.24977830e-02 1.78116634e-01 -5.01534045e-01
-9.24460357e-04 6.97674513e-01 5.05260348e-01 -6.89824224e-01
7.11412430e-01 3.30345035e-01 1.64152205e-01 -1.11573195e+00
-3.39924395e-01 -5.76529801e-01 -5.59544146e-01 -4.01671082e-01
4.29033488e-01 -4.00735795e-01 -7.25861073e-01 5.80905795e-01
-5.85480571e-01 1.74435209e-02 -4.81166095e-01 4.77081984e-01
-6.87799573e-01 1.84121281e-01 -1.88741833e-01 -1.08593166e+00
4.52991128e-02 -8.41164052e-01 6.35964096e-01 -4.04991694e-02
-6.92388415e-01 -1.17419398e+00 1.05254777e-01 -3.45240802e-01
1.69135258e-01 9.35554922e-01 1.06570089e+00 -1.01863325e+00
-2.39507794e-01 -7.79145479e-01 3.91341209e-01 1.33557141e-01
3.90128970e-01 5.55472136e-01 -7.47948050e-01 -2.38417551e-01
2.40410209e-01 -8.02872032e-02 2.85514712e-01 4.76166308e-01
1.43615484e+00 9.55335870e-02 -2.34892126e-02 4.04811203e-01
1.19842136e+00 5.56471467e-01 5.50991058e-01 6.03935182e-01
5.14167547e-02 1.10480726e+00 1.05708492e+00 8.91068935e-01
1.09994046e-01 4.97056574e-01 5.09558856e-01 9.02602151e-02
1.14076078e+00 2.65094459e-01 2.64615744e-01 9.50179994e-01
-4.30638403e-01 2.53159851e-01 -1.05703759e+00 6.05255365e-01
-1.90958786e+00 -1.24592662e+00 -4.37143445e-01 2.50890231e+00
6.18105292e-01 2.45039374e-01 6.16614223e-01 8.07828963e-01
5.30901790e-01 1.60528645e-01 -4.60223973e-01 -5.35274506e-01
3.51641104e-02 1.82561055e-01 1.46920487e-01 -1.87446885e-02
-1.07262039e+00 2.79720962e-01 6.98299026e+00 5.27233541e-01
-9.74247515e-01 -5.27560413e-01 5.80578625e-01 -7.21878782e-02
-5.72811253e-02 -6.34786487e-02 -2.62384471e-02 6.03153765e-01
1.21692848e+00 -7.54354894e-01 3.79324526e-01 9.21593904e-01
7.28810012e-01 -2.78830588e-01 -1.07953632e+00 9.93187904e-01
-5.42851985e-01 -8.54391396e-01 -4.77365345e-01 9.08096060e-02
4.12925750e-01 -7.78438300e-02 1.01096727e-01 1.65705651e-01
1.08658403e-01 -9.61048484e-01 6.07273817e-01 6.34385467e-01
7.15957820e-01 -1.16310012e+00 5.69117248e-01 2.82090217e-01
-1.02808511e+00 -3.27428639e-01 -1.27312168e-01 -5.63776195e-01
3.03623229e-02 1.04229021e+00 -6.23932064e-01 9.01358128e-01
8.98756325e-01 1.14885581e+00 -5.30163288e-01 1.33841956e+00
1.93397641e-01 6.32518411e-01 -5.21186590e-01 -9.91194248e-02
3.63911539e-02 -6.09449685e-01 6.38411880e-01 8.03448796e-01
5.61725974e-01 -6.54381439e-02 2.41552889e-02 5.24203539e-01
6.08605564e-01 3.64913344e-01 -7.50699461e-01 -5.07118642e-01
4.57159102e-01 1.17983544e+00 -9.49452996e-01 -2.51864910e-01
-4.97786939e-01 2.89111346e-01 -6.29202500e-02 9.62352455e-02
-8.37843537e-01 -7.39070892e-01 7.42657006e-01 2.84487575e-01
-7.36711100e-02 -5.48274398e-01 -4.75584179e-01 -9.62614119e-01
-2.62954049e-02 -1.06835568e+00 6.67437971e-01 -6.31701171e-01
-1.43543756e+00 3.99807066e-01 4.53515440e-01 -1.39694452e+00
-7.26043522e-01 -5.61573982e-01 -7.09102333e-01 8.07087302e-01
-7.65838146e-01 -9.07723457e-02 -1.66113436e-01 6.76786482e-01
2.90915668e-01 -1.66397467e-01 9.38290656e-01 1.71596125e-01
-6.28792048e-01 -1.40961185e-01 5.12495458e-01 -5.37740216e-02
7.17165053e-01 -1.41136885e+00 3.50975722e-01 5.09579420e-01
-3.19114715e-01 8.25607121e-01 1.12024438e+00 -6.28618658e-01
-1.11031687e+00 -6.21051788e-01 8.47815990e-01 -3.13569754e-01
1.27559042e+00 -2.02499807e-01 -9.15129483e-01 4.70306009e-01
-1.43578246e-01 -3.05736989e-01 9.17879939e-01 7.69296810e-02
-9.95913148e-02 -1.80793047e-01 -1.02347565e+00 6.21986628e-01
4.41561490e-01 -3.93702745e-01 -8.20978522e-01 1.31073445e-01
1.84239358e-01 9.24223736e-02 -1.41106784e+00 2.44797423e-01
6.42107546e-01 -1.26107752e+00 8.19487631e-01 -5.66936255e-01
3.44308525e-01 -4.05100733e-01 4.15955260e-02 -1.13375068e+00
-4.17688815e-03 -8.76190484e-01 6.12157956e-02 1.05673707e+00
2.86506228e-02 -1.05927229e+00 1.83234677e-01 5.77033937e-01
-7.30938539e-02 -6.86008275e-01 -8.76201272e-01 -9.36552823e-01
-4.84088585e-02 -7.59346426e-01 7.19554663e-01 1.16029489e+00
5.18333256e-01 -1.76363885e-01 1.54827898e-02 -3.15702766e-01
6.36962593e-01 4.02895838e-01 8.16311955e-01 -1.80631936e+00
-1.79904774e-01 -6.09579086e-01 -4.88926470e-01 -2.98819929e-01
-1.23780519e-01 -7.97084391e-01 -3.82304549e-01 -1.13159883e+00
-3.65047812e-01 -3.92881334e-01 -3.12967032e-01 4.44224417e-01
7.14356154e-02 -1.40787035e-01 -1.36375651e-01 4.06294554e-01
-8.11829716e-02 3.34416240e-01 8.23340952e-01 4.81395692e-01
-3.56960595e-01 3.98464710e-01 -3.69656742e-01 7.81261086e-01
9.53828752e-01 -5.75667977e-01 -3.34094673e-01 2.79953688e-01
3.60469848e-01 3.11316073e-01 2.88740575e-01 -8.25546861e-01
-8.23513120e-02 -2.72983164e-01 3.05681437e-01 -7.68501699e-01
6.36644959e-02 -1.14676774e+00 5.66220939e-01 3.50451767e-01
-4.55764942e-02 8.06307435e-01 1.89494804e-01 3.53871554e-01
-4.73807842e-01 -1.73498392e-01 5.90151370e-01 -7.25120604e-02
-3.89845014e-01 1.23558879e-01 -5.00736475e-01 -2.44436949e-01
1.29585850e+00 -4.36675727e-01 -1.82723701e-01 -4.87858862e-01
-7.60357320e-01 1.45513758e-01 6.83749199e-01 4.39273119e-01
3.28634351e-01 -1.12912762e+00 -4.81212378e-01 1.34909555e-01
1.76495910e-01 -1.82499141e-01 -2.25403845e-01 1.22809958e+00
-5.28594673e-01 3.86096865e-01 -4.87486422e-01 -6.01885855e-01
-1.01793230e+00 7.36821592e-01 8.54455382e-02 -1.80399999e-01
-6.27663851e-01 -3.03359479e-02 -3.77751350e-01 -1.56587556e-01
3.22555415e-02 -5.50462961e-01 -1.74954548e-01 5.93052566e-01
5.30104101e-01 7.89564610e-01 -1.02603845e-01 -3.47430944e-01
-4.03695315e-01 1.98951110e-01 2.87778318e-01 -1.61580563e-01
1.95853758e+00 7.49830455e-02 -5.18944681e-01 1.41517270e+00
9.87582862e-01 -1.71248704e-01 -8.11627746e-01 1.54210389e-01
6.46315098e-01 -4.91375923e-01 -4.76070046e-01 -4.74269241e-01
-4.71666336e-01 6.29854560e-01 9.31870490e-02 1.23183870e+00
1.48346925e+00 -2.82215506e-01 1.29074275e-01 7.33052269e-02
6.70872927e-01 -9.88202453e-01 -2.10461929e-01 5.63047588e-01
9.00554419e-01 -9.84135032e-01 8.48725140e-02 -9.68034491e-02
-3.30330878e-01 1.48783970e+00 -2.01723829e-01 -3.75844091e-01
6.97350621e-01 3.40154469e-01 -1.47420108e-01 -2.08088741e-01
-6.82238698e-01 -1.12203889e-01 8.09515193e-02 3.77594292e-01
7.97566533e-01 -7.73684829e-02 -4.46780413e-01 4.79101241e-01
-5.35653710e-01 7.71610737e-02 5.66460311e-01 9.09085572e-01
-3.73924673e-01 -1.10646665e+00 -7.10573792e-01 7.97095478e-01
-7.84744740e-01 3.52494535e-03 -3.29935193e-01 1.00055671e+00
-2.71781385e-01 1.08308613e+00 2.80015290e-01 -1.25776172e-01
2.51429349e-01 1.51688755e-01 7.23020136e-02 -4.58545446e-01
-9.58278596e-01 2.52517670e-01 8.81879032e-02 -5.04170001e-01
-6.46654129e-01 -1.31599021e+00 -1.10752141e+00 -5.04386127e-01
-4.76660430e-02 3.90638053e-01 7.00020552e-01 7.86917448e-01
-1.70888305e-01 5.06014109e-01 8.34402978e-01 -8.49602342e-01
-3.80565733e-01 -1.03844619e+00 -1.12573099e+00 2.84442872e-01
1.39590055e-01 -6.08854353e-01 -6.41765535e-01 -1.34603202e-01] | [7.2028350830078125, 3.3841192722320557] |
db9b0f86-5eef-4820-a735-66d7fb54fc5e | multiview-detection-with-feature-perspective | 2007.07247 | null | https://arxiv.org/abs/2007.07247v2 | https://arxiv.org/pdf/2007.07247v2.pdf | Multiview Detection with Feature Perspective Transformation | Incorporating multiple camera views for detection alleviates the impact of occlusions in crowded scenes. In a multiview system, we need to answer two important questions when dealing with ambiguities that arise from occlusions. First, how should we aggregate cues from the multiple views? Second, how should we aggregate unreliable 2D and 3D spatial information that has been tainted by occlusions? To address these questions, we propose a novel multiview detection system, MVDet. For multiview aggregation, existing methods combine anchor box features from the image plane, which potentially limits performance due to inaccurate anchor box shapes and sizes. In contrast, we take an anchor-free approach to aggregate multiview information by projecting feature maps onto the ground plane (bird's eye view). To resolve any remaining spatial ambiguity, we apply large kernel convolutions on the ground plane feature map and infer locations from detection peaks. Our entire model is end-to-end learnable and achieves 88.2% MODA on the standard Wildtrack dataset, outperforming the state-of-the-art by 14.1%. We also provide detailed analysis of MVDet on a newly introduced synthetic dataset, MultiviewX, which allows us to control the level of occlusion. Code and MultiviewX dataset are available at https://github.com/hou-yz/MVDet. | ['Yunzhong Hou', 'Stephen Gould', 'Liang Zheng'] | 2020-07-14 | null | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/116_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123520001.pdf | eccv-2020-8 | ['multiview-detection'] | ['computer-vision'] | [-1.04559429e-01 -1.77181408e-01 2.83766508e-01 -4.41433400e-01
-1.07830441e+00 -6.86239779e-01 3.26897353e-01 4.06041406e-02
-4.30406362e-01 3.51502925e-01 1.69675469e-01 8.79959986e-02
3.48742962e-01 -5.00443101e-01 -9.05320883e-01 -3.51649612e-01
1.03420012e-01 1.51239857e-01 7.17703819e-01 3.33106215e-03
2.46678963e-01 2.27988034e-01 -1.86198235e+00 3.69863123e-01
7.42311120e-01 1.00542498e+00 2.89217830e-01 7.19099045e-01
2.98967630e-01 5.54252028e-01 -6.51196659e-01 -4.18434113e-01
5.54673493e-01 3.54788035e-01 -4.84934337e-02 3.13895017e-01
1.29916298e+00 -8.38031948e-01 -3.06678623e-01 9.56004083e-01
5.97896397e-01 3.95388156e-02 3.97168428e-01 -1.22321153e+00
-4.11178052e-01 -2.68187135e-01 -1.16595697e+00 3.13822687e-01
6.40542150e-01 5.90610862e-01 9.76031959e-01 -1.56142616e+00
6.08232141e-01 1.35940969e+00 4.85243171e-01 6.64926693e-02
-1.12089157e+00 -5.78722298e-01 5.33777297e-01 -3.71324383e-02
-1.47810268e+00 -6.62752092e-01 5.65583587e-01 -7.82816947e-01
7.57956505e-01 2.33814985e-01 4.02181715e-01 8.59802127e-01
3.07403505e-02 8.80695224e-01 8.98464620e-01 -1.90199718e-01
-6.72108829e-02 -5.86943291e-02 2.55829990e-02 7.23139524e-01
3.77995938e-01 1.12187147e-01 -9.51298535e-01 -2.31014132e-01
5.19395709e-01 2.31407031e-01 -3.32536191e-01 -6.55219674e-01
-1.24390650e+00 7.75502384e-01 3.69139761e-01 -4.89836991e-01
-9.47359353e-02 -3.83669022e-03 1.13295630e-01 -8.52170214e-02
5.02026677e-01 1.77735195e-01 -2.25829542e-01 2.23059028e-01
-7.30677724e-01 4.52594191e-01 2.58359075e-01 1.18110800e+00
9.54968870e-01 -5.17539680e-02 -4.92362529e-02 7.12110281e-01
4.24079388e-01 7.37091541e-01 -1.66421473e-01 -1.32394195e+00
9.05822814e-01 4.72848058e-01 3.97564858e-01 -1.04047084e+00
-3.85520518e-01 -2.36534655e-01 -3.77920985e-01 5.29983819e-01
6.24795735e-01 -2.61393368e-01 -7.07603633e-01 1.43785620e+00
6.90013885e-01 -8.08305666e-02 -1.83364466e-01 1.20473874e+00
1.03533351e+00 3.47768426e-01 -3.98594826e-01 1.90179750e-01
1.48390973e+00 -1.00015056e+00 -4.78516608e-01 -7.15266228e-01
3.18422526e-01 -9.27159309e-01 9.58086848e-01 3.58228803e-01
-9.86012936e-01 -5.07709324e-01 -1.22453034e+00 -2.11926296e-01
-2.81550169e-01 4.09303397e-01 2.84933448e-01 3.38860303e-01
-8.59668314e-01 -7.77914829e-04 -7.27191448e-01 -2.51342297e-01
2.70607471e-01 1.54380471e-01 -5.23790896e-01 -2.83363670e-01
-8.94416869e-01 6.75806284e-01 -1.35860220e-01 9.99555066e-02
-7.54330695e-01 -7.23210037e-01 -1.22536564e+00 -1.54178232e-01
7.88567603e-01 -7.85103142e-01 1.15484905e+00 -3.80756259e-01
-9.12447989e-01 7.96403289e-01 -6.10320985e-01 -1.86521932e-01
7.30020225e-01 -6.33199215e-01 -1.86658397e-01 2.69793153e-01
5.81559062e-01 5.57688117e-01 8.21390510e-01 -1.39591360e+00
-8.46039355e-01 -6.20513141e-01 6.69364929e-02 3.39959413e-01
1.15684398e-01 -3.16724926e-02 -8.05938482e-01 -5.73647499e-01
3.21472496e-01 -7.96580553e-01 -3.97149809e-02 3.81808132e-01
-3.80317122e-01 2.20137045e-01 9.14151549e-01 -5.96197963e-01
7.39894867e-01 -2.31805873e+00 -1.49543077e-01 -1.32310152e-01
6.44527435e-01 7.63433203e-02 3.30294408e-02 4.43260401e-01
2.22140491e-01 -1.06081523e-01 -1.55402115e-02 -6.74416780e-01
-1.77466914e-01 -9.63184051e-03 -3.72073710e-01 7.90575624e-01
7.58758262e-02 6.25638604e-01 -6.75534725e-01 -4.78629619e-01
4.99380529e-01 4.05412525e-01 -6.45856440e-01 3.60211402e-01
1.84582248e-02 3.02026153e-01 -3.03964078e-01 1.00929391e+00
1.03381169e+00 -2.87190467e-01 -2.37450570e-01 -3.64905447e-01
-3.52593362e-01 -1.31086245e-01 -1.49135613e+00 1.66167426e+00
-1.24641687e-01 8.61114383e-01 2.66298056e-01 -3.28434706e-01
8.42813015e-01 -2.25018635e-02 2.10935220e-01 -4.99805987e-01
-1.55317530e-01 1.00443155e-01 -4.30860937e-01 -4.44062680e-01
7.17202723e-01 3.03691179e-01 5.28101409e-05 9.19795036e-02
5.06614847e-03 -1.03019802e-02 6.82202652e-02 2.77698815e-01
1.04252934e+00 2.18350887e-02 4.06679899e-01 9.70537439e-02
2.63445765e-01 -1.87954947e-01 7.70373106e-01 8.26613069e-01
-3.48763764e-01 1.21988094e+00 4.73784328e-01 -6.30784988e-01
-8.78088236e-01 -1.42890024e+00 2.06727125e-02 8.51400256e-01
5.26655376e-01 -5.88779509e-01 -3.46279889e-01 -5.74369788e-01
1.66522935e-01 2.48619780e-01 -5.84923983e-01 4.90289539e-01
-5.73635757e-01 -4.73051667e-01 1.16312809e-01 5.08935392e-01
4.32773650e-01 -2.63643950e-01 -1.06969762e+00 -1.61070079e-01
-4.69744265e-01 -1.49051452e+00 -7.74461031e-01 -1.48203686e-01
-3.92065585e-01 -1.35342908e+00 -6.63608730e-01 -1.79260205e-02
5.30357003e-01 9.20741200e-01 1.06851065e+00 -7.45099857e-02
-4.22883928e-01 6.47905648e-01 -1.95889786e-01 -3.63026589e-01
2.94698685e-01 -2.13756174e-01 1.11461841e-01 1.39765978e-01
2.78668642e-01 -4.79146451e-01 -8.52301180e-01 4.96831983e-01
-5.76670468e-01 -1.07166441e-02 3.77847522e-01 7.20998347e-01
6.47082031e-01 -4.95957792e-01 1.85084313e-01 -5.66870093e-01
5.21776415e-02 -2.41784900e-01 -1.13776767e+00 3.30977999e-02
-7.40715489e-02 -1.90259576e-01 2.62060791e-01 -1.16348378e-01
-1.03282952e+00 3.79151016e-01 1.19298398e-01 -7.83040583e-01
-2.59636700e-01 -7.97208305e-03 -3.21007758e-01 -2.28626858e-02
7.05166996e-01 -6.94907382e-02 -7.00236633e-02 -3.03038746e-01
3.84509087e-01 5.52075446e-01 7.47768104e-01 -4.83869612e-01
8.09389651e-01 8.89288843e-01 -9.30594280e-02 -1.13836586e+00
-1.01537144e+00 -7.36219883e-01 -6.24767303e-01 -4.56000984e-01
1.02319920e+00 -1.45270753e+00 -8.69716763e-01 3.12193036e-01
-1.44049382e+00 -9.83539410e-03 1.88424513e-01 4.48096991e-01
-4.88740534e-01 5.69090307e-01 -2.85546154e-01 -8.97076249e-01
-1.45724460e-01 -1.40759838e+00 1.57265460e+00 9.82248411e-02
-1.56945735e-01 -4.32319224e-01 -2.46785805e-01 6.30896926e-01
8.70559886e-02 5.95656097e-01 2.12270990e-01 -2.95848429e-01
-1.03462565e+00 -2.56053627e-01 -2.58932799e-01 8.55273604e-02
-2.21027330e-01 -3.61278467e-03 -1.43615556e+00 -3.64011884e-01
-1.19630702e-01 -2.54987001e-01 9.95507300e-01 4.47931975e-01
7.03691602e-01 -9.46165323e-02 -3.57041895e-01 8.66869628e-01
1.36794436e+00 -1.87995717e-01 2.94584125e-01 4.11660671e-01
7.99123168e-01 7.14389801e-01 8.38437378e-01 6.59609497e-01
8.59737754e-01 1.00418055e+00 7.50395298e-01 -1.32528484e-01
-6.87739402e-02 -2.98195034e-01 2.51761585e-01 1.61017075e-01
1.30603254e-01 -2.94983983e-01 -9.57847416e-01 3.96349341e-01
-1.98173642e+00 -7.58238673e-01 -3.24651092e-01 2.41593242e+00
1.86802387e-01 1.35785088e-01 1.47480354e-01 -2.28024900e-01
7.32974648e-01 5.80115855e-01 -5.38269222e-01 3.30296159e-01
-2.96064734e-01 -5.49289227e-01 6.96486712e-01 8.26783955e-01
-1.38379204e+00 7.36132503e-01 5.53249502e+00 4.30013895e-01
-9.55985069e-01 6.13143481e-02 4.48893309e-01 -3.54815602e-01
1.02010639e-02 2.33704434e-03 -1.17087901e+00 3.85331005e-01
7.69724324e-02 3.04451287e-01 1.93884388e-01 8.73332083e-01
1.24427065e-01 -5.57790518e-01 -1.10368669e+00 1.23763037e+00
3.37293535e-01 -1.26210630e+00 -3.32883447e-01 1.30185321e-01
4.77646798e-01 4.31870431e-01 1.14180431e-01 -1.02145649e-01
3.53958935e-01 -5.73035121e-01 9.37809169e-01 4.61897999e-01
5.83646655e-01 -5.89511156e-01 4.67018753e-01 3.88384998e-01
-1.42313993e+00 -2.32070070e-02 -2.86338657e-01 -8.84000063e-02
4.11020398e-01 6.44454539e-01 -6.80926502e-01 4.76096600e-01
9.45955634e-01 6.43857300e-01 -8.86018157e-01 1.14974070e+00
-2.59314001e-01 4.89074886e-02 -4.40319777e-01 5.07659018e-01
-4.68131565e-02 9.22262222e-02 8.93564224e-01 1.02020407e+00
2.85185218e-01 -1.01428673e-01 2.92906463e-01 8.55171800e-01
1.51897177e-01 -3.01457196e-01 -8.82406533e-01 8.26972485e-01
6.87536478e-01 1.16816187e+00 -4.76963073e-01 -2.70349324e-01
-7.85828054e-01 9.27409887e-01 4.41307932e-01 7.18735337e-01
-7.72856891e-01 -2.23291516e-01 7.76423752e-01 3.70312482e-01
5.18084288e-01 -4.29299712e-01 -2.54332513e-01 -1.51412213e+00
6.44213557e-01 -8.08063328e-01 3.47076565e-01 -9.02321458e-01
-1.09470832e+00 4.42915201e-01 1.12790793e-01 -1.49426246e+00
-2.30892792e-01 -5.06727755e-01 -5.64279020e-01 9.15725887e-01
-1.34249330e+00 -1.09038377e+00 -5.92741013e-01 4.42951262e-01
6.06691599e-01 -6.35447120e-03 3.27870131e-01 3.03460598e-01
-3.46294433e-01 5.38296819e-01 -1.84243485e-01 5.17095849e-02
1.16590357e+00 -1.11949766e+00 5.60997903e-01 1.20459783e+00
-9.62245539e-02 4.53069448e-01 7.80261397e-01 -4.62111354e-01
-1.48634827e+00 -1.00101936e+00 6.32230997e-01 -8.92128706e-01
2.40011737e-01 -6.71722829e-01 -8.41939747e-01 7.90433943e-01
7.77913183e-02 5.62942445e-01 4.76695091e-01 1.22999335e-02
-5.92148423e-01 -2.26944923e-01 -9.04857278e-01 5.48543155e-01
8.76151383e-01 -3.10680658e-01 -3.18629235e-01 1.30234882e-01
8.35340381e-01 -7.67096877e-01 -4.79080468e-01 4.79536772e-01
6.84948564e-01 -1.39342666e+00 1.27035284e+00 1.63404047e-02
1.92585334e-01 -7.55541146e-01 -5.80495059e-01 -1.17115974e+00
-2.16924064e-02 -3.92836213e-01 -7.36935213e-02 1.09039092e+00
2.47063309e-01 -7.79039621e-01 5.29250205e-01 6.31487548e-01
-1.19527541e-01 -6.38718247e-01 -9.84269679e-01 -6.78603113e-01
-3.78426015e-01 -4.67166662e-01 3.84531260e-01 6.13311231e-01
-3.12798470e-01 4.99736041e-01 -6.28261924e-01 5.85681200e-01
9.82121468e-01 2.15925023e-01 1.42906344e+00 -1.16883361e+00
-4.24820065e-01 -3.45869586e-02 -3.32077414e-01 -1.28085256e+00
-2.42150813e-01 -2.23868638e-01 -8.99687931e-02 -1.43255305e+00
3.12262457e-02 -6.11403920e-02 4.89507496e-01 1.07351646e-01
-3.80183876e-01 2.34175310e-01 6.36426747e-01 2.15291411e-01
-9.07449543e-01 5.77311754e-01 1.04023862e+00 1.16421375e-02
-1.65195353e-02 -1.63493112e-01 -4.89809304e-01 1.01387239e+00
5.38154781e-01 -2.89310306e-01 -1.74133316e-01 -8.50597560e-01
1.97131380e-01 3.58186662e-01 5.81229687e-01 -1.03452134e+00
4.79452550e-01 6.98559284e-02 6.36901915e-01 -1.15284109e+00
8.27382386e-01 -6.24767184e-01 -2.01083258e-01 -1.51487598e-02
8.49951208e-02 3.64458710e-01 2.00582430e-01 7.12286592e-01
-2.53930092e-01 2.60393560e-01 7.06299484e-01 -2.14882106e-01
-5.72824299e-01 3.09645206e-01 -3.57995093e-01 2.70131379e-01
1.01354933e+00 -9.06027183e-02 -6.92390859e-01 -4.80070353e-01
-5.11044681e-01 6.33059740e-01 7.37866521e-01 5.12402952e-01
8.72777343e-01 -1.14286017e+00 -6.42685950e-01 3.56877446e-01
4.44438636e-01 2.85704374e-01 4.95020211e-01 8.86057973e-01
-4.47140247e-01 2.55201459e-01 1.41864195e-01 -1.00513732e+00
-1.40363503e+00 6.36887729e-01 3.73177022e-01 1.95412621e-01
-6.95059538e-01 6.62261307e-01 8.38200152e-01 -3.82499933e-01
3.53271097e-01 -2.66711205e-01 6.80278316e-02 1.16813481e-01
7.90690482e-01 4.96539474e-01 -8.32485408e-02 -8.34917128e-01
-5.23309529e-01 8.32544386e-01 -9.51748714e-03 -2.21129462e-01
1.03540218e+00 -4.42132741e-01 2.98869491e-01 2.89721429e-01
1.04580152e+00 4.17097986e-01 -1.72975802e+00 -3.56297523e-01
-4.41917568e-01 -8.95503998e-01 -2.90608585e-01 -4.03800130e-01
-9.42915201e-01 1.09315789e+00 7.06942677e-01 -1.65722266e-01
8.02335501e-01 6.28625154e-02 3.15522760e-01 3.14023912e-01
2.26016223e-01 -8.28245997e-01 1.89664498e-01 3.97440225e-01
1.01733875e+00 -1.63933575e+00 1.93117440e-01 -5.86954355e-01
-7.03064919e-01 1.02245641e+00 8.94916952e-01 4.11518961e-02
4.99356538e-01 3.69947672e-01 2.82195210e-01 -4.58706617e-01
-6.70076907e-01 -3.16033483e-01 1.05406925e-01 5.99243701e-01
1.86377138e-01 -1.23145133e-02 2.06945002e-01 3.52193624e-01
-2.03597248e-02 -3.97777945e-01 5.33746898e-01 8.88540804e-01
-6.02176011e-01 -5.52527070e-01 -8.50705028e-01 2.46852130e-01
-3.47060084e-01 -1.27754835e-02 -2.85812587e-01 8.03036809e-01
1.64733633e-01 1.16206419e+00 -2.28747670e-02 -4.68794972e-01
5.86961210e-01 -3.22046757e-01 1.57398790e-01 -5.67406058e-01
-6.83446154e-02 3.94613743e-01 1.37175858e-01 -8.50499034e-01
-2.03136981e-01 -8.76771867e-01 -7.73570657e-01 -5.74197993e-02
-2.53042281e-01 -3.36583793e-01 3.32479417e-01 6.76018417e-01
4.89923328e-01 3.21974993e-01 3.21244091e-01 -1.16591179e+00
-3.51154804e-01 -8.02616715e-01 -3.72399539e-01 6.95197359e-02
8.79888356e-01 -9.65015590e-01 -5.19393861e-01 -2.06300430e-02] | [7.901167392730713, -2.054199695587158] |
a1bcb897-7965-4bb5-8311-5bfa7bc29563 | self-training-improves-pre-training-for-few | 2108.12589 | null | https://arxiv.org/abs/2108.12589v1 | https://arxiv.org/pdf/2108.12589v1.pdf | Self-training Improves Pre-training for Few-shot Learning in Task-oriented Dialog Systems | As the labeling cost for different modules in task-oriented dialog (ToD) systems is expensive, a major challenge is to train different modules with the least amount of labeled data. Recently, large-scale pre-trained language models, have shown promising results for few-shot learning in ToD. In this paper, we devise a self-training approach to utilize the abundant unlabeled dialog data to further improve state-of-the-art pre-trained models in few-shot learning scenarios for ToD systems. Specifically, we propose a self-training approach that iteratively labels the most confident unlabeled data to train a stronger Student model. Moreover, a new text augmentation technique (GradAug) is proposed to better train the Student by replacing non-crucial tokens using a masked language model. We conduct extensive experiments and present analyses on four downstream tasks in ToD, including intent classification, dialog state tracking, dialog act prediction, and response selection. Empirical results demonstrate that the proposed self-training approach consistently improves state-of-the-art pre-trained models (BERT, ToD-BERT) when only a small number of labeled data are available. | ['Boi Faltings', 'Minlie Huang', 'Lingjing Kong', 'Fengyu Cai', 'Wanhao Zhou', 'Fei Mi'] | 2021-08-28 | null | https://aclanthology.org/2021.emnlp-main.142 | https://aclanthology.org/2021.emnlp-main.142.pdf | emnlp-2021-11 | ['text-augmentation'] | ['natural-language-processing'] | [ 5.56370616e-02 4.44361448e-01 -4.29101974e-01 -9.05749798e-01
-7.33693063e-01 -3.69173348e-01 6.20083988e-01 1.01990834e-01
-5.83488584e-01 8.22130322e-01 4.19051111e-01 -4.75867093e-01
5.24419725e-01 -3.67974699e-01 6.69915900e-02 -3.23952347e-01
4.35851574e-01 8.40913355e-01 5.74006617e-01 -6.44183576e-01
1.78656727e-02 -6.58946708e-02 -1.17576015e+00 2.75315046e-01
1.20474052e+00 7.31384039e-01 2.97187299e-01 5.91080666e-01
-5.45671821e-01 1.32866061e+00 -5.81569612e-01 -4.15136784e-01
-7.59401470e-02 -5.64827919e-01 -1.08907592e+00 4.26594317e-01
5.90740591e-02 -6.57046199e-01 -2.58699447e-01 7.10089147e-01
6.29303336e-01 6.43932819e-01 5.10633826e-01 -1.01148891e+00
-5.11385381e-01 9.28526759e-01 -2.78079957e-01 3.96700539e-02
4.46194783e-02 4.14628595e-01 8.92086089e-01 -1.10453844e+00
5.41616619e-01 1.30671775e+00 2.75326759e-01 1.22900951e+00
-1.19450617e+00 -3.50456566e-01 3.10016036e-01 1.80757511e-03
-8.32247674e-01 -6.88987613e-01 7.82971144e-01 -3.69003803e-01
8.24829578e-01 -1.05637550e-01 7.15413019e-02 1.11991429e+00
-1.40199959e-01 1.01173890e+00 1.15007925e+00 -7.20180154e-01
5.91276586e-01 7.22293377e-01 9.00853634e-01 7.06234455e-01
-4.96879458e-01 -3.22347879e-01 -5.74886501e-01 -3.07604104e-01
1.98018357e-01 1.40688241e-01 1.12674996e-01 -3.75890464e-01
-7.98278451e-01 1.15325749e+00 5.82262576e-02 3.92632574e-01
-1.71979621e-01 -6.63216293e-01 4.18564558e-01 4.42986667e-01
8.27828109e-01 3.84445995e-01 -6.17508054e-01 -1.63542703e-01
-5.40743113e-01 -8.28861594e-02 1.14088845e+00 1.11854219e+00
8.52614224e-01 1.44776836e-01 -6.80054307e-01 1.21464586e+00
3.34032893e-01 4.22168262e-02 8.21709871e-01 -5.89110911e-01
6.08616233e-01 8.87328982e-01 1.11936599e-01 2.49651313e-01
-3.19882214e-01 2.53032207e-01 -5.19697428e-01 -1.92011986e-02
2.66749591e-01 -8.56196880e-01 -9.59424675e-01 1.70761633e+00
5.87928534e-01 -2.37654224e-02 4.00236189e-01 6.44414127e-01
9.14368331e-01 6.02806628e-01 4.67946291e-01 -3.70276242e-01
1.43774152e+00 -1.34208202e+00 -9.33193922e-01 -5.40247202e-01
9.90411162e-01 -5.92577934e-01 1.28321707e+00 -1.20524071e-01
-9.88221765e-01 -7.21154809e-01 -8.57540309e-01 -2.48387363e-02
-3.20249349e-01 1.13646224e-01 5.53917587e-01 7.07991302e-01
-7.95042336e-01 2.41296828e-01 -5.39421499e-01 -5.14756322e-01
7.67475814e-02 3.77759218e-01 -2.73694079e-02 -2.12076353e-03
-1.33488357e+00 1.10652423e+00 3.27313006e-01 -2.47253954e-01
-1.01911485e+00 -3.98053408e-01 -1.12412167e+00 2.15334892e-01
6.46801531e-01 -2.45306298e-01 1.94146574e+00 -4.43384260e-01
-2.08501577e+00 7.34096110e-01 -3.23849052e-01 -5.87130368e-01
3.31470460e-01 -2.40174681e-01 -1.17288224e-01 -1.66841105e-01
-1.45075426e-01 8.22207868e-01 6.97525680e-01 -1.05623055e+00
-5.85569501e-01 -3.83428812e-01 2.10257888e-01 2.91811436e-01
-8.19812357e-01 2.64760882e-01 -1.12147167e-01 -2.32193932e-01
-2.49345347e-01 -9.87578213e-01 -5.16571879e-01 -4.48560625e-01
-3.66113722e-01 -7.19979882e-01 9.48612630e-01 -2.82652915e-01
1.25923419e+00 -1.90435505e+00 6.03453815e-02 -4.95054901e-01
2.37181306e-01 6.15805268e-01 6.54989434e-03 5.36822557e-01
3.34841490e-01 -2.43903682e-01 -1.86319321e-01 -8.97676945e-01
3.09180189e-02 2.16692299e-01 -2.48502821e-01 -6.21326528e-02
3.62785608e-01 7.46091068e-01 -7.78075635e-01 -7.23698020e-01
4.70881790e-01 -1.79167017e-01 -3.91434729e-01 9.40684021e-01
-6.11677527e-01 7.02693701e-01 -5.74721634e-01 4.64061409e-01
4.92144406e-01 -3.97689551e-01 3.19336563e-01 2.56220043e-01
-1.57935634e-01 5.02489984e-01 -8.68661165e-01 1.80916166e+00
-5.93621254e-01 3.19100231e-01 1.20726295e-01 -7.11079895e-01
1.19120359e+00 5.18561959e-01 2.87975103e-01 -3.46152067e-01
3.94332558e-01 -1.42476365e-01 1.16097853e-01 -4.83114004e-01
6.32701933e-01 -2.27437586e-01 -3.84883434e-01 8.91461670e-01
6.23888314e-01 2.06774995e-02 1.76210273e-02 4.36219990e-01
9.79967594e-01 -2.41001680e-01 4.36736673e-01 -3.91357355e-02
7.37198234e-01 1.56747043e-01 4.75967944e-01 6.35168910e-01
-6.47163570e-01 1.66962102e-01 2.62989968e-01 -1.28912002e-01
-7.38357127e-01 -6.64189100e-01 5.03560714e-02 1.97505319e+00
2.39465106e-02 -1.91282928e-01 -9.54171658e-01 -1.06124187e+00
-2.55228221e-01 9.88286734e-01 -3.67024094e-01 -3.64826113e-01
-6.01624727e-01 -6.26857042e-01 4.32325065e-01 5.93225420e-01
6.24394596e-01 -1.10315168e+00 -3.02296132e-01 4.00562018e-01
-5.47265112e-02 -1.18010676e+00 -5.93358517e-01 6.94859564e-01
-9.11379635e-01 -6.27365768e-01 -6.54463708e-01 -1.00993323e+00
6.53362095e-01 4.95603353e-01 8.22808862e-01 -1.10458776e-01
1.92289114e-01 1.24312066e-01 -4.98248994e-01 -2.61754155e-01
-7.52763033e-01 2.57663280e-01 2.31309578e-01 -1.09910019e-01
8.19595993e-01 -1.21168993e-01 -1.17873214e-01 6.03574812e-01
-5.45412242e-01 3.30430903e-02 4.34588850e-01 1.29494059e+00
-1.05439417e-01 -5.51654339e-01 1.05939388e+00 -1.22378337e+00
8.99521828e-01 -3.69897604e-01 -3.05765957e-01 5.19817770e-01
-7.92457581e-01 3.02528590e-01 6.04489684e-01 -7.78998137e-01
-1.85222828e+00 1.46230608e-01 -9.22515765e-02 -4.16703969e-01
-4.34029222e-01 3.14270675e-01 -1.30369246e-01 1.54754803e-01
7.24236071e-01 1.13675319e-01 -8.96330699e-02 -7.38270700e-01
6.12942994e-01 1.19503391e+00 2.28533089e-01 -4.56981659e-01
4.61961389e-01 1.31379785e-02 -8.55425775e-01 -6.63642347e-01
-1.17603385e+00 -9.40598845e-01 -1.04640949e+00 -1.51614726e-01
9.86392617e-01 -9.99431431e-01 -4.87448901e-01 5.25176764e-01
-1.08141029e+00 -7.61430144e-01 -3.25536370e-01 4.91921484e-01
-3.16830277e-01 1.85466170e-01 -9.64219749e-01 -1.30729759e+00
-5.72682381e-01 -1.12563932e+00 8.32186103e-01 5.02075732e-01
-3.48545134e-01 -1.08768833e+00 3.73247206e-01 7.23518252e-01
2.92239606e-01 -8.08781385e-01 9.11567032e-01 -1.63223290e+00
-1.32917359e-01 -1.78030208e-02 1.12165533e-01 2.75896400e-01
2.55505741e-01 -5.05776107e-01 -1.48869562e+00 -1.56648472e-01
4.32537645e-01 -1.06351459e+00 6.55186951e-01 9.03585106e-02
4.61977094e-01 -9.01687369e-02 -2.34003320e-01 -2.25360304e-01
6.49236023e-01 4.36041772e-01 1.09831113e-02 -9.59876031e-02
5.44731796e-01 9.32099462e-01 1.08212912e+00 5.99254429e-01
5.47948003e-01 5.94658971e-01 -2.10588455e-01 1.71719268e-01
2.69927196e-02 -2.99240172e-01 5.31203508e-01 1.12721908e+00
5.41554272e-01 -3.35625321e-01 -8.46571386e-01 4.29661542e-01
-2.16360950e+00 -6.09783530e-01 3.10414821e-01 1.99272346e+00
1.22958422e+00 4.03708339e-01 2.25936741e-01 -3.86057824e-01
8.97197247e-01 2.83646375e-01 -6.99250579e-01 -2.88868666e-01
3.00277591e-01 1.42514825e-01 5.05012460e-02 6.52553380e-01
-1.22605908e+00 1.49320686e+00 5.88474226e+00 6.76547885e-01
-8.78672779e-01 4.45901603e-01 7.85652578e-01 2.67751545e-01
3.19016576e-02 1.48243457e-01 -1.36493611e+00 2.37750217e-01
1.30785918e+00 -1.95492625e-01 -1.44725740e-01 1.37947845e+00
2.48862758e-01 -1.60132423e-01 -1.17104340e+00 2.87665576e-01
1.37416601e-01 -1.00312519e+00 -3.38164032e-01 -1.24120057e-01
8.14672351e-01 -2.17337281e-01 -2.47302651e-01 1.15293539e+00
9.43024158e-01 -4.13375199e-01 1.02762081e-01 2.36915704e-02
3.21772277e-01 -3.05653542e-01 7.76131153e-01 8.28796148e-01
-9.96059477e-01 -1.89898312e-01 -5.83865523e-01 -5.35195321e-03
4.30905044e-01 2.00676266e-02 -1.44132042e+00 1.01764210e-01
1.11431189e-01 3.57698411e-01 -4.79391456e-01 5.91767967e-01
-2.64377326e-01 7.24395514e-01 1.94395527e-01 -3.78271133e-01
3.91564757e-01 1.68341883e-02 1.82521924e-01 8.99684846e-01
-2.42817894e-01 4.18175757e-01 6.67571366e-01 6.94481969e-01
-2.98936337e-01 1.29139930e-01 -4.99054790e-01 -5.29570654e-02
6.46171808e-01 1.41557193e+00 -3.48541170e-01 -6.56273007e-01
-6.98035061e-01 8.48221600e-01 6.03968382e-01 1.24631248e-01
-5.64366817e-01 1.23869386e-02 3.25866044e-01 -3.51580828e-01
1.28480718e-01 -1.79075226e-01 -1.50604829e-01 -1.17742622e+00
-4.12291378e-01 -7.48194396e-01 5.57142615e-01 -4.36208397e-01
-1.40807557e+00 6.93286717e-01 5.08457571e-02 -1.26893640e+00
-7.95903921e-01 -3.99122745e-01 -1.16615987e+00 7.88870275e-01
-1.28026497e+00 -1.05139899e+00 -1.97585493e-01 7.07508028e-01
1.40361810e+00 -4.61440980e-01 9.14824367e-01 1.74802795e-01
-8.86565626e-01 6.36123657e-01 -4.33192961e-02 2.45506018e-01
1.30745339e+00 -1.26687753e+00 4.72728074e-01 6.62381232e-01
-1.43696204e-01 6.85496271e-01 5.79331696e-01 -8.28043640e-01
-1.21666479e+00 -8.70867908e-01 8.77453208e-01 -2.56413490e-01
7.80892491e-01 -4.52970266e-01 -1.12599683e+00 8.33753347e-01
4.62133676e-01 -2.71596074e-01 8.98963809e-01 4.32362080e-01
-1.18325859e-01 1.95406914e-01 -1.14263546e+00 6.20660067e-01
5.64978540e-01 -5.65306425e-01 -1.02682889e+00 3.97184998e-01
1.12403297e+00 -2.88499951e-01 -6.77281678e-01 2.38476500e-01
7.66481161e-02 -6.18601739e-01 5.93850613e-01 -1.06589222e+00
-1.15983672e-02 3.65415603e-01 1.40071064e-01 -1.31860721e+00
1.91796705e-01 -7.85891533e-01 -4.15335298e-01 1.54634869e+00
5.03512323e-01 -1.57002196e-01 1.09195590e+00 1.29868102e+00
-4.35465693e-01 -5.65610170e-01 -7.58367181e-01 -6.23342395e-01
-6.38506487e-02 1.11015260e-01 1.07556835e-01 1.02908528e+00
7.82931209e-01 1.15185034e+00 -8.47159624e-01 -4.14488286e-01
2.54467964e-01 8.92312378e-02 1.02350497e+00 -1.39114738e+00
-1.92770109e-01 1.76382929e-01 2.08863586e-01 -1.50824475e+00
5.41234851e-01 -3.85429651e-01 5.55682242e-01 -1.11365759e+00
2.22021937e-01 -4.63202834e-01 -1.58020392e-01 5.92413187e-01
-7.30008185e-01 -5.84670365e-01 1.43608198e-01 2.39685759e-01
-9.29688990e-01 1.00601530e+00 1.06974602e+00 -1.77232355e-01
-7.58903921e-01 5.01250684e-01 -4.83138323e-01 7.71274090e-01
7.93039203e-01 -4.55415159e-01 -6.59751654e-01 -1.03622826e-03
-8.05211246e-01 3.88811678e-01 -3.01788807e-01 -8.01315188e-01
6.00310147e-01 -2.13080868e-01 -1.35964021e-01 -6.84881151e-01
7.70920992e-01 -6.23044848e-01 -8.18742156e-01 4.02430952e-01
-9.26634252e-01 -2.17236564e-01 -7.99875520e-03 5.11585176e-01
-1.74694121e-01 -8.75583589e-01 9.56080258e-01 -2.07642913e-01
-1.01284599e+00 1.53977305e-01 -6.64022088e-01 1.95293069e-01
1.08937967e+00 3.30139488e-01 -6.54602170e-01 -6.31576598e-01
-8.01282167e-01 5.85290313e-01 1.04245000e-01 4.81675476e-01
4.62027460e-01 -1.02835810e+00 -4.11386728e-01 1.21621964e-02
3.66070002e-01 -1.71746239e-01 4.01974529e-01 6.12380624e-01
2.96882242e-01 5.28635919e-01 -1.36041135e-01 -4.74737138e-01
-1.48758280e+00 6.52139485e-01 -3.28847207e-02 -7.27967262e-01
-4.16282117e-01 8.88923109e-01 1.33889005e-01 -8.85348678e-01
4.99047428e-01 -1.26946419e-01 -4.15648311e-01 2.52777696e-01
5.62259138e-01 1.73635498e-01 -8.98320079e-02 -3.36507976e-01
-4.16459627e-02 -2.59429723e-01 -7.02958107e-01 -4.96728182e-01
9.02009249e-01 -3.97381485e-01 5.16715109e-01 7.22666442e-01
6.68413937e-01 -2.94413447e-01 -1.32355785e+00 -9.20778930e-01
2.94909090e-01 -8.97288024e-02 -1.58701390e-01 -5.69760680e-01
-4.66223806e-01 1.14791894e+00 5.05688787e-01 1.45943999e-01
6.14800036e-01 1.39595762e-01 8.03849041e-01 1.00572395e+00
2.97571152e-01 -1.30833900e+00 6.88324928e-01 9.30741310e-01
1.37427181e-01 -1.87558722e+00 -2.51461118e-01 -3.69797289e-01
-1.33984339e+00 7.89934933e-01 1.49919140e+00 4.27933723e-01
5.97485304e-01 1.49984092e-01 3.39269131e-01 -5.03236800e-02
-1.13549352e+00 -4.08505827e-01 9.38492492e-02 2.75189310e-01
6.31470859e-01 -3.07566404e-01 -1.98292688e-01 8.08417320e-01
2.46691614e-01 -2.32805401e-01 4.32946712e-01 1.41721904e+00
-1.02628863e+00 -1.37866426e+00 -8.04143846e-02 2.92290062e-01
-1.11658953e-01 -3.13374460e-01 -7.45157480e-01 4.86484528e-01
-5.38373828e-01 1.52023673e+00 -2.35408798e-01 -5.88026226e-01
2.07830280e-01 8.59300494e-01 -1.15203828e-01 -1.31605339e+00
-9.55726743e-01 1.69365928e-01 3.96824211e-01 2.81116907e-02
-3.93153518e-01 -2.11880729e-01 -1.17167187e+00 -2.96440870e-02
-7.63099849e-01 4.82507885e-01 2.28962108e-01 1.16142380e+00
2.60015935e-01 4.37865406e-01 9.36235785e-01 -5.59819281e-01
-1.17373729e+00 -1.59701836e+00 -3.91830832e-01 1.15530066e-01
-2.48280913e-02 -7.33763456e-01 -2.10074916e-01 4.47101966e-02] | [12.803311347961426, 7.876428127288818] |
dca67423-84f0-4e4e-9183-9b6151ca8703 | melnet-a-generative-model-for-audio-in-the | 1906.01083 | null | https://arxiv.org/abs/1906.01083v1 | https://arxiv.org/pdf/1906.01083v1.pdf | MelNet: A Generative Model for Audio in the Frequency Domain | Capturing high-level structure in audio waveforms is challenging because a single second of audio spans tens of thousands of timesteps. While long-range dependencies are difficult to model directly in the time domain, we show that they can be more tractably modelled in two-dimensional time-frequency representations such as spectrograms. By leveraging this representational advantage, in conjunction with a highly expressive probabilistic model and a multiscale generation procedure, we design a model capable of generating high-fidelity audio samples which capture structure at timescales that time-domain models have yet to achieve. We apply our model to a variety of audio generation tasks, including unconditional speech generation, music generation, and text-to-speech synthesis---showing improvements over previous approaches in both density estimates and human judgments. | ['Mike Lewis', 'Sean Vasquez'] | 2019-06-04 | null | https://openreview.net/forum?id=r1gIa0NtDH | https://openreview.net/pdf?id=r1gIa0NtDH | null | ['audio-generation'] | ['audio'] | [ 1.65836170e-01 3.65402699e-02 9.52738523e-02 -3.37756544e-01
-1.54596698e+00 -7.52550125e-01 7.84450531e-01 2.71117333e-02
2.12852120e-01 8.97023857e-01 8.59494328e-01 -1.23763554e-01
-1.89710394e-01 -5.98378062e-01 -6.42257273e-01 -3.70388150e-01
-4.34631437e-01 4.20487702e-01 1.05368398e-01 -4.19319496e-02
1.08655214e-01 2.17771545e-01 -1.59498680e+00 4.04843122e-01
6.68708622e-01 9.13340986e-01 -1.52053330e-02 1.21787524e+00
6.80318177e-02 8.53777349e-01 -8.07270706e-01 3.49687189e-02
-2.02259690e-01 -4.99402076e-01 -5.39986670e-01 -1.49233848e-01
2.95469403e-01 -4.03026134e-01 -2.58771092e-01 5.67435980e-01
7.03016818e-01 3.58964592e-01 1.09407210e+00 -9.37115431e-01
-4.18051064e-01 1.13337612e+00 -2.00768471e-01 1.23047024e-01
6.97851479e-01 7.79794306e-02 1.18213010e+00 -7.80779183e-01
3.82762283e-01 1.41796708e+00 9.86372650e-01 4.62953418e-01
-1.67539716e+00 -6.13155901e-01 -1.60559006e-02 -2.77530789e-01
-1.20987475e+00 -8.49197268e-01 8.30839097e-01 -5.60633063e-01
9.75079298e-01 -1.00189291e-01 4.81799364e-01 1.33380508e+00
2.16049239e-01 7.18597293e-01 7.51900792e-01 -3.44337165e-01
1.75041422e-01 -3.72559994e-01 -2.85955310e-01 1.80182785e-01
-4.98130739e-01 5.85543573e-01 -1.03628778e+00 -5.94451487e-01
8.75892997e-01 -6.23035669e-01 -2.12242439e-01 1.64659098e-01
-1.26653957e+00 6.02147698e-01 -1.17450714e-01 1.79543659e-01
-4.56651717e-01 8.02641749e-01 4.20991749e-01 1.70542195e-01
6.80287361e-01 5.05239129e-01 -3.27204883e-01 -8.62721026e-01
-1.48034418e+00 6.16984606e-01 9.16151583e-01 9.31344092e-01
3.06339443e-01 7.43958116e-01 -1.39662340e-01 7.95804083e-01
1.41476497e-01 5.60369492e-01 4.67399001e-01 -1.00857329e+00
2.46353731e-01 -4.95651931e-01 4.46718931e-01 -5.54578960e-01
-4.12021764e-02 -3.65146607e-01 -6.55812800e-01 1.43433794e-01
4.81160462e-01 -2.64115483e-01 -8.67896378e-01 2.07447886e+00
1.07657231e-01 6.58185601e-01 -1.04079299e-01 5.35220385e-01
2.64849871e-01 1.14724410e+00 8.29263330e-02 -3.29809964e-01
1.12219083e+00 -4.23227787e-01 -7.63895571e-01 5.21770045e-02
5.71550392e-02 -9.83097255e-01 9.13006246e-01 6.83839619e-01
-1.57297516e+00 -7.75964737e-01 -9.02729213e-01 7.50433207e-02
6.72452897e-02 -1.58732936e-01 6.36357963e-01 6.77915156e-01
-1.10064435e+00 9.22203302e-01 -8.39328945e-01 1.75716624e-01
2.67990753e-02 -7.78422728e-02 -1.08995244e-01 4.53929931e-01
-1.25787783e+00 4.99086380e-01 1.07685052e-01 -3.06831867e-01
-1.39803672e+00 -1.14996207e+00 -8.12149525e-01 1.49106696e-01
-2.49558955e-01 -7.73728371e-01 1.91491544e+00 -4.05524880e-01
-1.62887394e+00 9.76224914e-02 -4.21171844e-01 -6.37572169e-01
2.44330347e-01 -3.37676436e-01 -5.85473299e-01 2.29331240e-01
-6.07679449e-02 6.37397826e-01 1.37158287e+00 -8.86750698e-01
-6.01607621e-01 1.80385411e-01 -1.04808591e-01 -4.05204780e-02
-2.36870587e-01 1.20916680e-01 2.01559171e-01 -1.35057080e+00
-2.86730200e-01 -7.13336170e-01 -2.48460010e-01 -1.57333821e-01
-3.05394948e-01 -1.47202805e-01 6.06911421e-01 -6.50501907e-01
1.36454165e+00 -2.10698223e+00 2.47408554e-01 -6.64024800e-02
-1.43901750e-01 -2.35731885e-01 -2.22985730e-01 7.67163098e-01
-1.75943851e-01 8.89721960e-02 -1.41351074e-01 -5.81443667e-01
3.98645639e-01 -1.20145157e-01 -1.16091824e+00 1.49886804e-02
4.02507216e-01 6.38005674e-01 -9.36278880e-01 -4.23775643e-01
7.98127875e-02 7.30694532e-01 -7.41999507e-01 3.82820070e-01
-5.72903693e-01 3.32638592e-01 -2.14931563e-01 4.00981665e-01
1.84203178e-01 -8.96323770e-02 -2.85984278e-02 -6.44486994e-02
-7.56112859e-02 8.83856416e-01 -1.17268491e+00 1.84022200e+00
-6.46331370e-01 5.82526684e-01 4.77039665e-02 -5.25089502e-01
6.32586598e-01 9.06093836e-01 2.94971526e-01 -1.85042322e-02
-1.83810696e-01 1.80248618e-01 -2.34938949e-01 -4.88128923e-02
6.21571839e-01 -7.26644695e-01 -3.70835513e-01 9.66745555e-01
3.01515311e-01 -1.00480664e+00 1.90510884e-01 2.71643221e-01
9.80271280e-01 2.66214311e-01 1.03013784e-01 -1.30131617e-01
-1.71929032e-01 -3.95170569e-01 2.60613620e-01 8.53150964e-01
1.94899365e-01 8.49729717e-01 3.49083483e-01 -6.49825484e-03
-1.27399707e+00 -1.62636709e+00 -2.03465849e-01 1.31575930e+00
-6.89265668e-01 -6.47565484e-01 -7.14241445e-01 1.12282440e-01
-3.55189443e-02 8.97878289e-01 -3.98438185e-01 -1.86610878e-01
-4.03084546e-01 -5.08607507e-01 1.06327522e+00 8.98043990e-01
-5.00254273e-01 -9.80490148e-01 -2.45131329e-01 8.54528904e-01
-2.02881873e-01 -9.02837694e-01 -7.43879080e-01 2.32079044e-01
-1.00453997e+00 -3.53070349e-01 -7.50522017e-01 -4.38902885e-01
-2.00693533e-01 -4.10666406e-01 1.48667359e+00 -6.05037153e-01
-4.23464090e-01 4.61376399e-01 -1.78787515e-01 -7.06008017e-01
-6.75744593e-01 -1.42919526e-01 4.05543953e-01 -2.68219233e-01
-1.51835442e-01 -1.30961347e+00 -3.11112911e-01 -6.01051301e-02
-9.13365841e-01 -6.14727773e-02 2.68628836e-01 9.29762602e-01
3.32009047e-01 2.14581460e-01 1.16316271e+00 -4.82645690e-01
1.02920914e+00 -4.79770154e-01 -5.17707109e-01 -7.76504949e-02
-2.52312243e-01 6.70343712e-02 6.71249211e-01 -9.19703364e-01
-1.04915679e+00 -1.03012763e-01 -2.90133297e-01 -3.49024057e-01
-1.34075925e-01 5.28676510e-01 2.76493132e-01 4.79636014e-01
8.92258644e-01 2.77589530e-01 -1.87628239e-01 -4.38281298e-01
7.61792123e-01 5.62697113e-01 7.17532039e-01 -1.17046881e+00
8.34283113e-01 2.12433547e-01 -2.22119913e-01 -7.00581670e-01
-8.70789170e-01 2.59954296e-03 -2.88641900e-01 -6.22176798e-03
5.75540245e-01 -1.13530970e+00 -4.83410150e-01 2.78339058e-01
-1.22179759e+00 -4.14553255e-01 -7.54916906e-01 7.89914250e-01
-1.08753729e+00 1.28398865e-01 -1.12042642e+00 -1.34209228e+00
-1.60637394e-01 -7.29727089e-01 1.31312573e+00 -8.02465901e-02
-7.35696554e-01 -1.05833411e+00 3.68988752e-01 -3.59803408e-01
5.83401740e-01 1.89610481e-01 1.20075703e+00 -2.21558630e-01
-4.59410548e-01 -6.35449290e-02 1.92129180e-01 3.89854163e-01
3.29872295e-02 3.27331692e-01 -1.26548326e+00 1.03642121e-01
-7.76907429e-02 -6.14263058e-01 7.26866841e-01 7.18746483e-01
9.75990355e-01 -2.04261586e-01 -6.14910871e-02 3.09047371e-01
8.30736279e-01 4.26472053e-02 3.60379308e-01 -5.85332811e-01
2.80479252e-01 4.13142115e-01 4.52497870e-01 9.72727597e-01
1.96444839e-01 4.56328899e-01 -2.33568475e-02 3.42940688e-01
-3.06800574e-01 -6.95853293e-01 4.75565493e-01 1.20448172e+00
9.20950063e-03 -2.88060606e-01 -6.87776029e-01 1.05750275e+00
-1.49591398e+00 -1.38071108e+00 2.67617285e-01 2.00279641e+00
1.38404238e+00 2.77612329e-01 3.67557406e-01 4.29336846e-01
4.89050031e-01 3.55386555e-01 -3.50374103e-01 -3.85573894e-01
9.18996483e-02 6.59758031e-01 -2.50254720e-01 6.69940829e-01
-9.07598794e-01 6.12189054e-01 8.75304985e+00 1.15303004e+00
-9.46330070e-01 -4.31625471e-02 4.04853582e-01 -2.64024645e-01
-8.11876118e-01 1.06687829e-01 -6.40963376e-01 4.23407793e-01
1.67086887e+00 -4.45775986e-01 6.96829319e-01 6.40422761e-01
4.17248368e-01 3.30768198e-01 -1.59049106e+00 7.64884889e-01
-2.47306719e-01 -1.41701710e+00 1.69261634e-01 -2.14279801e-01
7.58179069e-01 -2.43426383e-01 4.57106024e-01 3.59247208e-01
5.82635343e-01 -1.47192121e+00 1.12306619e+00 6.58336461e-01
1.18261254e+00 -7.90240049e-01 -2.47100182e-02 4.30472702e-01
-1.51292026e+00 -5.34556946e-03 -2.44447216e-01 -1.89247951e-01
7.69436181e-01 1.03464079e+00 -1.04682016e+00 3.91329229e-01
2.85817623e-01 6.24342442e-01 1.69281065e-01 9.62754548e-01
-6.37074709e-02 1.09949315e+00 -4.73548561e-01 1.83216184e-01
1.05613820e-01 2.82822996e-01 6.98620677e-01 1.40351391e+00
8.62081945e-01 8.11298639e-02 -2.94908043e-02 1.05302250e+00
6.79632202e-02 -4.08750981e-01 -6.42966866e-01 -5.54523528e-01
8.41166019e-01 8.71264398e-01 -2.26527508e-02 -4.27505970e-01
5.16791977e-02 4.37661827e-01 1.40359849e-01 5.03950179e-01
-9.30757761e-01 -5.45407414e-01 7.39569664e-01 1.63828269e-01
4.12231892e-01 -5.09185016e-01 -4.30906899e-02 -9.88397479e-01
-2.72952795e-01 -9.24016356e-01 9.93326455e-02 -1.18681550e+00
-1.69065678e+00 7.23254919e-01 2.89875358e-01 -1.36790395e+00
-1.35562789e+00 -2.88855970e-01 -6.37949467e-01 1.02305436e+00
-1.13200009e+00 -1.14289486e+00 5.12223542e-01 4.99919534e-01
5.78724682e-01 3.82446721e-02 1.23278594e+00 2.08664164e-01
1.99770719e-01 4.04958248e-01 -1.27426878e-01 -2.73226589e-01
6.79163575e-01 -1.41882002e+00 8.50232065e-01 5.35144925e-01
5.30254662e-01 6.97610915e-01 1.01276410e+00 -4.19040948e-01
-1.05731106e+00 -9.55214381e-01 9.50992584e-01 -6.76585376e-01
1.00094450e+00 -3.90807211e-01 -8.36695790e-01 5.60097575e-01
2.06381634e-01 -9.33593065e-02 9.05669212e-01 4.22023505e-01
-7.41173863e-01 1.71419516e-01 -7.90251434e-01 4.13868845e-01
8.75158250e-01 -1.25494349e+00 -9.12243664e-01 2.75088310e-01
9.20480311e-01 -3.76403064e-01 -1.14347029e+00 2.73246080e-01
7.80079126e-01 -7.27046192e-01 1.20793080e+00 -6.19075239e-01
4.69973713e-01 -2.98972517e-01 -2.37602651e-01 -1.51735473e+00
-2.78057963e-01 -1.41825104e+00 -4.17296916e-01 1.39267612e+00
5.18082619e-01 -2.56726295e-01 2.82013804e-01 2.80351520e-01
-1.88054278e-01 -3.66677612e-01 -1.00887334e+00 -8.18236411e-01
2.60298967e-01 -1.02052331e+00 8.65175009e-01 4.68131155e-01
2.09699646e-01 3.01008463e-01 -6.57121241e-01 1.20590501e-01
6.19988918e-01 1.95716858e-01 4.79200810e-01 -1.19741881e+00
-8.08633626e-01 -4.29216802e-01 -1.04960717e-01 -1.23755753e+00
2.39679515e-01 -4.42367166e-01 5.04820704e-01 -1.08063614e+00
-1.92511097e-01 -5.47093928e-01 -2.58000553e-01 7.23769069e-02
-3.43652517e-02 3.03766936e-01 3.14600132e-02 -2.00955227e-01
-2.52947211e-01 7.59309292e-01 8.94341052e-01 7.90667236e-02
9.93359461e-02 9.65030417e-02 -5.92528403e-01 8.16230536e-01
4.19904172e-01 -6.84625804e-01 -7.76841283e-01 -3.38464916e-01
1.23815536e-01 8.21527004e-01 2.95710802e-01 -1.21322012e+00
3.31570432e-02 -3.36676478e-01 3.73949826e-01 -6.22077525e-01
9.93873656e-01 -1.88968897e-01 4.34228599e-01 -6.76485375e-02
-7.17624664e-01 -8.05071443e-02 3.38382661e-01 8.78863573e-01
-4.78962123e-01 -1.75158486e-01 5.77557325e-01 1.01468101e-01
-4.35371734e-02 2.33537108e-01 -7.37254262e-01 3.62909555e-01
4.40339774e-01 1.23951405e-01 -4.41609770e-02 -9.61244404e-01
-8.63022327e-01 -4.12479609e-01 1.67968839e-01 2.30839789e-01
5.62434971e-01 -1.58628762e+00 -8.74139667e-01 5.60630746e-02
-9.62041467e-02 -3.25758338e-01 3.09272468e-01 2.37570524e-01
7.70332068e-02 3.38524193e-01 9.75752175e-02 -5.37326396e-01
-7.59996057e-01 1.27613559e-01 3.64367701e-02 -3.65374565e-01
-4.87316251e-01 1.07209754e+00 3.29178095e-01 -1.69546455e-01
2.42780253e-01 -6.08960092e-01 1.84786439e-01 -2.52876412e-02
8.88509154e-01 -5.07053919e-02 -3.00691456e-01 -3.17601532e-01
-1.31033763e-01 4.67601359e-01 2.76071489e-01 -1.00229394e+00
1.07902884e+00 -4.01194394e-02 2.90480345e-01 9.60637391e-01
7.77400494e-01 3.82742167e-01 -1.56461930e+00 6.93364739e-02
-2.35115752e-01 -4.02519137e-01 -1.14476249e-01 -7.75778890e-01
-2.29026169e-01 1.26224065e+00 2.41928976e-02 4.80262220e-01
1.08203340e+00 -1.14108868e-01 1.01676416e+00 1.69964582e-01
5.90278745e-01 -8.22531939e-01 3.80269468e-01 6.81601286e-01
9.18988049e-01 -5.62541902e-01 -2.49592528e-01 -6.34161383e-02
-4.71280515e-01 9.79770362e-01 -6.47028834e-02 -2.71949172e-01
9.04308736e-01 8.34489167e-01 -1.99237376e-01 3.22142422e-01
-1.47151732e+00 1.30864844e-01 5.08954406e-01 8.75657380e-01
7.54659951e-01 2.18173951e-01 3.30872506e-01 7.80649900e-01
-7.04033792e-01 4.61805575e-02 5.16024351e-01 7.34122038e-01
-5.03057241e-01 -1.15094674e+00 -3.23487490e-01 3.32669288e-01
-6.27335489e-01 -4.03574556e-01 1.08946539e-01 1.82010993e-01
-4.15824652e-01 1.17416620e+00 1.56002015e-01 -3.08308691e-01
-4.63771559e-02 4.00663257e-01 5.32535911e-01 -7.77218401e-01
-4.54264730e-01 5.35973608e-01 5.28554320e-01 -3.65777045e-01
-1.94065973e-01 -6.28430724e-01 -1.16756511e+00 -3.56244475e-01
-2.17222899e-01 4.04337406e-01 5.96398711e-01 6.43558919e-01
3.62218201e-01 7.80619621e-01 4.91027534e-01 -1.29128540e+00
-9.05199289e-01 -1.18660188e+00 -7.80127764e-01 1.57763213e-01
5.71565151e-01 -3.02402943e-01 -4.93772775e-01 5.57618916e-01] | [15.625884056091309, 5.8140645027160645] |
f129bfdd-a366-41ef-b002-94be9327ea9f | imlovenet-misaligned-image-supported | 2207.00826 | null | https://arxiv.org/abs/2207.00826v1 | https://arxiv.org/pdf/2207.00826v1.pdf | ImLoveNet: Misaligned Image-supported Registration Network for Low-overlap Point Cloud Pairs | Low-overlap regions between paired point clouds make the captured features very low-confidence, leading cutting edge models to point cloud registration with poor quality. Beyond the traditional wisdom, we raise an intriguing question: Is it possible to exploit an intermediate yet misaligned image between two low-overlap point clouds to enhance the performance of cutting-edge registration models? To answer it, we propose a misaligned image supported registration network for low-overlap point cloud pairs, dubbed ImLoveNet. ImLoveNet first learns triple deep features across different modalities and then exports these features to a two-stage classifier, for progressively obtaining the high-confidence overlap region between the two point clouds. Therefore, soft correspondences are well established on the predicted overlap region, resulting in accurate rigid transformations for registration. ImLoveNet is simple to implement yet effective, since 1) the misaligned image provides clearer overlap information for the two low-overlap point clouds to better locate overlap parts; 2) it contains certain geometry knowledge to extract better deep features; and 3) it does not require the extrinsic parameters of the imaging device with respect to the reference frame of the 3D point cloud. Extensive qualitative and quantitative evaluations on different kinds of benchmarks demonstrate the effectiveness and superiority of our ImLoveNet over state-of-the-art approaches. | ['Jun Wang', 'Mingqiang Wei', 'Yabin Xu', 'Zeyong Wei', 'Honghua Chen'] | 2022-07-02 | null | null | null | null | ['point-cloud-registration'] | ['computer-vision'] | [-1.07188247e-01 1.02182291e-01 -2.19330147e-01 -3.46344739e-01
-8.38508666e-01 -4.93578345e-01 6.74794734e-01 -6.14894368e-02
-1.49246931e-01 1.03649355e-01 -1.68133348e-01 7.45092854e-02
-1.82310551e-01 -5.41741252e-01 -9.69457865e-01 -4.91845965e-01
2.39116445e-01 6.60157621e-01 2.71510065e-01 -1.94716841e-01
1.20606475e-01 7.44243622e-01 -1.33721197e+00 -1.27486810e-02
9.18886364e-01 1.10139084e+00 1.82662711e-01 -1.28910601e-01
6.92648813e-02 -1.66377112e-01 -1.27943888e-01 -4.98111248e-01
8.44092369e-01 4.50695418e-02 -6.11662805e-01 1.54643757e-02
7.54993320e-01 -2.07227767e-01 -3.28657210e-01 1.12036133e+00
9.95317996e-02 -9.44250152e-02 6.99332297e-01 -1.30478489e+00
-5.68482637e-01 1.09404311e-01 -8.98397803e-01 -1.28285855e-01
4.41044331e-01 1.81844205e-01 9.23965037e-01 -1.25931108e+00
6.77857459e-01 8.86140883e-01 1.06267571e+00 1.73123591e-02
-1.15827012e+00 -8.56033444e-01 2.76795728e-03 -9.35196429e-02
-1.72833765e+00 -3.65822166e-01 9.20943558e-01 -5.41392744e-01
7.00285256e-01 1.46483451e-01 7.32490003e-01 7.65884459e-01
1.28817797e-01 2.59869963e-01 9.60476637e-01 -3.41769040e-01
-2.65695900e-01 -2.59965695e-02 -1.45717874e-01 7.04690814e-01
2.14427382e-01 5.45236349e-01 -4.50976938e-01 -1.57242030e-01
1.13874376e+00 3.86818171e-01 -4.98973429e-01 -7.30926871e-01
-1.52884007e+00 4.80648875e-01 9.73654151e-01 5.18181503e-01
-4.24215376e-01 -5.20299152e-02 -1.08128451e-01 5.31551875e-02
4.23839003e-01 5.16568899e-01 -2.59822994e-01 1.14701301e-01
-8.38763595e-01 -9.69277024e-02 2.07449034e-01 1.27030253e+00
1.09203732e+00 -2.30384484e-01 1.49056509e-01 4.78808552e-01
4.35887277e-01 5.81409812e-01 9.78329703e-02 -8.75943303e-01
6.73581898e-01 8.49921167e-01 7.70463720e-02 -1.35762382e+00
-4.47131634e-01 -4.72419322e-01 -9.81307864e-01 2.42889658e-01
3.95583987e-01 2.06478655e-01 -6.86114728e-01 1.37960339e+00
4.25814837e-01 7.01819718e-01 -8.09283778e-02 1.19858778e+00
9.59288776e-01 3.21096331e-01 -4.99394357e-01 1.00675737e-02
1.32742620e+00 -6.92254543e-01 -2.20851213e-01 -2.32534543e-01
5.17142236e-01 -1.05789387e+00 8.16886067e-01 -1.81026682e-01
-8.72942328e-01 -6.07943475e-01 -1.05189884e+00 -8.83233249e-02
-5.73188961e-02 1.92544892e-01 5.77790260e-01 2.00287059e-01
-8.33472431e-01 6.89851046e-01 -7.53977835e-01 -7.45733678e-02
5.60845494e-01 5.12379944e-01 -8.98847640e-01 -1.62906080e-01
-8.07825923e-01 8.71598244e-01 1.94613814e-01 2.67089337e-01
-2.09890574e-01 -1.05249369e+00 -9.66896236e-01 -4.18764129e-02
2.81969458e-01 -9.68591809e-01 7.03566730e-01 -8.87413979e-01
-1.23392415e+00 1.14598334e+00 -1.37983918e-01 3.94252278e-02
6.10920668e-01 -1.67594507e-01 -1.51262462e-01 7.02524334e-02
2.90454209e-01 7.28262842e-01 7.90396214e-01 -1.71920919e+00
-5.12721002e-01 -5.25521636e-01 -2.20174238e-01 4.04965460e-01
1.93952292e-01 -1.73658982e-01 -6.92604601e-01 -5.98798513e-01
7.83217013e-01 -1.14033401e+00 -1.45812765e-01 3.07262391e-01
-5.12443960e-01 -8.11999217e-02 8.08094144e-01 -2.09957942e-01
3.83738190e-01 -2.31862879e+00 -1.09887198e-01 4.58921283e-01
4.40003425e-01 1.79577693e-01 -2.24515438e-01 3.13708857e-02
-3.72953087e-01 -1.50181472e-01 -1.34587914e-01 -4.56561267e-01
-2.80085176e-01 1.11712642e-01 -2.77237177e-01 8.12889040e-01
1.98985055e-01 1.10316157e+00 -9.32065606e-01 -3.40045542e-01
4.87393200e-01 5.77154160e-01 -4.41004753e-01 1.88308418e-01
3.42989415e-01 8.11292529e-01 -4.37865466e-01 6.87314510e-01
1.10050285e+00 -3.32026303e-01 -5.21650016e-01 -8.85256410e-01
-1.29826799e-01 -6.83590472e-02 -1.25195575e+00 1.81077933e+00
-2.06259385e-01 5.05256295e-01 -2.94142872e-01 -6.12190306e-01
1.14032269e+00 2.03734711e-01 1.01222980e+00 -4.65022415e-01
2.97822863e-01 4.21570122e-01 -1.49406508e-01 -1.75916180e-01
4.43673611e-01 -1.49594650e-01 1.41704127e-01 -5.02975285e-02
2.09898464e-02 -5.32558978e-01 -6.47345185e-01 -1.50421977e-01
5.59652209e-01 2.40065470e-01 1.67887926e-01 5.50677814e-03
5.06419003e-01 -1.19816661e-01 6.16672218e-01 4.43850398e-01
-9.08443630e-02 1.30897224e+00 -8.10703784e-02 -4.18171346e-01
-1.02506554e+00 -1.19236565e+00 -4.27994132e-01 7.18328729e-02
8.17540348e-01 -3.27411145e-01 -5.20380080e-01 -5.84770143e-01
7.29865506e-02 2.01000705e-01 -3.21125746e-01 -2.41426632e-01
-5.96064746e-01 -3.62602293e-01 2.00302243e-01 4.76591885e-01
6.23315036e-01 -4.36631352e-01 -3.22603524e-01 -2.53525227e-01
-3.06982845e-01 -1.50134277e+00 -6.97409809e-01 -6.13012090e-02
-1.01683390e+00 -1.13793182e+00 -5.86733758e-01 -7.92225778e-01
9.60272491e-01 7.05125511e-01 1.07701468e+00 2.37744272e-01
2.05331922e-01 3.17520082e-01 -3.51450056e-01 -1.64187208e-01
-3.89305800e-02 -2.07385346e-01 3.06925714e-01 8.87813419e-02
3.25102180e-01 -6.91362321e-01 -7.69346654e-01 7.69240677e-01
-5.14509499e-01 1.39502913e-01 6.87395215e-01 8.82380068e-01
9.67588842e-01 -1.98933378e-01 3.98165695e-02 -4.10447538e-01
-1.28792644e-01 -2.06618696e-01 -7.18129694e-01 1.73430443e-01
-5.03875732e-01 -1.39507383e-01 3.39972109e-01 -3.57845843e-01
-6.94779873e-01 3.60097885e-01 -7.84723088e-02 -1.05598724e+00
2.74951626e-02 2.72726715e-01 -2.18853831e-01 -6.65137887e-01
5.33369005e-01 1.60945982e-01 1.53065681e-01 -3.40090752e-01
3.29646409e-01 4.08082426e-01 9.17192936e-01 -4.83689785e-01
1.38521588e+00 8.67959440e-01 2.29159504e-01 -6.15486026e-01
-7.58505344e-01 -6.88207626e-01 -1.20035732e+00 -1.76122457e-01
8.70153427e-01 -9.95325089e-01 -6.76322341e-01 4.30643380e-01
-1.33076966e+00 1.98135644e-01 -3.31241697e-01 7.90231347e-01
-6.66092396e-01 5.73209345e-01 -1.94321886e-01 -1.89136490e-01
-3.17667454e-01 -1.47951913e+00 1.59556746e+00 3.32627594e-01
3.58638540e-02 -7.60533392e-01 -5.56381494e-02 4.39399123e-01
-6.23947605e-02 1.98401317e-01 5.72716594e-01 -4.54371452e-01
-9.44398046e-01 -4.57811534e-01 -4.22287792e-01 1.26733094e-01
3.35584313e-01 9.52525735e-02 -9.51364756e-01 -2.67714441e-01
3.12854886e-01 1.72409117e-01 2.82068938e-01 2.52105474e-01
1.03245544e+00 1.08749174e-01 -6.14491224e-01 1.20738709e+00
1.40053248e+00 -2.11424172e-01 6.96078002e-01 4.39201266e-01
1.16089928e+00 3.83385092e-01 8.90979111e-01 6.37409836e-02
4.23593283e-01 1.23554039e+00 7.76739299e-01 -4.05284345e-01
-1.15570631e-02 -3.54873687e-01 1.14076972e-01 1.04102516e+00
-3.98156852e-01 3.98874730e-01 -9.37232196e-01 3.17049384e-01
-1.72040534e+00 -6.39658034e-01 -4.01085645e-01 2.42807913e+00
5.51142991e-01 -5.75786270e-02 -2.85388976e-01 -2.67157167e-01
7.71199822e-01 9.55582038e-03 -3.97473454e-01 3.47079843e-01
-1.78332016e-01 2.95174364e-02 4.74887311e-01 2.96630234e-01
-1.11504900e+00 8.54258239e-01 5.16466188e+00 7.68623054e-01
-1.22418797e+00 3.88640934e-03 3.85641277e-01 2.13616103e-01
-1.22553959e-01 3.81273597e-01 -7.64080167e-01 4.26546544e-01
2.68973380e-01 -8.74940585e-03 8.27508792e-02 8.08507502e-01
-1.28075868e-01 1.55818343e-01 -1.40959287e+00 1.46766555e+00
1.44855291e-01 -1.64029324e+00 9.43773426e-03 1.96271867e-01
8.69288564e-01 4.62186307e-01 4.86828499e-02 -8.87379646e-02
-7.81604205e-04 -9.31979597e-01 6.55213058e-01 7.32639790e-01
8.63529325e-01 -5.98977685e-01 9.41165090e-01 3.64170641e-01
-1.37145770e+00 4.06145632e-01 -5.32567322e-01 4.07018453e-01
2.30148539e-01 5.21028876e-01 -6.66040123e-01 1.16401017e+00
9.03124869e-01 8.91849577e-01 -6.38487577e-01 1.21240139e+00
5.39244078e-02 -1.56408072e-01 -5.84636033e-01 8.26523364e-01
1.20856263e-01 -6.46085918e-01 7.28625655e-01 6.08378232e-01
6.94236815e-01 1.23290129e-01 2.28130281e-01 1.08132756e+00
-4.50998880e-02 -1.40704915e-01 -8.53385150e-01 5.68634629e-01
6.99483335e-01 1.30599463e+00 -6.61979914e-01 -1.13829756e-02
-5.32385051e-01 7.99999237e-01 1.88473344e-01 8.35282952e-02
-7.13580906e-01 9.41125005e-02 8.22444141e-01 3.09723496e-01
1.57090321e-01 -3.28908026e-01 -3.60199362e-01 -1.42542613e+00
2.96309203e-01 -6.40230775e-01 1.53375462e-01 -1.02302182e+00
-1.51036108e+00 7.79243410e-01 -1.98624693e-02 -2.06940818e+00
2.42739698e-04 -5.26925087e-01 -6.56358302e-01 1.02890658e+00
-1.48377907e+00 -1.45572102e+00 -7.34214604e-01 7.16204405e-01
1.53224841e-01 -5.99497333e-02 6.49998307e-01 2.99894452e-01
-2.58304179e-01 7.26798236e-01 7.22635314e-02 3.49994242e-01
7.82665372e-01 -8.69742990e-01 3.40080023e-01 7.75619268e-01
4.39631760e-01 6.98755562e-01 3.58756274e-01 -5.94451308e-01
-1.46378136e+00 -1.09453905e+00 4.12668884e-01 -8.02542746e-01
3.62872213e-01 -1.80023104e-01 -1.23649871e+00 6.43286347e-01
-1.41547546e-01 5.66955984e-01 2.78229743e-01 -9.88615081e-02
-5.12762666e-01 -2.15230227e-01 -1.07566929e+00 4.89557713e-01
1.02584100e+00 -5.99834740e-01 -7.91953325e-01 2.92545229e-01
6.68515742e-01 -1.00827181e+00 -1.15954745e+00 7.87612021e-01
2.74931818e-01 -9.10167992e-01 1.35911405e+00 -1.03992939e-01
3.99850041e-01 -5.63418865e-01 -1.40725464e-01 -1.31125653e+00
-1.82625607e-01 -4.76957977e-01 2.19232693e-01 1.17364323e+00
8.36194456e-02 -8.00948381e-01 7.87647665e-01 5.82420886e-01
-4.46308613e-01 -8.72119367e-01 -1.27193248e+00 -9.78104651e-01
4.25347202e-02 -4.38521922e-01 8.78905892e-01 1.30075753e+00
-3.00616443e-01 1.37421057e-01 -6.95144534e-02 7.03543067e-01
6.11649811e-01 5.65322042e-01 1.02738297e+00 -1.48959351e+00
8.15225020e-02 -5.05982161e-01 -7.69766092e-01 -1.24809217e+00
1.82875037e-01 -1.13026762e+00 3.39133218e-02 -1.08553040e+00
-2.71906666e-02 -9.52825010e-01 -3.50938886e-02 4.93037283e-01
-1.66623265e-01 5.64839065e-01 2.78218418e-01 8.59636486e-01
-2.48804599e-01 7.92156994e-01 1.30337226e+00 -4.77592833e-02
-1.30086735e-01 1.24458902e-01 -5.30274153e-01 1.02471101e+00
2.67011911e-01 -5.48506498e-01 1.52095351e-02 -4.09087211e-01
1.16119482e-01 -5.72123379e-02 6.99596703e-01 -1.03128099e+00
3.88761997e-01 -2.07887664e-01 4.23126161e-01 -7.16524005e-01
4.06533539e-01 -1.37891996e+00 5.12317598e-01 2.10708077e-03
2.44766325e-01 1.17333174e-01 3.48835078e-04 4.43155229e-01
-3.43104362e-01 -2.39309207e-01 7.97101140e-01 1.40449524e-01
-5.35481036e-01 8.62075329e-01 5.92758715e-01 -1.63951680e-01
1.08365095e+00 -7.17705429e-01 -2.95646966e-01 -8.99334699e-02
-4.48469102e-01 1.14104569e-01 1.11150050e+00 6.48538530e-01
7.98512161e-01 -1.59166837e+00 -4.93316203e-01 4.60709751e-01
4.83064681e-01 6.76689565e-01 1.53548419e-01 1.13699090e+00
-4.47711796e-01 2.43381038e-01 -1.82605520e-01 -1.41642714e+00
-1.21377063e+00 4.21844184e-01 7.11362362e-01 1.39018074e-01
-1.13071513e+00 5.76493621e-01 5.29664934e-01 -7.74228394e-01
-2.20845133e-01 -4.09201473e-01 -7.41603668e-04 -2.22334042e-01
1.34586692e-01 -3.36987041e-02 2.95061558e-01 -1.29843235e+00
-4.75493968e-01 1.42224848e+00 9.34543833e-03 3.54028046e-01
1.30531538e+00 -9.73020494e-02 -8.38310272e-02 1.25946194e-01
1.37715232e+00 1.23494685e-01 -1.43555367e+00 -4.88749236e-01
-2.62121350e-01 -1.03502691e+00 7.24530593e-03 -1.94955334e-01
-1.38933611e+00 6.55179918e-01 7.88234055e-01 -1.37223437e-01
9.36289608e-01 2.69090176e-01 7.15180278e-01 1.48811117e-01
5.85898042e-01 -5.42385697e-01 -3.91488038e-02 4.52474236e-01
9.06638801e-01 -1.59640801e+00 2.76086330e-01 -9.08335984e-01
-6.23573124e-01 1.06606281e+00 7.30305374e-01 -2.60235846e-01
6.92276120e-01 1.16182286e-02 5.61814718e-02 -6.03184640e-01
-8.68200660e-02 -7.09137768e-02 8.39342952e-01 8.65283549e-01
-5.43394126e-03 -1.53909326e-01 2.64039218e-01 4.18698132e-01
-4.47728902e-01 -1.32490799e-01 1.03060484e-01 4.87255067e-01
8.16385821e-02 -1.00386512e+00 -6.14521921e-01 3.84845495e-01
1.22105861e-02 1.00148834e-01 -2.87296623e-01 1.00671422e+00
2.52834380e-01 5.12404382e-01 3.70202541e-01 -5.22565007e-01
5.38899601e-01 -4.89116192e-01 4.79851812e-01 -5.36000848e-01
-5.40815413e-01 -2.37342250e-02 -3.62554908e-01 -7.38490522e-01
-4.98884171e-01 -5.58088601e-01 -1.29752409e+00 -1.65112361e-01
-6.43935382e-01 -3.38461734e-02 5.19022405e-01 1.07853639e+00
7.30701327e-01 1.95658524e-02 8.72444689e-01 -1.41899538e+00
-3.60527903e-01 -4.82238770e-01 -2.09665403e-01 6.57241404e-01
2.61685073e-01 -1.05725098e+00 -3.43354076e-01 -2.75365800e-01] | [7.694187641143799, -3.0463266372680664] |
b8cc8b6c-c5ea-4911-a3f0-a65c0c95ab23 | neuro-symbolic-program-synthesis | 1611.01855 | null | http://arxiv.org/abs/1611.01855v1 | http://arxiv.org/pdf/1611.01855v1.pdf | Neuro-Symbolic Program Synthesis | Recent years have seen the proposal of a number of neural architectures for
the problem of Program Induction. Given a set of input-output examples, these
architectures are able to learn mappings that generalize to new test inputs.
While achieving impressive results, these approaches have a number of important
limitations: (a) they are computationally expensive and hard to train, (b) a
model has to be trained for each task (program) separately, and (c) it is hard
to interpret or verify the correctness of the learnt mapping (as it is defined
by a neural network). In this paper, we propose a novel technique,
Neuro-Symbolic Program Synthesis, to overcome the above-mentioned problems.
Once trained, our approach can automatically construct computer programs in a
domain-specific language that are consistent with a set of input-output
examples provided at test time. Our method is based on two novel neural
modules. The first module, called the cross correlation I/O network, given a
set of input-output examples, produces a continuous representation of the set
of I/O examples. The second module, the Recursive-Reverse-Recursive Neural
Network (R3NN), given the continuous representation of the examples,
synthesizes a program by incrementally expanding partial programs. We
demonstrate the effectiveness of our approach by applying it to the rich and
complex domain of regular expression based string transformations. Experiments
show that the R3NN model is not only able to construct programs from new
input-output examples, but it is also able to construct new programs for tasks
that it had never observed before during training. | ['Dengyong Zhou', 'Abdel-rahman Mohamed', 'Emilio Parisotto', 'Rishabh Singh', 'Pushmeet Kohli', 'Lihong Li'] | 2016-11-06 | null | null | null | null | ['program-induction'] | ['computer-code'] | [ 6.62014127e-01 2.44917586e-01 -3.36819291e-01 -4.32501525e-01
-2.41564184e-01 -5.53383827e-01 4.41637248e-01 1.63575530e-01
-8.23706165e-02 6.97157323e-01 -4.86404181e-01 -9.22449648e-01
-7.70157203e-02 -1.30912733e+00 -1.23171926e+00 -1.81355327e-01
-1.25950247e-01 4.39946294e-01 4.21450675e-01 -2.71513849e-01
2.27718875e-01 6.39258087e-01 -2.08195162e+00 4.14632708e-01
9.94642079e-01 5.48192680e-01 2.79194236e-01 1.01259816e+00
-3.16171616e-01 8.33673060e-01 -5.99621952e-01 -1.50294363e-01
9.03335214e-02 -6.91344142e-01 -1.01752710e+00 -1.93954334e-01
1.34409010e-01 -1.18729837e-01 -8.71245489e-02 1.27856457e+00
-1.26959860e-01 8.49704072e-03 4.13191646e-01 -1.32366240e+00
-6.94493234e-01 7.93199480e-01 1.24556385e-01 -2.80860871e-01
4.99345928e-01 1.70549855e-01 8.95882249e-01 -6.09830379e-01
6.96015179e-01 1.07147658e+00 6.69831753e-01 7.95186460e-01
-1.77119732e+00 -3.94744605e-01 -1.38919607e-01 -1.61715016e-01
-1.10061002e+00 -1.19343074e-02 6.61194921e-01 -4.87859458e-01
1.27281725e+00 4.20654774e-01 7.02926040e-01 6.78243279e-01
-5.51783331e-02 5.61813235e-01 1.09284294e+00 -8.77358198e-01
2.95832247e-01 6.40966475e-01 3.43494177e-01 9.73504782e-01
1.62007689e-01 4.96247768e-01 3.59283313e-02 -2.92434812e-01
7.27613389e-01 -1.14563167e-01 -2.16914162e-01 -5.16044021e-01
-9.06318128e-01 7.77901173e-01 3.31769735e-01 7.42013991e-01
4.58232164e-02 2.84316838e-01 5.36800683e-01 7.49118626e-01
-7.00971931e-02 7.53479123e-01 -6.88437581e-01 3.29316668e-02
-8.90989006e-01 4.37329918e-01 1.19110572e+00 1.05972016e+00
1.09399986e+00 2.66740680e-01 2.60222137e-01 4.73983198e-01
2.69895457e-02 3.32105488e-01 7.38435149e-01 -5.79582930e-01
4.08811957e-01 1.00171089e+00 -2.63414145e-01 -8.12850595e-01
-1.77335307e-01 -1.52576983e-01 -5.71254849e-01 6.91299558e-01
3.69508326e-01 -1.43986121e-01 -8.36671233e-01 2.03079700e+00
1.63757685e-03 -8.57000649e-02 2.43067876e-01 2.22899944e-01
5.82519889e-01 8.07555556e-01 -1.90048605e-01 -4.25822660e-02
7.98047304e-01 -7.87095428e-01 -2.41245523e-01 -2.53213704e-01
1.10599756e+00 -2.48711035e-01 1.28294098e+00 3.23606968e-01
-1.17635703e+00 -9.71180499e-01 -1.26007843e+00 3.02482247e-01
-8.12836766e-01 1.22477576e-01 6.47751451e-01 7.91461706e-01
-1.31488156e+00 8.07360709e-01 -4.95925337e-01 -1.95759341e-01
7.59015558e-03 7.61530101e-01 -3.07575524e-01 7.69369826e-02
-9.16156650e-01 7.66999185e-01 1.02808738e+00 -1.86955318e-01
-8.63232851e-01 -4.67917800e-01 -1.08563256e+00 3.39981496e-01
3.61624032e-01 -5.41309774e-01 1.42082858e+00 -1.64795780e+00
-1.54773390e+00 7.69422412e-01 -2.10979596e-01 -5.05387843e-01
1.45037726e-01 2.55003631e-01 -5.71397662e-01 -2.78451324e-01
-2.11657777e-01 4.77593839e-01 4.55665290e-01 -1.37222111e+00
-5.91388226e-01 -8.83043930e-02 2.12232679e-01 -4.84583586e-01
-1.84652969e-01 4.70872223e-02 -3.20650376e-02 -4.90041852e-01
-1.48013741e-01 -9.82074678e-01 -2.77067691e-01 -3.90796810e-01
-4.10895556e-01 -1.90122858e-01 7.46113360e-01 -3.28560442e-01
1.33602011e+00 -2.14315557e+00 2.44617522e-01 6.99879825e-01
-1.45001307e-01 5.79655290e-01 -2.02361062e-01 2.89703310e-01
-5.57803810e-01 3.19604009e-01 -6.46706104e-01 2.67378300e-01
1.25239760e-01 4.72179443e-01 -5.28172195e-01 -9.42707881e-02
6.05375469e-01 1.05034316e+00 -8.17980349e-01 -1.50307640e-01
-8.97670835e-02 -3.67704779e-02 -8.82032990e-01 4.83075470e-01
-8.94114733e-01 1.73548758e-01 -2.56120652e-01 2.61432856e-01
2.72074401e-01 -1.80543691e-01 3.61206591e-01 2.87725300e-01
-8.37716013e-02 2.38196060e-01 -1.16657043e+00 1.39457381e+00
-7.05221534e-01 6.67429745e-01 -4.70487356e-01 -1.36193681e+00
1.29315984e+00 3.92176092e-01 -1.80742756e-01 -4.43562567e-01
5.70265427e-02 5.29061079e-01 3.08541536e-01 -7.98090518e-01
4.31064755e-01 -1.19292781e-01 -2.39866063e-01 7.51525640e-01
2.74666727e-01 -3.51930201e-01 4.68019783e-01 -1.12066492e-01
1.22826445e+00 4.40104127e-01 4.54243630e-01 -1.01859927e-01
1.02281582e+00 2.35653818e-01 4.67700213e-01 9.02584434e-01
5.44105053e-01 2.69500822e-01 9.14690256e-01 -8.03556621e-01
-1.22175193e+00 -7.48893142e-01 2.13943794e-01 1.07278800e+00
-4.09341931e-01 -2.24644706e-01 -7.76317239e-01 -7.70058393e-01
-3.85310091e-02 1.05385435e+00 -7.14861333e-01 -3.15530837e-01
-1.02337289e+00 -2.51251131e-01 8.13857794e-01 6.59061551e-01
1.88056067e-01 -1.40508425e+00 -7.95536220e-01 2.36705810e-01
3.02597672e-01 -7.30720997e-01 -9.53106433e-02 6.46816909e-01
-1.22493947e+00 -1.14179003e+00 -2.93751836e-01 -1.30645370e+00
1.20338559e+00 -3.93101394e-01 1.21900368e+00 5.48937023e-01
-2.99866050e-02 2.18289986e-01 -2.08753824e-01 -2.83848912e-01
-1.26347160e+00 1.05410285e-01 -3.03004444e-01 -1.49284512e-01
3.10591370e-01 -7.24751055e-01 4.94068682e-01 1.10057175e-01
-1.39408886e+00 1.78919882e-02 7.56709397e-01 1.18896461e+00
4.50528890e-01 3.15722764e-01 6.65706635e-01 -1.41791022e+00
7.33171225e-01 -2.98368931e-01 -1.04361379e+00 4.60472316e-01
-7.71377683e-01 7.14824796e-01 1.21311641e+00 -6.70283139e-01
-9.19108510e-01 3.28332722e-01 -1.71354964e-01 -6.53635561e-02
-4.59996909e-01 8.87311220e-01 -1.56177193e-01 -9.24786553e-02
1.10793996e+00 3.98591727e-01 1.04446247e-01 -2.93068975e-01
2.42602497e-01 3.67821664e-01 7.24995315e-01 -1.04407489e+00
9.91384208e-01 -2.29901493e-01 9.94136408e-02 -4.33229178e-01
-1.85685828e-01 1.87276483e-01 -7.41713524e-01 1.23770379e-01
4.94786024e-01 -2.79944301e-01 -4.88495439e-01 2.76837766e-01
-1.40833676e+00 -7.66500473e-01 -4.79244500e-01 1.25291467e-01
-6.93324864e-01 1.23804621e-01 -2.19345808e-01 -7.48615444e-01
-6.42327741e-02 -1.37882948e+00 4.70791191e-01 2.09522694e-01
-4.45540041e-01 -1.08593678e+00 1.48515344e-01 -4.67716604e-01
3.36019874e-01 3.59068751e-01 1.76966715e+00 -1.08822811e+00
-5.57553649e-01 -4.94619817e-01 6.50733337e-02 7.71637380e-01
1.10283149e-02 1.19528197e-01 -6.55562520e-01 -3.29694599e-02
6.83730692e-02 -4.18775022e-01 2.07980841e-01 -8.52476880e-02
1.47024214e+00 -3.99111807e-01 -2.49663070e-01 5.33833265e-01
1.40011859e+00 5.57309330e-01 7.10793793e-01 3.22578937e-01
1.94429159e-01 4.43859607e-01 2.75404572e-01 -2.65438437e-01
-3.83093059e-02 5.57173193e-01 1.88421667e-01 -1.54046863e-02
1.30784348e-01 -3.55089873e-01 5.98073244e-01 5.58610857e-01
-1.36119857e-01 1.80274442e-01 -1.03584754e+00 5.69256365e-01
-1.66150498e+00 -8.06937277e-01 8.17127377e-02 2.25779366e+00
1.07306099e+00 4.02925313e-01 2.01259241e-01 3.87972444e-01
5.81447661e-01 -3.55615705e-01 -4.38689619e-01 -1.04385352e+00
1.32583633e-01 9.72111702e-01 7.13320151e-02 4.91382480e-01
-6.40538394e-01 7.82473028e-01 6.28814173e+00 3.19254488e-01
-1.21178496e+00 -3.03926796e-01 3.09044242e-01 4.26788747e-01
-4.92072403e-01 1.80311695e-01 -5.57746351e-01 7.02917799e-02
1.19959688e+00 -4.34382141e-01 7.36952245e-01 1.19898725e+00
-4.67009723e-01 1.78571828e-02 -1.85633934e+00 3.74842525e-01
9.46987271e-02 -1.31290495e+00 9.45597216e-02 -2.65032649e-01
6.46699309e-01 -4.15936261e-01 -2.49719657e-02 7.68771112e-01
4.14989233e-01 -1.25328791e+00 5.97092986e-01 3.58165264e-01
7.84548223e-01 -8.48533809e-01 6.91088378e-01 6.19310737e-01
-8.50208044e-01 -3.34967017e-01 -9.25924405e-02 -1.85810789e-01
-4.71277803e-01 2.99318850e-01 -1.08635557e+00 4.34789389e-01
3.04652780e-01 3.17589164e-01 -6.75138831e-01 8.08446527e-01
-5.85151911e-01 3.53277206e-01 -3.49742267e-03 -4.48909611e-01
1.15898289e-01 -4.10516895e-02 3.38419497e-01 1.30937588e+00
5.73162317e-01 -8.10682625e-02 -8.34015608e-02 1.43520188e+00
-2.09358837e-02 -8.90446305e-02 -1.18418574e+00 -1.72715500e-01
1.37933493e-01 8.68644118e-01 -4.60027605e-01 -5.67877889e-01
-4.20679748e-01 4.52506423e-01 4.08816725e-01 4.54059184e-01
-6.53203189e-01 -8.63436699e-01 6.95742816e-02 -1.39445916e-01
3.60446423e-01 -9.80281159e-02 -3.34753960e-01 -8.83967936e-01
2.02555999e-01 -1.35584998e+00 1.52573064e-01 -7.46952713e-01
-6.69234872e-01 8.79030824e-01 1.45444557e-01 -9.42589045e-01
-8.64153922e-01 -8.33810568e-01 -7.52658904e-01 1.09763265e+00
-1.13851416e+00 -7.26756394e-01 -8.68280455e-02 4.56768870e-01
1.58930019e-01 -5.10682225e-01 1.27066123e+00 1.72186226e-01
-2.40819499e-01 6.90750539e-01 -3.63462359e-01 3.55021834e-01
1.10705467e-02 -1.28837442e+00 5.21148682e-01 9.53004718e-01
4.09381576e-02 9.46797550e-01 6.85481846e-01 -4.28189069e-01
-1.55159509e+00 -1.16566813e+00 9.96280491e-01 -2.20267802e-01
6.83991134e-01 -2.81241089e-01 -1.22271669e+00 1.07561469e+00
-5.00499457e-02 -5.53912818e-02 5.78028202e-01 5.85954189e-02
-5.92593431e-01 -1.05769344e-01 -1.02532196e+00 6.31817162e-01
7.30155826e-01 -7.43507564e-01 -9.66660261e-01 1.71074733e-01
8.02090824e-01 -5.11045516e-01 -7.46255875e-01 4.90572274e-01
4.33703333e-01 -1.07845950e+00 6.79576159e-01 -1.04701972e+00
7.92518616e-01 -4.31502104e-01 -7.96972960e-02 -1.30847454e+00
8.99164602e-02 -3.96408767e-01 -2.56624222e-01 1.02615321e+00
9.05499756e-01 -9.18111384e-01 6.58873439e-01 6.04579747e-01
-2.80710638e-01 -8.54462504e-01 -5.93443692e-01 -9.50841248e-01
1.47687912e-01 -5.52661479e-01 1.01801419e+00 7.70880699e-01
1.56963125e-01 2.80574590e-01 -3.41429794e-03 4.47840020e-02
-7.79164881e-02 5.15649319e-01 1.11544752e+00 -1.15455151e+00
-9.68554676e-01 -6.22318387e-01 -3.55165392e-01 -8.76388490e-01
5.03753483e-01 -1.25196123e+00 1.66746557e-01 -9.97106016e-01
-1.05759956e-01 -6.87502980e-01 -5.18579222e-02 7.42844224e-01
1.42093092e-01 -4.00722355e-01 -9.84739214e-02 -1.73756152e-01
3.50377932e-02 2.75205765e-02 9.11718845e-01 -2.61436462e-01
-5.33215582e-01 3.34180683e-01 -5.47926009e-01 7.10183561e-01
7.39383876e-01 -7.40014493e-01 -4.18843091e-01 -2.74677545e-01
5.50430238e-01 2.92692900e-01 4.41716224e-01 -1.16278052e+00
2.54884839e-01 -1.74934193e-01 1.20514899e-01 -2.81651020e-01
-1.92433178e-01 -9.87080812e-01 5.17228961e-01 9.02959466e-01
-6.52181983e-01 3.72393101e-01 4.08671796e-01 1.71797216e-01
-3.83559763e-01 -9.51091528e-01 5.73443890e-01 -2.02090949e-01
-6.81436837e-01 -2.27905795e-01 -3.42066497e-01 -1.84920833e-01
9.00019050e-01 -1.53784707e-01 -1.14892699e-01 5.39523885e-02
-6.51308239e-01 -1.45468697e-01 5.84849000e-01 2.60478586e-01
5.14118612e-01 -1.28107250e+00 -3.14937472e-01 6.81353033e-01
2.00497404e-01 1.06286064e-01 -3.64933759e-01 4.59982783e-01
-6.16242588e-01 3.87294948e-01 -2.41952166e-01 -5.31401992e-01
-1.25788295e+00 8.81311595e-01 4.01428849e-01 -4.93420452e-01
-4.13172334e-01 3.95380497e-01 -3.69542986e-02 -1.03486311e+00
2.04498008e-01 -7.85821021e-01 -6.09051101e-02 -6.92098677e-01
4.26940829e-01 -1.72379673e-01 4.74422500e-02 -1.78272769e-01
1.74317896e-01 3.62108946e-01 1.53421819e-01 -4.84075882e-02
1.50942969e+00 8.86777878e-01 -5.60665786e-01 5.35508513e-01
1.18854868e+00 -1.92160219e-01 -5.35379350e-01 -4.09567267e-01
3.88507307e-01 -2.73319483e-01 -5.19651413e-01 -7.03838766e-01
-7.72686601e-01 8.49466264e-01 2.67772317e-01 5.56131899e-01
1.30973065e+00 -2.31952295e-01 4.97658938e-01 9.53444242e-01
3.25291693e-01 -8.55575204e-01 6.36397675e-02 8.70439768e-01
7.59543657e-01 -7.42084742e-01 -2.17458099e-01 -2.46982723e-01
-1.41134247e-01 1.61235976e+00 6.42394423e-01 -2.23621905e-01
2.77726710e-01 5.16171515e-01 -4.09048915e-01 -1.06268730e-02
-7.70432234e-01 -3.49436179e-02 2.83967167e-01 6.97990000e-01
4.64768499e-01 -7.64214694e-02 -2.14165393e-02 5.00861883e-01
-3.34149510e-01 4.36190456e-01 6.05369031e-01 1.15656900e+00
-3.89493495e-01 -1.63236117e+00 -4.29213196e-01 4.46952224e-01
-1.26839623e-01 9.29914191e-02 -4.05831277e-01 1.20301592e+00
2.21110299e-01 4.04385507e-01 -1.51411787e-01 -6.56263411e-01
5.00846148e-01 4.96258110e-01 6.05996609e-01 -9.82777536e-01
-7.50725031e-01 -5.74428737e-01 1.34081438e-01 -2.42103487e-01
-2.82451808e-01 -4.52449352e-01 -1.30735373e+00 -1.96947847e-02
-1.86968029e-01 3.15462530e-01 6.88587606e-01 9.22515512e-01
1.18897729e-01 7.19310045e-01 6.53618693e-01 -5.87405324e-01
-6.96409881e-01 -6.14015460e-01 -1.80739671e-01 3.87889028e-01
3.35765213e-01 -1.85764834e-01 -2.20533773e-01 3.45459014e-01] | [8.232442855834961, 7.387650012969971] |
82967c0d-ec36-4626-9ad2-2c050fb9636d | multi-stream-extension-of-variational | 2305.13580 | null | https://arxiv.org/abs/2305.13580v1 | https://arxiv.org/pdf/2305.13580v1.pdf | Multi-Stream Extension of Variational Bayesian HMM Clustering (MS-VBx) for Combined End-to-End and Vector Clustering-based Diarization | Combining end-to-end neural speaker diarization (EEND) with vector clustering (VC), known as EEND-VC, has gained interest for leveraging the strengths of both methods. EEND-VC estimates activities and speaker embeddings for all speakers within an audio chunk and uses VC to associate these activities with speaker identities across different chunks. EEND-VC generates thus multiple streams of embeddings, one for each speaker in a chunk. We can cluster these embeddings using constrained agglomerative hierarchical clustering (cAHC), ensuring embeddings from the same chunk belong to different clusters. This paper introduces an alternative clustering approach, a multi-stream extension of the successful Bayesian HMM clustering of x-vectors (VBx), called MS-VBx. Experiments on three datasets demonstrate that MS-VBx outperforms cAHC in diarization and speaker counting performance. | ['Shoko Araki', 'Lukas Burget', 'Tomohiro Nakatani', 'Atsunori Ogawa', 'Anna Silnova', 'Federico Landini', 'Mireia Diez', 'Naohiro Tawara', 'Marc Delcroix'] | 2023-05-23 | null | null | null | null | ['speaker-diarization'] | ['speech'] | [-1.41921476e-01 7.36838160e-03 4.65714969e-02 -5.70921659e-01
-1.00784576e+00 -4.65155900e-01 7.14008331e-01 2.36033261e-01
-3.56829733e-01 -4.29657251e-02 6.26204491e-01 2.48813815e-02
-1.17160104e-01 -3.13969165e-01 -3.20672601e-01 -6.94689512e-01
-4.16276842e-01 8.44767511e-01 2.23341689e-01 5.39515078e-01
8.01373646e-02 4.15893912e-01 -1.77944863e+00 2.42221266e-01
3.08454484e-01 6.67402327e-01 2.03211069e-01 9.41255569e-01
-4.32815045e-01 4.47685421e-01 -1.02445114e+00 -1.81767091e-01
-1.04720838e-01 -4.04717088e-01 -4.67549503e-01 1.59981906e-01
3.07976693e-01 -1.37658730e-01 -1.07473880e-01 4.81040955e-01
6.74789369e-01 4.06191170e-01 8.83313894e-01 -1.76796329e+00
-3.24101895e-01 1.04836702e+00 -6.37829542e-01 2.34065711e-01
1.61789894e-01 -5.18820584e-01 1.02568483e+00 -1.03647959e+00
4.04378206e-01 1.49708974e+00 7.57970273e-01 9.14427340e-01
-1.40859771e+00 -9.24171627e-01 1.25090450e-01 1.70860410e-01
-1.84790659e+00 -7.37372816e-01 6.90118074e-01 -6.79488957e-01
1.10035324e+00 4.11677271e-01 4.58776742e-01 1.06898177e+00
-5.54670513e-01 1.08175504e+00 4.72489476e-01 -5.05641341e-01
7.45132804e-01 1.92086175e-01 5.88306665e-01 1.62651166e-01
-1.21335551e-01 -2.44623590e-02 -1.12652338e+00 -3.15474957e-01
4.48740929e-01 -8.51136446e-02 7.73111805e-02 -2.75878221e-01
-1.16334069e+00 9.91255999e-01 -2.27690041e-01 3.72711092e-01
-2.95537472e-01 3.64738286e-01 4.98028874e-01 -3.53715979e-02
3.15894395e-01 -3.58196273e-02 2.34909113e-02 -4.70477372e-01
-1.36216438e+00 1.64961264e-01 7.23511636e-01 1.14541841e+00
7.04369724e-01 2.28062198e-01 -1.02071635e-01 1.15034986e+00
8.48001838e-01 2.12486789e-01 8.60758364e-01 -1.18614030e+00
3.40760022e-01 8.10287520e-02 -1.74548268e-01 -5.77634573e-01
-1.90587506e-01 -7.73273781e-02 -6.87789559e-01 6.87855557e-02
7.67668635e-02 -8.40123147e-02 -7.19482839e-01 1.89251602e+00
6.57432914e-01 6.73817694e-01 1.93466529e-01 5.26983678e-01
7.20375061e-01 9.46557462e-01 2.06377611e-01 -2.10569605e-01
1.15966475e+00 -5.50678194e-01 -7.16385603e-01 9.21513536e-04
2.99070805e-01 -4.19380397e-01 6.56000316e-01 5.40002286e-01
-9.62200701e-01 -7.53918886e-01 -9.72243369e-01 1.66707844e-01
-7.90135413e-02 -6.85380772e-02 4.35839385e-01 1.00452316e+00
-1.29142189e+00 1.72102839e-01 -1.19958031e+00 -4.03131962e-01
3.53659600e-01 5.62649786e-01 -1.77937090e-01 3.57204407e-01
-8.91018510e-01 1.48175359e-01 2.14025542e-01 -2.11510316e-01
-1.07613337e+00 -6.56284809e-01 -8.87826562e-01 2.13073924e-01
-1.72762349e-01 -9.02795196e-02 1.38935852e+00 -5.25311530e-01
-1.57680130e+00 6.99690878e-01 -7.75897384e-01 -5.92542291e-01
1.61969125e-01 -1.19685300e-01 -5.14219284e-01 1.00472145e-01
1.20527387e-01 1.06851947e+00 1.14073265e+00 -1.56034482e+00
-8.10688376e-01 -5.91296792e-01 -9.38562572e-01 2.78072245e-02
-4.88666326e-01 2.23340943e-01 -9.34459925e-01 -5.95824361e-01
2.68338412e-01 -8.29720080e-01 2.59573102e-01 -5.28820157e-01
-7.81159997e-01 -9.31532145e-01 8.57711613e-01 -5.43387413e-01
1.68609548e+00 -2.49262953e+00 4.46856230e-01 4.49915975e-01
5.20669878e-01 -7.43573010e-02 5.28623573e-02 4.89091963e-01
1.10083289e-01 -1.99876707e-02 -2.02678666e-01 -1.31375110e+00
5.31277239e-01 3.98548305e-01 -2.62901962e-01 5.03333986e-01
-6.50760308e-02 2.88226575e-01 -5.60650468e-01 -8.44419658e-01
1.59816682e-01 6.57891929e-01 -5.62557757e-01 2.86719590e-01
1.78277433e-01 -2.03812987e-01 2.38051593e-01 4.59991008e-01
7.71176755e-01 3.75869960e-01 4.50569659e-01 2.29844004e-01
-1.94749832e-01 4.99933586e-02 -1.64121807e+00 1.35848916e+00
-2.91816443e-01 1.00102627e+00 1.66017875e-01 -7.91891873e-01
7.64649808e-01 4.99408901e-01 4.49734718e-01 -4.64364514e-02
-3.15630399e-02 -1.42852202e-01 -3.72366071e-01 -2.69483685e-01
6.04078352e-01 -1.58480018e-01 -2.81126916e-01 7.02395380e-01
6.03242934e-01 3.99141759e-01 -2.85429191e-02 6.39879465e-01
9.42916274e-01 -4.65271115e-01 1.19150616e-01 1.21609338e-01
1.60759687e-01 -4.85162914e-01 6.52956486e-01 8.18477690e-01
-4.67371434e-01 6.09600127e-01 3.33269417e-01 4.85000014e-01
-8.31959307e-01 -1.58826458e+00 -1.84517562e-01 1.50806701e+00
-2.40140617e-01 -5.88954329e-01 -9.91322100e-01 -4.14863497e-01
2.75953531e-01 7.67296255e-01 -6.83545291e-01 1.92817688e-01
-4.36048239e-01 -4.82627004e-01 9.80435312e-01 6.89673007e-01
-9.93841812e-02 -8.68482769e-01 -3.72894734e-01 4.49207753e-01
-1.70637742e-01 -9.61138725e-01 -7.03862906e-01 5.29098690e-01
-5.36056936e-01 -4.51111019e-01 -6.08397424e-01 -9.76277173e-01
1.02583811e-01 4.28329825e-01 7.81538725e-01 -3.90472919e-01
-1.38882011e-01 6.89621925e-01 -4.43484098e-01 -5.87619007e-01
-7.99749732e-01 8.08785036e-02 6.31824613e-01 4.30069238e-01
9.33395624e-01 -6.07026339e-01 1.23629114e-03 3.63380164e-01
-6.77155733e-01 -6.09886169e-01 1.52669951e-01 2.89812207e-01
5.56800604e-01 -3.46352942e-02 7.61013269e-01 -5.48584938e-01
4.36631203e-01 -7.07418501e-01 -3.21609616e-01 -1.46897435e-01
-2.86792010e-01 -1.41544044e-01 -1.04781054e-02 -6.87731564e-01
-7.64021456e-01 1.76930189e-01 -3.21091205e-01 -8.15605223e-01
-4.88792151e-01 1.63081214e-01 -5.16284347e-01 5.68863809e-01
4.66309428e-01 4.07208890e-01 6.55228794e-02 -6.92735434e-01
5.49698889e-01 1.26198983e+00 8.43954802e-01 -4.69082817e-02
6.96069956e-01 3.42746943e-01 -7.46605814e-01 -1.44425404e+00
-1.69778496e-01 -1.14389455e+00 -7.40493238e-01 -3.94719690e-01
1.24909627e+00 -1.07398844e+00 -7.80712366e-01 6.00553334e-01
-1.01273084e+00 -3.76179695e-01 -2.96649098e-01 8.18525672e-01
-2.16714293e-01 5.57729959e-01 -4.33880448e-01 -1.41627753e+00
-2.60433778e-02 -7.88186073e-01 1.28782511e+00 -1.63115617e-02
-5.93770981e-01 -8.19475889e-01 4.45716411e-01 1.18786342e-01
-5.74036241e-02 -1.03638478e-01 5.83512962e-01 -1.13878167e+00
-1.02894522e-01 -6.56571612e-02 2.42475092e-01 3.60771775e-01
-1.78491712e-01 3.16516519e-01 -1.42487645e+00 -1.51035368e-01
-1.48483530e-01 6.52615577e-02 8.45121562e-01 6.31678879e-01
8.67725551e-01 -1.68117717e-01 -6.24598563e-01 2.18830973e-01
1.05386496e+00 3.82556915e-01 5.40169299e-01 -3.49936001e-02
8.35274339e-01 6.27915800e-01 2.20733315e-01 6.39294684e-01
5.86442411e-01 9.67390478e-01 9.44482312e-02 2.46130481e-01
-1.67355895e-01 -2.19888434e-01 7.72729635e-01 1.39077044e+00
3.98400992e-01 -3.44912678e-01 -8.30547869e-01 8.93798828e-01
-1.51519239e+00 -1.43645692e+00 -3.33470702e-01 2.24588919e+00
7.80421257e-01 -4.08385433e-02 8.83430719e-01 7.37743139e-01
1.15129542e+00 -1.50687262e-01 -3.95145059e-01 -4.52861190e-01
6.72235340e-02 1.77459672e-01 1.39543474e-01 7.42107034e-01
-1.03321028e+00 7.96570837e-01 5.95883512e+00 8.57674420e-01
-6.65859818e-01 5.42178094e-01 -4.54887040e-02 -1.87285841e-01
-8.37857872e-02 -4.59354907e-01 -1.40685022e+00 6.97186410e-01
1.40525103e+00 1.63294733e-01 2.67882735e-01 8.89892042e-01
1.39345825e-01 5.92800975e-03 -1.31087482e+00 1.20499027e+00
3.69434476e-01 -1.20890558e+00 -1.50537953e-01 3.92033964e-01
3.41118723e-01 1.53308108e-01 1.34972278e-02 5.39410055e-01
6.19450629e-01 -6.47272289e-01 9.86881971e-01 1.46950528e-01
6.05483294e-01 -8.53493273e-01 3.43788832e-01 2.43155763e-01
-1.40764558e+00 -8.78593624e-02 -3.15120310e-01 6.36138141e-01
3.10015976e-01 3.69541347e-01 -1.19036698e+00 1.02423795e-01
8.10602784e-01 3.40304494e-01 -4.19471532e-01 1.01657212e+00
2.03084171e-01 1.17066383e+00 -4.17584509e-01 7.87581503e-02
8.22720677e-02 -2.22296845e-02 6.19898260e-01 1.69629979e+00
2.92727947e-01 -1.85848802e-01 -1.85567960e-01 5.01340508e-01
1.98260434e-02 -1.84587866e-01 -1.48925498e-01 4.61145416e-02
1.46449661e+00 9.38268661e-01 -5.43550849e-01 -6.08557761e-01
-6.64370805e-02 1.09948003e+00 2.41783246e-01 1.89068168e-01
-1.07296610e+00 -5.71749628e-01 1.09738255e+00 -5.52540459e-02
8.10933650e-01 -4.25051689e-01 -4.24167179e-02 -5.81419766e-01
-4.63214219e-01 -5.64294159e-01 4.34364408e-01 -3.56597692e-01
-1.28615654e+00 4.69562560e-01 9.42309350e-02 -1.28235698e+00
-3.29044700e-01 4.75525223e-02 -8.03680241e-01 4.90999579e-01
-1.01050234e+00 -8.09930384e-01 -2.87113190e-01 7.38053679e-01
7.25581527e-01 -3.19360614e-01 9.02898848e-01 4.46240395e-01
-7.68693566e-01 1.02855456e+00 5.11506736e-01 1.51220307e-01
6.19351625e-01 -1.63047099e+00 4.76086348e-01 5.62805772e-01
6.33386374e-01 5.41423917e-01 5.67533255e-01 -2.85146892e-01
-1.01393199e+00 -1.47835338e+00 1.26656222e+00 -6.85586691e-01
2.60978401e-01 -9.20004606e-01 -1.13921320e+00 7.71463394e-01
1.71078220e-01 -7.69292474e-01 1.49347484e+00 3.70689750e-01
-4.39753711e-01 -5.93237542e-02 -9.04230535e-01 1.66630730e-01
8.10373664e-01 -7.25372434e-01 -7.71240294e-01 9.52656716e-02
8.34033072e-01 1.65192723e-01 -9.78154957e-01 -5.76733232e-01
4.20787036e-01 -9.90541995e-01 1.06091106e+00 -2.51130551e-01
-2.33356342e-01 -4.70496088e-01 -6.34014666e-01 -1.20170379e+00
-3.66175652e-01 -6.66627586e-01 -4.19203281e-01 2.06435585e+00
2.29027092e-01 -3.08971196e-01 6.25976324e-01 1.90142974e-01
-4.76516038e-01 -1.33405149e-01 -1.33878756e+00 -1.07807624e+00
3.02514713e-02 -1.02761698e+00 8.73517275e-01 1.12566113e+00
2.16493472e-01 4.21866685e-01 -2.09078074e-01 7.04992890e-01
9.58935320e-01 -3.07884991e-01 9.28052008e-01 -1.63024819e+00
-4.19098169e-01 -4.24063087e-01 -7.13633358e-01 -1.04486358e+00
4.00640965e-01 -1.06588435e+00 3.46513003e-01 -1.44241703e+00
-7.33066201e-02 -4.64414090e-01 1.49396211e-01 3.36660713e-01
-4.92957458e-02 2.48489290e-01 1.82761818e-01 5.74185610e-01
-6.89044714e-01 4.70393002e-01 -5.90225011e-02 -4.05303240e-02
-7.70918608e-01 4.46001766e-03 -3.96059036e-01 3.89337182e-01
6.44191921e-01 -6.47267759e-01 -2.57292002e-01 -3.59345049e-01
-5.46908319e-01 -3.18980850e-02 4.33918387e-02 -1.34128761e+00
3.81109804e-01 5.20600855e-01 3.29289347e-01 -9.47135389e-01
6.37376606e-01 -4.28733051e-01 2.75534332e-01 1.18716069e-01
-4.38255340e-01 -1.56725392e-01 4.62621450e-02 9.85986292e-01
-3.11987817e-01 -8.76567662e-02 7.51604021e-01 4.12153840e-01
-6.70291364e-01 -2.21139155e-02 -1.11142147e+00 -7.29787648e-02
1.11761391e+00 -5.50592482e-01 2.28466183e-01 -2.94853419e-01
-1.08625519e+00 2.53088504e-01 -1.00979358e-02 3.17636728e-01
7.14677572e-01 -1.37854469e+00 -6.28158331e-01 2.40232185e-01
2.58127153e-01 -1.38819560e-01 4.03498650e-01 5.88684976e-01
-1.22320585e-01 3.35051984e-01 2.72519231e-01 -1.07549369e+00
-1.84547246e+00 5.17391443e-01 -1.64256617e-02 1.71182692e-01
-4.16555792e-01 1.34866130e+00 2.89571453e-02 -5.12728870e-01
8.91493201e-01 -2.54490942e-01 -2.69105941e-01 6.87198579e-01
6.25625312e-01 5.76918662e-01 -2.44015511e-02 -8.46819699e-01
-6.04618847e-01 2.48289198e-01 -1.10981338e-01 -7.72459209e-01
1.17926788e+00 -3.36149156e-01 3.58359426e-01 8.43749583e-01
1.43624997e+00 5.90527616e-02 -1.17654622e+00 -2.25141093e-01
2.53937751e-01 -2.46229738e-01 9.05583352e-02 -2.38502517e-01
-7.10009217e-01 1.04336166e+00 7.98676908e-01 4.12356108e-01
6.38317108e-01 4.62459207e-01 8.31113636e-01 -2.41956543e-02
4.90702968e-03 -1.08410847e+00 1.49312213e-01 2.00970784e-01
4.73395467e-01 -7.11902797e-01 -4.54514146e-01 4.18180078e-02
-1.11327565e+00 7.14942873e-01 3.36051971e-01 2.10024133e-01
7.94566095e-01 7.71161318e-02 2.45415881e-01 -2.08455428e-01
-5.64274967e-01 -2.17160031e-01 -1.76650062e-01 1.00460207e+00
1.36898130e-01 3.20015162e-01 4.80112255e-01 7.13155925e-01
-2.96907634e-01 -5.50333261e-01 4.58864719e-01 8.27806890e-01
-7.28837430e-01 -9.86190259e-01 -7.60339022e-01 2.63966739e-01
-5.92031702e-02 9.15876552e-02 -4.28201586e-01 6.07145846e-01
2.11729705e-01 1.21829700e+00 6.40544176e-01 -7.19748020e-01
1.96842536e-01 5.29676259e-01 1.53115794e-01 -6.80943131e-01
-6.43723786e-01 5.81171274e-01 -4.25911397e-02 -2.11122021e-01
-3.99865836e-01 -1.22908521e+00 -1.52010930e+00 -2.76051044e-01
-3.32824111e-01 4.89177436e-01 1.04383421e+00 5.28712273e-01
4.53275561e-01 4.45425123e-01 8.92556429e-01 -1.17857015e+00
-3.91525000e-01 -1.09003305e+00 -1.17108846e+00 1.50071129e-01
4.94709492e-01 -6.26917839e-01 -6.81232393e-01 2.38112628e-01] | [14.503596305847168, 6.190752983093262] |
adc58927-b8e5-41d7-89b4-8d34f3159288 | a-normalized-bottleneck-distance-on | 2306.06727 | null | https://arxiv.org/abs/2306.06727v1 | https://arxiv.org/pdf/2306.06727v1.pdf | A Normalized Bottleneck Distance on Persistence Diagrams and Homology Preservation under Dimension Reduction | Persistence diagrams are used as signatures of point cloud data assumed to be sampled from manifolds, and represent their topology in a compact fashion. Further, two given clouds of points can be compared by directly comparing their persistence diagrams using the bottleneck distance, d_B. But one potential drawback of this pipeline is that point clouds sampled from topologically similar manifolds can have arbitrarily large d_B values when there is a large degree of scaling between them. This situation is typical in dimension reduction frameworks that are also aiming to preserve topology. We define a new scale-invariant distance between persistence diagrams termed normalized bottleneck distance, d_N, and study its properties. In defining d_N, we also develop a broader framework called metric decomposition for comparing finite metric spaces of equal cardinality with a bijection. We utilize metric decomposition to prove a stability result for d_N by deriving an explicit bound on the distortion of the associated bijective map. We then study two popular dimension reduction techniques, Johnson-Lindenstrauss (JL) projections and metric multidimensional scaling (mMDS), and a third class of general biLipschitz mappings. We provide new bounds on how well these dimension reduction techniques preserve homology with respect to d_N. For a JL map f that transforms input X to f(X), we show that d_N(dgm(X),dgm(f(X)) < e, where dgm(X) is the Vietoris-Rips persistence diagram of X, and 0 < e < 1 is the tolerance up to which pairwise distances are preserved by f. For mMDS, we present new bounds for both d_B and d_N between persistence diagrams of X and its projection in terms of the eigenvalues of the covariance matrix. And for k-biLipschitz maps, we show that d_N is bounded by the product of (k^2-1)/k and the ratio of diameters of X and f(X). | ['Nathan H. May', 'Bala Krishnamoorthy'] | 2023-06-11 | null | null | null | null | ['dimensionality-reduction'] | ['methodology'] | [-1.53238997e-01 -1.84864048e-02 1.09713301e-02 1.07571512e-01
-2.32965529e-01 -1.08647406e+00 6.69636607e-01 1.96238562e-01
-1.36405617e-01 4.00707006e-01 1.13088012e-01 -2.35106170e-01
-4.87068206e-01 -1.00950980e+00 -8.94377530e-01 -7.65931308e-01
-4.99858409e-01 4.71370518e-01 2.66956836e-01 -3.10884535e-01
3.04678589e-01 9.49557185e-01 -1.71851826e+00 -5.15951455e-01
7.64790177e-01 5.60443401e-01 -7.71425292e-02 9.31154609e-01
-3.06388754e-02 -1.22366503e-01 -2.93146819e-01 -9.52522531e-02
6.15677416e-01 -3.95807892e-01 -9.97287273e-01 -1.78141857e-03
7.46600330e-01 1.16562970e-01 -3.98173243e-01 1.32867360e+00
-3.95901427e-02 8.85069892e-02 9.76819634e-01 -1.52252758e+00
-4.95522916e-01 2.34398603e-01 -5.39875507e-01 1.95276096e-01
2.07937360e-01 -2.61302799e-01 7.44788110e-01 -7.34366775e-01
9.87813592e-01 1.29663515e+00 6.83998048e-01 3.60936701e-01
-1.71356106e+00 -3.78825366e-01 -2.36383528e-01 -2.98571795e-01
-1.41062737e+00 6.87687695e-02 5.35551786e-01 -8.53646517e-01
3.92130017e-01 6.33498251e-01 5.68293989e-01 2.79196978e-01
3.47263843e-01 -4.42938022e-02 1.05907059e+00 -3.94234419e-01
1.82334155e-01 2.10858643e-01 4.15572941e-01 6.81385458e-01
6.12889826e-01 -1.44532755e-01 1.56870531e-03 -1.81124598e-01
8.80635202e-01 -3.98618877e-02 -2.87087500e-01 -7.79763579e-01
-1.36576974e+00 7.98303545e-01 4.34176594e-01 4.07843202e-01
1.86961561e-01 -2.44150665e-02 5.38927428e-02 5.52142203e-01
6.66603744e-02 3.64413500e-01 -2.34412085e-02 5.08161895e-02
-2.18308106e-01 3.39974850e-01 8.63599479e-01 1.26166308e+00
1.05354679e+00 -6.60584867e-01 5.35690784e-01 3.55259836e-01
-5.74602261e-02 7.43653417e-01 -8.22301432e-02 -1.02584100e+00
3.11901301e-01 7.50173986e-01 -2.51985751e-02 -1.24993527e+00
-3.50129604e-01 4.10618633e-02 -9.65577245e-01 1.67151168e-01
6.40249610e-01 4.99133348e-01 -1.90718517e-01 2.25116944e+00
3.65383863e-01 -1.34023264e-01 2.42659282e-02 5.34457088e-01
1.20857701e-01 5.26278615e-01 -4.34300691e-01 -4.53600138e-01
1.05535257e+00 -2.14394838e-01 -1.11434199e-01 3.41628432e-01
8.77298772e-01 -5.02101839e-01 1.33210206e+00 -7.44743645e-02
-1.04726624e+00 -1.18726104e-01 -1.03200328e+00 1.11461915e-01
-3.96341741e-01 -2.13524997e-01 -7.59928152e-02 4.57984418e-01
-1.14062035e+00 9.07294035e-01 -8.05551767e-01 -8.20456803e-01
-1.23397484e-02 3.36624265e-01 -5.30711591e-01 2.37177581e-01
-6.63748682e-01 8.59890759e-01 -2.00981554e-02 -3.77464950e-01
-3.49477053e-01 -8.64243388e-01 -3.79470438e-01 -2.26310894e-01
-7.27511346e-02 -5.23259759e-01 4.68304396e-01 -2.36607358e-01
-8.42438936e-01 9.59130585e-01 6.84740767e-02 -2.97218561e-01
5.21971524e-01 1.41090810e-01 -3.19104403e-01 1.72749057e-01
2.67501593e-01 3.69729221e-01 6.28873825e-01 -1.14480770e+00
-4.53905970e-01 -7.11751223e-01 2.46222436e-01 2.29806721e-01
-3.52103949e-01 -8.24501365e-02 7.95513391e-02 -1.68242350e-01
4.01220798e-01 -1.32813537e+00 -4.41217460e-02 2.67683059e-01
-3.30198646e-01 -2.46686116e-01 1.14341366e+00 -2.37426490e-01
1.03158557e+00 -2.27699995e+00 5.46143532e-01 3.46965820e-01
6.01912856e-01 -6.30429983e-02 -6.46593869e-02 5.43933392e-01
5.25501035e-02 4.31383729e-01 -5.35692334e-01 1.47201270e-01
-3.67386341e-02 8.13273266e-02 -4.62113678e-01 1.04556632e+00
-6.70360550e-02 2.02041924e-01 -7.91907191e-01 -1.28286585e-01
7.06985891e-02 3.51964116e-01 -4.12967443e-01 -1.30577520e-01
7.31492937e-02 3.16915840e-01 -5.43776564e-02 1.24560490e-01
1.02151787e+00 2.00520512e-02 2.15726346e-02 -1.35786012e-01
-4.85406309e-01 7.48009831e-02 -1.45465374e+00 1.20088470e+00
-1.64359286e-01 5.86517990e-01 1.11085474e-01 -7.38319039e-01
9.44811165e-01 -2.51870185e-01 5.43109596e-01 -1.04774289e-01
9.16142296e-03 2.53088534e-01 -1.63676322e-01 6.14624918e-02
3.95714641e-01 -1.09314419e-01 -1.31259665e-01 5.23597360e-01
-4.24052268e-01 -9.83123705e-02 3.10752153e-01 4.23072219e-01
1.37021160e+00 -5.42792022e-01 4.09502424e-02 -1.06529558e+00
6.97545290e-01 -5.35645559e-02 4.12404716e-01 3.49395871e-01
-9.42824036e-02 3.60867828e-01 9.47599351e-01 -9.03612599e-02
-1.53082943e+00 -1.50804782e+00 -7.44494259e-01 4.46204424e-01
7.98745453e-01 -4.49097097e-01 -7.95446277e-01 -3.50360066e-01
4.06572074e-01 4.13639516e-01 -7.88767636e-01 -3.40358377e-01
-7.00951576e-01 -5.35844922e-01 4.49736297e-01 3.49034995e-01
3.85874003e-01 -2.61867821e-01 -5.12800992e-01 -2.54042655e-01
1.50274530e-01 -7.70203531e-01 -7.75964677e-01 -2.27125347e-01
-1.19609261e+00 -1.24985337e+00 -3.84760261e-01 -7.21305370e-01
8.73794794e-01 4.06141907e-01 7.76198030e-01 -2.20146134e-01
-1.52151793e-01 5.10425150e-01 1.04123794e-01 9.29643735e-02
-6.12408102e-01 4.19968460e-03 7.80990541e-01 -3.06437910e-01
5.06007150e-02 -7.79672503e-01 -5.60575485e-01 9.69986618e-01
-1.04761147e+00 -2.12215796e-01 7.59909526e-02 2.80185670e-01
7.06325412e-01 -8.93232133e-03 1.67771921e-01 -2.23205030e-01
5.11858225e-01 -2.85485804e-01 -9.28415835e-01 9.97774526e-02
-7.82593906e-01 4.03722346e-01 6.09541535e-01 -5.03587306e-01
-9.12020728e-02 -1.06299251e-01 6.94159389e-01 -4.83784765e-01
2.53833532e-01 8.29325467e-02 -2.58500218e-01 -4.91491437e-01
5.73076189e-01 -1.28254563e-01 3.46681386e-01 -4.97754216e-01
7.22255409e-01 4.69106615e-01 6.57432139e-01 -4.99048173e-01
8.93815696e-01 8.12145948e-01 5.89121282e-01 -1.06030643e+00
-9.24416184e-02 -3.11145484e-01 -9.00954127e-01 5.36840111e-02
6.86720133e-01 -3.27730954e-01 -9.66470122e-01 1.11558467e-01
-9.08152044e-01 1.07145764e-01 -4.48343724e-01 4.53958541e-01
-8.79215837e-01 3.44431639e-01 -3.74483705e-01 -5.35222590e-01
1.08748831e-01 -9.09208834e-01 5.82349181e-01 -2.65452057e-01
-8.61015692e-02 -1.00306046e+00 5.79548120e-01 -2.28288800e-01
3.36665750e-01 5.98353267e-01 1.35853922e+00 -3.99111003e-01
-5.31110406e-01 -1.96868524e-01 -3.18650305e-01 3.16320628e-01
8.92564133e-02 2.00621471e-01 -2.02450007e-01 -5.39044499e-01
2.06114382e-01 5.31838179e-01 4.24576402e-01 2.28551045e-01
5.30765533e-01 -6.35676801e-01 -5.84889531e-01 5.91345847e-01
1.53881562e+00 -3.72168459e-02 4.57014859e-01 4.05077189e-01
7.35894799e-01 5.15086412e-01 3.26544136e-01 -1.27621181e-02
2.12259382e-01 7.26102471e-01 3.39648724e-01 3.72557431e-01
-4.13620919e-02 -3.11687998e-02 3.90894264e-01 1.17613459e+00
-4.33111005e-02 3.48367065e-01 -9.77512598e-01 6.07504189e-01
-1.60447061e+00 -7.15059876e-01 -3.53260010e-01 2.77745795e+00
4.51894343e-01 4.18262593e-02 5.71949959e-01 8.91777724e-02
1.13840568e+00 -1.62111133e-01 -5.26527584e-01 -3.98362637e-01
-3.67498815e-01 5.53965308e-02 9.55687702e-01 7.46831536e-01
-7.36836731e-01 3.91552359e-01 6.15390921e+00 2.65988111e-01
-9.37312186e-01 -1.55488877e-02 -2.27763340e-01 -1.60768062e-01
-3.29690963e-01 3.06011140e-01 -7.12912321e-01 5.66662133e-01
1.05267906e+00 -7.42437482e-01 3.02810073e-01 8.21431696e-01
-7.72299394e-02 3.58174518e-02 -1.35907805e+00 8.05078626e-01
-2.10903943e-01 -1.32412183e+00 7.39610419e-02 7.72425532e-01
6.37071550e-01 -4.03980240e-02 1.09230518e-01 -3.26677084e-01
1.33556321e-01 -7.48943210e-01 3.70707810e-01 4.43675190e-01
9.50924039e-01 -9.86396015e-01 1.83460742e-01 2.15835914e-01
-1.56064987e+00 1.93854105e-02 -5.57449877e-01 1.88430220e-01
-1.25168581e-02 5.51411629e-01 -3.84422272e-01 4.72426713e-01
3.90311331e-01 5.73498368e-01 -3.58631432e-01 8.39984179e-01
5.78062892e-01 -4.93395850e-02 -6.05382383e-01 9.86084118e-02
-2.00581085e-02 -9.38760877e-01 1.08465350e+00 7.87133813e-01
5.46205878e-01 1.59966424e-01 -1.34252876e-01 9.09311175e-01
-1.32895455e-01 1.08238600e-01 -8.88595581e-01 -7.01470226e-02
7.10245132e-01 1.12867868e+00 -8.32797050e-01 -9.55986381e-02
-1.91470280e-01 5.97439885e-01 1.42371561e-02 -1.98642597e-01
-6.05772257e-01 -5.69341600e-01 1.32019579e+00 6.58915818e-01
1.72080204e-01 -6.02779806e-01 -1.23487264e-01 -1.03799164e+00
2.22380742e-01 -4.22114730e-01 5.51601425e-02 -4.08455849e-01
-1.08027411e+00 5.22337973e-01 2.05558777e-01 -1.43926334e+00
1.70180395e-01 -5.87105811e-01 -4.24218625e-01 7.34508157e-01
-5.85909843e-01 -6.45885766e-01 -9.18136016e-02 4.66247857e-01
-2.85005629e-01 6.79297075e-02 7.51891017e-01 1.07108399e-01
-3.29271764e-01 2.59113103e-01 6.65083170e-01 -5.03998846e-02
4.49981928e-01 -1.32709718e+00 4.58597124e-01 7.72638440e-01
-1.02448702e-01 6.72076046e-01 8.25003266e-01 -7.02911437e-01
-1.72813988e+00 -1.10877395e+00 6.85898840e-01 -8.14531147e-01
1.01289630e+00 -3.69511366e-01 -9.97313023e-01 7.14370668e-01
-4.26716059e-01 -4.84347455e-02 3.20166200e-01 -1.62742540e-01
-5.26045918e-01 -2.25802258e-01 -1.36375713e+00 4.89404410e-01
1.32887948e+00 -7.95153081e-01 -3.36685032e-01 4.17446762e-01
7.55527496e-01 -1.90036669e-02 -1.30315948e+00 3.99375409e-01
5.37161112e-01 -9.18790817e-01 7.35490501e-01 -4.41004932e-01
-4.97615784e-02 -6.44411922e-01 -4.44992214e-01 -9.88976598e-01
-2.95053184e-01 -7.19187856e-01 9.56162587e-02 1.05650949e+00
3.34973074e-02 -7.56684542e-01 4.98034477e-01 3.91992152e-01
1.36918172e-01 -5.99449098e-01 -1.15877533e+00 -1.45684636e+00
6.02559388e-01 4.94927168e-02 6.30533874e-01 9.47461307e-01
1.54920176e-01 2.86723584e-01 2.32846692e-01 3.67623031e-01
8.60479832e-01 -2.67397892e-02 8.31392229e-01 -1.72991312e+00
2.42921352e-01 -6.55670643e-01 -1.05587363e+00 -7.04791784e-01
9.01818369e-03 -1.23277628e+00 -3.90436947e-01 -1.21047735e+00
9.12653953e-02 -7.08527982e-01 8.11983496e-02 -3.36948365e-01
6.44393086e-01 -1.00154042e-01 1.55549988e-01 7.90726066e-01
-8.81350338e-02 2.99378723e-01 9.23170269e-01 2.39436015e-01
-5.97344637e-01 -2.33167544e-01 -3.70699674e-01 6.89109683e-01
5.87423384e-01 -3.90521675e-01 -2.47909427e-01 -1.21440232e-01
1.96561694e-01 3.56974779e-03 3.46570075e-01 -1.03361177e+00
5.63303679e-02 -5.95123656e-02 -2.53592372e-01 -6.76909566e-01
1.37005672e-01 -5.67254543e-01 7.00886071e-01 4.70796853e-01
-2.36522049e-01 4.07834589e-01 -4.69701029e-02 7.25664198e-01
2.58605629e-01 6.12344071e-02 1.13192594e+00 2.30570927e-01
-2.60639191e-02 3.82213145e-01 -5.11438362e-02 -6.51907772e-02
1.34341478e+00 -3.05907041e-01 -6.63635254e-01 -8.51550028e-02
-7.18609929e-01 5.72036542e-02 1.18757010e+00 1.86876103e-01
3.39952826e-01 -1.68412304e+00 -4.05286521e-01 2.66061157e-01
1.75196484e-01 -3.33992839e-02 -1.23878017e-01 1.20308614e+00
-6.47175789e-01 2.31774285e-01 -4.25637841e-01 -9.60304439e-01
-1.42848086e+00 6.63542926e-01 3.34712029e-01 2.73322374e-01
-7.66594410e-01 4.15696859e-01 2.98346937e-01 -3.76536012e-01
5.89972474e-02 -4.56991524e-01 4.52393949e-01 1.95111305e-01
3.77445966e-01 8.45141232e-01 7.77176693e-02 -9.83694255e-01
-5.33536255e-01 1.05135548e+00 8.97469744e-03 -3.08994502e-01
1.12070131e+00 -3.14025342e-01 -7.01545537e-01 5.45574725e-01
1.58878970e+00 7.23204762e-02 -1.04840147e+00 -6.30607530e-02
-3.51889794e-05 -4.94562924e-01 -5.20381808e-01 1.45185068e-01
-6.59315288e-01 6.39884233e-01 7.40596950e-01 8.16281378e-01
8.61798048e-01 5.34417629e-01 3.67831260e-01 2.85009325e-01
2.91799605e-01 -8.49194407e-01 -7.71255493e-02 4.97158855e-01
9.98520792e-01 -4.71369147e-01 -1.70651034e-01 -5.92593193e-01
1.40925143e-02 1.12258184e+00 2.51614988e-01 -5.40554702e-01
8.64895046e-01 5.45857474e-02 -4.18718934e-01 -1.43804535e-01
-4.16828811e-01 -5.55798970e-02 1.01327792e-01 3.74323875e-01
4.11535194e-03 2.20335722e-01 -4.83550578e-01 -1.47911116e-01
-7.40889192e-01 -3.40737820e-01 8.38188231e-01 8.46417308e-01
-7.69423246e-01 -1.08147848e+00 -4.85015959e-01 3.83673608e-01
2.08461121e-01 3.75652552e-01 -4.87730175e-01 9.85511184e-01
-1.26380548e-01 4.09128398e-01 4.85095203e-01 -7.08635271e-01
4.16636288e-01 -3.32250074e-02 7.74056673e-01 -1.98134422e-01
1.17187805e-01 -3.54887635e-01 -3.71217340e-01 -3.86962086e-01
-3.92819375e-01 -8.45513642e-01 -1.16913557e+00 -9.68456686e-01
-3.12267005e-01 2.11435094e-01 8.88927877e-01 6.14353299e-01
4.74401712e-01 -1.25150293e-01 1.02160883e+00 -4.40677971e-01
-5.96507311e-01 -4.53136325e-01 -1.10367811e+00 3.98742944e-01
4.86377776e-01 -7.59076118e-01 -9.25854385e-01 -1.60244048e-01] | [7.410229206085205, 4.303731441497803] |
69a0441d-dbc0-4f78-8229-da2ae9d5452a | action-matching-a-variational-method-for | 2210.06662 | null | https://arxiv.org/abs/2210.06662v3 | https://arxiv.org/pdf/2210.06662v3.pdf | Action Matching: Learning Stochastic Dynamics from Samples | Learning the continuous dynamics of a system from snapshots of its temporal marginals is a problem which appears throughout natural sciences and machine learning, including in quantum systems, single-cell biological data, and generative modeling. In these settings, we assume access to cross-sectional samples that are uncorrelated over time, rather than full trajectories of samples. In order to better understand the systems under observation, we would like to learn a model of the underlying process that allows us to propagate samples in time and thereby simulate entire individual trajectories. In this work, we propose Action Matching, a method for learning a rich family of dynamics using only independent samples from its time evolution. We derive a tractable training objective, which does not rely on explicit assumptions about the underlying dynamics and does not require back-propagation through differential equations or optimal transport solvers. Inspired by connections with optimal transport, we derive extensions of Action Matching to learn stochastic differential equations and dynamics involving creation and destruction of probability mass. Finally, we showcase applications of Action Matching by achieving competitive performance in a diverse set of experiments from biology, physics, and generative modeling. | ['Daniel Severo', 'Rob Brekelmans', 'Alireza Makhzani', 'Kirill Neklyudov'] | 2022-10-13 | null | null | null | null | ['colorization'] | ['computer-vision'] | [ 2.33474940e-01 -2.67995745e-02 1.30641431e-01 8.04588646e-02
-4.53865319e-01 -7.11283922e-01 8.98202956e-01 9.01700333e-02
-3.56826991e-01 1.07692361e+00 -1.19556129e-01 -2.66888171e-01
-1.95988685e-01 -8.15649390e-01 -9.54783201e-01 -1.20738316e+00
-1.57976776e-01 7.89464772e-01 5.16235270e-02 -7.54635558e-02
5.97938970e-02 7.20870137e-01 -8.51858437e-01 -1.54083461e-01
7.94466615e-01 4.30753201e-01 8.74835253e-02 1.06187856e+00
1.38503417e-01 7.21768200e-01 -1.79518342e-01 -2.92105996e-03
-6.69191033e-02 -9.08499897e-01 -8.11123729e-01 1.15275174e-01
-7.42392987e-03 -8.90599340e-02 -7.53034115e-01 8.86409402e-01
3.12616110e-01 3.65028650e-01 1.27792728e+00 -1.18048513e+00
-6.22138917e-01 4.24333602e-01 -1.80647731e-01 3.29397589e-01
3.97743322e-02 5.95711172e-01 7.51340687e-01 -3.46467584e-01
8.07029605e-01 1.21069646e+00 5.70224047e-01 7.80755818e-01
-2.02975154e+00 -2.58046478e-01 1.06260851e-01 -1.22196250e-01
-1.16632164e+00 -3.12043428e-01 5.00126481e-01 -7.91036010e-01
5.76745212e-01 -4.99783680e-02 7.94888735e-01 1.34477067e+00
6.93141937e-01 6.76873803e-01 1.17224360e+00 -1.66986898e-01
7.42707431e-01 -1.31277680e-01 9.89324227e-02 7.61554241e-01
3.78166363e-02 9.20507535e-02 -3.51878077e-01 -4.97256547e-01
9.35027182e-01 2.51095861e-01 -3.62550706e-01 -4.81394410e-01
-1.21201944e+00 7.09395111e-01 1.54148132e-01 6.06551878e-02
-3.85747313e-01 5.65888762e-01 -1.23680644e-01 1.66504413e-01
4.13781971e-01 3.19762468e-01 -2.98795342e-01 -3.80184688e-02
-7.84860671e-01 6.23466790e-01 1.12939644e+00 7.84309924e-01
9.04766679e-01 -1.26073971e-01 -1.62538722e-01 4.99303378e-02
1.61736682e-01 9.63437021e-01 4.10519168e-02 -1.36604917e+00
-5.70615418e-02 1.09746829e-01 5.76745749e-01 -3.32724839e-01
-3.47189426e-01 -3.46154720e-01 -1.13101530e+00 4.80517466e-03
7.59862006e-01 -3.34866166e-01 -1.15790033e+00 2.11594343e+00
4.47039753e-01 5.53832531e-01 -1.46697238e-02 5.51326096e-01
-2.84883976e-02 9.62133169e-01 -2.04899251e-01 -4.39216167e-01
7.67131984e-01 -4.11976576e-01 -5.68895400e-01 1.85087308e-01
7.21458554e-01 -2.52180874e-01 7.40697503e-01 2.80408174e-01
-1.26786220e+00 -9.61809009e-02 -5.53403497e-01 2.96981875e-02
-2.65783459e-01 -3.93176615e-01 6.46980643e-01 1.41869083e-01
-1.05442917e+00 1.29137361e+00 -1.61924624e+00 -6.10741615e-01
4.45151299e-01 2.76215762e-01 -1.41603455e-01 1.60098985e-01
-7.36483335e-01 6.28307521e-01 -1.80397138e-01 -1.05971113e-01
-1.59206998e+00 -9.08781171e-01 -4.77476776e-01 -6.91785589e-02
1.05828211e-01 -1.33203614e+00 1.20893061e+00 -3.39570284e-01
-1.68430805e+00 2.98977137e-01 -4.48543787e-01 -5.75432003e-01
6.67339921e-01 8.82888064e-02 3.93370092e-02 7.28394538e-02
5.58447391e-02 3.09093505e-01 7.06320405e-01 -8.88402045e-01
-2.31840387e-02 -3.48515183e-01 3.96059863e-02 -1.07572176e-01
1.81756854e-01 -6.44796073e-01 9.60805565e-02 -7.76211470e-02
1.28864944e-01 -1.30588949e+00 -6.05208576e-01 1.07105732e-01
-5.92691183e-01 1.44259542e-01 6.39320433e-01 -2.94864923e-01
6.49508834e-01 -1.72535014e+00 8.68374467e-01 6.47826307e-03
3.46558750e-01 -1.69509679e-01 1.48953542e-01 9.06184018e-01
2.55228907e-01 1.80043533e-01 -6.58738554e-01 -4.53222990e-01
1.12591192e-01 4.30137992e-01 -6.18866742e-01 7.50945210e-01
2.27281481e-01 1.01828098e+00 -1.10672200e+00 -2.65110701e-01
4.17168401e-02 5.89909911e-01 -6.10096157e-01 5.27414680e-02
-5.46609163e-01 1.16269588e+00 -6.64667726e-01 1.44639090e-01
4.77170557e-01 -7.40100265e-01 -2.79986616e-02 1.28577992e-01
-5.71700372e-02 1.43952519e-01 -1.06013668e+00 1.58437991e+00
-5.11573672e-01 3.25675309e-01 5.04701100e-02 -1.18137455e+00
2.05574945e-01 3.50657374e-01 5.00479758e-01 -1.34076357e-01
1.57343566e-01 1.66755468e-01 1.44689694e-01 -3.74059230e-01
8.45678151e-02 -5.82784534e-01 1.98812596e-03 9.09668803e-01
2.85189450e-01 -3.31429571e-01 2.27556616e-01 5.51120222e-01
1.43254781e+00 1.80986553e-01 4.37660404e-02 -3.62342417e-01
2.99509913e-01 -8.06903914e-02 4.61762965e-01 8.98484707e-01
-1.78656504e-02 3.75200093e-01 5.91725826e-01 -2.04105616e-01
-1.39282107e+00 -1.57788336e+00 -2.21839592e-01 2.54530579e-01
1.72462657e-01 -1.55349687e-01 -7.64668941e-01 -2.21985862e-01
1.04678191e-01 4.32154417e-01 -8.58235419e-01 -3.22823048e-01
-5.12571454e-01 -1.14345574e+00 2.38294423e-01 1.90793872e-01
8.26036092e-03 -7.93521166e-01 -2.64125556e-01 4.59577769e-01
-1.30424807e-02 -9.24078107e-01 -3.40253830e-01 1.30222067e-01
-1.09333467e+00 -9.09314930e-01 -8.52737129e-01 -7.96381533e-02
7.89449394e-01 -1.32311672e-01 8.55790138e-01 -3.12060177e-01
-6.35687768e-01 6.05711937e-01 2.48683736e-01 -1.52230129e-01
-6.93561494e-01 2.04084322e-01 2.05522224e-01 9.99638215e-02
-2.83860892e-01 -1.03757274e+00 -6.23889446e-01 9.70527064e-03
-8.83960664e-01 4.74103913e-03 2.91231871e-01 8.36390257e-01
7.55268931e-01 -1.31160825e-01 3.30731273e-01 -8.54273558e-01
3.29190344e-01 -6.86610818e-01 -7.38943636e-01 1.39705434e-01
-2.64696717e-01 5.72712421e-01 8.84393752e-01 -5.83269000e-01
-9.38374817e-01 -5.15921228e-02 8.07681009e-02 -3.59531313e-01
-6.47742674e-02 1.71122730e-01 1.56601757e-01 7.07077980e-02
4.49537128e-01 7.11685061e-01 1.77777529e-01 -4.41196233e-01
4.93198395e-01 4.08195099e-03 2.92844117e-01 -9.27717090e-01
7.26231754e-01 1.21119082e+00 6.81360006e-01 -1.00828576e+00
-7.25704789e-01 -1.52612776e-01 -7.37851918e-01 -8.78670514e-02
8.57563913e-01 -6.52570128e-01 -1.10264850e+00 6.26042724e-01
-1.10553133e+00 -8.48028660e-01 -6.74441695e-01 5.02174973e-01
-1.02486837e+00 4.92099933e-02 -8.82785618e-01 -1.14288163e+00
5.55964410e-02 -8.76331031e-01 1.13913786e+00 3.16890359e-01
-1.10443771e-01 -1.48497593e+00 6.11727655e-01 -3.79478395e-01
5.26119709e-01 3.94261003e-01 8.45407605e-01 -1.27937034e-01
-1.17416358e+00 -1.32770956e-01 3.04739088e-01 1.90602019e-02
1.07729204e-01 2.76945144e-01 -7.17695475e-01 -3.11681569e-01
3.13491076e-02 -1.33669451e-01 1.08100092e+00 7.55294919e-01
1.00318980e+00 -2.73216635e-01 -8.21872413e-01 7.88654923e-01
1.27798188e+00 5.43194115e-02 3.16340774e-01 -4.46900278e-01
7.12395608e-01 5.59453070e-01 -1.09386966e-01 4.84513164e-01
3.60531211e-01 2.26098523e-01 1.41233802e-01 3.07823390e-01
1.50869995e-01 -4.18999195e-01 3.76546472e-01 7.96316504e-01
-1.78775325e-01 -4.24994022e-01 -6.69977069e-01 4.72738057e-01
-1.81726050e+00 -1.42177510e+00 -2.03961223e-01 2.23937249e+00
7.85952628e-01 -1.29308000e-01 2.52082139e-01 -3.26677501e-01
5.43534338e-01 -1.11956388e-01 -1.17616856e+00 9.52641070e-02
-1.81041941e-01 2.88621932e-01 5.53223073e-01 9.17697608e-01
-7.73387134e-01 7.41588593e-01 6.85941935e+00 4.17587131e-01
-1.11161113e+00 3.55916888e-01 6.66524291e-01 -1.79053277e-01
-5.00537097e-01 4.03257847e-01 -9.36630785e-01 5.32991529e-01
1.21875727e+00 -4.53979343e-01 6.39182925e-01 2.02916130e-01
4.20170218e-01 -1.19783938e-01 -1.31158340e+00 4.73079264e-01
-6.21269286e-01 -1.56252301e+00 -5.91641590e-02 4.89268839e-01
9.83564854e-01 1.60996631e-01 7.27162138e-02 4.61334586e-02
8.87382925e-01 -8.70786726e-01 5.80359638e-01 1.12034833e+00
2.13501915e-01 -3.36961567e-01 1.43156901e-01 7.51222789e-01
-8.30649853e-01 2.21525550e-01 -4.26613331e-01 -1.82468653e-01
7.50767052e-01 9.21011329e-01 -5.10318279e-01 2.73857445e-01
1.73312142e-01 9.37847018e-01 1.71427846e-01 9.43197370e-01
1.24240890e-01 9.46539819e-01 -7.03862965e-01 -1.07461914e-01
-5.99076226e-02 -6.98443055e-01 7.34016716e-01 1.04790092e+00
5.25355518e-01 1.47841752e-01 -2.97161508e-02 1.47081876e+00
-1.04801180e-02 -4.43983644e-01 -7.73876369e-01 -4.65342164e-01
2.21015722e-01 1.19281840e+00 -8.51658642e-01 -2.66687542e-01
-5.98413907e-02 1.06466508e+00 3.33503038e-01 7.64712691e-01
-9.75979328e-01 -2.30831318e-02 8.05938542e-01 3.21223646e-01
3.15581799e-01 -6.28152668e-01 1.67509705e-01 -1.38575685e+00
-2.34614894e-01 -1.10422604e-01 -6.75661936e-02 -4.32239145e-01
-1.45191610e+00 2.36133233e-01 1.38778042e-03 -8.24180901e-01
-3.21423262e-01 -4.80128407e-01 -8.33106875e-01 1.07767546e+00
-1.15381539e+00 -7.77332902e-01 9.92354676e-02 4.77296859e-01
-6.52045384e-02 3.76825422e-01 5.66077292e-01 -1.05799690e-01
-7.30730534e-01 -1.67344838e-01 6.03176713e-01 -2.74484336e-01
3.15163761e-01 -1.50319302e+00 6.93639994e-01 7.21709609e-01
1.36911711e-02 7.60607362e-01 1.00914240e+00 -6.90379739e-01
-1.64568210e+00 -1.06908906e+00 5.52403450e-01 -7.19596326e-01
1.04104507e+00 -6.24425530e-01 -9.75147367e-01 7.52923608e-01
-5.56090511e-02 3.54249805e-01 4.24773395e-01 -2.21636042e-01
7.73743168e-02 9.54322591e-02 -8.61771166e-01 5.93196392e-01
1.27274597e+00 -6.62787259e-01 6.35854453e-02 5.95985830e-01
4.93800133e-01 -2.59597600e-01 -7.68723369e-01 -8.91456828e-02
4.72145468e-01 -5.91446817e-01 9.48134005e-01 -1.00678289e+00
3.00413430e-01 -3.24325919e-01 -7.45013729e-02 -1.43681610e+00
-2.01134011e-01 -1.01864302e+00 -4.49886411e-01 9.51563179e-01
5.01465738e-01 -7.89628386e-01 7.28894651e-01 5.11101663e-01
1.58535428e-02 -7.26423919e-01 -7.76282310e-01 -1.01741946e+00
5.48965573e-01 -1.90908253e-01 2.78917968e-01 6.16445363e-01
-2.46423483e-01 3.09531271e-01 -2.86062062e-01 3.97206783e-01
1.12145162e+00 2.20670193e-01 7.02803016e-01 -1.09822679e+00
-7.16357291e-01 -2.36529052e-01 -2.43329883e-01 -1.28153002e+00
2.82739639e-01 -8.84286225e-01 1.26249030e-01 -1.59597743e+00
4.14968103e-01 -2.77067453e-01 5.57050891e-02 -2.08633453e-01
-5.75883016e-02 -9.74540189e-02 -6.95336312e-02 4.25717056e-01
-4.34418976e-01 9.79981303e-01 1.49066329e+00 2.48143934e-02
-6.75378293e-02 2.38201603e-01 -2.12435365e-01 5.45378745e-01
5.47965169e-01 -6.83413029e-01 -5.41724443e-01 -1.44499287e-01
2.69788653e-01 3.38491440e-01 8.60001743e-01 -9.56546605e-01
3.89439195e-01 -4.37691867e-01 4.34516072e-02 -3.63211125e-01
5.74813247e-01 -1.70342565e-01 4.59384888e-01 7.23728061e-01
-3.79025728e-01 -1.00642636e-01 -2.06295773e-01 9.64571536e-01
2.49504626e-01 -4.75455709e-02 9.56358552e-01 -3.99653494e-01
1.04844272e-01 7.93951154e-01 -7.19878554e-01 4.58647221e-01
8.50518823e-01 2.76769757e-01 -3.87451231e-01 -6.26329720e-01
-1.25548244e+00 1.50030002e-01 8.11051428e-01 -4.24004644e-01
1.52580500e-01 -1.10422969e+00 -4.73027080e-01 -1.24830954e-01
-2.97594100e-01 1.35416063e-02 4.88435298e-01 1.19651628e+00
-4.30685133e-01 3.75718206e-01 5.12463041e-02 -6.92933261e-01
-3.94268811e-01 4.76355821e-01 5.95974386e-01 -2.02698618e-01
-5.08943737e-01 6.56655908e-01 4.93041515e-01 -4.23265278e-01
-3.84208530e-01 -4.33459669e-01 4.66424644e-01 -3.43890488e-01
4.36585665e-01 5.01755953e-01 -3.54876071e-01 -3.98955822e-01
-1.07330739e-01 5.45683026e-01 -4.39911969e-02 -5.62034905e-01
1.23377991e+00 -3.10269296e-01 -3.21543962e-02 1.15671194e+00
1.11133289e+00 -2.13101715e-01 -1.73195684e+00 -3.09881181e-01
-2.75938839e-01 5.08079566e-02 -1.82423562e-01 -3.89888793e-01
-7.63646483e-01 1.18338287e+00 9.73924398e-02 5.31029522e-01
5.57068765e-01 3.39930803e-01 8.45635712e-01 4.68882740e-01
5.23851991e-01 -7.09962308e-01 -1.03770420e-01 5.19447386e-01
3.54093760e-01 -9.67169821e-01 -2.25733012e-01 -1.49136335e-01
-1.50998286e-03 1.05274796e+00 2.66563762e-02 -5.01888335e-01
1.02514482e+00 3.89064848e-01 -5.11209071e-01 -1.03275403e-01
-1.15430176e+00 -2.29108587e-01 -7.18689114e-02 4.28782225e-01
2.84719527e-01 8.10784027e-02 4.33039898e-03 4.25258316e-02
1.37211278e-01 1.22653119e-01 8.52434516e-01 9.20478106e-01
-3.68474305e-01 -1.16523898e+00 1.24854743e-01 4.12649184e-01
-2.67415613e-01 4.34994102e-02 -2.14878082e-01 6.56327844e-01
-1.34715408e-01 3.30091447e-01 1.81608289e-01 1.39967129e-01
-8.51820186e-02 1.81499302e-01 8.56633842e-01 -5.88243186e-01
2.41902396e-01 -1.81458425e-02 -5.29666781e-01 -5.00281274e-01
-4.39663351e-01 -1.27665913e+00 -1.32763743e+00 -6.69691265e-01
-2.00187966e-01 2.38749787e-01 5.07165551e-01 1.29338193e+00
4.63598669e-01 6.24231875e-01 4.48588073e-01 -9.20789957e-01
-7.46471584e-01 -8.42828929e-01 -8.26096714e-01 1.73453316e-01
7.95051694e-01 -5.25187075e-01 -6.54692054e-01 3.24027956e-01] | [6.668600082397461, 3.9641172885894775] |
b8186b87-f9d7-4971-961c-98bf128c665a | ensemble-of-jointly-trained-deep-neural | 1608.04983 | null | http://arxiv.org/abs/1608.04983v1 | http://arxiv.org/pdf/1608.04983v1.pdf | Ensemble of Jointly Trained Deep Neural Network-Based Acoustic Models for Reverberant Speech Recognition | Distant speech recognition is a challenge, particularly due to the corruption
of speech signals by reverberation caused by large distances between the
speaker and microphone. In order to cope with a wide range of reverberations in
real-world situations, we present novel approaches for acoustic modeling
including an ensemble of deep neural networks (DNNs) and an ensemble of jointly
trained DNNs. First, multiple DNNs are established, each of which corresponds
to a different reverberation time 60 (RT60) in a setup step. Also, each model
in the ensemble of DNN acoustic models is further jointly trained, including
both feature mapping and acoustic modeling, where the feature mapping is
designed for the dereverberation as a front-end. In a testing phase, the two
most likely DNNs are chosen from the DNN ensemble using maximum a posteriori
(MAP) probabilities, computed in an online fashion by using maximum likelihood
(ML)-based blind RT60 estimation and then the posterior probability outputs
from two DNNs are combined using the ML-based weights as a simple average.
Extensive experiments demonstrate that the proposed approach leads to
substantial improvements in speech recognition accuracy over the conventional
DNN baseline systems under diverse reverberant conditions. | ['Joon-Hyuk Chang', 'Jeehye Lee', 'Myungin Lee'] | 2016-08-17 | null | null | null | null | ['distant-speech-recognition'] | ['speech'] | [ 1.33966342e-01 -4.24979150e-01 5.94437063e-01 -4.10970718e-01
-1.20587206e+00 -4.78543639e-01 4.22265410e-01 -2.90540367e-01
-4.74680662e-01 5.65979064e-01 5.18711507e-01 -3.82445157e-01
1.06309410e-02 -3.83013815e-01 -6.51747704e-01 -1.12334573e+00
1.80570036e-01 3.52376163e-01 4.95225638e-02 1.49843454e-01
-2.51684487e-01 6.63166285e-01 -1.52422881e+00 -5.89222822e-04
7.29238212e-01 1.23824632e+00 7.11712658e-01 1.02161264e+00
5.32444529e-02 5.38349569e-01 -1.15429533e+00 -4.68165874e-02
1.22927606e-01 -3.86130363e-01 1.04737515e-02 -1.74106896e-01
3.77688885e-01 -4.52683926e-01 -4.31097656e-01 1.03380859e+00
1.15605962e+00 7.48788536e-01 4.91334707e-01 -5.37095606e-01
-1.75379917e-01 7.10458159e-01 6.46494627e-02 3.16604346e-01
7.23124444e-02 3.39622535e-02 7.99601853e-01 -1.25921512e+00
-2.94098586e-01 1.38339591e+00 4.57593709e-01 6.11591578e-01
-9.24137652e-01 -8.11397552e-01 1.12252131e-01 4.44697410e-01
-1.24615204e+00 -1.06749916e+00 7.87463963e-01 -1.76936209e-01
1.03424668e+00 2.63600230e-01 2.61511683e-01 1.44058609e+00
-1.62401974e-01 5.14571249e-01 7.05132067e-01 -5.76215625e-01
4.59797710e-01 1.73796285e-02 3.96510251e-02 -5.49001507e-02
-2.41599903e-01 3.35553020e-01 -6.30645156e-01 -2.57291377e-01
4.79675174e-01 -1.80992037e-01 -5.80100119e-01 3.61323714e-01
-1.07145274e+00 3.82786840e-01 3.09191883e-01 4.37648892e-01
-6.93742454e-01 7.48447180e-02 9.96819958e-02 1.70644641e-01
3.70804131e-01 2.64364053e-02 -4.70879376e-01 -1.06610216e-01
-8.44413996e-01 -1.73920900e-01 1.02573061e+00 4.23859417e-01
4.46434319e-01 6.00302696e-01 -1.71277538e-01 1.56281888e+00
6.66697562e-01 7.28094339e-01 5.62044799e-01 -5.49897969e-01
7.10898697e-01 -4.83200073e-01 2.36199796e-01 -6.41241848e-01
-6.26564622e-02 -1.10501480e+00 -8.98831308e-01 2.13784009e-01
2.50408441e-01 -4.90721107e-01 -1.06308448e+00 1.82632160e+00
4.52034324e-01 6.00400209e-01 3.70479524e-01 8.76176476e-01
5.89264452e-01 1.18618882e+00 -4.52429764e-02 -3.38807017e-01
1.02981198e+00 -1.09732795e+00 -8.68487418e-01 -4.38881665e-01
9.76727232e-02 -1.20009434e+00 4.68640625e-01 7.47595966e-01
-8.79832625e-01 -9.34711337e-01 -1.08191597e+00 4.12762880e-01
-8.81088823e-02 1.76175460e-01 -3.46847057e-01 8.24474096e-01
-1.14443564e+00 3.74163538e-01 -6.99716330e-01 9.44219604e-02
5.07438891e-02 1.65398359e-01 1.75104721e-03 -1.90196872e-01
-1.29957616e+00 9.58268762e-01 1.31370321e-01 7.72350848e-01
-1.49611139e+00 -4.46472675e-01 -6.88748777e-01 1.62639201e-01
-4.98805977e-02 -4.85887170e-01 1.50723851e+00 -8.86755645e-01
-1.83938289e+00 1.37642562e-01 -4.84536320e-01 -5.28666139e-01
4.12125766e-01 -4.16774213e-01 -9.96653140e-01 -1.55045867e-01
-4.84033883e-01 5.23223951e-02 1.05736101e+00 -1.39967287e+00
-6.76305234e-01 -1.67827994e-01 -3.01314682e-01 5.28326988e-01
-3.55285764e-01 2.60979027e-01 -1.98690295e-01 -7.90046990e-01
2.19728693e-01 -5.54106414e-01 -2.74415046e-01 -2.81830043e-01
-4.53794718e-01 -1.69866025e-01 5.81752598e-01 -1.27205908e+00
1.19571459e+00 -2.42821145e+00 -1.58482678e-02 2.49666110e-01
-1.56489655e-01 7.85791934e-01 -2.66895354e-01 2.86025107e-01
-1.81294959e-02 -4.10922736e-01 8.79122317e-03 -8.43139410e-01
1.21500954e-01 1.13872059e-01 -5.07101834e-01 3.15244675e-01
-2.14975655e-01 7.41591156e-02 -7.30872333e-01 -8.73754248e-02
6.09105706e-01 9.66390789e-01 2.23824177e-02 6.54709935e-01
6.47092611e-02 5.09633720e-01 -5.82563244e-02 2.43555441e-01
8.14700603e-01 5.71318209e-01 1.75764058e-02 -3.43759656e-02
3.37405391e-02 7.07299709e-01 -1.28241587e+00 1.26253736e+00
-9.05631781e-01 6.19179964e-01 4.36629325e-01 -8.90082061e-01
1.19177151e+00 8.23373318e-01 -2.20677525e-01 -3.61008197e-01
2.21971683e-02 5.90507925e-01 2.19689727e-01 -1.67020768e-01
5.02266921e-02 -2.56996185e-01 4.27654654e-01 3.07550639e-01
2.95272738e-01 1.71653971e-01 -3.94090980e-01 -4.87928987e-01
9.24705565e-01 -3.68235201e-01 1.95233062e-01 2.95985460e-01
6.29678428e-01 -9.74333107e-01 6.23436928e-01 1.00056148e+00
-2.53957748e-01 7.45009959e-01 -2.91638732e-01 -1.82406962e-01
-7.15795279e-01 -1.38601565e+00 -1.41977400e-01 9.07648802e-01
-2.98677385e-01 1.34011865e-01 -7.53367007e-01 -3.46958011e-01
-3.31288815e-01 1.26204181e+00 -2.73769479e-02 -2.45742083e-01
-7.42987275e-01 -6.55891061e-01 4.06815559e-01 4.34565872e-01
2.71623641e-01 -1.17130399e+00 1.21298753e-01 6.19021058e-01
-1.85713351e-01 -1.11252594e+00 -4.07252669e-01 5.98038554e-01
-6.37381494e-01 -1.70445606e-01 -9.51468825e-01 -9.60578382e-01
4.14529115e-01 3.66190344e-01 7.44530499e-01 -3.57867450e-01
2.55346805e-01 -1.20539226e-01 -2.28940547e-01 -3.49605083e-01
-8.32141876e-01 -3.74592990e-01 4.95316297e-01 4.69938248e-01
2.51648456e-01 -9.43334520e-01 -5.61327875e-01 4.76764441e-01
-5.23452818e-01 -4.14583504e-01 5.71526408e-01 7.64885128e-01
5.07468522e-01 4.30843562e-01 7.70058155e-01 3.70781980e-02
5.80835342e-01 -4.61948484e-01 -6.15431547e-01 2.26928592e-01
7.81423971e-02 -3.82221818e-01 8.92017484e-01 -7.82041848e-01
-1.37501264e+00 -2.05119327e-01 -7.37437606e-01 -4.92798239e-01
-6.21930420e-01 3.39820653e-01 -6.96093142e-01 2.92124152e-01
6.39819384e-01 4.61817890e-01 -3.15173537e-01 -8.55327666e-01
1.66970715e-01 1.27972496e+00 4.96407539e-01 -2.02112660e-01
7.47474194e-01 -1.73918277e-01 -6.65683448e-01 -1.03237522e+00
-7.74753451e-01 -3.14591140e-01 -3.02078754e-01 -3.41986090e-01
6.64474666e-01 -1.03735614e+00 -3.01914424e-01 9.86548483e-01
-1.50076008e+00 -2.15783119e-01 1.07984664e-02 1.13592649e+00
-9.68404859e-02 1.52112409e-01 -5.54982483e-01 -1.34209728e+00
-3.64266187e-01 -1.18370342e+00 5.02132118e-01 3.75850320e-01
-8.30251351e-02 -8.77437413e-01 6.44611940e-02 3.95697325e-01
5.69752395e-01 -3.75357747e-01 6.84787631e-01 -1.11845458e+00
-3.68622988e-01 -4.41385001e-01 2.47636378e-01 1.36164308e+00
3.58864069e-01 -2.95448780e-01 -1.59479606e+00 -2.57395238e-01
5.30156493e-01 3.99452336e-02 6.52911663e-01 8.02640855e-01
8.97751629e-01 -2.34627739e-01 -1.81470394e-01 2.72808224e-01
1.16811657e+00 8.16037178e-01 4.53033328e-01 -1.60905704e-01
6.68581963e-01 2.73870200e-01 1.87244922e-01 2.96428084e-01
-5.21329828e-02 5.79811573e-01 3.22580606e-01 -8.78788009e-02
-4.63923991e-01 -4.28041182e-02 7.11295068e-01 1.43889487e+00
3.44157070e-01 -7.62774408e-01 -7.91245937e-01 6.00543439e-01
-1.40109968e+00 -7.16428518e-01 3.13234568e-01 2.70375848e+00
7.28059232e-01 1.41833767e-01 -2.79321462e-01 2.50239670e-01
1.12541044e+00 3.37166965e-01 -6.02576673e-01 -3.05993944e-01
-1.90595359e-01 3.31958830e-01 1.03108445e-02 8.43414605e-01
-7.34599113e-01 5.66611767e-01 5.99092484e+00 1.16630244e+00
-1.05788898e+00 3.33111525e-01 6.10074043e-01 -1.89397737e-01
-4.43562455e-02 -2.04046309e-01 -8.82993639e-01 4.03574526e-01
1.28229749e+00 2.09276408e-01 6.96824789e-01 6.65647924e-01
5.67731202e-01 9.95338932e-02 -1.02089608e+00 9.97822523e-01
1.85586452e-01 -6.27704740e-01 -1.31195158e-01 -1.10816471e-01
5.20420074e-01 3.95527482e-01 1.63409039e-01 1.63660884e-01
3.92599881e-01 -8.72826338e-01 9.17163789e-01 6.44426167e-01
2.97954321e-01 -9.52663600e-01 6.81170762e-01 5.49382687e-01
-9.56511378e-01 -2.82182395e-01 -4.75866467e-01 1.44308507e-01
4.10582751e-01 1.23899245e+00 -1.14719236e+00 4.41497535e-01
6.14293337e-01 8.77552703e-02 3.00352514e-01 1.38268304e+00
-5.20933747e-01 1.12107420e+00 -5.07296324e-01 -6.80597052e-02
-1.18681928e-02 -5.03206439e-02 9.76322949e-01 1.17412210e+00
6.67792678e-01 -2.08017185e-01 -1.74998417e-01 6.51114821e-01
-1.08889818e-01 -1.29434988e-01 -1.63690522e-01 2.27531582e-01
8.99121404e-01 1.13318849e+00 -2.88401276e-01 -3.21240842e-01
-1.14662856e-01 9.08358514e-01 1.21077612e-01 9.16039944e-01
-7.63996840e-01 -5.48559844e-01 8.24803948e-01 -5.93550384e-01
6.03128016e-01 -3.19583118e-01 1.41206414e-01 -8.26366544e-01
2.40542814e-01 -8.83434951e-01 -6.41743094e-02 -9.19493616e-01
-1.36014187e+00 1.14360023e+00 -4.56020236e-01 -1.26655114e+00
-2.30028182e-01 -6.12996578e-01 -7.22687364e-01 1.60890663e+00
-1.50743353e+00 -5.30546367e-01 1.18215889e-01 2.86346674e-01
7.68054128e-01 -2.17376605e-01 1.02508295e+00 5.86120725e-01
-7.11903214e-01 5.55154920e-01 5.94813645e-01 2.89911106e-02
6.80232167e-01 -1.07603860e+00 5.50797462e-01 1.10959613e+00
2.84527242e-01 4.88950461e-01 8.00733149e-01 -3.23665977e-01
-8.71414006e-01 -1.08635664e+00 1.15565526e+00 -1.55841589e-01
3.93943936e-01 -7.01852560e-01 -9.69725966e-01 3.66011649e-01
1.24840096e-01 -8.52095429e-03 8.05290282e-01 -6.58351630e-02
-2.00095639e-01 -6.01862550e-01 -8.02200735e-01 5.24958730e-01
7.01928198e-01 -6.49479151e-01 -6.65745676e-01 3.17942530e-01
8.30380082e-01 -3.93584371e-01 -4.84957427e-01 8.31461400e-02
3.49162430e-01 -8.12774956e-01 9.94704843e-01 -1.27833523e-02
-1.56799749e-01 -4.91659045e-01 -6.15820944e-01 -1.76728106e+00
-8.20510387e-02 -7.39201903e-01 -2.40324527e-01 1.61192477e+00
6.15186930e-01 -8.65080714e-01 1.53599858e-01 1.43807799e-01
-4.15605068e-01 -3.01780969e-01 -1.32489741e+00 -8.57565999e-01
-2.40511999e-01 -8.70155692e-01 6.29514873e-01 3.31793576e-01
-6.13362849e-01 2.39856616e-01 -5.01228213e-01 7.96370685e-01
5.95360696e-01 -3.95181090e-01 3.66941065e-01 -9.58323181e-01
-5.27640462e-01 -2.74981499e-01 -8.60914513e-02 -1.64699256e+00
7.40364790e-02 -4.66509432e-01 7.07866549e-01 -1.58782089e+00
-4.76426095e-01 -3.57294053e-01 -7.59777069e-01 5.46321087e-03
-2.13975653e-01 -2.68718421e-01 9.73100541e-04 -8.44095945e-02
-1.13383658e-01 7.50941336e-01 7.86142766e-01 3.98792028e-02
-1.68960378e-01 7.12246120e-01 -4.13682103e-01 7.21387327e-01
5.19978344e-01 -6.68185949e-01 -4.42029864e-01 -6.76744461e-01
-4.51278567e-01 3.84248585e-01 2.11122394e-01 -1.41854382e+00
3.65110219e-01 3.11157614e-01 4.67901766e-01 -7.00699806e-01
7.76149988e-01 -9.08054054e-01 2.61125177e-01 1.64389044e-01
-2.75624692e-01 -4.90701586e-01 6.02067113e-02 6.77153647e-01
-3.69440019e-01 -3.93831402e-01 7.29414761e-01 6.80968910e-02
-1.39429986e-01 1.04216285e-01 -7.02030480e-01 -5.87331712e-01
4.48747963e-01 -7.65927508e-02 7.23843277e-02 -6.03236079e-01
-9.40820932e-01 -1.46866068e-01 -4.46809947e-01 4.49488848e-01
6.25769854e-01 -1.30638981e+00 -7.07655787e-01 1.86945572e-01
-3.22690189e-01 8.67509916e-02 6.15467727e-01 5.54474831e-01
-1.66659453e-03 1.85185239e-01 3.77371579e-01 -5.59768617e-01
-1.13391054e+00 1.82725310e-01 8.11855257e-01 -5.87684847e-02
-2.73999780e-01 1.31321061e+00 3.87157559e-01 -5.95238626e-01
8.67663801e-01 -2.74572521e-01 -2.48869970e-01 -1.42327189e-01
7.58967042e-01 4.11281675e-01 4.85926539e-01 -7.41271079e-01
-3.36298555e-01 3.47461879e-01 5.82402498e-02 -6.53150499e-01
1.22634518e+00 -4.16162878e-01 1.66924194e-01 6.61734879e-01
1.26416540e+00 4.82089855e-02 -1.26985753e+00 -5.65037966e-01
-4.36466962e-01 -2.57392913e-01 4.12496150e-01 -1.15549767e+00
-9.06337082e-01 1.03835022e+00 7.56990731e-01 4.57521640e-02
1.30702436e+00 -2.78801978e-01 8.44319701e-01 2.65689284e-01
4.76464406e-02 -1.01322865e+00 -1.45223349e-01 5.29961944e-01
9.32661653e-01 -7.98685372e-01 -7.25838363e-01 1.77771434e-01
-3.10904175e-01 1.20466089e+00 3.75469923e-01 1.56904340e-01
9.24608588e-01 3.01229745e-01 5.78437269e-01 5.29724538e-01
-5.94656825e-01 7.58926645e-02 2.27259159e-01 6.85468137e-01
4.89586554e-02 1.42146245e-01 2.95567602e-01 7.25090206e-01
-4.08635408e-01 -4.37002122e-01 -3.18043642e-02 5.77575684e-01
-5.56602120e-01 -9.62665975e-01 -8.66261721e-01 1.88506424e-01
-3.18126470e-01 -3.90982568e-01 8.22872967e-02 -1.58473611e-01
-2.21062917e-02 1.50836456e+00 1.18927747e-01 -5.11427939e-01
4.28821802e-01 1.97097182e-01 1.22982942e-01 -4.68143970e-01
-5.21640658e-01 5.78478873e-01 2.64070481e-01 -8.58092234e-02
-2.12905649e-03 -5.52376926e-01 -1.01788402e+00 1.19618692e-01
-6.98628306e-01 1.84843630e-01 1.06865454e+00 1.18593371e+00
1.11008406e-01 7.62873054e-01 9.98013616e-01 -1.09640276e+00
-8.65632713e-01 -1.33320129e+00 -6.94289505e-01 -7.22968653e-02
9.01566803e-01 -2.58450747e-01 -1.03936779e+00 -3.66896577e-02] | [15.01656436920166, 5.894041061401367] |
f85b0cb6-4cce-4d5b-adce-bdc917805573 | joint-3d-instance-segmentation-and-object | null | null | http://openaccess.thecvf.com/content_CVPR_2020/html/Zhou_Joint_3D_Instance_Segmentation_and_Object_Detection_for_Autonomous_Driving_CVPR_2020_paper.html | http://openaccess.thecvf.com/content_CVPR_2020/papers/Zhou_Joint_3D_Instance_Segmentation_and_Object_Detection_for_Autonomous_Driving_CVPR_2020_paper.pdf | Joint 3D Instance Segmentation and Object Detection for Autonomous Driving | Currently, in Autonomous Driving (AD), most of the 3D object detection frameworks (either anchor- or anchor-free-based) consider the detection as a Bounding Box (BBox) regression problem. However, this compact representation is not sufficient to explore all the information of the objects. To tackle this problem, we propose a simple but practical detection framework to jointly predict the 3D BBox and instance segmentation. For instance segmentation, we propose a Spatial Embeddings (SEs) strategy to assemble all foreground points into their corresponding object centers. Base on the SE results, the object proposals can be generated based on a simple clustering strategy. For each cluster, only one proposal is generated. Therefore, the Non-Maximum Suppression (NMS) process is no longer needed here. Finally, with our proposed instance-aware ROI pooling, the BBox is refined by a second-stage network. Experimental results on the public KITTI dataset show that the proposed SEs can significantly improve the instance segmentation results compared with other feature embedding-based method. Meanwhile, it also outperforms most of the 3D object detectors on the KITTI testing benchmark.
| [' Ruigang Yang', ' Hongdong Li', ' Yuchao Dai', ' Junbo Yin', ' Liu Liu', ' Xibin Song', ' Jin Fang', 'Dingfu Zhou'] | 2020-06-01 | null | null | null | cvpr-2020-6 | ['3d-instance-segmentation-1'] | ['computer-vision'] | [-6.43049553e-02 2.16774479e-01 -2.22970769e-01 -4.18397516e-01
-4.83051509e-01 -1.86222985e-01 5.52738547e-01 7.99049139e-02
-4.91508693e-01 2.68884271e-01 -4.03905958e-01 -2.88683504e-01
6.54555932e-02 -6.92704320e-01 -7.11948931e-01 -8.89861107e-01
2.73439586e-01 3.68773937e-01 1.08395183e+00 3.05117927e-02
2.93822408e-01 5.29433787e-01 -1.66407239e+00 -1.80887818e-01
9.20865655e-01 1.45868385e+00 5.18007338e-01 8.55332017e-02
-4.45923150e-01 2.94932663e-01 -5.21441162e-01 -4.94844615e-02
4.58093166e-01 3.91234681e-02 7.70354469e-04 2.03801647e-01
2.01042801e-01 -3.81562084e-01 -1.80961698e-01 1.02017057e+00
3.63410592e-01 1.77272335e-01 6.47130251e-01 -1.38810861e+00
-1.51026964e-01 3.24713916e-01 -1.06917882e+00 2.05141738e-01
-2.41709471e-01 8.35312977e-02 8.36274266e-01 -1.41601610e+00
4.07603830e-01 1.28641427e+00 4.23485875e-01 3.44755143e-01
-8.35671902e-01 -8.12439501e-01 6.49658084e-01 4.13788289e-01
-1.42581654e+00 -2.50407338e-01 1.04527199e+00 -2.14762136e-01
5.18675327e-01 2.03056723e-01 7.44750321e-01 6.12433851e-01
1.34011952e-03 1.01952481e+00 8.70262921e-01 -1.15853377e-01
2.15795875e-01 4.10679132e-01 2.77799308e-01 7.68457174e-01
5.45221269e-01 -1.40669852e-01 -2.65694439e-01 1.70485914e-01
5.64184368e-01 9.80116948e-02 -2.24098880e-02 -7.76760757e-01
-1.23927593e+00 7.86027133e-01 7.98133314e-01 8.09354987e-03
-1.21253259e-01 4.34540734e-02 2.76298702e-01 -4.09767032e-01
6.10236049e-01 1.09009750e-01 -1.96830451e-01 3.99489760e-01
-9.61864233e-01 3.40257734e-01 3.47353011e-01 1.07713532e+00
1.06480610e+00 -2.30300114e-01 -3.50990623e-01 7.27107525e-01
5.64235628e-01 1.81685135e-01 1.90034047e-01 -5.44735432e-01
6.60902858e-01 1.18604529e+00 2.64556587e-01 -1.21370876e+00
-5.47398150e-01 -6.48272157e-01 -5.98538339e-01 3.23788881e-01
3.08705449e-01 5.36405109e-02 -1.15768409e+00 1.15930498e+00
9.69766200e-01 5.99901795e-01 -1.87310874e-01 1.23021948e+00
1.02678096e+00 7.73514211e-01 -1.75636053e-01 4.58759926e-02
1.37147284e+00 -1.17325485e+00 -5.43336987e-01 -5.48842371e-01
5.13910234e-01 -5.03759384e-01 6.68719530e-01 6.82470649e-02
-6.26946807e-01 -6.83456779e-01 -1.40901446e+00 -3.10493350e-01
-3.11566800e-01 5.06484449e-01 5.44591784e-01 6.05515480e-01
-6.19109869e-01 -6.25651479e-02 -9.59537089e-01 -6.98426589e-02
7.86461949e-01 1.16163887e-01 -1.66143134e-01 -2.18086809e-01
-6.91829324e-01 7.07292199e-01 7.15589762e-01 4.81752336e-01
-8.66775095e-01 -5.28465509e-01 -8.99573624e-01 -1.00571834e-01
7.78587341e-01 -2.95846820e-01 6.34841442e-01 -3.70651364e-01
-1.19888723e+00 7.16090620e-01 -2.65402168e-01 -4.55210775e-01
6.03086770e-01 -5.05662002e-02 -4.94200289e-02 1.63753644e-01
2.70716846e-01 1.06165659e+00 9.25951481e-01 -1.54727256e+00
-1.18199003e+00 -5.06523788e-01 4.41480242e-02 1.64764777e-01
-2.83997238e-01 -2.36609980e-01 -9.08461809e-01 -2.27033302e-01
7.67279625e-01 -9.13065314e-01 -4.88879502e-01 2.56571680e-01
-7.13071346e-01 -6.86473131e-01 1.01821065e+00 -4.18776602e-01
1.14814734e+00 -2.41495180e+00 7.71786273e-02 2.15686843e-01
4.68434274e-01 1.83830768e-01 1.28443778e-01 -2.55216956e-01
3.16455245e-01 9.90126282e-02 -3.61627936e-01 -6.09984457e-01
1.75960094e-01 1.10224105e-01 -2.72258997e-01 6.12019837e-01
6.56619668e-01 8.09143543e-01 -6.65533125e-01 -9.19017911e-01
5.19600809e-01 3.05634707e-01 -6.12011969e-01 1.18367963e-01
-1.96288615e-01 3.03010315e-01 -8.70842934e-01 7.97424376e-01
1.11257541e+00 3.03183740e-04 -2.54219085e-01 -2.25929931e-01
-3.50202769e-01 -2.43844260e-02 -1.41532314e+00 1.54614258e+00
-1.29099801e-01 4.36928868e-01 6.21533878e-02 -1.15995502e+00
1.28166628e+00 -3.04168046e-01 4.48904335e-01 -3.81326228e-01
8.22184682e-02 3.44631314e-01 -2.48087710e-03 -3.86086047e-01
4.37192947e-01 3.05660874e-01 -1.46360517e-01 -7.15895649e-03
-2.64358997e-01 7.18833134e-02 1.44285396e-01 -9.06593818e-03
9.26266491e-01 2.93164760e-01 -6.07424043e-02 -1.20431885e-01
6.43703640e-01 2.17859983e-01 1.04094648e+00 7.02548742e-01
-5.18986225e-01 7.23655045e-01 5.97812593e-01 -3.30149442e-01
-8.05697918e-01 -8.72040749e-01 -3.61459076e-01 6.96449518e-01
9.81702626e-01 -2.07476616e-01 -8.60418022e-01 -8.31587970e-01
1.64849162e-01 6.02086782e-01 -6.93116248e-01 -1.82209179e-01
-7.57187843e-01 -7.87616789e-01 1.05415411e-01 5.40540576e-01
6.50069654e-01 -7.94144332e-01 -8.91928554e-01 3.49286973e-01
-2.18266957e-02 -1.23662066e+00 -3.72437656e-01 2.41230711e-01
-8.58733594e-01 -8.93405318e-01 -6.65524244e-01 -7.83065379e-01
7.88292110e-01 7.29288995e-01 4.13325995e-01 1.48567230e-01
-3.48660111e-01 -1.50882974e-01 -4.44532365e-01 -4.48895931e-01
2.21373029e-02 1.40848488e-01 -1.68402314e-01 3.38422328e-01
5.57822883e-01 -1.82686716e-01 -9.15608108e-01 5.73107958e-01
-6.01499319e-01 1.57431573e-01 8.96431386e-01 5.40556550e-01
8.50830317e-01 3.62041630e-02 6.19774938e-01 -4.26957279e-01
6.05434477e-02 -4.77884471e-01 -7.85368979e-01 3.52354944e-02
-3.79534006e-01 -2.54681945e-01 2.28559554e-01 -4.57092613e-01
-9.12514150e-01 3.02932739e-01 4.40506265e-02 -6.77606165e-01
-2.28474736e-01 2.43113801e-01 -4.19334888e-01 -4.06955509e-03
2.80104429e-01 3.40329885e-01 -1.14996880e-01 -5.85461140e-01
2.97368228e-01 7.85073638e-01 2.44099602e-01 -1.96629241e-01
1.00318098e+00 6.83650076e-01 2.63219737e-02 -7.86819875e-01
-7.70090044e-01 -6.89040899e-01 -6.71249151e-01 -3.33523184e-01
1.04082835e+00 -9.36615348e-01 -4.82895911e-01 1.67482391e-01
-1.19940507e+00 7.54831061e-02 1.13310535e-02 4.75301772e-01
-2.39947215e-01 2.70297945e-01 -1.20660864e-01 -1.03733182e+00
4.14468795e-02 -1.51378644e+00 1.35007000e+00 3.85865569e-01
2.72802293e-01 -2.81557322e-01 -4.76775527e-01 5.76166868e-01
3.86433937e-02 3.35008860e-01 6.96928203e-01 -5.76209605e-01
-1.11952496e+00 -4.70401168e-01 -6.66970730e-01 9.31604803e-02
-4.60718237e-02 -2.32372265e-02 -1.01617122e+00 4.35534455e-02
-6.97522610e-02 2.24164695e-01 1.26625562e+00 3.58500898e-01
1.33137059e+00 1.76377207e-01 -8.17266226e-01 5.73356748e-01
1.14219153e+00 1.12631544e-01 3.21215183e-01 4.68673497e-01
9.54167545e-01 6.38616920e-01 1.15615654e+00 4.02085632e-01
5.04359782e-01 7.03437686e-01 8.51309657e-01 -8.50114599e-02
-9.13981646e-02 -2.10116938e-01 3.74918342e-01 3.61270845e-01
1.08695731e-01 -2.35906884e-01 -6.94131613e-01 5.58021784e-01
-1.95856512e+00 -4.81764883e-01 -3.59274268e-01 1.91209888e+00
4.20991927e-01 6.12749279e-01 1.69530019e-01 2.26477474e-01
9.57806110e-01 3.23300242e-01 -8.50541472e-01 9.95364413e-02
1.01200506e-01 -2.14662105e-01 6.67877734e-01 1.37312144e-01
-1.27241385e+00 9.43345070e-01 4.38374710e+00 9.87459779e-01
-8.63202274e-01 1.94981515e-01 6.54134035e-01 -8.73000920e-02
-2.99793035e-02 -4.28115688e-02 -1.44307292e+00 5.75929463e-01
2.92611361e-01 1.79713532e-01 -1.63151026e-01 1.13307428e+00
3.32557619e-01 -2.37895235e-01 -8.70823681e-01 8.49687159e-01
-1.67348776e-02 -1.19305325e+00 -9.40467864e-02 7.69656524e-02
4.63203222e-01 -1.09071888e-01 2.50149821e-03 4.72672284e-01
-3.19622308e-01 -6.21173680e-01 1.11519158e+00 2.87549615e-01
4.32702720e-01 -7.84929991e-01 7.33926833e-01 6.87536657e-01
-1.38868177e+00 -4.48518425e-01 -7.97658026e-01 3.18971425e-01
1.99889675e-01 8.90305459e-01 -7.75624037e-01 4.83582288e-01
8.57838392e-01 7.31652260e-01 -7.59180546e-01 1.37834013e+00
-1.58629313e-01 4.51433480e-01 -5.67805529e-01 -2.42715567e-01
4.45108503e-01 -2.88087845e-01 8.53040218e-01 9.56042767e-01
4.36260730e-01 3.91288698e-02 4.22038913e-01 1.16388059e+00
1.08223967e-01 1.86742157e-01 -2.91355073e-01 4.17929798e-01
4.92314845e-01 1.55792081e+00 -1.18001676e+00 -2.71251082e-01
-3.82000536e-01 6.29852712e-01 2.90879697e-01 1.73096389e-01
-1.11215556e+00 -4.67042625e-01 7.14541376e-01 2.30566710e-01
7.85124123e-01 -3.37986529e-01 -3.53773445e-01 -8.39280784e-01
4.19374555e-01 -3.93578142e-01 3.47570371e-04 -4.67440784e-01
-9.26949322e-01 3.29459488e-01 5.62014505e-02 -1.46098435e+00
3.81023526e-01 -6.36364460e-01 -6.71751738e-01 5.57138503e-01
-1.92714429e+00 -1.03101873e+00 -5.57173729e-01 1.52916223e-01
6.11918628e-01 2.02932894e-01 1.65240448e-02 4.39622372e-01
-1.21682084e+00 3.72313410e-01 -2.81289458e-01 3.37622017e-02
4.24673408e-01 -1.05681825e+00 1.31986603e-01 8.17260683e-01
-5.91274574e-02 2.99678415e-01 3.68761778e-01 -6.47729278e-01
-1.28457797e+00 -1.48171031e+00 4.95582789e-01 -3.70687246e-01
3.44111651e-01 -6.57246709e-01 -1.02869940e+00 3.55979294e-01
-2.67022550e-01 4.09991175e-01 1.27898768e-01 -3.01231861e-01
2.25028526e-02 -5.50529718e-01 -9.27482247e-01 5.11778116e-01
1.19366717e+00 2.43830159e-01 -4.57416564e-01 2.76108682e-01
9.46612000e-01 -5.92829943e-01 -4.94772226e-01 6.02463305e-01
1.98432773e-01 -9.86308575e-01 1.05215979e+00 -1.10080540e-01
2.49925449e-01 -9.11541939e-01 -1.70313697e-02 -6.93823695e-01
-1.49438038e-01 -2.04310507e-01 -2.38330260e-01 1.28781343e+00
3.19559872e-01 -5.91035426e-01 8.64305556e-01 4.04978305e-01
-6.31705463e-01 -8.70948434e-01 -1.32879210e+00 -7.00834095e-01
-3.77576530e-01 -5.68269610e-01 8.34069788e-01 4.44275379e-01
-5.36530077e-01 8.09371397e-02 2.32940353e-02 6.69444859e-01
7.97637463e-01 3.27363133e-01 1.05498958e+00 -1.39381099e+00
1.57009780e-01 -5.01851678e-01 -7.86572635e-01 -1.47242320e+00
6.83697686e-02 -8.81790757e-01 4.37975764e-01 -1.66661477e+00
-7.81011730e-02 -9.19043303e-01 -4.91200507e-01 3.93968225e-01
-4.41682249e-01 2.54717350e-01 1.02005256e-02 1.66914105e-01
-7.14130223e-01 7.42343426e-01 1.23591053e+00 -2.26929396e-01
-3.46888870e-01 7.77121857e-02 -4.37420100e-01 6.87259078e-01
7.35689104e-01 -5.84480703e-01 -2.07664981e-01 -2.10847303e-01
-1.59946769e-01 -2.76540965e-01 6.97244346e-01 -1.16717374e+00
2.85695553e-01 -1.56919643e-01 4.82829869e-01 -1.27960050e+00
4.60447222e-01 -8.49515676e-01 -2.98763841e-01 3.74189913e-01
6.72158897e-02 -4.72409695e-01 2.13932961e-01 9.32661772e-01
9.39953048e-03 -2.70411909e-01 7.85327494e-01 1.97563633e-01
-9.73672271e-01 4.88624573e-01 -3.20302360e-02 -1.87334731e-01
1.51751864e+00 -6.38007045e-01 -1.05190180e-01 3.75983983e-01
-3.65494698e-01 6.16062641e-01 1.88415483e-01 6.24270916e-01
7.24124551e-01 -1.32842052e+00 -5.87857604e-01 3.23043466e-01
4.04043496e-01 7.68508434e-01 1.38032496e-01 1.09488535e+00
-3.14380169e-01 3.25061917e-01 2.13908315e-01 -1.10084069e+00
-9.30072844e-01 4.90371376e-01 2.23025873e-01 1.16897352e-01
-9.20572460e-01 9.21506763e-01 5.47734976e-01 -2.90838063e-01
2.45159864e-01 -5.85424900e-01 -4.45338130e-01 2.09309578e-01
3.51202577e-01 3.39074194e-01 -5.93623780e-02 -6.35765612e-01
-5.52507639e-01 6.82108700e-01 -1.38151318e-01 2.08479151e-01
1.30056655e+00 -1.68503940e-01 6.21244349e-02 4.24775302e-01
9.25402880e-01 -2.57999569e-01 -1.63250768e+00 -6.91863075e-02
4.54271920e-02 -5.53308427e-01 3.72324377e-01 -9.27378312e-02
-1.24219012e+00 9.61157978e-01 7.32509434e-01 2.29904726e-01
7.66357839e-01 1.25136092e-01 7.69000530e-01 2.47461021e-01
2.25644603e-01 -1.12478399e+00 -1.72603410e-02 6.20667301e-02
6.40091300e-01 -1.39765966e+00 -1.88038591e-02 -7.07640648e-01
-2.92113483e-01 1.04483223e+00 9.85670209e-01 -2.03635648e-01
7.18849838e-01 -8.77088159e-02 -7.35063106e-02 -1.09886050e-01
-3.31886590e-01 -3.45919490e-01 3.29162061e-01 3.88370067e-01
-2.35153511e-01 -6.52195513e-02 -2.24334583e-01 7.57583082e-01
1.45823225e-01 -3.40688139e-01 2.89061695e-01 6.72118604e-01
-1.02865911e+00 -7.26684034e-01 -4.18648571e-01 5.92992425e-01
3.88061889e-02 2.86483228e-01 -1.17975205e-01 9.30525303e-01
4.57245708e-01 9.94099200e-01 2.65367895e-01 -2.76881874e-01
4.43318009e-01 -1.47391364e-01 1.49818525e-01 -7.33849764e-01
-3.78237031e-02 1.68292165e-01 -2.47936144e-01 -6.18345439e-01
-3.14855486e-01 -6.39086187e-01 -1.52229488e+00 2.16017783e-01
-8.57495487e-01 -2.21064538e-01 8.63611102e-01 8.75837088e-01
4.29685563e-01 5.67115664e-01 8.60657096e-01 -1.14656699e+00
-3.39586198e-01 -7.38131225e-01 -4.99619991e-01 -1.01289146e-01
1.98066339e-01 -1.17851603e+00 -5.56647122e-01 -4.36605960e-01] | [7.910488128662109, -2.4557979106903076] |
b0db6f1e-5a62-46bf-a0ca-bacb685d522e | when-combinatorial-thompson-sampling-meets | 2302.11182 | null | https://arxiv.org/abs/2302.11182v1 | https://arxiv.org/pdf/2302.11182v1.pdf | When Combinatorial Thompson Sampling meets Approximation Regret | We study the Combinatorial Thompson Sampling policy (CTS) for combinatorial multi-armed bandit problems (CMAB), within an approximation regret setting. Although CTS has attracted a lot of interest, it has a drawback that other usual CMAB policies do not have when considering non-exact oracles: for some oracles, CTS has a poor approximation regret (scaling linearly with the time horizon $T$) [Wang and Chen, 2018]. A study is then necessary to discriminate the oracles on which CTS could learn. This study was started by Kong et al. [2021]: they gave the first approximation regret analysis of CTS for the greedy oracle, obtaining an upper bound of order $\mathcal{O}(\log(T)/\Delta^2)$, where $\Delta$ is some minimal reward gap. In this paper, our objective is to push this study further than the simple case of the greedy oracle. We provide the first $\mathcal{O}(\log(T)/\Delta)$ approximation regret upper bound for CTS, obtained under a specific condition on the approximation oracle, allowing a reduction to the exact oracle analysis. We thus term this condition REDUCE2EXACT, and observe that it is satisfied in many concrete examples. Moreover, it can be extended to the probabilistically triggered arms setting, thus capturing even more problems, such as online influence maximization. | ['Pierre Perrault'] | 2023-02-22 | null | null | null | null | ['thompson-sampling'] | ['methodology'] | [ 2.44943663e-01 5.06731927e-01 -6.43749416e-01 -1.72210932e-01
-1.25430918e+00 -1.08373797e+00 -5.44817783e-02 1.82513446e-01
-4.74418461e-01 1.02542853e+00 -1.87428206e-01 -7.91569173e-01
-8.92428577e-01 -8.51540685e-01 -1.18721640e+00 -1.04780400e+00
-9.51760933e-02 6.73128307e-01 -1.66767556e-02 -2.75209621e-02
1.18746869e-01 3.70747119e-01 -1.22543490e+00 -2.04227939e-02
6.29165411e-01 1.59325624e+00 8.27350542e-02 6.00493312e-01
-9.67014357e-02 6.01149201e-01 -3.93881023e-01 -8.59088063e-01
6.45398617e-01 -6.34067714e-01 -1.09712803e+00 -7.84599185e-02
-1.06212579e-01 -3.87800843e-01 -1.94452986e-01 1.13921964e+00
4.02888626e-01 9.28503349e-02 2.37348720e-01 -1.18882227e+00
-1.66691214e-01 1.30981493e+00 -6.43236399e-01 1.61273316e-01
6.63924590e-02 -2.51590252e-01 1.49028897e+00 1.50388062e-01
4.51442778e-01 9.81648922e-01 3.74682456e-01 2.62463540e-01
-1.07909334e+00 -3.81971359e-01 4.56717640e-01 1.60415560e-01
-9.28578496e-01 -2.44834460e-02 3.41617733e-01 2.76289193e-05
3.83437097e-01 9.22651470e-01 6.99809372e-01 8.03967714e-01
-4.24087316e-01 1.31655800e+00 1.46108031e+00 -6.68750167e-01
4.73201185e-01 5.86333424e-02 1.63917780e-01 3.07192743e-01
6.38749182e-01 1.76273242e-01 -1.02870930e-02 -2.90356904e-01
5.42859793e-01 5.54248840e-02 -3.60435575e-01 -2.86255389e-01
-6.76550746e-01 1.07131374e+00 2.81211048e-01 2.76694298e-01
-3.10870647e-01 4.95621473e-01 8.64132345e-02 5.90517998e-01
3.48839790e-01 3.43466282e-01 -5.97226024e-01 -5.09985507e-01
-9.24026847e-01 2.84471720e-01 1.05131340e+00 1.09848797e+00
4.18376386e-01 -3.64847481e-01 -5.72800457e-01 5.04001856e-01
-1.09755591e-01 6.36755407e-01 -2.36379281e-01 -1.11922181e+00
9.23549771e-01 1.21126086e-01 7.45955944e-01 -3.76018018e-01
-2.53335655e-01 -8.06963027e-01 -5.31287432e-01 -1.18731253e-01
9.64186907e-01 -3.43926013e-01 -3.38173985e-01 1.81264317e+00
3.60138983e-01 -5.46882153e-01 -3.14983904e-01 9.10466909e-01
-2.50062823e-01 3.71585041e-01 -5.93270779e-01 -8.24484348e-01
1.20147347e+00 -7.32786834e-01 -5.73174477e-01 -1.23246148e-01
6.54612243e-01 -5.49094558e-01 8.93304408e-01 7.15316951e-01
-1.34213328e+00 4.46965516e-01 -8.09015453e-01 4.49838132e-01
9.70514491e-02 -4.17723805e-01 1.11206472e+00 1.27745247e+00
-7.33728230e-01 6.83389127e-01 -6.19976878e-01 -4.61761430e-02
5.76509058e-01 3.04816604e-01 2.74938017e-01 -3.66820663e-01
-7.85041809e-01 4.85979766e-01 7.01010078e-02 1.30280629e-01
-8.50788653e-01 -3.86284918e-01 -2.70552516e-01 1.45050824e-01
1.25474417e+00 -5.20322382e-01 1.61051297e+00 -8.67118239e-01
-1.50563550e+00 5.25307178e-01 -1.12289682e-01 -6.78796172e-01
1.11197805e+00 -1.15699172e-01 3.47150862e-01 3.13906558e-02
-4.81336340e-02 -2.98748314e-01 3.71949524e-01 -9.62342381e-01
-8.52116644e-01 -5.79321146e-01 9.76949096e-01 -2.66638193e-02
-1.44387975e-01 1.21729840e-02 -1.94033250e-01 -4.45175648e-01
1.20724529e-01 -1.02698302e+00 -6.52799845e-01 -4.21623409e-01
-4.95602578e-01 -6.60086125e-02 -1.65528119e-01 -1.47041336e-01
1.18392801e+00 -1.72667360e+00 1.35247037e-01 3.45611215e-01
-1.59631222e-01 -1.97525531e-01 4.27449457e-02 6.35532260e-01
2.36004919e-01 3.71160686e-01 -2.12350368e-01 -1.98127031e-01
4.47985440e-01 2.22361267e-01 -5.23126066e-01 6.36598527e-01
-6.13415778e-01 9.20059800e-01 -7.68032312e-01 -3.62222977e-02
-2.79573053e-01 -4.13153142e-01 -6.50227368e-01 -1.23455085e-01
-5.76013267e-01 9.65729803e-02 -8.17976773e-01 7.76545465e-01
8.02397966e-01 -3.54975879e-01 3.96667480e-01 4.15569544e-01
-1.82388544e-01 5.79159297e-02 -1.51258862e+00 1.24395621e+00
-4.78407294e-01 1.16172820e-01 4.69027549e-01 -1.39923704e+00
6.73125014e-02 1.33465216e-01 5.87317944e-01 -5.20490050e-01
4.36970174e-01 3.46093148e-01 -1.88976020e-01 -3.61491412e-01
2.21564636e-01 -4.59465683e-01 -2.92503566e-01 8.34946871e-01
-5.11083186e-01 1.14005767e-01 1.95023984e-01 -4.95790876e-02
1.33379853e+00 -2.50200570e-01 1.49573877e-01 -2.04280421e-01
-3.70076410e-02 -4.27327529e-02 4.13713485e-01 1.50935972e+00
-3.34853120e-02 3.17850083e-01 1.08598185e+00 -1.71775311e-01
-8.58207464e-01 -7.51334190e-01 -2.02730522e-01 1.23804867e+00
1.80591449e-01 1.40744308e-02 -7.94498503e-01 -8.19357574e-01
2.26575643e-01 8.96508455e-01 -9.10857201e-01 1.36572808e-01
-1.26225010e-01 -1.04857135e+00 4.79190826e-01 3.34865361e-01
2.88613915e-01 -4.33667481e-01 -5.78125536e-01 3.60460758e-01
-1.68337286e-01 -9.36867774e-01 -4.66108799e-01 5.85001945e-01
-9.73410189e-01 -1.07206559e+00 -7.55232632e-01 9.24291089e-02
4.02457654e-01 3.39297771e-01 6.70398951e-01 -2.10512727e-01
5.26467450e-02 6.76945329e-01 -4.82049763e-01 -6.59582794e-01
3.77513580e-02 1.21098831e-01 -1.32414296e-01 1.37724385e-01
-6.86116144e-02 -5.23858845e-01 -7.31543422e-01 4.29065138e-01
-9.77906406e-01 -1.93845838e-01 8.19072902e-01 7.37567663e-01
7.75899410e-01 1.35854194e-02 3.66084278e-01 -1.02233183e+00
3.71476442e-01 -4.50721622e-01 -1.23091137e+00 4.61983323e-01
-6.80477679e-01 3.70150983e-01 6.45101786e-01 -3.96136492e-01
-7.62174428e-01 -1.97266296e-01 3.15685682e-02 -1.46908849e-01
3.92636418e-01 6.27006650e-01 -1.63850546e-01 -3.46640237e-02
2.65734822e-01 3.28282624e-01 -3.51236850e-01 -6.31931484e-01
3.44262898e-01 4.65491474e-01 -7.75551423e-02 -1.08749700e+00
6.19445801e-01 5.46796858e-01 4.07449096e-01 -2.94929236e-01
-1.53836203e+00 -1.23710789e-01 1.35915563e-01 5.18678986e-02
1.75878882e-01 -3.27649087e-01 -1.46664441e+00 -1.60959482e-01
-8.32727790e-01 -3.82681787e-01 -5.53062916e-01 6.04626536e-01
-1.02368331e+00 3.13923389e-01 -3.54911506e-01 -1.63621950e+00
8.49396084e-03 -9.12303030e-01 4.57641035e-01 9.42122042e-02
3.99288058e-01 -6.57396793e-01 -1.34709999e-01 7.45242417e-01
4.58224386e-01 3.03882420e-01 7.02756763e-01 -5.97934425e-01
-1.03329194e+00 -2.43145525e-01 -1.34389564e-01 1.41649231e-01
-3.04467112e-01 -6.22866213e-01 -7.31478095e-01 -5.06061077e-01
3.75024766e-01 -3.69086832e-01 8.54783833e-01 7.79473126e-01
1.48047304e+00 -9.91286874e-01 -3.33372533e-01 5.97332478e-01
1.52934885e+00 2.02088296e-01 5.11659205e-01 6.55955374e-01
1.53863635e-02 2.68299192e-01 8.31169426e-01 8.51753831e-01
1.24189749e-01 7.07015514e-01 8.99117112e-01 3.61055315e-01
5.76905251e-01 -1.64588660e-01 2.20477954e-01 6.15586936e-02
-4.51299816e-01 -3.21660727e-01 -2.70103097e-01 6.28039062e-01
-1.92273283e+00 -8.90055299e-01 -3.15887369e-02 2.87840962e+00
9.72010195e-01 2.66898543e-01 5.02341688e-01 2.53829926e-01
6.44693315e-01 -5.21597862e-02 -6.99981570e-01 -6.85299277e-01
-2.66516536e-01 4.88548189e-01 1.23754323e+00 5.19870520e-01
-7.38775313e-01 4.59500015e-01 5.56880093e+00 1.37304723e+00
-7.55702615e-01 4.43078160e-01 8.06640327e-01 -8.22694540e-01
-4.57916677e-01 2.99495429e-01 -7.16200888e-01 5.33890128e-01
8.83308828e-01 -3.70171845e-01 1.02686584e+00 1.05828953e+00
9.61909443e-02 -3.74460459e-01 -1.23513258e+00 7.29985654e-01
-5.02137065e-01 -1.24851406e+00 -5.09234130e-01 5.93566060e-01
8.35044444e-01 -1.70447960e-01 1.07966453e-01 1.67789981e-01
6.82946622e-01 -9.93429959e-01 8.26234937e-01 2.46705577e-01
7.27885306e-01 -9.81095552e-01 7.77649760e-01 5.90758562e-01
-7.60213435e-01 -5.89115798e-01 -3.15479010e-01 -1.96118176e-01
2.32143313e-01 9.02038157e-01 -5.22888958e-01 1.01606488e+00
7.35434115e-01 -1.48880109e-01 1.34021670e-01 1.29053187e+00
-1.47407591e-01 8.40119660e-01 -8.66622686e-01 -5.25083482e-01
5.35113394e-01 -3.33955288e-01 6.29702151e-01 8.09575140e-01
4.45422530e-01 3.57847154e-01 -6.08291477e-02 6.07935250e-01
-1.63468391e-01 1.41375616e-01 -1.60778403e-01 -1.74874082e-01
4.38794345e-01 9.50606525e-01 -8.28791022e-01 -6.05060300e-03
-6.51560873e-02 8.13137174e-01 2.15689287e-01 2.19332024e-01
-1.13269365e+00 -2.96558559e-01 5.07471323e-01 7.33196810e-02
8.26152444e-01 8.73645693e-02 -5.36098897e-01 -7.75804520e-01
5.79345047e-01 -4.87159610e-01 6.19816959e-01 1.14347739e-02
-1.14131629e+00 4.96205613e-02 1.66692615e-01 -1.01350820e+00
-7.30390772e-02 -5.01959682e-01 -8.59520435e-02 5.52434504e-01
-1.37086546e+00 -7.60122120e-01 3.59036475e-01 4.01312590e-01
9.28325951e-02 4.41837609e-01 5.48436761e-01 1.30982772e-01
-4.83035713e-01 1.01826668e+00 8.58301938e-01 -4.12180930e-01
-7.44229928e-02 -1.28701329e+00 -4.31415170e-01 6.69173896e-01
-1.73805598e-02 3.75979692e-01 9.08708811e-01 -1.26355186e-01
-1.97192514e+00 -7.20158935e-01 4.93744314e-01 -2.69918919e-01
7.53436744e-01 -2.94806778e-01 -7.76882563e-03 8.39651644e-01
-4.43335921e-02 -1.28719181e-01 5.27566254e-01 3.34026098e-01
-1.38230473e-01 -4.30334538e-01 -1.39129996e+00 5.32419622e-01
1.40263641e+00 -3.41816656e-02 -2.53146123e-02 5.49525261e-01
6.58177257e-01 -5.19846857e-01 -8.51602972e-01 1.85734183e-01
6.48655534e-01 -1.08287942e+00 6.22500181e-01 -6.98058963e-01
1.43204495e-01 1.37031332e-01 -4.37780201e-01 -9.65855181e-01
6.17768168e-02 -1.17607343e+00 -2.72092640e-01 1.03896809e+00
5.71312726e-01 -9.27453458e-01 8.27787519e-01 7.13170230e-01
9.73235369e-02 -1.12496305e+00 -1.66385210e+00 -1.17003977e+00
3.05053324e-01 -9.28857923e-01 6.50201917e-01 6.59665525e-01
7.22209588e-02 -1.91919178e-01 -4.75792050e-01 8.91922265e-02
7.50381291e-01 5.36776423e-01 5.16888738e-01 -7.86036432e-01
-1.05555725e+00 -7.54448473e-01 1.58185169e-01 -1.48805094e+00
-4.74476963e-01 -7.92328954e-01 -1.84302345e-01 -1.28250444e+00
4.07375813e-01 -9.01571631e-01 -3.88814598e-01 4.87178802e-01
2.31944501e-01 -2.58012444e-01 3.19502532e-01 -8.25076401e-02
-8.27558219e-01 5.11560440e-01 1.33559155e+00 -2.14979842e-01
6.41842112e-02 9.36473191e-01 -1.23134947e+00 2.78161377e-01
5.62722623e-01 -7.54145265e-01 -2.15565071e-01 -4.19861645e-01
6.60894930e-01 6.71443701e-01 1.80538118e-01 -4.81432855e-01
9.25132707e-02 -7.03521371e-01 -2.44728014e-01 -4.85816836e-01
1.87823519e-01 -9.01410878e-01 3.02308857e-01 5.33744335e-01
-4.61272448e-01 -3.93051237e-01 -1.00787863e-01 8.85553479e-01
3.75069171e-01 -7.58760989e-01 4.65035081e-01 -1.85156718e-01
5.08038402e-01 3.57496202e-01 -1.74194768e-01 1.38174579e-01
1.12510467e+00 -1.33801103e-01 -2.13253215e-01 -7.37012267e-01
-5.99578500e-01 3.03305507e-01 2.30105475e-01 -2.66160548e-01
2.85021961e-02 -1.11566854e+00 -5.05971611e-01 -5.12330770e-01
-1.82727918e-01 8.21225420e-02 3.88134152e-01 1.38745213e+00
-6.90440834e-02 7.09651828e-01 4.95095342e-01 -3.31705600e-01
-6.97541893e-01 8.06302905e-01 2.90630996e-01 -7.29630291e-01
-1.83930680e-01 8.26738536e-01 -9.98544991e-02 1.08430326e-01
3.94899577e-01 -5.29739022e-01 4.81033266e-01 8.97804461e-03
4.71394986e-01 6.36251688e-01 2.82468032e-02 2.55846858e-01
-1.65512368e-01 2.42548957e-01 -3.36899720e-02 -4.98528868e-01
1.62346280e+00 -2.54744887e-01 -1.54975191e-01 3.55447471e-01
9.33879495e-01 2.57263243e-01 -1.15360856e+00 -2.97864705e-01
1.92867547e-01 -6.48457408e-01 -1.40762970e-01 -9.41868365e-01
-1.19450438e+00 5.07837296e-01 4.16762024e-01 9.46620822e-01
1.09268641e+00 2.91449577e-01 6.30794525e-01 5.24093390e-01
9.53147471e-01 -1.14188850e+00 -5.17917335e-01 4.35337156e-01
7.77731895e-01 -9.00589347e-01 -5.71908131e-02 -1.56634659e-01
-3.84721756e-01 9.75158274e-01 -8.89325216e-02 1.00901857e-01
3.74941021e-01 1.06555693e-01 -6.55780733e-01 6.83055371e-02
-6.09102786e-01 -5.24417996e-01 -2.22164303e-01 1.00994512e-01
-3.60776894e-02 4.56850052e-01 -8.59965146e-01 1.31968868e+00
-3.59433562e-01 1.36525139e-01 5.71180284e-01 9.77432013e-01
-3.89676362e-01 -1.29932928e+00 -5.01927435e-01 6.40201747e-01
-9.90603566e-01 -3.27647515e-02 -2.52455920e-01 6.93632364e-01
-2.45244965e-01 1.09274149e+00 -3.35776895e-01 3.87953408e-02
2.92398483e-01 -2.04252064e-01 8.87173653e-01 -1.28394663e-01
-3.80823821e-01 1.70408990e-02 2.93496996e-01 -5.30705452e-01
-2.59506971e-01 -7.91127980e-01 -7.27304399e-01 -5.25781572e-01
-5.52352846e-01 5.83299875e-01 6.83828413e-01 1.19318473e+00
2.92311795e-02 1.01795182e-01 1.07497740e+00 -4.88922536e-01
-1.18380475e+00 -9.92896914e-01 -8.87370944e-01 -9.56214592e-02
1.97223678e-01 -5.86267054e-01 -5.64307213e-01 -7.71160722e-01] | [4.563113689422607, 3.351210355758667] |
d74f3fb4-2ec6-48ce-a9b3-f08670ec50fc | quantitative-stock-investment-by-routing | 2207.07578 | null | https://arxiv.org/abs/2207.07578v1 | https://arxiv.org/pdf/2207.07578v1.pdf | Quantitative Stock Investment by Routing Uncertainty-Aware Trading Experts: A Multi-Task Learning Approach | Quantitative investment is a fundamental financial task that highly relies on accurate stock prediction and profitable investment decision making. Despite recent advances in deep learning (DL) have shown stellar performance on capturing trading opportunities in the stochastic stock market, we observe that the performance of existing DL methods is sensitive to random seeds and network initialization. To design more profitable DL methods, we analyze this phenomenon and find two major limitations of existing works. First, there is a noticeable gap between accurate financial predictions and profitable investment strategies. Second, investment decisions are made based on only one individual predictor without consideration of model uncertainty, which is inconsistent with the workflow in real-world trading firms. To tackle these two limitations, we first reformulate quantitative investment as a multi-task learning problem. Later on, we propose AlphaMix, a novel two-stage mixture-of-experts (MoE) framework for quantitative investment to mimic the efficient bottom-up trading strategy design workflow of successful trading firms. In Stage one, multiple independent trading experts are jointly optimized with an individual uncertainty-aware loss function. In Stage two, we train neural routers (corresponding to the role of a portfolio manager) to dynamically deploy these experts on an as-needed basis. AlphaMix is also a universal framework that is applicable to various backbone network architectures with consistent performance gains. Through extensive experiments on long-term real-world data spanning over five years on two of the most influential financial markets (US and China), we demonstrate that AlphaMix significantly outperforms many state-of-the-art baselines in terms of four financial criteria. | ['Bo An', 'Rundong Wang', 'Shuo Sun'] | 2022-06-07 | null | null | null | null | ['stock-prediction'] | ['time-series'] | [-4.65822965e-01 -1.11052714e-01 -5.01642644e-01 -2.78672725e-01
-4.98217195e-01 -6.39784455e-01 5.09176373e-01 -4.27388906e-01
-3.30104649e-01 6.81401253e-01 -8.79060179e-02 -6.57604694e-01
-4.38462555e-01 -9.70368326e-01 -6.87871099e-01 -4.02666569e-01
-1.35081619e-01 1.11101699e+00 1.45607114e-01 -1.93651915e-02
3.95782858e-01 4.19332027e-01 -7.98219621e-01 2.83624709e-01
6.90262496e-01 1.92783082e+00 -3.82532179e-01 1.08253039e-01
-3.82112980e-01 1.35566866e+00 -6.66557312e-01 -1.05096054e+00
1.01728153e+00 6.52081817e-02 -3.38137418e-01 -3.96718867e-02
-6.06521964e-02 -7.67632842e-01 -1.03893168e-01 9.19078350e-01
4.46546376e-01 -1.81476712e-01 6.55436456e-01 -1.44784451e+00
-3.25236112e-01 1.32579303e+00 -7.38830388e-01 3.78777891e-01
-9.02949035e-01 4.62600231e-01 1.56516862e+00 -6.55854583e-01
8.53038579e-02 1.13307738e+00 6.95156693e-01 -1.10584296e-01
-1.07496178e+00 -1.11795068e+00 4.27639335e-01 -3.09477985e-01
-4.86315817e-01 -3.37982893e-01 8.90592277e-01 -5.56424856e-01
8.84010315e-01 -2.67043382e-01 6.70185506e-01 9.07001436e-01
4.57000047e-01 8.62115383e-01 1.15683556e+00 1.67424098e-01
3.91917109e-01 2.11021975e-01 -2.81544089e-01 1.71743289e-01
3.14800352e-01 4.67390567e-01 -5.24961770e-01 -2.50933319e-01
1.09732258e+00 3.05578768e-01 3.75154585e-01 -2.28703812e-01
-1.14651513e+00 1.11809599e+00 3.44035447e-01 1.36674479e-01
-7.10214019e-01 5.37596881e-01 3.88378412e-01 9.22905028e-01
7.74766982e-01 8.23624790e-01 -9.77024257e-01 -1.50818020e-01
-1.36296701e+00 3.03247035e-01 1.19325280e+00 6.35741711e-01
4.53936905e-01 5.17042220e-01 -3.07266742e-01 4.66214240e-01
4.94313955e-01 2.06318662e-01 3.41495633e-01 -1.38722217e+00
7.84515262e-01 3.69385689e-01 1.39979646e-01 -9.29068029e-01
-2.50374138e-01 -1.25674605e+00 -8.03303182e-01 5.06172299e-01
4.31634575e-01 -5.25035679e-01 -3.75875860e-01 1.22233629e+00
-2.44496539e-01 3.52645010e-01 -1.01028942e-01 4.57444727e-01
4.20501642e-02 3.74235362e-01 -3.57312649e-01 -1.95020229e-01
9.70121324e-01 -1.56302643e+00 -2.30228782e-01 -3.40670586e-01
7.54566267e-02 -5.80842793e-01 4.51320112e-01 8.11461389e-01
-1.17720461e+00 -3.01136941e-01 -9.84991372e-01 7.73389161e-01
-2.94633836e-01 -3.36675420e-02 9.09581542e-01 4.12949979e-01
-8.70486975e-01 9.81212497e-01 -6.24803424e-01 7.45933056e-01
7.08926141e-01 2.78996110e-01 3.76289815e-01 5.29895961e-01
-1.23013163e+00 7.48082936e-01 1.45867735e-01 2.87905246e-01
-1.12054849e+00 -1.21743011e+00 -6.74577132e-02 3.66951734e-01
9.70829487e-01 -6.98063135e-01 1.75477588e+00 -1.24221468e+00
-1.86220980e+00 4.27299559e-01 7.13544190e-01 -1.22079897e+00
1.31533551e+00 -3.78557682e-01 -3.12225431e-01 -9.73711908e-02
-9.15173739e-02 2.32299075e-01 9.84085143e-01 -9.13843751e-01
-8.35365534e-01 1.51338382e-02 1.64326429e-01 -1.43683627e-01
-2.77613401e-01 1.25867710e-01 -2.45969385e-01 -1.06884265e+00
-2.33486399e-01 -5.52979290e-01 -5.77557921e-01 -1.98030472e-01
-4.06522721e-01 -1.93607081e-02 2.49014959e-01 -5.48408031e-01
1.30390501e+00 -1.56932235e+00 -1.72882766e-01 4.86496210e-01
3.71724069e-01 6.96907938e-02 1.03836603e-01 3.22680086e-01
2.85283253e-02 3.60041797e-01 -2.23402679e-03 -6.29985094e-01
6.88836813e-01 -1.76518872e-01 -7.25213885e-01 8.23011398e-02
2.64342606e-01 1.24133980e+00 -5.61502635e-01 -1.62806079e-01
1.23752153e-03 -1.89589828e-01 -3.64823967e-01 2.70886570e-01
-7.10194945e-01 1.93558097e-01 -6.51541114e-01 1.15966237e+00
5.31868219e-01 -5.94593465e-01 1.38161242e-01 -1.80716496e-02
-3.77139933e-02 2.82395124e-01 -1.16722703e+00 9.19676960e-01
-4.56859380e-01 2.21292794e-01 1.26591533e-01 -1.28281939e+00
8.53193223e-01 1.71007574e-01 7.41138577e-01 -7.08141088e-01
1.62189856e-01 6.83196902e-01 -5.46771381e-03 2.86868542e-01
3.19382213e-02 -3.50061148e-01 -1.24064185e-01 1.16537488e+00
-7.84656256e-02 5.35189994e-02 1.82315215e-01 -3.52525525e-02
1.10555303e+00 -1.20588295e-01 -3.05897389e-02 -1.21063113e-01
4.63638231e-02 -3.15880924e-01 1.00504375e+00 9.29001749e-01
-3.32965225e-01 2.19560146e-01 1.20455909e+00 -5.47968864e-01
-8.20821106e-01 -1.01295221e+00 2.62701605e-02 9.61485565e-01
-5.24437964e-01 2.21083283e-01 -1.96338966e-01 -9.90395308e-01
7.67712891e-01 6.70171440e-01 -5.47184050e-01 3.05128753e-01
-3.98008436e-01 -1.08441138e+00 2.43083268e-01 5.55366755e-01
7.98873305e-01 -1.26180792e+00 -5.51793635e-01 6.76175773e-01
5.18679798e-01 -9.17891741e-01 -4.45436418e-01 3.56070727e-01
-7.35459328e-01 -1.07234037e+00 -8.76626849e-01 5.35370596e-02
1.85818195e-01 -2.94701308e-01 1.86852562e+00 -2.30514139e-01
1.56340688e-01 1.83755711e-01 4.54972796e-02 -6.56705797e-01
-2.89172560e-01 1.36187643e-01 -1.23269059e-01 1.99252903e-01
2.03663677e-01 -8.04367065e-01 -7.96051025e-01 3.46217960e-01
-6.15644097e-01 -2.94659823e-01 1.03462327e+00 6.87615216e-01
5.23167610e-01 5.80326259e-01 1.04922307e+00 -1.07833183e+00
7.85678029e-01 -7.57607102e-01 -1.34984922e+00 4.41105932e-01
-1.30279839e+00 1.45734057e-01 5.25883317e-01 -3.98794889e-01
-8.96629095e-01 -5.81987143e-01 4.51615930e-01 -7.06405878e-01
7.59124637e-01 8.41313839e-01 1.48304254e-01 8.16849396e-02
-1.64294213e-01 1.92943700e-02 1.12515405e-01 -5.52795827e-01
-1.96925714e-03 2.84294546e-01 1.55043125e-01 -7.12592721e-01
9.58520770e-01 1.50750726e-01 4.78359610e-02 2.38074243e-01
-1.04905331e+00 1.50355041e-01 -4.14208889e-01 -1.70778960e-01
2.10510597e-01 -1.01714861e+00 -7.95982420e-01 9.42795157e-01
-6.46050155e-01 -6.47845387e-01 -3.95612925e-01 3.70438337e-01
-4.88657475e-01 -1.38825849e-01 -8.83084238e-01 -8.45158875e-01
-6.08037710e-01 -1.04652023e+00 4.87012506e-01 2.68678623e-03
1.71871752e-01 -1.34817791e+00 1.39359340e-01 4.63166833e-01
9.80772555e-01 3.93731594e-01 8.03261578e-01 -1.12059033e+00
-1.03394508e+00 -1.64806291e-01 -3.36391181e-01 9.21713769e-01
-2.42602378e-02 1.62444249e-01 -6.36004806e-01 -2.34333873e-01
1.99180767e-01 -4.74990755e-01 1.16238821e+00 5.58362305e-01
9.42213655e-01 -3.78478914e-01 -3.94666456e-02 6.90228343e-01
1.29692090e+00 2.60419965e-01 1.54616877e-01 6.85079694e-01
1.82862297e-01 7.55469978e-01 6.34525537e-01 7.78107464e-01
2.76201844e-01 2.04211548e-01 5.51120996e-01 8.01461935e-02
6.26285017e-01 -2.66177803e-01 6.04747295e-01 5.16131282e-01
1.29970714e-01 -1.34472996e-01 -8.59593570e-01 1.81262553e-01
-1.96830237e+00 -8.51413488e-01 6.97809875e-01 2.00318456e+00
9.24008727e-01 8.45048189e-01 4.66776252e-01 -2.53354847e-01
3.01021218e-01 4.86573428e-01 -1.18213046e+00 -3.18864807e-02
-1.63712859e-01 1.52067840e-01 9.22981381e-01 -4.26081084e-02
-1.18623316e+00 7.28592753e-01 6.30436993e+00 8.25476587e-01
-1.16408420e+00 -5.14913164e-02 1.43213880e+00 -6.38225555e-01
-4.47542131e-01 -7.50586763e-02 -9.15801346e-01 8.33966136e-01
1.07496107e+00 -3.81703854e-01 5.67784607e-01 9.71439600e-01
1.33424364e-02 4.93155718e-01 -1.05451798e+00 5.22264063e-01
-5.04280210e-01 -1.79513574e+00 -2.62362272e-01 3.54420185e-01
8.76067996e-01 3.25235128e-01 6.58046842e-01 6.50051892e-01
9.49909389e-01 -1.18410265e+00 9.64686334e-01 8.81732285e-01
3.43963027e-01 -1.13867915e+00 9.38721895e-01 1.77674040e-01
-1.04685152e+00 -5.41765511e-01 -3.44817281e-01 1.78512812e-01
1.95211381e-01 1.10051799e+00 -3.13311249e-01 4.32749182e-01
7.44270265e-01 9.00706649e-01 -2.28801563e-01 7.92643249e-01
-2.01071247e-01 5.89277685e-01 -4.16648656e-01 2.40179569e-01
7.63658881e-01 -5.20551801e-01 1.53475836e-01 7.09833920e-01
3.53135794e-01 -5.13387263e-01 1.01868428e-01 1.34138131e+00
-5.55803239e-01 -2.34516248e-01 -1.57222524e-01 -5.96825898e-01
4.87887442e-01 1.30125272e+00 -6.32084370e-01 -2.54014969e-01
-6.72301054e-01 4.04617459e-01 8.54382440e-02 3.12969297e-01
-8.88004541e-01 -1.41289815e-01 6.13626301e-01 1.27340987e-01
5.70849001e-01 5.66341691e-02 -4.03536886e-01 -1.14385664e+00
1.09133296e-01 -1.20222783e+00 2.75129408e-01 -5.54236323e-02
-1.82202065e+00 3.31957191e-01 -3.05448383e-01 -1.18774140e+00
-4.78670835e-01 -7.93044627e-01 -1.01263833e+00 8.19701552e-01
-1.99151373e+00 -8.91362488e-01 4.86211479e-01 1.58187121e-01
3.47862035e-01 -9.62879360e-01 1.35732725e-01 1.81227475e-01
-7.66596854e-01 5.82054377e-01 4.71289545e-01 3.53823423e-01
5.12733400e-01 -1.47519219e+00 7.01980352e-01 7.38370240e-01
1.36409447e-01 2.43951365e-01 2.28693932e-01 -6.69262528e-01
-9.49748278e-01 -1.08906662e+00 3.70322078e-01 -3.10549140e-01
1.41067684e+00 -1.13101415e-01 -5.60218096e-01 8.85971010e-01
3.11250925e-01 5.78126945e-02 6.48567915e-01 -2.74860859e-02
-2.86288500e-01 -6.44608438e-01 -1.08039010e+00 1.85387701e-01
6.09570205e-01 -1.99516907e-01 -2.62820095e-01 2.39461601e-01
9.17120099e-01 4.22004052e-02 -1.05569756e+00 5.12505829e-01
4.88803238e-01 -1.47611952e+00 9.79745448e-01 -6.36491776e-01
4.50068682e-01 2.38979116e-01 7.62259141e-02 -1.22137368e+00
-1.46024078e-01 -1.22972393e+00 -7.16041446e-01 1.25044215e+00
6.42243147e-01 -1.20638359e+00 1.01650429e+00 6.92028105e-01
3.57524365e-01 -1.14428365e+00 -8.97459567e-01 -8.82040143e-01
4.80331630e-01 -2.98437804e-01 1.04169297e+00 7.10224926e-01
-5.92149794e-01 -7.28254020e-02 -4.72250015e-01 -2.56533027e-01
1.04193485e+00 6.54274583e-01 5.31552672e-01 -1.42642403e+00
-9.39641654e-01 -1.09634626e+00 2.77103961e-01 -9.17057753e-01
3.77724230e-01 -4.42359477e-01 -4.02845293e-01 -9.72157121e-01
-8.25608373e-02 -7.58085787e-01 -1.10911167e+00 2.54875571e-01
2.62252927e-01 -2.43640095e-01 2.17681453e-01 4.81350332e-01
-6.09472573e-01 6.29792929e-01 9.79315579e-01 -2.58601248e-01
-2.87657846e-02 8.23197126e-01 -1.09865940e+00 6.59007370e-01
8.57413828e-01 -3.01403701e-01 -3.16512227e-01 -2.08470657e-01
7.10619748e-01 2.11541802e-01 1.18576050e-01 -6.78893864e-01
2.44286329e-01 -3.69819283e-01 2.19336525e-01 -7.52099872e-01
-4.43085395e-02 -8.01148057e-01 2.76766330e-01 6.90831780e-01
-2.41865933e-01 2.89108098e-01 -1.71718627e-01 6.65401995e-01
-4.85738784e-01 -2.21595928e-01 6.39010191e-01 -4.75520432e-01
-4.06444341e-01 7.62480557e-01 1.36761619e-02 3.29076409e-01
1.00789249e+00 2.89803922e-01 -2.91449070e-01 -3.32375169e-01
-4.05235469e-01 9.24206734e-01 6.75244033e-02 1.97540745e-01
2.25893900e-01 -1.12433803e+00 -9.20748055e-01 -3.79798450e-02
-4.90717620e-01 1.42127737e-01 -1.46136001e-01 8.85994673e-01
-5.08494377e-01 5.19287288e-01 -1.13847740e-02 4.07391712e-02
-2.12400213e-01 3.85134757e-01 9.62087691e-01 -1.25730276e+00
-2.20324665e-01 1.09750187e+00 -1.86662711e-02 -4.01978940e-01
5.05431831e-01 -5.53601444e-01 -6.51876330e-02 6.09880507e-01
4.00174856e-01 5.13183236e-01 -1.23702906e-01 7.81211257e-02
-4.89938166e-03 3.47891152e-01 -1.96980655e-01 -1.51000038e-01
1.84885216e+00 1.57674327e-01 -2.48970032e-01 5.94415486e-01
7.58157790e-01 -1.28724739e-01 -1.98626220e+00 -7.10697114e-01
6.29954040e-01 -3.29766840e-01 1.74735859e-01 -1.06896579e+00
-1.93851972e+00 8.72678041e-01 -2.77854316e-02 3.59379977e-01
9.62949693e-01 -4.63434130e-01 9.29492176e-01 3.63697320e-01
4.56086844e-01 -1.45012057e+00 3.66026878e-01 2.78209716e-01
7.56245673e-01 -1.39694989e+00 2.17316654e-02 3.65535021e-01
-8.07545900e-01 1.24636924e+00 4.84872818e-01 -3.44551533e-01
1.29614472e+00 6.21026099e-01 1.58654660e-01 -1.75049096e-01
-1.31296885e+00 2.89134979e-01 1.74366087e-01 -2.52552126e-02
-1.67354860e-03 1.21239051e-02 4.86309290e-01 1.26172364e+00
-1.49835229e-01 6.58690755e-04 2.44353309e-01 8.38534713e-01
-3.27869296e-01 -1.13917696e+00 9.31202993e-02 9.33189571e-01
-9.76689994e-01 -2.75947750e-01 -3.86087418e-01 7.33888328e-01
-2.48425782e-01 5.61299980e-01 3.03909510e-01 -1.98570058e-01
3.02019477e-01 -5.55874892e-02 -6.31862953e-02 -4.74300712e-01
-1.02522969e+00 2.22899452e-01 -6.13010600e-02 -6.75820231e-01
-1.49307415e-01 -7.64906228e-01 -7.94993877e-01 -6.36503637e-01
-3.42228310e-03 5.23859961e-03 1.68607652e-01 7.99009502e-01
3.41939539e-01 7.89099097e-01 1.14794064e+00 -6.87456787e-01
-1.69515920e+00 -8.14697742e-01 -1.16688764e+00 -2.88534701e-01
2.33291194e-01 -7.36850202e-01 -8.48860025e-01 -5.84783375e-01] | [4.523787975311279, 4.09350061416626] |
682bfd2f-3cc4-4af9-9952-4475187122d5 | constructing-natural-language-explanations | 2210.07222 | null | https://arxiv.org/abs/2210.07222v3 | https://arxiv.org/pdf/2210.07222v3.pdf | Saliency Map Verbalization: Comparing Feature Importance Representations from Model-free and Instruction-based Methods | Saliency maps can explain a neural model's predictions by identifying important input features. They are difficult to interpret for laypeople, especially for instances with many features. In order to make them more accessible, we formalize the underexplored task of translating saliency maps into natural language and compare methods that address two key challenges of this approach -- what and how to verbalize. In both automatic and human evaluation setups, using token-level attributions from text classification tasks, we compare two novel methods (search-based and instruction-based verbalizations) against conventional feature importance representations (heatmap visualizations and extractive rationales), measuring simulatability, faithfulness, helpfulness and ease of understanding. Instructing GPT-3.5 to generate saliency map verbalizations yields plausible explanations which include associations, abstractive summarization and commonsense reasoning, achieving by far the highest human ratings, but they are not faithfully capturing numeric information and are inconsistent in their interpretation of the task. In comparison, our search-based, model-free verbalization approach efficiently completes templated verbalizations, is faithful by design, but falls short in helpfulness and simulatability. Our results suggest that saliency map verbalization makes feature attribution explanations more comprehensible and less cognitively challenging to humans than conventional representations. | ['Sebastian Möller', 'Robert Schwarzenberg', 'Christopher Ebert', 'Maximilian Dustin Nasert', 'Leonhard Hennig', 'Nils Feldhus'] | 2022-10-13 | null | null | null | null | ['abstractive-text-summarization'] | ['natural-language-processing'] | [ 4.30158645e-01 8.88521910e-01 -2.22083077e-01 -5.23467600e-01
-4.18182731e-01 -4.57485884e-01 7.84669638e-01 7.45945334e-01
-2.99332470e-01 7.06957459e-01 7.29987085e-01 -5.38574636e-01
-3.33773941e-01 -4.12611842e-01 -5.58936775e-01 -2.08171755e-02
6.13879800e-01 5.14009058e-01 -1.02764390e-01 -3.82498533e-01
9.94910300e-01 1.35522559e-01 -1.76541817e+00 4.55639899e-01
1.39371872e+00 6.03202879e-01 1.71029776e-01 4.87898111e-01
-2.87868470e-01 1.01838529e+00 -5.92131436e-01 -6.24052227e-01
-3.63929778e-01 -5.46023250e-01 -1.28817642e+00 -2.02053964e-01
7.64369071e-01 -2.25025907e-01 2.58895546e-01 9.27702427e-01
6.93627000e-02 9.91736576e-02 1.13020420e+00 -1.47995138e+00
-1.35889578e+00 8.91921937e-01 -2.74944395e-01 3.80595058e-01
5.60334861e-01 3.29177707e-01 1.31624448e+00 -9.47916448e-01
4.36599791e-01 1.42462254e+00 7.61975884e-01 6.77940667e-01
-1.58525455e+00 -5.82773626e-01 3.19414020e-01 3.04082394e-01
-7.24210322e-01 -3.79339337e-01 4.93210644e-01 -6.31467164e-01
1.39905632e+00 6.62475526e-01 8.38472724e-01 1.10174263e+00
1.16064921e-01 6.54672921e-01 1.21273565e+00 -3.88664275e-01
2.91322917e-01 3.82906437e-01 6.38127029e-01 7.08170414e-01
6.79734468e-01 -7.53511041e-02 -9.39299405e-01 -3.28206122e-02
8.16988885e-01 2.28309166e-02 -3.68088126e-01 -1.39903471e-01
-1.59693348e+00 1.13552594e+00 8.26228023e-01 1.42973945e-01
-6.70283973e-01 3.74617308e-01 2.22872913e-01 8.28759074e-02
2.96191216e-01 1.39808774e+00 -2.42872432e-01 -2.39619806e-01
-1.11456919e+00 6.09181166e-01 5.81802666e-01 8.25966716e-01
6.60378993e-01 4.06368762e-01 -4.98862267e-01 4.52653050e-01
2.06623837e-01 4.14860636e-01 8.81597877e-01 -9.21344638e-01
2.83714682e-01 7.37422585e-01 5.36797047e-02 -1.44469213e+00
-7.28098810e-01 -4.87111360e-01 -3.99224102e-01 3.05518866e-01
1.58046797e-01 1.23793386e-01 -8.07373166e-01 1.74578953e+00
-2.29258507e-01 -4.96814966e-01 7.16134235e-02 1.07730818e+00
1.25346446e+00 4.77244109e-01 5.62282979e-01 9.51670706e-02
1.62264526e+00 -9.19446051e-01 -8.38212550e-01 -7.09579110e-01
5.13272405e-01 -4.54597324e-01 1.85738754e+00 1.30351499e-01
-1.31279147e+00 -5.25342405e-01 -1.17149913e+00 -4.75828141e-01
-4.72869247e-01 -1.65842205e-01 7.52733409e-01 4.59987253e-01
-1.21629119e+00 6.93310738e-01 -4.49691266e-01 -4.38979238e-01
5.67634523e-01 1.75972462e-01 -1.98025331e-01 4.54154253e-01
-1.15203094e+00 1.64970887e+00 4.58452314e-01 -5.76344252e-01
-4.99111265e-01 -9.87222850e-01 -1.19860697e+00 5.51712036e-01
1.51076809e-01 -1.07433975e+00 1.51211834e+00 -1.08472764e+00
-1.01327991e+00 8.00642967e-01 -2.55758971e-01 -5.11929035e-01
2.50466675e-01 -2.06045195e-01 -1.26082167e-01 1.24795094e-01
5.21586359e-01 1.24022019e+00 8.44756186e-01 -1.04732871e+00
-1.75508037e-01 -2.40348205e-02 2.14707360e-01 3.07505280e-01
-2.81989217e-01 -1.17946319e-01 4.79682833e-01 -7.72043467e-01
1.00104183e-01 -5.08309722e-01 -6.26511648e-02 9.30237919e-02
-7.81876624e-01 -2.30665669e-01 3.27386737e-01 -5.48228741e-01
1.07169127e+00 -1.74274826e+00 3.75238843e-02 -1.27478316e-01
7.33297646e-01 1.67312156e-02 7.74970725e-02 2.31595904e-01
-3.41616124e-01 6.67919815e-01 -1.08440571e-01 -7.32698590e-02
4.38781351e-01 -2.10431352e-01 -5.52633524e-01 -3.57676446e-02
5.28697014e-01 1.29652452e+00 -1.13315201e+00 -5.47062993e-01
1.99428573e-01 2.79686511e-01 -7.34122694e-01 2.24165991e-02
-2.25076359e-02 -3.58729959e-01 -1.23283699e-01 2.70995498e-01
1.64051160e-01 -5.17617702e-01 -2.24058613e-01 -2.01214612e-01
-1.54930636e-01 8.95661533e-01 -6.63149059e-01 1.20962059e+00
-3.89616996e-01 1.05050945e+00 -6.41423047e-01 -4.66117948e-01
8.04001570e-01 1.87217325e-01 -5.88661849e-01 -4.44473833e-01
5.27325273e-03 1.00369446e-01 -4.32055118e-03 -4.46346223e-01
8.07186604e-01 -5.09777486e-01 -1.42715678e-01 9.01591539e-01
1.88004881e-01 -6.69801056e-01 -5.83646372e-02 5.48422754e-01
7.70636737e-01 -6.51065260e-02 9.26465571e-01 -7.09386468e-01
-1.65628091e-01 4.32678908e-01 1.52202938e-02 7.89227962e-01
-1.01766728e-01 7.17956066e-01 7.74338305e-01 -5.02440393e-01
-9.99922633e-01 -1.05410087e+00 2.56822586e-01 1.17655158e+00
6.44189343e-02 -7.00900853e-01 -9.37548995e-01 -4.62747931e-01
-3.65379862e-02 1.80582631e+00 -1.06292665e+00 -5.27212679e-01
1.56945705e-01 -1.33341715e-01 4.13003534e-01 7.13630736e-01
2.00713173e-01 -1.35432422e+00 -1.53369999e+00 5.26722381e-03
-3.71912569e-01 -5.65930724e-01 -4.66820121e-01 3.09474736e-01
-8.71003151e-01 -8.35875332e-01 -3.86466891e-01 -4.24771726e-01
8.41457188e-01 4.93551463e-01 1.36698925e+00 4.27663863e-01
-9.76592898e-02 1.90507546e-01 -1.39969811e-01 -9.04758751e-01
-3.43482375e-01 -2.32769325e-01 -2.94478815e-02 -8.32456052e-01
4.84008342e-01 -2.20799640e-01 -4.01106000e-01 6.60300860e-03
-7.28606164e-01 8.37618649e-01 6.09640718e-01 9.17685807e-01
1.61992759e-01 -5.87647498e-01 5.94865799e-01 -1.04501450e+00
1.26061392e+00 -3.80594730e-01 1.85408462e-02 2.26900637e-01
-1.13036859e+00 3.95913392e-01 5.08273900e-01 -2.25048050e-01
-8.96642923e-01 -3.75859708e-01 2.83761114e-01 -4.53982353e-02
-2.39674911e-01 5.59283614e-01 3.96049231e-01 2.35967889e-01
1.23756540e+00 1.26720101e-01 1.58064827e-01 1.31665580e-02
5.07194221e-01 1.99613422e-01 6.66406453e-01 -4.98365998e-01
8.73225272e-01 -9.11465585e-02 -4.36652035e-01 -4.32132035e-01
-1.09520113e+00 6.93657324e-02 -3.99292141e-01 -1.52844891e-01
8.15402210e-01 -5.70152462e-01 -6.94318593e-01 -3.12252194e-01
-1.40861666e+00 -1.36192292e-01 -5.33314049e-01 2.24140272e-01
-7.90364742e-01 7.53484294e-02 -3.61686975e-01 -4.81116027e-01
-6.37868881e-01 -1.11544263e+00 9.00586009e-01 2.71349400e-01
-1.45652163e+00 -1.05655563e+00 -2.89074451e-01 3.00284415e-01
8.77852440e-01 3.09890658e-01 1.45548022e+00 -1.01105928e+00
2.20318194e-02 1.39673501e-01 -6.09987974e-01 -2.59520292e-01
5.73573308e-03 1.09913079e-02 -1.03466725e+00 4.42977637e-01
-7.03685209e-02 -6.78269982e-01 8.22858751e-01 3.95525485e-01
1.08909261e+00 -9.01382625e-01 -8.74446854e-02 1.64787784e-01
9.79114115e-01 -2.28573710e-01 4.78400797e-01 5.28788030e-01
4.89098608e-01 8.83596957e-01 6.53960645e-01 3.88960987e-01
6.50319815e-01 4.40735996e-01 4.39234287e-01 -3.50956708e-01
-7.38381520e-02 -5.89931548e-01 2.28285015e-01 1.96397439e-01
-2.61010025e-02 8.29261094e-02 -1.05379868e+00 3.79316926e-01
-1.82053638e+00 -1.11932778e+00 -3.06905836e-01 1.73766708e+00
1.04225492e+00 2.72445232e-01 -4.12450125e-03 2.03730136e-01
5.95554054e-01 8.24434385e-02 -5.59917092e-01 -9.89660800e-01
1.00178994e-01 6.45008609e-02 8.26702733e-03 6.69625401e-01
-6.35414362e-01 1.15106523e+00 7.10429907e+00 4.72081691e-01
-1.05031013e+00 -9.38975587e-02 6.74554288e-01 -2.97181308e-02
-1.06954873e+00 -7.69907907e-02 -2.21967265e-01 8.60768110e-02
9.04571950e-01 -5.87126136e-01 3.19410801e-01 9.43376541e-01
2.80519366e-01 -5.45619847e-03 -1.55591595e+00 9.05200541e-01
2.40581900e-01 -1.75327110e+00 6.04692578e-01 -5.66630900e-01
3.44201356e-01 -5.16190946e-01 4.91300672e-01 3.17057133e-01
2.65213758e-01 -1.62625134e+00 1.34160566e+00 4.25628960e-01
7.52498329e-01 -5.53957224e-01 6.68011665e-01 1.51384696e-01
-4.62528020e-01 2.01332316e-01 -3.25661331e-01 -6.95755422e-01
-1.24177270e-01 2.43435010e-01 -1.28917968e+00 -1.53292105e-01
5.10667741e-01 4.99924213e-01 -1.07631099e+00 5.45424879e-01
-6.91273689e-01 5.45155406e-01 2.09905282e-01 -5.89455128e-01
2.73588479e-01 5.69521785e-01 3.41733843e-01 1.32526135e+00
1.88508987e-01 1.47518173e-01 -3.42666984e-01 1.64040208e+00
2.41774887e-01 1.33410692e-01 -5.96661687e-01 -1.46449998e-01
6.46698296e-01 1.07765758e+00 -8.05871725e-01 -6.10315084e-01
1.50201082e-01 7.92421162e-01 5.00884771e-01 2.82344818e-01
-8.79653275e-01 -4.20035958e-01 4.00576144e-01 1.75353333e-01
-2.29875341e-01 2.11011186e-01 -1.06995428e+00 -8.83318007e-01
-3.93019229e-01 -8.89006793e-01 5.79515845e-02 -1.55502331e+00
-1.00119352e+00 6.67743325e-01 1.83212832e-01 -8.37417722e-01
-5.18663347e-01 -6.32176816e-01 -9.73590732e-01 9.86188233e-01
-1.17973948e+00 -9.54938769e-01 -5.09215355e-01 2.45383561e-01
8.05580854e-01 9.89248008e-02 1.10308123e+00 -8.16834509e-01
-2.65238434e-01 5.08608937e-01 -8.68916750e-01 -3.95062685e-01
5.38196921e-01 -1.68208039e+00 7.01632380e-01 5.81608772e-01
1.27817988e-01 1.15486109e+00 1.40062082e+00 -5.15012145e-01
-8.81220698e-01 -8.10041010e-01 1.09926307e+00 -5.81118345e-01
4.61430997e-01 -8.72445107e-02 -1.09986603e+00 6.39495134e-01
5.03094375e-01 -6.09009564e-01 9.68738139e-01 3.42470467e-01
-6.19208038e-01 7.09187806e-01 -1.20871270e+00 1.05254173e+00
8.28277707e-01 -5.21284223e-01 -1.59307301e+00 4.66520607e-01
9.39326286e-01 -9.75905731e-02 -3.78599524e-01 -8.62977058e-02
3.93019885e-01 -1.06971490e+00 8.57240856e-01 -8.47472131e-01
1.01035607e+00 -1.10375270e-01 1.40215665e-01 -1.79212546e+00
-8.10113251e-01 -3.89022231e-01 3.02012593e-01 9.62535143e-01
8.79228354e-01 -6.68917775e-01 3.99917275e-01 1.11643147e+00
-3.56786609e-01 -6.32745683e-01 -5.61489999e-01 -3.80337834e-01
3.47547904e-02 -3.46630156e-01 7.21841216e-01 1.14009130e+00
9.43164706e-01 7.89508760e-01 1.22905605e-01 -3.51363510e-01
3.99204373e-01 1.53962508e-01 4.11382377e-01 -1.43050683e+00
9.63489786e-02 -1.01610029e+00 -2.77709782e-01 -4.48499829e-01
2.79578894e-01 -1.04896140e+00 5.77798039e-02 -1.87052441e+00
5.49923897e-01 4.83562127e-02 1.58751160e-02 1.05070961e+00
-5.59380889e-01 4.49815616e-02 3.83458734e-01 1.05707891e-01
-5.37926733e-01 3.44449669e-01 1.03974807e+00 -1.41460538e-01
-3.47999148e-02 -5.74727356e-01 -1.61934459e+00 1.00429428e+00
8.88699949e-01 -2.33883932e-01 -5.92145503e-01 -4.35649842e-01
4.15381223e-01 -1.76361531e-01 8.71823668e-01 -8.61304522e-01
1.44103527e-01 -5.15196443e-01 7.23619878e-01 -2.66097665e-01
2.11867079e-01 -3.04065049e-01 6.54922100e-03 6.08781159e-01
-1.16970980e+00 4.94931221e-01 4.54328626e-01 1.56852335e-01
-2.01462936e-02 -3.44767898e-01 5.31443655e-01 -2.04158053e-01
-6.79824531e-01 -3.68241161e-01 -5.51725388e-01 1.49499714e-01
6.98745787e-01 -5.80049574e-01 -7.94778228e-01 -7.62502134e-01
-4.36596423e-01 2.11846028e-02 7.26751685e-01 3.80515963e-01
1.14519370e+00 -1.22135806e+00 -7.03911483e-01 -4.19485904e-02
4.27504927e-01 -3.95426542e-01 1.38919726e-01 7.40032256e-01
-4.84943062e-01 6.50826514e-01 -6.45234525e-01 -2.93016046e-01
-1.00608063e+00 6.37700140e-01 -6.00819476e-02 1.71031475e-01
-6.05644226e-01 9.76380765e-01 5.75093150e-01 -1.66253433e-01
1.19688269e-02 -7.25616693e-01 -5.47874451e-01 3.57635841e-02
7.18569100e-01 2.65332013e-01 -1.81491181e-01 -4.43207651e-01
-2.32861310e-01 2.93873459e-01 -1.51931420e-01 -2.32877836e-01
1.08322132e+00 1.98949039e-01 6.07883222e-02 5.67361116e-01
6.35364175e-01 -2.19070464e-01 -9.91119862e-01 2.62132704e-01
1.01497181e-01 -4.09803152e-01 -4.88134660e-02 -1.19098520e+00
-3.76759589e-01 1.17265046e+00 -1.69533882e-02 3.59106004e-01
6.18516088e-01 7.90927410e-02 2.30625898e-01 5.32507598e-01
-2.04294082e-02 -8.74558687e-01 3.21277529e-01 3.41142744e-01
1.50788021e+00 -1.11104178e+00 1.17523126e-01 -4.09569889e-01
-1.23825347e+00 1.22215724e+00 9.34041083e-01 2.35726058e-01
-8.68475884e-02 -1.88256517e-01 -7.70687163e-02 -5.91150641e-01
-1.07540882e+00 1.47843406e-01 5.38094461e-01 7.15027213e-01
8.22852075e-01 3.45759690e-01 -2.13090703e-01 1.03328061e+00
-9.81613219e-01 -4.58471090e-01 9.21840489e-01 5.96692681e-01
-8.48797798e-01 -1.32595394e-02 -4.24242347e-01 7.75867462e-01
-1.04078285e-01 -6.59671485e-01 -8.43927681e-01 7.51009047e-01
-4.88210618e-01 9.82798815e-01 -4.51442078e-02 -4.17670310e-01
1.66754052e-01 2.32131392e-01 1.05684228e-01 -1.01026082e+00
-8.06543827e-01 -6.33211970e-01 3.17894399e-01 -5.57270586e-01
-3.28855477e-02 -5.26778519e-01 -1.67119586e+00 -4.12499875e-01
-2.44099766e-01 2.97438949e-01 5.90886593e-01 1.09872711e+00
3.74680758e-01 7.43964911e-01 -1.69641241e-01 -1.05162072e+00
-7.46830225e-01 -1.10003185e+00 -9.38809738e-02 4.73358840e-01
3.05473924e-01 -8.79081130e-01 -4.90193218e-01 -3.73649672e-02] | [9.576805114746094, 6.74660587310791] |
2b25e920-b784-43ba-b9ac-e5882b66beae | order-planning-neural-text-generation-from | 1709.00155 | null | http://arxiv.org/abs/1709.00155v1 | http://arxiv.org/pdf/1709.00155v1.pdf | Order-Planning Neural Text Generation From Structured Data | Generating texts from structured data (e.g., a table) is important for
various natural language processing tasks such as question answering and dialog
systems. In recent studies, researchers use neural language models and
encoder-decoder frameworks for table-to-text generation. However, these neural
network-based approaches do not model the order of contents during text
generation. When a human writes a summary based on a given table, he or she
would probably consider the content order before wording. In a biography, for
example, the nationality of a person is typically mentioned before occupation
in a biography. In this paper, we propose an order-planning text generation
model to capture the relationship between different fields and use such
relationship to make the generated text more fluent and smooth. We conducted
experiments on the WikiBio dataset and achieve significantly higher performance
than previous methods in terms of BLEU, ROUGE, and NIST scores. | ['Zhifang Sui', 'Sujian Li', 'Lili Mou', 'Pascal Poupart', 'Lei Sha', 'Baobao Chang', 'Tianyu Liu'] | 2017-09-01 | null | null | null | null | ['table-to-text-generation'] | ['natural-language-processing'] | [ 5.20956874e-01 4.07572865e-01 -1.84952766e-01 -5.93917191e-01
-5.70093453e-01 -4.39032823e-01 9.97725606e-01 5.70543528e-01
-3.38668495e-01 1.04417932e+00 9.22066927e-01 -4.18467999e-01
2.15965837e-01 -1.19561327e+00 -7.18589067e-01 1.57639548e-01
7.45791078e-01 5.85448146e-01 -2.20424667e-01 -4.91207212e-01
6.11863375e-01 -1.11176774e-01 -1.18442535e+00 6.53536797e-01
1.13004065e+00 6.62527382e-01 5.16842306e-01 5.12445927e-01
-7.25978732e-01 9.67288435e-01 -1.02943838e+00 -8.04239869e-01
-6.78663179e-02 -8.63396287e-01 -1.07766974e+00 8.87198448e-02
3.78715038e-01 -4.66480345e-01 -5.14358461e-01 1.04230833e+00
1.93527669e-01 2.09033340e-01 8.92225683e-01 -7.72891283e-01
-8.80087614e-01 1.53290296e+00 -1.46852061e-01 -1.81776196e-01
5.88455260e-01 -1.55210257e-01 1.25074232e+00 -5.71791887e-01
7.87134767e-01 1.31613791e+00 1.22498050e-01 6.63794518e-01
-8.51090312e-01 -4.27868038e-01 1.71854988e-01 4.36225981e-02
-9.63231444e-01 -4.47823763e-01 5.66355884e-01 -4.12793726e-01
8.21649969e-01 2.70368040e-01 2.82370269e-01 9.93915439e-01
4.84740317e-01 8.93614233e-01 6.07502997e-01 -5.97000778e-01
7.89474174e-02 1.52888194e-01 1.51397109e-01 4.31199670e-01
4.27420169e-01 -5.30136764e-01 -6.52939796e-01 -6.07622508e-03
5.26403904e-01 -1.10531829e-01 -1.06637791e-01 4.19462860e-01
-1.51055062e+00 9.02537644e-01 2.01728821e-01 2.08599359e-01
-5.93935907e-01 -6.24569468e-02 4.10005957e-01 -6.77673295e-02
2.02278242e-01 8.68198633e-01 -1.25346571e-01 -2.08913118e-01
-9.35152113e-01 5.97773671e-01 1.23265219e+00 1.26136088e+00
4.57384974e-01 -1.35596886e-01 -6.81757569e-01 7.88144827e-01
2.90233314e-01 3.15987557e-01 5.74172556e-01 -7.90881991e-01
9.36815739e-01 7.10481524e-01 3.57945204e-01 -9.81395364e-01
-1.41806111e-01 -1.37378037e-01 -1.17013848e+00 -4.65238720e-01
4.84751344e-01 -4.11957085e-01 -7.69161642e-01 1.46695352e+00
-1.05268165e-01 -5.10642469e-01 3.10201734e-01 6.48231387e-01
1.21588039e+00 1.14689302e+00 -3.67472917e-02 -1.21192507e-01
1.54404497e+00 -8.50136161e-01 -1.08113730e+00 -4.81486320e-01
5.28090954e-01 -7.54847586e-01 9.86318827e-01 1.96957842e-01
-1.07288373e+00 -6.23662412e-01 -9.13240969e-01 -3.11310798e-01
-3.76961172e-01 4.41763610e-01 3.80487382e-01 4.25534308e-01
-6.86319172e-01 6.20855033e-01 -4.50943977e-01 -4.82373744e-01
1.02210939e-02 -1.73220992e-01 -4.03615125e-02 4.46789674e-02
-1.71652758e+00 5.93717217e-01 6.89255595e-01 7.30919838e-02
-3.78042608e-01 -3.74698907e-01 -9.55813169e-01 2.33149275e-01
4.00710076e-01 -8.14778566e-01 1.57437563e+00 -5.04535019e-01
-1.44451344e+00 5.66534042e-01 -3.07925224e-01 -7.09528089e-01
2.93948561e-01 -2.09935322e-01 -3.18212658e-01 -2.73897141e-01
2.04506874e-01 8.91195476e-01 4.18609530e-01 -8.91130388e-01
-7.29556322e-01 -1.87683925e-01 3.59341055e-01 3.43549341e-01
-1.78001955e-01 -7.98597783e-02 -3.41230601e-01 -6.64856553e-01
-1.12602375e-01 -6.92498326e-01 -2.46655524e-01 -4.22489852e-01
-7.91279852e-01 -5.23956537e-01 2.99245596e-01 -1.01142895e+00
1.64011157e+00 -1.59964967e+00 -1.14803515e-01 -5.81982061e-02
-1.49056464e-02 2.02979948e-02 -4.01864909e-02 9.14662421e-01
4.23616022e-01 4.84991133e-01 -1.30756229e-01 1.29159033e-01
3.13005298e-01 -6.49110600e-02 -5.03757060e-01 -3.53845239e-01
5.75963641e-03 8.32922935e-01 -8.14188719e-01 -4.89911169e-01
-2.62570828e-01 1.50011420e-01 -5.15482605e-01 2.55134672e-01
-7.26334035e-01 -6.42967299e-02 -4.60898370e-01 2.48575345e-01
7.63662308e-02 -3.81389022e-01 2.79941201e-01 -1.44592389e-01
2.82213297e-02 9.85820234e-01 -8.65216851e-01 1.67465913e+00
-8.38865697e-01 7.83543646e-01 -3.51925641e-01 -4.00031418e-01
1.10453212e+00 3.47122610e-01 -7.17535168e-02 -6.66805744e-01
-3.48611400e-02 5.08198217e-02 1.31901085e-01 -3.27249587e-01
1.35608304e+00 1.55346066e-01 -2.78644204e-01 7.45835662e-01
-4.13298905e-01 -2.64111698e-01 8.09583366e-01 5.70749640e-01
9.20570195e-01 9.20877904e-02 5.22313356e-01 -4.39281166e-02
5.88771880e-01 8.39477852e-02 2.64850110e-01 9.47952032e-01
4.31417942e-01 5.74784219e-01 6.80952907e-01 -2.62883812e-01
-1.22796094e+00 -6.32989347e-01 2.65118778e-01 8.36195827e-01
-5.93357757e-02 -8.01407278e-01 -9.29423034e-01 -5.90096116e-01
-2.48403862e-01 1.38597703e+00 -3.24147552e-01 -4.38721552e-02
-6.41194701e-01 -2.49996185e-01 5.76724589e-01 3.70929629e-01
4.97331798e-01 -1.56755781e+00 -3.97857100e-01 6.15904093e-01
-7.80874789e-01 -1.09407938e+00 -8.62269938e-01 -2.38427415e-01
-7.02720106e-01 -5.57765007e-01 -7.18497038e-01 -8.59541953e-01
7.26999402e-01 -1.47598326e-01 1.23101497e+00 -2.93566305e-02
2.63800561e-01 -2.01271266e-01 -5.30773640e-01 -5.65435410e-01
-8.49931061e-01 5.97777903e-01 -2.92511135e-01 -1.31859362e-01
3.62809032e-01 -8.59459266e-02 -2.29535148e-01 -1.53397433e-02
-1.17169821e+00 5.16760528e-01 7.24598587e-01 9.17118967e-01
3.20859134e-01 -3.13799605e-02 6.51510239e-01 -1.09263277e+00
1.40726054e+00 -3.24078083e-01 -5.54950893e-01 5.33498943e-01
-6.94659948e-01 5.29877961e-01 9.53185797e-01 -2.26222977e-01
-1.22947776e+00 -1.35689080e-01 -1.44835711e-01 3.85103941e-01
-1.42959386e-01 9.27403748e-01 -2.79775172e-01 8.57021213e-01
7.54544914e-01 5.86803317e-01 -1.81081519e-01 -2.55856752e-01
3.23109865e-01 1.06266570e+00 3.97348166e-01 -6.02515221e-01
6.09846413e-01 -1.65615857e-01 -3.95270735e-01 -6.89193308e-01
-8.16902518e-01 -2.48443678e-01 -3.89934689e-01 -1.98927075e-01
7.50551760e-01 -6.58908844e-01 -5.59339106e-01 2.03176901e-01
-1.72672582e+00 -2.69231528e-01 -9.79110412e-03 3.23623091e-01
-3.14878732e-01 3.36235166e-01 -5.54240227e-01 -6.24158561e-01
-7.91997075e-01 -9.71527696e-01 9.01505768e-01 5.49300730e-01
-7.63931036e-01 -8.88520420e-01 -1.84026286e-01 4.93215382e-01
3.04163069e-01 -1.54365534e-02 1.27545249e+00 -8.19032252e-01
-4.98242140e-01 -2.11740568e-01 -1.77145243e-01 6.00220598e-02
3.40692818e-01 -7.39423260e-02 -2.76737928e-01 2.98386961e-01
-4.23811078e-01 -2.06821904e-01 5.93993068e-01 3.39018293e-02
1.19705367e+00 -8.99505436e-01 -1.08448111e-01 8.63890201e-02
1.01325083e+00 6.19288564e-01 8.84528041e-01 2.31785372e-01
5.65608442e-01 1.00254452e+00 4.95619744e-01 5.55919230e-01
7.81791866e-01 5.07286906e-01 -1.21097593e-02 3.88008744e-01
1.85739361e-02 -7.11578071e-01 3.70076865e-01 9.80538785e-01
2.95113951e-01 -9.15646493e-01 -8.27973545e-01 6.14531219e-01
-1.58198261e+00 -1.06343281e+00 -1.69476256e-01 1.91395724e+00
1.18986154e+00 2.44525850e-01 -3.58166933e-01 -1.79390326e-01
7.44377673e-01 3.45245272e-01 -3.73641819e-01 -5.71311891e-01
8.00651535e-02 -8.24115798e-02 3.27479571e-01 6.09335959e-01
-6.16716087e-01 1.08744025e+00 5.66647530e+00 6.48626983e-01
-7.04146743e-01 -3.69071722e-01 8.35565090e-01 1.92292243e-01
-6.23664021e-01 -6.53925389e-02 -1.21795535e+00 4.78618592e-01
9.46415067e-01 -8.17831337e-01 4.59523082e-01 6.65260971e-01
2.60118216e-01 -1.06600977e-01 -1.27159011e+00 8.79125237e-01
3.37502778e-01 -1.55280948e+00 5.72470129e-01 7.30895251e-02
6.48087859e-01 -5.95784128e-01 -2.98487544e-01 4.31794673e-01
3.88633847e-01 -1.14412785e+00 8.40777874e-01 5.52209139e-01
5.23015320e-01 -6.17959499e-01 7.37263560e-01 5.27657866e-01
-7.89669275e-01 3.82669121e-01 -4.40890133e-01 -1.31115675e-01
4.77376103e-01 5.86329103e-01 -1.19282603e+00 5.08885264e-01
4.00205962e-02 3.26420277e-01 -3.30596924e-01 6.13523304e-01
-5.29276490e-01 5.80961049e-01 7.15151206e-02 -8.47348154e-01
3.15755785e-01 -3.03597122e-01 2.34050974e-01 1.11057305e+00
5.42817533e-01 2.23408461e-01 6.09422196e-03 1.08102608e+00
-7.28706956e-01 4.33165729e-01 -6.03477001e-01 -5.99092841e-01
4.70631003e-01 1.14974701e+00 -5.83206356e-01 -6.33793592e-01
-1.99044675e-01 9.35420752e-01 -5.01527935e-02 2.63735592e-01
-5.54138422e-01 -1.00866151e+00 3.70678097e-01 3.78378153e-01
-4.83115157e-03 -1.67521432e-01 -5.88939726e-01 -1.03170681e+00
2.01312140e-01 -1.18803573e+00 -1.81933809e-02 -1.05488443e+00
-9.16742802e-01 8.67830336e-01 -1.45614464e-02 -9.82484877e-01
-8.31319392e-01 -2.05348819e-01 -4.84870791e-01 1.06911516e+00
-9.34804916e-01 -6.16477787e-01 -2.69569725e-01 -1.01106592e-01
1.00934029e+00 -2.65534341e-01 6.35371506e-01 5.70794451e-04
-2.62003660e-01 3.80664289e-01 -1.21126451e-01 5.78691185e-01
7.19423831e-01 -1.15215409e+00 1.07828486e+00 1.05891895e+00
4.14797366e-01 9.49708283e-01 6.55781984e-01 -1.09454679e+00
-1.10514808e+00 -1.08069897e+00 1.71644437e+00 -1.66762978e-01
5.09659290e-01 -4.67792809e-01 -8.21757078e-01 4.97032791e-01
7.36875236e-01 -1.03974068e+00 6.14434361e-01 -1.05639538e-02
6.60438463e-02 -4.87466641e-02 -7.00690150e-01 9.43646789e-01
8.13466668e-01 -4.57357109e-01 -7.64620543e-01 5.91500878e-01
8.37096334e-01 -6.38682008e-01 -6.80867255e-01 -9.55037624e-02
5.33818364e-01 -5.58291495e-01 3.84262323e-01 -5.27290940e-01
1.11560559e+00 -2.40118116e-01 8.85164514e-02 -1.58216488e+00
-1.24763101e-01 -7.47565508e-01 -1.04893506e-01 1.36076152e+00
8.72153878e-01 -2.72808343e-01 8.40669572e-01 7.95235395e-01
-1.26218334e-01 -4.89259899e-01 -3.25191230e-01 -4.17444557e-01
-6.86547831e-02 -1.06127150e-01 1.04273021e+00 5.78471780e-01
1.67857245e-01 8.21463346e-01 -5.38969338e-01 -3.24500740e-01
1.69093981e-01 4.49587777e-02 8.48444819e-01 -1.12595725e+00
-6.87022731e-02 -5.07254004e-01 1.69183850e-01 -1.46887469e+00
1.74367815e-01 -9.77872968e-01 3.66012573e-01 -2.34084129e+00
4.42140698e-02 -1.40611053e-01 4.18166637e-01 3.12345237e-01
-3.00721973e-01 -4.23716486e-01 2.87886530e-01 -5.16507253e-02
-3.28027844e-01 5.00171602e-01 1.28670859e+00 -4.58023548e-01
-2.49232933e-01 8.49090293e-02 -1.09694195e+00 3.81552517e-01
8.96714985e-01 -4.96485025e-01 -3.94973874e-01 -5.77868998e-01
5.51635742e-01 6.06016994e-01 -1.89299971e-01 -8.82767439e-01
5.56710482e-01 -4.21071678e-01 3.11359316e-01 -7.08093166e-01
-7.51279108e-03 -3.77769351e-01 -1.37450667e-02 3.84300530e-01
-1.00624394e+00 4.09767926e-01 -1.49080157e-01 1.30666748e-01
-4.33056504e-01 -5.81748545e-01 2.51633316e-01 -3.54867101e-01
-4.36576545e-01 1.95809454e-01 -7.17747569e-01 1.90090328e-01
4.09484446e-01 -3.82353477e-02 -3.77571791e-01 -1.02477753e+00
-7.31551349e-02 2.75106877e-01 2.05647692e-01 6.98180676e-01
7.30312884e-01 -1.32641065e+00 -9.87915337e-01 -4.66785990e-02
1.83541715e-01 8.55291709e-02 -2.09298357e-03 1.56323373e-01
-8.51930559e-01 8.80779564e-01 -2.03220382e-01 9.00355056e-02
-1.04414618e+00 1.96920112e-02 8.46224558e-03 -7.68871784e-01
-2.48542845e-01 4.55427200e-01 1.20006442e-01 -5.32034039e-01
3.19113582e-01 -6.02671206e-01 -5.33188522e-01 1.47911444e-01
9.21436489e-01 1.62369266e-01 4.64602141e-03 -3.51100892e-01
1.23281650e-01 -1.20963641e-01 -4.51451927e-01 -5.35408735e-01
9.39295292e-01 5.39700836e-02 -1.47336364e-01 3.94272387e-01
7.07468927e-01 8.31997097e-02 -7.31600165e-01 -2.96183378e-01
2.23612174e-01 -2.41035987e-02 -3.17691654e-01 -9.71553266e-01
-6.93185091e-01 8.20502758e-01 -3.18677723e-01 3.18684161e-01
7.86826611e-01 -2.10386872e-01 1.18678522e+00 9.55461562e-01
7.43237138e-02 -1.19610858e+00 -3.05271968e-02 1.06434405e+00
1.01781225e+00 -9.62720394e-01 -2.63685193e-02 -1.47086099e-01
-9.76067662e-01 1.29123354e+00 8.68666053e-01 3.95976186e-01
7.82889575e-02 5.06138615e-02 7.52297789e-02 7.69241378e-02
-9.46650922e-01 6.96038827e-02 4.22322452e-01 4.60699022e-01
8.70025218e-01 5.93643002e-02 -6.76883101e-01 5.30572355e-01
-7.70438969e-01 -3.69231403e-02 1.00622416e+00 7.12550044e-01
-5.59244096e-01 -1.27627170e+00 -3.43935460e-01 7.40965843e-01
-6.22295618e-01 -4.74739313e-01 -7.42821753e-01 3.49456549e-01
-3.26625526e-01 1.05995762e+00 1.33621007e-01 -3.92408580e-01
3.02707940e-01 1.82081789e-01 1.33477569e-01 -9.25974309e-01
-8.24161768e-01 -2.79361755e-01 4.62485582e-01 8.60897824e-02
-8.08856338e-02 -5.31646490e-01 -1.35955811e+00 -4.12027478e-01
-2.70169359e-02 5.18982172e-01 7.32498407e-01 8.72600555e-01
2.31140122e-01 5.86834967e-01 3.12097251e-01 -1.66002110e-01
-4.37211394e-01 -1.41962028e+00 -3.38420898e-01 3.56126904e-01
-1.57622248e-01 1.00979589e-01 1.05423048e-01 4.35469031e-01] | [11.815862655639648, 8.88070011138916] |
d621dc8e-c9ab-4add-b52b-3e61784ea0b7 | short-and-long-range-relation-based-spatio | 2112.05851 | null | https://arxiv.org/abs/2112.05851v4 | https://arxiv.org/pdf/2112.05851v4.pdf | Short and Long Range Relation Based Spatio-Temporal Transformer for Micro-Expression Recognition | Being spontaneous, micro-expressions are useful in the inference of a person's true emotions even if an attempt is made to conceal them. Due to their short duration and low intensity, the recognition of micro-expressions is a difficult task in affective computing. The early work based on handcrafted spatio-temporal features which showed some promise, has recently been superseded by different deep learning approaches which now compete for the state of the art performance. Nevertheless, the problem of capturing both local and global spatio-temporal patterns remains challenging. To this end, herein we propose a novel spatio-temporal transformer architecture -- to the best of our knowledge, the first purely transformer based approach (i.e. void of any convolutional network use) for micro-expression recognition. The architecture comprises a spatial encoder which learns spatial patterns, a temporal aggregator for temporal dimension analysis, and a classification head. A comprehensive evaluation on three widely used spontaneous micro-expression data sets, namely SMIC-HS, CASME II and SAMM, shows that the proposed approach consistently outperforms the state of the art, and is the first framework in the published literature on micro-expression recognition to achieve the unweighted F1-score greater than 0.9 on any of the aforementioned data sets. | ['Guoying Zhao', 'Ognjen Arandjelovic', 'Xiaopeng Hong', 'Liangfei Zhang'] | 2021-12-10 | null | null | null | null | ['micro-expression-recognition'] | ['computer-vision'] | [ 8.74582753e-02 -2.16169104e-01 -2.00913008e-02 -4.88529563e-01
-6.24616146e-01 -4.60081100e-01 8.87575686e-01 1.37799203e-01
-4.09883082e-01 8.54030132e-01 7.25256186e-03 1.16697058e-01
-3.20016116e-01 -5.13543069e-01 -3.71676385e-01 -1.01965630e+00
-4.01139677e-01 2.05619261e-01 -2.12789804e-01 -4.35306817e-01
1.37003645e-01 6.29281342e-01 -1.83415866e+00 5.27283192e-01
4.42989588e-01 1.64365208e+00 -3.62874240e-01 4.56541032e-01
-2.14874059e-01 1.48072839e+00 -5.35114050e-01 -7.14475870e-01
-3.08133036e-01 -4.65720922e-01 -8.65360260e-01 -1.09424815e-01
8.00962746e-03 1.06608227e-01 5.08323833e-02 6.95054233e-01
5.31124055e-01 1.68418884e-01 5.14446855e-01 -1.18505013e+00
-5.77103555e-01 1.09204017e-01 -4.51040238e-01 2.62722522e-01
6.03566706e-01 -3.06072265e-01 1.02997351e+00 -8.77226710e-01
6.21819258e-01 6.42187655e-01 7.68359721e-01 2.18624368e-01
-9.13530290e-01 -5.26942551e-01 -9.32826251e-02 5.14174581e-01
-1.51669228e+00 -4.91370410e-01 9.40104723e-01 -4.95045811e-01
1.13098598e+00 2.60671943e-01 7.76834548e-01 1.51058722e+00
1.81901097e-01 9.08934712e-01 1.50021982e+00 -3.99231613e-01
2.54129022e-01 2.61322856e-01 -1.71895549e-01 5.76956809e-01
-6.91930830e-01 -4.18031365e-02 -7.60764658e-01 4.70275134e-02
2.79063612e-01 -3.65783870e-02 8.71818587e-02 -3.07841301e-01
-9.59799051e-01 5.55279016e-01 2.20355868e-01 9.89877641e-01
-6.03602648e-01 -7.46363997e-02 7.61789382e-01 4.91529882e-01
8.94177258e-01 2.79533118e-01 -4.53339398e-01 -9.10277307e-01
-1.15907574e+00 1.57363772e-01 5.54858804e-01 4.19092625e-01
5.90484619e-01 7.86129087e-02 -1.67725489e-01 8.82044852e-01
-2.77254254e-01 1.02801271e-01 6.36781275e-01 -4.40204412e-01
6.01917282e-02 7.87238538e-01 1.45937771e-01 -1.47257626e+00
-5.81361890e-01 -5.80551624e-01 -1.03158009e+00 1.43774763e-01
3.95849586e-01 -1.91848844e-01 -3.70206207e-01 1.88889611e+00
2.50108600e-01 4.58375454e-01 1.50848366e-02 8.56846929e-01
4.94475126e-01 5.55769205e-01 4.19471599e-03 -3.97100657e-01
1.11214244e+00 -7.31303453e-01 -1.02964473e+00 1.35938093e-01
7.70083845e-01 -5.52316070e-01 7.85956800e-01 5.18930733e-01
-9.21796024e-01 -4.03967738e-01 -9.50423121e-01 4.87456508e-02
-7.86463797e-01 2.20663503e-01 9.23819959e-01 3.34395736e-01
-1.17872334e+00 5.42766929e-01 -7.34454155e-01 -3.75399739e-01
3.62319529e-01 4.56623167e-01 -8.03285062e-01 4.78284836e-01
-1.38832700e+00 1.12660074e+00 -3.52096707e-02 4.23632652e-01
-3.76883119e-01 -5.37839532e-01 -6.97570443e-01 -1.62558675e-01
1.05759673e-01 -1.99589670e-01 1.16859102e+00 -1.56564760e+00
-1.75203264e+00 1.15949357e+00 -2.76466191e-01 -6.44206047e-01
4.63089794e-01 -1.81418285e-01 -7.66679645e-01 1.82061449e-01
-1.03455134e-01 2.03180864e-01 7.57182300e-01 -8.35092068e-01
-4.32493120e-01 -6.74457967e-01 6.61419630e-02 3.45286317e-02
-4.81220603e-01 2.89930016e-01 -7.92866871e-02 -5.63472450e-01
-2.78822899e-01 -9.47774231e-01 4.58300151e-02 -2.10137650e-01
-2.53021717e-01 -3.63962620e-01 8.49655747e-01 -2.73395628e-01
1.43350947e+00 -2.29584742e+00 3.66516858e-01 1.76851854e-01
-1.69927217e-02 2.73279428e-01 2.01981217e-01 5.45518160e-01
-3.86692017e-01 -2.30629742e-01 -1.40797049e-01 -6.43132746e-01
1.29720792e-01 1.76055565e-01 -3.10188293e-01 7.64346540e-01
1.16849191e-01 1.06767571e+00 -8.81757259e-01 -4.12704229e-01
4.65841502e-01 6.56822562e-01 4.36587222e-02 2.07737386e-01
1.76721498e-01 3.56166273e-01 -2.44427145e-01 5.61008513e-01
4.35273588e-01 -1.11524068e-01 1.12655565e-01 -5.30439653e-02
-4.27916259e-01 -3.48280817e-02 -7.52449095e-01 1.76037133e+00
-6.73362076e-01 8.23661625e-01 2.73771398e-02 -1.27203202e+00
1.00810492e+00 7.18265116e-01 1.12208855e+00 -1.15701962e+00
1.58099800e-01 3.21645796e-01 -3.86084467e-01 -5.48747003e-01
3.93327564e-01 -4.48761851e-01 -3.22836548e-01 3.22593689e-01
1.54123306e-01 3.93701583e-01 -3.34751047e-02 -1.23671137e-01
9.95674431e-01 6.50262460e-02 4.42638367e-01 -1.60109818e-01
6.56050324e-01 -2.73319036e-01 4.06437814e-01 2.09271505e-01
-3.21661890e-01 3.30212623e-01 5.94163060e-01 -7.95268595e-01
-8.33722711e-01 -7.76021898e-01 -2.08773255e-01 1.19745207e+00
-1.45957857e-01 -5.42935729e-01 -7.32286751e-01 -4.94509012e-01
-2.87098616e-01 4.38408554e-01 -1.16000330e+00 8.03418159e-02
-3.37107092e-01 -6.37772739e-01 8.97182941e-01 6.05875909e-01
4.43597466e-01 -1.29810023e+00 -8.95568967e-01 2.76982754e-01
-2.58742154e-01 -1.30218363e+00 1.19896024e-01 2.53445178e-01
-3.47970814e-01 -7.19268203e-01 -6.54531062e-01 -4.96988863e-01
8.69382471e-02 -2.98627973e-01 1.13495040e+00 -3.07987124e-01
-2.51224309e-01 2.00091422e-01 -5.01800060e-01 -4.64205176e-01
2.62681037e-01 5.25631681e-02 4.87111770e-02 6.28730178e-01
7.37843573e-01 -1.00444448e+00 -5.26683331e-01 1.03237033e-01
-7.41266072e-01 -1.70699567e-01 4.88467336e-01 7.20581353e-01
5.82300961e-01 -3.80718261e-02 4.81189758e-01 -6.62638605e-01
5.85797071e-01 -5.64413369e-01 -1.43040910e-01 5.83643876e-02
-1.74186885e-01 -2.18852967e-01 7.76001275e-01 -2.33719841e-01
-1.05669749e+00 1.41309097e-01 -3.34209204e-01 -4.79731143e-01
-3.51377934e-01 6.09859467e-01 2.38939270e-01 -6.25030994e-02
4.48058218e-01 5.40742755e-01 -1.00682773e-01 -3.59256595e-01
1.77877858e-01 5.70743501e-01 5.88014603e-01 -3.67544293e-01
2.74510354e-01 6.95010781e-01 1.31763101e-01 -9.70496476e-01
-9.45894003e-01 -5.82935154e-01 -8.92868221e-01 -2.72829950e-01
1.03128946e+00 -8.39820683e-01 -7.69748628e-01 6.84832871e-01
-1.08022249e+00 -1.71642944e-01 -2.27946028e-01 1.06321976e-01
-7.34153092e-01 1.34105682e-01 -5.53525686e-01 -1.15259290e+00
-2.45225430e-01 -7.75712371e-01 1.23619306e+00 -1.15164399e-01
-6.99118197e-01 -1.15267169e+00 2.50827640e-01 6.68885410e-02
6.48172915e-01 8.59948575e-01 4.20939356e-01 -4.57075715e-01
1.06743108e-02 -3.34777445e-01 3.06051392e-02 1.47792652e-01
1.26501456e-01 1.47176581e-02 -1.23120952e+00 1.12331003e-01
5.81971928e-02 -5.95530510e-01 5.51761508e-01 2.05215484e-01
1.20640814e+00 -3.44774246e-01 -3.92104089e-02 5.84908426e-01
1.20931995e+00 3.04912657e-01 7.35568881e-01 4.17907894e-01
2.72112638e-01 6.92615092e-01 6.07230663e-01 9.57571685e-01
3.97151560e-01 1.21731091e+00 3.68501872e-01 -2.40870073e-01
4.57079798e-01 7.41285756e-02 3.34356219e-01 7.49539018e-01
-3.51284236e-01 -3.53324190e-02 -7.80919135e-01 7.13709831e-01
-2.02489138e+00 -1.41076279e+00 1.02936946e-01 1.75312471e+00
5.77131510e-01 -1.33907795e-01 1.91350907e-01 5.33727109e-01
3.75546105e-02 5.98630309e-01 -1.92097515e-01 -9.45966482e-01
-3.47158700e-01 4.96658981e-01 -1.44872829e-01 1.98567942e-01
-1.41954172e+00 7.64779389e-01 5.60332060e+00 7.68611073e-01
-1.50447261e+00 4.51417230e-02 6.69573784e-01 -2.49130696e-01
8.16860273e-02 -5.58418930e-01 -2.38826618e-01 5.06615758e-01
1.18221259e+00 -2.44882610e-02 2.79915273e-01 8.45577359e-01
1.86828732e-01 6.36158884e-02 -9.68480527e-01 1.35871768e+00
1.39918536e-01 -1.18067563e+00 -3.44219863e-01 -1.66619346e-01
4.91196722e-01 -1.91651881e-01 3.78624141e-01 2.41092712e-01
-2.04759657e-01 -1.37328207e+00 8.06765556e-01 8.03536355e-01
8.41096282e-01 -1.05536437e+00 1.03849864e+00 2.58070290e-01
-1.33501625e+00 -1.13831423e-01 1.23510793e-01 -4.59961474e-01
2.01006398e-01 6.87831521e-01 -2.82007903e-01 5.57488561e-01
7.18881905e-01 1.03006184e+00 -3.28503460e-01 3.88276339e-01
-4.04885225e-02 3.96232873e-01 -2.49983996e-01 -2.11518377e-01
6.92558467e-01 -1.50506958e-01 2.44769990e-01 1.63774455e+00
4.21519786e-01 1.53380349e-01 -2.05606773e-01 4.23053831e-01
1.78811535e-01 3.80124241e-01 -8.14168036e-01 -7.06016943e-02
4.89603952e-02 1.28618717e+00 -2.74571180e-01 -1.79941937e-01
-4.42518234e-01 1.30992985e+00 5.42242467e-01 1.93133384e-01
-1.00977635e+00 -2.65759140e-01 8.52897465e-01 -9.72755030e-02
2.24603966e-01 -1.65267691e-01 -3.19162011e-01 -1.09920490e+00
3.00057501e-01 -9.02939916e-01 4.17023927e-01 -7.99433768e-01
-1.18107092e+00 1.01175642e+00 -2.94004440e-01 -1.07793462e+00
-7.28814721e-01 -6.10745490e-01 -4.23769385e-01 6.40602052e-01
-1.27894008e+00 -1.22704339e+00 -2.69204259e-01 9.97915447e-01
3.44554394e-01 -1.42106339e-01 1.33676469e+00 4.38678026e-01
-4.95531917e-01 6.30697906e-01 9.31631923e-02 1.14048354e-01
4.77714062e-01 -1.25002456e+00 -2.71520406e-01 4.77607667e-01
3.14994484e-01 4.94071245e-01 7.52988219e-01 1.37355179e-01
-1.35176659e+00 -8.09232891e-01 1.29083169e+00 -5.69624007e-01
8.48826170e-01 -5.10487676e-01 -7.34320700e-01 5.31687677e-01
3.18239659e-01 3.60735863e-01 9.95771766e-01 2.52984226e-01
-3.00115794e-01 -3.48122030e-01 -9.93570507e-01 3.82169574e-01
8.13211977e-01 -9.62507129e-01 -3.97611916e-01 -6.65669516e-02
1.44360987e-02 -2.08621114e-01 -1.12211108e+00 5.04531860e-01
9.30080473e-01 -1.50383663e+00 9.07615244e-01 -5.16012967e-01
6.76138103e-01 -5.08956909e-02 -3.32012057e-01 -1.22480726e+00
-1.58609971e-01 -5.77581048e-01 -2.31909409e-01 1.15979457e+00
1.37441888e-01 -2.73176163e-01 8.11729789e-01 5.20286620e-01
9.87176076e-02 -1.26183569e+00 -1.33698499e+00 -4.68179226e-01
-2.19923526e-01 -7.06404567e-01 5.18805444e-01 1.18518329e+00
4.12680387e-01 3.83361846e-01 -7.42275596e-01 -2.59862036e-01
2.15689987e-01 4.24555928e-01 6.52415693e-01 -9.28551555e-01
-5.14900461e-02 -5.57893157e-01 -8.74152899e-01 -6.76180720e-01
3.54468167e-01 -5.31488001e-01 -2.39894807e-01 -1.02725983e+00
-5.31659350e-02 -9.35765058e-02 -6.23227060e-01 4.58100259e-01
2.91001529e-01 4.41371650e-01 -4.85621318e-02 -6.94445372e-02
-8.70575070e-01 6.80489004e-01 9.08750057e-01 -4.74121012e-02
-4.18232568e-02 -1.48360357e-01 -4.25580978e-01 7.19039679e-01
7.43099928e-01 -1.05767176e-01 -4.91640151e-01 -5.59970625e-02
3.86240691e-01 9.33684334e-02 3.52368802e-01 -1.04674006e+00
2.18179256e-01 -4.62656543e-02 3.73805970e-01 -4.55929250e-01
7.37884164e-01 -8.66244793e-01 1.66204467e-01 -1.00742139e-01
-3.11159462e-01 4.05175775e-01 2.59481639e-01 3.04284811e-01
-8.32262993e-01 3.80939633e-01 7.21988022e-01 -1.15073770e-01
-1.07443953e+00 3.31090599e-01 -5.21062493e-01 -1.53615564e-01
1.25803185e+00 -2.90628314e-01 6.39721453e-02 -4.99621898e-01
-8.23788643e-01 -3.09197009e-01 8.84984955e-02 4.31356847e-01
5.61968029e-01 -1.42614484e+00 -5.55803359e-01 8.84589776e-02
3.16775471e-01 -5.22154093e-01 5.56697488e-01 1.30359757e+00
-9.76817682e-02 4.99601036e-01 -3.85895967e-01 -5.77628613e-01
-1.31070757e+00 5.79495907e-01 4.42957819e-01 -5.70320547e-01
-4.61236805e-01 7.62611330e-01 -1.95157621e-02 -1.95407286e-01
7.76737854e-02 -8.22186992e-02 -4.66535687e-01 4.68690485e-01
5.89212537e-01 2.24715412e-01 2.74693877e-01 -1.13042641e+00
-6.76592708e-01 5.78625202e-01 2.35289127e-01 -9.79475299e-05
1.55680978e+00 -9.81514975e-02 -4.06188071e-01 9.18657184e-01
1.61501014e+00 -1.64861791e-02 -7.75743783e-01 -1.53972238e-01
3.04266989e-01 -3.26701611e-01 -7.72005916e-02 -8.32201123e-01
-8.94796550e-01 1.09531295e+00 5.97538888e-01 3.50861549e-01
1.50044310e+00 -1.25381187e-01 6.78842723e-01 2.82030195e-01
5.21239161e-01 -1.13803315e+00 1.78448096e-01 7.15302587e-01
8.25010300e-01 -1.08708405e+00 -3.72213870e-01 5.88416979e-02
-9.84273612e-01 1.04553032e+00 3.61218899e-01 -7.79544935e-02
6.78190351e-01 5.61477661e-01 2.27703184e-01 -4.15732265e-01
-9.03596282e-01 -1.83033615e-01 2.28694737e-01 2.87608087e-01
7.79969990e-01 1.44615293e-01 -1.34706214e-01 6.99630797e-01
-2.89032966e-01 3.36612493e-01 -1.20405778e-01 7.29021311e-01
3.15958783e-02 -8.46285164e-01 7.24815875e-02 1.91168264e-01
-9.46994424e-01 1.13078833e-01 -5.93767285e-01 7.18755543e-01
2.94694215e-01 7.33739078e-01 4.10692133e-02 -6.28873408e-01
4.43874329e-01 2.64265478e-01 3.93945009e-01 5.52395545e-02
-7.03051865e-01 -3.29860657e-01 3.01756680e-01 -9.17736173e-01
-1.06502473e+00 -6.41962230e-01 -9.02886450e-01 -3.88272494e-01
2.78088599e-01 3.32457423e-01 4.95181710e-01 1.08730674e+00
3.54435205e-01 3.14193189e-01 7.12122023e-01 -7.60409713e-01
-5.99597059e-02 -8.55822265e-01 -6.74222052e-01 9.13660228e-01
4.35458422e-01 -8.32734466e-01 -3.34685296e-02 -1.41586348e-01] | [13.632964134216309, 1.9518814086914062] |
a205e753-44f2-4655-8a23-410fe6cde940 | npvforensics-jointing-non-critical-phonemes | 2306.06885 | null | https://arxiv.org/abs/2306.06885v1 | https://arxiv.org/pdf/2306.06885v1.pdf | NPVForensics: Jointing Non-critical Phonemes and Visemes for Deepfake Detection | Deepfake technologies empowered by deep learning are rapidly evolving, creating new security concerns for society. Existing multimodal detection methods usually capture audio-visual inconsistencies to expose Deepfake videos. More seriously, the advanced Deepfake technology realizes the audio-visual calibration of the critical phoneme-viseme regions, achieving a more realistic tampering effect, which brings new challenges. To address this problem, we propose a novel Deepfake detection method to mine the correlation between Non-critical Phonemes and Visemes, termed NPVForensics. Firstly, we propose the Local Feature Aggregation block with Swin Transformer (LFA-ST) to construct non-critical phoneme-viseme and corresponding facial feature streams effectively. Secondly, we design a loss function for the fine-grained motion of the talking face to measure the evolutionary consistency of non-critical phoneme-viseme. Next, we design a phoneme-viseme awareness module for cross-modal feature fusion and representation alignment, so that the modality gap can be reduced and the intrinsic complementarity of the two modalities can be better explored. Finally, a self-supervised pre-training strategy is leveraged to thoroughly learn the audio-visual correspondences in natural videos. In this manner, our model can be easily adapted to the downstream Deepfake datasets with fine-tuning. Extensive experiments on existing benchmarks demonstrate that the proposed approach outperforms state-of-the-art methods. | ['Haoliang Li', 'Yao Zhao', 'Rongrong Ni', 'Yang Yu', 'Yu Chen'] | 2023-06-12 | null | null | null | null | ['deepfake-detection', 'face-swapping'] | ['computer-vision', 'computer-vision'] | [-1.34933427e-01 -3.89777333e-01 -2.31866807e-01 -2.97508478e-01
-8.16982388e-01 -4.77022856e-01 5.52504420e-01 -5.30579567e-01
-7.63385184e-03 1.87244847e-01 4.11455214e-01 2.64235198e-01
-2.97648013e-01 -4.91137415e-01 -6.93417609e-01 -1.04674566e+00
5.50356470e-02 -1.76096529e-01 -1.59899563e-01 -3.13568056e-01
1.87903717e-01 2.55736053e-01 -1.87280154e+00 4.17830437e-01
8.33154440e-01 1.35273409e+00 -7.07106944e-03 5.49109459e-01
-1.55704813e-02 4.74292994e-01 -4.03890133e-01 -7.84004986e-01
4.23287973e-02 -5.88054299e-01 -5.41480124e-01 1.62005633e-01
7.44825006e-01 -3.70799512e-01 -4.65327293e-01 1.25060833e+00
8.42390776e-01 -2.13160202e-01 4.51566637e-01 -1.71493328e+00
-2.25015014e-01 4.60355699e-01 -8.39287877e-01 4.23913449e-02
2.55149245e-01 3.40797514e-01 7.49976695e-01 -1.12665689e+00
3.58194292e-01 1.58541250e+00 6.45510912e-01 4.19333190e-01
-7.29790092e-01 -1.05604947e+00 4.86370549e-02 8.52748871e-01
-1.66643834e+00 -8.68501723e-01 1.03715146e+00 -3.55964363e-01
5.55797696e-01 3.68763953e-01 7.11686552e-01 1.22921312e+00
2.15051904e-01 6.37070894e-01 6.17275953e-01 -1.73539087e-01
-2.36791983e-01 -4.68100607e-03 -4.02639717e-01 9.97364044e-01
-2.91971713e-01 2.39742234e-01 -9.61649895e-01 1.48883358e-01
3.52831036e-01 -1.79594010e-01 -4.14764196e-01 -3.68312389e-01
-1.20327914e+00 6.26074195e-01 2.69324541e-01 2.80077100e-01
-7.88976029e-02 3.39068647e-04 8.12255859e-01 3.41971964e-01
4.46954146e-02 -2.57768538e-02 -2.59162277e-01 -2.79653788e-01
-8.78358603e-01 -2.23760501e-01 2.33457446e-01 6.32768035e-01
7.83503294e-01 6.25022128e-02 -2.15321317e-01 8.68722260e-01
4.82075304e-01 4.15049732e-01 4.83885795e-01 -9.95230496e-01
5.80446482e-01 4.25437868e-01 -3.58468264e-01 -1.46713698e+00
-1.95675313e-01 -1.99806780e-01 -1.06667054e+00 -8.53993446e-02
3.95375602e-02 -1.94802787e-02 -5.04604936e-01 1.90819752e+00
5.80977321e-01 7.29435742e-01 8.02400485e-02 9.54845428e-01
8.10083389e-01 6.99985981e-01 -3.13012123e-01 -5.32976270e-01
1.51150751e+00 -8.35465431e-01 -1.00476182e+00 1.10751195e-02
4.21231002e-01 -8.11030090e-01 1.07861781e+00 4.84270960e-01
-8.69008064e-01 -8.16083133e-01 -1.10991681e+00 -3.79547328e-02
-1.07093394e-01 3.15615237e-01 3.87130231e-01 7.46010900e-01
-6.41587615e-01 1.39821097e-01 -4.51082975e-01 -1.41253158e-01
4.97442991e-01 4.33754891e-01 -5.05847096e-01 -1.10136513e-02
-1.40121746e+00 4.20121044e-01 5.67335963e-01 5.21613538e-01
-1.17006648e+00 -6.42181575e-01 -8.70242298e-01 1.16662011e-01
4.47042733e-01 -4.91652966e-01 9.17934299e-01 -1.11698925e+00
-1.60671735e+00 9.16691601e-01 -1.72121450e-01 4.53822091e-02
4.43883121e-01 -1.31530091e-01 -9.86259818e-01 1.88553423e-01
-2.44017899e-01 6.87441647e-01 1.45465946e+00 -1.22692430e+00
-7.02401638e-01 -2.33916670e-01 -1.82545960e-01 2.40542620e-01
-8.97442400e-01 1.92967027e-01 -8.06215644e-01 -6.91998661e-01
-1.97147444e-01 -6.60893261e-01 6.21682107e-01 1.63705193e-03
-4.91739750e-01 -2.53741264e-01 1.22099733e+00 -5.67041576e-01
1.49613357e+00 -2.43325496e+00 4.22383815e-01 1.86972216e-01
3.77223253e-01 4.19202000e-01 -4.61436778e-01 1.51952520e-01
-2.27245584e-01 -2.24587888e-01 -2.36134063e-02 -3.29826683e-01
-1.02152243e-01 -1.15365349e-01 -4.03399736e-01 5.26148021e-01
2.25731358e-01 8.33391130e-01 -8.29640031e-01 -6.90036178e-01
3.28964353e-01 5.42046428e-01 -5.99333286e-01 9.20350924e-02
2.01609522e-01 2.89282680e-01 9.95930377e-03 8.67968380e-01
1.08047521e+00 3.20171639e-02 9.83107910e-02 -7.42602348e-01
8.38742405e-02 -7.60091171e-02 -1.03944790e+00 1.91682935e+00
-4.03035969e-01 8.04215968e-01 2.62914985e-01 -9.24173415e-01
9.38028932e-01 5.39859496e-02 5.24196208e-01 -8.56337368e-01
4.51761514e-01 1.72864348e-01 -9.14661959e-02 -8.82411718e-01
3.83046240e-01 7.16870949e-02 -9.04671475e-02 6.26006648e-02
2.84978390e-01 2.04083830e-01 -2.77668953e-01 8.85615721e-02
4.44618136e-01 -1.40359536e-01 -1.03921562e-01 1.47164660e-02
1.03173566e+00 -6.07468545e-01 7.68956840e-01 6.01513200e-02
-3.03393692e-01 3.60533893e-01 3.92497927e-01 -1.12822540e-02
-6.80985987e-01 -1.04164493e+00 -7.93984458e-02 1.12213540e+00
5.59102654e-01 -7.94275224e-01 -1.08170855e+00 -6.81010664e-01
-1.38982624e-01 2.04551652e-01 -5.66845894e-01 -8.29513788e-01
-5.31684756e-01 -4.82441783e-01 7.96950936e-01 2.88599044e-01
9.11268651e-01 -7.62657940e-01 -1.73498407e-01 -7.46024325e-02
-6.42579675e-01 -1.02578032e+00 -7.57792175e-01 -4.07176346e-01
-2.81045258e-01 -9.96698081e-01 -4.82466817e-01 -1.01648581e+00
2.81310856e-01 3.13749284e-01 6.78203881e-01 -9.15361336e-04
-2.49674931e-01 3.61686528e-01 -1.75174356e-01 1.95892751e-02
-2.95242637e-01 -2.13751137e-01 3.17922682e-01 5.31701207e-01
5.94350159e-01 -4.91372943e-01 -7.24730730e-01 6.17605507e-01
-7.39604771e-01 -4.07354161e-02 2.95978963e-01 1.02654755e+00
3.96530986e-01 2.86856472e-01 5.45836568e-01 8.08060169e-02
3.78269881e-01 -3.89375031e-01 -5.22584379e-01 4.62008268e-01
-2.62995034e-01 -2.08481714e-01 5.02037406e-01 -7.25491583e-01
-1.09460175e+00 -1.86381266e-02 -7.20323548e-02 -8.21881950e-01
6.16668873e-02 3.68945032e-01 -1.02715123e+00 -2.19310075e-01
2.47169912e-01 3.35535496e-01 5.83665483e-02 -1.42342702e-01
4.19272244e-01 8.44154418e-01 1.10085654e+00 -2.44824350e-01
7.26644754e-01 4.59878862e-01 -3.15362290e-02 -7.47108102e-01
-2.38038823e-01 -1.88460112e-01 -2.58477390e-01 -6.44520402e-01
7.73027539e-01 -1.04206347e+00 -1.12970495e+00 9.67491508e-01
-1.08916235e+00 5.12036458e-02 1.23451933e-01 5.04722893e-01
-5.08850276e-01 7.96510577e-01 -4.94552344e-01 -5.42686045e-01
-1.91811070e-01 -1.36584556e+00 1.24358058e+00 2.23022774e-01
4.88253981e-02 -7.29705215e-01 5.32700606e-02 4.10343051e-01
-6.46599308e-02 -8.60761181e-02 8.03962708e-01 -1.36032879e-01
-5.96987545e-01 1.14974521e-01 -3.31542969e-01 4.39425409e-01
2.80980170e-01 2.32239395e-01 -1.09269702e+00 -4.00682509e-01
-1.03227064e-01 -4.62128788e-01 9.69126701e-01 2.15743184e-01
1.47852266e+00 -3.83337915e-01 -2.89177388e-01 1.04981697e+00
8.74070942e-01 7.30636641e-02 8.80951107e-01 1.89217806e-01
1.02756774e+00 7.27611244e-01 9.19184685e-01 5.88755846e-01
5.34486115e-01 9.00009394e-01 6.55819297e-01 1.40625253e-01
-8.46755058e-02 -4.94004279e-01 6.81104302e-01 1.13621128e+00
3.60121757e-01 -4.03928101e-01 -5.37625074e-01 4.11546260e-01
-1.74600255e+00 -1.01921666e+00 1.02283947e-01 1.93621850e+00
6.08374298e-01 -2.61756390e-01 3.69082727e-02 3.12177867e-01
1.01609719e+00 3.43110353e-01 -5.71075618e-01 -5.16331047e-02
-3.56325895e-01 -2.32794076e-01 1.60365120e-01 2.80869782e-01
-1.12092674e+00 9.82419670e-01 5.08192110e+00 1.41175842e+00
-1.48841476e+00 1.51407868e-01 3.49650919e-01 -5.61870746e-02
-3.68179291e-01 -3.00106704e-01 -7.92822421e-01 8.76850843e-01
6.36787593e-01 -2.97532231e-02 5.45558512e-01 4.97363359e-01
2.66171813e-01 3.04448932e-01 -1.06751585e+00 1.77149415e+00
2.88567543e-01 -1.33102202e+00 -1.36102382e-02 2.26931840e-01
4.38124001e-01 -4.13417190e-01 3.45552713e-01 9.11602154e-02
-2.41457522e-01 -8.99056494e-01 7.85971165e-01 5.76994717e-01
1.09931612e+00 -1.15853369e+00 4.80486989e-01 3.84484306e-02
-1.60850406e+00 -4.41092223e-01 -1.26430944e-01 6.45337939e-01
1.86074436e-01 3.44921857e-01 -4.14997190e-01 6.85299277e-01
1.00498652e+00 9.57041264e-01 -4.29246783e-01 9.52368200e-01
6.07368611e-02 3.52152139e-01 -1.60652414e-01 4.01479512e-01
-1.10022329e-01 8.11183825e-02 6.42799377e-01 1.21264148e+00
7.00101495e-01 -2.71023601e-01 -2.20228568e-01 5.13922870e-01
-2.83127487e-01 3.39092650e-02 -3.65728319e-01 -1.35333136e-01
9.44571555e-01 1.38209116e+00 -2.19179749e-01 -7.32376501e-02
-1.70317352e-01 1.20251489e+00 1.44245720e-03 1.41193569e-01
-1.13286459e+00 -4.86366868e-01 1.12441730e+00 -3.17900836e-01
4.20668006e-01 1.12680614e-01 1.79269463e-01 -1.29729974e+00
5.01602627e-02 -1.21645164e+00 4.74629730e-01 -7.80843377e-01
-1.30068016e+00 4.98963028e-01 -2.18773559e-01 -1.48824692e+00
-2.76783556e-01 -4.54687953e-01 -6.04660392e-01 3.71230781e-01
-1.45984423e+00 -1.19942999e+00 -7.13901997e-01 1.19006097e+00
4.01809394e-01 -5.22736192e-01 4.42932427e-01 5.97545326e-01
-7.00741827e-01 1.30938315e+00 -7.69102424e-02 1.13043018e-01
1.06827736e+00 -3.84373128e-01 9.64537635e-02 8.63448799e-01
-3.79911140e-02 3.75348359e-01 3.59141469e-01 -4.47630495e-01
-1.59223759e+00 -1.05849349e+00 4.52921271e-01 -1.60056591e-01
5.40596426e-01 -5.57433009e-01 -9.56974566e-01 2.12997749e-01
2.68275172e-01 -7.99575746e-02 5.92214584e-01 -2.32642159e-01
-6.42447829e-01 -7.82333910e-01 -9.69770730e-01 4.38760191e-01
1.14500880e+00 -1.08717692e+00 -3.15305799e-01 7.84308091e-02
7.14570940e-01 -2.60859519e-01 -7.64276564e-01 6.51544034e-01
6.98918760e-01 -1.13814640e+00 1.00577927e+00 -2.41444722e-01
3.46644491e-01 -5.11807263e-01 -1.90887585e-01 -1.20479977e+00
-1.21246092e-01 -9.95877624e-01 -4.56003606e-01 1.81876516e+00
-2.48909339e-01 -3.63777667e-01 5.59601009e-01 -2.60740012e-01
-1.71022102e-01 -4.03686821e-01 -1.21878147e+00 -6.61451459e-01
-3.16465944e-01 -3.14419180e-01 8.81861329e-01 1.30711782e+00
-8.41266569e-03 1.75510362e-01 -9.06888783e-01 2.94278204e-01
4.98450518e-01 -9.56364945e-02 8.03728759e-01 -1.19462049e+00
-1.22290030e-01 -6.36167526e-01 -6.41279101e-01 -1.07696259e+00
4.30453151e-01 -6.17635250e-01 2.28729658e-02 -8.33544195e-01
2.92799711e-01 -1.41466353e-02 -4.08513755e-01 3.54239196e-01
-6.63585141e-02 2.26993620e-01 -2.27209982e-02 2.49484219e-02
-5.28105974e-01 1.02050543e+00 1.22800529e+00 -2.85365820e-01
1.28404833e-02 -5.24339616e-01 -5.03395498e-01 6.19898438e-01
5.34584641e-01 -2.35086128e-01 -5.10860145e-01 -2.99963772e-01
2.04939201e-01 5.05631343e-02 6.03878319e-01 -9.18126881e-01
3.38422656e-01 -7.90044069e-02 3.23677391e-01 -7.08677530e-01
5.95772386e-01 -8.29904914e-01 6.55057952e-02 2.02228352e-01
-1.98691517e-01 4.33007069e-02 2.18935221e-01 6.12479925e-01
-5.46346903e-01 2.60982513e-01 7.85386682e-01 3.94094408e-01
-9.10383821e-01 5.13025463e-01 -2.88182646e-01 -1.47601664e-01
1.10082066e+00 -4.38191295e-01 -4.53709424e-01 -4.95566517e-01
-1.82303965e-01 3.86085480e-01 3.38187873e-01 8.64557445e-01
9.43175018e-01 -1.77791405e+00 -4.75548118e-01 3.99292558e-01
3.51246089e-01 -4.58026141e-01 8.77435565e-01 1.01286054e+00
-1.09494805e-01 3.06320516e-03 -3.84114474e-01 -8.50428641e-01
-1.66990995e+00 6.55725658e-01 5.22011220e-01 3.55709136e-01
-2.24548161e-01 1.09913921e+00 4.41379845e-01 -3.63700151e-01
4.35933560e-01 1.62208200e-01 -1.28774077e-01 5.09206951e-01
6.36957884e-01 4.18463439e-01 8.93984064e-02 -1.08889198e+00
-6.93304241e-01 8.21326256e-01 -6.92417398e-02 8.38154182e-02
9.23520267e-01 -6.35923445e-01 -2.13588253e-01 1.06789485e-01
1.64277732e+00 2.77760118e-01 -1.41030455e+00 -8.89740363e-02
-4.63712424e-01 -5.58923066e-01 3.04819047e-01 -4.80440378e-01
-1.41799200e+00 1.17761683e+00 1.03134418e+00 -2.88070410e-01
1.64440274e+00 -1.71878990e-02 9.80345726e-01 1.89568356e-01
-4.50003985e-03 -1.11588967e+00 6.03598416e-01 3.27051669e-01
6.81878746e-01 -1.04405260e+00 -3.45435768e-01 -3.98249120e-01
-7.32354164e-01 1.21780443e+00 7.90815115e-01 4.72338319e-01
5.54299831e-01 1.66487053e-01 7.45025054e-02 -1.64787069e-01
-5.59010148e-01 4.25542630e-02 5.28525233e-01 5.73707700e-01
-1.10738114e-01 -2.60556012e-01 2.31376514e-01 4.97707635e-01
-2.69105554e-01 -1.26330823e-01 9.94212255e-02 4.68373895e-01
-1.82793155e-01 -8.57283652e-01 -3.55492622e-01 -3.89436446e-02
-1.03385963e-01 -1.79845579e-02 -4.30189788e-01 5.62663376e-01
4.63527858e-01 1.06126261e+00 1.35451287e-01 -1.10460162e+00
-3.95057648e-02 -2.91135997e-01 5.03672361e-01 1.47528723e-01
-3.48537058e-01 4.18719351e-01 -1.07095793e-01 -8.01650047e-01
-4.35713589e-01 -5.42424977e-01 -9.80242133e-01 -5.71429729e-01
-4.41542685e-01 4.26258519e-02 4.84552652e-01 8.65004718e-01
5.48613369e-01 3.07749569e-01 1.08231163e+00 -8.98144841e-01
-2.37495139e-01 -5.70152700e-01 -5.04656255e-01 9.75702405e-02
6.84127212e-01 -8.45372617e-01 -3.75021815e-01 -6.23761788e-02] | [13.105860710144043, 1.1765352487564087] |
209b594a-5c26-40ae-a433-f048feeb199f | sequential-item-recommendation-in-the-moba | 2201.08724 | null | https://arxiv.org/abs/2201.08724v1 | https://arxiv.org/pdf/2201.08724v1.pdf | Sequential Item Recommendation in the MOBA Game Dota 2 | Multiplayer Online Battle Arena (MOBA) games such as Dota 2 attract hundreds of thousands of players every year. Despite the large player base, it is still important to attract new players to prevent the community of a game from becoming inactive. Entering MOBA games is, however, often demanding, requiring the player to learn numerous skills at once. An important factor of success is buying the correct items which forms a complex task depending on various in-game factors such as already purchased items, the team composition, or available resources. A recommendation system can support players by reducing the mental effort required to choose a suitable item, helping, e.g., newer players or players returning to the game after a longer break, to focus on other aspects of the game. Since Sequential Item Recommendation (SIR) has proven to be effective in various domains (e.g. e-commerce, movie recommendation or playlist continuation), we explore the applicability of well-known SIR models in the context of purchase recommendations in Dota 2. To facilitate this research, we collect, analyze and publish Dota-350k, a new large dataset based on recent Dota 2 matches. We find that SIR models can be employed effectively for item recommendation in Dota 2. Our results show that models that consider the order of purchases are the most effective. In contrast to other domains, we find RNN-based models to outperform the more recent Transformer-based architectures on Dota-350k. | ['Andreas Hotho', 'Daniel Zoller', 'Johannes Kohlmann', 'Alexander Dallmann'] | 2022-01-17 | null | null | null | null | ['movie-recommendation', 'dota-2'] | ['miscellaneous', 'playing-games'] | [-5.61027586e-01 -4.38729167e-01 -2.14717016e-01 -6.60619140e-02
-3.93107951e-01 -7.22373068e-01 -6.51163310e-02 1.40344352e-01
-7.78490901e-01 3.61107558e-01 2.90876001e-01 -4.22554195e-01
-6.42662942e-01 -1.20455158e+00 -5.01509428e-01 -2.82361001e-01
-1.57506675e-01 8.28173578e-01 2.21278727e-01 -1.12408483e+00
2.25091428e-01 1.44336715e-01 -1.55545974e+00 5.86653709e-01
3.89160126e-01 1.14599657e+00 4.11435038e-01 3.29381108e-01
2.50067953e-02 9.38745737e-01 -4.62412387e-01 -9.12344098e-01
6.67505920e-01 -4.96835142e-01 -7.79998899e-01 -2.45682374e-01
-5.51177375e-02 -5.65514505e-01 -3.13863099e-01 6.16023123e-01
4.33921367e-01 7.09741354e-01 3.00541043e-01 -9.27178919e-01
-2.67301828e-01 1.26856816e+00 -2.62566954e-01 4.52830344e-01
2.04688013e-01 -7.50127509e-02 1.59858406e+00 -8.16746533e-01
5.39683819e-01 5.85105062e-01 7.46543348e-01 2.07767114e-01
-8.87673140e-01 -6.86085939e-01 3.24105442e-01 4.87574816e-01
-1.06687391e+00 -1.05004042e-01 4.23999608e-01 -2.30565920e-01
9.97565508e-01 3.37304920e-01 1.16195321e+00 9.47678387e-01
2.09031969e-01 7.38670349e-01 5.68502545e-01 -4.56147194e-02
2.32915714e-01 -2.14872003e-01 6.65141717e-02 1.74381390e-01
1.75849482e-01 -2.13610560e-01 -7.56768167e-01 1.01389177e-01
7.63674915e-01 3.17311496e-01 1.53473809e-01 -1.72634527e-01
-7.97660112e-01 1.18666339e+00 6.04866982e-01 4.44811881e-01
-9.12343323e-01 1.72476042e-02 4.68622923e-01 7.56666839e-01
3.49778086e-01 1.02096999e+00 -2.75795519e-01 -7.52046824e-01
-9.18749571e-01 6.22191072e-01 7.76727855e-01 4.66582447e-01
5.17301500e-01 -6.14020452e-02 -2.69716457e-02 1.00186491e+00
4.46625687e-02 -6.59392774e-02 4.84714359e-01 -7.85487413e-01
3.97709072e-01 5.38713932e-01 1.42241493e-01 -1.22464883e+00
-4.89605546e-01 -8.24209929e-01 -8.61564934e-01 4.61675636e-02
6.82987094e-01 -8.53632390e-02 -6.32478967e-02 1.51767731e+00
-8.12370405e-02 -4.17323411e-02 -3.62320900e-01 1.19867611e+00
8.06479335e-01 4.99269426e-01 -1.32779062e-01 -9.02359560e-02
1.29184639e+00 -1.03495336e+00 -3.28704774e-01 -4.68689263e-01
4.44666326e-01 -7.56786168e-01 8.47608447e-01 8.70052695e-01
-1.38459706e+00 -6.63151920e-01 -7.74634957e-01 1.18622594e-01
-1.47151574e-01 -9.63870063e-02 1.04541123e+00 5.47160029e-01
-8.45253944e-01 8.82760167e-01 -3.94517839e-01 -3.33800554e-01
4.31825250e-01 6.63922250e-01 -2.73203522e-01 -3.07496697e-01
-1.31770110e+00 6.80568278e-01 9.61625576e-02 5.02006598e-02
-6.71392024e-01 -6.22367203e-01 -3.64287555e-01 3.63207251e-01
7.53918469e-01 -5.58642447e-01 1.35373867e+00 -8.21371198e-01
-1.59727776e+00 4.66388196e-01 7.38009036e-01 -5.60733616e-01
3.11975837e-01 -4.02852803e-01 -1.93474591e-02 -4.36358392e-01
8.66945386e-02 3.43806863e-01 3.43609005e-01 -2.84215957e-01
-1.08304989e+00 -3.17099124e-01 9.23804164e-01 6.87246680e-01
-5.29694140e-01 2.41207138e-01 -4.05817002e-01 -4.80458438e-01
-8.77614617e-02 -8.93022120e-01 -4.76227432e-01 -6.28560126e-01
4.25602915e-03 -3.54637235e-01 -2.32189476e-01 -6.84355080e-01
1.32194972e+00 -2.02903605e+00 4.23533976e-01 3.59856963e-01
3.87355685e-01 1.39010578e-01 -2.70242184e-01 7.12804079e-01
1.61512822e-01 -1.15017764e-01 8.53794336e-01 -1.90345407e-01
8.36647674e-02 3.23135518e-02 -1.87728643e-01 9.82981622e-02
-5.32132685e-01 9.80719328e-01 -6.74941957e-01 2.88144767e-01
-4.99384059e-03 -2.77121235e-02 -1.05044544e+00 1.77481055e-01
-7.90186226e-02 2.93672103e-02 -3.98846328e-01 1.83409527e-01
2.96196282e-01 -3.10979247e-01 3.00997883e-01 1.06518477e-01
-3.50424498e-02 9.20952916e-01 -1.31764436e+00 1.65889120e+00
-5.23381054e-01 3.48924279e-01 5.18474020e-02 -1.07165432e+00
7.78330624e-01 9.24421102e-02 6.45675540e-01 -9.51135516e-01
5.42043447e-01 1.59515098e-01 5.95773101e-01 -1.12393156e-01
1.17530990e+00 -1.10504940e-01 -2.61275858e-01 8.75951767e-01
-1.21047601e-01 4.44629282e-01 4.98601079e-01 5.55719972e-01
1.23371625e+00 -1.30034029e-01 3.53970468e-01 -9.94942412e-02
-5.26868664e-02 1.35386258e-01 4.17888671e-01 1.16026258e+00
1.07587390e-01 2.54728645e-01 2.62076467e-01 -6.02002025e-01
-9.50035691e-01 -6.02154374e-01 5.59455574e-01 1.83469880e+00
1.48364931e-01 -7.71228790e-01 -2.48853907e-01 -4.92723495e-01
-1.30267516e-01 3.64066571e-01 -5.48051298e-01 -1.64599612e-01
-5.60787499e-01 -4.30856764e-01 2.39691108e-01 4.88648117e-01
4.64106828e-01 -1.18624938e+00 -4.21634287e-01 8.03677678e-01
-5.66928923e-01 -6.58869982e-01 -7.23069847e-01 2.60786235e-01
-5.01355708e-01 -8.08938801e-01 -5.08509636e-01 -6.25696182e-01
-7.75384307e-02 7.58646131e-01 1.45601726e+00 1.82253331e-01
3.77670024e-03 2.64173537e-01 -7.26165593e-01 -2.80872770e-02
1.83301624e-02 5.52376270e-01 3.77564281e-01 -1.38359666e-01
4.01816607e-01 -9.29887950e-01 -6.64764524e-01 7.75869429e-01
-5.13648510e-01 9.58077386e-02 4.14666533e-01 7.23055720e-01
2.62444824e-01 4.20157313e-01 6.47877932e-01 -8.92573416e-01
1.06994998e+00 -7.69413948e-01 -1.84002459e-01 -2.02571958e-01
-3.09582323e-01 -7.25920022e-01 8.89906645e-01 -7.31462598e-01
-5.20433724e-01 -6.66248739e-01 -4.49783266e-01 -9.77722704e-02
2.48446152e-01 1.08488166e+00 1.45385072e-01 1.97634637e-01
9.31915283e-01 1.09101251e-01 -5.13210669e-02 -5.14848948e-01
1.39699534e-01 4.32692498e-01 1.34609357e-01 -3.22926044e-01
5.37979364e-01 3.50638554e-02 -4.56304431e-01 -3.38306457e-01
-8.17984104e-01 -7.01648831e-01 -1.96893990e-01 -6.18453979e-01
3.99429858e-01 -1.00544167e+00 -1.24743330e+00 3.22991788e-01
-6.37031615e-01 -3.81821126e-01 -5.09764373e-01 6.74879491e-01
-2.20941991e-01 -7.73162469e-02 -9.55753922e-01 -5.05764723e-01
-3.21584493e-01 -8.53460550e-01 1.11886382e-01 2.72489160e-01
-3.90646785e-01 -6.17353976e-01 1.45716265e-01 8.85315657e-01
7.43615031e-01 -7.04516351e-01 5.00290394e-01 -9.03061271e-01
-5.98028481e-01 -3.10807079e-01 1.62905455e-01 1.09100446e-01
2.88116690e-02 -6.31749570e-01 -2.11482197e-01 -4.19015259e-01
-1.89653337e-01 -5.33013761e-01 5.09773254e-01 4.75767195e-01
6.18099332e-01 -9.37717184e-02 8.51996541e-02 1.86799809e-01
8.51281583e-01 2.64203757e-01 8.15920115e-01 6.78698182e-01
4.15313751e-01 5.00198007e-01 9.06131387e-01 6.41619503e-01
5.98087549e-01 1.02543926e+00 5.13361692e-01 2.74994314e-01
2.37164304e-01 -3.55510443e-01 4.58991021e-01 1.25590384e+00
-6.70456052e-01 -4.03568357e-01 -3.22846770e-01 4.34183419e-01
-1.95197904e+00 -1.23663044e+00 3.04372776e-02 2.18783307e+00
5.34892023e-01 3.11917782e-01 8.44336331e-01 1.05649509e-01
2.60369182e-01 -1.29203290e-01 -4.21047151e-01 -3.60477984e-01
-6.01190925e-02 3.82931471e-01 4.04589802e-01 1.26574948e-01
-7.97444999e-01 9.72840071e-01 5.59892893e+00 1.44778526e+00
-8.48845363e-01 2.00364351e-01 5.91106474e-01 -6.59887910e-01
-2.55624443e-01 -2.19993770e-01 -6.68778360e-01 4.44644749e-01
6.18792772e-01 -1.91657007e-01 1.14072335e+00 8.21293235e-01
1.83312476e-01 4.93405480e-03 -9.28760767e-01 1.09825706e+00
-5.10018952e-02 -1.57933509e+00 -2.13318720e-01 4.48242247e-01
4.68498319e-01 1.65546536e-01 1.41861528e-01 9.30972099e-01
6.47823274e-01 -1.03337312e+00 7.52706170e-01 9.08517465e-02
3.31070572e-01 -8.86418760e-01 8.75245750e-01 6.71706676e-01
-1.15436602e+00 -3.34777325e-01 -7.19895482e-01 -9.33879733e-01
2.14194492e-01 2.38988966e-01 -3.79726648e-01 5.65575421e-01
1.02845860e+00 8.60645890e-01 -2.68528879e-01 1.16993284e+00
1.01670451e-01 7.07126558e-01 -3.95121276e-01 -3.57939452e-01
4.23964649e-01 -6.47841513e-01 3.25441986e-01 3.93870741e-01
5.64421296e-01 3.83135229e-01 1.90275699e-01 3.85377347e-01
-3.15562308e-01 7.33968616e-01 -3.76760721e-01 -1.38824075e-01
1.28455520e-01 1.37420750e+00 -8.97215068e-01 1.65336967e-01
-4.50665057e-01 8.48579049e-01 6.13662362e-01 -1.04402922e-01
-7.17703164e-01 1.01907225e-02 9.85671461e-01 5.98914087e-01
6.15442395e-01 -3.88371274e-02 5.86229712e-02 -1.06077206e+00
-1.03868507e-01 -1.41621363e+00 4.64906573e-01 -6.23537600e-01
-1.42153835e+00 6.25285089e-01 -5.34971535e-01 -1.26837337e+00
-1.98674530e-01 -4.57968712e-01 -6.20211065e-01 5.70450544e-01
-7.73607671e-01 -8.37794363e-01 -3.09155788e-02 5.81371188e-01
6.93463683e-01 -4.60351348e-01 5.46816647e-01 9.03048396e-01
-3.65437269e-01 6.62829399e-01 1.77487403e-01 2.27042884e-02
5.26968062e-01 -8.13915730e-01 4.62645680e-01 4.88134086e-01
6.45830035e-01 5.79320192e-01 4.67406303e-01 -5.41123092e-01
-1.31968617e+00 -4.38053250e-01 7.42832363e-01 -2.86131471e-01
6.48117959e-01 -4.86050159e-01 -4.83846366e-01 6.89112663e-01
-5.79426438e-02 -6.24931395e-01 1.01934218e+00 9.31722879e-01
-6.79990277e-02 -2.27407038e-01 -6.18976712e-01 7.30133235e-01
1.35223067e+00 -2.73638606e-01 -1.69333711e-01 1.30432785e-01
2.57980376e-01 -3.21436226e-01 -7.91104317e-01 -1.40862375e-01
6.61473811e-01 -1.20626473e+00 1.03784287e+00 -9.48179722e-01
5.33718228e-01 1.50995925e-01 -1.24969639e-01 -1.60216773e+00
-1.02599609e+00 -6.81917012e-01 3.33569795e-01 8.50162983e-01
4.21193123e-01 -3.40267926e-01 1.23624647e+00 6.48829043e-01
1.69133693e-02 -8.12444568e-01 -9.11279559e-01 -5.60123801e-01
6.95570260e-02 -8.48273873e-01 6.64577425e-01 8.04474711e-01
3.61905217e-01 7.12640941e-01 -1.05722535e+00 -3.55388284e-01
6.49800748e-02 2.60079026e-01 1.04165840e+00 -1.46575737e+00
-1.00157833e+00 -6.41534686e-01 -3.19743514e-01 -1.60595560e+00
-4.93152112e-01 -9.33425248e-01 -2.23743200e-01 -1.60574591e+00
2.89789408e-01 -7.29096174e-01 -5.03718376e-01 4.58478332e-01
4.69053760e-02 6.69252992e-01 4.85087931e-01 2.49511048e-01
-1.12123692e+00 2.60724366e-01 1.29534471e+00 2.83354148e-03
-3.56481761e-01 6.45206690e-01 -1.28314614e+00 4.33554918e-01
7.40289092e-01 -3.92228603e-01 -4.43395406e-01 -2.69040942e-01
1.36970830e+00 2.07383215e-01 -2.89886385e-01 -7.24598467e-01
4.38415736e-01 -1.18513532e-01 -1.28168494e-01 -3.73524934e-01
4.81783956e-01 -7.26334035e-01 3.42728108e-01 2.48773709e-01
-2.84229249e-01 1.71426699e-01 5.55113330e-03 3.49838585e-01
-7.68544003e-02 -3.79407823e-01 1.65388271e-01 -1.40595853e-01
-4.58735168e-01 3.77487212e-01 -1.02299178e+00 3.94053459e-02
7.86394119e-01 -4.00689512e-01 9.06827580e-03 -1.12088764e+00
-7.21433818e-01 1.37828961e-01 1.77966077e-02 5.98733783e-01
3.15099359e-01 -1.28330195e+00 -7.85765052e-01 -8.42119306e-02
-1.62081480e-01 -3.84331375e-01 7.77408957e-01 6.74600542e-01
-3.62239867e-01 2.19331920e-01 -6.12146437e-01 9.69021767e-02
-1.32286310e+00 2.84199238e-01 2.35435799e-01 -9.22504008e-01
-5.00355661e-01 1.17801082e+00 1.04933232e-01 -3.71980608e-01
1.79411590e-01 -7.85771608e-02 -7.19094157e-01 2.80199170e-01
6.22664034e-01 4.47895050e-01 1.96732864e-01 -4.11166668e-01
-3.15575376e-02 1.71213725e-03 -5.04033804e-01 1.38412908e-01
1.76121533e+00 -5.81179000e-02 4.48898189e-02 1.60323709e-01
3.75718564e-01 1.73471421e-01 -8.45583081e-01 -4.54043895e-01
-4.94676054e-01 -6.10947669e-01 7.98474550e-02 -8.86758983e-01
-1.44873726e+00 5.03457606e-01 1.38499513e-01 6.14577353e-01
1.03836250e+00 -1.36028677e-01 8.50456774e-01 5.64945638e-01
1.00886154e+00 -1.03026354e+00 3.54227364e-01 6.56595409e-01
4.83143389e-01 -9.73231912e-01 -5.66917285e-02 -1.56711489e-01
-8.15648437e-01 7.99689353e-01 6.63274169e-01 -2.40223706e-01
6.18556738e-01 -9.13859084e-02 -3.32435891e-02 -2.49325246e-01
-7.87955403e-01 -1.24903142e-01 6.90254197e-02 2.95064867e-01
1.76740542e-01 8.89325812e-02 -3.63532126e-01 1.50085342e+00
-6.09863162e-01 -5.46221212e-02 6.73305511e-01 4.82101500e-01
-3.36567134e-01 -1.32756853e+00 -4.85219620e-02 1.00314713e+00
-6.73731387e-01 -3.54027748e-01 -1.08367182e-01 5.85090756e-01
1.40562862e-01 1.16405630e+00 1.43748716e-01 -6.76258385e-01
5.36777914e-01 -3.66670012e-01 3.80195379e-01 -6.60651505e-01
-1.47947359e+00 6.71277270e-02 3.49521190e-01 -5.26818752e-01
-2.91446541e-02 -7.23402798e-01 -5.96301317e-01 -1.09828472e+00
-4.38889474e-01 4.29016411e-01 3.75035524e-01 9.23081219e-01
3.41143817e-01 6.22450948e-01 4.41777974e-01 -7.91971922e-01
-3.86351168e-01 -1.11641526e+00 -1.08333862e+00 3.39854658e-01
-5.90189159e-01 -7.04859853e-01 3.20379138e-02 -6.08347714e-01] | [6.580568313598633, 0.40051648020744324] |
33ccb270-dbb6-4d44-9c6e-a6771aa45571 | exploiting-the-generative-adversarial-network | 2301.11871 | null | https://arxiv.org/abs/2301.11871v1 | https://arxiv.org/pdf/2301.11871v1.pdf | Exploiting the Generative Adversarial Network Approach to Create a Synthetic Topography Corneal Image | Corneal diseases are the most common eye disorders. Deep learning techniques are used to per-form automated diagnoses of cornea. Deep learning networks require large-scale annotated datasets, which is conceded as a weakness of deep learning. In this work, a method for synthesizing medical images using conditional generative adversarial networks (CGANs), is presented. It also illustrates how produced medical images may be utilized to enrich medical data, improve clinical decisions, and boost the performance of the conventional neural network (CNN) for medical image diagnosis. The study includes using corneal topography captured using a Pentacam device from patients with corneal diseases. The dataset contained 3448 different corneal images. Furthermore, it shows how an unbalanced dataset affects the performance of classifiers, where the data are balanced using the resampling approach. Finally, the results obtained from CNN networks trained on the balanced dataset are compared to those obtained from CNN networks trained on the imbalanced dataset. For performance, the system estimated the diagnosis accuracy, precision, and F1-score metrics. Lastly, some generated images were shown to an expert for evaluation and to see how well experts could identify the type of image and its condition. The expert recognized the image as useful for medical diagnosis and for determining the severity class according to the shape and values, by generating images based on real cases that could be used as new different stages of illness between healthy and unhealthy patients. | ['P. S. JosephNg', 'Sinan Q. Salih', 'Tarik A. Rashid', 'Jafar Majidpour', 'Nebras H. Ghaeb', 'Sezgin Aydin', 'Samer Kais Jameel'] | 2022-12-25 | null | null | null | null | ['medical-diagnosis'] | ['medical'] | [ 7.50877336e-02 2.69225329e-01 1.92869529e-01 -2.45104179e-01
-3.32329750e-01 -1.89279452e-01 7.50751123e-02 -7.76707530e-02
-1.98660970e-01 8.85729730e-01 -2.87088770e-02 -2.25723192e-01
1.66965239e-02 -8.33270788e-01 -5.46004832e-01 -9.42090094e-01
2.33422413e-01 5.22785246e-01 -3.80419195e-01 1.38440579e-01
2.65154749e-01 6.85714602e-01 -1.98996794e+00 6.21843398e-01
1.17824209e+00 1.02641988e+00 -3.14584553e-01 7.47844517e-01
-7.88207948e-02 8.07571828e-01 -1.25611174e+00 -4.75100100e-01
2.53415853e-01 -5.40782928e-01 -4.56271380e-01 1.43313497e-01
8.04456115e-01 -4.32283223e-01 1.53938755e-01 1.07045865e+00
8.24273169e-01 -4.97342914e-01 8.18259478e-01 -1.08797705e+00
-7.64802754e-01 -6.74549416e-02 -9.26491097e-02 -2.39886120e-02
2.38282308e-01 4.62546557e-01 9.63221714e-02 -5.98489344e-01
7.73609221e-01 7.81654894e-01 6.91191733e-01 6.31918550e-01
-7.26165533e-01 -7.12877214e-01 -5.71214914e-01 3.90542179e-01
-9.67224419e-01 -2.06442419e-02 4.41215187e-01 -8.23272884e-01
6.30416751e-01 2.73193836e-01 1.22842574e+00 8.73914123e-01
4.94516969e-01 5.04091144e-01 1.53713429e+00 -5.16816735e-01
1.12205774e-01 3.56581241e-01 -1.35829300e-01 6.88836157e-01
5.60699224e-01 3.97598416e-01 -2.20746219e-01 -3.88261303e-02
6.16629601e-01 -6.60637245e-02 -4.58790243e-01 8.90977308e-02
-8.86171460e-01 8.47814441e-01 4.23908323e-01 3.21320236e-01
-6.26034021e-01 -3.35704416e-01 3.81222546e-01 4.23154891e-01
5.21564245e-01 8.35453868e-01 -3.21699768e-01 2.92639375e-01
-8.64129603e-01 5.78096882e-02 6.02029622e-01 4.08093333e-01
3.97105426e-01 1.89909115e-01 -2.23837242e-01 6.53462172e-01
-3.63828987e-02 4.73023713e-01 9.08175409e-01 -5.53157806e-01
-3.33001554e-01 9.30186629e-01 -8.10236931e-02 -8.34371269e-01
-3.72319281e-01 -5.66381812e-01 -1.01321769e+00 7.95743823e-01
5.48563957e-01 -4.36644197e-01 -1.47867441e+00 9.30862010e-01
9.74157900e-02 9.05194357e-02 3.59411359e-01 9.97398853e-01
1.05088437e+00 3.66210401e-01 -2.41626546e-01 -1.48142606e-01
1.23580134e+00 -8.09776723e-01 -9.02121484e-01 1.97923884e-01
4.75243539e-01 -1.00629663e+00 9.00126994e-01 9.21650350e-01
-1.01933074e+00 -6.97888434e-01 -9.92773414e-01 1.09712534e-01
-5.40530264e-01 4.58022088e-01 5.58349073e-01 6.30656779e-01
-1.28083730e+00 3.13582569e-01 -4.00573373e-01 -3.68788868e-01
6.43846810e-01 4.86441821e-01 -2.76419580e-01 -9.15690809e-02
-1.12955630e+00 1.11762011e+00 2.26115882e-01 8.16075355e-02
-6.18721843e-01 -7.40079820e-01 -6.06437385e-01 -1.66034400e-01
-3.91608447e-01 -9.83068347e-01 7.85857618e-01 -1.50576091e+00
-1.37884772e+00 1.24597061e+00 8.57672021e-02 -5.77568233e-01
6.81070149e-01 4.97955084e-02 -5.26791275e-01 3.37802917e-01
-2.55121350e-01 5.98856926e-01 8.16418827e-01 -1.17080903e+00
-7.23959446e-01 -5.49632609e-01 -1.80061430e-01 -4.15464751e-02
2.65948824e-03 -1.30385920e-01 -1.24814242e-01 -6.50667131e-01
-2.76745796e-01 -8.79331112e-01 3.19288298e-02 1.20508492e-01
-6.34518027e-01 1.19427942e-01 5.59973538e-01 -9.29958463e-01
6.85370266e-01 -1.89460814e+00 -2.64727205e-01 4.48649138e-01
2.88366944e-01 6.44415200e-01 -7.47067183e-02 1.14960477e-01
-2.44580507e-01 2.14306444e-01 3.97166573e-02 -6.45735487e-02
-5.65459132e-01 -1.13613354e-02 2.03013271e-01 3.92874628e-01
3.49623770e-01 7.14378953e-01 -4.80877489e-01 -3.07585776e-01
4.22417402e-01 7.81946957e-01 -3.18101764e-01 3.79299998e-01
-1.22964126e-03 4.88055885e-01 1.00616746e-01 9.67043042e-01
6.85019076e-01 -5.83419129e-02 -7.71886632e-02 -4.86851782e-01
3.37080896e-01 -3.57719600e-01 -8.17245305e-01 8.60088527e-01
-4.85802084e-01 1.11212623e+00 -3.76513422e-01 -7.41764784e-01
9.84430432e-01 4.54140723e-01 2.53055274e-01 -7.49585450e-01
3.38919133e-01 1.66430563e-01 2.69449085e-01 -1.06762028e+00
1.49535045e-01 -1.64013043e-01 6.68748140e-01 5.87445386e-02
-1.22806646e-01 -2.38753319e-01 4.52221334e-02 -4.09067839e-01
5.44215322e-01 -1.79814637e-01 1.28739625e-01 5.68785369e-02
4.06721503e-01 2.81866401e-01 1.89871386e-01 2.98514575e-01
-2.15548798e-01 7.28584051e-01 6.74000740e-01 -8.71055901e-01
-1.11680424e+00 -6.97847426e-01 -5.32830298e-01 1.26818866e-01
-1.91492453e-01 4.65602100e-01 -9.68668163e-01 -3.54558080e-01
3.74655686e-02 3.23247254e-01 -8.16015780e-01 -2.60633379e-01
5.34413233e-02 -1.09841251e+00 4.65351939e-01 2.97371864e-01
2.30875224e-01 -1.20901573e+00 -6.23987913e-01 -1.70374840e-01
-2.19589118e-02 -6.73372567e-01 1.64357394e-01 -2.73084342e-01
-7.72458375e-01 -1.58762944e+00 -1.02989948e+00 -1.11592937e+00
1.09445167e+00 -4.81585145e-01 1.09403825e+00 2.74300963e-01
-8.53071570e-01 -9.75278765e-02 -1.54015929e-01 -1.04289901e+00
-7.76927829e-01 -3.63201648e-01 9.75684728e-03 1.48456275e-01
7.00064719e-01 -1.28931329e-01 -9.49672163e-01 4.73795198e-02
-7.88856208e-01 3.15323360e-02 8.35743427e-01 1.10581744e+00
7.98470080e-01 2.78591067e-01 5.76669097e-01 -1.24847710e+00
9.55989420e-01 -1.19133241e-01 -5.11306345e-01 1.18554488e-01
-9.39734280e-01 -3.14878970e-01 4.16741222e-01 -3.95475298e-01
-7.42125690e-01 -2.55353183e-01 -4.84567806e-02 -6.10257208e-01
-2.96715796e-01 5.50119638e-01 1.71190649e-01 -2.65237033e-01
9.30234134e-01 4.07785103e-02 6.51487827e-01 3.96390911e-03
-1.05415560e-01 1.00550079e+00 5.14228523e-01 1.43869683e-01
1.96248189e-01 3.36506844e-01 -2.73016132e-02 -5.45236707e-01
-4.74427998e-01 -1.33456618e-01 -3.57466429e-01 -3.10819000e-01
8.61953437e-01 -8.98411870e-01 -7.52813041e-01 1.06645179e+00
-8.29833508e-01 -2.78863072e-01 -3.00826907e-01 5.95620215e-01
-2.65243143e-01 -1.26572743e-01 -6.62012279e-01 -6.82083130e-01
-6.33643746e-01 -1.24104798e+00 7.35059679e-01 5.01416326e-01
-1.16179079e-01 -1.10028529e+00 -7.92037249e-02 7.60969281e-01
4.50891346e-01 7.36678421e-01 1.19712150e+00 -5.15815616e-01
-3.56063128e-01 -6.88652217e-01 -1.51401460e-01 9.45910633e-01
2.53637910e-01 5.97954690e-01 -1.21958399e+00 -2.67281950e-01
-4.06374224e-02 -3.85488838e-01 5.24490058e-01 8.01710010e-01
1.37804043e+00 -4.87286508e-01 -5.34020104e-02 6.47415459e-01
1.67853963e+00 5.58519185e-01 1.25479770e+00 2.11723104e-01
4.43063259e-01 8.21970820e-01 4.59798872e-01 7.51270279e-02
7.69572034e-02 1.33553848e-01 5.68920493e-01 -8.01911891e-01
-4.39194590e-01 1.67802200e-01 1.07277073e-02 5.72885096e-01
-3.83697182e-01 -2.20892161e-01 -8.29894125e-01 6.48925066e-01
-1.08827269e+00 -7.35751212e-01 -1.55817881e-01 2.18501401e+00
8.23199093e-01 -4.88789827e-02 -1.07561536e-01 2.98164159e-01
7.87189662e-01 -6.67964697e-01 -6.47597969e-01 -6.68084621e-01
-2.94799805e-01 5.85461318e-01 3.66996020e-01 2.96657473e-01
-7.00486779e-01 5.81079781e-01 6.62818623e+00 4.43132222e-01
-1.78981590e+00 -3.35135907e-01 1.05297673e+00 -1.50112122e-01
3.09037715e-02 -5.97410917e-01 -2.71906048e-01 8.01047564e-01
9.79823768e-01 1.73946060e-02 6.78091645e-02 5.64666152e-01
3.08755577e-01 -3.78705680e-01 -8.45340312e-01 9.88872111e-01
3.38652134e-01 -1.37440693e+00 3.69051546e-01 1.27954885e-01
9.85985160e-01 -1.78698659e-01 2.55190969e-01 -8.46484601e-02
-1.08894199e-01 -1.38297224e+00 -2.18962617e-02 1.11072767e+00
1.31325209e+00 -7.37209201e-01 1.66495693e+00 -1.51229158e-01
-1.92252457e-01 1.56919897e-01 -2.63804406e-01 6.05911762e-02
-3.83297056e-01 6.47094488e-01 -1.36135221e+00 3.31789196e-01
6.76008165e-01 4.61238295e-01 -8.04892957e-01 1.25968897e+00
-1.99457884e-01 6.19567633e-01 1.46105170e-01 -1.67718068e-01
-2.09912807e-01 -2.76085347e-01 2.73618877e-01 8.49620521e-01
3.28009248e-01 -2.72942394e-01 -4.89289671e-01 6.73779130e-01
3.11986543e-02 3.43120247e-01 -6.19326353e-01 1.41652869e-02
1.71849757e-01 1.07349539e+00 -3.22640538e-01 -4.35344547e-01
-1.55636370e-01 7.02941358e-01 -3.19681108e-01 3.55532318e-01
-4.69516933e-01 -5.55130720e-01 5.71862936e-01 4.02430683e-01
-1.87696964e-01 8.09700072e-01 -5.44814825e-01 -1.01218939e+00
-6.44324273e-02 -1.17427194e+00 1.67496160e-01 -1.43392658e+00
-1.41179013e+00 9.49642122e-01 -5.97621202e-01 -1.50912476e+00
-3.26490790e-01 -8.60365093e-01 -5.28933585e-01 1.30180347e+00
-1.59872246e+00 -1.23062372e+00 -9.27405417e-01 5.81674695e-01
1.22763276e-01 -6.27874255e-01 9.64559436e-01 3.52642894e-01
-5.05683839e-01 6.62089348e-01 -7.33736996e-03 3.48540068e-01
9.65846300e-01 -1.28304529e+00 -3.04792523e-01 6.05649590e-01
-2.98264742e-01 3.73295993e-01 5.36798894e-01 -4.91835952e-01
-5.86177826e-01 -1.24355388e+00 7.38286853e-01 -1.82639822e-01
1.42493555e-02 3.61870050e-01 -8.64899039e-01 1.84201211e-01
3.97686630e-01 -3.85262072e-02 1.17169428e+00 -2.36396685e-01
5.25147393e-02 -1.59577101e-01 -1.72742617e+00 3.80567789e-01
2.63340473e-01 -4.13845599e-01 -4.49305803e-01 5.19688845e-01
3.05225670e-01 -7.15505719e-01 -1.06814742e+00 6.00567460e-01
6.37047887e-01 -1.30352497e+00 6.29365504e-01 -8.76062036e-01
9.26332712e-01 -1.46927580e-01 2.95454323e-01 -1.58511114e+00
1.41542017e-01 1.91153716e-02 2.28799865e-01 7.13789701e-01
5.01681089e-01 -8.51357341e-01 9.53709185e-01 3.93751532e-01
-8.16999078e-02 -1.32260132e+00 -4.74352092e-01 -2.35005692e-01
1.69417337e-01 1.53396502e-01 5.99227309e-01 1.02728415e+00
-5.51907241e-01 -1.51485488e-01 -1.09471589e-01 1.23665616e-01
3.70721996e-01 9.79392305e-02 8.06421638e-01 -1.39018428e+00
7.11590350e-02 -3.64158928e-01 -9.21746194e-01 1.74244672e-01
-1.18922494e-01 -6.15217805e-01 -4.41377521e-01 -1.75842810e+00
8.35332274e-03 -3.12028825e-01 -2.50862300e-01 5.36886871e-01
-2.11777270e-01 8.88400853e-01 -6.70143813e-02 3.42968166e-01
4.66652930e-01 -9.86958891e-02 1.65438664e+00 -2.71818340e-01
-1.97388545e-01 2.31495962e-01 -7.07243383e-01 6.88946724e-01
9.32502151e-01 -4.76954840e-02 -3.01300406e-01 -9.71612707e-02
2.22524986e-01 -8.00384805e-02 5.08966506e-01 -1.27922690e+00
6.14790209e-02 1.16434947e-01 8.95540714e-01 -4.28915858e-01
1.60975233e-01 -6.46906257e-01 3.31448555e-01 8.72927606e-01
-7.07811266e-02 -1.69369221e-01 3.08633357e-01 3.45945358e-02
-7.25531638e-01 -3.07184309e-02 1.07558763e+00 -1.06293507e-01
-5.50857246e-01 1.25054508e-01 -2.61762738e-01 -1.89050823e-01
1.20440280e+00 -7.25431740e-01 -4.70337898e-01 -4.61646885e-01
-8.98760319e-01 -7.87897110e-02 6.37538075e-01 1.10824071e-01
7.49816120e-01 -8.88306975e-01 -8.43337893e-01 4.43041474e-01
2.96406567e-01 2.08869964e-01 4.52575356e-01 7.96241283e-01
-1.26358318e+00 2.51863182e-01 -6.98146105e-01 -6.24124944e-01
-1.49925756e+00 5.24393380e-01 8.66331935e-01 2.83302590e-02
-1.38366073e-01 7.14483321e-01 -8.29339400e-02 -2.24109277e-01
1.23232551e-01 -6.52432680e-01 -5.37639737e-01 2.79203027e-01
4.31500584e-01 1.90438524e-01 5.10276675e-01 -2.34209433e-01
1.25401735e-01 7.56385624e-01 2.61020139e-02 4.98395771e-01
9.83189821e-01 2.18934074e-01 -3.84801120e-01 1.90470561e-01
9.58476305e-01 4.21230383e-02 -7.07707882e-01 2.22089991e-01
-6.87343895e-01 -4.09291148e-01 1.68538272e-01 -1.56794250e+00
-1.59014285e+00 8.86458039e-01 1.30430496e+00 1.02160931e-01
1.57108033e+00 -4.26819593e-01 5.07950068e-01 -1.51128788e-02
2.17526406e-01 -7.92337179e-01 -8.77462924e-02 -2.31048554e-01
8.08841109e-01 -1.44082081e+00 -1.41148537e-01 -2.90730923e-01
-7.19245017e-01 1.30674338e+00 7.37883627e-01 -4.74155284e-02
4.32334244e-01 3.01325411e-01 8.37221861e-01 -2.92151749e-01
-3.09455425e-01 1.00227827e-02 3.89849603e-01 9.37747359e-01
2.65455395e-01 1.39172986e-01 -2.80499250e-01 3.46053958e-01
-3.49018127e-01 6.02671623e-01 9.15973067e-01 4.36748654e-01
-1.05086959e-03 -9.28892434e-01 -4.14223522e-01 1.03049409e+00
-6.88860059e-01 -2.37919822e-01 -5.98842323e-01 8.76015961e-01
8.05338442e-01 6.57095730e-01 3.89058053e-01 -4.08403218e-01
2.08252907e-01 -3.69671099e-02 3.40077937e-01 -5.22761106e-01
-6.42147839e-01 -3.77617687e-01 -9.44876596e-02 -2.06199944e-01
-5.12456417e-01 -2.31567815e-01 -8.42227161e-01 -2.33883068e-01
-2.74178147e-01 3.63407545e-02 7.43439615e-01 4.51398373e-01
5.56741178e-01 7.18169928e-01 5.40475488e-01 -3.51177365e-01
-8.69614780e-02 -1.33758342e+00 -5.13451815e-01 5.43566227e-01
5.89091420e-01 -3.74683827e-01 -5.53247631e-01 4.92291927e-01] | [15.60062313079834, -3.317481517791748] |
3309ba12-d4c6-4d31-9dd3-17700d67434d | learning-adaptive-discriminative-correlation | 1807.11348 | null | https://arxiv.org/abs/1807.11348v3 | https://arxiv.org/pdf/1807.11348v3.pdf | Learning Adaptive Discriminative Correlation Filters via Temporal Consistency Preserving Spatial Feature Selection for Robust Visual Tracking | With efficient appearance learning models, Discriminative Correlation Filter (DCF) has been proven to be very successful in recent video object tracking benchmarks and competitions. However, the existing DCF paradigm suffers from two major issues, i.e., spatial boundary effect and temporal filter degradation. To mitigate these challenges, we propose a new DCF-based tracking method. The key innovations of the proposed method include adaptive spatial feature selection and temporal consistent constraints, with which the new tracker enables joint spatial-temporal filter learning in a lower dimensional discriminative manifold. More specifically, we apply structured spatial sparsity constraints to multi-channel filers. Consequently, the process of learning spatial filters can be approximated by the lasso regularisation. To encourage temporal consistency, the filter model is restricted to lie around its historical value and updated locally to preserve the global structure in the manifold. Last, a unified optimisation framework is proposed to jointly select temporal consistency preserving spatial features and learn discriminative filters with the augmented Lagrangian method. Qualitative and quantitative evaluations have been conducted on a number of well-known benchmarking datasets such as OTB2013, OTB50, OTB100, Temple-Colour, UAV123 and VOT2018. The experimental results demonstrate the superiority of the proposed method over the state-of-the-art approaches. | ['Josef Kittler', 'Xiao-Jun Wu', 'Zhen-Hua Feng', 'Tianyang Xu'] | 2018-07-30 | null | null | null | null | ['video-object-tracking'] | ['computer-vision'] | [-1.81606203e-01 -6.87737167e-01 -2.71243185e-01 5.99745922e-02
-4.04058784e-01 -3.47767800e-01 6.19959950e-01 -3.57723206e-01
-4.87766474e-01 7.17713594e-01 4.31782231e-02 3.76887560e-01
-4.39956725e-01 -9.04640704e-02 -6.72318101e-01 -8.94217670e-01
-2.34033659e-01 -1.74406752e-01 4.04281855e-01 1.88424170e-01
6.46522492e-02 4.73709732e-01 -1.38289285e+00 -1.59764349e-01
9.60348010e-01 1.21339405e+00 3.88191104e-01 1.75373673e-01
2.53425390e-01 5.51321149e-01 -1.64318115e-01 -2.21956968e-01
5.57528913e-01 -1.14378773e-01 -3.32797199e-01 3.65865022e-01
7.89533079e-01 -1.98154539e-01 -6.63660526e-01 1.29069614e+00
3.30347419e-01 3.81708622e-01 3.88285786e-01 -1.39743865e+00
-5.33892572e-01 -6.58179075e-02 -9.17464733e-01 3.58064920e-01
6.36813045e-02 3.62008184e-01 7.06314325e-01 -1.10739326e+00
7.60980248e-01 1.34744191e+00 8.00853550e-01 3.40323806e-01
-1.31107354e+00 -7.98712552e-01 6.38632119e-01 3.37652951e-01
-1.47720587e+00 -3.33523601e-01 8.01015019e-01 -7.59643197e-01
3.85068834e-01 2.16129929e-01 8.90778542e-01 9.49546158e-01
2.93877870e-01 8.36798608e-01 1.23622537e+00 -1.60278291e-01
-5.97962961e-02 4.79044169e-02 2.72654146e-02 8.20270956e-01
1.57401666e-01 5.48190653e-01 -6.82519317e-01 -8.39112103e-02
8.78951848e-01 2.10152015e-01 -4.18822974e-01 -8.19451630e-01
-1.30992830e+00 6.29282176e-01 4.60168213e-01 9.70188305e-02
-3.12355965e-01 1.89651251e-01 3.88174057e-01 1.05096586e-01
5.55217147e-01 -3.44951563e-02 -3.78765732e-01 1.88049376e-02
-9.17213142e-01 2.57228017e-01 2.14695245e-01 1.01575685e+00
5.40945172e-01 1.72440156e-01 -4.33630854e-01 8.13347280e-01
8.72424841e-01 6.39921844e-01 1.15369052e-01 -8.95572662e-01
3.79153579e-01 3.22696477e-01 3.32169175e-01 -1.30595779e+00
-1.58005580e-01 -6.36286378e-01 -9.10065532e-01 3.06041986e-01
4.15914536e-01 9.52052325e-02 -7.94665635e-01 1.66415405e+00
5.82140148e-01 8.08706760e-01 -2.93704391e-01 1.28391898e+00
6.38634562e-01 5.72511852e-01 2.10937098e-01 -5.09206653e-01
1.07865500e+00 -1.05813026e+00 -9.21035409e-01 1.25647634e-02
2.21846536e-01 -9.41894650e-01 5.50590158e-01 3.68572742e-01
-7.97983766e-01 -9.13644910e-01 -1.02496779e+00 4.07017022e-01
5.31158969e-03 5.85548580e-01 5.90779305e-01 3.97086531e-01
-7.16139793e-01 4.80915040e-01 -9.36048925e-01 -3.13778758e-01
4.53737974e-01 1.94986209e-01 -3.34372997e-01 -7.35827461e-02
-9.59678233e-01 6.73579097e-01 3.58756721e-01 4.67235804e-01
-9.97365475e-01 -7.39027202e-01 -6.69286609e-01 -3.34324270e-01
4.32001472e-01 -5.22640646e-01 6.03418112e-01 -9.12337959e-01
-1.37877560e+00 5.27392089e-01 -8.51945430e-02 -4.35367256e-01
6.50685072e-01 -5.36292493e-01 -4.92797613e-01 -1.62048743e-03
1.21192075e-01 4.67216492e-01 1.20294178e+00 -1.25730395e+00
-8.67946148e-01 -2.76401132e-01 -2.11599156e-01 1.01154834e-01
-5.09065568e-01 1.36649907e-01 -8.34513724e-01 -1.08556509e+00
1.36212364e-01 -1.02963340e+00 -1.97674707e-01 5.00620127e-01
-7.12816343e-02 -2.86173016e-01 1.15853441e+00 -8.07988465e-01
1.45056641e+00 -2.43974805e+00 4.61050779e-01 1.83565423e-01
1.44284423e-02 3.91928673e-01 -1.68221399e-01 -3.67772840e-02
1.72656834e-01 -3.95881414e-01 6.37804717e-02 -4.61390883e-01
-1.23207353e-01 1.58715341e-02 -1.21447049e-01 1.03038549e+00
1.66627079e-01 6.71889424e-01 -9.82467532e-01 -6.26340568e-01
6.10271513e-01 6.44645989e-01 -5.00559568e-01 1.03658266e-01
-1.36518672e-01 8.68062377e-01 -6.69441879e-01 7.22223759e-01
9.30067897e-01 -4.54651713e-02 -1.71394393e-01 -5.40629268e-01
-5.65747917e-01 -3.55753839e-01 -1.58991456e+00 1.94041240e+00
-7.19927996e-02 4.48640078e-01 3.48649621e-01 -8.28932524e-01
8.76875222e-01 1.29439428e-01 8.03490579e-01 -6.40548527e-01
-6.03472479e-02 1.70941353e-01 -2.46122986e-01 -4.89829421e-01
2.75622934e-01 2.63133049e-01 3.51790100e-01 -3.12523961e-01
1.66308343e-01 4.19405967e-01 2.19291463e-01 8.76385197e-02
6.46309555e-01 7.43426800e-01 -8.85474384e-02 -5.34787118e-01
9.35369313e-01 -9.61257517e-02 1.14539576e+00 5.34205079e-01
-4.63192940e-01 3.62596452e-01 -1.24603458e-01 -4.67531949e-01
-6.99901879e-01 -9.64009762e-01 -2.99156904e-01 7.31850684e-01
5.84156394e-01 -4.81325388e-01 -2.25159481e-01 -7.66505361e-01
1.96132258e-01 9.32822227e-02 -5.86165607e-01 -1.56807989e-01
-6.64247930e-01 -5.32261670e-01 -1.98297892e-02 5.06564796e-01
7.65581548e-01 -7.56298482e-01 -4.62596864e-01 3.03424984e-01
2.72191726e-02 -1.10976636e+00 -9.15998101e-01 -2.88628280e-01
-8.55192602e-01 -1.08007646e+00 -7.75300562e-01 -7.71308899e-01
5.96511006e-01 5.79461455e-01 5.80496490e-01 5.37888817e-02
-4.24710691e-01 6.25007331e-01 -4.08294946e-01 4.78708297e-02
2.38521308e-01 -3.99782598e-01 2.17706889e-01 5.45404971e-01
-1.18480977e-02 -3.04376811e-01 -8.48134398e-01 6.63741589e-01
-5.50022483e-01 -1.00712880e-01 6.62209332e-01 1.20875239e+00
8.90388310e-01 5.99009842e-02 2.26481199e-01 -1.54614478e-01
-2.86017847e-03 -3.50318223e-01 -1.05887187e+00 3.07613105e-01
-4.82732743e-01 -2.34470651e-01 4.26120281e-01 -7.23438740e-01
-1.19218516e+00 3.51200193e-01 4.28420633e-01 -9.64442194e-01
2.72529155e-01 3.38580012e-01 -2.37871066e-01 -5.00421286e-01
1.34542942e-01 3.46888751e-01 7.13263676e-02 -6.23663008e-01
2.51425236e-01 2.20769525e-01 4.91023391e-01 -5.18242836e-01
1.17272604e+00 7.37008572e-01 7.25909919e-02 -7.70532250e-01
-7.18066216e-01 -7.56630123e-01 -6.82973444e-01 -6.56867087e-01
9.15887058e-01 -1.13649333e+00 -6.02200449e-01 6.06292605e-01
-1.00597560e+00 -7.76469335e-02 -1.09132767e-01 9.23552811e-01
-3.63483936e-01 6.24838710e-01 -2.97996342e-01 -8.59417856e-01
-2.37413689e-01 -1.07556367e+00 9.80986834e-01 3.89263570e-01
1.95915371e-01 -8.49930167e-01 7.96683058e-02 9.18456689e-02
2.84188181e-01 3.29277277e-01 1.47713378e-01 3.07254754e-02
-8.78161132e-01 6.00178689e-02 -2.73465097e-01 4.58990991e-01
7.95833617e-02 1.13737658e-01 -5.27379572e-01 -8.20014894e-01
-1.48339599e-01 -1.44154206e-01 9.74064291e-01 4.60965514e-01
8.72305393e-01 -7.75234476e-02 -6.34243846e-01 9.32895184e-01
1.48148453e+00 1.27304107e-01 3.23203623e-01 5.93015611e-01
8.04266751e-01 2.76541650e-01 1.23220801e+00 5.63742340e-01
1.14097878e-01 1.13224590e+00 2.93647408e-01 1.21672310e-01
-2.94435680e-01 -9.63306576e-02 5.49140155e-01 6.00669563e-01
-2.47165009e-01 2.89034426e-01 -4.26245034e-01 5.67796052e-01
-2.20350575e+00 -9.11500812e-01 -2.18764052e-01 2.33949590e+00
4.71033365e-01 2.29772031e-02 1.22397706e-01 -2.44436190e-01
6.98228359e-01 2.73558587e-01 -3.05520266e-01 4.97894049e-01
-2.43691728e-01 -2.42627144e-01 4.75007564e-01 1.64879024e-01
-1.58923018e+00 7.71931410e-01 4.93133688e+00 1.04431319e+00
-1.19695294e+00 3.13526541e-01 5.70922121e-02 -1.17540590e-01
2.07028106e-01 1.45591451e-02 -9.46767151e-01 8.40646863e-01
2.23786071e-01 6.07278198e-02 3.80969107e-01 6.46724463e-01
4.35462892e-01 5.63736036e-02 -5.85616350e-01 1.25785708e+00
-2.91269366e-02 -1.22577822e+00 -3.44433159e-01 -5.45239262e-03
8.45718741e-01 -7.97968283e-02 2.95943767e-01 2.37965539e-01
-8.37934762e-03 -6.37764156e-01 9.61219728e-01 8.62052023e-01
5.78749359e-01 -5.37623405e-01 4.22669083e-01 2.70386860e-02
-1.77498627e+00 -2.51138657e-01 -5.07011354e-01 3.88900578e-01
1.84530422e-01 3.84358078e-01 9.49638411e-02 8.54142308e-01
1.12538218e+00 1.39768481e+00 -6.18878543e-01 1.54578769e+00
-1.45447515e-02 4.28508908e-01 -4.09687757e-01 1.94706097e-01
2.29224637e-01 -5.23781180e-01 9.03850198e-01 9.97704327e-01
2.54964679e-01 -7.02844858e-02 5.64912677e-01 7.00985909e-01
3.44146758e-01 1.27893955e-01 -2.92638332e-01 2.88137823e-01
4.10303593e-01 1.48687911e+00 -4.90647763e-01 -8.27268511e-02
-7.90584564e-01 8.28638613e-01 1.27478868e-01 5.00496566e-01
-1.20402610e+00 2.83448786e-01 6.64972782e-01 4.91161793e-02
6.81556761e-01 -4.62128669e-01 2.91604549e-01 -1.26367974e+00
2.74393916e-01 -7.98485816e-01 3.41205269e-01 -4.02232438e-01
-1.27336919e+00 3.77571046e-01 -8.17441419e-02 -1.80265212e+00
4.71340954e-01 -4.76037532e-01 -5.74073136e-01 7.26297796e-01
-1.50935495e+00 -1.41195190e+00 -4.39920515e-01 8.22270393e-01
5.63857794e-01 -3.99445266e-01 2.42247120e-01 7.85435319e-01
-7.78676152e-01 4.68494862e-01 4.21099514e-01 5.52424788e-02
7.94418812e-01 -9.57737684e-01 -1.39137894e-01 1.07122827e+00
7.26770759e-02 6.38016999e-01 6.07874036e-01 -7.51708150e-01
-1.78982449e+00 -1.44004798e+00 2.83033162e-01 -1.89359859e-01
7.07430303e-01 -2.44519800e-01 -8.02529335e-01 5.00063479e-01
5.72425760e-02 6.85643315e-01 1.18905537e-01 -2.15233549e-01
-1.97268277e-01 -3.36755693e-01 -8.01036119e-01 2.79445797e-01
1.16905260e+00 -1.51620910e-01 -2.49449104e-01 3.19805145e-01
3.05201977e-01 -3.73916477e-01 -9.38512802e-01 7.14422584e-01
7.28735566e-01 -6.48512065e-01 1.08020830e+00 -4.16615963e-01
-3.22372794e-01 -9.82265770e-01 -8.75792503e-02 -1.15550995e+00
-6.25616312e-01 -7.82718599e-01 -4.93229628e-01 1.49675977e+00
-1.89775929e-01 -6.07051253e-01 5.95845461e-01 1.58095837e-01
-1.15804747e-01 -7.57873237e-01 -1.25078487e+00 -1.22413349e+00
-3.89247507e-01 -8.10983554e-02 3.74605432e-02 6.80329978e-01
-5.13073862e-01 -1.41098782e-01 -8.03500712e-01 2.58576602e-01
1.12600589e+00 1.58457488e-01 8.45691323e-01 -1.18414629e+00
-3.07249993e-01 -2.53146797e-01 -5.09482503e-01 -1.45972896e+00
1.15952581e-01 -6.53546810e-01 -9.59694479e-03 -1.01069629e+00
2.53188014e-01 -6.57914162e-01 -4.50189292e-01 2.93096136e-02
-3.32056642e-01 1.66426748e-01 4.55214739e-01 4.50194329e-01
-9.40460324e-01 9.57569718e-01 1.44375885e+00 -3.41265857e-01
-4.38187830e-02 2.84495251e-03 2.45436262e-02 5.59938312e-01
3.57882977e-01 -3.51972729e-01 -1.69562221e-01 -2.95774639e-01
-1.43444702e-01 -2.01036528e-01 7.88780510e-01 -9.83155906e-01
3.47342253e-01 -1.81117550e-01 6.04774833e-01 -7.91519642e-01
3.56373459e-01 -1.10569358e+00 3.82474601e-01 5.52289367e-01
8.67535844e-02 -6.69557229e-02 2.63147712e-01 1.04113698e+00
-3.90679389e-01 1.45177528e-01 9.75124538e-01 2.79467642e-01
-1.19559562e+00 6.92100048e-01 2.35083271e-02 -1.14176400e-01
1.25328302e+00 -2.67905086e-01 7.61707947e-02 1.20199099e-01
-6.90473616e-01 5.57378352e-01 4.05464411e-01 7.34637141e-01
6.96660757e-01 -1.69108272e+00 -7.04280853e-01 1.47352636e-01
4.95706201e-02 -2.45492950e-01 4.70133483e-01 1.43068779e+00
-7.97645301e-02 4.10401136e-01 -2.11433455e-01 -9.95635152e-01
-1.37351751e+00 7.65594125e-01 3.57458293e-01 -1.82242364e-01
-7.92789340e-01 7.56329477e-01 2.29008704e-01 -4.55554761e-03
5.11405349e-01 -7.96773881e-02 -2.59381533e-01 3.92977186e-02
3.88145238e-01 4.63267773e-01 -3.66832465e-01 -9.45159197e-01
-6.10807002e-01 1.00876904e+00 8.26694667e-02 2.64065802e-01
1.27073705e+00 -3.52309376e-01 -4.50612828e-02 3.14763561e-02
1.04905343e+00 -2.93564759e-02 -1.94290459e+00 -4.24551755e-01
-1.33307930e-03 -9.47206318e-01 1.15937836e-01 -3.77975166e-01
-1.44254756e+00 4.71673042e-01 1.02162993e+00 -1.29139349e-01
1.10366476e+00 -2.47664645e-01 6.40218616e-01 -1.52954519e-01
5.41049182e-01 -1.12354279e+00 2.75927544e-01 3.40939641e-01
1.02337480e+00 -1.29630542e+00 2.51235336e-01 -5.67697883e-01
-3.20234686e-01 9.65639532e-01 8.58830810e-01 -3.26246232e-01
7.94640839e-01 -1.10594593e-02 -1.98897138e-01 -8.53901058e-02
-4.53937262e-01 -2.29266435e-01 7.55391657e-01 4.20512736e-01
1.88509017e-01 -2.19221115e-01 -5.22779584e-01 3.15346003e-01
6.12613440e-01 -4.25676405e-02 -2.52445132e-01 8.64716113e-01
-1.19818270e-01 -9.40124631e-01 -6.07798815e-01 1.88814387e-01
-3.43891710e-01 2.05504745e-01 9.53057259e-02 8.74113262e-01
4.20009941e-01 7.62776494e-01 -2.57783234e-01 -2.16137588e-01
3.70668948e-01 -4.27830160e-01 6.15827858e-01 -1.80233911e-01
-3.43648821e-01 5.88545561e-01 -2.44979560e-01 -8.71614277e-01
-8.51643384e-01 -1.12570775e+00 -8.68384063e-01 2.19535857e-01
-7.01552629e-01 1.00625180e-01 3.75793576e-01 7.21107900e-01
3.20590854e-01 4.88265872e-01 6.89140618e-01 -9.33165431e-01
-5.97288072e-01 -9.26763773e-01 -5.43928802e-01 5.93931854e-01
3.33901137e-01 -1.34381032e+00 -2.32222602e-01 9.11741331e-02] | [6.331315994262695, -2.182359457015991] |
5d560886-424a-4384-829f-ef4312c2968f | revisiting-rubik-s-cube-self-supervised | 2007.08826 | null | https://arxiv.org/abs/2007.08826v1 | https://arxiv.org/pdf/2007.08826v1.pdf | Revisiting Rubik's Cube: Self-supervised Learning with Volume-wise Transformation for 3D Medical Image Segmentation | Deep learning highly relies on the quantity of annotated data. However, the annotations for 3D volumetric medical data require experienced physicians to spend hours or even days for investigation. Self-supervised learning is a potential solution to get rid of the strong requirement of training data by deeply exploiting raw data information. In this paper, we propose a novel self-supervised learning framework for volumetric medical images. Specifically, we propose a context restoration task, i.e., Rubik's cube++, to pre-train 3D neural networks. Different from the existing context-restoration-based approaches, we adopt a volume-wise transformation for context permutation, which encourages network to better exploit the inherent 3D anatomical information of organs. Compared to the strategy of training from scratch, fine-tuning from the Rubik's cube++ pre-trained weight can achieve better performance in various tasks such as pancreas segmentation and brain tissue segmentation. The experimental results show that our self-supervised learning method can significantly improve the accuracy of 3D deep learning networks on volumetric medical datasets without the use of extra data. | ['Xing Tao', 'Yefeng Zheng', 'Wenhui Zhou', 'Yuexiang Li', 'Kai Ma'] | 2020-07-17 | null | null | null | null | ['rubik-s-cube', 'pancreas-segmentation'] | ['graphs', 'medical'] | [ 2.41280809e-01 4.60505664e-01 -8.64741728e-02 -7.13805437e-01
-5.24901688e-01 -1.70611128e-01 2.24250436e-01 3.42592895e-01
-5.39491832e-01 5.73504627e-01 1.75180286e-01 -3.69931579e-01
5.68033867e-02 -7.55092919e-01 -7.29046643e-01 -8.52704406e-01
-8.94748345e-02 3.70771825e-01 3.03960383e-01 4.84785773e-02
-1.09987661e-01 3.57216090e-01 -9.64958727e-01 3.05557288e-02
1.03028190e+00 1.17406368e+00 6.09654903e-01 1.81255206e-01
-5.55482745e-01 8.08109581e-01 -2.90586859e-01 5.04525006e-02
3.14364970e-01 -3.21251273e-01 -6.86992288e-01 3.41390789e-01
1.92780793e-02 -2.85456359e-01 -1.61331564e-01 1.06804407e+00
6.48173034e-01 -1.21136419e-01 5.33117294e-01 -7.82626748e-01
-4.80169117e-01 5.82527041e-01 -5.87528408e-01 3.61860782e-01
-1.91874519e-01 2.41843522e-01 5.74294388e-01 -8.02481771e-01
4.75103438e-01 7.76383400e-01 7.08603501e-01 6.57450855e-01
-1.01132405e+00 -6.48190737e-01 9.80992317e-02 -7.44298771e-02
-1.29205430e+00 -1.40149474e-01 1.18325770e+00 -5.53796589e-01
5.42566478e-01 1.06365658e-01 9.09492195e-01 8.74464691e-01
5.37513644e-02 8.94520402e-01 1.07041609e+00 -2.78643221e-01
3.04793417e-01 7.96468183e-03 9.59653407e-02 8.62172008e-01
2.36851230e-01 4.11429666e-02 -3.68039235e-02 -4.73020822e-02
9.94503438e-01 2.45897084e-01 -3.05758804e-01 -9.14152920e-01
-1.20278180e+00 7.68112481e-01 7.40702331e-01 3.72804046e-01
-3.36771101e-01 -6.06540665e-02 6.26995683e-01 -1.07601427e-01
7.72598147e-01 3.41937512e-01 -6.94618642e-01 3.04650396e-01
-9.61543858e-01 -2.57265061e-01 3.88543814e-01 7.85716116e-01
6.28675997e-01 2.66663861e-02 -2.18363747e-01 7.78497338e-01
4.10689026e-01 9.60653424e-02 7.31226265e-01 -4.45919007e-01
3.93082231e-01 1.02025604e+00 -3.48095566e-01 -5.98703921e-01
-7.91639864e-01 -7.08141804e-01 -1.14621246e+00 1.42173737e-01
3.90862912e-01 -2.24443719e-01 -1.21228480e+00 1.50948894e+00
6.42116964e-01 1.98728517e-01 1.25279445e-02 1.05991459e+00
1.22236717e+00 3.29415351e-01 1.56969726e-01 -3.47786605e-01
1.10328436e+00 -1.00886965e+00 -4.88428295e-01 -1.22559234e-01
8.42269063e-01 -4.24979895e-01 1.16257679e+00 6.10392317e-02
-8.26286316e-01 -3.66698563e-01 -9.26873922e-01 -1.50037999e-03
-3.42797153e-02 3.76182757e-02 9.68644857e-01 4.98673886e-01
-7.35851467e-01 5.49245059e-01 -1.03651428e+00 -2.55226474e-02
1.06691587e+00 3.11568439e-01 -3.41346264e-01 -1.42433390e-01
-8.88729274e-01 7.97567070e-01 5.48605442e-01 2.56870747e-01
-8.71901751e-01 -9.29812074e-01 -9.75375772e-01 -1.02678753e-01
5.48079967e-01 -6.55501783e-01 1.05268276e+00 -7.77854025e-01
-1.33365130e+00 1.07441282e+00 1.96911171e-01 -3.93878460e-01
5.84622741e-01 2.89656818e-02 4.86179180e-02 2.52583832e-01
5.36539480e-02 6.88189030e-01 7.50734985e-01 -1.23064399e+00
-2.20431685e-01 -6.24551952e-01 -1.34201512e-01 5.25205247e-02
-3.75980794e-01 -4.23411071e-01 -4.70918804e-01 -7.16167271e-01
4.61774319e-01 -8.07363212e-01 -6.17432952e-01 2.16674283e-01
-4.40693349e-01 -2.39518672e-01 7.41924882e-01 -5.77696145e-01
7.91727126e-01 -2.13661075e+00 2.68065576e-02 3.11037719e-01
6.21713400e-01 2.42065609e-01 1.17069393e-01 -3.98809254e-01
-1.59583226e-01 -4.72547784e-02 -6.73139215e-01 -2.21767440e-01
-3.45560700e-01 4.21581447e-01 3.98102641e-01 3.57890695e-01
3.46969813e-01 1.10186362e+00 -1.08758891e+00 -8.72814894e-01
2.58161068e-01 5.09086430e-01 -6.38941884e-01 2.60396481e-01
-3.13299686e-01 1.10184240e+00 -5.38195729e-01 5.23964047e-01
6.65012717e-01 -5.90571105e-01 1.37554139e-01 -3.92773092e-01
1.53618097e-01 1.51875272e-01 -5.89742959e-01 2.21829057e+00
-5.39104760e-01 2.03916311e-01 6.13439158e-02 -1.45480025e+00
9.90837872e-01 3.11646819e-01 7.96297371e-01 -8.37134004e-01
3.22632670e-01 2.65093267e-01 -2.52475068e-02 -7.56599426e-01
-1.54091865e-01 -2.77382821e-01 2.92199024e-04 2.04380736e-01
1.91747248e-01 -8.61173943e-02 2.86776498e-02 1.06703721e-01
1.04453123e+00 2.00505167e-01 2.67717510e-01 -4.35117632e-01
5.10181606e-01 6.68305382e-02 8.66875648e-01 3.24348301e-01
-3.05438071e-01 4.64236230e-01 4.91755635e-01 -6.52384698e-01
-9.28366780e-01 -8.93184960e-01 -2.55749375e-01 7.50448287e-01
3.19859870e-02 -1.15836024e-01 -7.38345206e-01 -1.04718387e+00
7.91875925e-03 3.53041172e-01 -7.70408034e-01 -1.52005404e-01
-6.68465793e-01 -1.04211473e+00 1.34632304e-01 7.50446200e-01
6.32852614e-01 -8.95195127e-01 -6.96103752e-01 2.80454397e-01
-4.28947434e-02 -1.11716902e+00 -4.78859097e-01 4.47902054e-01
-1.44229627e+00 -1.06267428e+00 -1.02123749e+00 -9.81704593e-01
1.13044369e+00 7.03643337e-02 1.04440188e+00 3.08841765e-01
-5.13362706e-01 2.40403399e-01 -3.33995640e-01 -4.06791538e-01
-1.87594801e-01 1.60577029e-01 -3.66162956e-01 -2.07612649e-01
1.52417436e-01 -8.20359468e-01 -9.33837771e-01 9.48659033e-02
-8.42518449e-01 3.14413965e-01 7.51936436e-01 9.38259065e-01
9.66023266e-01 -5.19974977e-02 6.44349813e-01 -1.14921129e+00
3.21334392e-01 -3.88400644e-01 -4.47721183e-01 9.84891355e-02
-5.55582523e-01 1.58126101e-01 4.55376357e-01 -4.18312073e-01
-8.79003763e-01 4.31957603e-01 -2.08142430e-01 -4.80509222e-01
-7.58408755e-02 5.28806865e-01 -1.72166631e-01 -1.62274033e-01
5.38126528e-01 2.91770369e-01 1.45481646e-01 -5.84410489e-01
1.20014817e-01 3.51271272e-01 4.39489633e-01 -4.46254015e-01
6.23870313e-01 4.70570862e-01 1.06000155e-01 -3.29235345e-01
-1.06530702e+00 -3.28756511e-01 -8.29581082e-01 -9.79582518e-02
1.09722865e+00 -8.27494979e-01 -3.73421341e-01 2.04035968e-01
-9.05886590e-01 -5.48270524e-01 -4.59117740e-01 5.86950660e-01
-2.91965336e-01 4.36860472e-01 -5.23877501e-01 -4.27300185e-01
-6.37196839e-01 -1.21041846e+00 8.39002311e-01 1.75110206e-01
2.32866913e-01 -1.03930283e+00 -1.08651407e-01 3.87001246e-01
4.55917269e-01 4.46716726e-01 1.05471098e+00 -7.23196268e-01
-5.21852970e-01 -4.03126441e-02 -4.45100546e-01 4.85425711e-01
4.62067574e-01 -7.53978789e-01 -7.70636797e-01 -1.76825717e-01
2.23748192e-01 -2.74739534e-01 8.49049509e-01 4.35896367e-01
1.75786185e+00 -2.98616022e-01 -2.26402596e-01 7.94941366e-01
1.32264185e+00 8.59992653e-02 2.25818843e-01 2.01367036e-01
1.03693616e+00 4.62432683e-01 3.47104728e-01 4.34054762e-01
6.08412564e-01 2.42927909e-01 6.91666067e-01 -5.47202528e-01
-1.35112539e-01 -7.76126683e-02 -4.62568671e-01 1.10561252e+00
-1.18481636e-01 3.83280426e-01 -1.15981936e+00 4.72434103e-01
-1.64050794e+00 -3.85298699e-01 1.16035249e-02 2.03483248e+00
1.11898899e+00 3.14326823e-01 -7.44783282e-02 9.18325707e-02
4.82565641e-01 1.08825276e-02 -7.38643467e-01 2.02754632e-01
2.71818250e-01 2.31100589e-01 5.37337124e-01 1.95268959e-01
-1.24352801e+00 6.66261077e-01 5.35525894e+00 6.92240894e-01
-1.34097934e+00 3.25536013e-01 7.26636410e-01 -9.53848734e-02
-2.80124098e-01 -3.52968454e-01 -4.92491394e-01 6.51401401e-01
2.79308051e-01 1.58083245e-01 1.24905892e-01 9.30150688e-01
2.18949363e-01 -1.14204539e-02 -1.18893695e+00 1.12676060e+00
5.78522906e-02 -1.29172504e+00 -3.95908877e-02 -1.20760813e-01
7.21822977e-01 9.99986194e-03 -2.17847377e-01 3.51739317e-01
7.67797465e-03 -8.69351149e-01 3.20681810e-01 4.08594429e-01
6.45973206e-01 -5.02249539e-01 7.52237201e-01 4.73445535e-01
-9.85634208e-01 9.10557136e-02 -3.89088362e-01 3.25841397e-01
4.21452411e-02 1.12110519e+00 -1.03554928e+00 4.50317353e-01
6.69578433e-01 6.28335476e-01 -5.52332819e-01 1.18499720e+00
-1.14143088e-01 6.00421011e-01 -2.33197719e-01 1.38954937e-01
2.38848642e-01 -1.15966946e-02 3.15175623e-01 1.07395756e+00
9.02062356e-02 2.00866386e-01 4.76569951e-01 7.99322486e-01
-1.54480070e-01 3.13702524e-01 -3.52956176e-01 2.67631888e-01
-1.14488313e-02 1.34329331e+00 -9.18527782e-01 -4.22779500e-01
-3.66930366e-01 8.19369376e-01 2.59989619e-01 -1.41039304e-02
-8.03565025e-01 -1.07830288e-02 -1.09586500e-01 2.48988643e-01
3.12288135e-01 -2.53893495e-01 -5.09054720e-01 -1.06270456e+00
8.20310935e-02 -4.14963245e-01 2.98698574e-01 -5.44610858e-01
-1.38436747e+00 5.48683524e-01 -1.58740878e-01 -1.25757623e+00
9.42091867e-02 -4.93451238e-01 -6.50971711e-01 5.46219587e-01
-1.94027114e+00 -1.16224670e+00 -6.30398393e-01 7.32352376e-01
3.25976551e-01 -6.55965060e-02 6.71919048e-01 4.26868141e-01
-4.90911990e-01 4.32340622e-01 -1.51067480e-01 4.17084336e-01
5.80201983e-01 -1.39135170e+00 -5.70101431e-03 3.47309202e-01
-1.71118185e-01 3.89660925e-01 2.38441482e-01 -6.36844695e-01
-1.23104525e+00 -1.24074459e+00 4.46780801e-01 -5.99825755e-02
3.40853453e-01 -2.10595205e-01 -1.07727540e+00 4.77661759e-01
-7.99398422e-02 7.97813892e-01 9.17986393e-01 -5.59680238e-02
-1.83565438e-01 -3.67199212e-01 -1.38536060e+00 3.53233695e-01
1.29428899e+00 -2.48260379e-01 -7.14035153e-01 5.90235770e-01
9.83384550e-01 -7.13425100e-01 -1.11251211e+00 7.91537642e-01
1.96768537e-01 -5.99629641e-01 1.03793323e+00 -5.22063255e-01
5.51892400e-01 -1.99981928e-01 1.12500340e-01 -1.17853391e+00
-3.44305336e-02 -2.19895929e-01 -1.07633308e-01 9.05269444e-01
3.53146702e-01 -5.05493164e-01 1.01949525e+00 6.45852089e-01
-4.98701811e-01 -1.07696474e+00 -1.07425714e+00 -6.30333543e-01
2.24723220e-01 -2.25337535e-01 6.15932524e-01 1.22105670e+00
-8.00992027e-02 1.98048770e-01 -4.80385609e-02 1.76682547e-02
9.16289866e-01 1.45723641e-01 5.67107141e-01 -1.28525114e+00
-2.60517895e-01 -1.98898762e-01 -4.43251729e-01 -9.46624398e-01
-5.68628963e-03 -1.28715742e+00 5.34849614e-02 -1.73533118e+00
3.37444961e-01 -8.10403049e-01 -6.36354148e-01 6.56288028e-01
-2.65037686e-01 2.47685313e-01 -6.67508971e-03 1.57354280e-01
-7.67175734e-01 6.41152680e-01 1.93588734e+00 -3.72631758e-01
-3.35995495e-01 -9.16388351e-03 -7.77669728e-01 8.91953468e-01
9.82701361e-01 -5.16502857e-01 -4.91843283e-01 -6.31928265e-01
-1.18488289e-01 2.04378858e-01 5.17270386e-01 -8.28354836e-01
2.06104949e-01 3.59954797e-02 6.85191154e-01 -6.89734578e-01
2.26303060e-02 -9.35409248e-01 -3.52862418e-01 5.23510695e-01
-1.62251770e-01 -2.33488828e-01 2.00588882e-01 4.68344957e-01
-9.90270451e-02 -2.24945247e-01 7.30438769e-01 -6.67753339e-01
-5.28303862e-01 5.36196232e-01 -3.84261236e-02 1.45690575e-01
9.48802173e-01 -9.92578864e-02 6.09340146e-02 1.56804591e-01
-1.04064798e+00 3.81144643e-01 8.36902782e-02 1.08895630e-01
6.84442997e-01 -1.26362002e+00 -4.62225199e-01 1.70007780e-01
6.27842024e-02 7.37459064e-01 4.44530666e-01 9.00129676e-01
-3.93541753e-01 2.74167985e-01 -1.78075165e-01 -1.04704702e+00
-8.92903984e-01 6.27277553e-01 3.42484921e-01 -5.96507251e-01
-9.94845927e-01 8.39856982e-01 4.13101196e-01 -7.11804032e-01
3.07047457e-01 -6.32328510e-01 -2.14242443e-01 -1.97053194e-01
2.43900269e-01 -2.04261288e-01 1.61989912e-01 -1.72129333e-01
-4.76150513e-01 4.69536453e-01 -2.70922095e-01 3.61606628e-01
1.63474202e+00 1.17804006e-01 -2.05910075e-02 2.30464935e-01
1.34161413e+00 -3.34835321e-01 -1.46668589e+00 -4.49510962e-01
7.54351020e-02 -3.03800732e-01 3.48607928e-01 -8.90531778e-01
-1.57245648e+00 8.68169248e-01 7.89814532e-01 -1.94807395e-01
1.19164956e+00 1.02435924e-01 9.32728529e-01 3.28072220e-01
3.99807274e-01 -9.03548896e-01 1.98511034e-01 2.02526793e-01
7.85772383e-01 -1.55040848e+00 8.73226151e-02 -5.43846667e-01
-5.02069831e-01 9.29823160e-01 6.27155364e-01 -1.83097526e-01
8.39864135e-01 2.51547754e-01 9.13233161e-02 -2.49355748e-01
-2.47488007e-01 -2.09319681e-01 3.04722399e-01 6.54780149e-01
3.71640891e-01 2.91342754e-02 -3.27583581e-01 8.36286664e-01
-1.64055303e-02 2.24261269e-01 1.88310549e-01 1.00906479e+00
-3.83734405e-01 -1.01339650e+00 -2.18563929e-01 7.61796296e-01
-4.04393435e-01 -1.43171310e-01 9.33714285e-02 6.80496693e-01
2.72202551e-01 4.11135077e-01 -6.83559058e-03 -1.35365024e-01
2.73335427e-01 -2.18508253e-03 6.64246023e-01 -7.71802187e-01
-4.81223434e-01 1.74048275e-01 -3.49841982e-01 -2.76425421e-01
-6.57920659e-01 -3.86415094e-01 -1.53906918e+00 1.89237416e-01
-3.59180242e-01 -9.89343449e-02 6.32903337e-01 1.06537855e+00
1.32694215e-01 8.24594319e-01 6.74006164e-01 -8.26843202e-01
-4.47938353e-01 -8.46729577e-01 -2.86244303e-01 4.19928700e-01
2.73189634e-01 -8.01556528e-01 -6.12925962e-02 -1.22332489e-02] | [14.741047859191895, -2.2317774295806885] |
5e69468d-20b0-480c-8c75-61c97d9f4331 | qpgesture-quantization-based-and-phase-guided-1 | 2305.11094 | null | https://arxiv.org/abs/2305.11094v1 | https://arxiv.org/pdf/2305.11094v1.pdf | QPGesture: Quantization-Based and Phase-Guided Motion Matching for Natural Speech-Driven Gesture Generation | Speech-driven gesture generation is highly challenging due to the random jitters of human motion. In addition, there is an inherent asynchronous relationship between human speech and gestures. To tackle these challenges, we introduce a novel quantization-based and phase-guided motion-matching framework. Specifically, we first present a gesture VQ-VAE module to learn a codebook to summarize meaningful gesture units. With each code representing a unique gesture, random jittering problems are alleviated effectively. We then use Levenshtein distance to align diverse gestures with different speech. Levenshtein distance based on audio quantization as a similarity metric of corresponding speech of gestures helps match more appropriate gestures with speech, and solves the alignment problem of speech and gestures well. Moreover, we introduce phase to guide the optimal gesture matching based on the semantics of context or rhythm of audio. Phase guides when text-based or speech-based gestures should be performed to make the generated gestures more natural. Extensive experiments show that our method outperforms recent approaches on speech-driven gesture generation. Our code, database, pre-trained models, and demos are available at https://github.com/YoungSeng/QPGesture. | ['Haolin Zhuang', 'Weihong Bao', 'Lei Hao', 'Zhensong Zhang', 'Minglei Li', 'Zhiyong Wu', 'Sicheng Yang'] | 2023-05-18 | qpgesture-quantization-based-and-phase-guided | http://openaccess.thecvf.com//content/CVPR2023/html/Yang_QPGesture_Quantization-Based_and_Phase-Guided_Motion_Matching_for_Natural_Speech-Driven_Gesture_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Yang_QPGesture_Quantization-Based_and_Phase-Guided_Motion_Matching_for_Natural_Speech-Driven_Gesture_CVPR_2023_paper.pdf | cvpr-2023-1 | ['gesture-generation'] | ['robots'] | [ 1.02983743e-01 -5.17758548e-01 -2.25069299e-01 -4.38644618e-01
-1.05061483e+00 -6.03806794e-01 6.62153125e-01 -3.37592274e-01
-3.13629657e-01 1.17518224e-01 8.50825489e-01 1.83508590e-01
-2.81256866e-02 -3.72475713e-01 -3.01085562e-01 -7.82630205e-01
-9.00267586e-02 3.30508322e-01 4.24079806e-01 -2.30873495e-01
3.63305748e-01 3.11300516e-01 -1.55967176e+00 1.91926613e-01
5.87415457e-01 6.07223451e-01 4.71876204e-01 1.05105734e+00
-3.89975399e-01 4.53123510e-01 -8.92608285e-01 1.29276559e-01
2.56908059e-01 -8.52835476e-01 -5.04687548e-01 -7.71853775e-02
1.02400795e-01 -5.26301861e-01 -6.29646897e-01 8.32518280e-01
8.38926077e-01 4.18303698e-01 7.64305890e-01 -1.53962374e+00
-3.25035930e-01 7.58816957e-01 -1.51663005e-01 -8.07837993e-02
6.89719021e-01 3.61076504e-01 9.42271948e-01 -8.12822640e-01
6.71199262e-01 1.31458306e+00 3.23200077e-01 9.63910341e-01
-5.65013349e-01 -8.72084141e-01 1.15019716e-02 2.55409479e-01
-1.63281727e+00 -6.40628517e-01 7.04203665e-01 -3.25221926e-01
6.04755461e-01 5.96183240e-01 8.60762000e-01 1.20151591e+00
-4.13047850e-01 9.75885451e-01 1.76340833e-01 -3.00269276e-01
3.75275224e-01 -6.46221519e-01 -2.82712609e-01 3.22260588e-01
-3.99704397e-01 1.96948439e-01 -9.79213893e-01 -7.68798366e-02
8.40775311e-01 6.69959411e-02 -3.25047255e-01 -2.60617167e-01
-1.49828541e+00 5.50365627e-01 2.71892607e-01 3.52271497e-01
-2.82764018e-01 5.91394484e-01 2.94720888e-01 3.34461890e-02
-2.45919779e-01 3.23185734e-02 -1.89922117e-02 -7.93430686e-01
-1.08039629e+00 3.95174116e-01 5.95780730e-01 1.29630840e+00
4.68176782e-01 4.35088016e-02 -4.12034094e-01 7.54698396e-01
7.28918493e-01 8.17598581e-01 7.20679283e-01 -1.17473590e+00
6.01508915e-01 3.18522751e-01 -2.46305186e-02 -8.90027583e-01
-1.54486760e-01 3.67269307e-01 -6.96674526e-01 -1.07741825e-01
4.11958426e-01 -5.71925566e-02 -9.60667133e-01 1.65042996e+00
3.26285213e-01 4.04729545e-01 -8.62035379e-02 1.18599820e+00
7.65368223e-01 7.61505783e-01 -4.55259383e-02 -2.30886489e-01
1.15028393e+00 -7.88357079e-01 -9.94235992e-01 7.22087771e-02
4.55913216e-01 -1.00149727e+00 1.14049864e+00 3.42928246e-02
-8.19662333e-01 -4.62827682e-01 -9.11931694e-01 1.75650671e-01
2.32574075e-01 5.38614206e-02 2.22223297e-01 5.04224777e-01
-8.65906239e-01 3.70104700e-01 -1.15364909e+00 -4.50861365e-01
1.24432251e-01 2.86269516e-01 1.15275763e-01 2.17514724e-01
-1.07187903e+00 1.85966298e-01 2.54280657e-01 5.61310649e-02
-5.53704381e-01 -2.01126695e-01 -8.45376849e-01 -2.42333159e-01
1.25742286e-01 -2.79610783e-01 1.72386217e+00 -7.21411049e-01
-1.93022907e+00 2.83954561e-01 -4.60331768e-01 -1.28214285e-01
5.04174352e-01 -8.21028501e-02 -3.42943728e-01 4.65597808e-01
4.98664305e-02 9.47250545e-01 8.61772716e-01 -1.00540531e+00
-7.01204658e-01 -6.53875992e-02 -4.47813451e-01 5.90924025e-01
-3.25635523e-01 1.63090840e-01 -9.12015438e-01 -1.00115728e+00
3.85582745e-01 -1.00708878e+00 -5.59443049e-02 1.80747986e-01
-3.51036519e-01 -2.52389163e-01 8.90906930e-01 -4.24856395e-01
1.66546726e+00 -2.29925942e+00 2.02897400e-01 3.05124342e-01
-1.30704075e-01 1.35363758e-01 -4.17493761e-01 6.87003434e-01
4.42889869e-01 -1.33777354e-02 -1.58547387e-01 -1.40856266e-01
1.76366061e-01 3.90923470e-01 -2.80178875e-01 1.40157506e-01
-1.30803725e-02 8.35803390e-01 -1.05447423e+00 -7.45363832e-01
3.43022466e-01 5.08506179e-01 -5.43989420e-01 4.62379128e-01
-1.02736771e-01 8.02464128e-01 -6.70974433e-01 6.86659575e-01
2.22797468e-01 8.11534673e-02 8.99192095e-02 -1.57120317e-01
-8.81002024e-02 5.82500756e-01 -1.45258546e+00 2.13199186e+00
-2.99494475e-01 7.67956257e-01 6.02446198e-02 -4.67463255e-01
9.91787732e-01 4.66502935e-01 6.04976177e-01 -4.35183525e-01
6.99735731e-02 2.51465321e-01 -6.91485107e-02 -7.31306434e-01
4.39811826e-01 3.31919938e-01 -1.37010321e-01 6.31833255e-01
-2.65551001e-01 -5.41536689e-01 1.90156430e-01 1.61064804e-01
1.08519077e+00 2.78010786e-01 1.98455110e-01 2.74820447e-01
2.86066622e-01 -2.82411184e-02 5.78952789e-01 4.94735032e-01
-3.53716403e-01 1.02331769e+00 -4.37039416e-03 -1.92094356e-01
-7.74176478e-01 -1.10080099e+00 3.41420561e-01 1.21848297e+00
4.50089753e-01 -7.50390470e-01 -8.55433762e-01 -3.09083134e-01
-3.95112664e-01 3.75962734e-01 -1.08115114e-01 1.43868044e-01
-1.03614056e+00 -2.18262509e-01 7.75591552e-01 5.69683313e-01
3.65885407e-01 -1.24685204e+00 -6.99750483e-01 2.87022352e-01
-5.06809592e-01 -9.35985804e-01 -1.22938573e+00 -1.51542008e-01
-6.42848134e-01 -9.91119504e-01 -1.08698189e+00 -9.93262649e-01
3.61811817e-01 3.77406240e-01 6.12228096e-01 1.30260110e-01
-1.79964513e-01 2.62987345e-01 -9.03820872e-01 -1.06492929e-01
-5.08072078e-01 4.60646376e-02 4.07147035e-02 -2.21823826e-01
4.30170774e-01 -5.68254888e-01 -7.71677375e-01 5.47111034e-01
-8.83096457e-01 -3.21408808e-02 3.99988443e-01 7.23493934e-01
3.77954811e-01 -4.67150956e-01 6.90973476e-02 2.66284525e-01
6.62056684e-01 -1.19527042e-01 -1.74171701e-01 9.47890207e-02
-9.85512882e-02 2.07918867e-01 2.35823750e-01 -9.00810599e-01
-5.24446845e-01 4.42620963e-01 -2.62635469e-01 -4.94113684e-01
-2.28017047e-01 -8.69002938e-03 -2.29788750e-01 2.70075858e-01
5.41614175e-01 4.38935488e-01 4.49164733e-02 -3.25597882e-01
5.22240222e-01 1.29186571e+00 6.43394232e-01 -4.30523545e-01
8.66520584e-01 2.53940105e-01 -4.27125782e-01 -1.03462207e+00
7.63100386e-02 -5.69841683e-01 -6.16651595e-01 -3.59072536e-01
8.42796147e-01 -6.20603919e-01 -7.23167002e-01 5.40131390e-01
-1.34507143e+00 -6.44609511e-01 -8.43380243e-02 7.57296860e-01
-6.80128932e-01 5.29364765e-01 -5.54574907e-01 -9.61072564e-01
-4.61181670e-01 -1.43818128e+00 1.41842806e+00 3.46912593e-01
-6.83920979e-01 -3.64526510e-01 1.40947625e-01 6.38382807e-02
1.43480226e-01 6.76547457e-03 2.75657684e-01 -4.13329899e-01
-6.41095221e-01 1.11858889e-01 1.85604140e-01 -1.78669021e-01
5.69195569e-01 2.51603752e-01 -5.35204828e-01 -1.12239823e-01
-4.42452848e-01 -1.17435738e-01 4.22198951e-01 3.65553737e-01
6.85705006e-01 -4.52077806e-01 -2.99978882e-01 7.09188998e-01
6.88916981e-01 6.86558485e-01 4.80650157e-01 1.56305760e-01
7.45756030e-01 5.26522696e-01 8.54578853e-01 6.72949791e-01
5.15526652e-01 1.19614089e+00 1.59364387e-01 4.18275118e-01
-4.11670566e-01 -6.09300911e-01 6.02144659e-01 1.21783173e+00
6.38092905e-02 -3.40737969e-01 -8.45624566e-01 5.31331480e-01
-1.88221574e+00 -1.04890227e+00 -4.55845520e-02 1.99049592e+00
8.60921502e-01 -2.78367251e-01 4.44863558e-01 1.78467929e-01
9.31416333e-01 2.27683023e-01 -4.07597035e-01 -6.53413013e-02
9.50159281e-02 2.18154848e-01 2.36310184e-01 5.97826540e-01
-1.01531899e+00 1.17267525e+00 6.05550241e+00 9.81916249e-01
-1.23233271e+00 -3.13847363e-02 -6.51483685e-02 -3.66964132e-01
-2.73934543e-01 -1.51288092e-01 -5.93029857e-01 7.46044278e-01
5.33802390e-01 -4.82728891e-02 5.67960441e-01 5.11011243e-01
6.23856246e-01 1.94101229e-01 -1.02430856e+00 1.38758779e+00
-1.10909186e-01 -1.13791180e+00 1.24044456e-01 -1.30917713e-01
4.71080273e-01 -5.08594587e-02 -1.20591499e-01 -5.36113568e-02
2.96982169e-01 -8.37609708e-01 9.35468376e-01 2.97382057e-01
9.16622102e-01 -4.83816385e-01 2.27845028e-01 2.65278846e-01
-1.87585545e+00 8.99018943e-02 8.64868052e-03 7.40976855e-02
4.25777584e-01 -1.44525021e-01 -8.42589676e-01 2.29883388e-01
5.36186695e-01 5.31810343e-01 -3.00734565e-02 9.48512316e-01
-5.71516871e-01 5.58719695e-01 -5.03639340e-01 -4.88775283e-01
2.11182237e-01 2.66665537e-02 7.23730087e-01 1.33860850e+00
7.89412558e-01 3.56043875e-01 1.70646414e-01 4.57707912e-01
3.29948664e-01 2.67969459e-01 -4.06513721e-01 -1.60296217e-01
1.22292554e+00 8.43119621e-01 -7.10421443e-01 -2.54161000e-01
-3.07916068e-02 1.31849682e+00 -4.24451649e-01 3.85475516e-01
-6.21540964e-01 -5.07462919e-01 1.01527727e+00 -1.18157133e-01
2.46626452e-01 -5.49630046e-01 2.61897054e-02 -1.00162745e+00
1.01557977e-01 -9.31109369e-01 1.21534497e-01 -6.86882377e-01
-7.44766057e-01 5.91622531e-01 -5.64142279e-02 -1.86093128e+00
-9.20174181e-01 -6.11570887e-02 -6.71460569e-01 5.33118784e-01
-8.63162994e-01 -8.59221160e-01 -4.85330522e-01 6.29846632e-01
1.01102996e+00 -2.27816373e-01 8.03417206e-01 2.14000136e-01
-2.32215449e-01 7.10079551e-01 -1.01948634e-01 5.21738112e-01
8.66626620e-01 -8.96749258e-01 7.61029005e-01 7.40399420e-01
5.10152459e-01 5.06663024e-01 6.76744103e-01 -6.88788652e-01
-1.41304100e+00 -9.16955531e-01 8.54997456e-01 -1.95671171e-01
5.49410701e-01 -3.12388748e-01 -7.12114692e-01 2.36114100e-01
-9.32122841e-02 -2.52863854e-01 6.39715672e-01 -5.06621599e-01
-1.74830899e-01 -9.37607419e-03 -6.43687963e-01 9.01649356e-01
1.45890176e+00 -5.27109504e-01 -6.65765584e-01 1.55635431e-01
7.84479856e-01 -5.95324337e-01 -5.19082010e-01 1.43142477e-01
8.83796573e-01 -7.57904708e-01 7.60737777e-01 1.95365362e-02
6.97962120e-02 -5.70639968e-01 -3.07824612e-01 -1.22441113e+00
5.50959334e-02 -1.18645036e+00 -6.81444108e-02 1.20283139e+00
7.61981606e-02 -4.21051197e-02 9.74682212e-01 3.38939220e-01
-2.05675419e-02 -3.82985443e-01 -1.03219557e+00 -8.71414721e-01
-4.05507654e-01 -6.33033991e-01 8.85132909e-01 6.85335815e-01
2.92869091e-01 -1.80238746e-02 -4.48951304e-01 5.49430773e-02
5.04820824e-01 7.35958898e-03 1.01144969e+00 -7.07762122e-01
-2.74268836e-01 -6.71357572e-01 -5.47584891e-01 -1.69868314e+00
-2.00976044e-01 -7.18315363e-01 5.74240506e-01 -1.43938291e+00
-1.22654028e-01 -4.37009573e-01 2.13810831e-01 4.64978158e-01
-1.17470980e-01 1.51031062e-01 5.50031602e-01 6.87976420e-01
-4.80134368e-01 7.31704533e-01 1.17316949e+00 -2.45241016e-01
-8.28665555e-01 6.50748387e-02 -1.00946479e-01 4.99275416e-01
8.79493594e-01 -4.43190008e-01 -4.37452585e-01 -5.91961026e-01
-2.82581747e-01 3.18744987e-01 1.77866831e-01 -1.04455578e+00
6.85776293e-01 -3.03493232e-01 -1.16146863e-01 -6.83769166e-01
4.62509215e-01 -6.25270665e-01 1.53291836e-01 6.59880817e-01
-4.85863119e-01 1.04166619e-01 -1.83973148e-01 3.47023875e-01
-2.91638166e-01 -7.22054541e-02 3.14283639e-01 1.87629104e-01
-7.32456565e-01 2.70848244e-01 -5.58686912e-01 -2.52619535e-02
7.96111822e-01 -4.40033704e-01 2.47906849e-01 -9.31300282e-01
-5.97694993e-01 3.53805542e-01 3.05093229e-01 7.38969207e-01
9.25838947e-01 -1.74745059e+00 -6.76831663e-01 2.85727590e-01
1.92259327e-01 3.95426638e-02 3.62711512e-02 5.78151107e-01
-6.02025151e-01 2.42555812e-01 -8.87738764e-02 -8.49347115e-01
-1.58034861e+00 -3.43513638e-02 5.41764088e-02 3.68301868e-01
-5.02009869e-01 9.07364726e-01 -1.24071524e-01 -2.87951589e-01
6.39983773e-01 -5.77879250e-01 4.16388884e-02 -5.19773625e-02
8.11838031e-01 3.65095824e-01 -3.64044666e-01 -8.87814879e-01
-5.13421118e-01 8.51627290e-01 4.15308386e-01 -7.86799967e-01
9.36018050e-01 -2.48548910e-01 3.77993047e-01 3.61239582e-01
1.07503986e+00 -3.46428752e-02 -1.49795759e+00 -1.02477960e-01
4.62373532e-02 -4.10686910e-01 -4.65865523e-01 -3.86815429e-01
-8.50022316e-01 8.48754108e-01 5.50590158e-01 -1.07575074e-01
1.15325856e+00 1.21720349e-02 1.14170563e+00 3.52199793e-01
4.31952178e-01 -1.08022964e+00 3.93070430e-01 6.02261364e-01
8.77387762e-01 -8.90827656e-01 -4.65920895e-01 -2.89645821e-01
-7.75317907e-01 1.19060981e+00 3.93812537e-01 1.75369278e-01
4.14839298e-01 4.89570707e-01 5.81023037e-01 1.91981182e-01
-3.87416005e-01 -4.07110512e-01 2.41072372e-01 8.57185662e-01
3.95677745e-01 6.00419082e-02 -5.09020329e-01 3.81760150e-01
-5.95240235e-01 -3.77817973e-02 9.74162668e-02 1.04480875e+00
-5.58587611e-01 -1.44624627e+00 -6.49942696e-01 -1.14567377e-01
7.35493600e-02 -1.03496030e-01 -5.05106568e-01 6.66762233e-01
-1.47194192e-01 1.14852345e+00 1.82411075e-01 -9.11101937e-01
3.01713109e-01 4.07519192e-02 3.81560475e-01 -5.78074396e-01
-4.40536708e-01 6.62268102e-01 -2.23410800e-01 -8.00083101e-01
-5.35662651e-01 -7.83393383e-01 -1.82242227e+00 -1.05957158e-01
-4.47516628e-02 1.65416300e-01 6.75299942e-01 8.33593190e-01
3.66644144e-01 1.77464887e-01 7.57448196e-01 -1.13917649e+00
-5.86594164e-01 -1.08827579e+00 -3.39862168e-01 4.76098090e-01
2.34537691e-01 -3.72514546e-01 -3.95408422e-01 2.05432177e-01] | [5.69223690032959, -0.1756144016981125] |
2c452979-512d-4a70-9d76-a2b3d616f511 | plenoptic-monte-carlo-object-localization-for | 1806.09769 | null | http://arxiv.org/abs/1806.09769v4 | http://arxiv.org/pdf/1806.09769v4.pdf | Plenoptic Monte Carlo Object Localization for Robot Grasping under Layered Translucency | In order to fully function in human environments, robot perception will need
to account for the uncertainty caused by translucent materials. Translucency
poses several open challenges in the form of transparent objects (e.g.,
drinking glasses), refractive media (e.g., water), and diffuse partial
occlusions (e.g., objects behind stained glass panels). This paper presents
Plenoptic Monte Carlo Localization (PMCL) as a method for localizing object
poses in the presence of translucency using plenoptic (light-field)
observations. We propose a new depth descriptor, the Depth Likelihood Volume
(DLV), and its use within a Monte Carlo object localization algorithm. We
present results of localizing and manipulating objects with translucent
materials and objects occluded by layers of translucency. Our PMCL
implementation uses observations from a Lytro first generation light field
camera to allow a Michigan Progress Fetch robot to perform grasping. | ['Zheming Zhou', 'Zhiqiang Sui', 'Odest Chadwicke Jenkins'] | 2018-06-26 | null | null | null | null | ['transparent-objects'] | ['computer-vision'] | [ 1.97980955e-01 -1.01455323e-01 4.66776162e-01 -1.07771590e-01
-4.52882379e-01 -6.81410611e-01 3.32553118e-01 -2.72068948e-01
-4.05698001e-01 7.99921691e-01 -1.16755612e-01 7.57620782e-02
9.67497155e-02 -6.39047623e-01 -1.07935858e+00 -6.36910737e-01
1.40952185e-01 1.13070953e+00 5.06925166e-01 2.88666129e-01
4.80165541e-01 7.76578069e-01 -1.65266669e+00 1.79466635e-01
5.68489254e-01 6.79572999e-01 1.01164806e+00 7.27206111e-01
1.35115370e-01 4.56275970e-01 -6.68768048e-01 7.78243914e-02
3.56716752e-01 3.39921355e-01 -3.93343478e-01 1.31714925e-01
4.90950853e-01 -1.01920342e+00 4.12355848e-02 7.15847909e-01
1.21866219e-01 2.62829542e-01 1.11961603e+00 -1.05456769e+00
-4.65493023e-01 2.70065814e-01 -5.47417343e-01 -3.84057522e-01
8.33023071e-01 5.31566620e-01 2.55755544e-01 -8.95126700e-01
9.52673078e-01 1.80546045e+00 2.06291124e-01 6.84166253e-01
-1.04460108e+00 -2.66618401e-01 -1.29219323e-01 -6.02000505e-02
-9.72642422e-01 -1.54151678e-01 4.27807271e-01 -7.15015173e-01
9.78003263e-01 1.43737420e-01 4.93104398e-01 1.24299335e+00
1.01143968e+00 6.52853549e-01 1.26116300e+00 -5.49743056e-01
7.87681878e-01 7.72377709e-03 -1.56314701e-01 7.59533167e-01
6.34960651e-01 3.82803291e-01 -6.11002982e-01 -5.13191819e-01
1.09931564e+00 3.98537368e-02 -1.79863125e-01 -9.34910476e-01
-1.25976348e+00 4.10376430e-01 4.67820376e-01 -3.92367095e-01
-2.09146455e-01 5.86555004e-01 -1.67000577e-01 -4.38449949e-01
2.70865053e-01 6.24059975e-01 -3.30112547e-01 -1.72428846e-01
-1.58438161e-02 1.96896896e-01 1.07094526e+00 1.31190407e+00
6.98057413e-01 -3.55174690e-01 2.99053431e-01 4.52641428e-01
8.94563138e-01 9.49496329e-01 -1.81620028e-02 -1.26536155e+00
3.52798462e-01 9.37772617e-02 8.69468331e-01 -1.87566131e-01
-3.05676281e-01 3.30743134e-01 1.32694721e-01 1.21336484e+00
6.69076502e-01 -1.21085450e-01 -1.11056674e+00 1.05393374e+00
6.63435340e-01 -1.80840597e-01 3.00732076e-01 1.12431097e+00
6.14529073e-01 5.74918330e-01 -1.75589859e-01 7.05594122e-02
1.30267441e+00 -6.79488778e-01 -4.59963053e-01 -3.83104324e-01
5.00803292e-01 -1.06332994e+00 9.32423472e-01 6.58098400e-01
-9.76336062e-01 6.65339530e-02 -8.02153945e-01 -2.54654318e-01
-1.81478664e-01 3.06155477e-02 8.52643669e-01 5.19358575e-01
-5.19082308e-01 2.82612503e-01 -1.37677121e+00 -4.49685484e-01
2.74669409e-01 2.73423761e-01 -4.14925128e-01 -5.80584526e-01
-2.92856902e-01 1.14586067e+00 2.88567156e-01 1.28017277e-01
-1.34504032e+00 -6.07891738e-01 -8.14105630e-01 -3.39538544e-01
3.08158189e-01 -4.17821079e-01 1.11274683e+00 -5.88229001e-02
-1.81792092e+00 7.00404823e-01 -1.34306014e-01 4.37888317e-02
6.13306284e-01 -7.68987060e-01 4.70121533e-01 3.85845423e-01
-2.32730120e-01 6.42306447e-01 6.44395530e-01 -2.01670742e+00
-6.84415996e-02 -6.23509526e-01 2.60976136e-01 6.88278437e-01
4.55094546e-01 2.12729685e-02 2.36721829e-01 -1.01507410e-01
6.94405437e-01 -1.05278194e+00 -8.77312347e-02 7.41099596e-01
-4.00250465e-01 -5.12104705e-02 9.65499043e-01 -3.19369048e-01
-2.45969191e-01 -1.87505174e+00 -1.96496665e-01 2.59511196e-03
4.33688685e-02 -2.60668069e-01 1.20258696e-01 3.62619221e-01
5.59006751e-01 -5.31975687e-01 -1.10861190e-01 -4.46217835e-01
3.32395211e-02 4.11580235e-01 -2.43246868e-01 9.27697957e-01
-2.02881888e-01 6.24554276e-01 -1.10070384e+00 -3.51792365e-01
5.54602861e-01 6.59200251e-01 -3.88703734e-01 8.86290509e-04
-6.69119835e-01 5.03954351e-01 -6.28208816e-01 1.06119084e+00
1.03266144e+00 1.63771510e-01 -2.40471542e-01 -9.61130485e-02
-2.75821120e-01 -1.70139924e-01 -1.08514977e+00 1.75817382e+00
-6.41085744e-01 5.43743849e-01 3.71767431e-01 3.53330016e-01
9.01930690e-01 1.16315633e-02 2.04957187e-01 2.24613864e-02
3.13455164e-01 5.80100417e-01 -3.29900235e-01 -8.01175654e-01
5.62652588e-01 -3.41430753e-01 4.25970554e-01 3.92422289e-01
-2.85711557e-01 -1.05518162e+00 -2.56836444e-01 -8.53852704e-02
1.33316457e+00 8.26677024e-01 8.01751483e-03 -2.37180740e-01
-4.40221667e-01 1.40241534e-01 1.36780858e-01 6.65460467e-01
-2.37715915e-01 1.05375123e+00 -2.26032078e-01 -4.32300985e-01
-1.02452326e+00 -1.57460594e+00 -4.83512521e-01 6.66373372e-01
8.37586343e-01 4.00150090e-01 -4.72679496e-01 -1.09111406e-01
2.03564048e-01 7.82310963e-01 -3.60111594e-01 4.75412272e-02
-5.82553089e-01 -5.63591957e-01 -2.15622529e-01 5.04794538e-01
2.25425676e-01 -1.18535161e+00 -1.45137596e+00 1.94959819e-01
-1.33895531e-01 -1.15597379e+00 1.43936768e-01 4.17663664e-01
-1.00380397e+00 -1.04044557e+00 -8.34904253e-01 -3.81240159e-01
8.30158472e-01 4.29749072e-01 9.18740988e-01 -5.12995183e-01
-8.17160010e-01 8.71945620e-01 -5.72486103e-01 -6.36762321e-01
-2.78047979e-01 -6.39760792e-01 8.96940455e-02 -5.13475776e-01
3.28404784e-01 -2.83906370e-01 -6.31900549e-01 4.31054652e-01
-8.18307996e-01 -2.04980105e-01 3.28410923e-01 4.43222582e-01
2.76802331e-01 -2.14285776e-01 -6.30217373e-01 -4.55723375e-01
1.96019441e-01 -2.58442938e-01 -9.73919272e-01 1.71713099e-01
1.86607018e-01 -9.15419608e-02 -5.81331588e-02 -7.10957587e-01
-1.43563855e+00 1.92325532e-01 6.72609746e-01 -5.26360810e-01
-4.96845752e-01 -1.54461622e-01 9.94812101e-02 -2.47303128e-01
1.17614102e+00 -4.05258566e-01 -1.64294243e-01 -2.16394261e-01
1.44850150e-01 5.80568910e-01 3.12355936e-01 -8.32111776e-01
7.05754697e-01 1.22851062e+00 3.42906922e-01 -8.65690887e-01
-3.20826650e-01 -3.59114915e-01 -6.27686381e-01 -4.03216869e-01
8.92317474e-01 -7.39661455e-01 -1.31483984e+00 5.36899209e-01
-1.62238503e+00 -6.45497561e-01 -2.77009904e-01 1.25969827e+00
-8.96407366e-01 2.41780847e-01 -5.18957913e-01 -1.52299523e+00
4.27208304e-01 -1.34797978e+00 1.59061790e+00 3.66871327e-01
-3.30093428e-02 -6.55738533e-01 -7.88672641e-03 5.40851057e-01
3.11345365e-02 4.41499799e-01 3.83389235e-01 3.79193783e-01
-1.34990156e+00 -1.39383122e-01 -8.83592591e-02 -1.03625156e-01
8.92517790e-02 2.80700624e-03 -1.32598448e+00 -2.90952772e-01
2.80883700e-01 -7.19885767e-01 7.77012110e-01 7.52940297e-01
6.36659980e-01 2.72051077e-02 -7.03018725e-01 4.95466501e-01
1.42345858e+00 4.37838107e-01 5.66229820e-01 3.07279706e-01
6.32940531e-01 8.67484987e-01 1.02705514e+00 7.06844628e-01
1.50382608e-01 4.52042609e-01 1.21916020e+00 4.36594367e-01
-8.37005451e-02 2.24957988e-01 2.78327703e-01 -1.68459177e-01
-3.24242353e-01 -7.02694356e-01 -8.69205594e-01 2.47108117e-01
-1.31338143e+00 -5.43188334e-01 -3.16505611e-01 2.16779947e+00
5.65517962e-01 -1.14902165e-02 -6.90786541e-01 -3.53490561e-01
4.43615258e-01 -3.70519906e-01 -8.06900561e-01 -2.78850198e-01
1.97654124e-02 -2.12260578e-02 8.98069441e-01 9.40613329e-01
-6.88341677e-01 8.70217919e-01 6.05110025e+00 3.76576483e-01
-5.78759253e-01 -7.32461214e-02 -1.00921422e-01 -6.05062395e-02
-4.16517794e-01 1.91514418e-01 -9.60630774e-01 1.77922904e-01
9.18293223e-02 5.97906053e-01 6.55219853e-01 9.26908016e-01
3.32863852e-02 -1.30080068e+00 -1.08939457e+00 7.86382020e-01
2.15950668e-01 -3.55113655e-01 -3.77583861e-01 8.63808114e-03
8.50167871e-01 2.84845769e-01 2.04268292e-01 -4.34930354e-01
7.83774436e-01 -7.77063191e-01 1.00784099e+00 5.80332220e-01
4.79587406e-01 3.58971022e-02 5.37207842e-01 3.77622336e-01
-4.78552938e-01 -1.34521365e-01 -7.65027702e-01 -4.67250049e-02
3.72226119e-01 8.15805435e-01 -1.47684777e+00 -1.59132909e-02
7.77837753e-01 -1.04121648e-01 1.40105993e-01 1.37965155e+00
-8.20272639e-02 -5.91876768e-02 -9.36570048e-01 -3.53622288e-01
-8.38016346e-02 -1.91652671e-01 8.11244130e-01 4.68818545e-01
2.84502119e-01 1.30554274e-01 5.95177077e-02 1.18843067e+00
3.86215240e-01 -3.93691331e-01 -5.74930310e-01 5.21923482e-01
3.13363642e-01 8.13915908e-01 -9.18943465e-01 3.86890396e-02
-2.65672971e-02 1.05000389e+00 -2.14294605e-02 5.93510926e-01
-4.78657007e-01 -2.10141018e-01 3.11802089e-01 8.53739008e-02
-3.88806202e-02 -7.48013735e-01 -4.65544879e-01 -1.18952394e+00
4.69658077e-01 -1.58051163e-01 -4.88900185e-01 -1.66987646e+00
-1.14276373e+00 2.82321155e-01 5.04075646e-01 -1.14184475e+00
3.21380757e-02 -1.24731827e+00 2.25471947e-02 9.51342404e-01
-1.32917643e+00 -1.18352294e+00 -8.25337768e-01 6.70003667e-02
6.72213912e-01 3.64762664e-01 6.94126785e-01 -5.20048499e-01
4.71800059e-01 -3.50530148e-01 4.58500624e-01 -6.75165772e-01
8.61287177e-01 -1.20174229e+00 -1.39204022e-02 3.46050739e-01
-2.59175271e-01 5.97145140e-01 1.11209178e+00 -1.04895663e+00
-1.55810177e+00 -5.36921799e-01 1.37187587e-02 -1.14983273e+00
3.19624603e-01 -6.54582143e-01 -4.98135656e-01 7.86263466e-01
-3.47689092e-01 2.01915562e-01 -3.64090279e-02 -3.62104267e-01
-2.81296670e-01 3.58531266e-01 -1.70161664e+00 5.80022693e-01
8.84448946e-01 -2.24622563e-01 -5.38565278e-01 7.93938279e-01
5.04282236e-01 -1.14421213e+00 -2.87022710e-01 4.45945561e-01
1.12760150e+00 -9.46535528e-01 9.96535957e-01 1.77273616e-01
3.21336538e-01 -3.06780457e-01 -3.01879197e-01 -1.25361860e+00
2.47048378e-01 -5.24015903e-01 -1.33801326e-01 4.98591363e-01
5.28043322e-02 -7.45634139e-01 9.29276109e-01 1.00396752e+00
-4.36512738e-01 -1.70716554e-01 -1.02711630e+00 -1.10020030e+00
-4.30233963e-02 -9.66068655e-02 1.16219006e-01 2.84421444e-01
-1.02066033e-01 -3.52492601e-01 5.81892058e-02 5.65904260e-01
1.14163780e+00 3.24145071e-02 6.15735948e-01 -1.37722492e+00
-2.39100352e-01 3.83449793e-01 -3.86569530e-01 -1.10905910e+00
6.79925159e-02 -4.35079306e-01 9.23161805e-01 -1.78899240e+00
3.57289940e-01 -7.14020371e-01 7.38698542e-01 7.02158734e-02
2.21688733e-01 3.74758057e-02 2.52575185e-02 3.23437572e-01
-1.74555048e-01 6.07456446e-01 1.76224041e+00 -1.66230723e-02
-2.52936423e-01 -2.88095661e-02 2.54570663e-01 1.04010975e+00
7.10638821e-01 -6.37601078e-01 -2.54958540e-01 -8.20666671e-01
1.99466154e-01 8.22365656e-02 6.84399247e-01 -1.27704811e+00
1.77563429e-01 -4.12356108e-01 3.93420339e-01 -7.27031767e-01
1.12889981e+00 -1.22706532e+00 1.18282162e-01 5.77270329e-01
-1.47766434e-02 -3.95722270e-01 5.99842258e-02 7.55539775e-01
5.82956076e-01 -8.20938349e-01 8.77735138e-01 -7.07411289e-01
-2.87533969e-01 -1.69470727e-01 -5.72247744e-01 -4.47205037e-01
1.14801395e+00 -4.89643186e-01 -7.97361910e-01 1.03439257e-01
-7.71003246e-01 -7.49145523e-02 1.01896501e+00 1.92330897e-01
9.30446744e-01 -7.97331870e-01 -3.73663604e-01 1.15212753e-01
2.13255897e-01 4.32102680e-01 -8.10524225e-02 2.83323348e-01
-1.35191214e+00 2.22975537e-01 -1.44470423e-01 -7.57680237e-01
-1.09583282e+00 3.90514940e-01 1.44219592e-01 3.42899770e-01
-6.13156199e-01 1.25045633e+00 7.80892789e-01 -4.50525671e-01
1.73427984e-01 -9.94572997e-01 4.24473822e-01 -8.04737389e-01
2.75436521e-01 7.57327378e-01 -1.70033634e-01 -1.72465727e-01
-1.89692423e-01 7.40047455e-01 1.14847377e-01 -4.63231534e-01
1.04177356e+00 -2.56855488e-01 -3.55216749e-02 7.81496584e-01
6.42035186e-01 -8.25004205e-02 -1.84094572e+00 1.51731804e-01
-5.69300652e-01 -7.77387738e-01 -3.07441860e-01 -1.01180291e+00
-3.52500349e-01 8.40759575e-01 6.84803426e-01 -2.76338458e-01
3.00639898e-01 3.47976297e-01 3.33921283e-01 8.11876833e-01
1.14857221e+00 -8.60953629e-01 2.83532679e-01 5.62797904e-01
1.05322814e+00 -1.30557263e+00 1.49902627e-01 -9.20799136e-01
-1.70986116e-01 1.36642766e+00 7.16348946e-01 -2.96336561e-01
5.31298399e-01 7.58201420e-01 1.18360072e-01 -3.85924786e-01
-3.99831802e-01 8.57085511e-02 -2.48609707e-01 1.04290628e+00
-2.16667522e-02 2.70301044e-01 1.84561193e-01 -5.06066024e-01
-3.81672569e-02 -1.20263584e-01 9.53429699e-01 1.53843188e+00
-8.57586563e-01 -4.25532430e-01 -1.08951235e+00 3.07768583e-01
4.84895371e-02 1.04595959e-01 -7.84946010e-02 4.32978213e-01
1.00276992e-01 9.87222850e-01 8.42040777e-02 2.79705346e-01
1.32092386e-01 -3.60785097e-01 1.16916955e+00 -8.84264708e-01
-1.99745186e-02 1.40181050e-01 -2.63595790e-01 -6.52086854e-01
-4.86920625e-01 -8.35317314e-01 -1.34371746e+00 2.46347576e-01
-7.69005120e-01 -4.63294327e-01 1.24406528e+00 8.79685581e-01
-2.16111779e-01 2.32633933e-01 1.36758134e-01 -1.34876466e+00
-4.80255038e-01 -1.12267148e+00 -9.97160196e-01 -1.27542928e-01
4.09213513e-01 -1.19432855e+00 -8.45507443e-01 8.59418139e-02] | [5.98126220703125, -1.0792629718780518] |
17792b84-2ee5-457a-8dcb-12f14289e6d1 | incipient-fault-detection-in-power | 2302.09332 | null | https://arxiv.org/abs/2302.09332v1 | https://arxiv.org/pdf/2302.09332v1.pdf | Incipient Fault Detection in Power Distribution System: A Time-Frequency Embedded Deep Learning Based Approach | Incipient fault detection in power distribution systems is crucial to improve the reliability of the grid. However, the non-stationary nature and the inadequacy of the training dataset due to the self-recovery of the incipient fault signal, make the incipient fault detection in power distribution systems a great challenge. In this paper, we focus on incipient fault detection in power distribution systems and address the above challenges. In particular, we propose an ADaptive Time-Frequency Memory(AD-TFM) cell by embedding wavelet transform into the Long Short-Term Memory (LSTM), to extract features in time and frequency domain from the non-stationary incipient fault signals.We make scale parameters and translation parameters of wavelet transform learnable to adapt to the dynamic input signals. Based on the stacked AD-TFM cells, we design a recurrent neural network with ATtention mechanism, named AD-TFM-AT model, to detect incipient fault with multi-resolution and multi-dimension analysis. In addition, we propose two data augmentation methods, namely phase switching and temporal sliding, to effectively enlarge the training datasets. Experimental results on two open datasets show that our proposed AD-TFM-AT model and data augmentation methods achieve state-of-the-art (SOTA) performance of incipient fault detection in power distribution system. We also disclose one used dataset logged at State Grid Corporation of China to facilitate future research. | ['Zhi Liu', 'Weitao Li', 'Wei Sun', 'Yuxing Deng', 'Hong Cheng', 'Huan Luo', 'Qiyue Li'] | 2023-02-18 | null | null | null | null | ['fault-detection'] | ['miscellaneous'] | [-2.41386935e-01 -7.10956335e-01 -2.15591080e-02 4.97362874e-02
-5.86058140e-01 -3.88544351e-01 1.19628415e-01 -4.75473017e-01
3.72981310e-01 5.70166111e-01 4.47050333e-02 -4.63531911e-01
-3.48504692e-01 -7.07677841e-01 -2.55265683e-01 -9.59265232e-01
-5.05482674e-01 7.99145997e-02 -9.60178748e-02 -1.39811978e-01
2.65271097e-01 6.11093819e-01 -1.20752144e+00 4.66371387e-01
1.06666625e+00 1.16880918e+00 3.75792116e-01 3.89511734e-01
2.32416600e-01 8.14216256e-01 -1.15525055e+00 6.65184617e-01
-6.20454662e-02 -4.36546385e-01 -6.41468406e-01 2.16441706e-01
-3.00067782e-01 -3.42765480e-01 -7.58623421e-01 1.11857677e+00
7.72432268e-01 5.85615188e-02 4.08512682e-01 -1.41030133e+00
-9.94508922e-01 5.55384219e-01 -7.11641550e-01 1.12244856e+00
1.18824735e-01 -1.75556205e-02 5.35609186e-01 -1.11527050e+00
-6.97998479e-02 8.54635119e-01 7.94931948e-01 1.23415098e-01
-6.97501600e-01 -7.09354222e-01 2.11530313e-01 9.94856358e-01
-1.37716067e+00 -2.47447267e-01 1.14967561e+00 -1.24338008e-01
1.62759602e+00 1.17253311e-01 6.67782962e-01 7.48719513e-01
6.87830091e-01 5.98129570e-01 7.99771547e-01 -3.25629354e-01
1.46920800e-01 -5.40797830e-01 2.70589799e-01 3.51777196e-01
-3.31811532e-02 8.68701283e-03 -4.77881968e-01 -1.67360753e-01
9.75872099e-01 2.34424040e-01 -7.45973647e-01 5.32314062e-01
-8.68342638e-01 6.01125181e-01 3.74817610e-01 9.11882043e-01
-5.04099548e-01 -1.10490359e-01 7.70376563e-01 8.16392839e-01
7.05014050e-01 2.73349643e-01 -7.99925745e-01 -1.56968623e-01
-8.30782175e-01 -2.72181153e-01 3.13890040e-01 4.34435487e-01
4.64857556e-02 1.10983217e+00 5.84861487e-02 8.13936949e-01
-7.94399902e-02 6.15833640e-01 1.48722827e+00 -4.29110676e-01
4.23510700e-01 3.69611114e-01 -7.63619915e-02 -8.89578879e-01
-6.41002953e-01 -8.24756384e-01 -1.15319300e+00 6.10194765e-02
-1.85098618e-01 -1.84765309e-01 -8.35125983e-01 1.34243381e+00
-1.06513500e-01 6.11042082e-01 1.49073541e-01 7.42501616e-01
4.43536520e-01 1.03542411e+00 -5.04489303e-01 -6.81191921e-01
1.27833092e+00 -8.63926709e-01 -1.30872786e+00 -1.00857198e-01
6.08649611e-01 -5.72836816e-01 8.69591415e-01 5.63278019e-01
-7.95181394e-01 -6.36970460e-01 -1.34124231e+00 5.11224329e-01
-3.07636619e-01 3.73751283e-01 1.99426204e-01 1.69226050e-01
-6.27703905e-01 8.38735282e-01 -9.27458584e-01 2.80716661e-02
2.22595036e-01 2.48082161e-01 -2.72461802e-01 2.35317856e-01
-1.62723863e+00 1.07312691e+00 2.72898614e-01 5.20904720e-01
-8.39847147e-01 -6.54652834e-01 -6.33668244e-01 1.67111009e-01
5.66375032e-02 -1.18384875e-01 9.16965425e-01 -8.47869575e-01
-1.36070263e+00 1.60534531e-02 -3.57316621e-02 -6.00595891e-01
-1.53873965e-01 -4.80291359e-02 -1.23885119e+00 3.95381749e-01
8.38599429e-02 -5.12458742e-01 1.23189294e+00 -6.31468654e-01
-6.92209661e-01 -4.56542701e-01 -7.55386174e-01 -8.10606033e-02
-6.05490267e-01 -1.26725391e-01 4.98199373e-01 -1.09847200e+00
4.21842575e-01 -2.54780710e-01 2.10741326e-01 -7.23735809e-01
-3.32786560e-01 -4.20223147e-01 1.88918138e+00 -1.17922366e+00
1.48402286e+00 -2.35724545e+00 -1.03526972e-01 1.32797927e-01
-1.63976908e-01 4.45736527e-01 -8.99440199e-02 4.20298398e-01
-6.91743374e-01 -1.92015305e-01 -1.29276067e-01 1.85892999e-01
-2.71782964e-01 3.79148543e-01 -7.84451485e-01 7.91497171e-01
3.34504128e-01 7.08595335e-01 -6.20430768e-01 1.47894979e-01
3.84400040e-01 4.39816117e-01 1.37892246e-01 1.03876330e-01
2.96148121e-01 4.04162914e-01 -4.10605043e-01 6.20758057e-01
6.88428342e-01 -2.10951179e-01 -2.98301995e-01 -5.50503492e-01
-1.21303923e-01 3.05373996e-01 -1.04794312e+00 1.39971173e+00
-4.74356771e-01 4.68163729e-01 -2.76198816e-02 -1.58346117e+00
1.05515003e+00 9.23499703e-01 6.79034531e-01 -1.02689910e+00
6.29121140e-02 4.60000396e-01 8.42930153e-02 -6.46589577e-01
5.49909100e-02 -7.05521554e-02 2.42399573e-01 4.85523254e-01
3.96973550e-01 1.34067371e-01 6.61624297e-02 -3.20651978e-01
1.00408447e+00 -3.16345721e-01 -1.38411745e-01 -6.10854506e-01
5.17424583e-01 -3.55759025e-01 9.91753936e-01 1.98866174e-01
-8.91015977e-02 4.50915009e-01 1.95526376e-01 -6.93276227e-01
-9.75927234e-01 -6.94717467e-01 -4.84466851e-01 3.80498350e-01
-2.78588533e-01 -1.22048825e-01 -4.64520991e-01 -5.45922339e-01
1.22685827e-01 5.35814643e-01 -3.83771122e-01 -5.51969051e-01
-8.29187632e-01 -1.22838235e+00 6.31160378e-01 8.13823879e-01
8.14505458e-01 -1.04953825e+00 -4.75437671e-01 6.39518142e-01
-1.93401575e-01 -8.00328076e-01 -5.74092507e-01 5.20984650e-01
-8.93653512e-01 -1.00108945e+00 -7.28088200e-01 -1.15221536e+00
5.73595345e-01 2.59656012e-01 8.40369165e-01 2.24121019e-01
-2.97530323e-01 -9.02402624e-02 -4.80914116e-01 9.31647718e-02
1.51318759e-01 -7.50131607e-02 4.25265640e-01 -6.37311041e-02
1.61240354e-01 -1.14105606e+00 -4.18279856e-01 3.06403339e-01
-7.57666051e-01 -3.18446785e-01 3.37770551e-01 1.22530377e+00
2.94673920e-01 1.15232599e+00 1.54019022e+00 -3.67619067e-01
7.46483862e-01 -7.08147109e-01 -5.28325796e-01 -7.31023625e-02
-8.19919586e-01 -3.22256267e-01 1.29443359e+00 -7.12393582e-01
-7.31131911e-01 -5.79799414e-01 -3.09304059e-01 -8.23577046e-01
4.03435647e-01 8.86003971e-01 -1.61367536e-01 -6.51356056e-02
2.00400755e-01 5.61462760e-01 -9.64488983e-02 -6.96002305e-01
-6.06161505e-02 8.04004848e-01 8.49190414e-01 -2.41420612e-01
7.33600616e-01 1.71344072e-01 -2.30486631e-01 -8.65866125e-01
-3.40768427e-01 3.01512820e-03 -3.52345705e-01 9.18919668e-02
2.06327692e-01 -8.48456323e-01 -6.04676962e-01 9.51293945e-01
-9.60996985e-01 -2.09781185e-01 -4.34430212e-01 4.00490522e-01
-2.39914015e-01 5.28445065e-01 -1.24500239e+00 -7.74405301e-01
-7.59333730e-01 -9.17163551e-01 5.86968064e-01 1.52588561e-01
7.63466731e-02 -1.02235079e+00 -2.87485719e-01 -1.89270586e-01
7.05993593e-01 1.21005207e-01 1.14795935e+00 -5.86966574e-01
-1.51420450e-02 -2.06201702e-01 2.01362431e-01 6.68962061e-01
6.02577269e-01 -1.70561090e-01 -7.78750658e-01 -7.52426624e-01
8.47424686e-01 -1.57152787e-02 6.04513288e-01 5.87478876e-01
1.08023250e+00 -5.68814635e-01 -1.50420487e-01 5.85516691e-01
1.30685580e+00 6.83676898e-01 5.86828411e-01 3.93967718e-01
5.25575519e-01 4.40678559e-02 4.75876123e-01 7.04526126e-01
1.97795480e-01 3.19303125e-01 1.31619900e-01 -1.60566866e-01
-6.86496869e-02 -2.54843701e-02 4.40159857e-01 1.47388136e+00
2.55221009e-01 -9.06920284e-02 -6.37909234e-01 8.41843963e-01
-1.58639586e+00 -9.03340220e-01 -8.55920389e-02 1.85401952e+00
6.34433508e-01 2.57986099e-01 -2.09908321e-01 1.11308241e+00
6.97184563e-01 6.85363412e-02 -7.13672638e-01 -1.49993107e-01
-5.99914730e-01 2.23509908e-01 1.94690809e-01 4.04699534e-01
-8.89643252e-01 2.96514392e-01 5.70331860e+00 1.17585599e+00
-1.60507977e+00 3.17457020e-01 5.87132335e-01 1.49110958e-01
3.02800033e-02 -2.40834713e-01 -3.03992748e-01 8.90904844e-01
1.06695497e+00 -4.02320147e-01 5.37057042e-01 5.02552807e-01
4.85050172e-01 2.09054723e-01 -7.71672666e-01 9.12183464e-01
-1.19027989e-02 -1.02879035e+00 -8.20666552e-02 -3.39207411e-01
5.98080754e-01 1.55454576e-02 7.19938949e-02 3.17875236e-01
-3.95790279e-01 -9.37019527e-01 4.42588806e-01 3.04911256e-01
6.64090216e-01 -1.14163148e+00 9.83664274e-01 3.68585795e-01
-1.51467836e+00 -6.91325665e-01 -2.77764440e-01 -2.01598570e-01
3.37694287e-01 1.07881594e+00 -1.67963430e-01 9.83880103e-01
1.05031550e+00 1.18067729e+00 -2.39002883e-01 6.43318772e-01
-2.26728764e-04 7.68644571e-01 -2.99735516e-01 6.38293743e-01
-3.78330797e-02 -1.13347135e-02 4.22205985e-01 6.28565550e-01
7.04915404e-01 6.60618171e-02 2.00222790e-01 5.30869126e-01
2.33114153e-01 -4.95051533e-01 -4.06516641e-01 -1.39824925e-02
8.62490952e-01 1.14334631e+00 -5.38907468e-01 -3.94581288e-01
-3.35880280e-01 8.80437315e-01 -2.03296244e-01 5.95737934e-01
-6.34015262e-01 -6.93455935e-01 4.46912587e-01 -2.59919107e-01
2.44406387e-01 -1.73313975e-01 -1.94707498e-01 -1.33314824e+00
2.40519002e-01 -9.18792844e-01 5.22123814e-01 -8.71220410e-01
-1.57238686e+00 9.13681149e-01 -3.93865943e-01 -1.39288557e+00
-3.96135926e-01 -2.43729860e-01 -1.08389425e+00 1.03323650e+00
-1.69701135e+00 -1.04400349e+00 1.24368735e-01 7.53751874e-01
9.01609540e-01 -3.79244447e-01 9.31806803e-01 6.34794474e-01
-8.28233242e-01 2.96654701e-01 2.27662846e-01 1.56332463e-01
1.61885798e-01 -9.70943928e-01 4.67402905e-01 1.13040102e+00
-1.33882970e-01 1.65637419e-01 6.26300335e-01 -6.94642425e-01
-1.68321741e+00 -1.07298839e+00 5.85239828e-01 1.76880360e-01
7.48285174e-01 1.45062432e-02 -1.43078482e+00 7.92955101e-01
3.99297208e-01 4.04101223e-01 2.61170417e-01 -3.78072053e-01
1.27875447e-01 -3.10889482e-01 -1.34048486e+00 8.74320343e-02
3.26468766e-01 -7.94575930e-01 -8.04287851e-01 3.66861343e-01
4.94405121e-01 -1.83510289e-01 -1.21844554e+00 8.06281805e-01
-1.13558201e-02 -4.51668471e-01 9.67285752e-01 -7.25276992e-02
-2.44883925e-01 -5.63836217e-01 -8.43613446e-02 -1.77350187e+00
-6.77555740e-01 -5.55196226e-01 -5.95272183e-01 1.23019242e+00
-2.38476391e-03 -1.21053779e+00 3.74866366e-01 -2.43252754e-01
-5.74310005e-01 -9.17557359e-01 -1.48717284e+00 -8.36115658e-01
1.03924617e-01 -7.98151568e-02 7.49587297e-01 1.26826274e+00
5.02247691e-01 2.09679216e-01 -4.08338398e-01 6.46445751e-01
3.22842389e-01 2.42254093e-01 -2.31547967e-01 -8.41981769e-01
-1.72769979e-01 -3.09501201e-01 -2.69854963e-01 -7.38879085e-01
1.98482573e-01 -6.68211341e-01 -2.40909293e-01 -1.24540067e+00
-3.62546504e-01 -2.15925425e-01 -7.35112190e-01 6.92536771e-01
-8.39078277e-02 2.12335885e-01 -2.24081084e-01 4.01349396e-01
3.63367163e-02 9.04793978e-01 1.01894689e+00 -2.82449812e-01
1.44197289e-02 -2.23615631e-01 -7.17528760e-02 5.55334568e-01
9.83004630e-01 -1.50541663e-01 -3.54191154e-01 -5.53961694e-01
-3.06183219e-01 5.65930367e-01 4.04714644e-02 -1.21110332e+00
3.10347825e-01 2.65239120e-01 7.49127984e-01 -8.46908391e-01
1.29729390e-01 -7.75766730e-01 6.48415163e-02 9.18614626e-01
3.69562566e-01 8.25744212e-01 3.66164118e-01 4.16149572e-02
-5.78103542e-01 -1.26475528e-01 5.58243811e-01 7.28384033e-02
-6.50279582e-01 2.37919807e-01 -8.20939124e-01 -2.93141067e-01
6.05715156e-01 1.53849095e-01 -5.83242118e-01 -1.10163294e-01
-5.75218260e-01 4.02734756e-01 -1.18659146e-01 5.60507834e-01
8.34759176e-01 -1.69910383e+00 -7.87211120e-01 9.51619685e-01
-4.48568225e-01 -2.94313073e-01 6.43588603e-01 8.25950444e-01
-8.35826397e-02 4.11582202e-01 -2.82417595e-01 -4.68203217e-01
-8.72214198e-01 5.47505498e-01 6.10967219e-01 -2.21350431e-01
-1.18634033e+00 3.78921419e-01 -3.89865637e-01 8.76743048e-02
-5.32667637e-02 -5.97585857e-01 -4.05278206e-01 2.09852327e-02
6.71995580e-01 4.79647785e-01 5.73880017e-01 -4.98754025e-01
-2.77708918e-01 6.17011964e-01 1.09289445e-01 3.30256790e-01
1.43597376e+00 -1.61350846e-01 -3.92084509e-01 6.25561059e-01
1.04830897e+00 -4.96532410e-01 -1.03417730e+00 -8.09779242e-02
-4.23099361e-02 -2.46638790e-01 4.48404014e-01 -8.48402679e-01
-1.62478471e+00 7.21863806e-01 8.38211656e-01 6.78168476e-01
1.61105549e+00 -5.69134414e-01 1.21023905e+00 1.95876718e-01
4.37963694e-01 -1.13604891e+00 3.01390141e-02 5.62615931e-01
8.65690470e-01 -5.72067082e-01 -2.57823765e-01 6.83160350e-02
-8.35537072e-03 1.43309736e+00 4.93383348e-01 -4.74277169e-01
8.25407505e-01 7.86125541e-01 5.52511141e-02 1.10985013e-02
-1.00314522e+00 4.69980359e-01 -1.92164421e-01 5.08968353e-01
8.55594650e-02 -1.43284455e-01 -1.39770254e-01 9.21392739e-01
-3.74200828e-02 6.86692223e-02 5.11275232e-01 1.22307110e+00
-4.93140489e-01 -9.31512952e-01 -7.23521531e-01 7.24001110e-01
-6.57793760e-01 6.64513931e-02 5.54297090e-01 4.13372934e-01
1.36054307e-02 1.09409308e+00 2.76045293e-01 -5.31586945e-01
3.50543946e-01 1.04162320e-01 4.18493569e-01 -1.86232850e-01
-5.09704292e-01 2.57564962e-01 -3.57937366e-01 -9.75177363e-02
9.22682807e-02 -3.79999787e-01 -1.56176698e+00 -1.54261380e-01
-6.74209774e-01 6.12801611e-01 2.68527120e-01 1.15011299e+00
3.10928077e-01 1.17847347e+00 1.15641153e+00 -9.04237688e-01
-7.14420974e-01 -1.61445856e+00 -1.12187183e+00 2.16376066e-01
7.24819779e-01 -6.94543660e-01 -6.59450412e-01 -2.12447003e-01] | [6.495669364929199, 2.7309763431549072] |
6ef4dcc0-6908-402f-9608-f399fb79b742 | jojogan-one-shot-face-stylization-1 | 2112.11641 | null | https://arxiv.org/abs/2112.11641v4 | https://arxiv.org/pdf/2112.11641v4.pdf | JoJoGAN: One Shot Face Stylization | A style mapper applies some fixed style to its input images (so, for example, taking faces to cartoons). This paper describes a simple procedure -- JoJoGAN -- to learn a style mapper from a single example of the style. JoJoGAN uses a GAN inversion procedure and StyleGAN's style-mixing property to produce a substantial paired dataset from a single example style. The paired dataset is then used to fine-tune a StyleGAN. An image can then be style mapped by GAN-inversion followed by the fine-tuned StyleGAN. JoJoGAN needs just one reference and as little as 30 seconds of training time. JoJoGAN can use extreme style references (say, animal faces) successfully. Furthermore, one can control what aspects of the style are used and how much of the style is applied. Qualitative and quantitative evaluation show that JoJoGAN produces high quality high resolution images that vastly outperform the current state-of-the-art. | ['David Forsyth', 'Min Jin Chong'] | 2021-12-22 | jojogan-one-shot-face-stylization | https://arxiv.org/abs/2112.11641 | https://arxiv.org/pdf/2112.11641.pdf | arxiv-2021-12 | ['image-stylization', 'one-shot-face-stylization'] | ['computer-vision', 'computer-vision'] | [ 5.17857850e-01 2.63937384e-01 2.30749875e-01 -5.01760662e-01
-6.60095215e-01 -6.97327256e-01 7.16840327e-01 -8.52851391e-01
-3.45466398e-02 8.75030339e-01 5.05380407e-02 -7.64674693e-03
3.87691200e-01 -1.01672614e+00 -7.64327943e-01 -6.41122997e-01
6.04930580e-01 6.42024398e-01 -1.23148136e-01 -5.22588074e-01
2.04517022e-01 4.20666248e-01 -1.16031396e+00 4.11904395e-01
5.60038149e-01 6.65272117e-01 3.21408689e-01 9.89404261e-01
-4.57446426e-01 7.06787288e-01 -8.89546633e-01 -5.61536014e-01
4.91756320e-01 -1.32674050e+00 -7.08252311e-01 1.28772676e-01
5.72841227e-01 -2.21065089e-01 1.36110500e-01 9.23898876e-01
3.99744242e-01 -3.41060683e-02 6.09382808e-01 -1.02860546e+00
-7.76766956e-01 2.99939007e-01 -6.82275891e-01 -2.90630221e-01
4.44182664e-01 3.03081512e-01 6.57687485e-01 -6.18059218e-01
1.01819623e+00 1.40825021e+00 4.60376292e-01 1.07049870e+00
-1.62073600e+00 -7.84540296e-01 -2.95482427e-01 -3.99999142e-01
-1.11147809e+00 -3.00494403e-01 1.05049384e+00 -3.03309083e-01
4.43917871e-01 3.64377975e-01 8.44785452e-01 1.15602791e+00
3.20004106e-01 3.67252171e-01 1.61989093e+00 -6.20595753e-01
3.47484380e-01 1.08762562e-01 -5.18716395e-01 6.22110188e-01
-1.82867557e-01 1.13756746e-01 -3.98138523e-01 4.77289595e-02
1.09100962e+00 -2.53320754e-01 -3.63883041e-02 -4.51812968e-02
-1.17598474e+00 7.84253657e-01 5.17512977e-01 1.96182117e-01
-2.28192285e-01 3.35936219e-01 1.76216319e-01 7.34106243e-01
4.26120579e-01 8.43783855e-01 -9.37866122e-02 -2.09241837e-01
-1.12260365e+00 1.99828282e-01 6.05040491e-01 1.22212136e+00
1.21261382e+00 3.83845508e-01 -8.72042626e-02 1.01700282e+00
-6.66550994e-02 8.29255998e-01 8.86384398e-02 -1.20563066e+00
1.75058335e-01 3.18941891e-01 -1.02421843e-01 -5.93150377e-01
1.74111918e-01 -2.73446366e-02 -9.88705695e-01 8.88776481e-01
3.34491491e-01 -3.70153695e-01 -1.21203732e+00 1.65598917e+00
8.76370966e-02 -1.36039183e-01 -1.19251080e-01 6.21041358e-01
6.51734889e-01 8.40526044e-01 -2.94925690e-01 1.13720968e-01
1.31761992e+00 -9.27455783e-01 -4.49762225e-01 -3.33723813e-01
2.18647823e-01 -8.64664912e-01 1.53452444e+00 3.54476124e-01
-1.40010643e+00 -6.70807242e-01 -9.92979765e-01 -5.12441769e-02
-2.85422832e-01 1.45535758e-02 3.20138365e-01 6.41117692e-01
-1.24741161e+00 5.07413983e-01 -3.54058802e-01 -5.05725265e-01
6.33140802e-01 2.12771654e-01 -4.01677877e-01 -2.10291430e-01
-8.36411059e-01 7.46773183e-01 3.57831597e-01 -2.20858261e-01
-8.46937418e-01 -9.97526586e-01 -7.40052402e-01 -1.64333120e-01
9.73615870e-02 -1.22449422e+00 1.33520269e+00 -1.63316417e+00
-2.05070782e+00 1.13535917e+00 -2.73795247e-01 -1.69302642e-01
6.75193667e-01 6.99802395e-03 -3.97259772e-01 2.03035876e-01
1.43108308e-01 1.13293815e+00 1.30331814e+00 -1.53787947e+00
-5.86326241e-01 4.79208380e-02 3.05397332e-01 1.50371000e-01
2.32132778e-01 1.80157740e-02 -5.40218592e-01 -9.40252364e-01
-2.27768734e-01 -9.95676517e-01 -1.29244579e-02 9.94591564e-02
-4.99721855e-01 2.80511290e-01 1.04821932e+00 -5.01647413e-01
8.39056194e-01 -2.28638458e+00 4.09622371e-01 2.85794437e-01
1.63111746e-01 -4.98379534e-03 -2.96287298e-01 2.90943921e-01
-4.45477515e-02 2.97280699e-01 -3.71144086e-01 -3.33121389e-01
-1.96711034e-01 2.73687363e-01 -2.25455821e-01 1.40374759e-02
2.57868856e-01 1.10844648e+00 -8.94300759e-01 -2.49216035e-01
2.56624699e-01 6.16469145e-01 -6.01589024e-01 5.11598527e-01
-3.55446547e-01 7.94001043e-01 -9.36781019e-02 4.40285057e-01
6.89187825e-01 -1.18054211e-01 6.08765185e-02 -3.27420235e-01
1.54873222e-01 -2.15277985e-01 -8.14983130e-01 1.76681685e+00
-8.32667589e-01 9.73038614e-01 -5.89526147e-02 -3.41343880e-01
1.20745897e+00 -7.49865621e-02 -1.39541984e-01 -7.79886782e-01
5.81664480e-02 2.65297502e-01 -2.05059603e-01 -1.76229462e-01
7.19320029e-02 -6.57670796e-01 -3.39647122e-02 6.49186909e-01
8.00459012e-02 -7.97348320e-01 2.84707814e-01 1.61107376e-01
8.19178939e-01 4.27399814e-01 3.81492883e-01 -4.76789296e-01
4.41669971e-01 -7.23957783e-03 3.78113657e-01 6.41606033e-01
5.71463168e-01 1.07015324e+00 6.79456115e-01 -6.93209827e-01
-1.67782652e+00 -1.04278779e+00 3.65702152e-01 9.52566445e-01
-2.00128287e-01 -1.70786083e-01 -1.28884900e+00 -6.97396517e-01
-2.96957850e-01 9.36134279e-01 -9.85535800e-01 9.48271453e-02
-7.16383874e-01 -1.98568285e-01 5.15602767e-01 4.73492742e-01
9.35266435e-01 -1.56727147e+00 -6.04707718e-01 -1.64768212e-02
2.23667622e-01 -6.73034489e-01 -7.57090986e-01 -8.86443630e-02
-7.52086878e-01 -6.89139724e-01 -1.03599179e+00 -7.02956557e-01
1.06197441e+00 -2.04143539e-01 1.61333132e+00 -1.23563185e-01
9.21721756e-03 2.38908455e-01 -2.20255822e-01 -3.09264332e-01
-7.89865136e-01 1.29401058e-01 -4.10503358e-01 4.66216169e-02
1.77911520e-02 -7.08446443e-01 -5.07799149e-01 1.54497221e-01
-1.04092789e+00 5.58824122e-01 6.49651468e-01 1.14896560e+00
6.70561433e-01 -1.44476578e-01 5.01559079e-01 -1.53591871e+00
4.40515012e-01 7.29697943e-02 -6.11463249e-01 2.10751697e-01
-4.06634331e-01 2.27267712e-01 8.93943965e-01 -4.34029013e-01
-1.37440693e+00 1.95273049e-02 -9.42016169e-02 -4.69107479e-01
-3.44400823e-01 -5.53014912e-02 -4.53225732e-01 -1.12256654e-01
7.40222394e-01 2.62642086e-01 1.53203234e-01 -3.67152542e-01
6.80785954e-01 2.84045160e-01 7.23316491e-01 -6.31054342e-01
1.07988119e+00 4.78524476e-01 -2.21119914e-02 -8.03707361e-01
-6.55628562e-01 4.11074549e-01 -5.09609103e-01 -1.38964429e-01
1.00521815e+00 -5.90500712e-01 -3.00367415e-01 5.69376528e-01
-9.72288430e-01 -9.37502503e-01 -6.99924469e-01 -2.30859101e-01
-7.30014026e-01 -2.89866328e-01 -4.66497481e-01 -3.98669809e-01
-3.40103805e-01 -9.84981418e-01 1.04224586e+00 4.00321007e-01
-4.03638005e-01 -1.15810490e+00 6.70355707e-02 1.26658212e-02
6.45854592e-01 4.27116394e-01 9.70092118e-01 9.21158493e-02
-4.17513490e-01 1.81093380e-01 -2.31543735e-01 3.23776662e-01
4.69792217e-01 2.04565506e-02 -1.10487747e+00 -3.78004014e-01
-2.26495966e-01 -1.81792185e-01 4.88624007e-01 2.48230264e-01
1.12110937e+00 -3.32351685e-01 -4.28901874e-02 1.06304121e+00
1.47437835e+00 5.08255899e-01 9.74197030e-01 3.97563577e-01
9.15388882e-01 2.25742623e-01 3.38782310e-01 -1.28527224e-01
-1.77229140e-02 6.83101118e-01 -2.81753372e-02 -5.82263410e-01
-4.01076883e-01 -5.14322698e-01 2.43114829e-01 5.30236304e-01
-1.35075301e-01 8.16387910e-05 -4.20144886e-01 1.65008470e-01
-1.23158336e+00 -1.09795690e+00 4.11547691e-01 1.85093057e+00
9.77720320e-01 8.68802220e-02 1.51049927e-01 -1.24848530e-01
6.62854493e-01 1.92846894e-01 -5.24040580e-01 -8.54803920e-01
-2.20034599e-01 4.15049583e-01 3.81256521e-01 5.60317516e-01
-6.24528468e-01 1.01149309e+00 7.11364317e+00 7.94826984e-01
-1.34588170e+00 -3.09968833e-02 1.11084020e+00 -1.06498845e-01
-7.14026988e-01 6.59041926e-02 -4.54355448e-01 6.70280397e-01
5.25948226e-01 -1.58836752e-01 9.72028315e-01 7.21315801e-01
-1.59579944e-02 -7.97579736e-02 -1.18157530e+00 1.22752261e+00
1.06362283e-01 -1.61086500e+00 2.68101901e-01 7.80163035e-02
1.12128496e+00 -5.12943387e-01 8.20294097e-02 1.73757821e-01
4.80453193e-01 -1.14990246e+00 6.91347361e-01 7.58756697e-01
1.62489557e+00 -8.23327124e-01 5.05629897e-01 -9.13578719e-02
-1.03082621e+00 1.76614866e-01 -3.56528521e-01 1.68544263e-01
2.33945623e-01 6.70331180e-01 -5.47244608e-01 3.73676211e-01
6.71651125e-01 4.94591624e-01 -6.89500630e-01 3.41675192e-01
-4.04081583e-01 6.38971269e-01 -2.65929788e-01 3.09230834e-01
1.39920160e-01 -4.35207844e-01 3.40138376e-01 1.12187243e+00
6.25056803e-01 1.53812751e-01 -1.84238449e-01 1.15734577e+00
-3.22443575e-01 -2.59470254e-01 -9.50922251e-01 8.80032964e-03
4.56805319e-01 1.30088520e+00 -7.68939257e-01 -6.68313086e-01
-2.32209757e-01 1.54655266e+00 1.86525006e-02 5.59620738e-01
-5.63372672e-01 -5.23922145e-01 5.90945482e-01 1.78905621e-01
3.90924931e-01 1.92705452e-01 -6.17061317e-01 -9.86944258e-01
-1.53950453e-01 -1.16318643e+00 4.04772498e-02 -1.33383548e+00
-1.18330348e+00 9.48369443e-01 -1.71544659e-03 -1.17937934e+00
-5.71162522e-01 -3.35966110e-01 -1.02470720e+00 1.12412941e+00
-9.79936302e-01 -1.46152508e+00 -5.20894170e-01 3.90097171e-01
6.55042827e-01 -5.11885643e-01 7.73698211e-01 -3.05777192e-02
-1.11196093e-01 7.47350216e-01 -7.23039359e-02 5.27446903e-02
8.12838554e-01 -1.44959164e+00 7.96832204e-01 5.99901259e-01
-4.37048860e-02 4.42254275e-01 7.84781933e-01 -5.81145406e-01
-1.30333102e+00 -9.94615972e-01 4.93556529e-01 -3.58108610e-01
2.41694480e-01 -4.42705214e-01 -7.01977730e-01 7.41845906e-01
7.00842798e-01 6.30705804e-02 3.88095379e-01 -3.53535086e-01
-3.15086126e-01 -4.00543600e-01 -1.59540224e+00 8.41163278e-01
1.15140617e+00 -5.65447390e-01 -4.31979269e-01 -3.30614388e-01
5.50971746e-01 -5.37041366e-01 -7.62624264e-01 -3.77984978e-02
7.71480441e-01 -1.16090024e+00 6.96701288e-01 -3.95324022e-01
7.53190517e-01 -5.36714554e-01 -5.67267723e-02 -1.85347855e+00
-4.42719102e-01 -8.74753714e-01 1.06605224e-01 1.49743044e+00
2.74164289e-01 -6.61019862e-01 6.62972391e-01 2.65712887e-01
2.10188687e-01 -6.16020381e-01 -4.13090795e-01 -6.46139443e-01
1.78752571e-01 1.77784398e-01 1.12147033e+00 7.62613177e-01
-6.39402568e-01 4.92769241e-01 -4.56007898e-01 -3.31336021e-01
7.27090120e-01 1.15720376e-01 1.18632746e+00 -9.38534319e-01
-4.20730472e-01 -3.89739335e-01 -3.66738915e-01 -6.74388051e-01
-3.92307527e-02 -5.45751691e-01 -1.62959620e-01 -1.35499418e+00
4.21990268e-02 -3.91880989e-01 1.55046865e-01 4.69046891e-01
-1.65735289e-01 8.61383080e-01 6.12812102e-01 -2.03383602e-02
-5.26195243e-02 1.74083501e-01 1.68710315e+00 -2.31039941e-01
-1.68907642e-01 -3.28753501e-01 -1.02627707e+00 5.21446466e-01
8.97003710e-01 -2.74091929e-01 -5.97099841e-01 -5.16370773e-01
3.50945324e-01 1.16923712e-02 2.77755231e-01 -1.05061722e+00
-3.21055561e-01 -3.11599493e-01 8.04461360e-01 -1.85078099e-01
2.41548210e-01 -7.34021783e-01 8.34289014e-01 2.60232717e-01
-4.14400399e-01 1.86288014e-01 1.78322256e-01 7.75962994e-02
-9.71638113e-02 -1.95684135e-01 1.21070349e+00 -4.52270627e-01
-6.53426468e-01 -4.34293523e-02 -1.84901178e-01 2.27690101e-01
8.87564063e-01 -4.05434728e-01 -5.60005233e-02 -6.03259265e-01
-8.16005349e-01 -2.70356834e-01 1.13350451e+00 2.48342365e-01
4.08391505e-01 -1.71495962e+00 -7.60534883e-01 6.17554724e-01
-1.65559128e-02 2.27623329e-01 9.58721116e-02 1.93251222e-01
-7.20772386e-01 -2.00545847e-01 -6.35018051e-01 -6.05188072e-01
-1.06791770e+00 3.36718142e-01 3.88136595e-01 -1.18244797e-01
-8.74547720e-01 6.98119938e-01 5.74435472e-01 -4.19525892e-01
-5.54242730e-01 7.98895024e-03 3.52658778e-01 -2.86059946e-01
7.43691564e-01 2.34395653e-01 -2.19786838e-01 -3.85132998e-01
-5.92186861e-02 9.40618336e-01 1.01056471e-01 -5.49656749e-01
1.22080827e+00 2.70689968e-02 -1.75164714e-01 4.80810434e-01
1.24471319e+00 1.39256582e-01 -1.62668836e+00 1.19261272e-01
-5.39866745e-01 -6.04183137e-01 -3.50211918e-01 -9.82456923e-01
-1.35099471e+00 7.99938202e-01 5.76795280e-01 5.09194620e-02
1.56745660e+00 -1.17664240e-01 5.59516847e-01 1.96253359e-01
4.03766036e-01 -8.88009906e-01 7.98797980e-02 3.48737866e-01
1.04325676e+00 -8.76960814e-01 -1.46749556e-01 -5.23774922e-02
-8.82397115e-01 1.13318384e+00 4.91642118e-01 -4.72198039e-01
4.39386427e-01 5.27757347e-01 4.80484009e-01 -9.36716199e-02
-4.74967301e-01 2.56741583e-01 3.43791038e-01 7.83236146e-01
2.86824852e-01 9.84149575e-02 7.34275877e-02 -6.66358620e-02
-6.99448287e-01 2.44569555e-01 4.41004217e-01 6.78157091e-01
-3.51520032e-02 -1.24030256e+00 -5.52451491e-01 6.39201880e-01
-2.49791935e-01 -3.56940292e-02 -5.14114320e-01 5.76546431e-01
6.78811371e-02 5.35844326e-01 2.79781163e-01 -3.67591679e-01
1.38427481e-01 1.82662085e-01 6.91617429e-01 -5.33481240e-01
-6.51675403e-01 5.90132326e-02 -3.34088951e-02 -6.19099975e-01
-3.29609424e-01 -3.04810852e-01 -8.39552164e-01 -6.03505731e-01
1.53701782e-01 -2.77980771e-02 4.06518281e-01 8.93428564e-01
3.54614079e-01 6.39774740e-01 7.71514297e-01 -1.20849431e+00
1.09120592e-01 -9.22075212e-01 -4.42418158e-01 6.31704032e-01
2.40578339e-01 -3.30699027e-01 -1.86498612e-01 4.75872278e-01] | [11.712017059326172, -0.44369709491729736] |
3a3a6c5c-87cf-4b91-b721-afbd72847daa | spatio-temporal-modeling-for-large-scale | 2103.07636 | null | https://arxiv.org/abs/2103.07636v1 | https://arxiv.org/pdf/2103.07636v1.pdf | Spatio-temporal Modeling for Large-scale Vehicular Networks Using Graph Convolutional Networks | The effective deployment of connected vehicular networks is contingent upon maintaining a desired performance across spatial and temporal domains. In this paper, a graph-based framework, called SMART, is proposed to model and keep track of the spatial and temporal statistics of vehicle-to-infrastructure (V2I) communication latency across a large geographical area. SMART first formulates the spatio-temporal performance of a vehicular network as a graph in which each vertex corresponds to a subregion consisting of a set of neighboring location points with similar statistical features of V2I latency and each edge represents the spatio-correlation between latency statistics of two connected vertices. Motivated by the observation that the complete temporal and spatial latency performance of a vehicular network can be reconstructed from a limited number of vertices and edge relations, we develop a graph reconstruction-based approach using a graph convolutional network integrated with a deep Q-networks algorithm in order to capture the spatial and temporal statistic of feature map pf latency performance for a large-scale vehicular network. Extensive simulations have been conducted based on a five-month latency measurement study on a commercial LTE network. Our results show that the proposed method can significantly improve both the accuracy and efficiency for modeling and reconstructing the latency performance of large vehicular networks. | ['H. Vincent Poor', 'Walid Saad', 'Guangming Shiyz', 'Yingyu Li', 'Yong Xiao', 'Juntong Liu'] | 2021-03-13 | null | null | null | null | ['graph-reconstruction'] | ['graphs'] | [-8.33747864e-01 -1.77926004e-01 -2.53036022e-01 -2.51725852e-01
-3.13094109e-01 -3.98794234e-01 6.49862289e-01 2.08007529e-01
1.26065820e-01 5.51696956e-01 -1.61432281e-01 -8.16620231e-01
-5.79453111e-01 -1.04018509e+00 -8.27977896e-01 -3.22928578e-01
-1.30708444e+00 4.25586402e-01 3.57250929e-01 -1.77191570e-01
-3.75526547e-02 8.41880500e-01 -1.37411320e+00 -2.01438606e-01
6.71781361e-01 1.52090895e+00 1.59492433e-01 6.69564545e-01
-1.81150716e-02 7.61201560e-01 -4.29141104e-01 -1.58815682e-01
1.91255301e-01 2.40528762e-01 -4.66685474e-01 -8.78966227e-02
-8.91951397e-02 -3.36049348e-01 -1.23483157e+00 5.68106651e-01
1.25488222e-01 1.71015859e-02 5.37904978e-01 -2.09377933e+00
-1.33035332e-01 3.31950247e-01 -2.06357852e-01 4.79907990e-01
-4.45412807e-02 1.97949242e-02 7.01062202e-01 -7.11024106e-01
5.99087775e-01 9.74500835e-01 7.27819324e-01 -2.57894367e-01
-8.39940727e-01 -8.01964521e-01 2.00546488e-01 6.18759513e-01
-2.14789033e+00 -4.75163966e-01 5.97011268e-01 -4.36905414e-01
9.48414505e-01 1.03345580e-01 8.80311370e-01 4.78825599e-01
7.02596426e-01 5.66793442e-01 2.79041201e-01 4.25683379e-01
1.30587354e-01 1.24565847e-02 -2.53933996e-01 5.23693144e-01
8.45541209e-02 4.00760055e-01 -2.10637659e-01 -1.21209614e-01
4.10490155e-01 4.92411107e-02 1.60213560e-01 -5.29843330e-01
-7.05010295e-01 7.27399528e-01 8.50266099e-01 7.39163086e-02
-6.61316514e-01 7.54377544e-01 5.87056696e-01 4.83252436e-01
6.20555997e-01 -2.63725221e-01 -3.68695319e-01 -7.89574236e-02
-1.27340257e+00 1.38342410e-01 6.97404683e-01 1.47141516e+00
1.04016912e+00 2.45305568e-01 -1.56411991e-01 1.35846645e-01
4.04974967e-01 8.48093569e-01 -5.49371839e-01 -9.26071405e-01
4.24973786e-01 5.01596928e-01 -3.83346342e-02 -1.34057224e+00
-7.01181710e-01 -4.88340527e-01 -1.00825739e+00 -1.18799165e-01
-4.19643000e-02 -2.34583676e-01 -7.23463476e-01 1.72750533e+00
2.96108127e-01 1.03824759e+00 -1.51683807e-01 6.70122206e-01
8.94973040e-01 9.32590961e-01 3.34452875e-02 -2.72080779e-01
8.93647492e-01 -4.92984772e-01 -6.15537822e-01 7.51150921e-02
6.21821702e-01 -2.84399360e-01 6.99712113e-02 -2.95320272e-01
-1.07922506e+00 -3.88312727e-01 -8.97666514e-01 5.09009242e-01
-4.80800718e-01 -2.47664005e-01 6.02912068e-01 5.55365443e-01
-1.56940532e+00 4.12569344e-01 -7.21795917e-01 -2.62103379e-01
4.03119624e-01 6.38416052e-01 -3.73032838e-01 -2.24814370e-01
-1.30765057e+00 4.55302775e-01 7.51710609e-02 3.24627161e-01
-1.40963304e+00 -1.02850568e+00 -7.04231262e-01 1.95606291e-01
4.78863806e-01 -4.49165791e-01 7.63266623e-01 -4.64648873e-01
-9.45171297e-01 1.43408507e-01 -4.12440777e-01 -6.31526649e-01
2.65508294e-01 3.72467965e-01 -1.09080112e+00 2.08293006e-01
3.05627584e-01 2.38058567e-01 4.42827791e-01 -1.25398934e+00
-8.53045821e-01 -1.68050766e-01 1.34585500e-01 -2.47466698e-01
1.05332069e-01 -7.13695362e-02 -1.18571520e+00 -2.41600703e-02
-1.30215839e-01 -9.56247628e-01 -2.29656532e-01 -3.24665099e-01
-3.75828683e-01 -1.78869352e-01 1.07040727e+00 -5.80549300e-01
1.38187540e+00 -1.97031808e+00 -2.10038558e-01 8.95254731e-01
7.45860815e-01 1.96105808e-01 -5.12059748e-01 8.69809628e-01
3.45877111e-01 1.33604437e-01 4.73893404e-01 -2.07451388e-01
-2.47142419e-01 8.96211490e-02 -5.53441159e-02 8.57956529e-01
-5.57754468e-03 1.26535356e+00 -1.03073454e+00 -3.45831454e-01
4.79719073e-01 6.78100586e-01 -3.74443680e-01 2.69052267e-01
4.52222563e-02 4.54333454e-01 -7.14942873e-01 5.77294827e-01
1.08006144e+00 -2.93089956e-01 2.86361337e-01 -1.28179073e-01
-3.50030780e-01 6.36005402e-02 -8.50026965e-01 1.22857761e+00
-6.41060293e-01 9.90409672e-01 3.14822495e-01 -1.05080295e+00
8.81980002e-01 3.77336234e-01 1.14717150e+00 -1.25468922e+00
2.59619445e-01 -1.18384548e-02 -2.54665166e-01 -3.89323890e-01
5.30752182e-01 6.37825549e-01 -1.51251838e-01 1.70831576e-01
-1.70028314e-01 2.75938421e-01 1.12822153e-01 6.17060900e-01
1.53214419e+00 -6.58432961e-01 -2.09971920e-01 -3.32287669e-01
4.24910575e-01 -1.04252346e-01 4.42991972e-01 5.45809090e-01
-2.68106610e-01 -8.19453672e-02 6.42849684e-01 -4.69086319e-01
-1.25373626e+00 -1.33734536e+00 4.12244117e-03 6.43700659e-01
6.03538811e-01 -5.59536338e-01 -3.42178911e-01 -3.22388947e-01
3.82132202e-01 4.77386296e-01 -6.49173141e-01 -2.40151092e-01
-5.04499674e-01 -2.09220722e-01 4.28071052e-01 1.50992855e-01
3.46386492e-01 -5.07041812e-01 1.36486262e-01 4.40868348e-01
1.27632944e-02 -1.46091521e+00 -4.10052657e-01 -2.53648549e-01
-3.75591695e-01 -1.16372752e+00 2.82318108e-02 -7.12015927e-01
2.19118699e-01 9.78823185e-01 1.20781374e+00 3.83158296e-01
2.35892490e-01 4.23772871e-01 -2.43627846e-01 -7.83343017e-02
-3.90372515e-01 2.17808992e-01 2.73532450e-01 2.35408723e-01
2.18790472e-01 -9.30042624e-01 -8.62603188e-01 5.26383698e-01
-4.67846811e-01 -1.41882092e-01 4.30794835e-01 1.36766821e-01
5.29608250e-01 3.73019367e-01 6.51638091e-01 -3.43060136e-01
3.38319749e-01 -1.49643672e+00 -8.46182704e-01 1.15391895e-01
-7.08410084e-01 -3.13802361e-01 6.13163531e-01 -2.13628098e-01
-2.67496854e-01 -3.01125139e-01 1.05910294e-01 -5.92486739e-01
2.17285156e-01 8.27600062e-01 -2.94492841e-01 -2.65956700e-01
2.76255943e-02 1.71731263e-01 3.87290330e-03 7.18488693e-02
5.52874148e-01 5.94341159e-01 4.23424274e-01 -1.54552087e-01
1.06513011e+00 6.91494465e-01 3.89658064e-01 -9.93198991e-01
1.98826436e-02 -7.98645377e-01 -4.46628124e-01 -7.30776727e-01
4.44503427e-01 -1.32301986e+00 -1.27909577e+00 1.75240591e-01
-1.15929651e+00 -4.47067440e-01 2.64664590e-01 5.22845268e-01
-4.19793427e-01 1.37026876e-01 -2.80140489e-01 -7.41703093e-01
-6.95398375e-02 -1.08851194e+00 9.91943717e-01 -1.69208407e-01
3.55878383e-01 -1.23716497e+00 1.15756467e-02 -2.42920145e-01
6.89910650e-01 1.48264840e-01 6.07377350e-01 -3.04797977e-01
-1.27886224e+00 -4.32909995e-01 -9.35104251e-01 -1.73534244e-01
-1.23398557e-01 1.45857647e-01 -5.43252349e-01 -6.42452061e-01
-6.48518145e-01 4.39184308e-01 3.07502866e-01 8.23884308e-01
9.92856383e-01 -2.49111235e-01 -7.86215782e-01 1.06345296e+00
1.57851946e+00 5.79760559e-02 6.78794682e-01 3.93697135e-02
9.32521880e-01 4.29782927e-01 4.56299007e-01 5.00051320e-01
1.02984130e+00 8.79877388e-01 1.04305911e+00 -7.98497275e-02
-2.05910459e-01 -3.68076861e-01 8.15980211e-02 7.86361814e-01
-9.19421092e-02 -6.30582154e-01 -9.38945353e-01 1.01040018e+00
-1.82578135e+00 -1.02323234e+00 -4.45841223e-01 2.06895328e+00
-3.41934562e-01 1.44897342e-01 3.23472768e-01 -1.96695864e-01
6.99610770e-01 2.40779907e-01 -4.39282358e-01 -3.12460989e-01
-8.77144337e-02 -3.52518171e-01 1.20282352e+00 4.90620524e-01
-7.77738094e-01 9.60691631e-01 6.31456470e+00 9.76074934e-01
-1.16047466e+00 1.82577044e-01 3.14200610e-01 4.24495563e-02
-3.98748040e-01 3.30819897e-02 -4.60089773e-01 5.78338742e-01
1.88135386e+00 -5.69112659e-01 7.42594004e-01 5.96643507e-01
8.70797992e-01 1.22503191e-01 -8.35808992e-01 7.86678910e-01
-4.79329646e-01 -1.86044800e+00 -3.24002877e-02 5.26928782e-01
7.08269119e-01 8.81221354e-01 1.74268454e-01 2.44332105e-01
3.15475643e-01 -1.21868336e+00 7.27151871e-01 6.91252589e-01
1.20783973e+00 -1.07706583e+00 7.66632318e-01 1.73346505e-01
-2.18635249e+00 -2.56259143e-01 -1.12022616e-01 1.62666544e-01
4.08561021e-01 4.91667330e-01 -1.04201627e+00 8.37801754e-01
5.88245511e-01 8.90604913e-01 -2.89635450e-01 1.21752679e+00
3.16776544e-01 5.88546276e-01 -3.19606245e-01 2.18016431e-02
5.46534717e-01 -1.83514267e-01 5.37783921e-01 9.88136113e-01
5.90181291e-01 -1.15309410e-01 1.25084698e-01 5.79644263e-01
-1.80893198e-01 6.83172867e-02 -1.08915055e+00 1.57726958e-01
1.07480502e+00 1.17887688e+00 -5.03489733e-01 -2.15633392e-01
-7.39219606e-01 2.48090237e-01 3.09357017e-01 8.31871092e-01
-1.05493534e+00 -1.71854347e-01 1.22387862e+00 5.23395956e-01
6.34878516e-01 -6.66845500e-01 3.97305228e-02 -4.86153334e-01
-1.61292568e-01 -1.75456971e-01 -1.74222678e-01 -6.55151844e-01
-7.81700313e-01 7.55058825e-01 -4.64534648e-02 -1.46293592e+00
-1.89409852e-01 -4.75872383e-02 -8.32566559e-01 7.89499283e-01
-1.67282426e+00 -1.26373529e+00 -3.66720796e-01 9.86228764e-01
1.15489416e-01 -3.34287763e-01 3.10424864e-01 6.17044091e-01
-4.31562215e-01 5.12929916e-01 4.83709008e-01 -3.92662622e-02
-1.91612914e-01 -6.39985323e-01 8.37663174e-01 8.50755394e-01
-8.88537243e-02 1.18462935e-01 6.80642843e-01 -8.17584634e-01
-1.84677494e+00 -1.72466636e+00 9.46485698e-01 -2.82521069e-01
9.64966297e-01 -3.64048481e-01 -6.64261699e-01 7.31475651e-01
-9.31921750e-02 5.99841535e-01 2.03769714e-01 -2.76116771e-03
-1.59176558e-01 -4.72702920e-01 -7.88050294e-01 3.82140845e-01
1.14106703e+00 -8.42897713e-01 4.08967108e-01 3.42782617e-01
8.64855587e-01 -8.02261084e-02 -9.19390678e-01 4.30638731e-01
5.56081116e-01 -7.51076341e-01 1.03261423e+00 -5.18449545e-01
-3.30535412e-01 -3.46066564e-01 -3.28168392e-01 -1.26509333e+00
-4.45736200e-01 -5.79055667e-01 -3.30345958e-01 8.39130580e-01
2.65602648e-01 -3.77501637e-01 8.48456621e-01 2.19678640e-01
-2.73585200e-01 -6.37219727e-01 -1.32176590e+00 -9.54346657e-01
-1.47988066e-01 -1.12116945e+00 1.22222030e+00 3.45925897e-01
-3.95809919e-01 1.67812139e-01 -4.48882431e-01 9.07493174e-01
6.58906758e-01 -6.24756701e-02 1.05466723e+00 -1.34688485e+00
6.30344927e-01 -3.70142013e-01 -9.69101548e-01 -1.17602015e+00
3.57438356e-01 -9.99209225e-01 -1.91815481e-01 -1.55031765e+00
-9.28658918e-02 -9.22653079e-01 -3.33815545e-01 -4.24241364e-01
4.58561361e-01 1.46447405e-01 -1.20034859e-01 2.51037270e-01
-1.04408252e+00 6.21790051e-01 8.71918619e-01 -1.70667708e-01
-1.82100713e-01 3.80520195e-01 -1.21267520e-01 2.06875682e-01
6.71709061e-01 -3.57494265e-01 -7.82254338e-01 -4.25761670e-01
3.65643144e-01 7.65822887e-01 4.73605305e-01 -9.68925655e-01
4.54347163e-01 -1.52674630e-01 -1.59526467e-01 -1.08943689e+00
3.43519062e-01 -1.11223269e+00 6.71505809e-01 2.65106529e-01
2.42678881e-01 4.18236881e-01 2.47664481e-01 9.52495933e-01
-2.42400572e-01 8.60879719e-01 2.94993013e-01 5.81957757e-01
-1.12690043e+00 1.13406003e+00 -5.71991861e-01 -1.41491011e-01
1.32625103e+00 -2.57671595e-01 -1.62694767e-01 -8.37162912e-01
-3.95158798e-01 7.63505518e-01 3.65987271e-01 4.80626434e-01
8.44156802e-01 -1.65772605e+00 -6.94777548e-01 1.56919435e-01
3.13034385e-01 -5.98021209e-01 6.80203378e-01 9.63402510e-01
-4.87418056e-01 9.29263175e-01 -1.75691873e-01 -7.36256599e-01
-9.56819117e-01 8.12467396e-01 3.43169838e-01 -1.34334221e-01
-5.75116515e-01 5.28875709e-01 3.05900555e-02 -2.00025350e-01
9.00295526e-02 -3.89144011e-02 -2.45224044e-01 -1.74193531e-01
3.05207044e-01 6.32534266e-01 1.83923796e-01 -1.40297103e+00
-6.15954936e-01 4.22614008e-01 4.80603486e-01 3.14627707e-01
1.19927025e+00 -8.19952488e-01 -5.18345460e-02 1.31406710e-01
1.71659100e+00 -9.71930549e-02 -1.18316102e+00 -3.66772771e-01
-2.02269241e-01 -3.12838316e-01 4.33881909e-01 -1.00853674e-01
-1.70989430e+00 5.04730821e-01 4.38886464e-01 5.74290931e-01
9.29636955e-01 2.05940664e-01 1.05054307e+00 -1.59853082e-02
7.30400205e-01 -8.18912566e-01 -4.04062152e-01 6.29372299e-01
4.26779002e-01 -1.04336286e+00 -2.71615744e-01 -5.57974935e-01
-1.25983149e-01 7.95359910e-01 2.67694086e-01 -2.10785285e-01
1.37120104e+00 1.65440246e-01 -2.25459799e-01 -7.21216440e-01
-9.62650716e-01 -4.55537558e-01 3.72262150e-01 6.83819115e-01
-2.07144395e-01 4.74074662e-01 1.77314147e-01 2.81320333e-01
-3.60478051e-02 -1.46252558e-01 3.23191762e-01 2.63390243e-01
-3.03599775e-01 -6.07459903e-01 1.56746596e-01 5.94023943e-01
-2.35809702e-02 -1.12799904e-03 3.65042448e-01 8.89111400e-01
1.46098575e-02 1.39015210e+00 5.31077504e-01 -1.07824683e+00
2.20142990e-01 -7.08304822e-01 -6.28263205e-02 -1.38290063e-01
-5.08727580e-02 -3.17575008e-01 2.90417165e-01 -1.07451868e+00
-1.43686935e-01 -4.29403722e-01 -1.22812927e+00 -1.24016416e+00
2.61841174e-02 3.70489240e-01 1.02072954e+00 8.92877877e-01
6.96066260e-01 6.28804028e-01 1.37294221e+00 -8.01413357e-01
2.80051708e-01 -5.08471489e-01 -9.97298837e-01 8.04706737e-02
6.62249744e-01 -7.43010938e-01 -2.41085023e-01 -7.84466267e-01] | [6.347537517547607, 1.8739615678787231] |
7f1ff4b4-05cb-4f4d-9f86-30714785fb78 | dut-learning-video-stabilization-by-simply | 2011.14574 | null | https://arxiv.org/abs/2011.14574v3 | https://arxiv.org/pdf/2011.14574v3.pdf | DUT: Learning Video Stabilization by Simply Watching Unstable Videos | Previous deep learning-based video stabilizers require a large scale of paired unstable and stable videos for training, which are difficult to collect. Traditional trajectory-based stabilizers, on the other hand, divide the task into several sub-tasks and tackle them subsequently, which are fragile in textureless and occluded regions regarding the usage of hand-crafted features. In this paper, we attempt to tackle the video stabilization problem in a deep unsupervised learning manner, which borrows the divide-and-conquer idea from traditional stabilizers while leveraging the representation power of DNNs to handle the challenges in real-world scenarios. Technically, DUT is composed of a trajectory estimation stage and a trajectory smoothing stage. In the trajectory estimation stage, we first estimate the motion of keypoints, initialize and refine the motion of grids via a novel multi-homography estimation strategy and a motion refinement network, respectively, and get the grid-based trajectories via temporal association. In the trajectory smoothing stage, we devise a novel network to predict dynamic smoothing kernels for trajectory smoothing, which can well adapt to trajectories with different dynamic patterns. We exploit the spatial and temporal coherence of keypoints and grid vertices to formulate the training objectives, resulting in an unsupervised training scheme. Experiment results on public benchmarks show that DUT outperforms state-of-the-art methods both qualitatively and quantitatively. The source code is available at https://github.com/Annbless/DUTCode. | ['Stephen J. Maybank', 'DaCheng Tao', 'Jing Zhang', 'Yufei Xu'] | 2020-11-30 | null | null | null | null | ['video-stabilization', 'homography-estimation'] | ['computer-vision', 'computer-vision'] | [-4.03640896e-01 -4.79347169e-01 -2.69044250e-01 9.21551287e-02
-5.37787557e-01 -4.67669010e-01 5.21593213e-01 -2.26189375e-01
-1.79022461e-01 6.06027603e-01 2.65327781e-01 -1.55278057e-01
3.05286236e-02 -5.88258088e-01 -8.48786414e-01 -1.15461159e+00
2.50248536e-02 2.86319964e-02 3.88234198e-01 -3.20271403e-01
1.17518954e-01 2.19520926e-01 -1.16496778e+00 -1.23004526e-01
1.02646327e+00 9.51038659e-01 1.37977421e-01 3.10176492e-01
2.26592839e-01 8.81650150e-01 1.28411934e-01 -2.25720592e-02
1.36923552e-01 -4.36730564e-01 -3.41715366e-01 2.45440453e-01
4.03486550e-01 -3.15432310e-01 -8.83314729e-01 1.07343471e+00
3.84655297e-01 3.90640318e-01 4.61344808e-01 -1.11347795e+00
-5.54496229e-01 3.52013618e-01 -6.90305114e-01 3.25938702e-01
2.74786949e-02 4.29908991e-01 6.62389398e-01 -9.44764197e-01
6.41690612e-01 1.07344782e+00 7.67138898e-01 5.59869230e-01
-1.14511192e+00 -6.81279659e-01 4.80987549e-01 3.98251235e-01
-1.43199253e+00 -6.86110139e-01 9.88895714e-01 -5.53520024e-01
4.16744500e-01 1.58231575e-02 7.61084974e-01 1.21547580e+00
1.07303619e-01 8.78853023e-01 6.68327928e-01 1.49450079e-01
1.50609374e-01 -3.86007816e-01 -1.73213467e-01 8.59412849e-01
-1.27019044e-02 2.05383338e-02 -4.12436336e-01 1.17511362e-01
1.17301059e+00 2.35200956e-01 -5.31362295e-01 -6.62738621e-01
-1.40120959e+00 7.05128849e-01 5.90724409e-01 3.04104328e-01
-3.24463516e-01 2.37550691e-01 5.23660481e-01 2.19605938e-01
5.95302403e-01 2.30612103e-02 -2.01862231e-01 -1.66672602e-01
-1.07397020e+00 2.33977601e-01 4.21770096e-01 8.99974942e-01
8.08850646e-01 3.15938324e-01 -1.91380545e-01 3.41877997e-01
2.31934115e-01 4.39673126e-01 4.46277350e-01 -8.89743805e-01
6.19353652e-01 6.26212120e-01 2.84146219e-01 -1.48731148e+00
-3.92911047e-01 -2.92063326e-01 -1.44594800e+00 -7.62075335e-02
5.00208378e-01 -2.76327759e-01 -7.83371508e-01 1.76397991e+00
6.69796884e-01 1.12081265e+00 -1.48230836e-01 1.07752335e+00
4.89233375e-01 8.95498037e-01 -2.35015661e-01 -6.30322814e-01
8.16968441e-01 -1.27261329e+00 -6.58029735e-01 3.01812410e-01
5.09462476e-01 -4.92182344e-01 1.14341056e+00 2.65607089e-01
-1.09871638e+00 -5.47548234e-01 -8.46527755e-01 -6.62111565e-02
6.22344837e-02 4.11539525e-01 3.34946156e-01 2.74230279e-02
-1.14512217e+00 8.28017771e-01 -1.40258193e+00 -1.35182187e-01
4.28866059e-01 2.88479030e-01 -2.00955719e-01 3.55895847e-01
-9.90306020e-01 4.20937002e-01 2.89581120e-01 5.05774558e-01
-1.03949511e+00 -5.68345845e-01 -9.97971833e-01 -4.22021896e-02
6.49644673e-01 -6.21308744e-01 9.12860274e-01 -1.01995540e+00
-1.74520087e+00 3.71807516e-01 -2.83168316e-01 -2.65093833e-01
7.67192423e-01 -2.48835310e-01 -2.46620506e-01 9.98210013e-02
-2.76098065e-02 1.40462399e-01 1.20791340e+00 -9.91635323e-01
-5.91974437e-01 -1.39824944e-02 -1.54717550e-01 1.80308223e-01
-4.22122866e-01 -2.72997111e-01 -7.79788494e-01 -8.26339364e-01
-2.88347458e-03 -9.32824492e-01 -4.52703118e-01 -7.42741525e-02
-3.33924949e-01 -1.87395588e-01 1.10029924e+00 -5.86142838e-01
1.57704341e+00 -2.10814595e+00 7.77580678e-01 4.58060019e-02
6.25189602e-01 4.49691296e-01 -2.10094545e-02 2.66198456e-01
4.24293056e-02 -2.39260033e-01 -5.35600912e-03 -5.06671727e-01
-1.96083099e-01 -5.90953119e-02 -5.18421352e-01 8.78102064e-01
1.81343891e-02 8.20987999e-01 -1.13904405e+00 -3.55911016e-01
3.78495544e-01 4.54345137e-01 -5.43015659e-01 1.57683372e-01
-2.38864854e-01 9.55057800e-01 -7.28607953e-01 5.71470201e-01
6.23381913e-01 -5.03027916e-01 5.11941314e-02 -2.88455933e-01
-4.68900323e-01 -1.24656215e-01 -1.35891438e+00 1.84503913e+00
-2.57303596e-01 5.07370532e-01 1.02622777e-01 -1.17148507e+00
7.31179237e-01 3.06419075e-01 7.11094201e-01 -3.58984858e-01
2.53896445e-01 2.85891980e-01 -4.37942803e-01 -5.30335188e-01
3.95397574e-01 2.31922254e-01 5.49773388e-02 1.15276441e-01
-4.54824865e-02 3.05519760e-01 1.74716920e-01 1.61132529e-01
9.60619569e-01 3.58893484e-01 -3.90008055e-02 -3.27840805e-01
8.79508853e-01 -1.45277336e-01 9.16776896e-01 3.02924454e-01
-2.16580257e-01 6.11818910e-01 5.91129720e-01 -8.33533466e-01
-1.15523946e+00 -9.50843990e-01 2.35335559e-01 6.41660035e-01
6.67363644e-01 -5.38847148e-01 -7.80403435e-01 -5.84565878e-01
-1.74049944e-01 1.71245839e-02 -6.30166471e-01 -1.30158529e-01
-9.48560357e-01 -6.33543968e-01 2.01705709e-01 4.79128450e-01
5.32510936e-01 -9.30967152e-01 -3.46324444e-01 3.52307200e-01
-2.37138331e-01 -9.65583086e-01 -9.64467645e-01 -3.17735672e-01
-7.57221401e-01 -1.05587888e+00 -8.86894047e-01 -8.52046430e-01
5.72990000e-01 3.91361803e-01 6.94741130e-01 3.26485306e-01
3.01289231e-01 1.10171191e-01 -2.86432177e-01 3.40349764e-01
1.79462843e-02 3.17396462e-01 5.09590447e-01 4.14705217e-01
-6.25508726e-02 -8.90292943e-01 -8.86385143e-01 5.38208306e-01
-8.23459506e-01 3.49026084e-01 3.36052477e-01 9.81802166e-01
7.04616070e-01 -6.80020601e-02 4.13093537e-01 -3.81157249e-01
3.34706843e-01 -6.69231355e-01 -8.42669606e-01 1.60728201e-01
-2.19531238e-01 -1.07452482e-01 1.00561523e+00 -8.16830754e-01
-7.37341583e-01 3.30834925e-01 8.40419456e-02 -1.16641402e+00
9.99067649e-02 4.93862063e-01 -2.12902665e-01 -2.23399445e-01
3.64745378e-01 5.87158263e-01 1.33326650e-01 -3.49713415e-01
4.26984370e-01 2.27039188e-01 6.38562024e-01 -5.86650431e-01
1.19882846e+00 6.65799916e-01 -4.98198383e-02 -7.11334944e-01
-7.09820509e-01 -4.51212287e-01 -6.99695706e-01 -4.05503929e-01
8.66375804e-01 -1.06064498e+00 -8.40976834e-01 9.85233426e-01
-9.62363064e-01 -7.64809668e-01 -2.03263704e-02 3.76746327e-01
-7.33982921e-01 5.35820246e-01 -7.08397925e-01 -3.41389000e-01
-2.75699705e-01 -1.28251827e+00 1.04161716e+00 5.00467956e-01
2.16093600e-01 -1.05178666e+00 2.66801268e-01 -1.39944926e-01
3.11006546e-01 5.39861321e-01 4.27481294e-01 -8.89154673e-02
-9.81265485e-01 1.44850507e-01 -7.80499652e-02 5.67547306e-02
1.98661134e-01 8.81491825e-02 -4.28039968e-01 -6.12655044e-01
8.82549286e-02 -2.50793755e-01 9.77034390e-01 5.70142329e-01
1.14499164e+00 -5.51215827e-01 -2.97963113e-01 1.20327616e+00
1.34302354e+00 -1.16868071e-01 6.10645831e-01 3.50871652e-01
1.23747027e+00 1.70789227e-01 5.91018856e-01 4.66247648e-01
5.40147841e-01 8.03845227e-01 3.12368214e-01 -8.85598361e-03
2.06101626e-01 -3.23521018e-01 5.77941120e-01 1.18380678e+00
-3.87278587e-01 -9.67237726e-02 -8.67643416e-01 5.57487369e-01
-2.50552130e+00 -1.14494717e+00 -1.67480215e-01 1.96002626e+00
8.42313528e-01 -1.21889964e-01 3.34009051e-01 -2.04700127e-01
7.12529421e-01 6.05607808e-01 -7.13521242e-01 3.72082472e-01
5.62970042e-02 -3.46946537e-01 3.13446313e-01 5.04629076e-01
-1.40090144e+00 1.29304111e+00 4.57085133e+00 1.10585368e+00
-1.55495238e+00 1.07135242e-02 7.40593374e-01 -1.39257565e-01
-1.44348562e-01 -5.26483469e-02 -7.17160821e-01 8.28769267e-01
7.01704443e-01 -1.92851916e-01 6.20927811e-01 6.55330658e-01
7.98807323e-01 2.96946138e-01 -8.59230340e-01 1.06058979e+00
-2.11915344e-01 -1.83544528e+00 -6.45919070e-02 -2.21774817e-01
8.67312431e-01 4.79720347e-02 2.06704810e-01 1.21541850e-01
3.40921618e-02 -7.63952553e-01 9.11570191e-01 9.24432516e-01
6.79064691e-01 -6.88308358e-01 3.89928073e-01 3.24247837e-01
-1.70442343e+00 -7.80928358e-02 -3.41686338e-01 -8.97875279e-02
3.45686316e-01 3.73495579e-01 6.33281246e-02 7.02911198e-01
7.99925387e-01 1.49565911e+00 -3.56644958e-01 1.13328242e+00
-7.92976841e-02 5.85009754e-01 -2.05168411e-01 1.56904042e-01
4.75830078e-01 -7.02733219e-01 6.95115566e-01 9.29379225e-01
5.43139279e-01 1.99526295e-01 3.91213536e-01 7.29673326e-01
4.39424403e-02 1.44307017e-02 -4.03944701e-01 1.20336637e-01
3.49434614e-01 1.35984206e+00 -7.77240276e-01 -3.63115788e-01
-4.25580472e-01 9.49959219e-01 4.37956125e-01 5.99613428e-01
-1.02267849e+00 -1.46791220e-01 5.75156033e-01 1.25515357e-01
3.96516293e-01 -5.65206170e-01 -7.58735910e-02 -1.89491773e+00
1.97253227e-01 -9.93625760e-01 1.72826454e-01 -4.76428330e-01
-1.07542562e+00 6.05165422e-01 -1.82902768e-01 -1.60049939e+00
-7.52605647e-02 -2.29396746e-01 -1.08138359e+00 6.29447222e-01
-1.28827000e+00 -1.39674270e+00 -5.91799140e-01 8.95088315e-01
6.06855869e-01 -1.73187271e-01 2.09343359e-01 3.53888869e-01
-1.06504464e+00 3.86766225e-01 5.47365785e-01 1.86884850e-01
6.06900215e-01 -1.12918639e+00 4.79471564e-01 1.20112050e+00
-9.73732620e-02 5.86591959e-01 5.80519378e-01 -6.24000907e-01
-1.52989674e+00 -1.38922429e+00 5.44242859e-01 -3.20425242e-01
1.09211302e+00 -3.51352751e-01 -1.08281171e+00 5.44191122e-01
1.27521269e-02 2.16612488e-01 1.10223452e-02 -2.28908494e-01
1.68257058e-02 -2.29754552e-01 -5.11284649e-01 8.42589796e-01
1.18304420e+00 -3.63150895e-01 -1.20801657e-01 2.40964070e-01
7.44136989e-01 -7.92409658e-01 -5.98379374e-01 2.89771438e-01
3.57184291e-01 -8.48506212e-01 9.38022316e-01 -5.16305149e-01
5.61532557e-01 -7.31410563e-01 3.26962501e-01 -1.18749273e+00
-4.09326881e-01 -1.21587348e+00 -5.68871379e-01 1.15161538e+00
1.97423548e-02 -3.50775838e-01 8.73409390e-01 2.05543473e-01
-3.20200324e-01 -1.08507752e+00 -9.52385128e-01 -7.24147379e-01
6.78222179e-02 -2.47121905e-03 4.78743643e-01 1.04070926e+00
-1.50696442e-01 1.32203490e-01 -7.68282115e-01 3.21327686e-01
5.49018621e-01 2.46309564e-01 8.67854357e-01 -7.97547936e-01
-1.01139523e-01 -4.77182478e-01 -3.51443261e-01 -1.39148200e+00
3.00380528e-01 -5.39972186e-01 1.11168781e-02 -1.04144239e+00
-1.58059839e-02 -3.57515484e-01 -2.23953664e-01 2.14591175e-01
-4.34778988e-01 6.94491640e-02 1.43375531e-01 4.99611944e-01
-9.40630138e-01 1.07331014e+00 1.24528742e+00 -1.27802804e-01
-6.34929180e-01 1.54744774e-01 -3.15209031e-01 8.80245328e-01
6.21276140e-01 -1.32413715e-01 -5.84403038e-01 -4.99329031e-01
6.11062124e-02 1.60766840e-01 5.84167719e-01 -1.15411329e+00
5.28766990e-01 -4.67136592e-01 1.33997276e-01 -5.83859444e-01
1.33854970e-01 -7.23983824e-01 1.15811221e-01 3.30152810e-01
-5.09574972e-02 3.39375585e-01 1.08256496e-01 9.65580285e-01
-1.49356291e-01 3.62556189e-01 8.65238011e-01 9.60350558e-02
-7.08774984e-01 1.02602530e+00 -2.18884647e-01 2.53416169e-02
1.07940912e+00 -4.24583629e-02 -1.96803808e-01 -3.64719629e-01
-7.49145806e-01 5.08255899e-01 7.74491668e-01 3.35606843e-01
5.73046386e-01 -1.62693584e+00 -4.48506474e-01 1.12402275e-01
-2.79583603e-01 3.17600310e-01 5.34650028e-01 1.41493785e+00
-6.69447422e-01 -9.61145852e-04 -1.02013431e-01 -7.20860898e-01
-9.24134135e-01 7.03014135e-01 5.02788782e-01 -2.65753090e-01
-8.98785174e-01 5.63178182e-01 2.60644972e-01 4.93838675e-02
3.57791394e-01 -4.65624422e-01 -2.59394258e-01 8.53763595e-02
5.50546825e-01 3.34561974e-01 -3.23674232e-01 -8.29614282e-01
-1.39140740e-01 7.77662098e-01 3.56343091e-02 2.89802045e-01
1.47459245e+00 -4.02142614e-01 -1.64273456e-01 3.54541957e-01
1.10298216e+00 -1.42250687e-01 -1.84218287e+00 -3.71574312e-01
-7.82750174e-02 -4.65927213e-01 -5.77390008e-02 9.90205556e-02
-1.31373394e+00 8.06727588e-01 4.04338121e-01 9.35878530e-02
1.16389430e+00 -3.01490992e-01 1.18162704e+00 1.68474048e-01
1.68139189e-01 -9.86659884e-01 2.34425470e-01 7.03313947e-01
6.52370989e-01 -1.09011078e+00 -2.27067307e-01 -1.59266040e-01
-5.88199198e-01 1.18724775e+00 7.64169157e-01 -4.86610144e-01
6.81594670e-01 -6.27232343e-03 -3.93176079e-02 -4.38829027e-02
-6.58567905e-01 3.96068506e-02 3.27233106e-01 2.85901934e-01
2.26247758e-02 -2.43152380e-01 -2.30810598e-01 5.13949335e-01
2.59070814e-01 9.65716019e-02 3.59575003e-01 5.27067423e-01
-4.83750015e-01 -8.65733683e-01 -6.73930570e-02 1.48919001e-01
-1.39745712e-01 -1.24048740e-02 6.69169202e-02 5.58362305e-01
7.68239275e-02 5.79528093e-01 -2.86343507e-02 -6.27727568e-01
2.10571200e-01 -4.68613446e-01 4.18758988e-02 -2.43395984e-01
-3.06129068e-01 3.62574369e-01 -2.66788185e-01 -9.00451362e-01
-5.05777121e-01 -7.52153575e-01 -1.00429189e+00 -6.81305647e-01
-1.44034803e-01 2.12361291e-02 -1.84877634e-01 9.73048866e-01
4.48159188e-01 2.97004402e-01 8.45405340e-01 -1.35827804e+00
-5.23969889e-01 -6.86652184e-01 -3.58303547e-01 3.19239795e-01
7.59712338e-01 -7.55733311e-01 -2.53036022e-01 3.55830848e-01] | [10.608251571655273, -1.186396598815918] |
ea9c7e5a-4ba8-4ec9-84e0-ab1669cb9af6 | robust-optimization-over-multiple-domains | 1805.07588 | null | http://arxiv.org/abs/1805.07588v2 | http://arxiv.org/pdf/1805.07588v2.pdf | Robust Optimization over Multiple Domains | In this work, we study the problem of learning a single model for multiple
domains. Unlike the conventional machine learning scenario where each domain
can have the corresponding model, multiple domains (i.e., applications/users)
may share the same machine learning model due to maintenance loads in cloud
computing services. For example, a digit-recognition model should be applicable
to hand-written digits, house numbers, car plates, etc. Therefore, an ideal
model for cloud computing has to perform well at each applicable domain. To
address this new challenge from cloud computing, we develop a framework of
robust optimization over multiple domains. In lieu of minimizing the empirical
risk, we aim to learn a model optimized to the adversarial distribution over
multiple domains. Hence, we propose to learn the model and the adversarial
distribution simultaneously with the stochastic algorithm for efficiency.
Theoretically, we analyze the convergence rate for convex and non-convex
models. To our best knowledge, we first study the convergence rate of learning
a robust non-convex model with a practical algorithm. Furthermore, we
demonstrate that the robustness of the framework and the convergence rate can
be further enhanced by appropriate regularizers over the adversarial
distribution. The empirical study on real-world fine-grained visual
categorization and digits recognition tasks verifies the effectiveness and
efficiency of the proposed framework. | ['Shenghuo Zhu', 'Qi Qian', 'Baigui Sun', 'Rong Jin', 'Jiasheng Tang', 'Hao Li'] | 2018-05-19 | null | null | null | null | ['fine-grained-visual-categorization'] | ['computer-vision'] | [ 1.88269038e-02 -3.55481684e-01 3.37277394e-04 -2.80221313e-01
-9.58368361e-01 -7.27215886e-01 1.29782215e-01 -1.13460265e-01
-3.35649312e-01 6.42127514e-01 -4.55458194e-01 -2.19540879e-01
-2.57015586e-01 -6.72107458e-01 -9.53545749e-01 -8.92926633e-01
8.22650194e-02 5.44487119e-01 -2.53712445e-01 2.06118926e-01
3.03274356e-02 5.96775115e-01 -1.07796061e+00 2.21258909e-01
1.09469855e+00 1.25467956e+00 2.68578440e-01 5.08508623e-01
1.89737603e-01 5.32150984e-01 -6.76344752e-01 -5.86167216e-01
6.44150496e-01 1.03335306e-01 -3.20215106e-01 3.62539142e-01
3.78347069e-01 -2.52430171e-01 -3.19432199e-01 1.61998641e+00
4.22898650e-01 7.63623789e-02 8.14086318e-01 -1.71151173e+00
-6.84685826e-01 2.32688755e-01 -7.00513780e-01 3.28905173e-02
-1.04705952e-01 7.63428882e-02 7.29050934e-01 -7.80139506e-01
1.51006579e-01 1.02905297e+00 3.70491147e-01 4.95947301e-01
-9.64763999e-01 -7.73533523e-01 6.15451634e-01 7.03006536e-02
-1.58113480e+00 -9.16970298e-02 5.75931787e-01 -5.18120170e-01
4.13157493e-01 1.85784161e-01 5.39440662e-02 8.88138652e-01
1.60595074e-01 8.90059829e-01 1.09207344e+00 -1.78139806e-01
3.99082631e-01 4.78678048e-01 -1.87490061e-01 3.27009827e-01
3.14940810e-01 3.66160236e-02 7.32373120e-03 -3.41770798e-01
5.53603411e-01 2.95863330e-01 -5.44463508e-02 -3.74072343e-01
-8.13796639e-01 7.56572843e-01 2.16692746e-01 -5.32058328e-02
-4.73691791e-01 5.34666069e-02 4.10952657e-01 3.79273444e-01
5.33605397e-01 1.53082356e-01 -5.64443409e-01 1.51223123e-01
-7.83256292e-01 3.43295813e-01 8.89734566e-01 1.27398539e+00
4.24834222e-01 1.73118338e-01 -3.73502150e-02 7.66663611e-01
1.82492837e-01 6.48896813e-01 3.60535204e-01 -6.79602623e-01
9.66327906e-01 2.28977740e-01 3.42724085e-01 -1.09899223e+00
-1.04091279e-02 -4.89910811e-01 -1.31448901e+00 2.06842959e-01
5.51827669e-01 -4.01435345e-01 -5.20920634e-01 1.65429437e+00
3.19435626e-01 6.21419609e-01 6.37235940e-02 9.98018086e-01
-5.46842627e-02 6.26518190e-01 -1.58954840e-02 -3.46654832e-01
9.24637794e-01 -7.48566628e-01 -4.73466903e-01 -2.22584590e-01
4.65665422e-02 -7.75991797e-01 1.01230109e+00 6.33532822e-01
-9.37527061e-01 -3.95351470e-01 -9.04245198e-01 3.96685481e-01
-1.97913483e-01 1.63792759e-01 3.06335092e-01 6.60024703e-01
-5.76688945e-01 3.73451620e-01 -6.53307319e-01 1.56019509e-01
5.16974509e-01 4.58761513e-01 -1.87188432e-01 -4.13988978e-01
-9.97753143e-01 6.52559042e-01 3.14729452e-01 -5.49253561e-02
-1.03180444e+00 -8.60872805e-01 -4.13106024e-01 1.30450353e-01
4.94510174e-01 -5.03354013e-01 1.09708631e+00 -1.31490231e+00
-1.44691193e+00 6.22105598e-01 -2.59646066e-02 -4.29850250e-01
9.27139461e-01 -2.08574057e-01 -2.99555779e-01 -1.90854874e-02
4.57490757e-02 -2.55151421e-01 1.22894490e+00 -1.33424783e+00
-8.32110405e-01 -4.34829533e-01 2.63816655e-01 3.91387343e-02
-7.11787701e-01 2.25992486e-01 -4.19623703e-01 -8.59067440e-01
-3.12381178e-01 -1.04486001e+00 -3.72835129e-01 -2.89637614e-02
-3.06385368e-01 8.72130245e-02 8.54736686e-01 -8.23508024e-01
9.32914734e-01 -2.40112209e+00 1.18347906e-01 4.21973437e-01
8.18588212e-02 1.52351663e-01 -7.33628422e-02 -4.55764160e-02
-2.15493105e-02 2.47678295e-01 -1.54094219e-01 -3.18931282e-01
2.14131072e-01 1.91659868e-01 -5.25320232e-01 8.86198997e-01
2.87279665e-01 5.65566778e-01 -7.72601485e-01 -2.77180642e-01
1.06009088e-01 2.71103472e-01 -6.41939461e-01 3.18095356e-01
-1.46025673e-01 1.38331935e-01 -7.18234003e-01 6.76468849e-01
1.15780199e+00 -3.66202652e-01 2.24616796e-01 1.30536661e-01
4.26709324e-01 -5.16800284e-01 -1.66211474e+00 1.10735059e+00
-9.84620750e-01 1.26817703e-01 5.71322262e-01 -1.38608551e+00
6.90222859e-01 3.33501935e-01 6.37606621e-01 -3.27537119e-01
-1.70728043e-01 3.68167996e-01 -2.06628621e-01 -2.11845383e-01
5.72147965e-02 -2.70275146e-01 -1.99436009e-01 3.27063173e-01
-1.58299267e-01 -2.22418085e-02 -4.23446149e-01 -6.72276467e-02
7.97945976e-01 -2.83907562e-01 3.43936473e-01 -2.00271875e-01
4.27254200e-01 -3.05987716e-01 7.65461504e-01 7.23146856e-01
-2.69679189e-01 5.76281905e-01 4.06274676e-01 -2.07359135e-01
-1.04500842e+00 -1.13104475e+00 -1.04318269e-01 9.68745232e-01
2.00916529e-01 1.31191701e-01 -6.08153462e-01 -7.90975273e-01
3.13500434e-01 4.13855463e-01 -1.90709621e-01 -6.00455403e-02
-3.66613090e-01 -8.96673918e-01 2.73088574e-01 3.63748252e-01
5.12242973e-01 -4.01334345e-01 -2.08744600e-01 1.73456103e-01
-1.28263712e-01 -1.37660646e+00 -7.92874753e-01 -2.15801578e-02
-5.86003304e-01 -9.65489626e-01 -7.92826414e-01 -7.02765822e-01
7.57719576e-01 1.81422234e-01 1.05092371e+00 -1.08259097e-01
-1.18197307e-01 4.24007297e-01 -1.22069113e-01 -6.75130486e-01
-2.47424439e-01 -1.17310636e-01 5.39437652e-01 5.40558994e-01
1.20268025e-01 -4.21090096e-01 -3.86519462e-01 3.52651536e-01
-1.02729177e+00 -3.36471349e-01 3.32461268e-01 9.62826550e-01
7.96187460e-01 7.30919242e-01 7.78104484e-01 -8.52631211e-01
7.66596794e-01 -9.10868466e-01 -1.05879843e+00 3.99573296e-01
-4.14399058e-01 -2.65299797e-01 1.28540480e+00 -8.36470604e-01
-7.91059911e-01 2.58794844e-01 2.49209985e-01 -9.63007152e-01
-3.96758392e-02 3.47785473e-01 -7.46531487e-01 -1.23351648e-01
4.01193410e-01 1.97349295e-01 -2.28755414e-01 -2.18784496e-01
2.34941512e-01 9.84648705e-01 2.89384335e-01 -1.03004539e+00
1.10325563e+00 4.55790997e-01 2.09886562e-02 -5.32632649e-01
-6.45735860e-01 -4.17400092e-01 -3.37564051e-01 -8.32993388e-02
4.61752295e-01 -1.05506825e+00 -7.58979380e-01 5.78260660e-01
-1.16574359e+00 -5.90862222e-02 -1.48007758e-02 3.91494930e-01
-7.47178793e-01 4.24441606e-01 -2.74597675e-01 -1.09082735e+00
-1.37393743e-01 -1.12452865e+00 8.86885643e-01 1.35840684e-01
4.91888463e-01 -1.01600182e+00 -4.15670216e-01 1.79372475e-01
3.10776055e-01 4.24289942e-01 7.36529708e-01 -7.39648521e-01
-6.64831698e-01 -3.87782156e-01 -2.98058271e-01 8.51449788e-01
8.80088210e-02 -1.73161387e-01 -8.23342204e-01 -5.24349093e-01
3.65523249e-01 -2.24495679e-01 3.89751315e-01 3.72878581e-01
1.74066293e+00 -7.69682884e-01 -1.01473734e-01 7.92335510e-01
1.78899586e+00 3.49643648e-01 3.18401575e-01 3.80419970e-01
6.38305426e-01 4.43472683e-01 8.01789045e-01 8.53543401e-01
1.31688178e-01 6.42782331e-01 5.67790627e-01 1.15224175e-01
5.62335789e-01 2.34836768e-02 3.45387191e-01 4.96649772e-01
1.56326324e-01 -2.84104139e-01 -9.71614957e-01 5.06776512e-01
-1.97903299e+00 -9.32547688e-01 3.39396894e-01 2.43971205e+00
6.90318346e-01 -7.32527748e-02 5.76498695e-02 -1.89256728e-01
9.37976182e-01 -2.31913432e-01 -1.11417484e+00 -2.51066148e-01
-1.13909908e-01 7.90255442e-02 6.66584790e-01 2.93855339e-01
-1.27625918e+00 6.51890874e-01 6.01503611e+00 1.12629330e+00
-1.17482877e+00 2.47501865e-01 8.70900035e-01 -3.02817792e-01
-9.90156531e-02 -4.17781770e-01 -5.90028107e-01 8.91079307e-01
7.04074264e-01 -6.86958373e-01 9.23982561e-01 1.27235532e+00
2.70973742e-01 5.03271282e-01 -1.07967913e+00 1.13640964e+00
-2.04079434e-01 -1.10027659e+00 -1.92699715e-01 2.15253755e-01
9.13802743e-01 3.11740697e-03 4.20813113e-01 2.76458681e-01
5.04164815e-01 -1.13662136e+00 6.49640620e-01 4.81792778e-01
1.03046167e+00 -1.03501499e+00 6.94258988e-01 7.27415562e-01
-1.06254959e+00 -4.08546388e-01 -5.98339558e-01 1.56918988e-01
-1.75627947e-01 6.79584622e-01 -4.79049534e-01 6.82719588e-01
5.75425029e-01 5.00523090e-01 -7.86579773e-02 8.95702362e-01
2.60294199e-01 3.31021965e-01 -4.17790055e-01 1.53983921e-01
8.43276922e-03 -3.55008215e-01 3.54458660e-01 1.09037852e+00
4.53700036e-01 5.96468747e-02 5.91505587e-01 6.96510315e-01
-3.77347767e-01 1.75553814e-01 -6.20217979e-01 -9.90258157e-02
5.98858476e-01 1.03676701e+00 -2.55998880e-01 -1.20895892e-01
-6.08250201e-01 1.04598117e+00 3.64977121e-01 5.51074505e-01
-1.06147957e+00 -2.61999935e-01 1.19813907e+00 -1.11921718e-02
1.58021763e-01 -1.84759423e-01 -5.91832817e-01 -1.41468990e+00
4.29951698e-01 -1.21192694e+00 3.75283390e-01 -2.02615321e-01
-1.91379130e+00 4.75031316e-01 -2.13354737e-01 -1.48167515e+00
-1.11944117e-01 -6.80468917e-01 -5.44833839e-01 1.13812482e+00
-1.57952881e+00 -9.61669564e-01 -6.68541482e-03 1.07181299e+00
4.02785271e-01 -6.42615318e-01 5.47515035e-01 4.09113497e-01
-5.60410917e-01 9.06804204e-01 8.21806550e-01 1.59153268e-01
6.65429533e-01 -1.26320004e+00 3.68192829e-02 1.10606587e+00
-4.82821316e-02 4.33354080e-01 4.58268613e-01 -4.32182848e-01
-1.49990666e+00 -1.61071455e+00 2.51115173e-01 -1.27211854e-01
8.14838409e-01 -2.61755109e-01 -7.69264162e-01 6.15699470e-01
-1.91523701e-01 4.35637027e-01 6.86234653e-01 -1.82189330e-01
-3.91300648e-01 -3.55896473e-01 -1.62352979e+00 4.12278622e-01
6.57169640e-01 -5.58344424e-01 -1.37064651e-01 8.76059592e-01
6.12070203e-01 -5.04623294e-01 -8.44036698e-01 2.05503583e-01
3.53815645e-01 -5.37413895e-01 1.10168326e+00 -1.02454650e+00
5.14811516e-01 -2.25608513e-01 -6.18797302e-01 -1.47450948e+00
-1.27928451e-01 -5.45266330e-01 -3.96515816e-01 1.26949549e+00
-4.52905335e-02 -8.45072210e-01 7.73256123e-01 8.75886500e-01
2.00799227e-01 -6.97908819e-01 -1.31122947e+00 -1.13005280e+00
4.69705075e-01 -5.60933352e-01 7.99505293e-01 9.04019117e-01
-4.09206539e-01 -3.07647079e-01 -6.32044375e-01 9.08035278e-01
9.82799888e-01 2.03722864e-01 6.15156710e-01 -8.79371405e-01
-7.30008006e-01 -2.89933175e-01 -3.63056213e-01 -9.10713792e-01
6.67181313e-01 -7.90744007e-01 -2.20252760e-03 -1.05631363e+00
2.41612032e-01 -8.69329154e-01 -5.26504993e-01 8.26660916e-02
-3.17177474e-01 -2.35275775e-01 6.94266260e-01 2.00744018e-01
-4.92852658e-01 3.15806180e-01 1.04391110e+00 -4.32601511e-01
1.07129648e-01 6.10241354e-01 -9.44327593e-01 6.24423981e-01
8.38733554e-01 -5.16562462e-01 -5.20976186e-01 -5.46199560e-01
-5.05385511e-02 5.65726310e-02 3.24445069e-01 -7.68311918e-01
3.41042072e-01 -7.38243580e-01 2.54314363e-01 1.78635810e-02
2.18204036e-01 -1.32928610e+00 4.69594374e-02 3.10963660e-01
-8.19505453e-02 -1.07161783e-01 5.29265553e-02 8.83864820e-01
-3.01105559e-01 -2.48385459e-01 1.13135540e+00 -7.84460828e-02
-4.19250011e-01 8.39608490e-01 -1.27640486e-01 3.33826274e-01
1.42555404e+00 2.26594761e-01 -1.45536184e-01 -4.45750475e-01
-6.46914661e-01 3.61962855e-01 3.42980117e-01 4.22927022e-01
7.69901693e-01 -1.47597563e+00 -9.32013333e-01 1.59506053e-01
-9.62053761e-02 1.76877528e-01 2.24063963e-01 3.14708412e-01
-2.10033670e-01 8.77752975e-02 -3.79221290e-02 -4.26323026e-01
-1.11519587e+00 9.62211251e-01 4.81927246e-01 -5.01471758e-01
-3.16472501e-02 5.53175926e-01 3.67774934e-01 -3.76268029e-01
2.85662621e-01 -7.28873461e-02 2.32336640e-01 -1.99693084e-01
6.08635843e-01 4.06150341e-01 3.35742570e-02 -3.42932910e-01
-4.56098229e-01 3.24990302e-01 -6.81333691e-02 1.10855259e-01
1.21951020e+00 -8.84376988e-02 -5.96456835e-03 7.90950283e-02
1.36045372e+00 4.70618606e-02 -1.30829918e+00 -4.74415511e-01
-4.43060473e-02 -9.26060319e-01 -1.16892904e-01 -5.50054014e-01
-1.38036633e+00 8.04832339e-01 6.75450623e-01 1.96069464e-01
1.45544410e+00 -3.04407537e-01 5.39477050e-01 2.14096576e-01
6.71287358e-01 -1.25274718e+00 3.95121127e-02 4.12322819e-01
8.66113842e-01 -1.27665973e+00 -2.47430548e-01 -2.75763601e-01
-9.44858551e-01 1.03262722e+00 6.37317717e-01 -3.74370098e-01
9.39783752e-01 5.77238441e-01 -1.34136960e-01 4.74889547e-01
-6.17284596e-01 1.72360763e-01 1.03873633e-01 9.38513160e-01
-7.45079964e-02 5.47509253e-01 1.14178248e-01 9.78166997e-01
1.17290616e-01 8.99556652e-02 4.34529692e-01 5.67512095e-01
3.35611217e-02 -1.04558754e+00 -6.82663977e-01 5.52279174e-01
-8.66060078e-01 -1.29612535e-02 1.04307175e-01 4.04987037e-01
1.07281283e-01 1.13265920e+00 -1.63157955e-01 -3.91458392e-01
4.24069405e-01 -2.15277791e-01 2.16233104e-01 -5.67653656e-01
-3.74171913e-01 -6.18027933e-02 -4.87783849e-01 -3.56154710e-01
-9.83626246e-02 -7.74592578e-01 -8.40213060e-01 -4.16471571e-01
-8.36408213e-02 -1.13547906e-01 5.60557067e-01 8.28460097e-01
3.18801403e-01 3.52325916e-01 1.21607780e+00 -5.72897911e-01
-1.40359950e+00 -4.94307727e-01 -1.02532279e+00 4.02412534e-01
3.32450539e-01 -4.48141843e-01 -4.38033640e-01 2.95797847e-02] | [6.884095668792725, 4.432368755340576] |
d57e94f9-ce26-451f-aec8-22b00f009b5b | posscore-a-simple-yet-effective-evaluation-of | 2109.03039 | null | https://arxiv.org/abs/2109.03039v1 | https://arxiv.org/pdf/2109.03039v1.pdf | POSSCORE: A Simple Yet Effective Evaluation of Conversational Search with Part of Speech Labelling | Conversational search systems, such as Google Assistant and Microsoft Cortana, provide a new search paradigm where users are allowed, via natural language dialogues, to communicate with search systems. Evaluating such systems is very challenging since search results are presented in the format of natural language sentences. Given the unlimited number of possible responses, collecting relevance assessments for all the possible responses is infeasible. In this paper, we propose POSSCORE, a simple yet effective automatic evaluation method for conversational search. The proposed embedding-based metric takes the influence of part of speech (POS) of the terms in the response into account. To the best knowledge, our work is the first to systematically demonstrate the importance of incorporating syntactic information, such as POS labels, for conversational search evaluation. Experimental results demonstrate that our metrics can correlate with human preference, achieving significant improvements over state-of-the-art baseline metrics. | ['Max L. Wilson', 'Jiaxin Mao', 'Ke Zhou', 'Zeyang Liu'] | 2021-09-07 | null | null | null | null | ['conversational-search'] | ['natural-language-processing'] | [ 3.50574367e-02 1.91365574e-02 -3.64056110e-01 -4.23597783e-01
-8.99997294e-01 -8.76314402e-01 1.04685795e+00 2.39440337e-01
-8.14563751e-01 5.94651401e-01 6.14953816e-01 -6.21730983e-01
-5.22292964e-02 -4.82690364e-01 -1.77955199e-02 -1.03379436e-01
1.96072802e-01 6.12750709e-01 2.55951554e-01 -6.18919671e-01
5.14568925e-01 2.51247287e-01 -1.42827356e+00 4.17426556e-01
8.63443136e-01 9.61127877e-01 3.91518354e-01 8.60196054e-01
-3.69536549e-01 5.75378239e-01 -6.96300507e-01 -6.39063537e-01
-9.75550637e-02 -6.48910552e-02 -1.25582170e+00 -4.47124034e-01
2.83357561e-01 -3.89673501e-01 -3.84461820e-01 7.78701544e-01
5.53104281e-01 3.96591276e-01 4.65917826e-01 -1.13532579e+00
-6.77965045e-01 4.95235711e-01 1.65057585e-01 3.69069606e-01
1.14132738e+00 1.23182140e-01 1.72146559e+00 -9.99279678e-01
5.15963256e-01 1.28287077e+00 7.42203221e-02 6.90484285e-01
-1.01556516e+00 -2.62312710e-01 8.47290158e-02 3.80620092e-01
-1.04244840e+00 -2.62131542e-01 6.88306987e-01 -3.27588469e-01
1.39195001e+00 8.88060153e-01 2.15936154e-01 1.25824404e+00
-1.46464854e-01 9.27722633e-01 1.04151845e+00 -7.91480899e-01
6.75334856e-02 4.83247906e-01 5.71725488e-01 3.47395360e-01
-2.13484347e-01 -1.66174933e-01 -5.50317168e-01 -7.77653694e-01
-1.15744822e-01 -3.11074555e-02 -4.42233890e-01 -1.74151495e-01
-1.07326841e+00 8.96511137e-01 4.04860348e-01 3.17252189e-01
-4.08004940e-01 -1.33717135e-01 4.05706257e-01 2.18678117e-01
3.58089119e-01 1.00617456e+00 -4.74832207e-01 -6.59285307e-01
-5.68660378e-01 2.17834607e-01 1.38060224e+00 9.81734276e-01
5.65452814e-01 -9.18461978e-01 -4.76166397e-01 1.22071350e+00
2.61554450e-01 3.69748831e-01 6.73858285e-01 -9.84769404e-01
5.56147397e-01 7.42137194e-01 5.02611637e-01 -1.00235355e+00
2.92529203e-02 -1.37074804e-02 -1.61857665e-01 -3.27845514e-01
1.57677233e-01 -1.81276843e-01 -2.81114072e-01 1.41434598e+00
1.03683650e-01 -5.73721111e-01 9.91116092e-02 9.15776789e-01
9.94185150e-01 4.91976678e-01 2.29448099e-02 -1.08474351e-01
1.61423457e+00 -1.26173878e+00 -6.98791146e-01 -3.88681471e-01
4.80882883e-01 -1.17241311e+00 1.58533823e+00 5.32944128e-03
-9.52222168e-01 -1.31885588e-01 -8.02763700e-01 -1.82508975e-01
-5.46049535e-01 -3.54802981e-02 8.24573040e-01 5.69171906e-01
-1.07647395e+00 1.53437093e-01 -3.41748148e-01 -5.75510085e-01
-3.38576704e-01 2.66090602e-01 -2.10467592e-01 -1.05986714e-01
-1.61688364e+00 1.18885005e+00 -2.61470795e-01 -3.59407142e-02
-2.75743306e-01 -1.16506457e-01 -6.73302591e-01 3.77190441e-01
3.83470237e-01 -5.73824644e-01 1.97331810e+00 -2.09785596e-01
-1.31650448e+00 6.62840962e-01 -6.34114027e-01 -1.16950534e-01
1.97566852e-01 -2.32927382e-01 -3.12941134e-01 1.53296515e-01
1.09660504e-02 5.61735690e-01 2.51030445e-01 -8.51938188e-01
-6.21304929e-01 -5.56941442e-02 6.94230556e-01 6.03154123e-01
-6.21625245e-01 4.82018143e-01 -7.05241084e-01 -1.50478646e-01
-4.02985334e-01 -1.20456219e+00 -2.40481943e-01 -1.43312961e-01
-2.97764748e-01 -9.63614702e-01 5.16106308e-01 -5.83831787e-01
1.54260695e+00 -1.95161307e+00 -7.66175091e-02 2.64209453e-02
1.09037094e-01 2.88454056e-01 -1.60887763e-01 1.00614977e+00
5.61429977e-01 4.29202557e-01 2.07553104e-01 -4.04881895e-01
4.70740199e-01 1.33988306e-01 -2.37964839e-01 -1.79028898e-01
-4.27642316e-02 1.03869867e+00 -1.11284614e+00 -6.40022993e-01
6.38211966e-02 2.35027030e-01 -3.47093761e-01 3.64005119e-01
-1.23417072e-01 -1.46203503e-01 -6.99629009e-01 4.76706088e-01
2.30348945e-01 -3.53018016e-01 1.06126137e-01 1.94642395e-01
8.45412537e-03 1.11252165e+00 -5.14437258e-01 1.38710129e+00
-9.85781252e-01 7.02313960e-01 6.58864602e-02 -4.10160333e-01
4.53733891e-01 3.41791689e-01 1.41290233e-01 -7.05762386e-01
-1.58068836e-01 2.39233002e-01 -5.27842864e-02 -4.34308439e-01
8.98158073e-01 4.82151359e-01 -4.55732167e-01 6.39569640e-01
-2.12802187e-01 -2.62378514e-01 4.74569410e-01 4.36994314e-01
1.43255019e+00 -6.01081371e-01 3.60981613e-01 -1.78073898e-01
7.82664716e-01 5.06734625e-02 -7.49165714e-02 1.07430363e+00
-2.13465646e-01 2.89973944e-01 3.61065656e-01 -1.23592801e-01
-6.12033963e-01 -7.08023667e-01 5.08206561e-02 1.30574465e+00
2.75335073e-01 -6.70950294e-01 -6.02207124e-01 -8.82976532e-01
2.00267006e-02 7.29298532e-01 -3.50952208e-01 -3.43695581e-02
-4.97067750e-01 -2.45692328e-01 3.49591166e-01 2.70434916e-01
1.92185834e-01 -1.27593470e+00 -3.89205515e-01 1.24237873e-01
-6.28115833e-01 -1.22849500e+00 -1.08285570e+00 8.63966048e-02
-5.51463187e-01 -1.00469422e+00 -6.63799345e-01 -9.00327384e-01
2.60589689e-01 5.05698860e-01 1.24466610e+00 3.22003186e-01
-1.58866033e-01 8.53726089e-01 -5.80237985e-01 5.42505644e-03
-1.67635098e-01 2.82182604e-01 -2.99143046e-01 -5.32667041e-01
7.83485055e-01 -3.24352086e-01 -9.28251982e-01 5.65929651e-01
-6.65062845e-01 -4.32516187e-01 5.26413858e-01 9.72815752e-01
-9.61401239e-02 -3.84810030e-01 2.78359383e-01 -6.90627277e-01
1.81524050e+00 -3.24329257e-01 -1.84285954e-01 6.38637543e-01
-9.41963255e-01 1.58438891e-01 4.14545715e-01 -4.23812687e-01
-9.15524602e-01 -3.24351549e-01 -1.30369306e-01 3.15672010e-01
-5.91729879e-02 8.28304470e-01 9.86432359e-02 -1.93960771e-01
7.77568817e-01 1.96441144e-01 -1.59814104e-01 -4.05654907e-01
4.75958377e-01 1.25186944e+00 1.93099871e-01 -3.90289873e-01
2.51167387e-01 -1.36920974e-01 -4.54146326e-01 -8.79879594e-01
-4.04456019e-01 -1.29824781e+00 -1.23326637e-01 -1.69217989e-01
4.91332442e-01 -5.20809233e-01 -9.99986529e-01 -8.84096771e-02
-1.49088240e+00 1.25046238e-01 1.96650207e-01 4.61964995e-01
-2.67311662e-01 7.80432820e-01 -6.22426867e-01 -1.10286880e+00
-6.06754065e-01 -1.30813920e+00 9.98961389e-01 1.50497377e-01
-8.38207543e-01 -8.48784447e-01 1.24089971e-01 8.14514399e-01
6.45846307e-01 -6.55017376e-01 8.44592035e-01 -9.97389615e-01
-4.66236383e-01 -6.09277666e-01 -1.23778358e-01 2.13204011e-01
1.87172443e-01 -2.63390392e-01 -8.55953455e-01 4.38958220e-02
-7.31587783e-02 -3.70760053e-01 6.96278155e-01 -1.46213546e-01
9.75638092e-01 -6.47145271e-01 -4.82370824e-01 -3.70227933e-01
8.01210344e-01 2.71344066e-01 5.12511432e-01 3.57639492e-01
-4.70565818e-02 7.68825948e-01 6.12397552e-01 2.96815068e-01
6.28338277e-01 1.09689963e+00 1.64202109e-01 3.09398651e-01
-2.59485431e-02 -1.96832791e-01 1.92856774e-01 6.88436568e-01
2.81514287e-01 -7.34324038e-01 -8.28948915e-01 6.27291381e-01
-1.77305353e+00 -9.84700561e-01 1.15393505e-01 2.12306857e+00
1.03472042e+00 -3.51111032e-02 -1.38506293e-01 -1.48964569e-01
6.10875010e-01 2.60857910e-01 -1.95346549e-01 -7.72730470e-01
1.51696578e-01 2.37684965e-01 1.61588430e-01 1.02002895e+00
-8.45272958e-01 7.39181459e-01 6.45089436e+00 5.33877611e-01
-8.39431226e-01 -2.37131864e-02 2.72197455e-01 7.89394304e-02
-5.34531116e-01 1.65321156e-01 -7.59745002e-01 5.48023164e-01
7.63258457e-01 -4.88230526e-01 6.81424797e-01 1.04608560e+00
-1.07751135e-02 -1.70112625e-01 -1.26366389e+00 9.69983995e-01
2.08374873e-01 -1.02533019e+00 -1.92326292e-01 3.90134417e-02
2.43584067e-01 4.88904752e-02 -6.77802553e-03 5.55321276e-01
3.75520706e-01 -8.24914932e-01 1.32341594e-01 2.66093731e-01
3.19065392e-01 -2.86711365e-01 8.61394942e-01 5.81035614e-01
-8.14310789e-01 -3.16274911e-02 -4.34184894e-02 -8.60663876e-02
3.48900765e-01 4.25590992e-01 -1.32881272e+00 5.40326834e-02
5.09510279e-01 6.47375360e-02 -5.98461330e-01 9.70118761e-01
-2.98495412e-01 4.26779121e-01 -3.24541479e-01 -1.08570075e+00
3.01177382e-01 4.31360044e-02 6.76199615e-01 1.42417824e+00
3.26593101e-01 2.19842374e-01 1.67263895e-01 5.72072983e-01
-2.55166203e-01 3.67583066e-01 -6.33579850e-01 -4.37819958e-01
7.45434225e-01 1.29129136e+00 -1.40245438e-01 -2.16297388e-01
-5.88179350e-01 1.11871159e+00 3.04524958e-01 2.53411651e-01
-2.46232778e-01 -6.63643718e-01 6.70866430e-01 -3.67473774e-02
6.14740066e-02 -2.85078228e-01 -1.95025709e-02 -1.02720213e+00
7.56774724e-01 -9.75133717e-01 3.16125095e-01 -7.05387652e-01
-1.54248011e+00 9.19441283e-01 1.89724937e-02 -9.79237318e-01
-8.48315001e-01 -5.06168306e-01 -6.21279776e-01 1.20053709e+00
-1.45638549e+00 -6.87549353e-01 -3.56748283e-01 1.30869195e-01
8.27531934e-01 -6.81398883e-02 1.13889587e+00 2.18763769e-01
-4.42550853e-02 6.73679352e-01 -8.68291706e-02 -8.14281553e-02
8.09960246e-01 -1.21888781e+00 4.84597415e-01 3.41391385e-01
1.99482277e-01 1.25142813e+00 7.96798050e-01 -3.49721521e-01
-1.34099638e+00 -1.80966601e-01 1.83976698e+00 -6.75149262e-01
7.59732366e-01 -4.43965793e-01 -5.84989488e-01 2.09873058e-02
7.35074520e-01 -4.05555457e-01 9.93891716e-01 6.95172668e-01
-2.53043413e-01 1.13506101e-01 -8.22847247e-01 9.36959207e-01
8.92973721e-01 -9.43176627e-01 -8.35945606e-01 5.88510692e-01
9.14833307e-01 -2.63055176e-01 -5.51232934e-01 2.47636974e-01
6.86654627e-01 -6.37649477e-01 1.06804228e+00 -5.11504471e-01
2.92400241e-01 1.88268140e-01 -2.37145662e-01 -1.41251767e+00
4.23249416e-02 -8.08997989e-01 -1.38679340e-01 9.40059364e-01
9.85487282e-01 -7.25715935e-01 5.91090262e-01 1.43093276e+00
-1.39450720e-02 -7.91937947e-01 -9.07794952e-01 -5.93606472e-01
-1.62742227e-01 -1.67330101e-01 6.04384542e-01 6.28163457e-01
9.15462971e-01 5.98222077e-01 -1.78158060e-01 -1.80398360e-01
2.32231757e-03 1.09509982e-01 5.77117145e-01 -1.24619055e+00
-2.90927738e-01 -6.38775766e-01 -1.11128107e-01 -1.58167684e+00
3.94824713e-01 -6.41622007e-01 1.45213172e-01 -1.68009663e+00
2.89746106e-01 -3.48519862e-01 -1.84126332e-01 2.17927307e-01
-3.47481161e-01 -3.15093517e-01 1.38462275e-01 2.75190681e-01
-7.77450502e-01 5.67318439e-01 1.01901877e+00 -3.61824453e-01
-1.07000992e-01 2.55880505e-01 -9.76357579e-01 3.32420617e-01
7.21753538e-01 -2.57455587e-01 -4.81040061e-01 -3.90097111e-01
3.32730353e-01 2.46531561e-01 1.32750660e-01 -4.19071048e-01
8.18589211e-01 -2.83997297e-01 -3.34098071e-01 -3.77440035e-01
7.22741067e-01 -7.98186719e-01 -4.58175361e-01 4.18609977e-02
-9.41999555e-01 3.91256064e-01 -3.27791572e-02 5.75707614e-01
-5.94359934e-01 -6.10368371e-01 -1.58835605e-01 -1.24407120e-01
-5.78367770e-01 -1.39845312e-01 -5.31447768e-01 1.49428576e-01
5.57351232e-01 1.19097054e-01 -5.38269818e-01 -9.80336428e-01
-3.40397090e-01 5.59166908e-01 2.04661444e-01 8.06776166e-01
6.38097227e-01 -1.15732634e+00 -3.46522570e-01 -3.44886959e-01
4.77967501e-01 -5.75309396e-01 -1.89588636e-01 4.57864434e-01
-1.09103665e-01 1.08701575e+00 4.00194615e-01 -2.14181617e-01
-1.60886264e+00 3.77402186e-01 -4.66369875e-02 -4.82186437e-01
1.91441663e-02 8.40323150e-01 -4.28281538e-02 -7.04533458e-01
4.87461716e-01 -2.51153231e-01 -4.23944563e-01 9.48360749e-03
5.22130370e-01 1.63796946e-01 3.00707608e-01 -2.29289174e-01
-5.89594841e-01 -3.89268212e-02 -3.47303152e-01 -5.79348505e-01
9.65872824e-01 -2.59109586e-01 -2.73185730e-01 1.18926466e-01
1.29574180e+00 2.51306921e-01 -3.27971995e-01 -3.67255867e-01
2.93420762e-01 -6.83357298e-01 -2.29902510e-02 -1.12021255e+00
-2.83732951e-01 7.31757283e-01 2.73214638e-01 3.96412671e-01
7.42336690e-01 1.80483088e-01 9.11161721e-01 1.17546749e+00
4.85669345e-01 -1.24305749e+00 1.92014560e-01 7.57294297e-01
1.00123870e+00 -1.56476474e+00 -2.48367518e-01 -4.00049388e-01
-8.17946970e-01 9.19538438e-01 5.42203963e-01 4.60827827e-01
4.24014807e-01 -2.14911655e-01 1.63871586e-01 -1.64067686e-01
-1.16107094e+00 -1.97066665e-01 6.87120080e-01 3.05107504e-01
8.13840926e-01 4.11043391e-02 -8.73816490e-01 4.72314090e-01
-3.50119829e-01 -3.45847696e-01 3.11158001e-01 1.02294528e+00
-3.85802537e-01 -1.63256133e+00 7.85111263e-02 5.68911493e-01
-4.24218029e-01 -6.10219479e-01 -1.10865164e+00 4.02918130e-01
-7.29777932e-01 1.63856447e+00 -3.03698868e-01 -3.43480200e-01
5.76626241e-01 3.69097888e-01 1.92614663e-02 -7.25876391e-01
-8.44786763e-01 -4.31987911e-01 6.85242057e-01 -4.03873801e-01
-2.01777533e-01 -5.48989117e-01 -6.26349688e-01 -1.80733755e-01
-6.79590046e-01 7.85371482e-01 7.75703788e-01 6.83250189e-01
5.87145567e-01 -1.43929854e-01 7.10054636e-01 -3.77696544e-01
-9.98348236e-01 -1.22976553e+00 -6.94456100e-02 4.37536538e-01
3.14504176e-01 -5.76113284e-01 -7.41784692e-01 -4.86450195e-01] | [12.132999420166016, 7.804937839508057] |
00fdf6d8-c64f-400e-a4a2-48b55e6de8ff | unlocking-masked-autoencoders-as-loss | 2303.16411 | null | https://arxiv.org/abs/2303.16411v1 | https://arxiv.org/pdf/2303.16411v1.pdf | Unlocking Masked Autoencoders as Loss Function for Image and Video Restoration | Image and video restoration has achieved a remarkable leap with the advent of deep learning. The success of deep learning paradigm lies in three key components: data, model, and loss. Currently, many efforts have been devoted to the first two while seldom study focuses on loss function. With the question ``are the de facto optimization functions e.g., $L_1$, $L_2$, and perceptual losses optimal?'', we explore the potential of loss and raise our belief ``learned loss function empowers the learning capability of neural networks for image and video restoration''. Concretely, we stand on the shoulders of the masked Autoencoders (MAE) and formulate it as a `learned loss function', owing to the fact the pre-trained MAE innately inherits the prior of image reasoning. We investigate the efficacy of our belief from three perspectives: 1) from task-customized MAE to native MAE, 2) from image task to video task, and 3) from transformer structure to convolution neural network structure. Extensive experiments across multiple image and video tasks, including image denoising, image super-resolution, image enhancement, guided image super-resolution, video denoising, and video enhancement, demonstrate the consistent performance improvements introduced by the learned loss function. Besides, the learned loss function is preferable as it can be directly plugged into existing networks during training without involving computations in the inference stage. Code will be publicly available. | ['Chongyi Li', 'Chunle Guo', 'Jie Huang', 'Naishan Zheng', 'Man Zhou'] | 2023-03-29 | null | null | null | null | ['image-super-resolution', 'video-denoising', 'image-enhancement', 'video-enhancement', 'video-restoration'] | ['computer-vision', 'computer-vision', 'computer-vision', 'computer-vision', 'computer-vision'] | [ 3.89578819e-01 -5.10615371e-02 4.22252566e-02 -3.03071320e-01
-5.89599073e-01 7.66732672e-04 2.70916581e-01 -4.53520566e-01
-4.33021843e-01 5.64097404e-01 2.31156006e-01 -7.09463507e-02
-2.70552933e-01 -6.85128748e-01 -9.60372031e-01 -9.22385037e-01
2.24636998e-02 -3.49138260e-01 4.84665185e-02 -3.76577228e-01
2.10884330e-03 2.11322084e-01 -1.46590674e+00 2.82671928e-01
9.52776968e-01 1.37705195e+00 3.80322963e-01 2.63983309e-01
2.39260588e-02 1.18329263e+00 -3.13258886e-01 -7.06675828e-01
3.45070809e-01 -4.10589755e-01 -6.01517558e-01 1.57431394e-01
2.83382148e-01 -6.65118814e-01 -5.58412194e-01 1.31702244e+00
4.98783678e-01 1.61120355e-01 2.32582137e-01 -1.01907218e+00
-9.99006569e-01 5.59636354e-01 -5.03623068e-01 3.25095296e-01
-6.44482151e-02 3.51880968e-01 8.68138611e-01 -9.11253691e-01
4.07946110e-01 1.06751215e+00 8.80525529e-01 5.92419446e-01
-1.16578019e+00 -5.85817695e-01 1.83824062e-01 3.54149729e-01
-1.23819840e+00 -5.23447633e-01 8.45846176e-01 -4.85124499e-01
6.07887566e-01 -2.91295280e-03 3.60960513e-01 1.19450271e+00
6.32600337e-02 7.71726966e-01 1.22608995e+00 -1.73251197e-01
2.35838350e-02 2.20299676e-01 -2.26865798e-01 7.32274115e-01
3.08923405e-02 3.96695346e-01 -5.77651083e-01 4.34652597e-01
1.12998986e+00 -3.48105393e-02 -6.08210027e-01 -6.43144771e-02
-8.46508920e-01 6.31386280e-01 5.26161313e-01 1.54721916e-01
-4.83540356e-01 1.95903763e-01 3.39959919e-01 4.90289599e-01
3.46567482e-01 3.06821913e-01 -4.36433524e-01 1.69110119e-01
-7.14403987e-01 1.20438812e-02 2.59491503e-01 6.21527016e-01
7.36572444e-01 4.55467790e-01 -1.21431455e-01 1.03147709e+00
1.61897019e-01 1.67426527e-01 4.61011857e-01 -1.20807683e+00
2.80523986e-01 2.89361447e-01 2.92006135e-02 -1.02423847e+00
-1.00527398e-01 -6.00962281e-01 -1.22570682e+00 4.14000213e-01
2.12828040e-01 -1.88976511e-01 -6.36803508e-01 1.95405126e+00
2.31195777e-03 2.67293364e-01 1.44314691e-01 1.02148807e+00
8.99854541e-01 5.04720569e-01 2.19136234e-02 -2.52366722e-01
1.19525218e+00 -8.93682241e-01 -7.80722976e-01 -2.45135695e-01
9.92986038e-02 -6.27934039e-01 1.15488708e+00 3.83293957e-01
-1.43718970e+00 -9.02553618e-01 -1.01057124e+00 -2.66267210e-01
-1.56775266e-02 7.32589886e-02 7.44507432e-01 4.12878990e-01
-1.19032288e+00 8.09727490e-01 -6.56420410e-01 -3.44182476e-02
6.53103590e-01 3.48410845e-01 -4.55466509e-01 -1.19606398e-01
-1.21625829e+00 8.39579046e-01 2.62904525e-01 4.26723123e-01
-9.57393587e-01 -7.31872976e-01 -6.85244620e-01 1.48310721e-01
4.04198736e-01 -9.84380603e-01 8.40582252e-01 -1.26923406e+00
-1.54865646e+00 8.23663533e-01 1.06660336e-01 -5.35125017e-01
5.22601783e-01 -4.10799921e-01 -3.46869498e-01 1.45548955e-01
-8.90126750e-02 6.90062582e-01 1.11588383e+00 -1.33601987e+00
-5.77460110e-01 -1.76805139e-01 2.08419502e-01 6.66711554e-02
-3.69986981e-01 4.18203436e-02 -3.18749636e-01 -9.79612350e-01
1.30535085e-02 -4.42196846e-01 -4.65142392e-02 3.68684947e-01
-3.45647894e-02 1.86240440e-03 6.25451565e-01 -1.06673229e+00
1.10903347e+00 -2.38523173e+00 2.19927773e-01 -1.37925491e-01
1.80622384e-01 4.40224707e-01 -2.34349981e-01 1.61657512e-01
-2.61240602e-01 9.81907398e-02 -4.08217162e-01 -1.18077695e-01
-2.36843806e-02 1.12407535e-01 -3.34791034e-01 3.25325817e-01
3.72991294e-01 9.23923731e-01 -5.82342744e-01 -2.61742234e-01
1.38797909e-01 8.33151221e-01 -7.59078205e-01 3.31734538e-01
-9.37001184e-02 5.71288347e-01 -2.17587337e-01 5.01503468e-01
7.43524969e-01 -2.65097350e-01 -7.67888352e-02 -7.87797630e-01
-7.99193159e-02 9.02109891e-02 -1.04722321e+00 1.59589660e+00
-4.49001372e-01 5.65980077e-01 5.77956736e-01 -1.21437907e+00
7.32341111e-01 3.27744335e-01 4.38464463e-01 -1.01780415e+00
1.05581284e-01 1.01766288e-02 -1.36048377e-01 -8.08408856e-01
2.19485581e-01 -4.06026065e-01 4.42099303e-01 1.57492831e-01
1.39201388e-01 3.25718403e-01 -5.00111692e-02 -1.18269160e-01
8.95692945e-01 3.17727506e-01 -7.04333186e-02 -5.19268177e-02
5.48498869e-01 -3.55363756e-01 6.73995554e-01 7.11232483e-01
-3.53422612e-01 6.74805224e-01 4.68795568e-01 -3.56742382e-01
-1.00714135e+00 -1.14820719e+00 -1.98540553e-01 1.19578934e+00
2.13696748e-01 2.81829704e-02 -7.17315495e-01 -3.95658314e-01
-3.12293291e-01 3.67899925e-01 -5.47082722e-01 -2.42433071e-01
-7.52248824e-01 -8.69037271e-01 3.29188526e-01 6.60700321e-01
1.01481998e+00 -1.01935136e+00 -3.66857260e-01 1.14049762e-01
-3.50146592e-01 -1.14373529e+00 -3.87271672e-01 7.34775513e-02
-8.36463749e-01 -8.82665813e-01 -4.90927130e-01 -8.47655356e-01
5.12654245e-01 3.35336000e-01 1.06791830e+00 2.57421106e-01
-1.08746134e-01 3.88445526e-01 -2.42369384e-01 4.25627269e-02
-1.75661117e-01 -3.89729857e-01 -7.75877610e-02 2.03040525e-01
-1.59965336e-01 -9.75571871e-01 -1.02132034e+00 2.39695773e-01
-1.18452227e+00 6.75693229e-02 8.23739707e-01 9.64380085e-01
5.31917930e-01 4.42540586e-01 4.68281060e-01 -7.43953943e-01
5.39222121e-01 -2.74331599e-01 -4.88452137e-01 1.76423416e-01
-5.73239386e-01 3.79102789e-02 5.52184880e-01 -3.52580249e-01
-1.29978800e+00 -2.63189405e-01 -5.89974165e-01 -5.27422488e-01
-9.13834592e-05 5.62334180e-01 -3.20185393e-01 -1.35708228e-01
4.19812024e-01 4.66600746e-01 2.73707896e-01 -5.07229865e-01
3.19095552e-01 2.01397419e-01 8.38951290e-01 -6.83701575e-01
8.45471740e-01 6.70929492e-01 -1.50781751e-01 -5.86788118e-01
-9.13994372e-01 6.91646338e-02 -1.95276946e-01 -1.77882671e-01
9.10465479e-01 -1.18131244e+00 -8.59590888e-01 5.90550005e-01
-1.05356848e+00 -3.41912478e-01 -3.63454551e-01 3.89505148e-01
-4.09736544e-01 4.41296548e-01 -8.56038094e-01 -4.24138635e-01
-4.27950293e-01 -1.20560765e+00 5.98494291e-01 1.88023865e-01
2.87162691e-01 -9.68128681e-01 -4.03780580e-01 5.31726182e-01
7.35988140e-01 2.42026523e-01 1.17845690e+00 -6.62358431e-03
-7.78939486e-01 2.05174088e-02 -6.82992518e-01 1.05349386e+00
4.03430639e-03 -2.97584474e-01 -1.03700089e+00 -4.85043377e-01
4.63629454e-01 -4.17208970e-01 1.03104436e+00 6.12766743e-01
1.50690126e+00 -5.78289628e-01 2.34480366e-01 1.12835217e+00
1.61458623e+00 7.90379718e-02 9.69710171e-01 4.01898801e-01
5.84796309e-01 6.06686532e-01 5.84094189e-02 3.18316191e-01
4.43284720e-01 5.93934119e-01 6.23748004e-01 -1.97954074e-01
-3.70596528e-01 -1.67253360e-01 5.94821334e-01 9.06268179e-01
-3.87750149e-01 7.33680576e-02 -4.38292444e-01 2.48330414e-01
-1.58560741e+00 -1.02125275e+00 3.04234564e-01 2.00581741e+00
8.04313064e-01 4.29471321e-02 -1.93705320e-01 -4.03181417e-03
6.18343532e-01 3.20163071e-01 -5.16382039e-01 -6.01004362e-02
-2.96409100e-01 1.89897269e-01 1.48460299e-01 4.66755778e-01
-1.12976611e+00 7.54041433e-01 5.87642336e+00 9.52331126e-01
-1.29036415e+00 2.57710844e-01 7.78331041e-01 1.55348210e-02
-3.55330974e-01 -8.74321535e-02 -5.37603617e-01 5.63428402e-01
6.04700208e-01 2.27422059e-01 7.99879432e-01 5.39805830e-01
2.38439992e-01 1.90948471e-01 -9.28000271e-01 1.04406166e+00
1.59300528e-02 -1.29195178e+00 2.45479196e-01 -8.88310596e-02
5.65374434e-01 -1.65318266e-01 4.51848805e-01 3.82119745e-01
9.48894620e-02 -1.13413417e+00 7.98956096e-01 4.90151048e-01
6.49111509e-01 -5.51890850e-01 5.57890177e-01 1.70927227e-01
-8.59801650e-01 -2.82755673e-01 -4.01236027e-01 -5.56320511e-02
9.39217359e-02 5.90210795e-01 -1.81086242e-01 3.49624276e-01
8.91971111e-01 7.15447545e-01 -3.92852187e-01 7.64671147e-01
-2.71338284e-01 4.99964356e-01 1.37137651e-01 6.74915373e-01
1.41184062e-01 -4.01429862e-01 4.23373669e-01 9.39158320e-01
2.15095162e-01 1.66438848e-01 -1.32431507e-01 1.06753552e+00
-2.54369348e-01 -2.23354369e-01 -1.99497789e-01 8.44488740e-02
2.31053829e-01 1.10303068e+00 -3.97180170e-01 7.01062903e-02
-5.20148575e-01 1.02190053e+00 3.13804984e-01 7.07294226e-01
-9.76660788e-01 -1.16214775e-01 7.66813993e-01 1.76857397e-01
5.13435006e-01 -6.15855902e-02 -2.92532951e-01 -1.14090621e+00
2.48069629e-01 -9.88050759e-01 9.22711343e-02 -8.44829738e-01
-1.45679224e+00 7.70032525e-01 -3.12293440e-01 -1.03722012e+00
1.38585195e-01 -8.07213128e-01 -4.84628588e-01 6.73202813e-01
-1.85296512e+00 -9.90634382e-01 -3.31552178e-01 8.43342721e-01
3.46260160e-01 -7.25978985e-02 4.24737841e-01 8.32684278e-01
-8.43904197e-01 6.23963535e-01 1.31097183e-01 2.32661188e-01
4.97116596e-01 -8.71441066e-01 -1.56310543e-01 9.81178701e-01
-2.47326002e-01 5.74998140e-01 6.39933586e-01 -1.82117000e-01
-1.40777624e+00 -1.06017864e+00 4.88232940e-01 -1.31648317e-01
5.96637547e-01 -9.05629471e-02 -1.05932117e+00 6.24498487e-01
3.50227624e-01 1.60860449e-01 5.02687156e-01 -1.90871641e-01
-4.42489892e-01 -4.92854863e-01 -1.08396530e+00 4.33252513e-01
1.06685925e+00 -6.17981851e-01 -1.88949019e-01 1.49517104e-01
7.67190576e-01 -3.83724988e-01 -1.02561772e+00 6.65544510e-01
4.60396618e-01 -1.28408289e+00 1.45444798e+00 -4.76882637e-01
9.72018123e-01 -1.75219774e-01 -3.93340975e-01 -9.89511490e-01
-4.30204153e-01 -3.99979502e-01 -8.44253898e-02 1.28304303e+00
1.29640609e-01 -4.78490800e-01 5.58448613e-01 5.94144821e-01
-3.26278538e-01 -9.99873638e-01 -9.05808926e-01 -4.97580558e-01
4.78233658e-02 -3.49544227e-01 4.45706069e-01 9.49380517e-01
-7.02431917e-01 1.25576586e-01 -7.54819691e-01 2.80228108e-01
6.56962335e-01 -5.36675639e-02 4.05611485e-01 -8.98281276e-01
-6.53566360e-01 -5.82699716e-01 -5.42783029e-02 -1.25108075e+00
6.68236986e-02 -7.97681808e-01 -8.31107199e-02 -1.37060809e+00
3.18493724e-01 -3.28230858e-01 -4.49866533e-01 4.42421615e-01
-3.04042429e-01 3.36601675e-01 2.61739135e-01 2.71957964e-01
-5.38271785e-01 6.65330589e-01 1.47429502e+00 -1.15673043e-01
3.55626680e-02 -1.35003164e-01 -9.82604563e-01 7.92817354e-01
5.65550625e-01 -1.56507343e-01 -3.81560355e-01 -9.25429106e-01
3.12533289e-01 1.22183926e-01 8.29766154e-01 -8.38905513e-01
1.66529775e-01 -4.58337702e-02 3.53543788e-01 5.27257547e-02
3.83997291e-01 -7.45330989e-01 1.36180222e-01 1.90562919e-01
-2.58012563e-01 -1.27181098e-01 1.09349534e-01 4.78279769e-01
-5.08381248e-01 -1.55029789e-01 1.13059103e+00 -3.62611294e-01
-9.71672952e-01 3.34377587e-01 -2.91608945e-02 1.25678405e-01
5.03827274e-01 -4.33617949e-01 -3.84330601e-01 -3.24077159e-01
-7.54743338e-01 -1.09782219e-01 2.02836603e-01 1.82165712e-01
7.48285770e-01 -1.11705017e+00 -7.18434989e-01 1.60463691e-01
-4.49104309e-01 -4.69197147e-02 6.21321917e-01 1.00997925e+00
-3.46724719e-01 -8.32197592e-02 -3.57408643e-01 -4.24642831e-01
-7.26938486e-01 5.14131963e-01 5.21769047e-01 -2.61740267e-01
-8.17176521e-01 9.78732586e-01 4.87189025e-01 -7.45924264e-02
4.84296918e-01 -2.95475237e-02 -6.44164160e-02 -1.05962247e-01
4.96581525e-01 4.03045595e-01 -8.41575563e-02 -4.67259914e-01
-7.42665231e-02 5.24838686e-01 -1.25483423e-01 2.87652791e-01
1.53391302e+00 -3.49202633e-01 -3.32360029e-01 -1.22641116e-01
1.19433773e+00 -3.63291502e-01 -1.75479758e+00 -3.42005312e-01
-2.95655370e-01 -4.66368258e-01 2.90678978e-01 -8.71207774e-01
-1.55115259e+00 8.80970359e-01 9.24806118e-01 -4.99777757e-02
1.56524396e+00 -1.80318236e-01 8.69713545e-01 1.30841546e-02
1.37751669e-01 -9.10548091e-01 4.45238858e-01 3.38624120e-01
1.00658453e+00 -1.38737249e+00 -1.30236149e-01 -2.81122804e-01
-5.83570123e-01 7.53624141e-01 6.39716029e-01 -1.15716629e-01
5.88342309e-01 2.07849577e-01 3.87738608e-02 -7.14647919e-02
-5.75705290e-01 -1.41395748e-01 1.56866118e-01 5.76555550e-01
3.92772645e-01 -3.59941930e-01 -1.00528896e-01 7.82845438e-01
-9.09089521e-02 2.44369462e-01 2.58538306e-01 4.92749929e-01
-3.68629962e-01 -8.62741113e-01 -2.22482473e-01 2.74324417e-01
-5.85080564e-01 -2.78540641e-01 2.40253195e-01 6.56360686e-01
6.48898065e-01 7.82297850e-01 -1.75633192e-01 -2.74455160e-01
3.40236396e-01 -3.06393445e-01 5.17614245e-01 -1.25699773e-01
-5.04221261e-01 -3.31985652e-02 -2.99598634e-01 -5.83840013e-01
-5.84191680e-01 -2.99782604e-01 -9.00367320e-01 -4.10811186e-01
-1.28772646e-01 -6.24635071e-02 4.78824168e-01 9.01856303e-01
3.68122399e-01 7.29699194e-01 5.00302136e-01 -7.20845640e-01
-6.77832425e-01 -6.11011386e-01 -5.49021602e-01 5.91090202e-01
4.88995403e-01 -4.76946115e-01 -3.05014879e-01 2.74983585e-01] | [11.308062553405762, -2.1008849143981934] |
83558431-3a9b-4e8e-9c51-3d46225432a9 | quran-qa-2022-overview-of-the-first-shared | null | null | https://aclanthology.org/2022.osact-1.9 | https://aclanthology.org/2022.osact-1.9.pdf | Qur’an QA 2022: Overview of The First Shared Task on Question Answering over the Holy Qur’an | Motivated by the resurgence of the machine reading comprehension (MRC) research, we have organized the first Qur’an Question Answering shared task, “Qur’an QA 2022”. The task in its first year aims to promote state-of-the-art research on Arabic QA in general and MRC in particular on the Holy Qur’an, which constitutes a rich and fertile source of knowledge for Muslim and non-Muslim inquisitors and knowledge-seekers. In this paper, we provide an overview of the shared task that succeeded in attracting 13 teams to participate in the final phase, with a total of 30 submitted runs. Moreover, we outline the main approaches adopted by the participating teams in the context of highlighting some of our perceptions and general trends that characterize the participating systems and their submitted runs. | ['Tamer Elsayed', 'Watheq Mansour', 'Rana Malhas'] | null | null | null | null | osact-lrec-2022-6 | ['machine-reading-comprehension'] | ['natural-language-processing'] | [-1.67029858e-01 6.83516562e-01 2.28637770e-01 -2.95448214e-01
-1.34036851e+00 -8.14132512e-01 6.09117448e-01 4.34491485e-01
-3.78764957e-01 9.83282208e-01 5.22069275e-01 -5.37617028e-01
-1.50334105e-01 -7.98468113e-01 -5.14562130e-01 -1.89113721e-01
2.29046270e-01 9.08752441e-01 1.29931167e-01 -1.26475501e+00
8.62919748e-01 -2.33589839e-02 -1.67796230e+00 7.13623285e-01
1.53319824e+00 6.81148946e-01 8.35613832e-02 6.47415221e-01
-3.84984873e-02 1.45868969e+00 -8.63772273e-01 -8.29571486e-01
-3.80138397e-01 -9.81559634e-01 -1.84456503e+00 -5.71561158e-01
5.33708155e-01 -6.16689064e-02 -3.31026036e-03 8.30894768e-01
4.36293870e-01 1.29830703e-01 5.19922793e-01 -9.47603106e-01
-1.28601909e+00 8.41525853e-01 1.80667832e-01 3.60135168e-01
1.07836246e+00 -1.73793569e-01 1.26793456e+00 -1.08576405e+00
8.14119637e-01 9.95078683e-01 4.51035827e-01 5.36869526e-01
-2.45143890e-01 -2.80760154e-02 -3.83359969e-01 1.13929355e+00
-1.37414443e+00 -5.04326344e-01 4.93683964e-01 -3.39419931e-01
9.66056108e-01 4.83279288e-01 3.98898810e-01 5.58287203e-01
-6.69570118e-02 1.03469670e+00 1.22958660e+00 -9.29914951e-01
2.26881295e-01 6.67866319e-02 2.62391090e-01 4.65474725e-01
-2.08829716e-01 -6.66960061e-01 -8.41816843e-01 -4.39518020e-02
-2.99733970e-02 -9.78239775e-01 -4.80159849e-01 2.38844603e-01
-1.18952823e+00 1.10080087e+00 2.53546655e-01 6.06421888e-01
-3.59284341e-01 -2.95985758e-01 2.35600442e-01 5.95432222e-01
2.10705418e-02 7.83298135e-01 -4.44829255e-01 -7.02854812e-01
-6.78258181e-01 5.77307045e-01 1.17839408e+00 5.76417625e-01
6.29233479e-01 -4.67582792e-01 1.42996153e-02 8.39722931e-01
3.95007640e-01 5.81179857e-01 4.40996319e-01 -1.11983204e+00
5.60921371e-01 7.37814248e-01 3.75370353e-01 -1.01840615e+00
-2.09646910e-01 -1.05138890e-01 -3.09314787e-01 -3.80396128e-01
6.05316639e-01 -6.98128268e-02 -4.94030535e-01 1.17951894e+00
2.10443363e-01 -7.54531145e-01 3.45609248e-01 9.37669635e-01
1.22646856e+00 8.00313950e-01 -2.48377800e-01 -2.30208218e-01
1.67655218e+00 -1.07102644e+00 -9.94959056e-01 2.79985033e-02
8.04928899e-01 -1.00420320e+00 1.15033627e+00 5.11297524e-01
-1.39179885e+00 -2.72558779e-01 -1.01366794e+00 -5.07850111e-01
-5.86903214e-01 3.60956602e-02 1.53631285e-01 8.31645310e-01
-1.08087039e+00 2.17708796e-01 -4.42431301e-01 -3.46986592e-01
1.46715522e-01 -2.13630721e-01 5.23604564e-02 -3.15386683e-01
-1.67253149e+00 1.79377997e+00 5.03380716e-01 3.47595453e-01
-8.64252925e-01 -4.57718521e-01 -7.22369075e-01 -2.74356067e-01
4.97726351e-01 -2.55225807e-01 1.46867955e+00 -9.64125156e-01
-1.73948216e+00 1.08313155e+00 -1.56824142e-01 -3.63540709e-01
3.30738038e-01 -5.61392784e-01 -7.06195533e-01 4.56266224e-01
1.39174968e-01 3.68595928e-01 2.58470953e-01 -9.93878484e-01
-5.46022415e-01 -2.58299381e-01 5.07372200e-01 5.94864011e-01
1.33689284e-01 5.78159630e-01 1.02412350e-01 -3.62753600e-01
-4.60250415e-02 -5.24701774e-01 3.32531691e-01 -6.51097119e-01
7.13264123e-02 -7.50524998e-01 5.67813993e-01 -1.20820808e+00
1.25139308e+00 -1.59723485e+00 4.26381975e-01 -2.71514475e-01
6.52037561e-02 3.06995243e-01 -7.77145773e-02 1.28965211e+00
3.53427261e-01 -8.71425271e-02 -4.70777601e-01 2.82621771e-01
1.85735792e-01 2.67990828e-01 -5.49617469e-01 2.27449864e-01
3.18902075e-01 1.25432265e+00 -1.30415511e+00 -4.24741119e-01
-1.55527040e-01 1.58386722e-01 6.67686090e-02 2.33749032e-01
-5.84622860e-01 4.96075720e-01 -4.52560484e-01 7.32880652e-01
4.69040632e-01 -3.73635530e-01 2.00130820e-01 2.43617266e-01
-3.78592938e-01 9.30685222e-01 -7.44475305e-01 1.80298650e+00
-1.70259729e-01 6.93515360e-01 3.58724594e-02 -8.98481011e-01
8.03201795e-01 5.35014987e-01 2.92137861e-02 -1.08465648e+00
1.48217186e-01 8.38086963e-01 1.82063252e-01 -7.66310751e-01
9.94043827e-01 -4.08016937e-03 -8.30357000e-02 8.55517626e-01
-5.10635152e-02 -4.20187801e-01 5.44441879e-01 3.53521347e-01
6.87445879e-01 2.14177221e-01 5.51197529e-01 -7.31681526e-01
1.07323086e+00 4.66700703e-01 -1.05599016e-01 5.21875203e-01
-3.18582714e-01 4.02541012e-01 4.69114631e-01 -4.35750425e-01
-7.30156004e-01 -8.70797098e-01 -3.06355149e-01 1.36147928e+00
1.76025257e-01 -4.85990703e-01 -8.73560250e-01 -7.84561157e-01
-7.33734131e-01 1.07526577e+00 -8.25153232e-01 2.36943245e-01
-8.89815748e-01 -5.30412316e-01 8.90158772e-01 1.68442279e-02
6.99982166e-01 -1.44495010e+00 -1.19153750e+00 1.83440298e-01
-1.12266660e+00 -7.19463229e-01 2.60213584e-01 -2.86845654e-01
-3.47136497e-01 -1.32295728e+00 -8.65586042e-01 -9.07166064e-01
4.74425629e-02 -1.08989917e-01 1.63395309e+00 4.11443800e-01
2.67206818e-01 2.72945732e-01 -1.33384264e+00 -7.18374491e-01
-7.73533762e-01 3.33003312e-01 -5.50455689e-01 -3.75442088e-01
3.59847665e-01 5.95433265e-02 -4.25218940e-01 2.12183818e-01
-8.40720952e-01 4.70987484e-02 1.98848113e-01 6.58749223e-01
-1.18994534e-01 -7.61740446e-01 1.15368307e+00 -6.73034787e-01
8.67674708e-01 -5.54714501e-01 -2.09310278e-01 9.36688840e-01
-2.87336707e-01 -2.72314966e-01 3.05942386e-01 4.06584591e-01
-1.03308272e+00 -7.61923730e-01 -5.31768024e-01 1.07882905e+00
6.35813996e-02 1.02973664e+00 -5.08923419e-02 -1.53050840e-01
1.10806012e+00 1.62915722e-01 -1.44655947e-02 -2.73328006e-01
6.90104365e-01 1.09943604e+00 3.49065274e-01 -5.41723073e-01
3.18117231e-01 7.32746720e-02 -5.10188699e-01 -9.40514803e-01
-1.19495058e+00 -3.23103189e-01 -5.43771744e-01 -4.31720734e-01
7.79226661e-01 -7.69098938e-01 -5.80184400e-01 5.12417734e-01
-1.24534225e+00 -2.52699018e-01 -2.93123096e-01 7.30465278e-02
-5.72748780e-01 5.49368978e-01 -6.73413932e-01 -8.32230866e-01
-4.91913378e-01 -1.12217081e+00 7.00110912e-01 4.73419905e-01
-2.98801184e-01 -9.69410658e-01 4.96708274e-01 1.16069186e+00
7.02993751e-01 8.00513104e-02 8.57332230e-01 -7.04611540e-01
-2.07961217e-01 2.98386365e-01 -2.06992701e-01 2.78008699e-01
-1.85034066e-01 -3.00904006e-01 -7.88762391e-01 3.79133746e-02
1.20797634e-01 -1.17337108e+00 2.84307420e-01 -2.18712851e-01
4.69392717e-01 -2.63145983e-01 5.61946988e-01 -5.44204235e-01
1.20933938e+00 1.62159279e-02 1.07152212e+00 6.93910897e-01
2.09837466e-01 8.21490407e-01 9.09491956e-01 8.03179815e-02
1.17594588e+00 6.02393568e-01 4.88648921e-01 5.14929652e-01
-7.76249543e-02 3.19484286e-02 3.48025143e-01 1.50292146e+00
-4.79510307e-01 -4.52405103e-02 -1.45598984e+00 1.00261617e+00
-1.82246983e+00 -5.99620998e-01 -5.73315203e-01 1.76421952e+00
1.17534745e+00 -5.30237079e-01 4.50161546e-02 4.10827436e-03
2.19444752e-01 -2.71358099e-02 1.63011238e-01 -7.57512152e-01
-5.41050076e-01 6.46624267e-01 -2.81107098e-01 8.20824742e-01
-5.87773442e-01 1.00587070e+00 6.93538284e+00 8.30207288e-01
-5.54405749e-01 2.11940944e-01 1.76047578e-01 7.62490034e-01
-4.79738921e-01 2.03065008e-01 -2.92894989e-01 8.11188817e-02
1.35554826e+00 -2.97867566e-01 4.34458494e-01 2.57504970e-01
-1.38150424e-01 -5.13962090e-01 -5.94602883e-01 3.54674995e-01
6.74230695e-01 -1.62569141e+00 2.33593389e-01 -3.09717327e-01
7.43906677e-01 3.26910436e-01 -1.81950256e-01 4.23324496e-01
1.21769324e-01 -1.45928228e+00 8.29410732e-01 5.33884883e-01
4.62348759e-01 -6.27949834e-01 1.22123992e+00 5.87135732e-01
-4.66126114e-01 -4.56658043e-02 -1.96048036e-01 -7.86710560e-01
5.04643559e-01 3.77780646e-02 -5.27951121e-01 1.22981763e+00
7.60359704e-01 2.80260444e-01 -6.56005025e-01 7.98627496e-01
-6.22705996e-01 8.69611442e-01 -1.04521058e-01 -4.83586371e-01
4.22769368e-01 -3.52796137e-01 3.31256658e-01 1.13792515e+00
-6.67069480e-02 4.07334328e-01 -1.80098087e-01 4.22024220e-01
3.31004262e-02 6.83372796e-01 9.23603028e-02 -8.04942995e-02
5.70827305e-01 8.76876295e-01 -4.43349928e-01 -3.93060178e-01
-1.62967414e-01 6.95893466e-01 4.45281923e-01 2.94260290e-02
-6.71288431e-01 -6.38295591e-01 -1.72587305e-01 -2.48977572e-01
2.25599796e-01 -2.42816284e-01 -4.06965196e-01 -1.06850100e+00
1.35731474e-01 -1.52374780e+00 5.71177661e-01 -9.10604954e-01
-1.15550089e+00 7.53257513e-01 7.26173306e-03 -6.19003713e-01
-3.31078351e-01 -4.76981699e-01 -9.37449858e-02 7.78859437e-01
-1.74597800e+00 -1.35928702e+00 -2.89526671e-01 3.88966471e-01
3.22778344e-01 1.52443992e-02 1.33700860e+00 1.94131896e-01
2.66697798e-02 2.96532512e-01 -2.57188827e-02 1.95063859e-01
6.43692493e-01 -1.15097046e+00 2.05233604e-01 7.25903034e-01
4.68581617e-01 6.71966732e-01 8.10069501e-01 -3.43582481e-01
-1.38745463e+00 -2.82176435e-01 1.69669020e+00 -1.08242238e+00
8.65390062e-01 9.47858840e-02 -9.96338010e-01 6.27164721e-01
9.39766467e-01 -8.55370641e-01 7.67678499e-01 -5.13329729e-02
-3.39028358e-01 3.64077747e-01 -9.68681037e-01 2.99148023e-01
4.68189597e-01 -6.89180553e-01 -1.52184081e+00 7.58294404e-01
4.02770936e-01 -6.49619937e-01 -9.03083265e-01 3.48754108e-01
3.85873258e-01 -9.36799705e-01 4.90385383e-01 -6.04951739e-01
7.68806159e-01 -4.53327358e-01 -3.75696123e-01 -1.07554531e+00
3.39504629e-01 -6.41364574e-01 -8.73092636e-02 7.90126264e-01
6.52016044e-01 -5.48899412e-01 4.99797642e-01 1.88940182e-01
-4.59124625e-01 -6.47047758e-01 -1.36687636e+00 -1.08611688e-01
8.10542762e-01 -4.55054790e-01 2.80015439e-01 9.95136619e-01
4.60989088e-01 5.95332801e-01 -2.15366483e-01 -2.69827038e-01
1.34699598e-01 1.71711892e-01 5.85376382e-01 -9.04033482e-01
-1.09036602e-01 -2.73703068e-01 -1.21157154e-01 -1.20599496e+00
-2.87246346e-01 -8.50321829e-01 1.77458435e-01 -1.93242753e+00
-7.02920556e-02 -1.21202692e-01 1.51509076e-01 4.23289835e-01
-3.08184087e-01 6.40635133e-01 4.47441280e-01 2.56610543e-01
-1.06220603e+00 4.94674772e-01 1.32587194e+00 1.17453463e-01
1.05497904e-01 -3.94755453e-01 -1.08885956e+00 4.73200738e-01
7.09145606e-01 -4.28433493e-02 -4.42141555e-02 -8.09264958e-01
1.09772432e+00 -8.36145356e-02 1.55339286e-01 -6.16223931e-01
1.74584046e-01 -1.58309296e-01 -2.79453337e-01 -7.56181300e-01
2.30357096e-01 -2.16961250e-01 -2.69224793e-01 4.90002573e-01
-1.92253202e-01 2.92277962e-01 -1.12011749e-02 -2.49997318e-01
-6.73315048e-01 -6.12980723e-01 5.55148780e-01 -3.51104915e-01
-6.91844821e-01 -5.04279017e-01 -6.90886974e-01 7.40262866e-01
9.81597722e-01 -1.40939847e-01 -7.18111813e-01 -6.22768939e-01
-4.62760836e-01 6.34562433e-01 2.94650137e-01 3.79014254e-01
5.62790215e-01 -8.34118247e-01 -1.50010657e+00 -4.52389270e-01
3.17206532e-01 -2.09841520e-01 4.83474463e-01 9.32638645e-01
-1.13970757e+00 6.89910352e-01 -4.34243053e-01 -2.64359623e-01
-7.98394263e-01 -1.81950280e-03 3.22079092e-01 -2.72399813e-01
-2.76395261e-01 8.80766392e-01 -7.85850048e-01 -7.92236507e-01
-3.81460428e-01 2.95101017e-01 -7.67164409e-01 2.61469603e-01
8.21184456e-01 7.52505124e-01 4.95821625e-01 -1.01607156e+00
-4.45339292e-01 2.88834631e-01 5.98700158e-02 -4.47090685e-01
1.16406047e+00 -3.19237649e-01 -9.03928578e-01 8.35522056e-01
5.96185148e-01 2.27010384e-01 -2.02676132e-01 8.48187432e-02
3.90991122e-01 -1.39669955e-01 -4.28505272e-01 -1.58277714e+00
-1.78225026e-01 7.74792612e-01 8.06678012e-02 4.64121193e-01
8.06692243e-01 2.66276896e-01 9.68224287e-01 5.87651670e-01
5.31429827e-01 -1.30336618e+00 -1.06030904e-01 1.28945136e+00
1.33898640e+00 -9.49047267e-01 -1.01156078e-01 -3.14454496e-01
-6.85047805e-01 1.31125903e+00 2.85722107e-01 6.14322647e-02
3.28703076e-01 -6.03344381e-01 8.19307029e-01 -5.70104718e-01
-4.52463388e-01 -1.83944434e-01 4.65690434e-01 8.70776951e-01
7.24493742e-01 1.88111648e-01 -1.14288616e+00 4.33207095e-01
-9.57831800e-01 1.75027810e-02 9.41241503e-01 1.26188982e+00
-7.46756196e-01 -1.16472793e+00 -3.88676405e-01 2.73487680e-02
-5.48328578e-01 -1.81701437e-01 -7.84282982e-01 8.17042291e-01
2.05867514e-02 1.50417030e+00 -3.43010634e-01 4.16859388e-02
4.73936439e-01 1.32890537e-01 7.74776340e-01 -5.98452330e-01
-1.09377849e+00 -7.94834256e-01 6.07268393e-01 -3.13563943e-01
-7.20643997e-01 -7.21258283e-01 -1.17679965e+00 -2.18447372e-01
-3.61079127e-01 8.81012142e-01 5.48608363e-01 1.40825582e+00
2.23755717e-01 2.75536291e-02 7.25143477e-02 -2.10054651e-01
-7.28121996e-01 -1.44037223e+00 -1.58677176e-01 2.31353313e-01
1.07370511e-01 -2.78499275e-01 -7.85860270e-02 2.72867791e-02] | [11.328455924987793, 8.184185028076172] |
9409ae80-5529-45c2-932a-266714261972 | zero-day-backdoor-attack-against-text-to | 2305.10701 | null | https://arxiv.org/abs/2305.10701v1 | https://arxiv.org/pdf/2305.10701v1.pdf | Zero-Day Backdoor Attack against Text-to-Image Diffusion Models via Personalization | Although recent personalization methods have democratized high-resolution image synthesis by enabling swift concept acquisition with minimal examples and lightweight computation, they also present an exploitable avenue for high accessible backdoor attacks. This paper investigates a critical and unexplored aspect of text-to-image (T2I) diffusion models - their potential vulnerability to backdoor attacks via personalization. Our study focuses on a zero-day backdoor vulnerability prevalent in two families of personalization methods, epitomized by Textual Inversion and DreamBooth.Compared to traditional backdoor attacks, our proposed method can facilitate more precise, efficient, and easily accessible attacks with a lower barrier to entry. We provide a comprehensive review of personalization in T2I diffusion models, highlighting the operation and exploitation potential of this backdoor vulnerability. To be specific, by studying the prompt processing of Textual Inversion and DreamBooth, we have devised dedicated backdoor attacks according to the different ways of dealing with unseen tokens and analyzed the influence of triggers and concept images on the attack effect. Our empirical study has shown that the nouveau-token backdoor attack has better attack performance while legacy-token backdoor attack is potentially harder to defend. | ['Felix Juefei-Xu', 'Qing Guo', 'Yihao Huang'] | 2023-05-18 | null | null | null | null | ['backdoor-attack'] | ['adversarial'] | [ 4.04601306e-01 -8.86925310e-02 -6.27914816e-02 3.03770989e-01
-6.15482569e-01 -9.97472465e-01 1.30185974e+00 4.53295298e-02
-6.98440313e-01 3.66509706e-01 1.19080372e-01 -4.21131492e-01
-4.52304870e-01 -8.51670027e-01 -5.42593002e-01 -6.41400874e-01
-4.60562676e-01 2.41316572e-01 3.37483525e-01 -4.52567458e-01
5.99046230e-01 6.28153980e-01 -1.14049363e+00 4.49796647e-01
4.57588851e-01 8.53784025e-01 -5.98765574e-02 7.49200165e-01
6.85694278e-04 2.65571505e-01 -1.15317464e+00 -7.81201422e-01
4.78461057e-01 7.77004957e-02 -5.73079050e-01 -2.29640007e-01
3.66259217e-01 -3.80468726e-01 -3.82741749e-01 1.09230709e+00
5.07456660e-01 -1.63459852e-01 5.59091032e-01 -1.15040100e+00
-7.95234978e-01 8.71326685e-01 -8.08822095e-01 7.09264040e-01
3.23995799e-01 2.37405851e-01 6.49250031e-01 -6.78018808e-01
9.19062734e-01 1.23783267e+00 4.99747515e-01 3.49899292e-01
-1.35222733e+00 -1.03217673e+00 -1.28093749e-01 1.75383940e-01
-1.69388914e+00 -1.17401749e-01 5.04864335e-01 -2.77618527e-01
7.82616794e-01 6.21011496e-01 4.95195359e-01 1.61714065e+00
6.06897831e-01 6.60413504e-01 1.38796616e+00 -1.95291981e-01
2.91160773e-02 6.96404219e-01 1.43253371e-01 3.59915584e-01
3.14218014e-01 4.97717947e-01 -7.08110690e-01 -5.08723855e-01
5.65613329e-01 -3.42153460e-01 -8.48455206e-02 1.22141510e-01
-1.17466855e+00 9.63677704e-01 2.97032464e-02 4.75476742e-01
-1.42511755e-01 9.55742747e-02 5.86631000e-01 2.94841588e-01
3.42950433e-01 5.97696960e-01 -1.59912467e-01 1.07660063e-01
-1.07572091e+00 3.54938358e-01 5.06166041e-01 7.51089334e-01
7.94507563e-01 -5.87490723e-02 -4.07437325e-01 1.17304459e-01
-1.80857599e-01 8.43433380e-01 4.75451291e-01 -3.75878662e-01
5.45593739e-01 -1.23119399e-01 -2.10911185e-01 -1.37033355e+00
-2.22952262e-01 -5.56679010e-01 -6.63678765e-01 2.27119103e-01
1.96810707e-01 -1.58444896e-01 -5.78907371e-01 1.32650816e+00
2.58506507e-01 1.11629277e-01 -1.17789485e-01 5.16648769e-01
1.00118406e-01 6.16469383e-01 5.33104874e-02 -1.23126939e-01
1.65496027e+00 -3.45412701e-01 -7.17201531e-01 3.02340001e-01
3.71835619e-01 -9.97328520e-01 1.11678839e+00 7.11889267e-01
-7.54392803e-01 -1.80352449e-01 -1.17041087e+00 4.02108282e-01
-6.47180796e-01 -3.33208621e-01 7.66280711e-01 1.62554514e+00
-7.34258175e-01 5.35047531e-01 -4.25922334e-01 -3.42898279e-01
6.01714790e-01 2.64437318e-01 -3.62898260e-01 3.45326364e-01
-1.46908712e+00 7.10977733e-01 3.88110071e-01 -3.17582756e-01
-1.24137926e+00 -8.62574100e-01 -1.48307085e-01 -4.61646877e-02
8.09360206e-01 -5.74595153e-01 5.18516064e-01 -5.17211556e-01
-1.72829783e+00 6.66485190e-01 5.46771884e-01 -9.82493460e-01
7.27375031e-01 -4.94364351e-01 -4.53305066e-01 5.43510258e-01
3.92368212e-02 3.22776079e-01 1.80670094e+00 -1.09064388e+00
-4.20231640e-01 -3.68377924e-01 3.75014842e-02 -1.18231535e-01
-1.11035109e+00 2.98105210e-01 -1.96410343e-01 -1.02469528e+00
-2.69987583e-01 -8.55000913e-01 -9.31972265e-02 -3.83362681e-01
-7.44368672e-01 1.72273993e-01 1.28062379e+00 -4.19710815e-01
1.53757131e+00 -2.19954991e+00 3.82527225e-02 5.27913809e-01
4.59553123e-01 4.44258600e-01 -8.40192661e-02 7.89258540e-01
-6.56586438e-02 7.25385487e-01 -2.06697434e-02 -1.49729773e-01
-7.90162608e-02 8.44696388e-02 -9.56913948e-01 5.85860133e-01
-2.60025889e-01 8.32389116e-01 -5.56011498e-01 -2.87967324e-01
1.09649561e-01 4.87974197e-01 -4.88093197e-01 -2.24536717e-01
-1.00543112e-01 1.99848369e-01 -5.13561308e-01 5.75621068e-01
8.97345960e-01 2.31852219e-01 5.19962423e-02 -3.86494070e-01
-2.13502288e-01 -4.39346135e-01 -1.02180183e+00 1.21793532e+00
-2.56826222e-01 6.32965326e-01 4.48510796e-02 -5.61585844e-01
7.53109992e-01 -2.25019711e-03 1.92924723e-01 -6.13725185e-01
2.07626462e-01 -1.24257334e-01 -5.37060618e-01 -3.99094194e-01
9.30516839e-01 -1.82929337e-01 -2.67641336e-01 6.43230021e-01
-2.45820791e-01 -1.72896475e-01 6.80020228e-02 8.04109693e-01
8.84152293e-01 -1.49591431e-01 1.76683694e-01 -4.92227316e-01
6.14938915e-01 -2.39297688e-01 1.43418223e-01 1.14697373e+00
-1.46894902e-01 1.82636529e-01 5.26550353e-01 -1.93045974e-01
-7.19075203e-01 -1.35047495e+00 -1.31869465e-01 8.14325035e-01
2.10056499e-01 -9.78605032e-01 -9.49656546e-01 -7.49626279e-01
-1.72609702e-01 6.89931750e-01 -6.68088675e-01 -4.59740162e-02
-3.97876978e-01 -9.87975657e-01 1.32580423e+00 -9.07129142e-03
6.05682492e-01 -4.17255491e-01 -7.89184809e-01 -9.70569849e-02
-3.59084979e-02 -1.22300708e+00 -4.96563137e-01 -3.82276237e-01
-7.80267179e-01 -8.29319656e-01 -6.84220552e-01 2.46552736e-01
3.63609672e-01 4.69257295e-01 6.51223361e-01 -1.13405280e-01
-5.51690161e-01 6.57232344e-01 -4.99790400e-01 -6.04575872e-01
-4.45145786e-01 5.80841340e-02 2.10150450e-01 3.93442422e-01
1.05393473e-02 -5.86351395e-01 -5.57250738e-01 4.03581142e-01
-1.50533164e+00 -4.89816785e-01 5.36646724e-01 6.65917814e-01
2.73900539e-01 6.76255405e-01 9.48294774e-02 -1.05225921e+00
8.92427504e-01 -3.45240653e-01 -7.24277675e-01 2.17448622e-01
-7.90134788e-01 1.50606513e-01 3.58909667e-01 -6.98225498e-01
-1.28369629e+00 -3.06840420e-01 3.51172755e-03 -5.26984811e-01
1.86724380e-01 3.12978268e-01 -1.33621907e-02 -2.84100622e-01
1.09457541e+00 2.38109484e-01 -1.25355273e-01 -4.04696643e-01
9.17059124e-01 3.69469464e-01 3.37666452e-01 -8.07531893e-01
1.45939577e+00 1.22937417e+00 1.12108052e-01 -1.45627749e+00
-2.51544148e-01 -2.30855308e-02 -2.43995324e-01 -2.68760353e-01
1.01715159e+00 -3.22682649e-01 -7.39289284e-01 5.78803539e-01
-9.63072121e-01 6.99610263e-02 -1.81428418e-01 1.70382589e-01
-1.86252639e-01 1.10871732e+00 -5.51469088e-01 -7.44217932e-01
-4.92489755e-01 -9.79735196e-01 7.00119257e-01 -2.08301336e-01
-2.45606706e-01 -8.31322968e-01 -6.60944209e-02 5.10493636e-01
6.92469418e-01 1.69039100e-01 1.12597179e+00 -6.77177668e-01
-9.56038475e-01 -1.91250801e-01 -7.75780603e-02 2.30322063e-01
-2.28773937e-01 9.10874009e-02 -9.94862437e-01 -4.86319214e-01
4.50533807e-01 3.35419960e-02 8.02199364e-01 -7.37677980e-03
8.63569856e-01 -5.30974686e-01 -4.63595241e-01 7.91815341e-01
1.35495281e+00 -1.13440685e-01 8.46793234e-01 6.91721499e-01
3.27387869e-01 6.33642375e-01 6.37099981e-01 7.65453696e-01
-3.14930856e-01 7.37496436e-01 3.72909874e-01 2.30136290e-01
1.39538705e-01 -2.21722707e-01 5.39772272e-01 4.76951003e-01
-4.04205546e-02 -3.40205401e-01 -6.44344747e-01 1.84887573e-01
-1.29080188e+00 -9.90832448e-01 3.99292633e-02 2.30370283e+00
4.55882162e-01 4.57885623e-01 4.28683273e-02 1.59759924e-01
7.16532588e-01 4.14124787e-01 2.19362173e-02 -3.47654998e-01
-2.84565628e-01 3.93636763e-01 1.06292975e+00 4.56852406e-01
-8.53436887e-01 1.23070860e+00 6.93855810e+00 1.65398061e+00
-1.18746257e+00 5.53354502e-01 2.69140780e-01 -3.00311625e-01
-5.03441393e-01 2.77897418e-01 -9.58761454e-01 2.95598090e-01
9.23984170e-01 -3.63425136e-01 3.05592299e-01 4.62318182e-01
-3.77277471e-02 -2.18782708e-01 -7.46719897e-01 7.94745803e-01
1.94012091e-01 -1.67646718e+00 5.36917746e-01 3.77564996e-01
3.45091403e-01 -2.96171784e-01 7.29096234e-01 -7.99884424e-02
-1.59033060e-01 -6.52348399e-01 8.37619245e-01 3.63619566e-01
9.14074361e-01 -8.09964716e-01 1.39855877e-01 9.06467512e-02
-8.97784352e-01 -4.07052517e-01 -1.49943933e-01 2.68041670e-01
1.86925337e-01 6.13552094e-01 -6.87669039e-01 7.58904815e-01
6.98092222e-01 9.37771648e-02 -6.60939515e-01 1.43805072e-01
-1.67739391e-01 5.88524759e-01 -5.98738790e-01 1.03202514e-01
2.26000279e-01 -1.15885712e-01 1.24061644e+00 1.24662375e+00
2.42518395e-01 -1.81482255e-03 -2.70004392e-01 7.34878957e-01
4.50851947e-01 -4.49269526e-02 -8.62349391e-01 -3.86210620e-01
4.83359188e-01 1.12362909e+00 -1.08526886e+00 -3.05972308e-01
-3.79266664e-02 1.12998128e+00 -2.51441866e-01 2.72515208e-01
-9.10765290e-01 -4.32761073e-01 5.72663605e-01 5.05433679e-01
2.77153283e-01 -5.29989719e-01 -2.89500833e-01 -1.09379303e+00
-4.88147497e-01 -1.08276892e+00 5.60158610e-01 -3.66313040e-01
-1.24180150e+00 7.70039260e-01 5.40562630e-01 -7.78396487e-01
1.33343980e-01 -5.51029921e-01 -4.39687103e-01 5.12287736e-01
-1.14985490e+00 -1.29982007e+00 3.52532536e-01 1.10560632e+00
3.57129723e-01 -5.70792973e-01 5.72874606e-01 2.90432453e-01
-5.93310535e-01 9.87976551e-01 1.46394104e-01 -3.03003430e-01
6.33343220e-01 -5.90401590e-01 5.00388741e-01 1.04212868e+00
3.47999066e-01 1.09212363e+00 9.61758614e-01 -1.01721442e+00
-1.61742663e+00 -5.86263299e-01 3.40911061e-01 -7.00578094e-01
1.03800178e+00 -6.27201498e-01 -5.95222771e-01 4.90174919e-01
2.76747674e-01 -6.83656991e-01 4.63967144e-01 -1.06002614e-01
-7.12553442e-01 -1.00962386e-01 -1.13459837e+00 9.34825420e-01
7.85520554e-01 -8.90584290e-01 -4.08816487e-01 3.61426145e-01
7.46801257e-01 8.22688118e-02 -6.50614440e-01 -2.39532739e-02
5.00733316e-01 -1.16775727e+00 1.32041073e+00 1.87572762e-02
1.33419216e-01 5.28987460e-02 -1.83416709e-01 -6.66220069e-01
-1.60733581e-01 -1.41614926e+00 4.10369821e-02 1.22921574e+00
1.62570968e-01 -9.31492388e-01 8.07528496e-01 2.21355647e-01
4.07290637e-01 -3.44699174e-01 -1.12702394e+00 -1.07188535e+00
1.75325587e-01 -7.05494523e-01 6.14740312e-01 7.43401289e-01
-1.02042913e-01 1.13945708e-01 -7.76312888e-01 3.31567615e-01
9.63009536e-01 -4.75024343e-01 9.33025360e-01 -5.65208673e-01
-4.85625416e-01 -3.55281383e-01 -3.54424745e-01 -6.69648707e-01
-1.85011879e-01 -8.39346826e-01 -8.33682477e-01 -4.77089971e-01
8.44625458e-02 -2.22093627e-01 -1.34136319e-01 -6.36227205e-02
2.87848979e-01 5.43129265e-01 5.26657343e-01 4.27032977e-01
-5.04956022e-03 3.01976025e-01 1.00456893e+00 -2.47291669e-01
-8.34792256e-02 -1.66850775e-01 -6.33394063e-01 5.24705529e-01
5.85173070e-01 -7.45478392e-01 -7.36624002e-01 -1.75554544e-01
5.84779561e-01 -2.06434727e-01 3.35615903e-01 -8.61745834e-01
1.80063188e-01 -7.38476962e-03 -2.64614284e-01 -6.63114727e-01
3.93478721e-01 -6.79463446e-01 1.44351423e-01 4.87749338e-01
-1.06954001e-01 6.24290295e-02 3.48885685e-01 8.70042324e-01
2.98864752e-01 -2.86224693e-01 5.83385229e-01 -2.65930742e-01
-6.79243505e-01 1.66109130e-01 -9.53332722e-01 -1.85347334e-01
1.24891794e+00 -4.88803804e-01 -3.84298444e-01 -3.24726999e-01
-6.81099176e-01 -2.87265062e-01 1.40940323e-01 3.88795823e-01
7.95667946e-01 -7.19308615e-01 -6.33307815e-01 3.80964756e-01
-2.04969868e-01 -9.64562953e-01 5.18369019e-01 6.24391258e-01
-3.74945253e-01 4.05462146e-01 -4.84678805e-01 -5.57042181e-01
-1.44701803e+00 1.11189806e+00 -3.48018371e-02 -5.54463685e-01
-8.49485338e-01 8.49852145e-01 5.02958298e-01 2.39474580e-01
-7.79048353e-02 2.89639324e-01 -4.87956367e-02 3.19286883e-01
6.15233004e-01 7.61123776e-01 -2.07048282e-01 -4.95613754e-01
-2.52532601e-01 5.51253200e-01 -6.42259240e-01 -6.30266726e-01
7.45092571e-01 -2.94911712e-01 -2.59098589e-01 -1.48547381e-01
9.57717240e-01 4.65122640e-01 -7.46443331e-01 -9.51900408e-02
-4.62700911e-02 -9.19622719e-01 1.56059966e-03 -6.42794311e-01
-7.69519210e-01 7.28758514e-01 6.80500150e-01 3.98843825e-01
9.50693905e-01 -3.20684344e-01 1.09472370e+00 3.30494016e-01
5.06741345e-01 -1.12031913e+00 4.61053431e-01 3.47269922e-01
8.54759455e-01 -4.35084581e-01 4.21349615e-01 -6.03575468e-01
-6.23695910e-01 8.25815439e-01 1.57431617e-01 9.61288586e-02
7.52193511e-01 4.53319162e-01 4.25028540e-02 -3.64009410e-01
-4.05755490e-01 3.74557436e-01 -1.46467328e-01 9.35911953e-01
-3.79470438e-01 -3.01508382e-02 -5.38607895e-01 3.73074055e-01
-1.31841749e-01 -3.30233812e-01 7.26287425e-01 6.60878301e-01
-9.58151296e-02 -1.36247814e+00 -9.33861732e-01 1.03982881e-01
-7.44521558e-01 -1.90623313e-01 -2.14124143e-01 8.86436522e-01
9.04651657e-02 8.95438671e-01 -3.39629441e-01 -7.13189781e-01
-1.02733483e-03 -1.48152217e-01 3.70194435e-01 -3.15179259e-01
-1.03953981e+00 5.37302420e-02 7.93346167e-02 -6.39408588e-01
-4.51418050e-02 -3.84063661e-01 -6.55261219e-01 -6.17403090e-01
-4.79378879e-01 1.32173330e-01 8.42103899e-01 8.59649658e-01
3.77713740e-01 1.71548277e-01 5.18954873e-01 -9.56294537e-01
-8.92557681e-01 -4.69402403e-01 -7.08567441e-01 3.68260503e-01
5.85390721e-03 -5.36057711e-01 -2.77645886e-01 -1.74145892e-01] | [5.746319770812988, 7.804510593414307] |
b641f053-ce2d-424d-8509-dffd2bd3e3fe | ukara-10-challenge-track-1-automatic-short | 2002.12540 | null | https://arxiv.org/abs/2002.12540v1 | https://arxiv.org/pdf/2002.12540v1.pdf | UKARA 1.0 Challenge Track 1: Automatic Short-Answer Scoring in Bahasa Indonesia | We describe our third-place solution to the UKARA 1.0 challenge on automated essay scoring. The task consists of a binary classification problem on two datasets | answers from two different questions. We ended up using two different models for the two datasets. For task A, we applied a random forest algorithm on features extracted using unigram with latent semantic analysis (LSA). On the other hand, for task B, we only used logistic regression on TF-IDF features. Our model results in F1 score of 0.812. | ['Yosef Ardhito Winatmoko', 'Ali Akbar Septiandri'] | 2020-02-28 | null | null | null | null | ['automated-essay-scoring'] | ['natural-language-processing'] | [ 6.39100596e-02 3.70332710e-02 -3.98500651e-01 -4.95695651e-01
-1.24421930e+00 -7.71407127e-01 5.72467506e-01 4.69465584e-01
-6.60218060e-01 9.97500122e-01 4.90785897e-01 -3.41750890e-01
-3.26660067e-01 -4.82365102e-01 -4.29755375e-02 -2.55259693e-01
5.38563550e-01 5.01058519e-01 1.50068685e-01 -6.92737997e-02
9.02230620e-01 -2.36552373e-01 -1.55277455e+00 6.94374442e-01
8.13083112e-01 9.48612928e-01 -1.00255087e-01 1.14709616e+00
-2.86131412e-01 1.10570550e+00 -7.56533921e-01 -8.22935343e-01
4.38297540e-03 -3.51599693e-01 -1.24948299e+00 -5.33611655e-01
6.77183092e-01 -5.71268089e-02 -1.90183014e-01 7.99716294e-01
2.00822845e-01 2.38614336e-01 1.26657212e+00 -1.10918844e+00
-6.17362380e-01 4.24904823e-01 -2.90331602e-01 2.57638186e-01
9.94358301e-01 -2.01825008e-01 1.60980165e+00 -9.39606786e-01
5.79664290e-01 8.34791660e-01 9.01589274e-01 3.50644410e-01
-1.41320229e+00 -5.32551765e-01 -4.65989083e-01 3.28328460e-01
-7.09810317e-01 -4.13266569e-01 2.98398674e-01 -7.66493320e-01
8.11434269e-01 4.15404171e-01 2.62243003e-01 1.08247960e+00
1.95958331e-01 7.63814747e-01 1.83148229e+00 -7.45826125e-01
9.99923944e-02 2.88001448e-01 1.18021345e+00 7.06385136e-01
-5.49506284e-02 -3.71853173e-01 -9.32461500e-01 -6.86169326e-01
1.40419364e-01 -3.54055941e-01 -6.53126463e-02 2.20119104e-01
-1.03884149e+00 1.21168363e+00 -1.40161678e-01 3.17953616e-01
-3.48869786e-02 -8.74731392e-02 3.57298493e-01 8.39500606e-01
4.72919941e-01 8.71460080e-01 -7.85481274e-01 -4.99515504e-01
-1.09498560e+00 5.12971997e-01 1.24415386e+00 3.68802309e-01
4.48335946e-01 -6.77724183e-01 -6.38843954e-01 9.62252557e-01
1.18050516e-01 2.12515727e-01 9.87119257e-01 -1.09665334e+00
5.10237455e-01 4.90390331e-01 2.21830159e-01 -9.96317685e-01
-4.96931791e-01 -1.58706546e-01 -4.19605523e-02 -1.71760514e-01
9.49435651e-01 -7.90831819e-02 -7.99486458e-01 1.35108137e+00
-3.34290713e-01 -3.30694407e-01 -1.35937542e-01 7.73233593e-01
1.09545517e+00 3.49001646e-01 6.92750290e-02 -1.31723762e-01
1.51459706e+00 -1.25371754e+00 -9.77001429e-01 2.97161844e-02
1.06565797e+00 -9.98875260e-01 1.43465710e+00 6.60915077e-01
-9.63446379e-01 -5.05210459e-01 -9.68400717e-01 -3.23255658e-01
-5.27865648e-01 2.87765563e-01 5.28029084e-01 8.66202712e-01
-9.26289678e-01 7.10555196e-01 -1.01177081e-01 -1.19265310e-01
-8.97676721e-02 1.68270126e-01 -4.35536176e-01 4.30907048e-02
-1.43360829e+00 1.12724996e+00 -2.25603774e-01 -5.56649625e-01
-6.09093904e-02 -4.99112040e-01 -6.22723460e-01 1.25461936e-01
-5.28693572e-02 -4.11679953e-01 1.57383251e+00 -6.79445446e-01
-1.71231270e+00 1.05296516e+00 -3.18897218e-01 -1.71809286e-01
1.80943772e-01 -3.08472276e-01 -8.24085921e-02 -1.23920336e-01
6.15447938e-01 2.38651782e-02 5.52166522e-01 -4.39326346e-01
-5.38502097e-01 -4.79540288e-01 -1.78511590e-01 2.82375659e-05
-6.28399491e-01 5.40610194e-01 1.22594312e-01 -5.79141140e-01
1.37636885e-01 -7.86323905e-01 1.19272687e-01 -6.36105239e-01
-4.28869538e-02 -7.03981221e-01 6.11805804e-02 -1.20290077e+00
1.64921498e+00 -1.66818857e+00 2.66631901e-01 2.07377732e-01
2.91500390e-01 -1.82734579e-01 -6.09740764e-02 1.01072796e-01
-1.48640843e-02 1.26345366e-01 -7.73721114e-02 -3.37202549e-01
1.67579561e-01 -4.27589267e-01 -3.16009074e-01 4.83402051e-02
-3.38833660e-01 1.01395047e+00 -8.96406651e-01 -5.43106794e-01
-5.80489397e-01 -2.58552551e-01 -2.97578096e-01 2.85409778e-01
-3.12857926e-02 -1.77119240e-01 -9.88107622e-02 4.80366677e-01
1.11307465e-01 -1.77353099e-01 1.83271334e-01 3.87233287e-01
-1.53036639e-01 1.06694114e+00 -6.93992555e-01 1.81102407e+00
-4.51791793e-01 9.16879892e-01 -1.55503094e-01 -5.65891147e-01
9.18649912e-01 2.35668138e-01 1.81540474e-01 -6.91440344e-01
-6.28816560e-02 5.31533718e-01 -1.14940099e-01 -5.70793629e-01
4.80328530e-01 -8.39393511e-02 -6.67215168e-01 9.10920441e-01
3.83722723e-01 -1.75482988e-01 3.36504638e-01 3.56046855e-01
1.52835166e+00 7.80859739e-02 2.05214411e-01 -3.80475432e-01
4.24557179e-01 2.20707268e-01 9.38932970e-02 9.80250061e-01
-2.53359318e-01 7.18894601e-01 8.41200531e-01 -3.48561943e-01
-6.68626606e-01 -8.78767967e-01 -1.90584436e-01 1.40066421e+00
-6.02007747e-01 -9.42871213e-01 -7.22675264e-01 -1.12815499e+00
1.43465489e-01 1.00707364e+00 -7.81206787e-01 1.67623430e-01
-3.21072549e-01 -6.63916111e-01 6.86773956e-01 1.90619960e-01
9.14361179e-02 -8.32542479e-01 -4.00189310e-01 8.28087628e-02
-5.13943076e-01 -7.10295379e-01 -4.89599794e-01 5.06126106e-01
-6.85175836e-01 -9.24846470e-01 -5.98905206e-01 -5.44481754e-01
-1.79329403e-02 2.04707235e-01 1.47941339e+00 1.11100554e-01
4.10274938e-02 2.71339387e-01 -4.64856029e-01 -3.24147463e-01
-7.12905917e-03 6.79602861e-01 -2.80630179e-02 -3.95683110e-01
8.64438355e-01 -2.84548193e-01 -1.29121199e-01 1.86018318e-01
-3.06363702e-01 -2.30793864e-01 3.19340825e-01 9.88303840e-01
-6.21511526e-02 -6.04086041e-01 7.66711116e-01 -1.06056821e+00
1.28336406e+00 -7.44270623e-01 -7.91967511e-02 4.63658482e-01
-7.60387182e-01 7.40939826e-02 4.50622201e-01 -2.69083589e-01
-7.34847248e-01 -9.15661231e-02 9.67697278e-02 2.26038158e-01
-2.86010578e-02 7.16341436e-01 2.69585490e-01 2.23902002e-01
1.03087556e+00 -1.18880369e-01 -8.50235447e-02 -6.62311912e-01
1.53268501e-01 1.22522438e+00 3.63482952e-01 -5.44766426e-01
4.55999494e-01 -2.61833251e-01 -2.73731709e-01 -5.59165776e-01
-1.57232726e+00 -7.42400825e-01 -8.20245266e-01 -7.63829947e-02
8.64842117e-01 -5.57670534e-01 -6.50217712e-01 3.44904512e-01
-1.10160363e+00 -2.51749396e-01 -1.98279880e-02 4.92178619e-01
-5.61505675e-01 3.34561735e-01 -6.62494421e-01 -7.39175260e-01
-3.51725638e-01 -8.29880595e-01 6.82121098e-01 1.84335306e-01
-1.03012764e+00 -1.03481054e+00 6.77611828e-01 1.11650479e+00
4.39983159e-01 -2.79741377e-01 9.38491821e-01 -1.04892039e+00
2.50361145e-01 -3.69536817e-01 -2.63481319e-01 1.09623164e-01
-8.02799016e-02 -2.57717311e-01 -1.14542770e+00 1.38158396e-01
3.23248446e-01 -9.74575996e-01 1.27241194e+00 5.33023179e-02
1.26837122e+00 -3.20895493e-01 1.20047905e-01 9.90341306e-02
8.81933391e-01 -4.34542298e-01 2.08856389e-01 5.67721248e-01
3.28285187e-01 8.34585905e-01 6.93218052e-01 6.46207929e-02
6.07754290e-01 8.36414218e-01 -4.19563770e-01 4.89909530e-01
8.08834955e-02 -3.91743004e-01 6.11837566e-01 9.31956828e-01
8.00907463e-02 1.45576164e-01 -1.23623109e+00 4.74208742e-01
-1.91185367e+00 -7.13040054e-01 -7.27528870e-01 2.14168334e+00
1.09752643e+00 1.42820522e-01 3.24023664e-01 3.23292285e-01
2.91861236e-01 6.37586415e-02 7.22815022e-02 -6.99187160e-01
-9.65383202e-02 6.92393661e-01 3.43452871e-01 7.07957208e-01
-1.03324342e+00 7.11000144e-01 7.00522470e+00 9.51431155e-01
-4.74984407e-01 3.05548251e-01 4.31703091e-01 -2.39782289e-01
-3.37743789e-01 2.21770614e-01 -8.28451276e-01 7.38440514e-01
1.54761541e+00 -2.75546402e-01 1.27277508e-01 6.21434271e-01
-3.82472008e-01 -5.76989949e-01 -6.98391616e-01 5.94448745e-01
3.47268313e-01 -9.65803683e-01 -2.04037502e-01 -2.67016560e-01
7.52888858e-01 -3.82316858e-01 2.23802552e-02 7.35277236e-01
4.77474153e-01 -1.36415780e+00 5.17834961e-01 7.60230958e-01
8.78748834e-01 -3.97510052e-01 8.09720695e-01 7.44159400e-01
-5.41786551e-01 -2.33164057e-01 -2.41414845e-01 -7.13662088e-01
-3.11321497e-01 5.69005311e-01 -6.32761776e-01 2.09016785e-01
6.33832932e-01 4.51819837e-01 -8.18072915e-01 8.33948076e-01
-5.98344684e-01 1.02448010e+00 9.76996571e-02 -3.01838100e-01
-5.94383851e-02 -6.67197928e-02 2.09563538e-01 1.30312037e+00
-6.33832812e-02 8.71797949e-02 -8.07736143e-02 2.92565107e-01
-1.70790583e-01 3.74064326e-01 -2.63351589e-01 1.11373849e-01
2.80505508e-01 1.36850226e+00 -5.25381327e-01 -3.99725318e-01
-4.41266030e-01 9.67349827e-01 8.03356051e-01 -7.81991258e-02
-6.05468452e-01 -7.09610522e-01 2.51353290e-02 1.58660889e-01
-2.81487256e-01 -1.73739806e-01 -1.01236784e+00 -1.49441063e+00
-2.95031995e-01 -9.15084124e-01 6.87644482e-01 -7.75933743e-01
-1.68065619e+00 2.23495126e-01 -3.50390494e-01 -4.30702180e-01
-2.08367765e-01 -6.92970753e-01 -6.16524339e-01 1.26228237e+00
-1.33962381e+00 -7.49236345e-01 -2.45598212e-01 1.55455604e-01
5.67782164e-01 -4.10430431e-01 1.25295305e+00 9.64474902e-02
-3.01539719e-01 7.30957806e-01 2.94342935e-01 1.84291124e-01
1.47881758e+00 -1.66446424e+00 2.26974905e-01 1.49202436e-01
3.19843739e-01 7.96325862e-01 4.65368599e-01 -6.29017591e-01
-6.93318009e-01 -4.49737847e-01 1.91830409e+00 -1.11125839e+00
9.54570591e-01 -2.89627761e-01 -8.71582806e-01 4.23736840e-01
2.16285855e-01 -6.18829787e-01 1.32811880e+00 5.39704740e-01
-6.04782701e-01 6.30171835e-01 -9.70830500e-01 1.19017228e-01
5.51074445e-01 -7.99620211e-01 -1.01434684e+00 3.81279200e-01
3.67733091e-01 -9.40715000e-02 -9.43358600e-01 2.17374340e-01
1.01665258e+00 -7.89085686e-01 5.36649585e-01 -1.26622128e+00
1.08255696e+00 2.12378427e-01 -1.84378266e-01 -1.33464241e+00
-5.73165417e-01 -3.58098149e-01 -3.37802380e-01 8.21323693e-01
7.36060083e-01 -2.98028290e-01 7.89432824e-01 8.42775106e-01
3.69833857e-01 -9.15916860e-01 -8.01161230e-01 -6.03443742e-01
5.42370856e-01 -6.31045848e-02 1.21438935e-01 1.11906934e+00
5.39017081e-01 8.11874390e-01 -2.64046974e-02 -6.66267216e-01
5.13580561e-01 2.59551257e-01 6.45433426e-01 -1.71745086e+00
-2.58270115e-01 -5.03348947e-01 6.74381107e-02 -8.71614337e-01
4.99065429e-01 -1.06572628e+00 -1.93601638e-01 -1.33900189e+00
6.80998683e-01 -1.96183935e-01 -3.99288803e-01 4.66041714e-01
-6.72119021e-01 4.26855117e-01 5.25719859e-02 4.79419529e-01
-6.72639012e-01 1.17825344e-01 4.73207682e-01 -6.04964085e-02
-7.78558850e-02 2.65638441e-01 -7.28399754e-01 6.11106992e-01
1.27536964e+00 -9.22401369e-01 8.26123208e-02 -2.37085983e-01
5.39333701e-01 2.23657206e-01 9.58192125e-02 -6.24711037e-01
3.44564259e-01 -2.43528858e-01 3.35480243e-01 -2.60698974e-01
2.09883183e-01 -1.02611303e-01 -4.72673088e-01 1.54235169e-01
-8.70446980e-01 1.34855554e-01 -1.57468334e-01 2.10812584e-01
-2.34143019e-01 -8.75419199e-01 4.22112197e-01 -2.10104331e-01
-3.05265971e-02 -4.12244380e-01 -6.23294771e-01 9.68966410e-02
5.47514558e-01 -6.00190349e-02 -4.16274399e-01 -4.34920579e-01
-7.57237077e-01 2.40802348e-01 2.94568062e-01 4.82068479e-01
3.59685630e-01 -1.08550692e+00 -9.50162590e-01 -3.21347341e-02
3.10057141e-02 -8.35948169e-01 -1.66294813e-01 9.19891000e-01
-2.44560137e-01 8.93222928e-01 -2.39087865e-01 -8.34701490e-03
-1.65190005e+00 -1.40783250e-01 8.65776390e-02 -9.43242669e-01
2.75534451e-01 8.95707846e-01 -5.19892812e-01 -8.83784235e-01
1.10909650e-02 4.26055074e-01 -8.39082062e-01 6.24498248e-01
3.53809565e-01 8.59909058e-01 2.72246927e-01 -3.62796336e-01
-2.10835356e-02 3.43165010e-01 -3.18498820e-01 -6.69804335e-01
1.39257073e+00 2.15948120e-01 -3.55446845e-01 8.52573752e-01
1.08248389e+00 3.57050776e-01 -3.28614771e-01 -1.81012094e-01
7.35251129e-01 -5.90424538e-01 2.10117593e-01 -1.17911685e+00
-2.09381118e-01 6.94107771e-01 -5.17742559e-02 3.86872262e-01
4.46817696e-01 -3.65958065e-01 6.42919838e-01 5.51098466e-01
1.82208091e-01 -1.37540483e+00 2.34955758e-01 1.08650029e+00
7.51872897e-01 -1.18986011e+00 2.89423078e-01 -1.51462078e-01
-6.49298728e-01 1.35952044e+00 5.83725274e-01 5.25643639e-02
4.80542153e-01 -8.87416005e-02 1.54133558e-01 -2.81610578e-01
-1.18857872e+00 6.36815578e-02 5.76511621e-01 1.30756050e-01
1.08238864e+00 1.06277645e-01 -1.00813890e+00 1.47065413e+00
-7.20455825e-01 -1.25025008e-02 8.50953162e-01 8.05164933e-01
-5.26855052e-01 -1.05488896e+00 -2.87350565e-01 1.04593039e+00
-7.94270158e-01 -9.06682238e-02 -1.00835121e+00 2.50675857e-01
-3.27756464e-01 1.15002644e+00 -2.51458287e-01 -5.35774291e-01
8.39465857e-02 8.14181626e-01 4.43329394e-01 -8.18858802e-01
-9.99900877e-01 -7.31979728e-01 3.91585261e-01 -2.67205149e-01
-2.79915147e-02 -9.72746015e-01 -7.13362157e-01 -1.91213802e-01
-5.99387944e-01 5.07061601e-01 8.75410974e-01 7.94206917e-01
1.28254712e-01 1.62918359e-01 6.30968928e-01 -1.82783693e-01
-1.13510227e+00 -1.41008520e+00 -4.17462021e-01 3.33349824e-01
5.49128167e-02 -6.51682317e-01 -7.80525565e-01 -1.51821554e-01] | [11.182600021362305, 9.40710735321045] |
3b7b8a97-8856-40e2-adbc-ec51418a01c8 | understanding-and-improving-zero-shot-multi | 2210.04234 | null | https://arxiv.org/abs/2210.04234v1 | https://arxiv.org/pdf/2210.04234v1.pdf | Understanding and Improving Zero-shot Multi-hop Reasoning in Generative Question Answering | Generative question answering (QA) models generate answers to questions either solely based on the parameters of the model (the closed-book setting) or additionally retrieving relevant evidence (the open-book setting). Generative QA models can answer some relatively complex questions, but the mechanism through which they do so is still poorly understood. We perform several studies aimed at better understanding the multi-hop reasoning capabilities of generative QA models. First, we decompose multi-hop questions into multiple corresponding single-hop questions, and find marked inconsistency in QA models' answers on these pairs of ostensibly identical question chains. Second, we find that models lack zero-shot multi-hop reasoning ability: when trained only on single-hop questions, models generalize poorly to multi-hop questions. Finally, we demonstrate that it is possible to improve models' zero-shot multi-hop reasoning capacity through two methods that approximate real multi-hop natural language (NL) questions by training on either concatenation of single-hop questions or logical forms (SPARQL). In sum, these results demonstrate that multi-hop reasoning does not emerge naturally in generative QA models, but can be encouraged by advances in training or modeling techniques. | ['Graham Neubig', 'Haibo Ding', 'Jun Araki', 'Zhengbao Jiang'] | 2022-10-09 | null | https://aclanthology.org/2022.coling-1.152 | https://aclanthology.org/2022.coling-1.152.pdf | coling-2022-10 | ['generative-question-answering'] | ['natural-language-processing'] | [ 1.81025546e-03 9.13308263e-01 2.38648146e-01 -3.35632950e-01
-1.63751280e+00 -9.78372335e-01 7.74916828e-01 1.05249591e-01
1.09729789e-01 8.75943244e-01 3.85667294e-01 -7.07130313e-01
-3.77858698e-01 -1.44378996e+00 -8.78648639e-01 4.21641432e-02
5.60733795e-01 1.32323968e+00 5.68900645e-01 -7.08573341e-01
6.19920790e-02 -7.27272853e-02 -1.42956364e+00 6.92437947e-01
1.30794764e+00 3.01512867e-01 -8.17323178e-02 1.11115801e+00
-7.39589930e-01 1.49915648e+00 -6.95215702e-01 -1.02802575e+00
-1.77909493e-01 -9.51661408e-01 -1.40670395e+00 -6.21960163e-01
5.24759591e-01 -4.53629941e-01 -3.62840950e-01 7.67588615e-01
1.34203747e-01 1.60398141e-01 5.72456539e-01 -1.26446581e+00
-1.16412652e+00 6.98479772e-01 4.39504981e-01 3.09948146e-01
1.07193339e+00 2.92195797e-01 1.34957492e+00 -6.99922085e-01
8.58632743e-01 1.55829442e+00 7.48284757e-01 6.58831358e-01
-1.28534710e+00 -7.16386065e-02 -3.71699929e-01 4.98156250e-01
-1.02613091e+00 -3.44547361e-01 3.83757979e-01 -1.61088437e-01
1.26638210e+00 5.06109238e-01 5.65922081e-01 9.17907834e-01
4.51894730e-01 5.59171021e-01 1.11686933e+00 -5.26670635e-01
3.00208092e-01 7.43392110e-02 4.31919783e-01 9.34852779e-01
2.63838857e-01 -1.95663318e-01 -3.89925778e-01 -5.92975020e-01
4.19049442e-01 -4.74277258e-01 -1.00646250e-01 -4.24922369e-02
-9.06444192e-01 1.34548533e+00 1.62119001e-01 2.98289597e-01
-3.19849044e-01 1.89629897e-01 -1.79912075e-01 6.42725170e-01
-6.92268014e-02 9.22392011e-01 -3.84324163e-01 -1.76735759e-01
-7.63230562e-01 8.66200209e-01 1.26088524e+00 1.16792130e+00
9.43252683e-01 -2.37933367e-01 -5.13365626e-01 4.22423810e-01
4.60296124e-01 5.13000131e-01 4.72742975e-01 -1.48407519e+00
6.01536989e-01 6.97453201e-01 2.43488997e-01 -7.12089360e-01
-1.58994839e-01 -1.29765078e-01 -1.27082586e-01 -5.10834157e-01
6.04145348e-01 -1.80857792e-01 -6.71418488e-01 1.79839063e+00
2.84329027e-01 -3.81737500e-01 4.39305395e-01 5.30385435e-01
1.07551694e+00 8.61467838e-01 2.48438969e-01 -6.90398272e-03
1.72067392e+00 -8.99196327e-01 -7.23278463e-01 -4.88736600e-01
8.19024980e-01 -6.02537572e-01 1.37164879e+00 -2.56830230e-02
-1.64659643e+00 -4.67727453e-01 -7.54559577e-01 -5.89428127e-01
-4.25890923e-01 -5.51430404e-01 6.47798538e-01 7.18399584e-01
-1.26766729e+00 1.14659742e-01 -3.46597135e-01 -4.27977800e-01
1.52636077e-02 -9.47614908e-02 9.27571114e-03 -6.79608583e-01
-1.64247680e+00 1.25550807e+00 1.47882447e-01 -2.75651187e-01
-7.66777456e-01 -7.34104037e-01 -9.54946041e-01 3.63495201e-01
4.78728086e-01 -1.42352700e+00 1.79473495e+00 -2.47646540e-01
-1.14990413e+00 4.11587805e-01 -6.49644792e-01 -4.34730649e-01
1.46559134e-01 -1.02311303e-03 -4.18588161e-01 4.80672538e-01
4.12187666e-01 7.54891455e-01 2.85339683e-01 -1.28939319e+00
-2.54424632e-01 -3.40091079e-01 7.80737996e-01 1.85860991e-01
2.37671286e-01 -1.59981698e-01 -2.80463099e-02 -1.89133272e-01
2.66696095e-01 -5.86289287e-01 3.70948315e-02 -3.89319867e-01
-2.35269368e-01 -5.83734334e-01 3.13685149e-01 -7.70483553e-01
1.11471081e+00 -1.33455002e+00 -1.54373676e-01 -3.69631797e-02
1.85841426e-01 -9.57905594e-03 -4.07580078e-01 1.00115836e+00
3.62783581e-01 2.25758567e-01 -2.92658657e-01 5.65077066e-02
5.92687011e-01 6.71154141e-01 -7.56323516e-01 -4.91956502e-01
4.18562382e-01 1.69889462e+00 -1.01717186e+00 -7.89707243e-01
-3.23706388e-01 2.34320350e-02 -8.47207308e-01 3.70482862e-01
-9.96772468e-01 -3.95955555e-02 -3.15047175e-01 6.21918440e-01
3.11077654e-01 -6.39893591e-01 2.71213919e-01 8.22082460e-02
5.92230916e-01 8.23045790e-01 -7.53037989e-01 1.53017282e+00
-4.54290897e-01 5.34452796e-01 -4.93187487e-01 -4.95841384e-01
6.50734603e-01 5.05147457e-01 -2.81221211e-01 -8.46181333e-01
-3.12152982e-01 2.09509969e-01 5.97302355e-02 -9.02911663e-01
7.15772927e-01 -6.88737094e-01 -1.59498245e-01 9.41479683e-01
3.47836494e-01 -7.09095895e-01 6.69399798e-01 7.11026132e-01
1.21798921e+00 1.49226571e-02 4.31816615e-02 7.94131532e-02
2.79946804e-01 5.73796213e-01 8.83784890e-02 1.32021713e+00
2.56287903e-01 5.47543526e-01 6.37156487e-01 -1.59744009e-01
-8.55100214e-01 -1.38262463e+00 5.82596920e-02 1.03566182e+00
-3.43919769e-02 -7.33409822e-01 -8.84969413e-01 -7.09708631e-01
-1.75921127e-01 1.67013741e+00 -4.42169815e-01 -3.26862097e-01
-5.63305438e-01 -4.69399363e-01 9.91188109e-01 6.19665146e-01
3.76643032e-01 -1.14962804e+00 -6.43166065e-01 3.64592791e-01
-8.25548112e-01 -9.41586077e-01 -1.00845834e-02 -1.03070296e-01
-9.75612342e-01 -1.07941651e+00 -4.14806873e-01 -5.49002826e-01
4.63703364e-01 9.24356207e-02 1.81367421e+00 2.88537562e-01
1.53857201e-01 1.05952311e+00 -3.15132260e-01 -2.86055744e-01
-8.88392508e-01 1.05554424e-01 -6.75615728e-01 -7.89608896e-01
6.38462901e-01 -4.07115608e-01 -3.36687028e-01 1.46689087e-01
-1.20071495e+00 -2.19266281e-01 4.92395639e-01 7.92183101e-01
2.57837206e-01 -3.71990323e-01 8.32870901e-01 -1.08025455e+00
1.25009096e+00 -8.80652547e-01 -2.94758648e-01 8.67339969e-01
-6.45858109e-01 4.83804405e-01 4.73444253e-01 -4.70700040e-02
-1.40893054e+00 -7.32904673e-01 -4.44821656e-01 2.16611534e-01
-1.13460526e-01 6.80954933e-01 7.31492415e-02 2.95245171e-01
9.66256917e-01 4.32239681e-01 -1.99178666e-01 -9.02835801e-02
9.02566433e-01 2.80755937e-01 5.64795375e-01 -1.03935969e+00
9.05542970e-01 1.20812602e-01 -2.85198748e-01 -4.88917738e-01
-1.05995440e+00 -2.56505366e-02 -5.38373142e-02 4.57909219e-02
8.95219207e-01 -6.14578545e-01 -5.58675826e-01 -2.85515845e-01
-1.50938344e+00 -3.49724323e-01 -6.55135393e-01 1.10263944e-01
-7.75750339e-01 4.57851380e-01 -8.81092250e-01 -7.48804808e-01
-2.72366703e-01 -8.04913104e-01 9.56338882e-01 3.19558501e-01
-7.59774923e-01 -1.23138940e+00 4.32770967e-01 8.02740633e-01
4.40622509e-01 -4.14901435e-01 1.85461175e+00 -9.45951939e-01
-1.02603722e+00 -1.06222257e-01 -5.05302958e-02 1.81365956e-03
-2.46964619e-01 -2.99593240e-01 -8.49587560e-01 2.57031441e-01
3.21109831e-01 -7.52718091e-01 5.14971316e-01 -2.63573835e-03
5.85792959e-01 -6.74066246e-01 4.91570123e-02 -4.87067811e-02
1.29627967e+00 1.12830378e-01 1.01566494e+00 1.24847963e-01
4.72849682e-02 8.08615685e-01 2.39959881e-01 -1.21678181e-01
1.10961413e+00 2.25375876e-01 1.45154044e-01 5.61268628e-01
-3.01484495e-01 -7.67436028e-01 2.88430639e-02 7.24853635e-01
2.16665685e-01 -3.79960328e-01 -9.66829002e-01 7.07529306e-01
-1.67802775e+00 -1.24203193e+00 -3.42818409e-01 1.90317726e+00
1.02867448e+00 -4.03376259e-02 -1.62645280e-01 -2.48692885e-01
1.90404132e-01 -1.09677121e-01 -3.47264767e-01 -5.55809319e-01
-3.04601818e-01 5.56882322e-01 -4.50217485e-01 8.55737686e-01
-2.87788231e-02 7.87414789e-01 7.43556213e+00 3.80563021e-01
-1.58396795e-01 2.14749247e-01 2.71167427e-01 1.86661243e-01
-1.39952481e+00 5.44968963e-01 -6.49805784e-01 1.80426240e-01
1.25291789e+00 -3.06641221e-01 4.57008928e-01 4.89086539e-01
-5.12259781e-01 -2.53553718e-01 -1.27665555e+00 4.07316923e-01
3.05309981e-01 -1.58111107e+00 5.83041549e-01 -1.46616444e-01
7.10744917e-01 -4.78280753e-01 -2.34268919e-01 8.87867630e-01
8.31273675e-01 -1.09197640e+00 5.82605362e-01 8.58088493e-01
2.88154304e-01 -5.24946153e-01 7.06188142e-01 6.68141484e-01
-7.09789991e-01 -1.91847801e-01 -5.06564856e-01 -2.04538684e-02
4.16619182e-01 1.46369934e-01 -8.90476286e-01 6.73884034e-01
3.50423276e-01 -4.00051385e-01 -8.62942755e-01 6.80258214e-01
-5.61389029e-01 5.82020760e-01 -1.45667776e-01 -2.66776234e-01
1.75425515e-01 -9.14356261e-02 2.62855411e-01 8.00032496e-01
5.81204712e-01 6.28268600e-01 -4.76570576e-01 1.39640355e+00
2.60670353e-02 -3.51137161e-01 -5.89860439e-01 -2.23069727e-01
4.81669724e-01 9.40996885e-01 -1.25115484e-01 -7.22435772e-01
-4.45442140e-01 7.06590891e-01 3.33294570e-01 5.25210798e-01
-7.25648522e-01 -3.77412260e-01 1.77426979e-01 1.62355751e-01
6.08592965e-02 -1.10379130e-01 -2.52502859e-01 -1.32519770e+00
-3.20122279e-02 -1.26940846e+00 6.89575732e-01 -1.58253229e+00
-1.47960460e+00 5.32042027e-01 2.27963448e-01 -3.49699527e-01
-1.20556760e+00 -3.03767622e-01 -6.34047389e-01 1.01830256e+00
-1.33272028e+00 -1.03183997e+00 -2.23394066e-01 6.24286652e-01
5.02020419e-01 2.55379736e-01 1.14732397e+00 -2.53896981e-01
1.03310458e-01 3.78935665e-01 -2.76443809e-01 -9.66043249e-02
4.50874537e-01 -1.39610076e+00 5.29801846e-01 5.67017138e-01
4.91086006e-01 1.12965727e+00 7.82656908e-01 -6.49357855e-01
-1.55478203e+00 -6.48904085e-01 1.59071839e+00 -1.11952293e+00
5.57415843e-01 -1.29328847e-01 -1.32904458e+00 1.06452835e+00
5.92250288e-01 -5.09103179e-01 8.55019391e-01 9.93155241e-02
-5.05817533e-01 2.43611276e-01 -1.24056649e+00 8.60248744e-01
7.94644296e-01 -8.71283352e-01 -1.60538626e+00 6.16267741e-01
1.04622006e+00 -2.32282370e-01 -7.41871834e-01 1.01061992e-01
3.02631468e-01 -1.14361572e+00 8.48838687e-01 -1.17571199e+00
7.90340722e-01 -2.14301467e-01 -3.38895857e-01 -1.21963131e+00
-1.92013443e-01 -5.78155339e-01 -5.32954693e-01 9.63759422e-01
6.72231317e-01 -8.55853975e-01 6.32828951e-01 1.07647741e+00
-2.33827516e-01 -6.13503039e-01 -8.66326630e-01 -5.61589062e-01
5.54883540e-01 -4.44680631e-01 9.09744680e-01 6.33490264e-01
1.74748600e-01 7.78663337e-01 3.52330118e-01 1.70795843e-02
4.45670694e-01 3.90892535e-01 7.37136543e-01 -1.14992964e+00
-6.52047157e-01 -1.52538922e-02 2.78115690e-01 -1.16648757e+00
1.10122330e-01 -7.95457602e-01 1.12066045e-01 -2.25074387e+00
8.05649087e-02 -2.06002817e-01 4.60549653e-01 2.50628591e-01
-3.62795204e-01 -4.99734096e-02 1.77365929e-01 7.18483329e-02
-6.40142322e-01 5.00746250e-01 1.31659198e+00 3.26927938e-03
2.63457090e-01 -2.41698295e-01 -1.09312415e+00 5.41635156e-01
5.55267870e-01 -3.98865014e-01 -8.91195357e-01 -7.34839141e-01
9.15780663e-01 6.27836585e-01 7.43321538e-01 -7.84813404e-01
6.15621924e-01 -1.23334289e-01 1.98015407e-01 -4.91330415e-01
4.67859447e-01 -2.84832627e-01 6.54713586e-02 2.94795275e-01
-5.37112653e-01 5.34031689e-01 1.59664765e-01 4.56836760e-01
-2.11791575e-01 -8.43152821e-01 2.51590163e-01 -6.71484053e-01
-3.52474898e-01 -3.86718273e-01 -5.46118379e-01 7.84970701e-01
5.49284875e-01 -6.84315339e-02 -9.71416056e-01 -9.24392283e-01
-6.41291022e-01 2.49045148e-01 1.80864662e-01 6.90983906e-02
6.76793575e-01 -1.17113483e+00 -6.16218448e-01 -4.01785597e-02
1.45565137e-01 2.41970159e-02 4.95145589e-01 5.04846573e-01
-4.80175614e-01 9.50258076e-01 1.60774127e-01 -2.69236147e-01
-5.45139790e-01 4.31180418e-01 4.30359393e-01 -4.74386871e-01
-1.71859905e-01 8.17066252e-01 -9.85731557e-02 -9.71141994e-01
-3.12833369e-01 -3.38297248e-01 7.47525170e-02 1.29682058e-02
3.92914712e-01 3.48164976e-01 1.50842607e-01 -1.24412872e-01
7.43766455e-03 3.53136331e-01 4.59566638e-02 -5.77955842e-01
8.36359859e-01 -9.09419805e-02 -2.91474879e-01 3.23123515e-01
7.12320924e-01 -8.06714296e-02 -6.02607548e-01 -2.24545687e-01
-8.99904966e-02 -1.95019722e-01 -6.56703413e-01 -1.07461178e+00
-9.64390635e-02 9.40936506e-01 -1.94609478e-01 6.40806317e-01
6.39547765e-01 5.58427215e-01 1.13820922e+00 9.43264544e-01
3.66153300e-01 -6.73312247e-01 3.89375627e-01 6.83277309e-01
9.26037550e-01 -9.29743946e-01 -4.56892908e-01 -1.52669638e-01
-5.27363241e-01 9.65734363e-01 7.37779558e-01 1.79880381e-01
2.13415418e-02 -1.18298702e-01 1.10830799e-01 -5.85197389e-01
-1.26432097e+00 -1.97639763e-01 1.64524928e-01 7.27224231e-01
3.43048096e-01 -2.68463522e-01 -6.29462153e-02 7.35531330e-01
-7.81284869e-01 7.15696206e-03 7.04124033e-01 9.24333096e-01
-6.37003779e-01 -1.00128460e+00 -5.36864519e-01 5.46357989e-01
-6.89044744e-02 -4.69873846e-01 -5.68268836e-01 8.82373393e-01
-2.33998448e-01 1.44976211e+00 -4.47062552e-02 3.58083546e-02
2.95958310e-01 9.21300769e-01 8.57717514e-01 -8.14946890e-01
-5.77638507e-01 -7.19508469e-01 3.77727181e-01 -2.83283532e-01
8.84726644e-02 -4.69408184e-01 -1.15376198e+00 -4.48117614e-01
-7.81630501e-02 5.29069424e-01 2.25906342e-01 1.32994342e+00
4.53733534e-01 4.14751381e-01 -4.65245470e-02 2.20067844e-01
-1.14330208e+00 -8.94205451e-01 -1.20218888e-01 2.04664260e-01
3.49935703e-02 -1.11119322e-01 -3.60753924e-01 -4.06964719e-02] | [11.05844497680664, 7.96763277053833] |
a10a0f76-301c-47d1-8bf1-b6d58045316f | a-report-on-the-first-native-language | null | null | https://aclanthology.org/W13-1706 | https://aclanthology.org/W13-1706.pdf | A Report on the First Native Language Identification Shared Task | null | ['Joel Tetreault', 'Daniel Blanchard', 'Aoife Cahill'] | 2013-06-01 | null | null | null | ws-2013-6 | ['native-language-identification'] | ['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] | [-7.330080032348633, 3.788383722305298] |
871c50f4-dc0e-4256-bce0-a789e4f41df8 | uif-an-objective-quality-assessment-for | 2205.09392 | null | https://arxiv.org/abs/2205.09392v1 | https://arxiv.org/pdf/2205.09392v1.pdf | UIF: An Objective Quality Assessment for Underwater Image Enhancement | Due to complex and volatile lighting environment, underwater imaging can be readily impaired by light scattering, warping, and noises. To improve the visual quality, Underwater Image Enhancement (UIE) techniques have been widely studied. Recent efforts have also been contributed to evaluate and compare the UIE performances with subjective and objective methods. However, the subjective evaluation is time-consuming and uneconomic for all images, while existing objective methods have limited capabilities for the newly-developed UIE approaches based on deep learning. To fill this gap, we propose an Underwater Image Fidelity (UIF) metric for objective evaluation of enhanced underwater images. By exploiting the statistical features of these images, we present to extract naturalness-related, sharpness-related, and structure-related features. Among them, the naturalness-related and sharpness-related features evaluate visual improvement of enhanced images; the structure-related feature indicates structural similarity between images before and after UIE. Then, we employ support vector regression to fuse the above three features into a final UIF metric. In addition, we have also established a large-scale UIE database with subjective scores, namely Underwater Image Enhancement Database (UIED), which is utilized as a benchmark to compare all objective metrics. Experimental results confirm that the proposed UIF outperforms a variety of underwater and general-purpose image quality metrics. | ['Tiesong Zhao', 'Rongfu Lin', 'Weiling Chen', 'Yannan Zheng'] | 2022-05-19 | null | null | null | null | ['uie'] | ['computer-vision'] | [ 2.78835207e-01 -6.81140304e-01 8.47029686e-01 -3.20032597e-01
-5.87064803e-01 -8.36433470e-02 1.57925278e-01 1.71081275e-01
-9.43957448e-01 7.04066992e-01 3.86970758e-01 2.95828640e-01
-3.53786975e-01 -8.52248132e-01 -4.51893598e-01 -1.01078224e+00
-3.44383836e-01 -7.26957083e-01 2.83312172e-01 -5.51700234e-01
5.85091949e-01 2.45430350e-01 -1.74036288e+00 -2.59800851e-01
1.14080954e+00 1.19426680e+00 5.23687243e-01 5.37249804e-01
9.20276344e-02 4.15441364e-01 -5.36179066e-01 -1.30967200e-01
2.30507925e-01 -4.41492647e-01 -1.96635678e-01 1.12198815e-01
1.67313039e-01 -7.06034839e-01 -4.60435748e-01 1.48611057e+00
8.86739969e-01 5.06940544e-01 5.19856870e-01 -8.59062791e-01
-7.60093510e-01 5.05629927e-02 -4.76448268e-01 3.76688480e-01
1.70507103e-01 4.38985638e-02 7.34811723e-01 -1.19400859e+00
2.52034247e-01 9.94095922e-01 6.73718154e-01 2.89731562e-01
-5.77115178e-01 -5.76060951e-01 -2.56235927e-01 4.20935243e-01
-1.08788610e+00 -4.14855838e-01 6.47481024e-01 -2.50761271e-01
4.46942121e-01 1.72989815e-01 5.61034858e-01 1.74865380e-01
4.05656487e-01 5.86689830e-01 1.43580985e+00 -1.34197533e-01
1.70785606e-01 -7.48438537e-02 -1.45719305e-01 8.17788064e-01
2.40830481e-01 3.50897372e-01 -2.52250433e-01 2.95181841e-01
6.10051036e-01 2.85816342e-01 -8.26431870e-01 2.19220504e-01
-7.07874894e-01 5.68331182e-01 6.29200637e-01 1.07430726e-01
-2.00871199e-01 -3.28511119e-01 4.50249702e-01 4.01813000e-01
1.90995663e-01 3.10078114e-01 -2.63882190e-01 -2.01858521e-01
-5.93419731e-01 -1.64099112e-01 3.93732846e-01 4.19168591e-01
1.11787987e+00 2.25289449e-01 5.28842434e-02 1.08646786e+00
5.85824668e-01 6.81518435e-01 6.30578995e-01 -9.39984024e-01
2.41699219e-01 3.37454170e-01 3.40044834e-02 -1.21774530e+00
-3.98287386e-01 -6.53296933e-02 -1.07767403e+00 6.32982075e-01
-2.60515243e-01 -9.62527618e-02 -8.25282931e-01 1.14671624e+00
3.76380049e-02 2.45785024e-02 5.69070935e-01 1.15417409e+00
1.34721577e+00 1.03150582e+00 -1.39751151e-01 -5.02436697e-01
9.88832593e-01 -7.51516223e-01 -9.47800636e-01 1.28047774e-02
8.32650885e-02 -8.86052132e-01 9.23094630e-01 2.66900778e-01
-9.73773301e-01 -6.71053231e-01 -1.44017470e+00 8.07344094e-02
-1.33934885e-01 1.29088378e-02 3.02523822e-01 5.31167150e-01
-8.77501547e-01 6.93653822e-01 -5.32137632e-01 6.99878782e-02
1.50162742e-01 1.33608922e-01 -6.25250578e-01 -3.50235164e-01
-1.30767250e+00 9.18779552e-01 3.25450152e-01 4.54568207e-01
-9.62233782e-01 -3.74286205e-01 -1.10155892e+00 -1.01271853e-01
7.35578537e-02 9.48307216e-02 8.28235447e-01 -5.57952642e-01
-1.55182731e+00 2.82700032e-01 4.87652048e-02 1.95797548e-01
2.53075749e-01 -1.51090845e-01 -6.94389522e-01 3.93717110e-01
-6.11788630e-02 1.53535903e-01 4.16230321e-01 -1.67076516e+00
-6.69451535e-01 -2.59727508e-01 1.19084651e-02 4.70960796e-01
-6.22925878e-01 1.67267546e-01 -5.23341715e-01 -5.07594168e-01
3.75992417e-01 -4.00669247e-01 -1.17614873e-01 4.32881981e-01
-9.80828255e-02 2.03915805e-01 9.62463021e-01 -9.04823303e-01
1.14101136e+00 -2.35371852e+00 3.68712246e-02 1.03818551e-01
3.79485339e-02 5.89286804e-01 -3.48830760e-01 2.63751537e-01
3.33542436e-01 6.38084412e-02 -6.04265809e-01 -2.28103667e-01
-2.91750312e-01 4.05525357e-01 2.65292883e-01 5.78837395e-01
-4.81681107e-03 3.51327866e-01 -1.04201782e+00 -6.82688892e-01
4.85518128e-01 5.51969886e-01 -3.04475337e-01 5.54270446e-01
5.73013783e-01 3.49318177e-01 -3.13686281e-01 9.16417062e-01
1.02204907e+00 3.87793571e-01 -4.15449053e-01 -6.17408633e-01
-5.32703340e-01 -5.64695120e-01 -1.23653448e+00 1.18517792e+00
-6.58016086e-01 6.96590185e-01 4.53184754e-01 -7.40726233e-01
1.21545136e+00 1.11566201e-01 2.98402488e-01 -9.74780202e-01
4.04091448e-01 5.27959049e-01 -3.35945159e-01 -1.09823811e+00
7.06552684e-01 -5.16162217e-01 3.76602471e-01 4.42848206e-02
-7.28666112e-02 5.06012142e-02 2.40618780e-01 -8.57047364e-02
6.47907078e-01 1.94962206e-03 1.33363217e-01 -2.04189315e-01
6.46182001e-01 -6.86660051e-01 7.39857674e-01 3.36187571e-01
-3.72887433e-01 8.45171571e-01 -2.16044694e-01 -1.71333492e-01
-8.52277160e-01 -9.04017866e-01 -4.38475847e-01 6.87679410e-01
1.08029783e+00 3.85847278e-02 -4.39661652e-01 -1.97107971e-01
-3.52373302e-01 1.04896940e-01 -5.93839526e-01 -2.76336432e-01
-2.90634096e-01 -1.04646969e+00 3.57086092e-01 4.79719996e-01
1.10040915e+00 -1.06182098e+00 -3.35952878e-01 4.11083668e-01
-2.43280202e-01 -9.09773171e-01 -4.18209970e-01 -1.79359078e-01
-8.44750285e-01 -8.89249206e-01 -1.19127297e+00 -8.59867096e-01
5.70301652e-01 7.30521917e-01 6.48105383e-01 4.31236655e-01
-3.23777437e-01 6.54955879e-02 -7.82718539e-01 -3.10860157e-01
-2.59205103e-01 -8.62471819e-01 1.19135439e-01 1.23600312e-01
-2.26169124e-01 -4.88640636e-01 -9.63536978e-01 5.27946770e-01
-1.42544854e+00 -4.21227396e-01 8.21734190e-01 8.99970770e-01
4.15833205e-01 2.86767930e-01 2.41425991e-01 3.95666771e-02
5.44591010e-01 -3.98214549e-01 -4.69677150e-01 1.86098322e-01
-6.16153955e-01 -1.53719142e-01 3.85632694e-01 -1.85609818e-01
-1.28640962e+00 -5.15441179e-01 -5.13841927e-01 -7.04012141e-02
1.87101141e-01 1.02763009e+00 -3.19712400e-01 -5.59751749e-01
2.69426346e-01 6.08187079e-01 2.07579806e-01 -4.86901462e-01
-3.60917300e-01 9.15549338e-01 6.49302721e-01 -1.92170441e-01
9.61484492e-01 4.59949076e-01 -3.33895832e-02 -1.14511991e+00
-6.21283591e-01 -4.88214135e-01 -1.00905351e-01 -5.53403974e-01
8.05021584e-01 -1.03834248e+00 -6.36214018e-01 9.82156515e-01
-9.54989254e-01 -2.46676747e-02 2.76431978e-01 7.89397001e-01
3.97334173e-02 1.08017445e+00 -8.31026435e-01 -8.94562662e-01
-4.67065454e-01 -1.42876565e+00 8.60427380e-01 1.00821257e+00
7.21785188e-01 -9.05729532e-01 -3.83123234e-02 2.02289507e-01
7.32653618e-01 2.31629759e-01 2.78989643e-01 2.23946661e-01
-3.69277179e-01 -1.17005564e-01 -7.03144372e-01 9.02337909e-01
2.74191290e-01 2.26872221e-01 -6.61966741e-01 -4.13317740e-01
-5.49746975e-02 -4.35637027e-01 9.69784260e-01 1.81077600e-01
9.84344363e-01 -2.86722090e-02 2.40656927e-01 9.57320452e-01
1.85111141e+00 4.57763582e-01 1.12321103e+00 7.23855853e-01
4.24684525e-01 4.56676275e-01 8.75220239e-01 6.09686077e-01
3.84471387e-01 3.04010838e-01 6.50147140e-01 -5.72263956e-01
1.58049509e-01 -2.77126953e-02 5.60825646e-01 9.77531910e-01
-7.10888028e-01 -1.97813839e-01 -3.68037015e-01 4.42991376e-01
-1.17568564e+00 -8.51017177e-01 -1.06538855e-01 1.80621672e+00
8.16685498e-01 -2.22511157e-01 -5.20456970e-01 3.27509522e-01
6.06611907e-01 1.41485780e-01 -3.57562959e-01 9.88017395e-02
-4.53002661e-01 4.46075611e-02 5.50810635e-01 4.10747439e-01
-1.13411868e+00 3.55525643e-01 5.52299881e+00 6.32306993e-01
-1.16074491e+00 -1.22381523e-01 2.15230495e-01 5.61106265e-01
-3.80932659e-01 -2.77770281e-01 -5.14486134e-01 4.92145687e-01
3.44392210e-01 -3.64234969e-02 1.48791969e-01 5.33345342e-01
4.59652364e-01 -4.29075733e-02 -2.39563748e-01 9.83938813e-01
3.26291472e-01 -1.00365233e+00 -1.17976598e-01 4.96944077e-02
8.77022743e-01 -1.32026002e-01 -4.88247313e-02 1.69561327e-01
-3.27653773e-02 -7.99392104e-01 5.20281553e-01 7.52704978e-01
9.65428710e-01 -5.83923817e-01 1.41844022e+00 -1.55377239e-01
-1.38836753e+00 -1.86901063e-01 -7.40157783e-01 -4.14122380e-02
3.36742848e-01 3.89008075e-01 9.29226205e-02 7.35527039e-01
1.12681985e+00 7.17099011e-01 -3.67832065e-01 1.63098717e+00
-2.91454345e-01 3.60850483e-01 -1.28549844e-01 -1.18233832e-02
3.76251698e-01 -5.64126492e-01 4.51801300e-01 1.13201368e+00
7.60633707e-01 5.89407325e-01 -3.93964425e-02 3.10847342e-01
2.61078589e-02 3.11130315e-01 -1.87192500e-01 1.77441016e-01
2.19419092e-01 1.26281834e+00 -4.42271203e-01 -1.57449260e-01
-5.54265618e-01 1.04648602e+00 -5.06125212e-01 3.67145091e-01
-5.62552452e-01 -9.93749499e-01 6.26564026e-01 -4.12516952e-01
8.74353200e-02 -2.95665920e-01 1.64412931e-01 -1.17014122e+00
5.45265898e-03 -7.21959472e-01 3.85337658e-02 -9.92646277e-01
-1.34143174e+00 7.87050784e-01 -2.21503153e-01 -1.67023313e+00
6.82509422e-01 -6.01630092e-01 -8.92488360e-01 7.97089100e-01
-2.15839005e+00 -5.58700860e-01 -1.11290848e+00 5.18843770e-01
5.09029925e-01 -8.36610720e-02 6.06931865e-01 6.27250135e-01
-6.62497044e-01 5.18479288e-01 6.30997956e-01 3.58189523e-01
5.55204391e-01 -1.02473855e+00 -3.54756534e-01 1.03211141e+00
-5.46555638e-01 3.76505911e-01 8.45446885e-01 -5.07815897e-01
-1.36191893e+00 -8.61107886e-01 3.01446795e-01 4.79173273e-01
5.00691712e-01 5.23527622e-01 -1.25666332e+00 5.09351008e-02
1.41002223e-01 1.43503293e-01 7.93700397e-01 -6.02549851e-01
7.70119056e-02 -3.91349912e-01 -1.15001202e+00 3.40559006e-01
5.82858682e-01 -1.95570573e-01 -4.94846344e-01 -1.47788122e-01
5.27588010e-01 -1.45717442e-01 -1.28727520e+00 8.16887140e-01
6.10897720e-01 -1.20638263e+00 9.34592783e-01 2.39898592e-01
7.46263087e-01 -7.89962709e-01 -4.49196219e-01 -1.40413713e+00
-6.79161027e-02 -1.79859236e-01 5.16266763e-01 1.32312167e+00
2.92343646e-01 -5.27585983e-01 2.56416112e-01 2.14481458e-01
-6.01266623e-01 -5.28565824e-01 -7.01653242e-01 -6.73385561e-01
-3.41735572e-01 1.60710197e-02 4.27541584e-01 6.19712055e-01
-1.82170138e-01 -1.30713340e-02 -7.37937987e-01 5.75180233e-01
9.47261810e-01 4.19128835e-02 5.42298913e-01 -1.17400122e+00
-1.21796224e-02 -3.59785259e-01 -5.97468674e-01 -1.00004530e+00
-3.72290999e-01 -2.83340752e-01 6.77793145e-01 -2.01198721e+00
3.22074592e-01 -2.28810921e-01 -4.56268698e-01 1.88171223e-01
-6.70984268e-01 7.23218322e-01 1.39075980e-01 2.61737704e-01
-4.33422327e-01 1.28190792e+00 1.60027790e+00 -2.25410998e-01
-8.07162449e-02 -1.79422766e-01 -5.22927582e-01 7.03186691e-01
6.43488705e-01 -3.62089486e-03 8.62652902e-03 -4.79874104e-01
3.11498567e-02 2.41172090e-01 5.80504946e-02 -1.23544633e+00
2.39890516e-01 -2.62846977e-01 3.62780333e-01 -2.30799183e-01
3.06322366e-01 -7.31820107e-01 -2.94286251e-01 5.67791104e-01
7.05188373e-03 -2.01439440e-01 -2.40262114e-02 6.50639594e-01
-9.31954741e-01 -5.46765804e-01 1.15019071e+00 -1.31027892e-01
-1.48353481e+00 5.20612955e-01 -3.43048275e-01 -2.24653617e-01
7.16195524e-01 -5.05541682e-01 -2.39278883e-01 -4.05992031e-01
-3.44890386e-01 2.63321638e-01 4.68047887e-01 1.17787287e-01
1.44466662e+00 -9.08988953e-01 -9.53583419e-01 4.37528379e-02
3.40580910e-01 -1.67567015e-01 7.07245946e-01 7.07217872e-01
-1.06368995e+00 -7.14544177e-01 -6.38222635e-01 -2.42564887e-01
-1.47720551e+00 5.47165796e-02 4.68079329e-01 1.43205866e-01
-4.35195237e-01 8.91028881e-01 7.62068201e-03 -1.85918242e-01
1.15160495e-01 5.02917171e-02 -7.15590000e-01 -2.80222278e-02
8.97140145e-01 6.57210588e-01 -1.90718204e-01 -7.82610416e-01
-9.18289274e-02 1.05482125e+00 3.74102950e-01 1.20114952e-01
1.66209829e+00 -6.14142478e-01 -2.78019935e-01 6.24818057e-02
1.41689050e+00 2.08424367e-02 -1.52675831e+00 -3.48782003e-01
-5.39828479e-01 -6.72099829e-01 3.99929047e-01 -5.25112391e-01
-1.33395016e+00 1.09323823e+00 1.00400174e+00 9.56412479e-02
1.59816468e+00 -4.21440989e-01 1.14556146e+00 3.08306277e-01
2.60474175e-01 -1.03115761e+00 3.65784138e-01 5.33109665e-01
9.46576774e-01 -1.61022186e+00 3.21337968e-01 -3.07273179e-01
-6.16785705e-01 1.30369127e+00 5.45431614e-01 1.71235010e-01
6.27295852e-01 3.06539357e-01 5.04422307e-01 -1.18203564e-02
1.51041955e-01 -3.96847010e-01 1.10630453e-01 6.00107968e-01
1.48211777e-01 -2.65643179e-01 -4.24208611e-01 3.67209136e-01
1.37490064e-01 -5.11603281e-02 7.26292014e-01 8.31876993e-01
-1.07888019e+00 -7.77489722e-01 -4.08542395e-01 2.08024025e-01
-5.22374570e-01 -1.49846762e-01 4.65936244e-01 4.84497309e-01
2.15424985e-01 1.22449899e+00 -3.25212091e-01 -7.42903113e-01
4.72698897e-01 -8.54064882e-01 1.00108132e-01 -1.38912678e-01
-2.36345846e-02 1.25851892e-02 -2.08671451e-01 -1.95964515e-01
-9.67724502e-01 -2.73346335e-01 -1.40458834e+00 4.61929291e-02
-3.90639871e-01 3.72267634e-01 7.44125664e-01 8.63576472e-01
-2.95876056e-01 3.97165984e-01 1.02958775e+00 -1.24978626e+00
-3.83292794e-01 -1.17377603e+00 -9.83885348e-01 4.52971071e-01
4.80299294e-01 -7.81553268e-01 -8.57137382e-01 -5.47493547e-02] | [10.697052001953125, -3.517683267593384] |
d138f78f-2840-48a7-93a3-6680e615e9c1 | keyphrase-generation-for-scientific-articles | 1909.12229 | null | https://arxiv.org/abs/1909.12229v1 | https://arxiv.org/pdf/1909.12229v1.pdf | Keyphrase Generation for Scientific Articles using GANs | In this paper, we present a keyphrase generation approach using conditional Generative Adversarial Networks (GAN). In our GAN model, the generator outputs a sequence of keyphrases based on the title and abstract of a scientific article. The discriminator learns to distinguish between machine-generated and human-curated keyphrases. We evaluate this approach on standard benchmark datasets. Our model achieves state-of-the-art performance in generation of abstractive keyphrases and is also comparable to the best performing extractive techniques. We also demonstrate that our method generates more diverse keyphrases and make our implementation publicly available. | ['Rajiv Ratn Shah', 'Haimin Zhang', 'Avinash Swaminathan', 'Rakesh Gosangi', 'Raj Kuwar Gupta', 'Debanjan Mahata'] | 2019-09-24 | null | null | null | null | ['keyphrase-generation'] | ['natural-language-processing'] | [ 4.22385752e-01 1.24292344e-01 -2.31332809e-01 3.49303126e-01
-1.39828300e+00 -1.33001649e+00 1.19649184e+00 2.14608535e-01
-4.77229804e-01 1.20112789e+00 2.98082858e-01 -5.90524316e-01
3.30422014e-01 -1.20064330e+00 -9.99979079e-01 -5.73911071e-01
3.75394225e-02 3.15089107e-01 -6.96437806e-02 -2.99301654e-01
5.20856202e-01 5.99294603e-01 -1.05561042e+00 3.84492308e-01
6.18681908e-01 7.71182418e-01 -3.77577990e-01 1.29957545e+00
-3.44809853e-02 7.20537484e-01 -1.23335385e+00 -7.37848520e-01
5.83832383e-01 -6.16913795e-01 -7.96431839e-01 -6.18024290e-01
8.29755962e-01 -5.35140157e-01 -6.15000069e-01 9.24722731e-01
3.75837505e-01 -2.24215567e-01 9.23444331e-01 -1.50794518e+00
-8.73605430e-01 9.49131846e-01 -1.52940825e-01 2.64441162e-01
5.43665826e-01 9.37358662e-02 1.21242273e+00 -7.60462344e-01
8.31882417e-01 9.03546393e-01 3.07706237e-01 5.83299220e-01
-1.35677719e+00 -1.16137624e+00 -3.33126485e-01 -2.23402575e-01
-1.15598321e+00 -5.84823489e-02 8.18072557e-01 -1.80619195e-01
5.12832999e-01 4.35853451e-01 7.65956759e-01 1.53414774e+00
7.37957716e-01 9.00483847e-01 1.38041127e+00 -3.60028803e-01
2.65154570e-01 -1.93747878e-01 -3.19142789e-01 5.79643011e-01
4.16805029e-01 3.55996788e-01 -5.74078321e-01 -8.37009668e-01
8.31805468e-01 -5.52012213e-02 -2.58329958e-01 1.42123908e-01
-1.79012406e+00 9.46897328e-01 8.72449875e-02 8.56270343e-02
-5.48650205e-01 4.40374076e-01 2.04131335e-01 6.80686891e-01
3.15740019e-01 1.12484550e+00 -3.38450760e-01 -9.83824283e-02
-1.21753180e+00 9.16437447e-01 1.20810080e+00 8.76881599e-01
5.77130139e-01 9.23537090e-02 -6.48117721e-01 1.26330927e-01
-3.50441694e-01 6.95021749e-01 2.27968752e-01 -7.32362688e-01
3.17651898e-01 2.98321813e-01 1.53470621e-01 -8.57253551e-01
3.32470238e-01 -5.74389212e-02 -9.56334889e-01 1.38996810e-01
1.21714100e-01 -4.51260597e-01 -1.30775154e+00 1.29910851e+00
-8.27763602e-02 2.85641432e-01 2.87568539e-01 1.57168627e-01
1.08153236e+00 9.17089105e-01 2.87389252e-02 5.64478412e-02
1.40845001e+00 -7.29278207e-01 -7.04600930e-01 1.57234251e-01
-1.13144428e-01 -1.07513678e+00 7.63864756e-01 4.94574070e-01
-1.08476961e+00 -3.31741035e-01 -1.19536269e+00 5.32416552e-02
-9.38074350e-01 1.34933889e-01 6.63324893e-01 5.50831676e-01
-1.01650143e+00 5.83015859e-01 -3.14072758e-01 1.48661494e-01
3.74519587e-01 1.30338624e-01 -4.41426992e-01 3.12101841e-01
-1.45394230e+00 4.22250032e-01 8.05356085e-01 -6.91869557e-01
-1.26429558e+00 -1.16275966e+00 -6.40334725e-01 -1.19561613e-01
4.37464029e-01 -9.77410972e-01 1.48221874e+00 -3.97398978e-01
-1.42045999e+00 9.43114221e-01 2.50262469e-01 -7.47034967e-01
4.68451977e-01 -3.59236985e-01 -5.84815919e-01 5.34353375e-01
4.12933417e-02 6.83030784e-01 1.37499213e+00 -1.22236073e+00
-6.41361713e-01 2.96638936e-01 2.75802672e-01 -3.32005739e-01
-3.20882112e-01 8.39000270e-02 -1.77568942e-01 -1.65040684e+00
-5.22274196e-01 -1.00786757e+00 -1.58118501e-01 -2.90370613e-01
-9.00691271e-01 -4.51079616e-03 8.64665926e-01 -6.96815252e-01
1.35276592e+00 -1.65844011e+00 2.63673682e-02 3.40566397e-01
4.98278171e-01 1.99235201e-01 -1.38380051e-01 1.07223582e+00
-8.81978348e-02 5.57270944e-01 -2.96020746e-01 1.96963586e-02
1.75241008e-02 1.22767575e-02 -1.24190462e+00 6.52490482e-02
1.42612576e-01 1.27303648e+00 -1.01843572e+00 -3.85774583e-01
-2.77260989e-01 2.39960641e-01 -3.37121189e-01 5.86688697e-01
-6.43798649e-01 2.54011124e-01 -5.51385343e-01 6.24659538e-01
5.26856363e-01 -2.42032886e-01 -8.59783888e-02 -2.11517334e-01
2.76604027e-01 4.28540319e-01 -8.69700432e-01 1.47258592e+00
-3.01897615e-01 7.07138240e-01 -2.67555237e-01 -4.65344220e-01
8.60999167e-01 4.58787531e-01 2.44863883e-01 3.70090653e-04
-1.06414512e-01 1.39896601e-01 -5.14893651e-01 4.62288260e-01
1.01102173e+00 9.35525447e-02 -6.17786288e-01 1.09314525e+00
3.13604534e-01 -1.02780700e+00 3.65392625e-01 8.88901472e-01
1.15411818e+00 2.39274628e-03 5.46098530e-01 -1.83380455e-01
3.95018727e-01 -2.28474095e-01 -1.48269966e-01 1.22629976e+00
5.05340815e-01 5.88990390e-01 7.31283963e-01 -6.53733373e-01
-1.23438239e+00 -1.21457481e+00 3.69327664e-01 7.64062285e-01
-3.25203180e-01 -1.10347080e+00 -8.24719191e-01 -1.21187663e+00
1.62889495e-01 7.68087089e-01 -1.01992261e+00 -2.77928889e-01
-6.36491299e-01 -4.17908192e-01 1.06212199e+00 4.02697116e-01
4.32938099e-01 -1.28212619e+00 -4.14071351e-01 -4.25320715e-02
1.84622649e-02 -1.10468411e+00 -8.54128659e-01 -8.43161531e-03
-4.23839569e-01 -1.08230364e+00 -1.11438417e+00 -5.54019332e-01
7.21150577e-01 -1.12248100e-01 1.62651885e+00 4.79424885e-03
-1.97388709e-01 6.14833832e-01 -3.36638302e-01 -8.14951718e-01
-1.18925726e+00 4.22459394e-01 -2.05540657e-01 -2.04586148e-01
3.51761952e-02 -5.08009970e-01 -6.00240886e-01 -3.22228283e-01
-1.55287123e+00 5.34496866e-02 7.02393770e-01 8.05698991e-01
6.36685073e-01 1.51133507e-01 1.09862067e-01 -1.11935520e+00
1.27822483e+00 -1.84728369e-01 -7.82786608e-01 2.33646825e-01
-8.99497032e-01 4.89822626e-01 1.03072751e+00 -6.54983640e-01
-5.27934015e-01 -3.66713047e-01 6.99739233e-02 -2.18742982e-01
-9.86787304e-02 1.93642244e-01 1.97667584e-01 -1.83922976e-01
7.33436406e-01 7.42091179e-01 -3.72848570e-01 -2.73023725e-01
6.57161057e-01 4.10382658e-01 9.17579949e-01 -9.46200788e-01
1.57517874e+00 3.28286499e-01 5.32993898e-02 -3.45585674e-01
-6.83729529e-01 -9.34644714e-02 -1.72060907e-01 1.86802551e-01
5.69679379e-01 -9.95734215e-01 -5.31702101e-01 5.90732515e-01
-1.21463609e+00 -2.01964289e-01 -5.27198672e-01 -2.72615850e-02
-4.89867121e-01 2.89834201e-01 -7.77765989e-01 -1.85622826e-01
-1.18599045e+00 -8.91068220e-01 1.38153934e+00 2.91026443e-01
-3.62753093e-01 -8.69902730e-01 3.33803594e-01 -1.83019638e-01
3.87607157e-01 8.23355794e-01 8.86490107e-01 -9.65626001e-01
-7.48840809e-01 -5.30474484e-01 4.61892560e-02 3.22818369e-01
2.25999802e-01 1.40413672e-01 -7.17515528e-01 -3.03404093e-01
-5.15930474e-01 -5.52404463e-01 1.15037465e+00 -1.50542557e-01
1.28757203e+00 -1.07126188e+00 -3.25295061e-01 6.74715459e-01
1.25565743e+00 1.73556153e-02 7.99737751e-01 4.41428155e-01
5.76130629e-01 -3.92672718e-02 1.83301702e-01 3.38766098e-01
1.81450531e-01 2.29335532e-01 4.76122834e-02 -1.53511286e-01
8.87733772e-02 -9.41717684e-01 2.53801048e-01 4.60842133e-01
-7.15601221e-02 -4.05364066e-01 -6.83595836e-01 4.69756871e-01
-1.27170515e+00 -1.12020564e+00 3.77256215e-01 1.85426271e+00
1.53399813e+00 4.13576454e-01 2.06397444e-01 8.72962251e-02
2.07415193e-01 5.45389712e-01 -1.93490177e-01 -5.45007288e-01
-1.81710005e-01 1.19008541e+00 9.62117910e-01 2.72208691e-01
-1.16563416e+00 1.20521855e+00 7.69063759e+00 8.80309701e-01
-9.97378826e-01 -2.90813923e-01 5.12062848e-01 2.62808979e-01
-7.87263811e-01 -1.66970603e-02 -7.65454352e-01 4.76238310e-01
1.01480913e+00 -7.48454213e-01 4.37335372e-01 8.09565127e-01
-5.87028086e-01 7.69775212e-02 -1.05284452e+00 8.76209140e-01
4.43058312e-02 -1.86095285e+00 6.76206350e-01 1.76154077e-01
1.12373745e+00 -4.20811057e-01 1.82731435e-01 2.45935127e-01
1.06736851e+00 -1.24270058e+00 5.73040962e-01 5.87822914e-01
1.05523062e+00 -1.00430441e+00 3.95206630e-01 -7.77494088e-02
-1.00469112e+00 3.36444885e-01 8.48404467e-02 3.85839701e-01
-3.19722384e-01 5.94965398e-01 -1.02607954e+00 6.41597033e-01
4.13925976e-01 3.47972929e-01 -6.74124181e-01 4.74933118e-01
-8.39477897e-01 7.79318213e-01 -9.18485597e-02 -2.74946630e-01
2.79542863e-01 1.02111466e-01 6.11138105e-01 1.50645983e+00
1.97603881e-01 7.81795010e-02 2.14971751e-01 1.04636776e+00
-7.09666967e-01 -1.12099171e-01 -8.11012983e-01 -8.97481620e-01
4.90324378e-01 1.40968060e+00 -6.71486616e-01 -8.43496263e-01
5.93260787e-02 1.32448292e+00 -2.27586657e-01 4.95496809e-01
-5.05873144e-01 -8.19346547e-01 6.73908293e-01 -1.48597777e-01
4.96085554e-01 -3.21898133e-01 3.91378328e-02 -1.24244094e+00
-1.86738521e-01 -1.55120981e+00 4.91018057e-01 -5.98389447e-01
-1.27285576e+00 6.71929121e-01 1.96457580e-01 -1.03403544e+00
-5.87365091e-01 -4.39252347e-01 -7.32113600e-01 1.01696384e+00
-1.15197062e+00 -1.37860012e+00 -5.74674606e-02 4.47822064e-01
1.85791969e-01 -4.96319383e-01 1.30089688e+00 -4.30480599e-01
2.35821307e-01 8.75549734e-01 1.48353036e-02 5.80920458e-01
8.67307305e-01 -1.62990975e+00 1.42237401e+00 8.12838554e-01
5.21385193e-01 8.83058369e-01 8.70739341e-01 -8.83867443e-01
-1.65069699e+00 -8.68273556e-01 6.52350187e-01 -6.00719333e-01
8.92745852e-01 -5.76647758e-01 -5.47071695e-01 6.30324185e-01
6.71521008e-01 -1.05260298e-01 8.13151598e-01 -4.59055841e-01
-8.07768822e-01 1.75172642e-01 -1.01360178e+00 8.91379237e-01
5.56369245e-01 -8.32690239e-01 -7.80372679e-01 2.94792533e-01
1.08240259e+00 -7.12352633e-01 -9.19148326e-01 3.56469929e-01
7.45913446e-01 -4.53777581e-01 1.18203890e+00 -7.94351518e-01
8.77710998e-01 -1.41512379e-01 2.60621786e-01 -1.58545172e+00
-1.01258941e-01 -1.30559468e+00 -7.34083235e-01 8.09981704e-01
2.54860103e-01 -6.42400265e-01 7.63734341e-01 3.29180360e-02
5.00164330e-01 -5.39384067e-01 -3.71603310e-01 -6.77291274e-01
3.56226265e-01 1.14999898e-02 8.75846982e-01 9.21410322e-01
-4.64828074e-01 3.06966156e-01 -6.16425276e-01 -1.42520651e-01
6.73533142e-01 3.90138358e-01 1.33853149e+00 -9.81916487e-01
-4.93884921e-01 -5.31231940e-01 -9.90874618e-02 -7.51951933e-01
2.75602221e-01 -8.65696430e-01 -3.73822272e-01 -1.31866109e+00
-5.52753285e-02 2.24768326e-01 -3.26168001e-01 6.59383953e-01
-5.40487230e-01 4.27841276e-01 2.07903191e-01 1.15248404e-01
-1.56343877e-01 2.83329785e-01 1.25500977e+00 -5.43492854e-01
-6.89932778e-02 1.64603535e-02 -9.65452850e-01 1.29174784e-01
9.80263710e-01 -7.51398623e-01 -1.84117436e-01 3.84256124e-01
4.28271383e-01 -1.75184146e-01 5.82573533e-01 -7.76583731e-01
4.92940024e-02 -4.51447278e-01 5.71233928e-01 -7.01732874e-01
-3.78605239e-02 -3.14530641e-01 7.78971687e-02 4.92814839e-01
-5.37043214e-01 3.48940998e-01 5.04950345e-01 4.87276316e-01
-9.99519601e-02 1.47933781e-01 3.53249013e-01 -4.54037756e-01
-2.34967455e-01 4.96257842e-01 -3.72414500e-01 3.22539657e-01
5.93608737e-01 3.21090400e-01 -3.65387619e-01 -7.29580998e-01
3.72707844e-03 -3.05600613e-01 5.49481690e-01 5.32403350e-01
7.44578362e-01 -1.31371391e+00 -1.04819274e+00 2.72800177e-01
2.34844223e-01 -3.95316511e-01 -4.92876351e-01 -8.51030871e-02
-6.05826974e-01 4.41572607e-01 -2.59800404e-01 3.38932842e-01
-9.46098030e-01 6.31623685e-01 -8.75091925e-02 -7.19795406e-01
-5.62793970e-01 6.73534393e-01 -1.96261890e-02 -1.55506626e-01
-2.82495283e-02 -4.03427780e-01 2.09926963e-01 -1.24256931e-01
9.24083471e-01 1.32220671e-01 -8.66373032e-02 -1.83152229e-01
-2.99586765e-02 1.18455425e-01 -4.44290906e-01 -4.04738486e-01
1.03901684e+00 9.05509174e-01 -1.49829030e-01 1.35239482e-01
1.26757538e+00 7.14436531e-01 -7.15706706e-01 -1.88389659e-01
-4.14567381e-01 -3.55024785e-02 8.97292234e-03 -1.01551569e+00
-8.93481493e-01 2.49881804e-01 1.82494834e-01 3.29003990e-01
1.09985673e+00 -1.46993205e-01 1.34316385e+00 4.91217464e-01
2.69307762e-01 -8.62476051e-01 3.16732049e-01 2.45018348e-01
1.02939177e+00 -9.22764421e-01 4.88502860e-01 -2.94820638e-03
-2.65883178e-01 1.22556877e+00 1.39711872e-01 -3.58843207e-01
5.23381233e-01 6.47660792e-01 2.39474699e-02 -9.41760764e-02
-7.67128766e-01 1.71609759e-01 8.16372871e-01 4.80021894e-01
1.92630604e-01 2.67977566e-01 -4.07763749e-01 3.85650933e-01
-8.24784756e-01 9.79019031e-02 6.60130203e-01 1.25685906e+00
4.76273708e-02 -1.49976897e+00 -6.27085805e-01 4.71152842e-01
-1.23238075e+00 -6.37790084e-01 -9.84450281e-01 6.17932320e-01
-3.59786451e-01 5.45304596e-01 -2.96363562e-01 -5.72633326e-01
-3.79202068e-02 1.65504605e-01 4.84219134e-01 -4.94909585e-01
-9.22878742e-01 -3.06647569e-01 -1.22630566e-01 -2.85787702e-01
-7.98458606e-02 -6.03918284e-02 -8.81850719e-01 -5.25017798e-01
1.59898102e-01 4.79289919e-01 6.45295024e-01 4.18411821e-01
4.67800230e-01 3.64222109e-01 6.47420347e-01 -5.91716170e-01
-5.36804080e-01 -8.48678470e-01 -1.71379492e-01 4.50715661e-01
4.53678697e-01 1.33795971e-02 -2.91601181e-01 5.49197614e-01] | [12.277429580688477, 8.89411735534668] |
362d473e-ac41-4e1b-9051-f6cc92230fa4 | probabilistic-assumptions-matter-improved | 2005.01898 | null | https://arxiv.org/abs/2005.01898v1 | https://arxiv.org/pdf/2005.01898v1.pdf | Probabilistic Assumptions Matter: Improved Models for Distantly-Supervised Document-Level Question Answering | We address the problem of extractive question answering using document-level distant super-vision, pairing questions and relevant documents with answer strings. We compare previously used probability space and distant super-vision assumptions (assumptions on the correspondence between the weak answer string labels and possible answer mention spans). We show that these assumptions interact, and that different configurations provide complementary benefits. We demonstrate that a multi-objective model can efficiently combine the advantages of multiple assumptions and out-perform the best individual formulation. Our approach outperforms previous state-of-the-art models by 4.3 points in F1 on TriviaQA-Wiki and 1.7 points in Rouge-L on NarrativeQA summaries. | ['Kristina Toutanova', 'Ming-Wei Chang', 'Kenton Lee', 'Hao Cheng'] | 2020-05-05 | probabilistic-assumptions-matter-improved-1 | https://aclanthology.org/2020.acl-main.501 | https://aclanthology.org/2020.acl-main.501.pdf | acl-2020-6 | ['triviaqa'] | ['miscellaneous'] | [-3.69432718e-02 2.11692989e-01 3.03669386e-02 -2.89457142e-01
-1.76337111e+00 -1.29889262e+00 7.29802728e-01 3.42091173e-01
-5.08324802e-01 5.87327778e-01 6.06373668e-01 -3.00157368e-01
-6.16762877e-01 -4.81663495e-01 -7.01004326e-01 -9.84645486e-02
3.99150163e-01 1.16689563e+00 6.05663896e-01 -4.90044951e-01
4.14206058e-01 1.41943535e-02 -1.41345370e+00 1.06446826e+00
9.99684870e-01 7.37919152e-01 2.19408609e-02 1.28210855e+00
-7.70021498e-01 1.06214786e+00 -6.76769733e-01 -1.00326550e+00
-1.33599386e-01 -3.01325142e-01 -1.44993424e+00 -3.56973052e-01
1.19089556e+00 -3.42552036e-01 -2.91467935e-01 8.68372262e-01
4.02713031e-01 5.20073511e-02 8.04695189e-01 -9.71503079e-01
-1.14751482e+00 5.26245892e-01 -2.28099525e-01 6.99882746e-01
9.30459857e-01 1.46307081e-01 1.70697546e+00 -1.05598164e+00
6.95152402e-01 1.40322518e+00 5.90997279e-01 4.41108674e-01
-1.12022519e+00 4.93787881e-03 1.35301286e-02 6.35975361e-01
-7.63864636e-01 -3.92929405e-01 4.31506217e-01 -4.42215145e-01
1.27319539e+00 6.10619605e-01 3.61001603e-02 8.62954736e-01
-1.71976462e-01 1.20494425e+00 1.09316015e+00 -4.49020743e-01
-5.39388731e-02 9.89436135e-02 9.66548860e-01 8.75542700e-01
-1.85513198e-01 -2.78253973e-01 -6.63935900e-01 -3.69445920e-01
1.10536613e-01 -4.46052611e-01 -3.11116457e-01 -1.39703467e-01
-1.09856677e+00 9.48947132e-01 2.57247657e-01 2.08802164e-01
-2.63323635e-01 2.19245419e-01 7.41508827e-02 3.85453850e-01
3.18188608e-01 9.22016025e-01 -3.67995858e-01 -2.89568126e-01
-9.37122107e-01 5.78613877e-01 9.32351828e-01 8.23958397e-01
4.63635266e-01 -3.95322323e-01 -8.82353544e-01 7.66036034e-01
1.40894547e-01 4.52437699e-01 1.32598162e-01 -1.52233028e+00
7.60178685e-01 5.47719717e-01 3.08703423e-01 -8.77133667e-01
-2.94256061e-01 -5.39777815e-01 -1.05978111e-02 -3.57233256e-01
8.47308338e-01 -4.61492687e-02 -7.60422051e-01 1.31411326e+00
1.31030276e-01 -2.06362218e-01 2.12977007e-01 8.34414303e-01
1.44689775e+00 8.48742545e-01 6.95332885e-02 -1.88825261e-02
1.61227441e+00 -1.29667151e+00 -8.48427713e-01 -3.12574446e-01
3.09861213e-01 -8.09188843e-01 1.30720985e+00 1.36784866e-01
-1.51298380e+00 -5.70483446e-01 -6.80544794e-01 -4.88461226e-01
-1.80626199e-01 1.47836238e-01 3.02548200e-01 4.74086136e-01
-1.15985930e+00 3.30026299e-01 -1.27258301e-01 -2.01667294e-01
2.49079898e-01 1.36694893e-01 -9.46087539e-02 -2.58740634e-01
-1.02303994e+00 1.16054988e+00 1.22178249e-01 -2.90509492e-01
-9.42265511e-01 -8.81767869e-01 -3.58813226e-01 2.96153456e-01
6.04892790e-01 -1.06469178e+00 1.64081454e+00 -3.78967732e-01
-1.13034117e+00 9.71173763e-01 -3.84334058e-01 -4.37603414e-01
3.03285450e-01 -5.24299741e-01 -1.86133385e-01 6.53025627e-01
1.12169005e-01 8.33182514e-01 6.38129115e-01 -1.31125879e+00
-7.55328357e-01 -4.27991599e-01 7.16081619e-01 4.13443506e-01
-3.14593256e-01 1.86851934e-01 -7.49884665e-01 -1.52153179e-01
-7.53674209e-02 -5.92398822e-01 1.44194990e-01 -2.14719906e-01
-1.79814726e-01 -7.26431668e-01 3.79056036e-01 -1.12666345e+00
1.12658107e+00 -1.52508366e+00 4.48816925e-01 -3.16301435e-01
3.56072813e-01 1.99438870e-01 -3.97709250e-01 4.45884496e-01
3.44649851e-01 1.69935506e-02 -9.20584649e-02 -3.01188111e-01
2.70593256e-01 1.81471631e-01 -5.30917287e-01 1.82531644e-02
2.07398385e-01 1.24983919e+00 -1.09756923e+00 -7.22060144e-01
-2.83852488e-01 3.27401562e-03 -4.07992959e-01 2.33649239e-01
-6.01898909e-01 -9.14853364e-02 -3.81236821e-01 4.98308182e-01
2.50051290e-01 -3.72195750e-01 -1.05780892e-01 -2.16676161e-01
1.08395174e-01 4.36089307e-01 -8.46939921e-01 1.60923755e+00
-3.83859724e-01 7.96852887e-01 7.12421089e-02 -6.06965840e-01
7.36279547e-01 2.22742185e-01 -1.80814490e-01 -7.66100049e-01
-1.78493738e-01 1.39113486e-01 -8.64224881e-02 -8.63628328e-01
6.61338627e-01 9.50537343e-03 -3.27006616e-02 4.26723629e-01
5.58737814e-01 -3.16501588e-01 3.31744164e-01 6.64966524e-01
1.16086996e+00 6.53484911e-02 -8.21157843e-02 -1.80530682e-01
5.40967226e-01 3.58360827e-01 -5.86917549e-02 1.07929552e+00
-2.04368681e-01 7.48634696e-01 7.48755217e-01 -1.56863302e-01
-8.23079348e-01 -1.35653746e+00 1.80360720e-01 1.56517649e+00
-8.27299282e-02 -3.86531800e-01 -6.64224207e-01 -1.20855474e+00
5.17357364e-02 1.37542057e+00 -5.26825249e-01 1.88841358e-01
-6.90118372e-01 -3.74834061e-01 7.54346609e-01 6.64138138e-01
1.00631662e-01 -8.85175347e-01 -3.92696619e-01 1.84470043e-02
-6.32137954e-01 -1.19047344e+00 -3.44877839e-01 -3.28092016e-02
-7.09377885e-01 -9.77948427e-01 -8.76656592e-01 -6.77732825e-01
1.35974824e-01 2.71062076e-01 1.85962605e+00 -2.31334306e-02
2.74925977e-02 9.09493387e-01 -3.79855305e-01 -4.07499433e-01
-3.10621977e-01 1.18289106e-01 -4.79298592e-01 -2.61248678e-01
5.21459937e-01 -1.31911650e-01 -3.97905022e-01 -7.47593120e-02
-5.68493545e-01 -2.35553071e-01 2.81666547e-01 8.52779448e-01
4.07702655e-01 -6.55485213e-01 6.96169555e-01 -8.15840900e-01
1.00114405e+00 -5.33692718e-01 -3.55762124e-01 9.18591678e-01
-3.20732474e-01 4.55485404e-01 3.28088760e-01 -1.43528208e-01
-1.26502335e+00 -3.37091804e-01 -1.93766981e-01 -3.09703410e-01
-7.17885494e-02 3.72401834e-01 3.89030986e-02 2.54778266e-01
8.75886321e-01 2.56468430e-02 -3.81908923e-01 -4.16531920e-01
1.05875778e+00 5.98600328e-01 7.87079632e-01 -6.55171275e-01
5.80855370e-01 3.92811775e-01 -4.81284380e-01 -5.49563766e-01
-1.54088247e+00 -7.95893490e-01 -5.01113892e-01 -2.40656704e-01
9.97717261e-01 -5.33880830e-01 -6.66783810e-01 -1.57275259e-01
-1.64807153e+00 2.71129310e-01 -4.26473677e-01 1.06607951e-01
-4.84763175e-01 5.36680639e-01 -5.88461459e-01 -8.31244111e-01
-5.17028511e-01 -8.67073655e-01 1.02328575e+00 3.44188243e-01
-4.17684257e-01 -9.72984672e-01 3.90726298e-01 1.29495001e+00
2.95462996e-01 -1.70737565e-01 1.17995822e+00 -9.32017207e-01
-6.43696666e-01 2.10584234e-02 -4.62737024e-01 1.99515611e-01
-3.86376351e-01 -1.09979488e-01 -1.11112094e+00 9.42390859e-02
-8.55664313e-02 -6.24937594e-01 1.32919919e+00 2.32698753e-01
9.21140432e-01 -4.40304339e-01 -6.25523403e-02 1.01238765e-01
1.28342927e+00 -6.82182610e-02 5.03455520e-01 2.06249133e-01
6.95348561e-01 1.07844663e+00 5.35670340e-01 -7.94282332e-02
6.67866766e-01 6.20439649e-01 5.18497348e-01 2.81100929e-01
-5.39729834e-01 -2.15629503e-01 2.14787558e-01 7.47197390e-01
2.01459870e-01 -4.70183134e-01 -1.15229285e+00 1.21404982e+00
-1.98146868e+00 -1.09311175e+00 -6.21183276e-01 1.66814232e+00
9.13192987e-01 2.21028011e-02 2.11736158e-01 -2.54889995e-01
4.35377538e-01 1.82832673e-01 -2.95559734e-01 -5.95682144e-01
-3.21316868e-01 3.26097101e-01 1.18024811e-01 9.26480651e-01
-9.81154978e-01 8.13979506e-01 6.97973728e+00 9.85990703e-01
-4.02149670e-02 3.55370492e-01 2.80630231e-01 -1.98137030e-01
-1.02325630e+00 5.30223139e-02 -9.21554804e-01 1.24177285e-01
9.53042626e-01 -2.68765874e-02 2.01230794e-01 4.63510007e-01
-5.22394359e-01 2.30825674e-02 -1.29708946e+00 8.37162971e-01
5.71103990e-01 -1.69469535e+00 3.48944873e-01 -3.72898638e-01
7.43163168e-01 -1.56572342e-01 2.52839357e-01 4.10440475e-01
4.42947477e-01 -1.11712253e+00 7.10868895e-01 8.13861728e-01
4.02178973e-01 -7.17298150e-01 9.02304113e-01 4.43257987e-01
-6.74239695e-01 -2.60200441e-01 -1.87009260e-01 4.86249141e-02
2.61967450e-01 3.01955789e-01 -8.79842103e-01 6.34396911e-01
7.83954382e-01 7.13759065e-02 -8.34073484e-01 1.00940239e+00
-3.32986981e-01 6.93888664e-01 -1.45930931e-01 -3.90528411e-01
5.92414737e-01 1.34172171e-01 7.52067626e-01 1.31945622e+00
6.54778033e-02 2.42145076e-01 -1.70434713e-01 9.36011791e-01
-2.08028108e-01 -1.76525414e-02 -2.96521068e-01 1.24116115e-01
4.28853631e-01 1.09825504e+00 -1.49318472e-01 -4.25398111e-01
-4.67595845e-01 9.68162596e-01 6.04904771e-01 2.49778047e-01
-6.65368617e-01 -3.58418733e-01 2.65498549e-01 -3.17630589e-01
5.87585926e-01 1.74596207e-03 -3.88628006e-01 -9.15094256e-01
1.96742624e-01 -1.14404881e+00 8.85236800e-01 -1.00634849e+00
-1.51484382e+00 4.62035060e-01 -6.08932646e-03 -5.93965650e-01
-3.73371720e-01 -6.69866443e-01 -4.31908965e-01 8.57636988e-01
-1.65452278e+00 -1.30230749e+00 -1.51999697e-01 3.66011053e-01
8.04656267e-01 -2.11848497e-01 9.12336230e-01 -6.50816783e-03
-1.98392019e-01 6.04009628e-01 3.05652171e-01 -5.46231084e-02
5.37842691e-01 -1.73538423e+00 4.20389146e-01 7.26813436e-01
8.26940358e-01 4.06313688e-01 8.14988256e-01 -3.76862973e-01
-1.28198040e+00 -6.33046687e-01 1.26144695e+00 -1.44577026e+00
7.08892822e-01 4.89694774e-02 -1.02199972e+00 4.53098863e-01
7.12841511e-01 -4.53925490e-01 6.44139290e-01 3.99050564e-01
-6.67850018e-01 -2.79228995e-03 -9.50068235e-01 4.93571401e-01
6.64212942e-01 -9.03386712e-01 -1.43883872e+00 6.28356159e-01
9.68921185e-01 -3.60348642e-01 -6.48152888e-01 2.69694805e-01
3.85911018e-01 -9.69814301e-01 1.22377491e+00 -1.06519032e+00
6.02018118e-01 -2.14659572e-01 -3.76466900e-01 -1.04898119e+00
-3.58823091e-01 -4.29601997e-01 -6.57131314e-01 1.26539600e+00
6.95830762e-01 1.49342835e-01 7.17915833e-01 4.39850479e-01
-5.98886386e-02 -6.93003654e-01 -1.11810350e+00 -6.39280736e-01
3.64985108e-01 -4.50337946e-01 3.60874861e-01 5.40641606e-01
-6.01462759e-02 9.10120130e-01 -1.80259738e-02 2.33205587e-01
5.23203552e-01 2.81360835e-01 3.48296851e-01 -1.15496480e+00
-4.52994466e-01 -5.23890972e-01 1.88993603e-01 -1.30752385e+00
2.41305411e-01 -8.78916383e-01 -4.28735316e-02 -2.05904245e+00
4.40270931e-01 6.92126006e-02 -1.43784076e-01 2.19194487e-01
-6.44858301e-01 -4.75970237e-03 4.00442034e-01 1.27804592e-01
-1.21681345e+00 4.37396228e-01 1.08343637e+00 -4.14898366e-01
9.61117744e-02 -1.82125330e-01 -9.24156010e-01 7.44258046e-01
5.66024363e-01 -4.93165851e-01 -4.69725162e-01 -1.08675301e+00
5.59127927e-01 3.16382110e-01 4.84550923e-01 -5.86545050e-01
4.84715670e-01 1.15882479e-01 1.49550578e-02 -8.65355670e-01
5.58423519e-01 -3.60713065e-01 -7.26155639e-01 1.58584863e-01
-8.55619907e-01 1.48216918e-01 1.79083243e-01 7.99636841e-01
-2.32889265e-01 -7.10620403e-01 2.09326565e-01 -3.05873781e-01
-6.67355478e-01 -2.43389323e-01 -6.55666068e-02 7.34851837e-01
7.01828063e-01 -1.42307818e-01 -8.85186732e-01 -7.62361884e-01
-5.24386823e-01 5.95528662e-01 -1.26283869e-01 5.49841344e-01
8.34381104e-01 -1.02753830e+00 -1.21857643e+00 -7.31999099e-01
3.12737942e-01 -4.17614728e-01 3.23262870e-01 5.55062234e-01
-4.02135760e-01 7.45808184e-01 1.00244991e-01 -5.89024186e-01
-1.46068048e+00 3.89469117e-01 2.90718615e-01 -6.23795092e-01
-6.88705891e-02 1.31695402e+00 -8.33079666e-02 -6.89977765e-01
2.53838807e-01 -3.43369804e-02 -7.35409975e-01 4.17376041e-01
6.08773053e-01 7.13652372e-01 2.13996187e-01 -2.88014531e-01
-4.32773173e-01 6.13818824e-01 -2.72272736e-01 -4.13159460e-01
1.16609621e+00 -1.77140027e-01 -8.77761692e-02 3.46369028e-01
1.04389191e+00 7.45910183e-02 -9.36153531e-01 -3.22762609e-01
2.50819415e-01 -2.47205228e-01 7.18994439e-02 -1.22535872e+00
-4.71021503e-01 1.14976370e+00 4.56579119e-01 3.44443113e-01
7.21352220e-01 5.86385787e-01 8.61018479e-01 7.26641417e-01
-1.51657864e-01 -9.36754465e-01 5.06060362e-01 7.77260125e-01
1.14323056e+00 -1.15035915e+00 -1.38653636e-01 -2.25320876e-01
-7.77652502e-01 1.05989087e+00 5.81767619e-01 2.52318606e-02
-2.20864192e-02 -1.43151581e-01 5.46733439e-02 -6.29037440e-01
-9.73563552e-01 -4.70223188e-01 9.01188612e-01 6.80351436e-01
2.93903470e-01 -9.42924991e-02 -1.05774626e-01 7.36852288e-01
-2.50894427e-01 -4.19259071e-01 3.46671343e-01 6.01160228e-01
-5.21897972e-01 -7.63320684e-01 -4.46963012e-01 4.09292459e-01
-6.70872152e-01 -3.07950497e-01 -7.47697830e-01 6.01154864e-01
-8.84205401e-02 1.37786913e+00 5.57602048e-02 -1.54273137e-01
4.71511692e-01 3.53462309e-01 6.55553401e-01 -5.61756670e-01
-8.16927075e-01 -4.46178854e-01 7.35427916e-01 -2.74558306e-01
-3.38677526e-01 -4.96906698e-01 -9.52700257e-01 -9.66493338e-02
-5.00669360e-01 3.42844754e-01 3.52285713e-01 1.13776827e+00
4.48764294e-01 4.68385518e-01 1.15950473e-01 -2.41499282e-02
-1.00276470e+00 -8.06602418e-01 5.84424138e-02 2.99084634e-01
5.24528146e-01 -1.73687145e-01 -3.08681279e-01 3.82596557e-03] | [11.254068374633789, 8.001913070678711] |
f2f8a0c7-3709-4931-bf6b-1f66b146b29c | towards-provably-secure-encrypted-control | 2210.08849 | null | https://arxiv.org/abs/2210.08849v1 | https://arxiv.org/pdf/2210.08849v1.pdf | Towards Provably Secure Encrypted Control Using Homomorphic Encryption | Encrypted control is a promising method for the secure outsourcing of controller computation to a public cloud. However, a feasible method for security proofs of control has not yet been developed in the field of encrypted control systems. Additionally, cryptography does not consider certain types of attacks on encrypted control systems; therefore, the security of such a system cannot be guaranteed using a secure cryptosystem. This study proposes a novel security definition for encrypted control under attack for control systems using cryptography. It applies the concept of provable security, which is the security of cryptosystems based on mathematical proofs, to encrypted control systems. Furthermore, this study analyzes the relation between the proposed security and the conventional security of cryptosystems. The results of the analysis demonstrated that the security of an encrypted control system can be enhanced by employing secure homomorphic encryption. | ['Kiminao Kogiso', 'Kaoru Teranishi'] | 2022-10-17 | null | null | null | null | ['mathematical-proofs'] | ['miscellaneous'] | [ 3.54645044e-01 -1.92126706e-01 -3.64469923e-02 -1.03125505e-01
2.72078902e-01 -1.20044887e+00 8.55627775e-01 1.14279516e-01
-3.73027653e-01 7.61181176e-01 -5.49464941e-01 -7.91698337e-01
-3.70014399e-01 -1.28237891e+00 -4.86707389e-01 -8.87466967e-01
-2.51284122e-01 -3.51539493e-01 1.14826620e-01 -4.82321650e-01
3.89887720e-01 5.70271850e-01 -1.60836375e+00 2.14336216e-01
4.58360612e-01 1.18818593e+00 -2.61469930e-01 4.83905375e-01
1.84776172e-01 4.82180595e-01 -7.31159151e-01 -3.62955272e-01
6.29858315e-01 -1.65076897e-01 -5.62842071e-01 -2.25981519e-01
-5.85808679e-02 -5.00075638e-01 -1.54989481e-01 1.30050528e+00
-1.72102898e-01 -5.94021618e-01 2.01194555e-01 -2.19376445e+00
-3.44423234e-01 1.83787316e-01 3.16249460e-01 -3.51572633e-01
1.87230080e-01 1.11429513e-01 7.17616558e-01 -1.11322172e-01
5.85835457e-01 6.90621376e-01 6.51739091e-02 5.00680029e-01
-1.07200050e+00 -1.15430379e+00 -1.01005891e-02 -5.27031235e-02
-1.58450985e+00 2.19342381e-01 2.58343101e-01 -1.23547956e-01
7.07217455e-01 7.83164501e-01 9.66255367e-01 2.77417332e-01
1.06483495e+00 -1.32037655e-01 1.77386367e+00 -4.92565006e-01
2.59667695e-01 8.47962439e-01 1.99822575e-01 2.06725106e-01
1.25435579e+00 6.29534662e-01 1.78619266e-01 -5.56392908e-01
3.66614670e-01 2.62688428e-01 -7.42613912e-01 -2.69922554e-01
-1.07968760e+00 7.65563130e-01 8.82240981e-02 5.12380183e-01
7.88677186e-02 2.23484620e-01 6.03515327e-01 9.41355348e-01
1.57240883e-01 6.75308347e-01 -7.40086317e-01 3.68668973e-01
-2.08583325e-01 5.08285344e-01 1.13100255e+00 1.15120649e+00
3.36371452e-01 -1.81193873e-02 2.99276501e-01 -8.08579206e-01
4.45593774e-01 7.99113810e-01 -1.83893993e-01 -8.75032902e-01
2.03587309e-01 7.06051171e-01 2.91357011e-01 -8.77239347e-01
3.60362053e-01 3.59724015e-01 -7.00152755e-01 8.77200305e-01
1.88638762e-01 -3.94614249e-01 -1.61835104e-01 1.48243487e+00
3.84148210e-01 -1.04618944e-01 8.13996911e-01 5.51479816e-01
-3.87221463e-02 9.95646298e-01 -2.08593443e-01 -5.49469531e-01
1.25731301e+00 -3.64801645e-01 -1.18715227e+00 7.57945299e-01
1.68224409e-01 -7.83913672e-01 3.32048178e-01 4.85151738e-01
-8.67511988e-01 -3.43904607e-02 -1.26753843e+00 4.78038549e-01
-1.08270395e+00 -5.40501535e-01 8.20540488e-01 1.15492415e+00
-1.02539241e+00 2.91964710e-01 -3.73466045e-01 -6.14956487e-03
-2.02514738e-01 1.14558756e+00 -6.78317666e-01 2.92555094e-01
-1.46382403e+00 9.84536946e-01 6.56774342e-01 -1.25041269e-02
-8.10061276e-01 -5.93538463e-01 -7.57264853e-01 2.13563442e-01
3.92945200e-01 -4.62154746e-01 9.35500562e-01 -7.64777601e-01
-1.33750749e+00 4.37608272e-01 3.11461955e-01 -5.69265962e-01
2.25129575e-01 3.19296926e-01 -4.31027621e-01 3.16560030e-01
-2.84375131e-01 -3.73777039e-02 7.81726599e-01 -1.16672564e+00
-8.84437919e-01 -3.40146959e-01 8.48640978e-01 -4.48368490e-01
-2.73254246e-01 2.86445439e-01 7.28298604e-01 -3.13269079e-01
-2.56711215e-01 -1.21372807e+00 -2.38783002e-01 1.95929468e-01
3.28897871e-02 2.97614813e-01 1.42469895e+00 -9.72885191e-02
1.07316017e+00 -2.14318371e+00 -4.95295882e-01 7.44033337e-01
1.58518299e-01 6.48904085e-01 3.58599693e-01 1.06915009e+00
-1.12269089e-01 8.04099739e-01 -2.33620927e-02 6.87484622e-01
4.66344178e-01 1.20839059e-01 -8.01664591e-01 6.60106599e-01
-3.02181151e-02 4.07999247e-01 -7.02394009e-01 -3.20519358e-01
4.26960826e-01 7.06996322e-01 -4.71385449e-01 2.45074674e-01
8.12293962e-03 3.19455385e-01 -7.69634306e-01 2.48266980e-01
1.19777334e+00 1.88946098e-01 7.95480907e-01 1.26268759e-01
-9.19225752e-01 3.01891956e-02 -1.22753322e+00 4.87748146e-01
-5.28755069e-01 2.82357693e-01 5.45207381e-01 -3.34648281e-01
5.65784872e-01 1.19827127e+00 5.97908437e-01 1.46815792e-01
5.69840968e-01 2.47585386e-01 1.20046638e-01 -2.75727421e-01
2.29220212e-01 -6.85784146e-02 -2.68856823e-01 7.54963398e-01
-4.47297752e-01 -6.13569260e-01 4.26052287e-02 -1.11165516e-01
7.55450428e-01 -1.26161143e-01 4.02685970e-01 -5.29993713e-01
1.20558751e+00 -9.20244977e-02 4.40051794e-01 2.28933081e-01
-6.33841008e-02 -4.62368667e-01 3.47693473e-01 -5.82225919e-01
-9.19147372e-01 -7.70974874e-01 -1.07371204e-01 8.11829604e-03
4.20447886e-01 -9.85943437e-01 -9.07436490e-01 -6.64033711e-01
2.51508504e-01 1.77084818e-01 2.50146743e-02 -2.04551026e-01
-2.71406412e-01 -2.35657424e-01 8.34806263e-01 -3.42753641e-02
1.04255676e+00 -6.49379432e-01 -6.70335948e-01 -2.92169526e-02
3.37325037e-01 -1.34484446e+00 -1.44003004e-01 -1.44859388e-01
-6.62014306e-01 -1.56042349e+00 2.00311571e-01 -6.83000207e-01
1.09818733e+00 4.66291904e-01 3.14148277e-01 9.17816699e-01
5.27036376e-02 4.32231665e-01 -1.38307795e-01 -9.71610010e-01
-8.13189745e-01 -3.03867996e-01 4.17488545e-01 6.30859733e-02
1.88579008e-01 -1.69135198e-01 -4.30401593e-01 1.69812649e-01
-1.28730917e+00 -2.54923731e-01 1.64702252e-01 5.48214138e-01
2.47614980e-01 1.06204963e+00 3.37376207e-01 -6.00592673e-01
7.96505034e-01 9.61444154e-02 -1.62786090e+00 4.42191839e-01
-1.37806594e+00 -1.33250533e-02 9.74663615e-01 -1.71044141e-01
-7.94191003e-01 4.22530575e-03 3.22243392e-01 -5.95168322e-02
-4.05393727e-02 9.83589143e-02 -4.26518321e-01 -8.26634228e-01
-3.36479187e-01 2.83328772e-01 3.46739560e-01 9.27156061e-02
-7.08232597e-02 8.74462068e-01 -1.79244861e-01 -5.65557659e-01
1.25108230e+00 7.11043000e-01 7.38644063e-01 -5.47887743e-01
1.91361144e-01 -4.71163280e-02 -3.38579118e-01 -3.24824918e-03
1.03785932e+00 -5.79180360e-01 -1.83150828e+00 6.83182299e-01
-1.35342348e+00 2.36337736e-01 2.45864112e-02 7.23127186e-01
-3.31795692e-01 3.53065908e-01 -6.15312338e-01 -1.16474617e+00
-3.96843314e-01 -1.23895133e+00 6.19520962e-01 -8.91840234e-02
4.61435057e-02 -8.75065804e-01 -1.49452463e-01 9.72418860e-02
5.76407969e-01 5.34270346e-01 9.27174568e-01 -3.82312953e-01
-1.20315647e+00 -8.01062644e-01 1.58593655e-01 7.64510274e-01
4.02271360e-01 4.27360445e-01 -8.73381257e-01 -6.46019518e-01
6.33637667e-01 4.93036769e-02 -3.50158662e-01 -1.84367403e-01
9.39510286e-01 -9.53119934e-01 -5.00778072e-02 3.45433563e-01
2.01356220e+00 6.21256173e-01 9.09678161e-01 2.68223554e-01
-1.01372235e-01 6.24185920e-01 8.38907242e-01 3.42404574e-01
1.29103601e-01 4.13601875e-01 4.85457271e-01 1.39717339e-02
9.73997533e-01 -1.79924499e-02 4.69083339e-01 5.92891753e-01
-4.50174868e-01 -2.69795842e-02 -3.58256429e-01 -1.05340570e-01
-1.41550934e+00 -1.09582043e+00 -4.13992912e-01 2.34911013e+00
5.98446846e-01 -4.41014953e-02 -6.04235888e-01 6.92128003e-01
8.47131193e-01 -4.55115400e-02 4.54582661e-01 -1.01254857e+00
2.47681931e-01 7.12935507e-01 9.21162426e-01 6.33204877e-01
-1.01135564e+00 7.46372163e-01 6.15707064e+00 1.89264849e-01
-1.42462134e+00 2.07919657e-01 -1.63987264e-01 5.56651354e-01
-7.57268146e-02 7.95239389e-01 -5.84385812e-01 5.43024242e-01
9.19460893e-01 -9.66555476e-01 5.55907369e-01 5.72659016e-01
3.69844377e-01 6.57922402e-02 -8.86565626e-01 2.22700045e-01
-4.98157054e-01 -1.28508472e+00 1.13082498e-01 6.11928344e-01
8.62775683e-01 -8.97473693e-01 8.41249004e-02 -1.71922892e-02
-2.66215861e-01 -7.30092525e-01 4.67672229e-01 1.81162760e-01
3.69771242e-01 -9.20258582e-01 1.10935712e+00 3.56109411e-01
-1.26614881e+00 -1.88480198e-01 -1.54725820e-01 -5.01578093e-01
-6.99878931e-02 -2.34517768e-01 -5.21234334e-01 8.43506873e-01
3.93443018e-01 -1.38967082e-01 2.86692381e-03 5.55947244e-01
-4.28084046e-01 2.17487663e-01 -2.15729401e-01 -2.31956095e-01
2.80345768e-01 -4.36643034e-01 2.17249513e-01 6.68007791e-01
3.30296010e-01 5.60865402e-01 3.41663584e-02 8.57438445e-01
2.78737426e-01 -8.44944268e-02 -1.27347863e+00 -2.27914929e-01
4.56336051e-01 9.45832431e-01 -4.78068262e-01 -2.84905404e-01
-4.09677804e-01 5.22195101e-01 -1.00501025e+00 1.65067151e-01
-7.24111557e-01 -6.54552221e-01 7.64164865e-01 3.28211218e-01
1.57948777e-01 -4.77932125e-01 -6.34986609e-02 -9.77047861e-01
4.18760292e-02 -1.01332700e+00 1.59623399e-01 -2.35698149e-01
-8.54458094e-01 4.29234475e-01 4.40093428e-01 -1.61250508e+00
-4.25419621e-02 -6.21433198e-01 -5.96559048e-01 8.21920753e-01
-1.70643473e+00 -1.28252256e+00 1.16191626e-01 1.10697699e+00
-4.02829081e-01 -1.46484897e-01 1.34965730e+00 2.69483835e-01
6.60802498e-02 2.38705873e-01 -2.35659450e-01 -2.65860856e-01
5.31042337e-01 -1.05427718e+00 7.28755370e-02 9.94193912e-01
-6.48149610e-01 1.18669796e+00 6.60044312e-01 -6.20152771e-01
-1.92492533e+00 -9.35011446e-01 1.41561627e+00 -1.27216697e-01
6.18502676e-01 -6.89593792e-01 -4.61570799e-01 7.24044800e-01
7.87036180e-01 1.02605097e-01 6.57680809e-01 -7.02827632e-01
-1.77215040e-01 -5.22312284e-01 -1.60656297e+00 3.68590862e-01
2.03295991e-01 -1.00248456e+00 -4.32268083e-01 3.19125116e-01
1.01617908e+00 -2.83356453e-03 -1.11094344e+00 3.53010654e-01
7.47987807e-01 -5.78229427e-01 5.32300532e-01 -5.20745158e-01
-1.60844266e-01 -8.70912850e-01 -3.23758543e-01 -6.10461414e-01
2.87530452e-01 -9.79865491e-01 1.72891110e-01 9.53484237e-01
1.55657396e-01 -1.39871633e+00 3.65636528e-01 9.60285246e-01
4.79827225e-01 -3.53299230e-01 -7.28953004e-01 -1.07102346e+00
1.09029934e-01 -2.07174018e-01 1.27825415e+00 1.21655738e+00
3.05521160e-01 -4.08776611e-01 -6.55177459e-02 7.86829472e-01
5.99256873e-01 1.87675461e-01 8.60750318e-01 -1.05631399e+00
2.01770455e-01 1.34666726e-01 -7.03883171e-01 1.58931568e-01
4.80485380e-01 -7.35504508e-01 -5.12358308e-01 -1.01052284e+00
-4.33984876e-01 -4.99282092e-01 -2.94491619e-01 2.50915498e-01
4.13781583e-01 1.09632410e-01 7.66530454e-01 5.84552959e-02
9.26284939e-02 2.51823992e-01 1.28631914e+00 -2.67523676e-01
1.05151627e-02 3.06667268e-01 -4.77821529e-01 3.28402728e-01
1.01088238e+00 -6.16326213e-01 -9.22935009e-01 3.51762414e-01
1.54781818e-01 1.12852780e-02 5.06898880e-01 -8.85766745e-01
2.50720829e-01 -7.60037601e-01 -6.12020493e-01 -1.27396241e-01
3.64248380e-02 -1.83213913e+00 7.04432189e-01 1.18496001e+00
1.06674723e-01 6.83656573e-01 2.81894691e-02 8.93720016e-02
-4.66966182e-01 -2.39997953e-01 3.93110335e-01 -8.40999335e-02
-5.23721814e-01 1.75174385e-01 -7.90239155e-01 -8.22434783e-01
1.71670592e+00 -2.12453172e-01 -5.02845645e-01 -3.15826625e-01
-4.45821464e-01 2.11658195e-01 7.70457506e-01 3.17313299e-02
5.92254162e-01 -1.09959555e+00 -2.23631486e-01 4.62616146e-01
-1.62237197e-01 -5.43116152e-01 -2.80062765e-01 4.90891606e-01
-9.37667549e-01 8.43513608e-01 -5.10726571e-01 -9.54069272e-02
-1.67814291e+00 9.92830873e-01 3.68584335e-01 -2.94206768e-01
-2.62418807e-01 -4.23693180e-01 1.38440281e-01 -2.74717361e-01
5.98047338e-02 -5.56279719e-01 5.19474000e-02 -6.42507553e-01
6.86583638e-01 1.48129538e-01 -1.56450823e-01 -4.01850730e-01
-3.94781351e-01 4.79705721e-01 3.12501699e-01 -3.95047754e-01
8.81842554e-01 -2.12698609e-01 -1.10180926e+00 -1.33818850e-01
1.04546177e+00 2.76210010e-01 -2.17927769e-01 7.23865628e-01
-2.57762969e-01 -7.68374920e-01 3.42748649e-02 -3.78122687e-01
-1.01824188e+00 5.89405775e-01 6.28391743e-01 6.10717952e-01
1.20032918e+00 -1.01498914e+00 6.79392040e-01 6.90833569e-01
1.00160956e+00 -1.08310795e+00 -7.32299149e-01 4.64804351e-01
5.95366597e-01 -8.24052095e-01 4.36414517e-02 -1.20535541e+00
-2.03378424e-01 1.20001554e+00 4.97574300e-01 -2.40693957e-01
1.16698706e+00 7.31420040e-01 -2.50311778e-03 -2.46584773e-01
-9.46188509e-01 9.65589285e-02 -4.46059614e-01 6.50267065e-01
9.35362875e-02 2.11786434e-01 -1.17436218e+00 5.10448873e-01
-1.44763365e-02 5.18307030e-01 1.06276035e+00 1.65837872e+00
1.39399245e-01 -1.85546064e+00 -6.24476075e-01 -3.70318830e-01
-9.72798347e-01 1.13055883e-02 -5.06537676e-01 1.38660777e+00
6.98080599e-01 1.10101426e+00 -1.58305109e-01 -5.85654020e-01
2.60148853e-01 -2.56548077e-01 3.95883769e-01 -4.37913090e-01
-1.12587583e+00 -4.49566036e-01 -1.63203012e-02 -2.86444962e-01
-8.57675731e-01 -1.61995023e-01 -1.29164672e+00 -9.27957952e-01
-7.74593651e-01 7.38888085e-01 9.06348348e-01 6.60282552e-01
5.78795522e-02 1.26878396e-01 1.19919372e+00 -1.48721039e-01
-6.33576989e-01 -1.25216935e-02 -8.21448445e-01 -4.12793569e-02
2.74809033e-01 -2.40766495e-01 -5.87840319e-01 6.61161616e-02] | [5.718376636505127, 4.861716270446777] |
01285115-c8aa-4b11-96a4-60111e03ca76 | petsgan-rethinking-priors-for-single-image | 2203.01488 | null | https://arxiv.org/abs/2203.01488v1 | https://arxiv.org/pdf/2203.01488v1.pdf | PetsGAN: Rethinking Priors for Single Image Generation | Single image generation (SIG), described as generating diverse samples that have similar visual content with the given single image, is first introduced by SinGAN which builds a pyramid of GANs to progressively learn the internal patch distribution of the single image. It also shows great potentials in a wide range of image manipulation tasks. However, the paradigm of SinGAN has limitations in terms of generation quality and training time. Firstly, due to the lack of high-level information, SinGAN cannot handle the object images well as it does on the scene and texture images. Secondly, the separate progressive training scheme is time-consuming and easy to cause artifact accumulation. To tackle these problems, in this paper, we dig into the SIG problem and improve SinGAN by fully-utilization of internal and external priors. The main contributions of this paper include: 1) We introduce to SIG a regularized latent variable model. To the best of our knowledge, it is the first time to give a clear formulation and optimization goal of SIG, and all the existing methods for SIG can be regarded as special cases of this model. 2) We design a novel Prior-based end-to-end training GAN (PetsGAN) to overcome the problems of SinGAN. Our method gets rid of the time-consuming progressive training scheme and can be trained end-to-end. 3) We construct abundant qualitative and quantitative experiments to show the superiority of our method on both generated image quality, diversity, and the training speed. Moreover, we apply our method to other image manipulation tasks (e.g., style transfer, harmonization), and the results further prove the effectiveness and efficiency of our method. | ['BoWen Zhou', 'Tiande Guo', 'Hailin Shi', 'Congying Han', 'Yinglu Liu', 'ZiCheng Zhang'] | 2022-03-03 | null | null | null | null | ['single-image-generation'] | ['computer-vision'] | [ 4.12369817e-01 9.02436078e-02 1.04115695e-01 7.45441169e-02
-6.32733762e-01 -3.36181432e-01 5.32260835e-01 -7.50513494e-01
4.92506847e-02 8.63928735e-01 -6.49545416e-02 1.58137366e-01
-8.44952986e-02 -8.94884050e-01 -7.56603837e-01 -9.87836540e-01
5.07863045e-01 1.64314523e-01 1.37279466e-01 -2.79628545e-01
-5.71031198e-02 3.08305234e-01 -1.54226518e+00 -2.21568439e-02
1.27996135e+00 9.26127017e-01 5.93546927e-01 2.59009033e-01
3.05405334e-02 6.48396850e-01 -6.74052000e-01 -4.24055934e-01
2.46275276e-01 -1.06963193e+00 -5.98369658e-01 5.26819050e-01
2.33897999e-01 -4.09589052e-01 -1.19657740e-01 9.55248892e-01
7.42980361e-01 4.27519791e-02 5.43705046e-01 -1.34119928e+00
-8.73058021e-01 3.21863115e-01 -7.57531524e-01 -4.62152511e-01
8.34634379e-02 2.68317848e-01 6.06532037e-01 -8.26218784e-01
6.63498342e-01 1.07161605e+00 4.70572412e-01 7.14048147e-01
-1.21758294e+00 -6.13757133e-01 1.49646923e-01 -3.47137712e-02
-1.15745533e+00 -3.44508678e-01 1.06219172e+00 -3.85633796e-01
1.71921268e-01 2.43498579e-01 7.11135268e-01 1.26378572e+00
-6.71617910e-02 1.07016766e+00 1.32979500e+00 -5.47972202e-01
2.06399888e-01 -5.91724515e-02 -4.85742539e-01 6.69466496e-01
9.03353188e-03 1.26686350e-01 -4.85328645e-01 1.80438310e-01
1.12884045e+00 -1.30680293e-01 -5.31163692e-01 -3.96688074e-01
-1.16516542e+00 7.04824686e-01 4.79866534e-01 2.88201720e-01
-3.99394870e-01 2.12428585e-01 1.93865467e-02 6.35625795e-02
4.63823587e-01 3.92126143e-01 -3.39256264e-02 -5.35994731e-02
-1.02454007e+00 1.44168526e-01 3.98758262e-01 1.15818894e+00
7.92600155e-01 3.96537900e-01 -3.88145387e-01 9.91716623e-01
1.01886243e-01 5.21023273e-01 5.13754308e-01 -1.01600134e+00
1.02914296e-01 2.81189919e-01 6.60969168e-02 -8.70866001e-01
6.96825683e-02 -5.39623678e-01 -1.12371027e+00 3.32953304e-01
2.57497281e-01 -2.11616814e-01 -1.01824915e+00 1.87681198e+00
2.88970798e-01 1.42217308e-01 -1.95650458e-01 8.70392501e-01
8.80897343e-01 7.17777491e-01 -1.29883379e-01 -4.14207608e-01
1.10206199e+00 -1.27500856e+00 -9.79022861e-01 -5.32191917e-02
9.81042981e-02 -9.59425747e-01 1.19044852e+00 3.89404744e-01
-1.43135035e+00 -7.43831575e-01 -9.03514624e-01 1.56214600e-02
7.99045637e-02 3.05591255e-01 6.85461044e-01 4.81929511e-01
-1.11870492e+00 4.07407045e-01 -8.08552444e-01 -2.50148147e-01
5.43224275e-01 8.97198766e-02 -9.35433805e-02 -1.55253679e-01
-9.74894166e-01 5.87948024e-01 2.35277489e-01 2.52152383e-01
-8.84614646e-01 -6.54445112e-01 -8.24956596e-01 -6.04455583e-02
5.65270245e-01 -1.16060901e+00 1.10434747e+00 -1.28860593e+00
-2.08143568e+00 5.84772110e-01 -2.65815526e-01 9.45465267e-02
6.70399368e-01 -2.26585478e-01 -3.37466113e-02 -9.82784014e-03
1.64548352e-01 7.89390504e-01 9.70290542e-01 -1.65935767e+00
-3.06615800e-01 -3.95304337e-02 7.41139576e-02 1.49892896e-01
-3.59709710e-01 -1.64163515e-01 -8.42988491e-01 -1.09831643e+00
8.60350393e-03 -8.79071712e-01 -2.01335549e-01 -1.03243878e-02
-4.29211795e-01 9.74857807e-02 8.26283813e-01 -6.18392348e-01
1.14405859e+00 -2.25937152e+00 5.69578707e-01 -5.77050373e-02
1.93935990e-01 4.01362419e-01 -2.37417787e-01 4.87403244e-01
1.59109548e-01 9.00315419e-02 -5.47344625e-01 -4.87497568e-01
2.92353053e-02 2.60251194e-01 -3.61475587e-01 1.15619399e-01
1.90987468e-01 1.21330941e+00 -9.10311401e-01 -4.98911768e-01
2.25861549e-01 7.00241923e-01 -5.13110042e-01 4.48479503e-01
-2.84411877e-01 7.77176738e-01 -4.20660764e-01 5.41467249e-01
7.49926150e-01 -2.72570521e-01 -7.39726499e-02 -3.85277003e-01
-1.13220084e-02 -1.82299539e-01 -1.07831168e+00 1.78140104e+00
-5.11188030e-01 3.33094597e-01 1.13193750e-01 -8.75393033e-01
8.83833587e-01 2.62853295e-01 2.57180810e-01 -7.13632643e-01
-1.21150147e-02 2.75776356e-01 -3.03330421e-01 -4.43266749e-01
3.68331283e-01 -2.45688170e-01 9.50086787e-02 2.89044470e-01
9.26975161e-02 -3.37085158e-01 2.42501453e-01 4.37262841e-02
5.49376190e-01 5.61666310e-01 2.21392646e-01 -4.68319021e-02
5.13689935e-01 -3.24453622e-01 7.22441375e-01 6.26168072e-01
1.39726356e-01 9.80442286e-01 4.49097663e-01 -1.61462054e-01
-1.01233613e+00 -9.98846829e-01 1.76864527e-02 7.20341086e-01
3.03025812e-01 -3.45713258e-01 -8.68634939e-01 -5.52323401e-01
-4.00751263e-01 5.51158786e-01 -4.84347731e-01 -3.88858989e-02
-5.53833663e-01 -8.05514455e-01 2.74193823e-01 5.28671026e-01
9.50308144e-01 -1.24080575e+00 -4.16658849e-01 1.08349189e-01
-4.04380322e-01 -1.05625713e+00 -4.98767674e-01 -4.24111754e-01
-7.07271695e-01 -7.57575035e-01 -1.04130661e+00 -9.83021915e-01
7.59052813e-01 1.89337447e-01 1.09952974e+00 1.70954883e-01
-2.54800618e-01 3.25561076e-01 -4.37983423e-01 -3.86269093e-01
-2.90962338e-01 -4.21007872e-02 -3.44957560e-01 2.05542132e-01
-4.20498222e-01 -7.60650754e-01 -7.66183197e-01 6.03781462e-01
-1.15954792e+00 6.24117732e-01 7.79724598e-01 1.16695940e+00
8.63387585e-01 2.72254467e-01 5.78241169e-01 -9.31890309e-01
5.03413737e-01 -4.73825596e-02 -6.45663381e-01 2.48679474e-01
-4.30407077e-01 -1.89075321e-01 6.86030149e-01 -4.44097966e-01
-1.29579723e+00 -2.15501748e-02 -1.65841773e-01 -4.02160108e-01
-7.11210594e-02 4.56075937e-01 -5.21401405e-01 -1.55669779e-01
4.11305994e-01 4.79834586e-01 1.13458902e-01 -4.89837021e-01
4.87281382e-01 3.74294102e-01 5.43911457e-01 -6.30782664e-01
1.00473464e+00 5.69783747e-01 1.06521830e-01 -7.62622774e-01
-9.12501097e-01 7.53379539e-02 -3.91738296e-01 -2.31808394e-01
7.87411451e-01 -8.45694423e-01 -4.35526222e-01 9.34730351e-01
-9.77061629e-01 -6.90244138e-01 -3.93905222e-01 1.85706079e-01
-8.31786394e-01 5.73178589e-01 -6.91847742e-01 -6.78406239e-01
-3.28811318e-01 -1.21953976e+00 1.16409731e+00 2.42243484e-01
1.09622516e-01 -9.58779454e-01 -2.43838593e-01 3.82704020e-01
6.29820287e-01 6.32949591e-01 7.53431737e-01 3.56983393e-01
-8.38730097e-01 2.35178351e-01 -1.26899183e-01 4.54899728e-01
2.21509233e-01 -1.17046639e-01 -9.08156753e-01 -4.87472802e-01
2.15534359e-01 -4.30045873e-01 8.22315335e-01 5.20095706e-01
1.37000561e+00 -2.05184400e-01 -1.55578732e-01 8.34504902e-01
1.56868696e+00 2.42437571e-01 1.03124380e+00 3.36370438e-01
8.66967976e-01 5.13443291e-01 6.33428514e-01 3.13537240e-01
1.29007712e-01 9.90943432e-01 3.55406523e-01 -6.57108366e-01
-6.11053586e-01 -3.77295107e-01 3.32463473e-01 8.17797184e-01
-2.68938601e-01 -5.02483189e-01 -4.45752412e-01 4.98610049e-01
-1.94217288e+00 -9.27022219e-01 -1.14559615e-02 1.95299709e+00
7.32466877e-01 -2.67487645e-01 7.03592747e-02 5.95654808e-02
7.40208924e-01 6.18516244e-02 -4.27653462e-01 1.56620622e-01
-2.97553927e-01 2.14541882e-01 5.01939096e-02 3.42961490e-01
-8.62689853e-01 9.89475131e-01 5.75852776e+00 1.14579034e+00
-1.17745125e+00 6.80405051e-02 7.78790295e-01 -5.88902719e-02
-4.49437022e-01 -2.84733512e-02 -5.33361435e-01 7.38486767e-01
2.28393510e-01 -4.32530083e-02 6.84009433e-01 6.49533749e-01
1.51699483e-01 3.62903476e-02 -8.72764230e-01 1.13956714e+00
1.94345400e-01 -1.23094964e+00 1.68680832e-01 1.75474663e-04
1.10657358e+00 -4.75719154e-01 2.80735493e-01 1.96220785e-01
1.37465507e-01 -1.05453742e+00 8.31099987e-01 3.91809016e-01
1.01763713e+00 -7.03913271e-01 6.70381606e-01 3.17321837e-01
-9.51170862e-01 1.48310825e-01 -2.70030767e-01 1.95368662e-01
5.11797011e-01 7.80392170e-01 -1.60608351e-01 9.95363295e-01
5.58696687e-01 7.27862060e-01 -3.98831874e-01 1.01448882e+00
-7.13591516e-01 6.34342253e-01 -6.11515827e-02 3.59398365e-01
6.79879338e-02 -4.25578177e-01 5.20793855e-01 7.80422509e-01
5.31835258e-01 6.26682341e-02 1.36055037e-01 1.18260610e+00
-7.47383758e-02 2.01695897e-02 -4.80731696e-01 9.29322392e-02
2.23124906e-01 1.21447062e+00 -6.04457021e-01 -1.64395720e-01
-2.82265723e-01 1.30455983e+00 1.94986224e-01 6.17945552e-01
-8.50889683e-01 -3.74416947e-01 1.63206831e-01 1.31167769e-01
5.67202151e-01 -5.04172593e-02 -3.18134964e-01 -1.21248055e+00
2.58108854e-01 -9.90105033e-01 2.81581134e-02 -9.62757349e-01
-1.35247636e+00 7.58329809e-01 6.94288835e-02 -1.25101519e+00
-2.56387740e-01 -2.45697290e-01 -6.74408257e-01 8.67883086e-01
-1.59889400e+00 -1.45202208e+00 -6.44555807e-01 5.82588017e-01
5.47404230e-01 -1.02554001e-01 7.08150625e-01 3.38115305e-01
-6.31802380e-01 6.91535592e-01 1.65941373e-01 -5.15940562e-02
6.83555424e-01 -1.04278946e+00 1.83774948e-01 1.04941821e+00
-1.47948876e-01 4.55453038e-01 5.19404471e-01 -5.37610769e-01
-1.25065422e+00 -1.09974027e+00 6.34938836e-01 -1.87832847e-01
2.98731476e-01 -4.16631967e-01 -7.60573208e-01 5.40602326e-01
3.58857363e-01 -1.91853419e-01 5.14841437e-01 -8.51567835e-02
1.14005320e-01 -1.46728799e-01 -9.83217061e-01 7.77206481e-01
1.23462403e+00 -1.86074302e-01 -1.94805130e-01 1.63730070e-01
6.24295712e-01 -5.58981895e-01 -8.05778444e-01 5.87153316e-01
3.39300424e-01 -1.13616133e+00 8.33235919e-01 3.16586643e-02
7.71195590e-01 -5.89152753e-01 7.40072429e-02 -1.42526269e+00
-4.45595205e-01 -9.87953544e-01 1.51677318e-02 1.71772492e+00
2.86943484e-02 -5.97735465e-01 6.10323012e-01 2.98425496e-01
-2.76457965e-01 -9.20844853e-01 -6.53882682e-01 -8.54173243e-01
-1.50625259e-02 -1.21634588e-01 8.25160444e-01 7.75224626e-01
-4.83694434e-01 3.58837426e-01 -8.21464717e-01 -5.17474711e-02
7.04770446e-01 5.05595565e-01 1.06195605e+00 -8.43425632e-01
-6.43349051e-01 -3.45706731e-01 -1.51278242e-01 -1.29682481e+00
-9.13211778e-02 -5.86165786e-01 1.00213706e-01 -1.67773533e+00
3.10267955e-01 -4.89753604e-01 3.42866927e-02 3.70012671e-01
-4.28624362e-01 4.90608245e-01 4.50511307e-01 3.79930079e-01
-4.09267306e-01 1.05856681e+00 1.87867665e+00 1.65990330e-02
-1.07245766e-01 -1.16593868e-01 -8.64029586e-01 5.04831374e-01
7.38017738e-01 -1.51720881e-01 -7.34753191e-01 -6.24418437e-01
7.25731328e-02 1.26137882e-01 3.89713585e-01 -7.92489409e-01
-1.63279861e-01 -1.89156547e-01 3.46979737e-01 -3.78433317e-01
3.10572058e-01 -6.45568192e-01 4.66966301e-01 2.93080270e-01
-1.32799577e-02 -2.45379075e-01 8.49514455e-02 4.53224480e-01
-3.93653750e-01 -6.83488175e-02 9.89615142e-01 -9.31359679e-02
-5.06234348e-01 3.54762584e-01 -1.40224606e-01 -1.51982689e-02
1.04686987e+00 -2.93465853e-01 -2.02035427e-01 -6.01689935e-01
-3.64596486e-01 1.39498323e-01 7.95993447e-01 2.07832217e-01
4.94008690e-01 -1.57178569e+00 -7.13785231e-01 3.31889361e-01
-8.21225047e-02 4.78867978e-01 4.48982745e-01 7.83866525e-01
-4.04174805e-01 5.27400449e-02 -2.81810731e-01 -6.49196088e-01
-1.00344682e+00 6.64240718e-01 1.74647480e-01 -2.87443876e-01
-6.36428058e-01 6.86773002e-01 8.41732025e-01 -2.01289371e-01
1.13011273e-02 1.65422499e-01 2.80220364e-03 -3.01325619e-01
4.17028874e-01 1.83528766e-01 -2.59051055e-01 -4.95457768e-01
6.90798461e-02 7.97541857e-01 6.56362250e-02 -5.98011762e-02
1.38905656e+00 -1.76210508e-01 -3.06773961e-01 1.53951317e-01
8.34601223e-01 1.36033362e-02 -1.52694798e+00 -6.32568523e-02
-6.70297503e-01 -7.30608463e-01 -9.45830997e-03 -7.95360923e-01
-1.43260801e+00 7.79316783e-01 4.22045022e-01 -9.73225459e-02
1.58582687e+00 -6.83183968e-02 9.79112148e-01 -3.08378190e-01
3.46806616e-01 -9.10306275e-01 3.61452609e-01 2.18592569e-01
1.11771059e+00 -9.55045700e-01 -1.02269858e-01 -7.57456422e-01
-6.44908130e-01 8.22612524e-01 7.06623673e-01 6.74821585e-02
2.26721674e-01 6.94529340e-02 8.59711245e-02 -9.50563028e-02
-3.72685194e-01 -1.43811285e-01 3.05082768e-01 7.17102885e-01
3.66378516e-01 -1.59158140e-01 -5.61221600e-01 3.41308057e-01
-2.13468179e-01 2.24391535e-01 3.57106268e-01 7.93869317e-01
8.47088464e-04 -1.38144469e+00 -3.60579222e-01 2.33469866e-02
-3.06879073e-01 -3.84240858e-02 -1.01485029e-01 6.75407767e-01
2.91200370e-01 8.76973927e-01 -2.73246139e-01 -1.54681668e-01
2.80649781e-01 -2.56279409e-01 6.70747399e-01 -4.17097151e-01
-2.71279097e-01 4.41205204e-01 -2.45572060e-01 -4.23676521e-01
-6.79031610e-01 -4.63375092e-01 -7.48266160e-01 -2.44792655e-01
-4.65527624e-01 -7.04658777e-02 4.71788049e-01 7.43761659e-01
4.51626390e-01 7.39906192e-01 6.56127572e-01 -1.03819025e+00
-3.14431667e-01 -7.75424838e-01 -5.84155977e-01 7.93620408e-01
6.34090304e-02 -7.06478775e-01 -1.74852461e-01 2.52039075e-01] | [11.470954895019531, -0.7894147038459778] |
fa33044f-9616-4ca9-9ef5-87a77a421a04 | picie-unsupervised-semantic-segmentation | 2103.17070 | null | https://arxiv.org/abs/2103.17070v1 | https://arxiv.org/pdf/2103.17070v1.pdf | PiCIE: Unsupervised Semantic Segmentation using Invariance and Equivariance in Clustering | We present a new framework for semantic segmentation without annotations via clustering. Off-the-shelf clustering methods are limited to curated, single-label, and object-centric images yet real-world data are dominantly uncurated, multi-label, and scene-centric. We extend clustering from images to pixels and assign separate cluster membership to different instances within each image. However, solely relying on pixel-wise feature similarity fails to learn high-level semantic concepts and overfits to low-level visual cues. We propose a method to incorporate geometric consistency as an inductive bias to learn invariance and equivariance for photometric and geometric variations. With our novel learning objective, our framework can learn high-level semantic concepts. Our method, PiCIE (Pixel-level feature Clustering using Invariance and Equivariance), is the first method capable of segmenting both things and stuff categories without any hyperparameter tuning or task-specific pre-processing. Our method largely outperforms existing baselines on COCO and Cityscapes with +17.5 Acc. and +4.5 mIoU. We show that PiCIE gives a better initialization for standard supervised training. The code is available at https://github.com/janghyuncho/PiCIE. | ['Bharath Hariharan', 'Kavita Bala', 'Utkarsh Mall', 'Jang Hyun Cho'] | 2021-03-30 | null | http://openaccess.thecvf.com//content/CVPR2021/html/Cho_PiCIE_Unsupervised_Semantic_Segmentation_Using_Invariance_and_Equivariance_in_Clustering_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Cho_PiCIE_Unsupervised_Semantic_Segmentation_Using_Invariance_and_Equivariance_in_Clustering_CVPR_2021_paper.pdf | cvpr-2021-1 | ['unsupervised-semantic-segmentation'] | ['computer-vision'] | [ 1.11756451e-01 -2.38772005e-01 -2.10089758e-01 -6.95458710e-01
-9.79772568e-01 -9.74696755e-01 4.75676954e-01 2.36217920e-02
-3.58617157e-01 1.32762238e-01 -2.45366663e-01 1.40996715e-02
1.33399934e-01 -4.65781629e-01 -8.38873506e-01 -7.39076555e-01
1.60192952e-01 6.12956345e-01 3.73616934e-01 2.56936938e-01
1.25642732e-01 2.31886253e-01 -1.51462626e+00 4.22763169e-01
8.86960089e-01 8.26296568e-01 4.33499217e-01 6.02719247e-01
1.68771502e-02 3.49921554e-01 -6.58355132e-02 -1.93227366e-01
5.89577138e-01 -2.29179189e-01 -9.69661772e-01 6.61552906e-01
1.03017187e+00 1.50875002e-01 1.69494390e-01 1.31363750e+00
2.63469189e-01 1.50713786e-01 8.97629619e-01 -1.35555124e+00
-7.80086040e-01 2.89700240e-01 -7.97505856e-01 -2.18485638e-01
-1.58348009e-01 2.55342275e-01 1.09803939e+00 -9.74163175e-01
7.08353162e-01 1.25576341e+00 7.47076690e-01 4.14071590e-01
-1.53459060e+00 -5.34964859e-01 4.59416598e-01 3.20896730e-02
-1.58799052e+00 -3.32040340e-01 6.86488867e-01 -5.74812472e-01
5.18632233e-01 2.20665157e-01 4.27763939e-01 7.80183017e-01
-3.61809134e-01 1.05330658e+00 1.38284218e+00 -2.78461933e-01
2.41398618e-01 3.55366617e-01 1.26685128e-01 8.26872587e-01
1.61473885e-01 -3.40034097e-01 -3.38404924e-02 2.16635168e-01
6.40415192e-01 2.53896475e-01 -5.89665435e-02 -6.38263702e-01
-1.19457841e+00 7.70452023e-01 6.67861283e-01 1.14109017e-01
-1.42609831e-02 3.08908194e-01 1.23382062e-01 -1.32618472e-01
5.36284447e-01 4.27745670e-01 -7.74305522e-01 2.92077452e-01
-1.08958054e+00 -8.14049393e-02 5.69541454e-01 1.18808901e+00
1.20453811e+00 -2.12432966e-01 8.95549282e-02 1.14665890e+00
1.87897801e-01 7.22273707e-01 2.56877810e-01 -1.56870532e+00
7.76583403e-02 5.54857671e-01 -1.34346455e-01 -8.85763884e-01
-5.51328480e-01 -4.75673079e-01 -5.42854846e-01 1.35542884e-01
5.09713948e-01 -2.31816322e-02 -1.42533565e+00 1.65842390e+00
5.43625891e-01 1.94849059e-01 -1.96086615e-01 8.83945704e-01
7.54793108e-01 4.61341351e-01 2.18272433e-01 3.48133117e-01
1.48366857e+00 -1.40501761e+00 -1.69836789e-01 -3.93222719e-01
7.12559462e-01 -8.45117211e-01 1.39875722e+00 3.22408319e-01
-8.63159955e-01 -5.10192156e-01 -7.68583119e-01 -2.24271506e-01
-6.85260594e-01 2.12698415e-01 7.46213377e-01 6.83020830e-01
-1.12765670e+00 2.93629825e-01 -8.47891986e-01 -6.42833829e-01
6.05668187e-01 1.30822718e-01 -2.70682186e-01 -3.26126724e-01
-6.01368070e-01 4.69797939e-01 5.28722525e-01 -3.87186944e-01
-7.12549448e-01 -8.14244688e-01 -9.01945949e-01 -2.15776488e-01
3.71655524e-01 -6.11120403e-01 9.86447990e-01 -1.31013131e+00
-1.17829835e+00 1.33335721e+00 -1.43616199e-01 -1.88610375e-01
4.24502432e-01 1.35276169e-02 -1.11073896e-01 4.27644581e-01
4.84891117e-01 1.45833790e+00 6.73835218e-01 -1.87924862e+00
-6.76492810e-01 -2.42321745e-01 -1.59307290e-02 3.92020345e-01
-7.01570287e-02 -1.45383015e-01 -1.04514456e+00 -6.62053227e-01
4.49818611e-01 -1.19386530e+00 -2.80362844e-01 7.36805275e-02
-4.72942054e-01 -4.28619087e-02 8.20756257e-01 -4.25583690e-01
6.00744367e-01 -2.16641259e+00 -2.16726869e-01 2.20155999e-01
1.37048075e-02 -8.80504474e-02 -2.32676923e-01 -1.61423817e-01
1.24730669e-01 3.21646094e-01 -5.80044568e-01 -4.70306635e-01
1.72615513e-01 2.88716078e-01 2.69481510e-01 6.44302011e-01
1.68083191e-01 1.03253365e+00 -1.00187874e+00 -8.51795137e-01
6.01780176e-01 4.74457085e-01 -7.62684345e-01 -8.98550376e-02
-3.58703375e-01 5.43275297e-01 -9.75355953e-02 1.05622840e+00
8.01314831e-01 -5.78276634e-01 -8.93258899e-02 -2.21278653e-01
2.19322811e-03 -7.84057081e-02 -1.25453651e+00 1.98674345e+00
-4.97112989e-01 5.63067257e-01 2.76270270e-01 -9.27863419e-01
5.67286491e-01 -1.58216909e-01 5.93208492e-01 -5.16332150e-01
9.60455984e-02 1.88368574e-01 -3.58195633e-01 -3.53757083e-01
2.46705413e-01 1.19405337e-01 -2.13013291e-01 3.53974640e-01
1.98868141e-01 -6.76336646e-01 2.35132143e-01 3.64858478e-01
7.27237642e-01 2.60781735e-01 6.90173730e-03 -5.27775705e-01
1.51642025e-01 2.39299431e-01 6.65797830e-01 7.65332460e-01
-3.17169458e-01 1.34819436e+00 -4.70777135e-03 -1.13649763e-01
-1.06313097e+00 -1.16586292e+00 -5.26163697e-01 1.35206234e+00
4.76175308e-01 -1.08063541e-01 -1.11220562e+00 -7.13534176e-01
6.71951696e-02 6.39511228e-01 -6.60595417e-01 2.68980771e-01
-1.82655513e-01 -9.08220828e-01 3.36737931e-01 4.91687328e-01
7.48323977e-01 -7.80747533e-01 -1.86252221e-01 -8.40418488e-02
-4.19492722e-01 -1.35600889e+00 -8.26816678e-01 3.17888379e-01
-6.76279902e-01 -1.13981140e+00 -5.16624272e-01 -9.75480318e-01
9.43553805e-01 6.22807086e-01 1.20052123e+00 2.08837376e-03
-5.17106652e-01 5.69667816e-01 -4.01344627e-01 -1.73275918e-01
6.91494718e-02 1.35665715e-01 -2.17577025e-01 8.77658799e-02
4.22196418e-01 -3.43862355e-01 -8.52609396e-01 5.77157080e-01
-8.27128649e-01 2.04003796e-01 5.29143035e-01 5.52810669e-01
9.67697382e-01 -1.13713242e-01 1.64081305e-01 -1.23653388e+00
-2.59156853e-01 -5.38751245e-01 -3.99225354e-01 2.46707648e-01
-4.92125720e-01 -6.63204864e-02 3.70928973e-01 -4.05007124e-01
-1.06137478e+00 6.40452385e-01 1.12056378e-02 -4.48888421e-01
-7.41963148e-01 -5.26658073e-03 -3.89441848e-01 -8.91686380e-02
6.49484277e-01 -1.12509072e-01 -4.10648495e-01 -5.09322882e-01
9.26024020e-01 5.95296144e-01 7.86253273e-01 -8.31019759e-01
7.17129469e-01 1.02755082e+00 -4.20413226e-01 -7.60183990e-01
-1.13334727e+00 -1.14899194e+00 -1.13700199e+00 -1.53026432e-01
1.29249740e+00 -1.25295222e+00 -1.92921862e-01 4.11501050e-01
-8.22601080e-01 -7.66682029e-01 -1.76735565e-01 3.64681333e-01
-7.54870713e-01 3.39083523e-01 -4.99015629e-01 -3.42870474e-01
-8.92963707e-02 -1.06179690e+00 1.32370484e+00 2.46883452e-01
-5.80313429e-02 -1.08388448e+00 -2.49918342e-01 6.90399528e-01
1.03539504e-01 3.91778886e-01 4.63312805e-01 -4.65751022e-01
-6.38882577e-01 2.89605651e-02 -6.18335187e-01 4.16653544e-01
1.57419801e-01 1.08330466e-01 -1.17888963e+00 -1.18521839e-01
-2.09606364e-01 -4.95273620e-01 1.16952360e+00 5.52344263e-01
1.51740813e+00 -1.30940825e-01 -3.07596505e-01 1.11368191e+00
1.70447826e+00 -2.28162125e-01 3.39419305e-01 3.67126733e-01
1.15137994e+00 6.68027461e-01 5.28058946e-01 1.08745120e-01
7.25667477e-01 4.67969567e-01 3.80120277e-01 -5.71744859e-01
-5.50368786e-01 -5.88264354e-02 1.22855559e-01 6.25230491e-01
9.37951207e-02 -8.84731486e-02 -9.69989121e-01 8.79719913e-01
-1.81581414e+00 -7.81155884e-01 -5.00924587e-01 1.97419333e+00
9.07245576e-01 -5.48414364e-02 1.29112691e-01 -4.15731072e-01
9.48077857e-01 -1.71842039e-01 -7.10864007e-01 -3.03658564e-02
-3.01651746e-01 -7.67813204e-03 9.72194135e-01 6.10535204e-01
-1.62914300e+00 1.39058769e+00 5.47211647e+00 9.11643803e-01
-9.02442515e-01 3.45590591e-01 9.43936348e-01 6.51144050e-03
-1.08777605e-01 1.17430650e-02 -6.09469533e-01 4.38948393e-01
4.28080440e-01 5.25992692e-01 3.38553667e-01 1.03493309e+00
1.25711918e-01 -1.42897174e-01 -9.26983356e-01 1.05428922e+00
2.59152949e-01 -1.00393522e+00 -1.99182108e-01 -8.84172693e-02
1.41135550e+00 4.09940004e-01 2.42486820e-01 -3.54779593e-04
8.30751359e-01 -9.68675494e-01 9.14518058e-01 1.79128811e-01
7.51910686e-01 -4.69051480e-01 4.76676553e-01 -1.38187427e-02
-1.29165053e+00 1.08313553e-01 -3.25630486e-01 2.95661896e-01
-8.29251483e-03 5.79596162e-01 -5.02521038e-01 3.18339497e-01
1.06067646e+00 8.81976843e-01 -1.02885377e+00 9.39021945e-01
-2.36245334e-01 7.37997770e-01 -6.20863140e-01 3.52321059e-01
5.42797387e-01 -3.03320467e-01 1.25881108e-02 1.65442872e+00
1.99184008e-02 1.35789469e-01 6.52353287e-01 7.66042233e-01
-1.21654831e-01 6.65494055e-02 -2.21159369e-01 4.35210824e-01
4.75328684e-01 1.59395599e+00 -1.39285171e+00 -4.92301762e-01
-3.73656094e-01 1.20203590e+00 1.67602658e-01 6.13963366e-01
-8.41134846e-01 -1.37078851e-01 6.12227619e-01 8.40275735e-02
4.96198803e-01 -2.64618278e-01 -8.25827301e-01 -1.34244752e+00
-2.27385953e-01 -6.28736913e-01 5.79324365e-01 -7.99869359e-01
-1.44435656e+00 1.35980532e-01 8.72267783e-02 -9.36824620e-01
2.29500577e-01 -6.79749548e-01 -4.08417374e-01 4.26877618e-01
-1.51788676e+00 -1.58422852e+00 -4.94918406e-01 6.60279036e-01
7.73123324e-01 3.12228948e-01 5.70072412e-01 2.70791560e-01
-4.61081773e-01 4.32669610e-01 4.27126884e-01 3.14631164e-01
1.04722488e+00 -1.76271582e+00 1.41194955e-01 8.02836001e-01
2.86122441e-01 3.97246093e-01 5.36061406e-01 -3.44294190e-01
-8.48381102e-01 -1.58788264e+00 4.86319214e-01 -8.53462815e-01
5.68761289e-01 -6.36837542e-01 -7.91556954e-01 6.75641179e-01
1.22974291e-01 2.99973190e-01 5.74411571e-01 8.53730589e-02
-7.51178503e-01 -4.38598730e-02 -1.24341309e+00 4.94650215e-01
1.09454167e+00 -4.27781314e-01 -1.69696316e-01 6.96468890e-01
8.09464753e-01 -2.09642336e-01 -8.05711985e-01 3.01010191e-01
2.84904480e-01 -7.34254360e-01 1.19382083e+00 -4.76757251e-02
2.25939766e-01 -6.17774189e-01 -3.79535586e-01 -1.06073165e+00
-3.78539413e-01 -1.57351226e-01 4.90373582e-01 1.36247718e+00
5.23846447e-01 -3.63123447e-01 7.50012696e-01 6.07950866e-01
-3.64023745e-01 -3.20144087e-01 -5.09356976e-01 -8.53442609e-01
4.27963793e-01 -5.60441852e-01 2.85751343e-01 1.33600271e+00
-4.33068275e-01 2.34458119e-01 -8.06223974e-02 3.75688225e-01
9.33396816e-01 3.14598262e-01 8.05415988e-01 -1.15540588e+00
-1.87040076e-01 -6.00140035e-01 -2.41007268e-01 -8.87760699e-01
1.76051944e-01 -1.24957979e+00 4.21051055e-01 -1.60900676e+00
5.31550765e-01 -8.34136724e-01 -1.92251518e-01 7.04271019e-01
-2.52770692e-01 8.74125659e-01 2.30899990e-01 3.57170969e-01
-1.13588202e+00 1.88540593e-01 9.81578708e-01 -3.73662293e-01
-1.63454324e-01 -3.36797923e-01 -7.58912563e-01 9.09674227e-01
1.16590369e+00 -5.31294525e-01 -2.90757567e-01 -5.51057041e-01
-1.83870971e-01 -9.31494355e-01 5.48485100e-01 -9.94501472e-01
1.66457996e-01 -3.16748083e-01 4.86404121e-01 -4.55418706e-01
1.60732254e-01 -8.08318079e-01 2.52940021e-02 1.56845212e-01
-2.51301885e-01 -3.73159349e-01 -2.89249886e-03 5.47715664e-01
8.56070891e-02 -1.33157670e-01 1.16268623e+00 -4.85589445e-01
-1.03225720e+00 3.68643731e-01 -1.34360403e-01 4.26347584e-01
1.00872028e+00 -2.40501642e-01 -1.83870181e-01 -3.93557847e-02
-6.93743169e-01 5.22534549e-01 1.01327980e+00 3.12654972e-01
1.38812169e-01 -1.06929755e+00 -6.49502337e-01 -1.69202358e-01
3.40549499e-01 3.39485824e-01 2.28795558e-01 6.20130181e-01
-6.11993551e-01 2.03221753e-01 7.88012370e-02 -1.15809131e+00
-1.28359795e+00 5.28833032e-01 3.56055558e-01 2.02910423e-01
-5.21224141e-01 8.61118257e-01 6.10573590e-01 -9.78798211e-01
1.54946357e-01 -3.38725954e-01 2.05803826e-01 8.50034878e-03
7.54095837e-02 6.98871538e-02 3.07319146e-02 -7.37140775e-01
-4.17993337e-01 9.84510958e-01 8.06170031e-02 -2.20583633e-01
1.07884645e+00 -4.92429316e-01 -7.85501674e-02 5.36773384e-01
1.45639277e+00 -2.95759916e-01 -1.55394626e+00 -2.06875414e-01
1.88157354e-02 -5.00044465e-01 1.13899246e-01 -9.46439266e-01
-1.15888393e+00 8.96557808e-01 6.01236165e-01 -1.24289885e-01
9.76666093e-01 4.58458424e-01 5.56338012e-01 3.71309251e-01
2.81498730e-01 -1.50428164e+00 2.57651329e-01 3.98751765e-01
3.94468337e-01 -1.68788719e+00 1.02578931e-01 -6.96244836e-01
-8.03736985e-01 6.72567964e-01 6.72717750e-01 -2.07356349e-01
7.16025531e-01 3.13239433e-02 4.69725788e-01 -2.05089569e-01
-4.09583867e-01 -8.11912417e-01 4.81920838e-01 8.16230059e-01
3.96735668e-01 3.72437298e-01 6.33698106e-02 1.80567473e-01
-1.48009643e-01 -6.31424189e-01 3.20479870e-01 6.35687172e-01
-6.54541969e-01 -8.46573293e-01 -4.82769847e-01 3.38348269e-01
-3.12279791e-01 -2.32284695e-01 -3.29809487e-01 8.04291844e-01
5.18953621e-01 1.05122995e+00 3.29422027e-01 -1.17419206e-01
1.32474005e-02 5.79625517e-02 3.47699165e-01 -8.18375707e-01
-3.15947503e-01 4.22288924e-01 -2.16821432e-01 -6.06678128e-01
-7.76867449e-01 -9.59315658e-01 -1.43745589e+00 -1.39619499e-01
-2.47065425e-01 -1.50640875e-01 7.91138053e-01 6.52640164e-01
3.31589103e-01 1.82769835e-01 5.66085994e-01 -9.94377792e-01
1.88942134e-01 -7.11632073e-01 -4.46368039e-01 8.74985099e-01
8.70940685e-02 -5.52501559e-01 -5.25156796e-01 6.17629707e-01] | [9.567609786987305, 0.672742486000061] |
a359cb10-8e70-474d-97fc-a2e5ff2b780a | learning-identity-invariant-motion | null | null | http://openaccess.thecvf.com/content_CVPR_2020/html/Huang_Learning_Identity-Invariant_Motion_Representations_for_Cross-ID_Face_Reenactment_CVPR_2020_paper.html | http://openaccess.thecvf.com/content_CVPR_2020/papers/Huang_Learning_Identity-Invariant_Motion_Representations_for_Cross-ID_Face_Reenactment_CVPR_2020_paper.pdf | Learning Identity-Invariant Motion Representations for Cross-ID Face Reenactment | Human face reenactment aims at transferring motion patterns from one face (from a source-domain video) to an-other (in the target domain with the identity of interest).While recent works report impressive results, they are notable to handle multiple identities in a unified model. In this paper, we propose a unique network of CrossID-GAN to perform multi-ID face reenactment. Given a source-domain video with extracted facial landmarks and a target-domain image, our CrossID-GAN learns the identity-invariant motion patterns via the extracted landmarks and such information to produce the videos whose ID matches that of the target domain. Both supervised and unsupervised settings are proposed to train and guide our model during training.Our qualitative/quantitative results confirm the robustness and effectiveness of our model, with ablation studies confirming our network design.
| [' Yu-Chiang Frank Wang', ' Fu-En Yang', 'Po-Hsiang Huang'] | 2020-06-01 | null | null | null | cvpr-2020-6 | ['face-reenactment'] | ['computer-vision'] | [ 3.54473412e-01 2.83603460e-01 -4.42265213e-01 -3.68342876e-01
-7.14862645e-01 -7.70658731e-01 6.51146591e-01 -1.19533038e+00
-7.21077397e-02 5.74168384e-01 5.26276767e-01 4.52383399e-01
1.54549390e-01 -4.00867850e-01 -7.74471343e-01 -7.13344157e-01
1.28445804e-01 6.32161796e-01 -3.76566380e-01 -1.09075233e-01
-1.66562751e-01 5.33206284e-01 -1.42831087e+00 1.65601537e-01
2.69509345e-01 7.74784088e-01 -1.64345369e-01 6.29725218e-01
2.60532260e-01 9.17644560e-01 -3.77734214e-01 -6.02063835e-01
6.60586655e-01 -9.68258500e-01 -9.66681302e-01 2.07418412e-01
1.12049592e+00 -6.68900311e-01 -5.96927762e-01 8.62789571e-01
5.15817881e-01 1.18617713e-01 5.66640019e-01 -1.78461051e+00
-8.22922349e-01 2.88394094e-01 -8.36912453e-01 -1.16323661e-02
5.28534353e-01 1.61379114e-01 6.34382665e-01 -7.84192026e-01
1.33845222e+00 1.32938588e+00 8.15595329e-01 1.39530718e+00
-1.22638977e+00 -1.03896117e+00 1.84915047e-02 -8.64576474e-02
-1.37395203e+00 -1.01350772e+00 1.03139293e+00 -3.22845161e-01
3.54450166e-01 -2.65933603e-01 6.53859556e-01 1.74506724e+00
-1.32776931e-01 5.18580973e-01 7.02175677e-01 -2.06385657e-01
-8.01694766e-02 -1.20486811e-01 -7.43978739e-01 7.69107938e-01
1.22892102e-02 4.01997447e-01 -8.42167616e-01 6.52406812e-02
1.19940913e+00 5.84094133e-03 -1.77254632e-01 -6.32580042e-01
-1.00030923e+00 8.50935996e-01 2.76347309e-01 4.21618015e-01
-3.43881190e-01 4.52744424e-01 2.64275730e-01 3.36809486e-01
3.37257504e-01 1.77390411e-01 -1.62765682e-01 -2.11621687e-01
-1.06667233e+00 2.48211455e-02 5.30081034e-01 1.19277942e+00
7.78258324e-01 5.07254779e-01 1.77422352e-02 6.40235007e-01
1.60026178e-01 6.62670910e-01 4.73981440e-01 -1.47058046e+00
4.16820198e-02 1.50876388e-01 -6.26083314e-02 -9.24457908e-01
-6.21907460e-03 -4.68385220e-02 -8.30855727e-01 3.98503006e-01
3.75353307e-01 -4.50852871e-01 -9.25525427e-01 2.44468904e+00
3.67092848e-01 7.33383596e-01 2.61212558e-01 8.56472790e-01
6.84195817e-01 3.13479125e-01 1.44548099e-02 -9.45094675e-02
1.01814067e+00 -9.58185732e-01 -6.06825650e-01 -3.05976748e-01
4.44755584e-01 -6.01600528e-01 8.05184782e-01 2.56110774e-03
-1.25134075e+00 -6.61399603e-01 -7.24897563e-01 -1.47887692e-01
7.21937791e-02 2.22536787e-01 3.06823552e-01 5.99293828e-01
-1.69008315e+00 4.82471675e-01 -6.01137757e-01 -6.72291279e-01
7.36692548e-01 5.50840318e-01 -9.56365407e-01 -1.90325379e-01
-8.91501367e-01 6.58795357e-01 4.09445725e-02 -1.80957541e-02
-1.47605705e+00 -7.44040847e-01 -9.52398360e-01 -3.29569846e-01
-1.69608882e-03 -1.01095736e+00 1.22489977e+00 -1.74848807e+00
-1.55725884e+00 1.27090931e+00 -3.65946531e-01 -2.61588305e-01
6.85178757e-01 -3.05607617e-02 -3.55782926e-01 4.72046971e-01
4.39648181e-01 1.32293940e+00 1.38817751e+00 -1.41274130e+00
-6.04318082e-01 -3.72261018e-01 -3.53321657e-02 2.31934145e-01
-3.69186789e-01 1.03456359e-02 -6.83949769e-01 -8.89664829e-01
-2.56767362e-01 -1.09630048e+00 1.66419461e-01 3.21962893e-01
-1.92608297e-01 2.32278988e-01 1.27516258e+00 -7.31483757e-01
6.80914521e-01 -2.26611161e+00 3.96982193e-01 -7.40331784e-03
3.40419263e-01 -4.89698276e-02 -5.33732057e-01 1.72102571e-01
-3.60304654e-01 8.26892555e-02 -6.26363829e-02 -7.41262376e-01
-3.06621194e-01 1.85578302e-01 -2.70370603e-01 6.65646493e-01
1.75019324e-01 1.02187645e+00 -9.51230228e-01 -5.02062321e-01
-5.59506938e-02 7.16251612e-01 -6.67345643e-01 4.27253604e-01
4.52020727e-02 1.02949464e+00 -2.77126133e-01 5.40412545e-01
7.18256295e-01 -1.31057695e-01 2.09646568e-01 -5.14457107e-01
3.12497705e-01 -3.53231877e-01 -7.39402771e-01 2.06968141e+00
-5.29869318e-01 7.65485108e-01 3.01843882e-01 -5.92260897e-01
6.38688087e-01 5.24346352e-01 8.45939755e-01 -5.57986975e-01
1.35397002e-01 -9.67519358e-03 -2.38428175e-01 -4.31975067e-01
2.43831396e-01 -3.92125398e-01 1.11396678e-01 7.59369612e-01
6.31792665e-01 2.02646717e-01 -1.71527371e-01 1.99015498e-01
7.85725117e-01 6.19855464e-01 2.67241970e-02 -1.66091055e-01
3.73934746e-01 -1.39730141e-01 6.32036328e-01 4.25759703e-01
-2.42974445e-01 8.95009637e-01 2.74229497e-01 -5.52801013e-01
-1.28319407e+00 -1.13003063e+00 2.93306142e-01 1.06826270e+00
3.54755893e-02 -8.22163820e-02 -9.69002724e-01 -9.93531704e-01
-9.50832665e-02 3.33880275e-01 -1.17027175e+00 -2.49714360e-01
-7.58762658e-01 5.31648621e-02 8.91262829e-01 5.79343498e-01
6.84768677e-01 -1.12462318e+00 -2.51371741e-01 -4.36544418e-02
-2.68651068e-01 -1.38157105e+00 -1.10785854e+00 -4.58995134e-01
-6.38289511e-01 -1.16603529e+00 -8.89156938e-01 -1.19742930e+00
1.00031376e+00 2.64474720e-01 1.00246930e+00 -2.10937876e-02
-1.13575876e-01 8.44902337e-01 -2.43284017e-01 8.01442936e-02
-5.44464529e-01 -1.15883775e-01 4.30338353e-01 3.99957299e-01
1.01356573e-01 -8.17685306e-01 -6.96232796e-01 4.51208472e-01
-7.94178903e-01 9.34787560e-03 2.77534306e-01 8.06326151e-01
1.80491805e-01 -3.00474375e-01 7.48137236e-01 -8.17498088e-01
1.21494003e-01 -4.56815839e-01 -2.68684834e-01 6.96898699e-02
-2.98639119e-01 -2.08906919e-01 7.11683810e-01 -5.38530350e-01
-1.13966668e+00 3.92839164e-01 1.58267599e-02 -1.18037832e+00
-1.47313565e-01 -2.99368709e-01 -3.70667517e-01 -3.77418041e-01
5.22849679e-01 3.16733271e-01 3.80989343e-01 -1.64119080e-01
6.05954945e-01 1.92623511e-01 1.03673685e+00 -6.01124704e-01
1.17215323e+00 9.76610422e-01 -1.57026619e-01 -4.92052406e-01
-5.53469002e-01 7.28035253e-03 -1.04715252e+00 -3.52291763e-01
1.13345945e+00 -1.09389818e+00 -6.25952363e-01 6.19447112e-01
-1.13314617e+00 -4.64480877e-01 -5.44159532e-01 2.00050935e-01
-9.02062297e-01 -9.02763754e-03 -3.95841539e-01 -4.07771319e-01
-2.75408119e-01 -9.46577370e-01 1.29961669e+00 4.70189363e-01
-3.36616457e-01 -1.33052397e+00 1.86615109e-01 2.19134554e-01
2.78149128e-01 4.78071600e-01 3.80409062e-01 -3.91995788e-01
-5.11722147e-01 1.87136710e-01 5.05650509e-03 2.22777784e-01
7.20898032e-01 2.14188881e-02 -1.15780687e+00 -6.58752024e-01
-5.93405031e-02 -3.54115695e-01 4.72698897e-01 4.54571187e-01
7.15046465e-01 -6.19613826e-01 -5.31340957e-01 1.02245724e+00
1.21549475e+00 3.02348256e-01 5.27240872e-01 9.80875343e-02
9.15751159e-01 6.34145439e-01 3.70605439e-01 2.53887713e-01
2.71032780e-01 6.87422574e-01 3.51235151e-01 -3.96533102e-01
-5.86971104e-01 -6.63949549e-01 6.80065513e-01 2.93374985e-01
-3.46555747e-02 -9.51743796e-02 -5.58141053e-01 8.91551495e-01
-1.47700202e+00 -1.24239028e+00 7.99336374e-01 2.02753854e+00
7.62252271e-01 -4.47887450e-01 4.25029516e-01 -5.86874425e-01
8.26972663e-01 2.00807586e-01 -7.80091822e-01 8.09833705e-02
-1.94921777e-01 1.83241054e-01 1.82928860e-01 4.80976611e-01
-1.00081384e+00 1.33477557e+00 6.78412104e+00 4.27582324e-01
-1.34977651e+00 3.04483473e-01 6.65837526e-01 -3.65903407e-01
-2.46136621e-01 -1.53261200e-01 -7.21804023e-01 3.49393398e-01
6.84409440e-01 -4.59091991e-01 6.43479824e-01 7.92277396e-01
7.69296736e-02 4.41297114e-01 -1.47406292e+00 1.13926291e+00
5.37106276e-01 -1.56871450e+00 2.13803917e-01 1.54985124e-02
1.09125531e+00 -2.82358259e-01 4.35902774e-01 2.73717986e-03
6.93939745e-01 -1.18156755e+00 7.19246387e-01 4.24373507e-01
1.44559503e+00 -7.63292611e-01 3.54285896e-01 -2.39873782e-01
-1.28932047e+00 4.66952957e-02 2.01794002e-02 3.56843233e-01
2.51424611e-01 -2.47648895e-01 -9.01713133e-01 5.32445252e-01
7.57264078e-01 1.27392900e+00 -5.01666784e-01 3.41795921e-01
5.89137059e-03 3.21875244e-01 -1.02634609e-01 8.22928905e-01
7.28945062e-02 -4.27787513e-01 4.80731815e-01 8.91588867e-01
5.40574074e-01 -4.89212275e-02 -9.56574604e-02 8.48783970e-01
-5.95610440e-01 -2.60031015e-01 -1.19884706e+00 6.77907988e-02
5.13022363e-01 1.25512922e+00 -4.89252359e-01 -2.55806983e-01
-3.73764247e-01 1.40768957e+00 1.51405871e-01 5.56029499e-01
-8.20734203e-01 7.49155432e-02 1.27424741e+00 7.45337754e-02
4.50484246e-01 3.37523259e-02 5.68848811e-02 -1.11666119e+00
-2.00895652e-01 -1.12400985e+00 4.44895327e-01 -8.71723890e-01
-1.34530556e+00 8.87847543e-01 3.40811200e-02 -1.45845842e+00
-7.89425671e-01 -2.45652407e-01 -6.78739548e-01 5.27641594e-01
-1.07395172e+00 -1.70441532e+00 -4.64804351e-01 9.60842550e-01
5.29713035e-01 -5.55636764e-01 7.05807626e-01 3.06325942e-01
-4.66148496e-01 9.95431304e-01 6.75359229e-03 6.14420593e-01
1.08372521e+00 -6.86208189e-01 5.90686381e-01 9.00235713e-01
2.25402504e-01 5.99403441e-01 4.19344127e-01 -6.40682757e-01
-1.40890169e+00 -1.31575632e+00 3.03842366e-01 -6.28322959e-01
3.80704284e-01 -3.42269927e-01 -5.14020503e-01 1.35171032e+00
4.86934870e-01 1.65092856e-01 5.58314085e-01 -4.03562546e-01
-5.34904540e-01 -1.46028131e-01 -1.45590460e+00 7.51502872e-01
1.63309264e+00 -7.53738344e-01 -1.98468879e-01 -4.64076176e-02
5.71839213e-01 -4.33117658e-01 -8.97989571e-01 7.36534745e-02
7.47521400e-01 -8.47977757e-01 1.04861963e+00 -7.56790221e-01
4.86879796e-01 -3.09729964e-01 6.34139869e-03 -1.34465420e+00
-2.02525362e-01 -9.69284236e-01 4.44277003e-02 1.45228052e+00
-9.43860859e-02 -5.29235542e-01 9.57409322e-01 3.09134722e-01
2.66764045e-01 -2.36603916e-01 -9.98423994e-01 -7.00521827e-01
1.71331123e-01 1.15945794e-01 6.56480193e-01 1.25043154e+00
-3.91384900e-01 2.92813182e-01 -8.52828562e-01 2.91246790e-02
7.96359956e-01 9.86874774e-02 1.01077259e+00 -1.07964373e+00
1.14533938e-02 -3.09525162e-01 -4.34147000e-01 -9.57320213e-01
7.62925804e-01 -8.73130918e-01 -2.19699621e-01 -1.08906341e+00
1.97525203e-01 -2.17528671e-01 2.07869053e-01 8.37083697e-01
2.16149852e-01 7.93976188e-01 2.31136307e-01 2.79883236e-01
-3.11874479e-01 6.30589068e-01 1.35947812e+00 -1.10480569e-01
-3.68324630e-02 -3.35010618e-01 -9.14012969e-01 5.90928376e-01
5.51154852e-01 -5.08535028e-01 -7.55571127e-01 -5.62874794e-01
-2.93588191e-01 6.87680617e-02 4.63753372e-01 -8.42755318e-01
1.89581603e-01 -1.08855866e-01 6.92048669e-01 7.18870014e-02
4.71952736e-01 -8.72166872e-01 5.66580415e-01 2.92454541e-01
-2.87411332e-01 2.39223495e-01 2.68937677e-01 5.26238859e-01
-1.52285591e-01 5.71164973e-02 1.04883933e+00 1.01128988e-01
-8.42406511e-01 7.81431615e-01 -3.71442437e-02 3.23563248e-01
1.16602612e+00 -5.16583145e-01 -1.79141447e-01 -9.65492189e-01
-9.02490973e-01 -2.71914825e-02 1.16094637e+00 7.71486700e-01
6.80156529e-01 -1.69816005e+00 -8.85047436e-01 5.37715554e-01
2.57032514e-02 -1.38037413e-01 3.32413167e-01 5.68261325e-01
-3.20889652e-01 5.94380014e-02 -8.16620409e-01 -6.72795951e-01
-1.22894537e+00 5.23367465e-01 7.33123362e-01 1.60556242e-01
-7.61136353e-01 7.44162560e-01 6.64504826e-01 -3.21880043e-01
-2.81705912e-02 4.15971547e-01 5.99421337e-02 -4.36957441e-02
4.47596759e-01 8.44505057e-02 -3.31358761e-01 -1.27327335e+00
-3.78068537e-01 1.05223072e+00 -9.77601185e-02 -4.81954664e-01
1.24025059e+00 -1.74338937e-01 4.92286962e-03 -4.17529494e-02
1.60912716e+00 1.12963468e-01 -1.83456028e+00 -1.87445462e-01
-3.03461254e-01 -7.26181030e-01 -3.74490172e-01 -4.70964938e-01
-1.63632011e+00 4.64095712e-01 6.31962180e-01 -6.60463035e-01
1.36774766e+00 1.69335991e-01 8.05690825e-01 -1.92605644e-01
4.86066312e-01 -7.38053501e-01 5.29754698e-01 1.50468901e-01
8.77918363e-01 -1.13827717e+00 -4.43204999e-01 -1.24036670e-01
-9.13939297e-01 9.94016409e-01 9.21578825e-01 -1.35104805e-01
4.89699781e-01 2.89512426e-01 3.24493498e-01 -1.55875295e-01
-4.76319790e-01 -6.25515729e-02 2.44970731e-02 1.14333606e+00
1.51247978e-01 -4.94594067e-01 5.01968443e-01 2.16577277e-01
-3.64493936e-01 1.58207133e-01 4.72911417e-01 6.71355963e-01
3.37821156e-01 -1.23263633e+00 -2.77069092e-01 4.32455204e-02
-2.18499079e-01 2.16488525e-01 -6.10699713e-01 9.96728837e-01
7.52984285e-02 7.87460923e-01 3.62526387e-01 -4.87560838e-01
9.21463594e-02 1.35905314e-02 8.41906071e-01 -4.34302181e-01
-4.22727972e-01 -8.28767419e-02 -1.65085226e-01 -8.30423653e-01
-6.47372425e-01 -8.45642447e-01 -1.20580649e+00 -5.16426504e-01
3.39965999e-01 2.38254983e-02 2.18789488e-01 6.27276838e-01
7.07521379e-01 1.58493266e-01 9.40188944e-01 -1.08733702e+00
-3.35976519e-02 -6.59534752e-01 -5.28749943e-01 7.30771899e-01
7.68701494e-01 -7.25404859e-01 -4.18126196e-01 6.45664573e-01] | [12.703385353088379, -0.1822313666343689] |
3c994d06-ce33-4893-bcca-3209417fc5c2 | improved-disentangled-speech-representations | 2211.08191 | null | https://arxiv.org/abs/2211.08191v2 | https://arxiv.org/pdf/2211.08191v2.pdf | Improved disentangled speech representations using contrastive learning in factorized hierarchical variational autoencoder | Leveraging the fact that speaker identity and content vary on different time scales, \acrlong{fhvae} (\acrshort{fhvae}) uses different latent variables to symbolize these two attributes. Disentanglement of these attributes is carried out by different prior settings of the corresponding latent variables. For the prior of speaker identity variable, \acrshort{fhvae} assumes it is a Gaussian distribution with an utterance-scale varying mean and a fixed variance. By setting a small fixed variance, the training process promotes identity variables within one utterance gathering close to the mean of their prior. However, this constraint is relatively weak, as the mean of the prior changes between utterances. Therefore, we introduce contrastive learning into the \acrshort{fhvae} framework, to make the speaker identity variables gathering when representing the same speaker, while distancing themselves as far as possible from those of other speakers. The model structure has not been changed in this work but only the training process, thus no additional cost is needed during testing. Voice conversion has been chosen as the application in this paper. Latent variable evaluations include speaker verification and identification for the speaker identity variable, and speech recognition for the content variable. Furthermore, assessments of voice conversion performance are on the grounds of fake speech detection experiments. Results show that the proposed method improves both speaker identity and content feature extraction compared to \acrshort{fhvae}, and has better performance than baseline on conversion. | ['Zheng-Hua Tan', 'Thomas Arildsen', 'Yuying Xie'] | 2022-11-15 | null | null | null | null | ['voice-conversion', 'voice-conversion', 'speaker-verification'] | ['audio', 'speech', 'speech'] | [ 5.09452522e-02 3.02114874e-01 -1.92639932e-01 -4.99942392e-01
-7.47009873e-01 -8.09316218e-01 8.19706619e-01 -1.62290409e-01
-3.02403331e-01 6.16844416e-01 1.18269525e-01 -2.42367342e-01
4.53798361e-02 -5.52311897e-01 -4.07765806e-01 -1.05402255e+00
1.40116826e-01 4.24074918e-01 -6.76688999e-02 -1.04603566e-01
-9.04456079e-02 4.71697956e-01 -1.47130287e+00 1.32288396e-01
5.47124028e-01 7.40288854e-01 9.25309658e-02 6.14592850e-01
-1.74792349e-01 2.29559332e-01 -1.08411205e+00 -4.17836845e-01
2.13189632e-01 -5.21375716e-01 -5.67346275e-01 1.92922890e-01
3.93490613e-01 -2.03802794e-01 -5.09656370e-02 1.20464396e+00
5.57261467e-01 2.00367033e-01 6.65552437e-01 -1.37753892e+00
-6.03232920e-01 8.42057347e-01 -2.20007762e-01 7.34153166e-02
3.59126180e-01 -1.33230045e-01 1.07863688e+00 -7.42429852e-01
4.44577694e-01 1.40390790e+00 5.28183758e-01 6.39823556e-01
-1.49955142e+00 -9.79903221e-01 4.09284472e-01 4.21561152e-02
-1.53704441e+00 -8.50665748e-01 8.44772816e-01 -6.34317338e-01
6.21888578e-01 6.24569476e-01 2.96263725e-01 1.46860516e+00
-1.97011292e-01 5.12124002e-01 1.15269315e+00 -6.76273465e-01
2.96757221e-01 6.10192358e-01 2.99513578e-01 1.94636315e-01
-2.49595538e-01 2.19443887e-01 -5.54675817e-01 -3.03329945e-01
5.18070877e-01 -4.46388572e-01 -3.91266137e-01 -9.61270928e-02
-9.21076477e-01 9.85220313e-01 -2.95329839e-01 4.46374625e-01
-2.15682358e-01 -2.31991678e-01 2.53853589e-01 5.33058763e-01
3.22477043e-01 1.19903445e-01 -2.82193512e-01 -2.50586301e-01
-1.03362370e+00 -1.31349592e-02 6.97527289e-01 1.04510498e+00
6.38372898e-01 4.04440731e-01 -2.16062889e-01 1.05442655e+00
4.36848283e-01 6.75659597e-01 6.88212454e-01 -8.48891795e-01
3.34078312e-01 1.38763553e-02 2.04821020e-01 -8.35109651e-01
3.73237878e-02 -3.44959825e-01 -5.97963393e-01 3.29785049e-01
4.25864011e-01 -1.63312778e-01 -8.97736013e-01 2.09608054e+00
2.66254038e-01 2.24338189e-01 1.68808818e-01 6.14239991e-01
6.65617943e-01 7.84371972e-01 -4.94344020e-03 -7.11391330e-01
1.26462567e+00 -6.98270202e-01 -1.23474801e+00 -1.68245062e-01
1.42196998e-01 -9.98671710e-01 1.25934350e+00 4.46450144e-01
-8.75666618e-01 -7.19067276e-01 -1.09542489e+00 2.51367897e-01
-2.01229662e-01 1.50694400e-01 3.07948917e-01 1.26603019e+00
-8.80310476e-01 2.40746215e-01 -8.38892221e-01 2.92367898e-02
-2.14088723e-01 4.20722336e-01 -3.96260083e-01 5.39642692e-01
-1.35739267e+00 6.99447334e-01 8.20488110e-02 1.24936588e-01
-8.07685554e-01 -2.77449638e-01 -8.63938451e-01 3.19862552e-02
6.03690520e-02 -1.58587962e-01 1.18761051e+00 -1.00906837e+00
-2.08701015e+00 7.67489672e-01 -3.21006864e-01 -3.37198555e-01
6.66013241e-01 -1.06583670e-01 -7.89158165e-01 -1.17224678e-01
5.83137153e-03 2.81798124e-01 1.48229969e+00 -1.33398187e+00
-3.71766210e-01 -1.72825411e-01 -1.23118475e-01 9.05272737e-02
-4.86624092e-01 2.73959726e-01 -3.42378110e-01 -7.87262917e-01
2.91257143e-01 -1.10981834e+00 5.19605637e-01 -4.97934312e-01
-3.87804151e-01 -8.38274956e-02 9.64134276e-01 -8.12847018e-01
1.27427125e+00 -2.65842652e+00 1.08352944e-01 1.83047637e-01
-1.18110888e-01 1.50977030e-01 -2.53089070e-02 2.64054954e-01
-2.80415714e-01 2.05575109e-01 -2.02654198e-01 -7.31752455e-01
1.89397901e-01 4.06365335e-01 -6.56456709e-01 6.02529228e-01
-1.09989047e-01 1.94292963e-01 -6.19960248e-01 -2.85294980e-01
1.54438987e-01 6.04607522e-01 -3.23953390e-01 2.04119623e-01
1.11268781e-01 4.31370407e-01 -3.49297114e-02 2.18537405e-01
7.40494072e-01 5.03032267e-01 1.73727304e-01 -2.68510133e-02
-2.77317643e-01 4.91595864e-01 -1.47511005e+00 1.07628536e+00
-5.90709031e-01 7.30072856e-01 3.99611056e-01 -7.17825651e-01
1.17616427e+00 8.03612530e-01 2.97462165e-01 -1.80752844e-01
1.32536054e-01 4.96838279e-02 2.43741259e-01 -2.86284536e-01
2.92764753e-01 -4.08494413e-01 -3.24317627e-02 2.24483892e-01
8.15507770e-02 -3.50236058e-01 -1.68298259e-02 -1.49941310e-01
5.25002360e-01 -2.04981014e-01 1.66522920e-01 -5.48249595e-02
7.11304903e-01 -6.73444211e-01 7.17830598e-01 5.64411223e-01
-3.25834602e-01 5.51125526e-01 4.51521218e-01 3.28522235e-01
-7.52730250e-01 -1.27426589e+00 -4.08494085e-01 1.14411998e+00
3.09281014e-02 -2.62667388e-01 -8.23726714e-01 -5.93242705e-01
-2.03945771e-01 1.18685329e+00 -4.71692681e-01 -1.82706833e-01
-6.25606239e-01 -5.17546475e-01 8.28751743e-01 1.72870874e-01
1.43543631e-01 -9.73037124e-01 -1.47250473e-01 2.28911161e-01
-4.02077734e-01 -1.04676294e+00 -5.32468438e-01 2.98948646e-01
-4.30179417e-01 -4.43672419e-01 -4.91591424e-01 -6.96649611e-01
3.08622658e-01 -1.46281466e-01 6.77636623e-01 -3.41443479e-01
2.60717243e-01 2.90628731e-01 -4.34747845e-01 -3.95001888e-01
-1.00917888e+00 -1.75591245e-01 3.90248716e-01 3.39392722e-01
2.68185675e-01 -5.88430822e-01 -1.55883823e-02 5.71742952e-01
-6.82313025e-01 -4.84221637e-01 1.92951754e-01 1.00125277e+00
4.47494984e-01 4.15644020e-01 5.86551905e-01 -6.31052673e-01
7.22898602e-01 -1.86960340e-01 -5.47910929e-01 9.62887555e-02
-5.61340213e-01 -1.29019573e-01 5.61494350e-01 -9.80919600e-01
-1.14111817e+00 -1.02102585e-01 -1.90219447e-01 -4.95179147e-01
-5.22986770e-01 1.92415133e-01 -7.13600516e-01 3.77295375e-01
5.50053179e-01 2.01576740e-01 6.94710761e-02 -4.92009789e-01
3.26695085e-01 1.04621577e+00 6.24050736e-01 -3.79345238e-01
8.70708704e-01 3.25411350e-01 -5.19829512e-01 -1.15253747e+00
-4.88620937e-01 -3.39087218e-01 -7.07997441e-01 2.78074690e-03
6.63884938e-01 -7.76943624e-01 -4.95781630e-01 6.15353346e-01
-1.11651492e+00 -1.58073291e-01 -4.27432925e-01 8.60684574e-01
-5.01948655e-01 4.29051965e-01 -4.33752120e-01 -1.02005124e+00
3.73899639e-02 -1.31611967e+00 8.82944405e-01 -1.82803959e-01
-6.72870219e-01 -1.03238153e+00 1.36011578e-02 3.69046539e-01
2.12836489e-01 -5.96875772e-02 9.06086564e-01 -8.62166166e-01
-1.02280155e-01 -2.43444338e-01 3.64279002e-01 7.76775062e-01
6.10086441e-01 2.77949035e-01 -1.40054691e+00 -5.03103256e-01
5.85242033e-01 1.82143331e-01 4.37343806e-01 3.30293834e-01
5.80010474e-01 -4.37418282e-01 9.62953083e-03 5.05268574e-01
6.79579198e-01 4.27552223e-01 3.60305160e-01 1.94148332e-01
4.51067328e-01 7.77935565e-01 4.29997355e-01 3.02786857e-01
-2.00795252e-02 9.38171804e-01 1.70246214e-01 7.60050863e-02
-2.41514206e-01 8.85742605e-02 8.95687580e-01 9.74353015e-01
2.09361345e-01 -2.74449229e-01 -6.15143001e-01 4.09917384e-01
-1.25980818e+00 -1.22485006e+00 -9.43120793e-02 2.49953580e+00
1.02452743e+00 2.38418236e-01 1.43430466e-02 4.39971358e-01
1.00653517e+00 3.22057545e-01 -3.42493653e-01 -6.30023122e-01
-1.87562674e-01 -2.06282753e-02 2.13740975e-01 9.98573959e-01
-1.12808037e+00 9.49812114e-01 5.74866486e+00 8.14727008e-01
-1.48975539e+00 1.51512742e-01 1.97528467e-01 2.57400684e-02
-3.77463341e-01 3.93750593e-02 -1.05566370e+00 5.64287722e-01
8.72823417e-01 -1.96901649e-01 4.83978987e-01 6.02767766e-01
4.27700251e-01 3.25925916e-01 -1.23960388e+00 9.32253122e-01
1.34715110e-01 -5.87549925e-01 -9.60011631e-02 7.18057826e-02
4.03347015e-01 -3.13560516e-01 4.91946369e-01 4.00205463e-01
4.46821116e-02 -9.20692980e-01 1.19453430e+00 1.27080590e-01
8.04523945e-01 -5.46168029e-01 7.06392586e-01 4.09205168e-01
-1.06791842e+00 9.67384875e-03 -1.67713210e-01 8.72291103e-02
1.72026739e-01 3.88777912e-01 -1.10115778e+00 2.72390872e-01
3.62221569e-01 9.21156406e-02 -2.26195365e-01 5.55436909e-01
-5.33923209e-01 1.07245708e+00 -2.40061983e-01 3.43113542e-01
-1.25267863e-01 -2.16515154e-01 9.99868631e-01 1.33017707e+00
3.93060356e-01 -2.51921922e-01 3.40569288e-01 9.46054876e-01
1.79983780e-01 2.05080226e-01 -3.75160933e-01 -7.96939284e-02
9.04553235e-01 9.49119866e-01 -5.20355821e-01 -2.78502524e-01
-2.23132789e-01 9.71791744e-01 -1.36214241e-01 5.16892850e-01
-9.80461717e-01 -3.16665381e-01 7.44798541e-01 4.23175022e-02
3.40757161e-01 -2.65740007e-01 -2.31430367e-01 -9.53143179e-01
3.95553038e-02 -1.01514292e+00 2.02387273e-01 -3.07166576e-01
-1.16871727e+00 9.73590195e-01 2.61616141e-01 -1.21473908e+00
-6.93233132e-01 -3.12832266e-01 -4.66986239e-01 1.28505528e+00
-1.28696966e+00 -1.11214888e+00 5.38322981e-03 7.02574432e-01
5.26477039e-01 -4.22231346e-01 1.12250876e+00 1.14852883e-01
-6.52244866e-01 1.08828664e+00 1.33000270e-01 1.33471787e-01
9.53278005e-01 -1.29087138e+00 2.15389594e-01 8.01812530e-01
3.34863067e-01 9.47541118e-01 1.01722789e+00 -5.66823423e-01
-7.03831553e-01 -7.21085072e-01 9.55697775e-01 -2.78989643e-01
6.64076507e-01 -7.05347180e-01 -1.07250333e+00 6.18224502e-01
1.64432466e-01 -2.41912484e-01 9.17274892e-01 2.47933701e-01
-4.46455002e-01 -2.17409171e-02 -1.11876297e+00 4.28052425e-01
4.81751949e-01 -9.84680176e-01 -7.70818174e-01 7.47330710e-02
7.23474324e-01 -1.30296588e-01 -7.46290445e-01 1.95260718e-01
4.05187845e-01 -8.98019552e-01 8.47691238e-01 -1.28907323e-01
-3.36012542e-01 -3.31278652e-01 -3.34436953e-01 -1.29907739e+00
-2.64693439e-01 -8.93671393e-01 -9.65793431e-02 1.89629650e+00
3.97754163e-01 -7.77213991e-01 6.66874230e-01 5.28138399e-01
-1.12903818e-01 -3.95436287e-02 -1.13068759e+00 -1.14104521e+00
1.49658531e-01 -4.81118828e-01 6.27969801e-01 1.20251429e+00
-2.87214428e-01 4.01363283e-01 -3.76530290e-01 6.97761953e-01
4.12332863e-01 1.78160202e-02 7.35228896e-01 -1.12284970e+00
-5.21128714e-01 -4.29239422e-01 -3.28737617e-01 -8.50550115e-01
4.77960527e-01 -7.25317717e-01 9.21774507e-02 -8.88241947e-01
-2.51626849e-01 -3.41276139e-01 -2.36803606e-01 4.83860910e-01
-9.67846513e-02 -1.32563105e-02 1.20400622e-01 1.75394788e-01
4.30228323e-01 6.37728810e-01 8.85402262e-01 -1.94198713e-01
-5.33624411e-01 4.66275483e-01 -2.49586612e-01 8.54561508e-01
7.74182200e-01 -4.60091501e-01 -5.79143822e-01 -1.39808461e-01
-4.09438252e-01 1.30275816e-01 2.13902161e-01 -7.26016521e-01
-1.99726075e-02 -4.65711579e-02 -4.30715717e-02 -3.13482523e-01
7.55499601e-01 -8.30374122e-01 3.02219510e-01 9.66254324e-02
-2.86285251e-01 -1.08007252e-01 1.91362470e-01 2.85607666e-01
-5.03337801e-01 -4.57739681e-01 9.80084360e-01 3.00960451e-01
-2.66516834e-01 -1.12555265e-01 -5.69534540e-01 -2.22315922e-01
8.38158429e-01 -4.09316540e-01 7.07712099e-02 -5.85203290e-01
-1.11708236e+00 -3.57436776e-01 2.69955784e-01 5.17553627e-01
3.30800891e-01 -1.20117497e+00 -6.95195615e-01 5.06184638e-01
-6.81400076e-02 -3.28129232e-01 1.01514541e-01 6.84574723e-01
1.42716363e-01 1.59835011e-01 -4.18052785e-02 -6.51313365e-01
-1.74367177e+00 4.77791905e-01 2.20258445e-01 5.44566996e-02
-3.38057399e-01 1.03835869e+00 2.62481838e-01 -5.38349032e-01
5.41596472e-01 -2.08190888e-01 -3.59532595e-01 6.00105286e-01
3.22045833e-01 2.78637856e-01 -6.32851807e-05 -1.22736931e+00
-2.50372559e-01 2.97186136e-01 -2.31107116e-01 -5.93866408e-01
9.98483837e-01 -4.71498966e-01 -9.82050449e-02 9.00501311e-01
1.18096256e+00 7.64925241e-01 -1.11233902e+00 -2.61738664e-03
-1.22817680e-01 -5.35158396e-01 9.48207006e-02 -6.71165943e-01
-6.23406768e-01 9.21195447e-01 8.51245165e-01 5.13394177e-01
8.75081718e-01 -2.52501756e-01 3.63708556e-01 -1.31170955e-02
2.08820254e-01 -1.10462713e+00 -1.24616839e-01 4.83723462e-01
1.04715860e+00 -1.00080478e+00 -1.90175816e-01 -6.13456547e-01
-6.25280321e-01 9.49526429e-01 3.49539608e-01 3.22250456e-01
5.47493577e-01 2.63975680e-01 4.74918783e-01 2.87623733e-01
-3.70457649e-01 1.48748696e-01 2.56351680e-01 5.70419431e-01
5.07559836e-01 2.90926099e-01 -1.66824296e-01 6.59187973e-01
-8.23278904e-01 -6.22328758e-01 4.25131083e-01 5.65405250e-01
-1.34020388e-01 -1.49928784e+00 -7.84125507e-01 -1.57755092e-01
-5.31708717e-01 -9.73062813e-02 -3.95677209e-01 7.72489607e-01
3.10387820e-01 1.23818135e+00 -1.89937297e-02 -2.42456257e-01
3.04011971e-01 4.15458649e-01 1.63449407e-01 -7.32830048e-01
-5.06458223e-01 5.33382535e-01 1.55032068e-01 3.93639272e-03
-1.77911326e-01 -9.83914852e-01 -1.32874036e+00 3.89557122e-03
-7.74835706e-01 4.14782763e-01 7.98539400e-01 8.53507996e-01
-8.94416198e-02 5.30430436e-01 9.62921262e-01 -5.53801954e-01
-9.91182685e-01 -1.20756912e+00 -7.03621566e-01 5.01814187e-01
5.01621664e-01 -6.91565752e-01 -7.85563111e-01 3.23163450e-01] | [14.682943344116211, 6.256300926208496] |
989057dc-2b2b-41e5-be02-450740205103 | accelerated-and-quantitative-3d-semisolid-mt | 2207.11297 | null | https://arxiv.org/abs/2207.11297v1 | https://arxiv.org/pdf/2207.11297v1.pdf | Accelerated and Quantitative 3D Semisolid MT/CEST Imaging using a Generative Adversarial Network (GAN-CEST) | Purpose: To substantially shorten the acquisition time required for quantitative 3D chemical exchange saturation transfer (CEST) and semisolid magnetization transfer (MT) imaging and allow for rapid chemical exchange parameter map reconstruction. Methods: Three-dimensional CEST and MT magnetic resonance fingerprinting (MRF) datasets of L-arginine phantoms, whole-brains, and calf muscles from healthy volunteers, cancer patients, and cardiac patients were acquired using 3T clinical scanners at 3 different sites, using 3 different scanner models and coils. A generative adversarial network supervised framework (GAN-CEST) was then designed and trained to learn the mapping from a reduced input data space to the quantitative exchange parameter space, while preserving perceptual and quantitative content. Results: The GAN-CEST 3D acquisition time was 42-52 seconds, 70% shorter than CEST-MRF. The quantitative reconstruction of the entire brain took 0.8 seconds. An excellent agreement was observed between the ground truth and GAN-based L-arginine concentration and pH values (Pearson's r > 0.97, NRMSE < 1.5%). GAN-CEST images from a brain-tumor subject yielded a semi-solid volume fraction and exchange rate NRMSE of 3.8$\pm$1.3% and 4.6$\pm$1.3%, respectively, and SSIM of 96.3$\pm$1.6% and 95.0$\pm$2.4%, respectively. The mapping of the calf-muscle exchange parameters in a cardiac patient, yielded NRMSE < 7% and SSIM > 94% for the semi-solid exchange parameters. In regions with large susceptibility artifacts, GAN-CEST has demonstrated improved performance and reduced noise compared to MRF. Conclusion: GAN-CEST can substantially reduce the acquisition time for quantitative semisolid MT/CEST mapping, while retaining performance even when facing pathologies and scanner models that were not available during training. | ['Or Perlman', 'Christian T. Farrar', 'Moritz Zaiss', 'Christopher Nguyen', 'Elizabeth R. Gerstner', 'Anna N. Foster', 'Jaume Coll-Font', 'Kai Herz', 'Maria Sedykh', 'Jonah Weigand-Whittier'] | 2022-07-22 | null | null | null | null | ['magnetic-resonance-fingerprinting'] | ['medical'] | [ 5.21519065e-01 2.61834979e-01 1.70983493e-01 -1.92157492e-01
-1.01588476e+00 -3.20026606e-01 1.80087209e-01 -8.82100612e-02
-6.01711988e-01 9.57020164e-01 -5.93058802e-02 -2.54039139e-01
1.07411832e-01 -2.69440025e-01 -6.62044287e-01 -9.95530128e-01
-3.63815010e-01 7.52803445e-01 2.75483638e-01 1.66559145e-01
-6.19417652e-02 4.98640209e-01 -3.86938542e-01 1.36056125e-01
1.04763567e+00 8.72664392e-01 3.40093017e-01 5.60371399e-01
1.55455843e-01 1.13603210e+00 -4.59662318e-01 1.10409223e-01
2.97378153e-01 -8.14266860e-01 -6.97185397e-01 -3.68280709e-01
3.58520210e-01 -3.09805810e-01 -3.67554009e-01 9.52832103e-01
7.55729318e-01 1.32112980e-01 7.54813135e-01 -8.63441467e-01
-4.64130104e-01 6.55019045e-01 -8.10172856e-01 2.92953372e-01
-1.46254480e-01 4.43182945e-01 -2.62583226e-01 -4.69915658e-01
8.17343056e-01 4.31668282e-01 7.40093708e-01 9.02308166e-01
-1.44674647e+00 -8.72276723e-01 -5.36045074e-01 -2.42324769e-01
-1.16818714e+00 -3.22876811e-01 3.22285831e-01 -6.45689726e-01
1.01352584e+00 3.90255392e-01 8.41174364e-01 6.80451632e-01
1.02342606e+00 1.29855141e-01 1.94229054e+00 -2.27266833e-01
2.28727445e-01 1.37208834e-01 -4.03488606e-01 6.47754192e-01
7.01497570e-02 -6.51773736e-02 -2.76049882e-01 -3.78949285e-01
1.17146242e+00 -3.14253569e-01 -5.01883209e-01 -3.59550208e-01
-1.47059822e+00 4.90093678e-01 2.87817627e-01 6.31049156e-01
-5.42430162e-01 2.40045354e-01 4.01440591e-01 9.64666680e-02
2.57074684e-01 4.88854289e-01 6.13414608e-02 -1.06733531e-01
-8.77276361e-01 -1.46810219e-01 3.79994661e-01 8.79365444e-01
4.04351205e-02 3.67091209e-01 -8.17811340e-02 7.48154044e-01
-3.86315864e-03 1.06156576e+00 6.22185647e-01 -1.13845778e+00
3.68540026e-02 -1.79616153e-01 1.58587158e-01 -4.47336406e-01
-3.91276866e-01 -5.18324196e-01 -1.07196796e+00 9.24287140e-02
4.51939106e-01 -1.19652107e-01 -1.22646177e+00 1.79766488e+00
4.28620577e-01 6.80881292e-02 -1.41694337e-01 1.20103788e+00
4.61953938e-01 4.06594872e-01 4.01140839e-01 -6.93746388e-01
1.13733387e+00 -6.03997767e-01 -8.04342151e-01 -1.41932309e-01
5.99636674e-01 -6.34925723e-01 9.53958869e-01 1.19322993e-01
-1.65608501e+00 -1.41135335e-01 -9.90039825e-01 3.86628330e-01
4.55643326e-01 -5.20721793e-01 2.76030540e-01 8.54414046e-01
-1.07872498e+00 6.39970124e-01 -1.35476732e+00 1.39492899e-01
3.91296357e-01 5.00450730e-01 -3.54217559e-01 -3.27324569e-01
-1.24410701e+00 1.30501282e+00 7.50022307e-02 -5.38725741e-02
-9.81692672e-01 -1.34923816e+00 -4.79645938e-01 -5.11889577e-01
-6.13392256e-02 -6.37757778e-01 1.05949426e+00 -6.60768270e-01
-1.55866027e+00 9.59369302e-01 -1.81589443e-02 -6.20813012e-01
7.28231072e-01 2.20100522e-01 -5.53201854e-01 6.94931984e-01
3.40713829e-01 5.79319477e-01 5.17700016e-01 -1.05394340e+00
3.34808797e-01 -4.42892045e-01 -7.86009490e-01 2.20568672e-01
5.02982557e-01 3.01399916e-01 1.58488050e-01 -4.86946851e-01
5.47581136e-01 -1.19425678e+00 -3.85792375e-01 -7.98526928e-02
-1.81251094e-01 8.87427986e-01 2.83219308e-01 -1.25528157e+00
4.97179508e-01 -1.65924168e+00 -9.01658237e-02 4.21864599e-01
5.94549954e-01 5.27218096e-02 1.33113116e-01 -3.05203110e-01
-3.00998956e-01 6.03425577e-02 -8.49012136e-01 3.33992511e-01
-3.89548153e-01 -3.30003589e-01 3.54493141e-01 9.87315536e-01
-4.39650327e-01 1.30328894e+00 -1.12882614e+00 -3.56931627e-01
1.93361208e-01 6.80332303e-01 -1.78517848e-01 -8.71234387e-02
3.75221342e-01 1.26434481e+00 -2.91716486e-01 5.07709265e-01
8.05106461e-01 -1.06378585e-01 6.82697237e-01 -3.06766242e-01
-1.13400161e-01 -4.98442426e-02 -4.44414467e-01 1.93897057e+00
-2.62263119e-01 4.01960731e-01 3.88670444e-01 -6.45925581e-01
8.50671530e-01 5.47718942e-01 1.28597724e+00 -1.52803659e+00
2.46750638e-01 6.82998896e-01 4.49930936e-01 -3.89972895e-01
-1.90151080e-01 -1.19899952e+00 1.04154557e-01 5.93244433e-01
-1.06765501e-01 -6.28303587e-01 -4.66870695e-01 2.25280389e-01
1.18401372e+00 -7.23971799e-02 -3.39121699e-01 -8.46467197e-01
2.76924819e-01 -1.23085780e-02 4.79618371e-01 6.79381251e-01
-5.61157167e-01 5.73287487e-01 2.93500334e-01 -1.96695372e-01
-1.50062978e+00 -1.24248576e+00 -4.08259749e-01 1.12397060e-01
1.51587501e-02 2.20995814e-01 -1.11729479e+00 -2.82821327e-01
-3.05952638e-01 6.64667726e-01 -4.89774466e-01 -2.94196963e-01
-8.46304297e-01 -9.06209469e-01 7.11907208e-01 3.78865033e-01
4.41399217e-01 -9.16671991e-01 -8.40228498e-01 4.98323768e-01
-5.45773745e-01 -1.06648922e+00 -5.38598001e-01 3.21015030e-01
-1.22613025e+00 -7.70638049e-01 -1.11869836e+00 -3.64606619e-01
6.27448976e-01 -3.16919386e-01 1.14289260e+00 -2.88073629e-01
-4.00535315e-01 2.72419721e-01 -5.29530011e-02 -3.28217596e-01
-6.90011919e-01 -3.62875253e-01 3.78582537e-01 -4.47270304e-01
-1.32894725e-01 -7.79588044e-01 -9.89216566e-01 4.72674340e-01
-8.21188331e-01 1.56115830e-01 4.04273957e-01 8.52378249e-01
1.13090599e+00 -5.19470215e-01 5.69246054e-01 -9.33445215e-01
1.75784603e-01 -3.89489830e-01 -2.92330295e-01 1.84958369e-01
-5.58098912e-01 3.34597491e-02 3.86868536e-01 -6.58225894e-01
-9.36899602e-01 -3.46828580e-01 6.04771934e-02 -5.53383291e-01
1.02163374e-01 1.28036574e-01 1.13592833e-01 -4.48931098e-01
9.82079327e-01 3.76095325e-01 3.32590967e-01 4.43465635e-02
2.60964539e-02 2.37065658e-01 7.28151083e-01 -7.30402946e-01
4.61239725e-01 4.82711762e-01 1.96316019e-01 -5.61377943e-01
-1.72467664e-01 1.44979656e-01 -6.11959636e-01 -2.78981119e-01
9.96160030e-01 -6.36209607e-01 -3.87056023e-01 5.90288281e-01
-4.98915374e-01 -9.10521984e-01 -5.37371993e-01 1.06746662e+00
-8.25643063e-01 4.79923278e-01 -1.05665720e+00 -4.02976394e-01
-9.42896366e-01 -1.53588593e+00 3.59356701e-01 -2.03001667e-02
-3.39389145e-01 -9.32902932e-01 -1.81822643e-01 6.57178104e-01
1.10519791e+00 9.25965428e-01 1.02572727e+00 -1.40758067e-01
-3.16545159e-01 1.33203030e-01 1.81430191e-01 3.89612824e-01
1.53272614e-01 -1.00096643e+00 -7.55477726e-01 -5.38232625e-01
6.67803347e-01 -2.72434413e-01 3.44359815e-01 1.07464123e+00
8.56503606e-01 1.40342936e-01 8.50928295e-03 7.27125466e-01
1.22213948e+00 9.95306909e-01 9.03856516e-01 1.27550617e-01
6.10985637e-01 2.74986953e-01 4.58907336e-01 9.08344388e-02
-1.95117369e-01 3.67104590e-01 1.77835062e-01 -2.63631701e-01
-2.19544679e-01 -1.22267112e-01 2.23047361e-01 9.20769513e-01
-8.00804198e-02 1.26691133e-01 -9.78969097e-01 3.27990055e-01
-9.29933310e-01 -6.71680272e-01 -3.31533819e-01 2.27661777e+00
9.60697889e-01 2.61495798e-03 9.86106880e-03 -2.45417133e-01
7.50073671e-01 -3.16576570e-01 -1.03419197e+00 -2.55308479e-01
-8.81861243e-03 6.94068789e-01 8.86864424e-01 6.37185037e-01
-4.04798388e-01 4.43653852e-01 6.05618715e+00 3.25114459e-01
-1.57307005e+00 6.27856851e-01 8.47729087e-01 -3.53258401e-01
-4.92537081e-01 -8.99776146e-02 1.79093570e-01 5.77808499e-01
1.43440723e+00 -3.73162366e-02 6.15673780e-01 1.03512369e-01
1.64771825e-01 -4.70584571e-01 -7.75824666e-01 1.10667896e+00
-1.37623951e-01 -1.08944905e+00 -5.30663967e-01 1.48841441e-01
6.06963813e-01 3.23709071e-01 2.13080674e-01 -8.34738389e-02
-1.00453697e-01 -1.29647171e+00 7.85283983e-01 6.24541879e-01
1.90192032e+00 -6.19709671e-01 5.04122257e-01 1.76845312e-01
-6.64843738e-01 5.44688642e-01 -7.11034890e-03 7.27392733e-01
3.72433811e-01 6.96315289e-01 -1.10385048e+00 3.84545326e-01
5.90719998e-01 -1.73972286e-02 2.03384571e-02 5.61267674e-01
2.04268500e-01 4.72079664e-01 -3.72755349e-01 4.67952371e-01
8.26258659e-02 -3.78598064e-01 5.06990850e-01 8.39944363e-01
3.00829500e-01 6.18323386e-01 -2.43779987e-01 1.09182060e+00
-2.07517296e-02 4.22854684e-02 -1.96445152e-01 1.30582169e-01
4.41054016e-01 9.42104757e-01 -9.39532757e-01 -4.06764269e-01
-1.83420740e-02 9.70294714e-01 -2.50432730e-01 3.53157490e-01
-1.12632346e+00 -1.59858644e-01 3.43878791e-02 7.34511435e-01
-3.61338019e-01 5.62011439e-04 -2.15960234e-01 -1.12601781e+00
-3.40814665e-02 -8.67256522e-01 -9.47178749e-04 -9.27526414e-01
-9.31403339e-01 7.80543804e-01 2.27103531e-01 -8.15625310e-01
-1.47202805e-01 -1.00991048e-01 -2.29448169e-01 1.38710546e+00
-1.11445785e+00 -6.93613768e-01 -2.54275292e-01 5.66117823e-01
-9.11445692e-02 2.00958371e-01 7.91313052e-01 4.00329709e-01
-2.89655048e-02 5.66491663e-01 2.40647852e-01 -8.53611305e-02
7.37023175e-01 -1.10244787e+00 2.07381681e-01 5.31412661e-01
-7.76777625e-01 7.61537135e-01 4.60396916e-01 -1.02693355e+00
-1.42369282e+00 -9.10694540e-01 2.79622674e-01 -3.45752060e-01
3.48374397e-01 -1.67505115e-01 -1.08653855e+00 6.04045391e-01
1.90907642e-01 3.99870723e-01 7.56944656e-01 -8.25708926e-01
6.34865612e-02 1.79180086e-01 -2.05020142e+00 1.55979410e-01
6.62919283e-01 -6.13997996e-01 -1.20339140e-01 2.21209422e-01
4.32553887e-01 -1.07223499e+00 -1.74974597e+00 3.56366724e-01
6.59049571e-01 -6.15038157e-01 8.16588819e-01 -3.17167729e-01
6.29522577e-02 -2.71229267e-01 -7.42493719e-02 -1.23124731e+00
-3.56894553e-01 -6.12377763e-01 1.31719366e-01 4.07700300e-01
1.97306320e-01 -8.35845709e-01 7.36970186e-01 1.17963052e+00
-4.88400459e-01 -5.44033587e-01 -1.17756975e+00 -4.14252311e-01
6.04931891e-01 -1.62704393e-01 2.95171618e-01 1.20592380e+00
-1.33109748e-01 -2.67582312e-02 -2.87669271e-01 -1.24388359e-01
1.09193563e+00 -4.06331152e-01 -1.44736394e-01 -9.02075469e-01
-1.17633216e-01 1.16725504e-01 -1.40038267e-01 -3.42224717e-01
4.46729697e-02 -1.25121200e+00 -1.67758316e-01 -1.27424681e+00
3.70528549e-01 -9.20976937e-01 -4.99132991e-01 3.82356316e-01
3.32596689e-01 1.57947525e-01 1.27823934e-01 2.38268554e-01
1.52084947e-01 2.77741522e-01 1.80265188e+00 -2.62440920e-01
-1.19616225e-01 -7.04950392e-01 -3.40818852e-01 2.06815854e-01
9.47888613e-01 -6.69510901e-01 -6.33645892e-01 -3.27761829e-01
-2.84922272e-01 8.94824982e-01 3.58469933e-01 -1.20284224e+00
3.02071609e-02 -1.61671489e-01 8.05538833e-01 -1.79126546e-01
1.49866819e-01 -6.91306949e-01 1.11655307e+00 9.07631814e-01
-9.50617790e-02 -1.05611637e-01 3.29354823e-01 -7.30988532e-02
2.44262621e-01 5.62574565e-02 1.42855561e+00 -6.35543704e-01
-3.61124128e-02 3.74498457e-01 -5.01988947e-01 2.31168941e-01
7.90923417e-01 -2.69071013e-01 -4.05067839e-02 -1.88293010e-01
-1.18327677e+00 -1.13126963e-01 6.71248198e-01 -1.27732456e-01
6.67124748e-01 -1.22730815e+00 -6.54303670e-01 1.59705684e-01
-5.12586713e-01 9.51836910e-03 1.09293151e+00 1.73136866e+00
-9.77024078e-01 1.82554111e-01 -6.33829474e-01 -7.69309878e-01
-6.08436108e-01 2.49002323e-01 1.24553061e+00 -3.26091468e-01
-7.31408060e-01 5.70745766e-01 1.01131365e-01 -4.18222249e-01
-5.53242207e-01 -3.47790569e-01 5.80682755e-01 -6.79546416e-01
3.71790409e-01 3.62937093e-01 4.23703104e-01 -7.24536002e-01
-4.57439214e-01 4.46406007e-01 9.34838504e-02 -3.83640617e-01
1.15180063e+00 -1.21179320e-01 -1.27650082e-01 8.68107900e-02
1.01787210e+00 -1.67394042e-01 -1.24946284e+00 -3.86836715e-02
-3.52655023e-01 -2.65162379e-01 3.21496606e-01 -1.35454798e+00
-1.35269010e+00 7.59212673e-01 1.16466820e+00 -4.14773583e-01
1.03540778e+00 -7.04749674e-02 9.70731735e-01 -4.72796559e-01
5.57688355e-01 -7.80243099e-01 -2.98513055e-01 8.62080604e-02
7.48135507e-01 -5.85520029e-01 -1.32657751e-01 -3.09502542e-01
-7.72522807e-01 7.06621885e-01 4.41591173e-01 2.35320628e-01
2.67079830e-01 6.88130379e-01 1.88204452e-01 -2.99218863e-01
-1.94012493e-01 8.73814464e-01 -4.70192321e-02 8.53742063e-01
7.14833379e-01 3.33011508e-01 -9.28529352e-02 2.33504340e-01
7.38318861e-02 2.41395403e-02 5.87222755e-01 1.13012564e+00
-8.16269591e-02 -7.85141051e-01 -3.96237433e-01 5.22284925e-01
-6.91134989e-01 -1.23549484e-01 1.32622808e-01 6.40497804e-01
-1.04752652e-01 4.85907853e-01 -1.20412022e-01 4.99175079e-02
4.37823176e-01 2.33801514e-01 1.08982825e+00 -1.93930507e-01
-7.54927576e-01 4.41347122e-01 -2.74637103e-01 -5.84519506e-01
-5.50770342e-01 -6.73823237e-01 -1.82973993e+00 -4.24155980e-01
1.09164696e-02 4.58535910e-01 8.09683204e-01 4.30790126e-01
2.12520272e-01 5.77601612e-01 4.19933707e-01 -5.68448424e-01
-2.29911432e-01 -1.06575572e+00 -1.14610326e+00 1.88608319e-01
1.51953533e-01 -5.42379558e-01 -2.07951695e-01 -1.44325361e-01] | [13.5971040725708, -2.418259859085083] |
883d082b-a86e-4f4f-8415-96bacb0bb865 | fine-grained-classification-with-noisy-labels | 2303.02404 | null | https://arxiv.org/abs/2303.02404v1 | https://arxiv.org/pdf/2303.02404v1.pdf | Fine-Grained Classification with Noisy Labels | Learning with noisy labels (LNL) aims to ensure model generalization given a label-corrupted training set. In this work, we investigate a rarely studied scenario of LNL on fine-grained datasets (LNL-FG), which is more practical and challenging as large inter-class ambiguities among fine-grained classes cause more noisy labels. We empirically show that existing methods that work well for LNL fail to achieve satisfying performance for LNL-FG, arising the practical need of effective solutions for LNL-FG. To this end, we propose a novel framework called stochastic noise-tolerated supervised contrastive learning (SNSCL) that confronts label noise by encouraging distinguishable representation. Specifically, we design a noise-tolerated supervised contrastive learning loss that incorporates a weight-aware mechanism for noisy label correction and selectively updating momentum queue lists. By this mechanism, we mitigate the effects of noisy anchors and avoid inserting noisy labels into the momentum-updated queue. Besides, to avoid manually-defined augmentation strategies in contrastive learning, we propose an efficient stochastic module that samples feature embeddings from a generated distribution, which can also enhance the representation ability of deep models. SNSCL is general and compatible with prevailing robust LNL strategies to improve their performance for LNL-FG. Extensive experiments demonstrate the effectiveness of SNSCL. | ['Yilong Yin', 'Chenhui Guo', 'Ren Wang', 'Haoliang Sun', 'Lei Feng', 'Qi Wei'] | 2023-03-04 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Wei_Fine-Grained_Classification_With_Noisy_Labels_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Wei_Fine-Grained_Classification_With_Noisy_Labels_CVPR_2023_paper.pdf | cvpr-2023-1 | ['learning-with-noisy-labels', 'learning-with-noisy-labels'] | ['computer-vision', 'natural-language-processing'] | [ 2.22347021e-01 -6.84663951e-02 -1.42719001e-01 -5.95747471e-01
-1.11825669e+00 -5.44754624e-01 4.22221661e-01 1.40531734e-01
-4.74946529e-01 7.16481686e-01 -1.01794198e-03 1.86807197e-02
-4.33016479e-01 -7.04656065e-01 -5.38255990e-01 -8.13134789e-01
3.41798246e-01 3.28790277e-01 6.45350367e-02 4.49312441e-02
1.56547427e-01 4.33346003e-01 -1.71240699e+00 2.35838696e-01
9.63000596e-01 1.15613210e+00 2.60854095e-01 2.47176349e-01
-3.98185343e-01 6.86357141e-01 -6.40476048e-01 -3.91331404e-01
2.83488929e-01 -4.09389257e-01 -5.71720183e-01 2.51782358e-01
5.48115134e-01 4.76450920e-02 1.84963778e-01 1.21563625e+00
7.66268492e-01 3.60140443e-01 8.76533449e-01 -1.39034295e+00
-7.30878532e-01 6.76846504e-01 -5.97412348e-01 2.07318470e-01
-1.66191623e-01 2.89609265e-02 1.06713092e+00 -1.21148252e+00
1.50427654e-01 1.42310846e+00 9.72624838e-01 9.57167983e-01
-1.24313068e+00 -7.12457120e-01 6.66512251e-01 6.78614527e-02
-1.30912590e+00 -2.45827809e-01 8.70365322e-01 -2.69752055e-01
3.21774304e-01 3.87359977e-01 2.71628857e-01 1.16851354e+00
-2.33157679e-01 7.57887959e-01 1.28210473e+00 -5.44539273e-01
5.28522193e-01 2.51556575e-01 4.45072949e-01 6.35012805e-01
2.81539023e-01 4.08274420e-02 -5.41570961e-01 -3.05409670e-01
4.62766379e-01 -2.42956565e-03 -2.31935903e-01 -2.61406451e-01
-9.78182793e-01 7.62328148e-01 3.64857495e-01 1.54621944e-01
-1.56686336e-01 2.15107232e-01 5.78963935e-01 3.02456260e-01
8.69009256e-01 3.11064094e-01 -6.05467737e-01 9.89689752e-02
-9.11472082e-01 1.43802494e-01 2.53379077e-01 8.74017894e-01
7.64847338e-01 -7.31002539e-03 -5.61070561e-01 1.12190890e+00
3.54185551e-01 1.39525965e-01 8.34851086e-01 -8.49397182e-01
2.89645493e-01 6.22085273e-01 -4.17300649e-02 -6.67964995e-01
-4.93446141e-01 -1.00287914e+00 -9.38055515e-01 7.47874975e-02
2.76072234e-01 4.44059074e-02 -8.29082012e-01 2.02435088e+00
4.77131724e-01 5.26428223e-01 2.67612487e-02 6.89347923e-01
8.02245855e-01 4.68129069e-01 3.68361443e-01 -2.40168139e-01
1.04721355e+00 -1.26034784e+00 -7.59012520e-01 -1.88436851e-01
9.38584805e-01 -5.11630535e-01 1.66761148e+00 3.52256775e-01
-8.35768461e-01 -6.51091397e-01 -7.90574670e-01 -7.12250918e-02
-3.97450238e-01 4.51779097e-01 5.25238276e-01 7.62789965e-01
-9.01062250e-01 5.57661951e-01 -5.70219576e-01 3.62972505e-02
5.35108149e-01 1.76320568e-01 -1.16135448e-01 -1.64419904e-01
-1.12924004e+00 5.86985826e-01 4.09540653e-01 4.45308685e-01
-6.66116238e-01 -7.97395170e-01 -9.21837986e-01 5.03023714e-02
4.02683258e-01 -5.37475944e-01 1.23998797e+00 -7.62598872e-01
-1.34272075e+00 8.49002123e-01 -1.44969150e-01 -1.92225263e-01
6.94981933e-01 -1.63105488e-01 -2.17793405e-01 -2.32585207e-01
2.82764494e-01 6.68497264e-01 9.43204939e-01 -1.57714200e+00
-5.87163925e-01 -2.76888371e-01 1.12103522e-02 2.74175286e-01
-8.05783808e-01 -2.29309946e-01 -9.68570337e-02 -1.04838324e+00
3.98914486e-01 -7.36334383e-01 -4.16754782e-01 1.52591720e-01
-3.10389996e-01 -7.25155711e-01 6.40204489e-01 -1.00178465e-01
1.43071806e+00 -2.32899857e+00 -2.79530913e-01 1.35385886e-01
3.03091615e-01 3.36418420e-01 -4.48566616e-01 1.39449099e-02
6.38515353e-02 2.21946448e-01 -3.10293347e-01 -8.00022066e-01
3.12158346e-01 4.63326722e-01 -2.70313382e-01 3.24963748e-01
4.32894349e-01 8.01523149e-01 -1.11076081e+00 -4.93228972e-01
6.09491318e-02 3.25830370e-01 -5.86586714e-01 6.01128079e-02
-2.94244558e-01 2.70498782e-01 -4.51810390e-01 7.04807341e-01
9.10693467e-01 -4.48295981e-01 -1.48056969e-01 -3.27381283e-01
1.82800889e-01 9.69495848e-02 -1.43774235e+00 1.26994383e+00
-6.61586821e-01 -6.19879887e-02 -8.02726150e-02 -1.01735258e+00
1.08161366e+00 -1.65927457e-03 4.87736911e-02 -4.44256634e-01
1.70676962e-01 3.82960677e-01 -4.74202603e-01 -3.51331025e-01
3.13438326e-01 -3.10499668e-01 -1.20204113e-01 3.57648611e-01
6.45478219e-02 -4.71933372e-02 1.01138979e-01 -6.89058825e-02
9.54763651e-01 -8.96416977e-02 1.57968178e-01 -2.66278118e-01
5.68754494e-01 -4.85543191e-01 8.24045837e-01 9.89868879e-01
-5.62976182e-01 7.43435383e-01 1.17873289e-01 -2.30106130e-01
-6.43281281e-01 -9.59484637e-01 -3.34399670e-01 1.49987257e+00
1.91678554e-01 -2.79683203e-01 -8.28117132e-01 -1.23385704e+00
-3.20246592e-02 5.09250700e-01 -7.19689667e-01 -3.82169724e-01
-3.95289689e-01 -1.03874934e+00 5.02430081e-01 6.17886901e-01
5.65842211e-01 -9.98488426e-01 8.19322169e-02 2.63619304e-01
-6.79856688e-02 -8.53882432e-01 -5.09642601e-01 7.00279176e-01
-7.54225612e-01 -8.59729886e-01 -5.86332023e-01 -1.09641373e+00
9.08419371e-01 2.56887436e-01 1.06528127e+00 -5.36417589e-02
5.15017062e-02 1.06594339e-01 -4.61776793e-01 -1.99484095e-01
-2.84468621e-01 1.77043810e-01 1.93317503e-01 1.45826086e-01
4.45991844e-01 -3.74362916e-01 -4.84183103e-01 4.71470565e-01
-1.10514188e+00 -3.84277999e-01 4.53985095e-01 1.18301249e+00
8.88880312e-01 3.30012143e-01 1.14419389e+00 -1.10181129e+00
8.05568635e-01 -5.40380359e-01 -4.50961977e-01 3.35348994e-01
-7.21183360e-01 1.29572660e-01 9.62014973e-01 -7.35109270e-01
-9.61782992e-01 -6.52825162e-02 -3.75728875e-01 -4.70978260e-01
-1.33964881e-01 2.94288725e-01 -3.35609376e-01 -2.91298497e-02
7.60010839e-01 1.59720942e-01 -3.33560675e-01 -7.85808802e-01
5.00981271e-01 8.11615586e-01 3.83216590e-01 -7.35557556e-01
7.06113696e-01 4.13610905e-01 -7.62976184e-02 -2.69227028e-01
-1.50790215e+00 -4.70251262e-01 -3.12132716e-01 -2.30295230e-02
2.09713876e-01 -9.16307449e-01 -3.70767504e-01 5.15588403e-01
-7.10025191e-01 -3.28887135e-01 -6.89724863e-01 3.58200014e-01
-5.05026579e-01 2.90706843e-01 -5.83225846e-01 -7.77413666e-01
-2.36177623e-01 -9.86257493e-01 1.21318853e+00 2.26497903e-01
-9.32725742e-02 -1.15225923e+00 -1.99167076e-02 2.76230574e-01
3.99980813e-01 -5.34842536e-02 8.67538214e-01 -7.95597374e-01
-1.89002797e-01 -9.24794376e-03 -3.44482034e-01 9.00332630e-01
2.05271736e-01 -3.22168529e-01 -1.20026886e+00 -4.15983140e-01
1.54386833e-01 -7.04244554e-01 1.07749975e+00 2.81309366e-01
1.52628243e+00 -2.28414610e-01 -1.82429612e-01 4.44695055e-01
1.26705480e+00 -1.13112703e-01 1.55863270e-01 4.62481320e-01
7.50075459e-01 5.51467478e-01 8.21463525e-01 5.53560674e-01
2.46450588e-01 4.01714146e-01 3.45411777e-01 -1.02451409e-03
-3.32813799e-01 -2.95847118e-01 1.21883847e-01 1.05011010e+00
4.73163009e-01 -4.01544958e-01 -4.52105284e-01 5.97527504e-01
-1.79154623e+00 -4.79541689e-01 4.49668728e-02 1.83202326e+00
1.09808505e+00 2.55432069e-01 -3.02886724e-01 5.38313687e-01
9.66662109e-01 1.99892238e-01 -6.70483470e-01 -1.87441051e-01
-2.71139950e-01 1.29833698e-01 2.08643138e-01 4.27625120e-01
-1.24734354e+00 9.54803824e-01 5.57474899e+00 1.61853230e+00
-9.15370405e-01 4.53933477e-01 8.23101461e-01 -1.62899062e-01
-6.11185312e-01 -3.99419844e-01 -1.20812368e+00 7.18034208e-01
5.91645002e-01 2.53064781e-01 7.19323307e-02 1.08241487e+00
2.42694274e-01 2.10299671e-01 -1.04054379e+00 1.10206223e+00
-5.37405834e-02 -1.12887704e+00 1.93335578e-01 -3.03864270e-01
9.80045617e-01 -2.57398963e-01 4.15223420e-01 6.74809635e-01
2.85787404e-01 -7.13645577e-01 8.19567382e-01 3.99009228e-01
7.47709811e-01 -7.75584280e-01 7.79036760e-01 5.64845026e-01
-9.53321278e-01 -3.09900910e-01 -6.15693033e-01 1.41638592e-02
3.41612697e-02 1.22695351e+00 -4.48108256e-01 2.87471980e-01
5.56693435e-01 5.22336781e-01 -7.38702953e-01 1.01827693e+00
-2.58580297e-01 7.17060268e-01 -3.32905173e-01 -3.45043279e-02
4.87415105e-01 1.12034768e-01 2.32617378e-01 1.15095127e+00
3.17330927e-01 -2.40519956e-01 4.83004212e-01 6.64680779e-01
-3.48731160e-01 1.21272758e-01 -2.51576155e-01 2.58964390e-01
7.26648390e-01 1.18628550e+00 -7.16749668e-01 -2.62959212e-01
-1.86348557e-01 8.17262590e-01 6.78591251e-01 1.85751960e-01
-7.74045825e-01 -1.10164024e-01 4.72483039e-01 -1.34464502e-02
6.15303256e-02 2.33490437e-01 -4.17070419e-01 -9.66663957e-01
3.35275322e-01 -7.38868535e-01 3.84410501e-01 -3.41099471e-01
-1.94573629e+00 7.68840313e-01 -3.67240429e-01 -1.27681589e+00
1.44764259e-01 -3.47336560e-01 -3.10047835e-01 5.19262791e-01
-1.95808578e+00 -1.07849324e+00 -3.37911248e-01 4.71340597e-01
7.55547822e-01 2.51232032e-02 7.04504550e-01 4.97835875e-01
-7.10306108e-01 1.09295237e+00 4.44940865e-01 -3.33117664e-01
8.35544765e-01 -1.30117476e+00 2.58700341e-01 5.88260472e-01
5.13940789e-02 4.26055610e-01 3.55745137e-01 -5.41985512e-01
-5.15947163e-01 -1.81814635e+00 9.83639061e-01 -2.34049767e-01
5.67099154e-01 -5.17530441e-01 -9.89683688e-01 3.97044748e-01
-4.89864171e-01 6.32549047e-01 7.42231071e-01 -6.62814528e-02
-4.47668493e-01 -2.20120743e-01 -1.49482071e+00 4.85578209e-01
1.30393732e+00 -5.70570886e-01 -3.82827163e-01 6.09316349e-01
8.16938102e-01 -1.67643681e-01 -5.40194631e-01 6.42344296e-01
8.61665979e-02 -7.21307218e-01 7.40672469e-01 -4.43195343e-01
9.71239880e-02 -5.00141263e-01 -3.24875563e-01 -1.41705012e+00
-4.56926197e-01 -3.13242972e-01 -1.00888185e-01 1.52000916e+00
1.81082726e-01 -6.34878755e-01 9.17659104e-01 2.73925513e-01
-3.06126922e-01 -1.03005719e+00 -9.96536553e-01 -1.11505318e+00
1.01107612e-01 -4.71569121e-01 5.74849308e-01 1.07048595e+00
-5.16926646e-01 -3.60973142e-02 -2.33886540e-01 2.07787931e-01
7.38211513e-01 -1.85587451e-01 2.97840565e-01 -1.26278210e+00
-1.99582204e-01 -4.14424896e-01 -1.56973079e-01 -1.12820196e+00
4.14027065e-01 -9.27113295e-01 2.94188380e-01 -1.39985526e+00
2.40195513e-01 -1.06299078e+00 -7.72585511e-01 5.19908428e-01
-5.58663607e-01 7.87717164e-01 4.87506241e-02 2.30980009e-01
-1.01167858e+00 8.70037496e-01 1.29925442e+00 -1.62480786e-01
1.71897002e-02 9.09275115e-02 -8.59344959e-01 8.32621276e-01
8.41089129e-01 -7.07655311e-01 -6.02463543e-01 -1.76999807e-01
1.60740584e-01 -7.34299362e-01 1.90944120e-01 -9.63932097e-01
3.46040502e-02 1.20442696e-01 1.69785723e-01 -4.99354333e-01
1.20846793e-01 -6.73439503e-01 -1.89052179e-01 1.11125685e-01
-7.43482351e-01 -2.36922234e-01 -1.08323500e-01 7.36944556e-01
-2.78231293e-01 -7.34566867e-01 9.17063653e-01 -2.90876981e-02
-5.81857860e-01 2.96744674e-01 -3.05326700e-01 3.91852885e-01
9.50008214e-01 -9.26620811e-02 -3.55658442e-01 1.54902026e-01
-9.20383394e-01 2.69472241e-01 2.66037643e-01 2.67563552e-01
3.87246460e-01 -1.53917003e+00 -5.49747765e-01 2.56900817e-01
3.81900102e-01 2.64215142e-01 3.48619193e-01 4.88796264e-01
-1.46387354e-01 3.80837061e-02 3.00235182e-01 -4.55534011e-01
-8.87805581e-01 5.40197611e-01 2.03062102e-01 -5.44756830e-01
-4.33884531e-01 1.33569837e+00 2.53977329e-01 -9.56559718e-01
5.82251489e-01 -4.63146031e-01 -2.96397477e-01 1.61139339e-01
5.77010036e-01 4.05847579e-01 3.97865057e-01 -3.43643695e-01
-2.13573277e-01 5.09457231e-01 -3.38361740e-01 3.79896492e-01
1.22303164e+00 -4.27330881e-01 4.52151299e-02 5.77436924e-01
1.34465408e+00 -1.58138312e-02 -1.56101787e+00 -5.74217975e-01
1.38362795e-01 -3.16419274e-01 2.10149363e-01 -7.67917275e-01
-1.04679310e+00 6.36605918e-01 8.73424113e-01 2.39177898e-01
1.20517576e+00 1.38411662e-02 8.68607998e-01 3.41746986e-01
2.12391123e-01 -1.11908102e+00 3.23956549e-01 5.08698523e-01
5.39434493e-01 -1.30650568e+00 -3.50484014e-01 -6.14394844e-01
-2.95399994e-01 7.97813892e-01 7.70069122e-01 -1.40336946e-01
7.43101954e-01 2.79742718e-01 2.74697334e-01 1.71134118e-02
-6.61241353e-01 -3.01548302e-01 -1.27532259e-02 6.02124035e-01
3.91346663e-02 -5.83057255e-02 -4.71104085e-01 9.87982035e-01
2.29864810e-02 -7.09877908e-02 2.87308842e-01 9.10374343e-01
-5.75322330e-01 -1.20999491e+00 -3.74709219e-01 4.62814808e-01
-3.30349714e-01 -1.96877703e-01 -2.39758119e-02 4.13882732e-01
6.59186661e-01 8.83297205e-01 9.93429031e-03 -1.60290450e-01
3.55542928e-01 1.11152738e-01 3.29240084e-01 -7.21674681e-01
-5.97823143e-01 4.63625230e-02 -1.64377853e-01 -3.07700098e-01
-4.90222603e-01 -3.77657473e-01 -1.05777359e+00 7.80111132e-03
-6.26537859e-01 2.47643948e-01 4.90593225e-01 9.41607058e-01
2.34268859e-01 6.67392731e-01 8.33796263e-01 -7.68171012e-01
-9.65977252e-01 -9.43729937e-01 -9.09322083e-01 4.63336736e-01
3.72873247e-01 -9.59600210e-01 -8.82965922e-01 -2.24048570e-01] | [9.358393669128418, 3.8900411128997803] |
0e596832-b5d4-4f5a-afa5-bccba903559e | revisiting-evaluation-of-knowledge-base | null | null | https://openreview.net/forum?id=1uufzxsxfL | https://openreview.net/pdf?id=1uufzxsxfL | Revisiting Evaluation of Knowledge Base Completion Models | Representing knowledge graphs (KGs) by learning embeddings for entities and relations has led to accurate models for existing KG completion benchmarks. However, due to the open-world assumption of existing KGs, evaluation of KG completion uses ranking metrics and triple classification with negative samples, and is thus unable to directly assess models on the goals of the task: completion. In this paper, we first study the shortcomings of these evaluation metrics. Specifically, we demonstrate that these metrics (1) are unreliable for estimating how calibrated the models are, (2) make strong assumptions that are often violated, and 3) do not sufficiently, and consistently, differentiate embedding methods from each other, or from simpler approaches. To address these issues, we gather a semi-complete KG referred as YAGO3-TC, using a random subgraph from the test and validation data of YAGO3-10, which enables us to compute accurate triple classification accuracy on this data. Conducting thorough experiments on existing models, we provide new insights and directions for the KG completion research. Along with the dataset and the open source implementation of the models, we also provide a leaderboard for knowledge graph completion that consists of a hidden, and growing, test set, available at https://pouyapez.github.io/yago3-tc/. | ['Sameer Singh', 'Yifan Tian', 'Pouya Pezeshkpour'] | 2020-02-14 | null | null | null | akbc-2020-6 | ['knowledge-base-completion', 'triple-classification', 'knowledge-base-completion'] | ['graphs', 'graphs', 'knowledge-base'] | [-3.56252342e-01 2.72402912e-01 -6.23027802e-01 -1.47528291e-01
-7.39499688e-01 -6.31028771e-01 5.79539239e-01 4.30953681e-01
-6.93008602e-02 8.10684979e-01 2.82346904e-01 -4.21140164e-01
-3.59729558e-01 -1.05745494e+00 -7.82009840e-01 -2.43642539e-01
-3.28272998e-01 5.61161101e-01 6.08259626e-02 -5.25650904e-02
-1.17735915e-01 -1.19215682e-01 -1.28654265e+00 1.41128048e-01
1.03106868e+00 1.06572711e+00 -3.93944055e-01 5.05284429e-01
1.83888689e-01 1.08047104e+00 -2.25777179e-01 -9.39745486e-01
2.49460220e-01 -1.39141664e-01 -1.08284104e+00 -2.61984229e-01
6.26883030e-01 -3.09447080e-01 -5.29880106e-01 9.25620675e-01
2.96700507e-01 -2.03310713e-01 8.85062456e-01 -1.74505031e+00
-1.10049284e+00 8.93483043e-01 -3.10918301e-01 -9.78165418e-02
3.17961454e-01 5.13832793e-02 1.61745644e+00 -1.08366394e+00
8.97157252e-01 9.70328689e-01 9.86596525e-01 9.57937762e-02
-1.22329926e+00 -4.90781695e-01 -3.32721844e-02 3.58474046e-01
-1.65980089e+00 -4.90792423e-01 5.14157295e-01 -4.94935125e-01
9.01405156e-01 1.38761044e-01 8.29343975e-01 1.20165861e+00
-2.22572207e-01 9.07075822e-01 9.59465563e-01 -3.79016042e-01
8.56743231e-02 1.99894994e-01 6.75498307e-01 1.09109771e+00
8.58990252e-01 4.75574955e-02 -5.80144167e-01 -3.48168373e-01
4.14378703e-01 -3.48430574e-01 -6.74902797e-01 -9.05047774e-01
-1.16158378e+00 8.96823406e-01 5.26077569e-01 -3.83952931e-02
-1.21803395e-01 2.85751909e-01 2.84381241e-01 4.21751112e-01
4.33499336e-01 4.33993459e-01 -5.56334555e-01 -1.96582481e-01
-5.17701149e-01 3.25359941e-01 1.26116312e+00 1.05743635e+00
9.56585824e-01 -3.99498940e-02 3.36335227e-02 6.53324902e-01
2.78592974e-01 1.90937921e-01 1.03752337e-01 -6.44791305e-01
6.02046907e-01 8.11880946e-01 -4.23357524e-02 -1.21161473e+00
-3.19343925e-01 -5.49371958e-01 -4.34648961e-01 -1.14771537e-01
6.05131924e-01 6.31464049e-02 -9.14361000e-01 1.71572411e+00
1.62942007e-01 1.22879602e-01 2.11650327e-01 6.65375173e-01
1.22886944e+00 8.19726139e-02 -2.07272265e-03 3.05438668e-01
1.25641644e+00 -1.04735970e+00 -4.86104220e-01 -8.60009268e-02
1.24259305e+00 -3.92335922e-01 1.25364769e+00 1.86854810e-01
-7.58124888e-01 -1.68791384e-01 -1.23307037e+00 -4.79395464e-02
-8.29020143e-01 4.43764441e-02 1.07880139e+00 5.90853572e-01
-1.10761809e+00 5.52998483e-01 -8.36376846e-01 -5.45750082e-01
3.13908905e-01 1.79558806e-02 -6.12093985e-01 -3.96347374e-01
-1.47254562e+00 1.01501286e+00 5.84686756e-01 -5.83580025e-02
-8.12038720e-01 -8.73696864e-01 -1.05981874e+00 7.49813020e-02
5.49480081e-01 -8.10645819e-01 9.36769545e-01 -5.19492865e-01
-7.88300693e-01 8.12015653e-01 3.79333198e-01 -2.32664093e-01
4.44957733e-01 -1.28802180e-01 -5.96067369e-01 -4.61521856e-02
7.51719996e-02 4.52021360e-01 2.46253535e-01 -1.36405194e+00
-2.97862530e-01 -4.30425674e-01 3.58852774e-01 1.07192308e-01
-5.56282580e-01 -5.06242156e-01 -8.08833063e-01 -4.66750890e-01
8.69408529e-03 -9.10612166e-01 1.41637743e-01 -4.27661508e-01
-7.18545377e-01 -1.98359579e-01 4.22301531e-01 -7.44679749e-01
1.46819150e+00 -1.99768841e+00 -3.15884091e-02 4.03483301e-01
7.07386911e-01 9.84803513e-02 -1.35280758e-01 6.77806735e-01
-9.92996544e-02 5.83558261e-01 -9.71700903e-03 -2.36235693e-01
1.59218922e-01 1.39159217e-01 -1.01166934e-01 4.29271579e-01
1.85245663e-01 1.33073688e+00 -1.03886950e+00 -5.40627599e-01
-7.42751062e-02 3.12991768e-01 -4.56800789e-01 -2.53170967e-01
-1.46419510e-01 -2.58541167e-01 -3.59093875e-01 1.02071285e+00
4.87787068e-01 -6.59270108e-01 4.41077918e-01 -6.56842828e-01
3.83816749e-01 3.28446299e-01 -1.15202367e+00 1.40445638e+00
-1.13254644e-01 4.38251466e-01 -3.48679394e-01 -8.87958407e-01
6.27107620e-01 4.51646820e-02 2.64100373e-01 -4.60464597e-01
-1.66560411e-02 2.69022793e-01 -2.26414308e-01 -2.41208076e-01
6.72138691e-01 3.48805010e-01 -2.88069751e-02 4.18941736e-01
3.04347873e-01 1.15555011e-01 3.72286618e-01 6.01531506e-01
1.47952533e+00 2.91093439e-02 2.57670462e-01 -8.32850263e-02
-1.24360561e-01 1.99132070e-01 6.39466822e-01 6.37630463e-01
-1.87277138e-01 5.33052027e-01 8.02876234e-01 -3.91686499e-01
-8.08212757e-01 -1.27742028e+00 -1.60307109e-01 6.58058286e-01
9.83925611e-02 -1.14274108e+00 -2.68853366e-01 -1.04625714e+00
5.72878122e-01 5.80350518e-01 -8.79971266e-01 -3.01146120e-01
3.92622203e-02 -8.00086439e-01 9.52880144e-01 7.67103553e-01
2.40509316e-01 -5.57377577e-01 2.22934276e-01 -2.83374954e-02
-2.19080582e-01 -1.16959929e+00 -1.40779376e-01 1.00170404e-01
-6.16155505e-01 -1.76052010e+00 -1.42740175e-01 -7.06398368e-01
6.66289926e-01 1.86419308e-01 1.63401377e+00 3.05252820e-01
-1.49742011e-02 6.62307799e-01 -6.30746186e-01 -2.15390608e-01
-2.96980411e-01 2.61972725e-01 8.06145817e-02 -3.00512284e-01
5.00170290e-01 -6.49160564e-01 -3.49791527e-01 1.95535123e-01
-7.45708287e-01 5.39835393e-02 6.44862950e-01 8.48596215e-01
4.51532543e-01 4.55273762e-02 6.48084462e-01 -1.24003696e+00
7.38227010e-01 -6.41994178e-01 -4.23982382e-01 6.13643348e-01
-1.08006144e+00 3.26340981e-02 3.69216174e-01 -6.90735802e-02
-4.47960258e-01 -3.88804585e-01 2.36392662e-01 -5.62444925e-01
4.21659559e-01 1.11920643e+00 -8.91835093e-02 -2.13270947e-01
8.88811588e-01 5.49543984e-02 -1.38532057e-01 -4.21925992e-01
6.64617658e-01 5.55364192e-01 5.21913886e-01 -7.26669729e-01
9.68041122e-01 2.34697327e-01 -3.90446782e-01 -5.25598168e-01
-8.30687642e-01 -2.69091666e-01 -4.22995925e-01 1.46971075e-02
3.44105512e-01 -1.17540157e+00 -5.08778572e-01 2.82260001e-01
-6.28133655e-01 -4.71520811e-01 -1.83677793e-01 4.81939375e-01
-3.42796296e-01 4.57242876e-01 -7.83481002e-01 -4.00459856e-01
-2.42360115e-01 -7.76223719e-01 5.88595748e-01 -9.40473005e-02
-2.28664503e-01 -1.36290610e+00 3.73227708e-02 5.68650842e-01
3.43261480e-01 4.71998692e-01 1.13463581e+00 -6.42659366e-01
-6.74625218e-01 -3.61615092e-01 -2.35502496e-01 4.14216518e-01
1.19641669e-01 3.07550341e-01 -9.80389178e-01 -3.76119107e-01
-8.18847895e-01 -9.53745186e-01 1.11499548e+00 -9.24297422e-03
9.69547093e-01 -2.38163114e-01 -4.61222321e-01 7.27189898e-01
1.48777986e+00 -6.31110787e-01 6.14349246e-01 2.43527874e-01
8.23376060e-01 2.55448431e-01 5.44336319e-01 2.66045481e-01
1.14737391e+00 5.23015082e-01 4.96393949e-01 3.16658877e-02
-1.77590847e-01 -7.88712859e-01 2.38953844e-01 1.06079161e+00
-1.02765530e-01 -3.91407669e-01 -1.04792953e+00 8.16451073e-01
-1.81461525e+00 -7.06060708e-01 -1.90511361e-01 2.19553328e+00
9.07483160e-01 1.03506595e-01 -2.40103975e-02 1.45520717e-01
4.36042815e-01 2.18919903e-01 -3.03923815e-01 4.15655747e-02
-2.78233021e-01 7.74426237e-02 6.92191482e-01 4.88717645e-01
-1.00619566e+00 1.03585565e+00 6.33812714e+00 8.02946150e-01
-6.68566585e-01 3.68531458e-02 3.60683829e-01 1.45186469e-01
-6.91194713e-01 3.85899931e-01 -7.05005407e-01 2.87172645e-01
7.78147876e-01 -4.76290822e-01 4.76834714e-01 7.52324283e-01
-4.67851907e-01 2.16540366e-01 -1.44751775e+00 8.97471726e-01
7.36425743e-02 -1.35538685e+00 9.03613642e-02 1.81637928e-01
7.02400863e-01 2.63396502e-01 -8.34647194e-02 8.40040088e-01
8.24292958e-01 -1.12175930e+00 5.37236691e-01 2.96493560e-01
8.82317305e-01 -4.65268582e-01 6.42759264e-01 1.04480743e-01
-1.21305776e+00 7.31403977e-02 -2.69117385e-01 7.67351761e-02
-1.49891570e-01 7.35340357e-01 -8.90135825e-01 1.11135399e+00
7.17615604e-01 8.37109864e-01 -1.05746114e+00 7.96509981e-01
-4.95567530e-01 8.56351554e-01 -3.30931813e-01 1.93533570e-01
-6.63560629e-02 -4.41599376e-02 1.77323297e-01 9.14086461e-01
8.85570571e-02 -2.43607223e-01 2.45273829e-01 8.84520471e-01
-4.28169847e-01 -3.97960423e-03 -8.87233555e-01 -4.97434586e-01
6.89790249e-01 1.42433035e+00 -3.70920271e-01 -1.98104531e-01
-6.87762141e-01 4.06723857e-01 9.58193600e-01 5.71386218e-01
-8.11642051e-01 -3.84784520e-01 7.09481657e-01 6.41664937e-02
1.29993677e-01 -2.57944822e-01 -5.25887236e-02 -1.48464429e+00
3.28920901e-01 -8.08224082e-01 7.09909081e-01 -6.84302390e-01
-1.47508943e+00 2.76659399e-01 -3.41184251e-02 -8.57655942e-01
-1.29608184e-01 -5.42073309e-01 -3.81658256e-01 5.06302238e-01
-1.54358876e+00 -1.42524147e+00 -6.03567660e-01 4.15472597e-01
-3.63123864e-01 -1.20591195e-02 9.33470547e-01 5.29394090e-01
-6.28289759e-01 9.31694627e-01 8.31306055e-02 4.54201043e-01
8.12370300e-01 -1.36509848e+00 5.06438732e-01 5.54502368e-01
4.17693377e-01 7.52823710e-01 4.81926054e-01 -7.75207758e-01
-1.54237700e+00 -1.14167655e+00 8.14868987e-01 -9.74497795e-01
9.62934077e-01 -3.71606648e-01 -9.73766327e-01 1.28408718e+00
-2.03516066e-01 2.94992954e-01 7.71727741e-01 8.02428901e-01
-5.44377208e-01 -6.15725294e-02 -8.47572803e-01 6.46842957e-01
1.30714762e+00 -4.20126736e-01 -4.09272164e-01 3.21897179e-01
7.13526785e-01 -3.56265396e-01 -1.38025069e+00 6.11078262e-01
5.36302090e-01 -9.01258826e-01 8.52887452e-01 -9.54678535e-01
3.91242206e-01 -1.98133171e-01 -2.92712778e-01 -1.57602847e+00
-3.41714680e-01 -5.45559190e-02 -5.49556911e-01 1.47728097e+00
7.71863401e-01 -8.94030988e-01 1.05920005e+00 6.85684860e-01
-1.41886339e-01 -1.07791567e+00 -5.76149821e-01 -8.89112592e-01
6.08150326e-02 -5.87903976e-01 5.75556695e-01 1.53912997e+00
2.07736209e-01 2.69007713e-01 -2.27332100e-01 2.35603049e-01
6.24794662e-01 9.18009877e-02 1.05089557e+00 -1.26444995e+00
-2.94448674e-01 -1.73074588e-01 -7.03711629e-01 -6.61906481e-01
2.58821249e-01 -1.32599843e+00 -5.35304964e-01 -1.83885932e+00
3.75619531e-01 -8.46818805e-01 -4.02613252e-01 1.05403781e+00
-2.65190870e-01 2.00681999e-01 -7.18635321e-02 2.52677560e-01
-6.51345611e-01 7.56447792e-01 9.04079378e-01 -2.25893080e-01
8.31638277e-02 -4.61219311e-01 -1.07876599e+00 4.91521776e-01
6.38195157e-01 -2.09890440e-01 -6.65891588e-01 -4.41590846e-01
6.13697946e-01 -1.25931427e-01 6.88908041e-01 -7.96991229e-01
1.20981537e-01 1.01431184e-01 3.14307541e-01 -3.03700089e-01
3.17802638e-01 -4.57262576e-01 2.41154984e-01 1.93481505e-01
-2.12864913e-02 3.73038650e-02 -7.39065185e-03 6.44617796e-01
-3.40698481e-01 -8.67151245e-02 1.61947936e-01 4.92870659e-02
-9.79717135e-01 4.62917268e-01 3.93457860e-01 4.85480368e-01
8.72392833e-01 -2.34824240e-01 -8.34022522e-01 -4.82074052e-01
-5.77816606e-01 7.91629314e-01 8.58332574e-01 4.93350714e-01
4.28434968e-01 -1.62610328e+00 -7.97224462e-01 -8.35656188e-03
7.60384440e-01 -6.58822879e-02 -9.33475420e-03 9.63252962e-01
-4.95958090e-01 2.52571881e-01 6.58206716e-02 -2.86532074e-01
-9.56179023e-01 4.84876186e-01 2.18447745e-01 -4.89428014e-01
-5.38130283e-01 5.50146997e-01 -5.47327884e-02 -6.55852020e-01
1.11575641e-01 -3.39025021e-01 -2.47509945e-02 -2.14190334e-02
8.87490213e-02 3.53712678e-01 2.33759016e-01 -3.02021444e-01
-4.32436496e-01 5.72412834e-02 -1.01383120e-01 2.52192527e-01
1.26852441e+00 2.22797692e-01 1.02026902e-01 3.07619274e-01
1.19906783e+00 1.91655591e-01 -6.17182016e-01 -3.33652854e-01
3.64654250e-02 -4.35712963e-01 -8.94146189e-02 -8.26112509e-01
-1.16892338e+00 3.98287654e-01 1.18594535e-01 3.27663124e-01
6.98241115e-01 1.76016137e-01 4.84342307e-01 4.45767015e-01
4.97396916e-01 -9.02212739e-01 -8.94232616e-02 3.95720452e-01
7.08060861e-01 -1.30890250e+00 2.58041203e-01 -6.06909215e-01
-6.58245623e-01 8.22237730e-01 7.53148198e-01 2.27556929e-01
6.62573695e-01 7.25426897e-02 -5.02717532e-02 -5.65114677e-01
-1.08521748e+00 -3.89470935e-01 3.40574682e-02 7.01335192e-01
4.47728366e-01 1.83089346e-01 -2.09370852e-01 7.59486258e-01
-4.11352336e-01 1.61821425e-01 5.34637928e-01 6.80822849e-01
-7.22866282e-02 -1.09730470e+00 5.32177500e-02 8.80962551e-01
-2.05898136e-01 -2.26427510e-01 -6.03348732e-01 1.32224631e+00
-1.90551534e-01 7.79883564e-01 -5.08427560e-01 -8.15495372e-01
4.01168525e-01 1.56472117e-01 5.14705718e-01 -4.93164420e-01
-1.27117783e-02 -7.70132720e-01 7.13825941e-01 -4.67317104e-01
2.26888880e-02 -3.28407526e-01 -8.89383256e-01 -6.78565443e-01
-5.30272841e-01 3.14345330e-01 2.58650601e-01 4.86119479e-01
4.97763991e-01 2.73152411e-01 2.75299072e-01 -3.63834262e-01
-8.31534386e-01 -1.08064115e+00 -9.44913149e-01 6.72123730e-01
-2.64201760e-01 -9.35652375e-01 -4.99543667e-01 -3.49827141e-01] | [8.809741973876953, 7.868429660797119] |
1130df8f-6737-4340-a732-45d0ce611d04 | wide-deep-learning-for-spatial-intensity | 2305.18708 | null | https://arxiv.org/abs/2305.18708v1 | https://arxiv.org/pdf/2305.18708v1.pdf | Wide & deep learning for spatial & intensity adaptive image restoration | Most existing deep learning-based image restoration methods usually aim to remove degradation with uniform spatial distribution and constant intensity, making insufficient use of degradation prior knowledge. Here we bootstrap the deep neural networks to suppress complex image degradation whose intensity is spatially variable, through utilizing prior knowledge from degraded images. Specifically, we propose an ingenious and efficient multi-frame image restoration network (DparNet) with wide & deep architecture, which integrates degraded images and prior knowledge of degradation to reconstruct images with ideal clarity and stability. The degradation prior is directly learned from degraded images in form of key degradation parameter matrix, with no requirement of any off-site knowledge. The wide & deep architecture in DparNet enables the learned parameters to directly modulate the final restoring results, boosting spatial & intensity adaptive image restoration. We demonstrate the proposed method on two representative image restoration applications: image denoising and suppression of atmospheric turbulence effects in images. Two large datasets, containing 109,536 and 49,744 images respectively, were constructed to support our experiments. The experimental results show that our DparNet significantly outperform SoTA methods in restoration performance and network efficiency. More importantly, by utilizing the learned degradation parameters via wide & deep learning, we can improve the PSNR of image restoration by 0.6~1.1 dB with less than 2% increasing in model parameter numbers and computational complexity. Our work suggests that degraded images may hide key information of the degradation process, which can be utilized to boost spatial & intensity adaptive image restoration. | ['Xiangzhi Bai', 'Yadong Wang'] | 2023-05-30 | null | null | null | null | ['image-restoration'] | ['computer-vision'] | [ 2.59545654e-01 -6.69704020e-01 3.48681569e-01 -2.29051113e-02
-2.88621426e-01 -2.94544965e-01 1.12158969e-01 -5.66687107e-01
-2.63118237e-01 8.29976022e-01 4.13794369e-01 -1.00377344e-01
-1.81553841e-01 -6.48210883e-01 -6.84909284e-01 -1.34572232e+00
3.05378642e-02 -6.57215953e-01 -9.74716805e-03 -4.11640346e-01
2.52055079e-01 5.58344424e-01 -1.28556609e+00 1.84687033e-01
1.06904352e+00 1.14832413e+00 5.76593816e-01 6.91174626e-01
2.37677723e-01 8.20085108e-01 -7.72007287e-01 2.37021208e-01
5.10449946e-01 -4.23739195e-01 -2.03845635e-01 3.60115141e-01
2.31625319e-01 -7.64750421e-01 -8.90556037e-01 1.13245058e+00
1.01356602e+00 2.99594462e-01 4.15884942e-01 -6.17499053e-01
-1.15627134e+00 8.84201080e-02 -8.50193977e-01 7.39989817e-01
-2.40528852e-01 6.36721313e-01 1.56958193e-01 -6.39904737e-01
1.50969595e-01 1.27798569e+00 6.76184714e-01 4.12704766e-01
-1.07138121e+00 -6.11381114e-01 2.35165972e-02 6.02080047e-01
-1.18330514e+00 -5.44582725e-01 9.85959888e-01 -9.81824100e-02
6.11397386e-01 2.00537711e-01 3.56182277e-01 7.22016931e-01
4.47692364e-01 4.43717360e-01 1.51526046e+00 -2.92245060e-01
-9.06095356e-02 -2.48438910e-01 6.00177981e-03 3.98532689e-01
1.91668063e-01 3.32301915e-01 -2.48026654e-01 2.97998548e-01
1.14973485e+00 2.09951848e-02 -8.81479561e-01 1.41095161e-01
-1.07305694e+00 3.71826470e-01 7.11344898e-01 1.17084481e-01
-5.08723736e-01 7.95528218e-02 4.82529640e-01 5.72921813e-01
4.14316595e-01 5.84315099e-02 -5.30939162e-01 3.19761574e-01
-4.14962381e-01 -1.71813563e-01 5.81883602e-02 4.96319026e-01
6.69793606e-01 6.04111314e-01 -1.79168448e-01 1.10710001e+00
2.23615065e-01 8.50496590e-01 6.66880727e-01 -1.13764966e+00
1.68550134e-01 1.18677489e-01 3.70032042e-01 -1.11072385e+00
-1.28812209e-01 -7.04719305e-01 -1.54706681e+00 5.10168016e-01
3.19814845e-03 -2.24326804e-01 -1.07972050e+00 1.55034566e+00
1.81563944e-01 2.85915196e-01 2.22051874e-01 1.26606178e+00
5.98305523e-01 9.45045948e-01 -1.05899833e-01 -8.27124774e-01
1.15802443e+00 -9.61323261e-01 -1.03872144e+00 -1.62845090e-01
-1.21909805e-01 -1.02426720e+00 1.28950870e+00 5.91990292e-01
-1.02390003e+00 -1.03188992e+00 -1.29228520e+00 -4.60254923e-02
1.56864434e-01 2.43029773e-01 4.68517512e-01 5.27935207e-01
-1.15878487e+00 7.30131984e-01 -5.94510198e-01 7.64897242e-02
5.71212590e-01 3.17162901e-01 -2.59665519e-01 -5.29813826e-01
-1.18834937e+00 7.04235971e-01 2.14490235e-01 6.64057970e-01
-1.27730274e+00 -6.18030250e-01 -4.81464207e-01 -1.69436429e-02
5.24838716e-02 -8.30429494e-01 6.64831221e-01 -1.13868260e+00
-1.59466577e+00 3.38793635e-01 1.69008318e-03 -2.63097376e-01
3.43832523e-01 -4.15427744e-01 -7.06541359e-01 3.97369832e-01
-1.24815248e-01 1.72828779e-01 1.27410924e+00 -1.51139057e+00
-3.60596865e-01 -2.47436926e-01 -1.13887392e-01 3.62020791e-01
-3.95269305e-01 -9.10322145e-02 -2.13063091e-01 -8.38098466e-01
1.57771721e-01 -3.75255615e-01 -1.08139999e-01 2.22115144e-01
-1.66592300e-01 3.31224650e-01 1.17501020e+00 -1.21456647e+00
9.72150862e-01 -2.24432278e+00 1.61447972e-01 -2.21882507e-01
5.75283356e-02 5.59557974e-01 -3.78982723e-01 2.04450488e-01
-2.34222367e-01 -5.04251793e-02 -3.64595473e-01 2.34455592e-03
-4.11479473e-01 3.59013349e-01 -4.21758771e-01 8.50754261e-01
-1.17375247e-01 4.74556565e-01 -5.04882514e-01 -1.74332768e-01
5.65064311e-01 8.45052660e-01 -2.27466196e-01 4.67221886e-01
9.56723541e-02 7.32054949e-01 -1.96781293e-01 5.95514715e-01
1.20511210e+00 6.93468973e-02 -3.69347855e-02 -8.39810610e-01
-1.09412059e-01 -3.74712020e-01 -1.10635686e+00 1.54884946e+00
-6.78882301e-01 6.03984654e-01 4.96887624e-01 -1.24212861e+00
8.85648727e-01 1.76755235e-01 2.93103874e-01 -1.09409416e+00
1.03190400e-01 1.45425618e-01 -1.35618359e-01 -8.70221794e-01
-1.04779162e-01 -3.64199907e-01 5.14867425e-01 5.00961691e-02
-2.68722117e-01 1.60584718e-01 -1.22491397e-01 -1.43699259e-01
9.72661138e-01 1.69728864e-02 -1.31144583e-01 -3.62071604e-01
9.45828736e-01 -3.39960098e-01 7.70081460e-01 7.55220056e-01
-3.96892101e-01 5.58163226e-01 1.09838741e-02 -5.13827503e-01
-1.13038290e+00 -1.06721294e+00 -1.27235368e-01 8.23853910e-01
5.33296824e-01 2.98841149e-01 -5.62874913e-01 -2.06808038e-02
-3.29043239e-01 3.25442433e-01 -3.30816060e-01 -2.92717785e-01
-9.29654419e-01 -1.25562310e+00 1.62203208e-01 1.63310423e-01
1.16931939e+00 -1.01403105e+00 -6.46360815e-02 1.32851258e-01
-1.64018214e-01 -8.22458029e-01 -4.95654732e-01 -1.09660052e-01
-1.08037555e+00 -9.85595286e-01 -7.74722159e-01 -9.24026430e-01
6.73822880e-01 7.70732164e-01 8.28069329e-01 2.77189165e-01
-3.29841673e-01 1.42706990e-01 -2.00146779e-01 1.73726484e-01
-2.05881238e-01 -6.68420315e-01 2.53205210e-01 1.88568890e-01
-2.40974650e-01 -1.20841169e+00 -1.30088866e+00 3.27062219e-01
-1.16569018e+00 -4.54087276e-03 8.83128107e-01 1.16971874e+00
4.89606082e-01 6.76528990e-01 6.78288400e-01 -2.42996618e-01
7.33051419e-01 -2.51954377e-01 -4.58624154e-01 1.62472893e-02
-5.95202446e-01 -1.57425642e-01 8.82623732e-01 -4.35024530e-01
-1.48239982e+00 -4.11354810e-01 -6.07902035e-02 -5.79159737e-01
-9.64465067e-02 4.33698088e-01 -4.00622815e-01 -4.11136240e-01
6.37730062e-01 7.26360559e-01 2.70314664e-01 -8.27063680e-01
2.58898944e-01 6.77426755e-01 9.72717881e-01 -4.66416478e-01
8.05534482e-01 5.76143622e-01 5.28052375e-02 -7.53833830e-01
-5.62063217e-01 -2.01008454e-01 -2.23629668e-01 -2.75108397e-01
5.93668580e-01 -1.33305705e+00 -8.42110693e-01 1.13226831e+00
-9.06415462e-01 -3.74959886e-01 4.36212234e-02 2.92473823e-01
-3.13474476e-01 8.48993182e-01 -1.16579866e+00 -4.95955139e-01
-6.28391922e-01 -1.12741244e+00 5.36752343e-01 4.89413142e-01
8.89620245e-01 -8.59750450e-01 -4.18688595e-01 5.32764316e-01
7.95437515e-01 1.99999124e-01 1.00292599e+00 3.78840983e-01
-7.22191274e-01 1.38953060e-01 -5.74645758e-01 1.07464325e+00
5.27816594e-01 -5.11694133e-01 -9.08589482e-01 -7.45580375e-01
4.94703680e-01 -1.14023723e-02 1.04925561e+00 7.36343920e-01
1.35424781e+00 -4.01457995e-01 1.11599714e-01 9.21477735e-01
1.86469603e+00 3.31432104e-01 1.16538882e+00 5.83495796e-01
7.22752869e-01 2.29190886e-01 3.90491456e-01 4.09421563e-01
-1.13089815e-01 2.89520979e-01 6.78488851e-01 -4.32987034e-01
-8.07978153e-01 2.75252879e-01 5.21037757e-01 1.15362835e+00
-1.65762365e-01 -2.56253034e-01 -4.86187398e-01 6.96476817e-01
-1.46884000e+00 -7.09649444e-01 -1.71176150e-01 1.88083708e+00
9.23414469e-01 -1.72203153e-01 -5.82657278e-01 1.63079783e-01
8.58299077e-01 2.66089708e-01 -8.63355279e-01 8.34874064e-02
-6.18733227e-01 1.05045639e-01 6.10441983e-01 5.56452334e-01
-9.52315509e-01 5.86787939e-01 5.65691328e+00 9.40943122e-01
-1.22996306e+00 2.16616452e-01 9.44203019e-01 -1.46265149e-01
-3.38064507e-02 -2.19594255e-01 -2.14576289e-01 5.70238113e-01
7.12459326e-01 2.76322607e-02 7.68925071e-01 3.18403542e-01
9.29124773e-01 -9.35031101e-02 -2.65881598e-01 1.17024899e+00
-7.62317553e-02 -1.14285517e+00 1.01146311e-01 -1.92297235e-01
7.82571256e-01 -1.16988525e-01 1.53884336e-01 1.98569372e-02
1.02747284e-01 -9.98518109e-01 2.63591260e-01 7.91911721e-01
7.99779892e-01 -8.41134787e-01 8.65440011e-01 1.29039899e-01
-6.79560781e-01 -4.28152800e-01 -6.08566701e-01 -2.16133699e-01
3.02331299e-01 8.98092508e-01 -3.77074331e-02 6.82020724e-01
9.88433540e-01 8.42967987e-01 -2.92034477e-01 9.39808071e-01
-3.27150583e-01 7.73794234e-01 3.13210003e-02 7.05323398e-01
-4.50438298e-02 -4.99563515e-01 5.85762322e-01 9.59786534e-01
4.34427053e-01 5.04159510e-01 7.40267485e-02 3.62452120e-01
4.78271255e-03 -2.34977946e-01 -1.11532651e-01 4.17164654e-01
1.85119271e-01 1.12766039e+00 -2.33216390e-01 -3.55847836e-01
-2.19343066e-01 1.22642994e+00 -3.07225972e-01 7.78921664e-01
-7.20015168e-01 -3.70163321e-01 9.29598093e-01 -6.56154826e-02
4.83170509e-01 -2.99970776e-01 -3.19275677e-01 -1.10734558e+00
1.19622849e-01 -1.08710182e+00 -2.71897782e-02 -1.16038036e+00
-1.41746485e+00 6.24530613e-01 -3.68938565e-01 -1.13966787e+00
4.94542629e-01 -6.78607285e-01 -5.04711092e-01 1.24547601e+00
-2.01898265e+00 -1.09230590e+00 -5.77173054e-01 8.20310414e-01
6.11703396e-01 -2.29144782e-01 3.16129655e-01 5.96096516e-01
-9.87559974e-01 1.23551026e-01 6.50559664e-01 -1.72694281e-01
8.95573080e-01 -9.40025032e-01 -1.23006687e-01 1.31603825e+00
-7.72861421e-01 5.13353765e-01 1.00148165e+00 -5.43469727e-01
-1.63327205e+00 -1.09933972e+00 -6.25992864e-02 3.20372820e-01
4.31095660e-01 2.19932839e-01 -1.04227829e+00 2.12755203e-01
5.48051119e-01 3.80435109e-01 2.32988179e-01 -5.33265591e-01
-1.14948034e-01 -8.04230273e-01 -1.27400255e+00 5.00772238e-01
9.18173730e-01 -3.87866288e-01 -1.68277696e-01 5.00725091e-01
9.46297288e-01 -3.56171966e-01 -1.01697969e+00 5.19420743e-01
2.66445637e-01 -1.00855243e+00 1.32083428e+00 -2.22084299e-01
5.08851647e-01 -8.45500112e-01 -3.50544035e-01 -1.39791346e+00
-6.18578613e-01 -5.85999727e-01 -1.97626308e-01 1.18665600e+00
-1.46292999e-01 -6.37630522e-01 2.79812068e-01 9.37066153e-02
-6.38074696e-01 -5.31250894e-01 -8.08257401e-01 -5.83376706e-01
-1.10923387e-01 9.33277234e-02 3.86938214e-01 8.65160942e-01
-9.37143743e-01 5.13525568e-02 -9.67431128e-01 8.03239465e-01
9.26652670e-01 -6.06955215e-02 4.67618227e-01 -6.62938237e-01
-2.57395297e-01 -3.90735939e-02 -1.72458336e-01 -1.29656279e+00
-8.96485671e-02 -1.71202093e-01 5.97777776e-02 -1.53712559e+00
1.71631128e-01 -1.87983125e-01 -7.32432246e-01 3.53636831e-01
-3.40312809e-01 5.33409536e-01 -1.18835539e-01 4.78878587e-01
-1.12914421e-01 8.02086592e-01 1.80579185e+00 -3.70077133e-01
1.74368806e-02 -3.07832927e-01 -8.82326782e-01 3.79725754e-01
8.16151977e-01 -4.78996523e-02 -5.96321702e-01 -9.24042463e-01
-2.93510616e-01 1.61857337e-01 4.87061203e-01 -1.08484519e+00
1.43446743e-01 -3.17554712e-01 7.91846693e-01 -1.58165261e-01
2.52260387e-01 -6.59813046e-01 2.31305704e-01 4.99434501e-01
1.30103067e-01 -2.07992554e-01 4.56036448e-01 7.47600794e-01
-3.94254833e-01 4.19980809e-02 1.05594158e+00 -2.67826676e-01
-1.03984797e+00 2.75231570e-01 -2.33105734e-01 -6.22822106e-01
6.47551060e-01 -2.38003746e-01 -5.64526975e-01 -2.97037870e-01
-6.80732012e-01 -8.90952423e-02 2.89297253e-01 1.10457160e-01
8.28638613e-01 -1.26742899e+00 -8.84927750e-01 2.31963500e-01
-5.54711163e-01 -1.12664267e-01 1.00631583e+00 7.07212150e-01
-8.85309696e-01 3.25408718e-03 -5.35604179e-01 -4.93680328e-01
-1.11312973e+00 7.53567815e-01 6.54204786e-01 7.58626536e-02
-8.31225157e-01 6.78905249e-01 3.53371561e-01 -9.99523699e-03
-3.40912864e-02 -6.27743825e-02 -1.22329168e-01 -4.38027561e-01
8.19142997e-01 5.37086546e-01 -2.04304904e-02 -4.53165293e-01
-1.30058359e-03 5.76252759e-01 -1.61805078e-01 1.84555039e-01
1.65684390e+00 -8.77371252e-01 -5.36439896e-01 -1.52429342e-01
1.03852642e+00 -5.93662448e-02 -1.83116257e+00 -2.08123088e-01
-6.76492810e-01 -7.54367769e-01 7.19731033e-01 -1.18621182e+00
-1.59388995e+00 7.31215596e-01 1.31318331e+00 -1.44207552e-02
2.12811351e+00 -6.79475307e-01 9.76672292e-01 2.73708142e-02
1.17924204e-02 -7.37716198e-01 3.50451648e-01 1.51595667e-01
1.14860296e+00 -1.20214939e+00 2.11206585e-01 -2.61754960e-01
-2.14333266e-01 1.17977369e+00 7.29227781e-01 -1.91566080e-01
3.58296633e-01 1.40826516e-02 4.10721630e-01 1.03129423e-03
-4.25620645e-01 3.05040061e-01 -9.70937610e-02 7.76050329e-01
9.27941948e-02 -2.06902891e-01 -3.09952080e-01 3.59924406e-01
4.68372852e-01 -1.52870016e-02 6.40746951e-01 6.51055574e-01
-6.78145587e-01 -9.37044382e-01 -7.41650879e-01 2.23511100e-01
-4.94300932e-01 -2.70244956e-01 5.00179172e-01 4.08135116e-01
3.38477194e-01 1.15781748e+00 -2.59953409e-01 -2.11333618e-01
2.47487903e-01 -6.10651851e-01 4.06855285e-01 1.77011881e-02
-1.25374630e-01 2.57951140e-01 -1.88934535e-01 -3.66930902e-01
-6.05980396e-01 -3.43779147e-01 -6.59703314e-01 -4.37916815e-01
-2.44764790e-01 -9.65846479e-02 5.16531229e-01 8.76325130e-01
4.12252218e-01 8.57460022e-01 8.78194690e-01 -8.75431836e-01
-5.00163972e-01 -1.15853727e+00 -7.40402877e-01 3.70646119e-01
8.14436197e-01 -4.68460351e-01 -6.55509889e-01 3.52481782e-01] | [11.089553833007812, -2.443153142929077] |
c1fb6a1b-9a9c-45f9-a890-aed475912d84 | complex-valued-time-frequency-self-attention | 2211.12632 | null | https://arxiv.org/abs/2211.12632v1 | https://arxiv.org/pdf/2211.12632v1.pdf | Complex-Valued Time-Frequency Self-Attention for Speech Dereverberation | Several speech processing systems have demonstrated considerable performance improvements when deep complex neural networks (DCNN) are coupled with self-attention (SA) networks. However, the majority of DCNN-based studies on speech dereverberation that employ self-attention do not explicitly account for the inter-dependencies between real and imaginary features when computing attention. In this study, we propose a complex-valued T-F attention (TFA) module that models spectral and temporal dependencies by computing two-dimensional attention maps across time and frequency dimensions. We validate the effectiveness of our proposed complex-valued TFA module with the deep complex convolutional recurrent network (DCCRN) using the REVERB challenge corpus. Experimental findings indicate that integrating our complex-TFA module with DCCRN improves overall speech quality and performance of back-end speech applications, such as automatic speech recognition, compared to earlier approaches for self-attention. | ['John H. L. Hansen', 'Vinay Kothapally'] | 2022-11-22 | null | null | null | null | ['speech-dereverberation'] | ['speech'] | [ 1.06008612e-01 1.38808368e-02 3.49314421e-01 -3.39717031e-01
-6.92308128e-01 -2.70108640e-01 5.85371912e-01 -9.91984531e-02
-2.73255736e-01 1.54847950e-01 7.96589911e-01 -7.82471061e-01
-2.33443361e-02 -2.79128879e-01 -5.09434938e-01 -4.67595011e-01
-1.71489254e-01 -1.74363077e-01 -1.11703485e-01 -5.11654198e-01
9.39036012e-02 5.27549982e-01 -1.54350483e+00 3.64011288e-01
7.80696154e-01 9.95573759e-01 2.72198856e-01 1.17103207e+00
2.79148340e-01 7.56170750e-01 -1.03749549e+00 8.51846300e-03
-2.09814206e-01 -4.84315872e-01 -8.13695669e-01 -3.86821151e-01
2.35248402e-01 2.74515338e-02 -3.48574311e-01 8.94375980e-01
1.00793791e+00 6.96027100e-01 4.52102572e-01 -8.16962063e-01
-9.85988736e-01 8.28122079e-01 -1.03713693e-02 1.10361648e+00
3.03184003e-01 1.61469012e-01 1.13157284e+00 -1.11351442e+00
4.13298979e-03 1.48787820e+00 8.93314242e-01 5.33810139e-01
-1.11334133e+00 -5.95053434e-01 2.72986859e-01 5.65862179e-01
-1.07179224e+00 -8.89428616e-01 7.72033334e-01 -3.67991254e-02
2.14366102e+00 4.12087351e-01 4.27875042e-01 1.22571063e+00
1.65848836e-01 6.58276856e-01 5.41592240e-01 -6.70748591e-01
3.84532027e-02 -3.54681671e-01 4.96594042e-01 1.17042802e-01
-5.44664562e-01 3.65096509e-01 -6.55101180e-01 2.15956584e-01
4.19860095e-01 -5.82826972e-01 -3.60263735e-01 6.20363653e-01
-1.04599094e+00 6.71135306e-01 5.76687694e-01 6.80703878e-01
-5.67560494e-01 2.74798870e-01 7.05055416e-01 4.06001478e-01
8.89724970e-01 5.06941020e-01 -4.70296144e-01 -3.44521582e-01
-6.02358520e-01 -2.51967520e-01 4.12138104e-01 6.84843302e-01
1.82592899e-01 9.25423563e-01 -5.38220406e-01 1.21663022e+00
1.59141570e-01 5.05546093e-01 1.03233755e+00 -5.99423289e-01
3.20618242e-01 -2.44112313e-01 -4.06943947e-01 -7.83438325e-01
-5.13299286e-01 -8.17649126e-01 -8.01104963e-01 2.06468962e-02
-1.56811133e-01 -3.24320853e-01 -9.60364401e-01 1.82159388e+00
-1.13967456e-01 2.29692549e-01 4.23182994e-01 9.71716702e-01
8.30229044e-01 9.18650031e-01 9.51105580e-02 -1.79828748e-01
1.16919851e+00 -1.17147458e+00 -1.45571077e+00 -1.91545144e-01
3.05505782e-01 -8.74065459e-01 1.15706694e+00 1.90869257e-01
-1.25556099e+00 -8.55177641e-01 -1.20756269e+00 -4.37644452e-01
-5.33424854e-01 1.77003875e-01 2.97718227e-01 8.51849139e-01
-1.50366282e+00 5.73292375e-01 -6.85400188e-01 -2.34219745e-01
3.97521481e-02 3.67265046e-01 1.69773251e-02 6.50184751e-01
-1.65111172e+00 9.95216250e-01 -2.64230873e-02 4.11326081e-01
-7.39602864e-01 -8.66501451e-01 -8.08381021e-01 5.58615148e-01
-1.07776463e-01 -3.78022492e-01 1.62813330e+00 -1.02196991e+00
-2.00863338e+00 1.93864793e-01 -3.93547356e-01 -9.48193669e-01
-2.86599338e-01 -6.09529853e-01 -9.72692192e-01 5.93977123e-02
-2.32367724e-01 4.56369877e-01 1.03797925e+00 -5.93795776e-01
-2.62556314e-01 2.44190861e-02 -2.62227178e-01 2.93560117e-01
-4.22001392e-01 2.16318935e-01 1.05331108e-01 -1.14525819e+00
-3.23212631e-02 -6.15194559e-01 5.27378283e-02 -6.06040776e-01
-4.18916017e-01 -5.20932794e-01 1.14066601e+00 -7.03387320e-01
1.57601607e+00 -2.55492640e+00 1.11788481e-01 -4.46841121e-02
-1.62462384e-01 9.35937941e-01 -3.22102368e-01 3.09643030e-01
-6.12240374e-01 1.99002668e-01 8.18141326e-02 -6.00681901e-01
6.97028711e-02 -2.05232166e-02 -5.53402901e-01 2.13112101e-01
7.75820553e-01 7.94028461e-01 -7.23286688e-01 3.79753225e-02
3.47509384e-01 1.04202068e+00 -8.58181834e-01 1.84023038e-01
-6.62840530e-02 1.23971760e-01 2.60141224e-01 1.67900726e-01
3.82546484e-01 1.84701487e-01 -1.35318041e-01 -3.16666692e-01
-4.07103419e-01 1.15986907e+00 -4.02224600e-01 1.44942820e+00
-8.55633497e-01 1.38634348e+00 4.96943370e-02 -7.11931407e-01
8.52427900e-01 9.31640983e-01 1.64766014e-01 -9.25421357e-01
2.30297148e-01 2.41737608e-02 5.50527930e-01 -3.20266724e-01
5.28491020e-01 -1.66275315e-02 4.30580765e-01 3.56886804e-01
4.68986541e-01 -9.26288962e-02 -2.28857905e-01 -2.76904413e-03
1.10617876e+00 -4.16245759e-01 -1.78541895e-02 -3.15049440e-01
7.14431703e-01 -6.36309028e-01 3.15287441e-01 4.53353047e-01
-3.49973619e-01 7.93625355e-01 2.28200555e-01 -1.69504464e-01
-1.00268316e+00 -9.82574284e-01 -9.04206559e-02 1.49168420e+00
-5.74847519e-01 -3.80558729e-01 -1.01959193e+00 -1.48904130e-01
-2.89704740e-01 8.59159887e-01 -5.04359901e-01 -3.57075810e-01
-5.62388241e-01 -4.40428615e-01 8.78596246e-01 6.55255497e-01
1.49400488e-01 -1.46878374e+00 -3.24191689e-01 5.03495395e-01
-3.00358832e-01 -1.10743344e+00 -1.02954888e+00 5.04649937e-01
-3.78780305e-01 -3.83026212e-01 -6.87211931e-01 -7.58413255e-01
4.94855829e-02 3.82769972e-01 7.32518077e-01 -8.06856602e-02
-1.28534809e-01 2.89836079e-01 -5.58602691e-01 -4.22757506e-01
-3.39357525e-01 4.23911244e-01 2.65679985e-01 1.92316547e-01
2.68766880e-01 -8.38721633e-01 -3.98592770e-01 -4.41660173e-02
-6.00816071e-01 -3.05535585e-01 4.15641159e-01 8.39187205e-01
9.24327523e-02 -1.05208077e-01 1.31430340e+00 -3.87165338e-01
1.20064247e+00 -4.01125789e-01 -2.73666918e-01 -6.10979684e-02
-4.11440998e-01 1.80540040e-01 6.40895545e-01 -6.52203083e-01
-1.18996358e+00 -4.21212524e-01 -6.39589131e-01 -6.47698462e-01
1.04497656e-01 6.05283141e-01 5.57299033e-02 3.58905703e-01
7.02043891e-01 -6.81821778e-02 -2.61806138e-02 -3.02917868e-01
2.93102354e-01 1.07367432e+00 6.22214794e-01 -1.47683039e-01
2.94254392e-01 -2.59598911e-01 -5.06486595e-01 -1.27767694e+00
-6.75398350e-01 -5.01603425e-01 -3.29653054e-01 -1.66525528e-01
9.59267855e-01 -7.44606078e-01 -8.19688320e-01 5.24358213e-01
-1.51873481e+00 -3.98807436e-01 -3.02316844e-01 8.64593685e-01
-4.96908963e-01 6.85405806e-02 -7.01286495e-01 -1.18138719e+00
-8.01972270e-01 -1.11092496e+00 9.06707942e-01 3.24101336e-02
-3.33175480e-01 -8.98979723e-01 -5.43003380e-02 1.51116267e-01
1.22156978e+00 -4.59738821e-01 9.35727656e-01 -6.07392848e-01
-1.27470747e-01 2.55643338e-01 -6.69860020e-02 6.69487298e-01
1.71215177e-01 -4.21364158e-02 -1.74242175e+00 -1.32511079e-01
5.57436831e-02 5.04433326e-02 7.20076919e-01 6.33412302e-01
1.06643224e+00 -3.72700393e-01 2.27738172e-01 4.72499162e-01
6.86593950e-01 6.43949628e-01 6.64127886e-01 -8.92750993e-02
4.40895200e-01 3.36788923e-01 6.96853101e-02 3.09642881e-01
1.53671980e-01 6.53306782e-01 1.15580358e-01 -2.80181646e-01
-6.93467617e-01 1.18678324e-01 5.95549107e-01 1.74397099e+00
-3.17288423e-03 -4.22287226e-01 -8.76714826e-01 8.56622159e-01
-1.44824255e+00 -1.08529425e+00 6.81886077e-02 1.69096863e+00
8.70725393e-01 2.09447935e-01 -1.10389501e-01 5.78638375e-01
8.16514969e-01 2.11988404e-01 -6.04474843e-01 -1.11198843e+00
-1.84177414e-01 6.41488314e-01 -2.47592553e-02 8.44208419e-01
-9.59281027e-01 9.22397316e-01 6.55483961e+00 7.64822960e-01
-1.52121389e+00 3.50996166e-01 4.20088798e-01 -1.59571007e-01
-1.60124570e-01 -5.74080825e-01 -4.57315803e-01 9.48140100e-02
1.74561644e+00 -2.83092439e-01 7.69987345e-01 5.77406406e-01
4.44238305e-01 4.84068573e-01 -1.00996304e+00 6.59254014e-01
-1.12252729e-02 -1.04171634e+00 -2.49804407e-01 -1.43139601e-01
4.47520286e-01 3.52974892e-01 5.78151762e-01 5.13115168e-01
1.03835285e-01 -1.19135559e+00 7.25998819e-01 3.35460156e-01
8.41871679e-01 -9.73881543e-01 5.69275558e-01 -1.43110842e-01
-1.29821599e+00 -1.90324083e-01 1.57784391e-02 -6.71689771e-03
-1.26401847e-03 4.38362658e-01 -1.19000506e+00 3.46733600e-01
8.09229195e-01 6.45099580e-01 -2.55634606e-01 6.36239469e-01
1.46911861e-02 9.18092847e-01 -1.26563236e-01 -1.90293804e-01
2.45370463e-01 2.80405819e-01 6.63885474e-01 1.50806844e+00
1.75232187e-01 1.94050148e-01 -7.23320484e-01 6.89494371e-01
-5.22676855e-02 -1.15391821e-01 -2.91807443e-01 -4.57339257e-01
5.81257522e-01 7.02531755e-01 -2.39674285e-01 -1.32226542e-01
-2.80392349e-01 8.30342472e-01 3.89094025e-01 6.63972139e-01
-7.65539110e-01 -8.66292000e-01 1.27321255e+00 -2.25997135e-01
6.69082224e-01 -4.17652637e-01 -2.36078605e-01 -7.80671418e-01
-1.23525880e-01 -9.26156044e-01 -6.57545328e-02 -1.21773291e+00
-1.23614299e+00 1.40444624e+00 -4.67637986e-01 -8.59063208e-01
-5.09001672e-01 -5.40279031e-01 -6.48948848e-01 1.23420298e+00
-1.53354883e+00 -7.91503668e-01 2.86698461e-01 6.24765038e-01
9.84045982e-01 -2.18803808e-01 1.07396924e+00 4.18223113e-01
-6.68522835e-01 8.98372889e-01 -1.78157434e-01 1.77392121e-02
3.23684365e-01 -1.11419320e+00 1.05190659e+00 1.05392253e+00
1.28389806e-01 7.16480851e-01 5.29157221e-01 -2.17196956e-01
-1.15265822e+00 -1.07386780e+00 1.07344067e+00 -8.99908766e-02
6.32273257e-01 -5.50225735e-01 -1.16079092e+00 5.93317151e-01
8.28875601e-01 2.37546220e-01 4.58543956e-01 3.55828285e-01
-4.97606367e-01 -1.51832193e-01 -6.85755491e-01 5.21678567e-01
8.78863454e-01 -1.35273337e+00 -6.08148992e-01 1.26730815e-01
1.38739169e+00 -2.23390922e-01 -7.09691763e-01 3.45421135e-01
2.77020603e-01 -8.41448843e-01 1.08731520e+00 -4.69423473e-01
1.46889286e-02 -1.13812231e-01 -2.81360120e-01 -1.79775631e+00
-3.80855471e-01 -9.35861647e-01 -9.48545709e-02 1.19497502e+00
6.12368584e-01 -6.76391840e-01 -1.63165424e-02 -6.73109852e-03
-9.39545333e-01 -3.52107227e-01 -1.31923425e+00 -7.04298258e-01
2.67867837e-02 -7.69451022e-01 6.48856103e-01 8.80926192e-01
2.42025182e-01 6.30562365e-01 -1.49276838e-01 5.38137257e-01
-9.86734703e-02 -5.81016481e-01 1.05527952e-01 -8.83097589e-01
-1.15465008e-01 -7.14984238e-01 -3.69846106e-01 -8.20572078e-01
3.22029859e-01 -6.99470162e-01 2.90110260e-01 -1.19192326e+00
-7.29517937e-01 -9.71871912e-02 -4.91037041e-01 3.75471354e-01
-4.48250137e-02 -1.34867802e-01 1.57833874e-01 -3.39485079e-01
-1.65916711e-01 8.70151043e-01 8.92589986e-01 -4.42511067e-02
-2.16190130e-01 -5.98485246e-02 -2.73658007e-01 2.27735117e-01
8.22297633e-01 -1.29165262e-01 -5.82637012e-01 -6.33566439e-01
-2.57513821e-01 1.64512068e-01 6.88097104e-02 -1.37636673e+00
6.06603742e-01 4.27625030e-01 2.67862417e-02 -6.74972832e-01
6.73700273e-01 -5.29610991e-01 -2.37146035e-01 1.54916078e-01
-6.45653307e-01 3.13774139e-01 7.15055346e-01 3.46190661e-01
-4.83616561e-01 1.95775598e-01 7.13826478e-01 3.15844655e-01
-2.71705598e-01 -1.74965158e-01 -1.03514659e+00 -6.21718094e-02
2.72401243e-01 1.86756909e-01 -4.51903671e-01 -6.53846025e-01
-9.30051267e-01 -2.86012292e-01 -7.03018785e-01 7.75463164e-01
7.15287149e-01 -1.43588579e+00 -5.99123955e-01 8.20609331e-01
-3.00223917e-01 -3.58642161e-01 4.04866248e-01 6.66787446e-01
-1.10840574e-01 8.74298573e-01 -1.68098733e-02 -5.04558861e-01
-1.08033991e+00 5.11347175e-01 6.99445903e-01 9.53940824e-02
-4.97853339e-01 9.63578522e-01 1.61790177e-01 -3.67974877e-01
6.91680133e-01 -9.62680399e-01 -2.69834012e-01 5.38629480e-02
5.47101438e-01 2.61577159e-01 6.84485435e-01 -8.21835816e-01
-3.40570927e-01 2.06418753e-01 -1.91342235e-01 -5.07650733e-01
1.25292289e+00 -2.48836681e-01 3.11835378e-01 5.95638990e-01
1.28076637e+00 -2.36730903e-01 -1.05859256e+00 -3.24339539e-01
-3.62611674e-02 1.19161263e-01 6.49533272e-01 -9.51253593e-01
-1.15637672e+00 1.38944089e+00 7.11915314e-01 6.94007456e-01
1.38351822e+00 -3.26313108e-01 6.80038810e-01 3.05004269e-01
-2.84677953e-01 -1.17036366e+00 2.69178748e-01 1.23436534e+00
1.46989596e+00 -8.54765832e-01 -6.19731724e-01 -9.51587930e-02
-5.81220508e-01 1.06462765e+00 5.13484299e-01 -9.70151927e-03
1.04478467e+00 3.56608897e-01 3.37183803e-01 5.21884263e-02
-1.19102633e+00 -3.54242682e-01 3.64124060e-01 7.08187759e-01
7.40048409e-01 -7.15084821e-02 1.09118238e-01 6.42399251e-01
-6.24067485e-01 -5.92729747e-01 4.01645929e-01 6.40243888e-01
-3.22362661e-01 -8.18231821e-01 -3.76605421e-01 -1.07790507e-01
-6.53687239e-01 -6.11990154e-01 -2.39519194e-01 2.99903780e-01
-2.74969697e-01 1.36035097e+00 5.20103395e-01 -5.92182755e-01
5.39564610e-01 2.47192904e-01 -5.39665222e-02 -6.08199716e-01
-1.27853715e+00 4.75257754e-01 1.42243162e-01 -3.18628192e-01
-2.84187347e-01 -5.22726238e-01 -1.32045150e+00 -3.34926248e-02
-5.69998980e-01 1.34489968e-01 1.06498754e+00 9.14064407e-01
6.92126155e-01 1.40824270e+00 8.07110310e-01 -9.19795990e-01
-3.35827738e-01 -1.58609521e+00 -1.72656655e-01 -2.07872503e-02
1.23026705e+00 -3.58110428e-01 -5.10042965e-01 -3.10328901e-02] | [14.71248722076416, 6.028441429138184] |
ef21ca56-c762-4dc2-b9ed-f8a3f37d6858 | backproptools-a-fast-portable-deep | 2306.03530 | null | https://arxiv.org/abs/2306.03530v1 | https://arxiv.org/pdf/2306.03530v1.pdf | BackpropTools: A Fast, Portable Deep Reinforcement Learning Library for Continuous Control | Deep Reinforcement Learning (RL) has been demonstrated to yield capable agents and control policies in several domains but is commonly plagued by prohibitively long training times. Additionally, in the case of continuous control problems, the applicability of learned policies on real-world embedded devices is limited due to the lack of real-time guarantees and portability of existing deep learning libraries. To address these challenges, we present BackpropTools, a dependency-free, header-only, pure C++ library for deep supervised and reinforcement learning. Leveraging the template meta-programming capabilities of recent C++ standards, we provide composable components that can be tightly integrated by the compiler. Its novel architecture allows BackpropTools to be used seamlessly on a heterogeneous set of platforms, from HPC clusters over workstations and laptops to smartphones, smartwatches, and microcontrollers. Specifically, due to the tight integration of the RL algorithms with simulation environments, BackpropTools can solve popular RL problems like the Pendulum-v1 swing-up about 7 to 15 times faster in terms of wall-clock training time compared to other popular RL frameworks when using TD3. We also provide a low-overhead and parallelized interface to the MuJoCo simulator, showing that our PPO implementation achieves state of the art returns in the Ant-v4 environment while achieving a 25 to 30 percent faster wall-clock training time. Finally, we also benchmark the policy inference on a diverse set of microcontrollers and show that in most cases our optimized inference implementation is much faster than even the manufacturer's DSP libraries. To the best of our knowledge, BackpropTools enables the first-ever demonstration of training a deep RL algorithm directly on a microcontroller, giving rise to the field of Tiny Reinforcement Learning (TinyRL). Project page: https://backprop.tools | ['Giuseppe Loianno', 'Dario Albani', 'Jonas Eschmann'] | 2023-06-06 | null | null | null | null | ['continuous-control'] | ['playing-games'] | [-5.06213605e-01 -1.48329875e-02 -4.77424532e-01 -1.07654139e-01
-3.90437216e-01 -6.46626174e-01 4.93010759e-01 -2.09855720e-01
-6.31516278e-01 9.11744595e-01 -4.28431153e-01 -8.70227993e-01
9.97192562e-02 -8.90182436e-01 -8.46195936e-01 -6.37332976e-01
-3.49202126e-01 4.36364919e-01 1.32492617e-01 -4.75121439e-01
-2.39365727e-01 2.70708203e-01 -1.89001727e+00 -3.47412825e-01
5.78176200e-01 9.42213655e-01 1.34290487e-01 8.88758123e-01
3.11140448e-01 9.84457612e-01 -7.75197208e-01 3.75327468e-01
3.67919505e-01 2.98626535e-02 -3.70735675e-01 -6.46617889e-01
4.26968709e-02 -8.23365629e-01 -6.54215887e-02 6.38062179e-01
8.11079741e-01 -1.34149551e-01 -1.56503081e-01 -1.59712243e+00
2.33527467e-01 6.69063926e-01 -2.62411624e-01 -9.03735533e-02
2.63681263e-01 7.12856054e-01 6.00789845e-01 1.15276590e-01
2.99004763e-01 9.66249108e-01 5.91415703e-01 6.39733255e-01
-1.12590456e+00 -1.03891456e+00 9.27618220e-02 -5.46656772e-02
-9.40021515e-01 -4.51105535e-01 4.37697113e-01 3.59616093e-02
1.54879141e+00 7.95025155e-02 8.02877963e-01 1.64740813e+00
3.75783801e-01 8.24444354e-01 1.36878598e+00 -2.37093315e-01
8.41573834e-01 4.76937518e-02 -1.20981425e-01 6.80142820e-01
2.23624602e-01 5.42496383e-01 -2.18732923e-01 -2.02703074e-01
7.57839561e-01 -4.00117338e-01 8.56314451e-02 -1.81468710e-01
-1.11092830e+00 8.12355459e-01 2.71176010e-01 1.79831177e-01
-3.88523549e-01 1.06573486e+00 9.09302056e-01 4.64252830e-01
-2.47204509e-02 4.03924555e-01 -9.05961812e-01 -7.55569160e-01
-5.34396708e-01 5.16129971e-01 1.38157058e+00 8.83307874e-01
7.25805461e-01 7.21237957e-01 4.06847715e-01 2.72217274e-01
3.44979078e-01 6.57427728e-01 6.29975259e-01 -1.32048595e+00
1.82410121e-01 1.43302307e-01 2.14704290e-01 -3.82358044e-01
-7.43130744e-01 -4.99509186e-01 -3.90553266e-01 8.86394024e-01
2.23145783e-01 -8.80882025e-01 -4.42359030e-01 1.86003554e+00
6.48535728e-01 1.51995793e-01 3.59861523e-01 8.39523673e-01
4.21065956e-01 5.93859911e-01 1.20472014e-01 -1.17115891e-02
1.24023807e+00 -7.78451800e-01 -2.72635549e-01 -4.49929506e-01
5.16089022e-01 -4.44422215e-01 1.18040252e+00 5.51006794e-01
-7.85373926e-01 -4.22424883e-01 -1.48002911e+00 2.59924561e-01
-4.25246388e-01 -6.00263849e-02 1.21799171e+00 8.88853073e-01
-1.17629230e+00 7.95801640e-01 -1.34268880e+00 -2.82344937e-01
-3.21826403e-04 9.16840136e-01 1.42604169e-02 4.26894546e-01
-1.02882075e+00 9.83276069e-01 4.05098438e-01 -3.63799125e-01
-1.23150551e+00 -8.74705195e-01 -7.25073874e-01 -4.50123064e-02
6.06819570e-01 -7.56085813e-01 1.92790127e+00 -9.07414675e-01
-2.52080178e+00 1.43791124e-01 5.28117716e-01 -8.59154940e-01
4.27854419e-01 -2.57858008e-01 -3.23376626e-01 -1.11328781e-01
-8.23418424e-02 6.94371700e-01 8.42491448e-01 -8.62881064e-01
-7.53308296e-01 -1.30083948e-01 6.27043724e-01 1.59698203e-01
-1.36008427e-01 -2.90915906e-01 4.07895669e-02 -1.58202901e-01
-8.52470875e-01 -1.07576048e+00 -3.54864210e-01 -3.27969223e-01
1.10639080e-01 -8.65495428e-02 1.24867213e+00 -1.09839395e-01
6.05498075e-01 -1.96759307e+00 -2.44541481e-01 3.77305970e-02
-1.14192650e-01 3.63143325e-01 -4.62630391e-02 3.30631435e-01
2.45440841e-01 -2.80798048e-01 1.18341312e-01 8.09734985e-02
7.05008864e-01 7.38515079e-01 -3.19041908e-01 4.87719536e-01
-3.07921618e-01 6.33493066e-01 -1.07489407e+00 -2.43219241e-01
6.30625904e-01 4.20312792e-01 -6.73360765e-01 2.40828589e-01
-7.14917183e-01 4.26405936e-01 -5.37275970e-01 4.97614324e-01
2.63094366e-01 3.92183214e-02 5.99246204e-01 3.09085339e-01
-4.85466212e-01 6.98604047e-01 -1.21701908e+00 1.75690567e+00
-1.03507733e+00 4.51550007e-01 5.97764254e-01 -8.86313677e-01
6.52725875e-01 3.82320285e-01 5.67595065e-01 -9.10540342e-01
4.58209813e-01 4.17680115e-01 -7.34074637e-02 -2.44757667e-01
4.18170214e-01 1.86713830e-01 -3.05989236e-01 6.80107236e-01
1.77939013e-01 -5.04708469e-01 2.82750458e-01 -3.40934008e-01
1.30307400e+00 7.76539505e-01 3.14505696e-01 -3.50633472e-01
5.97640350e-02 2.22990409e-01 5.58517277e-01 7.21332550e-01
-2.15816244e-01 -5.46662211e-01 4.51733828e-01 -5.18650830e-01
-8.89177740e-01 -8.46571445e-01 3.76650840e-02 1.49689078e+00
-2.84593046e-01 -5.55329263e-01 -7.59467065e-01 -5.13324678e-01
1.87161803e-01 9.11913395e-01 -1.93380713e-01 2.95344125e-02
-6.40665114e-01 -6.50720775e-01 1.00214183e+00 5.31188130e-01
8.45411181e-01 -1.12967825e+00 -1.48931849e+00 5.48833609e-01
6.78189695e-01 -1.05675256e+00 1.07613187e-02 7.01556742e-01
-6.70530379e-01 -9.68811214e-01 -2.63813995e-02 -6.17886603e-01
-1.17063254e-01 -9.66500938e-02 1.12770748e+00 -1.61668405e-01
-1.52394786e-01 6.81963682e-01 -1.83461294e-01 -4.68463838e-01
-5.29863238e-01 1.68523699e-01 3.40846717e-01 -9.30265009e-01
7.96862170e-02 -9.63965476e-01 -5.89640260e-01 1.07578747e-01
-4.91027832e-01 -1.06015667e-01 3.47712904e-01 6.87187254e-01
2.01426953e-01 -1.57653280e-02 5.92031062e-01 -3.42185318e-01
7.20019817e-01 -4.04045582e-01 -1.47863281e+00 -2.93742478e-01
-7.83015013e-01 2.36223042e-01 1.10505807e+00 -5.02044439e-01
-7.48453438e-01 1.68769121e-01 -4.57586080e-01 -1.80564344e-01
-2.29799733e-01 1.54883415e-01 2.02469394e-01 -2.19089404e-01
5.84446669e-01 1.46572039e-01 3.70175600e-01 -4.83344086e-02
4.00346220e-01 5.77430308e-01 4.32863891e-01 -1.02941155e+00
3.22213471e-01 7.88103640e-02 -1.48899496e-01 -6.72878206e-01
-1.34203985e-01 2.26101756e-01 2.81211197e-01 -1.85790166e-01
6.21035695e-01 -1.02817667e+00 -1.74323606e+00 2.63421297e-01
-6.93936765e-01 -1.29472780e+00 -3.45380485e-01 4.95454758e-01
-8.25567424e-01 -5.03752672e-04 -6.91331029e-01 -6.01870418e-01
-7.58264363e-01 -1.56922293e+00 8.71174276e-01 4.72810537e-01
-2.02059463e-01 -9.55090702e-01 3.73650789e-01 -5.78907281e-02
6.80512846e-01 2.57997423e-01 5.96398473e-01 -3.01294893e-01
-3.07792485e-01 1.65081754e-01 3.79428834e-01 3.72357905e-01
-2.36674070e-01 5.24335988e-02 -1.05455017e+00 -6.76032126e-01
-1.39975250e-01 -8.00457537e-01 2.08139062e-01 3.62144619e-01
9.40179050e-01 -3.30142498e-01 -1.36328712e-01 6.03645742e-01
1.35715556e+00 2.89693654e-01 2.36750394e-01 7.98410118e-01
2.32986689e-01 -6.95996881e-02 5.99099398e-01 7.50001609e-01
4.78536248e-01 7.32608676e-01 8.63438725e-01 8.59074965e-02
2.54735708e-01 -1.05695017e-01 1.00779223e+00 6.91506743e-01
1.16797999e-01 2.93516010e-01 -7.31861591e-01 1.29311532e-01
-1.81142259e+00 -7.54043281e-01 1.95893303e-01 2.24416924e+00
8.58589292e-01 1.91309094e-01 3.30567956e-01 -2.87572877e-03
1.84022486e-01 5.43900728e-02 -9.15092230e-01 -9.50683355e-01
5.24145901e-01 7.28817999e-01 7.92071402e-01 5.23809850e-01
-1.00275934e+00 1.03380346e+00 6.58454990e+00 8.06187987e-01
-1.59208226e+00 1.91828027e-01 -3.14465426e-02 -2.88643390e-01
3.05766761e-01 2.08638962e-02 -7.60454595e-01 4.43519741e-01
1.47966838e+00 -1.67698175e-01 9.55118358e-01 1.49361801e+00
2.68695921e-01 -3.15182835e-01 -1.16002309e+00 6.31349564e-01
-5.75761020e-01 -1.15438759e+00 -7.51397431e-01 8.64646658e-02
3.23322982e-01 6.19396508e-01 2.48306151e-02 1.08851302e+00
1.10919333e+00 -8.82030129e-01 6.96879625e-01 -3.29201758e-01
6.29335821e-01 -1.04324603e+00 5.05271018e-01 4.28735942e-01
-9.63587940e-01 -1.81679070e-01 -2.28649378e-01 -4.99280334e-01
-2.80090421e-01 2.25969568e-01 -1.21726549e+00 2.83826232e-01
8.44684601e-01 4.86414760e-01 -1.19190678e-01 6.19241476e-01
-4.51652497e-01 7.86925197e-01 -7.56065488e-01 -4.62036103e-01
5.22542775e-01 7.12526143e-02 1.68787390e-01 1.02895737e+00
9.23897400e-02 -2.76605189e-01 2.61935622e-01 5.96054018e-01
1.33549467e-01 -3.50249350e-01 -4.78368431e-01 1.47731051e-01
5.11812627e-01 1.53390443e+00 -3.70242506e-01 -2.09513366e-01
-3.11798066e-01 5.84707081e-01 3.14999610e-01 1.52843863e-01
-1.30927300e+00 -3.64654124e-01 1.18014061e+00 -3.92162830e-01
2.60343820e-01 -5.61766684e-01 2.14495286e-02 -8.15171540e-01
-2.34312683e-01 -1.38873017e+00 -1.44103058e-02 -4.81676817e-01
-7.69835949e-01 3.49651694e-01 -1.01334460e-01 -9.08986926e-01
-6.91447318e-01 -7.79226482e-01 -5.03502488e-01 3.03199738e-01
-1.48685026e+00 -7.52931774e-01 -1.54445007e-01 7.47140348e-01
3.19069624e-01 -3.46813202e-01 1.28171742e+00 2.46899635e-01
-6.47075534e-01 3.51001263e-01 1.89959720e-01 -1.76827058e-01
4.68924314e-01 -1.48268819e+00 4.14925218e-01 3.13041031e-01
-3.12081754e-01 4.59023416e-01 8.83603930e-01 -1.95126578e-01
-2.13981080e+00 -9.44816351e-01 -1.83210686e-01 1.61624372e-01
9.13428724e-01 -5.01859546e-01 -2.88598061e-01 8.82958472e-01
5.85029602e-01 6.82442635e-02 3.96441221e-01 1.14892438e-01
-2.39396468e-01 -5.36924481e-01 -1.02078998e+00 7.59802997e-01
4.02963042e-01 -2.95823395e-01 -3.07263643e-01 4.62882251e-01
6.92892253e-01 -7.76938081e-01 -1.00007451e+00 1.59036353e-01
6.72806144e-01 -1.00757623e+00 7.85028756e-01 -1.21502608e-01
8.06576014e-02 -2.92216420e-01 -9.06370431e-02 -1.48641217e+00
2.13985875e-01 -1.12617290e+00 -3.02621931e-01 8.55262041e-01
4.44500297e-02 -1.01051259e+00 7.70969510e-01 2.60414958e-01
-2.17025891e-01 -5.91690183e-01 -1.11486185e+00 -7.95813620e-01
1.25474632e-01 -7.64749408e-01 7.78559327e-01 6.50735736e-01
4.28052098e-01 3.31530601e-01 -1.97088301e-01 1.83261707e-01
4.43622768e-01 3.41249704e-02 1.05463910e+00 -5.55329323e-01
-1.03996372e+00 -3.99915636e-01 -1.26814485e-01 -9.42166150e-01
4.64649290e-01 -7.10218608e-01 1.31174743e-01 -1.10070658e+00
-4.65179533e-01 -7.81996727e-01 4.43328209e-02 9.40069854e-01
4.72010136e-01 -1.56922922e-01 3.18164647e-01 -2.42356598e-01
-6.47608876e-01 5.94178915e-01 8.77980828e-01 -8.58868733e-02
-3.27383071e-01 1.65143935e-03 -2.04910517e-01 8.36590350e-01
1.25912309e+00 -3.98890823e-01 -5.10396242e-01 -5.36235988e-01
1.43738329e-01 1.60254210e-01 4.66478258e-01 -1.51865292e+00
1.95530728e-01 -1.25605553e-01 1.16047718e-01 7.56723285e-02
4.33911413e-01 -9.24037158e-01 -7.74627402e-02 9.31081951e-01
1.32704778e-02 2.72278041e-01 6.72690809e-01 1.21964186e-01
3.16067308e-01 6.96704015e-02 7.30471313e-01 -1.06589667e-01
-7.87621796e-01 -7.84076303e-02 -9.44765210e-01 -3.17128301e-02
1.39050889e+00 3.51232976e-01 -6.78002775e-01 -3.14631999e-01
-2.41903588e-01 3.49755794e-01 3.69116724e-01 2.25363195e-01
1.21292889e-01 -8.43720853e-01 -1.51817545e-01 1.59680694e-01
-3.74924064e-01 -1.60692155e-01 -1.69776514e-01 6.85779572e-01
-7.61595964e-01 4.79291797e-01 -5.45931458e-01 -5.25642037e-01
-9.99708652e-01 5.12117684e-01 6.19037747e-01 -3.36710542e-01
-8.13307941e-01 1.77553579e-01 -5.83156884e-01 -7.60897458e-01
3.72761220e-01 -6.77758217e-01 2.85024196e-01 -4.57157016e-01
4.05741036e-01 3.65620345e-01 1.98162273e-01 2.53965706e-01
-4.99756664e-01 2.07372546e-01 1.91977710e-01 -2.09873915e-01
1.44653451e+00 5.69999576e-01 2.56738327e-02 1.93986923e-01
7.77639925e-01 -1.25932246e-01 -1.61368680e+00 4.36026216e-01
-1.98583841e-01 2.41371170e-01 2.85206586e-01 -9.26255822e-01
-1.00017357e+00 4.57493722e-01 7.49357939e-01 -2.33880300e-02
9.21434700e-01 -5.73126137e-01 7.22999215e-01 7.25666285e-01
1.11988688e+00 -1.33326602e+00 -1.17221095e-01 5.91672003e-01
5.11812985e-01 -1.00405252e+00 8.56416449e-02 2.79763371e-01
-5.24376571e-01 1.13037574e+00 7.44266987e-01 -4.75484192e-01
3.12247306e-01 1.17952442e+00 6.73431754e-02 2.00610846e-01
-1.10982025e+00 -1.67011365e-01 -8.21444571e-01 7.87071824e-01
2.89662123e-01 4.36115801e-01 -2.01020941e-01 3.49104911e-01
-4.72960204e-01 3.32932174e-01 6.37640893e-01 1.30884790e+00
-3.80850852e-01 -1.31541312e+00 -4.79569167e-01 -9.66373459e-02
-4.51519012e-01 2.05820009e-01 1.95890442e-01 1.19189465e+00
4.88938093e-02 1.03209484e+00 -2.26405729e-02 -4.09992874e-01
5.79043776e-02 -2.24016011e-01 6.03790879e-01 -2.62441218e-01
-1.21224821e+00 -1.15760669e-01 3.83426070e-01 -1.02838719e+00
-7.77141005e-02 -2.66899139e-01 -1.61021602e+00 -7.57045448e-01
2.44932190e-01 9.28252339e-02 1.16030324e+00 7.08118200e-01
5.34430027e-01 7.41986215e-01 3.82234603e-01 -1.31672096e+00
-1.00193882e+00 -5.90354741e-01 -4.54122066e-01 -4.30765063e-01
2.77500957e-01 -6.98033392e-01 -3.97591889e-02 -6.41551256e-01] | [4.02696418762207, 1.4802101850509644] |
67e562bb-266c-4945-8287-30645814dbaa | infinite-physical-monkey-do-deep-learning | 2304.10494 | null | https://arxiv.org/abs/2304.10494v1 | https://arxiv.org/pdf/2304.10494v1.pdf | Infinite Physical Monkey: Do Deep Learning Methods Really Perform Better in Conformation Generation? | Conformation Generation is a fundamental problem in drug discovery and cheminformatics. And organic molecule conformation generation, particularly in vacuum and protein pocket environments, is most relevant to drug design. Recently, with the development of geometric neural networks, the data-driven schemes have been successfully applied in this field, both for molecular conformation generation (in vacuum) and binding pose generation (in protein pocket). The former beats the traditional ETKDG method, while the latter achieves similar accuracy compared with the widely used molecular docking software. Although these methods have shown promising results, some researchers have recently questioned whether deep learning (DL) methods perform better in molecular conformation generation via a parameter-free method. To our surprise, what they have designed is some kind analogous to the famous infinite monkey theorem, the monkeys that are even equipped with physics education. To discuss the feasibility of their proving, we constructed a real infinite stochastic monkey for molecular conformation generation, showing that even with a more stochastic sampler for geometry generation, the coverage of the benchmark QM-computed conformations are higher than those of most DL-based methods. By extending their physical monkey algorithm for binding pose prediction, we also discover that the successful docking rate also achieves near-best performance among existing DL-based docking models. Thus, though their conclusions are right, their proof process needs more concern. | ['Yafeng Deng', 'Dejun Jiang', 'Huifeng Zhao', 'Jintu Zhang', 'Haotian Zhang'] | 2023-03-08 | null | null | null | null | ['pose-prediction', 'drug-discovery', 'molecular-docking'] | ['computer-vision', 'medical', 'medical'] | [-7.46553624e-03 1.90950781e-01 -9.25545543e-02 1.66899641e-03
-5.75064242e-01 -6.52268469e-01 4.40319449e-01 1.29306629e-01
-2.08062306e-01 1.54099381e+00 -2.71011174e-01 -6.94892585e-01
-8.29583779e-02 -9.76612747e-01 -1.13020778e+00 -1.15497446e+00
-2.29454786e-01 6.33661926e-01 -2.56675947e-02 -4.90381271e-01
5.12483954e-01 8.07344496e-01 -1.40533948e+00 1.48521900e-01
1.21075344e+00 5.10936320e-01 2.86237538e-01 2.61160672e-01
-3.78737114e-02 3.02683324e-01 -3.14791501e-01 -4.80056673e-01
1.49552643e-01 -4.28640634e-01 -8.09386492e-01 -6.61808789e-01
3.12289029e-01 5.88165298e-02 4.40351218e-02 8.85755241e-01
8.27074647e-01 1.32329231e-02 8.30038190e-01 -5.53457022e-01
-7.38852501e-01 2.18973070e-01 -2.18029574e-01 -1.05447926e-01
6.62629008e-01 3.41643035e-01 7.11452007e-01 -8.35614145e-01
7.96810091e-01 1.02020335e+00 6.88879728e-01 7.28165388e-01
-1.45846689e+00 -8.15019608e-01 -1.31045207e-01 3.45273077e-01
-1.43714666e+00 1.53022995e-02 5.60349941e-01 -5.43386936e-01
1.07508445e+00 2.24698290e-01 8.24695408e-01 1.46737969e+00
7.63465941e-01 3.43261182e-01 1.11769485e+00 -3.74989569e-01
6.60298884e-01 1.21349853e-03 -2.45970011e-01 5.21860003e-01
2.97469944e-01 4.59000260e-01 -3.68391842e-01 -5.14123559e-01
6.88618600e-01 -2.76185367e-02 -4.79509711e-01 -6.66397452e-01
-1.07871091e+00 1.11463070e+00 6.56153142e-01 1.90063491e-01
-4.37824696e-01 -3.86128984e-02 -1.09778754e-01 7.64623797e-03
1.62112385e-01 1.11324251e+00 -6.74893260e-01 1.39949992e-01
-6.85851038e-01 8.31863165e-01 8.82730126e-01 7.43997276e-01
4.78647321e-01 -1.19082987e-01 4.15803008e-02 7.25234300e-02
3.94035637e-01 2.89981872e-01 2.62133598e-01 -5.28662801e-01
1.67388111e-01 4.42313045e-01 3.00294906e-01 -8.61516953e-01
-7.00905323e-01 -4.49517906e-01 -9.84604895e-01 4.60166991e-01
5.19305229e-01 -1.57194406e-01 -6.60990298e-01 1.71995318e+00
4.64560568e-01 -1.57698587e-01 9.95975435e-02 7.53618360e-01
7.90870488e-01 8.62164378e-01 3.30709577e-01 -5.44025898e-01
1.10933948e+00 -3.40713829e-01 -3.10589612e-01 5.28113484e-01
6.08077526e-01 -7.09006310e-01 7.80087471e-01 8.03563535e-01
-8.17174673e-01 -3.55206370e-01 -1.22836018e+00 2.74642825e-01
-4.43269283e-01 -3.60025674e-01 1.28426206e+00 8.21723282e-01
-9.30408239e-01 1.26711738e+00 -5.90383410e-01 -3.58775944e-01
4.58913594e-01 8.08251441e-01 -5.06271124e-01 3.91787708e-01
-1.27833843e+00 1.16684675e+00 5.15106022e-01 -5.61478138e-02
-7.69081652e-01 -6.98988080e-01 -3.31032187e-01 -2.58040071e-01
1.30582944e-01 -9.92188275e-01 9.27137434e-01 -6.95268929e-01
-1.82957864e+00 6.25212669e-01 -1.90111980e-01 -5.32231390e-01
3.99805605e-01 2.05429360e-01 -7.61174783e-02 -1.80632517e-01
-2.97964275e-01 8.31546903e-01 2.99148649e-01 -1.06255376e+00
4.42863554e-02 -5.25676727e-01 1.80277914e-01 6.79002777e-02
4.24213082e-01 -2.57973522e-01 4.40632999e-01 -2.69065738e-01
1.38283581e-01 -8.62136960e-01 -7.43414044e-01 -1.87788948e-01
-4.76408005e-01 -5.28463364e-01 5.87070361e-02 -2.03083605e-01
8.73950005e-01 -1.36650026e+00 3.97449940e-01 2.54212677e-01
3.07175368e-01 3.44447613e-01 2.35107187e-02 8.29326034e-01
-4.63878900e-01 1.99824423e-01 -2.81176846e-02 3.34811270e-01
-2.42964119e-01 -2.25462005e-01 -3.89699966e-01 5.35032868e-01
7.97996223e-02 8.37271750e-01 -8.02298665e-01 -1.86824769e-01
-4.88867275e-02 5.42811573e-01 -1.00787425e+00 -6.92284033e-02
-6.86875224e-01 1.00187004e+00 -7.90066957e-01 5.12896061e-01
8.27223718e-01 -3.24748904e-01 5.43021739e-01 -1.12854101e-01
-2.88067043e-01 1.54712588e-01 -7.44288146e-01 1.61912167e+00
2.32804399e-02 9.43684056e-02 -6.37719393e-01 -8.09670508e-01
8.56270254e-01 3.20682049e-01 5.27505755e-01 -8.55553627e-01
-2.77021956e-02 3.79349858e-01 4.36789989e-01 -3.87633532e-01
2.61083663e-01 -5.72284639e-01 3.92499298e-01 -5.33609279e-02
-4.24498357e-02 -2.38665231e-02 -1.39551520e-01 -1.24146968e-01
7.24425495e-01 7.65909910e-01 4.91119623e-01 -4.02576834e-01
5.16343474e-01 2.73179978e-01 3.86585325e-01 4.07561332e-01
1.60586730e-01 4.94648576e-01 4.83414799e-01 -8.27017367e-01
-1.16980827e+00 -9.36633348e-01 -6.57057166e-01 6.67629719e-01
2.68397301e-01 -4.04239684e-01 -9.98237073e-01 -3.37174922e-01
-8.68553668e-02 4.62167770e-01 -6.11723125e-01 -1.49367318e-01
-5.76129735e-01 -1.27716267e+00 3.50428045e-01 -2.55112723e-02
-1.96198598e-01 -1.42039621e+00 -3.47311944e-01 5.83620787e-01
2.52858251e-01 -4.13783669e-01 1.38573691e-01 4.46493268e-01
-9.61981952e-01 -1.23166037e+00 -8.43627691e-01 -5.59534848e-01
3.98504883e-01 -2.89283395e-01 9.33335900e-01 -9.67006460e-02
-3.87603313e-01 -5.47863483e-01 -2.53181130e-01 -5.23837030e-01
-5.72580636e-01 1.76533341e-01 3.42001826e-01 -4.69822556e-01
3.74999672e-01 -9.63252246e-01 -9.59952474e-01 1.42836720e-01
-5.91457605e-01 9.84948128e-02 6.51775360e-01 6.63637757e-01
7.32853174e-01 -5.67533970e-01 7.74675965e-01 -6.97569549e-01
4.95849729e-01 -4.67182636e-01 -6.33069932e-01 4.98515032e-02
-5.87652028e-01 4.31802750e-01 8.63468051e-01 -3.34409058e-01
-5.15380919e-01 1.92419350e-01 -7.49964178e-01 1.81967020e-01
-2.93494225e-01 3.46392572e-01 -3.14268500e-01 -5.06726265e-01
9.71890807e-01 3.90445173e-01 4.17433027e-03 -6.11190617e-01
1.96380317e-01 2.98204750e-01 -4.67494316e-02 -8.31416905e-01
5.62632084e-01 1.60466909e-01 4.10182536e-01 -7.42753386e-01
-5.00285447e-01 1.86343119e-01 -6.11024678e-01 2.81157255e-01
1.00697613e+00 -5.61955869e-01 -1.56515324e+00 7.89922103e-02
-1.52537894e+00 -9.68598947e-02 -3.70270647e-02 5.96987963e-01
-8.44558120e-01 6.50612831e-01 -1.94050550e-01 -6.87600374e-01
-3.48477244e-01 -1.55341887e+00 8.48594606e-01 2.58352071e-01
-2.02268556e-01 -7.65042901e-01 4.65471685e-01 5.98334409e-02
3.07218790e-01 7.77718782e-01 1.43486655e+00 -6.91720366e-01
-7.67811954e-01 2.64007822e-02 2.07399875e-01 -1.54734820e-01
-1.58119291e-01 -3.68951783e-02 -8.25699270e-01 -2.55082905e-01
-1.72816426e-01 -2.73737639e-01 4.67660636e-01 5.64695835e-01
1.19902802e+00 -3.10594738e-01 -6.04831755e-01 7.57359326e-01
1.56861234e+00 7.26094902e-01 9.24225509e-01 4.26694483e-01
4.80984271e-01 4.62897360e-01 1.86699510e-01 3.23098421e-01
6.05890080e-02 9.30215597e-01 7.37916827e-01 -1.06276967e-01
2.87128299e-01 -3.14976871e-01 2.55362689e-01 2.61168748e-01
-7.06413269e-01 -2.20569238e-01 -7.27655768e-01 -2.23179728e-01
-1.64014781e+00 -1.21776259e+00 -4.08394545e-01 2.45549583e+00
1.31858969e+00 -1.12377383e-01 3.28763306e-01 -2.37532496e-01
3.23702633e-01 -3.35059911e-01 -7.87402272e-01 -4.66432929e-01
-3.07240814e-01 7.71074176e-01 3.05039525e-01 4.06105101e-01
-8.05987477e-01 9.49834466e-01 7.36403465e+00 9.76548970e-01
-1.07499099e+00 -3.21332812e-01 7.41107106e-01 7.12446868e-02
-3.73705238e-01 1.41804487e-01 -9.31480825e-01 4.48501021e-01
9.43595350e-01 -6.16089180e-02 2.82055110e-01 9.47970450e-01
4.07710284e-01 -9.65212360e-02 -1.42278016e+00 9.23547149e-01
-4.04455841e-01 -1.93691003e+00 4.33955908e-01 5.63295901e-01
6.06265783e-01 -2.50167906e-01 4.39908095e-02 1.74373746e-01
-7.67638609e-02 -1.75321376e+00 4.76123720e-01 7.64579117e-01
5.83823740e-01 -9.49875474e-01 5.79724669e-01 4.99082625e-01
-6.28788412e-01 3.66181225e-01 -7.80002654e-01 -1.45777285e-01
-8.96433815e-02 3.30219746e-01 -7.36668229e-01 6.29499912e-01
1.48864672e-01 2.33790010e-01 -4.40181613e-01 1.26405108e+00
1.49543852e-01 2.14418620e-01 -1.54008240e-01 -5.67583382e-01
2.34242752e-01 -6.20857120e-01 3.97775173e-01 6.98374927e-01
1.73855230e-01 2.29433700e-01 -5.65432804e-03 1.29983568e+00
1.62797645e-01 4.00808334e-01 -6.01280034e-01 8.86950716e-02
2.31906101e-01 9.30923343e-01 -4.47724432e-01 9.21009257e-02
4.18042764e-02 6.37127280e-01 1.16344750e-01 2.85281420e-01
-1.04360223e+00 -2.77128220e-01 7.95832455e-01 2.50380307e-01
9.34478045e-02 -1.92936972e-01 -3.32446806e-02 -9.13450539e-01
-2.00851008e-01 -9.80864227e-01 -3.96934122e-01 -5.80397844e-01
-1.11421287e+00 6.15209460e-01 -1.50491208e-01 -1.02138984e+00
-4.38725613e-02 -9.99956071e-01 -5.17091572e-01 1.04281771e+00
-1.21013355e+00 -6.78951561e-01 2.42082760e-01 4.14513737e-01
1.19177967e-01 -2.72885144e-01 1.17435062e+00 1.77732170e-01
-3.98306489e-01 5.89576304e-01 4.70045418e-01 -5.06091058e-01
4.83293295e-01 -1.16166580e+00 2.83775210e-01 5.34559116e-02
-7.95275643e-02 1.07832754e+00 9.61000025e-01 -5.98895729e-01
-1.80029619e+00 -5.73879123e-01 6.45815551e-01 -6.44130409e-01
3.62279385e-01 -2.28185758e-01 -7.46726215e-01 8.04028809e-02
9.10705030e-02 -4.72964108e-01 7.61366963e-01 -1.70268714e-02
-6.74192153e-04 3.79844218e-01 -1.24037957e+00 6.99241936e-01
1.14110529e+00 2.17986051e-02 -5.05138397e-01 7.94053197e-01
7.84770906e-01 -4.17403221e-01 -7.95300245e-01 4.50429022e-01
7.04682350e-01 -1.27985382e+00 1.19126797e+00 -1.18759727e+00
3.89840305e-01 -4.25035536e-01 -1.16478480e-01 -9.46090102e-01
-4.00701523e-01 -8.67506862e-01 1.98103581e-02 4.61049467e-01
6.57048583e-01 -6.43570244e-01 1.08670223e+00 1.87095016e-01
-1.39832161e-02 -1.30924046e+00 -1.04711545e+00 -9.35101807e-01
7.75740564e-01 -3.25955488e-02 6.89798355e-01 7.89562583e-01
6.84173405e-02 4.95093614e-01 -2.48053268e-01 -1.49610518e-02
4.56173599e-01 1.78544015e-01 7.41594613e-01 -1.56928253e+00
-4.27886426e-01 -4.45428848e-01 -4.41621125e-01 -9.20581698e-01
4.62048128e-02 -1.11963034e+00 -3.89957041e-01 -1.41335905e+00
3.63240689e-01 -3.50460529e-01 1.59045339e-01 1.33677304e-01
5.47927469e-02 -3.88683937e-02 -1.20819524e-01 2.06299752e-01
-2.05529869e-01 6.76687121e-01 1.68374157e+00 1.04270570e-01
-3.67380589e-01 1.81510791e-01 -8.07707191e-01 6.65296376e-01
7.07103193e-01 -5.63102603e-01 -1.33069739e-01 4.12508905e-01
7.85109878e-01 -7.74310762e-03 3.21052790e-01 -9.74886298e-01
-4.72049229e-02 -3.52494091e-01 5.81327379e-01 -6.34406745e-01
2.11399913e-01 -4.74555016e-01 5.10538459e-01 7.56692410e-01
1.47737876e-01 -9.70554501e-02 7.53364265e-02 5.51916838e-01
1.72185168e-01 -1.27612829e-01 9.56111491e-01 -2.81192064e-01
-1.54060170e-01 6.51435614e-01 -5.06398618e-01 -2.47191012e-01
1.04551268e+00 -6.46806359e-01 -1.12482615e-01 -8.06306489e-03
-9.57948506e-01 -2.78860003e-01 5.62255502e-01 -8.07653889e-02
5.17114222e-01 -1.09542108e+00 -4.96307075e-01 -6.62132055e-02
2.40954775e-02 6.06693476e-02 8.26521292e-02 8.32735658e-01
-9.65369642e-01 7.56624520e-01 -1.46824107e-01 -4.48296517e-01
-8.24227512e-01 7.16497242e-01 6.35007501e-01 -2.14201644e-01
-5.33114791e-01 7.08044171e-01 4.74219412e-01 -4.12968069e-01
1.99205250e-01 -3.60350430e-01 -9.96353328e-02 -8.47594589e-02
3.90216053e-01 2.25282237e-02 1.31440237e-01 -2.11545154e-01
-4.99910980e-01 7.18114495e-01 -2.49891505e-02 4.76702213e-01
1.60997939e+00 6.38292134e-01 -1.36713862e-01 -5.83140813e-02
9.13149238e-01 4.15503643e-02 -1.10562801e+00 3.54807675e-01
5.95951453e-02 9.59400982e-02 -5.21028340e-01 -9.24517751e-01
-3.32488418e-01 8.30471039e-01 7.34228373e-01 1.50111103e-02
5.65426826e-01 -7.19695017e-02 5.72266698e-01 8.33729565e-01
9.24950004e-01 -4.81453568e-01 -7.05152154e-02 5.24457872e-01
1.01510870e+00 -1.18777645e+00 2.02465236e-01 -1.95641950e-01
-3.05077016e-01 1.36868072e+00 4.29850519e-01 -2.19951436e-01
4.09579068e-01 -2.11355127e-02 -4.41607028e-01 -3.67448837e-01
-5.14156938e-01 -7.07861409e-03 2.18630895e-01 8.27855587e-01
8.83979023e-01 -4.28690352e-02 -6.89977050e-01 6.77656353e-01
-4.68992025e-01 1.87439434e-02 3.90861750e-01 5.62976301e-01
-7.52015233e-01 -1.58760977e+00 -3.73348832e-01 -1.18561558e-01
-3.81301492e-01 -3.37677807e-01 -6.84515238e-01 9.69877422e-01
3.40615928e-01 5.17480373e-01 -4.60987866e-01 -1.99271634e-01
3.13714385e-01 1.22351281e-01 1.11932766e+00 -5.98236144e-01
-6.68636441e-01 -2.51704961e-01 -2.56321430e-01 -5.59394538e-01
-3.24279904e-01 -1.61513016e-01 -1.34389162e+00 -6.45991862e-01
-4.76190776e-01 6.00067258e-01 6.18247926e-01 9.24824774e-01
5.09776413e-01 2.27689058e-01 5.65930665e-01 -1.19305396e+00
-5.16090453e-01 -7.56517947e-01 -6.34854794e-01 -7.12033212e-02
8.15976560e-02 -6.52413368e-01 -5.90349659e-02 -2.41225794e-01] | [4.962181568145752, 5.563250541687012] |
21074099-1239-43c1-96e0-d8ed4d6bd5a9 | 3d-object-reconstruction-from-hand-object | 1704.00529 | null | http://arxiv.org/abs/1704.00529v1 | http://arxiv.org/pdf/1704.00529v1.pdf | 3D Object Reconstruction from Hand-Object Interactions | Recent advances have enabled 3d object reconstruction approaches using a
single off-the-shelf RGB-D camera. Although these approaches are successful for
a wide range of object classes, they rely on stable and distinctive geometric
or texture features. Many objects like mechanical parts, toys, household or
decorative articles, however, are textureless and characterized by minimalistic
shapes that are simple and symmetric. Existing in-hand scanning systems and 3d
reconstruction techniques fail for such symmetric objects in the absence of
highly distinctive features. In this work, we show that extracting 3d hand
motion for in-hand scanning effectively facilitates the reconstruction of even
featureless and highly symmetric objects and we present an approach that fuses
the rich additional information of hands into a 3d reconstruction pipeline,
significantly contributing to the state-of-the-art of in-hand scanning. | ['Dimitrios Tzionas', 'Juergen Gall'] | 2017-04-03 | 3d-object-reconstruction-from-hand-object-1 | http://openaccess.thecvf.com/content_iccv_2015/html/Tzionas_3D_Object_Reconstruction_ICCV_2015_paper.html | http://openaccess.thecvf.com/content_iccv_2015/papers/Tzionas_3D_Object_Reconstruction_ICCV_2015_paper.pdf | iccv-2015-12 | ['3d-object-reconstruction'] | ['computer-vision'] | [ 1.49483770e-01 -3.33749771e-01 -1.29841641e-01 -3.14723492e-01
-2.92853504e-01 -7.52048433e-01 3.98720324e-01 -3.26896280e-01
3.46839428e-02 9.05740559e-02 -8.43009651e-02 -1.77876040e-01
-1.81488663e-01 -5.56789577e-01 -3.20785493e-01 -2.38626227e-01
2.92359233e-01 1.00882709e+00 8.34271789e-01 -4.07210588e-01
1.15084879e-01 1.17000401e+00 -1.79160452e+00 2.82874644e-01
1.99291836e-02 1.12440133e+00 3.47349435e-01 6.30366921e-01
-3.05066824e-01 2.03309536e-01 1.53783634e-02 -3.87184530e-01
3.70274752e-01 -3.21554393e-01 -7.33286977e-01 2.88313270e-01
5.20548940e-01 -6.71452582e-01 -3.00890744e-01 5.38822889e-01
6.67222142e-01 -2.66386867e-01 5.65456986e-01 -7.21461594e-01
-2.01685414e-01 -1.46914452e-01 -7.14694381e-01 -4.80977476e-01
9.67431247e-01 1.49046689e-01 8.04825902e-01 -9.83785331e-01
1.09409082e+00 1.27539742e+00 8.70194554e-01 6.81172550e-01
-1.21999002e+00 -2.43370950e-01 -6.16712309e-02 -2.24194169e-01
-1.24160147e+00 -3.42094302e-01 1.09358966e+00 -3.60687912e-01
1.02818286e+00 3.75032395e-01 1.11524808e+00 9.86217141e-01
-9.05774906e-02 9.32741404e-01 1.13484156e+00 -6.49846554e-01
-4.43121828e-02 -1.65963843e-02 -2.19581619e-01 8.39533031e-01
2.02290848e-01 -4.80479933e-02 -7.17205405e-01 1.00457162e-01
1.41082716e+00 3.07644546e-01 -6.78372309e-02 -1.07006884e+00
-1.28722990e+00 3.16628128e-01 1.92577228e-01 2.26018935e-01
-3.35774124e-01 1.40903249e-01 3.73222120e-02 -8.41279104e-02
2.37457052e-01 -8.27309191e-02 -6.09468222e-01 -5.82534373e-01
-9.13248003e-01 2.87860602e-01 7.50705898e-01 1.13156652e+00
6.80587769e-01 -2.68429548e-01 4.24932033e-01 4.91520315e-01
3.42260629e-01 7.72049367e-01 4.89945598e-02 -8.68495643e-01
3.57383877e-01 7.75989354e-01 2.48288691e-01 -6.88140690e-01
-6.55086815e-01 1.56692043e-01 -4.26781178e-01 5.29243350e-01
4.11341161e-01 4.92679030e-01 -9.57883239e-01 8.67480874e-01
7.28835881e-01 -6.75710738e-01 -6.01426959e-01 1.14640725e+00
8.31499159e-01 -2.21675828e-01 -5.45587838e-01 4.15365040e-01
1.34686112e+00 -6.74917817e-01 -4.34647411e-01 -1.28315398e-02
8.90004858e-02 -1.12288558e+00 1.43505168e+00 6.08105600e-01
-1.39998746e+00 -3.03548098e-01 -5.95474958e-01 -2.54305869e-01
-1.22199051e-01 -1.55694187e-01 7.03797698e-01 8.07520986e-01
-6.32261932e-01 8.17398787e-01 -9.77594554e-01 -4.58923966e-01
5.12268424e-01 5.06140471e-01 -7.11599767e-01 -1.87743604e-01
-2.37429395e-01 1.05823195e+00 4.66972217e-02 1.11821350e-02
-3.31989169e-01 -6.65796936e-01 -5.83526552e-01 -5.15108347e-01
4.69127625e-01 -7.41223812e-01 1.46175313e+00 -5.13463438e-01
-1.90272856e+00 1.36359119e+00 -5.47476709e-02 3.76476169e-01
8.85980964e-01 -4.91047740e-01 5.19418940e-02 4.62079674e-01
-3.49321902e-01 4.37206596e-01 1.09073853e+00 -1.36405134e+00
-1.35407299e-01 -7.66559422e-01 1.59651086e-01 6.03505317e-03
1.81497678e-01 2.22673770e-02 -6.52658641e-01 -6.14373922e-01
6.60443366e-01 -9.75321174e-01 3.50433029e-02 6.16940498e-01
-4.70470369e-01 1.56065300e-01 1.30211473e+00 -6.73575044e-01
6.85165167e-01 -1.88037026e+00 2.31153414e-01 3.56149584e-01
-7.62914866e-02 6.08419105e-02 1.50542974e-01 3.48956108e-01
2.91870326e-01 -1.51210561e-01 1.99273285e-02 -5.66774905e-01
1.24891005e-01 2.27337390e-01 5.47823645e-02 5.64718485e-01
1.68774560e-01 9.38418269e-01 -7.89708376e-01 -5.53432763e-01
7.44860709e-01 8.65840971e-01 -4.25300211e-01 1.52231365e-01
-2.40152881e-01 6.60326421e-01 -5.89121878e-01 1.39476669e+00
6.05771005e-01 -1.32689118e-01 2.09339052e-01 -3.22529048e-01
-1.29420102e-01 2.25651860e-01 -1.49354064e+00 2.23721862e+00
-3.92901242e-01 2.39767551e-01 3.68475229e-01 -3.63638550e-01
6.77865386e-01 2.10071623e-01 6.75458252e-01 -6.06118023e-01
2.66084999e-01 5.60984671e-01 -2.74372250e-01 -4.01592851e-01
3.75488281e-01 -5.55818498e-01 2.90522784e-01 7.68488109e-01
-2.71552086e-01 -9.57388461e-01 -4.84060198e-01 -1.67970911e-01
7.85802186e-01 8.78003120e-01 3.35430682e-01 7.20914453e-02
1.25057042e-01 1.24240465e-01 -1.27990633e-01 3.43694508e-01
1.80311665e-01 1.30879974e+00 9.74802207e-03 -4.63456035e-01
-1.36374366e+00 -1.17708349e+00 -3.04403514e-01 7.27274954e-01
2.36920446e-01 -2.54605323e-01 -5.80701590e-01 -5.01038849e-01
4.57279325e-01 -1.88552006e-03 -5.07169008e-01 3.89679670e-01
-6.80431426e-01 5.38265146e-02 2.23729715e-01 7.61956990e-01
2.07348645e-01 -9.53311741e-01 -1.37038970e+00 1.07152171e-01
3.28660876e-01 -1.13051438e+00 -2.36868382e-01 1.93584129e-01
-1.17944336e+00 -1.21764386e+00 -9.54822004e-01 -4.06865180e-01
6.64815784e-01 5.47340989e-01 1.25401008e+00 7.88689852e-02
-7.75031388e-01 9.58478332e-01 -4.75562125e-01 -3.45395803e-01
-1.04091868e-01 -5.96264526e-02 2.30421856e-01 -3.55681956e-01
-5.25497682e-02 -7.64960110e-01 -7.33206570e-01 5.31287313e-01
-6.71567798e-01 1.23858601e-01 5.01730382e-01 4.97275323e-01
9.57583606e-01 -3.88773590e-01 -1.53313503e-01 -5.23598731e-01
7.47104213e-02 6.79385215e-02 -4.82048213e-01 1.73646897e-01
-1.35414660e-01 4.27942835e-02 1.93841636e-01 -6.60220861e-01
-1.09056747e+00 4.45742071e-01 -2.50014961e-01 -7.15241015e-01
-2.78331876e-01 6.64031319e-03 -1.39071688e-01 -3.50221545e-01
3.70685816e-01 8.87146890e-02 2.59218425e-01 -1.09449136e+00
4.20342326e-01 4.90339309e-01 5.88021696e-01 -5.03428340e-01
8.74747813e-01 1.04043138e+00 3.47042888e-01 -9.91889358e-01
-5.56221426e-01 -5.82279325e-01 -1.51682174e+00 -3.44415009e-01
5.58256805e-01 -3.29204857e-01 -7.07645535e-01 6.37778759e-01
-1.14927781e+00 -3.83490235e-01 -5.92150390e-01 4.12716150e-01
-8.65730524e-01 4.64911789e-01 -2.95045793e-01 -1.07784462e+00
-1.80759057e-01 -1.05086493e+00 1.66912341e+00 -1.09549491e-02
-4.89749789e-01 -6.90523505e-01 -3.79680656e-02 3.32640082e-01
2.38790855e-01 3.48859906e-01 7.35104978e-01 4.00094949e-02
-6.93846166e-01 -5.34972966e-01 8.95983800e-02 -4.92375204e-03
3.04514766e-01 1.92113191e-01 -9.99318063e-01 -1.24739766e-01
-3.15592676e-01 -2.88790733e-01 3.08505684e-01 2.44538739e-01
9.53990221e-01 2.73886681e-01 -2.58325100e-01 5.89322031e-01
1.29982483e+00 -2.36367434e-01 5.46017647e-01 3.16401750e-01
7.89600492e-01 7.38741219e-01 7.66117990e-01 6.02390230e-01
7.07829222e-02 1.03404188e+00 6.82006896e-01 -6.50208294e-02
-4.89770800e-01 -1.71569481e-01 4.34759110e-02 5.78239381e-01
-7.66914248e-01 3.53108943e-01 -9.32214677e-01 2.30324686e-01
-1.30875969e+00 -6.80092335e-01 -3.98372978e-01 2.27184725e+00
7.15319693e-01 9.13047865e-02 4.02317166e-01 4.29241866e-01
1.38547212e-01 -2.42037058e-01 -6.31052852e-01 -2.09992304e-01
-1.85245857e-01 6.98625803e-01 3.86098683e-01 1.11980557e-01
-7.41652429e-01 8.39706600e-01 6.72334290e+00 2.84821212e-01
-1.10634828e+00 1.02136448e-01 -4.87565607e-01 -4.05056745e-01
-2.71640569e-01 -1.52955249e-01 -6.93992496e-01 8.22578371e-02
2.15928331e-01 5.05898178e-01 3.52804601e-01 9.96916056e-01
-3.57104912e-02 -4.21204686e-01 -1.20675528e+00 1.24590087e+00
1.33155018e-01 -1.15943265e+00 -1.69782400e-01 7.11707622e-02
6.19183123e-01 -3.23504694e-02 7.98955411e-02 -5.24852812e-01
-1.15103215e-01 -9.33637679e-01 1.37086499e+00 6.41534030e-01
1.06565082e+00 -5.50922334e-01 3.90907258e-01 4.45030779e-01
-1.17421305e+00 3.17600906e-01 -1.53243719e-02 9.05914679e-02
3.25314522e-01 4.87728626e-01 -6.34327829e-01 4.29049939e-01
1.01791537e+00 3.94438058e-01 -5.12543261e-01 8.27153504e-01
-5.55103384e-02 -8.79352316e-02 -7.06749618e-01 1.04206130e-01
6.65961504e-02 -9.90719944e-02 4.06221032e-01 1.14503598e+00
2.04436317e-01 2.86490202e-01 -1.81222990e-01 7.05926418e-01
1.30714312e-01 -1.66651562e-01 -6.26242995e-01 3.62371653e-02
1.81046501e-01 1.00520599e+00 -1.14775717e+00 -7.18548596e-02
-4.72374231e-01 1.24944615e+00 -2.46828701e-02 -1.56920195e-01
-4.53452498e-01 -2.08497703e-01 5.03315866e-01 7.22597897e-01
7.03812897e-01 -7.84478247e-01 -4.25821483e-01 -1.09916818e+00
5.18346906e-01 -5.91904104e-01 -1.12395771e-01 -9.41549897e-01
-1.05948603e+00 1.31735131e-01 2.22095065e-02 -1.40104258e+00
-2.02713460e-01 -1.00883627e+00 -1.44628748e-01 6.11999452e-01
-1.33017588e+00 -1.66288340e+00 -5.58159709e-01 8.54712248e-01
5.90191603e-01 2.01864451e-01 9.82750773e-01 -2.59665642e-02
4.20886166e-02 1.12017557e-01 1.85551688e-01 -2.53578305e-01
5.63959241e-01 -1.05651891e+00 3.03055972e-01 2.49755234e-01
1.83051825e-01 6.41734838e-01 5.04301012e-01 -6.75607979e-01
-2.25693822e+00 -2.87192911e-01 3.71349990e-01 -9.43439484e-01
2.83352286e-01 -5.35731196e-01 -6.28997982e-01 5.54169655e-01
-3.81107181e-01 3.36653382e-01 3.12830925e-01 1.29187778e-01
-5.00137269e-01 3.25318217e-01 -1.36304724e+00 3.42174649e-01
1.51196432e+00 -7.73525119e-01 -7.47740805e-01 2.25380659e-01
-6.70180991e-02 -8.10278893e-01 -1.07744813e+00 2.21776724e-01
1.53964841e+00 -1.31656015e+00 1.52135503e+00 -6.33629322e-01
2.89535373e-01 -2.34622627e-01 -3.34787548e-01 -7.59307921e-01
-9.79389399e-02 -5.30206501e-01 -4.53640759e-01 6.84333146e-01
-1.76485285e-01 -2.31547043e-01 1.15480363e+00 7.45113969e-01
4.67035696e-02 -6.60994530e-01 -1.04914248e+00 -9.76191998e-01
-2.69029707e-01 -6.90472722e-01 5.51770926e-01 5.72696626e-01
-1.53420269e-01 -1.04959212e-01 -2.11966082e-01 -2.14873016e-01
7.79387712e-01 6.37744009e-01 1.00163007e+00 -1.63087618e+00
-1.37016162e-01 -4.36213672e-01 -5.73018730e-01 -9.67416704e-01
-1.55470043e-01 -5.14884114e-01 -8.85827020e-02 -1.48169935e+00
1.23248532e-01 -7.16843605e-01 4.36244577e-01 4.57390010e-01
3.52952778e-01 5.71405172e-01 2.53953904e-01 4.27385956e-01
-3.64934593e-01 1.39388248e-01 1.40986502e+00 1.48254842e-01
-2.24940360e-01 5.38244732e-02 -4.54652533e-02 9.70146298e-01
4.00840282e-01 -1.75632849e-01 -2.79974025e-02 -5.70067406e-01
2.60922134e-01 -2.58791417e-01 6.07924640e-01 -6.21077538e-01
-7.85804242e-02 -2.00358167e-01 5.96211731e-01 -1.15781486e+00
6.67654037e-01 -1.37702203e+00 5.74619830e-01 5.30907333e-01
3.44904661e-01 -3.39342430e-02 3.41974907e-02 1.96933240e-01
2.15278849e-01 -1.64575636e-01 8.97800386e-01 -6.40220523e-01
-6.61309123e-01 2.22988978e-01 6.24769107e-02 -2.51967072e-01
8.20920467e-01 -9.50858653e-01 2.69217342e-01 -1.69074163e-01
-7.52893806e-01 -5.22850990e-01 1.11224759e+00 3.97192001e-01
9.42368507e-01 -1.29796183e+00 -3.44182253e-01 5.42133987e-01
-9.30098221e-02 5.35665572e-01 1.30578652e-01 8.58490705e-01
-8.06801975e-01 5.22081852e-01 -4.83154118e-01 -9.20370340e-01
-1.50122786e+00 2.69322962e-01 2.21711203e-01 2.22212762e-01
-9.93585527e-01 5.64087093e-01 -2.44816020e-01 -5.39135516e-01
4.95393649e-02 -5.47341585e-01 5.70439994e-01 -1.57745369e-02
3.12320858e-01 6.24293625e-01 4.74792868e-01 -6.95513785e-01
-5.47661483e-01 1.30090952e+00 4.18567598e-01 -1.16361998e-01
1.62686241e+00 3.77586000e-02 8.62464830e-02 4.14356649e-01
9.64218080e-01 2.42790356e-01 -1.28056669e+00 -1.35394111e-01
-2.75932044e-01 -9.70498264e-01 -1.02208890e-01 -7.45822370e-01
-9.54848289e-01 1.26639640e+00 4.08307880e-01 -1.77261680e-02
9.00013864e-01 5.15595019e-01 8.36738348e-01 3.71340990e-01
1.04989588e+00 -1.00407290e+00 2.28930041e-01 4.34773892e-01
1.00300431e+00 -1.01990569e+00 5.23123562e-01 -5.80425143e-01
-4.17369723e-01 1.48627639e+00 6.29448369e-02 8.76567811e-02
5.31728387e-01 5.69863677e-01 -2.68206984e-01 -5.36342442e-01
9.23135579e-02 -2.91167974e-01 4.44512606e-01 9.15888369e-01
3.51998359e-01 2.97372770e-02 2.53017038e-01 3.29447120e-01
-1.91467240e-01 3.74680497e-02 -2.93976236e-02 1.66046035e+00
-2.22819328e-01 -1.18745971e+00 -6.32994294e-01 1.78599611e-01
-2.23330632e-01 2.88857579e-01 -6.43673837e-01 1.18818033e+00
-1.09232262e-01 2.03055114e-01 6.95785880e-02 -4.02195007e-01
8.81340861e-01 1.11549467e-01 1.47591901e+00 -5.47422171e-01
-5.73016047e-01 2.38724694e-01 -8.23346302e-02 -8.85460317e-01
-6.51155412e-01 -9.19275820e-01 -1.09354317e+00 -3.94090474e-01
-3.50382388e-01 -8.23969722e-01 9.66055512e-01 6.44941509e-01
1.97483331e-01 -5.41902743e-02 3.84685457e-01 -1.61354542e+00
-5.04420042e-01 -6.33088112e-01 -9.39320862e-01 4.46265966e-01
2.86798000e-01 -1.20613575e+00 3.57457697e-02 1.24507740e-01] | [6.692664623260498, -1.0341027975082397] |
7f4464c4-407b-452d-b4ab-34461f546a2f | multi-branch-neural-networks-for-video | null | null | https://biblio.ugent.be/publication/8735199 | https://openaccess.thecvf.com/content/WACV2022/papers/Leroux_Multi-Branch_Neural_Networks_for_Video_Anomaly_Detection_in_Adverse_Lighting_WACV_2022_paper.pdf | Multi-branch Neural Networks for Video Anomaly Detection in Adverse Lighting and Weather Conditions | Automated anomaly detection in surveillance videos has attracted much interest as it provides a scalable alternative to manual monitoring. Most existing approaches achieve good performance on clean benchmark datasets recorded in well-controlled environments. However, detecting anomalies is much more challenging in the real world. Adverse weather conditions like rain or changing brightness levels cause a significant shift in the input data distribution, which in turn can lead to the detector model incorrectly reporting high anomaly scores. Additionally, surveillance cameras are usually deployed in evolving environments such as a city street of which the appearance changes over time because of seasonal changes or roadworks. The anomaly detection model will need to be updated periodically to deal with these issues. In this paper, we introduce a multi-branch model that is equipped with a trainable preprocessing step and multiple identical branches for detecting anomalies during day and night as well as in sunny and rainy conditions. We experimentally validate our approach on a distorted version of the Avenue dataset and provide qualitative results on real-world surveillance camera data. Experimental results show that our method outperforms the existing methods in terms of detection accuracy while being faster and more robust on scenes with varying visibility. | ['Pieter Simoens', 'Bo Li', 'Sam Leroux'] | 2021-10-10 | null | null | null | wacv-2021-10 | ['anomaly-detection-in-surveillance-videos', 'anomaly-detection-in-surveillance-videos'] | ['computer-vision', 'methodology'] | [ 3.91770482e-01 -6.30007386e-01 2.31269717e-01 -4.49821889e-01
-7.40194470e-02 -6.22495830e-01 5.29016972e-01 2.99000204e-01
-3.73912781e-01 3.13801527e-01 -1.42331421e-01 -2.61019677e-01
2.36219227e-01 -6.74245000e-01 -7.39612818e-01 -7.79683888e-01
-3.45376670e-01 5.33743761e-02 8.21652174e-01 -1.23724848e-01
-4.39429469e-02 5.23479342e-01 -1.78904331e+00 -1.90286934e-02
7.48621047e-01 8.56243908e-01 -1.17871225e-01 9.08425927e-01
2.15375155e-01 6.49012327e-01 -8.38170528e-01 -2.50733376e-01
5.78852832e-01 -8.84161443e-02 -1.21850610e-01 6.41411424e-01
7.67400980e-01 -6.52818859e-01 -1.83073163e-01 1.19030058e+00
8.74260068e-02 7.99896419e-02 2.48420224e-01 -1.34337091e+00
1.08256906e-01 -3.09914172e-01 -8.44130516e-01 9.11589265e-01
4.60635573e-01 4.73638564e-01 5.31464279e-01 -4.60701138e-01
3.59453529e-01 1.00540173e+00 6.37606859e-01 3.21453393e-01
-1.11563456e+00 -3.57935876e-01 6.65535331e-01 4.60939765e-01
-9.10911739e-01 -3.10403109e-01 6.56547546e-01 -3.76578540e-01
7.27493346e-01 3.10812563e-01 7.17516065e-01 1.33925235e+00
2.48536184e-01 6.68889165e-01 8.20168912e-01 -3.51575739e-03
3.19195777e-01 -1.26593679e-01 -2.66621262e-02 3.97944897e-01
6.00139797e-01 1.78782582e-01 -6.84721395e-02 -3.24531138e-01
3.26134443e-01 5.63884437e-01 -4.26837921e-01 -4.30233032e-01
-1.12573063e+00 5.87067842e-01 3.49887282e-01 5.26739750e-03
-7.05546677e-01 -1.62558228e-01 4.23346549e-01 6.47415817e-01
4.97331589e-01 2.55008936e-01 -4.23147678e-01 -1.30321220e-01
-8.52356374e-01 3.48420620e-01 5.96611917e-01 5.63457668e-01
3.70200396e-01 6.58668354e-02 -1.13670781e-01 6.31947041e-01
1.86047629e-01 7.57900298e-01 6.05649203e-02 -7.39830554e-01
3.22984487e-01 5.29558718e-01 4.25090551e-01 -1.36363018e+00
-5.33393562e-01 -3.45217139e-01 -9.65345860e-01 3.41665477e-01
6.29498243e-01 -1.77405909e-01 -1.06957328e+00 1.51699507e+00
6.02996528e-01 7.40281045e-01 -7.84778595e-02 1.01305783e+00
4.18771446e-01 8.68207276e-01 -2.51959190e-02 -4.48927969e-01
1.14526474e+00 -5.66394627e-01 -9.09253895e-01 -4.25390989e-01
3.56946886e-01 -7.72883117e-01 7.46658862e-01 5.79593360e-01
-5.32839894e-01 -3.19955945e-01 -7.57189333e-01 7.69865215e-01
-3.85454088e-01 -1.96849018e-01 2.07844153e-01 4.29829687e-01
-8.17975879e-01 1.10789109e-03 -1.07171738e+00 -8.59282374e-01
3.47114116e-01 -2.79465667e-03 -4.16079223e-01 -3.70665431e-01
-9.45997119e-01 8.66616726e-01 1.51346833e-01 4.06443566e-01
-9.76285219e-01 -4.02432233e-01 -9.97482598e-01 -3.09065819e-01
6.95361674e-01 -2.25562945e-01 1.21630251e+00 -1.09252179e+00
-1.02972794e+00 7.08882749e-01 -2.48258010e-01 -6.39634669e-01
6.31046772e-01 -3.22681129e-01 -8.84773195e-01 -2.71825325e-02
-6.09534271e-02 1.02965631e-01 1.06881428e+00 -1.10109234e+00
-1.02823758e+00 -4.52868909e-01 2.11354539e-01 1.41067207e-02
-7.47168362e-02 2.55810022e-01 -4.12688941e-01 -7.81368494e-01
-4.00562286e-02 -9.71400559e-01 -4.22067821e-01 8.78987610e-02
-7.40962550e-02 1.40227109e-01 1.30645216e+00 -6.88652575e-01
1.24459636e+00 -2.16194797e+00 -1.59411907e-01 3.33088398e-01
5.57389036e-02 5.64236939e-01 -4.81087305e-02 1.45425469e-01
8.27966034e-02 -4.40530121e-01 -4.88970518e-01 9.61179205e-04
-6.16995633e-01 5.55451810e-01 -2.73477226e-01 6.82260931e-01
4.19567972e-01 1.32220104e-01 -1.10093391e+00 -7.75210783e-02
3.81847799e-01 1.31949231e-01 -4.97621775e-01 4.82295811e-01
-1.41718060e-01 7.36119747e-01 -1.38756290e-01 9.30228055e-01
7.58465230e-01 7.39858821e-02 -2.69829422e-01 2.60045260e-01
-1.66121796e-01 -2.02083796e-01 -1.25105703e+00 1.16657090e+00
-8.76427218e-02 8.98593843e-01 2.34784111e-01 -9.91985202e-01
6.91671312e-01 2.89363533e-01 2.99661666e-01 -5.73001266e-01
-1.48656834e-02 -5.91598377e-02 7.71725327e-02 -8.94756734e-01
4.24866468e-01 4.19972450e-01 1.64515838e-01 9.72937867e-02
-5.15715599e-01 1.76711008e-01 4.21565264e-01 -2.98493691e-02
1.51592588e+00 -2.80989349e-01 2.41211578e-01 1.15731247e-01
6.57966495e-01 3.20990011e-02 9.23618257e-01 8.94384384e-01
-5.94419241e-01 4.35026199e-01 5.72078347e-01 -9.65851367e-01
-7.54721582e-01 -1.11332536e+00 -6.50326982e-02 8.69166851e-01
2.21676782e-01 -1.16283029e-01 -5.54878175e-01 -6.87332332e-01
-2.49320507e-01 6.64743483e-01 -6.76166177e-01 -2.05255419e-01
-5.91001809e-01 -1.26606739e+00 1.68129206e-01 2.97191650e-01
6.52962327e-01 -1.03569460e+00 -1.16291904e+00 1.73129141e-01
-1.97039083e-01 -1.42479539e+00 -1.12618506e-01 -2.40280151e-01
-7.58268595e-01 -1.42774010e+00 -4.22309816e-01 -2.97722578e-01
9.39755082e-01 6.03008747e-01 1.06401074e+00 1.59294531e-01
-6.13346815e-01 5.54398060e-01 -5.13483703e-01 -7.07052946e-01
-3.82305712e-01 -4.04616117e-01 1.46256268e-01 3.87719750e-01
6.70754194e-01 -1.90100774e-01 -5.73193431e-01 5.09020388e-01
-1.37192500e+00 -4.03676540e-01 3.95090342e-01 7.54815578e-01
2.80077010e-01 4.09450710e-01 1.31222442e-01 -6.64509058e-01
2.96311438e-01 -7.58104503e-01 -1.23009908e+00 1.35186892e-02
-2.17737615e-01 -3.41561794e-01 5.16273081e-01 -3.53424817e-01
-1.00677586e+00 1.38096660e-01 1.73927262e-01 -3.62618953e-01
-6.48950517e-01 1.67197645e-01 5.33038229e-02 2.72434294e-01
6.09252274e-01 7.92293623e-02 -1.13205276e-01 -1.74899921e-01
-2.68018395e-01 4.54936534e-01 7.05563426e-01 6.43793344e-02
1.24519038e+00 8.25120509e-01 -1.05328932e-01 -1.14459229e+00
-8.31464648e-01 -6.49271727e-01 -4.62832600e-01 -4.29288954e-01
8.37706089e-01 -7.75543869e-01 -1.22409627e-01 9.94963586e-01
-1.11900723e+00 -1.90026045e-01 1.31019190e-01 4.01993573e-01
9.48688909e-02 4.69801515e-01 -1.89291105e-01 -8.76332939e-01
-5.38298511e-04 -1.01926184e+00 9.61976528e-01 4.14776802e-01
1.86516549e-02 -9.74400043e-01 3.19485217e-01 2.23544631e-02
6.26559794e-01 7.53775001e-01 3.13078284e-01 -4.03308421e-01
-4.86360520e-01 -5.74179471e-01 -2.77819447e-02 4.14279908e-01
4.94130403e-01 3.67034912e-01 -9.73867893e-01 -5.41317165e-01
-1.29336834e-01 2.50425130e-01 8.73155117e-01 4.21488047e-01
1.17826664e+00 -2.82795697e-01 -1.98235929e-01 5.68763137e-01
1.16290426e+00 2.15154886e-01 5.11261702e-01 5.62793851e-01
4.23684418e-01 5.64553261e-01 8.94414604e-01 6.82426631e-01
1.97142169e-01 7.00130641e-01 1.00965130e+00 -3.12147886e-01
5.08084953e-01 3.71575624e-01 6.69845283e-01 7.58595020e-02
-2.19780244e-02 -3.49019766e-01 -1.00302362e+00 7.84782767e-01
-1.91016972e+00 -1.34549046e+00 -2.37727106e-01 2.64587235e+00
3.06490839e-01 1.63571149e-01 2.53479302e-01 2.23608941e-01
6.99982047e-01 3.03046584e-01 -5.17726481e-01 -2.87099868e-01
-2.03776658e-01 -3.85177076e-01 4.90558714e-01 2.44517609e-01
-1.59158731e+00 5.04120350e-01 6.09889984e+00 -1.78813979e-01
-1.18531156e+00 -2.26433605e-01 5.34599006e-01 -3.26840699e-01
4.23036605e-01 -2.95557737e-01 -4.42241013e-01 7.81784058e-01
7.80542076e-01 3.47364217e-01 1.94076285e-01 6.85408890e-01
5.51214755e-01 -4.29759324e-01 -7.65733898e-01 8.84400904e-01
9.51994956e-02 -6.22113645e-01 -1.81676969e-01 -1.20498404e-01
6.97265625e-01 3.13963234e-01 -5.22705019e-02 -2.18284260e-02
1.95923254e-01 -7.16206074e-01 1.94651306e-01 3.88599604e-01
1.89131916e-01 -6.34427726e-01 1.13880742e+00 3.05031329e-01
-9.98192906e-01 -2.08192602e-01 -3.00593287e-01 -1.46063536e-01
9.10495147e-02 7.12722242e-01 -7.04738259e-01 1.50441796e-01
1.21527719e+00 9.01027858e-01 -8.21111381e-01 1.44520986e+00
-3.29062045e-01 7.42489159e-01 -6.03644848e-01 3.53503615e-01
4.07399476e-01 -1.58905670e-01 1.06561279e+00 1.19977522e+00
3.32902223e-01 -2.25608163e-02 2.81418055e-01 2.03654662e-01
3.98487449e-01 -3.02005857e-01 -1.05302167e+00 5.17328739e-01
1.12226427e-01 1.21502388e+00 -5.52409530e-01 -2.94054002e-01
-7.67132401e-01 1.00632524e+00 -1.83343560e-01 5.17280638e-01
-1.08601248e+00 -7.93067813e-02 9.71330822e-01 1.58476934e-01
4.25442308e-01 -8.26345533e-02 4.09149110e-01 -1.39216709e+00
4.01971996e-01 -1.03238606e+00 6.49095953e-01 -3.53824347e-01
-1.24178219e+00 6.06811166e-01 4.70084250e-02 -1.50151908e+00
-2.88469166e-01 -6.51759148e-01 -9.58470345e-01 2.72756368e-01
-1.74275470e+00 -5.78754306e-01 -8.82480741e-01 7.03434885e-01
7.49797881e-01 -1.18061304e-01 6.47116721e-01 3.39208126e-01
-1.00068998e+00 2.53374577e-01 1.74668431e-01 1.43859491e-01
7.70828784e-01 -1.19951093e+00 4.19825673e-01 1.65547848e+00
1.49896191e-02 -2.23891120e-02 1.16417778e+00 -3.78378212e-01
-1.15232742e+00 -1.39655387e+00 3.01509887e-01 -4.94296223e-01
7.23782063e-01 -1.78795233e-01 -1.25754333e+00 6.24885499e-01
7.78997242e-02 4.18748766e-01 5.42848945e-01 -1.94308862e-01
-2.10361078e-01 -2.89327443e-01 -1.19452989e+00 4.08432454e-01
6.91469133e-01 9.04082209e-02 -5.20261288e-01 3.59031498e-01
4.45528030e-02 -4.30756420e-01 -2.19348788e-01 6.15266263e-01
2.42007941e-01 -1.23408806e+00 5.76009214e-01 -5.72702169e-01
-3.18870842e-02 -7.08296359e-01 -1.58935096e-02 -1.52708912e+00
-6.36736155e-02 -5.65483570e-01 -3.08005810e-01 7.32401907e-01
1.88957319e-01 -8.29750121e-01 3.21597070e-01 4.48358953e-01
2.06741363e-01 -3.52462083e-01 -8.40826273e-01 -8.21688414e-01
-8.03242087e-01 -5.10470748e-01 3.42541158e-01 7.41059721e-01
-5.19233823e-01 -1.80958897e-01 -5.54409862e-01 8.61468792e-01
6.58923924e-01 -1.36517033e-01 1.07720935e+00 -1.38777018e+00
5.61634116e-02 -2.44763941e-01 -9.18506503e-01 -6.08198583e-01
-2.11018085e-01 4.37645800e-02 3.57249767e-01 -1.15368915e+00
-7.34842420e-02 -5.08604385e-02 -3.44976634e-01 2.70480841e-01
-4.52214271e-01 2.47390866e-01 -2.29522079e-01 4.89558801e-02
-7.68707097e-01 3.06405216e-01 5.57606936e-01 -1.09477334e-01
-2.44709790e-01 4.60142821e-01 -9.81396344e-03 1.06718242e+00
9.17547464e-01 -3.62048358e-01 -2.59083301e-01 -6.42095268e-01
1.51016027e-01 -3.73273253e-01 5.79480708e-01 -1.32752812e+00
-3.58456671e-02 -2.68316925e-01 3.52409780e-01 -5.53865135e-01
8.61838162e-02 -1.42458725e+00 -1.20594941e-01 5.95314026e-01
1.55833766e-01 6.39945507e-01 4.12903249e-01 9.98276174e-01
-3.29627842e-01 1.78533807e-01 8.81237805e-01 1.48755044e-01
-9.46019709e-01 3.15614939e-01 -7.73052454e-01 -1.57732263e-01
1.42103755e+00 -2.62345344e-01 -3.27748537e-01 -5.37693262e-01
-4.48244900e-01 5.75079262e-01 6.39580667e-01 8.89907718e-01
5.25856495e-01 -9.45509136e-01 -8.65418196e-01 5.47198176e-01
4.60922569e-01 8.40985179e-02 1.63457751e-01 1.03248477e+00
-6.76280737e-01 1.17959939e-01 -2.26822883e-01 -9.65600908e-01
-1.51327217e+00 5.00325620e-01 4.04916376e-01 -2.39217669e-01
-5.96094191e-01 4.24199253e-01 1.02783166e-01 -7.23638805e-03
5.02382040e-01 -5.97022116e-01 -3.87214065e-01 -2.50091348e-02
1.02834034e+00 4.06495422e-01 1.53783886e-02 -7.08932102e-01
-3.68764549e-01 4.01522219e-01 -2.10912749e-01 3.93615246e-01
1.15353882e+00 -4.01790589e-02 -2.88514607e-03 3.87823403e-01
5.61424077e-01 -1.81653321e-01 -1.45555997e+00 -4.37872559e-01
9.21062902e-02 -7.07951188e-01 2.83278301e-02 -4.59253013e-01
-1.28345323e+00 4.93673086e-01 1.00387561e+00 5.58874905e-01
1.62152696e+00 -2.31173724e-01 7.27907479e-01 6.01478457e-01
8.38361084e-02 -1.05064785e+00 -1.22867413e-01 4.12700683e-01
6.10046387e-01 -1.79980779e+00 -1.60919055e-01 -1.53089643e-01
-5.59952915e-01 9.50837910e-01 7.66920149e-01 -1.30401522e-01
5.39836824e-01 8.18036422e-02 4.16535348e-01 -2.75485039e-01
-8.29694211e-01 -1.95812553e-01 1.44680053e-01 7.35182226e-01
-1.46738151e-02 -1.06271915e-01 1.77406907e-01 -3.42896372e-01
2.79805988e-01 -3.72513354e-01 7.00652242e-01 1.12290788e+00
-4.09190834e-01 -5.85761130e-01 -8.45641017e-01 5.64718544e-01
-6.54865742e-01 3.18644404e-01 -2.50446111e-01 6.41435623e-01
8.16219226e-02 1.20656455e+00 5.22322416e-01 -3.44646610e-02
6.06720805e-01 -1.46893561e-01 -1.83735490e-02 -2.86473095e-01
-2.43740037e-01 -9.75348055e-02 -1.50940672e-01 -1.03099036e+00
-7.15355575e-01 -9.29052591e-01 -8.08530748e-01 -1.69695914e-01
-1.72868133e-01 -1.60670429e-01 5.03546000e-01 9.79553699e-01
2.17787251e-01 5.48917711e-01 8.32919896e-01 -6.67890847e-01
-2.22151667e-01 -8.86238754e-01 -2.19136521e-01 6.08864307e-01
9.88482594e-01 -6.96798146e-01 -5.65268517e-01 1.48909345e-01] | [7.858386516571045, 1.5775470733642578] |
ccaa68de-acb7-4fec-afaf-9c717a393b54 | defining-data-science-a-new-field-of-inquiry | 2306.16177 | null | https://arxiv.org/abs/2306.16177v2 | https://arxiv.org/pdf/2306.16177v2.pdf | Defining data science: a new field of inquiry | Data science is not a science. It is a research paradigm. Its power, scope, and scale will surpass science, our most powerful research paradigm, to enable knowledge discovery and change our world. We have yet to understand and define it, vital to realizing its potential and managing its risks. Modern data science is in its infancy. Emerging slowly since 1962 and rapidly since 2000, it is a fundamentally new field of inquiry, one of the most active, powerful, and rapidly evolving 21st century innovations. Due to its value, power, and applicability, it is emerging in 40+ disciplines, hundreds of research areas, and thousands of applications. Millions of data science publications contain myriad definitions of data science and data science problem solving. Due to its infancy, many definitions are independent, application-specific, mutually incomplete, redundant, or inconsistent, hence so is data science. This research addresses this data science multiple definitions challenge by proposing the development of coherent, unified definition based on a data science reference framework using a data science journal for the data science community to achieve such a definition. This paper provides candidate definitions for essential data science artifacts that are required to discuss such a definition. They are based on the classical research paradigm concept consisting of a philosophy of data science, the data science problem solving paradigm, and the six component data science reference framework (axiology, ontology, epistemology, methodology, methods, technology) that is a frequently called for unifying framework with which to define, unify, and evolve data science. It presents challenges for defining data science, solution approaches, i.e., means for defining data science, and their requirements and benefits as the basis of a comprehensive solution. | ['Michael L Brodie'] | 2023-06-28 | null | null | null | null | ['philosophy'] | ['miscellaneous'] | [-2.00306132e-01 -2.47031480e-01 -6.51865125e-01 -1.51777595e-01
1.33818984e-01 -7.83988297e-01 8.00166547e-01 4.10408318e-01
-4.11779374e-01 2.56650954e-01 3.98331940e-01 -5.99683046e-01
-9.41355705e-01 -8.01656008e-01 -3.96559209e-01 -1.78600475e-01
3.26727033e-01 4.69915986e-01 1.09474482e-02 -2.82916307e-01
8.67121279e-01 3.98853719e-01 -2.20245910e+00 2.47066598e-02
8.95257950e-01 1.07561541e+00 2.63856590e-01 -6.65833727e-02
-7.85003126e-01 6.46551013e-01 -6.53137028e-01 -1.60478592e-01
1.48101404e-01 -9.30875465e-02 -1.02766657e+00 -5.20694137e-01
1.39960885e-01 9.00102854e-02 1.59186855e-01 9.15187716e-01
1.71659470e-01 -4.95531708e-01 4.06320766e-02 -1.61690533e+00
-1.19444561e+00 3.09373319e-01 -3.55302960e-01 2.30666712e-01
5.73707342e-01 -1.95892110e-01 5.90849161e-01 -8.65332067e-01
9.27729309e-01 1.00941086e+00 6.61067426e-01 4.87942100e-01
-1.01319945e+00 -8.66480172e-01 -1.11410990e-01 2.37782270e-01
-1.21657658e+00 -3.71348649e-01 4.70260829e-01 -8.71113777e-01
7.53935695e-01 7.14302301e-01 1.04340553e+00 7.99610376e-01
2.26714551e-01 2.12990493e-01 1.04750478e+00 -6.41155124e-01
2.54354149e-01 5.06301105e-01 8.89361560e-01 -5.30808792e-02
1.08595490e+00 1.70463800e-01 -6.07964754e-01 -2.70176828e-01
5.49300373e-01 5.27201533e-01 -1.76717956e-02 -1.69583336e-01
-1.10607564e+00 4.88231063e-01 -1.32490993e-01 8.23104799e-01
-4.49534833e-01 -3.23989481e-01 2.76883125e-01 4.69221652e-01
4.21726704e-02 6.81569517e-01 -4.85040784e-01 -6.60878658e-01
-3.49437565e-01 4.43324059e-01 1.08207214e+00 1.04826796e+00
5.28743863e-01 -1.21764295e-01 3.86398613e-01 3.44315946e-01
6.30936444e-01 4.61347401e-01 6.35117590e-01 -7.95692742e-01
-8.83176476e-02 1.45006168e+00 5.81975654e-02 -1.00008237e+00
-4.66683537e-01 -2.88919181e-01 -5.98626733e-01 1.92509428e-01
1.01682149e-01 1.98351920e-01 -7.40758240e-01 1.41140687e+00
4.84236658e-01 -1.99577063e-01 2.68183738e-01 6.88828826e-01
1.24496567e+00 3.14688474e-01 2.74554074e-01 -3.44022959e-01
1.38217974e+00 -1.12336427e-01 -1.15770197e+00 8.97837356e-02
4.79085326e-01 -7.76901305e-01 1.11873186e+00 7.38432050e-01
-1.16701937e+00 -3.90689760e-01 -8.59977067e-01 -4.95246202e-01
-9.09634233e-01 -7.23908484e-01 6.83604181e-01 8.86354685e-01
-8.17778766e-01 5.27327508e-02 -4.09916878e-01 -7.98577905e-01
3.10181230e-01 -1.82183117e-01 -2.08758742e-01 1.52998403e-01
-1.17085648e+00 1.23599505e+00 5.85607529e-01 -2.80702472e-01
-4.27803516e-01 -1.14163148e+00 -3.01263154e-01 -2.77028352e-01
3.46952289e-01 -8.29119444e-01 5.03951073e-01 -5.03811061e-01
-7.15951085e-01 8.53543401e-01 -1.55890837e-01 -2.39961937e-01
-1.11298896e-01 -2.72593588e-01 -1.07737243e+00 -4.87687528e-01
1.29712731e-01 -3.98393989e-01 1.50190387e-02 -1.61870253e+00
-1.04323339e+00 -8.82028461e-01 -2.82782853e-01 -5.07621393e-02
-5.21676838e-01 5.15110612e-01 -1.72085956e-01 -3.46779644e-01
1.18865632e-01 -1.88901603e-01 -9.67778265e-03 -6.18927330e-02
4.35724020e-01 -3.25911582e-01 1.09113300e+00 -3.17368299e-01
1.93291056e+00 -2.38265920e+00 -1.65728405e-01 3.20559263e-01
6.36711895e-01 4.45863694e-01 4.14995521e-01 1.05915654e+00
1.71342760e-01 8.05438399e-01 1.08695731e-01 4.99661535e-01
6.03522137e-02 4.47787762e-01 -4.24384236e-01 1.81732997e-01
-2.99789608e-01 6.09298468e-01 -9.33602631e-01 3.83439884e-02
2.84921527e-01 4.68402386e-01 -1.68336004e-01 8.64889324e-02
-8.10902938e-02 1.66411057e-01 -5.38888454e-01 7.86017478e-01
8.75580430e-01 -1.62094280e-01 3.78434777e-01 -1.95161089e-01
-9.51448202e-01 7.08531216e-02 -1.40921044e+00 1.48363781e+00
-2.81462938e-01 3.44485193e-01 1.01919703e-01 -8.15767050e-01
1.40995049e+00 5.39276659e-01 6.69989467e-01 -9.54309404e-01
-7.47987777e-02 5.79429686e-01 1.49384782e-01 -1.24440897e+00
4.29087073e-01 -1.19498661e-02 6.00841790e-02 4.94213939e-01
-3.80158186e-01 -1.15725927e-01 2.62063816e-02 4.85869423e-02
1.11483824e+00 -1.48556054e-01 4.37361598e-01 -6.12541974e-01
3.41000617e-01 2.97949463e-01 8.68061364e-01 3.92508775e-01
-8.19034651e-02 3.11180025e-01 1.31413311e-01 -8.66311491e-01
-9.83269274e-01 -1.03360283e+00 -5.73438466e-01 7.29266167e-01
4.50481802e-01 -7.50464618e-01 -2.70697445e-01 -1.62722602e-01
3.33102882e-01 7.95853615e-01 -5.33952534e-01 2.40065709e-01
-9.58345011e-02 -4.57141161e-01 3.29472125e-01 -5.75288478e-03
5.33061385e-01 -9.70833361e-01 -6.20685339e-01 2.16653556e-01
1.36124879e-01 -6.70581341e-01 5.72935402e-01 -3.08663100e-01
-9.40117240e-01 -1.42063630e+00 2.91643500e-01 -5.00002027e-01
3.01959127e-01 5.38417399e-01 1.33609903e+00 4.36114967e-01
-1.06763877e-01 3.87405962e-01 -6.07480347e-01 -1.18640316e+00
-3.70893687e-01 -2.54488230e-01 1.53814346e-01 -5.23167372e-01
1.29506600e+00 -5.12200892e-01 -3.61896694e-01 3.09398421e-03
-1.46725285e+00 -2.22229093e-01 5.28114855e-01 2.26780534e-01
4.12482530e-01 2.39503130e-01 1.03774643e+00 -9.60719883e-01
9.45870697e-01 -1.20920944e+00 -2.16538593e-01 6.64271951e-01
-1.46220016e+00 -4.86252725e-01 2.65672475e-01 2.36271665e-01
-9.49440837e-01 -7.82099128e-01 1.07194118e-01 9.08538476e-02
-2.95283109e-01 9.35724556e-01 -3.44774425e-01 2.96597004e-01
9.82340753e-01 1.35519728e-01 4.79983568e-01 -9.32515502e-01
3.52642685e-01 1.19882476e+00 3.30086797e-01 -7.40129054e-01
4.68602985e-01 6.53415620e-01 -2.88696826e-01 -9.03035402e-01
-2.85763890e-01 -5.82768500e-01 -4.87241238e-01 -2.22083554e-01
6.90808773e-01 -5.83761871e-01 -8.78240108e-01 1.55067399e-01
-1.18452072e+00 6.88579261e-01 -6.70470595e-01 4.07137603e-01
2.78381892e-02 -4.97822203e-02 3.12820047e-01 -8.22430551e-01
-5.16557693e-01 -8.07742119e-01 3.80885273e-01 3.28379571e-01
-3.09418201e-01 -1.07481098e+00 1.32528335e-01 5.93861639e-01
6.65416956e-01 7.07319319e-01 9.38429952e-01 -6.33191943e-01
-3.60606611e-01 -8.88988525e-02 -2.17531726e-01 1.25626713e-01
4.90871608e-01 4.00056124e-01 -5.77349424e-01 1.55331884e-02
4.87373412e-01 1.36437744e-01 2.81336755e-02 -2.03281015e-01
1.10335720e+00 -5.10558486e-01 -2.50771612e-01 4.94909346e-01
1.81546974e+00 7.58326411e-01 6.89570069e-01 7.15974152e-01
4.41627800e-01 9.48018789e-01 6.70193315e-01 6.47451878e-01
8.43814611e-01 1.26819640e-01 2.51351655e-01 2.39286348e-01
-2.61003040e-02 1.06557772e-01 -1.59932956e-01 1.02385890e+00
-4.18290287e-01 3.54410648e-01 -1.59529567e+00 6.71872318e-01
-1.99514186e+00 -1.06156528e+00 -9.18502569e-01 2.31946111e+00
8.05066943e-01 1.23215942e-02 2.35865161e-01 1.32637575e-01
3.20574254e-01 -2.94718564e-01 -3.41503739e-01 -6.02420807e-01
-1.04340799e-01 1.74587622e-01 2.12127954e-01 -5.36005059e-03
-3.13178360e-01 5.17904639e-01 6.66259432e+00 2.27785558e-01
-9.48138416e-01 2.48366863e-01 -2.10896045e-01 2.09359840e-01
-1.14110053e+00 1.95870534e-01 -6.45051301e-01 7.08031416e-01
9.94256616e-01 -9.81247425e-01 2.09986791e-01 6.46512210e-01
3.75629216e-01 -1.27322286e-01 -9.61866975e-01 9.36403215e-01
-6.46344498e-02 -1.93021405e+00 1.78343862e-01 2.78939784e-01
5.32653153e-01 5.38930073e-02 -2.06525065e-02 -2.04123452e-01
2.42995501e-01 -9.71308947e-01 6.69695258e-01 8.19678307e-01
4.94685233e-01 -4.38568294e-01 6.86432838e-01 2.48469070e-01
-6.16747558e-01 -2.30769172e-01 -1.80227011e-01 -7.08661854e-01
1.40704382e-02 7.28882611e-01 -8.16322863e-03 1.09704912e+00
1.21714282e+00 7.51157343e-01 -2.19369009e-01 9.83564377e-01
2.48026907e-01 3.99770707e-01 -1.73127353e-01 -8.71054269e-03
-2.62994975e-01 -3.31068069e-01 4.35474008e-01 9.07531381e-01
2.14222208e-01 1.85508236e-01 -1.65187642e-01 7.78876364e-01
4.30206031e-01 -3.94713730e-02 -9.62510884e-01 -2.33743832e-01
1.45827651e+00 8.10292661e-01 -1.23166479e-01 -6.14406988e-02
-7.96925724e-01 -2.19462186e-01 -8.05967376e-02 2.19539538e-01
-1.54059365e-01 -2.73311704e-01 9.02041852e-01 3.98449510e-01
-5.78387618e-01 -2.08085537e-01 -1.22155166e+00 -8.12200248e-01
5.70055693e-02 -1.24886298e+00 5.08731663e-01 -1.57273278e-01
-1.64085877e+00 2.50240117e-01 1.29977375e-01 -1.19927835e+00
1.26369804e-01 -3.47766966e-01 -2.47770131e-01 9.00942683e-01
-1.23019147e+00 -1.34753633e+00 -7.32074440e-01 3.97485793e-01
-2.02333570e-01 -2.64591813e-01 1.02212572e+00 4.49455947e-01
-5.13487458e-01 -1.22226305e-01 4.62893397e-01 -2.74872124e-01
5.13493776e-01 -9.83360887e-01 2.87699878e-01 5.86855948e-01
-4.05802280e-01 1.22225606e+00 5.84696412e-01 -7.56464303e-01
-2.00493670e+00 -4.93056327e-01 1.02550972e+00 -9.30790603e-01
8.71462643e-01 -3.81430268e-01 -1.19576037e+00 6.79611981e-01
1.02472594e-02 -4.28920567e-01 1.24579155e+00 4.12475318e-01
-3.62160027e-01 -6.99276328e-01 -1.42418993e+00 2.94717401e-01
1.06062102e+00 -1.94509923e-01 -1.28528559e+00 -6.31016642e-02
7.38497019e-01 1.46006674e-01 -1.43825471e+00 4.95701283e-01
9.23782408e-01 -1.04928815e+00 7.29437053e-01 -5.98044217e-01
8.07183683e-02 -4.26681399e-01 -2.18989521e-01 -4.90817577e-01
-3.62024456e-01 -6.69547021e-01 -1.66135669e-01 1.58136594e+00
1.10864088e-01 -1.05769682e+00 4.36371922e-01 1.24266398e+00
-2.65271157e-01 -5.62516212e-01 -8.32318246e-01 -8.56311500e-01
2.27876872e-01 -5.80412686e-01 1.28859830e+00 1.60826635e+00
2.38457933e-01 -7.22004147e-03 2.25471050e-01 -9.71933007e-02
8.68807256e-01 3.70690934e-02 1.07457495e+00 -1.91512001e+00
5.49449801e-01 -6.63325429e-01 -5.14792860e-01 -4.54305977e-01
-7.13416696e-01 -7.01886475e-01 -7.47685552e-01 -2.18211031e+00
7.99869224e-02 -8.62407088e-01 -2.57498980e-01 1.86004579e-01
1.76369414e-01 -3.18245679e-01 1.01852916e-01 8.33573878e-01
7.50758424e-02 -1.11760333e-01 1.04861248e+00 4.40955609e-01
-2.99186826e-01 -5.80946565e-01 -1.39296591e+00 6.06520295e-01
6.83836341e-01 -2.42355749e-01 -6.02174461e-01 -5.32630920e-01
8.81003797e-01 -4.87282813e-01 2.42096052e-01 -8.03284764e-01
4.91671473e-01 -8.74773860e-01 -7.50303119e-02 -5.30961871e-01
-2.80588925e-01 -1.46697557e+00 9.89574671e-01 4.32600826e-01
1.83371291e-01 9.44282934e-02 5.81085049e-02 8.38230029e-02
-3.68170947e-01 -3.33899349e-01 3.00919980e-01 3.21914963e-02
-1.14115763e+00 1.36341602e-01 7.65195489e-02 2.28777960e-01
1.52306402e+00 -6.71481013e-01 -9.85288203e-01 3.13072652e-01
-3.91975254e-01 3.26685250e-01 7.12599516e-01 9.30913568e-01
5.46590626e-01 -1.20850515e+00 -6.11403763e-01 4.16609168e-01
3.35556060e-01 2.44208395e-01 9.09297541e-02 4.50774938e-01
-5.19095004e-01 3.55691701e-01 -5.18135786e-01 -1.83175996e-01
-8.20716262e-01 8.44825864e-01 -1.00447319e-01 4.05478418e-01
-4.66904849e-01 2.35638082e-01 6.53941557e-02 -2.07567036e-01
3.17036003e-01 -1.32556379e-01 -4.15909529e-01 1.44131139e-01
9.14531589e-01 8.53762984e-01 -9.39859971e-02 -3.57564121e-01
-6.02567017e-01 5.83066463e-01 1.14221781e-01 1.69204935e-01
1.53118277e+00 -3.18679780e-01 -7.47205913e-01 9.63407934e-01
6.08610272e-01 -1.12743519e-01 -1.52479485e-01 -5.87580204e-02
5.30939460e-01 -8.74445856e-01 -2.49201106e-03 -1.34820819e+00
-4.79272425e-01 3.36466014e-01 7.07461953e-01 9.95570064e-01
1.06878543e+00 9.78268310e-02 3.45238388e-01 -3.30175936e-01
4.47996259e-01 -1.21666300e+00 -4.94508177e-01 2.35667989e-01
1.29498351e+00 -1.11410594e+00 2.24519581e-01 -3.22435886e-01
-1.97092488e-01 1.13900936e+00 6.11188054e-01 3.21579278e-01
1.42284191e+00 5.69783747e-01 2.56133854e-01 -8.58505368e-01
-5.86141348e-01 -7.34311044e-02 5.50494008e-02 9.74744320e-01
6.84731960e-01 -9.17060375e-02 -1.33735192e+00 8.39887857e-01
-1.84136987e-01 6.82701290e-01 2.09071428e-01 1.22153044e+00
-6.04655445e-01 -1.49638402e+00 -6.00502312e-01 5.52369177e-01
-6.03335202e-01 1.09392971e-01 -5.61500490e-01 1.18230760e+00
7.78184354e-01 1.11399066e+00 4.33189452e-01 -6.73927486e-01
7.35644400e-01 -1.61342233e-01 -4.21493575e-02 -4.07198817e-01
-9.38715637e-01 -5.20892859e-01 -2.36965287e-02 -3.01662415e-01
-8.21929991e-01 -7.05915451e-01 -1.31894290e+00 -9.65840042e-01
-3.42177562e-02 4.19318825e-01 1.21462524e+00 7.47107029e-01
1.07047963e+00 1.66579962e-01 4.25489515e-01 4.12595689e-01
-5.62463552e-02 -8.05392981e-01 -4.78478432e-01 5.22622406e-01
3.41839977e-02 -6.18871152e-01 -1.30174503e-01 -1.64870527e-02] | [9.043131828308105, 7.143357753753662] |
f4736690-25c2-41ab-af86-2e0dec332ef7 | reference-based-image-and-video-super | 2212.09581 | null | https://arxiv.org/abs/2212.09581v2 | https://arxiv.org/pdf/2212.09581v2.pdf | Reference-based Image and Video Super-Resolution via C2-Matching | Reference-based Super-Resolution (Ref-SR) has recently emerged as a promising paradigm to enhance a low-resolution (LR) input image or video by introducing an additional high-resolution (HR) reference image. Existing Ref-SR methods mostly rely on implicit correspondence matching to borrow HR textures from reference images to compensate for the information loss in input images. However, performing local transfer is difficult because of two gaps between input and reference images: the transformation gap (e.g., scale and rotation) and the resolution gap (e.g., HR and LR). To tackle these challenges, we propose C2-Matching in this work, which performs explicit robust matching crossing transformation and resolution. 1) To bridge the transformation gap, we propose a contrastive correspondence network, which learns transformation-robust correspondences using augmented views of the input image. 2) To address the resolution gap, we adopt teacher-student correlation distillation, which distills knowledge from the easier HR-HR matching to guide the more ambiguous LR-HR matching. 3) Finally, we design a dynamic aggregation module to address the potential misalignment issue between input images and reference images. In addition, to faithfully evaluate the performance of Reference-based Image Super-Resolution under a realistic setting, we contribute the Webly-Referenced SR (WR-SR) dataset, mimicking the practical usage scenario. We also extend C2-Matching to Reference-based Video Super-Resolution task, where an image taken in a similar scene serves as the HR reference image. Extensive experiments demonstrate that our proposed C2-Matching significantly outperforms state of the arts on the standard CUFED5 benchmark and also boosts the performance of video SR by incorporating the C2-Matching component into Video SR pipelines. | ['Ziwei Liu', 'Chen Change Loy', 'Xintao Wang', 'Kelvin C. K. Chan', 'Yuming Jiang'] | 2022-12-19 | null | null | null | null | ['video-super-resolution', 'reference-based-video-super-resolution', 'reference-based-super-resolution'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 6.29277170e-01 -2.44544744e-01 -8.89724940e-02 -2.29026139e-01
-1.31308472e+00 -2.14136004e-01 5.62184691e-01 -5.75950325e-01
-2.33869165e-01 6.46539927e-01 4.05137688e-01 1.34249598e-01
-3.08271758e-02 -7.90638685e-01 -9.21332598e-01 -6.80790246e-01
3.48560303e-01 -1.13890387e-01 4.82303888e-01 -4.06880498e-01
2.25687295e-01 4.24093992e-01 -1.66635358e+00 4.57445800e-01
1.04586637e+00 9.73864138e-01 5.29647052e-01 3.98925215e-01
9.73593444e-02 5.63641071e-01 -3.99154842e-01 -1.83949009e-01
6.22132242e-01 -5.19045472e-01 -7.21605003e-01 9.51023847e-02
9.89177942e-01 -6.42315388e-01 -4.64035302e-01 1.22503304e+00
5.09632707e-01 3.79863769e-01 9.74236950e-02 -6.69854522e-01
-9.35767829e-01 4.44246948e-01 -1.07895958e+00 5.80341399e-01
5.88907480e-01 1.28268048e-01 7.31904089e-01 -1.15562940e+00
8.07538927e-01 1.48789966e+00 4.00210768e-01 5.25748253e-01
-1.39088225e+00 -9.41850543e-01 2.10787177e-01 2.31755286e-01
-1.46073627e+00 -5.64103007e-01 8.45164359e-01 -1.22625582e-01
4.00896847e-01 1.97617337e-01 2.20752954e-01 1.15630698e+00
-1.33227378e-01 3.82499099e-01 1.34800506e+00 -2.09671661e-01
-1.12149179e-01 2.10597669e-03 -2.38705903e-01 3.14051241e-01
-2.39425898e-02 5.56217968e-01 -6.07059658e-01 2.47667491e-01
1.39385867e+00 -1.88207887e-02 -7.58573055e-01 -1.69589937e-01
-1.28525651e+00 4.56560016e-01 6.47087991e-01 2.59313971e-01
-1.13160320e-01 -2.89487869e-01 1.09125614e-01 3.59306216e-01
4.32769597e-01 2.98764318e-01 -3.01309228e-01 2.28949666e-01
-1.03594530e+00 3.58033590e-02 3.72116379e-02 9.81025577e-01
7.73985684e-01 1.37943670e-01 -3.16165179e-01 1.13523436e+00
-1.54317349e-01 3.95098627e-01 5.17458618e-01 -1.11972988e+00
7.65481591e-01 1.49451599e-01 2.22195968e-01 -9.68144655e-01
2.98594944e-02 -5.34055173e-01 -1.12481582e+00 2.33443812e-01
3.44710976e-01 4.56066459e-01 -7.99965739e-01 1.81292927e+00
4.05691832e-01 6.14340007e-01 1.50731429e-01 1.39197469e+00
1.01307821e+00 4.18243825e-01 -2.42867991e-01 -4.26794767e-01
1.22571838e+00 -1.08394086e+00 -5.63487232e-01 -6.87110098e-03
-2.94380579e-02 -9.07192469e-01 1.18861449e+00 2.70080060e-01
-1.45926976e+00 -1.16471910e+00 -1.09255457e+00 -3.96685630e-01
5.36262000e-04 5.38727492e-02 -1.41058955e-02 1.04829758e-01
-1.11448514e+00 7.17017889e-01 -5.02264619e-01 -9.60461050e-02
1.88238770e-01 4.45410162e-02 -5.18161178e-01 -4.60785180e-01
-1.37317991e+00 7.72250295e-01 2.59916544e-01 1.00976847e-01
-6.09614849e-01 -9.06043351e-01 -9.29203629e-01 -1.82377949e-01
5.48037112e-01 -5.99033058e-01 8.23586464e-01 -8.82394612e-01
-1.50885260e+00 8.39176118e-01 -2.18552783e-01 -2.24112093e-01
6.14915073e-01 -2.57836074e-01 -6.74691200e-01 2.96119392e-01
8.82560983e-02 5.89557588e-01 1.04526877e+00 -1.49954784e+00
-8.63260508e-01 -3.01363289e-01 1.28943399e-01 4.54455256e-01
1.32239297e-01 3.94263789e-02 -6.01453841e-01 -9.26603258e-01
3.73443276e-01 -5.47024071e-01 -1.01581722e-01 -1.44895047e-01
2.95279402e-04 1.53726637e-01 8.22219193e-01 -9.76971745e-01
1.09263480e+00 -2.28425503e+00 3.53519320e-01 -9.52967107e-02
3.47573340e-01 2.08165288e-01 -4.49698508e-01 -2.11179942e-01
-3.22341502e-01 -9.85603407e-02 -1.47280842e-01 -2.33910203e-01
-5.55682480e-01 7.31950775e-02 -4.41063643e-01 3.52742195e-01
3.54937255e-01 7.24292874e-01 -9.35155630e-01 -5.55538058e-01
1.62349343e-01 8.63507569e-01 -3.32725793e-01 3.45887542e-01
1.60042122e-01 9.97912526e-01 -1.40692145e-01 4.76441026e-01
1.03893530e+00 -3.09828907e-01 -3.99750695e-02 -8.30800831e-01
-2.55525678e-01 1.64740235e-01 -1.40090549e+00 1.82888198e+00
-6.12169325e-01 4.15748179e-01 1.98060483e-01 -6.62125409e-01
1.18371677e+00 3.19561101e-02 4.01285321e-01 -1.17836452e+00
-3.66160125e-01 1.20811544e-01 -2.37164482e-01 -9.80377719e-02
7.00054944e-01 1.82996038e-02 3.41001987e-01 1.96346417e-01
-8.04665983e-02 -6.38198331e-02 6.45478629e-03 7.08790869e-02
7.98403919e-01 5.38675070e-01 2.90882885e-01 3.53172272e-02
9.27208245e-01 -5.17434061e-01 8.37432861e-01 4.46912378e-01
-5.77107444e-02 1.21091223e+00 -3.61248925e-02 -3.77567559e-01
-1.16757965e+00 -1.32251799e+00 -2.46486515e-01 9.49780285e-01
5.87319136e-01 -1.28563225e-01 -5.20112872e-01 -4.70638812e-01
-3.27134639e-01 2.12321326e-01 -5.43994486e-01 8.11594725e-02
-9.30209935e-01 -4.46650505e-01 5.28961569e-02 5.46718836e-01
9.23711777e-01 -7.91981637e-01 -2.85059214e-01 7.71334991e-02
-6.47037327e-01 -1.47427416e+00 -8.81491959e-01 -3.41030806e-01
-8.81116986e-01 -1.00116050e+00 -8.54799211e-01 -5.65062821e-01
5.42974949e-01 9.04540658e-01 1.19678509e+00 1.16577700e-01
-1.18604161e-01 1.95307598e-01 -3.24100971e-01 5.52409887e-01
-3.11432868e-01 -2.56742865e-01 2.65848916e-02 1.52514085e-01
-8.97894204e-02 -7.14813232e-01 -9.19183493e-01 7.34085858e-01
-1.02149498e+00 3.75910580e-01 6.58705235e-01 1.04041052e+00
1.03093791e+00 1.25921711e-01 4.29023445e-01 -5.71461618e-01
2.52665192e-01 -2.12382808e-01 -7.97439337e-01 2.96304584e-01
-6.30164027e-01 -1.41563481e-02 6.19388044e-01 -6.47723079e-01
-1.41045666e+00 -7.93816373e-02 -2.75735673e-03 -7.57877469e-01
-2.74093002e-02 6.15986548e-02 -2.94318646e-01 -2.51082122e-01
4.95406926e-01 3.51676434e-01 -1.72148496e-01 -5.86300492e-01
4.87773657e-01 5.21752298e-01 1.11938524e+00 -6.04529619e-01
1.23540699e+00 5.52628756e-01 -8.09570849e-02 -5.44930816e-01
-1.07531929e+00 -4.38357264e-01 -7.30395794e-01 2.42154617e-02
8.28120887e-01 -1.38034761e+00 -2.28234857e-01 1.14670463e-01
-9.76360142e-01 -2.01996878e-01 -2.44113028e-01 4.65126961e-01
-6.20411456e-01 5.73579788e-01 -8.19634795e-01 -4.20378149e-01
-4.23787981e-01 -1.31859410e+00 1.21594715e+00 5.15231431e-01
2.41849124e-01 -5.51374614e-01 -6.68123588e-02 6.02473855e-01
5.78847945e-01 1.63410991e-01 4.69754040e-01 -7.23863766e-02
-8.75962555e-01 3.93987805e-01 -8.47235322e-01 3.72788101e-01
3.27369630e-01 -2.94624507e-01 -8.96583855e-01 -5.16301155e-01
4.49339896e-02 -1.03699200e-01 7.79200733e-01 1.04416162e-01
1.07360518e+00 -9.86476019e-02 7.89101422e-02 1.02031898e+00
1.44872284e+00 5.53619601e-02 1.02197230e+00 5.91711998e-01
7.57535458e-01 5.23554325e-01 1.04839301e+00 6.92451745e-02
4.95737433e-01 1.21855843e+00 9.95622799e-02 -4.72165763e-01
-6.85732424e-01 -3.42216611e-01 4.32203829e-01 7.70415068e-01
-3.95810336e-01 3.98778260e-01 -2.89334059e-01 2.04815120e-01
-1.60254073e+00 -1.17136979e+00 5.26050366e-02 2.57585931e+00
1.15495908e+00 -7.09005520e-02 4.42328863e-02 -1.16589680e-01
8.77134621e-01 2.76445627e-01 -5.73748171e-01 1.26890317e-01
-4.61993337e-01 2.26931944e-01 2.64705032e-01 5.73065162e-01
-9.29875195e-01 9.32194889e-01 5.18933105e+00 9.84343231e-01
-1.21838033e+00 2.30071262e-01 6.80630982e-01 2.51138173e-02
-2.40183532e-01 -1.89427674e-01 -8.65867853e-01 5.38823962e-01
5.52022517e-01 -1.35930136e-01 7.82347023e-01 4.93388653e-01
2.70195603e-01 3.10096759e-02 -1.05552745e+00 1.21450126e+00
1.85463354e-01 -1.12949920e+00 1.52341709e-01 -7.81118795e-02
8.82341087e-01 -1.46657661e-01 2.91279972e-01 3.16518009e-01
1.41752765e-01 -1.02025187e+00 4.25785124e-01 5.60802877e-01
1.41462994e+00 -6.53225541e-01 5.35248041e-01 1.20677818e-02
-1.61281979e+00 -1.14101302e-02 -4.77721751e-01 4.37724292e-01
9.66212600e-02 3.40545744e-01 -1.53803453e-01 9.89027500e-01
1.01430511e+00 7.58710146e-01 -7.11104095e-01 5.56838691e-01
-1.26610667e-01 -2.72676740e-02 8.60618055e-02 1.21106005e+00
-2.50731707e-01 -5.08211195e-01 5.00776827e-01 9.45233405e-01
4.36661839e-01 4.35446799e-01 1.50771499e-01 1.06617594e+00
-5.69977723e-02 -1.61438480e-01 -3.69411200e-01 4.94727492e-01
5.72092772e-01 1.30234587e+00 -3.57645512e-01 -2.89630771e-01
-6.22197866e-01 1.14437079e+00 1.98696196e-01 5.52346230e-01
-8.03459644e-01 -7.27703124e-02 6.33645236e-01 2.12988079e-01
4.01248217e-01 -4.12701741e-02 1.57181453e-02 -1.49425554e+00
1.50399283e-01 -1.18665886e+00 3.31964254e-01 -1.07535541e+00
-1.20383573e+00 8.19192886e-01 -2.94351950e-02 -1.69996691e+00
-2.16114432e-01 6.31266832e-02 -2.87237555e-01 1.20148563e+00
-1.91409183e+00 -1.13540494e+00 -7.85541058e-01 7.24480629e-01
8.19260299e-01 9.17777345e-02 2.18728900e-01 4.99447584e-01
-4.09302860e-01 7.74013042e-01 -9.77036804e-02 1.80212297e-02
1.15226305e+00 -9.09279943e-01 3.45755309e-01 1.15063298e+00
-7.90128484e-03 6.57854021e-01 6.11280262e-01 -6.04921997e-01
-1.23691702e+00 -1.20012772e+00 3.54710996e-01 -3.76312584e-01
4.14051324e-01 -5.72344810e-02 -1.44026673e+00 4.61819440e-01
-1.93731442e-01 4.41472232e-01 1.31678998e-01 -3.06172580e-01
-6.76312327e-01 -4.29382622e-01 -1.24875462e+00 6.49912655e-01
1.32539082e+00 -6.73892200e-01 -5.15530646e-01 -2.15427652e-01
1.14029336e+00 -7.44004846e-01 -1.27082169e+00 7.12588847e-01
5.07968843e-01 -1.24267530e+00 1.49002159e+00 -7.45786503e-02
7.97691822e-01 -6.92776263e-01 -3.97232652e-01 -1.14768434e+00
-4.24886107e-01 -5.04918694e-01 -1.94143370e-01 1.41390276e+00
-9.27504227e-02 -4.58783954e-01 4.03101504e-01 3.89945269e-01
1.05397433e-01 -6.12068892e-01 -8.06630909e-01 -7.61972427e-01
-5.70399426e-02 1.22894511e-01 7.96112657e-01 1.05018425e+00
-5.53124011e-01 1.36731163e-01 -5.46510935e-01 4.50889617e-01
9.70275700e-01 3.89490992e-01 9.11657095e-01 -9.28331614e-01
-4.85862702e-01 -3.55215520e-01 -1.22689359e-01 -1.32476759e+00
-2.32339539e-02 -4.35916662e-01 1.29288025e-02 -1.10860479e+00
4.34201002e-01 -3.50702971e-01 -4.22168314e-01 -8.32452253e-02
-5.13113260e-01 5.41311741e-01 2.52305567e-01 4.12764132e-01
-5.81126690e-01 5.22442758e-01 1.68562210e+00 1.54209867e-01
-3.27085793e-01 -1.55873775e-01 -6.99179053e-01 4.35076624e-01
5.03678381e-01 4.93193828e-02 -3.41356575e-01 -3.56298149e-01
-1.10731132e-01 6.19807661e-01 4.13323939e-01 -7.21724033e-01
1.13005027e-01 -2.16929138e-01 5.64252019e-01 -5.81055880e-01
3.03390950e-01 -7.01588809e-01 2.80537665e-01 -9.00767222e-02
-4.23065007e-01 9.11090970e-02 -1.53349102e-01 6.27383649e-01
-4.22054410e-01 2.09931657e-01 1.14063931e+00 -9.54727530e-02
-7.04125583e-01 5.19804835e-01 5.26730359e-01 4.80499789e-02
6.37149692e-01 -3.99139404e-01 -6.64632380e-01 -2.80668497e-01
-5.28338075e-01 1.46914542e-01 8.75716925e-01 6.27320707e-01
9.61384177e-01 -1.38978446e+00 -8.49412382e-01 2.70591408e-01
5.98395057e-02 2.47897744e-01 5.75448096e-01 9.44606006e-01
-7.71578178e-02 4.99479398e-02 -3.48908722e-01 -5.31388700e-01
-1.43164074e+00 7.53735065e-01 3.35226566e-01 -3.22103173e-01
-1.05885696e+00 5.27663171e-01 6.22677445e-01 -9.47423950e-02
1.88902810e-01 -8.68272930e-02 -3.72592300e-01 -2.77084649e-01
9.87214327e-01 3.87991458e-01 -7.50974938e-02 -8.27184558e-01
-4.09595817e-02 1.01201177e+00 -2.66281813e-01 -1.16739519e-01
1.12674749e+00 -7.47954547e-01 4.40881550e-02 2.00796515e-01
1.02785468e+00 1.20220266e-01 -1.63906670e+00 -7.75940597e-01
-1.66437730e-01 -9.90414143e-01 8.55403915e-02 -6.03774369e-01
-1.35959685e+00 6.38810754e-01 7.53344655e-01 -3.12492222e-01
1.59239614e+00 -1.43549800e-01 8.26371431e-01 -2.15104982e-01
5.44663250e-01 -8.61018360e-01 2.41186962e-01 1.31028429e-01
1.06172752e+00 -1.47373664e+00 2.30127260e-01 -6.26054108e-01
-5.14432073e-01 1.14820719e+00 9.01349187e-01 3.14531513e-02
1.84051916e-02 1.16666913e-01 4.55924720e-02 2.21272901e-01
-7.03852832e-01 -3.38990182e-01 4.66211081e-01 6.80874169e-01
4.38395411e-01 -3.00277531e-01 -2.45586008e-01 3.41355175e-01
-5.21984883e-02 3.59781622e-03 6.02433801e-01 3.67742389e-01
-1.31646082e-01 -9.64936674e-01 -5.78407347e-01 5.47604598e-02
-3.52650076e-01 -2.67450929e-01 1.11746252e-01 7.27004230e-01
1.55648850e-02 9.36754644e-01 5.00930808e-02 -4.72049087e-01
3.60574692e-01 -5.88371217e-01 6.26868844e-01 -3.13747346e-01
-2.97495633e-01 3.50486130e-01 -3.26927751e-01 -9.60905790e-01
-6.26212001e-01 -4.29365695e-01 -1.03049111e+00 -2.49827728e-01
-1.86467886e-01 -1.00267358e-01 3.40527147e-01 6.26609862e-01
2.95976877e-01 5.58503747e-01 8.80444348e-01 -1.11489737e+00
-5.98184109e-01 -7.61043549e-01 -5.07227123e-01 6.80531621e-01
4.63529617e-01 -7.17122674e-01 -5.06903052e-01 1.51547849e-01] | [10.903375625610352, -2.0870606899261475] |
80021808-715c-485e-98b8-50699de71fe2 | detecting-troll-tweets-in-a-bilingual-corpus | null | null | https://aclanthology.org/2020.lrec-1.766 | https://aclanthology.org/2020.lrec-1.766.pdf | Detecting Troll Tweets in a Bilingual Corpus | During the past several years, a large amount of troll accounts has emerged with efforts to manipulate public opinion on social network sites. They are often involved in spreading misinformation, fake news, and propaganda with the intent of distracting and sowing discord. This paper aims to detect troll tweets in both English and Russian assuming that the tweets are generated by some {``}troll farm.{''} We reduce this task to the authorship verification problem of determining whether a single tweet is authored by a {``}troll farm{''} account or not. We evaluate a supervised classification approach with monolingual, cross-lingual, and bilingual training scenarios, using several machine learning algorithms, including deep learning. The best results are attained by the bilingual learning, showing the area under the ROC curve (AUC) of 0.875 and 0.828, for tweet classification in English and Russian test sets, respectively. It is noteworthy that these results are obtained using only raw text features, which do not require manual feature engineering efforts. In this paper, we introduce a resource of English and Russian troll tweets containing original tweets and translation from English to Russian, Russian to English. It is available for academic purposes. | ['Marina Litvak', 'Mark Last', 'Lin Miao'] | 2020-05-01 | null | null | null | lrec-2020-5 | ['authorship-verification'] | ['natural-language-processing'] | [-2.38227285e-03 1.05890259e-02 -3.31858307e-01 -1.13728782e-02
-4.65060353e-01 -7.35855341e-01 1.04380846e+00 4.80725020e-01
-6.56507313e-01 9.03345406e-01 1.48950979e-01 -6.11025691e-01
3.85801733e-01 -8.47486913e-01 -4.90879714e-01 -4.28845078e-01
2.67733306e-01 4.11688626e-01 -3.25172186e-01 -5.90676486e-01
5.49084246e-01 4.98418063e-01 -9.78817165e-01 2.75634736e-01
8.34120274e-01 7.54651368e-01 -5.29574454e-01 4.35440242e-01
-2.41843566e-01 1.12437963e+00 -1.10009181e+00 -8.79895210e-01
7.13902712e-02 -5.04284382e-01 -7.46275187e-01 -1.10644169e-01
2.70639867e-01 -2.51525212e-02 -3.49502116e-01 1.16083407e+00
3.04217666e-01 -5.34909725e-01 6.00826204e-01 -1.22956097e+00
-8.64094913e-01 8.75200629e-01 -6.78795278e-01 5.64625442e-01
3.26831311e-01 -4.01674621e-02 9.27654982e-01 -7.86394894e-01
7.01949716e-01 1.16660094e+00 6.86649323e-01 8.51595998e-02
-1.05274904e+00 -1.16266024e+00 -3.36079866e-01 -1.47017017e-01
-1.43865275e+00 -2.56264895e-01 5.24152517e-01 -9.46651876e-01
5.74274600e-01 1.40670687e-01 4.72079962e-01 1.54803002e+00
5.70236504e-01 5.16447783e-01 1.54644895e+00 -3.83531541e-01
-3.32908094e-01 7.30438054e-01 1.36913016e-01 6.89865589e-01
4.69544113e-01 6.37759119e-02 -7.00124741e-01 -4.31950241e-01
1.65530682e-01 -2.40421891e-01 2.89212633e-02 3.98019820e-01
-1.26729727e+00 1.37048781e+00 3.02803308e-01 6.13004446e-01
-2.01393694e-01 -2.57612824e-01 5.99019468e-01 8.05013716e-01
1.03853977e+00 6.40365601e-01 -1.97182968e-01 3.54987532e-02
-8.17680061e-01 -3.35856080e-02 8.21673810e-01 5.87091446e-01
7.14228809e-01 7.08225965e-02 2.10961640e-01 6.18097723e-01
3.23242359e-02 1.05269766e+00 7.50026047e-01 -4.48184982e-02
5.45205295e-01 5.54333210e-01 2.05486596e-01 -1.67115593e+00
-5.52253127e-01 -7.66707182e-01 -9.34547842e-01 -3.65336865e-01
4.99530643e-01 -4.75008249e-01 -3.73645902e-01 1.20312405e+00
1.13514550e-01 -2.87714601e-02 2.29781777e-01 6.16985023e-01
8.78439009e-01 6.87122762e-01 -5.04130498e-02 -2.07579479e-01
1.57051563e+00 -6.48898363e-01 -8.19365859e-01 -1.84802234e-01
8.86575162e-01 -1.14758253e+00 6.48927093e-01 1.48205340e-01
-5.14098287e-01 -1.70855448e-01 -8.36404383e-01 1.90123513e-01
-7.57273614e-01 1.99351549e-01 2.72177339e-01 9.92929757e-01
-4.83574361e-01 5.34802318e-01 -1.84983969e-01 -3.14783096e-01
4.97348756e-01 2.31595673e-02 -4.88628566e-01 4.23337847e-01
-1.65244579e+00 1.02759671e+00 8.11220035e-02 2.97763906e-02
-2.84749240e-01 -3.12584579e-01 -6.00568891e-01 -3.44939590e-01
2.34540999e-02 1.11915106e-02 8.95917475e-01 -1.29846859e+00
-1.18426609e+00 1.36103427e+00 7.58692110e-03 -6.27839386e-01
7.78486848e-01 -1.53237358e-01 -9.08178151e-01 -9.02993679e-02
5.97328782e-01 -1.03088543e-01 1.13664162e+00 -8.48089039e-01
-5.93054771e-01 -3.59576821e-01 -1.55646011e-01 -2.56551445e-01
-4.26745176e-01 5.15281141e-01 3.85411829e-01 -8.68409157e-01
1.74038988e-02 -1.01974702e+00 3.92187864e-01 -7.28967488e-01
-6.60762727e-01 -1.38114870e-01 7.20221043e-01 -8.84140253e-01
1.41998756e+00 -1.93158722e+00 -2.33667746e-01 3.89518619e-01
3.72534215e-01 4.90158409e-01 2.49215826e-01 6.66673064e-01
7.86075667e-02 6.09905720e-01 5.81887625e-02 1.94911491e-02
-1.89775288e-01 -2.08834648e-01 -5.41679204e-01 1.01082683e+00
3.98457684e-02 8.79416406e-01 -1.08853698e+00 -2.17091277e-01
-1.77152693e-01 9.80368033e-02 3.11011672e-02 -3.13079119e-01
2.61635274e-01 6.58234000e-01 -4.42850977e-01 4.21083957e-01
5.47965705e-01 -3.38632584e-01 2.34827027e-01 3.20564397e-02
-4.09101605e-01 5.07328451e-01 -6.00088775e-01 6.20296001e-01
-4.36384082e-01 1.21186697e+00 -9.13966894e-02 -5.98926544e-01
1.14038014e+00 2.49430045e-01 3.12762886e-01 -8.26063037e-01
7.26689279e-01 5.99643230e-01 -5.50470650e-02 -4.73024368e-01
6.36602104e-01 -3.30238640e-01 -5.04931748e-01 7.78306484e-01
-4.33607519e-01 2.07631662e-01 1.11045279e-01 6.45687506e-02
8.04754615e-01 -5.41573822e-01 4.09647971e-01 -1.84470192e-01
6.65560246e-01 2.51773715e-01 1.22826742e-02 7.37486959e-01
-1.23047203e-01 1.46515429e-01 6.85790122e-01 -5.32197058e-01
-1.03551006e+00 -5.78582346e-01 -3.20054829e-01 9.51001585e-01
-4.09972630e-02 -9.39065218e-02 -4.40684766e-01 -7.60609329e-01
1.73937585e-02 7.67824650e-01 -5.47948897e-01 -5.31317480e-02
-6.01159394e-01 -8.82896781e-01 1.15423441e+00 -2.07442969e-01
6.25244439e-01 -9.54151332e-01 -3.81173730e-01 1.88371867e-01
-5.92177927e-01 -1.14973223e+00 -3.44501108e-01 -1.44565433e-01
-3.60497653e-01 -1.15035999e+00 -4.99220312e-01 -5.81777692e-01
4.30505395e-01 2.09962547e-01 8.53908539e-01 9.18468535e-02
8.99283886e-02 -2.18290463e-01 -5.53880811e-01 -4.99165326e-01
-9.80362236e-01 4.21995461e-01 1.97877392e-01 3.19299221e-01
5.89609623e-01 -2.29430631e-01 -2.77976915e-02 3.30298185e-01
-8.36530089e-01 -1.42448083e-01 5.72507501e-01 7.67311335e-01
-2.50836700e-01 -1.07311629e-01 5.99753022e-01 -1.37005126e+00
8.78009796e-01 -8.80220830e-01 -5.75560033e-01 -2.61230350e-01
-5.27170599e-01 -9.55030769e-02 7.04613745e-01 -4.09496635e-01
-4.56690997e-01 -4.72033530e-01 -1.24002174e-01 1.72723327e-02
-1.25724941e-01 7.53341734e-01 4.75285083e-01 1.84235558e-01
9.92022514e-01 1.97470933e-01 1.00757033e-01 -3.17947865e-01
1.94382332e-02 1.30552471e+00 1.53901763e-02 5.79455197e-02
1.09132874e+00 5.38989186e-01 -2.16632187e-01 -9.74916399e-01
-1.10799706e+00 -5.94545245e-01 -5.38235188e-01 -1.78579390e-01
6.16872132e-01 -9.01599109e-01 -8.32289875e-01 7.36732662e-01
-1.34606135e+00 2.12663278e-01 2.30464295e-01 4.42011505e-01
1.01974249e-01 3.48578572e-01 -6.08776927e-01 -5.12567997e-01
-3.75381500e-01 -8.59099746e-01 6.93856180e-01 -2.46592939e-01
-6.05409980e-01 -1.10278535e+00 1.29381001e-01 5.65742552e-01
4.31024373e-01 5.62866211e-01 7.50385702e-01 -1.18601787e+00
1.20424435e-01 -7.57711768e-01 -4.21243906e-01 2.49732345e-01
3.85378808e-01 -1.86572731e-01 -8.25287998e-01 -3.10430348e-01
-8.77328888e-02 -2.91676909e-01 4.76087421e-01 -2.63435930e-01
3.29898566e-01 -9.07773554e-01 -2.17119053e-01 1.08548842e-01
1.04933870e+00 -1.67275473e-01 4.87606138e-01 6.06254637e-01
7.05365062e-01 6.35451078e-01 3.34074914e-01 5.05791247e-01
2.49976814e-01 5.05144715e-01 1.98051944e-01 3.11452281e-02
2.54888147e-01 -4.20735151e-01 5.02581239e-01 6.72218263e-01
1.32236436e-01 -3.93451482e-01 -1.01759350e+00 3.33842874e-01
-1.49774086e+00 -1.21450317e+00 -6.53113365e-01 1.94890583e+00
6.99530423e-01 1.51993334e-01 1.91149533e-01 1.66320026e-01
9.71281886e-01 3.87097448e-01 3.07893381e-02 -6.12202823e-01
-3.10142428e-01 3.92871685e-02 9.79648590e-01 6.22113585e-01
-1.25708377e+00 1.05741870e+00 5.14566326e+00 6.66783690e-01
-1.65573180e+00 3.84157449e-01 5.82192421e-01 3.41232538e-01
-1.14354052e-01 -2.23323196e-01 -8.22337925e-01 6.24940872e-01
1.14511335e+00 -2.69406796e-01 1.76580861e-01 4.46564794e-01
5.78791678e-01 6.97712004e-02 -4.95657086e-01 8.21034491e-01
3.07064593e-01 -1.43873549e+00 -7.54037732e-03 3.25740963e-01
8.09493303e-01 1.78811476e-01 1.97066307e-01 2.36934245e-01
2.60126382e-01 -1.22303808e+00 1.01487815e+00 2.11891443e-01
7.34340310e-01 -6.15261674e-01 1.07384908e+00 5.51882088e-01
-5.17741501e-01 7.61915073e-02 -3.35532799e-02 -1.66643724e-01
1.92385912e-02 7.26960421e-01 -1.02191103e+00 4.54983205e-01
4.08656567e-01 8.50124359e-01 -5.70080638e-01 4.28183019e-01
-3.76605481e-01 6.71185374e-01 -1.23397000e-01 -4.75039989e-01
4.90269393e-01 -6.65819272e-02 6.72157228e-01 1.38305926e+00
2.18063772e-01 -2.51270831e-01 -9.43994336e-03 5.92008531e-01
-4.21980917e-01 4.56018627e-01 -8.64098132e-01 -4.55464005e-01
3.08541358e-01 1.15022695e+00 -6.98707402e-01 -2.38978088e-01
-2.51856804e-01 9.41387832e-01 2.43582338e-01 1.68976933e-02
-1.03132975e+00 -3.71296108e-01 1.57945618e-01 5.75878620e-01
-8.61271620e-02 -2.24099115e-01 -3.14781547e-01 -1.21086979e+00
-3.01679254e-01 -1.05639601e+00 1.21800594e-01 -3.92593563e-01
-1.45973063e+00 8.26091111e-01 -2.49257401e-01 -1.21776724e+00
-1.89019874e-01 -5.65223813e-01 -2.18524978e-01 8.24851573e-01
-1.37604845e+00 -1.00801551e+00 -4.94609401e-03 2.73578286e-01
1.51586130e-01 -4.65278029e-01 5.51146507e-01 4.23981130e-01
-4.14853722e-01 6.44022703e-01 3.25844795e-01 6.94154620e-01
6.77121639e-01 -6.80086017e-01 2.20243037e-01 5.39016783e-01
1.67770252e-01 5.56783974e-01 9.43381488e-01 -7.24593699e-01
-9.72643435e-01 -1.13366759e+00 1.75174594e+00 -6.36485398e-01
1.36423111e+00 -3.32872391e-01 -4.14863855e-01 6.42332971e-01
3.08405042e-01 -4.69396770e-01 8.00601721e-01 2.89882403e-02
-5.79574108e-01 2.68652976e-01 -1.12431705e+00 3.95847976e-01
5.47451496e-01 -5.50050139e-01 -4.42303658e-01 7.04704046e-01
1.48413494e-01 -2.15458229e-01 -6.40143275e-01 -1.54754475e-01
5.41362822e-01 -8.16062450e-01 4.26641583e-01 -7.56056428e-01
6.45805657e-01 -1.02496855e-01 1.64634347e-01 -1.23540449e+00
-1.18597358e-01 -6.37042403e-01 5.37644744e-01 1.08149731e+00
7.44273841e-01 -1.02942169e+00 4.80944753e-01 -1.33015901e-01
2.47052595e-01 -1.69775322e-01 -1.05700564e+00 -5.82616389e-01
3.82350892e-01 -1.46653131e-01 3.13024282e-01 1.49471271e+00
5.95263913e-02 7.89191544e-01 -7.89245427e-01 9.98574868e-03
2.72957444e-01 1.92456059e-02 9.24003303e-01 -1.31284237e+00
1.73913136e-01 -4.18946624e-01 -2.53595799e-01 -6.82041943e-01
4.76767719e-01 -1.12037027e+00 -3.29615384e-01 -9.45556402e-01
1.95737220e-02 -5.09445310e-01 2.09491745e-01 3.79600137e-01
2.58559346e-01 6.78691328e-01 -8.76705274e-02 6.68336809e-01
-1.64372727e-01 2.12190107e-01 1.05561078e+00 -3.62279266e-01
-1.25793412e-01 2.95487285e-01 -7.37525046e-01 7.84985065e-01
9.80831385e-01 -8.69019866e-01 2.59910256e-01 -2.40813762e-01
7.74351001e-01 -2.00043693e-01 4.49147046e-01 -5.28185964e-01
-1.47449464e-01 1.96554903e-02 6.07181378e-02 -2.63451010e-01
-3.64281312e-02 -5.99847078e-01 4.26396094e-02 8.00425351e-01
-3.02414715e-01 4.53887999e-01 4.71046707e-03 5.33585906e-01
-3.52798820e-01 -1.70286059e-01 9.97858226e-01 -2.33238548e-01
-1.68311521e-01 -4.65539247e-02 -9.19575453e-01 1.62398756e-01
8.50344300e-01 3.48680578e-02 -6.50552809e-01 -5.92465401e-01
-5.24183273e-01 -2.38189951e-01 2.99131483e-01 5.91656268e-01
1.58985436e-01 -9.86134231e-01 -1.05097771e+00 1.27136663e-01
2.21248388e-01 -9.96047139e-01 -2.50680864e-01 1.16269803e+00
-8.51100922e-01 6.35579526e-01 -5.92389964e-02 -3.06980699e-01
-1.18647218e+00 3.58237088e-01 1.78269878e-01 -4.07050878e-01
-2.48964474e-01 4.42202508e-01 -4.40219194e-01 -6.24057949e-01
-3.73822659e-01 1.80035114e-01 -3.78692627e-01 5.64521253e-01
4.22813952e-01 5.28338671e-01 1.60317615e-01 -1.58643794e+00
-3.45743388e-01 2.22351909e-01 -1.54110819e-01 -1.90852564e-02
9.09538090e-01 -1.59795016e-01 -4.27811503e-01 6.06723428e-01
1.39329135e+00 8.24413240e-01 6.65168017e-02 -2.85106868e-01
-4.06067558e-02 -4.79914367e-01 -5.90021610e-02 -6.39078319e-01
-7.55659401e-01 6.02106273e-01 3.13058197e-01 7.15979278e-01
3.38154525e-01 1.32716969e-02 8.68602693e-01 1.93442523e-01
3.53837103e-01 -8.86112392e-01 -3.07042953e-02 8.82735431e-01
6.89915001e-01 -1.22678792e+00 -1.90335978e-02 -2.73465812e-01
-6.91725850e-01 1.10190713e+00 5.72898760e-02 -1.86879206e-02
7.08271325e-01 -1.30214453e-01 3.55456024e-01 -1.85705051e-01
-1.49056181e-01 -4.95592318e-02 -4.52017561e-02 1.28107488e-01
5.79110086e-01 1.78509355e-01 -8.12477887e-01 1.79369941e-01
-6.47657335e-01 -2.89768606e-01 8.15716028e-01 6.78789735e-01
-4.75625962e-01 -9.31969404e-01 -6.14333868e-01 5.38522899e-01
-9.51571405e-01 -3.22800159e-01 -8.40261281e-01 8.43705535e-01
1.83193535e-01 1.24824703e+00 -6.43782876e-03 -5.33902884e-01
2.85591073e-02 -1.17057696e-01 -1.33965284e-01 -5.34604847e-01
-1.06619263e+00 -2.14104205e-01 4.41232681e-01 1.31002143e-01
-4.45264816e-01 -6.32348478e-01 -9.19288158e-01 -1.07389235e+00
-4.50874388e-01 4.61035371e-01 5.47495365e-01 1.16614103e+00
2.11020976e-01 -6.26893044e-02 8.97477150e-01 -2.61759847e-01
-4.13512170e-01 -1.18885076e+00 -5.10560155e-01 5.17752767e-01
3.24234366e-01 -4.55159962e-01 -6.35752857e-01 -1.80821911e-01] | [8.262774467468262, 10.289885520935059] |
d3bb71af-edb6-459b-8a00-ac44c6b103aa | openie6-iterative-grid-labeling-and | 2010.03147 | null | https://arxiv.org/abs/2010.03147v1 | https://arxiv.org/pdf/2010.03147v1.pdf | OpenIE6: Iterative Grid Labeling and Coordination Analysis for Open Information Extraction | A recent state-of-the-art neural open information extraction (OpenIE) system generates extractions iteratively, requiring repeated encoding of partial outputs. This comes at a significant computational cost. On the other hand, sequence labeling approaches for OpenIE are much faster, but worse in extraction quality. In this paper, we bridge this trade-off by presenting an iterative labeling-based system that establishes a new state of the art for OpenIE, while extracting 10x faster. This is achieved through a novel Iterative Grid Labeling (IGL) architecture, which treats OpenIE as a 2-D grid labeling task. We improve its performance further by applying coverage (soft) constraints on the grid at training time. Moreover, on observing that the best OpenIE systems falter at handling coordination structures, our OpenIE system also incorporates a new coordination analyzer built with the same IGL architecture. This IGL based coordination analyzer helps our OpenIE system handle complicated coordination structures, while also establishing a new state of the art on the task of coordination analysis, with a 12.3 pts improvement in F1 over previous analyzers. Our OpenIE system, OpenIE6, beats the previous systems by as much as 4 pts in F1, while being much faster. | ['Soumen Chakrabarti', 'Mausam', 'Samarth Aggarwal', 'Vaibhav Adlakha', 'Keshav Kolluru'] | 2020-10-07 | null | https://aclanthology.org/2020.emnlp-main.306 | https://aclanthology.org/2020.emnlp-main.306.pdf | emnlp-2020-11 | ['open-information-extraction'] | ['natural-language-processing'] | [-1.45204976e-01 6.56420708e-01 -2.52059102e-01 -2.88886391e-02
-1.13325179e+00 -8.74028027e-01 3.34534526e-01 3.74926776e-01
-3.64981323e-01 8.71128023e-01 1.05913341e-01 -5.07885635e-01
7.41258785e-02 -8.61694515e-01 -6.97611988e-01 -2.78603673e-01
-2.28873625e-01 8.39562774e-01 2.48168945e-01 -2.88233936e-01
5.88797927e-02 2.33788133e-01 -1.27372062e+00 5.63383698e-01
7.38028705e-01 8.04271519e-01 -6.96338713e-02 7.58731425e-01
-5.23334086e-01 9.41006660e-01 -7.38802671e-01 -4.72765237e-01
2.95162112e-01 -2.09447458e-01 -1.38655818e+00 -3.79379570e-01
3.49974185e-01 9.36405808e-02 1.04804210e-01 9.16482151e-01
5.20780623e-01 -3.36334884e-01 5.93060732e-01 -1.15836656e+00
-4.20359701e-01 1.36259258e+00 -4.44562644e-01 -1.40850678e-01
5.26322186e-01 -1.46974236e-01 1.57464552e+00 -7.55248308e-01
9.09740150e-01 7.75309265e-01 9.43742990e-01 4.48703974e-01
-1.06758618e+00 -6.69299185e-01 2.92664707e-01 2.89066285e-01
-1.77423835e+00 -3.68257701e-01 4.91656214e-01 -1.80541128e-01
1.73377645e+00 6.14589900e-02 4.50897038e-01 6.30967140e-01
-5.96369803e-02 9.95571494e-01 7.83168554e-01 -7.66240895e-01
1.91138223e-01 -3.53080034e-01 4.64705467e-01 8.61747622e-01
4.01792884e-01 -2.92135030e-01 -4.75185305e-01 9.65320617e-02
6.97189212e-01 -6.41577661e-01 -1.04513533e-01 -1.11432768e-01
-1.25725865e+00 6.08762205e-01 2.75014341e-01 5.39745808e-01
-1.30574659e-01 8.14283921e-05 8.67860079e-01 6.19136214e-01
3.30501616e-01 6.84420407e-01 -7.53236294e-01 -4.41527098e-01
-9.86933053e-01 3.15644026e-01 1.46517766e+00 1.32599175e+00
7.57449508e-01 -1.74856171e-01 -1.19555146e-01 6.30400598e-01
6.69202805e-02 1.08444959e-01 1.45473540e-01 -7.96781421e-01
8.93319547e-01 7.59468138e-01 -1.70337722e-01 -5.17795086e-01
-8.97029400e-01 -7.17532277e-01 -7.58577466e-01 1.96120873e-01
7.61059940e-01 -3.06971312e-01 -6.38228714e-01 1.67293882e+00
-3.25375190e-03 -3.52641672e-01 1.95070237e-01 4.54824328e-01
6.25728548e-01 5.63116372e-01 -2.23644197e-01 -3.46150964e-01
1.42470324e+00 -1.27039993e+00 -8.28106523e-01 9.06000808e-02
1.28148186e+00 -5.70147991e-01 9.14919257e-01 9.37171459e-01
-1.20363319e+00 -3.81373554e-01 -1.40791738e+00 -1.53901413e-01
-5.57514131e-01 2.49634802e-01 7.87659764e-01 7.40192056e-01
-1.27975237e+00 6.01027250e-01 -8.28018188e-01 -7.77091309e-02
2.88773239e-01 5.17136693e-01 -4.82826293e-01 2.38588676e-01
-1.10280871e+00 9.02315676e-01 7.13081539e-01 -2.23083764e-01
-2.64698774e-01 -6.85926795e-01 -1.10060608e+00 3.48005831e-01
1.02306139e+00 -2.78523654e-01 1.55175054e+00 -4.99477208e-01
-1.44681847e+00 6.57381356e-01 -1.12954311e-01 -7.04404473e-01
2.23507166e-01 -2.39748314e-01 -3.58536601e-01 6.24654517e-02
2.62371331e-01 6.09990299e-01 1.62833348e-01 -9.01649117e-01
-4.98154938e-01 -1.29826784e-01 1.29563391e-01 8.24385807e-02
-1.37645662e-01 8.16273764e-02 -6.44396722e-01 -7.06465602e-01
-1.57878667e-01 -8.42592835e-01 -1.86824769e-01 -3.28694791e-01
-5.24384439e-01 -7.77432084e-01 4.55900252e-01 -6.39831066e-01
1.83574283e+00 -1.81098950e+00 1.52494922e-01 4.30855602e-01
6.19609952e-01 4.20621067e-01 -1.65323898e-01 6.16483390e-01
-1.35870323e-01 3.97568762e-01 -3.30135077e-01 -5.43854713e-01
3.68663251e-01 2.25209206e-01 -2.65056062e-02 1.18415885e-01
2.51014471e-01 9.39334035e-01 -8.00395310e-01 -6.15409911e-01
-2.94828653e-01 -1.00241132e-01 -8.58423173e-01 -8.58832970e-02
-6.78501844e-01 2.76647024e-02 5.94669208e-02 6.67657197e-01
4.74186450e-01 -3.17854047e-01 4.19049472e-01 4.21319790e-02
-4.70627695e-01 5.94884336e-01 -1.31865036e+00 2.26933122e+00
-5.42906046e-01 5.76983988e-01 -1.70362350e-02 -9.97573197e-01
9.17577565e-01 4.92416918e-01 4.98546034e-01 -5.16351759e-01
2.81193703e-01 4.14771974e-01 2.19576523e-01 -1.17906392e-01
6.08787656e-01 4.83461112e-01 -4.63722408e-01 9.46680725e-01
4.17879492e-01 1.03399739e-01 7.65361845e-01 3.48600805e-01
1.34176993e+00 2.68173337e-01 5.83055258e-01 -3.11798185e-01
4.00830954e-01 1.22809388e-01 6.52490437e-01 8.05513620e-01
3.60459387e-02 5.80148876e-01 7.76924491e-01 -6.32689774e-01
-1.22993147e+00 -7.73482680e-01 1.16436575e-02 8.09165359e-01
-3.43866110e-01 -1.26586330e+00 -1.32377732e+00 -9.67395484e-01
-4.46752280e-01 3.48266453e-01 -4.92274255e-01 3.20764929e-01
-9.37764585e-01 -4.90718931e-01 1.38182950e+00 8.45353842e-01
5.90319633e-01 -1.12078011e+00 -5.41427970e-01 4.30803448e-01
-4.53864545e-01 -1.21624637e+00 -2.57146329e-01 6.91316724e-01
-3.04319352e-01 -1.04792452e+00 -4.04024512e-01 -8.12174618e-01
3.83313835e-01 -4.12244469e-01 1.47829652e+00 2.64553338e-01
-1.34682208e-01 -3.50862533e-01 -7.87372887e-01 -4.62244630e-01
-5.31371355e-01 1.08087790e+00 9.33431238e-02 -6.04167283e-01
5.64591527e-01 -5.10739386e-01 2.49423265e-01 1.48087099e-01
-7.80559301e-01 2.45612115e-01 5.82839310e-01 9.57390070e-01
5.27895451e-01 2.07611352e-01 6.68958426e-01 -9.49579656e-01
5.04297018e-01 -1.60735205e-01 -6.27399266e-01 3.01650822e-01
-6.59846842e-01 5.39272845e-01 9.61577237e-01 -7.17603136e-03
-8.83188963e-01 6.68094158e-02 -7.78622985e-01 -1.21067017e-01
-2.97995329e-01 7.26058066e-01 -3.87254447e-01 8.22871178e-02
8.56899440e-01 -1.99405342e-01 -1.28298566e-01 -6.55531824e-01
4.90450472e-01 7.46661782e-01 7.26991892e-01 -7.07908571e-01
5.94340146e-01 1.05271153e-01 -1.14417829e-01 -5.33806801e-01
-9.63827789e-01 -3.49144816e-01 -1.01860225e+00 3.71076167e-01
7.11682498e-01 -7.86303282e-01 -7.62873411e-01 5.53802252e-01
-1.25694621e+00 -5.72911441e-01 -3.47170949e-01 -5.31787463e-02
-5.82508206e-01 5.17600000e-01 -8.22772324e-01 -3.26257229e-01
-5.27700722e-01 -9.42623854e-01 8.82943809e-01 -1.54433608e-01
-6.25895917e-01 -7.79408872e-01 1.40082315e-01 7.96380863e-02
1.45484865e-01 8.25284421e-02 9.70314503e-01 -9.91427600e-01
-5.20941138e-01 1.51483983e-01 -3.00053746e-01 2.10362405e-01
-2.32965186e-01 -2.11456180e-01 -7.22741067e-01 -7.70755857e-02
-4.06348377e-01 -2.93765515e-01 6.01552486e-01 7.60812387e-02
7.97948420e-01 -1.95669696e-01 -3.02605897e-01 6.56171024e-01
1.34257710e+00 1.48600787e-01 6.62785351e-01 6.02532089e-01
6.81156456e-01 4.21505541e-01 5.93978763e-01 4.18832511e-01
5.02625465e-01 9.14583266e-01 1.76629886e-01 -2.85705566e-01
-2.92354822e-01 -2.19458073e-01 2.72808462e-01 8.45926940e-01
-1.21051379e-01 -9.97514725e-02 -1.27198327e+00 5.96180737e-01
-2.04238153e+00 -8.04984152e-01 -1.30310863e-01 1.69397497e+00
1.22590470e+00 4.58182991e-01 2.90971011e-01 8.44766974e-01
2.12316170e-01 -1.82051286e-01 -2.66729206e-01 -7.57149041e-01
-7.94486627e-02 6.59720123e-01 5.53042054e-01 3.68067533e-01
-1.03627717e+00 1.23203337e+00 6.40718603e+00 1.11910141e+00
-8.91233921e-01 5.28140850e-02 3.15977633e-01 7.47026429e-02
-9.67407774e-04 1.60027910e-02 -1.27387035e+00 1.21772811e-01
9.54120934e-01 1.77416161e-01 4.78567719e-01 7.54910588e-01
-3.38707477e-01 -1.09454446e-01 -1.33433235e+00 1.05397749e+00
9.83914435e-02 -1.61849475e+00 -1.13181435e-01 6.56754971e-02
4.63158011e-01 1.84268117e-01 -4.88445550e-01 4.77544248e-01
5.83395302e-01 -1.01491952e+00 9.85261440e-01 2.29296342e-01
1.07348335e+00 -9.99841392e-01 9.71762002e-01 4.66406345e-01
-1.47209346e+00 7.32205517e-04 -3.14204544e-02 -5.24288297e-01
3.13285768e-01 4.80877817e-01 -8.16772521e-01 7.90231645e-01
6.06150925e-01 5.81359148e-01 -6.32922888e-01 7.35310137e-01
-6.01045370e-01 5.95432341e-01 -4.93872344e-01 -4.44929376e-02
3.06731015e-01 9.07333717e-02 3.72611701e-01 1.58022606e+00
-1.25426158e-01 1.69445407e-02 3.99551600e-01 8.78527582e-01
-1.14052586e-01 9.71512273e-02 -6.16061211e-01 9.66669768e-02
5.31142592e-01 1.03896523e+00 -8.35883915e-01 -5.60566247e-01
-3.62987131e-01 7.99658835e-01 7.25863218e-01 -2.11440235e-01
-7.92079270e-01 -8.97711813e-01 2.31382683e-01 -2.76604205e-01
4.39116776e-01 -3.89894813e-01 -7.04100847e-01 -1.20074248e+00
1.63937137e-01 -1.06059384e+00 4.64902759e-01 -5.05038261e-01
-7.95000553e-01 9.64288533e-01 5.10790981e-02 -1.00520813e+00
-6.62340105e-01 -5.21367967e-01 -1.37579039e-01 6.59012496e-01
-1.26638687e+00 -1.16065443e+00 1.08466774e-01 4.01576728e-01
6.44970894e-01 -1.10202566e-01 1.18044055e+00 5.11507750e-01
-5.65384150e-01 1.05818260e+00 -3.15697551e-01 6.35195136e-01
4.53104407e-01 -1.62010014e+00 9.87564802e-01 9.81257915e-01
5.24593949e-01 6.88035905e-01 2.65191138e-01 -5.28448343e-01
-1.15197086e+00 -9.95111525e-01 1.35035038e+00 -4.39986944e-01
6.06837690e-01 -6.46939874e-01 -6.85686946e-01 7.19058037e-01
3.79284859e-01 -3.36218417e-01 6.77642584e-01 6.02319658e-01
-6.65953279e-01 3.20631087e-01 -8.92373502e-01 5.58971226e-01
1.39875972e+00 -5.05402267e-01 -6.85378134e-01 6.62892535e-02
9.98788178e-01 -6.80251539e-01 -1.07350111e+00 3.47638220e-01
4.55226094e-01 -9.13558185e-01 6.27526164e-01 -1.66357026e-01
2.80318499e-01 -5.39502800e-01 2.24438845e-03 -1.36804473e+00
-3.14149439e-01 -8.76708150e-01 -9.75812599e-02 1.10837257e+00
9.24956799e-01 -4.34370279e-01 8.21898520e-01 2.01066118e-02
-6.84434533e-01 -9.38917994e-01 -7.48474002e-01 -9.35436845e-01
9.03318971e-02 -7.17140019e-01 7.43857324e-01 8.97191405e-01
8.68008733e-01 6.62180841e-01 -1.77174792e-01 -9.14607197e-02
3.38434070e-01 3.56866531e-02 8.78642499e-01 -1.23074746e+00
-4.83671099e-01 -4.13361043e-01 -1.48595154e-01 -1.11948550e+00
2.87619174e-01 -1.15454471e+00 -1.45182803e-01 -1.45837593e+00
-5.87944947e-02 -7.64439046e-01 7.73048252e-02 1.17238319e+00
3.23341370e-01 5.12723923e-01 3.28673542e-01 1.20599896e-01
-8.67417634e-01 -6.38798624e-02 7.32743442e-01 -3.23918015e-02
-2.83240139e-01 -4.12348300e-01 -9.31725562e-01 7.75596201e-01
8.04426789e-01 -3.75085056e-01 -5.73550723e-02 -5.83871007e-01
6.91344321e-01 -9.09184292e-02 -2.84815758e-01 -1.33773875e+00
4.86080915e-01 5.39273858e-01 -3.66251059e-02 -9.54575062e-01
9.44174174e-03 -4.51971829e-01 -3.57678309e-02 3.00372899e-01
-5.88851087e-02 2.44378492e-01 3.81599903e-01 -5.34272455e-02
-3.63079637e-01 -2.84252852e-01 4.25008893e-01 -2.01205581e-01
-5.12043655e-01 1.28647342e-01 -5.19372880e-01 1.91707253e-01
6.66222870e-01 -2.63705879e-01 -3.04382205e-01 -6.69764131e-02
-9.18192267e-01 2.61141032e-01 3.96982044e-01 1.20109633e-01
5.14579304e-02 -8.56622279e-01 -6.05390787e-01 4.34143037e-01
8.94768238e-02 5.28739154e-01 -2.47199446e-01 7.42497623e-01
-7.19218791e-01 7.57546782e-01 -2.14772195e-01 -4.40321565e-01
-1.22275949e+00 4.37838942e-01 3.08933631e-02 -1.24030042e+00
-7.27089167e-01 7.92580128e-01 -3.51920843e-01 -5.72325230e-01
4.84550059e-01 -6.65665984e-01 -2.79035121e-01 7.62381032e-02
6.80849195e-01 3.27761829e-01 6.67276740e-01 -3.28464538e-01
-1.28811747e-01 4.45316106e-01 -4.39460874e-01 -1.31673515e-01
1.21918380e+00 2.78390497e-01 -2.59549588e-01 8.91889781e-02
8.54255557e-01 1.82929948e-01 -7.72887588e-01 -3.00521821e-01
3.19661230e-01 2.08404705e-01 -2.90420949e-01 -1.05600798e+00
-7.28405118e-01 6.59658253e-01 -1.61434039e-02 2.80919760e-01
1.07425928e+00 4.26218770e-02 9.99520898e-01 6.96123898e-01
6.97016895e-01 -1.18048954e+00 -2.31935084e-01 1.19747710e+00
5.73091388e-01 -7.79561937e-01 -1.58731118e-01 -7.15295374e-01
-4.17833716e-01 1.09960938e+00 4.66864377e-01 -8.48318860e-02
4.23102081e-01 9.75604713e-01 6.17019571e-02 -3.53447556e-01
-8.31098318e-01 -2.70890623e-01 -2.43377596e-01 5.47595322e-01
4.22152668e-01 1.65009230e-01 -5.04702628e-01 9.85169768e-01
-6.81635559e-01 -1.03233524e-01 6.08131945e-01 9.44677174e-01
-2.49408871e-01 -1.60880601e+00 -2.94015706e-01 2.79312909e-01
-6.90880895e-01 -5.03100455e-01 -4.74231690e-01 8.80877614e-01
3.21478248e-01 7.84363151e-01 1.43254101e-01 -4.42332923e-01
2.34174654e-01 2.33573109e-01 5.04459262e-01 -8.51681054e-01
-9.06173527e-01 -6.16409294e-02 5.88711500e-01 -6.84619963e-01
-2.51820982e-01 -6.95260823e-01 -1.69592702e+00 -1.41679928e-01
-6.58391654e-01 4.82657969e-01 4.91102815e-01 9.82235551e-01
3.68755370e-01 5.38341641e-01 2.13323727e-01 -6.94294631e-01
-3.87257963e-01 -1.09452546e+00 -4.95085001e-01 -5.79269268e-02
5.36131710e-02 -5.77713668e-01 1.06395349e-01 -5.07345721e-02] | [9.64733600616455, 8.703424453735352] |
fa484b2f-d184-4d7a-8b86-a68dcf5bd29c | biogpt-generative-pre-trained-transformer-for | 2210.10341 | null | https://arxiv.org/abs/2210.10341v3 | https://arxiv.org/pdf/2210.10341v3.pdf | BioGPT: Generative Pre-trained Transformer for Biomedical Text Generation and Mining | Pre-trained language models have attracted increasing attention in the biomedical domain, inspired by their great success in the general natural language domain. Among the two main branches of pre-trained language models in the general language domain, i.e., BERT (and its variants) and GPT (and its variants), the first one has been extensively studied in the biomedical domain, such as BioBERT and PubMedBERT. While they have achieved great success on a variety of discriminative downstream biomedical tasks, the lack of generation ability constrains their application scope. In this paper, we propose BioGPT, a domain-specific generative Transformer language model pre-trained on large scale biomedical literature. We evaluate BioGPT on six biomedical NLP tasks and demonstrate that our model outperforms previous models on most tasks. Especially, we get 44.98%, 38.42% and 40.76% F1 score on BC5CDR, KD-DTI and DDI end-to-end relation extraction tasks respectively, and 78.2% accuracy on PubMedQA, creating a new record. Our case study on text generation further demonstrates the advantage of BioGPT on biomedical literature to generate fluent descriptions for biomedical terms. Code is available at https://github.com/microsoft/BioGPT. | ['Tie-Yan Liu', 'Hoifung Poon', 'Sheng Zhang', 'Tao Qin', 'Yingce Xia', 'Liai Sun', 'Renqian Luo'] | 2022-10-19 | null | null | null | null | ['document-classification'] | ['natural-language-processing'] | [ 1.70292467e-01 4.99413967e-01 -3.07434052e-01 -2.76244640e-01
-1.10514069e+00 -3.57828081e-01 5.89548707e-01 2.18386248e-01
-3.13555121e-01 1.33231258e+00 3.04050684e-01 -4.50660706e-01
-7.90047497e-02 -8.20769906e-01 -5.88384211e-01 -5.09316206e-01
2.32687026e-01 9.27380204e-01 -1.10574707e-01 -2.55985975e-01
-1.98371544e-01 -1.61849894e-02 -6.86928570e-01 3.82248044e-01
1.17723978e+00 6.56661749e-01 2.52844274e-01 3.37692022e-01
-1.88395694e-01 6.00960970e-01 -6.94242895e-01 -9.33909118e-01
-3.24302286e-01 -6.49471879e-01 -9.95307982e-01 -5.22037983e-01
-2.22184196e-01 1.38108671e-01 -4.35143560e-01 8.44913244e-01
8.28824341e-01 -3.04447889e-01 9.23090518e-01 -1.05266237e+00
-9.93582129e-01 1.08986747e+00 -4.84044254e-01 2.06227601e-01
3.63298386e-01 -3.05559933e-02 9.26702261e-01 -7.67733455e-01
8.96876335e-01 1.25196385e+00 6.88904762e-01 8.63535404e-01
-1.20490241e+00 -9.25542772e-01 -3.40662509e-01 2.62756646e-02
-1.54336989e+00 -3.26693565e-01 2.96549737e-01 -3.74072134e-01
1.19901824e+00 1.05746575e-01 2.02496842e-01 1.49544752e+00
6.26728833e-01 7.57267654e-01 1.03343785e+00 -3.36788923e-01
-1.14249021e-01 4.35647406e-02 -1.28480449e-01 6.22708380e-01
3.49742264e-01 -1.13071360e-01 -4.99441862e-01 -2.67005235e-01
4.34209824e-01 -3.04963976e-01 -3.85193974e-01 5.39157391e-01
-1.25573254e+00 8.36704969e-01 3.62118334e-01 5.29896259e-01
-4.18667436e-01 -3.06423847e-02 4.57859069e-01 -3.75556201e-02
7.73812771e-01 6.97316825e-01 -6.07027888e-01 -8.12439471e-02
-8.13842297e-01 2.64267206e-01 7.83912838e-01 1.26602840e+00
1.63099244e-01 -2.03278571e-01 -5.66813469e-01 1.13617027e+00
1.81905761e-01 4.98363525e-01 8.09581757e-01 -3.20528477e-01
5.58628857e-01 4.61444587e-01 -5.00917614e-01 -7.46166587e-01
-5.02650023e-01 -7.59989619e-01 -1.29075909e+00 -8.72675717e-01
1.80870786e-01 -3.04262757e-01 -1.06173420e+00 1.79504848e+00
8.62952136e-03 8.69668871e-02 4.51061040e-01 4.90837604e-01
1.45947444e+00 8.03319037e-01 5.30270100e-01 -1.70714349e-01
1.62574303e+00 -6.55713975e-01 -9.29109991e-01 -2.86219537e-01
6.86116278e-01 -8.12113166e-01 5.46698630e-01 3.51929486e-01
-1.05481327e+00 -3.62344831e-01 -6.49309576e-01 -3.55018556e-01
-5.16508043e-01 3.23850304e-01 7.32776165e-01 2.10267469e-01
-1.03424823e+00 4.75616932e-01 -7.54586816e-01 -4.67260122e-01
6.24008119e-01 2.40717202e-01 -3.92289758e-01 -2.96454102e-01
-1.71573961e+00 1.13434505e+00 6.46713376e-01 -1.02363445e-03
-8.09082150e-01 -9.89717543e-01 -6.80072188e-01 -3.60730402e-02
1.97246686e-01 -1.25603414e+00 1.21786022e+00 8.13412145e-02
-1.26463616e+00 1.13013613e+00 -2.30406687e-01 -7.00558662e-01
4.29602355e-01 -1.99399650e-01 -4.23229396e-01 -4.14778367e-02
4.35459793e-01 8.60327244e-01 1.52143627e-01 -6.96199536e-01
-4.63684678e-01 -2.46138662e-01 -2.80607760e-01 -4.66564521e-02
-2.20157772e-01 -1.05122454e-01 -4.51870382e-01 -8.41412663e-01
-3.42230141e-01 -8.60992312e-01 -2.89973885e-01 -5.10109723e-01
-8.43176484e-01 -6.97643280e-01 2.19972566e-01 -8.59497666e-01
1.31794691e+00 -1.71313250e+00 1.88128754e-01 -2.22407326e-01
1.84360623e-01 3.57380390e-01 -2.76791211e-02 6.96163356e-01
-1.85255691e-01 4.85316813e-01 -4.16943014e-01 -2.28660241e-01
-1.43228859e-01 2.47608498e-01 -2.82559514e-01 2.40408946e-02
4.86542046e-01 1.32276857e+00 -1.15201652e+00 -6.62174881e-01
-3.12568843e-01 6.66051567e-01 -3.85925055e-01 9.59567577e-02
-1.95098072e-01 5.28009355e-01 -7.49729693e-01 5.61718225e-01
4.06406671e-01 -3.83319557e-01 1.16168156e-01 -2.06821129e-01
2.48333126e-01 6.64457917e-01 -1.69430062e-01 1.90733182e+00
-5.82906067e-01 2.45716244e-01 -4.84525293e-01 -8.89579833e-01
8.65889788e-01 7.80279875e-01 5.15855014e-01 -4.97168452e-01
9.41488519e-02 4.74473149e-01 2.39220440e-01 -4.98304427e-01
1.44336864e-01 -6.02608979e-01 -1.72432169e-01 8.43977705e-02
4.47806746e-01 -3.37183893e-01 4.29672927e-01 2.83396900e-01
1.24266863e+00 2.24024639e-01 6.60722256e-01 -2.52197385e-01
4.64908689e-01 1.58810124e-01 5.30625403e-01 5.41781783e-01
2.82287717e-01 6.98473573e-01 3.96444678e-01 6.92781284e-02
-5.66395998e-01 -8.04817200e-01 -5.94339490e-01 5.92888653e-01
-2.65156627e-01 -6.30134404e-01 -6.03706181e-01 -7.61161745e-01
-2.28138193e-02 9.41625893e-01 -5.79247832e-01 -4.18265820e-01
-5.55637002e-01 -1.36448002e+00 1.07435226e+00 2.98549831e-01
5.38959086e-01 -1.07023728e+00 -3.37384120e-02 4.59844142e-01
-4.87042993e-01 -1.24582469e+00 -3.42834592e-01 1.28765017e-01
-8.94716322e-01 -8.25934410e-01 -1.14834201e+00 -7.34539330e-01
4.81455475e-01 -4.19739336e-01 1.39007223e+00 -3.69232297e-01
-4.44219559e-01 -1.22180290e-03 -3.32899958e-01 -6.73776627e-01
-7.88212359e-01 4.88688529e-01 -2.73155808e-01 -4.29797709e-01
3.81380618e-01 -4.42236811e-01 -5.35902739e-01 -5.63131794e-02
-8.09817076e-01 4.17299569e-01 1.02235866e+00 1.07926011e+00
6.84977710e-01 -2.16744930e-01 1.10227966e+00 -1.24991298e+00
9.05545533e-01 -8.60156178e-01 -8.27580318e-02 3.02864283e-01
-7.98334539e-01 1.37794688e-01 4.58355337e-01 -2.82742620e-01
-1.09495091e+00 -3.30768049e-01 -6.42558336e-01 -5.69408573e-02
-6.24823906e-02 1.07260597e+00 -1.36084050e-01 6.11494780e-01
6.69854462e-01 3.69134098e-01 -2.10489765e-01 -5.05654573e-01
3.44924927e-01 6.80497944e-01 4.49725270e-01 -5.21951020e-01
4.09900278e-01 -5.31028733e-02 6.73673227e-02 -5.38860559e-01
-9.21938062e-01 -3.00291985e-01 -2.60937005e-01 3.95460933e-01
1.05945933e+00 -9.33318973e-01 -3.33712012e-01 2.28929356e-01
-1.39035237e+00 -1.42507166e-01 -7.16455430e-02 4.41528618e-01
-1.44502386e-01 1.30960852e-01 -8.79014432e-01 -4.71779853e-01
-9.73822713e-01 -1.08101976e+00 1.25808024e+00 4.82690632e-02
-5.91094077e-01 -1.15605927e+00 7.20146969e-02 3.51830214e-01
2.97866315e-01 3.40045661e-01 1.24698102e+00 -9.11031723e-01
-5.95925339e-02 -1.60511583e-01 -2.04004914e-01 1.62947550e-01
4.02159929e-01 -3.40867132e-01 -8.67047608e-01 -2.92469263e-02
-2.03046694e-01 -2.20534876e-01 9.30306971e-01 3.43564212e-01
1.09310234e+00 -3.41217160e-01 -9.49469328e-01 4.58580136e-01
1.15875149e+00 4.95214999e-01 6.03174508e-01 -4.03883904e-02
5.93743026e-01 4.70467269e-01 5.06627679e-01 6.49176389e-02
4.12090391e-01 6.91385269e-01 -3.43583561e-02 -2.33934045e-01
-4.37626094e-01 -3.76920700e-01 1.23171709e-01 9.88442302e-01
-8.21028948e-02 -6.57477260e-01 -1.20921624e+00 7.04129815e-01
-1.59843290e+00 -6.00509226e-01 -1.65906519e-01 1.68829346e+00
1.66216600e+00 -8.74268543e-03 -1.63628772e-01 -2.10355088e-01
5.05501986e-01 -3.46727997e-01 -4.90744114e-01 -2.30612099e-01
-2.36787617e-01 7.66527355e-01 1.90019667e-01 2.32210100e-01
-9.37092602e-01 9.92966950e-01 5.19135666e+00 1.18133616e+00
-8.93831551e-01 2.95427889e-01 1.07683480e+00 1.83917090e-01
-1.76834106e-01 -2.46275559e-01 -1.05114150e+00 5.32149136e-01
1.09675515e+00 -6.97731555e-01 -1.95259109e-01 5.76348722e-01
-2.77220216e-02 2.39854082e-01 -1.26470029e+00 9.31569099e-01
-6.41215593e-02 -1.32324958e+00 4.95272875e-01 1.14660047e-01
5.34904718e-01 5.06072715e-02 -1.74813773e-02 6.34356856e-01
3.36369336e-01 -1.42025709e+00 3.35957468e-01 4.01677161e-01
1.17184258e+00 -5.53100884e-01 1.07081318e+00 3.14558178e-01
-6.61950469e-01 3.78315687e-01 -3.72799248e-01 4.57458198e-01
3.41344863e-01 1.08132541e+00 -1.38644314e+00 1.02069867e+00
3.74865472e-01 9.05212164e-01 -4.47727472e-01 8.45496178e-01
-4.60633129e-01 9.19012904e-01 -1.07357115e-01 -1.23854615e-01
1.03524439e-01 1.51072051e-02 3.48155499e-01 1.65457141e+00
5.97321212e-01 3.36668670e-01 -5.00343740e-02 1.04925513e+00
-4.64976490e-01 1.77891493e-01 -5.45724154e-01 -2.84310490e-01
2.77304739e-01 1.30642021e+00 -4.10223991e-01 -4.06701446e-01
-7.74409920e-02 7.99698830e-01 1.58177227e-01 2.57121742e-01
-9.59878922e-01 -5.62457800e-01 3.15683752e-01 -1.13181183e-02
-1.96873337e-01 3.19809526e-01 -2.05935687e-01 -1.00625515e+00
-2.40177408e-01 -9.40260947e-01 5.30420065e-01 -7.47204065e-01
-1.61140776e+00 9.94614005e-01 1.93152592e-01 -9.78554845e-01
-5.36964774e-01 -5.48241496e-01 -9.19874683e-02 1.16266668e+00
-1.43387806e+00 -1.41157758e+00 1.11892834e-01 2.22253412e-01
4.67890024e-01 -1.75444826e-01 1.13675475e+00 6.20633721e-01
-5.27544737e-01 7.72756755e-01 -1.17720373e-01 2.56663144e-01
9.14320409e-01 -1.19763231e+00 5.22850990e-01 3.89886737e-01
-4.15054560e-02 1.03596926e+00 5.67666650e-01 -8.82816434e-01
-1.19348073e+00 -1.44894683e+00 1.45670927e+00 -5.39026916e-01
4.77460146e-01 -3.97421420e-01 -9.44914103e-01 6.08733356e-01
3.54265809e-01 -4.26090270e-01 7.57224560e-01 -5.63450493e-02
-5.63569590e-02 3.46477665e-02 -1.08059537e+00 6.45578504e-01
1.25115967e+00 -2.27415204e-01 -5.83218455e-01 7.70610094e-01
7.60939837e-01 -5.66512465e-01 -1.30827177e+00 6.76102459e-01
3.21083516e-01 -2.80522466e-01 1.01684988e+00 -6.19686246e-01
9.04769719e-01 2.76000444e-02 2.61660010e-01 -1.50846386e+00
-3.92089456e-01 -6.25198722e-01 8.13038498e-02 1.43840325e+00
9.77780938e-01 -8.06245387e-01 3.26633841e-01 2.75225699e-01
-3.29088211e-01 -1.25374150e+00 -1.04063690e+00 -6.54801428e-01
5.42155743e-01 -1.75764710e-01 4.51384306e-01 9.19154465e-01
1.49221599e-01 9.71160531e-01 -1.48272783e-01 -3.14901352e-01
1.57779604e-01 5.40648177e-02 3.62773985e-01 -1.07623446e+00
-4.08790827e-01 -4.34765309e-01 -1.31195173e-01 -9.32766557e-01
1.49088502e-01 -1.51648295e+00 5.48633235e-03 -1.93437660e+00
4.94926691e-01 -5.41226804e-01 -1.91497490e-01 9.09421563e-01
-4.76442248e-01 2.63128072e-01 -2.73562968e-01 8.46135840e-02
-6.41810372e-02 4.38074321e-01 1.34933305e+00 -3.13129097e-01
7.93632120e-02 -1.03379600e-01 -9.81713295e-01 3.06607336e-01
9.50992405e-01 -6.52402461e-01 -4.56870854e-01 -4.63773221e-01
1.98531866e-01 1.51438817e-01 7.21809715e-02 -6.26304805e-01
-5.45589700e-02 1.48057595e-01 2.79095650e-01 -4.47471708e-01
2.51402259e-01 -1.20394051e-01 3.00792545e-01 6.64552867e-01
-4.94394392e-01 1.13666505e-01 4.10973519e-01 3.80221844e-01
-3.13552797e-01 -1.34256169e-01 5.19758880e-01 -2.72660583e-01
-1.86420083e-01 3.84720743e-01 -2.33425975e-01 3.34317952e-01
7.06128716e-01 3.21962357e-01 -5.38821876e-01 -2.70782888e-01
-6.45107567e-01 1.98293209e-01 -2.80637175e-01 5.16235828e-01
6.28977478e-01 -1.04568326e+00 -1.30515540e+00 -2.19161183e-01
2.21389800e-01 1.66701317e-01 8.80988166e-02 1.02527118e+00
-2.63323039e-01 1.02410769e+00 2.23101050e-01 -4.74045396e-01
-1.07408726e+00 3.92161995e-01 2.31369704e-01 -8.53213251e-01
-4.97607470e-01 1.07901227e+00 5.84546328e-01 -2.44246975e-01
-1.44990534e-01 -6.05910718e-01 -2.89951116e-01 -7.61376023e-02
3.36255044e-01 7.09893778e-02 3.16575587e-01 -3.78920138e-01
-5.05761683e-01 2.92867333e-01 -2.89097250e-01 -1.71877500e-02
1.44447851e+00 2.88236469e-01 -3.66585165e-01 2.17805386e-01
1.20137155e+00 -9.45241675e-02 -2.43068501e-01 -9.85506997e-02
2.54759997e-01 2.75389969e-01 -1.83724225e-01 -1.36907387e+00
-9.15650189e-01 9.46649075e-01 1.17194273e-01 -8.85295346e-02
1.09032404e+00 2.77286202e-01 1.05242372e+00 1.32954270e-01
4.77236867e-01 -4.00345147e-01 -3.52050334e-01 5.58537781e-01
1.17107308e+00 -9.63625789e-01 1.04762383e-01 -6.28976226e-01
-5.74833691e-01 8.35949779e-01 3.67288083e-01 2.14855924e-01
4.97170866e-01 3.51055235e-01 -1.14574909e-01 -2.72639841e-01
-8.84254277e-01 -2.56096441e-02 5.17929554e-01 4.95800406e-01
1.06171525e+00 3.00076641e-02 -8.17428708e-01 9.96442080e-01
-5.13435304e-01 3.75836015e-01 7.93021843e-02 5.37318647e-01
2.94292361e-01 -1.31983614e+00 1.37361288e-01 6.72385097e-01
-1.03531778e+00 -6.31521225e-01 -4.92730379e-01 8.09239686e-01
1.94006130e-01 9.54814374e-01 -4.45418864e-01 -2.19202146e-01
3.55032474e-01 3.22817594e-01 4.57986087e-01 -9.90217507e-01
-7.01957822e-01 1.20849922e-01 3.99160504e-01 -4.97475602e-02
-2.98053652e-01 -4.10227895e-01 -1.36751842e+00 2.33858339e-02
-2.06197813e-01 5.35805166e-01 3.62804264e-01 7.58168817e-01
6.60430133e-01 9.82973039e-01 -1.56420335e-01 -1.70800790e-01
-2.99487203e-01 -1.49278557e+00 -1.83348373e-01 2.19110876e-01
-4.43387441e-02 -4.75512713e-01 1.21810429e-01 2.63703585e-01] | [8.58874797821045, 8.726927757263184] |
16eba130-2504-4c61-abf5-cedbd5d1d953 | unsupervised-monocular-depth-and-ego-motion | 1906.05717 | null | https://arxiv.org/abs/1906.05717v1 | https://arxiv.org/pdf/1906.05717v1.pdf | Unsupervised Monocular Depth and Ego-motion Learning with Structure and Semantics | We present an approach which takes advantage of both structure and semantics for unsupervised monocular learning of depth and ego-motion. More specifically, we model the motion of individual objects and learn their 3D motion vector jointly with depth and ego-motion. We obtain more accurate results, especially for challenging dynamic scenes not addressed by previous approaches. This is an extended version of Casser et al. [AAAI'19]. Code and models have been open sourced at https://sites.google.com/corp/view/struct2depth. | ['Anelia Angelova', 'Reza Mahjourian', 'Vincent Casser', 'Soeren Pirk'] | 2019-06-12 | null | null | null | null | ['depth-and-camera-motion'] | ['computer-vision'] | [-5.00810444e-01 -6.97566345e-02 -3.81744117e-01 -6.69613004e-01
-3.01643372e-01 -7.29494631e-01 6.90119803e-01 -2.68923491e-01
-2.31782123e-01 6.95323765e-01 5.40272057e-01 8.18661228e-02
3.42872471e-01 -5.37997365e-01 -6.57463968e-01 -3.79879028e-01
-1.16659924e-01 4.49611217e-01 5.35823762e-01 1.59212202e-01
4.48989213e-01 6.93735898e-01 -1.65185678e+00 4.03920233e-01
2.01515540e-01 4.68335390e-01 5.13469040e-01 1.10488939e+00
2.07547192e-02 1.16546226e+00 1.67191222e-01 5.96777834e-02
2.83270031e-01 -1.99301451e-01 -1.14684951e+00 3.29926580e-01
7.33675957e-01 -8.31327140e-01 -8.59579980e-01 8.95106494e-01
3.19006920e-01 6.49302900e-02 5.04165769e-01 -1.23253393e+00
-2.53862739e-01 1.42299622e-01 -4.97191966e-01 5.33215404e-01
6.67598367e-01 2.38688201e-01 7.97005713e-01 -9.81834590e-01
1.15190268e+00 1.29858339e+00 1.69414505e-01 7.77350307e-01
-9.13612843e-01 -3.20527703e-01 2.36474887e-01 5.35188258e-01
-1.16921461e+00 -8.52209270e-01 8.84921193e-01 -6.44808471e-01
1.09150982e+00 -1.10899583e-01 8.78278732e-01 1.09395111e+00
2.58948177e-01 1.36242878e+00 8.44655395e-01 -2.89180994e-01
2.25471124e-01 -7.75056705e-02 8.72094631e-02 8.00737500e-01
1.92942962e-01 1.70525715e-01 -5.91165066e-01 2.13525385e-01
1.15319097e+00 5.57537489e-02 -9.85537991e-02 -1.13642001e+00
-1.31760180e+00 6.62304878e-01 3.95558506e-01 1.89379498e-01
-3.29589427e-01 5.93709648e-01 -4.31431644e-02 5.88610582e-02
6.33266628e-01 9.40025747e-02 -5.27170777e-01 -5.05216420e-01
-5.87509036e-01 6.24421000e-01 8.50012600e-01 1.16564524e+00
1.12554932e+00 -1.58809498e-01 5.82148015e-01 2.95751393e-01
5.02833068e-01 2.75226444e-01 4.37637657e-01 -1.83676732e+00
2.43067116e-01 3.98200423e-01 3.31212848e-01 -6.19936943e-01
-4.69383180e-01 9.44946259e-02 -1.46624878e-01 2.61449397e-01
4.14114475e-01 -1.65220752e-01 -8.55019987e-01 1.47534359e+00
5.73044896e-01 4.56034958e-01 1.57281488e-01 8.25741589e-01
9.83696640e-01 3.32348645e-01 -2.33560801e-01 1.09898984e-01
9.30159330e-01 -1.31906211e+00 -6.90279424e-01 -4.73134249e-01
1.04330003e+00 -6.45908833e-01 6.43025815e-01 2.46024340e-01
-1.34728682e+00 -4.37360048e-01 -8.55301619e-01 -4.99495178e-01
-2.87255824e-01 -7.31806308e-02 1.04876757e+00 1.67506889e-01
-1.35386837e+00 4.56795961e-01 -1.26342750e+00 -8.03017735e-01
5.34408569e-01 1.55129194e-01 -5.54284751e-01 -3.11317086e-01
-6.80379391e-01 8.25466335e-01 3.39120924e-01 -2.50777900e-01
-1.13902879e+00 -2.49525458e-01 -1.08728528e+00 -6.30912304e-01
2.65472054e-01 -1.06217945e+00 1.79272246e+00 -6.17774010e-01
-1.43414795e+00 1.18371880e+00 -5.95255911e-01 -3.66692185e-01
4.92173672e-01 -6.97316289e-01 1.51139930e-01 2.80141354e-01
1.26774028e-01 1.10678911e+00 3.33221495e-01 -1.22058189e+00
-6.86518848e-01 -7.46098220e-01 5.90086579e-01 8.00533354e-01
9.43436027e-02 -6.98679686e-02 -6.80220068e-01 -1.49820656e-01
4.80179816e-01 -9.16727781e-01 -3.24197739e-01 2.79387325e-01
-7.40164891e-02 9.94294696e-03 8.37326765e-01 -3.35135758e-01
7.27660596e-01 -1.97928393e+00 3.87771189e-01 -4.76673216e-01
3.06949705e-01 -1.57670826e-01 -1.66766252e-02 4.23146904e-01
1.18853681e-01 -2.67375618e-01 5.07961214e-02 -6.87314987e-01
-1.96488604e-01 4.64896679e-01 7.73084238e-02 7.89382577e-01
-2.78806746e-01 1.06419420e+00 -1.06576526e+00 -4.68188047e-01
5.56364655e-01 3.68388355e-01 -6.50591016e-01 1.82137340e-01
-2.98365355e-01 7.11146891e-01 -5.62587917e-01 6.95218503e-01
4.77945507e-01 -3.42892885e-01 2.04194020e-02 1.01989888e-01
-2.99512774e-01 4.45911378e-01 -1.46006846e+00 2.21661353e+00
-3.82366553e-02 7.33987570e-01 1.25819445e-01 -6.23728871e-01
3.40900660e-01 2.28553414e-01 5.60100675e-01 -3.70859742e-01
1.09421484e-01 4.47257161e-02 -3.49829376e-01 -6.74666822e-01
5.35801768e-01 1.10783568e-02 1.72863141e-01 3.78857583e-01
9.84726101e-03 -6.43467844e-01 1.95004955e-01 4.86333668e-01
9.39321220e-01 9.52615142e-01 5.58878601e-01 -1.82971284e-01
3.05570364e-01 3.08453262e-01 3.26297045e-01 5.89509666e-01
-4.47017699e-01 7.20210791e-01 1.28255218e-01 -5.04645467e-01
-1.08331645e+00 -1.10569596e+00 -6.51266351e-02 7.16767251e-01
5.20916104e-01 -4.14388418e-01 -4.79998350e-01 -3.69872242e-01
3.59239280e-02 5.34427702e-01 -5.31494558e-01 2.19665006e-01
-5.77491939e-01 -9.75909084e-02 2.39917248e-01 8.52202177e-01
3.57445866e-01 -1.01059484e+00 -1.02792406e+00 -5.46733327e-02
-3.95400524e-01 -1.27231276e+00 -2.99137622e-01 -5.91338612e-03
-1.27871823e+00 -9.45352316e-01 -4.82961953e-01 -6.02821410e-01
3.34502012e-01 6.33461297e-01 1.12738109e+00 -2.22826228e-01
-4.89844263e-01 1.04322457e+00 -3.71623635e-01 -3.41319025e-01
-4.07334045e-02 -2.17039771e-02 1.49024546e-01 -3.78240794e-01
5.97632229e-01 -6.66769028e-01 -7.97437012e-01 3.68413590e-02
-6.19207144e-01 3.44511360e-01 7.50624835e-02 1.90236762e-01
6.68675661e-01 -2.71364659e-01 -1.44135937e-01 -8.38075161e-01
-3.34783256e-01 -6.11770332e-01 -7.08957553e-01 -4.54252660e-01
-2.12461963e-01 -1.57766491e-01 -1.91828102e-01 1.28547591e-03
-1.09613883e+00 5.04025400e-01 -5.33525161e-02 -5.47170997e-01
-7.26428926e-01 2.26339493e-02 -1.83892876e-01 4.39417064e-02
4.25485730e-01 6.89906627e-02 -8.60909596e-02 -6.27674699e-01
5.85132539e-01 2.93266952e-01 5.32782555e-01 -3.61853600e-01
6.48984790e-01 1.22904420e+00 -1.91779375e-01 -8.39650095e-01
-8.96952391e-01 -1.07940888e+00 -1.35763335e+00 -2.07931310e-01
1.04471302e+00 -1.38913453e+00 -3.45215619e-01 7.28816926e-01
-1.23564839e+00 -6.44171953e-01 -2.84578443e-01 9.41127956e-01
-1.07673478e+00 5.80240011e-01 -6.88819170e-01 -6.60005152e-01
2.04190105e-01 -7.79985547e-01 1.19074786e+00 3.12774509e-01
-2.40140960e-01 -1.31618917e+00 3.64200741e-01 3.50511223e-01
-2.14441586e-02 2.87800789e-01 1.30009428e-01 -2.16501117e-01
-1.06224477e+00 5.60622439e-02 -1.00266770e-01 1.04128964e-01
9.85352322e-02 -7.28898421e-02 -1.08921504e+00 -2.61696279e-01
2.10781395e-02 -4.21124339e-01 8.41792405e-01 7.55512476e-01
7.13165641e-01 -8.34642947e-02 -4.92810667e-01 8.39065373e-01
1.65415657e+00 1.26019031e-01 5.04982054e-01 5.76075852e-01
1.07131934e+00 9.41710532e-01 7.47003198e-01 6.35498941e-01
9.17290866e-01 6.29319072e-01 7.56774724e-01 2.77283221e-01
-1.84260249e-01 -2.59800434e-01 3.08504850e-01 4.91002411e-01
-2.01156229e-01 -1.19955465e-01 -1.04770517e+00 9.71370101e-01
-1.95660567e+00 -1.23007524e+00 -3.80274504e-01 1.68181264e+00
4.82212305e-01 -7.00496044e-03 1.73835009e-01 -2.94284374e-01
1.28581569e-01 6.26107693e-01 -8.00037801e-01 -2.03200355e-01
-1.94193140e-01 -1.55344501e-01 5.11291683e-01 9.44935501e-01
-1.08626544e+00 1.29832280e+00 6.46036816e+00 -6.42411932e-02
-7.54942358e-01 1.51142284e-01 1.39149353e-01 -4.07615840e-01
-4.09665138e-01 4.73333776e-01 -1.01140964e+00 -8.88582096e-02
7.45234370e-01 -1.13868147e-01 2.58519322e-01 9.71743226e-01
3.00708920e-01 -5.48715591e-01 -1.29155874e+00 9.76418555e-01
2.18440935e-01 -1.33855593e+00 -2.77011901e-01 4.17244166e-01
8.78716111e-01 5.85770309e-01 -2.14721918e-01 -6.87558055e-02
5.83943784e-01 -6.73840225e-01 8.60418677e-01 5.87836504e-01
3.06866467e-01 -3.08463454e-01 4.18807000e-01 5.96153855e-01
-1.22062159e+00 -3.95974778e-02 -4.09208685e-01 -5.86964726e-01
3.20486248e-01 4.15967286e-01 -5.80904007e-01 5.36936104e-01
9.94741023e-01 1.38487470e+00 -4.90577847e-01 9.04731035e-01
-2.95443237e-01 1.11508071e-01 -3.47300649e-01 3.53685707e-01
2.03744203e-01 -1.82851866e-01 6.76728845e-01 1.16723657e+00
5.60152084e-02 3.32459718e-01 3.06916296e-01 6.07027590e-01
1.64015368e-01 -1.39697552e-01 -1.07266903e+00 8.95669460e-02
3.99915785e-01 1.05793846e+00 -5.28443694e-01 -5.46779513e-01
-4.80023623e-01 1.11641634e+00 3.19001228e-01 4.06946301e-01
-5.40547073e-01 9.59305912e-02 8.98905993e-01 1.95668146e-01
5.21027625e-01 -7.64308453e-01 -2.11034268e-01 -1.68769908e+00
-7.34146237e-02 -3.35069209e-01 2.82280535e-01 -1.15758276e+00
-7.61781931e-01 2.06082702e-01 3.17961484e-01 -1.19332671e+00
-7.26133525e-01 -6.22249961e-01 -3.48517329e-01 5.63613296e-01
-1.48323154e+00 -1.10591614e+00 -5.47790110e-01 7.47784555e-01
8.81411612e-01 2.91166931e-01 5.16283035e-01 -6.08907491e-02
-8.95200148e-02 -2.26226881e-01 -3.77126713e-03 -1.25482678e-01
5.46895444e-01 -1.38906193e+00 7.74089515e-01 8.80812585e-01
3.16293627e-01 2.37679765e-01 7.79658914e-01 -6.18125200e-01
-1.49319959e+00 -7.47326493e-01 9.99573350e-01 -1.21172071e+00
6.37285888e-01 -4.05210376e-01 -6.41035974e-01 1.26754010e+00
2.11352423e-01 8.81033838e-02 2.63804525e-01 -2.27655292e-01
-1.47704571e-01 2.92105943e-01 -8.92262816e-01 6.10947490e-01
1.51819336e+00 -6.14655018e-01 -5.63729823e-01 3.23076397e-01
5.88769913e-01 -8.44251215e-01 -6.01052701e-01 3.78781170e-01
7.28173614e-01 -1.31038845e+00 1.05634046e+00 -4.08627063e-01
4.92366582e-01 -3.17678988e-01 -5.64909756e-01 -8.02197456e-01
-2.81926572e-01 -1.75754711e-01 -5.50031185e-01 7.63512015e-01
-1.02740817e-01 -4.05501902e-01 1.11356568e+00 2.59313613e-01
3.99849890e-03 -4.91786391e-01 -6.99783325e-01 -6.15555942e-01
-5.70268407e-02 -5.28296053e-01 2.92746499e-02 9.12392378e-01
-1.98009029e-01 1.44787848e-01 -3.02599907e-01 2.94249773e-01
8.76910210e-01 2.59975493e-01 1.14005649e+00 -1.07891011e+00
-3.10983509e-01 -1.73030853e-01 -6.89943016e-01 -1.57709360e+00
1.65405378e-01 -7.06911504e-01 -1.01301412e-03 -1.88430977e+00
1.17898285e-01 1.49874106e-01 1.26409397e-01 3.89726222e-01
2.21428528e-01 3.46188247e-01 1.71125010e-01 5.26071727e-01
-7.32691407e-01 4.63314712e-01 1.29435337e+00 3.07023793e-01
-1.06556639e-01 -1.16960950e-01 -3.33715588e-01 1.12721336e+00
1.04055882e+00 -3.33006263e-01 -6.63449824e-01 -8.06217670e-01
-1.91313416e-01 1.86549500e-01 6.22271657e-01 -1.07165861e+00
3.87603015e-01 -3.94582659e-01 4.45175260e-01 -1.00540721e+00
8.30520511e-01 -5.43847084e-01 4.90717180e-02 5.24577200e-01
-5.12810051e-02 3.00381213e-01 2.06912532e-01 6.61407888e-01
-1.63873717e-01 -1.00046553e-01 6.84489310e-01 -7.65634000e-01
-1.41114068e+00 4.33013439e-01 -3.30853611e-01 -1.20733388e-01
1.05756688e+00 -4.87548381e-01 -3.26735407e-01 -7.01338649e-01
-9.73990798e-01 2.43952185e-01 1.00219667e+00 5.23150027e-01
7.21345603e-01 -1.09229398e+00 -6.40497863e-01 1.20687842e-01
1.79576606e-01 2.01872811e-01 4.58584279e-01 8.39708984e-01
-8.55772316e-01 5.70278049e-01 -3.07059020e-01 -8.78977418e-01
-1.27158082e+00 5.01870275e-01 3.49864274e-01 2.06655458e-01
-8.48362863e-01 9.12567496e-01 6.17014050e-01 -4.55490023e-01
1.11428447e-01 -1.01432748e-01 -2.06521049e-01 -1.94912434e-01
5.85123241e-01 5.02134264e-01 -4.65684593e-01 -8.65571439e-01
-4.90923017e-01 8.83971453e-01 -5.76951355e-02 -3.60550523e-01
1.33667684e+00 -7.86226988e-01 -1.23413004e-01 9.50939059e-01
1.23753905e+00 -1.21842632e-02 -1.65104795e+00 -4.20118660e-01
3.97905782e-02 -7.08026171e-01 -1.55722862e-02 -2.56015569e-01
-7.40585864e-01 9.77444768e-01 5.29080331e-01 -3.18335682e-01
8.60274732e-01 4.32469189e-01 4.24469233e-01 3.22608203e-01
5.86031914e-01 -8.40454519e-01 1.80999041e-01 7.82689273e-01
4.97568518e-01 -1.23075175e+00 1.44998655e-01 -3.60649884e-01
-5.76311946e-01 8.18709493e-01 8.83892953e-01 -2.80140072e-01
8.19594681e-01 3.15962136e-01 1.89411789e-01 -3.77447754e-01
-1.09904039e+00 -4.33335930e-01 -6.17640913e-02 6.01488352e-01
4.31212306e-01 -9.64187160e-02 4.97385971e-02 -3.09028625e-01
-5.25050983e-02 2.66648054e-01 8.23600590e-01 1.32959998e+00
-5.63005924e-01 -8.78041625e-01 -1.12696670e-01 -1.33566827e-01
-1.48118794e-01 2.99742334e-02 -3.86212915e-01 8.63318801e-01
-5.78655105e-04 6.48171604e-01 1.52253553e-01 -2.47910872e-01
2.38308087e-01 1.36809334e-01 8.58226001e-01 -1.01136553e+00
1.49792835e-01 1.22754797e-01 1.52364984e-01 -1.04923987e+00
-9.58875418e-01 -1.08852851e+00 -1.32397377e+00 -3.93903881e-01
2.07823351e-01 -3.09480190e-01 7.04519570e-01 7.65977979e-01
3.36558789e-01 1.06843881e-01 4.38213110e-01 -1.29737377e+00
-1.15508117e-01 -6.92768455e-01 -7.04332650e-01 3.38638037e-01
5.22459626e-01 -6.26611888e-01 -5.28449476e-01 4.37430054e-01] | [8.456494331359863, -2.014842987060547] |
ba19eb0e-fa2a-4ee3-8e47-d76ac06ac744 | blockchain-platform-for-covid-19-vaccine | 2101.00983 | null | https://arxiv.org/abs/2101.00983v1 | https://arxiv.org/pdf/2101.00983v1.pdf | Blockchain platform for COVID-19 vaccine supply management | In the context of the COVID-19 pandemic, the rapid roll-out of a vaccine and the implementation of a worldwide immunization campaign is critical, but its success will depend on the availability of an operational and transparent distribution chain that can be audited by all relevant stakeholders. In this paper, we discuss how blockchain technology can be used for assuring the transparent tracing of COVID-19 vaccine registration, storage and delivery, and side effects self-reporting. We present such system implementation in which blockchain technology is used for assuring data integrity and immutability in case of beneficiary registration for vaccination, eliminating identity thefts and impersonations. Smart contracts are defined to monitor and track the proper vaccine distribution conditions against the safe handling rules defined by vaccine producers enabling the awareness of all network peers. For vaccine administration, a transparent and tamper-proof side effects self-reporting solution is provided considering person identification and administrated vaccine association. A prototype was implemented using the Ethereum test network, Ropsten, considering the COVID-19 vaccine distribution tracking conditions. The results obtained for each on-chain operation can be checked and validated on the Etherscan, demonstrating various aspects of the proposed system such as immunization actors and safe rules registration, vaccine tracking, and administration. In terms of throughput and scalability, the proposed blockchain system shows promising results. | ['Ionut Anghel', 'Marcel Antal', 'Tudor Cioara', 'Claudia Daniela Antal'] | 2021-01-04 | null | null | null | null | ['person-identification'] | ['computer-vision'] | [ 1.03669629e-01 1.26704305e-01 -3.84257823e-01 -2.65041031e-02
1.46782428e-01 -1.40658534e+00 7.42615163e-01 7.27756023e-01
-4.25437957e-01 1.00529921e+00 -5.25947241e-03 -7.82976031e-01
-1.54652640e-01 -9.27703261e-01 -6.95794582e-01 -6.03929698e-01
-3.21479738e-01 7.88172483e-01 2.33019099e-01 -1.58543870e-01
-2.12977286e-02 4.63910013e-01 -9.62678671e-01 7.85212293e-02
7.25463867e-01 9.04544294e-01 -4.46303695e-01 4.95084345e-01
8.52367654e-02 6.89737082e-01 -6.98472381e-01 -3.22697669e-01
3.47319216e-01 5.04546659e-03 -5.65822482e-01 -7.59778500e-01
-3.00277054e-01 -9.99921083e-01 5.01436114e-01 1.01325011e+00
1.56129286e-01 -9.52328205e-01 2.93166935e-01 -1.75965190e+00
-3.56355101e-01 8.32895756e-01 -3.41078341e-01 -3.57744426e-01
4.97201025e-01 4.96169388e-01 4.93644923e-01 6.21085167e-02
1.30737352e+00 5.78060210e-01 6.42174304e-01 2.92007238e-01
-9.84748304e-01 -9.70641673e-01 -4.84071940e-01 2.64645666e-01
-9.23591018e-01 -2.27213781e-02 2.10199952e-01 -6.21324599e-01
7.39257574e-01 4.86958236e-01 8.32253993e-01 8.41146886e-01
9.71029580e-01 -6.77129701e-02 1.66520929e+00 -3.40606481e-01
4.17041093e-01 1.59565046e-01 4.62250441e-01 1.89687639e-01
1.35348129e+00 3.87657374e-01 -7.11721107e-02 -7.41659582e-01
3.81683618e-01 1.84691116e-01 -3.11717421e-01 -3.11844319e-01
-1.21885431e+00 7.37937152e-01 -2.18730107e-01 6.48586452e-01
-9.18746531e-01 2.36317664e-01 7.06555426e-01 5.90983808e-01
6.36406392e-02 -2.29128301e-01 -5.35611689e-01 -1.63788889e-02
-7.10346818e-01 2.55236506e-01 1.08675146e+00 6.55947268e-01
2.67630547e-01 -2.64741331e-01 -1.09868065e-01 -9.18673098e-01
8.66752505e-01 1.27258193e+00 -3.30059469e-01 -7.47742653e-01
6.10001013e-02 6.13937020e-01 6.73421741e-01 -9.11431730e-01
-3.74505550e-01 -2.14877352e-01 -6.33928895e-01 3.51012588e-01
4.71349806e-01 -3.35194588e-01 -5.70114076e-01 1.70727801e+00
7.83787549e-01 -1.76345110e-01 1.36442572e-01 5.72813928e-01
3.42403740e-01 4.90821242e-01 2.21681103e-01 -6.54384553e-01
1.81804967e+00 8.35891441e-02 -1.39759171e+00 9.34942067e-01
5.99702656e-01 -7.47047722e-01 -2.79191345e-01 4.23141181e-01
-9.89571393e-01 2.68660516e-01 -9.25766289e-01 6.05276227e-01
-4.78745282e-01 -8.75948727e-01 3.74815166e-01 1.12994540e+00
-9.15126204e-01 3.46736647e-02 -8.14964354e-01 -2.55271554e-01
4.69434857e-02 9.12290514e-01 -5.11381388e-01 3.10775060e-02
-1.61905086e+00 8.99907231e-01 4.13553178e-01 1.40557587e-01
-1.20784640e+00 -4.88674313e-01 -4.79262263e-01 -2.49356136e-01
-2.12319031e-01 -7.16607749e-01 7.16667831e-01 -7.11721659e-01
-1.03689408e+00 6.86983526e-01 3.77291203e-01 -8.90594065e-01
6.67788863e-01 1.55719265e-01 -7.14182436e-01 3.49694759e-01
7.05687776e-02 -1.01833791e-01 4.12061036e-01 -8.62582028e-01
-6.43954217e-01 -4.91377443e-01 5.60215637e-02 -7.75207520e-01
2.64591455e-01 5.07474482e-01 5.75613260e-01 -1.98304802e-01
-6.87252223e-01 -9.72333789e-01 1.99974611e-01 -6.07878029e-01
-6.34357613e-03 4.32238951e-02 1.17560995e+00 -1.04118049e+00
9.62456703e-01 -1.65430355e+00 -7.59637296e-01 8.53816867e-01
9.24152695e-03 7.33103037e-01 1.83151290e-01 9.21077430e-01
1.14660397e-01 2.71487962e-02 5.19273430e-02 6.45094633e-01
1.88609004e-01 5.77604398e-02 -3.23714077e-01 9.93586302e-01
-2.42055446e-01 8.48398745e-01 -9.44520473e-01 -1.85424328e-01
1.99389204e-01 7.13133812e-01 -3.21956187e-01 8.65292326e-02
-4.49190289e-01 9.22230721e-01 -4.20219630e-01 5.90134919e-01
1.21904206e+00 -9.59556997e-02 1.02149892e+00 3.38442400e-02
-5.21979392e-01 6.59033880e-02 -1.01324880e+00 1.00818026e+00
-5.98269589e-02 7.24343807e-02 6.78539693e-01 -1.01337880e-01
6.86038017e-01 1.08668077e+00 6.38931334e-01 -7.07369030e-01
5.47918737e-01 3.57290357e-01 2.99980547e-02 -4.07627225e-01
2.02107906e-01 3.28352869e-01 -3.92494649e-02 1.30352700e+00
-3.48328739e-01 7.31199920e-01 -2.76559163e-02 1.39932886e-01
9.33890522e-01 3.69614482e-01 3.42273295e-01 -4.29302514e-01
6.50465429e-01 2.48987705e-01 5.32249749e-01 3.06718290e-01
-3.30235362e-01 -7.41198540e-01 3.02536994e-01 -5.69349110e-01
-1.01717174e+00 -5.94551086e-01 1.97291359e-01 4.00277883e-01
-6.79698726e-03 -1.56509191e-01 -1.05017602e+00 -7.66521633e-01
1.05699115e-01 4.79480743e-01 -3.49407077e-01 4.63684320e-01
-8.64121974e-01 -2.23916605e-01 6.19710684e-01 -1.02024846e-01
4.47904289e-01 -9.53971267e-01 -1.25715351e+00 5.41615963e-01
-2.58022726e-01 -9.66562510e-01 -1.75778702e-01 -3.54994506e-01
-4.84275341e-01 -1.73765719e+00 -1.69449598e-01 -3.68790776e-01
7.63859212e-01 -1.29366249e-01 5.09981036e-01 4.72227454e-01
1.37621641e-01 9.19799730e-02 -2.02312216e-01 -6.27861440e-01
-1.12768042e+00 -3.02013487e-01 9.21186432e-02 -2.42033955e-02
3.38047236e-01 -1.82396069e-01 -1.16801739e+00 3.99316698e-01
-1.07207048e+00 -3.68971169e-01 6.06822632e-02 1.60548627e-01
1.99421242e-01 -3.84540975e-01 8.46351087e-01 -1.00802052e+00
4.98482883e-01 -6.42007709e-01 -1.45968461e+00 5.77921569e-01
-1.11664450e+00 -2.65837878e-01 6.08651102e-01 4.78134025e-03
-8.91321778e-01 -3.87982577e-01 -1.58762231e-01 4.56415743e-01
-1.25236422e-01 2.28711158e-01 -1.55392876e-02 -2.31421262e-01
-5.63072674e-02 -1.17329806e-01 4.68487650e-01 -1.03381984e-01
5.29762864e-01 7.43262053e-01 3.66411991e-02 -2.67584383e-01
6.88340545e-01 8.94397259e-01 1.88791275e-01 -3.09064351e-02
3.52763385e-01 -1.15316950e-01 2.40495875e-01 -4.68796730e-01
7.81503677e-01 -6.38980925e-01 -1.87495828e+00 9.08706367e-01
-1.67741776e+00 1.66980430e-01 -7.74322152e-02 6.44690156e-01
-1.11850008e-01 4.95902538e-01 -6.47592783e-01 -7.76268005e-01
-1.13048816e+00 -1.09468877e+00 9.34659913e-02 -8.83143619e-02
-4.96899039e-01 -8.91171515e-01 6.51629090e-01 4.06839490e-01
9.59594607e-01 7.15500236e-01 6.47209644e-01 -1.02519774e+00
-9.28110242e-01 -4.82476741e-01 -8.63788128e-02 2.86725070e-02
2.85266280e-01 1.94089130e-01 -5.82718611e-01 -8.05686355e-01
3.09557259e-01 1.82138905e-02 -2.85025120e-01 1.62148759e-01
-1.98485076e-01 -1.03149974e+00 -2.13583797e-01 1.03332266e-01
1.54569364e+00 5.60498834e-01 6.44719601e-01 2.03957051e-01
1.41167203e-02 8.56053114e-01 7.49279737e-01 4.99112934e-01
6.52252436e-01 4.53124791e-01 7.52460480e-01 1.23966858e-01
4.33487371e-02 -2.19778851e-01 4.97252464e-01 7.05346346e-01
-1.33905128e-01 5.27465576e-03 -7.52464175e-01 4.79908943e-01
-1.48233712e+00 -9.95501935e-01 -6.49198055e-01 2.24081922e+00
7.33594179e-01 -2.19456404e-01 2.76830882e-01 -2.56373882e-02
8.28549087e-01 -2.51001298e-01 5.96185178e-02 -7.09936500e-01
1.35726050e-01 3.61515015e-01 1.01887000e+00 7.33487248e-01
-4.13558841e-01 4.00772303e-01 5.69899082e+00 1.47572607e-01
-1.15730405e+00 8.65224242e-01 -6.57844096e-02 5.60389996e-01
-7.85351098e-01 5.59469640e-01 -5.64832449e-01 5.61271191e-01
1.30812728e+00 -9.55110043e-02 3.67233604e-01 5.43797314e-02
6.95705712e-01 -1.41779035e-01 -5.95134079e-01 1.69599265e-01
-3.94733757e-01 -1.71728981e+00 -3.11046630e-01 1.66971758e-01
8.66477072e-01 6.67267591e-02 -4.90704179e-01 -5.78663468e-01
5.43556690e-01 -4.99265969e-01 8.69321048e-01 5.22685945e-01
7.27677405e-01 -8.32160950e-01 1.12978423e+00 9.66900438e-02
-9.58521605e-01 1.44250384e-02 4.10680771e-01 1.47092223e-01
5.73005319e-01 3.39025706e-01 -8.64681005e-01 8.47176492e-01
3.40069562e-01 -1.88580111e-01 2.72609472e-01 6.16525888e-01
-5.32148659e-01 6.82904720e-01 -3.34725261e-01 -7.32426718e-02
-6.15244694e-02 -4.46723163e-01 4.84659433e-01 1.00560105e+00
8.05812031e-02 -3.84884290e-02 -6.13831103e-01 6.73552990e-01
2.65361778e-02 -1.37274116e-01 -6.80127203e-01 3.95899564e-02
5.90094090e-01 8.40171754e-01 -6.62249029e-01 -2.35569626e-01
-1.43791139e-01 4.66653258e-02 -6.35903716e-01 2.67168909e-01
-7.67556071e-01 -2.45838985e-01 2.94570655e-01 3.71411175e-01
6.62533283e-01 -1.30613580e-01 -1.63785130e-01 -4.70235586e-01
-3.80854718e-02 -1.13861537e+00 5.31979978e-01 -1.89127952e-01
-5.58693707e-01 2.03343347e-01 2.24707797e-02 -8.72407198e-01
-3.75617743e-01 -1.87416561e-02 -4.55517024e-01 6.66046143e-01
-1.65042078e+00 -1.60657036e+00 3.54418397e-01 8.21241558e-01
-8.15099657e-01 9.30214152e-02 9.10963118e-01 3.27953845e-01
9.58764777e-02 3.20088565e-01 1.30054176e-01 1.21110436e-02
4.69584197e-01 -2.51336962e-01 2.49536514e-01 1.02910709e+00
-6.15194201e-01 1.07716286e+00 9.18769121e-01 -1.41718221e+00
-1.78902388e+00 -7.99832046e-01 1.35231829e+00 -2.24733949e-01
7.34185398e-01 -4.31743145e-01 -6.14182532e-01 1.01086164e+00
8.56795013e-01 -4.67472464e-01 4.84470040e-01 -5.23435354e-01
-3.91032398e-01 -6.12791777e-01 -1.76249158e+00 7.04591125e-02
2.66367406e-01 -5.94810426e-01 -3.77987653e-01 3.82559925e-01
7.53786206e-01 2.36740753e-01 -1.09071028e+00 3.37173849e-01
9.71062243e-01 -7.83907950e-01 2.41684064e-01 -5.22299349e-01
-6.55084312e-01 -8.52067769e-01 3.17234546e-01 -4.58640397e-01
4.36975002e-01 -1.27559221e+00 -1.31159529e-01 1.40882993e+00
-1.05416596e-01 -1.47201216e+00 4.77333307e-01 5.94521880e-01
7.45705009e-01 8.31975192e-02 -1.21287382e+00 -7.09325850e-01
-3.85300368e-01 -2.96732336e-02 1.20950079e+00 1.06752336e+00
2.10007161e-01 -4.68351781e-01 -4.38318074e-01 4.07804608e-01
1.04463863e+00 2.66926557e-01 7.09156275e-01 -8.59499753e-01
-1.04033509e-02 4.47715133e-01 -2.45909125e-01 1.25097439e-01
-4.08726484e-01 -6.28509760e-01 -4.50794578e-01 -1.38683522e+00
1.14324719e-01 -5.64822912e-01 -2.19463602e-01 4.19793963e-01
6.12175405e-01 -8.00941437e-02 3.93020988e-01 3.77081722e-01
-2.33094320e-01 -2.30773732e-01 6.80878043e-01 -2.50971299e-02
2.03204289e-01 -6.68370500e-02 -8.42595026e-02 -2.13478394e-02
7.44586527e-01 -1.19893909e+00 -2.79004008e-01 3.35559435e-02
7.71132231e-01 3.38261396e-01 4.76089805e-01 -3.61787617e-01
6.01091862e-01 -2.78534621e-01 -5.15154123e-01 -8.30075443e-01
-6.30867779e-01 -1.56637609e+00 1.64155042e+00 1.74075866e+00
3.35929543e-01 3.91000241e-01 -3.12230229e-01 2.74888486e-01
-2.68572895e-03 3.07681561e-02 2.77712137e-01 1.23902045e-01
1.29149035e-01 1.09927155e-01 -4.52707529e-01 -4.17821020e-01
1.57452166e+00 -2.30115876e-01 -1.24530923e+00 -4.04134281e-02
-3.15085292e-01 4.36806202e-01 7.54883528e-01 7.20692202e-02
3.96536738e-01 -9.13848460e-01 -9.44911003e-01 5.26360750e-01
-2.69772857e-01 -5.89265227e-01 2.39442423e-01 1.24513030e+00
-9.71874416e-01 6.91224754e-01 -7.56279826e-01 -4.36942756e-01
-1.21207261e+00 9.01245534e-01 2.99127102e-02 -5.76502919e-01
4.34168032e-04 -4.58915830e-01 -4.26785499e-01 -1.21522047e-01
7.28648379e-02 -2.21391395e-01 -2.82002967e-02 1.25999004e-01
7.56253600e-01 6.00780129e-01 1.22628115e-01 -5.16295195e-01
-8.61279607e-01 4.20441240e-01 1.19121194e-01 -2.63714761e-01
1.15239108e+00 -2.15346605e-01 -9.95091558e-01 -1.97651401e-01
6.71809316e-01 4.60857391e-01 -6.59679770e-01 2.29896963e-01
3.76165360e-02 -1.55613378e-01 -5.32519162e-01 -1.14134121e+00
-8.88507426e-01 3.81741136e-01 7.49004662e-01 3.25066596e-01
7.01374114e-01 -5.96388817e-01 1.02519262e+00 -2.90978789e-01
9.29024458e-01 -7.50744045e-01 -1.04563856e+00 1.09816507e-01
4.24175352e-01 -5.24492204e-01 -9.58428457e-02 -1.59737125e-01
-2.02704579e-01 8.47597957e-01 -2.81607598e-01 2.07417265e-01
7.12935328e-01 4.94474441e-01 1.81088880e-01 -5.63145518e-01
-6.31514788e-01 6.35033071e-01 -5.57570696e-01 9.63204801e-01
-1.48418665e-01 6.36505902e-01 -1.21657133e+00 2.30515525e-01
5.36150038e-01 9.00244057e-01 6.06549919e-01 1.25096154e+00
-1.20770723e-01 -1.71267843e+00 -6.81580842e-01 -2.93847054e-01
-9.11573589e-01 1.50831804e-01 -2.90833384e-01 9.71888304e-01
5.80130994e-01 1.07382762e+00 -3.10754895e-01 1.67255476e-01
2.22782478e-01 -2.95483023e-02 1.72075659e-01 1.39672577e-01
-1.46429980e+00 -2.94844627e-01 5.00873089e-01 -4.17103618e-01
-7.60635674e-01 -5.25007010e-01 -1.38092601e+00 -1.07741141e+00
-3.31851810e-01 7.06462026e-01 1.26028347e+00 8.20448875e-01
5.81373930e-01 -1.65259063e-01 8.20846021e-01 5.34582078e-01
-4.34060872e-01 -3.79341990e-01 -3.71013194e-01 4.19446677e-01
5.33625484e-01 -2.50252783e-02 -2.20051575e-02 -1.55621478e-02] | [5.980743408203125, 6.434967041015625] |
3bab3e9a-8175-49b3-ad3f-46b240db93c1 | adversarial-text-generation-without | 1810.06640 | null | http://arxiv.org/abs/1810.06640v2 | http://arxiv.org/pdf/1810.06640v2.pdf | Adversarial Text Generation Without Reinforcement Learning | Generative Adversarial Networks (GANs) have experienced a recent surge in
popularity, performing competitively in a variety of tasks, especially in
computer vision. However, GAN training has shown limited success in natural
language processing. This is largely because sequences of text are discrete,
and thus gradients cannot propagate from the discriminator to the generator.
Recent solutions use reinforcement learning to propagate approximate gradients
to the generator, but this is inefficient to train. We propose to utilize an
autoencoder to learn a low-dimensional representation of sentences. A GAN is
then trained to generate its own vectors in this space, which decode to
realistic utterances. We report both random and interpolated samples from the
generator. Visualization of sentence vectors indicate our model correctly
learns the latent space of the autoencoder. Both human ratings and BLEU scores
show that our model generates realistic text against competitive baselines. | ['Anna Rumshisky', 'David Donahue'] | 2018-10-11 | null | null | null | null | ['adversarial-text'] | ['adversarial'] | [ 2.89299130e-01 2.93726593e-01 5.56437187e-02 -3.98355335e-01
-9.92239356e-01 -8.06605458e-01 1.03042638e+00 -3.95590156e-01
-1.91778511e-01 9.16679323e-01 7.54850030e-01 -1.46168545e-01
7.63414323e-01 -8.19734335e-01 -8.28955531e-01 -6.11482263e-01
3.04335684e-01 6.51207328e-01 -4.65072095e-01 -2.42119193e-01
-6.39521405e-02 1.62068054e-01 -1.01334929e+00 3.83214802e-01
7.43168175e-01 3.53822768e-01 1.12306684e-01 1.21781707e+00
-1.30443424e-01 1.08942103e+00 -1.25163221e+00 -6.02807224e-01
2.27406994e-01 -1.23874140e+00 -4.74670798e-01 4.39865403e-02
3.73053133e-01 -7.92769313e-01 -5.88659763e-01 9.14918303e-01
4.76462871e-01 6.06131218e-02 1.00303173e+00 -1.19929588e+00
-1.02391732e+00 6.41629815e-01 -3.26760471e-01 -1.07200623e-01
4.52885687e-01 3.53078216e-01 1.05707836e+00 -7.85884917e-01
8.07281435e-01 1.19196522e+00 5.73920012e-01 1.03857374e+00
-1.34801698e+00 -7.55245268e-01 -1.94190338e-01 -4.00779814e-01
-1.01024377e+00 -4.82741266e-01 6.45209014e-01 -3.58133197e-01
9.95404661e-01 1.95251498e-02 6.21879637e-01 1.75294232e+00
4.48351353e-01 8.88241470e-01 9.78946209e-01 -3.90029639e-01
4.27461118e-01 1.15179941e-01 -7.06049800e-01 5.70067585e-01
-2.01941520e-01 1.04664691e-01 -4.22578484e-01 -5.63255139e-02
8.83017778e-01 -8.06217641e-02 -4.00474481e-02 5.57638332e-03
-1.00356686e+00 1.30346751e+00 4.95305508e-01 7.31699243e-02
-6.73189700e-01 4.31016147e-01 2.12919146e-01 3.86663586e-01
6.09425485e-01 6.98576629e-01 1.24183752e-01 -4.86013740e-01
-1.02522242e+00 5.41563094e-01 8.55877042e-01 9.37840700e-01
3.96606803e-01 5.93234599e-01 -4.36469167e-01 8.35105956e-01
4.32747602e-02 6.79320216e-01 7.62718260e-01 -9.66078758e-01
6.28795326e-01 1.72631428e-01 1.20419428e-01 -7.79040098e-01
1.60618767e-01 -3.35763812e-01 -1.06165469e+00 4.95283544e-01
3.69811296e-01 -5.81608176e-01 -1.13861918e+00 1.74764669e+00
-1.62071168e-01 1.75934568e-01 3.01272541e-01 7.36850083e-01
4.06156003e-01 1.28927863e+00 8.82716030e-02 7.98008889e-02
7.12881088e-01 -1.04548669e+00 -6.16127014e-01 -6.79947436e-01
2.82160670e-01 -7.28506565e-01 1.13647258e+00 3.18412244e-01
-1.41664195e+00 -5.96967697e-01 -1.09426451e+00 -1.92642398e-02
-1.39529198e-01 1.13024019e-01 3.34715039e-01 6.32470608e-01
-1.29476297e+00 4.33003843e-01 -8.49029899e-01 -2.21736450e-02
5.21327138e-01 1.31607696e-01 -2.52393007e-01 -1.28342286e-01
-1.16675520e+00 8.33499610e-01 1.82003275e-01 -1.39885515e-01
-1.27492917e+00 -2.44924590e-01 -9.13969100e-01 1.48261636e-01
-3.61206174e-01 -9.21078622e-01 1.54634285e+00 -1.35265076e+00
-1.77549672e+00 4.84377712e-01 -1.69997439e-01 -8.27547252e-01
6.92053020e-01 -1.62966177e-01 -3.18070978e-01 1.49357110e-01
9.50797275e-02 1.05462694e+00 1.32204401e+00 -1.15638947e+00
-3.39790612e-01 1.98536381e-01 -6.86295033e-02 3.86889547e-01
-4.10701632e-01 -3.31149369e-01 -1.60523072e-01 -9.77113366e-01
-3.64358932e-01 -8.85478377e-01 -2.88249701e-01 -3.01439613e-01
-4.11977619e-01 -9.93379578e-02 6.24100626e-01 -9.11178827e-01
9.38634396e-01 -1.89156687e+00 1.22237608e-01 -1.14381745e-01
6.43646941e-02 1.48724958e-01 -3.59080493e-01 7.35094607e-01
-3.91731411e-02 1.80187270e-01 -2.99906582e-01 -7.48327494e-01
4.42973748e-02 2.21656814e-01 -7.93645740e-01 1.45435840e-01
3.94312888e-01 1.26777923e+00 -9.80641842e-01 -1.41952455e-01
6.91941828e-02 7.12878108e-01 -6.61223829e-01 5.21298170e-01
-5.72555900e-01 4.07389641e-01 -1.78431675e-01 -4.92278636e-02
3.34626794e-01 -1.87871173e-01 -3.50230522e-02 2.90652364e-01
4.20114905e-01 4.68269706e-01 -5.05543768e-01 1.72478282e+00
-6.64439976e-01 9.67035234e-01 -3.60976130e-01 -8.24860036e-01
9.50475752e-01 5.23682535e-01 -2.57477257e-03 -5.09446025e-01
-7.82408938e-02 -8.86060297e-02 -8.10963809e-02 -1.97687000e-01
5.70800483e-01 -3.95326763e-01 -1.03902258e-01 7.80261159e-01
1.98347330e-01 -6.56689882e-01 1.17418334e-01 4.98574257e-01
1.14059734e+00 1.86343506e-01 2.23145764e-02 3.72723997e-01
1.25075132e-01 2.20648721e-02 1.10885434e-01 8.88029575e-01
2.67460436e-01 1.00499010e+00 5.34859538e-01 -2.15505362e-01
-1.63767719e+00 -1.27447128e+00 5.11089206e-01 7.70438910e-01
-4.90148872e-01 -2.89412260e-01 -1.18822896e+00 -8.94153833e-01
-2.52391279e-01 1.20910060e+00 -7.63677299e-01 -4.35863376e-01
-5.41872919e-01 -3.18792731e-01 6.79646909e-01 7.10610211e-01
2.67860979e-01 -1.45111287e+00 -1.93033919e-01 4.16878939e-01
-1.06763817e-01 -9.38987494e-01 -6.61617875e-01 3.06061376e-02
-8.71554673e-01 -4.05896336e-01 -1.03599358e+00 -8.66816580e-01
8.96386147e-01 -1.33504063e-01 1.40989673e+00 -2.77464062e-01
-3.83322723e-02 2.64265031e-01 -2.75706589e-01 -5.01419961e-01
-1.32207358e+00 1.36878237e-01 -2.47893676e-01 -1.18450008e-01
3.45188707e-01 -3.87134939e-01 -5.43214917e-01 -4.22733016e-02
-9.59837198e-01 1.33269057e-01 6.21577144e-01 1.14344704e+00
2.80926615e-01 -8.58038217e-02 7.43663847e-01 -9.14954722e-01
1.24333131e+00 -5.84891498e-01 -4.33623612e-01 -1.05276383e-01
-5.01405478e-01 1.16520397e-01 1.17501187e+00 -3.36069316e-01
-1.12730861e+00 -2.96707321e-02 -4.60502535e-01 -4.57431138e-01
-3.37919325e-01 3.25203240e-01 1.40142396e-01 6.77975833e-01
9.88415062e-01 4.18285578e-01 6.57229051e-02 -7.22329840e-02
4.85527396e-01 6.88500702e-01 5.98732173e-01 -2.96380788e-01
1.03051484e+00 1.89940616e-01 -5.00526249e-01 -6.71051443e-01
-7.11101532e-01 2.46851519e-01 -1.26374647e-01 6.11471431e-03
9.89964604e-01 -1.10602045e+00 -1.81035474e-01 3.09769958e-01
-1.30891430e+00 -6.70215070e-01 -6.88013613e-01 4.49934602e-01
-7.62318134e-01 -2.69048721e-01 -7.93665409e-01 -6.75691247e-01
-4.16178197e-01 -9.35612381e-01 1.05862331e+00 3.85932565e-01
-4.80822891e-01 -1.19008160e+00 3.01919639e-01 9.23905894e-02
5.86619258e-01 2.01826081e-01 6.97400928e-01 -5.77040434e-01
-3.65470916e-01 -5.53686976e-01 2.36945882e-01 7.10968673e-01
1.56829849e-01 -6.46916553e-02 -1.01227880e+00 -4.12972540e-01
1.28908053e-01 -5.05227506e-01 6.35513008e-01 4.71834987e-01
1.03958464e+00 -5.92221439e-01 -7.96418041e-02 4.40470546e-01
1.13945460e+00 3.16548854e-01 8.31921220e-01 -6.49053305e-02
5.47473133e-01 3.36594731e-01 -1.00507680e-02 1.80917099e-01
5.67793958e-02 2.99519688e-01 2.49323398e-01 -2.33414963e-01
-3.28792095e-01 -9.63970244e-01 6.59423053e-01 9.57506120e-01
1.52378261e-01 -7.54931808e-01 -7.12171257e-01 3.15617085e-01
-1.29609191e+00 -1.24241388e+00 5.49084961e-01 1.85979104e+00
8.98354173e-01 4.40631390e-01 1.62157282e-01 -1.30868927e-01
5.20008206e-01 2.56326467e-01 -7.23303735e-01 -6.10425055e-01
-1.39365718e-01 4.90729570e-01 1.53737053e-01 6.27538443e-01
-7.31646419e-01 1.04016089e+00 7.17440844e+00 5.60557246e-01
-1.18816829e+00 9.03195515e-03 9.61615086e-01 -1.77741081e-01
-6.21792078e-01 -2.19399855e-01 -5.30557990e-01 5.89381397e-01
1.23496711e+00 -3.14126790e-01 6.27490222e-01 8.95163894e-01
1.75565913e-01 2.79450148e-01 -1.12642086e+00 8.12535703e-01
3.08173954e-01 -1.46386588e+00 2.58834809e-01 -1.93544235e-02
1.22153497e+00 6.89261705e-02 3.26396495e-01 5.45959055e-01
9.70486760e-01 -1.50721478e+00 5.63663185e-01 4.37407821e-01
9.29606616e-01 -8.44926655e-01 5.05212307e-01 4.32652801e-01
-5.52443862e-01 2.85746127e-01 -4.19415802e-01 -8.44408050e-02
2.70460278e-01 2.69545436e-01 -1.44622374e+00 -1.16645701e-01
1.57618940e-01 5.70155680e-01 -4.52704430e-01 5.25896609e-01
-7.09050953e-01 9.93783116e-01 2.51623299e-02 -2.75293410e-01
3.86891067e-01 -2.54423171e-01 3.23581666e-01 1.26354635e+00
4.64627057e-01 -2.08870485e-01 -6.80108294e-02 1.17002809e+00
-5.71735978e-01 -9.85122472e-02 -1.02071834e+00 -4.90151733e-01
3.56166214e-01 1.02364743e+00 -5.63680649e-01 -5.46572804e-01
-3.19843739e-01 1.54275894e+00 3.01236302e-01 7.56786287e-01
-8.12914610e-01 -4.28792447e-01 6.21107936e-01 -1.81080699e-01
3.03976506e-01 -3.13906968e-01 -3.58971834e-01 -1.06788588e+00
-9.47175547e-02 -1.16870797e+00 -2.73333467e-03 -1.05779135e+00
-1.34781432e+00 9.67502475e-01 -3.78873318e-01 -1.24313533e+00
-1.29114592e+00 -1.92732796e-01 -7.57322133e-01 1.37639737e+00
-8.44460309e-01 -8.51340115e-01 -1.29845798e-01 4.12378460e-01
9.54996407e-01 -5.63040793e-01 1.25142956e+00 -2.25680947e-01
-1.41518846e-01 7.99046338e-01 3.47697467e-01 4.75776672e-01
5.55394113e-01 -1.37506545e+00 1.24385738e+00 8.93109143e-01
6.19821310e-01 4.54087257e-01 9.68714118e-01 -6.33037508e-01
-1.10888374e+00 -1.00458014e+00 6.94048703e-01 -5.36070526e-01
3.56454164e-01 -6.86705768e-01 -7.58033931e-01 8.75952005e-01
7.88746953e-01 -2.20045611e-01 5.49141824e-01 -3.68864655e-01
-4.26199585e-01 1.90299198e-01 -1.10337818e+00 8.21333945e-01
5.63976347e-01 -7.81396270e-01 -4.24850851e-01 3.60334069e-01
6.70346916e-01 -5.20921528e-01 -5.78539729e-01 -2.65650839e-01
4.23125833e-01 -9.53112900e-01 8.05561364e-01 -7.82870054e-01
1.16770387e+00 1.55136390e-02 6.21422082e-02 -2.01628351e+00
-1.97872490e-01 -6.48878038e-01 -2.22670063e-01 1.20020473e+00
5.65513849e-01 -4.48705882e-01 1.15407717e+00 4.29614097e-01
-4.70478684e-02 -5.78325868e-01 -5.16478717e-01 -6.58245146e-01
4.93016869e-01 -1.31058872e-01 5.38606822e-01 7.25747645e-01
-3.01718533e-01 7.93296099e-01 -5.65802455e-01 -2.71922201e-01
3.85984927e-01 -1.43997252e-01 8.67992818e-01 -7.05458343e-01
-7.29018152e-01 -4.75395977e-01 -1.52722150e-01 -1.39447618e+00
3.21228206e-01 -9.34620976e-01 2.83637434e-01 -1.74527609e+00
-1.17117912e-01 -1.12868816e-01 1.98734373e-01 2.79736280e-01
-2.76395768e-01 4.25113648e-01 1.31997824e-01 1.76312923e-01
3.06065790e-02 7.37503052e-01 1.22741985e+00 -3.51727486e-01
-1.60583809e-01 8.15775202e-05 -6.92490041e-01 5.41017711e-01
1.02104831e+00 -5.06490886e-01 -6.27702594e-01 -6.91664398e-01
1.94956064e-01 1.58328399e-01 1.12025954e-01 -1.03757083e+00
-1.11830113e-02 1.11605111e-03 9.41438615e-01 -2.17513993e-01
5.37633717e-01 -4.21476215e-01 7.75319487e-02 3.67364377e-01
-7.10364938e-01 2.95584023e-01 2.60227993e-02 5.50293028e-01
-3.40724081e-01 -3.26733500e-01 7.14933395e-01 -2.35505298e-01
-1.33408323e-01 1.57085687e-01 -5.81067741e-01 4.08593267e-01
7.86520362e-01 4.88130637e-02 -1.91695556e-01 -1.25612152e+00
-5.39631307e-01 -1.19115464e-01 6.03958309e-01 5.35461009e-01
7.12205231e-01 -1.40429866e+00 -9.96039093e-01 3.59496683e-01
-3.78674865e-01 3.18923369e-02 -8.36729351e-03 -5.13206348e-02
-5.90878308e-01 2.11295232e-01 -1.79891914e-01 -4.37421173e-01
-6.93238020e-01 4.37669903e-01 2.73552388e-01 -3.15687954e-01
-6.30407333e-01 9.39163029e-01 1.27030805e-01 -1.41028196e-01
1.35306060e-01 1.65020108e-01 1.52477369e-01 -2.05952153e-01
5.80062270e-01 1.34010660e-02 -1.53464511e-01 -3.93779218e-01
1.41553655e-01 8.98781344e-02 -3.07170391e-01 -6.75870538e-01
1.34331405e+00 2.71328688e-01 2.74861217e-01 4.07159179e-01
1.39039397e+00 1.40547693e-01 -1.64584100e+00 7.96769708e-02
-3.79979968e-01 -1.82048038e-01 -3.27226371e-01 -7.75652647e-01
-8.81939173e-01 1.01678252e+00 2.28425607e-01 4.71704066e-01
9.60383773e-01 -8.92505199e-02 9.31393504e-01 2.41507217e-01
5.55118434e-02 -8.55298102e-01 4.16568011e-01 5.87200761e-01
1.03139806e+00 -1.05773997e+00 -3.08753997e-01 2.33097717e-01
-9.95429218e-01 1.09647608e+00 5.38234651e-01 -5.86479247e-01
2.36823902e-01 5.06797552e-01 3.56145948e-01 1.98576435e-01
-8.23545456e-01 3.46260995e-01 -2.18284782e-02 6.68628037e-01
7.09877610e-01 5.52504323e-02 1.67183906e-01 9.63831246e-02
-7.82395661e-01 -1.14538357e-01 6.97754443e-01 6.10652149e-01
-1.59843773e-01 -1.30856800e+00 -2.65997261e-01 5.79531252e-01
-4.40298766e-01 -2.74810463e-01 -4.66504246e-01 4.16065216e-01
-4.31800455e-01 8.53862464e-01 3.10827106e-01 -3.54769379e-01
4.57759462e-02 3.25449705e-01 3.35052103e-01 -7.46927083e-01
-5.72533786e-01 2.80995481e-02 -4.87671494e-02 -2.43964046e-01
9.01028793e-03 -5.93317688e-01 -1.12348866e+00 -2.28999302e-01
1.12571061e-01 4.01927382e-01 9.09075320e-01 6.30325019e-01
2.54759222e-01 6.32151067e-01 8.18458974e-01 -7.79965997e-01
-9.15686846e-01 -1.15756083e+00 -2.68292844e-01 6.19265079e-01
3.45299065e-01 8.35409164e-02 -4.16475326e-01 3.93165857e-01] | [11.86758804321289, 9.350616455078125] |
e663a1c0-7e7a-4ab1-b510-6a321b646f01 | semi-tensor-product-based-tensordecomposition | 2109.15200 | null | https://arxiv.org/abs/2109.15200v1 | https://arxiv.org/pdf/2109.15200v1.pdf | Semi-tensor Product-based TensorDecomposition for Neural Network Compression | The existing tensor networks adopt conventional matrix product for connection. The classical matrix product requires strict dimensionality consistency between factors, which can result in redundancy in data representation. In this paper, the semi-tensor product is used to generalize classical matrix product-based mode product to semi-tensor mode product. As it permits the connection of two factors with different dimensionality, more flexible and compact tensor decompositions can be obtained with smaller sizes of factors. Tucker decomposition, Tensor Train (TT) and Tensor Ring (TR) are common decomposition for low rank compression of deep neural networks. The semi-tensor product is applied to these tensor decompositions to obtained their generalized versions, i.e., semi-tensor Tucker decomposition (STTu), semi-tensor train(STT) and semi-tensor ring (STR). Experimental results show the STTu, STT and STR achieve higher compression factors than the conventional tensor decompositions with the same accuracy but less training times in ResNet and WideResNetcompression. With 2% accuracy degradation, the TT-RN (rank = 14) and the TR-WRN (rank = 16) only obtain 3 times and99t times compression factors while the STT-RN (rank = 14) and the STR-WRN (rank = 16) achieve 9 times and 179 times compression factors, respectively. | ['Ce Zhu', 'Xiaolin Huang', 'Yipeng Liu', 'Hengling Zhao'] | 2021-09-30 | null | null | null | null | ['low-rank-compression', 'neural-network-compression', 'tensor-networks', 'neural-network-compression'] | ['computer-code', 'methodology', 'methodology', 'miscellaneous'] | [-2.49436304e-01 -3.53105754e-01 -8.97275582e-02 3.69836316e-02
9.33149233e-02 -2.71295071e-01 4.05451894e-01 -4.62133557e-01
-4.68309104e-01 4.03094500e-01 4.37699497e-01 -5.44043303e-01
-7.64656067e-01 -4.63341027e-01 -5.48871934e-01 -7.84086466e-01
-3.35176557e-01 2.62961686e-01 2.77909756e-01 -3.26294899e-01
-1.20759442e-01 4.71836030e-01 -1.38268566e+00 4.63747442e-01
6.07341290e-01 1.37328911e+00 1.91394150e-01 4.24766988e-01
-9.76464748e-02 8.33420694e-01 -3.54699343e-01 -8.38020384e-01
5.89406550e-01 1.08832859e-01 -7.02486634e-01 -8.53009894e-02
2.18373537e-01 -7.55967855e-01 -1.22098899e+00 1.00623894e+00
2.23170117e-01 2.61351943e-01 4.44996119e-01 -1.25738215e+00
-7.41627872e-01 1.20745182e+00 -8.07101607e-01 4.00280178e-01
-9.60033238e-02 -4.62634534e-01 1.06568241e+00 -1.17502892e+00
1.79242313e-01 1.47355485e+00 7.04480171e-01 7.47982487e-02
-1.04981267e+00 -8.19796145e-01 -8.39313120e-02 4.28168535e-01
-1.58393490e+00 -2.36654639e-01 4.55404311e-01 -2.97760814e-01
1.18505788e+00 5.58538735e-01 7.11094499e-01 7.49594152e-01
1.62274972e-01 7.72301018e-01 8.81223261e-01 1.14812821e-01
-1.63987428e-01 -4.95465308e-01 3.68919998e-01 6.30387187e-01
5.57166994e-01 3.71770933e-03 -5.99130213e-01 1.23521918e-02
1.00554585e+00 6.73271418e-01 -4.13087964e-01 6.76349178e-02
-1.64832664e+00 5.79512417e-01 5.80262244e-01 5.69490135e-01
-5.81416488e-01 2.11475730e-01 9.24983621e-01 6.96139932e-01
1.93545491e-01 -5.92962764e-02 -5.52327335e-01 -3.44054013e-01
-9.38505292e-01 2.40956023e-01 5.18114090e-01 1.31725967e+00
4.60897416e-01 6.52167916e-01 -1.48352593e-01 1.16975379e+00
2.29898289e-01 4.36876118e-01 1.19284749e+00 -8.28514338e-01
8.49610507e-01 5.59936166e-01 -5.19734144e-01 -1.24886763e+00
-2.84811974e-01 -6.79425061e-01 -1.57864022e+00 -5.67533195e-01
4.69197966e-02 5.23804240e-02 -9.81930912e-01 1.30621755e+00
7.33528510e-02 3.21654886e-01 1.69258699e-01 7.54070997e-01
9.85153556e-01 6.78142071e-01 -5.12037635e-01 -2.04199359e-01
1.73669851e+00 -1.00281203e+00 -1.00200975e+00 2.93162733e-01
9.14650619e-01 -9.88260031e-01 5.80276728e-01 6.40563071e-01
-8.61978710e-01 -6.10747874e-01 -1.24188912e+00 -1.88680172e-01
-2.87585557e-01 1.62780166e-01 7.20898747e-01 4.68757749e-01
-1.04042625e+00 9.94361103e-01 -8.23097229e-01 6.56294227e-02
1.82765916e-01 3.79536271e-01 -7.11859226e-01 -5.06694853e-01
-1.30370009e+00 5.60842514e-01 7.13385642e-01 3.12861502e-01
-5.98126769e-01 -8.18764091e-01 -4.74254280e-01 9.51186642e-02
1.11013904e-01 -5.09144366e-01 9.40659225e-01 -1.50096759e-01
-1.23391187e+00 1.02588825e-01 2.59495795e-01 -4.14751291e-01
2.31432408e-01 -3.74327242e-01 -5.78271568e-01 3.26702237e-01
-1.31477714e-01 2.80532837e-01 8.23630631e-01 -6.01806641e-01
-4.70993817e-01 -4.53080982e-01 1.32708754e-02 1.18119173e-01
-7.25634813e-01 7.13315904e-02 -2.68460423e-01 -9.93064106e-01
8.53197873e-01 -8.81972909e-01 -1.23890312e-02 -2.71770597e-01
-5.85671604e-01 -3.25217664e-01 1.07316744e+00 -7.14102507e-01
1.36461365e+00 -2.32694387e+00 5.99291265e-01 1.34090140e-01
8.05023670e-01 4.10192966e-01 -2.75289059e-01 4.56367612e-01
-5.69392979e-01 1.46351427e-01 2.94425577e-01 -3.50597709e-01
-1.32052079e-01 8.15333843e-01 -4.30828631e-01 3.64625454e-01
-5.18118367e-02 4.29223537e-01 -7.05959320e-01 -4.94523972e-01
-1.10410430e-01 6.25112891e-01 -6.38162792e-01 8.91696513e-02
4.61100757e-01 -2.68483967e-01 -3.91078949e-01 5.07782519e-01
1.03972960e+00 3.61085348e-02 -7.10336193e-02 -1.07836711e+00
-1.25625506e-01 2.31930822e-01 -1.25228858e+00 1.27906108e+00
-1.66801304e-01 3.53243202e-01 -1.17654942e-01 -8.94814491e-01
1.07644427e+00 5.31862259e-01 8.37972641e-01 -3.40699166e-01
3.18980247e-01 3.82129610e-01 2.75416285e-01 -4.78508443e-01
1.04254711e+00 -3.35235931e-02 4.20328110e-01 4.41075206e-01
3.53634179e-01 4.30768877e-01 7.09820628e-01 4.48740959e-01
1.10944247e+00 -1.58245564e-01 -3.99310440e-01 -2.06388280e-01
3.65633965e-01 -5.57799101e-01 7.90651321e-01 3.03517524e-02
1.35300115e-01 5.48394799e-01 6.50558650e-01 -6.49025857e-01
-1.50267255e+00 -8.55476797e-01 -3.99714261e-01 1.03980982e+00
-2.14281857e-01 -9.30528522e-01 -3.41417462e-01 -2.95037895e-01
-8.71189907e-02 1.21368475e-01 -3.44655633e-01 -1.73937723e-01
-6.41364455e-01 -7.45173931e-01 8.72469544e-01 6.44992352e-01
1.05431533e+00 -2.50376552e-01 1.69546567e-02 -1.90422498e-02
-3.29652607e-01 -1.29516637e+00 -8.08296740e-01 1.28954709e-01
-1.41094673e+00 -7.83546448e-01 -9.35843647e-01 -5.21254659e-01
6.55817568e-01 7.15443134e-01 5.56707084e-01 9.48510766e-02
3.17234814e-01 -1.03696212e-01 -8.37705553e-01 3.33297074e-01
-9.77366865e-02 -2.07430050e-01 6.29591227e-01 2.44259328e-01
1.25030234e-01 -9.90909755e-01 -4.66672152e-01 5.33621967e-01
-1.44770336e+00 1.00510471e-01 8.51240456e-01 1.04300618e+00
4.58717078e-01 2.69745857e-01 -1.01990007e-01 -4.55324203e-01
8.76947582e-01 -3.06471199e-01 -1.59106016e-01 2.95380340e-03
-7.70681083e-01 3.26397836e-01 9.66544330e-01 -9.07014847e-01
-4.46476191e-01 -3.87324542e-01 1.84290677e-01 -1.45491862e+00
8.17712545e-01 9.85644341e-01 1.88886076e-01 -1.47240218e-02
6.63994253e-01 4.81525093e-01 1.25546545e-01 -8.79886031e-01
3.06676507e-01 4.51502055e-01 2.81696528e-01 -6.38804376e-01
1.00779212e+00 6.42461702e-02 3.68664563e-01 -7.70237565e-01
-3.91996145e-01 -2.54138559e-01 -5.57155788e-01 1.26470877e-02
5.52305341e-01 -1.00346351e+00 -6.40557945e-01 4.35035586e-01
-1.17003286e+00 4.67254758e-01 -2.61804014e-01 1.06457508e+00
-2.26807192e-01 6.92940056e-01 -9.69136357e-01 -3.79197359e-01
-5.67235768e-01 -1.16410339e+00 6.12207830e-01 -2.16340154e-01
4.85907853e-01 -6.63877428e-01 -4.00989443e-01 2.84527153e-01
5.32268822e-01 3.57050039e-02 7.42428184e-01 -6.22738957e-01
-4.06825006e-01 -3.12904149e-01 -5.58865726e-01 8.94542754e-01
-1.95598394e-01 7.94342533e-02 -4.49919701e-01 -4.43440288e-01
1.85275927e-01 -4.77541722e-02 8.32889915e-01 -1.10303491e-01
1.01025975e+00 -7.75416970e-01 1.57920614e-01 1.10172725e+00
1.21565044e+00 1.63460732e-01 8.72068882e-01 2.21385792e-01
1.06076443e+00 1.06631607e-01 3.01555246e-01 4.10871893e-01
3.32172692e-01 4.28640723e-01 2.25125775e-01 2.23673627e-01
-1.76848546e-01 -2.51008987e-01 6.04950488e-01 1.99955904e+00
-8.98375690e-01 3.59921724e-01 -5.60496807e-01 1.90348610e-01
-1.69693315e+00 -9.56269443e-01 -4.57305372e-01 2.10568500e+00
8.02898228e-01 4.96298671e-02 -6.21505431e-04 8.24476957e-01
5.18932164e-01 2.77581006e-01 -2.83601612e-01 -4.67267573e-01
-1.55173108e-01 -5.18291108e-02 9.31150794e-01 3.51213962e-02
-6.57739937e-01 5.01685202e-01 5.73563194e+00 1.03638530e+00
-1.00683701e+00 2.50426650e-01 3.08455974e-01 1.74137168e-02
-2.11218804e-01 -1.03256740e-01 -6.99777603e-01 2.87618041e-01
9.10568595e-01 -1.72931120e-01 8.47990513e-01 6.85248435e-01
-2.38953717e-02 5.62044024e-01 -1.03953624e+00 1.36064088e+00
-2.17399925e-01 -1.10503840e+00 4.81940329e-01 2.42475793e-01
5.07881939e-01 2.30916813e-01 7.68455863e-02 5.28110206e-01
1.71069965e-01 -7.59157002e-01 6.50320530e-01 4.51994687e-01
8.93007040e-01 -7.86580861e-01 8.07726204e-01 1.52589604e-02
-1.62129748e+00 -1.34467661e-01 -1.07135916e+00 8.77271667e-02
-9.66200605e-02 8.75304282e-01 -4.16917443e-01 1.07042074e+00
7.78247297e-01 1.11393774e+00 -4.37606514e-01 7.34410465e-01
3.32725257e-01 4.43826318e-01 -5.32320261e-01 3.46700698e-02
3.53709519e-01 -4.92262572e-01 5.92474461e-01 9.46436107e-01
5.55889189e-01 8.90487507e-02 2.65795388e-03 4.66205329e-01
-9.80520919e-02 6.49820343e-02 -2.99641937e-01 -2.21115634e-01
5.52515984e-01 1.22899365e+00 -3.15152824e-01 -3.90753150e-01
-2.47388512e-01 6.52125239e-01 2.41573781e-01 4.03258979e-01
-6.77651584e-01 -6.13109231e-01 8.16570401e-01 -2.53252871e-02
5.70757449e-01 -6.90592229e-01 3.11659593e-02 -1.38766682e+00
1.90023512e-01 -8.20880651e-01 2.38665625e-01 -6.76220298e-01
-1.18259287e+00 1.04638171e+00 3.86792123e-01 -1.70542908e+00
1.46394158e-02 -8.16795409e-01 -1.02690503e-01 7.81715930e-01
-1.19004345e+00 -1.06364977e+00 -8.13524723e-02 9.53483045e-01
1.55741930e-01 -5.57024956e-01 5.72715342e-01 8.68292451e-01
-8.46942008e-01 8.82411897e-01 4.85490769e-01 2.19946787e-01
7.50349909e-02 -8.23842406e-01 6.40948787e-02 7.74325132e-01
-2.21966088e-01 9.98627245e-01 5.86600602e-01 -4.09682661e-01
-2.03871965e+00 -1.23239911e+00 6.43198788e-01 3.24831575e-01
1.03843153e+00 3.33700888e-02 -9.05855477e-01 7.11924851e-01
-1.08035393e-01 1.49629563e-01 5.50429106e-01 2.29553841e-02
-1.06641328e+00 -7.09750295e-01 -1.02409029e+00 6.19503617e-01
1.06046021e+00 -5.29875278e-01 -4.45632577e-01 5.83800316e-01
1.34966505e+00 -3.54856670e-01 -1.73506916e+00 3.66438568e-01
7.00027227e-01 -1.13110733e+00 1.02501106e+00 -4.49353993e-01
6.62547767e-01 -3.51824671e-01 -6.29805923e-01 -9.80132639e-01
-7.69463122e-01 -5.74230433e-01 -3.14935863e-01 1.00244105e+00
1.71324492e-01 -8.49611282e-01 4.56831157e-01 3.63594174e-01
-5.00531733e-01 -9.50687051e-01 -1.37430429e+00 -1.29229319e+00
-1.20810799e-01 -3.85978043e-01 9.69826937e-01 1.15019500e+00
4.30829339e-02 3.02195728e-01 -5.99593520e-01 -2.14366242e-01
5.55728257e-01 -4.89694893e-01 4.94164944e-01 -9.31493104e-01
-3.62983823e-01 -4.39061195e-01 -7.06334233e-01 -1.42341352e+00
-4.80101317e-01 -1.16439950e+00 -5.55327475e-01 -1.30748546e+00
8.81088674e-02 -3.48623157e-01 -4.46958959e-01 5.88479698e-01
3.72093588e-01 2.69409627e-01 4.12065655e-01 7.49678433e-01
-1.10726133e-01 8.62345636e-01 1.78933716e+00 -1.28623158e-01
-1.11225536e-02 -3.03554714e-01 -4.54585761e-01 2.99741447e-01
4.87182617e-01 -2.61612505e-01 -6.34757578e-01 -7.49559164e-01
2.42512628e-01 5.21518588e-02 -3.76185387e-01 -1.00479484e+00
3.41080874e-01 1.33590654e-01 1.28136560e-01 -8.27865481e-01
6.02939904e-01 -8.49906921e-01 4.49656993e-01 6.60697281e-01
-5.95185533e-02 7.11859226e-01 -9.80213284e-02 3.51840466e-01
-3.54465693e-01 -2.57712811e-01 3.62727523e-01 9.99416858e-02
4.79401031e-04 7.61708379e-01 -6.93360344e-02 -3.36941689e-01
4.87248182e-01 -4.62384880e-01 -6.37097538e-01 -6.50350153e-02
-6.71661556e-01 -2.47949213e-02 -4.17317599e-01 4.10950750e-01
1.24207640e+00 -1.98406291e+00 -8.25829089e-01 4.04251695e-01
-2.15589568e-01 3.02507013e-01 4.13402051e-01 1.15136814e+00
-7.53707647e-01 5.28616726e-01 -2.92644233e-01 -5.47858953e-01
-1.15701449e+00 5.27991354e-01 1.46206528e-01 -5.73350608e-01
-7.71794677e-01 7.16442168e-01 -4.18175273e-02 -2.17677578e-01
1.51873723e-01 -5.83221138e-01 -4.04308289e-01 4.17351536e-03
6.91976488e-01 7.48196244e-01 1.42669156e-01 -9.52189326e-01
5.15649356e-02 4.75351185e-01 -4.78102982e-01 2.24207833e-01
1.61203754e+00 2.28445753e-01 -7.04476058e-01 2.19445288e-01
1.68794990e+00 -5.09778798e-01 -5.90827882e-01 -6.62131906e-01
-2.99769580e-01 -6.84739947e-01 3.85540724e-01 -2.36027054e-02
-1.65986609e+00 6.94573641e-01 4.83012438e-01 4.83766466e-01
1.24110389e+00 -4.78925526e-01 1.09757161e+00 6.07892573e-01
2.64651537e-01 -6.40992343e-01 2.49456659e-01 8.06892514e-01
1.26938438e+00 -2.97467887e-01 4.49727058e-01 -3.07144225e-01
-4.20898557e-01 1.27556980e+00 3.91814083e-01 -2.75761127e-01
9.83330965e-01 8.39532465e-02 -6.04068339e-01 -1.14030831e-01
-1.02752018e+00 2.36539349e-01 4.99724001e-01 3.42111945e-01
4.34202343e-01 1.85194522e-01 -4.41236317e-01 4.44836944e-01
-8.78034174e-01 -3.36557508e-01 5.63275456e-01 6.26494825e-01
-2.57460177e-01 -9.84637380e-01 -5.72524011e-01 9.91793633e-01
-4.38623756e-01 -2.65456885e-01 2.79786885e-01 4.28916246e-01
1.08066954e-01 8.43077958e-01 8.01287219e-02 -1.50460958e+00
4.03557777e-01 -4.02008444e-01 3.05000305e-01 -1.62345693e-01
-4.99615103e-01 3.71484190e-01 1.54317185e-01 -6.38780236e-01
-2.72231102e-01 -4.97182429e-01 -9.90112543e-01 -9.98009622e-01
-3.81894380e-01 1.04011640e-01 9.33545113e-01 6.99858189e-01
2.48304069e-01 6.27656341e-01 7.43688285e-01 -6.87969863e-01
-8.18663418e-01 -1.44073868e+00 -1.00518620e+00 4.23181355e-01
2.13740915e-01 -7.06141710e-01 -4.08628315e-01 -1.14742033e-01] | [8.178206443786621, 3.428417205810547] |
d14c93db-fb27-4ee3-9721-e30bd24fb5f4 | human-in-the-loop-optimization-for-deep | 2306.13104 | null | https://arxiv.org/abs/2306.13104v1 | https://arxiv.org/pdf/2306.13104v1.pdf | Human-in-the-Loop Optimization for Deep Stimulus Encoding in Visual Prostheses | Neuroprostheses show potential in restoring lost sensory function and enhancing human capabilities, but the sensations produced by current devices often seem unnatural or distorted. Exact placement of implants and differences in individual perception lead to significant variations in stimulus response, making personalized stimulus optimization a key challenge. Bayesian optimization could be used to optimize patient-specific stimulation parameters with limited noisy observations, but is not feasible for high-dimensional stimuli. Alternatively, deep learning models can optimize stimulus encoding strategies, but typically assume perfect knowledge of patient-specific variations. Here we propose a novel, practically feasible approach that overcomes both of these fundamental limitations. First, a deep encoder network is trained to produce optimal stimuli for any individual patient by inverting a forward model mapping electrical stimuli to visual percepts. Second, a preferential Bayesian optimization strategy utilizes this encoder to optimize patient-specific parameters for a new patient, using a minimal number of pairwise comparisons between candidate stimuli. We demonstrate the viability of this approach on a novel, state-of-the-art visual prosthesis model. We show that our approach quickly learns a personalized stimulus encoder, leads to dramatic improvements in the quality of restored vision, and is robust to noisy patient feedback and misspecifications in the underlying forward model. Overall, our results suggest that combining the strengths of deep learning and Bayesian optimization could significantly improve the perceptual experience of patients fitted with visual prostheses and may prove a viable solution for a range of neuroprosthetic technologies. | ['Michael Beyeler', 'Matthew Chalk', 'Tristan Fauvel', 'Jacob Granley'] | 2023-06-16 | null | null | null | null | ['bayesian-optimization'] | ['methodology'] | [ 4.84316856e-01 3.38623598e-02 -1.22812539e-01 -2.51835585e-01
-8.42302144e-01 -6.21720552e-01 6.64867461e-02 -2.15975851e-01
-5.46590507e-01 9.18792844e-01 5.57678461e-01 -3.67630161e-02
-4.78691280e-01 -2.33854726e-01 -7.50652850e-01 -6.84136331e-01
1.36930436e-01 5.10475934e-01 1.78722497e-02 1.44700542e-01
2.08011538e-01 5.64743638e-01 -1.54466832e+00 5.92758834e-01
6.22663736e-01 1.39969909e+00 7.58257151e-01 4.89059299e-01
2.67816782e-01 -3.20854820e-02 -4.68032688e-01 6.85848147e-02
1.73731565e-01 -1.97476894e-01 -3.93545628e-01 -2.12593660e-01
1.82503134e-01 -5.28882861e-01 -3.02071154e-01 9.73220468e-01
1.08820522e+00 -2.71761298e-01 6.59037232e-01 -5.66449225e-01
-5.92202604e-01 4.18635458e-01 1.77791223e-01 -1.58302248e-01
1.90529808e-01 6.53864920e-01 6.29597127e-01 -5.30931592e-01
3.91397893e-01 8.14254105e-01 5.52873731e-01 8.95353556e-01
-1.86146784e+00 -4.15683389e-01 -1.27293065e-01 2.39179209e-01
-9.74512696e-01 -6.44054115e-01 7.00062096e-01 -6.98882639e-01
1.01644611e+00 4.62361544e-01 1.06841207e+00 1.57810521e+00
6.75481498e-01 5.10050535e-01 1.48879194e+00 -8.79129022e-02
5.17932117e-01 2.34361991e-01 -3.87794465e-01 1.29851460e-01
1.29877299e-01 7.30051398e-01 -6.23913288e-01 -2.04259768e-01
8.34519565e-01 -2.23751217e-01 -6.30553424e-01 -6.08481824e-01
-1.09462142e+00 4.32928443e-01 5.11517107e-01 2.20090672e-01
-7.89941311e-01 2.65268117e-01 2.00910300e-01 -1.10978022e-01
-2.00379312e-01 1.02220452e+00 -7.29823709e-01 -2.09651589e-01
-8.03568184e-01 1.68529958e-01 6.91172183e-01 4.34455067e-01
9.82232392e-02 -1.07634617e-02 -6.28753245e-01 1.02546334e+00
3.88288826e-01 4.58792001e-01 6.77109838e-01 -1.22909141e+00
-7.28511587e-02 -7.66467303e-02 2.17391953e-01 -6.42154455e-01
-7.12398767e-01 -8.69351625e-01 -6.07102156e-01 7.11312950e-01
1.02132931e-01 -9.40332860e-02 -1.34668434e+00 1.88870585e+00
-3.06958020e-01 -2.35500425e-01 -1.49612457e-01 1.06060016e+00
5.08784294e-01 1.12389095e-01 2.49719962e-01 -3.12763661e-01
1.17215693e+00 -3.46416801e-01 -6.37838125e-01 -7.62167811e-01
-4.40910533e-02 -5.98250449e-01 1.11876965e+00 6.87257826e-01
-1.23040771e+00 -1.29334629e-01 -9.87653255e-01 2.59612799e-01
8.59309062e-02 1.16198644e-01 7.65734792e-01 6.31016552e-01
-1.08125508e+00 8.59308004e-01 -9.03741896e-01 -3.63613129e-01
7.50653684e-01 9.54147577e-01 -1.64276049e-01 -7.43616149e-02
-7.35955596e-01 1.23467398e+00 3.41093481e-01 3.60018760e-01
-9.22309458e-01 -9.12672102e-01 -4.22774613e-01 1.97203025e-01
7.98880607e-02 -1.30356026e+00 1.30763245e+00 -7.21261799e-01
-2.01785636e+00 5.54784238e-01 -4.27815430e-02 -2.05319479e-01
3.39036465e-01 5.74726239e-02 -1.26057461e-01 -2.22673029e-01
-2.36011729e-01 9.02909636e-01 7.78243363e-01 -1.27704668e+00
3.48649397e-02 -4.13597584e-01 -5.35507977e-01 1.65261492e-01
-2.20441073e-01 -1.26545668e-01 -1.44881383e-01 -4.98159021e-01
3.07900667e-01 -6.94013536e-01 -5.00660479e-01 4.59240884e-01
-2.35412657e-01 4.32739258e-01 1.71820242e-02 -7.54644752e-01
7.50425041e-01 -2.18068767e+00 3.35416436e-01 3.00322384e-01
4.66149524e-02 8.06588233e-02 -2.43421048e-01 2.44486123e-01
9.39332247e-02 -8.46341997e-02 -3.80071908e-01 -1.12099536e-01
-5.85471466e-02 4.42279786e-01 -1.09548017e-01 1.69092059e-01
-3.93651053e-02 1.09781349e+00 -7.66683698e-01 -7.71097243e-02
4.11832392e-01 6.74389482e-01 -8.60801578e-01 1.25631854e-01
-2.41717204e-01 7.67258108e-01 -1.08542271e-01 7.16838598e-01
4.79725540e-01 -7.53484890e-02 3.19781393e-01 -6.90335751e-01
1.60716578e-01 4.72161978e-01 -7.87648320e-01 1.94435775e+00
-5.38630545e-01 4.63906109e-01 1.58187047e-01 -7.04183936e-01
5.53686321e-01 2.78795123e-01 6.48940384e-01 -7.79957652e-01
3.71861815e-01 5.48328698e-01 2.99618334e-01 -6.89393878e-01
-1.63541690e-01 -8.51372242e-01 2.13569015e-01 -2.76920170e-01
2.36348778e-01 -4.22759384e-01 -5.19243121e-01 -5.26676893e-01
1.14769530e+00 4.96886186e-02 2.07437187e-01 -1.09222837e-01
-2.45266154e-01 -3.87905240e-01 5.64179540e-01 9.26453173e-01
-2.44930416e-01 8.33316982e-01 2.06207216e-01 -8.63159355e-03
-8.09885204e-01 -1.62770343e+00 -5.16389191e-01 3.30188423e-01
-8.85585397e-02 2.04032417e-02 -2.41605937e-01 -9.07655135e-02
3.08712661e-01 7.81921923e-01 -3.97104234e-01 -4.89677101e-01
-1.59121320e-01 -7.18393028e-01 8.80726278e-02 7.08913088e-01
-2.92255338e-02 -1.19661641e+00 -7.36182511e-01 4.75167662e-01
-8.44286010e-02 -8.52724791e-01 -1.69779062e-01 5.91576815e-01
-1.17287695e+00 -5.16439199e-01 -8.66464078e-01 -6.89527452e-01
5.91727078e-01 -5.51081002e-01 6.59322917e-01 -7.83509731e-01
-4.95857418e-01 2.55957097e-01 2.54854023e-01 -4.37265486e-01
-1.77377343e-01 -3.20464462e-01 3.26921016e-01 -3.25312048e-01
-8.15305784e-02 -9.33837652e-01 -1.02232885e+00 1.16706587e-01
-4.95383084e-01 -4.24231179e-02 1.17323375e+00 1.19637811e+00
8.15320611e-01 -2.09474683e-01 4.55469340e-01 -3.82287145e-01
9.61856902e-01 -1.46279767e-01 -3.90431762e-01 2.58502308e-02
-6.92266822e-01 3.90324861e-01 3.68294388e-01 -6.03011727e-01
-8.43545914e-01 3.35750788e-01 -2.44648889e-01 -4.14892673e-01
-6.57976121e-02 6.03430986e-01 -1.53179988e-01 -3.27154458e-01
9.70879436e-01 1.03553258e-01 2.47049332e-01 -6.56703889e-01
2.41114855e-01 7.10765064e-01 8.16112876e-01 -4.21769708e-01
9.80386883e-02 2.80060917e-01 -3.41543145e-02 -4.23526138e-01
-3.57022554e-01 -7.23315477e-02 -1.24739736e-01 -1.77015558e-01
6.49323642e-01 -4.94450420e-01 -1.14207470e+00 5.32443941e-01
-1.15992451e+00 -5.39743423e-01 -4.18596089e-01 9.57873583e-01
-9.39324558e-01 -2.97561325e-02 -2.56661057e-01 -6.87943399e-01
-2.10605294e-01 -1.42554760e+00 9.73403335e-01 6.61778152e-02
-5.49906492e-01 -5.30098855e-01 -1.92443468e-02 2.50903875e-01
6.18758559e-01 -5.56579195e-02 1.14665139e+00 -4.94570807e-02
-6.40216708e-01 -2.26675808e-01 -9.68478993e-02 5.93581378e-01
3.88730854e-01 -6.02970421e-01 -1.08619165e+00 -3.89397383e-01
1.91976935e-01 -3.54933053e-01 8.56653035e-01 1.03176630e+00
1.43560314e+00 -3.27186495e-01 -5.05629778e-01 8.01837444e-01
1.36003423e+00 2.80043900e-01 7.55912602e-01 -4.84809466e-02
2.38874555e-01 3.87982458e-01 -1.04586855e-01 3.21196318e-01
-9.58616883e-02 8.86474848e-01 4.10517156e-01 2.12507188e-01
-3.79945695e-01 -2.69997150e-01 1.50476351e-01 4.69361514e-01
9.42415595e-02 1.21980697e-01 -6.32294953e-01 5.23135960e-01
-1.42320251e+00 -6.83992267e-01 6.53953731e-01 2.49410534e+00
9.27576840e-01 1.97366700e-01 -2.24136084e-01 -1.22695073e-01
1.80298090e-01 -2.74616718e-01 -1.00759041e+00 -3.41309160e-01
-1.71834990e-01 6.09757781e-01 8.50785732e-01 4.55528766e-01
-4.19723928e-01 5.26230633e-01 7.76051044e+00 3.47474813e-01
-1.39726102e+00 3.00607327e-02 1.50848657e-01 -6.26256347e-01
-4.91215438e-01 -1.35402828e-01 -3.47675532e-01 6.46018207e-01
8.38090539e-01 1.54929653e-01 9.41406548e-01 4.97784287e-01
4.32549715e-01 -1.77121326e-01 -1.23432863e+00 1.28212118e+00
6.87916204e-02 -1.45133114e+00 -2.88056105e-01 1.60402387e-01
4.92702305e-01 2.43543386e-01 3.61300379e-01 -8.88895541e-02
-1.37957200e-01 -1.19820273e+00 6.58580363e-01 9.60848987e-01
9.09867644e-01 -2.14728594e-01 5.41617453e-01 9.29579586e-02
-3.02045584e-01 -5.22495389e-01 -2.22752497e-01 1.86616614e-01
3.74652088e-01 5.20253301e-01 -1.07207286e+00 -8.98851007e-02
8.40132833e-01 2.53619254e-01 -1.08481079e-01 1.71217513e+00
-1.40008301e-01 1.97194844e-01 -5.27883649e-01 -2.98601538e-01
-4.66040432e-01 3.47336024e-01 7.48151720e-01 6.42350256e-01
5.15108645e-01 -5.62178232e-02 -3.15004975e-01 1.09674132e+00
2.16745153e-01 -2.07860991e-01 -4.24151063e-01 -1.33231357e-01
3.22443634e-01 7.51690149e-01 -2.38719195e-01 2.21525997e-01
9.86666828e-02 1.04532361e+00 1.41646206e-01 5.94966829e-01
-2.97070205e-01 -3.06191985e-02 9.17631686e-01 1.74026757e-01
3.54144067e-01 -1.05467275e-01 -1.01447272e+00 -7.64480710e-01
4.08866763e-01 -9.02871072e-01 1.88587066e-02 -1.20403004e+00
-1.29411745e+00 4.21106875e-01 -1.16590478e-01 -1.06983280e+00
-3.29547167e-01 -1.12886572e+00 -1.00060761e-01 1.09827459e+00
-1.15329683e+00 -7.75863469e-01 -6.97025433e-02 3.22948754e-01
1.69491559e-01 1.52112581e-02 9.84781086e-01 1.01816572e-01
-1.74169447e-02 5.21424115e-01 2.24880055e-01 -7.28073776e-01
8.07313681e-01 -9.47826266e-01 -3.20667893e-01 3.67716670e-01
-1.30051553e-01 5.72548807e-01 9.98304129e-01 -7.19984055e-01
-1.25547266e+00 -4.75326478e-01 5.87503254e-01 -2.55848825e-01
3.05046618e-01 -2.56621063e-01 -6.20965004e-01 3.21611136e-01
-1.17650330e-02 -1.57836825e-01 6.65593743e-01 2.32380033e-01
7.88235757e-03 -4.62113470e-01 -1.30668211e+00 9.55200493e-01
1.02471673e+00 -6.14699244e-01 -4.95923012e-01 3.82788956e-01
2.54484206e-01 -5.50173759e-01 -8.71353388e-01 7.38954604e-01
1.10053241e+00 -6.37033701e-01 9.62823212e-01 -5.06197989e-01
5.28296195e-02 -6.68981001e-02 -2.56378889e-01 -1.73229301e+00
-5.09373486e-01 -6.39201343e-01 9.88834128e-02 5.56233227e-01
5.23263276e-01 -8.50457788e-01 8.63505125e-01 9.61819470e-01
-5.25528193e-01 -1.02493024e+00 -1.24933779e+00 -7.32231975e-01
-1.61046907e-02 -6.70012474e-01 2.90170759e-01 1.00717053e-01
3.37888688e-01 -3.58737558e-02 -4.71294820e-01 1.90533563e-01
5.30126691e-01 -3.32904369e-01 6.24022679e-03 -1.08637106e+00
-9.36925530e-01 -7.09650397e-01 -6.56699181e-01 -9.76686060e-01
-2.12411985e-01 -9.53593373e-01 6.15764737e-01 -1.89106083e+00
6.75891787e-02 -3.95081609e-01 -6.32441759e-01 5.07803977e-01
1.66792274e-01 -3.95058235e-03 -2.22795159e-02 5.20498084e-04
4.71762419e-01 6.57209992e-01 1.37596524e+00 -2.35921040e-01
-4.55361605e-01 1.80416450e-01 -9.90047276e-01 3.41554165e-01
6.50122285e-01 -5.12879491e-01 -4.38796669e-01 -5.62281191e-01
1.32325500e-01 2.05770414e-02 6.90467894e-01 -1.09500611e+00
4.10679668e-01 -5.67843914e-02 7.07160234e-01 -1.35121852e-01
7.66613424e-01 -8.86354864e-01 3.94048214e-01 5.87729871e-01
-2.21613407e-01 -5.47430515e-01 5.69923222e-01 5.52281976e-01
1.91368937e-01 -4.43899855e-02 8.29057932e-01 -1.63460076e-02
-3.84404659e-01 9.22239646e-02 -4.79506671e-01 -3.16750556e-01
6.19467258e-01 -4.39849675e-01 -8.30481127e-02 -1.68619007e-01
-1.18500423e+00 -8.74829739e-02 2.29748860e-01 3.90874445e-01
9.16479170e-01 -1.30789566e+00 -4.05896753e-01 4.73086268e-01
7.54399672e-02 -4.30822015e-01 4.56653655e-01 9.03685391e-01
-1.28684431e-01 4.74191517e-01 -6.03741884e-01 -8.63732100e-01
-9.77074981e-01 3.76015574e-01 6.76630735e-01 2.67912626e-01
-4.30525631e-01 9.39813673e-01 1.65538386e-01 -1.87157199e-01
3.22699726e-01 -5.44566989e-01 1.62812173e-01 -3.69151115e-01
8.36354569e-02 -2.76596338e-01 2.60198772e-01 1.26421051e-02
-5.10673523e-01 5.31150579e-01 3.35115828e-02 -5.43744087e-01
1.37535238e+00 1.38380855e-01 1.80332303e-01 3.30837458e-01
8.76367629e-01 -2.89769322e-01 -1.42594564e+00 1.22345835e-01
-5.06029367e-01 -5.85832477e-01 4.49364930e-01 -1.73843491e+00
-8.90969813e-01 6.92108274e-01 1.14882660e+00 -6.05457842e-01
1.48626447e+00 2.32384935e-01 5.75636864e-01 2.70350397e-01
5.39626956e-01 -9.48307335e-01 4.42365417e-03 -1.39236316e-01
1.27621281e+00 -9.86625135e-01 -2.22356454e-01 -2.19144210e-01
-5.16525090e-01 8.21728468e-01 4.13659602e-01 1.61094889e-01
6.77772284e-01 4.45480138e-01 1.16159916e-01 -4.51623239e-02
-5.82630694e-01 -2.48593744e-02 5.21525085e-01 9.70793188e-01
1.07415348e-01 3.00486773e-01 -4.55945104e-01 7.08942890e-01
-1.86665773e-01 5.20211995e-01 3.78007889e-02 7.22371519e-01
-2.15842098e-01 -1.28300738e+00 -1.64223999e-01 8.99650455e-01
-1.09423801e-01 -2.95640618e-01 -1.02207698e-01 3.21945608e-01
1.64388731e-01 5.50128043e-01 -9.48715806e-02 -5.04075944e-01
6.33364677e-01 1.07789114e-01 1.10053813e+00 -6.02187216e-01
-3.31872523e-01 2.64776796e-01 9.71720591e-02 -8.07324052e-01
1.95984721e-01 -8.44139278e-01 -9.00671601e-01 4.14308876e-01
-6.45814762e-02 -5.45957148e-01 1.07271802e+00 8.35838377e-01
6.90804601e-01 7.95134783e-01 1.74981967e-01 -1.08333850e+00
-7.93736279e-01 -7.43948281e-01 -5.36618471e-01 -6.18947670e-02
3.92820597e-01 -9.74339426e-01 -3.72737885e-01 -2.61097908e-01] | [6.880618572235107, 0.17196016013622284] |
005ed018-f5d2-4751-8c7e-5443a1240e17 | corrmatch-label-propagation-via-correlation | 2306.04300 | null | https://arxiv.org/abs/2306.04300v2 | https://arxiv.org/pdf/2306.04300v2.pdf | CorrMatch: Label Propagation via Correlation Matching for Semi-Supervised Semantic Segmentation | In this paper, we present a simple but performant semi-supervised semantic segmentation approach, termed CorrMatch. Our goal is to mine more high-quality regions from the unlabeled images to leverage the unlabeled data more efficiently via consistency regularization. The key contributions of our CorrMatch are two novel and complementary strategies. First, we introduce an adaptive threshold updating strategy with a relaxed initialization to expand the high-quality regions. Furthermore, we propose to propagate high-confidence predictions through measuring the pairwise similarities between pixels. Despite its simplicity, we show that CorrMatch achieves great performance on popular semi-supervised semantic segmentation benchmarks. Taking the DeepLabV3+ framework with ResNet-101 backbone as our segmentation model, we receive a 76%+ mIoU score on the Pascal VOC 2012 segmentation benchmark with only 92 annotated images provided. We also achieve a consistent improvement over previous semi-supervised semantic segmentation models. Code will be made publicly available. | ['Weifeng Yuan', 'Qibin Hou', 'Ming-Ming Cheng', 'Le Zhang', 'YuQi Yang', 'Boyuan Sun'] | 2023-06-07 | null | null | null | null | ['semi-supervised-semantic-segmentation'] | ['computer-vision'] | [ 4.35402125e-01 4.67481166e-01 -4.08590794e-01 -8.41899157e-01
-1.21730840e+00 -6.17309570e-01 2.33082265e-01 -2.16277137e-01
-5.93556404e-01 6.02035403e-01 -5.18930629e-02 -6.11329963e-03
3.11916620e-01 -4.20785457e-01 -9.29687738e-01 -5.06642461e-01
3.69458109e-01 6.27528846e-01 7.31944561e-01 1.37203962e-01
2.47096300e-01 6.48910329e-02 -1.38534868e+00 1.75638571e-01
1.14643335e+00 1.15478754e+00 3.52688402e-01 3.95168275e-01
-2.85409912e-02 5.92035115e-01 -2.23675147e-01 -2.48003781e-01
4.72053170e-01 -4.38734591e-01 -1.41405702e+00 5.25053978e-01
7.37238526e-01 -1.43526807e-01 1.89409479e-01 1.31149888e+00
9.92315337e-02 1.74156994e-01 5.03100932e-01 -1.00111377e+00
-2.54144460e-01 7.85338521e-01 -8.08707058e-01 -8.38143826e-02
-3.50597091e-02 -1.13155204e-03 1.27319634e+00 -8.80124986e-01
8.75108004e-01 9.96017396e-01 7.62783766e-01 5.00546932e-01
-1.36408448e+00 -5.73243380e-01 2.65574425e-01 -2.17134114e-02
-1.43826103e+00 -2.62675822e-01 6.91115022e-01 -2.36752361e-01
7.85163045e-01 3.28959376e-02 4.50219065e-01 6.57511294e-01
-4.65100080e-01 1.26199508e+00 1.38282025e+00 -2.96695918e-01
3.75239342e-01 4.95193042e-02 4.58008707e-01 8.98817778e-01
5.38889617e-02 -1.40082508e-01 -2.48220071e-01 1.11250207e-01
5.87661266e-01 -2.04524145e-01 1.08658627e-01 -6.77016616e-01
-1.06623268e+00 7.30495274e-01 7.14806199e-01 1.50481805e-01
-1.52587980e-01 5.91464452e-02 2.32949033e-01 -1.58159748e-01
7.06797183e-01 4.40478951e-01 -6.67616129e-01 4.93355142e-03
-1.44279385e+00 -3.72948498e-02 6.50512576e-01 1.09944904e+00
1.19976056e+00 -1.76899076e-01 -3.17163765e-02 9.82022166e-01
3.70824337e-01 3.32722425e-01 2.24732295e-01 -1.43390703e+00
2.80796438e-01 4.79853243e-01 -5.18500507e-02 -6.56638265e-01
-2.12151036e-01 -4.23340321e-01 -4.59446281e-01 1.08667687e-01
3.34365845e-01 -7.98610821e-02 -1.46380520e+00 1.56720543e+00
4.44212645e-01 3.63137454e-01 -9.37251672e-02 1.01430309e+00
7.00479388e-01 3.38146180e-01 2.57186204e-01 2.17159197e-01
9.51292098e-01 -1.74114215e+00 -4.62626606e-01 -3.53414983e-01
6.86688364e-01 -6.60424054e-01 1.03421772e+00 3.17900568e-01
-1.00707078e+00 -6.08017385e-01 -1.13861609e+00 -1.25662223e-01
-3.30868781e-01 1.81900740e-01 8.06380033e-01 6.06202722e-01
-1.24306619e+00 7.62754738e-01 -1.13166416e+00 -2.42658734e-01
8.49660516e-01 2.31140211e-01 -2.80958295e-01 -8.70900229e-02
-6.96114182e-01 4.56528187e-01 7.21521080e-01 -6.37057498e-02
-7.62281537e-01 -5.80818355e-01 -1.05856001e+00 -1.32714316e-01
6.34317458e-01 -2.68607855e-01 1.22243309e+00 -1.27695858e+00
-1.30296063e+00 1.32987154e+00 -1.93443745e-01 -6.72613502e-01
7.27950096e-01 -3.06905329e-01 1.46600336e-01 4.13538784e-01
4.84685332e-01 1.46544409e+00 4.37836826e-01 -1.48527467e+00
-7.23454595e-01 -2.92388797e-01 -3.70462120e-01 2.08718508e-01
-2.39851270e-02 -9.13110524e-02 -9.71976399e-01 -6.64632320e-01
4.90455627e-01 -1.13526142e+00 -6.28691435e-01 6.86770380e-02
-7.50332594e-01 -1.50861815e-01 7.12968290e-01 -5.24309754e-01
6.28942549e-01 -2.06746602e+00 5.41099533e-02 2.97015727e-01
1.73857838e-01 3.50600511e-01 -1.89487830e-01 -3.02063525e-01
1.21443503e-01 2.76340991e-01 -9.41162825e-01 -7.60129392e-01
-2.50523929e-02 3.18386734e-01 -9.66893286e-02 3.36359501e-01
3.27988714e-01 1.20043707e+00 -9.31558669e-01 -8.71977806e-01
3.76836360e-01 2.15995356e-01 -4.89933372e-01 9.66441408e-02
-3.81033868e-01 4.44095701e-01 -4.35109794e-01 8.59707773e-01
8.88441145e-01 -5.99944234e-01 -6.28700107e-02 -1.01229466e-01
1.03631549e-01 8.97099450e-02 -1.04093587e+00 2.25154400e+00
3.46586690e-03 4.01223719e-01 3.50274026e-01 -1.07574677e+00
7.97250330e-01 -1.76121086e-01 5.93576968e-01 -6.06892288e-01
6.23719692e-02 2.25052446e-01 -5.36484897e-01 7.34420419e-02
4.97639179e-01 1.77621275e-01 2.37738322e-02 3.14057440e-01
2.91889340e-01 -4.49630648e-01 3.06244552e-01 4.70811039e-01
8.77280056e-01 6.11773849e-01 -2.35773828e-02 -4.89437640e-01
3.55856955e-01 5.45347095e-01 8.78431916e-01 7.64085650e-01
-6.24761701e-01 1.10923481e+00 3.31496716e-01 -2.02731147e-01
-8.40190291e-01 -1.23306191e+00 -2.48731434e-01 9.59159911e-01
5.30638814e-01 -2.41526276e-01 -1.11128139e+00 -1.20940495e+00
-1.03387892e-01 4.39134896e-01 -6.41093910e-01 2.47905433e-01
-3.56792659e-01 -7.73810208e-01 5.11197329e-01 9.17157412e-01
9.37079906e-01 -1.01110041e+00 -2.05854744e-01 8.21575057e-03
-2.82259524e-01 -1.42869699e+00 -5.10215938e-01 3.68143886e-01
-1.07466853e+00 -1.13995445e+00 -7.63666987e-01 -1.05132306e+00
8.48832309e-01 3.63078088e-01 1.19509363e+00 5.77794649e-02
-3.29665363e-01 1.86659932e-01 -4.30352062e-01 -2.91470513e-02
-1.58197731e-01 2.53152728e-01 -4.50822860e-01 -2.88462162e-01
3.35890919e-01 -1.66198015e-01 -6.50946736e-01 5.39240479e-01
-7.50101030e-01 2.56945193e-01 4.74614233e-01 7.60595679e-01
1.39059615e+00 -3.30368042e-01 4.08308774e-01 -1.40573609e+00
-2.14089215e-01 -2.69433856e-01 -6.34341240e-01 1.56183645e-01
-9.11907792e-01 -1.20832384e-01 3.22672933e-01 2.95458417e-02
-1.17835629e+00 7.29119480e-01 -4.10518318e-01 -3.60309839e-01
-5.01500487e-01 1.08863078e-01 -1.39568657e-01 -1.84509069e-01
4.12658036e-01 6.63890317e-02 -1.57346785e-01 -5.31641304e-01
6.29729092e-01 4.77174461e-01 7.77828634e-01 -7.42179573e-01
6.84370995e-01 7.42892385e-01 -3.85437101e-01 -5.03723860e-01
-1.37121499e+00 -1.00774848e+00 -9.88624871e-01 5.21536022e-02
1.17500043e+00 -1.08428097e+00 -3.40703912e-02 5.74811161e-01
-7.15485096e-01 -8.47380459e-01 -3.52886945e-01 2.75453806e-01
-7.63172746e-01 6.67170763e-01 -7.96571672e-01 -3.41181487e-01
-3.74893606e-01 -1.33013999e+00 1.38260829e+00 3.97142947e-01
-1.29896238e-01 -9.96984661e-01 -2.65724175e-02 9.90114272e-01
7.92371705e-02 1.82845563e-01 2.28570312e-01 -9.48446870e-01
-6.72187567e-01 5.09075187e-02 -6.82673514e-01 7.09615827e-01
4.47555743e-02 -4.95561585e-02 -1.07748830e+00 -1.63442716e-01
-4.24847126e-01 -7.88814723e-01 1.37395024e+00 5.22897184e-01
1.50744867e+00 2.31730759e-01 -3.56391996e-01 1.02235162e+00
1.43344963e+00 -2.01325715e-01 4.83383596e-01 2.21758604e-01
9.99566317e-01 5.39662838e-01 1.01984084e+00 -2.69875377e-02
4.28463221e-01 3.76819134e-01 2.93865681e-01 -5.87808907e-01
-2.66359568e-01 -3.24912429e-01 -1.17317237e-01 6.92496300e-01
1.74637631e-01 -1.74221806e-02 -8.42513800e-01 7.37669826e-01
-1.94266701e+00 -4.98538941e-01 -2.01014802e-01 1.85545266e+00
9.48342085e-01 3.17721158e-01 4.26696613e-02 -1.65278226e-01
7.83263206e-01 1.50279090e-01 -7.29185283e-01 -5.62051907e-02
-3.42908412e-01 4.62737620e-01 8.35422277e-01 6.15758598e-01
-1.54741883e+00 1.65207207e+00 6.61895418e+00 8.84496629e-01
-7.36275375e-01 2.44396031e-01 1.08355677e+00 2.19106555e-01
-1.25447273e-01 6.65580481e-02 -8.08364391e-01 3.17899615e-01
6.06284738e-01 4.37033683e-01 -1.23381680e-02 1.01661551e+00
1.03099551e-03 -4.06150013e-01 -7.84455538e-01 7.32605278e-01
6.42050356e-02 -1.35217834e+00 -1.86616689e-01 -1.62664980e-01
1.31287241e+00 5.13040364e-01 -4.50168736e-02 1.66768536e-01
7.48301268e-01 -9.11541224e-01 6.25669897e-01 3.53836268e-02
5.74908555e-01 -6.54258668e-01 7.11109579e-01 1.38070896e-01
-1.19090164e+00 4.83656079e-01 -2.94229954e-01 3.38383228e-01
2.29369059e-01 6.37229741e-01 -6.34940505e-01 5.26688576e-01
8.77342522e-01 1.06137180e+00 -6.96668923e-01 1.00908136e+00
-4.56194907e-01 9.69863415e-01 -3.33334982e-01 3.64621431e-01
5.50487339e-01 -3.64404351e-01 5.52397706e-02 1.35205960e+00
-2.30332449e-01 2.75332574e-02 5.82990170e-01 1.14157641e+00
-1.77535832e-01 4.70579900e-02 -1.43674225e-01 1.65693328e-01
2.70947635e-01 1.38745749e+00 -1.46570539e+00 -5.59229612e-01
-2.64596373e-01 1.30523646e+00 3.68663311e-01 3.40007424e-01
-7.45844603e-01 -2.26799294e-01 3.77552688e-01 -3.16659957e-01
3.94161969e-01 -1.21118791e-01 -7.65630662e-01 -1.06045651e+00
-4.48680632e-02 -5.04915774e-01 3.71336907e-01 -7.39899516e-01
-1.17786312e+00 4.88180578e-01 -1.13057256e-01 -7.34403551e-01
9.92728919e-02 -5.48963368e-01 -3.49966824e-01 6.19624674e-01
-1.73421788e+00 -1.20780587e+00 -2.65443206e-01 3.53376359e-01
7.71462321e-01 3.19297343e-01 5.58923781e-01 6.33514822e-02
-5.51290512e-01 4.51860636e-01 -1.12965880e-02 4.84432429e-01
6.91995502e-01 -1.67851579e+00 6.55109704e-01 9.63445544e-01
3.54334891e-01 3.29767108e-01 3.42277318e-01 -7.77851105e-01
-6.54957235e-01 -1.35974073e+00 5.37478566e-01 -3.82506102e-01
4.74551409e-01 -3.05285245e-01 -1.02161729e+00 8.02876592e-01
3.78789902e-02 3.38210613e-01 4.44731325e-01 -1.34942487e-01
-4.36715066e-01 1.89810485e-01 -1.29196441e+00 3.66245598e-01
1.17861784e+00 -4.03657377e-01 -6.27552688e-01 4.32351768e-01
8.77911270e-01 -4.35863525e-01 -6.87533915e-01 7.18536556e-01
2.52293795e-01 -9.21724379e-01 1.01743603e+00 -2.43038297e-01
3.56304944e-01 -3.12041700e-01 -1.76463023e-01 -1.00290000e+00
8.31885710e-02 -4.56520826e-01 3.69715899e-01 1.13090634e+00
6.28708720e-01 -4.78936017e-01 1.43610013e+00 6.62357628e-01
-2.15878382e-01 -6.70796096e-01 -6.91037178e-01 -7.97260523e-01
1.53709203e-01 -4.37836498e-01 1.31934941e-01 9.21669006e-01
-3.24539661e-01 -4.18649949e-02 -1.54328540e-01 7.20586479e-02
9.92156088e-01 1.96667328e-01 5.70567608e-01 -1.12008262e+00
-2.25550994e-01 -3.15717936e-01 -2.88214535e-01 -1.39430285e+00
5.36514461e-01 -1.02474976e+00 6.19153798e-01 -1.53168440e+00
5.84319651e-01 -6.68098390e-01 -4.98775929e-01 7.61527598e-01
-2.77290493e-01 7.35862017e-01 -1.04821501e-02 1.97898313e-01
-1.27002072e+00 5.04855514e-01 1.21118891e+00 -2.60787100e-01
-1.61058411e-01 -3.86202596e-02 -5.73453367e-01 8.68444562e-01
8.85768294e-01 -5.02947450e-01 -3.63060415e-01 -3.73274773e-01
-3.51866424e-01 -3.27995867e-01 3.18638206e-01 -1.01617074e+00
1.56514838e-01 -5.48646487e-02 2.97447622e-01 -8.71755183e-01
1.57142594e-01 -6.62990987e-01 -3.04037303e-01 2.83653557e-01
-5.35160601e-01 -2.87468910e-01 3.92520279e-02 6.42922640e-01
-4.16112810e-01 -2.07591951e-01 1.07689297e+00 -1.97880805e-01
-1.13333488e+00 3.15168679e-01 -1.29825005e-03 4.43187028e-01
1.02953696e+00 -2.53885269e-01 -1.07853644e-01 -6.14298396e-02
-7.91398704e-01 5.47047615e-01 8.66140902e-01 3.30006063e-01
4.96161103e-01 -8.22842062e-01 -3.74398023e-01 -4.05156501e-02
1.90573037e-01 4.50758964e-01 1.92423522e-01 5.93167365e-01
-6.96321547e-01 2.63217032e-01 -6.76334426e-02 -8.75620008e-01
-1.18376207e+00 2.38949999e-01 2.94744790e-01 -8.27689096e-02
-5.57635665e-01 1.24668610e+00 1.14620209e-01 -6.80732548e-01
3.60404730e-01 -2.47679859e-01 1.70030355e-01 -3.04536283e-01
2.11718380e-01 2.14612067e-01 3.80545482e-02 -7.47634053e-01
-4.19536352e-01 6.69469476e-01 -1.50384888e-01 -2.38205895e-01
1.20766318e+00 -2.03568056e-01 -6.45364374e-02 2.10017443e-01
1.35017955e+00 -2.22393885e-01 -1.78459811e+00 -4.31512356e-01
2.82916248e-01 -2.66971737e-01 1.95419833e-01 -9.18433070e-01
-1.42241216e+00 7.18005657e-01 5.72413683e-01 -2.93834448e-01
9.54237700e-01 3.17312092e-01 9.08443928e-01 2.98438668e-01
3.54694843e-01 -1.53397274e+00 -4.23134752e-02 3.74618530e-01
1.42997071e-01 -1.80601406e+00 -3.31641920e-02 -8.71480405e-01
-8.72003734e-01 5.91711819e-01 8.57684374e-01 -1.77499324e-01
5.69137633e-01 2.61693537e-01 3.29146087e-01 -2.06756219e-01
-2.28210181e-01 -6.09892607e-01 3.47549647e-01 5.13260841e-01
3.70477021e-01 1.82394981e-02 -2.36834988e-01 4.69056219e-01
-9.77257360e-03 -3.97598147e-02 1.77906156e-01 9.25610423e-01
-5.91655195e-01 -1.06100011e+00 -7.43627250e-02 3.10645252e-01
-5.52470863e-01 -1.20823167e-01 -5.78748763e-01 5.95890105e-01
-8.31184760e-02 1.05148530e+00 1.06628202e-01 -2.35590547e-01
-2.73558628e-02 5.22879586e-02 2.44321808e-01 -8.29802871e-01
-3.49016637e-01 2.74935186e-01 -1.30929261e-01 -8.90454531e-01
-8.12754393e-01 -6.12788498e-01 -1.72088706e+00 2.31953993e-01
-3.42894763e-01 1.42111793e-01 7.45652735e-01 1.00322950e+00
3.24900806e-01 3.31904411e-01 4.29128289e-01 -8.16721261e-01
-3.23654681e-01 -7.43868947e-01 -4.93713677e-01 5.12156546e-01
-2.94294655e-02 -4.32850897e-01 -2.93222338e-01 2.24880204e-01] | [9.558653831481934, 0.5828022360801697] |
87d5bd42-a862-4a99-9fc9-9f484546e24b | fast-fixed-backbone-protein-sequence-and | null | null | https://openreview.net/forum?id=jLClifZ6YER | https://openreview.net/pdf?id=jLClifZ6YER | Fast fixed-backbone protein sequence and rotamer design | Protein fixed-backbone sequence design is an important task in computational protein design, and being able to quickly and accurately modify or redesign side-chains is a useful subroutine in the context of functional design problems such as ligand binding site, enzyme, and binder design. We present a fast and accurate learned method for protein fixed-backbone sequence and rotamer design. We find that a graph attention model for joint rotamer and sequence prediction trained on-policy via imitation learning can produce a distributions of accurate sequences for target backbones. We show that this method generalizes to design sequences onto novel generated backbones. | ['Tudor Achim', 'Namrata Anand'] | 2021-09-29 | null | null | null | null | ['protein-design'] | ['medical'] | [ 4.65557218e-01 4.67843354e-01 -4.25127417e-01 -4.32111382e-01
-5.10016978e-01 -8.61561835e-01 2.19762530e-02 8.32232535e-02
-3.02180678e-01 1.47854078e+00 2.58012682e-01 -8.38344514e-01
4.06880751e-02 -2.92507976e-01 -1.34020770e+00 -8.70822072e-01
-1.62457898e-01 8.06040943e-01 -9.88462847e-03 -2.71044970e-01
4.30450857e-01 1.06600451e+00 -5.90284348e-01 3.82751711e-02
7.05243766e-01 -8.75464268e-03 4.79779571e-01 6.60574198e-01
5.60009554e-02 4.15218771e-01 -2.79211491e-01 -1.11577861e-01
3.77188809e-02 -6.61589086e-01 -9.59265053e-01 -4.54323590e-02
-2.19839420e-02 1.26681238e-01 -1.77506506e-01 7.06224978e-01
5.45687258e-01 4.20758218e-01 9.06787872e-01 -3.27663302e-01
-7.91264772e-01 3.65217566e-01 -4.22393888e-01 -1.00366287e-01
3.08607906e-01 6.27733529e-01 1.13670349e+00 -7.20295250e-01
1.24670506e+00 1.22106194e+00 5.07502854e-01 8.26749563e-01
-1.69651294e+00 -4.87217873e-01 2.17965320e-01 5.94350994e-02
-8.01311910e-01 -2.11574167e-01 4.17478144e-01 -4.98636156e-01
1.44778371e+00 -1.93496853e-01 7.27710605e-01 1.12084806e+00
7.72579789e-01 5.80396652e-01 5.10497034e-01 -1.36472479e-01
4.82754171e-01 -6.49915278e-01 1.24366321e-01 6.99586689e-01
-1.56228980e-02 7.04021007e-02 -4.05988634e-01 -6.60458207e-01
7.32459903e-01 1.94986954e-01 -5.83550751e-01 -9.93920922e-01
-1.09791076e+00 1.14218068e+00 5.44496953e-01 -1.04097262e-01
-6.15282118e-01 4.24021125e-01 2.17049018e-01 1.03818767e-01
-2.89994001e-01 1.12466884e+00 -8.01230788e-01 -2.36187562e-01
-2.58595377e-01 5.83904326e-01 9.99403536e-01 7.81704307e-01
7.73664057e-01 9.01379914e-04 1.88376397e-01 4.22255665e-01
1.22638762e-01 4.40713078e-01 1.28687114e-01 -7.73540258e-01
-2.58721739e-01 9.49745029e-02 5.17213881e-01 -2.38458723e-01
-6.18504405e-01 -5.82354292e-02 -3.53225410e-01 2.45938823e-01
2.28307664e-01 -2.36512706e-01 -1.05963469e+00 1.72955418e+00
5.17727494e-01 -5.41775465e-01 1.44588910e-02 7.96838045e-01
1.70527577e-01 7.59479880e-01 2.28112891e-01 -7.73798883e-01
8.95761907e-01 -8.87082160e-01 -5.92973053e-01 2.82856822e-01
5.69027662e-01 -5.30167341e-01 6.18896306e-01 2.92949826e-01
-9.05017734e-01 -1.93787202e-01 -9.65517163e-01 -1.23398624e-01
6.78637549e-02 -3.39203835e-01 7.35753536e-01 1.08796068e-01
-7.45656610e-01 1.16239512e+00 -6.93802655e-01 -2.88889855e-01
2.14423910e-01 8.78151715e-01 -4.78011519e-01 1.57637686e-01
-8.77874732e-01 1.26673961e+00 6.99876130e-01 -2.92663991e-01
-9.35871780e-01 -6.98889315e-01 -4.18807030e-01 -2.06883088e-01
1.85466036e-01 -8.24339449e-01 1.36080885e+00 -1.04426265e+00
-1.74958837e+00 4.04410183e-01 -1.79599792e-01 -4.01437163e-01
3.20755132e-02 -7.62848780e-02 2.53015339e-01 -2.53601611e-01
-2.03675061e-01 8.46445262e-01 8.73449028e-01 -6.89818561e-01
-1.30067602e-01 -3.74448836e-01 -4.60062087e-01 4.35941517e-01
5.45252979e-01 -1.24415418e-03 4.85187471e-01 -5.43036282e-01
-2.23727569e-01 -1.04777181e+00 -8.65998089e-01 1.37782648e-01
-2.86336452e-01 -3.71481657e-01 6.12859428e-01 -6.27948105e-01
4.96135771e-01 -1.46696031e+00 1.07718778e+00 4.70448285e-01
2.88576484e-01 3.94530594e-01 -6.36678338e-02 8.96875024e-01
-2.66498655e-01 -1.81044608e-01 -7.68633783e-02 6.27589583e-01
-2.69991666e-01 2.69465208e-01 -3.14260542e-01 5.83721936e-01
1.59457639e-01 1.04192793e+00 -1.13321316e+00 3.49867791e-01
9.18260496e-03 5.27617812e-01 -7.95181155e-01 3.13172549e-01
-1.08923018e+00 8.94465208e-01 -6.73714280e-01 2.53018111e-01
1.47290379e-01 -5.27372956e-01 1.03987396e+00 -1.83619246e-01
1.20556680e-02 2.98208117e-01 -3.59359145e-01 1.69618392e+00
4.97356541e-02 2.41690218e-01 -2.41460025e-01 -5.77397704e-01
1.09092212e+00 1.76062465e-01 4.68855500e-01 -2.43265256e-01
1.34858852e-02 2.10373566e-01 4.36293632e-01 -1.24823660e-01
6.00675419e-02 -4.50929135e-01 3.09885889e-01 6.39630497e-01
-5.68505260e-04 8.60486031e-02 -9.98103395e-02 -2.80517284e-02
1.20192337e+00 8.33243728e-01 6.63270652e-01 -3.67784172e-01
2.51829505e-01 2.21083984e-01 5.98752201e-01 2.49625802e-01
6.43615872e-02 2.37034187e-01 5.26035666e-01 -9.49923754e-01
-1.86724257e+00 -7.96692729e-01 1.10319860e-01 1.29923594e+00
-2.13437259e-01 -1.78755283e-01 -6.05501235e-01 -6.26642883e-01
3.08870256e-01 3.38012189e-01 -3.37190568e-01 -3.13912719e-01
-1.03075147e+00 -5.94265282e-01 6.18691891e-02 3.49923760e-01
-4.86169130e-01 -1.49489582e+00 -4.98155475e-01 1.03787327e+00
3.10553908e-01 -2.55634427e-01 -9.92576122e-01 7.98849046e-01
-6.63178563e-01 -1.31209767e+00 -1.02799201e+00 -1.09555018e+00
5.42973340e-01 -2.81317551e-02 8.40539038e-01 -1.82237357e-01
-4.90434468e-01 -1.41629264e-01 -1.78483993e-01 -1.25204518e-01
-6.76542103e-01 2.11101398e-01 1.52037472e-01 -2.29218453e-01
3.32754821e-01 -7.77398229e-01 -7.03332365e-01 2.13058457e-01
-5.43850482e-01 1.27979696e-01 5.22816062e-01 1.17321408e+00
9.47927296e-01 -9.78748143e-01 7.28800714e-01 -9.95618463e-01
9.47354198e-01 -1.61519065e-01 -6.92365587e-01 3.42962414e-01
-6.08984530e-01 9.60190535e-01 8.16377759e-01 -4.49165851e-01
-7.68198252e-01 6.23334110e-01 -3.72977912e-01 -2.21020095e-02
7.28591308e-02 1.54220462e-01 -3.07286233e-01 -4.58003819e-01
8.08752477e-01 4.15400505e-01 4.10571247e-01 -6.43736899e-01
5.28578460e-01 2.96998888e-01 1.83447778e-01 -9.00602996e-01
3.46868068e-01 -4.02930155e-02 2.18680233e-01 -9.08870518e-01
-2.47003123e-01 -2.23158360e-01 -5.30589044e-01 3.53531778e-01
8.74686003e-01 -4.49535608e-01 -1.16492736e+00 -3.10975593e-02
-1.33234155e+00 -5.58479667e-01 -8.02431852e-02 2.39671275e-01
-1.09778118e+00 6.20193124e-01 -5.34315884e-01 -2.11293504e-01
-5.53648531e-01 -1.61905003e+00 1.07389092e+00 -1.80140138e-02
-6.57826185e-01 -7.77321458e-01 6.45235240e-01 -3.71722877e-02
7.76484162e-02 3.19913507e-01 1.43314385e+00 -6.26571059e-01
-7.98035026e-01 4.01411116e-01 7.44454488e-02 -8.75494182e-02
2.11342067e-01 -5.73531259e-03 -1.60064176e-01 -4.33647841e-01
-5.65809488e-01 -5.60254514e-01 6.76263213e-01 4.37095195e-01
8.18662763e-01 -4.71620560e-01 -5.66796720e-01 4.50052768e-01
1.00843477e+00 6.37457490e-01 4.80206072e-01 2.04564556e-01
8.49464059e-01 4.27427173e-01 4.25840676e-01 3.14752072e-01
-3.42691123e-01 8.25977921e-01 1.68184578e-01 1.60852000e-01
1.15108028e-01 -5.10308087e-01 3.31352293e-01 1.18963458e-01
-7.44437054e-02 -2.78060913e-01 -7.07589328e-01 5.80331087e-02
-2.07294416e+00 -8.09265256e-01 1.98635593e-01 1.88184154e+00
1.43950224e+00 -2.20015839e-01 5.17272353e-01 -4.70877111e-01
6.80487752e-01 -3.92709225e-01 -1.36726284e+00 -6.21785998e-01
3.01801786e-02 5.33016980e-01 7.97582865e-01 8.97267222e-01
-5.04175663e-01 1.19792151e+00 7.62021255e+00 4.48172122e-01
-8.42869341e-01 -4.22379524e-01 2.63106734e-01 1.29478157e-01
-5.54944813e-01 3.01437110e-01 -8.38894069e-01 2.71087766e-01
7.99273849e-01 1.56245694e-01 7.69020677e-01 7.59995997e-01
3.53194684e-01 4.94467705e-01 -1.17158997e+00 5.93475699e-01
-5.99671543e-01 -1.99339962e+00 3.38654488e-01 1.73083797e-01
8.60914230e-01 -2.23045439e-01 -1.89904228e-01 -2.65679270e-01
7.48286784e-01 -1.22880352e+00 2.20388919e-01 3.67373049e-01
5.87709367e-01 -1.09797275e+00 5.55361696e-02 1.89578056e-01
-6.23748302e-01 3.06743264e-01 -4.99428302e-01 2.47896731e-01
2.66466320e-01 2.12192133e-01 -1.27331686e+00 -3.00957024e-01
-1.28777146e-01 6.28644466e-01 3.76151592e-01 6.74594641e-01
-2.76275963e-01 1.43518135e-01 -1.77680701e-01 -5.98091722e-01
1.26752064e-01 -5.55749357e-01 4.16599900e-01 8.86783004e-01
-3.54785800e-01 3.83325130e-01 4.99678522e-01 8.66256833e-01
-2.86324173e-01 2.23246604e-01 -4.91791129e-01 -4.90304142e-01
1.53402463e-01 7.92680562e-01 -4.47174489e-01 1.70136656e-04
-5.56360148e-02 9.47098255e-01 6.56741917e-01 7.25896001e-01
-7.41782367e-01 -6.12533748e-01 1.23718596e+00 2.73389649e-02
6.63166165e-01 -4.36960518e-01 3.85439754e-01 -7.93612957e-01
-4.46135789e-01 -1.34424341e+00 -1.02769762e-01 -6.33488297e-01
-9.64925289e-01 7.92101175e-02 -4.65065867e-01 -8.70086253e-02
-2.71477789e-01 -8.78949940e-01 -1.06734954e-01 9.58585680e-01
-1.08678758e+00 -8.93858731e-01 5.78211725e-01 1.55740529e-01
6.10684693e-01 -1.68879390e-01 6.86381876e-01 -3.62219602e-01
-2.98641950e-01 4.66833174e-01 6.13106132e-01 -4.48863238e-01
6.11240983e-01 -1.35706532e+00 8.26535702e-01 1.66427717e-01
-3.27610731e-01 9.18759882e-01 1.15199542e+00 -9.80995595e-01
-1.84163594e+00 -1.03261328e+00 5.07057130e-01 -2.63371080e-01
5.15921891e-01 -1.08256228e-01 -9.67581987e-01 8.17638993e-01
6.19170442e-02 -4.90254074e-01 7.13000894e-01 -7.76361525e-02
-1.70597196e-01 4.38669622e-01 -9.72409606e-01 7.74871767e-01
1.14904416e+00 -2.29717612e-01 -3.92580837e-01 9.07345176e-01
9.99179482e-01 -4.58937973e-01 -9.79328752e-01 1.98522493e-01
5.29799581e-01 -3.59599710e-01 9.67059612e-01 -1.40893579e+00
-1.60022229e-01 -4.06432122e-01 1.11151725e-01 -1.18804836e+00
-7.41371810e-01 -1.58421278e+00 -3.77553374e-01 1.01535901e-01
8.12440574e-01 -5.31206846e-01 9.85691667e-01 4.17753160e-01
-1.77777380e-01 -7.94912636e-01 -4.71234471e-01 -6.04429185e-01
2.82710284e-01 5.35459936e-01 4.95089859e-01 4.11675543e-01
4.56039190e-01 8.02777648e-01 -5.66356897e-01 -3.81502688e-01
5.10786593e-01 3.69650394e-01 7.99084485e-01 -7.69201040e-01
-8.02278042e-01 -2.91174233e-01 -1.42458856e-01 -1.58098471e+00
3.73359233e-01 -1.04167092e+00 1.03368357e-01 -1.24983907e+00
1.84942603e-01 -6.46072328e-02 1.56261802e-01 4.74272639e-01
4.90386551e-03 -4.51853752e-01 -1.51246250e-01 1.41959280e-01
-4.13016409e-01 9.44785714e-01 1.53231418e+00 4.53114212e-02
-4.36607599e-01 -6.25694618e-02 -4.07507569e-01 1.34415075e-01
8.07121634e-01 -5.86907566e-01 -2.28403926e-01 3.56236249e-01
5.05676985e-01 1.26802489e-01 -2.20115662e-01 -2.12164670e-01
-3.48146111e-01 -6.27453208e-01 3.21916908e-01 -2.35715315e-01
1.80636019e-01 -5.44535160e-01 4.11523402e-01 9.06542659e-01
-5.41705549e-01 1.92907304e-01 -1.09036371e-01 9.28341389e-01
5.85265934e-01 -1.84687134e-02 1.03313327e+00 -2.96636760e-01
-3.79766554e-01 4.87774283e-01 -8.20832133e-01 -2.51833238e-02
1.02496159e+00 -2.36049220e-01 1.78003386e-02 -2.10064828e-01
-1.00930119e+00 1.39226899e-01 9.59837973e-01 2.25808665e-01
7.20357716e-01 -9.06683624e-01 -3.55336398e-01 1.88917264e-01
-6.41488284e-02 -3.48214060e-01 -4.20879424e-01 1.47497639e-01
-1.15738559e+00 6.07258260e-01 -2.41727635e-01 -4.60235029e-01
-1.19036078e+00 7.79794037e-01 6.47891521e-01 4.12726030e-03
-5.20500600e-01 6.30042255e-01 2.38566652e-01 -4.39957857e-01
2.02454284e-01 -1.48552552e-01 3.21066648e-01 -6.32747889e-01
4.19693977e-01 4.33487073e-02 -2.51659721e-01 -2.81640649e-01
-2.55026937e-01 6.12647414e-01 -5.27128696e-01 2.68913269e-01
1.46116173e+00 3.40199202e-01 -1.37164388e-02 -2.08942980e-01
1.09102881e+00 -5.82708776e-01 -1.69725037e+00 -1.34090334e-01
1.83116272e-01 3.53809707e-02 -4.92943108e-01 -6.91680014e-01
-3.46019208e-01 4.81096685e-01 4.50785726e-01 -7.86967099e-01
2.45976716e-01 -1.65397767e-02 9.12307799e-01 1.15070701e+00
6.20770633e-01 -8.15341711e-01 2.72329777e-01 5.23026645e-01
9.85410690e-01 -8.07214439e-01 6.87427307e-03 5.19600362e-02
-4.39858377e-01 1.37135196e+00 2.91917533e-01 -1.02199547e-01
3.54396403e-01 4.60171029e-02 -2.96932757e-01 -2.29321599e-01
-7.40767539e-01 6.96903840e-02 -8.29054229e-03 9.20379519e-01
6.36994958e-01 4.42988873e-02 -4.50970650e-01 -3.52797210e-02
2.45387629e-01 -1.10172451e-01 3.34395498e-01 1.24231672e+00
-9.93505895e-01 -1.80623496e+00 -1.17423750e-01 8.08710456e-02
-3.47129643e-01 4.50881869e-02 -9.15602744e-01 3.51667166e-01
-4.29472834e-01 3.71638149e-01 -5.79828441e-01 -8.49810243e-03
1.26439676e-01 2.95833439e-01 1.03909326e+00 -7.03982770e-01
-6.59762025e-01 3.79904509e-02 3.99051420e-02 -4.56150800e-01
-5.86356930e-02 -3.86767983e-01 -1.73674917e+00 -4.03851241e-01
-6.08269453e-01 5.42646706e-01 4.84527975e-01 7.41804361e-01
8.42546284e-01 2.40775615e-01 6.06293559e-01 -8.70930076e-01
-9.00454640e-01 -4.76824582e-01 -5.28682053e-01 2.63224512e-01
5.13476789e-01 -4.55251276e-01 3.09210747e-01 6.80295303e-02] | [4.735812664031982, 5.602683067321777] |
9c9b8c6d-c08a-4a00-957c-469b2f6d02b6 | conservative-bayesian-model-based-value | 2210.03802 | null | https://arxiv.org/abs/2210.03802v2 | https://arxiv.org/pdf/2210.03802v2.pdf | Conservative Bayesian Model-Based Value Expansion for Offline Policy Optimization | Offline reinforcement learning (RL) addresses the problem of learning a performant policy from a fixed batch of data collected by following some behavior policy. Model-based approaches are particularly appealing in the offline setting since they can extract more learning signals from the logged dataset by learning a model of the environment. However, the performance of existing model-based approaches falls short of model-free counterparts, due to the compounding of estimation errors in the learned model. Driven by this observation, we argue that it is critical for a model-based method to understand when to trust the model and when to rely on model-free estimates, and how to act conservatively w.r.t. both. To this end, we derive an elegant and simple methodology called conservative Bayesian model-based value expansion for offline policy optimization (CBOP), that trades off model-free and model-based estimates during the policy evaluation step according to their epistemic uncertainties, and facilitates conservatism by taking a lower bound on the Bayesian posterior value estimate. On the standard D4RL continuous control tasks, we find that our method significantly outperforms previous model-based approaches: e.g., MOPO by $116.4$%, MOReL by $23.2$% and COMBO by $23.7$%. Further, CBOP achieves state-of-the-art performance on $11$ out of $18$ benchmark datasets while doing on par on the remaining datasets. | ['Scott Sanner', 'Baher Abdulhai', 'Hyunwoo Kim', 'Michael Gimelfarb', 'Xiaoyu Wang', 'Jihwan Jeong'] | 2022-10-07 | null | null | null | null | ['d4rl'] | ['robots'] | [-2.08601594e-01 3.14902425e-01 -6.75874829e-01 -1.69545725e-01
-1.22005427e+00 -6.69736564e-01 4.26343650e-01 2.27769807e-01
-8.39463174e-01 1.17767513e+00 -8.11651349e-02 -5.77302158e-01
-4.07995671e-01 -5.62849224e-01 -1.05960274e+00 -7.45235980e-01
-4.27993983e-01 5.61851144e-01 -9.55553949e-02 -5.66454642e-02
3.36875618e-01 1.26429006e-01 -1.08681321e+00 -3.65736097e-01
8.04492593e-01 1.39703953e+00 9.90726426e-02 5.25976121e-01
5.00599504e-01 1.04052341e+00 -4.84721541e-01 -1.52805835e-01
5.70826173e-01 -2.58570880e-01 -4.26493227e-01 -2.00291723e-01
-9.58383903e-02 -8.78242254e-01 -2.60031730e-01 1.18243468e+00
5.10590911e-01 3.55621248e-01 4.28338140e-01 -1.03313255e+00
-3.60551942e-03 9.02587354e-01 -6.32042468e-01 -7.34765257e-04
-2.31521912e-02 5.67890644e-01 1.12250996e+00 -2.87744462e-01
3.37477833e-01 1.25785732e+00 2.83650845e-01 4.64466482e-01
-1.39401901e+00 -6.64103627e-01 7.17340231e-01 1.01964064e-01
-9.95085657e-01 -4.71897006e-01 5.66848934e-01 -3.00670832e-01
8.42260242e-01 -3.39231007e-02 6.04668081e-01 1.09696853e+00
2.55408406e-01 8.80780220e-01 1.60514283e+00 -2.92153895e-01
7.66170204e-01 9.76780877e-02 -1.50678694e-01 6.51888728e-01
4.57812458e-01 9.42911267e-01 -6.98521495e-01 -3.14403832e-01
4.52007622e-01 -3.71781588e-01 -1.74354658e-01 -4.51989293e-01
-7.04544365e-01 8.33021522e-01 8.29943046e-02 -4.61644024e-01
-5.17124236e-01 7.19547093e-01 3.58656853e-01 4.06992316e-01
3.89169842e-01 7.59268641e-01 -7.12798059e-01 -6.29599214e-01
-7.11429894e-01 6.57989919e-01 7.38172293e-01 8.97091627e-01
4.49030578e-01 1.86135039e-01 -3.10750276e-01 3.43540013e-01
4.69822735e-01 6.87071502e-01 1.25408053e-01 -1.44906878e+00
6.40142083e-01 -4.16486338e-02 8.82519841e-01 -4.93074059e-01
-1.86605945e-01 -5.07651389e-01 -1.92218974e-01 4.91226286e-01
6.90210283e-01 -5.03826499e-01 -8.19336832e-01 2.13211083e+00
3.33213717e-01 8.12328756e-02 2.97147632e-02 7.13207781e-01
-3.98124278e-01 5.08667827e-01 -9.87863094e-02 -4.48191404e-01
9.44163799e-01 -6.84157312e-01 -6.30841434e-01 -4.69988018e-01
4.48629558e-01 -9.36167762e-02 1.07129073e+00 8.99757147e-01
-1.11562514e+00 -1.32138506e-01 -1.09053361e+00 5.53071022e-01
8.28509182e-02 -2.88605075e-02 4.63079780e-01 7.17295229e-01
-6.77095354e-01 8.59339178e-01 -1.21037757e+00 2.18181789e-01
4.97794330e-01 4.25912142e-01 2.36201420e-01 1.84281349e-01
-1.07707357e+00 1.09157646e+00 2.96700329e-01 6.81763515e-02
-1.68724608e+00 -8.19760680e-01 -6.24741852e-01 8.57857764e-02
1.14649785e+00 -3.00332248e-01 1.90725446e+00 -5.44651151e-01
-2.05794859e+00 1.42208353e-01 2.25879386e-01 -1.09305012e+00
8.88962686e-01 -5.28160870e-01 2.54405010e-02 1.18240230e-01
-2.86555737e-01 2.48109698e-01 1.06022489e+00 -1.41059494e+00
-9.02312517e-01 -2.65893251e-01 4.05935377e-01 1.06453799e-01
-7.91023746e-02 -3.84577155e-01 -1.04624458e-01 -3.28946650e-01
-4.35610414e-01 -1.03700423e+00 -5.19183755e-01 -5.82477264e-02
-2.97004342e-01 -3.31555828e-02 3.44628870e-01 -6.00324273e-01
1.29275572e+00 -1.75987208e+00 -7.22099543e-02 3.76896143e-01
1.62280366e-01 8.71114060e-02 1.08934723e-01 3.86985511e-01
4.92020667e-01 1.30745441e-01 -1.75807744e-01 -3.77738208e-01
4.83247787e-01 2.64072180e-01 -6.28646314e-01 6.85033083e-01
-3.82291041e-02 6.30649507e-01 -1.08756030e+00 -2.03516334e-01
1.87892228e-01 -1.23506457e-01 -8.13092947e-01 3.64911944e-01
-6.98327005e-01 4.42029417e-01 -6.70489311e-01 4.52724367e-01
3.77283335e-01 2.92522125e-02 5.74268699e-01 1.62292123e-01
-1.47560805e-01 3.15835893e-01 -1.21709824e+00 1.40231943e+00
-7.15395331e-01 2.11726099e-01 4.02102232e-01 -1.26628876e+00
6.11645579e-01 1.18017465e-01 5.82102537e-01 -6.84524417e-01
3.32286507e-01 3.41811895e-01 7.22322240e-02 -3.39053214e-01
2.30740458e-01 -2.62796819e-01 -2.84457684e-01 5.72796285e-01
-2.92152371e-02 -3.50703120e-01 9.48452670e-03 -7.62623623e-02
1.25362694e+00 4.83587027e-01 4.52821255e-01 -3.82145047e-01
1.96650773e-02 -1.23500891e-01 7.67312348e-01 1.12787342e+00
-5.17962933e-01 -2.31406391e-01 8.29611242e-01 -1.14627950e-01
-7.06452429e-01 -8.89243603e-01 5.07761464e-02 9.09168422e-01
6.95208758e-02 -2.04268813e-01 -6.66652679e-01 -9.06062901e-01
4.40416992e-01 1.00232363e+00 -7.32325613e-01 -4.01729554e-01
-3.94006044e-01 -5.88015854e-01 2.31960729e-01 4.41399068e-01
4.66459453e-01 -4.57510173e-01 -9.59450543e-01 4.74440098e-01
1.77096725e-01 -7.87892461e-01 -3.01782042e-01 4.58192706e-01
-7.18747258e-01 -9.05691028e-01 -4.13504422e-01 1.81205764e-01
4.66710150e-01 -2.63417900e-01 9.65867221e-01 -3.62137347e-01
3.89866203e-01 5.88162720e-01 -1.32969037e-01 -8.27381611e-01
-3.40427488e-01 -1.43941775e-01 4.59537596e-01 -4.99087647e-02
-2.37157359e-03 -3.50655377e-01 -5.80123723e-01 1.91808373e-01
-4.21671629e-01 -3.48440707e-01 4.31900322e-01 1.05639160e+00
5.92523277e-01 1.43693686e-01 6.45757496e-01 -8.72548461e-01
7.37983167e-01 -4.39429671e-01 -1.48376811e+00 2.92756259e-01
-1.24114275e+00 4.71803904e-01 9.54565883e-01 -5.57549059e-01
-1.16996837e+00 -9.59819257e-02 2.53815562e-01 -5.85956216e-01
3.76514196e-01 5.15729666e-01 4.41904254e-02 9.90209877e-02
6.03064716e-01 -2.75874976e-02 1.89235941e-01 -4.07244086e-01
3.29120964e-01 5.22660017e-01 3.37324858e-01 -1.32162702e+00
7.56766319e-01 3.06185663e-01 5.97224720e-02 -1.95282340e-01
-1.12083590e+00 5.40326461e-02 -8.90109967e-03 -3.79060030e-01
3.95208895e-01 -9.67914045e-01 -1.14641511e+00 2.31394440e-01
-6.24612868e-01 -1.03137028e+00 -6.04534149e-01 6.26182914e-01
-1.16794944e+00 2.19835918e-02 -2.31414080e-01 -1.42332149e+00
-8.00598711e-02 -1.18652511e+00 6.56029820e-01 1.43196136e-01
6.05014563e-02 -9.39131260e-01 1.76093489e-01 1.89914137e-01
3.48384410e-01 2.59685636e-01 8.29547465e-01 -4.93558943e-01
-5.96481860e-01 -1.72960475e-01 1.88302130e-01 6.29692137e-01
-1.57674521e-01 -3.40994954e-01 -1.04520512e+00 -6.90972269e-01
2.10884184e-01 -7.16025054e-01 7.11458027e-01 5.73649704e-01
1.41420317e+00 -7.24368691e-01 -9.90279764e-02 3.67874801e-01
1.48027778e+00 4.26519006e-01 5.68910763e-02 4.86558229e-01
1.37106394e-02 4.29494679e-01 1.15699637e+00 1.14292240e+00
2.68648535e-01 5.22059262e-01 7.95932889e-01 4.84820306e-01
4.62634414e-01 -6.48695529e-01 7.28309274e-01 2.44053990e-01
7.08999783e-02 -1.69246495e-01 -7.25753963e-01 3.25386256e-01
-2.12343431e+00 -7.98748791e-01 6.85784221e-01 2.69830108e+00
1.21422696e+00 5.81111431e-01 2.69302517e-01 -2.97393650e-01
2.91852534e-01 4.26128358e-02 -1.32606280e+00 -2.92267263e-01
4.18560296e-01 2.41054147e-01 1.13628232e+00 7.14860857e-01
-7.75629938e-01 6.93257868e-01 6.25653362e+00 9.43953454e-01
-8.74801099e-01 1.14037946e-01 7.72223175e-01 -5.09700537e-01
-8.13805237e-02 2.26720929e-01 -8.88953507e-01 6.60631537e-01
1.31800783e+00 -6.04739428e-01 1.00734615e+00 1.08701909e+00
6.04652941e-01 -6.31063163e-01 -1.33079326e+00 8.02299440e-01
-4.15375292e-01 -1.17097878e+00 -5.27077377e-01 3.39002460e-01
7.87899613e-01 2.49538347e-02 2.09188774e-01 7.64985263e-01
9.70923185e-01 -8.96185935e-01 1.21643174e+00 5.96689939e-01
5.41368425e-01 -9.03176010e-01 6.55125320e-01 6.71914339e-01
-7.63682127e-01 -5.80110908e-01 -1.10027827e-01 -2.41923541e-01
1.32798478e-02 4.70173568e-01 -7.49478281e-01 4.07849818e-01
5.92082024e-01 3.74563634e-01 -2.64441855e-02 8.06019723e-01
-4.08474773e-01 9.49909985e-01 -5.19946635e-01 -1.35056764e-01
4.07013237e-01 -1.59256876e-01 3.62178057e-01 6.20861292e-01
1.52033120e-01 -1.46535993e-01 4.45200205e-01 9.80257392e-01
-8.96512792e-02 -3.67497891e-01 -3.49295259e-01 -1.81991920e-01
7.10176587e-01 8.02879691e-01 -1.30955055e-01 -3.24577212e-01
2.06464548e-02 2.24211872e-01 4.12467152e-01 4.09560651e-01
-1.01146746e+00 -2.44268641e-01 6.62151694e-01 -2.96127260e-01
3.05474907e-01 -3.93577039e-01 -1.72875732e-01 -8.35082650e-01
1.42536670e-01 -1.17160285e+00 3.21276426e-01 -1.34091467e-01
-1.08038497e+00 -5.38504049e-02 4.38828707e-01 -1.18024075e+00
-6.70919716e-01 -7.00192153e-01 -2.45664611e-01 6.76276505e-01
-1.72826600e+00 -4.64746356e-01 3.34171504e-01 1.78522527e-01
4.37721610e-01 5.32942973e-02 4.79210556e-01 -1.61708012e-01
-6.60317838e-01 5.75761020e-01 5.85536659e-01 -2.47776702e-01
6.59569979e-01 -1.38652658e+00 -1.38362661e-01 7.14576185e-01
-2.99628556e-01 3.72798085e-01 9.51388240e-01 -3.90928984e-01
-1.78513253e+00 -8.61234546e-01 -1.97282527e-02 -5.46823621e-01
9.14634228e-01 -2.96430707e-01 -4.15106446e-01 6.49067640e-01
-9.57072228e-02 1.77766532e-01 1.59213856e-01 2.27188319e-02
-2.51530886e-01 -4.41501647e-01 -1.33745718e+00 6.92454875e-01
8.42369854e-01 -3.31248969e-01 -3.44503015e-01 1.61348194e-01
7.81810760e-01 -6.15247369e-01 -9.88921285e-01 3.16812724e-01
6.80095375e-01 -7.08471179e-01 7.04438508e-01 -8.77480507e-01
1.25442550e-01 -7.95138180e-02 -3.36317211e-01 -1.57485318e+00
1.36419430e-01 -1.31450450e+00 -6.68448985e-01 9.32153165e-01
3.91819507e-01 -8.50822628e-01 7.52134919e-01 9.01162803e-01
-8.79942160e-03 -1.02996242e+00 -1.12289596e+00 -1.33218122e+00
3.82951558e-01 -6.87975049e-01 5.13710499e-01 4.37730432e-01
-2.49121841e-02 -2.72644818e-01 -6.49511337e-01 2.30003580e-01
9.79977012e-01 1.04618087e-01 8.01443398e-01 -6.54011011e-01
-8.33671749e-01 -4.30281878e-01 3.14395726e-01 -1.25545394e+00
3.48745108e-01 -2.81117618e-01 4.05690521e-01 -1.09340918e+00
-6.42023385e-02 -7.92193294e-01 -6.03216946e-01 4.94124711e-01
-1.71744693e-02 -5.92028022e-01 4.16734219e-01 -1.00871556e-01
-6.55413866e-01 9.46147144e-01 1.10338378e+00 -1.34933606e-01
-3.41034919e-01 1.48393422e-01 -7.52251744e-01 6.94465756e-01
1.06216514e+00 -6.55863404e-01 -6.50674403e-01 -2.40152493e-01
2.71728337e-01 3.77302736e-01 2.61820495e-01 -7.71665871e-01
9.06034112e-02 -8.51591468e-01 -8.42109248e-02 -2.60008395e-01
3.04089218e-01 -7.83998609e-01 -3.52581650e-01 7.62590885e-01
-7.71508336e-01 -2.94938505e-01 2.27246046e-01 1.08312786e+00
2.37890691e-01 -2.76458859e-01 8.45517993e-01 -2.19926462e-01
-5.22053063e-01 3.48427236e-01 -3.64903331e-01 4.96652961e-01
9.04927373e-01 3.70848179e-01 -3.87252480e-01 -5.52133739e-01
-4.51552927e-01 6.55532241e-01 1.27566606e-01 1.48435190e-01
2.61921734e-01 -9.74674702e-01 -3.34277809e-01 -1.78909078e-01
-1.26924887e-01 -9.36571881e-02 -2.55215857e-02 8.68165791e-01
5.91828898e-02 4.88868624e-01 1.50418997e-01 -3.05488944e-01
-4.86983031e-01 4.97327805e-01 6.30900562e-01 -6.26931489e-01
-1.78219438e-01 6.50890827e-01 -1.24181300e-01 -2.46620029e-01
5.70164323e-01 -6.35357738e-01 4.17476207e-01 -1.71789438e-01
3.88130456e-01 5.31496167e-01 -2.60073487e-02 2.45915055e-01
-1.06732525e-01 1.36079147e-01 4.85402867e-02 -6.87465847e-01
1.25240719e+00 -1.15255229e-01 4.19618160e-01 4.72084463e-01
7.36381710e-01 -4.87665683e-02 -2.29671526e+00 -3.86458367e-01
2.70043403e-01 -6.44879103e-01 4.70816284e-01 -1.19921970e+00
-7.53104150e-01 6.04195714e-01 5.53779006e-01 7.60379806e-02
9.96807814e-01 -3.00300330e-01 4.27684903e-01 6.82476938e-01
9.06229317e-01 -1.74118519e+00 4.81990241e-02 3.25780630e-01
7.40500450e-01 -1.29196537e+00 2.90854067e-01 4.13150072e-01
-6.32642746e-01 7.44351268e-01 6.71526134e-01 -2.26996809e-01
6.39877379e-01 2.35124364e-01 -2.72837520e-01 1.37915984e-01
-1.33642256e+00 -1.62456498e-01 -1.41854629e-01 5.43462276e-01
-3.68388683e-01 2.49581352e-01 -8.20345506e-02 7.71744847e-01
-3.49809453e-02 1.22841857e-01 4.91496861e-01 1.26141751e+00
-6.55860364e-01 -1.12408280e+00 -3.70831460e-01 5.86614370e-01
-4.12373573e-01 1.80947304e-01 4.04766127e-02 8.81902277e-01
-2.60797352e-01 1.07280254e+00 -2.24988133e-01 -2.04269752e-01
2.52775282e-01 -1.28489316e-01 5.86309135e-01 -4.07463461e-01
-2.70859838e-01 6.77103922e-02 3.79273534e-01 -9.99498069e-01
-1.47193417e-01 -6.90600634e-01 -1.09777176e+00 -1.97357967e-01
-2.62760311e-01 1.72954246e-01 6.72401488e-01 8.94547522e-01
3.36179733e-01 2.42516249e-01 1.04865956e+00 -7.74676740e-01
-2.03807735e+00 -7.41877913e-01 -6.87773466e-01 -1.56187430e-01
4.90324229e-01 -1.06451714e+00 -5.60649693e-01 -4.12506193e-01] | [4.204972267150879, 2.454413890838623] |
e6c6c8de-0ca6-4a8f-8c4b-6c5a514b8abb | on-uncertainty-calibration-and-selective | 2304.08653 | null | https://arxiv.org/abs/2304.08653v1 | https://arxiv.org/pdf/2304.08653v1.pdf | On Uncertainty Calibration and Selective Generation in Probabilistic Neural Summarization: A Benchmark Study | Modern deep models for summarization attains impressive benchmark performance, but they are prone to generating miscalibrated predictive uncertainty. This means that they assign high confidence to low-quality predictions, leading to compromised reliability and trustworthiness in real-world applications. Probabilistic deep learning methods are common solutions to the miscalibration problem. However, their relative effectiveness in complex autoregressive summarization tasks are not well-understood. In this work, we thoroughly investigate different state-of-the-art probabilistic methods' effectiveness in improving the uncertainty quality of the neural summarization models, across three large-scale benchmarks with varying difficulty. We show that the probabilistic methods consistently improve the model's generation and uncertainty quality, leading to improved selective generation performance (i.e., abstaining from low-quality summaries) in practice. We also reveal notable failure patterns of probabilistic methods widely-adopted in NLP community (e.g., Deep Ensemble and Monte Carlo Dropout), cautioning the importance of choosing appropriate method for the data setting. | ['Jeremiah Liu', 'Jie Ren', 'Shashi Narayan', 'Joshua Maynez', 'Du Phan', 'Polina Zablotskaia'] | 2023-04-17 | null | null | null | null | ['probabilistic-deep-learning'] | ['computer-vision'] | [ 8.40299502e-02 3.13257575e-01 -2.27221876e-01 -3.98828536e-01
-1.47514081e+00 -4.44964886e-01 7.88808048e-01 1.88191548e-01
-1.95655003e-01 1.31254709e+00 6.48432255e-01 -1.77351594e-01
-5.03132679e-02 -8.18625152e-01 -1.13185108e+00 -8.70971143e-01
2.66665131e-01 7.51332164e-01 -1.43330768e-01 1.19389534e-01
4.30302173e-01 -9.85122323e-02 -1.29217851e+00 2.87046462e-01
1.48070359e+00 6.93324804e-01 -2.73362726e-01 6.93562388e-01
2.50909268e-03 8.46485257e-01 -1.09841013e+00 -6.91627145e-01
-3.17003459e-01 -2.57450253e-01 -6.30047321e-01 -5.52853525e-01
5.35287261e-01 -5.91674030e-01 -4.61560190e-01 1.11570263e+00
7.44279563e-01 1.26236886e-01 1.21163368e+00 -1.18780661e+00
-8.57965946e-01 1.32443416e+00 -5.45933306e-01 1.24880351e-01
1.67327702e-01 1.76427200e-01 1.28939438e+00 -6.87099576e-01
3.24154228e-01 1.25757444e+00 8.70028079e-01 4.02662158e-01
-1.29034173e+00 -7.22896695e-01 9.50069726e-02 1.29114956e-01
-1.31130660e+00 -7.76048362e-01 4.75677818e-01 -4.09039199e-01
8.90230715e-01 2.06452966e-01 5.53679690e-02 1.78971469e+00
5.79328656e-01 9.22726393e-01 6.55687690e-01 -1.84043078e-03
5.75791359e-01 8.36021602e-02 4.36839163e-01 1.13639347e-01
8.18872809e-01 7.64380246e-02 -8.13780010e-01 -5.12061298e-01
2.67720431e-01 -3.39339077e-01 -3.76419663e-01 1.08643018e-01
-1.06913364e+00 7.98511684e-01 2.12153196e-01 1.18016452e-02
-4.63805139e-01 4.97127801e-01 4.50915635e-01 -1.67899162e-01
6.89179778e-01 5.74141741e-01 -4.28352326e-01 -4.72411573e-01
-1.15885222e+00 8.47775280e-01 9.17059422e-01 1.03381073e+00
2.56120265e-01 2.95833319e-01 -8.50675404e-01 7.98594892e-01
2.19487250e-01 6.76952958e-01 3.76447260e-01 -9.16776955e-01
7.18626261e-01 9.10598189e-02 4.36814308e-01 -1.05135667e+00
-2.12533176e-01 -6.44635022e-01 -1.28868115e+00 -1.90348238e-01
3.50451380e-01 -3.51602018e-01 -8.23957682e-01 1.81661701e+00
-1.16403364e-01 2.33816400e-01 3.00242513e-01 6.52307332e-01
1.00302649e+00 8.79111707e-01 2.19641790e-01 -1.87502764e-02
1.05773759e+00 -6.64631963e-01 -7.21216559e-01 -2.65477419e-01
3.57291698e-01 -4.61493343e-01 9.90186274e-01 5.18259585e-01
-1.10198903e+00 -2.15967491e-01 -1.15184772e+00 6.37936518e-02
-1.63722076e-02 1.19298272e-01 4.24007535e-01 5.34654021e-01
-8.57912242e-01 1.19986463e+00 -9.56928670e-01 8.76625702e-02
7.32499480e-01 7.07469732e-02 -4.37928326e-02 -3.62355188e-02
-1.38688350e+00 9.39166069e-01 5.23916543e-01 7.99086615e-02
-1.08517766e+00 -9.69607711e-01 -6.37307763e-01 4.33866143e-01
2.21822560e-01 -9.50923681e-01 1.65019584e+00 -2.41505399e-01
-1.51751554e+00 2.17098087e-01 -2.41953462e-01 -7.71307528e-01
7.87626266e-01 -7.34076679e-01 -1.22089237e-01 -3.04684639e-01
-4.82236966e-02 3.57790709e-01 6.77734494e-01 -1.42621028e+00
-5.34167409e-01 -1.23314686e-01 -3.64621818e-01 1.86469898e-01
-4.58080471e-02 -3.54649186e-01 -2.12376732e-02 -5.81844151e-01
-1.53128123e-02 -6.63995385e-01 -2.52860367e-01 -7.90731668e-01
-1.04157829e+00 -4.15949792e-01 2.09322616e-01 -6.93300605e-01
1.38957465e+00 -1.65488183e+00 -6.62857713e-03 -1.78806275e-01
2.74974942e-01 2.55346179e-01 -8.12880322e-03 4.30777133e-01
3.58083189e-01 6.57334626e-01 -4.14193064e-01 -6.40400648e-01
2.78061271e-01 4.76224720e-02 -6.88065588e-01 3.68897766e-01
4.10001040e-01 9.18514311e-01 -1.07039201e+00 -4.83412683e-01
-7.57148266e-02 5.40464044e-01 -4.11986679e-01 2.42184430e-01
-5.80502987e-01 3.12751561e-01 -6.23697817e-01 4.87097085e-01
6.07430398e-01 -3.84675384e-01 -2.48599440e-01 -7.58804157e-02
2.35739395e-01 7.89577663e-01 -8.15262794e-01 1.32524681e+00
-3.26603591e-01 6.86865091e-01 -5.28419852e-01 -6.37051523e-01
7.05523312e-01 3.39011788e-01 -7.87859634e-02 -1.48456350e-01
1.53724581e-01 1.50379628e-01 -6.19433820e-02 -9.18124244e-02
1.13468647e+00 1.73860133e-01 -1.23894624e-01 5.95506310e-01
8.06393400e-02 -2.81418175e-01 -1.53819388e-02 4.30578113e-01
9.89378870e-01 5.27407341e-02 1.67173475e-01 -1.97615802e-01
-1.38109505e-01 -4.38354760e-02 6.85342073e-01 1.51353395e+00
9.99035127e-03 1.01875913e+00 8.20853531e-01 2.84489989e-02
-1.27931058e+00 -1.12171018e+00 -2.22059309e-01 8.67873430e-01
-1.82958111e-01 -3.91921818e-01 -7.36971438e-01 -5.75866282e-01
-3.24850809e-03 1.43154883e+00 -5.07659197e-01 -4.84598964e-01
-2.26951241e-01 -1.26060152e+00 7.81099439e-01 4.61102724e-01
2.89682746e-01 -1.00065529e+00 -4.39018250e-01 4.06073540e-01
-4.23593998e-01 -9.13621008e-01 -1.48752794e-01 -4.63120602e-02
-8.26368928e-01 -7.36497998e-01 -8.42384219e-01 1.60842091e-01
2.32650176e-01 -1.03515640e-01 1.65890241e+00 -3.71309400e-01
4.84974086e-01 1.04216123e-02 -2.53675073e-01 -5.88757932e-01
-9.41176295e-01 5.73303461e-01 1.62058562e-01 -5.34383357e-01
1.95554897e-01 -6.18687510e-01 -5.87478518e-01 1.55494940e-02
-7.64040947e-01 -1.02427743e-01 7.02947557e-01 9.02140439e-01
4.31354851e-01 4.32806164e-02 9.86646354e-01 -1.09403062e+00
1.00341773e+00 -7.07729936e-01 -3.71023208e-01 2.42768645e-01
-9.01479959e-01 2.97433555e-01 6.29363358e-01 -3.18792641e-01
-1.24804366e+00 -7.30456054e-01 -2.96029281e-02 -1.13791853e-01
-4.24873903e-02 7.75038362e-01 -8.47789347e-02 5.90167999e-01
9.45044160e-01 3.12201589e-01 -4.38951045e-01 -3.83317083e-01
3.83128732e-01 6.93451524e-01 7.41168201e-01 -6.98328674e-01
2.99106628e-01 1.10453762e-01 -2.81047195e-01 -5.53162038e-01
-1.20593214e+00 1.40080839e-01 -1.04461331e-02 1.14917502e-01
3.95876169e-01 -1.08997607e+00 -3.45299214e-01 7.59157896e-01
-1.50493014e+00 -1.28387213e-01 -1.06969491e-01 3.50333154e-01
-4.31746721e-01 4.72012550e-01 -5.99497318e-01 -9.74278390e-01
-8.07672143e-01 -1.08085012e+00 1.18617153e+00 3.92396986e-01
-6.06944740e-01 -6.98129833e-01 2.02613533e-01 2.05556467e-01
5.36551416e-01 1.96972892e-01 8.88385475e-01 -9.32920039e-01
-4.31260169e-01 -1.96406305e-01 -2.28453979e-01 3.81307244e-01
-2.79081047e-01 5.92455268e-01 -1.29208732e+00 -6.37224168e-02
-2.69428372e-01 -3.02765042e-01 1.20210600e+00 8.12920988e-01
1.18943810e+00 -4.93532896e-01 -4.04021889e-01 2.11672097e-01
1.03588092e+00 -2.43230879e-01 7.32369661e-01 -3.34237069e-02
5.64000666e-01 3.59672099e-01 4.37919825e-01 6.34441435e-01
3.27474385e-01 3.24416250e-01 2.20774263e-01 5.22978842e-01
1.72580421e-01 -5.99120617e-01 3.93636853e-01 7.89173543e-01
1.44986838e-01 -7.98844576e-01 -9.51901197e-01 5.45097113e-01
-2.18987536e+00 -9.31267142e-01 -1.76484391e-01 2.36737299e+00
1.25659537e+00 4.07164752e-01 -2.06756905e-01 -1.36746287e-01
7.60076523e-01 3.65131348e-01 -8.33836496e-01 -2.48665780e-01
-2.89157271e-01 -1.96928263e-01 4.35496390e-01 2.41074786e-01
-1.03536737e+00 8.87071729e-01 6.37288046e+00 1.18833482e+00
-6.74293280e-01 -3.24604213e-02 1.12194037e+00 -2.33050495e-01
-6.54082954e-01 -2.21056223e-01 -1.00423539e+00 7.55831063e-01
1.28532076e+00 -4.20937091e-01 -4.75690328e-03 8.60260367e-01
2.44429320e-01 -3.06709290e-01 -1.31663096e+00 7.63272941e-01
-1.03513248e-01 -1.58938193e+00 2.75503308e-01 -1.58806071e-01
1.08993876e+00 2.43437365e-01 1.16693527e-01 5.74799359e-01
8.27147007e-01 -1.40716839e+00 7.36753404e-01 7.59427190e-01
5.63440442e-01 -9.02057827e-01 1.05877769e+00 4.02596742e-01
-2.51163781e-01 3.22444469e-01 -6.65115476e-01 2.61613458e-01
3.26921940e-01 1.26200008e+00 -7.83564091e-01 5.33712745e-01
6.40982568e-01 5.28109312e-01 -3.31759304e-01 9.59467351e-01
-4.59743470e-01 1.04815733e+00 -4.17803675e-01 -2.38884464e-01
1.12779394e-01 -1.60593335e-02 7.54162252e-01 1.10893095e+00
4.02707666e-01 -1.85466766e-01 -4.41386193e-01 1.33084548e+00
-4.54107523e-01 -3.84318084e-01 -5.51771402e-01 -1.52117640e-01
8.87519300e-01 8.68550479e-01 -3.96801941e-02 -5.07429421e-01
1.47803828e-01 5.38700044e-01 3.63135576e-01 4.01459962e-01
-9.16990280e-01 -2.30260462e-01 4.52132136e-01 -2.86506832e-01
7.85736963e-02 1.11488998e-01 -7.02074409e-01 -1.16316164e+00
2.23281085e-02 -8.50004315e-01 2.27266550e-01 -7.77514637e-01
-1.58793771e+00 6.60993099e-01 1.19852945e-01 -8.06106687e-01
-4.96045262e-01 -6.42509833e-02 -8.69517207e-01 8.76666129e-01
-1.35296428e+00 -7.38813400e-01 -4.62502129e-02 -1.65502459e-01
3.98513198e-01 -1.67152748e-01 6.31051302e-01 -1.18429892e-01
-7.94552624e-01 8.13442409e-01 5.96534729e-01 -2.33899683e-01
8.86247337e-01 -1.28903174e+00 7.86211431e-01 7.83017635e-01
1.99860483e-02 5.59855938e-01 1.30821621e+00 -7.40453959e-01
-1.16514790e+00 -1.21211958e+00 8.80267680e-01 -8.66550565e-01
5.48959911e-01 -7.01735616e-02 -1.24625552e+00 4.29830909e-01
2.18457609e-01 -5.48717916e-01 4.52648669e-01 4.02815461e-01
-2.99753636e-01 1.44428313e-01 -1.11749625e+00 7.42499948e-01
7.11955249e-01 -2.37038374e-01 -7.31586993e-01 3.69691819e-01
1.01327956e+00 -5.12983143e-01 -7.28739500e-01 4.33277309e-01
5.17037213e-01 -1.07612288e+00 8.38000953e-01 -7.16624916e-01
1.16191459e+00 1.03287779e-01 2.11076858e-03 -1.71761429e+00
-2.35882640e-01 -6.40624404e-01 -6.05511665e-01 1.63412130e+00
8.23620915e-01 -5.82810044e-01 8.66878271e-01 8.55191171e-01
-1.84213936e-01 -7.23148823e-01 -8.95702839e-01 -4.65409935e-01
5.63338399e-01 -3.95620257e-01 7.27668166e-01 6.25838816e-01
-2.38221392e-01 4.74441051e-01 -5.49288630e-01 2.13126451e-01
8.16260576e-01 -1.54937413e-02 7.36239314e-01 -1.14822114e+00
-1.88895419e-01 -6.80720568e-01 7.43872523e-02 -9.00385439e-01
3.09970856e-01 -4.14061397e-01 2.64690548e-01 -1.53695285e+00
3.78355384e-01 -2.85509735e-01 -2.70347655e-01 6.71888068e-02
-6.32343113e-01 -2.97132760e-01 -2.91726410e-01 3.03030610e-01
-7.49370039e-01 1.04524279e+00 7.51588345e-01 -1.28731340e-01
-1.41266316e-01 2.05248073e-01 -1.07350516e+00 5.42018652e-01
9.42550242e-01 -7.57141232e-01 -3.10090572e-01 -6.04229271e-01
5.40037513e-01 -3.00576445e-02 3.19189996e-01 -9.77022588e-01
2.26449162e-01 1.94591656e-02 2.78419942e-01 -9.57287014e-01
4.56202887e-02 -2.19784468e-01 1.56481713e-01 1.35954190e-02
-7.19889164e-01 -9.74190682e-02 6.86297193e-02 9.41890597e-01
-1.69751927e-01 -3.86853814e-01 6.35257840e-01 -8.19935128e-02
-2.21019816e-02 1.80173740e-01 -2.76129305e-01 4.33993846e-01
4.19422597e-01 1.94294706e-01 -7.67811537e-01 -6.65614605e-01
-2.70457685e-01 4.54783171e-01 2.87263006e-01 2.77033657e-01
3.93776983e-01 -1.03970551e+00 -1.32489932e+00 -4.00778562e-01
5.37605658e-02 5.33016920e-01 5.34534216e-01 6.32737815e-01
-1.62308395e-01 4.57992852e-01 3.62924159e-01 -6.63664222e-01
-5.95238209e-01 1.31659666e-02 2.11415753e-01 -5.43398619e-01
-4.70794320e-01 8.84865344e-01 -6.46671876e-02 -2.82470524e-01
3.98362309e-01 -5.50229430e-01 1.94567982e-02 1.15570156e-02
5.31953096e-01 7.02790558e-01 2.52659976e-01 -1.16210863e-01
-1.34991616e-01 -1.10003524e-01 -4.24827427e-01 -1.49886742e-01
1.24019170e+00 3.84957902e-02 1.01728946e-01 5.84067106e-01
7.82259345e-01 -2.16992900e-01 -1.33527827e+00 -2.74944454e-01
1.16204917e-02 -1.15761921e-01 2.32176542e-01 -9.39238966e-01
-7.10725904e-01 8.92900229e-01 -4.41179201e-02 2.35758975e-01
3.82595748e-01 -2.19627574e-01 8.36845934e-01 5.02371669e-01
2.30673149e-01 -9.70203936e-01 -3.66812050e-01 5.94889998e-01
9.83701646e-01 -1.38559437e+00 3.51424813e-01 1.14714630e-01
-9.91740465e-01 7.22410202e-01 4.83338088e-01 -1.23863786e-01
3.58469486e-01 9.93234962e-02 -3.44656020e-01 6.11332469e-02
-1.20895386e+00 6.21151924e-01 4.13326085e-01 5.22846520e-01
4.77307349e-01 2.90020078e-01 -4.78464253e-02 1.19933414e+00
-4.94649529e-01 -2.67947286e-01 7.87795782e-01 3.05621237e-01
-2.94589996e-01 -4.34073240e-01 -2.11478218e-01 9.45097864e-01
-6.21847987e-01 -3.60759437e-01 -3.37478161e-01 4.84228164e-01
-5.93940496e-01 1.06873012e+00 -1.06915664e-02 -2.15749353e-01
1.57164723e-01 5.18084951e-02 1.46919906e-01 -4.14361358e-01
-5.49600899e-01 -2.84863949e-01 4.52535659e-01 -2.84831524e-01
-4.34365608e-02 -7.41586745e-01 -8.24921012e-01 -6.76912427e-01
-4.05821055e-01 3.10839176e-01 5.76813698e-01 8.08404684e-01
7.79895961e-01 5.84014893e-01 1.81973964e-01 -7.52988040e-01
-1.34904027e+00 -1.36120069e+00 -3.80106062e-01 2.31408939e-01
3.23041469e-01 -5.86543322e-01 -6.07992530e-01 -3.45057935e-01] | [12.103280067443848, 9.174949645996094] |
fce82e33-da04-40fb-af59-f5168fd15658 | are-you-really-looking-at-me-a-framework-for | 1906.12175 | null | https://arxiv.org/abs/1906.12175v2 | https://arxiv.org/pdf/1906.12175v2.pdf | Are you really looking at me? A Feature-Extraction Framework for Estimating Interpersonal Eye Gaze from Conventional Video | Despite a revolution in the pervasiveness of video cameras in our daily lives, one of the most meaningful forms of nonverbal affective communication, interpersonal eye gaze, i.e. eye gaze relative to a conversation partner, is not available from common video. We introduce the Interpersonal-Calibrating Eye-gaze Encoder (ICE), which automatically extracts interpersonal gaze from video recordings without specialized hardware and without prior knowledge of participant locations. Leveraging the intuition that individuals spend a large portion of a conversation looking at each other enables the ICE dynamic clustering algorithm to extract interpersonal gaze. We validate ICE in both video chat using an objective metric with an infrared gaze tracker (F1=0.846, N=8), as well as in face-to-face communication with expert-rated evaluations of eye contact (r= 0.37, N=170). We then use ICE to analyze behavior in two different, yet important affective communication domains: interrogation-based deception detection, and communication skill assessment in speed dating. We find that honest witnesses break interpersonal gaze contact and look down more often than deceptive witnesses when answering questions (p=0.004, d=0.79). In predicting expert communication skill ratings in speed dating videos, we demonstrate that interpersonal gaze alone has more predictive power than facial expressions. | ['Mohammad Rafayet Ali', 'Minh Tran', 'Taylan Sen', 'Mohammed Ehsan Hoque', 'Kurtis Haut'] | 2019-06-21 | null | null | null | null | ['deception-detection'] | ['miscellaneous'] | [-3.54662724e-02 3.45578343e-01 -1.20414011e-01 -7.65386522e-01
-3.58352810e-01 -8.64989340e-01 3.99717718e-01 -3.36445212e-01
-5.61778367e-01 3.71139705e-01 -2.00638059e-03 1.89193532e-01
-4.74455506e-02 1.62891418e-01 -2.99344033e-01 -3.35731655e-01
3.43844444e-02 -1.32365555e-01 -3.31632227e-01 -3.89283313e-03
6.47010982e-01 3.34603399e-01 -1.52541482e+00 -2.48167347e-02
4.02838916e-01 1.15976393e+00 -5.26092649e-01 1.10057509e+00
5.10775745e-01 1.02210617e+00 -8.06831181e-01 -1.08436012e+00
-9.59828123e-02 -6.15988076e-01 -6.17636204e-01 -1.95496287e-02
9.51125860e-01 -8.30705822e-01 -1.73559144e-01 7.70901322e-01
2.40022302e-01 2.15038404e-01 3.88764948e-01 -1.58235002e+00
-6.62005901e-01 4.83972300e-03 -1.09602797e+00 7.42132008e-01
1.06244552e+00 2.32311189e-01 9.53475177e-01 -5.99012792e-01
7.22979069e-01 1.08701742e+00 8.08717549e-01 9.88962114e-01
-1.19769907e+00 -1.12824774e+00 -2.64515162e-01 2.80431330e-01
-1.38751066e+00 -1.19064009e+00 8.02261174e-01 -4.98690009e-01
6.13020480e-01 4.21772718e-01 6.50469720e-01 1.75224161e+00
5.05663268e-02 3.32779497e-01 1.23662424e+00 -6.01326562e-02
-3.08207870e-01 5.10555267e-01 4.17945623e-01 5.84504306e-01
-4.18622345e-02 3.12746689e-02 -1.33958983e+00 -1.86462209e-01
4.57658499e-01 -1.75136328e-01 -2.73182958e-01 4.89140928e-01
-5.87072015e-01 5.36991596e-01 1.72891378e-01 1.01952806e-01
-1.18324704e-01 7.07704872e-02 1.44665435e-01 5.66361129e-01
6.47271454e-01 3.94129485e-01 3.21069807e-01 -1.14459133e+00
-8.73395503e-01 7.82170892e-02 8.46576035e-01 6.73395097e-01
5.46135783e-01 -4.14570123e-01 1.78337812e-01 7.21915662e-01
3.20458740e-01 5.74253559e-01 2.61283398e-01 -1.55431747e+00
1.28811002e-01 3.97016734e-01 8.49922448e-02 -1.45777440e+00
-4.66748476e-01 2.46999279e-01 -9.13082249e-03 9.52432603e-02
7.13227928e-01 -4.23812926e-01 1.16047174e-01 1.98166454e+00
3.66575599e-01 2.82818556e-01 -5.19703388e-01 1.13028622e+00
6.24660492e-01 1.41014069e-01 -1.10658050e-01 -6.53670371e-01
1.37846518e+00 -2.92540044e-01 -8.36982667e-01 -3.28004926e-01
5.52092135e-01 -5.93573451e-01 8.86377513e-01 6.38385475e-01
-1.23494637e+00 -1.43838465e-01 -8.55300128e-01 -3.32796395e-01
1.67866886e-01 -3.45493466e-01 5.83676159e-01 1.25801885e+00
-1.04585838e+00 4.28143054e-01 -5.63527584e-01 -5.50790191e-01
5.05503058e-01 4.31468695e-01 -7.50519454e-01 2.04646245e-01
-7.63340950e-01 8.53567898e-01 -7.29688168e-01 2.52127320e-01
-4.62438345e-01 -7.04898834e-01 -6.83119953e-01 -4.11234125e-02
2.62784541e-01 -2.26852536e-01 1.44646275e+00 -1.89527655e+00
-1.75683296e+00 1.44417775e+00 -6.02653444e-01 6.65429607e-02
2.67917484e-01 -3.45449299e-01 -7.10045159e-01 6.84793174e-01
-8.38219076e-02 5.29398799e-01 1.32950926e+00 -8.99836659e-01
-3.97285193e-01 -8.10208917e-01 9.40695107e-02 3.27574641e-01
-6.81400716e-01 7.45941699e-01 -3.22843306e-02 6.81271553e-02
-5.32727003e-01 -1.13885319e+00 9.81099248e-01 2.77848274e-01
-7.45161399e-02 -3.42807859e-01 1.09663928e+00 -7.85187662e-01
1.34137034e+00 -2.39354634e+00 6.70884550e-02 5.41038275e-01
9.58234787e-01 7.32811466e-02 3.65476161e-01 1.15689807e-01
-7.55768940e-02 1.56406507e-01 3.99675518e-01 -7.09754825e-01
-3.19306068e-02 -1.92485556e-01 1.79430872e-01 9.52598751e-01
-1.88180760e-01 1.03384054e+00 -7.70686328e-01 -6.55230939e-01
-2.46243253e-01 3.62227559e-01 -4.33971137e-01 1.87088385e-01
5.65534413e-01 4.89655107e-01 1.65638283e-01 6.35933638e-01
5.18364012e-01 -3.54613036e-01 5.89991473e-02 3.72159749e-01
-2.71883328e-02 -5.26195094e-02 -3.38325441e-01 1.38887656e+00
-1.89290032e-01 1.68976963e+00 6.83785081e-01 -3.04825664e-01
6.92782700e-01 1.78637356e-01 1.93458468e-01 -5.26339412e-01
5.43364823e-01 -2.51780540e-01 2.63950437e-01 -8.63212526e-01
5.75678706e-01 -2.21704498e-01 -1.25426771e-02 8.88318181e-01
3.05859745e-02 5.00757992e-01 -9.21585858e-02 3.74101996e-01
1.16634834e+00 -3.12475741e-01 -6.75936742e-03 8.41445625e-02
2.92024195e-01 -4.01443213e-01 2.76258141e-01 3.17634463e-01
-8.18522632e-01 5.17316520e-01 9.76394773e-01 2.69036181e-02
-5.46297789e-01 -8.00270557e-01 -4.65533212e-02 1.42415869e+00
1.82276607e-01 -3.68144989e-01 -1.12866127e+00 -5.69280267e-01
8.41392800e-02 5.54630458e-01 -8.53916883e-01 -3.93870443e-01
-3.17909569e-01 9.07581225e-02 8.04000318e-01 1.58437997e-01
2.76011564e-02 -7.24265635e-01 -4.59387988e-01 -6.12185001e-01
-2.14153692e-01 -1.27354324e+00 -8.05513144e-01 -9.13663328e-01
-3.44474226e-01 -1.23932588e+00 -4.04598504e-01 -1.30774125e-01
5.64417660e-01 4.88436371e-01 9.27992761e-01 5.92017472e-01
-1.36896729e-01 1.09270298e+00 -3.27128381e-01 -2.89854079e-01
-8.81624743e-02 -2.22267330e-01 5.47396839e-01 5.55999815e-01
1.11910093e+00 -5.54148376e-01 -6.14816964e-01 5.22530675e-01
-2.69721806e-01 -5.06535053e-01 7.53492191e-02 3.12991083e-01
-6.26045942e-01 -8.94776046e-01 2.07503811e-01 -7.70242393e-01
9.27822888e-01 -7.20123231e-01 -2.59895742e-01 -1.34787187e-01
-3.99315774e-01 -7.87352026e-01 4.73367386e-02 -4.61805761e-01
-9.99764025e-01 -6.16307855e-01 3.53874415e-01 -8.68108094e-01
-2.20357791e-01 -4.65868451e-02 5.99458039e-01 -3.25806439e-01
8.91183496e-01 -3.98582071e-01 5.17487526e-01 4.65075746e-02
-3.79697651e-01 1.22273397e+00 5.99872828e-01 -2.35608980e-01
5.86299181e-01 4.96215016e-01 -1.53049916e-01 -1.47697973e+00
-8.17152798e-01 -6.05813563e-01 -4.64094251e-01 -1.08099914e+00
9.67252851e-01 -7.39924371e-01 -2.29933405e+00 5.22485375e-01
-1.00182784e+00 -3.47006470e-02 5.14135659e-01 6.11619711e-01
-4.20446903e-01 4.62375343e-01 -6.58096075e-01 -1.21003318e+00
-2.01444611e-01 -6.28731132e-01 1.06441283e+00 4.34204936e-01
-1.08821642e+00 -7.60816753e-01 7.30992481e-02 9.83683705e-01
1.07036151e-01 1.76832661e-01 -2.01338336e-01 -4.86994892e-01
-1.41685486e-01 -4.40748394e-01 -1.21170625e-01 2.11962879e-01
-4.57265554e-03 4.56659108e-01 -1.42508030e+00 -6.60766289e-02
2.22058162e-01 -6.45971358e-01 2.25311786e-01 1.67940497e-01
9.03006911e-01 -1.67356566e-01 -3.59459519e-01 5.27720273e-01
5.17278969e-01 -2.31900904e-02 6.72668099e-01 -1.34367138e-01
6.54321313e-01 9.36853051e-01 4.15685356e-01 4.63356286e-01
3.71929526e-01 6.54115558e-01 3.32068503e-01 6.93407834e-01
4.52018350e-01 -1.58213913e-01 6.85980082e-01 3.53218973e-01
-3.05770427e-01 -2.12887362e-01 -9.16607499e-01 1.18139878e-01
-1.32912207e+00 -1.44933319e+00 -4.00966942e-01 2.10070539e+00
5.71097374e-01 -7.57957250e-02 5.63270867e-01 4.58609685e-02
9.04068530e-01 -5.31794578e-02 -6.11609340e-01 -7.71029770e-01
3.53965312e-01 -1.41221479e-01 1.90527841e-01 5.50246298e-01
-4.94020939e-01 5.12748063e-01 5.94542027e+00 1.75121456e-01
-1.02609909e+00 2.33079389e-01 6.70143723e-01 -1.03347993e+00
3.28642577e-02 -3.04812938e-01 -5.92507124e-01 9.37328339e-01
1.47589469e+00 -1.15837060e-01 8.30042660e-01 5.19983232e-01
1.92551523e-01 -5.93203664e-01 -1.54247296e+00 1.70329499e+00
8.47562551e-01 -9.23125327e-01 -1.12800920e+00 3.01783025e-01
1.94349557e-01 -4.22427714e-01 3.58963013e-01 -1.02519933e-02
-3.05501610e-01 -1.19326937e+00 4.85368192e-01 9.16181743e-01
1.04397357e+00 -6.46742225e-01 3.36886972e-01 4.61994410e-01
-3.89686346e-01 -2.45679557e-01 2.11383596e-01 -5.80008030e-01
1.25897676e-01 -4.77905832e-02 -5.68710983e-01 -2.99013942e-01
1.06799674e+00 9.55824018e-01 -3.44139636e-01 3.01227629e-01
-1.12174198e-01 5.47239840e-01 -4.88515615e-01 -2.43563250e-01
-1.46461889e-01 -1.37726083e-01 5.80282688e-01 6.90906882e-01
7.84640834e-02 5.30866921e-01 -9.78256345e-01 7.96087980e-01
-4.84500855e-01 -4.09952521e-01 -6.35032475e-01 -2.16065735e-01
5.14378428e-01 1.64969432e+00 -1.78751335e-01 1.09526098e-01
-5.54164767e-01 1.04572546e+00 5.11393905e-01 2.72153884e-01
-8.44167233e-01 -2.73532718e-01 1.21512544e+00 4.27645892e-01
-3.09506178e-01 -1.39114618e-01 -2.07175344e-01 -9.68429029e-01
1.81674391e-01 -7.56028354e-01 4.11261246e-02 -1.17966998e+00
-1.34357548e+00 1.29488990e-01 -1.40685245e-01 -1.05249929e+00
-4.79435682e-01 -4.72452611e-01 -7.27997720e-01 8.76729429e-01
-9.41991627e-01 -6.08130634e-01 -8.33646357e-01 6.86843336e-01
9.57405940e-02 8.20492357e-02 3.91903341e-01 1.90691590e-01
-8.50518465e-01 1.18670905e+00 -5.23233473e-01 2.43110776e-01
1.34934080e+00 -8.73381615e-01 -5.06150760e-02 3.97000819e-01
-1.45294890e-01 9.40179288e-01 5.38432360e-01 -2.20244199e-01
-1.51274598e+00 -4.87629510e-02 8.47245097e-01 -1.16939986e+00
9.28498387e-01 -6.02898717e-01 -5.95809042e-01 1.05701673e+00
3.72209907e-01 -1.69815972e-01 1.26989412e+00 4.68028933e-01
-4.70323414e-01 -1.80365101e-01 -1.33224595e+00 4.62101907e-01
1.31525922e+00 -1.10144877e+00 -6.53269887e-01 1.15515895e-01
4.41364385e-02 -5.10225534e-01 -1.01789558e+00 -3.56094003e-01
1.18911815e+00 -1.48153353e+00 5.30367196e-01 -4.75284040e-01
5.91590524e-01 4.16847616e-01 5.12071788e-01 -8.87545109e-01
-3.02048605e-02 -1.17721117e+00 -2.06763744e-01 1.30555713e+00
-2.52808612e-02 -6.68013036e-01 8.51865947e-01 1.44724214e+00
1.19854853e-01 -1.19828515e-01 -9.44579959e-01 -2.86504626e-01
-3.39720517e-01 -3.92587215e-01 5.00004962e-02 1.19505191e+00
8.43106329e-01 5.30596137e-01 -5.47995150e-01 2.39162873e-02
4.75593299e-01 -4.68418181e-01 1.09733450e+00 -1.55250859e+00
-5.90494350e-02 -3.56169641e-01 -7.23320842e-01 -8.43518972e-01
7.17363775e-01 -2.48194769e-01 -4.84106213e-01 -3.38190943e-01
3.02230775e-01 1.31335989e-01 2.48415709e-01 8.33497941e-02
7.57659599e-02 5.79525411e-01 1.34777859e-01 4.20801163e-01
-8.02812338e-01 8.98608044e-02 8.67596090e-01 3.55265766e-01
-2.27083471e-02 -1.19611949e-01 -8.83455634e-01 7.20041335e-01
4.58885372e-01 -3.15189183e-01 -2.55673349e-01 -1.63321510e-01
6.55380666e-01 2.89892912e-01 6.09124303e-01 -6.86746120e-01
3.91422153e-01 9.86797065e-02 4.84077066e-01 3.78070474e-02
8.37172091e-01 -7.96710014e-01 -2.17745826e-01 -2.14981005e-01
-3.43182296e-01 1.90178901e-01 -6.31934404e-02 5.71421027e-01
2.40408015e-02 -3.56827348e-01 6.42336011e-01 2.08220989e-01
-2.74497986e-01 -2.56301761e-02 -4.69353557e-01 2.95527130e-01
1.23630345e+00 -8.43482912e-01 -7.70318270e-01 -1.07993102e+00
-6.83763504e-01 1.88366532e-01 7.34782577e-01 5.12376964e-01
3.49384010e-01 -9.20701981e-01 -3.29956889e-01 2.50305980e-01
8.93097222e-02 -7.63755798e-01 4.01239812e-01 1.50899279e+00
-2.46674284e-01 3.56422551e-02 -2.71943033e-01 -7.79186249e-01
-1.85713160e+00 3.39778811e-01 3.14150900e-01 8.80113304e-01
-3.91678587e-02 1.24335265e+00 -5.05020134e-02 3.44614655e-01
5.56375179e-03 2.81097740e-01 -1.57507122e-01 3.46650243e-01
9.94477689e-01 8.38457584e-01 -1.60798848e-01 -1.17789221e+00
-4.60684925e-01 5.23159981e-01 -1.63381979e-01 -1.20518111e-01
9.37649190e-01 -6.10564172e-01 -1.95369452e-01 6.83222592e-01
1.53119493e+00 3.31949413e-01 -1.11162949e+00 2.33917117e-01
-2.41084263e-01 -1.08491659e+00 -1.19647384e-01 -2.55620539e-01
-7.41061509e-01 7.40887225e-01 2.59680390e-01 5.98674238e-01
8.72623265e-01 6.06522039e-02 6.30508065e-01 2.92233586e-01
9.58896801e-02 -1.24545789e+00 2.20739946e-01 1.10326856e-01
5.37311554e-01 -1.53217745e+00 1.10591747e-01 -3.40016693e-01
-9.21540082e-01 8.89081717e-01 8.59815001e-01 6.76201731e-02
7.06041873e-01 -6.98714927e-02 1.64098274e-02 -5.43940902e-01
-8.78825426e-01 3.47464710e-01 1.89935789e-01 5.43538630e-01
4.35541689e-01 -3.17316711e-01 -7.18035400e-02 4.85902697e-01
-6.11218929e-01 -1.04153983e-01 6.48735523e-01 6.63789451e-01
-1.54666215e-01 -3.39635551e-01 -4.35026437e-01 6.29822135e-01
-8.50876153e-01 2.84569323e-01 -1.05930996e+00 7.67123580e-01
-4.24573809e-01 1.46640170e+00 7.36456871e-01 -5.80591857e-01
6.93912059e-02 2.51131147e-01 5.31945467e-01 -4.71199781e-01
-8.27143908e-01 -5.20075381e-01 4.01582688e-01 -9.99930382e-01
-8.21080446e-01 -1.15878999e+00 -6.79388106e-01 -1.21350396e+00
-3.13945889e-01 -7.86812827e-02 4.35675412e-01 9.75719333e-01
4.51763600e-01 -3.78038049e-01 8.06650877e-01 -1.01755071e+00
-9.09578130e-02 -1.16581762e+00 -7.37429500e-01 5.98071933e-01
7.36394107e-01 -8.49609911e-01 -8.44868958e-01 4.93287891e-02] | [13.357603073120117, 1.9706246852874756] |
6b54320e-58ea-497c-bfe4-f2bf8709f84a | transformer-based-multimodal-information | 2203.12367 | null | https://arxiv.org/abs/2203.12367v2 | https://arxiv.org/pdf/2203.12367v2.pdf | Transformer-based Multimodal Information Fusion for Facial Expression Analysis | Human affective behavior analysis has received much attention in human-computer interaction (HCI). In this paper, we introduce our submission to the CVPR 2022 Competition on Affective Behavior Analysis in-the-wild (ABAW). To fully exploit affective knowledge from multiple views, we utilize the multimodal features of spoken words, speech prosody, and facial expression, which are extracted from the video clips in the Aff-Wild2 dataset. Based on these features, we propose a unified transformer-based multimodal framework for Action Unit detection and also expression recognition. Specifically, the static vision feature is first encoded from the current frame image. At the same time, we clip its adjacent frames by a sliding window and extract three kinds of multimodal features from the sequence of images, audio, and text. Then, we introduce a transformer-based fusion module that integrates the static vision features and the dynamic multimodal features. The cross-attention module in the fusion module makes the output integrated features focus on the crucial parts that facilitate the downstream detection tasks. We also leverage some data balancing techniques, data augmentation techniques, and postprocessing methods to further improve the model performance. In the official test of ABAW3 Competition, our model ranks first in the EXPR and AU tracks. The extensive quantitative evaluations, as well as ablation studies on the Aff-Wild2 dataset, prove the effectiveness of our proposed method. | ['Suzhen Wang', 'Yu Ding', 'Rudong An', 'Hao Zeng', 'Bowen Ma', 'Feng Qiu', 'Zhimeng Zhang', 'Wei zhang'] | 2022-03-23 | null | null | null | null | ['action-unit-detection'] | ['computer-vision'] | [ 2.23390624e-01 -3.25621605e-01 5.82775548e-02 -5.94030142e-01
-9.15908694e-01 -3.26348186e-01 4.39566731e-01 -1.52413413e-01
-5.81944823e-01 2.11180434e-01 5.01170397e-01 4.45226759e-01
2.69665331e-01 -9.04239714e-02 -1.46569923e-01 -8.63889396e-01
1.34048596e-01 -2.33351842e-01 3.59431794e-03 -1.99685529e-01
-1.57973215e-01 -3.88387293e-02 -1.55957556e+00 7.12966979e-01
4.69091177e-01 1.67213082e+00 -1.89635307e-01 6.17334723e-01
8.07076320e-02 1.03298223e+00 -1.44344732e-01 -4.26032096e-01
-1.97904676e-01 -5.03902853e-01 -4.59244043e-01 2.53560394e-01
9.35090259e-02 -3.51842970e-01 -2.50822335e-01 8.94014776e-01
7.43550003e-01 2.84034371e-01 4.90318567e-01 -1.49030161e+00
-2.04519674e-01 2.32304141e-01 -8.46299946e-01 6.59158006e-02
6.36410415e-01 1.81751266e-01 8.49010944e-01 -1.26244020e+00
4.62054133e-01 1.23271215e+00 4.60310966e-01 6.28396571e-01
-7.42361009e-01 -6.02664113e-01 3.60393703e-01 6.48513794e-01
-1.25253820e+00 -7.54160821e-01 9.19739008e-01 -3.93943906e-01
7.15541720e-01 2.90023685e-01 7.72080481e-01 1.36114085e+00
-1.54196128e-01 1.34921944e+00 8.38776946e-01 -3.33161294e-01
4.70460206e-02 4.80434299e-02 2.03815684e-01 7.82599032e-01
-8.19208026e-01 -2.89779872e-01 -8.75418007e-01 -4.96344678e-02
8.56913775e-02 1.37202397e-01 -4.40410107e-01 -6.58012107e-02
-1.10383797e+00 5.61964154e-01 1.84906423e-01 2.35264078e-01
-6.30203247e-01 -6.18465878e-02 7.68481076e-01 1.08215719e-01
4.84044343e-01 -1.45823568e-01 -1.91229075e-01 -5.50299585e-01
-4.84915316e-01 1.97017733e-02 3.81029576e-01 5.03297091e-01
4.89297003e-01 -3.26642603e-01 -6.49656355e-01 1.22397447e+00
3.62547606e-01 4.84282792e-01 5.78020036e-01 -9.64445174e-01
4.17804241e-01 7.59737551e-01 -8.02696422e-02 -1.17418146e+00
-3.93859476e-01 3.15943122e-01 -8.71898711e-01 -2.89666653e-01
-8.61943588e-02 -2.54008174e-01 -6.02798820e-01 2.00274014e+00
3.91289473e-01 7.79453889e-02 -1.11057032e-02 1.07429314e+00
8.66274953e-01 6.60528004e-01 2.74918288e-01 -6.41388774e-01
1.54360807e+00 -1.11069846e+00 -1.10337949e+00 -5.74185625e-02
4.72352624e-01 -5.59319556e-01 9.31804061e-01 4.41289544e-01
-9.57879603e-01 -5.74737132e-01 -8.75996590e-01 1.37275746e-02
-2.00606495e-01 4.13604319e-01 3.82197648e-01 3.11446875e-01
-6.24293029e-01 -9.31438953e-02 -8.35634708e-01 -3.33931774e-01
2.63196260e-01 2.69923717e-01 -5.56975543e-01 1.48801088e-01
-1.28197825e+00 5.43084085e-01 -7.85700828e-02 4.76222903e-01
-6.16692543e-01 -1.68041542e-01 -1.04850566e+00 -1.33710299e-02
3.30110520e-01 -2.20991686e-01 1.35525167e+00 -1.45461929e+00
-1.66555488e+00 9.13535655e-01 -3.90023440e-01 -1.09789684e-01
2.86925048e-01 -4.11887467e-01 -5.61616540e-01 3.81583303e-01
-2.75982141e-01 5.83520412e-01 8.42320919e-01 -8.15724671e-01
-6.81201696e-01 -6.91465735e-01 -1.24876037e-01 4.85138118e-01
-8.30230832e-01 2.69973814e-01 -1.02774620e+00 -4.68483567e-01
-2.70613402e-01 -9.03630674e-01 1.09686308e-01 -1.61975399e-01
-2.95421630e-01 -3.07955712e-01 7.93652654e-01 -6.38739288e-01
1.45758283e+00 -2.65941143e+00 4.06768084e-01 2.21188784e-01
3.76576543e-01 1.39814273e-01 -3.10888648e-01 1.50535122e-01
-7.20784590e-02 -4.15319890e-01 1.18460044e-01 -7.20491529e-01
1.01194933e-01 -1.88974321e-01 -2.52383679e-01 3.13914359e-01
2.18514323e-01 9.29977000e-01 -6.56588554e-01 -6.44955397e-01
2.36009359e-01 4.39957201e-01 -3.88294280e-01 4.78584498e-01
6.51257709e-02 3.70224386e-01 -4.88899171e-01 6.21457934e-01
3.58336151e-01 8.54096860e-02 7.35057145e-02 -6.18365765e-01
-4.12333943e-03 -2.47236207e-01 -8.50107968e-01 1.77624035e+00
-3.64517987e-01 5.53769886e-01 2.08574697e-01 -8.87578845e-01
7.87000537e-01 5.78567147e-01 7.32737243e-01 -9.47286725e-01
4.60244954e-01 -2.12933332e-01 -1.54543787e-01 -9.64424729e-01
1.03853092e-01 1.50348917e-01 -2.70878255e-01 1.72706828e-01
2.75384545e-01 3.03668290e-01 1.20345235e-01 1.87209472e-01
1.03350532e+00 1.36763379e-01 2.32409358e-01 2.89887995e-01
9.03991699e-01 -4.98598725e-01 6.92025065e-01 1.34688020e-01
-5.97853959e-01 5.64364970e-01 5.58643341e-01 -3.17994982e-01
-3.97647083e-01 -7.80195475e-01 3.45253170e-01 1.48353171e+00
1.54334068e-01 -8.00684631e-01 -8.64375055e-01 -7.06561327e-01
-3.66055250e-01 2.03617945e-01 -1.08055520e+00 -6.08531833e-01
-1.99077398e-01 -7.37848461e-01 3.46507221e-01 5.95245242e-01
6.39951110e-01 -1.19811332e+00 -6.81799054e-01 -5.90382107e-02
-6.70838416e-01 -1.37897217e+00 -7.41912723e-01 -1.08218618e-01
-6.72159120e-02 -9.30352092e-01 -6.74724638e-01 -5.06622493e-01
2.95590103e-01 9.40769240e-02 5.06709337e-01 -2.50439137e-01
-3.24101299e-01 6.89970076e-01 -5.23753226e-01 -2.56611019e-01
9.95842665e-02 -3.55712026e-01 6.31957278e-02 9.41557109e-01
5.69269240e-01 -3.98143679e-01 -5.85916340e-01 2.62717664e-01
-5.19629478e-01 1.72693521e-01 4.75066036e-01 7.86455631e-01
5.28206706e-01 -4.06811386e-01 4.62617189e-01 -3.93936455e-01
6.62626445e-01 -3.04404885e-01 -6.71835840e-02 3.42455924e-01
-3.60239185e-02 -3.82324487e-01 3.56969655e-01 -5.36130130e-01
-1.16689157e+00 3.31416994e-01 -1.39041185e-01 -8.57822359e-01
-4.54170294e-02 5.24414182e-01 -5.91883004e-01 3.21839273e-01
1.37102827e-01 1.87129840e-01 1.34346876e-02 -7.17963725e-02
2.71964550e-01 9.20993090e-01 6.82048500e-01 -2.68048227e-01
1.23329550e-01 3.92962873e-01 -4.50524986e-01 -8.61022055e-01
-9.61784542e-01 -5.58277726e-01 -3.23418379e-01 -7.37591624e-01
1.22420514e+00 -1.03945613e+00 -1.22358906e+00 7.26902902e-01
-1.06500554e+00 1.91369411e-02 1.06078684e-01 4.89975482e-01
-6.40764773e-01 2.47276634e-01 -5.98825574e-01 -1.17834353e+00
-5.21294177e-01 -1.14940667e+00 1.38879228e+00 3.16735297e-01
-5.01861572e-01 -4.90298271e-01 3.05223435e-01 5.61929286e-01
8.54195878e-02 1.57860950e-01 6.43595338e-01 -6.47252858e-01
1.80089876e-01 -2.82159925e-01 -1.55119061e-01 4.72591043e-01
1.03595331e-01 1.39455885e-01 -1.24901879e+00 1.14713788e-01
3.83524597e-03 -8.24706018e-01 9.70495403e-01 1.26260787e-01
1.26214612e+00 -3.56221981e-02 -2.03171909e-01 3.80775660e-01
7.59529173e-01 4.49253976e-01 5.81871629e-01 -1.34580240e-01
6.51988626e-01 7.98609614e-01 8.16241622e-01 6.76176667e-01
4.38156307e-01 8.04487646e-01 3.53595525e-01 -1.28173023e-01
3.43546480e-01 2.39565726e-02 7.58285880e-01 1.04333520e+00
-1.75745040e-01 -1.14900827e-01 -6.98711812e-01 3.77436548e-01
-2.24137902e+00 -1.15647125e+00 2.75312334e-01 1.91162217e+00
7.48066008e-01 -1.50566325e-01 3.13440472e-01 5.24477214e-02
6.34169817e-01 1.54375717e-01 -4.13313955e-01 -2.48376861e-01
-9.08828378e-02 -1.19052693e-01 -4.26710993e-01 3.33912462e-01
-1.36594641e+00 6.81441426e-01 5.13408852e+00 8.01387012e-01
-1.14672863e+00 1.12239271e-01 7.18327284e-01 -4.99951482e-01
4.89046164e-02 -5.95388055e-01 -4.20638829e-01 5.73415935e-01
7.47340202e-01 4.57353443e-02 2.75898933e-01 6.53133392e-01
3.90223265e-01 -1.42313972e-01 -1.07395816e+00 1.50611079e+00
4.38592434e-01 -7.20472276e-01 -1.83574542e-01 -2.14637503e-01
1.37618825e-01 -2.14260235e-01 2.19236657e-01 6.09695137e-01
-2.50298083e-01 -8.08772326e-01 6.23785496e-01 7.89113224e-01
6.20778918e-01 -9.55956221e-01 7.96419859e-01 1.92702666e-01
-1.24179173e+00 -2.39904746e-01 1.75163791e-01 1.59975007e-01
1.05455771e-01 2.74131924e-01 -2.32020155e-01 4.37116414e-01
9.08158779e-01 9.75389898e-01 -4.82681930e-01 4.81094092e-01
1.10833691e-02 4.40629631e-01 -1.34315938e-01 -5.11825271e-02
9.62548703e-02 -1.53938726e-01 2.61915416e-01 1.31991673e+00
7.59398118e-02 4.38791186e-01 1.33944005e-01 4.36479360e-01
-6.64584339e-02 2.63108015e-01 -4.78797734e-01 -2.24084526e-01
7.17804506e-02 1.72584879e+00 -1.47260308e-01 -2.95949489e-01
-7.03685045e-01 1.24767983e+00 3.48634720e-01 3.46745074e-01
-1.06881166e+00 -2.96543896e-01 7.37131894e-01 -5.34940302e-01
1.55350089e-01 1.49938792e-01 3.24276626e-01 -1.43206418e+00
1.31572112e-01 -9.47222173e-01 4.45024937e-01 -9.13565576e-01
-1.18133748e+00 7.55224764e-01 -1.96807206e-01 -1.20282412e+00
-3.94673854e-01 -4.65339065e-01 -6.54781342e-01 5.02234697e-01
-9.48424399e-01 -1.18547547e+00 -6.17334485e-01 9.24471200e-01
6.21154428e-01 -1.98487043e-01 9.37356949e-01 5.27614594e-01
-1.07471669e+00 6.39194429e-01 -2.67233640e-01 3.48611444e-01
9.93512928e-01 -7.89517999e-01 -3.02703649e-01 5.20921469e-01
1.04429685e-01 3.81548256e-01 3.27720582e-01 -2.00307459e-01
-1.38268483e+00 -8.72429788e-01 5.43168724e-01 -1.79455563e-01
6.23989522e-01 -5.17034769e-01 -6.81387603e-01 4.81979907e-01
4.56963509e-01 2.34626625e-02 8.69570255e-01 1.39197176e-02
-3.11582088e-01 -2.44284287e-01 -6.25823677e-01 5.79085767e-01
6.92018032e-01 -6.95277691e-01 -5.16099036e-01 -1.09968640e-01
4.93755758e-01 -1.22070707e-01 -6.68106675e-01 7.26689756e-01
9.68942344e-01 -1.04530537e+00 6.93011105e-01 -6.66522682e-01
5.85310698e-01 -4.38645110e-02 -4.80873942e-01 -1.09887981e+00
-4.18042466e-02 -5.59779525e-01 -9.72207189e-02 1.58185339e+00
6.70474246e-02 4.89075435e-03 4.01343584e-01 6.50765598e-01
7.11971670e-02 -9.64162350e-01 -7.75911450e-01 -7.85723552e-02
-7.43956804e-01 -5.64027309e-01 8.98360088e-03 8.02881360e-01
4.88363594e-01 8.32069278e-01 -6.74834788e-01 -3.37645233e-01
2.54460990e-01 2.16253996e-01 7.05221713e-01 -8.63113344e-01
-1.62686929e-01 -3.31094444e-01 -4.45767969e-01 -7.22137153e-01
3.90715182e-01 -4.52518970e-01 2.90091246e-01 -9.52109039e-01
6.14829600e-01 4.85382855e-01 -6.27128959e-01 6.22326791e-01
-3.59521270e-01 3.56170297e-01 3.70902151e-01 -4.86365035e-02
-1.10257089e+00 1.14658916e+00 8.82278979e-01 -2.82088012e-01
-1.39305010e-01 -2.20110640e-01 -4.39905316e-01 1.05264282e+00
4.73040789e-01 5.03307655e-02 -3.84336114e-01 -2.22849682e-01
-4.84580547e-02 1.50959298e-01 3.18371385e-01 -8.22222650e-01
1.70219034e-01 -9.79911312e-02 4.12283719e-01 -5.21278739e-01
9.45246875e-01 -8.71225715e-01 -3.97543877e-01 6.73196539e-02
-5.61317444e-01 4.08016406e-02 2.34677404e-01 4.96075094e-01
-5.26248753e-01 3.06231588e-01 7.47342765e-01 2.45348096e-01
-8.64256084e-01 3.90950471e-01 -4.96601582e-01 -4.22089100e-02
1.15511322e+00 1.93362176e-01 -1.66528791e-01 -6.17571414e-01
-1.04450810e+00 5.85681021e-01 -2.26571094e-02 6.61742091e-01
8.17813933e-01 -1.56836188e+00 -4.88261312e-01 2.66354024e-01
5.09502649e-01 -5.93943536e-01 6.83767498e-01 1.43111598e+00
2.48543113e-01 5.08751683e-02 -3.35527241e-01 -5.74932814e-01
-1.66690612e+00 5.33580780e-01 3.49929422e-01 -3.02593857e-02
-7.90607706e-02 7.97040403e-01 6.24015212e-01 -1.39993727e-02
4.69564915e-01 4.13571075e-02 -6.01748228e-01 5.94762981e-01
8.99111271e-01 7.93210045e-02 -7.75238797e-02 -8.52089286e-01
-7.07149565e-01 5.76203823e-01 1.55135468e-02 -4.35948402e-01
1.19839191e+00 -4.17672187e-01 -1.89353883e-01 8.85518432e-01
1.38544774e+00 -8.35965481e-03 -1.01411009e+00 -3.04364055e-01
-2.75736004e-01 -1.42904133e-01 -1.49596268e-02 -6.39359295e-01
-1.21648157e+00 1.21905088e+00 7.58593798e-01 -3.27534154e-02
1.53298533e+00 -1.42879477e-02 7.30178654e-01 2.96156138e-01
-1.17621645e-01 -1.24458969e+00 4.46916938e-01 3.77669603e-01
9.73293185e-01 -1.16907418e+00 -4.86302048e-01 -1.94874957e-01
-1.18838084e+00 9.15924013e-01 9.06002462e-01 2.31478408e-01
5.40405035e-01 3.53719980e-01 2.22716764e-01 -1.07113659e-01
-1.15810823e+00 -3.41749847e-01 3.95811290e-01 2.67738134e-01
3.78398657e-01 -2.27828488e-01 -6.34369776e-02 1.32228601e+00
3.63960803e-01 1.38007164e-01 -1.89201087e-02 5.60048759e-01
-2.23626375e-01 -7.11579442e-01 -1.45307496e-01 1.24876007e-01
-3.32896501e-01 6.20217100e-02 -7.61753321e-01 3.85531873e-01
8.36333707e-02 9.67170238e-01 8.69623646e-02 -9.11844850e-01
5.08207917e-01 4.86643732e-01 4.17642981e-01 -9.32637379e-02
-5.92193723e-01 5.42598009e-01 6.53558150e-02 -1.05995095e+00
-7.77197063e-01 -7.40683675e-01 -1.24086273e+00 9.85844061e-02
1.23631498e-02 1.01958312e-01 2.68344373e-01 1.03153896e+00
4.70235914e-01 5.74413717e-01 8.33776295e-01 -9.00099516e-01
-1.41807035e-01 -1.08347380e+00 -4.24166530e-01 6.03693485e-01
4.06948775e-01 -6.91952825e-01 -1.95094645e-01 2.91860998e-01] | [13.274090766906738, 4.970888137817383] |
79f3a088-03b6-4b13-808a-7ddc1df3d154 | parameter-efficient-local-implicit-image | 2303.15122 | null | https://arxiv.org/abs/2303.15122v1 | https://arxiv.org/pdf/2303.15122v1.pdf | Parameter Efficient Local Implicit Image Function Network for Face Segmentation | Face parsing is defined as the per-pixel labeling of images containing human faces. The labels are defined to identify key facial regions like eyes, lips, nose, hair, etc. In this work, we make use of the structural consistency of the human face to propose a lightweight face-parsing method using a Local Implicit Function network, FP-LIIF. We propose a simple architecture having a convolutional encoder and a pixel MLP decoder that uses 1/26th number of parameters compared to the state-of-the-art models and yet matches or outperforms state-of-the-art models on multiple datasets, like CelebAMask-HQ and LaPa. We do not use any pretraining, and compared to other works, our network can also generate segmentation at different resolutions without any changes in the input resolution. This work enables the use of facial segmentation on low-compute or low-bandwidth devices because of its higher FPS and smaller model size. | ['Balaji Krishnamurthy', 'Rishabh Jain', 'Mayur Hemani', 'Nikitha SR', 'Mausoom Sarkar'] | 2023-03-27 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Sarkar_Parameter_Efficient_Local_Implicit_Image_Function_Network_for_Face_Segmentation_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Sarkar_Parameter_Efficient_Local_Implicit_Image_Function_Network_for_Face_Segmentation_CVPR_2023_paper.pdf | cvpr-2023-1 | ['face-parsing'] | ['computer-vision'] | [ 2.66793847e-01 6.68677747e-01 5.06350026e-02 -5.93023360e-01
-3.10423881e-01 -2.61338234e-01 2.51677901e-01 -5.61697423e-01
-3.31578583e-01 5.22655249e-01 -3.90868038e-01 -1.45179734e-01
4.82987195e-01 -9.13403749e-01 -8.82337630e-01 -4.15781409e-01
2.24604964e-01 4.79321420e-01 6.95176542e-01 9.67030674e-02
-7.38402531e-02 7.46358931e-01 -1.86383343e+00 5.38379312e-01
1.90111354e-01 1.28761637e+00 2.29045544e-02 6.89470768e-01
-2.45379746e-01 4.51623082e-01 -3.40823591e-01 -6.04819238e-01
2.55695432e-01 -2.28161082e-01 -8.47327292e-01 1.60982653e-01
8.62793028e-01 -7.22635865e-01 1.05002420e-02 8.49383593e-01
5.26068091e-01 -4.27073538e-01 3.58481675e-01 -1.07754111e+00
-2.36679196e-01 3.51158977e-01 -6.34758115e-01 -2.03952104e-01
2.93815345e-01 -4.65730354e-02 3.49500597e-01 -7.48037398e-01
6.55134380e-01 1.59276652e+00 8.43510211e-01 1.02671039e+00
-9.76908922e-01 -8.93527985e-01 7.29854498e-03 -1.95609540e-01
-1.31666970e+00 -8.17039132e-01 4.18827116e-01 -2.30518207e-01
8.44434142e-01 2.76063159e-02 2.81894803e-01 9.16676283e-01
-5.13165966e-02 1.89407751e-01 1.11529267e+00 -5.68569303e-01
3.14909257e-02 1.00376785e-01 -1.24014072e-01 1.36307061e+00
2.96910554e-02 -2.22074479e-01 -3.67538929e-01 -1.90756042e-02
1.11648750e+00 -1.68812707e-01 4.79219705e-02 6.86433092e-02
-5.37865818e-01 5.69896579e-01 1.76490352e-01 1.51938379e-01
-1.14424959e-01 4.32543904e-01 2.53257364e-01 -1.95400372e-01
3.62045914e-01 -3.08856815e-01 -5.82102776e-01 1.15983807e-01
-1.19854009e+00 -2.18766853e-01 7.40200996e-01 8.78278911e-01
1.11944544e+00 -1.08242027e-01 1.86913371e-01 5.84328473e-01
5.39286613e-01 4.34132904e-01 2.54905194e-01 -1.31117451e+00
2.03296304e-01 5.94572365e-01 -2.15174064e-01 -6.84121728e-01
-5.81101596e-01 2.10781172e-01 -5.57826936e-01 5.21434903e-01
3.33352715e-01 -3.75940263e-01 -1.36075258e+00 1.60211217e+00
7.36078918e-01 5.23001015e-01 -1.11805715e-01 6.60206318e-01
1.13856447e+00 5.69456935e-01 2.44986743e-01 -3.46275158e-02
1.68528712e+00 -9.85973477e-01 -4.98034656e-01 -1.13599412e-01
2.59404659e-01 -8.54605973e-01 9.10646200e-01 3.33454192e-01
-1.23924434e+00 -8.45251381e-01 -8.27011645e-01 -2.60977238e-01
-4.43062961e-01 3.75194162e-01 6.34422839e-01 1.14967501e+00
-1.42322218e+00 6.50624633e-01 -8.87054145e-01 -4.28872615e-01
6.01257026e-01 8.09715867e-01 -6.54430807e-01 2.50888258e-01
-5.90421140e-01 4.23187941e-01 2.79614955e-01 5.01159318e-02
-7.87000835e-01 -5.58359981e-01 -8.18025112e-01 4.66958955e-02
2.42747426e-01 -5.91466844e-01 1.26910675e+00 -1.29580283e+00
-2.09156036e+00 1.17894208e+00 -4.09914643e-01 -2.19848990e-01
3.65594983e-01 -1.56713754e-01 -2.99624592e-01 5.74585497e-01
-1.88345432e-01 1.38419890e+00 1.09092748e+00 -1.04802966e+00
-5.26834011e-01 -2.70323336e-01 9.69389156e-02 -3.75736505e-01
-3.66995707e-02 3.89276385e-01 -8.34060609e-01 -1.38000920e-01
6.14719689e-02 -9.02849913e-01 1.18104838e-01 2.38655493e-01
-2.64609665e-01 -1.67945206e-01 1.19301152e+00 -8.41295719e-01
7.61004448e-01 -2.21985126e+00 -3.79965186e-01 1.21619485e-01
-1.52226537e-01 5.44556677e-01 -1.41610250e-01 -1.02252088e-01
1.37872007e-02 2.83489615e-01 -2.34320372e-01 -6.96457088e-01
-2.03890726e-01 2.71738291e-01 1.15645677e-01 3.82928044e-01
3.74425530e-01 5.82464993e-01 -1.28540784e-01 -8.94004226e-01
1.80552289e-01 9.27949727e-01 -5.64687073e-01 7.07599744e-02
-3.35684359e-01 2.55551189e-01 -3.25557381e-01 8.22566628e-01
1.17218018e+00 -1.29169837e-01 2.89331555e-01 -3.13516647e-01
-9.41787362e-02 9.92186442e-02 -1.28084600e+00 1.64652646e+00
-4.39424574e-01 3.21144313e-01 6.41201854e-01 -6.43300056e-01
7.71248460e-01 5.45089364e-01 1.55974165e-01 -6.98506713e-01
3.37928176e-01 2.12014705e-01 -2.76021123e-01 -4.19905037e-01
6.52797446e-02 1.18995845e-01 3.72227788e-01 3.20781916e-01
3.21935654e-01 1.86704725e-01 1.31972834e-01 -3.10546994e-01
7.15432763e-01 5.25676370e-01 -8.12169816e-03 -2.94755489e-01
7.78042197e-01 -5.04498839e-01 6.61491215e-01 1.47133708e-01
-7.21387193e-02 7.21842647e-01 6.60653412e-01 -4.74125713e-01
-8.74766767e-01 -7.03417540e-01 -3.15201640e-01 1.13431120e+00
-5.43143377e-02 -3.42778325e-01 -1.65584874e+00 -6.99914932e-01
-2.37211332e-01 4.20996808e-02 -5.21192014e-01 6.12429559e-01
-8.54934931e-01 -4.12361175e-01 7.69392729e-01 5.02538025e-01
8.13331544e-01 -1.25418663e+00 -8.39985728e-01 -9.13897157e-03
3.34828556e-01 -1.54168403e+00 -3.05501044e-01 -9.91487876e-02
-8.44732583e-01 -1.10980868e+00 -4.80277151e-01 -1.13893938e+00
7.57908106e-01 -3.57500434e-01 1.05696189e+00 1.81248277e-01
-6.10863566e-01 2.11519018e-01 -5.15875928e-02 -1.77915230e-01
-3.18265676e-01 6.48833588e-02 -1.80960864e-01 1.29561812e-01
1.53614551e-01 -4.98585999e-01 -7.04295099e-01 3.96782249e-01
-7.46804595e-01 1.84284016e-01 4.58749592e-01 4.24539685e-01
7.93568671e-01 -1.76159263e-01 3.29115301e-01 -1.21799934e+00
-1.18201815e-01 -5.37377298e-02 -8.41214538e-01 2.85787523e-01
-3.25967729e-01 -6.16928143e-03 4.97926146e-01 -2.20001280e-01
-1.13616312e+00 5.72924197e-01 -4.46680576e-01 -2.11595029e-01
-4.63268697e-01 -3.58754218e-01 -4.12744194e-01 -3.93434346e-01
1.33770481e-01 -2.44382218e-01 4.55432869e-02 -6.81490541e-01
3.08301091e-01 7.92772055e-01 5.10847509e-01 -5.58632314e-01
2.73244172e-01 6.03294313e-01 3.15964252e-01 -8.86122704e-01
-4.04056758e-01 -7.75641650e-02 -8.53188217e-01 -1.03222035e-01
1.14009309e+00 -8.90988946e-01 -1.10929251e+00 5.27202070e-01
-1.45114374e+00 -4.46320921e-01 7.30495453e-02 -5.87283866e-03
-5.28912723e-01 2.24301636e-01 -9.33153212e-01 -7.35982418e-01
-7.03968287e-01 -1.30123997e+00 1.39322495e+00 6.87382519e-01
1.29145920e-01 -6.10645235e-01 -4.57430512e-01 4.24988657e-01
3.77999485e-01 3.68754029e-01 6.44807875e-01 -9.63227600e-02
-6.19114876e-01 2.32575789e-01 -3.97789508e-01 4.10612524e-01
-1.01196408e-01 4.29575920e-01 -1.41756582e+00 -8.26920941e-02
-2.45824739e-01 -2.53877938e-01 7.00680137e-01 4.59053546e-01
1.44049513e+00 -3.52254033e-01 -2.91623473e-01 8.42351496e-01
1.43851113e+00 1.06879368e-01 8.51985335e-01 -4.60998453e-02
5.06152987e-01 7.66054928e-01 1.92652971e-01 2.58230358e-01
3.19624275e-01 6.68208599e-01 5.88906229e-01 -3.77168655e-01
-4.93705153e-01 -8.56133401e-02 4.54815537e-01 3.52566093e-01
-1.23481505e-01 1.64871998e-02 -7.55643070e-01 2.65648901e-01
-1.50123441e+00 -7.25148797e-01 -1.48395121e-01 2.00269151e+00
7.82521725e-01 8.95877555e-03 1.74066707e-01 -1.40605703e-01
8.78211915e-01 -1.05150096e-01 -2.66231686e-01 -8.45927358e-01
1.56434193e-01 9.29425359e-01 5.86122632e-01 5.88957608e-01
-1.29227412e+00 1.18388355e+00 6.69969559e+00 7.57907689e-01
-1.39178622e+00 4.21217889e-01 1.03374898e+00 -4.35863696e-02
3.14333528e-01 -1.72454536e-01 -1.22358620e+00 4.47844833e-01
1.19799650e+00 7.98405468e-01 3.67172480e-01 9.77837741e-01
-1.01846509e-01 -2.07822099e-01 -1.04886270e+00 9.83348072e-01
1.57515734e-01 -1.11996984e+00 -1.68626815e-01 -2.22669356e-02
4.15908098e-01 -1.06368996e-01 -5.88832125e-02 -6.06013536e-02
-1.17086947e-01 -1.35858214e+00 6.82799101e-01 4.65022266e-01
1.22181880e+00 -7.96341360e-01 6.90911829e-01 1.19188741e-01
-1.23566020e+00 1.98416144e-01 -4.45848495e-01 2.25612774e-01
4.16882373e-02 3.00066829e-01 -8.28731060e-01 1.44973278e-01
9.28980052e-01 1.17186168e-02 -6.50704682e-01 5.15541494e-01
-1.46631867e-01 5.88455975e-01 -6.46317720e-01 4.16259199e-01
3.21201272e-02 -1.73900109e-02 -3.52576524e-01 1.38074350e+00
4.80872601e-01 1.04901642e-01 1.43045142e-01 6.52162075e-01
-3.00664514e-01 6.47493601e-02 -1.19829521e-01 1.89969629e-01
2.15114340e-01 1.51026058e+00 -9.64386761e-01 -2.80489683e-01
-6.04136646e-01 1.09514546e+00 4.63914461e-02 -4.59005721e-02
-9.14127171e-01 -1.57580405e-01 5.18945336e-01 4.73923117e-01
5.75886250e-01 3.84428091e-02 -9.81609300e-02 -5.60690820e-01
-2.29958794e-03 -7.12190330e-01 1.26253650e-01 -6.63737297e-01
-7.41534293e-01 8.54423702e-01 -2.17406511e-01 -5.44688642e-01
-1.97737336e-01 -9.73921835e-01 -3.22381854e-01 8.40789795e-01
-1.65539742e+00 -1.63317335e+00 -4.25320476e-01 5.74817061e-01
3.92458498e-01 -4.90413718e-02 1.02205920e+00 4.59972560e-01
-6.42069995e-01 6.97699964e-01 -5.12373626e-01 4.08386976e-01
6.16515815e-01 -8.11311245e-01 5.00653625e-01 7.06600666e-01
-1.84257507e-01 3.45965058e-01 1.58968106e-01 -4.69663620e-01
-1.09236157e+00 -1.07170999e+00 6.93837643e-01 -1.22303283e-02
-5.74203990e-02 -5.94068825e-01 -6.34714603e-01 6.96101606e-01
4.11636263e-01 4.58589107e-01 5.47613382e-01 -2.86816955e-01
-2.89131612e-01 -4.30154592e-01 -1.64936423e+00 3.22526932e-01
1.07302678e+00 -3.94785464e-01 -7.56921060e-03 8.69560838e-02
6.24907434e-01 -5.04135191e-01 -6.97751522e-01 4.49745536e-01
7.25629747e-01 -1.28740823e+00 8.53794813e-01 -1.15867376e-01
3.00717026e-01 -3.13936293e-01 -1.09959051e-01 -4.05004025e-01
-1.35490254e-01 -6.69945419e-01 -6.75267801e-02 1.57374191e+00
3.89650971e-01 -6.76874816e-01 1.09952724e+00 6.45914257e-01
-6.36560516e-03 -5.79847872e-01 -1.23825121e+00 -2.62531966e-01
-1.99796692e-01 3.32854763e-02 8.16802680e-01 4.63120371e-01
-4.51982021e-01 4.20183726e-02 -1.05383322e-01 3.68069828e-01
5.75074971e-01 -1.81621574e-02 6.84026539e-01 -1.17746139e+00
-2.02914506e-01 -1.65889680e-01 -4.95884478e-01 -7.30363429e-01
3.22030038e-01 -4.90998268e-01 -1.32638648e-01 -1.15436530e+00
-7.18701109e-02 -4.79622394e-01 1.31754443e-01 1.05636168e+00
3.71397197e-01 7.37436533e-01 1.68400183e-01 -2.13147417e-01
-4.27795649e-01 -2.27100980e-02 1.07617867e+00 2.03408569e-01
1.12909198e-01 -3.27088803e-01 -3.44004303e-01 9.80852246e-01
7.31321275e-01 -5.02466679e-01 -1.87051922e-01 -4.98469919e-01
-1.57149300e-01 1.68644711e-02 2.82281190e-01 -1.28248537e+00
1.37670606e-01 9.86187607e-02 6.26428306e-01 -3.87393624e-01
5.11327147e-01 -9.82638657e-01 3.89547795e-01 5.20214140e-01
2.47255728e-01 4.93044518e-02 3.99118781e-01 -2.30076555e-02
-5.08891791e-02 -4.95623797e-01 1.18785155e+00 -3.65125984e-01
-7.86675990e-01 2.26593688e-01 -3.18052396e-02 -3.22685987e-01
1.22480631e+00 -3.25909108e-01 -3.36924255e-01 1.81345373e-01
-6.65025651e-01 -2.85389036e-01 5.88491559e-01 1.56275153e-01
3.94696176e-01 -8.00749600e-01 -5.30749559e-01 6.80234313e-01
-5.09832263e-01 2.06751883e-01 2.36521706e-01 4.92856354e-01
-1.17287648e+00 4.60375458e-01 -2.90253162e-01 -6.29128337e-01
-1.58941400e+00 2.49828085e-01 5.42608917e-01 6.66124150e-02
-5.54994106e-01 9.16439056e-01 2.26760194e-01 -3.77157509e-01
1.30029634e-01 -2.74361670e-01 -8.46374258e-02 -7.36451820e-02
6.50757849e-01 2.49154657e-01 2.68423289e-01 -8.03080082e-01
-5.93571961e-01 1.00963664e+00 1.31385729e-01 2.12358683e-02
1.13210177e+00 8.51199105e-02 -3.52128506e-01 -1.41579613e-01
1.10308516e+00 2.81209778e-02 -1.51668096e+00 1.87628001e-01
-2.40041927e-01 -3.01276416e-01 8.34558606e-02 -7.56330431e-01
-1.40170205e+00 9.28263128e-01 1.08742118e+00 -1.35914415e-01
1.34020519e+00 -8.80105048e-03 7.44363248e-01 -3.01441867e-02
5.67071855e-01 -1.10820472e+00 -2.26569414e-01 3.29842001e-01
5.16729057e-01 -1.14816189e+00 -8.05002153e-02 -9.20254827e-01
-3.90237212e-01 1.38388515e+00 7.56131053e-01 -7.63867944e-02
7.77443051e-01 7.50025630e-01 1.01884775e-01 4.06943224e-02
-4.22680587e-01 -2.51549274e-01 1.92951616e-02 6.18116975e-01
5.67281306e-01 3.84105146e-02 -2.11526588e-01 4.03621167e-01
-1.48163795e-01 2.18479648e-01 1.94442272e-01 6.52255058e-01
-3.71193856e-01 -1.35243499e+00 -3.65077943e-01 1.10628016e-01
-8.99790525e-01 9.39283241e-03 -1.87173635e-01 8.52309287e-01
7.70478904e-01 8.94785345e-01 4.44852591e-01 8.24804828e-02
-2.58744732e-02 3.65647435e-01 6.63254261e-01 -6.33473516e-01
-7.40470946e-01 1.20192140e-01 7.65902549e-02 -8.43397379e-01
-6.12504542e-01 -3.90425205e-01 -1.54952550e+00 -4.44497377e-01
-1.06051631e-01 -2.81285524e-01 1.03328693e+00 6.74962640e-01
7.01572239e-01 3.16217124e-01 3.06718796e-01 -1.02148902e+00
1.03094243e-01 -9.45779443e-01 -5.79899073e-01 1.64389417e-01
1.02737159e-01 -4.77069765e-01 1.10085234e-02 3.99064183e-01] | [13.446515083312988, 0.6027265191078186] |
5304c899-9643-4c3c-8a1b-e053bc411c74 | network-of-experts-for-large-scale-image | 1604.06119 | null | http://arxiv.org/abs/1604.06119v3 | http://arxiv.org/pdf/1604.06119v3.pdf | Network of Experts for Large-Scale Image Categorization | We present a tree-structured network architecture for large scale image
classification. The trunk of the network contains convolutional layers
optimized over all classes. At a given depth, the trunk splits into separate
branches, each dedicated to discriminate a different subset of classes. Each
branch acts as an expert classifying a set of categories that are difficult to
tell apart, while the trunk provides common knowledge to all experts in the
form of shared features. The training of our "network of experts" is completely
end-to-end: the partition of categories into disjoint subsets is learned
simultaneously with the parameters of the network trunk and the experts are
trained jointly by minimizing a single learning objective over all classes. The
proposed structure can be built from any existing convolutional neural network
(CNN). We demonstrate its generality by adapting 4 popular CNNs for image
categorization into the form of networks of experts. Our experiments on
CIFAR100 and ImageNet show that in every case our method yields a substantial
improvement in accuracy over the base CNN, and gives the best result achieved
so far on CIFAR100. Finally, the improvement in accuracy comes at little
additional cost: compared to the base network, the training time is only
moderately increased and the number of parameters is comparable or in some
cases even lower. | ['Karim Ahmed', 'Mohammad Haris Baig', 'Lorenzo Torresani'] | 2016-04-20 | null | null | null | null | ['image-categorization'] | ['computer-vision'] | [-1.13145605e-01 3.24556142e-01 1.54052051e-02 -5.48650861e-01
-6.00957870e-01 -8.66289794e-01 1.53933540e-01 4.77470718e-02
-5.68559706e-01 4.76931661e-01 -2.40360111e-01 -2.51115978e-01
-1.22473523e-01 -7.79335082e-01 -6.13933802e-01 -6.73287153e-01
-1.52026072e-01 7.46588051e-01 6.14318669e-01 4.89140041e-02
1.45412728e-01 6.12755477e-01 -1.48473120e+00 5.64292729e-01
6.15963995e-01 1.68705225e+00 9.42155346e-02 6.78411543e-01
-2.34414153e-02 1.07141411e+00 -6.37028813e-01 -4.27536339e-01
4.59661603e-01 -4.77434881e-02 -1.30348504e+00 2.71255434e-01
6.28174603e-01 -1.33538514e-01 -2.07731426e-01 9.80974317e-01
3.09686095e-01 2.27981363e-03 6.03297114e-01 -1.18845820e+00
-4.35746700e-01 5.71063459e-01 -1.67161644e-01 9.08548608e-02
-4.13404047e-01 -4.55041714e-02 1.18481612e+00 -6.17513478e-01
2.49384016e-01 1.20436192e+00 8.70802104e-01 5.10993481e-01
-1.38833141e+00 -4.84989733e-01 5.99831641e-01 3.06230355e-02
-1.27491677e+00 -2.55271018e-01 2.81593800e-01 -6.33066356e-01
8.01848054e-01 6.62393644e-02 2.97842413e-01 5.50957918e-01
-9.07523558e-02 6.57238603e-01 7.21790493e-01 -6.39955401e-02
3.92245531e-01 3.37346792e-01 6.58535242e-01 8.18285823e-01
1.33105978e-01 -3.26576442e-01 1.97354257e-01 -3.45954001e-02
7.83565938e-01 -3.14605189e-04 -1.10652596e-01 -5.86086571e-01
-1.07483602e+00 8.80821705e-01 9.48135853e-01 3.24619740e-01
-4.42828089e-02 1.61851332e-01 3.85218978e-01 6.85967386e-01
5.06330311e-01 5.48443377e-01 -7.14811325e-01 6.69091046e-01
-8.12560558e-01 8.69340971e-02 1.10130060e+00 6.60287380e-01
9.50492680e-01 -8.15846846e-02 -6.56193718e-02 9.88952875e-01
5.99016882e-02 8.47183466e-02 5.53065002e-01 -1.25820541e+00
3.99753690e-01 8.63988340e-01 1.81179151e-01 -8.96173954e-01
-5.80429971e-01 -1.02690196e+00 -8.96642208e-01 4.35251653e-01
6.43371344e-01 -2.72029996e-01 -1.07276702e+00 1.70203161e+00
6.86822236e-02 4.94865514e-02 -6.61740229e-02 7.73267746e-01
7.88282990e-01 4.02429312e-01 7.39793256e-02 5.12212396e-01
1.26920259e+00 -1.28735352e+00 1.22591723e-02 -5.03591418e-01
4.34899926e-01 -5.90846598e-01 6.83496237e-01 5.17185569e-01
-9.99161541e-01 -8.34612608e-01 -9.52965617e-01 1.00817950e-02
-6.78726971e-01 5.01521349e-01 5.04131436e-01 4.71020848e-01
-1.40490580e+00 6.86038911e-01 -4.97171104e-01 -1.11988872e-01
7.21762359e-01 7.73066580e-01 -3.84113222e-01 8.63488019e-02
-9.44530249e-01 8.84538054e-01 5.55281937e-01 2.71269791e-02
-1.20140898e+00 -5.92900753e-01 -7.05294073e-01 4.76358443e-01
1.95236616e-02 -5.84121883e-01 1.44632840e+00 -1.52530742e+00
-1.27440071e+00 1.17795932e+00 4.11319211e-02 -5.72164476e-01
5.56107283e-01 1.03768684e-01 -9.74117666e-02 2.52807140e-01
2.26406470e-01 1.01362669e+00 7.06760526e-01 -1.32992268e+00
-9.54820454e-01 -1.44877687e-01 5.16677022e-01 -2.51177829e-02
-4.40665722e-01 9.88571197e-02 -5.73418319e-01 -5.78265429e-01
4.18443643e-02 -9.38355505e-01 -4.82064098e-01 2.08528161e-01
-3.44690442e-01 -4.01023954e-01 5.40486753e-01 -3.30852479e-01
9.50409472e-01 -2.17923403e+00 9.99804810e-02 2.17646867e-01
4.55153763e-01 2.75966495e-01 -3.88714492e-01 -5.54309040e-03
-3.82770509e-01 1.04318760e-01 -3.08935106e-01 -4.14053798e-01
-1.39962435e-01 1.91269949e-01 -2.55693913e-01 3.59153748e-01
1.42784655e-01 7.61686504e-01 -6.77360713e-01 -1.63351998e-01
7.03056082e-02 1.43161923e-01 -5.28114080e-01 2.67493129e-01
-1.48512021e-01 1.08730398e-01 -3.39868307e-01 2.61434555e-01
6.43945813e-01 -7.44103670e-01 6.14735670e-02 -4.12715822e-02
2.93670781e-02 2.47932792e-01 -1.27526581e+00 1.33881319e+00
-5.30542910e-01 6.11755669e-01 3.62527400e-01 -1.55386090e+00
7.32055426e-01 3.39871347e-01 1.14501283e-01 -1.18454307e-01
3.13920885e-01 2.33288035e-01 2.62208760e-01 -1.08196519e-01
-2.37700697e-02 1.26440078e-01 -9.51654986e-02 4.59326893e-01
4.84146297e-01 3.12155508e-03 2.60402381e-01 1.38140857e-01
9.02606070e-01 -5.00103533e-01 6.25113398e-02 -5.47257960e-01
5.76023698e-01 8.51490721e-02 3.19185525e-01 8.42002213e-01
-7.83656165e-02 5.49411714e-01 6.55305922e-01 -9.08448756e-01
-9.42759573e-01 -9.72913444e-01 -1.15394890e-01 1.35366833e+00
1.17748335e-01 -8.91875550e-02 -8.75005841e-01 -9.55848813e-01
1.02260217e-01 6.06748760e-02 -8.73063445e-01 4.56551643e-04
-4.77681071e-01 -6.32880032e-01 4.61370528e-01 7.54141629e-01
7.45423794e-01 -9.72682357e-01 -5.16892850e-01 5.92059195e-02
-6.27489761e-02 -1.09561932e+00 -1.45436391e-01 5.66893458e-01
-9.18827593e-01 -1.42312741e+00 -6.83235228e-01 -1.26999378e+00
9.60762382e-01 3.08721721e-01 1.35074294e+00 2.48774886e-01
-1.94556430e-01 5.88457547e-02 -6.12260960e-02 -3.22198272e-01
-1.87343031e-01 3.69191110e-01 -3.07956517e-01 3.22098881e-01
6.73516691e-02 -4.90958333e-01 -6.70259178e-01 6.62202895e-01
-6.35280907e-01 -3.82824540e-01 5.27451813e-01 8.24203908e-01
2.78556198e-01 2.45230138e-01 3.73037338e-01 -1.04716253e+00
2.12736517e-01 -6.53158844e-01 -6.18355811e-01 1.48340657e-01
-7.76795000e-02 -3.95432785e-02 1.04428005e+00 -5.06651580e-01
-7.24888802e-01 1.92763761e-01 -6.48201704e-02 -3.38044673e-01
-5.03409326e-01 2.24282846e-01 -1.24092698e-01 -3.56024384e-01
6.28450751e-01 -1.03285804e-01 -1.59016103e-01 -7.20234275e-01
2.29884967e-01 7.87855148e-01 5.60338080e-01 -3.82102162e-01
5.24585903e-01 4.52646405e-01 -3.44032586e-01 -5.15236557e-01
-1.25793791e+00 -5.03361106e-01 -8.81972134e-01 1.28912181e-01
8.46509933e-01 -1.06619954e+00 -6.55211210e-01 6.83129311e-01
-1.08467591e+00 -6.67954624e-01 -2.83460408e-01 1.56201065e-01
-2.93609977e-01 4.41618674e-02 -7.79189408e-01 -1.47195354e-01
-1.25675559e-01 -1.12773681e+00 7.61312783e-01 2.13164300e-01
-2.90654190e-02 -1.17598915e+00 -3.21093202e-01 2.37944245e-01
6.01683676e-01 7.31279375e-03 1.07372808e+00 -1.10583055e+00
-5.24057448e-01 -4.32093471e-01 -4.82552081e-01 8.48714590e-01
4.17353027e-02 -2.98399985e-01 -1.17841220e+00 -4.66370046e-01
-1.39949277e-01 -7.07214534e-01 1.27591181e+00 2.53714144e-01
1.66397750e+00 -1.28323138e-01 -3.80205959e-01 5.74549556e-01
1.36742163e+00 1.11980647e-01 3.66734385e-01 2.36382812e-01
6.01141334e-01 5.44615984e-01 -7.18155056e-02 2.78128847e-03
1.42903417e-01 3.78144920e-01 6.28813684e-01 -4.42616791e-01
-1.36912063e-01 2.98626006e-01 -9.56470370e-02 2.76975036e-01
5.56768756e-03 -3.01231712e-01 -9.39040720e-01 7.47636497e-01
-1.80593228e+00 -6.86514616e-01 2.05135450e-01 2.22478056e+00
5.56259036e-01 2.92526484e-01 1.26890644e-01 1.99455529e-01
7.88097918e-01 3.38260569e-02 -5.72929263e-01 -3.14818770e-01
-1.03283115e-02 3.34018916e-01 4.22021985e-01 6.03435397e-01
-1.41825473e+00 8.50226760e-01 7.23221970e+00 8.87722909e-01
-1.22269535e+00 4.18437459e-03 9.86808717e-01 1.39478773e-01
1.69573501e-01 -9.93387401e-02 -7.45281458e-01 2.71626741e-01
7.76026964e-01 1.12586550e-01 1.94808751e-01 1.10525787e+00
-2.72026122e-01 4.82172035e-02 -1.22500312e+00 6.50459945e-01
-1.10342421e-01 -1.24276304e+00 7.67615587e-02 -9.63491201e-02
6.41420901e-01 4.48257327e-01 -2.57094037e-02 2.69466072e-01
8.35713923e-01 -1.05612731e+00 8.71665418e-01 7.44860843e-02
7.94091344e-01 -8.50599170e-01 7.48689353e-01 5.49315631e-01
-1.14929163e+00 -5.69634795e-01 -6.67741835e-01 -2.19635457e-01
-4.06870693e-01 4.44152385e-01 -4.26974356e-01 1.82443291e-01
7.61480212e-01 6.11954153e-01 -6.25952482e-01 1.22654521e+00
-2.78681397e-01 5.84655464e-01 -2.09402993e-01 2.65490830e-01
6.54668808e-01 1.68773279e-01 1.44600466e-01 1.30219531e+00
-1.70806676e-01 1.22417798e-02 5.80922842e-01 5.45849204e-01
-4.35835272e-01 -2.69997507e-01 -2.77481079e-01 2.88343936e-01
3.02711278e-01 1.43006277e+00 -1.03614092e+00 -7.01101363e-01
-4.15176660e-01 1.03393435e+00 8.35524321e-01 3.51938993e-01
-5.34141064e-01 -7.99275875e-01 7.07718611e-01 -9.67172533e-02
6.59923255e-01 1.15568958e-01 -2.48281837e-01 -1.25009513e+00
-1.55892298e-01 -8.12216401e-01 6.85493588e-01 -4.41423774e-01
-1.42894197e+00 1.03522289e+00 -2.27456212e-01 -1.01249719e+00
5.96700534e-02 -1.16333556e+00 -7.23535597e-01 9.34598565e-01
-1.46914601e+00 -9.48583484e-01 -3.73776346e-01 7.47169256e-01
5.88503420e-01 -2.53840327e-01 9.89531517e-01 2.31989771e-01
-5.52583933e-01 6.79354250e-01 2.03228086e-01 6.43602967e-01
4.43305731e-01 -1.34538043e+00 3.03511351e-01 6.89031005e-01
1.66010689e-02 5.21619081e-01 2.78932571e-01 -8.16456601e-02
-3.65205467e-01 -1.25230610e+00 9.57887948e-01 -3.74962956e-01
5.85735261e-01 -5.55529535e-01 -8.47447395e-01 7.77650595e-01
1.43877864e-01 2.86599457e-01 6.16698980e-01 2.03042999e-01
-8.20241034e-01 -3.81564021e-01 -1.17262292e+00 1.39789715e-01
7.51065791e-01 -4.60865051e-01 -5.36309958e-01 3.72158498e-01
5.35825312e-01 -2.17816502e-01 -6.49699032e-01 1.58866435e-01
4.90699738e-01 -1.06598365e+00 1.05506170e+00 -8.08536708e-01
4.94811594e-01 -2.80156493e-01 -1.73310693e-02 -1.47704887e+00
-7.50342071e-01 -1.07563019e-01 2.83333689e-01 8.81107926e-01
6.00856602e-01 -7.92535543e-01 8.51278007e-01 5.30317843e-01
-1.32873774e-01 -7.20026493e-01 -6.75108433e-01 -8.11504960e-01
2.95276523e-01 -2.24473223e-01 4.68899637e-01 8.22004855e-01
-3.65054727e-01 5.61273992e-01 -6.28280938e-02 2.45108992e-01
5.92673957e-01 4.09234494e-01 4.74988520e-01 -1.52467990e+00
-4.57920820e-01 -6.88898802e-01 -3.48065406e-01 -1.26390898e+00
2.48968273e-01 -9.58746433e-01 3.44718620e-02 -1.44212019e+00
2.83701688e-01 -8.12048316e-01 -5.84810555e-01 8.83527040e-01
-3.77421491e-02 4.79934722e-01 3.07699770e-01 3.06842983e-01
-7.45181978e-01 -5.03825247e-02 1.02354980e+00 -2.67346174e-01
1.13168105e-01 3.71624708e-01 -9.00205135e-01 1.03311443e+00
8.14548314e-01 -3.85219127e-01 -4.53236312e-01 -8.64937365e-01
-6.49292246e-02 -3.28420341e-01 4.50965434e-01 -1.25192237e+00
3.09961885e-01 2.91859120e-01 6.08601749e-01 -1.93825677e-01
4.55555841e-02 -8.66425037e-01 -1.38347700e-01 5.39358199e-01
-5.54749489e-01 -2.48517796e-01 1.83741331e-01 3.62125218e-01
-4.35221702e-01 -1.97200000e-01 1.30255985e+00 -5.13803840e-01
-6.82201624e-01 4.80732381e-01 -3.22574675e-01 -1.15149412e-02
8.03165495e-01 -1.80023462e-01 -2.53494203e-01 -1.74182579e-01
-9.67812359e-01 4.52372342e-01 2.89717436e-01 2.22764447e-01
1.75816596e-01 -9.80840325e-01 -5.05740821e-01 1.10259794e-01
-8.91686827e-02 1.73345253e-01 5.01279384e-02 4.30264592e-01
-4.11668628e-01 6.11852348e-01 -2.75190622e-01 -5.73275924e-01
-1.07108045e+00 6.43794477e-01 8.74815464e-01 -4.81748432e-01
-1.48050025e-01 1.18749487e+00 7.48911917e-01 -6.66597784e-01
6.07007742e-01 -3.30444515e-01 -4.61369604e-01 2.69342989e-01
5.97166002e-01 2.15417802e-01 1.60827607e-01 -4.99678463e-01
-3.14226925e-01 6.63548946e-01 -1.71588212e-01 2.85309196e-01
1.50565672e+00 -1.17376624e-02 -2.90556997e-01 2.99982071e-01
1.36471295e+00 -4.17106599e-01 -1.32263505e+00 -3.13813835e-01
-5.40471151e-02 -1.43917799e-01 -5.77295311e-02 -1.12017632e+00
-1.39725387e+00 1.11561072e+00 3.74872297e-01 3.37074846e-01
1.21279669e+00 1.56864867e-01 4.50152755e-01 6.25532269e-01
4.07652467e-01 -7.90149808e-01 1.45309299e-01 8.13188910e-01
7.15131938e-01 -1.18962550e+00 -4.42121714e-01 -3.66520494e-01
-4.67914730e-01 1.18733370e+00 6.81211233e-01 -5.87404013e-01
8.50543916e-01 1.22331686e-01 2.68382341e-01 -1.94392934e-01
-7.23529637e-01 -2.07611829e-01 2.91736633e-01 3.95844162e-01
2.04939529e-01 -6.12652153e-02 1.73363328e-01 7.32592165e-01
-1.10068033e-03 -1.72239110e-01 9.45656598e-02 6.00304008e-01
-8.20855618e-01 -9.72997785e-01 -8.26764107e-02 3.87999833e-01
-6.02480352e-01 -1.93364397e-02 -4.14380401e-01 7.63034821e-01
5.77575207e-01 8.76515329e-01 4.36281949e-01 -2.38425598e-01
2.84427732e-01 -5.80739975e-02 1.67841092e-01 -7.94118345e-01
-8.45792830e-01 -1.90747052e-01 1.05162129e-01 -5.87178469e-01
-2.06367582e-01 -6.68070242e-02 -1.03102899e+00 -1.82794437e-01
-2.33523279e-01 3.79538447e-01 2.98915893e-01 8.61067891e-01
3.24946135e-01 3.90693933e-01 8.48350465e-01 -8.60293210e-01
-5.56406558e-01 -8.87393177e-01 -6.14305377e-01 3.21238071e-01
5.26139259e-01 -4.94378895e-01 -5.20727456e-01 7.25612938e-02] | [9.406293869018555, 2.442139148712158] |
0b9d8bb0-cf8d-4ba1-a76b-9c06a950cec0 | document-translation-vs-query-translation-for | null | null | https://aclanthology.org/2020.acl-main.613 | https://aclanthology.org/2020.acl-main.613.pdf | Document Translation vs. Query Translation for Cross-Lingual Information Retrieval in the Medical Domain | We present a thorough comparison of two principal approaches to Cross-Lingual Information Retrieval: document translation (DT) and query translation (QT). Our experiments are conducted using the cross-lingual test collection produced within the CLEF eHealth information retrieval tasks in 2013{--}2015 containing English documents and queries in several European languages. We exploit the Statistical Machine Translation (SMT) and Neural Machine Translation (NMT) paradigms and train several domain-specific and task-specific machine translation systems to translate the non-English queries into English (for the QT approach) and the English documents to all the query languages (for the DT approach). The results show that the quality of QT by SMT is sufficient enough to outperform the retrieval results of the DT approach for all the languages. NMT then further boosts translation quality and retrieval quality for both QT and DT for most languages, but still, QT provides generally better retrieval results than DT. | ['Shadi Saleh', 'Pavel Pecina'] | 2020-07-01 | null | null | null | acl-2020-6 | ['cross-lingual-information-retrieval'] | ['natural-language-processing'] | [ 7.04765096e-02 -2.55367130e-01 -7.19345748e-01 -6.87572584e-02
-2.21077275e+00 -1.10389125e+00 1.16684496e+00 1.71848193e-01
-6.79428756e-01 9.62918699e-01 5.62889457e-01 -7.43664384e-01
-1.81940854e-01 -3.39634985e-01 -6.36865258e-01 -2.60665894e-01
6.44670248e-01 1.14827609e+00 -9.62157398e-02 -4.32375401e-01
1.66401908e-01 2.75754601e-01 -9.30830956e-01 7.33968198e-01
9.48953807e-01 5.63827991e-01 2.76233032e-02 5.26780367e-01
-3.88958186e-01 1.78475410e-01 -3.85549217e-01 -6.67326689e-01
2.44459689e-01 -4.50014859e-01 -1.27178144e+00 -5.24648488e-01
4.22007173e-01 -8.69575366e-02 -4.27484848e-02 7.76985705e-01
8.54256213e-01 -2.49302238e-01 8.13358188e-01 -6.14638388e-01
-1.19806719e+00 4.37872738e-01 -3.48781317e-01 4.76076901e-02
6.71952605e-01 -3.39043438e-01 1.20682251e+00 -1.42562604e+00
1.14967763e+00 1.41515172e+00 3.72170359e-01 4.87782598e-01
-1.15800738e+00 -4.23572600e-01 -5.03176570e-01 -1.54403076e-01
-1.43684554e+00 -5.92539608e-01 7.78319910e-02 -2.70483017e-01
1.43253863e+00 3.38764012e-01 1.52684003e-01 1.11021781e+00
4.55928952e-01 8.13455105e-01 1.26959801e+00 -1.11571980e+00
-9.00235102e-02 6.88655734e-01 -1.70434609e-01 1.25443548e-01
1.92940459e-01 1.59735709e-01 -6.07356071e-01 -3.54811579e-01
3.17069292e-01 -3.52357954e-01 -9.80258808e-02 2.19957590e-01
-1.45225048e+00 8.27499866e-01 1.09467767e-01 6.02552295e-01
-5.01416028e-01 -1.55780673e-01 7.39210069e-01 9.00110424e-01
7.24438608e-01 6.54662371e-01 -8.14907074e-01 4.70257699e-02
-1.06878710e+00 1.88982725e-01 9.05057847e-01 1.28787363e+00
6.66394591e-01 -4.18540597e-01 -6.70016646e-01 1.05702519e+00
3.20248812e-01 1.22124481e+00 6.79552913e-01 -6.73194945e-01
9.68617737e-01 2.76530236e-01 4.02269036e-01 -5.98467648e-01
1.80372242e-02 -4.35691476e-01 -3.09074879e-01 -6.72423065e-01
1.37647018e-01 -1.55688629e-01 -8.73968840e-01 1.41008675e+00
4.94826492e-03 -1.14226544e+00 5.53405941e-01 6.73637211e-01
9.99343157e-01 8.58406425e-01 1.33295640e-01 -4.19079661e-01
1.39168513e+00 -1.08831346e+00 -9.50517714e-01 -3.16131175e-01
1.42909634e+00 -1.50179601e+00 1.20628524e+00 -1.85111433e-01
-1.23544800e+00 -4.68529612e-01 -6.61000431e-01 -3.33122671e-01
-6.48416519e-01 5.57748258e-01 1.27741829e-01 4.00434285e-01
-1.64177871e+00 -3.10479533e-02 -5.27637124e-01 -9.15725529e-01
-2.04705030e-01 2.43763342e-01 -4.21836376e-01 -6.29561424e-01
-1.53426516e+00 1.46614611e+00 2.06667811e-01 -1.78663552e-01
-6.21642411e-01 -3.59392285e-01 -5.61206818e-01 -2.67220885e-01
-1.44384995e-01 -7.82450914e-01 1.32879579e+00 -8.62474561e-01
-1.12562633e+00 1.40614879e+00 -5.58002293e-01 -1.91712633e-01
4.89322513e-01 -3.84396434e-01 -4.28343832e-01 2.17207104e-01
7.64842033e-01 7.74244487e-01 3.45823228e-01 -1.06943691e+00
-5.49358547e-01 -3.07656348e-01 -4.49980438e-01 5.03311992e-01
-1.82840645e-01 8.87631595e-01 -8.48013222e-01 -4.97056395e-01
-5.68077676e-02 -1.07588625e+00 2.04945251e-01 -5.38210452e-01
-3.39819729e-01 -4.59108174e-01 3.10880661e-01 -1.04271579e+00
1.31969154e+00 -1.72883832e+00 1.83521211e-01 -7.49063045e-02
-4.32346374e-01 1.90964386e-01 -4.96810913e-01 1.12331820e+00
1.12272859e-01 3.55804831e-01 6.01514243e-02 -4.41919893e-01
1.14340119e-01 3.58077288e-01 -5.03534496e-01 3.96664850e-02
1.23603560e-01 1.52455997e+00 -5.95329583e-01 -7.12624431e-01
-2.83181280e-01 5.06469488e-01 -1.13923974e-01 2.50998568e-02
-9.28297564e-02 4.01944727e-01 -6.14168823e-01 8.64063203e-01
2.01284811e-01 -2.41562654e-03 4.70353439e-02 2.38718744e-02
-5.03664054e-02 9.27421510e-01 -1.34732664e-01 1.76529324e+00
-7.08887577e-01 5.23747981e-01 -1.62592068e-01 -3.70734036e-01
8.40692461e-01 8.25820148e-01 3.20802212e-01 -1.43510962e+00
-3.06182623e-01 1.04480112e+00 -4.04891878e-01 -5.10294616e-01
5.96574545e-01 -1.89248174e-01 -3.14178228e-01 7.79397190e-01
-1.40995756e-02 -3.07111353e-01 2.15918541e-01 9.41954553e-02
7.72659361e-01 3.49064559e-01 2.03556493e-02 -4.35194165e-01
4.97926146e-01 6.23753488e-01 9.47820693e-02 6.74502492e-01
2.34490946e-01 3.16775173e-01 -1.81629539e-01 -9.97986645e-02
-1.28754592e+00 -8.25797200e-01 -1.45112917e-01 1.36992121e+00
-3.97697985e-01 -3.73525053e-01 -8.12577367e-01 -7.12506711e-01
-1.20132543e-01 9.03410912e-01 -3.44988734e-01 -2.39794791e-01
-6.17294371e-01 -6.23772979e-01 6.60416186e-01 1.76488087e-01
3.06787670e-01 -1.19293320e+00 -2.62388531e-02 2.63256997e-01
-7.28912532e-01 -1.22751117e+00 -9.15053606e-01 1.90418661e-01
-9.43023205e-01 -2.75786608e-01 -1.27227449e+00 -9.74592686e-01
3.28694105e-01 1.65672302e-01 1.36194193e+00 -1.39877707e-01
2.66998615e-02 6.09685242e-01 -5.15234470e-01 -5.28828681e-01
-9.47157025e-01 4.54163343e-01 7.95149952e-02 -5.01910865e-01
9.19909775e-01 1.52657460e-03 -3.80628377e-01 2.03726426e-01
-9.46195483e-01 -2.97945201e-01 9.53872681e-01 7.75934458e-01
6.46178544e-01 -5.29602885e-01 6.94638312e-01 -5.75933039e-01
1.20241296e+00 -2.13081867e-01 -3.48108053e-01 8.28116536e-01
-1.19129467e+00 1.77144453e-01 1.45188361e-01 -3.10965836e-01
-7.30198145e-01 -4.17627901e-01 -9.19803884e-03 -6.64129183e-02
1.74292415e-01 1.03083968e+00 1.64002180e-01 1.19998284e-01
8.89185131e-01 4.64273423e-01 -3.21999460e-01 -6.56843066e-01
3.77586126e-01 9.99191523e-01 1.79128796e-01 -6.76863372e-01
6.60049975e-01 -2.24413387e-02 -4.55652714e-01 -2.50752449e-01
-7.69099474e-01 -6.41648412e-01 -7.67154455e-01 9.61214676e-02
9.75990415e-01 -1.04236150e+00 1.03934109e-01 -1.87816799e-01
-1.38851571e+00 -1.32822216e-01 -1.52813315e-01 6.84896171e-01
-4.94307131e-01 -9.80709493e-02 -7.84988940e-01 -6.46794736e-01
-9.52409565e-01 -1.31153595e+00 1.77466416e+00 -3.89631152e-01
-4.05714482e-01 -1.19580078e+00 3.65565538e-01 4.08181906e-01
6.44217789e-01 -2.13869080e-01 1.31054008e+00 -8.30372632e-01
-1.89706847e-01 -5.51652551e-01 -2.74043828e-01 1.19168036e-01
2.91299403e-01 -2.18196511e-01 -5.28025746e-01 -5.59780657e-01
-6.83893189e-02 -5.52732646e-01 5.65883875e-01 1.76339969e-01
-1.41588062e-01 -5.04360318e-01 -3.41224939e-01 -2.72728242e-02
1.54040313e+00 2.76712865e-01 5.07774770e-01 7.85790861e-01
2.80448139e-01 7.93094873e-01 7.13149607e-01 -4.22102392e-01
5.34047425e-01 1.08454514e+00 -5.01079082e-01 -5.27031422e-01
-1.99212760e-01 -3.50814551e-01 7.06874371e-01 1.17328382e+00
2.33944371e-01 -1.65470794e-01 -1.15501261e+00 7.19302833e-01
-1.68282092e+00 -5.58012247e-01 8.98730475e-04 2.33761907e+00
1.31049299e+00 -2.26443246e-01 -5.14592156e-02 -4.13179398e-01
3.67744714e-01 -4.17909116e-01 -9.22138914e-02 -8.62058640e-01
-3.58863443e-01 4.11094785e-01 5.13488710e-01 6.32286191e-01
-6.58185899e-01 1.35651731e+00 7.41864538e+00 8.61242354e-01
-1.00223029e+00 3.67445379e-01 4.18230027e-01 9.51771662e-02
-5.43321490e-01 2.61391662e-02 -1.08196068e+00 -1.01045556e-01
1.44903564e+00 -5.68535447e-01 4.42224115e-01 4.35170650e-01
2.06825256e-01 8.02597031e-02 -1.20700431e+00 7.66555846e-01
2.64539570e-02 -1.28003788e+00 5.67251146e-01 5.30666336e-02
9.59507227e-01 5.45493662e-01 1.44202739e-01 6.74288869e-01
2.14291230e-01 -9.82633471e-01 5.87423325e-01 3.56026828e-01
1.26776123e+00 -4.50855970e-01 9.27959681e-01 4.86473441e-01
-7.40490437e-01 3.92678708e-01 -2.96695620e-01 6.65666163e-01
-6.14494048e-02 1.45207882e-01 -8.47120404e-01 9.99924004e-01
7.59057701e-01 4.91082370e-01 -6.90222561e-01 4.75551575e-01
-8.66492912e-02 4.57766175e-01 -1.63476244e-01 -2.97839772e-02
6.15230083e-01 -2.79192090e-01 4.45917606e-01 1.61803913e+00
3.48623574e-01 -4.78803813e-01 1.02176905e-01 5.09331703e-01
-3.06798905e-01 7.91063070e-01 -9.02683437e-01 -2.87000060e-01
3.83858860e-01 8.83953214e-01 -2.14389384e-01 -6.59405649e-01
-3.65354955e-01 1.16868556e+00 5.45286238e-02 7.76097834e-01
-1.49437085e-01 -3.06522816e-01 8.70731845e-02 -8.13086480e-02
-1.62219703e-02 -6.10323697e-02 -1.92620561e-01 -1.04839551e+00
3.60938817e-01 -1.42739880e+00 5.81187248e-01 -8.98605347e-01
-1.22140920e+00 1.04438543e+00 1.39432877e-01 -1.39725983e+00
-8.42391253e-01 -2.83503324e-01 2.36095116e-01 1.57560456e+00
-1.64640927e+00 -1.28945458e+00 6.72326207e-01 4.96905297e-01
5.01376033e-01 -2.64370441e-01 1.30506611e+00 6.03614569e-01
-3.60909244e-03 6.24987066e-01 7.00358868e-01 4.01587859e-02
1.37354481e+00 -9.15853202e-01 3.58788997e-01 4.94751304e-01
2.48303041e-01 1.13434446e+00 4.67159241e-01 -7.35021412e-01
-1.55774236e+00 -1.04070485e+00 1.97822487e+00 -8.55569839e-01
5.48009694e-01 -2.80636996e-01 -6.58960581e-01 7.40006268e-01
7.25718021e-01 -7.35230267e-01 6.34297907e-01 -1.06779762e-01
-3.81924897e-01 3.58211063e-02 -1.02414989e+00 6.18062556e-01
3.56912851e-01 -1.19180453e+00 -7.68474221e-01 1.05855656e+00
1.02919924e+00 -5.51839136e-02 -1.16831195e+00 4.28588152e-01
6.24459386e-01 -1.93585560e-01 9.35602665e-01 -9.32888687e-01
4.07889336e-01 -1.05217077e-01 -5.38115919e-01 -1.12305748e+00
-5.07861152e-02 -5.91268539e-01 4.69889879e-01 1.14727199e+00
1.13862860e+00 -7.98716068e-01 2.15792134e-02 6.74706101e-01
-1.50173446e-02 -5.65185845e-01 -1.25970268e+00 -8.84065986e-01
8.11211467e-01 -2.37495124e-01 3.60228330e-01 9.45438802e-01
-9.82618481e-02 8.35260808e-01 -8.31048489e-02 -4.05520469e-01
1.64410427e-01 1.70695812e-01 5.67896843e-01 -7.46639490e-01
-1.97489876e-02 -4.81674820e-01 3.62285793e-01 -8.90528738e-01
1.01331182e-01 -1.45633972e+00 1.85502339e-02 -2.00439429e+00
3.11970681e-01 -6.12277687e-02 -1.16387434e-01 5.47681630e-01
-1.33744717e-01 2.95242161e-01 -1.62963212e-01 9.20447886e-01
-3.11039090e-01 3.68546158e-01 1.28854454e+00 -1.63505778e-01
-2.34658018e-01 -9.57875177e-02 -8.03735375e-01 -1.20542593e-01
4.05495673e-01 -7.87034571e-01 -1.46576360e-01 -9.56816554e-01
2.31373414e-01 2.41183937e-01 -1.14013083e-01 -1.62627220e-01
1.62358686e-01 7.01712221e-02 2.67907411e-01 -7.06663311e-01
2.77286805e-02 -6.03717983e-01 1.85177308e-02 4.53927755e-01
-7.15626419e-01 8.08197320e-01 5.04749417e-01 7.26639060e-03
-5.06652057e-01 2.81245448e-02 4.61143255e-01 -1.64793938e-01
9.77083668e-02 8.31785575e-02 -4.71069068e-01 1.25061303e-01
1.69171125e-01 1.05701424e-01 -3.73658270e-01 -5.86442113e-01
-5.33493340e-01 3.29291403e-01 2.83781618e-01 6.49153352e-01
3.61451924e-01 -1.53502011e+00 -1.22662246e+00 -1.04307368e-01
5.58209479e-01 -6.85299575e-01 -6.34303153e-01 1.23170960e+00
-2.33661130e-01 1.50655079e+00 1.92114890e-01 -6.11867011e-01
-9.91847575e-01 5.08133590e-01 2.04704985e-01 -7.60163844e-01
-1.90330073e-01 4.06347603e-01 -6.37473688e-02 -1.06398678e+00
6.10950142e-02 1.15758434e-01 -3.95049565e-02 1.76337082e-02
4.39787179e-01 -1.41792580e-01 5.30001104e-01 -9.54992354e-01
-3.41489732e-01 7.97131062e-01 -3.24026555e-01 -9.48145151e-01
1.09448159e+00 -3.50181907e-01 -4.63848174e-01 5.02788424e-01
1.63136578e+00 -2.75794007e-02 2.38492176e-01 -7.62142420e-01
7.03286350e-01 -1.49888352e-01 4.50634994e-02 -1.37906432e+00
-4.10914570e-01 8.65195096e-01 5.18734217e-01 -1.51407972e-01
1.09178758e+00 1.47384658e-01 8.21226060e-01 7.01522887e-01
5.70668638e-01 -1.24098039e+00 -3.40292662e-01 5.84017098e-01
1.25819063e+00 -1.13113654e+00 -1.99574783e-01 5.87880649e-02
-7.07550526e-01 1.06124377e+00 -7.96692297e-02 4.99092966e-01
2.60545790e-01 -2.34625101e-01 7.22551763e-01 -2.51005679e-01
-8.11637819e-01 -1.82282984e-01 9.12010014e-01 2.80305564e-01
9.22092319e-01 -1.18787974e-01 -6.90105557e-01 -4.20578867e-02
-2.26403967e-01 -6.85042143e-02 -2.68663883e-01 9.07728314e-01
-2.53711529e-02 -1.55389237e+00 -3.50414932e-01 2.21406668e-01
-8.39110494e-01 -6.99829400e-01 -9.67795193e-01 8.66622090e-01
-6.43629193e-01 1.16667652e+00 -4.60142314e-01 -2.28606552e-01
3.99339795e-01 5.05054355e-01 2.55925924e-01 -6.96151316e-01
-1.00139070e+00 7.45771348e-01 4.02931094e-01 -4.19703841e-01
-4.70197052e-01 -6.99043393e-01 -8.78555059e-01 1.42387941e-01
-4.45991784e-01 7.02222168e-01 1.07129037e+00 1.02571774e+00
4.66603428e-01 -1.11233048e-01 6.09703600e-01 -2.00537309e-01
-7.25324631e-01 -1.38582671e+00 5.11831418e-02 1.80709094e-01
1.96010336e-01 -1.08751319e-01 -2.98659861e-01 1.37171401e-02] | [11.567727088928223, 9.9383544921875] |
8fa328ca-8b59-431a-914a-10875fd532b7 | keypoint-less-camera-calibration-for-sports | 2207.11709 | null | https://arxiv.org/abs/2207.11709v2 | https://arxiv.org/pdf/2207.11709v2.pdf | TVCalib: Camera Calibration for Sports Field Registration in Soccer | Sports field registration in broadcast videos is typically interpreted as the task of homography estimation, which provides a mapping between a planar field and the corresponding visible area of the image. In contrast to previous approaches, we consider the task as a camera calibration problem. First, we introduce a differentiable objective function that is able to learn the camera pose and focal length from segment correspondences (e.g., lines, point clouds), based on pixel-level annotations for segments of a known calibration object. The calibration module iteratively minimizes the segment reprojection error induced by the estimated camera parameters. Second, we propose a novel approach for 3D sports field registration from broadcast soccer images. Compared to the typical solution, which subsequently refines an initial estimation, our solution does it in one step. The proposed method is evaluated for sports field registration on two datasets and achieves superior results compared to two state-of-the-art approaches. | ['Ralph Ewerth', 'Jonas Theiner'] | 2022-07-24 | null | null | null | null | ['homography-estimation'] | ['computer-vision'] | [ 3.68095577e-01 1.14950342e-02 -8.95909742e-02 -2.81189948e-01
-1.03358316e+00 -7.49248207e-01 3.48533303e-01 3.78674567e-01
-5.90866506e-01 4.11847591e-01 -8.42676610e-02 4.17623341e-01
4.96293455e-02 -5.93241453e-01 -1.19118392e+00 -4.51519638e-01
3.78611416e-01 7.51777232e-01 5.56071639e-01 -1.88268572e-01
5.44699430e-01 4.50049490e-01 -1.21039760e+00 -2.27241397e-01
5.71927905e-01 7.26441562e-01 9.65352505e-02 5.84759116e-01
2.77857065e-01 3.64102423e-01 -3.03443015e-01 -6.92633927e-01
3.43511999e-01 -3.51975471e-01 -8.32331896e-01 7.66209900e-01
9.41457927e-01 -1.41846621e-02 -4.04831231e-01 1.09610713e+00
1.18767761e-01 1.44796595e-01 4.24701393e-01 -1.01983750e+00
1.63080767e-01 2.01801240e-01 -7.64189005e-01 1.45151513e-02
8.10544908e-01 -1.54383913e-01 7.79367447e-01 -6.70609534e-01
1.04122615e+00 7.34637856e-01 8.74213576e-01 7.90434331e-02
-1.59820151e+00 -3.34920138e-01 -7.87555501e-02 1.12350948e-01
-1.53145564e+00 -1.76746979e-01 9.15539742e-01 -7.61993408e-01
3.40404093e-01 -4.76538241e-02 1.01386166e+00 5.96653759e-01
2.08443627e-01 4.46248859e-01 9.56602216e-01 -5.62339783e-01
1.77012086e-01 -5.32829650e-02 2.13491425e-01 5.37510574e-01
2.26912543e-01 -7.95370415e-02 -7.40182638e-01 -3.90509479e-02
1.18506861e+00 -2.06454486e-01 -4.37597305e-01 -9.57651556e-01
-1.37445927e+00 6.59593463e-01 2.06299379e-01 -1.19314678e-02
-3.97529274e-01 2.73247927e-01 5.95993847e-02 -6.14898689e-02
3.23789686e-01 2.91375518e-01 1.52528062e-01 -1.77415460e-01
-1.14580536e+00 4.95583981e-01 7.63484359e-01 1.11822772e+00
1.09997594e+00 -2.65264422e-01 3.12957853e-01 5.60739040e-01
1.03042625e-01 4.04661208e-01 -1.31285533e-01 -1.20326281e+00
5.90498209e-01 4.91392821e-01 9.31853130e-02 -1.12829864e+00
-2.65100002e-01 -4.02547657e-01 -2.77787745e-01 5.99299520e-02
8.52312148e-01 5.86558208e-02 -5.07596016e-01 1.25847113e+00
6.73226655e-01 6.60436809e-01 -7.65007436e-02 1.10360074e+00
4.49550927e-01 3.24149281e-01 -4.17476684e-01 -2.16797628e-02
1.34655285e+00 -8.00425887e-01 -4.68628109e-01 -5.04813313e-01
1.74879462e-01 -1.09800839e+00 6.30989254e-01 5.53612947e-01
-1.25168145e+00 -4.05546367e-01 -1.04025710e+00 -7.79318437e-02
2.46857718e-01 4.04572517e-01 2.06908196e-01 5.02968669e-01
-8.06774020e-01 4.92087901e-01 -8.10390115e-01 -4.72160786e-01
-7.82264620e-02 3.56719673e-01 -7.09209502e-01 -4.74399514e-03
-6.19723678e-01 7.31185555e-01 4.16207969e-01 1.90784782e-01
-6.93509758e-01 -6.99398577e-01 -9.77841914e-01 -1.07176915e-01
6.86357737e-01 -7.54482985e-01 1.25294220e+00 -1.02799392e+00
-1.75991130e+00 1.28795660e+00 -7.28110299e-02 -4.96455699e-01
7.18784451e-01 -4.28958893e-01 5.45698404e-02 4.05465811e-01
1.84159026e-01 4.92886603e-01 6.95845187e-01 -1.49574113e+00
-6.94414437e-01 -3.61751676e-01 2.61878401e-01 4.40152556e-01
2.67388463e-01 -1.10173471e-01 -1.05388975e+00 -5.50475240e-01
7.48943686e-01 -1.22617662e+00 -3.17051351e-01 5.21629713e-02
-2.99439967e-01 3.37988973e-01 4.37292695e-01 -8.26957107e-01
9.26534653e-01 -1.95452261e+00 5.12748659e-01 5.26040494e-01
3.20317037e-02 -2.25683108e-01 2.73314416e-01 3.18802029e-01
-8.94060731e-03 -6.10773444e-01 -3.98095727e-01 -3.57240319e-01
-3.75034869e-01 1.80091739e-01 3.13538834e-02 1.15275371e+00
-1.72517955e-01 4.84019279e-01 -7.84035325e-01 -6.78530514e-01
3.60558063e-01 4.05305028e-01 -4.70830083e-01 3.26131023e-02
4.14799191e-02 7.72204578e-01 -1.94560573e-01 3.72115910e-01
8.60160768e-01 1.47930130e-01 2.43521377e-01 -3.67967695e-01
-4.28973973e-01 -1.61768764e-01 -1.78353083e+00 2.33519697e+00
-2.35724464e-01 7.06928730e-01 1.62944943e-01 -1.14221501e+00
1.06706583e+00 3.22540522e-01 8.28299105e-01 -1.51547253e-01
3.21897656e-01 2.21169263e-01 -4.66215700e-01 -3.91237080e-01
6.73339427e-01 8.98127630e-02 -1.48432761e-01 1.01319691e-02
3.41918260e-01 -4.34304059e-01 2.42431432e-01 -3.99620039e-03
6.67098403e-01 5.51956534e-01 2.47326106e-01 -2.77500629e-01
7.42044091e-01 3.59203875e-01 5.92025995e-01 3.59557122e-01
2.67946005e-01 9.45105195e-01 5.29235363e-01 -4.38806176e-01
-1.17511034e+00 -8.96560729e-01 -1.10980928e-01 2.59140760e-01
7.75567472e-01 -5.50761223e-01 -1.22845912e+00 -3.87664884e-01
-1.08532079e-01 1.49169147e-01 -4.44280803e-01 1.62908137e-01
-9.44451928e-01 -2.10570678e-01 2.31493428e-01 5.10732591e-01
4.41097885e-01 -2.37048894e-01 -7.23700762e-01 2.52060831e-01
-4.87184584e-01 -1.53409338e+00 -8.92849326e-01 -2.17878029e-01
-9.84789193e-01 -1.24613285e+00 -7.42295444e-01 -8.63404036e-01
7.40604997e-01 3.12285632e-01 1.15474963e+00 -1.22732759e-01
-2.79424991e-02 6.54278636e-01 -2.46036112e-01 -2.42292546e-02
-1.09963983e-01 5.43829799e-02 -1.04679629e-01 4.44325566e-01
9.37977061e-02 -4.36912179e-01 -6.88648105e-01 5.92928827e-01
-7.70813823e-01 4.67182975e-03 2.98349082e-01 6.64802432e-01
1.09136844e+00 -3.22352171e-01 -3.84741098e-01 -8.05709362e-01
6.99836463e-02 -4.79425043e-02 -1.18352234e+00 3.19697261e-01
-2.33691633e-01 -1.07741468e-01 8.36039078e-04 -3.93594235e-01
-8.93872678e-01 1.06876624e+00 1.23385727e-01 -5.66996396e-01
-2.94092447e-01 2.73749322e-01 -9.14893076e-02 -5.35819829e-01
6.70921266e-01 1.42854318e-01 -7.89021775e-02 -5.47121525e-01
2.58837163e-01 3.68453026e-01 1.17137945e+00 -6.32746339e-01
8.96564960e-01 7.21247256e-01 3.15227419e-01 -8.66541564e-01
-6.52773857e-01 -1.00878680e+00 -1.40273583e+00 -6.50046825e-01
1.12258530e+00 -9.62374449e-01 -8.41037571e-01 2.95803636e-01
-1.36782718e+00 -5.50208874e-02 -3.21791589e-01 9.82485533e-01
-9.26934183e-01 6.44632459e-01 -3.50273430e-01 -4.54930872e-01
1.15667038e-01 -1.26217163e+00 1.38801980e+00 2.56197274e-01
4.19310182e-02 -9.88646328e-01 4.99066174e-01 5.88040650e-01
-4.69801039e-01 6.70126379e-01 1.05415940e-01 1.91278178e-02
-8.46692502e-01 -3.95254940e-01 3.10626477e-01 1.73160642e-01
-4.10546184e-01 -1.34279137e-03 -8.50734174e-01 -1.23557985e-01
-2.34452751e-03 2.04320103e-01 4.17500496e-01 6.34882569e-01
5.64151406e-01 9.25984979e-02 -3.22338372e-01 1.10477006e+00
1.58552992e+00 -2.31566094e-03 7.02101171e-01 7.26122439e-01
6.54361010e-01 6.92542136e-01 9.90712643e-01 3.00705850e-01
4.08429623e-01 1.30521655e+00 2.95880407e-01 -2.39126787e-01
-3.78653631e-02 -4.66008008e-01 1.24176972e-01 4.38778937e-01
-4.29188401e-01 2.20142812e-01 -8.55946183e-01 3.69075090e-01
-2.02304435e+00 -8.29689443e-01 -6.69421971e-01 2.68515849e+00
3.21917683e-01 6.66913241e-02 4.10044849e-01 1.72534853e-01
8.73830318e-01 -2.48839140e-01 -2.26084009e-01 6.33089012e-03
-5.15003726e-02 -7.95174681e-04 1.02982950e+00 6.84627891e-01
-1.22201705e+00 8.78657520e-01 6.28605413e+00 5.11376679e-01
-1.08202171e+00 8.74533802e-02 6.40949309e-02 2.62752026e-01
6.53927848e-02 5.12524486e-01 -8.03894520e-01 1.53556019e-01
3.17982405e-01 -2.03622808e-03 2.72124410e-01 6.78524554e-01
3.54741924e-02 -4.26443040e-01 -1.19859231e+00 1.18498230e+00
4.81797576e-01 -1.40772605e+00 -5.17259836e-01 1.80675939e-01
7.66729951e-01 -3.37101161e-01 -3.45728010e-01 -4.33365315e-01
-1.15761518e-01 -5.10337412e-01 1.12964487e+00 7.25563765e-01
5.98574460e-01 -6.70943558e-01 6.79202735e-01 3.35566133e-01
-1.36889684e+00 2.55997568e-01 -2.44364873e-01 8.90557095e-02
6.26066506e-01 2.12078914e-01 -6.61423862e-01 7.48130441e-01
6.68763876e-01 9.03688014e-01 -4.81501579e-01 1.45460379e+00
-3.38823944e-01 4.27009344e-01 -2.69160300e-01 6.45111978e-01
-4.46298858e-03 -9.11803544e-01 7.07527518e-01 1.07781971e+00
3.19016248e-01 1.08566634e-01 3.56566817e-01 6.70023620e-01
2.16923565e-01 2.73732126e-01 -4.81061369e-01 5.48423946e-01
1.78524684e-02 1.34410441e+00 -1.12060118e+00 -2.18910590e-01
-4.16514188e-01 9.82535243e-01 3.78631381e-03 2.74579167e-01
-7.89698124e-01 -6.01021238e-02 2.33852997e-01 6.10168099e-01
2.57521421e-01 -3.69925439e-01 -1.51021183e-01 -1.46124971e+00
2.76921719e-01 -6.94465041e-01 3.27226996e-01 -8.48272204e-01
-7.36928940e-01 4.99229103e-01 2.88832724e-01 -1.71400070e+00
-2.40961179e-01 -3.37697715e-01 -4.34297144e-01 6.34917200e-01
-1.22856665e+00 -1.18603683e+00 -6.15970910e-01 6.01583064e-01
5.73479056e-01 2.57324189e-01 3.64999354e-01 3.03890198e-01
-2.25012004e-01 2.80434728e-01 1.23506933e-01 1.67655170e-01
7.55699754e-01 -1.20857275e+00 1.40739933e-01 9.87267673e-01
3.48328203e-01 2.54539400e-01 8.91236246e-01 -6.20584071e-01
-1.27778184e+00 -5.18871605e-01 6.53992176e-01 -4.59892362e-01
5.12802064e-01 -4.51263309e-01 -7.43074954e-01 9.37090635e-01
4.56864387e-02 -4.83220741e-02 3.02979290e-01 -2.59458065e-01
1.60497189e-01 -3.40689659e-01 -7.26401210e-01 4.84093502e-02
8.27285409e-01 -2.09154010e-01 -4.92535919e-01 3.10581297e-01
9.63321999e-02 -1.12050426e+00 -1.11791217e+00 -4.73151682e-03
6.20164633e-01 -7.71009982e-01 1.25238633e+00 -8.60353261e-02
1.42700613e-01 -5.89999914e-01 6.23555966e-02 -1.16220391e+00
7.47833773e-02 -7.94485748e-01 5.25674701e-01 1.10646713e+00
3.13573591e-02 -1.47211879e-01 1.03863037e+00 4.48528975e-01
-2.30391905e-01 -1.42046332e-01 -1.05015516e+00 -5.73294640e-01
-2.83671618e-01 -1.45794168e-01 2.46778771e-01 7.56275952e-01
-8.96557197e-02 3.28459293e-01 -4.59391147e-01 4.58743840e-01
1.10471308e+00 2.14844242e-01 1.23604071e+00 -1.45351946e+00
-3.99826705e-01 1.49381626e-02 -9.84003782e-01 -1.25066471e+00
1.89689681e-01 -6.16904616e-01 2.09545225e-01 -1.23788285e+00
-5.02281711e-02 -2.39498556e-01 5.65205336e-01 -7.21971691e-02
8.52974281e-02 4.49550331e-01 2.32495099e-01 2.69365728e-01
-3.78933251e-01 2.89317816e-02 1.16402626e+00 1.18698604e-01
-2.81748950e-01 3.18057716e-01 -4.02171798e-02 9.89201128e-01
3.39774668e-01 -5.89351952e-01 -9.24162939e-02 -4.36006367e-01
1.66669354e-01 6.03710353e-01 6.09444380e-01 -1.22767568e+00
5.49413383e-01 -1.95310280e-01 4.47598547e-02 -5.43956399e-01
5.26657581e-01 -9.25532699e-01 7.25922763e-01 2.88410813e-01
-2.07845539e-01 2.05445930e-01 -1.75265074e-01 6.10286474e-01
-4.65959698e-01 -5.78849614e-01 6.80798769e-01 -1.35343134e-01
-6.51919425e-01 5.93618937e-02 -1.58018753e-01 3.70004252e-02
1.22268713e+00 -6.37656152e-01 4.32013012e-02 -3.93313587e-01
-9.74708498e-01 6.52511343e-02 8.51274908e-01 9.31046009e-02
6.19860113e-01 -1.10541928e+00 -7.42928922e-01 1.33269697e-01
2.50348300e-01 3.31253529e-01 1.74304470e-01 1.21337032e+00
-1.14902389e+00 1.31540194e-01 -2.12408900e-01 -1.18377352e+00
-1.51128101e+00 3.48739415e-01 6.53459013e-01 -6.78843958e-03
-7.49936819e-01 3.91808122e-01 9.06053185e-02 -1.75541461e-01
8.56270567e-02 -3.50136459e-01 -2.78832525e-01 -1.63808838e-01
1.46804228e-01 5.61078370e-01 1.50651425e-01 -1.33146405e+00
-2.35308826e-01 1.47630906e+00 4.25111622e-01 -4.59736735e-01
1.26615202e+00 -2.54509747e-01 1.57082632e-01 2.76633054e-01
9.96029735e-01 7.84139559e-02 -1.60383654e+00 -3.10618728e-01
-3.71142365e-02 -8.29107285e-01 1.48055211e-01 -7.22098649e-02
-1.21489811e+00 7.62863815e-01 6.94564641e-01 -2.68831164e-01
8.65195811e-01 8.24157149e-02 5.41085005e-01 5.52033149e-02
3.86179358e-01 -1.26829886e+00 -1.01796240e-01 3.45011987e-02
6.30944252e-01 -1.03015828e+00 2.50075966e-01 -9.44615543e-01
-7.06714332e-01 1.44738972e+00 4.35992986e-01 -5.02100229e-01
3.55263531e-01 1.87058434e-01 1.72255635e-01 -2.68562704e-01
4.83369529e-02 -1.51224479e-01 5.42722285e-01 4.63302821e-01
2.10920647e-01 -2.06074849e-01 -4.87910599e-01 -4.23159264e-02
-2.71620095e-01 1.86137781e-01 7.33563721e-01 8.98396134e-01
-2.96346843e-01 -1.32793522e+00 -8.85990262e-01 -2.35304520e-01
-2.45524645e-01 3.93794894e-01 -2.51542658e-01 8.44241917e-01
1.75901309e-01 7.78616726e-01 3.08734894e-01 -2.39378974e-01
8.78353298e-01 -3.70470762e-01 7.11409867e-01 -5.05341411e-01
-6.43864751e-01 6.53120100e-01 -5.25625497e-02 -6.40647709e-01
-9.16542768e-01 -1.05649531e+00 -9.24239635e-01 5.92707135e-02
-5.37103951e-01 2.19104648e-01 8.36269855e-01 7.86518157e-01
-1.61793351e-01 6.01709895e-02 4.76062983e-01 -1.00454223e+00
-1.62186489e-01 -4.72774714e-01 -7.98540413e-01 6.00188196e-01
1.00663103e-01 -8.13915789e-01 -1.37901334e-02 7.00632393e-01] | [7.887402057647705, -1.655833125114441] |
a42c96db-8fa6-463e-aedb-bf6c8b6cf663 | a-transformer-based-generative-adversarial | 2207.14134 | null | https://arxiv.org/abs/2207.14134v2 | https://arxiv.org/pdf/2207.14134v2.pdf | A Transformer-based Generative Adversarial Network for Brain Tumor Segmentation | Brain tumor segmentation remains a challenge in medical image segmentation tasks. With the application of transformer in various computer vision tasks, transformer blocks show the capability of learning long-distance dependency in global space, which is complementary with CNNs. In this paper, we proposed a novel transformer-based generative adversarial network to automatically segment brain tumors with multi-modalities MRI. Our architecture consists of a generator and a discriminator, which are trained in min-max game progress. The generator is based on a typical "U-shaped" encoder-decoder architecture, whose bottom layer is composed of transformer blocks with resnet. Besides, the generator is trained with deep supervision technology. The discriminator we designed is a CNN-based network with multi-scale $L_{1}$ loss, which is proved to be effective for medical semantic image segmentation. To validate the effectiveness of our method, we conducted experiments on BRATS2015 dataset, achieving comparable or better performance than previous state-of-the-art methods. | ['Senchun Chai', 'Baihai Zhang', 'Long Chen', 'Liqun Huang'] | 2022-07-28 | null | null | null | null | ['brain-tumor-segmentation'] | ['medical'] | [ 3.97340924e-01 3.81126314e-01 3.38027269e-01 -3.32151294e-01
-9.72957671e-01 -8.74267817e-02 2.84868330e-01 -4.44663852e-01
-6.05075479e-01 7.35112965e-01 7.30361603e-03 -4.08017993e-01
2.11575642e-01 -1.02327013e+00 -8.87665093e-01 -9.80007946e-01
2.44751215e-01 4.18053538e-01 4.85051036e-01 -3.33144546e-01
-1.70563459e-01 6.81291595e-02 -7.92447805e-01 2.45927319e-01
1.09938085e+00 1.18816531e+00 4.00354415e-01 3.23967069e-01
-4.04571742e-02 1.03085661e+00 -6.25097334e-01 -5.81386328e-01
1.13002144e-01 -7.06411183e-01 -8.99364829e-01 -1.94056347e-01
-2.68310547e-01 -2.18215644e-01 -6.59351885e-01 1.24177074e+00
7.24935472e-01 -3.88417721e-01 5.65757275e-01 -1.03069830e+00
-6.04829311e-01 9.21356142e-01 -3.10459703e-01 2.11046234e-01
-1.36027694e-01 2.86412776e-01 6.87437475e-01 -3.13125968e-01
3.76842022e-01 8.60104263e-01 6.67998195e-01 8.50403011e-01
-9.16864574e-01 -9.29472983e-01 3.73039395e-02 6.96828663e-02
-1.26614785e+00 7.65531659e-02 8.51970494e-01 -3.93871844e-01
5.32912076e-01 6.84493184e-02 7.64188647e-01 1.37893772e+00
6.45654440e-01 9.94759083e-01 1.16553307e+00 8.01970884e-02
-2.43525598e-02 -1.10632725e-01 -2.22808242e-01 8.74665499e-01
-2.06542850e-01 1.35441840e-01 1.65652514e-01 2.90159553e-01
1.03840876e+00 1.21842571e-01 -2.74643213e-01 -5.97061180e-02
-1.08878446e+00 9.92489219e-01 1.21529961e+00 5.05539417e-01
-1.88698709e-01 2.51086473e-01 4.38003004e-01 3.63340415e-02
5.24238825e-01 1.34121031e-01 -2.47358233e-02 2.70703256e-01
-7.34727144e-01 -6.49335608e-02 3.01011056e-01 8.48725259e-01
1.26570463e-01 1.20426118e-01 -5.12914896e-01 7.78976917e-01
2.88442552e-01 2.79952168e-01 1.04611957e+00 -1.62492439e-01
4.43812162e-01 6.82195783e-01 -4.78954077e-01 -3.82049590e-01
-5.92905164e-01 -9.42771435e-01 -1.28953803e+00 6.46663457e-02
2.34060332e-01 -3.19954634e-01 -1.33598220e+00 1.96511209e+00
1.27924204e-01 3.96384776e-01 1.72912285e-01 8.57084692e-01
1.06334233e+00 4.51047033e-01 2.57674128e-01 5.16885668e-02
1.28610408e+00 -1.26066744e+00 -4.95965064e-01 -2.66182631e-01
4.74113196e-01 -5.39842725e-01 1.02451146e+00 1.33037433e-01
-1.02656734e+00 -4.47757483e-01 -1.01843441e+00 -1.93652827e-02
-2.39372820e-01 -4.05922979e-02 5.50716758e-01 6.07674897e-01
-1.06456280e+00 3.99435312e-01 -8.80385876e-01 6.67571798e-02
8.52394104e-01 4.19893384e-01 -1.93983376e-01 1.42229319e-01
-1.43595958e+00 6.67323828e-01 6.15046680e-01 2.71653742e-01
-1.48589027e+00 -5.61517715e-01 -7.75695205e-01 7.97075778e-02
1.21179096e-01 -8.93256366e-01 1.31303310e+00 -1.20034552e+00
-1.69041383e+00 9.32196975e-01 3.82166207e-01 -6.57566309e-01
1.03242791e+00 1.05428480e-01 -3.99813384e-01 9.18618590e-02
2.66199231e-01 7.25909531e-01 7.70809472e-01 -9.36876953e-01
-4.03337806e-01 -3.89937371e-01 -1.87764317e-02 1.79227874e-01
-2.01720178e-01 -2.11953253e-01 -3.71207356e-01 -9.67795968e-01
-4.38116640e-02 -9.27662969e-01 -5.40620863e-01 -4.28821892e-01
-6.90229356e-01 -8.75960104e-03 6.92006946e-01 -8.36534619e-01
1.02266264e+00 -2.07154727e+00 1.67481855e-01 1.49095103e-01
2.74542481e-01 2.91635603e-01 -4.77270707e-02 -4.05525304e-02
-2.21156359e-01 1.63430139e-01 -6.65644109e-01 -1.74561799e-01
-1.98289111e-01 1.88128352e-02 -1.03183597e-01 3.68739486e-01
1.24511607e-01 1.36460662e+00 -8.27890277e-01 -6.02856576e-01
1.19209997e-01 4.99194354e-01 -3.46440643e-01 3.37429047e-01
-2.68548965e-01 9.01303887e-01 -7.54446089e-01 5.38405418e-01
6.24151051e-01 -2.85761565e-01 -2.59668738e-01 -2.77224388e-02
2.61993647e-01 1.39565149e-03 -3.69031876e-01 2.05049133e+00
-3.51134419e-01 2.17521787e-01 -1.02039129e-01 -1.34684181e+00
8.65918338e-01 3.16893041e-01 5.54688811e-01 -1.07728207e+00
5.22704959e-01 1.41902164e-01 1.32553920e-01 -4.57345039e-01
-9.55532119e-02 -3.81091952e-01 -2.50010163e-01 -1.42067131e-02
1.00851290e-01 -2.26468652e-01 -7.23764822e-02 1.66433621e-02
1.25092816e+00 -7.06949178e-03 -1.63158983e-01 -2.31809989e-01
7.38000333e-01 -2.46810421e-01 5.10833263e-01 4.34111387e-01
-6.31739870e-02 7.11743057e-01 7.07503200e-01 -1.39883980e-01
-9.15597677e-01 -1.15903807e+00 -2.14477196e-01 5.77774644e-01
2.50267446e-01 1.86542757e-02 -1.18187428e+00 -9.87967551e-01
-4.89529073e-01 6.09449685e-01 -8.64800334e-01 -4.82443631e-01
-6.11989915e-01 -1.12067342e+00 8.41466963e-01 7.38600910e-01
1.11972451e+00 -1.22982359e+00 -4.80989575e-01 3.02366525e-01
-2.77860165e-01 -1.07167470e+00 -6.60714865e-01 1.63576350e-01
-8.00453842e-01 -1.11063540e+00 -1.11565125e+00 -1.09577191e+00
8.29437554e-01 -3.06153148e-01 8.92830729e-01 -1.09633029e-01
-3.81402254e-01 -2.39285693e-01 -3.23874146e-01 -4.62722629e-01
-4.46545541e-01 4.02033985e-01 -6.63270712e-01 6.12436309e-02
-1.02848090e-01 -6.01125121e-01 -9.77015316e-01 4.97005612e-01
-1.04730058e+00 4.66611236e-01 9.01172578e-01 1.20243478e+00
6.93363249e-01 -1.41016031e-02 6.22409940e-01 -1.02807963e+00
6.70087636e-01 -4.67524260e-01 -5.33655643e-01 1.10651851e-01
-3.81467819e-01 -2.42120214e-02 8.83100450e-01 -2.54919738e-01
-9.49307442e-01 1.51853785e-01 -7.48600900e-01 -3.66908431e-01
1.46953657e-01 4.15739089e-01 -3.71276110e-01 -7.67041966e-02
4.33468252e-01 4.97914672e-01 1.63118958e-01 -2.11264178e-01
1.54279619e-01 5.11081338e-01 6.82243109e-01 -3.13013852e-01
6.06222034e-01 2.65976250e-01 -7.64880851e-02 -1.54428408e-01
-7.92928219e-01 7.50016347e-02 -2.42582366e-01 -1.59248009e-01
1.36280394e+00 -9.20297801e-01 -7.30390072e-01 9.56289411e-01
-9.33885157e-01 -5.60669601e-01 -3.06532085e-01 4.00800526e-01
-6.36428535e-01 -1.42837062e-01 -9.61504638e-01 -3.38445269e-02
-7.37091959e-01 -1.57185578e+00 9.10023391e-01 5.89901507e-01
4.91612822e-01 -9.39476550e-01 -5.56059405e-02 6.28772914e-01
6.24474585e-01 5.88675261e-01 8.05996895e-01 -7.72152543e-01
-5.59618056e-01 -9.56562981e-02 -2.83772796e-01 5.68098843e-01
-3.71242873e-02 -5.59363306e-01 -8.22801352e-01 -4.05513793e-01
1.38491467e-01 -3.52941394e-01 1.08394754e+00 5.08966208e-01
1.76319182e+00 -2.22606584e-01 -5.20139098e-01 9.55086231e-01
1.45121992e+00 4.46974963e-01 1.02492976e+00 3.02114069e-01
8.96365166e-01 1.40759060e-02 2.52092462e-02 -2.93577891e-02
3.81790161e-01 4.81715292e-01 7.55016506e-01 -5.56973040e-01
-1.68801248e-01 -2.95533210e-01 2.23313794e-01 7.68478215e-01
3.58159021e-02 -2.40173206e-01 -9.47033525e-01 4.21307206e-01
-1.65127909e+00 -8.44103277e-01 9.30770636e-02 1.76226687e+00
9.31255996e-01 4.11233157e-01 -2.88210288e-02 -8.57633129e-02
6.95347190e-01 3.00618764e-02 -8.07412863e-01 -4.00317609e-02
1.24910483e-02 6.30685866e-01 6.18909061e-01 9.82942507e-02
-1.25838053e+00 1.00642824e+00 5.46949577e+00 1.27087498e+00
-1.35679412e+00 5.10077000e-01 1.02588308e+00 2.86285877e-01
-2.67873287e-01 -4.65101391e-01 -3.55374962e-01 8.24852049e-01
6.51292443e-01 8.54332149e-02 3.25863361e-02 6.74428463e-01
-1.97142139e-01 2.53787160e-01 -7.47429371e-01 6.80896521e-01
4.80376277e-03 -1.16902411e+00 -4.27908897e-02 -8.91685039e-02
6.53027534e-01 2.34531179e-01 2.82962978e-01 3.84130597e-01
4.89983648e-01 -1.54993606e+00 6.81262612e-01 3.54030222e-01
1.04017687e+00 -8.79385114e-01 9.63760316e-01 2.84936309e-01
-1.11500216e+00 1.61672980e-02 -2.32332110e-01 5.01870453e-01
1.56722009e-01 5.09348869e-01 -7.37035334e-01 8.07052255e-01
5.76955616e-01 6.41307175e-01 -5.14237821e-01 1.01359200e+00
-5.25544345e-01 6.77653134e-01 -9.44002345e-02 1.55076146e-01
5.07617652e-01 -2.56674230e-01 2.02192843e-01 1.17365026e+00
3.67470264e-01 1.06736913e-01 2.12087870e-01 9.04288173e-01
-4.09633726e-01 6.88096881e-02 -4.36341822e-01 1.55755579e-01
5.97916432e-02 1.34330523e+00 -9.41227198e-01 -2.18303725e-01
-1.85134292e-01 9.92950261e-01 1.57714725e-01 4.70925495e-02
-1.35893297e+00 -5.65217137e-01 2.09457278e-01 -9.12956166e-05
3.39800328e-01 2.52951294e-01 -3.11864138e-01 -1.09437799e+00
-4.79591414e-02 -6.73105597e-01 2.63274610e-01 -6.54146075e-01
-1.21462357e+00 1.07645214e+00 -2.88874984e-01 -1.28383362e+00
3.05998530e-02 -4.27968085e-01 -9.18962538e-01 7.77292311e-01
-1.41334569e+00 -1.57154489e+00 -4.95643497e-01 1.01130033e+00
3.91589373e-01 -3.92657608e-01 6.80580735e-01 4.33726668e-01
-7.15307295e-01 8.81074190e-01 9.00477991e-02 6.57346487e-01
3.41235220e-01 -1.11263609e+00 3.43917251e-01 7.77841151e-01
-2.83663243e-01 6.70079142e-03 2.58394659e-01 -4.91116613e-01
-7.77151346e-01 -1.50318420e+00 7.72330686e-02 6.68072775e-02
6.15221739e-01 -4.22766298e-01 -6.88894808e-01 7.57362545e-01
3.57160240e-01 1.53130442e-01 4.69340295e-01 -5.68249583e-01
-6.40737414e-02 -2.18349293e-01 -1.33091772e+00 4.36449617e-01
1.00225580e+00 -2.69507408e-01 -4.08941686e-01 4.61075515e-01
9.32237566e-01 -9.53809798e-01 -8.60943794e-01 7.10673153e-01
2.17415497e-01 -7.98847854e-01 1.09571648e+00 -3.08047026e-01
8.19240510e-01 -1.35976627e-01 3.47736090e-01 -1.55108130e+00
-1.96377411e-01 -3.22751313e-01 6.76270425e-01 9.48503554e-01
4.01943445e-01 -8.51653099e-01 7.78180361e-01 7.45667368e-02
-6.07180595e-01 -1.17910612e+00 -1.12875080e+00 -6.64139986e-01
5.24414957e-01 -5.66927232e-02 6.84503496e-01 6.66949451e-01
-3.41042161e-01 4.47135091e-01 -1.72844023e-01 -1.19435959e-01
4.69158262e-01 -1.94010526e-01 2.60870814e-01 -8.95034790e-01
-2.33397871e-01 -7.27818966e-01 -4.78660315e-01 -8.61375988e-01
2.74022967e-01 -1.23943698e+00 5.70408888e-02 -1.51397288e+00
3.13136637e-01 -5.12888372e-01 -6.32160246e-01 5.34488499e-01
-1.02725685e-01 2.51208663e-01 -2.35237903e-03 -1.35182396e-01
-3.44216168e-01 8.09442103e-01 1.80713582e+00 -6.47162616e-01
1.78882673e-01 1.48745626e-01 -7.23009169e-01 6.53471828e-01
9.32003677e-01 -4.31832790e-01 -5.96628129e-01 -5.42393267e-01
-2.20140457e-01 2.96107680e-01 5.25938988e-01 -1.16166627e+00
2.91525364e-01 6.98225945e-02 4.64306742e-01 -2.29532242e-01
1.37030303e-01 -6.96924865e-01 2.51669943e-01 8.16306472e-01
-3.74853700e-01 7.78104663e-02 1.61789954e-01 2.25276127e-01
-3.65940005e-01 -4.07187976e-02 1.04813337e+00 -2.79683739e-01
-3.33852589e-01 8.42469752e-01 -9.93344188e-02 3.26622099e-01
1.40126610e+00 1.51203945e-01 -3.37432325e-01 5.26415035e-02
-6.90671086e-01 3.39583635e-01 1.22926153e-01 2.88772285e-01
7.07247615e-01 -1.24335945e+00 -8.45625639e-01 1.83048964e-01
-8.84651393e-02 3.96512061e-01 3.99923742e-01 8.48196566e-01
-7.52942085e-01 2.07087860e-01 -4.92864519e-01 -4.59058791e-01
-8.81033301e-01 4.81437683e-01 7.53900707e-01 -7.91536272e-01
-8.17048907e-01 1.10818732e+00 6.66851819e-01 -3.18558574e-01
8.65695029e-02 -3.20058554e-01 -2.82889426e-01 -3.15718859e-01
1.40562594e-01 -1.99090078e-01 4.51557934e-02 -4.20414418e-01
-2.61584371e-01 3.74573350e-01 -2.83162892e-01 -1.48735836e-01
1.30840194e+00 3.62218469e-01 -1.29425049e-01 -2.83797402e-02
1.18231606e+00 -4.84288782e-01 -1.28074312e+00 -1.88834921e-01
-4.91593629e-01 1.08892329e-01 -2.56280564e-02 -9.80217159e-01
-1.82138300e+00 9.42370415e-01 8.10758412e-01 4.35798466e-02
1.42607927e+00 1.83862597e-02 1.27228594e+00 -2.12628886e-01
3.09227973e-01 -8.33776593e-01 1.82482660e-01 3.76153648e-01
7.82871246e-01 -1.20839846e+00 -6.65260077e-01 -1.28977239e-01
-6.01567864e-01 8.17896247e-01 8.72202396e-01 -2.71519870e-01
6.97064698e-01 4.41559881e-01 4.39712331e-02 -2.24587381e-01
-3.67780447e-01 -2.17786103e-01 6.43822327e-02 4.35021162e-01
2.46865779e-01 2.87536711e-01 -4.65239346e-01 9.66993868e-01
-5.04210293e-01 1.45061076e-01 1.87530309e-01 5.61059594e-01
-8.07368904e-02 -1.08457112e+00 9.56904739e-02 5.11210799e-01
-7.40752816e-01 -1.28175452e-01 3.44239213e-02 7.60028064e-01
2.56725699e-01 5.65899432e-01 3.91236320e-02 -5.23368359e-01
2.85160929e-01 -2.76966125e-01 3.82281214e-01 -3.43337536e-01
-8.51711631e-01 -2.80961711e-02 -3.81208092e-01 -4.77672964e-01
-1.22368909e-01 -3.20343405e-01 -1.41566896e+00 2.07642987e-02
-1.92346692e-01 2.09704116e-01 5.13808966e-01 1.03013182e+00
-2.46937331e-02 1.27319455e+00 6.34552479e-01 -2.99443781e-01
-4.43252265e-01 -1.05621302e+00 -4.65822250e-01 4.65448052e-01
1.03006884e-01 -4.49989080e-01 1.40884772e-01 -8.00467655e-02] | [14.536612510681152, -2.4438610076904297] |
761d2a2b-4f39-4876-adff-2ad9dc1728d5 | black-box-bayesian-inference-for-economic | 2202.00625 | null | https://arxiv.org/abs/2202.00625v1 | https://arxiv.org/pdf/2202.00625v1.pdf | Black-box Bayesian inference for economic agent-based models | Simulation models, in particular agent-based models, are gaining popularity in economics. The considerable flexibility they offer, as well as their capacity to reproduce a variety of empirically observed behaviours of complex systems, give them broad appeal, and the increasing availability of cheap computing power has made their use feasible. Yet a widespread adoption in real-world modelling and decision-making scenarios has been hindered by the difficulty of performing parameter estimation for such models. In general, simulation models lack a tractable likelihood function, which precludes a straightforward application of standard statistical inference techniques. Several recent works have sought to address this problem through the application of likelihood-free inference techniques, in which parameter estimates are determined by performing some form of comparison between the observed data and simulation output. However, these approaches are (a) founded on restrictive assumptions, and/or (b) typically require many hundreds of thousands of simulations. These qualities make them unsuitable for large-scale simulations in economics and can cast doubt on the validity of these inference methods in such scenarios. In this paper, we investigate the efficacy of two classes of black-box approximate Bayesian inference methods that have recently drawn significant attention within the probabilistic machine learning community: neural posterior estimation and neural density ratio estimation. We present benchmarking experiments in which we demonstrate that neural network based black-box methods provide state of the art parameter inference for economic simulation models, and crucially are compatible with generic multivariate time-series data. In addition, we suggest appropriate assessment criteria for future benchmarking of approximate Bayesian inference procedures for economic simulation models. | ['Sebastian Schmon', 'J. Doyne Farmer', 'Patrick Cannon', 'Joel Dyer'] | 2022-02-01 | null | null | null | null | ['density-ratio-estimation'] | ['methodology'] | [-1.08769417e-01 -4.02646989e-01 -3.40748727e-01 -1.69344321e-01
-3.23424459e-01 -4.00758237e-01 1.19541717e+00 2.14331537e-01
-6.81743979e-01 1.17173135e+00 -3.53282422e-01 -8.72142851e-01
-6.29756093e-01 -8.33478272e-01 -5.68369210e-01 -7.63571978e-01
-3.66033107e-01 6.49987876e-01 1.21935392e-02 2.20442072e-01
2.03699112e-01 6.57205284e-01 -1.59579539e+00 -5.43417037e-01
8.55786800e-01 8.36004496e-01 7.54411370e-02 4.58101064e-01
1.02474220e-01 5.70373535e-01 -5.30786216e-01 -4.07523006e-01
-7.55263120e-02 -3.36196274e-01 -1.30940944e-01 -2.54314631e-01
-3.69442940e-01 -6.38327777e-01 -2.60030389e-01 9.20521915e-01
5.24015725e-01 3.29727829e-01 1.18010688e+00 -1.42976296e+00
-2.32465670e-01 4.49033022e-01 -4.20667827e-01 2.49307156e-01
2.77881056e-01 4.24143791e-01 6.57281160e-01 -2.97095537e-01
2.18998209e-01 1.32042134e+00 7.06849217e-01 7.44292140e-02
-1.61669660e+00 -6.04448318e-01 -1.69718098e-02 1.85071584e-02
-1.52135944e+00 -4.36995208e-01 5.02406240e-01 -4.31114107e-01
1.18796039e+00 7.47248679e-02 8.15079153e-01 1.11620998e+00
5.31851470e-01 4.00307685e-01 1.42864192e+00 -4.79889959e-01
7.97830939e-01 1.31496638e-01 -3.75420123e-01 2.93741286e-01
5.85039735e-01 7.82652199e-01 -1.32555112e-01 -5.88079274e-01
1.14393902e+00 -6.86970130e-02 1.21217377e-01 -3.17114294e-01
-9.40543652e-01 1.17015970e+00 4.39695865e-02 7.99877644e-02
-6.15788341e-01 6.05850816e-01 4.63769794e-01 3.23754027e-02
6.65968001e-01 7.16011226e-02 -4.09028292e-01 -2.53286332e-01
-1.01201153e+00 8.51817131e-01 1.06007791e+00 4.77539271e-01
3.17951739e-01 2.75189489e-01 2.82462984e-01 6.59792781e-01
5.38001776e-01 7.26421237e-01 2.68274188e-01 -1.43385458e+00
3.79459187e-02 8.09903815e-02 4.55230027e-01 -9.52313781e-01
-2.95580924e-01 -2.00995371e-01 -1.02282679e+00 5.87726474e-01
8.21427882e-01 -3.42910469e-01 -5.11397362e-01 1.87554669e+00
2.31316850e-01 2.88961112e-01 -4.07771952e-02 5.27900875e-01
-3.41062471e-02 8.13052893e-01 3.63907516e-01 -4.94370937e-01
1.13854349e+00 -1.36709675e-01 -4.65045184e-01 1.24526903e-01
3.44014913e-01 -5.53344190e-01 7.78695762e-01 2.87720650e-01
-1.16702497e+00 -8.65828544e-02 -7.41018653e-01 7.59451926e-01
-4.78803128e-01 -4.49508518e-01 8.99519086e-01 8.46214592e-01
-8.14551771e-01 7.69829094e-01 -1.15679026e+00 -2.76021153e-01
2.08714321e-01 2.95808256e-01 3.04795727e-02 1.56929240e-01
-1.17540896e+00 1.21054053e+00 2.63678968e-01 1.80949077e-01
-6.31044686e-01 -4.86657679e-01 -8.38490844e-01 4.64426845e-01
3.44743073e-01 -7.99648941e-01 1.43679249e+00 -8.02002430e-01
-1.65516794e+00 2.01158389e-01 1.41055986e-01 -5.99160433e-01
7.00971842e-01 4.04472649e-01 -3.11282009e-01 -1.01693399e-01
-1.70251966e-01 3.78328413e-01 6.95278168e-01 -9.93991435e-01
-4.81429249e-01 1.97211821e-02 -2.41523162e-02 1.12247117e-01
2.56593972e-01 1.88650772e-01 2.05713585e-01 -6.45995617e-01
-4.17326391e-01 -6.85992658e-01 -5.01665115e-01 -1.85299307e-01
1.02507278e-01 -3.15133899e-01 2.81014591e-01 -3.66896808e-01
1.27291787e+00 -1.61551046e+00 -3.37061852e-01 4.99346554e-01
-1.91884071e-01 6.90247715e-02 1.51710808e-01 7.02465594e-01
1.48510516e-01 1.09017819e-01 -4.43298846e-01 -3.08351009e-03
5.92154264e-01 2.36203283e-01 -2.62951046e-01 6.27761781e-01
4.49192263e-02 8.51330459e-01 -7.72548258e-01 -5.85429728e-01
6.98839068e-01 3.46386015e-01 -4.27557170e-01 -6.53041527e-02
-1.83619022e-01 -8.04733578e-03 -4.55379426e-01 2.98961431e-01
4.20422584e-01 -1.61148012e-01 3.54386032e-01 3.43998998e-01
-1.26256496e-01 1.82862416e-01 -1.29637623e+00 8.89149129e-01
-5.40205479e-01 4.46372002e-01 -7.24979341e-02 -1.25446582e+00
4.49753910e-01 4.18165028e-01 3.03386986e-01 -5.77264369e-01
1.93546355e-01 3.05948526e-01 3.36719334e-01 -2.17534497e-01
1.61697611e-01 -5.23162127e-01 -2.13159412e-01 7.84613192e-01
-2.55565047e-02 -5.31004727e-01 4.04244721e-01 -1.70287862e-01
8.02367449e-01 4.28536445e-01 6.53378248e-01 -4.12311405e-01
4.32658121e-02 -8.67928118e-02 4.41224247e-01 1.04245090e+00
-1.61409289e-01 4.57947738e-02 6.04142249e-01 -3.36578846e-01
-1.34828210e+00 -1.31506169e+00 -4.69298810e-01 5.42194903e-01
-2.09604979e-01 8.44688043e-02 -7.97252357e-01 -2.02448949e-01
3.25222403e-01 8.96555603e-01 -4.35941041e-01 -9.02460795e-03
-3.45253199e-01 -1.47529185e+00 4.69222218e-01 5.23399174e-01
1.59471869e-01 -1.10617840e+00 -9.00733471e-01 5.83292186e-01
2.70273089e-01 -7.73008943e-01 2.73108304e-01 2.18708783e-01
-1.00568879e+00 -8.68027806e-01 -8.53449643e-01 -7.96639994e-02
3.59795511e-01 -2.86907822e-01 1.07030189e+00 -7.27680102e-02
-1.01954028e-01 3.34697455e-01 1.22002207e-01 -4.39562798e-01
-7.63664842e-01 -1.04164667e-01 4.60051507e-01 -4.58641440e-01
3.97486210e-01 -8.52528930e-01 -3.22199374e-01 2.60253996e-01
-8.96586239e-01 -2.65453547e-01 6.40806019e-01 8.20936203e-01
1.41144350e-01 3.24005723e-01 1.08088243e+00 -5.85438907e-01
1.01571715e+00 -6.85405850e-01 -1.28693473e+00 2.21620142e-01
-9.52840328e-01 8.23047757e-03 5.10194838e-01 -5.95299244e-01
-1.11937296e+00 -3.44364673e-01 -2.93802656e-02 -1.14239581e-01
-3.30800593e-01 1.06172812e+00 2.80745834e-01 2.15011552e-01
4.49379861e-01 -7.05809742e-02 2.63158590e-01 -2.61769027e-01
-4.92345393e-02 5.85178614e-01 3.08949560e-01 -8.30553353e-01
6.03116095e-01 9.89375412e-02 3.11019957e-01 -7.35773683e-01
-6.20595887e-02 1.47153229e-01 -2.12989762e-01 -2.59573787e-01
4.40483034e-01 -7.57561147e-01 -9.37242389e-01 5.37456751e-01
-7.66370118e-01 -7.46201336e-01 -1.91560075e-01 9.17365134e-01
-8.30891967e-01 2.28200942e-01 -5.78674138e-01 -1.43499982e+00
2.83969492e-01 -1.13333631e+00 4.63085383e-01 2.95372367e-01
-5.42730451e-01 -1.44374359e+00 1.05855152e-01 -3.04328978e-01
5.62648237e-01 2.48426363e-01 1.14820349e+00 -4.88380730e-01
-3.02931160e-01 -3.23863089e-01 -2.35568300e-01 5.13833389e-02
-9.78458226e-02 4.87494618e-01 -8.13521087e-01 -7.45471418e-02
-1.52295649e-01 -1.92177862e-01 4.72282797e-01 1.12410271e+00
8.63867700e-01 -2.07400292e-01 -2.94726223e-01 2.36773953e-01
1.41295218e+00 2.44613379e-01 3.44859600e-01 2.50541717e-01
-8.95418301e-02 6.18796229e-01 2.58332819e-01 6.23805285e-01
2.54424751e-01 6.58029556e-01 1.74786210e-01 1.37381300e-01
5.15013456e-01 -1.96880057e-01 1.62556469e-01 7.26413608e-01
-2.38203496e-01 -3.80903214e-01 -8.96164358e-01 4.35840100e-01
-2.00087357e+00 -1.19054973e+00 1.13137342e-01 2.45938373e+00
7.63237596e-01 4.27314669e-01 3.63653123e-01 -2.55147293e-02
6.95226371e-01 -6.04671426e-02 -6.24557376e-01 -7.04675853e-01
7.84424096e-02 2.27953598e-01 6.09473765e-01 4.07782108e-01
-7.59861708e-01 3.91014248e-01 7.43022251e+00 1.28450453e+00
-6.18537068e-01 -2.27177981e-02 9.32425857e-01 -3.74703109e-02
-1.18972830e-01 6.69433922e-02 -3.98469120e-01 7.62402296e-01
1.57952178e+00 -4.92630899e-01 6.47213697e-01 5.93996048e-01
6.33943319e-01 -9.33307886e-01 -9.96068776e-01 6.81915343e-01
-5.74998975e-01 -1.20881939e+00 -1.88628763e-01 3.53375733e-01
6.32896662e-01 -1.61501300e-02 5.67735769e-02 4.62631881e-01
8.13984036e-01 -1.32207286e+00 6.93316996e-01 5.36332011e-01
6.34024382e-01 -1.05099344e+00 9.57602084e-01 6.16471112e-01
-8.36836576e-01 -1.26611935e-02 -4.67574775e-01 -5.64166903e-01
4.42726523e-01 7.57091463e-01 -3.69344294e-01 2.12822407e-01
4.56780821e-01 1.49349958e-01 -5.36210183e-03 1.24583125e+00
8.84219015e-04 8.89039576e-01 -9.26806986e-01 -3.37847322e-01
3.67340982e-01 -3.76788378e-01 3.06829691e-01 1.05880427e+00
4.42492902e-01 -1.08775958e-01 -2.85052538e-01 1.18812132e+00
2.03633130e-01 -3.53263140e-01 -6.77379251e-01 -1.03311799e-02
8.19682121e-01 8.53650451e-01 -1.02954543e+00 -2.42241174e-01
-4.71247375e-01 1.18807852e-01 4.52695042e-02 6.64939761e-01
-9.45622742e-01 -1.82377040e-01 5.74601769e-01 -6.02528341e-02
3.58094908e-02 -3.40606570e-01 -1.07394822e-01 -9.00272906e-01
-1.93880469e-01 -7.55655229e-01 1.61836326e-01 -6.11951470e-01
-1.36600351e+00 -1.00250557e-01 8.71035516e-01 -7.64973938e-01
-1.11015975e+00 -7.00074852e-01 -7.02225029e-01 9.98752832e-01
-1.33244503e+00 -5.84382534e-01 4.61144239e-01 7.78296813e-02
2.72589087e-01 -9.94331986e-02 8.12515199e-01 4.04400080e-02
-5.65952241e-01 1.31920397e-01 7.61401415e-01 -1.71207577e-01
1.91581205e-01 -1.17610431e+00 4.74371582e-01 4.77442414e-01
-6.78533539e-02 4.77057427e-01 9.85508084e-01 -5.82834601e-01
-1.05374038e+00 -5.21878421e-01 6.84486389e-01 -1.81778178e-01
9.22415376e-01 -2.50253648e-01 -6.71977162e-01 4.39297467e-01
-8.00075307e-02 -2.43424475e-01 3.32531840e-01 1.10262856e-01
4.17464599e-02 1.31860569e-01 -1.15344656e+00 6.97864056e-01
4.19307947e-01 -4.83213484e-01 -4.90782708e-01 -2.39672209e-03
6.93547502e-02 2.06543565e-01 -8.54489982e-01 3.14136297e-01
8.74161780e-01 -9.74834561e-01 1.05276084e+00 -4.31891888e-01
2.81054646e-01 3.36272153e-03 -7.12028593e-02 -1.32998049e+00
-1.17358118e-01 -5.53320348e-01 -2.30940580e-01 1.14653194e+00
3.62914920e-01 -1.10669029e+00 4.15833026e-01 8.59978020e-01
5.17206848e-01 -7.12627947e-01 -1.22414398e+00 -1.04228449e+00
5.83513260e-01 -5.89478731e-01 7.65227377e-01 7.88434982e-01
1.24504335e-01 -1.15733989e-01 -1.17042260e-02 -1.34430110e-01
8.35612714e-01 3.63140702e-02 4.51289982e-01 -1.31668377e+00
-3.94138545e-01 -1.00578320e+00 -3.28469813e-01 -6.16445243e-01
4.55329657e-01 -2.78393686e-01 -4.79575768e-02 -1.16491449e+00
2.83834785e-01 -6.04819059e-01 -7.36166984e-02 -5.77673167e-02
-1.06232926e-01 1.62757978e-01 -1.67348400e-01 1.34077400e-01
-1.34033598e-02 5.39331436e-01 5.53870440e-01 3.72083515e-01
6.86522797e-02 2.59872973e-01 -1.39419377e-01 9.51314330e-01
8.90472233e-01 -5.29509187e-01 -4.24924493e-01 2.26312671e-02
3.99001926e-01 3.93360019e-01 8.68616104e-01 -6.44665062e-01
4.31750678e-02 -6.54502571e-01 5.21171927e-01 -2.24312752e-01
2.51368791e-01 -7.90470779e-01 3.97496969e-01 4.84906882e-01
-1.12198234e-01 3.09433073e-01 2.79227614e-01 6.37943566e-01
2.16745920e-02 -4.75122422e-01 6.98001206e-01 -1.54381841e-01
-2.63107151e-01 -1.76306311e-02 -1.09295571e+00 -1.04463786e-01
7.94833899e-01 -2.26985440e-01 -3.81840654e-02 -7.15645254e-01
-4.78398710e-01 4.00394015e-02 5.68520248e-01 -2.57195860e-01
1.49604931e-01 -1.10628200e+00 -7.31093645e-01 -2.03710660e-01
-3.83228421e-01 -2.66863942e-01 -6.33586524e-03 8.58955085e-01
-5.24586260e-01 4.12554353e-01 -1.19404480e-01 -4.49352264e-01
-4.40013140e-01 6.18887484e-01 4.16285634e-01 -3.74486238e-01
-4.72062141e-01 1.67824715e-01 1.34030893e-01 -4.64396298e-01
-2.64076665e-02 -4.37056392e-01 2.64938146e-01 -1.73852623e-01
3.02095473e-01 6.09378099e-01 -2.84979761e-01 -4.01123017e-01
-1.51868194e-01 8.34592655e-02 4.98767704e-01 -6.54960990e-01
1.33961034e+00 -2.27085859e-01 1.30152225e-01 6.14647329e-01
7.07492173e-01 -5.10856807e-01 -1.51036859e+00 6.10531196e-02
1.39548331e-01 -2.01075464e-01 2.68869817e-01 -7.75957763e-01
-6.02334380e-01 8.21930230e-01 1.93048090e-01 5.95736980e-01
8.49766314e-01 -3.96868765e-01 3.50535512e-01 4.28236067e-01
3.73286784e-01 -1.15576613e+00 -6.62753880e-01 7.81982243e-02
3.73245239e-01 -1.08377945e+00 3.50933999e-01 8.02010968e-02
-1.01452433e-01 8.78809035e-01 1.51588330e-02 -3.56014997e-01
7.65467882e-01 5.48457980e-01 -3.59556794e-01 1.20197557e-01
-9.97998536e-01 8.59192479e-03 -7.87564442e-02 6.54349506e-01
1.82793111e-01 2.54633427e-01 -3.66002828e-01 4.46673691e-01
-4.61176112e-02 2.63954878e-01 5.00731885e-01 8.45871985e-01
-2.61914849e-01 -8.95970702e-01 -4.16870117e-01 6.82729423e-01
-5.33803225e-01 -1.33585185e-02 2.19052106e-01 1.32307339e+00
-4.72713709e-01 9.49993074e-01 3.20826739e-01 4.14995551e-01
-1.39532417e-01 2.12622315e-01 5.50689757e-01 -6.76689148e-02
-4.02628154e-01 2.60080665e-01 2.13349059e-01 -2.84371853e-01
-4.52066392e-01 -1.07880259e+00 -6.59473121e-01 -9.75375891e-01
-5.88839948e-01 2.86557168e-01 8.09250891e-01 1.24018073e+00
-9.92074907e-02 2.35827416e-01 2.93433040e-01 -1.22264564e+00
-1.15565729e+00 -9.88208532e-01 -9.03505802e-01 -9.89504606e-02
5.91024868e-02 -1.01244855e+00 -6.57187104e-01 -1.72886223e-01] | [6.561779975891113, 3.9990882873535156] |
c45ca0b0-a79e-4d3b-8499-610be3645107 | interest-point-detection-based-on-adaptive | 1901.00031 | null | http://arxiv.org/abs/1901.00031v1 | http://arxiv.org/pdf/1901.00031v1.pdf | Interest Point Detection based on Adaptive Ternary Coding | In this paper, an adaptive pixel ternary coding mechanism is proposed and a
contrast invariant and noise resistant interest point detector is developed on
the basis of this mechanism. Every pixel in a local region is adaptively
encoded into one of the three statuses: bright, uncertain and dark. The blob
significance of the local region is measured by the spatial distribution of the
bright and dark pixels. Interest points are extracted from this blob
significance measurement. By labeling the statuses of ternary bright,
uncertain, and dark, the proposed detector shows more robustness to image noise
and quantization errors. Moreover, the adaptive strategy for the ternary
cording, which relies on two thresholds that automatically converge to the
median of the local region in measurement, enables this coding to be
insensitive to the image local contrast. As a result, the proposed detector is
invariant to illumination changes. The state-of-the-art results are achieved on
the standard datasets, and also in the face recognition application. | ['Xudong Jiang', 'Kim-Hui Yap', 'Zhenwei Miao'] | 2018-12-31 | null | null | null | null | ['interest-point-detection'] | ['computer-vision'] | [ 4.16765600e-01 -3.40512812e-01 -1.11417763e-01 -3.30595434e-01
-6.48682117e-02 -2.35341579e-01 5.51022649e-01 9.83790532e-02
-4.92513388e-01 5.89066982e-01 -3.87220740e-01 2.41239779e-02
-8.46164450e-02 -1.00587261e+00 -2.74355412e-01 -1.10547030e+00
2.41494790e-01 -1.10715203e-01 5.85066020e-01 1.68953449e-01
5.38075209e-01 8.77858877e-01 -1.95140183e+00 2.05283642e-01
6.16906464e-01 1.59529805e+00 1.94136888e-01 4.77880448e-01
-2.78795779e-01 3.86931390e-01 -6.53593838e-01 1.85621418e-02
4.00409430e-01 -6.89119935e-01 -1.60122156e-01 2.87983030e-01
5.29050291e-01 -1.92806780e-01 1.52839541e-01 1.57832706e+00
5.12393117e-01 -1.47237852e-01 7.93103218e-01 -9.96260345e-01
-5.63952565e-01 1.35583445e-01 -8.06198299e-01 3.36324334e-01
1.20995969e-01 9.42814909e-03 4.77066964e-01 -1.08478904e+00
5.64723492e-01 1.09345901e+00 2.83929050e-01 4.21246499e-01
-1.16306531e+00 -6.92583621e-01 -1.42154068e-01 5.42322278e-01
-1.75388348e+00 -3.30503196e-01 9.77149904e-01 -3.34492028e-01
2.55230606e-01 3.35677296e-01 5.74298501e-01 3.77456456e-01
4.95381087e-01 3.15096706e-01 1.47157502e+00 -8.11127841e-01
5.96848309e-01 1.69669881e-01 -3.36908661e-02 8.67410600e-01
3.80165666e-01 7.31085762e-02 -4.06808019e-01 1.09814003e-01
7.54532754e-01 1.25480756e-01 -4.05614376e-01 -5.16103208e-01
-8.16923141e-01 4.28578824e-01 4.00183707e-01 7.37279534e-01
-3.12654674e-01 -4.62331660e-02 2.74397768e-02 -1.17011867e-01
9.40936357e-02 -2.00675011e-01 -1.76133763e-03 1.91048831e-01
-1.07358575e+00 -4.03546989e-01 4.28730637e-01 6.50034308e-01
8.82882178e-01 1.25064924e-01 -1.92493752e-01 6.82493508e-01
5.50872087e-01 8.06998909e-01 4.86576170e-01 -9.16086376e-01
7.04414994e-02 8.34283948e-01 5.44044822e-02 -1.21142209e+00
-2.21883938e-01 -3.06627870e-01 -9.72485304e-01 9.24710155e-01
4.85410064e-01 9.53384936e-02 -1.05381203e+00 1.35644436e+00
4.93909448e-01 5.54913431e-02 2.56699882e-02 7.58664608e-01
6.83316112e-01 6.35253429e-01 -2.27160558e-01 -4.95979816e-01
1.58778512e+00 -3.16498667e-01 -1.07863307e+00 5.85811287e-02
-2.87069213e-02 -8.41610312e-01 8.07135701e-01 5.13745844e-01
-1.05744195e+00 -8.86912346e-01 -1.40485990e+00 1.79833651e-01
-4.95449215e-01 4.27958280e-01 1.05654253e-02 1.06133270e+00
-1.11013877e+00 2.47206420e-01 -6.58208668e-01 -3.65262449e-01
9.70773622e-02 2.45803639e-01 -6.96646422e-02 7.85669312e-03
-8.55195045e-01 6.13023698e-01 4.06951427e-01 4.23994124e-01
-3.51905435e-01 -6.57947659e-02 -5.97238541e-01 1.77737877e-01
-1.62759706e-01 -1.42293841e-01 4.45233464e-01 -1.06465423e+00
-1.47709095e+00 1.00285935e+00 -3.11904103e-01 -3.39645147e-01
4.87069607e-01 4.93369550e-01 -5.57210803e-01 5.65600872e-01
-1.06894024e-01 4.65322614e-01 1.12183917e+00 -1.21957791e+00
-9.23958898e-01 -5.47968686e-01 -6.16726935e-01 7.22262412e-02
-5.29519916e-01 1.11198485e-01 -4.96295273e-01 -4.41920489e-01
5.22668839e-01 -4.50819910e-01 1.56935990e-01 4.36803311e-01
9.08820704e-03 -1.56239554e-01 1.18919456e+00 -2.82622546e-01
1.29394960e+00 -2.57908773e+00 -3.84174377e-01 7.08537340e-01
-9.24116150e-02 1.28460705e-01 2.68728703e-01 -2.24300385e-01
1.17368745e-02 -7.05683380e-02 -2.77088135e-01 1.17007129e-01
-2.17542425e-01 7.96567872e-02 1.64031938e-01 7.12299585e-01
-9.57132969e-03 1.48999155e-01 -5.22699594e-01 -1.01951349e+00
5.75199246e-01 5.19471943e-01 -7.32760727e-02 -7.34869540e-02
3.47506553e-01 8.84450004e-02 -3.48918617e-01 8.35937738e-01
1.24930596e+00 5.46852909e-02 1.66981831e-01 -4.99368191e-01
-5.60708165e-01 -4.06801969e-01 -1.51560318e+00 9.07400250e-01
-8.48767757e-02 6.41612887e-01 2.90895998e-01 -7.48099744e-01
1.46196723e+00 2.04717785e-01 3.66849869e-01 -9.19863582e-01
4.66234624e-01 3.17644238e-01 -1.88773379e-01 -2.19420433e-01
4.59383875e-01 1.15824506e-01 3.63553703e-01 1.28385220e-02
-3.09278429e-01 -6.80659264e-02 3.23215276e-01 -2.56249577e-01
6.88742340e-01 -1.22776031e-01 5.79375446e-01 -5.24691582e-01
1.04976070e+00 -4.48182225e-01 6.51435852e-01 4.98691767e-01
-5.37697315e-01 5.24371266e-01 3.55067760e-01 -1.79873168e-01
-8.25170934e-01 -1.14734459e+00 -7.79627204e-01 5.30888319e-01
6.88192189e-01 6.25141058e-03 -9.71975207e-01 -3.24728042e-01
-1.28408596e-01 4.57266778e-01 -5.88492334e-01 8.57932679e-03
-5.25064349e-01 -5.76971889e-01 8.16390291e-02 2.62768596e-01
1.05802536e+00 -9.49239910e-01 -1.00874019e+00 -1.18734846e-02
5.24292588e-02 -1.07378173e+00 -1.97801813e-01 5.43302149e-02
-8.20490837e-01 -9.96561527e-01 -6.49036527e-01 -9.86123681e-01
9.21601295e-01 2.69817382e-01 5.19960344e-01 7.78939649e-02
-3.41340542e-01 2.38344967e-01 -4.74502087e-01 1.10634856e-01
-3.95856559e-01 -5.86041093e-01 -2.68834412e-01 5.14696062e-01
3.50982666e-01 -1.57141566e-01 -9.17655170e-01 6.79519594e-01
-8.09590101e-01 -3.69112790e-01 5.80507576e-01 7.28923976e-01
8.64217043e-01 2.78951913e-01 3.74710709e-02 -3.28692853e-01
2.78368890e-01 6.33574352e-02 -8.84314835e-01 3.92281651e-01
-7.98003674e-01 9.13740769e-02 5.16828775e-01 -3.25343937e-01
-1.21245384e+00 1.98935837e-01 1.77126199e-01 2.99454108e-03
-2.63630390e-01 3.17612477e-02 -2.76847988e-01 -4.33327347e-01
6.78422630e-01 3.13741565e-01 -9.17634219e-02 -1.95239797e-01
1.72230184e-01 8.92988861e-01 7.16018617e-01 -1.31150499e-01
6.89663768e-01 7.74123669e-01 4.29616392e-01 -1.10277069e+00
1.76316619e-01 -4.18881357e-01 -5.33754289e-01 -5.52720964e-01
9.24451292e-01 -5.13029575e-01 -7.84801722e-01 6.14660978e-01
-1.01321507e+00 1.73208281e-01 -2.56049544e-01 3.59446108e-01
-3.15644234e-01 8.29027891e-01 -5.42967796e-01 -9.16597188e-01
-4.34773117e-01 -1.28572536e+00 7.95796275e-01 6.77931547e-01
1.84140414e-01 -7.13188350e-01 -5.31619549e-01 -1.13343172e-01
2.61017084e-01 2.65818924e-01 7.97958553e-01 -2.16398686e-02
-7.96350062e-01 -3.13954204e-01 -4.65083390e-01 5.42934179e-01
2.62417465e-01 4.51404005e-01 -7.73918509e-01 -1.14374846e-01
1.80068970e-01 1.67058319e-01 7.22001791e-01 4.51563478e-01
1.09498191e+00 -1.43746790e-02 -2.77454108e-01 4.88158554e-01
1.63608325e+00 6.05454624e-01 9.07543600e-01 4.09427315e-01
2.98559684e-02 2.81842202e-01 9.07293141e-01 5.63406765e-01
-1.74150124e-01 6.46091342e-01 5.65476239e-01 -3.53913963e-01
-3.15899342e-01 1.60272971e-01 3.26385140e-01 1.14468977e-01
8.73164013e-02 -2.37731934e-01 -5.86082995e-01 3.16584766e-01
-1.31625819e+00 -1.08694541e+00 -2.13558018e-01 2.30279255e+00
7.57398903e-01 1.54179484e-01 -4.47111219e-01 6.13082111e-01
1.10948241e+00 1.96083561e-01 -2.21308216e-01 -5.66443026e-01
-5.55905104e-01 3.32479686e-01 8.02777052e-01 3.86999816e-01
-1.07026613e+00 6.84544206e-01 5.93695593e+00 1.02861929e+00
-1.44370663e+00 -1.47910058e-01 7.41103232e-01 5.06405473e-01
6.62375912e-02 -2.53712863e-01 -9.97519553e-01 7.65339196e-01
3.59557241e-01 1.53155163e-01 1.60250306e-01 8.57147157e-01
3.57143253e-01 -7.87424922e-01 -5.28986573e-01 1.14904511e+00
3.67472917e-01 -8.12916636e-01 1.00438660e-02 3.89507040e-02
7.65929461e-01 -4.97707546e-01 5.36551714e-01 -4.85157162e-01
-2.20560282e-01 -5.52571356e-01 8.83839905e-01 4.81279045e-01
8.21815312e-01 -6.85970008e-01 8.82356346e-01 5.57109565e-02
-1.34806752e+00 -3.50102544e-01 -3.05898488e-01 3.04285198e-01
-1.13601856e-01 6.30029976e-01 -8.81892383e-01 1.02626070e-01
6.78734899e-01 3.29321802e-01 -7.51815975e-01 1.04129648e+00
-3.13075334e-01 4.38263267e-01 -4.57795978e-01 -2.06929207e-01
-5.49626239e-02 -4.72835124e-01 3.46941829e-01 1.15640128e+00
5.70044935e-01 7.37482905e-02 -1.18753701e-01 7.86932766e-01
1.93983644e-01 3.36161077e-01 -2.49565959e-01 4.14342523e-01
6.50293708e-01 1.31465375e+00 -1.13569033e+00 -4.17169094e-01
-2.46872351e-01 1.00982547e+00 -3.39618355e-01 3.78452033e-01
-5.82660079e-01 -6.87678576e-01 1.55597940e-01 8.60531628e-02
5.54273784e-01 -2.45639309e-01 -7.14322209e-01 -5.29318154e-01
7.05200285e-02 -5.39921403e-01 3.47299248e-01 -8.02838922e-01
-6.82668507e-01 5.62063396e-01 -1.21098392e-01 -1.41978085e+00
-8.00160170e-02 -8.84687364e-01 -5.02685845e-01 8.40147257e-01
-1.51277387e+00 -6.50569439e-01 -5.87814629e-01 7.21297443e-01
2.19105333e-01 1.07271597e-02 3.92993659e-01 2.50429988e-01
-3.93869281e-01 6.12366974e-01 3.01500946e-01 1.21730618e-01
5.54627597e-01 -1.00565541e+00 -2.44047254e-01 1.22775328e+00
-2.38968626e-01 5.11209130e-01 7.37718165e-01 -4.55752283e-01
-7.73166299e-01 -6.40246868e-01 6.99747562e-01 2.79832929e-01
2.94151753e-01 -1.54462948e-01 -7.20539451e-01 -5.82415424e-02
9.11572725e-02 2.69799829e-01 2.54414111e-01 -5.48790038e-01
-1.76003247e-01 -6.53572619e-01 -1.63949692e+00 4.40393925e-01
4.86537486e-01 -2.64723688e-01 -3.37337345e-01 -5.61446957e-02
-2.37806998e-02 -8.44016373e-02 -7.93773890e-01 4.92692381e-01
6.90168142e-01 -1.31468451e+00 9.64708805e-01 4.95059699e-01
-6.98979869e-02 -8.73582542e-01 -2.40009055e-01 -6.50041521e-01
-2.95624703e-01 -1.05331264e-01 1.78392947e-01 1.21745431e+00
3.24300975e-01 -5.94586909e-01 9.23960388e-01 1.00173794e-01
1.66187271e-01 -3.77255857e-01 -1.23306549e+00 -5.60210228e-01
-5.03529549e-01 1.12637833e-01 4.55490917e-01 5.42477369e-01
-2.19539523e-01 -1.53562233e-01 -9.01155919e-03 2.77238905e-01
8.21429729e-01 1.61335275e-01 3.82493794e-01 -1.17405891e+00
1.72635555e-01 -5.60218513e-01 -9.85364199e-01 -1.01472592e+00
-3.53315562e-01 -3.81429970e-01 1.84076771e-01 -1.27583671e+00
3.22769247e-02 -3.79765958e-01 -4.97001678e-01 2.39423126e-01
-1.48549434e-02 7.38965631e-01 1.15612112e-01 1.53906405e-01
-2.27478072e-01 1.76305085e-01 1.04367840e+00 -2.54167140e-01
-1.32319972e-01 -3.03946930e-04 8.92402306e-02 6.65194392e-01
8.51539254e-01 -8.09790790e-02 -1.13981795e-02 9.51031074e-02
-1.37660921e-01 -1.87742129e-01 3.07621658e-01 -1.35683489e+00
3.46833527e-01 -1.52665868e-01 7.34490216e-01 -9.07959998e-01
4.83187973e-01 -1.27872598e+00 -8.07099640e-02 7.01503217e-01
9.70289763e-03 -4.61688973e-02 -1.25049964e-01 4.47909266e-01
-2.28927791e-01 -4.78063315e-01 1.49372149e+00 1.30118832e-01
-9.34737623e-01 -1.27913594e-01 -3.10800135e-01 -5.35435259e-01
1.16982591e+00 -1.12398839e+00 -2.71936417e-01 -1.11749172e-01
-4.81197298e-01 -3.30496848e-01 5.64621866e-01 -1.29272580e-01
8.54144633e-01 -1.24426043e+00 -4.05416191e-01 7.80079126e-01
1.01213150e-01 -5.23237109e-01 3.61831188e-01 7.57066667e-01
-8.44713688e-01 6.75726309e-02 -4.97303575e-01 -8.52099478e-01
-1.74041784e+00 5.88662863e-01 4.17179674e-01 2.57529944e-01
-1.93527535e-01 4.85216558e-01 -1.78422973e-01 3.89523983e-01
3.37079048e-01 -5.37156999e-01 -5.35476685e-01 -7.73869827e-02
6.96950793e-01 5.73340952e-01 7.86385220e-03 -9.70982850e-01
-2.19237372e-01 1.15446353e+00 2.91661173e-01 -1.78733945e-01
6.04874194e-01 -3.52779955e-01 -3.96407217e-01 1.46137059e-01
1.15407062e+00 2.97990829e-01 -1.05083394e+00 6.08514389e-03
-1.09289289e-01 -7.55947769e-01 2.02318400e-01 -7.30609655e-01
-1.10822940e+00 8.78531516e-01 1.39674199e+00 4.36124325e-01
1.60870111e+00 -4.08824354e-01 3.52783293e-01 2.32524499e-02
3.09953690e-01 -1.20591998e+00 -1.20336324e-01 9.82394665e-02
5.60644448e-01 -8.25975955e-01 8.81169960e-02 -5.37143946e-01
-2.23570183e-01 1.42657316e+00 3.74426961e-01 1.47126138e-01
6.23127878e-01 4.62964892e-01 3.83210555e-02 7.46661350e-02
-1.31132111e-01 -2.97270238e-01 1.75954834e-01 7.10906744e-01
2.14050889e-01 -1.00005083e-02 -8.61765623e-01 -2.21813381e-01
1.30229294e-01 -4.74449545e-02 2.78951406e-01 7.51184106e-01
-1.07612848e+00 -8.54211390e-01 -1.02783990e+00 -3.13863419e-02
-3.45276207e-01 1.47713125e-01 -2.13618249e-01 7.15240002e-01
6.73148334e-01 1.15194225e+00 3.44455630e-01 -1.39323235e-01
2.05171615e-01 -1.92551523e-01 4.32646036e-01 5.46360314e-02
-1.63790941e-01 2.98280835e-01 -5.21739841e-01 -3.72973949e-01
-4.65162992e-01 -6.61180019e-01 -1.58122361e+00 -1.78441793e-01
-3.75234723e-01 2.19040468e-01 1.01402175e+00 4.04338211e-01
4.89325747e-02 1.58651456e-01 8.48365426e-01 -4.51808870e-01
-3.23385119e-01 -7.09988475e-01 -8.01195085e-01 2.48221576e-01
2.49570534e-01 -5.97803354e-01 -5.58057487e-01 1.18224137e-01] | [10.936622619628906, -2.343397855758667] |
aa233c60-cd41-4f43-84f3-d4258c5d0cea | automatic-audio-captioning-using-attention | 2201.12352 | null | https://arxiv.org/abs/2201.12352v1 | https://arxiv.org/pdf/2201.12352v1.pdf | Automatic Audio Captioning using Attention weighted Event based Embeddings | Automatic Audio Captioning (AAC) refers to the task of translating audio into a natural language that describes the audio events, source of the events and their relationships. The limited samples in AAC datasets at present, has set up a trend to incorporate transfer learning with Audio Event Detection (AED) as a parent task. Towards this direction, in this paper, we propose an encoder-decoder architecture with light-weight (i.e. with lesser learnable parameters) Bi-LSTM recurrent layers for AAC and compare the performance of two state-of-the-art pre-trained AED models as embedding extractors. Our results show that an efficient AED based embedding extractor combined with temporal attention and augmentation techniques is able to surpass existing literature with computationally intensive architectures. Further, we provide evidence of the ability of the non-uniform attention weighted encoding generated as a part of our model to facilitate the decoder glance over specific sections of the audio while generating each token. | ['Sunil Kumar Kopparapu', 'Rupayan Chakraborty', 'Swapnil Bhosale'] | 2022-01-28 | null | null | null | null | ['audio-captioning'] | ['audio'] | [ 4.53257799e-01 3.90817076e-01 3.78592223e-01 -1.18345223e-01
-1.21730959e+00 -3.59532416e-01 6.58046424e-01 1.18654341e-01
-3.20694268e-01 5.33478260e-01 1.08137476e+00 -1.83506608e-01
2.24609688e-01 -5.19462228e-01 -7.85916746e-01 -3.41312885e-01
-2.22323373e-01 2.59143203e-01 -4.21844609e-02 -1.32247373e-01
-8.61569792e-02 1.70282528e-01 -1.53974283e+00 7.42754459e-01
3.27290148e-01 1.03524184e+00 2.67755061e-01 1.06184101e+00
6.05033524e-02 9.53201532e-01 -8.61976326e-01 -3.84774208e-01
-3.12885255e-01 -5.85813403e-01 -9.10709143e-01 -1.30902052e-01
3.20811659e-01 -4.47494656e-01 -5.61792910e-01 3.74614477e-01
8.94756377e-01 1.06182322e-01 7.13875532e-01 -1.01242805e+00
-6.80191338e-01 9.46838677e-01 5.22427866e-03 3.49507779e-01
6.64158285e-01 1.23643577e-01 1.30952489e+00 -9.64057326e-01
5.83694935e-01 9.31964040e-01 7.00670540e-01 6.52834475e-01
-1.17703283e+00 -4.94511813e-01 -5.58224022e-02 4.64146554e-01
-1.14334297e+00 -6.78194761e-01 1.03479290e+00 -2.98698276e-01
1.49320638e+00 2.45881334e-01 6.22339547e-01 1.70146537e+00
8.59982669e-02 7.98300743e-01 6.22435153e-01 -6.47520006e-01
1.77874729e-01 6.02531135e-02 -2.60660648e-01 2.37897515e-01
-2.73419052e-01 1.02984486e-02 -1.00206757e+00 5.37176467e-02
4.25809175e-01 -4.80011433e-01 -2.99464256e-01 7.74652511e-02
-1.18465316e+00 6.06541395e-01 3.42784703e-01 4.86203820e-01
-7.50773072e-01 4.66318458e-01 8.95119965e-01 2.34632120e-01
6.67914391e-01 6.31566584e-01 -3.68973404e-01 -5.96554637e-01
-1.14754272e+00 1.11380786e-01 6.47720098e-01 8.28161061e-01
1.75171286e-01 5.29063642e-01 -7.15037286e-01 7.43107021e-01
2.85312891e-01 -1.73399355e-02 6.34780049e-01 -5.33414364e-01
6.40874743e-01 1.87814966e-01 -1.10374587e-02 -5.02374947e-01
-1.61160275e-01 -6.38724566e-01 -4.99045521e-01 1.09317340e-01
4.57487628e-02 -4.09094512e-01 -8.89570355e-01 1.93862915e+00
-4.21799123e-02 4.16459590e-01 1.31501317e-01 7.14701355e-01
5.73137581e-01 1.12095976e+00 4.37316805e-01 1.42880186e-01
1.61892712e+00 -9.13610935e-01 -9.53632772e-01 -3.26200902e-01
5.07046402e-01 -9.15966094e-01 1.16544759e+00 2.27578059e-01
-1.34590769e+00 -6.42658353e-01 -1.14444184e+00 -3.82911414e-01
-3.97700101e-01 3.62432659e-01 5.04406393e-01 3.01013112e-01
-1.16204631e+00 4.66820538e-01 -6.87084615e-01 -1.50695816e-01
3.02712381e-01 1.87779665e-01 -2.31275201e-01 3.93453628e-01
-1.60043561e+00 1.05889499e+00 4.48765516e-01 1.96019150e-02
-1.34772062e+00 -8.96979868e-01 -9.45803165e-01 4.69599664e-01
-1.49352953e-01 -8.20755720e-01 1.58018625e+00 -1.00466311e+00
-1.79041052e+00 6.40789270e-01 -2.43224949e-01 -9.95304048e-01
2.18675166e-01 -5.85999191e-01 -5.88345468e-01 3.35766405e-01
-7.26428404e-02 9.61211622e-01 1.12144887e+00 -9.80312586e-01
-5.66774368e-01 1.50273621e-01 -1.03192374e-01 2.27444082e-01
-5.64689755e-01 1.75852478e-01 1.50830066e-02 -9.88952875e-01
-6.73856795e-01 -7.86840558e-01 2.28560790e-01 -1.80689648e-01
-3.08690876e-01 -2.63985723e-01 7.61626840e-01 -1.14904988e+00
1.56382513e+00 -2.43574572e+00 3.58470440e-01 -2.24690467e-01
-5.17468564e-02 3.35282087e-01 -3.21240902e-01 8.21137786e-01
-4.53481019e-01 -8.39626137e-03 -2.00312287e-01 -8.35332811e-01
3.60129237e-01 -9.58866924e-02 -8.68892908e-01 -2.01752316e-02
8.06509137e-01 7.69005120e-01 -8.26515555e-01 -1.41640022e-01
1.08616620e-01 1.07952392e+00 -5.58117986e-01 4.56370741e-01
-2.78837889e-01 2.59831607e-01 -1.19409569e-01 2.34256461e-01
-2.96736713e-02 1.05401635e-01 -3.04383844e-01 -2.13868171e-01
-2.37898096e-01 1.19707108e+00 -7.60150075e-01 2.03873253e+00
-7.70960748e-01 1.18089950e+00 -7.49272108e-02 -5.41713059e-01
6.26643002e-01 1.15496826e+00 1.26524448e-01 -5.14648139e-01
1.70430779e-01 3.35634351e-01 1.35864720e-01 -3.71234059e-01
5.39234877e-01 -1.26624212e-01 3.31895687e-02 5.49763560e-01
5.77707589e-01 -7.58397626e-03 1.17517747e-01 2.36186236e-02
1.16970456e+00 2.97611594e-01 -1.76164880e-02 1.81315482e-01
4.63546753e-01 -3.35788935e-01 -7.86484554e-02 4.28618163e-01
9.90121365e-02 8.75024438e-01 2.27621332e-01 -1.12467706e-01
-1.33007479e+00 -8.80024076e-01 1.10896669e-01 1.24144828e+00
-7.72248864e-01 -5.42526722e-01 -9.37324107e-01 -1.90187767e-01
-4.46865141e-01 1.07871628e+00 -6.02601230e-01 -4.80947137e-01
-6.19332731e-01 -2.10532457e-01 1.02956307e+00 6.38507724e-01
1.57106146e-01 -1.43182707e+00 -7.31542945e-01 7.75669396e-01
-3.54597747e-01 -1.16976249e+00 -5.84506750e-01 4.59234178e-01
-6.68280363e-01 -3.05730313e-01 -1.05426824e+00 -9.10377681e-01
1.62033692e-01 -4.71735746e-01 1.10831356e+00 -5.65619409e-01
-1.69377044e-01 3.68245959e-01 -5.87637126e-01 -8.13982010e-01
-4.82969075e-01 4.69689488e-01 -1.62119955e-01 2.43760914e-01
3.34536761e-01 -9.02880251e-01 -4.77155834e-01 -3.06981683e-01
-9.62221324e-01 3.11615407e-01 7.10317612e-01 5.34347177e-01
3.40997577e-01 -4.96191591e-01 6.87688768e-01 -2.03841537e-01
1.01226234e+00 -4.23273087e-01 2.49057878e-02 -2.09569961e-01
-2.80453861e-01 2.86906958e-01 5.44220984e-01 -6.40486598e-01
-1.07885635e+00 -1.26534086e-02 -4.04096663e-01 -3.47313821e-01
-2.24329427e-01 4.11587596e-01 -5.93266275e-04 4.80954736e-01
5.19915760e-01 3.40723991e-01 -3.05928022e-01 -5.77679634e-01
3.42088550e-01 8.52014363e-01 4.56622392e-01 -3.67293000e-01
6.51973248e-01 2.14158267e-01 -2.81265557e-01 -5.89432895e-01
-7.94946313e-01 -1.95763633e-01 -2.69247860e-01 -3.14765155e-01
1.04270852e+00 -1.21719527e+00 -2.36066669e-01 2.38648281e-01
-1.53235555e+00 -3.98160309e-01 -6.74137235e-01 6.95680857e-01
-8.04864347e-01 -1.55112937e-01 -8.59566092e-01 -8.64937782e-01
-6.45172119e-01 -9.01492238e-01 1.37176991e+00 -1.71385735e-01
-8.03999543e-01 -7.48944819e-01 1.95985824e-01 1.02355562e-01
6.31661475e-01 1.99348599e-01 9.66887712e-01 -7.83589065e-01
-3.26059759e-01 -1.86392888e-01 5.21567054e-02 4.38302666e-01
-1.59974061e-02 -1.78447053e-01 -1.52649319e+00 1.14377425e-03
-4.47342247e-02 -2.47635022e-01 7.17821300e-01 2.28579655e-01
1.06296051e+00 -4.69921082e-01 7.97763690e-02 3.08657318e-01
9.31652665e-01 3.08607846e-01 7.79176414e-01 2.60254920e-01
4.11913037e-01 4.30928290e-01 2.10086763e-01 4.89168704e-01
1.79805741e-01 7.92524338e-01 2.75889724e-01 -2.13553503e-01
-7.86558747e-01 -5.81863582e-01 8.72354329e-01 1.22313523e+00
5.22195408e-03 -3.99269730e-01 -7.59509921e-01 8.50872874e-01
-1.45857954e+00 -8.92655492e-01 2.07284108e-01 1.84504950e+00
1.15067923e+00 2.64650762e-01 -3.19768302e-02 5.53444564e-01
5.88711381e-01 2.19280720e-01 -2.29874551e-01 -8.97587597e-01
1.08580850e-01 6.62970364e-01 5.30903321e-03 4.75601196e-01
-8.30843389e-01 6.79361463e-01 6.01375103e+00 7.13243663e-01
-1.13866103e+00 2.37580612e-01 2.87985712e-01 -2.45574534e-01
-2.50461400e-01 -2.53177464e-01 -5.59967637e-01 4.43431139e-01
1.76126575e+00 -2.41972343e-03 4.60547686e-01 5.08728325e-01
3.61303210e-01 4.46394891e-01 -1.54444861e+00 8.18699062e-01
1.33133456e-01 -1.16389811e+00 4.58539575e-01 -5.83736338e-02
3.00505877e-01 -1.03846028e-01 2.92509049e-01 4.79175061e-01
-1.81234807e-01 -1.01759684e+00 1.12487972e+00 4.44731832e-01
9.00512516e-01 -6.39440596e-01 6.02580965e-01 -2.31556356e-01
-1.33174479e+00 -1.47225307e-02 1.11144878e-01 -3.29545528e-01
4.76005048e-01 2.37303808e-01 -1.28195977e+00 4.86601949e-01
5.34944594e-01 6.01714015e-01 -3.64916801e-01 1.00771284e+00
-2.72050589e-01 9.94207323e-01 -1.71415344e-01 -2.99391393e-02
5.46893239e-01 3.78765404e-01 8.01147223e-01 1.59991634e+00
6.38939559e-01 -4.32008147e-01 -6.70803308e-01 8.32599938e-01
-2.69383520e-01 1.07877240e-01 -4.35041279e-01 -4.37675983e-01
3.72209638e-01 8.32291663e-01 -2.37748757e-01 -2.49148652e-01
-2.26747558e-01 1.14506400e+00 1.36263013e-01 4.12502348e-01
-9.28867757e-01 -7.33173907e-01 6.49959564e-01 1.28416806e-01
5.26096940e-01 -1.40226811e-01 -8.35101232e-02 -8.07050526e-01
1.12960055e-01 -7.57927656e-01 2.18681887e-01 -1.37553895e+00
-9.73030746e-01 9.23202038e-01 -1.30631119e-01 -1.32936966e+00
-6.66805685e-01 -4.98512626e-01 -7.24537730e-01 9.65963840e-01
-1.67437589e+00 -1.14757276e+00 7.19374865e-02 4.72965389e-01
9.24475253e-01 -7.00539127e-02 1.23774755e+00 6.57782018e-01
-3.32132459e-01 4.69708622e-01 -3.45260918e-01 2.61069655e-01
7.38010883e-01 -1.15127468e+00 5.01872480e-01 7.82073677e-01
5.22655606e-01 3.21597904e-01 9.88697767e-01 -2.27475405e-01
-1.02881920e+00 -1.14772868e+00 1.34440637e+00 -3.99215132e-01
7.89017260e-01 -6.39714777e-01 -8.33752751e-01 8.03651750e-01
8.40414226e-01 -4.69671816e-01 7.94542909e-01 -1.28110468e-01
-2.86730766e-01 -1.93084314e-01 -5.92303157e-01 7.10478842e-01
7.61988282e-01 -1.07959545e+00 -9.47224617e-01 3.86714265e-02
1.12042761e+00 -3.34421039e-01 -7.46081114e-01 1.53494641e-01
3.63161236e-01 -5.24999261e-01 9.37677681e-01 -5.77688992e-01
6.81314826e-01 -2.29309008e-01 9.50926170e-02 -1.35698140e+00
-1.52421236e-01 -1.00066698e+00 -4.44905102e-01 1.56693006e+00
7.26930261e-01 -2.16832370e-01 3.34398478e-01 2.43990421e-01
-7.47135162e-01 -4.58082914e-01 -1.25407338e+00 -4.48146313e-01
-3.13471287e-01 -7.44512677e-01 5.06595790e-01 4.54306930e-01
8.40006918e-02 6.03217781e-01 -5.35501838e-01 1.20349728e-01
-2.38114148e-02 -4.32823420e-01 2.67727792e-01 -1.07503903e+00
-2.63710439e-01 -4.40781415e-01 -4.01265591e-01 -6.05114758e-01
2.58990288e-01 -8.31384003e-01 2.91947033e-02 -1.66035652e+00
-3.59425038e-01 2.21925780e-01 -4.28382158e-01 5.71603656e-01
1.60669461e-01 2.64683932e-01 1.71650723e-01 -1.38399482e-01
-2.05684707e-01 7.33805239e-01 8.35680187e-01 -1.62645459e-01
-2.06281766e-01 -1.44628555e-01 -4.53299254e-01 4.34514046e-01
8.06226194e-01 -6.55755103e-01 -4.13399309e-01 -6.55432463e-01
2.98211247e-01 -9.09110829e-02 3.26206803e-01 -1.35258389e+00
2.15584144e-01 6.77340806e-01 2.54805684e-01 -2.18816444e-01
7.44437695e-01 -8.38205755e-01 2.00284719e-01 4.63000804e-01
-8.87413919e-01 1.20547481e-01 5.78511059e-01 3.50838095e-01
-6.20290041e-01 -2.54390359e-01 2.90059507e-01 9.28400680e-02
-4.37989652e-01 -5.60884736e-03 -8.41948450e-01 9.76465549e-03
6.34530008e-01 -1.63057745e-01 7.51933828e-02 -6.89012229e-01
-8.77117753e-01 -2.81452239e-01 -2.91999251e-01 5.83157599e-01
5.95188916e-01 -1.40132713e+00 -9.77861166e-01 8.70004371e-02
7.95023292e-02 -3.48504901e-01 1.56475380e-01 4.89756227e-01
-1.63699746e-01 7.29137957e-01 -2.87271917e-01 -9.95667875e-02
-1.12113178e+00 2.78760672e-01 1.99660435e-01 -3.26452255e-01
-6.41810298e-01 8.18442166e-01 -1.96835235e-01 2.77782530e-01
5.69706738e-01 -5.11021674e-01 -2.66361475e-01 4.35798824e-01
5.29239237e-01 2.24257469e-01 2.56389409e-01 -4.88865942e-01
-2.06819043e-01 2.21531242e-01 2.35214382e-02 -6.05968177e-01
1.52018619e+00 4.41586226e-02 2.73146868e-01 7.24706888e-01
1.20411909e+00 -1.41909905e-02 -1.29639471e+00 -8.80500823e-02
2.27675512e-02 2.28128806e-01 2.08452061e-01 -9.00502801e-01
-6.20185733e-01 1.39113426e+00 6.33506596e-01 2.28812143e-01
1.23428297e+00 -4.01526056e-02 1.01588047e+00 2.87383720e-02
-1.57394007e-01 -1.12832439e+00 1.50463119e-01 5.34216940e-01
1.15865982e+00 -6.07079744e-01 -5.25506675e-01 1.20787323e-02
-7.14986384e-01 1.26033676e+00 2.94301808e-01 -1.30139336e-01
2.44172439e-01 3.12249243e-01 -1.71763748e-01 3.40099707e-02
-1.00616729e+00 -3.10146898e-01 5.53793192e-01 5.82045496e-01
6.78789079e-01 -2.46694982e-01 1.59197599e-02 5.20145714e-01
-4.31844622e-01 6.27006516e-02 5.31156003e-01 6.82550788e-01
-1.38778567e-01 -1.07339752e+00 -8.37688744e-02 1.64533466e-01
-7.59790182e-01 -4.83808368e-01 -4.16731298e-01 5.79128921e-01
-3.84650826e-02 8.13687146e-01 2.74068445e-01 -2.46868044e-01
3.81119996e-01 7.25563765e-01 1.88161880e-01 -7.21340179e-01
-9.95670974e-01 -2.40774248e-02 3.34582001e-01 -3.37676018e-01
-3.65970850e-01 -4.75874335e-01 -1.03566766e+00 5.28558791e-01
-3.73802222e-02 1.78706869e-01 7.67674744e-01 8.61973226e-01
7.35312700e-01 1.30652964e+00 3.04012328e-01 -1.07609999e+00
-3.84914309e-01 -1.33437157e+00 -2.14797884e-01 2.24218816e-01
6.77552164e-01 -1.97323471e-01 -3.24102134e-01 3.95871490e-01] | [15.278390884399414, 4.9274678230285645] |
8f9b87e8-cc39-4367-abed-6361ce1c864f | autoavatar-autoregressive-neural-fields-for | 2203.13817 | null | https://arxiv.org/abs/2203.13817v1 | https://arxiv.org/pdf/2203.13817v1.pdf | AutoAvatar: Autoregressive Neural Fields for Dynamic Avatar Modeling | Neural fields such as implicit surfaces have recently enabled avatar modeling from raw scans without explicit temporal correspondences. In this work, we exploit autoregressive modeling to further extend this notion to capture dynamic effects, such as soft-tissue deformations. Although autoregressive models are naturally capable of handling dynamics, it is non-trivial to apply them to implicit representations, as explicit state decoding is infeasible due to prohibitive memory requirements. In this work, for the first time, we enable autoregressive modeling of implicit avatars. To reduce the memory bottleneck and efficiently model dynamic implicit surfaces, we introduce the notion of articulated observer points, which relate implicit states to the explicit surface of a parametric human body model. We demonstrate that encoding implicit surfaces as a set of height fields defined on articulated observer points leads to significantly better generalization compared to a latent representation. The experiments show that our approach outperforms the state of the art, achieving plausible dynamic deformations even for unseen motions. https://zqbai-jeremy.github.io/autoavatar | ['Shunsuke Saito', 'Ping Tan', 'Michael Zollhöfer', 'Javier Romero', 'Timur Bagautdinov', 'Ziqian Bai'] | 2022-03-25 | null | null | null | null | ['3d-human-dynamics', '3d-human-reconstruction', 'human-dynamics'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 8.49577785e-02 4.53110099e-01 2.32771318e-02 -2.49069676e-01
-5.81779242e-01 -4.31875676e-01 6.14692926e-01 -3.66637081e-01
-7.59252459e-02 4.39107686e-01 5.77473104e-01 1.01743333e-01
2.29841154e-02 -8.05546343e-01 -9.61527765e-01 -6.01699769e-01
-2.39092439e-01 8.35997224e-01 2.27974087e-01 -2.34664813e-01
-2.53972821e-02 4.94810522e-01 -1.33611941e+00 1.34547189e-01
5.78229010e-01 6.95903301e-01 -8.09171982e-03 6.70866370e-01
-9.69919711e-02 5.47388613e-01 -2.82764852e-01 -3.45746756e-01
1.93952501e-01 -1.00587085e-01 -8.55217516e-01 9.81941167e-03
7.28198767e-01 -4.36110884e-01 -5.85336745e-01 6.14368975e-01
3.12264144e-01 3.42255592e-01 8.57255697e-01 -8.49756539e-01
-6.69604182e-01 2.10689470e-01 -4.25349236e-01 -2.76916027e-01
1.56336352e-01 1.16839647e-01 7.98136950e-01 -7.71926224e-01
7.82934904e-01 1.32624900e+00 8.42928231e-01 8.57588410e-01
-1.46243382e+00 -3.90103042e-01 4.44847673e-01 -1.51514541e-02
-1.17189348e+00 -5.54413557e-01 1.00473142e+00 -6.19714200e-01
8.57832015e-01 4.46306229e-01 8.16075921e-01 1.32975185e+00
2.55768239e-01 9.01502609e-01 7.78272390e-01 -2.47890651e-01
2.80948132e-01 -5.06851315e-01 2.58133262e-01 9.72339928e-01
-2.56839097e-01 4.21728045e-02 -5.50187230e-01 -3.35049152e-01
1.34749889e+00 6.55691326e-02 -1.68841839e-01 -7.98057020e-01
-1.10491467e+00 7.41997242e-01 5.51815808e-01 -1.65811136e-01
-5.95255733e-01 7.66974926e-01 2.12852940e-01 -7.16067255e-02
6.26406908e-01 2.84223616e-01 -3.38817298e-01 -2.49562547e-01
-8.30338180e-01 3.74559999e-01 7.88843155e-01 8.05364788e-01
6.11542165e-01 3.81050944e-01 2.54434735e-01 7.42676258e-01
3.81756425e-01 5.45873821e-01 3.59499872e-01 -1.25275767e+00
4.21040766e-02 2.80522674e-01 -5.44500798e-02 -9.47557211e-01
-3.46885562e-01 -1.62053451e-01 -9.25986826e-01 3.24869961e-01
3.87944907e-01 1.69041514e-01 -1.34504032e+00 1.78888834e+00
3.25870365e-01 4.32233661e-01 -1.30096048e-01 1.09501815e+00
6.72507286e-01 6.55559003e-01 2.20990226e-01 1.26387961e-02
1.24908662e+00 -1.03863537e+00 -4.43461061e-01 -1.44445270e-01
3.52917939e-01 -3.98226738e-01 1.10627151e+00 4.02819276e-01
-1.12905788e+00 -2.73645222e-01 -6.52643204e-01 -2.25401506e-01
1.19820103e-01 -1.48197547e-01 9.32606995e-01 2.74215162e-01
-9.46983814e-01 6.08008206e-01 -1.71831942e+00 -3.18861187e-01
2.05545966e-02 4.07192379e-01 -3.57576877e-01 4.59261328e-01
-9.48828101e-01 6.48006439e-01 -1.54436633e-01 3.00784022e-01
-1.01449680e+00 -7.18012512e-01 -1.06793499e+00 -1.56689569e-01
3.29403907e-01 -1.02268648e+00 1.16973007e+00 -1.05060720e+00
-1.98376095e+00 7.52598226e-01 -2.92118430e-01 -5.03494382e-01
7.37945974e-01 -6.92520559e-01 8.50525349e-02 2.58763991e-02
-4.59057301e-01 6.91945672e-01 7.65649080e-01 -1.16391516e+00
4.71227318e-01 -4.25007313e-01 1.33287892e-01 7.53668249e-02
-2.13299677e-01 -3.42270762e-01 -5.76275706e-01 -7.58356810e-01
3.88686359e-01 -1.34928071e+00 -5.74938715e-01 3.33158910e-01
-3.58394861e-01 1.01652317e-01 7.57190883e-01 -8.50856066e-01
9.57376122e-01 -1.85875058e+00 6.00238144e-01 2.46617317e-01
2.21305847e-01 1.20642483e-02 -5.50106317e-02 9.40504596e-02
1.58091590e-01 1.40764117e-02 -4.15978581e-01 -4.00928974e-01
-3.99070866e-02 5.76289058e-01 -5.25044024e-01 4.97018725e-01
1.17783919e-01 1.11872244e+00 -5.94031811e-01 -3.58381420e-01
2.14919761e-01 9.53222275e-01 -7.78757393e-01 1.40466347e-01
-4.95915949e-01 7.33154178e-01 -7.35662043e-01 4.60376263e-01
3.72537315e-01 -2.49134898e-01 1.00563265e-01 -1.39300764e-01
5.86376116e-02 3.01416665e-01 -9.54442918e-01 1.97919297e+00
-4.65058148e-01 3.57652903e-01 1.83335274e-01 -8.80317628e-01
9.49919522e-01 3.50444227e-01 6.52660847e-01 -2.43313789e-01
-8.35654791e-03 7.03401165e-03 -2.51304507e-01 -2.61388421e-01
5.55755019e-01 -1.62444204e-01 -1.70124680e-01 1.47990203e-02
-2.04940155e-01 -2.68176347e-01 -4.59591180e-01 -1.11316957e-01
9.44017768e-01 7.40620852e-01 7.87420571e-02 -2.50103086e-01
2.27935642e-01 -1.04505889e-01 5.70146799e-01 2.50492692e-01
1.56425506e-01 8.66187572e-01 1.55609816e-01 -8.24328780e-01
-1.07810736e+00 -1.48670995e+00 -6.30756794e-03 8.77786696e-01
1.05584338e-01 -3.78986478e-01 -6.19679868e-01 -2.96761155e-01
9.52921901e-03 4.55301225e-01 -8.66995513e-01 -7.91882277e-02
-1.01595497e+00 -3.77876878e-01 4.98699129e-01 8.59776258e-01
1.13111682e-01 -9.39819932e-01 -9.20500517e-01 2.62620896e-01
9.20336042e-03 -8.23076248e-01 -4.37371016e-01 -1.81532204e-01
-1.30716956e+00 -5.56722939e-01 -9.20782864e-01 -3.64763826e-01
5.95462084e-01 -4.37760592e-01 1.14403296e+00 5.75145446e-02
-4.44775224e-01 5.26957929e-01 -6.59382269e-02 -8.93606767e-02
-3.39005440e-01 1.78261343e-02 1.07751101e-01 -1.04974143e-01
-2.26457998e-01 -9.08304989e-01 -6.85192823e-01 2.94272989e-01
-7.09516704e-01 4.94284749e-01 3.45939040e-01 8.02882373e-01
1.08037829e+00 -8.80802572e-01 7.68015860e-03 -9.19560969e-01
5.47424555e-02 -1.40518203e-01 -4.89963591e-01 7.83156455e-02
-2.67058194e-01 5.51050484e-01 2.66847104e-01 -8.35225165e-01
-1.11277950e+00 3.92369330e-01 -2.68511176e-01 -9.48678434e-01
-1.15919456e-01 5.31475723e-01 4.74894233e-03 1.54862747e-01
5.18621027e-01 1.79065645e-01 1.12995200e-01 -7.33385444e-01
5.41867256e-01 1.57219350e-01 7.82070637e-01 -9.87614155e-01
6.55283034e-01 7.52359748e-01 2.54585117e-01 -1.02749920e+00
-4.99096841e-01 -9.94643643e-02 -8.89557719e-01 -1.93165287e-01
1.03256035e+00 -7.46321261e-01 -6.60174966e-01 3.20166945e-01
-1.13031209e+00 -8.91221166e-01 -3.47786278e-01 4.34426725e-01
-1.10507274e+00 5.49260020e-01 -8.76231015e-01 -9.74499941e-01
-4.16702569e-01 -1.05977690e+00 1.40146291e+00 -7.46109188e-02
-6.15171671e-01 -1.02002752e+00 3.01328719e-01 2.85818756e-01
2.61364818e-01 7.63805628e-01 6.83608770e-01 -6.60445243e-02
-7.88455069e-01 -9.83919278e-02 4.71535921e-01 -2.30186462e-01
-8.78746435e-02 2.05120966e-01 -8.42229724e-01 -1.31755099e-01
-2.90886313e-01 -1.87414318e-01 7.94928610e-01 5.53743839e-01
1.20301580e+00 -4.59623277e-01 -2.41116479e-01 9.11764681e-01
9.70871925e-01 -2.94493735e-01 8.91568601e-01 2.51340747e-01
8.71436477e-01 5.28476536e-01 3.15500587e-01 4.18082446e-01
3.63187104e-01 1.05205119e+00 5.01017153e-01 -8.37401971e-02
-1.62164077e-01 -3.74377400e-01 4.83983546e-01 9.36944366e-01
-7.02708185e-01 5.13485447e-02 -1.03871179e+00 4.76199180e-01
-1.79558969e+00 -8.10669661e-01 -3.22159231e-01 2.15978479e+00
7.72199750e-01 -8.82163718e-02 8.49669576e-02 -4.84701663e-01
2.38678366e-01 2.96166390e-01 -6.11420155e-01 -3.48223448e-01
1.55988291e-01 2.73607761e-01 3.03702474e-01 7.42627501e-01
-9.35445130e-01 1.15331161e+00 6.06131363e+00 2.46365070e-01
-1.28307879e+00 2.12648138e-01 3.89942050e-01 -2.73972332e-01
-4.95810777e-01 8.83937776e-02 -4.09630477e-01 1.38218492e-01
9.53855336e-01 7.66490325e-02 3.19246382e-01 9.25033212e-01
2.32618809e-01 3.59266609e-01 -9.91933644e-01 7.58030653e-01
-1.33553103e-01 -1.22290599e+00 1.52359545e-01 2.46331289e-01
6.00403249e-01 -1.36201009e-02 6.66574389e-02 1.40768930e-01
3.32478344e-01 -1.04633236e+00 9.84007776e-01 9.80340064e-01
9.12013352e-01 -4.35759634e-01 1.19511522e-01 3.82882833e-01
-1.33238518e+00 4.53442782e-01 -3.75126153e-01 -8.49986225e-02
4.15729463e-01 1.41683340e-01 -4.55930561e-01 2.63222426e-01
5.47493815e-01 5.59679747e-01 -2.40523174e-01 6.10774159e-01
-6.89406395e-02 6.89980328e-01 -6.00356996e-01 4.26880717e-01
-1.29436264e-02 -2.06774145e-01 7.25422680e-01 1.00466442e+00
3.84678483e-01 1.94652930e-01 2.24799201e-01 1.01582134e+00
1.56583741e-01 1.32528562e-02 -5.15645325e-01 4.69281450e-02
6.90958947e-02 9.70622003e-01 -5.08671105e-01 -2.38625482e-01
-1.40467003e-01 1.15421855e+00 3.97208691e-01 4.33410525e-01
-9.03192520e-01 4.06970590e-01 8.91526461e-01 4.94800299e-01
8.66732970e-02 -8.93448830e-01 -2.49993980e-01 -1.35995722e+00
-9.23776999e-02 -5.18262982e-01 3.23547661e-01 -7.62016058e-01
-9.98679996e-01 5.59833229e-01 1.32020384e-01 -1.00070000e+00
-6.01753891e-01 -4.86910969e-01 -4.63271767e-01 8.43701243e-01
-8.84867370e-01 -1.51315582e+00 -4.99596477e-01 4.13318753e-01
7.28332877e-01 3.49842846e-01 1.06936944e+00 -6.70207217e-02
-1.27525046e-01 4.81828004e-01 -2.12865636e-01 1.42435849e-01
4.34610277e-01 -1.10013342e+00 7.42337823e-01 6.70139074e-01
3.57177079e-01 7.79875696e-01 9.45010185e-01 -8.95159483e-01
-1.67873728e+00 -8.43639553e-01 1.13623746e-01 -5.36969006e-01
5.26188374e-01 -4.53296781e-01 -1.34212923e+00 9.86663103e-01
-2.41664365e-01 1.57408684e-01 4.58878428e-01 1.92000598e-01
-4.65509474e-01 3.26706231e-01 -5.67298412e-01 7.55506933e-01
1.33510852e+00 -5.03910542e-01 -5.74044943e-01 1.58344358e-02
6.25235379e-01 -7.80259907e-01 -1.21741807e+00 5.87946236e-01
7.95579672e-01 -5.13550401e-01 1.21991968e+00 -7.79232740e-01
4.60449159e-01 -1.54273778e-01 -5.86556084e-02 -1.00687623e+00
-2.22981766e-01 -7.44362056e-01 -7.29332805e-01 8.66126835e-01
1.82369307e-01 -4.22499239e-01 1.27190447e+00 1.14586663e+00
-2.27698311e-01 -8.51792872e-01 -9.06756759e-01 -8.30214679e-01
3.17674011e-01 -4.27816927e-01 4.17961597e-01 7.85397530e-01
-3.20732415e-01 -8.60798284e-02 -6.81339383e-01 3.49421471e-01
5.11418402e-01 1.88155830e-01 1.06990719e+00 -1.31459427e+00
-8.05319428e-01 -3.45288336e-01 -6.21424854e-01 -1.52327383e+00
5.02240479e-01 -7.17297673e-01 1.47031292e-01 -1.49355936e+00
3.44499387e-02 -5.66215038e-01 6.88935816e-02 5.21023214e-01
1.06320735e-02 3.06067228e-01 2.12084502e-01 5.01899600e-01
-2.70050913e-01 8.44285011e-01 1.30814183e+00 -2.82896608e-02
-3.15537512e-01 -5.64431511e-02 -1.61755145e-01 9.64793861e-01
4.93630230e-01 -1.88998342e-01 -3.64605159e-01 -7.34779179e-01
5.81493825e-02 4.03842539e-01 7.22812057e-01 -8.70620251e-01
3.71419213e-04 -3.01642537e-01 2.02487782e-01 -2.76841223e-01
9.68297780e-01 -6.14923060e-01 8.78274560e-01 4.08978611e-01
-3.44393313e-01 7.79725760e-02 1.67814836e-01 7.86053002e-01
-4.11096849e-02 1.51430205e-01 5.53772986e-01 -4.30137515e-02
-6.74717784e-01 4.94178444e-01 -2.63399631e-01 -1.81088939e-01
6.98741436e-01 -3.10659707e-01 -1.11695882e-02 -5.57856083e-01
-1.15881550e+00 -1.08746238e-01 7.99625695e-01 4.01382953e-01
6.86144173e-01 -1.08933914e+00 -5.49595952e-01 1.24734245e-01
-1.39904022e-01 3.41702014e-01 4.77150291e-01 8.93855631e-01
-7.25187719e-01 6.47604167e-02 -1.99739799e-01 -7.76943803e-01
-1.42236888e+00 3.62295806e-01 5.22071123e-01 -2.46895105e-01
-1.22767437e+00 7.29685485e-01 6.89229429e-01 -3.95150065e-01
2.33669914e-02 -4.55756307e-01 1.70146242e-01 -4.10351366e-01
1.13566950e-01 1.81069165e-01 -1.74772874e-01 -9.03796077e-01
-2.71613896e-01 8.75625670e-01 2.08684668e-01 -2.12747499e-01
1.50878000e+00 2.82198973e-02 1.64898336e-01 6.52126908e-01
8.84601295e-01 1.12950534e-01 -1.76484787e+00 -4.25542481e-02
-4.11306918e-02 -3.25348675e-01 -4.68471721e-02 -4.45259154e-01
-1.07645416e+00 1.12492037e+00 5.58715582e-01 -2.92368144e-01
8.20372224e-01 2.24941224e-01 8.07893336e-01 3.31936210e-01
5.70217729e-01 -5.88769853e-01 1.25225261e-01 6.29687428e-01
1.16389298e+00 -7.71366477e-01 -4.00608629e-02 -7.02657521e-01
-6.26821458e-01 1.22304809e+00 4.65030640e-01 -4.83420789e-01
6.04654670e-01 4.27349150e-01 1.42576739e-01 -2.31656253e-01
-7.57562101e-01 4.31378447e-02 5.72973311e-01 4.77133244e-01
5.24898052e-01 3.23070556e-01 1.40657583e-02 4.40091759e-01
-5.32303929e-01 -1.05293833e-01 2.98651785e-01 7.82827437e-01
-1.23134494e-01 -1.06212533e+00 -4.17820662e-01 1.88378707e-01
-3.55100751e-01 1.52206331e-01 -2.78749526e-01 8.03830266e-01
-4.19553995e-01 1.74261078e-01 3.53977680e-01 -2.58327574e-01
3.85684401e-01 1.87155813e-01 7.20805407e-01 -5.29777110e-01
-2.06098035e-01 3.00680667e-01 -6.03503212e-02 -7.72981763e-01
-2.40324259e-01 -7.50447333e-01 -1.53090739e+00 -1.39446810e-01
-1.89623311e-02 -8.68429765e-02 6.37061775e-01 6.53931439e-01
3.79964411e-01 5.16663313e-01 -1.48562804e-01 -1.29584968e+00
-5.04237831e-01 -7.60560513e-01 -2.35493228e-01 6.30315244e-01
2.42479995e-01 -1.01001298e+00 4.59111519e-02 4.77159888e-01] | [7.022848606109619, -1.254431962966919] |
3f5811f0-2f8d-4686-88a7-700276524cff | neighborhood-contrastive-learning-for-novel | 2106.10731 | null | https://arxiv.org/abs/2106.10731v1 | https://arxiv.org/pdf/2106.10731v1.pdf | Neighborhood Contrastive Learning for Novel Class Discovery | In this paper, we address Novel Class Discovery (NCD), the task of unveiling new classes in a set of unlabeled samples given a labeled dataset with known classes. We exploit the peculiarities of NCD to build a new framework, named Neighborhood Contrastive Learning (NCL), to learn discriminative representations that are important to clustering performance. Our contribution is twofold. First, we find that a feature extractor trained on the labeled set generates representations in which a generic query sample and its neighbors are likely to share the same class. We exploit this observation to retrieve and aggregate pseudo-positive pairs with contrastive learning, thus encouraging the model to learn more discriminative representations. Second, we notice that most of the instances are easily discriminated by the network, contributing less to the contrastive loss. To overcome this issue, we propose to generate hard negatives by mixing labeled and unlabeled samples in the feature space. We experimentally demonstrate that these two ingredients significantly contribute to clustering performance and lead our model to outperform state-of-the-art methods by a large margin (e.g., clustering accuracy +13% on CIFAR-100 and +8% on ImageNet). | ['Nicu Sebe', 'Elisa Ricci', 'Zhiming Luo', 'Subhankar Roy', 'Enrico Fini', 'Zhun Zhong'] | 2021-06-20 | null | http://openaccess.thecvf.com//content/CVPR2021/html/Zhong_Neighborhood_Contrastive_Learning_for_Novel_Class_Discovery_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Zhong_Neighborhood_Contrastive_Learning_for_Novel_Class_Discovery_CVPR_2021_paper.pdf | cvpr-2021-1 | ['novel-class-discovery', 'novel-class-discovery'] | ['computer-vision', 'methodology'] | [ 1.13969691e-01 -5.39663024e-02 -4.01085049e-01 -6.50874078e-01
-1.07291996e+00 -7.22418070e-01 7.39820540e-01 2.32303143e-01
-3.63064706e-01 7.34243691e-01 -2.85220835e-02 1.96063697e-01
-1.24398209e-01 -6.85527802e-01 -7.32328951e-01 -9.18536127e-01
-1.63246199e-01 6.84345067e-01 5.59383444e-02 2.14077920e-01
2.04414800e-02 6.73607528e-01 -1.69315338e+00 5.92566311e-01
7.83986211e-01 9.78700519e-01 -3.42823491e-02 7.36032948e-02
2.74200618e-01 7.46020377e-01 -5.04888833e-01 -1.91306457e-01
4.01423782e-01 -6.10963583e-01 -9.59648490e-01 2.33782843e-01
6.60259783e-01 6.60768971e-02 -1.03880979e-01 1.05814171e+00
3.65031362e-01 2.00045615e-01 1.25665975e+00 -1.19872940e+00
-6.81071281e-01 7.72301495e-01 -7.18334615e-01 2.15897411e-01
-1.24522433e-01 -4.12866585e-02 1.35730255e+00 -1.33001661e+00
6.94054186e-01 1.20029891e+00 4.25174147e-01 6.66150331e-01
-1.44929659e+00 -8.47438753e-01 3.21911126e-01 2.91928619e-01
-1.74663186e+00 -4.72540736e-01 9.18302596e-01 -3.67452800e-01
3.25517386e-01 3.11083674e-01 3.47859949e-01 9.22310531e-01
-6.15483105e-01 1.15330577e+00 1.14313197e+00 -5.17445922e-01
4.16359305e-01 4.06252444e-01 5.83779633e-01 4.68557537e-01
2.70505786e-01 8.17454681e-02 -4.15608346e-01 -7.22386837e-02
2.91010022e-01 1.76152185e-01 -2.74490982e-01 -7.97037244e-01
-1.01873505e+00 9.98335838e-01 8.62026334e-01 5.23708701e-01
-1.96118832e-01 -5.67367114e-02 7.30246082e-02 1.78310484e-01
5.25855780e-01 7.56474853e-01 -4.14953858e-01 4.01222378e-01
-7.63834953e-01 3.53379473e-02 4.99744296e-01 6.52303696e-01
1.20670629e+00 -2.97675699e-01 -2.71153629e-01 8.54203939e-01
2.15361416e-01 4.31706458e-01 4.93860692e-01 -7.91934967e-01
3.04234505e-01 7.52094209e-01 -3.43489349e-01 -8.22212934e-01
-3.63822669e-01 -1.08835256e+00 -9.21178877e-01 1.26587644e-01
3.60561639e-01 5.77154942e-02 -8.89701009e-01 1.94209898e+00
2.97751248e-01 5.25978208e-01 2.61178225e-01 7.69334853e-01
5.56202710e-01 4.05703336e-01 -9.50901359e-02 -2.47890279e-01
9.14214969e-01 -1.03345227e+00 -2.39356264e-01 -1.87133506e-01
8.03310931e-01 -4.10588413e-01 8.97325158e-01 2.92350322e-01
-7.11223066e-01 -6.05993271e-01 -1.06923938e+00 4.54967380e-01
-3.62497896e-01 3.04377675e-01 5.29276311e-01 4.85599905e-01
-8.40876698e-01 8.58974814e-01 -6.75830841e-01 -1.32980183e-01
8.75073612e-01 3.94574255e-01 -3.20312053e-01 -3.69145781e-01
-7.19558418e-01 5.15423536e-01 5.69331408e-01 -1.90310761e-01
-9.79825318e-01 -6.52725458e-01 -5.76941371e-01 1.17584608e-01
3.78725559e-01 -2.75488794e-01 8.87836337e-01 -1.19463468e+00
-7.51272321e-01 9.54998136e-01 -2.84373671e-01 -5.91076016e-01
3.42509001e-01 -5.32332733e-02 -4.05334115e-01 2.87906170e-01
3.35165471e-01 8.81356061e-01 7.82398880e-01 -1.72224557e+00
-7.36193240e-01 -6.46277547e-01 -3.00914437e-01 2.08189666e-01
-5.68821490e-01 -5.42335808e-01 -3.94312024e-01 -7.09929526e-01
3.38236630e-01 -1.03837037e+00 -3.13815057e-01 -1.26657769e-01
-5.84755719e-01 -5.41662574e-01 8.79067123e-01 6.62460253e-02
8.19051921e-01 -2.25439000e+00 5.78901395e-02 6.98176324e-01
5.43422937e-01 4.15465415e-01 -3.29479098e-01 2.34599989e-02
-3.03846180e-01 1.19376674e-01 -2.40104541e-01 -3.83338273e-01
-1.53958380e-01 1.70952290e-01 -3.55711818e-01 4.36896592e-01
5.91272473e-01 1.00634122e+00 -1.07601392e+00 -2.37917289e-01
1.42900959e-01 3.10768306e-01 -4.32204574e-01 8.75597447e-02
-1.81845687e-02 5.32364368e-01 -3.43546361e-01 5.52709579e-01
8.13082516e-01 -6.16337359e-01 4.20155168e-01 -2.30675131e-01
4.36405361e-01 2.93580145e-01 -1.21314943e+00 1.38252950e+00
-6.82411343e-02 3.60074639e-01 -3.99417609e-01 -1.42447257e+00
8.30567777e-01 -2.70748809e-02 3.06870729e-01 -6.80825412e-01
-1.05475500e-01 3.91609401e-01 -6.44712448e-02 -1.24725774e-01
9.68177393e-02 6.28977194e-02 2.36455932e-01 5.37240684e-01
2.87699789e-01 3.19509119e-01 1.01222873e-01 3.06500256e-01
8.40548813e-01 -3.19014609e-01 1.85355410e-01 -4.07062382e-01
3.21139961e-01 -2.13846117e-01 6.86553597e-01 1.05062175e+00
-2.36498371e-01 5.54003298e-01 2.79262632e-01 -2.78429240e-01
-7.51505911e-01 -1.30177879e+00 -2.29364812e-01 1.06701815e+00
1.57386392e-01 -2.04488248e-01 -5.36245286e-01 -1.24038064e+00
-7.70712569e-02 6.13253534e-01 -8.03111434e-01 -4.29043412e-01
-4.94730383e-01 -9.74052846e-01 1.64268047e-01 6.91593349e-01
3.19677979e-01 -7.98703909e-01 -8.17738939e-03 -6.63657859e-02
-4.44351286e-02 -8.56408060e-01 -1.27671748e-01 6.16350055e-01
-9.22553480e-01 -1.31350279e+00 -6.50978088e-01 -1.13045621e+00
9.01277423e-01 5.50125837e-01 1.33243072e+00 1.13599740e-01
-1.88867763e-01 2.47571409e-01 -5.44177234e-01 -2.55842302e-02
-2.23768204e-01 1.94355428e-01 1.69333756e-01 3.78315121e-01
6.42596424e-01 -6.12831831e-01 -6.26273215e-01 4.35046285e-01
-8.06434393e-01 -1.79318339e-01 7.61035562e-01 8.91174078e-01
8.41487527e-01 1.09152339e-01 8.10684919e-01 -1.23655951e+00
8.12374577e-02 -6.98050559e-01 -3.41327548e-01 3.91276002e-01
-6.66236162e-01 2.45734274e-01 6.99591219e-01 -6.18041754e-01
-5.93068421e-01 4.60579097e-01 2.19253883e-01 -6.57764852e-01
-2.53039569e-01 3.47387701e-01 -2.21782789e-01 3.54948528e-02
9.33835983e-01 1.74082503e-01 -1.91718295e-01 -5.38800418e-01
6.09147787e-01 4.10281003e-01 5.80212295e-01 -5.06099582e-01
1.11181235e+00 6.54859900e-01 -6.74578100e-02 -5.79133809e-01
-1.20402396e+00 -8.18384945e-01 -8.67226183e-01 -8.51082243e-03
6.62439764e-01 -1.22251749e+00 -2.93670833e-01 7.62339011e-02
-6.61912799e-01 -2.01808840e-01 -5.63460946e-01 7.28832006e-01
-2.49840841e-01 1.59045801e-01 -3.32355708e-01 -6.22433722e-01
6.11221008e-02 -1.09701955e+00 9.17202890e-01 2.20507219e-01
-1.12181202e-01 -7.57413089e-01 4.91996594e-02 1.54568329e-01
-1.91012397e-02 1.23431131e-01 1.01013529e+00 -1.35474145e+00
-6.31968439e-01 -5.31648770e-02 -2.45493382e-01 4.90652472e-01
1.79807663e-01 -1.93658695e-01 -1.49263549e+00 -5.21483958e-01
-1.26152202e-01 -6.53134346e-01 1.39404345e+00 1.82372853e-01
1.42180324e+00 -2.68607974e-01 -6.84128523e-01 5.29100120e-01
1.32106650e+00 -6.32451400e-02 3.47662717e-01 -1.10720426e-01
7.67758369e-01 7.04116642e-01 5.48606217e-01 2.17634007e-01
-8.01798236e-03 6.70186222e-01 4.03082281e-01 -1.73832238e-01
-3.27913672e-01 -2.23778516e-01 -3.63394022e-02 7.28647768e-01
2.35011518e-01 -5.35648428e-02 -8.58311236e-01 7.00921893e-01
-1.86790824e+00 -8.44860494e-01 -1.17371403e-01 2.22959495e+00
8.45438838e-01 2.84217112e-02 1.10838369e-01 2.65123636e-01
9.29753006e-01 -1.25339702e-01 -6.57024503e-01 4.57184672e-01
-3.22109640e-01 4.83757883e-01 1.60965443e-01 3.57368082e-01
-1.39624536e+00 8.83719742e-01 5.44038725e+00 1.28669965e+00
-1.00691521e+00 -1.26241520e-01 1.20670497e+00 7.04036132e-02
-4.48068738e-01 -1.83026031e-01 -8.83538544e-01 2.06742421e-01
7.32378781e-01 9.88469422e-02 2.86619186e-01 9.62013960e-01
-2.37783670e-01 2.51223624e-01 -1.52642131e+00 1.05615056e+00
2.41175786e-01 -1.43727756e+00 3.57580364e-01 5.13448566e-02
1.08751845e+00 7.81387612e-02 4.54168618e-01 4.60051328e-01
3.72389019e-01 -1.07142782e+00 4.80274141e-01 1.26476347e-01
5.79181314e-01 -8.78760993e-01 6.31273985e-01 3.17591518e-01
-1.01141667e+00 -1.52543306e-01 -5.42862952e-01 1.97044507e-01
-4.65361625e-01 9.02139664e-01 -1.07936203e+00 4.91163880e-01
6.70930326e-01 1.08819890e+00 -1.00568438e+00 1.15419471e+00
-2.42521822e-01 7.88819969e-01 -1.83189675e-01 7.31999427e-02
2.95422167e-01 -1.39689714e-01 1.65354937e-01 8.33168983e-01
8.28967392e-02 -1.62712380e-01 2.65850097e-01 1.05848813e+00
-6.78311288e-01 4.29339148e-02 -5.66707730e-01 1.10397086e-01
5.42124689e-01 1.24746573e+00 -8.25467050e-01 -6.20995581e-01
-1.92094222e-01 9.18960333e-01 8.00034881e-01 4.74728703e-01
-6.70381784e-01 -1.22757390e-01 4.60148364e-01 -2.41647437e-01
5.42409718e-01 3.01857054e-01 -1.68344498e-01 -1.16661024e+00
1.28963619e-01 -7.86036193e-01 4.49906796e-01 -2.69036353e-01
-1.83286357e+00 5.90245664e-01 -3.19610655e-01 -1.49131036e+00
-2.65923858e-01 -4.12680537e-01 -4.04520541e-01 5.91538489e-01
-1.71328616e+00 -9.68516409e-01 -1.96698114e-01 6.15479469e-01
3.20398897e-01 -1.99152663e-01 8.20492148e-01 4.73580807e-01
-4.53285158e-01 8.99275362e-01 5.02915323e-01 4.43197101e-01
7.95035124e-01 -1.27045751e+00 7.79090747e-02 6.94816470e-01
8.41840684e-01 5.98956108e-01 1.15810364e-01 -3.10258657e-01
-7.24769711e-01 -1.63098133e+00 7.32355535e-01 -5.92983425e-01
3.62016350e-01 -5.41637242e-01 -1.06413925e+00 5.37015200e-01
-2.39032939e-01 5.24076104e-01 1.07972407e+00 2.28854403e-01
-9.32654321e-01 -2.44439602e-01 -1.06957757e+00 4.95944321e-01
1.05190861e+00 -6.09818280e-01 -5.96623838e-01 4.78856593e-01
8.19335401e-01 2.68084735e-01 -4.45718676e-01 6.07180834e-01
2.02919200e-01 -9.63907599e-01 1.01961446e+00 -8.18376184e-01
2.59790093e-01 -3.72012526e-01 -3.20230007e-01 -1.34401572e+00
-4.46151555e-01 -1.06510051e-01 -8.47111717e-02 1.27962089e+00
8.17575812e-01 -5.29759228e-01 9.83959138e-01 2.43303552e-01
-8.38112906e-02 -7.39745021e-01 -8.78857315e-01 -9.37785149e-01
3.20269793e-01 -2.16461882e-01 4.64062184e-01 1.45685756e+00
-3.16122741e-01 7.01879680e-01 -1.10164948e-01 2.92156130e-01
7.46037900e-01 3.66036147e-01 5.26056647e-01 -1.59859443e+00
-3.66937697e-01 -2.78992325e-01 -3.34207058e-01 -1.11874747e+00
3.40667933e-01 -1.27828193e+00 -8.93739089e-02 -8.52135658e-01
6.78936660e-01 -9.25843775e-01 -6.87378585e-01 5.39437592e-01
-4.14823532e-01 6.73135757e-01 1.50321767e-01 5.81858456e-01
-1.15320051e+00 5.84871411e-01 8.34758222e-01 -4.15890247e-01
-2.03703165e-01 1.15267530e-01 -8.56521308e-01 5.87416232e-01
6.71961546e-01 -6.43674016e-01 -3.65476161e-01 -2.91768730e-01
-5.49212210e-02 -5.62320471e-01 5.23208737e-01 -1.14119649e+00
2.59975716e-02 1.78857669e-01 7.67093897e-01 -6.41236067e-01
3.58220667e-01 -9.34978664e-01 -3.39193553e-01 5.37188590e-01
-8.06455612e-01 -4.35519964e-01 -1.96941659e-01 8.59250009e-01
-3.17982167e-01 -2.75826722e-01 9.71819341e-01 -2.53384304e-03
-7.20337808e-01 3.57450426e-01 -3.35152484e-02 3.53949159e-01
1.10926545e+00 8.12967643e-02 -4.34900492e-01 -1.94415182e-01
-7.87328959e-01 1.02467231e-01 3.68347108e-01 2.20223114e-01
5.47901154e-01 -1.50838363e+00 -6.77886844e-01 3.13752323e-01
5.06063163e-01 3.74907888e-02 6.22194484e-02 6.11333072e-01
-5.91999963e-02 1.97374582e-01 1.69814616e-01 -9.90391195e-01
-1.10693872e+00 7.04598606e-01 2.07957581e-01 -2.54409492e-01
-2.70503104e-01 1.12718284e+00 4.87531662e-01 -6.95023477e-01
3.47113043e-01 -2.50928476e-02 -1.85123414e-01 1.81229085e-01
6.07182920e-01 1.45148650e-01 1.33405998e-01 -5.61609924e-01
-4.63861406e-01 4.81429607e-01 -4.24468011e-01 1.94262281e-01
1.40774524e+00 1.80552736e-01 -7.12289140e-02 4.87884343e-01
1.78903282e+00 4.56324555e-02 -1.12822163e+00 -5.62008202e-01
2.13951901e-01 -4.19276834e-01 -2.69761205e-01 -7.44541049e-01
-1.28366017e+00 8.56954038e-01 9.88280952e-01 5.66008575e-02
1.02890623e+00 4.44396585e-01 3.37698728e-01 6.95161104e-01
1.59319714e-01 -9.46764529e-01 4.89835113e-01 3.79795760e-01
2.42426932e-01 -1.58584201e+00 -2.90317357e-01 -4.96689707e-01
-5.35099804e-01 8.18668365e-01 5.42049170e-01 -2.71505475e-01
6.50115669e-01 -2.65100479e-01 5.29772826e-02 -2.13907853e-01
-5.60353994e-01 -4.85073358e-01 4.74595636e-01 7.77162671e-01
2.23730236e-01 1.87536702e-01 1.51261151e-01 4.09903586e-01
2.15186760e-01 -4.70528483e-01 1.65199757e-01 5.79691291e-01
-5.41892350e-01 -1.10662711e+00 -1.53586701e-01 6.45831823e-01
-2.36858025e-01 -1.95257023e-01 -6.74036801e-01 6.90419853e-01
1.12574942e-01 1.00581205e+00 2.12067977e-01 -4.25969809e-01
-5.42369932e-02 9.70721394e-02 2.67067879e-01 -8.32622349e-01
-3.63997012e-01 3.49814259e-02 -2.89745957e-01 -3.95413607e-01
-6.53887928e-01 -3.68608534e-01 -1.16189051e+00 2.28731915e-01
-4.33033824e-01 5.41343033e-01 1.27974480e-01 8.83550406e-01
3.96602184e-01 3.65488678e-01 1.16502309e+00 -6.33966208e-01
-6.67862833e-01 -7.32533634e-01 -8.11068594e-01 8.32963943e-01
2.96519548e-01 -7.15660870e-01 -6.44545138e-01 -1.84161022e-01] | [9.537176132202148, 3.0136330127716064] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.