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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
5736659b-94f9-4cc8-abfd-48bf5b1b66fa | controllability-of-coarsely-measured | 2206.10569 | null | https://arxiv.org/abs/2206.10569v1 | https://arxiv.org/pdf/2206.10569v1.pdf | Controllability of Coarsely Measured Networked Linear Dynamical Systems (Extended Version) | We consider the controllability of large-scale linear networked dynamical systems when complete knowledge of network structure is unavailable and knowledge is limited to coarse summaries. We provide conditions under which average controllability of the fine-scale system can be well approximated by average controllability of the (synthesized, reduced-order) coarse-scale system. To this end, we require knowledge of some inherent parametric structure of the fine-scale network that makes this type of approximation possible. Therefore, we assume that the underlying fine-scale network is generated by the stochastic block model (SBM) -- often studied in community detection. We then provide an algorithm that directly estimates the average controllability of the fine-scale system using a coarse summary of SBM. Our analysis indicates the necessity of underlying structure (e.g., in-built communities) to be able to quantify accurately the controllability from coarsely characterized networked dynamics. We also compare our method to that of the reduced-order method and highlight the regimes where both can outperform each other. Finally, we provide simulations to confirm our theoretical results for different scalings of network size and density, and the parameter that captures how much community-structure is retained in the coarse summary. | ['Stark C. Draper', 'Gautam Dasarathy', 'Rajasekhar Anguluri', 'Nafiseh Ghoroghchian'] | 2022-06-21 | null | null | null | null | ['stochastic-block-model'] | ['graphs'] | [ 3.76153812e-02 1.73403844e-01 5.29530644e-02 5.73378980e-01
7.86927789e-02 -1.04290307e+00 6.63344026e-01 1.54127061e-01
8.73627961e-02 9.61915016e-01 8.70443822e-04 -3.49081308e-01
-5.72300494e-01 -9.92338479e-01 -5.76102734e-01 -8.92787218e-01
-7.14860320e-01 6.66841447e-01 4.34544951e-01 -5.97297609e-01
-7.48822615e-02 7.25972056e-01 -9.78660643e-01 -2.85882056e-01
8.66845429e-01 3.14596713e-01 4.93395813e-02 1.17444420e+00
3.17997038e-01 6.29474878e-01 -4.68462378e-01 2.54355937e-01
5.17833948e-01 -4.98753399e-01 -6.83630824e-01 2.72646099e-01
2.14399695e-02 -9.75237507e-03 -5.70361435e-01 1.08253789e+00
1.75772846e-01 -2.43313566e-01 8.69018912e-01 -1.40970802e+00
-1.54775843e-01 8.11391473e-01 -4.64346826e-01 3.35976750e-01
1.20348282e-01 1.64010137e-01 9.15983975e-01 -3.81980985e-01
9.87133801e-01 1.24812198e+00 9.01102006e-01 1.17316455e-01
-1.93046105e+00 -5.84091485e-01 5.74235804e-02 -5.03370404e-01
-1.77602673e+00 -3.59205216e-01 5.06301761e-01 -8.10575545e-01
5.08078277e-01 1.59546986e-01 9.69482064e-01 8.16654861e-01
3.28442633e-01 -1.32210612e-01 1.17149186e+00 -1.90988854e-01
3.17810595e-01 1.07579408e-02 7.00646862e-02 6.71094894e-01
1.05214059e+00 6.23176508e-02 9.94574837e-03 -6.80842996e-01
1.23417592e+00 -1.58007473e-01 -4.36317593e-01 -7.01019824e-01
-1.30798876e+00 8.01828206e-01 4.99967992e-01 5.86732447e-01
-3.70180339e-01 2.43309274e-01 1.92613050e-01 7.96821952e-01
3.88151914e-01 6.30390882e-01 -7.57636800e-02 1.77522063e-01
-9.76417601e-01 1.60836846e-01 1.58688068e+00 1.03582561e+00
8.10467899e-01 -5.31275645e-02 1.30455434e-01 2.87488587e-02
-4.41478007e-02 7.21275091e-01 -4.94353801e-01 -1.06358624e+00
8.65518972e-02 5.31801462e-01 4.37036276e-01 -1.07596457e+00
-4.66057956e-01 -9.14197922e-01 -1.51658332e+00 5.88595122e-02
6.91895425e-01 -2.65766472e-01 -2.53105074e-01 1.80613875e+00
2.34240070e-01 2.17823416e-01 -8.68966728e-02 4.30412441e-01
-4.68757302e-02 6.74574137e-01 -5.85575521e-01 -8.32902133e-01
6.82654202e-01 -6.37961745e-01 -4.59616274e-01 2.36363366e-01
3.97295624e-01 -3.23540658e-01 5.73428631e-01 3.15682888e-02
-9.99927878e-01 -1.87471434e-02 -9.72361147e-01 7.52259672e-01
-5.57302013e-02 -3.17076564e-01 9.31122899e-02 4.20104682e-01
-1.62349594e+00 8.10515463e-01 -8.26738119e-01 -9.31185365e-01
-8.24025720e-02 2.92623520e-01 -3.15219492e-01 5.91820069e-02
-1.08268273e+00 6.77762568e-01 1.43070459e-01 1.84206918e-01
-1.04478014e+00 -6.47029579e-01 -2.32108340e-01 2.38464639e-01
5.70870876e-01 -8.53262782e-01 6.67293608e-01 -7.19911814e-01
-7.22153366e-01 3.68236661e-01 -7.08488002e-02 -3.12998116e-01
7.52445519e-01 6.01406157e-01 5.32133840e-02 6.01776600e-01
1.92475796e-01 -1.19662888e-01 8.57170880e-01 -1.51980102e+00
-3.94176915e-02 1.34642839e-01 4.29167897e-01 -1.68603987e-01
-2.78311223e-01 -2.12820023e-01 1.20005809e-01 -3.62343103e-01
-4.75009568e-02 -1.25360489e+00 -5.42616963e-01 1.08055882e-01
-7.57171333e-01 2.34523892e-01 5.94632030e-01 -1.15536831e-01
1.42160261e+00 -1.77043235e+00 4.41660076e-01 8.46220076e-01
8.92004251e-01 2.89254133e-02 -3.38116646e-01 1.31883454e+00
-8.77006054e-02 6.60774589e-01 -8.22456479e-02 -1.57549940e-02
-2.13242695e-01 2.25271419e-01 -1.08362220e-01 8.31749320e-01
1.17514886e-01 4.83549476e-01 -1.00170970e+00 -3.96708220e-01
-8.12570825e-02 2.85353601e-01 -5.72359145e-01 -2.48290375e-02
2.05584079e-01 5.58735013e-01 -6.85548127e-01 3.51639092e-01
4.46373820e-01 -8.62467229e-01 5.51432610e-01 -5.44914752e-02
-2.64375776e-01 -3.22849929e-01 -1.43038130e+00 7.28275836e-01
-3.46298337e-01 4.89504874e-01 7.91980445e-01 -9.85555947e-01
5.67237973e-01 3.94361079e-01 4.46996063e-01 4.22277182e-01
-1.37597686e-02 2.91490644e-01 5.18067956e-01 -8.49794000e-02
3.03191356e-02 -1.41049132e-01 -1.56935062e-02 6.97669446e-01
-2.31637686e-01 -6.07524030e-02 5.46064496e-01 8.30082834e-01
1.77510524e+00 -8.22312832e-01 7.76305854e-01 -1.11756468e+00
4.06839818e-01 1.12120792e-01 2.23572031e-01 1.21146345e+00
-3.16040099e-01 -3.49128284e-02 1.23682761e+00 3.69294137e-02
-1.62401319e+00 -1.09264433e+00 4.03386019e-02 4.46361721e-01
1.19913548e-01 -5.49760580e-01 -7.76417732e-01 2.31631380e-02
2.86026716e-01 -2.86644191e-01 -8.99637401e-01 -1.79367159e-02
-4.35347706e-01 -6.04345381e-01 3.66767973e-01 5.65488301e-02
9.31037366e-02 -4.64594930e-01 1.10542543e-01 3.71004134e-01
-5.93011156e-02 -1.03713965e+00 -4.28185910e-01 6.56538308e-02
-7.98394203e-01 -1.44465399e+00 -5.61269224e-01 -6.03039742e-01
8.80220354e-01 2.25276902e-01 1.01304471e+00 5.36229491e-01
-9.20006856e-02 4.49440718e-01 -1.69087112e-01 4.89215106e-01
-1.02379370e+00 4.83986527e-01 4.72057909e-01 -1.10839255e-01
-6.21934772e-01 -1.14381802e+00 -4.26571786e-01 3.82412016e-01
-7.41984069e-01 -8.81188288e-02 6.14708364e-01 6.96657419e-01
7.22066760e-02 4.38168615e-01 6.49295151e-01 -7.64806807e-01
9.16454375e-01 -6.22806668e-01 -7.53704846e-01 4.13508832e-01
-4.97697294e-01 3.25189903e-02 9.05513287e-01 -6.34445012e-01
-2.46073857e-01 -1.68096438e-01 6.57267392e-01 -2.90184706e-01
2.60891974e-01 6.34336114e-01 1.30050123e-01 -5.00557363e-01
5.67627251e-01 2.01074287e-01 3.15184951e-01 -1.80101335e-01
2.31539235e-01 5.10641456e-01 2.15230390e-01 -6.46218598e-01
1.41585791e+00 6.45813286e-01 5.93059659e-01 -1.15203214e+00
-3.57362986e-01 -2.44181976e-01 -8.69744301e-01 -4.05713707e-01
2.57295460e-01 -8.69991779e-01 -1.10106826e+00 3.97371143e-01
-9.96937335e-01 -5.53353071e-01 -2.29695871e-01 7.37669766e-02
-5.00743926e-01 9.32605267e-02 -9.77999270e-01 -1.22618794e+00
1.56327724e-01 -7.11253703e-01 6.80970073e-01 -2.66479880e-01
-2.52848506e-01 -1.41479468e+00 3.78560811e-01 -4.09340203e-01
7.84982502e-01 6.73768103e-01 7.80188799e-01 -4.89012361e-01
-8.23546767e-01 -1.65139183e-01 -2.99047589e-01 -9.76844057e-02
1.91917077e-01 5.73881507e-01 -2.33988300e-01 -7.80864775e-01
-1.44324258e-01 1.68570593e-01 6.13321841e-01 5.81041932e-01
2.81477571e-01 -6.65249944e-01 -7.19717979e-01 1.59857914e-01
1.65443015e+00 -5.82880557e-01 1.81260128e-02 -7.20964372e-02
7.60812342e-01 5.86771965e-01 5.54977879e-02 6.02554142e-01
1.90896079e-01 3.40577841e-01 2.78401524e-01 1.14961341e-01
9.67193469e-02 -2.93924421e-01 1.82554320e-01 1.30140340e+00
-2.98873872e-01 -1.91766173e-01 -1.04825687e+00 6.31848633e-01
-1.85115123e+00 -1.16825569e+00 -3.88948977e-01 2.07197237e+00
8.75520170e-01 3.18879485e-01 5.11698782e-01 -8.39227252e-03
1.27454758e+00 1.09514169e-01 -5.63629746e-01 8.66256058e-02
-3.39097053e-01 -2.48949662e-01 7.01401174e-01 1.03606546e+00
-8.83907378e-01 5.46348035e-01 7.25886488e+00 6.78067923e-01
-7.40335763e-01 -5.19804005e-03 2.82173991e-01 1.70892570e-02
-2.10870489e-01 3.27043176e-01 -5.97088099e-01 2.15151846e-01
1.04447842e+00 -8.20123672e-01 7.88027108e-01 4.27727968e-01
7.62271047e-01 -5.60448542e-02 -1.01104081e+00 2.22137257e-01
-5.72713077e-01 -1.35839534e+00 -2.17545331e-01 7.57691562e-01
1.15826547e+00 1.04217799e-02 -4.33636457e-01 -1.73351198e-01
8.77664328e-01 -8.00714850e-01 3.55653137e-01 7.34277487e-01
8.89960468e-01 -4.24820632e-01 5.33137858e-01 6.37073755e-01
-1.80836129e+00 -8.03062171e-02 -2.19195589e-01 -4.27453190e-01
2.15876356e-01 1.03528488e+00 -5.58264911e-01 3.69695455e-01
1.80898920e-01 9.64558244e-01 -3.63133609e-01 9.55457866e-01
1.61483362e-01 6.76579952e-01 -7.79682994e-01 -1.13931291e-01
1.20065287e-01 -3.81141514e-01 1.03922009e+00 8.95258665e-01
2.96475083e-01 1.38955832e-01 4.22172338e-01 9.50121582e-01
6.84246719e-02 -2.24529460e-01 -8.48817170e-01 -3.49768639e-01
8.78231883e-01 1.28908813e+00 -1.00834727e+00 -2.80459225e-01
-7.62665421e-02 4.60727930e-01 2.70380139e-01 6.28266275e-01
-2.68755615e-01 -2.17885077e-01 7.97268391e-01 5.52001774e-01
4.36128914e-01 -4.65484798e-01 2.54333373e-02 -1.40187037e+00
-2.42780417e-01 -6.78519964e-01 -1.12670250e-01 -5.87585747e-01
-1.52139473e+00 5.58862329e-01 4.32459377e-02 -1.10798490e+00
-3.14011335e-01 -1.98293060e-01 -7.51045823e-01 7.32018948e-01
-9.22604442e-01 -9.10405517e-01 -1.69800315e-02 4.95857745e-01
-3.10568631e-01 2.74184793e-01 5.95969915e-01 8.10450539e-02
-6.17909670e-01 -1.86991133e-02 6.23995423e-01 1.19279942e-03
3.91359925e-01 -1.33899939e+00 2.01529369e-01 9.31464195e-01
-6.16641343e-01 9.30245638e-01 1.25757146e+00 -1.00836658e+00
-1.38026667e+00 -1.21547353e+00 7.86504865e-01 -3.66077214e-01
1.61929536e+00 -7.88337231e-01 -6.50981307e-01 5.68377316e-01
-3.55409570e-02 3.12477857e-01 3.64946537e-02 -5.31536415e-02
-2.22091060e-02 -1.09947540e-01 -1.13917911e+00 6.36689603e-01
1.32483435e+00 -6.20331228e-01 -6.27764687e-02 2.95905292e-01
7.02831447e-01 2.96223342e-01 -1.15534604e+00 7.38962367e-02
5.45364618e-01 -7.37117708e-01 7.17833877e-01 -4.28722501e-01
1.24644205e-01 -4.34338570e-01 -6.29268214e-02 -1.46129632e+00
-5.65287054e-01 -9.70332325e-01 -1.83470741e-01 1.02416551e+00
2.62566715e-01 -1.03789294e+00 4.18583393e-01 1.48959085e-01
5.80485642e-01 -5.67408442e-01 -7.92751610e-01 -1.10431516e+00
3.40127230e-01 4.50827867e-01 1.25065282e-01 1.08883286e+00
1.26730695e-01 3.05921644e-01 -2.47351095e-01 3.93728614e-01
9.02967751e-01 1.43505529e-01 6.70471013e-01 -1.73972213e+00
-2.11232767e-01 -4.82417881e-01 -3.26403797e-01 -7.45850980e-01
1.60414413e-01 -5.05491555e-01 -3.13239396e-01 -1.36489785e+00
5.06033540e-01 -4.36471343e-01 2.02798784e-01 -8.95598903e-03
1.71198130e-01 2.71474630e-01 1.57228485e-01 6.82012677e-01
-5.20805538e-01 1.57962292e-01 1.41651464e+00 1.35060504e-01
-1.59403622e-01 6.41217902e-02 -6.83756053e-01 3.68323207e-01
5.88344514e-01 -5.33616066e-01 -3.92329305e-01 5.48530161e-01
5.64885318e-01 4.30999070e-01 5.33623993e-01 -8.75081539e-01
4.95340705e-01 -2.30628476e-01 -2.54288614e-01 -3.20981443e-01
2.33460087e-02 -7.06017733e-01 5.40868521e-01 9.62570369e-01
-3.34552199e-01 6.61663488e-02 -2.43772700e-01 9.40532863e-01
1.00609399e-01 3.35070305e-02 6.40782595e-01 -5.42804077e-02
1.53043523e-01 3.00121814e-01 -8.65754724e-01 2.21628338e-01
8.61389399e-01 -9.44637787e-03 -5.39614618e-01 -8.78355920e-01
-1.09577310e+00 3.15275967e-01 9.77902055e-01 -3.45507711e-01
-1.04387335e-01 -1.19430709e+00 -9.14427280e-01 -4.44931425e-02
-1.64862290e-01 -4.79566097e-01 -3.51274759e-02 1.27544081e+00
-5.84240496e-01 3.13846886e-01 -5.31417727e-02 -5.97834110e-01
-1.10100865e+00 6.88396096e-01 5.03097773e-01 -5.65310001e-01
-2.14831308e-01 9.59141999e-02 2.27080211e-01 -2.51940101e-01
-2.41971254e-01 -3.54481220e-01 2.48583004e-01 6.90065790e-03
2.50189036e-01 4.64491934e-01 -5.29576123e-01 -5.88493109e-01
-3.36460650e-01 5.58667243e-01 2.74668634e-01 3.41894217e-02
1.31380737e+00 -6.90555513e-01 -7.51117587e-01 6.70756578e-01
1.03206325e+00 2.26207003e-01 -1.25210977e+00 -2.07629234e-01
-3.71344030e-01 -6.81403801e-02 -2.83632874e-01 -1.31243065e-01
-8.74832809e-01 5.16765952e-01 -2.22037464e-01 1.25911331e+00
7.94914126e-01 2.88272291e-01 -5.25614284e-02 6.43655896e-01
5.94489276e-01 -4.33512986e-01 -2.02866718e-01 5.51096439e-01
8.08101833e-01 -7.01729357e-01 1.54275343e-01 -7.83830643e-01
1.42638132e-01 1.09473932e+00 1.64810658e-01 -8.10485959e-01
1.12805343e+00 6.18559182e-01 -5.28485715e-01 -1.57132819e-01
-1.11041856e+00 -2.31141210e-01 -1.24389805e-01 6.08575165e-01
-1.50553081e-02 1.58036470e-01 -2.30093434e-01 7.43084997e-02
-9.32457596e-02 -2.98737347e-01 1.23021519e+00 5.83586395e-01
-6.47140324e-01 -8.69637549e-01 -2.85878003e-01 7.12033987e-01
-9.22036469e-02 -1.64279133e-01 -7.43798196e-01 1.16719270e+00
-4.31550443e-01 9.75619435e-01 3.00562587e-02 -6.85733333e-02
-2.40759761e-03 -4.91994441e-01 3.76405001e-01 -5.95175743e-01
-4.17596698e-01 2.25633502e-01 1.82216659e-01 -3.68237436e-01
-6.31418705e-01 -6.01538062e-01 -5.23515284e-01 -1.13487828e+00
-5.72597742e-01 3.84962231e-01 -5.91523573e-02 8.27872396e-01
3.01470160e-01 4.05305952e-01 8.45765054e-01 -9.67730701e-01
-7.08355725e-01 -9.12048221e-01 -1.26397252e+00 -2.02638015e-01
9.28059340e-01 -5.66038370e-01 -1.06945193e+00 -1.73537955e-01] | [6.7261786460876465, 4.960445880889893] |
6031f465-2077-4b28-93e9-589b349ea256 | a-memory-efficient-baseline-for-open-domain | 2012.15156 | null | https://arxiv.org/abs/2012.15156v1 | https://arxiv.org/pdf/2012.15156v1.pdf | A Memory Efficient Baseline for Open Domain Question Answering | Recently, retrieval systems based on dense representations have led to important improvements in open-domain question answering, and related tasks. While very effective, this approach is also memory intensive, as the dense vectors for the whole knowledge source need to be kept in memory. In this paper, we study how the memory footprint of dense retriever-reader systems can be reduced. We consider three strategies to reduce the index size: dimension reduction, vector quantization and passage filtering. We evaluate our approach on two question answering benchmarks: TriviaQA and NaturalQuestions, showing that it is possible to get competitive systems using less than 6Gb of memory. | ['Edouard Grave', 'Sebastian Riedel', 'Nicola De Cao', 'Lucas Hosseini', 'Fabio Petroni', 'Gautier Izacard'] | 2020-12-30 | null | null | null | null | ['triviaqa'] | ['miscellaneous'] | [-1.41348839e-01 -1.14969373e-01 1.77488290e-02 -3.97326052e-02
-1.34566522e+00 -8.34711611e-01 4.83378887e-01 7.90094972e-01
-7.15597570e-01 8.23906600e-01 4.83238101e-01 -4.43798900e-01
-5.36693692e-01 -1.12908411e+00 -5.25040567e-01 -2.23754257e-01
1.99037045e-01 7.89901972e-01 8.07864249e-01 -7.36001670e-01
4.95350122e-01 3.79815996e-01 -1.90819001e+00 4.58239794e-01
8.12708914e-01 9.62608635e-01 6.63107112e-02 9.71671700e-01
-7.25764334e-01 1.03393769e+00 -8.89055908e-01 -5.04940152e-01
-1.56242140e-02 -6.73268586e-02 -1.56181157e+00 -6.40734851e-01
5.74040115e-01 -4.48018074e-01 -6.20035470e-01 7.52416730e-01
6.19856238e-01 6.73874140e-01 6.62061274e-01 -5.61227679e-01
-8.95487547e-01 5.32293797e-01 -2.42379636e-01 7.70212829e-01
6.61325216e-01 -4.64369208e-01 1.34298635e+00 -5.69582939e-01
7.57790804e-01 1.29342163e+00 1.12969548e-01 3.34826767e-01
-9.71678972e-01 -2.49767661e-01 -9.49713811e-02 8.86322379e-01
-1.45327497e+00 -5.07994652e-01 2.00314105e-01 4.83496189e-02
1.44988370e+00 6.84549928e-01 3.32024544e-01 6.08665884e-01
-1.18956767e-01 7.39647865e-01 6.12870336e-01 -6.58006787e-01
2.91110098e-01 -7.58659989e-02 9.90874290e-01 5.48325002e-01
4.14250910e-01 -3.22798103e-01 -4.61260587e-01 -4.55106139e-01
2.20137075e-01 -6.40033036e-02 -4.07171488e-01 2.10772343e-02
-8.21690500e-01 1.19929826e+00 5.65644026e-01 4.30757701e-01
-3.62262815e-01 9.48479548e-02 5.17440319e-01 7.27873802e-01
2.81652063e-01 8.66078973e-01 -4.02894884e-01 -2.65037596e-01
-7.95529008e-01 5.35093665e-01 1.35320413e+00 8.92785847e-01
8.26285720e-01 -6.62209988e-01 -5.17108023e-01 1.10714054e+00
3.56965214e-02 5.70433676e-01 6.51296020e-01 -1.19240940e+00
4.52541023e-01 5.68978071e-01 1.20174833e-01 -1.15334725e+00
-2.68859744e-01 -1.33190915e-01 -4.19168472e-01 -6.24939203e-01
2.91506410e-01 4.18263078e-02 -8.51365447e-01 1.26912773e+00
1.79322064e-01 -2.91095674e-01 1.11884832e-01 8.04900587e-01
1.08811748e+00 1.07903337e+00 -6.92388564e-02 -1.70528412e-01
1.59190631e+00 -1.03430283e+00 -9.45337832e-01 -4.27696481e-02
9.35687184e-01 -8.91025662e-01 8.81254554e-01 1.57292426e-01
-1.17961407e+00 -2.27637023e-01 -9.11717176e-01 -7.36388981e-01
-7.34711051e-01 -3.92099738e-01 6.18655026e-01 4.59182173e-01
-1.10446572e+00 4.76140112e-01 -5.75317621e-01 -5.42542815e-01
2.46221244e-01 2.17673361e-01 -1.57356650e-01 -4.30344999e-01
-1.41657710e+00 1.00706697e+00 4.92112100e-01 -2.75824189e-01
-4.56399798e-01 -6.32750809e-01 -5.17514586e-01 6.32793903e-01
4.56374586e-01 -7.65520573e-01 1.32075596e+00 -1.59665823e-01
-1.23559415e+00 6.25934839e-01 -2.51240969e-01 -6.93641901e-01
-2.84818262e-01 -6.70115530e-01 -3.07245135e-01 5.55890977e-01
-2.49730602e-01 4.41403478e-01 4.13100302e-01 -5.63265324e-01
-5.90656698e-01 -4.42009062e-01 3.60231042e-01 2.67949224e-01
-8.41112256e-01 1.34929121e-01 -7.71562576e-01 -3.69749784e-01
-4.69433069e-02 -6.81320131e-01 -2.03311816e-01 -2.68466294e-01
-1.79300196e-02 -7.91326046e-01 5.27085185e-01 -6.58572257e-01
1.66398716e+00 -1.80081117e+00 1.48066148e-01 2.63146758e-01
4.28102106e-01 5.74784875e-01 -2.91922450e-01 5.86342275e-01
5.45158148e-01 8.45490620e-02 1.94308519e-01 2.81243742e-01
1.04189664e-01 4.55157548e-01 -7.62911141e-01 2.54908949e-03
-1.85194522e-01 1.20153689e+00 -7.77710557e-01 -6.94031000e-01
-1.11484423e-01 3.89459133e-01 -7.10660338e-01 3.26712608e-01
-4.01092708e-01 -5.77983975e-01 -6.40928745e-01 3.89675260e-01
2.96439350e-01 -4.99551624e-01 2.64407657e-02 -4.43221852e-02
3.24141860e-01 7.42359638e-01 -9.11025405e-01 1.79191101e+00
-4.02592540e-01 7.64935851e-01 -1.31787509e-01 -7.49111652e-01
8.13395858e-01 2.06250429e-01 1.89471334e-01 -1.20072222e+00
7.89854862e-03 2.00680912e-01 -4.25931931e-01 -4.66999739e-01
1.02791071e+00 1.62655652e-01 1.37321651e-04 6.62726343e-01
1.66404933e-01 -1.02608398e-01 6.27907097e-01 6.20306611e-01
1.50413644e+00 -5.56924820e-01 8.54549333e-02 -2.89709240e-01
6.40827477e-01 3.46789002e-01 -9.56776142e-02 9.89995956e-01
1.78991124e-01 2.75970817e-01 3.01766843e-01 -3.48137677e-01
-8.26647401e-01 -9.89335299e-01 -3.55681218e-02 1.65954435e+00
1.87808067e-01 -9.00990903e-01 -7.88598061e-01 -4.29295838e-01
1.96595281e-01 6.29509211e-01 -3.76154155e-01 -3.84810627e-01
-8.38722825e-01 -4.23424184e-01 6.50321841e-01 5.77658117e-01
9.80432183e-02 -9.94241893e-01 -6.94062054e-01 1.71345115e-01
-3.39944065e-01 -7.41512120e-01 -1.22561216e-01 1.99986592e-01
-8.75782371e-01 -1.04857397e+00 -8.91445756e-01 -7.19848394e-01
8.45026672e-02 5.40970922e-01 1.70215511e+00 4.44893926e-01
-3.56397033e-01 4.83892024e-01 -6.97742581e-01 -2.13144496e-01
-5.72282914e-03 6.75093055e-01 -3.54848534e-01 -6.42826200e-01
8.30463529e-01 -2.36377642e-01 -6.23052955e-01 1.73732921e-01
-1.11203742e+00 -6.51557267e-01 3.12295020e-01 8.20673227e-01
7.24984765e-01 -2.01235995e-01 5.86222470e-01 -8.67098212e-01
1.13498247e+00 -7.41147459e-01 -5.97934365e-01 7.16587186e-01
-7.29187608e-01 4.78870660e-01 2.67498493e-01 -2.01795757e-01
-8.45210910e-01 -4.95929360e-01 -5.00613213e-01 -2.91622460e-01
-2.60336138e-03 6.17253602e-01 4.04374719e-01 6.12972081e-02
9.64066267e-01 1.72238667e-02 -1.79492116e-01 -8.32115412e-01
8.25298548e-01 8.80457878e-01 1.98622257e-01 -5.54322004e-01
4.42277908e-01 1.76306963e-01 -3.31096172e-01 -8.92380774e-01
-1.03126061e+00 -9.80671644e-01 -1.59330785e-01 2.46778160e-01
6.59562111e-01 -6.47654772e-01 -7.36549199e-01 -2.36181110e-01
-1.14302540e+00 1.90462172e-01 -6.70576453e-01 2.39737406e-01
-3.49630296e-01 3.63972187e-01 -8.23205888e-01 -4.63072717e-01
-7.26703882e-01 -7.33502507e-01 1.06666493e+00 2.00413033e-01
-2.68751979e-01 -7.68451571e-01 5.91855884e-01 5.51540375e-01
7.11298823e-01 -4.85036105e-01 1.23079991e+00 -9.53728914e-01
-5.91917634e-01 -3.23782086e-01 -4.01615292e-01 1.49005085e-01
-3.71557891e-01 -6.20917439e-01 -7.45201528e-01 -3.47357512e-01
-1.45761400e-01 -8.16638827e-01 1.25309753e+00 -1.05902992e-01
1.26814914e+00 -2.42018715e-01 -2.31795609e-01 3.22327793e-01
1.40419531e+00 -3.69660463e-03 8.64463389e-01 2.45048776e-01
5.09600222e-01 4.98025119e-01 5.32858372e-01 2.15770215e-01
2.75813699e-01 5.49644351e-01 -5.43770455e-02 3.95134389e-01
-3.87460709e-01 -2.74198294e-01 2.19711699e-02 1.15288591e+00
1.22227758e-01 -6.45787716e-01 -9.48243856e-01 7.56007791e-01
-1.78328252e+00 -9.01708722e-01 -3.24494913e-02 2.08349705e+00
7.86731601e-01 -2.59360880e-01 -1.08433604e-01 2.00769156e-01
2.95901924e-01 2.85766333e-01 -3.72726083e-01 -6.48652434e-01
-1.16034433e-01 8.50007415e-01 4.10048783e-01 5.31806052e-01
-7.41289139e-01 9.66006696e-01 7.74219799e+00 1.01890635e+00
-5.31134129e-01 2.15378806e-01 2.23958269e-01 -5.97711146e-01
-5.36636472e-01 -1.76629454e-01 -8.96111250e-01 4.63768020e-02
1.45573783e+00 -4.97579426e-01 3.32153857e-01 6.48258030e-01
-7.69039631e-01 -1.51428252e-01 -9.48251307e-01 1.19276714e+00
2.62301266e-01 -1.64280605e+00 5.12560129e-01 -1.16140708e-01
4.42735374e-01 2.18322888e-01 -2.83922553e-01 7.60161281e-01
2.20317602e-01 -1.05554616e+00 7.68143609e-02 6.75627649e-01
3.48520458e-01 -1.00403476e+00 8.52775872e-01 2.58233100e-01
-9.28904414e-01 -1.30737960e-01 -9.32587028e-01 1.14055723e-01
6.45370260e-02 4.33481872e-01 -3.98724198e-01 3.87946397e-01
8.71483207e-01 2.72720397e-01 -8.41217399e-01 9.49845016e-01
6.20551892e-02 6.50866270e-01 -4.72632825e-01 -6.12076163e-01
2.35195696e-01 2.09745407e-01 3.08867514e-01 1.10924196e+00
1.06637537e-01 4.00341392e-01 -7.21277967e-02 4.39503849e-01
-4.17151421e-01 5.36068082e-01 -4.98673916e-01 -1.82192683e-01
4.81599867e-01 9.04119074e-01 -3.55484605e-01 -5.94653487e-01
-3.91923517e-01 9.14706349e-01 7.91504145e-01 2.27270097e-01
-4.43143725e-01 -9.81542230e-01 5.40542126e-01 1.38085485e-01
6.05306506e-01 -9.05559212e-02 2.27155834e-01 -1.14615369e+00
9.28564295e-02 -9.73900199e-01 9.82827306e-01 -5.92557490e-01
-1.24634814e+00 6.37233496e-01 -1.34805441e-01 -4.61964607e-01
-4.18389559e-01 -4.72283751e-01 1.06610663e-01 7.79126644e-01
-1.74449301e+00 -2.25172266e-01 -1.81198835e-01 8.16598594e-01
4.16103214e-01 -4.30022255e-02 1.24705648e+00 6.95258260e-01
-2.20173180e-01 6.61756396e-01 5.99401832e-01 4.90680896e-02
6.00583673e-01 -1.25846672e+00 1.89251691e-01 3.25825393e-01
6.51949406e-01 8.49521756e-01 3.34036082e-01 -1.82603389e-01
-1.94640815e+00 -6.15752220e-01 1.03850520e+00 -6.40700877e-01
6.90970361e-01 -8.24727565e-02 -1.29082131e+00 2.74075270e-01
4.05061364e-01 -4.82944921e-02 9.04044986e-01 6.81243181e-01
-6.51625872e-01 -1.35872811e-01 -8.20832610e-01 3.62629920e-01
8.12074721e-01 -7.54050136e-01 -1.23519599e+00 6.79616570e-01
9.42849040e-01 -1.67251840e-01 -1.20408237e+00 1.03492886e-01
3.27698916e-01 -6.16734445e-01 1.22211170e+00 -9.61334467e-01
2.03420013e-01 -8.05968046e-03 -4.04358983e-01 -9.88837540e-01
-3.58771235e-01 -4.72235322e-01 -9.14771378e-01 9.10962582e-01
3.88774872e-01 -4.52250570e-01 5.85166454e-01 5.16641617e-01
3.35627526e-01 -8.13784480e-01 -9.60163951e-01 -7.44010329e-01
4.96452034e-01 -1.38860956e-01 7.07296968e-01 5.02257943e-01
2.43765980e-01 1.03789854e+00 1.03120148e-01 -2.70794630e-01
1.77189529e-01 5.35384834e-01 5.65166116e-01 -1.34028053e+00
-2.32631564e-01 -5.01673520e-01 -3.60378385e-01 -1.61272788e+00
5.65264933e-02 -9.71890569e-01 -8.00691247e-02 -1.77810144e+00
2.37086579e-01 -2.70987779e-01 -5.72672606e-01 2.07574576e-01
-3.83175164e-01 1.45971447e-01 6.97577149e-02 3.29198003e-01
-1.29369259e+00 4.70964849e-01 9.32551861e-01 -3.34863901e-01
9.94095579e-02 -5.31536102e-01 -7.25882173e-01 2.08513707e-01
5.02676189e-01 -4.38373387e-01 -5.69463372e-01 -9.09314871e-01
5.84105253e-01 1.84988663e-01 1.17262781e-01 -9.32348013e-01
5.12629092e-01 3.96739006e-01 1.13172732e-01 -7.91157961e-01
4.67612684e-01 -6.13723397e-01 -5.29502749e-01 2.46795326e-01
-7.34119654e-01 2.54682004e-01 2.88076818e-01 6.15437210e-01
-5.66921115e-01 -4.99050021e-01 4.46647495e-01 -4.04381119e-02
-6.65155470e-01 1.09474108e-01 -2.82412291e-01 5.99943757e-01
6.40467584e-01 4.06283528e-01 -6.70925677e-01 -4.19955790e-01
-4.18346316e-01 3.58228207e-01 3.80984768e-02 5.13594866e-01
6.81308150e-01 -1.13649046e+00 -7.02807665e-01 -1.70127273e-01
2.58711994e-01 -2.76118785e-01 2.40808606e-01 2.62523860e-01
-7.82634497e-01 1.13036251e+00 1.70215607e-01 -2.76965767e-01
-1.21689737e+00 9.96527791e-01 -7.00154249e-03 -5.96204698e-01
-6.30172491e-01 9.15121377e-01 -1.95328444e-01 -2.16732472e-01
4.67803657e-01 -1.69482827e-01 -5.91426849e-01 3.73843133e-01
1.12981200e+00 7.61270285e-01 4.23509538e-01 -1.67715475e-01
-2.49572769e-01 4.54456925e-01 -5.92118740e-01 1.38043053e-02
1.16015029e+00 -1.25869885e-01 -3.41086984e-01 1.66904688e-01
1.58972383e+00 -1.41319990e-01 -2.12081730e-01 -5.26111841e-01
3.00189406e-01 -3.72794032e-01 4.44901109e-01 -6.71294689e-01
-8.60542238e-01 9.10658836e-01 6.21218383e-01 6.83961272e-01
1.18377233e+00 1.94037527e-01 1.44425738e+00 1.40034819e+00
5.01710415e-01 -1.20050836e+00 -1.97306424e-01 8.55317593e-01
7.88766026e-01 -9.54579473e-01 1.53625280e-01 -5.67476451e-02
-1.11527845e-01 8.97715688e-01 2.73232847e-01 -1.94526568e-01
8.07192147e-01 -2.34945372e-01 -8.11670721e-02 -6.21483564e-01
-1.09234667e+00 -4.75387961e-01 5.64728439e-01 2.08414525e-01
3.65945399e-01 -1.03102155e-01 -5.76734662e-01 3.67571652e-01
-3.15150082e-01 -1.55360773e-01 4.60669100e-02 1.18038142e+00
-9.76853967e-01 -1.04227638e+00 -3.40764582e-01 8.46860886e-01
-5.45698404e-01 -3.19896817e-01 -4.42452371e-01 4.13186729e-01
-5.04008234e-01 1.22083247e+00 2.54160017e-01 -1.41984478e-01
4.31489795e-01 2.88869500e-01 5.73619068e-01 -6.79937720e-01
-6.58858776e-01 -3.77119124e-01 2.02554703e-01 -8.22238445e-01
-1.11819513e-01 -2.01115072e-01 -1.08799052e+00 -4.65244591e-01
-6.54830694e-01 7.78754115e-01 3.20359170e-01 7.43901491e-01
7.48297691e-01 4.41376090e-01 2.13574857e-01 9.14499089e-02
-8.05264533e-01 -9.52659726e-01 -3.41681004e-01 3.23233724e-01
1.86227694e-01 -4.06934917e-01 -2.62496829e-01 -3.28080535e-01] | [11.400516510009766, 7.703930854797363] |
f1e666c5-b9ac-4274-af90-09e70dc65d0c | affine-correspondences-between-multi-camera | 2306.12996 | null | https://arxiv.org/abs/2306.12996v1 | https://arxiv.org/pdf/2306.12996v1.pdf | Affine Correspondences between Multi-Camera Systems for Relative Pose Estimation | We present a novel method to compute the relative pose of multi-camera systems using two affine correspondences (ACs). Existing solutions to the multi-camera relative pose estimation are either restricted to special cases of motion, have too high computational complexity, or require too many point correspondences (PCs). Thus, these solvers impede an efficient or accurate relative pose estimation when applying RANSAC as a robust estimator. This paper shows that the 6DOF relative pose estimation problem using ACs permits a feasible minimal solution, when exploiting the geometric constraints between ACs and multi-camera systems using a special parameterization. We present a problem formulation based on two ACs that encompass two common types of ACs across two views, i.e., inter-camera and intra-camera. Moreover, the framework for generating the minimal solvers can be extended to solve various relative pose estimation problems, e.g., 5DOF relative pose estimation with known rotation angle prior. Experiments on both virtual and real multi-camera systems prove that the proposed solvers are more efficient than the state-of-the-art algorithms, while resulting in a better relative pose accuracy. Source code is available at https://github.com/jizhaox/relpose-mcs-depth. | ['Ji Zhao', 'Banglei Guan'] | 2023-06-22 | null | null | null | null | ['pose-estimation'] | ['computer-vision'] | [-9.66660157e-02 -2.97885299e-01 -1.54961467e-01 -4.16276678e-02
-9.68254209e-01 -8.98486435e-01 3.52590889e-01 -2.61896700e-01
-2.00122550e-01 4.17446733e-01 -2.82206684e-01 2.39349827e-02
-2.30389282e-01 -4.93127197e-01 -8.63036633e-01 -6.07265770e-01
3.86059344e-01 5.15607059e-01 2.47852772e-01 -2.87588269e-01
4.18669432e-01 6.43114865e-01 -1.35695684e+00 -4.61261898e-01
6.56007886e-01 8.12700808e-01 3.63781154e-01 7.26872504e-01
5.35394728e-01 1.59785539e-01 -3.03151280e-01 -3.02778870e-01
6.04988933e-01 -2.16598839e-01 -6.52138948e-01 4.86204177e-01
8.86216819e-01 -4.75328982e-01 -8.06739777e-02 1.01736093e+00
4.56608593e-01 1.20350890e-01 1.05109558e-01 -1.44329238e+00
-7.27333501e-02 -2.52047032e-01 -1.00766647e+00 -5.96112199e-02
1.05889845e+00 -2.15715379e-01 7.19570160e-01 -1.04234409e+00
6.77585661e-01 1.00403154e+00 7.50134706e-01 2.71861047e-01
-9.44561243e-01 -2.89443016e-01 1.32633850e-01 6.13443879e-03
-1.80878806e+00 -2.61083812e-01 7.61037886e-01 -3.91435623e-01
6.39312387e-01 6.18454516e-01 7.55158186e-01 8.88459802e-01
3.17478955e-01 3.86316657e-01 9.02649999e-01 -3.21833283e-01
-1.12692982e-01 -1.08164378e-01 -1.71564370e-01 5.65359056e-01
5.26825428e-01 -1.01426475e-01 -3.44971061e-01 -3.97566378e-01
1.33002782e+00 2.32378289e-01 -5.54679692e-01 -8.65397274e-01
-1.61277044e+00 7.46019006e-01 2.00778201e-01 4.80457675e-03
-5.75432181e-02 1.24893002e-01 -5.98199815e-02 -1.37843743e-01
2.61693299e-01 4.51848716e-01 -4.51658100e-01 -5.26650362e-02
-6.37953222e-01 4.30091828e-01 5.45119405e-01 1.62185752e+00
8.97378385e-01 -2.37641305e-01 7.14545906e-01 5.47920763e-01
3.56302351e-01 7.66730189e-01 2.42087364e-01 -1.25768542e+00
6.72217667e-01 2.78359205e-01 4.41710889e-01 -1.38775134e+00
-4.29822505e-01 -2.72901326e-01 -6.64209723e-01 -7.05091432e-02
3.43348891e-01 -1.85193911e-01 -3.34380537e-01 1.30281138e+00
7.66872346e-01 2.93896616e-01 -4.38077003e-02 1.10413432e+00
5.34188330e-01 5.42620540e-01 -9.19422805e-01 -4.61661428e-01
1.27987909e+00 -9.56898391e-01 -7.82409191e-01 -4.72544044e-01
5.66265047e-01 -1.33189607e+00 5.92795610e-01 4.59038526e-01
-1.18367493e+00 -4.02190328e-01 -1.15975368e+00 -8.88780504e-02
-2.43201386e-02 3.88318688e-01 4.61850375e-01 4.26926851e-01
-9.78595555e-01 3.15891653e-01 -8.08104157e-01 -4.21995282e-01
-4.12615418e-01 5.24714828e-01 -5.69294035e-01 -5.36555275e-02
-6.51353717e-01 1.13487864e+00 2.22991616e-01 3.16624731e-01
-4.32455122e-01 -4.55377996e-01 -9.87663209e-01 -3.49205822e-01
7.62948632e-01 -9.33323860e-01 1.21797073e+00 -6.96435571e-01
-1.62241316e+00 7.35285044e-01 -2.10151270e-01 2.16154873e-01
7.19660163e-01 -5.32829344e-01 -1.09917643e-02 2.27315739e-01
2.34381542e-01 3.85287523e-01 5.63993633e-01 -1.37017131e+00
-5.11479676e-01 -3.84173751e-01 3.26894134e-01 6.47280753e-01
2.46849060e-02 8.05020928e-02 -9.17852044e-01 -4.43754762e-01
7.01471031e-01 -1.46876657e+00 -6.00286722e-01 8.07306319e-02
-4.21791226e-01 3.25986445e-01 8.71436357e-01 -4.69281793e-01
1.01784623e+00 -1.90427458e+00 5.96991777e-01 1.28151357e-01
-3.98188494e-02 -1.84858412e-01 5.08683883e-02 4.25914168e-01
-1.27131745e-01 -2.25020021e-01 6.66407645e-02 -4.01596904e-01
-4.59331244e-01 2.06696123e-01 1.58311613e-02 1.12062132e+00
-3.80440444e-01 2.42918029e-01 -7.95549870e-01 -4.72030222e-01
5.59895098e-01 4.71293181e-01 -4.56500292e-01 2.65967101e-01
2.20340133e-01 6.02307260e-01 -4.46989000e-01 5.29359400e-01
1.09241092e+00 -1.26303017e-01 1.51801497e-01 -4.59286004e-01
-4.18122888e-01 -3.01826984e-01 -2.01564407e+00 1.86864555e+00
-4.64021444e-01 2.78222054e-01 2.99481928e-01 -4.51889753e-01
7.86283374e-01 3.22685450e-01 7.18859375e-01 -5.18417656e-02
3.62138599e-01 3.56368601e-01 -3.49510968e-01 -2.06452444e-01
8.00855219e-01 1.00138254e-01 -1.02591373e-01 6.65803626e-02
-1.28566459e-01 -5.64631104e-01 2.53853351e-01 1.19803706e-02
5.23514271e-01 4.05664563e-01 7.44950235e-01 -2.30739921e-01
8.30513656e-01 -8.32485501e-03 8.39442551e-01 1.13706283e-01
6.10030405e-02 1.03840446e+00 1.23560034e-01 -3.72052968e-01
-1.05082345e+00 -7.81055868e-01 -1.43021852e-01 2.69870967e-01
8.15044105e-01 -6.05667174e-01 -7.31856406e-01 -1.02263510e-01
-1.14919804e-01 1.44809112e-01 -1.52233452e-01 3.12361598e-01
-7.49409437e-01 -5.37683129e-01 -5.98232783e-02 4.79218245e-01
3.98205459e-01 1.21404611e-01 -8.26676071e-01 -1.21125273e-01
-4.90681529e-01 -1.66268218e+00 -7.37183034e-01 -3.67577225e-01
-1.01293206e+00 -1.38644743e+00 -9.00353312e-01 -5.82106590e-01
9.11473811e-01 8.75965416e-01 6.74860954e-01 -6.64295554e-02
-4.04703468e-02 8.38112414e-01 -2.79579192e-01 6.14033565e-02
4.83569503e-02 -1.81196287e-01 4.41996098e-01 -5.29557690e-02
-2.53876865e-01 -3.29865277e-01 -6.03823662e-01 1.05223548e+00
-6.26508653e-01 2.53212273e-01 3.52520525e-01 5.51431000e-01
9.58725274e-01 -3.27335030e-01 -3.10751885e-01 -4.91010576e-01
9.54744890e-02 -2.51371205e-01 -1.11557209e+00 1.34932995e-01
-1.64096028e-01 -3.71261448e-01 4.95793611e-01 -5.06884634e-01
-1.00849438e+00 6.84498549e-01 7.12809563e-02 -8.35930288e-01
-2.40089558e-02 2.90491074e-01 -2.65807003e-01 -4.83758360e-01
4.45329666e-01 -2.50317510e-02 -6.51586130e-02 -2.81002074e-01
2.71072865e-01 4.39502865e-01 4.05637354e-01 -5.06261587e-01
1.00175047e+00 7.57585943e-01 4.20692384e-01 -1.03761780e+00
-7.13333666e-01 -1.00384700e+00 -1.03504014e+00 -2.75015801e-01
8.49605322e-01 -1.15039468e+00 -7.69739032e-01 4.06954527e-01
-1.39478874e+00 2.39759400e-01 3.25824231e-01 9.24178541e-01
-9.25842047e-01 7.95480907e-01 -4.14781719e-01 -5.09899676e-01
-6.01550043e-02 -1.56767845e+00 1.44381869e+00 1.03223599e-01
-1.94511160e-01 -1.02630663e+00 2.47096032e-01 5.21706104e-01
-1.18812531e-01 5.96063137e-01 -2.03293692e-02 1.86068937e-01
-8.65984380e-01 -3.16704959e-01 2.73793161e-01 -9.05085579e-02
-3.72524373e-02 4.73266453e-01 -5.85601807e-01 -6.37451053e-01
1.83411181e-01 1.40689284e-01 -7.62232468e-02 4.84641165e-01
7.65833616e-01 -1.72039360e-01 -5.56591928e-01 9.90119040e-01
1.81049395e+00 8.31694435e-03 5.18614173e-01 6.17552280e-01
8.82146001e-01 2.75769770e-01 1.20392418e+00 6.47786438e-01
5.22832274e-01 1.26336634e+00 8.63116086e-01 7.69499838e-02
4.04194862e-01 4.77206111e-02 3.05692017e-01 1.11792588e+00
-4.17140186e-01 -4.53719981e-02 -9.57519889e-01 3.70641023e-01
-2.04468918e+00 -5.84674716e-01 -7.29877651e-01 2.51776958e+00
2.62987733e-01 -4.65861827e-01 7.27683902e-02 -8.03856850e-02
9.78742540e-01 6.12725578e-02 -2.21200556e-01 -1.43637180e-01
1.17335748e-02 -4.26358968e-01 9.04349506e-01 7.44252086e-01
-1.09604609e+00 7.56242216e-01 5.66653776e+00 4.37932014e-01
-9.00836408e-01 8.75746086e-02 3.46343964e-02 -1.01157472e-01
-2.34119385e-03 3.56278598e-01 -1.16935909e+00 1.93518832e-01
5.94642937e-01 -1.22028962e-01 3.03295732e-01 1.02632904e+00
9.98917073e-02 -3.90879005e-01 -9.74151671e-01 1.52514434e+00
4.18059766e-01 -1.27297556e+00 -1.77845404e-01 2.31338099e-01
8.69655728e-01 -8.37963820e-02 -1.67151973e-01 -4.30129468e-01
-1.32846117e-01 -3.36096108e-01 9.07920659e-01 3.11270714e-01
7.31130898e-01 -6.71514630e-01 7.07306325e-01 6.45812690e-01
-1.44281685e+00 7.66909271e-02 -6.79270685e-01 -3.24108936e-02
3.47878128e-01 4.09711748e-01 -2.99403429e-01 1.11867583e+00
6.61989629e-01 7.88544178e-01 -4.81811583e-01 1.17524552e+00
-7.28253499e-02 -5.03295243e-01 -4.72522140e-01 2.52107203e-01
1.21505216e-01 -6.66102469e-01 7.64967144e-01 5.98361909e-01
8.41183305e-01 2.01249272e-01 3.10982198e-01 4.24689770e-01
4.23526287e-01 1.85166538e-01 -7.60179222e-01 5.83555162e-01
4.04378802e-01 1.51883650e+00 -7.57139921e-01 -5.22967689e-02
-6.53514564e-01 1.08121645e+00 8.38118643e-02 2.43398234e-01
-1.10816264e+00 -3.09376214e-02 6.77184284e-01 1.64673716e-01
1.53892368e-01 -6.88068926e-01 -9.11625475e-02 -1.60357082e+00
2.77405173e-01 -8.50149095e-01 2.82152414e-01 -9.93671775e-01
-6.98980808e-01 5.20554245e-01 3.25173944e-01 -2.03778291e+00
-3.56291622e-01 -7.32162654e-01 -4.63145107e-01 6.84946895e-01
-1.16685891e+00 -1.19326317e+00 -5.46912432e-01 7.50233233e-01
5.34328699e-01 2.75993139e-01 6.60742640e-01 3.01251441e-01
-5.90448618e-01 2.59832531e-01 1.26685232e-01 -1.89382479e-01
7.58398533e-01 -9.68661606e-01 -1.94372553e-02 1.07959294e+00
-1.36562929e-01 7.40557253e-01 7.17030406e-01 -3.97467434e-01
-2.09172511e+00 -7.62840629e-01 4.47944134e-01 -7.47351825e-01
4.97133791e-01 -2.64924943e-01 -3.18180293e-01 1.22964513e+00
6.00853041e-02 2.94656545e-01 4.91102070e-01 -1.89914808e-01
7.32469931e-02 -1.15506098e-01 -1.03209591e+00 5.30047536e-01
8.38496327e-01 -2.64465749e-01 -2.98744828e-01 4.57039714e-01
4.64243442e-01 -1.21367502e+00 -1.04715157e+00 2.30667308e-01
4.78287339e-01 -9.00002062e-01 1.30368412e+00 1.86714709e-01
1.70830742e-01 -7.40317762e-01 -3.86935025e-01 -1.22857165e+00
-1.35078609e-01 -8.96561563e-01 3.89001885e-04 9.92589414e-01
-7.88504183e-02 -7.10867643e-01 5.23745060e-01 3.74290854e-01
-1.57578915e-01 -6.24660313e-01 -1.07239318e+00 -9.20331657e-01
-1.87625930e-01 -2.42157906e-01 4.02825475e-01 1.13525653e+00
-1.10470325e-01 3.01560014e-01 -5.56110263e-01 8.27296257e-01
7.12259233e-01 2.45051369e-01 1.31339765e+00 -1.05048311e+00
-3.19129139e-01 -9.44694206e-02 -6.07504308e-01 -1.28731978e+00
-1.22826733e-01 -3.23345393e-01 -9.56045836e-02 -1.31483078e+00
6.79196939e-02 -2.42536038e-01 6.04476571e-01 -9.08750370e-02
-1.42705841e-02 1.33410290e-01 3.55361283e-01 3.63327444e-01
-4.56218392e-01 2.63286263e-01 1.46711636e+00 3.50208104e-01
7.43890181e-02 1.77149609e-01 -3.28038007e-01 1.12035227e+00
5.19541264e-01 -2.55982101e-01 -3.51795077e-01 -6.53509617e-01
5.21089911e-01 5.77926695e-01 3.49511147e-01 -1.02943182e+00
4.42675143e-01 -3.93274099e-01 2.88247705e-01 -8.38424265e-01
8.24868083e-01 -1.15987813e+00 6.98856533e-01 3.61944497e-01
4.56312537e-01 5.61068654e-01 6.75956756e-02 3.44476104e-01
-4.04939562e-01 -4.45188135e-01 7.31813490e-01 -3.38362455e-01
-6.29587114e-01 4.10585642e-01 -8.57022926e-02 -2.74793386e-01
1.52718055e+00 -5.90259254e-01 -2.95278400e-01 -5.11937320e-01
-6.75055921e-01 1.93644330e-01 9.12395298e-01 3.39986354e-01
7.33168781e-01 -1.49228752e+00 -3.40951771e-01 2.11954806e-02
1.52599186e-01 3.93438220e-01 3.28955680e-01 1.17127657e+00
-1.10068440e+00 4.98771310e-01 -1.11723162e-01 -1.11030734e+00
-1.58749294e+00 5.79201043e-01 4.67098862e-01 -7.72893196e-04
-3.90250146e-01 5.39775550e-01 1.79201916e-01 -6.62076712e-01
-2.05676168e-01 -4.14314836e-01 -5.30768000e-02 -1.24301851e-01
2.28675291e-01 8.51750612e-01 -6.24294356e-02 -1.18765247e+00
-5.69707990e-01 1.60867655e+00 2.56085336e-01 -2.10312270e-02
1.14781511e+00 -5.25429785e-01 -1.42364621e-01 9.03134868e-02
1.32589257e+00 3.19846660e-01 -1.17190373e+00 1.63348451e-01
-3.82661223e-01 -1.03397346e+00 -3.94292772e-02 1.46140859e-01
-1.02699935e+00 6.20486736e-01 4.71873790e-01 -1.70512125e-01
9.69577134e-01 -1.36730418e-01 4.99740928e-01 3.94170433e-01
7.52223074e-01 -9.70923424e-01 -4.21713218e-02 4.39139396e-01
1.18570030e+00 -1.19711626e+00 6.14284575e-01 -1.15910006e+00
-5.17056227e-01 1.42634439e+00 8.14246058e-01 -3.07221413e-01
4.71612245e-01 3.22892427e-01 1.02353744e-01 -5.47102187e-03
-3.02597046e-01 1.06119961e-01 2.55976915e-01 2.29414687e-01
3.79783660e-01 -6.70538247e-02 -2.36622110e-01 -5.71383499e-02
-1.26599655e-01 -3.95701379e-01 1.03656960e+00 1.11880088e+00
-3.20015699e-02 -1.06947362e+00 -9.59432125e-01 -2.37049282e-01
-1.88738719e-01 3.63902420e-01 -2.28859633e-01 1.11505210e+00
7.49853952e-03 7.89169967e-01 6.72979802e-02 -2.08535284e-01
5.36035120e-01 -6.57916605e-01 7.26295769e-01 -5.45784056e-01
-2.62257725e-01 5.92534304e-01 -5.16032893e-03 -9.97758865e-01
-7.50366092e-01 -8.82886708e-01 -9.60436404e-01 -3.91117036e-01
-8.14285100e-01 -6.02508932e-02 5.95418274e-01 6.81227028e-01
3.04507434e-01 1.20642327e-03 7.15462029e-01 -1.25587904e+00
-4.30167943e-01 -6.79298580e-01 -5.26358783e-01 1.39589190e-01
3.97708982e-01 -8.95079494e-01 -3.88052016e-01 -1.29504070e-01] | [7.833804130554199, -2.292947292327881] |
6e65ad9d-18f5-4429-843d-036a2aac9a89 | frequency-aware-face-hallucination-generative | 2110.01880 | null | https://arxiv.org/abs/2110.01880v1 | https://arxiv.org/pdf/2110.01880v1.pdf | Frequency Aware Face Hallucination Generative Adversarial Network with Semantic Structural Constraint | In this paper, we address the issue of face hallucination. Most current face hallucination methods rely on two-dimensional facial priors to generate high resolution face images from low resolution face images. These methods are only capable of assimilating global information into the generated image. Still there exist some inherent problems in these methods; such as, local features, subtle structural details and missing depth information in final output image. Present work proposes a Generative Adversarial Network (GAN) based novel progressive Face Hallucination (FH) network to address these issues present among current methods. The generator of the proposed model comprises of FH network and two sub-networks, assisting FH network to generate high resolution images. The first sub-network leverages on explicitly adding high frequency components into the model. To explicitly encode the high frequency components, an auto encoder is proposed to generate high resolution coefficients of Discrete Cosine Transform (DCT). To add three dimensional parametric information into the network, second sub-network is proposed. This network uses a shape model of 3D Morphable Models (3DMM) to add structural constraint to the FH network. Extensive experimentation results in the paper shows that the proposed model outperforms the state-of-the-art methods. | ['Vinay Kumar', 'Abhinav Dhall', 'Shailza Sharma'] | 2021-10-05 | null | null | null | null | ['face-hallucination'] | ['computer-vision'] | [ 5.32844365e-01 8.09653103e-01 2.78791159e-01 -1.35042593e-01
-5.15881181e-01 -1.14885248e-01 7.85458744e-01 -1.08710825e+00
2.93596953e-01 8.37051690e-01 4.98511434e-01 4.25306052e-01
1.85063317e-01 -1.08106554e+00 -6.19692564e-01 -6.48128450e-01
4.89650220e-02 8.55699554e-02 -2.94477612e-01 -3.50831926e-01
-6.79516941e-02 7.04284787e-01 -1.43458974e+00 4.68850195e-01
6.51563704e-01 9.19131398e-01 1.64718088e-03 4.39634711e-01
9.65479463e-02 7.47247100e-01 -6.15343034e-01 -4.00312245e-01
5.83474278e-01 -7.78480649e-01 -4.38086390e-01 3.14407170e-01
3.44097435e-01 -6.38399482e-01 -4.26433623e-01 1.09855390e+00
6.14985704e-01 7.49964416e-02 8.76204193e-01 -1.30985427e+00
-1.24665821e+00 1.66953608e-01 -7.44131207e-01 -1.08793966e-01
6.83057547e-01 7.30649009e-02 3.02670181e-01 -1.27293766e+00
7.59802938e-01 1.77668583e+00 8.02729905e-01 9.47167158e-01
-1.31857860e+00 -8.36365700e-01 -4.16244566e-01 -1.44135267e-01
-1.61005771e+00 -8.21956873e-01 1.28388798e+00 -2.92607516e-01
6.55124843e-01 1.06927559e-01 5.77748239e-01 1.21208799e+00
3.39685738e-01 8.14534202e-02 1.29777718e+00 -3.42609227e-01
6.33479729e-02 1.48732066e-01 -1.00373137e+00 8.75480771e-01
-1.03505030e-01 5.12239039e-01 -5.83134472e-01 -1.23342365e-01
1.54418850e+00 -5.85477017e-02 -2.90558875e-01 3.07073817e-02
-5.15170455e-01 9.50363696e-01 4.80282933e-01 3.25758308e-02
-6.39965117e-01 -1.04028553e-01 -3.49449247e-01 2.36726701e-01
5.12505472e-01 2.30893373e-01 1.84256956e-01 4.53014165e-01
-1.13067031e+00 2.26249918e-01 6.31075144e-01 8.85326326e-01
9.06044543e-01 6.48043573e-01 -2.35920489e-01 9.31710243e-01
6.08668625e-01 2.66651839e-01 4.47561651e-01 -1.19390750e+00
1.69757232e-01 3.22228611e-01 -2.21748039e-01 -1.24769676e+00
1.52344152e-01 -3.36981326e-01 -1.14575958e+00 4.84448880e-01
-4.55909222e-01 -1.57482877e-01 -1.26117194e+00 1.88069534e+00
4.02718186e-01 6.19358122e-01 1.98896393e-01 9.30721641e-01
1.02080429e+00 7.09756970e-01 -2.26415232e-01 -3.27349603e-01
1.08504474e+00 -8.35655153e-01 -8.69661391e-01 -5.58307394e-02
-4.51616615e-01 -8.35897923e-01 6.25619471e-01 3.47887091e-02
-1.52271152e+00 -8.14297855e-01 -1.11319900e+00 -6.75687147e-03
-1.11985557e-01 -1.84788167e-01 2.60883927e-01 6.60794795e-01
-1.43176365e+00 4.07507747e-01 -3.03301096e-01 -8.12812448e-02
6.74750686e-01 4.16167915e-01 -6.62485003e-01 -1.70431823e-01
-1.22004855e+00 8.40869486e-01 1.22967504e-01 2.30048239e-01
-1.21466041e+00 -8.35922301e-01 -1.15828037e+00 -5.96333742e-02
-1.52924761e-01 -8.92518163e-01 7.53235281e-01 -1.10718715e+00
-1.88024640e+00 7.39758313e-01 -1.34768441e-01 -4.29965556e-02
3.83304656e-01 1.59658924e-01 -5.62725246e-01 5.19656837e-01
9.62755159e-02 1.05807900e+00 1.36060047e+00 -1.72554684e+00
8.95869434e-02 -2.14861214e-01 -2.14471385e-01 2.37683192e-01
-5.71472198e-02 -2.03659549e-01 -1.99380279e-01 -9.72807229e-01
5.69391027e-02 -7.14703500e-01 -7.05539435e-02 1.24733791e-01
-3.60817879e-01 3.96995097e-01 1.27178311e+00 -9.02900100e-01
7.68289030e-01 -1.99086034e+00 1.42999679e-01 9.35200602e-02
2.55546331e-01 2.74506778e-01 -4.50741380e-01 3.64345253e-01
-3.17424834e-01 2.33663842e-01 -3.47127765e-01 -6.32777810e-01
-1.71294272e-01 2.56759316e-01 -3.22767705e-01 4.37981993e-01
6.80677056e-01 1.01381087e+00 -5.11863649e-01 -3.69782805e-01
2.29058295e-01 1.35863721e+00 -8.30587447e-01 5.50567389e-01
-1.01298913e-01 8.45292449e-01 -3.06226283e-01 6.43152654e-01
1.26028287e+00 6.68081120e-02 -1.63656786e-01 -3.76088649e-01
1.34454772e-01 -2.45435253e-01 -9.74514186e-01 1.57257020e+00
-3.32868040e-01 1.95830762e-01 3.40772629e-01 -8.15020740e-01
1.09699738e+00 6.73371255e-01 3.34516823e-01 -6.15834951e-01
1.89859539e-01 -8.82729888e-02 -2.55409122e-01 -5.13904393e-01
2.24982664e-01 -6.58892214e-01 4.00159508e-01 2.41332144e-01
2.59084225e-01 -4.46399271e-01 -4.93539602e-01 -1.27430081e-01
8.12416553e-01 3.32634509e-01 1.51351064e-01 2.93798167e-02
6.92335367e-01 -6.82717264e-01 6.00211680e-01 1.11335233e-01
2.29137819e-02 1.19739592e+00 1.59938157e-01 -2.47411102e-01
-1.24147487e+00 -1.15713263e+00 7.89000764e-02 2.29576111e-01
-2.12492481e-01 -7.53131285e-02 -9.59381759e-01 -3.02542418e-01
-2.74190277e-01 3.04837376e-01 -1.00749516e+00 -3.04049760e-01
-4.77849245e-01 -3.65008026e-01 7.04450130e-01 3.34097117e-01
9.98643935e-01 -1.16406322e+00 -3.20458531e-01 2.58462969e-02
-9.85884890e-02 -1.23772323e+00 -4.69060749e-01 -5.57884455e-01
-6.30966723e-01 -6.86852932e-01 -8.24967206e-01 -8.91385734e-01
9.56546426e-01 -1.32437674e-02 8.04298162e-01 -6.87445998e-02
-5.25608599e-01 1.76030084e-01 -1.62431538e-01 -1.99576408e-01
-5.74526787e-01 -5.24211168e-01 5.22684259e-03 3.33200365e-01
-5.44054359e-02 -1.13558209e+00 -7.07232118e-01 1.04369096e-01
-1.15875363e+00 1.75624147e-01 7.91269779e-01 8.19518805e-01
5.19468367e-01 2.05026984e-01 7.93194413e-01 -8.39804471e-01
6.19847000e-01 -6.25101984e-01 -1.38172016e-01 -1.96965143e-01
-3.28555137e-01 -5.42307692e-03 7.58064985e-01 -5.30035317e-01
-1.33980525e+00 -4.87972051e-02 -3.86568934e-01 -9.06776071e-01
-1.21937469e-01 1.68807030e-01 -5.09443939e-01 -2.86885828e-01
4.97269183e-01 4.44786727e-01 4.47238028e-01 -4.07229275e-01
3.38526726e-01 4.60512608e-01 6.37982845e-01 -3.13165307e-01
1.25442076e+00 6.20501041e-01 2.60991454e-01 -7.99973428e-01
-6.27086818e-01 3.84692281e-01 -5.11820614e-01 -2.31424674e-01
8.93542051e-01 -1.07889664e+00 -4.03517693e-01 3.24737310e-01
-1.12416291e+00 -1.53954387e-01 -3.01817447e-01 7.30788186e-02
-8.22120190e-01 7.96420500e-02 -6.29175961e-01 -7.99749374e-01
-4.07131851e-01 -1.07753682e+00 1.09733415e+00 4.17690039e-01
-1.20825749e-02 -9.47457731e-01 -3.59721929e-02 4.98049587e-01
9.09874558e-01 9.14332211e-01 8.28394234e-01 1.29811078e-01
-5.37339926e-01 5.83703071e-02 -1.00115046e-01 4.60022837e-01
5.00208735e-01 -2.86047369e-01 -1.19473314e+00 -4.08307314e-01
3.15694273e-01 -2.74505407e-01 5.29722750e-01 3.14899296e-01
1.01738596e+00 -7.93014705e-01 -8.75786617e-02 9.09890354e-01
1.50579393e+00 1.83902115e-01 1.10551953e+00 -1.93609849e-01
7.68151104e-01 6.44140720e-01 1.30210355e-01 4.66361880e-01
2.38524854e-01 5.86833954e-01 5.61219394e-01 -2.95853108e-01
-6.51262879e-01 -7.15083361e-01 4.74523813e-01 5.46982706e-01
-2.91347027e-01 4.02930640e-02 -3.49882394e-01 2.75829375e-01
-1.35326111e+00 -1.18497252e+00 5.44972122e-01 1.77932525e+00
9.07618821e-01 -4.34209585e-01 -2.26260245e-01 5.93674369e-02
7.56988883e-01 2.04552218e-01 -4.64077264e-01 -2.83245832e-01
-1.25526950e-01 6.62028730e-01 -1.64507791e-01 6.88904703e-01
-6.10180080e-01 8.87820959e-01 6.55795717e+00 5.98475397e-01
-1.15398896e+00 9.46813971e-02 6.24490023e-01 1.23530971e-02
-5.11331499e-01 -1.29578263e-01 -5.60871720e-01 3.46018285e-01
7.85387635e-01 -1.28497398e-02 5.98890662e-01 5.44142604e-01
3.59483242e-01 2.92671859e-01 -7.92832553e-01 9.93008614e-01
6.13985598e-01 -1.22684276e+00 6.31229520e-01 3.30614626e-01
9.94422376e-01 -7.44305849e-01 5.65676153e-01 1.40098140e-01
7.50998706e-02 -1.79881382e+00 4.88084704e-01 7.95724332e-01
1.41854370e+00 -1.20829809e+00 4.07441288e-01 1.40576258e-01
-1.10116470e+00 9.25937220e-02 -4.91112828e-01 1.58563748e-01
1.74724281e-01 4.38125730e-01 -7.77722478e-01 5.05145431e-01
4.76806074e-01 4.32852447e-01 -2.61925817e-01 3.16555589e-01
-1.91778913e-01 1.87124461e-01 1.79246794e-02 9.06393111e-01
-9.81358513e-02 -2.63885438e-01 6.88313663e-01 7.19459713e-01
6.12464249e-01 3.94896746e-01 -1.50484577e-01 1.39798236e+00
-3.14909071e-01 -3.17254573e-01 -1.01566851e+00 9.65649188e-02
4.23141301e-01 1.32590997e+00 -2.68188119e-01 -5.52522466e-02
-3.27505618e-01 1.20585895e+00 1.34547472e-01 3.97069812e-01
-7.69176126e-01 -1.17006242e-01 6.41335905e-01 5.79625726e-01
2.93683380e-01 -7.02704713e-02 8.17718506e-02 -9.45137084e-01
-1.52369410e-01 -1.07015359e+00 -2.89854798e-02 -8.83398652e-01
-1.26968420e+00 8.96922350e-01 -1.07774943e-01 -1.07702065e+00
-5.45042634e-01 -1.32446289e-01 -6.52866364e-01 1.22848082e+00
-1.83775914e+00 -1.57913816e+00 -5.42232931e-01 9.49871659e-01
6.33973420e-01 -5.31906009e-01 9.26579714e-01 2.29908958e-01
-3.22670996e-01 7.34237313e-01 -5.76233327e-01 1.23249181e-01
6.36484027e-01 -6.71271324e-01 2.25908816e-01 8.30249727e-01
-4.09218222e-01 7.86508858e-01 6.24735832e-01 -8.24216485e-01
-1.52145672e+00 -1.44614136e+00 5.02870321e-01 -1.05889976e-01
-1.34775877e-01 -2.70955950e-01 -8.23781013e-01 7.19121397e-01
5.01452148e-01 3.08147162e-01 7.97328889e-01 -8.21232677e-01
-3.87844265e-01 1.99641928e-01 -1.95360947e+00 5.07174194e-01
1.03593397e+00 -5.49205601e-01 -4.79029983e-01 -1.22596614e-01
6.28298342e-01 -3.38306606e-01 -1.08258498e+00 6.03427768e-01
5.09302974e-01 -1.07215834e+00 1.10810280e+00 -1.80770785e-01
8.86981189e-01 -3.84855211e-01 -2.19766974e-01 -1.43919790e+00
-3.80300909e-01 -9.55177963e-01 -4.47096407e-01 1.27724838e+00
8.02365690e-02 -5.40199578e-01 8.28974783e-01 4.45629239e-01
-6.60397783e-02 -7.15787172e-01 -7.92564988e-01 -5.52147865e-01
1.08896263e-01 2.08031505e-01 8.78274918e-01 1.10562968e+00
-4.84000087e-01 4.09500390e-01 -6.78042352e-01 3.05855125e-01
8.76132011e-01 -2.57475138e-01 6.03919029e-01 -1.05805600e+00
-3.00627023e-01 3.73028219e-02 -3.73207718e-01 -6.42829001e-01
3.62213016e-01 -7.59709060e-01 -1.43394858e-01 -1.42221940e+00
2.25603536e-01 -1.07917540e-01 2.69805878e-01 3.22909296e-01
1.57092392e-01 8.42171848e-01 1.48283362e-01 8.38619098e-02
3.21497113e-01 9.48967278e-01 1.71395183e+00 2.34417111e-01
-1.24369308e-01 -3.57683003e-01 -9.10658598e-01 6.76001608e-01
7.34224856e-01 -2.42983580e-01 -7.01725960e-01 -2.45672256e-01
-1.99690223e-01 4.95475799e-01 5.44537365e-01 -1.07015610e+00
1.13779493e-01 -1.63603902e-01 1.04592109e+00 -2.60478050e-01
9.29841816e-01 -7.37568378e-01 6.70211732e-01 1.82816833e-01
-7.60057420e-02 5.30707389e-02 2.50644982e-01 3.88279229e-01
-4.48253006e-01 1.83570281e-01 1.35392737e+00 -3.46858174e-01
-1.56823903e-01 5.92091978e-01 -1.78948686e-01 -1.87423393e-01
1.03006876e+00 -5.66369534e-01 -1.02014825e-01 -6.56467438e-01
-7.91616917e-01 -3.68667722e-01 6.03937924e-01 5.09614050e-01
1.12062192e+00 -1.71831703e+00 -9.46631968e-01 8.11757982e-01
-4.35705721e-01 1.69481322e-01 3.65968108e-01 4.59230036e-01
-3.76350105e-01 9.10600796e-02 -6.73493564e-01 -2.81251334e-02
-1.03689313e+00 3.39207768e-01 3.87219697e-01 -1.11943986e-02
-6.97243392e-01 7.20602334e-01 4.13803935e-01 -3.45740557e-01
-2.04559162e-01 4.70337868e-01 -4.17399794e-01 -1.23152800e-01
6.42334700e-01 1.37810647e-01 -3.59569579e-01 -1.12644494e+00
-6.55595064e-02 8.29708040e-01 4.53794859e-02 -4.39708143e-01
1.59245515e+00 -1.04993396e-01 -3.02228302e-01 -2.28896692e-01
1.34027076e+00 -6.49814261e-03 -1.46814775e+00 -5.80614470e-02
-7.84546435e-01 -6.86574399e-01 7.81800747e-02 -6.74426734e-01
-1.45775676e+00 7.88045883e-01 4.84287888e-01 -2.68571466e-01
1.49708652e+00 -3.37189704e-01 7.80300498e-01 -4.08942103e-01
2.08841667e-01 -6.98686242e-01 5.83566070e-01 3.21444780e-01
1.29816687e+00 -8.75859439e-01 -6.55807182e-02 -6.92146063e-01
-4.62536961e-01 9.73819017e-01 7.29216099e-01 -4.37928289e-01
7.33372450e-01 5.13002753e-01 -1.39151111e-01 -1.79869249e-01
-6.62315071e-01 2.32735977e-01 3.08441699e-01 1.06481314e+00
3.46639067e-01 -3.63645971e-01 5.91223454e-03 5.11162281e-01
-5.32659352e-01 3.77483703e-02 6.35319889e-01 7.66655505e-01
-2.19884798e-01 -9.36216891e-01 -5.43098211e-01 1.81425422e-01
-5.96969306e-01 -1.46943033e-01 -2.83284187e-01 5.00630736e-01
3.10686648e-01 9.66455519e-01 7.02828392e-02 -4.02517229e-01
1.31903887e-01 -6.41225427e-02 6.81891918e-01 -6.21698022e-01
-2.50560105e-01 9.41941217e-02 -3.34694117e-01 -4.93913174e-01
-3.99843395e-01 -2.72054046e-01 -9.75479662e-01 -4.03541952e-01
7.00728223e-02 -5.70706204e-02 4.52381879e-01 5.32810032e-01
6.39813006e-01 5.45697153e-01 8.68820190e-01 -1.26668334e+00
-2.34796256e-01 -1.08171225e+00 -6.45660698e-01 2.86800563e-01
6.25618815e-01 -7.45101869e-01 -4.50998724e-01 3.68303478e-01] | [12.782869338989258, -0.09824017435312271] |
683abddb-5899-4207-bc49-f3ccfca66562 | type-prediction-with-program-decomposition | 2305.17145 | null | https://arxiv.org/abs/2305.17145v1 | https://arxiv.org/pdf/2305.17145v1.pdf | Type Prediction With Program Decomposition and Fill-in-the-Type Training | TypeScript and Python are two programming languages that support optional type annotations, which are useful but tedious to introduce and maintain. This has motivated automated type prediction: given an untyped program, produce a well-typed output program. Large language models (LLMs) are promising for type prediction, but there are challenges: fill-in-the-middle performs poorly, programs may not fit into the context window, generated types may not type check, and it is difficult to measure how well-typed the output program is. We address these challenges by building OpenTau, a search-based approach for type prediction that leverages large language models. We propose a new metric for type prediction quality, give a tree-based program decomposition that searches a space of generated types, and present fill-in-the-type fine-tuning for LLMs. We evaluate our work with a new dataset for TypeScript type prediction, and show that 47.4% of files type check (14.5% absolute improvement) with an overall rate of 3.3 type errors per file. All code, data, and models are available at: https://github.com/GammaTauAI/opentau. | ['Steven Holtzen', 'Arjun Guha', 'Noah Shinn', 'Ming-Ho Yee', 'Federico Cassano'] | 2023-05-25 | null | null | null | null | ['type-prediction'] | ['computer-code'] | [-2.20919952e-01 -3.95219438e-02 -6.24135792e-01 -4.26419973e-01
-7.68293917e-01 -8.59158158e-01 2.08768889e-01 5.46555042e-01
-1.59066468e-02 5.19920707e-01 -2.76832525e-02 -7.16727674e-01
3.98706526e-01 -1.08314478e+00 -1.00296497e+00 -1.28905997e-02
-5.59102707e-02 1.39118984e-01 4.16997552e-01 9.14161652e-02
5.34884155e-01 -2.35968769e-01 -1.78276956e+00 9.46700931e-01
1.21003568e+00 9.16660666e-01 1.94802806e-01 8.37354183e-01
-5.80901742e-01 9.03658092e-01 -5.45053124e-01 -4.44443554e-01
-5.55239245e-02 8.83975849e-02 -1.10980308e+00 -7.35528171e-01
5.18734574e-01 -1.20072708e-01 3.45477551e-01 9.64306116e-01
2.04839736e-01 -4.24917966e-01 3.87646914e-01 -1.61504519e+00
-4.55162913e-01 7.94727027e-01 -4.06305522e-01 -1.40736565e-01
6.58535302e-01 3.52682918e-01 1.24644291e+00 -9.18997169e-01
8.48513305e-01 1.16115165e+00 1.06964815e+00 5.68478942e-01
-1.41656625e+00 -3.94184202e-01 -2.30343938e-01 5.65738752e-02
-1.13889110e+00 -3.29667538e-01 1.48970261e-02 -1.09898245e+00
1.51692891e+00 7.88938701e-01 4.41804528e-01 6.02695584e-01
3.89419913e-01 7.35593438e-01 1.17912769e+00 -4.23719227e-01
2.62836367e-01 2.30459601e-01 5.19865751e-01 8.10329080e-01
1.53690398e-01 -9.84535366e-02 -3.96109998e-01 -8.57351243e-01
1.44690841e-01 -3.00714046e-01 3.93320397e-02 -1.92144692e-01
-1.05994582e+00 5.25394380e-01 8.52789953e-02 1.82315253e-03
1.78333372e-01 2.22398877e-01 7.02150822e-01 3.12197626e-01
3.27190608e-01 5.36932528e-01 -9.62037742e-01 -8.25938106e-01
-8.99085641e-01 5.26367843e-01 1.13763499e+00 1.45492387e+00
1.05926025e+00 -1.49440348e-01 -2.92519748e-01 1.07326293e+00
1.75449207e-01 3.71460438e-01 3.00126433e-01 -8.60290766e-01
4.04774725e-01 1.23481548e+00 -1.87146917e-01 -3.68631989e-01
-2.36115023e-01 5.01273870e-02 -3.66430551e-01 4.12094474e-01
6.48097873e-01 8.52077156e-02 -4.47521865e-01 1.28147972e+00
2.81599700e-01 -2.49948308e-01 -2.98464626e-01 3.75914335e-01
6.94884717e-01 4.77215976e-01 -4.18155678e-02 1.26959532e-01
1.33141899e+00 -8.62469256e-01 -1.45023778e-01 -2.17783481e-01
1.35871482e+00 -9.84296501e-01 1.42409909e+00 3.60331833e-01
-1.16210544e+00 -2.36484975e-01 -7.24093854e-01 -1.83240354e-01
-5.04186630e-01 6.49803653e-02 6.00076079e-01 8.43134165e-01
-1.17183864e+00 6.47216856e-01 -8.69025826e-01 -1.85529083e-01
3.02520096e-01 1.01541437e-01 -1.35874331e-01 2.09811285e-01
-4.71131355e-01 5.02695382e-01 3.93163919e-01 -3.45054567e-01
-4.57555920e-01 -1.25228786e+00 -5.64923286e-01 -5.27441576e-02
4.36394662e-01 -6.38409078e-01 1.52703846e+00 -7.14359462e-01
-7.03556001e-01 8.10644031e-01 -6.58022225e-01 -2.42539480e-01
4.49782580e-01 -1.15869403e-01 -1.36576533e-01 -6.37767851e-01
9.86999497e-02 -3.70860137e-02 3.56179088e-01 -1.12278569e+00
-9.06150639e-01 -3.13193768e-01 1.50969669e-01 -5.12575209e-01
-1.80368990e-01 4.00949299e-01 -3.23192865e-01 -6.67819440e-01
-2.28650153e-01 -9.42724109e-01 5.42158559e-02 -9.41778496e-02
-2.85807848e-01 -4.58361566e-01 5.42947054e-01 -8.77906919e-01
1.98337770e+00 -1.91937041e+00 -1.37364075e-01 8.45230073e-02
3.83945793e-01 1.10188335e-01 7.41982162e-02 2.58651197e-01
3.01821083e-02 8.86203647e-01 -2.86769152e-01 3.07301935e-02
3.52452934e-01 1.14637345e-01 -2.51708210e-01 -3.58778164e-02
-1.02636077e-01 7.43773341e-01 -7.92988777e-01 -3.93913388e-01
-3.65802914e-01 -1.26023486e-01 -1.38660240e+00 5.19252300e-01
-9.28729355e-01 -2.05340654e-01 -2.54438728e-01 1.00797689e+00
7.69697666e-01 -4.13171388e-02 1.21699765e-01 5.87048493e-02
-7.22044826e-01 7.95374036e-01 -1.17622375e+00 1.22173631e+00
-6.27121449e-01 4.00033295e-01 2.79854108e-02 -4.25545156e-01
8.49731266e-01 -3.71667817e-02 1.84702396e-01 -5.91993511e-01
-4.90303427e-01 6.29537940e-01 -1.23949431e-01 -7.77094007e-01
8.17966223e-01 5.33874273e-01 -4.74097610e-01 7.15323985e-01
-2.74249643e-01 -1.77202091e-01 4.60411608e-01 1.16091080e-01
1.42952311e+00 5.32580554e-01 1.84051111e-01 -3.32619071e-01
2.08331019e-01 2.36688823e-01 9.50701594e-01 1.09415913e+00
2.33892962e-01 3.26359659e-01 1.09451973e+00 -4.30437863e-01
-1.27127028e+00 -6.67861104e-01 -8.06930438e-02 1.47113383e+00
-4.24792081e-01 -1.16354978e+00 -7.72568405e-01 -6.90861046e-01
1.10991195e-01 7.87971079e-01 -3.00916016e-01 1.02545165e-01
-7.92269528e-01 -4.73864377e-01 8.69688451e-01 5.79893947e-01
1.06046036e-01 -8.83693218e-01 -4.07066494e-01 1.26089618e-01
-3.84923667e-01 -5.70597410e-01 -4.54725504e-01 9.37766675e-03
-7.58656859e-01 -1.22292650e+00 -1.28387600e-01 -6.49909735e-01
4.09770638e-01 -2.90656328e-01 1.51914287e+00 6.30507886e-01
-2.11522296e-01 6.82732835e-02 -4.98936713e-01 -2.06055358e-01
-9.92653430e-01 2.17925668e-01 -2.65108168e-01 -6.39175415e-01
2.54285187e-01 -3.09577346e-01 -1.60363242e-01 4.56436515e-01
-5.30770838e-01 8.70432258e-02 1.61462903e-01 1.01559150e+00
4.31522220e-01 -1.08006269e-01 -2.91306730e-02 -1.26375377e+00
2.55178064e-01 -6.04324341e-01 -9.36523438e-01 5.11504531e-01
-9.05525565e-01 3.10773373e-01 7.73960292e-01 -8.69065002e-02
-9.81799781e-01 -2.25336298e-01 -4.58448619e-01 -9.67728533e-03
-1.43909622e-02 8.14095199e-01 -1.41141713e-01 -4.45955992e-02
7.73722291e-01 2.40778357e-01 -4.73726727e-02 -7.58390844e-01
7.34240981e-03 7.90053010e-01 3.01014692e-01 -1.17102778e+00
5.97385883e-01 -2.78262049e-01 -3.74894708e-01 -5.77614605e-01
-2.20698535e-01 -4.25044745e-01 -6.27026975e-01 -7.14906603e-02
2.67192453e-01 -6.76153779e-01 -8.93799424e-01 7.83092797e-01
-1.09755540e+00 -9.33296204e-01 2.67249376e-01 -2.61580735e-01
-5.88364124e-01 5.21991014e-01 -6.04944229e-01 -7.83536255e-01
-4.53872740e-01 -1.37479424e+00 9.83463168e-01 -1.65682361e-01
-6.52722001e-01 -8.50590646e-01 -1.76033884e-01 2.90732920e-01
5.81586182e-01 -1.29632145e-01 1.76472688e+00 -4.52879876e-01
-8.33789945e-01 1.12245761e-01 -1.98446527e-01 8.60916823e-02
-2.72250295e-01 5.46901941e-01 -7.52965808e-01 -7.61441439e-02
-5.05968094e-01 -1.55603796e-01 4.75662470e-01 1.92568637e-02
1.62577391e+00 -9.38463509e-01 -4.12406445e-01 8.74699652e-01
1.43123198e+00 8.48848559e-03 7.87894607e-01 5.72515070e-01
8.60312939e-01 2.18308836e-01 7.94319212e-01 8.38523805e-01
7.44155049e-01 7.97924757e-01 1.64925501e-01 4.40720826e-01
3.94159518e-02 -3.80380094e-01 4.78984714e-01 7.32004046e-01
1.29221249e-02 3.21847759e-02 -1.38253760e+00 4.32743520e-01
-1.84260273e+00 -1.03702569e+00 -8.22646201e-01 2.53037071e+00
1.40742552e+00 2.17770636e-01 5.67238212e-01 1.32513210e-01
3.35551679e-01 -3.51734698e-01 -2.99889803e-01 -8.37636411e-01
1.47248134e-01 2.18129665e-01 6.62396610e-01 5.18847287e-01
-7.69121766e-01 8.26901078e-01 6.21568632e+00 8.29848647e-01
-1.08094168e+00 4.70931418e-02 3.40915591e-01 1.85171023e-01
-8.32355559e-01 6.51169538e-01 -1.14971042e+00 8.24570417e-01
1.02020144e+00 -5.17033875e-01 5.42306006e-01 1.19873703e+00
2.77211845e-01 -4.11268830e-01 -1.32531643e+00 5.86381614e-01
-4.05465752e-01 -1.39695179e+00 -2.55775958e-01 1.28301829e-01
3.60274196e-01 -1.13284960e-01 -2.45254457e-01 7.16381550e-01
2.56050229e-01 -7.65536845e-01 1.17797720e+00 5.65264285e-01
7.49594986e-01 -3.09150249e-01 3.34839106e-01 3.97327572e-01
-1.26752865e+00 -1.67021826e-01 -2.30186000e-01 -1.84768945e-01
-4.24477369e-01 7.82777488e-01 -1.11002636e+00 2.13726282e-01
1.04543638e+00 5.29793561e-01 -1.07234967e+00 1.30363667e+00
6.85914084e-02 7.70246565e-01 -2.64263272e-01 -3.80673647e-01
-5.39677441e-01 2.06565097e-01 4.58575189e-01 1.47092700e+00
4.31997120e-01 -2.33557582e-01 3.35339695e-01 1.26788568e+00
1.18407104e-02 8.76501650e-02 -2.62040585e-01 -7.78073445e-02
8.40978920e-01 9.75698352e-01 -9.08689871e-02 -5.50792933e-01
-5.07980406e-01 4.01966304e-01 3.36743593e-01 -2.26267446e-02
-7.57926285e-01 -3.89439940e-01 7.83226073e-01 4.36713755e-01
6.91965520e-02 -1.50745571e-01 -8.64946067e-01 -1.35122120e+00
6.18999124e-01 -1.29672718e+00 5.40164292e-01 -6.68027699e-01
-9.24371541e-01 2.09350318e-01 2.19197832e-02 -1.05988348e+00
-3.93218815e-01 -5.52815139e-01 -4.60192770e-01 9.27235007e-01
-1.06162405e+00 -1.03758705e+00 -3.12855780e-01 -1.41328983e-02
3.63478750e-01 7.97765777e-02 9.22385514e-01 4.98409212e-01
-6.03661239e-01 1.34873927e+00 1.98377192e-01 4.49491553e-02
5.61680555e-01 -1.47301948e+00 7.98248410e-01 8.17672968e-01
-7.22981632e-01 1.02745616e+00 7.29276001e-01 -8.42829704e-01
-2.01153636e+00 -1.19819462e+00 1.08922291e+00 -6.21331394e-01
9.50555682e-01 -6.69418037e-01 -1.25107241e+00 6.64882183e-01
-1.99818477e-01 -1.62930042e-01 5.30613363e-01 4.74993169e-01
-9.59073901e-01 9.89019498e-02 -1.01120377e+00 4.00728554e-01
9.83782351e-01 -5.43964863e-01 -2.88540930e-01 1.85261533e-01
4.17080522e-01 -5.37966728e-01 -1.40808225e+00 1.44402727e-01
1.11701298e+00 -1.16357982e+00 6.14337206e-01 -2.98224926e-01
6.82788014e-01 -1.87044889e-01 -1.90969855e-01 -1.25060284e+00
-1.32340953e-01 -6.19113266e-01 -1.03919275e-01 1.65163982e+00
8.51437211e-01 -8.15722525e-01 6.40375316e-01 1.10588229e+00
-3.97058725e-01 -7.62285650e-01 -3.50832343e-01 -9.96667385e-01
3.57626587e-01 -5.79171777e-01 8.28081310e-01 9.06001568e-01
5.47234952e-01 -8.33572820e-02 4.82839486e-03 -2.90787388e-02
4.14295286e-01 2.69739121e-01 1.14078081e+00 -1.01293516e+00
-6.59457028e-01 -7.63889015e-01 -1.83719784e-01 -7.40236700e-01
9.28219780e-02 -1.24842501e+00 -3.49513553e-02 -1.25964880e+00
3.02845031e-01 -1.28443086e+00 4.44286644e-01 1.16820467e+00
-1.96682796e-01 -2.67087698e-01 5.25940172e-02 2.87511617e-01
-2.32276157e-01 -1.30379990e-01 4.03048009e-01 -3.80263746e-01
-4.53742445e-02 1.62244290e-01 -4.19135451e-01 7.31559932e-01
8.40234160e-01 -5.59750319e-01 3.20222408e-01 -4.51680481e-01
7.57965028e-01 3.24003041e-01 3.11340153e-01 -8.45556438e-01
-6.63989112e-02 -4.55038577e-01 -1.61482975e-01 -7.24101782e-01
-2.30179597e-02 -2.28738368e-01 4.37656164e-01 4.56180543e-01
-4.93731022e-01 2.51340389e-01 3.25019479e-01 -1.42742798e-01
2.27918401e-01 -6.87267840e-01 6.05822980e-01 -1.90591201e-01
-1.02944219e+00 1.93524152e-01 -3.17706764e-01 3.56318295e-01
5.88723063e-01 -1.71483636e-01 -8.52250695e-01 4.15552348e-01
-4.01423842e-01 1.95231989e-01 1.13981962e+00 5.74601293e-01
1.62897766e-01 -9.91867781e-01 -4.36137974e-01 3.40899229e-01
5.57676494e-01 -2.73806036e-01 -1.00482985e-01 6.71312094e-01
-5.89677036e-01 1.01265505e-01 5.65445125e-02 -6.19389892e-01
-1.66923535e+00 4.55294251e-01 1.59215912e-01 -2.74346117e-02
-2.21268311e-01 8.06856334e-01 -1.49680257e-01 -9.05767381e-01
-7.90044293e-02 -6.50126815e-01 3.67441356e-01 -1.29330773e-02
4.99664664e-01 5.11415362e-01 4.22145545e-01 -3.05823743e-01
-4.27959055e-01 1.66076660e-01 -2.43368912e-02 2.35869706e-01
1.30343044e+00 4.02608603e-01 -9.09896255e-01 6.05888426e-01
1.33820844e+00 2.16784492e-01 -7.53401518e-01 -1.24910407e-01
3.83738130e-01 -8.20623279e-01 -3.52836967e-01 -9.96548235e-01
-5.69046736e-01 6.50227427e-01 3.06980729e-01 5.59891880e-01
7.53423631e-01 -9.46876854e-02 5.79377234e-01 2.48923630e-01
7.91725099e-01 -9.54715669e-01 -3.21892202e-01 9.26843166e-01
6.75902903e-01 -9.85379040e-01 -5.92263602e-02 -6.21665061e-01
-1.57448024e-01 9.98574972e-01 8.90831053e-01 4.11139756e-01
4.67595041e-01 8.98606479e-01 -3.32662344e-01 1.66294247e-01
-1.25337958e+00 1.72030225e-01 -3.94453444e-02 4.83975321e-01
9.72399771e-01 3.76895607e-01 -3.98424178e-01 6.61050439e-01
-5.81784844e-01 2.50203520e-01 9.10270631e-01 1.39279318e+00
-3.38635921e-01 -1.58891523e+00 -6.61029756e-01 1.05612075e+00
-5.14570951e-01 -4.28073853e-01 -1.85666010e-01 3.79521370e-01
-7.50164762e-02 8.40149879e-01 -9.48885605e-02 -5.39643943e-01
3.13060254e-01 2.98697442e-01 5.26388943e-01 -8.32125962e-01
-8.97937953e-01 -4.63698171e-02 5.82743645e-01 -5.84400654e-01
3.21290880e-01 -8.20560932e-01 -1.10440993e+00 -7.75017500e-01
-3.24626148e-01 -5.47943339e-02 5.63878417e-01 6.06473625e-01
7.77264774e-01 8.74318108e-02 5.02269745e-01 -3.76600921e-01
-5.26698351e-01 -7.59670556e-01 -1.17265902e-01 2.18022436e-01
2.52665997e-01 -4.84704792e-01 -2.64625490e-01 3.81226927e-01] | [7.831954479217529, 7.7145795822143555] |
1e0340a9-02e7-483f-916b-03615b0f1bda | chatabl-abductive-learning-via-natural | 2304.11107 | null | https://arxiv.org/abs/2304.11107v1 | https://arxiv.org/pdf/2304.11107v1.pdf | ChatABL: Abductive Learning via Natural Language Interaction with ChatGPT | Large language models (LLMs) such as ChatGPT have recently demonstrated significant potential in mathematical abilities, providing valuable reasoning paradigm consistent with human natural language. However, LLMs currently have difficulty in bridging perception, language understanding and reasoning capabilities due to incompatibility of the underlying information flow among them, making it challenging to accomplish tasks autonomously. On the other hand, abductive learning (ABL) frameworks for integrating the two abilities of perception and reasoning has seen significant success in inverse decipherment of incomplete facts, but it is limited by the lack of semantic understanding of logical reasoning rules and the dependence on complicated domain knowledge representation. This paper presents a novel method (ChatABL) for integrating LLMs into the ABL framework, aiming at unifying the three abilities in a more user-friendly and understandable manner. The proposed method uses the strengths of LLMs' understanding and logical reasoning to correct the incomplete logical facts for optimizing the performance of perceptual module, by summarizing and reorganizing reasoning rules represented in natural language format. Similarly, perceptual module provides necessary reasoning examples for LLMs in natural language format. The variable-length handwritten equation deciphering task, an abstract expression of the Mayan calendar decoding, is used as a testbed to demonstrate that ChatABL has reasoning ability beyond most existing state-of-the-art methods, which has been well supported by comparative studies. To our best knowledge, the proposed ChatABL is the first attempt to explore a new pattern for further approaching human-level cognitive ability via natural language interaction with ChatGPT. | ['Tuo Zhang', 'Tianming Liu', 'Dinggang Shen', 'Junwei Han', 'Xi Jiang', 'Dajiang Zhu', 'Xiang Li', 'Chong Ma', 'Junjie Yao', 'Wenjun Li', 'Xiaozheng Wei', 'Zhengliang Liu', 'Zihao Wu', 'Li Yang', 'Yaonai Wei', 'Tianyang Zhong'] | 2023-04-21 | null | null | null | null | ['decipherment', 'logical-reasoning'] | ['natural-language-processing', 'reasoning'] | [-2.35800356e-01 4.33033168e-01 3.36776786e-02 -1.38918087e-01
-8.11424032e-02 -5.36616862e-01 4.15483028e-01 3.34139287e-01
-1.98437378e-01 6.16476357e-01 1.03907734e-02 -6.64496958e-01
-7.02009201e-01 -1.06041646e+00 -5.68541050e-01 2.35159392e-03
5.47807179e-02 6.16256475e-01 2.58460402e-01 -6.29133165e-01
5.97408295e-01 4.24043208e-01 -1.73354816e+00 6.67439103e-01
1.35609519e+00 7.84402251e-01 6.43384039e-01 4.51542646e-01
-7.83410609e-01 1.69567621e+00 -5.27768493e-01 -4.68820751e-01
-7.24754557e-02 -2.83829331e-01 -1.25380671e+00 -4.19037580e-01
7.42216855e-02 -4.53887254e-01 -2.77950875e-02 9.91624355e-01
3.07248998e-02 2.52921823e-02 5.72665095e-01 -1.23918283e+00
-9.89579201e-01 8.26528966e-01 3.95854190e-02 4.78145201e-03
1.10110795e+00 -7.13997856e-02 7.06754923e-01 -5.96668124e-01
2.22929984e-01 1.63326693e+00 5.73789895e-01 3.69381070e-01
-7.08759069e-01 -5.94478309e-01 1.48158520e-01 7.47677624e-01
-1.34361351e+00 1.86857238e-01 6.92608476e-01 -4.76916224e-01
1.44059932e+00 1.95935860e-01 8.56872141e-01 4.71563339e-01
4.31446880e-01 6.43752933e-01 1.51418972e+00 -7.88904786e-01
2.51940668e-01 5.28507888e-01 4.65548962e-01 9.36253667e-01
2.39014894e-01 -1.58688232e-01 -7.70117581e-01 2.84486026e-01
9.10509348e-01 -3.08813721e-01 1.34893849e-01 -2.91344702e-01
-1.06758618e+00 4.69313979e-01 3.15248668e-01 6.59284770e-01
-3.31785589e-01 -1.10896818e-01 1.63897261e-01 6.03934824e-01
-4.77922052e-01 7.95382917e-01 -3.21549863e-01 -2.92151690e-01
-5.43221354e-01 2.74075389e-01 1.04145801e+00 8.48827004e-01
5.10296404e-01 7.85217285e-02 -4.93110381e-02 6.43406868e-01
6.37439132e-01 5.97896576e-01 6.04825258e-01 -1.10719752e+00
6.05354965e-01 1.24891198e+00 1.27682790e-01 -1.16482294e+00
-5.08255601e-01 -1.34668425e-01 -4.75617707e-01 4.90795583e-01
5.22072375e-01 1.53202355e-01 -4.18720216e-01 1.71858788e+00
7.37053901e-02 -2.49378011e-01 6.68671608e-01 7.51473129e-01
1.06198883e+00 6.89975441e-01 3.61128688e-01 -2.07308959e-02
1.71946740e+00 -7.07932472e-01 -8.85501146e-01 -2.78247893e-01
5.27715087e-01 -6.16611958e-01 1.33720291e+00 7.91138589e-01
-1.34264636e+00 -8.53961170e-01 -1.20432258e+00 -4.30714607e-01
-7.71724999e-01 1.74426600e-01 9.08435404e-01 6.34277403e-01
-1.02642179e+00 1.59298331e-01 -4.82787639e-01 -3.65629315e-01
1.35519981e-01 3.97088259e-01 -2.41824493e-01 -2.01458126e-01
-1.54180050e+00 1.37418699e+00 8.93569827e-01 2.21430302e-01
-6.03012204e-01 -6.53419912e-01 -9.03815091e-01 3.31429988e-01
6.04479373e-01 -7.69014180e-01 1.29581535e+00 -6.23014212e-01
-1.60009181e+00 6.23054445e-01 -3.14979553e-02 -4.69539583e-01
4.78199065e-01 -3.91294688e-01 -4.31303322e-01 1.37543812e-01
6.34245872e-02 5.63316226e-01 2.84277290e-01 -1.14248371e+00
-3.12883556e-01 -2.76802272e-01 6.82707429e-01 5.28316379e-01
-2.84647763e-01 -1.62531257e-01 3.15834694e-02 -4.24133748e-01
3.92373383e-01 -3.55887920e-01 3.18037033e-01 -1.21623456e-01
2.06426363e-02 -6.37790143e-01 4.32664573e-01 -6.90221131e-01
1.30916262e+00 -1.77946424e+00 2.86724955e-01 7.06173927e-02
1.68092042e-01 4.57444996e-01 1.34095326e-01 7.61465847e-01
1.52943268e-01 1.86284855e-02 1.56218871e-01 5.02341092e-02
4.43932116e-01 5.07827461e-01 -4.67739433e-01 -4.15037274e-01
-4.91953343e-02 9.06166732e-01 -8.38389814e-01 -6.76990271e-01
3.61968488e-01 -4.52520996e-02 -4.49830085e-01 2.69786507e-01
-4.03382778e-01 6.83746710e-02 -4.88911122e-01 5.45600355e-01
6.02617800e-01 -1.12554454e-03 3.79220724e-01 1.65442638e-02
-1.97364464e-01 2.54962564e-01 -1.50014603e+00 1.70371497e+00
-6.88871324e-01 2.94382870e-01 -1.71812341e-01 -1.09301674e+00
1.06326270e+00 4.09709871e-01 -2.24720582e-01 -9.50533271e-01
2.73067262e-02 1.73143178e-01 3.17155838e-01 -1.01177084e+00
2.23182902e-01 -3.21077824e-01 -3.47591303e-02 4.24106032e-01
-5.91741689e-02 -6.02325916e-01 2.80272841e-01 2.79067576e-01
5.55895984e-01 4.69709128e-01 7.69322097e-01 -1.29280552e-01
9.85991061e-01 1.76287711e-01 8.14472511e-02 9.79836822e-01
-9.85222459e-02 -2.50009716e-01 4.72797453e-01 -4.54070807e-01
-6.44478738e-01 -1.20971155e+00 2.31340632e-01 9.01243925e-01
2.34923035e-01 -4.68158007e-01 -6.57218218e-01 -9.10392106e-02
-3.03075820e-01 1.07939148e+00 -1.16928816e-01 -8.81409720e-02
-4.19629514e-01 -7.41362870e-02 8.83654773e-01 5.02297342e-01
1.24134815e+00 -1.40187573e+00 -9.66764033e-01 1.38382062e-01
-2.19655231e-01 -1.32006395e+00 6.67636931e-01 -1.35528177e-01
-8.18044841e-01 -1.11925995e+00 -4.09301221e-02 -9.72523272e-01
5.10777593e-01 1.05603360e-01 9.54349577e-01 5.03119111e-01
1.49565106e-02 5.43706715e-01 -3.85443032e-01 -6.93033636e-01
-4.92078841e-01 -3.60442102e-01 7.02372193e-03 -5.95005035e-01
4.61759001e-01 -5.83748281e-01 -9.22646187e-03 1.60225034e-02
-7.40943611e-01 6.72975481e-01 8.17526281e-01 4.64244485e-01
5.47848009e-02 6.90058351e-01 5.79942524e-01 -4.03351486e-01
1.07928681e+00 -1.74025759e-01 -4.27513957e-01 5.68108380e-01
-5.74335277e-01 2.30625272e-01 8.40728879e-01 -1.77808270e-01
-1.61661851e+00 -4.91704553e-01 2.39963294e-03 1.47396088e-01
-4.04212743e-01 7.20916331e-01 -2.01066598e-01 -5.56602739e-02
4.61639076e-01 6.54643834e-01 8.14964548e-02 -1.60426080e-01
3.06765079e-01 5.13409078e-01 6.33024693e-01 -1.26540422e+00
7.52090454e-01 -3.25309262e-02 1.58592075e-01 -4.95816946e-01
-7.50598311e-01 -7.16798529e-02 -6.90957129e-01 -1.33786246e-01
8.76080334e-01 -8.59532535e-01 -1.40878785e+00 4.88196313e-01
-1.26909709e+00 -3.12711746e-01 -5.21968864e-02 6.28782511e-01
-7.06626952e-01 6.15723073e-01 -6.36391342e-01 -1.04876614e+00
-6.94945306e-02 -1.05536568e+00 4.21416759e-01 3.96215200e-01
-4.55634445e-01 -9.30805564e-01 -3.87670398e-01 9.21478868e-01
2.96837509e-01 -1.98759407e-01 1.41008449e+00 -6.00166202e-01
-5.78479886e-01 -4.48170528e-02 -3.27527195e-01 3.69709462e-01
-1.02925017e-01 -3.80198866e-01 -6.20686233e-01 2.57609874e-01
2.60843933e-01 -6.84042990e-01 1.95624545e-01 -1.44847110e-01
9.16757584e-01 -2.20701903e-01 7.19866306e-02 -4.33182940e-02
1.30219686e+00 5.59348404e-01 8.25019598e-01 5.86368084e-01
2.69501448e-01 7.31987119e-01 8.16844404e-01 2.00970173e-01
9.83828723e-01 4.17838603e-01 1.01028465e-01 2.56379902e-01
-1.79580122e-01 -4.15903717e-01 2.71823585e-01 6.90486431e-01
-3.87178600e-01 2.78596073e-01 -1.31809652e+00 1.07743330e-01
-1.89037192e+00 -1.12396514e+00 1.10844094e-02 1.80659592e+00
9.73175466e-01 2.99558908e-01 -2.94164777e-01 4.21563268e-01
1.67511627e-01 -3.58783901e-01 -2.74127543e-01 -8.79837155e-01
7.43392333e-02 1.36542782e-01 -3.95762056e-01 7.74635315e-01
-4.29449141e-01 1.15223157e+00 6.00264645e+00 8.18138361e-01
-6.60009861e-01 -1.74788237e-01 8.68901797e-03 4.31799144e-01
-1.36678129e-01 1.78966746e-02 -6.25433207e-01 7.15441406e-02
6.74391747e-01 -8.50231051e-02 7.32115567e-01 5.48571944e-01
2.78968036e-01 -5.07675767e-01 -1.14634073e+00 1.02315450e+00
1.48723647e-01 -1.26186299e+00 5.63752711e-01 -4.62461382e-01
3.01400453e-01 -9.63422835e-01 -1.40001699e-01 7.86457837e-01
1.66048080e-01 -1.35176659e+00 9.49547052e-01 8.01546156e-01
1.26804322e-01 -5.57822526e-01 9.66541350e-01 9.14953411e-01
-1.13424301e+00 -5.30573189e-01 -2.95274377e-01 -1.17649126e+00
-1.50615647e-01 -7.02512115e-02 -8.48196447e-01 7.88919628e-01
6.40698552e-01 2.00505033e-01 -7.08347201e-01 6.79431379e-01
-6.31715059e-01 8.36882964e-02 -9.07683223e-02 -3.92543882e-01
-6.62348345e-02 -1.66511640e-01 2.16256931e-01 7.87201107e-01
2.83486784e-01 5.31077802e-01 1.71126321e-01 1.17149615e+00
5.17394781e-01 2.62849241e-01 -5.04300475e-01 -1.12549506e-01
7.55040169e-01 6.82351887e-01 -6.70660853e-01 -6.52542233e-01
-5.76231420e-01 4.66872841e-01 5.50793290e-01 2.54110605e-01
-7.48085856e-01 -2.36546293e-01 1.13417119e-01 -4.16129082e-02
-1.78680107e-01 -5.05834043e-01 -7.47779608e-01 -9.61462736e-01
1.14898607e-01 -1.07315922e+00 3.47044557e-01 -1.47797096e+00
-9.35966551e-01 2.41880015e-01 5.66933215e-01 -8.58663380e-01
-1.94676951e-01 -9.56853569e-01 -5.69883466e-01 9.81586218e-01
-1.33805907e+00 -1.36751711e+00 -5.79340816e-01 7.52706230e-01
5.85203886e-01 -4.40885514e-01 1.02373648e+00 -1.89900547e-01
-2.94428915e-01 2.17381671e-01 -6.92093492e-01 -4.66109179e-02
1.86054572e-01 -1.26543331e+00 -3.64917099e-01 7.18297660e-01
-1.94850564e-01 1.03493130e+00 6.38111591e-01 -5.99887252e-01
-1.57697952e+00 -1.99515387e-01 9.71596181e-01 -2.89107114e-01
7.01112926e-01 -2.77269423e-01 -9.82150435e-01 6.69221103e-01
2.95451462e-01 -7.35497236e-01 6.72721744e-01 -1.11084180e-02
-3.48687530e-01 -1.32303461e-01 -1.15716159e+00 9.13474143e-01
9.08511221e-01 -5.14998794e-01 -1.75388622e+00 1.83977485e-01
6.33015633e-01 -3.66493434e-01 -7.77118921e-01 1.98954269e-01
5.66677332e-01 -1.17989039e+00 1.16963673e+00 -4.36257243e-01
5.12020528e-01 -6.01260900e-01 -1.80779263e-01 -9.44265783e-01
-2.12318808e-01 -2.77003646e-01 -6.30240962e-02 1.25557089e+00
1.00026079e-01 -7.37890005e-01 3.73676687e-01 7.51775861e-01
-7.27175176e-02 -5.75289071e-01 -7.86319554e-01 -6.51533365e-01
1.43111289e-01 -7.98962593e-01 6.56136274e-01 7.76263416e-01
6.58409536e-01 3.55102539e-01 -4.42268476e-02 4.17947084e-01
3.01647902e-01 2.44979709e-01 6.64193392e-01 -1.46763873e+00
-4.13062125e-01 -5.45755744e-01 -3.27147514e-01 -9.32608366e-01
3.60410035e-01 -8.89824748e-01 -2.82220334e-01 -1.94009340e+00
4.56740186e-02 -3.41232717e-01 1.36782318e-01 6.30007148e-01
1.20485224e-01 -2.46387318e-01 4.74127650e-01 6.41190261e-02
-5.19144833e-01 2.25667670e-01 1.56207132e+00 -5.93719296e-02
-2.12122917e-01 -3.95054132e-01 -8.68432581e-01 1.12866294e+00
1.04284632e+00 1.11836791e-01 -1.00682628e+00 -4.60267097e-01
5.12348711e-01 3.59381944e-01 5.34094036e-01 -1.36808014e+00
6.92801774e-01 -5.36710501e-01 3.75320017e-01 -4.99121010e-01
3.71149331e-01 -9.68724608e-01 -1.92459896e-02 6.50974095e-01
-2.44092420e-01 2.28435472e-01 4.80175912e-01 6.63895607e-02
-2.77886927e-01 -5.18714190e-01 3.66818070e-01 -5.24942338e-01
-1.36299205e+00 -6.71520472e-01 -7.29299009e-01 -1.45512804e-01
1.15144324e+00 -5.42306125e-01 -5.21852136e-01 -2.45970905e-01
-7.96027899e-01 2.64495254e-01 -3.21236737e-02 3.18376452e-01
8.13748419e-01 -9.66255546e-01 -1.68520331e-01 2.09067330e-01
-9.61693078e-02 -5.68409823e-02 3.88614535e-01 7.79247761e-01
-9.11248088e-01 8.28596354e-01 -7.74175644e-01 -8.73310640e-02
-1.22680020e+00 6.82428181e-01 4.97972488e-01 -4.23014849e-01
-5.34517288e-01 5.51177680e-01 4.84318333e-03 -5.91029584e-01
3.49505156e-01 -7.09273100e-01 -6.09798074e-01 -1.45419240e-01
5.85604608e-01 3.57982308e-01 -1.92593187e-01 -1.84045866e-01
-3.60785425e-01 7.79250801e-01 1.86621889e-01 -4.19950560e-02
7.96404779e-01 -2.87083536e-01 -5.22113800e-01 5.42326689e-01
1.41526654e-01 6.59944192e-02 -4.83270139e-01 -8.14927965e-02
9.05699953e-02 -1.58193752e-01 -3.84289205e-01 -1.24676347e+00
-7.85645097e-02 8.83625686e-01 2.44469658e-01 2.73586065e-01
1.02275765e+00 -2.20057163e-02 3.32856923e-01 1.05158556e+00
6.71609879e-01 -1.02432740e+00 4.01767552e-01 7.35893905e-01
1.03166962e+00 -1.05598986e+00 1.18681692e-01 -6.38369322e-01
-5.24511158e-01 1.50175607e+00 1.11020386e+00 1.78259566e-01
1.42724618e-01 3.39627504e-01 7.12943077e-02 -2.48284012e-01
-7.19562113e-01 -5.31367995e-02 1.14028811e-01 8.33815098e-01
4.00605857e-01 7.06726983e-02 -5.03282428e-01 9.57759678e-01
-7.66674280e-01 4.64436322e-01 5.13431251e-01 1.10074008e+00
-6.80613697e-01 -9.80514407e-01 -8.71803403e-01 -1.17337778e-01
-4.44029309e-02 -1.90683499e-01 -2.50138253e-01 1.11238623e+00
4.21610743e-01 1.10299659e+00 -1.60757467e-01 -1.19902967e-02
4.20094669e-01 3.87041777e-01 8.02679002e-01 -4.98514563e-01
-3.43135625e-01 -6.60134792e-01 7.36292079e-02 -3.23049605e-01
-3.96665633e-01 -2.73732990e-01 -1.81943393e+00 -3.77815038e-01
4.42116022e-01 2.19182938e-01 2.28071377e-01 1.53052127e+00
-9.14974138e-02 6.35809839e-01 -3.44056189e-01 -3.86356533e-01
-4.81988132e-01 -8.32795918e-01 -2.52878100e-01 2.53556430e-01
-8.17084238e-02 -8.63527894e-01 6.49653524e-02 -1.45980805e-01] | [9.546823501586914, 7.352589130401611] |
5a2386b6-a06e-4e38-ba2c-bed7ea2ce6d2 | an-automated-approach-for-the-recognition-of | 2109.00906 | null | https://arxiv.org/abs/2109.00906v1 | https://arxiv.org/pdf/2109.00906v1.pdf | An Automated Approach for the Recognition of Bengali License Plates | Automatic Number Plate Recognition (ALPR) is a system for automatically identifying the license plates of any vehicle. This process is important for tracking, ticketing, and any billing system, among other things. With the use of information and communication technology (ICT), all systems are being automated, including the vehicle tracking system. This study proposes a hybrid method for detecting license plates using characters from them. Our captured image information was used for the recognition procedure in Bangladeshi vehicles, which is the topic of this study. Here, for license plate detection, the YOLO model was used where 81% was correctly predicted. And then, for license plate segmentation, Otsu's Thresholding was used and eventually, for character recognition, the CNN model was applied. This model will allow the vehicle's automated license plate detection system to avoid any misuse. | ['Faisal Muhammad Shah', 'Jannatun Naeem Muna', 'Atiqul Islam Chowdhury', 'MD Abdullah Al Nasim'] | 2021-09-01 | null | null | null | null | ['license-plate-detection'] | ['computer-vision'] | [-4.55945842e-02 -6.05624974e-01 -1.86036631e-01 -1.03999712e-01
-3.41741383e-01 -8.81800711e-01 2.53302962e-01 -1.77223936e-01
-3.78140271e-01 5.72963715e-01 -4.01632577e-01 -5.60920775e-01
5.06277204e-01 -8.66000891e-01 -3.74650747e-01 -6.26104057e-01
6.19372189e-01 3.90143335e-01 5.80805779e-01 -1.54728144e-01
1.01351428e+00 9.41760063e-01 -1.18179297e+00 1.23691127e-01
4.51331556e-01 7.07115412e-01 1.12828799e-01 7.92615950e-01
-3.06354582e-01 8.32735479e-01 -4.11638707e-01 -4.28722352e-01
5.23229480e-01 -1.64515048e-01 -2.45839477e-01 3.84323508e-01
-5.75252883e-02 -6.99187994e-01 -3.55987072e-01 1.00443685e+00
1.95949990e-02 1.32134467e-01 1.03236830e+00 -1.04642904e+00
-4.65625018e-01 -1.38737127e-01 -5.89742720e-01 8.72543976e-02
3.04791182e-02 7.08150789e-02 1.77143246e-01 -1.06795168e+00
2.84627467e-01 8.19776237e-01 7.97474384e-01 4.06700671e-01
-3.00805390e-01 -1.03690124e+00 -8.79210353e-01 3.01156312e-01
-1.63412428e+00 -4.01925862e-01 6.45331621e-01 -7.38852322e-01
6.67410254e-01 5.21953218e-02 4.82773006e-01 -1.71329960e-01
4.45583016e-01 7.13409126e-01 1.07035637e+00 -6.89304471e-01
3.14138792e-02 6.62838340e-01 6.10680282e-01 5.97532332e-01
5.02568185e-01 -2.86772639e-01 2.91098446e-01 2.87504643e-01
7.78002918e-01 2.71786004e-01 6.02648914e-01 5.64857721e-01
-4.43135053e-01 7.72101521e-01 -3.48917723e-01 4.96737033e-01
-2.32873350e-01 -2.62990594e-01 8.06749915e-04 -1.54205739e-01
1.24140710e-01 -6.25346452e-02 1.78487934e-02 -1.05771340e-01
-1.21023750e+00 -1.41700521e-01 4.93281752e-01 8.82083118e-01
7.78562486e-01 2.91855663e-01 1.22123137e-01 9.08516943e-01
7.41801322e-01 7.69606292e-01 4.41216528e-01 -4.42264140e-01
4.84656990e-01 9.27482545e-01 1.67952418e-01 -1.12024999e+00
-1.90738082e-01 2.56269991e-01 -4.12740558e-01 2.98856944e-01
3.38827133e-01 -3.47557873e-01 -1.34702373e+00 2.58176923e-01
-1.71553746e-01 2.71339059e-01 5.31204529e-02 6.36243284e-01
7.48851240e-01 1.23132551e+00 9.57539752e-02 2.16811910e-01
1.54372478e+00 -7.32910931e-01 -6.86643600e-01 -1.46121442e-01
5.44185221e-01 -1.11307716e+00 3.07695687e-01 1.63362220e-01
-7.33278215e-01 -4.94133621e-01 -1.11651933e+00 3.20491582e-01
-7.54024208e-01 6.58844888e-01 2.20905110e-01 1.05924785e+00
-7.80585229e-01 -2.39061952e-01 -4.72570956e-01 -4.25611258e-01
3.21300060e-01 7.33319223e-01 -2.96817213e-01 -2.98808485e-01
-7.77993619e-01 1.13218677e+00 1.64618254e-01 2.51524687e-01
-2.34005928e-01 3.43728900e-01 -6.40884161e-01 -1.89342767e-01
-6.40088394e-02 3.50860208e-01 8.90171170e-01 -1.09037471e+00
-1.41217220e+00 8.90041173e-01 -2.14572966e-01 -1.53904349e-01
1.36037290e-01 3.30349356e-01 -7.42495716e-01 1.52585149e-01
7.07194507e-02 4.43548679e-01 7.22459257e-01 -8.47291112e-01
-8.62641990e-01 -4.72740471e-01 -4.45255220e-01 2.98774511e-01
1.68889582e-01 6.18893862e-01 -6.37798727e-01 -1.14488699e-01
8.24444070e-02 -1.10026288e+00 2.56539732e-01 -7.74648488e-01
-2.84509957e-01 -4.42612588e-01 1.11306465e+00 -1.26532984e+00
9.46658850e-01 -2.45072031e+00 -1.15690494e+00 8.31981122e-01
-8.53700042e-02 8.16886485e-01 3.78174454e-01 2.07372919e-01
2.12635458e-01 3.30148667e-01 -2.02906787e-01 2.04235017e-01
-1.91036597e-01 -2.61211812e-01 1.72765836e-01 5.77977538e-01
3.64808261e-01 5.83767533e-01 2.25492179e-01 -6.55277014e-01
5.78076839e-01 2.60120511e-01 1.53992236e-01 -8.73381495e-02
4.46008056e-01 -9.55620781e-02 -4.63045359e-01 1.03810322e+00
1.08870673e+00 3.59104037e-01 -1.55932933e-01 2.03941435e-01
-6.45936728e-01 -3.98093551e-01 -1.05931365e+00 8.43841806e-02
1.19007058e-01 1.17834520e+00 -1.39374375e-01 -7.87010670e-01
1.45500088e+00 4.18766767e-01 3.66172820e-01 -7.23766744e-01
4.43718702e-01 5.68414211e-01 -3.82362269e-02 -8.07379305e-01
6.59237683e-01 1.31904185e-01 1.10318214e-01 -1.42118648e-01
-5.59132993e-01 2.32536301e-01 2.68552423e-01 -2.41187871e-01
3.91535610e-01 -1.52277693e-01 -4.64165257e-03 9.69770923e-02
8.40058625e-01 6.80406630e-01 3.67048979e-01 3.84774804e-01
-4.53178287e-01 3.49021703e-01 -2.34655645e-02 -4.10604209e-01
-1.21440256e+00 -4.26450312e-01 -1.96926653e-01 5.46189427e-01
2.07952127e-01 7.42356241e-01 -9.24392104e-01 -1.41204327e-01
-4.57100496e-02 6.06110036e-01 8.41731112e-03 3.46444279e-01
-6.95524871e-01 -4.95828152e-01 7.60551572e-01 3.46661031e-01
9.80846226e-01 -1.16294086e+00 -6.73209131e-02 2.72390097e-01
2.58310676e-01 -9.61436927e-01 -4.77057517e-01 -3.17566305e-01
-7.07795501e-01 -1.08791709e+00 -8.63547087e-01 -1.23871696e+00
8.39317977e-01 6.90635383e-01 2.30094448e-01 2.86171198e-01
1.64126568e-02 6.91742450e-02 -1.22488379e-01 -5.63388288e-01
-6.58418357e-01 -2.47352704e-01 -3.00334960e-01 1.34776399e-01
9.92696464e-01 3.00804347e-01 -3.91184747e-01 4.24652219e-01
-7.76724815e-01 -1.42464146e-01 8.60788822e-01 -3.07096597e-02
6.06114790e-02 4.31573868e-01 4.98734325e-01 -7.88365424e-01
6.20185018e-01 -3.69666964e-01 -1.05428779e+00 2.58669168e-01
-4.11396891e-01 -6.74714446e-01 3.37499171e-01 9.46516991e-02
-1.02317679e+00 3.43194872e-01 -2.45740518e-01 6.38166908e-03
-7.00549662e-01 1.94441631e-01 -2.21075043e-01 -4.34241623e-01
-2.23638475e-01 6.53950691e-01 2.09774911e-01 -1.42568916e-01
-4.06214029e-01 1.43167722e+00 3.13884705e-01 3.23427737e-01
7.13870585e-01 3.81187797e-02 -6.71789497e-02 -1.36057842e+00
2.86004484e-01 -1.00658596e+00 -5.39707541e-01 -9.16303158e-01
1.27169037e+00 -6.88621700e-01 -9.06841993e-01 1.12528789e+00
-1.15757883e+00 3.70222837e-01 7.33773649e-01 8.59184921e-01
1.26480043e-01 2.38894299e-01 -5.89743316e-01 -1.38287890e+00
-2.45743357e-02 -1.15511751e+00 5.22204936e-01 8.04455757e-01
2.25591511e-01 -7.85748243e-01 8.47770795e-02 8.67198110e-01
3.44244242e-01 -1.93258539e-01 5.33520758e-01 -1.05883741e+00
-7.54127026e-01 -9.01618123e-01 -5.26602566e-01 6.06255412e-01
-3.18152383e-02 5.91951489e-01 -6.76427662e-01 3.85638773e-01
-1.48918644e-01 3.39255512e-01 5.82016706e-01 5.26580572e-01
4.81978744e-01 -4.37883019e-01 -1.16601333e-01 8.68485123e-02
1.59400237e+00 1.28974891e+00 1.29647493e+00 6.97562337e-01
6.66598141e-01 4.01836544e-01 6.07439220e-01 8.75767246e-02
4.70110089e-01 3.31576169e-01 -2.29631700e-02 -1.32457092e-01
1.79880708e-01 3.52605395e-02 5.85488498e-01 8.30861032e-01
-3.34585994e-01 -1.65747955e-01 -1.25030267e+00 2.11804390e-01
-1.19887185e+00 -1.07688534e+00 -6.96856678e-01 1.96586132e+00
2.13904396e-01 1.26604224e-02 2.27543473e-01 3.16023052e-01
1.34396076e+00 -5.79916000e-01 -9.37350839e-02 -8.26596379e-01
1.88301399e-01 -6.11191690e-02 1.19636345e+00 3.53374720e-01
-1.11390173e+00 1.08963001e+00 5.78051043e+00 8.21918845e-01
-1.38185346e+00 -2.71126956e-01 7.66548038e-01 8.40241194e-01
3.09455007e-01 -2.74125040e-01 -1.25980675e+00 7.54236579e-01
8.57740641e-01 2.61856109e-01 3.35637599e-01 6.49730504e-01
5.26867330e-01 -4.98336792e-01 -7.46599510e-02 7.57233620e-01
2.77928770e-01 -1.26332998e+00 -1.79952204e-01 2.04412431e-01
4.90378499e-01 -3.43559057e-01 -8.38936567e-02 2.41696134e-01
1.93689279e-02 -8.04661810e-01 5.24407446e-01 5.90184987e-01
4.93923694e-01 -1.06662154e+00 1.29465187e+00 3.77791524e-01
-1.00140750e+00 6.53969198e-02 -3.94350916e-01 2.29473729e-02
-2.12001756e-01 -5.65787964e-02 -1.32618487e+00 -1.60648346e-01
5.42564541e-02 3.16727877e-01 -5.53759217e-01 1.41285861e+00
4.20513034e-01 9.57725048e-01 -3.52333724e-01 -4.74912554e-01
4.71147239e-01 -6.35213673e-01 1.63885102e-01 1.47513843e+00
5.92244983e-01 2.32269287e-01 -1.51915029e-01 4.29024190e-01
-1.56233951e-01 2.93678224e-01 -6.97966218e-01 -1.21760428e-01
6.51885688e-01 1.04421878e+00 -1.31231070e+00 -3.43575984e-01
-6.02089524e-01 4.95127439e-01 -6.95168674e-01 2.09475294e-01
-9.01970208e-01 -7.47110188e-01 -6.53629825e-02 4.28831935e-01
1.76582932e-01 -4.60937858e-01 -5.57549357e-01 -5.95800817e-01
-2.98377186e-01 -4.92185086e-01 -3.40053393e-03 -6.59742236e-01
-6.93872690e-01 -6.57168403e-02 -9.63831320e-02 -1.32113421e+00
6.70758337e-02 -1.04221439e+00 -8.48224461e-01 7.82970607e-01
-1.31601787e+00 -1.09782040e+00 5.80346249e-02 6.05466843e-01
7.89543450e-01 -7.30917871e-01 1.92881092e-01 5.52034318e-01
-8.65987778e-01 2.50737995e-01 6.94916666e-01 9.83070850e-01
3.14162195e-01 -6.32072806e-01 3.29153165e-02 1.01935053e+00
-2.46390522e-01 2.37863600e-01 1.83110952e-01 -9.48270500e-01
-1.35305834e+00 -1.05641901e+00 9.43791807e-01 -9.00867954e-02
1.99593276e-01 -4.82409224e-02 -5.51347136e-01 6.15957379e-01
1.81863859e-01 -5.76324821e-01 7.03065932e-01 -8.12527657e-01
2.76937753e-01 -1.74822271e-01 -1.52167821e+00 3.11756134e-01
-4.27930832e-01 -1.67612031e-01 -3.24900359e-01 2.36266628e-02
-2.00834751e-01 -3.43816690e-02 -2.45922908e-01 -2.75290340e-01
7.60334313e-01 -5.47829032e-01 5.48737347e-01 -2.22268421e-02
1.49907887e-01 -5.92030644e-01 -7.73591176e-02 -3.69015306e-01
-3.36865246e-01 2.09582776e-01 6.61863267e-01 1.51029277e+00
5.93164086e-01 -7.33921289e-01 9.23181474e-01 1.17077589e+00
-1.11714445e-01 9.68676656e-02 -5.79695582e-01 -4.60567325e-01
-3.00741196e-01 -3.72911245e-01 3.18225473e-01 7.11676717e-01
-3.31714243e-01 -2.08624184e-01 -2.92318553e-01 3.84369433e-01
4.98246789e-01 -6.58310175e-01 4.77531731e-01 -8.31605613e-01
3.96277726e-01 -4.18407589e-01 -8.44702184e-01 -6.14798725e-01
-1.83484644e-01 -4.46465701e-01 2.02099845e-01 -1.43224156e+00
1.89654261e-01 -2.37802878e-01 8.84753466e-02 3.10095817e-01
2.43163973e-01 8.22798550e-01 4.98269111e-01 7.79717445e-01
-2.68145144e-01 -3.77162546e-01 8.38623047e-01 -2.41572157e-01
-2.60307610e-01 5.48870921e-01 -4.03144032e-01 8.69803786e-01
1.32003355e+00 -5.49475074e-01 1.61133811e-01 -1.39780179e-01
-4.03097644e-02 1.33019805e-01 1.69895157e-01 -9.27749634e-01
6.03559971e-01 -3.00686717e-01 7.50736535e-01 -1.05788636e+00
7.05390945e-02 -1.03086603e+00 3.97457927e-02 3.33937436e-01
1.76784828e-01 1.24369480e-01 3.66692364e-01 2.79391240e-02
-2.15915963e-01 -7.98990726e-01 8.42461288e-01 -4.41990048e-02
-1.04906213e+00 -7.59800598e-02 -1.46001506e+00 -6.34079993e-01
1.52291608e+00 -1.12805235e+00 -1.16844125e-01 -1.93557248e-01
-1.86425015e-01 -4.71765064e-02 3.60083789e-01 -6.28098473e-02
8.05454373e-01 -1.22334552e+00 -7.20204234e-01 2.87191510e-01
-1.84293404e-01 -3.89983952e-01 1.54722318e-01 4.76853758e-01
-1.56836367e+00 7.57931650e-01 -5.42663038e-01 -2.69028604e-01
-1.43417287e+00 -7.75624216e-02 1.69856101e-01 1.90975368e-01
-1.63844246e-02 2.86805987e-01 -4.19014901e-01 -8.95563886e-02
-2.08433673e-01 1.33138850e-01 -9.38267589e-01 -1.08079083e-01
4.32113200e-01 7.70027220e-01 6.77569062e-02 -1.10157454e+00
-4.60672826e-01 9.69884872e-01 -4.70470963e-03 -2.26514667e-01
1.02040017e+00 -1.05202466e-01 -1.92106411e-01 2.15193853e-01
8.93212199e-01 1.78645089e-01 -8.86166573e-01 4.30862784e-01
1.76429972e-01 -4.08579051e-01 -2.08166298e-02 -6.56258643e-01
-8.52333665e-01 7.25379348e-01 6.01267159e-01 2.47540697e-01
7.60638475e-01 -5.56883216e-01 6.97860479e-01 4.68886048e-01
1.85986206e-01 -1.47907948e+00 -6.73017561e-01 5.88018179e-01
7.55219683e-02 -1.28520834e+00 -3.26862991e-01 -1.83870703e-01
-9.35649216e-01 1.61192727e+00 4.27640229e-01 -2.32051119e-01
5.78726172e-01 5.04073679e-01 5.82267456e-02 1.69314891e-01
-2.38337461e-02 -4.89862189e-02 2.52211511e-01 6.77027583e-01
4.74525660e-01 2.20851749e-01 -3.34873319e-01 4.45517510e-01
2.26011217e-01 2.50445753e-01 8.82059038e-01 8.97344589e-01
-1.08157277e+00 -6.90213501e-01 -1.03283525e+00 7.28323519e-01
-7.44490266e-01 -6.03372790e-02 -5.73512197e-01 9.72385168e-01
3.51232022e-01 1.21218264e+00 2.31010169e-01 -4.05730635e-01
-2.02215146e-02 2.20411107e-01 1.25260269e-02 -2.42687643e-01
-3.11416507e-01 -7.62039497e-02 4.48836274e-02 4.74902868e-01
-2.85262436e-01 -7.34608591e-01 -1.41551542e+00 -8.35235953e-01
-5.70450723e-01 1.64543837e-01 1.32213938e+00 1.07588255e+00
-1.98924914e-01 6.57933056e-02 8.54129732e-01 -4.19891357e-01
-2.10324232e-03 -7.30222404e-01 -8.19286168e-01 -6.20321594e-02
-1.43854290e-01 -2.37269133e-01 8.67101888e-04 4.10496354e-01] | [9.809768676757812, -4.980074882507324] |
fd76b4d2-9444-41a3-bf59-18b078510ede | gaitpart-temporal-part-based-model-for-gait | null | null | http://openaccess.thecvf.com/content_CVPR_2020/html/Fan_GaitPart_Temporal_Part-Based_Model_for_Gait_Recognition_CVPR_2020_paper.html | http://openaccess.thecvf.com/content_CVPR_2020/papers/Fan_GaitPart_Temporal_Part-Based_Model_for_Gait_Recognition_CVPR_2020_paper.pdf | GaitPart: Temporal Part-Based Model for Gait Recognition | Gait recognition, applied to identify individual walking patterns in a long-distance, is one of the most promising video-based biometric technologies. At present, most gait recognition methods take the whole human body as a unit to establish the spatio-temporal representations. However, we have observed that different parts of human body possess evidently various visual appearances and movement patterns during walking. In the latest literature, employing partial features for human body description has been verified being beneficial to individual recognition. Taken above insights together, we assume that each part of human body needs its own spatio-temporal expression. Then, we propose a novel part-based model GaitPart and get two aspects effect of boosting the performance: On the one hand, Focal Convolution Layer, a new applying of convolution, is presented to enhance the fine-grained learning of the part-level spatial features. On the other hand, the Micro-motion Capture Module (MCM) is proposed and there are several parallel MCMs in the GaitPart corresponding to the pre-defined parts of the human body, respectively. It is worth mentioning that the MCM is a novel way of temporal modeling for gait task, which focuses on the short-range temporal features rather than the redundant long-range features for cycle gait. Experiments on two of the most popular public datasets, CASIA-B and OU-MVLP, richly exemplified that our method meets a new state-of-the-art on multiple standard benchmarks. The source code will be available on https://github.com/ChaoFan96/GaitPart.
| [' Zhiqiang He', ' Qing Li', ' Yongzhen Huang', ' Jiannan Chi', ' Saihui Hou', ' Xu Liu', ' Chunshui Cao', ' Yunjie Peng', 'Chao Fan'] | 2020-06-01 | null | null | null | cvpr-2020-6 | ['multiview-gait-recognition'] | ['computer-vision'] | [-6.52671531e-02 -6.54774904e-01 -3.47118467e-01 -1.39856204e-01
-2.00651303e-01 1.41000105e-02 4.84123290e-01 -3.71479511e-01
-3.24149460e-01 5.38112879e-01 1.73203826e-01 3.44256520e-01
-3.11001074e-02 -8.41687560e-01 -3.81269872e-01 -1.05309093e+00
-3.97060812e-01 1.50825351e-01 3.76718521e-01 -1.80299118e-01
1.79426983e-01 4.27697420e-01 -1.77574050e+00 2.39280909e-01
4.49717790e-01 1.07405269e+00 -8.92804191e-02 4.69640106e-01
2.21631303e-01 5.05993187e-01 -3.56561661e-01 -3.24786574e-01
-3.56769040e-02 -3.74569774e-01 -6.11620247e-01 1.69544071e-01
2.18534097e-01 -2.47767568e-01 -6.53508425e-01 8.01794827e-01
7.20611632e-01 5.93637899e-02 7.44556367e-01 -1.21050143e+00
-4.72986966e-01 2.28430167e-01 -9.51320827e-01 3.14326376e-01
4.73520160e-01 5.47778070e-01 6.19138300e-01 -8.32901835e-01
4.37016606e-01 1.16206098e+00 8.22014570e-01 5.03410697e-01
-7.40419388e-01 -6.15409613e-01 -1.66491047e-02 7.30882347e-01
-1.65070391e+00 -3.09190273e-01 9.39719260e-01 -4.11775261e-01
7.39205658e-01 1.62873581e-01 9.83705938e-01 1.22451317e+00
5.82612038e-01 9.17875290e-01 8.96644175e-01 -2.13466570e-01
-8.98385793e-02 -6.64530337e-01 2.15437934e-01 8.12803805e-01
2.71628171e-01 1.44775122e-01 -6.13359213e-01 8.13019350e-02
8.86460185e-01 1.50564164e-01 -3.74302357e-01 -3.81327450e-01
-1.25582004e+00 2.74050117e-01 3.12916279e-01 5.89878500e-01
-3.42306167e-01 2.89674580e-01 5.78627408e-01 -1.06188864e-01
9.20611992e-02 -5.86510122e-01 2.66287755e-03 -4.13842052e-01
-1.00549316e+00 1.86739475e-01 5.47571123e-01 6.16692066e-01
7.09697485e-01 -1.39801279e-02 -3.27443451e-01 7.53776431e-01
3.68600100e-01 6.29301250e-01 7.99215376e-01 -5.52123368e-01
4.37781513e-01 6.46841526e-01 1.20303093e-03 -1.16251445e+00
-4.47487652e-01 -4.33451533e-01 -1.36878324e+00 5.93718216e-02
5.16457200e-01 1.83777764e-01 -9.09055114e-01 1.57221520e+00
2.83326983e-01 6.85624599e-01 -3.06790709e-01 1.04543126e+00
7.46105909e-01 3.84240985e-01 2.37177193e-01 -1.16497234e-01
1.67624342e+00 -8.22754085e-01 -6.70553446e-01 -6.81600496e-02
3.69325966e-01 -5.12285173e-01 8.47927630e-01 2.22416237e-01
-8.09896111e-01 -9.06196654e-01 -1.15925384e+00 1.98518261e-01
-5.06498992e-01 4.94647324e-01 5.31515062e-01 7.67955959e-01
-7.46656001e-01 7.15608418e-01 -1.07657051e+00 -6.23383880e-01
2.96861559e-01 3.69676501e-01 -5.78168511e-01 -2.66098715e-02
-1.29542363e+00 5.78159094e-01 2.59458661e-01 6.19161367e-01
-5.92426300e-01 -1.41988471e-01 -9.00495410e-01 -4.19498533e-02
5.99822812e-02 -7.92362034e-01 7.28295088e-01 -6.87093556e-01
-1.32878304e+00 1.03914762e+00 -1.82830334e-01 -3.12307537e-01
6.74522698e-01 -1.36582494e-01 -6.96289659e-01 1.41247153e-01
6.73313290e-02 3.08901072e-01 9.63877082e-01 -7.79889047e-01
-5.45114875e-01 -4.64772433e-01 -3.83396655e-01 4.84965295e-02
-3.61014456e-01 -5.82664162e-02 -8.85991693e-01 -8.45636070e-01
-2.86010187e-02 -7.92144179e-01 4.89245392e-02 -3.20228897e-02
-2.14891016e-01 -2.69026428e-01 8.17875147e-01 -8.49880874e-01
1.63048756e+00 -2.13744903e+00 5.53190969e-02 2.09831640e-01
3.15620899e-02 4.64216501e-01 7.45902508e-02 5.03103971e-01
-1.35063872e-01 -1.41556799e-01 -4.07485008e-01 -3.00626546e-01
-1.45137310e-02 1.58799127e-01 1.46794572e-01 8.04746509e-01
6.79427385e-02 9.45954621e-01 -6.71923280e-01 -7.96927869e-01
4.62175757e-01 5.02523780e-01 -1.58993408e-01 -7.07225353e-02
5.05715787e-01 4.46935594e-01 -4.10925269e-01 9.27205682e-01
8.77132595e-01 -2.32690245e-01 1.42851293e-01 -4.41346467e-01
1.20739860e-03 -3.56507361e-01 -1.40462220e+00 1.87785220e+00
4.68144789e-02 3.91908251e-02 -2.42091432e-01 -1.22746539e+00
7.82124162e-01 3.92089784e-01 9.41941559e-01 -6.82096183e-01
1.34825870e-01 2.48152241e-01 1.33249508e-02 -6.79192305e-01
1.74497381e-01 1.66973084e-01 -1.18953116e-01 2.34190404e-01
1.19814053e-01 6.97506785e-01 3.98230702e-01 -4.77132171e-01
1.05383682e+00 4.98031318e-01 5.79640210e-01 -9.28616002e-02
8.78497839e-01 -2.65192449e-01 7.32979536e-01 4.07212555e-01
-5.91762900e-01 7.68597186e-01 -2.54828501e-02 -6.35534525e-01
-6.99037015e-01 -1.19154060e+00 -9.95852798e-03 5.63558877e-01
3.34205955e-01 -3.25773180e-01 -6.26073122e-01 -4.15659010e-01
-2.42779776e-02 -2.52578974e-01 -7.22227097e-01 -2.57153213e-01
-1.00909257e+00 -9.74825323e-01 9.19829726e-01 8.33211839e-01
1.18769336e+00 -1.19713914e+00 -9.24580395e-01 3.52382034e-01
-3.96952182e-01 -9.87763882e-01 -6.64892435e-01 -1.91013873e-01
-7.49817193e-01 -1.13806188e+00 -1.13930237e+00 -8.22014749e-01
2.77450264e-01 1.12504601e-01 8.60162020e-01 3.26377809e-01
-5.42163193e-01 2.77833194e-01 -3.34339797e-01 2.14973688e-01
2.76418209e-01 -3.94750647e-02 1.55651525e-01 4.07879323e-01
6.39023364e-01 -8.79312396e-01 -8.74503493e-01 5.27150273e-01
-5.64985454e-01 5.21214455e-02 6.47329926e-01 9.54869866e-01
4.47648346e-01 6.12406954e-02 1.59771755e-01 -1.50255725e-01
2.31027812e-01 -4.33513850e-01 -2.72842981e-02 3.32232267e-01
-8.34813341e-02 -2.72027791e-01 4.01831776e-01 -5.65608442e-01
-8.63774836e-01 7.22502917e-02 -2.02387005e-01 -4.54197347e-01
-4.54094410e-01 4.64958727e-01 -4.14702028e-01 1.48005188e-02
2.84232169e-01 7.22817302e-01 1.01724669e-01 -5.39370358e-01
1.63103156e-02 5.36384940e-01 6.67231023e-01 -7.56321669e-01
8.50289583e-01 6.48091614e-01 2.37737641e-01 -9.13791656e-01
-6.79267272e-02 -7.24370122e-01 -9.34341133e-01 -4.42614675e-01
9.24517930e-01 -8.09836149e-01 -8.98059368e-01 1.10719919e+00
-8.57538640e-01 -1.88810050e-01 -2.77107768e-02 4.17830169e-01
-7.63731480e-01 8.54324698e-01 -7.69935727e-01 -7.57814825e-01
-3.45757753e-01 -8.96875918e-01 1.21215832e+00 3.13424408e-01
-1.69751197e-01 -7.31259167e-01 1.63286820e-01 1.35171056e-01
8.95645320e-02 4.28309172e-01 5.76553702e-01 -1.15986109e-01
-3.00083548e-01 -2.54870445e-01 -1.38853401e-01 9.93886739e-02
4.10176128e-01 7.02744722e-02 -8.43895316e-01 -3.59315246e-01
-2.00668842e-01 -4.32123197e-03 8.46325755e-01 6.39208794e-01
9.61771548e-01 -4.33224551e-02 -5.85629880e-01 6.32372081e-01
1.38962710e+00 3.01273793e-01 1.06114769e+00 2.98433065e-01
6.86406374e-01 4.87955213e-01 8.08852494e-01 6.20211065e-01
3.15396577e-01 1.02971900e+00 1.20355919e-01 -4.56911847e-02
-3.09759289e-01 -1.55824959e-01 5.41015923e-01 7.75914967e-01
-7.27068305e-01 -6.35844320e-02 -9.46911097e-01 5.28476834e-01
-2.08720469e+00 -1.37228358e+00 -5.71354963e-02 2.23009801e+00
3.79655838e-01 -3.24840434e-02 5.33477426e-01 4.83914167e-01
8.71334791e-01 3.11397523e-01 -2.52058268e-01 1.03138499e-01
-2.04393968e-01 2.24621490e-01 3.51071447e-01 -4.11743857e-02
-1.49709773e+00 6.09296858e-01 5.45313787e+00 1.03429317e+00
-1.14708936e+00 -2.50348840e-02 4.32001263e-01 1.17456906e-01
5.22949696e-01 -3.30227613e-01 -7.77548075e-01 8.47955108e-01
6.94870472e-01 -1.99520309e-02 5.12859225e-02 5.34280658e-01
4.29125428e-01 1.51821575e-03 -8.48841250e-01 1.25132871e+00
3.87263596e-02 -1.09734046e+00 7.35065565e-02 2.18828380e-01
2.23875150e-01 -4.75785255e-01 -7.53916148e-03 1.14175379e-01
-2.71414936e-01 -1.00919366e+00 6.35985255e-01 8.27660501e-01
8.97910237e-01 -7.42977202e-01 8.98112178e-01 2.59162903e-01
-2.20628929e+00 -4.73940931e-02 -2.02758789e-01 -1.55013263e-01
2.53346950e-01 3.28302383e-01 -5.74454367e-02 1.05516255e+00
9.67061937e-01 1.18022358e+00 -7.19157517e-01 1.29524744e+00
3.19883302e-02 5.39902568e-01 -2.23315299e-01 1.64438456e-01
6.83612674e-02 -1.59873679e-01 3.25235337e-01 1.36278379e+00
5.27730465e-01 7.17868432e-02 2.31757641e-01 5.63324273e-01
3.94737959e-01 9.86643583e-02 -5.41912317e-01 9.89238471e-02
9.99384597e-02 1.02758574e+00 -6.78971589e-01 -2.65560031e-01
-5.62426090e-01 1.09358728e+00 1.17997900e-02 2.85768807e-01
-1.03306282e+00 -2.57487655e-01 7.27557123e-01 1.98264033e-01
3.48156154e-01 -3.14777732e-01 -8.12005103e-02 -1.36732805e+00
3.13342929e-01 -7.55911350e-01 5.72895050e-01 -4.22198802e-01
-1.39119816e+00 1.27790645e-01 3.18753421e-02 -1.81288326e+00
-2.61877418e-01 -6.98819816e-01 -7.89025545e-01 9.34954584e-01
-1.07823217e+00 -1.54753995e+00 -5.90242565e-01 9.01164889e-01
5.26985526e-01 4.11732793e-02 7.44853497e-01 7.80644476e-01
-7.52548695e-01 6.71328664e-01 -1.99398130e-01 6.05539739e-01
6.12088263e-01 -8.09665442e-01 4.59417582e-01 1.00026751e+00
3.97750698e-02 6.73246086e-01 5.29916942e-01 -7.74683714e-01
-1.46354926e+00 -8.27255368e-01 7.48978794e-01 -1.12231970e-01
4.24690187e-01 -4.21689637e-02 -8.13714921e-01 4.88221228e-01
-3.18696536e-02 2.14364052e-01 5.41244566e-01 -1.90135613e-01
-6.63932115e-02 -1.70755118e-01 -1.03474844e+00 5.22807956e-01
1.41814995e+00 -2.92148381e-01 -4.87127632e-01 -2.48213455e-01
-8.12795013e-02 -1.91473380e-01 -9.36809599e-01 6.65027797e-01
1.07070708e+00 -1.02160013e+00 1.24020374e+00 -3.05248439e-01
3.20356071e-01 -7.76819646e-01 -1.88182622e-01 -7.55680799e-01
-5.52559257e-01 -1.62563071e-01 -5.30154765e-01 1.20176578e+00
-4.53164876e-01 -4.76543397e-01 1.08486998e+00 5.97691387e-02
1.30048608e-02 -8.49896491e-01 -1.12416685e+00 -1.07895970e+00
-1.60492897e-01 -3.21797818e-01 6.03337824e-01 6.71046972e-01
7.13564381e-02 -7.51718208e-02 -7.61229336e-01 7.77612850e-02
8.40275109e-01 1.82213351e-01 8.33212018e-01 -1.12749350e+00
-2.95438737e-01 -5.43008864e-01 -1.02799678e+00 -1.08897054e+00
-1.00643434e-01 -5.39784193e-01 -1.68349147e-01 -1.37159956e+00
2.41660923e-01 -8.55796039e-02 -5.64756215e-01 4.17884290e-01
-1.65596128e-01 4.90675449e-01 4.58609760e-02 3.21506768e-01
-3.21021974e-01 4.88637686e-01 1.26593339e+00 -4.77054626e-01
6.57333061e-02 1.24843068e-01 -1.59773231e-02 6.52557015e-01
7.31757402e-01 -8.28577951e-03 -2.07573414e-01 -5.15699200e-02
-4.76925462e-01 -2.62720115e-03 6.77018762e-01 -1.64336395e+00
3.60543340e-01 -1.72273204e-01 6.11549020e-01 -8.29313815e-01
6.00737691e-01 -7.79982269e-01 6.26557052e-01 9.36308682e-01
3.57448161e-01 6.44422770e-02 5.21336757e-02 5.63096702e-01
-4.37460363e-01 1.57383129e-01 7.46440172e-01 -2.47523949e-01
-1.31913686e+00 6.26934767e-01 -4.77516979e-01 -2.10647866e-01
9.12368119e-01 -7.86624610e-01 -1.13062397e-01 -1.43949032e-01
-8.19119930e-01 -1.86096542e-02 3.84811461e-01 4.45652574e-01
7.09238231e-01 -1.61385477e+00 -5.16714633e-01 2.55981386e-01
2.43806466e-01 -6.13607168e-01 7.93141961e-01 1.11553586e+00
-4.62499112e-01 4.07558888e-01 -7.49576688e-01 -8.19603384e-01
-1.44491935e+00 5.16792834e-01 5.17117143e-01 -5.24959147e-01
-7.88916528e-01 3.83148015e-01 1.06633566e-01 2.26987638e-02
1.15741000e-01 -2.24524587e-01 -4.95155722e-01 1.08338431e-01
5.13863385e-01 5.69479704e-01 -8.17644596e-02 -1.20083714e+00
-7.00892627e-01 1.22507346e+00 3.64762098e-01 7.61796385e-02
1.02288604e+00 -2.26658508e-01 3.97940837e-02 4.21518296e-01
1.00563562e+00 -3.78986448e-01 -1.18208265e+00 -7.60613009e-02
6.28335914e-03 -4.79754657e-01 -6.19162798e-01 -4.26278591e-01
-1.13231885e+00 1.09031999e+00 1.00374043e+00 -2.79720008e-01
1.31454444e+00 -5.12006819e-01 8.83397162e-01 4.26497757e-02
9.20523584e-01 -9.18578327e-01 5.81275187e-02 2.97450542e-01
6.70790672e-01 -9.34071302e-01 -3.50975357e-02 -4.94371325e-01
-3.49689692e-01 1.17076969e+00 7.36021817e-01 -1.58691883e-01
7.47347772e-01 1.24770783e-01 -2.46156588e-01 -1.18501484e-01
-3.01582009e-01 -3.09195369e-01 3.88330430e-01 7.94984877e-01
4.79467243e-01 1.78273037e-01 -5.85575938e-01 7.15676963e-01
1.10903569e-01 3.84335071e-01 -4.29299697e-02 1.01008463e+00
-2.98390567e-01 -1.11433530e+00 -4.48378563e-01 3.33307236e-01
-2.92632997e-01 1.78500310e-01 6.00360185e-02 9.34880853e-01
5.24032712e-01 6.70844138e-01 -1.01634808e-01 -6.41601443e-01
2.76998729e-01 2.09753692e-01 6.17176533e-01 -2.58564442e-01
-4.16650832e-01 -4.72909398e-02 -3.55267227e-02 -8.46917868e-01
-6.80687666e-01 -7.30369210e-01 -1.13641107e+00 -5.39905787e-01
1.72205746e-01 -8.84678885e-02 1.50322914e-02 8.65703642e-01
1.61244065e-01 3.93332332e-01 2.08456412e-01 -1.14977753e+00
-2.36573711e-01 -9.46799099e-01 -9.74534333e-01 6.71517491e-01
8.88826400e-02 -1.13652062e+00 -3.81100476e-02 3.31845701e-01] | [14.289898872375488, 1.4305458068847656] |
b490243d-80e8-4a20-b2ba-786d97910495 | cross-stitched-multi-task-dual-recursive | 2211.08290 | null | https://arxiv.org/abs/2211.08290v1 | https://arxiv.org/pdf/2211.08290v1.pdf | Cross-Stitched Multi-task Dual Recursive Networks for Unified Single Image Deraining and Desnowing | We present the Cross-stitched Multi-task Unified Dual Recursive Network (CMUDRN) model targeting the task of unified deraining and desnowing in a multi-task learning setting. This unified model borrows from the basic Dual Recursive Network (DRN) architecture developed by Cai et al. The proposed model makes use of cross-stitch units that enable multi-task learning across two separate DRN models, each tasked for single image deraining and desnowing, respectively. By fixing cross-stitch units at several layers of basic task-specific DRN networks, we perform multi-task learning over the two separate DRN models. To enable blind image restoration, on top of these structures we employ a simple neural fusion scheme which merges the output of each DRN. The separate task-specific DRN models and the fusion scheme are simultaneously trained by enforcing local and global supervision. Local supervision is applied on the two DRN submodules, and global supervision is applied on the data fusion submodule of the proposed model. Consequently, we both enable feature sharing across task-specific DRN models and control the image restoration behavior of the DRN submodules. An ablation study shows the strength of the hypothesized CMUDRN model, and experiments indicate that its performance is comparable or better than baseline DRN models on the single image deraining and desnowing tasks. Moreover, CMUDRN enables blind image restoration for the two underlying image restoration tasks, by unifying task-specific image restoration pipelines via a naive parametric fusion scheme. The CMUDRN implementation is available at https://github.com/VCL3D/CMUDRN. | ['Dimitrios Zarpalas', 'Konstantinos Konstantoudakis', 'Alexandros Doumanoglou', 'Sotiris Karavarsamis'] | 2022-11-15 | null | null | null | null | ['single-image-deraining'] | ['computer-vision'] | [ 3.87522399e-01 4.58928570e-02 5.86755797e-02 -2.57712044e-02
-9.99686301e-01 -4.90431130e-01 7.91310191e-01 -4.13390934e-01
-4.61152881e-01 3.81210893e-01 3.49485070e-01 -4.75686133e-01
-2.17772216e-01 -3.29729170e-01 -7.89034665e-01 -1.01164556e+00
4.33066159e-01 -4.57232073e-02 -2.36510709e-02 -4.33007367e-02
5.33649065e-02 5.50391853e-01 -1.58319545e+00 4.78597164e-01
9.88480568e-01 8.95560801e-01 4.16461706e-01 9.47044969e-01
2.98374921e-01 9.49535251e-01 -6.06483817e-01 -3.77824754e-01
6.07586324e-01 -1.21495863e-02 -5.46219647e-01 1.94186613e-01
1.05804944e+00 -5.67946672e-01 -5.97017586e-01 9.71988976e-01
7.84832656e-01 2.33263806e-01 4.25930619e-01 -1.29024947e+00
-1.22040772e+00 4.99325722e-01 -8.06953371e-01 1.24602534e-01
2.39666048e-02 2.09794119e-01 1.04179013e+00 -9.15463686e-01
8.60805586e-02 1.42505717e+00 4.72873062e-01 7.97244370e-01
-1.69213867e+00 -8.06791663e-01 3.34338248e-01 1.18079018e-02
-1.04989660e+00 -7.78082967e-01 4.38683301e-01 -4.21155095e-01
9.18730199e-01 -7.33703375e-02 1.11956544e-01 1.28835428e+00
2.25806281e-01 1.08438575e+00 1.16611433e+00 -1.89520657e-01
-5.36313355e-02 -1.86965600e-01 1.99496284e-01 3.92056108e-01
1.95762500e-01 4.07981306e-01 -4.93306696e-01 8.83499533e-03
1.02030861e+00 2.94856697e-01 -7.03593433e-01 -3.82058620e-01
-1.12367308e+00 7.16339350e-01 6.40116811e-01 3.21695894e-01
-1.72094956e-01 2.57714242e-01 4.11873043e-01 4.15071905e-01
5.46384156e-01 5.41173890e-02 -5.55089593e-01 6.65870488e-01
-1.13718605e+00 8.71643610e-03 2.52378613e-01 5.69803655e-01
8.80712271e-01 2.84752369e-01 -6.25078619e-01 1.00959444e+00
3.73698115e-01 4.09376293e-01 6.36919558e-01 -1.03505230e+00
3.49344909e-01 2.39268020e-01 -1.20276578e-01 -5.27454197e-01
-2.18431130e-01 -5.69882214e-01 -1.06784236e+00 8.12313914e-01
1.41165689e-01 -2.97495693e-01 -1.56398189e+00 1.93824673e+00
-2.56602801e-02 4.63994056e-01 3.47065836e-01 8.60443473e-01
8.22602868e-01 4.29041624e-01 8.75251368e-02 4.33827527e-02
1.44977987e+00 -1.30468082e+00 -6.63479030e-01 -4.14235324e-01
5.23650408e-01 -8.57459366e-01 9.65869904e-01 4.90739882e-01
-1.02989411e+00 -9.49925721e-01 -1.04607832e+00 -7.05195546e-01
-2.88109124e-01 5.96975327e-01 4.27979112e-01 4.80727792e-01
-1.66374183e+00 2.08434910e-01 -6.61731660e-01 -1.41965419e-01
6.13862455e-01 5.10033309e-01 -4.88989532e-01 -4.07075465e-01
-8.84977877e-01 9.66535389e-01 2.28772372e-01 5.77425241e-01
-1.46990585e+00 -6.86564267e-01 -1.04417777e+00 -1.26791790e-01
2.56104946e-01 -1.20905864e+00 1.23260224e+00 -9.73461986e-01
-1.25361121e+00 8.97756875e-01 -2.24919528e-01 -5.80172181e-01
4.01542664e-01 -5.22133291e-01 -3.47893536e-01 6.33932203e-02
1.03947327e-01 7.67856956e-01 1.52248704e+00 -1.70421827e+00
-5.95276952e-01 -2.79583335e-01 9.19418782e-02 3.12340796e-01
-1.01657115e-01 -9.32335481e-02 -3.48979503e-01 -1.00281394e+00
-1.21775970e-01 -5.63745022e-01 -1.66118383e-01 6.84800968e-02
-4.04374838e-01 2.63184961e-03 1.11233854e+00 -9.52804506e-01
9.38597739e-01 -2.49568462e+00 4.66121316e-01 -2.58513570e-01
5.17958105e-01 4.94039059e-01 -7.38199055e-01 -1.05345003e-01
-5.48033178e-01 -1.26997344e-02 -5.24108708e-01 -1.04681301e+00
-1.52492613e-01 3.69388551e-01 -6.26380503e-01 6.03555739e-01
4.10410725e-02 8.46995950e-01 -4.94178236e-01 1.53792530e-01
4.80297774e-01 6.90544426e-01 -3.69869918e-01 3.64670128e-01
-9.12010446e-02 3.30022722e-01 1.52586559e-02 5.41412294e-01
9.68705416e-01 -1.05654605e-01 -3.80569138e-02 -5.87940097e-01
2.07478516e-02 9.80671048e-02 -1.12895608e+00 1.85794747e+00
-7.02175915e-01 4.63171870e-01 5.07126749e-01 -9.73731399e-01
6.20281518e-01 6.35391831e-01 8.29501376e-02 -6.73474073e-01
-3.60916033e-02 9.81089771e-02 -1.29545495e-01 -1.52094796e-01
3.28398019e-01 -1.40196726e-01 1.19954534e-02 6.30247831e-01
5.52963436e-01 7.20659792e-02 -5.87318651e-02 3.66866112e-01
9.88869309e-01 3.11908692e-01 2.10188374e-01 -5.05868867e-02
7.88957357e-01 -6.44394100e-01 5.62413573e-01 9.31763649e-01
-2.29864955e-01 8.66476715e-01 2.91542083e-01 -3.51709157e-01
-9.26630139e-01 -1.23053145e+00 -4.89469320e-02 1.24340141e+00
-1.58715844e-02 -1.24968283e-01 -5.33801615e-01 -6.91954017e-01
1.63177609e-01 6.56270444e-01 -8.52142274e-01 -1.52350426e-01
-4.87844050e-01 -8.69746983e-01 5.63915908e-01 2.23356187e-01
6.61560237e-01 -1.10049701e+00 -1.29556835e-01 -5.95537573e-02
-1.51541248e-01 -1.31042027e+00 -7.15888739e-01 4.05597717e-01
-7.01409459e-01 -9.84503329e-01 -8.89980197e-01 -8.12344909e-01
7.05530465e-01 8.87695849e-01 8.80057871e-01 1.38384655e-01
-1.31157324e-01 5.00238419e-01 -1.74586982e-01 1.51107073e-01
-3.53512347e-01 -2.08669249e-02 1.00594595e-01 4.79213953e-01
-1.07669108e-01 -8.83303404e-01 -6.01129830e-01 2.35243112e-01
-1.23375225e+00 1.60422981e-01 7.77406991e-01 9.40283120e-01
3.82271647e-01 -1.99131310e-01 5.16874790e-01 -5.48420429e-01
5.57641685e-01 -1.86380938e-01 -6.36635065e-01 3.72962564e-01
-5.12732387e-01 1.09863050e-01 3.69847029e-01 -4.74015027e-01
-1.05792558e+00 -9.87510569e-03 1.81540683e-01 -8.42610538e-01
-2.47460499e-01 -6.86313435e-02 -3.88197631e-01 -1.70559525e-01
4.68425632e-01 2.36897588e-01 1.09147735e-01 -6.73521101e-01
7.19496787e-01 6.77272737e-01 8.70745778e-01 -3.45931083e-01
1.07741284e+00 3.22451681e-01 -5.68418682e-01 -3.25921118e-01
-9.80338454e-01 -2.44385242e-01 -4.35751647e-01 1.02094606e-01
1.12187564e+00 -1.46315992e+00 -7.73720264e-01 1.08233202e+00
-1.24038911e+00 -6.49915695e-01 -2.25177437e-01 2.24178866e-01
-4.88096833e-01 4.85049486e-01 -5.75318575e-01 -6.42412305e-01
-4.40088689e-01 -1.43970454e+00 1.33705163e+00 2.13761121e-01
3.12542021e-01 -1.02863991e+00 -8.44222754e-02 7.91955054e-01
2.67706215e-01 -3.11294138e-01 8.65526497e-01 -6.14884019e-01
-6.43281162e-01 2.54789442e-01 -5.84113777e-01 9.97881711e-01
3.35960597e-01 -4.25325960e-01 -1.42263663e+00 -7.41772652e-01
2.07998738e-01 -4.72896069e-01 1.55927789e+00 7.30688095e-01
1.12196362e+00 -2.05419227e-01 -7.29390010e-02 9.65077400e-01
1.40452898e+00 -1.87002897e-01 7.74530232e-01 3.50949496e-01
8.18698466e-01 3.40228140e-01 -7.75982440e-03 -7.48350099e-02
4.59129781e-01 5.75709343e-01 7.42690325e-01 -6.09785259e-01
-7.35216439e-01 8.09813812e-02 8.89096797e-01 2.91219622e-01
2.41542071e-01 -3.04218590e-01 -6.46216154e-01 7.46471882e-01
-2.03133845e+00 -7.47395694e-01 2.03894123e-01 2.10376787e+00
7.48022377e-01 -8.38309228e-02 -2.29888216e-01 -6.24141544e-02
9.28662300e-01 4.80951011e-01 -6.96412444e-01 -2.91888475e-01
-4.63151842e-01 1.64897069e-01 4.61530060e-01 9.72751737e-01
-1.27195692e+00 1.04511356e+00 5.59235144e+00 8.11958015e-01
-7.93686271e-01 3.83725077e-01 5.00383794e-01 -2.88070589e-01
-2.05518737e-01 -4.18417947e-03 -8.64138424e-01 1.09084934e-01
5.89581132e-01 1.67694479e-01 7.72929192e-01 2.88969517e-01
1.80641323e-01 -1.84651881e-01 -9.63104606e-01 9.42010403e-01
2.87747681e-01 -1.10207641e+00 3.92368376e-01 1.39766365e-01
6.65995717e-01 2.84410089e-01 4.40807968e-01 1.39569163e-01
8.20275903e-01 -1.06511664e+00 5.67277074e-01 3.83625388e-01
7.84005642e-01 -4.41941708e-01 5.03502607e-01 5.45011200e-02
-1.03043818e+00 -5.87641954e-01 -9.23982337e-02 3.32894295e-01
2.76941627e-01 8.40309381e-01 -3.63867790e-01 9.33824301e-01
7.81903803e-01 9.18344557e-01 -6.10843480e-01 8.17588747e-01
-5.58064997e-01 2.02405944e-01 3.96453924e-02 1.26548910e+00
2.05571018e-02 -1.90591961e-01 5.77321708e-01 1.02488136e+00
2.45185629e-01 -3.71696442e-01 5.99118024e-02 7.34360874e-01
-2.63686329e-01 -5.46400309e-01 -4.78595227e-01 4.63788837e-01
2.18073249e-01 1.59777689e+00 -1.19683512e-01 -2.49653861e-01
-5.53723931e-01 1.33234203e+00 4.17895675e-01 8.28315258e-01
-7.30370283e-01 -4.94447872e-02 1.07459939e+00 -3.40993553e-01
6.85380399e-01 -1.92969933e-01 -4.87926364e-01 -1.41081834e+00
-4.31331918e-02 -1.04855430e+00 4.45528835e-01 -9.83735502e-01
-1.41442859e+00 7.57759690e-01 -4.04545844e-01 -1.06099606e+00
5.52274659e-02 -6.47909462e-01 -5.58589101e-01 1.33505297e+00
-2.13629007e+00 -1.66518414e+00 -1.42089620e-01 1.12826395e+00
4.21045631e-01 -3.31244081e-01 7.18546987e-01 3.27378094e-01
-9.95924771e-01 6.69885695e-01 -1.69486567e-01 1.91057831e-01
1.12385809e+00 -1.22316945e+00 3.83798391e-01 1.53053463e+00
-7.71185979e-02 6.74881220e-01 3.42341572e-01 -6.06623769e-01
-9.12811100e-01 -1.37643659e+00 4.88110840e-01 -2.24825501e-01
6.38962924e-01 -5.93924880e-01 -9.99260426e-01 9.98079479e-01
3.96254539e-01 2.82243073e-01 6.30329430e-01 3.67168449e-02
-8.68549109e-01 -3.34608227e-01 -1.09253383e+00 5.83286881e-01
7.86275983e-01 -9.20993626e-01 -6.38121128e-01 2.76925057e-01
9.44119275e-01 -2.11849973e-01 -5.45339346e-01 4.06815916e-01
2.37768620e-01 -9.52160060e-01 1.14629114e+00 -1.26149401e-01
3.11328024e-01 -6.37433112e-01 -4.41195577e-01 -1.21053445e+00
-3.21113706e-01 -7.84762323e-01 -3.09732199e-01 1.18511069e+00
3.55671138e-01 -7.72793293e-01 1.34850875e-01 2.13348478e-01
-4.00733382e-01 -1.78923935e-01 -1.12088358e+00 -5.95463872e-01
1.21938856e-02 -3.21468711e-01 1.53022811e-01 7.09190428e-01
-7.59684861e-01 4.16548163e-01 -6.79254234e-01 6.55160308e-01
8.31489801e-01 -1.22863986e-01 5.74222445e-01 -1.01208961e+00
-4.70292181e-01 -3.16036195e-01 1.60392135e-01 -1.10054934e+00
2.29460195e-01 -1.13974953e+00 3.28854062e-02 -1.70574582e+00
1.21518128e-01 -2.11405456e-02 -3.97585452e-01 1.10012734e+00
-3.25700819e-01 3.76957387e-01 4.27947074e-01 3.40956211e-01
-2.79884130e-01 6.96359932e-01 1.25903428e+00 -7.41333589e-02
-2.18758479e-01 -8.64094682e-03 -1.22372746e+00 5.02564192e-01
5.19059896e-01 -2.99457878e-01 -4.69487160e-01 -1.00986183e+00
-2.78552979e-01 -9.82117802e-02 9.31652844e-01 -8.89559507e-01
4.98769134e-01 4.49191362e-01 3.31917197e-01 -2.92358577e-01
3.21819544e-01 -8.01006377e-01 -4.93498221e-02 2.51235276e-01
-1.36964619e-01 -1.64059341e-01 3.91054481e-01 4.80708778e-01
-1.21452831e-01 1.81910351e-01 1.13551891e+00 7.85675794e-02
-4.98859346e-01 4.31381047e-01 -3.85056317e-01 -5.21496713e-01
7.65646279e-01 -3.37056726e-01 -5.94559669e-01 -1.57618195e-01
-9.79370296e-01 3.46363783e-01 2.61101037e-01 5.47816575e-01
7.37454653e-01 -1.12638521e+00 -8.10339689e-01 4.44367409e-01
-1.23865664e-01 1.93131864e-01 5.44002473e-01 7.42038906e-01
2.21055254e-01 9.59515795e-02 -1.88188121e-01 -4.14862573e-01
-1.15295434e+00 6.93261802e-01 7.63646960e-01 -4.96960968e-01
-5.95037401e-01 8.03056657e-01 7.74531007e-01 -6.89335406e-01
3.00967842e-01 -1.53306976e-01 -1.08619079e-01 2.60383606e-01
6.12571120e-01 2.00048655e-01 1.82273299e-01 -4.54486817e-01
-9.34628695e-02 5.82447708e-01 -5.73293865e-01 -1.67606294e-01
1.38667524e+00 -4.63309854e-01 -4.20888931e-01 1.71826817e-02
1.09300089e+00 -2.78484970e-01 -1.62025535e+00 -5.28054357e-01
-3.85652393e-01 -8.81013349e-02 5.51315665e-01 -1.17506349e+00
-1.39554667e+00 8.28639627e-01 7.77890265e-01 -2.16744632e-01
1.76957881e+00 -1.52404457e-01 6.82477057e-01 -2.70787384e-02
-8.74691978e-02 -5.56937516e-01 1.64859429e-01 4.76945072e-01
1.16649890e+00 -1.11522055e+00 -4.20170762e-02 -4.93426286e-02
-4.15169418e-01 8.62476289e-01 7.12496758e-01 -1.74678147e-01
5.13164580e-01 3.54883581e-01 2.30209827e-01 -3.07253264e-02
-7.49477267e-01 -2.55742610e-01 3.32137704e-01 7.59833276e-01
1.14489026e-01 -2.41688192e-01 2.51756519e-01 4.51372594e-01
1.43448785e-01 -4.21479419e-02 5.40712893e-01 5.25721312e-01
-1.84418112e-01 -1.37116492e+00 -4.85659957e-01 1.98178977e-01
-3.39077801e-01 -5.63461661e-01 8.46037492e-02 3.93785387e-01
2.30693191e-01 1.04701269e+00 -6.76765516e-02 -4.28453863e-01
2.16554716e-01 -9.91253629e-02 3.06980848e-01 -6.66908979e-01
-9.55228209e-01 2.49877051e-01 -2.41073221e-01 -5.30922949e-01
-5.73636949e-01 -5.77880561e-01 -7.97155499e-01 -1.78325564e-01
-1.79373890e-01 -2.46593580e-01 2.73399651e-01 9.82779384e-01
5.74634910e-01 6.21654928e-01 6.69148922e-01 -1.11478519e+00
-3.21629882e-01 -9.34135199e-01 -6.03694618e-01 -4.09140661e-02
1.21980047e+00 -5.08313596e-01 -6.28938735e-01 1.19584702e-01] | [11.228264808654785, -2.2875568866729736] |
de6ba88c-a91c-40ad-b366-93802f99b0f5 | df-platter-multi-face-heterogeneous-deepfake | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Narayan_DF-Platter_Multi-Face_Heterogeneous_Deepfake_Dataset_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Narayan_DF-Platter_Multi-Face_Heterogeneous_Deepfake_Dataset_CVPR_2023_paper.pdf | DF-Platter: Multi-Face Heterogeneous Deepfake Dataset | Deepfake detection is gaining significant importance in the research community. While most of the research efforts are focused around high-quality images and videos, deepfake generation algorithms today have the capability to generate low-resolution videos, occluded deepfakes, and multiple-subject deepfakes. In this research, we emulate the real-world scenario of deepfake generation and spreading, and propose the DF-Platter dataset, which contains (i) both low-resolution and high-resolution deepfakes generated using multiple generation techniques and (ii) single-subject and multiple-subject deepfakes, with face images of Indian ethnicity. Faces in the dataset are annotated for various attributes such as gender, age, skin tone, and occlusion. The database is prepared in 116 days with continuous usage of 32 GPUs accounting to 1,800 GB cumulative memory. With over 500 GBs in size, the dataset contains a total of 133,260 videos encompassing three sets. To the best of our knowledge, this is one of the largest datasets containing vast variability and multiple challenges. We also provide benchmark results under multiple evaluation settings using popular and state-of-the-art deepfake detection models. Further, benchmark results under c23 and c40 compression are provided. The results demonstrate a significant performance reduction in the deepfake detection task on low-resolution deepfakes and show that the existing techniques fail drastically on multiple-subject deepfakes. It is our assertion that this database will improve the state-of-the-art by extending the capabilities of deepfake detection algorithms to real-world scenarios. The database is available at: http://iab-rubric.org/df-platter-database. | ['Richa Singh', 'Mayank Vatsa', 'Surbhi Mittal', 'Kartik Thakral', 'Harsh Agarwal', 'Kartik Narayan'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['deepfake-detection', 'face-swapping'] | ['computer-vision', 'computer-vision'] | [-1.50060177e-01 -2.88814634e-01 -1.46590501e-01 -2.13290602e-01
-7.15737045e-01 -5.30340374e-01 6.65334702e-01 -6.04998052e-01
-8.51386487e-02 5.80161095e-01 1.68707401e-01 1.37960792e-01
-5.10914857e-03 -8.66528630e-01 -6.41089916e-01 -7.54207075e-01
-3.20469528e-01 2.51298845e-01 1.25755280e-01 -3.73109318e-02
1.44187108e-01 6.35151088e-01 -1.96451294e+00 4.26806092e-01
4.26300317e-01 1.23628592e+00 -1.91343576e-02 8.45174193e-01
2.80868530e-01 5.47793448e-01 -7.86917210e-01 -7.92876184e-01
4.99297023e-01 -1.60189733e-01 -6.02420092e-01 -4.39235084e-02
1.41263866e+00 -9.88811910e-01 -5.97908318e-01 1.00217223e+00
9.73558486e-01 -2.96283662e-01 3.66744727e-01 -1.26506042e+00
-6.79341316e-01 2.15146810e-01 -1.08142161e+00 5.70746899e-01
3.30198497e-01 2.70740122e-01 5.59857786e-01 -1.34690988e+00
8.46467733e-01 1.68629491e+00 7.73602605e-01 6.31907225e-01
-8.76248777e-01 -1.15422809e+00 -2.44308710e-01 4.02358145e-01
-1.56159568e+00 -7.62470126e-01 2.94686168e-01 -6.22483432e-01
9.67433512e-01 1.61861740e-02 5.37393808e-01 1.29148567e+00
1.69658780e-01 5.90723932e-01 8.82034600e-01 -2.15298861e-01
-1.44530516e-02 -7.19147772e-02 -1.09067537e-01 7.64846385e-01
3.53765696e-01 2.83361107e-01 -6.96041405e-01 -3.35579991e-01
9.99019086e-01 -3.89532626e-01 -1.88243836e-01 -2.99722124e-02
-9.56556380e-01 1.05765903e+00 -4.71836925e-02 -6.67000264e-02
-2.09258363e-01 5.55328839e-02 4.99564618e-01 1.04396895e-01
5.24136841e-01 -6.61706552e-02 -2.00209796e-01 -1.69871360e-01
-1.15301144e+00 6.61516666e-01 6.13743365e-01 7.90098190e-01
6.02000773e-01 3.00765246e-01 -2.22966358e-01 9.59035218e-01
5.82343787e-02 7.25820243e-01 3.05958688e-01 -1.07495189e+00
4.02717561e-01 1.16036780e-01 3.28511628e-03 -1.30768824e+00
-1.74970075e-01 -2.22040802e-01 -7.82857060e-01 3.52179646e-01
5.96648335e-01 -2.49526158e-01 -1.02671945e+00 1.59242868e+00
5.38173139e-01 4.20758128e-01 -2.60976613e-01 1.07026160e+00
1.20892549e+00 7.43673503e-01 -9.07845125e-02 -7.37466812e-02
1.65959275e+00 -7.72244632e-01 -6.04770124e-01 -5.60182631e-02
1.74751773e-01 -1.08242667e+00 8.51286352e-01 7.87712753e-01
-1.13772702e+00 -6.81964397e-01 -9.54258025e-01 -1.09406210e-01
-1.91299587e-01 4.39756662e-01 6.05889499e-01 1.19062150e+00
-1.22220874e+00 1.10887066e-01 -3.71071368e-01 -4.95663911e-01
8.69775176e-01 2.66653091e-01 -5.65761983e-01 -2.10310847e-01
-1.07773459e+00 5.97636521e-01 8.51817988e-03 1.15709740e-03
-1.46002996e+00 -9.48866904e-01 -4.64950591e-01 -1.33790806e-01
4.06715363e-01 -4.80588794e-01 1.02247083e+00 -9.71754193e-01
-1.06016338e+00 1.19233251e+00 -8.58617201e-02 -3.63981694e-01
7.18512893e-01 -4.68295187e-01 -5.98049760e-01 4.21097398e-01
1.06492341e-01 8.67585242e-01 1.31404817e+00 -9.11482751e-01
-5.64381361e-01 -4.21390176e-01 -5.55452779e-02 -1.25282660e-01
-4.22912955e-01 5.79548955e-01 -5.78116775e-01 -9.71667409e-01
-7.13654935e-01 -8.42142344e-01 3.76855940e-01 1.48737326e-01
-4.08760518e-01 -2.41322637e-01 1.29935646e+00 -8.37092817e-01
1.25987732e+00 -2.13662887e+00 -2.65893131e-01 -1.51814669e-01
3.76445293e-01 5.65799475e-01 -4.54102695e-01 6.75374293e-04
-1.57471478e-01 1.01554744e-01 2.41223738e-01 -2.73113340e-01
-1.87364787e-01 -4.95813135e-03 -3.55374008e-01 6.25757217e-01
1.89910740e-01 6.81519091e-01 -5.30372918e-01 -2.82161385e-01
1.70921996e-01 7.34827816e-01 -8.43979418e-01 1.79861426e-01
1.88173890e-01 4.80776466e-02 -1.21343717e-01 9.76279259e-01
1.35013044e+00 -1.45469010e-01 -5.99678606e-02 -3.71511906e-01
-4.02032360e-02 -2.17713460e-01 -1.20470369e+00 1.41727567e+00
-3.04714292e-02 8.69760394e-01 4.12272334e-01 -5.93901336e-01
7.70918787e-01 9.41474214e-02 4.47900862e-01 -6.44652367e-01
2.07400575e-01 1.47263914e-01 -1.75235122e-02 -3.95302922e-01
6.80597842e-01 2.80345201e-01 7.64351934e-02 3.88568729e-01
2.16101080e-01 2.87147403e-01 5.23422301e-01 4.07820612e-01
1.02045667e+00 -1.85945146e-02 9.11046937e-02 -4.36802328e-01
4.57892239e-01 -2.76867807e-01 4.47869658e-01 7.68908262e-01
-3.74707460e-01 6.41498029e-01 5.24548709e-01 -6.80301905e-01
-1.20027816e+00 -1.27372527e+00 -4.58194286e-01 1.19811249e+00
-2.27213502e-01 -6.18438482e-01 -1.18687570e+00 -3.84231955e-01
1.43650353e-01 -5.07740257e-03 -7.93949544e-01 1.45284489e-01
-5.84613442e-01 -1.04286325e+00 1.04013467e+00 3.07987839e-01
9.75783229e-01 -1.00499570e+00 -6.68311000e-01 -9.26540121e-02
-2.94997811e-01 -1.36222148e+00 -4.74936604e-01 -7.18475938e-01
-2.99123675e-01 -1.31041968e+00 -1.03608716e+00 -6.43449128e-01
2.47859195e-01 1.60229817e-01 1.28550220e+00 -8.68843198e-02
-1.05414021e+00 3.11688662e-01 -2.52930641e-01 -1.10548459e-01
-2.96432048e-01 -3.18550080e-01 2.13355258e-01 4.02412489e-02
6.57436430e-01 -4.18282785e-02 -9.77751493e-01 4.93885756e-01
-6.47813976e-01 -2.17141777e-01 3.92312795e-01 8.03041816e-01
3.92422169e-01 4.34403233e-02 7.43061364e-01 -5.20194530e-01
4.49764341e-01 -6.13699555e-01 -6.01469457e-01 -2.46028230e-02
-2.03374714e-01 -5.99629760e-01 2.19085827e-01 -5.34302115e-01
-1.31507683e+00 -2.01276675e-01 -2.47104034e-01 -6.10017002e-01
-3.53463620e-01 -1.10493660e-01 -1.29309013e-01 -2.06672132e-01
7.27127910e-01 2.13833645e-01 -1.08621106e-01 -4.93967772e-01
2.30966717e-01 7.07597256e-01 7.72279739e-01 -5.50834656e-01
3.27586025e-01 5.47819078e-01 -2.52510369e-01 -1.06346357e+00
-3.45410794e-01 -1.90617740e-01 -2.46952698e-01 -3.90367717e-01
6.70414567e-01 -1.35134566e+00 -7.10736573e-01 1.24689400e+00
-1.03057861e+00 -2.61365116e-01 6.71460405e-02 1.20661296e-01
-3.79941612e-01 4.59293365e-01 -8.45571399e-01 -5.17250896e-01
-5.13232231e-01 -1.28690565e+00 1.36580241e+00 1.94785640e-01
3.51358242e-02 -4.43285704e-01 -2.79672481e-02 4.79685664e-01
3.46460134e-01 1.30274057e-01 5.08262575e-01 -5.03025427e-02
-6.57237172e-01 2.06949741e-01 -6.12328768e-01 3.72862220e-01
7.48738721e-02 2.68524766e-01 -1.21972585e+00 -5.70637405e-01
-3.86743039e-01 -5.40407538e-01 9.48222220e-01 6.97992623e-01
1.35211182e+00 -2.20441177e-01 -1.47065118e-01 8.07180285e-01
1.33913016e+00 7.81919733e-02 1.08905470e+00 2.52041429e-01
6.44875944e-01 5.82637250e-01 6.23279274e-01 8.99628639e-01
2.35193700e-01 8.90871286e-01 4.02358979e-01 -3.86356463e-04
-6.50481164e-01 6.87082335e-02 4.69543546e-01 8.33994821e-02
-1.15926705e-01 -3.65968108e-01 -8.80112946e-01 6.70482695e-01
-1.39284134e+00 -1.21025300e+00 -2.55490124e-01 1.96046197e+00
7.42157698e-01 -5.13116002e-01 4.79570508e-01 3.51836197e-02
8.26804876e-01 2.27104366e-01 -5.42070568e-01 -5.04035175e-01
-3.49744290e-01 3.44695061e-01 3.28603327e-01 1.04395650e-01
-1.41722012e+00 9.84434664e-01 6.44923449e+00 1.08086741e+00
-1.21018243e+00 2.18744472e-01 9.34265673e-01 -3.65676522e-01
3.26265514e-01 -7.73308098e-01 -1.39296055e+00 7.40446210e-01
9.45881188e-01 -8.24898258e-02 3.55128437e-01 9.45334315e-01
2.34145358e-01 -1.03936628e-01 -7.01510012e-01 1.22484231e+00
4.90029901e-01 -1.47454953e+00 1.22550525e-01 2.86315948e-01
8.29477429e-01 1.28005415e-01 3.39562833e-01 1.53487742e-01
1.77003905e-01 -1.26429617e+00 7.68036664e-01 8.65609124e-02
1.44459152e+00 -8.86051536e-01 6.28214061e-01 -2.01468751e-01
-1.38381767e+00 -1.35381222e-01 -4.52368706e-01 2.27921933e-01
-2.47173794e-02 6.10889792e-01 -5.80238521e-01 1.52332172e-01
1.42619038e+00 7.25558817e-01 -6.03506923e-01 8.78690660e-01
1.95262969e-01 5.14653027e-01 -4.23833579e-01 4.15213794e-01
-1.62649751e-01 1.73075214e-01 3.61352265e-01 1.48434365e+00
7.15721548e-01 2.92357747e-02 1.23973396e-02 5.13918638e-01
-1.99403197e-01 7.14313164e-02 -5.08204877e-01 4.87992726e-02
5.76632380e-01 1.35912812e+00 -5.60986638e-01 -4.75744843e-01
-5.86956143e-01 7.13734508e-01 -3.32917608e-02 2.01366395e-01
-1.03824532e+00 -1.08703516e-01 1.35362840e+00 3.69979888e-01
4.76613343e-01 7.40890577e-02 -2.24820320e-02 -1.12561059e+00
-2.67823935e-01 -1.17196488e+00 5.75155795e-01 -7.19962895e-01
-1.40883088e+00 6.65898025e-01 7.08090067e-02 -9.33170855e-01
-2.66732484e-01 -7.79447317e-01 -1.40683025e-01 7.61575401e-01
-1.43131506e+00 -1.17166507e+00 -7.25428760e-01 1.02866697e+00
7.62054801e-01 -5.09934962e-01 6.97681487e-01 7.96759844e-01
-5.77659607e-01 8.64334524e-01 -2.15583220e-02 2.49769196e-01
9.96249735e-01 -7.34606624e-01 6.32633090e-01 9.20646071e-01
-1.97314516e-01 4.59843874e-01 5.18092453e-01 -7.30512142e-01
-1.41666198e+00 -1.12548530e+00 3.17087948e-01 -2.27737471e-01
6.11131907e-01 -5.59164166e-01 -7.81134903e-01 4.81436670e-01
2.18636796e-01 2.75636017e-01 5.37698865e-01 -3.14368844e-01
-4.39979404e-01 -1.73771665e-01 -1.39435267e+00 3.44957650e-01
1.12726665e+00 -2.57020175e-01 -6.87081963e-02 1.26178965e-01
-1.13548584e-01 -5.78364491e-01 -8.62103760e-01 5.27734399e-01
7.62096643e-01 -1.33845472e+00 1.31793547e+00 -9.32419226e-02
6.31822348e-01 -1.37510464e-01 -1.13732338e-01 -1.03289497e+00
-3.01922083e-01 -5.00330091e-01 -4.70875770e-01 1.29355001e+00
-3.53898346e-01 -4.90335345e-01 9.51669753e-01 1.33165196e-01
1.78721547e-01 -4.93938059e-01 -9.95311558e-01 -6.60654366e-01
1.58551484e-01 -1.95196912e-01 6.80126071e-01 7.92485833e-01
-7.28036225e-01 -1.20700195e-01 -8.91070068e-01 4.14496958e-02
8.79022360e-01 -3.52904275e-02 1.21101272e+00 -1.10264182e+00
-6.18299432e-02 -3.94825220e-01 -4.83236849e-01 -9.44142222e-01
-3.90525050e-02 -3.89793068e-01 -3.03321421e-01 -9.73592103e-01
4.19614106e-01 -2.57002801e-01 2.08309099e-01 6.41755104e-01
1.14237279e-01 9.46907699e-01 1.17403477e-01 -2.64706388e-02
-3.56219038e-02 3.14644605e-01 1.24703634e+00 4.69907224e-02
3.10725778e-01 -6.26463532e-01 -4.86792207e-01 7.51610935e-01
6.67181492e-01 -8.90847892e-02 -4.36995625e-01 -3.59613061e-01
-2.17753667e-02 -1.53156921e-01 6.29522800e-01 -9.08399343e-01
-1.15978055e-01 -6.33973852e-02 7.59502828e-01 -5.57507992e-01
5.07931352e-01 -2.18435138e-01 2.16499329e-01 2.89311707e-01
1.04418203e-01 7.30136968e-03 4.94920671e-01 1.91629529e-01
-1.75711691e-01 2.43690863e-01 1.11615109e+00 -8.09515640e-02
-1.29920387e+00 5.24550974e-01 -3.76123905e-01 2.18680084e-01
1.09605193e+00 -3.71967942e-01 -9.96459007e-01 -3.59127045e-01
-4.56295967e-01 -2.69849271e-01 3.05482805e-01 8.17627907e-01
8.34245026e-01 -1.26688051e+00 -1.09607697e+00 6.70962870e-01
-1.10458424e-02 -4.84452426e-01 7.40446329e-01 5.05906403e-01
-7.30338395e-01 2.51951247e-01 -6.32783771e-01 -7.39696443e-01
-1.83848953e+00 2.46678725e-01 2.86595225e-01 2.18243778e-01
-6.10587597e-01 1.08911741e+00 6.94051087e-01 4.76029590e-02
1.05894223e-01 3.23881149e-01 -2.30615884e-01 2.74319887e-01
1.12197888e+00 5.87601423e-01 1.58487126e-01 -1.07300258e+00
-5.09407640e-01 6.96944296e-01 -4.08949465e-01 3.84979069e-01
1.06332195e+00 -1.42670363e-01 7.23372102e-02 -5.51276445e-01
1.12292528e+00 1.96917653e-01 -1.42356765e+00 1.73332691e-01
-5.71114182e-01 -1.07447410e+00 1.97656989e-01 -9.66228545e-01
-1.47785258e+00 9.16661918e-01 1.13472116e+00 -1.77760884e-01
1.40378666e+00 7.58677945e-02 8.56646895e-01 -2.35405445e-01
5.45671284e-01 -7.46550500e-01 4.00265366e-01 3.64943087e-01
1.01890254e+00 -9.99693274e-01 -3.19843995e-03 -5.37041962e-01
-3.59190077e-01 1.06105697e+00 9.50232506e-01 -7.26973489e-02
5.45565784e-01 7.59328425e-01 7.76057243e-02 -7.91590735e-02
-7.24031031e-01 1.16832785e-01 8.14788416e-02 6.66663766e-01
2.98055768e-01 4.79891058e-03 6.02598593e-04 3.63594294e-01
-3.85374904e-01 2.51511604e-01 5.75917900e-01 4.33508694e-01
-2.74941444e-01 -8.07430625e-01 -5.02288342e-01 5.24296463e-01
-8.85288477e-01 -9.35172737e-02 -2.16725901e-01 7.09116936e-01
3.51079106e-01 1.01136136e+00 2.38489106e-01 -2.03439772e-01
3.12496927e-02 -4.34622109e-01 6.40833735e-01 -2.33817294e-01
-4.80648100e-01 1.51905701e-01 1.13733999e-01 -7.99475312e-01
-2.45357826e-01 -6.06575549e-01 -7.26265967e-01 -1.07597542e+00
-6.82672160e-03 -3.21198493e-01 4.75287378e-01 4.43533659e-01
5.22541761e-01 9.99518782e-02 2.93396950e-01 -1.09872448e+00
-8.76366943e-02 -8.82670045e-01 -6.65214956e-01 2.95708269e-01
2.57841408e-01 -9.41466570e-01 -2.84310013e-01 1.58644438e-01] | [13.033408164978027, 0.9057313203811646] |
570bdfe6-d53e-40c0-a2b2-a8a1b70158f1 | rethinking-resolution-in-the-context-of | 2209.12797 | null | https://arxiv.org/abs/2209.12797v1 | https://arxiv.org/pdf/2209.12797v1.pdf | Rethinking Resolution in the Context of Efficient Video Recognition | In this paper, we empirically study how to make the most of low-resolution frames for efficient video recognition. Existing methods mainly focus on developing compact networks or alleviating temporal redundancy of video inputs to increase efficiency, whereas compressing frame resolution has rarely been considered a promising solution. A major concern is the poor recognition accuracy on low-resolution frames. We thus start by analyzing the underlying causes of performance degradation on low-resolution frames. Our key finding is that the major cause of degradation is not information loss in the down-sampling process, but rather the mismatch between network architecture and input scale. Motivated by the success of knowledge distillation (KD), we propose to bridge the gap between network and input size via cross-resolution KD (ResKD). Our work shows that ResKD is a simple but effective method to boost recognition accuracy on low-resolution frames. Without bells and whistles, ResKD considerably surpasses all competitive methods in terms of efficiency and accuracy on four large-scale benchmark datasets, i.e., ActivityNet, FCVID, Mini-Kinetics, Something-Something V2. In addition, we extensively demonstrate its effectiveness over state-of-the-art architectures, i.e., 3D-CNNs and Video Transformers, and scalability towards super low-resolution frames. The results suggest ResKD can serve as a general inference acceleration method for state-of-the-art video recognition. Our code will be available at https://github.com/CVMI-Lab/ResKD. | ['Xiaojuan Qi', 'Ping Luo', 'Zehuan Yuan', 'Yi Jiang', 'Qiushan Guo', 'Chuofan Ma'] | 2022-09-26 | null | null | null | null | ['video-recognition'] | ['computer-vision'] | [-1.42201176e-02 -5.17663181e-01 -4.48246211e-01 -9.08040181e-02
-5.37859976e-01 -2.26032883e-01 4.31025237e-01 -4.37975228e-01
-7.23552406e-01 6.49410784e-01 2.22310618e-01 -1.05697222e-01
6.99151084e-02 -6.08124375e-01 -9.01124477e-01 -5.71425557e-01
9.93868187e-02 -9.12570208e-02 4.36820060e-01 -2.08894163e-02
1.48083687e-01 6.05347514e-01 -1.60699689e+00 4.52330381e-01
2.67376870e-01 1.00732994e+00 1.97125182e-01 7.55060732e-01
-1.39030600e-02 1.10318303e+00 -6.00466490e-01 -3.65249872e-01
2.64473021e-01 -3.33727688e-01 -7.52231538e-01 -2.17557937e-01
6.11618519e-01 -7.58539557e-01 -8.24728727e-01 8.10872018e-01
6.35506988e-01 9.22010317e-02 3.45854521e-01 -1.07604611e+00
-5.13613343e-01 5.26976764e-01 -6.59887373e-01 1.03006732e+00
1.17389619e-01 2.84831971e-01 7.33638108e-01 -1.09647906e+00
6.65411532e-01 1.30434370e+00 6.91107333e-01 7.73014963e-01
-1.12012517e+00 -7.52527177e-01 2.24411711e-01 5.72619557e-01
-1.66948080e+00 -8.07305872e-01 4.40395325e-01 -2.72811890e-01
1.43359244e+00 6.90349340e-02 6.71359241e-01 1.34564865e+00
3.30270715e-02 8.73059928e-01 8.33645046e-01 -2.35813782e-01
1.13522299e-01 -3.36549670e-01 -1.66959185e-02 6.20097160e-01
3.89181405e-01 1.32244036e-01 -1.09793460e+00 2.25707188e-01
1.31885481e+00 -1.33023923e-02 -5.03852725e-01 -4.12980579e-02
-1.15194166e+00 4.66245919e-01 1.83857337e-01 3.88792902e-01
-3.76879573e-01 4.29507405e-01 5.53021789e-01 3.45400304e-01
4.65835512e-01 3.21244568e-01 -5.55633068e-01 -7.07200646e-01
-1.00285852e+00 6.46151677e-02 5.90096712e-01 6.38498485e-01
4.44700956e-01 4.04431522e-01 -1.57601267e-01 7.60619283e-01
-8.65866616e-02 1.94118500e-01 4.67021674e-01 -1.26618230e+00
3.45437050e-01 2.79649824e-01 -6.74487799e-02 -9.35168087e-01
-2.74950743e-01 -3.64862204e-01 -1.01815033e+00 1.39253929e-01
6.82439029e-01 3.56571525e-02 -8.68819714e-01 1.72232187e+00
-4.07799520e-02 5.31784415e-01 6.37326390e-02 1.18569875e+00
8.63865316e-01 5.99775970e-01 8.08643475e-02 -2.75010437e-01
1.36372328e+00 -1.04262960e+00 -6.25897884e-01 1.19832670e-02
3.97135496e-01 -4.01003569e-01 1.08839095e+00 4.65385675e-01
-1.35731709e+00 -7.73714066e-01 -1.03515422e+00 -1.10621743e-01
-2.12533891e-01 2.02079579e-01 6.49103045e-01 3.18098485e-01
-1.11367297e+00 8.11350584e-01 -1.12646365e+00 -4.01905000e-01
7.86511958e-01 3.53062183e-01 -3.98756504e-01 -1.07722782e-01
-1.10712039e+00 7.78759837e-01 3.06191057e-01 -1.02771088e-01
-8.85479808e-01 -8.78735721e-01 -5.00070989e-01 1.95003852e-01
5.73662877e-01 -6.71021104e-01 1.16423774e+00 -1.01815665e+00
-1.42573214e+00 7.37986922e-01 -1.81723014e-01 -8.22178483e-01
5.03826976e-01 -4.99428511e-01 -3.30351532e-01 4.09773499e-01
-3.81038368e-01 9.55496907e-01 9.08097148e-01 -7.22187757e-01
-6.11192703e-01 -1.46045282e-01 1.81229472e-01 2.18849421e-01
-5.67653775e-01 8.88488665e-02 -9.07116175e-01 -8.04606020e-01
-2.59250611e-01 -6.83840096e-01 1.39142096e-01 -1.10221123e-02
1.65413618e-02 -2.85608888e-01 8.02680373e-01 -5.91132224e-01
1.36277485e+00 -2.20572257e+00 1.55876607e-01 -2.04175606e-01
3.89792204e-01 6.49982631e-01 -1.63575679e-01 -7.86769092e-02
-6.24120571e-02 2.68351436e-01 2.71421909e-01 -1.02661587e-01
-3.25915664e-01 2.56512135e-01 -1.71624988e-01 4.00084376e-01
3.98691624e-01 1.11716318e+00 -6.84382319e-01 -3.14157844e-01
2.48989895e-01 9.30547833e-01 -5.61827064e-01 -4.64357529e-03
-1.89752225e-02 2.36596689e-01 -1.55737564e-01 6.94555521e-01
4.25694019e-01 -6.83234930e-01 1.30621925e-01 -6.46219611e-01
-9.65916291e-02 3.02359372e-01 -1.12838650e+00 1.81359637e+00
-4.25153188e-02 8.92180800e-01 -1.46057189e-01 -9.96628284e-01
6.35769606e-01 2.51200289e-01 4.45899159e-01 -1.03026247e+00
9.45773721e-02 -1.34086637e-02 -3.21855918e-02 -3.11120629e-01
4.19735253e-01 3.04553330e-01 4.93177652e-01 2.19531193e-01
1.30655393e-01 5.16030490e-01 4.58396494e-01 8.17314759e-02
1.21238589e+00 2.59325683e-01 2.51850307e-01 -1.60593733e-01
2.90413320e-01 -2.02761814e-01 6.09903693e-01 7.54755199e-01
-2.79446632e-01 5.51413655e-01 5.26891768e-01 -7.05606878e-01
-1.22301388e+00 -9.01225567e-01 1.12761766e-01 1.16411734e+00
3.81268598e-02 -6.06431127e-01 -8.60129297e-01 -3.79617512e-01
-2.04534322e-01 6.81848824e-02 -4.87927228e-01 -1.72394261e-01
-7.96959758e-01 -6.97019160e-01 9.04180944e-01 7.98123419e-01
9.53564703e-01 -8.22522461e-01 -8.16653669e-01 7.51203299e-02
-2.48964563e-01 -1.41767561e+00 -3.59465837e-01 3.78645584e-02
-1.01270711e+00 -1.01814091e+00 -6.50081396e-01 -3.77330810e-01
3.32457870e-01 5.83645642e-01 1.29298425e+00 2.10392863e-01
-2.67926931e-01 1.92277238e-01 -2.87510961e-01 -6.26157075e-02
-9.32445079e-02 2.22289145e-01 2.51333565e-01 -3.44945192e-01
5.68942547e-01 -6.83772027e-01 -8.55035484e-01 5.03060520e-01
-8.82030070e-01 3.19472343e-01 6.84896290e-01 8.01615179e-01
7.43913472e-01 2.95902826e-02 3.42666715e-01 -4.41431433e-01
4.06971782e-01 -2.28009686e-01 -4.60905552e-01 3.99733474e-03
-5.58092535e-01 1.58007741e-01 5.98661423e-01 -6.76957369e-01
-8.21466982e-01 -5.92759326e-02 -1.94079503e-01 -1.05685771e+00
-1.67180881e-01 2.95872837e-01 1.67975739e-01 -1.19230330e-01
7.05797195e-01 3.22579563e-01 3.17589045e-02 -4.42005277e-01
1.05471775e-01 2.25277498e-01 6.23195171e-01 -5.49032688e-01
4.26020414e-01 5.79690397e-01 -5.10538220e-02 -9.21035767e-01
-6.46285713e-01 -1.97332010e-01 -4.57971543e-01 -2.33019173e-01
8.56088698e-01 -1.40459442e+00 -8.70495677e-01 5.92055619e-01
-9.81596351e-01 -5.29509544e-01 -2.32396945e-01 4.86575961e-01
-3.80380630e-01 1.67462230e-01 -1.03115249e+00 -4.68307644e-01
-3.56826901e-01 -1.06219602e+00 7.42192090e-01 2.86872804e-01
-1.18468650e-01 -5.97848117e-01 -2.41076484e-01 3.92124027e-01
5.79786599e-01 -2.53175888e-02 5.01852512e-01 -4.02673841e-01
-8.84846807e-01 2.95286953e-01 -5.87420940e-01 4.25352722e-01
-1.08436994e-01 1.10291764e-01 -9.80325878e-01 -3.56486142e-01
-3.75392050e-01 -4.19639170e-01 1.12071657e+00 5.78730464e-01
1.34837902e+00 -2.79597729e-01 -1.09065086e-01 8.98041725e-01
1.25438499e+00 1.57723740e-01 9.86640394e-01 3.74181986e-01
7.40989149e-01 2.91931592e-02 3.54424447e-01 4.53162521e-01
3.45533252e-01 8.73046994e-01 2.09315747e-01 -4.01852876e-02
-5.18504083e-01 -7.27990940e-02 5.25004327e-01 8.76166761e-01
-6.24623537e-01 -1.71323493e-01 -8.58378112e-01 3.50537270e-01
-1.90854800e+00 -1.30055177e+00 1.93673134e-01 1.97511959e+00
8.55060339e-01 3.66632283e-01 3.76060158e-01 6.89060614e-02
5.64383030e-01 2.72507519e-01 -7.20265687e-01 -9.09916088e-02
-4.10222381e-01 1.57326028e-01 5.99512398e-01 1.64334163e-01
-9.63292718e-01 1.00062633e+00 6.10615492e+00 1.13123810e+00
-1.20067179e+00 2.26418719e-01 8.55357647e-01 -5.39398015e-01
3.56218457e-01 -3.52113456e-01 -1.05887508e+00 4.99475837e-01
1.25933135e+00 1.15044795e-01 6.98315799e-01 7.97129154e-01
2.94745922e-01 3.91471907e-02 -1.19257617e+00 1.36066151e+00
4.53631254e-03 -1.93904984e+00 1.46653786e-01 -5.26969656e-02
5.62507987e-01 2.75991470e-01 -1.42974094e-01 3.09825718e-01
5.82531132e-02 -1.14784074e+00 6.68057501e-01 4.79370981e-01
1.03674543e+00 -6.11259103e-01 5.84436953e-01 3.62021998e-02
-1.37315166e+00 3.22807543e-02 -4.75897193e-01 -2.40970537e-01
-8.24003518e-02 4.36947137e-01 -3.56286675e-01 1.53428972e-01
1.11504471e+00 9.12620008e-01 -4.22595799e-01 6.84441447e-01
-1.59349083e-03 6.87761307e-01 -2.96732873e-01 2.41926655e-01
9.58782434e-02 3.08422089e-01 1.89910382e-01 1.58195865e+00
2.09950089e-01 3.36864352e-01 -5.03092483e-02 6.97841346e-01
-3.96749973e-01 -3.60121101e-01 -4.15461600e-01 -1.83359981e-01
5.40540040e-01 1.01848900e+00 -6.85688198e-01 -5.02050817e-01
-7.55889714e-01 8.60940516e-01 4.10439819e-01 3.59050304e-01
-1.07451904e+00 9.19954106e-02 1.05221272e+00 9.47396755e-02
6.24621212e-01 -2.72416711e-01 -1.50087968e-01 -1.40266836e+00
5.09713739e-02 -1.07148039e+00 4.06670421e-01 -7.68617749e-01
-8.63511682e-01 4.71027493e-01 -1.20707050e-01 -1.03103912e+00
-2.92973947e-02 -5.97033381e-01 -9.63632539e-02 4.67185795e-01
-1.70256817e+00 -6.61132216e-01 -5.17851710e-01 8.37884367e-01
9.03888941e-01 -9.00445133e-02 5.40046096e-01 7.04296827e-01
-7.56565809e-01 7.19862342e-01 -1.43042684e-01 3.90676528e-01
5.79082429e-01 -7.42996812e-01 6.11723244e-01 9.74801421e-01
2.15121821e-01 4.61524636e-01 4.69866395e-01 -3.76036823e-01
-1.72562850e+00 -9.02862549e-01 4.39615309e-01 -3.58975649e-01
5.32904565e-01 -2.17915297e-01 -1.02579737e+00 3.84349942e-01
-2.18497254e-02 3.88473570e-01 3.90273899e-01 -1.91002294e-01
-6.48198426e-01 -1.90674394e-01 -7.86063492e-01 6.59920752e-01
1.40645099e+00 -6.37502253e-01 -2.02571407e-01 -2.72680447e-02
7.44715452e-01 -5.56281686e-01 -9.49313939e-01 4.25736636e-01
7.77450740e-01 -1.14278281e+00 1.29424977e+00 -6.13010883e-01
6.82218313e-01 -2.26625934e-01 -2.37837449e-01 -8.95728707e-01
-5.43324411e-01 -4.78817165e-01 -8.52817357e-01 9.54240739e-01
1.58249930e-01 -2.03718215e-01 1.00418568e+00 4.43158090e-01
1.15914546e-01 -7.89360940e-01 -9.83761787e-01 -8.71994495e-01
-2.23297715e-01 -5.59191346e-01 3.19322705e-01 8.31298530e-01
-2.26856232e-01 3.19493771e-01 -6.59702420e-01 2.14561205e-02
4.91337985e-01 -2.82801419e-01 7.56006181e-01 -8.94737005e-01
-3.82999599e-01 -6.19285703e-01 -4.55938369e-01 -1.29423416e+00
-6.91730008e-02 -3.84717882e-01 -3.05300593e-01 -1.17745292e+00
4.38816816e-01 -2.37256545e-03 -5.83813429e-01 6.02170289e-01
9.07991230e-02 5.96347570e-01 4.22764421e-01 4.77762520e-01
-8.55234802e-01 2.05974117e-01 1.24302971e+00 7.55880252e-02
-2.65571415e-01 -5.06860793e-01 -5.98900080e-01 6.64477944e-01
1.06039500e+00 -2.01503173e-01 -3.66896659e-01 -5.86506367e-01
1.46060646e-01 1.02050267e-01 5.15985429e-01 -1.19227040e+00
3.22991341e-01 -1.58986092e-01 6.22737586e-01 -3.61592025e-01
4.39611375e-01 -5.43784142e-01 2.23371804e-01 3.93044561e-01
-2.89385706e-01 2.66503543e-01 4.61182415e-01 3.52496445e-01
-1.97546065e-01 3.29281837e-01 8.66734266e-01 -1.73681349e-01
-1.17012107e+00 3.87700319e-01 -3.22100043e-01 2.46609479e-01
6.20655537e-01 -4.87279594e-01 -8.51262867e-01 -2.46861771e-01
-4.94317114e-01 -1.27702340e-01 4.06720728e-01 5.42168498e-01
8.05786312e-01 -1.28086352e+00 -6.15035415e-01 1.63749784e-01
-1.52603820e-01 -2.33338848e-01 3.29903245e-01 1.00548625e+00
-4.93443131e-01 5.74409068e-01 -4.74237978e-01 -6.10444665e-01
-1.44559813e+00 3.42613786e-01 3.84245396e-01 -2.53914446e-01
-6.77380443e-01 9.69606996e-01 1.07759118e-01 4.10707593e-01
5.84132910e-01 -4.59805220e-01 -3.35166492e-02 -2.61649936e-02
9.07380283e-01 4.35252249e-01 -1.82600189e-02 -4.34184194e-01
-5.02566576e-01 5.27538359e-01 -2.98388541e-01 2.14477688e-01
1.45591724e+00 -9.52897444e-02 2.26440221e-01 1.91559821e-01
1.10322404e+00 -5.98682821e-01 -1.82994795e+00 -3.44621241e-01
-2.87607700e-01 -6.21917069e-01 2.39808202e-01 -6.00444198e-01
-1.45662785e+00 7.38172412e-01 6.94913685e-01 -3.80556881e-02
1.29520905e+00 -4.95968461e-02 7.52677321e-01 4.85138237e-01
2.67333597e-01 -1.18771195e+00 3.47309768e-01 5.68569183e-01
6.46384120e-01 -1.25695050e+00 1.52221262e-01 -1.10549495e-01
-4.75344568e-01 1.06870830e+00 8.81876230e-01 -8.43047500e-02
3.66549611e-01 5.05997598e-01 -2.10447878e-01 -1.85184196e-01
-1.11268854e+00 -1.42970562e-01 1.97415665e-01 3.57514590e-01
4.98856366e-01 -2.73139417e-01 -1.50835877e-02 3.83879840e-01
1.62965804e-01 4.98676896e-01 4.07370597e-01 8.25180829e-01
-3.83568048e-01 -6.50189281e-01 -1.95069030e-01 4.60321426e-01
-8.32086146e-01 -2.60894299e-01 -1.10490523e-01 8.77175093e-01
4.42596748e-02 6.36201560e-01 1.87525868e-01 -6.34197176e-01
3.24822038e-01 -2.65680440e-02 5.92446446e-01 -4.63023642e-03
-3.54861468e-01 6.74032420e-02 8.34789202e-02 -9.92758691e-01
-7.16505826e-01 -4.39162046e-01 -1.14298570e+00 -9.38986242e-01
-1.60049096e-01 -2.01828152e-01 4.64252830e-01 7.51849890e-01
7.26720035e-01 7.10446000e-01 4.25950326e-02 -9.83826935e-01
-2.16392830e-01 -6.66037500e-01 -2.95655340e-01 3.40871572e-01
3.69588852e-01 -7.06780016e-01 -2.52087951e-01 3.05613071e-01] | [9.010186195373535, 0.5332000255584717] |
328a94fe-c3eb-4682-82d9-0c1ebc340f99 | target-really-matters-target-aware | null | null | https://aclanthology.org/2022.coling-1.605 | https://aclanthology.org/2022.coling-1.605.pdf | Target Really Matters: Target-aware Contrastive Learning and Consistency Regularization for Few-shot Stance Detection | Stance detection aims to identify the attitude from an opinion towards a certain target. Despite the significant progress on this task, it is extremely time-consuming and budget-unfriendly to collect sufficient high-quality labeled data for every new target under fully-supervised learning, whereas unlabeled data can be collected easier. Therefore, this paper is devoted to few-shot stance detection and investigating how to achieve satisfactory results in semi-supervised settings. As a target-oriented task, the core idea of semi-supervised few-shot stance detection is to make better use of target-relevant information from labeled and unlabeled data. Therefore, we develop a novel target-aware semi-supervised framework. Specifically, we propose a target-aware contrastive learning objective to learn more distinguishable representations for different targets. Such an objective can be easily applied with or without unlabeled data. Furthermore, to thoroughly exploit the unlabeled data and facilitate the model to learn target-relevant stance features in the opinion content, we explore a simple but effective target-aware consistency regularization combined with a self-training strategy. The experimental results demonstrate that our approach can achieve state-of-the-art performance on multiple benchmark datasets in the few-shot setting. | ['Weiping Wang', 'Peng Fu', 'Jiangnan Li', 'Huishan Ji', 'Zheng Lin', 'Rui Liu'] | null | null | null | null | coling-2022-10 | ['stance-detection'] | ['natural-language-processing'] | [ 2.88512081e-01 4.50251028e-02 -7.73234546e-01 -7.76023448e-01
-1.13852429e+00 -2.43701428e-01 6.09666467e-01 1.43624991e-01
-3.05054486e-01 6.90249324e-01 2.32422724e-01 1.99163184e-01
2.45236650e-01 -7.33638048e-01 -4.26176429e-01 -9.07163143e-01
4.32180643e-01 4.92378950e-01 3.07574362e-01 -4.04424965e-01
2.81416416e-01 -2.23162249e-01 -1.59982073e+00 2.15230048e-01
1.18313992e+00 1.08034492e+00 5.82590885e-02 1.59140229e-02
1.49959577e-02 8.84086311e-01 -3.90023112e-01 -4.37288821e-01
9.84526649e-02 -8.81441355e-01 -5.28319240e-01 7.52895892e-01
2.69574851e-01 -2.47806758e-02 2.76172578e-01 1.23833704e+00
4.74354178e-01 9.88824815e-02 8.61947715e-01 -1.01802087e+00
-7.20238805e-01 4.80099440e-01 -7.10289836e-01 2.57611006e-01
-3.54540683e-02 1.10998482e-01 1.30540419e+00 -1.17303693e+00
6.61759913e-01 1.04203475e+00 2.25552872e-01 3.85589749e-01
-9.54764903e-01 -4.60037857e-01 4.10352796e-01 3.90490174e-01
-9.18888032e-01 -4.92827624e-01 1.17449605e+00 -2.84482777e-01
1.29594266e-01 -2.90446077e-02 6.28436387e-01 1.30577135e+00
-5.47991395e-02 1.11933887e+00 1.35955548e+00 -6.49212182e-01
5.36653161e-01 4.77506667e-01 7.44511843e-01 6.36286139e-01
3.43815833e-01 -7.90390372e-02 -5.65492868e-01 -1.81864649e-02
5.97991869e-02 1.65976167e-01 -3.51244897e-01 -7.38654256e-01
-9.84266639e-01 1.20772231e+00 4.87552106e-01 5.14522076e-01
-3.25173706e-01 -3.98755997e-01 5.58275759e-01 4.60705370e-01
1.15103590e+00 3.59014630e-01 -3.32828611e-01 -3.43078040e-02
-9.73753989e-01 7.31991455e-02 7.09343791e-01 6.20490909e-01
8.75138044e-01 2.44100407e-01 -3.51913661e-01 8.46747994e-01
1.50947541e-01 6.27872348e-01 7.04038084e-01 -6.58409476e-01
4.23393965e-01 8.43212008e-01 1.34746522e-01 -7.63443112e-01
-1.08351208e-01 -4.66215611e-01 -6.79887712e-01 5.99620976e-02
4.01542753e-01 -2.17784673e-01 -9.10362363e-01 1.63543808e+00
6.23614848e-01 -1.22057155e-01 9.32397880e-03 1.05489278e+00
7.44411647e-01 5.43522060e-01 -1.34926990e-01 -8.22157443e-01
1.38286126e+00 -1.31975365e+00 -8.34546208e-01 -5.01341343e-01
7.58455992e-01 -6.54031277e-01 1.22802079e+00 1.75669745e-01
-8.09781551e-01 -4.64183420e-01 -1.24013650e+00 2.53068358e-01
-3.29202771e-01 7.92158619e-02 4.44220066e-01 6.07988894e-01
-2.57646650e-01 4.63957161e-01 -5.86138844e-01 -3.34967136e-01
4.97903436e-01 -1.01084910e-01 -5.55531234e-02 -3.12082648e-01
-1.36607587e+00 1.07275212e+00 4.32307124e-01 -9.60172862e-02
-7.93441474e-01 -5.43019414e-01 -1.12852144e+00 -1.90196574e-01
9.75184321e-01 -3.31661403e-01 1.21857738e+00 -1.29875004e+00
-1.32833326e+00 1.04569662e+00 -2.25194916e-01 -4.14495111e-01
3.65016431e-01 -8.78014192e-02 -3.09934676e-01 1.33125201e-01
3.42859507e-01 1.12239145e-01 1.32446289e+00 -1.31473184e+00
-4.97532099e-01 -5.68965197e-01 3.22580859e-02 3.74404371e-01
-8.11568677e-01 9.25955363e-03 -1.88847050e-01 -6.43834293e-01
7.19158724e-02 -8.59919310e-01 3.01253665e-02 -8.79423022e-02
-1.52675867e-01 -3.84471387e-01 9.25863743e-01 -2.59643823e-01
1.07370794e+00 -1.86457956e+00 1.97848886e-01 -3.00896347e-01
1.58009127e-01 4.82922047e-01 3.89333926e-02 8.57742429e-02
1.15427129e-01 -3.82488191e-01 -2.33693302e-01 -2.86760569e-01
-1.58813864e-01 1.12506256e-01 -2.65872329e-01 6.11920118e-01
3.52867782e-01 1.04159927e+00 -1.10302651e+00 -7.11973071e-01
4.81664017e-02 1.74292684e-01 -2.35095114e-01 1.60457909e-01
-3.08453679e-01 4.49738741e-01 -6.64438605e-01 9.57318783e-01
4.24683690e-01 -5.85280716e-01 2.60549486e-01 -3.48427087e-01
5.38566560e-02 1.32901505e-01 -7.27421045e-01 1.35000527e+00
-3.71415883e-01 3.15192401e-01 9.33919381e-03 -1.49323606e+00
1.09493482e+00 1.87726036e-01 3.08013111e-01 -6.69846237e-01
5.78152239e-01 4.26872581e-01 -1.37153298e-01 -4.24787909e-01
3.61428410e-01 -5.70547104e-01 -3.05685967e-01 6.06932938e-01
-3.73112895e-02 -9.55880657e-02 3.93570036e-01 8.65519717e-02
5.59979200e-01 4.23116535e-02 7.43911266e-01 -3.80594909e-01
6.12455010e-01 9.94872451e-02 8.02175879e-01 4.02589023e-01
-5.14473498e-01 4.06379551e-01 2.06648067e-01 -3.15965146e-01
-7.64017820e-01 -6.35406256e-01 8.79069511e-03 1.35786068e+00
2.63425201e-01 -2.58240342e-01 -4.57117528e-01 -1.17708671e+00
-2.99601942e-01 6.61844552e-01 -9.76242900e-01 -3.98225784e-01
-4.97789562e-01 -9.29405868e-01 -2.59692758e-01 5.37967384e-01
6.02422476e-01 -9.32589173e-01 -5.70811808e-01 6.94150701e-02
-3.93436521e-01 -8.25975358e-01 -5.52570343e-01 4.91009474e-01
-8.06635141e-01 -1.06433833e+00 -9.70844507e-01 -7.65681386e-01
6.66407943e-01 7.93285489e-01 1.02646160e+00 -1.89635679e-01
1.21825196e-01 6.52286485e-02 -7.32382596e-01 -4.72377390e-01
-2.99352854e-01 -1.10231645e-01 9.29335281e-02 3.42108816e-01
7.35241175e-01 -2.81291515e-01 -4.30396885e-01 3.83623779e-01
-7.35915065e-01 -2.97132283e-01 4.37232912e-01 1.22521770e+00
5.99339366e-01 -8.35682824e-02 8.09679210e-01 -1.18533432e+00
6.22149825e-01 -4.09585834e-01 -3.65994185e-01 4.16088969e-01
-8.19839954e-01 -8.24070945e-02 7.03918219e-01 -5.53402781e-01
-1.28016710e+00 -2.84974068e-01 2.56483585e-01 -4.58619505e-01
2.98909664e-01 7.20169902e-01 -2.58102179e-01 2.25713402e-01
8.42062593e-01 2.96811342e-01 8.32054242e-02 -1.85782716e-01
4.71561134e-01 7.70104408e-01 2.26482507e-02 -4.77177918e-01
7.70432711e-01 6.64453328e-01 -3.62880379e-01 -9.00396287e-01
-2.03502035e+00 -7.61860788e-01 -7.24646270e-01 -2.64412135e-01
7.33784616e-01 -1.06214178e+00 8.30471218e-02 4.71187770e-01
-6.95891082e-01 -2.08456591e-01 -4.42480922e-01 2.71614343e-01
-5.10099649e-01 6.03995621e-01 -2.77645171e-01 -8.02049637e-01
-5.97443640e-01 -8.63944590e-01 1.02553332e+00 2.34444857e-01
-2.12545156e-01 -1.10764635e+00 3.84833902e-01 7.13683307e-01
1.40114218e-01 5.88860847e-02 5.79973638e-01 -1.03552175e+00
-1.13816090e-01 -4.42142934e-01 -1.70896769e-01 6.95444107e-01
4.30028588e-01 -2.65199184e-01 -8.68540704e-01 -4.29693341e-01
5.96353173e-01 -1.14095366e+00 1.10914314e+00 2.66197413e-01
6.30600452e-01 -2.27237746e-01 -1.04601070e-01 9.08717141e-02
1.23395336e+00 -1.79299042e-01 1.35084406e-01 3.09646100e-01
5.43998301e-01 8.41549218e-01 1.48899209e+00 4.82723564e-01
2.94900924e-01 6.61456585e-01 2.13896021e-01 1.04001939e-01
-5.12961149e-02 1.79911014e-02 3.71522427e-01 1.04119277e+00
-3.28689441e-02 -4.24432978e-02 -5.45001745e-01 6.44370973e-01
-2.12993693e+00 -1.08717155e+00 1.18372045e-01 2.09039783e+00
1.06590521e+00 4.16778564e-01 2.48395696e-01 2.83766985e-01
8.49406779e-01 5.59976697e-01 -7.36792803e-01 2.46263556e-02
-1.25775129e-01 3.40416096e-04 -6.50856197e-02 2.06478536e-01
-1.39648867e+00 9.01956320e-01 5.19997025e+00 9.71993446e-01
-1.11453438e+00 3.52277517e-01 6.37665927e-01 -2.22269833e-01
-1.40799239e-01 -1.52569309e-01 -9.81309891e-01 5.34075677e-01
8.33451450e-01 -4.82971519e-01 -2.24853918e-01 1.24181378e+00
3.14320862e-01 -1.32980213e-01 -9.00798380e-01 7.36960053e-01
6.15191162e-01 -1.10144341e+00 -2.06357002e-01 -8.85329768e-02
1.01433110e+00 -2.39041120e-01 5.77154346e-02 7.36703217e-01
1.72492191e-01 -3.95641804e-01 6.17893875e-01 6.52041053e-03
5.25685668e-01 -6.85384691e-01 7.78858185e-01 5.44124126e-01
-1.00892854e+00 -1.37833267e-01 -5.17845094e-01 -1.68335766e-01
9.28192735e-02 8.90114844e-01 -4.50009137e-01 4.83280689e-01
4.56835449e-01 1.20637012e+00 -4.49924618e-01 6.60776019e-01
-5.99547982e-01 6.36586070e-01 1.83460638e-01 -4.03361171e-01
3.66444409e-01 -1.63097560e-01 3.67781729e-01 7.62385368e-01
2.57499912e-03 1.66886836e-01 5.82718611e-01 5.19686520e-01
-7.40571544e-02 3.57393444e-01 -5.08044183e-01 -7.39406049e-02
2.79545546e-01 1.42787504e+00 -7.35259950e-01 -6.78883612e-01
-4.11624104e-01 7.58534849e-01 5.92668295e-01 8.55922326e-03
-8.87323201e-01 -1.65569354e-02 2.16079235e-01 5.27689271e-02
5.79981804e-01 2.20801204e-01 -2.24130258e-01 -1.55796611e+00
8.24950449e-03 -1.12902999e+00 5.00426531e-01 -4.65979099e-01
-1.65238190e+00 4.11467254e-01 -7.88175166e-02 -1.63074040e+00
-3.06425482e-01 -4.53577310e-01 -7.79807389e-01 4.93604809e-01
-1.70770395e+00 -1.33238566e+00 -2.69690871e-01 5.51451027e-01
1.00277042e+00 2.51405500e-02 6.47842586e-01 -5.45035070e-03
-6.60664558e-01 3.55884492e-01 1.75031304e-01 5.37512898e-02
1.10443437e+00 -1.07026398e+00 -2.45433703e-01 8.68537188e-01
1.98484976e-02 5.13044238e-01 8.22460473e-01 -6.43105447e-01
-1.14363873e+00 -1.21924603e+00 8.41850102e-01 -2.50385612e-01
9.19424474e-01 -9.89332646e-02 -1.06703794e+00 4.42895263e-01
2.03385234e-01 1.04053810e-01 8.95493865e-01 3.32013071e-01
-5.50711691e-01 -1.32691652e-01 -8.51062775e-01 5.59914172e-01
8.40011835e-01 -3.55681926e-01 -1.14189720e+00 5.87220132e-01
5.41245759e-01 -5.86075261e-02 -5.17965019e-01 5.37166536e-01
1.27539158e-01 -9.07132268e-01 6.54051960e-01 -5.79047561e-01
6.56573772e-01 -4.91682962e-02 -1.49261668e-01 -1.48514605e+00
-1.75342441e-01 -3.49488556e-01 -4.70512927e-01 1.05408227e+00
2.86977232e-01 -3.65421742e-01 9.29452419e-01 3.67180556e-01
-1.65199935e-02 -9.40614402e-01 -6.86887920e-01 -8.94838929e-01
2.04617664e-01 -2.30523217e-02 -1.10524371e-01 9.93658960e-01
3.94224286e-01 9.42915797e-01 -8.11401010e-01 -2.76097745e-01
8.49328041e-01 9.35717583e-01 6.94937527e-01 -1.36769712e+00
-2.74684072e-01 -1.05941817e-01 2.41610054e-02 -1.00544286e+00
2.92769074e-01 -6.46294892e-01 3.91539335e-01 -1.39399731e+00
5.75500190e-01 -8.27097446e-02 -2.66278118e-01 3.96326035e-01
-7.06562459e-01 3.64502817e-01 1.09904915e-01 4.54872221e-01
-1.10894930e+00 1.04334545e+00 1.41774321e+00 -4.95073527e-01
-1.05466299e-01 2.64652133e-01 -9.32066441e-01 9.35478449e-01
7.69414306e-01 -4.83639181e-01 -5.93638361e-01 7.62667432e-02
-2.25470308e-02 -8.81706849e-02 1.36658158e-02 -6.10274255e-01
-5.30362651e-02 -2.44850725e-01 -7.64647499e-02 -8.15587282e-01
5.31613231e-01 -5.51716864e-01 -7.03720868e-01 4.87086862e-01
-2.77667701e-01 -6.32400811e-01 -3.99862915e-01 8.50235939e-01
-4.77447718e-01 -4.64432955e-01 1.07188821e+00 -4.05775756e-01
-8.17879736e-01 2.09302649e-01 -1.07513145e-01 4.93696928e-01
1.23034060e+00 1.26842093e-02 -6.29947126e-01 -4.57017303e-01
-6.84120834e-01 2.77335078e-01 5.78374982e-01 3.98093104e-01
5.60681641e-01 -1.37235796e+00 -6.61790848e-01 -8.40228051e-02
7.02267468e-01 -3.89490068e-01 1.92506954e-01 1.03735650e+00
2.78155029e-01 3.78857285e-01 -7.62507766e-02 -6.71661139e-01
-1.36011314e+00 8.43877733e-01 1.52904857e-02 -5.21830440e-01
-3.05164695e-01 7.32664406e-01 2.00462371e-01 -3.50774914e-01
9.44352448e-02 1.14479735e-01 -4.50488359e-01 6.40207231e-01
7.25856721e-01 1.54348984e-01 3.20590921e-02 -5.93181074e-01
-1.74711406e-01 5.47878444e-01 -4.15183604e-01 1.46506861e-01
1.22552466e+00 -2.32764438e-01 1.05406186e-02 9.40015018e-01
1.19715893e+00 -2.08714053e-01 -1.30147803e+00 -6.52014196e-01
-2.04264838e-02 -4.29653913e-01 5.87615259e-02 -4.93039250e-01
-8.85488927e-01 8.92493486e-01 2.79734641e-01 2.65326381e-01
8.80625308e-01 3.96087021e-02 6.79264784e-01 5.13050199e-01
4.80982959e-01 -1.40637851e+00 7.97122657e-01 4.56416696e-01
5.46915293e-01 -1.89394736e+00 1.58615306e-01 -5.43201566e-01
-9.09345686e-01 7.56712615e-01 7.34030247e-01 -5.25641665e-02
6.88995242e-01 -2.62450486e-01 2.84917831e-01 -3.81079078e-01
-7.44604528e-01 -4.65559483e-01 3.98977816e-01 4.08872515e-01
4.24291968e-01 -1.48959041e-01 -6.49744391e-01 4.78555858e-01
3.62760633e-01 -1.98544078e-02 4.62605029e-01 1.23642588e+00
-9.97527182e-01 -1.08268476e+00 -2.04140648e-01 4.92644310e-01
-3.40007216e-01 -8.30319375e-02 -4.67257529e-01 5.25557220e-01
-1.65862456e-01 1.10943604e+00 -4.05646116e-01 -8.48418102e-02
2.30607539e-01 1.65646464e-01 3.47197801e-01 -1.08820891e+00
-2.60504037e-01 2.88406163e-01 2.08131641e-01 -2.83590674e-01
-1.02301252e+00 -6.78718388e-01 -8.57938409e-01 2.48955116e-02
-6.98900521e-01 2.02319399e-01 2.14097500e-01 1.21473396e+00
5.93116041e-03 4.77553874e-01 1.01579916e+00 -7.57998645e-01
-1.14538908e+00 -1.12222731e+00 -5.93794823e-01 6.03315234e-01
-3.59687693e-02 -9.03927624e-01 -6.28454030e-01 -3.89134213e-02] | [11.21363639831543, 6.5255656242370605] |
ad8f80c8-3e58-4de0-b25b-231f470d1307 | learning-interpretable-causal-networks-from | 2303.06423 | null | https://arxiv.org/abs/2303.06423v1 | https://arxiv.org/pdf/2303.06423v1.pdf | Learning interpretable causal networks from very large datasets, application to 400,000 medical records of breast cancer patients | Discovering causal effects is at the core of scientific investigation but remains challenging when only observational data is available. In practice, causal networks are difficult to learn and interpret, and limited to relatively small datasets. We report a more reliable and scalable causal discovery method (iMIIC), based on a general mutual information supremum principle, which greatly improves the precision of inferred causal relations while distinguishing genuine causes from putative and latent causal effects. We showcase iMIIC on synthetic and real-life healthcare data from 396,179 breast cancer patients from the US Surveillance, Epidemiology, and End Results program. More than 90\% of predicted causal effects appear correct, while the remaining unexpected direct and indirect causal effects can be interpreted in terms of diagnostic procedures, therapeutic timing, patient preference or socio-economic disparity. iMIIC's unique capabilities open up new avenues to discover reliable and interpretable causal networks across a range of research fields. | ['Hervé Isambert', 'Anne-Sophie Hamy', 'Liza Hettal', 'Franck Simon', 'Louise Dupuis', 'Vincent Cabeli', 'Honghao Li', 'Marcel da Câmara Ribeiro-Dantas'] | 2023-03-11 | null | null | null | null | ['causal-discovery', 'epidemiology'] | ['knowledge-base', 'medical'] | [ 5.94260216e-01 4.94201183e-01 -9.51743901e-01 -4.56099898e-01
-3.08202088e-01 -4.81247604e-01 6.68986738e-01 6.18004799e-01
-1.26520712e-02 1.25804508e+00 7.08415449e-01 -8.97802889e-01
-8.87333572e-01 -8.23228836e-01 -8.05381596e-01 -5.06145060e-01
-8.29969883e-01 4.35216159e-01 -2.42571041e-01 3.75543386e-01
-7.38223195e-02 2.48726055e-01 -9.27230835e-01 3.46484065e-01
8.83316696e-01 2.98946053e-01 -2.02371314e-01 4.79699880e-01
3.05699766e-01 9.06915188e-01 -2.63207138e-01 -3.25265586e-01
-2.79750377e-01 -6.30806863e-01 -6.72111213e-01 -5.35833776e-01
-1.55042350e-01 -9.91997793e-02 -4.32839602e-01 6.64544344e-01
3.15915197e-01 -6.18036151e-01 8.29643309e-01 -1.38031268e+00
-6.47084594e-01 1.16129458e+00 -4.95929033e-01 1.85075298e-01
4.09150332e-01 1.48066292e-02 1.28076136e+00 -3.46471697e-01
8.60210359e-01 1.37914622e+00 6.60215616e-01 4.43235427e-01
-1.50823283e+00 -9.49747503e-01 2.15607136e-01 1.84372529e-01
-9.53462601e-01 -2.62790531e-01 3.08325261e-01 -3.76556069e-01
3.52727950e-01 6.85607195e-01 4.70943063e-01 1.26335001e+00
4.78689224e-01 5.30500770e-01 1.09432709e+00 -2.59614468e-01
1.87328458e-01 -4.18124139e-01 5.15138321e-02 8.74621272e-01
7.72269249e-01 6.25852585e-01 -4.49213624e-01 -1.05508840e+00
6.97560549e-01 3.67835522e-01 -4.34775651e-01 -3.49515155e-02
-1.64011776e+00 9.17539358e-01 3.42917353e-01 7.11677447e-02
-4.32734191e-01 3.02333742e-01 2.42764622e-01 1.30629703e-01
3.03912342e-01 4.24500972e-01 -7.61854649e-01 2.40491122e-01
-4.15940285e-01 1.28531083e-01 8.49891901e-01 6.92005277e-01
1.45123899e-01 -5.60225368e-01 -7.53692910e-02 3.74628007e-01
2.97400892e-01 7.53247857e-01 -7.25377500e-02 -8.56580734e-01
2.02358723e-01 8.41309369e-01 2.86042187e-02 -1.21917486e+00
-6.86372519e-01 -1.81681618e-01 -1.19655776e+00 -4.58901465e-01
2.33653739e-01 -4.25903410e-01 -5.35611808e-01 1.82478344e+00
3.80684823e-01 3.48491192e-01 -8.43186527e-02 7.34383821e-01
8.41137052e-01 1.11754224e-01 4.72229838e-01 -7.13725388e-01
1.38398695e+00 -6.41549826e-02 -8.51429820e-01 1.11928336e-01
6.40299678e-01 -3.74267459e-01 3.85579228e-01 7.04904273e-02
-6.83603048e-01 2.80638963e-01 -5.47873378e-01 5.89157522e-01
-1.11691197e-02 -1.77727610e-01 1.37329900e+00 3.13798994e-01
-5.57520509e-01 5.00688255e-01 -8.28890681e-01 -2.65919894e-01
6.35992050e-01 3.20273697e-01 -4.67342168e-01 -3.38185042e-01
-1.45457101e+00 4.09718007e-01 2.22376168e-01 -7.07163364e-02
-7.65703201e-01 -1.35260427e+00 -5.98946393e-01 6.51827082e-02
6.39476836e-01 -1.16032863e+00 7.84969449e-01 -4.81752425e-01
-5.88472605e-01 3.66981119e-01 -4.85977739e-01 -4.00449216e-01
3.17215830e-01 2.13685364e-01 -6.06862605e-01 -7.46537894e-02
3.22616696e-01 1.39316693e-01 6.97098896e-02 -1.02655828e+00
-7.08445191e-01 -5.70863724e-01 -4.47539359e-01 -1.01126939e-01
1.10482052e-02 8.08529258e-02 1.46696195e-01 -7.76817739e-01
-7.66354753e-03 -8.83400679e-01 -6.48020267e-01 -1.41690582e-01
-8.36549163e-01 -3.13579470e-01 5.53734958e-01 -2.89716184e-01
1.29669344e+00 -1.74620008e+00 3.72390561e-02 2.52223700e-01
5.90669751e-01 -2.96932489e-01 9.44991931e-02 5.51322281e-01
-6.43395662e-01 5.77012062e-01 -3.97305131e-01 4.32403475e-01
-5.33721387e-01 3.32346588e-01 -1.76640719e-01 6.00562453e-01
2.90237308e-01 9.50673640e-01 -1.21034622e+00 -4.29468572e-01
1.12714209e-01 1.38744656e-02 -5.90070486e-01 -1.71317887e-02
-1.42231584e-01 5.72001100e-01 -7.99327016e-01 5.26524246e-01
2.31233612e-01 -6.99221969e-01 8.81506920e-01 2.70601571e-01
-1.20284624e-01 3.25561434e-01 -9.13346648e-01 9.86545265e-01
-6.97332993e-02 4.56249952e-01 -1.59258053e-01 -1.12591124e+00
3.82442117e-01 6.13739908e-01 7.34774351e-01 -3.07886600e-01
1.39784114e-02 -8.17291066e-02 2.74380475e-01 -6.26363575e-01
-4.59036469e-01 -2.62668997e-01 -7.74317309e-02 5.10434508e-01
-4.71776336e-01 4.88884062e-01 -2.61359569e-02 4.13556665e-01
1.71199667e+00 -8.02385569e-01 8.01278949e-01 -4.67887223e-01
-1.97249368e-01 1.99104548e-01 1.01205897e+00 8.74435484e-01
1.08506195e-01 2.25650862e-01 8.98370862e-01 -5.01393199e-01
-7.50326455e-01 -1.16305506e+00 -4.26713556e-01 4.25006300e-01
-2.37501144e-01 -2.66161203e-01 -4.30156924e-02 -7.98113346e-01
3.73931795e-01 7.08078027e-01 -1.09930110e+00 -2.76827812e-01
-2.56339937e-01 -1.39257812e+00 5.67597866e-01 2.62349933e-01
-1.06754061e-02 -7.29602456e-01 -2.24467710e-01 3.40410620e-01
-3.96539241e-01 -8.03598285e-01 -2.19830483e-01 6.06316812e-02
-9.62095857e-01 -1.83919168e+00 -2.29145616e-01 -1.30029500e-01
7.28001654e-01 2.96855215e-02 1.17458749e+00 1.39828607e-01
-6.16073430e-01 -1.59146637e-01 -1.96285531e-01 -5.89636087e-01
-4.79804724e-01 -3.77763778e-01 2.87686110e-01 -2.84824252e-01
1.50148720e-01 -5.33234715e-01 -8.07315886e-01 2.93817192e-01
-5.73497355e-01 1.71702340e-01 7.71391153e-01 9.72664297e-01
4.18738931e-01 3.77796859e-01 8.62618029e-01 -1.45023549e+00
4.14803118e-01 -1.06251109e+00 -5.64008772e-01 3.42280030e-01
-1.16656041e+00 7.84476325e-02 3.35449785e-01 -3.87831569e-01
-1.23320472e+00 1.99309979e-02 3.50257486e-01 1.54444262e-01
-1.74118280e-01 8.84441018e-01 1.21279763e-05 6.00432694e-01
6.47471964e-01 -3.37813526e-01 2.38797329e-02 -1.84096992e-01
2.81299382e-01 4.22430456e-01 2.94571757e-01 -3.15030068e-01
3.18119317e-01 6.72074556e-01 3.31805825e-01 -3.95621300e-01
-8.41501653e-01 -3.53404850e-01 -2.00976148e-01 1.09703667e-01
5.98255932e-01 -8.07650506e-01 -1.44206929e+00 -2.30239928e-01
-1.04579210e+00 -2.44348451e-01 6.89523369e-02 8.21749210e-01
-1.46332189e-01 -1.19048141e-01 -5.56936502e-01 -7.65993953e-01
-1.39860287e-01 -6.91111505e-01 8.22086692e-01 -4.59499061e-02
-5.50468862e-01 -1.31581235e+00 2.18776748e-01 1.36255488e-01
-2.23508388e-01 6.03628397e-01 1.54178083e+00 -2.81829745e-01
-4.10291970e-01 -2.13197306e-01 -5.42824566e-01 -5.53766668e-01
6.54640496e-01 3.96277845e-01 -3.36131185e-01 1.27534851e-01
-6.06091022e-01 1.96839005e-01 8.62094104e-01 9.15104926e-01
1.29156709e+00 -8.26350868e-01 -1.11403978e+00 2.39598647e-01
1.12083077e+00 3.40760797e-01 3.99650663e-01 -3.38455111e-01
6.48297131e-01 7.54850984e-01 3.60603660e-01 5.86327255e-01
4.83694792e-01 4.02195901e-01 4.04298395e-01 -4.35652494e-01
1.34620070e-01 -4.55688000e-01 -2.43997440e-01 1.28281653e-01
-4.10201669e-01 -3.00659388e-01 -1.03407419e+00 5.60227275e-01
-2.01725411e+00 -1.11258936e+00 -8.82648766e-01 2.13075900e+00
1.18652022e+00 -6.43560588e-02 -7.03693926e-03 -6.60506031e-03
7.45939732e-01 -5.52339315e-01 -5.42311072e-01 -1.08777188e-01
-2.79767275e-01 -5.68581112e-02 9.32460606e-01 2.96521991e-01
-7.46363938e-01 4.28851724e-01 7.76443958e+00 4.61185187e-01
-7.01539755e-01 -1.86912082e-02 1.05232513e+00 -4.12831530e-02
-6.47822797e-01 1.44846082e-01 -3.05390179e-01 3.09594303e-01
1.05293334e+00 -1.14243872e-01 1.28065065e-01 4.02524620e-01
9.17390227e-01 -1.59878701e-01 -1.23280120e+00 3.34960014e-01
-6.63726389e-01 -1.75619876e+00 -1.39279529e-01 3.48651528e-01
9.30600643e-01 -1.85019165e-01 -1.98433325e-01 -3.51495832e-01
1.26585937e+00 -1.19034541e+00 1.68519348e-01 7.32173979e-01
1.12797439e+00 -6.12502158e-01 6.88335240e-01 2.03352556e-01
-5.37873745e-01 -2.60346293e-01 -4.04136926e-02 -3.48690152e-01
1.64386719e-01 1.43390489e+00 -1.25142276e+00 6.25555992e-01
7.71363914e-01 8.96996081e-01 -5.36330231e-02 5.61551392e-01
-4.05576497e-01 1.28163409e+00 -5.94029762e-02 -1.76120356e-01
-2.22357422e-01 1.97598711e-01 4.08647239e-01 9.93745506e-01
6.95175231e-02 6.96955860e-01 -2.31885731e-01 8.16144049e-01
-1.86468422e-01 -2.49478385e-01 -6.89103603e-01 -3.08859438e-01
5.73839247e-01 8.86469305e-01 -5.44359624e-01 -2.55514979e-01
-2.78908104e-01 3.66904318e-01 1.69459712e-02 2.70592272e-01
-7.51582921e-01 1.50036946e-01 6.01443589e-01 1.38556167e-01
-4.75992620e-01 2.37621039e-01 -6.95255935e-01 -9.51640546e-01
-4.93058741e-01 -7.65322089e-01 9.57167685e-01 -3.49002302e-01
-1.44506836e+00 -1.23429351e-01 1.70197219e-01 -6.82554245e-01
-3.13982069e-01 -3.40464503e-01 -4.56731141e-01 7.23307371e-01
-1.08309710e+00 -7.18657672e-01 9.12592858e-02 4.22748655e-01
1.08986802e-01 9.28616002e-02 1.02116239e+00 1.54043183e-01
-8.31098557e-01 2.80606240e-01 8.03055838e-02 9.66530852e-03
6.90939665e-01 -1.18291354e+00 1.65028527e-01 3.18018198e-01
-3.91511887e-01 8.10768604e-01 7.22206533e-01 -1.09353912e+00
-1.25487328e+00 -1.20988452e+00 1.25129735e+00 -5.24822831e-01
1.10207546e+00 -2.10708424e-01 -5.64344168e-01 6.51367903e-01
-1.08226627e-01 -4.34733592e-02 1.00172412e+00 8.21838200e-01
-4.83522505e-01 2.92490162e-02 -9.50986862e-01 7.75383711e-01
1.36201215e+00 -1.59426555e-02 -3.85401189e-01 5.18327892e-01
8.71324420e-01 2.85022199e-01 -1.03510606e+00 7.79026210e-01
5.66510558e-01 -6.73507512e-01 1.03650904e+00 -1.08381975e+00
9.04806197e-01 2.38885498e-03 2.20903113e-01 -1.24112439e+00
-4.97642070e-01 -5.00733495e-01 2.21338093e-01 8.62918437e-01
9.38017368e-01 -7.03463078e-01 5.74334145e-01 8.00064266e-01
2.42428795e-01 -7.23878384e-01 -6.53592229e-01 -4.00232643e-01
-2.15117931e-01 -3.95680279e-01 5.84177613e-01 1.58894062e+00
3.92194301e-01 3.66393089e-01 -3.50976557e-01 3.42805654e-01
9.33078408e-01 4.04867381e-01 2.11169615e-01 -1.56829417e+00
-3.00709963e-01 -1.56503811e-01 -8.52406174e-02 -1.31613538e-01
-5.65011869e-04 -6.39835417e-01 -2.87599117e-01 -1.51973593e+00
9.06768501e-01 -1.02318382e+00 -1.45518377e-01 7.23834634e-01
-5.44975340e-01 -2.36055449e-01 -5.14556229e-01 1.49932936e-01
-1.67599291e-01 2.72345334e-01 1.20838523e+00 -2.39489183e-01
-9.67970416e-02 2.79500306e-01 -8.57930064e-01 7.44918168e-01
6.35947108e-01 -7.92939007e-01 -4.60052878e-01 -5.75774238e-02
4.70210999e-01 7.93599069e-01 7.49557495e-01 8.13800842e-02
8.45976993e-02 -8.36003244e-01 4.57588196e-01 -1.58015549e-01
-4.13766086e-01 -5.90348005e-01 7.71755576e-01 1.02462685e+00
-7.77752221e-01 -1.44770339e-01 -3.61698419e-02 1.03112054e+00
1.16717249e-01 4.34897065e-01 1.66723371e-01 -7.30649680e-02
-1.83870494e-01 1.51358530e-01 -3.00169826e-01 -1.75101876e-01
9.03397083e-01 4.44751471e-01 -7.46826649e-01 -2.63025612e-01
-7.50817895e-01 3.09404284e-01 -1.25784114e-01 3.47723871e-01
6.02843761e-01 -1.23821127e+00 -1.16470718e+00 -3.31332266e-01
2.60907114e-01 -1.90914690e-01 3.76472443e-01 1.05086279e+00
1.03648596e-01 6.53340101e-01 3.91637266e-01 -2.74898916e-01
-1.32077456e+00 7.65136540e-01 -5.31106703e-02 -3.35994393e-01
-5.05472898e-01 5.09064138e-01 5.59096634e-01 -3.82513940e-01
-1.87456757e-01 -4.00995202e-02 2.83430121e-03 -1.68400303e-01
5.53243637e-01 5.03352582e-01 -2.54036278e-01 -1.04837092e-02
-5.95896125e-01 -2.23416418e-01 1.98490664e-01 1.98908493e-01
1.53635907e+00 -4.04611509e-03 -5.85014045e-01 3.43893826e-01
1.03938353e+00 -2.82285698e-02 -7.76158452e-01 -6.71650544e-02
1.58903256e-01 -2.18216226e-01 -8.27484131e-02 -1.12650728e+00
-8.26571584e-01 2.21653923e-01 1.75746262e-01 2.89557099e-01
1.06489909e+00 4.30313230e-01 -2.05859207e-02 -1.84837319e-02
2.48993337e-01 -3.42719227e-01 -4.66336548e-01 -1.79900303e-01
8.60086441e-01 -1.34508610e+00 1.41587764e-01 -9.01988566e-01
-1.64694682e-01 7.07425833e-01 -5.92208654e-02 1.68328851e-01
7.59736955e-01 3.51765603e-01 -3.48074973e-01 -4.96835649e-01
-1.38432252e+00 2.86093503e-01 3.35263073e-01 3.86770397e-01
7.46668220e-01 7.73959041e-01 -7.76586413e-01 6.69761837e-01
8.16027969e-02 2.41647195e-02 4.00766879e-01 3.38809133e-01
2.21092224e-01 -9.36118662e-01 -3.24005395e-01 1.04487276e+00
-6.59377038e-01 -3.51033747e-01 -5.49645185e-01 8.13112259e-01
7.62669295e-02 1.18504012e+00 1.39465081e-02 -9.73232836e-02
2.21455604e-01 -4.64446872e-01 6.91734701e-02 -3.70131165e-01
2.50316579e-02 1.05280150e-02 4.99799341e-01 -6.41690612e-01
-5.82452178e-01 -1.04054296e+00 -1.38339376e+00 -6.81222856e-01
-2.76314378e-01 1.66741237e-01 3.11487734e-01 1.06278145e+00
6.01460934e-01 7.78691530e-01 7.00478017e-01 1.91030741e-01
-1.86137930e-02 -5.79307258e-01 -3.73346955e-01 2.25913271e-01
4.15110350e-01 -6.49206221e-01 -3.23735982e-01 2.45447502e-01] | [7.881875514984131, 5.405465126037598] |
caa81ad4-688e-4115-bef7-ed69241f295e | gates-are-not-what-you-need-in-rnns | 2108.00527 | null | https://arxiv.org/abs/2108.00527v1 | https://arxiv.org/pdf/2108.00527v1.pdf | Gates are not what you need in RNNs | Recurrent neural networks have flourished in many areas. Consequently, we can see new RNN cells being developed continuously, usually by creating or using gates in a new, original way. But what if we told you that gates in RNNs are redundant? In this paper, we propose a new recurrent cell called Residual Recurrent Unit (RRU) which beats traditional cells and does not employ a single gate. It is based on the residual shortcut connection together with linear transformations, ReLU, and normalization. To evaluate our cell's effectiveness, we compare its performance against the widely-used GRU and LSTM cells and the recently proposed Mogrifier LSTM on several tasks including, polyphonic music modeling, language modeling, and sentiment analysis. Our experiments show that RRU outperforms the traditional gated units on most of these tasks. Also, it has better robustness to parameter selection, allowing immediate application in new tasks without much tuning. We have implemented the RRU in TensorFlow, and the code is made available at https://github.com/LUMII-Syslab/RRU . | ['Karlis Freivalds', 'Emils Ozolins', 'Eliza Gaile', 'Andis Draguns', 'Ronalds Zakovskis'] | 2021-08-01 | null | null | null | null | ['music-modeling'] | ['music'] | [-8.12690880e-04 -1.81220531e-01 6.64169863e-02 -2.33453676e-01
-2.98516184e-01 -3.92777830e-01 3.94427031e-01 -4.36207056e-01
-5.78701854e-01 6.84526742e-01 3.67580116e-01 -4.50743556e-01
4.49941546e-01 -6.27213657e-01 -6.08599782e-01 -7.24653125e-01
3.72046858e-01 -6.12617582e-02 1.16448037e-01 -3.88523072e-01
1.14720128e-01 3.57400656e-01 -1.18492806e+00 4.48200315e-01
4.76993531e-01 8.08772385e-01 3.72758985e-01 4.91600722e-01
-9.55669805e-02 1.22129309e+00 -5.88334322e-01 -1.20787427e-01
5.42217568e-02 -5.51761448e-01 -5.48833728e-01 -4.28973436e-01
-1.59730706e-02 3.96981463e-02 -5.80262959e-01 8.17827046e-01
7.98095167e-01 5.11145949e-01 1.42365992e-01 -5.93039870e-01
-8.37951541e-01 1.23649561e+00 -4.25611019e-01 3.72498423e-01
3.15170102e-02 -1.76770821e-01 1.08870769e+00 -1.09195018e+00
4.66490924e-01 1.21864903e+00 8.13202024e-01 7.35192299e-01
-1.01112294e+00 -9.47558105e-01 2.27362946e-01 -5.04739806e-02
-1.42898583e+00 -5.58476090e-01 5.92697203e-01 -4.73772250e-02
1.32397377e+00 2.07959324e-01 4.86580878e-01 1.44927955e+00
2.44383410e-01 1.05096745e+00 8.71643782e-01 -3.93459171e-01
-6.28606230e-02 -1.34034961e-01 2.87050366e-01 6.05512381e-01
-1.39332321e-02 -9.45832208e-02 -4.50028062e-01 2.10916653e-01
8.87146413e-01 3.74889672e-01 -4.03712869e-01 4.11771119e-01
-1.38358891e+00 7.33526170e-01 5.46419799e-01 8.04117918e-01
-1.69132367e-01 6.42918050e-01 5.23786366e-01 4.11258280e-01
5.24426281e-01 4.05017048e-01 -4.31269109e-01 -1.59382209e-01
-7.60237992e-01 -1.93692923e-01 6.52512252e-01 8.65091741e-01
5.87923825e-01 6.38402879e-01 -2.90082693e-01 1.17351556e+00
6.88434243e-02 2.02861786e-01 1.22350979e+00 -6.75670803e-01
2.15649635e-01 4.75807488e-01 -4.86374378e-01 -7.29599655e-01
-5.33704638e-01 -8.39203179e-01 -1.08521724e+00 -3.17216575e-01
-8.85405317e-02 -3.45082700e-01 -1.18368256e+00 1.63566709e+00
-3.29008996e-01 4.56043035e-01 -3.84488259e-03 6.46775663e-01
1.21438050e+00 8.43582451e-01 -2.90995449e-01 -6.28148094e-02
9.52689946e-01 -1.21159220e+00 -9.04953301e-01 -2.80344754e-01
8.51658762e-01 -8.05245519e-01 1.03824341e+00 4.71552253e-01
-9.56510484e-01 -7.34731853e-01 -9.32420194e-01 -2.47579351e-01
-4.53250706e-01 4.47513551e-01 7.72436500e-01 5.03677070e-01
-1.29214883e+00 6.64876759e-01 -8.72820973e-01 -3.85044873e-01
7.99761415e-02 5.03598213e-01 -2.06279501e-01 2.48463288e-01
-1.40735209e+00 7.80675828e-01 4.27347749e-01 5.99885941e-01
-5.80802321e-01 -2.56868780e-01 -8.18398416e-01 1.04714572e-01
2.29696631e-01 -5.85986495e-01 1.33404112e+00 -8.49543273e-01
-1.88486040e+00 4.73496407e-01 -3.08521718e-01 -7.24711716e-01
1.60335198e-01 -2.70656943e-01 -4.82719839e-01 -4.84462768e-01
-4.02682424e-01 8.08942139e-01 6.30430937e-01 -7.50829458e-01
-4.20420647e-01 -1.08333215e-01 -9.03419331e-02 9.42925066e-02
-3.58671278e-01 5.55908829e-02 -3.64792198e-01 -1.11940956e+00
2.77666688e-01 -1.08514643e+00 -3.17603767e-01 -9.82009828e-01
-4.46181923e-01 -2.78799057e-01 6.87343121e-01 -4.78689373e-01
1.50507784e+00 -2.35526323e+00 5.15888445e-02 1.09463431e-01
1.10334545e-01 4.40241218e-01 -2.15477422e-01 3.00580740e-01
-1.98933661e-01 3.22044611e-01 -1.10243060e-01 -4.94505495e-01
-2.48631969e-01 2.44612843e-01 -5.60727656e-01 1.46311775e-01
-1.76016539e-01 1.16342247e+00 -7.31519878e-01 2.60691233e-02
-5.38186133e-02 7.25503683e-01 -3.10359776e-01 -2.15725556e-01
2.68363822e-02 2.87024260e-01 -2.03728765e-01 6.92281425e-01
1.85006127e-01 -6.40557855e-02 2.54992004e-02 -2.13757548e-02
-3.37542981e-01 6.30125880e-01 -8.91439974e-01 1.70377290e+00
-5.95413089e-01 9.88833964e-01 -4.18411970e-01 -8.07289720e-01
1.38845551e+00 5.18576086e-01 -3.25934589e-02 -7.65142918e-01
4.80945438e-01 4.52045172e-01 2.04246417e-01 -4.18707915e-02
7.30241776e-01 -2.45717987e-02 1.41894715e-02 4.81248260e-01
2.17589527e-01 2.26995200e-01 2.67101020e-01 5.53071275e-02
1.15834594e+00 3.05079781e-02 5.97104952e-02 -3.86345349e-02
3.88570100e-01 -6.42412007e-01 8.61437500e-01 7.65000403e-01
1.06322683e-01 8.56325567e-01 3.50883693e-01 -3.96107435e-01
-5.93903422e-01 -7.76411235e-01 5.60317282e-03 1.25242424e+00
-3.30962718e-01 -4.61208880e-01 -4.53098387e-01 -2.67913282e-01
-3.16404164e-01 6.70980155e-01 -6.22637272e-01 -1.65360138e-01
-7.99017489e-01 -9.51070189e-01 8.72414947e-01 7.11927116e-01
5.98394990e-01 -1.63333774e+00 -3.30287099e-01 3.00910532e-01
-2.34470800e-01 -1.02830970e+00 -6.71663404e-01 5.55211186e-01
-1.13513660e+00 -3.56983364e-01 -8.58742058e-01 -8.52491558e-01
5.16821444e-01 4.78682518e-01 9.00695801e-01 -2.75470056e-02
-5.06549329e-02 -9.05458406e-02 -6.10546887e-01 -3.22492391e-01
-1.22445546e-01 6.95746005e-01 8.27651769e-02 8.04974139e-03
3.31936419e-01 -6.86437070e-01 -3.70347559e-01 2.66893417e-01
-8.19616377e-01 3.10609695e-02 7.05312133e-01 8.55771065e-01
6.12567306e-01 -1.83972791e-01 8.01079273e-01 -1.19680679e+00
8.20518970e-01 -2.90246546e-01 -2.74472833e-01 -8.38119350e-03
-3.07381421e-01 3.96769717e-02 5.70145071e-01 -4.73832518e-01
-9.63548779e-01 -1.36816666e-01 -2.59776711e-01 -5.57356238e-01
2.14677453e-01 7.28864431e-01 1.70721784e-01 1.42693415e-01
6.19044781e-01 1.41180024e-01 -3.33560407e-01 -6.39577031e-01
2.48420104e-01 5.97403288e-01 4.31809336e-01 -1.03399381e-01
3.79477650e-01 3.02932262e-01 -6.50522947e-01 -8.58695388e-01
-6.94184422e-01 -3.45145255e-01 -3.86163890e-01 -3.78610939e-02
5.82711399e-01 -9.82828081e-01 -5.38303852e-01 5.93549907e-01
-1.25915921e+00 -5.54442704e-01 -2.83457398e-01 7.43751228e-01
-1.48464888e-01 -3.95356230e-02 -1.04938483e+00 -7.05859184e-01
-6.18963599e-01 -1.04135418e+00 4.75210607e-01 3.97499561e-01
-2.46482581e-01 -1.02119875e+00 2.00809985e-01 1.26538038e-01
6.90422595e-01 -3.96160185e-02 5.18613279e-01 -6.42122269e-01
-3.08326215e-01 -1.83163732e-01 -2.09318474e-02 7.31266797e-01
2.12433934e-01 1.13791630e-01 -1.20452642e+00 -3.35494936e-01
6.99137300e-02 -6.13652766e-02 1.63453650e+00 5.02977490e-01
1.20111299e+00 -5.70873991e-02 -1.85577974e-01 8.08002830e-01
1.12278581e+00 5.54687440e-01 9.51129496e-01 2.14549005e-01
9.40341771e-01 2.47255400e-01 2.03336366e-02 1.37789696e-01
1.91814631e-01 2.58688807e-01 1.38294831e-01 -3.23513359e-01
-4.69043702e-02 -1.13095678e-01 8.63233864e-01 1.66887021e+00
-3.91322732e-01 -3.60745788e-01 -7.90545225e-01 3.83216441e-01
-1.98107910e+00 -9.81722057e-01 -3.67817469e-02 1.96290004e+00
7.04765797e-01 3.55445385e-01 -1.45416766e-01 8.31066370e-02
7.36547709e-01 2.26063877e-01 -4.75803941e-01 -7.53180981e-01
-3.86565208e-01 4.80879396e-01 6.03457570e-01 4.00721967e-01
-8.77788067e-01 1.32626069e+00 6.18592787e+00 8.29263628e-01
-1.54180527e+00 2.48108879e-01 7.95933187e-01 -3.32049280e-01
-2.54745811e-01 -1.51706278e-01 -1.03170848e+00 2.08752722e-01
9.37290847e-01 1.03421591e-01 5.20991564e-01 5.16424775e-01
2.88188696e-01 1.56443238e-01 -8.07031214e-01 9.56740856e-01
8.25536028e-02 -1.34900248e+00 5.47151454e-02 -1.70701817e-01
7.27409244e-01 6.57379985e-01 1.63522080e-01 7.05947816e-01
3.96835536e-01 -1.21868479e+00 5.60712874e-01 5.20612657e-01
5.96006215e-01 -8.88052821e-01 9.97553647e-01 2.05238853e-02
-1.28561938e+00 -4.83162552e-02 -4.53229338e-01 -2.20853016e-01
-1.53492078e-01 6.42803907e-01 -6.42984509e-01 3.49352986e-01
6.46331370e-01 1.04155874e+00 -4.90429491e-01 8.12580228e-01
-3.72515529e-01 7.36218572e-01 -1.21211320e-01 -1.02986932e-01
3.68117481e-01 -2.23222777e-01 3.39169919e-01 1.50918901e+00
5.34376323e-01 -9.97825190e-02 -2.75423408e-01 7.14911044e-01
-4.26887810e-01 6.43145368e-02 -5.91282129e-01 -1.26931548e-01
3.33675742e-01 1.39523721e+00 -9.36516941e-01 -1.38506159e-01
-2.63108045e-01 8.52268696e-01 1.49966225e-01 6.81461692e-01
-9.11405504e-01 -7.44580746e-01 4.68810320e-01 -1.56822041e-01
3.19927454e-01 -1.35709807e-01 -2.65246630e-01 -1.31508207e+00
-5.79465665e-02 -7.58377194e-01 2.01198071e-01 -7.43869543e-01
-8.73565793e-01 1.12894511e+00 -4.97634739e-01 -1.03819001e+00
-2.13532031e-01 -6.05666697e-01 -7.54321754e-01 6.95550859e-01
-1.38206422e+00 -7.80896068e-01 -3.93225141e-02 4.68105823e-01
5.79778492e-01 -3.59569281e-01 6.84928179e-01 5.10550261e-01
-1.09956801e+00 7.26862550e-01 8.94289911e-02 3.79546821e-01
7.32097149e-01 -9.64986622e-01 7.78299809e-01 7.84256935e-01
4.07013267e-01 9.42791581e-01 4.67427492e-01 -3.30915272e-01
-1.04008436e+00 -1.16751719e+00 1.00707459e+00 -2.33066920e-02
6.65939331e-01 -6.13530159e-01 -1.05126202e+00 1.11558306e+00
2.86786735e-01 -2.12046318e-02 6.23171389e-01 2.47213468e-01
-2.93504000e-01 -2.48883203e-01 -5.14345288e-01 7.58122206e-01
1.07656658e+00 -5.19811869e-01 -5.17913163e-01 1.40592400e-02
9.55514431e-01 -3.69345367e-01 -3.89793843e-01 5.19262850e-01
5.24170041e-01 -8.40023637e-01 6.40016258e-01 -2.15277821e-01
9.83768404e-02 -3.45191956e-01 -2.29719758e-01 -1.41254878e+00
-4.44670260e-01 -7.83862770e-01 2.68485397e-02 1.21269321e+00
7.44461894e-01 -1.02725637e+00 5.05971909e-01 -5.34945540e-02
-5.06488919e-01 -8.90160203e-01 -8.28137338e-01 -8.22644055e-01
-1.61904365e-01 -5.12988389e-01 3.63658905e-01 9.81849372e-01
2.39255130e-02 6.54452145e-01 -4.67518151e-01 -3.25485408e-01
-6.71664551e-02 -1.42596677e-01 5.16255498e-01 -9.74198461e-01
-3.79729450e-01 -7.27400243e-01 -1.48762479e-01 -1.06582046e+00
1.09976910e-01 -1.14223003e+00 -5.44645861e-02 -1.54628909e+00
-8.74816999e-02 -2.57736534e-01 -8.89983714e-01 8.85703921e-01
3.32134329e-02 5.00470519e-01 2.80739069e-01 1.91231012e-01
-4.49109852e-01 6.84850752e-01 1.03317976e+00 -7.15645850e-02
-6.08035743e-01 2.04155743e-02 -8.13428104e-01 8.01529229e-01
1.14701104e+00 -4.16739136e-01 -2.41450921e-01 -5.02264678e-01
4.77539927e-01 -3.13129097e-01 -8.75886753e-02 -1.13507140e+00
2.92330325e-01 4.23816204e-01 2.39037529e-01 -6.97891116e-01
3.40016454e-01 -1.48193389e-01 1.29786134e-01 5.50305009e-01
-3.45967650e-01 3.42910081e-01 2.57577032e-01 8.13844875e-02
-4.45582896e-01 -1.39119774e-01 6.03857279e-01 -2.14477181e-01
-6.11234844e-01 1.82006568e-01 -4.92347389e-01 -1.67933270e-01
5.03466308e-01 -2.28458032e-01 -3.36552501e-01 -3.41812998e-01
-8.17892969e-01 1.40554532e-01 1.82582322e-03 6.15089595e-01
7.84083843e-01 -1.31394780e+00 -7.32000589e-01 2.16905922e-01
-2.52891988e-01 -1.10890567e-01 5.08488566e-02 8.12725902e-01
-4.52593386e-01 6.36443377e-01 1.87410302e-02 -3.01466942e-01
-1.11472380e+00 1.81688666e-01 4.72832739e-01 -3.60976756e-01
-6.58434451e-01 1.06651914e+00 4.11403328e-01 -6.76154375e-01
4.01144862e-01 -8.86515260e-01 -3.91351014e-01 4.07035351e-01
4.73177284e-01 3.22212785e-01 2.83786446e-01 -3.69346976e-01
-1.79860011e-01 3.04096133e-01 -3.39951992e-01 -6.04191683e-02
1.31350327e+00 6.99379146e-02 -3.14892232e-01 1.14957333e+00
1.01396561e+00 -4.36836518e-02 -7.49601066e-01 -3.67743492e-01
-4.94181439e-02 1.64931387e-01 1.41830593e-01 -5.31877756e-01
-1.56693947e+00 9.29696977e-01 3.34667772e-01 1.36512026e-01
1.15730357e+00 -3.32660377e-01 1.10381496e+00 5.92506409e-01
2.35909820e-01 -1.05767393e+00 5.91071835e-03 1.22137570e+00
8.47035348e-01 -7.64454782e-01 -3.09005529e-01 9.90258530e-02
-5.79090238e-01 1.15195525e+00 5.17359257e-01 -2.24966705e-01
6.15526736e-01 2.94109792e-01 2.33730420e-01 1.01555362e-01
-1.10886800e+00 -2.75884539e-01 2.83316593e-03 -4.61909696e-02
1.03881145e+00 5.96517548e-02 -2.23562509e-01 4.68975514e-01
-4.33933944e-01 5.98483533e-02 6.45022392e-01 7.02908218e-01
-4.08456773e-01 -1.12317860e+00 -1.76900461e-01 4.54282582e-01
-6.24166310e-01 -5.39885581e-01 -2.58387804e-01 6.31614327e-01
3.01331338e-02 7.24843025e-01 1.81022108e-01 -7.58594096e-01
2.69764245e-01 4.31795381e-02 2.30058968e-01 -7.99604893e-01
-9.61709559e-01 4.02355939e-01 -1.10659912e-01 -3.42621058e-01
-2.98836440e-01 -4.82304484e-01 -1.55194175e+00 -2.75765389e-01
-4.29105341e-01 1.68004081e-01 6.22771382e-01 8.49153936e-01
3.43090713e-01 1.02902687e+00 5.58128059e-01 -8.46766353e-01
-5.61799332e-02 -1.27050948e+00 -4.80031133e-01 -2.43040845e-01
2.98909932e-01 -3.58532012e-01 -3.50719631e-01 -2.59934008e-01] | [10.88154125213623, 6.362493991851807] |
15034821-ae78-4b06-aa49-111bcadf96c1 | detection-of-uncertainty-in-exceedance-of | 2303.10291 | null | https://arxiv.org/abs/2303.10291v1 | https://arxiv.org/pdf/2303.10291v1.pdf | Detection of Uncertainty in Exceedance of Threshold (DUET): An Adversarial Patch Localizer | Development of defenses against physical world attacks such as adversarial patches is gaining traction within the research community. We contribute to the field of adversarial patch detection by introducing an uncertainty-based adversarial patch localizer which localizes adversarial patch on an image, permitting post-processing patch-avoidance or patch-reconstruction. We quantify our prediction uncertainties with the development of \textit{\textbf{D}etection of \textbf{U}ncertainties in the \textbf{E}xceedance of \textbf{T}hreshold} (DUET) algorithm. This algorithm provides a framework to ascertain confidence in the adversarial patch localization, which is essential for safety-sensitive applications such as self-driving cars and medical imaging. We conducted experiments on localizing adversarial patches and found our proposed DUET model outperforms baseline models. We then conduct further analyses on our choice of model priors and the adoption of Bayesian Neural Networks in different layers within our model architecture. We found that isometric gaussian priors in Bayesian Neural Networks are suitable for patch localization tasks and the presence of Bayesian layers in the earlier neural network blocks facilitates top-end localization performance, while Bayesian layers added in the later neural network blocks contribute to better model generalization. We then propose two different well-performing models to tackle different use cases. | ['Jun Zhao', 'Wenhan Yu', 'Terence Jie Chua'] | 2023-03-18 | null | null | null | null | ['self-driving-cars'] | ['computer-vision'] | [ 2.69363135e-01 3.19161147e-01 1.51042953e-01 -1.08435243e-01
-1.05839670e+00 -9.32833374e-01 6.14918113e-01 -1.51608050e-01
-1.24432303e-01 3.49494278e-01 -8.03416595e-02 -4.33225065e-01
-2.31628552e-01 -7.31099844e-01 -1.27785766e+00 -7.93957293e-01
-2.92121887e-01 7.19961002e-02 5.47916174e-01 -1.37116149e-01
1.24376245e-01 7.40776718e-01 -8.22621882e-01 3.23215038e-01
5.01476347e-01 1.32302403e+00 -9.90595371e-02 6.46902323e-01
7.95286715e-01 8.05376351e-01 -8.62147450e-01 -6.04732215e-01
7.72839189e-01 2.35915497e-01 -1.83090582e-01 -5.33002973e-01
5.88329971e-01 -3.62409711e-01 -8.01100492e-01 1.11709166e+00
6.54656768e-01 3.10730450e-02 8.86016309e-01 -1.17532670e+00
-6.53417170e-01 4.48283792e-01 -3.60213667e-01 3.16147625e-01
1.89221337e-01 3.79431069e-01 2.95462281e-01 -5.93729079e-01
4.43671912e-01 1.08833706e+00 1.13156295e+00 3.29575270e-01
-9.85488296e-01 -9.17661548e-01 2.23465309e-01 2.49294862e-01
-1.76748109e+00 -4.48541552e-01 9.02063847e-01 -3.08667421e-01
8.82048726e-01 3.59549880e-01 -9.48558673e-02 1.46529508e+00
9.94476676e-01 6.01057887e-01 1.04209089e+00 -1.59803301e-01
3.47828537e-01 2.05185160e-01 -4.13909316e-01 3.27044934e-01
-1.19696081e-01 8.91703308e-01 -3.22852880e-01 -5.80321670e-01
8.65182221e-01 -1.94038525e-02 -9.23033282e-02 2.15079486e-02
-7.23245382e-01 8.56877267e-01 7.65320599e-01 8.80789310e-02
-5.91231763e-01 5.87433219e-01 1.37889460e-01 1.12409085e-01
4.67336506e-01 3.90321910e-01 -2.71490127e-01 4.60091621e-01
-1.01166105e+00 1.55409619e-01 6.21765077e-01 1.07215238e+00
4.35995847e-01 1.31546602e-01 -4.29981083e-01 5.16246140e-01
3.08518946e-01 7.98094213e-01 -3.15130353e-01 -9.97329772e-01
2.29135096e-01 -2.02974930e-01 3.68793793e-02 -1.32537329e+00
-1.72200188e-01 -4.92514133e-01 -8.41953039e-01 5.28858721e-01
1.12573847e-01 -3.36185515e-01 -1.07432270e+00 1.63099158e+00
2.96154350e-01 7.04483449e-01 -6.64444044e-02 5.53500354e-01
3.67891580e-01 5.41784883e-01 1.83293253e-01 8.69078413e-02
1.18822610e+00 -4.00473624e-01 -2.97832251e-01 -1.69385403e-01
-1.16605468e-01 -1.03433108e+00 2.64002055e-01 6.49451256e-01
-7.61681139e-01 -7.49007881e-01 -1.22236025e+00 6.28095984e-01
-3.80951136e-01 -2.60729015e-01 2.57121861e-01 9.67246652e-01
-9.37825501e-01 7.24381030e-01 -7.28906512e-01 1.85323656e-01
7.74882734e-01 3.91349941e-01 -3.18644047e-01 -6.24759234e-02
-1.48896277e+00 1.06582320e+00 1.46799251e-01 1.73623398e-01
-1.65890586e+00 -7.16802776e-01 -8.10126603e-01 -2.71033496e-01
3.43326628e-01 -5.10956526e-01 7.58518219e-01 -7.32046366e-01
-1.19303739e+00 4.17609692e-01 3.64756405e-01 -8.43978345e-01
3.58155072e-01 -1.16995670e-01 -7.22583294e-01 3.00382465e-01
-1.40956283e-01 8.16126049e-01 1.73760998e+00 -1.37105834e+00
-4.67041999e-01 -1.70261249e-01 2.41510764e-01 -2.80512452e-01
-6.81595057e-02 1.18830688e-01 -2.18823016e-01 -1.24038935e+00
-1.63484290e-01 -1.05716813e+00 -4.15981889e-01 -2.78806891e-02
-4.26634490e-01 1.15727261e-01 1.04849756e+00 -7.17171192e-01
1.03971136e+00 -2.19356894e+00 -2.13930637e-01 6.74982190e-01
2.18404338e-01 2.45380193e-01 -1.06795609e-01 3.26052368e-01
-2.49794886e-01 1.14658335e-02 -2.81343013e-01 -1.15943171e-01
9.18621048e-02 1.81079566e-01 -6.98633850e-01 9.09579039e-01
3.01592678e-01 6.03296041e-01 -4.58424628e-01 -2.97773123e-01
2.72055298e-01 6.53750718e-01 -4.46964085e-01 1.38876334e-01
-1.81435078e-01 4.25646812e-01 -5.15703976e-01 8.09356451e-01
1.01665878e+00 4.72474396e-01 -6.32615507e-01 -1.52158275e-01
3.43956441e-01 -2.37156719e-01 -1.14236772e+00 1.42754769e+00
-3.21505040e-01 4.23881024e-01 4.28731859e-01 -7.86808014e-01
7.22419858e-01 5.14891863e-01 3.86530906e-01 -5.00709772e-01
2.64919370e-01 -2.33371302e-01 -1.10631920e-01 -1.80531502e-01
3.87787402e-01 -1.07613921e-01 -2.59278059e-01 2.93364406e-01
1.11434571e-02 -3.28657091e-01 -5.08473098e-01 1.79068938e-01
1.42040896e+00 1.22673459e-01 -6.64842576e-02 -4.45092559e-01
3.60303819e-01 -8.97797197e-02 5.59111655e-01 1.11394072e+00
-4.76213127e-01 5.71702659e-01 2.61220574e-01 -4.41750675e-01
-7.67876685e-01 -1.38665092e+00 -5.52778184e-01 9.79292870e-01
2.21205994e-01 -7.65931979e-02 -7.30696082e-01 -1.11928403e+00
8.97679329e-02 8.40339899e-01 -8.80444527e-01 -3.75433147e-01
-5.67620516e-01 -6.33673966e-01 1.14211392e+00 6.07907236e-01
3.05355459e-01 -9.99288321e-01 -3.81744713e-01 6.02003559e-02
2.38367006e-01 -9.84987617e-01 -4.32628125e-01 4.29134041e-01
-3.69533300e-01 -1.01205862e+00 -4.27236438e-01 -4.62785512e-01
6.43064857e-01 -2.05272973e-01 1.01892614e+00 -2.37735540e-01
-4.13951963e-01 6.40183806e-01 -3.78993928e-01 -9.79228675e-01
-6.79564476e-01 -3.42632890e-01 2.51378119e-01 -1.36989132e-01
1.43852353e-01 -6.49697423e-01 -8.01189661e-01 7.26519048e-01
-1.05253839e+00 -7.42343128e-01 6.78104639e-01 5.75458348e-01
6.89874291e-01 4.58152503e-01 5.36132574e-01 -6.06237829e-01
5.01503110e-01 -7.51993001e-01 -8.87421012e-01 7.67700374e-02
-2.64408737e-01 -2.40820706e-01 6.58466518e-01 -7.10638702e-01
-7.96790779e-01 -1.14523739e-01 -6.09326005e-01 -8.98767114e-01
-4.54216450e-01 1.41405150e-01 -2.36390874e-01 -5.69446623e-01
1.17057979e+00 -1.29610047e-01 -1.22505851e-01 -1.26325022e-02
3.61333966e-01 4.11385298e-01 7.34689534e-01 -7.19464183e-01
1.10678136e+00 5.54125667e-01 1.12199083e-01 -5.26530623e-01
-3.51700693e-01 -1.57240584e-01 -3.58698279e-01 -2.91334897e-01
8.21266651e-01 -8.75590861e-01 -6.41611218e-01 3.99258405e-01
-1.11449146e+00 -2.15519562e-01 -2.24462196e-01 1.65879816e-01
-5.35905063e-01 6.25228584e-01 -4.36436415e-01 -7.80510724e-01
-2.57138550e-01 -1.40853715e+00 1.16814280e+00 -2.14150399e-02
-1.52757883e-01 -1.04464149e+00 2.44235564e-02 1.88312516e-01
6.34846926e-01 5.67455411e-01 6.57462478e-01 -8.05513263e-01
-5.87445438e-01 -7.52769411e-01 1.42531276e-01 7.62673795e-01
-3.35156769e-01 -3.00223649e-01 -1.39865172e+00 -3.91838074e-01
4.56472576e-01 4.47263796e-04 7.43203163e-01 6.39860511e-01
1.27246678e+00 -3.85648787e-01 -4.88584965e-01 6.38935506e-01
1.24824882e+00 2.53692508e-01 9.58186328e-01 4.13102992e-02
5.77084661e-01 4.23592418e-01 4.82990324e-01 2.12741017e-01
-1.40599638e-01 6.82954550e-01 9.45007980e-01 -3.87391560e-02
-9.09832716e-02 -2.58819610e-02 5.49780190e-01 -3.53146568e-02
2.41323471e-01 -4.47556734e-01 -8.95517707e-01 5.49620688e-01
-1.53072679e+00 -1.00259483e+00 4.01849955e-01 2.17269278e+00
5.96476674e-01 4.50934321e-01 -2.34135970e-01 -1.96530804e-01
7.14066088e-01 2.18310609e-01 -6.20735705e-01 -3.16823751e-01
1.31208166e-01 4.89977509e-01 9.54418778e-01 5.87979019e-01
-1.70454550e+00 8.71108830e-01 6.80297565e+00 1.46880412e+00
-9.25801098e-01 2.57529289e-01 6.58710539e-01 1.90364540e-01
-7.47169554e-02 -3.17787305e-02 -6.93581522e-01 4.57007647e-01
9.50551152e-01 3.65675837e-01 3.24446410e-01 9.58889782e-01
-1.07383288e-01 -7.70078152e-02 -1.18016243e+00 6.00602388e-01
2.28044689e-01 -1.35422277e+00 -2.13060752e-01 1.15918793e-01
6.90728009e-01 1.13100111e-01 5.63114643e-01 2.13500217e-01
6.99732959e-01 -1.28701794e+00 8.50106776e-01 5.67237020e-01
6.18987501e-01 -9.04832304e-01 8.49743664e-01 4.02284026e-01
-9.94881690e-01 -1.23327993e-01 -5.53081334e-01 3.52763474e-01
1.01530157e-01 4.75546569e-01 -1.03249300e+00 6.79804742e-01
8.50008786e-01 3.35736871e-01 -4.34961915e-01 7.08132923e-01
-3.10493261e-01 6.90221250e-01 -4.70991105e-01 5.51346421e-01
1.44667089e-01 3.30918878e-01 9.84826803e-01 1.05023503e+00
1.30944490e-01 4.59813029e-02 3.01385403e-01 9.52080667e-01
7.91656226e-02 -5.09619653e-01 -6.78570032e-01 5.83747864e-01
6.64049268e-01 1.00377262e+00 -6.12086773e-01 9.47467461e-02
2.03813892e-02 7.35052228e-01 -1.50535151e-01 3.86412472e-01
-1.38196313e+00 -2.52391905e-01 6.88357949e-01 1.70478716e-01
6.24254644e-01 -8.26365799e-02 -3.91150206e-01 -4.97442722e-01
-4.82833758e-02 -1.00151646e+00 3.43881726e-01 -5.46708524e-01
-1.61049724e+00 8.17588627e-01 3.80855948e-01 -1.26591682e+00
-2.11006388e-01 -6.90739214e-01 -7.76195705e-01 9.78390753e-01
-1.11936378e+00 -1.29011667e+00 1.65655389e-01 8.90875995e-01
2.46974990e-01 -4.06532079e-01 9.31673288e-01 2.03779355e-01
-3.73655766e-01 9.32547212e-01 8.15161597e-03 8.73456895e-02
8.27827156e-01 -9.57332015e-01 4.51837718e-01 1.35642672e+00
4.79992591e-02 7.36120582e-01 8.89054716e-01 -1.00963795e+00
-1.20865488e+00 -1.44538319e+00 -7.09391385e-02 -9.75651801e-01
6.32384360e-01 -3.04990321e-01 -5.79541266e-01 6.39424503e-01
3.69036466e-01 1.72231764e-01 6.04961872e-01 -2.60181844e-01
-5.33362150e-01 -2.05658719e-01 -1.66709208e+00 4.17761087e-01
4.64550853e-01 -8.21513951e-01 -4.92557496e-01 3.96376312e-01
6.88769341e-01 -4.62649405e-01 -1.18306446e+00 6.84522569e-01
4.16005045e-01 -9.15436149e-01 1.41932321e+00 -5.39833419e-02
-2.36207843e-02 -4.82762516e-01 -4.10140157e-01 -1.04008389e+00
-3.72975767e-01 -8.49693358e-01 -1.20311961e-01 9.27802861e-01
4.20101017e-01 -6.63158178e-01 7.98592925e-01 5.35730779e-01
-4.18000847e-01 -6.14227414e-01 -1.44619024e+00 -7.81551957e-01
4.80525851e-01 -9.12223816e-01 4.13082629e-01 6.16393507e-01
-5.70726812e-01 -5.10280430e-01 -4.93890405e-01 1.11676741e+00
7.49643266e-01 -5.27151704e-01 6.21762693e-01 -7.82370806e-01
-5.46531796e-01 -2.08030015e-01 -6.26814187e-01 -7.42182374e-01
4.00522389e-02 -5.55655837e-01 2.85013527e-01 -1.05239582e+00
-2.56162941e-01 -6.43524706e-01 -4.81435239e-01 4.75775003e-01
-1.69409841e-01 5.85185170e-01 -2.54850816e-02 6.52116835e-02
-4.82220411e-01 1.61004439e-01 6.15322649e-01 -4.39850599e-01
2.66719759e-01 2.25977823e-01 -5.48815012e-01 8.80979180e-01
5.75750232e-01 -9.54870701e-01 -3.67066085e-01 -3.11889023e-01
-4.77852821e-02 -7.45050516e-03 7.25351930e-01 -1.18704641e+00
4.36423123e-01 1.64931968e-01 7.15015471e-01 -5.94542861e-01
4.21732843e-01 -1.27852178e+00 4.06963557e-01 2.66168356e-01
1.26179770e-01 2.32141651e-02 6.32282436e-01 8.61688018e-01
-4.31225896e-02 -1.62320927e-01 9.27016377e-01 -5.82335852e-02
-5.05789578e-01 3.55043530e-01 -5.70771039e-01 -2.12439537e-01
1.45148504e+00 -2.53002882e-01 -2.91036785e-01 -3.86805981e-01
-8.53043318e-01 -5.00751659e-02 4.25071418e-01 4.47088122e-01
7.19768286e-01 -1.23545039e+00 -6.57882929e-01 4.20341581e-01
-8.33384693e-02 -3.52857709e-02 4.99046862e-01 6.83057487e-01
-6.06124699e-01 6.08097948e-02 -7.09563047e-02 -7.14287817e-01
-9.37369585e-01 9.26102519e-01 4.41909790e-01 -2.27612883e-01
-3.12521517e-01 1.10499942e+00 3.60092521e-01 -2.04811290e-01
6.28106892e-01 -1.09823741e-01 1.02450818e-01 -3.63335818e-01
3.99551123e-01 3.73134375e-01 8.29145312e-02 -7.68846750e-01
-4.50004041e-01 3.37289631e-01 -2.41128400e-01 2.75045726e-02
1.05149364e+00 1.81805164e-01 1.62419453e-02 -4.18870360e-01
8.84442866e-01 1.97518915e-01 -1.49934232e+00 1.17153004e-01
-2.91525751e-01 -3.14940125e-01 1.42072439e-01 -9.57511902e-01
-9.76072192e-01 4.92612034e-01 9.33166981e-01 1.05316274e-01
1.18911326e+00 3.07746157e-02 6.51242495e-01 2.55533457e-01
6.18602574e-01 -8.82517636e-01 -4.51376438e-02 2.60709941e-01
1.10695267e+00 -1.06246126e+00 1.46901116e-01 -3.11309993e-01
-5.14839411e-01 7.00789988e-01 2.19416127e-01 -6.41290188e-01
1.09762990e+00 5.91840744e-01 1.91435605e-01 -3.07337910e-01
-3.15679073e-01 3.18956614e-01 5.35427749e-01 1.01374006e+00
-2.63543904e-01 1.72612648e-02 3.57371777e-01 6.88424587e-01
-8.00798833e-02 -6.37072444e-01 7.07127601e-02 1.00181067e+00
-1.21377617e-01 -1.15525115e+00 -9.14914072e-01 1.44507959e-01
-8.24291706e-01 -2.20056787e-01 -1.79751098e-01 6.14694357e-01
5.99803090e-01 1.26152086e+00 -3.87511134e-01 -5.89831114e-01
2.88657755e-01 -2.81968564e-01 3.78313452e-01 -2.64566898e-01
-8.49349082e-01 1.71326697e-01 -1.56941459e-01 -7.68502653e-01
-8.15316066e-02 -6.26107991e-01 -6.46044731e-01 -4.21539880e-02
-4.57929939e-01 4.67496701e-02 5.93028843e-01 7.88823783e-01
4.45734113e-01 6.79087400e-01 7.06745565e-01 -1.05890620e+00
-5.89679420e-01 -9.75802362e-01 -4.29937541e-01 1.53599977e-02
2.95669049e-01 -6.60451710e-01 -4.41915631e-01 5.31988814e-02] | [5.539936542510986, 7.9185662269592285] |
5b567c8a-f7db-4ed3-b337-4cd834007082 | e-convrec-a-large-scale-conversational | null | null | https://aclanthology.org/2022.lrec-1.622 | https://aclanthology.org/2022.lrec-1.622.pdf | E-ConvRec: A Large-Scale Conversational Recommendation Dataset for E-Commerce Customer Service | There has been a growing interest in developing conversational recommendation system (CRS), which provides valuable recommendations to users through conversations. Compared to the traditional recommendation, it advocates wealthier interactions and provides possibilities to obtain users’ exact preferences explicitly. Nevertheless, the corresponding research on this topic is limited due to the lack of broad-coverage dialogue corpus, especially real-world dialogue corpus. To handle this issue and facilitate our exploration, we construct E-ConvRec, an authentic Chinese dialogue dataset consisting of over 25k dialogues and 770k utterances, which contains user profile, product knowledge base (KB), and multiple sequential real conversations between users and recommenders. Next, we explore conversational recommendation in a real scene from multiple facets based on the dataset. Therefore, we particularly design three tasks: user preference recognition, dialogue management, and personalized recommendation. In the light of the three tasks, we establish baseline results on E-ConvRec to facilitate future studies. | ['Xiaodong He', 'Jinhui Pang', 'Meng Chen', 'Xin Shen', 'Haobin Li', 'Zexi Xi', 'Yang song', 'Peiying Wang', 'Ruixue Liu', 'Meihuizi Jia'] | null | null | null | null | lrec-2022-6 | ['dialogue-management'] | ['natural-language-processing'] | [-1.30448937e-01 1.54798061e-01 -2.89084643e-01 -7.06926465e-01
-5.44800222e-01 -6.23878241e-01 7.34272838e-01 -4.09053296e-01
-1.08217120e-01 7.57547617e-01 1.09606278e+00 -2.68806934e-01
-7.02932924e-02 -4.40465957e-01 1.87650025e-01 -3.33634645e-01
1.54292271e-01 4.40082639e-01 7.82351289e-03 -8.19082201e-01
5.17900527e-01 -1.03647269e-01 -1.26124036e+00 8.50385666e-01
8.36694717e-01 9.82718408e-01 4.62295800e-01 5.81252158e-01
-1.64602086e-01 7.05910444e-01 -3.90100449e-01 -7.45035231e-01
-2.69231260e-01 -3.38745207e-01 -1.40675068e+00 1.63136467e-01
-2.11015016e-01 -7.93688416e-01 -3.37135077e-01 5.11737168e-01
8.37615073e-01 7.08212912e-01 5.00688732e-01 -8.34725320e-01
-9.82882380e-01 1.15333581e+00 4.34227921e-02 -1.71431184e-01
8.76824319e-01 -2.42151275e-01 1.31824231e+00 -7.45002687e-01
5.16467750e-01 1.22950315e+00 2.24127382e-01 8.08913708e-01
-6.36477351e-01 -2.22440615e-01 2.51632810e-01 1.44003239e-02
-8.40615809e-01 -5.73950708e-01 5.78258336e-01 -1.53967872e-01
9.77894604e-01 6.53508306e-01 4.92157936e-01 1.63944316e+00
-5.41125476e-01 1.06196976e+00 6.62457168e-01 -1.64808929e-01
-7.17512518e-02 5.73159575e-01 6.07759774e-01 1.93266824e-01
-6.17394745e-01 -3.22378457e-01 -6.00492120e-01 -4.55225438e-01
5.81152081e-01 5.13634644e-02 -6.49667025e-01 2.03761742e-01
-1.14589417e+00 8.57972920e-01 2.30403226e-02 8.27909559e-02
-2.14440882e-01 -6.62842691e-01 5.16833484e-01 7.08976686e-01
6.04791701e-01 7.83440411e-01 -8.32132339e-01 -7.66886115e-01
4.84061651e-02 2.47330695e-01 1.43699253e+00 1.28955603e+00
3.41952205e-01 -4.63855892e-01 -2.18629509e-01 1.47879815e+00
5.52301168e-01 2.51675665e-01 7.36158729e-01 -9.52439010e-01
5.11064172e-01 5.44044793e-01 4.60129291e-01 -1.08710051e+00
-3.46538335e-01 -6.15305863e-02 -6.19301200e-01 -9.34733748e-01
1.70040071e-01 -7.84035504e-01 1.52461156e-01 1.31526864e+00
1.81719363e-01 -2.21706092e-01 6.58644676e-01 1.04026890e+00
1.55336833e+00 8.90074909e-01 -3.91559511e-01 -4.09335196e-01
1.52906764e+00 -1.34165299e+00 -6.80209458e-01 2.13659450e-01
8.15820694e-01 -8.55921149e-01 1.47476304e+00 3.29215795e-01
-6.72433913e-01 -4.19288427e-01 -4.88736272e-01 -1.49724483e-01
-1.52372405e-01 4.44843501e-01 8.42174828e-01 6.41990900e-01
-7.31860936e-01 2.49104515e-01 -1.64406613e-01 -6.94509029e-01
-3.46198708e-01 2.12944880e-01 -8.89047906e-02 1.77891716e-01
-1.58089805e+00 6.05208695e-01 -2.36147925e-01 3.51155162e-01
-2.80982614e-01 -3.14173430e-01 -5.67387998e-01 -1.16085678e-01
5.53006470e-01 -3.10386002e-01 1.73904359e+00 -6.48481548e-01
-2.40847898e+00 2.06570908e-01 6.18694834e-02 8.95438641e-02
9.55438316e-02 -4.32310760e-01 -8.32014024e-01 -4.23163056e-01
-4.04497832e-01 2.31624186e-01 2.38328055e-01 -8.85402739e-01
-9.00983214e-01 -4.86182831e-02 5.53738892e-01 6.73455715e-01
-5.35950840e-01 3.16876411e-01 -8.62891257e-01 -4.18051243e-01
-4.29114670e-01 -1.10993314e+00 -2.58879900e-01 -9.26006913e-01
-5.16618967e-01 -5.58645964e-01 4.93143618e-01 -5.66644609e-01
1.31065178e+00 -2.29337454e+00 9.28510949e-02 -2.65020840e-02
-3.40197831e-02 1.72553703e-01 -2.76145697e-01 8.24755669e-01
6.68110311e-01 -2.86515197e-03 4.38516557e-01 -3.41961890e-01
2.16933385e-01 1.48116335e-01 -4.99844939e-01 3.59289050e-02
-5.10217667e-01 7.78402269e-01 -9.93427813e-01 -1.41322643e-01
2.74049677e-02 2.30206445e-01 -7.49242246e-01 7.56352127e-01
-2.97990471e-01 5.37938535e-01 -1.09631395e+00 2.30680302e-01
2.06389561e-01 -3.40420902e-01 4.69483078e-01 -2.45603859e-01
-1.02417782e-01 1.00305617e+00 -8.79602671e-01 1.59964204e+00
-6.54225707e-01 4.00800794e-01 1.00825038e-02 -2.83839643e-01
1.09385109e+00 6.47778511e-01 2.80349135e-01 -5.80601156e-01
2.29193360e-01 -2.12790772e-01 -1.04760192e-01 -8.09741914e-01
1.29271615e+00 5.37745357e-01 -2.90935844e-01 9.95611131e-01
-1.79307640e-01 3.35309923e-01 -1.45730987e-01 3.73319417e-01
8.33411396e-01 -2.14340165e-01 1.68485343e-01 -1.13551736e-01
6.53228045e-01 -1.79508090e-01 1.55998558e-01 6.91589832e-01
1.68139219e-01 3.50318283e-01 2.82595932e-01 -1.75312892e-01
-2.14370906e-01 -3.22174549e-01 8.06726664e-02 1.65311766e+00
3.60510051e-01 -8.08280289e-01 -4.28765208e-01 -8.76209855e-01
-3.25932026e-01 6.06066644e-01 -4.33940351e-01 1.31616458e-01
-4.54177350e-01 -7.18906105e-01 2.89541572e-01 1.82800129e-01
6.45014703e-01 -1.41990209e+00 1.14296243e-01 1.74279630e-01
-8.15648198e-01 -1.03087533e+00 -1.02871156e+00 -5.10575593e-01
-5.56636214e-01 -9.98976588e-01 -6.18597925e-01 -8.39382946e-01
3.80128145e-01 8.07003438e-01 1.26192486e+00 1.09776072e-01
4.97680336e-01 6.54064417e-01 -1.17102957e+00 1.61518216e-01
-3.82547110e-01 5.32109201e-01 1.26851462e-02 -8.66086315e-03
3.90529990e-01 -3.40041965e-01 -6.17601395e-01 1.12947309e+00
-2.69862086e-01 3.15342218e-01 2.62280494e-01 7.46037066e-01
-1.37872681e-01 -2.54393697e-01 9.71695483e-01 -1.29412150e+00
1.43163335e+00 -6.39957786e-01 -2.79259011e-02 3.82199764e-01
-4.90236759e-01 -3.97270352e-01 6.15175426e-01 -5.05240738e-01
-1.65645981e+00 -5.14165223e-01 -4.58008200e-01 7.22679377e-01
-2.41870835e-01 8.85093629e-01 -2.70871073e-01 3.44776928e-01
5.45552254e-01 2.05491409e-02 -7.42840245e-02 -8.03982556e-01
8.25473607e-01 1.72838271e+00 3.13718319e-01 -7.21459210e-01
-1.36224419e-01 -1.67143464e-01 -1.21476054e+00 -1.06649768e+00
-8.87242436e-01 -7.51872897e-01 -3.00525010e-01 -3.10371161e-01
4.95751143e-01 -7.30521917e-01 -1.08178580e+00 1.37669250e-01
-9.49694037e-01 -4.01242554e-01 1.28807604e-01 6.83448255e-01
-3.68438393e-01 5.04658461e-01 -1.13248134e+00 -8.92664731e-01
-5.79835236e-01 -1.18658042e+00 5.41991293e-01 2.59352773e-01
-5.58741331e-01 -9.00177479e-01 1.44267872e-01 8.02621007e-01
5.26390612e-01 -7.39331603e-01 6.42853498e-01 -9.34152842e-01
1.70877669e-02 -7.98718482e-02 -2.12673873e-01 1.81668803e-01
5.30650020e-01 -1.32403165e-01 -7.98666418e-01 -1.46205917e-01
-1.18028700e-01 -6.42278075e-01 2.49675676e-01 -1.54914603e-01
1.06826305e+00 -5.74237764e-01 -2.04883412e-01 -1.81969792e-01
5.02658784e-01 3.44331324e-01 6.49934113e-01 1.01580568e-01
5.52160144e-01 1.00380874e+00 9.45386171e-01 6.60582185e-01
9.42016661e-01 8.89765620e-01 -1.00078937e-02 1.98452905e-01
2.78747678e-01 -1.61255583e-01 4.37662721e-01 1.46287203e+00
-3.76165956e-01 -7.32894838e-01 -1.26788363e-01 1.82258919e-01
-2.16286063e+00 -8.51009071e-01 1.63545534e-01 1.80975330e+00
9.70094681e-01 -3.37047219e-01 4.58415955e-01 -5.83899260e-01
6.03404284e-01 1.46372214e-01 -3.70181054e-01 -3.91276926e-01
2.78944850e-01 -6.80449188e-01 -4.05147932e-02 5.17501414e-01
-7.13660002e-01 9.85694170e-01 6.20814371e+00 6.62656307e-01
-8.68267417e-01 -2.61408985e-01 6.47604764e-01 1.89411327e-01
-4.42989945e-01 -3.67583901e-01 -9.04977560e-01 4.50769484e-01
8.71230423e-01 -1.54396996e-01 7.09754407e-01 8.77540112e-01
1.33715674e-01 2.31837600e-01 -1.16918218e+00 9.70313370e-01
6.33492246e-02 -1.25958610e+00 -1.29674196e-01 1.19243659e-01
3.29524040e-01 5.80749102e-02 7.60268569e-02 7.80682385e-01
6.63792253e-01 -6.95063591e-01 7.73290545e-03 3.37027788e-01
4.28531438e-01 -5.71541607e-01 9.33040738e-01 5.80837309e-01
-8.80945563e-01 1.19704239e-01 -5.65112531e-01 -1.79815307e-01
3.25844884e-01 1.33506820e-01 -8.62868011e-01 5.23401737e-01
5.40916324e-01 1.17844403e+00 -3.29836160e-02 5.42045414e-01
1.67635009e-01 6.71708703e-01 -5.42038158e-02 -5.17737389e-01
6.22805282e-02 -6.55459523e-01 1.02648646e-01 1.44195771e+00
2.52438009e-01 7.64143586e-01 2.62032896e-01 2.06785768e-01
-3.40755343e-01 5.48411429e-01 -3.00397933e-01 -1.89781845e-01
6.15403116e-01 1.54492760e+00 -2.89837897e-01 -7.64873251e-02
-9.07993138e-01 1.03367627e+00 3.50798309e-01 3.46694946e-01
-4.60711300e-01 -7.55066946e-02 8.03526342e-01 -4.51877058e-01
4.28604186e-02 -1.68665554e-02 3.85619849e-01 -1.28975475e+00
-4.00275081e-01 -1.22697902e+00 2.50177175e-01 -3.82219464e-01
-1.68011463e+00 8.32932234e-01 -3.24020416e-01 -1.09634829e+00
-4.82301146e-01 -2.96062201e-01 -5.50717235e-01 7.07185984e-01
-1.10193694e+00 -7.89992511e-01 -1.52351215e-01 6.04826331e-01
9.91523325e-01 -4.22842056e-01 1.22309923e+00 5.86039245e-01
-7.10472286e-01 7.13727117e-01 2.17868730e-01 9.46954712e-02
1.14233005e+00 -1.04089999e+00 3.80276799e-01 -7.05101117e-02
-1.74422309e-01 9.13531303e-01 6.10095203e-01 -4.68893826e-01
-1.66952503e+00 -7.62644112e-01 7.31774390e-01 -5.29614270e-01
5.84848225e-01 -2.60982841e-01 -6.21084094e-01 6.56473517e-01
4.60542262e-01 -7.17212498e-01 1.29203522e+00 8.59680474e-01
9.72689241e-02 1.86336040e-01 -8.67447913e-01 9.13770795e-01
1.08535886e+00 -5.05993962e-01 -5.33575892e-01 2.52386302e-01
9.46311414e-01 -5.00949085e-01 -1.25227308e+00 1.57979429e-01
8.78891885e-01 -7.57968307e-01 8.29698563e-01 -6.58141434e-01
2.72603422e-01 1.62402794e-01 -2.82430112e-01 -1.57898486e+00
-2.71473348e-01 -9.23421264e-01 -6.58605099e-02 1.62446141e+00
8.30965638e-01 -5.19324958e-01 6.86668456e-01 1.12231648e+00
-3.21039915e-01 -7.80596375e-01 -1.49839789e-01 -1.60359358e-03
-5.43054342e-01 -3.06075841e-01 1.05915809e+00 1.24034035e+00
9.22741354e-01 9.96634901e-01 -1.11725950e+00 -1.13331862e-01
-2.80314803e-01 6.03667319e-01 1.02877867e+00 -1.05223417e+00
-5.52946448e-01 -2.26954609e-01 6.01787210e-01 -2.06114912e+00
8.31872970e-02 -5.08451521e-01 1.94434851e-01 -1.37296879e+00
9.56041366e-02 -6.71996415e-01 8.93956870e-02 1.57291621e-01
-1.96500532e-02 -1.25431776e-01 -1.60082027e-01 3.96543533e-01
-1.02665603e+00 7.84877002e-01 1.66561139e+00 2.18709499e-01
-6.82256341e-01 5.15398204e-01 -1.40304422e+00 5.47465980e-01
7.16078281e-01 1.20961443e-01 -7.56002307e-01 -1.71449125e-01
3.88246506e-01 5.59843302e-01 -6.41622841e-01 7.99862295e-02
1.47781566e-01 -3.71377677e-01 -1.90600321e-01 -4.50669289e-01
3.81896198e-01 -6.30399466e-01 -1.76673084e-01 -4.33754683e-01
-1.04009521e+00 -5.78821041e-02 -3.67537975e-01 4.70167190e-01
-4.41218540e-02 -1.44790217e-01 -4.10758480e-02 -8.28808472e-02
-7.66740799e-01 4.04078096e-01 -7.51503885e-01 -2.90659994e-01
3.11337501e-01 2.03939423e-01 -6.50560498e-01 -1.01653051e+00
-6.92112863e-01 6.84805810e-01 1.15386687e-01 8.25591326e-01
5.08363724e-01 -1.17696571e+00 -6.53936267e-01 -9.13083106e-02
3.39924812e-01 -2.73265123e-01 4.01125044e-01 4.18190837e-01
8.79558474e-02 5.11034608e-01 -1.04020676e-03 5.78559786e-02
-1.49001229e+00 3.00028920e-01 1.00449398e-01 -7.51611292e-02
-7.50477612e-01 6.21191204e-01 8.80652592e-02 -1.03002715e+00
6.21034086e-01 -1.46584168e-01 -1.27915597e+00 1.87092274e-01
9.70109284e-01 4.82226431e-01 -2.10506082e-01 -6.39139175e-01
1.39242768e-01 -1.45730734e-01 -5.61686099e-01 -1.26136884e-01
1.09553707e+00 -9.63802040e-01 -1.95429698e-01 4.15010005e-01
9.64302003e-01 3.38223934e-01 -8.27638924e-01 -6.35665596e-01
-1.21545486e-01 -5.58394909e-01 -1.63500100e-01 -8.77413094e-01
-9.84013677e-01 1.52023926e-01 1.49343848e-01 7.40424693e-01
6.58363998e-01 2.77262647e-02 1.09938455e+00 9.89974797e-01
3.38310778e-01 -1.17721558e+00 1.30695075e-01 8.29046011e-01
1.06607068e+00 -1.24355245e+00 -2.95627117e-01 -5.40786743e-01
-1.30210745e+00 9.90582585e-01 7.38954425e-01 3.64928961e-01
9.54453290e-01 -3.60319078e-01 5.19223630e-01 -1.20417893e-01
-9.91597593e-01 -1.01224154e-01 3.75910461e-01 4.27965850e-01
9.99696076e-01 1.45275965e-01 -2.96501726e-01 1.33259058e+00
-5.92193902e-01 -5.04874140e-02 6.21132076e-01 5.33836544e-01
-4.58859801e-01 -1.29932952e+00 3.89819741e-01 6.59880280e-01
-1.97455555e-01 -1.03266999e-01 -5.23583412e-01 3.09221476e-01
-7.09077358e-01 1.63208532e+00 -2.55107075e-01 -8.04737628e-01
5.11922002e-01 -2.87836254e-01 -1.66215952e-02 -7.83203840e-01
-7.48520970e-01 6.70679808e-02 1.05133188e+00 -5.06113172e-01
-6.20479524e-01 -4.98312891e-01 -9.36730385e-01 -6.36789382e-01
-7.90813565e-01 8.45722079e-01 3.46253514e-01 1.13949668e+00
4.86687660e-01 1.95114091e-01 9.97790456e-01 -7.48299837e-01
-4.55596775e-01 -1.32869852e+00 -8.03862870e-01 2.03291595e-01
5.35248928e-02 -3.06436151e-01 -7.78867900e-02 -3.11084032e-01] | [12.339434623718262, 7.519394874572754] |
c7a902df-bbb1-4804-a777-eb71889ae8b5 | bayesian-regression-approach-for-building-and | 2201.02034 | null | https://arxiv.org/abs/2201.02034v1 | https://arxiv.org/pdf/2201.02034v1.pdf | Bayesian Regression Approach for Building and Stacking Predictive Models in Time Series Analytics | The paper describes the use of Bayesian regression for building time series models and stacking different predictive models for time series. Using Bayesian regression for time series modeling with nonlinear trend was analyzed. This approach makes it possible to estimate an uncertainty of time series prediction and calculate value at risk characteristics. A hierarchical model for time series using Bayesian regression has been considered. In this approach, one set of parameters is the same for all data samples, other parameters can be different for different groups of data samples. Such an approach allows using this model in the case of short historical data for specified time series, e.g. in the case of new stores or new products in the sales prediction problem. In the study of predictive models stacking, the models ARIMA, Neural Network, Random Forest, Extra Tree were used for the prediction on the first level of model ensemble. On the second level, time series predictions of these models on the validation set were used for stacking by Bayesian regression. This approach gives distributions for regression coefficients of these models. It makes it possible to estimate the uncertainty contributed by each model to stacking result. The information about these distributions allows us to select an optimal set of stacking models, taking into account the domain knowledge. The probabilistic approach for stacking predictive models allows us to make risk assessment for the predictions that are important in a decision-making process. | ['Bohdan M. Pavlyshenko'] | 2022-01-06 | null | null | null | null | ['time-series-prediction'] | ['time-series'] | [-1.70431331e-01 -1.58341572e-01 1.00389376e-01 -4.94453967e-01
-3.26035142e-01 -3.42956007e-01 6.89958513e-01 5.45015037e-01
-1.92392349e-01 6.47929430e-01 1.52044088e-01 -4.42263931e-01
-7.23636150e-01 -1.18792391e+00 -3.61103088e-01 -8.15916121e-01
-2.68222570e-01 7.17344165e-01 3.57581139e-01 -1.88270122e-01
3.04856002e-01 7.85046518e-01 -1.75901115e+00 3.18802208e-01
5.34023285e-01 8.86063457e-01 2.86760390e-01 4.27580923e-01
-4.32196587e-01 5.45745075e-01 -4.67153132e-01 1.70419797e-01
1.24741964e-01 -3.52301896e-02 -1.95397183e-01 -4.96741712e-01
-4.02396232e-01 -2.46649936e-01 5.05492389e-01 6.04371786e-01
1.25193475e-02 3.31976980e-01 1.30908704e+00 -1.18246794e+00
1.87888652e-01 1.00468612e+00 -2.93405473e-01 1.49200559e-01
2.42705330e-01 -2.47048587e-01 4.35771227e-01 -3.03377390e-01
1.97736442e-01 1.32637393e+00 6.23637795e-01 -2.87008733e-01
-1.44490731e+00 -5.88532209e-01 -4.00122590e-02 2.50300229e-01
-1.17323637e+00 3.54983419e-01 5.94936550e-01 -1.10491014e+00
7.26807415e-01 2.87449092e-01 7.59913087e-01 4.77829307e-01
7.24095166e-01 4.04290296e-02 1.52976787e+00 -4.99216974e-01
3.30456883e-01 3.33819300e-01 8.98895860e-01 -2.12540403e-01
2.94720083e-01 5.11691272e-01 -4.57353815e-02 -2.02444017e-01
1.41943857e-01 3.14127475e-01 1.58456549e-01 2.06659332e-01
-4.45126444e-01 8.74853969e-01 1.10489145e-01 8.22203934e-01
-7.29371369e-01 1.03580281e-02 3.96890640e-01 1.26115605e-01
6.01350248e-01 3.45661640e-02 -5.73113263e-01 2.86577642e-02
-9.98458564e-01 4.90812570e-01 9.55326140e-01 3.07509392e-01
4.35895711e-01 9.39030647e-02 -1.12286307e-01 2.34063640e-01
5.03530979e-01 5.58682621e-01 4.00653481e-01 -2.79357016e-01
1.41561687e-01 3.21732700e-01 2.71840125e-01 -9.30173039e-01
-5.80595493e-01 -3.44235152e-01 -4.78926182e-01 4.52054173e-01
4.22329307e-01 -9.85399261e-02 -9.61026132e-01 1.22862828e+00
2.40123823e-01 1.36198357e-01 1.11096650e-02 1.68755397e-01
3.55964959e-01 1.34283090e+00 3.28434885e-01 -6.64833248e-01
1.19031394e+00 -1.84571203e-02 -8.04645956e-01 7.60129571e-01
4.07414556e-01 -8.59169781e-01 2.98809916e-01 8.81010711e-01
-3.69424820e-01 -6.36177957e-01 -9.10878599e-01 7.86670804e-01
-6.90433145e-01 8.33639875e-02 1.34535292e-02 5.73735058e-01
-4.77055997e-01 1.00602889e+00 -9.29321826e-01 -3.84861320e-01
-4.30083126e-01 3.97951394e-01 1.03372537e-01 1.67498618e-01
-1.04721594e+00 1.37808561e+00 8.35462213e-01 2.16659665e-01
-5.19720793e-01 -5.61485291e-01 -3.50374460e-01 1.74259976e-01
-1.77199438e-01 -1.65302306e-01 9.32865441e-01 -6.27855957e-01
-1.22942960e+00 5.52768074e-02 -2.54779086e-02 -5.79784334e-01
5.29473066e-01 -6.33290708e-02 -6.71260238e-01 -4.62065190e-01
-8.42650458e-02 -1.28180280e-01 7.00423896e-01 -1.04658115e+00
-4.11289811e-01 -4.67839390e-01 -4.31042403e-01 -4.02210385e-01
3.35156262e-01 3.07495505e-01 3.97894025e-01 -4.26166832e-01
9.21343490e-02 -8.25971842e-01 -3.70228112e-01 -9.68773723e-01
-1.40652448e-01 -3.26583058e-01 7.67013371e-01 -1.09510314e+00
1.41782784e+00 -1.92827845e+00 -3.35869472e-03 9.83974755e-01
-5.03950238e-01 -2.10335687e-01 5.28080881e-01 8.43775094e-01
-3.45370889e-01 7.48878419e-02 -2.06990227e-01 1.26954451e-01
-2.06294283e-01 2.14358047e-01 -4.78163034e-01 6.99996436e-03
-2.26762787e-01 -2.45594442e-01 -1.99609533e-01 -3.15562546e-01
5.97091496e-01 4.14681107e-01 -5.75324101e-03 2.67384261e-01
-2.91387767e-01 4.08717453e-01 -6.27744853e-01 3.54687013e-02
7.89881349e-01 4.77493346e-01 1.51132405e-01 -3.47146332e-01
-6.94931149e-01 -1.26509033e-02 -1.29868948e+00 4.85401481e-01
-5.85328639e-01 1.58893794e-01 -6.19430244e-01 -1.17620313e+00
1.45786035e+00 5.50123036e-01 4.91945982e-01 -1.24660634e-01
2.55596638e-01 3.63052875e-01 9.19623002e-02 -4.07429338e-01
2.35534981e-01 -3.64268124e-01 1.43027499e-01 3.12130570e-01
-1.09719940e-01 -2.05755487e-01 1.59591526e-01 -5.06696820e-01
5.14702022e-01 4.67415333e-01 3.77812952e-01 -2.81627566e-01
4.72816557e-01 6.06114343e-02 4.20088053e-01 1.99708670e-01
5.85227549e-01 1.36591062e-01 7.26052821e-01 -6.16685927e-01
-9.48156655e-01 -7.76971698e-01 -5.54998219e-01 6.26344264e-01
-6.64533019e-01 -6.71339333e-02 -3.03654164e-01 -9.92666408e-02
3.14613655e-02 1.26683331e+00 -4.61088568e-01 1.91142887e-01
-2.62100458e-01 -1.04279149e+00 -1.96549594e-01 4.56838645e-02
1.09427892e-01 -8.20334673e-01 -5.43801963e-01 6.82224214e-01
2.66737014e-01 -3.96408141e-01 6.87670290e-01 5.34439802e-01
-1.45888662e+00 -8.17026258e-01 -3.55158269e-01 1.21613346e-01
2.61662334e-01 -5.10191262e-01 8.30539823e-01 -3.95817935e-01
-3.56429741e-02 1.71210811e-01 -4.88064170e-01 -1.02068925e+00
-8.18528414e-01 -1.26951426e-01 -2.45875224e-01 -5.46845682e-02
3.46372426e-01 -7.01391637e-01 -5.51146921e-03 2.43163660e-01
-6.66716516e-01 -3.84013385e-01 1.37201086e-01 3.27581137e-01
5.21256447e-01 4.54831332e-01 6.48017764e-01 -7.48530686e-01
6.18451476e-01 -8.45049441e-01 -1.24460733e+00 4.42678630e-01
-7.87161767e-01 2.05912367e-01 5.02188683e-01 -2.23409429e-01
-1.16706848e+00 5.49640357e-02 -1.87471583e-01 -1.47998393e-01
-2.51469165e-01 9.46138442e-01 1.52908638e-01 4.05499876e-01
5.53931594e-01 -2.62374021e-02 4.46867198e-02 -7.67150223e-01
-2.16269344e-01 4.12882835e-01 -1.28536746e-01 -4.74112928e-01
4.71717000e-01 -1.53161824e-01 4.11366016e-01 -7.13377178e-01
-1.78189695e-01 -3.13382059e-01 -7.23565102e-01 -6.90664649e-01
1.00117528e+00 -5.54619491e-01 -4.93049979e-01 4.56380665e-01
-1.06422722e+00 8.59575942e-02 -4.79550548e-02 9.71354544e-01
-4.70369369e-01 -1.63640350e-01 -9.37372670e-02 -1.57510459e+00
-2.60710359e-01 -1.07296193e+00 2.55832821e-01 1.19553320e-01
-3.97887170e-01 -1.12416697e+00 1.91870749e-01 -9.32402685e-02
2.91995227e-01 5.63275814e-01 1.17421770e+00 -1.07855201e+00
-4.36969280e-01 -6.32605374e-01 3.42088282e-01 5.84406316e-01
-1.88063234e-01 7.43494868e-01 -6.73614860e-01 3.21689010e-01
1.55766383e-01 4.03838813e-01 7.49517381e-01 8.99126709e-01
1.01639462e+00 -4.91888262e-02 -4.24607694e-01 -3.26516479e-02
1.59236538e+00 8.33046079e-01 6.63129568e-01 2.31286019e-01
2.09397390e-01 1.01606572e+00 1.19512987e+00 6.54659033e-01
3.67368758e-02 8.98730814e-01 2.38382980e-01 5.29726326e-01
6.62627399e-01 1.03580020e-01 4.04795885e-01 3.12882692e-01
-5.59638619e-01 3.23764980e-02 -1.12780583e+00 1.39935136e-01
-1.77311277e+00 -1.31386507e+00 -7.01825321e-01 2.64248610e+00
2.63360888e-01 3.68636638e-01 4.08214748e-01 4.63167042e-01
8.31175983e-01 -1.58747733e-01 1.31088853e-01 -8.22210133e-01
4.92565393e-01 1.66626006e-01 6.89207613e-01 6.32786393e-01
-9.01980460e-01 1.96667299e-01 6.01877260e+00 6.98242605e-01
-1.09911025e+00 -9.84149352e-02 5.56181073e-01 5.30610010e-02
-4.01445299e-01 3.05176973e-01 -1.00433528e+00 8.68696511e-01
1.38573480e+00 -2.57082045e-01 1.91257656e-01 6.72478557e-01
7.68353462e-01 -7.07606971e-01 -7.99057603e-01 4.12264049e-01
-4.15662587e-01 -9.09503281e-01 -1.32240891e-01 1.94494262e-01
4.57191944e-01 -1.03581429e-01 -2.90959001e-01 1.39076576e-01
4.41134810e-01 -8.89138222e-01 6.10808432e-01 1.32316732e+00
-2.88337227e-02 -9.93141234e-01 1.17972934e+00 4.98026848e-01
-8.79214346e-01 -2.80364662e-01 -1.09242775e-01 -8.84047598e-02
4.72163230e-01 1.10792053e+00 -8.33689988e-01 8.02411020e-01
6.26479506e-01 4.20793295e-01 -1.03890657e-01 1.16253078e+00
2.48460189e-01 9.36870396e-01 -9.53564048e-01 -2.51583189e-01
-1.28183318e-02 -6.60797715e-01 5.06877422e-01 8.56867671e-01
1.03775501e+00 3.75494137e-02 7.75557384e-03 9.07900870e-01
1.07084119e+00 3.94810170e-01 -7.32686043e-01 1.96111575e-01
5.31745374e-01 7.84600854e-01 -6.64042473e-01 -1.40692532e-01
6.76503405e-02 -1.50373012e-01 -3.79746825e-01 2.49351621e-01
-6.28135800e-01 -9.20352191e-02 -8.96492824e-02 4.85560417e-01
1.53209299e-01 -3.18369538e-01 -2.45200679e-01 -4.55192894e-01
-2.41113707e-01 -3.09556693e-01 5.20016611e-01 -7.77795732e-01
-1.08379495e+00 3.72847259e-01 1.02641296e+00 -1.26595688e+00
-6.89834774e-01 -6.99222863e-01 -9.89247739e-01 1.43310642e+00
-7.12334156e-01 -9.71227705e-01 6.24041148e-02 2.63076007e-01
1.48944438e-01 -2.38793805e-01 6.95814252e-01 -1.39994264e-01
-4.32603747e-01 -5.32576323e-01 5.42608261e-01 -4.38989490e-01
3.79898787e-01 -1.06966996e+00 -4.42703247e-01 6.24190867e-01
-4.18653220e-01 2.63359845e-01 1.17230809e+00 -9.50761855e-01
-4.61373061e-01 -6.69096291e-01 9.87023473e-01 -1.90458566e-01
6.91416860e-01 2.15805084e-01 -9.91181195e-01 5.07436633e-01
-1.13972776e-01 -6.74978733e-01 8.57871711e-01 4.01049972e-01
4.33796421e-02 -5.32918155e-01 -1.22852933e+00 5.26286773e-02
-3.42080384e-01 5.20078577e-02 -7.71726847e-01 1.15665689e-01
3.12080801e-01 1.68523997e-01 -1.39440250e+00 2.10794851e-01
7.93017745e-01 -1.16410434e+00 6.65280461e-01 -2.35076457e-01
1.87831461e-01 -3.27060074e-01 -9.31680426e-02 -1.32833302e+00
-1.11080565e-01 -1.11572342e-02 3.72563004e-01 1.48637688e+00
5.51089108e-01 -7.75922000e-01 4.12751501e-03 7.88009584e-01
1.52730867e-01 -3.58826518e-01 -8.55163813e-01 -7.14108527e-01
-2.45279856e-02 -6.41501844e-01 8.15961182e-01 5.74631035e-01
-2.96958715e-01 -1.78567454e-01 -1.79453760e-01 2.88584501e-01
6.21018708e-01 7.39847422e-02 5.17253637e-01 -1.73612690e+00
-3.69218677e-01 -2.57625252e-01 -5.70472658e-01 2.59136409e-01
-7.89052434e-03 -3.30796361e-01 -3.50090563e-01 -1.55715656e+00
-3.19371015e-01 -7.18078017e-01 -1.72251195e-01 3.81616428e-02
1.73245758e-01 -4.87704128e-01 3.02285999e-01 2.09980562e-01
8.17798018e-01 1.12947360e-01 5.07207215e-01 3.06999981e-01
-4.40885156e-01 8.06220055e-01 2.32088700e-01 6.47833049e-01
8.74384105e-01 -7.45531619e-01 -6.21417701e-01 4.39518064e-01
2.84366488e-01 5.01175761e-01 3.02151650e-01 -1.01358151e+00
-2.06477419e-01 -4.30836469e-01 4.07988876e-01 -1.20152056e+00
7.69864991e-02 -1.49702716e+00 1.16995931e+00 7.22151697e-01
-1.94246083e-01 1.56337053e-01 1.56537369e-01 5.16334355e-01
-3.81798208e-01 -8.88838172e-01 4.96344119e-01 1.22471660e-01
-3.78874272e-01 -1.24386840e-01 -7.46205389e-01 -9.53509271e-01
1.19443905e+00 -4.49514002e-01 2.35631913e-01 -5.04734755e-01
-1.43820989e+00 1.27362326e-01 -1.87207878e-01 -3.63042839e-02
-3.92135121e-02 -1.06463909e+00 -7.09684312e-01 -2.23932922e-01
-1.23625249e-01 -2.57069975e-01 4.39545482e-01 9.32948351e-01
-7.10456431e-01 2.27919534e-01 -5.69585621e-01 -4.30131704e-01
-1.37635386e+00 6.91867828e-01 4.49844688e-01 -5.88445008e-01
8.59631039e-03 1.33581817e-01 -3.29978138e-01 1.38930976e-02
-1.31424770e-01 -7.31129706e-01 -1.07256687e+00 6.78126991e-01
1.92618817e-01 6.32592559e-01 8.12468603e-02 -4.93586779e-01
-1.20647557e-01 7.93873370e-01 2.47156367e-01 -4.96017218e-01
1.53013384e+00 1.35749325e-01 -4.36437756e-01 1.14122832e+00
9.19704854e-01 -1.09813832e-01 -7.58896887e-01 3.05765897e-01
3.98842514e-01 -1.27991736e-01 2.30413675e-01 -9.37965393e-01
-5.07046998e-01 7.86933184e-01 7.59734869e-01 6.27955556e-01
1.23484266e+00 -4.53355908e-01 -1.67876750e-01 3.48666087e-02
4.33112621e-01 -9.79210258e-01 -9.28272307e-01 3.23333234e-01
1.09444833e+00 -9.69302833e-01 1.86225191e-01 -4.95587766e-01
-4.78677630e-01 1.79442775e+00 -4.15423959e-02 -1.59428313e-01
1.30993378e+00 2.73804069e-01 -4.78385836e-01 2.29819231e-02
-7.25256681e-01 -7.94951990e-02 3.27404201e-01 6.04778111e-01
6.98608398e-01 6.12981081e-01 -8.74514997e-01 8.62748146e-01
-2.57925034e-01 5.14345288e-01 4.08821464e-01 6.00611925e-01
-5.24487674e-01 -1.27056658e+00 -9.58001673e-01 7.21354842e-01
-3.86916727e-01 1.46814406e-01 1.81708232e-01 9.19187367e-01
2.58121043e-01 9.20437753e-01 2.35761255e-01 -3.96466792e-01
3.81874233e-01 5.11112630e-01 2.96976477e-01 -3.49482536e-01
-7.85669446e-01 1.90947741e-01 5.59905529e-01 2.86393496e-03
-2.15073735e-01 -1.15350699e+00 -1.11020041e+00 -3.54140490e-01
-3.65709960e-01 3.76750678e-01 1.23564899e+00 8.74418020e-01
-1.42283380e-01 4.83441561e-01 7.19827950e-01 -7.56036878e-01
-5.53108275e-01 -1.18538618e+00 -8.75086546e-01 -1.87338844e-01
-4.12152857e-02 -7.91513205e-01 -3.45769256e-01 1.27842605e-01] | [6.8020477294921875, 3.3426454067230225] |
bc2f3186-c798-4606-91b7-3dd82b3afcf1 | freematch-self-adaptive-thresholding-for-semi | 2205.07246 | null | https://arxiv.org/abs/2205.07246v3 | https://arxiv.org/pdf/2205.07246v3.pdf | FreeMatch: Self-adaptive Thresholding for Semi-supervised Learning | Semi-supervised Learning (SSL) has witnessed great success owing to the impressive performances brought by various methods based on pseudo labeling and consistency regularization. However, we argue that existing methods might fail to utilize the unlabeled data more effectively since they either use a pre-defined / fixed threshold or an ad-hoc threshold adjusting scheme, resulting in inferior performance and slow convergence. We first analyze a motivating example to obtain intuitions on the relationship between the desirable threshold and model's learning status. Based on the analysis, we hence propose FreeMatch to adjust the confidence threshold in a self-adaptive manner according to the model's learning status. We further introduce a self-adaptive class fairness regularization penalty to encourage the model for diverse predictions during the early training stage. Extensive experiments indicate the superiority of FreeMatch especially when the labeled data are extremely rare. FreeMatch achieves 5.78%, 13.59%, and 1.28% error rate reduction over the latest state-of-the-art method FlexMatch on CIFAR-10 with 1 label per class, STL-10 with 4 labels per class, and ImageNet with 100 labels per class, respectively. Moreover, FreeMatch can also boost the performance of imbalanced SSL. The codes can be found at https://github.com/microsoft/Semi-supervised-learning. | ['Xing Xie', 'Bernt Schiele', 'Yue Fan', 'Jindong Wang', 'Zhen Wu', 'Bhiksha Raj', 'Takahiro Shinozaki', 'Marios Savvides', 'Wenxin Hou', 'Qiang Heng', 'Hao Chen', 'Yidong Wang'] | 2022-05-15 | null | null | null | null | ['semi-supervised-image-classification'] | ['computer-vision'] | [-1.06495515e-01 7.91129097e-02 -6.15686893e-01 -9.77458954e-01
-8.01100194e-01 -2.64435202e-01 1.76065758e-01 1.38895735e-01
-6.10767722e-01 8.19112301e-01 -1.58774853e-01 -2.01216489e-01
-6.74397051e-02 -3.79467785e-01 -5.64943850e-01 -7.93958485e-01
2.03229338e-01 4.57229227e-01 9.34021324e-02 8.03266317e-02
1.34500042e-01 -7.77186407e-03 -1.39769554e+00 2.25883260e-01
1.24264538e+00 1.22168767e+00 -6.07208945e-02 1.00282570e-02
8.99806470e-02 8.21428359e-01 -3.23782742e-01 -4.23151642e-01
4.39810127e-01 -3.55427235e-01 -5.42457581e-01 1.21347211e-01
4.57025558e-01 -1.39189452e-01 -2.64091324e-02 1.17672670e+00
4.73624825e-01 2.49728411e-02 5.84849775e-01 -1.52542818e+00
-6.06685460e-01 8.62367153e-01 -8.21697354e-01 1.81164429e-01
-2.63580710e-01 1.54677019e-01 1.06677604e+00 -9.90523696e-01
1.88727945e-01 1.13412547e+00 6.37572050e-01 6.01248622e-01
-1.16664517e+00 -1.17784309e+00 3.04822892e-01 2.44514883e-01
-1.62549138e+00 -4.36349690e-01 6.72480702e-01 -3.50614578e-01
3.66209060e-01 1.54221514e-02 1.37391239e-01 8.64792645e-01
-1.33884460e-01 8.91149819e-01 1.38589609e+00 -4.67415333e-01
3.48675042e-01 4.32829350e-01 5.87259829e-01 6.24915183e-01
2.98937291e-01 1.83724672e-01 -5.53664804e-01 -2.60500401e-01
4.26022232e-01 -8.40151012e-02 -8.22490975e-02 -3.18264008e-01
-8.89633119e-01 9.64351714e-01 4.85112250e-01 -7.94155672e-02
-1.70918047e-01 -1.56884894e-01 5.61041951e-01 2.63733029e-01
7.13820398e-01 2.43192673e-01 -7.30542123e-01 1.90645620e-01
-1.05111814e+00 8.56680945e-02 4.90255356e-01 7.64574647e-01
6.64375126e-01 4.24440056e-02 -3.04750919e-01 1.12021434e+00
5.36025643e-01 3.62237722e-01 5.22440553e-01 -9.82638478e-01
4.57282454e-01 5.97766340e-01 6.69295192e-02 -6.43837571e-01
-3.04065555e-01 -8.88773322e-01 -9.21727955e-01 1.82363838e-01
5.19496799e-01 -1.59738630e-01 -1.04710352e+00 1.93733573e+00
3.70450735e-01 1.30924329e-01 -6.76174015e-02 9.84920979e-01
6.35125399e-01 3.72761458e-01 3.70596379e-01 -4.30110037e-01
1.04663265e+00 -1.18339539e+00 -6.83752000e-01 -2.14149266e-01
7.69997776e-01 -6.75906956e-01 1.41109931e+00 4.10559207e-01
-7.99526691e-01 -5.33822596e-01 -9.72498536e-01 3.35973054e-01
3.68737872e-03 4.56333607e-01 5.10530770e-01 7.42635250e-01
-6.19919896e-01 4.98852104e-01 -7.49621689e-01 4.56117168e-02
8.72310698e-01 3.90830398e-01 -3.00329234e-02 -6.83080927e-02
-1.30257988e+00 5.56998491e-01 4.39958781e-01 5.70631474e-02
-5.36926031e-01 -7.17006624e-01 -5.10516942e-01 4.01450582e-02
4.54867542e-01 -2.82421350e-01 1.34638751e+00 -1.36019337e+00
-1.19077516e+00 9.57108021e-01 5.89254051e-02 -6.08867049e-01
6.82313323e-01 -1.87913507e-01 -2.72637188e-01 -5.82371801e-02
1.22097671e-01 7.64051378e-01 7.27487445e-01 -1.17928958e+00
-6.52863503e-01 -3.93480778e-01 -1.38494655e-01 2.33455256e-01
-5.02451122e-01 -6.90555200e-02 -3.42248708e-01 -7.31302738e-01
1.95701435e-01 -1.00081301e+00 -3.89628410e-01 -1.12884715e-02
-5.04032195e-01 -3.95122975e-01 5.55299044e-01 -1.63248286e-01
1.45435333e+00 -2.23011184e+00 -5.65773726e-01 2.32121065e-01
2.07237184e-01 6.44872785e-01 -6.18034154e-02 -5.67387566e-02
-1.72660410e-01 2.27002189e-01 -3.22707623e-01 -3.46523225e-01
-1.95526574e-02 5.74073894e-03 -1.05594449e-01 4.38026905e-01
4.04878072e-02 5.79637170e-01 -8.26100588e-01 -6.02658033e-01
-2.66910605e-02 3.65683526e-01 -6.59651160e-01 1.19745508e-01
3.19729447e-02 3.91811192e-01 -4.62794453e-01 6.97269082e-01
8.67428839e-01 -6.12713516e-01 2.05698207e-01 -3.34646225e-01
1.65798932e-01 2.93137997e-01 -1.24296808e+00 1.10571682e+00
-1.63914889e-01 2.07760260e-01 -1.88415825e-01 -1.01641524e+00
1.04128754e+00 1.92675427e-01 3.80906224e-01 -7.27559149e-01
1.93120390e-01 2.94749469e-01 1.36368796e-01 -1.19116820e-01
1.63517877e-01 -2.45642915e-01 1.22610532e-01 4.15318489e-01
-1.71415240e-01 5.33976376e-01 1.89281672e-01 2.12779954e-01
7.26220250e-01 -6.08399399e-02 2.99781501e-01 -3.93362701e-01
4.47958082e-01 -1.14227541e-01 1.01900518e+00 7.41221488e-01
-6.65172935e-01 5.72288036e-01 3.53953987e-01 -4.19709653e-01
-8.26006889e-01 -9.04786646e-01 -4.63525683e-01 1.26444674e+00
1.23716511e-01 -2.46890724e-01 -7.14445829e-01 -9.85108972e-01
6.85170665e-02 6.64202571e-01 -5.45265198e-01 -3.21087927e-01
-1.92373887e-01 -1.23199522e+00 4.16863531e-01 5.09285808e-01
7.36060321e-01 -9.17836189e-01 -2.80598193e-01 5.76610714e-02
-1.70523375e-01 -1.01219332e+00 -4.55837190e-01 3.53825808e-01
-9.24942493e-01 -9.54049170e-01 -4.08626795e-01 -7.36226022e-01
1.05613804e+00 1.85856998e-01 1.15232575e+00 1.80242196e-01
1.39944077e-01 -1.02909088e-01 -4.39842522e-01 -2.88765758e-01
-3.03940326e-01 1.79242894e-01 3.31809521e-01 1.37730956e-01
5.12676120e-01 -3.46760601e-01 -6.64505959e-01 7.83060431e-01
-6.64330721e-01 1.16698250e-01 3.94464433e-01 1.02703011e+00
7.93068290e-01 3.78624313e-02 9.74267602e-01 -1.44001639e+00
3.37386757e-01 -5.96168756e-01 -4.97695416e-01 2.96222299e-01
-1.29236591e+00 -7.80816674e-02 7.60479033e-01 -6.15681052e-01
-1.17415047e+00 1.72949418e-01 -4.25375849e-02 -6.42495573e-01
1.74786337e-02 3.75852436e-01 -1.19092532e-01 8.34371001e-02
7.90646911e-01 -1.94340512e-01 -1.35059461e-01 -4.02529567e-01
1.72116205e-01 8.62140477e-01 2.37137094e-01 -5.79449892e-01
6.94041848e-01 3.00567806e-01 -5.11049390e-01 -3.22311372e-02
-1.59349716e+00 -4.51779813e-01 -3.89053851e-01 -1.61141142e-01
3.58249575e-01 -1.22685504e+00 -5.92971385e-01 5.20425916e-01
-4.71581638e-01 -6.27910793e-01 -2.27193773e-01 5.70690632e-01
-3.13799620e-01 1.60912305e-01 -6.92358494e-01 -7.76375473e-01
-6.00111961e-01 -9.57121372e-01 5.47060788e-01 4.16129082e-01
-5.42493388e-02 -7.90478051e-01 -2.87680268e-01 5.80969214e-01
4.34164554e-01 -1.46453083e-01 6.73052073e-01 -9.59670722e-01
-2.43424103e-01 -1.65225700e-01 -5.40254474e-01 6.93334639e-01
-1.19205164e-02 -1.05401561e-01 -1.02637625e+00 -4.81389493e-01
-1.42149359e-01 -7.99329281e-01 9.11054552e-01 4.66176540e-01
1.38369691e+00 -2.77701467e-01 -7.99893364e-02 6.24057889e-01
1.45691645e+00 8.12412351e-02 3.66585225e-01 2.69298702e-01
6.84962749e-01 4.40181285e-01 1.11261857e+00 6.22964144e-01
3.99609983e-01 4.58043188e-01 4.42216277e-01 -1.79208264e-01
-5.46396244e-03 -2.08676651e-01 1.15646459e-01 8.81069481e-01
2.04373270e-01 -1.13938697e-01 -9.73628581e-01 1.83483392e-01
-1.97423077e+00 -7.10163891e-01 -1.51389435e-01 2.42714429e+00
1.10009444e+00 4.58910763e-01 8.37702230e-02 2.52569705e-01
9.34544563e-01 -2.34152302e-02 -7.93752193e-01 -1.07215628e-01
-9.28158164e-02 -1.28104940e-01 7.26900995e-01 3.82272661e-01
-1.21406698e+00 9.88845348e-01 5.81226397e+00 1.22581482e+00
-1.07219660e+00 2.35695988e-01 1.39084744e+00 -1.77898958e-01
-1.33483410e-02 -1.11051008e-01 -1.10812616e+00 6.41223788e-01
9.11060929e-01 -1.51168406e-01 2.02045977e-01 9.35042143e-01
2.76619375e-01 -1.23714559e-01 -8.79230678e-01 1.00704753e+00
-6.85401633e-02 -1.05695379e+00 -3.01931836e-02 -1.17628716e-01
8.00117493e-01 2.34045357e-01 7.29242712e-02 5.01500309e-01
3.39967906e-01 -7.87385046e-01 7.47907341e-01 2.53799289e-01
7.52495885e-01 -7.98630178e-01 7.43896723e-01 5.87546766e-01
-9.18678403e-01 -2.05269918e-01 -5.44163585e-01 -1.06333494e-01
-1.79250628e-01 9.65438664e-01 -5.75684786e-01 1.99050933e-01
7.14041889e-01 6.54605091e-01 -6.23744905e-01 1.08583796e+00
-2.38520458e-01 1.24384427e+00 -2.91579127e-01 1.48063883e-01
1.76436290e-01 -1.99649736e-01 7.42535964e-02 1.11701810e+00
-3.90023738e-02 1.40487283e-01 4.27229196e-01 5.81905365e-01
-2.37258554e-01 3.15790474e-01 -2.47681085e-02 3.30791324e-01
7.37146318e-01 1.20371318e+00 -7.38912225e-01 -5.02536535e-01
-2.54733950e-01 4.41312462e-01 5.68565249e-01 2.67286181e-01
-1.08449125e+00 -1.49365336e-01 2.66177148e-01 1.54094458e-01
7.76171610e-02 2.04597443e-01 -5.76325893e-01 -1.03920674e+00
-1.27332825e-02 -8.84243369e-01 5.76320112e-01 -5.66016614e-01
-1.62455904e+00 6.47969961e-01 -1.40381142e-01 -1.49299383e+00
1.75820038e-01 -2.34494716e-01 -4.14692312e-01 5.69701791e-01
-1.63126850e+00 -7.80970633e-01 -2.44133830e-01 4.50410485e-01
5.82777739e-01 -2.57878751e-01 5.44032216e-01 6.03069663e-01
-8.90778244e-01 1.05269802e+00 1.59004018e-01 1.66697741e-01
1.07006991e+00 -1.04693031e+00 3.71170491e-02 6.72728062e-01
-2.76636295e-02 4.47890252e-01 5.50035059e-01 -3.76442432e-01
-6.24238789e-01 -1.30444109e+00 8.60535085e-01 -1.88306436e-01
4.76251662e-01 -1.92658573e-01 -1.13862789e+00 4.96123493e-01
-1.59496412e-01 5.39929211e-01 9.10932243e-01 1.31975368e-01
-5.54082394e-01 -4.12090868e-01 -1.19334483e+00 3.64484698e-01
9.34416175e-01 -2.28547439e-01 -2.51720071e-01 4.46408927e-01
5.48585951e-01 -4.33671236e-01 -8.11807513e-01 7.92220831e-01
5.02304554e-01 -1.02180886e+00 7.09131181e-01 -4.25760329e-01
4.63528067e-01 -2.41921797e-01 -2.10545510e-01 -1.09730732e+00
-4.01710361e-01 -2.60706335e-01 8.86926949e-02 1.43862593e+00
6.34954751e-01 -7.38160610e-01 1.02483273e+00 7.17361629e-01
2.33249329e-02 -1.19366145e+00 -7.91570187e-01 -8.39263797e-01
5.41527234e-02 -4.41827565e-01 3.65024239e-01 1.27168739e+00
-2.16415510e-01 3.42480958e-01 -4.99537289e-01 7.21207559e-02
9.24012244e-01 1.06640778e-01 4.74075258e-01 -1.28657019e+00
-2.13238224e-01 -2.69484758e-01 6.01903796e-02 -7.96009839e-01
3.40078324e-01 -9.39546108e-01 8.26180205e-02 -9.47878182e-01
4.68424529e-01 -1.11269307e+00 -9.03501391e-01 9.44134533e-01
-5.56117415e-01 5.48022866e-01 2.46588349e-01 6.12958491e-01
-1.02796137e+00 5.84875703e-01 1.15787697e+00 1.24903895e-01
-2.51469940e-01 2.63137549e-01 -9.26591933e-01 7.74717927e-01
1.24707544e+00 -8.24090004e-01 -5.33138871e-01 -1.09894671e-01
1.13064162e-01 -3.14541519e-01 1.18485995e-01 -8.62821817e-01
8.29085186e-02 -2.42339090e-01 3.01044852e-01 -2.98772663e-01
-1.59778982e-01 -5.72267294e-01 -1.96717083e-01 4.42959666e-01
-7.56805778e-01 -3.29053402e-01 4.32781279e-02 5.36154926e-01
-2.97977895e-01 -2.14724660e-01 1.20858479e+00 1.66209012e-01
-5.23331106e-01 4.92808402e-01 -1.45244393e-02 3.83203894e-01
8.90854418e-01 -4.44502272e-02 -2.74584651e-01 -1.93112820e-01
-7.27161884e-01 5.40086567e-01 3.30766410e-01 2.94138342e-01
3.25076193e-01 -1.34081566e+00 -7.00906456e-01 1.39110953e-01
2.50413418e-01 -1.03098780e-01 1.63165554e-01 8.94832015e-01
-1.26984343e-01 1.65353522e-01 -1.27081484e-01 -6.52947962e-01
-1.13438213e+00 2.95456946e-01 3.24528486e-01 -4.72474039e-01
-2.20585421e-01 9.26466107e-01 2.56565839e-01 -5.90723217e-01
4.70497102e-01 3.61406915e-02 -3.35584879e-02 -3.74154784e-02
3.64052594e-01 3.41810197e-01 8.53621811e-02 -3.97603929e-01
-4.63389933e-01 3.47581059e-01 -3.13989401e-01 3.58272165e-01
9.89622712e-01 -2.73676425e-01 2.44315967e-01 4.59208548e-01
9.85706031e-01 -1.69303194e-01 -1.61798429e+00 -6.01312220e-01
7.89019186e-03 -4.58985537e-01 2.07870737e-01 -1.00023544e+00
-1.36941898e+00 5.66112399e-01 7.73807466e-01 -3.32082212e-02
1.19314981e+00 -1.73970968e-01 5.98901451e-01 1.80054128e-01
3.37292522e-01 -1.27010846e+00 2.10799202e-02 4.09217805e-01
4.48741019e-01 -1.71054840e+00 1.11142598e-01 -5.19307971e-01
-9.69454765e-01 6.46753490e-01 9.26838815e-01 -1.00648321e-01
7.62319803e-01 1.99464694e-01 4.06713665e-01 1.24358378e-01
-9.57185268e-01 -1.46535356e-02 2.59208471e-01 1.24352723e-01
5.63364744e-01 1.72678009e-01 -4.71623659e-01 7.73470640e-01
3.14936601e-02 1.02879323e-01 2.39834741e-01 6.78533494e-01
-5.06016314e-01 -1.01326597e+00 -2.38485888e-01 7.95053303e-01
-6.00974441e-01 -1.62398994e-01 6.81160465e-02 5.78935683e-01
1.62326962e-01 9.43279266e-01 -8.24010521e-02 -3.26930523e-01
1.74912497e-01 8.66589472e-02 7.84564093e-02 -6.23278558e-01
-5.41472197e-01 1.27214551e-01 -5.66232055e-02 -4.79396969e-01
-6.43194795e-01 -4.80880916e-01 -1.45628548e+00 -1.69835627e-01
-6.89234078e-01 2.95840055e-01 3.78743052e-01 8.59466493e-01
3.51610541e-01 2.64106572e-01 9.40076888e-01 -3.99807811e-01
-9.27558124e-01 -8.70715022e-01 -6.71321034e-01 4.95990545e-01
-6.07970282e-02 -7.44152606e-01 -6.41595960e-01 4.89293560e-02] | [9.3927583694458, 3.7916150093078613] |
5b3fbc72-15a9-4b9a-a778-122d8e229047 | emotionflow-capture-the-dialogue-level | null | null | https://github.com/fpcsong/emotionflow/blob/master/EmotionFlow.pdf | https://github.com/fpcsong/emotionflow/blob/master/EmotionFlow.pdf | EmotionFlow: Capture the Dialogue Level Emotion Transitions | Emotion recognition in conversations (ERC) has attracted increasing interests in recent years, due to its wide range of applications, such as customer service analysis, health-care consultation, etc. One key challenge of ERC is that users' emotions would change due to the impact of others' emotions. That is, the emotions within the conversation can spread among the communication participants. However, the spread impact of emotions in a conversation is rarely addressed in existing researches. To this end, we propose \textbf{EmotionFlow} for ERC with the consideration of the spread of participants' emotions during a conversation. EmotionFlow first encodes users' utterance by concatenating the context with an auxiliary question, which helps to learn user-specific features. Then, conditional random field is applied to capture the sequential information at emotional level. We conduct extensive experiments on a public dataset Multimodal EmotionLines Dataset (MELD), and the results demonstrate the effectiveness of our proposed model. | ['Longtao Huang', 'Songlin Hu', 'Rong Zhang', 'Liangjun Zang', 'Xiaohui Song'] | 2022-05-07 | null | null | null | icassp-2022-5 | ['emotion-recognition-in-conversation'] | ['natural-language-processing'] | [ 1.82336923e-02 -4.01226342e-01 1.92394316e-01 -7.44076610e-01
-3.09345573e-01 -2.86395073e-01 3.34815353e-01 -5.51400855e-02
-1.61510095e-01 5.91712117e-01 7.46415854e-01 2.32816726e-01
1.92743748e-01 -4.84200746e-01 -1.71181753e-01 -6.83961570e-01
1.23635128e-01 -1.47142127e-01 -2.80623376e-01 -3.75958055e-01
3.34235914e-02 7.79084489e-02 -1.43062258e+00 6.47939861e-01
8.82696033e-01 1.16193056e+00 -9.25820088e-04 4.33346212e-01
-5.98798513e-01 8.26274335e-01 -8.39551091e-01 -5.42320788e-01
-6.98312581e-01 -6.80838704e-01 -6.23539090e-01 3.50810379e-01
-8.69287848e-01 8.71248990e-02 -5.27113825e-02 8.97906184e-01
5.88006079e-01 4.45464402e-01 4.21159506e-01 -1.38286901e+00
-3.30583513e-01 5.65461814e-01 -5.57749867e-01 -1.26561403e-01
5.28034806e-01 -3.69565636e-02 7.20999360e-01 -6.74966931e-01
3.44985098e-01 1.63178313e+00 3.36009711e-01 7.41643429e-01
-5.69648087e-01 -8.62932742e-01 4.69462842e-01 3.82670611e-01
-1.03599095e+00 -3.21082950e-01 1.04798615e+00 -2.09542856e-01
6.51295900e-01 3.77007276e-01 5.19660175e-01 1.28416014e+00
4.80495580e-02 1.16627848e+00 9.90003228e-01 -1.45338491e-01
2.01868951e-01 4.00191545e-01 1.73651874e-01 -1.01061189e-03
-6.61569953e-01 -5.19738138e-01 -4.80093986e-01 -1.38971344e-01
-1.11112565e-01 1.29083782e-01 -4.27654773e-01 4.87822771e-01
-8.58395755e-01 6.83674097e-01 1.95725560e-01 3.83781701e-01
-4.60988641e-01 -1.69546962e-01 7.89101541e-01 2.41692379e-01
7.21551478e-01 -2.67590642e-01 -5.27007163e-01 -9.04557943e-01
-2.05879390e-01 -7.11799785e-02 7.94304907e-01 8.24597001e-01
5.16811371e-01 -4.62573916e-01 5.04900031e-02 1.16718650e+00
2.93186843e-01 3.64287257e-01 5.67998827e-01 -6.11779511e-01
4.57522810e-01 6.92889750e-01 7.85889030e-02 -1.59077621e+00
-4.75825489e-01 -1.51288733e-01 -1.21064961e+00 -7.48101115e-01
-3.24171990e-01 -8.22457135e-01 -4.82809171e-02 1.83572423e+00
4.73292232e-01 1.79296032e-01 3.54837537e-01 7.75274575e-01
1.01683700e+00 1.11591458e+00 1.47753805e-01 -6.89435244e-01
1.41090012e+00 -6.46152794e-01 -1.13083017e+00 4.47849669e-02
5.52688181e-01 -1.00964296e+00 9.16362166e-01 5.03502667e-01
-5.87912858e-01 -3.95383298e-01 -3.48160237e-01 3.00389081e-01
-1.77075565e-01 -3.73387262e-02 7.37130523e-01 6.06711686e-01
-4.27174151e-01 -8.16106722e-02 -3.80810231e-01 -3.35115820e-01
2.07691133e-01 8.33375752e-02 -1.66488774e-02 4.23959866e-02
-1.60230720e+00 4.55768973e-01 2.85259709e-02 7.10396469e-01
-8.08159262e-02 -1.03257343e-01 -6.81910038e-01 1.40152663e-01
3.07928771e-01 1.01671065e-03 1.29426932e+00 -1.34689474e+00
-1.73752153e+00 3.75168979e-01 -7.32828736e-01 2.79321134e-01
3.08868021e-01 -9.88395438e-02 -9.93964314e-01 -2.26994246e-01
-2.43878573e-01 3.16414684e-01 3.90716583e-01 -1.15689051e+00
-7.04627931e-01 -3.75505000e-01 -2.35395506e-01 4.39723521e-01
-4.82167155e-01 3.15244019e-01 -7.68150270e-01 -3.28130722e-01
-2.93081015e-01 -9.09650087e-01 -5.76216988e-02 -7.33706832e-01
-4.76918995e-01 -5.99092424e-01 8.85995090e-01 -5.06544173e-01
1.50477862e+00 -2.56084228e+00 2.24058196e-01 1.60236657e-01
-2.38052964e-01 -2.12428629e-01 -1.09921299e-01 5.43265522e-01
-1.93909165e-02 2.87386924e-01 -3.45886275e-02 -4.47934598e-01
6.68786988e-02 2.99199194e-01 -3.70110691e-01 -3.52582112e-02
1.15817778e-01 6.47209704e-01 -7.05011606e-01 -5.50386786e-01
8.79191421e-03 4.21493620e-01 -3.79711121e-01 4.36681926e-01
-2.07418337e-01 6.84938550e-01 -7.08216310e-01 4.13886040e-01
8.46062839e-01 -3.83589640e-02 1.06917575e-01 -2.30610684e-01
-2.32779950e-01 -1.21717751e-01 -9.89584386e-01 1.32036126e+00
-6.86662376e-01 4.42200333e-01 1.76978990e-01 -9.20276701e-01
1.00784600e+00 4.91191417e-01 5.64425766e-01 -6.18512869e-01
5.48844934e-01 -1.01383731e-01 -5.38799390e-02 -8.22970450e-01
4.70787257e-01 -2.89928228e-01 -7.16205359e-01 5.75681686e-01
-2.66506225e-01 1.64042741e-01 -2.14031681e-01 1.69605017e-01
7.46606529e-01 -3.96313906e-01 9.83732939e-02 3.51065516e-01
7.78399825e-01 -3.66961777e-01 7.97049940e-01 8.68695825e-02
-3.93671989e-01 1.77001551e-01 6.89308345e-01 -1.13003537e-01
-1.30198687e-01 -2.68211365e-01 1.06776230e-01 1.14290714e+00
4.13655519e-01 -2.13322118e-01 -6.16631567e-01 -6.43072546e-01
-3.56059492e-01 7.49396265e-01 -6.54921114e-01 -3.73627722e-01
-1.51739806e-01 -8.75537157e-01 2.93907195e-01 2.40421146e-02
6.44770861e-01 -1.67244852e+00 -2.00588360e-01 4.12576795e-01
-8.90402853e-01 -1.06450140e+00 -5.70999205e-01 -1.97344527e-01
-3.54638517e-01 -7.25605845e-01 -5.76009393e-01 -6.64529979e-01
3.77268434e-01 8.27455074e-02 7.60295331e-01 -9.00745690e-02
-1.79153994e-01 4.43732291e-01 -7.74881065e-01 -3.53904128e-01
-1.54885665e-01 1.23445511e-01 -1.77394733e-01 6.14886761e-01
5.98464608e-01 -3.15288365e-01 -5.65217674e-01 4.53944892e-01
-9.86002803e-01 -5.71036376e-02 2.35693142e-01 5.87168097e-01
1.23238795e-01 5.01445651e-01 1.10237205e+00 -8.25650334e-01
1.26400697e+00 -9.18897271e-01 2.15973333e-01 3.12550485e-01
5.63411694e-03 -4.60430086e-01 5.45464456e-01 -6.30762517e-01
-1.70304501e+00 -3.13798219e-01 -5.10822177e-01 8.33734125e-02
-3.19267958e-01 6.93421841e-01 -6.49808049e-01 6.64992988e-01
-2.20707729e-01 1.63453713e-01 -1.59204289e-01 -2.47345209e-01
2.68554717e-01 1.30794358e+00 2.16209576e-01 -6.10948503e-01
2.48726979e-02 1.47232115e-01 -6.40698135e-01 -7.51036108e-01
-7.13694274e-01 -4.85569775e-01 1.17151685e-01 -6.41214728e-01
8.94680202e-01 -6.97422981e-01 -1.29726231e+00 6.98536694e-01
-1.51561928e+00 8.56147557e-02 3.10453653e-01 5.19046247e-01
-9.82639939e-02 2.78343707e-01 -6.75940931e-01 -1.12906814e+00
-3.04891914e-01 -1.23708975e+00 8.78225684e-01 8.54931772e-01
-3.91257554e-01 -9.95190561e-01 -6.10759333e-02 2.56236583e-01
2.82848746e-01 7.92168453e-02 8.24156463e-01 -5.21552324e-01
-2.61795651e-02 -6.31726161e-02 -9.90689546e-02 3.49604756e-01
5.74587345e-01 7.92102888e-03 -7.56888211e-01 1.97933555e-01
7.00001195e-02 -4.34162110e-01 4.69834358e-01 1.75867025e-02
1.30364466e+00 -3.23461324e-01 -2.63251513e-01 -9.29822493e-03
7.96653450e-01 7.04014063e-01 7.00552344e-01 -2.27143407e-01
2.91915685e-01 1.00345182e+00 7.76509523e-01 9.08133864e-01
6.63201213e-01 2.90072322e-01 3.14553559e-01 1.66575164e-02
7.07063198e-01 -2.46334057e-02 4.12526101e-01 1.31643486e+00
1.69818208e-01 -6.52259171e-01 -5.20825505e-01 3.37063730e-01
-1.97655296e+00 -1.04953492e+00 -1.09176971e-01 1.48803461e+00
9.24320459e-01 -1.63309589e-01 -2.55167365e-01 -1.58703387e-01
9.75991189e-01 2.03490496e-01 -6.70025766e-01 -7.10817635e-01
3.62007320e-02 -1.40418768e-01 -4.43088859e-01 2.28102952e-01
-6.04965925e-01 7.41626143e-01 4.67980719e+00 7.69924343e-01
-1.07981169e+00 3.49215604e-02 1.02136147e+00 -4.08752225e-02
-4.85671431e-01 -2.18469024e-01 -5.81321895e-01 9.40996408e-01
5.00189424e-01 -3.27655524e-02 5.23190439e-01 5.62836349e-01
4.99679327e-01 -1.18159734e-01 -5.10436594e-01 1.23872590e+00
2.69703537e-01 -6.17346048e-01 -2.77862638e-01 -2.21191838e-01
5.86654305e-01 -4.47938919e-01 5.12691168e-03 7.71379769e-01
5.54958321e-02 -7.49184251e-01 7.71919712e-02 8.82822454e-01
5.27042508e-01 -1.30251884e+00 9.82091546e-01 5.58732569e-01
-1.19706261e+00 1.22894146e-01 -7.83820227e-02 -1.24785453e-02
5.10346830e-01 7.99665213e-01 -4.02149498e-01 4.73950714e-01
7.86758959e-01 7.44951189e-01 -6.75702989e-02 3.97386700e-01
-3.83361056e-02 7.85620272e-01 -1.46429926e-01 -6.00789666e-01
7.15149567e-02 -4.18528259e-01 2.49482289e-01 1.36135626e+00
3.25882912e-01 5.74284136e-01 8.77955332e-02 3.66727889e-01
-3.86346102e-01 5.71330905e-01 -1.21855699e-01 -2.13922963e-01
4.52667356e-01 1.44936454e+00 -4.10387099e-01 -1.45271793e-01
-3.85307997e-01 1.32190776e+00 1.33353453e-02 4.54523146e-01
-1.04367912e+00 -6.12288654e-01 4.90747720e-01 -6.88166320e-01
8.91624019e-02 3.81766915e-01 2.05624253e-01 -1.20868647e+00
3.48236263e-02 -9.22623694e-01 2.32007638e-01 -7.75151253e-01
-1.51236165e+00 5.66880941e-01 -5.37321270e-01 -9.76194084e-01
-4.05244343e-02 -8.27567801e-02 -1.05173993e+00 9.20715988e-01
-1.16881061e+00 -7.12919891e-01 -5.41222274e-01 6.13350928e-01
5.57330310e-01 7.28352368e-02 8.15927327e-01 3.25276554e-01
-9.32027876e-01 5.52081108e-01 -1.12187915e-01 1.13597110e-01
1.01005733e+00 -6.23708844e-01 -2.65335590e-01 2.41306379e-01
-6.36746764e-01 6.25658810e-01 5.11679530e-01 -4.94141847e-01
-1.14810371e+00 -9.36551690e-01 1.16538703e+00 1.92334548e-01
5.30970991e-01 -2.69391656e-01 -8.50983679e-01 3.57830614e-01
5.24248064e-01 -2.15264872e-01 1.11007559e+00 4.52266783e-01
7.14898854e-02 -1.56235173e-01 -9.83235717e-01 7.72847235e-01
7.18488216e-01 -5.04384816e-01 -4.02197480e-01 -5.37804700e-02
8.45642984e-01 -1.10437147e-01 -8.56806159e-01 3.41322154e-01
5.78803480e-01 -8.71323109e-01 3.32342565e-01 -5.04710317e-01
5.87932169e-01 -1.38297770e-02 -1.46735236e-01 -1.58916855e+00
1.22989081e-01 -5.99316597e-01 3.53336006e-01 1.94266748e+00
1.03134818e-01 -7.11225390e-01 3.26595038e-01 7.74023056e-01
4.37289923e-02 -7.66227543e-01 -7.06625283e-01 2.32854813e-01
-3.09437692e-01 -6.35075092e-01 9.24287856e-01 1.00996232e+00
6.32323563e-01 6.23001039e-01 -6.03278697e-01 3.23194601e-02
-1.40373006e-01 2.42560610e-01 7.95001805e-01 -8.82186949e-01
-1.29980549e-01 -4.13367480e-01 -3.23591381e-02 -1.19721043e+00
4.93268818e-01 -4.28262830e-01 3.95495832e-01 -1.34447265e+00
2.58014351e-01 -5.26684701e-01 -1.43287197e-01 2.58048087e-01
-5.02014339e-01 -3.62582505e-01 1.77662577e-02 -6.89041913e-02
-1.00702739e+00 1.10648429e+00 1.24161148e+00 5.23111364e-03
-3.40032041e-01 1.21727087e-01 -7.48908341e-01 6.50231242e-01
9.03554916e-01 -2.20445827e-01 -3.95060629e-01 2.16201767e-02
4.01231587e-01 4.58422244e-01 -9.13683921e-02 -4.87434745e-01
1.44654632e-01 -2.97155946e-01 1.36714682e-01 -7.37480998e-01
3.71311754e-01 -8.67014587e-01 -4.24020551e-02 5.97692877e-02
-4.22047824e-01 8.09518695e-02 1.61420226e-01 5.98157406e-01
-6.40472472e-01 -2.32375544e-02 3.19328427e-01 2.67219935e-02
-5.23857892e-01 3.04161221e-01 -7.22602546e-01 1.84892014e-01
8.16119015e-01 3.51411164e-01 -1.69345498e-01 -8.43949080e-01
-6.61688507e-01 6.64265871e-01 -1.39493896e-02 7.12576747e-01
5.41253626e-01 -1.32126832e+00 -5.52469254e-01 -2.89187450e-02
2.78454006e-01 3.20131779e-02 1.05982447e+00 7.87535906e-01
2.88776517e-01 -5.30328602e-02 2.81074554e-01 -3.88888419e-01
-1.40762365e+00 1.74922764e-01 2.08869472e-01 -2.59790719e-01
-5.07121794e-02 6.99085474e-01 1.33295655e-01 -7.92810023e-01
3.91776919e-01 -1.19389273e-01 -4.91182208e-01 5.19214809e-01
5.74186981e-01 9.09313411e-02 -2.48489991e-01 -7.72173941e-01
-4.21129853e-01 2.45540604e-01 -1.13019153e-01 -9.28018615e-02
1.12319875e+00 -6.21280015e-01 -3.69627327e-01 8.42558801e-01
1.42859578e+00 1.72826782e-01 -7.66524374e-01 -3.29337418e-01
-2.48491496e-01 -1.31899983e-01 -2.27858543e-01 -8.22499037e-01
-9.33372378e-01 8.90741527e-01 3.61412495e-01 2.86928415e-01
1.32628334e+00 3.22032012e-02 1.12470245e+00 1.86639652e-01
1.21450990e-01 -1.23359466e+00 1.54852048e-01 7.01938927e-01
7.65281022e-01 -1.11841869e+00 -5.63059807e-01 -2.60646492e-01
-1.19170451e+00 9.82224405e-01 4.59006995e-01 4.10647601e-01
8.60887527e-01 1.44709468e-01 2.93234706e-01 8.88049304e-02
-8.96023631e-01 -6.67328655e-04 -1.25646695e-01 2.08572552e-01
4.75621283e-01 2.30365798e-01 -4.82905805e-01 1.18895555e+00
-2.51682609e-01 7.63131604e-02 1.99726045e-01 6.63735211e-01
-1.59880221e-01 -1.06215465e+00 -2.28594825e-01 1.86070293e-01
-2.96617538e-01 3.10225952e-02 -4.07666534e-01 3.22136223e-01
4.10178691e-01 1.48241043e+00 1.13403961e-01 -5.70282161e-01
3.39291364e-01 1.89784318e-01 -2.38148347e-01 -2.64216453e-01
-5.78732729e-01 2.10605264e-01 2.52705723e-01 -2.59159684e-01
-7.54103899e-01 -8.14612091e-01 -1.48274922e+00 -2.34960124e-01
-2.30667666e-01 5.54736733e-01 5.80568433e-01 1.09958684e+00
4.68012691e-01 7.48413563e-01 1.33211374e+00 -3.27502072e-01
-3.92151661e-02 -1.13242757e+00 -5.15547276e-01 8.17553639e-01
8.18372890e-02 -4.22812045e-01 -6.26295388e-01 -6.43149987e-02] | [13.024555206298828, 6.074990749359131] |
8700c64a-3de6-48fd-bb1c-6a6c83234029 | procedural-urban-environments-for-fps-games | 1604.05791 | null | http://arxiv.org/abs/1604.05791v1 | http://arxiv.org/pdf/1604.05791v1.pdf | Procedural urban environments for FPS games | This paper presents a novel approach to procedural generation of urban maps
for First Person Shooter (FPS) games. A multi-agent evolutionary system is
employed to place streets, buildings and other items inside the Unity3D game
engine, resulting in playable video game levels. A computational agent is
trained using machine learning techniques to capture the intent of the game
designer as part of the multi-agent system, and to enable a semi-automated
aesthetic selection for the underlying genetic algorithm. | ['Ricardo Sosa', 'Jan Kruse', 'Andy M. Connor'] | 2016-04-20 | null | null | null | null | ['fps-games'] | ['playing-games'] | [ 1.16320610e-01 3.25243056e-01 3.81476730e-01 2.55099505e-01
-2.41762906e-01 -5.05069375e-01 5.38151503e-01 -1.43903077e-01
-1.01602428e-01 5.72230041e-01 -7.80578330e-02 -3.59596670e-01
-2.45750576e-01 -1.52009583e+00 -1.53699353e-01 -3.81705523e-01
-1.19506316e-02 1.00696564e+00 5.45554638e-01 -8.08283746e-01
4.02093232e-01 5.03701270e-01 -1.94980824e+00 2.91353375e-01
7.71121025e-01 5.20160675e-01 5.05782485e-01 1.26166940e+00
-4.98744324e-02 1.03484857e+00 -9.13876534e-01 -5.00736952e-01
4.47538286e-01 -7.05792785e-01 -5.08623898e-01 3.08223248e-01
-5.59449673e-01 2.44415522e-01 2.09501162e-01 7.25661933e-01
5.51090717e-01 2.54568756e-01 9.33498204e-01 -1.47577214e+00
1.49375677e-01 4.19634461e-01 -2.70236880e-01 -2.05876365e-01
8.17079484e-01 3.32037449e-01 8.30264568e-01 -4.77751881e-01
7.62416303e-01 1.10513449e+00 3.52790564e-01 2.65428454e-01
-7.52325535e-01 -2.84375846e-01 -4.70331401e-01 1.27884403e-01
-1.69650161e+00 1.68844452e-03 8.94343495e-01 -5.08854687e-01
1.12088466e+00 6.87353313e-01 1.67963874e+00 6.41738296e-01
5.24285972e-01 4.41323489e-01 1.07021391e+00 -6.45357192e-01
8.96747351e-01 1.26977593e-01 -7.57770717e-01 1.09620380e+00
-9.54659209e-02 3.59769285e-01 -4.74496245e-01 -3.42489511e-01
1.05649829e+00 -6.83811426e-01 5.44417739e-01 -4.14968491e-01
-3.50994825e-01 1.12945318e+00 2.76252836e-01 1.65435329e-01
-8.87812495e-01 -1.71556976e-03 1.08466208e-01 6.16307519e-02
3.45818549e-01 8.70306313e-01 -8.60373750e-02 -4.82152343e-01
-7.45585680e-01 9.86394942e-01 9.62601185e-01 7.19187677e-01
5.96416056e-01 2.14683235e-01 1.16639547e-01 4.97643054e-01
4.56741631e-01 4.38578166e-02 3.37080270e-01 -9.10272717e-01
-3.96678187e-02 1.04700303e+00 -9.00269896e-02 -1.03074801e+00
-3.84387106e-01 -8.68683830e-02 -3.15864205e-01 1.13638461e+00
-2.44643524e-01 -4.72459525e-01 -7.10643530e-01 6.75738573e-01
5.65389693e-01 4.61076237e-02 -2.31356714e-02 6.67015672e-01
6.48553908e-01 3.72908831e-01 1.04157746e-01 3.53833795e-01
1.24105716e+00 -8.97379398e-01 -1.36284381e-01 -3.92228812e-01
3.59134644e-01 -4.16866601e-01 6.00960851e-01 4.51901525e-01
-1.36995733e+00 -2.51894593e-01 -1.18144643e+00 5.89174330e-01
-5.49497783e-01 -1.92416832e-01 8.10181618e-01 1.07755041e+00
-1.20804858e+00 2.43332058e-01 -2.44132727e-01 -1.32211357e-01
3.26320857e-01 3.62678885e-01 3.23148936e-01 3.72774631e-01
-1.00249875e+00 1.30764604e+00 7.62011528e-01 -4.16706830e-01
-6.69442952e-01 -4.09528494e-01 -6.85629606e-01 -5.23459464e-02
1.36659518e-01 -1.04272509e+00 8.79100263e-01 -1.08200717e+00
-2.08984327e+00 1.15298676e+00 5.10646760e-01 -3.34689707e-01
7.94843793e-01 5.05697548e-01 -1.11779206e-01 -1.02822617e-01
2.42020842e-02 5.19303799e-01 7.30614662e-01 -1.37149155e+00
-1.10596943e+00 4.40084003e-02 3.24417859e-01 8.88719380e-01
5.20776212e-01 1.91731602e-01 -2.22239897e-01 -4.77497399e-01
-3.02100778e-01 -7.05488384e-01 -8.20762157e-01 -5.08844018e-01
-1.95932180e-01 -7.02107251e-02 4.76604372e-01 -5.29234469e-01
9.88693833e-01 -1.63254857e+00 5.69324791e-01 7.26904750e-01
1.73830017e-02 5.85971214e-02 3.22809488e-01 4.84377384e-01
1.85201958e-01 -1.43442407e-01 3.82913859e-03 1.02795884e-01
8.78837332e-02 1.11489393e-01 3.91017109e-01 -2.05352321e-01
-1.21985711e-01 9.56830263e-01 -8.01283717e-01 -5.71404099e-01
6.38059437e-01 3.38884331e-02 -5.72862208e-01 6.86203539e-02
-5.96124709e-01 1.23829074e-01 -6.03527486e-01 6.41598046e-01
2.18658060e-01 3.53020728e-01 2.40405381e-01 6.00217342e-01
-3.74857754e-01 -3.09091303e-02 -1.33206165e+00 1.57654524e+00
-6.52428746e-01 2.98091739e-01 1.38450200e-02 -4.90742505e-01
1.19451249e+00 2.17076674e-01 6.54124975e-01 -7.63877690e-01
4.52379823e-01 2.19389006e-01 -1.13319024e-01 -3.40827048e-01
9.68831122e-01 -1.51020601e-01 -3.08987498e-01 6.30776346e-01
-3.90047878e-01 -9.41926897e-01 2.86336035e-01 -2.61285514e-01
1.21803832e+00 4.66602087e-01 7.25388646e-01 -2.91892380e-01
1.95979491e-01 6.13542795e-01 -5.50088100e-02 6.00236177e-01
2.02259749e-01 3.86855096e-01 5.45269959e-02 -6.69570506e-01
-1.22108769e+00 -9.76398885e-01 6.21980965e-01 1.14317143e+00
1.12608045e-01 -2.38585100e-01 -9.27434206e-01 -1.31705087e-02
-3.49066406e-01 9.55494344e-01 -3.74461830e-01 -2.13433698e-01
-4.30610776e-01 -8.98023486e-01 4.44218069e-01 -1.20028630e-01
3.48452359e-01 -1.27331066e+00 -1.39676917e+00 8.65918815e-01
8.35191384e-02 -5.38795173e-01 1.75131544e-01 9.52424332e-02
-3.81855041e-01 -9.28862453e-01 -2.35738710e-01 -7.05990136e-01
3.92357528e-01 -3.17064375e-01 1.23489583e+00 3.72585684e-01
-5.15848875e-01 2.04224735e-01 -4.64264303e-01 -8.90165687e-01
-8.98659527e-01 1.70020238e-01 -5.10311842e-01 -3.43739152e-01
-2.45851930e-02 -7.84406841e-01 -2.08255336e-01 2.44677171e-01
-6.62833393e-01 7.49049067e-01 2.87809134e-01 3.46242875e-01
2.02203751e-01 6.82926238e-01 -1.42290350e-02 -2.53074735e-01
1.10735190e+00 -2.53212392e-01 -7.92252779e-01 2.91111976e-01
-9.27206278e-02 -2.26458490e-01 2.35061496e-01 -1.71012610e-01
-8.75675440e-01 3.44896525e-01 -3.45200181e-01 1.71691269e-01
-1.95561543e-01 4.12432641e-01 -3.07850927e-01 -5.68728149e-01
8.48616004e-01 9.70359370e-02 -2.64101088e-01 2.09193096e-01
3.93910557e-01 5.47108412e-01 2.56845355e-01 -6.33907974e-01
7.67394543e-01 2.20243320e-01 2.11903807e-02 -8.88934672e-01
2.15024054e-01 -1.41963601e-01 -6.63199902e-01 -1.03622150e+00
9.34224784e-01 -6.39780760e-01 -7.71474242e-01 7.13022172e-01
-9.92090881e-01 -5.02163529e-01 -4.69776452e-01 -4.02759351e-02
-1.19739389e+00 -6.85814083e-01 -1.85761470e-02 -9.67992663e-01
-1.52558625e-01 -1.00112474e+00 7.21220374e-01 5.91349065e-01
-6.18881464e-01 -8.62771511e-01 4.38617289e-01 4.19736952e-01
5.01971543e-02 7.06845522e-01 9.39566135e-01 -2.92985052e-01
-8.03312778e-01 -1.16542175e-01 4.24759924e-01 -6.02613866e-01
-2.58126836e-02 2.93107510e-01 -5.63221574e-01 4.54185009e-01
-2.89673924e-01 8.95617157e-02 -1.08731866e-01 4.75056171e-01
4.76982266e-01 -2.31362402e-01 -4.41003144e-01 4.30811048e-01
1.62412095e+00 9.05151188e-01 1.03139269e+00 1.02691841e+00
3.30482483e-01 4.97340947e-01 6.84615552e-01 8.48415554e-01
5.93133926e-01 8.77873957e-01 5.86934865e-01 -6.46400824e-02
1.97660014e-01 -1.56249493e-01 -1.05463319e-01 1.00994103e-01
-6.86336577e-01 -5.47261834e-01 -1.22751415e+00 2.86133170e-01
-1.80683577e+00 -1.13908470e+00 1.09085113e-01 1.56367433e+00
3.35269034e-01 2.24134713e-01 5.58307171e-01 3.34382653e-01
7.03272760e-01 8.35835487e-02 -7.24644139e-02 -9.87877369e-01
-2.59345531e-01 5.59516490e-01 5.57389498e-01 5.24400413e-01
-7.63282776e-01 1.26010799e+00 7.05098486e+00 8.60663354e-01
-6.80422485e-01 -1.51018560e-01 4.63771045e-01 -1.46001697e-01
-4.25346524e-01 -2.05300093e-01 -2.46212691e-01 3.51827353e-01
7.25093305e-01 -7.99039006e-01 9.31080580e-01 8.42483521e-01
4.79591906e-01 -5.28004766e-01 -2.59493351e-01 7.17161834e-01
9.99982730e-02 -1.81018972e+00 -6.22920245e-02 3.16813976e-01
8.47377837e-01 -3.83361965e-01 -3.92158329e-02 -4.80532758e-02
1.17042553e+00 -1.02243257e+00 1.23377419e+00 5.25247276e-01
2.55130649e-01 -1.62181497e+00 2.59755492e-01 5.24758220e-01
-1.32052863e+00 -1.75885916e-01 -5.57566993e-02 -4.43300009e-01
5.70131242e-01 -2.77207702e-01 -1.20361137e+00 6.60721004e-01
5.09330332e-01 -2.18601972e-01 -4.93422538e-01 1.29256928e+00
-1.71009645e-01 -7.71587566e-02 -3.38631213e-01 -7.37601936e-01
4.67985034e-01 -5.90842128e-01 8.73054206e-01 6.56074345e-01
4.61226791e-01 5.02520025e-01 1.90682620e-01 7.38226831e-01
4.94583160e-01 2.54033983e-01 -7.63073742e-01 3.39255519e-02
2.68986553e-01 1.05459440e+00 -1.31572115e+00 -1.42668262e-01
3.07629049e-01 1.13838434e+00 8.85158330e-02 -1.10365517e-01
-8.36033702e-01 -1.67462155e-01 6.99628413e-01 2.92121917e-01
1.67019591e-01 -2.32162088e-01 -8.34234893e-01 -2.38021731e-01
-4.36133862e-01 -1.00328207e+00 4.44072261e-02 -1.20439339e+00
-4.32682395e-01 7.16814399e-01 -1.77807748e-01 -8.05165946e-01
-9.54453290e-01 -2.99587607e-01 -1.06336868e+00 7.67563641e-01
-3.94332796e-01 -1.39081466e+00 -4.37426120e-02 3.27273041e-01
5.81851065e-01 -8.49325597e-01 9.69745696e-01 -2.47059256e-01
-7.36744553e-02 -2.26460159e-01 2.12858412e-02 -2.40559369e-01
-4.85515893e-01 -1.06075227e+00 8.03082287e-01 8.50209773e-01
1.23094849e-01 -2.88928121e-01 9.79302704e-01 -1.02409077e+00
-9.90913332e-01 -7.49821305e-01 4.98860925e-01 -1.87400773e-01
5.28805315e-01 -2.42350683e-01 -1.37334587e-02 1.35412455e-01
3.17504019e-01 -1.02868593e+00 4.98977035e-01 -2.18071178e-01
7.34620631e-01 5.33903956e-01 -1.34539473e+00 1.05081165e+00
9.98917401e-01 -1.37614414e-01 -3.83234829e-01 1.21279269e-01
1.26953453e-01 -9.16885972e-01 -3.85461330e-01 -1.83495119e-01
4.89806265e-01 -1.13873780e+00 1.16239691e+00 -3.66489559e-01
5.87567449e-01 -1.89332500e-01 6.14266843e-02 -1.61375523e+00
-4.55998570e-01 -6.71524644e-01 2.93092579e-01 6.87386036e-01
4.15693402e-01 -2.30982482e-01 1.28554595e+00 6.48308754e-01
-1.59148782e-01 -4.67599899e-01 -1.07553744e+00 -2.25448996e-01
-3.23958963e-01 -4.00486469e-01 1.06840181e+00 4.67610300e-01
-2.12259844e-01 -3.11396867e-02 -3.67673755e-01 1.39441654e-01
5.08479536e-01 -9.59284827e-02 6.90103352e-01 -1.30092001e+00
-4.37874109e-01 -9.58895445e-01 -8.92672420e-01 -1.23590276e-01
-1.89632047e-02 -5.81880152e-01 1.86093006e-04 -1.75574577e+00
-4.21878427e-01 -6.01998329e-01 4.74845648e-01 1.01181798e-01
-3.69781069e-02 1.31369963e-01 2.34239697e-01 -4.03194785e-01
-3.21014941e-01 3.68673563e-01 1.04467273e+00 -9.10106748e-02
-6.48605108e-01 4.20502007e-01 -5.93667269e-01 9.49233294e-01
9.99151707e-01 -5.14182568e-01 -4.84884292e-01 -4.92935479e-02
6.60921156e-01 1.82969496e-01 4.70169216e-01 -1.26647818e+00
3.66115928e-01 -5.02206087e-01 1.15360141e-01 -5.51062047e-01
4.52968121e-01 -7.98733711e-01 9.92434204e-01 9.44454491e-01
2.42724985e-01 3.22863221e-01 1.70228258e-01 1.39131159e-01
6.95049986e-02 -4.64286506e-01 6.88970685e-01 -6.33202910e-01
-9.67484772e-01 -1.38893008e-01 -1.17036211e+00 -2.21342027e-01
1.62915540e+00 -9.46763933e-01 2.78469890e-01 -3.10410082e-01
-7.20146656e-01 2.51528900e-02 8.17015648e-01 4.31630880e-01
7.27243304e-01 -1.26264298e+00 -6.91063464e-01 2.07538590e-01
-2.50191092e-01 -1.37437806e-01 3.23481441e-01 -3.04932684e-01
-1.51945353e+00 -1.16799377e-01 -9.47559893e-01 -9.80344415e-02
-1.41057038e+00 1.53305335e-02 7.74266005e-01 -3.98506880e-01
-5.13721704e-01 9.80529964e-01 -4.72030729e-01 -4.57292706e-01
-3.26196164e-01 3.16561848e-01 -5.25963545e-01 7.99468011e-02
3.55537415e-01 7.10017562e-01 1.81331504e-02 -6.61852598e-01
-1.69564694e-01 2.43869707e-01 7.95716524e-01 -6.93766057e-01
1.52681029e+00 1.09864332e-01 1.53508618e-01 9.84366834e-02
1.07622460e-01 -2.87735641e-01 -1.07439590e+00 5.30599177e-01
3.00520975e-02 -4.00998294e-01 -1.44968126e-02 -9.37925637e-01
-6.94232106e-01 -3.31725702e-02 3.42211813e-01 2.74876863e-01
1.16115868e+00 -3.05540472e-01 2.25628093e-01 -1.09636830e-02
1.15358639e+00 -1.38742435e+00 -6.52831420e-03 3.36946517e-01
8.10586333e-01 -6.31915748e-01 1.42834627e-03 -3.47956449e-01
-1.01843476e+00 9.83611584e-01 4.73625273e-01 -2.12539718e-01
2.91463554e-01 7.40324974e-01 -2.63784155e-02 -6.43966854e-01
-6.37628973e-01 -2.32760489e-01 6.71796724e-02 1.29199386e+00
-3.41362298e-01 2.66351521e-01 -2.54438788e-01 3.51736844e-01
-9.65951860e-01 5.23619391e-02 5.77015579e-01 1.18032384e+00
-7.60041535e-01 -9.92219567e-01 -6.06873512e-01 2.88292915e-01
1.39586270e-01 5.23015186e-02 -6.99666142e-01 7.20438123e-01
6.90362871e-01 8.20651054e-01 5.49402386e-02 -4.08209234e-01
5.38195848e-01 -2.55314469e-01 6.12992644e-01 -5.59495687e-01
-1.26886737e+00 -2.05425337e-01 2.65652627e-01 -2.73758024e-01
-6.95979521e-02 -8.12291980e-01 -1.04661250e+00 -5.59220552e-01
8.52910057e-02 2.06982717e-01 9.39373851e-01 7.89287508e-01
-1.51706571e-02 8.14327657e-01 6.13614440e-01 -1.08490610e+00
6.72320902e-01 -2.28808999e-01 -7.17184007e-01 -2.95117676e-01
-6.92205906e-01 -5.94755292e-01 3.27037096e-01 -1.14720561e-01] | [3.5151755809783936, 1.5186182260513306] |
d0661e97-e93b-495b-afec-34444ecccedd | a-sentence-judgment-system-for-grammatical | null | null | https://aclanthology.org/C14-2015 | https://aclanthology.org/C14-2015.pdf | A Sentence Judgment System for Grammatical Error Detection | null | ['Hsin-Hsi Chen', 'Li-Ping Chang', 'Yuen-Hsien Tseng', 'Kuei-Ching Lee', 'Liang-Chih Yu', 'Lung-Hao Lee'] | 2014-08-01 | a-sentence-judgment-system-for-grammatical-1 | https://aclanthology.org/C14-2015 | https://aclanthology.org/C14-2015.pdf | coling-2014-8 | ['grammatical-error-detection'] | ['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] | [-7.263407230377197, 3.6874289512634277] |
fa95c620-7bee-4ea4-a8e3-7117c92b535a | a-sampling-theory-perspective-of-graph-based | 1705.09518 | null | http://arxiv.org/abs/1705.09518v2 | http://arxiv.org/pdf/1705.09518v2.pdf | A Sampling Theory Perspective of Graph-based Semi-supervised Learning | Graph-based methods have been quite successful in solving unsupervised and
semi-supervised learning problems, as they provide a means to capture the
underlying geometry of the dataset. It is often desirable for the constructed
graph to satisfy two properties: first, data points that are similar in the
feature space should be strongly connected on the graph, and second, the class
label information should vary smoothly with respect to the graph, where
smoothness is measured using the spectral properties of the graph Laplacian
matrix. Recent works have justified some of these smoothness conditions by
showing that they are strongly linked to the semi-supervised smoothness
assumption and its variants. In this work, we reinforce this connection by
viewing the problem from a graph sampling theoretic perspective, where class
indicator functions are treated as bandlimited graph signals (in the
eigenvector basis of the graph Laplacian) and label prediction as a bandlimited
reconstruction problem. Our approach involves analyzing the bandwidth of class
indicator signals generated from statistical data models with separable and
nonseparable classes. These models are quite general and mimic the nature of
most real-world datasets. Our results show that in the asymptotic limit, the
bandwidth of any class indicator is also closely related to the geometry of the
dataset. This allows one to theoretically justify the assumption of
bandlimitedness of class indicator signals, thereby providing a sampling
theoretic interpretation of graph-based semi-supervised classification. | ['Aly El Gamal', 'Salman Avestimehr', 'Aamir Anis', 'Antonio Ortega'] | 2017-05-26 | null | null | null | null | ['graph-sampling'] | ['graphs'] | [ 3.76902014e-01 4.70727682e-01 -3.34932327e-01 -2.94862539e-01
-1.93632767e-01 -4.96159852e-01 4.26681846e-01 2.81952232e-01
2.73964554e-01 5.90382040e-01 2.58573126e-02 -8.39053243e-02
-6.85750127e-01 -8.90421808e-01 -5.25912702e-01 -8.69878352e-01
-3.94030064e-01 3.04467916e-01 3.38161923e-02 -8.14535543e-02
1.74842089e-01 6.59172118e-01 -1.48826599e+00 -2.51240700e-01
8.51564586e-01 9.47432518e-01 -7.01970682e-02 5.43759704e-01
-6.57621473e-02 5.65290809e-01 -2.22121686e-01 -4.94368002e-02
2.60579616e-01 -7.10315228e-01 -9.63191748e-01 6.32119358e-01
2.29269788e-01 5.70733130e-01 -2.44850785e-01 1.42350674e+00
5.65942377e-02 2.06896320e-01 9.91982996e-01 -1.42222655e+00
-5.48114240e-01 3.33573997e-01 -4.62841809e-01 -1.41398413e-02
4.03818637e-01 -5.82287192e-01 1.28146195e+00 -4.59437400e-01
6.53147936e-01 8.71934056e-01 7.35497594e-01 1.03412606e-01
-1.71118474e+00 4.46964353e-02 -2.42960289e-01 1.59690604e-01
-1.41978252e+00 -2.55996883e-01 1.31755376e+00 -7.01832950e-01
1.18148789e-01 4.09782648e-01 6.90152824e-01 6.42216861e-01
5.91958798e-02 3.87571841e-01 9.89547670e-01 -8.63964260e-01
4.55604285e-01 2.67048478e-01 5.65317988e-01 9.00550187e-01
4.02361542e-01 -1.04071207e-01 -1.47661552e-01 -3.67563754e-01
8.76720428e-01 -7.17051476e-02 -6.67195559e-01 -1.13341570e+00
-1.09293234e+00 1.11156690e+00 3.28501731e-01 6.58878922e-01
-3.95203680e-01 -1.17915787e-01 1.54589742e-01 4.70869690e-01
6.69270635e-01 3.01557362e-01 8.24457854e-02 4.22994405e-01
-6.07500970e-01 -2.87653357e-01 1.05688632e+00 1.11262989e+00
9.30739760e-01 1.27845295e-02 4.33140159e-01 6.65207624e-01
4.68231380e-01 4.37146723e-01 3.69319290e-01 -6.49583638e-01
8.45267624e-02 5.26225090e-01 -1.87657937e-01 -1.22861350e+00
-5.51883638e-01 -6.05601847e-01 -9.70277667e-01 -6.20407723e-02
8.30760300e-01 2.42071092e-01 -4.26477760e-01 1.77597487e+00
2.01558262e-01 2.50924468e-01 -2.16550175e-02 7.48723567e-01
3.90236259e-01 3.76088709e-01 -3.17658544e-01 -6.00523651e-01
9.53532636e-01 -2.00219080e-01 -6.34267271e-01 3.23020846e-01
7.13381946e-01 -4.64230597e-01 1.22733545e+00 3.55370849e-01
-8.02731335e-01 -4.90859717e-01 -1.00103045e+00 4.59703773e-01
-4.38968837e-02 -5.59427254e-02 5.42041957e-01 8.28495681e-01
-1.07797348e+00 7.65140712e-01 -5.84373832e-01 -5.66362560e-01
9.33598131e-02 1.69356421e-01 -4.44324285e-01 1.87604763e-02
-9.23586071e-01 3.92955065e-01 2.27976561e-01 -2.34189164e-03
-2.57015228e-01 -4.09561038e-01 -8.44054937e-01 1.33402511e-01
-1.14223175e-02 -2.48064041e-01 6.37512445e-01 -1.31192541e+00
-1.11545289e+00 9.72639382e-01 -2.52063554e-02 -2.83694267e-01
3.04658502e-01 5.17555475e-01 -6.38631761e-01 5.31067431e-01
-8.55191872e-02 -3.11081201e-01 1.14491212e+00 -1.19878089e+00
-6.28656670e-02 -4.72203493e-01 -7.60987103e-02 -1.47281453e-01
-2.18114451e-01 -5.80082417e-01 1.36540830e-01 -4.92813617e-01
6.07551277e-01 -7.81669140e-01 -1.06071629e-01 -1.95696637e-01
-5.29385507e-01 -1.35779649e-01 7.97100306e-01 -3.29111516e-01
9.94942963e-01 -2.48412657e+00 2.21486241e-01 8.10710251e-01
3.41950238e-01 -1.96151063e-01 -2.04279616e-01 6.53864503e-01
-5.27078629e-01 8.06999281e-02 -3.99416059e-01 1.30479380e-01
-1.49981111e-01 2.91701779e-02 -2.88007498e-01 1.35292077e+00
2.02896316e-02 5.87140143e-01 -9.72544074e-01 -3.42689544e-01
1.09216183e-01 3.92163426e-01 -3.93751860e-01 -2.88501512e-02
-5.95823601e-02 6.74397826e-01 -5.45872629e-01 2.86155790e-01
6.09865487e-01 -5.87644935e-01 3.36496592e-01 -1.27032593e-01
2.03780025e-01 -3.50137092e-02 -1.37439716e+00 1.41400278e+00
-2.35433951e-01 7.50495911e-01 2.19398901e-01 -1.82666349e+00
1.08243287e+00 3.61476392e-01 8.39188159e-01 -1.99111059e-01
-1.67258945e-03 1.89727649e-01 -8.68173838e-02 -4.77944463e-01
-1.61439136e-01 -3.68539780e-01 3.10162932e-01 4.01360065e-01
1.24925070e-01 -2.03853562e-01 5.94964251e-03 1.73970133e-01
9.37400579e-01 -3.53675127e-01 5.73010862e-01 -9.91327584e-01
6.99590743e-01 -1.65120065e-01 1.85959905e-01 5.00395358e-01
8.05580989e-02 4.70761508e-01 7.15384603e-01 -1.64395958e-01
-7.86273956e-01 -1.19923615e+00 -5.85872650e-01 3.39775383e-01
2.21738115e-01 -1.60939962e-01 -8.87015939e-01 -6.07110202e-01
-4.47827242e-02 4.67110306e-01 -6.16935730e-01 -2.55871952e-01
-1.84738889e-01 -4.65458274e-01 1.27363473e-01 -4.08955961e-02
2.53676862e-01 -5.61429739e-01 -2.10345343e-01 -7.11622089e-03
-1.19867146e-01 -8.65030408e-01 -3.63372028e-01 1.13738507e-01
-1.13118732e+00 -1.59496391e+00 -6.37381673e-01 -9.25516546e-01
9.41120565e-01 5.17505109e-01 1.07488859e+00 1.36177182e-01
-3.39995295e-01 1.02144742e+00 -3.93905491e-01 -7.44595453e-02
-4.82378632e-01 -2.01206177e-01 -2.18945672e-04 6.94580913e-01
9.06543955e-02 -8.90617251e-01 -2.89914638e-01 5.16755700e-01
-1.06552017e+00 -2.03304768e-01 7.77154490e-02 7.65412748e-01
7.21814394e-01 3.11761320e-01 6.71070158e-01 -9.79767203e-01
5.89885652e-01 -6.31415248e-01 -6.89093113e-01 3.52926314e-01
-5.08300006e-01 1.88844666e-01 8.80979478e-01 -5.28890729e-01
-4.48858857e-01 2.13185579e-01 4.05830204e-01 -2.09707975e-01
-2.13222429e-01 6.97288573e-01 -3.07904571e-01 -4.50143814e-01
8.03339422e-01 3.41337532e-01 2.08179191e-01 -3.85343730e-01
2.39828467e-01 5.35731018e-01 1.84249029e-01 -5.28074503e-01
7.83436596e-01 6.79485738e-01 6.71062708e-01 -1.68783355e+00
-6.56295359e-01 -8.58523846e-01 -8.12773407e-01 -3.40382338e-01
6.80569947e-01 -3.60156000e-01 -6.89597964e-01 1.17336974e-01
-7.34535217e-01 -1.95862010e-01 -7.87422419e-01 7.49229908e-01
-1.03892827e+00 7.08892703e-01 -2.12210000e-01 -1.07672977e+00
3.11597526e-01 -7.38443732e-01 8.15231800e-01 -2.18575120e-01
-6.49349717e-03 -1.70382977e+00 1.00696340e-01 -1.27611235e-01
2.50687152e-01 4.83242124e-01 1.14111638e+00 -7.34622836e-01
-1.27440900e-01 -4.27513957e-01 -2.62393430e-02 2.73158550e-01
4.09010768e-01 -2.16654882e-01 -7.92274594e-01 -3.74555111e-01
5.34402609e-01 3.36023085e-02 5.45039654e-01 7.59749055e-01
1.09961116e+00 -2.40386322e-01 -1.55483961e-01 5.75126052e-01
1.62489331e+00 -2.68710941e-01 2.98148453e-01 -3.32632393e-01
5.75487077e-01 8.71565700e-01 1.55479088e-01 3.06025267e-01
-2.71122783e-01 7.19910741e-01 3.92126769e-01 -8.15247893e-02
-3.90624478e-02 -2.52292782e-01 9.03246105e-02 8.69208634e-01
-2.06537411e-01 -7.32781366e-02 -7.41360486e-01 3.43416601e-01
-1.71761513e+00 -9.31580067e-01 -7.41112769e-01 2.73710322e+00
3.32466424e-01 -1.33271873e-01 5.33796072e-01 6.24963462e-01
1.02921867e+00 6.12065792e-02 -2.27046579e-01 -4.21212390e-02
-2.68983394e-01 5.94593287e-02 5.64358592e-01 7.12542772e-01
-9.28991795e-01 2.61498243e-01 6.43249464e+00 5.77292383e-01
-1.09810305e+00 -1.54530257e-01 3.36779058e-01 5.32816052e-01
-4.01833117e-01 1.00202501e-01 -2.04810798e-01 3.01572978e-01
8.27218533e-01 -4.25362915e-01 4.84150559e-01 8.40261459e-01
1.41532302e-01 -3.50675210e-02 -1.27461517e+00 8.76378119e-01
4.45470400e-02 -1.11166954e+00 -2.20890313e-01 4.82717484e-01
4.76007223e-01 -2.24184245e-01 -3.34869698e-02 -3.71196389e-01
4.67227511e-02 -9.36575115e-01 5.00602365e-01 5.62108040e-01
8.69186580e-01 -6.08886719e-01 3.78219754e-01 4.98817593e-01
-1.40167260e+00 1.95091948e-01 -4.18038070e-01 -5.06172664e-02
1.26920745e-01 1.01053309e+00 -6.51360571e-01 6.86077893e-01
6.15982525e-02 1.03743303e+00 -3.31705421e-01 1.06155825e+00
2.30584294e-02 8.66920412e-01 -3.27344030e-01 1.53889004e-02
-3.41141261e-02 -8.66822958e-01 7.74775088e-01 9.88068163e-01
2.28464216e-01 9.19972733e-02 2.12996587e-01 7.53745556e-01
1.72673672e-01 4.06219751e-01 -9.06505406e-01 -1.76158920e-01
5.74206039e-02 9.51421618e-01 -1.08470869e+00 -4.64981124e-02
-6.41220689e-01 6.34628177e-01 1.65575549e-01 5.92006326e-01
-4.38938826e-01 -4.56608623e-01 2.80296415e-01 4.65378702e-01
1.33989006e-01 -3.00242126e-01 -1.73484787e-01 -1.33378994e+00
5.12743108e-02 -5.20255327e-01 3.80261332e-01 -3.22019249e-01
-1.44852579e+00 3.33584934e-01 2.88018230e-02 -1.44828987e+00
-4.32437025e-02 -6.36073411e-01 -5.30983746e-01 7.28379071e-01
-1.27859914e+00 -9.19223070e-01 -1.34852499e-01 1.02684057e+00
1.22463889e-01 1.03420332e-01 9.32196796e-01 1.91530883e-02
-1.20364718e-01 2.08155796e-01 3.04740638e-01 1.27766028e-01
1.28140211e-01 -1.40048409e+00 -1.87365741e-01 7.32094169e-01
6.26279056e-01 6.40335798e-01 8.09466720e-01 -4.79673505e-01
-1.55495024e+00 -8.61595631e-01 6.62900686e-01 -1.55130504e-02
9.58965480e-01 -3.22663844e-01 -9.64171231e-01 5.07507801e-01
-4.01208341e-01 4.13032264e-01 8.45400274e-01 1.26298517e-01
-2.09605008e-01 3.79375927e-02 -1.11106622e+00 1.40264079e-01
1.07574415e+00 -5.72598517e-01 -3.29430848e-01 8.73777688e-01
2.16643274e-01 2.34499529e-01 -9.10074055e-01 1.39173582e-01
1.99663848e-01 -1.01314914e+00 8.31194639e-01 -7.38842010e-01
4.49456014e-02 -1.01265326e-01 -3.07767034e-01 -1.31689227e+00
-4.14510548e-01 -7.26976991e-01 8.56894162e-03 8.91275346e-01
2.15482876e-01 -9.39580977e-01 7.78290570e-01 1.42231137e-01
1.38300076e-01 -4.84443933e-01 -9.11778986e-01 -1.13713741e+00
-1.60303682e-01 -2.14885920e-01 1.96621835e-01 1.22673261e+00
5.73468029e-01 3.63653243e-01 -1.15379162e-01 2.13721469e-01
9.36500907e-01 2.51907915e-01 4.29120243e-01 -2.01095796e+00
-5.02862573e-01 -3.76114875e-01 -9.54845488e-01 -1.02674472e+00
3.79042447e-01 -1.25121152e+00 -1.54148355e-01 -1.26930439e+00
-1.97724983e-01 -6.96382165e-01 -7.13724941e-02 -2.12192550e-01
4.09170955e-01 1.64154414e-02 -2.09630445e-01 5.68819284e-01
-1.40220672e-01 2.61101991e-01 1.07422984e+00 1.58859193e-01
-3.35696638e-01 5.51002324e-01 -4.00444567e-01 9.13164675e-01
5.85597277e-01 -1.76803932e-01 -6.95996642e-01 9.25650969e-02
2.96064824e-01 3.51049751e-01 3.81679952e-01 -8.76894355e-01
6.60591722e-02 -1.81493774e-01 -1.95088118e-01 -2.99934726e-02
1.20210305e-01 -1.21159542e+00 3.09223115e-01 4.56636906e-01
-5.62931836e-01 -4.78582621e-01 -3.99134904e-01 1.00349629e+00
-2.27802977e-01 -4.65339810e-01 9.18837845e-01 1.85247749e-01
-3.79755616e-01 3.44758958e-01 -2.13184729e-01 2.09048897e-01
9.99394238e-01 -4.31253403e-01 6.12486601e-02 -6.62981689e-01
-8.89676869e-01 -3.41018736e-01 6.18944943e-01 -2.36260056e-01
4.59181458e-01 -1.26838887e+00 -5.13891697e-01 4.79166687e-01
1.86951175e-01 -3.00452411e-01 -8.82978141e-02 1.09758890e+00
-3.88447851e-01 3.63257349e-01 4.93164361e-02 -8.27965796e-01
-1.02775228e+00 7.82757759e-01 2.83933580e-01 4.19938639e-02
-8.22671592e-01 4.73359615e-01 1.68797061e-01 -1.08185589e-01
1.43981606e-01 -2.40542784e-01 -2.21493915e-01 9.08273645e-03
2.68780261e-01 4.65241760e-01 2.52677798e-02 -8.71942103e-01
-1.18787423e-01 7.81929314e-01 6.28201246e-01 1.89804062e-02
1.31373739e+00 -1.33970708e-01 -2.18247145e-01 6.76409781e-01
1.53577280e+00 3.78071904e-01 -9.17927623e-01 -4.67834562e-01
2.28727251e-01 -3.75758052e-01 -5.21781817e-02 2.56167930e-02
-9.95222926e-01 8.22375417e-01 1.75565541e-01 1.25267899e+00
1.23598945e+00 3.93462241e-01 7.88059384e-02 8.85899812e-02
4.32896018e-01 -7.51934409e-01 -1.75268725e-02 1.16120815e-01
8.02345634e-01 -9.94894087e-01 -5.06629162e-02 -1.14895558e+00
-1.99393049e-01 1.41564763e+00 -3.95221323e-01 -6.06620550e-01
1.06309783e+00 -2.07599372e-01 -3.59361112e-01 -4.34828430e-01
-6.40288889e-02 -2.54441261e-01 6.95610583e-01 9.55931485e-01
5.44580638e-01 2.03082934e-01 -4.24210548e-01 2.25475609e-01
-1.55021787e-01 -3.08478028e-01 5.00892460e-01 5.49438953e-01
-5.89630246e-01 -7.96833336e-01 -3.58310908e-01 4.45649475e-01
-1.44939035e-01 1.58429310e-01 -4.60611701e-01 7.44091272e-01
-4.66446370e-01 1.08522236e+00 -1.76054001e-01 -6.74776807e-02
2.05014914e-01 7.36429319e-02 5.04077315e-01 -7.40577519e-01
3.39280367e-01 1.07056618e-01 -1.04047008e-01 -4.29925501e-01
-6.86599851e-01 -6.39242947e-01 -1.05977380e+00 -1.45797148e-01
-5.81579506e-01 6.79995120e-01 6.08892441e-01 9.55451965e-01
4.01531197e-02 1.61446810e-01 8.92874360e-01 -5.94908655e-01
-6.51025653e-01 -7.09369719e-01 -1.41392529e+00 5.95351577e-01
4.39372152e-01 -6.34535611e-01 -8.72746646e-01 2.58302093e-01] | [7.004343509674072, 5.245625019073486] |
ddec5d75-1d20-4386-8a01-8781c318ad1b | if-at-first-you-don-t-succeed-test-time-re | 2303.17703 | null | https://arxiv.org/abs/2303.17703v1 | https://arxiv.org/pdf/2303.17703v1.pdf | If At First You Don't Succeed: Test Time Re-ranking for Zero-shot, Cross-domain Retrieval | In this paper we propose a novel method for zero-shot, cross-domain image retrieval in which we make two key contributions. The first is a test-time re-ranking procedure that enables query-gallery pairs, without meaningful shared visual features, to be matched by incorporating gallery-gallery ranks into an iterative re-ranking process. The second is the use of cross-attention at training time and knowledge distillation to encourage cross-attention-like features to be extracted at test time from a single image. When combined with the Vision Transformer architecture and zero-shot retrieval losses, our approach yields state-of-the-art results on the Sketchy and TU-Berlin sketch-based image retrieval benchmarks. However, unlike many previous methods, none of the components in our approach are engineered specifically towards the sketch-based image retrieval task - it can be generally applied to any cross-domain, zero-shot retrieval task. We therefore also show results on zero-shot cartoon-to-photo retrieval using the Office-Home dataset. | ['William A. P. Smith', 'Finlay G. C. Hudson'] | 2023-03-30 | null | null | null | null | ['sketch-based-image-retrieval'] | ['computer-vision'] | [ 3.20823878e-01 -3.39051783e-01 -2.24535748e-01 -2.74832636e-01
-1.52237856e+00 -5.46369553e-01 1.12362266e+00 2.73501985e-02
-2.59834945e-01 3.45150232e-01 1.90064996e-01 5.88732399e-02
-5.38334906e-01 -5.32513857e-01 -6.78930223e-01 -2.94336438e-01
2.60403275e-01 6.94705307e-01 2.18187779e-01 -5.02619743e-01
4.82875794e-01 4.34086174e-01 -1.83226001e+00 4.73037392e-01
3.91491503e-01 1.19422412e+00 2.06792712e-01 7.07001746e-01
8.95562842e-02 7.77936637e-01 -2.53362536e-01 -5.46502233e-01
4.48319048e-01 -2.32286945e-01 -1.01385927e+00 5.83404265e-02
1.08926201e+00 -8.22252810e-01 -4.61959362e-01 7.12872088e-01
7.48637557e-01 3.87971044e-01 8.60739470e-01 -1.26494253e+00
-1.15610564e+00 -2.59160437e-03 -4.64868993e-01 -2.98081208e-02
6.54097199e-01 3.36131901e-02 1.43071270e+00 -1.40123630e+00
1.01885450e+00 1.44788277e+00 3.14650208e-01 4.89453852e-01
-1.29786217e+00 -5.56245208e-01 -5.80015369e-02 3.77465367e-01
-1.36800039e+00 -6.21552348e-01 7.95845985e-01 -3.27315360e-01
8.73005867e-01 2.67423481e-01 4.46361780e-01 1.04116046e+00
-2.74845719e-01 1.09437692e+00 9.75901961e-01 -6.98884010e-01
-1.57735348e-01 1.79796278e-01 2.08946187e-02 7.95887172e-01
-1.89575449e-01 1.79603040e-01 -5.12031555e-01 -3.28166783e-01
7.03278184e-01 3.73387843e-01 -1.70249224e-01 -1.08455026e+00
-1.30177689e+00 8.35733831e-01 5.68339765e-01 5.14093995e-01
-1.74115673e-01 3.87208462e-01 5.72780609e-01 7.44723737e-01
3.86886597e-01 5.40589929e-01 -4.88939285e-02 2.08474413e-01
-1.27915156e+00 2.99748898e-01 6.36221111e-01 1.06498110e+00
9.21516180e-01 -4.05215263e-01 -5.93891621e-01 1.14306664e+00
1.59190699e-01 4.75169748e-01 3.92031193e-01 -1.00388491e+00
2.85331607e-01 3.97883207e-01 1.86104715e-01 -9.49269831e-01
3.60746771e-01 9.45990086e-02 -4.24557269e-01 2.60626614e-01
2.90653855e-02 7.21918404e-01 -1.15051436e+00 1.47167265e+00
-1.37227744e-01 -9.56420321e-03 7.46095702e-02 1.07075799e+00
9.34742153e-01 3.06280643e-01 -7.82620385e-02 4.16420847e-01
1.27995563e+00 -1.24809074e+00 -1.79400772e-01 -2.47956887e-02
2.85939515e-01 -1.10362744e+00 1.31614411e+00 8.29088092e-02
-1.19213426e+00 -6.53166711e-01 -1.20681047e+00 -6.33056045e-01
-8.35072458e-01 6.91596717e-02 3.40936750e-01 1.85625717e-01
-1.22304213e+00 7.92127788e-01 -1.23849630e-01 -8.53524089e-01
3.28369290e-01 1.02172032e-01 -5.38031816e-01 -7.21186996e-01
-1.10373127e+00 1.14452326e+00 3.29058953e-02 -4.86371756e-01
-1.16975677e+00 -7.37707973e-01 -6.74653530e-01 2.61506617e-01
3.69938016e-01 -8.38424981e-01 1.14645481e+00 -9.31570828e-01
-1.23016441e+00 1.31098294e+00 8.90485123e-02 -8.97620693e-02
4.80476171e-01 -1.67162657e-01 -2.22507372e-01 5.13944685e-01
2.98148423e-01 1.05929911e+00 1.26087725e+00 -1.49500513e+00
-3.48121434e-01 -2.41113782e-01 1.63056388e-01 1.88009098e-01
-2.33107179e-01 6.57092631e-02 -9.56437886e-01 -7.80212641e-01
-1.96151480e-01 -8.94659042e-01 1.84971020e-01 5.69021225e-01
-1.08505376e-02 -3.20409179e-01 9.15071726e-01 -4.18375701e-01
8.89923692e-01 -2.13141918e+00 2.47912705e-01 3.27105552e-01
3.45198214e-02 3.99849415e-01 -7.58499026e-01 8.67272258e-01
-1.21876799e-01 3.87756736e-03 -1.19666435e-01 -5.16040564e-01
3.29899579e-01 8.59289169e-02 -6.55589044e-01 2.76423424e-01
4.70235884e-01 1.32813740e+00 -1.10365963e+00 -6.60818219e-01
4.62438375e-01 4.10791278e-01 -3.57760072e-01 1.77670866e-01
-9.91307944e-03 -1.58012390e-01 -2.69495964e-01 8.22808087e-01
4.90022391e-01 -4.04126704e-01 -1.13501094e-01 -1.96722060e-01
2.12053716e-01 -4.65549342e-02 -7.51309991e-01 2.23303056e+00
-5.97190976e-01 7.05797851e-01 -1.99316204e-01 -8.62692595e-01
7.04699099e-01 2.86126792e-01 3.81421447e-01 -1.05235171e+00
-2.60984659e-01 2.91076809e-01 -6.99447334e-01 -1.64226383e-01
7.22630024e-01 3.62173878e-02 -1.82882831e-01 6.27798975e-01
4.66138095e-01 -3.45247716e-01 3.34470361e-01 6.19291067e-01
9.75870073e-01 2.45644242e-01 2.20804632e-01 -1.18015155e-01
5.83158135e-01 -5.98095693e-02 -1.84922546e-01 1.03320730e+00
6.82663694e-02 1.01738644e+00 1.84728459e-01 -4.12022978e-01
-1.43386579e+00 -1.16588211e+00 1.02065332e-01 1.41932237e+00
1.94515780e-01 -4.91351485e-01 -1.04985617e-01 -7.23017633e-01
2.65338510e-01 2.59785950e-01 -6.33182824e-01 -3.15532088e-01
-2.26134405e-01 1.88350350e-01 3.81371021e-01 3.80623966e-01
2.25341618e-01 -9.82924104e-01 -5.23268282e-01 2.14079693e-02
-2.98660845e-02 -9.02227759e-01 -8.51144075e-01 -1.90776616e-01
-5.91521561e-01 -1.18726850e+00 -1.50941598e+00 -9.63333488e-01
5.66472948e-01 8.12120974e-01 1.45488358e+00 4.20378506e-01
-7.76525557e-01 1.26940620e+00 -3.66531670e-01 -6.52615400e-03
-1.11935072e-01 -1.68499351e-01 -2.71943092e-01 -8.28671604e-02
3.56339216e-01 -3.68830383e-01 -9.00467575e-01 3.87321144e-01
-9.54699993e-01 -3.33054870e-01 7.61641502e-01 1.31422186e+00
5.54318309e-01 -7.35945225e-01 4.06753391e-01 -4.77143914e-01
6.81432188e-01 -1.66576236e-01 -5.15986025e-01 8.97595108e-01
-8.06145549e-01 2.08835602e-01 2.58498996e-01 -4.76136237e-01
-6.27105713e-01 6.11572973e-02 3.99185836e-01 -1.13430667e+00
1.50860295e-01 7.46835070e-03 1.93856031e-01 -4.75502014e-01
4.02399153e-01 2.65601844e-01 5.69174625e-02 -4.38594997e-01
8.26714218e-01 3.94247264e-01 4.39142108e-01 -5.23048341e-01
9.38885391e-01 4.92077351e-01 -1.41230643e-01 -6.70517385e-01
-6.12440586e-01 -9.75352764e-01 -5.36050200e-01 -4.18721773e-02
7.60094285e-01 -1.09019625e+00 -5.25562704e-01 1.62653551e-01
-1.08459425e+00 -3.52836028e-02 -4.93771642e-01 5.65597303e-02
-7.97968268e-01 6.23722911e-01 -3.72426003e-01 -7.73110032e-01
-6.56559169e-01 -9.78063166e-01 1.70722520e+00 -1.58565134e-01
-7.67239463e-03 -6.36044800e-01 1.25698134e-01 2.55831718e-01
4.14325774e-01 -2.14911759e-01 9.70256507e-01 -6.51679873e-01
-9.71627355e-01 -3.66109729e-01 -8.20790350e-01 3.62867326e-01
-2.93708682e-01 -2.43769363e-01 -9.71585274e-01 -6.19271755e-01
-6.82842076e-01 -1.05808485e+00 1.21533954e+00 -1.06639154e-01
1.00606215e+00 -8.78397152e-02 -2.84870058e-01 3.89123499e-01
1.62770808e+00 -2.51106739e-01 7.85586655e-01 2.16683462e-01
3.78741145e-01 4.97130305e-01 8.68532419e-01 3.01398665e-01
2.88326770e-01 9.32200611e-01 3.30983609e-01 -3.18075776e-01
-3.99054050e-01 -4.76949364e-01 1.22094825e-02 4.51830417e-01
1.66437417e-01 -1.21987015e-01 -5.82578182e-01 1.01694357e+00
-1.90088177e+00 -1.25974512e+00 4.37224597e-01 2.39751792e+00
5.38734078e-01 -4.29977804e-01 2.47607380e-03 -1.33342266e-01
5.11470377e-01 4.58358347e-01 -3.95626754e-01 -8.86429474e-02
-3.37617546e-02 5.11374056e-01 3.28100055e-01 4.07344997e-01
-1.06168115e+00 1.04276264e+00 6.34900665e+00 1.05114293e+00
-8.12631369e-01 1.89473256e-01 3.96559894e-01 -2.93752879e-01
-3.17803323e-01 9.24623460e-02 -4.66265708e-01 8.94050077e-02
2.11001560e-01 -4.80422042e-02 6.56574368e-01 8.21876228e-01
-6.73036635e-01 -5.09333648e-02 -1.45934319e+00 1.28750074e+00
5.97540975e-01 -1.28505433e+00 3.02428693e-01 -2.84929834e-02
7.34184623e-01 -1.92447472e-02 2.53568858e-01 5.29769123e-01
3.41036975e-01 -8.31255913e-01 5.31628668e-01 7.43806481e-01
1.14134657e+00 -5.39875925e-01 2.48504475e-01 -1.95189863e-01
-1.15788639e+00 -4.12538759e-02 -5.44761360e-01 4.94682133e-01
-1.20489813e-01 1.29145786e-01 -6.49368048e-01 7.34042227e-01
6.85087740e-01 7.43940592e-01 -6.32683277e-01 1.03752983e+00
1.39924407e-01 -2.15311423e-01 -9.14858878e-02 -4.45300620e-03
4.36639309e-01 1.09026901e-01 4.30927575e-01 1.07591105e+00
3.86608899e-01 -1.72403887e-01 1.83863878e-01 8.19385946e-01
-1.58305973e-01 -4.08020802e-03 -9.61397588e-01 -3.69055867e-02
3.13434660e-01 1.15095079e+00 -2.96023250e-01 -5.54984808e-01
-3.37820113e-01 1.53511751e+00 3.78084660e-01 4.75652099e-01
-4.51556116e-01 -7.08325326e-01 5.91259599e-01 -3.76984589e-02
7.20857024e-01 1.10279351e-01 4.31569397e-01 -1.22320974e+00
1.41709521e-01 -8.20924759e-01 4.12667066e-01 -1.15570784e+00
-1.74441910e+00 3.70515436e-01 7.32902437e-02 -1.34348536e+00
-4.84366596e-01 -5.43852746e-01 -5.71428537e-01 9.28709090e-01
-1.96117020e+00 -1.61471248e+00 -2.61267364e-01 9.29498971e-01
6.52196467e-01 -3.43260676e-01 1.09952164e+00 4.87979293e-01
1.09918758e-01 6.43855870e-01 3.15270543e-01 8.67023617e-02
1.29031956e+00 -1.08302307e+00 4.85132724e-01 4.33935940e-01
4.40360457e-01 7.14384437e-01 2.86283493e-01 -3.26196432e-01
-1.37492073e+00 -7.65887380e-01 9.50174987e-01 -5.67997575e-01
8.03250194e-01 -4.41324919e-01 -6.66178882e-01 3.80238473e-01
3.19691807e-01 3.26635271e-01 4.05194581e-01 -2.37613153e-02
-8.52726460e-01 -1.41461402e-01 -9.03450727e-01 5.02888381e-01
1.01827848e+00 -1.15308535e+00 -7.52943993e-01 4.54742044e-01
4.73113507e-01 -4.04307991e-03 -8.79125416e-01 1.87021866e-01
1.10146356e+00 -7.18877614e-01 1.55237472e+00 -7.71490693e-01
6.62067950e-01 1.76646095e-02 -3.92334312e-01 -9.76981044e-01
-3.62670183e-01 -5.14350593e-01 1.00516595e-01 1.15687466e+00
1.52960017e-01 -2.85772979e-01 4.19505864e-01 3.82621437e-01
2.39517778e-01 -5.29230893e-01 -8.29984009e-01 -8.92174244e-01
-1.04416102e-01 2.14815810e-02 4.12645936e-01 7.47443438e-01
-1.60134643e-01 3.80908638e-01 -8.01696897e-01 -4.42841500e-01
7.50477731e-01 4.42053497e-01 9.82242048e-01 -1.30639505e+00
-2.77048320e-01 -5.03704131e-01 -4.49136525e-01 -1.15189064e+00
9.45535004e-02 -9.35246706e-01 5.80454990e-02 -1.60846281e+00
4.53982562e-01 -2.91719288e-01 -6.69897079e-01 3.87895018e-01
5.25079668e-02 6.24431014e-01 7.24581540e-01 4.19453263e-01
-1.11877966e+00 5.28674662e-01 1.16847563e+00 -4.38208044e-01
1.99373335e-01 -3.98628473e-01 -4.72749650e-01 7.03521967e-02
1.92343280e-01 -3.48764569e-01 -4.26861703e-01 -3.83787602e-01
-7.80899227e-02 8.67453665e-02 9.21506643e-01 -7.86992788e-01
3.43561739e-01 3.50958526e-01 3.51166725e-01 -7.66899109e-01
8.19408715e-01 -9.76528406e-01 -2.69637316e-01 2.38010123e-01
-4.93975848e-01 9.39335003e-02 -5.33829592e-02 8.62803936e-01
-2.73128420e-01 -3.12971741e-01 6.09688044e-01 -3.88099611e-01
-1.00898385e+00 3.36429298e-01 -4.70894529e-03 1.08394669e-02
8.77144635e-01 -1.76597163e-01 -4.79614764e-01 -5.65379500e-01
-5.90168655e-01 3.00863296e-01 7.36005723e-01 7.37540543e-01
8.47937346e-01 -1.66856444e+00 -6.97315812e-01 2.17109755e-01
8.63013446e-01 -6.58786237e-01 2.57549256e-01 5.48997819e-01
-1.86315075e-01 8.06766748e-01 -3.43511879e-01 -5.58798671e-01
-1.39704454e+00 9.10593092e-01 -6.39836043e-02 -5.60927629e-01
-5.96643507e-01 7.26301134e-01 1.75547495e-01 -3.04138243e-01
4.06589627e-01 1.27053395e-01 2.52500236e-01 1.33524388e-01
5.17473459e-01 2.24770918e-01 -2.49926895e-02 -4.61300611e-01
-3.21292192e-01 9.09904957e-01 -2.91209430e-01 -4.04466927e-01
1.22904301e+00 -1.07489295e-01 7.07817674e-02 3.86122078e-01
1.65330565e+00 -3.87795031e-01 -9.58110154e-01 -3.90707105e-01
8.34455416e-02 -8.10338676e-01 1.46767767e-02 -1.08082986e+00
-8.78346264e-01 9.24577653e-01 8.40519011e-01 -1.16334178e-01
1.06969047e+00 1.90574318e-01 7.24791050e-01 7.05309451e-01
3.73139918e-01 -1.16155863e+00 5.44732273e-01 3.85287195e-01
1.25242662e+00 -1.61439180e+00 3.12942937e-02 1.52182877e-01
-6.56558454e-01 9.22680795e-01 2.17123657e-01 -4.61480170e-01
3.57146859e-01 -4.31612074e-01 -2.30861768e-01 -3.01472813e-01
-8.56409490e-01 -5.41581511e-01 7.69500017e-01 6.07876599e-01
1.59577310e-01 -4.38776523e-01 -1.41661927e-01 -7.07682222e-02
4.94506866e-01 2.13658452e-01 -1.61421709e-02 9.99018312e-01
-3.03143382e-01 -1.26336551e+00 -1.76369712e-01 3.57938558e-01
-1.75294116e-01 -4.48867738e-01 -7.04645157e-01 8.98428798e-01
-4.01649415e-01 7.74766028e-01 6.19117171e-02 -1.25320911e-01
3.76625240e-01 3.05451155e-01 7.67104089e-01 -4.68511283e-01
-6.24191880e-01 -1.23203173e-01 -7.09846243e-02 -8.09851527e-01
-4.31136131e-01 -3.30290973e-01 -3.97506535e-01 -9.99981537e-02
-3.59708548e-01 -3.56053151e-02 6.93848014e-01 4.83851880e-01
6.37423217e-01 1.46144703e-01 5.78266501e-01 -1.08090353e+00
-6.79517090e-01 -7.06951737e-01 -5.11808813e-01 8.21110249e-01
4.93874311e-01 -7.87490129e-01 -2.89734513e-01 -1.76831201e-01] | [11.423848152160645, 0.7352107763290405] |
42825fb2-02d3-4152-9a74-431a487b5b8e | learning-from-synthetic-shadows-for-shadow | 2101.01713 | null | https://arxiv.org/abs/2101.01713v2 | https://arxiv.org/pdf/2101.01713v2.pdf | Learning from Synthetic Shadows for Shadow Detection and Removal | Shadow removal is an essential task in computer vision and computer graphics. Recent shadow removal approaches all train convolutional neural networks (CNN) on real paired shadow/shadow-free or shadow/shadow-free/mask image datasets. However, obtaining a large-scale, diverse, and accurate dataset has been a big challenge, and it limits the performance of the learned models on shadow images with unseen shapes/intensities. To overcome this challenge, we present SynShadow, a novel large-scale synthetic shadow/shadow-free/matte image triplets dataset and a pipeline to synthesize it. We extend a physically-grounded shadow illumination model and synthesize a shadow image given an arbitrary combination of a shadow-free image, a matte image, and shadow attenuation parameters. Owing to the diversity, quantity, and quality of SynShadow, we demonstrate that shadow removal models trained on SynShadow perform well in removing shadows with diverse shapes and intensities on some challenging benchmarks. Furthermore, we show that merely fine-tuning from a SynShadow-pre-trained model improves existing shadow detection and removal models. Codes are publicly available at https://github.com/naoto0804/SynShadow. | ['Toshihiko Yamasaki', 'Naoto Inoue'] | 2021-01-05 | null | null | null | null | ['shadow-removal', 'shadow-detection-and-removal', 'shadow-detection'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 6.71670139e-01 -1.65379077e-01 4.19374913e-01 -5.83577812e-01
-4.05434281e-01 -3.88171107e-01 3.99831474e-01 -6.39000058e-01
-6.30962402e-02 8.68659079e-01 1.30880043e-01 -4.77274030e-01
5.89422941e-01 -6.74021006e-01 -9.49007452e-01 -8.83715451e-01
4.53962743e-01 3.41811627e-01 6.65244162e-01 -4.64606673e-01
-1.11270405e-01 7.37765312e-01 -1.54090941e+00 2.43738532e-01
1.06496632e+00 8.28716397e-01 4.93677795e-01 9.13367867e-01
1.22766212e-01 3.07407528e-01 -7.93431818e-01 -9.49399546e-02
6.31556511e-01 -3.93599480e-01 6.08115457e-03 -5.79160526e-02
1.01617026e+00 -8.04061353e-01 -5.63604653e-01 4.48029339e-01
5.51746666e-01 5.71823828e-02 4.65281725e-01 -1.41501343e+00
-6.54202700e-01 -2.03495711e-01 -6.03752494e-01 -1.42982185e-01
9.18162763e-02 4.26892579e-01 4.95981455e-01 -1.05318511e+00
5.40096879e-01 1.28312850e+00 9.47844505e-01 3.35096717e-01
-1.22019041e+00 -9.28502500e-01 6.07719272e-02 -8.95913169e-02
-1.22085178e+00 -4.07796204e-01 6.84956789e-01 -7.83626065e-02
5.55673599e-01 4.46655631e-01 7.57753968e-01 1.52871180e+00
5.50660551e-01 8.99809062e-01 1.45387244e+00 -2.51772940e-01
3.65121737e-02 -3.52379791e-02 -7.51361921e-02 8.88733566e-01
4.95242268e-01 3.99098784e-01 -6.28314316e-01 -2.15637051e-02
6.39536977e-01 8.63006935e-02 -7.15465784e-01 -6.01494074e-01
-9.73153293e-01 5.02343059e-01 6.50857627e-01 -4.50112164e-01
4.02289070e-02 3.71960312e-01 -1.49605379e-01 -5.32994457e-02
4.20589745e-01 3.12832832e-01 -4.05834675e-01 3.90219808e-01
-1.00176346e+00 3.51359665e-01 8.28729808e-01 9.64905918e-01
9.83344674e-01 3.12503010e-01 -3.27674985e-01 5.87481141e-01
-9.27684158e-02 1.40666556e+00 -1.21991009e-01 -1.02798402e+00
3.36517096e-02 2.39121377e-01 4.14625108e-01 -9.69551921e-01
-4.21926558e-01 -3.44625860e-01 -6.57898962e-01 7.72247851e-01
3.76312554e-01 -1.35520786e-01 -1.38113415e+00 1.48731875e+00
3.13881069e-01 3.10762286e-01 1.62448678e-02 1.04161632e+00
1.03223062e+00 5.50200045e-01 -4.83108729e-01 1.38035864e-01
1.02425802e+00 -1.23039448e+00 -5.78487039e-01 -6.87097728e-01
1.30991757e-01 -1.01708579e+00 1.50572050e+00 2.79874533e-01
-6.09961390e-01 -3.20047975e-01 -1.23278868e+00 -2.98158765e-01
-4.73102152e-01 3.09852928e-01 6.91953361e-01 6.09634519e-01
-9.53927875e-01 3.32475901e-01 -5.00522614e-01 -1.62626237e-01
5.52070677e-01 2.11044610e-03 -1.99461300e-02 -4.59641427e-01
-9.33234215e-01 7.15831161e-01 -1.89080521e-01 2.26604834e-01
-1.15723157e+00 -1.02011156e+00 -8.24788034e-01 6.61529973e-02
6.32306337e-01 -5.22939742e-01 1.17868400e+00 -1.07458305e+00
-1.10724401e+00 5.21564007e-01 -2.38965809e-01 -3.13965917e-01
6.30701303e-01 -5.15888870e-01 -2.98602194e-01 -1.85485974e-01
8.78807530e-02 4.27771866e-01 1.16172516e+00 -2.09266019e+00
-2.39577815e-01 4.05947864e-02 -2.78801881e-02 2.72605747e-01
1.78790570e-03 -3.99701923e-01 -5.00759780e-01 -6.42651498e-01
-2.39229858e-01 -1.04404950e+00 -9.21988115e-02 5.21716356e-01
-7.79776871e-01 7.75184751e-01 1.38068426e+00 -5.69207430e-01
7.02426910e-01 -2.10666966e+00 -4.50695902e-01 7.88737163e-02
2.33895332e-01 3.94725382e-01 -3.19465011e-01 2.76367545e-01
2.14944914e-01 -2.06196502e-01 -7.94286013e-01 -6.67454660e-01
-2.37933807e-02 4.74574119e-01 -6.46925688e-01 4.28248763e-01
-2.16517001e-02 9.91167367e-01 -8.35309982e-01 -4.02014941e-01
4.14279461e-01 8.81653965e-01 -5.55184335e-02 4.78822470e-01
-4.15027648e-01 2.48813406e-01 -6.38364330e-02 9.25891995e-01
1.23021030e+00 -6.68757483e-02 6.35046735e-02 -2.78539062e-01
9.63136479e-02 -1.54547334e-01 -8.94827068e-01 1.20091939e+00
-7.35837638e-01 1.20527124e+00 3.74767095e-01 9.21378210e-02
8.12132180e-01 -2.99404383e-01 1.91698417e-01 -9.34040666e-01
1.06168866e-01 2.45009601e-01 -1.96161956e-01 -1.58086121e-01
6.98878706e-01 -2.73555946e-02 1.39820397e-01 6.17738664e-01
-6.01562619e-01 -6.31107986e-01 -2.01344565e-01 3.77383947e-01
1.28432941e+00 2.78916448e-01 -2.41564531e-02 -1.85236201e-01
-1.44507280e-02 -1.05683565e-01 5.82929611e-01 7.94508815e-01
-5.98309338e-02 1.37654710e+00 2.64030099e-01 -3.67949665e-01
-9.18484092e-01 -1.39341080e+00 -5.63835539e-02 1.14074600e+00
5.48053324e-01 -3.77984606e-02 -6.80131793e-01 -6.05778217e-01
5.15063524e-01 7.69773543e-01 -9.86297548e-01 -5.82224987e-02
-6.55095279e-01 -8.07107270e-01 5.72186828e-01 4.78618860e-01
6.10315025e-01 -1.15298140e+00 -8.98105204e-01 -2.47110650e-01
-1.57429576e-02 -1.20777929e+00 -6.39196336e-01 2.24296033e-01
-9.64270234e-02 -1.28720951e+00 -7.12031782e-01 -3.12963456e-01
6.18802965e-01 1.10260868e+00 1.49570119e+00 5.04348516e-01
-6.86655760e-01 1.33205533e-01 -2.55021602e-01 -6.98885381e-01
-2.91531324e-01 -3.67975950e-01 -1.92202643e-01 2.30714217e-01
-4.07536656e-01 -6.56280756e-01 -1.05115604e+00 4.39231277e-01
-1.16781247e+00 5.59959233e-01 6.97803438e-01 7.16663718e-01
2.82973230e-01 -4.11080182e-01 -4.12042625e-02 -1.23827982e+00
4.10230130e-01 -2.02419788e-01 -7.01125383e-01 2.83397257e-01
-6.62002802e-01 -2.09045544e-01 5.19920766e-01 -1.72957435e-01
-1.55109954e+00 2.00022653e-01 2.22118840e-01 -6.36803865e-01
-3.84411663e-02 -3.63945037e-01 -4.10955369e-01 -3.48808885e-01
9.54308867e-01 2.01185673e-01 -2.24298611e-01 -1.83570310e-01
4.44639802e-01 2.10067585e-01 6.70273304e-01 -3.98241997e-01
1.28229499e+00 1.10564446e+00 5.34995366e-03 -7.91086555e-01
-1.17517471e+00 -2.94813901e-01 -5.07581115e-01 -2.00032353e-01
6.17187083e-01 -6.94401503e-01 -2.84924150e-01 9.06420827e-01
-9.91986096e-01 -1.46884084e+00 -1.42547607e-01 -2.58279324e-01
-1.83851540e-01 3.60106617e-01 -1.14630893e-01 -6.33239746e-01
-3.26155722e-01 -1.00196874e+00 1.46653378e+00 3.37666631e-01
2.31404245e-01 -7.73001850e-01 -3.28660347e-02 3.40924561e-01
7.17958510e-01 6.23337984e-01 5.54776371e-01 2.47492552e-01
-1.06372213e+00 1.41969085e-01 -7.99857199e-01 4.33341742e-01
4.87319112e-01 1.37731031e-01 -1.24896038e+00 -2.43046075e-01
-3.69584054e-01 -4.56874222e-01 1.37019503e+00 2.89517194e-01
1.20766068e+00 -3.51915061e-02 -4.24191922e-01 1.07650006e+00
1.42460108e+00 -1.75230518e-01 6.15227938e-01 3.48557569e-02
1.17680001e+00 1.42617628e-01 7.66276181e-01 2.98695028e-01
3.50755423e-01 5.88257551e-01 5.36780298e-01 -8.51004541e-01
-9.86175179e-01 7.86999688e-02 1.71345383e-01 4.17673541e-03
1.04053132e-01 -7.48301864e-01 -8.93038511e-01 3.73753220e-01
-1.65669203e+00 -5.43401122e-01 -3.18521470e-01 2.01061320e+00
7.74311066e-01 1.53551325e-01 -4.77435142e-01 -2.45589808e-01
3.14361930e-01 6.12375438e-01 -8.00363243e-01 -1.36522233e-01
-7.58559704e-01 3.97453904e-01 9.22940850e-01 1.01652837e+00
-8.63670468e-01 1.33455777e+00 5.38016939e+00 6.84775651e-01
-1.35213709e+00 5.02547855e-03 4.16203916e-01 -4.73019779e-02
-7.28283942e-01 -1.62274987e-02 -5.16829908e-01 3.16456199e-01
5.03417671e-01 3.05677772e-01 5.04958034e-01 6.43571019e-01
2.20790580e-01 -7.25062609e-01 -6.92272961e-01 7.07826734e-01
4.87424344e-01 -1.33114195e+00 -1.56150132e-01 2.31915247e-02
1.06749749e+00 1.53060600e-01 1.43399969e-01 1.72696441e-01
5.53976953e-01 -1.21196389e+00 6.87304676e-01 5.42304277e-01
1.06253266e+00 -1.78714916e-01 5.42384267e-01 -1.13924451e-01
-1.17494655e+00 1.71415552e-01 -9.43766236e-02 1.68984875e-01
2.03798890e-01 7.82016635e-01 -1.24246848e+00 4.05751437e-01
7.58776188e-01 3.55869830e-01 -9.10879493e-01 8.09235215e-01
-4.63207066e-01 5.45244634e-01 -2.73527056e-01 3.17355007e-01
-9.95461047e-02 -1.64502919e-01 3.11671704e-01 1.58340442e+00
-8.45176261e-03 5.92194460e-02 1.69974908e-01 8.14654768e-01
-1.88949853e-01 -4.45404857e-01 -5.87633252e-01 3.59031945e-01
4.42744553e-01 1.42633557e+00 -9.64550436e-01 -3.06314886e-01
-1.16402321e-01 1.29740500e+00 -4.66947407e-02 7.40225136e-01
-1.18263423e+00 -2.14272171e-01 8.56157720e-01 1.67670742e-01
1.25638366e-01 -2.98952460e-01 -5.86912453e-01 -9.16878045e-01
-2.10019629e-02 -8.32339346e-01 -3.14420819e-01 -1.31773412e+00
-8.85707617e-01 6.28608882e-01 -9.71681550e-02 -9.13585722e-01
3.62032026e-01 -5.89892745e-01 -7.93771207e-01 8.52114260e-01
-1.85988081e+00 -1.49134338e+00 -1.40690780e+00 5.57274222e-01
6.05290413e-01 1.46870658e-01 7.12046623e-01 4.52955998e-02
-4.09944594e-01 4.27218676e-01 2.14495435e-01 -4.68017126e-04
1.18285453e+00 -1.42641163e+00 8.87234330e-01 9.26359475e-01
-2.40011364e-01 1.22260429e-01 9.24882591e-01 -8.38956416e-01
-1.47256136e+00 -1.29333580e+00 7.85523355e-02 -6.42280638e-01
1.96741506e-01 -8.18637967e-01 -1.14857042e+00 4.66749609e-01
1.81659743e-01 2.97762722e-01 6.27704635e-02 -2.59182155e-01
-6.10808372e-01 -4.04296875e-01 -9.78686869e-01 9.57781076e-01
1.33928061e+00 -3.57567579e-01 1.54290507e-02 5.20993650e-01
8.42358053e-01 -8.92713368e-01 -9.12752077e-02 6.65324092e-01
8.41520131e-01 -1.64251328e+00 1.10700428e+00 5.22472449e-02
4.00446653e-01 -3.69617879e-01 -2.93326855e-01 -1.31366110e+00
2.10586831e-01 -6.03211462e-01 -1.22739457e-01 8.41307580e-01
3.20110917e-01 -7.45890856e-01 9.61451650e-01 3.62363696e-01
-5.27016878e-01 -9.44057345e-01 -3.74052465e-01 -6.63963675e-01
-2.66319960e-01 -2.97009975e-01 6.84862912e-01 5.11522174e-01
-1.24341607e+00 -4.63864207e-03 -5.88224351e-01 3.43499601e-01
7.93533623e-01 8.24028492e-01 1.42342854e+00 -9.60752904e-01
-2.27965474e-01 -1.02778159e-01 4.36206281e-01 -8.43510151e-01
1.43383786e-01 -3.52539510e-01 6.85185969e-01 -1.76812184e+00
1.89332813e-01 -9.35097754e-01 5.95198572e-02 7.10768104e-01
-3.80230576e-01 6.64075613e-01 1.66143805e-01 1.61769450e-01
-3.11860800e-01 8.24941754e-01 1.53075254e+00 -4.48102988e-02
-2.35658050e-01 1.39143348e-01 -4.50498581e-01 5.90715051e-01
9.18342888e-01 -2.94739634e-01 -4.93763596e-01 -5.03368258e-01
-1.65105462e-01 -2.31666505e-01 7.75287330e-01 -1.03427815e+00
-2.46874720e-01 -5.68870306e-01 6.91691160e-01 -6.32101595e-01
1.25480127e+00 -5.36572099e-01 6.04082942e-02 3.69648069e-01
1.54733360e-01 -2.85194159e-01 3.48664433e-01 6.12734675e-01
4.16854560e-01 4.77728039e-01 7.97529340e-01 1.39705893e-02
-6.80193901e-01 3.44387591e-01 -1.07824579e-02 6.41108537e-03
6.89745843e-01 -3.71889740e-01 -7.47012734e-01 -5.53449273e-01
4.28370126e-02 2.84272805e-02 1.01490724e+00 4.43946749e-01
9.40444708e-01 -7.84206510e-01 -5.65098941e-01 4.12089080e-01
3.02452333e-02 2.86805987e-01 2.64805406e-01 6.10858738e-01
-8.62628162e-01 2.45203753e-03 -3.24821562e-01 -4.53436702e-01
-1.45862317e+00 1.92003697e-01 4.89453822e-01 1.94504157e-01
-8.00239325e-01 9.31803346e-01 1.05755901e+00 -4.96049523e-01
2.61098027e-01 -3.96937013e-01 6.61106646e-01 -4.66206044e-01
1.66299030e-01 1.94835067e-01 2.28365257e-01 -4.98487353e-01
-2.13873491e-01 3.29451710e-01 3.32243085e-01 3.46089192e-02
1.12491679e+00 -8.31596330e-02 1.76045001e-01 2.21086398e-01
7.93634892e-01 5.47735929e-01 -2.00692892e+00 -1.87767968e-01
-5.42597890e-01 -8.93998265e-01 -7.14068115e-02 -1.04997623e+00
-1.30409646e+00 7.25131154e-01 6.15740299e-01 -2.18010485e-01
1.10451257e+00 -1.80960029e-01 1.21154034e+00 4.25598949e-01
1.38853565e-01 -8.57860446e-01 4.09210324e-01 5.87392807e-01
1.27313495e+00 -1.42891157e+00 2.85246372e-01 -7.49839187e-01
-6.35815084e-01 7.87530899e-01 8.54709744e-01 -1.45233378e-01
5.45214474e-01 7.95742035e-01 5.68110466e-01 -1.62626788e-01
-3.66460919e-01 -2.54075229e-01 1.81955919e-01 7.80950129e-01
-1.31465867e-01 3.19908768e-01 3.73421788e-01 8.51997584e-02
-3.96687835e-01 -3.66534144e-01 8.30687344e-01 9.99483883e-01
-5.14050424e-01 -9.08326447e-01 -7.41509795e-01 4.49392587e-01
2.47750506e-01 -3.21862936e-01 -7.50433028e-01 9.63331699e-01
2.20662266e-01 7.75435448e-01 -3.44757169e-01 -2.89898515e-01
1.57718793e-01 -3.59219581e-01 3.95173043e-01 -6.10271156e-01
-1.92901418e-01 -1.50246277e-01 1.78798795e-01 -8.80434990e-01
8.49952623e-02 -3.65811646e-01 -1.28991973e+00 -3.32169801e-01
-1.15950681e-01 -3.95358622e-01 7.26226449e-01 8.01615655e-01
3.48093063e-01 6.89542174e-01 5.15292585e-01 -1.41340458e+00
7.39160776e-02 -7.45772243e-01 -4.70435888e-01 2.91103005e-01
7.58993149e-01 -9.70446348e-01 -3.98639709e-01 -8.44873190e-02] | [10.845457077026367, -4.1003899574279785] |
0db58ee6-b785-4f13-8b2b-4d447665829f | explicit-attention-enhanced-fusion-for-rgb | 2303.15710 | null | https://arxiv.org/abs/2303.15710v1 | https://arxiv.org/pdf/2303.15710v1.pdf | Explicit Attention-Enhanced Fusion for RGB-Thermal Perception Tasks | Recently, RGB-Thermal based perception has shown significant advances. Thermal information provides useful clues when visual cameras suffer from poor lighting conditions, such as low light and fog. However, how to effectively fuse RGB images and thermal data remains an open challenge. Previous works involve naive fusion strategies such as merging them at the input, concatenating multi-modality features inside models, or applying attention to each data modality. These fusion strategies are straightforward yet insufficient. In this paper, we propose a novel fusion method named Explicit Attention-Enhanced Fusion (EAEF) that fully takes advantage of each type of data. Specifically, we consider the following cases: i) both RGB data and thermal data, ii) only one of the types of data, and iii) none of them generate discriminative features. EAEF uses one branch to enhance feature extraction for i) and iii) and the other branch to remedy insufficient representations for ii). The outputs of two branches are fused to form complementary features. As a result, the proposed fusion method outperforms state-of-the-art by 1.6\% in mIoU on semantic segmentation, 3.1\% in MAE on salient object detection, 2.3\% in mAP on object detection, and 8.1\% in MAE on crowd counting. The code is available at https://github.com/FreeformRobotics/EAEFNet. | ['Tin Lun Lam', 'Fuqin Deng', 'Hua Feng', 'Chenyu Bao', 'Junjie Hu', 'Mingjian Liang'] | 2023-03-28 | null | null | null | null | ['thermal-image-segmentation'] | ['computer-vision'] | [ 1.53097764e-01 -2.46233210e-01 1.59116283e-01 -1.44996852e-01
-6.59579396e-01 -4.67043996e-01 4.89888966e-01 1.18412133e-02
-7.28272080e-01 6.89692199e-01 -7.34795555e-02 -7.08183721e-02
1.40564770e-01 -6.96892023e-01 -7.81061828e-01 -9.43239212e-01
4.56662118e-01 -8.36540665e-03 5.85638165e-01 -3.03256392e-01
1.86650574e-01 1.61707014e-01 -1.93757212e+00 1.56708494e-01
1.17213762e+00 1.39851522e+00 5.98923028e-01 6.64170384e-01
-2.46586934e-01 5.14162064e-01 -4.74924058e-01 -2.20588699e-01
3.53637218e-01 -3.19644630e-01 -5.12989819e-01 -5.48870526e-02
1.43801436e-01 -3.82008225e-01 -1.62014440e-01 1.14338970e+00
5.66959083e-01 3.33350182e-01 4.22369629e-01 -1.48263454e+00
-4.90348756e-01 1.80925086e-01 -8.88031960e-01 3.56645525e-01
3.76275271e-01 2.89274782e-01 5.61932623e-01 -7.58121967e-01
1.81218266e-01 1.00355649e+00 3.39311391e-01 4.55284804e-01
-6.43848121e-01 -5.56999683e-01 3.19940805e-01 3.61123145e-01
-1.27213359e+00 -4.58139926e-01 6.93073630e-01 -4.07517225e-01
9.42537189e-01 2.67742664e-01 7.59904683e-01 8.23120117e-01
-1.35345189e-02 9.24530685e-01 1.37018931e+00 -2.64062941e-01
1.07849732e-01 1.83024496e-01 2.07988128e-01 6.38900936e-01
3.60750467e-01 2.25165929e-03 -5.23333788e-01 6.93157688e-02
4.70741808e-01 2.10557327e-01 -3.47912967e-01 3.12390834e-01
-1.06596828e+00 6.59380496e-01 7.57722735e-01 2.71814317e-01
-5.33675611e-01 8.61631259e-02 -6.15964644e-02 -1.62585065e-01
3.11222643e-01 1.69295296e-01 -2.99639434e-01 -1.23434164e-01
-7.00735211e-01 8.63879398e-02 3.11998010e-01 8.08187723e-01
1.07779002e+00 2.41274107e-02 1.04012281e-01 7.90011764e-01
3.49388242e-01 8.48475575e-01 5.60577631e-01 -8.02161157e-01
4.94266689e-01 7.18499124e-01 3.58867466e-01 -8.74998212e-01
-5.18133938e-01 -9.42074656e-02 -6.72564685e-01 1.66984826e-01
3.94088179e-01 -3.13668519e-01 -1.32025468e+00 1.68030810e+00
4.27667707e-01 1.29108310e-01 2.33792052e-01 1.27517962e+00
1.35011363e+00 7.10891485e-01 -4.14317250e-02 -7.07538426e-02
1.40129375e+00 -9.75843430e-01 -7.95220971e-01 -5.43007195e-01
3.43062043e-01 -9.13168311e-01 8.62199962e-01 2.84371436e-01
-1.08394313e+00 -6.26299977e-01 -1.07449198e+00 -1.68819368e-01
-6.75793707e-01 1.95584893e-01 5.65175176e-01 4.97509629e-01
-1.08144581e+00 2.42931336e-01 -8.46984327e-01 -4.41264600e-01
3.57306033e-01 4.75486279e-01 -2.38444507e-01 -3.20041150e-01
-1.16484892e+00 8.39016914e-01 4.65465367e-01 3.62013608e-01
-5.43433964e-01 -1.17729478e-01 -8.16513062e-01 -2.98913062e-01
6.29927278e-01 -7.15872884e-01 1.06139302e+00 -9.00107086e-01
-1.33674145e+00 4.76495057e-01 -4.03703600e-01 -1.06816061e-01
4.07815158e-01 -5.71911693e-01 -2.87428588e-01 2.45947242e-01
1.91824824e-01 9.05100346e-01 7.41873801e-01 -1.41804373e+00
-9.49603975e-01 -4.66070741e-01 2.53550947e-01 5.36157787e-01
-2.03976288e-01 -5.61234951e-02 -6.09637678e-01 -3.35657269e-01
2.67483503e-01 -7.45590448e-01 -2.41404548e-01 -1.48699224e-01
-3.48481238e-01 -1.77677557e-01 9.83447313e-01 -8.55078101e-01
1.02494586e+00 -1.96032286e+00 1.40054002e-01 -5.55921812e-03
1.21923804e-01 5.33153772e-01 1.41848296e-01 8.30363035e-02
3.64565611e-01 2.12689608e-01 -3.37437660e-01 -3.90309006e-01
-2.41778448e-01 3.05817038e-01 8.30901787e-02 3.06559205e-01
2.74470299e-01 8.50550592e-01 -8.98482084e-01 -5.76453567e-01
5.46678662e-01 5.31149149e-01 -2.98726588e-01 2.19878644e-01
1.20718792e-01 6.06243610e-01 -5.37413895e-01 9.59986567e-01
9.22319293e-01 2.81995591e-02 -2.95415550e-01 -3.15780073e-01
-3.66942197e-01 -3.55939642e-02 -1.24317992e+00 1.57851005e+00
-3.12255740e-01 3.40691864e-01 1.36946216e-01 -6.41597331e-01
7.93541133e-01 2.14702070e-01 3.27231199e-01 -8.40492308e-01
5.08234203e-01 4.53730494e-01 -1.16194181e-01 -7.21726239e-01
6.78155422e-01 -1.09760955e-01 -6.88213557e-02 2.44033888e-01
3.76753584e-02 -1.77800417e-01 1.96027502e-01 3.09598837e-02
7.95355439e-01 1.61270991e-01 2.60500968e-01 1.83047593e-01
3.82951826e-01 -1.51166424e-01 8.01570535e-01 7.52463400e-01
-4.90011483e-01 8.52172315e-01 8.51978511e-02 -1.36873111e-01
-6.39390945e-01 -8.33088458e-01 3.37735891e-01 9.89432931e-01
7.85577714e-01 -1.23945892e-01 -7.37450898e-01 -5.42153180e-01
-1.08122282e-01 5.46891153e-01 -5.97886324e-01 -4.16001417e-02
-4.07803446e-01 -7.75696456e-01 3.19165051e-01 8.03152680e-01
1.08538604e+00 -9.60241139e-01 -1.10280049e+00 -1.44476935e-01
-6.08807147e-01 -1.21164286e+00 -2.51595348e-01 3.37691933e-01
-6.02162719e-01 -1.04242969e+00 -5.83229423e-01 -3.76890451e-01
5.60738087e-01 8.45509350e-01 7.13350534e-01 1.62398472e-01
-7.85946324e-02 2.87708104e-01 -6.72306418e-01 -6.43466532e-01
2.86893338e-01 -1.33521467e-01 -4.99958210e-02 1.11113898e-01
3.58048916e-01 -2.69744992e-01 -7.57924914e-01 4.16836143e-01
-9.01585281e-01 9.17569622e-02 6.29939556e-01 6.01158202e-01
4.42585379e-01 -2.05762371e-01 1.96144074e-01 -2.53584296e-01
1.59244165e-01 -6.24249458e-01 -3.68122190e-01 2.21295238e-01
-9.13257897e-02 -1.53230265e-01 3.54627758e-01 -2.69407898e-01
-1.21002138e+00 2.26521522e-01 -6.94739223e-02 -6.54188573e-01
-5.25143623e-01 3.21372807e-01 -3.24456960e-01 -1.79503988e-02
4.99852151e-01 1.31764352e-01 -2.05236465e-01 -3.44533294e-01
1.81638867e-01 8.57915938e-01 5.98675072e-01 -4.34825182e-01
6.36516690e-01 6.55418038e-01 -3.01721334e-01 -8.47897291e-01
-6.69659376e-01 -6.56411469e-01 -5.73925734e-01 -3.79362881e-01
1.25030303e+00 -8.03122163e-01 -5.63270926e-01 7.24500895e-01
-1.08077681e+00 -1.04350947e-01 5.30277751e-03 6.54783666e-01
-2.19879970e-01 3.83172303e-01 -2.86133319e-01 -1.10283506e+00
-3.29482883e-01 -1.37991619e+00 1.19536793e+00 8.66351545e-01
3.04663241e-01 -6.07748151e-01 -2.88034916e-01 6.64793432e-01
2.81431079e-01 3.03387940e-01 7.04365820e-02 -3.68451834e-01
-4.85728115e-01 -8.75519961e-02 -3.52575988e-01 3.32140595e-01
1.38388917e-01 -2.08178014e-02 -1.28636742e+00 -1.11497641e-01
-2.44779494e-02 -3.17154825e-01 1.18972325e+00 3.47898215e-01
7.78152406e-01 1.89936999e-03 -3.07399213e-01 6.86811984e-01
1.30230868e+00 5.22816718e-01 5.22181451e-01 2.91535944e-01
8.98856640e-01 5.95906734e-01 7.75371730e-01 4.10569042e-01
5.60652316e-01 5.95631659e-01 7.58269846e-01 -2.00689718e-01
4.81916182e-02 1.34488910e-01 3.83201808e-01 7.01979995e-01
-5.05746782e-01 -4.00912464e-01 -1.04675710e+00 5.76598525e-01
-2.04051900e+00 -6.62026763e-01 -2.08395317e-01 2.00803685e+00
3.66391897e-01 1.50516853e-02 1.84434265e-01 1.77221254e-01
8.95669580e-01 5.78324310e-02 -5.84633231e-01 -2.28496656e-01
-2.97591299e-01 4.91938740e-02 5.86540103e-01 1.92449957e-01
-1.30502367e+00 9.41136956e-01 4.51957130e+00 6.63815379e-01
-1.23947942e+00 3.54799151e-01 3.95988971e-01 -3.92016172e-02
-1.35792285e-01 1.15684204e-01 -5.85209846e-01 6.82859302e-01
4.59793210e-01 3.04721445e-01 3.79781604e-01 4.69488174e-01
3.84408645e-02 -7.88085520e-01 -4.90041554e-01 1.19185424e+00
1.50832325e-01 -6.82725251e-01 -1.46082237e-01 -4.53119688e-02
5.53487897e-01 2.26559713e-02 -5.56965657e-02 1.17036723e-01
1.58929750e-01 -8.32890749e-01 9.30813789e-01 7.11462796e-01
3.80969316e-01 -6.16744876e-01 1.18134415e+00 4.48188037e-01
-1.45252919e+00 -1.82998180e-01 -1.72250122e-01 -8.68530571e-02
3.08171332e-01 5.04054487e-01 -3.83123457e-01 1.05680704e+00
1.06201816e+00 5.50685227e-01 -6.34138227e-01 1.20411420e+00
-3.36129487e-01 4.30174977e-01 -7.05784976e-01 6.56452924e-02
3.41996998e-01 -2.15264216e-01 4.94476527e-01 1.08287954e+00
5.33756793e-01 4.24883991e-01 3.60800266e-01 5.21308661e-01
2.43366301e-01 -1.81384906e-01 -4.46171910e-01 2.42949158e-01
3.33608717e-01 1.32602167e+00 -9.64876592e-01 -4.48364139e-01
-5.59455693e-01 9.94796634e-01 1.67141899e-01 5.81621706e-01
-1.01578593e+00 -4.58512843e-01 5.49857855e-01 -2.38335937e-01
5.37069440e-01 -2.21336052e-01 -3.70796144e-01 -1.19438887e+00
1.56851128e-01 -3.26945275e-01 3.92583191e-01 -1.06293368e+00
-8.97823632e-01 7.54085720e-01 1.16509832e-01 -1.20451200e+00
5.24557494e-02 -5.25423408e-01 -5.14372468e-01 8.75766158e-01
-1.54872382e+00 -1.14671576e+00 -7.94560671e-01 6.69516265e-01
3.50710839e-01 2.60262132e-01 2.86416411e-01 3.25295419e-01
-9.16372359e-01 3.56076002e-01 -2.22508594e-01 -1.23057373e-01
5.28894067e-01 -1.20445514e+00 -1.30415246e-01 1.18761349e+00
-1.71201125e-01 3.22802871e-01 5.55416346e-01 -6.46338105e-01
-1.20830190e+00 -9.23629880e-01 5.77484488e-01 -2.13956639e-01
2.11359352e-01 -1.49679884e-01 -9.13557470e-01 2.96704948e-01
2.65353501e-01 2.14120969e-01 4.46766198e-01 -3.60246927e-01
-2.23568171e-01 -1.27546057e-01 -1.27428377e+00 3.83983523e-01
8.57787371e-01 -2.09307790e-01 -5.21368921e-01 -6.19018339e-02
6.14383042e-01 -5.76788187e-01 -6.71378613e-01 4.98275965e-01
4.02012646e-01 -1.19125450e+00 7.80894876e-01 1.76783781e-02
1.97841302e-01 -7.97412276e-01 -5.00477135e-01 -1.14466703e+00
-9.59631950e-02 -4.61675674e-01 -8.47332999e-02 1.11101186e+00
1.47140741e-01 -8.01394939e-01 3.55065554e-01 5.05062342e-01
-4.38737601e-01 -7.47711837e-01 -9.59271610e-01 -4.31846172e-01
-4.07724738e-01 -4.57740188e-01 5.74057937e-01 7.91339338e-01
-2.61002332e-01 1.50960907e-01 -3.53806287e-01 1.71054140e-01
3.75147730e-01 1.85862750e-01 6.07811928e-01 -1.01896322e+00
-5.38554192e-02 -5.01745701e-01 -3.68585914e-01 -8.96719754e-01
-2.47879609e-01 -4.48084474e-01 2.64782995e-01 -1.89644217e+00
2.48667583e-01 -2.74993032e-01 -3.88997614e-01 7.85152793e-01
-7.40075409e-01 4.74703699e-01 5.48109889e-01 2.02799499e-01
-6.73590183e-01 6.57765031e-01 1.12456083e+00 1.05145641e-01
-2.47829139e-01 -1.91193029e-01 -7.10400581e-01 8.02075684e-01
1.19653583e+00 -1.02958530e-01 -2.83378392e-01 -4.97109056e-01
2.96322592e-02 -7.47360662e-02 5.50104320e-01 -1.13733661e+00
2.95538604e-01 -1.87035009e-01 4.97258753e-01 -6.45559788e-01
6.93372071e-01 -7.56534040e-01 -3.20940912e-02 3.65731686e-01
3.70196968e-01 7.79185584e-03 2.98176706e-01 4.97572154e-01
-3.22650045e-01 -1.59833431e-01 7.11617410e-01 -2.77940333e-01
-1.07274175e+00 4.89100926e-02 -2.77865380e-01 5.60874073e-03
1.11905158e+00 -6.06697142e-01 -6.16005123e-01 -3.55398208e-01
-6.21270835e-01 4.92599607e-01 4.40467954e-01 4.49315310e-01
7.51299858e-01 -1.14147329e+00 -4.91759092e-01 3.61245722e-02
6.63838116e-03 1.52721301e-01 5.11341870e-01 1.16863656e+00
-2.44897991e-01 1.44458458e-01 -4.01723415e-01 -7.00130701e-01
-1.33893287e+00 4.71373230e-01 2.57146209e-01 2.40223333e-02
-8.50272328e-02 9.57079709e-01 2.71976292e-01 -1.61221579e-01
-9.71969124e-03 -5.16487896e-01 -3.75970542e-01 3.34545642e-01
3.39177042e-01 5.24833441e-01 2.59976182e-02 -1.09305370e+00
-6.27187133e-01 9.21647906e-01 1.17023870e-01 -2.16391832e-01
1.10495651e+00 -3.52623314e-01 2.41129496e-03 4.45090920e-01
9.98045146e-01 -3.68258774e-01 -1.31854188e+00 -2.61292577e-01
-4.23433006e-01 -5.39236248e-01 1.05634332e-01 -8.18156004e-01
-1.19612050e+00 9.88412619e-01 7.22066283e-01 7.17439577e-02
1.66851294e+00 -2.86428817e-02 8.40751290e-01 1.03323407e-01
4.40948188e-01 -1.04624259e+00 4.13153581e-02 4.26036000e-01
7.59029269e-01 -1.45309103e+00 3.63258049e-02 -4.32026207e-01
-8.77398431e-01 8.12766612e-01 8.92627120e-01 7.06462488e-02
4.51527536e-01 1.74477533e-01 5.44402525e-02 -2.96910591e-02
-5.52415192e-01 -8.84403884e-01 3.05860400e-01 5.08137286e-01
1.11591578e-01 6.06424622e-02 -2.09984913e-01 6.90082371e-01
-7.13578761e-02 -1.07019782e-01 2.46096626e-01 1.38385761e+00
-6.36599243e-01 -6.32483959e-01 -7.58067608e-01 5.34107745e-01
-3.60012978e-01 -5.32804504e-02 -4.40685183e-01 7.07429707e-01
6.62689567e-01 1.38171327e+00 -3.25112864e-02 -7.69358873e-01
3.41591895e-01 -3.36041935e-02 4.35629755e-01 -2.44485915e-01
-6.89523578e-01 3.46149415e-01 -1.45590855e-02 -6.33836031e-01
-7.72867322e-01 -7.06517279e-01 -1.24535322e+00 -1.00268140e-01
-5.96784294e-01 -1.31169735e-02 7.52718925e-01 9.70307171e-01
2.46453822e-01 5.44034719e-01 3.60914439e-01 -1.31383109e+00
6.25045821e-02 -1.06492925e+00 -3.37923378e-01 2.87441522e-01
4.36951905e-01 -1.13281548e+00 -3.51083666e-01 -6.38995394e-02] | [9.488916397094727, -0.9647848010063171] |
c50158e5-49b8-4697-9edf-3d379b593d7f | conditional-density-estimation-with-bayesian | 1802.04908 | null | http://arxiv.org/abs/1802.04908v1 | http://arxiv.org/pdf/1802.04908v1.pdf | Conditional Density Estimation with Bayesian Normalising Flows | Modeling complex conditional distributions is critical in a variety of
settings. Despite a long tradition of research into conditional density
estimation, current methods employ either simple parametric forms or are
difficult to learn in practice. This paper employs normalising flows as a
flexible likelihood model and presents an efficient method for fitting them to
complex densities. These estimators must trade-off between modeling
distributional complexity, functional complexity and heteroscedasticity without
overfitting. We recognize these trade-offs as modeling decisions and develop a
Bayesian framework for placing priors over these conditional density estimators
using variational Bayesian neural networks. We evaluate this method on several
small benchmark regression datasets, on some of which it obtains state of the
art performance. Finally, we apply the method to two spatial density modeling
tasks with over 1 million datapoints using the New York City yellow taxi
dataset and the Chicago crime dataset. | ['Brian L. Trippe', 'Richard E. Turner'] | 2018-02-14 | null | null | null | null | ['normalising-flows'] | ['methodology'] | [-1.77537456e-01 -1.58115417e-01 -5.85262477e-01 -7.62144625e-01
-9.49479580e-01 -3.91667247e-01 7.45555341e-01 -1.61906481e-01
-6.33382082e-01 1.11997199e+00 2.44676664e-01 -7.14280605e-01
-2.92504430e-01 -9.56439793e-01 -7.23818898e-01 -4.12474990e-01
-1.95349857e-01 7.15658069e-01 -9.00810137e-02 4.31284159e-01
2.68899351e-01 4.08082962e-01 -1.07677794e+00 -3.25696677e-01
1.14468598e+00 9.44301724e-01 -1.62586328e-02 5.31249344e-01
-4.85356636e-02 7.69029617e-01 -2.73781151e-01 -4.83740538e-01
-4.69578244e-02 1.51947901e-01 -4.11983073e-01 -3.52012999e-02
4.69366759e-01 -7.22672462e-01 -2.24442482e-01 9.53118145e-01
2.09980801e-01 3.30907673e-01 1.73977911e+00 -1.33614659e+00
-8.02493751e-01 6.45107448e-01 -8.12947094e-01 5.13034999e-01
-1.79019481e-01 -2.01302450e-02 1.02007842e+00 -9.79451180e-01
8.64289328e-02 1.42057395e+00 9.41395938e-01 4.14559431e-02
-1.79423392e+00 -6.93726599e-01 8.89300741e-03 -9.43611637e-02
-1.79318261e+00 -6.38882339e-01 2.62744069e-01 -9.88603771e-01
1.13383615e+00 -4.44451064e-01 3.21395874e-01 1.26392198e+00
4.16929394e-01 6.59624159e-01 1.00374699e+00 -4.20155451e-02
5.29236019e-01 3.23112845e-01 2.40409032e-01 2.37727687e-01
4.15414661e-01 2.53126025e-01 -2.74130464e-01 -5.56639135e-01
1.11423576e+00 2.30882205e-02 -2.83252280e-02 -2.82353342e-01
-4.75008100e-01 1.44731367e+00 -7.00603938e-03 -1.84562832e-01
-2.34381348e-01 6.06351674e-01 -1.47913052e-02 -1.20599464e-01
7.87982285e-01 -1.93073526e-01 -3.10634255e-01 -3.06154460e-01
-1.07495689e+00 4.32133973e-01 1.04031515e+00 9.04409528e-01
9.18752253e-01 1.95371434e-01 -9.50037688e-02 1.05415666e+00
7.72474229e-01 6.85284734e-01 -1.62923440e-01 -9.44945037e-01
4.93036360e-01 5.87637126e-02 3.42355818e-01 -8.73237908e-01
-3.16840798e-01 1.45071726e-02 -8.99293244e-01 9.69278365e-02
6.56559289e-01 -5.35011292e-01 -8.89441073e-01 1.84366083e+00
1.22094870e-01 2.75196463e-01 -4.18751568e-01 4.11328971e-01
2.74630427e-01 8.12712729e-01 4.71312970e-01 -2.72831321e-02
1.03314626e+00 -2.84154087e-01 -6.66531563e-01 -3.28956366e-01
4.41184193e-01 -4.39164758e-01 8.76694262e-01 2.87000924e-01
-1.17201710e+00 -2.32534513e-01 -8.04711699e-01 2.45549697e-02
-2.45289817e-01 -3.59754004e-02 7.75914013e-01 1.03202486e+00
-1.03762639e+00 5.92741609e-01 -1.22144639e+00 -3.22790474e-01
7.10590959e-01 -4.66906801e-02 1.14368219e-02 -2.68883677e-03
-1.02455735e+00 9.74231422e-01 2.93727070e-01 -2.21481994e-01
-8.42690349e-01 -1.26561546e+00 -1.08864558e+00 5.37850678e-01
2.19277248e-01 -6.03035450e-01 1.39437997e+00 -5.11595428e-01
-1.28704023e+00 3.85589063e-01 -1.76051393e-01 -4.40904468e-01
5.44774950e-01 -1.64664820e-01 -3.50304902e-01 -2.24493623e-01
2.46009469e-01 6.95869148e-01 8.15622926e-01 -7.71337748e-01
-6.40330732e-01 -1.64218217e-01 -3.19641799e-01 -8.71425942e-02
-3.16951245e-01 -4.78683151e-02 -2.26914167e-01 -5.75654984e-01
-3.66403311e-01 -6.31689847e-01 -1.14599332e-01 1.06410541e-01
-3.15009028e-01 -4.58268106e-01 3.72824371e-01 -5.16221702e-01
1.36639726e+00 -2.03550553e+00 -1.86283275e-01 4.78736281e-01
5.10413162e-02 -3.89153391e-01 4.05375399e-02 2.18966424e-01
1.08675987e-01 2.78729767e-01 -6.10403180e-01 -5.32745957e-01
4.41441774e-01 2.76568264e-01 -4.39292461e-01 7.03843355e-01
4.01078939e-01 7.60510683e-01 -3.97174507e-01 -5.17696261e-01
2.52181470e-01 7.62122512e-01 -8.98028195e-01 1.57151535e-01
-1.33519262e-01 -7.97281489e-02 -1.32669061e-01 5.19496143e-01
9.00301218e-01 -4.15642500e-01 2.73963869e-01 1.14074484e-01
-4.10180818e-03 2.70699233e-01 -1.44307613e+00 1.17419577e+00
-2.64285922e-01 5.94477713e-01 9.17394236e-02 -1.10431790e+00
6.47288501e-01 -2.46864697e-03 3.03657174e-01 -3.34777832e-01
1.05284177e-01 -2.72921957e-02 -3.89908373e-01 -2.42853463e-01
6.10043108e-01 -4.63089973e-01 -1.01637803e-01 5.06059468e-01
2.96756268e-01 -8.70591786e-04 2.79456943e-01 2.79913656e-02
8.85236740e-01 2.47637793e-01 5.00881374e-01 -6.17230952e-01
-2.83150375e-01 -2.84839332e-01 5.08685112e-01 1.10364258e+00
-1.94698691e-01 5.85433304e-01 8.60362291e-01 -2.07287356e-01
-8.96782041e-01 -1.79173851e+00 -7.87651479e-01 1.13222802e+00
-4.97663736e-01 -1.25550315e-01 -4.78680044e-01 -3.25392783e-01
3.74399245e-01 9.51098740e-01 -6.91676199e-01 2.10720524e-02
-1.30590126e-01 -1.12556291e+00 6.99165404e-01 8.95799875e-01
2.69226104e-01 -7.32933283e-01 -3.99760395e-01 2.95449108e-01
1.56506319e-02 -8.81814182e-01 -3.21561992e-01 2.96384782e-01
-7.68149495e-01 -8.08136523e-01 -8.14628541e-01 -2.80942470e-01
1.62265390e-01 -9.73066613e-02 1.53838027e+00 -3.07241440e-01
-1.10646449e-01 5.46949267e-01 1.99602708e-01 -4.62197900e-01
-2.19361797e-01 4.11356330e-01 1.44197389e-01 -2.94612795e-01
8.12564015e-01 -8.07112277e-01 -3.59008819e-01 8.36443231e-02
-9.21817780e-01 -5.20409763e-01 3.68472397e-01 7.53539920e-01
3.15066785e-01 -4.31192368e-02 8.17870259e-01 -8.15191865e-01
8.31945479e-01 -1.20809102e+00 -1.15777278e+00 1.46509513e-01
-5.25593638e-01 -8.78678411e-02 7.97554925e-02 -4.28496599e-01
-1.54714131e+00 -2.55596131e-01 -1.69287398e-01 -2.03161314e-01
-4.40509498e-01 1.04664481e+00 1.11457869e-01 4.69321430e-01
6.74983561e-01 -1.92942157e-01 -1.73027083e-01 -4.56228048e-01
3.33807617e-01 5.35602689e-01 4.80767339e-01 -8.60010087e-01
4.81810361e-01 3.47172529e-01 -1.10762469e-01 -9.47271407e-01
-7.45610178e-01 -2.07382545e-01 -6.45093143e-01 1.54938146e-01
9.90941584e-01 -1.12325668e+00 -6.87508285e-01 4.12284076e-01
-9.15908098e-01 -6.50207996e-01 -5.12063503e-02 6.40642703e-01
-6.95913255e-01 2.58213937e-01 -7.00500131e-01 -1.24019516e+00
1.87361583e-01 -8.46289456e-01 8.33180249e-01 2.70789713e-01
-2.62931645e-01 -1.49209571e+00 4.77138847e-01 -1.86384350e-01
5.94427168e-01 -1.08197458e-01 1.17426741e+00 -3.17856997e-01
-5.54132342e-01 1.81305051e-01 -5.39781868e-01 2.75057197e-01
1.15765512e-01 3.82072389e-01 -9.42978024e-01 -6.98647052e-02
-3.51056844e-01 -4.22511667e-01 1.13038504e+00 1.19819319e+00
1.18001640e+00 -1.60147399e-01 -3.85196090e-01 7.90099025e-01
1.43536615e+00 3.05498987e-02 8.52384508e-01 -1.59421474e-01
4.60025489e-01 5.61929643e-01 1.11424409e-01 7.89561629e-01
7.35510468e-01 2.56075978e-01 9.11582708e-02 2.20678821e-01
6.49727762e-01 -3.84406149e-01 3.32752943e-01 2.31761113e-01
2.43423238e-01 -2.34862879e-01 -1.21351779e+00 7.36129284e-01
-1.84885740e+00 -1.21131825e+00 4.27552685e-02 2.09799528e+00
9.87855613e-01 9.05585140e-02 4.62468803e-01 -5.11509061e-01
6.13567114e-01 -1.04886116e-02 -6.11361444e-01 -1.30163401e-01
-6.26164079e-02 2.90410608e-01 5.74471712e-01 6.60760283e-01
-1.36652303e+00 8.62793505e-01 7.95399046e+00 1.01304948e+00
-6.99993312e-01 -6.17565960e-03 1.07044053e+00 -1.48179084e-01
-2.99202055e-01 -2.00454369e-02 -9.94865358e-01 4.63257372e-01
1.21663654e+00 -5.85144721e-02 4.65615541e-01 8.95869613e-01
1.13563076e-01 -5.39055169e-01 -1.23938549e+00 8.38863671e-01
-2.92220920e-01 -1.37218344e+00 -2.51193196e-01 2.93115258e-01
6.75301671e-01 2.61256129e-01 4.24841613e-01 6.41635835e-01
1.12476695e+00 -1.69984496e+00 5.83779037e-01 7.51921594e-01
8.86201203e-01 -1.04237640e+00 4.59161162e-01 3.90801430e-01
-9.63395238e-01 -1.64808668e-02 -6.67350590e-01 -3.07390392e-01
5.11728883e-01 6.70074761e-01 -5.76955020e-01 -1.38960332e-01
9.19171035e-01 8.59442890e-01 -3.16792458e-01 1.15285242e+00
3.66717391e-02 1.09588087e+00 -7.42629170e-01 1.27560109e-01
1.73196763e-01 -5.65740526e-01 4.74028401e-02 1.35895860e+00
5.69710255e-01 -1.58501983e-01 -8.69839489e-02 1.59058452e+00
3.37649025e-02 -3.07141870e-01 -8.07680547e-01 -6.72003403e-02
8.31586242e-01 1.00373018e+00 -5.53458214e-01 -2.28185415e-01
-7.44671226e-01 1.59695074e-01 7.35551894e-01 8.49370718e-01
-1.02281559e+00 -8.37820619e-02 8.80431235e-01 9.41703692e-02
5.41804314e-01 -3.64376307e-01 -3.90863895e-01 -1.32742834e+00
-3.28023553e-01 -3.69935781e-01 4.67870981e-01 -6.62948191e-01
-2.16376710e+00 -1.65493041e-01 8.31278443e-01 -5.43346167e-01
-6.10902727e-01 -7.70612180e-01 -6.62334561e-01 1.11754036e+00
-1.50498617e+00 -7.42748618e-01 1.02626607e-01 4.11377639e-01
1.11061931e-01 -4.52739038e-02 5.73508739e-01 4.85896647e-01
-5.68571448e-01 4.00612593e-01 3.35926443e-01 1.06275477e-01
6.53754592e-01 -1.36819148e+00 5.73970616e-01 5.91633558e-01
-3.90400030e-02 6.57841563e-01 3.56477827e-01 -7.20937490e-01
-8.78226280e-01 -9.01527524e-01 7.54761815e-01 -6.15348399e-01
8.20875108e-01 -4.94939059e-01 -1.07171953e+00 1.24006975e+00
2.25737616e-01 -3.81480783e-01 9.20986652e-01 5.73607922e-01
-4.02274311e-01 2.70740628e-01 -1.14795661e+00 4.86456156e-01
6.42446995e-01 -4.74246860e-01 -4.51387852e-01 -5.01719266e-02
2.51674026e-01 5.01526818e-02 -9.61650133e-01 1.77341461e-01
5.97934008e-01 -9.50747311e-01 1.02728808e+00 -5.62780678e-01
4.07899857e-01 2.30339002e-02 -5.19499540e-01 -1.43836665e+00
-4.43904221e-01 -3.36144537e-01 -2.15556562e-01 1.40364826e+00
4.92247373e-01 -6.44610703e-01 7.79238701e-01 1.08568680e+00
2.48319551e-01 -4.78237182e-01 -1.24049282e+00 -6.86647654e-01
9.45571899e-01 -8.80879343e-01 5.79342067e-01 9.40307498e-01
-1.65218100e-01 2.20698386e-01 -4.91747022e-01 -6.51489720e-02
8.40003312e-01 -5.87968886e-01 7.78522491e-01 -1.58391321e+00
-1.03663668e-01 -5.80144823e-01 -5.24873100e-02 -1.38661277e+00
5.58427930e-01 -6.25029504e-01 -2.45330483e-02 -1.30410600e+00
4.05426919e-01 -4.27984715e-01 9.89511013e-02 3.10003698e-01
-1.47407889e-01 -3.77042294e-02 -2.59096891e-01 -1.67401984e-01
-2.88787276e-01 8.27475667e-01 3.99492264e-01 -1.05406202e-01
-1.29786044e-01 -9.08665417e-04 -7.72133350e-01 9.85148072e-01
7.37808645e-01 -6.21604800e-01 -6.47783875e-01 -4.47222382e-01
3.59248102e-01 1.93669736e-01 6.45564616e-01 -6.57259881e-01
1.56214327e-01 -6.39892042e-01 7.54833281e-01 -9.25237179e-01
3.85780185e-01 -7.66582429e-01 9.44229215e-02 -2.28169814e-01
-1.63747370e-01 1.34031266e-01 3.06599110e-01 7.97660410e-01
5.07725105e-02 -2.58369803e-01 9.23607111e-01 1.09498054e-01
-4.92599487e-01 4.55211043e-01 -9.94309664e-01 4.16227341e-01
6.21892512e-01 -1.92590639e-01 -3.65117222e-01 -6.03990078e-01
-6.24664962e-01 3.52874607e-01 1.83857098e-01 1.02250069e-01
4.80330437e-01 -1.38292718e+00 -7.50875592e-01 2.22062901e-01
-1.27249628e-01 -7.50958622e-02 2.62817919e-01 7.86477089e-01
-2.58977294e-01 2.42130235e-01 -1.00449264e-01 -8.82012606e-01
-4.44656909e-01 2.21674174e-01 4.39463496e-01 -2.90735960e-01
-2.72687823e-01 4.66018766e-01 4.28498447e-01 -6.63957298e-01
3.19622248e-01 -4.14681435e-01 -1.25699550e-01 1.16794519e-01
4.54543263e-01 5.01866102e-01 -4.91482615e-01 -4.60628718e-01
-1.71835616e-01 1.57445878e-01 7.58828744e-02 -4.15973544e-01
1.35849917e+00 -3.13553125e-01 1.95079148e-01 7.14189827e-01
1.15733111e+00 -5.08545518e-01 -1.71208274e+00 -2.17672750e-01
1.39204502e-01 -5.40218472e-01 8.12329054e-02 -4.23321426e-01
-8.36237431e-01 1.02630055e+00 3.49159002e-01 2.02242970e-01
6.41742408e-01 -9.18917879e-02 1.59523726e-01 5.37061214e-01
1.00733988e-01 -1.22492683e+00 -6.17115200e-02 7.63389111e-01
4.52913851e-01 -1.38689899e+00 6.66994676e-02 -2.12288294e-02
-5.74133039e-01 8.73042822e-01 6.30561650e-01 -3.67406577e-01
1.38658965e+00 5.67100167e-01 -2.92979538e-01 -7.48874024e-02
-7.53489971e-01 2.65040826e-02 3.11812729e-01 9.18449104e-01
5.44057071e-01 1.08056493e-01 2.27721557e-01 7.72467554e-01
-1.10645361e-01 1.31807700e-01 5.59571683e-01 6.94070697e-01
-4.00153339e-01 -6.01441145e-01 -2.60809839e-01 8.85246873e-01
-5.91092825e-01 -3.66144031e-01 2.37354845e-01 1.07913184e+00
-4.91113961e-01 9.69588935e-01 6.57002568e-01 3.78934950e-01
-6.36009276e-02 2.00595424e-01 2.24283472e-01 -5.03972232e-01
-4.90766158e-03 3.21389079e-01 2.66878195e-02 -1.77879453e-01
-4.16540831e-01 -7.92247176e-01 -7.09704876e-01 -8.02118242e-01
-1.47476733e-01 -1.61959425e-01 2.80828536e-01 9.54596639e-01
2.33077914e-01 1.32159784e-01 7.47913420e-02 -9.28125679e-01
-8.74975801e-01 -1.19604588e+00 -1.13126564e+00 9.60959718e-02
3.25534970e-01 -1.07785654e+00 -4.20703143e-01 1.18724011e-01] | [7.134440898895264, 3.968278646469116] |
1df3186a-f0d4-4df7-b1e8-a9b0c8401b17 | niff-alleviating-forgetting-in-generalized | 2303.04958 | null | https://arxiv.org/abs/2303.04958v1 | https://arxiv.org/pdf/2303.04958v1.pdf | NIFF: Alleviating Forgetting in Generalized Few-Shot Object Detection via Neural Instance Feature Forging | Privacy and memory are two recurring themes in a broad conversation about the societal impact of AI. These concerns arise from the need for huge amounts of data to train deep neural networks. A promise of Generalized Few-shot Object Detection (G-FSOD), a learning paradigm in AI, is to alleviate the need for collecting abundant training samples of novel classes we wish to detect by leveraging prior knowledge from old classes (i.e., base classes). G-FSOD strives to learn these novel classes while alleviating catastrophic forgetting of the base classes. However, existing approaches assume that the base images are accessible, an assumption that does not hold when sharing and storing data is problematic. In this work, we propose the first data-free knowledge distillation (DFKD) approach for G-FSOD that leverages the statistics of the region of interest (RoI) features from the base model to forge instance-level features without accessing the base images. Our contribution is three-fold: (1) we design a standalone lightweight generator with (2) class-wise heads (3) to generate and replay diverse instance-level base features to the RoI head while finetuning on the novel data. This stands in contrast to standard DFKD approaches in image classification, which invert the entire network to generate base images. Moreover, we make careful design choices in the novel finetuning pipeline to regularize the model. We show that our approach can dramatically reduce the base memory requirements, all while setting a new standard for G-FSOD on the challenging MS-COCO and PASCAL-VOC benchmarks. | ['Juergen Beyerer', 'Bin Yang', 'Matthias Kayser', 'George Eskandar', 'Johannes Meier', 'Karim Guirguis'] | 2023-03-09 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Guirguis_NIFF_Alleviating_Forgetting_in_Generalized_Few-Shot_Object_Detection_via_Neural_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Guirguis_NIFF_Alleviating_Forgetting_in_Generalized_Few-Shot_Object_Detection_via_Neural_CVPR_2023_paper.pdf | cvpr-2023-1 | ['few-shot-object-detection'] | ['computer-vision'] | [ 3.69568676e-01 2.75671482e-01 -1.50109246e-01 -4.37982589e-01
-8.51966262e-01 -6.40122056e-01 7.23998427e-01 6.10711798e-02
-7.17053831e-01 9.70813453e-01 -1.58211365e-02 -1.36147350e-01
-4.27039191e-02 -9.44386542e-01 -1.07572663e+00 -6.42253697e-01
5.44609018e-02 3.69548321e-01 2.31133431e-01 5.24515007e-03
1.82488158e-01 6.37974262e-01 -1.85286748e+00 4.99916881e-01
6.78279877e-01 1.04499698e+00 -4.34896462e-02 7.44244993e-01
1.96325928e-01 7.87640154e-01 -6.74359560e-01 -7.06786275e-01
6.64677322e-01 -3.01728219e-01 -7.45732486e-01 -1.36393517e-01
8.42951238e-01 -7.11498916e-01 -4.93710458e-01 9.66357887e-01
6.07267261e-01 1.35458529e-01 4.91866827e-01 -1.52714121e+00
-6.73395634e-01 6.01368070e-01 -3.06292206e-01 2.71236658e-01
-1.95837051e-01 6.57198250e-01 8.04245770e-01 -7.76697814e-01
9.02739942e-01 1.13865578e+00 7.22666860e-01 9.41083550e-01
-1.28505194e+00 -8.06944728e-01 9.20389071e-02 3.43441606e-01
-1.42450023e+00 -8.63196313e-01 3.22604626e-01 -4.11577851e-01
9.46928382e-01 4.71316695e-01 5.25171041e-01 1.33702385e+00
3.35799977e-02 8.00254822e-01 1.03705645e+00 -3.76901448e-01
5.30056536e-01 5.15722811e-01 2.70681888e-01 6.72527790e-01
6.07936263e-01 2.80571550e-01 -7.29627550e-01 -2.09963709e-01
3.13373774e-01 2.08823442e-01 -2.66699176e-02 -6.61205232e-01
-8.99222732e-01 8.13224018e-01 4.96959925e-01 -2.21337359e-02
-1.97150633e-01 2.43872955e-01 3.68384928e-01 2.16587424e-01
3.92132550e-01 5.41675806e-01 -4.06752527e-01 -2.06434596e-02
-9.62989986e-01 6.00911617e-01 1.07116795e+00 1.02581382e+00
1.03903508e+00 -3.08159411e-01 -5.96999347e-01 4.79504317e-01
-3.71626168e-01 3.35707784e-01 3.73948932e-01 -9.83363986e-01
2.94547558e-01 4.42939103e-01 -2.81333458e-02 -8.72746468e-01
-5.85490391e-02 -3.14586192e-01 -6.28962755e-01 3.15888673e-01
5.03004789e-01 -2.27538034e-01 -1.32456493e+00 1.94306743e+00
5.29585421e-01 1.85812503e-01 1.70295954e-01 6.30085111e-01
7.28141427e-01 5.14022768e-01 -4.10880819e-02 1.84594169e-01
1.05476630e+00 -9.05982018e-01 -2.71631449e-01 -4.00979161e-01
5.32928109e-01 4.07498404e-02 1.12688291e+00 3.69615585e-01
-8.42876434e-01 -2.66654521e-01 -1.20040071e+00 -3.45477670e-01
-8.84186804e-01 -1.16735488e-01 8.94761384e-01 8.15310359e-01
-8.74332488e-01 6.89047575e-01 -7.10920155e-01 -2.51008630e-01
1.13918293e+00 4.65423673e-01 -4.62935627e-01 -2.89602667e-01
-9.80285466e-01 8.12428176e-01 5.01423478e-01 -1.03196196e-01
-1.27673602e+00 -1.05663145e+00 -7.00099051e-01 2.43780427e-02
7.07425237e-01 -6.95116937e-01 1.09345841e+00 -1.02853107e+00
-1.05508947e+00 9.09718633e-01 3.90821919e-02 -8.14319134e-01
7.36095190e-01 -2.39454985e-01 -1.83255330e-01 7.38582835e-02
5.93582876e-02 1.10799551e+00 1.15264738e+00 -1.14995658e+00
-8.46087277e-01 -3.80147606e-01 7.86860064e-02 -3.61835957e-02
-5.00891805e-01 -5.10964513e-01 -3.03412259e-01 -6.09327555e-01
-5.60140550e-01 -9.02310371e-01 -1.49755850e-01 4.84403312e-01
-5.62738717e-01 -1.15224309e-02 9.70083773e-01 -3.80916834e-01
1.02925491e+00 -2.39811850e+00 -2.29421124e-01 1.08201377e-01
4.53583121e-01 5.74923992e-01 -2.17041790e-01 2.15506703e-02
2.00950354e-01 1.15359826e-02 -2.91814089e-01 -1.49909258e-01
9.11163613e-02 3.67272168e-01 -5.19648194e-01 2.63934404e-01
4.82302934e-01 1.10350442e+00 -8.62304807e-01 -1.13640465e-01
-1.84720501e-01 3.28449845e-01 -8.00407171e-01 -6.60560429e-02
-5.53554595e-01 -1.54458098e-02 -2.11623132e-01 5.04261076e-01
8.05197418e-01 -1.97230801e-01 -1.65365174e-01 -8.42787102e-02
5.19204140e-02 2.05868691e-01 -1.11794651e+00 1.63540530e+00
4.44370620e-02 3.46398115e-01 -2.94798780e-02 -9.06503856e-01
6.38854504e-01 -1.72368616e-01 8.18010941e-02 -6.68746471e-01
-7.65017420e-02 1.89125016e-01 -1.45192876e-01 -3.25980663e-01
5.37833452e-01 -2.19464563e-02 -2.58074880e-01 4.35235888e-01
3.19203019e-01 3.21475044e-02 2.32737958e-01 4.13750648e-01
1.47478676e+00 -5.76256663e-02 2.51299828e-01 -1.33739442e-01
1.61144346e-01 2.00495496e-01 7.52983093e-01 1.26262403e+00
-1.79329172e-01 5.57157636e-01 5.19135535e-01 -7.78494895e-01
-1.06605458e+00 -1.15704346e+00 1.93842426e-02 1.15044236e+00
-1.29386470e-01 -2.76924103e-01 -8.46946359e-01 -1.07497084e+00
4.27322239e-01 8.94448042e-01 -8.69796991e-01 -5.51555216e-01
-2.70535499e-01 -7.42857814e-01 7.07754910e-01 3.28833520e-01
5.65134525e-01 -9.67427850e-01 -8.81728709e-01 6.65818676e-02
2.49831185e-01 -7.17815995e-01 -4.85352397e-01 3.86124760e-01
-4.08871382e-01 -9.65808332e-01 -4.51658040e-01 -3.82092535e-01
6.82449043e-01 2.00975016e-01 9.86562252e-01 -6.69463277e-02
-8.77038002e-01 3.72839987e-01 -4.43720818e-02 -8.98523748e-01
-3.10798436e-01 2.98377931e-01 -6.94350526e-02 8.69273767e-02
7.13930130e-01 -6.01114988e-01 -5.70620894e-01 -1.71749610e-02
-1.06190944e+00 9.77615714e-02 7.70927250e-01 8.45802724e-01
6.49156809e-01 -1.03872560e-01 7.14963257e-01 -1.51216805e+00
3.46096426e-01 -6.12712204e-01 -7.48621523e-01 1.93944111e-01
-7.91344643e-01 1.78873062e-01 4.78582174e-01 -6.03807449e-01
-9.42005038e-01 2.61776060e-01 2.56752998e-01 -5.70871592e-01
-6.35078102e-02 9.85263363e-02 -2.43807822e-01 -4.84743416e-02
1.05372441e+00 1.41884774e-01 1.50991768e-01 -3.50698769e-01
6.52522504e-01 6.77320123e-01 8.89401019e-01 -5.91861188e-01
7.10229576e-01 5.98692298e-01 -1.67269215e-01 -5.74797928e-01
-1.07031178e+00 -6.74743652e-02 -5.12432814e-01 2.12601483e-01
5.94168007e-01 -1.06031096e+00 -5.03920436e-01 5.29044688e-01
-9.05053139e-01 -3.42717379e-01 -9.32191968e-01 1.21618278e-01
-5.48841417e-01 -1.52652085e-01 -3.38504136e-01 -5.03542900e-01
-4.93776470e-01 -8.83628130e-01 7.69464731e-01 3.44132006e-01
-1.75194651e-01 -4.31268543e-01 -8.82147402e-02 3.21367413e-01
4.43454415e-01 3.17084819e-01 8.51112902e-01 -8.88456225e-01
-9.89253759e-01 -1.95695803e-01 -3.11490208e-01 4.94450927e-01
-4.74706404e-02 -4.69840258e-01 -1.44715393e+00 -3.59068692e-01
-2.24371552e-01 -5.96867263e-01 1.16373098e+00 -2.96055041e-02
1.47458446e+00 -6.99432313e-01 -4.78323340e-01 7.53275812e-01
1.18233895e+00 2.51145605e-02 6.40069842e-01 2.38279238e-01
6.16532445e-01 4.28363115e-01 3.89931172e-01 6.06967032e-01
3.24061036e-01 3.12819421e-01 2.56343842e-01 6.96955994e-02
-2.95286268e-01 -3.93337220e-01 3.64128411e-01 -1.63644031e-02
4.22415942e-01 -2.04887599e-01 -6.21464849e-01 8.68529558e-01
-1.81429994e+00 -9.91216779e-01 4.42624152e-01 2.38219285e+00
1.26070726e+00 1.38326392e-01 1.12422243e-01 -2.11652756e-01
6.06771946e-01 -4.49374616e-02 -1.14329147e+00 -3.52554828e-01
-6.39682356e-03 4.61598426e-01 8.03984046e-01 2.13122964e-01
-1.11653078e+00 8.83814275e-01 5.77656651e+00 8.75895023e-01
-1.06466341e+00 1.71398014e-01 8.44298005e-01 -6.53959513e-01
-2.23619208e-01 2.29806155e-02 -1.30898833e+00 3.60231012e-01
1.07575262e+00 -4.61962491e-01 6.83737397e-01 1.02212012e+00
-4.34918046e-01 -1.07223488e-01 -1.49543965e+00 9.85771835e-01
2.46997818e-01 -1.60869586e+00 2.09912598e-01 1.87690128e-02
6.55638337e-01 1.01940393e-01 2.58723289e-01 3.70698661e-01
5.99538803e-01 -1.09531450e+00 7.01107144e-01 6.67175472e-01
9.78762448e-01 -8.77773702e-01 2.16485679e-01 4.35117334e-01
-4.89528805e-01 -3.46597314e-01 -5.23986161e-01 6.42177239e-02
-4.00650948e-01 9.28514183e-01 -1.17231274e+00 2.08636671e-01
9.58305657e-01 4.63944227e-01 -8.89846742e-01 1.01802897e+00
5.91008142e-02 4.84081686e-01 -4.79497135e-01 1.34022251e-01
1.45975560e-01 3.47006440e-01 5.44068992e-01 1.11225915e+00
2.44734049e-01 -5.53008653e-02 -2.41023749e-01 1.23730695e+00
-5.10909379e-01 -4.10508126e-01 -7.79506385e-01 -3.20811510e-01
6.78270936e-01 1.27758276e+00 -3.06627512e-01 -4.61013347e-01
-3.16703349e-01 1.00627613e+00 5.43750584e-01 -2.38545407e-02
-5.49384654e-01 -5.82233667e-01 7.17476070e-01 2.06763923e-01
6.67964995e-01 1.64302766e-01 -1.58691779e-01 -1.10284388e+00
9.55575332e-02 -1.07213080e+00 6.78124368e-01 -3.52832288e-01
-1.57884753e+00 2.44915709e-01 -2.55261753e-02 -7.22044170e-01
-2.43630901e-01 -4.65000600e-01 -3.58004689e-01 6.86776280e-01
-1.49287736e+00 -1.19344425e+00 -2.77346969e-01 7.53594220e-01
1.63621411e-01 -1.37127846e-01 8.47667634e-01 3.03949386e-01
-4.63067174e-01 1.08480227e+00 1.13535412e-01 2.09686637e-04
7.46619225e-01 -1.03919876e+00 5.26081741e-01 8.60620081e-01
1.88440070e-01 5.80086529e-01 3.62610400e-01 -7.28200972e-01
-1.57485974e+00 -1.45181596e+00 7.69852817e-01 -6.78453326e-01
2.75807828e-01 -8.27915668e-01 -9.00201082e-01 8.47752869e-01
-2.01996624e-01 4.95792478e-01 5.62810123e-01 -1.48136348e-01
-5.99057078e-01 -4.61606145e-01 -1.52921164e+00 4.61932480e-01
1.16721821e+00 -3.88587475e-01 -3.78291577e-01 2.15732470e-01
8.09875131e-01 -3.30188513e-01 -5.47916353e-01 1.04083136e-01
5.73759496e-01 -7.66477883e-01 9.66006160e-01 -7.45175123e-01
3.38422388e-01 -2.66664326e-01 -1.04189992e-01 -1.05996966e+00
-2.21799701e-01 -8.77995968e-01 -3.51513743e-01 1.36756837e+00
4.38525051e-01 -6.51422620e-01 1.07645667e+00 1.11882138e+00
1.34823620e-01 -4.73518819e-01 -9.68597293e-01 -7.36589193e-01
-1.53327929e-02 -1.95972711e-01 8.10696602e-01 8.94977570e-01
-5.20427465e-01 1.27700940e-01 -4.30632353e-01 1.05866566e-01
6.56768322e-01 -8.46883878e-02 1.13027191e+00 -1.14491677e+00
-4.07798231e-01 -1.45659233e-02 -3.94119680e-01 -6.35888517e-01
-1.43688127e-01 -9.68889296e-01 9.29278433e-02 -9.84836638e-01
4.69937891e-01 -6.77598417e-01 -4.46232438e-01 1.03756261e+00
-1.89785928e-01 1.89259917e-01 2.02075347e-01 -1.16404910e-02
-5.89262009e-01 2.98283398e-01 9.70562637e-01 -1.97622404e-01
-1.81480154e-01 -1.04886852e-01 -1.16028523e+00 4.11180377e-01
5.43465912e-01 -7.47321725e-01 -5.97829521e-01 -3.46581310e-01
1.59152046e-01 -5.84825873e-01 7.57648289e-01 -1.15856481e+00
4.74950343e-01 -2.18070269e-01 6.87580943e-01 -6.00065649e-01
4.07218635e-01 -7.83152878e-01 2.65694112e-01 3.29382300e-01
-5.38599372e-01 -4.23069388e-01 2.38052502e-01 7.74706483e-01
2.54723132e-01 -3.41451347e-01 7.90128529e-01 -4.44803059e-01
-8.79602194e-01 5.12911916e-01 8.84373412e-02 2.59434938e-01
1.31621051e+00 -1.89673647e-01 -6.38374388e-01 -6.24781586e-02
-5.38296998e-01 1.38262644e-01 5.64856172e-01 4.57312316e-01
5.29294312e-01 -1.08229268e+00 -5.86322844e-01 5.78288496e-01
2.96096116e-01 3.47168386e-01 1.26367062e-01 3.90742958e-01
-2.05121934e-01 2.19503090e-01 -2.64686614e-01 -1.89723894e-01
-1.00388348e+00 8.68246555e-01 1.91246256e-01 -1.46392435e-01
-7.28700101e-01 1.08374214e+00 3.08911055e-01 -4.75586474e-01
4.33899105e-01 2.61819689e-03 3.86993736e-01 1.01377033e-01
9.10338283e-01 2.69177914e-01 1.60330117e-01 3.08759947e-04
-3.31106961e-01 -2.58132160e-01 -7.92758882e-01 7.04141287e-03
1.70137680e+00 1.63607389e-01 -8.11186656e-02 2.82875448e-01
1.11702418e+00 -1.21120073e-01 -1.53488326e+00 -3.96312535e-01
-4.62896749e-02 -6.78791046e-01 3.55259143e-02 -1.12267649e+00
-9.58819389e-01 7.47025549e-01 7.26784468e-01 -2.74322778e-01
9.65300858e-01 -8.73113722e-02 9.40647244e-01 7.37861037e-01
3.39699000e-01 -1.28848016e+00 -4.33814079e-02 2.34981731e-01
6.44989312e-01 -1.00049353e+00 1.44376308e-01 -1.51982144e-01
-6.16060853e-01 8.89977276e-01 8.59290421e-01 3.43176872e-02
4.57527637e-01 3.43245238e-01 -3.80510926e-01 -1.72928303e-01
-9.65414703e-01 2.93281209e-02 5.60543314e-02 7.80553937e-01
-3.31928849e-01 -1.06638066e-01 1.77832544e-01 8.82979751e-01
-2.09769174e-01 4.71425742e-01 4.39927578e-01 1.27664518e+00
-4.24864680e-01 -1.00053728e+00 -1.87668666e-01 9.76651371e-01
-3.73458356e-01 -1.85618311e-01 -4.83317137e-01 5.67379594e-01
3.96351665e-01 5.02896070e-01 1.63706675e-01 -3.22818696e-01
1.81461319e-01 1.51132897e-01 3.44618201e-01 -7.81713068e-01
-5.35414279e-01 -6.76035762e-01 6.69954121e-02 -6.27399862e-01
4.61146869e-02 -6.17857337e-01 -9.88918424e-01 -4.16810542e-01
-2.84945160e-01 -9.12144035e-02 4.56716597e-01 6.64939642e-01
9.85585809e-01 1.14430048e-01 4.07494068e-01 -6.11097753e-01
-7.85181344e-01 -4.99489486e-01 -4.93159384e-01 4.60653663e-01
5.62010646e-01 -5.48604429e-01 -3.50619376e-01 1.19326793e-01] | [9.79869270324707, 3.2990260124206543] |
21195ab8-1b26-4329-a894-bfa4d010de51 | localize-me-anywhere-anytime-a-multi-task | null | null | http://openaccess.thecvf.com/content_iccv_2015/html/Lu_Localize_Me_Anywhere_ICCV_2015_paper.html | http://openaccess.thecvf.com/content_iccv_2015/papers/Lu_Localize_Me_Anywhere_ICCV_2015_paper.pdf | Localize Me Anywhere, Anytime: A Multi-Task Point-Retrieval Approach | Image-based localization is an essential complement to GPS localization. Current image-based localization methods are based on either 2D-to-3D or 3D-to-2D to find the correspondences, which ignore the real scene geometric attributes. The main contribution of our paper is that we use a 3D model reconstructed by a short video as the query to realize 3D-to-3D localization under a multi-task point retrieval framework. Firstly, the use of a 3D model as the query enables us to efficiently select location candidates. Furthermore, the reconstruction of 3D model exploits the correlations among different images, based on the fact that images captured from different views for SfM share information through matching features. By exploring shared information (matching features) across multiple related tasks (images of the same scene captured from different views), the visual feature's view-invariance property can be improved in order to get to a higher point retrieval accuracy. More specifically, we use multi-task point retrieval framework to explore the relationship between descriptors and the 3D points, which extracts the discriminant points for more accurate 3D-to-3D correspondences retrieval. We further apply multi-task learning (MTL) retrieval approach on thermal images to prove that our MTL retrieval framework also provides superior performance for the thermal domain. This application is exceptionally helpful to cope with the localization problem in an environment with limited light sources. | ['Nicu Sebe', 'Guoyu Lu', 'Chandra Kambhamettu', 'Jingkuan Song', 'Yan Yan', 'Li Ren'] | 2015-12-01 | null | null | null | iccv-2015-12 | ['image-based-localization'] | ['computer-vision'] | [-8.11733156e-02 -8.97391200e-01 -3.55196118e-01 -1.92998916e-01
-1.16374683e+00 -7.04700291e-01 6.09119177e-01 -6.97710812e-02
-3.07166696e-01 1.45375699e-01 -1.42398864e-01 2.05962121e-01
-5.25509953e-01 -5.77137530e-01 -4.86513555e-01 -8.53363454e-01
6.49720132e-02 3.14867198e-01 3.38414967e-01 -3.23738128e-01
7.09564924e-01 7.71592975e-01 -1.58143675e+00 -2.88464632e-02
5.91650248e-01 1.12672460e+00 5.23464918e-01 2.15073824e-01
-2.04762146e-01 -1.09205525e-02 -4.74451810e-01 1.03923649e-01
5.99500895e-01 6.13498017e-02 -5.05013049e-01 -3.05123962e-02
5.30985713e-01 -6.90930411e-02 -1.74684785e-02 1.01777637e+00
5.45296967e-01 2.50305653e-01 7.44845390e-01 -1.33125198e+00
-4.99844223e-01 -5.59921861e-01 -8.64267588e-01 -1.27580911e-01
9.16201293e-01 -3.09750348e-01 8.87059510e-01 -1.10581613e+00
4.16063249e-01 1.10304689e+00 5.50812006e-01 -1.44961327e-01
-9.23959196e-01 -4.28503841e-01 -1.82004347e-01 3.55816007e-01
-1.92845452e+00 -2.73850232e-01 1.07761967e+00 -2.71319598e-01
5.57906032e-01 1.43648759e-01 5.46230674e-01 8.99714589e-01
4.82804090e-01 4.21283662e-01 1.27097940e+00 -5.38635552e-01
-5.85880280e-02 1.67807430e-01 -1.65963233e-01 7.56967425e-01
1.03306174e-01 4.50376794e-02 -7.77674675e-01 -2.75904804e-01
9.36965644e-01 4.66250837e-01 -3.01654190e-01 -7.14227676e-01
-1.37742519e+00 6.62611663e-01 5.05964756e-01 5.72442472e-01
-1.93326056e-01 4.37102839e-02 1.11509055e-01 3.25517595e-01
3.23432714e-01 3.23156834e-01 -2.09167331e-01 1.29203290e-01
-7.96964765e-01 -1.24434137e-03 4.91105437e-01 1.13490403e+00
1.37833452e+00 -1.80356205e-01 5.11133559e-02 7.79580235e-01
5.56205869e-01 1.07871032e+00 6.27838254e-01 -7.90857553e-01
4.32388812e-01 5.02826154e-01 2.47740954e-01 -1.44322264e+00
-2.82253623e-01 -3.28280240e-01 -7.25202441e-01 -3.47632691e-02
1.63216114e-01 4.47698325e-01 -4.57276016e-01 1.33711684e+00
3.76620889e-01 1.02952868e-01 2.18366887e-02 1.15285766e+00
5.93356729e-01 5.47054052e-01 -4.52795357e-01 6.10851496e-03
1.42127848e+00 -7.05769420e-01 -3.49987060e-01 -1.51953995e-01
7.57274926e-01 -1.12995791e+00 8.20069075e-01 1.04125470e-01
-4.53628182e-01 -9.32706177e-01 -9.16375041e-01 5.97014427e-02
-6.08405650e-01 4.19718474e-01 2.55416721e-01 4.71409053e-01
-9.48817134e-01 1.77065104e-01 -2.90070742e-01 -5.47294378e-01
-5.29624149e-02 3.75494540e-01 -8.14693809e-01 -2.78542310e-01
-1.03228664e+00 1.10765374e+00 4.06891197e-01 1.36584982e-01
-6.41859055e-01 -1.86421871e-01 -8.82579684e-01 -2.40749598e-01
4.61161971e-01 -5.78812301e-01 5.73716164e-01 -6.48920357e-01
-1.08684349e+00 9.26784813e-01 -5.64065993e-01 1.15242667e-01
2.13004574e-01 -1.37075439e-01 -3.61824155e-01 4.41366374e-01
5.85605323e-01 3.14414114e-01 1.06439185e+00 -1.32193160e+00
-5.71048081e-01 -6.04354978e-01 -1.55520901e-01 4.86508518e-01
-8.67435783e-02 -2.14200184e-01 -5.51218569e-01 -3.34894538e-01
7.37619162e-01 -9.72791970e-01 -1.64368320e-02 -5.53361885e-03
-6.19668104e-02 -1.84071124e-01 9.53271031e-01 -2.85721183e-01
5.06948769e-01 -2.14311385e+00 1.24941468e-01 3.98976326e-01
-1.32660642e-01 -1.25802785e-01 -2.96317697e-01 6.15378916e-01
1.57493323e-01 -1.19236723e-01 2.87036747e-01 -3.97502810e-01
-3.25554579e-01 1.64194196e-01 -2.94545032e-02 7.25880206e-01
-5.13440222e-02 7.36561477e-01 -8.37857544e-01 -7.03009903e-01
6.07958496e-01 2.74629086e-01 4.33284901e-02 3.08493152e-03
3.79000723e-01 6.79980993e-01 -8.85854065e-01 9.40008879e-01
8.56248677e-01 -6.16563223e-02 -4.26233649e-01 -6.84759557e-01
-2.87945539e-01 -3.89922917e-01 -1.24757910e+00 2.10970974e+00
-6.54371381e-01 3.01678419e-01 -2.29394272e-01 -1.02525198e+00
1.37704802e+00 1.93397567e-01 8.36822271e-01 -9.15273964e-01
-1.10530801e-01 5.38757682e-01 -5.09692430e-01 -5.34967959e-01
6.33707166e-01 1.08072214e-01 -3.49830210e-01 2.19130859e-01
1.02736972e-01 -3.28237802e-01 -4.12307322e-01 -2.65413076e-01
5.57353258e-01 3.33814740e-01 4.24541593e-01 -2.51555711e-01
8.72608423e-01 -4.02329899e-02 3.31353813e-01 8.43133271e-01
-3.43939513e-02 5.81910014e-01 -1.71447322e-01 -4.91273761e-01
-1.02438295e+00 -9.32356715e-01 -2.36162290e-01 6.10317647e-01
7.29913235e-01 -2.31909484e-01 2.84218919e-02 -5.26127815e-01
5.07172979e-02 1.28603622e-01 -2.60249466e-01 -1.23530202e-01
-3.61155927e-01 -4.23241407e-01 3.39944422e-01 1.02648482e-01
8.31153691e-01 -2.85275370e-01 -5.94807863e-01 -4.60254066e-02
-4.30718154e-01 -1.03656530e+00 -4.92985725e-01 7.32965469e-02
-7.51417100e-01 -1.12070334e+00 -9.29869115e-01 -7.58889318e-01
6.32482827e-01 1.21992540e+00 4.50001448e-01 -5.32474332e-02
-5.52858673e-02 9.24594998e-01 -5.84693789e-01 -4.63021696e-02
1.65687770e-01 1.07395286e-02 3.45965862e-01 3.58869702e-01
4.20849800e-01 -3.84224683e-01 -4.68600810e-01 7.59581804e-01
-6.98140264e-01 -2.60610700e-01 7.16846764e-01 8.68836105e-01
8.93759608e-01 9.10436809e-02 2.05982924e-01 -6.14761300e-02
2.05952436e-01 -3.45400304e-01 -7.11690664e-01 5.13792336e-01
-3.34556758e-01 1.13312483e-01 4.17928100e-01 -2.91524053e-01
-7.78138041e-01 3.48069280e-01 1.80697173e-01 -8.17517400e-01
-1.66859567e-01 4.09251064e-01 -3.12830359e-02 -9.76034045e-01
4.78138596e-01 7.03440607e-01 8.01283792e-02 -4.19136077e-01
1.75884873e-01 6.78411782e-01 1.94295779e-01 -5.91402292e-01
1.14294183e+00 6.66728020e-01 7.16185033e-01 -9.29514885e-01
-8.09170187e-01 -1.12910545e+00 -1.29514062e+00 -3.21263552e-01
7.98733413e-01 -1.12925065e+00 -7.54304469e-01 2.54926503e-01
-1.24097490e+00 3.56936157e-01 1.89156085e-01 7.88861156e-01
-5.71054518e-01 8.04659843e-01 -7.54793501e-03 -8.38378370e-01
-2.27258444e-01 -1.33972311e+00 1.73465967e+00 8.69446322e-02
3.65972966e-01 -9.97899890e-01 4.54446524e-02 3.23071063e-01
2.20031306e-01 -1.58288796e-02 6.52814627e-01 -5.25993109e-01
-9.61108923e-01 -4.17620331e-01 -2.91195452e-01 4.19127829e-02
2.95681477e-01 -3.01020414e-01 -8.27671349e-01 -4.56753135e-01
3.61413449e-01 -6.53628111e-02 6.29503787e-01 2.93707430e-01
7.52826333e-01 2.52598524e-01 -5.61644256e-01 7.24227786e-01
1.80327249e+00 2.80380905e-01 4.30943847e-01 5.63998699e-01
7.92337179e-01 5.89058876e-01 1.15690863e+00 2.38121510e-01
4.27080691e-01 1.14701796e+00 4.51857418e-01 -8.45010132e-02
2.56941944e-01 -2.51473486e-01 2.08010241e-01 7.83970118e-01
-1.14948474e-01 2.86573078e-03 -8.19515109e-01 4.27890092e-01
-1.86277092e+00 -7.57874727e-01 -2.13673383e-01 2.52239108e+00
-4.14008908e-02 -2.82703578e-01 -1.10913485e-01 -9.93613675e-02
7.52892613e-01 3.17167252e-01 -1.38822824e-01 1.52967721e-01
-1.64927378e-01 -4.30521816e-02 7.18515813e-01 2.32373208e-01
-1.11575842e+00 7.84478307e-01 5.56367350e+00 1.20309436e+00
-1.13391769e+00 1.09865382e-01 -3.51999537e-03 3.64957392e-01
-1.79647148e-01 3.76821250e-01 -9.96212423e-01 3.25487763e-01
3.23855758e-01 -1.21836141e-02 1.69300303e-01 8.09979737e-01
8.39653164e-02 -4.34432298e-01 -8.55468214e-01 1.52744150e+00
4.90465641e-01 -1.00264597e+00 2.57250160e-01 2.44831026e-01
4.85349238e-01 -4.53012697e-02 1.49994448e-01 -1.40401497e-01
-5.55148780e-01 -5.84605992e-01 6.37875795e-01 6.27857506e-01
7.97950983e-01 -5.42666733e-01 8.97082329e-01 5.25705695e-01
-1.59177935e+00 1.97317731e-02 -6.81534708e-01 2.17256159e-01
1.58527285e-01 5.02550423e-01 -8.06083441e-01 1.34348977e+00
7.41757393e-01 9.33101356e-01 -7.68825710e-01 1.19675910e+00
9.81609672e-02 -4.04067665e-01 -3.39923322e-01 -7.33643547e-02
2.39483833e-01 -5.88974059e-01 5.71312547e-01 7.23571539e-01
9.62764323e-01 -3.09948862e-01 3.79076719e-01 5.52854359e-01
4.11495328e-01 2.22240925e-01 -1.30696619e+00 5.73433638e-01
5.54288208e-01 1.21685648e+00 -6.41472876e-01 -3.22884433e-02
-4.44871694e-01 1.14900661e+00 -6.39469326e-02 3.30991149e-01
-5.72395325e-01 -3.88709277e-01 3.53208482e-01 -1.45351902e-01
1.84707120e-01 -8.09625864e-01 2.25311756e-01 -1.16959047e+00
5.98183796e-02 -4.19163287e-01 3.28055955e-02 -1.13563490e+00
-1.24507737e+00 3.36472124e-01 2.53276944e-01 -1.81847978e+00
-1.62425116e-01 -4.41111982e-01 -3.55249792e-01 1.05246127e+00
-1.65295458e+00 -1.45699883e+00 -4.78812337e-01 9.44200218e-01
3.47251207e-01 -2.45655805e-01 7.77830184e-01 3.00927043e-01
-1.68349836e-02 2.59705812e-01 3.26146603e-01 -3.47964950e-02
1.01572001e+00 -8.83985519e-01 -2.25345418e-02 6.09780669e-01
4.88960952e-01 4.92346764e-01 2.02981140e-02 -4.78694111e-01
-1.80964148e+00 -7.94154704e-01 6.32959604e-01 -6.54945433e-01
4.03377861e-01 -2.88711965e-01 -5.72539747e-01 4.03332025e-01
-2.40378782e-01 -1.17690176e-01 5.03656268e-01 5.27708083e-02
-2.73678362e-01 -5.53117454e-01 -9.88046646e-01 2.97245860e-01
8.89568329e-01 -9.44021881e-01 -4.62787360e-01 4.79177147e-01
5.15559733e-01 -3.49609733e-01 -1.04921925e+00 3.43332618e-01
5.46131909e-01 -8.39030921e-01 1.30334914e+00 2.29768008e-01
-2.46094868e-01 -6.65473223e-01 -6.67165041e-01 -1.18677211e+00
-2.71868020e-01 -2.14365587e-01 4.29591864e-01 1.10330689e+00
-1.72039032e-01 -8.86187196e-01 4.80011135e-01 -1.55073613e-01
-6.93065003e-02 -3.18281442e-01 -1.32538950e+00 -1.02667046e+00
-4.45980817e-01 -3.67835522e-01 5.54283500e-01 7.89996028e-01
-5.37567735e-01 2.66493589e-01 -5.34357488e-01 5.03571272e-01
7.67287314e-01 6.40438020e-01 9.21319723e-01 -1.38138843e+00
2.19439790e-02 -5.64506203e-02 -7.46501565e-01 -1.37808609e+00
2.34297171e-01 -8.85173738e-01 -9.46263596e-02 -1.19595146e+00
5.62180802e-02 -7.35412657e-01 -2.15143308e-01 2.61864692e-01
2.60196328e-01 3.73388886e-01 2.65921116e-01 8.77694547e-01
-6.66203737e-01 6.06590629e-01 1.38259673e+00 -6.99552372e-02
-5.91759011e-02 1.69763327e-01 -2.83587188e-01 3.60751569e-01
4.38380241e-01 -4.54048574e-01 -2.97487557e-01 -5.81772923e-01
1.10183828e-01 3.10547948e-01 7.19643176e-01 -1.13860989e+00
5.05841017e-01 -1.88286826e-01 5.37307024e-01 -9.72845078e-01
8.64947200e-01 -1.19147515e+00 2.42269143e-01 2.47826517e-01
2.12623596e-01 1.15298271e-01 -1.08052239e-01 7.58910596e-01
-5.22882760e-01 -5.61529875e-01 3.83284569e-01 -3.27219278e-01
-1.13800597e+00 4.83381152e-01 -5.14057651e-02 -5.35907686e-01
1.06795585e+00 -6.62270248e-01 -1.89795956e-01 -3.97281468e-01
-2.92013049e-01 1.44874498e-01 6.71841502e-01 4.37152773e-01
8.79962921e-01 -1.72941542e+00 -2.97524333e-01 4.10196900e-01
5.60090125e-01 -6.69919252e-02 3.29255939e-01 1.04094529e+00
-4.13776368e-01 6.65310562e-01 -3.17434907e-01 -1.23810732e+00
-1.26921129e+00 5.62672079e-01 3.33694965e-01 1.91778481e-01
-2.78509736e-01 3.28860551e-01 9.00798142e-02 -4.63625968e-01
-1.33605659e-01 2.21357718e-02 -1.07499838e-01 2.24083424e-01
1.20961070e-01 2.92230576e-01 1.08645067e-01 -1.13235092e+00
-5.94745100e-01 1.74480271e+00 1.95323020e-01 -1.00894190e-01
1.02219677e+00 -4.38964009e-01 -1.76296726e-01 5.05050242e-01
1.62559807e+00 8.73950198e-02 -7.08177030e-01 -4.65765268e-01
5.54169528e-03 -1.05482006e+00 6.89065307e-02 -1.69084072e-01
-7.27572739e-01 9.32385325e-01 7.80132353e-01 3.13115641e-02
1.15471220e+00 -7.16557773e-03 4.59391743e-01 5.33278048e-01
1.15712523e+00 -8.58928084e-01 4.25195485e-01 3.96057367e-01
7.08027661e-01 -1.55231011e+00 1.77262232e-01 -2.83883750e-01
-4.77466911e-01 1.31702721e+00 2.77092397e-01 6.50013983e-02
5.74161887e-01 -5.85982442e-01 -1.06484637e-01 -2.68592596e-01
-3.22591588e-02 -4.91803646e-01 5.26483119e-01 8.92319560e-01
-1.07265487e-01 -1.90798625e-01 7.15784132e-02 -1.71071053e-01
2.85666943e-01 -3.45032841e-01 1.61639184e-01 7.63021946e-01
-4.80411887e-01 -1.30674005e+00 -7.42181182e-01 9.90795568e-02
7.70231336e-02 1.17011458e-01 -5.24649583e-02 8.15245688e-01
2.27834061e-01 9.76137698e-01 -1.87671795e-01 -5.53682864e-01
2.92298526e-01 -1.27802743e-02 7.24967241e-01 -4.96407092e-01
-2.25345477e-01 2.86375403e-01 -4.43155885e-01 -6.11807585e-01
-8.07045579e-01 -5.86674452e-01 -6.97427154e-01 -8.55273306e-02
-5.99379301e-01 1.77170455e-01 1.05018187e+00 7.76600599e-01
4.30852473e-01 -8.76886994e-02 1.10847092e+00 -1.16847479e+00
-4.59625185e-01 -5.22608519e-01 -7.98825145e-01 2.50657767e-01
4.23967987e-01 -9.90834355e-01 -4.83171105e-01 -4.06545490e-01] | [7.6164937019348145, -2.195307493209839] |
28e9a1d9-db37-45d4-9afb-38a4ec4acb6b | contrastive-learning-of-global-and-local-1 | 2104.05418 | null | https://arxiv.org/abs/2104.05418v2 | https://arxiv.org/pdf/2104.05418v2.pdf | Contrastive Learning of Global-Local Video Representations | Contrastive learning has delivered impressive results for various tasks in the self-supervised regime. However, existing approaches optimize for learning representations specific to downstream scenarios, i.e., \textit{global} representations suitable for tasks such as classification or \textit{local} representations for tasks such as detection and localization. While they produce satisfactory results in the intended downstream scenarios, they often fail to generalize to tasks that they were not originally designed for. In this work, we propose to learn video representations that generalize to both the tasks which require global semantic information (e.g., classification) and the tasks that require local fine-grained spatio-temporal information (e.g., localization). We achieve this by optimizing two contrastive objectives that together encourage our model to learn global-local visual information given audio signals. We show that the two objectives mutually improve the generalizability of the learned global-local representations, significantly outperforming their disjointly learned counterparts. We demonstrate our approach on various tasks including action/sound classification, lip reading, deepfake detection, event and sound localization (https://github.com/yunyikristy/global\_local). | ['Yale Song', 'Daniel McDuff', 'Zhaoyang Zeng', 'Shuang Ma'] | 2021-04-07 | null | null | null | null | ['sound-classification'] | ['audio'] | [ 3.49219441e-01 -2.94720352e-01 -3.99791121e-01 -2.74652094e-01
-1.22450972e+00 -5.82133412e-01 4.76332426e-01 8.19724724e-02
-2.23828778e-01 5.33185840e-01 4.26316559e-01 -1.27554968e-01
-9.84573215e-02 -4.04906064e-01 -8.67126226e-01 -6.61482990e-01
-1.24111928e-01 -1.44843319e-02 8.46829936e-02 6.89807534e-02
1.33683875e-01 2.84465373e-01 -1.91422844e+00 5.10831535e-01
4.43649083e-01 1.28210628e+00 2.85083205e-01 9.18310463e-01
1.66207448e-01 7.63483763e-01 -4.94379073e-01 2.87546098e-01
4.17279527e-02 -4.16063637e-01 -8.67179811e-01 1.30177243e-02
7.36881495e-01 -2.06820369e-01 -5.02955377e-01 8.94110203e-01
6.48745179e-01 3.92737687e-01 7.23581016e-01 -1.56604362e+00
-8.19198489e-01 3.41635436e-01 -5.31579971e-01 5.16496062e-01
3.64187002e-01 2.71569461e-01 1.30771041e+00 -8.54844987e-01
8.52222517e-02 1.20429134e+00 6.30668998e-01 6.56724334e-01
-1.08704555e+00 -6.39998257e-01 4.47565824e-01 4.05001432e-01
-1.25301957e+00 -9.13692415e-01 6.45117402e-01 -4.75734860e-01
7.99550653e-01 2.76360154e-01 1.25357658e-01 1.40481603e+00
-2.15099886e-01 1.12191331e+00 9.44658756e-01 -3.26397479e-01
1.25848293e-01 -1.62489474e-01 1.02830999e-01 4.16174561e-01
-2.08738878e-01 6.75233593e-03 -9.69271302e-01 9.17165503e-02
7.29579508e-01 -2.26482237e-03 -4.82285112e-01 -1.74134269e-01
-1.21584034e+00 7.44757593e-01 3.27827424e-01 4.68632400e-01
-2.66901672e-01 5.76210320e-01 5.79205811e-01 3.07693154e-01
6.74336195e-01 1.51681989e-01 -6.88550830e-01 -2.07060948e-01
-9.91571724e-01 -1.11484185e-01 2.71204025e-01 8.67048621e-01
6.39671743e-01 2.83748388e-01 -3.69731188e-01 1.04733825e+00
2.10331231e-01 3.50697994e-01 7.48755157e-01 -8.76754940e-01
4.61075127e-01 6.80160895e-02 -5.59671735e-03 -7.95908034e-01
-5.27012289e-01 -3.97999883e-01 -6.85733795e-01 6.31105304e-02
3.84283096e-01 -1.55030251e-01 -1.01515400e+00 2.19163084e+00
9.52676088e-02 4.96833771e-01 5.21167368e-02 8.97870362e-01
1.18400741e+00 7.80886650e-01 2.23715246e-01 -1.75044507e-01
1.20578039e+00 -1.11632228e+00 -6.41777515e-01 -2.88018376e-01
4.51946914e-01 -6.21928811e-01 1.25546598e+00 2.37963945e-01
-9.93864417e-01 -7.62166500e-01 -7.52372921e-01 -1.06894359e-01
-2.20509291e-01 3.17423970e-01 4.99738097e-01 2.08561152e-01
-1.35623109e+00 3.21797967e-01 -7.19235182e-01 -5.85118949e-01
3.75146836e-01 2.14532539e-01 -3.34143937e-01 2.24231079e-01
-1.06150234e+00 4.77436125e-01 1.43981382e-01 -5.49736619e-02
-1.15417457e+00 -4.71629053e-01 -9.10598814e-01 1.31638378e-01
4.25971955e-01 -4.74180609e-01 1.51336265e+00 -1.16492724e+00
-1.41171110e+00 8.25109720e-01 -2.56906569e-01 -3.93032640e-01
2.49981150e-01 -3.19104254e-01 -2.62400985e-01 3.21529865e-01
2.52040893e-01 7.64286101e-01 1.17709363e+00 -9.92659569e-01
-6.77931070e-01 -2.51053512e-01 2.26922303e-01 3.77049774e-01
-5.85162342e-01 7.70013183e-02 -3.69181126e-01 -9.47038233e-01
-5.68861589e-02 -6.68554485e-01 2.33152911e-01 2.06326321e-01
-5.10472059e-01 -5.48444986e-01 9.85472500e-01 -7.48256743e-01
9.71961439e-01 -2.35860586e+00 1.11549281e-01 -1.71971709e-01
2.90042814e-02 2.07965910e-01 -3.71248215e-01 3.71672094e-01
-2.31040224e-01 4.54224348e-02 -3.83785628e-02 -4.92940426e-01
3.67928930e-02 1.42717669e-02 -2.56442100e-01 5.41605771e-01
2.05052525e-01 9.56825614e-01 -1.02380419e+00 -3.12399179e-01
1.28412664e-01 4.79689211e-01 -5.09294391e-01 2.74902463e-01
-2.45065868e-01 6.11645281e-01 -4.30328816e-01 6.79255664e-01
2.12922797e-01 -3.71654212e-01 -1.59245953e-01 -1.57071203e-01
6.30384758e-02 3.65447670e-01 -1.03362083e+00 1.84340143e+00
-8.17253768e-01 9.77239788e-01 4.28294659e-01 -1.48334122e+00
6.58060730e-01 5.25380909e-01 6.32681310e-01 -5.87056160e-01
-2.27267027e-01 -2.61607952e-02 -3.48654896e-01 -6.15591228e-01
2.31244951e-01 -2.01715082e-01 -1.03146568e-01 4.86561209e-01
3.30050975e-01 2.04328507e-01 -2.66592860e-01 8.35506544e-02
1.10158992e+00 2.84197867e-01 3.45192909e-01 -1.54763252e-01
3.49296302e-01 -3.64562571e-01 3.72818381e-01 7.37079084e-01
-3.00023854e-01 8.33965957e-01 3.50326151e-01 -9.86217801e-03
-4.40307319e-01 -1.08137631e+00 -2.93013975e-02 1.82285357e+00
1.60712585e-01 -5.10939598e-01 -5.74733078e-01 -8.25465739e-01
-3.38571593e-02 3.68761867e-01 -4.84344065e-01 -3.12392116e-01
-3.11812043e-01 -3.85925919e-01 7.01888919e-01 8.86022866e-01
3.73262435e-01 -1.17963231e+00 -6.16342485e-01 1.61340117e-01
-3.08055371e-01 -1.06576073e+00 -6.39897943e-01 3.02859873e-01
-5.52358687e-01 -9.63654041e-01 -8.46034467e-01 -8.61439586e-01
2.65876621e-01 3.80727977e-01 9.64543879e-01 -7.39880465e-03
-4.26595628e-01 1.02942300e+00 -3.97720844e-01 -2.85575718e-01
8.08910206e-02 -1.25709236e-01 1.29832342e-01 3.77696186e-01
1.18936189e-01 -6.28341734e-01 -7.65125096e-01 3.33672971e-01
-8.92277837e-01 -1.21709973e-01 2.67243862e-01 9.78487074e-01
5.02257764e-01 -2.09884450e-01 9.02361274e-01 -2.76596129e-01
5.29972374e-01 -5.94278514e-01 -2.64651626e-02 2.22669765e-01
-5.18717468e-02 -1.86736941e-01 4.95296627e-01 -7.49717593e-01
-6.94508195e-01 -4.41412330e-02 -2.80724972e-01 -6.55683637e-01
-5.02797782e-01 3.18545401e-01 -1.20704852e-01 1.81023270e-01
5.91770470e-01 3.46930236e-01 -6.69684708e-02 -5.65001428e-01
3.41653198e-01 8.59094262e-01 5.74164271e-01 -4.76604402e-01
4.98681813e-01 2.98984498e-01 -3.14597875e-01 -9.90409255e-01
-9.65669334e-01 -7.26691246e-01 -3.90108824e-01 -1.51025519e-01
1.01352775e+00 -1.11678123e+00 -6.90940499e-01 5.15991211e-01
-9.42552686e-01 -6.03168905e-01 -4.13524270e-01 5.01975238e-01
-9.74610984e-01 2.36905336e-01 -3.39160860e-01 -9.20559645e-01
-5.67669086e-02 -1.05043495e+00 1.54570532e+00 5.97072691e-02
-2.46775001e-01 -1.11069810e+00 -1.22540079e-01 2.20813453e-01
4.55004513e-01 1.01973861e-01 6.77050650e-01 -6.57682657e-01
-3.42393607e-01 1.35448873e-01 -1.77910417e-01 5.95883906e-01
3.64158839e-01 -2.33555347e-01 -1.46777785e+00 -2.84379721e-01
-3.17442238e-01 -7.01750457e-01 1.31750882e+00 5.55334091e-01
1.70888448e+00 -4.93124932e-01 -2.94407874e-01 6.57956660e-01
9.45668161e-01 2.63546892e-02 3.14431459e-01 2.45469362e-02
6.39581442e-01 5.72153509e-01 6.41968548e-01 4.56770867e-01
3.77148181e-01 1.13711882e+00 5.53534627e-01 -2.77771443e-01
-5.47154188e-01 -3.63783807e-01 5.40052593e-01 4.37379181e-01
6.36654943e-02 -4.14396584e-01 -8.28620493e-01 7.84967959e-01
-2.15378213e+00 -9.68864202e-01 3.18554908e-01 1.93632603e+00
7.43642926e-01 -3.25110853e-01 2.78122514e-01 1.83200240e-01
8.46100211e-01 3.68834198e-01 -6.46631300e-01 -2.93907762e-01
2.33765040e-02 3.23817134e-01 1.65520683e-02 3.68670642e-01
-1.49967849e+00 1.04222929e+00 5.62478495e+00 8.63813519e-01
-1.36342978e+00 5.12018025e-01 6.00739121e-01 -2.73161709e-01
-1.12513281e-01 -3.38346988e-01 -5.31540990e-01 4.23371613e-01
8.85555983e-01 1.34332687e-01 3.51511687e-01 6.23928666e-01
3.08305711e-01 5.83420433e-02 -1.27053046e+00 1.30195534e+00
1.28627837e-01 -1.16327524e+00 -9.46406871e-02 -2.04305962e-01
5.92610896e-01 -3.79786417e-02 4.47451472e-01 3.90437037e-01
-7.06230029e-02 -1.19782615e+00 8.86172116e-01 5.00780284e-01
1.11205530e+00 -4.17143285e-01 1.80226162e-01 3.82364959e-01
-1.36977494e+00 -2.49197692e-01 5.87332575e-03 4.85719973e-03
-4.18259390e-02 2.59285092e-01 -5.86835086e-01 3.21034342e-01
9.79852557e-01 9.21860576e-01 -3.93291175e-01 9.57472444e-01
-2.58297294e-01 7.64949679e-01 -2.60473013e-01 1.75007999e-01
2.72677869e-01 2.79914916e-01 6.86522841e-01 1.23473668e+00
3.64762157e-01 -1.52618274e-01 2.20632941e-01 6.07753813e-01
-2.08581477e-01 1.89277336e-01 -7.64046371e-01 -5.26156873e-02
3.72259229e-01 1.03733385e+00 -3.63950014e-01 -1.50955290e-01
-4.55130696e-01 9.92397487e-01 3.31421703e-01 6.03686869e-01
-8.74355257e-01 -3.11692029e-01 1.02344429e+00 5.03768809e-02
4.48771954e-01 -1.97956473e-01 -1.25581309e-01 -1.13229537e+00
2.27473930e-01 -8.06323588e-01 6.40091836e-01 -8.39349687e-01
-1.25496066e+00 4.10525799e-01 5.90833612e-02 -1.29725015e+00
-4.68004376e-01 -6.48189664e-01 -8.58894527e-01 5.84072649e-01
-1.55952775e+00 -1.33840263e+00 -2.33630031e-01 1.16168714e+00
7.72481501e-01 -1.91717058e-01 7.49090016e-01 1.73711970e-01
-4.87956673e-01 8.84541988e-01 2.00645681e-02 9.89679992e-02
9.82390046e-01 -1.24481475e+00 -8.22497010e-02 6.64417803e-01
5.88301539e-01 3.50370497e-01 4.11357075e-01 -1.96570069e-01
-1.23908520e+00 -1.25021195e+00 6.27521276e-01 -1.86213002e-01
6.77773714e-01 -4.10435021e-01 -5.60882568e-01 8.28772485e-01
1.18156441e-01 3.90364230e-01 6.27210915e-01 3.16040993e-01
-4.88874108e-01 -2.07574457e-01 -9.74798977e-01 3.41568887e-01
1.33287549e+00 -1.00653720e+00 -3.14110875e-01 6.38284326e-01
6.77290559e-01 -1.99791580e-01 -6.21437550e-01 3.59667093e-01
3.89277965e-01 -8.33003461e-01 1.04538918e+00 -8.26914966e-01
2.08008528e-01 -1.91931963e-01 -4.49335665e-01 -1.22175097e+00
-7.79226795e-02 -7.62204409e-01 -4.09055620e-01 1.38093901e+00
3.64273310e-01 -5.83084226e-01 3.73211026e-01 -1.29538745e-01
-3.58609557e-01 -6.94446445e-01 -1.19875312e+00 -8.29109013e-01
2.57276334e-02 -6.85322940e-01 3.24845552e-01 7.47428536e-01
-1.32951774e-02 3.46842438e-01 -5.52364111e-01 2.64946759e-01
3.09227079e-01 2.05869973e-01 5.99928916e-01 -1.06003368e+00
-4.97886628e-01 -4.81546819e-01 -6.80072486e-01 -1.48504364e+00
4.42960143e-01 -9.20585155e-01 2.59325445e-01 -1.51280904e+00
1.99338779e-01 -4.10509259e-01 -4.94567424e-01 8.85709703e-01
-1.30419195e-01 5.25068641e-01 1.21026896e-01 1.87108621e-01
-8.29309702e-01 5.10374606e-01 9.93387640e-01 -2.71864176e-01
-1.67434499e-01 2.76646435e-01 -9.22063053e-01 6.31073356e-01
9.66125727e-01 -5.22721112e-02 -4.87604141e-01 -5.48203051e-01
-2.66925037e-01 1.03100225e-01 8.75351787e-01 -1.00775397e+00
2.41882801e-01 -3.64513099e-02 2.81069458e-01 -2.47060522e-01
6.83527589e-01 -3.90840948e-01 -5.41087151e-01 -1.39139071e-01
-6.80744052e-01 -2.16250524e-01 3.28200907e-01 7.82692313e-01
-5.21230340e-01 8.57063197e-03 8.63264561e-01 8.08838680e-02
-9.87075865e-01 4.76954132e-01 -2.44834393e-01 1.90937802e-01
1.03084564e+00 -8.65857303e-02 -2.71977037e-01 -9.21858788e-01
-1.21881878e+00 6.42283112e-02 2.40845187e-03 6.90171182e-01
7.11246192e-01 -1.36107826e+00 -6.99648619e-01 1.25416428e-01
2.34112635e-01 -1.92879379e-01 2.09082276e-01 1.09727383e+00
1.72339275e-01 4.53913718e-01 -1.80579592e-02 -1.00046575e+00
-1.36958802e+00 4.45735067e-01 2.78840214e-01 6.44813180e-02
-7.27851152e-01 1.17344248e+00 7.52920568e-01 -2.34026134e-01
7.40294755e-01 -4.12519038e-01 -2.72237599e-01 1.21943071e-01
4.86186355e-01 1.59019530e-01 6.21684156e-02 -6.38266623e-01
-4.76297647e-01 6.68992519e-01 2.09143281e-01 -6.60594041e-03
1.28955960e+00 -2.89667964e-01 3.85037988e-01 6.04828119e-01
1.40550816e+00 -8.13703015e-02 -1.50323772e+00 -2.92027801e-01
-9.12530795e-02 -2.90803105e-01 2.04752862e-01 -7.42179394e-01
-1.18288875e+00 1.31107700e+00 7.18436360e-01 2.79303104e-01
1.42075598e+00 5.43223202e-01 5.41629076e-01 1.25600353e-01
2.47223154e-01 -8.99115741e-01 5.03616393e-01 5.36062658e-01
1.03417373e+00 -1.19326997e+00 -4.56214249e-01 -9.51686800e-02
-7.13887632e-01 1.07066798e+00 4.92642820e-01 1.54559419e-01
5.58886111e-01 1.80793524e-01 -5.23807816e-02 7.74611905e-03
-8.48488688e-01 -7.63172686e-01 6.02566898e-01 8.42130601e-01
4.51238364e-01 5.66657446e-02 2.28882119e-01 5.84340572e-01
1.38592646e-01 -2.55713642e-01 5.13575822e-02 8.97308230e-01
-4.46637303e-01 -7.92788923e-01 -3.26179326e-01 3.75231743e-01
-4.22358274e-01 -6.34741187e-02 -5.02821207e-01 7.50619709e-01
1.97430655e-01 1.26224327e+00 1.86812431e-01 -3.77651036e-01
9.88145769e-02 1.98468253e-01 4.22416061e-01 -7.57879496e-01
-3.38801444e-01 3.57837528e-01 1.22913502e-01 -6.94078743e-01
-5.02468050e-01 -8.05180252e-01 -1.06695020e+00 1.69514924e-01
-1.59386396e-01 -8.93084630e-02 4.31236982e-01 8.42886806e-01
2.96595365e-01 5.56526423e-01 6.20993018e-01 -9.39593494e-01
-6.04945719e-01 -9.53266025e-01 -5.45877814e-01 3.63968641e-01
8.05990517e-01 -8.33405435e-01 -4.35589820e-01 7.91283995e-02] | [14.66942024230957, 4.8974738121032715] |
9e8e8bf3-d311-4c8e-98bf-73048f5d3c9e | openstance-real-world-zero-shot-stance | 2210.14299 | null | https://arxiv.org/abs/2210.14299v1 | https://arxiv.org/pdf/2210.14299v1.pdf | OpenStance: Real-world Zero-shot Stance Detection | Prior studies of zero-shot stance detection identify the attitude of texts towards unseen topics occurring in the same document corpus. Such task formulation has three limitations: (i) Single domain/dataset. A system is optimized on a particular dataset from a single domain; therefore, the resulting system cannot work well on other datasets; (ii) the model is evaluated on a limited number of unseen topics; (iii) it is assumed that part of the topics has rich annotations, which might be impossible in real-world applications. These drawbacks will lead to an impractical stance detection system that fails to generalize to open domains and open-form topics. This work defines OpenStance: open-domain zero-shot stance detection, aiming to handle stance detection in an open world with neither domain constraints nor topic-specific annotations. The key challenge of OpenStance lies in the open-domain generalization: learning a system with fully unspecific supervision but capable of generalizing to any dataset. To solve OpenStance, we propose to combine indirect supervision, from textual entailment datasets, and weak supervision, from data generated automatically by pre-trained Language Models. Our single system, without any topic-specific supervision, outperforms the supervised method on three popular datasets. To our knowledge, this is the first work that studies stance detection under the open-domain zero-shot setting. All data and code are publicly released. | ['Wenpeng Yin', 'Slobodan Vucetic', 'Hanzi Xu'] | 2022-10-25 | null | null | null | null | ['stance-detection'] | ['natural-language-processing'] | [ 3.27581853e-01 5.03989935e-01 -5.60152471e-01 -2.50767231e-01
-8.82036865e-01 -6.72661006e-01 9.05191362e-01 1.86572701e-01
-2.52532125e-01 8.48505080e-01 2.90799618e-01 -1.38623253e-01
2.32475419e-02 -7.62769341e-01 -5.90475321e-01 -5.84579706e-01
3.13308775e-01 9.40258741e-01 6.26637399e-01 -6.59673989e-01
1.69666857e-01 -3.58620435e-01 -1.57771838e+00 4.99531955e-01
1.02450323e+00 7.87593603e-01 2.13347599e-01 2.30539635e-01
-2.47930691e-01 5.83551824e-01 -8.93408179e-01 -5.21610081e-01
1.03486106e-01 -2.37620503e-01 -1.12176263e+00 3.34778547e-01
5.07402956e-01 -1.77516654e-01 1.93463981e-01 9.21472013e-01
3.92215908e-01 -2.09621474e-01 8.85226965e-01 -1.22484863e+00
-5.67803264e-01 8.31463873e-01 -3.17536592e-01 2.39805624e-01
4.55484629e-01 -1.64348409e-01 1.29374540e+00 -7.92042375e-01
1.02487087e+00 1.22720611e+00 4.74369019e-01 5.04081786e-01
-1.19204414e+00 -2.42018700e-01 3.95061910e-01 3.59314755e-02
-1.00790226e+00 -1.21782854e-01 9.32956815e-01 -6.65836036e-01
8.82926643e-01 1.53839454e-01 3.87404203e-01 1.89232397e+00
-5.73199727e-02 9.47366059e-01 1.06809759e+00 -4.75259423e-01
3.35867763e-01 5.34189284e-01 6.91973507e-01 1.24201484e-01
3.07461858e-01 -4.30009425e-01 -5.22188544e-01 -3.99517059e-01
4.40238975e-02 -2.62283176e-01 -3.38148624e-01 -4.43718433e-01
-1.22978175e+00 1.21299362e+00 -2.66431458e-02 6.47433579e-01
-9.61331204e-02 -6.34471416e-01 9.21533942e-01 5.51417112e-01
1.01119792e+00 6.22436285e-01 -7.68516481e-01 -1.20101355e-01
-1.07957983e+00 6.74529552e-01 1.20609891e+00 1.05268359e+00
7.15246081e-01 -2.54078954e-01 1.10939825e-02 8.04575741e-01
8.16374347e-02 5.43305218e-01 9.07108784e-01 -5.46416819e-01
8.39907527e-01 8.25185776e-01 4.41158116e-02 -7.78495550e-01
-3.07055086e-01 -3.89646471e-01 -1.65545106e-01 -1.96998119e-01
5.36502838e-01 -3.42463821e-01 -5.51881075e-01 1.78184652e+00
4.94143903e-01 -3.20521742e-01 4.52902019e-01 7.97950268e-01
1.16788447e+00 6.87519252e-01 -1.78231239e-01 -3.86048019e-01
1.67732692e+00 -8.63633335e-01 -9.23524678e-01 -6.11927927e-01
9.96124804e-01 -9.28707242e-01 1.36823630e+00 5.23408651e-01
-7.75180101e-01 -1.61610886e-01 -1.19429111e+00 -1.25489801e-01
-7.53911912e-01 1.13800867e-02 1.26824513e-01 5.63795865e-01
-4.84871328e-01 1.67792022e-01 -3.03656042e-01 -7.09856331e-01
4.81602736e-02 7.29444996e-03 -6.34486601e-02 1.99101165e-01
-1.82625830e+00 1.03777313e+00 7.31768012e-01 -5.65859377e-01
-6.94616497e-01 -6.83891058e-01 -9.41546798e-01 -3.29209387e-01
1.06949425e+00 -4.92674947e-01 1.45239651e+00 -1.18456471e+00
-1.29127586e+00 1.21907580e+00 -1.71647996e-01 -6.83042884e-01
4.94036227e-01 -5.22616386e-01 -2.73966730e-01 2.55015697e-02
6.88072920e-01 1.65605530e-01 1.05699611e+00 -1.06684399e+00
-2.07268849e-01 -4.32893723e-01 3.62332523e-01 2.15571553e-01
-6.84059262e-01 2.03412682e-01 8.05424675e-02 -7.86793053e-01
-8.22091550e-02 -8.75210464e-01 2.42335439e-01 -4.85675007e-01
-4.29344684e-01 -6.27424717e-01 1.41538358e+00 -4.17221904e-01
1.30701673e+00 -1.88443553e+00 1.96473509e-01 -4.03463662e-01
4.07165587e-02 4.04571533e-01 2.66627848e-01 5.77216268e-01
-2.47223745e-03 -1.22435540e-02 -2.30659783e-01 -1.30078793e-01
2.34517798e-01 4.39428836e-01 -7.78170466e-01 3.16415131e-01
6.35057911e-02 6.05193496e-01 -1.04266644e+00 -9.33294833e-01
-9.20063928e-02 -8.45245644e-02 -5.45907080e-01 8.24523419e-02
-7.46250629e-01 2.97463257e-02 -4.55017775e-01 4.00130242e-01
4.59304482e-01 -2.07471907e-01 2.17057258e-01 -7.31831044e-02
-1.44396320e-01 6.16284192e-01 -1.03047144e+00 1.61773801e+00
-4.65582520e-01 5.52970588e-01 -1.50753722e-01 -1.30852091e+00
9.72282827e-01 8.10561538e-01 4.27644461e-01 -2.82331526e-01
2.78835773e-01 4.28053975e-01 -2.03705251e-01 -7.00478911e-01
6.25545502e-01 -2.98035473e-01 -6.11590445e-01 7.98676610e-01
4.00709391e-01 -3.02439660e-01 5.20085156e-01 4.15158004e-01
7.56473362e-01 3.97083983e-02 5.32027900e-01 -5.25705636e-01
4.81340975e-01 1.75615475e-01 5.71357608e-01 5.60905874e-01
-1.96424395e-01 5.07187247e-01 7.47174621e-01 -2.80082673e-01
-1.01107585e+00 -8.55515957e-01 -5.59845984e-01 1.56640518e+00
2.78953798e-02 -7.04033613e-01 -7.85597742e-01 -9.76223171e-01
-2.99517393e-01 7.30798006e-01 -6.45124018e-01 1.41118005e-01
-5.64545453e-01 -6.93837345e-01 5.70802629e-01 1.99129552e-01
2.52024770e-01 -8.78113806e-01 -7.88751006e-01 2.19867602e-01
-8.71901870e-01 -1.39209890e+00 -2.47972116e-01 3.15119714e-01
-6.79923415e-01 -1.13468027e+00 -7.03083217e-01 -8.26860905e-01
2.40997285e-01 1.10221043e-01 1.23008263e+00 -4.00037259e-01
2.03643665e-01 7.09488317e-02 -6.86191380e-01 -6.78847373e-01
-6.02315843e-01 3.42954904e-01 4.96192798e-02 -1.66533828e-01
7.23300934e-01 -2.03515410e-01 -9.22565237e-02 3.11750412e-01
-9.34502959e-01 -1.55364826e-01 1.98965669e-01 1.07042933e+00
8.16938058e-02 -9.62372720e-02 6.87970042e-01 -1.24962330e+00
7.84242690e-01 -7.71028996e-01 -1.25006378e-01 1.48582727e-01
-4.21395093e-01 -6.97047710e-02 5.85513830e-01 -6.37511551e-01
-1.28332722e+00 -3.97146970e-01 -1.52321917e-03 -5.59366681e-02
-2.42042795e-01 5.20567775e-01 -2.82061666e-01 7.23086715e-01
1.11752617e+00 -1.03093684e-02 -1.50627881e-01 -4.28943843e-01
4.42964494e-01 9.51240897e-01 2.22151637e-01 -7.88692117e-01
5.93884051e-01 6.17262542e-01 -4.86346722e-01 -1.15742958e+00
-1.61328149e+00 -7.87468493e-01 -9.16174352e-01 -6.30405033e-03
8.99042070e-01 -9.77949798e-01 1.24859907e-01 2.19575047e-01
-1.29851663e+00 -2.38574564e-01 -2.05931589e-01 1.71274379e-01
-6.89637125e-01 6.79993689e-01 -5.47201633e-01 -5.01034081e-01
-5.72116673e-01 -8.13151419e-01 1.27417254e+00 -2.48895794e-01
-9.95860100e-01 -1.26382267e+00 2.48032495e-01 5.89146435e-01
-2.26163585e-02 3.14549327e-01 7.22113907e-01 -1.22479618e+00
1.50750488e-01 -1.10433541e-01 2.57155120e-01 4.70669001e-01
1.11731797e-01 5.36878221e-03 -9.81980741e-01 -2.29275331e-01
5.40416241e-01 -1.01208866e+00 6.59853578e-01 1.92665696e-01
4.89633977e-01 -6.30765498e-01 -2.54839242e-01 6.01144182e-03
1.03296900e+00 -3.20003599e-01 6.34341910e-02 6.82751477e-01
1.94654316e-01 8.26201379e-01 1.17337751e+00 4.89367455e-01
3.20612043e-01 7.69595027e-01 1.75174564e-01 8.07805173e-03
6.79164156e-02 -4.40894030e-02 5.86238742e-01 6.96168363e-01
1.49878934e-01 -2.70587653e-01 -9.43276107e-01 8.68106306e-01
-1.96779490e+00 -1.03786242e+00 -4.38972503e-01 1.89582348e+00
1.18061006e+00 5.56740761e-01 2.32630104e-01 3.27398062e-01
5.23475468e-01 4.07861084e-01 -1.78208143e-01 -5.96508861e-01
-2.99478352e-01 2.89714094e-02 -4.38490957e-02 3.02907854e-01
-1.49268115e+00 9.32759404e-01 5.33907318e+00 6.29992306e-01
-1.04963350e+00 6.35052085e-01 -1.08701014e-03 -1.13987133e-01
-1.37041986e-01 1.80250272e-01 -1.08004522e+00 5.31852067e-01
8.89574111e-01 -5.02636135e-01 -5.64166963e-01 1.23481667e+00
7.60774538e-02 4.54903543e-02 -1.21937025e+00 3.15441400e-01
6.06385827e-01 -8.64586890e-01 -3.59351747e-02 4.35363203e-02
8.55315566e-01 4.92943339e-02 -1.63971949e-02 5.42980254e-01
2.36853406e-01 -3.53557646e-01 8.49754095e-01 -2.03164116e-01
5.00298738e-01 -3.06268066e-01 9.09854829e-01 8.98495018e-01
-6.38267040e-01 3.44429836e-02 -4.44509149e-01 -1.99395388e-01
1.80990353e-01 8.06914210e-01 -1.08081818e+00 4.70451653e-01
5.91882110e-01 8.37456822e-01 -2.93945312e-01 3.02109152e-01
-5.29472768e-01 7.36644268e-01 -3.32067281e-01 -1.88317522e-01
4.02390718e-01 1.56923026e-01 8.97838771e-01 1.36603904e+00
2.68034507e-02 -1.41556308e-01 5.21855593e-01 8.22620749e-01
2.20537350e-01 3.22954446e-01 -8.95961106e-01 2.05880791e-01
3.32872540e-01 1.00766790e+00 -4.81441170e-01 -6.65196121e-01
-5.16372621e-01 7.32439399e-01 1.95142522e-01 1.37325853e-01
-8.13309968e-01 -1.57984152e-01 5.19392252e-01 3.98238510e-01
2.79118419e-01 1.04347102e-01 -3.89137954e-01 -1.37644958e+00
8.82011577e-02 -1.00030839e+00 7.29527950e-01 -5.02723634e-01
-1.56395781e+00 5.29446959e-01 4.39300001e-01 -1.36435020e+00
-5.95974445e-01 -5.39045990e-01 -4.86639589e-01 5.35870433e-01
-1.36729300e+00 -1.30043566e+00 -6.96755499e-02 5.20539522e-01
1.08539498e+00 9.22843162e-03 8.67067039e-01 1.49337992e-01
-3.60745162e-01 2.50598043e-01 -9.43908021e-02 9.65068638e-02
1.28007066e+00 -1.32914901e+00 6.14604726e-02 7.11603284e-01
-1.18921399e-01 7.18353391e-01 1.18204451e+00 -7.94055521e-01
-7.16511130e-01 -1.09596157e+00 1.39771938e+00 -7.68427610e-01
1.15841031e+00 -7.00493693e-01 -1.21215260e+00 8.85963380e-01
4.54412550e-01 -4.37524706e-01 8.69762897e-01 5.61661124e-01
-5.00548124e-01 3.60425264e-01 -7.92358458e-01 4.52182084e-01
8.06613088e-01 -4.50517565e-01 -1.63322091e+00 8.24137688e-01
8.50008786e-01 -5.19703507e-01 -8.17345500e-01 3.37688178e-01
1.29979923e-01 -6.60019457e-01 8.87299120e-01 -6.14777982e-01
8.35680246e-01 -3.25471833e-02 -9.71885100e-02 -1.22800004e+00
1.65313095e-01 -4.04659063e-01 -3.65187287e-01 1.39830387e+00
5.19459903e-01 -5.42686522e-01 4.73203599e-01 2.10319370e-01
-1.66493520e-01 -6.32102549e-01 -9.14298654e-01 -8.63885939e-01
6.32049203e-01 -2.95998365e-01 3.35080385e-01 1.22956455e+00
5.75798035e-01 9.36325133e-01 -4.61307436e-01 1.12308459e-02
4.69387531e-01 5.48096061e-01 8.72256756e-01 -1.48310995e+00
-2.51176000e-01 -1.37823299e-01 -1.38082013e-01 -1.00278759e+00
4.59160030e-01 -7.43744910e-01 1.42742163e-02 -1.36341250e+00
3.06648016e-01 -2.99266487e-01 3.23255539e-01 4.32541490e-01
-1.71006590e-01 3.99204046e-02 -9.24618393e-02 3.41847748e-01
-7.25204229e-01 6.14299476e-01 9.72007275e-01 -2.61974871e-01
-1.06519610e-01 9.86868218e-02 -8.10077310e-01 1.07041776e+00
8.40136230e-01 -6.09894037e-01 -6.39587998e-01 -2.10131288e-01
1.99064031e-01 -2.58039415e-01 2.41027862e-01 -6.37484252e-01
-1.21562611e-02 -2.01771870e-01 -3.13668936e-01 -7.75299430e-01
3.75901759e-01 -4.43733275e-01 -5.47070265e-01 2.68770456e-01
-2.64187992e-01 -4.02332574e-01 -1.95792168e-01 4.44577903e-01
-3.84776890e-01 -7.00642288e-01 6.80826366e-01 -3.69463354e-01
-6.70840025e-01 -1.59542322e-01 -5.10845900e-01 7.48574913e-01
1.23692000e+00 -2.22615704e-01 -6.71315372e-01 -2.25645497e-01
-7.27982759e-01 1.11406341e-01 5.65019071e-01 6.25602901e-01
1.15029886e-01 -1.00944114e+00 -8.35623443e-01 -1.18523508e-01
5.14500201e-01 8.54447559e-02 -8.20127223e-03 7.52843797e-01
8.54702145e-02 6.20825231e-01 2.24094614e-01 -8.84206414e-01
-1.37230849e+00 7.83316970e-01 -2.67437641e-02 -6.25575900e-01
-6.84760392e-01 4.80382383e-01 4.46070582e-01 -7.39201128e-01
2.56650299e-02 -2.15853944e-01 -4.07382101e-01 5.65389276e-01
4.18184012e-01 4.90073999e-03 2.70331502e-01 -8.23940456e-01
-1.69368491e-01 4.36847419e-01 -2.44388163e-01 -8.34829435e-02
1.05658126e+00 -2.23021716e-01 -1.32876644e-02 9.39833641e-01
1.13175869e+00 -1.60863653e-01 -7.85622239e-01 -7.33624995e-01
3.11023682e-01 -2.43140742e-01 -1.80978239e-01 -5.38780749e-01
-2.12254897e-01 9.41560268e-01 -1.07251406e-01 3.80763680e-01
6.03588700e-01 5.01535475e-01 8.61088753e-01 4.38071847e-01
2.55554080e-01 -1.65318096e+00 5.11991203e-01 1.02847588e+00
9.45078552e-01 -1.43160129e+00 1.29978061e-01 -5.92503250e-01
-8.57763469e-01 1.05362272e+00 7.67449975e-01 1.10346407e-01
6.09580219e-01 3.23103487e-01 2.94150561e-01 -4.73788798e-01
-1.02898562e+00 -1.89459056e-01 1.85421184e-01 4.05544609e-01
6.91531897e-01 1.27389783e-03 -7.94252574e-01 7.22914279e-01
-6.31540596e-01 -1.98822364e-01 7.07713902e-01 1.05714035e+00
-6.19555712e-01 -1.06556237e+00 -4.66983855e-01 2.82564610e-01
-5.80972314e-01 2.83171814e-02 -7.92526066e-01 8.26588929e-01
2.40721405e-01 1.04095495e+00 -1.22565024e-01 1.29780889e-01
3.38668436e-01 3.62292975e-01 9.69035551e-02 -1.13697684e+00
-4.70441818e-01 4.72675115e-02 4.92112577e-01 -1.68875724e-01
-6.21796727e-01 -9.58528161e-01 -1.00050092e+00 5.49104391e-03
-5.12956321e-01 2.84882694e-01 2.59873241e-01 1.23035002e+00
-3.07522640e-02 3.61775041e-01 3.32998365e-01 -4.45554465e-01
-8.98036122e-01 -1.25396979e+00 -4.07109618e-01 7.81123877e-01
2.37131268e-01 -8.99686515e-01 -4.17118907e-01 1.96886495e-01] | [9.018631935119629, 9.944207191467285] |
f338908d-ca6f-4bca-837e-c4ade9961860 | adversarial-examples-for-electrocardiograms | 1905.05163 | null | https://arxiv.org/abs/1905.05163v2 | https://arxiv.org/pdf/1905.05163v2.pdf | Adversarial Examples for Electrocardiograms | In recent years, the electrocardiogram (ECG) has seen a large diffusion in both medical and commercial applications, fueled by the rise of single-lead versions. Single-lead ECG can be embedded in medical devices and wearable products such as the injectable Medtronic Linq monitor, the iRhythm Ziopatch wearable monitor, and the Apple Watch Series 4. Recently, deep neural networks have been used to automatically analyze ECG tracings, outperforming even physicians specialized in cardiac electrophysiology in detecting certain rhythm irregularities. However, deep learning classifiers have been shown to be brittle to adversarial examples, which are examples created to look incontrovertibly belonging to a certain class to a human eye but contain subtle features that fool the classifier into misclassifying them into the wrong class. Very recently, adversarial examples have also been created for medical-related tasks. Yet, traditional attack methods to create adversarial examples, such as projected gradient descent (PGD) do not extend directly to ECG signals, as they generate examples that introduce square wave artifacts that are not physiologically plausible. Here, we developed a method to construct smoothed adversarial examples for single-lead ECG. First, we implemented a neural network model achieving state-of-the-art performance on the data from the 2017 PhysioNet/Computing-in-Cardiology Challenge for arrhythmia detection from single lead ECG classification. For this model, we utilized a new technique to generate smoothed examples to produce signals that are 1) indistinguishable to cardiologists from the original examples and 2) incorrectly classified by the neural network. Finally, we show that adversarial examples are not unique and provide a general technique to collate and perturb known adversarial examples to create new ones. | ['Rajesh Ranganath', 'Luca Foschini', 'Yuxuan Hu', 'Xintian Han', 'Lior Jankelson', 'Larry Chinitz'] | 2019-05-13 | null | null | null | null | ['arrhythmia-detection', 'ecg-classification', 'electrocardiography-ecg'] | ['medical', 'medical', 'methodology'] | [ 5.27908027e-01 3.11804503e-01 4.38649327e-01 -2.24900678e-01
-8.59072983e-01 -9.08494592e-01 -1.31351337e-01 6.10066429e-02
-1.50356099e-01 9.81241703e-01 -2.66366839e-01 -6.06915534e-01
-5.86665757e-02 -5.66999495e-01 -8.11435819e-01 -6.15188718e-01
-4.92698729e-01 2.67000586e-01 -2.76356906e-01 -1.50462165e-01
-2.48880088e-01 5.75450242e-01 -7.98427403e-01 3.82834792e-01
7.88055599e-01 8.43101382e-01 -7.56548166e-01 9.27115142e-01
6.46511197e-01 4.86659467e-01 -1.24145758e+00 -3.79679590e-01
4.40925807e-01 -8.64719510e-01 -3.14542681e-01 -5.25174022e-01
3.63353014e-01 -2.02612936e-01 -1.85507774e-01 8.60033154e-01
1.14033222e+00 -5.76969981e-01 5.19332826e-01 -1.05837095e+00
-4.49246615e-01 7.87440658e-01 -2.62317181e-01 3.19243431e-01
3.37170690e-01 3.80930960e-01 4.53113407e-01 -5.00780761e-01
5.85302174e-01 7.32811987e-01 1.31713593e+00 7.35036135e-01
-1.36473143e+00 -7.57256389e-01 -4.73914415e-01 -8.63165110e-02
-1.24713659e+00 1.35219563e-02 1.05737889e+00 -3.05620313e-01
4.65172410e-01 7.95251369e-01 9.46916640e-01 1.62407982e+00
5.56730092e-01 4.58377331e-01 1.18066740e+00 -2.24400058e-01
2.31000453e-01 3.97603847e-02 -7.51302093e-02 3.52871448e-01
2.87343115e-01 5.19001782e-01 -6.51244819e-02 -6.65964603e-01
7.86550820e-01 -6.22346401e-02 -7.13805437e-01 -1.29400402e-01
-1.32454848e+00 7.01583385e-01 3.32545936e-01 3.10506999e-01
-4.83690351e-01 2.51882493e-01 6.13333106e-01 4.74281073e-01
2.83427089e-01 1.04914570e+00 -3.78973097e-01 -7.19765127e-02
-7.09022045e-01 3.95688295e-01 9.23445225e-01 2.81569779e-01
8.37846398e-02 5.41166484e-01 -2.98579365e-01 4.58555609e-01
-1.39486149e-01 5.29581249e-01 6.21442139e-01 -8.65368783e-01
2.24692598e-01 1.91936657e-01 1.67563111e-01 -1.09471178e+00
-6.61979377e-01 -7.73052514e-01 -1.24468458e+00 3.89349133e-01
5.91363966e-01 -6.04739010e-01 -8.08584809e-01 1.60705447e+00
1.29321426e-01 4.02850389e-01 1.79875299e-01 9.60829139e-01
7.85310328e-01 1.71230584e-01 -2.70689756e-01 -7.22680464e-02
9.96846437e-01 -2.47815609e-01 -6.13805294e-01 1.04547059e-02
5.80347419e-01 -5.45984507e-01 9.60916281e-01 7.78900385e-01
-1.18854320e+00 -5.98661005e-01 -1.39303076e+00 5.39917529e-01
-1.32843837e-01 -3.87787730e-01 4.16420609e-01 1.17801607e+00
-7.85526454e-01 1.11702836e+00 -7.09981740e-01 2.11749911e-01
7.82128572e-01 2.32412145e-01 -2.25673616e-01 4.28072363e-01
-1.64099073e+00 8.92385602e-01 1.24648027e-01 1.22955278e-01
-1.03238285e+00 -1.13691318e+00 -5.93129396e-01 6.88227639e-03
-3.60489711e-02 -7.53714621e-01 8.60263467e-01 -1.22381401e+00
-1.33016825e+00 1.02595520e+00 4.85700041e-01 -9.81561184e-01
8.57184768e-01 -1.58171490e-01 -7.96322405e-01 1.34856418e-01
-3.12316120e-01 8.96734893e-02 1.14845586e+00 -1.12366164e+00
1.67924527e-03 -2.28876576e-01 -5.02483211e-02 -3.58078390e-01
-5.65228909e-02 -1.87520385e-01 4.65909183e-01 -1.12265050e+00
-4.44588847e-02 -8.55032921e-01 -4.21292335e-01 -1.02563515e-01
-8.35922539e-01 4.50512618e-01 5.14486194e-01 -6.29190743e-01
1.14444149e+00 -2.29787445e+00 -8.41045678e-02 4.88355666e-01
4.99932528e-01 6.50961041e-01 1.37369201e-01 1.25052258e-01
-6.13469958e-01 4.91133213e-01 -6.04015350e-01 1.44324407e-01
-2.16442496e-01 1.27058089e-01 -6.40323162e-01 5.98480642e-01
1.67900354e-01 1.11555541e+00 -9.69582260e-01 1.31183770e-02
-9.40332282e-03 5.27882457e-01 -3.83635730e-01 2.94157326e-01
2.00854242e-01 9.12150562e-01 -1.85228825e-01 3.99305284e-01
5.20336211e-01 -2.02243086e-02 4.31769229e-02 -1.43881649e-01
5.99392295e-01 -3.43448594e-02 -9.57378864e-01 1.39269137e+00
-1.63270473e-01 4.16888893e-01 -2.02142835e-01 -1.08898938e+00
8.84936810e-01 6.94678664e-01 5.44673860e-01 -1.54110029e-01
1.58187002e-01 3.17981958e-01 4.19664681e-01 -5.26466072e-01
-2.44980127e-01 -4.31264162e-01 -6.03770949e-02 4.72146153e-01
-3.92009288e-01 -3.35837632e-01 -4.70790595e-01 -4.20755222e-02
1.42754078e+00 -2.31573381e-03 2.98479080e-01 -5.78578077e-02
3.16414565e-01 -1.79245695e-01 7.26823032e-01 1.08135796e+00
-2.82192916e-01 1.09868407e+00 4.91333365e-01 -1.00759912e+00
-7.80661404e-01 -1.43735492e+00 -2.82661706e-01 1.09868348e-02
-1.91281706e-01 -1.54888421e-01 -9.11608577e-01 -7.72492230e-01
4.06973101e-02 5.34118235e-01 -5.47088981e-01 -7.06930041e-01
-7.51339555e-01 -8.47697973e-01 1.47512615e+00 4.69450593e-01
1.02243274e-01 -1.28147912e+00 -1.04106951e+00 5.03119469e-01
5.95341064e-02 -8.60383391e-01 -3.02378207e-01 2.64828503e-01
-8.74633431e-01 -1.23159456e+00 -9.28122759e-01 -4.14602131e-01
5.61071098e-01 -8.42007279e-01 1.35875452e+00 1.35264099e-01
-8.30647111e-01 2.46037126e-01 -1.87233195e-01 -1.02835310e+00
-1.01260030e+00 -2.94123828e-01 4.00002599e-01 2.10188389e-01
2.21107658e-02 -8.43603671e-01 -8.56068850e-01 2.34137028e-01
-8.05780232e-01 -3.16866398e-01 3.49482298e-01 1.02774012e+00
5.69802463e-01 -2.32295290e-01 1.05819070e+00 -1.26907134e+00
1.00754201e+00 -3.33628863e-01 -8.85575488e-02 -1.78953618e-01
-6.22386813e-01 -3.19779515e-01 1.10163915e+00 -7.34332025e-01
-3.41878265e-01 -6.98654726e-02 -5.72208285e-01 -6.69765472e-01
-3.34120274e-01 4.12101537e-01 1.22775413e-01 -1.47322372e-01
1.30423760e+00 9.86958221e-02 5.58875874e-02 -2.21441075e-01
3.52914724e-03 5.09387434e-01 8.13216269e-01 -3.85255039e-01
9.37106252e-01 4.01107579e-01 8.17583948e-02 -5.65837264e-01
-6.13790989e-01 3.14796239e-01 -1.11961372e-01 -9.65573173e-03
6.26072228e-01 -5.16949415e-01 -6.88251793e-01 6.21007860e-01
-1.13017476e+00 -3.49277854e-01 -7.70808637e-01 3.13328177e-01
-5.30827343e-01 5.13713658e-01 -5.96496880e-01 -5.90808809e-01
-7.36147642e-01 -9.26981807e-01 6.14200354e-01 -8.79280791e-02
-7.45841980e-01 -9.00399864e-01 9.41168666e-02 -1.54315591e-01
6.06874466e-01 1.22483575e+00 1.00285411e+00 -9.03834999e-01
-4.36796919e-02 -6.27831221e-01 5.67785621e-01 7.54788935e-01
3.13292891e-01 -1.81693569e-01 -1.14669240e+00 -3.74915242e-01
4.15178865e-01 -1.45512119e-01 1.92772493e-01 4.81605113e-01
1.44552195e+00 -3.40169340e-01 -2.97435224e-01 1.08832371e+00
9.18938994e-01 3.86748970e-01 9.36014295e-01 -7.38897324e-02
7.35245466e-01 8.29304084e-02 2.53010876e-02 8.17806125e-02
-3.20761442e-01 4.80615586e-01 3.82232875e-01 -6.38049841e-01
9.18565094e-02 -3.03981714e-02 3.83952484e-02 2.81236082e-01
-2.96367496e-01 3.13082486e-02 -8.04076910e-01 2.33909070e-01
-1.39550257e+00 -9.62733686e-01 -2.31562600e-01 2.38332415e+00
9.64183688e-01 4.66351271e-01 8.37066174e-02 7.39928305e-01
5.67238450e-01 -2.03852624e-01 -8.50212574e-01 -6.56731963e-01
-2.24418446e-01 8.19954216e-01 2.09193558e-01 8.43141377e-02
-1.00038862e+00 3.96936566e-01 6.46920204e+00 3.65887851e-01
-1.42715085e+00 8.18340182e-02 9.48453963e-01 1.40544176e-02
-1.75659850e-01 -4.26758856e-01 -7.37898201e-02 6.20020449e-01
9.81693625e-01 -1.84975371e-01 2.58002996e-01 7.43021846e-01
1.84265092e-01 6.12591684e-01 -1.35292792e+00 1.26961350e+00
-2.85255127e-02 -1.47084534e+00 -1.19113110e-01 -2.46395528e-01
5.53597867e-01 -2.78942317e-01 1.46584168e-01 1.44165471e-01
-2.97590584e-01 -1.46822536e+00 4.70227748e-01 4.95060503e-01
1.10011554e+00 -7.82409668e-01 7.99933910e-01 2.66260207e-01
-4.70771015e-01 2.11615525e-02 -6.09924458e-02 -1.48307113e-02
1.64633363e-01 7.27142632e-01 -9.02097464e-01 4.25635457e-01
6.50882661e-01 5.77537358e-01 -2.59219646e-01 9.12057877e-01
-2.23432794e-01 9.03100371e-01 -2.37567440e-01 2.95744479e-01
-2.23378718e-01 1.18390769e-01 9.90778446e-01 8.38387311e-01
3.29538554e-01 -8.37103426e-02 -1.43398181e-01 1.13157785e+00
-7.14036450e-02 -1.44403666e-01 -1.00242352e+00 2.12380961e-01
2.17382580e-01 1.09164894e+00 -4.13674027e-01 -2.74100184e-01
7.30663016e-02 9.88535225e-01 -2.41259173e-01 5.42242646e-01
-1.06480646e+00 -8.09618115e-01 5.41718423e-01 4.20237124e-01
-1.14733249e-01 3.22983563e-01 -6.99223995e-01 -9.51831281e-01
2.24813864e-01 -1.44344962e+00 2.54095972e-01 -6.31596029e-01
-1.45866013e+00 1.04015493e+00 -3.94177824e-01 -1.53548908e+00
-5.30164480e-01 -4.57749754e-01 -7.96502531e-01 1.26949632e+00
-9.41006303e-01 -6.54758215e-01 -2.08589032e-01 6.15167737e-01
-6.24949262e-02 -2.23612383e-01 1.37456429e+00 3.32192391e-01
-8.12427700e-02 9.93161082e-01 -2.92897642e-01 3.97726893e-01
6.98255241e-01 -1.32218373e+00 6.26737416e-01 8.04855824e-01
2.32532293e-01 6.44124448e-01 8.36223006e-01 -3.96841079e-01
-1.02182317e+00 -1.11488116e+00 4.80457783e-01 -8.14511657e-01
3.46824706e-01 -2.44138360e-01 -1.06535017e+00 5.13079166e-01
-9.87615883e-02 4.13719893e-01 5.98168612e-01 -3.44584018e-01
-1.79043874e-01 -2.83937216e-01 -1.43693650e+00 7.36882567e-01
8.12164366e-01 -4.52203244e-01 -6.88648522e-01 3.19392413e-01
3.53885174e-01 -8.30710649e-01 -1.08674157e+00 7.16371298e-01
7.12682426e-01 -1.05682480e+00 1.17733622e+00 -9.57453609e-01
4.37684029e-01 -9.67545062e-02 3.51575494e-01 -1.69035351e+00
-1.89402308e-02 -1.39141572e+00 -3.92657906e-01 6.68987989e-01
4.61794287e-01 -9.95562196e-01 6.52388990e-01 3.33507597e-01
-2.65673488e-01 -1.16026223e+00 -9.78750229e-01 -7.84352958e-01
2.51237452e-01 -3.73598874e-01 6.39173925e-01 1.07692051e+00
7.76810274e-02 1.26239344e-01 -6.06448770e-01 1.52270868e-01
4.87255692e-01 -7.95858949e-02 6.47473454e-01 -1.14583266e+00
-6.96906924e-01 -3.08252722e-01 -7.54879832e-01 -2.85099506e-01
-1.98201522e-01 -9.54454899e-01 -1.50437862e-01 -9.12268221e-01
-3.87588292e-01 -6.52738571e-01 -6.11722708e-01 3.18454921e-01
-6.02467120e-01 7.11831450e-01 -7.29404809e-03 -7.48763829e-02
3.03953111e-01 -2.22815186e-01 1.11018956e+00 -2.90158331e-01
-3.08236092e-01 4.28468525e-01 -8.18450570e-01 8.36727321e-01
7.87178993e-01 -7.87019491e-01 -1.91592708e-01 5.86477518e-02
4.59294587e-01 2.16667846e-01 6.53083205e-01 -1.27835131e+00
-4.37278479e-01 5.96118569e-01 6.28770411e-01 3.77087779e-02
1.61278158e-01 -8.37765336e-01 5.98500550e-01 9.51323509e-01
-2.92944521e-01 1.32447138e-01 2.91784346e-01 3.98782015e-01
-8.49894658e-02 -7.56516820e-03 8.91977251e-01 -5.80725074e-02
1.95049927e-01 2.95115560e-01 -4.60712314e-01 4.78993982e-01
1.01507497e+00 -2.93603390e-01 -1.56523734e-01 -5.25020242e-01
-1.08855939e+00 -2.85280794e-01 1.40099272e-01 2.61380941e-01
7.72549987e-01 -1.16825128e+00 -9.65481400e-01 4.55134124e-01
-7.64312744e-02 -6.78179832e-03 2.88579464e-01 9.13500547e-01
-7.80885518e-01 -1.70748040e-01 -1.82104990e-01 -7.96504080e-01
-1.14544940e+00 6.72959387e-01 8.69696736e-01 -2.48764575e-01
-9.82555747e-01 6.72312498e-01 8.09128955e-03 -2.80089468e-01
2.41666093e-01 -4.76739377e-01 8.87335613e-02 -3.61453205e-01
5.00762105e-01 9.03856084e-02 2.57329166e-01 1.19554624e-02
-3.91659379e-01 4.24163431e-01 2.97990650e-01 2.95877844e-01
1.10259950e+00 5.41759312e-01 1.95958257e-01 5.11160731e-01
8.99927378e-01 4.44374867e-02 -8.61238241e-01 2.56960362e-01
-2.68155664e-01 -1.62922919e-01 -4.62523878e-01 -1.13020563e+00
-1.15481865e+00 1.01911283e+00 1.04668629e+00 6.09221220e-01
1.10358202e+00 -5.48179805e-01 9.40166593e-01 3.44556987e-01
3.52126479e-01 -5.96984446e-01 -3.03718150e-02 4.55423221e-02
9.15000141e-01 -7.96694160e-01 -2.02906981e-01 -1.16171010e-01
-7.21659660e-01 9.90224004e-01 4.95082699e-02 -3.27006519e-01
6.95795894e-01 4.92418438e-01 7.02002883e-01 -3.31670530e-02
-5.04220389e-02 6.36437595e-01 1.65481657e-01 9.26855624e-01
4.14586604e-01 2.11527035e-01 -3.14409643e-01 9.58503902e-01
-2.53895313e-01 2.88956851e-01 6.48318827e-01 7.80348063e-01
3.27954412e-01 -9.91934836e-01 -4.68266428e-01 7.42540777e-01
-1.11014616e+00 -2.00832233e-01 -3.31393749e-01 5.95562935e-01
2.10424870e-01 6.95300937e-01 -2.37587124e-01 -4.01947945e-01
4.68928933e-01 2.64699399e-01 4.37730491e-01 -4.67848480e-01
-1.17365324e+00 -2.37504914e-01 1.10428043e-01 -5.05327225e-01
8.68681371e-02 -2.12482572e-01 -1.01405942e+00 1.04544334e-01
6.02904521e-02 1.72488824e-01 3.41352612e-01 7.19325602e-01
4.52775747e-01 9.93609190e-01 6.85469508e-01 -8.61922026e-01
-9.03773963e-01 -9.54030991e-01 -7.19337642e-01 7.69112885e-01
5.95536351e-01 -1.26884222e-01 -6.39710963e-01 1.52607352e-01] | [14.330135345458984, 3.1059048175811768] |
8f975327-93e4-468d-86a8-7849d1823999 | peach-tree-a-multiple-sequence-alignment-and | 2112.07422 | null | https://arxiv.org/abs/2112.07422v1 | https://arxiv.org/pdf/2112.07422v1.pdf | PEACH Tree: A Multiple Sequence Alignment and Tree Display Tool for Epidemiologists | PEACH Tree is an easy-to-use, online tool for displaying multiple sequence alignments and phylogenetic trees side-by-side. PEACH Tree is powerful for rapidly tracing evolutionary and transmission histories by filtering invariant sites out of the display, and allowing samples to readily be filtered out of the display. These features, coupled with the ability to display epidemiological metadata, make the tool suitable for infectious disease epidemiology. PEACH Tree further enables much needed communication between the fields of genomics and infectious disease epidemiology, as exemplified by the COVID-19 pandemic. | ['David Welch', 'Jordan Douglas'] | 2021-12-12 | null | null | null | null | ['multiple-sequence-alignment', 'epidemiology'] | ['medical', 'medical'] | [ 1.00035369e-01 -7.11882830e-01 -5.06523391e-03 -1.54137343e-01
-2.40117759e-01 -9.23208177e-01 2.67720193e-01 7.90648878e-01
-3.65441859e-01 6.95163608e-01 7.65871629e-03 -8.74153912e-01
-2.08647296e-01 -5.67235053e-01 -9.75782052e-02 -7.79894590e-01
-7.22945452e-01 5.58306932e-01 3.54334891e-01 -2.44066641e-01
2.00669318e-01 6.36396527e-01 -1.26585543e+00 2.35653818e-01
7.34843493e-01 -8.40112418e-02 5.13707280e-01 9.21752572e-01
-2.67840713e-01 5.60399033e-02 -6.35315359e-01 -3.05818588e-01
-1.63896963e-01 -3.42459589e-01 -3.84773791e-01 -7.99152076e-01
-2.61581779e-01 -1.87368870e-01 4.34417695e-01 5.27377903e-01
6.06783509e-01 -3.65179181e-01 4.47395176e-01 -1.02591240e+00
-1.70482054e-01 -1.52420014e-01 -4.72485244e-01 6.63975537e-01
6.30094230e-01 4.94829386e-01 7.56965041e-01 -5.93101561e-01
1.48686099e+00 1.16912234e+00 9.50415790e-01 1.80886477e-01
-1.22374165e+00 -5.39698780e-01 -2.68691808e-01 2.16432601e-01
-1.05839884e+00 -4.15881127e-02 7.37764835e-02 -9.36424255e-01
1.42341793e+00 7.46109962e-01 9.58628118e-01 7.36756086e-01
7.04816103e-01 8.51541460e-02 1.16845226e+00 1.81134745e-01
8.88242200e-02 -2.99405873e-01 -3.57915424e-02 6.74446046e-01
3.64860445e-01 1.82115138e-02 -3.41572613e-01 -1.15215719e+00
6.83076262e-01 3.81201327e-01 -2.60276228e-01 -3.39994758e-01
-1.06205368e+00 7.23912835e-01 1.12317078e-01 8.43650922e-02
-4.16115552e-01 -2.55228788e-01 7.80506313e-01 1.28939569e-01
4.80087727e-01 2.48012513e-01 -5.61060846e-01 -4.85947400e-01
-5.09451926e-01 3.16364437e-01 3.73975933e-01 3.77493441e-01
3.86625111e-01 -4.01043534e-01 3.50565761e-01 6.36638761e-01
2.73831069e-01 4.06302035e-01 -1.77615985e-01 -5.20071328e-01
-2.56925672e-01 4.05712426e-01 1.93711251e-01 -6.37339711e-01
-6.84094787e-01 -1.22064523e-01 -5.85015237e-01 2.92430013e-01
3.46801400e-01 -9.08299461e-02 -7.03928173e-01 1.44289255e+00
9.41518068e-01 6.06915280e-02 -3.62743855e-01 4.22764421e-01
4.73126054e-01 5.80072582e-01 4.19694752e-01 -4.01424229e-01
1.66826129e+00 -7.95455649e-02 -5.75822592e-01 3.52063328e-01
6.42321467e-01 -9.29906189e-01 6.51713908e-01 2.00395152e-01
-5.86480200e-01 1.37424558e-01 -8.95698428e-01 6.92900866e-02
-6.67084575e-01 -7.39685655e-01 3.32686961e-01 4.76018578e-01
-1.13840961e+00 7.28130937e-01 -1.18541968e+00 -9.26456869e-01
3.33510637e-01 6.90559372e-02 -4.88391340e-01 -6.30089566e-02
-9.42738414e-01 1.19457030e+00 5.16571343e-01 -3.13339710e-01
-2.14608997e-01 -9.31950033e-01 -4.05918717e-01 -1.31882206e-01
-8.82824212e-02 -1.10381150e+00 8.31182539e-01 9.30374190e-02
-8.51560116e-01 9.88400757e-01 -3.54790539e-01 5.89013733e-02
3.78285319e-01 -1.04164898e-01 -6.09990716e-01 1.85480863e-01
1.58405229e-01 1.93710223e-01 1.59140125e-01 -4.83389169e-01
-5.16562402e-01 -6.64481044e-01 -6.34570241e-01 -1.12244979e-01
5.29115200e-01 8.75265718e-01 -2.88910747e-01 -5.32606840e-01
-5.16285241e-01 -6.57033682e-01 -5.74728608e-01 1.23517320e-01
6.66274279e-02 -2.26765096e-01 1.07680774e+00 -1.22870040e+00
1.13100600e+00 -2.17013907e+00 -1.37973085e-01 1.59675121e-01
3.80628556e-01 3.95459414e-01 9.79149640e-02 1.29814780e+00
-2.48777986e-01 3.56847554e-01 -1.66268930e-01 5.73252559e-01
-2.51080871e-01 2.44618833e-01 2.04164907e-01 6.03958011e-01
2.30494931e-01 6.44472361e-01 -1.33165419e+00 -3.09345961e-01
2.16001838e-01 6.98134363e-01 -4.61787373e-01 2.67895669e-01
-1.43676683e-01 7.96906471e-01 -4.40044552e-01 4.14353460e-01
7.08324492e-01 -4.65090185e-01 7.26574302e-01 5.48900664e-01
-6.22988880e-01 4.59682554e-01 -5.92285037e-01 1.33493543e+00
2.67440170e-01 3.87306154e-01 3.91298831e-01 -1.47232980e-01
8.42814624e-01 4.16502863e-01 3.60996753e-01 -3.74089070e-02
-2.18937263e-01 2.44233198e-02 1.81998491e-01 -4.89921540e-01
1.05786450e-01 -1.09577052e-01 3.01403701e-01 7.19093263e-01
-3.66551727e-01 3.03076535e-01 3.28549623e-01 1.93548694e-01
1.21138513e+00 2.91146070e-01 8.37216616e-01 -3.48563880e-01
-6.79582283e-02 4.65531677e-01 6.16156518e-01 4.61154789e-01
-7.71989748e-02 -4.04916853e-02 3.23935598e-01 -7.38153458e-01
-1.31402040e+00 -1.23821640e+00 -5.35439610e-01 1.01349878e+00
-4.09806788e-01 -8.32195640e-01 -5.08085489e-01 -2.69488752e-01
-8.03424045e-03 4.37784284e-01 -4.65693831e-01 4.17850047e-01
-4.91694599e-01 -8.53510737e-01 4.34879035e-01 4.30757403e-02
-1.24438725e-01 -9.19829071e-01 -1.01349628e+00 3.12855214e-01
-2.41040990e-01 -2.99622983e-01 -4.06169206e-01 9.72011462e-02
-7.73131490e-01 -1.29697657e+00 -7.63851821e-01 -4.99515444e-01
2.83856720e-01 1.64691508e-01 8.59546542e-01 2.23404393e-01
-1.02752137e+00 -7.13273734e-02 -1.85670972e-01 -5.10695577e-01
-6.66098297e-01 -1.48934320e-01 1.32774441e-02 -6.88929379e-01
4.36221629e-01 -6.32254481e-01 -6.58173203e-01 5.16792893e-01
-6.60194218e-01 1.77586421e-01 5.01648225e-02 6.96301997e-01
6.30011559e-01 -3.17285568e-01 4.87100869e-01 -9.57561672e-01
4.49237853e-01 -8.66901517e-01 -5.91384292e-01 3.92447144e-01
-4.19022769e-01 -2.96863794e-01 5.11338770e-01 -1.69448871e-02
-8.09003651e-01 -2.57084757e-01 -6.54672861e-01 3.14834327e-01
-2.30117127e-01 4.54254538e-01 2.67201096e-01 3.21313202e-01
3.59380811e-01 1.42217176e-02 2.88622409e-01 -8.76680255e-01
2.71469861e-01 8.76385450e-01 4.45642769e-01 1.12130031e-01
3.75176996e-01 4.19762462e-01 -1.48827761e-01 -1.08522201e+00
8.23603570e-02 -8.79899025e-01 -6.59519136e-01 -2.18595490e-01
1.08260155e+00 -4.97855723e-01 -1.14251268e+00 5.28633058e-01
-8.12716842e-01 -1.55121654e-01 3.44810694e-01 2.48531312e-01
-2.29183391e-01 4.01077360e-01 -6.71574950e-01 -5.72765410e-01
-5.26265264e-01 -8.14026594e-01 8.09084117e-01 8.68592560e-02
-6.49248302e-01 -9.49197829e-01 7.86996663e-01 -2.82652839e-03
4.39928025e-01 8.24758112e-01 1.35228884e+00 -6.78667426e-01
-3.15244257e-01 -3.34187113e-02 1.28805377e-02 -5.82641482e-01
3.70709538e-01 5.60656846e-01 -5.74164510e-01 -3.46749038e-01
-3.12277198e-01 1.84653997e-01 4.43487048e-01 2.65380442e-01
4.98367369e-01 -2.34746188e-01 -7.77987778e-01 7.30541170e-01
1.18874788e+00 5.96631706e-01 3.88363481e-01 4.06050771e-01
1.67732596e-01 6.92145288e-01 4.67257380e-01 5.46530843e-01
-3.77629437e-02 7.53290057e-01 1.13734920e-02 -1.88888177e-01
5.01015067e-01 -3.27979505e-01 -1.70768082e-01 6.87736928e-01
-6.52558655e-02 2.76359860e-02 -1.14785659e+00 8.59152153e-02
-1.65118301e+00 -1.30955875e+00 -3.33585471e-01 2.15326929e+00
7.44670212e-01 -2.64057815e-01 6.01866066e-01 -3.79840612e-01
8.30406547e-01 -1.95750240e-02 -7.64467239e-01 -8.07740211e-01
-2.08486915e-02 1.00949928e-01 3.20830911e-01 2.93775767e-01
-6.84355021e-01 4.72203493e-01 8.20167923e+00 5.48129022e-01
-1.19513106e+00 4.86360490e-02 2.65262991e-01 -1.64027773e-02
-2.76649356e-01 -1.30044624e-01 -4.29650635e-01 6.12227440e-01
1.09533298e+00 -5.30036032e-01 3.34658504e-01 5.39701164e-01
5.42712867e-01 -6.19356893e-02 -6.40880704e-01 6.38602853e-01
-6.76215112e-01 -1.74769247e+00 -3.31953406e-01 3.98934811e-01
-2.98351217e-02 5.63855052e-01 -3.21020186e-01 -4.57037985e-01
9.54138339e-01 -8.58092487e-01 -3.63385044e-02 2.89836526e-01
1.11403084e+00 -9.36567247e-01 5.75178325e-01 1.90802976e-01
-1.04440558e+00 5.05750418e-01 -3.55865210e-01 1.16417222e-01
8.52234840e-01 4.48845744e-01 -1.27661383e+00 6.99309051e-01
9.72229600e-01 5.37938893e-01 -3.40062559e-01 1.38988209e+00
2.28496924e-01 4.34035301e-01 -4.22833741e-01 -2.08837464e-01
-6.74143806e-02 -3.55080396e-01 7.01083839e-01 1.60760081e+00
2.45744780e-01 1.23524778e-01 2.59344041e-01 3.71691734e-01
7.59660900e-01 2.37950608e-01 -7.96203494e-01 -2.38209903e-01
6.39678121e-01 1.01167381e+00 -9.78700340e-01 -3.17015558e-01
-6.33762255e-02 7.32357085e-01 5.36359325e-02 4.59042788e-02
-4.13716346e-01 -5.93804181e-01 1.20980978e+00 1.14094280e-01
5.52667499e-01 -2.82538831e-01 2.10934669e-01 -5.39392114e-01
-5.14993727e-01 -9.98927295e-01 9.40936267e-01 -8.78284574e-01
-1.18675792e+00 5.27676344e-01 2.02216178e-01 -6.76862955e-01
-5.40350437e-01 -5.02446055e-01 -6.08103395e-01 1.18437946e+00
-6.87351823e-01 -8.25872481e-01 3.72201324e-01 3.72748822e-01
1.37062535e-01 4.19625640e-01 1.10409844e+00 2.21238825e-02
-6.04159236e-01 2.74520572e-02 8.25702846e-01 -3.18503350e-01
5.09048641e-01 -1.05737543e+00 1.07924891e+00 3.26417059e-01
-4.31398898e-01 1.19143260e+00 8.50674331e-01 -1.19568682e+00
-1.20064175e+00 -7.03536212e-01 8.60435963e-01 -4.84424353e-01
8.78210247e-01 -5.52894473e-01 -1.12684941e+00 4.53221053e-01
3.78990561e-01 -7.90325403e-01 1.26167417e+00 2.76055545e-01
-5.68509519e-01 6.46249115e-01 -1.36836994e+00 8.21507633e-01
8.70532036e-01 -3.20330590e-01 -6.12762511e-01 4.45625395e-01
5.34295857e-01 -1.90749899e-01 -1.02696753e+00 1.28124475e-01
1.15827608e+00 -9.84351099e-01 9.19423640e-01 -8.75190318e-01
-2.23939702e-01 -4.05274987e-01 2.63168454e-01 -1.18415821e+00
-5.98568261e-01 -1.06235802e+00 6.41415596e-01 9.34265673e-01
2.43584946e-01 -9.94924247e-01 2.84621835e-01 -7.67645091e-02
7.77149051e-02 -4.19093013e-01 -9.29283857e-01 -6.58072591e-01
-2.90807515e-01 1.78017139e-01 8.73221576e-01 1.00852895e+00
3.64743859e-01 2.02501655e-01 -3.88397902e-01 1.77995563e-02
4.00813371e-01 -5.22611625e-02 5.80895424e-01 -1.47855985e+00
-5.31756997e-01 -4.07463104e-01 -2.60734022e-01 -3.76786143e-01
-9.43408847e-01 -6.79552555e-01 -2.39936367e-01 -1.44729352e+00
3.08351129e-01 -4.00928527e-01 -6.38549402e-02 2.98539758e-01
-4.78932023e-01 3.18653360e-02 1.62238143e-02 4.15621996e-01
-9.04818326e-02 -2.90979475e-01 8.70676339e-01 5.55536568e-01
-1.29187293e-02 -2.31273353e-01 -2.77618915e-01 5.15787959e-01
8.65891814e-01 -6.46124721e-01 -4.23492566e-02 1.07940897e-01
2.18880817e-01 -8.09862614e-02 -3.18380818e-02 -2.74900168e-01
-3.44765335e-01 -4.68012333e-01 4.32213217e-01 -1.07384408e+00
1.84254408e-01 -3.85168046e-01 1.19030619e+00 1.18237579e+00
1.90873981e-01 9.97515202e-01 4.89615738e-01 3.85001987e-01
5.11848807e-01 5.68375252e-02 7.23873019e-01 -1.94521144e-01
-5.17508149e-01 1.88962519e-02 -9.12255049e-01 -1.01986758e-01
1.05280674e+00 -4.83465791e-01 -8.76389086e-01 -8.19294900e-02
-7.12824106e-01 3.23129475e-01 1.05777991e+00 2.76220739e-01
2.12504134e-01 -5.90180695e-01 -7.34942436e-01 4.14625347e-01
1.49929687e-01 -6.74785495e-01 5.42460740e-01 8.31193089e-01
-1.20757556e+00 4.55915153e-01 -7.01352119e-01 -7.77685761e-01
-1.92811036e+00 1.04910398e+00 -8.44039023e-02 -2.12016746e-01
-1.01094568e+00 4.70558017e-01 1.53738394e-01 -5.48117697e-01
-1.40921488e-01 4.53546882e-01 -6.53551295e-02 -1.17044657e-01
1.11727464e+00 4.47120488e-01 -2.18461692e-01 -4.97242004e-01
-8.53236556e-01 2.60270894e-01 -2.39327848e-02 -7.56622404e-02
1.73009372e+00 -1.76765993e-01 -4.54377353e-01 6.13836050e-01
7.74046063e-01 2.45552629e-01 -7.00895488e-01 3.11613292e-01
3.23125660e-01 -5.18673182e-01 -8.61207426e-01 -1.12024665e+00
-3.77186574e-02 7.23771095e-01 5.58334410e-01 2.71833837e-01
9.25608993e-01 -4.58889566e-02 5.50422370e-01 -1.21423826e-01
4.44385767e-01 -4.18910176e-01 -8.01782906e-01 1.39386371e-01
5.38650215e-01 -3.32993001e-01 -1.14725135e-01 -5.66991746e-01
-3.32855076e-01 7.98774421e-01 -8.23953822e-02 4.68445361e-01
2.34325036e-01 6.08614683e-01 4.45994675e-01 -5.59248447e-01
-1.27383363e+00 1.09189503e-01 -1.91599295e-01 1.11891294e+00
4.63290662e-01 1.27512291e-01 -5.63259304e-01 -3.84095341e-01
-8.81509259e-02 6.36326969e-02 7.29646608e-02 1.07615662e+00
-6.27072990e-01 -1.51418293e+00 -6.02537572e-01 5.59844553e-01
-6.55092716e-01 -2.97410429e-01 -6.64204717e-01 8.48189473e-01
-5.06141074e-02 6.63333237e-01 2.54431099e-01 2.05093306e-02
-9.09086913e-02 4.62189674e-01 1.83897108e-01 -4.61491138e-01
-1.03194129e+00 4.46734250e-01 4.81309026e-01 -4.20731515e-01
9.88670662e-02 -7.45881796e-01 -8.75285804e-01 -8.67396176e-01
-1.63659528e-02 4.83161032e-01 7.17168570e-01 5.17119646e-01
8.48579764e-01 1.56621441e-01 1.43944845e-01 -3.49022001e-01
-8.99591669e-02 -7.60781944e-01 -4.86994624e-01 2.15730235e-01
2.84365118e-01 -3.10970038e-01 2.14896575e-01 -1.67129442e-01] | [4.935035228729248, 5.168646812438965] |
4a0f30c4-ea8f-4325-b6e0-4c4dd29a37fa | clear-causal-explanations-from-attention-in | 2210.10621 | null | https://arxiv.org/abs/2210.10621v1 | https://arxiv.org/pdf/2210.10621v1.pdf | CLEAR: Causal Explanations from Attention in Neural Recommenders | We present CLEAR, a method for learning session-specific causal graphs, in the possible presence of latent confounders, from attention in pre-trained attention-based recommenders. These causal graphs describe user behavior, within the context captured by attention, and can provide a counterfactual explanation for a recommendation. In essence, these causal graphs allow answering "why" questions uniquely for any specific session. Using empirical evaluations we show that, compared to naively using attention weights to explain input-output relations, counterfactual explanations found by CLEAR are shorter and an alternative recommendation is ranked higher in the original top-k recommendations. | ['Gal Novik', 'Guy Koren', 'Yaniv Gurwicz', 'Raanan Y. Rohekar', 'Shami Nisimov'] | 2022-10-07 | null | null | null | null | ['counterfactual-explanation'] | ['miscellaneous'] | [ 2.56792039e-01 8.08833957e-01 -7.21678317e-01 -3.68824065e-01
-4.54697967e-01 -5.54825425e-01 5.89470744e-01 -3.92682105e-02
2.80969590e-01 1.15316653e+00 1.33194637e+00 -6.81908309e-01
-8.81341994e-01 -6.47814214e-01 -1.16726923e+00 -2.26021186e-01
-3.43065053e-01 2.85751820e-01 -2.64923483e-01 -2.60580480e-02
4.70597535e-01 1.62911892e-03 -1.54964566e+00 7.21060753e-01
8.21245134e-01 2.07213998e-01 -6.21812604e-02 7.66863883e-01
-8.89739245e-02 9.46834624e-01 -4.45666820e-01 -7.76530862e-01
-1.18298665e-01 -6.18174255e-01 -9.73780811e-01 -3.36682886e-01
9.27037001e-01 -2.24399045e-01 -6.07163846e-01 4.70790088e-01
-9.43405181e-02 5.77833414e-01 6.10277951e-01 -1.40648317e+00
-1.51978219e+00 1.50713325e+00 -1.41604707e-01 6.13927066e-01
7.51758933e-01 1.76289827e-02 1.64837015e+00 -6.75563574e-01
4.99737114e-01 1.55803466e+00 5.43590069e-01 7.87969351e-01
-1.72807515e+00 -7.85135210e-01 9.31456804e-01 2.56231874e-01
-4.67148185e-01 -2.15795100e-01 4.07486916e-01 -3.38853598e-01
9.78854954e-01 9.20069456e-01 6.23086452e-01 1.51917624e+00
3.85507911e-01 4.74518597e-01 6.91998661e-01 -9.07472372e-02
7.19638243e-02 -9.74894781e-03 6.48079872e-01 3.75092417e-01
7.10199654e-01 5.65278590e-01 -7.21842647e-01 -5.90420187e-01
5.33890605e-01 2.70316958e-01 -6.23962760e-01 -3.53809655e-01
-1.09149528e+00 1.02282131e+00 7.37910211e-01 1.26949951e-01
-7.71807194e-01 6.22401237e-01 -2.09899172e-01 2.71202326e-01
5.33450663e-01 1.22857475e+00 -8.94788206e-01 2.43719906e-01
-3.85389268e-01 4.29002285e-01 9.56549227e-01 7.29863405e-01
4.52713698e-01 -1.28867760e-01 -6.59170330e-01 1.74964055e-01
3.29046309e-01 4.75662112e-01 4.48121190e-01 -8.84872973e-01
1.56869888e-01 4.75494683e-01 3.85415107e-01 -6.90646589e-01
-4.18227226e-01 -5.36029756e-01 -3.79276246e-01 -3.70003581e-01
3.98611218e-01 -3.93926650e-01 -7.67439008e-01 1.74229205e+00
9.99749824e-02 7.20337152e-01 -3.76224481e-02 9.10606146e-01
8.30512941e-01 3.41512918e-01 5.23940444e-01 -5.51134884e-01
1.07147908e+00 -7.91343749e-01 -9.48561013e-01 -1.23553135e-01
4.37783808e-01 -3.45021814e-01 1.32689703e+00 1.45042151e-01
-6.85004771e-01 -4.47472900e-01 -7.18373418e-01 1.77550390e-01
-2.20181331e-01 -4.66500908e-01 1.09268379e+00 6.06073439e-01
-9.71004009e-01 9.26643789e-01 3.33249494e-02 -3.64996940e-01
1.10831544e-01 4.90065396e-01 6.00412264e-02 2.06943695e-02
-1.65519524e+00 6.30589545e-01 -6.16284423e-02 -1.50228456e-01
-9.18008268e-01 -1.52127004e+00 -4.72982109e-01 5.26792586e-01
6.43035293e-01 -1.04106033e+00 1.42265761e+00 -7.63367236e-01
-1.18349266e+00 -5.30766882e-02 -2.91665822e-01 -5.10249019e-01
-8.30803588e-02 -5.04620492e-01 -6.82256341e-01 -5.94473779e-01
-3.63851562e-02 4.36026193e-02 6.49224401e-01 -1.30650532e+00
-7.18619287e-01 -3.83461386e-01 5.40555775e-01 1.58369482e-01
1.26729071e-01 -2.77811140e-01 -4.26339731e-02 -3.95149946e-01
-4.79997337e-01 -1.05672765e+00 -5.78654051e-01 -7.08052039e-01
-5.71072280e-01 -5.22250414e-01 5.52293837e-01 -3.62567753e-01
1.48551464e+00 -1.69015360e+00 3.23179774e-02 2.35300437e-01
4.50983882e-01 -1.47600546e-01 -1.34782046e-01 2.56027550e-01
-6.60081565e-01 8.03917229e-01 5.07855415e-01 2.26048857e-01
-7.45556951e-02 1.74756512e-01 -9.29993153e-01 1.17949247e-01
-1.87052667e-01 1.11307681e+00 -1.22411561e+00 2.28847295e-01
1.17246479e-01 8.87204260e-02 -9.07671750e-01 3.41705173e-01
-5.32939434e-01 1.29177660e-01 -4.13895577e-01 -3.03493112e-01
1.47142768e-01 -5.22235572e-01 6.39043808e-01 -1.68612823e-02
-1.45914227e-01 8.01890373e-01 -7.74674296e-01 1.19178712e+00
-7.18507528e-01 5.81886351e-01 -6.80878282e-01 -4.47043270e-01
2.77622759e-01 6.12132132e-01 1.79412607e-02 -4.96857792e-01
-1.10325605e-01 -3.51113081e-01 7.21228421e-02 -4.67891902e-01
4.72258031e-01 -3.25800329e-02 -5.35358638e-02 8.19815695e-01
1.56813189e-01 6.02475584e-01 -1.08389882e-02 7.62146771e-01
1.00603247e+00 4.39248569e-02 3.87947202e-01 -3.36742520e-01
3.52036059e-02 -6.65637925e-02 3.49694461e-01 1.40915811e+00
3.66796374e-01 2.70328671e-01 7.13479936e-01 -5.18023312e-01
-6.59346879e-01 -9.59124982e-01 2.83518255e-01 1.52043724e+00
-1.34465378e-02 -4.90077883e-01 -3.06836665e-01 -1.16618776e+00
4.17703211e-01 1.66973448e+00 -1.43846643e+00 -3.03390563e-01
-1.24287076e-01 -5.16923547e-01 -4.52777036e-02 4.68813866e-01
-3.43876392e-01 -1.20813406e+00 3.45463529e-02 2.43162900e-01
-1.97344184e-01 -3.42166930e-01 -9.85132575e-01 -8.42151493e-02
-1.01053834e+00 -1.53511453e+00 -3.76028508e-01 6.13700189e-02
4.39682871e-01 6.36439800e-01 1.42999029e+00 3.74810666e-01
1.52607948e-01 5.55947125e-01 3.37199569e-02 -3.76581103e-01
-2.91590113e-02 -2.95309514e-01 2.50244826e-01 -1.32984772e-01
3.64845842e-01 -6.59149945e-01 -7.15688586e-01 1.08884595e-01
-2.61896312e-01 -3.27784382e-02 5.32798350e-01 7.03086436e-01
1.69429079e-01 -5.86697161e-01 8.51806819e-01 -1.89592874e+00
1.04727244e+00 -1.09076405e+00 -2.93396115e-01 4.85346317e-01
-1.33575583e+00 4.87809688e-01 7.53073156e-01 -5.90007722e-01
-1.55687392e+00 -5.58275998e-01 3.89390737e-01 -1.75161690e-01
-3.09607118e-01 6.21992230e-01 -4.64297459e-02 6.22474730e-01
1.21018207e+00 -3.71468157e-01 -4.61549908e-01 -5.47388375e-01
1.14185941e+00 5.87275997e-02 3.48652065e-01 -2.65284836e-01
6.12572253e-01 1.04693264e-01 -3.27388674e-01 -1.43960595e-01
-1.34339023e+00 -5.07410109e-01 -1.49191499e-01 -2.04317614e-01
7.84315586e-01 -5.98326325e-01 -1.17660439e+00 -8.16319942e-01
-1.07396364e+00 -2.95834482e-01 -3.78406137e-01 6.72059000e-01
-3.96770269e-01 -3.53146791e-01 -1.06111825e-01 -9.36271250e-01
-1.88174441e-01 -3.90296578e-01 3.63163650e-01 3.61793578e-01
-7.83814549e-01 -1.31347001e+00 2.83378601e-01 3.40891212e-01
2.05603421e-01 -2.23967001e-01 1.14698625e+00 -9.25405264e-01
-5.37174702e-01 4.34827246e-03 -2.47604340e-01 -6.21397555e-01
2.56822288e-01 4.66183387e-02 -8.86466920e-01 1.79608494e-01
-5.55647314e-01 5.47245502e-01 7.30483770e-01 6.23104870e-01
1.16203237e+00 -1.21530616e+00 -7.82856584e-01 8.01233575e-03
1.16401386e+00 2.62337029e-01 4.92660820e-01 -3.13693851e-01
8.11222196e-01 5.17435312e-01 2.47547016e-01 1.87116325e-01
2.74432659e-01 5.26677251e-01 2.67255723e-01 -1.44216344e-01
-3.11378896e-01 -8.28743935e-01 2.86397278e-01 -7.84416217e-04
-2.96658009e-01 -4.96661991e-01 -3.36670280e-01 6.87295675e-01
-2.16020608e+00 -1.46411216e+00 -8.78560901e-01 2.31621671e+00
5.23961127e-01 3.93530875e-02 1.41993612e-01 -4.45624262e-01
7.93724358e-01 -2.20383301e-01 -7.14554191e-01 -6.77891552e-01
2.77630836e-01 1.89441442e-01 6.65519178e-01 8.44830573e-01
-5.23345351e-01 4.87380326e-01 7.69411898e+00 1.74793869e-01
-3.88787240e-01 6.22057430e-02 3.93676162e-01 -4.89922315e-01
-1.35039186e+00 4.04043406e-01 -3.78771842e-01 2.54141539e-01
1.39155102e+00 -7.20734417e-01 5.68118870e-01 7.34613895e-01
5.82011819e-01 2.33977020e-01 -1.42581928e+00 1.12254150e-01
-2.80661166e-01 -1.77576697e+00 4.71088648e-01 2.55446523e-01
1.16526532e+00 -3.15147847e-01 3.08099359e-01 5.07093787e-01
1.24328148e+00 -9.18606460e-01 3.11917931e-01 9.08316791e-01
5.13503790e-01 -6.50209546e-01 5.36502779e-01 2.03976750e-01
-5.56607664e-01 -4.90901679e-01 -3.60140055e-01 -4.06367123e-01
-1.18479565e-01 6.25634432e-01 -1.09630847e+00 6.00355566e-01
6.44479871e-01 5.02089798e-01 -2.35793144e-01 8.26200306e-01
-7.79067278e-01 1.36362863e+00 4.35659498e-01 -2.92118043e-01
-1.61866084e-01 2.49844044e-01 5.61968565e-01 9.81157243e-01
8.51211697e-02 6.37872934e-01 -4.27713931e-01 8.23705316e-01
-3.57733935e-01 1.11382954e-01 -8.36952984e-01 9.00008753e-02
3.12423021e-01 8.42261612e-01 -1.40406549e-01 -5.24245024e-01
-3.48337322e-01 5.68538427e-01 3.20214719e-01 7.59792387e-01
-7.25891829e-01 2.45761171e-01 8.66087198e-01 3.05639833e-01
1.00970961e-01 5.98635852e-01 -4.68452632e-01 -1.14956033e+00
-8.98919582e-01 -6.76173508e-01 8.43255103e-01 -1.20205009e+00
-1.46562183e+00 -3.99251375e-03 -8.69321811e-05 -4.88814771e-01
-2.24679559e-01 -4.69759181e-02 -1.10475123e+00 9.85028684e-01
-1.04846907e+00 -6.38322413e-01 -4.63624634e-02 3.29978079e-01
6.63704753e-01 2.32968360e-01 1.12157738e+00 -1.41740382e-01
-4.38421011e-01 5.77639043e-01 -3.93998027e-02 -5.03806770e-01
5.95494330e-01 -1.58175313e+00 4.81634229e-01 7.57949948e-01
5.15792847e-01 1.39780211e+00 1.22732699e+00 -9.82494473e-01
-1.16710687e+00 -1.01284337e+00 1.19704390e+00 -9.53056872e-01
6.62507951e-01 -3.03674582e-02 -1.01823354e+00 1.10303402e+00
4.69199628e-01 -6.70777857e-01 1.27313447e+00 1.49825001e+00
-5.52029669e-01 2.04722479e-01 -8.21238816e-01 1.02925313e+00
1.51233077e+00 -2.46613920e-01 -9.72353697e-01 3.68112504e-01
1.28065503e+00 1.81615259e-03 -4.65795100e-01 -1.23541027e-01
7.64089346e-01 -8.24548602e-01 8.75228286e-01 -1.71266043e+00
7.39776313e-01 -8.49498261e-04 3.87497187e-01 -1.77396429e+00
-1.10255766e+00 -9.10795033e-01 -5.27582645e-01 7.63441026e-01
1.01448870e+00 -3.92565012e-01 7.05357432e-01 1.09090042e+00
-1.33539885e-01 -3.31983358e-01 -4.56301033e-01 -2.34521538e-01
-2.32249796e-01 -4.55592155e-01 1.10143459e+00 1.13684905e+00
4.10877496e-01 7.97476172e-01 -6.79095209e-01 3.79689068e-01
6.76927269e-01 4.28769112e-01 5.96438944e-01 -1.51210761e+00
-3.60700816e-01 -2.96625048e-01 3.94323796e-01 -8.26177478e-01
1.19546771e-01 -8.17297339e-01 -1.94600999e-01 -1.81136453e+00
3.36996824e-01 -1.54544577e-01 -6.74372196e-01 4.69894499e-01
-8.25629056e-01 2.51207203e-02 -1.27170123e-02 5.86462654e-02
-5.41525960e-01 1.89300507e-01 1.30270338e+00 3.68848704e-02
-3.12467337e-01 3.02174777e-01 -1.57698655e+00 5.04724085e-01
6.25905275e-01 -8.02363515e-01 -7.53961921e-01 6.78015873e-02
6.42276168e-01 3.19747567e-01 3.40187192e-01 -1.70392379e-01
3.74051705e-02 -6.73733294e-01 3.69825125e-01 -1.85517237e-01
-3.16306710e-01 -5.61637819e-01 5.13155282e-01 2.55055636e-01
-1.31314409e+00 -2.46658549e-01 1.32342532e-01 1.28094947e+00
5.33511817e-01 6.33894429e-02 -1.62176222e-01 -5.62811382e-02
-5.25271416e-01 3.34697753e-01 -6.02994084e-01 -1.46409377e-01
5.61617911e-01 1.25297357e-03 -4.46322948e-01 -8.75538170e-01
-1.04335129e+00 1.96263909e-01 -2.14437082e-01 8.69682491e-01
4.44686711e-01 -1.38347280e+00 -5.10457695e-01 -3.51769954e-01
2.07199976e-01 -9.19889808e-01 5.67004919e-01 3.83710027e-01
6.55153036e-01 8.52336407e-01 1.69391826e-01 4.21602428e-01
-1.03378415e+00 1.12305033e+00 2.02261463e-01 -3.30725402e-01
-3.17535400e-01 8.53588998e-01 6.27360046e-01 -6.50133640e-02
-1.64753079e-01 -3.45198393e-01 -6.62794948e-01 -1.07951231e-01
6.54208302e-01 3.59339952e-01 -2.33610153e-01 1.53864995e-01
-1.84601083e-01 -2.95958072e-01 -4.83051538e-02 9.57015250e-03
1.23512471e+00 -2.12418824e-01 1.23740844e-01 5.34216344e-01
6.24937177e-01 4.38822567e-01 -1.01811492e+00 -9.65806097e-02
1.51301935e-01 -6.46185100e-01 7.60592520e-02 -1.35657465e+00
-8.25594485e-01 3.97620618e-01 2.09215760e-01 7.97557652e-01
4.52462465e-01 1.28716961e-01 8.69635418e-02 1.82046369e-01
-1.17074080e-01 -4.55440164e-01 -1.91913947e-01 1.11905046e-01
1.23670173e+00 -7.89961576e-01 7.04567209e-02 -4.74189818e-01
-5.58556020e-01 8.25254321e-01 7.49916255e-01 -1.81523979e-01
7.10024178e-01 -2.53389984e-01 -9.03479010e-02 -4.25499022e-01
-1.52806699e+00 -9.85306427e-02 9.37810302e-01 7.25769103e-01
9.25947130e-01 3.70482802e-01 -4.54598635e-01 1.09564257e+00
-1.58453792e-01 -2.86239572e-02 9.06796932e-01 -1.19200990e-01
-1.22789793e-01 -6.32577121e-01 -2.30638266e-01 1.06892264e+00
-5.18010259e-01 -3.72579604e-01 -5.57014048e-01 6.23410821e-01
-2.44546495e-02 1.39340258e+00 7.06419721e-02 -5.45014024e-01
6.93045795e-01 2.10229725e-01 2.93815464e-01 -9.96475160e-01
-8.18518460e-01 -3.97437871e-01 4.69497353e-01 -8.02668631e-01
-1.70915097e-01 -5.53213596e-01 -1.11260462e+00 -5.60142934e-01
-6.27028704e-01 6.59285724e-01 1.84239417e-01 1.14724052e+00
6.24928832e-01 1.00259340e+00 5.60724199e-01 -2.14267060e-01
-2.88954765e-01 -1.03390670e+00 -2.77865887e-01 6.72821164e-01
3.78943861e-01 -8.23220849e-01 -6.23338521e-01 -3.44816372e-02] | [9.424358367919922, 5.699220180511475] |
c9d129c8-485d-4395-b8a6-b15c0f4bb6f0 | lightness-modulated-deep-inverse-tone-mapping | 2107.07907 | null | https://arxiv.org/abs/2107.07907v1 | https://arxiv.org/pdf/2107.07907v1.pdf | Lightness Modulated Deep Inverse Tone Mapping | Single-image HDR reconstruction or inverse tone mapping (iTM) is a challenging task. In particular, recovering information in over-exposed regions is extremely difficult because details in such regions are almost completely lost. In this paper, we present a deep learning based iTM method that takes advantage of the feature extraction and mapping power of deep convolutional neural networks (CNNs) and uses a lightness prior to modulate the CNN to better exploit observations in the surrounding areas of the over-exposed regions to enhance the quality of HDR image reconstruction. Specifically, we introduce a Hierarchical Synthesis Network (HiSN) for inferring a HDR image from a LDR input and a Lightness Adpative Modulation Network (LAMN) to incorporate the the lightness prior knowledge in the inferring process. The HiSN hierarchically synthesizes the high-brightness component and the low-brightness component of the HDR image whilst the LAMN uses a lightness adaptive mask that separates detail-less saturated bright pixels from well-exposed lower light pixels to enable HiSN to better infer the missing information, particularly in the difficult over-exposed detail-less areas. We present experimental results to demonstrate the effectiveness of the new technique based on quantitative measures and visual comparisons. In addition, we present ablation studies of HiSN and visualization of the activation maps inside LAMN to help gain a deeper understanding of the internal working of the new iTM algorithm and explain why it can achieve much improved performance over state-of-the-art algorithms. | ['Guoping Qiu', 'Jiang Duan', 'Gaofeng Cao', 'Kanglin Liu'] | 2021-07-16 | null | null | null | null | ['hdr-reconstruction', 'tone-mapping', 'inverse-tone-mapping'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 6.48468494e-01 1.30716309e-01 1.91256776e-01 -2.87040561e-01
-4.36154157e-01 -2.22276852e-01 3.46184134e-01 -4.80048954e-01
-8.65992084e-02 6.72287881e-01 8.73821601e-02 -1.31166011e-01
1.41013637e-01 -9.76764858e-01 -7.78552949e-01 -9.29044545e-01
1.80026397e-01 -6.50068223e-02 3.13865542e-01 -4.55702424e-01
1.37383118e-01 6.52659774e-01 -1.74232447e+00 4.64017063e-01
9.13452983e-01 1.01862037e+00 6.61726236e-01 5.31578064e-01
-2.49252878e-02 8.16861033e-01 -4.80834752e-01 7.88162127e-02
5.65497398e-01 -6.04912221e-01 -5.28024793e-01 4.15153801e-02
5.56736231e-01 -6.59173787e-01 -6.45588994e-01 1.00572121e+00
4.14724141e-01 1.51593328e-01 3.63669813e-01 -7.93853879e-01
-7.25843370e-01 1.05036840e-01 -6.86076224e-01 2.08572343e-01
6.46580458e-02 4.01175708e-01 6.32364333e-01 -8.62594366e-01
4.18916166e-01 1.04104745e+00 3.37848723e-01 3.55862051e-01
-1.46969962e+00 -7.54752576e-01 -1.49534196e-01 2.51652688e-01
-1.30824447e+00 -4.34721082e-01 1.12470508e+00 -1.20961025e-01
7.34401464e-01 3.55658293e-01 6.75361931e-01 5.48542559e-01
1.50456324e-01 5.05991995e-01 1.72501707e+00 -4.34461981e-01
4.02664654e-02 1.72588080e-01 -3.79576772e-01 6.99546576e-01
-3.89173557e-03 4.50958550e-01 -6.09904170e-01 4.16875243e-01
1.30330372e+00 7.12006772e-03 -6.61757290e-01 7.43734613e-02
-9.62230444e-01 5.57677031e-01 8.43803763e-01 2.91994095e-01
-3.53670865e-01 -9.38865319e-02 -3.19283992e-01 1.53081432e-01
4.75730777e-01 5.98921239e-01 -1.28556073e-01 3.02967727e-01
-1.20196378e+00 -2.32813418e-01 2.76513696e-01 4.66331750e-01
1.27925241e+00 1.78549841e-01 -8.49693120e-02 1.00055838e+00
9.28453133e-02 3.30536276e-01 -2.74212062e-02 -1.30084956e+00
2.24197686e-01 6.11524880e-01 -3.22494619e-02 -7.19244957e-01
-2.93709934e-01 -3.50619316e-01 -1.07896030e+00 8.19636643e-01
3.03588480e-01 1.08286202e-01 -1.04322183e+00 1.57266140e+00
1.05272472e-01 -8.75271566e-04 -1.46332845e-01 1.36642122e+00
6.05889022e-01 8.47938359e-01 -2.85763979e-01 -1.34434998e-01
1.25999129e+00 -7.34781623e-01 -7.50826001e-01 -3.97446573e-01
-2.70405374e-02 -6.78934157e-01 1.34640539e+00 2.74898112e-01
-1.34237254e+00 -8.49094629e-01 -1.35755301e+00 -5.49811065e-01
-2.32602358e-01 1.98561266e-01 3.12198907e-01 1.29463926e-01
-1.17252088e+00 6.59488022e-01 -4.43637043e-01 -4.45951074e-02
5.54504871e-01 1.95589945e-01 -2.60230690e-01 -4.88195986e-01
-1.42349207e+00 9.63568687e-01 2.67035306e-01 4.09310549e-01
-8.73098016e-01 -9.47880626e-01 -7.64466465e-01 1.12761520e-01
2.19178751e-01 -6.56179190e-01 7.31044352e-01 -1.11438346e+00
-1.67719638e+00 8.94943595e-01 -2.88501140e-02 -1.47281691e-01
2.24796757e-01 5.77436648e-02 -2.63039291e-01 4.96630639e-01
-2.08638877e-01 7.05561399e-01 1.00522387e+00 -1.41443229e+00
-3.77827972e-01 -3.25424254e-01 2.41650362e-02 3.53767812e-01
-1.21568562e-02 -3.40560913e-01 -3.35180461e-01 -7.88753510e-01
1.85933173e-01 -6.17819130e-01 2.08964288e-01 4.02105153e-01
-3.67760301e-01 4.99708325e-01 9.89491880e-01 -9.59035635e-01
1.05977118e+00 -2.29899120e+00 2.14449495e-01 3.97477746e-02
3.82911056e-01 3.65223467e-01 -2.26849601e-01 2.10892692e-01
-1.99412078e-01 -2.71012634e-01 -5.58717787e-01 -1.11861147e-01
-3.40475053e-01 1.40715212e-01 -3.62160802e-01 4.98715430e-01
3.91663611e-01 8.10809374e-01 -5.34019351e-01 -2.31143534e-01
6.79406881e-01 8.37339401e-01 -2.69227743e-01 6.11368299e-01
-2.59136379e-01 7.15902269e-01 1.21793523e-01 5.88289797e-01
8.91662300e-01 -1.49890348e-01 6.94981217e-02 -7.29668856e-01
-4.49990541e-01 1.03362978e-01 -9.48751330e-01 1.48154914e+00
-8.23186755e-01 1.02331865e+00 3.53522360e-01 -7.02863932e-01
1.10523748e+00 -4.83700074e-02 2.35644758e-01 -1.32630599e+00
-1.47217568e-02 1.62662596e-01 -6.60937130e-02 -5.25304496e-01
4.05370861e-01 -4.05493349e-01 4.18453395e-01 3.10058713e-01
-2.45986864e-01 -2.67933518e-01 -3.20469409e-01 4.48035635e-03
7.82985985e-01 4.65435758e-02 1.93857819e-01 -1.62273362e-01
4.37792331e-01 -4.07081485e-01 4.14350092e-01 5.23460448e-01
-2.05784701e-02 1.02676570e+00 3.52708876e-01 -5.99241853e-01
-1.22208965e+00 -1.16739285e+00 -1.82966247e-01 8.13686252e-01
3.51165980e-01 1.14969825e-02 -4.94887739e-01 -2.72495031e-01
-3.19553435e-01 7.67852604e-01 -7.18820274e-01 -2.35055700e-01
-7.29412019e-01 -7.09473848e-01 1.40651688e-01 2.69681126e-01
1.15263700e+00 -1.32198906e+00 -6.68105006e-01 -6.22934401e-02
-3.70526552e-01 -1.13268030e+00 -2.27299601e-01 2.67369747e-01
-5.56557596e-01 -8.33635449e-01 -7.71441102e-01 -6.18658960e-01
6.11384630e-01 4.09361959e-01 1.17303419e+00 1.35992125e-01
-5.39955378e-01 -3.53286654e-01 -1.47798136e-01 -1.21862553e-02
-3.43509525e-01 -2.13814721e-01 -5.79536378e-01 1.57852650e-01
-1.43105507e-01 -6.89638019e-01 -9.27444935e-01 3.73900235e-01
-1.15963757e+00 6.87403321e-01 9.09762859e-01 7.64365315e-01
5.84102690e-01 4.74591613e-01 1.94012284e-01 -8.26666594e-01
-1.26042692e-02 -1.58540815e-01 -6.15414917e-01 4.65353392e-02
-4.68958229e-01 5.55633679e-02 5.47459245e-01 -3.10965419e-01
-1.36444187e+00 6.59565255e-03 -1.18891411e-01 -4.70795989e-01
-1.23625875e-01 -1.08061448e-01 -4.97150421e-01 -2.91321754e-01
5.55951297e-01 4.86221075e-01 9.10385549e-02 -4.12506878e-01
3.06959003e-01 5.00043690e-01 7.27321625e-01 -2.75542170e-01
9.14369464e-01 8.75998318e-01 7.38523230e-02 -9.45443809e-01
-9.39609766e-01 -7.23976791e-02 -4.20833498e-01 -3.57062161e-01
9.43859339e-01 -9.12538230e-01 -4.65966493e-01 6.18322432e-01
-8.07976067e-01 -9.94005263e-01 -3.85485470e-01 1.38093084e-01
-5.13925254e-01 3.99724878e-02 -7.62404144e-01 -6.19108915e-01
-2.45718375e-01 -1.18362665e+00 1.12793732e+00 5.69784760e-01
1.47782400e-01 -8.66129339e-01 -1.70795560e-01 3.83798480e-01
7.56020248e-01 1.06197864e-01 1.32342005e+00 2.33706832e-01
-1.00054681e+00 3.85320663e-01 -6.80262625e-01 4.61870849e-01
1.78995296e-01 -1.29926428e-01 -1.32617605e+00 -1.17756240e-01
2.55579930e-02 -1.61925435e-01 1.11471391e+00 6.05590403e-01
1.44120538e+00 -1.11184716e-01 7.41287991e-02 9.74021971e-01
1.61339176e+00 -6.40188009e-02 1.41590083e+00 5.14280915e-01
7.19949841e-01 8.03334653e-01 6.01842582e-01 1.10644199e-01
2.52448320e-01 1.00754523e+00 5.60360253e-01 -1.04105687e+00
-8.51709366e-01 -9.13560167e-02 3.61020803e-01 2.61645079e-01
1.79724246e-02 8.75137895e-02 -2.57594556e-01 3.02941054e-01
-1.31045055e+00 -7.91238308e-01 1.92937441e-03 2.17999768e+00
1.19243729e+00 -7.45257661e-02 -5.75145371e-02 1.94084212e-01
6.68503881e-01 4.19756502e-01 -6.28474057e-01 -3.27935159e-01
-5.35942972e-01 4.80009466e-01 4.32659477e-01 7.03302085e-01
-7.18698919e-01 8.06006730e-01 5.76254177e+00 8.20073903e-01
-1.20294690e+00 -1.37313128e-01 8.10536087e-01 -6.33582398e-02
-4.10094172e-01 2.01765820e-03 -5.68957686e-01 5.12538075e-01
4.54590559e-01 5.67427158e-01 9.81447935e-01 1.49002656e-01
2.98350155e-01 -4.59745407e-01 -8.03458810e-01 1.09549022e+00
4.00272831e-02 -1.19948387e+00 -1.12480745e-01 1.07118666e-01
8.14700603e-01 -1.28059223e-01 4.17938530e-01 -1.34769166e-02
1.17944650e-01 -1.03178298e+00 5.46579719e-01 7.18874514e-01
1.32156301e+00 -6.72779799e-01 5.26155770e-01 1.86420426e-01
-1.03421569e+00 -9.49496254e-02 -4.22307521e-01 1.11029811e-01
3.67582366e-02 1.01130295e+00 -5.13631761e-01 2.95394748e-01
9.67254043e-01 6.60983086e-01 -6.25942886e-01 5.86189687e-01
-4.87072021e-01 2.07209125e-01 -1.17419124e-01 7.71062970e-01
-9.01210159e-02 -3.74312550e-01 3.78724366e-01 1.07996893e+00
1.59873530e-01 3.32024097e-01 -2.31988028e-01 1.49024487e+00
-9.54387039e-02 -4.92432058e-01 -4.61750031e-01 1.63643435e-01
2.62629718e-01 1.55095947e+00 -6.01717532e-01 -1.98280662e-01
-4.03384238e-01 1.24826467e+00 2.75916010e-01 6.80959344e-01
-7.21947193e-01 -5.69060445e-01 6.29737020e-01 5.20833313e-01
3.25768411e-01 -1.38752446e-01 -3.98823947e-01 -9.72269893e-01
-1.04277402e-01 -8.29158485e-01 1.07177682e-01 -1.23657691e+00
-1.00865042e+00 5.48399508e-01 -2.69473910e-01 -1.11399651e+00
1.88391730e-01 -4.53522027e-01 -5.72701216e-01 1.25580919e+00
-2.12890840e+00 -9.53764081e-01 -6.56861842e-01 6.21017933e-01
5.61446309e-01 1.85921729e-01 4.86148179e-01 3.45054209e-01
-5.56305408e-01 2.21691191e-01 -2.43704870e-01 5.59049845e-03
7.77043223e-01 -1.12918222e+00 8.51134285e-02 9.40258086e-01
-2.54053414e-01 4.25502062e-01 5.75522244e-01 -4.39276218e-01
-1.14388669e+00 -1.16044247e+00 3.20592850e-01 -1.18062310e-01
1.88311681e-01 -4.20041502e-01 -1.18566084e+00 3.25224876e-01
2.44928285e-01 1.12870142e-01 2.81629443e-01 -4.11085635e-01
-3.58944297e-01 -4.56646621e-01 -1.24391675e+00 4.40298855e-01
6.31188333e-01 -7.70848691e-01 -2.90594220e-01 -6.14470430e-02
5.43646932e-01 -3.56838256e-01 -7.46588886e-01 5.38481832e-01
5.19684732e-01 -1.46338379e+00 1.06948960e+00 3.00541192e-01
6.59247458e-01 -7.43853509e-01 -3.38119149e-01 -1.36972976e+00
-2.99308985e-01 -3.83238435e-01 -1.11017711e-01 1.13873756e+00
1.79372385e-01 -5.87494791e-01 5.41525960e-01 3.35482091e-01
-1.49433121e-01 -5.63365996e-01 -6.59802735e-01 -3.36338401e-01
-1.66624859e-01 3.01820226e-02 3.67778003e-01 8.80056024e-01
-5.95017314e-01 1.79322109e-01 -4.74964708e-01 3.10164541e-01
7.51095355e-01 3.91909093e-01 4.24483687e-01 -9.70277309e-01
-3.46851200e-01 -1.47861272e-01 -2.91388649e-02 -8.96717727e-01
4.89918003e-03 -6.72291756e-01 2.20868349e-01 -1.50143850e+00
2.46793896e-01 -4.37112391e-01 -6.89977482e-02 4.86679316e-01
-2.28416115e-01 8.16728830e-01 3.41956019e-01 3.23621809e-01
-2.32247010e-01 6.09477878e-01 1.64043641e+00 -5.87235391e-02
-2.47978821e-01 -2.32449651e-01 -6.39330089e-01 5.49902558e-01
7.59586215e-01 -1.57848522e-01 -3.36656123e-01 -4.03381348e-01
1.66231409e-01 2.01179483e-03 7.90581346e-01 -9.05445457e-01
-8.61299504e-03 7.16264397e-02 1.00007629e+00 -4.75724816e-01
4.67030287e-01 -8.86837423e-01 2.27958977e-01 2.25198045e-01
-3.29506397e-01 -5.33832550e-01 1.65135518e-01 1.66704655e-01
-2.04469383e-01 2.05925032e-01 1.40422094e+00 -1.94689929e-01
-7.82974184e-01 1.53001592e-01 -2.11947680e-01 -1.78979486e-01
7.45202184e-01 -2.47206867e-01 -5.76716602e-01 -4.41969156e-01
-4.78469193e-01 -2.11639613e-01 8.31373811e-01 1.70881435e-01
9.37494993e-01 -1.18909240e+00 -4.93690521e-01 8.09070230e-01
-1.24627113e-01 6.58956617e-02 6.23424828e-01 8.86740804e-01
-6.33129060e-01 -1.69414729e-02 -4.98870552e-01 -5.09111583e-01
-9.38637793e-01 5.89971662e-01 6.13882661e-01 -4.05086949e-03
-9.59301353e-01 4.70608890e-01 8.04124892e-01 -1.47709087e-01
6.12661541e-02 -1.79091960e-01 -1.25449151e-01 -3.74386728e-01
7.07761228e-01 3.75441790e-01 8.45409650e-03 -4.90738362e-01
3.24156806e-02 7.49763846e-01 6.59478679e-02 -6.87104315e-02
1.39454341e+00 -6.03759587e-01 -2.95324981e-01 3.93896997e-01
1.29877651e+00 -9.76562276e-02 -1.76142025e+00 -2.68014729e-01
-7.49846280e-01 -7.31697381e-01 6.24554515e-01 -1.00723302e+00
-1.57091892e+00 1.18904543e+00 8.24250758e-01 -2.64428519e-02
1.79997909e+00 3.77608500e-02 8.40517163e-01 -2.03809097e-01
-1.83799546e-02 -1.14009356e+00 2.01013818e-01 4.14123051e-02
8.76902044e-01 -1.18654513e+00 5.82943708e-02 -4.99080837e-01
-5.37923574e-01 1.26201379e+00 6.75393581e-01 -5.43194115e-02
3.92960310e-01 4.32722449e-01 9.18958560e-02 -2.64423221e-01
-4.66648519e-01 -3.42106611e-01 3.17900509e-01 6.79328024e-01
1.22272678e-01 -2.42782623e-01 3.57770711e-01 1.21062525e-01
3.56775969e-02 -1.95239574e-01 4.44337189e-01 4.31483418e-01
-5.25923967e-01 -6.23485744e-01 -4.00828004e-01 1.60860062e-01
-1.53252885e-01 -3.18297446e-01 -2.38611802e-01 7.37526417e-01
3.55171680e-01 7.22887814e-01 1.80564940e-01 -3.84728670e-01
1.88232809e-01 -3.00108105e-01 6.65236235e-01 -3.62655699e-01
-2.48937026e-01 2.96533674e-01 -2.74088651e-01 -8.51881802e-01
-4.39004213e-01 -2.55124450e-01 -1.23881710e+00 -3.17618936e-01
-2.92756651e-02 -5.01731396e-01 6.75034881e-01 6.88937843e-01
2.10300282e-01 8.07143033e-01 8.54478419e-01 -1.06186199e+00
1.43622681e-01 -7.36413956e-01 -9.88955557e-01 3.37670952e-01
7.25303113e-01 -5.84107399e-01 -5.57776630e-01 -1.98169779e-02] | [10.858057022094727, -2.2092080116271973] |
d8b9c953-61e4-4a68-91a9-aca5eb2628ef | locate-then-generate-bridging-vision-and | 2304.01603 | null | https://arxiv.org/abs/2304.01603v1 | https://arxiv.org/pdf/2304.01603v1.pdf | Locate Then Generate: Bridging Vision and Language with Bounding Box for Scene-Text VQA | In this paper, we propose a novel multi-modal framework for Scene Text Visual Question Answering (STVQA), which requires models to read scene text in images for question answering. Apart from text or visual objects, which could exist independently, scene text naturally links text and visual modalities together by conveying linguistic semantics while being a visual object in an image simultaneously. Different to conventional STVQA models which take the linguistic semantics and visual semantics in scene text as two separate features, in this paper, we propose a paradigm of "Locate Then Generate" (LTG), which explicitly unifies this two semantics with the spatial bounding box as a bridge connecting them. Specifically, at first, LTG locates the region in an image that may contain the answer words with an answer location module (ALM) consisting of a region proposal network and a language refinement network, both of which can transform to each other with one-to-one mapping via the scene text bounding box. Next, given the answer words selected by ALM, LTG generates a readable answer sequence with an answer generation module (AGM) based on a pre-trained language model. As a benefit of the explicit alignment of the visual and linguistic semantics, even without any scene text based pre-training tasks, LTG can boost the absolute accuracy by +6.06% and +6.92% on the TextVQA dataset and the ST-VQA dataset respectively, compared with a non-pre-training baseline. We further demonstrate that LTG effectively unifies visual and text modalities through the spatial bounding box connection, which is underappreciated in previous methods. | ['Linli Xu', 'Changcun Bao', 'Hao liu', 'Xin Li', 'Yukang Liang', 'Zhen Liu', 'Yongxin Zhu'] | 2023-04-04 | null | null | null | null | ['answer-generation'] | ['natural-language-processing'] | [ 3.63037825e-01 3.25572908e-01 2.07905412e-01 -5.10251284e-01
-8.58146489e-01 -7.95152009e-01 8.86808395e-01 2.42801741e-01
-3.78719240e-01 1.34381846e-01 2.86526829e-01 -4.21152025e-01
3.52906913e-01 -8.59663665e-01 -1.03105521e+00 -4.14071143e-01
5.77257693e-01 6.10494137e-01 6.11929893e-01 -2.73561448e-01
2.15154141e-01 1.04678676e-01 -1.44774270e+00 8.15662026e-01
7.59999573e-01 1.11168444e+00 5.86556673e-01 7.13316798e-01
-7.88592398e-01 1.04245484e+00 -4.60433096e-01 -5.32655716e-01
2.92369034e-02 -5.33031762e-01 -1.22730565e+00 2.67039567e-01
9.33513761e-01 -3.01726282e-01 -3.13529611e-01 7.75992811e-01
2.75479645e-01 1.66702718e-01 7.76237786e-01 -1.25793397e+00
-1.03590524e+00 3.09656501e-01 -5.24313509e-01 -1.18281797e-01
7.17581570e-01 2.89247155e-01 1.27258754e+00 -1.03617394e+00
7.73743331e-01 1.56362557e+00 8.21666121e-02 6.43461585e-01
-1.18613219e+00 -2.81588495e-01 3.52808148e-01 2.00710475e-01
-1.30190837e+00 -3.50283831e-01 8.08798730e-01 -4.47180390e-01
1.06100869e+00 2.56840616e-01 4.53504026e-01 8.51211369e-01
1.75390109e-01 1.05891824e+00 8.64497483e-01 -4.00908977e-01
1.48721663e-02 1.80646062e-01 -9.37960595e-02 8.78401279e-01
-2.91073531e-01 -2.66156495e-01 -5.75293660e-01 2.79810667e-01
7.52324402e-01 -6.15209155e-02 -2.12397397e-01 -7.75636375e-01
-1.28488982e+00 8.40841770e-01 7.66593218e-01 4.32204567e-02
-1.38883978e-01 3.81221771e-01 3.46421480e-01 1.16349973e-01
2.28892595e-01 2.78835058e-01 -2.93012857e-01 5.64800918e-01
-7.53714859e-01 9.70419124e-02 3.91472667e-01 1.01365960e+00
1.00522780e+00 -6.99016303e-02 -7.17402339e-01 6.32500529e-01
5.65200806e-01 7.32799053e-01 1.87642381e-01 -8.23635995e-01
7.77951658e-01 1.13732588e+00 7.28505626e-02 -9.81296718e-01
-1.29008695e-01 -1.52987197e-01 -7.60690868e-01 -6.94394037e-02
4.35468763e-01 3.32565159e-01 -1.15390182e+00 1.70218349e+00
4.77726519e-01 5.20862080e-03 2.99020678e-01 1.08078742e+00
1.34934711e+00 9.57823873e-01 3.15818250e-01 1.88044995e-01
1.72530997e+00 -1.30955946e+00 -5.21280169e-01 -5.36025763e-01
4.84938860e-01 -7.53634989e-01 1.76171517e+00 -3.75603586e-02
-1.03229308e+00 -6.63527966e-01 -6.75052524e-01 -7.55226910e-01
-4.69636202e-01 1.84432529e-02 3.29610676e-01 1.56159744e-01
-1.45231843e+00 -3.99552763e-01 -3.05746734e-01 -6.31473780e-01
3.51731151e-01 -6.31384924e-03 -3.96802545e-01 -2.03786761e-01
-9.98696804e-01 8.20693612e-01 4.73809838e-01 -3.80201861e-02
-1.13840723e+00 -5.57803810e-01 -1.19574022e+00 3.14413123e-02
5.11896729e-01 -1.19098747e+00 1.17383742e+00 -1.17177153e+00
-1.29417074e+00 1.25280941e+00 -6.70357287e-01 -3.50655288e-01
3.51195633e-01 -3.17282267e-02 -1.89519003e-01 6.16562009e-01
3.78499657e-01 1.23255754e+00 1.04395962e+00 -1.58795130e+00
-4.69781160e-01 -3.72565627e-01 4.06360120e-01 5.00793278e-01
1.65148139e-01 -3.01508456e-02 -8.97492528e-01 -4.91600126e-01
2.22587794e-01 -4.13334340e-01 -9.51035507e-03 2.57569581e-01
-3.67806315e-01 -4.40947384e-01 9.28361416e-01 -6.38213277e-01
9.18061376e-01 -2.02385283e+00 2.35399365e-01 -4.67178859e-02
2.37467766e-01 -5.75926937e-02 -5.01974761e-01 4.00917500e-01
5.18790409e-02 -5.71032576e-02 -2.88494468e-01 -4.09921199e-01
2.29422189e-02 3.04772377e-01 -7.10324764e-01 2.79299974e-01
3.76705229e-01 1.67203927e+00 -9.01174724e-01 -6.81570292e-01
3.74589294e-01 3.00276935e-01 -4.53017175e-01 2.97182679e-01
-9.39432681e-01 4.45631385e-01 -4.74785537e-01 6.01623535e-01
5.51065087e-01 -7.30107427e-01 -7.01014623e-02 -3.32646430e-01
7.47889057e-02 1.79742008e-01 -7.84648895e-01 1.98341393e+00
-4.37868714e-01 5.81822813e-01 -3.06591019e-02 -9.99550164e-01
9.96947169e-01 1.84265748e-01 -1.40077276e-02 -1.10215271e+00
1.32367723e-02 -8.31938609e-02 -4.76237804e-01 -6.19755626e-01
5.71310103e-01 -1.16034448e-01 -2.11411476e-01 4.96101439e-01
1.75330371e-01 -4.39098448e-01 -2.09785700e-02 7.63055801e-01
6.99661374e-01 2.55067378e-01 2.37867758e-01 -8.31219330e-02
7.55819201e-01 2.23301068e-01 -1.18545040e-01 7.63696134e-01
-7.63456970e-02 8.59064460e-01 4.16878760e-01 -2.63618261e-01
-8.91524374e-01 -1.33208632e+00 1.89439416e-01 1.32618392e+00
6.17050707e-01 -2.88895190e-01 -5.66590548e-01 -8.61580968e-01
-1.14282675e-01 8.57252240e-01 -6.10333145e-01 4.09253165e-02
-4.30404425e-01 9.91625637e-02 4.37653542e-01 4.10317689e-01
8.04144263e-01 -1.24312103e+00 -7.07985401e-01 -2.06899509e-01
-5.76796770e-01 -1.33067811e+00 -6.67359114e-01 -1.19872227e-01
-4.71610785e-01 -9.53766584e-01 -5.94601512e-01 -1.02770507e+00
7.49063373e-01 5.09514928e-01 1.34679282e+00 2.62214899e-01
5.56419417e-02 8.67462337e-01 -4.76646423e-01 -1.35636076e-01
-2.89492577e-01 -1.41234398e-01 -6.03486538e-01 2.68315911e-01
1.67890906e-01 -9.27351564e-02 -7.11488724e-01 1.72220767e-01
-1.12080252e+00 6.82889760e-01 3.96198004e-01 5.10675609e-01
7.58138120e-01 -4.77680475e-01 4.14461672e-01 -7.35087156e-01
1.11747935e-01 -4.34896529e-01 -5.74650943e-01 6.66889191e-01
-1.54007554e-01 2.28250533e-01 4.01678085e-01 -2.66072482e-01
-1.16606593e+00 2.50599474e-01 -6.40626252e-02 -4.48433191e-01
-2.61168122e-01 2.68883914e-01 -4.24425840e-01 1.97563440e-01
5.48372090e-01 5.94011962e-01 -6.43548071e-02 -9.23852399e-02
9.32057142e-01 3.19978416e-01 5.94904661e-01 -5.41101038e-01
8.80524755e-01 7.54738212e-01 -1.91875592e-01 -6.22375667e-01
-1.08247972e+00 -6.04461789e-01 -6.66605175e-01 -2.03526720e-01
1.47798049e+00 -9.44531381e-01 -8.21744382e-01 2.25118473e-01
-1.56768560e+00 -4.34318304e-01 -2.09032416e-01 -8.27032253e-02
-6.87398791e-01 4.30824429e-01 -2.96303123e-01 -6.22086942e-01
-3.51277947e-01 -1.03582907e+00 1.59854054e+00 1.40970930e-01
-1.03011243e-02 -9.41807866e-01 -4.89634603e-01 8.04033577e-01
1.83476374e-01 1.52954534e-01 1.22588909e+00 -4.66912538e-01
-9.82779562e-01 2.47747064e-01 -7.84259617e-01 3.78216840e-02
-9.20737386e-02 -4.09995139e-01 -1.00422013e+00 -1.60521582e-01
-5.40389912e-04 -4.57387060e-01 1.00984621e+00 2.42017150e-01
1.06265724e+00 -2.89687097e-01 -2.50051647e-01 5.26160657e-01
1.52770197e+00 8.73950198e-02 6.42406940e-01 2.35941932e-01
1.05354440e+00 8.78336966e-01 5.05750000e-01 5.59624881e-02
9.06917214e-01 5.60844064e-01 8.09096634e-01 -3.41808707e-01
-5.38944066e-01 -6.14959002e-01 3.12401682e-01 4.08270657e-01
5.36606550e-01 -4.73425448e-01 -1.04482603e+00 6.06795371e-01
-1.86370230e+00 -8.18813205e-01 -4.19580817e-01 1.88273787e+00
7.91362405e-01 -2.84040272e-01 -1.03851430e-01 -4.24819171e-01
5.69110394e-01 3.97059023e-01 -5.94109237e-01 -3.39576602e-01
-4.31172818e-01 1.23831250e-01 2.25492820e-01 6.78295672e-01
-9.80277658e-01 1.47723174e+00 5.53961515e+00 7.64848888e-01
-1.00025392e+00 1.64392427e-01 6.02677107e-01 1.78124994e-01
-7.77527213e-01 1.69249982e-01 -6.67066455e-01 -6.90524951e-02
3.90967518e-01 1.30417049e-01 4.59248036e-01 5.40023685e-01
1.06783584e-01 -2.79280990e-01 -1.13228965e+00 9.76060390e-01
4.95355785e-01 -1.50714874e+00 8.50524366e-01 -3.69883209e-01
5.76036930e-01 -1.66149050e-01 9.63293239e-02 2.25274056e-01
1.83520049e-01 -1.27000833e+00 1.18925357e+00 5.02906501e-01
9.44580793e-01 -4.33785886e-01 3.56010109e-01 2.88080454e-01
-1.56127954e+00 -3.12736221e-02 -3.48124564e-01 -7.20261550e-03
3.75427872e-01 1.72728390e-01 -8.00663531e-01 7.57978797e-01
6.53501689e-01 6.23831630e-01 -8.37583721e-01 5.16029716e-01
-6.17207050e-01 3.77106667e-01 6.82324171e-02 1.01806387e-01
3.84159982e-01 -5.47713228e-02 4.16072220e-01 9.00912881e-01
5.99798299e-02 7.02909827e-02 2.00238019e-01 1.21325886e+00
-8.40051770e-02 1.58525780e-01 -6.48053944e-01 1.87521562e-01
2.10649535e-01 1.12260270e+00 -7.19280720e-01 -5.34791768e-01
-5.15910625e-01 1.36874926e+00 3.19584459e-01 8.64176095e-01
-8.56719613e-01 -1.77332789e-01 9.71050039e-02 2.22703859e-01
3.27578306e-01 -1.40328407e-01 -2.76303917e-01 -1.26898646e+00
8.24332833e-02 -7.96729863e-01 4.98980343e-01 -1.47618389e+00
-1.15968943e+00 6.30492985e-01 -3.70925777e-02 -9.77990985e-01
-1.51453450e-01 -5.83639324e-01 -3.57045799e-01 1.03646564e+00
-1.51202500e+00 -1.85503066e+00 -2.84800142e-01 9.61334229e-01
8.42122376e-01 1.32991046e-01 5.88510871e-01 -7.83936828e-02
-2.33278237e-02 3.29731166e-01 -5.28561473e-01 1.22429959e-01
6.22792184e-01 -1.20889115e+00 3.87853205e-01 9.31755066e-01
4.94356334e-01 5.11238754e-01 4.52943802e-01 -5.45112491e-01
-1.45506394e+00 -1.37227416e+00 1.13424885e+00 -8.51271868e-01
6.21465027e-01 -7.07306981e-01 -1.05979574e+00 8.03554893e-01
5.30297041e-01 -1.26063675e-01 1.68920338e-01 -2.55540133e-01
-7.38727093e-01 7.52673745e-02 -8.31085384e-01 7.54701078e-01
1.03528559e+00 -9.28654015e-01 -8.77166867e-01 2.93778390e-01
1.17872500e+00 -4.10664201e-01 -2.31706619e-01 3.38330418e-01
8.48588496e-02 -8.83545995e-01 1.15147722e+00 -4.94758159e-01
5.43007433e-01 -7.95462847e-01 -5.10927379e-01 -7.72328854e-01
8.80093221e-03 -1.10436358e-01 1.60407946e-01 1.31651747e+00
3.77704948e-01 -3.06153566e-01 4.57052052e-01 3.06777447e-01
-2.27909043e-01 -4.74879324e-01 -9.42763269e-01 -3.58582556e-01
-1.85674373e-02 -6.07643783e-01 5.26867092e-01 7.82343030e-01
-3.90146077e-01 9.70696330e-01 -1.92346528e-01 2.99630165e-01
2.96975017e-01 4.06436741e-01 9.33236599e-01 -9.14027989e-01
-2.34151661e-01 -3.95894378e-01 -8.44224766e-02 -1.75185645e+00
3.05473089e-01 -1.22440505e+00 2.65243381e-01 -2.13817167e+00
3.08249772e-01 -6.65265247e-02 9.67374295e-02 6.18643701e-01
-2.41368026e-01 2.73539215e-01 2.81389713e-01 2.90662795e-02
-1.09774220e+00 7.31908619e-01 1.53195894e+00 -3.76492351e-01
-6.67712539e-02 -5.49513161e-01 -6.60700321e-01 6.55786753e-01
5.40589988e-01 -8.02385435e-02 -7.95403421e-01 -9.11449075e-01
3.41064543e-01 2.49548271e-01 8.97403598e-01 -4.43817407e-01
2.89544612e-01 -2.75917530e-01 2.85610825e-01 -8.76531899e-01
4.16248053e-01 -7.72193789e-01 -2.25493208e-01 1.38442591e-01
-3.77595276e-01 5.55482358e-02 2.84701437e-01 6.09862983e-01
-3.57633352e-01 7.45727420e-02 5.56564689e-01 -1.43714800e-01
-1.15610754e+00 1.59634292e-01 -1.96000233e-01 2.06588298e-01
9.51003671e-01 -4.32826996e-01 -6.14019096e-01 -5.64548969e-01
-4.63070631e-01 6.21096373e-01 4.76986647e-01 5.50088704e-01
1.00783968e+00 -1.20643437e+00 -5.78784049e-01 1.05268784e-01
5.77941775e-01 3.14169168e-01 5.91677010e-01 6.55296922e-01
-4.20382589e-01 5.50192237e-01 2.33765468e-01 -8.87921810e-01
-1.14494526e+00 7.87729144e-01 4.44153517e-01 -8.82190168e-02
-5.33425391e-01 9.67614591e-01 1.07839286e+00 -6.38770759e-01
1.15892805e-01 -3.58139992e-01 -2.01886967e-01 -8.27976093e-02
3.30875665e-01 -2.76925445e-01 -3.20421338e-01 -9.96559322e-01
-3.41306716e-01 8.07542324e-01 2.20704764e-01 -3.13267171e-01
7.30989933e-01 -5.59693456e-01 -4.26432967e-01 5.25010228e-01
1.05673373e+00 -1.47798866e-01 -1.07417440e+00 -3.42394441e-01
-1.04543231e-01 -2.66247660e-01 -2.64724463e-01 -9.34209287e-01
-8.34143043e-01 1.20921063e+00 3.51255000e-01 -9.81635600e-03
1.20964563e+00 7.21806645e-01 6.78448737e-01 2.26451233e-01
1.48391962e-01 -5.68989694e-01 7.67848253e-01 7.26061881e-01
1.15999019e+00 -1.30573475e+00 -5.14424801e-01 -4.99241054e-01
-9.26118493e-01 7.85491645e-01 9.59335685e-01 1.34611845e-01
2.21252576e-01 -3.89764696e-01 2.24962533e-01 -4.45389330e-01
-7.55967379e-01 -4.77886558e-01 8.22292864e-01 5.79079032e-01
2.58186489e-01 -3.70036438e-02 5.34945726e-02 3.70333403e-01
-2.76690833e-02 -4.63653833e-01 1.28441572e-01 4.97182488e-01
-6.48219228e-01 -8.44601929e-01 -3.27946812e-01 1.89646974e-01
2.74043251e-02 -4.05976892e-01 -4.90355998e-01 7.57094562e-01
1.01023927e-01 1.14945161e+00 3.83460850e-01 -1.87827781e-01
2.36508489e-01 1.40166327e-01 3.35124314e-01 -7.86671340e-01
-4.85623688e-01 -8.54167789e-02 -1.65005431e-01 -5.87076604e-01
-4.02252823e-01 -1.58522859e-01 -1.76772344e+00 6.02630079e-02
-5.64606637e-02 -1.13438502e-01 3.37659150e-01 1.11110973e+00
3.84431422e-01 4.53657210e-01 1.75212562e-01 -3.26408386e-01
-4.26004604e-02 -6.82155252e-01 -2.33106524e-01 5.46633840e-01
3.38679820e-01 -3.67226779e-01 -1.43043295e-01 3.40594441e-01] | [10.81318187713623, 1.629797339439392] |
77735f7d-4d03-41ed-9ad6-3af06754f4dc | ct-image-harmonization-for-enhancing | 2107.01337 | null | https://arxiv.org/abs/2107.01337v1 | https://arxiv.org/pdf/2107.01337v1.pdf | CT Image Harmonization for Enhancing Radiomics Studies | While remarkable advances have been made in Computed Tomography (CT), capturing CT images with non-standardized protocols causes low reproducibility regarding radiomic features, forming a barrier on CT image analysis in a large scale. RadiomicGAN is developed to effectively mitigate the discrepancy caused by using non-standard reconstruction kernels. RadiomicGAN consists of hybrid neural blocks including both pre-trained and trainable layers adopted to learn radiomic feature distributions efficiently. A novel training approach, called Dynamic Window-based Training, has been developed to smoothly transform the pre-trained model to the medical imaging domain. Model performance evaluated using 1401 radiomic features show that RadiomicGAN clearly outperforms the state-of-art image standardization models. | ['Jin Chen', 'Guo-Qiang Zhang', 'Baowei Fei', 'Jie Zhang', 'Md Selim'] | 2021-07-03 | null | null | null | null | ['image-harmonization'] | ['computer-vision'] | [ 1.82668313e-01 8.20660964e-02 -3.77712131e-01 -6.22130275e-01
-1.30240428e+00 -5.66694103e-02 5.47180235e-01 -2.83242583e-01
-6.80588365e-01 5.49414337e-01 4.13985133e-01 -3.64711106e-01
-6.49920925e-02 -6.62300885e-01 -6.09791040e-01 -7.65765727e-01
-5.27678952e-02 5.60077667e-01 1.95887953e-01 1.83798885e-03
-2.67669350e-01 4.05741960e-01 -4.44977105e-01 3.47021371e-01
4.38217849e-01 9.12450492e-01 3.60132456e-01 7.58107424e-01
-8.95415768e-02 1.11123848e+00 -4.13232714e-01 -1.35658100e-01
3.54415298e-01 -6.46188080e-01 -8.42928529e-01 -1.55767158e-01
2.35121459e-01 -4.47540641e-01 -5.67802608e-01 6.98165298e-01
7.05470145e-01 -2.98417091e-01 6.60587549e-01 -6.94360018e-01
-6.42166913e-01 7.02771842e-01 -4.69991624e-01 4.22291189e-01
4.26706225e-02 9.96441767e-02 4.25211072e-01 -5.41832626e-01
6.53580487e-01 7.37827003e-01 9.13599610e-01 6.02720916e-01
-1.10563779e+00 -8.41255426e-01 -5.93419969e-01 -1.04711227e-01
-1.01451874e+00 3.20485123e-02 5.45083165e-01 -4.25638497e-01
6.45566165e-01 2.26950780e-01 7.72989213e-01 1.01849008e+00
9.01847661e-01 2.44960651e-01 1.24507701e+00 -3.34315568e-01
-8.31340626e-02 4.12290096e-02 -1.33836567e-01 9.05772090e-01
1.30279794e-01 2.78004795e-01 -5.11040203e-02 -2.50040799e-01
1.19899487e+00 2.25490540e-01 -3.40214223e-01 -4.94497567e-01
-1.14661479e+00 7.96223760e-01 6.45349562e-01 7.31164396e-01
-7.06032276e-01 3.74470264e-01 6.41673386e-01 3.29669237e-01
3.91174287e-01 3.26828778e-01 -2.30231702e-01 1.51157379e-01
-8.91918778e-01 -2.01188549e-01 3.23319256e-01 5.76486528e-01
4.58061367e-01 1.72790095e-01 -1.72696665e-01 7.73983777e-01
2.07727388e-01 4.00057107e-01 1.09453368e+00 -6.64803028e-01
2.32581347e-01 4.62568700e-01 -5.53536236e-01 -4.95726377e-01
-6.15742207e-01 -5.71699739e-01 -9.75630105e-01 2.76589245e-01
2.75648862e-01 1.10891089e-01 -1.46265113e+00 1.50126708e+00
3.62641603e-01 2.58086652e-01 -1.38572752e-01 8.63637865e-01
9.28079009e-01 3.75885338e-01 4.93883371e-01 2.82257169e-01
1.37815034e+00 -7.94204950e-01 -3.24763685e-01 1.51679054e-01
7.85914600e-01 -8.76953185e-01 9.57398117e-01 2.10481789e-02
-1.10212421e+00 -4.05756801e-01 -1.16854346e+00 1.91872232e-02
-1.37538031e-01 -1.72348976e-01 7.44892836e-01 9.73542333e-01
-1.04686821e+00 4.32269871e-01 -1.06513345e+00 -2.45237708e-01
7.24916637e-01 5.74197471e-01 -7.56626248e-01 -3.99065614e-01
-9.90155280e-01 1.15138841e+00 3.47192138e-01 -3.41601409e-02
-1.17936003e+00 -1.42974293e+00 -8.46856594e-01 -1.78358510e-01
1.97381422e-01 -9.47312236e-01 1.23113048e+00 -9.62808967e-01
-1.86104643e+00 9.58201170e-01 5.08928657e-01 -5.26243091e-01
7.43886054e-01 6.23132400e-02 -5.43464899e-01 5.04008591e-01
5.98992109e-02 5.40604591e-01 9.09892142e-01 -1.09339237e+00
-3.19632351e-01 2.94610132e-02 -3.01757365e-01 2.86586434e-02
9.63592678e-02 2.72311062e-01 -5.22865057e-01 -8.01796973e-01
1.83718979e-01 -1.05385292e+00 -5.96131742e-01 1.22336738e-01
-5.99235520e-02 2.13232264e-01 5.53287089e-01 -7.31642008e-01
7.40591705e-01 -1.73784137e+00 -2.88777798e-01 4.38381404e-01
4.71141726e-01 1.50181130e-01 -1.20576739e-01 -1.27388224e-01
-6.51233792e-01 -6.31711334e-02 -5.77353179e-01 -1.07034020e-01
-5.41927993e-01 1.87858045e-01 4.98766173e-03 8.61843944e-01
-8.98432732e-02 1.08719730e+00 -7.74000466e-01 -7.43564308e-01
7.07457781e-01 6.77691996e-01 -4.74464267e-01 3.87828052e-01
2.85888821e-01 8.89122784e-01 -5.54931879e-01 6.29245579e-01
9.46925461e-01 -4.64815795e-01 1.78135231e-01 -1.92762360e-01
2.04905421e-01 1.26237005e-01 -4.58866566e-01 2.14819598e+00
-9.21224296e-01 2.28117406e-01 -1.00783147e-01 -8.33296001e-01
3.67518485e-01 5.79478681e-01 1.25708556e+00 -9.19919431e-01
4.89044935e-01 1.92876056e-01 4.40962873e-02 -4.65514719e-01
2.52611578e-01 -7.49014318e-01 2.73123328e-02 5.85790277e-01
3.21831137e-01 -3.70166719e-01 -3.20290178e-01 -1.15800388e-02
1.21062481e+00 -8.89555439e-02 2.86912650e-01 -3.09822708e-01
5.20400703e-01 6.10239469e-02 2.55642414e-01 7.07275808e-01
-3.20998132e-01 1.25411236e+00 1.29590482e-02 -6.21055424e-01
-1.10712075e+00 -1.40723515e+00 -3.73183548e-01 7.23623037e-01
-1.74059331e-01 -1.27595589e-01 -6.50291204e-01 -9.44818258e-01
-2.92486697e-01 1.89917296e-01 -1.02205658e+00 -2.73502350e-01
-9.22115207e-01 -1.06707752e+00 7.51248360e-01 5.63738525e-01
4.44716007e-01 -9.15645123e-01 -6.36148214e-01 4.54263866e-01
-7.23759159e-02 -1.11947966e+00 -5.19896030e-01 3.85928631e-01
-1.14622915e+00 -1.30498219e+00 -1.24158716e+00 -5.94884276e-01
7.57935762e-01 2.36126661e-01 1.22190368e+00 1.98776960e-01
-5.95014155e-01 4.63009000e-01 -2.54019678e-01 -2.71084160e-01
-6.85256481e-01 2.99073249e-01 -4.91397470e-01 -2.50530720e-01
-1.52389389e-02 -4.71228540e-01 -9.20545876e-01 2.31266499e-01
-1.24985170e+00 3.40156332e-02 1.00885308e+00 9.65558589e-01
5.65476716e-01 -3.56359899e-01 1.76973298e-01 -1.23934245e+00
1.84157878e-01 -5.94762206e-01 -2.90979117e-01 2.61279136e-01
-6.94802284e-01 1.16875842e-02 5.94259739e-01 -2.05067977e-01
-1.06037009e+00 -6.65761083e-02 -3.59234869e-01 -2.18421742e-01
-1.22814542e-02 3.66397321e-01 4.31488663e-01 -5.30490279e-01
7.65349388e-01 2.48397514e-01 9.61965993e-02 -3.13811094e-01
6.98547885e-02 3.10640842e-01 6.77283466e-01 -5.23235440e-01
9.09747541e-01 8.95996332e-01 -9.89621803e-02 -7.18983650e-01
-5.39869487e-01 -5.45340002e-01 -5.09219468e-01 -9.10831243e-02
1.08522916e+00 -9.67482746e-01 -4.60371114e-02 5.33625245e-01
-7.51893342e-01 -4.97021228e-01 -4.18460280e-01 9.99857306e-01
-4.96029079e-01 2.28921473e-01 -7.68814862e-01 1.44328937e-01
-6.43103540e-01 -1.30161130e+00 7.25289285e-01 1.96207672e-01
-1.06224157e-01 -1.00881469e+00 6.88495755e-01 1.90215960e-01
9.13216650e-01 3.21848750e-01 8.22925866e-01 -3.29009622e-01
-5.61170459e-01 -2.48851702e-01 -3.14348966e-01 3.92823428e-01
3.12283188e-01 -3.99453610e-01 -9.46685135e-01 -2.69957304e-01
3.22537750e-01 -3.48923981e-01 8.18327367e-01 8.67389500e-01
1.44102657e+00 5.38467348e-01 -2.48214722e-01 1.22428691e+00
1.37416029e+00 3.79072949e-02 8.70122015e-01 6.54394329e-01
6.79732621e-01 2.91717164e-02 -2.90965643e-02 1.59248203e-01
3.76207590e-01 4.12703872e-01 3.18918586e-01 -6.71381354e-01
-4.17994082e-01 -9.00326520e-02 -2.27207020e-02 8.61648321e-01
-2.04463437e-01 1.14843272e-01 -8.25735211e-01 2.72000253e-01
-1.19890845e+00 -6.20369673e-01 1.47933885e-01 1.83826435e+00
8.40371490e-01 8.13649036e-03 -2.22451463e-01 -2.88285524e-01
5.48488498e-01 3.52033913e-01 -2.90304393e-01 -8.52256194e-02
-7.76311532e-02 8.17353904e-01 9.34755862e-01 1.73713744e-01
-1.13789225e+00 6.16636455e-01 7.67388439e+00 7.91021287e-01
-1.64806294e+00 6.39054596e-01 6.33226097e-01 2.00734049e-01
-2.80185491e-01 -8.66236389e-02 4.68495898e-02 2.86051780e-01
9.93510842e-01 4.25144285e-02 -9.10119563e-02 9.90600348e-01
-1.31768927e-01 -1.79692805e-01 -7.67023683e-01 1.00468421e+00
7.31257349e-02 -1.51842308e+00 -1.75497718e-02 1.14732593e-01
6.42315924e-01 4.66235816e-01 2.77742445e-01 2.86156327e-01
5.82651198e-01 -1.18083680e+00 5.16236126e-01 2.70417929e-01
1.22250569e+00 -3.86405557e-01 8.72523904e-01 -4.27223146e-01
-1.09298813e+00 4.78890419e-01 -3.89962465e-01 5.97561777e-01
2.06456140e-01 4.14230943e-01 -1.16913164e+00 8.42332840e-01
6.70912206e-01 5.49498796e-01 -4.91763353e-01 1.15063429e+00
2.19224882e-03 6.69789016e-01 -2.16522574e-01 7.39970684e-01
3.87384593e-01 -6.54863119e-02 4.76954319e-02 1.19014227e+00
3.28046829e-01 1.16423629e-01 -1.01193994e-01 5.43351293e-01
-1.52290910e-01 1.03661709e-01 -2.99936563e-01 2.74840832e-01
-2.13946134e-01 1.35383129e+00 -9.52810884e-01 -4.50162113e-01
-6.27242088e-01 9.58666384e-01 3.74782830e-03 5.39275557e-02
-1.16459119e+00 1.47098675e-01 2.88516879e-01 1.55467778e-01
5.58203505e-03 -1.62080884e-01 -3.69347036e-01 -1.07878065e+00
-4.60715622e-01 -7.30678260e-01 8.03051829e-01 -5.27840137e-01
-1.13000154e+00 9.27541852e-01 -1.72764603e-02 -1.28250146e+00
-7.69457370e-02 -5.68137348e-01 -7.46777594e-01 7.77256787e-01
-1.78925908e+00 -1.44492662e+00 -3.64953518e-01 8.52962732e-01
3.82887512e-01 -9.88646522e-02 8.62809062e-01 5.42540014e-01
-3.00580055e-01 6.36813819e-01 1.10919580e-01 2.08746478e-01
9.32885230e-01 -1.24939013e+00 1.78479925e-02 5.74147820e-01
-2.14438230e-01 4.98824537e-01 3.38144004e-01 -5.47897637e-01
-1.23385370e+00 -9.61985528e-01 3.45209271e-01 -4.37246174e-01
5.10552168e-01 -1.21606011e-02 -8.47013772e-01 8.42609942e-01
2.85659850e-01 5.38250506e-01 8.65710378e-01 -4.06099170e-01
-2.93213457e-01 -7.54021332e-02 -1.22892857e+00 2.95753181e-01
5.53357661e-01 -6.48088455e-01 -6.56511426e-01 1.08084776e-01
4.06051993e-01 -7.80986249e-01 -1.22403085e+00 3.04159582e-01
7.63166368e-01 -8.31633270e-01 1.27655745e+00 -3.41396660e-01
4.90230590e-01 -1.37587236e-02 8.61418843e-02 -1.19448531e+00
-1.76409930e-01 -3.09165567e-01 5.35896540e-01 5.08887947e-01
2.60506034e-01 -6.62086666e-01 1.07357204e+00 6.65191710e-01
-4.23159838e-01 -4.82172012e-01 -1.09363401e+00 -6.01557136e-01
3.46224636e-01 -5.30893683e-01 5.76491892e-01 1.10302949e+00
-4.57916051e-01 -2.44621783e-01 -4.16339636e-01 -8.92791748e-02
5.03723919e-01 -2.35597163e-01 6.52462423e-01 -6.95848167e-01
-3.62483591e-01 -4.56539184e-01 -3.63132179e-01 -5.48669338e-01
-8.28078315e-02 -1.00736821e+00 -1.41201794e-01 -1.07812750e+00
6.28439367e-01 -8.40315819e-01 -9.05440331e-01 2.77116299e-01
-7.30854347e-02 6.85593247e-01 1.35862920e-02 2.58737892e-01
-2.60663718e-01 4.43792790e-01 1.64038086e+00 -1.84120163e-01
7.17195123e-02 -1.46584541e-01 -5.19607425e-01 6.24823213e-01
6.48130953e-01 -9.54126120e-01 -2.19853327e-01 -3.43173087e-01
-2.14065865e-01 1.04343906e-01 2.82765120e-01 -1.24192047e+00
4.61968184e-02 3.71425808e-03 6.43951535e-01 -4.08438712e-01
7.32228085e-02 -1.09844720e+00 4.52400327e-01 8.76939595e-01
-3.49384129e-01 2.63231814e-01 7.87122622e-02 2.82150477e-01
-3.42096120e-01 -2.44230349e-02 1.16478145e+00 -4.20836776e-01
-4.55849081e-01 5.95948696e-01 -3.62780392e-01 -9.61050019e-02
9.31307375e-01 -8.62613991e-02 -1.11553758e-01 -1.81992024e-01
-4.98933703e-01 -1.65039971e-01 3.85541260e-01 1.39165074e-01
6.15175128e-01 -1.23722160e+00 -6.69267356e-01 1.62568867e-01
1.13924518e-01 -6.81520486e-03 7.27991939e-01 1.27499604e+00
-1.15503335e+00 3.72489184e-01 -5.06014228e-01 -7.66937673e-01
-8.62173915e-01 2.18614221e-01 6.75511420e-01 -9.66558933e-01
-1.12765539e+00 6.15858912e-01 3.89126867e-01 -7.04502881e-01
-1.95208952e-01 -6.51085019e-01 1.54345930e-01 -4.92379576e-01
4.14465040e-01 1.23758063e-01 4.14352775e-01 -5.76715171e-01
-2.54561484e-01 7.19857097e-01 -3.62938285e-01 -1.23487592e-01
1.60527873e+00 2.13926136e-02 5.81320375e-02 1.24535756e-02
1.40941823e+00 -1.49213761e-01 -1.09335279e+00 -3.69099975e-01
-1.97175428e-01 -3.51735204e-01 4.04810905e-01 -8.43227029e-01
-1.41802514e+00 5.94672024e-01 1.07581353e+00 -3.85789216e-01
1.05164135e+00 2.21943650e-02 1.08198786e+00 -9.27595869e-02
1.16857424e-01 -7.76145995e-01 1.03514545e-01 1.76724270e-01
6.47478223e-01 -1.54180157e+00 2.35423803e-01 -2.38988176e-01
-5.25504529e-01 1.24860585e+00 3.23230833e-01 -4.31779385e-01
8.33424687e-01 3.90847892e-01 6.24728084e-01 -3.30574512e-01
-1.21799707e-01 1.60481110e-01 2.84005284e-01 4.58843857e-01
6.06717587e-01 2.34829891e-03 -2.77059883e-01 3.15349400e-01
-1.47106916e-01 3.22601527e-01 3.51860821e-01 1.15757930e+00
3.07027735e-02 -1.29646456e+00 -6.23607695e-01 4.45799559e-01
-9.08957422e-01 -3.96784991e-02 3.20785850e-01 1.28407645e+00
-5.01238555e-02 2.08921537e-01 -1.13714285e-01 -2.52710134e-01
5.65711521e-02 -2.49010310e-01 6.92439020e-01 -5.61775446e-01
-1.04139447e+00 1.15288846e-01 -3.61104280e-01 -7.17674971e-01
-6.19136572e-01 -2.46471673e-01 -1.19767749e+00 -1.72285646e-01
-1.15574189e-01 -4.12298590e-02 8.81086409e-01 8.71490717e-01
-1.72253922e-01 9.43494022e-01 5.03497362e-01 -6.94936037e-01
-4.09342229e-01 -9.07049417e-01 -5.83606482e-01 4.68344688e-01
2.36749843e-01 -4.79394019e-01 -2.71663908e-02 -1.16046809e-01] | [14.19947624206543, -2.393794536590576] |
e9878def-65c6-4797-93e0-58f31b762f1c | interactive-conversational-head-generation | 2307.02090 | null | https://arxiv.org/abs/2307.02090v1 | https://arxiv.org/pdf/2307.02090v1.pdf | Interactive Conversational Head Generation | We introduce a new conversation head generation benchmark for synthesizing behaviors of a single interlocutor in a face-to-face conversation. The capability to automatically synthesize interlocutors which can participate in long and multi-turn conversations is vital and offer benefits for various applications, including digital humans, virtual agents, and social robots. While existing research primarily focuses on talking head generation (one-way interaction), hindering the ability to create a digital human for conversation (two-way) interaction due to the absence of listening and interaction parts. In this work, we construct two datasets to address this issue, ``ViCo'' for independent talking and listening head generation tasks at the sentence level, and ``ViCo-X'', for synthesizing interlocutors in multi-turn conversational scenarios. Based on ViCo and ViCo-X, we define three novel tasks targeting the interaction modeling during the face-to-face conversation: 1) responsive listening head generation making listeners respond actively to the speaker with non-verbal signals, 2) expressive talking head generation guiding speakers to be aware of listeners' behaviors, and 3) conversational head generation to integrate the talking/listening ability in one interlocutor. Along with the datasets, we also propose corresponding baseline solutions to the three aforementioned tasks. Experimental results show that our baseline method could generate responsive and vivid agents that can collaborate with real person to fulfil the whole conversation. Project page: https://vico.solutions/. | ['Tiejun Zhao', 'Ting Yao', 'Wei zhang', 'Yalong Bai', 'Mohan Zhou'] | 2023-07-05 | null | null | null | null | ['talking-head-generation'] | ['computer-vision'] | [ 2.59524435e-02 7.94747829e-01 3.66589695e-01 -4.00995731e-01
-6.47812247e-01 -6.39783859e-01 8.65337551e-01 -7.33898997e-01
3.19944710e-01 7.31405854e-01 8.53967488e-01 7.99214467e-02
2.87725091e-01 -4.69383359e-01 -1.52987570e-01 -8.03700686e-01
1.81083173e-01 8.38267088e-01 -4.81359959e-01 -6.66863084e-01
-2.24252865e-01 5.03412224e-02 -1.83294415e+00 5.94182670e-01
6.80089235e-01 7.64685273e-01 2.79972196e-01 1.06084144e+00
-4.69943322e-02 1.21628022e+00 -9.79638875e-01 -3.04229379e-01
-2.28827804e-01 -8.18545461e-01 -9.70015347e-01 3.15228075e-01
-2.67133981e-01 -4.43874419e-01 -3.14331017e-02 3.72993350e-01
1.16936231e+00 4.77502495e-01 5.21671236e-01 -2.01124859e+00
-4.11244065e-01 1.21381605e+00 -1.27043575e-01 -5.20820141e-01
1.44933367e+00 6.99203849e-01 8.03373158e-01 -5.33459544e-01
6.64967120e-01 1.92360485e+00 3.98175359e-01 1.22145975e+00
-9.11323011e-01 -7.73874879e-01 5.38864136e-02 -1.67148978e-01
-9.67677772e-01 -1.17528677e+00 1.03852630e+00 -4.19811219e-01
8.23801100e-01 5.24598598e-01 5.23133934e-01 1.89753377e+00
-2.78811634e-01 9.05615747e-01 7.81912506e-01 -1.20122589e-01
-8.37654527e-03 2.89517164e-01 -1.49576455e-01 3.26380543e-02
-8.84806871e-01 6.30933866e-02 -8.93621504e-01 -2.20902234e-01
3.54393512e-01 -6.52489305e-01 -4.74770218e-01 4.39505160e-01
-1.50956857e+00 8.50324571e-01 1.10836469e-01 2.28576154e-01
-4.95090365e-01 -9.48159844e-02 7.15075791e-01 3.29522699e-01
3.59020323e-01 4.41883028e-01 1.63841695e-01 -6.54464960e-01
-1.60453051e-01 4.91721034e-01 1.44081998e+00 1.15413189e+00
2.98890173e-01 -2.57718544e-02 -6.07915819e-01 1.12935913e+00
2.16168568e-01 4.74747896e-01 4.18892860e-01 -1.42177331e+00
4.31077003e-01 4.98046815e-01 3.20219159e-01 -9.02303219e-01
-6.79900110e-01 1.84396178e-01 -8.17882478e-01 -1.78420335e-01
1.69607967e-01 -7.44290709e-01 8.50556269e-02 2.05163097e+00
6.79041922e-01 -1.01912327e-01 6.10950410e-01 8.95388126e-01
1.65005589e+00 1.01737559e+00 -3.07211041e-01 -7.84746408e-01
1.36034977e+00 -1.39148951e+00 -1.19293296e+00 -2.21382648e-01
6.94619417e-01 -8.55307460e-01 1.23854089e+00 3.55400741e-01
-1.27773535e+00 -5.36792874e-01 -5.21866500e-01 -6.40979037e-02
2.17921048e-01 -1.60779238e-01 5.59751749e-01 4.39248294e-01
-1.24931300e+00 -2.52774894e-01 -4.24361415e-02 -3.95215899e-01
-3.10252517e-01 2.70116568e-01 -4.00951028e-01 2.26247519e-01
-1.32150841e+00 6.04761541e-01 -1.12888709e-01 8.93085524e-02
-8.89499545e-01 -4.39510763e-01 -9.51669455e-01 -1.02931909e-01
2.46370107e-01 -7.87044466e-01 1.85541952e+00 -1.07432020e+00
-2.29002452e+00 7.99513221e-01 -3.04651260e-01 -9.73565970e-03
7.90240884e-01 -2.00543720e-02 -3.72243673e-01 6.39682496e-03
1.95686713e-01 1.04359317e+00 4.46805269e-01 -1.75755298e+00
-4.03495252e-01 -1.94413051e-01 3.05342436e-01 6.79806530e-01
-5.47089577e-02 2.50016004e-01 -3.54143754e-02 -4.23763573e-01
-3.03739011e-01 -1.04440224e+00 1.73946142e-01 -4.84936625e-01
-1.03152359e+00 -6.91446424e-01 1.05171680e+00 -5.06092131e-01
8.15595686e-01 -2.16366816e+00 3.68554085e-01 -2.89055586e-01
3.21985453e-01 -5.02154976e-02 -3.08718890e-01 8.55372310e-01
7.84271210e-02 -1.65814772e-01 1.12585828e-01 -1.02064633e+00
2.80137807e-01 -7.41593167e-02 -3.27061474e-01 1.03044964e-01
-4.19480801e-01 9.38386321e-01 -8.69953394e-01 -4.50091690e-01
1.38020203e-01 6.10343039e-01 -5.39016008e-01 9.86416280e-01
-1.31178319e-01 1.11723292e+00 -2.12485686e-01 2.55124748e-01
4.30593222e-01 8.04556012e-02 2.63430536e-01 1.82044029e-01
5.49975894e-02 4.30815190e-01 -9.31073606e-01 1.39667976e+00
-8.03298712e-01 5.36279321e-01 7.59151042e-01 -3.49932671e-01
1.08776772e+00 9.79334295e-01 2.51008391e-01 -4.48949993e-01
4.69433993e-01 -1.43925279e-01 1.17785372e-01 -8.08157325e-01
4.62231755e-01 -2.51221001e-01 -5.87528586e-01 9.61636484e-01
-1.41379505e-01 -5.12881398e-01 3.85107100e-02 3.14494342e-01
8.21752846e-01 -2.25080818e-01 1.55001208e-01 1.91502094e-01
5.94934881e-01 -6.15252078e-01 4.38260823e-01 4.76182193e-01
-4.01768118e-01 4.85054493e-01 6.89720571e-01 -1.98865552e-02
-2.48032212e-01 -7.41129696e-01 6.08573735e-01 1.69904590e+00
1.94112614e-01 -2.34317243e-01 -1.16337955e+00 -2.50335664e-01
-4.97982144e-01 1.07365584e+00 -5.22148132e-01 -9.46045071e-02
-4.84609932e-01 -1.54435083e-01 8.21248055e-01 1.42629936e-01
5.03543437e-01 -1.90446413e+00 -4.49539840e-01 4.23185468e-01
-1.18206036e+00 -1.24232042e+00 -9.17369008e-01 -4.71219867e-01
8.89211595e-02 -7.33256400e-01 -4.82466996e-01 -7.96466827e-01
3.11494559e-01 3.69465768e-01 9.68577087e-01 -9.25346389e-02
1.90411165e-01 6.49258316e-01 -5.29799521e-01 -2.93994278e-01
-9.92203832e-01 -2.05976777e-02 2.79409975e-01 3.14989060e-01
-2.45619137e-02 -8.15673470e-01 -4.56412882e-01 8.53552341e-01
-4.24469233e-01 5.87811887e-01 2.60303286e-03 5.52434146e-01
-4.84615833e-01 -5.44041395e-01 1.23746252e+00 -6.23073220e-01
1.42644358e+00 -5.58518231e-01 3.74310464e-01 -6.18491881e-03
3.44632924e-01 -5.31167090e-01 7.10207403e-01 -8.11568558e-01
-1.51811790e+00 -1.25622720e-01 -2.27686301e-01 -4.86870483e-02
-3.92428935e-01 6.57561868e-02 -7.54746616e-01 4.89889175e-01
5.34484088e-01 1.40171349e-01 2.77816296e-01 7.44866058e-02
6.53006315e-01 1.31313992e+00 6.93683982e-01 -6.44295275e-01
2.83942044e-01 1.54396310e-01 -7.78595984e-01 -7.96357512e-01
-3.78249884e-01 -3.80385339e-01 -1.73070088e-01 -8.28870058e-01
7.82309055e-01 -8.14810216e-01 -1.78820920e+00 7.14229524e-01
-1.64033067e+00 -7.84238219e-01 2.49315053e-02 1.34029955e-01
-1.01987696e+00 7.08515197e-02 -7.21629143e-01 -1.17831171e+00
-6.73415542e-01 -1.41016626e+00 1.15956080e+00 2.90359259e-01
-9.83433068e-01 -8.49760652e-01 2.07937092e-01 1.10813642e+00
3.58938992e-01 2.57368654e-01 3.36880803e-01 -6.10980690e-01
1.05931930e-01 1.28843963e-01 2.02760383e-01 -3.04283164e-02
2.89471239e-01 -4.69035804e-02 -1.34020185e+00 -8.81985873e-02
1.13982342e-01 -8.89436901e-01 -6.70124665e-02 2.53801852e-01
5.50436139e-01 -8.97332251e-01 -2.91333199e-01 1.46058187e-01
2.99814250e-02 4.75341290e-01 6.36812508e-01 -4.60697562e-01
6.28289104e-01 1.50969374e+00 6.00043833e-01 7.52887845e-01
1.09455168e+00 8.48751307e-01 2.92622894e-01 1.41837403e-01
-5.41411154e-02 -4.81297314e-01 8.14146876e-01 9.73898888e-01
-6.38350397e-02 -8.12724590e-01 -5.39080560e-01 4.02270466e-01
-2.05681610e+00 -1.25206614e+00 -1.54682279e-01 1.62401688e+00
7.92747200e-01 -3.85771543e-01 4.89319712e-01 1.10313237e-01
1.02831376e+00 3.45144659e-01 -5.15584886e-01 -8.46795261e-01
-5.60470633e-02 -5.71153462e-01 -7.59342432e-01 9.07358468e-01
-5.61311603e-01 9.11600292e-01 5.12652397e+00 5.80203474e-01
-1.04864073e+00 7.27813691e-02 7.60623634e-01 -2.75328815e-01
-3.15537125e-01 -4.39904481e-01 -6.25351429e-01 4.54754442e-01
8.29143107e-01 -2.20500514e-01 8.56780708e-01 5.96493006e-01
8.93941104e-01 8.11819136e-02 -1.55845344e+00 1.34175730e+00
4.46904123e-01 -8.48037362e-01 -6.50588125e-02 -1.10056296e-01
4.95211929e-01 -6.52512670e-01 -1.60812791e-02 5.27010679e-01
3.42782497e-01 -1.24519920e+00 8.40024650e-01 2.39032328e-01
6.77711368e-01 -6.75026059e-01 5.56163430e-01 7.55218685e-01
-1.20553684e+00 -1.46957532e-01 4.97665793e-01 -3.48173708e-01
9.64943290e-01 1.30612820e-01 -1.08863163e+00 1.12557247e-01
3.59985232e-01 3.41560394e-01 2.62322545e-01 2.01614499e-01
-3.19166183e-01 3.11136454e-01 2.00804248e-02 -2.26009876e-01
-6.69925958e-02 -1.33821383e-01 9.43209469e-01 9.72460926e-01
2.10623249e-01 4.48905706e-01 1.98543161e-01 8.86582196e-01
-2.97996581e-01 5.81361540e-02 -7.32217968e-01 1.66632175e-01
8.88891816e-01 1.37630463e+00 -3.39600116e-01 -1.67794406e-01
1.59027249e-01 1.06071854e+00 1.26480117e-01 4.05847490e-01
-8.56390476e-01 -2.90351599e-01 6.56985700e-01 4.70876209e-02
-4.64625269e-01 2.13135839e-01 3.13480310e-02 -6.58467770e-01
-4.24600504e-02 -1.41268325e+00 1.48288840e-02 -1.19582117e+00
-1.16213024e+00 9.28528011e-01 -1.55618563e-01 -1.04845774e+00
-9.20443594e-01 2.17176974e-01 -1.12172306e+00 5.96782863e-01
-6.60344183e-01 -1.48002207e+00 -5.94373345e-01 5.61866760e-01
9.15383160e-01 -6.26460761e-02 8.14964890e-01 5.45041300e-02
-5.84682047e-01 6.98105037e-01 -6.46048427e-01 -1.52409062e-01
9.36753869e-01 -7.45074034e-01 1.35685056e-01 2.88473722e-02
-5.46196103e-01 4.72506016e-01 9.67641532e-01 -1.73432007e-01
-1.38152337e+00 -7.79996097e-01 1.00135338e+00 -3.71359050e-01
1.88300937e-01 -8.10823262e-01 -5.63596666e-01 7.92568624e-01
6.20053828e-01 -6.30851090e-01 9.08431053e-01 -4.48344946e-02
1.26433924e-01 1.20999724e-01 -1.41826367e+00 7.31011629e-01
1.24434459e+00 -3.66125226e-01 -5.02402902e-01 4.19219077e-01
1.10078561e+00 -3.62404108e-01 -4.69271898e-01 7.80125111e-02
5.35838723e-01 -1.24217975e+00 6.61254466e-01 -1.72852516e-01
5.32908916e-01 -3.60363573e-02 1.06499568e-01 -1.54294038e+00
2.61531860e-01 -1.69185376e+00 1.68451667e-01 1.82136393e+00
2.09768519e-01 -9.27815139e-01 2.62600809e-01 8.32797706e-01
-4.07842159e-01 -3.28748912e-01 -8.64038765e-01 -2.87193716e-01
-2.25608349e-01 -3.45997453e-01 8.50433052e-01 9.32990313e-01
8.35966766e-01 9.69199359e-01 -7.89676726e-01 -1.06997959e-01
7.94508830e-02 -9.50239692e-03 1.55920732e+00 -1.10108721e+00
-3.65925074e-01 -4.52725112e-01 2.78668791e-01 -1.37940538e+00
6.54011309e-01 -6.02744043e-01 6.59937799e-01 -1.52870560e+00
1.68646708e-01 -2.70398438e-01 9.24519420e-01 2.37446338e-01
-6.11215942e-02 -5.85110076e-02 3.67184699e-01 -3.21669062e-03
-5.55716574e-01 9.86995518e-01 1.58247304e+00 7.30481446e-02
-7.33305812e-01 2.04471022e-01 -9.50107634e-01 6.87250972e-01
6.22203767e-01 3.24967057e-02 -7.52824724e-01 -3.75749753e-03
6.39434811e-03 9.03697610e-01 2.63715178e-01 -6.54698312e-01
3.05178940e-01 -2.17038065e-01 -5.42024136e-01 -3.10240388e-01
9.10909474e-01 -2.07091585e-01 3.23686987e-01 1.39856935e-02
-5.46984792e-01 -2.74415463e-01 -4.35130596e-02 1.28450513e-01
-1.15919158e-01 2.36628801e-01 5.55163622e-01 2.50985846e-03
-6.37149718e-03 -1.77466795e-01 -8.56559396e-01 1.43317550e-01
1.23460150e+00 -1.92118779e-01 -6.71591938e-01 -1.64485645e+00
-7.91531801e-01 6.72479570e-01 2.33947709e-01 7.43977606e-01
5.41457653e-01 -1.27640820e+00 -9.02560472e-01 -1.89247608e-01
9.60905850e-02 1.31398961e-01 6.67784929e-01 6.36177659e-01
-1.49428591e-01 2.56904781e-01 -4.88401838e-02 -4.84275639e-01
-1.50625110e+00 2.31436610e-01 3.56173187e-01 -5.56077622e-02
-3.33388627e-01 9.86352026e-01 7.94933736e-01 -1.02580714e+00
4.79836524e-01 7.53073990e-02 -4.28175867e-01 2.71215856e-01
7.59681463e-01 6.25733614e-01 -5.84581316e-01 -1.25335777e+00
-1.39150769e-01 1.44961119e-01 2.57971138e-01 -5.33916056e-01
9.64071691e-01 -5.58485508e-01 -2.05080301e-01 5.63357413e-01
1.06678331e+00 4.62097302e-02 -9.45253074e-01 8.56169686e-02
-7.19572961e-01 -4.45651673e-02 -5.14267981e-01 -7.54219413e-01
-6.98098481e-01 7.31377840e-01 -3.48305106e-02 4.91277635e-01
8.04565966e-01 3.57839912e-01 1.19120598e+00 2.14301959e-01
3.36452305e-01 -9.06420887e-01 5.63315153e-01 5.44325948e-01
1.69694662e+00 -1.15182281e+00 -7.10695624e-01 -6.62496030e-01
-1.30939126e+00 6.77265584e-01 9.67607439e-01 6.57531142e-01
2.10651278e-01 2.77483016e-01 5.37556589e-01 -7.85389915e-02
-1.23949814e+00 1.35464385e-01 -1.06787249e-01 9.82495725e-01
5.36865532e-01 4.60629582e-01 6.13335595e-02 9.72905576e-01
-1.01151752e+00 -2.31510594e-01 6.71387196e-01 4.93523508e-01
-6.40185475e-02 -7.72454441e-01 -5.55258691e-01 -3.26557904e-02
1.95941091e-01 3.03070217e-01 -1.06156945e+00 3.09409469e-01
-6.46886602e-02 2.07553911e+00 3.40662748e-02 -5.33667922e-01
4.59118545e-01 4.37889174e-02 -1.50204394e-02 -6.26658618e-01
-9.20510590e-01 6.55341074e-02 6.93601489e-01 -4.87846792e-01
-5.26073277e-01 -5.02752602e-01 -1.30993664e+00 -7.44405210e-01
-1.64571717e-01 2.41961926e-01 3.01688164e-01 9.58600163e-01
4.09266412e-01 4.46072340e-01 1.17100894e+00 -1.66977310e+00
-1.53313637e-01 -1.53695834e+00 -2.69860804e-01 5.94323814e-01
2.64171869e-01 -4.61263597e-01 -6.94833875e-01 1.02981534e-02] | [12.933695793151855, 7.8924241065979] |
35981c03-00b9-4e91-9b28-39a7e0dceff7 | challenges-in-domain-specific-abstractive | 2307.00963 | null | https://arxiv.org/abs/2307.00963v1 | https://arxiv.org/pdf/2307.00963v1.pdf | Challenges in Domain-Specific Abstractive Summarization and How to Overcome them | Large Language Models work quite well with general-purpose data and many tasks in Natural Language Processing. However, they show several limitations when used for a task such as domain-specific abstractive text summarization. This paper identifies three of those limitations as research problems in the context of abstractive text summarization: 1) Quadratic complexity of transformer-based models with respect to the input text length; 2) Model Hallucination, which is a model's ability to generate factually incorrect text; and 3) Domain Shift, which happens when the distribution of the model's training and test corpus is not the same. Along with a discussion of the open research questions, this paper also provides an assessment of existing state-of-the-art techniques relevant to domain-specific text summarization to address the research gaps. | ['Florian Matthes', 'Daniel Braun', 'Juraj Vladika', 'Anum Afzal'] | 2023-07-03 | null | null | null | null | ['abstractive-text-summarization', 'text-summarization'] | ['natural-language-processing', 'natural-language-processing'] | [ 5.92856228e-01 5.38390756e-01 -3.37031126e-01 -5.15772142e-02
-1.01582325e+00 -4.50870663e-01 7.36221671e-01 4.92357761e-01
-3.46331477e-01 1.18311512e+00 9.08992231e-01 -2.81597704e-01
8.59973356e-02 -3.99353594e-01 -4.05886799e-01 -2.07763717e-01
4.15054500e-01 9.14386749e-01 3.68299410e-02 -4.82748002e-01
9.85921204e-01 9.87930968e-02 -1.39238894e+00 4.65614140e-01
1.47787344e+00 2.89830327e-01 2.50085831e-01 9.71080780e-01
-4.63787019e-01 9.28435147e-01 -1.46565819e+00 -2.36240178e-01
-2.53534198e-01 -8.78078282e-01 -1.22821844e+00 1.60324693e-01
8.11824799e-01 -1.38908669e-01 -2.39048123e-01 1.11749208e+00
7.78494656e-01 3.67527902e-01 9.23478723e-01 -1.00472283e+00
-7.33932316e-01 6.97842419e-01 -2.82692373e-01 6.08378828e-01
7.82032192e-01 -1.72159765e-02 8.33313525e-01 -4.91253436e-01
6.78142309e-01 1.49479342e+00 4.85721946e-01 8.48192155e-01
-1.08582330e+00 -3.53438169e-01 2.32760042e-01 1.16859168e-01
-8.29325080e-01 -6.54999912e-01 6.39993846e-01 -3.27505946e-01
1.59818017e+00 5.32226443e-01 4.72212732e-01 1.15040565e+00
8.06242406e-01 9.12811399e-01 6.42877400e-01 -4.51638460e-01
3.62436384e-01 2.08083063e-01 2.74386108e-01 7.56906793e-02
6.79512322e-01 -4.80076283e-01 -7.51328766e-01 -3.88657093e-01
2.93066740e-01 -7.98394501e-01 -2.62216985e-01 2.09100857e-01
-1.01484609e+00 9.82035220e-01 -3.75653654e-01 5.16833127e-01
-5.34160256e-01 -1.77905351e-01 7.01653123e-01 4.61517811e-01
9.56529081e-01 1.15154600e+00 -5.78013837e-01 -4.39011276e-01
-1.57045090e+00 7.69611955e-01 1.15788579e+00 1.18073308e+00
2.82855034e-01 6.05923712e-01 -4.16105151e-01 9.33040619e-01
-1.01014152e-01 4.30756539e-01 1.04523623e+00 -9.21782911e-01
1.01976943e+00 5.30526996e-01 2.69837558e-01 -7.22146392e-01
-3.41112971e-01 -3.70536208e-01 -8.70149732e-01 -4.57497984e-01
1.24794148e-01 -3.98583740e-01 -5.73716760e-01 1.56265533e+00
-3.03153723e-01 -2.83629179e-01 5.66796303e-01 3.79149705e-01
1.18111408e+00 9.44770932e-01 2.66806856e-02 -8.23287606e-01
1.18773615e+00 -1.05484104e+00 -1.08367825e+00 -9.09064531e-01
7.25345016e-01 -8.64071012e-01 9.98678565e-01 2.40978166e-01
-1.71532321e+00 -3.41562092e-01 -8.40326846e-01 -2.53721684e-01
-2.00595975e-01 1.18522786e-01 2.88127482e-01 3.74416649e-01
-1.10680449e+00 4.85036910e-01 -3.86551112e-01 -8.47168088e-01
1.36353493e-01 1.47847444e-01 1.27574280e-01 1.06223449e-01
-1.23893392e+00 1.41651499e+00 6.31180346e-01 -3.65782022e-01
-2.27435529e-01 -7.71703243e-01 -9.46032524e-01 2.53372371e-01
1.32375523e-01 -1.15834475e+00 1.85405660e+00 -8.86653066e-01
-1.55114019e+00 7.72844136e-01 -6.60591125e-01 -7.44328439e-01
5.06689429e-01 -4.26539928e-01 -3.67657959e-01 1.34259537e-01
4.72132683e-01 4.12422508e-01 5.21964490e-01 -1.05647421e+00
-6.31810606e-01 -3.05244297e-01 -3.95582348e-01 7.11455703e-01
-2.36712143e-01 -5.81251504e-03 1.27464443e-01 -8.63186538e-01
-2.21111640e-01 -4.40631568e-01 -2.85806566e-01 -9.19305682e-01
-7.36550987e-01 -5.75674415e-01 6.98215067e-01 -9.68420684e-01
1.69744098e+00 -1.51900470e+00 2.25362793e-01 -4.81729299e-01
-9.16351527e-02 5.46620190e-01 -1.79220259e-01 1.16751826e+00
7.91661143e-02 4.12747324e-01 -4.08945590e-01 -2.80238956e-01
-9.87793654e-02 -7.44894743e-02 -7.12122619e-01 -5.00049591e-02
1.61528111e-01 8.24014723e-01 -1.06760943e+00 -7.32409000e-01
-8.43268167e-03 -3.79360378e-01 -3.01151842e-01 3.06376047e-03
-4.04372901e-01 1.30764812e-01 -5.57686806e-01 8.38864502e-03
4.84310508e-01 1.29612148e-01 -1.90765597e-02 2.02783301e-01
-2.49115914e-01 8.09590816e-01 -8.25087130e-01 1.72924459e+00
-2.30281293e-01 8.75772178e-01 -1.39455214e-01 -1.02216148e+00
7.84472466e-01 4.93962228e-01 1.27406836e-01 -4.83019322e-01
-8.43421146e-02 2.30522752e-01 -7.14451224e-02 -6.34383619e-01
1.34656036e+00 -2.28014469e-01 -1.49754390e-01 6.39981091e-01
1.60457551e-01 -7.31956005e-01 7.73534417e-01 5.81890881e-01
1.01069021e+00 -3.00200611e-01 7.77033508e-01 -4.94062394e-01
4.53532279e-01 5.47617435e-01 3.25862974e-01 1.17605698e+00
1.34554654e-01 4.93758053e-01 5.66383660e-01 -7.18157738e-02
-1.12177730e+00 -6.95684850e-01 8.01558420e-02 9.30028915e-01
2.82749962e-02 -5.56067824e-01 -1.00662994e+00 -4.54466522e-01
-2.41929576e-01 1.77560484e+00 -3.94236207e-01 -4.13577914e-01
-5.63831329e-01 -7.43574858e-01 7.64623523e-01 3.66925836e-01
5.74735165e-01 -1.21027422e+00 -7.47022986e-01 4.71342474e-01
-7.12325752e-01 -8.58338356e-01 -5.12091339e-01 -1.17484100e-01
-1.41494322e+00 -8.37910891e-01 -6.53794408e-01 -9.10692155e-01
5.70266366e-01 3.91077161e-01 1.26659811e+00 -1.52341291e-01
1.35820001e-01 5.91977596e-01 -1.49349257e-01 -8.47416341e-01
-1.06567955e+00 4.30585444e-01 -1.77823633e-01 -7.52246261e-01
3.99944276e-01 -2.23014340e-01 -1.62372977e-01 -2.45928124e-01
-1.15948355e+00 2.46488258e-01 4.80523676e-01 8.43961477e-01
6.40137941e-02 1.28831327e-01 1.12963033e+00 -1.06554496e+00
1.83721292e+00 -4.36301321e-01 1.16291441e-01 4.67670500e-01
-6.10929847e-01 5.53213060e-02 8.03202748e-01 -4.24602658e-01
-1.46451485e+00 -6.90683544e-01 -2.69122068e-02 3.85777026e-01
-1.88063398e-01 6.70842469e-01 1.61383644e-01 6.24682426e-01
1.07344472e+00 7.64309347e-01 7.77719170e-02 -2.84547061e-01
1.57007366e-01 7.82311440e-01 4.10910785e-01 -3.24187905e-01
5.33280969e-01 1.15011558e-02 -3.32945436e-01 -1.33004260e+00
-1.05494201e+00 -4.88720417e-01 -4.48121518e-01 1.58913970e-01
5.02319217e-01 -8.01206112e-01 1.90177456e-01 3.95861179e-01
-1.61208367e+00 -2.85417080e-01 -6.37761772e-01 2.90425807e-01
-8.03121209e-01 7.23648071e-01 -6.42607510e-01 -7.15657771e-01
-9.95694816e-01 -6.20924175e-01 9.54833686e-01 4.51043367e-01
-9.35826719e-01 -1.30806959e+00 2.25252002e-01 4.01151568e-01
3.97435695e-01 -3.61680845e-03 1.12873149e+00 -1.10140991e+00
2.91237384e-01 -2.88092315e-01 2.88973987e-01 4.57331628e-01
9.00572315e-02 -9.36666429e-02 -5.50219059e-01 -3.01502109e-01
3.61529082e-01 -3.11435431e-01 7.55285084e-01 7.33464420e-01
7.23101974e-01 -9.69424605e-01 -2.62101501e-01 -2.13298127e-01
9.50406373e-01 8.94030631e-02 6.60452664e-01 3.54173750e-01
2.54542828e-01 7.09802091e-01 8.46953809e-01 4.41359282e-01
3.30803990e-01 2.37491637e-01 -1.72900677e-01 9.64334011e-02
-1.69802621e-01 -4.70447928e-01 3.25044930e-01 1.08721316e+00
1.94474787e-01 -8.06763649e-01 -6.97971165e-01 7.00952888e-01
-1.97917962e+00 -1.38454950e+00 -3.42000037e-01 2.02220416e+00
9.42640245e-01 5.12759648e-02 -1.18314989e-01 -2.02765167e-02
8.54921818e-01 4.02828246e-01 -6.28337383e-01 -1.06284750e+00
-2.76173443e-01 -2.50670630e-02 2.51873642e-01 5.45930207e-01
-6.23927891e-01 1.17366481e+00 7.19964027e+00 8.44850302e-01
-8.74217749e-01 -7.48692006e-02 2.75310785e-01 -1.18343703e-01
-3.86593014e-01 5.27646318e-02 -5.77043831e-01 4.44930792e-01
1.12533855e+00 -1.09259689e+00 9.75714251e-02 5.56703627e-01
5.24389863e-01 -4.16259348e-01 -1.18377590e+00 6.22521996e-01
6.00635529e-01 -1.31114948e+00 6.32328928e-01 -1.51041001e-01
9.98779416e-01 -2.17189714e-01 -2.46140912e-01 4.89297211e-01
9.96672660e-02 -7.97894955e-01 7.37727702e-01 1.82294309e-01
5.59000731e-01 -5.75138807e-01 7.98975170e-01 9.76252913e-01
-4.56785351e-01 3.59480157e-02 -7.60094166e-01 -3.09108913e-01
3.21086615e-01 3.47346604e-01 -1.06407785e+00 6.89435244e-01
1.27936855e-01 6.48428082e-01 -6.05846226e-01 8.92143786e-01
-1.93831235e-01 7.89884508e-01 3.40701751e-02 -2.82790571e-01
1.67281941e-01 -3.27331084e-03 1.12425339e+00 1.55879831e+00
3.61951053e-01 3.72504443e-02 -2.97507346e-02 7.09866703e-01
2.89937165e-02 1.87990576e-01 -8.52220774e-01 -6.73072785e-02
6.31163061e-01 6.28985524e-01 -2.94867486e-01 -7.44076014e-01
8.17195699e-03 1.14238024e+00 -6.83904961e-02 5.26139438e-01
-1.87816903e-01 -4.84120607e-01 4.47039865e-02 1.04418203e-01
-1.33642942e-01 1.51840486e-02 -6.65172100e-01 -1.25501788e+00
3.17380838e-02 -1.14806855e+00 3.44941676e-01 -8.26088190e-01
-1.15755427e+00 3.47308546e-01 4.13355500e-01 -1.11940169e+00
-7.64404356e-01 2.95557752e-02 -1.07718611e+00 7.58209229e-01
-1.16611683e+00 -8.05243015e-01 7.34801963e-02 1.96827143e-01
1.44182158e+00 -4.25053090e-01 9.60402071e-01 -3.58724862e-01
-3.75440508e-01 2.94849992e-01 3.36998761e-01 -3.92211556e-01
7.72747457e-01 -1.28781760e+00 7.24541605e-01 9.12087560e-01
-2.87226349e-01 6.73458695e-01 1.40081000e+00 -9.06671047e-01
-1.06297231e+00 -1.08308220e+00 1.67784882e+00 -4.89151597e-01
3.24608654e-01 9.50925574e-02 -8.45487475e-01 7.68369615e-01
6.02601051e-01 -1.14815021e+00 7.99137712e-01 -4.24070098e-02
2.78678328e-01 2.79928863e-01 -1.19041502e+00 8.44536781e-01
7.26241529e-01 -7.93451220e-02 -1.39481664e+00 7.29242325e-01
6.10512555e-01 -4.86519039e-01 -5.93614221e-01 -1.44236423e-02
1.12253502e-01 -6.72189772e-01 4.96915698e-01 -8.85871828e-01
7.79505670e-01 2.06690863e-01 3.38025540e-01 -1.78920090e+00
-2.80697405e-01 -7.21926630e-01 -2.50139356e-01 1.29102933e+00
4.41221803e-01 -6.91910625e-01 3.92500341e-01 5.19615650e-01
-4.64126676e-01 -3.75943422e-01 -8.96736205e-01 -7.49965012e-01
7.04162598e-01 5.30101880e-02 3.39948565e-01 7.38534808e-01
4.10676628e-01 1.01050246e+00 -2.17059880e-01 -4.26058084e-01
3.26760739e-01 -2.11384624e-01 9.07637179e-01 -1.52446234e+00
3.01530272e-01 -5.73712409e-01 -3.00886482e-03 -1.29321015e+00
4.35181409e-01 -6.49641633e-01 2.65898407e-02 -2.27792859e+00
3.19153249e-01 4.43146408e-01 4.26709414e-01 3.89759690e-02
-1.42061159e-01 -5.76072931e-01 2.46809140e-01 2.82637447e-01
-6.76960170e-01 6.89036250e-01 1.35024524e+00 -3.55393916e-01
-5.25141478e-01 3.03494662e-01 -1.40394282e+00 6.47708297e-01
1.12484276e+00 -6.14033997e-01 -6.46171331e-01 -5.09673595e-01
1.05087236e-01 4.74481165e-01 -3.18160132e-02 -9.26063955e-01
4.67609465e-01 -3.21630180e-01 1.05717666e-01 -7.55235732e-01
2.14416590e-02 -1.99111804e-01 -2.89855599e-01 5.66347599e-01
-8.68916988e-01 3.07421863e-01 4.84623313e-01 4.84045118e-01
-3.83542478e-01 -7.69524395e-01 5.91173947e-01 -4.12587672e-01
-3.44407409e-01 -3.67008060e-01 -1.19621038e+00 7.08505869e-01
7.67553151e-01 -5.62058032e-01 -7.03633547e-01 -7.91035116e-01
-1.26209557e-01 4.26246881e-01 4.91734356e-01 5.63117862e-01
5.53808808e-01 -7.87604511e-01 -1.20863140e+00 -3.08251083e-01
-3.57959489e-03 1.34494141e-01 3.46581876e-01 4.80355382e-01
-3.14320415e-01 9.30990398e-01 -2.51387283e-02 -2.31583938e-01
-1.30192089e+00 1.49402946e-01 1.30749345e-01 -7.38058269e-01
-4.95937645e-01 3.00801724e-01 4.75660339e-02 -3.42378378e-01
-2.15574391e-02 -2.18671814e-01 -3.31375033e-01 1.15940005e-01
6.59655929e-01 6.88970387e-01 1.44365877e-01 -4.28380430e-01
-5.70318401e-02 1.39519945e-01 -3.50206792e-01 -4.22588199e-01
1.02071416e+00 -2.18114674e-01 -1.99142545e-01 6.44903541e-01
6.87868357e-01 -1.69901118e-01 -6.51505649e-01 -2.29851022e-01
1.31846905e-01 9.31571871e-02 -1.37369588e-01 -1.08568060e+00
-1.15576543e-01 7.88061321e-01 -1.65698498e-01 3.58017117e-01
9.63447630e-01 -1.81968823e-01 7.19405413e-01 7.36573577e-01
-1.11967191e-01 -1.80395567e+00 1.33210659e-01 1.03721356e+00
1.31508362e+00 -8.17293584e-01 3.48168314e-01 1.04986392e-02
-1.06301248e+00 1.19760060e+00 6.70538664e-01 1.41153365e-01
-1.15552865e-01 -1.03416443e-01 -9.04807821e-02 -1.16370775e-01
-1.11026394e+00 2.29335204e-01 1.69203550e-01 7.10074604e-01
5.74419796e-01 -7.47476816e-02 -1.01362717e+00 5.25020599e-01
-7.27995872e-01 -1.50255077e-02 1.03660941e+00 1.01384366e+00
-8.35279524e-01 -8.17196906e-01 -3.87027889e-01 9.07604218e-01
-5.66131353e-01 -2.44926527e-01 -9.20149386e-01 7.21649706e-01
-6.18604898e-01 1.48626709e+00 1.16750777e-01 1.31199583e-01
4.47100729e-01 3.12770069e-01 3.79436523e-01 -1.24117434e+00
-9.28761363e-01 -1.75020486e-01 4.20646071e-01 7.03070909e-02
-2.05541387e-01 -7.96704590e-01 -1.18440139e+00 -4.07786548e-01
-4.03277040e-01 6.38538063e-01 5.31528890e-01 1.00501597e+00
4.33959961e-01 4.92404908e-01 1.31277010e-01 -7.96633244e-01
-1.09694135e+00 -1.64376187e+00 -4.70762819e-01 3.48708808e-01
2.73013294e-01 -3.79138142e-02 -3.57257128e-01 1.18321702e-01] | [12.406805992126465, 9.464607238769531] |
841d1e0f-1286-4d7e-bb02-43592b685588 | face-image-retrieval-with-attribute | null | null | http://openaccess.thecvf.com//content/ICCV2021/html/Zaeemzadeh_Face_Image_Retrieval_With_Attribute_Manipulation_ICCV_2021_paper.html | http://openaccess.thecvf.com//content/ICCV2021/papers/Zaeemzadeh_Face_Image_Retrieval_With_Attribute_Manipulation_ICCV_2021_paper.pdf | Face Image Retrieval With Attribute Manipulation | Current face image retrieval solutions are limited, since they treat different facial attributes the same and cannot incorporate user's preference for a subset of attributes in their search criteria. This paper introduces a new face image retrieval framework, where the input face query is augmented by both an adjustment vector that specifies the desired modifications to the facial attributes, and a preference vector that assigns different levels of importance to different attributes. For example, a user can ask for retrieving images similar to a query image, but with a different hair color, and no preference for absence/presence of eyeglasses in the results. To achieve this, we propose to disentangle the semantics, corresponding to various attributes, by learning a set of sparse and orthogonal basis vectors in the latent space of StyleGAN. Such basis vectors are then employed to decompose the dissimilarity between face images in terms of dissimilarity between their attributes, assign preference to the attributes, and adjust the attributes in the query. Enforcing sparsity on the basis vectors helps us to disentangle the latent space and adjust each attribute independently from other attributes, while enforcing orthogonality facilitates preference assignment and the dissimilarity decomposition. The effectiveness of our approach is illustrated by achieving state-of-the-art results for the face image retrieval task. | ['Ratheesh Kalarot', 'Mubarak Shah', 'Nazanin Rahnavard', 'Zhe Lin', 'Baldo Faieta', 'Shabnam Ghadar', 'Alireza Zaeemzadeh'] | 2021-01-01 | null | null | null | iccv-2021-1 | ['face-image-retrieval'] | ['computer-vision'] | [ 1.49937302e-01 -3.78049016e-01 -4.27733064e-01 -7.31936157e-01
-2.76827186e-01 -7.47959852e-01 5.59012711e-01 -1.91624518e-02
-1.94164604e-01 2.61395872e-01 3.72339785e-01 3.08316946e-01
-5.61653674e-01 -6.07727766e-01 -1.49685919e-01 -7.84644127e-01
2.04137594e-01 8.14105928e-01 -3.83403838e-01 -3.28704752e-02
2.82359511e-01 8.34497094e-01 -1.96156752e+00 3.64342004e-01
3.60282958e-01 1.11186266e+00 5.71172498e-02 1.90438181e-01
-5.76782674e-02 1.82009563e-01 -3.71058643e-01 -3.76091182e-01
3.88822824e-01 -1.09013185e-01 -7.41165400e-01 5.28921008e-01
8.24754238e-01 -4.44469184e-01 -3.80146444e-01 1.06315804e+00
4.21499431e-01 1.83865532e-01 1.10531151e+00 -1.52486432e+00
-1.07940972e+00 8.75183474e-03 -5.86338818e-01 -8.10395777e-02
4.80867207e-01 -3.59554708e-01 1.40461648e+00 -1.06778884e+00
5.89469790e-01 1.39297569e+00 -1.25748932e-01 7.23012805e-01
-1.61010802e+00 -7.35979259e-01 2.70594537e-01 3.61074239e-01
-1.79951811e+00 -8.39872181e-01 1.07266521e+00 -4.15791184e-01
4.29455906e-01 6.29105985e-01 4.75754082e-01 8.78208876e-01
-1.94812045e-01 5.96294165e-01 7.67611086e-01 -2.91032404e-01
1.12732589e-01 3.75124544e-01 -7.67778233e-02 6.57689512e-01
5.25951274e-02 -1.04261793e-01 -7.45981395e-01 -6.26709342e-01
6.82078958e-01 3.81821841e-01 -2.45261088e-01 -9.88567352e-01
-1.00651574e+00 9.51353788e-01 2.74139822e-01 1.25395209e-01
-4.81535405e-01 -2.43354380e-01 7.31347501e-02 3.84123832e-01
2.00585619e-01 2.31577069e-01 -2.83930957e-01 5.41842043e-01
-7.90778756e-01 2.09070206e-01 5.14977872e-01 1.01562452e+00
1.11364532e+00 -3.29808623e-01 -4.74234551e-01 1.17301762e+00
6.74158573e-01 5.60815573e-01 4.38205004e-01 -1.25883424e+00
1.02150522e-01 5.97923160e-01 1.66233942e-01 -1.41187346e+00
2.01253928e-02 -1.01709692e-02 -7.79968798e-01 5.74391522e-02
3.05566425e-03 4.57778573e-01 -9.51396644e-01 2.21088457e+00
2.99719006e-01 -7.37393126e-02 -1.51173905e-01 9.68683779e-01
6.02639914e-01 4.20263976e-01 -1.08115584e-01 -3.34473640e-01
1.60097754e+00 -5.55840135e-01 -7.10300148e-01 -7.84901977e-02
2.16761932e-01 -1.00658286e+00 1.05435312e+00 2.25594297e-01
-1.09841216e+00 -5.27875125e-01 -8.32116127e-01 -5.08972183e-02
-1.88417852e-01 3.38522762e-01 4.76323158e-01 6.07403338e-01
-1.22880971e+00 9.01898146e-02 -2.84174651e-01 -3.42454106e-01
9.80498120e-02 6.91260099e-01 -7.69370794e-01 -2.26981238e-01
-1.11528909e+00 5.95803022e-01 7.06687346e-02 -6.38815537e-02
-6.45162642e-01 -4.80230600e-01 -9.01151240e-01 3.18765849e-01
3.27252239e-01 -8.42618763e-01 6.86531365e-01 -1.00186169e+00
-1.24218166e+00 1.29242337e+00 -4.40777689e-01 3.70468765e-01
9.14974809e-02 1.25390559e-01 -5.54968596e-01 3.05495650e-01
3.56721908e-01 9.19977963e-01 1.42265594e+00 -1.34815180e+00
-4.95616525e-01 -8.36913526e-01 1.15493938e-01 3.27448756e-01
-6.77609503e-01 1.76848084e-01 -9.37950194e-01 -8.12178612e-01
4.44842577e-01 -1.19877911e+00 1.59383774e-01 3.47653925e-01
-7.32345283e-02 -2.28400841e-01 1.02299893e+00 -2.47094557e-01
1.24990547e+00 -2.38692164e+00 6.29507303e-01 7.04434335e-01
2.04384238e-01 -2.65371561e-01 -5.88912845e-01 1.52342752e-01
-3.06214839e-01 3.61104384e-02 -1.36244103e-01 -2.88020790e-01
1.22603439e-01 5.26614249e-01 -1.15751497e-01 4.36937243e-01
1.60299301e-01 3.89930844e-01 -8.26677382e-01 -4.89333868e-01
-6.98193908e-02 6.36323452e-01 -8.22420418e-01 2.66282707e-01
2.03072712e-01 1.38899967e-01 -6.18938029e-01 7.29045272e-01
7.19556093e-01 -1.37265891e-01 3.38431090e-01 -4.99384940e-01
2.67548382e-01 -1.26463026e-01 -1.63517427e+00 1.51235712e+00
-1.12128183e-01 2.55618274e-01 2.79217243e-01 -8.21018696e-01
8.70056629e-01 4.94241297e-01 7.71040797e-01 -4.44641501e-01
-1.37533009e-01 9.56812650e-02 -1.52025670e-01 -4.49199677e-01
3.56298000e-01 -6.10434674e-02 8.38733837e-02 7.16391444e-01
7.63715506e-02 -4.97338967e-03 1.79143548e-01 2.61225134e-01
5.22099435e-01 -3.37728590e-01 1.05921119e-01 -3.72087777e-01
9.71396804e-01 -7.28836298e-01 4.81785536e-01 4.35528100e-01
-6.97381794e-02 6.11075222e-01 3.55285525e-01 -5.73409557e-01
-8.37535203e-01 -1.06241155e+00 -2.07706824e-01 1.38808131e+00
1.63878247e-01 -4.50250089e-01 -4.44268584e-01 -6.52202129e-01
2.28388473e-01 2.21909210e-01 -8.21619093e-01 -3.75399172e-01
-3.15705121e-01 -4.48508739e-01 2.49958292e-01 2.44712189e-01
1.86164156e-01 -9.40967321e-01 -2.90657371e-01 -2.62124866e-01
-3.86076421e-01 -7.50076056e-01 -1.13537180e+00 -1.92117602e-01
-4.94621992e-01 -9.12523329e-01 -6.28909111e-01 -8.45536709e-01
1.14787006e+00 4.46874171e-01 9.87793088e-01 3.58442426e-01
-3.20861131e-01 7.20126748e-01 -2.04451412e-01 1.09689020e-01
5.76024130e-02 -1.76197246e-01 5.42701364e-01 7.56528437e-01
4.44377810e-01 -3.81469160e-01 -6.63033545e-01 5.78155160e-01
-1.35880077e+00 -4.87272412e-01 3.58033359e-01 9.96835768e-01
5.78770995e-01 -1.49916500e-01 1.36576638e-01 -7.06927180e-01
5.56708694e-01 -4.94934946e-01 -2.59377629e-01 4.78902727e-01
-6.29526317e-01 4.23347831e-01 5.54886311e-02 -7.00321376e-01
-7.65362442e-01 1.16203301e-01 2.53338009e-01 -6.57287359e-01
-9.48783848e-03 3.61024767e-01 -6.22881651e-01 -1.32677048e-01
3.56646270e-01 7.17083886e-02 7.41562024e-02 -3.75407606e-01
4.10508305e-01 4.12884742e-01 2.47239217e-01 -9.00450230e-01
6.34217858e-01 5.35215974e-01 4.86077890e-02 -6.63649261e-01
-6.41137183e-01 -6.47565663e-01 -8.81162465e-01 3.85118239e-02
6.55248940e-01 -6.89222753e-01 -6.34245872e-01 1.43827394e-01
-9.14243639e-01 5.03561676e-01 -1.94317892e-01 2.82090127e-01
-4.96165603e-01 4.97011930e-01 -3.15967798e-01 -4.04788792e-01
6.04444034e-02 -1.34464633e+00 1.12406099e+00 -1.55891888e-02
-5.28469086e-01 -6.18990958e-01 -1.82581231e-01 3.29411328e-01
3.64683658e-01 -3.15485507e-01 1.45899606e+00 -5.51659942e-01
-4.23483998e-01 -2.18690485e-01 -2.15896502e-01 9.41529348e-02
6.97230756e-01 5.95423654e-02 -8.34714055e-01 -6.40669227e-01
-1.61360148e-02 -1.99442700e-01 7.34362423e-01 1.97400272e-01
1.26735139e+00 -6.84287667e-01 -2.91918397e-01 5.26360273e-01
1.09570122e+00 1.18514463e-01 3.67771298e-01 7.42097646e-02
6.20157540e-01 9.10330117e-01 3.66142690e-01 5.93213558e-01
1.77562550e-01 1.01789904e+00 3.94811124e-01 6.07852116e-02
2.07898468e-01 -5.63758612e-03 1.70995444e-01 5.09971201e-01
-3.25315483e-02 1.84387818e-03 -4.19649452e-01 4.96622354e-01
-1.69890559e+00 -9.77294385e-01 4.69381154e-01 2.45121908e+00
8.73690784e-01 -4.92849410e-01 4.90943827e-02 -1.59197226e-02
7.84307539e-01 3.21763694e-01 -5.91024935e-01 -2.76686072e-01
-9.96225253e-02 3.59553993e-02 -1.47861302e-01 7.63432562e-01
-9.85063434e-01 8.22701216e-01 6.42572021e+00 6.25513732e-01
-1.13636112e+00 -1.99338526e-01 5.33302307e-01 -3.15151453e-01
-8.55324686e-01 5.93677051e-02 -6.19382322e-01 2.06331164e-01
1.79846734e-01 -7.01741427e-02 7.50075459e-01 4.83391792e-01
-2.16259450e-01 3.36801529e-01 -1.52693272e+00 1.20773327e+00
4.92794335e-01 -9.01644647e-01 8.99150312e-01 1.74068660e-01
5.18036544e-01 -6.14408851e-01 5.57989717e-01 8.29791743e-03
-1.37063563e-01 -9.35678542e-01 6.17006600e-01 5.84697187e-01
8.36632729e-01 -5.81809044e-01 2.58194417e-01 1.40979635e-02
-1.05328548e+00 -7.09024891e-02 -5.10749638e-01 1.89014956e-01
-4.59096879e-01 2.28259116e-02 -4.33316380e-01 1.38931632e-01
7.38406599e-01 3.29532981e-01 -5.42507350e-01 4.97705847e-01
-5.78241013e-02 -1.47673577e-01 -3.21733087e-01 4.09493029e-01
-5.47479317e-02 -5.69783688e-01 5.36041260e-01 8.24353516e-01
5.27601719e-01 3.88821006e-01 3.07123214e-01 7.51724660e-01
-2.96611011e-01 3.25802028e-01 -6.03700876e-01 1.67863041e-01
7.09516406e-01 1.27110887e+00 -3.70392293e-01 -3.92331392e-01
-4.34104741e-01 1.23491848e+00 1.86746657e-01 8.21497142e-01
-3.16502213e-01 -8.10468495e-02 1.27270651e+00 1.42747253e-01
1.07367374e-01 4.64176051e-02 1.64517090e-01 -1.17616332e+00
1.25966161e-01 -1.04527044e+00 7.65004337e-01 -7.27108955e-01
-1.39878213e+00 6.84092999e-01 1.32379115e-01 -1.11451602e+00
-2.91179925e-01 -4.71017271e-01 -1.63342774e-01 1.17223001e+00
-1.40713310e+00 -1.15487862e+00 -6.66208789e-02 1.12589407e+00
3.88019621e-01 -5.21760881e-01 1.05207622e+00 3.77103895e-01
-3.07114273e-01 7.71150351e-01 -1.04951680e-01 -3.80312018e-02
1.19421315e+00 -8.14460874e-01 -3.97500604e-01 3.01189393e-01
4.44733471e-01 1.10101271e+00 6.61088407e-01 -2.86087096e-01
-1.43020117e+00 -7.79536009e-01 1.09767199e+00 -4.17171597e-01
4.09124821e-01 -1.23209067e-01 -9.82560158e-01 5.99670351e-01
5.52672371e-02 2.38516644e-01 1.12280834e+00 1.68744817e-01
-8.60780776e-01 -2.93820202e-01 -1.26786566e+00 7.20696867e-01
6.99526250e-01 -8.87573957e-01 -3.67948800e-01 1.01384178e-01
2.57200927e-01 1.08975641e-01 -7.48570979e-01 3.77432853e-01
9.98889804e-01 -7.21908867e-01 1.29445362e+00 -9.86799061e-01
1.04801238e-01 -4.19387877e-01 -6.84947729e-01 -1.06806779e+00
-9.10801709e-01 -1.18437678e-01 9.79221761e-02 1.25783789e+00
2.47578546e-01 -4.98624623e-01 8.50556791e-01 1.00316024e+00
5.19784808e-01 -5.97177088e-01 -9.18818772e-01 -3.44203889e-01
-3.28897089e-01 1.06439188e-01 9.45188820e-01 1.22611868e+00
-2.63055503e-01 3.73393983e-01 -5.61863840e-01 3.96792144e-01
6.54252529e-01 3.82843018e-01 3.72170091e-01 -1.56088161e+00
-1.22294039e-01 -6.18160963e-01 -4.01149988e-01 -7.59131610e-01
5.00544488e-01 -9.13690686e-01 -3.12120616e-01 -9.86872792e-01
5.89632213e-01 -4.72988904e-01 -6.24321342e-01 8.14301372e-01
-1.84824511e-01 5.07636726e-01 3.39280993e-01 5.38782239e-01
-2.61922657e-01 6.05852365e-01 9.37735498e-01 -5.67141116e-01
-1.64478824e-01 -7.14927390e-02 -9.34042573e-01 6.71331048e-01
4.82227296e-01 -4.70722675e-01 -6.03194714e-01 -6.82110190e-01
2.53394127e-01 1.58639401e-01 1.72859311e-01 -4.12802249e-01
2.25829110e-01 -3.56503308e-01 6.54215515e-01 -1.52018458e-01
6.74964786e-01 -1.18482602e+00 1.50012031e-01 5.75424470e-02
-5.02713859e-01 2.45839238e-01 3.94350328e-02 4.31448460e-01
-4.13172573e-01 -2.19237939e-01 6.76564574e-01 -1.80647504e-02
-5.67864895e-01 7.63460040e-01 -2.26184487e-01 -3.34479451e-01
7.13531613e-01 -3.33429515e-01 1.27078965e-01 -4.53533113e-01
-1.07822740e+00 2.15887189e-01 6.87108576e-01 6.86882496e-01
8.57734621e-01 -1.97281671e+00 -7.63190806e-01 8.21868062e-01
5.06719291e-01 -6.61587536e-01 2.01092079e-01 3.61012429e-01
1.26884252e-01 2.20805109e-01 -6.08007729e-01 -6.05506778e-01
-1.76719630e+00 7.88296878e-01 -4.29361947e-02 1.00726858e-01
1.06568061e-01 8.98642600e-01 6.23112142e-01 -4.46339816e-01
2.38738582e-01 2.77455360e-01 -5.20426571e-01 5.49231231e-01
3.60790968e-01 5.72392531e-02 1.15337307e-02 -1.16883481e+00
-4.47037786e-01 9.95367408e-01 -2.62493432e-01 -2.43400052e-01
1.01437223e+00 -2.82692552e-01 -5.81557572e-01 2.22231299e-01
1.35090089e+00 2.16935381e-01 -8.44455779e-01 -6.52870655e-01
-1.07980773e-01 -9.32258010e-01 -2.16630865e-02 -5.00784457e-01
-1.25719786e+00 7.67271876e-01 6.89720750e-01 -1.17821217e-01
1.42491126e+00 2.05157593e-01 2.22518921e-01 4.76946980e-01
2.49772191e-01 -8.90759170e-01 2.97925740e-01 3.47164273e-01
1.21179819e+00 -1.20324099e+00 6.46145083e-03 -4.16582555e-01
-6.01889074e-01 9.97407436e-01 4.75995928e-01 1.16620794e-01
7.99836516e-01 -2.95563787e-01 5.00254184e-02 -3.28064203e-01
-7.24555612e-01 4.91995402e-02 8.11749756e-01 3.30128908e-01
5.42998314e-01 1.14940263e-01 -2.04851598e-01 2.51883805e-01
8.72153416e-02 -5.84881008e-01 -1.28364125e-02 6.54100180e-01
-2.19408855e-01 -1.57632935e+00 -5.79709768e-01 3.42647254e-01
-2.91139066e-01 -5.32226413e-02 -5.70557654e-01 2.69033045e-01
7.07041845e-02 7.66864121e-01 1.23367101e-01 -1.62385568e-01
3.41824204e-01 2.29213163e-01 6.68712080e-01 -7.21608400e-01
-5.68064228e-02 2.85083950e-01 -4.52586144e-01 -4.41891044e-01
-3.68449539e-01 -7.62256682e-01 -8.21356177e-01 -8.49184096e-02
-1.32891148e-01 4.40926075e-01 2.95161754e-01 7.62578070e-01
4.36726093e-01 -1.28059298e-01 8.27716708e-01 -7.55638003e-01
-4.41690147e-01 -6.06246829e-01 -8.44582379e-01 8.67349148e-01
4.41470206e-01 -8.52193356e-01 -1.89828604e-01 3.15253794e-01] | [13.1051607131958, 0.4751087427139282] |
2199282e-0e69-4697-84ba-2be6e2772a9d | bayesian-non-linear-latent-variable-modeling | 2306.08352 | null | https://arxiv.org/abs/2306.08352v1 | https://arxiv.org/pdf/2306.08352v1.pdf | Bayesian Non-linear Latent Variable Modeling via Random Fourier Features | The Gaussian process latent variable model (GPLVM) is a popular probabilistic method used for nonlinear dimension reduction, matrix factorization, and state-space modeling. Inference for GPLVMs is computationally tractable only when the data likelihood is Gaussian. Moreover, inference for GPLVMs has typically been restricted to obtaining maximum a posteriori point estimates, which can lead to overfitting, or variational approximations, which mischaracterize the posterior uncertainty. Here, we present a method to perform Markov chain Monte Carlo (MCMC) inference for generalized Bayesian nonlinear latent variable modeling. The crucial insight necessary to generalize GPLVMs to arbitrary observation models is that we approximate the kernel function in the Gaussian process mappings with random Fourier features; this allows us to compute the gradient of the posterior in closed form with respect to the latent variables. We show that we can generalize GPLVMs to non-Gaussian observations, such as Poisson, negative binomial, and multinomial distributions, using our random feature latent variable model (RFLVM). Our generalized RFLVMs perform on par with state-of-the-art latent variable models on a wide range of applications, including motion capture, images, and text data for the purpose of estimating the latent structure and imputing the missing data of these complex data sets. | ['Barbara E. Engelhardt', 'Gregory W. Gundersen', 'Michael Minyi Zhang'] | 2023-06-14 | null | null | null | null | ['dimensionality-reduction'] | ['methodology'] | [ 4.89965752e-02 -1.47155032e-01 -3.30799013e-01 -6.51685596e-02
-6.79205477e-01 -3.29307377e-01 8.91679168e-01 -2.61858314e-01
-1.34421706e-01 7.69715250e-01 2.65042037e-01 -2.87896872e-01
-2.08164841e-01 -6.70770764e-01 -5.61205208e-01 -9.23152506e-01
2.11242381e-02 8.44117701e-01 -9.56775844e-02 5.83698690e-01
6.84030503e-02 3.57393861e-01 -1.11038983e+00 -4.29223359e-01
6.86585188e-01 3.27737123e-01 1.24338418e-01 7.44662285e-01
1.67371128e-02 4.27565724e-01 -1.47390619e-01 -3.14520895e-01
-2.49820679e-01 -3.44607830e-02 -4.64599967e-01 5.22500239e-02
1.44316452e-02 -4.30553585e-01 -1.90847561e-01 9.77748632e-01
1.08769588e-01 3.32015574e-01 1.44801605e+00 -1.44267690e+00
-5.55455267e-01 2.41486996e-01 -6.97023511e-01 -1.05007097e-01
1.05228983e-01 -5.50487563e-02 7.99207807e-01 -1.09898508e+00
5.35599768e-01 1.76749742e+00 7.09987700e-01 3.30933303e-01
-1.80940974e+00 -5.04193902e-01 1.41039630e-02 -9.64980423e-02
-1.63104796e+00 -4.47504908e-01 5.76947272e-01 -9.99052405e-01
7.54048288e-01 2.80912053e-02 4.02980417e-01 1.36946905e+00
6.39290631e-01 8.32030296e-01 7.68759668e-01 -4.02538210e-01
4.20415878e-01 -1.60583034e-01 2.58265644e-01 6.43267691e-01
2.39009738e-01 5.13936691e-02 -5.96894443e-01 -9.51765656e-01
1.12033188e+00 3.13856930e-01 -9.66487974e-02 -6.35037005e-01
-1.10446429e+00 1.24318671e+00 -5.87480903e-01 -2.44256020e-01
-4.85401005e-01 6.14582121e-01 8.99961591e-02 -1.56487688e-01
6.10368669e-01 -1.96888357e-01 -3.06043804e-01 -4.09898400e-01
-1.12009621e+00 5.09040773e-01 8.10393572e-01 8.28864276e-01
8.14890921e-01 2.20264018e-01 -1.87408626e-01 6.16365910e-01
9.56050694e-01 1.02703214e+00 2.19644949e-01 -1.24874067e+00
2.12306291e-01 -1.57123715e-01 3.82861316e-01 -8.04608524e-01
-8.03910121e-02 -1.65644065e-01 -1.03222084e+00 4.84513864e-02
3.48001748e-01 -1.32843584e-01 -1.10086405e+00 1.81290007e+00
1.58776209e-01 3.22287738e-01 -8.21288675e-02 5.33375323e-01
2.82946974e-01 9.14907992e-01 4.60154712e-02 -6.63372159e-01
1.05216801e+00 -4.81021613e-01 -1.01963866e+00 -4.64386903e-02
2.91369349e-01 -7.28237987e-01 7.10019231e-01 3.44258875e-01
-9.02735531e-01 -3.43475670e-01 -5.18136919e-01 1.17585219e-01
3.02364737e-01 -4.83177602e-02 7.03356862e-01 5.71440697e-01
-8.33739281e-01 5.95582724e-01 -1.67377436e+00 -7.16245472e-02
1.67910263e-01 1.60815835e-01 -2.02716917e-01 -1.61016896e-01
-8.87888968e-01 6.63619876e-01 -1.46987602e-01 1.21672750e-01
-1.10644436e+00 -5.61103761e-01 -9.86271620e-01 1.09201364e-01
2.49924004e-01 -8.61298442e-01 1.13315868e+00 -2.19905362e-01
-1.61065853e+00 2.39729658e-01 -8.92610312e-01 -2.98557699e-01
2.70514369e-01 -1.92489296e-01 -6.99893013e-02 -6.50642160e-03
1.05975650e-01 3.56348574e-01 1.43726885e+00 -9.48050201e-01
3.32394987e-02 -1.68508410e-01 -6.76545441e-01 5.19436970e-02
1.28516808e-01 -1.03300095e-01 -6.29718840e-01 -6.57488763e-01
4.78546113e-01 -1.11104631e+00 -3.34323794e-01 -8.65391642e-03
-2.91064262e-01 -2.44843423e-01 6.58502281e-01 -6.85194969e-01
1.08240521e+00 -2.18643260e+00 6.19729459e-01 2.74181992e-01
2.76638359e-01 -2.59864867e-01 1.33329853e-01 4.70206231e-01
1.44418418e-01 1.38848349e-02 -4.35784817e-01 -8.74797583e-01
1.88165471e-01 4.16098654e-01 -6.07589364e-01 8.30456734e-01
-5.96171804e-02 8.58259737e-01 -7.16403842e-01 -4.80337411e-01
3.29288751e-01 6.43723845e-01 -6.74411118e-01 -1.20506480e-01
-1.39640763e-01 4.82401013e-01 -4.21753347e-01 4.34334576e-01
5.81897736e-01 -4.32568550e-01 5.27602322e-02 -8.69198516e-03
-1.04661971e-01 -6.61457926e-02 -1.45787561e+00 1.32674873e+00
-1.71732381e-01 7.35675573e-01 4.81456332e-02 -6.26819372e-01
5.11474609e-01 5.52227497e-01 6.17605090e-01 3.16985548e-01
-1.81530237e-01 -1.48199752e-01 -4.08825725e-01 -1.84488580e-01
3.85028452e-01 -4.19636041e-01 -1.36095107e-01 5.35139918e-01
2.13002905e-01 -2.07136139e-01 5.51717728e-02 3.33675176e-01
6.90331638e-01 2.58416533e-01 3.23416591e-01 -1.31039232e-01
1.41558042e-02 -4.37161833e-01 7.78817892e-01 9.89849329e-01
8.18192661e-02 5.84221005e-01 5.92066646e-01 6.53652325e-02
-9.98273253e-01 -1.74668682e+00 -4.45066631e-01 5.20291090e-01
-2.29212269e-01 -5.12042284e-01 -3.43852788e-01 3.77449617e-02
1.34113654e-01 7.16622591e-01 -3.49380195e-01 -1.17282085e-01
-2.34132633e-01 -1.16076970e+00 2.29333535e-01 4.21005458e-01
-8.39701965e-02 -4.57605183e-01 -3.59054282e-02 2.78736651e-01
-2.41213292e-01 -9.49522793e-01 -2.92208761e-01 1.35534564e-02
-9.55780089e-01 -6.85891330e-01 -7.64723182e-01 -2.02986374e-01
5.90516388e-01 -1.18548103e-01 8.33774269e-01 -5.16641557e-01
-6.15042001e-02 5.53958416e-01 1.11218989e-01 -1.16889596e-01
-5.32352567e-01 -4.05788302e-01 5.53730428e-01 -8.64481255e-02
3.44746768e-01 -5.94650507e-01 -2.49124274e-01 3.34680945e-01
-8.58185530e-01 1.12180062e-01 3.16125453e-01 9.42981780e-01
7.86132514e-01 2.43716925e-01 7.03335404e-02 -6.66531801e-01
5.25371253e-01 -5.49717486e-01 -8.79025042e-01 8.41221660e-02
-4.83156800e-01 3.84049088e-01 4.39940859e-03 -9.28035855e-01
-1.03369188e+00 5.96726425e-02 1.07147276e-01 -8.42722356e-01
6.75854906e-02 7.82590151e-01 -8.80026817e-02 2.66574830e-01
1.99053451e-01 1.79574773e-01 1.90820321e-01 -6.31366014e-01
4.50614691e-01 4.22954828e-01 4.80150223e-01 -6.01933062e-01
7.30647862e-01 7.69648015e-01 4.34384286e-01 -9.74207461e-01
-4.88380760e-01 -4.47423190e-01 -5.02323687e-01 1.57983944e-01
1.02270865e+00 -1.00393867e+00 -8.77940476e-01 6.05568767e-01
-1.06230748e+00 -3.32579076e-01 -1.99391022e-01 1.02920055e+00
-8.06000650e-01 6.19978309e-01 -7.65336633e-01 -1.28368795e+00
1.75123259e-01 -1.34269845e+00 1.20914412e+00 -1.60159677e-01
-5.54299414e-01 -1.33851051e+00 2.92388648e-01 5.92965297e-02
1.96754307e-01 -9.85278636e-02 1.11342132e+00 5.61844967e-02
-5.57745874e-01 -2.10988715e-01 5.45523539e-02 1.74320281e-01
1.15227766e-01 3.35364163e-01 -5.78650296e-01 -2.74053127e-01
3.06070775e-01 1.84435084e-01 9.87772286e-01 1.21746933e+00
7.49459028e-01 -1.27831250e-01 -5.99144101e-01 6.80407524e-01
1.02627563e+00 -2.09039405e-01 4.48116630e-01 -3.79844308e-01
8.23395073e-01 1.78940818e-01 2.86165118e-01 6.39390469e-01
2.82522917e-01 7.21027851e-01 5.21725677e-02 3.12474102e-01
3.10559720e-01 -2.77047217e-01 5.74926734e-01 7.77956188e-01
1.03772014e-01 -2.74777502e-01 -9.53450620e-01 4.37905312e-01
-2.06092024e+00 -1.03510308e+00 -4.34871107e-01 2.23688889e+00
6.76346421e-01 -2.04723757e-02 -4.38128300e-02 -1.73849314e-01
7.83192098e-01 6.42431900e-03 -5.55231750e-01 1.44658787e-02
5.77853285e-02 -5.79376444e-02 6.98439837e-01 1.05896509e+00
-1.12261534e+00 8.51156294e-01 7.50644207e+00 9.51816499e-01
-5.85327208e-01 2.81276584e-01 8.51221457e-02 -7.41394833e-02
-4.93330270e-01 4.07046467e-01 -1.19564044e+00 5.93201160e-01
9.33459520e-01 -1.38977081e-01 4.71169353e-01 5.35106897e-01
4.71375108e-01 -2.13250503e-01 -8.90017986e-01 1.13898015e+00
-1.72306776e-01 -1.01173675e+00 4.67293151e-02 5.15982389e-01
8.31653237e-01 -4.05751020e-02 2.32076928e-01 2.49087453e-01
7.33450234e-01 -9.20048952e-01 5.77737689e-01 9.92720604e-01
7.70589173e-01 -5.59745491e-01 3.73113245e-01 5.98611534e-01
-8.51720273e-01 1.45085797e-01 -5.20999551e-01 -1.55003488e-01
8.09428334e-01 9.76986289e-01 -6.88850701e-01 4.75795427e-03
2.96926320e-01 5.73431671e-01 1.68426454e-01 8.07660878e-01
-2.34216467e-01 9.50189054e-01 -6.33886158e-01 3.66310477e-01
-2.20023096e-01 -4.84834075e-01 1.01214266e+00 7.16856956e-01
4.21712011e-01 -2.16322705e-01 1.48411855e-01 1.25153327e+00
3.88701171e-01 -3.57065588e-01 -3.12557787e-01 -4.30094391e-01
4.96462673e-01 8.41761351e-01 -6.63223863e-01 -2.97145754e-01
-3.93135309e-01 9.86496866e-01 5.21000512e-02 9.59136724e-01
-7.34233439e-01 2.01361313e-01 6.98863804e-01 1.80489458e-02
5.49652934e-01 -9.64735746e-01 -1.25764817e-01 -1.59775400e+00
-3.22205871e-01 -4.38416690e-01 1.23604372e-01 -7.62499928e-01
-1.35105836e+00 -2.35663913e-03 6.09154284e-01 -8.13846409e-01
-1.02050722e+00 -4.88798946e-01 -3.40275109e-01 1.41091418e+00
-9.99623477e-01 -1.01696599e+00 3.05337906e-01 5.05225837e-01
3.93140227e-01 -1.78436577e-01 8.59730303e-01 -4.11335528e-02
-5.06861389e-01 8.98972154e-02 6.02632344e-01 -1.84947431e-01
6.31635725e-01 -1.08368909e+00 5.79799354e-01 8.86215389e-01
3.25882770e-02 1.09884012e+00 1.16135311e+00 -1.20721745e+00
-1.49896944e+00 -7.87327468e-01 6.31752551e-01 -5.84204257e-01
7.90916443e-01 -4.55938458e-01 -8.35520923e-01 1.01986730e+00
-5.95742047e-01 -8.79229605e-02 7.58527756e-01 4.97315347e-01
-2.86802929e-02 6.26036525e-01 -8.05734515e-01 5.29515088e-01
4.36578989e-01 -6.12688422e-01 -3.16205651e-01 1.97263345e-01
6.23991728e-01 -1.53708354e-01 -7.79940724e-01 2.75397062e-01
6.87852859e-01 -2.79554367e-01 1.08422852e+00 -5.24108469e-01
1.38996631e-01 -3.65835726e-01 -4.22293127e-01 -1.14265347e+00
-5.30137539e-01 -8.12563121e-01 -8.60984266e-01 1.15476871e+00
4.01000604e-02 -4.72311556e-01 7.36670673e-01 8.83882165e-01
2.37608358e-01 -4.18262064e-01 -1.12502325e+00 -7.48273969e-01
1.09046891e-01 -9.08031464e-01 2.83986419e-01 7.54877985e-01
-5.55542409e-01 2.36348733e-01 -7.87095129e-01 3.76983047e-01
1.07468963e+00 7.55243227e-02 8.19959998e-01 -1.38689196e+00
-7.76725292e-01 -9.04012397e-02 -4.42534953e-01 -1.46916556e+00
4.13663805e-01 -6.20235145e-01 4.66577709e-02 -1.47392571e+00
5.69486499e-01 -2.13673666e-01 3.90099078e-01 8.08151364e-02
-3.07317525e-01 -1.33681437e-02 -1.08016416e-01 5.35254538e-01
-2.61537313e-01 8.63875270e-01 9.83301222e-01 9.71019268e-02
-2.96537071e-01 3.84059191e-01 -3.08886558e-01 1.04476047e+00
2.18947098e-01 -6.89460456e-01 -3.28876317e-01 -2.26283103e-01
4.04077977e-01 3.92894477e-01 6.70930564e-01 -4.80082124e-01
1.31388187e-01 -4.55584437e-01 4.03704286e-01 -9.89746094e-01
8.08872163e-01 -5.22888780e-01 4.67406064e-01 1.86946481e-01
-4.91547473e-02 -2.47178018e-01 -4.96638380e-02 9.04940546e-01
4.30160463e-02 -4.11702484e-01 5.52922666e-01 1.68577611e-01
-2.78646857e-01 3.48351181e-01 -9.93660271e-01 -2.92472124e-01
5.14085650e-01 -5.41664194e-03 -1.11666946e-02 -9.13037896e-01
-1.08006704e+00 1.54055595e-01 4.02280807e-01 -4.51035500e-02
6.00840271e-01 -1.45741200e+00 -6.71197116e-01 3.05668503e-01
-3.08892131e-01 2.45346464e-02 3.86879683e-01 8.92726004e-01
-2.07242053e-02 4.00224328e-01 4.26776081e-01 -9.63629723e-01
-1.15919995e+00 5.53940475e-01 2.16405671e-02 -2.70545870e-01
-4.64141905e-01 4.84195322e-01 3.82601291e-01 -3.90364647e-01
-6.16039261e-02 -2.96586484e-01 2.19801411e-01 -1.01106465e-01
3.94280732e-01 6.45795941e-01 -5.53829789e-01 -7.47006774e-01
-1.61159933e-01 5.31080484e-01 8.87287110e-02 -6.88100994e-01
1.11993337e+00 -5.13352394e-01 -7.30609298e-02 9.49397504e-01
1.18443990e+00 -1.76945552e-02 -1.61830997e+00 -2.96803683e-01
-4.11065221e-01 -3.72921526e-01 2.84176290e-01 -3.61090936e-02
-3.58633220e-01 1.03371036e+00 3.32876891e-01 -8.78927261e-02
5.32992959e-01 2.02607259e-01 3.57143134e-01 1.62932932e-01
2.48346671e-01 -7.64803410e-01 -1.72398001e-01 7.60726333e-01
5.39113402e-01 -1.01021850e+00 4.00010020e-01 -4.97348636e-01
-4.15144652e-01 8.92713070e-01 -7.75379837e-02 -2.72051357e-02
1.02600825e+00 3.63855362e-01 -5.28073311e-01 -1.15240715e-01
-7.65488744e-01 2.41319105e-01 4.80496794e-01 5.88781238e-01
2.32610822e-01 2.11803406e-01 -4.89898808e-02 5.93125939e-01
-1.00773878e-01 -7.11476139e-04 5.39454639e-01 7.62148440e-01
-3.09180915e-01 -1.15766025e+00 -6.97198212e-01 4.70378071e-01
-5.52506864e-01 -2.85395831e-01 3.97199631e-01 4.51680064e-01
-4.19115007e-01 7.74177670e-01 3.79027039e-01 1.60717845e-01
-3.68853480e-01 2.37145960e-01 5.25678039e-01 -6.18355572e-01
6.08286738e-01 7.62159407e-01 -2.82480180e-01 -2.60674834e-01
-2.53697366e-01 -1.13906336e+00 -1.05012977e+00 -3.29939365e-01
-5.16117394e-01 8.39329734e-02 8.19155991e-01 1.17458379e+00
3.74706239e-02 2.19830677e-01 1.33469462e-01 -9.63252723e-01
-6.71006322e-01 -1.02081275e+00 -7.44110703e-01 1.76732004e-01
3.76678616e-01 -9.88008976e-01 -6.16299093e-01 2.88449615e-01] | [6.909604549407959, 3.833765983581543] |
53e373cc-983c-4b85-bcd1-120dde04896b | using-wiktionary-to-create-specialized | null | null | https://aclanthology.org/2022.lrec-1.370 | https://aclanthology.org/2022.lrec-1.370.pdf | Using Wiktionary to Create Specialized Lexical Resources and Datasets | This paper describes an approach aiming at utilizing Wiktionary data for creating specialized lexical datasets which can be used for enriching other lexical (semantic) resources or for generating datasets that can be used for evaluating or improving NLP tasks, like Word Sense Disambiguation, Word-in-Context challenges, or Sense Linking across lexicons and dictionaries. We have focused on Wiktionary data about pronunciation information in English, and grammatical number and grammatical gender in German. | ['Thierry Declerck', 'Lenka Bajčetić'] | null | null | null | null | lrec-2022-6 | ['word-sense-disambiguation'] | ['natural-language-processing'] | [-2.81442571e-02 4.47908759e-01 -4.07341242e-01 -4.65899020e-01
-3.20746809e-01 -6.21641636e-01 6.68437660e-01 1.17224252e+00
-9.95131195e-01 1.17908633e+00 7.02000618e-01 -2.33255818e-01
-1.93526432e-01 -9.47894573e-01 -1.38907716e-01 -1.58578679e-01
4.37483460e-01 7.87870407e-01 1.85145453e-01 -7.32770503e-01
4.24080729e-01 -3.13380100e-02 -1.85797310e+00 8.16592202e-02
1.01938689e+00 6.16940379e-01 6.18063271e-01 1.35595852e-03
-9.28194046e-01 2.65837997e-01 -7.61645615e-01 -7.52825260e-01
-5.48729897e-01 -4.32356559e-02 -8.91031802e-01 -7.54675865e-01
3.10738206e-01 7.39116609e-01 1.60398021e-01 1.48425591e+00
4.75921005e-01 1.63420931e-01 4.66438532e-01 -6.76811278e-01
-7.23003924e-01 1.49074173e+00 4.46730375e-01 2.88006574e-01
7.59530365e-01 -5.61328709e-01 9.16570961e-01 -4.95006233e-01
1.47750533e+00 1.38029778e+00 3.55529785e-01 7.72875011e-01
-7.50099838e-01 -4.91285086e-01 3.22380327e-02 2.97814399e-01
-1.39652812e+00 -2.07984015e-01 5.23360252e-01 -3.09586853e-01
1.32031441e+00 5.54402173e-02 5.53307235e-01 1.28324914e+00
-3.49377006e-01 4.09080267e-01 9.43310440e-01 -1.04745162e+00
-8.72532204e-02 4.91585374e-01 5.24871886e-01 4.52869326e-01
7.46239364e-01 -1.81806400e-01 -7.44186401e-01 4.91522579e-03
5.33205032e-01 -6.47582352e-01 -2.25014295e-02 2.10711941e-01
-1.30216599e+00 7.22752333e-01 -1.95795223e-01 1.02521873e+00
-4.00676340e-01 -1.87889636e-01 7.47835696e-01 3.13094586e-01
5.57686925e-01 1.34448254e+00 -1.07155573e+00 -3.95676851e-01
-5.03105581e-01 5.71381629e-01 9.72003937e-01 8.13204467e-01
6.59852326e-01 -1.53880939e-01 4.47869301e-02 1.26800644e+00
1.99127987e-01 6.98733509e-01 1.26011956e+00 -6.24273300e-01
3.61454636e-01 1.00610876e+00 1.97688997e-01 -9.65709686e-01
-6.78050518e-01 6.01806678e-02 -6.87791482e-02 -6.16107225e-01
2.83165008e-01 -1.53397068e-01 -7.87454128e-01 1.91791093e+00
4.34322506e-01 -7.16880783e-02 6.05165243e-01 3.77865881e-01
1.61191857e+00 3.09943676e-01 6.27780855e-01 -3.81928921e-01
1.82916641e+00 3.01984306e-02 -1.11627281e+00 -3.31455529e-01
9.22073901e-01 -1.13980949e+00 1.27660763e+00 1.70791999e-01
-8.38908970e-01 -5.39724231e-01 -7.37737417e-01 -2.40748510e-01
-1.58029747e+00 9.39566121e-02 6.47603631e-01 6.11811697e-01
-7.49076784e-01 4.24599648e-01 -2.41036132e-01 -7.22012520e-01
-5.17749637e-02 -6.15270734e-02 -5.35306394e-01 -4.11375165e-02
-1.95766544e+00 1.43767154e+00 1.34599197e+00 -6.80182457e-01
-2.12096676e-01 -8.81885707e-01 -1.45992482e+00 -2.60360032e-01
4.53776360e-01 -3.54628295e-01 6.43891633e-01 -4.07419950e-01
-8.98384273e-01 1.54166615e+00 2.69098431e-02 -4.80305701e-01
-5.23305118e-01 -2.04255536e-01 -1.23190629e+00 -3.65854114e-01
5.27464449e-01 6.20492637e-01 2.28148267e-01 -8.02560806e-01
-7.66181469e-01 -4.91282076e-01 5.45208156e-02 3.72829765e-01
-5.58541000e-01 3.54003996e-01 -1.53562739e-01 -1.04018295e+00
-1.22460701e-01 -4.94003624e-01 -1.61841586e-02 -1.07429957e+00
-3.31380069e-01 -7.10458338e-01 4.67110664e-01 -7.92017162e-01
1.64447629e+00 -1.79305851e+00 1.51234806e-01 2.40974143e-01
-6.33382201e-02 7.00459480e-01 -5.33809885e-03 4.59324986e-01
-7.22638518e-02 4.03780848e-01 3.63677412e-01 -2.93481573e-02
2.39317611e-01 5.92534363e-01 -7.86578581e-02 -5.12318552e-01
-5.43769412e-02 8.13821793e-01 -1.40166128e+00 -6.16528571e-01
3.77158731e-01 2.59302050e-01 -2.70464960e-02 -2.78819084e-01
-2.63433963e-01 2.48133969e-02 -4.24599648e-01 5.01673520e-01
-7.02303573e-02 5.36079347e-01 4.47793871e-01 -1.59491569e-01
-1.75865710e-01 1.18881953e+00 -1.38200045e+00 1.64437544e+00
-1.22491086e+00 3.71744305e-01 -5.62672019e-01 -7.05325186e-01
1.02936471e+00 2.13148817e-01 1.57255217e-01 -9.02849853e-01
3.76511782e-01 5.21229565e-01 -2.47289553e-01 -4.99655247e-01
1.08668125e+00 -1.20677441e-01 -6.60255551e-01 5.14586568e-02
6.38729155e-01 -5.23135304e-01 8.86264324e-01 -4.64179218e-02
5.98437309e-01 -1.15450891e-02 8.20992529e-01 -6.82419121e-01
6.39160872e-01 2.42998242e-01 3.29266131e-01 9.39240158e-02
4.02996391e-01 -6.80593923e-02 1.94729403e-01 -3.10779899e-01
-8.68810475e-01 -1.03487885e+00 -6.53579712e-01 1.10688770e+00
3.59356403e-02 -1.03420770e+00 -7.24232256e-01 -4.66872782e-01
-1.44059643e-01 1.12969112e+00 -5.87942123e-01 -1.12646826e-01
-3.30685496e-01 -3.40410322e-01 6.98105037e-01 4.34559315e-01
2.56552368e-01 -1.33790112e+00 -1.80611685e-01 4.39086825e-01
-2.10040852e-01 -1.17331088e+00 -4.66920622e-02 3.01391691e-01
-2.45826304e-01 -1.22441137e+00 7.17809796e-02 -1.08619738e+00
2.90589243e-01 -5.36242723e-01 1.66469669e+00 2.83268411e-02
-4.39913183e-01 2.57580906e-01 -7.50041008e-01 -1.18302774e+00
-4.76826102e-01 4.75016624e-01 3.07154089e-01 -7.36948729e-01
9.43180501e-01 -1.96699098e-01 2.02572301e-01 -9.22299847e-02
-1.03511679e+00 -5.01637936e-01 -5.99281974e-02 6.44377768e-01
6.54308200e-01 -2.54057646e-01 6.63178146e-01 -1.43803489e+00
9.95363772e-01 -5.47681093e-01 -4.28527594e-01 3.36942285e-01
-7.50877380e-01 3.83042097e-01 2.76542723e-01 -3.74299198e-01
-1.16607845e+00 -2.50681639e-01 -7.35564888e-01 5.85424364e-01
-4.04794842e-01 8.04760277e-01 -3.42266083e-01 1.08302854e-01
7.86796570e-01 -6.41967431e-02 -4.33897972e-01 -6.09507799e-01
7.41094410e-01 6.13292336e-01 5.47955751e-01 -8.54062259e-01
2.82680035e-01 -2.15427324e-01 -1.98111415e-01 -1.01912153e+00
-8.69803131e-01 -8.51887524e-01 -7.55217373e-01 1.11120664e-01
8.76328766e-01 -7.11149871e-01 -2.54327148e-01 1.60288513e-01
-1.00784159e+00 8.21447447e-02 -7.48840153e-01 5.07538319e-01
-1.10372983e-01 -9.18349624e-02 -2.19102383e-01 -1.45433590e-01
-2.48995498e-01 -4.82370228e-01 1.04828584e+00 4.22406435e-01
-8.34526360e-01 -1.69182301e+00 4.26267952e-01 1.53456628e-01
3.64262432e-01 1.17443800e-01 1.07650864e+00 -1.22962129e+00
4.10472065e-01 -1.56140640e-01 7.65615180e-02 3.04068416e-01
4.43985052e-02 -9.03514400e-02 -4.92372632e-01 4.55324560e-01
-6.45936012e-01 -4.99006420e-01 5.10360658e-01 4.98276018e-02
8.67498875e-01 -3.84147078e-01 -4.15598124e-01 2.03888282e-01
1.54375041e+00 2.77980119e-02 7.03070462e-01 5.09853065e-01
5.46639025e-01 7.60256946e-01 8.91717494e-01 3.33977461e-01
6.28734589e-01 8.76185000e-01 -1.54365331e-01 3.16965342e-01
-3.26490432e-01 -4.23186988e-01 -3.26991715e-02 7.38335371e-01
1.65392403e-02 -1.28781661e-01 -1.11524069e+00 1.22779524e+00
-1.51896179e+00 -6.50107086e-01 -2.56720871e-01 1.96990216e+00
1.47044086e+00 -5.62679069e-03 -1.76854685e-01 -5.65399714e-02
6.28609776e-01 7.27082863e-02 1.35618135e-01 -8.06898892e-01
-3.50855112e-01 9.27193940e-01 3.76772493e-01 7.68089116e-01
-8.14347088e-01 1.78542519e+00 5.87927771e+00 1.22269976e+00
-7.81636894e-01 2.80717909e-01 -1.29277125e-01 3.27572674e-01
-5.95859468e-01 -6.66044578e-02 -1.33455694e+00 3.71111959e-01
8.12581003e-01 -7.23679066e-01 1.03760958e-01 9.17757332e-01
-3.62648338e-01 -2.22489938e-01 -6.97663367e-01 9.78290737e-01
1.46005228e-01 -1.50225735e+00 3.35265964e-01 -4.71608877e-01
6.97027266e-01 -8.60988870e-02 -3.12454969e-01 4.39445823e-01
7.18614876e-01 -7.82915890e-01 3.11383158e-01 3.55830193e-01
9.54085171e-01 -7.74112999e-01 1.14511073e+00 -1.24312192e-01
-9.93779182e-01 6.31296873e-01 -5.43815494e-01 -1.28427222e-01
1.52869269e-01 6.57829285e-01 -8.58474731e-01 5.54166555e-01
5.98720014e-01 5.42401850e-01 -7.16512680e-01 5.82912028e-01
-7.91937053e-01 3.56660396e-01 -2.94411480e-01 -5.72705090e-01
5.74572384e-02 1.89307973e-01 6.97942257e-01 1.52591419e+00
3.28232437e-01 -2.19135746e-01 1.91654086e-01 3.77177089e-01
-1.98411703e-01 1.10956800e+00 -8.52739275e-01 -3.91189128e-01
9.73301351e-01 1.36397684e+00 -8.86170685e-01 -6.35673940e-01
-6.76565245e-02 6.16896212e-01 2.09008917e-01 -6.72551170e-02
-2.77258724e-01 -8.94567430e-01 1.15332711e+00 1.22496307e-01
9.21893399e-03 -6.52400777e-02 -1.79508880e-01 -1.04998529e+00
-1.78253278e-01 -4.41674322e-01 5.85924089e-01 -6.36766851e-01
-1.36932027e+00 6.84021711e-01 3.01139235e-01 -5.08534014e-01
-6.96553707e-01 -1.08463562e+00 -1.96265742e-01 8.35636795e-01
-1.39902830e+00 -1.16082704e+00 -6.92037307e-03 5.22547007e-01
2.23340333e-01 -4.34528023e-01 1.30661404e+00 3.34356815e-01
-2.87036091e-04 4.17171091e-01 -1.75703108e-01 2.27412567e-01
1.00780046e+00 -1.55738401e+00 2.80940503e-01 6.03512228e-01
6.39726341e-01 7.33724594e-01 8.92374694e-01 -8.81932557e-01
-5.14476717e-01 -1.13334119e+00 1.92536783e+00 -7.93848813e-01
9.92356360e-01 -2.16672227e-01 -8.46604526e-01 4.04312462e-01
1.89604566e-01 -3.98060054e-01 1.09604931e+00 7.50228107e-01
-3.35024744e-01 1.28100589e-01 -1.19938183e+00 5.51543236e-01
1.21678483e+00 -5.73648751e-01 -1.18142378e+00 3.25045198e-01
8.11270297e-01 -6.57869697e-01 -1.07121718e+00 1.34951413e-01
2.01465592e-01 -1.05290204e-01 1.06207323e+00 -1.07194567e+00
1.64437279e-01 -3.24812643e-02 -1.64350823e-01 -2.05545068e+00
2.25651264e-01 -4.26096290e-01 2.07523227e-01 1.68924916e+00
9.00450885e-01 -5.17004907e-01 8.55502561e-02 1.31010592e-01
-3.09650332e-01 -1.40067175e-01 -9.03402805e-01 -8.73078167e-01
1.89265296e-01 -8.01313579e-01 9.61935222e-01 1.30437279e+00
5.73392928e-01 5.98721266e-01 2.78551131e-01 -1.48368999e-01
-2.76539382e-02 -2.37133101e-01 1.88600019e-01 -1.50583661e+00
2.01108813e-01 -5.78982174e-01 -6.57047987e-01 -1.11421019e-01
7.20297515e-01 -1.11109936e+00 -4.14510854e-02 -1.52440178e+00
-2.83978105e-01 -5.77190578e-01 -1.61830127e-01 7.03362763e-01
-4.07384425e-01 3.30052763e-01 -6.73720837e-02 -6.37349248e-01
-6.09042943e-01 1.03228845e-01 8.43371153e-01 2.73516923e-01
-1.07369967e-01 -6.18739069e-01 -7.91606963e-01 8.81097794e-01
7.39030600e-01 -6.74296081e-01 -1.92969516e-01 -3.55035216e-01
9.98778164e-01 -4.74297732e-01 -5.87845035e-02 -7.60525703e-01
-1.29504919e-01 -2.45237097e-01 -6.38970882e-02 1.04556076e-01
-1.50169328e-01 -5.32919943e-01 -2.64929146e-01 1.85184643e-01
-4.46564198e-01 3.46763223e-01 3.62717301e-01 -3.72378714e-02
-6.47390068e-01 -7.92334378e-01 5.08485615e-01 -2.96910733e-01
-1.23855317e+00 -3.97479758e-02 -4.08439904e-01 8.22153628e-01
9.15802419e-01 -5.23093157e-03 -3.71279627e-01 1.15954340e-01
-9.76348341e-01 1.09546006e-01 5.42000353e-01 1.04315603e+00
3.01563472e-01 -1.47391605e+00 -6.32302940e-01 3.51590030e-02
8.51020098e-01 -5.60092866e-01 -7.33800754e-02 3.05313971e-02
-5.39631128e-01 6.13589227e-01 -4.55028772e-01 3.52090836e-01
-1.48590934e+00 5.48532426e-01 -7.80614540e-02 -4.85226721e-01
1.07280230e-02 1.09346044e+00 -4.62497264e-01 -7.39551067e-01
7.16173723e-02 -3.47263247e-01 -9.32223260e-01 5.82382023e-01
7.16866732e-01 1.66903570e-01 3.52090985e-01 -7.13075817e-01
-4.82751876e-01 2.72107631e-01 2.38356397e-01 -1.00940168e-01
1.26140058e+00 -2.57029891e-01 -4.11363631e-01 7.65517056e-01
7.07519770e-01 5.00169814e-01 1.03698231e-01 -2.76876211e-01
5.61113536e-01 -8.75971913e-02 1.16526708e-02 -1.20864904e+00
-5.09206235e-01 6.37831151e-01 2.62383312e-01 2.29698375e-01
7.19266593e-01 4.52856332e-01 7.70472527e-01 5.54801941e-01
7.14627087e-01 -1.69183064e+00 -4.11246032e-01 1.14788997e+00
6.96632624e-01 -1.08865297e+00 -5.44398010e-01 -6.40493453e-01
-6.25107110e-01 1.20984018e+00 5.30011892e-01 1.41468957e-01
8.80033493e-01 3.01913559e-01 2.11968780e-01 -5.76727450e-01
-4.32396144e-01 -9.57891405e-01 5.65205276e-01 1.11435282e+00
1.02492702e+00 3.40320379e-01 -1.26166761e+00 9.48303223e-01
-6.90980613e-01 -3.08080107e-01 3.57879549e-01 4.74229217e-01
-4.66692448e-01 -1.99822223e+00 7.98792467e-02 4.46369767e-01
-6.06108785e-01 -7.59221554e-01 -8.29221904e-01 9.01095867e-01
7.57311642e-01 5.46396077e-01 2.31656563e-02 -1.22605644e-01
3.00208151e-01 4.74305719e-01 6.11904860e-01 -1.42100823e+00
-6.29304588e-01 -6.96275711e-01 9.09955978e-01 -5.68595827e-01
-4.80736762e-01 -7.52500236e-01 -1.25576365e+00 6.17242269e-02
-1.92939639e-01 5.81722081e-01 8.48766625e-01 1.39549029e+00
9.98249277e-03 5.79285681e-01 -1.19018704e-01 -1.71236679e-01
3.44001383e-01 -1.15944147e+00 -6.68288708e-01 6.06991351e-01
-5.44463754e-01 -8.67200792e-01 2.30323032e-01 4.67410646e-02] | [10.265355110168457, 9.318961143493652] |
9b1c40e1-dbeb-4545-945f-00bf33521040 | a-machine-learning-framework-for-neighbor | 2212.11451 | null | https://arxiv.org/abs/2212.11451v1 | https://arxiv.org/pdf/2212.11451v1.pdf | A machine learning framework for neighbor generation in metaheuristic search | This paper presents a methodology for integrating machine learning techniques into metaheuristics for solving combinatorial optimization problems. Namely, we propose a general machine learning framework for neighbor generation in metaheuristic search. We first define an efficient neighborhood structure constructed by applying a transformation to a selected subset of variables from the current solution. Then, the key of the proposed methodology is to generate promising neighbors by selecting a proper subset of variables that contains a descent of the objective in the solution space. To learn a good variable selection strategy, we formulate the problem as a classification task that exploits structural information from the characteristics of the problem and from high-quality solutions. We validate our methodology on two metaheuristic applications: a Tabu Search scheme for solving a Wireless Network Optimization problem and a Large Neighborhood Search heuristic for solving Mixed-Integer Programs. The experimental results show that our approach is able to achieve a satisfactory trade-off between the exploration of a larger solution space and the exploitation of high-quality solution regions on both applications. | ['Andrea Lodi', 'Alain Hertz', 'Vincent Perreault', 'Defeng Liu'] | 2022-12-22 | null | null | null | null | ['variable-selection'] | ['methodology'] | [ 5.00334144e-01 7.24908337e-02 -8.48104656e-01 -1.61592096e-01
-1.05834472e+00 -4.33593452e-01 1.87795281e-01 4.19891000e-01
-3.29408050e-01 1.15656018e+00 -3.35647553e-01 -3.60895693e-01
-7.22381592e-01 -1.19302702e+00 -5.16876459e-01 -1.15708172e+00
-1.90617949e-01 6.93924665e-01 -2.63278961e-01 -3.45432788e-01
7.06178427e-01 7.19745100e-01 -1.82625377e+00 4.31541473e-01
1.30538774e+00 1.16316545e+00 3.05242419e-01 2.79346526e-01
-4.83132511e-01 1.43169135e-01 -6.97621405e-01 1.11959144e-01
3.64237785e-01 -6.19640231e-01 -1.15212572e+00 7.41927400e-02
-1.32725805e-01 9.03479934e-01 5.65211058e-01 1.11110473e+00
3.00537258e-01 3.54587793e-01 6.31925285e-01 -1.44370568e+00
4.87478487e-02 7.15885699e-01 -5.25002241e-01 3.10623527e-01
3.93653899e-01 -2.00643986e-01 1.26825261e+00 -1.13625817e-01
8.15684080e-01 9.96692538e-01 3.08930725e-01 3.26276779e-01
-1.36022747e+00 -4.17730600e-01 9.01384577e-02 6.14473403e-01
-1.32622492e+00 -1.62837997e-01 1.08739626e+00 1.27490744e-01
7.52482057e-01 9.62733746e-01 8.64665151e-01 5.27522922e-01
2.94862747e-01 7.47113585e-01 1.00590980e+00 -8.68393004e-01
1.04875004e+00 2.28546068e-01 -2.03332398e-02 5.67028999e-01
2.26556003e-01 3.08439940e-01 -3.34485143e-01 -6.51886821e-01
1.04317985e-01 -4.53208894e-01 -3.84960860e-01 -9.50926304e-01
-7.86580324e-01 1.26009488e+00 5.53606331e-01 6.66586995e-01
-7.09904253e-01 -2.43368909e-01 9.53702182e-02 4.42422837e-01
2.43718579e-01 1.09897435e+00 -6.58135355e-01 -5.51826395e-02
-1.15027726e+00 2.72144288e-01 9.57377851e-01 4.68459278e-01
1.07506192e+00 -1.16721034e-01 4.39082971e-03 6.24119043e-01
-1.26293421e-01 2.37000138e-01 4.04611379e-01 -8.20092976e-01
5.18125057e-01 6.71440423e-01 1.60293236e-01 -1.28850639e+00
-4.34349984e-01 -7.21190393e-01 -2.81669170e-01 4.61613327e-01
4.65278812e-02 -2.52051711e-01 -5.94174743e-01 1.46336854e+00
5.90852380e-01 6.50534704e-02 7.26878941e-02 5.89480162e-01
3.64509732e-01 8.93648624e-01 -2.97332138e-01 -1.04863691e+00
6.71651185e-01 -1.31442857e+00 -6.21074378e-01 9.24826488e-02
8.10904205e-01 -2.99747199e-01 4.46470857e-01 5.48462927e-01
-1.26750791e+00 -3.48964542e-01 -1.04606867e+00 8.24596822e-01
-4.30057347e-01 -6.21864609e-02 7.66999662e-01 8.93540740e-01
-1.05527067e+00 6.53798878e-01 -5.04967988e-01 -2.71574974e-01
1.63489372e-01 7.65579462e-01 -9.82948393e-02 2.42276639e-02
-7.30706930e-01 7.14796603e-01 7.13341653e-01 1.46406725e-01
-2.30295137e-01 -3.11156988e-01 -7.41664231e-01 3.06181133e-01
7.32945561e-01 -6.32256806e-01 6.58339322e-01 -1.59468997e+00
-1.30539227e+00 5.87170839e-01 -4.12882179e-01 -3.03424418e-01
2.84227818e-01 4.40083116e-01 -4.86668408e-01 -5.69543168e-02
2.46348288e-02 1.93187386e-01 1.00746620e+00 -1.63087523e+00
-1.19815028e+00 -3.87526095e-01 -8.73485394e-03 2.40127176e-01
-5.30524373e-01 -3.94334942e-01 -3.28691840e-01 -4.12188649e-01
1.93450034e-01 -7.27479875e-01 -8.81944180e-01 -5.63031971e-01
-2.99201459e-01 -3.94052118e-02 5.46946526e-01 -1.78735778e-01
1.59791052e+00 -1.46141088e+00 8.74734163e-01 1.30685163e+00
-9.88870934e-02 2.12311864e-01 -5.29939175e-01 3.49834025e-01
-2.39319652e-01 2.75819153e-01 -1.66909695e-01 1.71576589e-01
-3.19934309e-01 3.70909810e-01 1.26049325e-01 3.46226096e-01
-2.23410800e-01 6.89005554e-01 -1.01609600e+00 -4.18610126e-01
5.55983558e-02 -4.97019440e-02 -7.56405771e-01 5.18022627e-02
-4.82764304e-01 2.01109186e-01 -1.05368435e+00 9.30935800e-01
6.36112094e-01 7.71321729e-02 5.62592626e-01 -1.63708344e-01
-3.33630741e-01 -8.72070938e-02 -1.59812260e+00 1.39052272e+00
-5.89315295e-01 1.14124119e-01 3.66029143e-01 -1.61800086e+00
1.09207106e+00 -1.98358163e-01 1.01828408e+00 -8.38923693e-01
4.33099329e-01 3.84650081e-01 -3.29161376e-01 -5.26923954e-01
3.72317046e-01 1.06003985e-01 1.18485689e-01 4.34785515e-01
-2.16080397e-01 -6.21061772e-02 6.21000528e-01 -6.31319225e-01
1.02470410e+00 -2.83796787e-01 4.95069414e-01 -4.55013454e-01
1.11402619e+00 3.62708509e-01 7.63730586e-01 9.00422215e-01
2.47308880e-01 3.46367545e-02 4.65262711e-01 -8.16354930e-01
-6.04238927e-01 -4.85909134e-01 8.34756941e-02 8.62532318e-01
1.52592838e-01 -2.33324036e-01 -6.98474050e-01 -8.14679623e-01
-1.94072202e-01 9.10289288e-01 -7.26376653e-01 -2.02283964e-01
-7.96120942e-01 -1.06797206e+00 -1.64250687e-01 -1.52806759e-01
1.20605759e-01 -1.15169930e+00 -7.67753541e-01 2.52389342e-01
-2.37066686e-01 -4.39605266e-01 1.28878783e-02 6.84268594e-01
-1.16301417e+00 -1.17428839e+00 -3.41584176e-01 -1.15971661e+00
6.75773025e-01 5.99791147e-02 1.22994936e+00 4.28548276e-01
-4.03126478e-01 1.67646930e-02 -5.44402897e-01 6.66660443e-02
-5.02634764e-01 5.98581433e-01 -2.45691746e-01 -6.30802242e-03
1.33359537e-01 -5.52354336e-01 1.30107969e-01 2.75379092e-01
-7.07074523e-01 -4.48650718e-01 7.23855793e-01 1.01191092e+00
9.84005868e-01 9.54431117e-01 3.73442978e-01 -6.74677730e-01
7.64045656e-01 -6.44164264e-01 -9.94582295e-01 6.34871364e-01
-7.85034418e-01 6.18101478e-01 6.74219549e-01 -3.61263722e-01
-5.57654083e-01 2.24669009e-01 2.66670231e-02 -1.00348800e-01
-1.51525540e-02 7.21614659e-01 -4.24626499e-01 -5.73982716e-01
4.90291685e-01 2.24649549e-01 -9.15511847e-02 -2.82464147e-01
2.03608945e-01 2.48491257e-01 1.76942259e-01 -9.11796510e-01
9.51176643e-01 2.02250436e-01 4.13138986e-01 -6.72207475e-01
-3.54166925e-01 -5.32246649e-01 -4.75734860e-01 -1.44988438e-02
3.89698803e-01 1.19322734e-02 -7.95638025e-01 -1.55195281e-01
-7.55912066e-01 7.36697316e-02 -4.44210649e-01 1.02293245e-01
-8.38972330e-01 9.59568992e-02 3.97624344e-01 -8.45692456e-01
-3.05997372e-01 -1.35019124e+00 6.41239703e-01 4.72558558e-01
-8.95154402e-02 -1.17325246e+00 4.19343382e-01 3.33553761e-01
2.60357887e-01 5.92681170e-01 1.31614435e+00 -6.98527455e-01
-5.16768932e-01 5.84281571e-02 3.92393172e-01 -2.40623429e-01
4.07544076e-01 -2.10975856e-02 -2.98993468e-01 -4.84397888e-01
1.54224545e-01 -6.39213622e-02 6.33474350e-01 6.53122723e-01
1.13154900e+00 -5.13973176e-01 -9.13244486e-01 9.64088023e-01
1.74760818e+00 7.31927693e-01 4.88359392e-01 1.05287886e+00
3.56430560e-02 5.01333654e-01 1.11375916e+00 5.50765693e-01
-1.38930753e-01 9.67562735e-01 6.58833921e-01 -2.16598198e-01
5.36818564e-01 2.25309536e-01 -9.58887413e-02 -9.03294981e-02
-7.63667971e-02 -5.41036785e-01 -9.41158354e-01 5.13089955e-01
-1.92003226e+00 -1.04166090e+00 2.21531659e-01 2.12240434e+00
5.98541439e-01 6.49514422e-02 2.95052648e-01 5.87223053e-01
9.02463496e-01 1.03176557e-01 -3.64134431e-01 -9.90854979e-01
-1.90818515e-02 5.56443393e-01 3.05704057e-01 8.23954999e-01
-1.21792746e+00 8.23438942e-01 5.80863619e+00 1.01347852e+00
-1.32943618e+00 -5.71910739e-01 6.50710046e-01 -2.10118398e-01
-5.83354533e-01 -3.10361743e-01 -5.84130347e-01 2.10281640e-01
6.97376788e-01 -1.29497230e-01 8.05601776e-01 7.84860492e-01
3.63851160e-01 -6.11097962e-02 -8.49918664e-01 8.43060195e-01
1.53147027e-01 -1.76387155e+00 1.89116940e-01 1.07930303e-01
1.22546232e+00 -5.87752759e-01 2.62150973e-01 7.73129985e-03
-1.68048546e-01 -1.23455417e+00 3.55763048e-01 2.33108297e-01
1.21330738e-01 -1.53542495e+00 8.20781291e-01 3.02977085e-01
-1.10410964e+00 -6.95521355e-01 -9.13520530e-02 2.36170739e-01
-8.12751651e-02 4.42765981e-01 -6.21843517e-01 8.79364014e-01
6.09905064e-01 3.23548704e-01 -3.76455188e-01 1.54489470e+00
1.96591809e-01 -1.99210102e-04 -2.53244191e-01 -4.09123421e-01
6.91122830e-01 -5.81405282e-01 8.51546466e-01 8.54838669e-01
3.54005426e-01 -1.22455182e-02 2.88256526e-01 7.86042690e-01
3.04818422e-01 5.40470362e-01 -3.36056650e-01 6.75107390e-02
4.56665665e-01 1.27918541e+00 -1.01157987e+00 3.86702657e-01
1.09073274e-01 6.28251016e-01 3.36064100e-01 4.27198559e-01
-7.57147729e-01 -4.02627617e-01 4.77655560e-01 -3.13694626e-01
6.79533362e-01 3.11368018e-01 -6.52323961e-01 -6.96076453e-01
2.30835434e-02 -1.14868367e+00 6.17051542e-01 -1.34835720e-01
-6.93377018e-01 1.02895975e+00 -1.09095700e-01 -9.56434727e-01
-5.84346056e-01 -4.90866512e-01 -7.84582019e-01 7.16677248e-01
-1.75575233e+00 -8.57838988e-01 -1.11270189e-01 4.86648411e-01
4.82599199e-01 -5.62309861e-01 7.65300512e-01 4.24144641e-02
-4.22858298e-01 4.99837190e-01 4.93547916e-01 -6.35099232e-01
-5.99945672e-02 -1.03712094e+00 -3.01873952e-01 5.82122266e-01
2.24356800e-01 3.43746513e-01 8.83486211e-01 -3.23396653e-01
-1.47514224e+00 -6.69656813e-01 9.41324353e-01 1.23203792e-01
4.04875785e-01 2.60262549e-01 -3.25773627e-01 6.67678192e-02
-7.74624124e-02 -3.46297443e-01 6.24401867e-01 3.69459271e-01
3.54613572e-01 -3.01567405e-01 -1.38160050e+00 4.38205242e-01
6.46608293e-01 3.58922839e-01 -3.65458488e-01 2.19018757e-01
7.65195861e-02 -1.88203931e-01 -5.26674449e-01 5.32934427e-01
2.68712252e-01 -7.74794221e-01 1.06472492e+00 -8.19242060e-01
1.58709735e-01 -1.91119060e-01 -1.93195343e-01 -1.47524810e+00
-4.23040807e-01 -8.16295385e-01 -1.73524544e-01 1.04219007e+00
6.52239442e-01 -3.48066479e-01 1.39395857e+00 2.90819108e-01
2.27706850e-01 -1.25514102e+00 -1.22093081e+00 -6.75451994e-01
9.57114846e-02 1.39953882e-01 7.37327456e-01 8.96910191e-01
-8.06779340e-02 -3.40554900e-02 -8.27603880e-03 2.26395652e-01
6.58987045e-01 8.55809212e-01 3.91912878e-01 -1.34397554e+00
-3.49698156e-01 -7.49866962e-01 -2.41731733e-01 -4.21265274e-01
5.58902323e-01 -6.54502749e-01 -8.07914659e-02 -1.37515724e+00
-3.41759585e-02 -9.64006841e-01 -4.75100577e-01 2.88515985e-01
-1.54646914e-02 -1.88117400e-01 1.02895964e-02 -2.78275460e-01
-5.11673272e-01 5.16130209e-01 8.72364461e-01 -4.56506670e-01
-8.38125050e-01 6.06935143e-01 -7.26780415e-01 5.39818108e-01
8.42293501e-01 -5.85914850e-01 -3.85509133e-01 1.13384938e-02
2.09351003e-01 4.05730218e-01 -3.16377759e-01 -9.89966393e-01
3.11506595e-02 -8.77002597e-01 6.51063174e-02 -4.69225705e-01
6.22564964e-02 -1.31528282e+00 2.08310828e-01 7.50401914e-01
-3.27338606e-01 1.48697823e-01 1.28503218e-01 4.11405474e-01
-1.95892826e-01 -6.55611098e-01 7.98961878e-01 4.28040661e-02
-8.11535835e-01 -1.15488984e-01 -3.57711107e-01 -4.83015664e-02
1.33086336e+00 -4.94958490e-01 9.83927324e-02 -2.33981729e-01
-7.94883847e-01 4.47086185e-01 4.91910696e-01 3.94316405e-01
4.61784750e-01 -1.06996286e+00 -3.40359092e-01 2.82652885e-01
-1.41128406e-01 -5.46434581e-01 -3.01180393e-01 3.12111050e-01
-5.22223413e-01 4.41674531e-01 -4.11021024e-01 -3.39716643e-01
-1.50611651e+00 7.81338036e-01 6.06644869e-01 -6.01930082e-01
-9.71468613e-02 7.62556553e-01 -8.06005657e-01 -4.60177481e-01
4.48903441e-01 -2.84299165e-01 -5.56438565e-01 3.01654100e-01
3.62775028e-01 7.19204545e-01 2.27950528e-01 -5.51604092e-01
-5.58354974e-01 7.17326105e-01 2.92168230e-01 1.18896224e-01
1.65800619e+00 -2.89876889e-02 -4.78383511e-01 -1.40364915e-01
1.39763618e+00 2.19376966e-01 -4.51267183e-01 2.27451995e-02
2.99896210e-01 -4.44789022e-01 2.91462451e-01 -8.81197393e-01
-1.52670670e+00 -5.11432141e-02 6.41784728e-01 3.44211668e-01
1.67164278e+00 -2.11054742e-01 5.26661038e-01 7.57300735e-01
6.46533430e-01 -1.26669323e+00 -1.22426994e-01 2.48972833e-01
5.68063915e-01 -9.99169588e-01 9.68199968e-02 -5.21726131e-01
-2.28384450e-01 1.47646606e+00 4.64394271e-01 1.92898929e-01
4.99929845e-01 2.25351959e-01 -1.59433093e-02 -5.81674278e-02
-6.78648829e-01 -2.75881082e-01 3.69546026e-01 6.78510785e-01
-7.61660859e-02 -9.27248448e-02 -9.30290878e-01 3.50089461e-01
-2.09259182e-01 -2.87330747e-01 1.82614699e-01 1.03221309e+00
-7.36939847e-01 -1.69985759e+00 -6.82623625e-01 2.08072528e-01
-1.10210627e-01 7.98023567e-02 -3.98507476e-01 6.08332217e-01
4.26008910e-01 1.29920983e+00 -1.58551291e-01 -2.09621653e-01
1.46174625e-01 -2.38001317e-01 4.07361209e-01 -3.54221314e-01
-8.24536502e-01 -1.54727167e-02 2.12800235e-01 -9.04514313e-01
-5.24403811e-01 -6.44772351e-01 -1.16721511e+00 -2.68486571e-02
-4.10299480e-01 1.13261592e+00 5.33899724e-01 1.06835043e+00
8.99767950e-02 5.30879676e-01 1.06445670e+00 -7.59915471e-01
-2.58126467e-01 7.94781372e-02 -4.75795805e-01 -1.71007544e-01
3.76684368e-01 -7.56341934e-01 -3.15935403e-01 -4.14800972e-01] | [5.767823696136475, 3.619922399520874] |
def39620-422f-446a-b6d1-a47a6bb4a434 | 191013166 | 1910.13166 | null | https://arxiv.org/abs/1910.13166v2 | https://arxiv.org/pdf/1910.13166v2.pdf | Towards a Model for Spoken Conversational Search | Conversation is the natural mode for information exchange in daily life, a spoken conversational interaction for search input and output is a logical format for information seeking. However, the conceptualisation of user-system interactions or information exchange in spoken conversational search (SCS) has not been explored. The first step in conceptualising SCS is to understand the conversational moves used in an audio-only communication channel for search. This paper explores conversational actions for the task of search. We define a qualitative methodology for creating conversational datasets, propose analysis protocols, and develop the SCSdata. Furthermore, we use the SCSdata to create the first annotation schema for SCS: the SCoSAS, enabling us to investigate interactivity in SCS. We further establish that SCS needs to incorporate interactivity and pro-activity to overcome the complexity that the information seeking process in an audio-only channel poses. In summary, this exploratory study unpacks the breadth of SCS. Our results highlight the need for integrating discourse in future SCS models and contributes the advancement in the formalisation of SCS models and the design of SCS systems. | ['Damiano Spina', 'Hideo Joho', 'Paul Thomas', 'Lawrence Cavedon', 'Johanne R. Trippas', 'Mark Sanderson'] | 2019-10-29 | null | null | null | null | ['conversational-search'] | ['natural-language-processing'] | [ 3.65632862e-01 7.27111399e-01 -2.76401788e-01 -2.96497732e-01
-6.85020804e-01 -7.87589729e-01 1.15333903e+00 1.96128368e-01
-3.07718486e-01 3.33303034e-01 1.19677901e+00 -6.27790451e-01
-6.51049793e-01 -2.66133696e-01 7.71316215e-02 -5.26413880e-02
1.41833559e-01 3.32418531e-01 8.95887837e-02 -4.09229130e-01
4.97770876e-01 2.27292284e-01 -1.87671483e+00 8.34492803e-01
5.62048495e-01 7.80579090e-01 6.42291248e-01 9.85906899e-01
-7.25365520e-01 1.11890709e+00 -8.64944637e-01 1.35439053e-01
-4.47957367e-01 -7.85682499e-01 -1.84579539e+00 2.78580990e-02
-2.36386076e-01 -2.40210354e-01 -2.66275462e-02 2.49392688e-01
6.37713015e-01 1.08610302e-01 3.36254269e-01 -1.40718842e+00
-2.28552252e-01 9.67539251e-01 6.29283786e-01 2.52293795e-01
1.46826339e+00 -1.29373567e-02 9.63960826e-01 -4.30601358e-01
1.05222392e+00 1.48602664e+00 3.00866038e-01 5.24765551e-01
-1.05317688e+00 -3.81518126e-01 -8.62302855e-02 3.06295514e-01
-1.09189785e+00 -8.03717017e-01 4.73183185e-01 -5.56141973e-01
1.14731324e+00 1.08169627e+00 9.93168056e-01 1.17439294e+00
-4.60440874e-01 1.12253618e+00 9.29830313e-01 -9.76527035e-01
4.75286394e-02 7.12593257e-01 3.02538633e-01 -2.82543693e-02
-7.84448683e-01 -2.40242019e-01 -1.18072963e+00 -4.71967459e-01
4.24154937e-01 -5.67477942e-01 -4.14151162e-01 2.38412783e-01
-1.08797455e+00 5.36815643e-01 5.02844155e-02 8.50945234e-01
-2.59239346e-01 -1.82701260e-01 5.13787091e-01 6.75009429e-01
2.04912871e-01 7.27474868e-01 -1.45253628e-01 -1.34979916e+00
-5.63348532e-01 2.83435732e-01 1.65343344e+00 8.57084751e-01
1.84268698e-01 -6.90494299e-01 -2.70728886e-01 9.61583078e-01
5.44593036e-01 6.50091320e-02 2.36093655e-01 -1.25777972e+00
3.28987278e-02 8.21466923e-01 3.41977775e-01 -8.77135277e-01
-3.68543059e-01 2.36683171e-02 5.25741242e-02 -5.21148860e-01
2.37629637e-01 5.77825233e-02 -1.37097299e-01 1.26076758e+00
9.76859406e-02 -6.19198203e-01 2.14177176e-01 7.90155530e-01
1.15388882e+00 4.10986364e-01 6.89043999e-02 -3.00711572e-01
1.41014791e+00 -5.81655800e-01 -1.13933384e+00 1.35618746e-01
9.43533778e-01 -9.41634238e-01 1.35231209e+00 2.81129181e-01
-1.27078664e+00 -1.03209533e-01 -5.20358145e-01 -3.33772540e-01
-2.55820543e-01 -2.24299967e-01 6.33906186e-01 7.79235065e-01
-1.10408497e+00 2.43465193e-02 -6.56233966e-01 -8.73990595e-01
-2.68827379e-01 1.29616726e-02 2.98895016e-02 6.64834738e-01
-1.38091612e+00 9.72685635e-01 1.20889246e-02 6.30701706e-02
-1.40529513e-01 -4.56665456e-01 -5.63632965e-01 1.76689140e-02
3.15958560e-01 -3.46663475e-01 1.91196978e+00 -9.60142136e-01
-1.52677882e+00 6.39863789e-01 -3.32113892e-01 -7.52977505e-02
1.46343455e-01 -1.29685238e-01 -5.32140195e-01 3.39150071e-01
1.40453547e-01 4.18750167e-01 -1.72175705e-01 -1.52389538e+00
-7.65516341e-01 -3.66619378e-02 1.38424784e-01 6.38602257e-01
-2.37004831e-01 6.98498487e-01 -5.37394762e-01 1.37379784e-02
-9.81713235e-02 -6.37551963e-01 4.32829082e-01 -1.89560488e-01
-2.00848907e-01 -5.07078230e-01 8.71384025e-01 -5.89399755e-01
2.13603735e+00 -2.34320712e+00 7.39522874e-02 3.54029030e-01
7.60797188e-02 4.92449142e-02 2.74289399e-01 1.66125000e+00
4.32256609e-01 5.32700002e-01 3.38141263e-01 -3.98240238e-01
4.79958713e-01 1.53559729e-01 -3.09659719e-01 -5.58314659e-02
-4.85763282e-01 6.39375150e-01 -9.60649014e-01 -5.94277084e-01
7.05154315e-02 5.19619405e-01 -2.95037240e-01 2.23090366e-01
-1.48581475e-01 3.31688166e-01 -4.46468174e-01 4.99740958e-01
-2.27754731e-02 -1.38529554e-01 3.23270500e-01 3.98842454e-01
-7.88413882e-01 1.03147686e+00 -1.00244033e+00 1.60869515e+00
-5.73645353e-01 1.01822543e+00 7.56067216e-01 -2.60426074e-01
5.72698712e-01 7.00674176e-01 3.18423927e-01 -7.75599182e-01
-1.98595580e-02 3.45719635e-01 1.70610975e-02 -1.01311707e+00
5.63962340e-01 4.51025844e-01 -2.17377156e-01 1.07694530e+00
-2.58401185e-01 -4.28722113e-01 -9.71176550e-02 4.18584853e-01
9.66284692e-01 -1.65672347e-01 1.72700420e-01 -2.54980952e-01
4.91647810e-01 1.82919815e-01 -3.55370082e-02 8.28144193e-01
-3.53879146e-02 2.56427258e-01 6.63641214e-01 2.73950561e-03
-3.48476499e-01 -3.76358598e-01 -1.38327584e-01 1.18712580e+00
1.28942400e-01 -8.61141860e-01 -9.48889971e-01 -1.65432617e-01
-2.50268608e-01 9.23706353e-01 -2.47971103e-01 2.03041211e-01
-8.11862424e-02 1.36119574e-01 8.77618015e-01 -1.38066813e-01
4.21021253e-01 -1.46032250e+00 -1.11177003e+00 2.75416195e-01
-6.89967632e-01 -7.63258994e-01 -4.42136586e-01 -1.17912307e-01
-1.39882863e-01 -1.12585402e+00 -2.00829834e-01 -8.56778324e-01
-2.55919211e-02 2.53764868e-01 8.47897291e-01 5.25727034e-01
-1.24865137e-01 1.08013022e+00 -8.58829558e-01 -4.47826266e-01
-6.36626899e-01 1.78737015e-01 -3.66204381e-01 -3.47611308e-01
5.90750575e-01 -3.88490349e-01 -6.46911263e-01 5.38304925e-01
-9.85502064e-01 1.86642304e-01 -5.72893471e-02 4.76234347e-01
-4.07988638e-01 -1.46301687e-01 4.73183244e-01 -4.61874187e-01
1.75674593e+00 -5.66101193e-01 2.83055216e-01 2.10332423e-01
-5.62690258e-01 -3.35118800e-01 -2.39608288e-01 -2.37155274e-01
-1.27468455e+00 -4.05496508e-01 -1.14706658e-01 4.70342070e-01
-3.31665695e-01 9.16724384e-01 1.07120834e-01 -2.33815163e-02
7.56492078e-01 2.27426052e-01 5.48397899e-01 -4.23852801e-01
8.91414806e-02 1.46736801e+00 1.75254390e-01 -4.97122765e-01
-1.90028176e-01 2.01958045e-01 -7.52905667e-01 -1.40479648e+00
9.96370390e-02 -9.91186380e-01 -2.76909351e-01 -8.19290638e-01
6.42207146e-01 -5.59127033e-01 -1.44173706e+00 1.27605155e-01
-1.07996666e+00 -6.93223655e-01 -1.52513653e-01 3.98955047e-01
-6.29997075e-01 4.02641714e-01 -3.99654269e-01 -1.74946046e+00
-1.76101938e-01 -1.13540363e+00 1.02748787e+00 7.07482621e-02
-1.48024869e+00 -9.55463409e-01 -6.91346303e-02 6.90006196e-01
5.26762724e-01 -8.78772065e-02 7.66227245e-01 -8.38060141e-01
-2.82056838e-01 1.67406976e-01 1.81551259e-02 -4.18682933e-01
1.18578717e-01 -1.98270269e-02 -8.75648618e-01 2.51818925e-01
9.22416076e-02 -5.46704352e-01 -1.02051370e-01 -1.28576443e-01
7.59543836e-01 -5.91031730e-01 -3.87883782e-01 -2.58268923e-01
7.99623251e-01 8.03992033e-01 5.24886429e-01 5.08413970e-01
-1.79213300e-01 1.36554313e+00 5.56962729e-01 4.97868270e-01
7.41255403e-01 9.07832086e-01 -5.77234216e-02 2.16516539e-01
8.77134055e-02 -2.99291193e-01 1.64934099e-01 7.53930807e-01
1.97514683e-01 -3.79877120e-01 -1.22254384e+00 7.06688106e-01
-1.91056001e+00 -1.04401183e+00 -2.28262648e-01 1.67700362e+00
1.00773060e+00 -2.23827899e-01 2.57701784e-01 3.62398386e-01
4.26722392e-02 -5.25560491e-02 1.73772663e-01 -8.01009715e-01
3.14295739e-01 -4.77409624e-02 -2.63624460e-01 1.14677870e+00
-2.34356552e-01 6.33592606e-01 7.14142942e+00 2.49963909e-01
-7.45907605e-01 -1.01941742e-03 1.22881927e-01 -1.10108741e-01
-8.15352619e-01 9.90888551e-02 -3.35475683e-01 4.84836161e-01
8.47163439e-01 -1.14560053e-01 7.40832448e-01 4.12125796e-01
7.60848105e-01 -6.81292832e-01 -1.19749570e+00 8.04031014e-01
-8.46551917e-03 -1.30122924e+00 -2.91553587e-01 6.52471036e-02
-1.30312055e-01 -4.99322951e-01 -3.42502445e-01 1.88789561e-01
2.62369141e-02 -9.29290473e-01 6.92287266e-01 6.11935556e-01
5.03290057e-01 -1.54966146e-01 5.14027119e-01 7.20373631e-01
-9.50509787e-01 -1.32064566e-01 7.81555295e-01 -3.73148710e-01
7.16815591e-01 -3.00704062e-01 -1.14997137e+00 1.45317972e-01
8.54961336e-01 6.44471273e-02 -1.53979525e-01 7.68179893e-01
3.76072884e-01 3.93295348e-01 -4.90116179e-01 -5.52355468e-01
1.10449485e-01 -8.18043649e-02 7.67441809e-01 1.51153421e+00
3.74216810e-02 3.07308704e-01 4.73491587e-02 7.49365807e-01
5.38400412e-01 -4.04796526e-02 -5.06760776e-01 -4.96822178e-01
1.13217950e+00 6.78049326e-01 -5.50266445e-01 8.97345766e-02
-2.38467619e-01 9.18384552e-01 -4.03649300e-01 4.48547572e-01
1.13427021e-01 -2.91804790e-01 5.48828661e-01 3.89861882e-01
-4.28837597e-01 -9.94810238e-02 -3.04168135e-01 -4.89135593e-01
5.70986718e-02 -1.36336327e+00 1.05490513e-01 -9.35027540e-01
-5.80496728e-01 4.29097712e-01 5.08715570e-01 -6.51199222e-01
-9.03969049e-01 2.95469314e-01 -5.82115471e-01 1.08730042e+00
-6.53692484e-01 -8.66832495e-01 -5.84552169e-01 3.29884589e-01
7.02964008e-01 1.80792376e-01 1.23020256e+00 1.53475657e-01
6.03475347e-02 1.39189944e-01 -3.83828968e-01 -3.82142603e-01
4.76792276e-01 -1.06065130e+00 1.35143951e-01 8.08442757e-02
-2.49306187e-01 1.04072618e+00 8.63730431e-01 -6.58006132e-01
-1.61802876e+00 3.57039005e-01 1.53912079e+00 -5.14737487e-01
7.31469393e-01 -6.09704077e-01 -7.85657346e-01 3.79201412e-01
7.67888784e-01 -1.13024104e+00 1.10566866e+00 2.85678118e-01
3.14153224e-01 4.33266073e-01 -1.04986906e+00 6.64839387e-01
1.17937732e+00 -1.32537210e+00 -8.53118956e-01 9.31565836e-02
6.60294354e-01 -3.69512022e-01 -9.64956164e-01 -1.31944999e-01
9.54944611e-01 -1.04544699e+00 7.27811515e-01 -4.67246398e-02
-5.23180217e-02 1.07053280e-01 4.85607646e-02 -1.04487419e+00
3.59514862e-01 -1.30588984e+00 2.09954172e-01 1.43200696e+00
5.36237359e-01 -6.56519115e-01 5.59217989e-01 1.33244121e+00
-2.49165311e-01 -4.41930771e-01 -1.05945516e+00 -2.39385162e-02
-4.73103404e-01 -7.79207885e-01 5.88777304e-01 9.03645456e-01
1.28424954e+00 6.46658018e-02 1.25074551e-01 -3.86126071e-01
-1.21908024e-01 -5.34946859e-01 8.00497055e-01 -1.41723466e+00
-1.05310276e-01 -7.04545498e-01 -3.25108171e-02 -1.28488660e+00
-1.86412722e-01 -4.66727614e-01 6.39347956e-02 -1.87510049e+00
-1.79308996e-01 -4.32762802e-01 7.31639028e-01 1.10186070e-01
5.91824651e-01 -6.64961994e-01 3.19402844e-01 5.04343629e-01
-3.75306964e-01 3.74102145e-01 1.02997327e+00 3.54936630e-01
-9.24596608e-01 3.28151323e-02 -9.84639049e-01 2.28128850e-01
5.88114619e-01 -5.08934036e-02 -8.67917776e-01 -7.72944316e-02
5.65315783e-01 4.83633250e-01 2.12959528e-01 -2.77091920e-01
8.58368635e-01 -1.75203487e-01 -4.45546806e-01 -6.19470239e-01
5.03327131e-01 -1.00579822e+00 5.57313144e-01 1.32060781e-01
-9.26828623e-01 -1.25175789e-01 1.64157882e-01 1.51502371e-01
-5.78342259e-01 -4.02419180e-01 -2.76562154e-01 -7.39928186e-02
-3.42279851e-01 -7.62489676e-01 -1.33172953e+00 -1.69438049e-01
6.46989644e-01 -7.80431271e-01 -1.00126393e-01 -1.10553563e+00
-8.90091300e-01 6.67881072e-01 2.92873621e-01 5.87115943e-01
3.37489188e-01 -9.48317528e-01 6.14061058e-02 1.90805778e-01
6.84163580e-03 -1.33165196e-02 -2.03264683e-01 6.93021178e-01
-4.08730209e-01 8.22318792e-01 1.07621640e-01 -4.26172525e-01
-1.53272367e+00 -3.15354228e-01 2.58828431e-01 4.33727801e-01
-3.53178114e-01 7.03317523e-01 -4.22108233e-01 -3.93830299e-01
8.97063911e-01 -3.02061766e-01 -5.64739347e-01 5.83613813e-01
6.22934580e-01 6.24831378e-01 -1.65463343e-01 -4.74724412e-01
-2.16951355e-01 -1.61408365e-01 2.49327719e-01 -9.97216880e-01
7.96476185e-01 -8.54485929e-01 -3.61964792e-01 9.06419158e-01
1.08534038e+00 -7.79949501e-02 -4.06837910e-01 1.54605564e-02
5.14252007e-01 -3.26659322e-01 -2.35722568e-02 -1.02900124e+00
7.08899722e-02 2.83969790e-01 1.99153021e-01 1.16302872e+00
8.20049524e-01 2.90801823e-01 3.37624311e-01 3.07646364e-01
1.00449659e-01 -1.54348171e+00 -1.71292111e-01 4.69418406e-01
1.39410901e+00 -7.91050851e-01 -3.14009368e-01 -6.39137506e-01
-8.86767209e-01 1.15678477e+00 3.48336726e-01 9.91733313e-01
8.41400445e-01 2.25710422e-01 2.16223881e-01 -7.12349892e-01
-9.06503916e-01 -9.31261852e-02 7.25531057e-02 5.41714072e-01
8.69419575e-01 -1.16476312e-01 -8.69020164e-01 5.26804209e-01
-5.36092281e-01 1.70032710e-01 3.03415298e-01 1.39130747e+00
-2.87068576e-01 -1.20600116e+00 -3.16670477e-01 2.87513077e-01
-3.29251319e-01 6.51900545e-02 -1.46514046e+00 8.83899629e-01
-3.75657529e-01 1.70997453e+00 2.06673011e-01 -4.81790870e-01
4.56554443e-01 4.85529125e-01 -1.67673826e-01 -5.80884695e-01
-1.38763487e+00 8.82048532e-02 1.09758556e+00 -7.47516632e-01
-8.98788393e-01 -8.14687967e-01 -1.06419981e+00 -2.02243850e-01
-3.79005820e-01 8.86160016e-01 8.97995174e-01 9.13820148e-01
6.38702691e-01 1.79653838e-01 3.40915948e-01 -5.19108534e-01
2.61566602e-02 -1.20344484e+00 -3.06313373e-02 1.41298920e-01
4.13116246e-01 -3.74724507e-01 -7.18612373e-01 -1.26962692e-01] | [12.268839836120605, 7.761037349700928] |
731ed057-9a9e-487b-8a85-0179b2103b29 | image-and-encoded-text-fusion-for-multi-modal | 1810.02001 | null | http://arxiv.org/abs/1810.02001v1 | http://arxiv.org/pdf/1810.02001v1.pdf | Image and Encoded Text Fusion for Multi-Modal Classification | Multi-modal approaches employ data from multiple input streams such as
textual and visual domains. Deep neural networks have been successfully
employed for these approaches. In this paper, we present a novel multi-modal
approach that fuses images and text descriptions to improve multi-modal
classification performance in real-world scenarios. The proposed approach
embeds an encoded text onto an image to obtain an information-enriched image.
To learn feature representations of resulting images, standard Convolutional
Neural Networks (CNNs) are employed for the classification task. We demonstrate
how a CNN based pipeline can be used to learn representations of the novel
fusion approach. We compare our approach with individual sources on two
large-scale multi-modal classification datasets while obtaining encouraging
results. Furthermore, we evaluate our approach against two famous multi-modal
strategies namely early fusion and late fusion. | ['Ignazio Gallo', 'Alessandro Calefati', 'Shah Nawaz', 'Muhammad Kamran Janjua'] | 2018-10-03 | null | null | null | null | ['multi-modal-classification'] | ['miscellaneous'] | [ 4.26144689e-01 -1.98615640e-01 -1.95184439e-01 -4.43584144e-01
-1.56176198e+00 -5.23695827e-01 1.20902085e+00 2.60390639e-01
-4.94209230e-01 5.08262694e-01 4.53177899e-01 4.39494640e-01
9.92771834e-02 -8.11802149e-01 -7.51450658e-01 -4.11269128e-01
5.47101676e-01 4.28819329e-01 2.26371348e-01 -3.75598431e-01
2.31470793e-01 2.38323629e-01 -2.04531741e+00 1.27629602e+00
2.89027691e-01 1.38976192e+00 -1.78741105e-02 9.41490293e-01
-3.28980774e-01 9.07579482e-01 -4.80037898e-01 -5.96306920e-01
2.58715171e-02 -9.09024477e-02 -1.00158429e+00 3.68818164e-01
6.02800250e-01 -2.83089161e-01 -4.45553720e-01 7.28937387e-01
4.23671991e-01 2.01738942e-02 8.33551288e-01 -1.33670557e+00
-7.36779809e-01 6.04863524e-01 -5.51825166e-01 5.02344295e-02
6.66645110e-01 1.09512158e-01 9.44647789e-01 -1.06052351e+00
5.30791104e-01 1.23253429e+00 6.69663787e-01 2.86827236e-01
-1.00656927e+00 -4.07931238e-01 -2.99938172e-01 2.18507618e-01
-1.29096401e+00 -7.58337975e-01 8.79522264e-01 -3.31019372e-01
9.03654337e-01 1.54014621e-02 3.43162268e-01 1.18587685e+00
2.84845471e-01 1.35477769e+00 1.07915950e+00 -6.41524971e-01
-1.60140097e-01 2.84839839e-01 -8.60903040e-02 6.55018866e-01
-9.06800181e-02 9.87537280e-02 -9.48177636e-01 5.11464030e-02
1.02361329e-01 3.06991786e-01 -4.15752456e-02 2.64738929e-02
-1.43654215e+00 8.76692176e-01 7.00811505e-01 6.91670239e-01
-6.60809815e-01 4.33406949e-01 5.88821173e-01 1.24694675e-01
5.41761696e-01 -1.24832571e-01 -1.21000096e-01 2.55724669e-01
-1.32067609e+00 3.40748012e-01 4.44322765e-01 7.93416321e-01
9.65800524e-01 -1.83269322e-01 -4.58136231e-01 9.18054402e-01
5.65601528e-01 3.77653956e-01 5.19273877e-01 -7.07248390e-01
5.98978996e-01 7.80674338e-01 -2.12853193e-01 -8.52469325e-01
-5.08244574e-01 -7.50704482e-02 -8.60581934e-01 1.32615507e-01
1.81151405e-01 1.58217549e-01 -1.12559390e+00 1.18184292e+00
1.80048067e-02 -5.30312210e-02 4.47667450e-01 5.51400840e-01
1.30900896e+00 5.59753597e-01 3.77840549e-01 3.84752750e-01
1.50703776e+00 -1.08013010e+00 -6.72365487e-01 -1.42659396e-01
3.81292999e-01 -9.94774640e-01 6.22851193e-01 2.22725242e-01
-1.34616852e+00 -6.13069475e-01 -9.85786557e-01 -1.38677195e-01
-8.94359410e-01 4.06438649e-01 1.49224699e-01 3.41918260e-01
-1.28003144e+00 3.89749438e-01 -5.04306674e-01 -6.92599118e-01
7.02766716e-01 3.74332041e-01 -6.38711333e-01 -3.46969396e-01
-1.01691437e+00 8.50419044e-01 8.90798509e-01 -1.95371896e-01
-1.09683907e+00 -3.62797320e-01 -1.04195285e+00 2.31596059e-03
6.61652759e-02 -7.84465611e-01 1.20915270e+00 -1.13622200e+00
-1.14846909e+00 1.06239331e+00 -1.10706054e-01 -4.40472096e-01
3.01708013e-01 -3.22609991e-02 -4.24447358e-01 6.08398199e-01
1.09355807e-01 1.22466636e+00 1.15542316e+00 -1.53858030e+00
-9.26974773e-01 -2.71681070e-01 2.48812720e-01 1.42415628e-01
-5.49606025e-01 9.08318758e-02 -4.45558548e-01 -6.24773562e-01
-1.35077223e-01 -6.12873673e-01 1.31075278e-01 -1.54525995e-01
-4.21393275e-01 -2.49685854e-01 1.00077295e+00 -5.09039342e-01
8.20380569e-01 -2.06692553e+00 1.92561015e-01 -2.30375901e-01
2.28531510e-01 -2.74699554e-02 -3.79388094e-01 6.89810872e-01
1.58182219e-01 4.05825637e-02 -2.49450400e-01 -8.95710230e-01
8.38273913e-02 2.14794576e-01 -8.15795884e-02 4.19035107e-01
4.65660602e-01 1.31770349e+00 -4.62394953e-01 -7.28338420e-01
5.53997457e-01 9.06830788e-01 -2.41392583e-01 2.05075622e-01
-2.54894406e-01 1.97589040e-01 1.86250862e-02 9.52866435e-01
4.95264590e-01 -4.14616764e-01 -2.78533161e-01 -9.47541237e-01
2.02338338e-01 -3.00121665e-01 -8.46244514e-01 2.04083276e+00
-6.36987746e-01 7.11406291e-01 -5.00096753e-02 -1.15202832e+00
5.44989109e-01 6.80809319e-01 5.80367327e-01 -7.91797161e-01
5.82909584e-01 4.91888002e-02 -7.46570349e-01 -4.52333629e-01
1.01115930e+00 -4.47015464e-01 -2.91951358e-01 5.41983604e-01
7.52958059e-01 1.05699869e-02 2.24840090e-01 2.89183795e-01
8.01111102e-01 2.01040372e-01 2.00618625e-01 1.98496804e-01
7.93545425e-01 2.87603259e-01 -2.02833652e-01 7.15430498e-01
-2.54494369e-01 8.53736281e-01 -1.01219431e-01 -4.92194265e-01
-1.09889197e+00 -7.90255666e-01 6.36181235e-02 1.31749916e+00
3.49558257e-02 -3.42083931e-01 -2.49943078e-01 -8.32797170e-01
-6.29186928e-02 3.37678820e-01 -9.19605970e-01 -7.96642154e-02
-1.90317437e-01 -7.11610377e-01 8.81148458e-01 7.33600438e-01
6.66599095e-01 -1.01782358e+00 -5.30449450e-01 1.28101572e-01
-5.06605208e-01 -1.32094884e+00 -7.12444559e-02 1.90212071e-01
-3.83799613e-01 -1.16024017e+00 -8.52636337e-01 -6.82018340e-01
3.45477164e-01 3.49644095e-01 1.13858759e+00 5.00028767e-03
-2.52445608e-01 9.38475430e-01 -6.73715770e-01 -2.58138955e-01
-6.12976253e-01 1.01897359e-01 -4.68040615e-01 3.70129168e-01
4.04190063e-01 -1.41324967e-01 -4.43553388e-01 -1.56499222e-01
-1.59037304e+00 2.60906368e-01 6.25308394e-01 7.85914004e-01
2.43718162e-01 3.09138503e-02 5.47115505e-01 -4.21134889e-01
5.04145205e-01 -6.48193002e-01 -6.22723326e-02 3.65718037e-01
-2.22338125e-01 1.31131381e-01 3.30028266e-01 -2.81211793e-01
-1.10542655e+00 3.27188849e-01 -2.61852711e-01 -5.74245036e-01
-6.28252268e-01 8.19674432e-01 2.40061492e-01 -1.44251615e-01
5.98202527e-01 4.54734236e-01 -1.85599372e-01 -3.26719195e-01
6.41853690e-01 1.06312788e+00 7.72983551e-01 -3.87670547e-01
6.24224663e-01 9.22428191e-01 -1.84167996e-01 -5.28182030e-01
-9.92338419e-01 -5.74043334e-01 -8.83545935e-01 -5.17783701e-01
1.11028552e+00 -1.25850630e+00 -5.66414475e-01 4.47898716e-01
-1.31197321e+00 9.91371498e-02 -4.59643006e-02 2.57112771e-01
-8.07637453e-01 1.37977570e-01 -5.39718747e-01 -6.44879043e-01
-4.87674087e-01 -1.18428516e+00 1.78094339e+00 2.96783090e-01
1.78984523e-01 -1.24577427e+00 1.72628835e-02 7.57404029e-01
6.50165558e-01 3.51273596e-01 4.60457444e-01 -8.32094133e-01
-3.79412651e-01 -4.36378747e-01 -4.37306583e-01 1.22482024e-01
-3.20447907e-02 -5.87103097e-03 -1.37150633e+00 -2.32664719e-01
-6.10256612e-01 -8.33457649e-01 1.22569752e+00 1.48479998e-01
7.99370050e-01 1.95863601e-02 -3.08881968e-01 3.99186432e-01
1.85190964e+00 -1.15646824e-01 4.90580320e-01 5.92432559e-01
6.95586503e-01 5.60908675e-01 4.55112904e-01 4.96995509e-01
8.01927209e-01 5.44539094e-01 5.91958821e-01 -1.40323475e-01
-3.56236458e-01 1.65141121e-01 5.04592538e-01 4.83644336e-01
2.63770908e-01 -5.83913386e-01 -1.08326209e+00 8.01947951e-01
-1.97278512e+00 -9.81929183e-01 4.60236073e-02 1.50874388e+00
5.74083149e-01 -1.51434273e-01 5.06923310e-02 2.66865551e-01
6.02779567e-01 1.95369139e-01 -1.39512479e-01 -7.76669830e-02
-4.41575706e-01 1.32424578e-01 2.51155138e-01 1.65466487e-01
-1.56196856e+00 7.22665370e-01 6.66179276e+00 7.35049486e-01
-1.12028897e+00 3.33624005e-01 3.70111555e-01 -2.63251103e-02
-1.06699787e-01 -3.63171458e-01 -6.06012166e-01 7.93606266e-02
1.29138887e+00 -4.80847098e-02 8.54872391e-02 3.57895225e-01
-3.83711368e-01 -2.41467819e-01 -1.12188506e+00 1.23384774e+00
5.57036817e-01 -1.88433802e+00 3.37659985e-01 -2.92920023e-01
7.20511198e-01 3.01565230e-01 3.37353200e-02 2.74973989e-01
4.17106211e-01 -9.93154705e-01 9.46053267e-01 7.44874060e-01
8.35060060e-01 -7.25958705e-01 1.00327945e+00 9.56954509e-02
-1.60558605e+00 -9.40716267e-02 1.78524196e-01 4.28074509e-01
2.31924281e-01 2.44579658e-01 -6.74760461e-01 1.10697627e+00
8.46868694e-01 1.07790697e+00 -1.00605404e+00 7.71881759e-01
3.35411191e-01 6.22776709e-02 -2.42683366e-01 4.40850586e-01
4.42554742e-01 5.32556534e-01 2.15629026e-01 1.53053379e+00
2.45654210e-01 -3.84224892e-01 2.52226710e-01 7.94805348e-01
-2.47553855e-01 4.31079566e-02 -7.90812016e-01 -2.33551160e-01
-5.76429181e-02 1.59777105e+00 -7.15970576e-01 -7.06408024e-01
-7.70020664e-01 9.93275523e-01 7.85658807e-02 2.58468837e-01
-7.24018335e-01 -2.11467311e-01 8.48311111e-02 -3.76221865e-01
3.90290916e-01 1.16422154e-01 -1.65636525e-01 -1.33176482e+00
-3.72585475e-01 -6.67462647e-01 7.44945645e-01 -1.30317378e+00
-1.43820655e+00 8.98088872e-01 1.25301853e-01 -1.26341856e+00
-3.24115366e-01 -5.47050714e-01 -3.02176028e-01 6.95208848e-01
-1.71155965e+00 -2.09194493e+00 -5.86375952e-01 1.06371284e+00
7.38814771e-01 -4.09952790e-01 7.15707839e-01 5.51828861e-01
-2.69600093e-01 2.81592667e-01 -1.31803855e-01 2.65294075e-01
7.92674124e-01 -1.15080738e+00 -1.44817591e-01 7.08682835e-01
1.62698373e-01 -1.96363833e-02 3.53722543e-01 -4.05648053e-01
-1.47723603e+00 -1.35750639e+00 5.63647926e-01 -3.88703972e-01
6.70737326e-01 -9.37445015e-02 -7.57983565e-01 7.76946664e-01
1.01882696e+00 6.06220812e-02 8.23283076e-01 -3.71489227e-01
-5.21090567e-01 4.50600237e-02 -1.23527420e+00 1.82200879e-01
2.84323245e-01 -8.18410218e-01 -6.75867796e-01 2.04109222e-01
5.45897961e-01 -2.25383729e-01 -1.13811409e+00 5.05768597e-01
5.13660431e-01 -9.21772182e-01 9.65280414e-01 -4.09651518e-01
8.95487070e-01 -2.94426620e-01 -7.25152016e-01 -1.23482907e+00
4.65066321e-02 8.15516263e-02 -8.21007788e-02 1.38915312e+00
1.08099028e-01 -1.44083411e-01 5.08925676e-01 1.77373186e-01
-1.05937868e-01 -2.60104060e-01 -8.76266301e-01 -2.05263779e-01
7.37452507e-02 -6.21272206e-01 4.38477576e-01 7.62813807e-01
-1.78524897e-01 3.97156864e-01 -4.32400703e-01 2.82032415e-02
6.51718676e-01 1.60135999e-01 6.85710013e-01 -1.13779056e+00
1.67672634e-01 -4.25310344e-01 -5.29962361e-01 -3.45276058e-01
3.94900560e-01 -1.19044554e+00 -2.84567568e-02 -1.78176570e+00
5.08719265e-01 2.15108693e-02 -4.24432456e-01 7.37812459e-01
1.39576644e-01 9.22270596e-01 5.10234714e-01 1.55711949e-01
-9.98658538e-01 5.83990455e-01 7.39090741e-01 -7.05643356e-01
3.48098516e-01 -5.09590089e-01 -6.36770606e-01 5.70044339e-01
6.41619623e-01 -3.17296833e-01 4.31271642e-02 -4.24686402e-01
2.97000110e-01 1.68133527e-01 7.49665260e-01 -1.17252219e+00
4.96485382e-01 1.47605270e-01 6.33773923e-01 -1.16436875e+00
5.49449205e-01 -1.12213910e+00 1.81101114e-01 1.83645681e-01
-5.58572829e-01 -1.73558900e-03 5.09847760e-01 5.09575009e-01
-7.69947708e-01 -1.70700531e-02 8.39136004e-01 -1.80693358e-01
-9.07093048e-01 1.25290498e-01 -3.45406622e-01 -4.40061361e-01
1.07404947e+00 1.56365484e-02 -6.45327926e-01 -3.80628496e-01
-9.43060696e-01 8.67780969e-02 4.14990246e-01 6.70680225e-01
9.07603323e-01 -1.69762039e+00 -8.42617512e-01 1.06144093e-01
6.90584660e-01 -4.21934217e-01 2.97964990e-01 1.07061672e+00
-2.85773158e-01 5.12186706e-01 -3.94377083e-01 -9.61308956e-01
-1.38064682e+00 6.23601675e-01 4.03688788e-01 -2.35733166e-01
-3.96295160e-01 3.48653466e-01 -6.92034736e-02 -3.80861223e-01
3.99379432e-02 -1.86895579e-02 -4.05724734e-01 5.31204879e-01
7.52163112e-01 3.92923430e-02 2.37149835e-01 -1.17863441e+00
-3.33851844e-01 5.58668196e-01 -2.56748255e-02 -3.95044237e-01
1.53006291e+00 -5.43232560e-01 -8.15938264e-02 5.54265916e-01
1.59310412e+00 -5.79410076e-01 -8.67846668e-01 -6.34785473e-01
-5.83728449e-03 -1.81165516e-01 3.21302086e-01 -8.71965826e-01
-1.30737865e+00 8.75461757e-01 8.10553193e-01 3.07774961e-01
1.50507808e+00 3.62934679e-01 6.81811810e-01 2.32550874e-01
9.62166637e-02 -8.82521451e-01 3.26644808e-01 3.53454351e-01
8.69718611e-01 -1.87473941e+00 -1.28821075e-01 8.10762122e-02
-6.96586072e-01 1.47573054e+00 3.22236508e-01 4.32564244e-02
7.20162928e-01 2.24213660e-01 5.78587279e-02 -5.44362187e-01
-9.84074473e-01 -6.97708786e-01 6.77335322e-01 4.40489054e-01
4.71263170e-01 -3.26726228e-01 1.74072802e-01 3.50095958e-01
3.54346246e-01 2.72297144e-01 4.71407533e-01 1.19604659e+00
-4.47472900e-01 -1.00935674e+00 -7.10498214e-01 4.43640500e-01
-6.59105778e-01 -4.18779887e-02 -3.49111497e-01 7.29678929e-01
1.74918592e-01 1.06712186e+00 1.05509743e-01 -6.09528720e-01
1.67765021e-01 3.68437320e-01 4.50604051e-01 -4.20883536e-01
-9.84281778e-01 2.86355257e-01 -9.34694782e-02 -5.81015110e-01
-1.35669959e+00 -5.22315443e-01 -1.00353527e+00 -2.59274662e-01
-4.24352176e-02 -3.91424984e-01 8.26978087e-01 1.24254775e+00
2.63129860e-01 8.04152489e-01 4.50863659e-01 -1.25511205e+00
2.31732473e-01 -9.51458752e-01 -3.12446475e-01 7.25525081e-01
4.99658793e-01 -6.82887971e-01 -1.25816062e-01 6.11837029e-01] | [10.768221855163574, 1.516209363937378] |
05074f29-808b-435c-99c1-32ae4fe94d94 | towards-learning-discrete-representations-via | 2306.01108 | null | https://arxiv.org/abs/2306.01108v1 | https://arxiv.org/pdf/2306.01108v1.pdf | Towards Learning Discrete Representations via Self-Supervision for Wearables-Based Human Activity Recognition | Human activity recognition (HAR) in wearable computing is typically based on direct processing of sensor data. Sensor readings are translated into representations, either derived through dedicated preprocessing, or integrated into end-to-end learning. Independent of their origin, for the vast majority of contemporary HAR, those representations are typically continuous in nature. That has not always been the case. In the early days of HAR, discretization approaches have been explored - primarily motivated by the desire to minimize computational requirements, but also with a view on applications beyond mere recognition, such as, activity discovery, fingerprinting, or large-scale search. Those traditional discretization approaches, however, suffer from substantial loss in precision and resolution in the resulting representations with detrimental effects on downstream tasks. Times have changed and in this paper we propose a return to discretized representations. We adopt and apply recent advancements in Vector Quantization (VQ) to wearables applications, which enables us to directly learn a mapping between short spans of sensor data and a codebook of vectors, resulting in recognition performance that is generally on par with their contemporary, continuous counterparts - sometimes surpassing them. Therefore, this work presents a proof-of-concept for demonstrating how effective discrete representations can be derived, enabling applications beyond mere activity classification but also opening up the field to advanced tools for the analysis of symbolic sequences, as they are known, for example, from domains such as natural language processing. Based on an extensive experimental evaluation on a suite of wearables-based benchmark HAR tasks, we demonstrate the potential of our learned discretization scheme and discuss how discretized sensor data analysis can lead to substantial changes in HAR. | ['Thomas Ploetz', 'Irfan Essa', 'Harish Haresamudram'] | 2023-06-01 | null | null | null | null | ['human-activity-recognition', 'quantization', 'human-activity-recognition'] | ['computer-vision', 'methodology', 'time-series'] | [ 7.45448828e-01 6.92314729e-02 -2.82798409e-01 -4.75384235e-01
-8.52086484e-01 -5.81306696e-01 4.90094841e-01 5.91496646e-01
-3.77135515e-01 8.23617995e-01 3.39814574e-01 -5.76201454e-02
-3.46493691e-01 -8.79207671e-01 -6.80536032e-01 -5.88323295e-01
-3.59325290e-01 -1.02976738e-02 -8.72008428e-02 -1.45905733e-01
-7.13543147e-02 4.22924906e-01 -1.75729716e+00 1.50987595e-01
3.45887750e-01 1.42370558e+00 -4.70715702e-01 4.10852134e-01
-1.29962578e-01 5.25390863e-01 -5.21580040e-01 -1.09048627e-01
1.98270962e-01 -5.67976177e-01 -4.33855593e-01 6.08630590e-02
1.86149582e-01 4.17813100e-02 -1.16981119e-01 6.17422640e-01
4.56599414e-01 2.18755394e-01 5.79167128e-01 -1.03293085e+00
-2.66636431e-01 5.38537562e-01 -2.29478002e-01 -1.31945893e-01
6.88811839e-01 -7.64272362e-02 8.90607715e-01 -4.90507990e-01
3.43067825e-01 9.09886479e-01 1.20664036e+00 3.46590281e-01
-1.59856188e+00 -2.85104007e-01 3.54419313e-02 -6.90512881e-02
-1.59657097e+00 -5.21704733e-01 6.62536085e-01 -4.25796509e-01
9.13913488e-01 5.39107680e-01 8.60178232e-01 1.48373044e+00
-8.41623023e-02 5.98561347e-01 9.98708963e-01 -4.82621849e-01
7.51777232e-01 -5.34157194e-02 1.26612082e-01 2.91551024e-01
5.02094924e-01 -1.38330400e-01 -6.52911842e-01 -2.79711872e-01
5.21625817e-01 3.16684276e-01 -3.10687184e-01 -4.51482952e-01
-1.23134005e+00 7.28899717e-01 1.88186973e-01 4.40775245e-01
-6.14585340e-01 2.98982978e-01 6.49397790e-01 1.44278333e-01
1.72586501e-01 3.40914011e-01 -3.44554812e-01 -5.96044481e-01
-1.21584380e+00 3.98413688e-01 9.58145082e-01 6.31581724e-01
6.24957442e-01 9.76460353e-02 -2.43497074e-01 6.80453300e-01
-6.45890832e-02 2.80575246e-01 6.74402952e-01 -7.40696251e-01
2.05754444e-01 5.95544457e-01 1.69000670e-01 -9.88715231e-01
-3.80382210e-01 -3.45571369e-01 -1.02174866e+00 -1.66071519e-01
3.69876206e-01 -4.57018875e-02 -6.04874671e-01 1.80564618e+00
2.38631636e-01 1.30780444e-01 -1.60277680e-01 7.15731502e-01
7.82374814e-02 3.41869146e-01 2.17883185e-01 -4.10238355e-01
1.16543376e+00 -2.14164048e-01 -6.84674561e-01 8.82135853e-02
5.70249200e-01 -6.78072795e-02 1.10083163e+00 8.52180898e-01
-7.89778769e-01 -4.42147881e-01 -1.28924751e+00 9.71078575e-02
-6.43922567e-01 -2.47266680e-01 8.12154233e-01 1.09095192e+00
-8.12283754e-01 8.99469733e-01 -1.00497532e+00 -6.72223091e-01
6.14311457e-01 3.93433690e-01 -3.67304206e-01 6.04389012e-02
-1.22315407e+00 5.24672031e-01 2.75743455e-01 6.32415190e-02
-5.14295161e-01 -6.46973908e-01 -8.95073593e-01 8.83172527e-02
1.88188627e-01 -4.88472521e-01 9.66077685e-01 -9.35712636e-01
-1.45979047e+00 4.17366326e-01 9.42619219e-02 -1.01794851e+00
6.02024376e-01 -2.50663489e-01 -4.77919787e-01 -3.42048891e-02
-9.77377817e-02 3.27422261e-01 8.36268008e-01 -6.22310221e-01
-3.45013529e-01 -4.39665258e-01 -1.16834365e-01 -1.36416644e-01
-8.78781199e-01 -3.54023755e-01 5.31508215e-02 -6.95756137e-01
-1.35814279e-01 -8.01988423e-01 -2.53763288e-01 1.19387746e-01
-1.99742779e-01 -6.44935817e-02 4.83822197e-01 -5.06381571e-01
1.42409217e+00 -2.15240121e+00 1.12658173e-01 3.43078107e-01
7.73394853e-02 1.59621298e-01 2.50287354e-01 7.22490549e-01
9.20724851e-05 2.19506603e-02 -5.26143432e-01 -2.56587863e-01
1.78086996e-01 2.98451543e-01 -4.50092912e-01 7.66637743e-01
1.70479104e-01 8.97588670e-01 -8.90361726e-01 -2.44511276e-01
2.00830802e-01 6.96842968e-01 -3.19043130e-01 8.89308900e-02
-1.87667504e-01 3.35465282e-01 -4.61043745e-01 6.02281511e-01
1.11370794e-01 -7.33116567e-02 3.16048056e-01 -1.88433185e-01
-1.01158939e-01 2.23208830e-01 -1.23262525e+00 1.95083082e+00
-4.69375670e-01 4.06186104e-01 -2.60339379e-01 -1.57180142e+00
8.68349612e-01 2.68114239e-01 1.04223168e+00 -7.19172478e-01
4.00023349e-02 8.18454474e-02 -3.52402508e-01 -2.23709106e-01
3.32744002e-01 -6.90021645e-03 -4.72540855e-01 2.76780188e-01
-6.70873001e-02 2.55135894e-01 7.77241737e-02 -2.80673742e-01
1.33244646e+00 4.16553497e-01 7.03912377e-01 -1.27981663e-01
3.14185381e-01 8.01881659e-04 4.57741320e-01 7.65551388e-01
-8.85731056e-02 5.39525807e-01 1.97407320e-01 -3.38213652e-01
-7.52540112e-01 -1.33191240e+00 -4.47710067e-01 9.17651176e-01
-2.05809042e-01 -7.47532189e-01 -8.08146238e-01 -3.66167277e-01
2.17574090e-01 5.06744325e-01 -7.54677534e-01 -3.24582785e-01
-5.18327057e-01 -6.57714069e-01 9.31608319e-01 6.47636831e-01
2.30873361e-01 -7.75046587e-01 -1.33345163e+00 6.07970893e-01
1.98534027e-01 -1.14868557e+00 6.40457496e-02 5.05514681e-01
-1.11896610e+00 -8.14993441e-01 -6.59791291e-01 -1.20819239e-02
1.62252292e-01 -2.44520027e-02 8.81077945e-01 -2.68430352e-01
-6.00141883e-01 8.32498848e-01 -4.81249243e-01 -4.66547996e-01
2.45728064e-02 1.29964858e-01 3.58303845e-01 3.59736115e-01
4.42895472e-01 -1.13356280e+00 -7.33575761e-01 -6.31093457e-02
-9.94729757e-01 -3.99949580e-01 6.16433203e-01 6.12240911e-01
6.80926323e-01 -2.51389563e-01 7.89095819e-01 -6.80747032e-01
7.12275386e-01 -7.32910097e-01 -2.97802567e-01 1.39324978e-01
-7.53786266e-01 7.70083889e-02 8.66363049e-01 -6.48935378e-01
-4.44022089e-01 2.60664910e-01 -1.91625897e-02 -5.04763961e-01
-2.64510363e-01 6.09530687e-01 -4.53938125e-03 9.08701941e-02
8.68386984e-01 4.46367085e-01 -5.03568500e-02 -5.46819985e-01
5.46531796e-01 8.22373807e-01 6.66454911e-01 -7.15856075e-01
4.35867965e-01 5.42877316e-01 1.40668407e-01 -1.23951674e+00
-5.20792603e-01 -4.27081406e-01 -5.53065002e-01 2.93229651e-02
7.64959335e-01 -7.72294104e-01 -5.88798285e-01 -1.19272172e-02
-6.81031942e-01 -1.93289846e-01 -8.17741513e-01 3.17873120e-01
-7.07931399e-01 4.67313200e-01 -2.13348806e-01 -1.07601321e+00
-3.10904711e-01 -6.58920884e-01 1.17414784e+00 2.05247590e-04
-7.17388749e-01 -8.30720603e-01 1.02763645e-01 1.11790545e-01
4.92355943e-01 7.30157077e-01 7.51822531e-01 -6.78573370e-01
-1.23416774e-01 -3.93568575e-01 2.59007812e-01 4.90458280e-01
3.63403052e-01 -5.18577397e-01 -1.14782429e+00 -2.44464070e-01
-1.69262156e-01 -3.71249110e-01 6.11960769e-01 7.16441795e-02
1.34047914e+00 -2.36677155e-01 -2.95857757e-01 5.12495041e-01
1.28135753e+00 -3.62071618e-02 7.27408528e-01 9.96031091e-02
4.77401584e-01 2.51300901e-01 5.43476403e-01 8.24442506e-01
1.19895503e-01 9.22104955e-01 2.00864166e-01 8.15345868e-02
2.07089763e-02 -2.61384666e-01 4.72346812e-01 5.12146652e-01
-1.55969217e-01 1.33479774e-01 -7.11678922e-01 5.97497582e-01
-1.84851658e+00 -1.01322639e+00 2.25730121e-01 2.55732679e+00
9.19040442e-01 7.94181004e-02 5.68114877e-01 8.19972038e-01
2.34252229e-01 3.64313275e-02 -5.58049202e-01 -5.27167141e-01
1.91220760e-01 6.06202185e-01 5.61578155e-01 -7.76832327e-02
-1.13360107e+00 2.74518460e-01 5.77721739e+00 6.35376275e-01
-1.25794935e+00 1.17126510e-01 2.78650165e-01 -2.53956258e-01
5.73925965e-04 -3.22893471e-01 -5.70536435e-01 5.81383884e-01
1.46810889e+00 -7.22693354e-02 5.89141548e-01 1.01028144e+00
1.69138014e-01 -3.55434790e-02 -1.53509605e+00 1.41489983e+00
-1.27172932e-01 -1.20744407e+00 -1.63635127e-02 3.03841949e-01
3.76112282e-01 -2.50917196e-01 -1.45458132e-01 5.08259833e-01
-2.39542425e-01 -1.24989140e+00 7.59094834e-01 5.83148658e-01
9.57035542e-01 -5.57828248e-01 3.11689913e-01 3.61172378e-01
-1.30980670e+00 -2.84385651e-01 -1.28566831e-01 -3.34440976e-01
-4.59346548e-02 7.78969824e-01 -7.47529268e-01 6.33245885e-01
6.02837324e-01 6.70538843e-01 -3.41399163e-01 7.54206419e-01
3.59923065e-01 8.12407434e-01 -5.01576424e-01 -1.78519011e-01
9.50834975e-02 -9.70851481e-02 2.63538808e-01 1.44079316e+00
3.57036233e-01 2.57660821e-02 2.47598782e-01 6.98692322e-01
3.16732973e-02 2.24857796e-02 -8.07085991e-01 -3.74100626e-01
5.59862912e-01 9.63766992e-01 -5.91393113e-01 -1.90126181e-01
-4.42413092e-01 9.66086805e-01 2.28948504e-01 2.54212707e-01
-9.26140904e-01 -4.74774927e-01 8.47600102e-01 5.16656578e-01
4.20120984e-01 -4.23089027e-01 -3.97165239e-01 -9.79056478e-01
2.37432808e-01 -9.00823712e-01 3.56202394e-01 -6.41400069e-02
-1.13024426e+00 1.99664235e-01 1.55026034e-01 -1.34329259e+00
-5.39978504e-01 -4.19991314e-01 -2.52776057e-01 6.08906329e-01
-1.07146621e+00 -8.49041283e-01 -3.64600182e-01 6.62568986e-01
4.11592275e-01 2.45153993e-01 1.22936761e+00 4.87696201e-01
-3.00992489e-01 6.75967813e-01 1.30553514e-01 1.45581916e-01
3.38563710e-01 -1.06756449e+00 2.23832861e-01 6.01015449e-01
4.50855136e-01 8.41442883e-01 5.41804433e-01 -3.59806627e-01
-1.88238811e+00 -1.15204453e+00 3.71158153e-01 -3.84186119e-01
5.13388395e-01 -7.31925905e-01 -7.09709823e-01 5.68994701e-01
-2.74073750e-01 3.02115291e-01 8.79780829e-01 1.97243273e-01
-3.60980153e-01 -4.35872525e-01 -1.12810957e+00 4.17768985e-01
1.30468595e+00 -6.32460177e-01 -5.60025096e-01 5.36806770e-02
3.21876138e-01 -3.77541855e-02 -1.23201871e+00 4.37788785e-01
9.45092797e-01 -8.97985399e-01 1.17720270e+00 -7.61819780e-01
-1.28738051e-02 -2.38861695e-01 -5.67959189e-01 -9.57934201e-01
-1.03223383e-01 -6.11424863e-01 -6.40346885e-01 1.21527648e+00
-6.73692450e-02 -5.50757289e-01 6.58204556e-01 5.40799499e-01
7.13026300e-02 -9.58318830e-01 -1.15549767e+00 -9.69542921e-01
-2.15071857e-01 -7.54082203e-01 4.88807082e-01 9.59295392e-01
1.04462139e-01 1.70246229e-01 -2.59941608e-01 -1.43022746e-01
7.12063670e-01 6.67288825e-02 6.44116998e-01 -1.44972312e+00
-5.53943515e-01 -2.50092715e-01 -9.78419304e-01 -9.34802711e-01
-2.46856734e-01 -6.73194766e-01 -5.15448712e-02 -1.16744304e+00
-3.49703401e-01 -3.02820563e-01 -6.08245969e-01 5.83343506e-01
4.17397559e-01 4.85941350e-01 -1.06065273e-01 8.71240944e-02
-5.18576562e-01 4.43878561e-01 3.66566688e-01 -2.25394756e-01
-3.81297082e-01 -7.90922865e-02 -6.34264708e-01 4.98566717e-01
7.02980042e-01 -3.44874442e-01 -6.60128653e-01 8.25878233e-02
2.15964004e-01 6.97466657e-02 4.95574683e-01 -1.63835561e+00
-5.12548834e-02 -8.49960819e-02 5.50985694e-01 4.53265980e-02
5.27745306e-01 -8.54018390e-01 2.85878628e-01 3.58107805e-01
-3.88805300e-01 -2.92157590e-01 -2.62089949e-02 9.03161645e-01
-1.12011448e-01 2.49175891e-01 4.59146082e-01 -4.30691615e-02
-5.99467218e-01 7.72101209e-02 -2.37542719e-01 -8.30021035e-03
1.00946701e+00 -6.40214145e-01 3.07082236e-01 -2.68648058e-01
-9.60457683e-01 -2.73296773e-01 2.07140073e-01 3.93790424e-01
4.95708197e-01 -1.28386188e+00 -3.17768872e-01 2.06548035e-01
2.63961613e-01 -2.05858856e-01 -8.41177627e-02 1.04956937e+00
-4.94672321e-02 4.39704388e-01 -1.83422491e-01 -7.16555715e-01
-8.38473201e-01 5.21511316e-01 2.02542022e-01 5.58992960e-02
-8.56849253e-01 1.97246045e-01 -1.98322475e-01 5.43374531e-02
5.26869476e-01 -8.19166005e-01 3.95125039e-02 3.38016391e-01
5.75674057e-01 3.69736105e-01 4.18879449e-01 -2.53178447e-01
-5.89668691e-01 4.34655398e-01 3.48179340e-01 -1.24896120e-03
1.38946307e+00 7.97051098e-03 3.52978110e-01 9.21247184e-01
1.14888179e+00 -1.65606156e-01 -1.18262041e+00 6.04553558e-02
2.89283693e-01 -2.37619728e-01 -1.24165371e-01 -5.92117727e-01
-5.79285026e-01 8.31751406e-01 9.73752081e-01 4.93350923e-01
1.19562721e+00 -1.71961382e-01 9.19243515e-01 6.24542177e-01
8.68518710e-01 -1.03113997e+00 -2.29603842e-01 2.89671239e-03
5.75263083e-01 -7.71029592e-01 5.10217771e-02 -8.21820125e-02
-2.36904934e-01 1.04384279e+00 -1.74567044e-01 -1.76247329e-01
5.10206699e-01 3.44748259e-01 -2.08962634e-01 6.48981333e-02
-4.12637949e-01 -1.12427391e-01 -2.95135435e-02 9.44784045e-01
5.66905260e-01 1.46610454e-01 -3.88955116e-01 9.73776877e-01
-3.79523225e-02 4.47433889e-01 2.53815830e-01 1.22023988e+00
-3.36511046e-01 -1.09954846e+00 -3.91543180e-01 5.89941144e-01
-3.94488722e-01 1.49616942e-01 -2.15074301e-01 6.39839590e-01
2.93042064e-01 8.51554036e-01 -1.33286923e-01 -4.15175796e-01
5.25532663e-01 4.12984341e-01 6.11455917e-01 -4.64063853e-01
-5.54897547e-01 -3.41197401e-01 1.57411322e-01 -9.17901695e-01
-3.98396790e-01 -9.24220622e-01 -1.28210711e+00 2.34299954e-02
2.67859459e-01 2.10705101e-01 7.56933987e-01 8.94945920e-01
5.16817987e-01 4.70780373e-01 3.68856847e-01 -9.50507879e-01
-1.01200533e+00 -6.78643823e-01 -6.64682329e-01 5.42504370e-01
3.00445527e-01 -8.17761242e-01 -4.78606522e-02 1.77732483e-01] | [7.702676296234131, 1.0347967147827148] |
b8073561-3e69-4ba9-86e4-c5710bed09a1 | romanization-based-large-scale-adaptation-of | 2304.08865 | null | https://arxiv.org/abs/2304.08865v1 | https://arxiv.org/pdf/2304.08865v1.pdf | Romanization-based Large-scale Adaptation of Multilingual Language Models | Large multilingual pretrained language models (mPLMs) have become the de facto state of the art for cross-lingual transfer in NLP. However, their large-scale deployment to many languages, besides pretraining data scarcity, is also hindered by the increase in vocabulary size and limitations in their parameter budget. In order to boost the capacity of mPLMs to deal with low-resource and unseen languages, we explore the potential of leveraging transliteration on a massive scale. In particular, we explore the UROMAN transliteration tool, which provides mappings from UTF-8 to Latin characters for all the writing systems, enabling inexpensive romanization for virtually any language. We first focus on establishing how UROMAN compares against other language-specific and manually curated transliterators for adapting multilingual PLMs. We then study and compare a plethora of data- and parameter-efficient strategies for adapting the mPLMs to romanized and non-romanized corpora of 14 diverse low-resource languages. Our results reveal that UROMAN-based transliteration can offer strong performance for many languages, with particular gains achieved in the most challenging setups: on languages with unseen scripts and with limited training data without any vocabulary augmentation. Further analyses reveal that an improved tokenizer based on romanized data can even outperform non-transliteration-based methods in the majority of languages. | ['Ivan Vulić', 'Iryna Gurevych', 'Jonas Pfeiffer', 'Sebastian Ruder', 'Sukannya Purkayastha'] | 2023-04-18 | null | null | null | null | ['transliteration', 'cross-lingual-transfer'] | ['natural-language-processing', 'natural-language-processing'] | [-9.96497273e-02 -3.60754758e-01 -2.94601500e-01 -2.18807235e-01
-1.19900477e+00 -1.11675060e+00 5.35475910e-01 -2.59568632e-01
-9.88696098e-01 1.02060008e+00 8.62614587e-02 -7.98290253e-01
3.48743677e-01 -3.83063883e-01 -8.74434710e-01 -2.46897563e-01
3.17412853e-01 9.78468955e-01 -2.25957185e-01 -5.27428567e-01
-4.07614142e-01 3.42545152e-01 -7.66876876e-01 1.88072577e-01
1.24964273e+00 1.93637133e-01 4.32168901e-01 3.99826527e-01
-2.79016912e-01 1.90829799e-01 -5.21245778e-01 -7.39717901e-01
3.67649078e-01 -2.60290951e-01 -8.36007416e-01 -3.60667557e-01
8.01241577e-01 -2.12304175e-01 -2.40113456e-02 7.98685610e-01
6.74553037e-01 3.12497355e-02 4.97262299e-01 -5.96137822e-01
-1.09737754e+00 1.13366663e+00 -1.38093561e-01 2.37953961e-01
3.38322163e-01 1.98531687e-01 9.23680544e-01 -1.14032543e+00
9.14355457e-01 1.33178139e+00 6.73251688e-01 6.97397709e-01
-1.21735096e+00 -7.73483574e-01 1.81846067e-01 1.88949164e-02
-1.70656633e+00 -8.01066935e-01 2.79428005e-01 -2.28428140e-01
1.50963438e+00 9.95435789e-02 1.14420854e-01 1.21002555e+00
-1.46914244e-01 8.16264331e-01 1.17256927e+00 -8.58338237e-01
-3.32933724e-01 6.69918478e-01 -3.18136811e-01 6.33245647e-01
1.61935404e-01 -1.47729203e-01 -4.91486251e-01 2.39292726e-01
5.47592223e-01 -5.61777949e-01 -2.80173004e-01 1.56191245e-01
-1.47432995e+00 7.31211901e-01 2.14448366e-02 5.95475376e-01
7.65177980e-03 -1.67968288e-01 6.71024978e-01 7.90629387e-01
7.24060237e-01 8.07563841e-01 -1.01565182e+00 -1.51767239e-01
-8.87490690e-01 -1.54223397e-01 8.32441986e-01 1.29646146e+00
8.52955401e-01 4.84532714e-01 -3.54995904e-03 1.21747673e+00
1.79543514e-02 1.08243906e+00 7.62847900e-01 -2.98962414e-01
1.06717873e+00 2.79222995e-01 -2.40332708e-01 -2.00457573e-01
-1.53310403e-01 -3.62250298e-01 -5.11140943e-01 -3.34362388e-01
4.75880414e-01 -3.95602375e-01 -9.04642940e-01 1.82199192e+00
1.75599173e-01 -4.28382903e-01 3.01547945e-01 4.58920449e-01
6.11732066e-01 7.13870704e-01 2.03427821e-01 -1.62319943e-01
1.26851618e+00 -1.06196606e+00 -6.22015357e-01 -4.68050987e-01
1.12079167e+00 -1.12925339e+00 1.65469599e+00 1.06180884e-01
-8.79621029e-01 -5.91912210e-01 -8.38192880e-01 -3.54993224e-01
-7.21416473e-01 3.62073123e-01 5.05787253e-01 7.64675856e-01
-1.18932462e+00 3.03608209e-01 -7.16805637e-01 -7.46672988e-01
-2.06004623e-02 3.21663707e-01 -6.90505803e-01 -4.16803241e-01
-1.38798594e+00 1.37403393e+00 6.45322442e-01 -2.07395088e-02
-8.07814777e-01 -6.80213690e-01 -1.06465364e+00 -2.29839131e-01
1.21036626e-01 -3.34123999e-01 1.08755755e+00 -1.04112875e+00
-1.66619432e+00 1.06267691e+00 -1.23741098e-01 -3.32233667e-01
5.44709325e-01 -3.34388167e-01 -5.37939787e-01 -3.73933077e-01
5.33730239e-02 7.10102558e-01 5.14849722e-01 -8.68643165e-01
-5.46746671e-01 1.05582751e-01 -1.54383928e-01 5.77904761e-01
-8.54862154e-01 4.54832584e-01 -6.31835878e-01 -5.83036005e-01
-4.18802083e-01 -9.36920702e-01 -5.86968213e-02 -6.36221588e-01
-2.24209428e-01 -3.29428196e-01 4.67649549e-01 -1.04956782e+00
1.26115298e+00 -1.83096385e+00 2.04245895e-01 -5.10739349e-02
-5.55575848e-01 7.66844332e-01 -5.43542922e-01 6.20435894e-01
2.25593373e-01 3.66931289e-01 -1.27489537e-01 -7.37449050e-01
1.32381126e-01 5.82164884e-01 -3.31831515e-01 2.38071159e-01
2.78753996e-01 1.05300450e+00 -9.31830347e-01 -5.55987060e-01
2.57780254e-01 5.95821381e-01 -4.88228142e-01 1.39180526e-01
-1.94432706e-01 7.49066353e-01 -9.54125896e-02 6.93026483e-01
2.99315661e-01 2.73689389e-01 4.27394509e-01 1.93936855e-01
-2.84866065e-01 6.50994897e-01 -8.08464408e-01 1.87694216e+00
-1.14615309e+00 4.91104811e-01 -1.64276198e-01 -5.93950212e-01
7.88289249e-01 5.30024886e-01 -2.49464298e-03 -7.53746569e-01
-4.61411849e-02 9.51093256e-01 8.36187508e-03 -2.42507726e-01
7.23849416e-01 -2.49890089e-01 -3.75544071e-01 5.26602209e-01
5.81243694e-01 -3.19273442e-01 4.19768482e-01 -1.37785226e-01
7.16859519e-01 4.23439413e-01 4.04802293e-01 -4.41814721e-01
7.55106628e-01 1.47545427e-01 3.56888324e-01 5.76170385e-01
1.29910884e-02 3.67174208e-01 -2.67879218e-01 -2.56187350e-01
-1.18066275e+00 -1.02531123e+00 -2.76509553e-01 1.50610352e+00
-5.16118944e-01 -3.71754318e-01 -8.45426679e-01 -7.83127785e-01
-1.06353462e-01 7.07016766e-01 -6.57325760e-02 1.90677837e-01
-1.16571414e+00 -6.14751697e-01 9.11037505e-01 2.12504491e-01
2.20743597e-01 -1.33749211e+00 3.28471303e-01 3.88583332e-01
-3.87253761e-01 -1.61842573e+00 -8.36208224e-01 2.70793080e-01
-5.77031851e-01 -5.80004752e-01 -8.05652201e-01 -1.18062997e+00
6.77599132e-01 -1.63585544e-01 1.31072915e+00 -2.23354802e-01
4.08750251e-02 2.59300172e-01 -3.49449009e-01 -2.87797749e-01
-8.36992681e-01 9.00971413e-01 5.57787240e-01 -3.03492159e-01
5.33328414e-01 -3.25125277e-01 2.21455351e-01 1.73976809e-01
-6.84565723e-01 -1.86726570e-01 5.92696548e-01 5.92880726e-01
7.02587545e-01 -4.14266825e-01 5.44072866e-01 -1.16498911e+00
5.59107065e-01 -3.59291732e-01 -6.35457754e-01 4.93023574e-01
-5.10503054e-01 2.89654970e-01 1.27997303e+00 -8.01667571e-01
-1.02677882e+00 -9.22912359e-02 -3.71253312e-01 -2.08870038e-01
-9.93491244e-03 5.49706161e-01 -2.90358573e-01 -1.60778090e-01
7.72066772e-01 3.10429394e-01 -6.75615549e-01 -8.29413533e-01
6.65961385e-01 8.23340058e-01 5.74614525e-01 -9.25287485e-01
1.01100492e+00 1.96710546e-02 -5.52731991e-01 -8.83875489e-01
-7.25760043e-01 -3.52842897e-01 -9.53052640e-01 2.32575640e-01
6.68810785e-01 -1.10649776e+00 -9.90312845e-02 4.50454980e-01
-1.23997629e+00 -6.95131660e-01 -3.35303187e-01 6.48391604e-01
-3.09966505e-01 1.95518821e-01 -7.68400788e-01 -3.62728417e-01
-6.15556896e-01 -1.21540380e+00 8.23115408e-01 -8.96095112e-02
-2.49331102e-01 -1.57245624e+00 3.97705346e-01 3.40686142e-01
7.26984501e-01 -2.24819332e-01 1.02268016e+00 -9.05722857e-01
-3.23492378e-01 -2.13449281e-02 5.18724956e-02 6.87607586e-01
2.65686154e-01 -3.53928179e-01 -8.31369519e-01 -6.79699361e-01
-4.41192478e-01 -6.77212954e-01 2.47334585e-01 -1.77297950e-01
2.59209335e-01 -3.37664306e-01 1.37709212e-02 1.12356865e+00
1.43125772e+00 -1.52761221e-01 2.27461830e-01 5.42542577e-01
1.07819819e+00 3.93394589e-01 3.48544031e-01 -1.60828888e-01
5.96968710e-01 6.78676903e-01 -1.48041219e-01 -3.63929600e-01
-4.54148144e-01 -5.29228091e-01 9.56555367e-01 1.79283106e+00
1.14173286e-01 -3.37597549e-01 -1.16654134e+00 8.23476911e-01
-1.41022003e+00 -2.02276051e-01 6.63265660e-02 2.43740869e+00
1.45682657e+00 -2.15031967e-01 -2.94260889e-01 -5.44816375e-01
5.73245347e-01 -9.08477046e-03 -3.76868933e-01 -7.92671442e-01
-5.32504201e-01 6.77823007e-01 6.57519102e-01 8.79230976e-01
-8.48926723e-01 2.01086450e+00 6.33370686e+00 1.06149244e+00
-1.28841209e+00 5.00856817e-01 2.41084203e-01 7.29092583e-02
-3.52142692e-01 5.05820513e-02 -1.42337382e+00 1.97835803e-01
1.30377519e+00 -2.52308995e-01 8.50328207e-01 8.05590928e-01
8.46857578e-03 3.16978574e-01 -1.29031146e+00 7.52386928e-01
2.86933452e-01 -8.27906787e-01 3.68257225e-01 -1.19574994e-01
1.05302644e+00 7.08726645e-01 5.48597425e-02 7.45695114e-01
6.87930882e-01 -1.08814716e+00 5.89996576e-01 -7.29496852e-02
1.48914874e+00 -7.02029765e-01 6.83665574e-01 3.85261506e-01
-1.15719199e+00 5.29918015e-01 -5.42372286e-01 1.28082439e-01
3.33515674e-01 9.65314209e-02 -1.12216115e+00 5.70778966e-01
5.24773538e-01 4.69618887e-01 -6.25699878e-01 4.66519475e-01
-5.81284046e-01 8.54150474e-01 -6.36983156e-01 1.72071457e-01
3.57360989e-01 -1.91930637e-01 3.56727719e-01 1.74381506e+00
3.59351546e-01 -5.78676343e-01 2.80363262e-01 4.00847703e-01
-3.45473051e-01 7.85905242e-01 -7.38642395e-01 -1.91767529e-01
5.46500146e-01 1.30588138e+00 -1.98656172e-01 -3.21448714e-01
-4.80868697e-01 1.23554552e+00 9.74831164e-01 4.44828838e-01
-4.53102499e-01 -2.74851650e-01 6.48340702e-01 -1.72021575e-02
1.60594970e-01 -6.52005076e-01 1.06612690e-01 -1.47208118e+00
8.23890343e-02 -1.37377775e+00 3.55977803e-01 -2.61897027e-01
-1.32285738e+00 1.05798197e+00 -7.55760372e-02 -9.27781701e-01
-4.12755966e-01 -7.99530804e-01 -2.01423526e-01 1.32779026e+00
-1.90584457e+00 -1.69067168e+00 3.13705236e-01 8.03548574e-01
7.06489801e-01 -5.45879602e-01 1.14283943e+00 6.82404220e-01
-5.28359711e-01 1.13170016e+00 4.16611940e-01 3.59339476e-01
1.21749187e+00 -1.13105047e+00 7.27832735e-01 1.04188466e+00
5.29586315e-01 8.31217766e-01 4.41024572e-01 -5.13951778e-01
-1.42520833e+00 -1.18932211e+00 1.58747447e+00 -7.42685139e-01
1.01702285e+00 -8.64117861e-01 -9.51015651e-01 1.13883364e+00
3.71380031e-01 -1.16940469e-01 6.26645386e-01 4.40061450e-01
-3.51294816e-01 -5.12872916e-03 -7.89393902e-01 7.98914850e-01
9.66596425e-01 -8.33581448e-01 -4.45130318e-01 6.87212408e-01
7.99015760e-01 -5.42308807e-01 -1.02261925e+00 1.91849247e-01
3.11704427e-01 -1.12806328e-01 7.80217946e-01 -7.88156509e-01
3.99592854e-02 -6.84224591e-02 -2.03671619e-01 -1.64155483e+00
-1.27659202e-01 -9.12509203e-01 3.62987339e-01 1.70661879e+00
8.07393074e-01 -8.98642361e-01 2.63176531e-01 1.87554985e-01
-2.08519861e-01 -2.82761574e-01 -1.08277464e+00 -1.22241807e+00
7.82306254e-01 -5.08455575e-01 4.53649044e-01 1.25590420e+00
-6.52949065e-02 6.27408981e-01 -5.58079243e-01 -7.24886172e-03
2.38991618e-01 -3.64864677e-01 8.97899210e-01 -5.15021205e-01
-4.57551241e-01 -1.43061370e-01 2.11748155e-03 -8.39345038e-01
6.23173475e-01 -1.47955561e+00 1.09755524e-01 -1.16559637e+00
-7.87037164e-02 -9.21786368e-01 -2.00626165e-01 7.64059246e-01
-3.57139826e-01 5.21127343e-01 2.69751579e-01 1.81971803e-01
-3.81215513e-01 4.37534928e-01 1.10312581e+00 7.13989884e-02
-2.97807187e-01 -3.40709746e-01 -4.36780006e-01 5.61596870e-01
7.11607933e-01 -5.89623690e-01 -2.22629741e-01 -1.25846100e+00
3.04112256e-01 -4.75336283e-01 -5.21948993e-01 -8.37220132e-01
7.39820078e-02 -2.97657680e-02 -3.01684961e-02 -6.24643127e-03
-3.89087573e-03 -6.43928051e-01 -6.90567642e-02 1.54428437e-01
-5.29925227e-02 4.13213789e-01 6.01367533e-01 -2.48865873e-01
-1.58750623e-01 -2.81579226e-01 7.23050952e-01 -2.78896540e-01
-6.42652035e-01 3.63987386e-01 -3.31953347e-01 6.49673939e-01
5.53354025e-01 1.58635452e-01 -3.30804557e-01 -8.24898630e-02
-3.01676929e-01 2.01439142e-01 5.13793945e-01 6.73867226e-01
-6.47068173e-02 -1.30574083e+00 -1.16101563e+00 3.27424884e-01
2.25837275e-01 -1.29033178e-01 -1.56997576e-01 5.64862192e-01
-6.71458006e-01 7.84491837e-01 -1.20890141e-01 -4.19231206e-01
-1.04822743e+00 4.95550424e-01 3.21421295e-01 -8.60564113e-01
-2.61745393e-01 9.31063890e-01 3.78842428e-02 -1.28631747e+00
-1.85394082e-02 -2.02574044e-01 6.50611743e-02 -1.32644370e-01
2.40449771e-01 2.84952056e-02 3.50936860e-01 -1.00452340e+00
-4.07337397e-01 6.43170297e-01 -4.62229937e-01 -3.43408763e-01
1.13352609e+00 -1.89790756e-01 -1.42872911e-02 4.71751630e-01
1.13465452e+00 4.44298238e-01 -8.39614153e-01 -7.53465295e-01
6.55767545e-02 -3.26672830e-02 -2.66461611e-01 -9.23716128e-01
-7.25367606e-01 8.93369198e-01 1.35106057e-01 -5.41065395e-01
7.72947311e-01 -8.70774537e-02 9.90792036e-01 6.75718307e-01
7.05968618e-01 -1.31289375e+00 -4.10743833e-01 1.20542634e+00
6.67589128e-01 -1.48878455e+00 -1.87679067e-01 -2.08610259e-02
-6.78460598e-01 8.92653942e-01 6.15882874e-01 1.38805866e-01
2.61833310e-01 3.34798455e-01 6.13620043e-01 2.98834920e-01
-5.79367518e-01 -2.14853153e-01 4.46854979e-01 6.31949484e-01
7.88483441e-01 2.77043730e-01 -3.22025865e-01 2.82445312e-01
-6.97149515e-01 -4.13094640e-01 1.41849399e-01 6.11913502e-01
-7.69838691e-02 -1.71696413e+00 -3.25321853e-01 -4.93698604e-02
-6.41043901e-01 -8.99304330e-01 -2.30737865e-01 1.07993805e+00
1.80819854e-01 7.26026356e-01 -1.49205938e-01 -3.71038318e-02
3.65567178e-01 3.02061111e-01 5.06928444e-01 -1.02914655e+00
-9.60261822e-01 -6.37525171e-02 2.73225576e-01 -2.66988695e-01
6.18585572e-02 -7.03585148e-01 -9.15701270e-01 -2.86272019e-01
-2.47655451e-01 7.01998547e-02 8.81078899e-01 1.16746950e+00
1.05537670e-02 1.81111664e-01 2.70782143e-01 -7.04021096e-01
-6.26917064e-01 -1.17581892e+00 -1.84273601e-01 3.11088890e-01
7.38653615e-02 -4.42156531e-02 -3.61104816e-01 1.97373956e-01] | [11.053694725036621, 10.050030708312988] |
7392cc2e-74d6-45ae-ac99-2a33be0a6c6d | parallel-detection-and-segmentation-learning | null | null | http://openaccess.thecvf.com//content/ICCV2021/html/Shen_Parallel_Detection-and-Segmentation_Learning_for_Weakly_Supervised_Instance_Segmentation_ICCV_2021_paper.html | http://openaccess.thecvf.com//content/ICCV2021/papers/Shen_Parallel_Detection-and-Segmentation_Learning_for_Weakly_Supervised_Instance_Segmentation_ICCV_2021_paper.pdf | Parallel Detection-and-Segmentation Learning for Weakly Supervised Instance Segmentation | Weakly supervised instance segmentation (WSIS) with only image-level labels has recently drawn much attention. To date, bottom-up WSIS methods refine discriminative cues from classifiers with sophisticated multi-stage training procedures, which also suffer from inconsistent object boundaries. And top-down WSIS methods are formulated as cascade detection-to-segmentation pipeline, in which the quality of segmentation learning heavily depends on pseudo masks generated from detectors. In this paper, we propose a unified parallel detection-and-segmentation learning (PDSL) framework to learn instance segmentation with only image-level labels, which draws inspiration from both top-down and bottom-up instance segmentation approaches. The detection module is the same as the typical design of any weakly supervised object detection, while the segmentation module leverages self-supervised learning to model class-agnostic foreground extraction, following by self-training to refine class-specific segmentation. We further design instance-activation correlation module to improve the coherence between detection and segmentation branches. Extensive experiments verify that the proposed method outperforms baselines and achieves the state-of-the-art results on PASCAL VOC and MS COCO. | ['Rongrong Ji', 'Feiyue Huang', 'Yongjian Wu', 'Chi Su', 'Baochang Zhang', 'Zhiwei Chen', 'Liujuan Cao', 'Yunhang Shen'] | 2021-01-01 | null | null | null | iccv-2021-1 | ['weakly-supervised-instance-segmentation'] | ['computer-vision'] | [ 7.62046337e-01 2.97800988e-01 -6.32790506e-01 -4.71414298e-01
-1.02573442e+00 -5.65390110e-01 6.03339553e-01 1.29095092e-01
-5.07660031e-01 3.03763717e-01 -4.61591750e-01 -7.70858377e-02
4.70628738e-01 -5.39091468e-01 -9.41183448e-01 -7.93314457e-01
3.79963428e-01 3.39154363e-01 1.20343721e+00 1.39971644e-01
2.55009919e-01 1.44973353e-01 -1.35471988e+00 6.88838780e-01
9.61005747e-01 1.11748433e+00 3.44429940e-01 7.49134004e-01
-3.34553719e-01 7.23769426e-01 -5.50891757e-01 -3.49133253e-01
2.07405925e-01 -3.96742254e-01 -8.38401139e-01 4.04568195e-01
5.85547566e-01 -7.93000236e-02 3.24622840e-02 1.13238287e+00
1.47063687e-01 -4.12854791e-01 5.60843647e-01 -1.04472280e+00
-3.51209223e-01 6.16506398e-01 -1.06795847e+00 2.89729744e-01
-2.65581071e-01 4.66839254e-01 1.00658357e+00 -1.01792967e+00
4.19558942e-01 1.06682265e+00 6.50780976e-01 5.60207009e-01
-1.26146579e+00 -4.13912475e-01 7.88215041e-01 1.28497286e-02
-1.20943904e+00 -1.48429245e-01 7.48630047e-01 -4.85400498e-01
4.80046183e-01 1.31059125e-01 6.12362504e-01 8.25578153e-01
-3.96266341e-01 1.56230760e+00 1.16428173e+00 -4.15748328e-01
2.22413644e-01 3.04328084e-01 5.49815536e-01 8.72786164e-01
3.80473197e-01 -8.48191306e-02 -4.39099461e-01 3.94684941e-01
6.89100802e-01 -3.50181311e-02 -5.39482338e-03 -4.81896758e-01
-1.17552769e+00 5.48357666e-01 5.90759933e-01 2.74179131e-01
-6.78313570e-03 5.58783077e-02 2.94064939e-01 -2.90833324e-01
6.19604111e-01 1.49019361e-01 -7.97046006e-01 3.78665298e-01
-1.56984723e+00 -5.65395206e-02 6.58216238e-01 1.14146173e+00
9.84494865e-01 2.75216973e-03 -6.03389978e-01 7.41726398e-01
2.64756173e-01 2.30128735e-01 1.78561911e-01 -8.15184176e-01
3.18867445e-01 1.02119565e+00 -9.25888717e-02 -3.90933633e-01
-2.85609156e-01 -8.93888235e-01 -4.84577149e-01 1.73748076e-01
6.82493865e-01 -1.63668200e-01 -1.39442861e+00 1.37939298e+00
6.97839141e-01 5.33075333e-01 -2.05595270e-01 9.30036962e-01
8.90644431e-01 5.27912617e-01 2.28859186e-01 -8.12159628e-02
1.37605011e+00 -1.45955026e+00 -3.37498426e-01 -6.40063882e-01
5.44805944e-01 -4.50799257e-01 1.11027408e+00 3.04554909e-01
-1.17493582e+00 -8.77258718e-01 -1.17355537e+00 -8.18381086e-02
-1.97041303e-01 5.87959707e-01 4.21903729e-01 6.51594341e-01
-8.10471296e-01 3.75417262e-01 -1.15825462e+00 -2.26768553e-01
9.67199326e-01 2.24476531e-01 1.20741621e-01 1.35307983e-01
-5.80915570e-01 4.39349502e-01 7.79215336e-01 3.41022670e-01
-1.17147923e+00 -7.97518790e-01 -7.78812349e-01 -1.57953233e-01
1.03645670e+00 -4.04442012e-01 1.22585738e+00 -1.19937396e+00
-1.52993882e+00 1.26340222e+00 -3.36062491e-01 -5.09440184e-01
6.45735323e-01 -3.52013677e-01 7.62899145e-02 2.60975540e-01
3.15438092e-01 9.50758159e-01 1.06762612e+00 -1.70983469e+00
-1.18165445e+00 -3.41756821e-01 -2.07071137e-02 -4.62021567e-02
7.93355554e-02 -6.71420917e-02 -1.22149217e+00 -5.14698148e-01
3.22987407e-01 -7.78092444e-01 -4.19762880e-01 2.63332650e-02
-7.90835261e-01 -3.02109510e-01 1.17370617e+00 -4.56226468e-01
1.16988051e+00 -2.00306797e+00 3.31270285e-02 -4.98221368e-02
1.57706887e-01 5.06122589e-01 -1.27652094e-01 -2.65608519e-01
2.52121091e-01 2.45371982e-02 -6.85927629e-01 -5.43248713e-01
-1.02000386e-01 3.23517650e-01 -1.03934936e-01 4.43423480e-01
9.60923910e-01 1.16851330e+00 -9.50184405e-01 -9.50355649e-01
2.09388137e-01 5.99351339e-02 -6.51166558e-01 3.97048563e-01
-5.98210037e-01 5.43788016e-01 -3.78503114e-01 1.02200401e+00
6.84732080e-01 -4.62376684e-01 3.11628152e-02 -4.02738214e-01
-2.56258488e-01 8.33128765e-02 -1.12370038e+00 1.75843155e+00
5.12795057e-03 3.26606780e-01 3.89727056e-01 -1.37201262e+00
4.28140759e-01 -9.55422968e-02 1.20612964e-01 -4.49578822e-01
-4.05787081e-02 2.25099027e-01 3.09908502e-02 -3.68292153e-01
7.13367015e-02 9.48038399e-02 -1.10858873e-01 -1.29525647e-01
3.80578190e-01 -3.70108217e-01 5.28009951e-01 3.51145983e-01
8.11583102e-01 8.15147638e-01 1.07461520e-01 -2.70547181e-01
7.14092910e-01 2.09358081e-01 1.06765330e+00 1.04838407e+00
-4.29403037e-01 8.04607749e-01 5.34474730e-01 4.63859774e-02
-6.66468680e-01 -1.01340663e+00 -3.32651198e-01 1.36668932e+00
5.74377656e-01 -2.70223528e-01 -1.30887973e+00 -1.09428906e+00
-2.18473077e-01 3.28473896e-01 -6.08710170e-01 1.40709922e-01
-6.58688068e-01 -9.75504339e-01 4.96121675e-01 7.72244513e-01
7.49521554e-01 -1.01472664e+00 -5.49098313e-01 2.17937365e-01
9.39262286e-02 -1.49698126e+00 -4.43516225e-01 5.17473638e-01
-8.34259629e-01 -1.12624323e+00 -7.94069290e-01 -1.00675142e+00
7.16033399e-01 2.07878113e-01 1.13907123e+00 2.64129311e-01
-6.51856840e-01 1.40908182e-01 -1.87143967e-01 -4.06394184e-01
-2.09524170e-01 2.05165282e-01 -4.45811301e-01 2.34678149e-01
1.57607004e-01 -1.08316831e-01 -8.51098180e-01 3.68562847e-01
-8.65493953e-01 2.65733451e-01 9.28600311e-01 7.69115269e-01
9.77219045e-01 -2.69039243e-01 4.78202760e-01 -1.35690033e+00
-2.47707024e-01 -1.80134609e-01 -7.63741970e-01 3.89612317e-01
-4.03281987e-01 -1.13252766e-01 2.69301534e-01 -3.11971098e-01
-1.24298573e+00 4.86086607e-01 -9.04595181e-02 -2.08507881e-01
-5.19620180e-01 4.72074188e-02 -5.29940009e-01 1.99451204e-02
4.81176913e-01 2.49190986e-01 -4.98652101e-01 -4.51163083e-01
6.26034737e-01 5.22744596e-01 8.99232030e-01 -6.47425175e-01
7.66465783e-01 6.85036063e-01 -4.36157703e-01 -7.51587272e-01
-1.66555369e+00 -8.76226723e-01 -1.13819003e+00 -1.16616964e-01
1.22309041e+00 -1.00150740e+00 -2.47478440e-01 7.37926304e-01
-1.00305283e+00 -8.42361093e-01 -4.03435528e-01 -1.09722733e-03
-3.83756280e-01 1.76639140e-01 -8.66778135e-01 -7.91195095e-01
-1.44776985e-01 -1.24768412e+00 1.54921854e+00 5.25850832e-01
1.98262766e-01 -6.70431614e-01 -4.93696481e-01 6.40251577e-01
-1.40544111e-02 2.40815863e-01 4.39589322e-01 -6.26962721e-01
-9.54563797e-01 -8.59792344e-03 -7.41181970e-01 6.42787457e-01
-1.33539349e-01 1.73731819e-01 -1.28829432e+00 -8.78912881e-02
-2.46733785e-01 -3.25478911e-01 1.34322548e+00 4.63545114e-01
1.36329579e+00 -4.33618277e-02 -5.96284151e-01 8.03585410e-01
1.33028996e+00 -1.69112325e-01 2.58453012e-01 1.01336822e-01
1.08148277e+00 6.46884441e-01 7.67476916e-01 8.82128105e-02
3.00815970e-01 4.50912684e-01 2.69264817e-01 -6.18642569e-01
-5.16781330e-01 -1.29846051e-01 3.22033584e-01 4.24225450e-01
2.69979507e-01 1.62402928e-01 -8.47827017e-01 6.91678226e-01
-1.87635183e+00 -6.63225591e-01 -4.27352577e-01 1.77152300e+00
1.10763848e+00 7.10912049e-01 3.85775357e-01 1.32730395e-01
7.95249164e-01 1.63455769e-01 -8.77650976e-01 1.56493813e-01
-8.08177963e-02 2.59306043e-01 5.22645056e-01 4.95587945e-01
-1.46696484e+00 1.53421879e+00 5.23150396e+00 9.55310822e-01
-9.31029081e-01 2.87421227e-01 1.03748643e+00 5.54841980e-02
2.69942313e-01 1.89620957e-01 -1.25697958e+00 4.37339753e-01
3.46647978e-01 5.04122913e-01 -1.99901536e-01 9.08352256e-01
4.14414369e-02 -3.44026715e-01 -1.19678938e+00 6.18317723e-01
-1.16699390e-01 -1.33558118e+00 -1.80852026e-01 -3.81483793e-01
9.53387678e-01 1.58171669e-01 -1.70688584e-01 5.06612301e-01
1.61549285e-01 -4.84059751e-01 9.85972583e-01 1.14667751e-01
6.58371866e-01 -4.58896786e-01 4.79214132e-01 5.52349925e-01
-1.52001286e+00 -1.03514791e-01 -1.29170984e-01 1.94690257e-01
1.41604424e-01 7.53595352e-01 -5.33573627e-01 3.05041462e-01
8.10200214e-01 7.61187136e-01 -7.18583465e-01 1.03672552e+00
-7.02503204e-01 1.26821411e+00 -3.14724922e-01 2.65247405e-01
4.73408759e-01 -1.76025718e-01 2.71490008e-01 1.86875951e+00
-4.31030184e-01 8.84298980e-02 6.98529422e-01 1.19362807e+00
-6.69305399e-02 -2.77916402e-01 1.48101211e-01 1.58346385e-01
2.54148781e-01 1.63350654e+00 -1.34727681e+00 -7.54169822e-01
-4.12222892e-01 1.15613091e+00 2.66261965e-01 4.93318975e-01
-8.88328969e-01 -7.39451572e-02 2.78613240e-01 3.13577831e-01
5.13886273e-01 -7.63948858e-02 -6.19233310e-01 -1.13079405e+00
-2.28909571e-02 -6.22195005e-01 4.55440998e-01 -3.06756556e-01
-1.16116285e+00 1.60821587e-01 -3.22830155e-02 -7.65185773e-01
2.46163175e-01 -7.08381832e-01 -8.53486717e-01 5.16134024e-01
-2.00812125e+00 -1.50126255e+00 -3.73268098e-01 3.26249421e-01
9.57956135e-01 2.97725827e-01 1.27556100e-01 9.01519433e-02
-1.16961539e+00 4.92505491e-01 -3.86356175e-01 4.60870147e-01
4.15332407e-01 -1.56596732e+00 2.85480201e-01 1.26465142e+00
2.15844527e-01 3.71195644e-01 2.15935469e-01 -7.53634274e-01
-1.01448381e+00 -1.46140409e+00 1.82655290e-01 -3.40609521e-01
5.66754341e-01 -6.36363089e-01 -9.41250205e-01 3.93124580e-01
7.14293271e-02 6.17363393e-01 3.87936980e-01 -3.46750543e-02
-2.43906662e-01 -9.29094702e-02 -8.68176401e-01 5.07784963e-01
1.23426604e+00 -3.42216998e-01 -5.21013737e-01 4.48371947e-01
7.85247684e-01 -4.64993864e-01 -2.78718263e-01 6.43463671e-01
1.02658354e-01 -9.07287836e-01 1.06352437e+00 -3.14882755e-01
2.41695821e-01 -8.38270783e-01 2.18802616e-01 -6.43346250e-01
-1.19733267e-01 -5.73954046e-01 -2.64425427e-01 1.48347521e+00
5.49616933e-01 -6.39174804e-02 1.05002141e+00 2.86153466e-01
-3.88061196e-01 -8.82108748e-01 -5.10848761e-01 -6.55572772e-01
-1.20273292e-01 -5.01316249e-01 7.04451948e-02 5.81561089e-01
-3.57124835e-01 4.33871806e-01 1.84712574e-01 5.72093248e-01
9.41372931e-01 3.33530933e-01 8.12951386e-01 -1.09774911e+00
-6.62611127e-01 -4.97453928e-01 -1.39170989e-01 -1.44386935e+00
1.28652588e-01 -9.12518620e-01 5.21326125e-01 -1.61528850e+00
4.35940981e-01 -3.88212740e-01 -3.12149107e-01 5.39459646e-01
-8.86393726e-01 6.00462139e-01 1.99942976e-01 1.62263498e-01
-1.19549000e+00 2.74054199e-01 1.34169638e+00 -2.43735045e-01
-3.09203804e-01 2.59799995e-02 -6.30837798e-01 1.19205761e+00
4.47317243e-01 -4.29454029e-01 -2.70632386e-01 -1.98131412e-01
-3.25322837e-01 -2.47706488e-01 5.21333456e-01 -1.09040570e+00
2.92042851e-01 -6.08354174e-02 4.53041762e-01 -7.84714282e-01
5.51670678e-02 -4.09871191e-01 -7.51950145e-01 4.21283603e-01
-2.88573205e-01 -7.25258946e-01 1.47989482e-01 7.90398836e-01
-1.46584377e-01 -2.51821637e-01 1.21355557e+00 -2.39780575e-01
-1.07368743e+00 3.97070467e-01 -1.01312526e-01 4.60394442e-01
1.13870788e+00 -4.22326922e-01 -4.21894453e-02 3.95229548e-01
-8.73472929e-01 4.55364943e-01 2.62139827e-01 1.95001900e-01
3.31381738e-01 -6.77529156e-01 -7.08092928e-01 1.50984988e-01
2.28847176e-01 8.29875052e-01 -8.25003907e-02 9.41703975e-01
-3.22801948e-01 3.32065821e-02 4.89887208e-01 -1.30883479e+00
-1.04264820e+00 4.49734479e-01 3.14387947e-01 -2.87969917e-01
-6.25822902e-01 1.45870054e+00 7.70643294e-01 -3.43558967e-01
3.95811200e-01 -5.36814690e-01 1.74334273e-02 9.62970257e-02
4.53153640e-01 1.30793646e-01 -7.45753944e-02 -3.55584621e-01
-4.57055300e-01 4.85634476e-01 -3.26808006e-01 3.52520980e-02
1.15152979e+00 -7.81908408e-02 1.16548464e-01 4.94176447e-01
8.85789216e-01 -2.92841613e-01 -1.92746222e+00 -3.36615145e-01
5.44170141e-01 -1.31037608e-01 9.07213315e-02 -9.65663552e-01
-1.19358957e+00 1.11514091e+00 4.68909115e-01 -1.04683645e-01
9.49943602e-01 1.95332095e-01 7.88844287e-01 8.07347000e-02
1.81261942e-01 -1.38939643e+00 4.41731155e-01 3.65971059e-01
2.78281301e-01 -1.57579398e+00 -1.64290354e-01 -8.13006878e-01
-5.71904421e-01 7.90118992e-01 9.88205373e-01 -2.53280163e-01
4.88008350e-01 6.74060524e-01 3.39567326e-02 4.65649851e-02
-4.22849774e-01 -7.41515577e-01 5.24798989e-01 5.00866234e-01
4.83992070e-01 -1.17189847e-01 -1.87891468e-01 8.67341876e-01
5.80525219e-01 -1.45251945e-01 5.62445931e-02 9.07613337e-01
-6.84695423e-01 -1.03852928e+00 -3.33874762e-01 3.77055079e-01
-4.09598678e-01 -1.38882414e-01 -5.22538483e-01 7.34270394e-01
5.59290946e-01 1.04015636e+00 1.00249164e-02 1.33067340e-01
1.85174644e-01 8.42773169e-02 4.62117314e-01 -1.01439822e+00
-6.60450041e-01 5.20394325e-01 -1.57700017e-01 -7.69095600e-01
-6.22664511e-01 -7.17634737e-01 -1.66167068e+00 5.94038844e-01
-6.98364019e-01 -1.26536906e-01 4.40682083e-01 1.11148202e+00
1.58060938e-01 7.95237184e-01 2.66758025e-01 -1.20545816e+00
-2.24481225e-01 -7.82140851e-01 -3.39734465e-01 4.86896634e-01
3.57168734e-01 -4.43904579e-01 -1.40668958e-01 4.01762307e-01] | [9.542442321777344, 0.6353179812431335] |
5de0d5c1-7249-4d1c-bd26-ebd06b37ebb4 | the-user-behind-the-abuse-a-position-on | 2103.17191 | null | https://arxiv.org/abs/2103.17191v2 | https://arxiv.org/pdf/2103.17191v2.pdf | Modeling Users and Online Communities for Abuse Detection: A Position on Ethics and Explainability | Abuse on the Internet is an important societal problem of our time. Millions of Internet users face harassment, racism, personal attacks, and other types of abuse across various platforms. The psychological effects of abuse on individuals can be profound and lasting. Consequently, over the past few years, there has been a substantial research effort towards automated abusive language detection in the field of NLP. In this position paper, we discuss the role that modeling of users and online communities plays in abuse detection. Specifically, we review and analyze the state of the art methods that leverage user or community information to enhance the understanding and detection of abusive language. We then explore the ethical challenges of incorporating user and community information, laying out considerations to guide future research. Finally, we address the topic of explainability in abusive language detection, proposing properties that an explainable method should aim to exhibit. We describe how user and community information can facilitate the realization of these properties and discuss the effective operationalization of explainability in view of the properties. | ['Ekaterina Shutova', 'Helen Yannakoudakis', 'Pushkar Mishra'] | 2021-03-31 | null | https://aclanthology.org/2021.findings-emnlp.287 | https://aclanthology.org/2021.findings-emnlp.287.pdf | findings-emnlp-2021-11 | ['abuse-detection'] | ['natural-language-processing'] | [-4.88041714e-02 1.85930252e-01 -4.90618974e-01 -3.73293281e-01
-2.49365702e-01 -6.99545801e-01 3.07277650e-01 2.79185742e-01
-2.55277812e-01 7.96952009e-01 2.42714614e-01 -3.55464876e-01
-8.37797001e-02 -5.15737593e-01 -4.25777696e-02 2.90124100e-02
-1.86966017e-01 8.20784569e-02 -3.32887799e-01 -2.28572026e-01
5.56951165e-01 6.31424069e-01 -8.88409853e-01 1.55192614e-01
1.02583504e+00 2.68181682e-01 -6.93985283e-01 3.91168088e-01
-1.97478279e-01 7.20160425e-01 -8.86507988e-01 -1.39797950e+00
4.48665693e-02 -4.02945220e-01 -9.13469732e-01 2.72614747e-01
3.78319740e-01 -7.35379875e-01 -3.42693985e-01 1.39040184e+00
3.43713820e-01 -1.01922303e-02 4.02346969e-01 -1.74038517e+00
-8.16236377e-01 9.31320608e-01 -8.82761121e-01 6.45601988e-01
9.70632434e-01 1.19097874e-01 1.04324901e+00 -5.25873244e-01
7.39433169e-01 1.37187898e+00 6.90887153e-01 6.76251471e-01
-1.00727034e+00 -1.03863895e+00 1.45106450e-01 6.12170473e-02
-1.39734697e+00 -3.81487668e-01 9.35713708e-01 -5.64186454e-01
8.19098711e-01 4.84750301e-01 9.35418963e-01 1.63729715e+00
-1.97560489e-01 7.72747636e-01 7.06002474e-01 -3.39876711e-01
-2.10104212e-02 4.67147499e-01 7.04185188e-01 7.57138908e-01
9.58768249e-01 -4.92225438e-01 -6.68073535e-01 -1.22166193e+00
3.47829133e-01 -1.63392752e-01 -2.01843113e-01 -4.82272282e-02
-9.36227366e-02 1.36102700e+00 8.22458118e-02 5.33925653e-01
-9.34382752e-02 -2.08560869e-01 5.50429225e-01 -5.56987850e-03
6.59616768e-01 6.41808927e-01 4.56719920e-02 -4.31486577e-01
-8.62823367e-01 3.39757740e-01 1.17332637e+00 6.52912378e-01
2.72674114e-01 -1.79288641e-01 6.05205297e-01 9.40789759e-01
4.32205051e-01 1.20226860e-01 2.94234961e-01 -8.77839804e-01
3.86429936e-01 6.87092125e-01 -1.39746562e-01 -1.48097849e+00
-1.88899964e-01 -2.64623880e-01 -3.99771184e-01 -3.56730193e-01
1.99297175e-01 -2.32928187e-01 -9.70184132e-02 1.65826142e+00
3.60224515e-01 1.12007543e-01 -5.24813652e-01 6.33815348e-01
3.31996411e-01 7.13512227e-02 4.09627318e-01 -5.60556769e-01
1.20018208e+00 -3.46597552e-01 -1.14665973e+00 -4.36644763e-01
1.00819945e+00 -6.24621451e-01 7.18894899e-01 4.68494415e-01
-1.16794920e+00 4.46082592e-01 -1.03334010e+00 -4.36264947e-02
-3.46452355e-01 -6.36431158e-01 9.08034086e-01 1.68501496e+00
-5.85146487e-01 6.14905417e-01 -4.69737560e-01 -7.55476594e-01
5.80809236e-01 2.33642027e-01 -4.04234618e-01 2.60990579e-02
-1.31393564e+00 1.00093508e+00 -1.44555330e-01 -2.34835491e-01
-2.26288170e-01 -4.31560665e-01 -8.75020146e-01 -2.32718542e-01
3.60157430e-01 -4.42501664e-01 9.09471929e-01 -1.07031178e+00
-7.67516673e-01 1.17785263e+00 -1.90906096e-02 -2.74094641e-01
4.78002578e-01 -4.69575286e-01 -6.56271338e-01 2.01925531e-01
3.42871696e-01 -1.17478423e-01 7.21188545e-01 -9.07633722e-01
8.56294706e-02 -8.91296983e-01 2.64994144e-01 -3.36147472e-02
-1.04036486e+00 7.76417375e-01 6.31652698e-02 -6.13627255e-01
-1.87843323e-01 -7.40576684e-01 -1.09368801e-01 5.29877767e-02
-8.03512573e-01 -1.64496183e-01 9.41680074e-01 -9.36522722e-01
1.70904732e+00 -2.28112507e+00 9.44662839e-02 4.07527119e-01
7.28464723e-01 4.17383999e-01 2.42748097e-01 7.94000268e-01
1.53507898e-02 1.28964555e+00 -3.92186701e-01 -6.07746780e-01
-1.39962852e-01 2.50477821e-01 -1.95159525e-01 8.27094436e-01
-9.48446542e-02 4.75550860e-01 -7.84092665e-01 -5.67319214e-01
-2.31111228e-01 4.02518332e-01 -6.80410802e-01 1.61348045e-01
3.08550179e-01 1.06260575e-01 -6.30012572e-01 7.49497533e-01
7.77274013e-01 2.84461561e-03 4.49081749e-01 4.25838172e-01
7.75650665e-02 3.99524391e-01 -6.91151679e-01 9.92521524e-01
-3.65883578e-03 7.11949468e-01 6.21971428e-01 -3.98595452e-01
6.03766561e-01 1.37574181e-01 3.98410141e-01 -1.86863877e-02
2.62650728e-01 1.84731975e-01 3.27928692e-01 -8.89917672e-01
4.91707742e-01 -2.86138803e-01 -6.10694550e-02 9.61532474e-01
-1.41994774e-01 2.00947061e-01 8.41486007e-02 7.19499052e-01
1.21297979e+00 -5.89640617e-01 8.73193741e-01 1.24713853e-01
3.25828195e-01 -3.60686898e-01 4.58197117e-01 6.84776366e-01
-9.81617808e-01 2.51790881e-02 8.14298034e-01 -2.72811234e-01
-8.60136628e-01 -6.91189885e-01 -1.18444785e-01 1.05831313e+00
2.31935605e-02 -7.73082435e-01 -9.18028712e-01 -9.31400061e-01
2.66449034e-01 9.32439268e-01 -5.12920916e-01 -3.72078568e-01
-3.45825851e-01 -8.69302988e-01 1.23234797e+00 8.58302191e-02
3.89215678e-01 -7.35535264e-01 -1.64834082e-01 -1.41530603e-01
-5.38987696e-01 -1.32060444e+00 -1.57727644e-01 -7.82481551e-01
-7.67161429e-01 -1.14481425e+00 7.86993746e-03 -6.68097362e-02
3.64240855e-01 3.81729335e-01 7.39846766e-01 9.26555514e-01
-5.93399584e-01 5.27166784e-01 -4.69520539e-01 -1.94677413e-01
-5.05936682e-01 9.76990312e-02 5.11533856e-01 8.98727104e-02
8.42523158e-01 -9.81103241e-01 -1.32336721e-01 -1.07535429e-01
-8.83741677e-01 -5.10236800e-01 -1.05009250e-01 1.70977086e-01
-5.96129417e-01 -2.57030368e-01 4.70459521e-01 -1.23734641e+00
1.24650276e+00 -1.06103659e+00 2.31920317e-01 -1.22128598e-01
-4.38048720e-01 -6.08284891e-01 8.69322717e-02 -4.83388871e-01
-7.92804718e-01 -6.35385633e-01 -3.11456978e-01 -6.52791336e-02
-1.64092720e-01 4.19241875e-01 -1.81909457e-01 -3.89804900e-01
8.70254099e-01 -2.60566115e-01 6.51670769e-02 -5.63614488e-01
2.43813336e-01 9.94641662e-01 -3.83479118e-01 -3.71902913e-01
7.20431566e-01 5.59090137e-01 -4.11108077e-01 -1.43559206e+00
-9.91135299e-01 -6.15636230e-01 -5.97438753e-01 -3.26944649e-01
5.97667456e-01 -4.34095502e-01 -6.48436546e-01 2.80469447e-01
-1.46618974e+00 3.19119722e-01 3.67769569e-01 1.34298235e-01
-7.49615878e-02 1.15440440e+00 -1.17054284e+00 -1.43780053e+00
-2.64491916e-01 -7.26826549e-01 4.19824392e-01 -8.70875940e-02
-1.13278186e+00 -1.06037092e+00 1.35705203e-01 1.03685367e+00
1.87049538e-01 3.81467611e-01 9.73704994e-01 -1.31622052e+00
-6.38135225e-02 -3.21745664e-01 -1.16503112e-01 1.43322721e-01
-1.29708042e-02 3.23227823e-01 -7.70580947e-01 -2.50829637e-01
2.49413073e-01 -4.36897546e-01 1.75222501e-01 -1.83566242e-01
8.06534469e-01 -8.19868505e-01 -4.58236396e-01 7.05338866e-02
1.02739060e+00 -2.22766727e-01 5.06537437e-01 2.22072631e-01
6.50109828e-01 8.32300484e-01 1.53363228e-01 8.68967175e-01
-3.78816985e-02 3.77359986e-01 3.68409902e-01 4.92621571e-01
3.80519360e-01 -3.47002149e-01 4.32471067e-01 5.22078633e-01
5.34542874e-02 -1.24906570e-01 -9.38206255e-01 3.86940658e-01
-1.56237674e+00 -1.04154003e+00 -4.78708625e-01 1.88676071e+00
6.18902624e-01 1.05159149e-01 5.20510435e-01 2.41560712e-01
9.94052470e-01 1.99600697e-01 -3.50520074e-01 -9.02326584e-01
1.28507763e-01 -1.18505292e-01 5.06269000e-02 8.72857630e-01
-7.45265365e-01 8.84125352e-01 6.98167181e+00 4.15744424e-01
-6.86025321e-01 3.56685579e-01 4.29447025e-01 -3.26158017e-01
-2.73043752e-01 -1.26013923e-02 -6.00477397e-01 4.65744525e-01
8.69552493e-01 -3.95480841e-01 5.65057218e-01 9.66886103e-01
1.65287092e-01 1.00386821e-01 -9.26346242e-01 7.79241443e-01
5.39735377e-01 -6.73104823e-01 1.73320528e-02 6.02425754e-01
2.57512063e-01 -2.56761104e-01 -1.32051972e-03 2.25451607e-02
2.79508363e-02 -1.02456594e+00 2.71580637e-01 -9.11852941e-02
2.40723372e-01 -8.34536612e-01 4.74010408e-01 7.40801215e-01
-3.01597178e-01 -2.38087758e-01 -2.18346491e-01 -6.61038637e-01
4.38529909e-01 7.74355233e-01 -7.72562921e-01 1.02174871e-01
4.57691640e-01 7.23318815e-01 -4.14152086e-01 8.44605148e-01
-4.38818932e-01 7.90316284e-01 -4.41905856e-01 -1.46293730e-01
3.04231793e-03 -4.28190559e-01 9.80284750e-01 1.24595571e+00
4.50688191e-02 4.39772189e-01 2.58249007e-02 9.19111371e-01
-3.51011336e-01 2.65824676e-01 -1.25589490e+00 -6.55037940e-01
7.22412109e-01 1.28741860e+00 -4.05990899e-01 -4.94522303e-02
-4.02454674e-01 1.06896698e+00 7.23170519e-01 8.89656842e-02
-5.58482409e-01 -5.44966385e-02 1.11739302e+00 4.87933427e-01
-4.36533630e-01 -3.37415904e-01 -4.25995946e-01 -1.38688278e+00
4.39195102e-03 -1.13681591e+00 6.39619112e-01 -5.08116722e-01
-1.69514346e+00 2.23787919e-01 -4.09199856e-03 -4.93374765e-01
-3.67061257e-01 -4.56714749e-01 -5.21212935e-01 4.94455606e-01
-6.97944105e-01 -9.78675127e-01 2.03966469e-01 2.46604979e-01
2.48338729e-01 2.01653749e-01 8.40149999e-01 3.64645898e-01
-9.92419600e-01 6.26644433e-01 -6.59163058e-01 4.42566812e-01
5.10226607e-01 -6.18857622e-01 1.96243688e-01 9.29157317e-01
-2.35900786e-02 1.27506042e+00 8.54107797e-01 -1.14390564e+00
-1.12023902e+00 -5.33935547e-01 1.15043080e+00 -7.70337760e-01
1.18119669e+00 -7.37490058e-01 -1.03679669e+00 1.17083514e+00
2.30715498e-01 -3.55751395e-01 1.57368767e+00 6.36592746e-01
-8.32220614e-01 6.21849060e-01 -1.62878597e+00 7.60854423e-01
1.46915245e+00 -8.01647186e-01 -5.58130503e-01 5.53480983e-01
4.38835800e-01 1.72348917e-01 -7.07513392e-01 -2.04077139e-01
4.86158729e-01 -1.22479427e+00 6.58926904e-01 -1.20476890e+00
4.09816802e-01 6.42498434e-01 1.78462878e-01 -9.09706533e-01
-3.32369804e-01 -1.00716078e+00 -2.19348699e-01 1.45599127e+00
2.22992241e-01 -6.76913202e-01 6.97360694e-01 1.43306673e+00
5.17276227e-01 -4.47618127e-01 -8.70274603e-01 -5.17874181e-01
2.84995764e-01 -5.63413084e-01 1.59436285e-01 1.76627326e+00
1.00140584e+00 3.59698802e-01 -5.90183675e-01 2.21813470e-01
8.05631757e-01 -3.68420482e-01 6.08846068e-01 -1.28789663e+00
-1.27594620e-01 -4.29727763e-01 -4.89177555e-01 -4.43182528e-01
4.65434223e-01 -6.53863490e-01 -9.27152693e-01 -9.24261808e-01
7.23324239e-01 1.88863315e-02 3.76844913e-01 2.73630887e-01
-9.96323898e-02 3.50256473e-01 3.81441593e-01 2.08194330e-01
-5.59872091e-01 1.90927595e-01 7.25029588e-01 1.93144262e-01
-7.22544640e-02 -1.47857130e-01 -1.17614055e+00 1.09666455e+00
8.67576599e-01 -7.28964210e-01 -7.52468780e-02 -8.79000723e-02
7.26692080e-01 -1.03079014e-01 1.30737945e-01 -3.94388169e-01
-1.36534467e-01 -3.24701607e-01 -1.29875345e-02 1.59782380e-01
6.75725341e-01 -6.56713367e-01 -2.81821460e-01 5.40090799e-01
-3.02813560e-01 5.57342768e-02 -3.27849351e-02 5.53576529e-01
2.57857502e-01 -7.55974412e-01 6.58349872e-01 -2.58442074e-01
8.58164951e-02 2.56541580e-01 -8.16941381e-01 2.21274376e-01
1.14247382e+00 -2.99389269e-02 -2.42350519e-01 -8.56789470e-01
-6.94226563e-01 -2.05051191e-02 4.46797490e-01 3.66614014e-01
5.96577406e-01 -9.74470854e-01 -5.93544483e-01 -1.20831635e-02
-2.49145515e-02 -1.15931153e+00 9.03259590e-03 7.31504142e-01
-1.84303969e-01 -4.33732904e-02 -1.16688102e-01 1.39375791e-01
-1.73897159e+00 5.32568812e-01 2.10398525e-01 -8.53515714e-02
-4.88502383e-02 6.80132329e-01 -3.11853856e-01 -2.93333352e-01
-9.07037631e-02 6.00310147e-01 -2.45734721e-01 3.50926146e-02
7.82109857e-01 7.60219276e-01 -3.40161651e-01 -8.33265662e-01
-5.25390506e-01 -5.68863787e-02 -5.11321783e-01 -2.08454102e-01
1.11646867e+00 -2.38320351e-01 -7.65870333e-01 2.01661766e-01
1.23368824e+00 6.22730851e-01 -1.78245142e-01 1.65170699e-01
1.69102833e-01 -8.86327505e-01 -4.25087273e-01 -5.58403015e-01
-6.78629637e-01 7.87081480e-01 -2.28322640e-01 9.63353992e-01
4.34999764e-01 5.90956882e-02 8.25840890e-01 1.24159299e-01
4.79485780e-01 -8.17881942e-01 1.57526046e-01 2.69453406e-01
8.01472306e-01 -9.70029294e-01 3.87157738e-01 -1.40972245e+00
-5.01433730e-01 8.41544688e-01 7.05973983e-01 -8.02332610e-02
7.02614129e-01 1.39979407e-01 -2.00081989e-01 -3.43815953e-01
-1.85383245e-01 2.91380763e-01 -2.11447805e-01 6.48889422e-01
6.29084647e-01 8.08819234e-02 -9.34846282e-01 9.22476411e-01
-3.85604799e-01 -3.16483527e-01 1.03207612e+00 9.36978400e-01
-3.96376312e-01 -1.31697023e+00 -4.51135844e-01 4.18269783e-01
-9.77348685e-01 -1.72740668e-01 -1.57386279e+00 9.49657202e-01
-1.30069464e-01 1.18620729e+00 -2.85270244e-01 -4.12066460e-01
-8.03817809e-02 2.89832503e-01 2.80130357e-01 -7.89637208e-01
-9.24657702e-01 -5.01452446e-01 6.63485885e-01 -6.54210985e-01
-1.29446715e-01 -8.63388181e-01 -8.79115403e-01 -1.01136005e+00
-4.42904651e-01 1.39614433e-01 4.19100881e-01 1.23769367e+00
4.41430241e-01 -4.62252319e-01 3.88045013e-01 -2.49282762e-01
-6.13051534e-01 -1.01550055e+00 -7.72963226e-01 7.18088984e-01
1.50726423e-01 -5.19467592e-01 -7.96077192e-01 -4.65760857e-01] | [8.638358116149902, 10.415502548217773] |
c604f661-5b21-4be5-8a9e-26b5d2d9b0b4 | differentially-private-algorithms-for-the | 2302.12909 | null | https://arxiv.org/abs/2302.12909v2 | https://arxiv.org/pdf/2302.12909v2.pdf | Differentially Private Algorithms for the Stochastic Saddle Point Problem with Optimal Rates for the Strong Gap | We show that convex-concave Lipschitz stochastic saddle point problems (also known as stochastic minimax optimization) can be solved under the constraint of $(\epsilon,\delta)$-differential privacy with \emph{strong (primal-dual) gap} rate of $\tilde O\big(\frac{1}{\sqrt{n}} + \frac{\sqrt{d}}{n\epsilon}\big)$, where $n$ is the dataset size and $d$ is the dimension of the problem. This rate is nearly optimal, based on existing lower bounds in differentially private stochastic optimization. Specifically, we prove a tight upper bound on the strong gap via novel implementation and analysis of the recursive regularization technique repurposed for saddle point problems. We show that this rate can be attained with $O\big(\min\big\{\frac{n^2\epsilon^{1.5}}{\sqrt{d}}, n^{3/2}\big\}\big)$ gradient complexity, and $\tilde{O}(n)$ gradient complexity if the loss function is smooth. As a byproduct of our method, we develop a general algorithm that, given a black-box access to a subroutine satisfying a certain $\alpha$ primal-dual accuracy guarantee with respect to the empirical objective, gives a solution to the stochastic saddle point problem with a strong gap of $\tilde{O}(\alpha+\frac{1}{\sqrt{n}})$. We show that this $\alpha$-accuracy condition is satisfied by standard algorithms for the empirical saddle point problem such as the proximal point method and the stochastic gradient descent ascent algorithm. Further, we show that even for simple problems it is possible for an algorithm to have zero weak gap and suffer from $\Omega(1)$ strong gap. We also show that there exists a fundamental tradeoff between stability and accuracy. Specifically, we show that any $\Delta$-stable algorithm has empirical gap $\Omega\big(\frac{1}{\Delta n}\big)$, and that this bound is tight. This result also holds also more specifically for empirical risk minimization problems and may be of independent interest. | ['Michael Menart', 'Cristóbal Guzmán', 'Raef Bassily'] | 2023-02-24 | null | null | null | null | ['stochastic-optimization'] | ['methodology'] | [ 7.16555789e-02 3.63510668e-01 6.96850894e-03 -1.51397049e-01
-1.42242396e+00 -7.26487279e-01 -3.98931503e-01 3.37111324e-01
-8.91681671e-01 8.40649307e-01 -4.25244391e-01 -6.42770946e-01
-4.71814871e-01 -7.86592603e-01 -8.01964164e-01 -1.05164385e+00
-5.27699351e-01 1.52731776e-01 -1.75553858e-01 -2.60881066e-01
1.30623221e-01 2.82736331e-01 -1.25042605e+00 -3.72366250e-01
1.07337153e+00 1.47991168e+00 -3.25021416e-01 6.95162296e-01
1.15231663e-01 2.25377634e-01 -3.70921880e-01 -6.56141400e-01
7.32941747e-01 -7.30015993e-01 -6.66606665e-01 -1.69054672e-01
3.02871704e-01 -9.09858793e-02 2.81068888e-02 1.50919437e+00
5.53823233e-01 3.83431315e-01 5.65139771e-01 -1.05440199e+00
-1.57816604e-01 3.49608570e-01 -1.00278556e+00 2.20771447e-01
-4.38677482e-02 7.71571398e-02 9.56077516e-01 -6.58832312e-01
2.99135447e-01 6.95767760e-01 1.08637071e+00 4.78941858e-01
-1.21269536e+00 -8.82233143e-01 1.26983002e-01 -9.59410846e-01
-1.54801691e+00 -5.03775813e-02 3.64392698e-01 -4.98795271e-01
3.90488356e-01 6.10613763e-01 2.70288527e-01 2.43412748e-01
1.49759958e-02 3.64988893e-01 9.97149706e-01 -3.70428413e-01
3.95465076e-01 4.21878487e-01 8.80398974e-02 1.12749541e+00
2.56390214e-01 -6.47693966e-03 -1.17804922e-01 -4.86161917e-01
7.01601863e-01 -4.31307256e-02 -3.58679444e-01 -2.77021378e-02
-5.77490211e-01 1.26289284e+00 1.38092637e-01 2.58643538e-01
2.27424745e-02 4.63552743e-01 1.53790981e-01 3.42344195e-01
6.85957193e-01 2.19203562e-01 -5.48138201e-01 -2.07177967e-01
-9.41548824e-01 3.98889869e-01 9.27469909e-01 1.02622139e+00
7.22584367e-01 1.31797194e-01 -1.02715045e-02 6.87727928e-01
1.08883709e-01 8.56305063e-01 -1.96916237e-03 -1.49528706e+00
7.81302214e-01 2.76142985e-01 2.98871607e-01 -9.52836752e-01
-3.23720574e-01 -5.41988552e-01 -9.22708809e-01 4.05936033e-01
9.79016185e-01 -5.99430740e-01 -2.20615372e-01 2.17983747e+00
2.44249612e-01 -3.25166970e-01 -2.32104674e-01 6.39308870e-01
1.58977751e-02 5.60303092e-01 -1.37726322e-01 -8.11582923e-01
1.23443270e+00 -1.98397949e-01 -9.34893489e-02 -3.49420980e-02
1.02028596e+00 -6.15947187e-01 1.37827718e+00 3.06137115e-01
-1.53372347e+00 4.25683036e-02 -8.82219017e-01 1.05320975e-01
5.43884896e-02 -2.83335894e-01 6.01676524e-01 1.24191618e+00
-9.06681955e-01 7.78567791e-01 -6.82082474e-01 2.65790701e-01
5.44556677e-01 7.23109007e-01 -2.19941989e-01 2.43824333e-01
-6.59880459e-01 1.15142301e-01 -1.19965687e-01 -1.01131229e-02
-4.71311927e-01 -1.13346767e+00 -6.66266859e-01 6.67989627e-02
4.50601369e-01 -3.45014334e-01 7.56829262e-01 -6.86010957e-01
-1.07124150e+00 1.03725350e+00 -4.42521498e-02 -4.45298523e-01
8.84249210e-01 2.81114966e-01 1.71934351e-01 8.38795304e-02
2.42546335e-01 4.80982028e-02 5.07730126e-01 -9.64808047e-01
-7.43343174e-01 -1.07425857e+00 1.73506096e-01 -6.51851371e-02
-5.05553961e-01 1.82140857e-01 -1.06400497e-01 -6.17412806e-01
1.34136423e-01 -1.01969028e+00 -5.07178843e-01 1.52523398e-01
-1.07254408e-01 1.52244136e-01 3.28950375e-01 -5.61017275e-01
1.39462042e+00 -2.28797841e+00 -1.77585870e-01 7.54748583e-01
3.56079042e-01 8.16737711e-02 5.18011034e-01 -5.04946969e-02
1.20439291e-01 4.27991003e-01 -8.37414205e-01 -5.48297107e-01
2.79361419e-02 -9.25924331e-02 -5.04862666e-02 9.72054660e-01
-5.68663299e-01 3.43442082e-01 -7.27933884e-01 -1.66705295e-01
-3.73335272e-01 3.94144356e-01 -8.82534027e-01 -3.07191610e-01
-3.42913084e-02 2.40662962e-01 -6.12697899e-01 2.86599785e-01
1.06914580e+00 -3.55159223e-01 8.86459947e-02 3.40984374e-01
-1.64833158e-01 -3.06629568e-01 -1.72786415e+00 1.45905364e+00
-5.14392376e-01 1.52524397e-01 7.66474068e-01 -1.30242312e+00
6.56733811e-01 3.77682559e-02 8.57790470e-01 -3.78794700e-01
2.83553153e-01 5.34422755e-01 -5.60686350e-01 -1.20452881e-01
3.52715366e-02 -9.39666390e-01 -2.09951624e-01 5.85460663e-01
-3.39672416e-01 -4.68086712e-02 -2.48384178e-02 -4.05482911e-02
1.11930501e+00 -3.04746479e-01 -7.82671273e-02 -7.46001065e-01
4.66850013e-01 -2.69704401e-01 8.89855325e-01 7.84240365e-01
-2.80186296e-01 5.84670722e-01 1.32024550e+00 -1.63514227e-01
-9.70287323e-01 -9.92479026e-01 -3.80123079e-01 1.13555825e+00
6.04514517e-02 -2.57275432e-01 -1.07410693e+00 -6.46218359e-01
6.29773960e-02 5.68545461e-01 -7.62276411e-01 5.01572229e-02
-5.80862224e-01 -1.35755718e+00 7.17695832e-01 4.99631494e-01
5.49661040e-01 -4.62426573e-01 -6.40441239e-01 -1.83884259e-02
1.75926566e-01 -6.16498709e-01 -9.26993251e-01 4.03629094e-01
-9.81932819e-01 -7.15576589e-01 -9.35463727e-01 -4.96391147e-01
9.49884951e-01 -6.85006902e-02 7.86037385e-01 -7.53931999e-02
-1.76243603e-01 3.55704695e-01 1.60310850e-01 -4.34082001e-01
-4.56949882e-02 -1.71713978e-01 9.94565189e-02 1.35481237e-02
-7.27426587e-03 -7.58011639e-01 -9.64359403e-01 2.50729859e-01
-9.35359061e-01 -6.08838737e-01 3.21751013e-02 8.32789779e-01
1.10433102e+00 8.87956917e-02 2.20916033e-01 -1.04046524e+00
6.04707062e-01 -3.58125806e-01 -1.39563692e+00 -1.03005342e-01
-1.01429212e+00 2.13263556e-01 8.87858629e-01 -1.85542077e-01
-5.91229916e-01 2.16228645e-02 -3.07602227e-01 -4.11621898e-01
4.67579246e-01 3.67230713e-01 4.88504432e-02 -3.35757583e-01
8.40418696e-01 2.60697454e-01 2.45601013e-01 -6.20769560e-01
4.48893011e-02 4.04030055e-01 2.88968951e-01 -8.20738018e-01
5.23561716e-01 8.83513868e-01 4.41298604e-01 -6.46134555e-01
-9.14990246e-01 -1.13182895e-01 5.01066931e-02 3.03303659e-01
6.86449707e-01 -6.56072974e-01 -1.56080818e+00 -1.35179549e-01
-4.40558136e-01 -3.71377975e-01 -8.35474730e-01 4.35125589e-01
-8.70766044e-01 3.87162566e-01 -2.59995371e-01 -1.38100135e+00
-3.85693967e-01 -1.15586615e+00 6.69932604e-01 1.08898796e-01
1.45621002e-01 -1.07651830e+00 -2.84885377e-01 5.11069775e-01
3.33134085e-01 5.48146188e-01 8.13814402e-01 -4.08346683e-01
-2.18903303e-01 -3.42997372e-01 -2.59145111e-01 8.05895507e-01
-3.06939453e-01 -4.32714790e-01 -5.33480644e-01 -4.78176206e-01
6.39547765e-01 3.32827270e-02 7.12002814e-01 5.72683036e-01
1.49238837e+00 -9.31854725e-01 -9.35868472e-02 1.06239986e+00
1.69502079e+00 6.98910728e-02 3.91217083e-01 -5.92565536e-03
2.21541867e-01 2.28965759e-01 2.57250369e-01 8.57694507e-01
1.26036525e-01 3.74828368e-01 3.07002664e-01 -3.94837670e-02
4.73760247e-01 3.35000716e-02 3.40683311e-01 2.36712962e-01
-1.08761631e-01 1.99149325e-01 -7.48552740e-01 6.16843700e-01
-1.46462834e+00 -7.98032343e-01 -1.14700124e-01 2.77115512e+00
1.03402615e+00 2.60206223e-01 3.65586966e-01 1.26729369e-01
5.12442470e-01 -4.69372086e-02 -4.72461969e-01 -8.19738626e-01
-2.12099984e-01 8.70935798e-01 1.07166231e+00 8.15631747e-01
-7.89026737e-01 4.13413703e-01 4.29632807e+00 1.27114964e+00
-7.81452060e-01 3.49814415e-01 1.05407012e+00 -4.33123410e-01
-4.38773394e-01 1.05365507e-01 -8.41247499e-01 8.81773710e-01
9.13085580e-01 -3.43407542e-01 4.20386761e-01 1.14184296e+00
1.19314730e-01 -3.54060680e-01 -8.62813413e-01 9.88411307e-01
-2.14644477e-01 -1.40130985e+00 -6.64306879e-01 6.81610107e-01
9.17459428e-01 -4.24499452e-01 5.91187537e-01 -2.91147847e-02
4.60599661e-01 -1.08548629e+00 3.86207104e-01 -1.97566316e-01
1.17602885e+00 -1.21385694e+00 4.78561014e-01 4.59171504e-01
-1.17018509e+00 -3.25264126e-01 -2.33857885e-01 1.31424576e-01
7.25309402e-02 9.36390460e-01 -7.70350546e-02 3.70045811e-01
8.66038382e-01 -1.97992623e-01 3.21483374e-01 6.08677208e-01
2.45677978e-01 3.28532308e-01 -9.79296863e-01 1.44845396e-01
3.90864968e-01 -5.94959676e-01 5.60719013e-01 1.11861169e+00
4.54730809e-01 6.63891315e-01 1.55675203e-01 7.36289322e-01
-4.57587868e-01 3.81338060e-01 -3.96113753e-01 3.37264419e-01
1.42190084e-01 6.74977243e-01 -6.87670171e-01 1.22904591e-01
-2.45195493e-01 7.84383059e-01 1.23460159e-01 3.52345109e-01
-9.30310726e-01 -7.77568877e-01 1.01045835e+00 3.94169450e-01
4.95627940e-01 -2.61422634e-01 -8.71077478e-01 -9.10115063e-01
5.78301370e-01 -5.66857278e-01 8.57442677e-01 1.39430985e-01
-1.07844591e+00 2.96484619e-01 -1.18516631e-01 -8.22388172e-01
5.06840385e-02 -5.32856226e-01 -2.49089301e-01 1.08599281e+00
-9.72632825e-01 -3.84067982e-01 1.98231131e-01 9.26772654e-01
-1.78791970e-01 -4.61069681e-02 7.39703834e-01 3.77507001e-01
-3.40430439e-01 1.26493967e+00 6.87000751e-01 1.33146107e-01
-1.63091291e-02 -1.04511094e+00 -4.18483824e-01 7.87249982e-01
-3.00608158e-01 6.30182922e-01 6.95669889e-01 -2.83398926e-01
-1.43013322e+00 -9.09057498e-01 7.68320143e-01 -1.74954683e-01
5.48630178e-01 -3.77086550e-01 -6.37394369e-01 7.20094800e-01
-3.85748535e-01 4.84668344e-01 8.77532721e-01 -1.18726850e-01
-1.73825383e-01 -4.64511901e-01 -1.90685451e+00 5.42022705e-01
9.77046072e-01 -3.12266409e-01 2.24149615e-01 5.48169136e-01
5.08824646e-01 -5.27446985e-01 -1.21229362e+00 2.83122957e-01
3.58091205e-01 -1.10157430e+00 8.45555246e-01 -5.37414908e-01
9.50328112e-02 4.91770133e-02 -5.23060203e-01 -4.98411864e-01
3.45107585e-01 -1.41870379e+00 -3.98007371e-02 7.91950583e-01
7.56441772e-01 -8.77798140e-01 1.09876585e+00 1.11917198e+00
2.34323189e-01 -1.11393166e+00 -1.57551670e+00 -8.77538502e-01
8.64029467e-01 -5.83268046e-01 3.69930943e-03 7.11417437e-01
3.39411460e-02 -3.38293135e-01 -2.25532725e-01 9.24547538e-02
8.23931396e-01 -2.06439555e-01 3.76131594e-01 -8.64253104e-01
-7.01912284e-01 -4.54904437e-01 -1.53494045e-01 -9.24756885e-01
-1.29899859e-01 -8.00364971e-01 -1.94911137e-01 -7.90223718e-01
1.97861373e-01 -1.13784158e+00 -3.37221354e-01 3.55238378e-01
3.68604995e-02 6.88685998e-02 4.81603220e-02 -1.73680022e-01
-2.42427111e-01 3.45685363e-01 9.33579087e-01 1.64247692e-01
-5.04211783e-01 3.44598234e-01 -1.02974379e+00 8.36181223e-01
5.01396775e-01 -6.30629182e-01 -3.10677588e-01 -4.92778808e-01
6.72066331e-01 5.66538274e-01 1.36071309e-01 -7.19271183e-01
8.20689946e-02 -2.12624505e-01 -8.81887823e-02 -2.65149437e-02
3.28890085e-01 -5.06195366e-01 5.21344393e-02 5.74591756e-01
-2.92392135e-01 6.87573031e-02 6.90146908e-02 4.91309434e-01
3.09218347e-01 -4.13324594e-01 1.29278862e+00 -2.23578006e-01
3.93162936e-01 4.97776628e-01 -2.85970904e-02 5.72615623e-01
1.10724747e+00 -1.24056876e-01 7.24885464e-02 -5.22607625e-01
-6.92393541e-01 2.55092084e-01 4.09538031e-01 -5.68418145e-01
2.27159783e-01 -9.91270602e-01 -7.10633159e-01 1.40961662e-01
-3.08634281e-01 2.31396645e-01 3.57348889e-01 1.20071101e+00
-5.88608623e-01 8.02693795e-03 4.92857665e-01 -3.84607673e-01
-7.96491981e-01 4.57583189e-01 6.21205866e-01 -3.69971573e-01
-4.84287918e-01 1.48235059e+00 8.71006995e-02 -3.69109847e-02
2.95327336e-01 -2.38346651e-01 7.99563885e-01 -1.52538508e-01
4.64486331e-01 8.14201593e-01 5.95656894e-02 -2.97738105e-01
-3.64208728e-01 4.88800049e-01 8.64685401e-02 -5.99008143e-01
1.34553099e+00 -1.52201518e-01 -2.57698685e-01 -1.83379017e-02
1.78076506e+00 3.13855976e-01 -1.44703245e+00 9.69058350e-02
-3.42874885e-01 -4.36026096e-01 -2.24678442e-01 -3.95615518e-01
-1.34171844e+00 8.15819561e-01 8.62938106e-01 1.98597714e-01
1.13549316e+00 -5.46321012e-02 8.01429987e-01 2.93357149e-02
3.99524987e-01 -1.44077110e+00 -3.34197789e-01 3.30372274e-01
4.30422843e-01 -1.10620821e+00 3.56576033e-02 -2.85786837e-01
-4.04854298e-01 7.55677879e-01 1.15318663e-01 -3.14339578e-01
1.12285948e+00 4.09503281e-01 -3.59787732e-01 -6.57247528e-02
-6.47395104e-02 1.24832079e-01 -1.99172333e-01 1.20843656e-01
2.87540913e-01 2.18387675e-02 -8.35651159e-01 1.04015577e+00
-3.18032980e-01 -7.85484090e-02 3.07095826e-01 9.54940736e-01
-3.57010156e-01 -1.02368498e+00 -2.11099148e-01 4.64411974e-01
-1.10430813e+00 -1.39913827e-01 2.40599066e-01 7.15570688e-01
2.48717517e-01 7.89717138e-01 -1.49781570e-01 1.48872107e-01
1.55414298e-01 4.77719679e-02 4.42460537e-01 -1.63440570e-01
-6.84126854e-01 1.06087588e-01 -2.74573058e-01 -5.27613401e-01
2.01248139e-01 -7.32309282e-01 -1.66365683e+00 -7.16580391e-01
3.38764973e-02 4.64384228e-01 6.56447291e-01 7.99648762e-01
2.93791413e-01 -2.42075115e-01 9.00803566e-01 -2.27219313e-01
-8.86350274e-01 -3.61883134e-01 -1.03168011e+00 3.02693218e-01
1.68657914e-01 -2.01413959e-01 -7.74438620e-01 -1.44673288e-01] | [6.394703388214111, 4.522167205810547] |
6e34c00d-b595-4efb-b5b3-0e4ee96aa468 | multi-scale-sparse-convolution-point-cloud | 2205.01550 | null | https://arxiv.org/abs/2205.01550v6 | https://arxiv.org/pdf/2205.01550v6.pdf | Point Cloud Semantic Segmentation using Multi Scale Sparse Convolution Neural Network | In recent years, with the development of computing resources and LiDAR, point cloud semantic segmentation has attracted many researchers. For the sparsity of point clouds, although there is already a way to deal with sparse convolution, multi-scale features are not considered. In this letter, we propose a feature extraction module based on multi-scale sparse convolution and a feature selection module based on channel attention and build a point cloud segmentation network framework based on this. By introducing multi-scale sparse convolution, the network could capture richer feature information based on convolution kernels with different sizes, improving the segmentation result of point cloud segmentation. Experimental results on Stanford large-scale 3-D Indoor Spaces(S3DIS) dataset and outdoor dataset(SemanticKITTI), demonstrate effectiveness and superiority of the proposed mothod. | ['Jie Cao', 'Lei Jiang', 'Yunzheng Su'] | 2022-05-03 | null | null | null | null | ['point-cloud-segmentation'] | ['computer-vision'] | [-8.66712108e-02 -6.83587790e-01 3.10912371e-01 -6.62627041e-01
-1.25752077e-01 -2.22271815e-01 1.88890606e-01 -9.56310779e-02
-4.50246871e-01 3.30438763e-01 -6.91686496e-02 -7.97727704e-02
-2.50271380e-01 -1.34005964e+00 -7.77091622e-01 -4.16887164e-01
-6.80872947e-02 2.25723118e-01 4.89302427e-01 -7.47699663e-03
3.39348406e-01 8.26317310e-01 -1.55514765e+00 9.04921964e-02
1.27059102e+00 1.12657499e+00 6.77225411e-01 1.43596545e-01
-8.05339515e-01 -1.24403045e-01 -4.12985206e-01 2.68174056e-02
5.11332333e-01 2.66582817e-01 -5.71058452e-01 2.36783788e-01
4.22193885e-01 -1.65282682e-01 -3.11311454e-01 1.16954029e+00
4.05517101e-01 3.23346883e-01 1.61326542e-01 -1.11834025e+00
-5.84668696e-01 4.65111166e-01 -6.00415051e-01 2.50007004e-01
-1.33887887e-01 2.33255774e-01 6.81899190e-01 -8.70124161e-01
2.12409630e-01 1.50457788e+00 9.23138201e-01 4.22079004e-02
-7.85230815e-01 -9.26047206e-01 5.41947544e-01 1.99387342e-01
-1.77485406e+00 1.95762932e-01 8.74285281e-01 -2.76197553e-01
8.20350528e-01 1.60937399e-01 8.77316058e-01 5.50150514e-01
-3.85988712e-01 8.03463638e-01 8.24287236e-01 2.10088968e-01
9.72292721e-02 1.28459170e-01 1.52799517e-01 6.48696065e-01
3.31177831e-01 -6.92798523e-03 8.18517432e-02 -2.76519377e-02
1.18980670e+00 7.66488254e-01 -1.41979188e-01 -2.54892826e-01
-1.27140188e+00 8.96647096e-01 1.13017249e+00 4.68096286e-01
-2.97374040e-01 1.54630437e-01 2.21050173e-01 -6.35215789e-02
5.28108418e-01 1.67503431e-01 -5.44527054e-01 1.10033698e-01
-1.01803625e+00 1.47651836e-01 4.67303991e-01 1.32257426e+00
1.18150985e+00 1.13346493e-02 3.57510075e-02 8.07989120e-01
3.98457348e-01 6.98598146e-01 3.86120021e-01 -7.67256021e-01
5.59148550e-01 9.99759912e-01 -1.76543206e-01 -1.10860920e+00
-4.17695552e-01 -5.07635176e-01 -1.09755838e+00 -1.80215389e-02
-2.44315919e-02 -1.23128772e-01 -1.13697183e+00 1.24448061e+00
4.50389564e-01 8.32241774e-01 -1.15009576e-01 1.38042986e+00
1.03132069e+00 8.67566764e-01 4.42025885e-02 3.42749745e-01
1.18112516e+00 -8.61942708e-01 -2.93419778e-01 -6.93327487e-02
2.52689421e-01 -5.43156326e-01 1.00326562e+00 7.27103725e-02
-6.02043748e-01 -8.23017955e-01 -8.73875499e-01 -1.07759744e-01
-4.70543921e-01 1.44411445e-01 1.24391913e+00 4.93987352e-01
-7.32136905e-01 7.35430062e-01 -9.01259899e-01 -5.26327908e-01
1.07725239e+00 5.20955324e-01 -1.31603330e-01 -3.70503575e-01
-8.47148836e-01 2.06476241e-01 4.61207420e-01 3.66811067e-01
-4.95286971e-01 -6.84880435e-01 -7.17696846e-01 3.63416791e-01
3.05956692e-01 -8.72582555e-01 7.60643840e-01 -6.90134704e-01
-1.25004828e+00 2.86002398e-01 -6.48422688e-02 -2.18849570e-01
1.80253670e-01 -3.11976075e-01 -2.67067581e-01 2.00708792e-01
3.30056995e-01 8.99820626e-01 6.60549760e-01 -1.07250226e+00
-8.44157159e-01 -5.32092929e-01 8.45898092e-02 3.21715385e-01
-1.97740242e-01 -1.42359510e-01 -6.26451790e-01 -3.62691909e-01
5.22329211e-01 -7.72177458e-01 -7.90652096e-01 1.65778939e-02
-4.17042345e-01 -3.63551140e-01 1.12562525e+00 -2.69428372e-01
7.96234012e-01 -2.50888014e+00 -1.41229376e-01 4.49457228e-01
1.98422655e-01 1.81864440e-01 -8.97783563e-02 1.80189218e-02
7.19924644e-02 3.66059124e-01 -5.94947934e-01 -1.52356759e-01
-3.07200309e-02 2.15840355e-01 -2.09552035e-01 3.93875986e-01
4.15869504e-01 9.13892925e-01 -6.92784727e-01 -6.00240469e-01
6.31813765e-01 4.75189388e-01 -7.03192115e-01 9.74876061e-02
-1.48922682e-01 3.82968694e-01 -1.08724344e+00 9.99600351e-01
1.29593360e+00 -3.91829401e-01 -7.08412349e-01 -2.28149220e-01
-5.48603892e-01 -2.34875306e-01 -1.31386697e+00 2.21790385e+00
-4.34862942e-01 2.25104362e-01 1.50761187e-01 -9.14657772e-01
1.03709280e+00 -1.33536845e-01 6.71065807e-01 -2.84903497e-01
1.51548743e-01 2.93120980e-01 -9.93046239e-02 -3.93511623e-01
3.78757626e-01 -4.00351100e-02 3.39782089e-02 -1.70721754e-01
-1.43021960e-02 -5.53375840e-01 -2.22223718e-02 1.55516282e-01
9.73601162e-01 3.19097266e-02 -4.14444923e-01 -1.79702610e-01
6.20318353e-01 3.06731701e-01 5.60437202e-01 4.58470494e-01
9.36662182e-02 7.93382406e-01 -2.31551498e-01 -4.29697484e-01
-6.68612242e-01 -9.28348601e-01 -4.62710887e-01 4.64192688e-01
7.39621699e-01 -2.64082313e-01 -6.13095343e-01 -5.15977085e-01
4.17761296e-01 1.23333819e-01 -2.50277162e-01 5.28163537e-02
-6.35486722e-01 -5.78154385e-01 2.95879275e-01 6.32320821e-01
1.07006609e+00 -1.19998801e+00 -4.40882713e-01 2.43519753e-01
1.16356768e-01 -1.24977779e+00 -3.28462631e-01 1.20918989e-01
-1.26172364e+00 -8.34667921e-01 -6.24491036e-01 -8.23227704e-01
7.32896566e-01 8.08428943e-01 7.49595404e-01 2.72462010e-01
-3.65665615e-01 3.12366158e-01 -4.05407965e-01 -4.54374909e-01
6.68373287e-01 2.36050069e-01 -1.11520723e-01 -9.80566815e-02
4.82739329e-01 -9.83113170e-01 -6.14479959e-01 6.37308136e-02
-1.03433681e+00 -5.41510396e-02 8.77561331e-01 4.07200694e-01
8.15908670e-01 3.55683953e-01 1.83952361e-01 -7.28940725e-01
3.66971910e-01 -6.55278146e-01 -7.06656039e-01 -2.21876651e-01
-2.86449999e-01 -4.42227364e-01 6.84511244e-01 -2.10443452e-01
-9.82042551e-01 1.65233001e-01 -3.64755005e-01 -9.66672480e-01
-5.26303351e-01 4.85758394e-01 -4.13474560e-01 -3.63498688e-01
7.16034099e-02 3.80261481e-01 -3.35379958e-01 -7.39461660e-01
4.61252630e-01 6.63418591e-01 2.45639950e-01 -5.85713983e-01
1.15878189e+00 6.72996640e-01 -1.78031310e-01 -9.90468681e-01
-5.74651003e-01 -6.59093738e-01 -6.50562227e-01 2.16470748e-01
1.09018970e+00 -1.19732928e+00 -6.67764604e-01 5.02387643e-01
-1.24061394e+00 1.70558512e-01 -4.66212392e-01 6.30167484e-01
-1.52029857e-01 3.39637965e-01 -2.89688498e-01 -5.12664378e-01
-3.02047521e-01 -1.17673361e+00 1.24506354e+00 7.49026060e-01
5.66495001e-01 -5.35122931e-01 -2.50288755e-01 8.94076899e-02
5.13316751e-01 1.09514765e-01 7.08037078e-01 -4.49847162e-01
-1.37576163e+00 -1.45919949e-01 -7.55401611e-01 3.22662354e-01
5.52495718e-02 -1.31699473e-01 -8.35810721e-01 -7.79003203e-02
-1.00731939e-01 3.55256721e-02 1.00567758e+00 5.21923423e-01
1.86775076e+00 -1.15707312e-02 -6.11831844e-01 1.20737267e+00
1.76152253e+00 -4.07192623e-03 5.05635679e-01 1.17028601e-01
1.16772234e+00 1.96855307e-01 6.01556540e-01 4.33851808e-01
5.41027427e-01 1.99427471e-01 7.04121947e-01 -2.89417088e-01
1.78736597e-01 -1.99973345e-01 -2.25461096e-01 8.35059822e-01
-1.14943735e-01 2.21678451e-01 -7.20889509e-01 6.28925264e-01
-1.71274996e+00 -6.69275284e-01 -3.39874566e-01 1.70712471e+00
2.95109361e-01 5.05937859e-02 -1.74491882e-01 -1.75297275e-01
7.35687912e-01 9.78164077e-02 -6.64632916e-01 6.01710789e-02
-1.66291982e-01 5.63207626e-01 6.06940925e-01 -3.92571837e-02
-1.26031065e+00 1.15934122e+00 5.12532806e+00 1.05324602e+00
-1.34907961e+00 1.21233568e-01 2.28700995e-01 7.19235614e-02
-4.04386580e-01 5.12009449e-02 -8.13414931e-01 7.12735236e-01
3.79125178e-01 2.85404362e-02 3.41443062e-01 1.13201737e+00
7.25315288e-02 9.30368230e-02 -4.63922471e-01 1.15535378e+00
-4.41885680e-01 -1.48658669e+00 -5.64079881e-02 5.55742495e-02
7.08521783e-01 5.52760065e-01 -1.75630257e-01 4.99811620e-01
1.64094314e-01 -9.33448136e-01 4.75246966e-01 4.63354528e-01
6.08556747e-01 -9.04801548e-01 6.83591604e-01 3.91122252e-01
-1.63284111e+00 -1.53684661e-01 -9.46836233e-01 4.61010233e-04
-1.42592122e-03 9.37242925e-01 -3.38645071e-01 6.99368417e-01
1.11223578e+00 1.07556498e+00 -4.66507822e-01 1.65319705e+00
-2.90356763e-02 5.68009079e-01 -7.44386137e-01 -1.21298365e-01
5.41625798e-01 -6.00245774e-01 4.31276351e-01 1.23571086e+00
5.02818704e-01 5.21448433e-01 6.52506232e-01 1.23314989e+00
-7.52865374e-02 2.58075178e-01 -5.38385630e-01 -8.59770849e-02
6.35352731e-01 1.54570103e+00 -9.73766088e-01 -2.93446928e-01
-5.71050644e-01 8.04307401e-01 9.31005925e-02 3.56038362e-01
-6.50136054e-01 -7.45536149e-01 9.75616455e-01 -8.10982212e-02
7.34145463e-01 -5.93066335e-01 -5.32585442e-01 -1.17527163e+00
-2.39530765e-02 -3.08466703e-01 -1.43253982e-01 -5.57440937e-01
-1.32388747e+00 4.48250920e-01 -1.32503465e-01 -1.30738950e+00
6.41571701e-01 -3.01772892e-01 -9.21380639e-01 1.07765162e+00
-1.80366611e+00 -1.23797429e+00 -6.75043225e-01 8.06396663e-01
6.00372434e-01 9.71128941e-02 5.58163345e-01 5.09990036e-01
-5.29342949e-01 -7.64430538e-02 6.68853670e-02 1.10613413e-01
1.13661811e-01 -1.06963706e+00 5.18630207e-01 5.87816954e-01
-5.84670566e-02 9.03856814e-01 3.02585736e-02 -8.27510297e-01
-1.32360005e+00 -1.48985112e+00 2.66703576e-01 4.59602512e-02
3.23027015e-01 -3.21713418e-01 -1.06583941e+00 3.29100490e-01
-1.79179162e-01 4.12312269e-01 4.73167121e-01 1.22014955e-01
-4.72743530e-03 -1.20752610e-01 -1.38504553e+00 2.04565123e-01
1.55977643e+00 -2.69289881e-01 -4.95151460e-01 4.79333699e-01
1.22232127e+00 -5.21151662e-01 -8.46778870e-01 6.36332273e-01
2.77876016e-02 -7.17690647e-01 1.30049717e+00 -3.42049122e-01
2.87880093e-01 -6.57539904e-01 -2.03883186e-01 -1.07492757e+00
-5.99262476e-01 4.24356237e-02 3.11480731e-01 1.16542125e+00
-3.77523452e-02 -4.63814944e-01 9.17195678e-01 1.96215004e-01
-5.31322718e-01 -7.50729978e-01 -9.81674671e-01 -7.40019560e-01
-1.72047123e-01 -5.89584529e-01 1.23285091e+00 8.88500869e-01
-7.38866687e-01 1.40304193e-01 1.49215043e-01 5.83068907e-01
4.55809683e-01 7.35481143e-01 8.10816884e-01 -1.58208334e+00
1.80787921e-01 -4.51315045e-01 -5.89613140e-01 -1.41644609e+00
1.66644007e-01 -1.01991928e+00 -8.29243511e-02 -1.73894644e+00
-9.21946838e-02 -9.97376740e-01 -2.72496343e-01 4.15620625e-01
-2.27881566e-01 8.07595551e-02 2.16785237e-01 2.47451156e-01
-7.48786449e-01 9.05281067e-01 1.37513506e+00 -2.12954462e-01
-1.52441531e-01 1.78303927e-01 -6.68257892e-01 7.44037747e-01
6.02901220e-01 -3.62678975e-01 -2.29288757e-01 -7.11207807e-01
-3.00017297e-01 -2.19567388e-01 5.28083324e-01 -1.49525785e+00
1.74122706e-01 -2.80070573e-01 7.20114112e-01 -1.13304043e+00
4.37491506e-01 -1.19136035e+00 -3.21883112e-02 2.84301758e-01
2.79391706e-01 -2.57290155e-01 4.05823052e-01 6.77557230e-01
-3.19783896e-01 -6.76295608e-02 5.03765643e-01 -5.04258811e-01
-1.01915002e+00 9.55028772e-01 1.29131123e-01 -1.81827545e-01
8.75506222e-01 -5.28021276e-01 1.45202667e-01 3.05931807e-01
-4.83372450e-01 6.21766031e-01 4.65441406e-01 4.52473134e-01
1.05136502e+00 -1.45139706e+00 -4.86257911e-01 4.19179320e-01
-1.23919368e-01 9.20440674e-01 5.29020667e-01 5.44869900e-01
-5.63308418e-01 3.48452300e-01 -2.15911582e-01 -9.36623096e-01
-8.59099209e-01 2.98452556e-01 5.71083054e-02 1.75947025e-01
-9.27271307e-01 1.23759389e+00 3.17032814e-01 -7.76033401e-01
-6.24994151e-02 -7.69376516e-01 -3.92884582e-01 -2.00305104e-01
1.16628505e-01 1.07249334e-01 -6.57648891e-02 -4.36165631e-01
-4.38250870e-01 8.85946929e-01 1.62758484e-01 4.24873888e-01
1.69374931e+00 -9.13695693e-02 -1.20195433e-01 3.04011434e-01
1.04375660e+00 -1.45767322e-02 -1.19813335e+00 -3.91248345e-01
-2.87674904e-01 -9.85352337e-01 2.11351737e-01 -4.50739324e-01
-1.44524384e+00 1.04864407e+00 8.48111391e-01 6.46592751e-02
9.90981698e-01 -8.22918192e-02 1.28244650e+00 5.06852746e-01
7.10379004e-01 -7.91758001e-01 -2.54067421e-01 7.32443750e-01
5.13871908e-01 -1.35542226e+00 -6.31308407e-02 -8.11713874e-01
-2.38883391e-01 9.91443098e-01 8.47473621e-01 -6.50986612e-01
8.33899796e-01 4.52584848e-02 -2.33736113e-01 -3.96122664e-01
3.51284742e-02 -5.36222875e-01 1.52443284e-02 6.09489799e-01
-7.74869546e-02 1.14113554e-01 -2.69064624e-02 9.53259408e-01
-2.74987519e-01 1.44617051e-01 3.68688479e-02 9.18024004e-01
-7.73403287e-01 -8.90447855e-01 -4.79811758e-01 7.42348075e-01
-9.85786878e-03 -3.00819397e-01 8.77430961e-02 6.12208724e-01
5.99259794e-01 6.97545230e-01 2.86897898e-01 -5.35504460e-01
4.49419081e-01 -4.28148627e-01 2.79017508e-01 -9.28778350e-01
-6.41766012e-01 -3.78562249e-02 -5.34700215e-01 -6.46211267e-01
-4.03683841e-01 -6.43796861e-01 -1.73341894e+00 -2.90969729e-01
-6.37986600e-01 2.90967315e-01 9.40062225e-01 9.11605239e-01
4.66238707e-01 7.24758208e-01 7.53453135e-01 -1.06627786e+00
-1.33869842e-01 -8.47172499e-01 -8.52938235e-01 1.95462704e-01
3.77727114e-02 -7.06182778e-01 -1.80076048e-01 -5.07046819e-01] | [7.943401336669922, -3.4353489875793457] |
3cb30562-e939-4aa2-8f8e-7b7caff4ecd2 | aragpt2-pre-trained-transformer-for-arabic | 2012.15520 | null | https://arxiv.org/abs/2012.15520v2 | https://arxiv.org/pdf/2012.15520v2.pdf | AraGPT2: Pre-Trained Transformer for Arabic Language Generation | Recently, pre-trained transformer-based architectures have proven to be very efficient at language modeling and understanding, given that they are trained on a large enough corpus. Applications in language generation for Arabic are still lagging in comparison to other NLP advances primarily due to the lack of advanced Arabic language generation models. In this paper, we develop the first advanced Arabic language generation model, AraGPT2, trained from scratch on a large Arabic corpus of internet text and news articles. Our largest model, AraGPT2-mega, has 1.46 billion parameters, which makes it the largest Arabic language model available. The Mega model was evaluated and showed success on different tasks including synthetic news generation, and zero-shot question answering. For text generation, our best model achieves a perplexity of 29.8 on held-out Wikipedia articles. A study conducted with human evaluators showed the significant success of AraGPT2-mega in generating news articles that are difficult to distinguish from articles written by humans. We thus develop and release an automatic discriminator model with a 98% percent accuracy in detecting model-generated text. The models are also publicly available, hoping to encourage new research directions and applications for Arabic NLP. | ['Hazem Hajj', 'Fady Baly', 'Wissam Antoun'] | 2020-12-31 | null | https://aclanthology.org/2021.wanlp-1.21 | https://aclanthology.org/2021.wanlp-1.21.pdf | eacl-wanlp-2021-4 | ['news-generation'] | ['natural-language-processing'] | [-1.22055992e-01 4.74928737e-01 -4.08067834e-03 1.38661014e-02
-1.28468442e+00 -7.04660356e-01 1.11176252e+00 1.39519885e-01
-1.69155106e-01 1.02916157e+00 4.19862390e-01 -5.44516802e-01
5.96769273e-01 -9.51433539e-01 -6.11626387e-01 -2.30991200e-01
1.00781389e-01 1.23661280e+00 2.82523278e-02 -1.02453232e+00
2.72511452e-01 -6.90121725e-02 -1.15847528e+00 7.43624508e-01
1.45621419e+00 4.70134526e-01 1.46452382e-01 8.12549949e-01
-2.72719115e-01 8.72992754e-01 -1.11928236e+00 -7.84682751e-01
-2.17525542e-01 -8.74273598e-01 -1.00655019e+00 -2.06750199e-01
1.30676240e-01 -4.79970664e-01 -1.13414060e-02 4.55345601e-01
6.26156688e-01 -1.37794092e-01 9.32480752e-01 -1.08652771e+00
-1.11755478e+00 1.34720743e+00 -1.92883343e-01 -1.48872724e-02
6.52579427e-01 -3.83941345e-02 1.21122205e+00 -1.22789562e+00
8.28004181e-01 1.44151282e+00 4.89884645e-01 8.52641881e-01
-1.03920782e+00 -3.28989327e-01 -2.57889569e-01 -8.17175061e-02
-1.08337736e+00 -4.67637390e-01 3.31208438e-01 -9.66088846e-02
1.40844655e+00 7.39062279e-02 3.99224460e-01 1.47051239e+00
4.37263250e-01 1.17548907e+00 9.36168075e-01 -9.26086724e-01
9.60529819e-02 1.15340017e-01 -1.45751804e-01 6.39941454e-01
1.92492664e-01 -4.46766466e-01 -5.27721524e-01 -1.95400491e-01
3.34261596e-01 -9.75926936e-01 -1.35386005e-01 4.33208317e-01
-1.50358880e+00 1.35769939e+00 -3.30910180e-03 4.34182018e-01
-2.45937303e-01 -9.94093195e-02 3.00902814e-01 5.13251603e-01
6.83042109e-01 7.78442502e-01 -2.18442157e-01 -2.40274608e-01
-8.10516119e-01 5.52482486e-01 1.20885849e+00 1.01139998e+00
1.45404220e-01 4.08776999e-01 -1.93203196e-01 9.81191576e-01
3.86837900e-01 8.44396710e-01 1.10449934e+00 -3.98115963e-01
8.28672528e-01 4.96884048e-01 4.08342034e-01 -9.55678105e-01
-2.22464964e-01 3.03390007e-02 -5.86595476e-01 -1.92508072e-01
7.16001868e-01 -4.93171185e-01 -1.09144104e+00 1.37605047e+00
-1.78941637e-01 -7.34696865e-01 5.97818196e-01 4.19996560e-01
8.25000882e-01 1.39525282e+00 -1.19720213e-02 -6.45636534e-03
1.47004008e+00 -1.06625223e+00 -6.41567051e-01 -7.35019207e-01
8.18516791e-01 -1.12234592e+00 1.21208775e+00 4.12820578e-01
-1.27829695e+00 -3.44347000e-01 -1.03759694e+00 -2.78251152e-02
-6.14084423e-01 5.29881001e-01 4.48135674e-01 8.20611238e-01
-1.06008291e+00 1.15068220e-01 -6.90353394e-01 -5.13271511e-01
-6.72251657e-02 -6.13261722e-02 -2.03532562e-01 -1.23246543e-01
-1.66826010e+00 1.30079722e+00 7.02174008e-01 1.10327736e-01
-1.03566885e+00 -1.32143050e-01 -8.69410396e-01 -2.45411471e-01
2.37557311e-02 -5.41901827e-01 1.50054002e+00 -8.97608161e-01
-1.63201177e+00 6.48903370e-01 -2.21727509e-02 -8.14171851e-01
2.41458923e-01 -2.19754338e-01 -6.49751902e-01 1.11671783e-01
2.83213228e-01 8.05921495e-01 8.20586801e-01 -1.07406783e+00
-5.36521792e-01 6.65281489e-02 1.53344646e-02 3.30081582e-01
-3.24885160e-01 1.58432454e-01 -2.94486396e-02 -9.27038014e-01
-1.59877375e-01 -9.18769240e-01 -9.33749452e-02 -8.13102543e-01
-3.52788240e-01 -4.93379235e-01 3.79803866e-01 -1.08766937e+00
1.30380535e+00 -1.26097560e+00 -8.36669654e-02 -1.23965971e-01
-3.56619984e-01 2.95004696e-01 -7.07303286e-01 1.02081990e+00
2.92651772e-01 3.15312237e-01 -9.46664512e-02 -8.83562788e-02
2.55582690e-01 -1.38304094e-02 -8.58085632e-01 -3.34790081e-01
6.28705680e-01 1.11630750e+00 -1.11740875e+00 -3.26736957e-01
-2.68960208e-01 2.71118730e-01 -3.34806442e-01 7.86317512e-02
-9.19060469e-01 -7.79276192e-02 -2.78395325e-01 7.39331543e-01
2.06038058e-01 -1.81468844e-01 1.76360577e-01 3.30948830e-01
1.59936696e-01 7.52846241e-01 -5.74689031e-01 1.49103928e+00
-5.63229680e-01 6.41849220e-01 -4.38967466e-01 -4.27695811e-01
1.11284173e+00 6.60428703e-01 -2.06397370e-01 -6.98620319e-01
1.32479250e-01 7.06495464e-01 1.82379097e-01 -1.16168417e-01
1.05582082e+00 -1.29636988e-01 -4.27790344e-01 1.00774086e+00
3.18877310e-01 -7.11260021e-01 1.03109097e+00 5.61341763e-01
7.08322227e-01 1.19865984e-01 1.85631782e-01 -1.80477098e-01
4.29861009e-01 5.40038228e-01 -3.16090375e-01 8.45172286e-01
3.76207739e-01 5.97883821e-01 4.91046995e-01 -3.18648130e-01
-1.01701593e+00 -9.01977181e-01 2.73684442e-01 1.35334361e+00
-4.53551382e-01 -6.40862048e-01 -1.11354017e+00 -7.36087620e-01
-4.49018896e-01 1.32199395e+00 -3.55873972e-01 -3.89498845e-02
-8.17137063e-01 -1.29639947e+00 1.05594146e+00 2.81412929e-01
2.69714475e-01 -1.45748043e+00 -1.77221373e-01 6.72595799e-01
-8.92183125e-01 -9.41844225e-01 -3.88120621e-01 -3.09584707e-01
-6.83942020e-01 -8.21502566e-01 -1.04612982e+00 -1.13240516e+00
5.43592274e-01 -1.90340996e-01 1.57994235e+00 -1.85799658e-01
8.73477533e-02 1.13593042e-01 -7.46292174e-01 -7.69523859e-01
-1.35614896e+00 5.74695349e-01 -1.55067191e-01 -4.12702888e-01
4.52788025e-01 1.40408173e-01 5.33072017e-02 7.30955526e-02
-1.00003302e+00 1.50593057e-01 6.16996467e-01 1.05997324e+00
2.96930261e-02 -3.48088562e-01 1.10274923e+00 -7.33496308e-01
1.47799051e+00 -5.46116233e-01 -3.30444455e-01 4.96995211e-01
-6.83411241e-01 2.07046285e-01 6.86984837e-01 -4.06835616e-01
-1.22615695e+00 -5.14537930e-01 -3.26671094e-01 6.10343337e-01
1.56318694e-01 1.01271260e+00 2.23033994e-01 5.53231716e-01
1.13546765e+00 5.82394660e-01 1.51997820e-01 -1.28524989e-01
5.19787967e-01 7.88471758e-01 1.55813530e-01 -5.57446718e-01
5.02790093e-01 -1.17113061e-01 -5.68091571e-01 -8.26329410e-01
-7.71763384e-01 2.41555035e-01 -1.47212565e-01 5.64458668e-02
5.18758893e-01 -1.08125603e+00 -3.24543007e-02 8.54598999e-01
-1.34320140e+00 -5.71718931e-01 -2.45643836e-02 1.31455854e-01
-3.73895824e-01 2.61522472e-01 -1.00167441e+00 -6.80096984e-01
-8.96771073e-01 -7.92469859e-01 1.07795537e+00 1.16872499e-02
-4.42772716e-01 -1.13985348e+00 3.12983125e-01 2.87738562e-01
6.60955131e-01 2.19816957e-02 1.07361662e+00 -8.86308253e-01
-1.00652963e-01 -3.27377647e-01 6.57588020e-02 3.31207961e-01
1.99490841e-02 1.12456940e-01 -6.26058042e-01 -3.12205225e-01
-2.94346809e-01 -7.47515619e-01 5.96283317e-01 -3.80334742e-02
2.42258132e-01 -5.61866820e-01 -6.47597685e-02 -2.45900214e-01
8.92648280e-01 1.84406340e-01 7.79349089e-01 4.36032325e-01
2.76508391e-01 5.47163188e-01 7.21254766e-01 2.16976255e-01
6.88717008e-01 4.03543711e-01 1.84492543e-01 3.20080549e-01
-1.32692084e-01 -4.63834405e-01 1.04162455e+00 1.27931380e+00
2.96298345e-03 -1.02527893e+00 -1.24499166e+00 7.45349407e-01
-1.45695007e+00 -8.78006101e-01 -2.27884516e-01 1.88782787e+00
1.24504232e+00 1.60850167e-01 2.93959342e-02 -1.15047060e-01
4.53207195e-01 2.67603826e-02 3.31057273e-02 -7.19155490e-01
-4.38407600e-01 3.76426399e-01 -1.42429858e-01 7.01499403e-01
-9.45982218e-01 1.35825205e+00 6.88129520e+00 8.92479420e-01
-9.88374412e-01 -6.48626909e-02 7.55384147e-01 2.16464996e-01
-4.63829339e-01 -9.12338346e-02 -1.09330463e+00 5.20040572e-01
1.49407291e+00 -5.82017481e-01 2.58259118e-01 7.28344083e-01
6.21657781e-02 -7.83045366e-02 -8.41969788e-01 4.61974800e-01
6.78996980e-01 -1.26741052e+00 8.72058034e-01 -3.11186194e-01
9.37772751e-01 3.36071923e-02 -1.17224820e-01 6.98233545e-01
6.53207123e-01 -1.28414762e+00 8.40531588e-01 3.07379007e-01
6.33237779e-01 -8.41377616e-01 8.44428718e-01 4.67591256e-01
-7.28966594e-01 8.78688470e-02 -4.68829513e-01 -1.30014643e-02
3.98027122e-01 2.48897031e-01 -1.57560408e+00 3.06427389e-01
1.13271296e-01 3.66415441e-01 -7.55941927e-01 3.67800266e-01
-5.04655063e-01 8.66803586e-01 -1.73231244e-01 -5.27576566e-01
5.86617887e-01 -1.17077000e-01 4.49505150e-01 1.38005579e+00
6.98342383e-01 -2.67259479e-01 -6.60694093e-02 6.75784588e-01
-1.51002511e-01 4.02544022e-01 -4.60156113e-01 -6.98142409e-01
4.00804877e-01 1.15160251e+00 -6.51486158e-01 -4.82313722e-01
-1.20653056e-01 1.15488327e+00 1.34539023e-01 1.20718949e-01
-7.03165531e-01 -7.37001121e-01 2.11037584e-02 -2.70267092e-02
1.10272698e-01 -3.24985683e-01 1.23770565e-01 -1.24685025e+00
-9.70026776e-02 -1.63218045e+00 3.33213985e-01 -1.02128911e+00
-1.35083938e+00 1.23847473e+00 -1.35757878e-01 -9.97453749e-01
-1.23594713e+00 -8.86934578e-01 -5.04259229e-01 1.07335865e+00
-1.26790202e+00 -1.56207645e+00 6.01460859e-02 3.60476285e-01
8.74073386e-01 -6.34456038e-01 1.34834611e+00 -7.95438737e-02
-2.65887618e-01 6.46947920e-01 1.96058378e-01 3.63308907e-01
8.51685345e-01 -1.31278753e+00 9.09744442e-01 9.76335585e-01
3.18492591e-01 7.04760373e-01 5.92643857e-01 -7.31468439e-01
-1.26464713e+00 -9.99499559e-01 1.59641802e+00 -7.96202064e-01
9.38107789e-01 -4.77806956e-01 -6.41889393e-01 7.66560674e-01
6.97381079e-01 -8.74367893e-01 5.45212388e-01 -2.75015801e-01
-2.96150088e-01 2.85210729e-01 -9.00885761e-01 8.28028619e-01
2.51152724e-01 -3.54238003e-01 -8.84821236e-01 6.24425232e-01
6.08148277e-01 -4.04826432e-01 -7.10965455e-01 3.53012457e-02
4.52706724e-01 -4.62538600e-01 6.48395300e-01 -6.38158619e-01
8.60774398e-01 -7.76704848e-02 1.53613657e-01 -1.76152575e+00
1.31633312e-01 -9.02653277e-01 -3.02278876e-01 1.17295516e+00
1.06036031e+00 -6.86579943e-01 4.53561902e-01 2.09707677e-01
-2.01329634e-01 -4.11512107e-01 -5.33383906e-01 -6.40204370e-01
5.71087003e-01 -1.26471728e-01 5.26027441e-01 7.90198565e-01
3.18147779e-01 1.01027536e+00 -3.27748507e-01 -3.22589189e-01
1.02326712e-02 5.34025691e-02 6.99799180e-01 -9.46902812e-01
-1.50240093e-01 -4.33027565e-01 2.54671901e-01 -1.08538222e+00
6.96580485e-02 -1.00049698e+00 3.20052743e-01 -1.86128855e+00
-2.04878926e-01 -3.34951192e-01 5.80494165e-01 6.48412287e-01
-2.75473624e-01 5.20713568e-01 7.19542578e-02 3.92891653e-02
-1.37751713e-01 5.61099470e-01 1.10873890e+00 -2.92483300e-01
-1.55174330e-01 -1.37810960e-01 -8.66530120e-01 3.88834238e-01
1.10860193e+00 -3.09618920e-01 -4.26277906e-01 -7.71412611e-01
5.35921276e-01 -4.13524806e-02 -1.85028419e-01 -8.06220174e-01
-1.25587329e-01 -1.37903392e-02 3.02846700e-01 -5.40235460e-01
3.47048163e-01 1.10227369e-01 -2.67162770e-01 5.14528632e-01
-3.18898082e-01 5.64773440e-01 3.66406173e-01 2.95875035e-02
-3.49757820e-01 -5.82804501e-01 3.77438605e-01 -3.50635886e-01
-5.85143983e-01 -8.79228190e-02 -1.03166986e+00 3.90239596e-01
6.09825075e-01 1.13154687e-01 -6.96659565e-01 -8.44246745e-01
-4.28184301e-01 3.81454565e-02 1.50773510e-01 8.16705167e-01
6.03018880e-01 -1.17431545e+00 -1.42168057e+00 8.11935067e-02
1.79307193e-01 -3.96214128e-01 -3.30603242e-01 4.15303439e-01
-9.86208498e-01 8.77426505e-01 -2.34366268e-01 -1.16020046e-01
-8.22944701e-01 7.01197088e-02 1.23871833e-01 -7.34557927e-01
-1.21189184e-01 6.50365293e-01 -4.80979592e-01 -6.66185081e-01
-2.45329157e-01 -1.65711194e-01 -3.56442183e-01 2.68467188e-01
9.30771112e-01 3.45652699e-01 2.91277677e-01 -6.15147948e-01
1.17753029e-01 -1.50374502e-01 -3.58009189e-01 -6.37082875e-01
9.86827970e-01 1.55865818e-01 -4.85178798e-01 3.43541235e-01
4.59749550e-01 2.99921840e-01 -4.07932132e-01 1.89953119e-01
2.05401734e-01 1.55129641e-01 -4.38197970e-01 -1.18094039e+00
-3.79309058e-01 7.34637618e-01 1.83325782e-02 4.80705321e-01
6.98137879e-01 -1.61127985e-01 1.06902921e+00 7.92598844e-01
4.87400621e-01 -1.19541419e+00 3.88365537e-01 1.37502444e+00
1.32048869e+00 -1.19042671e+00 -3.26297581e-01 7.14805797e-02
-9.16584313e-01 1.33524549e+00 6.60134494e-01 -3.69268544e-02
1.25559881e-01 7.51972347e-02 3.67958069e-01 1.32153511e-01
-9.50268924e-01 1.39401749e-01 3.44909221e-01 6.84837461e-01
1.01094663e+00 8.13740194e-02 -5.84471345e-01 5.81230640e-01
-7.22052217e-01 -2.57638961e-01 9.86286283e-01 7.69048095e-01
-5.17602980e-01 -1.48369825e+00 -3.70496482e-01 4.55715239e-01
-4.82079834e-01 -6.44389272e-01 -6.54278457e-01 9.53996360e-01
-4.33129042e-01 1.28931630e+00 1.65975913e-02 -8.29985365e-02
-1.85362641e-02 4.57317352e-01 4.54586565e-01 -8.37598085e-01
-6.27868354e-01 -1.90072119e-01 6.21214986e-01 2.46477518e-02
-4.73576337e-02 -3.99311572e-01 -1.11904001e+00 -2.04436392e-01
-3.94398421e-01 6.50139928e-01 6.35413468e-01 7.47196913e-01
3.09489012e-01 1.98648162e-02 3.09056818e-01 -7.26670146e-01
-8.32308412e-01 -1.63526607e+00 -1.81606293e-01 3.54883224e-02
-1.81120232e-01 7.51108080e-02 -1.20693259e-01 2.09733531e-01] | [11.691597938537598, 9.480245590209961] |
27777c98-7c37-43e2-9f4b-7428fb162155 | ad-nerf-audio-driven-neural-radiance-fields | 2103.11078 | null | https://arxiv.org/abs/2103.11078v3 | https://arxiv.org/pdf/2103.11078v3.pdf | AD-NeRF: Audio Driven Neural Radiance Fields for Talking Head Synthesis | Generating high-fidelity talking head video by fitting with the input audio sequence is a challenging problem that receives considerable attentions recently. In this paper, we address this problem with the aid of neural scene representation networks. Our method is completely different from existing methods that rely on intermediate representations like 2D landmarks or 3D face models to bridge the gap between audio input and video output. Specifically, the feature of input audio signal is directly fed into a conditional implicit function to generate a dynamic neural radiance field, from which a high-fidelity talking-head video corresponding to the audio signal is synthesized using volume rendering. Another advantage of our framework is that not only the head (with hair) region is synthesized as previous methods did, but also the upper body is generated via two individual neural radiance fields. Experimental results demonstrate that our novel framework can (1) produce high-fidelity and natural results, and (2) support free adjustment of audio signals, viewing directions, and background images. Code is available at https://github.com/YudongGuo/AD-NeRF. | ['Yong-Jin Liu', 'Juyong Zhang', 'Hujun Bao', 'Sen Liang', 'Keyu Chen', 'Yudong Guo'] | 2021-03-20 | null | http://openaccess.thecvf.com//content/ICCV2021/html/Guo_AD-NeRF_Audio_Driven_Neural_Radiance_Fields_for_Talking_Head_Synthesis_ICCV_2021_paper.html | http://openaccess.thecvf.com//content/ICCV2021/papers/Guo_AD-NeRF_Audio_Driven_Neural_Radiance_Fields_for_Talking_Head_Synthesis_ICCV_2021_paper.pdf | iccv-2021-1 | ['talking-face-generation'] | ['computer-vision'] | [ 3.40680599e-01 1.33641928e-01 4.05527771e-01 -3.63777310e-01
-7.96282291e-01 -2.75514394e-01 3.07875037e-01 -5.78667462e-01
2.25476816e-01 5.20859063e-01 4.52898651e-01 1.04586974e-01
4.46554780e-01 -7.26326585e-01 -8.56767118e-01 -6.79520071e-01
3.17414939e-01 -1.88082889e-01 1.26631428e-02 -2.04584211e-01
-1.02079049e-01 3.34783733e-01 -1.78315282e+00 3.80383015e-01
7.27003217e-01 1.22671604e+00 1.82126731e-01 7.83358574e-01
-7.88818896e-02 7.75366962e-01 -5.94841421e-01 -1.49387360e-01
2.35307470e-01 -5.82582355e-01 -2.95154274e-01 1.52231127e-01
5.06532490e-01 -4.95161027e-01 -5.42118013e-01 8.41565609e-01
7.75221050e-01 3.33257824e-01 4.73904103e-01 -1.21968067e+00
-6.23319387e-01 2.53257483e-01 -5.64011335e-01 -4.28124756e-01
8.65095615e-01 1.76595643e-01 5.46315968e-01 -1.07881713e+00
4.66429412e-01 1.46256089e+00 5.82676351e-01 8.55527937e-01
-1.11899900e+00 -9.73812878e-01 1.83035046e-01 5.73955402e-02
-1.64513028e+00 -8.11848938e-01 1.14018631e+00 -3.46783400e-01
3.69373143e-01 3.99339765e-01 8.19745362e-01 1.07794261e+00
6.39680326e-02 6.32928789e-01 9.19676125e-01 -4.73943770e-01
6.41499013e-02 1.02865376e-01 -4.58895355e-01 7.68492639e-01
-6.46531463e-01 1.49595886e-01 -6.90419257e-01 -2.33600244e-01
9.41949725e-01 -1.89118460e-01 -9.15740848e-01 -1.34726912e-01
-9.56100047e-01 5.69316447e-01 4.06248212e-01 4.53635976e-02
-2.29417443e-01 3.27399373e-01 1.20572792e-02 -2.26783186e-01
6.62307799e-01 -2.57223874e-01 1.20923825e-01 4.20145169e-02
-9.77638841e-01 2.17630580e-01 5.25895655e-01 7.83639729e-01
6.10999048e-01 4.00061041e-01 -8.45725536e-02 1.04812551e+00
4.94274110e-01 5.79324067e-01 3.41882408e-01 -1.29551327e+00
2.65300184e-01 -5.61761437e-03 -5.69930300e-02 -9.40829456e-01
-2.40810961e-01 -1.68124229e-01 -7.53520787e-01 3.59117389e-01
2.23466918e-01 -3.83017957e-01 -6.41140342e-01 1.86027241e+00
7.71456540e-01 6.46530628e-01 -1.48309126e-01 1.20209670e+00
1.16505754e+00 1.06156921e+00 -3.25187296e-01 -3.29461038e-01
1.22118342e+00 -7.94386804e-01 -1.01051939e+00 6.08088747e-02
2.10714061e-02 -8.23991537e-01 1.06810760e+00 3.73442650e-01
-1.43068910e+00 -7.72622466e-01 -8.30492496e-01 -2.82459468e-01
1.49298683e-01 -1.02061421e-01 3.48904967e-01 5.29914141e-01
-1.33590043e+00 2.11921453e-01 -4.67650503e-01 1.10344090e-01
6.42848313e-02 1.38932720e-01 -1.87302575e-01 -1.24676041e-02
-1.19973326e+00 2.91131228e-01 -1.42137423e-01 3.21930647e-01
-8.18842292e-01 -9.05443192e-01 -1.07948053e+00 -8.34001228e-02
2.51294702e-01 -9.23846424e-01 1.46765828e+00 -1.32434928e+00
-2.11373067e+00 5.60017884e-01 -3.44562978e-01 1.17277041e-01
5.32761633e-01 -1.58865914e-01 -3.16341102e-01 3.21726143e-01
-1.77623853e-01 9.01592791e-01 9.69443262e-01 -1.57701826e+00
-3.15080822e-01 -2.55143285e-01 -8.38807747e-02 4.81773853e-01
1.78215895e-02 5.65379113e-02 -6.13885581e-01 -8.20699751e-01
4.64948975e-02 -7.29202271e-01 -1.43982917e-02 5.89248180e-01
-4.25715208e-01 1.25816941e-01 7.64104187e-01 -8.71124804e-01
9.48899031e-01 -2.24225163e+00 7.44568408e-02 1.20057426e-01
1.71767250e-01 -6.85537532e-02 -1.60082690e-02 7.83635750e-02
-1.69600219e-01 -1.86398104e-01 -3.12486321e-01 -4.10429597e-01
-1.73425481e-01 -1.97989702e-01 -5.09553313e-01 5.76466680e-01
-4.02244106e-02 5.49782395e-01 -7.05249250e-01 -4.67679769e-01
3.26377392e-01 1.36791086e+00 -8.62053633e-01 4.24248934e-01
-1.04783826e-01 8.56550217e-01 -2.95268536e-01 6.66257262e-01
8.21572661e-01 2.70912766e-01 -5.01043461e-02 -2.59739131e-01
-1.46393344e-01 1.63702995e-01 -1.29266953e+00 1.76578259e+00
-6.72430158e-01 6.20190203e-01 5.99840045e-01 -5.25530934e-01
1.05377316e+00 6.28446102e-01 5.10787427e-01 -7.89023757e-01
2.03255907e-01 -8.94956142e-02 -5.31635225e-01 -5.27350068e-01
2.90020466e-01 -2.88075566e-01 2.70921476e-02 2.42877036e-01
-5.72454184e-02 -3.96167964e-01 -4.53096062e-01 -1.12697044e-02
5.35483241e-01 5.05066216e-01 -9.20594931e-02 1.42700791e-01
5.56942225e-01 -6.04594052e-01 7.24650621e-01 1.58340380e-01
7.70790651e-02 1.18403554e+00 2.80420393e-01 -6.84600149e-04
-7.10801959e-01 -1.19287634e+00 -1.72587216e-01 8.86686504e-01
1.74760863e-01 -4.00970697e-01 -1.21346760e+00 -3.44714597e-02
-3.67457539e-01 7.23779976e-01 -4.88798112e-01 -4.59985062e-02
-6.09916270e-01 -6.52094558e-03 3.20233047e-01 3.70316923e-01
3.24787080e-01 -1.06563425e+00 -4.75464135e-01 1.75834134e-01
-7.19746470e-01 -9.56378341e-01 -8.31875920e-01 -2.35463157e-01
-6.25334024e-01 -7.69378662e-01 -1.05819404e+00 -6.02157891e-01
5.55531263e-01 3.20948958e-01 8.30358744e-01 -5.87383620e-02
-4.21779603e-01 6.93964660e-01 -3.16170305e-02 -4.83049482e-01
-3.37725252e-01 -6.66049302e-01 8.26403275e-02 5.49958587e-01
-2.56676942e-01 -7.49313295e-01 -6.47221744e-01 3.39327931e-01
-8.74458253e-01 4.31962758e-01 -2.45060116e-01 5.99176347e-01
7.25268304e-01 -2.02667236e-01 4.07300442e-01 -4.50730562e-01
5.44574559e-01 -4.09385651e-01 -6.47331297e-01 -1.20549828e-01
2.69284159e-01 -3.71484667e-01 6.17928505e-01 -5.66485286e-01
-1.57886016e+00 1.96316034e-01 -4.39796925e-01 -6.90574169e-01
-3.39213878e-01 -2.60664094e-02 -4.66809839e-01 1.39956892e-01
5.37412703e-01 9.84348804e-02 2.48400364e-02 -4.21751410e-01
4.41453159e-01 7.93039680e-01 7.71359444e-01 -5.24102986e-01
6.66710973e-01 6.51211500e-01 -1.57462016e-01 -1.05399227e+00
-4.59813952e-01 -5.99497817e-02 -4.71637219e-01 -8.32770407e-01
8.49463880e-01 -1.09195495e+00 -7.63039529e-01 5.31386614e-01
-1.29689622e+00 -5.70641339e-01 -2.64381796e-01 6.36243582e-01
-6.96073949e-01 2.78252572e-01 -6.59944832e-01 -1.06467724e+00
-2.23583043e-01 -1.21661413e+00 1.20888734e+00 3.63505721e-01
-1.84710994e-01 -7.02305257e-01 8.38499740e-02 4.90216702e-01
2.19259590e-01 4.26536262e-01 7.62897372e-01 4.56181288e-01
-5.72487652e-01 3.98243293e-02 6.30249754e-02 2.02476740e-01
1.60880461e-01 1.88060850e-01 -1.62837231e+00 -7.42331669e-02
2.32230559e-01 -2.86636323e-01 4.39361572e-01 7.72818923e-01
1.39450073e+00 -3.72676045e-01 -8.86504725e-02 1.05056202e+00
1.08792031e+00 3.15543979e-01 8.16115081e-01 -2.52432257e-01
6.83981419e-01 9.92193282e-01 4.29846883e-01 6.27426684e-01
3.87093693e-01 1.01561320e+00 4.99224246e-01 -3.49211693e-01
-5.47130525e-01 -5.38827419e-01 6.07345462e-01 8.54808033e-01
-5.78897186e-02 -1.57099441e-01 -4.44625407e-01 2.14818493e-01
-1.63688672e+00 -9.38717425e-01 -7.75175821e-03 2.13732815e+00
8.15567732e-01 -3.86326224e-01 7.51199275e-02 1.43956661e-01
8.10308397e-01 6.07893132e-02 -5.14923275e-01 -3.54092032e-01
7.30399713e-02 1.44235358e-01 -2.72193491e-01 7.75559306e-01
-6.30371153e-01 5.78744292e-01 5.88368797e+00 7.06899047e-01
-1.40079558e+00 5.37179708e-02 7.20282793e-01 -5.12393892e-01
-5.85467756e-01 -3.06195796e-01 -6.00984573e-01 4.23532307e-01
7.63665020e-01 6.47636428e-02 5.88259637e-01 6.36097252e-01
6.10949457e-01 9.55780894e-02 -9.40092087e-01 1.20336258e+00
3.77208650e-01 -1.02240443e+00 -8.25200826e-02 -7.37316012e-02
3.88519436e-01 -4.93597865e-01 1.56699464e-01 6.79231212e-02
-1.22454770e-01 -1.03392398e+00 1.19599462e+00 7.05251992e-01
1.35950148e+00 -7.77459860e-01 -8.12416226e-02 2.43236616e-01
-1.28561044e+00 2.05472201e-01 -2.40048110e-01 1.92200631e-01
4.24799740e-01 4.22158778e-01 -6.98711753e-01 2.59745032e-01
9.32555974e-01 3.44158471e-01 -3.63167562e-02 1.06408870e+00
-2.81868577e-01 5.57164729e-01 -2.23294273e-01 3.58447999e-01
-1.94082513e-01 -2.15399697e-01 6.48604393e-01 1.12756073e+00
5.64514577e-01 3.63345593e-01 1.90114714e-02 1.03891730e+00
-5.71160205e-02 3.38103503e-01 -8.23414445e-01 5.68032503e-01
1.60638481e-01 1.20596516e+00 -3.49449843e-01 -7.33493194e-02
-3.34211290e-01 9.16566610e-01 -4.14486378e-02 7.09883451e-01
-1.18952930e+00 -4.67467844e-01 7.60381997e-01 3.94937128e-01
1.28795117e-01 -9.35302768e-03 9.61345956e-02 -1.12858546e+00
1.51068136e-01 -6.45780265e-01 -2.51265522e-02 -1.56469691e+00
-6.91146910e-01 7.50387013e-01 -5.70970662e-02 -1.35808480e+00
-3.50163490e-01 -2.71915019e-01 -6.81379974e-01 1.02689457e+00
-1.44624484e+00 -1.03314424e+00 -5.77073216e-01 1.00311553e+00
6.20070696e-01 1.85262114e-01 8.29776883e-01 4.82501268e-01
-3.93695861e-01 6.78706527e-01 -9.39299464e-02 2.81054396e-02
6.70706451e-01 -6.97591186e-01 2.31445767e-02 5.53227425e-01
4.02060337e-02 2.00169474e-01 6.68079674e-01 -3.53460431e-01
-1.25324690e+00 -1.00309157e+00 3.99185389e-01 -1.49255425e-01
1.24337889e-01 -6.59656942e-01 -1.01363349e+00 5.05886912e-01
2.82572985e-01 5.47381490e-02 7.61495650e-01 -4.36368227e-01
-3.25902462e-01 -3.48913461e-01 -1.31278419e+00 7.17079341e-01
9.21257496e-01 -5.23958206e-01 -2.00319514e-01 1.28908500e-01
8.60316515e-01 -7.24046230e-01 -5.30933380e-01 2.72965401e-01
6.03000343e-01 -1.10569310e+00 1.01773906e+00 -1.57798119e-02
3.80524307e-01 -4.62282360e-01 -2.14586496e-01 -1.30795753e+00
-9.22543406e-02 -8.79493773e-01 -1.16064802e-01 1.26648700e+00
9.18050185e-02 -4.80688661e-01 4.86290723e-01 5.25315583e-01
-2.63961285e-01 -6.55669987e-01 -8.09088171e-01 -1.35310471e-01
-1.50044069e-01 -6.98122978e-01 6.82516634e-01 7.18534708e-01
-1.32856861e-01 2.43661210e-01 -4.37478900e-01 2.12855369e-01
6.72453940e-01 6.45241514e-02 7.73044646e-01 -9.37982023e-01
-4.65923220e-01 -2.22787008e-01 -7.74987116e-02 -1.25093627e+00
2.98952639e-01 -7.39522994e-01 2.74528325e-01 -1.49680340e+00
-1.57528326e-01 -3.78972977e-01 1.34480596e-01 3.27305555e-01
1.97774038e-01 4.69989538e-01 2.39952803e-01 -1.47513717e-01
2.80619264e-02 8.72785330e-01 1.43726993e+00 -1.62104983e-02
-2.86542028e-01 1.69504240e-01 -6.70148134e-01 9.40212905e-01
5.94975233e-01 -1.87285677e-01 -4.97582108e-01 -4.80908632e-01
-8.53139237e-02 6.76896095e-01 5.47462583e-01 -9.51277912e-01
8.34661350e-02 -1.26757070e-01 5.69285750e-01 -4.29273129e-01
9.22396302e-01 -8.55298638e-01 4.34272289e-01 8.59547555e-02
-3.09635907e-01 -2.60786891e-01 2.70755738e-01 3.83806586e-01
-2.37080723e-01 -5.25369644e-02 9.07417655e-01 -7.60506541e-02
-2.29236454e-01 2.04755977e-01 -4.74278122e-01 -1.11050732e-01
7.48995602e-01 -3.76024723e-01 1.20245805e-02 -1.00188267e+00
-6.71491206e-01 -1.04695834e-01 2.47847930e-01 4.74414647e-01
9.60570455e-01 -1.53527093e+00 -6.93143785e-01 4.36026275e-01
-2.24282309e-01 2.40407318e-01 6.11241400e-01 4.60753024e-01
-4.51475471e-01 8.39142501e-02 -2.49923561e-02 -8.47501993e-01
-1.28743303e+00 2.95276910e-01 6.32336199e-01 5.76435804e-01
-8.57891321e-01 8.78602386e-01 9.06659484e-01 -4.51008141e-01
5.30265152e-01 -1.16929427e-01 -1.02917790e-01 -1.21491924e-01
8.26745510e-01 2.05815494e-01 -1.55757904e-01 -1.01387489e+00
-7.71198794e-02 9.22104597e-01 4.82631922e-01 -3.60744506e-01
1.19537449e+00 -2.16463089e-01 9.27551165e-02 4.71419990e-01
1.31876218e+00 3.69163066e-01 -1.57409465e+00 -1.64743885e-01
-1.00192356e+00 -6.67267799e-01 2.29275212e-01 -4.93773162e-01
-1.43454957e+00 1.17490447e+00 6.30621850e-01 -1.75582349e-01
1.50364947e+00 -1.22207433e-01 8.09873164e-01 -1.71927199e-01
2.99935639e-01 -7.77461827e-01 2.47791663e-01 1.87561750e-01
1.28503883e+00 -6.76033914e-01 -3.41346949e-01 -6.19066417e-01
-5.93421578e-01 1.13509870e+00 6.48466349e-01 1.43356889e-01
6.59438252e-01 5.57797253e-01 4.28481430e-01 6.01810776e-02
-5.16780436e-01 1.75140306e-01 2.85319090e-01 7.90645301e-01
6.39993429e-01 -6.79923669e-02 4.24692035e-01 5.48598409e-01
-4.14660692e-01 -8.99710283e-02 4.44632351e-01 6.42575502e-01
-2.46576354e-01 -6.71899498e-01 -8.29834402e-01 -1.04496807e-01
-3.84451628e-01 -3.60282771e-02 -1.99187100e-01 3.68655026e-01
1.43249825e-01 8.98715734e-01 1.75656334e-01 -3.23095471e-01
4.47802365e-01 4.16133329e-02 5.46999216e-01 -4.71244961e-01
-4.22388047e-01 6.60458565e-01 -2.38194361e-01 -6.83978260e-01
-2.33371630e-01 -3.61302495e-01 -1.38221908e+00 -1.37709364e-01
-2.75121033e-01 -5.19249449e-03 7.91073680e-01 3.59783202e-01
3.24117690e-01 6.89290226e-01 8.57118785e-01 -1.40414250e+00
-1.69404782e-02 -6.44518197e-01 -6.05673075e-01 1.92907587e-01
5.25095165e-01 -6.38855815e-01 -4.29735392e-01 2.82155842e-01] | [13.16397762298584, -0.4364239275455475] |
4c3bc958-d439-4e5a-879f-2fcac664d693 | korean-timebank-including-relative-temporal | null | null | https://aclanthology.org/L18-1326 | https://aclanthology.org/L18-1326.pdf | Korean TimeBank Including Relative Temporal Information | null | ['Ho-Jin Choi', 'Young-Seob Jeong', 'Chae-Gyun Lim'] | 2018-05-01 | korean-timebank-including-relative-temporal-1 | https://aclanthology.org/L18-1326 | https://aclanthology.org/L18-1326.pdf | lrec-2018-5 | ['temporal-information-extraction'] | ['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] | [-7.358663558959961, 3.7095444202423096] |
54baea31-30b2-4fb5-bc11-4b7e6cf1ff09 | variable-complexity-weighted-tempered-gibbs | 2304.02899 | null | https://arxiv.org/abs/2304.02899v1 | https://arxiv.org/pdf/2304.02899v1.pdf | Variable-Complexity Weighted-Tempered Gibbs Samplers for Bayesian Variable Selection | Subset weighted-Tempered Gibbs Sampler (wTGS) has been recently introduced by Jankowiak to reduce the computation complexity per MCMC iteration in high-dimensional applications where the exact calculation of the posterior inclusion probabilities (PIP) is not essential. However, the Rao-Backwellized estimator associated with this sampler has a high variance as the ratio between the signal dimension and the number of conditional PIP estimations is large. In this paper, we design a new subset weighted-Tempered Gibbs Sampler (wTGS) where the expected number of computations of conditional PIPs per MCMC iteration can be much smaller than the signal dimension. Different from the subset wTGS and wTGS, our sampler has a variable complexity per MCMC iteration. We provide an upper bound on the variance of an associated Rao-Blackwellized estimator for this sampler at a finite number of iterations, $T$, and show that the variance is $O\big(\big(\frac{P}{S}\big)^2 \frac{\log T}{T}\big)$ for a given dataset where $S$ is the expected number of conditional PIP computations per MCMC iteration. Experiments show that our Rao-Blackwellized estimator can have a smaller variance than its counterpart associated with the subset wTGS. | ['Lan V. Truong'] | 2023-04-06 | null | null | null | null | ['variable-selection'] | ['methodology'] | [ 2.60935158e-01 -1.86679431e-03 3.46768707e-01 -2.04437539e-01
-1.37806594e+00 -2.04263628e-01 2.80927807e-01 2.95225065e-02
-7.65163898e-01 7.39971697e-01 -3.08796644e-01 -3.42294961e-01
-2.05848530e-01 -9.68964219e-01 -6.36683524e-01 -9.49802697e-01
-4.68946517e-01 7.30449438e-01 5.61086237e-01 3.87226641e-01
1.06570169e-01 1.91154897e-01 -1.23979139e+00 -5.32899797e-01
8.97569418e-01 1.01762366e+00 1.35196120e-01 1.00444090e+00
8.67648572e-02 1.47517072e-02 -5.62918901e-01 -2.00926319e-01
1.34880900e-01 -8.19102466e-01 -2.49261469e-01 -6.96578547e-02
-1.73783749e-01 -1.51778072e-01 3.07548970e-01 1.28584063e+00
5.66102803e-01 1.96649328e-01 8.81665647e-01 -9.40253794e-01
4.92391765e-01 6.09449625e-01 -1.08843005e+00 1.16458289e-01
-1.49582356e-01 -2.95314789e-01 7.20901787e-01 -9.88495588e-01
2.06522331e-01 1.09163296e+00 8.19766879e-01 1.87769830e-02
-1.45037222e+00 -1.04321802e+00 -1.18421324e-01 -3.62982899e-01
-1.88910317e+00 -2.25968644e-01 2.88754016e-01 -4.39923197e-01
2.52958387e-01 2.36281618e-01 6.41088724e-01 4.78802115e-01
1.34611815e-01 5.64267397e-01 9.53292549e-01 -4.48046714e-01
8.78362298e-01 -9.48158186e-03 1.45898357e-01 6.03256941e-01
7.15003967e-01 -1.07923746e-01 -3.78004193e-01 -4.97323036e-01
8.33855450e-01 -6.20615110e-02 -1.54295728e-01 -8.15766752e-02
-1.09481943e+00 1.02880883e+00 -4.24534202e-01 1.51170716e-01
-2.51573473e-01 6.75739646e-01 1.03670396e-02 -9.98810157e-02
4.59649831e-01 9.99879837e-02 -4.02924567e-01 -4.47951287e-01
-1.27774191e+00 3.14640045e-01 7.88155615e-01 1.13595223e+00
8.19968462e-01 7.76454359e-02 -3.66618857e-02 6.01296067e-01
4.52381372e-01 1.29430926e+00 -9.21553187e-03 -9.92996037e-01
3.93749237e-01 -4.55362201e-02 4.18609977e-01 -6.28153026e-01
-2.65152603e-01 -4.78024572e-01 -1.10041559e+00 -1.11350186e-01
7.83091903e-01 -5.90751052e-01 -4.84734178e-01 1.68737924e+00
3.55612457e-01 -6.36502577e-04 -3.20709467e-01 4.63847250e-01
-1.91794813e-01 6.75796926e-01 -2.88450301e-01 -6.35657012e-01
1.43092740e+00 -5.30718081e-02 -5.84038377e-01 -3.14222902e-01
3.72598499e-01 -7.71665335e-01 8.45171750e-01 4.79493916e-01
-1.12081337e+00 -1.35354713e-01 -1.12693989e+00 5.40491819e-01
2.41992652e-01 3.18163959e-03 3.80270004e-01 1.21141326e+00
-4.91280556e-01 4.17548388e-01 -1.20254564e+00 -8.48954841e-02
2.39339769e-01 1.56462610e-01 3.08633357e-01 -2.62816204e-03
-8.04171324e-01 3.33301395e-01 1.48255333e-01 7.64388442e-02
-8.48938465e-01 -4.29136753e-01 -5.50292492e-01 1.58717409e-01
4.11947995e-01 -1.99673831e-01 9.55031633e-01 -4.28298526e-02
-1.38292086e+00 1.87684342e-01 -4.09958035e-01 -4.96457458e-01
6.02833629e-01 -4.67487872e-02 2.75531728e-02 1.00802906e-01
3.02719444e-01 -2.72471607e-02 1.04633892e+00 -8.19296062e-01
-5.43237329e-01 -6.44212604e-01 -6.45043790e-01 -1.91293955e-01
-5.82281277e-02 -2.06858546e-01 -6.66605294e-01 -6.41815305e-01
6.19993508e-01 -1.10162425e+00 -4.65347648e-01 -2.43261188e-01
-2.19528049e-01 1.20141106e-02 3.30714703e-01 -1.09764926e-01
1.20129716e+00 -2.28638148e+00 -3.02217513e-01 6.86187804e-01
3.43122743e-02 -1.72576159e-01 4.48220551e-01 3.58617365e-01
3.47167373e-01 -7.62891695e-02 -5.69750607e-01 -3.20608526e-01
-2.34465376e-02 2.08883323e-02 -2.16646194e-01 1.00351906e+00
-3.71959716e-01 -7.82603677e-03 -9.01681066e-01 -5.57069600e-01
-8.20944533e-02 4.73903745e-01 -7.80164838e-01 -1.08111933e-01
9.99616385e-02 1.38534442e-01 -6.69309318e-01 2.02992365e-01
1.10256147e+00 -5.40600955e-01 4.80883598e-01 8.69080126e-02
-1.04700930e-01 -3.73498909e-02 -1.76720548e+00 1.20603645e+00
-2.94272155e-01 4.12373304e-01 2.57415175e-01 -9.06230330e-01
8.77430618e-01 3.09707254e-01 5.73734581e-01 -2.15837657e-02
1.96036980e-01 4.23584312e-01 -5.08821011e-01 2.27060899e-01
4.60564464e-01 -7.05144227e-01 -2.30088502e-01 7.56738663e-01
-1.96111307e-01 -2.65104085e-01 2.42506072e-01 2.32230991e-01
1.11406493e+00 -1.94610447e-01 4.95252848e-01 -5.32770813e-01
3.07279140e-01 -4.08911765e-01 8.98626983e-01 1.06152344e+00
-1.39135867e-01 5.19212604e-01 7.01866388e-01 1.00848541e-01
-8.36127579e-01 -1.31339908e+00 -5.34050584e-01 8.03416610e-01
1.52179301e-01 -5.26476264e-01 -8.39804888e-01 -1.98508531e-01
-2.03690991e-01 6.08628154e-01 -4.26412165e-01 1.21586390e-01
-3.68253440e-01 -1.32673144e+00 5.68083346e-01 3.52388889e-01
5.70862651e-01 -3.56570244e-01 -1.00549495e+00 2.44855478e-01
-1.16022021e-01 -9.57432926e-01 -6.14246726e-01 4.21220243e-01
-1.13739491e+00 -1.05423355e+00 -9.62694824e-01 -1.69812128e-01
7.67488420e-01 1.66038886e-01 6.68018460e-01 -5.18013954e-01
1.33159403e-02 3.16050082e-01 -3.30520749e-01 -6.11924469e-01
2.60879043e-02 -2.49777257e-01 1.60342664e-01 1.12421602e-01
2.66181916e-01 -5.33345759e-01 -7.30285525e-01 5.17003477e-01
-8.02780509e-01 -1.62648827e-01 3.46582204e-01 8.14893961e-01
7.36355007e-01 3.22507024e-01 3.08072090e-01 -8.13610911e-01
2.98476338e-01 -1.75294220e-01 -1.22298694e+00 -2.13318706e-01
-4.89807993e-01 3.50417584e-01 3.46071690e-01 -4.11176324e-01
-8.95017266e-01 8.55759829e-02 -1.88044794e-02 -1.22388773e-01
4.42949325e-01 4.46365654e-01 1.27512932e-01 2.77796090e-01
5.30626535e-01 2.76691437e-01 -1.03038490e-01 -6.96996212e-01
2.89160043e-01 6.26472175e-01 2.84878045e-01 -6.63459480e-01
5.47789514e-01 7.83626974e-01 4.49895054e-01 -1.13769293e+00
-5.56442261e-01 -2.77133554e-01 7.64483819e-03 -2.54782531e-02
7.33246148e-01 -7.89530575e-01 -1.27250266e+00 2.57938564e-01
-7.49771655e-01 -2.47053280e-01 -3.87792170e-01 1.11968625e+00
-7.29351759e-01 4.74095494e-01 -2.48446345e-01 -1.49521339e+00
-1.82394847e-01 -1.08748138e+00 8.69209766e-01 -1.10730097e-01
-1.53697938e-01 -6.16886318e-01 1.14354469e-01 -1.47308717e-02
1.93674266e-01 3.07857662e-01 6.87763631e-01 -3.74950290e-01
-4.28483933e-01 -5.95446885e-01 -2.63865799e-01 2.26687297e-01
-1.10993899e-01 -3.09098959e-01 -4.48807329e-01 -4.49470788e-01
4.67722446e-01 2.73367524e-01 7.42005467e-01 9.22612131e-01
8.66506994e-01 -2.84401417e-01 -4.27799493e-01 3.42136502e-01
1.49396873e+00 2.20623478e-01 6.31317616e-01 -1.08564638e-01
3.30846272e-02 -3.40281650e-02 7.59573221e-01 1.08694243e+00
-6.00916259e-02 4.90392894e-01 6.71167299e-02 3.81487638e-01
3.28544736e-01 -4.90389280e-02 4.77470607e-01 7.56756246e-01
-4.57521752e-02 -2.31358614e-02 -9.07981932e-01 5.86570859e-01
-1.62943268e+00 -8.26097012e-01 -3.03543389e-01 2.88283396e+00
8.29383790e-01 3.41452599e-01 1.52015373e-01 3.37598354e-01
8.95819724e-01 2.66338270e-02 -3.43600541e-01 -5.21964813e-03
1.25872180e-01 5.69880545e-01 1.03714335e+00 6.67862892e-01
-6.88826144e-01 3.73411149e-01 6.07047224e+00 1.13172233e+00
-4.39096272e-01 2.35398933e-01 2.89734870e-01 -3.80038500e-01
2.57891323e-02 2.41615489e-01 -1.09543061e+00 8.87277305e-01
1.04658866e+00 -1.22603625e-01 1.44314200e-01 9.27860379e-01
2.12800622e-01 -8.81175876e-01 -9.62740898e-01 1.00672245e+00
-1.66353151e-01 -9.64232922e-01 -4.64980930e-01 4.81959999e-01
5.35061479e-01 -1.12689370e-02 -4.95963097e-02 7.18209893e-02
7.28040457e-01 -5.13562381e-01 4.71780062e-01 2.83015668e-01
8.61631095e-01 -9.54129577e-01 7.16037512e-01 5.18088341e-01
-1.47807705e+00 1.08345218e-01 -5.00878036e-01 1.96692929e-01
4.51304138e-01 1.36187065e+00 -6.53898418e-01 1.81835338e-01
6.85076952e-01 4.76298435e-03 1.28968075e-01 7.71362901e-01
1.37556061e-01 6.94998205e-01 -1.02961779e+00 -3.28227311e-01
1.34786395e-02 -4.83646393e-01 7.51151562e-01 1.12691498e+00
8.16785395e-01 3.80187154e-01 -2.00369190e-02 5.38711965e-01
1.95052519e-01 -1.51974484e-01 -6.56954572e-02 1.00265943e-01
8.60544622e-01 6.86383247e-01 -1.07632983e+00 -4.82615739e-01
-8.08555484e-02 5.77838838e-01 -2.17386246e-01 3.55784088e-01
-8.19108009e-01 -7.49714553e-01 4.24416631e-01 2.97347158e-01
8.80488276e-01 -5.05571008e-01 -4.23395783e-01 -1.00899160e+00
-1.52816281e-01 -4.30060416e-01 4.58220303e-01 -2.02535287e-01
-9.87755597e-01 1.14938408e-01 1.87129602e-01 -1.33229101e+00
-4.37542945e-02 -2.77316362e-01 -3.09737444e-01 8.55527282e-01
-8.13612700e-01 -2.89671510e-01 1.30101830e-01 6.44613981e-01
1.58520505e-01 3.70312706e-02 6.77286446e-01 2.41701677e-01
-4.89796847e-01 5.32019734e-01 6.85728133e-01 2.36314107e-02
4.14843321e-01 -1.09800279e+00 3.93436663e-02 8.69865894e-01
-4.39516485e-01 7.40868330e-01 1.09295702e+00 -6.67812586e-01
-1.44268548e+00 -7.55539775e-01 6.76086545e-01 5.89885423e-03
4.75939691e-01 -3.56539220e-01 -4.00575429e-01 5.72110236e-01
-4.00362253e-01 -5.69541678e-02 8.48168433e-01 2.99153179e-01
-9.48457867e-02 -3.26126277e-01 -1.25765455e+00 5.45820594e-01
4.87730116e-01 -2.76063442e-01 -2.27473959e-01 1.57547668e-01
1.27229691e-01 -1.03414141e-01 -1.04604685e+00 3.02538514e-01
7.83883750e-01 -1.00672352e+00 8.16483080e-01 3.85064453e-01
-3.65932226e-01 -4.62508976e-01 -4.91502523e-01 -7.35666871e-01
3.86271137e-03 -8.92746627e-01 -7.63504580e-02 8.33541572e-01
3.02046299e-01 -7.45569229e-01 9.11897898e-01 4.37485993e-01
4.66561794e-01 -5.18419027e-01 -1.37167394e+00 -1.12227035e+00
-1.41727701e-02 -8.70997608e-01 2.68130839e-01 2.60909498e-01
2.82458514e-02 1.78818375e-01 -3.43338281e-01 1.86281160e-01
1.28361523e+00 -2.16278937e-02 8.53726983e-01 -1.30317903e+00
-6.55224264e-01 -1.52288198e-01 -8.30209553e-02 -1.29287958e+00
-7.83257127e-01 -4.24124002e-01 3.29088479e-01 -1.08064270e+00
3.33790064e-01 -5.73280454e-01 4.88861762e-02 -7.84350112e-02
-1.09057039e-01 3.27988602e-02 1.03171825e-01 1.40049160e-01
-4.41116452e-01 5.71518421e-01 9.28221464e-01 2.55085170e-01
-2.82510936e-01 2.65935212e-01 -2.80277282e-01 9.57702875e-01
3.92607391e-01 -9.86237586e-01 -2.52597839e-01 8.38533342e-02
5.00742018e-01 5.83846331e-01 -5.61056323e-02 -1.07858872e+00
1.44122809e-01 1.10260688e-03 1.25616670e-01 -1.18488550e+00
5.54053843e-01 -5.68262041e-01 3.14844012e-01 7.18614519e-01
-1.90268680e-02 -3.38218808e-01 -1.28942922e-01 1.06932402e+00
2.55539209e-01 -3.39085698e-01 1.17589402e+00 -1.39275700e-01
4.78240818e-01 1.33077919e-01 -8.21196496e-01 1.38524383e-01
8.87550712e-01 -1.60057321e-01 2.68949300e-01 -6.06129169e-01
-5.95973313e-01 6.62127286e-02 9.65035334e-02 -5.80309868e-01
5.08724511e-01 -1.24389720e+00 -5.06741881e-01 1.24058343e-01
-1.85676813e-01 2.36393899e-01 3.28642607e-01 1.12940109e+00
-3.96016777e-01 2.45880350e-01 4.88272578e-01 -7.31305003e-01
-9.49044883e-01 1.63233235e-01 3.44077288e-03 -4.20322388e-01
-5.24344444e-01 9.51256275e-01 1.67832315e-01 2.37472933e-02
2.54750103e-01 -3.74495089e-01 5.76395810e-01 1.95494846e-01
7.20604539e-01 9.19535637e-01 -2.83369601e-01 -7.99157247e-02
-4.21247989e-01 6.04540765e-01 9.33432858e-03 -8.14532280e-01
1.25896406e+00 -1.94690377e-01 -1.48514420e-01 6.06083751e-01
1.23697376e+00 1.80100560e-01 -1.35666811e+00 -2.37536922e-01
-8.57706815e-02 -5.94795048e-01 -8.27642828e-02 -1.64640471e-01
-7.84778893e-01 9.12240744e-01 6.59004927e-01 2.03124806e-01
9.44684327e-01 -8.73965845e-02 7.09170699e-01 2.35721663e-01
7.21749425e-01 -1.29705429e+00 -4.56647426e-02 4.98797059e-01
3.83560032e-01 -7.35177457e-01 3.74816149e-01 -2.32429385e-01
-3.99272710e-01 7.82889068e-01 -4.21825834e-02 -2.54054099e-01
1.16285861e+00 2.24570662e-01 -5.76685548e-01 -1.07589543e-01
-2.32841074e-01 -1.59049884e-03 -1.66370153e-01 3.46992701e-01
3.12258571e-01 2.99165845e-01 -6.57754958e-01 7.15106547e-01
-2.10280463e-01 1.38214584e-02 5.93361855e-01 9.12767708e-01
-8.62389326e-01 -9.84744191e-01 -6.80611908e-01 5.65487325e-01
-5.85224986e-01 -3.19600441e-02 3.51245165e-01 6.37109101e-01
-1.90186113e-01 1.03291762e+00 1.43233910e-01 2.17956007e-02
-5.82222492e-02 3.13035063e-02 4.45195466e-01 -3.86284083e-01
5.70005849e-02 7.82446384e-01 -7.06482828e-02 -3.74732167e-01
-2.78754473e-01 -8.66501391e-01 -1.28207338e+00 -5.73254526e-01
-4.85009640e-01 6.01151228e-01 6.72937095e-01 8.97472739e-01
2.03508474e-02 6.13810495e-02 6.18076861e-01 -5.79749048e-01
-4.95823324e-01 -1.03965318e+00 -1.32802463e+00 -3.26843202e-01
1.11127816e-01 -5.40519238e-01 -5.85729837e-01 -2.03016222e-01] | [6.704499244689941, 4.191630840301514] |
214e038c-b4f5-4a6d-aaed-63b882c28efd | reverse-kl-divergence-training-of-prior | 1905.13472 | null | https://arxiv.org/abs/1905.13472v2 | https://arxiv.org/pdf/1905.13472v2.pdf | Reverse KL-Divergence Training of Prior Networks: Improved Uncertainty and Adversarial Robustness | Ensemble approaches for uncertainty estimation have recently been applied to the tasks of misclassification detection, out-of-distribution input detection and adversarial attack detection. Prior Networks have been proposed as an approach to efficiently \emph{emulate} an ensemble of models for classification by parameterising a Dirichlet prior distribution over output distributions. These models have been shown to outperform alternative ensemble approaches, such as Monte-Carlo Dropout, on the task of out-of-distribution input detection. However, scaling Prior Networks to complex datasets with many classes is difficult using the training criteria originally proposed. This paper makes two contributions. First, we show that the appropriate training criterion for Prior Networks is the \emph{reverse} KL-divergence between Dirichlet distributions. This addresses issues in the nature of the training data target distributions, enabling prior networks to be successfully trained on classification tasks with arbitrarily many classes, as well as improving out-of-distribution detection performance. Second, taking advantage of this new training criterion, this paper investigates using Prior Networks to detect adversarial attacks and proposes a generalized form of adversarial training. It is shown that the construction of successful \emph{adaptive} whitebox attacks, which affect the prediction and evade detection, against Prior Networks trained on CIFAR-10 and CIFAR-100 using the proposed approach requires a greater amount of computational effort than against networks defended using standard adversarial training or MC-dropout. | ['Andrey Malinin', 'Mark Gales'] | 2019-05-31 | reverse-kl-divergence-training-of-prior-1 | http://papers.nips.cc/paper/9597-reverse-kl-divergence-training-of-prior-networks-improved-uncertainty-and-adversarial-robustness | http://papers.nips.cc/paper/9597-reverse-kl-divergence-training-of-prior-networks-improved-uncertainty-and-adversarial-robustness.pdf | neurips-2019-12 | ['adversarial-attack-detection', 'adversarial-attack-detection'] | ['computer-vision', 'knowledge-base'] | [ 3.73632014e-01 6.62076324e-02 2.13819429e-01 -4.32838500e-01
-8.87842953e-01 -9.13846850e-01 6.80844188e-01 -1.44401446e-01
-5.57211876e-01 9.97415662e-01 -2.68567204e-01 -6.12457633e-01
-2.14330465e-01 -8.42661619e-01 -7.34799445e-01 -8.54211092e-01
-9.02718380e-02 6.01673424e-01 1.83374077e-01 -1.53177902e-01
1.95232764e-01 8.74078393e-01 -1.30052876e+00 3.55933905e-01
6.52817845e-01 8.96320343e-01 -6.68210328e-01 9.14763749e-01
3.53015155e-01 4.44077611e-01 -1.43325889e+00 -8.19898129e-01
1.53077602e-01 -4.32074398e-01 -3.24383825e-01 -3.29738021e-01
5.99480748e-01 -5.58247447e-01 -4.61673468e-01 1.39898920e+00
7.79145002e-01 1.88996032e-01 1.03683209e+00 -1.38371515e+00
-5.63356340e-01 6.95557415e-01 3.12723517e-02 5.26106417e-01
-2.01250210e-01 6.41444400e-02 7.73935616e-01 -7.11726010e-01
7.72946328e-02 1.35395837e+00 6.61711872e-01 9.43585992e-01
-1.49766171e+00 -1.09283316e+00 -1.64354481e-02 -1.70166701e-01
-1.39893126e+00 -3.55189055e-01 6.42440140e-01 -4.57199425e-01
7.82432020e-01 2.72951812e-01 -1.97946936e-01 1.79344201e+00
1.36951685e-01 5.89622259e-01 1.15514195e+00 -3.26848686e-01
4.90761071e-01 6.25171602e-01 1.42838463e-01 1.64979979e-01
3.37926507e-01 5.73280394e-01 -1.08456559e-01 -6.76728249e-01
5.91433167e-01 -2.23413691e-01 -3.25214893e-01 7.26226941e-02
-2.33982220e-01 1.22514474e+00 5.01426160e-02 1.90913886e-01
1.86768658e-02 3.48936766e-01 3.74453038e-01 4.38472033e-01
7.34578192e-01 4.96244282e-01 -3.97867769e-01 1.27911359e-01
-9.02748644e-01 8.63905922e-02 1.09655380e+00 6.57442570e-01
2.02057183e-01 6.73109114e-01 -2.11417645e-01 5.45940042e-01
3.47414374e-01 7.63120353e-01 1.19971205e-02 -5.95093548e-01
5.45182228e-01 -2.41892740e-01 2.92849571e-01 -6.50898457e-01
6.80833608e-02 -8.55473399e-01 -8.23994696e-01 9.78756666e-01
7.51590550e-01 -8.02398980e-01 -1.07898808e+00 1.76482558e+00
2.64408290e-02 2.83018887e-01 2.59620219e-01 3.53118867e-01
2.49074429e-01 2.34590322e-01 2.05895096e-01 2.17989013e-01
1.04861152e+00 -5.22167683e-02 -4.77125674e-01 4.85822335e-02
3.39897275e-01 -5.70980430e-01 4.25082505e-01 7.83569992e-01
-8.12985361e-01 -4.17066157e-01 -1.41001415e+00 7.64917374e-01
-6.61882997e-01 -1.12926848e-01 3.03403258e-01 1.76082051e+00
-6.83521152e-01 9.62638438e-01 -6.24776363e-01 3.59507084e-01
7.75442123e-01 7.28997707e-01 -3.10794096e-02 2.69162923e-01
-1.65435314e+00 1.00961363e+00 4.55524713e-01 -1.43416733e-01
-1.51073146e+00 -6.15559578e-01 -4.71478969e-01 1.78150862e-01
-1.86767012e-01 -2.39574432e-01 8.82062376e-01 -1.12074172e+00
-1.51913798e+00 5.12807250e-01 5.01732945e-01 -9.00744557e-01
6.98275328e-01 -2.56032139e-01 -5.98125398e-01 2.47519221e-02
-6.71223581e-01 3.31469715e-01 1.40118575e+00 -1.11513841e+00
-5.44131637e-01 -2.37534210e-01 1.32365838e-01 -1.65823981e-01
-6.10463560e-01 2.75505930e-01 4.43573296e-01 -9.65849996e-01
-5.26532173e-01 -6.49222374e-01 -2.20968500e-01 -4.53347266e-01
-3.76847833e-01 -1.63560465e-01 9.68868732e-01 -5.18775403e-01
1.09043109e+00 -2.11796951e+00 -8.46713781e-02 7.23435640e-01
-4.10670005e-02 6.25428855e-01 1.27581522e-01 1.61279768e-01
-3.03602010e-01 4.48719710e-01 -3.92604351e-01 -4.93317664e-01
3.93288523e-01 2.73618966e-01 -1.05492783e+00 5.23234725e-01
3.56822491e-01 1.75047949e-01 -5.96908808e-01 1.55476229e-02
1.30158633e-01 8.32732081e-01 -5.41354418e-01 5.14954664e-02
-1.85844511e-01 3.35330307e-01 -1.83208555e-01 2.55276680e-01
7.36784399e-01 4.10900831e-01 -1.66321546e-01 2.49471098e-01
5.76892376e-01 -2.55514551e-02 -1.39700377e+00 6.10958457e-01
-3.10746789e-01 5.41909575e-01 1.77537113e-01 -1.05091035e+00
9.72990632e-01 6.77878797e-01 -1.45733394e-02 1.24612190e-01
5.17766953e-01 3.18954408e-01 5.29037155e-02 2.71847397e-01
2.17839684e-02 -2.71687835e-01 -2.49622479e-01 3.29717249e-01
4.06532437e-01 -8.89511779e-02 1.27526857e-02 1.03884928e-01
9.82291102e-01 -2.07317084e-01 -1.94076732e-01 -1.58567101e-01
5.18457472e-01 -3.67745668e-01 3.84320736e-01 1.23807991e+00
-1.45841166e-01 4.30323422e-01 6.65016949e-01 -1.13877326e-01
-1.05237961e+00 -1.40579259e+00 -5.84191203e-01 1.05932629e+00
-2.92994618e-01 1.62895218e-01 -9.68476176e-01 -1.13104939e+00
4.12922129e-02 1.24358141e+00 -5.85209608e-01 -3.13845575e-01
-4.55183536e-01 -1.35955620e+00 1.42374432e+00 7.40501046e-01
4.01426792e-01 -7.96790242e-01 -9.91659984e-02 2.44278118e-01
3.77497166e-01 -9.13716078e-01 -7.54356757e-02 6.93901002e-01
-7.66688049e-01 -8.17397237e-01 -6.31374538e-01 -3.67808610e-01
6.79977357e-01 -6.96026742e-01 1.03638947e+00 -2.15541705e-01
-2.46389851e-01 5.59551418e-01 -6.52651414e-02 -8.24662626e-01
-8.82969737e-01 -3.32205176e-01 3.33288878e-01 1.21247157e-01
4.46118176e-01 -9.53129411e-01 -2.68056482e-01 5.23667693e-01
-1.05281496e+00 -9.79713678e-01 4.61045444e-01 1.06085908e+00
3.08629543e-01 5.40299773e-01 1.00405729e+00 -1.05363381e+00
7.86461413e-01 -5.27405202e-01 -8.27504039e-01 1.37987480e-01
-4.93056029e-01 1.40957415e-01 7.62281060e-01 -8.72716904e-01
-1.07665348e+00 -1.93966344e-01 -6.17843688e-01 -7.95876622e-01
-5.02278745e-01 -3.21163610e-02 -4.51557100e-01 -3.69713455e-01
1.17307508e+00 -2.06054077e-01 -1.94042668e-01 -3.06180596e-01
2.83538163e-01 5.75791836e-01 5.58840454e-01 -7.22202003e-01
1.05860269e+00 3.37970555e-01 4.66328450e-02 -6.04748666e-01
-7.39995837e-01 -9.42777172e-02 -5.00365794e-01 1.03781246e-01
8.23582053e-01 -6.43785000e-01 -4.48435575e-01 6.05338216e-01
-9.72413182e-01 -1.12740315e-01 -3.69322479e-01 5.67396879e-01
-3.32102567e-01 3.21027040e-01 -6.75603628e-01 -1.17214084e+00
-3.19445163e-01 -9.86159980e-01 4.45575029e-01 -5.57195507e-02
-6.01687431e-02 -1.44331837e+00 -7.77078569e-02 -1.52460067e-03
6.46147013e-01 3.72261584e-01 8.64726543e-01 -1.57371354e+00
-1.94136307e-01 -6.90328360e-01 1.33642167e-01 1.21939111e+00
-4.02907521e-01 -1.81248616e-02 -1.53195679e+00 -4.94186133e-01
2.80305088e-01 -2.76364297e-01 8.69267702e-01 2.26896420e-01
1.20297694e+00 -2.56825030e-01 -1.89977080e-01 5.64064622e-01
1.29493940e+00 2.52951354e-01 7.64933169e-01 -2.57016048e-02
3.29725951e-01 1.99071839e-01 5.85140325e-02 4.12185192e-01
-7.03225791e-01 2.97384113e-01 6.80291295e-01 2.36532480e-01
1.19153440e-01 -7.12975813e-03 6.71889067e-01 -8.12383145e-02
1.63225815e-01 -6.25691056e-01 -7.39530444e-01 2.37544313e-01
-1.26328325e+00 -1.27848434e+00 9.67471153e-02 2.42253089e+00
6.78173661e-01 7.70725429e-01 -2.69594509e-02 3.31150323e-01
9.52729583e-01 -2.35360831e-01 -4.54482168e-01 -4.99625206e-01
-1.59580961e-01 8.45602751e-01 7.80228555e-01 6.61284029e-01
-1.67344034e+00 4.98520106e-01 6.42987061e+00 1.24962783e+00
-6.70513451e-01 2.76513755e-01 8.70366693e-01 7.31661618e-02
-5.81737645e-02 -1.40857503e-01 -1.37437952e+00 5.35375595e-01
1.21727037e+00 2.43329123e-01 2.11059242e-01 9.29356515e-01
-3.85431916e-01 1.88716084e-01 -1.31826615e+00 5.04385591e-01
2.98069119e-01 -9.57255423e-01 1.92202590e-02 1.07383996e-01
8.04365575e-01 -2.46957541e-01 7.05669284e-01 7.14766204e-01
8.97987187e-01 -1.34206676e+00 4.11305815e-01 4.42917883e-01
6.80831134e-01 -1.15635443e+00 1.01503873e+00 3.99543107e-01
-5.31391323e-01 -1.44554749e-01 -3.16642284e-01 2.81538337e-01
-8.45066160e-02 6.50767684e-01 -8.71022999e-01 2.87360191e-01
5.09062767e-01 -9.84554812e-02 -3.31901252e-01 1.02275193e+00
-5.40323615e-01 1.08125401e+00 -6.92932069e-01 5.16160913e-02
2.00966209e-01 3.19860071e-01 9.16879594e-01 1.27935803e+00
2.33567849e-01 -2.14429796e-01 9.30576026e-02 8.05203319e-01
-1.18303165e-01 -5.25613546e-01 -5.00035167e-01 1.81450129e-01
5.58884859e-01 8.35334182e-01 -5.83793998e-01 -3.11806321e-01
1.70138538e-01 7.44337082e-01 1.79646481e-02 5.21086633e-01
-1.05532742e+00 -8.39304209e-01 5.73594511e-01 -2.59594113e-01
5.71708858e-01 1.16597697e-01 -2.04503819e-01 -8.46361101e-01
-3.14947546e-01 -9.85298455e-01 7.10689068e-01 -2.86394477e-01
-1.85041428e+00 6.56073689e-01 4.11020726e-01 -1.05868590e+00
-3.56756777e-01 -1.07092142e+00 -9.48621273e-01 1.20997453e+00
-1.04569638e+00 -8.08308244e-01 3.80247533e-01 7.16014147e-01
1.56417638e-01 -5.69772422e-01 1.11652541e+00 3.19819152e-01
-5.79550862e-01 1.13584960e+00 4.98405576e-01 5.16159892e-01
6.46545708e-01 -1.52953970e+00 1.42651379e-01 1.04872227e+00
1.88620582e-01 4.18032110e-01 8.25064838e-01 -7.03768492e-01
-4.23808664e-01 -1.15651762e+00 2.22896427e-01 -9.85256910e-01
7.22919166e-01 -4.56986666e-01 -8.14234793e-01 5.66960633e-01
7.02985004e-02 -1.13087222e-02 8.49747777e-01 7.52064362e-02
-5.01800179e-01 1.82647016e-02 -1.59135365e+00 3.11388552e-01
4.74027663e-01 -6.36580884e-01 -7.56156802e-01 4.33920503e-01
2.55000114e-01 -3.14960063e-01 -8.89308274e-01 4.10643876e-01
1.33808523e-01 -8.23769510e-01 1.19134176e+00 -8.42119217e-01
-1.66231617e-01 -1.86576117e-02 -2.29802445e-01 -1.41240680e+00
1.76119749e-02 -9.28360820e-01 -3.92934561e-01 1.47270977e+00
5.29498219e-01 -8.82218301e-01 7.35010266e-01 4.50230032e-01
-1.95473135e-02 -4.08088833e-01 -1.12618804e+00 -8.25575590e-01
7.16927409e-01 -5.78311086e-01 2.89261699e-01 6.25615299e-01
-5.27032435e-01 6.52780384e-02 -3.90062660e-01 8.33008111e-01
1.07368720e+00 -6.62148893e-01 5.07170439e-01 -1.32070482e+00
-5.61192989e-01 -5.33364832e-01 -5.75049341e-01 -6.78935409e-01
2.35869855e-01 -7.73899257e-01 -7.84164816e-02 -6.35532022e-01
-4.83701557e-01 -5.32300413e-01 -6.09630525e-01 2.12255463e-01
-8.30875486e-02 3.61806840e-01 -8.69360119e-02 4.22438830e-02
-2.23475173e-01 2.61566222e-01 4.83520627e-01 -1.26928940e-01
2.11202890e-01 6.96299255e-01 -4.09645796e-01 9.87326503e-01
1.02431643e+00 -8.81816804e-01 -4.04574364e-01 1.79666863e-03
4.24297303e-02 -2.71341175e-01 7.16219783e-01 -1.31097150e+00
-1.95418787e-03 3.79653603e-01 8.47142100e-01 -5.55882215e-01
4.90815669e-01 -8.01255345e-01 -1.28468335e-01 3.41796547e-01
-4.24209476e-01 -2.53422111e-01 2.91750848e-01 9.17650878e-01
-1.88282300e-02 -7.52291620e-01 1.12296975e+00 -1.76404696e-02
-3.05900633e-01 1.08051464e-01 -6.56318784e-01 3.26714188e-01
1.15900445e+00 7.76049346e-02 -2.48851299e-01 -2.92963803e-01
-1.33196294e+00 -7.76808197e-03 -1.32867441e-01 -4.00908384e-03
5.74228585e-01 -1.02785110e+00 -8.15165639e-01 3.68125916e-01
-5.24846911e-01 -3.10597211e-01 1.07794710e-01 4.72604394e-01
-2.25198045e-01 7.28499815e-02 -1.22315906e-01 -4.33063298e-01
-1.29915261e+00 4.82575744e-01 6.51984870e-01 -4.80239600e-01
-1.46847963e-01 1.26793790e+00 2.51863487e-02 -3.20086360e-01
7.63598025e-01 2.45362565e-01 -2.17927337e-01 -4.59726639e-02
4.75228310e-01 3.11870724e-01 -1.94038812e-03 -2.15425432e-01
-8.42969120e-02 -1.03017591e-01 -2.55136162e-01 -3.22056055e-01
1.01802409e+00 3.90528500e-01 2.58980185e-01 4.10126783e-02
9.46596801e-01 4.79075722e-02 -1.29408908e+00 3.70627269e-02
-2.46821847e-02 -3.10212135e-01 1.63008235e-02 -1.02266204e+00
-7.07722545e-01 1.37278104e+00 8.94725084e-01 3.31149906e-01
9.04823542e-01 -4.16492552e-01 2.98394114e-01 4.91371632e-01
1.37400866e-01 -9.90767598e-01 1.68924004e-01 5.86206138e-01
5.41959703e-01 -8.64252985e-01 -1.06392533e-01 -3.31553556e-02
-4.16048229e-01 1.20689774e+00 5.45213521e-01 -5.32280982e-01
1.00371218e+00 6.39617085e-01 -2.42050767e-01 1.31765574e-01
-4.88885909e-01 1.49225667e-01 3.44204485e-01 8.75662565e-01
-6.95047081e-02 -2.85935234e-02 3.26815337e-01 6.79612398e-01
-8.76390785e-02 -5.66142797e-01 3.15339953e-01 7.90041804e-01
-4.93299246e-01 -1.14916563e+00 -7.94607460e-01 4.67993826e-01
-1.17396247e+00 -2.47632384e-01 -3.92843872e-01 7.40952492e-01
2.51318902e-01 9.67487514e-01 -1.68584771e-02 -2.44390249e-01
1.84265777e-01 4.74109769e-01 4.12262172e-01 -4.81411785e-01
-8.13133776e-01 -2.48538136e-01 2.82934695e-01 1.79792911e-01
5.05220555e-02 -4.71910298e-01 -6.79467082e-01 -1.62426203e-01
-6.85929179e-01 4.17228550e-01 5.48429012e-01 8.78840327e-01
8.52462426e-02 6.51526392e-01 6.40962005e-01 -8.88063669e-01
-1.62995768e+00 -1.09327865e+00 -5.81729949e-01 4.16258484e-01
1.88037291e-01 -7.30960548e-01 -1.16254139e+00 -1.54165238e-01] | [5.735271453857422, 7.700480937957764] |
5aa4bccd-dea3-4d4d-9049-0d05ecad174b | a-neural-pairwise-ranking-model-for | 2203.07450 | null | https://arxiv.org/abs/2203.07450v1 | https://arxiv.org/pdf/2203.07450v1.pdf | A Neural Pairwise Ranking Model for Readability Assessment | Automatic Readability Assessment (ARA), the task of assigning a reading level to a text, is traditionally treated as a classification problem in NLP research. In this paper, we propose the first neural, pairwise ranking approach to ARA and compare it with existing classification, regression, and (non-neural) ranking methods. We establish the performance of our model by conducting experiments with three English, one French and one Spanish datasets. We demonstrate that our approach performs well in monolingual single/cross corpus testing scenarios and achieves a zero-shot cross-lingual ranking accuracy of over 80% for both French and Spanish when trained on English data. Additionally, we also release a new parallel bilingual readability dataset in English and French. To our knowledge, this paper proposes the first neural pairwise ranking model for ARA, and shows the first results of cross-lingual, zero-shot evaluation of ARA with neural models. | ['Sowmya Vajjala', 'Justin Lee'] | 2022-03-14 | null | https://aclanthology.org/2022.findings-acl.300 | https://aclanthology.org/2022.findings-acl.300.pdf | findings-acl-2022-5 | ['cross-corpus'] | ['computer-vision'] | [ 2.89338946e-01 6.02648668e-02 -1.23968706e-01 -5.72179556e-01
-1.52670455e+00 -6.58723056e-01 9.24538493e-01 5.99124968e-01
-6.68105304e-01 9.39702749e-01 6.45722449e-01 -3.69545221e-01
-4.46844846e-01 -7.10072875e-01 -6.93773031e-01 -1.50689960e-01
2.56560832e-01 9.70057964e-01 -1.87482461e-01 -2.77623475e-01
4.03462499e-01 4.02567498e-02 -1.82841837e+00 4.50791180e-01
1.39339674e+00 6.56046152e-01 6.38085157e-02 6.80871487e-01
2.98863798e-01 1.07750332e+00 -4.12761867e-01 -6.86768830e-01
-5.66233764e-04 -2.69691646e-01 -1.36363375e+00 -6.60695553e-01
1.22679126e+00 -2.73979574e-01 -2.45605186e-02 8.21698725e-01
7.53764749e-01 3.24258804e-01 1.00421488e+00 -7.96666503e-01
-1.08455968e+00 8.92209530e-01 -1.39605254e-01 2.46088922e-01
6.90095842e-01 -2.98307657e-01 1.54846799e+00 -1.05185091e+00
5.59067547e-01 1.02276194e+00 7.22099781e-01 4.06282991e-01
-1.09471583e+00 -2.61181086e-01 -2.69900799e-01 2.76718795e-01
-9.62353170e-01 -4.67378199e-01 4.23868001e-01 -5.31060159e-01
1.19006097e+00 2.66576111e-01 2.38321826e-01 1.18033504e+00
4.40736115e-02 8.39281023e-01 1.85191190e+00 -9.49289680e-01
-1.37828793e-02 -6.56495616e-02 8.02677512e-01 6.77267015e-01
4.51055616e-02 1.85704380e-01 -8.03729951e-01 2.01182887e-02
8.07430744e-02 -6.67400956e-01 -3.99778366e-01 -1.46119803e-01
-1.40522170e+00 6.98707581e-01 7.53413215e-02 4.80193585e-01
4.23805974e-02 -2.36227497e-01 5.74618816e-01 8.80147934e-01
8.47663462e-01 6.89326644e-01 -7.63561487e-01 -4.00184542e-01
-1.16747820e+00 -4.63065878e-02 1.05484366e+00 8.12062323e-01
4.72313970e-01 -5.02832472e-01 -7.00662255e-01 1.38334584e+00
-8.14073607e-02 5.93903601e-01 8.39445472e-01 -6.92108929e-01
6.80357397e-01 4.27747905e-01 -3.42476219e-02 -7.42725849e-01
-6.00851357e-01 -3.80257815e-01 -5.74669778e-01 -7.19557633e-04
6.01902783e-01 2.88024426e-01 -4.63932544e-01 1.83252919e+00
-4.91136789e-01 -4.34657276e-01 1.99472412e-01 4.71359342e-01
1.19502485e+00 5.43316066e-01 -1.42789796e-01 -3.93303990e-01
1.30613995e+00 -1.17576694e+00 -5.08185983e-01 1.10096931e-01
7.59363592e-01 -6.48253322e-01 1.63998318e+00 6.09755576e-01
-1.32041287e+00 -7.56016433e-01 -9.72067952e-01 -5.34010053e-01
-5.44127941e-01 7.05692112e-01 1.36311844e-01 6.30712986e-01
-1.40941191e+00 5.46481311e-01 -3.14416856e-01 -6.31133378e-01
-4.33340184e-02 2.74150401e-01 -5.58964252e-01 -2.21867606e-01
-1.39812768e+00 1.25433493e+00 1.59731969e-01 -1.23617910e-01
-7.19595373e-01 -4.37747389e-01 -8.06659877e-01 1.65903661e-02
-1.04321130e-01 -3.52286339e-01 1.25819635e+00 -1.26400769e+00
-1.55928636e+00 1.45927560e+00 -1.49677083e-01 -9.66263786e-02
4.51469243e-01 -4.84877378e-01 -5.14676094e-01 -3.06109130e-01
3.43275756e-01 5.63483119e-01 6.55132458e-02 -8.23746383e-01
-5.54670930e-01 -3.50568116e-01 3.51689547e-01 4.51829582e-01
-5.98962605e-01 1.36650190e-01 -4.08363752e-02 -5.04236221e-01
-4.30751383e-01 -6.73053384e-01 4.24521893e-01 -5.90061247e-01
-2.93881744e-02 -7.61938691e-01 -3.41562092e-01 -1.10989761e+00
1.23680890e+00 -1.74466717e+00 3.28286141e-01 6.40025884e-02
-1.81191154e-02 3.38398404e-02 -5.46347499e-01 3.14028263e-01
-1.34498533e-03 6.22879528e-02 -8.68951231e-02 -2.32097298e-01
2.47944579e-01 -3.06316584e-01 1.58803761e-01 3.24727327e-01
-2.25562789e-02 9.66627717e-01 -1.16284418e+00 -4.16274279e-01
-1.55694306e-01 2.44226992e-01 -3.32319885e-01 2.12365359e-01
7.76812285e-02 6.00060895e-02 1.43236578e-01 4.62623417e-01
3.22645605e-01 -9.85806510e-02 2.11644053e-01 -5.17269410e-02
-2.40040779e-01 6.85489833e-01 -5.15521169e-01 1.89298844e+00
-8.63270581e-01 1.04976583e+00 -5.25980890e-01 -7.35766947e-01
9.53382730e-01 2.68515080e-01 -1.26364872e-01 -1.32790375e+00
6.30239695e-02 4.84516054e-01 6.74239686e-03 -4.11400348e-01
7.42520213e-01 2.45665967e-01 -3.75908941e-01 7.48930395e-01
2.93235004e-01 1.38772115e-01 7.46877193e-01 -7.90845379e-02
1.12603450e+00 4.48551416e-01 4.25985008e-01 -8.13435733e-01
7.99805820e-01 -1.13845348e-01 4.52669375e-02 9.27450240e-01
-2.62195170e-01 6.89948261e-01 4.78080839e-01 -3.38322461e-01
-1.16632128e+00 -1.16256535e+00 -4.11925286e-01 1.67463255e+00
-2.98923671e-01 -3.59292060e-01 -9.66937184e-01 -7.24515021e-01
-4.11631972e-01 8.80579412e-01 -6.50007427e-01 8.36742984e-04
-5.60838878e-01 -7.51469493e-01 5.79182506e-01 4.51016515e-01
4.77296621e-01 -1.13303280e+00 -1.92474991e-01 1.74987420e-01
-7.10665405e-01 -9.28548753e-01 -2.43311614e-01 2.55877405e-01
-5.65405071e-01 -1.12000215e+00 -6.17864847e-01 -1.37328243e+00
3.76276463e-01 -7.56315440e-02 1.92753935e+00 2.85638799e-03
8.41751695e-02 4.82560873e-01 -5.45785487e-01 -1.96243878e-02
-5.10185182e-01 4.80042070e-01 1.00535467e-01 -5.68741679e-01
7.87616909e-01 -1.74755588e-01 -1.32459044e-01 1.27823398e-01
-4.55483347e-01 1.48988232e-01 5.30067265e-01 1.13918233e+00
4.89780337e-01 -2.55318016e-01 7.32350469e-01 -8.47836614e-01
1.23683333e+00 -1.99872807e-01 -4.17340815e-01 7.64415503e-01
-7.22396433e-01 3.06247827e-02 6.95418656e-01 -7.17957169e-02
-9.73852932e-01 -2.99196690e-01 -8.82070586e-02 4.26882774e-01
-2.65719712e-01 8.48228097e-01 -8.88849348e-02 2.43456095e-01
9.95091736e-01 9.11875069e-02 -3.16311836e-01 -4.71624345e-01
4.77836192e-01 9.77206707e-01 5.25139630e-01 -8.31163168e-01
4.08791333e-01 -2.61095494e-01 -3.59937459e-01 -4.98814136e-01
-1.18483198e+00 -3.87628049e-01 -1.20056736e+00 -3.71831477e-01
7.40403771e-01 -1.20594454e+00 -4.28848058e-01 5.61799228e-01
-1.04901206e+00 -5.11284471e-01 -7.17311129e-02 4.67077762e-01
-7.95783639e-01 1.63087755e-01 -7.01891601e-01 -4.64541674e-01
-4.97364134e-01 -9.81167197e-01 1.00914454e+00 -1.01400271e-01
-5.27228951e-01 -1.14566386e+00 5.87446451e-01 7.41289198e-01
4.16579872e-01 -1.11476839e-01 1.19110501e+00 -9.29680467e-01
-2.65749432e-02 3.84583673e-03 -3.42272997e-01 4.25225496e-01
-1.16606228e-01 -2.40037799e-01 -1.07691801e+00 -5.29962182e-01
-2.98372507e-01 -1.09080863e+00 8.60017300e-01 2.26218969e-01
1.01245892e+00 -1.11216463e-01 2.58849829e-01 1.12971701e-01
1.30798256e+00 -1.02469154e-01 6.20975435e-01 7.01265097e-01
5.63949168e-01 8.16662192e-01 8.09754729e-01 -1.25508025e-01
6.57379270e-01 5.69285452e-01 -1.93059459e-01 -1.21102750e-01
-2.20771790e-01 -5.50432317e-02 6.73014641e-01 1.48158813e+00
-3.79544795e-01 -5.58732212e-01 -9.95666564e-01 5.19395590e-01
-1.67095172e+00 -1.04008329e+00 -1.81761950e-01 2.43454814e+00
1.16145492e+00 9.13813990e-03 5.10265818e-03 1.11695476e-01
6.14691794e-01 -1.03684485e-01 4.62623164e-02 -7.22269654e-01
-3.43524307e-01 5.00608921e-01 9.24129039e-02 7.63945401e-01
-1.32168329e+00 1.01061785e+00 6.53413868e+00 8.36277008e-01
-6.08276129e-01 3.78307849e-01 8.20481122e-01 -8.43963400e-02
-3.27147067e-01 -4.02896374e-01 -6.18288636e-01 2.05368936e-01
1.32329595e+00 -6.15368830e-03 6.72628582e-01 3.85862529e-01
-3.37303663e-03 -4.51936573e-02 -1.50281751e+00 7.12993860e-01
7.27421463e-01 -5.20198047e-01 5.59907872e-04 -2.25343987e-01
8.92341018e-01 3.76476735e-01 -9.89316776e-02 6.54632747e-01
5.06658196e-01 -1.42363203e+00 6.10801458e-01 5.74223340e-01
1.21167195e+00 -8.98895442e-01 9.99672294e-01 2.41058543e-01
-7.71160185e-01 -1.10324312e-04 -4.22847837e-01 -1.98028564e-01
-3.33166569e-01 3.88844877e-01 -3.46681744e-01 4.18060601e-01
7.92433202e-01 8.45186412e-01 -1.05564916e+00 8.35110188e-01
-3.66365671e-01 5.87943792e-01 1.82956204e-01 -3.93892109e-01
-5.85864224e-02 -7.54229948e-02 2.16067985e-01 1.34937096e+00
3.82863075e-01 -2.80060768e-01 6.53246865e-02 4.82584417e-01
-4.46574748e-01 6.22077167e-01 -5.75529099e-01 1.06791601e-01
2.41517767e-01 1.10064828e+00 -4.17759448e-01 -2.49322474e-01
-5.29979289e-01 9.80896175e-01 1.22634959e+00 1.72885567e-01
-5.86357594e-01 -6.30726278e-01 -1.41905487e-01 -3.09906453e-01
-3.47956985e-01 -5.29339025e-03 -2.48500079e-01 -1.38424504e+00
1.88477203e-01 -1.16242349e+00 3.08322072e-01 -8.89770985e-01
-1.68902290e+00 7.88009286e-01 -1.41659066e-01 -1.17955291e+00
-2.73667425e-01 -1.10863566e+00 -2.50055850e-01 9.69864070e-01
-1.59284985e+00 -1.13986349e+00 -4.57874864e-01 2.56308258e-01
6.98304892e-01 -6.81974292e-01 1.00137484e+00 2.36528173e-01
-4.11019653e-01 7.89870441e-01 7.23118067e-01 2.31104389e-01
1.19669628e+00 -1.74787140e+00 2.25398809e-01 6.48673117e-01
3.07287663e-01 5.95796168e-01 4.02272046e-01 -4.29085553e-01
-7.88244486e-01 -8.77357721e-01 1.77175641e+00 -8.03861678e-01
7.10260808e-01 -2.45384052e-01 -6.40542090e-01 5.04843116e-01
7.64292002e-01 -7.91055918e-01 9.70932186e-01 7.53297865e-01
-5.05123854e-01 -7.04497844e-02 -9.47237074e-01 4.11647379e-01
1.08298826e+00 -9.34986293e-01 -1.04415560e+00 6.85741365e-01
4.32497203e-01 -2.20396250e-01 -1.21092343e+00 3.91362637e-01
8.01898777e-01 -1.04052830e+00 8.52852762e-01 -3.93721908e-01
9.58836854e-01 2.90245172e-02 -2.49577612e-01 -1.76397336e+00
-5.60874045e-01 -1.47573296e-02 4.45255667e-01 1.31155050e+00
7.49674737e-01 -3.33878845e-01 1.97571635e-01 1.50980115e-01
-2.52330333e-01 -4.52604115e-01 -1.00335836e+00 -7.89307296e-01
8.13538253e-01 -1.06979288e-01 3.81398410e-01 1.22642827e+00
3.02428216e-01 6.83681667e-01 -2.93544263e-01 -2.67383665e-01
3.67469162e-01 -1.04110472e-01 4.46934491e-01 -1.60212779e+00
-2.74874657e-01 -8.74768436e-01 -1.33759171e-01 -5.95006883e-01
6.64773107e-01 -1.42654157e+00 2.99851179e-01 -1.73499799e+00
8.31911743e-01 -5.98441102e-02 -5.52859426e-01 3.87019604e-01
-2.51712799e-01 3.58840287e-01 -2.48682231e-01 1.05049364e-01
-8.11406314e-01 2.64232814e-01 8.35294127e-01 -4.21533972e-01
8.74020979e-02 -4.05971140e-01 -5.60443878e-01 3.08158249e-01
9.24498022e-01 -1.79223582e-01 -4.00947064e-01 -7.73639858e-01
5.83598256e-01 -1.35222897e-01 1.45315221e-02 -1.07816601e+00
1.19769998e-01 2.12309316e-01 4.09894913e-01 -3.67865264e-01
-9.12614092e-02 -3.57601732e-01 -4.31026459e-01 2.00105518e-01
-9.36701000e-01 6.05232060e-01 -5.17298207e-02 1.07384078e-01
-2.91604906e-01 -3.80391657e-01 6.57388389e-01 -8.88890475e-02
-5.65390468e-01 -1.17878482e-01 -2.72921324e-01 3.86670321e-01
5.19183517e-01 1.48303717e-01 -8.16620886e-01 -9.52682495e-02
-5.85239828e-01 2.62038231e-01 5.50241292e-01 7.27116108e-01
2.99109161e-01 -1.39609480e+00 -1.28478754e+00 1.79952048e-02
6.39760375e-01 -6.28195167e-01 -6.26036450e-02 7.35751331e-01
-6.75348401e-01 7.00519741e-01 -3.62226874e-01 -3.63518506e-01
-1.39058971e+00 2.62862086e-01 3.88394654e-01 -6.52788281e-01
-7.20616132e-02 4.92415607e-01 -1.54680476e-01 -1.11015177e+00
3.05739373e-01 5.44734113e-02 -9.02059972e-01 3.02215397e-01
5.81509471e-01 7.29146063e-01 4.20651227e-01 -8.35412085e-01
1.47634093e-03 7.28547394e-01 -1.79098338e-01 -3.93799275e-01
1.17872024e+00 -9.49103478e-03 -5.72495759e-01 9.45362747e-01
1.20764148e+00 1.93590164e-01 -3.30525160e-01 -2.05057308e-01
2.55613089e-01 -7.34556392e-02 1.36377975e-01 -1.50092542e+00
-4.36722040e-01 9.77567732e-01 7.41821051e-01 -2.91083008e-02
9.84345973e-01 -1.83782965e-01 3.10747176e-01 7.39613891e-01
3.96535218e-01 -1.59592068e+00 1.05559371e-01 1.07737994e+00
9.59348321e-01 -1.42074382e+00 -2.37559423e-01 5.09972572e-02
-3.49310696e-01 1.25435889e+00 7.91948736e-01 1.81385979e-01
2.42050454e-01 -2.12313056e-01 1.04934901e-01 -5.70976594e-03
-7.40250230e-01 -1.48070469e-01 8.36040497e-01 4.47313517e-01
1.44294989e+00 4.02654529e-01 -8.71337891e-01 4.26809877e-01
-5.97394288e-01 1.84086673e-02 6.12284780e-01 6.99657023e-01
-4.35367763e-01 -1.25452518e+00 7.21392706e-02 6.69458389e-01
-5.35324991e-01 -7.43917704e-01 -5.92310846e-01 7.87417233e-01
-7.17386082e-02 1.13289440e+00 1.58221394e-01 -3.40497017e-01
2.88062245e-01 2.95006305e-01 6.15709901e-01 -5.50377905e-01
-8.88174057e-01 -4.81571168e-01 5.62743843e-01 -3.29344183e-01
-5.99912465e-01 -8.60133052e-01 -5.40693104e-01 -2.54969507e-01
-2.09378496e-01 2.51064450e-01 3.57002467e-01 9.29006994e-01
-1.03357106e-01 3.86402577e-01 5.13723433e-01 -3.76573235e-01
-5.80462992e-01 -1.32867634e+00 -3.75825882e-01 5.60461462e-01
1.41626447e-01 -4.22786266e-01 -5.16317248e-01 -2.83959419e-01] | [11.04572582244873, 10.137849807739258] |
ab8e8880-ac3b-4fab-ba3c-ae72635bc867 | an-improved-baseline-framework-for-pose | 2303.07141 | null | https://arxiv.org/abs/2303.07141v1 | https://arxiv.org/pdf/2303.07141v1.pdf | An Improved Baseline Framework for Pose Estimation Challenge at ECCV 2022 Visual Perception for Navigation in Human Environments Workshop | This technical report describes our first-place solution to the pose estimation challenge at ECCV 2022 Visual Perception for Navigation in Human Environments Workshop. In this challenge, we aim to estimate human poses from in-the-wild stitched panoramic images. Our method is built based on Faster R-CNN for human detection, and HRNet for human pose estimation. We describe technical details for the JRDB-Pose dataset, together with some experimental results. In the competition, we achieved 0.303 $\text{OSPA}_{\text{IOU}}$ and 64.047\% $\text{AP}_{\text{0.5}}$ on the test set of JRDB-Pose. | ['Jianqin Yin', 'Wending Zhao', 'Feng Zhou', 'Shaojie Zhang', 'Ruoqi Yin', 'Yonghao Dang', 'Jiajun Fu'] | 2023-03-13 | null | null | null | null | ['human-detection'] | ['computer-vision'] | [-1.20410673e-01 7.42969364e-02 3.68706584e-01 -3.31160367e-01
-9.49723780e-01 -3.85813534e-01 1.08152144e-01 -4.09798652e-01
-1.07165360e+00 6.00993454e-01 -3.61393504e-02 -2.79544234e-01
1.78765386e-01 -4.31814522e-01 -8.41538668e-01 -2.43530273e-01
-3.32817882e-01 3.99479538e-01 3.96859974e-01 -5.41744590e-01
-1.65811986e-01 3.44804972e-01 -1.39727104e+00 3.23573798e-02
5.21918386e-02 1.40387893e+00 9.70228016e-02 9.76703227e-01
7.95478523e-01 5.63857913e-01 -8.18262160e-01 -5.66397250e-01
5.41083574e-01 9.49202925e-02 -5.92199802e-01 -3.58119100e-01
1.08763087e+00 -6.22651756e-01 -9.33509111e-01 8.99838150e-01
1.04856849e+00 2.27448776e-01 5.21373868e-01 -1.10448182e+00
1.04302280e-01 -5.52233532e-02 -8.76451552e-01 5.08711576e-01
7.19864368e-01 3.23158324e-01 5.84799349e-01 -1.04988539e+00
7.98981249e-01 1.29713356e+00 1.00963891e+00 4.83206958e-01
-5.96069753e-01 -7.72147417e-01 9.86079425e-02 7.98497722e-02
-1.56339943e+00 -2.54272103e-01 8.25466961e-02 -3.80969912e-01
1.08798718e+00 4.91998732e-01 6.88157082e-01 1.43714428e+00
2.82239795e-01 8.98249865e-01 7.97423899e-01 -2.21848771e-01
-6.51667267e-02 -3.12191933e-01 9.41014215e-02 1.03400671e+00
4.03797746e-01 4.57888812e-01 -4.84137982e-01 1.56873651e-02
9.93566930e-01 -4.06883180e-01 -4.13084507e-01 -5.47346652e-01
-1.14808834e+00 6.28058732e-01 8.80399704e-01 -3.21029723e-01
-1.55895412e-01 5.15007019e-01 3.66074562e-01 -3.27747464e-02
-1.07521359e-02 1.92757174e-01 -3.89421195e-01 -1.17398545e-01
-4.66264933e-01 5.61877370e-01 3.73204291e-01 1.38794827e+00
5.73113188e-02 -8.44264850e-02 -9.54780206e-02 7.58209825e-01
4.16894495e-01 9.19106841e-01 -1.21003561e-01 -1.19818568e+00
7.79188573e-01 -7.53745362e-02 6.11431837e-01 -9.90123630e-01
-8.42974305e-01 -6.15823150e-01 -5.63130856e-01 3.87842745e-01
6.45265341e-01 -4.17336851e-01 -1.14926529e+00 1.52721095e+00
1.56091765e-01 -5.83801687e-01 -3.09576005e-01 1.23867023e+00
9.95330334e-01 6.11189187e-01 1.59443691e-01 3.38098943e-01
1.67486215e+00 -1.00702560e+00 -1.73774034e-01 -7.39123285e-01
3.50147337e-01 -7.26396680e-01 8.74374390e-01 5.04853904e-01
-9.64346945e-01 -7.91197538e-01 -1.27269948e+00 -1.29858389e-01
-1.93036333e-01 5.87209225e-01 3.47055495e-01 7.30381072e-01
-1.04908907e+00 -1.66172814e-02 -5.99662364e-01 -5.34543753e-01
1.67176455e-01 2.74329096e-01 -5.18523693e-01 -1.38320148e-01
-1.19537389e+00 1.03714073e+00 3.47258896e-01 4.23807949e-01
-9.21157241e-01 -1.90549672e-01 -1.00765789e+00 -2.98711121e-01
5.64538836e-01 -5.60017049e-01 1.36275315e+00 6.76542521e-02
-9.45586562e-01 1.20363116e+00 1.26287445e-01 -7.25213170e-01
9.06409025e-01 -7.60608256e-01 -5.31465530e-01 1.28287941e-01
2.97039390e-01 1.06584013e+00 4.36155528e-01 -1.24860501e+00
-9.16966438e-01 -7.45145321e-01 3.12579870e-02 3.32745850e-01
2.82798797e-01 4.70530428e-02 -1.22816098e+00 -4.96715188e-01
4.19826359e-01 -1.14830232e+00 -5.53499877e-01 9.42238793e-02
-5.70717871e-01 2.35744774e-01 4.85342979e-01 -9.11977768e-01
9.42793310e-01 -1.68334484e+00 1.22553995e-02 4.28762078e-01
3.22300941e-01 1.86883882e-01 2.66824424e-01 4.28736843e-02
2.37674832e-01 -4.29422945e-01 2.10096389e-01 -4.10946548e-01
4.97022457e-02 -3.66483182e-01 -1.26589224e-01 6.57306135e-01
-6.93582892e-01 8.27537775e-01 -5.64542711e-01 -3.82209778e-01
3.70059818e-01 5.47036231e-01 -2.39648849e-01 6.55684173e-02
7.66776949e-02 3.30161095e-01 -3.47087145e-01 7.53773153e-01
7.64934599e-01 1.16058914e-02 -1.14533670e-01 -3.96922112e-01
-2.22120613e-01 -2.18389541e-01 -1.27507615e+00 1.78440762e+00
-2.79330090e-02 8.04893672e-01 2.67497510e-01 -2.90948540e-01
7.31612980e-01 -6.45158067e-02 2.59737074e-01 -1.00249219e+00
5.09055495e-01 1.29097089e-01 -2.11811706e-01 -9.14964676e-02
8.49830925e-01 4.36477214e-01 -4.94888306e-01 -1.74840108e-01
5.18662073e-02 -5.93767054e-02 1.85271397e-01 -8.01389851e-03
1.18629348e+00 4.05421257e-01 1.93580151e-01 -2.22234756e-01
3.99109453e-01 6.37057528e-04 3.76288086e-01 1.11641526e+00
-7.13600516e-01 8.89735222e-01 2.74644464e-01 -8.16869736e-01
-1.18374491e+00 -1.52042031e+00 -1.50793105e-01 1.14032006e+00
4.06506538e-01 -4.44578201e-01 -8.09679329e-01 -5.61421752e-01
-1.81155980e-01 4.05422419e-01 -7.69836307e-01 6.47445992e-02
-7.12505043e-01 -4.10864830e-01 9.96901095e-01 9.21836972e-01
8.08031857e-01 -1.09250128e+00 -1.36954427e+00 -7.94551894e-02
-4.35895562e-01 -1.27567792e+00 -5.17829776e-01 1.70296654e-01
-1.87736169e-01 -1.08965683e+00 -1.16651821e+00 -8.18813741e-01
4.09403414e-01 2.52219766e-01 9.70471084e-01 -4.21618462e-01
-8.65831256e-01 4.84839350e-01 -1.34981126e-01 -5.78255475e-01
4.56142783e-01 -1.43473148e-01 3.07177693e-01 -5.77076793e-01
3.55480552e-01 6.53690994e-02 -1.09252572e+00 8.13493669e-01
-7.39084184e-02 -2.03779191e-01 4.46252942e-01 5.71570277e-01
6.73916638e-01 -5.53160548e-01 -2.71995217e-01 -2.69943357e-01
9.05889347e-02 2.55220383e-01 -9.75430369e-01 9.37180892e-02
-1.70417875e-01 -3.32699567e-01 7.76261389e-02 5.39483912e-02
-8.84873629e-01 3.20465147e-01 -4.66483563e-01 -4.31475937e-01
-2.16316327e-01 -1.55470446e-02 -4.75149788e-03 -5.57982586e-02
1.12895131e+00 1.99338436e-01 -1.88638300e-01 -3.36275041e-01
2.23873064e-01 4.34213519e-01 1.27170312e+00 -3.21189463e-01
7.03751147e-01 6.61192596e-01 -8.44318867e-02 -8.40436637e-01
-7.64687002e-01 -6.72231317e-01 -5.40982604e-01 -3.97672474e-01
1.33345902e+00 -1.41515326e+00 -1.10785532e+00 4.27819103e-01
-1.00423586e+00 -2.86939204e-01 3.40664387e-02 6.37869716e-01
-6.82568073e-01 2.80338645e-01 -5.91210008e-01 -8.41521204e-01
-5.08199453e-01 -1.14116406e+00 1.32332873e+00 3.43291402e-01
-3.25715154e-01 -2.26149619e-01 3.95868346e-02 7.71821916e-01
3.72354001e-01 3.46456319e-01 -1.43902108e-01 -9.55250636e-02
-4.64460313e-01 -7.36373961e-01 -4.34993535e-01 3.56115177e-02
-5.39391875e-01 -6.37129903e-01 -8.60290945e-01 -6.56107545e-01
-4.46693718e-01 -6.11824453e-01 9.62078154e-01 5.96441686e-01
1.04779387e+00 1.43674433e-01 -5.14936745e-01 6.33122623e-01
9.71563756e-01 3.82680178e-01 1.00236845e+00 5.72078228e-01
7.32984841e-01 3.61443162e-01 8.53123069e-01 5.78059912e-01
4.32891756e-01 1.08158159e+00 5.44161022e-01 -3.02541722e-02
-3.79472822e-02 -2.67733693e-01 1.91628844e-01 -3.54514122e-01
-4.14603800e-01 -3.12924087e-01 -1.10444403e+00 1.92757711e-01
-1.45240939e+00 -8.65643203e-01 -1.89789087e-01 2.10483074e+00
2.64367849e-01 5.14379621e-01 2.47825921e-01 -2.66521394e-01
5.95798075e-01 2.39796087e-01 -1.25098839e-01 7.59028494e-02
1.17360931e-02 2.72191495e-01 8.33876193e-01 5.54238558e-01
-1.56718242e+00 1.05805826e+00 5.59514093e+00 5.68473637e-01
-7.77989149e-01 -1.21216364e-01 4.67257708e-01 -1.51598021e-01
5.89328110e-01 -3.85713577e-01 -1.16662729e+00 1.74725965e-01
5.21576345e-01 4.90127325e-01 -1.49593055e-01 1.27593601e+00
-2.16941059e-01 -5.60797572e-01 -8.75478387e-01 1.24419343e+00
2.30002940e-01 -9.60017204e-01 -5.01973152e-01 1.37823522e-01
2.79509544e-01 1.73814923e-01 3.18991929e-01 5.87753296e-01
1.99901983e-01 -1.30889845e+00 1.00547016e+00 1.84212402e-01
1.04305720e+00 -8.07191432e-01 8.64865899e-01 2.79169708e-01
-1.52635872e+00 -9.93571654e-02 -4.78529453e-01 9.45785567e-02
3.18822742e-01 3.63512933e-02 -6.08427823e-01 4.30500835e-01
1.53784716e+00 1.10610500e-01 -7.97849178e-01 1.14596200e+00
-2.53564417e-01 -1.51752234e-01 -3.92471761e-01 -1.25610694e-01
-3.35032307e-02 3.27694118e-01 5.93390822e-01 1.24991989e+00
2.52421916e-01 2.86870331e-01 2.01183781e-01 3.19060653e-01
1.20256782e-01 -2.80516714e-01 -4.93497819e-01 7.97275484e-01
1.64538413e-01 9.91195202e-01 -7.70342112e-01 -2.43656352e-01
1.09887607e-01 1.06778228e+00 1.47015914e-01 2.37746745e-01
-1.18960750e+00 -5.80625951e-01 5.56955397e-01 3.19957376e-01
5.43244004e-01 -4.87734407e-01 -1.79369405e-01 -8.67127836e-01
3.03563595e-01 -6.65189743e-01 4.24263835e-01 -1.08196700e+00
-7.98868835e-01 9.77239847e-01 2.58893043e-01 -1.36435378e+00
-1.65194914e-01 -1.03466809e+00 6.28389865e-02 8.06052387e-01
-4.39302117e-01 -1.08322620e+00 -8.07365000e-01 5.96337795e-01
2.99286842e-01 -3.07272673e-01 7.33295918e-01 3.85342181e-01
-1.78528756e-01 9.90706921e-01 -2.58387566e-01 4.57166165e-01
6.55092239e-01 -9.11132216e-01 9.32134032e-01 7.53916800e-01
-2.04446450e-01 5.35632253e-01 1.01209486e+00 -6.39693260e-01
-1.21596646e+00 -7.55534410e-01 6.27272248e-01 -1.00915408e+00
8.74264389e-02 -6.07356131e-01 -8.89363885e-02 8.51920605e-01
5.87329566e-02 1.85413167e-01 2.41743084e-02 -1.95705798e-02
-3.99907112e-01 -5.47404587e-02 -1.20231462e+00 9.38529968e-01
1.34886169e+00 -1.50190800e-01 -4.18526530e-01 1.72698349e-01
4.70878512e-01 -1.10560834e+00 -5.45807958e-01 6.14674032e-01
9.86486077e-01 -9.99471068e-01 1.53163981e+00 -1.66320488e-01
2.89148856e-02 -4.12184447e-01 -4.86824483e-01 -6.81520045e-01
-3.11520606e-01 -1.22741997e-01 8.30743313e-02 2.08533525e-01
3.87394458e-01 -2.82153994e-01 1.18993485e+00 2.12761626e-01
-8.57188925e-02 -3.99825305e-01 -1.33579183e+00 -7.48065531e-01
-2.96696544e-01 -5.41325390e-01 -2.19097093e-01 1.32914081e-01
-4.44197774e-01 1.77592114e-01 -7.95296013e-01 2.68869698e-01
1.11380148e+00 -6.35740697e-01 1.13468218e+00 -8.61734986e-01
-4.81403559e-01 -1.11765355e-01 -6.80761695e-01 -1.32482934e+00
-6.50612772e-01 3.95699367e-02 3.60937059e-01 -1.42479563e+00
-1.01536475e-01 -8.77178833e-02 3.30853160e-03 3.17718267e-01
4.53971177e-02 5.51169753e-01 3.23199183e-01 -3.39461043e-02
-8.68866563e-01 2.43511468e-01 8.30144107e-01 -6.05784766e-02
1.50478333e-01 2.22590461e-01 -2.49250159e-01 6.76819921e-01
7.04877257e-01 -1.04426332e-01 -1.75002694e-01 -2.91344404e-01
7.27316514e-02 3.67542714e-01 9.06408846e-01 -1.61519420e+00
2.22020045e-01 2.13508055e-01 1.01864982e+00 -1.30056894e+00
9.20050681e-01 -5.70800364e-01 7.24575520e-02 9.69331920e-01
-1.04469173e-01 3.01739275e-01 4.42977011e-01 4.95370656e-01
2.55124420e-01 2.89410561e-01 9.68681216e-01 -4.09051269e-01
-1.13206744e+00 2.03788936e-01 -1.87547609e-01 1.73585281e-01
1.05551541e+00 -2.66785532e-01 -5.04581571e-01 -6.33959949e-01
-8.19917858e-01 4.58703041e-01 2.06414849e-01 7.60759354e-01
7.37200737e-01 -1.20566583e+00 -5.68762541e-01 8.01980682e-03
4.47386742e-01 -5.36992634e-03 4.83803391e-01 6.32478893e-01
-1.10030568e+00 8.18110645e-01 -5.75332165e-01 -6.95170820e-01
-1.36083269e+00 2.69116491e-01 6.16933167e-01 -1.74190253e-01
-4.97496068e-01 1.47976089e+00 4.23017796e-03 -4.56264436e-01
7.37552106e-01 5.92376031e-02 1.23914033e-02 -2.81503856e-01
5.40927947e-01 6.23855412e-01 -3.70024219e-02 -8.63353193e-01
-8.10393274e-01 5.30893624e-01 -9.04401094e-02 -6.86964869e-01
7.85009980e-01 -1.83223456e-01 4.44058716e-01 -1.25306115e-01
1.41243374e+00 -7.95490742e-02 -1.53899479e+00 -9.56366770e-03
-2.67872900e-01 -4.10526186e-01 -4.05401170e-01 -1.06708038e+00
-7.11994827e-01 7.82926202e-01 1.38516736e+00 -5.68426609e-01
5.85736215e-01 1.69716939e-01 6.19253695e-01 6.48400664e-01
1.02221525e+00 -1.37383807e+00 3.47427517e-01 7.79519141e-01
1.05277050e+00 -1.32075560e+00 1.46260589e-01 -3.92546803e-01
-6.45277619e-01 8.35142732e-01 1.13415790e+00 -1.55756399e-01
4.54673380e-01 4.32868563e-02 2.77927667e-01 -3.95588130e-01
2.51479112e-02 -2.26215854e-01 4.42048520e-01 7.12553024e-01
1.71781316e-01 1.38055980e-01 -7.40314424e-02 4.92006481e-01
-5.99311113e-01 -4.52601045e-01 1.17078759e-01 1.32761586e+00
-6.28283799e-01 -4.63873327e-01 -6.71539187e-01 4.20176983e-02
-2.52696484e-01 6.15177415e-02 -3.88979733e-01 1.05004549e+00
1.86825022e-01 8.13757777e-01 -1.97813481e-01 -7.13551044e-01
8.39738131e-01 -1.85521185e-01 6.83968604e-01 -1.65447727e-01
-4.18494493e-01 1.59291297e-01 3.19710761e-01 -1.06453276e+00
4.16728914e-01 -5.93655646e-01 -1.16654849e+00 -1.82102948e-01
3.15177292e-01 -2.57965595e-01 6.13025308e-01 4.54194069e-01
5.86153800e-03 4.58074540e-01 8.39007124e-02 -1.17417181e+00
-5.46448648e-01 -7.96101570e-01 -4.10192668e-01 2.33094357e-02
1.69475704e-01 -7.61128187e-01 -4.31464165e-02 -3.26020569e-01] | [7.179612636566162, -0.8098330497741699] |
31d9fcf6-418b-44ab-b7d3-a518a2de77b9 | voting-for-deceptive-opinion-spam-detection | 1409.4504 | null | http://arxiv.org/abs/1409.4504v1 | http://arxiv.org/pdf/1409.4504v1.pdf | Voting for Deceptive Opinion Spam Detection | Consumers' purchase decisions are increasingly influenced by user-generated
online reviews. Accordingly, there has been growing concern about the potential
for posting deceptive opinion spam fictitious reviews that have been
deliberately written to sound authentic, to deceive the readers. Existing
approaches mainly focus on developing automatic supervised learning based
methods to help users identify deceptive opinion spams.
This work, we used the LSI and Sprinkled LSI technique to reduce the
dimension for deception detection. We make our contribution to demonstrate what
LSI is capturing in latent semantic space and reveal how deceptive opinions can
be recognized automatically from truthful opinions. Finally, we proposed a
voting scheme which integrates different approaches to further improve the
classification performance. | ['Tao Wang', 'Hua Zhu'] | 2014-09-16 | null | null | null | null | ['deception-detection', 'spam-detection'] | ['miscellaneous', 'natural-language-processing'] | [ 1.13937743e-01 1.59200773e-01 -2.04195101e-02 -7.35211790e-01
-5.12959421e-01 -7.57442296e-01 7.76532531e-01 3.47821377e-02
9.78065655e-02 8.25566471e-01 9.48100090e-02 -4.33137119e-01
1.94958687e-01 -6.09951138e-01 -2.28840709e-01 -6.13708854e-01
4.32950974e-01 -9.50401574e-02 2.66434234e-02 -1.94674030e-01
1.28339183e+00 1.67181924e-01 -1.32870436e+00 5.59116840e-01
1.38814974e+00 9.48479712e-01 -7.91867375e-01 4.92437214e-01
-1.74982718e-03 1.25527310e+00 -1.08940589e+00 -1.33217084e+00
1.77427649e-01 -6.82604730e-01 -5.64977109e-01 3.09654355e-01
6.97904706e-01 -4.34660554e-01 1.49471894e-01 1.64418268e+00
1.60787180e-01 -3.44451964e-02 7.87365317e-01 -1.50849485e+00
-1.53433192e+00 3.25781167e-01 -4.37397778e-01 1.88421533e-01
4.04087692e-01 -6.91584200e-02 8.71720552e-01 -9.69356239e-01
2.55721271e-01 1.77436352e+00 4.84729171e-01 6.94554269e-01
-7.65772939e-01 -7.16252267e-01 1.98441371e-01 5.90528250e-01
-7.53156900e-01 -6.08932376e-01 8.22233975e-01 -1.27116919e-01
5.90329826e-01 3.74257356e-01 6.02011442e-01 1.25397074e+00
6.10912323e-01 1.24663866e+00 1.40644884e+00 -4.17826146e-01
5.03401637e-01 8.98469806e-01 8.21300507e-01 5.50986111e-01
7.52661228e-01 -1.55984340e-02 -6.97435915e-01 -6.62503362e-01
-2.76389837e-01 -1.88813582e-01 3.38888876e-02 -3.79450917e-01
-4.57102023e-02 1.46408796e+00 1.24975361e-01 8.12698454e-02
-7.94728175e-02 -1.57834858e-01 5.95092058e-01 5.55555284e-01
1.11975205e+00 5.35463870e-01 -5.16346693e-01 -7.78324157e-02
-5.63125074e-01 -8.22726637e-02 1.11745965e+00 5.17672122e-01
4.30371344e-01 1.20464854e-01 2.32339635e-01 7.57524610e-01
8.47405255e-01 7.87492633e-01 1.03790903e+00 -6.19603872e-01
9.01252776e-03 9.55359280e-01 4.69971448e-01 -1.87524188e+00
3.16723168e-01 -2.88647175e-01 -2.65492260e-01 2.18695521e-01
-8.53164792e-02 5.57041354e-03 -6.19346678e-01 8.48515034e-01
6.41726032e-02 -3.13736975e-01 2.81767070e-01 1.02988970e+00
3.19295198e-01 4.08283681e-01 1.37518674e-01 -2.77331118e-02
1.16682148e+00 -1.28211164e+00 -9.88614023e-01 -4.47434843e-01
9.49104488e-01 -9.54117894e-01 1.10406208e+00 7.63745964e-01
-8.25677395e-01 -2.08658352e-01 -1.34950435e+00 -2.37107351e-02
-8.03017259e-01 2.40957007e-01 8.31951916e-01 1.41220999e+00
-6.26641631e-01 4.72894877e-01 -2.16224879e-01 -9.25716534e-02
7.43358135e-01 6.60714358e-02 -3.68965678e-02 1.59904547e-03
-1.51196790e+00 1.36337435e+00 -2.08065420e-01 3.21577519e-01
-3.99141163e-01 -1.27591491e-01 -8.20162773e-01 -1.38926774e-01
1.68864988e-02 -3.15082371e-01 1.40586483e+00 -1.80384564e+00
-1.47734189e+00 8.91972303e-01 -3.99398178e-01 -7.00668454e-01
5.42838037e-01 -4.57711428e-01 -9.25678551e-01 1.48978099e-01
2.30469510e-01 2.64893789e-02 1.65559244e+00 -1.31553769e+00
-5.36000907e-01 -9.40147579e-01 -2.53161311e-01 1.85739800e-01
-6.18923604e-01 9.01715606e-02 7.59156048e-01 -4.55789089e-01
1.01847932e-01 -7.83770919e-01 2.95119971e-01 -2.94043124e-01
-1.62425414e-01 -4.13030595e-01 1.42983603e+00 -5.74119508e-01
1.15376532e+00 -1.92987561e+00 -6.11148179e-01 1.78869620e-01
1.53967291e-01 1.05624664e+00 1.83733672e-01 2.94053346e-01
2.21850589e-01 4.93015885e-01 -5.50747570e-03 -1.71150595e-01
2.27174595e-01 5.81638068e-02 -5.86035728e-01 6.85227096e-01
-1.50776953e-01 1.13586211e+00 -1.09935939e+00 -1.27293497e-01
1.86074764e-01 1.06254548e-01 1.38536677e-01 -2.05747351e-01
3.23092282e-01 -4.14499044e-01 -5.99820435e-01 6.91385090e-01
9.04360831e-01 -2.09457666e-01 -1.38722420e-01 -3.73795372e-03
2.08567232e-01 6.55110300e-01 -4.76658463e-01 5.87175786e-01
-3.85601163e-01 7.61324227e-01 9.42647979e-02 -8.40123892e-01
1.05516434e+00 4.82771136e-02 -4.41835821e-01 -5.58315277e-01
3.13897461e-01 4.86935228e-01 -1.73741505e-01 -5.56832910e-01
6.00321233e-01 -4.63286668e-01 -5.47701307e-02 6.34127200e-01
-4.22315955e-01 -7.58853368e-03 -1.62789389e-01 4.61157918e-01
6.78081155e-01 -3.49490583e-01 4.93467331e-01 -4.15225685e-01
1.26881206e+00 2.49707237e-01 -3.31222601e-02 7.40789235e-01
-8.95302951e-01 -2.82936059e-02 4.33878392e-01 -5.51434338e-01
-5.77315569e-01 -8.42563212e-01 8.63003731e-02 8.57626259e-01
3.53515089e-01 -5.81419037e-04 -7.29778945e-01 -1.59590971e+00
2.65317917e-01 1.45350897e+00 -5.28510213e-01 -8.53224874e-01
-2.01928262e-02 -8.40039611e-01 4.03170437e-01 -3.78648266e-02
7.87995279e-01 -8.31899941e-01 -1.97471276e-01 -1.35427460e-01
5.06918803e-02 -3.88867855e-01 -5.90963542e-01 -2.29573324e-01
-7.99879491e-01 -1.21220756e+00 -3.09411049e-01 -5.28679132e-01
1.03106391e+00 8.26814413e-01 8.52666378e-01 1.41626641e-01
-3.61102000e-02 2.19651267e-01 -7.73170531e-01 -5.91514647e-01
-6.94633484e-01 -4.34037894e-02 1.14115961e-01 3.49651307e-01
1.46076095e+00 -4.81280573e-02 -7.28137493e-01 4.71564144e-01
-7.68676519e-01 -4.92713451e-01 4.52278405e-01 6.46159649e-01
-5.89138865e-01 1.25008464e-01 1.11427534e+00 -1.34981990e+00
1.47572374e+00 -4.43904698e-01 -2.38165498e-01 3.08765531e-01
-1.12717509e+00 7.50443339e-02 6.36494756e-01 -3.56395692e-01
-1.40490818e+00 -4.87378865e-01 6.94819167e-02 3.71524423e-01
-1.06208578e-01 9.34290439e-02 7.15922937e-02 -3.03791136e-01
7.01787710e-01 2.91574150e-01 2.73199856e-01 -2.09402308e-01
5.70542037e-01 1.26079655e+00 -2.40238115e-01 3.37969065e-01
6.12485349e-01 4.96085763e-01 -6.03636205e-01 -5.24983883e-01
-1.43549883e+00 -6.65208757e-01 -4.84361649e-01 -2.18827352e-01
2.22726315e-01 -4.46471959e-01 -7.62108564e-01 8.92967701e-01
-1.05166662e+00 5.56688309e-01 1.81525916e-01 8.29242989e-02
1.01525173e-01 9.57153738e-01 -7.60637999e-01 -1.32990742e+00
-7.99958408e-01 -5.54113746e-01 7.98884213e-01 -6.44036308e-02
-6.52388573e-01 -1.30136263e+00 -1.73891395e-01 9.45930660e-01
5.90382040e-01 -4.15900171e-01 7.11238384e-01 -1.20789945e+00
-1.21661492e-01 -7.96683908e-01 -2.76162207e-01 1.08520007e+00
4.47090924e-01 -2.42276952e-01 -7.36998141e-01 -1.15635693e-01
8.28594208e-01 -3.72767329e-01 7.12011576e-01 -2.56553572e-02
3.34065944e-01 -1.05840099e+00 -1.71881765e-01 -3.06502283e-01
1.04928410e+00 2.86789745e-01 6.69783533e-01 2.42309019e-01
4.32354063e-01 7.92247295e-01 7.13660300e-01 1.67515457e-01
1.61447465e-01 8.69287923e-02 1.96188375e-01 4.40113127e-01
2.83323616e-01 -4.04481024e-01 7.90739536e-01 1.03741050e+00
7.79184043e-01 -2.92527139e-01 -2.29236364e-01 4.46296215e-01
-1.92593801e+00 -9.43817139e-01 -8.05340707e-01 1.67707992e+00
5.55174530e-01 2.66204059e-01 -2.58283466e-01 2.58595377e-01
7.48742700e-01 1.80944085e-01 -7.28114188e-01 -1.12794757e+00
-5.11799380e-02 -9.13796425e-02 6.40566885e-01 9.15463150e-01
-9.55839396e-01 1.26728845e+00 6.56688356e+00 8.37997139e-01
-7.92594254e-01 2.54500866e-01 8.35864365e-01 2.12934718e-01
-5.19482672e-01 -3.33468504e-02 -8.24884474e-01 6.13799155e-01
9.96492803e-01 -3.00635040e-01 -9.36527699e-02 1.26883209e+00
6.50432289e-01 -3.96440834e-01 -4.81940210e-01 6.72187150e-01
1.01069832e+00 -6.20880485e-01 3.58800203e-01 -5.19122742e-02
8.37249517e-01 -8.37988496e-01 3.24846506e-01 2.05385521e-01
5.21481156e-01 -7.60567963e-01 3.54070783e-01 4.35001284e-01
-4.27524865e-01 -7.46319592e-01 1.07982814e+00 4.69982177e-01
-1.07258767e-01 -7.15789422e-02 -5.63462436e-01 -6.95168734e-01
1.90932080e-01 8.41508985e-01 -8.06686580e-01 -8.55020732e-02
2.41818309e-01 9.88948822e-01 -1.05184042e+00 4.20306087e-01
-8.17773342e-01 6.88658416e-01 3.65389109e-01 -8.21758091e-01
3.59144866e-01 -4.74316031e-01 3.60521168e-01 6.55965567e-01
-4.79379855e-02 -2.30875582e-01 -5.35154343e-01 8.09622169e-01
1.13130704e-01 1.93671450e-01 -7.69427299e-01 -1.25773519e-01
-3.28915417e-02 1.22346163e+00 -5.23261845e-01 -5.55697262e-01
-2.97940373e-01 1.69134831e+00 -1.69163957e-01 1.07203484e-01
-5.16018689e-01 -3.21922302e-01 5.27430773e-01 2.46248171e-02
2.41785049e-02 8.36644024e-02 -6.25744641e-01 -1.49181163e+00
4.30583023e-02 -9.94466126e-01 5.50813265e-02 -8.46195698e-01
-1.98722351e+00 2.80761778e-01 -5.21691144e-01 -7.01522052e-01
4.91587520e-02 -8.75931323e-01 -4.93307441e-01 6.35051727e-01
-1.31709719e+00 -9.63542044e-01 4.14083935e-02 8.53117183e-02
8.28073740e-01 -2.71894425e-01 6.78639352e-01 -8.07225853e-02
-2.50867784e-01 4.49879199e-01 1.63685843e-01 -2.29085043e-01
6.54770076e-01 -1.20052874e+00 3.96248400e-01 7.96791196e-01
-1.13612629e-01 9.27770257e-01 8.24124932e-01 -9.69522595e-01
-8.41830015e-01 -6.18573666e-01 1.41934931e+00 -8.99563313e-01
9.39127028e-01 -2.51104206e-01 -8.28617096e-01 3.56161773e-01
1.14267521e-01 -6.50180280e-01 9.15356994e-01 -1.24270365e-01
-4.22695398e-01 -3.77141871e-02 -1.48562729e+00 5.68123996e-01
5.17322302e-01 -6.37113690e-01 -1.06580985e+00 4.10898894e-01
1.68769225e-01 6.03722036e-01 -6.36226237e-02 -1.38768494e-01
6.12070918e-01 -1.17074084e+00 6.14615798e-01 -6.51690543e-01
3.50622177e-01 -1.03966638e-01 3.63301307e-01 -1.45670760e+00
-1.80548891e-01 -2.78106362e-01 -1.72595695e-01 1.00331080e+00
4.00789499e-01 -9.90312755e-01 1.19207549e+00 1.08570027e+00
3.30380827e-01 -3.44185442e-01 -5.60163915e-01 -5.07487774e-01
7.95772821e-02 3.83598767e-02 2.78100893e-02 7.78729737e-01
4.72342551e-01 8.45438242e-01 -6.31382883e-01 -7.13661537e-02
7.09323525e-01 -6.04749881e-02 3.72622669e-01 -1.15480077e+00
3.96474689e-01 -3.00925463e-01 -4.01256442e-01 -1.30368865e+00
2.69860953e-01 -5.55848837e-01 -2.02589631e-01 -1.10360038e+00
8.63633975e-02 2.92311817e-01 -8.01097155e-02 2.02013832e-02
-3.36337388e-01 3.67888361e-01 -9.21883211e-02 4.41427618e-01
-9.29537535e-01 6.41458154e-01 1.13115108e+00 -1.44565508e-01
9.71021354e-02 5.02926350e-01 -1.35997164e+00 1.03200710e+00
1.06667817e+00 -6.04507744e-01 -6.49815321e-01 7.27328286e-03
6.36537254e-01 -6.98199928e-01 3.36329609e-01 -3.33451331e-01
1.11103885e-01 1.42444476e-01 1.91271842e-01 -5.75949669e-01
3.65799889e-02 -7.74614930e-01 -6.18566811e-01 6.29753709e-01
-6.66119695e-01 -1.03134274e-01 -3.53410035e-01 7.98964918e-01
-2.18905047e-01 -9.71321166e-01 7.56595612e-01 -5.87854445e-01
-3.97199929e-01 -3.66592079e-01 -1.06692123e+00 -2.99998194e-01
1.01617551e+00 -2.43707851e-01 -6.14355683e-01 -8.09556127e-01
-4.08204406e-01 7.69691542e-02 3.40726852e-01 7.15828896e-01
7.56527841e-01 -1.13471508e+00 -4.54975635e-01 1.33196756e-01
2.44241320e-02 -1.22875750e+00 -9.22489017e-02 4.43175226e-01
-4.41248834e-01 7.39380181e-01 9.30374935e-02 2.10886255e-01
-1.55709267e+00 7.74509192e-01 2.44994015e-01 6.90119434e-03
6.96089640e-02 8.37861598e-01 -1.76069647e-01 -4.50019002e-01
-5.07562645e-02 1.85244843e-01 -2.68562376e-01 2.88769364e-01
7.19429553e-01 9.19477999e-01 1.46754488e-01 -9.16971743e-01
-3.42880428e-01 -2.21476369e-02 -6.13022387e-01 6.67961687e-02
7.22626448e-01 -6.45559132e-01 -5.54231942e-01 4.33221817e-01
1.54037011e+00 2.44399071e-01 -5.51456451e-01 1.41867638e-01
4.35359329e-01 -9.50246692e-01 1.45729125e-01 -1.32569063e+00
-5.47828674e-01 8.00067186e-01 5.22717714e-01 8.72884214e-01
6.17580593e-01 -4.40634608e-01 9.78260517e-01 3.60070497e-01
2.61032045e-01 -1.35257566e+00 4.30817664e-01 2.98832029e-01
6.02678955e-01 -1.60518861e+00 1.84680551e-01 -6.17088914e-01
-9.63267267e-01 1.00145197e+00 4.66267645e-01 -3.59448940e-01
7.42800832e-01 -5.54488897e-01 5.12717307e-01 -4.33188170e-01
-6.30242288e-01 2.89798349e-01 4.35736030e-01 5.85331142e-01
3.71055812e-01 -1.97001104e-03 -1.10878503e+00 5.88308513e-01
-2.63046980e-01 -1.92474782e-01 7.92254865e-01 8.05895686e-01
-8.61143768e-01 -1.22198427e+00 -2.41103992e-01 6.19025230e-01
-6.11718357e-01 -3.65416199e-01 -1.29029226e+00 2.36409128e-01
-1.54066041e-01 1.58781171e+00 -5.10740936e-01 -2.16415450e-01
5.29106334e-02 3.37263852e-01 -5.33498898e-02 -6.63295865e-01
-6.35658324e-01 -4.37569082e-01 3.00543785e-01 -1.70239210e-01
-5.29764771e-01 -6.51125789e-01 -5.25947809e-01 -4.68386739e-01
-9.97654259e-01 5.13378620e-01 6.31284833e-01 1.07656848e+00
4.61380273e-01 -3.61396998e-01 7.14817762e-01 -8.05344135e-02
-1.20962751e+00 -1.09046972e+00 -5.12000382e-01 8.33071887e-01
2.38251448e-01 -3.18303525e-01 -1.29526365e+00 -8.63977745e-02] | [7.860377311706543, 10.047374725341797] |
8eb32046-4375-4f31-adc4-41f875680af3 | towards-generating-diverse-audio-captions-via | 2212.02033 | null | https://arxiv.org/abs/2212.02033v1 | https://arxiv.org/pdf/2212.02033v1.pdf | Towards Generating Diverse Audio Captions via Adversarial Training | Automated audio captioning is a cross-modal translation task for describing the content of audio clips with natural language sentences. This task has attracted increasing attention and substantial progress has been made in recent years. Captions generated by existing models are generally faithful to the content of audio clips, however, these machine-generated captions are often deterministic (e.g., generating a fixed caption for a given audio clip), simple (e.g., using common words and simple grammar), and generic (e.g., generating the same caption for similar audio clips). When people are asked to describe the content of an audio clip, different people tend to focus on different sound events and describe an audio clip diversely from various aspects using distinct words and grammar. We believe that an audio captioning system should have the ability to generate diverse captions, either for a fixed audio clip, or across similar audio clips. To this end, we propose an adversarial training framework based on a conditional generative adversarial network (C-GAN) to improve diversity of audio captioning systems. A caption generator and two hybrid discriminators compete and are learned jointly, where the caption generator can be any standard encoder-decoder captioning model used to generate captions, and the hybrid discriminators assess the generated captions from different criteria, such as their naturalness and semantics. We conduct experiments on the Clotho dataset. The results show that our proposed model can generate captions with better diversity as compared to state-of-the-art methods. | ['Wenwu Wang', 'Mark D. Plumbley', 'Jianyuan Sun', 'Xubo Liu', 'Xinhao Mei'] | 2022-12-05 | null | null | null | null | ['audio-captioning'] | ['audio'] | [ 4.75005329e-01 4.62923981e-02 2.39063352e-01 -4.28594410e-01
-1.45756829e+00 -8.73518169e-01 3.73942763e-01 -2.82432318e-01
1.66687220e-01 8.77329469e-01 5.39678395e-01 2.99393028e-01
5.64254403e-01 -5.76642990e-01 -1.00145376e+00 -4.86004442e-01
2.42849633e-01 6.61515534e-01 -4.84638438e-02 -3.41284901e-01
-2.57730961e-01 -1.83424070e-01 -1.35451913e+00 5.78453720e-01
7.62371898e-01 9.03006673e-01 3.96380246e-01 8.14462781e-01
-2.68216543e-02 5.86235225e-01 -1.05767441e+00 -5.76519370e-01
1.09837409e-02 -1.04040241e+00 -5.91296315e-01 -7.70904571e-02
3.50932032e-01 -2.65981518e-02 -3.18067253e-01 1.02853179e+00
7.49566853e-01 -2.19116732e-01 6.62892342e-01 -1.45058417e+00
-1.00113547e+00 9.63815093e-01 -8.83395448e-02 -2.30832860e-01
8.03695679e-01 1.98784649e-01 1.10177481e+00 -5.55915952e-01
3.34521949e-01 1.12295783e+00 4.15708244e-01 9.90901589e-01
-1.21449947e+00 -9.28869545e-01 6.36854172e-02 1.14643294e-02
-1.57818317e+00 -3.72796953e-01 9.95896995e-01 -4.14692312e-01
3.00035272e-02 4.20038700e-01 5.45496523e-01 1.75972748e+00
1.32059142e-01 7.10591435e-01 7.55268395e-01 -2.26375028e-01
3.27337146e-01 2.31982544e-01 -5.98821998e-01 1.70670766e-02
-2.38283306e-01 -2.36977860e-01 -6.27024055e-01 -2.68257082e-01
5.58785617e-01 -3.90639246e-01 -4.63095278e-01 7.32182041e-02
-1.39140701e+00 8.85132194e-01 2.08717629e-01 1.08427957e-01
-5.06686807e-01 2.81760782e-01 5.77236533e-01 2.13541985e-01
1.25032693e-01 5.70555747e-01 -8.50197449e-02 -3.23407561e-01
-8.61059666e-01 5.31808019e-01 7.02748477e-01 1.15670204e+00
3.30366939e-01 2.10151792e-01 -7.12772012e-01 9.82954025e-01
2.04741985e-01 1.02251649e+00 6.61754906e-01 -8.12680840e-01
7.04632282e-01 -1.58022493e-01 1.61551625e-01 -8.98279548e-01
2.64212459e-01 -2.00286582e-01 -9.69989955e-01 -2.98319489e-01
5.36939427e-02 -6.35324955e-01 -9.01893735e-01 2.23238087e+00
-1.29761666e-01 3.52653801e-01 3.89389515e-01 1.17138362e+00
9.77886498e-01 1.21859169e+00 1.73235342e-01 -1.23195536e-02
1.31245136e+00 -8.59212995e-01 -7.69513309e-01 -3.74954790e-01
-1.03249028e-01 -1.09075105e+00 1.14904463e+00 1.52771503e-01
-1.16430879e+00 -9.32920277e-01 -8.97759140e-01 3.24515969e-01
9.72090513e-02 -7.77618587e-02 8.26575384e-02 5.57726741e-01
-7.27343440e-01 6.04609540e-03 -3.30594450e-01 4.97928634e-02
1.53051317e-03 -6.01008758e-02 -1.83569431e-01 -6.49984255e-02
-1.76599193e+00 3.35031033e-01 4.41756338e-01 -8.47091526e-02
-1.40808880e+00 -6.10952735e-01 -9.53714013e-01 2.22155020e-01
-7.07788914e-02 -8.09749603e-01 1.55868101e+00 -1.61696315e+00
-1.56026375e+00 4.93706644e-01 -4.11115363e-02 -5.13264179e-01
3.79297614e-01 -3.34742181e-02 -7.34955788e-01 3.35668296e-01
3.33219081e-01 1.09887099e+00 1.07277513e+00 -1.43165255e+00
-4.00325209e-01 3.39202225e-01 1.48892418e-01 3.08945268e-01
-2.37133235e-01 1.14664175e-01 -4.36806202e-01 -1.02997208e+00
-4.10812765e-01 -1.26610148e+00 1.33145362e-01 -2.40057021e-01
-5.64473450e-01 2.30336979e-01 5.87897480e-01 -7.59318650e-01
1.06374240e+00 -2.35283232e+00 2.40073532e-01 -1.27613414e-02
-3.89157712e-01 1.17595896e-01 -4.73177254e-01 5.70874691e-01
-1.33623689e-01 2.77777255e-01 -4.65013087e-01 -3.34797353e-01
1.48691863e-01 1.61778361e-01 -6.96193755e-01 -1.27242774e-01
4.62019056e-01 7.79049158e-01 -1.09999299e+00 -5.26256979e-01
-2.36726835e-01 8.05912018e-01 -4.05171007e-01 7.28691280e-01
-5.38873196e-01 6.86462224e-01 -4.01134968e-01 4.11758900e-01
3.49544883e-01 -2.64797267e-02 -8.99116322e-02 1.84611324e-02
3.76818240e-01 2.51814544e-01 -1.08036482e+00 1.72059023e+00
-7.73948789e-01 7.08195388e-01 -1.38906330e-01 -6.90441430e-01
1.20343089e+00 9.58375752e-01 8.86271447e-02 -3.84775460e-01
-1.04450434e-02 3.24055105e-01 -2.33788099e-02 -5.00654399e-01
3.08610201e-01 -3.61063391e-01 -7.15741634e-01 5.06120503e-01
4.78293002e-02 -5.24712265e-01 1.26454145e-01 8.46554190e-02
8.56609821e-01 -2.33849287e-02 -5.02558723e-02 3.54838133e-01
5.29931426e-01 -1.44633606e-01 3.91446620e-01 7.14551568e-01
3.64945307e-02 1.60708618e+00 3.38581622e-01 -1.89767946e-02
-1.23251629e+00 -1.21021461e+00 1.99651539e-01 1.08405852e+00
1.57938287e-01 -2.89710581e-01 -7.84143746e-01 -4.50738251e-01
-5.00709891e-01 7.45179355e-01 -5.57625592e-01 -3.25373262e-01
-5.05231738e-01 -8.61477032e-02 8.64506364e-01 4.82765019e-01
5.35241246e-01 -1.55604172e+00 -1.88011125e-01 5.09676635e-01
-7.95730352e-01 -1.20697999e+00 -1.06306624e+00 -2.10752368e-01
-1.67647675e-01 -4.60941225e-01 -1.32283318e+00 -1.19425714e+00
3.74811947e-01 7.06573352e-02 1.26280081e+00 -5.03456950e-01
1.60220146e-01 3.58941168e-01 -7.92310119e-01 -6.37849987e-01
-9.82864857e-01 2.11988971e-01 -1.84543859e-02 4.12808418e-01
3.29592898e-02 -5.84300995e-01 -3.47567260e-01 4.14677709e-01
-1.24110579e+00 3.12233627e-01 6.46671176e-01 8.33360374e-01
7.25160241e-01 -3.29921156e-01 1.01501918e+00 -5.72154641e-01
1.02903199e+00 -7.80053139e-01 -1.44543331e-02 1.76858813e-01
1.98265284e-01 -9.76102427e-02 1.26571155e+00 -9.89620388e-01
-8.83534253e-01 1.49067089e-01 -2.96114117e-01 -7.56580651e-01
-2.89025158e-01 5.83562434e-01 -3.76305044e-01 4.26980168e-01
6.36440396e-01 6.97139740e-01 -1.89801648e-01 -9.71859694e-02
2.63794482e-01 1.06689167e+00 8.93322587e-01 -7.43981421e-01
9.07183051e-01 2.82240864e-02 -4.56148922e-01 -4.66904074e-01
-7.85740018e-01 -3.55775282e-02 7.78086856e-02 -3.18504870e-01
9.85781550e-01 -1.23518598e+00 -1.30499631e-01 3.92766327e-01
-1.52047658e+00 -6.99431461e-04 -2.16595799e-01 4.39595312e-01
-8.05818200e-01 -1.07728206e-01 -3.87258589e-01 -5.95736623e-01
-5.30211627e-01 -1.23716319e+00 1.43739438e+00 3.81922096e-01
-5.79112947e-01 -6.42185688e-01 3.25517833e-01 2.53686965e-01
4.29956436e-01 5.39466321e-01 6.38958812e-01 -7.79114127e-01
-2.69048423e-01 -2.58466959e-01 8.73850733e-02 4.87590969e-01
2.36240059e-01 -1.87276810e-01 -9.27964032e-01 -1.57513052e-01
-2.46848255e-01 -6.35888696e-01 2.09862277e-01 1.05854295e-01
9.97315526e-01 -5.75177848e-01 1.49727330e-01 2.29574919e-01
1.11232710e+00 3.29189241e-01 8.19673657e-01 3.83541100e-02
4.88006294e-01 4.19910431e-01 4.87863541e-01 2.88047493e-01
1.29632115e-01 8.50844741e-01 3.80614102e-01 -5.88252954e-02
-1.88907698e-01 -7.87578702e-01 7.06277490e-01 9.03175175e-01
2.90877253e-01 -5.41334510e-01 -6.76885784e-01 8.38114381e-01
-1.58130634e+00 -1.14278901e+00 2.48205021e-01 1.98451042e+00
1.14570642e+00 1.22147664e-01 1.27526715e-01 1.48705989e-01
1.31026709e+00 1.35540590e-01 -5.31779289e-01 -5.60355246e-01
-1.92572087e-01 2.49898359e-01 -1.44136667e-01 1.76399797e-01
-8.39677036e-01 6.42408192e-01 5.36705685e+00 8.17083538e-01
-1.31473422e+00 2.40294747e-02 5.17470181e-01 -6.43411726e-02
-4.69444692e-01 -2.80288965e-01 -4.28841263e-01 1.03492928e+00
1.03928494e+00 -6.02217197e-01 4.43589985e-01 7.83612847e-01
2.22151190e-01 6.35313630e-01 -1.20456791e+00 1.18004823e+00
4.00582701e-01 -1.14089346e+00 5.13477325e-01 -4.14485157e-01
9.73467052e-01 -2.87532568e-01 2.75304854e-01 4.53653574e-01
4.59695980e-02 -1.21988738e+00 1.18453062e+00 4.40832704e-01
9.73090768e-01 -7.23101616e-01 9.93250847e-01 1.34929955e-01
-1.22336721e+00 2.54297107e-01 -1.06261954e-01 1.51203945e-01
6.43803000e-01 2.67337710e-01 -8.26363385e-01 5.46338260e-01
6.07299507e-01 1.60759717e-01 -2.44013518e-01 1.10240626e+00
-3.92926276e-01 9.30580318e-01 -6.14911690e-02 -1.80906385e-01
2.61326939e-01 7.79914260e-02 6.96680844e-01 1.19058251e+00
6.61194444e-01 -1.49917200e-01 1.45404786e-01 8.21386158e-01
-3.29295874e-01 2.29594558e-01 -5.59951186e-01 -2.70382106e-01
6.73796833e-01 9.24915969e-01 -2.51977853e-02 -2.73451656e-01
-2.89917648e-01 1.06430709e+00 -2.51618832e-01 3.64167839e-01
-1.27960527e+00 -6.55291736e-01 5.44519544e-01 6.63220435e-02
3.22864026e-01 1.73759103e-01 -5.19681862e-03 -9.35557723e-01
2.97879040e-01 -1.21109819e+00 1.30428553e-01 -1.33096182e+00
-1.55806768e+00 1.22938848e+00 -7.07145110e-02 -1.78430903e+00
-6.92343771e-01 1.16671011e-01 -8.62103105e-01 1.01966739e+00
-1.17805505e+00 -1.17513776e+00 -3.72354090e-01 5.39004564e-01
7.75588989e-01 -3.55870724e-01 9.52757537e-01 3.03004175e-01
-3.53387967e-02 7.75004506e-01 -1.46075040e-01 3.68495166e-01
1.00173295e+00 -9.49621201e-01 4.30430114e-01 6.25730515e-01
3.51510465e-01 3.17527205e-01 1.12170136e+00 -4.29913342e-01
-8.77107978e-01 -1.51847804e+00 9.23610806e-01 -3.09844073e-02
4.79685873e-01 -5.19818962e-01 -1.05843902e+00 4.74764407e-01
4.82941449e-01 -1.88335419e-01 9.64478314e-01 -5.13803840e-01
-2.92913646e-01 -1.44807681e-01 -1.01060402e+00 6.47077858e-01
6.25977814e-01 -5.77573538e-01 -8.40803325e-01 2.50820816e-01
1.04868352e+00 -5.26613116e-01 -5.96877217e-01 1.65171936e-01
6.05171561e-01 -5.83223164e-01 6.62675083e-01 -5.32538891e-01
8.45403790e-01 -5.07579982e-01 -2.59539485e-01 -1.65257037e+00
-1.52339473e-01 -8.73896956e-01 2.18089610e-01 1.68848574e+00
6.18753552e-01 -1.95463717e-01 4.05875117e-01 2.84839749e-01
-3.12060982e-01 -3.17061633e-01 -8.31321716e-01 -9.00941074e-01
-1.32270232e-02 -3.14719260e-01 9.54271495e-01 6.88819528e-01
-7.69778416e-02 7.03399777e-01 -8.47071469e-01 2.68842369e-01
1.05412960e-01 2.67169476e-01 8.49155128e-01 -9.96603370e-01
-4.65010315e-01 -1.17911145e-01 -3.42743576e-01 -9.51043785e-01
3.80876750e-01 -7.42892921e-01 4.83435780e-01 -1.34056294e+00
9.89880785e-02 -1.90863371e-01 2.98047774e-02 3.28620851e-01
-2.96687782e-01 6.41213298e-01 4.08025742e-01 2.14965627e-01
-5.19585729e-01 7.76319504e-01 1.33239400e+00 -5.53834140e-01
-2.72259086e-01 1.30194128e-01 -8.36122155e-01 2.73873270e-01
8.76152813e-01 -6.94983423e-01 -4.19996560e-01 -4.73066092e-01
2.16373920e-01 4.23153251e-01 2.98804939e-01 -1.22406983e+00
2.64200661e-02 -8.57737139e-02 1.45966575e-01 4.78375843e-03
4.92761940e-01 -6.50739074e-01 5.61648250e-01 1.34675413e-01
-7.83583760e-01 -4.05156985e-02 1.57747880e-01 4.48585808e-01
-8.01569939e-01 -2.60743648e-01 7.65162408e-01 -8.87874961e-02
-1.40000641e-01 2.69853741e-01 -3.24672639e-01 3.86452138e-01
9.65230942e-01 5.90648651e-02 -3.83382924e-02 -1.14520133e+00
-7.28460133e-01 9.87395346e-02 2.96389043e-01 8.09549153e-01
6.46873474e-01 -1.90200198e+00 -1.38343871e+00 -3.60889286e-02
4.12313551e-01 -3.36998105e-02 2.13912815e-01 1.20092314e-02
-2.37602547e-01 3.54463249e-01 -3.27402055e-01 -4.88782108e-01
-9.87150967e-01 3.97912204e-01 9.74778756e-02 -2.34691482e-02
-1.49114996e-01 4.79191959e-01 3.84622455e-01 -2.48402599e-02
1.98394254e-01 2.34560259e-02 -1.73647240e-01 -1.19547032e-01
6.41406655e-01 -2.07754463e-01 -3.02549809e-01 -9.08499599e-01
-9.18226093e-02 5.39486945e-01 1.90853328e-01 -3.99321437e-01
1.01022577e+00 2.88888458e-02 3.16951662e-01 5.23527145e-01
1.26394856e+00 1.13451622e-01 -1.03095961e+00 4.01013456e-02
-8.35641861e-01 -2.28028730e-01 -4.01835620e-01 -7.75963128e-01
-9.30239201e-01 7.12344527e-01 3.93464267e-01 6.03297591e-01
1.14378965e+00 1.65116459e-01 1.14626145e+00 -5.50318062e-02
3.64709049e-01 -6.86947286e-01 3.04914385e-01 3.40820462e-01
1.38297749e+00 -1.02155232e+00 -7.77402639e-01 -1.97875381e-01
-1.23624420e+00 8.80953312e-01 6.35175586e-01 -1.10930517e-01
-6.32515550e-02 -1.05750840e-02 3.33477318e-01 4.31739867e-01
-6.18708909e-01 -5.09809470e-03 3.62557888e-01 9.18181241e-01
3.49668920e-01 2.00872377e-01 -1.25767574e-01 9.58926797e-01
-6.61628723e-01 -1.21646546e-01 6.69166684e-01 4.09223408e-01
-8.32743496e-02 -1.20093668e+00 -6.98141336e-01 1.03367880e-01
-4.76018250e-01 -1.93757176e-01 -4.83761400e-01 3.46481949e-01
7.22110793e-02 1.12130320e+00 6.29016161e-02 -4.80953515e-01
2.61439800e-01 1.20875239e-01 3.63321826e-02 -8.12801540e-01
-5.36146045e-01 -5.32057174e-02 -8.43402445e-02 6.23028800e-02
-3.24904263e-01 -5.45778930e-01 -1.00747430e+00 1.97340414e-01
1.61097907e-02 7.37677217e-01 5.28024793e-01 6.66165471e-01
4.08748239e-01 5.84988773e-01 9.98875976e-01 -8.51437926e-01
-3.94672900e-01 -1.08734584e+00 -4.00604904e-01 6.85193121e-01
3.01019132e-01 -1.93780467e-01 -4.69473928e-01 5.94186246e-01] | [15.269145011901855, 4.918099403381348] |
5aca7a86-642a-4ba9-bb05-1c897f4f5c9e | perceiving-3d-human-object-spatial | 2007.15649 | null | https://arxiv.org/abs/2007.15649v2 | https://arxiv.org/pdf/2007.15649v2.pdf | Perceiving 3D Human-Object Spatial Arrangements from a Single Image in the Wild | We present a method that infers spatial arrangements and shapes of humans and objects in a globally consistent 3D scene, all from a single image in-the-wild captured in an uncontrolled environment. Notably, our method runs on datasets without any scene- or object-level 3D supervision. Our key insight is that considering humans and objects jointly gives rise to "3D common sense" constraints that can be used to resolve ambiguity. In particular, we introduce a scale loss that learns the distribution of object size from data; an occlusion-aware silhouette re-projection loss to optimize object pose; and a human-object interaction loss to capture the spatial layout of objects with which humans interact. We empirically validate that our constraints dramatically reduce the space of likely 3D spatial configurations. We demonstrate our approach on challenging, in-the-wild images of humans interacting with large objects (such as bicycles, motorcycles, and surfboards) and handheld objects (such as laptops, tennis rackets, and skateboards). We quantify the ability of our approach to recover human-object arrangements and outline remaining challenges in this relatively domain. The project webpage can be found at https://jasonyzhang.com/phosa. | ['Jason Y. Zhang', 'Hanbyul Joo', 'Deva Ramanan', 'Sam Pepose', 'Angjoo Kanazawa', 'Jitendra Malik'] | 2020-07-30 | null | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/1474_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123570035.pdf | eccv-2020-8 | ['3d-shape-reconstruction-from-a-single-2d'] | ['computer-vision'] | [ 5.36916889e-02 -8.93136561e-02 2.13320881e-01 -4.98701125e-01
-4.51982915e-01 -8.48926961e-01 4.00528312e-01 5.74470088e-02
-4.34600711e-01 2.04064935e-01 5.02078570e-02 -3.80245224e-02
-1.02029137e-01 -3.05089474e-01 -1.03616309e+00 -1.64436370e-01
-9.52608287e-02 9.12458599e-01 2.87213415e-01 2.39906795e-02
3.20750862e-01 7.45027184e-01 -1.60020638e+00 1.13161854e-01
3.66143912e-01 1.02279794e+00 3.87706399e-01 8.40581238e-01
3.68716896e-01 5.28389215e-01 -3.99336934e-01 -2.44292155e-01
8.58338237e-01 1.02000766e-01 -6.00730777e-01 5.94703317e-01
9.15932655e-01 -3.65811795e-01 -4.21148866e-01 7.17166245e-01
4.05344337e-01 3.76052260e-01 6.49143875e-01 -1.38500452e+00
-4.31991339e-01 -1.51575506e-01 -7.61443436e-01 8.28678012e-02
5.52122355e-01 5.52190185e-01 9.39674735e-01 -7.43553519e-01
7.32336998e-01 1.52905047e+00 5.77648103e-01 1.03242047e-01
-1.39481270e+00 -3.62890363e-01 3.98259282e-01 -8.98743868e-02
-1.48297989e+00 -4.58887577e-01 6.07300997e-01 -7.82052279e-01
9.44493592e-01 2.12306321e-01 7.40882576e-01 1.04831243e+00
3.64533849e-02 7.60652423e-01 8.31694126e-01 -3.83381039e-01
2.08983704e-01 1.20818997e-02 -2.35145409e-02 7.40584850e-01
4.85116988e-01 -4.40169033e-03 -7.09063530e-01 -1.01803221e-01
1.07631314e+00 2.06331432e-01 3.86100337e-02 -1.12617528e+00
-1.15481901e+00 3.21698517e-01 3.93287003e-01 -1.76842004e-01
-1.75022379e-01 7.03080297e-02 -9.18730721e-02 -1.80040181e-01
2.15684012e-01 5.30731320e-01 -5.60382605e-01 -8.55343342e-02
-3.57396215e-01 9.45878863e-01 5.53271115e-01 1.38589239e+00
6.98430836e-01 -5.16797602e-01 1.16134584e-01 5.23065746e-01
7.22926557e-02 7.19015121e-01 -1.96717292e-01 -1.37563920e+00
7.08563864e-01 6.68053508e-01 6.53845727e-01 -1.19970489e+00
-6.02547646e-01 -1.23356953e-01 -3.29481691e-01 3.81968498e-01
5.63932359e-01 1.18327364e-02 -8.31163287e-01 1.62754095e+00
6.44439757e-01 -2.99618561e-02 -4.85200673e-01 1.31225204e+00
3.58374208e-01 1.12634838e-01 -1.08130045e-01 5.05226254e-01
1.62553287e+00 -9.43926156e-01 -2.63998687e-01 -8.01482797e-01
1.74797758e-01 -5.75267255e-01 1.23730147e+00 2.48573422e-01
-1.23725629e+00 -5.71442962e-01 -7.76127815e-01 -4.46520478e-01
-3.13491225e-01 1.84185356e-01 5.65437198e-01 3.29530388e-01
-5.76636314e-01 4.73892599e-01 -9.96658564e-01 -4.92028147e-01
3.56810868e-01 2.29363248e-01 -5.08007824e-01 -4.86918241e-02
-2.91862488e-01 9.73961532e-01 1.82092324e-01 -2.89810188e-02
-7.59417534e-01 -7.23505676e-01 -1.11605251e+00 -1.29367560e-01
8.99115682e-01 -9.30261910e-01 1.28058875e+00 -4.84125048e-01
-1.10546219e+00 1.27587104e+00 -1.05988473e-01 -2.12783337e-01
7.16540277e-01 -6.45783067e-01 1.69028997e-01 1.17429845e-01
3.32881510e-01 6.76280379e-01 6.12546742e-01 -1.52872956e+00
-6.21272266e-01 -9.91332531e-01 2.20256776e-01 5.82865357e-01
2.37638772e-01 -2.16474250e-01 -8.07068944e-01 -5.31624317e-01
3.86906207e-01 -1.15948617e+00 -1.89213097e-01 4.22586530e-01
-6.01928115e-01 1.56057537e-01 5.39464414e-01 -5.60185850e-01
7.12765574e-01 -2.04247022e+00 3.93039972e-01 2.45331690e-01
7.22569972e-02 -1.66481674e-01 -1.91929445e-01 1.06680475e-01
2.47168854e-01 -7.12015629e-02 -2.69278526e-01 -6.72815323e-01
2.90807039e-01 1.06571943e-01 -7.48091340e-02 5.40812790e-01
1.51059955e-01 1.02978086e+00 -8.29492629e-01 -3.30862194e-01
3.99213165e-01 4.47733462e-01 -6.14467680e-01 3.75877678e-01
-3.03145885e-01 5.49660385e-01 -3.59413773e-01 7.74908304e-01
8.05541396e-01 -3.30798805e-01 2.08253890e-01 -1.35574564e-01
8.35641697e-02 2.06937298e-01 -1.42022049e+00 1.95069551e+00
-1.97115764e-01 3.28787088e-01 3.45072836e-01 -5.38155794e-01
5.74801385e-01 -3.27770233e-01 3.61992776e-01 -5.41690648e-01
4.97907139e-02 -1.59842074e-01 -4.28670019e-01 -5.56871593e-01
5.79504609e-01 1.96381852e-01 -1.31198645e-01 3.41088235e-01
-1.98310390e-01 -5.31144142e-01 6.47445694e-02 8.04793462e-02
1.06497192e+00 4.06509161e-01 2.07334027e-01 -1.48043439e-01
-1.27643168e-01 6.62146369e-03 4.18236285e-01 8.10832977e-01
-1.13216534e-01 8.65740120e-01 2.87385494e-01 -7.40093172e-01
-1.40676486e+00 -1.42431617e+00 7.48197734e-02 1.07416928e+00
4.66178775e-01 -2.81404704e-01 -5.07577777e-01 -5.03938675e-01
4.51302677e-01 4.12495136e-01 -5.91810226e-01 1.82850361e-01
-6.43826604e-01 -1.55642346e-01 2.25954577e-02 6.02784395e-01
1.61735803e-01 -8.93964767e-01 -1.14391887e+00 -2.50491977e-01
-1.94073051e-01 -1.39920604e+00 -8.41946781e-01 1.28446996e-01
-6.43081427e-01 -1.45795822e+00 -3.96244705e-01 -6.07456923e-01
9.64759767e-01 5.64128935e-01 1.22266245e+00 -7.54513517e-02
-7.72667229e-01 7.31272399e-01 -1.75083771e-01 -4.53121006e-01
2.66684443e-01 -1.26532599e-01 3.12200159e-01 -2.47802317e-01
3.16210121e-01 -6.39531553e-01 -6.78878307e-01 5.69665790e-01
-5.51634908e-01 1.17403999e-01 2.64008790e-01 4.27863121e-01
6.97146714e-01 -6.34709448e-02 -3.04349691e-01 -6.63896441e-01
8.98203179e-02 -2.45043650e-01 -9.85454857e-01 1.70557141e-01
-5.44523671e-02 -8.41855258e-02 9.32697281e-02 -3.68136525e-01
-8.29644799e-01 5.24614096e-01 5.78634202e-01 -7.30345488e-01
-4.36645031e-01 -1.16213635e-01 -3.03310007e-01 7.44482428e-02
6.96378529e-01 -1.18369795e-01 -1.17540769e-01 -7.23760843e-01
3.12442452e-01 3.61177266e-01 8.86933804e-01 -1.00486541e+00
8.39960039e-01 8.50239575e-01 1.06689349e-01 -7.20540345e-01
-8.36748302e-01 -5.58343410e-01 -1.04339218e+00 8.19114875e-03
8.90785396e-01 -1.05686510e+00 -1.10283530e+00 4.84157324e-01
-1.15026438e+00 -5.95720291e-01 -3.00385058e-01 3.74629617e-01
-7.30062485e-01 1.75704017e-01 -2.80911624e-01 -9.72229719e-01
3.32753837e-01 -9.16280627e-01 1.63598740e+00 1.15292110e-01
-3.51220816e-01 -6.28935635e-01 -8.39839056e-02 7.03023612e-01
-1.34357542e-01 3.78165901e-01 5.89137793e-01 -1.49819940e-01
-9.49414074e-01 -2.82559847e-03 -1.70062691e-01 -2.14709640e-02
4.71878052e-02 -2.49418527e-01 -7.04908550e-01 -3.14663827e-01
-2.65982121e-01 -3.91537488e-01 5.27096987e-01 5.77543736e-01
1.22565675e+00 -2.38397941e-01 -4.15449768e-01 7.71440148e-01
9.67366755e-01 -3.29158194e-02 2.53303587e-01 1.73442870e-01
7.83836007e-01 9.12670910e-01 6.81576431e-01 7.68844962e-01
6.90638483e-01 1.02761054e+00 4.29544538e-01 9.86298323e-02
5.24291769e-02 -5.44056058e-01 -6.52933866e-03 -7.61714624e-03
4.00581658e-02 -3.61695558e-01 -1.00106907e+00 5.15145481e-01
-1.78836381e+00 -7.46390104e-01 8.14722255e-02 2.35833025e+00
3.93938720e-01 1.26920745e-01 3.36479902e-01 -3.96465689e-01
7.27529526e-01 5.42161195e-03 -9.31785405e-01 2.52431452e-01
-1.70839429e-02 -1.56212330e-01 6.37670279e-01 6.96125269e-01
-1.28731811e+00 7.92451382e-01 6.15933657e+00 1.82155818e-01
-5.67037106e-01 -2.83314228e-01 5.04990578e-01 -5.57740033e-01
-1.41293585e-01 -1.19642522e-02 -8.19811761e-01 3.12399477e-01
9.53408033e-02 2.28319392e-01 6.61762714e-01 9.21489537e-01
1.67735085e-01 -2.78936803e-01 -1.37764287e+00 1.21084774e+00
7.66788498e-02 -1.00243270e+00 -2.98371911e-01 1.94870874e-01
5.45209110e-01 -9.73885506e-02 2.06283659e-01 -1.40992835e-01
5.13674796e-01 -1.01541114e+00 1.20426631e+00 5.59245765e-01
6.95899129e-01 -4.23610985e-01 2.64946997e-01 4.81695592e-01
-1.09557331e+00 -1.04975671e-01 -3.19889188e-01 -3.17146212e-01
1.69754878e-01 3.16966593e-01 -7.12672889e-01 1.92005053e-01
1.02421546e+00 4.33544964e-01 -6.20756388e-01 8.94247174e-01
-1.44167051e-01 -2.13363871e-01 -6.91786289e-01 2.99339324e-01
-1.66292161e-01 -2.50511467e-01 6.60046279e-01 8.67902160e-01
1.69704258e-01 4.90239173e-01 4.35138762e-01 1.11338758e+00
6.46256655e-02 -3.59938860e-01 -5.56330264e-01 1.74360126e-01
6.75228059e-01 1.02462935e+00 -7.26013541e-01 -1.12485677e-01
1.60351656e-02 1.02859855e+00 3.98150295e-01 3.83504719e-01
-6.34392381e-01 -1.68864548e-01 9.17490423e-01 6.58680201e-01
5.58685303e-01 -5.47885835e-01 -4.11594331e-01 -1.23017728e+00
6.93110406e-01 -7.01843321e-01 1.93272725e-01 -1.07909071e+00
-1.37646365e+00 2.31498703e-01 3.67208391e-01 -1.09294498e+00
-8.21301267e-02 -7.38573194e-01 -1.42588273e-01 6.50965512e-01
-9.14657891e-01 -1.10994709e+00 -6.60439610e-01 4.72741455e-01
4.23161894e-01 -7.81738758e-02 3.28228056e-01 3.01057007e-02
-1.86718985e-01 3.46081257e-01 -1.89971536e-01 -1.21624395e-02
6.89363837e-01 -1.18824875e+00 7.39362955e-01 5.43216050e-01
2.47789294e-01 6.92896664e-01 7.44332194e-01 -7.05424130e-01
-1.48072529e+00 -7.39876688e-01 5.58723092e-01 -1.21628797e+00
1.30665749e-01 -9.26692843e-01 -5.89074254e-01 9.49303627e-01
-3.26142341e-01 2.39105761e-01 2.23153144e-01 2.93145925e-01
-5.46428800e-01 5.64917661e-02 -1.14814973e+00 6.45032585e-01
1.65740955e+00 -2.90629685e-01 -5.75690866e-01 4.05949593e-01
6.75665677e-01 -8.99258077e-01 -4.27252501e-01 2.15990514e-01
8.91399682e-01 -1.06103277e+00 1.33666682e+00 -7.19577372e-01
1.98454425e-01 -5.10210454e-01 -3.43207687e-01 -1.02457392e+00
-2.37421334e-01 -5.02846003e-01 3.58679751e-03 6.24988139e-01
-2.07867473e-02 -3.43648672e-01 8.99596155e-01 1.08414590e+00
-8.65584181e-04 -4.79437739e-01 -8.11798871e-01 -7.71801174e-01
-3.23845237e-01 -3.40282440e-01 5.43243408e-01 4.88060653e-01
-3.02247971e-01 2.12326705e-01 -3.30693841e-01 3.49858850e-01
8.81232500e-01 8.35214406e-02 1.34488213e+00 -1.33531046e+00
-4.25492376e-01 -2.25219399e-01 -5.10477722e-01 -1.42860126e+00
1.26988113e-01 -4.78177071e-01 1.66633606e-01 -1.07397866e+00
3.51346254e-01 -5.74177980e-01 2.42763355e-01 4.60444450e-01
-9.11935940e-02 3.72104496e-02 4.86672670e-01 1.48643225e-01
-8.70833755e-01 3.39252353e-01 1.18159103e+00 5.91268986e-02
-2.09911644e-01 -1.01497978e-01 -5.89940250e-01 9.91407990e-01
4.92606878e-01 -3.94755721e-01 -3.06106567e-01 -7.89873123e-01
5.88650517e-02 -2.88112052e-02 9.33441818e-01 -9.35877204e-01
2.28523642e-01 -3.89748037e-01 6.14895761e-01 -6.18111908e-01
7.33584702e-01 -8.67610335e-01 2.23386958e-01 7.96806887e-02
-3.47323418e-01 1.65945351e-01 2.11687624e-01 5.93068302e-01
4.22638535e-01 9.19671953e-02 5.42014837e-01 -4.99835074e-01
-5.56116998e-01 4.01161194e-01 1.64628968e-01 2.65108675e-01
9.79061604e-01 -3.13006401e-01 -1.65976897e-01 -2.74780691e-01
-6.00767076e-01 4.97169316e-01 1.05597723e+00 6.56361639e-01
5.48195839e-01 -1.15373588e+00 -5.26801765e-01 4.07528907e-01
1.47111803e-01 5.42928517e-01 3.13877463e-01 4.02102768e-01
-6.44522429e-01 3.27799112e-01 -2.80492216e-01 -7.91701376e-01
-1.32412112e+00 5.25717974e-01 2.85679102e-01 4.68742661e-02
-6.43987596e-01 8.73591840e-01 5.71272969e-01 -8.16109419e-01
4.78915572e-01 -4.32894617e-01 5.12653887e-01 -4.36828583e-01
3.79910022e-01 5.19827366e-01 -1.20887861e-01 -6.07834041e-01
-5.61885417e-01 8.64862382e-01 1.82052702e-01 -2.44343415e-01
1.32914150e+00 -3.17285240e-01 1.19862705e-01 3.99288774e-01
9.53333676e-01 6.51312545e-02 -1.88974416e+00 -3.02091867e-01
-2.67574281e-01 -1.05427194e+00 -4.80063617e-01 -8.09435666e-01
-6.54693842e-01 7.13776171e-01 4.86496776e-01 -2.79693961e-01
6.78509653e-01 3.58950973e-01 4.85466450e-01 5.90051651e-01
7.16375887e-01 -1.08692110e+00 4.14323866e-01 3.84359568e-01
9.72077370e-01 -1.36511362e+00 1.91317081e-01 -3.83988440e-01
-7.32704401e-01 7.25817978e-01 9.02573586e-01 -1.52299166e-01
3.19696546e-01 4.62670177e-01 -1.83689296e-01 -3.14443618e-01
-4.86693501e-01 -6.82394877e-02 3.82415771e-01 7.02660799e-01
-2.12697256e-02 2.21437126e-01 5.18611014e-01 5.27331293e-01
-3.81505728e-01 -2.57254124e-01 1.66857570e-01 9.84346688e-01
-3.45322400e-01 -6.92085028e-01 -6.50549293e-01 2.70028651e-01
4.30940464e-02 2.73697406e-01 -4.62471902e-01 6.43381774e-01
3.09813410e-01 6.80758297e-01 2.85994321e-01 -3.05274755e-01
8.13450396e-01 -1.18245371e-01 8.95193219e-01 -7.47504056e-01
-1.76668555e-01 -7.82811642e-02 -1.93913743e-01 -8.43488693e-01
-2.53488034e-01 -8.39602888e-01 -1.07299495e+00 -3.21673542e-01
5.54533303e-02 -3.07860941e-01 5.71095407e-01 7.19163001e-01
6.84005380e-01 7.58746415e-02 1.49936646e-01 -1.27657068e+00
-3.61551791e-01 -6.66462302e-01 -7.19627857e-01 7.43025541e-01
4.97191668e-01 -9.30992603e-01 -3.14748585e-01 1.89305976e-01] | [6.912295341491699, -1.0810794830322266] |
fd7fca6b-35ae-4bf0-82e6-acd9ad13a21b | modelps-an-interactive-and-collaborative | 2105.08275 | null | https://arxiv.org/abs/2105.08275v3 | https://arxiv.org/pdf/2105.08275v3.pdf | ModelPS: An Interactive and Collaborative Platform for Editing Pre-trained Models at Scale | AI engineering has emerged as a crucial discipline to democratize deep neural network (DNN) models among software developers with a diverse background. In particular, altering these DNN models in the deployment stage posits a tremendous challenge. In this research, we propose and develop a low-code solution, ModelPS (an acronym for "Model Photoshop"), to enable and empower collaborative DNN model editing and intelligent model serving. The ModelPS solution embodies two transformative features: 1) a user-friendly web interface for a developer team to share and edit DNN models pictorially, in a low-code fashion, and 2) a model genie engine in the backend to aid developers in customizing model editing configurations for given deployment requirements or constraints. Our case studies with a wide range of deep learning (DL) models show that the system can tremendously reduce both development and communication overheads with improved productivity. | ['Yong Luo', 'Yonggang Wen', 'Fan Yang', 'Shanshan Jiang', 'Huaizheng Zhang', 'Yuanming Li'] | 2021-05-18 | null | null | null | null | ['model-editing'] | ['natural-language-processing'] | [-1.51638076e-01 -6.25711530e-02 1.55044362e-01 -5.37623048e-01
1.37155995e-01 -7.68278837e-01 3.03342640e-01 -2.93662071e-01
-2.05847561e-01 1.00417092e-01 -2.97528327e-01 -5.99341750e-01
-1.75755352e-01 -5.58236539e-01 -5.64041495e-01 3.75233404e-02
4.93979901e-01 3.28988045e-01 -3.05574443e-02 -3.35972577e-01
1.91519573e-01 8.27686548e-01 -1.46496809e+00 1.10878497e-01
9.32274163e-01 7.32754052e-01 4.21131939e-01 6.20821357e-01
-4.40149128e-01 1.20291924e+00 -8.31388772e-01 -5.49823403e-01
5.08159459e-01 -9.42839161e-02 -5.74982464e-01 -3.96139890e-01
5.23933291e-01 -5.76155722e-01 -1.17232993e-01 1.16701317e+00
6.68748081e-01 -5.00681937e-01 1.18818305e-01 -1.62830389e+00
-7.36175716e-01 9.27403808e-01 -3.71210873e-01 1.48277208e-01
-1.79141700e-01 3.17182600e-01 3.34198624e-01 -6.58805847e-01
4.33066398e-01 9.75038052e-01 8.88257086e-01 1.00513148e+00
-1.09417462e+00 -1.13393736e+00 1.56751379e-01 1.21952109e-01
-1.32121801e+00 -6.95269525e-01 9.13169324e-01 -6.47540092e-01
1.39554262e+00 1.79997116e-01 1.02433336e+00 1.16694438e+00
3.66079122e-01 4.62621838e-01 2.73753345e-01 -5.45201540e-01
3.49562168e-01 1.31737769e-01 1.62592337e-01 5.09727657e-01
5.58701575e-01 -1.82043150e-01 -3.85408849e-01 -7.78455660e-02
9.95372117e-01 1.46197394e-01 2.79381514e-01 -3.47556472e-01
-6.06162488e-01 2.29386300e-01 1.45164400e-01 2.58217365e-01
-4.83483374e-01 8.06200266e-01 3.83748591e-01 3.30602586e-01
2.06623301e-01 7.09641516e-01 -7.76275873e-01 -7.59575427e-01
-8.19067061e-01 1.72967896e-01 9.59426641e-01 1.42522550e+00
7.67483950e-01 6.26278937e-01 4.35795099e-01 8.74646902e-01
3.71547788e-01 3.27572793e-01 3.47244710e-01 -1.15483773e+00
3.75528514e-01 9.06155586e-01 -1.44061267e-01 -8.39221835e-01
-2.29496896e-01 -4.67699766e-01 -6.12098098e-01 5.56707144e-01
-2.08858773e-01 -6.39558017e-01 -8.96019101e-01 1.65089118e+00
6.78563118e-02 -1.24611808e-02 -6.80091884e-03 4.39558208e-01
6.94997907e-01 3.95976931e-01 1.51093394e-01 5.21591067e-01
9.25296485e-01 -1.07479882e+00 -6.34141147e-01 -6.41322374e-01
7.68349588e-01 -5.35065472e-01 1.02419591e+00 4.31881577e-01
-1.14669991e+00 -5.43028593e-01 -1.16454780e+00 -2.03486621e-01
-4.11155552e-01 1.93780363e-01 6.81174099e-01 7.45933414e-01
-1.48927343e+00 5.21829128e-01 -1.18370724e+00 -4.58128601e-01
6.59092367e-01 5.24107099e-01 -8.15248713e-02 -7.42986798e-03
-6.62100375e-01 1.08992028e+00 2.07523391e-01 1.85162410e-01
-1.12961698e+00 -1.37387836e+00 -6.49641633e-01 3.27632278e-01
5.71197495e-02 -8.53950083e-01 1.82098103e+00 -1.42941439e+00
-1.57735801e+00 5.23994327e-01 3.91572922e-01 -2.78810292e-01
5.06592095e-01 -1.79454625e-01 -5.22318900e-01 -3.74773145e-01
-2.47399092e-01 7.01516747e-01 7.08585262e-01 -1.12646139e+00
-6.07721448e-01 -2.15925537e-02 5.25175273e-01 1.89595353e-02
-7.39478648e-01 3.77337873e-01 -5.88001907e-01 -6.56167448e-01
-4.70203817e-01 -6.95606411e-01 -3.28166366e-01 4.59458351e-01
-1.57339022e-01 9.31483880e-02 1.09993529e+00 -7.51778185e-01
1.52193880e+00 -2.13319397e+00 -2.33330186e-02 3.30700487e-01
7.87610352e-01 6.91489279e-01 -3.26836556e-01 4.09513026e-01
-2.26431206e-01 2.98164338e-01 1.07452743e-01 -2.85272360e-01
3.95713925e-01 2.77619779e-01 3.21025670e-01 -3.24493915e-01
-4.06817868e-02 8.81100953e-01 -6.60876751e-01 -2.59230196e-01
4.79823500e-02 6.12189472e-01 -7.81640470e-01 1.35925010e-01
-2.59842545e-01 -1.97380573e-01 -1.09398238e-01 5.65165162e-01
6.89110398e-01 -1.78337753e-01 2.85780668e-01 -1.64326146e-01
-3.32741529e-01 -5.12082763e-02 -9.85309243e-01 1.79569137e+00
-8.14562261e-01 1.10879731e+00 4.22739238e-01 -3.80792350e-01
9.96467888e-01 2.56084234e-01 1.58235922e-01 -6.77743614e-01
1.65974781e-01 -7.02314153e-02 2.51495272e-01 -5.91895282e-01
2.12397307e-01 4.75181729e-01 3.64681035e-02 6.71746075e-01
3.72329983e-03 -1.90941971e-02 3.92679900e-01 1.93984956e-02
1.31008101e+00 5.07969148e-02 7.15702996e-02 -3.08531463e-01
-5.40464558e-02 -7.78402537e-02 6.58642590e-01 6.32319093e-01
5.23856394e-02 2.53506571e-01 6.78662300e-01 -8.26933861e-01
-1.14722490e+00 -8.77223551e-01 4.09915954e-01 1.09523845e+00
-3.56179208e-01 -5.37275851e-01 -1.37345624e+00 -5.60403824e-01
-3.17799300e-01 1.01958108e+00 -5.21415770e-01 -4.12250698e-01
-4.24498230e-01 -1.64054602e-01 7.99891353e-01 8.86366427e-01
5.97780704e-01 -1.11814392e+00 -1.06512463e+00 2.39794344e-01
4.28220063e-01 -7.64791787e-01 -5.64721465e-01 3.61272603e-01
-6.86271966e-01 -7.41504431e-01 -3.95108938e-01 -1.03603876e+00
9.57457364e-01 1.60244733e-01 1.19581759e+00 3.44297171e-01
-3.11681658e-01 3.00524205e-01 -3.04057658e-01 -8.54409337e-01
-7.22325921e-01 3.31011713e-01 -5.67239784e-02 -5.65126359e-01
5.65972686e-01 -1.01075363e+00 -2.34799609e-01 8.04399326e-02
-1.17438805e+00 7.76583374e-01 4.62187856e-01 2.52753288e-01
2.10749894e-01 3.34892571e-02 4.98316526e-01 -9.16027665e-01
1.10251892e+00 -2.18563646e-01 -1.17695928e+00 6.77371860e-01
-8.63293529e-01 -1.21930256e-01 1.02348781e+00 -5.43016136e-01
-9.64181006e-01 4.36090268e-02 -2.64227718e-01 -4.71984118e-01
-1.89719751e-01 8.63377750e-01 -5.44721365e-01 -3.09029818e-01
8.27371061e-01 -1.04440719e-01 -2.92767249e-02 -7.72402227e-01
2.38243327e-01 7.62605429e-01 5.22399187e-01 -5.65547585e-01
7.81081438e-01 -1.16724782e-01 -5.40080369e-01 -3.76531154e-01
-9.53605548e-02 3.21179926e-01 -7.05632389e-01 -3.31301302e-01
4.92225766e-01 -8.50665689e-01 -3.99546593e-01 8.21777999e-01
-1.52845967e+00 -9.03262377e-01 -2.97489554e-01 7.54878223e-02
-1.24811055e-02 -1.80397958e-01 -3.85533154e-01 -2.76858658e-01
-7.63843715e-01 -1.24031091e+00 1.26120389e-01 6.30675972e-01
-3.68315607e-01 -1.01971340e+00 -1.41870603e-03 1.52103677e-01
1.11248493e+00 1.91344116e-02 1.21722865e+00 -5.03211439e-01
-4.06232834e-01 -3.24566156e-01 -4.30094421e-01 7.82054245e-01
1.57245010e-01 5.49374878e-01 -9.11780059e-01 -1.35840625e-01
-4.05346841e-01 2.41300017e-01 -3.17320436e-01 7.56088346e-02
1.41575170e+00 -2.69161552e-01 -2.80046552e-01 8.97703350e-01
1.40627921e+00 7.99160004e-01 5.42534411e-01 2.80737221e-01
1.16718388e+00 1.11198507e-01 -1.10075310e-01 6.01024628e-01
4.36314285e-01 4.23719347e-01 5.44000626e-01 -4.57376361e-01
-1.58194661e-01 -1.15188561e-01 2.09895551e-01 1.07805324e+00
8.35437551e-02 -1.60457075e-01 -1.52161503e+00 4.09020633e-01
-1.80361640e+00 -4.85273838e-01 1.63620517e-01 1.41519308e+00
8.62518132e-01 1.00232013e-01 -3.07105899e-01 -3.71762127e-01
4.22238499e-01 -2.85588503e-01 -1.04750848e+00 -7.02489913e-01
4.97140050e-01 4.18786854e-01 2.71257102e-01 1.23297922e-01
-5.06653488e-01 1.01190066e+00 6.41536760e+00 5.30209303e-01
-1.31738806e+00 1.69561535e-01 1.55875921e-01 -3.07307780e-01
-4.30015922e-01 -2.52992027e-02 -6.53760970e-01 4.09480393e-01
1.11198914e+00 -8.91396403e-01 8.40583861e-01 1.50027812e+00
4.24343407e-01 4.40615475e-01 -1.42993581e+00 1.25422263e+00
4.33032438e-02 -1.86663389e+00 3.43332499e-01 -7.62513131e-02
6.49003804e-01 1.87018991e-01 -1.25886545e-01 2.45843560e-01
6.33634746e-01 -9.38372850e-01 1.10918057e+00 6.54748797e-01
9.21139002e-01 -1.07986951e+00 5.97445309e-01 2.50688761e-01
-8.37981939e-01 -1.07938729e-01 -2.79352129e-01 -3.67064625e-01
1.02594502e-01 2.21774191e-01 -8.72368038e-01 1.97592974e-02
1.00886214e+00 8.11537266e-01 -7.57720292e-01 1.17197967e+00
-7.57624507e-02 3.93706530e-01 -2.01104715e-01 -1.03335574e-01
-1.62263483e-01 8.82207975e-02 -3.58121656e-02 1.20052993e+00
3.21394682e-01 -1.08211309e-01 -4.30043012e-01 1.35429358e+00
-3.30356389e-01 -3.16187680e-01 -4.14839506e-01 -2.84159303e-01
8.50533843e-01 1.39519811e+00 -5.17950177e-01 -1.77049041e-01
-4.09684658e-01 7.47684360e-01 2.52156913e-01 4.91748631e-01
-1.07078922e+00 -8.05259109e-01 1.05790234e+00 1.40615448e-01
1.76803306e-01 -3.42821300e-01 -3.62418205e-01 -7.32353330e-01
8.70918930e-02 -1.08963788e+00 -5.11080801e-01 -1.09093547e+00
-8.53338003e-01 9.11012888e-01 7.80648217e-02 -9.31916058e-01
-1.35628611e-01 -5.08683920e-01 -6.85929120e-01 8.43686104e-01
-1.12216616e+00 -1.17789400e+00 -7.60552227e-01 3.43981266e-01
3.60176831e-01 -4.41680461e-01 9.67402577e-01 9.39288735e-01
-1.06350303e+00 8.97783875e-01 8.35024267e-02 1.74139738e-01
2.20203176e-01 -9.84278500e-01 1.16801369e+00 1.13051760e+00
-1.89436391e-01 1.02297044e+00 5.54344952e-01 -3.86699826e-01
-1.30018926e+00 -1.22525454e+00 6.13083303e-01 -4.21423942e-01
4.18285817e-01 -8.76436353e-01 -5.21291494e-01 1.06975865e+00
4.20204043e-01 -2.52683342e-01 8.34864736e-01 -3.36822301e-01
-2.12704465e-01 -4.83426392e-01 -1.04518998e+00 9.21975970e-01
1.23003316e+00 -5.63245535e-01 -5.87708689e-02 8.03628564e-02
8.24615479e-01 -4.96437192e-01 -8.82091522e-01 -1.04424953e-01
6.20265543e-01 -8.67167056e-01 6.23659611e-01 -5.47411919e-01
4.11558777e-01 -3.60254705e-01 1.28620625e-01 -1.33659017e+00
-4.20112103e-01 -8.99662018e-01 -1.85947672e-01 1.35426307e+00
6.64472163e-01 -5.13029754e-01 8.16688120e-01 1.49800777e+00
-6.66382134e-01 -6.17046714e-01 -5.24452865e-01 -5.61987579e-01
-2.66824931e-01 -8.55130434e-01 1.26443815e+00 9.10010993e-01
-3.28639179e-01 1.65205047e-01 -5.56281880e-02 6.76934272e-02
1.92848504e-01 -6.45398855e-01 1.00882518e+00 -1.38074076e+00
-2.20583409e-01 -5.60450375e-01 -3.61192703e-01 -1.09317565e+00
1.00249447e-01 -6.08530998e-01 -2.58542120e-01 -1.83302832e+00
3.87306847e-02 -5.97503364e-01 -1.32511154e-01 1.04748309e+00
7.23610461e-01 -3.18284869e-01 3.26656908e-01 -2.74667260e-03
-5.07302582e-01 1.20823547e-01 8.19492996e-01 -9.49755162e-02
-2.33219981e-01 -1.82022095e-01 -1.27931976e+00 9.13922846e-01
6.86060250e-01 -7.45495081e-01 -8.76770675e-01 -1.50918150e+00
5.17230451e-01 -4.86998022e-01 2.29433909e-01 -1.42754579e+00
6.64422274e-01 -1.14273190e-01 7.41462782e-02 -3.18640172e-01
-4.68417294e-02 -1.30355775e+00 8.02296877e-01 3.77360076e-01
-1.89276338e-01 6.40042305e-01 6.48118913e-01 -4.83047031e-02
7.56924748e-02 -5.35081446e-01 6.12404883e-01 1.33077279e-01
-1.06437421e+00 3.63477260e-01 -4.96478349e-01 -2.27491692e-01
1.10661995e+00 -5.68297803e-01 -5.45833886e-01 -1.68594256e-01
-2.91310847e-01 1.83442160e-01 8.52246881e-01 6.56455338e-01
5.52816153e-01 -1.16921806e+00 -1.47745296e-01 5.59745252e-01
9.41518322e-02 2.10104421e-01 3.25970471e-01 3.24795812e-01
-1.38070107e+00 1.15266806e-02 -7.23933578e-01 -1.41618386e-01
-1.24375784e+00 1.90694526e-01 8.19069564e-01 5.37890417e-04
-4.82792646e-01 1.12813282e+00 -2.62153987e-02 -6.05301917e-01
4.23305482e-01 -6.79286301e-01 1.64114907e-01 -3.27121884e-01
6.38135970e-01 3.60386550e-01 2.40472332e-01 7.23047107e-02
-3.12967390e-01 1.47231951e-01 -5.61381340e-01 1.14175245e-01
1.75803781e+00 -6.08213886e-04 -3.07818502e-01 1.19061485e-01
7.96939969e-01 -5.51617920e-01 -1.76371884e+00 7.28173777e-02
-4.32714485e-02 -1.48462877e-02 2.97129273e-01 -1.32092762e+00
-1.47893023e+00 8.22584033e-01 6.74835682e-01 1.36890501e-01
1.19887185e+00 -3.52054268e-01 7.32300341e-01 5.13190746e-01
2.89394677e-01 -1.27149439e+00 -1.62552804e-01 4.91550624e-01
8.71611416e-01 -5.63435197e-01 -2.38438070e-01 1.52500391e-01
-3.91885728e-01 1.34062314e+00 1.15499794e+00 2.47399062e-01
9.20023203e-01 9.40216064e-01 4.41468149e-01 -3.52809638e-01
-9.65402722e-01 7.26263940e-01 -3.32198173e-01 1.06587505e+00
2.65187234e-01 -1.08561747e-01 3.71550500e-01 1.17208898e+00
-2.29112253e-01 5.15259326e-01 7.08973587e-01 1.12992227e+00
-2.70248037e-02 -1.21499813e+00 6.51989430e-02 6.04785919e-01
-9.33412910e-02 -3.67194206e-01 -4.51704502e-01 8.04506302e-01
4.22437429e-01 6.95369959e-01 1.80512354e-01 -7.65276134e-01
7.24999845e-01 -2.51892209e-01 2.42490798e-01 -8.01639795e-01
-9.47675765e-01 -3.85814011e-01 2.25377932e-01 -5.29777110e-01
1.25792325e-01 -1.87930003e-01 -1.13177729e+00 -7.01958060e-01
-2.33448986e-02 -2.89130151e-01 1.03754270e+00 8.52316558e-01
1.00123489e+00 7.82506704e-01 9.72059742e-02 -1.01091492e+00
-2.80963242e-01 -7.70389080e-01 -5.69484770e-01 -1.71517938e-01
5.50347231e-02 -2.98084289e-01 8.25822353e-02 4.72102553e-01] | [7.927487850189209, 7.620371341705322] |
4eca24a2-f36c-406d-ba81-2f61c1d2deb9 | panoptic-video-scene-graph-generation | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Yang_Panoptic_Video_Scene_Graph_Generation_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Yang_Panoptic_Video_Scene_Graph_Generation_CVPR_2023_paper.pdf | Panoptic Video Scene Graph Generation | Towards building comprehensive real-world visual perception systems, we propose and study a new problem called panoptic scene graph generation (PVSG). PVSG is related to the existing video scene graph generation (VidSGG) problem, which focuses on temporal interactions between humans and objects localized with bounding boxes in videos. However, the limitation of bounding boxes in detecting non-rigid objects and backgrounds often causes VidSGG systems to miss key details that are crucial for comprehensive video understanding. In contrast, PVSG requires nodes in scene graphs to be grounded by more precise, pixel-level segmentation masks, which facilitate holistic scene understanding. To advance research in this new area, we contribute a high-quality PVSG dataset, which consists of 400 videos (289 third-person + 111 egocentric videos) with totally 150K frames labeled with panoptic segmentation masks as well as fine, temporal scene graphs. We also provide a variety of baseline methods and share useful design practices for future work. | ['Ziwei Liu', 'Chen Change Loy', 'Wayne Zhang', 'Kaiyang Zhou', 'Zheng Ma', 'Bo Li', 'Liangyu Chen', 'Zujin Guo', 'Xiangtai Li', 'Wenxuan Peng', 'Jingkang Yang'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['panoptic-segmentation', 'scene-graph-generation', 'video-understanding', 'panoptic-scene-graph-generation', 'scene-understanding'] | ['computer-vision', 'computer-vision', 'computer-vision', 'computer-vision', 'computer-vision'] | [ 4.23568100e-01 -1.09811488e-03 2.17639860e-02 -2.49503717e-01
-3.06610763e-01 -7.22315073e-01 6.17840469e-01 -2.33495444e-01
3.58449034e-02 3.96504968e-01 5.26638567e-01 2.95039285e-02
1.37541100e-01 -5.38078785e-01 -8.66956294e-01 -4.35097128e-01
-1.14874721e-01 -1.09508000e-02 6.91871762e-01 -8.66568610e-02
1.05076052e-01 2.11122215e-01 -1.43420970e+00 6.31048083e-01
5.36105573e-01 5.51223516e-01 2.87057549e-01 9.84766126e-01
2.86204100e-01 9.43152905e-01 -5.07498503e-01 -5.47809124e-01
4.65635419e-01 -6.06904924e-01 -8.34666789e-01 7.29727805e-01
1.14569461e+00 -5.43545187e-01 -8.17916036e-01 1.09050441e+00
1.53538331e-01 5.89744389e-01 5.33084869e-01 -1.70553422e+00
-4.84130621e-01 4.58423585e-01 -8.35857749e-01 5.02877176e-01
8.07268858e-01 5.75363338e-01 9.51340795e-01 -6.89498186e-01
1.11541033e+00 1.63111067e+00 4.91540968e-01 4.06671524e-01
-1.27822530e+00 -4.12411094e-01 7.26491570e-01 4.37936395e-01
-1.18649268e+00 -6.12206280e-01 7.70327210e-01 -6.66632712e-01
1.04965591e+00 3.43491912e-01 1.12014997e+00 1.24804878e+00
1.76439956e-01 9.76977110e-01 8.48317742e-01 -2.03936502e-01
7.08001480e-02 -2.53213167e-01 4.47526872e-02 7.33474016e-01
3.56389165e-01 2.02042684e-02 -6.56710386e-01 1.69997543e-01
1.21858442e+00 -1.89135335e-02 -4.00311321e-01 -7.16994286e-01
-1.52538097e+00 4.31863934e-01 2.94871658e-01 -6.76053315e-02
-3.48156810e-01 3.55140507e-01 1.96305171e-01 -5.00834212e-02
2.04274759e-01 3.70582610e-01 -8.90774503e-02 -6.95549995e-02
-9.30066943e-01 4.60116148e-01 6.31143391e-01 1.39092398e+00
7.25340545e-01 2.63945997e-01 -4.07530040e-01 4.84524846e-01
1.00299992e-01 4.56619024e-01 -2.77401626e-01 -1.65005267e+00
5.11099815e-01 5.49942195e-01 4.10106890e-02 -1.57037759e+00
-3.42939109e-01 6.42427504e-02 -5.74918926e-01 -2.54508287e-01
5.05513668e-01 -1.47878632e-01 -1.04077113e+00 1.59699893e+00
4.06431764e-01 4.49031681e-01 -1.28739879e-01 1.13224101e+00
1.20005190e+00 8.25531781e-01 2.41427034e-01 -2.64923900e-01
1.53924632e+00 -1.30207419e+00 -6.57300949e-01 -4.95762825e-01
3.24967206e-01 -5.50260723e-01 1.14982414e+00 3.65367144e-01
-1.17232895e+00 -6.53648078e-01 -6.82405055e-01 -2.65969068e-01
-1.82857603e-01 -1.51302934e-01 7.83198237e-01 4.70506459e-01
-1.45781422e+00 1.33844733e-01 -6.15010142e-01 -9.05010104e-01
4.47880864e-01 -7.16060176e-02 -4.88594979e-01 -3.88029993e-01
-8.80124271e-01 3.71536016e-01 5.51515818e-01 -6.72647879e-02
-1.20890403e+00 -6.14249825e-01 -1.26299059e+00 -1.25919059e-01
9.07029152e-01 -8.29369426e-01 1.10940540e+00 -9.06684518e-01
-1.03937256e+00 8.79016221e-01 -2.30237082e-01 -2.07779139e-01
4.26273406e-01 -2.33829483e-01 -3.39630514e-01 7.88810611e-01
1.96793884e-01 1.16664028e+00 8.86252642e-01 -1.42579985e+00
-6.20706320e-01 -3.15714888e-02 4.05791759e-01 5.96691906e-01
1.38577884e-02 2.17465416e-01 -1.22977889e+00 -9.59139645e-01
-8.31607580e-02 -8.56022418e-01 -4.21584547e-01 -1.82054833e-01
-8.27128589e-01 7.60568446e-03 9.98738110e-01 -8.59552860e-01
1.11620295e+00 -2.06920791e+00 2.39800006e-01 1.66048128e-02
6.12021446e-01 -1.36902004e-01 -3.70363116e-01 3.95202875e-01
-1.80087835e-01 -1.39846560e-02 -3.09145227e-02 -2.33309716e-01
-1.60532579e-01 1.21567167e-01 -3.21000814e-01 2.72486836e-01
-3.15592177e-02 1.24223340e+00 -1.15289927e+00 -8.46980512e-01
5.46333134e-01 1.91074550e-01 -9.13293362e-01 7.96199664e-02
-4.11959022e-01 2.98041940e-01 -3.92795920e-01 8.94359529e-01
5.29326916e-01 -4.51747149e-01 1.69025704e-01 -5.20953357e-01
1.94465831e-01 -1.64979473e-01 -1.19988656e+00 1.67546260e+00
4.95624602e-01 9.80533838e-01 -1.25489011e-01 -7.00169683e-01
3.09736669e-01 8.29695433e-04 7.61394083e-01 -4.45478797e-01
-5.11454120e-02 -7.07571805e-01 -4.43741798e-01 -5.96047580e-01
9.30024445e-01 4.28949177e-01 -6.22394085e-02 2.41092995e-01
1.49351910e-01 -3.42764258e-01 6.62963092e-01 8.56868207e-01
1.03857958e+00 3.77955407e-01 2.69511074e-01 -2.15463594e-01
7.23078847e-02 3.67514521e-01 5.78256428e-01 9.17738855e-01
-6.12903655e-01 9.81283724e-01 6.42523825e-01 -3.05553943e-01
-9.95847762e-01 -1.24570692e+00 3.93012643e-01 1.00503111e+00
6.58688784e-01 -9.01947021e-01 -1.01072574e+00 -5.68722665e-01
-2.45623007e-01 6.18885338e-01 -4.26147878e-01 2.02630565e-01
-5.61477184e-01 -5.16207039e-01 1.50232211e-01 5.79192400e-01
7.33205736e-01 -1.15731549e+00 -4.46967810e-01 3.00616715e-02
-7.61891186e-01 -1.82403600e+00 -9.33999002e-01 -5.04729688e-01
-5.74611187e-01 -1.44343114e+00 -6.18129373e-01 -5.39797604e-01
7.87453592e-01 1.08102798e+00 1.44149613e+00 -2.69623217e-03
-4.71061528e-01 1.01114511e+00 -5.50212741e-01 -6.22446947e-02
-5.88545538e-02 -4.64897275e-01 1.22695416e-01 -2.40585534e-03
1.98115006e-01 -2.64078587e-01 -8.08389723e-01 4.58659887e-01
-7.48715162e-01 7.45933115e-01 2.37112984e-01 2.88823187e-01
5.58238328e-01 2.01360390e-01 3.41063598e-03 -8.75772953e-01
1.88229382e-01 -2.99273401e-01 -4.76555765e-01 3.37106884e-01
7.78077021e-02 -6.53719366e-01 3.30355793e-01 -2.72488952e-01
-1.19224477e+00 7.52794463e-03 3.62723678e-01 -5.83173633e-01
-2.72835135e-01 1.65754687e-02 -1.19635336e-01 -5.84388338e-02
6.57437265e-01 2.86531389e-01 -3.98773044e-01 5.69240116e-02
8.50970805e-01 -9.93508697e-02 9.29067373e-01 -4.61974531e-01
8.32557201e-01 5.77018201e-01 -2.04376683e-01 -1.14928770e+00
-7.35722125e-01 -6.60612881e-01 -7.68799007e-01 -7.12606609e-01
1.15133786e+00 -1.14778614e+00 -6.09774947e-01 3.88340324e-01
-9.90450323e-01 -7.62121081e-01 -3.40862066e-01 2.74277985e-01
-8.94895315e-01 8.54317009e-01 -6.07440710e-01 -5.85824013e-01
1.68278500e-01 -9.66903448e-01 1.32310092e+00 2.70614833e-01
-3.11557025e-01 -8.31214070e-01 -3.04779917e-01 5.55818737e-01
-2.29376003e-01 4.84431982e-01 5.27953088e-01 -1.20672733e-01
-1.12974226e+00 2.16903746e-01 -5.32109261e-01 4.29214239e-02
3.73526774e-02 2.34670565e-01 -6.91870391e-01 -4.44176227e-01
-3.05600017e-01 -2.24347770e-01 7.91758955e-01 7.36589909e-01
1.14618182e+00 -3.52735400e-01 -5.31091154e-01 8.11904430e-01
1.16606081e+00 2.85965323e-01 7.88365722e-01 1.72413468e-01
1.27558732e+00 7.00566173e-01 7.51930177e-01 3.01565319e-01
7.70348251e-01 7.26210535e-01 1.18518963e-01 -3.36063296e-01
-5.57462096e-01 -4.61261690e-01 6.47096515e-01 4.81648862e-01
-4.02671129e-01 -5.86633027e-01 -7.90230989e-01 4.85559911e-01
-1.81291950e+00 -1.36170864e+00 -1.81919858e-01 1.67156291e+00
4.53657895e-01 5.23188524e-03 4.09361631e-01 -2.04511806e-01
7.24860787e-01 5.36228299e-01 -3.49358469e-01 1.37938485e-01
-3.86611938e-01 -4.34969783e-01 3.11590910e-01 3.88561487e-01
-1.35635900e+00 1.41507280e+00 7.40288544e+00 6.47565961e-01
-5.43577254e-01 -2.40738064e-01 7.72556663e-01 -6.67494535e-02
-1.08708054e-01 8.68698955e-02 -7.07048297e-01 2.38912284e-01
3.21017146e-01 -2.63241738e-01 5.48248112e-01 8.27212393e-01
4.58192080e-01 -4.45116371e-01 -1.09882784e+00 1.36252356e+00
3.57736200e-01 -1.33846533e+00 3.99488062e-01 -1.41179115e-01
1.07250082e+00 -1.88159302e-01 -1.74562275e-01 4.70608324e-02
5.54681718e-01 -7.12568700e-01 7.46987879e-01 5.12023091e-01
7.22447574e-01 -3.39241564e-01 1.14352234e-01 -1.13127857e-01
-1.63168132e+00 6.99782073e-02 -3.03972244e-01 -1.80334654e-02
5.17291963e-01 1.57949165e-01 -6.07934356e-01 6.43229902e-01
1.04111242e+00 1.20809901e+00 -8.96142483e-01 1.02366996e+00
-1.81560263e-01 5.83558738e-01 -2.18478635e-01 3.21061432e-01
2.01382324e-01 -3.86248589e-01 7.69041240e-01 1.44112182e+00
3.38318348e-02 4.47919816e-01 5.94429195e-01 9.04806197e-01
2.28445604e-01 -7.96523243e-02 -9.40623343e-01 -2.16195375e-01
2.73172170e-01 1.22136247e+00 -1.35922718e+00 -7.47065842e-01
-4.63550866e-01 1.12506104e+00 -1.63865045e-01 1.02475297e+00
-9.00805652e-01 7.62224644e-02 7.10141778e-01 2.15196297e-01
1.40827239e-01 -6.46135449e-01 6.20311350e-02 -1.60112453e+00
-1.72377884e-01 -1.06524611e+00 6.13956749e-01 -1.33777380e+00
-1.11974454e+00 3.94205540e-01 5.92927635e-01 -1.09034157e+00
-2.72799909e-01 -4.43018019e-01 -4.36255932e-01 2.54348278e-01
-1.01325214e+00 -1.25047779e+00 -8.77366245e-01 1.01158726e+00
1.05129862e+00 1.96522683e-01 1.65874630e-01 1.62791952e-01
-5.90890348e-01 1.88596517e-01 -4.57761049e-01 2.49075428e-01
6.38902903e-01 -1.09046340e+00 8.46475363e-01 1.38844776e+00
3.49260896e-01 4.81591463e-01 7.33984709e-01 -1.02310085e+00
-1.36271119e+00 -1.35866690e+00 4.20255929e-01 -8.83664489e-01
3.99099678e-01 -5.86792111e-01 -5.65038025e-01 1.05195272e+00
2.85323203e-01 -4.11942750e-02 1.21046357e-01 -1.08909003e-01
-1.10416032e-01 1.96875051e-01 -6.63416803e-01 1.18706310e+00
1.75862265e+00 -4.33597147e-01 -4.63871419e-01 5.00514328e-01
9.56642091e-01 -5.38694263e-01 -5.00305772e-01 3.61315459e-01
4.03870165e-01 -1.12306976e+00 1.38496625e+00 -6.22920334e-01
3.87995213e-01 -5.42174578e-01 -2.15735689e-01 -9.77695525e-01
-4.76791918e-01 -9.28914964e-01 -1.86609358e-01 1.05883563e+00
-2.84637868e-01 -1.72995657e-01 8.53270531e-01 8.38605285e-01
-2.11567014e-01 -2.01029181e-01 -2.56745130e-01 -7.05906451e-01
-7.47529030e-01 -6.22131824e-01 2.20123231e-01 1.05862999e+00
-1.66410029e-01 1.55922532e-01 -5.90212762e-01 1.23676732e-02
8.64769161e-01 -2.77501792e-02 1.33862054e+00 -6.85653746e-01
-1.59245893e-01 -3.67061317e-01 -6.49553478e-01 -1.38810396e+00
-1.96323410e-01 -4.16765392e-01 3.40782702e-02 -1.71483481e+00
5.25628030e-01 8.73488933e-02 3.78772616e-02 4.39897716e-01
-3.88169467e-01 6.67230129e-01 5.26102126e-01 -1.61070041e-02
-1.29524624e+00 3.84672523e-01 1.67631316e+00 -1.61104687e-02
-3.05649102e-01 -4.21659529e-01 -7.12193072e-01 9.12229121e-01
1.79642841e-01 4.03346634e-03 -8.30970705e-01 -2.19414607e-01
-1.00572206e-01 1.80932134e-01 8.15860093e-01 -1.00838625e+00
3.39646935e-01 -5.86089909e-01 5.98666370e-01 -8.89170289e-01
5.57945728e-01 -4.17516947e-01 3.70218098e-01 1.28292397e-01
-3.06270719e-02 2.31189534e-01 1.40781134e-01 7.89519548e-01
-8.66206959e-02 5.30323565e-01 4.98560548e-01 -2.91033715e-01
-1.60785496e+00 4.36968505e-01 -3.87473404e-01 2.27536693e-01
1.29817319e+00 -8.48834515e-01 -6.34439766e-01 -7.79087663e-01
-8.00852835e-01 4.70001757e-01 6.96097136e-01 6.78781092e-01
8.65377069e-01 -1.19983339e+00 -5.69420695e-01 6.44526407e-02
1.77238852e-01 -2.83014625e-02 7.69333720e-01 6.98670506e-01
-7.51361668e-01 3.70913982e-01 -3.14058632e-01 -9.03124988e-01
-1.43825686e+00 8.43330979e-01 1.59973487e-01 3.25927198e-01
-1.16970766e+00 8.77857327e-01 1.11786401e+00 2.03580707e-01
6.38042688e-02 -4.17178869e-01 -2.14267284e-01 -4.49970178e-02
5.36616147e-01 3.62064958e-01 -6.37024283e-01 -9.43718910e-01
-2.34806180e-01 5.78593433e-01 9.06924829e-02 -3.85971852e-02
8.97121549e-01 -5.92423499e-01 7.59577900e-02 2.21643403e-01
6.82292163e-01 -1.27899855e-01 -1.76236856e+00 -1.89816505e-01
-3.40490907e-01 -8.54095638e-01 -2.83219188e-01 -4.30521041e-01
-1.12018514e+00 4.47441846e-01 1.75187469e-01 1.77723497e-01
1.25978947e+00 2.01631516e-01 6.32123232e-01 3.22938323e-01
4.90727931e-01 -9.46553826e-01 6.84664011e-01 3.32017034e-01
9.43165064e-01 -1.19431055e+00 3.17549348e-01 -9.56707358e-01
-9.39910710e-01 8.54247451e-01 9.80867386e-01 -4.76741008e-02
2.85244465e-01 -3.56996059e-02 1.59113221e-02 -5.11487603e-01
-6.03546321e-01 -4.30579185e-01 6.35123789e-01 9.20562923e-01
-5.10900803e-02 -3.05007808e-02 2.65055031e-01 3.51229787e-01
-2.54153013e-01 -1.88015819e-01 6.16125882e-01 5.96258283e-01
-2.14934662e-01 -5.72142959e-01 -3.83561999e-01 3.96436661e-01
-8.27013552e-02 -1.35986596e-01 -4.21159774e-01 1.02437735e+00
-6.96330890e-02 1.20835233e+00 1.11655101e-01 -4.76805866e-01
2.75924832e-01 -4.59056616e-01 5.88195562e-01 -6.59482956e-01
-3.68768126e-02 3.73921454e-01 3.61047179e-01 -1.24192333e+00
-6.91124260e-01 -8.25635552e-01 -9.97830331e-01 -3.98898512e-01
1.55382052e-01 -3.04249555e-01 6.23525418e-02 7.21342623e-01
2.84961522e-01 6.95283055e-01 1.40529379e-01 -1.20367432e+00
4.16571647e-01 -6.44555628e-01 -6.19960129e-01 8.08310807e-01
1.07325323e-01 -8.22988749e-01 1.29440790e-02 8.82821321e-01] | [9.570330619812012, -0.1118595078587532] |
2ec80593-29f8-4fa2-9499-70ef93e467f1 | towards-a-unified-compositional-model-for | null | null | http://openaccess.thecvf.com/content_iccv_2017/html/Tang_Towards_a_Unified_ICCV_2017_paper.html | http://openaccess.thecvf.com/content_ICCV_2017/papers/Tang_Towards_a_Unified_ICCV_2017_paper.pdf | Towards a Unified Compositional Model for Visual Pattern Modeling | Compositional models represent visual patterns as hierarchies of meaningful and reusable parts. They are attractive to vision modeling due to their ability to decompose complex patterns into simpler ones and resolve the low-level ambiguities in high-level image interpretations. However, current compositional models separate structure and part discovery from parameter estimation, which generally leads to suboptimal learning and fitting of the model. Moreover, the commonly adopted latent structural learning is not scalable for deep architectures. To address these difficult issues for compositional models, this paper quests for a unified framework for compositional pattern modeling, inference and learning. Represented by And-Or graphs (AOGs), it jointly models the compositional structure, parts, features, and composition/sub-configuration relationships. We show that the inference algorithm of the proposed framework is equivalent to a feed-forward network. Thus, all the parameters can be learned efficiently via the highly-scalable back-propagation (BP) in an end-to-end fashion. We validate the model via the task of handwritten digit recognition. By visualizing the processes of bottom-up composition and top-down parsing, we show that our model is fully interpretable, being able to learn the hierarchical compositions from visual primitives to visual patterns at increasingly higher levels. We apply this new compositional model to natural scene character recognition and generic object detection. Experimental results have demonstrated its effectiveness.
| ['Jiahuan Zhou', 'Wei Tang', 'Ying Wu', 'Pei Yu'] | 2017-10-01 | null | null | null | iccv-2017-10 | ['handwritten-digit-recognition'] | ['computer-vision'] | [ 5.55640876e-01 1.92597762e-01 -1.52685165e-01 -5.20097017e-01
-1.61683872e-01 -5.07510900e-01 7.29527950e-01 8.91058054e-03
1.71738967e-01 2.58053452e-01 1.78824529e-01 -4.07485396e-01
-5.34069166e-02 -7.10144222e-01 -1.06084502e+00 -6.80641413e-01
8.85112211e-02 6.47897422e-01 3.11411232e-01 1.59698233e-01
2.52390504e-01 7.08996773e-01 -1.79905808e+00 9.12581801e-01
6.08351767e-01 9.96054769e-01 4.01974916e-01 6.41121626e-01
-4.48156416e-01 1.29825759e+00 -1.57313347e-01 -5.53201914e-01
3.99240218e-02 -3.21479350e-01 -7.42369771e-01 7.64377594e-01
9.97352302e-01 -3.29986185e-01 -5.65381013e-02 1.03341413e+00
-1.49553552e-01 -4.02961463e-01 6.21359408e-01 -1.25579476e+00
-8.76148820e-01 8.42790067e-01 -4.00215805e-01 -4.05414045e-01
1.04567721e-01 7.68791959e-02 1.36927283e+00 -1.16293907e+00
5.12755573e-01 1.60793042e+00 6.32377625e-01 4.46493655e-01
-1.58777893e+00 -2.23332033e-01 9.02984142e-01 2.06379086e-01
-9.57387447e-01 -3.17894548e-01 8.88416708e-01 -6.97454333e-01
9.30938721e-01 2.77313471e-01 8.54115546e-01 1.05502462e+00
-1.32682517e-01 1.17972791e+00 1.01185429e+00 -5.13818681e-01
2.65626758e-01 -1.20436832e-01 4.91611391e-01 1.18651712e+00
3.36678326e-01 -1.80298999e-01 -6.46189332e-01 -9.43314806e-02
9.68219578e-01 2.33495921e-01 -1.92351639e-02 -7.66190946e-01
-8.89880180e-01 5.86201549e-01 5.97918272e-01 -3.47489081e-02
-2.79737681e-01 5.38238764e-01 2.05938127e-02 5.28518409e-02
3.12357266e-02 1.92980483e-01 -3.29465568e-01 5.26236236e-01
-6.18397176e-01 2.02985331e-01 8.55920672e-01 1.13567173e+00
8.43550682e-01 9.12592486e-02 4.72298674e-02 8.37681830e-01
6.37291193e-01 2.55805790e-01 1.40092850e-01 -8.53779733e-01
2.75521040e-01 1.03436375e+00 -1.51105002e-01 -9.75036383e-01
-2.61654705e-01 -5.38352132e-01 -9.15801406e-01 2.92645782e-01
3.41859013e-01 3.95510882e-01 -1.16481400e+00 1.71363902e+00
5.40402755e-02 -1.49878524e-02 -1.82852656e-01 8.46695065e-01
6.77910566e-01 6.72120631e-01 2.30006367e-01 9.88960937e-02
1.72645271e+00 -1.27044261e+00 -4.47960496e-01 -6.92309380e-01
7.26843858e-03 -5.09126961e-01 1.19735658e+00 4.69837606e-01
-1.26419020e+00 -8.23106766e-01 -1.18828511e+00 -3.07195544e-01
-1.49074942e-01 3.80481511e-01 7.72879601e-01 2.85106659e-01
-9.72102225e-01 4.32899326e-01 -9.44496334e-01 -2.44684368e-01
5.46424568e-01 3.76111239e-01 -2.59108812e-01 -3.40864994e-02
-3.38742316e-01 5.75035095e-01 4.16033328e-01 4.46370721e-01
-1.13529992e+00 -5.63502371e-01 -9.57507610e-01 2.25352585e-01
4.96733040e-01 -1.17401719e+00 1.27245557e+00 -1.21716511e+00
-1.53734112e+00 8.21867406e-01 -4.59129781e-01 -4.67718959e-01
3.03744316e-01 -2.02571958e-01 2.05419809e-01 1.45928472e-01
-1.36771813e-01 7.04247296e-01 1.11071277e+00 -1.75092995e+00
-7.22693563e-01 -4.06565756e-01 1.27277702e-01 7.35008791e-02
-9.95973498e-02 -3.32190841e-02 -6.40010059e-01 -5.78786135e-01
6.51410639e-01 -6.26631618e-01 -1.96146801e-01 2.95076102e-01
-4.58004951e-01 -2.62182741e-03 5.91018796e-01 -5.58343530e-01
9.22826171e-01 -1.99508071e+00 6.21758461e-01 9.54010412e-02
3.54220897e-01 -7.96405226e-02 -8.87552276e-02 4.29843068e-01
-1.95706952e-02 7.09784478e-02 -6.86713398e-01 -7.01895177e-01
2.80972987e-01 5.77715456e-01 -5.55869460e-01 1.62922993e-01
5.69620788e-01 1.15659463e+00 -5.68706393e-01 -3.30578893e-01
2.89087147e-01 2.57671446e-01 -5.94227910e-01 3.17137569e-01
-6.76330745e-01 1.86996147e-01 -9.40435901e-02 8.56370211e-01
6.46525443e-01 -5.66408932e-01 5.67882478e-01 -3.79726440e-01
-7.52257109e-02 3.12032670e-01 -1.20460701e+00 1.47848642e+00
-3.61767143e-01 3.67051512e-01 4.15025391e-02 -1.07235789e+00
8.25850010e-01 -1.22974463e-01 -1.55222505e-01 -4.46006417e-01
-2.19220176e-01 7.98155069e-02 -5.66707999e-02 -4.74270970e-01
1.27974823e-01 -1.93847120e-01 -1.04743764e-01 2.46035695e-01
1.43650949e-01 3.30161862e-02 1.22174956e-01 2.17137281e-02
8.16361010e-01 4.52222615e-01 4.24500942e-01 -5.96147738e-02
6.69163823e-01 -5.05238622e-02 4.00215566e-01 7.89586127e-01
4.31402117e-01 4.85551447e-01 7.02650905e-01 -7.93931425e-01
-1.06388199e+00 -1.30595756e+00 1.27393559e-01 1.23337984e+00
1.08355507e-01 -4.87462789e-01 -5.17362714e-01 -3.89023453e-01
6.76429132e-03 5.22412598e-01 -7.44416416e-01 1.53018638e-01
-9.77528036e-01 -6.62351966e-01 3.92156780e-01 7.61560321e-01
4.04176503e-01 -1.26077032e+00 -5.66247582e-01 1.18118472e-01
5.71226217e-02 -1.28005064e+00 -8.96581337e-02 3.39017421e-01
-1.09246349e+00 -9.84311700e-01 -1.99370831e-01 -1.19522333e+00
1.07826459e+00 1.26354873e-01 1.21267164e+00 1.45634443e-01
-3.54718417e-01 2.68825442e-01 -8.68814364e-02 -5.36122583e-02
-5.48302889e-01 -3.16087365e-01 -4.55819458e-01 3.72788757e-01
4.94597256e-02 -7.68532753e-01 -3.67293358e-01 1.05774730e-01
-8.61868501e-01 6.62614226e-01 9.18422937e-01 9.28281188e-01
9.00326788e-01 -2.60404855e-01 8.91332626e-02 -9.85023856e-01
3.76354307e-01 8.28121230e-03 -8.33629847e-01 6.21776760e-01
-2.88717091e-01 3.94035995e-01 8.41480613e-01 -3.75376463e-01
-1.11341822e+00 4.22685742e-01 1.56644642e-01 -4.86756265e-01
-4.83244777e-01 5.15311956e-01 -4.59559917e-01 8.12415108e-02
2.28388429e-01 6.34265423e-01 -1.79387644e-01 -6.42629445e-01
7.28064001e-01 1.48068056e-01 7.25066841e-01 -9.34234738e-01
7.29266167e-01 6.08593404e-01 1.55196011e-01 -9.47769225e-01
-8.15214932e-01 -1.57938272e-01 -9.53539968e-01 6.74707890e-02
8.29120636e-01 -8.78189683e-01 -6.64127111e-01 6.07145548e-01
-1.36449122e+00 -2.53488094e-01 -2.70845771e-01 -1.61118209e-01
-6.45283282e-01 7.03959048e-01 -8.30675304e-01 -8.44782472e-01
-1.93227455e-01 -1.17597246e+00 1.20550156e+00 -9.81154591e-02
-2.32045278e-01 -7.90247619e-01 -2.40025595e-01 3.45633447e-01
-8.40831921e-03 1.89318031e-01 1.64668024e+00 -1.70225456e-01
-1.19900393e+00 4.88554388e-02 -4.73453641e-01 2.50314415e-01
1.79979187e-02 1.72268629e-01 -9.75557983e-01 2.50934549e-02
-8.57659150e-03 -3.56681943e-01 9.55426872e-01 2.10410103e-01
1.51672792e+00 -5.31139374e-01 -9.41752046e-02 7.85359979e-01
1.46907640e+00 1.30302440e-02 4.27189976e-01 1.10666774e-01
1.00509524e+00 7.98753560e-01 -9.42558646e-02 2.38521874e-01
4.92085934e-01 4.78589892e-01 7.16043055e-01 4.58985418e-02
-3.79432619e-01 -6.49343908e-01 2.85241425e-01 8.12257767e-01
-1.26011102e-02 5.64358616e-03 -8.54844868e-01 3.40457201e-01
-2.16167879e+00 -8.54050577e-01 -1.92850709e-01 1.86536098e+00
6.43828988e-01 1.45219028e-01 3.85840386e-02 5.68350125e-03
5.52958012e-01 1.04933806e-01 -4.95874822e-01 -3.33996832e-01
-2.74132580e-01 1.90567553e-01 4.12110873e-02 6.36341512e-01
-1.09871078e+00 1.11823189e+00 6.35647202e+00 4.39756453e-01
-9.16992843e-01 -3.52542818e-01 5.42278886e-01 3.55643868e-01
-4.54167008e-01 2.70161301e-01 -7.81668067e-01 1.04955725e-01
3.20976049e-01 4.46937829e-01 5.26596606e-01 8.67174208e-01
-1.39072567e-01 2.17044398e-01 -1.61985910e+00 1.07673895e+00
1.09663494e-01 -1.58971965e+00 9.09235656e-01 -1.51403874e-01
3.66180807e-01 -4.41496253e-01 2.43817661e-02 2.65643094e-02
3.72290701e-01 -1.12535214e+00 1.32345164e+00 4.18778837e-01
3.91564399e-01 -3.55932027e-01 1.94955334e-01 5.20335376e-01
-1.37335265e+00 -4.61767823e-01 -5.03027856e-01 -2.53100187e-01
8.51787254e-02 1.44164011e-01 -6.45967484e-01 3.87133688e-01
3.51102442e-01 7.56288588e-01 -6.67661667e-01 8.08911622e-01
-6.18682563e-01 4.26661789e-01 -1.01008162e-01 -4.42004390e-03
2.77585626e-01 -2.67695129e-01 2.36080185e-01 1.30235243e+00
2.94465218e-02 -1.51693285e-01 2.41869718e-01 1.35551488e+00
-1.19942769e-01 -2.33851746e-01 -3.90650600e-01 3.51783447e-03
2.44611844e-01 1.08110762e+00 -7.91469455e-01 -4.25830364e-01
-4.50204670e-01 9.71216202e-01 8.21568310e-01 4.03956980e-01
-6.91558719e-01 2.28199810e-01 4.57122236e-01 1.04023300e-01
5.79815507e-01 -3.25872958e-01 -6.70815885e-01 -1.26724410e+00
2.23345846e-01 -1.16089296e+00 2.80592293e-01 -7.93246150e-01
-1.47466588e+00 4.51301575e-01 1.64555952e-01 -9.15183306e-01
-5.12352884e-02 -1.21262646e+00 -4.28251028e-01 6.29644096e-01
-1.32918513e+00 -1.58596325e+00 -3.09654534e-01 4.45427924e-01
7.90159225e-01 5.66810146e-02 8.78873348e-01 3.46171670e-02
-5.90451181e-01 3.80968332e-01 -5.09523869e-01 2.55503982e-01
-3.35958786e-02 -1.42451239e+00 5.39350450e-01 1.15242410e+00
5.74536920e-01 7.89334595e-01 5.32619238e-01 -4.94805753e-01
-1.65937388e+00 -9.17656720e-01 6.11379147e-01 -3.66402775e-01
8.11085999e-01 -9.16228592e-01 -8.41701806e-01 8.04960012e-01
1.53765902e-01 -1.55603394e-01 6.04298711e-01 9.60668996e-02
-9.08151150e-01 -7.12045208e-02 -5.09999990e-01 8.58635366e-01
1.38357282e+00 -6.16630435e-01 -8.73991907e-01 7.74144754e-03
6.61467969e-01 -9.79910791e-02 -3.63211572e-01 3.69032621e-01
6.90660655e-01 -1.06531644e+00 1.10381877e+00 -8.09079349e-01
7.41929352e-01 -5.23709655e-01 -3.66992444e-01 -7.64813304e-01
-7.54275262e-01 -2.79428750e-01 -5.54245055e-01 9.24325764e-01
4.68993425e-01 -3.75097334e-01 7.94955909e-01 5.34721375e-01
-2.75220990e-01 -7.49377608e-01 -5.28634131e-01 -3.84343863e-01
-1.82371885e-01 -5.17369509e-01 5.56481957e-01 5.71867883e-01
-3.55153143e-01 6.86165214e-01 -4.35139656e-01 5.83557665e-01
9.86966431e-01 9.05423105e-01 8.54990602e-01 -1.24775839e+00
-8.99582684e-01 -6.82334244e-01 -4.17418659e-01 -1.66188645e+00
9.50854793e-02 -8.81420732e-01 6.04189895e-02 -1.77154267e+00
4.46661115e-01 -1.70611680e-01 -1.15999311e-01 6.21439934e-01
-9.09253210e-02 7.97393150e-04 3.96884918e-01 3.75651985e-01
-5.30806363e-01 4.26837742e-01 1.21121252e+00 -5.27381897e-01
2.07280397e-01 -2.68227786e-01 -7.15144396e-01 1.12158310e+00
4.76215333e-01 -1.55580252e-01 -5.77962220e-01 -8.27264428e-01
3.28732580e-01 -8.12052190e-02 7.92841196e-01 -7.02504873e-01
2.30026171e-01 -8.16481560e-02 5.24254560e-01 -7.48259008e-01
4.66384321e-01 -9.92938638e-01 1.42382115e-01 5.18962145e-01
-3.60401630e-01 -4.88764830e-02 1.45458579e-01 6.51671588e-01
-3.38691384e-01 -2.23337531e-01 4.96030927e-01 -3.82316947e-01
-9.22300577e-01 2.94358999e-01 -3.49943727e-01 -2.98635840e-01
6.05132818e-01 -4.54030007e-01 -4.85566765e-01 -1.31933212e-01
-1.07067323e+00 1.86543483e-02 4.43099201e-01 3.38332951e-01
9.03218806e-01 -1.14791453e+00 -4.12169069e-01 4.15424943e-01
2.06304535e-01 3.38128328e-01 1.44456580e-01 5.35798252e-01
-8.62627327e-01 1.59590706e-01 -3.12528104e-01 -9.24615264e-01
-1.35400176e+00 7.44419694e-01 4.09314424e-01 -2.01546505e-01
-7.91488945e-01 9.41101313e-01 8.37683320e-01 -4.25285071e-01
5.66719830e-01 -8.77033412e-01 -5.08541102e-03 -1.43049687e-01
5.10821581e-01 6.08709920e-03 -3.27549160e-01 -5.29229641e-01
-1.99150726e-01 8.57954383e-01 -3.91492508e-02 1.10658430e-01
1.36446249e+00 -1.52705340e-02 -7.13623822e-01 5.19804776e-01
7.61170447e-01 -2.09630668e-01 -1.54095399e+00 -2.78548419e-01
2.96862930e-01 -1.34278551e-01 -5.19499719e-01 -5.53042531e-01
-7.16266096e-01 1.21141601e+00 1.08332068e-01 1.78005934e-01
1.04714704e+00 1.90311849e-01 2.68715918e-01 6.17781281e-01
3.43537718e-01 -4.78719324e-01 3.01152289e-01 4.96956795e-01
1.09008813e+00 -8.61034989e-01 -3.97684276e-02 -7.45408833e-01
-3.95086199e-01 1.39116085e+00 6.60901606e-01 -2.57507831e-01
4.63695526e-01 5.42907119e-01 -1.03617162e-01 -3.23092282e-01
-1.04502547e+00 -5.90662956e-02 4.36253786e-01 3.38556677e-01
2.27193326e-01 1.93763465e-01 2.75498956e-01 6.52909398e-01
-6.66701868e-02 -2.52223045e-01 1.95703655e-01 7.85715580e-01
-4.41860348e-01 -1.23976672e+00 -2.69313663e-01 4.18940485e-02
-1.43073186e-01 -1.46890864e-01 -7.44787037e-01 7.45475888e-01
3.59026074e-01 4.92153883e-01 4.25109714e-02 -2.13238314e-01
2.51641899e-01 3.12100857e-01 9.28722799e-01 -6.71186626e-01
-3.57027292e-01 2.43775845e-01 -7.38223363e-03 -5.27232587e-01
-5.32581866e-01 -5.65267503e-01 -1.17228961e+00 1.73344821e-01
1.48329258e-01 -3.31678718e-01 4.26352471e-01 8.65906894e-01
1.94219559e-01 7.43513942e-01 2.26164639e-01 -8.23638618e-01
-5.71362317e-01 -6.85269535e-01 -2.36885324e-01 5.69976330e-01
2.77941912e-01 -4.19034630e-01 -1.81837842e-01 4.74116027e-01] | [9.952221870422363, 0.921543538570404] |
3293e11b-75c2-4b47-968d-c3d8918ce2a2 | a-comprehensive-overview-and-comparative | 2305.17473 | null | https://arxiv.org/abs/2305.17473v2 | https://arxiv.org/pdf/2305.17473v2.pdf | A Comprehensive Overview and Comparative Analysis on Deep Learning Models: CNN, RNN, LSTM, GRU | Deep learning (DL) has emerged as a powerful subset of machine learning (ML) and artificial intelligence (AI), outperforming traditional ML methods, especially in handling unstructured and large datasets. Its impact spans across various domains, including speech recognition, healthcare, autonomous vehicles, cybersecurity, predictive analytics, and more. However, the complexity and dynamic nature of real-world problems present challenges in designing effective deep learning models. Consequently, several deep learning models have been developed to address different problems and applications. In this article, we conduct a comprehensive survey of various deep learning models, including Convolutional Neural Networks (CNNs), Recurrent Neural Networks (RNNs), Generative Models, Deep Reinforcement Learning (DRL), and Deep Transfer Learning. We examine the structure, applications, benefits, and limitations of each model. Furthermore, we perform an analysis using three publicly available datasets: IMDB, ARAS, and Fruit-360. We compare the performance of six renowned deep learning models: CNN, Simple RNN, Long Short-Term Memory (LSTM), Bidirectional LSTM, Gated Recurrent Unit (GRU), and Bidirectional GRU. | ['Raihani Mohamed', 'Norwati Mustapha', 'Thinagaran Perumal', 'Farhad MortezaPour Shiri'] | 2023-05-27 | null | null | null | null | ['autonomous-vehicles'] | ['computer-vision'] | [-1.55539781e-01 -4.92222309e-02 -3.41326594e-01 -1.88951626e-01
-3.54516059e-01 -6.18542917e-02 7.42989540e-01 -2.19097912e-01
-1.68902338e-01 8.17545056e-01 2.24503711e-01 -7.39266515e-01
6.45160303e-02 -9.25836980e-01 -7.33850956e-01 -7.23161399e-01
-1.14078790e-01 4.08044100e-01 -8.24935287e-02 -3.01545709e-01
7.69240335e-02 6.29437327e-01 -1.09022725e+00 3.93013388e-01
5.91748118e-01 1.23105168e+00 1.89598665e-01 6.74175024e-01
-3.51946890e-01 1.75273061e+00 -8.27291787e-01 -2.26283327e-01
-3.92552763e-01 -2.10533306e-01 -8.09462011e-01 -3.68056476e-01
-6.27727628e-01 -2.87050039e-01 -7.87659824e-01 4.40352082e-01
8.03609312e-01 1.04478598e-01 5.64248025e-01 -1.33791625e+00
-1.16185975e+00 6.69413984e-01 -1.99151725e-01 2.95584768e-01
2.04753265e-01 2.60046154e-01 5.59719145e-01 -8.70624185e-01
5.70240796e-01 1.43122888e+00 8.62594903e-01 6.67385340e-01
-5.47943532e-01 -7.68115759e-01 8.04798454e-02 3.31423521e-01
-9.10013378e-01 -3.21589589e-01 6.14652514e-01 -3.02026421e-01
1.45486760e+00 -2.75551200e-01 3.50254208e-01 1.87239993e+00
6.44931734e-01 1.19054079e+00 7.78932691e-01 -3.39187652e-01
1.23210549e-01 -8.29903334e-02 1.79233477e-01 5.58801889e-01
-3.65560316e-02 3.73327255e-01 -4.88537520e-01 -1.43326789e-01
6.67628407e-01 4.84893471e-01 7.62601346e-02 2.06837475e-01
-1.08382404e+00 1.06107426e+00 5.51314592e-01 5.61511517e-01
-5.41009367e-01 1.41449496e-01 6.02635562e-01 5.44300020e-01
6.99887097e-01 1.23877257e-01 -6.72665596e-01 -2.92011082e-01
-2.61917412e-01 -1.00577965e-01 8.40055346e-01 8.83232772e-01
5.51628172e-01 8.20338249e-01 -4.21274379e-02 1.18850768e+00
1.97850555e-01 4.73812371e-01 1.19145012e+00 -3.77371252e-01
4.03072596e-01 5.88553071e-01 -5.06637335e-01 -9.93229449e-01
-7.07166195e-01 -4.12720382e-01 -1.53164864e+00 -3.49138588e-01
-5.15558422e-01 -4.61800992e-01 -1.04187155e+00 1.40259326e+00
-1.20589420e-01 3.32541287e-01 5.64821959e-01 3.08177382e-01
1.54981840e+00 1.09957302e+00 2.13320047e-01 -1.61471609e-02
7.03480840e-01 -1.13134682e+00 -7.99477577e-01 -5.50681710e-01
9.42991197e-01 -5.04328549e-01 4.49290574e-01 3.28189284e-01
-8.14216435e-01 -7.29555666e-01 -7.05547988e-01 -2.59031743e-01
-8.14806998e-01 5.76389842e-02 6.76437438e-01 2.32840866e-01
-1.01694787e+00 3.82263571e-01 -7.92893708e-01 -4.05531704e-01
6.31527677e-02 3.23532104e-01 1.05127478e-02 -5.01455627e-02
-1.58582079e+00 9.60449278e-01 4.45092350e-01 2.56327242e-01
-1.03696907e+00 -2.10359260e-01 -8.56085181e-01 -1.07633583e-01
1.24483913e-01 -5.68850636e-01 1.34148717e+00 -7.77696967e-01
-1.73873007e+00 4.95233148e-01 4.93546054e-02 -9.69042838e-01
-2.28145160e-02 -2.82819062e-01 -7.61417210e-01 -4.26271379e-01
-3.26157480e-01 4.85195458e-01 5.06153345e-01 -8.51972401e-01
-4.79019791e-01 -3.10451567e-01 -3.07175130e-01 -1.72362670e-01
-1.72461301e-01 1.00401744e-01 5.17688841e-02 -6.82307780e-01
-3.35094929e-01 -9.39385355e-01 -2.87594914e-01 -7.62060940e-01
-4.00916427e-01 -7.43813097e-01 1.23547912e+00 -7.77270257e-01
1.44113410e+00 -2.36650443e+00 2.29877889e-01 -1.76097795e-01
1.75728202e-01 7.22906530e-01 -3.33955795e-01 7.25533485e-01
8.74505639e-02 1.65727854e-01 8.77628699e-02 -1.25261739e-01
-2.73107767e-01 7.19311237e-01 -4.43130046e-01 -1.75409764e-02
1.80235326e-01 1.61567414e+00 -8.29005659e-01 -8.07773992e-02
2.25167170e-01 6.08154953e-01 2.01391578e-02 6.25524282e-01
-2.91285962e-01 3.24547887e-01 -5.70556462e-01 9.56790090e-01
2.77837589e-02 -4.56561238e-01 9.19831023e-02 5.38728088e-02
9.51321051e-02 2.22488075e-01 -4.26865757e-01 1.32135642e+00
-6.91108108e-01 8.24161232e-01 -2.89970875e-01 -1.32422161e+00
1.31793702e+00 5.38845539e-01 3.89911175e-01 -1.17361403e+00
2.12006032e-01 2.81567574e-01 -1.60978407e-01 -6.41046524e-01
3.30036908e-01 4.95232940e-02 -5.75271361e-02 3.57788324e-01
3.32761288e-01 3.92317533e-01 -2.15643972e-01 -4.63824868e-02
1.06454349e+00 -3.92516851e-01 2.28586480e-01 4.69773263e-01
1.94257706e-01 -4.71665353e-01 5.88621736e-01 5.47533631e-01
-1.41067266e-01 4.40520227e-01 3.93655777e-01 -9.62527931e-01
-7.70313799e-01 -7.58376122e-01 2.89151579e-01 1.33335066e+00
-3.09176564e-01 -2.61304826e-01 -4.12609398e-01 -3.43187451e-01
2.09806450e-02 8.11212659e-01 -6.68381214e-01 -7.05378711e-01
-6.32036507e-01 -9.88372386e-01 8.73508573e-01 8.86854053e-01
7.31223226e-01 -1.88829863e+00 -3.93880367e-01 4.09878790e-01
-2.08776429e-01 -1.33548868e+00 1.62603170e-01 3.10587764e-01
-9.15007830e-01 -8.26331198e-01 -4.36123133e-01 -8.21114063e-01
-2.43488643e-02 1.00038655e-03 1.30362833e+00 -3.15447867e-01
-1.86203718e-01 2.09169850e-01 -4.61913854e-01 -5.97994387e-01
-6.08515620e-01 3.83987099e-01 1.42321169e-01 -2.72489667e-01
3.66012067e-01 -2.13098019e-01 -2.02244863e-01 1.49428889e-01
-7.28896856e-01 -4.45286930e-02 1.08989441e+00 1.10555315e+00
4.31401908e-01 -1.61531776e-01 1.10499001e+00 -9.84101713e-01
9.41231191e-01 -1.12901688e+00 -9.08216760e-02 4.46319014e-01
-7.30709016e-01 -7.50708431e-02 6.70591652e-01 -5.13632774e-01
-9.85976994e-01 -4.96596187e-01 -5.86504281e-01 -7.68539965e-01
-9.83054340e-02 8.93302441e-01 2.30918661e-01 1.49854198e-01
4.76322830e-01 6.32290900e-01 1.84334263e-01 -6.16883516e-01
3.31813037e-01 1.07227933e+00 3.65335673e-01 -2.50628203e-01
9.79455635e-02 -1.24424011e-01 -4.15715069e-01 -9.42044318e-01
-7.98792839e-01 -1.21709749e-01 -4.50374454e-01 -7.85306748e-03
8.56777430e-01 -7.90222466e-01 -5.18986583e-01 8.16285849e-01
-1.22381067e+00 -6.71195507e-01 -2.05424875e-01 3.23456794e-01
-3.60688359e-01 -9.76238996e-02 -1.17980027e+00 -7.74745286e-01
-8.64206791e-01 -1.13762677e+00 8.67371798e-01 2.16143176e-01
-1.42698865e-02 -1.37511766e+00 1.14644364e-01 3.21696848e-01
8.50245714e-01 4.71266299e-01 1.02923608e+00 -9.55210447e-01
-1.92065954e-01 -2.86083758e-01 -1.46880582e-01 7.66514480e-01
1.44253284e-01 1.39817193e-01 -9.05137420e-01 -2.22650081e-01
-2.40951046e-01 -7.97854960e-01 1.00446892e+00 5.20314038e-01
1.46565056e+00 -5.28882444e-01 -4.04672116e-01 6.79399788e-01
8.49917948e-01 6.90653443e-01 6.77143872e-01 2.36147061e-01
8.47791076e-01 1.53617367e-01 9.16558430e-02 3.39134067e-01
4.93975252e-01 -2.02731919e-02 4.52933550e-01 -2.38270178e-01
2.94659194e-02 -2.36893862e-01 6.26854420e-01 1.33520770e+00
-1.53073773e-01 -4.41615850e-01 -1.03570092e+00 1.65448800e-01
-1.94264686e+00 -8.50547135e-01 2.17328340e-01 1.66698503e+00
4.92568523e-01 1.56187430e-01 6.48867860e-02 -1.85490362e-02
4.01804477e-01 2.10102662e-01 -8.68928552e-01 -8.55803788e-01
-3.31752241e-01 2.61592418e-01 2.30760351e-01 -5.33767119e-02
-9.48352754e-01 1.15202034e+00 7.26867771e+00 8.53551447e-01
-1.53003216e+00 1.04514979e-01 1.16300392e+00 1.81863278e-01
4.43798210e-03 -7.12910354e-01 -8.18974912e-01 3.87417287e-01
1.70394206e+00 5.96770495e-02 3.94012809e-01 1.05597854e+00
-7.87838176e-02 3.56316388e-01 -9.04541612e-01 9.34190750e-01
8.53521153e-02 -1.48426449e+00 3.82106632e-01 -1.45183075e-02
6.37084901e-01 9.42236006e-01 4.16447282e-01 9.78433430e-01
5.51502526e-01 -1.38663220e+00 1.04259200e-01 5.60585201e-01
7.12553740e-01 -1.05209935e+00 9.48564112e-01 5.06776750e-01
-7.78564155e-01 -5.67417920e-01 -4.87003535e-01 -5.93357794e-02
-2.51301408e-01 5.75493753e-01 -7.78144896e-01 5.83582699e-01
7.94394195e-01 1.13375604e+00 -3.22580159e-01 4.54192072e-01
7.34531358e-02 8.51142645e-01 1.38141692e-01 -2.67275423e-01
7.01178372e-01 -8.25119540e-02 4.38480787e-02 1.40512025e+00
2.67827868e-01 -1.86317123e-03 1.48760989e-01 6.33671284e-01
-3.17922920e-01 -1.86547980e-01 -9.29022729e-01 -5.21379709e-01
3.12704057e-01 1.05477810e+00 -1.14479549e-01 -2.98863053e-01
-6.14789665e-01 5.79003274e-01 1.96372032e-01 6.53750718e-01
-7.14556515e-01 -1.89509630e-01 5.74038625e-01 -2.76390493e-01
-7.97898993e-02 -2.07887858e-01 -1.27224922e-01 -1.03351498e+00
-5.23272991e-01 -1.01830852e+00 5.49368143e-01 -8.13014925e-01
-1.41301608e+00 9.79874909e-01 -2.63204664e-01 -9.10954118e-01
-9.33593988e-01 -6.83025718e-01 -6.14837527e-01 6.04506969e-01
-1.79617286e+00 -1.29384816e+00 -2.16510713e-01 6.95302963e-01
7.62889862e-01 -7.91093051e-01 8.32965195e-01 4.32379872e-01
-1.00374877e+00 3.70463789e-01 3.99237812e-01 4.79728132e-01
3.75641644e-01 -7.36941278e-01 5.85458875e-01 2.86961168e-01
-1.28414467e-01 3.91851187e-01 9.94864255e-02 -3.72924566e-01
-1.61142266e+00 -1.57028496e+00 7.43843913e-01 -3.54884490e-02
7.16041863e-01 -2.68257946e-01 -9.83281851e-01 1.20159447e+00
3.60186577e-01 1.71354711e-01 6.54588640e-01 9.66366976e-02
-1.61917105e-01 -2.28332147e-01 -8.03242028e-01 2.37919927e-01
6.12138987e-01 -4.87844914e-01 -2.17651129e-01 4.18691874e-01
1.04803061e+00 -3.08898926e-01 -8.92312407e-01 7.39969850e-01
5.05045772e-01 -8.16079319e-01 1.07212400e+00 -8.83073509e-01
5.66449761e-01 5.59181631e-01 -1.28921926e-01 -1.47732818e+00
-3.81803483e-01 -4.99105632e-01 -8.19475293e-01 9.16349292e-01
2.83102751e-01 -9.38395202e-01 4.08273309e-01 3.04527134e-01
-4.68285114e-01 -1.44985819e+00 -8.52167904e-01 -7.43538439e-01
2.69951075e-01 -4.28389490e-01 5.83012283e-01 8.93202007e-01
-3.18421006e-01 5.46063960e-01 -6.90905869e-01 -3.41136962e-01
9.01453793e-02 -3.53489816e-02 6.22652590e-01 -1.11735916e+00
-1.35015130e-01 -5.71943820e-01 -3.83345783e-01 -9.24122512e-01
3.01087141e-01 -5.35154104e-01 -4.06207472e-01 -1.80194402e+00
-6.77023754e-02 -3.91402900e-01 -7.18544364e-01 7.48807430e-01
1.80489197e-01 -3.04043502e-01 -5.82844168e-02 2.39601761e-01
-5.93433857e-01 6.79617107e-01 1.04086840e+00 -2.60618716e-01
-1.52415171e-01 2.70537108e-01 -3.83729190e-01 6.91044152e-01
1.06039774e+00 6.15500920e-02 -3.42795759e-01 -5.76337516e-01
1.61806703e-01 2.31772751e-01 1.48722857e-01 -8.42854083e-01
3.76999825e-01 -1.24782398e-01 4.59059894e-01 -7.28862882e-01
2.81001300e-01 -3.13564241e-01 -6.84376508e-02 7.84557641e-01
-4.02459651e-01 3.81175905e-01 8.98350030e-02 3.01116288e-01
-5.57399273e-01 2.30895564e-01 6.81603849e-01 -2.23410234e-01
-1.02668560e+00 6.63735271e-01 -5.74295282e-01 -1.10828117e-01
9.04036999e-01 9.14651081e-02 -4.03640509e-01 -5.49281001e-01
-6.50142908e-01 2.66990513e-01 -4.45557833e-01 9.23205674e-01
1.09119868e+00 -1.26008296e+00 -7.12629676e-01 3.78642231e-01
8.08458868e-03 1.33077845e-01 1.25737295e-01 5.87348461e-01
-2.95761824e-01 7.85671830e-01 -5.92996553e-02 -3.07021081e-01
-8.16438496e-01 7.33053267e-01 3.96830469e-01 -4.30831969e-01
-5.44681609e-01 7.05468118e-01 1.83908731e-01 -8.47258568e-01
6.14181519e-01 -5.78329146e-01 -3.62526417e-01 -3.18290554e-02
3.87043238e-01 5.72749615e-01 1.44231394e-01 -3.94852757e-01
-1.05389796e-01 1.05351254e-01 -1.07770227e-01 6.62451923e-01
1.52234936e+00 1.51413515e-01 -7.87721723e-02 9.10296679e-01
1.39099383e+00 -9.41252232e-01 -6.74479783e-01 -4.36626285e-01
1.21542521e-01 5.20491362e-01 1.09397724e-01 -8.96887600e-01
-1.40290892e+00 1.22233105e+00 3.82880360e-01 3.54643643e-01
1.03246105e+00 -9.02017280e-02 1.54236281e+00 6.81134403e-01
1.89663589e-01 -1.00413394e+00 3.78589958e-01 9.53617096e-01
9.14031327e-01 -1.19746923e+00 -6.39125347e-01 3.57051939e-01
-6.70523405e-01 1.33661950e+00 7.23450959e-01 -1.43027939e-02
1.02206898e+00 4.43754554e-01 -2.49313023e-02 1.86865218e-02
-1.33895552e+00 2.31043417e-02 2.57163737e-02 4.48260784e-01
6.67871892e-01 7.97927827e-02 2.81922072e-01 4.75884825e-01
-8.35135132e-02 2.59225130e-01 1.50842398e-01 1.01924396e+00
-4.08342063e-01 -8.75318527e-01 -8.98123439e-03 6.17816091e-01
-4.52802032e-01 -1.55530035e-01 -5.66189885e-01 6.02004051e-01
-1.37302741e-01 8.94815505e-01 1.26540586e-01 -7.92917669e-01
1.54853985e-01 2.26225629e-02 -5.25213405e-02 -3.72455448e-01
-5.86318493e-01 -1.09404996e-01 -3.30505073e-02 -4.01063114e-01
-1.53744414e-01 -1.61006555e-01 -1.24560809e+00 -5.06796479e-01
-1.11540675e-01 2.18305184e-04 7.52563477e-01 1.05929291e+00
7.90470719e-01 7.63580978e-01 7.55698383e-01 -4.94721562e-01
-4.48031694e-01 -1.19852817e+00 -3.45929980e-01 4.68043057e-04
3.66965920e-01 -5.68341851e-01 -2.01841250e-01 -2.40528062e-01] | [10.859955787658691, 6.290380477905273] |
03ebc842-09dc-436b-81f3-a6e399eb56cd | taches-auxiliaires-pour-lanalyse-biaffine-en | null | null | https://aclanthology.org/2022.jeptalnrecital-taln.42 | https://aclanthology.org/2022.jeptalnrecital-taln.42.pdf | Tâches auxiliaires pour l’analyse biaffine en graphes de dépendances (Auxiliary tasks to boost Biaffine Semantic Dependency Parsing) | L’analyseur biaffine de Dozat & Manning (2017), qui produit des arbres de dépendances syntaxiques, a été étendu avec succès aux graphes de dépendances syntaxico-sémantiques (Dozat & Manning, 2018). Ses performances sur les graphes sont étonnamment hautes étant donné que, sans la contrainte de devoir produire un arbre, les arcs pour une phrase donnée sont prédits indépendamment les uns des autres. Pour y remédier partiellement, tout en conservant la complexité O(n2 ) et l’architecture hautement parallélisable, nous proposons d’utiliser des tâches auxiliaires qui introduisent une forme d’interdépendance entre les arcs. Les expérimentations sur les trois jeux de données anglaises de la tâche 18 SemEval-2015 (Oepen et al., 2015), et sur des graphes syntaxiques profonds en français (Ribeyre et al., 2014) montrent une amélioration modeste mais systématique, par rapport à un système de base performant, utilisant un modèle de langue pré-entraîné. Notre méthode s’avère ainsi un moyen simple et robuste d’améliorer l’analyse vers graphes de dépendances. | ['Marie Candito'] | null | null | null | null | jep-taln-recital-2022-6 | ['semantic-dependency-parsing'] | ['natural-language-processing'] | [-2.86921769e-01 -5.28180739e-03 4.39911991e-01 -5.20349383e-01
-6.30848110e-02 -1.12228394e+00 1.07644117e+00 9.16079283e-01
-5.12698472e-01 6.44682825e-01 1.03316501e-01 -3.17427427e-01
5.39094657e-02 -1.10825908e+00 -1.04126060e+00 -1.38529912e-01
-3.20802510e-01 3.78135949e-01 8.33446980e-02 -7.44420826e-01
-1.43581986e-01 4.71194714e-01 -7.85655558e-01 1.20084502e-01
5.01786232e-01 7.64243066e-01 7.30257481e-02 5.00656903e-01
8.39266777e-02 8.25324059e-01 -3.02324057e-01 -6.36228204e-01
6.60509050e-01 -3.31450194e-01 -5.30093133e-01 -1.33126140e-01
7.39840448e-01 9.09201562e-01 -4.16268021e-01 1.02679360e+00
-8.60047638e-02 -1.43962950e-01 1.25737906e+00 -8.14398706e-01
-9.62319434e-01 1.07820594e+00 -3.89978826e-01 1.02288210e+00
2.95429677e-01 3.53575647e-02 1.55873251e+00 -5.41765153e-01
3.09438914e-01 8.57753634e-01 8.89806688e-01 4.29262370e-01
-1.22347021e+00 -2.87393689e-01 3.70311081e-01 3.69942077e-02
-1.01465666e+00 -3.67412001e-01 -8.66802111e-02 -4.38297570e-01
8.14378381e-01 5.12308419e-01 3.34590614e-01 1.00304770e+00
5.23602486e-01 6.91487253e-01 1.16235304e+00 1.50181528e-04
9.52270329e-02 4.04278845e-01 9.69899818e-02 1.08931649e+00
2.08870173e-01 2.36996084e-01 -4.01104577e-02 -4.23053384e-01
6.89208269e-01 1.94765598e-01 -8.31715763e-02 -1.35039777e-01
-9.40039873e-01 8.76066387e-01 1.10148120e+00 3.60924691e-01
-3.60752761e-01 1.35629565e-01 7.44316339e-01 5.30505598e-01
9.59867463e-02 8.02206457e-01 -4.15145159e-01 -5.81032395e-01
-3.00030440e-01 2.62028068e-01 1.42069006e+00 1.12175310e+00
6.39104307e-01 2.70231981e-02 6.36604190e-01 4.65588748e-01
1.33090258e-01 1.78027362e-01 -6.99280426e-02 -7.65775964e-02
2.73871630e-01 5.49615622e-01 -3.60455662e-01 -9.24165547e-01
-4.94864643e-01 -8.25063586e-01 -7.73584485e-01 -2.65772760e-01
3.66926014e-01 -1.89500734e-01 -3.11430931e-01 1.43684316e+00
-2.08378077e-01 -2.03491241e-01 5.11019826e-02 9.03352261e-01
-4.44514275e-01 3.73281837e-01 4.89533246e-01 -2.73560733e-01
1.53101957e+00 -5.73603570e-01 2.41755038e-01 -4.36346203e-01
5.55211782e-01 -8.16762805e-01 8.92490506e-01 9.13794339e-01
-6.19686365e-01 -3.31481963e-01 -1.12801683e+00 1.18420785e-02
-2.18839526e-01 -1.95624024e-01 3.74778628e-01 4.05328780e-01
-7.94737637e-01 5.38767576e-01 -6.77798986e-01 -2.67195284e-01
-1.90996006e-01 2.51566231e-01 -5.33427119e-01 5.68830371e-01
-8.60396087e-01 1.43090606e+00 7.52068102e-01 6.51888788e-01
-6.79322004e-01 9.43240151e-02 -1.07625401e+00 3.89705658e-01
1.24543500e+00 -2.28867054e-01 1.39480627e+00 -1.29265118e+00
-1.04929101e+00 3.29592288e-01 3.21530364e-02 -7.45534062e-01
6.31851733e-01 3.46352220e-01 -4.58264023e-01 -5.26976883e-02
-3.79329026e-01 9.81227146e-04 4.93093789e-01 -9.96188641e-01
-8.80970716e-01 6.52835518e-02 7.27176309e-01 7.20533207e-02
-5.10857761e-01 -1.36470124e-01 -3.71797122e-02 -9.25284147e-01
-1.00489065e-01 -1.39966512e+00 1.57703444e-01 4.50650416e-02
-5.43220818e-01 2.65646409e-02 8.62899840e-01 -3.34913194e-01
9.69804406e-01 -2.06137347e+00 8.89729738e-01 2.88247317e-01
9.67248380e-01 5.39681315e-01 -7.39668846e-01 5.79187810e-01
-6.80327564e-02 3.27975482e-01 -8.92869115e-01 1.12737000e-01
4.49725032e-01 4.21209969e-02 4.31832135e-01 2.68171906e-01
-2.18097627e-01 8.63123059e-01 -1.12519431e+00 -2.48203471e-01
-1.59234747e-01 1.37975886e-01 -6.92137420e-01 -1.33976668e-01
-6.58378243e-01 9.02583450e-02 -6.55014992e-01 -2.61593252e-01
5.35524786e-01 -6.42472804e-01 4.87851679e-01 -4.38971877e-01
-3.42034698e-01 3.89971137e-01 -1.16740465e+00 1.30045593e+00
-8.45572352e-01 1.64149269e-01 4.87544358e-01 -7.67952621e-01
1.34428918e+00 9.38012227e-02 4.67709273e-01 -1.20842129e-01
2.60910969e-02 3.61131936e-01 1.45307202e-02 2.22872458e-02
2.06014797e-01 -3.77827764e-01 1.60329133e-01 1.01528275e+00
-2.95017123e-01 -1.56216174e-01 1.10131514e+00 -1.12191744e-01
1.66324592e+00 4.32602689e-03 1.49661601e-01 -4.60370481e-01
1.02867460e+00 -1.42503574e-01 2.46586483e-02 4.50830981e-02
1.04581276e-02 -2.20273376e-01 9.52943027e-01 -4.31748658e-01
-1.52392566e+00 -7.67235041e-01 2.37253346e-02 1.00038910e+00
-6.72095537e-01 -6.84545040e-01 -4.49201822e-01 -1.55079651e+00
2.65480787e-01 1.00212574e+00 -6.42382383e-01 2.38030236e-02
-2.38438547e-01 -5.18240750e-01 7.05350578e-01 5.58138013e-01
4.93021309e-01 -5.16602695e-01 -6.10481858e-01 4.74815995e-01
-9.60276648e-02 -1.07726407e+00 -3.27135473e-01 -6.88990280e-02
-3.25238436e-01 -1.10767567e+00 -3.40449572e-01 5.37994131e-03
3.98167551e-01 -8.08371902e-02 1.32968307e+00 -3.09560567e-01
-2.48759657e-01 -1.78630501e-02 -6.51461005e-01 -3.93283725e-01
-1.10899949e+00 6.19991362e-01 5.35091385e-02 7.11169466e-02
4.80153650e-01 -4.69242126e-01 -6.05174839e-01 5.93238115e-01
-9.88119423e-01 -3.05842161e-01 7.68647850e-01 4.99338239e-01
2.47940570e-02 -1.34912014e-01 2.03998432e-01 -1.29299963e+00
9.56458807e-01 -8.16308200e-01 -6.80117965e-01 -7.48379603e-02
-5.00054061e-01 4.09521371e-01 1.89331293e+00 2.80263513e-01
-6.60567284e-01 -1.77186102e-01 -2.50512600e-01 -7.75849074e-02
-4.24280196e-01 1.06237626e+00 1.00182081e-02 4.19451982e-01
1.14090407e+00 3.74691516e-01 -2.20052719e-01 -4.77752328e-01
4.05725867e-01 3.17596376e-01 1.01221755e-01 -9.78290737e-01
6.91370666e-01 -2.82072462e-02 -2.29324311e-01 -3.08636039e-01
-3.26789856e-01 1.64765552e-01 -7.89002776e-01 1.90455377e-01
7.55741656e-01 -6.12511694e-01 -6.96715772e-01 -4.58489731e-02
-9.88440454e-01 -4.88737404e-01 1.86724395e-01 5.95475197e-01
-2.60190099e-01 -3.73328000e-01 -1.04809058e+00 -1.99507430e-01
3.30851257e-01 -6.87253714e-01 5.20291090e-01 -4.11834508e-01
-1.26060516e-01 -1.09422576e+00 4.28425640e-01 -9.61062163e-02
3.47233564e-01 1.20535895e-01 1.14276302e+00 -3.96749467e-01
-1.04180622e+00 -1.77995175e-01 1.43251896e-01 5.43152452e-01
7.98423052e-01 5.23891635e-02 -2.92609632e-01 -5.54394603e-01
-5.26182234e-01 5.03261447e-01 3.43212664e-01 -3.35021853e-01
6.70598149e-01 -7.70445228e-01 3.62453759e-01 2.25019127e-01
1.98176003e+00 -4.21139300e-02 1.34676350e-02 5.60034253e-02
1.68953225e-01 7.17769265e-01 6.41359985e-01 -3.96368504e-01
4.89554882e-01 5.79959631e-01 4.89326149e-01 4.48425710e-01
8.66897643e-01 -3.34037952e-02 3.04403186e-01 1.28611541e+00
-6.04067683e-01 -4.13095742e-01 -1.09006310e+00 4.04863238e-01
-1.60626268e+00 -5.87308705e-02 -6.37398660e-01 1.79525709e+00
7.33497500e-01 3.28426033e-01 5.52783728e-01 -2.71081448e-01
3.94167334e-01 5.87119818e-01 -3.36829871e-01 -9.57124352e-01
-2.03153700e-01 2.33679131e-01 5.32729447e-01 6.73902154e-01
-4.57943738e-01 8.65606964e-01 4.11555386e+00 1.10319704e-01
-1.34303236e+00 -2.41718553e-02 4.05363202e-01 -2.84135826e-02
-3.30326229e-01 -1.72698960e-01 -3.70637774e-01 6.98129654e-01
1.58846855e+00 -2.53272563e-01 9.44252014e-01 4.59068418e-01
-1.38950363e-01 9.13831741e-02 -1.47455382e+00 3.86223733e-01
2.98160344e-01 -1.25447059e+00 -9.44115102e-01 -4.69616115e-01
-7.21215531e-02 5.17558634e-01 -1.31087467e-01 7.94128001e-01
7.95880318e-01 -1.02015817e+00 6.82580948e-01 2.61915505e-01
2.93960720e-01 -8.13011169e-01 -2.96351135e-01 2.44749859e-01
-8.47166538e-01 3.16797942e-02 -2.26251349e-01 -7.43436348e-03
4.08120453e-01 3.87138277e-02 -5.03566265e-01 1.14059544e+00
3.78401242e-02 9.33110535e-01 -6.68619812e-01 7.18768537e-01
-6.04131937e-01 8.02443266e-01 -6.87142909e-01 -4.74994034e-01
6.52237475e-01 -5.09753466e-01 6.64956868e-01 1.24604142e+00
9.58319232e-02 3.62413943e-01 -5.18708885e-01 8.99035573e-01
-1.00588672e-01 4.99536961e-01 -1.69395432e-01 -3.78194958e-01
-2.80255675e-02 8.52275431e-01 -8.34442675e-01 1.13247946e-01
-6.61857724e-01 7.15986550e-01 4.38453645e-01 -4.63959463e-02
-5.14715970e-01 -6.75505698e-01 1.18870176e-01 -4.41614613e-02
1.45205930e-01 -8.94450963e-01 2.03293741e-01 -1.16800570e+00
-3.73825938e-01 -1.53225756e+00 8.61776322e-02 -4.98256743e-01
-1.17483282e+00 9.00608599e-01 2.62021929e-01 -4.61790115e-01
-9.07765508e-01 -4.79499072e-01 -6.17533773e-02 6.03116691e-01
-8.01240742e-01 -6.05473280e-01 5.19913793e-01 4.25593287e-01
2.67082125e-01 -7.62924105e-02 1.32804203e+00 6.92201257e-01
-9.00319219e-01 -9.49188396e-02 -1.80060297e-01 -1.23810265e-02
4.61640835e-01 -1.42279565e+00 1.02048290e+00 8.30271661e-01
7.97970474e-01 8.43119144e-01 6.97597563e-01 -6.08743131e-01
-1.48337781e+00 -1.01167107e+00 6.67125642e-01 -5.49192846e-01
1.13419223e+00 -6.43674135e-01 -9.88663077e-01 2.05805826e+00
3.10740203e-01 -3.11522000e-02 -1.68443769e-01 5.44203162e-01
-5.03307343e-01 -3.26074392e-01 -4.67553794e-01 7.31793880e-01
1.91624910e-01 -3.83529961e-01 -2.27792934e-01 8.61514211e-02
-8.28185007e-02 -6.85841024e-01 -1.22079539e+00 2.63686419e-01
-1.77713796e-01 -1.34575832e+00 1.78446665e-01 -4.09389883e-01
4.39822286e-01 -4.55805928e-01 2.97841430e-01 -1.37281311e+00
2.84821719e-01 -5.53406537e-01 -4.42436039e-01 7.19869435e-01
4.37513381e-01 -5.70132911e-01 2.00245082e-01 9.85216126e-02
1.00557610e-01 -8.32258835e-02 -8.97141755e-01 -4.20616239e-01
2.24450812e-01 -7.68205404e-01 6.52850568e-01 7.84386694e-01
-3.38530481e-01 1.06118298e+00 -1.97400331e-01 6.02123477e-02
1.14462666e-01 -2.38367349e-01 7.81400383e-01 -8.49226713e-01
-7.32271552e-01 -3.36674243e-01 -4.20784682e-01 -1.02057874e+00
3.60692769e-01 -1.12825859e+00 -6.69302523e-01 -1.23357439e+00
-7.98424706e-03 -3.67502868e-01 -3.86463106e-01 8.33888888e-01
3.50506216e-01 1.05623573e-01 9.81024563e-01 -1.68449447e-01
-5.68987370e-01 1.68528870e-01 1.12292254e+00 9.66285616e-02
2.94882447e-01 -1.80518115e-03 -6.11842394e-01 7.31512189e-01
4.40545171e-01 -5.77474296e-01 -4.58846599e-01 -4.35214877e-01
1.03794193e+00 -2.82305837e-01 -2.21523698e-02 -6.38781846e-01
5.33427715e-01 -3.61792535e-01 1.13996312e-01 -2.66898006e-01
8.59617710e-01 -9.68524396e-01 9.10499319e-02 7.18247294e-02
-3.80199015e-01 4.65791404e-01 4.70005363e-01 3.49788576e-01
-2.93761790e-01 -1.21621825e-01 7.57585168e-01 -9.61218402e-02
-7.97699869e-01 8.12585890e-01 -4.51743305e-01 5.95972896e-01
9.45819259e-01 5.34709916e-02 -3.51814814e-02 -4.24944945e-02
-7.71253228e-01 2.75739998e-01 4.36939269e-01 2.84603626e-01
1.98730767e-01 -8.04419458e-01 -1.47223878e+00 1.56499103e-01
-1.76669061e-01 -6.16128594e-02 7.09018171e-01 6.31342769e-01
-1.16016948e+00 3.23696762e-01 -4.49219048e-01 -9.17074323e-01
-1.46527994e+00 5.26991665e-01 4.25957978e-01 1.91320360e-01
-7.53341019e-01 9.78803754e-01 -5.92745364e-01 -3.16952467e-01
-7.26395190e-01 -3.67839515e-01 6.27166390e-01 -3.94619286e-01
-2.52598256e-01 4.59201545e-01 -1.11189581e-01 -4.65844989e-01
-2.52749115e-01 2.53510743e-01 -4.43271965e-01 -2.58257747e-01
1.26466322e+00 3.43784764e-02 -7.65117943e-01 3.70608330e-01
1.34011328e+00 8.19642603e-01 -1.59592187e+00 6.51306331e-01
1.24710567e-01 5.20796217e-02 -4.47254479e-01 -9.85207379e-01
-1.17065167e+00 7.27058351e-01 1.33127302e-01 8.45584035e-01
5.64002454e-01 -3.63131344e-01 4.32136655e-01 5.92613041e-01
5.34492791e-01 -6.93143606e-01 -6.45076990e-01 1.05044925e+00
1.15853238e+00 -9.80856359e-01 -2.68448949e-01 -7.12723956e-02
-6.84590876e-01 1.37334085e+00 4.39073443e-02 -2.04326704e-01
6.28846407e-01 2.87713498e-01 3.44987392e-01 -3.31810087e-01
-8.59727621e-01 -1.16313053e-02 4.43061918e-01 4.48292028e-03
5.73561132e-01 1.51306555e-01 -7.54942417e-01 1.81508511e-01
-9.28200364e-01 -3.06450222e-02 6.95037365e-01 8.74168754e-01
4.04841214e-01 -1.16563809e+00 -9.81598273e-02 2.07267299e-01
-2.71382123e-01 -3.90919656e-01 -6.95794106e-01 9.37376201e-01
-1.39761463e-01 7.83783615e-01 6.32405058e-02 -4.90042806e-01
9.99097228e-01 -3.11203331e-01 4.73204255e-01 -9.84936357e-01
-1.33837700e+00 8.21236297e-02 3.07615310e-01 -3.37741017e-01
-6.06621563e-01 -2.85098612e-01 -6.88514292e-01 -1.05269384e+00
-1.45465642e-01 6.39291763e-01 6.52617097e-01 1.16302752e+00
1.97757319e-01 8.00718486e-01 1.81764960e-01 -9.88567844e-02
-6.00588799e-01 -1.00739563e+00 -1.05993974e+00 3.13533889e-03
3.83080840e-01 -3.18512559e-01 -1.52143642e-01 -3.01511943e-01] | [14.101972579956055, 13.316201210021973] |
955a7564-0444-4275-8c3e-bcc705d8d99f | graph-to-tree-neural-networks-for-learning | 2004.13781 | null | https://arxiv.org/abs/2004.13781v2 | https://arxiv.org/pdf/2004.13781v2.pdf | Graph-to-Tree Neural Networks for Learning Structured Input-Output Translation with Applications to Semantic Parsing and Math Word Problem | The celebrated Seq2Seq technique and its numerous variants achieve excellent performance on many tasks such as neural machine translation, semantic parsing, and math word problem solving. However, these models either only consider input objects as sequences while ignoring the important structural information for encoding, or they simply treat output objects as sequence outputs instead of structural objects for decoding. In this paper, we present a novel Graph-to-Tree Neural Networks, namely Graph2Tree consisting of a graph encoder and a hierarchical tree decoder, that encodes an augmented graph-structured input and decodes a tree-structured output. In particular, we investigated our model for solving two problems, neural semantic parsing and math word problem. Our extensive experiments demonstrate that our Graph2Tree model outperforms or matches the performance of other state-of-the-art models on these tasks. | ['Lingfei Wu', 'Shucheng Li', 'Fengyuan Xu', 'Fangli Xu', 'Sheng Zhong', 'Shiwei Feng'] | 2020-04-07 | null | https://aclanthology.org/2020.findings-emnlp.255 | https://aclanthology.org/2020.findings-emnlp.255.pdf | findings-of-the-association-for-computational | ['math-word-problem-solving', 'math-word-problem-solving', 'math-word-problem-solving'] | ['knowledge-base', 'reasoning', 'time-series'] | [ 6.61629975e-01 5.02552450e-01 -2.01245546e-01 -4.93235648e-01
-6.13559008e-01 -7.52160370e-01 1.19561397e-01 2.59087056e-01
-2.00141817e-01 6.14477456e-01 2.83807009e-01 -7.72192776e-01
4.24115658e-01 -1.19087005e+00 -1.09886777e+00 -2.35937700e-01
1.47385215e-02 3.85666519e-01 9.21935141e-02 1.83661282e-02
1.60696730e-01 -7.43072927e-02 -1.17472589e+00 5.43613493e-01
1.00233376e+00 8.03949893e-01 3.51349711e-01 9.39906657e-01
-9.11515176e-01 1.23380864e+00 -5.31705678e-01 -8.57024789e-01
-7.86008686e-02 -7.55779862e-01 -9.10227716e-01 -3.17466170e-01
5.40720940e-01 -2.26892158e-01 -4.38157409e-01 1.35691881e+00
3.08934152e-01 1.57086492e-01 4.14801091e-01 -8.88663471e-01
-1.37265694e+00 1.03876865e+00 -2.52618611e-01 -3.67944827e-03
4.73068416e-01 -6.98262220e-03 1.24112856e+00 -5.32208264e-01
5.70806921e-01 1.50380516e+00 5.57170033e-01 7.90448844e-01
-1.09427452e+00 -5.59174418e-01 3.10076088e-01 1.76823810e-01
-1.06669462e+00 -2.33191982e-01 6.29100680e-01 -2.80377656e-01
1.32833922e+00 1.63449179e-02 5.36399424e-01 9.22003090e-01
4.63716984e-01 9.07528996e-01 6.73794925e-01 -2.20645219e-01
2.21501648e-01 -5.59462190e-01 3.56870770e-01 1.35167420e+00
5.40403612e-02 -7.36260787e-02 -3.48672509e-01 7.12528229e-02
9.04697359e-01 -2.34206736e-01 5.47670340e-03 1.75338492e-01
-1.03086078e+00 9.00574386e-01 5.18897533e-01 6.23120926e-02
-5.24805784e-02 7.30441451e-01 4.56604630e-01 5.66851199e-01
5.60076773e-01 3.55531603e-01 -4.24850225e-01 -1.08455867e-02
-3.59628886e-01 1.15209661e-01 9.57553864e-01 1.17795599e+00
7.65770853e-01 2.68902659e-01 -3.80705953e-01 6.21174574e-01
8.71262550e-02 2.55709976e-01 3.77196580e-01 -9.18978930e-01
1.03840542e+00 7.36905277e-01 -5.96211791e-01 -8.54120791e-01
-2.64925092e-01 -4.32513297e-01 -1.00876391e+00 -4.34040189e-01
2.45715022e-01 -1.43893883e-01 -1.20869827e+00 1.94995320e+00
5.10850549e-02 3.73856634e-01 3.10087293e-01 7.14676082e-01
1.39675772e+00 1.13086522e+00 3.34323078e-01 1.36636764e-01
1.24555576e+00 -1.39125085e+00 -7.14592457e-01 -3.91782373e-01
8.93009961e-01 -2.94023663e-01 1.10804725e+00 1.46099567e-01
-1.39902151e+00 -6.84540331e-01 -7.46584654e-01 -5.10546982e-01
-4.00507480e-01 -1.36544660e-01 8.04958701e-01 3.47413391e-01
-1.33408284e+00 7.61384130e-01 -5.84094584e-01 -1.42058462e-01
2.39459097e-01 2.26310134e-01 -3.15136582e-01 -3.08115900e-01
-1.19145882e+00 5.11551023e-01 5.88226795e-01 6.39100596e-02
-6.97309732e-01 -5.63653052e-01 -1.46025133e+00 4.41438496e-01
3.80505413e-01 -1.17598999e+00 1.52348816e+00 -1.03092623e+00
-1.65978205e+00 9.09678280e-01 -3.68292540e-01 -4.85925466e-01
-8.79089290e-04 -1.17101021e-01 -1.54836893e-01 5.31542785e-02
1.02070961e-02 7.11941540e-01 5.62195599e-01 -6.21605635e-01
-2.07321167e-01 -3.65197837e-01 1.58145338e-01 1.82139352e-01
-4.21618037e-02 -6.74419478e-03 -6.24396980e-01 -8.37136090e-01
9.01203081e-02 -5.54014921e-01 -4.41488534e-01 -2.56595939e-01
-6.32923365e-01 -4.63288367e-01 2.00560972e-01 -7.25157440e-01
1.21671307e+00 -2.05111194e+00 6.20205820e-01 -1.76625088e-01
3.91092539e-01 1.37059584e-01 -6.82170749e-01 6.24282122e-01
-5.18860780e-02 3.11526954e-01 -5.40599406e-01 -1.56454191e-01
-6.47789985e-02 4.30355787e-01 -2.44326904e-01 -1.22059666e-01
4.19696093e-01 1.66061366e+00 -1.17105985e+00 -4.02981430e-01
-2.01980591e-01 7.32271224e-02 -7.78768599e-01 5.47367334e-01
-7.86734760e-01 2.46143803e-01 -6.06152833e-01 5.06569564e-01
4.56208557e-01 -5.33044398e-01 3.10974807e-01 1.98146522e-01
1.90183714e-01 6.68220699e-01 -5.86130261e-01 2.03471756e+00
-4.25968707e-01 4.76753831e-01 3.09927128e-02 -1.14262509e+00
8.81722569e-01 1.27891988e-01 -4.03367355e-02 -8.33853900e-01
-3.84495258e-02 -8.88948813e-02 -2.47715842e-02 -5.96393585e-01
4.90060955e-01 -3.85991074e-02 -2.39453897e-01 3.45889091e-01
1.88622326e-01 -2.11302757e-01 9.28829908e-02 3.32390070e-01
1.51681864e+00 2.46727765e-01 1.34106234e-01 -8.30967948e-02
4.25428391e-01 -1.41836584e-01 4.68406141e-01 7.41617262e-01
2.57605821e-01 6.02542162e-01 1.01121485e+00 -3.45220417e-01
-9.49612379e-01 -1.13996911e+00 6.94822490e-01 1.26169586e+00
8.75397027e-02 -7.42740035e-01 -9.86759126e-01 -7.90268242e-01
-1.06238656e-01 8.00984442e-01 -4.63528007e-01 -3.12798917e-01
-9.45354342e-01 -3.87915701e-01 9.05254245e-01 7.45998085e-01
4.59623128e-01 -1.22657478e+00 -1.61714613e-01 4.80343670e-01
-1.45625770e-01 -1.34858489e+00 -6.71848238e-01 1.40503868e-01
-1.14744675e+00 -9.15322661e-01 -4.95265722e-01 -1.21150601e+00
6.40355110e-01 5.24513498e-02 1.53049624e+00 2.66051024e-01
5.73770422e-03 -5.10651246e-02 -3.93202305e-01 -1.52155176e-01
-5.52691817e-01 3.97250235e-01 -6.71935201e-01 -2.88896888e-01
-2.38990206e-02 -5.41323721e-01 -2.33004838e-01 -3.00407469e-01
-9.71349299e-01 5.78031838e-01 5.89316070e-01 7.06911027e-01
4.39345300e-01 -1.61702082e-01 6.91503108e-01 -1.43445492e+00
7.85716593e-01 -5.13644397e-01 -6.88486099e-01 4.29283917e-01
-2.02760711e-01 4.35635090e-01 1.22222960e+00 -1.37371823e-01
-7.27661312e-01 -4.42751683e-02 -5.29026747e-01 -4.01029855e-01
-8.82016048e-02 8.03294659e-01 -2.19584703e-01 2.00743854e-01
2.25571454e-01 4.78274375e-01 -1.69870809e-01 -5.27997017e-01
5.30722082e-01 3.87216091e-01 7.43840456e-01 -8.69711339e-01
4.99236971e-01 -2.62852192e-01 1.30156398e-01 -5.67615807e-01
-1.00958133e+00 -3.59620526e-02 -4.27079707e-01 1.87585220e-01
1.17647314e+00 -7.82615781e-01 -6.22347057e-01 4.56739366e-01
-1.68407035e+00 -6.42231166e-01 -1.45688072e-01 1.56040024e-02
-5.78638971e-01 3.20162147e-01 -1.08259368e+00 -6.87367916e-01
-4.94469553e-01 -1.11428499e+00 1.30421996e+00 1.13466024e-01
1.05942950e-01 -1.12103617e+00 -1.72074094e-01 2.74696499e-01
3.12163085e-01 1.73375085e-01 1.73159719e+00 -5.61762750e-01
-8.89196694e-01 1.52169466e-01 -4.62377131e-01 1.80314228e-01
-1.14371479e-01 -4.11865592e-01 -6.02260411e-01 8.01415816e-02
-4.17202145e-01 -4.35875267e-01 1.07383907e+00 1.53567240e-01
1.83370185e+00 -4.94628072e-01 -5.45756966e-02 9.28812742e-01
1.41437960e+00 9.76836979e-02 7.04139411e-01 -2.59012610e-01
9.94793177e-01 5.26859343e-01 -1.60371408e-01 -2.15576589e-01
6.23173535e-01 3.05663705e-01 5.68143904e-01 1.17993145e-03
-2.85544425e-01 -7.90697455e-01 5.48896194e-01 1.26312506e+00
1.75941557e-01 -7.09647417e-01 -7.84664869e-01 4.80938882e-01
-1.93512440e+00 -5.59116483e-01 -3.80284101e-01 1.65868843e+00
7.50045478e-01 -2.99098212e-02 -2.02998906e-01 -2.85481066e-01
6.74348176e-01 4.33740139e-01 -4.79211539e-01 -9.01731193e-01
4.62979786e-02 7.25959599e-01 6.05955184e-01 4.78102773e-01
-7.64647067e-01 1.60386908e+00 6.96679258e+00 6.25193715e-01
-7.84088850e-01 2.89036185e-02 6.01309180e-01 2.75144607e-01
-6.15603447e-01 -2.09376946e-01 -4.86543864e-01 2.84666151e-01
1.16060638e+00 -6.46820813e-02 7.89528370e-01 6.66402936e-01
-2.24122837e-01 2.65371442e-01 -1.27321351e+00 7.94412374e-01
-1.00254137e-02 -1.54697335e+00 4.56652343e-01 -2.49013200e-01
5.47707856e-01 -3.64468098e-02 -1.53576642e-01 5.43448508e-01
7.07806826e-01 -1.48552370e+00 5.29065132e-01 3.77509326e-01
9.78754997e-01 -4.65229481e-01 5.67035317e-01 3.34153444e-01
-1.38844347e+00 7.35967010e-02 -4.04645830e-01 -4.42053437e-01
1.67448461e-01 3.62298161e-01 -4.60491300e-01 8.35828424e-01
3.17087770e-01 9.48366284e-01 -3.70406061e-01 6.55225217e-01
-7.68992901e-01 8.67979467e-01 1.85805321e-01 -3.12132031e-01
6.24974430e-01 -2.88202077e-01 2.66735375e-01 1.34696662e+00
3.29671890e-01 6.04008794e-01 3.20883393e-01 1.09547162e+00
-6.07210279e-01 9.23270732e-02 -9.26494598e-01 -4.96450186e-01
2.09645525e-01 8.59500468e-01 -5.57924390e-01 -5.82340181e-01
-6.19173765e-01 1.14328039e+00 8.32945049e-01 4.70035344e-01
-7.41363049e-01 -5.68215907e-01 8.06959987e-01 -3.08980614e-01
3.37311208e-01 -5.17261684e-01 -4.89947140e-01 -1.38749373e+00
-2.20320627e-01 -7.46123254e-01 6.21787310e-01 -9.24884081e-01
-1.11651826e+00 5.55853844e-01 -2.85274982e-01 -4.52104956e-01
-2.87508011e-01 -8.40130627e-01 -7.41868198e-01 8.62346172e-01
-1.43006361e+00 -8.88266921e-01 -1.78668164e-02 3.39171708e-01
8.35843563e-01 -6.50088340e-02 8.50995660e-01 1.76513240e-01
-6.22655988e-01 6.12605453e-01 -1.40776798e-01 5.72917759e-01
8.29912536e-03 -1.39260888e+00 1.21080554e+00 7.86634326e-01
2.48677835e-01 4.79561269e-01 2.94404507e-01 -6.97852552e-01
-2.05455256e+00 -1.43586981e+00 1.14596748e+00 -2.56608248e-01
6.75779641e-01 -7.05023289e-01 -1.07549477e+00 9.38965499e-01
3.94702673e-01 -1.77849427e-01 3.86022866e-01 -5.69816753e-02
-4.18670207e-01 3.11575294e-01 -7.79408514e-01 5.78705907e-01
1.70552289e+00 -5.64583778e-01 -7.32161224e-01 3.02895457e-01
1.58224976e+00 -8.04669797e-01 -6.29404843e-01 3.17771345e-01
3.40761513e-01 -5.71679294e-01 8.36153686e-01 -1.28318489e+00
1.11396682e+00 7.34025985e-02 -2.43051901e-01 -1.40125287e+00
-4.72809941e-01 -5.15367925e-01 -3.07708323e-01 1.01542783e+00
4.52562243e-01 -5.58615923e-01 9.36118007e-01 8.52734596e-02
-4.67106164e-01 -8.45622361e-01 -7.41138279e-01 -7.89487541e-01
2.53719300e-01 -4.90937203e-01 7.44149208e-01 7.99713433e-01
-2.02935636e-01 8.06511998e-01 -2.16555148e-01 5.79757355e-02
5.45883179e-01 4.84718412e-01 5.49059629e-01 -9.46463406e-01
-4.35503930e-01 -4.98066604e-01 -3.12473059e-01 -1.42819512e+00
7.48299241e-01 -1.49779344e+00 6.23221360e-02 -1.92334318e+00
1.21820875e-01 -3.75745259e-02 -1.36507690e-01 6.53909564e-01
-4.15610641e-01 1.25896797e-01 2.09309176e-01 -3.39708030e-01
-6.07964635e-01 6.57144547e-01 1.45857191e+00 -3.82631242e-01
8.33831802e-02 -2.57023841e-01 -8.14984739e-01 4.19016570e-01
6.33261502e-01 -5.22636533e-01 -4.61518884e-01 -1.19105077e+00
5.12608707e-01 5.76025307e-01 2.11997524e-01 -6.09576225e-01
1.71353787e-01 -3.13005447e-01 -2.57578865e-02 -2.94492871e-01
9.33864415e-02 -4.68516052e-01 -1.28979042e-01 5.61007023e-01
-7.43775666e-01 4.84348744e-01 1.97425753e-01 5.83911717e-01
-3.11650932e-01 -2.91490644e-01 4.11101460e-01 -6.32545590e-01
-7.81587362e-01 4.71721172e-01 -3.64949524e-01 3.78115594e-01
5.94698310e-01 -1.71325486e-02 -5.33668637e-01 -5.72760820e-01
-6.24387681e-01 3.92511666e-01 1.17913745e-01 4.84580815e-01
7.82331645e-01 -1.22829354e+00 -7.95596182e-01 2.78957784e-01
-8.20858330e-02 3.15380841e-01 -1.07729435e-01 3.63462865e-01
-6.80045366e-01 5.94397008e-01 2.91957008e-03 -2.73401678e-01
-1.08719122e+00 6.34646058e-01 1.98487610e-01 -3.82764757e-01
-6.58027709e-01 1.02450728e+00 7.07125962e-01 -8.88067007e-01
3.30826759e-01 -7.82251120e-01 -1.52684040e-02 -4.43467885e-01
3.04226965e-01 9.13468599e-02 -6.74692839e-02 -2.17406422e-01
-6.57192394e-02 5.94786525e-01 2.13446707e-01 2.61262864e-01
1.25572729e+00 2.59597600e-01 -5.60188234e-01 3.66209418e-01
1.29010451e+00 -3.61553073e-01 -7.21135378e-01 -3.65603209e-01
1.59902424e-01 -8.74821767e-02 -2.98652530e-01 -7.05578029e-01
-1.16493595e+00 1.24058223e+00 -2.85399556e-01 1.53938770e-01
1.14629471e+00 1.53290063e-01 1.28397679e+00 5.90309501e-01
3.66497844e-01 -5.80583572e-01 9.68048647e-02 1.13045144e+00
4.97779250e-01 -7.64345527e-01 -5.28439164e-01 -8.95008028e-01
-4.13325340e-01 1.21918392e+00 8.25139940e-01 -1.96526423e-01
1.77636683e-01 3.88501108e-01 -2.75371343e-01 -1.05890766e-01
-1.27764368e+00 -3.28926027e-01 3.34836543e-01 4.38305140e-01
7.10955977e-01 1.57236114e-01 -3.22798580e-01 8.17387819e-01
-4.26859796e-01 4.21730541e-02 3.42357785e-01 7.33818829e-01
-4.96864825e-01 -1.12285805e+00 -1.60438158e-02 4.86408383e-01
-5.81409037e-01 -6.12103879e-01 -7.29801893e-01 4.73296493e-01
-1.92122802e-01 7.17859149e-01 2.18525276e-01 -5.15594363e-01
3.37765038e-01 3.48698646e-01 7.24892437e-01 -1.11446619e+00
-8.80693376e-01 -4.01404500e-01 3.53950143e-01 -7.41019785e-01
9.67006311e-02 -1.28364488e-01 -1.48490143e+00 -3.32750916e-01
1.01621307e-01 2.40084574e-01 5.85352123e-01 9.26867664e-01
4.09857005e-01 1.00261319e+00 1.93903297e-01 -3.11027467e-01
-4.67938632e-01 -6.91677809e-01 -2.78748184e-01 1.77573204e-01
2.41781592e-01 5.73254079e-02 2.68111557e-01 8.48074779e-02] | [10.068853378295898, 8.072834968566895] |
b90e760c-f5ec-4657-aa2c-82b40494a476 | some-algorithms-on-exact-approximate-and | 2012.15279 | null | https://arxiv.org/abs/2012.15279v1 | https://arxiv.org/pdf/2012.15279v1.pdf | Some Algorithms on Exact, Approximate and Error-Tolerant Graph Matching | The graph is one of the most widely used mathematical structures in engineering and science because of its representational power and inherent ability to demonstrate the relationship between objects. The objective of this work is to introduce the novel graph matching techniques using the representational power of the graph and apply it to structural pattern recognition applications. We present an extensive survey of various exact and inexact graph matching techniques. Graph matching using the concept of homeomorphism is presented. A category of graph matching algorithms is presented, which reduces the graph size by removing the less important nodes using some measure of relevance. We present an approach to error-tolerant graph matching using node contraction where the given graph is transformed into another graph by contracting smaller degree nodes. We use this scheme to extend the notion of graph edit distance, which can be used as a trade-off between execution time and accuracy. We describe an approach to graph matching by utilizing the various node centrality information, which reduces the graph size by removing a fraction of nodes from both graphs based on a given centrality measure. The graph matching problem is inherently linked to the geometry and topology of graphs. We introduce a novel approach to measure graph similarity using geometric graphs. We define the vertex distance between two geometric graphs using the position of their vertices and show it to be a metric over the set of all graphs with vertices only. We define edge distance between two graphs based on the angular orientation, length and position of the edges. Then we combine the notion of vertex distance and edge distance to define the graph distance between two geometric graphs and show it to be a metric. Finally, we use the proposed graph similarity framework to perform exact and error-tolerant graph matching. | ['Shri Prakash Dwivedi'] | 2020-12-30 | null | null | null | null | ['graph-similarity'] | ['graphs'] | [ 3.23066652e-01 3.45273525e-01 2.89822936e-01 -3.07222456e-01
2.52992921e-02 -8.06319356e-01 3.98078382e-01 8.46875608e-01
-2.32625216e-01 2.15169057e-01 -3.33723456e-01 -2.02466130e-01
-6.92400336e-01 -1.42781878e+00 -4.91071224e-01 -3.20167869e-01
-3.97773802e-01 4.85209674e-01 4.79163259e-01 -4.79209483e-01
4.67954546e-01 9.88710880e-01 -1.49463832e+00 -3.12130004e-01
7.40783572e-01 5.26893198e-01 5.24832085e-02 7.18666732e-01
-1.44093499e-01 9.85447466e-02 -2.91972995e-01 -3.72980207e-01
6.26254499e-01 -5.78426957e-01 -8.85960758e-01 7.97045678e-02
5.04173160e-01 3.86190504e-01 -3.03842813e-01 1.23852718e+00
4.95822281e-01 2.68612325e-01 4.78300273e-01 -1.60409856e+00
-3.55860204e-01 3.91942650e-01 -3.90044808e-01 7.99713470e-03
8.73955429e-01 -5.28441846e-01 1.04538965e+00 -1.93640172e-01
8.77691388e-01 1.09576321e+00 7.03897834e-01 6.08034320e-02
-1.15761697e+00 -2.47224689e-01 -3.14151317e-01 3.46123070e-01
-1.62956178e+00 1.93289071e-01 1.00698805e+00 -4.47647423e-01
7.71927893e-01 7.07885325e-01 7.72401750e-01 -2.22049788e-01
2.71483243e-01 -2.49618262e-01 8.55092645e-01 -9.28659379e-01
1.34730533e-01 -1.48242220e-01 3.73549372e-01 1.11029506e+00
4.86936569e-01 2.70393882e-02 1.44241944e-01 -2.08249047e-01
6.31060362e-01 1.68329313e-01 -2.93697447e-01 -8.56623113e-01
-1.02010012e+00 7.58685470e-01 7.75209546e-01 7.79055357e-01
7.07440749e-02 2.42577240e-01 3.03750962e-01 7.00105429e-01
3.02053150e-02 6.68583095e-01 1.71324208e-01 3.13586950e-01
-5.76634407e-01 2.88136248e-02 1.04832757e+00 1.10179150e+00
1.22226965e+00 -3.96124274e-01 3.40425193e-01 5.63534319e-01
1.38347922e-02 2.44148940e-01 9.34322923e-02 -6.06749415e-01
1.18313931e-01 1.18467641e+00 -5.61196148e-01 -1.91533148e+00
-4.24000233e-01 4.96860594e-03 -7.48839676e-01 4.00908500e-01
2.09725976e-01 4.86466736e-01 -5.73923469e-01 1.65802407e+00
4.46511149e-01 1.09664358e-01 -3.13107044e-01 5.25458336e-01
6.48619950e-01 2.29504824e-01 -4.51049030e-01 -1.26967862e-01
1.16645730e+00 -5.91709197e-01 -2.93001771e-01 3.85666430e-01
8.56984735e-01 -8.12231719e-01 7.19631255e-01 -1.04183748e-01
-1.00572741e+00 -3.65018368e-01 -1.32995057e+00 1.07023522e-01
-8.26295197e-01 -5.59160113e-01 2.24452808e-01 7.39212453e-01
-1.27755916e+00 1.08520019e+00 -5.12949765e-01 -8.45335901e-01
-3.05015206e-01 5.20164371e-01 -7.42339194e-01 -1.79420244e-02
-8.32737148e-01 8.21909726e-01 5.41422427e-01 -1.91653147e-01
9.14922804e-02 -4.99480933e-01 -1.02236962e+00 1.21300392e-01
2.84852624e-01 -7.05441117e-01 4.65025425e-01 -8.96054566e-01
-9.28196430e-01 1.17479706e+00 1.88951463e-01 -2.58377314e-01
4.19695199e-01 6.74397111e-01 -5.88796675e-01 2.65525252e-01
5.74813783e-02 1.29846066e-01 4.25610393e-01 -9.08431888e-01
-3.85413617e-01 -4.79498267e-01 1.59571305e-01 4.49132845e-02
-1.11130700e-01 -1.88402638e-01 -2.31161743e-01 -5.16234279e-01
5.07872343e-01 -9.78480339e-01 -1.26005083e-01 7.56007284e-02
-3.22792560e-01 -2.00045466e-01 8.82051945e-01 -4.89482999e-01
1.41193509e+00 -1.95413339e+00 3.71447682e-01 1.14462793e+00
8.04324389e-01 -4.28689197e-02 -1.04830340e-01 9.61316347e-01
-4.75329548e-01 1.80905655e-01 -5.39077699e-01 3.50711852e-01
-8.87415111e-02 1.91977456e-01 2.62192875e-01 7.14383125e-01
-1.78036556e-01 6.42917216e-01 -1.04704332e+00 -5.24377406e-01
3.04626465e-01 2.31986493e-01 -3.06096792e-01 1.52107313e-01
2.56900638e-01 -8.38458091e-02 -3.34103286e-01 1.98313594e-01
7.78717220e-01 1.18502922e-01 5.04267514e-01 -4.70098436e-01
5.30491807e-02 -1.94732249e-01 -1.62672913e+00 1.42175388e+00
-2.50507683e-01 4.06907320e-01 -8.01369399e-02 -1.25187981e+00
1.49365842e+00 7.30500668e-02 5.87796509e-01 -5.24268866e-01
2.14837700e-01 3.88277531e-01 1.00918420e-01 -1.74770609e-01
2.93309182e-01 2.27654129e-02 -1.14285372e-01 6.68274462e-01
-3.23172241e-01 -3.53141963e-01 5.01835287e-01 5.43994725e-01
1.50560117e+00 -3.89671922e-01 6.98037148e-01 -6.10625684e-01
9.57930982e-01 -2.70966649e-01 2.31049880e-02 3.56255978e-01
8.96424949e-02 5.13163745e-01 4.89047676e-01 -5.69351554e-01
-1.22357142e+00 -1.15554166e+00 1.10442497e-01 5.01516461e-01
3.28883141e-01 -7.06362903e-01 -7.82065392e-01 -5.09999216e-01
2.59831697e-01 9.38919261e-02 -6.94135666e-01 -5.50212145e-01
-7.48367310e-01 -6.74683005e-02 2.86163419e-01 1.46642193e-01
2.44055554e-01 -7.17723966e-01 -4.37165946e-01 -2.40103435e-02
2.37896085e-01 -8.37401628e-01 -5.99554360e-01 -2.94935137e-01
-8.79490852e-01 -1.46892512e+00 -3.52153063e-01 -1.03081274e+00
1.17233133e+00 3.04535598e-01 1.13665521e+00 9.12742019e-01
-5.69521308e-01 6.39008820e-01 -5.63891768e-01 5.11524864e-02
-7.41354585e-01 8.30131099e-02 1.07678743e-02 -3.47011872e-02
-4.58089858e-02 -9.29521203e-01 -3.42782021e-01 4.20652241e-01
-1.12006772e+00 -1.18890405e-01 1.80615455e-01 3.99532765e-01
4.93592858e-01 1.57552496e-01 5.59739545e-02 -9.61373150e-01
7.78206170e-01 -1.95505843e-01 -7.07258403e-01 5.36558807e-01
-8.07870626e-01 5.58986664e-01 6.38998449e-01 -9.35018957e-02
-1.06199689e-01 5.99710792e-02 2.25868955e-01 -1.57626837e-01
4.29629475e-01 5.31732321e-01 -2.91347951e-01 -8.91862869e-01
4.54957783e-01 -9.19820070e-02 2.39532843e-01 -3.44546020e-01
5.02881646e-01 3.23486716e-01 4.58765864e-01 -4.30523127e-01
9.45615113e-01 3.38164002e-01 8.53398323e-01 -6.82586074e-01
7.62307271e-02 -6.96935952e-01 -8.21820259e-01 -3.10208887e-01
4.19693947e-01 1.59123495e-01 -9.67169583e-01 8.90938044e-02
-1.06563520e+00 3.63277972e-01 -3.02802354e-01 2.52394080e-01
-6.68680966e-01 9.78331029e-01 -1.50026038e-01 -4.63116765e-01
-4.17144030e-01 -8.47105443e-01 7.24421322e-01 -1.82632521e-01
-2.05093250e-01 -1.23251748e+00 4.21137452e-01 -3.48379910e-01
1.27033964e-01 9.64309514e-01 1.24097431e+00 -7.29231298e-01
-5.01277387e-01 -5.30678034e-01 -2.76496053e-01 -4.44895364e-02
3.51244271e-01 1.67928502e-01 -8.42594355e-02 -6.11756921e-01
-3.44078392e-01 5.93314648e-01 3.89298022e-01 -5.29744215e-02
5.98486364e-01 -1.47660449e-01 -6.56314611e-01 5.20395219e-01
1.93554509e+00 1.96090177e-01 7.67161191e-01 3.27009112e-01
8.94220054e-01 6.98227763e-01 3.30425858e-01 6.29451796e-02
-3.55565846e-02 9.28835571e-01 3.00042897e-01 1.59904864e-02
-1.57048717e-01 -2.73345441e-01 -5.27358130e-02 1.27274656e+00
-2.04394862e-01 -3.70849147e-02 -8.83382261e-01 3.23491037e-01
-1.76911771e+00 -8.17598999e-01 -4.83980745e-01 2.58479428e+00
2.18456432e-01 -8.65077153e-02 2.82314748e-01 6.11144006e-01
1.35427308e+00 -1.69358492e-01 1.44298241e-01 -9.35742497e-01
1.50925770e-01 4.35490489e-01 7.65639186e-01 9.76228476e-01
-4.85455096e-01 5.02477288e-01 5.93171072e+00 5.70637941e-01
-8.49314868e-01 -1.27165347e-01 -1.62018701e-01 5.11708319e-01
-3.64271790e-01 4.61651504e-01 -1.02552027e-01 3.21610004e-01
6.33419156e-01 -8.27766299e-01 4.80948091e-01 6.11936212e-01
-3.02782327e-01 6.20053560e-02 -1.25933754e+00 1.01685143e+00
2.58912235e-01 -1.16957104e+00 1.35234877e-01 1.21054165e-01
4.21024531e-01 -5.44189095e-01 -5.31755984e-01 -3.24659199e-01
5.14773056e-02 -1.05932128e+00 3.31904799e-01 6.77669764e-01
7.65345871e-01 -7.71930695e-01 5.95398128e-01 -9.59007591e-02
-1.99535692e+00 3.03626925e-01 -3.03196281e-01 -1.39700081e-02
1.27417684e-01 4.27075863e-01 -9.33179677e-01 1.09569979e+00
3.35738808e-01 5.18752575e-01 -6.66410863e-01 1.26404107e+00
1.19538061e-01 -1.85061470e-01 -4.57923383e-01 -1.34325996e-01
-4.43577059e-02 -8.11959088e-01 8.74123037e-01 8.89320970e-01
5.46974480e-01 -3.95503603e-02 1.59956694e-01 8.50216806e-01
-2.45355386e-02 4.63739306e-01 -1.11830318e+00 9.64434147e-02
6.90684140e-01 1.21489620e+00 -1.10623717e+00 -1.77794740e-01
-1.93731844e-01 9.86237288e-01 4.01855230e-01 -2.30848417e-01
-5.70297658e-01 -1.05688035e+00 2.16778576e-01 4.59861249e-01
-1.83928348e-02 -3.75407487e-01 6.58596531e-02 -5.00735104e-01
2.05224022e-01 -5.72742105e-01 5.58919847e-01 -6.99765682e-01
-1.03807712e+00 8.05259645e-01 1.74909338e-01 -1.25381446e+00
-1.08526699e-01 -5.63985407e-01 -6.42542601e-01 7.62252212e-01
-8.38977933e-01 -8.61353040e-01 -5.95795810e-01 7.07848608e-01
-3.42007846e-01 1.05654940e-01 8.15529108e-01 3.92610013e-01
2.91453488e-02 6.67560399e-01 1.98148452e-02 2.11305633e-01
4.21926856e-01 -1.24674964e+00 5.78897059e-01 7.90642262e-01
1.17596425e-01 6.08937442e-01 7.59505153e-01 -7.07737923e-01
-1.46343040e+00 -7.93258965e-01 1.06538224e+00 -3.18756044e-01
5.87484777e-01 -2.80592918e-01 -8.34108829e-01 6.04681134e-01
-2.61866391e-01 2.03549832e-01 1.73886150e-01 -2.54188418e-01
-5.63854218e-01 -3.77037823e-01 -1.48512328e+00 4.53947067e-01
1.36501360e+00 -5.51032722e-01 -5.63807070e-01 1.74440265e-01
7.65306234e-01 -6.14390559e-02 -1.26313698e+00 4.79108065e-01
5.79962611e-01 -9.33617711e-01 9.29621696e-01 -3.68423700e-01
-2.92668074e-01 -6.33159459e-01 -8.23900178e-02 -1.18494844e+00
-4.84037071e-01 -7.67104268e-01 4.92120653e-01 1.14880681e+00
1.80529341e-01 -8.66950333e-01 6.38470650e-01 2.49228328e-01
1.47017032e-01 -4.49501991e-01 -8.74772608e-01 -1.03314877e+00
-3.25210094e-01 3.62934656e-02 7.60480344e-01 1.13725889e+00
4.13050205e-01 1.73459992e-01 1.12401336e-01 1.24225300e-02
6.04724646e-01 3.54306608e-01 7.38118589e-01 -1.57036436e+00
-1.78758442e-01 -5.51048756e-01 -1.57131207e+00 -2.46062368e-01
-9.28604826e-02 -1.35611272e+00 -4.00731057e-01 -1.47326732e+00
2.22464232e-03 -5.09957790e-01 5.21027446e-02 1.65815838e-02
3.62654030e-01 1.84970424e-01 2.68746257e-01 -7.36251622e-02
-1.80661619e-01 -1.48168700e-02 1.01453531e+00 3.67936306e-02
-1.94170952e-01 -1.69862285e-01 -1.90127328e-01 3.46795261e-01
6.72153711e-01 -5.31358659e-01 -3.68861139e-01 3.07231545e-02
5.12662768e-01 -6.95091188e-02 2.50373214e-01 -1.00926554e+00
2.82266825e-01 7.94168115e-02 -4.03495729e-01 -1.38194887e-02
-6.48247972e-02 -1.31493819e+00 8.33345890e-01 9.33998406e-01
-5.52430227e-02 8.08984280e-01 -1.81785837e-01 5.91674507e-01
-1.32157147e-01 -3.89826566e-01 8.85739088e-01 -2.13122815e-01
-6.27267420e-01 2.28901386e-01 2.61649758e-01 -1.82861403e-01
1.37599707e+00 -8.24789703e-01 -1.02026157e-01 -3.28677654e-01
-8.86315703e-01 -2.55861908e-01 8.85928929e-01 3.36689234e-01
5.47659218e-01 -1.54013669e+00 -4.77888286e-01 1.72946692e-01
2.83798307e-01 -5.27979612e-01 -1.86862394e-01 7.39659369e-01
-1.06216133e+00 1.53545469e-01 -5.07415056e-01 -3.87828201e-01
-1.85861576e+00 8.02240074e-01 5.38721800e-01 -1.74118116e-01
-5.42274177e-01 2.55922794e-01 -2.41915673e-01 -5.37811160e-01
-1.44173205e-01 -2.84031928e-01 7.84205794e-02 -4.00441349e-01
1.31142735e-01 8.50142777e-01 4.18756127e-01 -8.13509643e-01
-5.08128881e-01 1.26095331e+00 3.45495880e-01 1.68115050e-01
1.01240003e+00 -1.10926636e-01 -7.61527419e-01 1.46893442e-01
1.65097570e+00 1.73496440e-01 -3.71749662e-02 6.66104630e-02
2.13350862e-01 -5.48986316e-01 -3.56387049e-01 -2.01540604e-01
-9.81574714e-01 5.22410035e-01 6.60471499e-01 8.22296321e-01
1.07887399e+00 -1.94388796e-02 6.83179677e-01 2.59394586e-01
4.51565862e-01 -9.22185361e-01 -2.47601286e-01 1.72397912e-01
8.50890100e-01 -6.98446214e-01 1.57587022e-01 -1.01552629e+00
2.24953473e-01 1.41101146e+00 2.29977846e-01 -5.91294408e-01
8.40140343e-01 1.57214612e-01 -5.44302583e-01 -5.41605294e-01
7.54969642e-02 -3.45139682e-01 5.53893745e-01 6.27672553e-01
3.90276194e-01 2.26006418e-01 -1.07729220e+00 -5.43741882e-02
-4.72613215e-01 -3.12969238e-01 4.30713326e-01 9.82886970e-01
-6.93586767e-01 -1.53042316e+00 -3.48832220e-01 2.86693901e-01
1.34604335e-01 9.48649943e-02 -8.34512949e-01 1.03597999e+00
2.42273994e-02 7.05577433e-01 1.48804933e-02 -6.20675683e-01
6.87911868e-01 -3.47210467e-01 8.21133733e-01 -5.21200418e-01
-4.38361257e-01 -7.41114020e-01 8.17418841e-05 -4.55144972e-01
-4.60002512e-01 -2.03139827e-01 -1.25317919e+00 -8.76986861e-01
-4.75572795e-01 3.76465082e-01 6.21352434e-01 6.26812041e-01
4.91368473e-01 1.90718412e-01 8.20885122e-01 -3.65808129e-01
-1.19734243e-01 -3.98944706e-01 -9.08327401e-01 9.16921496e-01
-1.04090683e-01 -5.73548675e-01 -4.35214013e-01 -2.06589699e-01] | [7.2398858070373535, 5.7393412590026855] |
30101149-d295-44ae-85b8-f916d38fc477 | ai-marker-based-large-scale-ai-literature | 2011.00518 | null | https://arxiv.org/abs/2011.00518v2 | https://arxiv.org/pdf/2011.00518v2.pdf | AI Marker-based Large-scale AI Literature Mining | The knowledge contained in academic literature is interesting to mine. Inspired by the idea of molecular markers tracing in the field of biochemistry, three named entities, namely, methods, datasets and metrics are used as AI markers for AI literature. These entities can be used to trace the research process described in the bodies of papers, which opens up new perspectives for seeking and mining more valuable academic information. Firstly, the entity extraction model is used in this study to extract AI markers from large-scale AI literature. Secondly, original papers are traced for AI markers. Statistical and propagation analysis are performed based on tracing results. Finally, the co-occurrences of AI markers are used to achieve clustering. The evolution within method clusters and the influencing relationships amongst different research scene clusters are explored. The above-mentioned mining based on AI markers yields many meaningful discoveries. For example, the propagation of effective methods on the datasets is rapidly increasing with the development of time; effective methods proposed by China in recent years have increasing influence on other countries, whilst France is the opposite. Saliency detection, a classic computer vision research scene, is the least likely to be affected by other research scenes. | ['Ou wu', 'Shuxiao Li', 'Ji Zhang', 'Yingchun Ye', 'Rujing Yao'] | 2020-11-01 | null | null | null | null | ['literature-mining'] | ['natural-language-processing'] | [ 5.89694232e-02 -2.42832378e-01 -3.26554388e-01 1.90732315e-01
-6.28535151e-02 -3.27082753e-01 7.44762361e-01 5.32584190e-01
-4.95555699e-01 8.93251956e-01 8.56458768e-02 1.84191782e-02
-5.17930090e-01 -8.52986038e-01 -5.84063053e-01 -8.55304062e-01
-3.63730788e-02 1.47542715e-01 2.12494597e-01 2.02807084e-01
1.05286908e+00 7.32433140e-01 -1.47831857e+00 -6.23209067e-02
1.23408842e+00 5.42091787e-01 4.61943507e-01 3.02421272e-01
-6.47449374e-01 6.44182265e-01 -1.00567782e+00 -5.21226108e-01
-2.92043269e-01 -6.21924996e-01 -6.61558926e-01 -1.22732997e-01
-1.86244473e-01 4.71521914e-01 -1.88937217e-01 1.27933156e+00
4.30997044e-01 -2.52450705e-01 8.78206611e-01 -1.25617480e+00
-1.02365077e+00 8.85696530e-01 -8.50129843e-01 4.32942480e-01
2.76740819e-01 4.58513759e-03 7.60806024e-01 -8.47745657e-01
1.15212274e+00 1.08732569e+00 1.21746995e-01 9.50661674e-02
-8.02542984e-01 -7.73373544e-01 -1.24215648e-01 8.04278672e-01
-1.60957921e+00 1.43730454e-02 9.44557428e-01 -7.84470022e-01
2.14240432e-01 3.80337358e-01 8.57494235e-01 9.30248499e-01
1.64042562e-01 8.06726456e-01 1.08949876e+00 -5.23976624e-01
2.10526511e-01 6.34351492e-01 2.87802845e-01 5.46080470e-01
8.71054888e-01 -3.68584812e-01 -8.61854613e-01 9.24180895e-02
2.86769092e-01 1.56552911e-01 -3.38643104e-01 1.67006686e-01
-1.47564816e+00 6.29622221e-01 3.86281997e-01 9.41950679e-01
-4.74723279e-01 -4.46198702e-01 2.62584120e-01 -1.24453403e-01
3.92069936e-01 1.14801741e+00 -2.57277340e-01 -2.83170223e-01
-1.20765924e+00 6.83679804e-02 5.08240104e-01 1.01510012e+00
7.58580327e-01 -7.52137676e-02 -2.55860656e-01 5.67558348e-01
1.21508129e-01 3.15706402e-01 5.71462989e-01 -7.29798675e-01
2.20184494e-02 1.36345696e+00 -2.55140185e-01 -1.46110415e+00
-5.43248296e-01 -4.63710099e-01 -6.74355805e-01 -2.54055738e-01
3.86269957e-01 -8.93771742e-03 -5.63399315e-01 1.10991597e+00
1.35004655e-01 2.29534850e-01 -1.40169814e-01 7.10404098e-01
1.19032669e+00 6.45658970e-01 2.38969833e-01 -5.50726235e-01
1.64248490e+00 -7.31504261e-01 -1.12232578e+00 4.20893133e-01
4.68276292e-01 -7.33037472e-01 4.56509709e-01 5.00042558e-01
-7.97745705e-01 -4.78630722e-01 -9.36108291e-01 1.22577004e-01
-1.07396615e+00 3.19257528e-01 6.57209933e-01 3.18449080e-01
-7.75290728e-01 5.99420309e-01 -4.66296971e-01 -5.99145710e-01
9.04210567e-01 3.32529540e-03 1.07444100e-01 4.58828136e-02
-1.18426967e+00 1.06857955e+00 4.81282920e-01 1.13757938e-01
-6.57261074e-01 -9.03519094e-01 -2.49171287e-01 -2.17786238e-01
6.08641803e-01 -5.45243979e-01 2.33712792e-01 -4.43924427e-01
-1.11017132e+00 1.00946927e+00 -2.47744948e-01 -1.96026385e-01
3.66365671e-01 -1.93049476e-01 -8.35126281e-01 3.08014512e-01
4.22641248e-01 3.91018093e-01 2.77770817e-01 -1.10131466e+00
-9.69610751e-01 -6.10921979e-01 -4.29460287e-01 -1.35565531e-02
-7.26639152e-01 3.96178514e-01 -6.39964640e-01 -6.95566356e-01
-2.05599710e-01 -7.15846300e-01 -1.51087552e-01 -2.40336284e-01
-5.91874182e-01 -6.33114994e-01 7.79071033e-01 -6.13874078e-01
1.62968886e+00 -1.83438814e+00 5.50258636e-01 2.82834530e-01
7.78555810e-01 1.45513862e-01 4.00241852e-01 2.27568984e-01
2.63017148e-01 5.49243510e-01 -6.08547665e-02 7.47312665e-01
-2.25891158e-01 -4.05161381e-01 1.72792375e-01 5.35521090e-01
2.34785557e-01 8.16532433e-01 -1.15474796e+00 -9.85871673e-01
-9.57196429e-02 2.02292111e-02 1.98827833e-01 2.70158108e-02
1.96685437e-02 3.80582064e-01 -8.24516833e-01 1.01518285e+00
4.42478597e-01 -2.75242597e-01 -1.77311182e-01 -2.59289801e-01
-7.14454174e-01 -9.60440338e-02 -6.68382347e-01 1.52251399e+00
2.07088068e-01 9.36967015e-01 -1.92345262e-01 -1.05839324e+00
1.27691495e+00 1.20119378e-01 7.30742872e-01 -5.32073319e-01
1.86711445e-01 3.26367974e-01 3.98298174e-01 -5.95303595e-01
3.98859233e-01 4.85324055e-01 8.15383419e-02 7.17645735e-02
-1.91076770e-01 5.49553372e-02 8.38925064e-01 4.51831579e-01
9.43070471e-01 8.52942616e-02 2.90668130e-01 -4.72161412e-01
8.11442614e-01 5.22373676e-01 7.80336738e-01 4.99059319e-01
-3.21875662e-01 -1.15259118e-01 4.56299037e-01 -2.44708955e-01
-9.02970433e-01 -6.25781119e-01 -5.41823804e-01 6.87942386e-01
3.54473829e-01 -4.73917663e-01 -6.22443974e-01 -2.89393753e-01
-1.48111051e-02 4.89187688e-01 -7.49193370e-01 -1.35074705e-01
-3.15698892e-01 -8.36555421e-01 6.29957139e-01 5.28789721e-02
7.96505392e-01 -1.43982494e+00 -7.13564336e-01 1.67382434e-01
9.33168158e-02 -8.32843125e-01 2.53244191e-01 -1.40046194e-01
-6.49426877e-01 -1.14179528e+00 -1.26504946e+00 -7.21690416e-01
5.81606507e-01 2.60981590e-01 9.63705838e-01 4.48574163e-02
-7.10092902e-01 -1.20349757e-01 -3.10110271e-01 -1.17999196e+00
-3.68086964e-01 1.76594794e-01 -1.71972495e-02 -2.06033096e-01
8.07589829e-01 -1.14202768e-01 -4.41922456e-01 9.95381698e-02
-5.10565400e-01 -5.39570898e-02 8.03114891e-01 3.04393798e-01
4.99802977e-01 2.81706184e-01 6.53428078e-01 -9.70772147e-01
5.62585831e-01 -7.28101254e-01 -4.84758049e-01 4.94611770e-01
-1.03518629e+00 -5.81616685e-02 2.96370655e-01 -3.92219156e-01
-9.16420460e-01 -4.27421182e-01 5.99816442e-01 -2.07331195e-01
-3.58848959e-01 6.66914523e-01 -2.97846705e-01 2.27866143e-01
5.13493359e-01 2.30323166e-01 -2.87166476e-01 -3.74740869e-01
2.87195414e-01 7.71128774e-01 2.35141844e-01 -4.71786499e-01
5.76964200e-01 1.95059001e-01 2.91839361e-01 -1.16087449e+00
-5.58444738e-01 -6.77865863e-01 -6.60101175e-01 -4.52953935e-01
1.15205133e+00 -3.23791504e-01 -9.93976772e-01 2.85562694e-01
-1.09571028e+00 4.41126347e-01 8.86472315e-03 6.51471078e-01
7.84293190e-02 3.46011072e-01 -4.21297848e-01 -8.14796567e-01
-2.97873884e-01 -1.22212219e+00 5.69749236e-01 8.88022959e-01
-4.16792154e-01 -7.93937087e-01 -1.04716793e-01 2.31388450e-01
9.33218449e-02 2.79393524e-01 9.54496920e-01 -5.94820380e-01
-5.33336163e-01 1.88838691e-01 -4.25365567e-01 -4.09759194e-01
3.08811754e-01 4.83426362e-01 -5.47070861e-01 3.59600961e-01
-1.34511858e-01 4.72895861e-01 8.26496840e-01 4.18450922e-01
1.42805707e+00 -1.49103686e-01 -9.36186433e-01 4.35991228e-01
1.00446916e+00 8.53908896e-01 5.50505579e-01 9.68998373e-01
1.06310332e+00 7.48835087e-01 7.89460957e-01 2.19831556e-01
-3.84222083e-02 4.81021106e-01 -1.17392160e-01 -2.05004096e-01
9.16413218e-02 2.01850206e-01 1.65884510e-01 1.05317807e+00
-6.13026738e-01 -2.22041570e-02 -1.17627275e+00 6.17562711e-01
-1.56868792e+00 -1.13436460e+00 -7.48188436e-01 1.84531879e+00
1.11116862e+00 6.47197217e-02 2.48816535e-01 3.46025117e-02
8.28499496e-01 -1.35620788e-01 -6.32692933e-01 -3.55667979e-01
-4.79892462e-01 -2.28161570e-02 5.58369756e-01 -3.10764551e-01
-9.39763844e-01 9.73060548e-01 6.01757336e+00 9.25610363e-01
-9.65865135e-01 -2.61387438e-01 6.95087314e-01 1.71744853e-01
-2.44835168e-01 -5.25533520e-02 -7.83427656e-01 8.60020757e-01
1.04892933e+00 -9.03856754e-01 9.47732702e-02 6.51626527e-01
2.37791419e-01 -2.67188132e-01 -7.82455027e-01 8.91402543e-01
8.26079845e-02 -1.64634264e+00 2.51519650e-01 3.62843335e-01
6.36724293e-01 -4.43400115e-01 -1.91007659e-01 -1.59552231e-01
4.92827982e-01 -7.95248866e-01 4.24100131e-01 1.02628446e+00
6.02548242e-01 -8.57235372e-01 7.74930716e-01 1.16929360e-01
-9.27879333e-01 1.49596542e-01 -5.19543409e-01 1.84510931e-01
-1.63146973e-01 1.04169488e+00 -8.79509985e-01 8.15738320e-01
8.84238422e-01 1.11038041e+00 -9.36778843e-01 1.29975379e+00
-3.48143488e-01 8.41732800e-01 2.59399980e-01 -5.70421576e-01
7.15818852e-02 -7.59740293e-01 7.00081587e-01 1.48816586e+00
3.07619780e-01 -1.20867305e-01 -1.18932627e-01 1.32084382e+00
-2.43880689e-01 3.61313105e-01 -7.58385181e-01 -7.32848346e-01
6.10740185e-01 1.33623230e+00 -1.23390293e+00 -4.47584659e-01
-1.75273731e-01 8.39621127e-01 5.93340322e-02 3.54944021e-01
-7.65901923e-01 -8.18780363e-01 3.28085184e-01 -8.70368779e-02
-1.36303157e-01 -1.22018784e-01 -5.30862510e-01 -6.45861804e-01
-4.69187975e-01 -6.22463167e-01 3.58869910e-01 -6.41434371e-01
-1.21233785e+00 2.65236974e-01 -1.85400099e-01 -1.03019738e+00
4.53643084e-01 -4.44983035e-01 -3.94785553e-01 7.42239535e-01
-1.20379758e+00 -5.33517659e-01 -2.06467032e-01 1.62250325e-01
5.37140727e-01 -6.02106750e-01 6.40883088e-01 1.74832895e-01
-1.10742378e+00 3.18136722e-01 2.97236025e-01 2.27708176e-01
6.51350677e-01 -1.25956988e+00 -1.00469507e-01 7.39495575e-01
3.10740024e-01 9.40611720e-01 5.02764106e-01 -9.72248256e-01
-1.39835954e+00 -8.28796029e-01 7.82299757e-01 -6.25946224e-01
1.06494772e+00 -4.52826619e-02 -9.31888819e-01 -2.91554034e-02
1.80625513e-01 -6.24861002e-01 9.57973838e-01 8.26589689e-02
9.50178206e-02 -7.13007972e-02 -7.24602103e-01 8.34175289e-01
9.72492278e-01 -1.35116890e-01 -6.04231060e-01 2.12433204e-01
6.88826442e-01 1.11288801e-01 -1.03612292e+00 -1.86496582e-02
3.39539289e-01 -2.97913402e-01 8.34485233e-01 -6.10128164e-01
7.99736500e-01 -4.31780547e-01 4.59090382e-01 -1.14535499e+00
-6.74133539e-01 -4.23764706e-01 -2.59765126e-02 1.65509701e+00
6.73872709e-01 -2.46410757e-01 5.91477394e-01 2.52615750e-01
-7.23031908e-02 -6.60518765e-01 -4.41146046e-01 -4.93429661e-01
-5.21912612e-02 1.83254145e-02 5.00493646e-01 1.30991316e+00
2.08728790e-01 4.37726438e-01 3.99885774e-02 -3.65873694e-01
6.98919594e-01 1.97069034e-01 4.97753799e-01 -1.58144724e+00
4.68962491e-01 -1.15925860e+00 -7.13386774e-01 -3.34910125e-01
-1.12437159e-01 -1.03222203e+00 -2.36717984e-01 -1.79966784e+00
5.10110617e-01 -2.64554113e-01 -4.46424425e-01 7.88675770e-02
-5.41185021e-01 -3.35584518e-05 2.32508019e-01 6.25164270e-01
-7.28419483e-01 2.08635181e-01 1.45128071e+00 -3.06958020e-01
-3.85575026e-01 -2.50904381e-01 -9.23648357e-01 8.34551930e-01
6.54900193e-01 -3.92084479e-01 3.09980102e-02 1.82203382e-01
3.99869859e-01 -5.33081651e-01 -1.06942102e-01 -9.03803706e-01
4.00168836e-01 -4.98369426e-01 5.56808054e-01 -7.73179889e-01
-3.58108401e-01 -7.59868741e-01 1.04903214e-01 4.11321729e-01
-3.44806910e-01 5.22687770e-02 9.35467407e-02 3.88041615e-01
-6.56585246e-02 -3.30293804e-01 3.28727365e-01 -2.86904812e-01
-8.36552083e-01 1.07792690e-01 -4.97586787e-01 4.87565389e-03
1.21642232e+00 -5.48160136e-01 -3.73606354e-01 3.25240761e-01
-4.67135489e-01 3.28703672e-01 2.32413843e-01 3.24613035e-01
4.63670611e-01 -1.05561697e+00 -7.73938537e-01 -4.17943537e-01
2.00696766e-01 -2.70753294e-01 -7.13835433e-02 1.03083813e+00
-5.74370384e-01 4.35152680e-01 -4.58669156e-01 -5.69118321e-01
-1.20885623e+00 9.96532381e-01 -3.36257875e-01 -8.23356137e-02
-5.41258454e-01 6.73216701e-01 2.12276459e-01 3.15070331e-01
2.74293661e-01 9.88982469e-02 -1.00080955e+00 6.87084973e-01
6.42686009e-01 9.49457765e-01 -2.32110888e-01 -6.24046087e-01
-6.31785214e-01 5.66407502e-01 -2.84817308e-01 1.89590648e-01
1.59591544e+00 8.34887549e-02 -5.19063711e-01 9.79924917e-01
8.00350666e-01 2.01648533e-01 -4.42494482e-01 -5.41795827e-02
7.03781903e-01 -3.09118062e-01 2.19599642e-02 -9.57303107e-01
-9.77139056e-01 8.00919652e-01 4.20307785e-01 2.81540960e-01
7.90989339e-01 1.83545396e-01 3.94013107e-01 6.97757071e-03
1.79082245e-01 -1.31862974e+00 -1.30574420e-01 1.37355089e-01
7.74912536e-01 -9.45300281e-01 4.37399983e-01 -4.68692094e-01
-5.29778242e-01 1.11948943e+00 6.02620900e-01 4.97411907e-01
4.32311386e-01 1.50573626e-01 -2.89623737e-01 -6.90952182e-01
-3.87345068e-02 -2.93819129e-01 4.75846022e-01 3.99090409e-01
6.10942006e-01 2.32588515e-01 -1.03137755e+00 5.46153307e-01
-1.26711950e-01 1.64848596e-01 5.63187480e-01 8.02163720e-01
-5.54502428e-01 -7.02508688e-01 -5.56904972e-01 6.46902442e-01
-6.61960781e-01 -7.39270598e-02 -1.12079847e+00 7.09872425e-01
4.68948632e-01 8.65378082e-01 2.94120349e-02 -3.27396899e-01
1.57146156e-01 -2.89758909e-02 1.68506756e-01 -2.98235089e-01
-4.02342170e-01 8.44333619e-02 -9.53586549e-02 -1.04335763e-01
-8.97847176e-01 -5.77905595e-01 -1.58184254e+00 -2.91992813e-01
-5.22299051e-01 5.91511071e-01 6.98912501e-01 8.43673825e-01
4.76453066e-01 9.39079285e-01 4.41810220e-01 -3.06290805e-01
5.30031443e-01 -1.06369054e+00 -6.89362407e-01 3.41178358e-01
-4.05439794e-01 -9.27420855e-01 -3.05327713e-01 2.53715754e-01] | [9.4851713180542, 8.141538619995117] |
141480d0-1193-46f7-944c-64125e1a4e4b | unsupervised-opinion-summarization-with | 2004.10150 | null | https://arxiv.org/abs/2004.10150v1 | https://arxiv.org/pdf/2004.10150v1.pdf | Unsupervised Opinion Summarization with Noising and Denoising | The supervised training of high-capacity models on large datasets containing hundreds of thousands of document-summary pairs is critical to the recent success of deep learning techniques for abstractive summarization. Unfortunately, in most domains (other than news) such training data is not available and cannot be easily sourced. In this paper we enable the use of supervised learning for the setting where there are only documents available (e.g.,~product or business reviews) without ground truth summaries. We create a synthetic dataset from a corpus of user reviews by sampling a review, pretending it is a summary, and generating noisy versions thereof which we treat as pseudo-review input. We introduce several linguistically motivated noise generation functions and a summarization model which learns to denoise the input and generate the original review. At test time, the model accepts genuine reviews and generates a summary containing salient opinions, treating those that do not reach consensus as noise. Extensive automatic and human evaluation shows that our model brings substantial improvements over both abstractive and extractive baselines. | ['Mirella Lapata', 'Reinald Kim Amplayo'] | 2020-04-21 | unsupervised-opinion-summarization-with-1 | https://aclanthology.org/2020.acl-main.175 | https://aclanthology.org/2020.acl-main.175.pdf | acl-2020-6 | ['unsupervised-opinion-summarization'] | ['natural-language-processing'] | [ 5.57315707e-01 6.97757602e-01 -1.11042909e-01 -4.57298249e-01
-1.59739828e+00 -7.61258364e-01 9.72655237e-01 6.25973403e-01
-4.32067931e-01 1.08145976e+00 9.96153474e-01 -3.48429084e-01
4.25445586e-01 -6.27289474e-01 -8.28152776e-01 -3.24638516e-01
5.12225270e-01 5.97749710e-01 -2.94655383e-01 -2.68389940e-01
4.66825634e-01 -1.27655566e-01 -1.30642080e+00 6.70414329e-01
9.78775442e-01 6.83934927e-01 1.01737179e-01 1.00017905e+00
-3.81094128e-01 8.05207312e-01 -1.49414861e+00 -7.20174611e-01
2.24887773e-01 -8.19186985e-01 -8.24587584e-01 3.75234544e-01
6.55195892e-01 -5.25629997e-01 5.01317345e-02 1.02776706e+00
6.27787948e-01 1.55831218e-01 8.47320974e-01 -7.14051068e-01
-8.55105877e-01 1.19411004e+00 -6.09086812e-01 2.92416692e-01
4.27230537e-01 4.09932062e-02 1.34435296e+00 -9.35693681e-01
8.11386704e-01 1.31638193e+00 3.89108658e-01 6.16916776e-01
-1.26926720e+00 -2.33647242e-01 3.11566323e-01 -4.35338616e-01
-5.04693627e-01 -8.25920224e-01 6.87432587e-01 -9.30217430e-02
9.80126262e-01 2.48930290e-01 4.41684484e-01 1.50065446e+00
2.99540669e-01 9.81730402e-01 5.61958194e-01 -5.78312933e-01
5.23722291e-01 1.39258489e-01 2.42391765e-01 5.83957136e-03
7.00821698e-01 -5.99458218e-01 -5.76484740e-01 -2.94909626e-01
1.67359272e-03 -2.62732148e-01 -1.52069196e-01 1.88955039e-01
-1.14582098e+00 8.33204210e-01 2.30285108e-01 6.41391501e-02
-7.92405963e-01 1.36277184e-01 8.37507784e-01 5.02899110e-01
1.07138765e+00 7.15537429e-01 -2.44708151e-01 -5.21863662e-02
-1.31459665e+00 7.62186229e-01 1.04907560e+00 9.00010109e-01
4.56611514e-01 2.42001936e-01 -3.27117473e-01 7.62075424e-01
-1.74485043e-01 6.24639094e-01 6.52341843e-01 -1.00737917e+00
7.96341419e-01 4.09810513e-01 5.82125962e-01 -9.43301499e-01
-4.21012938e-02 -5.39734542e-01 -8.67873549e-01 -5.06736822e-02
-3.29685472e-02 -6.18563235e-01 -9.41096425e-01 1.41518354e+00
-1.78718999e-01 -2.90700316e-01 6.31527960e-01 6.37273848e-01
1.14080119e+00 8.97725642e-01 -7.74121061e-02 -5.06713510e-01
9.04103220e-01 -1.02779257e+00 -9.25900161e-01 -7.90596068e-01
5.71842730e-01 -8.64191592e-01 1.05740345e+00 5.78462601e-01
-1.39148808e+00 -2.80563325e-01 -1.30074453e+00 -2.84807950e-01
-1.75219610e-01 3.31642359e-01 2.10026726e-01 8.91133863e-03
-1.16841340e+00 6.40252888e-01 -5.48620403e-01 -3.17372292e-01
4.27446157e-01 2.45043859e-02 -2.99145490e-01 -1.53782070e-01
-1.02882338e+00 8.70562077e-01 9.90956053e-02 6.97663799e-02
-7.17796266e-01 -3.20705116e-01 -9.96368825e-01 9.57560316e-02
3.58781993e-01 -8.80079806e-01 1.90156019e+00 -1.40622807e+00
-1.16387999e+00 7.67331898e-01 -3.52454603e-01 -7.79174685e-01
5.23663700e-01 -4.58693296e-01 -1.85602888e-01 2.26004109e-01
6.52222574e-01 6.00091517e-01 9.00956273e-01 -1.38773441e+00
-5.77453971e-01 -4.57510911e-02 1.26347244e-01 3.64846826e-01
-1.13160163e-01 1.67184383e-01 5.99332564e-02 -7.41714537e-01
-2.71978885e-01 -5.02289951e-01 -4.07587409e-01 -5.96738219e-01
-8.10850263e-01 -1.83836639e-01 5.05984426e-01 -8.76739085e-01
1.22583306e+00 -1.82341838e+00 6.66343644e-02 -9.21026096e-02
2.56555319e-01 3.43082368e-01 -5.18046737e-01 6.58704221e-01
-2.52689328e-02 5.12430668e-01 -3.84618253e-01 -7.62175143e-01
7.47851059e-02 7.19437152e-02 -8.61753941e-01 3.53989378e-02
4.74058688e-01 1.03514040e+00 -1.22768295e+00 -2.61722326e-01
-3.24429721e-01 1.01554491e-01 -3.02315414e-01 1.81510448e-01
-6.15778685e-01 2.00883928e-03 -4.29560721e-01 1.70292214e-01
3.48809183e-01 -2.20911682e-01 -7.81206265e-02 5.09259477e-02
1.22171648e-01 9.73292410e-01 -7.53484786e-01 1.48347926e+00
-5.04780293e-01 8.76275957e-01 8.72033536e-02 -8.72354567e-01
9.63577986e-01 3.67650867e-01 -7.62991533e-02 -3.31689894e-01
1.43233806e-01 3.34492981e-01 -2.34350160e-01 -2.58206248e-01
1.14480317e+00 -2.72881061e-01 -3.18325698e-01 9.77351308e-01
5.84123805e-02 -6.17896974e-01 5.49755394e-01 8.96266997e-01
1.34617865e+00 -2.91742295e-01 3.95460010e-01 -2.33000033e-02
1.83943227e-01 3.12764376e-01 3.11445326e-01 1.03539896e+00
2.01738074e-01 1.08997321e+00 1.03613985e+00 -2.16447800e-01
-1.42015934e+00 -7.64919400e-01 4.79557991e-01 7.51901209e-01
-3.31934124e-01 -6.29060924e-01 -8.09189498e-01 -6.38946831e-01
-1.99654028e-01 1.20310438e+00 -6.04733050e-01 -2.00148165e-01
-4.42267060e-01 -4.84995455e-01 1.98561460e-01 4.39180434e-01
1.36008143e-01 -1.29090559e+00 -3.87604415e-01 3.51193249e-01
-2.91578025e-01 -8.40877295e-01 -5.44583440e-01 1.86276555e-01
-5.86456895e-01 -7.64036477e-01 -6.68097198e-01 -7.59825647e-01
6.73781097e-01 2.22842470e-01 1.60196674e+00 -4.93299738e-02
1.59898296e-01 1.29457593e-01 -4.03353602e-01 -8.38938475e-01
-1.10560346e+00 3.26374501e-01 -4.10099067e-02 -2.67150313e-01
4.45643991e-01 -4.42552984e-01 -3.89979005e-01 -5.99296987e-01
-1.15178061e+00 -3.09309177e-02 7.92694092e-01 8.35455894e-01
4.28765178e-01 -1.27482966e-01 1.34518921e+00 -1.26539981e+00
1.55826604e+00 -4.99227166e-01 4.56250757e-02 1.64117321e-01
-2.56804436e-01 9.13164020e-02 8.72437119e-01 -3.35999042e-01
-1.10362840e+00 -4.44042653e-01 3.84527706e-02 -4.51129265e-02
-4.20368053e-02 8.71287584e-01 8.24553147e-02 1.05133069e+00
1.08670068e+00 9.53380316e-02 -5.06874211e-02 -3.07057083e-01
5.78708887e-01 9.60044742e-01 6.80325031e-01 -2.92715609e-01
5.62665105e-01 1.42192364e-01 -7.43769109e-01 -7.78565228e-01
-1.51577675e+00 -2.37487435e-01 -2.40656644e-01 1.20048523e-01
5.19043565e-01 -1.09330869e+00 2.76960552e-01 1.16182111e-01
-1.52201498e+00 -1.53834268e-01 -8.50374401e-01 1.08327158e-01
-2.88299054e-01 1.71048462e-01 -5.87361097e-01 -7.22394168e-01
-9.51274991e-01 -7.74105728e-01 1.27088356e+00 2.34991699e-01
-7.87457228e-01 -7.02174008e-01 7.39796907e-02 1.82408258e-01
4.17404532e-01 2.45967194e-01 7.32835293e-01 -1.15427673e+00
-1.50012106e-01 -6.04914248e-01 5.74108325e-02 8.56508851e-01
2.22046912e-01 3.34668249e-01 -8.32731128e-01 -8.91002361e-03
1.16089925e-01 -7.49631822e-01 1.09413636e+00 4.36736584e-01
6.55189216e-01 -8.42036545e-01 -7.72796795e-02 -3.16556424e-01
8.78032207e-01 -2.37300813e-01 4.03725654e-01 6.16683774e-02
3.27805579e-01 7.23145962e-01 4.52727765e-01 3.51197124e-01
3.62647027e-01 -8.20972472e-02 6.21445328e-02 -5.62440883e-03
9.93730687e-03 -2.84574330e-01 4.34154630e-01 1.05098748e+00
5.73532164e-01 -7.67176092e-01 -5.77888072e-01 1.13320196e+00
-1.84749949e+00 -1.07630467e+00 2.34360531e-01 1.89422607e+00
1.19789672e+00 5.47145545e-01 -1.27940804e-01 5.92616573e-02
6.98805690e-01 5.43173850e-01 -7.06355810e-01 -9.16203201e-01
-2.88275391e-01 2.72132039e-01 1.13757662e-01 6.79862916e-01
-1.00394666e+00 8.12927365e-01 6.00890255e+00 4.73642260e-01
-9.82282877e-01 -2.07967490e-01 9.73249257e-01 -6.05806649e-01
-9.43248391e-01 -1.82272166e-01 -6.24593139e-01 2.78489292e-01
1.25150096e+00 -6.89353466e-01 -2.43161414e-02 7.81496823e-01
5.41997671e-01 -2.48488367e-01 -1.17775452e+00 5.24897814e-01
4.43453640e-01 -1.51800883e+00 4.53421652e-01 -2.70508885e-01
1.15461946e+00 1.64886504e-01 -1.95074067e-01 3.27653706e-01
7.26754844e-01 -1.01065850e+00 7.90022492e-01 3.65746528e-01
6.40884340e-01 -7.52032697e-01 1.06400394e+00 5.05036771e-01
-3.06084573e-01 2.20706448e-01 -5.13334036e-01 -1.71444640e-01
3.43424082e-01 9.41080093e-01 -8.62978160e-01 4.30194080e-01
2.13557661e-01 9.32361186e-01 -5.25859118e-01 6.19885504e-01
-5.51431239e-01 7.01293707e-01 -1.85307324e-01 -2.26471812e-01
3.60229492e-01 -1.60545960e-01 6.09951019e-01 1.21950591e+00
2.05459267e-01 -1.77398976e-02 8.85157660e-02 7.48218477e-01
-6.30235851e-01 4.17693965e-02 -7.88531601e-01 -3.73952508e-01
4.28785026e-01 1.26875675e+00 -5.88004827e-01 -7.96910286e-01
-2.81553686e-01 8.74177396e-01 2.31674552e-01 5.48789859e-01
-2.20999882e-01 -6.36433303e-01 3.85009557e-01 2.72467677e-02
2.66820490e-01 1.67490065e-01 -4.83494401e-01 -1.36414933e+00
3.33789885e-01 -1.22841322e+00 -6.25684783e-02 -1.02664137e+00
-1.35374212e+00 8.23241174e-01 -2.44256109e-01 -7.88769484e-01
-8.57806325e-01 -4.49114740e-02 -1.08920681e+00 9.55090702e-01
-1.27254450e+00 -7.67813146e-01 1.34682789e-01 -3.28414857e-01
9.77782488e-01 -1.01569012e-01 6.74630523e-01 -1.69742554e-01
-2.48546019e-01 2.11718485e-01 8.55440274e-02 -5.25065977e-03
8.89030695e-01 -1.49761033e+00 1.10254061e+00 1.08833218e+00
4.07237858e-02 7.24748015e-01 1.24952507e+00 -8.04356873e-01
-7.97181487e-01 -1.20532227e+00 1.32760346e+00 -5.34491897e-01
7.40194380e-01 -4.21652853e-01 -9.06119168e-01 7.35727608e-01
7.31600285e-01 -4.28402692e-01 5.15408099e-01 -1.46487758e-01
-1.26089767e-01 -3.05373333e-02 -9.40035999e-01 8.49755287e-01
6.06255054e-01 -4.66899723e-01 -1.05695200e+00 4.80143160e-01
1.08096647e+00 -2.58668214e-01 -1.99675217e-01 5.83216399e-02
1.55423924e-01 -6.97192132e-01 3.89730304e-01 -8.20529282e-01
1.13378072e+00 -8.72453079e-02 1.33759663e-01 -1.89103341e+00
2.04605281e-01 -8.06186020e-01 -1.63280010e-01 1.36780882e+00
7.79513955e-01 -2.49647468e-01 6.20098472e-01 6.40060484e-01
-3.65006059e-01 -6.66292012e-01 -5.28431356e-01 -2.04518795e-01
1.91565961e-01 -4.42361645e-02 3.97633314e-01 4.82699066e-01
6.77140430e-02 9.76603448e-01 -1.58635125e-01 -3.89302254e-01
3.33536893e-01 1.13575958e-01 9.21227813e-01 -1.02244818e+00
-2.92088855e-02 -4.71286654e-01 1.96823969e-01 -1.08865273e+00
3.08133394e-01 -5.75363576e-01 3.37875903e-01 -2.25613666e+00
8.65180418e-02 1.71510264e-01 1.74689457e-01 2.91100562e-01
-3.78486097e-01 7.05397641e-03 -1.33433357e-01 7.42464811e-02
-8.07619691e-01 6.11225367e-01 1.06513631e+00 -2.25408673e-01
-2.04416126e-01 6.17302433e-02 -1.61794817e+00 5.82610428e-01
8.07379544e-01 -3.74229580e-01 -4.96453583e-01 -5.58577418e-01
4.69020754e-01 -3.76593210e-02 3.31145599e-02 -5.56117773e-01
4.20508236e-02 2.83803307e-02 2.78248191e-01 -7.98152089e-01
3.18108797e-02 -1.05264351e-01 -3.20322871e-01 -2.38140114e-02
-1.06572700e+00 2.65899986e-01 6.66873232e-02 6.86468124e-01
-4.34681505e-01 -4.16987479e-01 3.86513114e-01 -4.90923613e-01
7.41865709e-02 -2.37414360e-01 -5.44054747e-01 4.46828097e-01
3.64993900e-01 1.33828461e-01 -7.63366461e-01 -9.76884186e-01
-4.06257242e-01 2.86868900e-01 5.55123866e-01 3.70829880e-01
5.31027436e-01 -9.60358381e-01 -1.22235775e+00 -6.81178942e-02
1.25701283e-03 5.25764585e-01 -1.64445922e-01 2.39939705e-01
-3.80923927e-01 4.14087862e-01 3.30775261e-01 -6.84944317e-02
-8.47140849e-01 2.98035562e-01 -5.73947430e-02 -4.43425626e-01
-6.45730793e-01 6.67406142e-01 7.40064010e-02 -3.49533945e-01
2.40216330e-01 -6.61233842e-01 -2.17591122e-01 3.23864371e-01
8.99399161e-01 4.13617864e-02 3.22672039e-01 -4.04641181e-01
4.46557403e-02 -1.63760617e-01 -5.29837251e-01 -4.87797022e-01
1.49403298e+00 -2.30116934e-01 -1.53710455e-01 6.05938137e-01
9.93887067e-01 2.75545478e-01 -1.09632432e+00 -2.33035162e-01
7.33243227e-02 8.12245756e-02 -1.33637354e-01 -9.37418878e-01
-5.32700717e-01 7.21262753e-01 -4.23849195e-01 4.98913974e-01
7.28005946e-01 2.62864511e-02 9.73013520e-01 7.51456320e-01
-2.98152626e-01 -1.16942155e+00 1.71241790e-01 5.58364093e-01
1.22970712e+00 -1.32992256e+00 2.78340042e-01 1.00084037e-01
-8.54998052e-01 8.64680529e-01 4.18698221e-01 -4.98615831e-01
1.14232808e-01 2.58917660e-01 3.28263164e-01 -1.90858379e-01
-1.40798187e+00 2.03105569e-01 1.90190464e-01 3.76980811e-01
5.54712355e-01 -6.35269359e-02 -4.65805501e-01 8.59458089e-01
-5.94998479e-01 -2.16621235e-01 1.23678195e+00 8.99249911e-01
-7.70515561e-01 -8.63679290e-01 -6.14482500e-02 1.02428925e+00
-7.65588820e-01 -3.41491073e-01 -9.16748285e-01 3.10151130e-01
-5.79926968e-01 1.14901912e+00 9.76562873e-02 6.98572397e-02
3.05810392e-01 1.18567325e-01 -5.22682406e-02 -1.20825243e+00
-7.98141241e-01 1.46215752e-01 5.84402680e-01 6.24476112e-02
-3.01890939e-01 -5.54904819e-01 -1.11137593e+00 -2.09879473e-01
-1.54760212e-01 5.78711867e-01 7.47905493e-01 1.03882813e+00
5.32337487e-01 5.73947251e-01 4.80218679e-01 -9.58150446e-01
-9.98030961e-01 -1.26115954e+00 -2.57646024e-01 4.82201874e-01
8.09035778e-01 1.43890560e-01 -6.82958007e-01 3.20377558e-01] | [12.492395401000977, 9.373167991638184] |
e607d219-dfbe-43cb-8ea7-a00796a58634 | towards-language-models-that-can-see-computer | 2306.16410 | null | https://arxiv.org/abs/2306.16410v1 | https://arxiv.org/pdf/2306.16410v1.pdf | Towards Language Models That Can See: Computer Vision Through the LENS of Natural Language | We propose LENS, a modular approach for tackling computer vision problems by leveraging the power of large language models (LLMs). Our system uses a language model to reason over outputs from a set of independent and highly descriptive vision modules that provide exhaustive information about an image. We evaluate the approach on pure computer vision settings such as zero- and few-shot object recognition, as well as on vision and language problems. LENS can be applied to any off-the-shelf LLM and we find that the LLMs with LENS perform highly competitively with much bigger and much more sophisticated systems, without any multimodal training whatsoever. We open-source our code at https://github.com/ContextualAI/lens and provide an interactive demo. | ['Amanpreet Singh', 'Douwe Kiela', 'Tristan Thrush', 'Gautam Mittal', 'William Berrios'] | 2023-06-28 | null | null | null | null | ['object-recognition'] | ['computer-vision'] | [-9.86425057e-02 -4.59248051e-02 -2.33255699e-02 -4.54880387e-01
-9.44120169e-01 -8.57439935e-01 9.19515014e-01 -9.56704170e-02
-6.26285374e-01 3.62339453e-03 2.90337265e-01 -2.30602771e-01
2.05457553e-01 -3.47214311e-01 -7.57829487e-01 -4.23626840e-01
2.50470072e-01 4.00197446e-01 2.43965566e-01 -3.94651759e-03
1.57114431e-01 5.24311364e-02 -1.70178139e+00 5.09131849e-01
3.12603891e-01 8.45710516e-01 5.51321208e-01 1.25521421e+00
-3.41646671e-02 1.12258315e+00 -8.79365429e-02 -7.35684752e-01
1.23236902e-01 1.02513529e-01 -7.30300248e-01 3.20647895e-01
1.05005276e+00 -2.62680054e-01 -6.11277580e-01 1.00866520e+00
4.58954871e-01 -1.97034732e-01 5.49337268e-01 -1.26131535e+00
-1.19454980e+00 4.16455686e-01 -2.96450019e-01 1.26663640e-01
4.33522403e-01 7.30709136e-01 1.02548671e+00 -1.34693360e+00
4.60265666e-01 1.55454111e+00 6.09641373e-01 6.60881042e-01
-1.21965575e+00 -1.34108678e-01 1.30095199e-01 7.96085894e-02
-1.20887601e+00 -9.20244932e-01 2.19490319e-01 -6.04256868e-01
1.05945587e+00 1.28404126e-01 3.00582409e-01 1.16142809e+00
3.73222562e-03 9.63464618e-01 1.09960556e+00 -4.14299160e-01
1.61459506e-01 2.88346410e-01 4.66135412e-01 1.17491627e+00
8.40333477e-02 -2.66273290e-01 -7.30133772e-01 -9.65520069e-02
4.61489588e-01 1.12782449e-01 -1.68444365e-01 -2.68600821e-01
-1.35114825e+00 7.05095291e-01 3.37387860e-01 1.93838682e-02
-1.97174937e-01 4.46829617e-01 1.50783494e-01 2.86561042e-01
1.92691714e-01 3.39609414e-01 -3.72186840e-01 -7.66536966e-02
-8.07691038e-01 -3.14811133e-02 1.11078167e+00 1.12580919e+00
8.12379539e-01 -3.73919964e-01 1.72492564e-02 6.73968077e-01
5.64440072e-01 7.02438653e-01 3.40194076e-01 -1.50706196e+00
1.20820262e-01 4.65449631e-01 -9.02174786e-02 -5.05068243e-01
-1.80527821e-01 -1.88768268e-01 -5.82415819e-01 1.52887091e-01
3.00063789e-01 -1.05780289e-01 -1.13508844e+00 1.63219452e+00
-1.17363483e-01 4.23791222e-02 2.59352654e-01 6.82906210e-01
1.27216387e+00 5.11388838e-01 1.79070398e-01 2.29013354e-01
1.49485385e+00 -1.14099586e+00 -3.42290938e-01 -8.40320468e-01
3.86601090e-01 -8.30012023e-01 1.29642534e+00 4.04358357e-01
-1.19528103e+00 -4.69112664e-01 -6.46640599e-01 -5.43517709e-01
-5.57030439e-01 3.02684397e-01 7.42507517e-01 1.40503973e-01
-1.61286747e+00 -1.16812102e-01 -9.34645653e-01 -8.13504934e-01
6.10185742e-01 1.31666034e-01 -4.73100692e-01 -3.78281474e-01
-3.82341713e-01 9.28860903e-01 1.58328652e-01 -1.65119335e-01
-1.19091666e+00 -3.67994279e-01 -9.94181037e-01 -3.07342082e-01
7.27025628e-01 -1.08066380e+00 1.68575859e+00 -8.98739874e-01
-1.18481874e+00 1.30468190e+00 -4.45381790e-01 -4.08993959e-01
2.79175758e-01 -4.23754841e-01 -5.37698828e-02 1.86687037e-01
-2.51491778e-02 9.67942536e-01 7.47117281e-01 -1.37218738e+00
-4.36663061e-01 -4.16069657e-01 5.70139706e-01 7.39920959e-02
-3.77717584e-01 1.34379148e-01 -1.14130127e+00 -3.67483869e-02
-2.97591299e-01 -1.03392172e+00 -2.55761534e-01 3.85265231e-01
-3.89954925e-01 -2.90458471e-01 4.31320250e-01 -3.86904180e-01
8.72674227e-01 -2.25889874e+00 3.36209714e-01 -3.31841111e-01
5.07332563e-01 2.07629263e-01 -4.45258766e-01 6.10815763e-01
3.73573005e-01 -1.06186450e-01 -3.50828767e-01 -8.41951191e-01
2.09820524e-01 3.53697360e-01 -3.57681274e-01 3.14652503e-01
1.41616479e-01 1.33582842e+00 -8.22225571e-01 -4.87411499e-01
2.95020610e-01 3.66237491e-01 -3.29076707e-01 3.64225924e-01
-5.67419350e-01 -1.06419615e-01 -1.41569823e-01 7.94513941e-01
1.88230008e-01 -7.72219896e-01 -1.79174483e-01 -2.47126967e-01
-1.77132726e-01 -5.46359606e-02 -8.40993285e-01 2.20675659e+00
-4.60088611e-01 9.18228805e-01 2.73709625e-01 -6.73687518e-01
3.50565583e-01 8.31151009e-02 3.87669168e-02 -3.07382762e-01
-9.28607583e-03 -1.64744377e-01 -1.55793712e-01 -8.00885618e-01
2.73044944e-01 9.46168080e-02 -6.84766248e-02 4.30815339e-01
4.75118846e-01 -1.90791935e-01 2.01468542e-01 6.61095023e-01
1.15047693e+00 2.01510582e-02 1.78578898e-01 -1.52497413e-02
2.63418198e-01 7.04104900e-02 -2.78366730e-02 1.10583901e+00
-5.62438667e-02 6.18568718e-01 1.21498168e-01 -1.44601643e-01
-7.47288465e-01 -1.34332454e+00 5.91223538e-02 1.34260142e+00
-9.52787772e-02 -8.24922800e-01 -6.52356684e-01 -2.37226382e-01
8.56799930e-02 6.12398267e-01 -2.92714208e-01 6.51070550e-02
1.61795780e-01 -5.87242186e-01 6.91232562e-01 5.78007281e-01
4.86375064e-01 -9.02890563e-01 -6.67255640e-01 -3.09075296e-01
9.11539868e-02 -1.43961751e+00 -3.85242909e-01 6.71422854e-02
-4.06807780e-01 -9.84935284e-01 -5.76442361e-01 -8.41244817e-01
5.41087866e-01 4.87973154e-01 1.44972360e+00 7.81787559e-03
-6.36609912e-01 1.39223325e+00 -2.34227479e-02 -6.06002390e-01
-3.21632385e-01 -2.38918379e-01 8.84675905e-02 -1.08605281e-01
5.94363630e-01 -3.61392200e-01 -4.79499608e-01 -8.54065567e-02
-1.01019442e+00 2.33950809e-01 8.98432434e-01 2.67423719e-01
3.95217687e-01 -4.51313376e-01 6.98996857e-02 -7.98964024e-01
5.85844398e-01 -5.21257222e-01 -5.96817911e-01 4.73566175e-01
-4.24250126e-01 1.36643916e-01 1.76705196e-01 -4.62698758e-01
-6.94772303e-01 2.53389746e-01 7.11534470e-02 -5.88809073e-01
-5.11847138e-01 5.80717802e-01 -4.66768965e-02 -9.11888331e-02
5.69099009e-01 4.99468386e-01 7.96230957e-02 -5.41327298e-01
1.13018930e+00 8.28949928e-01 9.90629554e-01 -4.29682553e-01
8.52229834e-01 7.11853981e-01 -1.93669498e-01 -9.62832868e-01
-1.01252246e+00 -6.87442780e-01 -6.66209340e-01 -1.78013638e-01
8.70750308e-01 -1.34741390e+00 -9.55690801e-01 5.77519417e-01
-1.22175336e+00 -4.31879818e-01 6.03352189e-02 3.57479125e-01
-6.89532518e-01 4.76907045e-01 -7.00031817e-01 -8.27330947e-01
-2.55077034e-01 -8.77226055e-01 1.18520105e+00 3.84476691e-01
1.24715213e-02 -9.71789896e-01 1.36343211e-01 4.34296876e-01
2.59809852e-01 -3.80848825e-01 5.30308604e-01 -5.57727098e-01
-8.60848606e-01 -8.20495337e-02 -3.12668175e-01 5.05929708e-01
-9.51198861e-02 8.40302631e-02 -1.25800681e+00 -1.76607341e-01
-7.48229548e-02 -9.25508320e-01 1.37785041e+00 1.45138755e-01
1.01460290e+00 -2.85389721e-01 -2.18009055e-01 6.52181387e-01
1.68312252e+00 -3.61111641e-01 4.50547904e-01 1.23191856e-01
8.39488149e-01 5.59803963e-01 5.77612594e-02 2.03869522e-01
9.69522357e-01 2.58723229e-01 4.14097697e-01 -1.25989094e-01
-2.65616179e-01 -1.75176375e-02 6.93571627e-01 6.98149979e-01
8.27957094e-02 -3.78853619e-01 -1.23494196e+00 5.76685548e-01
-2.06504488e+00 -1.00510204e+00 1.28031075e-01 1.79942310e+00
8.53741109e-01 4.91986200e-02 2.84711272e-02 -7.35975742e-01
3.00985754e-01 1.35672227e-01 -7.20023453e-01 -1.16156101e-01
-2.32954398e-01 -5.42509779e-02 4.07055795e-01 6.50672853e-01
-1.21744430e+00 9.96156514e-01 6.97536516e+00 5.81308842e-01
-9.27766085e-01 1.43222228e-01 3.04346472e-01 -4.29126680e-01
-2.58484483e-01 1.27817243e-01 -7.43736207e-01 1.86036319e-01
1.08123875e+00 -5.41100949e-02 6.46951973e-01 6.84477031e-01
-1.08594395e-01 -3.44747514e-01 -1.52208447e+00 1.43875337e+00
5.24216056e-01 -1.43930387e+00 1.69124037e-01 1.85581148e-02
2.82928407e-01 9.51016963e-01 1.94114730e-01 8.64615738e-02
6.58072889e-01 -1.04635072e+00 7.18592405e-01 1.05550098e+00
6.18481755e-01 -1.31892741e-01 1.96020275e-01 6.60719633e-01
-9.30968285e-01 -7.92836025e-02 -3.59185249e-01 -1.09363208e-02
-2.01425299e-01 2.48949453e-01 -4.11281407e-01 2.89550483e-01
6.52492881e-01 9.17782068e-01 -1.09470558e+00 1.09006250e+00
-9.38064698e-03 4.46807653e-01 -4.02919710e-01 1.74403638e-01
1.92398652e-01 -3.45404632e-02 4.76779044e-01 1.52913904e+00
5.48781678e-02 7.67396837e-02 4.18077350e-01 1.02034867e+00
-1.25490859e-01 -2.87626207e-01 -8.97821128e-01 -3.29375356e-01
3.24311793e-01 1.45670688e+00 -3.24253827e-01 -5.52434444e-01
-1.01275396e+00 1.03918576e+00 4.84011114e-01 5.59005439e-01
-4.12036568e-01 -6.00055456e-02 6.57966733e-01 -2.65004098e-01
3.42859536e-01 -5.48629701e-01 -3.87799516e-02 -1.66788125e+00
4.02244963e-02 -8.24217319e-01 3.47726077e-01 -1.31040108e+00
-1.62641478e+00 5.49025774e-01 -1.39215365e-01 -7.92123199e-01
-3.53133768e-01 -1.01550007e+00 -5.13368905e-01 6.19161963e-01
-1.48198485e+00 -1.52088010e+00 -3.12391877e-01 8.57626557e-01
8.14395189e-01 -3.58012795e-01 9.00886476e-01 -5.82282059e-02
-4.27550972e-01 1.98436931e-01 -9.01778191e-02 3.75725031e-01
8.24019492e-01 -1.35098624e+00 4.67450559e-01 7.85262167e-01
7.92963505e-01 7.62389362e-01 6.62870765e-01 -2.48754859e-01
-2.11045933e+00 -9.56777871e-01 6.82991147e-01 -1.27740359e+00
9.01492298e-01 -6.85610414e-01 -5.81374705e-01 1.04752696e+00
6.80483282e-01 2.98227631e-02 8.01259995e-01 2.32411906e-01
-8.63917768e-01 1.29593655e-01 -6.05144799e-01 6.93785012e-01
1.15040934e+00 -1.10240042e+00 -7.93279529e-01 5.56029618e-01
6.15783691e-01 -1.62779868e-01 -6.17939949e-01 5.58427162e-03
5.72389066e-01 -9.58588243e-01 1.11428583e+00 -5.34247518e-01
4.65980202e-01 -2.12848589e-01 -5.68602264e-01 -8.18820059e-01
-2.35650569e-01 -3.70533586e-01 -3.07601959e-01 1.16858661e+00
3.54607016e-01 -5.88388026e-01 1.99191421e-01 7.36776948e-01
4.84722592e-02 -6.33223712e-01 -5.24645209e-01 -5.99958062e-01
-1.86274141e-01 -7.78976262e-01 -8.81764442e-02 4.48392153e-01
-3.04549962e-01 6.85908973e-01 -9.13013667e-02 3.87585282e-01
9.54251349e-01 2.14842483e-01 8.70354116e-01 -1.14919698e+00
-6.42360926e-01 -4.14827824e-01 -4.19776589e-01 -1.07860088e+00
3.31659138e-01 -1.13901174e+00 8.03484023e-02 -1.77524137e+00
7.47939885e-01 1.66293025e-01 -2.23879576e-01 7.80540645e-01
9.29707363e-02 6.01880193e-01 5.78109145e-01 5.48301697e-01
-1.20155132e+00 2.27478713e-01 7.13986635e-01 -3.29228044e-01
2.44882420e-01 -2.10973844e-01 -1.00213170e+00 1.17190385e+00
4.53135192e-01 -1.29224941e-01 -3.80798668e-01 -9.15166557e-01
2.98777968e-01 -2.32906759e-01 9.48340178e-01 -8.36541831e-01
6.47166610e-01 3.66902277e-02 2.98568070e-01 -3.23615879e-01
6.67343974e-01 -6.43052220e-01 -2.87267417e-01 1.07308917e-01
-5.74119925e-01 -2.94547174e-02 3.28087062e-01 5.30248404e-01
-5.49302436e-02 -1.84297431e-02 4.19165850e-01 -5.86609006e-01
-1.11313176e+00 4.52215582e-01 -4.27892983e-01 -9.24681779e-03
7.18054116e-01 2.20523641e-01 -7.05359757e-01 -5.66903353e-01
-8.30006719e-01 4.48779941e-01 7.92376101e-01 5.57813883e-01
8.11106205e-01 -9.30768371e-01 -6.96212590e-01 5.58986068e-02
6.69135332e-01 -2.19091386e-01 -2.20583647e-01 7.56976366e-01
-1.11271851e-01 5.87473810e-01 9.63944569e-02 -7.54368603e-01
-1.31337428e+00 6.48221374e-01 3.46137941e-01 4.46348935e-01
-7.09409118e-01 8.56715918e-01 4.82129246e-01 -3.34428519e-01
4.83083248e-01 -3.58100981e-01 -3.41816316e-03 -6.35195524e-02
6.82568252e-01 -7.92469010e-02 -3.64815056e-01 -5.91250718e-01
-5.28465211e-01 7.73254931e-01 6.69547021e-02 -4.09924656e-01
1.25596833e+00 -4.92981166e-01 -4.05025750e-01 1.01005757e+00
1.01980317e+00 -2.04391137e-01 -1.14144325e+00 -6.66078925e-01
4.84374650e-02 -2.23709956e-01 2.01642931e-01 -1.00333536e+00
-6.15253925e-01 8.96323502e-01 5.35835326e-01 5.80875315e-02
1.12732208e+00 7.25984633e-01 4.37357128e-01 9.85444486e-01
2.82100409e-01 -7.07829773e-01 1.16811924e-01 5.74905574e-01
8.49811316e-01 -1.64096904e+00 -2.21021250e-01 -6.30155765e-03
-7.42637515e-01 1.06989002e+00 4.87661481e-01 -1.02514058e-01
8.10598075e-01 5.18167615e-01 3.26678425e-01 -3.92345548e-01
-1.48462462e+00 -7.45460629e-01 5.55161178e-01 4.64840978e-01
3.01928222e-01 -5.98336793e-02 2.36954704e-01 4.63677764e-01
1.42849579e-01 2.21859336e-01 4.50802356e-01 8.18049312e-01
-5.62870204e-01 -8.57657373e-01 -2.34946504e-01 2.73992896e-01
-1.74065858e-01 -4.17536885e-01 -7.27709174e-01 3.33172709e-01
5.99612445e-02 1.05649889e+00 1.24057464e-01 -3.17702949e-01
-1.01538107e-01 2.72692531e-01 6.94395185e-01 -8.24479997e-01
-2.21665323e-01 1.07621044e-01 -1.10828150e-02 -8.68135870e-01
-7.30262816e-01 -5.23572326e-01 -1.15754986e+00 -7.13008568e-02
3.23625028e-01 -4.80279148e-01 7.80570567e-01 1.03253508e+00
5.46183467e-01 -4.88844328e-03 2.26346791e-01 -7.67901838e-01
-4.36814427e-01 -8.14406633e-01 -2.83066392e-01 2.83723980e-01
7.97553122e-01 -2.13572726e-01 -3.35845411e-01 5.14924884e-01] | [10.79066276550293, 1.6278600692749023] |
77ed2fb6-a141-4577-81d2-69017f96c124 | ds-at-semeval-2019-task-9-from-suggestion | null | null | https://aclanthology.org/S19-2209 | https://aclanthology.org/S19-2209.pdf | DS at SemEval-2019 Task 9: From Suggestion Mining with neural networks to adversarial cross-domain classification | Suggestion Mining is the task of classifying sentences into suggestions or non-suggestions. SemEval-2019 Task 9 sets the task to mine suggestions from online texts. For each of the two subtasks, the classification has to be applied on a different domain. Subtask A addresses the domain of posts in suggestion online forums and comes with a set of training examples, that is used for supervised training. A combination of LSTM and CNN networks is constructed to create a model which uses BERT word embeddings as input features. For subtask B, the domain of hotel reviews is regarded. In contrast to subtask A, no labeled data for supervised training is provided, so that additional unlabeled data is taken to apply a cross-domain classification. This is done by using adversarial training of the three model parts label classifier, domain classifier and the shared feature representation. For subtask A, the developed model archives a F1-score of 0.7273, which is in the top ten of the leader board. The F1-score for subtask B is 0.8187 and is ranked in the top five of the submissions for that task. | ['Tobias Cabanski'] | 2019-06-01 | null | null | null | semeval-2019-6 | ['suggestion-mining'] | ['natural-language-processing'] | [ 7.64813200e-02 6.13441288e-01 -2.39998132e-01 -6.09553814e-01
-6.26328647e-01 -3.50095779e-01 8.46733034e-01 3.42936456e-01
-5.62722325e-01 7.47526228e-01 3.31407487e-01 -6.04100227e-01
1.05539873e-01 -5.65685332e-01 -4.06668544e-01 -4.43165988e-01
3.25019896e-01 4.05004263e-01 5.90215735e-02 -2.33002275e-01
4.61991966e-01 5.27228750e-02 -1.21258378e+00 6.11196756e-01
6.86300933e-01 1.11220741e+00 1.87449232e-01 4.84876752e-01
-4.93480772e-01 5.55817246e-01 -7.54973054e-01 -4.51508552e-01
2.53431380e-01 -3.76675010e-01 -8.98860514e-01 1.40389428e-01
2.40668207e-01 -2.90642142e-01 8.85980786e-04 7.69510627e-01
3.37075382e-01 4.91832495e-01 8.96637380e-01 -1.00976288e+00
-5.85985363e-01 8.24461699e-01 -7.33049512e-02 2.54518893e-02
2.45673999e-01 -2.93682456e-01 1.11153162e+00 -1.29831386e+00
5.90789557e-01 7.91475356e-01 2.19942227e-01 7.02525198e-01
-1.03274977e+00 -5.98627985e-01 2.37433508e-01 8.17010179e-03
-8.16496372e-01 -1.06073983e-01 7.64037311e-01 -5.80411613e-01
9.13310170e-01 2.95396000e-02 1.84364364e-01 1.16873586e+00
1.36975899e-01 5.41338325e-01 1.09959507e+00 -6.42560959e-01
6.93174958e-01 9.54643309e-01 4.44048554e-01 2.90812571e-02
-1.47285864e-01 -8.20219889e-02 -4.78844076e-01 -1.09790049e-01
2.20912740e-01 9.45540518e-02 -1.60018038e-02 5.04152961e-02
-8.21578741e-01 1.09634006e+00 3.15028399e-01 5.08170128e-01
-3.98131847e-01 -5.94485581e-01 6.66418493e-01 6.79345846e-01
8.24909449e-01 6.62932575e-01 -6.39268041e-01 1.64172277e-01
-7.65736401e-01 2.71824926e-01 1.22847807e+00 7.01295137e-01
8.57593358e-01 -7.31059611e-02 -2.69887567e-01 1.06588280e+00
4.38073054e-02 -1.93033919e-01 1.02307081e+00 -4.86384541e-01
5.05906463e-01 7.92209983e-01 3.35523069e-01 -7.81382382e-01
-3.25691879e-01 -4.32015032e-01 -6.64704561e-01 9.39709246e-02
2.21152335e-01 -5.30796766e-01 -5.79569578e-01 1.32465315e+00
2.33553737e-01 -1.60374001e-01 2.66722709e-01 9.31418002e-01
1.30617762e+00 6.55122817e-01 -6.14826493e-02 -1.13899954e-01
1.18696630e+00 -1.36832094e+00 -5.67512333e-01 -3.30112427e-01
6.78803384e-01 -8.08798254e-01 9.99448597e-01 5.19435942e-01
-5.44872880e-01 -1.09096491e+00 -1.08630443e+00 6.91425279e-02
-7.19059467e-01 5.64387262e-01 1.50469109e-01 2.76703984e-01
-9.46252823e-01 3.00726861e-01 -1.30482867e-01 -3.68292183e-01
-4.92508151e-02 3.43626827e-01 -4.90608126e-01 -6.72845021e-02
-1.34170902e+00 1.01598477e+00 2.14411631e-01 -2.49213025e-01
-5.72955251e-01 -4.26573724e-01 -7.81335771e-01 1.68446630e-01
1.72485977e-01 -1.76865593e-01 1.33247018e+00 -1.24364030e+00
-1.51424932e+00 8.77096295e-01 9.23078135e-02 -5.35485089e-01
4.49540019e-01 -1.50886431e-01 -5.45103431e-01 -2.43450329e-01
4.08502996e-01 6.31874621e-01 1.07550001e+00 -1.03792942e+00
-7.89932847e-01 -1.68702170e-01 2.29776144e-01 1.62732422e-01
-5.88452458e-01 9.83531624e-02 -1.67148128e-01 -4.29789484e-01
-1.78587958e-01 -7.63152957e-01 -2.38622263e-01 -7.00048447e-01
-6.05108976e-01 -6.96473718e-01 7.06749141e-01 -6.26106083e-01
9.39850390e-01 -2.18281412e+00 -3.14280570e-01 1.54639825e-01
1.52926818e-01 1.52444601e-01 -1.19818620e-01 5.65380335e-01
-2.77666003e-01 3.87676395e-02 1.72905684e-01 -4.79909867e-01
1.23434618e-01 -1.08003259e-01 -4.77261066e-01 2.08852082e-01
3.05652499e-01 3.73938501e-01 -6.71952724e-01 1.34452522e-01
1.31742023e-02 -1.36766031e-01 -3.07355165e-01 5.70760608e-01
-2.07824081e-01 4.13502991e-01 -5.32367229e-01 4.31900434e-02
6.62442863e-01 -8.15977752e-02 1.06258892e-01 1.80787683e-01
-2.41031542e-01 6.87691033e-01 -1.00568771e+00 1.42451417e+00
-7.14456618e-01 4.94833797e-01 -9.96219143e-02 -1.34669900e+00
1.60018969e+00 2.81260073e-01 3.61497432e-01 -5.71721375e-01
4.14395034e-02 2.35420927e-01 9.35035050e-02 -4.80143160e-01
6.31118655e-01 -8.13143328e-02 -2.16766879e-01 9.36682820e-01
2.59982318e-01 -1.06229752e-01 9.74134952e-02 2.76077598e-01
1.18831873e+00 -4.12512571e-02 1.15182526e-01 -2.11400390e-01
8.40442359e-01 5.80409467e-02 3.16874892e-01 5.88843584e-01
2.05907952e-02 2.85176665e-01 4.66374516e-01 -6.07140064e-01
-9.64786947e-01 -5.15677512e-01 -1.59195606e-02 1.35057652e+00
-1.88620806e-01 -4.06358331e-01 -6.16596043e-01 -1.12794030e+00
1.97137822e-03 1.01462865e+00 -7.73514986e-01 -1.61797062e-01
-1.20496027e-01 -7.91962743e-02 5.08212531e-03 2.25276500e-01
3.30792874e-01 -1.32451057e+00 -4.83087450e-01 2.57566601e-01
8.92727450e-02 -9.51239944e-01 -3.33574146e-01 6.79393709e-01
-6.67187572e-01 -1.18699968e+00 -7.75116920e-01 -1.01226616e+00
8.13983500e-01 1.00358099e-01 1.06899631e+00 -1.76940978e-01
7.80231133e-02 2.33621523e-01 -7.22886443e-01 -7.19218910e-01
-5.01741469e-01 1.54095769e-01 1.73301756e-01 6.46083727e-02
8.70847285e-01 -9.03484076e-02 -4.19818103e-01 3.63743901e-01
-5.47022700e-01 -1.36053383e-01 6.24337733e-01 1.14631665e+00
3.30215216e-01 -1.39581233e-01 1.17782331e+00 -1.11137486e+00
1.16503572e+00 -4.88169789e-01 -3.22255075e-01 -7.84384236e-02
-5.77936947e-01 -3.25738311e-01 1.15023589e+00 -3.35226566e-01
-1.03392959e+00 5.76518364e-02 -1.78464472e-01 -7.13436678e-02
-6.64107382e-01 6.78101480e-01 1.23042569e-01 4.85753447e-01
8.97790492e-01 1.25208020e-01 1.14061199e-01 -5.44880331e-01
2.23571509e-01 1.19534147e+00 2.15459421e-01 -1.96600303e-01
4.77610588e-01 -1.51814073e-01 -6.07715845e-01 -7.73580432e-01
-1.08537805e+00 -7.18497634e-01 -5.82614362e-01 -2.25598693e-01
5.87070107e-01 -6.39266849e-01 -8.73920396e-02 -7.09754080e-02
-1.04399657e+00 -3.42676014e-01 -4.86824363e-01 5.76915622e-01
-4.21407729e-01 4.57702689e-02 -2.37572655e-01 -8.63348007e-01
-5.20854652e-01 -1.04564679e+00 5.31858385e-01 6.26648143e-02
-4.99618888e-01 -1.03737938e+00 1.04646273e-01 4.95997429e-01
3.25755924e-01 -1.75948679e-01 8.51492703e-01 -1.89450777e+00
2.79774845e-01 -6.86385870e-01 -1.73724703e-02 9.27715063e-01
2.72177011e-01 -4.15872037e-01 -1.19162273e+00 -1.66674510e-01
5.34833789e-01 -7.30995059e-01 8.89824450e-01 2.22231299e-01
1.26335049e+00 -4.51906115e-01 -1.27086654e-01 -2.86815166e-01
8.84626508e-01 2.55450249e-01 1.95855245e-01 4.59161103e-01
-3.22233438e-02 1.03401780e+00 1.10180020e+00 4.54918176e-01
2.29689196e-01 5.76593637e-01 2.61886686e-01 -5.03923232e-03
2.94920534e-01 -2.18886584e-01 5.56119204e-01 7.28522480e-01
5.98821044e-01 -7.28267953e-02 -7.68397272e-01 5.65824509e-01
-1.74115431e+00 -6.99688137e-01 -5.32924421e-02 2.17720723e+00
5.32961965e-01 3.16095293e-01 2.02430040e-01 1.43289119e-01
7.01008439e-01 8.32940266e-02 -5.07839799e-01 -1.10535395e+00
3.68336260e-01 1.74979582e-01 9.73330885e-02 2.62406588e-01
-1.35650551e+00 8.04785550e-01 5.48774910e+00 5.77027082e-01
-1.14286745e+00 7.61996806e-02 8.19962323e-01 3.13327104e-01
-3.39789718e-01 -8.37917253e-02 -7.79531062e-01 5.84464073e-01
1.25013494e+00 -8.83840248e-02 7.23683916e-04 1.19451511e+00
4.68917131e-01 -2.35113263e-01 -1.10140526e+00 5.30253708e-01
3.28471810e-01 -1.28848732e+00 -1.73682019e-01 -2.70749684e-02
7.45919645e-01 8.14209133e-02 7.76356831e-02 7.77759373e-01
3.76900882e-01 -1.02715552e+00 2.79345006e-01 3.21664602e-01
5.59051037e-01 -6.99644566e-01 1.32819021e+00 9.15299535e-01
-4.24511790e-01 -2.99136877e-01 -6.76415086e-01 -3.37270886e-01
-1.08021460e-01 7.30797410e-01 -1.32595432e+00 5.71164727e-01
5.57481945e-01 8.41640055e-01 -2.26225913e-01 7.92733192e-01
-3.86600137e-01 7.06893504e-01 2.68965196e-02 -4.29528654e-01
3.09557438e-01 -4.26223487e-01 1.44832641e-01 1.16861224e+00
1.99708521e-01 -2.99249947e-01 5.08904397e-01 7.04035938e-01
-3.43587905e-01 5.23336291e-01 -8.13868046e-01 6.19493835e-02
4.09447551e-01 1.43455648e+00 -2.97195882e-01 -3.86970311e-01
-5.02603292e-01 8.72841179e-01 3.55515242e-01 2.67392546e-01
-3.74366313e-01 -3.58119249e-01 8.11334923e-02 -1.48271844e-02
2.86648989e-01 2.13815421e-01 -4.04240549e-01 -7.35737920e-01
-1.53478250e-01 -7.70541489e-01 1.27274096e-01 -3.58930051e-01
-1.53773558e+00 7.39044189e-01 -2.52520740e-01 -1.52329326e+00
-4.64369357e-01 -7.06437409e-01 -1.13966691e+00 1.15190852e+00
-1.32949460e+00 -9.79033172e-01 -2.68177092e-01 3.10686022e-01
9.98764217e-01 -8.60722363e-01 1.05649710e+00 7.03988671e-02
-4.47265416e-01 4.68682647e-01 2.52277195e-01 2.10220367e-02
1.06560087e+00 -1.47121859e+00 1.72821984e-01 4.59905893e-01
5.24289720e-02 6.39418900e-01 5.99635482e-01 -4.41904813e-01
-7.23578572e-01 -1.21054268e+00 1.42315853e+00 -1.96931601e-01
7.33305037e-01 -2.77076066e-01 -7.25080848e-01 3.88319135e-01
2.66730189e-01 -2.12956294e-01 1.32027757e+00 1.57651931e-01
-4.68332656e-02 -5.77700976e-03 -1.06570375e+00 1.12765633e-01
1.52233914e-01 -5.53434849e-01 -8.84063303e-01 6.72240973e-01
6.51253581e-01 -1.82150692e-01 -9.07418787e-01 -2.08647937e-01
2.46826366e-01 -8.22068512e-01 4.83188510e-01 -7.93817461e-01
8.88976872e-01 1.17959097e-01 -8.53709877e-03 -1.53763127e+00
-8.68627355e-02 -1.66027069e-01 1.88599885e-01 1.32543230e+00
6.59740210e-01 -5.80010593e-01 7.31008649e-01 6.86746955e-01
-3.48238468e-01 -8.63363147e-01 -7.54854441e-01 -6.24861360e-01
2.64054179e-01 -4.04664457e-01 3.55631113e-01 8.82882535e-01
2.01592997e-01 7.22512126e-01 -1.54322386e-01 -6.08596265e-01
-4.42070477e-02 2.36679435e-01 1.11708891e+00 -1.21375871e+00
-2.77674329e-02 -1.99142426e-01 1.62736773e-01 -1.09944093e+00
5.52093387e-01 -9.85921204e-01 1.26951844e-01 -1.64413178e+00
-1.84230044e-01 -4.45331812e-01 -4.93598610e-01 5.59455276e-01
1.21570326e-01 -6.41893968e-02 -1.69564225e-02 8.18037838e-02
-3.17526788e-01 6.67038620e-01 9.33817327e-01 -3.57709706e-01
-3.30441654e-01 6.33681417e-01 -9.23830509e-01 6.14082813e-01
1.05144930e+00 -4.23729271e-01 -4.68369186e-01 5.87948933e-02
-3.68608423e-02 -1.76373601e-01 4.50012982e-02 -7.02621460e-01
2.01702848e-01 -1.29590486e-03 2.97660917e-01 -6.48087859e-01
3.48500580e-01 -8.13597083e-01 -5.68907201e-01 2.66064018e-01
-9.13828552e-01 -1.55756176e-01 3.23476046e-02 6.07306719e-01
-4.51416075e-01 -7.36146092e-01 2.76553035e-01 -1.30835250e-01
-4.97001648e-01 -7.65691772e-02 -6.58769786e-01 -2.29145363e-01
8.09201717e-01 -2.59566247e-01 -1.55199692e-01 -5.10866463e-01
-8.84994090e-01 1.22148484e-01 1.27642840e-01 6.20742738e-01
7.99293637e-01 -1.01326728e+00 -7.83877254e-01 1.70354396e-01
3.27071428e-01 1.34739876e-02 1.82528540e-01 6.61307633e-01
5.77922314e-02 5.36536098e-01 -4.82223965e-02 -1.60700679e-01
-9.09497380e-01 2.90888458e-01 2.38424190e-03 -4.56135094e-01
-5.32919705e-01 8.95714641e-01 4.31249626e-02 -8.41519892e-01
3.24827164e-01 -9.76936817e-02 -9.71787691e-01 2.79634804e-01
5.86757064e-01 1.54646024e-01 3.02540272e-01 -4.44912791e-01
-1.49382278e-01 -1.27273768e-01 -4.30849910e-01 -1.13507673e-01
1.52242446e+00 1.30161822e-01 -4.75934818e-02 5.01064777e-01
1.07193327e+00 1.95534304e-02 -8.40613127e-01 -1.98683485e-01
1.33196250e-01 -1.22824743e-01 1.33899655e-02 -9.94525850e-01
-5.97085238e-01 9.27689552e-01 4.18595701e-01 6.20754182e-01
6.98135793e-01 -1.26658693e-01 5.18087447e-01 4.98752743e-01
5.91392927e-02 -1.57633317e+00 2.79521346e-01 8.90849292e-01
1.25303423e+00 -1.50112629e+00 -2.03113735e-01 1.33548295e-02
-9.73894835e-01 1.32180226e+00 9.99309719e-01 -2.41920963e-01
7.10222423e-01 -1.85439914e-01 1.60141170e-01 -1.57241493e-01
-7.96589375e-01 3.14868361e-01 5.04691243e-01 5.78941882e-01
6.87014461e-01 8.69584009e-02 -7.77330160e-01 1.25354052e+00
-3.32104415e-01 -6.77714869e-02 5.31310558e-01 5.13467431e-01
-5.53931475e-01 -1.21584094e+00 -4.88840900e-02 9.46511805e-01
-3.29267204e-01 1.34572357e-01 -8.44669580e-01 4.37014878e-01
7.04347193e-02 1.28429735e+00 1.90063596e-01 -6.76097393e-01
1.96072340e-01 2.57319570e-01 -5.58754444e-01 -1.06561673e+00
-1.08487749e+00 -1.88887849e-01 3.97162586e-01 -9.73863825e-02
-2.09490523e-01 -3.18987787e-01 -1.03048396e+00 7.82068297e-02
-4.87401724e-01 7.04870045e-01 8.87084484e-01 1.12479067e+00
2.38020256e-01 5.02030790e-01 9.37671006e-01 -7.56631970e-01
-9.51554239e-01 -1.43596959e+00 -5.89980543e-01 4.52020705e-01
1.53704062e-01 -5.83508015e-01 -4.13306862e-01 -2.82614261e-01] | [10.895408630371094, 7.5550007820129395] |
abdfa1b6-69f2-46bb-bdd3-ff61480c4395 | minkloc-lidar-and-monocular-image-fusion-for | 2104.05327 | null | https://arxiv.org/abs/2104.05327v2 | https://arxiv.org/pdf/2104.05327v2.pdf | MinkLoc++: Lidar and Monocular Image Fusion for Place Recognition | We introduce a discriminative multimodal descriptor based on a pair of sensor readings: a point cloud from a LiDAR and an image from an RGB camera. Our descriptor, named MinkLoc++, can be used for place recognition, re-localization and loop closure purposes in robotics or autonomous vehicles applications. We use late fusion approach, where each modality is processed separately and fused in the final part of the processing pipeline. The proposed method achieves state-of-the-art performance on standard place recognition benchmarks. We also identify dominating modality problem when training a multimodal descriptor. The problem manifests itself when the network focuses on a modality with a larger overfit to the training data. This drives the loss down during the training but leads to suboptimal performance on the evaluation set. In this work we describe how to detect and mitigate such risk when using a deep metric learning approach to train a multimodal neural network. Our code is publicly available on the project website: https://github.com/jac99/MinkLocMultimodal. | ['Tomasz Trzcinski', 'Monika Wysoczanska', 'Jacek Komorowski'] | 2021-04-12 | null | null | null | null | ['visual-place-recognition'] | ['computer-vision'] | [ 8.89987424e-02 -1.16062969e-01 3.20798680e-02 -5.75229287e-01
-1.19644058e+00 -7.39213705e-01 9.87218857e-01 4.50255096e-01
-7.93503881e-01 3.67522717e-01 -1.29792213e-01 3.40559594e-02
-6.21265545e-03 -5.06661892e-01 -9.09909487e-01 -7.07747996e-01
2.00975135e-01 5.31639338e-01 2.74856180e-01 -6.64872006e-02
4.16936457e-01 7.72363424e-01 -1.74070990e+00 7.25090876e-02
3.84505898e-01 1.22105491e+00 1.91406161e-01 4.78488982e-01
6.87159086e-03 5.74088812e-01 -1.58321396e-01 -2.63193786e-01
4.60326374e-01 -5.85057512e-02 -6.24803007e-01 -1.48524687e-01
8.86382043e-01 -1.15038924e-01 -4.20857430e-01 1.09776878e+00
5.40822804e-01 2.06667945e-01 4.24683899e-01 -1.55898273e+00
-3.12871993e-01 1.96749359e-01 -4.46413368e-01 -1.12613261e-01
5.40378034e-01 -9.54478979e-02 9.06866312e-01 -1.43172038e+00
7.17182577e-01 1.19248092e+00 7.01713622e-01 3.07647556e-01
-1.12408793e+00 -3.79561186e-01 -2.00515494e-01 3.88929963e-01
-1.68753421e+00 -6.75533891e-01 8.35172176e-01 -3.19263816e-01
7.74497867e-01 6.87821433e-02 3.44523609e-01 9.76011455e-01
4.92616221e-02 6.86699688e-01 9.30596173e-01 -2.61371166e-01
2.75828302e-01 6.22912217e-03 4.95238826e-02 8.19445133e-01
1.02015384e-01 2.67882466e-01 -8.02795291e-01 -9.73851606e-02
2.93652356e-01 2.66872942e-01 9.30571705e-02 -7.21131682e-01
-1.39113390e+00 6.98635578e-01 7.13853121e-01 5.02002873e-02
-2.88333923e-01 3.62873852e-01 2.14005500e-01 3.47830325e-01
2.30693743e-02 2.23040879e-01 -1.70769885e-01 -1.66483968e-01
-9.19095099e-01 2.31629983e-01 5.42961359e-01 9.63263392e-01
1.32021952e+00 -5.21719635e-01 6.44621700e-02 8.59760165e-01
6.75001502e-01 6.58802509e-01 2.27807015e-01 -1.09830225e+00
3.61810356e-01 6.49703920e-01 2.50190292e-02 -7.31849313e-01
-5.30911744e-01 1.46116734e-01 -3.31153095e-01 4.27624315e-01
4.28039253e-01 2.36920938e-02 -1.04518402e+00 1.51933205e+00
3.54592204e-01 1.53737098e-01 8.14620852e-02 1.01209497e+00
8.32443595e-01 3.80766541e-01 -8.12055618e-02 5.20689905e-01
1.16616261e+00 -8.62073958e-01 -3.80840212e-01 -2.95158207e-01
5.51493764e-01 -1.07079840e+00 6.28006697e-01 2.31661931e-01
-6.99861765e-01 -4.34295177e-01 -1.28486252e+00 -4.08243507e-01
-8.20493340e-01 2.11137101e-01 2.78013796e-01 2.92178661e-01
-1.26972938e+00 6.38957798e-01 -1.05399406e+00 -8.87436807e-01
2.32818499e-01 5.38339853e-01 -8.86632919e-01 -3.71053636e-01
-7.44285643e-01 1.16639078e+00 2.91571081e-01 1.74344555e-01
-8.62076581e-01 -2.94056118e-01 -1.23666108e+00 -5.08167088e-01
-8.60403180e-02 -2.89890945e-01 1.12937379e+00 -6.51306748e-01
-1.09356189e+00 1.01044881e+00 -2.50600785e-01 -2.74307013e-01
5.39744079e-01 -2.30222642e-01 -2.26554111e-01 -4.47528437e-02
2.49785513e-01 1.08681583e+00 7.91346371e-01 -1.38827085e+00
-8.16200376e-01 -6.24097407e-01 -4.51298133e-02 2.17028290e-01
1.80518270e-01 -1.43645540e-01 -4.96122032e-01 -1.27718315e-01
4.15283859e-01 -1.27239406e+00 -4.72244881e-02 1.42730251e-01
-2.91869998e-01 -1.82958618e-01 9.96048033e-01 -2.96174407e-01
5.26750505e-01 -2.28845263e+00 1.13228425e-01 4.27037150e-01
2.82470584e-02 -1.29313260e-01 -3.44154388e-01 6.27467453e-01
7.33731538e-02 -2.52198428e-01 -3.40243071e-01 -9.31496501e-01
2.56551832e-01 3.45903158e-01 -6.03586547e-02 1.01059163e+00
4.29402024e-01 6.91667497e-01 -8.74671757e-01 -2.94582784e-01
5.36895514e-01 7.50485599e-01 -2.20214605e-01 1.20117061e-01
1.99244484e-01 3.80883187e-01 -9.25452635e-03 1.09646654e+00
8.83570552e-01 2.32117027e-01 -2.92107254e-01 -3.48071873e-01
-3.33065212e-01 5.07788770e-02 -1.14308369e+00 2.32431722e+00
-2.60715187e-01 8.46888363e-01 -1.00426078e-01 -7.34502196e-01
1.04566145e+00 -1.47062773e-02 4.98672158e-01 -7.88821578e-01
2.00678468e-01 6.13604248e-01 -2.58554041e-01 -3.65240216e-01
8.54169905e-01 7.17708468e-02 -1.74897850e-01 5.93380369e-02
4.10574734e-01 -8.04314837e-02 1.97217628e-01 1.53631456e-02
1.12026310e+00 3.08639735e-01 6.02130480e-02 -1.76017508e-01
5.30722857e-01 1.51149690e-01 2.15980351e-01 6.14474356e-01
-3.32995415e-01 8.81799340e-01 2.38430262e-01 -2.30141312e-01
-1.01874471e+00 -1.03691828e+00 -3.97805601e-01 1.11256909e+00
5.18853247e-01 -3.14668506e-01 -2.32738003e-01 -5.59724092e-01
3.43620449e-01 4.52195615e-01 -6.70231521e-01 -2.85883933e-01
-3.23685795e-01 -3.09749097e-01 7.61014819e-01 5.38085222e-01
2.16921106e-01 -7.56083786e-01 -6.29154384e-01 -8.53630528e-02
1.06961928e-01 -1.22291934e+00 -5.83310835e-02 5.03910780e-01
-6.04876637e-01 -1.11845636e+00 -6.38725758e-01 -7.04976559e-01
6.72596395e-01 2.97833204e-01 6.74980223e-01 -1.87182069e-01
6.67293072e-02 7.45969474e-01 -3.47570747e-01 -3.32403958e-01
-1.60340339e-01 1.12308852e-01 5.34754023e-02 2.33226284e-01
6.90653801e-01 -4.05283511e-01 -5.72921634e-01 2.08088085e-01
-8.88098180e-01 -4.56925511e-01 4.56989527e-01 6.78746998e-01
7.49292791e-01 -6.80845022e-01 -6.60458356e-02 -2.73194999e-01
3.09441596e-01 -5.29395282e-01 -8.62603724e-01 7.18967095e-02
-5.14188468e-01 2.92686850e-01 1.36562318e-01 -8.65271017e-02
-4.39561903e-01 6.88894331e-01 -1.36099830e-01 -7.42280662e-01
-3.52091849e-01 5.28008342e-01 -7.76909739e-02 -5.43609202e-01
5.82066357e-01 -2.29803752e-02 1.40783697e-01 -5.64026713e-01
3.69501829e-01 6.05318427e-01 6.47959232e-01 -4.13817257e-01
6.62164211e-01 7.17891693e-01 2.95956880e-01 -6.91081166e-01
-3.89829576e-01 -8.83565247e-01 -8.92617762e-01 -3.20128560e-01
6.09241605e-01 -1.02140808e+00 -4.76793051e-01 5.06171703e-01
-1.31063771e+00 -9.23695564e-02 -1.70272425e-01 5.77979445e-01
-5.26381433e-01 1.88940898e-01 -1.31384328e-01 -4.98321503e-01
8.62963572e-02 -1.28714001e+00 1.64871430e+00 3.50529790e-01
1.41474560e-01 -8.57042551e-01 5.36181450e-01 1.86078385e-01
1.44563779e-01 3.59584838e-01 1.37412816e-01 -7.42578626e-01
-6.07668221e-01 -4.60034102e-01 -3.01193893e-01 2.91461766e-01
-7.07139075e-02 4.32818942e-02 -1.39971006e+00 -2.47268274e-01
-5.07522166e-01 -3.39431405e-01 1.10810494e+00 -1.82886183e-01
5.60595214e-01 3.81379962e-01 -2.65633881e-01 6.86252236e-01
1.55151486e+00 -9.21887085e-02 5.13076842e-01 5.32465994e-01
9.32506800e-01 5.42505026e-01 7.18107939e-01 3.58636439e-01
7.28051722e-01 6.79807544e-01 7.55584300e-01 -7.22415047e-03
-5.68925776e-02 -2.21576691e-01 5.14972448e-01 4.84416187e-01
1.27758667e-01 7.31284320e-02 -1.44973433e+00 6.52701318e-01
-2.11216784e+00 -5.42994916e-01 -1.18413880e-01 2.43684840e+00
2.69555718e-01 -2.52557755e-01 -8.15954879e-02 3.40058804e-02
6.70857728e-01 1.28286807e-02 -3.73189867e-01 -3.39087963e-01
-2.32967764e-01 -2.33440828e-02 9.13484693e-01 7.14684308e-01
-1.45000374e+00 9.23444808e-01 5.27463484e+00 3.02994877e-01
-1.47858489e+00 2.31323570e-01 -4.29428220e-02 3.59031907e-03
-1.33864675e-02 1.09941922e-01 -6.61659181e-01 2.24168524e-01
9.47697520e-01 2.12054312e-01 3.77126157e-01 7.52713740e-01
-1.55273184e-01 -4.18976307e-01 -1.54542744e+00 1.40382898e+00
3.68395507e-01 -1.01317191e+00 -2.45632365e-01 1.28095806e-01
4.51405823e-01 8.93087447e-01 2.75988244e-02 1.02371894e-01
-1.47302613e-01 -9.18122530e-01 9.61063683e-01 6.97234690e-01
6.01562679e-01 -7.01656044e-01 8.43218505e-01 -6.20108447e-04
-1.16781557e+00 9.41435918e-02 -3.66295397e-01 1.88963950e-01
7.28661492e-02 3.13289523e-01 -8.79208088e-01 5.33268809e-01
6.02477968e-01 9.58623588e-01 -1.03049207e+00 1.38481748e+00
-3.51431780e-02 -6.02885149e-02 -6.97766125e-01 2.52835274e-01
3.79426569e-01 -2.87194806e-03 6.53933585e-01 1.17793429e+00
5.95888793e-01 -5.36016107e-01 2.21995741e-01 6.21082664e-01
-5.54389507e-02 -1.30075186e-01 -1.05997872e+00 9.55196992e-02
5.05654395e-01 1.34089899e+00 -5.49128115e-01 -3.37617449e-03
-4.98421699e-01 1.05966449e+00 3.98233145e-01 2.19177023e-01
-7.00910091e-01 -4.47229117e-01 6.13368154e-01 -2.10227832e-01
3.00427109e-01 -4.07096088e-01 -2.33468994e-01 -1.08490074e+00
8.07366446e-02 -3.28126758e-01 2.42224202e-01 -8.02205384e-01
-1.10481894e+00 4.50567365e-01 3.61867808e-02 -1.35958624e+00
-2.73468405e-01 -8.98304760e-01 -3.25763971e-01 7.53224373e-01
-1.66718614e+00 -1.26740134e+00 -6.68273985e-01 8.28256130e-01
8.33281130e-02 -1.01862751e-01 7.60213315e-01 6.56421542e-01
-3.18849117e-01 5.46871960e-01 8.61917511e-02 1.08572021e-01
1.02140200e+00 -1.33953357e+00 1.31461740e-01 7.91422307e-01
2.20575944e-01 3.65941167e-01 5.92117488e-01 -1.71920240e-01
-1.66466427e+00 -1.05879366e+00 1.02229571e+00 -7.25151956e-01
6.83511674e-01 -3.26341361e-01 -6.04847014e-01 5.89173377e-01
2.10556582e-01 4.05829459e-01 5.85363507e-01 -1.26931325e-01
-5.15321851e-01 -2.70732492e-01 -1.15584934e+00 1.05566800e-01
7.66693652e-01 -8.85928929e-01 -4.54401910e-01 1.15357757e-01
3.00555497e-01 -5.43622792e-01 -8.23302567e-01 4.63631064e-01
6.60598397e-01 -9.20136809e-01 7.09441066e-01 -5.97443245e-02
2.17915162e-01 -6.85094059e-01 -6.33283556e-01 -1.10301435e+00
4.11356464e-02 -2.48003319e-01 -2.19081696e-02 1.08087778e+00
3.74641955e-01 -5.21334291e-01 4.45544451e-01 4.27478611e-01
-1.16630971e-01 -3.60730618e-01 -1.28948045e+00 -6.15742087e-01
-1.09203510e-01 -5.09519756e-01 4.17340010e-01 8.11085820e-01
-1.89678580e-01 1.88833386e-01 -7.74408355e-02 4.43997532e-01
7.64546096e-01 -1.12666078e-01 7.63755679e-01 -1.23774385e+00
3.66415352e-01 -2.57883817e-01 -1.08876872e+00 -7.60215163e-01
1.08603008e-01 -1.11988378e+00 4.24597591e-01 -1.41236985e+00
-1.08194478e-01 -3.40872526e-01 -5.57968676e-01 6.88827753e-01
4.27285343e-01 6.14438415e-01 5.54900050e-01 3.14495385e-01
-9.25813615e-01 5.94327211e-01 6.78569138e-01 -2.59513378e-01
-6.69203291e-05 -3.12946230e-01 -2.64059961e-01 4.96167332e-01
9.08432245e-01 -5.34073472e-01 -2.12300997e-02 -5.72854161e-01
1.56127498e-01 -3.57692659e-01 8.79550755e-01 -1.27342021e+00
6.27642155e-01 2.56274343e-01 2.11137176e-01 -7.37033963e-01
7.20304191e-01 -1.18322861e+00 1.11989476e-01 2.38403901e-01
-1.04394957e-01 5.12747049e-01 2.23495409e-01 4.86604840e-01
-4.16237235e-01 -1.68567404e-01 7.27308810e-01 2.50695407e-01
-1.11694467e+00 2.03059763e-01 -9.78726670e-02 -4.75486785e-01
9.08188641e-01 -2.18590945e-01 -4.04870540e-01 -1.15888990e-01
-7.12311506e-01 2.57824898e-01 8.23387682e-01 7.52620280e-01
7.70435214e-01 -1.69552350e+00 -4.67704564e-01 2.93440372e-01
7.28033900e-01 -3.54547165e-02 -1.34195551e-01 1.17998993e+00
-7.97278762e-01 3.33100975e-01 -2.89969891e-01 -9.41878617e-01
-1.11552966e+00 2.23464310e-01 5.80741823e-01 3.08623761e-01
-3.26173842e-01 7.37999141e-01 -3.61207873e-01 -7.75502503e-01
4.74548936e-01 -2.97163755e-01 -2.61400118e-02 3.55345964e-01
2.64655799e-01 3.99682373e-01 4.08689737e-01 -1.23873305e+00
-8.77342105e-01 6.26666486e-01 9.01573598e-02 -3.93883079e-01
1.16637230e+00 -2.53226310e-01 -2.78597564e-01 7.11104274e-01
1.68067777e+00 -2.16051221e-01 -1.24896944e+00 -2.41025627e-01
1.69049770e-01 -3.75169367e-01 3.24841112e-01 -6.18631363e-01
-9.41120207e-01 7.04009116e-01 1.21846628e+00 -9.01139751e-02
7.92261899e-01 2.12880895e-01 4.82513636e-01 7.04459727e-01
4.86594170e-01 -1.11130941e+00 -2.54383445e-01 7.86563993e-01
9.94649470e-01 -1.65184820e+00 -1.13485821e-01 1.27569050e-01
-5.50189495e-01 1.10297096e+00 2.84899324e-01 -3.53757948e-01
7.90108263e-01 1.01829596e-01 3.55176121e-01 -2.48848706e-01
-3.52773666e-01 -5.55389762e-01 3.69391441e-01 5.22209704e-01
1.97566077e-01 2.96377037e-02 1.84974689e-02 6.84111044e-02
-1.29303798e-01 -1.39148563e-01 3.10568988e-01 1.27467465e+00
-3.54862988e-01 -1.23248434e+00 -5.95245600e-01 6.21410385e-02
-1.03841126e-01 -1.83067489e-02 -6.98039830e-01 6.55419588e-01
4.06288952e-01 8.14356625e-01 1.60075352e-01 -6.15819156e-01
4.11982268e-01 6.94999397e-02 5.91123521e-01 -3.06870788e-01
-3.20712626e-01 -6.87480345e-02 -8.48773941e-02 -9.93926227e-01
-7.71915734e-01 -1.03053689e+00 -1.24444783e+00 -3.23652416e-01
-6.15968667e-02 -1.73302278e-01 1.11071336e+00 8.26907098e-01
4.63938892e-01 4.30509821e-02 4.88653004e-01 -1.36963999e+00
-2.68121004e-01 -8.00204575e-01 -4.19345111e-01 3.62845093e-01
7.77463496e-01 -1.00167596e+00 -2.32061580e-01 -9.70609561e-02] | [7.60085391998291, -1.9898031949996948] |
b75f98a9-7b11-4e00-8fff-069e576ba1ce | scene-text-image-super-resolution-via | null | null | https://dl.acm.org/doi/10.1145/3474085.3475469 | https://dl.acm.org/doi/10.1145/3474085.3475469 | Scene Text Image Super-Resolution via Parallelly Contextual Attention Network | Optical degradation makes text shapes and edges blurred, so the existing scene text recognition methods are difficult to achieve desirable results on low-resolution (LR) scene text images acquired in natural scenes. Therefore, efficiently extracting the sequential information for reconstructing super-resolution (SR) text images is challenging. In this paper, we propose a Parallel Contextual Attention Network (PCAN), which can effectively learn sequence-dependent features and focus more on high-frequency information of the reconstruction in text images. First, we explore the importance of sequence-dependent features in horizontal and vertical directions \emph{parallelly} for text SR, and design a parallel contextual attention block to adaptively select the key information in the text sequence that contributes to image reconstruction. Secondly, we propose a Hierarchically orthogonal texture-aware attention module and an edge guidance loss function, which can help to reconstruct high-frequency information in text images. Finally, we conduct extensive experiments on different test sets of TextZoom, and the results can easily incorporate into mainstream text algorithms to further improve their performance in LR image recognition. Compared with the SR images obtained by BICUBIC up-sampling, our method can respectively improve the recognition accuracy of ASTER, MORAN, and CRNN by 14.29%, 14.35%, and 20.60%. Besides, our PCAN outperforms 8 state-of-the-art (SOTA) SR methods in improving the recognition performance of LR images. Most importantly, it outperforms the current optimal text-orient SR method TSRN by 3.19%, 3.65%, and 6.0% on the recognition accuracy of ASTER, MORAN, and CRNN respectively. | ['Shen Heng Tao', 'Shen Fumin', 'Wu Jun', 'Ding Zhijun', 'Zhao Brian Nlong', 'Feng Shuyang', 'Zhao Cairong'] | 2021-10-17 | null | null | null | acm-international-conference-on-multimedia-2 | ['scene-text-recognition'] | ['computer-vision'] | [ 6.94352031e-01 -6.65468097e-01 1.84875187e-02 -2.94137504e-02
-6.72237098e-01 -4.82459813e-02 3.87495518e-01 -8.49744558e-01
-1.43187135e-01 5.79644501e-01 4.51953530e-01 2.82422174e-02
-2.82888919e-01 -6.17160380e-01 -6.11997604e-01 -1.12370145e+00
6.43232703e-01 3.89049798e-02 2.27093726e-01 -2.41600797e-01
5.65077007e-01 5.42056739e-01 -1.64972639e+00 5.04564285e-01
1.13775074e+00 1.08266008e+00 7.67117679e-01 6.06459737e-01
-3.08010399e-01 8.70306313e-01 -3.91083211e-01 8.68645385e-02
-1.29661593e-03 -3.52136552e-01 -5.56933701e-01 3.84771287e-01
6.90449595e-01 -4.90499020e-01 -7.12337017e-01 1.03489745e+00
5.92179894e-01 2.15194613e-01 6.43431008e-01 -3.92016023e-01
-1.37560940e+00 6.23739898e-01 -1.22269309e+00 4.52189416e-01
1.00902133e-01 4.76185866e-02 7.46148229e-01 -1.29580271e+00
4.93724674e-01 1.25696158e+00 5.51326096e-01 4.97130841e-01
-9.71473396e-01 -6.79804027e-01 3.67776155e-02 6.09302402e-01
-1.40245116e+00 -5.73069155e-01 8.55483174e-01 -3.66055593e-02
6.24696434e-01 5.19835532e-01 1.98352769e-01 1.18648589e+00
1.39972642e-01 1.06242394e+00 1.24695361e+00 -3.02730501e-01
-3.75797540e-01 -2.23059312e-01 1.29247099e-01 4.51565742e-01
1.49886519e-01 -1.73027813e-02 -8.26155365e-01 5.77396691e-01
1.15166163e+00 3.95281017e-02 -8.60946119e-01 2.79770374e-01
-1.33835626e+00 2.73029238e-01 3.40941340e-01 5.06613731e-01
-1.38682798e-01 -2.39956602e-01 1.46881625e-01 -3.99193466e-02
4.57402408e-01 6.61639124e-02 -2.79952288e-01 1.49261458e-02
-7.41778195e-01 -2.07392350e-01 8.39735344e-02 8.57415438e-01
3.53078097e-01 6.11830831e-01 -4.66958970e-01 1.40521908e+00
4.19578701e-03 8.22772563e-01 7.31831193e-01 -5.59898376e-01
5.57449639e-01 3.89085829e-01 1.59256652e-01 -1.04242003e+00
-1.06216490e-01 -4.71705973e-01 -1.28017449e+00 -2.62507081e-01
-7.95231294e-03 3.25960100e-01 -9.92015719e-01 9.47383463e-01
8.22867230e-02 1.90059736e-01 1.32683277e-01 1.32596672e+00
1.05097830e+00 8.70928705e-01 -4.47497904e-01 -4.54454094e-01
1.23194659e+00 -1.08445179e+00 -9.66077089e-01 -2.29392782e-01
-7.77663141e-02 -1.15134382e+00 1.38046503e+00 5.53003967e-01
-9.15046394e-01 -8.96195531e-01 -9.85949159e-01 -3.40146780e-01
1.68047860e-01 8.04309189e-01 3.08683336e-01 3.45367193e-01
-7.55635798e-01 4.51940745e-01 -5.88614047e-01 -2.14544043e-01
4.92301255e-01 1.15637951e-01 -1.48101449e-01 -5.65710783e-01
-8.66169095e-01 5.69528580e-01 1.85948744e-01 4.65950370e-01
-6.78409755e-01 -4.96630549e-01 -5.08444726e-01 2.25256011e-02
5.05472541e-01 -2.94594467e-01 7.37402499e-01 -7.41223752e-01
-1.76452100e+00 5.70516586e-01 -4.07887816e-01 -5.92757389e-02
4.39935207e-01 -3.46964836e-01 -6.38982952e-01 3.66831571e-01
5.71972802e-02 2.12444037e-01 1.16930664e+00 -1.29030406e+00
-6.58818424e-01 -5.17512083e-01 -5.17299950e-01 7.18984067e-01
-3.82973969e-01 1.58410504e-01 -6.63250744e-01 -1.04182458e+00
4.04949427e-01 -2.88497061e-01 2.23737538e-01 6.37104223e-03
-3.80766928e-01 -3.90423052e-02 1.35881901e+00 -1.11654758e+00
1.03404844e+00 -2.19363809e+00 4.07679200e-01 -5.19251585e-01
7.09556788e-02 2.82648206e-01 -1.69875517e-01 -5.00962101e-02
4.45740484e-02 -1.34270061e-02 -1.85594037e-01 -1.35463580e-01
-3.88627708e-01 -9.27589089e-02 -4.46833253e-01 4.91178542e-01
3.70214484e-03 9.16984379e-01 -2.97918558e-01 -3.75503391e-01
5.05783975e-01 6.32356167e-01 6.02670908e-02 6.96431249e-02
-7.44787902e-02 5.55755019e-01 -6.07620597e-01 7.46338069e-01
9.24964786e-01 -3.93720329e-01 -1.85720503e-01 -6.09385490e-01
-2.99423993e-01 -4.59145419e-02 -1.07690632e+00 1.38847792e+00
-4.84194726e-01 8.15964103e-01 9.99781191e-02 -1.11254096e+00
1.16776502e+00 -1.15240209e-01 3.12247902e-01 -1.11415493e+00
1.00941926e-01 4.65615094e-03 -4.10385847e-01 -7.30232954e-01
7.76110113e-01 1.17368542e-01 4.20036644e-01 4.09267366e-01
-4.41674322e-01 2.91952670e-01 -1.52006239e-01 -1.02789193e-01
4.97573793e-01 1.85207620e-01 -1.94168538e-01 -2.49678537e-01
1.02900863e+00 -2.84278065e-01 4.29225057e-01 6.57387435e-01
2.89330423e-01 9.25421298e-01 -1.64684504e-02 -4.21805263e-01
-1.23973429e+00 -6.83857203e-01 -5.05476654e-01 9.16925967e-01
4.96487170e-01 9.58007723e-02 -4.72133547e-01 -2.89581805e-01
-3.67160648e-01 5.56909978e-01 -4.52520967e-01 -2.10038815e-02
-6.92905903e-01 -9.65840876e-01 3.79518211e-01 5.59110761e-01
1.28841269e+00 -1.04438746e+00 -1.05795234e-01 -4.06744033e-02
-6.15365684e-01 -1.37740219e+00 -7.28539586e-01 -3.53932142e-01
-9.15925622e-01 -8.33825231e-01 -8.84790242e-01 -9.32207286e-01
6.47560239e-01 9.26565886e-01 5.11037171e-01 8.33395571e-02
-3.51371527e-01 9.42662507e-02 -5.12296617e-01 3.10870081e-01
1.69443652e-01 -1.95314974e-01 -1.11203462e-01 4.59725469e-01
1.86654896e-01 -2.69074261e-01 -4.18141901e-01 7.35399127e-01
-9.20728981e-01 5.48046350e-01 8.02083194e-01 1.24582291e+00
7.29301929e-01 5.26068747e-01 2.80670315e-01 -5.46595037e-01
2.94162333e-01 7.17082769e-02 -6.16810620e-01 2.74737149e-01
-5.51285088e-01 -5.58914058e-02 9.44505572e-01 -5.23050368e-01
-1.62315655e+00 -2.09241211e-01 8.73700343e-03 -6.46884382e-01
-8.79214257e-02 3.29921812e-01 -1.72991335e-01 -3.48683387e-01
4.25177544e-01 1.19303477e+00 -2.17727408e-01 -7.12731242e-01
2.00660285e-02 1.12222075e+00 5.87129653e-01 -4.28400636e-01
8.46291125e-01 6.54292583e-01 -1.38220578e-01 -1.28400898e+00
-9.92992818e-01 -3.72848570e-01 -3.18048775e-01 -7.56602883e-02
7.37776875e-01 -9.45956886e-01 -6.83659554e-01 9.80389237e-01
-8.36602449e-01 -2.96672583e-01 2.80626547e-02 4.17566955e-01
-4.25352663e-01 9.25497413e-01 -9.38965499e-01 -8.01034749e-01
-6.38209045e-01 -1.08910465e+00 1.47035754e+00 4.03300613e-01
7.77936161e-01 -5.52166581e-01 -7.85431981e-01 8.66833150e-01
4.77604002e-01 -3.69744062e-01 6.53170705e-01 7.75165781e-02
-1.05905700e+00 2.62741655e-01 -1.01525986e+00 2.96853662e-01
8.03158507e-02 -2.38618553e-01 -8.53439331e-01 -3.54604542e-01
1.92404404e-01 -1.54634506e-01 1.05281818e+00 5.93276858e-01
1.60041535e+00 -3.78159285e-01 -1.29844978e-01 9.36952055e-01
1.45240223e+00 1.81889728e-01 9.57784235e-01 3.14605802e-01
1.06569910e+00 3.89561564e-01 8.93126726e-01 3.03958833e-01
1.78508356e-01 8.62825930e-01 7.87808895e-02 1.18201887e-02
-4.97031569e-01 -1.37343481e-01 5.27744710e-01 8.54586065e-01
-2.08968088e-01 -2.16767415e-01 -5.48282444e-01 3.81562799e-01
-1.69144785e+00 -9.72761512e-01 -4.62407559e-01 1.99134791e+00
6.92576528e-01 -2.05108330e-01 -4.69446868e-01 1.15488447e-01
1.01633906e+00 3.66972536e-01 -6.88071728e-01 2.43859231e-01
-6.95348203e-01 1.46218225e-01 5.78560650e-01 3.06832194e-01
-9.45444107e-01 1.17430544e+00 5.13898277e+00 1.54042399e+00
-1.24883330e+00 -1.65496394e-01 7.25117743e-01 1.05874278e-01
-2.50575572e-01 -3.63130987e-01 -1.08375549e+00 5.66803336e-01
3.94872159e-01 -2.02003181e-01 9.43959475e-01 5.22658646e-01
4.64319348e-01 -1.45490477e-02 -4.47808713e-01 1.34500647e+00
5.22325754e-01 -1.29608369e+00 3.11644614e-01 -1.29438937e-01
8.98309588e-01 -1.83881655e-01 3.47766042e-01 -4.15052362e-02
3.93025242e-02 -1.35723662e+00 3.79522353e-01 6.29808366e-01
1.34422553e+00 -7.29109168e-01 7.08612978e-01 3.41535151e-01
-1.38040316e+00 -1.66438252e-01 -7.84241617e-01 3.59924555e-01
-7.27080479e-02 6.31948292e-01 -2.99052656e-01 1.01600850e+00
1.08841729e+00 1.31168306e+00 -4.77051765e-01 6.54972136e-01
-2.15824261e-01 5.13366401e-01 -1.19029023e-01 -6.25734702e-02
4.18717675e-02 -4.17471617e-01 6.73963010e-01 8.29747796e-01
3.87435496e-01 3.93275976e-01 -5.87531254e-02 8.92637730e-01
-6.19947948e-02 9.68356431e-02 -2.51449525e-01 7.27733299e-02
1.86108232e-01 1.32104206e+00 -4.95873988e-01 -1.94846213e-01
-3.04976821e-01 1.23595679e+00 -9.45325717e-02 5.80243349e-01
-7.38186836e-01 -4.65856045e-01 2.37709545e-02 -1.24859281e-01
6.68907344e-01 -1.20628402e-01 -5.42425096e-01 -1.53864169e+00
3.31451386e-01 -1.11727583e+00 2.11062104e-01 -1.52891707e+00
-1.22820687e+00 5.66636503e-01 -6.03158414e-01 -1.31784332e+00
6.03699386e-01 -6.36348248e-01 -3.27737719e-01 1.12548053e+00
-1.69736254e+00 -1.45697331e+00 -7.20184028e-01 6.68179929e-01
1.20903111e+00 -2.54838467e-01 8.60778391e-02 2.46835262e-01
-7.18639970e-01 4.46421444e-01 5.54885030e-01 1.63140316e-02
5.58832407e-01 -7.91081846e-01 1.64788440e-01 1.05020106e+00
-1.47167981e-01 3.70406568e-01 4.49765742e-01 -6.64464355e-01
-1.92172074e+00 -1.17906284e+00 3.01849604e-01 -2.40370810e-01
5.03208995e-01 -1.07593358e-01 -1.21092474e+00 5.20371437e-01
-5.76353930e-02 -1.57902643e-01 -1.44565001e-01 -2.26150721e-01
-2.53461629e-01 -3.00784647e-01 -1.01377165e+00 8.02458465e-01
1.23021686e+00 -4.17975217e-01 -4.79603231e-01 2.66959339e-01
8.64584148e-01 -5.83753049e-01 -8.25973332e-01 4.95510727e-01
5.46032310e-01 -8.38575661e-01 1.09802341e+00 -1.41620999e-02
7.18520343e-01 -5.31797230e-01 -2.77863830e-01 -7.96982825e-01
-4.41625983e-01 -2.81651765e-01 1.06906869e-01 1.13161099e+00
-6.43032938e-02 -6.60273015e-01 7.55009651e-01 -1.37682930e-01
-2.73990959e-01 -5.51081359e-01 -8.50090921e-01 -5.05721033e-01
-9.57785025e-02 -1.29656106e-01 3.80977303e-01 1.05874383e+00
-4.53860104e-01 4.88475353e-01 -7.85561621e-01 3.16685230e-01
9.15134311e-01 4.97511297e-01 5.21679342e-01 -9.97972190e-01
-9.35920253e-02 -5.20348132e-01 5.89115210e-02 -1.65599227e+00
3.50825377e-02 -5.74777365e-01 -1.08274771e-02 -1.43447852e+00
4.72560018e-01 -3.33108008e-01 -4.85784048e-03 1.65994599e-01
-3.41404647e-01 3.07677865e-01 -1.56349435e-01 4.15869504e-01
-4.96155232e-01 9.67379212e-01 1.85811174e+00 -2.85776466e-01
1.10960409e-01 -3.08034062e-01 -6.28617406e-01 3.26499254e-01
6.13782644e-01 1.40628070e-01 -1.19869992e-01 -6.41126812e-01
-1.08803876e-01 3.57825249e-01 2.85806537e-01 -7.34044790e-01
3.22105855e-01 -2.80680329e-01 8.58422697e-01 -1.04182255e+00
1.82901025e-01 -6.82897925e-01 1.07128322e-01 5.74182793e-02
-1.71550721e-01 -4.21448380e-01 2.12398142e-01 6.68634653e-01
-1.42173082e-01 -1.33040026e-01 1.03634250e+00 2.69108415e-02
-8.00501406e-01 4.57854152e-01 -1.99039623e-01 -2.45617211e-01
5.91921389e-01 -5.07213950e-01 -7.00141788e-01 -2.16993362e-01
-2.67477095e-01 2.15573117e-01 5.43187678e-01 5.56748927e-01
1.10521066e+00 -1.13645518e+00 -9.45662081e-01 4.72619623e-01
-9.74752009e-02 2.09471315e-01 9.43464279e-01 9.62041676e-01
-4.90323067e-01 4.93891895e-01 -2.93204457e-01 -7.57616758e-01
-1.39256680e+00 4.94509786e-01 3.79983515e-01 -1.31981924e-01
-1.05224156e+00 5.90817630e-01 3.86783093e-01 -1.42941117e-01
6.39844313e-02 3.14712450e-02 -5.83300233e-01 -3.95874709e-01
6.95997238e-01 5.70910096e-01 -1.62373148e-02 -7.62049615e-01
1.24610595e-01 1.36127567e+00 -4.35722888e-01 1.95250720e-01
1.38210547e+00 -5.40208697e-01 -2.22383499e-01 2.00755045e-01
7.15969145e-01 1.17666073e-01 -1.27092755e+00 -4.97895807e-01
-4.92818624e-01 -8.07553828e-01 3.99094880e-01 -8.55432689e-01
-1.26817489e+00 9.27736163e-01 5.07365823e-01 -4.23991829e-02
1.57489157e+00 -3.23297173e-01 8.39436173e-01 2.04131454e-01
2.21260518e-01 -9.12976623e-01 3.17702532e-01 4.57953721e-01
1.06114733e+00 -9.88961995e-01 2.10566178e-01 -4.82829869e-01
-7.40516961e-01 1.34397221e+00 8.05807114e-01 6.40472621e-02
-1.31667078e-01 1.15203813e-01 -2.30101913e-01 -9.51838586e-03
-6.46940410e-01 -1.24293178e-01 3.65543664e-01 4.42495227e-01
2.62242138e-01 -2.41478816e-01 -2.39279464e-01 4.12002027e-01
4.71566878e-02 -3.01382989e-02 7.18636155e-01 3.76479924e-01
-6.03946984e-01 -6.80901766e-01 -6.35828376e-01 7.14391351e-01
-5.54526627e-01 -4.49200451e-01 -5.09692132e-02 3.97622138e-01
-4.25589740e-01 7.79216528e-01 -1.42897572e-03 -3.22424531e-01
2.00060010e-01 -4.01335239e-01 3.64298910e-01 -2.21492499e-01
8.13271180e-02 4.66265470e-01 -1.43601252e-02 -2.29723141e-01
-3.70598346e-01 -6.72872484e-01 -9.72693563e-01 -4.65947807e-01
-6.39664769e-01 -1.55421227e-01 3.73919040e-01 7.87946165e-01
3.12953115e-01 6.89887404e-01 8.28131974e-01 -6.83067024e-01
-3.71291876e-01 -1.07464135e+00 -7.45354772e-01 1.43922806e-01
2.87829012e-01 -5.42280555e-01 -6.44740582e-01 2.04794541e-01] | [11.357799530029297, -1.8216440677642822] |
0a11b164-1c7b-419a-88c5-482825ec8085 | opinion-spam-detection-a-new-approach-using | 2205.13422 | null | https://arxiv.org/abs/2205.13422v1 | https://arxiv.org/pdf/2205.13422v1.pdf | Opinion Spam Detection: A New Approach Using Machine Learning and Network-Based Algorithms | E-commerce is the fastest-growing segment of the economy. Online reviews play a crucial role in helping consumers evaluate and compare products and services. As a result, fake reviews (opinion spam) are becoming more prevalent and negatively impacting customers and service providers. There are many reasons why it is hard to identify opinion spammers automatically, including the absence of reliable labeled data. This limitation precludes an off-the-shelf application of a machine learning pipeline. We propose a new method for classifying reviewers as spammers or benign, combining machine learning with a message-passing algorithm that capitalizes on the users' graph structure to compensate for the possible scarcity of labeled data. We devise a new way of sampling the labels for the training step (active learning), replacing the typical uniform sampling. Experiments on three large real-world datasets from Yelp.com show that our method outperforms state-of-the-art active learning approaches and also machine learning methods that use a much larger set of labeled data for training. | ['Dan Vilenchik', 'Michael Segal', 'Kiril Danilchenko'] | 2022-05-26 | null | null | null | null | ['spam-detection'] | ['natural-language-processing'] | [ 5.05552441e-02 3.36148053e-01 -6.39400780e-01 -5.70254922e-01
-6.16064727e-01 -6.40727520e-01 7.23021150e-01 6.25191808e-01
-3.93154800e-01 5.04739165e-01 -1.08825058e-01 -5.93429029e-01
2.65396923e-01 -9.14830744e-01 -4.91911292e-01 -3.87197971e-01
1.85152486e-01 8.74383926e-01 6.88568056e-01 -2.93417603e-01
5.98503947e-01 1.35953397e-01 -1.46981382e+00 4.14158046e-01
1.06762397e+00 8.69593084e-01 -3.24465394e-01 1.12569302e-01
-5.26209354e-01 7.21343398e-01 -5.40114045e-01 -1.09619868e+00
1.93059579e-01 -4.98427004e-01 -6.42175555e-01 4.99421060e-01
2.88178474e-01 -8.11060220e-02 1.65241852e-01 1.17520344e+00
5.30688353e-02 -3.46006006e-02 3.94212335e-01 -1.50504768e+00
-6.85655057e-01 4.86553371e-01 -6.49249911e-01 -8.61143693e-02
2.04128057e-01 -8.97884741e-02 1.42106140e+00 -8.75381768e-01
7.00563610e-01 1.20755112e+00 5.36501527e-01 4.35801089e-01
-1.01746571e+00 -4.57070678e-01 5.69820285e-01 3.48403126e-01
-8.58518600e-01 -4.54032838e-01 7.99998581e-01 -2.98973769e-01
5.74175894e-01 1.78102165e-01 8.29488039e-01 1.00307727e+00
-1.70100793e-01 1.13175678e+00 9.92321670e-01 -6.25639856e-01
8.07243168e-01 8.28180015e-01 4.63429332e-01 5.26231170e-01
6.93646610e-01 -2.97722518e-01 -5.35579383e-01 -7.24091887e-01
1.25547126e-01 1.41685545e-01 8.18939507e-03 -1.03002441e+00
-5.67591190e-01 1.32057834e+00 2.75364369e-01 1.74910501e-01
-3.45914900e-01 -2.53728598e-01 4.48532671e-01 5.66947281e-01
1.04482830e+00 5.22778809e-01 -6.18115127e-01 -3.98821272e-02
-7.29457676e-01 1.02130547e-01 1.27716231e+00 6.16408288e-01
7.56976724e-01 -4.12935406e-01 3.51647407e-01 7.82841802e-01
6.89947605e-01 2.12156177e-01 3.94859880e-01 -5.14340878e-01
1.92886248e-01 1.23966002e+00 4.64672744e-01 -9.66865361e-01
1.49238884e-01 -4.42291111e-01 -2.33464062e-01 2.47770980e-01
6.98924243e-01 1.12975247e-01 -6.06288552e-01 9.47912991e-01
3.21943760e-01 -5.30482735e-03 -5.17967224e-01 1.03781688e+00
4.48869139e-01 3.01710188e-01 4.78071421e-02 -2.93892354e-01
1.14044571e+00 -1.32885003e+00 -6.88787937e-01 -4.41336066e-01
7.80660272e-01 -7.96802461e-01 1.11823201e+00 7.92472005e-01
-7.33282685e-01 -4.86842431e-02 -1.01663172e+00 1.76241457e-01
-9.27954614e-01 -1.73349872e-01 1.20670593e+00 1.03219569e+00
-6.02424324e-01 7.00189710e-01 -7.12616980e-01 -3.96775484e-01
6.29177988e-01 3.81261051e-01 -2.13804662e-01 -9.62519273e-02
-1.07354414e+00 1.00083828e+00 -2.20432490e-01 -2.64101308e-02
-2.62901247e-01 -2.15153456e-01 -5.60030699e-01 -2.40379088e-02
6.51743591e-01 -1.83814615e-01 1.56071365e+00 -1.68563795e+00
-1.28634417e+00 8.65777671e-01 -2.35240534e-01 -4.53546643e-01
5.80809891e-01 -2.09273294e-01 -4.96515095e-01 -2.95998454e-02
-8.14365745e-02 2.92130262e-02 1.07306933e+00 -1.47520351e+00
-6.58711016e-01 -5.67088962e-01 -3.63119543e-02 -5.13264611e-02
-5.51818430e-01 7.89921731e-02 -2.49756038e-01 -1.76686794e-01
2.67188758e-01 -8.32855105e-01 -3.62251222e-01 9.76542234e-02
-2.32375026e-01 -4.81564134e-01 9.95856225e-01 -2.92317808e-01
1.08983839e+00 -1.79700434e+00 -2.76591688e-01 4.79528189e-01
4.68242317e-01 7.25081205e-01 -6.78429827e-02 3.69556636e-01
2.16913640e-01 3.33340108e-01 -5.00809290e-02 -5.68767071e-01
1.28226005e-03 1.06073484e-01 -1.82996258e-01 6.12694919e-01
2.19710976e-01 9.36092377e-01 -1.50371826e+00 -3.01032275e-01
4.75631244e-02 2.14375079e-01 -1.43038079e-01 3.25502045e-02
-2.67713010e-01 1.54564694e-01 -4.25106913e-01 7.91913748e-01
5.74510276e-01 -5.63565731e-01 3.36312771e-01 2.73433864e-01
7.16410950e-02 8.64378333e-01 -1.01537907e+00 1.11572349e+00
-3.92707229e-01 4.23689604e-01 1.84433758e-01 -1.22556615e+00
9.87353027e-01 1.28801391e-01 2.38432288e-01 -7.78864264e-01
2.86912084e-01 7.13355184e-01 5.41058965e-02 -3.44197124e-01
4.20134544e-01 9.48237479e-02 2.23279983e-01 7.95342624e-01
-7.56669417e-02 3.39776799e-02 4.95402575e-01 5.21692574e-01
1.17263782e+00 -1.66039437e-01 3.83306682e-01 -9.65336412e-02
5.44220448e-01 2.31808007e-01 4.53437179e-01 7.40574360e-01
-5.96352220e-01 4.39418256e-01 7.29353726e-01 -5.40662706e-01
-9.74711835e-01 -6.70387924e-01 3.29843611e-01 1.17464578e+00
8.05732459e-02 -4.34847623e-01 -6.28442824e-01 -1.41481864e+00
3.14667255e-01 7.28235066e-01 -5.74589372e-01 -1.95501551e-01
-3.11696589e-01 -8.78142178e-01 -2.04584494e-01 -2.44192779e-02
7.56601691e-02 -8.71669948e-01 1.20391892e-02 1.74173996e-01
6.15562759e-02 -8.01837921e-01 -1.27957433e-01 9.71979275e-02
-1.01479924e+00 -1.36615622e+00 -4.29526210e-01 -6.37888968e-01
1.00949216e+00 6.55050218e-01 1.50183439e+00 3.39043140e-01
-4.52171490e-02 5.07538691e-02 -7.82211065e-01 -5.01969635e-01
-6.48011446e-01 4.02854711e-01 -1.98852077e-01 2.46876761e-01
1.05031192e+00 -2.57725209e-01 -7.11672843e-01 5.35498381e-01
-5.76595485e-01 -3.47732395e-01 6.94326818e-01 7.42907107e-01
7.11326748e-02 -1.01053521e-01 1.06525600e+00 -1.86913419e+00
8.01911235e-01 -6.44611239e-01 -7.44423270e-01 2.47305602e-01
-1.12203217e+00 -2.51575232e-01 5.07969201e-01 -5.87862015e-01
-8.88712406e-01 1.73123270e-01 8.90263394e-02 2.68182367e-01
1.48394376e-01 2.48347372e-01 7.35115185e-02 -8.79999027e-02
7.72359371e-01 -3.83672953e-01 2.21663028e-01 -4.45061415e-01
4.58100259e-01 1.11130857e+00 -2.91473746e-01 2.18725801e-01
8.93700838e-01 5.99036932e-01 -5.33335745e-01 -4.82703179e-01
-1.08846653e+00 -1.04938662e+00 -5.27254581e-01 -1.24968000e-01
-3.02796364e-02 -5.64848840e-01 -5.04145682e-01 3.34452152e-01
-1.02873278e+00 8.54599774e-02 -3.57498944e-01 3.75969499e-01
-5.67975901e-02 3.63567322e-01 -6.24735653e-01 -1.06108809e+00
-3.94022256e-01 -8.86795104e-01 7.40718961e-01 8.66318569e-02
-4.44265753e-01 -1.00323498e+00 1.43497750e-01 7.50534117e-01
6.49725378e-01 -5.21188796e-01 6.06309533e-01 -1.23954856e+00
-6.48565888e-01 -8.28048944e-01 -2.58884609e-01 5.12442589e-01
1.13869108e-01 7.08926022e-02 -9.08854306e-01 -2.14400604e-01
-1.20807746e-02 -3.36823612e-01 9.01668966e-01 -1.04441158e-01
5.82115054e-01 -3.05051088e-01 -4.28763688e-01 -3.49763721e-01
1.20186675e+00 2.89414749e-02 6.48076534e-01 3.39458913e-01
3.71585131e-01 9.02738392e-01 7.46997833e-01 1.32419467e-01
3.07719141e-01 3.41203958e-01 5.54052353e-01 -1.20740086e-01
2.12116152e-01 -3.23887855e-01 2.90507436e-01 8.70382726e-01
4.64366972e-01 -1.94223702e-01 -7.19929874e-01 6.71320796e-01
-2.09334946e+00 -7.31087208e-01 -4.84912485e-01 2.52656245e+00
7.04648376e-01 5.86812139e-01 2.43065730e-01 4.51339662e-01
8.67989063e-01 -1.04411103e-01 -5.33328831e-01 -4.84621435e-01
1.81536511e-01 1.64331928e-01 6.73460126e-01 4.70676363e-01
-1.15429616e+00 9.65207160e-01 5.77205038e+00 6.61099911e-01
-9.30337608e-01 3.80965978e-01 8.02328348e-01 2.05265298e-01
-3.00356895e-01 3.43358606e-01 -7.64376283e-01 4.49255377e-01
9.20872808e-01 8.79591405e-02 1.88941494e-01 1.26109302e+00
3.89797479e-01 -4.14942801e-01 -9.43013072e-01 6.62212849e-01
1.98796630e-01 -1.23169088e+00 -2.92139858e-01 2.36424897e-02
8.06634605e-01 9.07269567e-02 -2.74728775e-01 2.65771061e-01
4.19168651e-01 -4.93163258e-01 4.68801349e-01 2.11825475e-01
2.65066698e-02 -4.10812974e-01 9.68246102e-01 5.01718283e-01
-5.36724210e-01 -2.07416996e-01 -3.06943804e-01 -3.07454497e-01
3.03261101e-01 1.18952286e+00 -9.27311718e-01 2.62853652e-02
4.35146302e-01 7.31446028e-01 -8.87681901e-01 1.37458956e+00
-4.21805412e-01 9.13123071e-01 -8.31373855e-02 -5.41594923e-01
-3.72610278e-02 -4.53081429e-01 2.61280477e-01 8.02471817e-01
-1.16004370e-01 -4.93855566e-01 4.59004454e-02 7.01786578e-01
-2.95650512e-01 3.98956150e-01 -6.14604175e-01 -4.96946692e-01
4.15306568e-01 1.61582088e+00 -1.13157701e+00 -4.68312055e-01
-7.65167236e-01 9.56839144e-01 2.66602993e-01 1.03334852e-01
-3.65332127e-01 -2.89449811e-01 1.51332989e-01 5.92825055e-01
1.68675736e-01 -6.14859499e-02 -3.72680008e-01 -1.26035857e+00
4.66055572e-02 -1.03195941e+00 6.29995540e-02 -3.00463349e-01
-1.84028864e+00 2.99012005e-01 -5.76894760e-01 -1.12796891e+00
-2.55051494e-01 -7.30605543e-01 -5.96116602e-01 6.01368427e-01
-1.56031632e+00 -8.48150015e-01 -1.07802995e-01 -3.44432378e-03
4.13610280e-01 -1.92030877e-01 6.93748355e-01 5.31134665e-01
-4.38495517e-01 3.20612133e-01 6.14893399e-02 8.23088810e-02
9.39511478e-01 -1.25674391e+00 7.93153048e-01 5.97214401e-01
2.88840264e-01 8.15217316e-01 6.43969357e-01 -6.50985003e-01
-1.21856976e+00 -6.68066561e-01 1.33786058e+00 -5.44281363e-01
1.01053369e+00 -7.93954432e-01 -8.80661607e-01 3.52794886e-01
3.33470069e-02 -5.43136448e-02 7.81636119e-01 4.23589140e-01
-5.97186446e-01 -1.93009570e-01 -1.17072666e+00 3.84756088e-01
7.66443253e-01 -4.40552354e-01 -4.18408602e-01 6.29421532e-01
4.27370310e-01 3.29087496e-01 -1.16508260e-01 1.83354765e-01
4.32657719e-01 -1.07442546e+00 4.47325617e-01 -5.76595128e-01
1.62822261e-01 -7.62841254e-02 4.89219099e-01 -1.35717607e+00
-4.65331040e-02 -5.98844528e-01 -3.97049278e-01 1.18634975e+00
8.44697893e-01 -9.84843135e-01 1.35114205e+00 7.17609286e-01
2.81862497e-01 -9.24005926e-01 -4.04013962e-01 -4.86808091e-01
-2.46208027e-01 -2.00029537e-01 1.97472468e-01 1.02696049e+00
2.72622168e-01 7.05505133e-01 6.08115718e-02 -4.82328266e-01
5.77066183e-01 1.29274838e-02 7.68937409e-01 -1.77237833e+00
-2.06780151e-01 -3.09631050e-01 -3.71796608e-01 -1.09486532e+00
1.28275171e-01 -4.68657851e-01 -1.44516602e-01 -1.63098228e+00
1.75492898e-01 -7.18199551e-01 7.25719100e-03 3.61730516e-01
-1.03103392e-01 4.88470405e-01 -2.78678298e-01 3.93789768e-01
-1.00948393e+00 2.73500711e-01 1.17235661e+00 -1.13761127e-01
-1.73269108e-01 5.60978293e-01 -8.92377615e-01 8.19533110e-01
9.84318376e-01 -5.39286911e-01 -4.84550297e-01 4.41623852e-02
9.32657599e-01 -7.96716750e-01 1.16410395e-02 -2.75288433e-01
2.96912402e-01 5.65263964e-02 -8.41476694e-02 -4.40439552e-01
1.51884407e-01 -7.48994708e-01 -4.83530313e-01 3.27664942e-01
-2.87542671e-01 -1.08022481e-01 -5.65786779e-01 7.61164784e-01
-2.88872868e-01 -5.88896096e-01 6.93677545e-01 -4.84811395e-01
-4.80402142e-01 6.83425814e-02 -4.87242073e-01 -2.63720118e-02
7.79467583e-01 -9.23879594e-02 -7.50486672e-01 -5.88579595e-01
-4.13760811e-01 1.68284908e-01 7.15621591e-01 7.44039059e-01
2.64273852e-01 -8.24684203e-01 -3.95249009e-01 1.41625226e-01
3.46855283e-01 -4.07310456e-01 -4.79237556e-01 8.07212591e-01
-4.68524903e-01 2.36691013e-01 2.80075878e-01 -2.19187573e-01
-1.02237225e+00 4.89842981e-01 1.67653173e-01 -4.38104630e-01
-1.37182459e-01 7.48851418e-01 -5.23171127e-01 -7.49491513e-01
2.44327888e-01 2.64214009e-01 -1.55838117e-01 4.02195811e-01
4.31347430e-01 5.76452732e-01 6.94211125e-01 -5.06183326e-01
-2.16554388e-01 -2.38480985e-01 -5.69646776e-01 -7.12155551e-02
1.28583300e+00 -1.60673246e-01 -5.02504349e-01 5.86973727e-01
1.11663258e+00 1.67302042e-01 -6.10848367e-01 -3.73374313e-01
5.39748609e-01 -8.17712009e-01 9.84667242e-02 -7.67127812e-01
-1.02156091e+00 6.90641224e-01 4.05379146e-01 9.67479765e-01
5.24107873e-01 -6.08554557e-02 7.13900506e-01 4.17897850e-01
6.18133485e-01 -1.44472170e+00 2.57277340e-01 2.49852315e-01
2.25385651e-01 -1.72824144e+00 2.89600402e-01 -1.00881755e+00
-6.61567450e-01 7.64820874e-01 4.00294870e-01 -2.91711003e-01
6.98974192e-01 -1.19125977e-01 4.24101740e-01 -2.65771419e-01
-9.04700696e-01 -1.47645548e-01 1.49090022e-01 4.17055964e-01
7.97158301e-01 -7.61212856e-02 -9.93169785e-01 3.62396091e-01
2.24525526e-01 -1.55904830e-01 6.01478457e-01 1.02724993e+00
-5.91171563e-01 -1.84312856e+00 -1.08659916e-01 8.39068413e-01
-4.11988705e-01 2.98648570e-02 -9.64749694e-01 5.78035831e-01
1.02426037e-01 1.40383720e+00 -8.51587504e-02 -9.17636380e-02
1.29218429e-01 2.50752091e-01 1.24781862e-01 -9.78166223e-01
-8.66679013e-01 -1.50794432e-01 3.87925088e-01 -5.26708126e-01
-5.50995827e-01 -6.04882598e-01 -7.95648158e-01 -3.54983479e-01
-1.05347788e+00 4.72535223e-01 9.32440341e-01 8.51816416e-01
4.31882292e-01 -5.38676679e-02 8.59142482e-01 -4.92689431e-01
-8.73121262e-01 -8.58901322e-01 -6.31949246e-01 8.48676741e-01
-9.67758819e-02 -6.76493824e-01 -9.19218004e-01 -2.32287109e-01] | [7.8513617515563965, 10.03599739074707] |
5c6aa70c-74ed-4e2e-a1f9-def53a4f91db | aircraft-engines-remaining-useful-life | null | null | https://www.sciencedirect.com/science/article/abs/pii/S095219762030258X | https://www.sciencedirect.com/science/article/pii/S095219762030258X?casa_token=sl4Tdv5eIjYAAAAA:91VC-_dXZoqUR4iKbieXHebBKvqb7yZAoOdQCGnYfsD14jqwCS3ZJxb5kZZBiVO41wWzgqErPGI | Aircraft engines Remaining Useful Life prediction with an adaptive denoising online sequential Extreme Learning Machine | Remaining Useful Life (RUL) prediction for aircraft engines based on the available run-to-failure measurements of similar systems becomes more prevalent in Prognostic Health Management (PHM) thanks to the new advanced methods of estimation. However, feature extraction and RUL prediction are challenging tasks, especially for data-driven prognostics. The key issue is how to design a suitable feature extractor that is able to give a raw of time-varying sensors measurements more meaningful representation to enhance prediction accuracy with low computational costs. In this paper, a new Denoising Online Sequential Extreme Learning Machine (DOS-ELM) with double dynamic forgetting factors (DDFF) and Updated Selection Strategy (USS) is proposed. First, depending on the characteristics of the training data that comes from aircraft sensors, robust feature extraction using a modified Denoising Autoencoder (DAE) is introduced to learn important patterns from data. Then, USS is integrated to ensure that only the useful data sequences pass through the training process. Finally, OS-ELM is used to fit the non-accumulative linear degradation function of the engine and to address dynamic programming by trucking the new coming data and forgetting gradually the old ones based on the proposed DDFF. The proposed DOS-ELM is tested on the public dataset of commercial modular aeropropulsion system simulation (C-MAPSS) of a turbofan engine and compared with OS-ELM trained with ordinary Autoencoder (AE), basic OS-ELM and previous works from the literature. Comparison results prove the effectiveness of the new integrated robust feature extraction scheme by showing more stability of the network responses even under random solutions. | ['Mohamed Benbouzid', 'Lotfi Saïdi', 'Ouahab Kadri', 'Leïla-Haye Mouss', 'Tarek Berghout'] | 2020-09-08 | null | null | null | null | ['exponential-degradation'] | ['time-series'] | [-5.90487979e-02 -3.91017675e-01 3.98831695e-01 -2.43804932e-01
-6.65004402e-02 9.83668268e-02 4.17056903e-02 2.77074635e-01
-4.03306514e-01 8.15169930e-01 -3.90089482e-01 -1.45151347e-01
-8.83191466e-01 -6.22783720e-01 -4.72478509e-01 -1.08755910e+00
-2.85028666e-01 4.89706099e-01 1.40590698e-01 -4.27173942e-01
-1.53440135e-02 8.63712788e-01 -1.92404473e+00 3.11738979e-02
7.16652453e-01 1.27693081e+00 3.11347634e-01 5.99711478e-01
2.97154218e-01 6.95250571e-01 -4.73499149e-01 2.93527126e-01
1.07580990e-01 -1.62489966e-01 -2.54388768e-02 -1.14990346e-01
-3.61800879e-01 -2.00376138e-01 -2.42592230e-01 5.43434680e-01
6.65289164e-01 5.28571129e-01 7.84171760e-01 -1.04618847e+00
7.49811307e-02 1.58209145e-01 -1.02083711e-02 2.33239055e-01
-6.68176487e-02 3.35681677e-01 2.23419905e-01 -9.76940215e-01
3.92107397e-01 7.93585539e-01 9.60495532e-01 4.44288820e-01
-8.78914714e-01 -2.05013707e-01 -2.26013556e-01 5.48018873e-01
-1.24667978e+00 -6.44201711e-02 8.98700416e-01 -4.25228953e-01
8.44291687e-01 2.37962499e-01 7.01531351e-01 9.55536902e-01
1.01140308e+00 1.36809781e-01 1.08635008e+00 -2.15041906e-01
6.33941770e-01 3.07984859e-01 3.11445177e-01 4.90813255e-01
3.64623129e-01 5.49951553e-01 -3.33090901e-01 -1.16958126e-01
1.06611505e-01 3.59646618e-01 -2.84696609e-01 -8.17330554e-02
-5.08337557e-01 5.14996827e-01 6.02707788e-02 7.71925867e-01
-1.00376606e+00 -2.14161977e-01 4.59618062e-01 6.61186099e-01
5.22814512e-01 4.28768337e-01 -8.13110530e-01 -3.53975259e-02
-1.07813561e+00 -9.28152818e-03 8.24883699e-01 3.07816267e-01
5.37155449e-01 7.64118493e-01 -8.13917741e-02 5.18761277e-01
1.20664895e-01 5.31483650e-01 6.94148719e-01 -6.31759346e-01
-1.83883145e-01 6.39346898e-01 3.18398662e-02 -8.35448086e-01
-6.77688956e-01 -9.98575807e-01 -1.02498865e+00 7.63998389e-01
-1.74164250e-01 -5.10473907e-01 -6.82859123e-01 1.28960037e+00
4.64983016e-01 -1.62195936e-02 2.69179791e-01 5.69410861e-01
2.70910382e-01 9.14826930e-01 -2.70011481e-02 -5.79728246e-01
1.03652787e+00 -2.75952607e-01 -9.04543400e-01 5.40625565e-02
2.47422233e-01 -2.15342760e-01 5.56162477e-01 8.36413443e-01
-6.26402557e-01 -9.43897009e-01 -1.37331927e+00 6.62491441e-01
-5.26182711e-01 2.85511762e-01 4.35100012e-02 3.80598605e-01
-8.19245100e-01 1.36823976e+00 -8.69432211e-01 -1.56494811e-01
-8.87182653e-02 3.97245288e-01 -3.21306288e-01 9.01429430e-02
-1.33264422e+00 1.06556475e+00 6.73987389e-01 5.29802084e-01
-1.26850069e+00 -8.90661597e-01 -4.48754519e-01 1.18473582e-01
2.39682049e-01 -7.86448658e-01 7.18031168e-01 -9.97268200e-01
-1.68692636e+00 -1.21498801e-01 1.70519710e-01 -6.40431941e-01
5.06477833e-01 -2.60136455e-01 -7.77451038e-01 2.27789432e-01
-6.24316096e-01 -4.38001081e-02 1.43273652e+00 -1.08847022e+00
-4.69259620e-01 -3.33099335e-01 -7.05050111e-01 -1.42177552e-01
-7.08815038e-01 -2.39326149e-01 4.95359570e-01 -4.19359863e-01
-9.17904377e-02 -6.63854003e-01 -3.89401317e-02 -3.64697665e-01
1.10269181e-01 -1.35885164e-01 1.17541564e+00 -1.22100866e+00
1.32363224e+00 -2.24968529e+00 3.86451542e-01 3.31612736e-01
-5.72904274e-02 2.35495687e-01 3.27267617e-01 6.36852980e-01
-4.45214778e-01 -6.40223324e-01 -4.53628182e-01 -3.68817210e-01
-2.61805564e-01 3.71783644e-01 -9.05946195e-02 4.68441516e-01
2.47333005e-01 2.41938576e-01 -4.86124247e-01 -2.32927799e-01
3.35362524e-01 5.34021735e-01 -1.14406839e-01 4.76387471e-01
-1.78851828e-01 5.80367148e-01 -3.80785108e-01 6.02505803e-01
4.75301862e-01 3.92677367e-01 -2.69446880e-01 -3.59408826e-01
-2.97429651e-01 -8.71044457e-01 -1.29435074e+00 9.63574171e-01
-7.86115706e-01 3.01966816e-01 1.24693930e-01 -1.14268637e+00
1.23644149e+00 6.70749962e-01 7.44270980e-01 -3.47585618e-01
5.57340801e-01 4.44801062e-01 -1.53266996e-01 -7.86466181e-01
1.44214258e-01 -4.51351672e-01 3.54138166e-01 -8.00955743e-02
3.95486861e-01 3.94140817e-02 -1.50906695e-02 -3.04950923e-01
1.08700264e+00 8.81318823e-02 1.75248068e-02 -3.60674053e-01
9.86400545e-01 -2.29299262e-01 8.22184205e-01 3.38684708e-01
1.90106072e-02 5.01146950e-02 3.00529033e-01 -6.06213570e-01
-1.06568813e+00 -5.70143044e-01 -1.43011689e-01 4.74102110e-01
-4.72246885e-01 3.99317220e-02 -6.03819430e-01 -3.50549936e-01
6.53389245e-02 1.07581627e+00 -4.25505757e-01 -8.03205729e-01
-3.96612227e-01 -8.71033609e-01 8.81507173e-02 2.46148929e-01
2.06410721e-01 -9.78970349e-01 -9.12389100e-01 5.11643350e-01
5.78015327e-01 -5.53152263e-01 3.70047092e-01 6.08850300e-01
-1.39411068e+00 -8.30432177e-01 -5.39103985e-01 -3.26690167e-01
5.28817654e-01 -6.63581491e-01 7.95848429e-01 9.32627991e-02
-3.07319224e-01 5.12521148e-01 -2.48409688e-01 -4.86332864e-01
-7.87699521e-01 -4.24955003e-02 4.49884534e-01 3.52203608e-01
-1.07596800e-01 -6.79830551e-01 -3.48750025e-01 9.55414623e-02
-9.33334351e-01 -3.65319371e-01 7.63830006e-01 1.11971223e+00
4.14170474e-01 8.22287798e-01 1.04277301e+00 -6.14358306e-01
7.05489933e-01 -6.95847690e-01 -6.78342700e-01 2.13195577e-01
-1.36468017e+00 2.68956900e-01 1.17650568e+00 -3.96575779e-01
-1.19584119e+00 -5.87132685e-02 -3.27841282e-01 -8.17524910e-01
-2.01077610e-01 5.67219555e-01 -6.58518448e-02 -5.71885742e-02
4.08207446e-01 3.31126273e-01 3.78265053e-01 -6.21116459e-01
-3.53454202e-01 7.22217083e-01 3.62186283e-01 -3.76565188e-01
5.17286301e-01 5.98634630e-02 5.55027723e-01 -8.71660352e-01
-3.11832666e-01 -1.94872484e-01 -4.40046608e-01 -8.12444091e-01
4.61417586e-01 -5.63731015e-01 -9.07676876e-01 8.39138925e-01
-7.71731913e-01 -2.25011520e-02 -6.81447923e-01 6.02268159e-01
-4.43805635e-01 1.76965281e-01 -6.46797895e-01 -1.15886772e+00
-7.64526188e-01 -9.75959599e-01 5.13190508e-01 2.75222629e-01
2.55507976e-01 -1.03891087e+00 9.51808915e-02 -1.22568935e-01
6.76034510e-01 5.65773189e-01 1.00337541e+00 -6.81220949e-01
4.34000939e-02 -8.75366390e-01 7.60313213e-01 1.36416698e+00
-1.10296018e-01 1.61796644e-01 -8.18919361e-01 -5.99793136e-01
9.19247806e-01 1.44160330e-01 7.87800610e-01 2.87320018e-01
8.23811173e-01 -1.28347307e-01 -1.45772740e-01 3.40586364e-01
1.76652503e+00 3.98667932e-01 3.08885634e-01 2.29194865e-01
2.07995594e-01 4.78982985e-01 6.96022332e-01 7.19671249e-01
-3.44124168e-01 1.33577004e-01 5.12239933e-01 2.12972015e-01
1.53534189e-01 2.11191639e-01 7.25175440e-01 1.07932878e+00
-1.22123353e-01 -6.76921755e-02 -7.17479587e-01 3.18609029e-01
-1.50886106e+00 -6.78902566e-01 -3.69279385e-02 2.26713252e+00
4.64548349e-01 4.36298966e-01 -2.70570815e-01 8.11482430e-01
6.58453524e-01 -2.00172096e-01 -6.00307167e-01 -4.64021653e-01
-1.22555837e-01 3.75918984e-01 4.41678613e-01 2.82475173e-01
-7.71967351e-01 -3.17899063e-02 4.55999041e+00 8.58247697e-01
-1.12693286e+00 2.76139528e-01 4.08610255e-01 3.34697850e-02
1.66705489e-01 5.56279300e-03 -6.92649126e-01 6.28070533e-01
1.68710935e+00 9.48122144e-02 3.42659861e-01 9.53197896e-01
3.50631118e-01 -1.74910069e-01 -7.95087636e-01 6.27905250e-01
1.08614601e-01 -5.62848985e-01 -2.04037338e-01 -3.86215776e-01
5.77707589e-01 -3.26178908e-01 -1.23789366e-02 4.17261839e-01
-5.52981436e-01 -4.51573342e-01 5.71546853e-01 1.58118963e+00
4.22160774e-01 -9.55345452e-01 1.35041475e+00 5.81552386e-01
-8.24649572e-01 -1.02814496e+00 -1.92534119e-01 6.20089024e-02
2.65290648e-01 1.16117120e+00 -7.16441035e-01 1.05394757e+00
5.84021688e-01 5.77034414e-01 -5.99519193e-01 9.35970068e-01
1.65892124e-01 7.72046447e-01 -4.13380235e-01 -1.22871839e-01
-6.24595881e-02 -1.20018259e-01 9.69627678e-01 5.52713990e-01
7.87395895e-01 -3.09904546e-01 -2.66961694e-01 6.28619194e-01
4.58291113e-01 -9.77045968e-02 -4.46173668e-01 -2.19376497e-02
1.70794427e-01 1.37317860e+00 -4.02768940e-01 -3.04217905e-01
-3.07166260e-02 6.69966936e-01 -3.12403768e-01 3.70034456e-01
-6.88926995e-01 -4.51244026e-01 2.14021847e-01 3.65186572e-01
2.67900914e-01 -3.48816842e-01 1.13336727e-01 -5.92020452e-01
-1.40980542e-01 -7.39268303e-01 4.78244454e-01 -7.95394063e-01
-1.24381673e+00 9.12331462e-01 1.08574899e-02 -1.50629365e+00
-4.87240434e-01 -8.29355657e-01 -5.46805024e-01 7.42514610e-01
-1.26554668e+00 -8.07093441e-01 -3.32873672e-01 4.57813501e-01
4.83757287e-01 -4.36515749e-01 6.00460708e-01 4.49029297e-01
-7.84859240e-01 2.31698696e-02 6.72326505e-01 -6.69563711e-01
6.02450073e-01 -1.02462542e+00 -5.34812510e-01 7.94308126e-01
-6.22314632e-01 1.33616880e-01 1.12493038e+00 -8.91719580e-01
-1.54319656e+00 -9.44415092e-01 5.13902605e-01 4.25997786e-02
4.45209563e-01 1.33414015e-01 -9.52983260e-01 2.39660785e-01
1.61814004e-01 4.41752635e-02 1.73189744e-01 -4.58183855e-01
5.78047395e-01 -8.59252274e-01 -1.27654815e+00 3.39058675e-02
1.93809226e-01 -1.29860014e-01 -5.65801442e-01 1.43866077e-01
4.00360465e-01 -1.90591842e-01 -1.26458299e+00 9.07384753e-01
3.29536080e-01 -8.47903073e-01 7.61294186e-01 -6.10629439e-01
8.93606916e-02 -3.81132513e-01 1.28778547e-01 -1.45459044e+00
-1.89439744e-01 -3.21126401e-01 -6.60564840e-01 1.10872233e+00
1.08664054e-02 -7.56374478e-01 1.94436044e-01 4.55840319e-01
-4.76420760e-01 -1.16760516e+00 -1.19155097e+00 -9.18407679e-01
-2.61518478e-01 -1.81671813e-01 3.95799577e-01 3.94731045e-01
-5.53084016e-01 -1.62146837e-01 -4.94122714e-01 2.65166938e-01
5.97689629e-01 -2.46027842e-01 1.67210594e-01 -1.61506510e+00
-4.86533016e-01 1.23665914e-01 -4.50448632e-01 2.66078085e-01
1.01025224e-01 -5.85050762e-01 3.64414463e-03 -1.06722403e+00
-4.42026347e-01 -9.15028825e-02 -7.58442044e-01 1.43081114e-01
2.67966855e-02 -4.05976444e-01 -6.95419535e-02 -8.18254724e-02
-8.83713365e-02 1.06376386e+00 7.23163009e-01 -4.21412289e-02
-1.20046616e-01 3.35159987e-01 1.43260032e-01 5.19663632e-01
6.94045126e-01 -6.96106434e-01 -3.76111507e-01 2.03232139e-01
3.10988694e-01 4.36002433e-01 4.69427466e-01 -1.91673446e+00
4.30886686e-01 2.66749650e-01 7.47744381e-01 -6.67595029e-01
2.95366526e-01 -1.31709874e+00 7.78170407e-01 9.23012972e-01
2.39219531e-01 2.59090930e-01 2.47110799e-01 6.61984146e-01
-4.76112127e-01 -7.25542903e-01 7.23191500e-01 1.16446838e-01
-7.38546789e-01 1.00925632e-01 -5.52790701e-01 -6.50649548e-01
1.14306939e+00 -1.63022488e-01 1.55877978e-01 -3.70814241e-02
-1.08894289e+00 1.69611033e-02 7.58651644e-02 1.02612734e-01
6.93914354e-01 -9.90077555e-01 -6.89547360e-01 4.19615149e-01
-2.62123346e-01 -3.37530911e-01 9.26168978e-01 1.08785307e+00
-3.06827605e-01 1.18310414e-01 -2.86515355e-01 -2.83463001e-01
-9.78004336e-01 8.38054776e-01 6.40897632e-01 -2.37779528e-01
-5.68804204e-01 2.94437677e-01 -7.66518950e-01 1.24915633e-02
-6.28993101e-03 -2.58635342e-01 -4.38458264e-01 3.21538836e-01
2.41269350e-01 9.16975379e-01 7.43814290e-01 -2.60508627e-01
-1.25375241e-01 3.35860431e-01 3.69999170e-01 2.27295384e-01
1.79386771e+00 1.68369234e-01 -3.18012714e-01 8.46569002e-01
9.20682073e-01 -3.39814693e-01 -1.30215478e+00 3.40502650e-01
1.59172684e-01 2.65049428e-01 4.72472697e-01 -6.96133137e-01
-1.12091374e+00 7.31046915e-01 1.26119912e+00 1.42378539e-01
1.64475405e+00 -6.87469959e-01 7.38676548e-01 5.00392795e-01
2.47684002e-01 -1.35442412e+00 -1.38805717e-01 4.04244214e-01
1.05116177e+00 -5.52367747e-01 3.81652862e-02 4.93817896e-01
-4.85187203e-01 1.58007514e+00 3.65305245e-01 -2.12268829e-01
1.09880435e+00 4.68292415e-01 -4.24034297e-01 -2.15207636e-01
-8.78880680e-01 3.38569701e-01 2.53781751e-02 1.43579459e-02
-1.23640060e-01 -1.88117340e-01 -5.88645756e-01 1.15672910e+00
4.38297421e-01 3.88199210e-01 3.01951498e-01 1.05386615e+00
-5.12641013e-01 -9.53631580e-01 -6.69661760e-01 5.89731574e-01
-2.60358781e-01 2.88610458e-01 3.63800704e-01 6.52092636e-01
4.48800325e-01 6.48923874e-01 -2.82050490e-01 -6.08418465e-01
5.33041000e-01 5.83230495e-01 1.92943718e-02 -1.86885688e-02
-6.58684671e-01 -4.08360541e-01 -1.60138071e-01 -3.02959949e-01
-4.29443903e-02 -7.32411265e-01 -1.03888619e+00 2.98062060e-02
-4.57519114e-01 4.25978690e-01 1.04007733e+00 9.32912171e-01
4.87609297e-01 1.10435104e+00 9.70258415e-01 -7.51060009e-01
-1.08888102e+00 -1.23528850e+00 -9.70798016e-01 4.53728028e-02
4.61258113e-01 -9.51331079e-01 -8.43729496e-01 -1.50978506e-01] | [6.766389846801758, 2.4263031482696533] |
cf9e31a7-945f-4f79-832c-46c3648f2063 | an-unbiased-transformer-source-code-learning | 2304.11072 | null | https://arxiv.org/abs/2304.11072v1 | https://arxiv.org/pdf/2304.11072v1.pdf | An Unbiased Transformer Source Code Learning with Semantic Vulnerability Graph | Over the years, open-source software systems have become prey to threat actors. Even as open-source communities act quickly to patch the breach, code vulnerability screening should be an integral part of agile software development from the beginning. Unfortunately, current vulnerability screening techniques are ineffective at identifying novel vulnerabilities or providing developers with code vulnerability and classification. Furthermore, the datasets used for vulnerability learning often exhibit distribution shifts from the real-world testing distribution due to novel attack strategies deployed by adversaries and as a result, the machine learning model's performance may be hindered or biased. To address these issues, we propose a joint interpolated multitasked unbiased vulnerability classifier comprising a transformer "RoBERTa" and graph convolution neural network (GCN). We present a training process utilizing a semantic vulnerability graph (SVG) representation from source code, created by integrating edges from a sequential flow, control flow, and data flow, as well as a novel flow dubbed Poacher Flow (PF). Poacher flow edges reduce the gap between dynamic and static program analysis and handle complex long-range dependencies. Moreover, our approach reduces biases of classifiers regarding unbalanced datasets by integrating Focal Loss objective function along with SVG. Remarkably, experimental results show that our classifier outperforms state-of-the-art results on vulnerability detection with fewer false negatives and false positives. After testing our model across multiple datasets, it shows an improvement of at least 2.41% and 18.75% in the best-case scenario. Evaluations using N-day program samples demonstrate that our proposed approach achieves a 93% accuracy and was able to detect 4, zero-day vulnerabilities from popular GitHub repositories. | ['Peyman Najafirad', 'Elias Bou-Harb', 'Dylan Manuel', 'Gonzalo De La Torre Parra', 'Nafis Tanveer Islam'] | 2023-04-17 | null | null | null | null | ['vulnerability-detection'] | ['miscellaneous'] | [-8.99339616e-02 -2.25263406e-02 -1.87014669e-01 -6.40056655e-02
-9.44833040e-01 -9.23902392e-01 1.45600900e-01 4.15146470e-01
1.52955905e-01 3.42836976e-01 -1.91316858e-01 -7.05591261e-01
5.11736348e-02 -9.86683249e-01 -8.56976390e-01 -3.04482996e-01
-5.24190426e-01 -2.71273315e-01 4.36119705e-01 -2.50945687e-01
6.44093990e-01 3.19617875e-02 -1.08371580e+00 4.12919551e-01
1.02398586e+00 7.17164218e-01 -3.14703435e-01 5.07039487e-01
-3.07004005e-01 6.62844360e-01 -7.88830638e-01 -7.30569243e-01
2.97061354e-01 2.13023379e-01 -9.22857046e-01 -7.56743312e-01
4.13819075e-01 -1.06638476e-01 3.82083878e-02 1.49581420e+00
3.40004683e-01 -3.80763769e-01 1.30346105e-01 -1.43731213e+00
-8.67274404e-01 1.06978631e+00 -9.99031246e-01 3.32105339e-01
4.67695832e-01 1.90741017e-01 1.07690835e+00 -5.93927622e-01
5.43860137e-01 9.81339097e-01 1.12897551e+00 5.52725911e-01
-1.16575897e+00 -7.92412341e-01 1.81380689e-01 1.88368838e-02
-9.89605427e-01 1.56421423e-01 9.15085912e-01 -8.48177671e-01
1.24547219e+00 1.96585447e-01 1.54434234e-01 1.49941659e+00
5.40294051e-01 3.45313996e-01 7.94510663e-01 -3.06228817e-01
1.29857287e-01 2.43936390e-01 5.54034531e-01 6.40671551e-01
4.68751490e-01 8.59976709e-02 -1.04316413e-01 -9.36002433e-01
-6.73746690e-02 1.37735056e-02 -2.31428236e-01 -1.86343696e-02
-7.53684878e-01 8.89590681e-01 3.31397682e-01 2.98632592e-01
2.03522593e-02 1.80533111e-01 9.59247112e-01 5.60850620e-01
5.62392890e-01 5.69396019e-01 -5.91872096e-01 -2.70459026e-01
-6.92317247e-01 5.58874458e-02 8.91211689e-01 6.69520319e-01
7.47381687e-01 1.80882171e-01 6.18554540e-02 5.64226866e-01
2.75090814e-01 2.86606342e-01 3.81191820e-01 -2.76525527e-01
7.99579680e-01 1.12530565e+00 -3.46085906e-01 -1.25777650e+00
-3.20441395e-01 -5.02518892e-01 -4.12294179e-01 3.69701684e-01
3.83125454e-01 -2.30156243e-01 -7.10050285e-01 1.68118620e+00
2.57695138e-01 4.84394357e-02 -1.30058363e-01 3.22400957e-01
4.61644590e-01 3.16651493e-01 2.20961615e-01 2.21351922e-01
1.06707394e+00 -6.03080094e-01 -3.10261339e-01 -1.55876085e-01
9.57994759e-01 -5.58831215e-01 1.17769599e+00 5.44504762e-01
-5.79804599e-01 -2.29485840e-01 -1.04741061e+00 4.68305081e-01
-4.68926728e-01 -2.88043886e-01 6.19198859e-01 1.24822795e+00
-8.80801439e-01 5.01787424e-01 -8.42621386e-01 -3.89076918e-02
6.02793932e-01 5.80577627e-02 -5.23200095e-01 1.41861126e-01
-1.11995661e+00 6.25068665e-01 1.72185212e-01 -1.20352298e-01
-1.17362988e+00 -1.12087488e+00 -7.62269616e-01 3.15963626e-01
4.38430011e-01 -6.57388866e-02 8.35655093e-01 -1.02449381e+00
-8.27938914e-01 5.22874713e-01 4.58998352e-01 -2.36846671e-01
4.20534104e-01 -3.48667204e-01 -5.59331715e-01 -1.90464780e-01
2.57337913e-02 -3.58347714e-01 7.21935034e-01 -1.10134256e+00
-4.74601537e-01 -2.52163321e-01 4.24456149e-01 -7.08531976e-01
-8.89118433e-01 5.34250736e-01 2.99901605e-01 -4.91458178e-01
-3.60016137e-01 -7.16462195e-01 -3.70944254e-02 -4.08323467e-01
-6.69463456e-01 -4.89769541e-02 8.90195906e-01 -9.39926386e-01
1.74744964e+00 -2.20584226e+00 -9.18854102e-02 4.60113078e-01
4.43901509e-01 4.89190221e-01 -2.65580684e-01 5.58399916e-01
-4.74115670e-01 7.28329718e-01 -4.75419730e-01 6.20904677e-02
-1.08765207e-01 -5.60324252e-01 -6.00575566e-01 3.85167271e-01
2.85369068e-01 7.27190137e-01 -8.75442624e-01 -7.43695721e-02
-3.77187073e-01 3.02335948e-01 -5.95863342e-01 8.33998322e-02
-1.82547674e-01 -1.85373407e-02 -5.91036558e-01 1.01331437e+00
6.55925632e-01 -1.71390295e-01 7.41852447e-02 3.38627607e-01
6.60272548e-03 6.31147847e-02 -7.72969067e-01 1.41143310e+00
-6.12954855e-01 5.26554227e-01 -1.51344419e-01 -8.77818644e-01
1.11049783e+00 2.65500426e-01 1.54450521e-01 -5.15779018e-01
8.60328823e-02 3.02534074e-01 1.73311964e-01 -7.63752222e-01
1.06721453e-01 3.68392318e-01 -3.47905666e-01 5.67571998e-01
-2.90951747e-02 5.55706739e-01 -1.12992756e-01 2.25545242e-01
1.85625637e+00 -3.19595113e-02 1.12716809e-01 -3.43777239e-01
6.59207404e-01 -1.22116931e-01 7.08306730e-01 5.27899265e-01
-2.25889936e-01 2.54124463e-01 1.32251740e+00 -6.47863269e-01
-7.36347556e-01 -1.02219081e+00 -6.19524606e-02 8.96439373e-01
-2.98670471e-01 -3.72642785e-01 -9.86314476e-01 -1.41539443e+00
5.25296181e-02 5.51107585e-01 -9.19501722e-01 -5.27831793e-01
-5.55854023e-01 -9.01049793e-01 1.02008152e+00 5.44238389e-01
2.77906626e-01 -1.02730560e+00 -6.36439264e-01 7.22969770e-02
-3.07057500e-02 -6.80159092e-01 -4.29407775e-01 1.22657064e-02
-5.32219470e-01 -1.52025461e+00 -1.70090631e-01 -3.87585044e-01
5.73319077e-01 -5.57985902e-02 1.11966443e+00 4.55931038e-01
-6.12473786e-01 1.17005780e-01 -6.35714769e-01 -1.67597502e-01
-6.19738221e-01 2.59293050e-01 -3.14089000e-01 -1.89422160e-01
3.75614345e-01 -7.95333445e-01 -2.89847493e-01 1.40784070e-01
-8.30053329e-01 -7.60543108e-01 2.82815605e-01 7.93599844e-01
7.87490234e-02 -6.24147058e-02 1.02335167e+00 -1.22939181e+00
7.16195107e-01 -1.16304386e+00 -8.55609298e-01 4.24227208e-01
-7.81507373e-01 -1.56661689e-01 8.24873388e-01 -5.57299197e-01
-1.10714912e+00 -2.58572578e-01 -9.15214047e-02 -3.68507892e-01
1.48006350e-01 8.36896598e-01 -2.10677281e-01 -4.71840382e-01
1.23369217e+00 -1.50246918e-01 -1.05740696e-01 -3.54956418e-01
1.04231045e-01 5.68184257e-01 8.70732442e-02 -7.47524083e-01
1.00148797e+00 9.01109427e-02 -2.04999879e-01 -3.00416738e-01
-2.20808297e-01 -7.59582072e-02 -2.70139366e-01 -9.05136466e-02
5.84916949e-01 -5.76089621e-01 -7.09087431e-01 6.39365554e-01
-1.06936347e+00 -2.01354191e-01 2.60008931e-01 -1.17635220e-01
1.17582574e-01 6.87860847e-01 -4.83245999e-01 -9.06879783e-01
-6.72240257e-01 -1.36586440e+00 5.54407895e-01 5.69417924e-02
-1.20451033e-01 -1.03990602e+00 4.52462554e-01 9.35714468e-02
7.05745876e-01 9.32203174e-01 1.42002571e+00 -9.73117411e-01
-4.93995786e-01 -5.88747859e-01 -1.58728823e-01 5.66209137e-01
1.05038911e-01 5.29421210e-01 -9.79736865e-01 -4.54118550e-01
-1.32999897e-01 -2.66070962e-01 6.09192491e-01 -1.27500281e-01
1.15249455e+00 -3.16975385e-01 -4.75513548e-01 6.08591735e-01
1.81143963e+00 1.96747378e-01 6.54191911e-01 5.28049469e-01
8.71119440e-01 8.18860948e-01 3.36872339e-01 4.45868731e-01
2.50487685e-01 3.05317402e-01 9.80894446e-01 3.29810023e-01
2.27721304e-01 3.05419173e-02 6.24844611e-01 3.36234897e-01
8.70673582e-02 -3.82287838e-02 -1.67408502e+00 7.40603566e-01
-1.56270933e+00 -8.10083926e-01 -4.49112713e-01 2.38191128e+00
7.22816646e-01 3.80584151e-01 1.28568053e-01 1.13437735e-01
7.30762303e-01 -2.15304680e-02 -5.14205754e-01 -7.38635957e-01
1.49214178e-01 2.15775996e-01 3.20875883e-01 1.61965564e-01
-1.05318308e+00 5.71381569e-01 5.62538195e+00 5.71562707e-01
-1.26172960e+00 3.61567855e-01 4.37813014e-01 1.38131902e-01
-5.79604745e-01 1.72794163e-01 -5.23921430e-01 5.77000380e-01
1.05395687e+00 -4.28326011e-01 2.76556224e-01 1.27300954e+00
-4.04327840e-01 1.97597995e-01 -9.32582855e-01 2.96092808e-01
5.86351976e-02 -1.16439700e+00 -3.44968557e-01 1.20696723e-02
7.41852224e-01 1.94863155e-01 2.22776592e-01 4.85631973e-01
3.33277076e-01 -9.11442935e-01 7.49554157e-01 2.35321552e-01
6.50457203e-01 -8.65179181e-01 8.96487474e-01 2.01907024e-01
-1.14148331e+00 -6.99993789e-01 -6.58447370e-02 1.30678326e-01
-3.58686566e-01 7.15586126e-01 -5.84576488e-01 6.88193381e-01
1.07596946e+00 3.45692366e-01 -9.69276190e-01 9.70087111e-01
-1.40683323e-01 7.51079977e-01 1.61708459e-01 9.28279236e-02
4.52881046e-02 3.28426808e-01 6.37244999e-01 1.18898487e+00
2.56056905e-01 -5.25489390e-01 -1.32166306e-02 1.13686109e+00
-2.97377035e-02 -4.09890804e-03 -7.60649741e-01 -1.98736861e-01
4.88692731e-01 1.43454576e+00 -6.98119402e-01 2.64808089e-01
-7.44474947e-01 1.93491235e-01 5.76847553e-01 2.68310755e-01
-9.99068439e-01 -8.19401920e-01 7.07379222e-01 2.51996592e-02
-1.99729968e-02 1.15766272e-01 -3.43765795e-01 -1.08904672e+00
4.67284143e-01 -1.03053474e+00 4.06337440e-01 9.25156567e-03
-1.50086582e+00 9.50026035e-01 -6.29562363e-02 -1.09381819e+00
2.07561273e-02 -4.54353362e-01 -1.17960000e+00 9.86389041e-01
-1.39459980e+00 -1.24990022e+00 -2.18688220e-01 1.85167789e-01
1.32646844e-01 -5.13779104e-01 7.32237875e-01 4.41166222e-01
-7.66180694e-01 1.18488026e+00 -9.58956853e-02 3.64671260e-01
6.05991185e-01 -1.29337597e+00 8.52761865e-01 1.29704165e+00
-3.36949259e-01 9.24161732e-01 3.77363056e-01 -1.07183444e+00
-1.28407490e+00 -1.18193758e+00 3.55820209e-01 -7.61307240e-01
1.13202178e+00 -5.50640941e-01 -1.44352496e+00 7.68423319e-01
-5.53321391e-02 2.51877606e-01 7.78511643e-01 2.23931059e-01
-1.16438448e+00 1.60302445e-01 -1.34910083e+00 2.92567343e-01
8.84689629e-01 -5.95572770e-01 -3.13709557e-01 8.82398784e-02
7.77291298e-01 -2.50855207e-01 -8.86564612e-01 3.59465301e-01
4.23482776e-01 -1.06613135e+00 7.94738054e-01 -6.65227115e-01
7.63036609e-01 -1.19638741e-02 -7.14513436e-02 -1.12148476e+00
-1.91851512e-01 -5.89898407e-01 -1.93346873e-01 1.56747723e+00
7.03480184e-01 -9.48121130e-01 4.95441139e-01 4.91318673e-01
-1.85843527e-01 -8.24336648e-01 -8.95286560e-01 -7.94898927e-01
4.55472797e-01 -4.50955510e-01 7.06886768e-01 1.28827846e+00
3.75871539e-01 -3.08020562e-01 -3.44510078e-02 3.82190585e-01
7.57016599e-01 -1.48270607e-01 4.65999722e-01 -1.26489663e+00
-3.87778759e-01 -6.30632699e-01 -4.59935188e-01 3.21913213e-01
5.74003041e-01 -9.65538621e-01 -2.10948288e-01 -9.35650706e-01
3.47060323e-01 -4.22124803e-01 -4.52574223e-01 8.74258459e-01
-5.24014294e-01 -4.74416688e-02 -1.16518892e-01 -3.70039605e-02
-1.55253857e-01 1.49903225e-03 4.98726726e-01 -2.95279801e-01
-8.63777697e-02 4.70298063e-03 -7.83150673e-01 6.44083023e-01
7.40557134e-01 -8.06436360e-01 -4.38513190e-01 -3.94959718e-01
8.20263028e-01 1.07442318e-02 4.04272646e-01 -8.48044157e-01
-5.18001281e-02 5.45872003e-02 -1.41663969e-01 5.15522398e-02
-5.67310214e-01 -5.24610817e-01 1.63776428e-01 5.64809978e-01
-1.95013776e-01 2.26510853e-01 5.22689462e-01 5.72541714e-01
2.10677702e-02 -4.09348488e-01 5.50046980e-01 4.15337011e-02
-4.68521625e-01 2.25263432e-01 -1.29597902e-01 1.73286334e-01
1.25875795e+00 -1.11939430e-01 -9.81824636e-01 4.35279876e-01
-3.18753153e-01 2.56577551e-01 5.45114338e-01 8.79801214e-01
5.48542798e-01 -9.64766562e-01 -6.14780188e-01 2.05452442e-01
3.74523610e-01 -4.16599840e-01 5.71107984e-01 6.39149547e-01
-4.24305707e-01 -5.45779616e-02 -4.01482880e-01 -2.61306822e-01
-1.31244183e+00 7.76159286e-01 2.96182424e-01 -3.86305809e-01
-6.29543245e-01 9.64923561e-01 3.95757966e-02 -6.23054087e-01
1.63979679e-01 -1.02123827e-01 -2.62604028e-01 9.51625593e-03
6.86319828e-01 6.22827888e-01 2.16531605e-01 -2.84972072e-01
-5.89921772e-01 4.04514134e-01 -2.88870305e-01 4.16569948e-01
1.26477742e+00 4.74708885e-01 -5.12878716e-01 3.24961543e-01
1.17473924e+00 8.97805467e-02 -9.53147888e-01 1.17836669e-01
3.54666173e-01 -7.14825451e-01 -1.58579990e-01 -1.15590930e+00
-1.21833682e+00 9.74476278e-01 5.32553494e-01 6.36823118e-01
1.01927590e+00 -2.06855357e-01 6.09659910e-01 5.84394159e-03
6.78470373e-01 -4.88437206e-01 2.02966198e-01 3.52891505e-01
6.72641337e-01 -1.23239124e+00 -4.12454009e-01 -3.58739465e-01
-2.70949841e-01 1.15024352e+00 9.28653181e-01 -1.89160600e-01
6.15097702e-01 5.02682447e-01 -2.76448317e-02 -2.50935107e-01
-7.89689422e-01 6.40656173e-01 1.01219900e-02 7.56676674e-01
3.71889353e-01 3.67396288e-02 -8.16233233e-02 8.92558813e-01
1.47373348e-01 -4.46147233e-01 9.23101604e-01 8.27473283e-01
-1.64529413e-01 -1.18876767e+00 -3.44497174e-01 4.53820378e-01
-9.85861361e-01 -2.46487290e-01 -2.82932520e-01 6.65054679e-01
1.39547605e-03 8.20737183e-01 -4.44017738e-01 -8.06967318e-01
3.49061102e-01 1.75207257e-01 6.55227005e-02 -8.34742188e-01
-1.35815847e+00 -5.88978529e-01 -1.14980899e-01 -7.63067544e-01
3.19134057e-01 -5.85834265e-01 -1.11502206e+00 -2.97552139e-01
-4.99288976e-01 1.62722453e-01 6.76766992e-01 6.14629507e-01
5.23222506e-01 8.01211476e-01 8.20459723e-01 -3.29340458e-01
-9.13590074e-01 -8.00974011e-01 -2.61447400e-01 3.10989827e-01
3.50822151e-01 -6.87808275e-01 -6.65240288e-01 -1.16985708e-01] | [7.083827018737793, 7.771669864654541] |
39cb17c0-bb06-4bec-b7c2-da768f3f70ee | formality-style-transfer-with-shared-latent | null | null | https://aclanthology.org/2020.coling-main.203 | https://aclanthology.org/2020.coling-main.203.pdf | Formality Style Transfer with Shared Latent Space | Conventional approaches for formality style transfer borrow models from neural machine translation, which typically requires massive parallel data for training. However, the dataset for formality style transfer is considerably smaller than translation corpora. Moreover, we observe that informal and formal sentences closely resemble each other, which is different from the translation task where two languages have different vocabularies and grammars. In this paper, we present a new approach, Sequence-to-Sequence with Shared Latent Space (S2S-SLS), for formality style transfer, where we propose two auxiliary losses and adopt joint training of bi-directional transfer and auto-encoding. Experimental results show that S2S-SLS (with either RNN or Transformer architectures) consistently outperforms baselines in various settings, especially when we have limited data. | ['WenHan Chao', 'Zhoujun Li', 'Lili Mou', 'Yu Wu', 'Yunli Wang'] | 2020-12-01 | null | null | null | coling-2020-8 | ['formality-style-transfer'] | ['natural-language-processing'] | [ 4.34530616e-01 1.56491533e-01 -4.48259354e-01 -4.89278048e-01
-1.19345117e+00 -8.63866031e-01 9.72306550e-01 -5.04956245e-01
-3.22371930e-01 1.22892249e+00 5.14881134e-01 -6.34156287e-01
4.06589299e-01 -6.55284822e-01 -1.13107884e+00 -5.19637525e-01
6.71332061e-01 7.66671956e-01 -1.83674783e-01 -5.02405226e-01
-6.91847801e-02 -4.07941341e-02 -5.35642028e-01 3.55600834e-01
8.91431928e-01 5.75157225e-01 1.14059165e-01 3.46647829e-01
-3.10602427e-01 5.38277328e-01 -3.39068204e-01 -8.69120359e-01
3.34067911e-01 -8.49896371e-01 -1.00806236e+00 -2.64989138e-01
5.52142322e-01 -2.94557184e-01 -3.07158798e-01 1.06935441e+00
4.05159622e-01 -3.93903702e-02 8.11428308e-01 -9.73410308e-01
-1.36888254e+00 7.81232357e-01 -3.53519738e-01 -2.55998019e-02
1.79070994e-01 1.88738853e-01 1.25949037e+00 -1.20054615e+00
6.35296226e-01 1.63289964e+00 6.30945325e-01 1.00331485e+00
-1.69980180e+00 -8.70558262e-01 5.24152070e-02 -1.35101229e-01
-1.18920231e+00 -5.55006444e-01 5.88896275e-01 -2.76519269e-01
9.78037119e-01 -8.82027671e-02 2.73342669e-01 1.78183949e+00
2.09239230e-01 1.04527080e+00 1.32738030e+00 -5.75059593e-01
-5.90543821e-02 2.69861221e-01 -1.43169791e-01 4.47652787e-01
1.14026643e-01 2.25660354e-01 -5.48176527e-01 -4.65670228e-02
1.01710522e+00 -1.23784937e-01 -2.73500085e-01 -3.70181382e-01
-1.53232396e+00 8.86647463e-01 1.28906831e-01 4.11288947e-01
-6.56272098e-02 3.20279241e-01 6.03178144e-01 9.33557451e-01
6.75741076e-01 4.25278753e-01 -8.01383615e-01 -6.73537254e-02
-8.46674562e-01 1.76474497e-01 7.84276247e-01 1.65104377e+00
8.61608863e-01 1.88797429e-01 -3.39248478e-01 8.18064451e-01
1.28053099e-01 8.30042303e-01 7.11530149e-01 -7.20648229e-01
9.53066051e-01 1.62144139e-01 7.99793005e-02 -3.46266061e-01
2.40912467e-01 -3.79334122e-01 -1.06720960e+00 -4.87679750e-01
2.28822842e-01 -2.45037809e-01 -6.63537741e-01 2.15285158e+00
-1.81251958e-01 2.43689045e-02 3.42286468e-01 6.83550894e-01
5.18616557e-01 7.83750057e-01 2.40930710e-02 -1.62777752e-01
1.16711867e+00 -1.31995630e+00 -8.57450604e-01 -3.66376221e-01
7.97874987e-01 -7.08953679e-01 1.73344374e+00 -5.25886826e-02
-1.37437749e+00 -6.04784667e-01 -8.42731357e-01 -4.89244342e-01
-2.32260928e-01 1.15670472e-01 4.71363038e-01 3.01322758e-01
-1.21748888e+00 6.25034571e-01 -7.32379138e-01 -3.87337029e-01
2.40421459e-01 2.50170588e-01 -3.86290520e-01 2.93119475e-02
-1.66237402e+00 9.18402612e-01 3.25601488e-01 -1.02391437e-01
-7.63436735e-01 -8.54236186e-01 -1.17442393e+00 9.50126443e-03
8.01846981e-02 -1.16191149e+00 1.45286727e+00 -1.31587493e+00
-1.82570386e+00 9.58707631e-01 -2.80757785e-01 -4.10660088e-01
5.53376317e-01 -3.69930089e-01 -1.82101995e-01 -3.49910378e-01
2.17437834e-01 7.03628659e-01 8.26688945e-01 -1.05178475e+00
-3.51634532e-01 -1.28217563e-01 1.37072548e-01 3.08457643e-01
-3.84045571e-01 3.20730567e-01 -2.77400017e-01 -9.73949075e-01
-3.35773110e-01 -9.05528545e-01 -9.37113538e-02 -6.16827086e-02
-1.13660313e-01 -4.62460101e-01 4.45947498e-01 -7.42749095e-01
1.10629058e+00 -1.99273455e+00 7.66703367e-01 -2.38102078e-01
-5.82798794e-02 4.03316796e-01 -4.63989556e-01 5.98744273e-01
4.17626202e-02 1.91967249e-01 -4.14127052e-01 -5.19141197e-01
3.01603138e-01 5.18703461e-01 -8.36892843e-01 5.42237721e-02
3.76634359e-01 1.41849089e+00 -9.91548121e-01 -3.65050316e-01
-3.24175239e-01 2.52856225e-01 -5.90756357e-01 3.27217460e-01
-3.22783560e-01 6.08312309e-01 -3.41152430e-01 3.38997453e-01
3.94892871e-01 -3.73766571e-01 3.16487700e-01 1.49839148e-01
3.04996371e-01 9.67039049e-01 -3.95477146e-01 2.07414389e+00
-8.34693849e-01 5.22484362e-01 -1.12038620e-01 -9.74853635e-01
9.85416174e-01 7.11875975e-01 -1.02193184e-01 -6.71740532e-01
-5.72526529e-02 5.56339741e-01 -1.55253872e-01 5.23548527e-03
3.04816931e-01 -7.28699088e-01 -2.27472365e-01 7.59050906e-01
5.15789449e-01 -3.24281484e-01 1.46232307e-01 1.05142538e-02
8.04888606e-01 6.10470355e-01 2.89433271e-01 -3.63583565e-01
6.42184675e-01 -2.88927108e-01 7.67472267e-01 4.80125248e-01
-1.16759151e-01 3.30638260e-01 4.88012940e-01 -3.66596878e-01
-1.35902417e+00 -1.13895977e+00 1.63362205e-01 1.18534768e+00
-1.12344220e-01 -3.37137878e-01 -8.68427753e-01 -9.80934262e-01
-1.83825761e-01 9.23224151e-01 -4.12652612e-01 -2.37026662e-01
-1.15443444e+00 -2.65168190e-01 7.90762424e-01 8.68685782e-01
1.98461622e-01 -1.24293828e+00 2.57767797e-01 2.86495805e-01
-4.29305792e-01 -1.26090682e+00 -1.01552689e+00 1.70494933e-02
-1.06192315e+00 -4.17065084e-01 -1.02117717e+00 -1.02357721e+00
5.34789443e-01 1.01195343e-01 1.59793258e+00 -1.67922631e-01
5.40289462e-01 3.25947255e-02 -1.63010493e-01 -3.36829811e-01
-7.14858055e-01 5.59226155e-01 2.72759587e-01 -2.77824581e-01
6.65363669e-01 -8.41074288e-01 -2.29333624e-01 1.95241377e-01
-8.35316896e-01 2.86527991e-01 7.44468331e-01 1.18586016e+00
4.06385630e-01 -7.59991050e-01 7.53996313e-01 -9.39432085e-01
6.70270383e-01 -2.32685417e-01 -5.29960573e-01 3.45141679e-01
-5.64314485e-01 2.08659753e-01 9.33022380e-01 -4.24248964e-01
-1.01636994e+00 -4.90368664e-01 2.08047107e-02 -5.21921933e-01
-1.44300535e-01 3.22401613e-01 -3.49147618e-01 2.81280667e-01
4.77160871e-01 6.24876797e-01 -3.40071842e-02 -5.46189129e-01
4.87770230e-01 7.66844928e-01 2.43993580e-01 -1.00992811e+00
1.09718275e+00 7.08799213e-02 -3.09665054e-01 -3.64645571e-01
-9.44755793e-01 -6.20973371e-02 -8.54860127e-01 5.01580358e-01
6.12757623e-01 -1.18120372e+00 -1.60147130e-01 2.32822672e-01
-1.56824565e+00 -7.15256989e-01 -4.98315275e-01 5.58830142e-01
-1.02437353e+00 1.86310843e-01 -1.03711045e+00 -2.77993113e-01
-6.75873101e-01 -1.08500421e+00 1.35552609e+00 -3.37832868e-01
-3.00455898e-01 -1.30833733e+00 2.16750264e-01 2.53105551e-01
5.87603092e-01 -2.09194675e-01 1.19060957e+00 -5.44854701e-01
-5.81131041e-01 2.92786241e-01 -2.38638073e-01 7.20695376e-01
4.14948404e-01 -5.70735931e-01 -5.57548642e-01 -5.12766898e-01
-3.11457179e-02 -7.05723763e-01 5.26009977e-01 2.24796981e-02
7.12046921e-01 -5.93287945e-01 -1.90317407e-01 8.03981066e-01
1.17100954e+00 -9.79416370e-02 6.49427116e-01 2.11355120e-01
7.65474975e-01 5.88102877e-01 4.16305840e-01 -2.10850522e-01
4.39442217e-01 7.25174606e-01 -3.52442861e-01 -3.52793425e-01
-3.51264477e-01 -5.71304917e-01 1.02014172e+00 1.45276880e+00
-1.37742516e-02 -1.51456878e-01 -7.15484619e-01 6.55709028e-01
-1.80943000e+00 -7.21285403e-01 2.80945003e-01 1.81504059e+00
1.57103133e+00 -5.54505996e-02 -6.58076778e-02 -3.35883915e-01
6.42499864e-01 8.76189321e-02 -4.31894630e-01 -6.37921929e-01
-1.83778331e-01 4.28277820e-01 4.92332518e-01 7.40110219e-01
-8.32760632e-01 1.56437242e+00 6.59128714e+00 9.63385165e-01
-1.06359553e+00 4.23425317e-01 3.37439984e-01 -3.54944915e-02
-7.00240970e-01 1.24994963e-01 -9.86532271e-01 4.56940323e-01
1.16181874e+00 -2.47686610e-01 6.73391402e-01 3.92463535e-01
-8.40183422e-02 8.79195213e-01 -1.70971072e+00 6.66087925e-01
6.26566485e-02 -1.25876427e+00 3.89768869e-01 -5.83362058e-02
9.48263705e-01 2.28720859e-01 1.08000986e-01 5.97883224e-01
7.26147294e-01 -9.19633389e-01 7.29714692e-01 2.34979317e-01
1.32951427e+00 -4.81562823e-01 5.35477161e-01 4.54853356e-01
-9.12406981e-01 3.93161237e-01 -4.53969717e-01 -5.99407814e-02
3.47591460e-01 1.52758425e-02 -5.04345894e-01 8.05523515e-01
3.19755137e-01 9.16032016e-01 -1.62272036e-01 -1.92626510e-02
-6.10210657e-01 8.52253020e-01 3.97087308e-03 -5.68406247e-02
5.27577400e-01 -5.57147563e-01 4.69948441e-01 1.34605563e+00
3.88163447e-01 -2.19853699e-01 6.06595762e-02 1.13936222e+00
-5.71198583e-01 2.50035435e-01 -9.36003625e-01 -2.00587466e-01
1.66009918e-01 7.31955886e-01 -4.37312201e-02 -5.72882533e-01
-8.80940139e-01 1.35341918e+00 5.66166103e-01 6.30904019e-01
-7.64865756e-01 -3.11224699e-01 7.25767553e-01 -3.99861671e-02
3.69973451e-01 -3.54298651e-01 -1.51891381e-01 -1.68502140e+00
1.91433042e-01 -1.24566996e+00 1.22874640e-02 -6.25225246e-01
-1.51734865e+00 7.24188149e-01 -8.68480876e-02 -1.20757937e+00
-5.13226390e-01 -6.05643809e-01 -3.75556171e-01 1.20966029e+00
-1.71275795e+00 -1.61064005e+00 5.43864250e-01 5.76270223e-01
8.02258253e-01 -4.11913007e-01 1.01601136e+00 3.45506608e-01
-1.57292888e-01 1.05635989e+00 3.51343125e-01 4.09559369e-01
1.13116241e+00 -1.38341892e+00 9.90477562e-01 7.08998382e-01
2.55718887e-01 9.37003195e-01 4.98788208e-01 -5.42637229e-01
-1.32653177e+00 -1.40040731e+00 1.51052284e+00 -7.08672404e-01
9.00159955e-01 -8.14384341e-01 -8.93166482e-01 1.30340326e+00
7.36269653e-01 -4.59984466e-02 8.28906059e-01 1.35926366e-01
-6.25102162e-01 7.61443451e-02 -8.16537797e-01 9.23034847e-01
1.37132609e+00 -9.36331987e-01 -9.62437391e-01 4.12857234e-01
1.22906399e+00 -1.99062303e-01 -9.31462646e-01 3.62681299e-01
4.24651444e-01 -3.91960800e-01 7.28419304e-01 -1.11379755e+00
7.01377094e-01 -7.81644508e-02 -5.19351006e-01 -1.48479319e+00
-4.40032959e-01 -9.08752799e-01 -1.04872070e-01 1.33286417e+00
6.77755594e-01 -7.05776572e-01 4.67046618e-01 1.48969218e-01
-8.82572159e-02 -6.11471415e-01 -7.12813377e-01 -1.18490148e+00
9.99259055e-01 9.49371532e-02 8.76619339e-01 1.15564132e+00
-2.07602262e-01 1.06311095e+00 -5.97556829e-01 -3.35340142e-01
4.64387983e-01 2.37069830e-01 9.90366757e-01 -8.77818584e-01
-6.03923142e-01 -4.11704451e-01 4.57821302e-02 -1.49677777e+00
7.22406447e-01 -1.32070947e+00 -1.08516812e-01 -1.26964808e+00
2.55571723e-01 -1.56941414e-01 -2.44272217e-01 5.10267556e-01
-2.40039989e-01 3.71879756e-01 4.27666418e-02 3.35679233e-01
-3.75472665e-01 1.10477901e+00 1.47955215e+00 -2.96712130e-01
-1.73656326e-02 -2.64478028e-01 -8.02705944e-01 6.11179113e-01
6.31627977e-01 -3.65225643e-01 -3.61728281e-01 -9.24868464e-01
3.39018226e-01 5.40920980e-02 -2.40464155e-02 -3.47640932e-01
-2.21969754e-01 -2.26721063e-01 -2.34647859e-02 -3.50255400e-01
7.41309375e-02 -6.48973525e-01 -2.32496083e-01 3.10344875e-01
-6.68299437e-01 1.26911655e-01 1.06274694e-01 4.61238116e-01
-4.26721334e-01 -1.83440894e-02 5.77156484e-01 -1.43691957e-01
-9.90216881e-02 4.70145136e-01 -2.14845508e-01 3.03977460e-01
5.33049762e-01 1.41580999e-01 -9.55347791e-02 -4.22589839e-01
-5.83963454e-01 1.50024727e-01 5.51066756e-01 6.06199682e-01
2.59874821e-01 -1.95447040e+00 -1.07819891e+00 2.97981769e-01
2.35383362e-01 -8.45186785e-02 -2.18567938e-01 7.61076570e-01
-9.41788778e-02 7.47609556e-01 -7.30681792e-02 -4.62476879e-01
-8.14936578e-01 6.40632391e-01 1.32196635e-01 -6.02164626e-01
-6.17135167e-01 7.68755496e-01 8.23184371e-01 -1.13393271e+00
-1.77953374e-02 -4.04269546e-01 3.99408132e-01 -3.30435276e-01
3.80316734e-01 6.70458796e-03 -3.33331972e-02 -5.97032070e-01
-1.36551052e-01 6.15384817e-01 -4.11324412e-01 -3.40258569e-01
1.24438155e+00 -2.34375194e-01 -4.04547304e-01 6.75302863e-01
1.19916952e+00 -1.72831774e-01 -1.21561754e+00 -1.02390885e+00
1.08896315e-01 -1.50711834e-01 -3.83407086e-01 -6.19749188e-01
-6.63805425e-01 1.21819758e+00 2.47505642e-02 -3.84445906e-01
9.49191630e-01 -2.21130718e-03 1.30084121e+00 5.91198683e-01
5.16367495e-01 -1.00982642e+00 -2.86261481e-03 1.03478420e+00
1.18577695e+00 -1.06705415e+00 -4.98236239e-01 -1.97665825e-01
-5.84177911e-01 7.38887727e-01 5.49131334e-01 -2.58395582e-01
3.68882537e-01 2.04502627e-01 1.80323586e-01 2.55396664e-01
-1.03362417e+00 1.06723048e-01 2.78547019e-01 5.13925850e-01
8.18547130e-01 5.21319956e-02 -4.70830888e-01 6.56549513e-01
-3.75673056e-01 6.19052611e-02 1.54005259e-01 6.63421750e-01
3.09898593e-02 -1.75117099e+00 -1.15913957e-01 2.19720453e-02
-6.16461515e-01 -6.02135479e-01 -5.44434607e-01 6.25583649e-01
-2.11893067e-01 4.69231844e-01 7.13150715e-03 -1.02765776e-01
2.21053466e-01 4.78613824e-01 8.13417494e-01 -8.50723147e-01
-5.55308342e-01 1.40048832e-01 -5.71933761e-02 -5.42183340e-01
-3.47857028e-01 -4.91326183e-01 -9.06695426e-01 -3.46236229e-01
-2.27182135e-01 2.14734167e-01 4.09468502e-01 1.02914166e+00
3.84902090e-01 4.41862494e-01 5.81028700e-01 -6.82283998e-01
-9.48931038e-01 -1.30835664e+00 -3.99920672e-01 4.11395162e-01
4.15478587e-01 -3.12242478e-01 -7.62590170e-02 3.82426113e-01] | [11.644266128540039, 9.872642517089844] |
31064e65-375d-4e77-a926-77344d73f4a1 | 3d-matting-a-benchmark-study-on-soft | 2210.05104 | null | https://arxiv.org/abs/2210.05104v1 | https://arxiv.org/pdf/2210.05104v1.pdf | 3D Matting: A Benchmark Study on Soft Segmentation Method for Pulmonary Nodules Applied in Computed Tomography | Usually, lesions are not isolated but are associated with the surrounding tissues. For example, the growth of a tumour can depend on or infiltrate into the surrounding tissues. Due to the pathological nature of the lesions, it is challenging to distinguish their boundaries in medical imaging. However, these uncertain regions may contain diagnostic information. Therefore, the simple binarization of lesions by traditional binary segmentation can result in the loss of diagnostic information. In this work, we introduce the image matting into the 3D scenes and use the alpha matte, i.e., a soft mask, to describe lesions in a 3D medical image. The traditional soft mask acted as a training trick to compensate for the easily mislabelled or under-labelled ambiguous regions. In contrast, 3D matting uses soft segmentation to characterize the uncertain regions more finely, which means that it retains more structural information for subsequent diagnosis and treatment. The current study of image matting methods in 3D is limited. To address this issue, we conduct a comprehensive study of 3D matting, including both traditional and deep-learning-based methods. We adapt four state-of-the-art 2D image matting algorithms to 3D scenes and further customize the methods for CT images to calibrate the alpha matte with the radiodensity. Moreover, we propose the first end-to-end deep 3D matting network and implement a solid 3D medical image matting benchmark. Its efficient counterparts are also proposed to achieve a good performance-computation balance. Furthermore, there is no high-quality annotated dataset related to 3D matting, slowing down the development of data-driven deep-learning-based methods. To address this issue, we construct the first 3D medical matting dataset. The validity of the dataset was verified through clinicians' assessments and downstream experiments. | ['ZongYuan Ge', 'Xin Zhao', 'Kaimin Song', 'Wei Feng', 'Xin Wang', 'Huan Luo', 'Yi Luo', 'Lie Ju', 'Wanji He', 'Donghao Zhang', 'Xiufen Ye', 'Lin Wang'] | 2022-10-11 | null | null | null | null | ['image-matting'] | ['computer-vision'] | [ 0.37847474 0.4111595 -0.2355155 -0.36500698 -0.7181227 -0.2740386
0.3617241 -0.04723692 -0.12009643 0.48680165 0.012515 -0.4813982
0.12493899 -0.782377 -0.44466418 -1.0439912 0.31737867 0.74700254
0.4195925 0.17087118 -0.30678195 0.5561162 -0.9795513 0.6377139
1.1102868 1.0168098 0.51499784 0.34523383 -0.35190946 0.41054478
-0.5496848 -0.19264233 0.40327454 -0.61528444 -0.84355426 0.41407678
0.23295893 -0.42896947 -0.3275514 1.0672184 0.4627694 -0.510137
0.8193266 -0.95400405 -0.4210728 0.5596068 -0.7868347 0.06982888
-0.04911776 0.39368504 0.4026004 -0.6692863 0.44403625 0.89832103
0.7054893 0.61095566 -0.9378501 -0.4170844 0.02349465 -0.12211797
-1.1476073 -0.01993454 0.67505634 -0.5727177 0.42430452 0.6029746
0.8734538 1.0431428 0.5053528 1.0576142 1.3957189 -0.21428955
0.06589209 -0.06704219 -0.04241617 0.88643277 0.2561157 0.02106711
0.13452516 0.12964079 0.934246 0.3277975 -0.45619908 -0.36508426
-1.3345076 0.5021152 0.7577351 0.42051122 -0.32888773 -0.10160204
0.28420484 -0.00892113 0.671572 0.09612221 -0.0741858 0.22167696
-1.1648917 -0.20061317 0.38876536 0.46901727 0.3884514 -0.09442153
-0.36791337 0.80035645 0.38625062 0.34882995 0.887816 -0.50645196
0.19891952 0.909172 -0.41936526 -0.7564673 -0.6757571 -0.7537386
-1.3912963 0.33165804 0.58465004 0.22404964 -1.7529031 1.3845572
0.5385537 0.28148273 -0.12490921 1.192113 1.1194811 0.32568455
-0.25851008 -0.22166564 1.2055103 -1.0765591 -0.6859322 -0.27293998
0.82765377 -0.6600012 1.0197967 0.2160469 -1.2319989 -0.38112402
-0.89949733 0.1827812 -0.21827774 -0.04969925 0.708576 0.7149163
-0.9807065 0.2439677 -1.2121471 -0.19474794 0.64998275 0.2580371
-0.32940313 -0.2506355 -1.1255798 1.0345899 0.39243406 0.21372281
-1.0922672 -0.6702606 -0.7575252 -0.4245416 0.3450163 -0.97073543
1.1740695 -1.0121028 -1.2294434 1.3879795 0.04597729 -0.2566063
0.94310933 0.3367857 -0.14382283 0.2101061 0.25711232 0.5493995
0.84673834 -1.3129374 -0.3545931 -0.36923182 -0.15027913 0.29992342
0.05144244 -0.5531534 -0.76292944 -0.9757981 0.45515227 -0.8862728
-0.45836616 0.22967197 -0.6984353 0.49568653 0.6942113 -0.58519125
0.966209 -2.163527 0.16921456 0.33341336 0.6847831 0.27161223
0.09214287 -0.39020324 -0.12146581 0.17682715 -0.7067535 -0.32831264
-0.2793712 0.44049728 0.18914784 0.6018517 0.0628165 1.200574
-0.8191872 -0.8429011 0.3669738 0.26694998 -0.34060726 0.19776984
-0.01030031 0.6371502 -0.524235 0.94191736 0.8780295 -0.42712918
-0.05411841 -0.3940298 0.22742464 0.0124002 -0.84998155 1.7865381
-0.3935044 0.15004772 0.3834957 -1.1783699 0.53294045 0.27528363
0.9244524 -0.61404765 0.36734837 0.4492258 0.12941068 -0.55561703
-0.03309406 -0.35415167 0.03831193 0.49388316 -0.24296233 -0.49074185
-0.16789359 0.02107566 1.149233 -0.18662061 0.11060793 -0.0463514
0.24777795 0.21469165 0.5479608 0.73154616 -0.32303867 1.1426444
0.2905262 -0.22636987 -0.8089598 -1.2956036 -0.5281324 0.22176082
0.4467683 0.00696662 -0.93303543 -1.1347018 -0.07273188 0.30703816
-0.89573205 -0.34872863 -0.47641283 -1.1990278 0.5451521 0.43274036
0.73297423 -0.69677585 -0.4640226 0.21963894 -0.319928 -0.897467
-0.52970034 0.36660308 -1.0437284 -1.1621376 -1.1516519 -0.7946821
0.8575385 0.27250004 1.0634112 0.31930602 -0.35146528 0.07839141
-0.36087865 -0.2456269 -0.7831685 0.07290895 -0.2112227 -0.08940228
0.03581622 -0.37838712 -0.6911523 0.55303067 -1.1941744 0.6575674
1.0817494 1.1022983 0.93540305 0.29164618 0.30471835 -0.9784928
0.3291665 -0.44108725 -0.08455401 0.24278948 -0.40152857 -0.07123104
0.20702137 -0.4385343 -0.95664275 0.22835864 -0.34480482 -0.44953316
-0.21916929 0.6248018 -0.13767277 -0.10258932 0.49702036 0.34780154
0.36595187 -0.3161794 0.18501718 0.702914 0.5505278 -0.2737064
0.857468 0.79869664 0.08456067 -0.43489954 -1.0880039 -0.2982238
-0.67464054 -0.27441883 0.99788773 -0.5969429 -0.09188583 0.685704
-0.8915453 -0.5649432 -0.4341822 0.46187 -0.48942903 0.54243004
-0.9309039 -0.1766146 -0.4825711 -1.7176586 1.2435312 0.05421481
-0.18280394 -0.8992585 -0.02712703 0.5839118 0.3100539 0.57333916
0.95640254 -0.35742712 -0.49251089 -0.15669207 -0.33670193 0.36454856
0.4940155 -0.14227371 -0.93249977 -0.09499067 0.26078194 -0.14658214
1.1377968 0.74990433 1.3312958 0.00822944 -0.62250525 0.95708185
0.9420528 0.14792241 0.6054377 0.36305392 0.8968478 0.38465628
0.6353945 0.08266164 0.09778897 0.4324248 0.7230552 -0.8170739
-0.5014937 -0.04621922 -0.07437771 0.74370533 0.12532295 -0.24592151
-1.0561315 0.39396298 -1.6535017 -0.500648 -0.21879819 1.7508746
1.165684 0.459228 -0.17905307 0.25339773 0.709473 -0.04548902
-0.62876004 0.0268095 -0.14488076 0.17015587 0.43725005 0.37842822
-1.2682115 0.577313 6.091092 1.133826 -1.4418031 0.19709483
1.0985984 0.06864827 -0.33929542 -0.4157608 -0.5026409 0.5898644
0.2370679 0.30439144 -0.16441473 0.60464525 0.29768008 -0.31886142
-1.1421947 0.95879406 0.03770895 -1.3523567 0.05703443 0.21089861
0.70616114 -0.03219113 0.15135752 0.10689044 0.02139777 -1.2717144
0.5933562 0.43381542 0.9757442 -0.4008705 1.017042 0.50079644
-0.79687166 0.33115408 -0.15637408 0.42844892 0.2818761 1.080174
-1.149462 0.65698946 0.5500207 0.6710965 -0.6635281 1.2869022
-0.11752516 0.5984226 -0.37995353 0.48219672 0.45468563 0.01071038
0.54394376 1.1612608 0.36042663 0.1614398 0.2672211 0.9984163
0.017468 -0.17161103 -0.32581478 0.21196873 -0.14324225 1.2340045
-1.1189817 -0.3806744 -0.20243515 1.093505 -0.21093123 0.13733093
-0.94117373 0.19979641 0.21123652 0.4439693 -0.18564405 -0.04729264
-0.6341727 -1.1693282 -0.1495527 -0.8289026 0.47069559 -0.6860284
-1.4001398 0.70604944 0.03302505 -1.2827836 0.08765294 -0.47882578
-0.6435873 0.7607853 -1.4319618 -1.0848957 -0.73775184 0.44108155
0.41344237 0.13349134 0.54581785 0.36769193 -0.5135312 0.64595157
0.07411817 0.20117874 0.7317662 -1.3982338 0.04538568 0.59683084
-0.305808 0.03792459 0.1784359 -0.8503339 -0.9657621 -1.2655473
0.16513734 -0.4247391 0.42594406 -0.22450382 -1.1061053 0.40798613
-0.05848858 0.18770756 0.5859888 -0.5609302 0.19127913 0.01246236
-1.4913774 0.6209935 0.9800238 -0.22819313 -0.5274116 0.39170203
0.68697476 -0.85867274 -0.9129468 0.90924114 0.23658098 -0.8535819
1.0412762 -0.06776336 0.32756814 -0.40160066 0.11492411 -1.3837194
-0.3153786 -0.04415325 -0.02161243 0.8601567 0.39776435 -0.4924358
1.1241367 0.4533351 -0.7340683 -1.0692393 -1.0497842 -0.48696908
0.14970532 -0.39129412 0.46374035 1.1146319 -0.32213444 -0.11147938
0.06806927 0.10917421 0.629984 0.21440314 0.393146 -0.7991845
-0.42976633 -0.6979668 -0.37077594 -1.1106372 -0.09332488 -1.2630917
-0.01476464 -1.760111 0.38434958 -0.8530298 -0.25114504 0.5665218
-0.24242237 0.5351148 -0.29298976 0.40487126 -0.21113467 0.4817678
2.1065283 -0.66941196 -0.07072362 0.16832703 -0.43557215 0.82059026
0.7704087 -0.311096 -0.3774021 -0.37668478 -0.27627322 0.12167241
0.48785996 -0.78785247 -0.05555999 -0.26838443 0.6327118 -0.9658021
0.2881708 -0.84859776 0.10125421 0.6764135 -0.11826875 -0.365832
0.20495863 0.29675874 -0.3472728 -0.2506626 1.0727288 -0.49079382
-0.33052638 0.57499635 -0.50065184 -0.02523081 1.3142221 -0.65068936
-0.12717776 -0.08572178 -0.8619162 0.20430723 0.57200813 0.02823636
0.8096363 -1.2453063 -0.71125144 0.22715831 -0.11428546 0.77160645
0.4198967 1.3033962 -0.8025161 0.17239487 -0.19764262 -1.0840186
-1.157399 0.559264 0.8928879 -0.47430977 -0.82428014 0.67710227
0.80470926 -0.38600117 0.22483858 -0.77756196 0.04856661 -0.2562251
0.3419644 -0.19899455 0.35170564 -0.41633484 -0.3380949 0.5040824
-0.2682932 0.16721386 1.0667233 -0.06692465 -0.1034743 0.16995813
1.0238744 -0.37524936 -1.0405751 -0.27384776 -0.347327 -0.48742357
0.39683565 -0.9109255 -1.5335838 0.8148461 0.8475823 0.14168978
1.3272388 0.18796653 1.019457 -0.23896731 0.06080434 -0.6362738
0.19631371 0.10262219 0.70562106 -1.3242536 -0.07706561 -0.6564577
-0.510786 0.92991817 0.73616064 0.24167272 0.6354413 0.5087578
0.293889 -0.28594068 -0.2097295 -0.16246949 0.34541303 0.6387711
0.26320213 0.16431317 -0.10310394 0.5955973 0.03690364 -0.13000931
0.41326848 0.90522206 -0.30137387 -1.0304235 -0.61291015 0.62148505
-0.4647721 -0.06415352 -0.4556122 0.7469083 0.36248672 0.47852662
-0.09605184 -0.34048975 0.19026268 -0.3726486 0.5234026 -0.703818
-0.55564636 0.3335079 -0.16375539 -0.26066816 -0.37512046 -0.37167993
-1.3739734 -0.12380309 -0.29330024 -0.13619 0.4391815 0.903128
0.01303726 0.81305563 0.5739061 -0.72144127 -0.30322826 -0.8073554
-0.51379746 0.4370027 0.25481656 -0.7987064 -0.2665829 -0.11776259] | [14.604718208312988, -2.196493148803711] |
7722c34c-a25c-4f34-b4db-7fba579f0952 | ai-based-arabic-language-and-speech-tutor | 2210.12346 | null | https://arxiv.org/abs/2210.12346v1 | https://arxiv.org/pdf/2210.12346v1.pdf | AI-based Arabic Language and Speech Tutor | In the past decade, we have observed a growing interest in using technologies such as artificial intelligence (AI), machine learning, and chatbots to provide assistance to language learners, especially in second language learning. By using AI and natural language processing (NLP) and chatbots, we can create an intelligent self-learning environment that goes beyond multiple-choice questions and/or fill in the blank exercises. In addition, NLP allows for learning to be adaptive in that it offers more than an indication that an error has occurred. It also provides a description of the error, uses linguistic analysis to isolate the source of the error, and then suggests additional drills to achieve optimal individualized learning outcomes. In this paper, we present our approach for developing an Artificial Intelligence-based Arabic Language and Speech Tutor (AI-ALST) for teaching the Moroccan Arabic dialect. The AI-ALST system is an intelligent tutor that provides analysis and assessment of students learning the Moroccan dialect at University of Arizona (UA). The AI-ALST provides a self-learned environment to practice each lesson for pronunciation training. In this paper, we present our initial experimental evaluation of the AI-ALST that is based on MFCC (Mel frequency cepstrum coefficient) feature extraction, bidirectional LSTM (Long Short-Term Memory), attention mechanism, and a cost-based strategy for dealing with class-imbalance learning. We evaluated our tutor on the word pronunciation of lesson 1 of the Moroccan Arabic dialect class. The experimental results show that the AI-ALST can effectively and successfully detect pronunciation errors and evaluate its performance by using F_1-score, accuracy, precision, and recall. | ['Abdessamad Mbarki', 'Sonia Shiri', 'Pratik Satam', 'Salim Hariri', 'Saleem Alharir', 'Sicong Shao'] | 2022-10-22 | null | null | null | null | ['self-learning'] | ['natural-language-processing'] | [-2.29183912e-01 1.22803502e-01 1.86555460e-01 -2.33512536e-01
-6.95896924e-01 -6.23424709e-01 1.07375994e-01 5.47090113e-01
-2.66469747e-01 7.88754761e-01 2.51870714e-02 -8.84364605e-01
-3.43776375e-01 -7.75195122e-01 -5.20669639e-01 -3.58462632e-01
-2.65052728e-02 4.41260397e-01 2.74917722e-01 -8.31326783e-01
6.69271588e-01 6.75007641e-01 -1.66582894e+00 4.58085001e-01
1.54783583e+00 6.45108163e-01 4.97000694e-01 9.52260137e-01
-5.09642541e-01 1.22871542e+00 -1.16040301e+00 -1.00971974e-01
-4.17893797e-01 -5.75723588e-01 -1.40263844e+00 -5.15977919e-01
1.10628530e-01 -1.16943270e-01 1.57358110e-01 8.05361450e-01
6.64652228e-01 2.63576925e-01 6.41631901e-01 -1.04789126e+00
-3.66342694e-01 7.94751704e-01 1.20072670e-01 4.01758492e-01
8.01690400e-01 5.36170304e-02 3.13324422e-01 -6.33082569e-01
9.28283408e-02 1.35811305e+00 8.62502694e-01 6.24920070e-01
-5.81324458e-01 -6.11635268e-01 -2.14075461e-01 4.81174409e-01
-9.61164057e-01 -1.50770815e-02 3.87322903e-01 -4.24119473e-01
1.18360078e+00 -6.97313845e-02 8.97720397e-01 5.87867677e-01
3.88673395e-01 9.28585052e-01 1.14108348e+00 -1.27719688e+00
1.79219730e-02 6.76525235e-01 3.54408622e-01 8.39529634e-01
-4.53995407e-01 -1.17103325e-03 -5.51496625e-01 2.96258390e-01
3.47068757e-01 -3.03528339e-01 -7.14953020e-02 3.84637356e-01
-7.96786726e-01 8.98556471e-01 2.02248976e-01 7.41398513e-01
-3.09336215e-01 -2.42900655e-01 3.74737680e-01 8.78711998e-01
7.17850253e-02 6.76734686e-01 -8.34064662e-01 -7.66104579e-01
-3.82856905e-01 -1.10822901e-01 9.98858809e-01 7.47819006e-01
1.86291933e-01 4.61791039e-01 -7.91131053e-03 1.14802003e+00
4.90310252e-01 3.03653300e-01 1.30115557e+00 -6.56528652e-01
7.01265931e-02 7.83387959e-01 -4.06640708e-01 -7.42591321e-01
-8.17185789e-02 4.57944982e-02 -3.12596783e-02 4.79722023e-01
5.90781629e-01 -7.16681123e-01 -7.31474221e-01 1.42159462e+00
2.47501269e-01 -8.00410227e-04 4.56606030e-01 2.75069922e-01
9.82713282e-01 9.31406677e-01 6.38145655e-02 -1.88026026e-01
1.40227556e+00 -8.01227629e-01 -1.09616911e+00 2.13336557e-01
9.75854397e-01 -1.16687179e+00 1.34830368e+00 8.06520641e-01
-1.07157004e+00 -5.64540446e-01 -9.84751582e-01 1.76983684e-01
-8.91007781e-01 -2.09486619e-01 1.99469835e-01 1.02861536e+00
-9.48063552e-01 6.84724212e-01 -5.50189018e-01 -5.08930981e-01
-2.37404361e-01 4.30588573e-01 -1.23016953e-01 2.58908868e-01
-1.33275151e+00 1.33612013e+00 3.23708236e-01 -3.14060062e-01
-6.89423025e-01 -7.10261822e-01 -9.48908746e-01 1.40540361e-01
-2.59897649e-01 2.84376115e-01 1.51553309e+00 -1.18660414e+00
-2.24665380e+00 5.45595765e-01 1.75139159e-01 -4.28487927e-01
3.48368920e-02 -2.45351925e-01 -5.27326643e-01 1.78883269e-01
-1.37380004e-01 5.08743703e-01 2.34499246e-01 -8.46035957e-01
-8.67463529e-01 -2.45245337e-01 -3.13477308e-01 4.23532426e-01
-4.42263514e-01 2.91469902e-01 3.48329812e-01 -6.73772454e-01
-2.72616055e-02 -5.34610808e-01 3.14131916e-01 -7.29138196e-01
3.15333068e-01 -7.46149957e-01 9.47386861e-01 -9.39761043e-01
1.49139333e+00 -1.98584652e+00 -5.10351896e-01 2.21542180e-01
-4.24923390e-01 8.57095897e-01 2.50082929e-02 6.36369348e-01
-1.41989410e-01 4.78212163e-02 1.79952666e-01 3.23912442e-01
-1.08732514e-01 2.59249449e-01 -1.31427661e-01 -2.98059702e-01
2.22775891e-01 3.24502081e-01 -9.91599381e-01 -4.95396316e-01
2.48676926e-01 4.53935057e-01 -2.50896186e-01 5.60923934e-01
-4.38193455e-02 1.78443506e-01 -2.28682049e-02 5.22963285e-01
1.13577850e-01 4.89344150e-01 -1.08662061e-01 4.30924505e-01
-3.12743783e-01 7.86437809e-01 -1.16148734e+00 1.37972641e+00
-7.73999512e-01 1.01711321e+00 -2.90317759e-02 -1.21052921e+00
1.17007363e+00 8.95788670e-01 5.77712953e-02 -6.95130646e-01
1.86552703e-01 5.94946206e-01 3.38220626e-01 -1.06692648e+00
2.56537259e-01 -4.26034955e-03 2.59222269e-01 5.99719524e-01
6.06671810e-01 -4.57569629e-01 1.81458797e-02 2.78647896e-02
7.62411416e-01 1.97917297e-01 1.35207668e-01 -2.86317348e-01
8.51971269e-01 1.67073980e-02 1.13307111e-01 5.58764815e-01
-4.36366200e-01 -4.77300520e-04 9.00785811e-03 -2.25577876e-01
-3.90131980e-01 -6.99192286e-01 -9.70322788e-02 1.47778630e+00
-5.33269942e-01 2.01932099e-02 -1.21676981e+00 -5.85653961e-01
-2.91962326e-01 1.22275651e+00 -7.56683275e-02 -2.90630251e-01
-6.71004117e-01 -1.32805586e-01 6.36591017e-01 2.13099808e-01
5.30336976e-01 -1.62729549e+00 -5.45563281e-01 6.03848994e-01
-5.37442341e-02 -5.51130652e-01 -3.06499720e-01 5.49656510e-01
-8.57005298e-01 -7.03803778e-01 -3.31421167e-01 -1.58111060e+00
4.20252055e-01 -1.09595582e-01 9.00842845e-01 2.43616983e-01
-1.84281588e-01 6.43142939e-01 -4.80314016e-01 -1.12873614e+00
-8.95431459e-01 -5.54486178e-02 7.99133163e-03 -4.62655723e-01
8.77098024e-01 -4.96229619e-01 -1.17220022e-01 3.19250301e-03
-5.43709815e-01 -3.07897717e-01 4.75896299e-01 8.07916164e-01
-2.00790554e-01 2.91596979e-01 1.25240195e+00 -6.26882195e-01
1.11557209e+00 -4.18624103e-01 -2.81746536e-01 3.64148945e-01
-7.05574870e-01 9.83158052e-02 6.83756590e-01 -5.69292426e-01
-1.15513098e+00 -2.16638207e-01 -6.95017457e-01 3.32086086e-02
-6.59448743e-01 7.46401668e-01 -1.27330692e-02 -4.29970682e-01
8.24330330e-01 2.85150826e-01 4.08289760e-01 -2.75632769e-01
-1.98191494e-01 1.43560183e+00 3.59793186e-01 -6.45115316e-01
1.31643280e-01 -6.53310418e-01 -6.40812039e-01 -1.12061834e+00
-5.33223450e-01 -2.67144680e-01 -6.29728436e-01 -7.62529850e-01
5.64590156e-01 -6.76339209e-01 -1.22904575e+00 6.58033907e-01
-1.03986406e+00 -5.62231600e-01 -1.80884749e-01 7.57571042e-01
-3.32753360e-01 -9.92252901e-02 -8.95711601e-01 -1.22499430e+00
-3.29708248e-01 -1.19008756e+00 1.31865278e-01 7.76644111e-01
-3.84436220e-01 -1.19344938e+00 -1.44965230e-02 6.13873661e-01
5.60286582e-01 -1.54545516e-01 1.21850872e+00 -1.08299208e+00
1.09511316e-01 -5.86692095e-02 4.94860470e-01 6.75163865e-01
1.09828822e-01 1.24658778e-01 -9.40229177e-01 9.21024382e-02
1.96594760e-01 -6.33353531e-01 5.26986569e-02 1.08454861e-01
7.67235696e-01 -5.57012558e-01 2.96995789e-01 -1.49750784e-01
1.10157633e+00 1.07585895e+00 4.19126600e-01 5.60428083e-01
2.22913280e-01 9.96557117e-01 7.21593022e-01 -1.70921534e-01
4.94905889e-01 2.95492887e-01 -1.31307319e-02 3.10485005e-01
-2.31373515e-02 -1.94794908e-01 9.19142902e-01 1.43547499e+00
2.78738558e-01 1.71866372e-01 -1.22916126e+00 9.47846949e-01
-1.31689644e+00 -7.23959625e-01 -1.05785608e-01 2.21389723e+00
1.46664286e+00 -3.22726853e-02 1.82995245e-01 6.70985997e-01
4.83222663e-01 -6.99339032e-01 1.21843502e-01 -1.46954072e+00
4.21454817e-01 7.66167104e-01 1.26617588e-02 9.83056307e-01
-6.30398512e-01 1.01374936e+00 5.47859001e+00 8.90178502e-01
-1.35548747e+00 -9.88116637e-02 3.65276366e-01 5.69387972e-01
-1.90271959e-02 -4.96605486e-01 -7.40522146e-01 3.29315424e-01
1.62795281e+00 -1.33047789e-01 4.41857159e-01 5.97307026e-01
1.31584033e-01 -2.94050783e-01 -7.28036046e-01 4.07931447e-01
4.04176973e-02 -8.64687145e-01 -3.17672677e-02 -4.83118951e-01
4.94182527e-01 -3.71136248e-01 -1.73522800e-01 6.69291437e-01
4.32488292e-01 -1.22441089e+00 4.20814723e-01 2.65107125e-01
2.10046798e-01 -1.14403033e+00 8.78543317e-01 6.12864077e-01
-7.08563030e-01 -2.66988635e-01 3.45353596e-02 -3.25882852e-01
-7.07882702e-01 -2.76026666e-01 -1.48307240e+00 2.37282231e-01
7.64218450e-01 3.33065465e-02 -4.14476424e-01 1.04243982e+00
-2.61880994e-01 1.07351470e+00 -2.80418456e-01 -8.37956250e-01
2.36480847e-01 -1.71687946e-01 3.29324931e-01 1.21655655e+00
5.64015448e-01 4.98763651e-01 -1.30810052e-01 3.31112057e-01
4.06210572e-01 5.91617227e-01 -4.57260251e-01 -9.21916068e-02
6.36344314e-01 7.18100011e-01 -4.58549529e-01 -7.26970956e-02
-2.10856482e-01 5.93560219e-01 1.17647387e-01 1.76618233e-01
-1.84090868e-01 -1.09548378e+00 3.37925315e-01 -1.10222496e-01
-9.43440124e-02 1.28279686e-01 -4.25273061e-01 -6.14861071e-01
-4.72772121e-01 -1.41703057e+00 2.73339838e-01 -7.19750404e-01
-9.37006891e-01 5.24950981e-01 -2.13493213e-01 -8.77891064e-01
-4.77667809e-01 -6.68644369e-01 -9.93308365e-01 1.07834017e+00
-1.14561844e+00 -5.64822197e-01 -7.46068582e-02 5.35034418e-01
8.06873381e-01 -6.42550170e-01 1.31717491e+00 2.23038688e-01
-4.58682954e-01 7.30873525e-01 4.35273647e-02 2.97366470e-01
6.62516475e-01 -1.67197323e+00 -3.62401098e-01 4.88208443e-01
-8.78598392e-02 3.74080211e-01 7.77668834e-01 -4.01009679e-01
-1.20405185e+00 -5.56255698e-01 1.34906864e+00 -1.17171541e-01
5.16733706e-01 5.71233779e-02 -1.08025968e+00 4.62834775e-01
6.80204391e-01 -7.21156597e-01 1.00271797e+00 -3.44573677e-01
3.24235082e-01 -2.40891367e-01 -1.43515682e+00 5.59715211e-01
-1.01621225e-01 -4.13650006e-01 -1.11190128e+00 4.16726768e-01
7.56789565e-01 -3.73271286e-01 -1.10499084e+00 2.14573860e-01
4.00746912e-01 -8.16573501e-01 6.57762408e-01 -6.47564948e-01
2.31052682e-01 7.71425590e-02 2.98901320e-01 -1.60823154e+00
4.54728901e-01 -7.39117265e-01 2.44996190e-01 1.43990600e+00
5.54532230e-01 -6.60131037e-01 4.82940495e-01 4.26372737e-01
-2.54712760e-01 -7.52570391e-01 -5.44182062e-01 -3.15537035e-01
5.04799128e-01 -4.80669945e-01 3.14594567e-01 1.29427826e+00
5.81907630e-01 2.67985225e-01 2.00185135e-01 4.77702916e-02
-1.05151925e-02 -4.21595722e-01 3.44581634e-01 -1.20962179e+00
1.82419438e-02 -4.66710240e-01 -2.66024649e-01 -6.98438764e-01
-3.74568962e-02 -4.80538577e-01 3.16201419e-01 -1.14572704e+00
-8.64357471e-01 -3.55251372e-01 -7.88580999e-02 5.49554467e-01
-2.21955642e-01 -3.30972373e-01 9.06862989e-02 -4.45298761e-01
6.21629655e-02 2.29315639e-01 1.03773975e+00 2.15488493e-01
-6.29498839e-01 3.24314773e-01 -3.44644248e-01 8.11084270e-01
1.07508802e+00 -4.49948251e-01 -3.26949090e-01 -7.44868889e-02
-9.32868719e-02 3.88835669e-01 -3.66659045e-01 -1.12571633e+00
3.69172901e-01 -1.29030481e-01 3.39484632e-01 -3.48827392e-01
-1.58319220e-01 -9.91913021e-01 -5.93364716e-01 7.97375500e-01
-6.26545608e-01 2.18429163e-01 6.83237791e-01 -3.25167507e-01
-5.32449305e-01 -9.14968967e-01 9.63573933e-01 -1.44541696e-01
-4.24963295e-01 -5.37912846e-01 -1.16176605e+00 -9.46170390e-02
1.11058378e+00 -1.25566080e-01 -1.95715055e-01 -5.97983003e-01
-6.68717742e-01 4.74450678e-01 -3.03572208e-01 4.22668070e-01
6.54682100e-01 -1.04371893e+00 -5.85987568e-01 5.10985076e-01
-3.65552992e-01 -1.28604636e-01 -8.56292322e-02 6.43384695e-01
-1.15358961e+00 5.67834914e-01 -5.65655470e-01 -3.14646691e-01
-1.65667033e+00 6.58275932e-02 5.57124019e-01 -1.35120004e-01
-7.54789338e-02 9.92722154e-01 -6.68197870e-01 -1.03411257e+00
8.65231633e-01 -2.18219131e-01 -1.08587945e+00 2.69168049e-01
7.39142537e-01 2.99534857e-01 2.79326260e-01 -3.82074684e-01
7.49617517e-02 2.11365223e-01 1.09479897e-01 -3.10866266e-01
1.22757947e+00 -1.27078295e-01 -8.87041613e-02 9.30822849e-01
7.95056522e-01 1.31650776e-01 -2.80766636e-01 5.21374345e-02
4.20464545e-01 5.72106354e-02 -7.52847791e-02 -1.38947237e+00
-5.80126762e-01 1.16402805e+00 9.55651760e-01 5.52500129e-01
1.12372839e+00 -6.11558378e-01 7.39529133e-01 5.91908574e-01
-8.78054574e-02 -1.41421175e+00 1.24205507e-01 1.18065655e+00
7.52139688e-01 -1.13391840e+00 -6.69501722e-01 -4.81774807e-02
-5.27997971e-01 1.64182258e+00 1.00415492e+00 1.55154064e-01
8.31929207e-01 1.30717903e-01 7.14150548e-01 1.13473840e-01
-7.45393574e-01 9.12867114e-03 2.45413005e-01 6.24146700e-01
1.14976537e+00 -5.55667840e-03 -4.68585342e-01 5.42623937e-01
-6.72071218e-01 -1.06360443e-01 6.76602900e-01 1.21286750e+00
-1.01743734e+00 -1.15538669e+00 -9.12421942e-01 2.29150802e-01
-6.81029618e-01 -1.78359404e-01 -5.53378165e-01 6.22388482e-01
6.85923323e-02 1.31167817e+00 -2.44695395e-02 -3.92082602e-01
3.66626292e-01 8.01979005e-01 3.01329583e-01 -6.74501479e-01
-1.45236969e+00 -1.48833498e-01 6.24481589e-04 4.02096845e-02
-1.55709177e-01 -5.14624476e-01 -1.64536440e+00 -2.21738711e-01
-4.21418667e-01 8.26133609e-01 1.05069518e+00 1.19288003e+00
-5.61704002e-02 7.56120861e-01 5.83712578e-01 -2.15691105e-01
-5.07517815e-01 -1.56849992e+00 -2.18366444e-01 -1.42081529e-01
2.82997549e-01 -1.89447865e-01 -2.84895182e-01 -9.12902653e-02] | [11.85006332397461, 8.391040802001953] |
e6876080-18b7-474d-8187-b0cb49517b60 | dynamic-mixed-membership-stochastic-block | 2304.05894 | null | https://arxiv.org/abs/2304.05894v1 | https://arxiv.org/pdf/2304.05894v1.pdf | Dynamic Mixed Membership Stochastic Block Model for Weighted Labeled Networks | Most real-world networks evolve over time. Existing literature proposes models for dynamic networks that are either unlabeled or assumed to have a single membership structure. On the other hand, a new family of Mixed Membership Stochastic Block Models (MMSBM) allows to model static labeled networks under the assumption of mixed-membership clustering. In this work, we propose to extend this later class of models to infer dynamic labeled networks under a mixed membership assumption. Our approach takes the form of a temporal prior on the model's parameters. It relies on the single assumption that dynamics are not abrupt. We show that our method significantly differs from existing approaches, and allows to model more complex systems --dynamic labeled networks. We demonstrate the robustness of our method with several experiments on both synthetic and real-world datasets. A key interest of our approach is that it needs very few training data to yield good results. The performance gain under challenging conditions broadens the variety of possible applications of automated learning tools --as in social sciences, which comprise many fields where small datasets are a major obstacle to the introduction of machine learning methods. | ['Sabine Loudcher', 'Julien Velcin', 'Gaël Poux-Médard'] | 2023-04-12 | null | null | null | null | ['stochastic-block-model'] | ['graphs'] | [ 1.80334181e-01 2.74847955e-01 -5.41290164e-01 -2.98794150e-01
1.29936844e-01 -7.08822489e-01 9.85766053e-01 -5.75779472e-03
-2.73646444e-01 9.03155744e-01 -4.73373353e-01 -5.69804192e-01
-4.94279265e-01 -8.02707434e-01 -7.23054647e-01 -8.97890389e-01
-3.95013422e-01 8.92855644e-01 7.02149928e-01 -2.38996625e-01
-1.28206119e-01 4.03773516e-01 -1.29854631e+00 -1.35100171e-01
7.62338996e-01 3.85575503e-01 -1.00448348e-01 5.46827674e-01
7.44816440e-04 8.75434697e-01 -2.97789901e-01 -3.23878706e-01
3.55183244e-01 -3.36917043e-01 -9.88888144e-01 2.51606941e-01
-5.85063174e-02 2.33453721e-01 -4.54153597e-01 1.05483365e+00
-3.92389372e-02 2.26770695e-02 9.57300425e-01 -1.75857282e+00
-2.24415749e-01 9.63672519e-01 -6.02244020e-01 1.93662569e-01
-1.66257232e-01 -7.03856945e-02 8.74722660e-01 -2.19767511e-01
8.22895765e-01 1.32171059e+00 9.39578056e-01 5.80867350e-01
-1.68348885e+00 -5.13437569e-01 6.02031171e-01 3.34267110e-01
-1.43488812e+00 -3.48626584e-01 7.34393239e-01 -7.15254426e-01
3.90953928e-01 9.86783728e-02 4.89601642e-01 9.98893559e-01
-1.70111448e-01 6.32041454e-01 1.42331731e+00 -4.57353204e-01
4.10678416e-01 1.58977583e-01 5.76085269e-01 5.70276916e-01
5.20991683e-01 -1.05796479e-01 5.51612750e-02 -2.90097028e-01
7.25053489e-01 9.96307433e-02 -5.30367345e-02 -7.89645374e-01
-1.19254696e+00 8.39917183e-01 8.44832957e-02 6.09821677e-01
-1.00970559e-01 2.01855525e-01 1.58766448e-01 5.40028691e-01
6.84641302e-01 3.41286622e-02 -4.80167896e-01 -5.19904904e-02
-9.65720773e-01 5.33226132e-02 1.31423342e+00 7.63361931e-01
8.60711873e-01 -1.87885389e-01 3.32296550e-01 6.43208981e-01
3.42432439e-01 2.10253581e-01 1.81416348e-01 -9.99815941e-01
1.50693849e-01 5.04660964e-01 2.62254447e-01 -8.78973484e-01
-3.91287923e-01 -3.56414735e-01 -9.52188730e-01 -5.32132247e-03
8.94418776e-01 -1.89701200e-01 -9.93399143e-01 2.17279696e+00
4.66153890e-01 6.88340485e-01 2.68929498e-03 2.94927061e-01
3.31066608e-01 5.37447035e-01 -2.01000482e-01 -7.02524841e-01
7.15967059e-01 -8.83012950e-01 -7.47966886e-01 1.18902996e-01
5.62789381e-01 -1.54642195e-01 6.53826356e-01 4.57388997e-01
-9.15896356e-01 -1.16179176e-01 -7.54411876e-01 6.38269126e-01
-5.27801931e-01 -4.68074262e-01 6.34839058e-01 7.96119869e-01
-1.42645836e+00 8.28607917e-01 -1.15755546e+00 -8.27168643e-01
2.02777520e-01 6.52561903e-01 -9.31750983e-02 -5.50584234e-02
-1.12339365e+00 7.92361677e-01 3.95180911e-01 3.27713974e-02
-7.98622489e-01 -3.76373202e-01 -4.83051509e-01 -2.91072726e-02
7.78244317e-01 -6.12315297e-01 1.34105003e+00 -1.21199846e+00
-1.44144940e+00 6.08765364e-01 -2.16311350e-01 -5.95220149e-01
8.58642876e-01 4.65640426e-01 -2.76670188e-01 1.22752115e-01
-1.46069467e-01 3.52203846e-01 7.54282475e-01 -1.59924316e+00
-3.82146657e-01 -1.91206783e-01 2.24111900e-01 -2.23713323e-01
-5.14509201e-01 -1.17789887e-01 -2.79811442e-01 -4.99822557e-01
3.40483747e-02 -1.32326138e+00 -5.27956963e-01 -2.86467820e-01
-3.91286522e-01 -4.22342032e-01 1.10196567e+00 -1.56166339e-02
1.27358997e+00 -1.89652371e+00 3.27013284e-01 5.19904971e-01
3.27117622e-01 1.32290408e-01 1.07534319e-01 7.46306658e-01
-2.23101914e-01 5.24722278e-01 -4.70433444e-01 -3.86253774e-01
-4.62444313e-03 6.87664807e-01 -1.20817460e-01 6.08585954e-01
-5.03428802e-02 7.70692289e-01 -9.56204951e-01 -5.24334431e-01
1.70837194e-02 1.90107465e-01 -1.70514971e-01 -2.85075039e-01
-3.47804517e-01 4.03397858e-01 -2.05618054e-01 3.45132291e-01
4.54393595e-01 -7.08036482e-01 5.98620057e-01 3.27616602e-01
-1.27299717e-02 -1.71631962e-01 -1.52706933e+00 1.12794197e+00
-8.27086046e-02 5.44841707e-01 3.20139766e-01 -1.46379626e+00
3.82385194e-01 4.04670089e-01 6.61973000e-01 2.32156396e-01
2.27628142e-01 5.70331402e-02 4.73099738e-01 -2.42537871e-01
-9.11763161e-02 -2.83258051e-01 6.45515621e-02 7.49662519e-01
-5.96950650e-02 8.89794976e-02 6.41462326e-01 4.86311585e-01
1.18631971e+00 -2.17675194e-01 3.23222190e-01 -6.37413204e-01
4.20511752e-01 -1.06858529e-01 7.43666351e-01 9.95787263e-01
-2.45119557e-01 1.04659058e-01 8.47517550e-01 -2.17120945e-01
-1.01644969e+00 -9.02310729e-01 -5.59025183e-02 8.44282329e-01
1.68916821e-01 -1.63003325e-01 -7.19107628e-01 -7.46231079e-01
7.04379827e-02 2.97012478e-01 -8.57861757e-01 7.09796846e-02
-4.89406407e-01 -1.13865829e+00 3.34889084e-01 2.54425347e-01
4.27432433e-02 -8.22118163e-01 1.03323117e-01 1.48538023e-01
-1.70248494e-01 -1.03360701e+00 -4.24383581e-03 8.68333727e-02
-1.08115721e+00 -1.25882614e+00 -4.88088489e-01 -7.56642222e-01
8.44801784e-01 1.77218989e-01 9.84578729e-01 1.03005402e-01
-7.34176859e-02 5.83662748e-01 -2.96835005e-01 -3.05668741e-01
-7.96810269e-01 3.91732991e-01 3.31388265e-01 3.80023986e-01
1.69895023e-01 -9.41546202e-01 -2.71419257e-01 5.38632452e-01
-1.15686727e+00 -3.45013551e-02 3.16541106e-01 6.67879760e-01
2.01184824e-01 5.49642980e-01 8.79915655e-01 -1.35161519e+00
4.29472476e-01 -8.72954309e-01 -6.68197513e-01 3.67793322e-01
-9.26181734e-01 -3.85959726e-03 6.62119746e-01 -9.17960346e-01
-9.16401505e-01 -1.42816594e-03 4.93779302e-01 -3.60368818e-01
-1.61926925e-01 6.91331625e-01 -1.02267928e-01 2.16547139e-02
4.08683151e-01 -3.83061878e-02 1.33366644e-01 -4.05351996e-01
3.29825968e-01 4.95088756e-01 2.75143832e-01 -5.95096409e-01
1.14263654e+00 8.53465736e-01 2.53532052e-01 -8.62156212e-01
-6.42035842e-01 -4.85568404e-01 -8.93289089e-01 -4.95250732e-01
3.44260544e-01 -6.92146361e-01 -7.30252087e-01 6.84158921e-01
-6.31944418e-01 -8.39399040e-01 -2.63469398e-01 2.62010306e-01
-5.31526327e-01 6.27451599e-01 -9.41131175e-01 -1.02169049e+00
2.31159627e-01 -7.63061106e-01 3.12899411e-01 1.16849095e-01
-1.95685476e-01 -1.55510402e+00 2.51124233e-01 5.84622622e-02
2.85615653e-01 2.53605545e-01 9.28331792e-01 -7.60223925e-01
-5.44543266e-01 -1.72314346e-01 6.66839853e-02 1.82114273e-01
2.91033208e-01 6.30074561e-01 -8.19825530e-01 -4.99595284e-01
-2.29464769e-01 1.92227948e-03 9.46691811e-01 4.29138869e-01
1.00777459e+00 -2.21669719e-01 -6.86348140e-01 7.46356621e-02
1.34528244e+00 1.97978646e-01 3.07820112e-01 1.49476439e-01
6.01022422e-01 8.49245667e-01 4.27379310e-01 2.46572688e-01
3.97893757e-01 5.06466269e-01 5.57800829e-01 -9.32436138e-02
3.29034954e-01 -1.06007606e-02 2.78822213e-01 1.14834666e+00
-1.48361400e-01 -4.45733219e-01 -1.18448424e+00 8.54644239e-01
-2.36851597e+00 -1.19665313e+00 -3.54658991e-01 2.14828897e+00
8.80055904e-01 3.00879806e-01 4.49425161e-01 3.30015600e-01
1.11298227e+00 -1.33169210e-02 -5.17489612e-01 2.78922729e-02
-5.33048324e-02 -5.72037362e-02 5.07452309e-01 7.02645123e-01
-1.21100247e+00 8.87481570e-01 6.60447454e+00 6.80456161e-01
-8.30872297e-01 2.38160640e-01 5.10629296e-01 8.38836953e-02
-1.34658232e-01 3.16010058e-01 -7.20651865e-01 4.86131340e-01
1.23209894e+00 -2.83267349e-01 2.91453719e-01 5.79418957e-01
3.96760464e-01 -1.01625569e-01 -1.15755022e+00 3.88905168e-01
-2.13594735e-01 -1.01481974e+00 -3.45019549e-01 4.03422385e-01
1.08144188e+00 1.95012558e-02 -3.40842158e-02 1.34725675e-01
1.05169666e+00 -7.80875266e-01 5.74042320e-01 5.62539160e-01
5.81845164e-01 -6.88165724e-01 4.99073565e-01 7.68457115e-01
-1.09929848e+00 -2.09302038e-01 -4.14556265e-02 -2.70275652e-01
1.28433272e-01 6.72780991e-01 -8.50383043e-01 5.32068729e-01
4.47135091e-01 1.02863455e+00 -3.99455369e-01 1.14462960e+00
-8.06627125e-02 1.16928005e+00 -5.99579573e-01 3.91592719e-02
1.50020167e-01 -3.88312340e-01 4.54689860e-01 1.09462309e+00
-5.10329148e-03 -1.89922914e-01 4.04531270e-01 5.00005901e-01
6.91915751e-02 -7.04754665e-02 -7.39983499e-01 -7.47864470e-02
4.43117231e-01 1.25022209e+00 -1.40493059e+00 -3.86950463e-01
-4.07503843e-01 6.27903819e-01 2.62142807e-01 5.29237807e-01
-7.76134729e-01 -5.99013381e-02 3.14486116e-01 3.58713925e-01
3.52602541e-01 -3.81453365e-01 9.20101702e-02 -1.38633025e+00
-2.79525369e-01 -6.73228920e-01 4.73203033e-01 -2.11333409e-01
-1.58099973e+00 2.27644011e-01 3.91410738e-01 -9.54340041e-01
-4.14670378e-01 -4.16941345e-01 -5.36285579e-01 3.74506116e-01
-1.39685690e+00 -1.21622860e+00 9.74963792e-03 7.05175877e-01
3.18453193e-01 1.84545323e-01 4.39556152e-01 3.08653653e-01
-8.51188302e-01 1.26148358e-01 4.70721811e-01 9.88427997e-02
6.35483325e-01 -1.43974590e+00 2.14356467e-01 1.00181293e+00
1.70302495e-01 5.72607517e-01 8.77726495e-01 -6.82116389e-01
-8.66135180e-01 -1.22347593e+00 8.41547072e-01 -5.19037247e-01
1.05343986e+00 -6.01136982e-01 -9.76281404e-01 1.01118076e+00
4.84272540e-02 1.32686663e-02 7.49994218e-01 2.22013459e-01
-2.40556315e-01 -3.84634845e-02 -9.25062716e-01 5.79054356e-01
1.09196901e+00 -1.47803441e-01 -3.36341202e-01 4.58215594e-01
6.30598605e-01 3.27443957e-01 -9.20278192e-01 5.56552827e-01
2.55476296e-01 -6.47189021e-01 7.15788424e-01 -7.39634633e-01
6.65437803e-02 -4.10030127e-01 1.99579284e-01 -1.18779719e+00
-3.60053480e-01 -7.89407909e-01 -4.64186430e-01 1.39598191e+00
4.87340331e-01 -8.58464301e-01 9.04999077e-01 5.09365439e-01
3.75944495e-01 -4.46187019e-01 -8.69582117e-01 -1.14087570e+00
2.13718385e-01 -2.73226440e-01 2.94206560e-01 1.34360480e+00
7.65055045e-02 1.77382827e-01 -2.97802836e-01 3.84306051e-02
7.62127042e-01 7.82042667e-02 6.83281541e-01 -1.94634819e+00
-4.03287798e-01 -5.15141726e-01 -2.74081349e-01 -7.71587133e-01
5.61944723e-01 -6.46809220e-01 -7.27724209e-02 -1.35354555e+00
4.16235298e-01 -8.38709354e-01 -2.81960785e-01 4.90312397e-01
-1.64925791e-02 2.27647692e-01 5.15318029e-02 5.11005640e-01
-7.45201290e-01 3.02706867e-01 7.83093631e-01 -5.43520693e-03
-2.91485637e-01 5.03848910e-01 -5.40829659e-01 9.15560603e-01
9.23190296e-01 -6.31719112e-01 -7.88853645e-01 1.46495417e-01
3.53361607e-01 -1.76723469e-02 2.34258980e-01 -6.91931427e-01
3.46866310e-01 -4.21363354e-01 -2.60778517e-01 -4.27498758e-01
2.49646887e-01 -9.80666041e-01 4.81860876e-01 6.82530880e-01
-1.36911765e-01 2.93418616e-02 -1.93145961e-01 1.06987393e+00
-5.85026667e-02 -2.80579388e-01 8.03660810e-01 -2.70753920e-01
-4.67430145e-01 3.99698883e-01 -5.71629107e-01 9.01291817e-02
1.42887700e+00 -1.49049431e-01 -1.47779584e-01 -6.83028340e-01
-1.15087235e+00 3.72730494e-01 7.53761351e-01 2.10048780e-01
-4.80831675e-02 -9.05747712e-01 -5.16141474e-01 -2.24491760e-01
-2.91857809e-01 -5.14952466e-02 8.29062611e-02 1.02722692e+00
-3.20018560e-01 2.23893344e-01 1.90119430e-01 -8.87723088e-01
-1.37621164e+00 8.00047457e-01 3.87817025e-01 -4.95367527e-01
-3.16337109e-01 3.96211445e-01 6.75824331e-03 -6.03064477e-01
1.15243554e-01 -7.26871984e-03 -2.15287179e-01 2.85155505e-01
8.97052363e-02 5.26802003e-01 -2.60997683e-01 -6.18965089e-01
-1.74395725e-01 2.58865923e-01 -2.29285702e-01 -2.61759639e-01
1.47877967e+00 -3.26102465e-01 -4.44080651e-01 1.04363143e+00
7.73738444e-01 -2.19875202e-01 -1.15928662e+00 -5.79433262e-01
3.10160935e-01 -2.51100864e-02 -4.25287515e-01 -4.33758080e-01
-9.61796224e-01 6.85337722e-01 1.42310157e-01 8.11568141e-01
9.08760250e-01 2.24355474e-01 4.36749384e-02 3.70795459e-01
4.65587705e-01 -9.92324889e-01 -7.50295818e-02 4.86094713e-01
1.23472966e-01 -1.09248376e+00 -1.07033089e-01 -7.29587197e-01
-1.26502067e-01 8.71309459e-01 5.19545674e-01 -6.27378151e-02
1.09545577e+00 2.47089088e-01 -1.40033066e-01 -1.81234162e-02
-9.87597466e-01 -2.64907002e-01 -2.74822295e-01 4.75704581e-01
4.34719808e-02 9.77728888e-02 -3.41408819e-01 2.55464494e-01
4.42552604e-02 -2.93374248e-02 9.92352724e-01 1.00370276e+00
-3.19215208e-01 -1.44197655e+00 -2.71613002e-01 5.01820564e-01
-5.16547263e-01 1.64466441e-01 -5.83400965e-01 1.13444543e+00
1.53527027e-02 1.10459423e+00 -1.09105716e-02 -1.01053514e-01
-1.07895784e-01 2.84613460e-01 3.60501945e-01 -7.16909409e-01
-3.64594340e-01 1.32087156e-01 -1.07874230e-01 6.00820631e-02
-1.05260706e+00 -1.06198645e+00 -7.89063215e-01 -5.98074853e-01
-7.04378307e-01 3.18013757e-01 3.61496866e-01 9.65111256e-01
2.30724774e-02 3.20383608e-01 6.57496214e-01 -5.68817139e-01
-5.03794611e-01 -1.03429317e+00 -8.27810585e-01 4.33179975e-01
1.94476545e-01 -8.40364158e-01 -5.98458290e-01 3.64774108e-01] | [6.90802001953125, 5.190961837768555] |
3b5b042b-ca82-4819-a998-9cf5788ac290 | upc-s-speech-translation-system-for-iwslt | 2105.04512 | null | https://arxiv.org/abs/2105.04512v2 | https://arxiv.org/pdf/2105.04512v2.pdf | End-to-End Speech Translation with Pre-trained Models and Adapters: UPC at IWSLT 2021 | This paper describes the submission to the IWSLT 2021 offline speech translation task by the UPC Machine Translation group. The task consists of building a system capable of translating English audio recordings extracted from TED talks into German text. Submitted systems can be either cascade or end-to-end and use a custom or given segmentation. Our submission is an end-to-end speech translation system, which combines pre-trained models (Wav2Vec 2.0 and mBART) with coupling modules between the encoder and decoder, and uses an efficient fine-tuning technique, which trains only 20% of its total parameters. We show that adding an Adapter to the system and pre-training it, can increase the convergence speed and the final result, with which we achieve a BLEU score of 27.3 on the MuST-C test set. Our final model is an ensemble that obtains 28.22 BLEU score on the same set. Our submission also uses a custom segmentation algorithm that employs pre-trained Wav2Vec 2.0 for identifying periods of untranscribable text and can bring improvements of 2.5 to 3 BLEU score on the IWSLT 2019 test set, as compared to the result with the given segmentation. | ['Marta R. Costa-jussà', 'José A. R. Fonollosa', 'Carlos Escolano', 'Ioannis Tsiamas', 'Gerard I. Gállego'] | 2021-05-10 | null | https://aclanthology.org/2021.iwslt-1.11 | https://aclanthology.org/2021.iwslt-1.11.pdf | acl-iwslt-2021-8 | ['speech-to-text-translation'] | ['natural-language-processing'] | [ 7.28270337e-02 3.07616532e-01 2.14948416e-01 -3.15185875e-01
-1.67845476e+00 -8.33934128e-01 5.89763582e-01 -2.71775067e-01
-4.20111150e-01 5.97863615e-01 3.53331834e-01 -9.01027143e-01
6.01003051e-01 -2.57349461e-01 -7.83584058e-01 -3.46108288e-01
2.39573345e-01 9.32881117e-01 2.00337470e-01 -5.45884907e-01
-3.44473362e-01 -6.31622002e-02 -6.82236373e-01 5.60430527e-01
4.78236318e-01 8.23091269e-01 1.58121675e-01 1.17808080e+00
3.98882031e-02 2.75460273e-01 -8.43740344e-01 -6.04009688e-01
2.79508144e-01 -8.16406786e-01 -1.16956413e+00 -1.00605942e-01
2.14862868e-01 -1.74851492e-01 -4.35009331e-01 5.69070339e-01
9.62947726e-01 -1.68768004e-01 4.55855906e-01 -7.54173458e-01
-3.61919433e-01 1.15157163e+00 -4.99858009e-03 2.76611090e-01
2.90751636e-01 2.48604845e-02 9.84957993e-01 -1.06147337e+00
5.89720130e-01 9.50481534e-01 5.39925814e-01 7.64819026e-01
-1.18560910e+00 -5.55974126e-01 -4.00947481e-01 2.51256168e-01
-1.34910226e+00 -9.01368976e-01 3.21446896e-01 -2.60735333e-01
1.72577763e+00 4.39172179e-01 4.50128257e-01 1.28120017e+00
1.69617519e-01 6.91847563e-01 8.06102097e-01 -6.52218401e-01
1.14878677e-01 3.30010921e-01 -2.32601479e-01 3.16426933e-01
-6.18936718e-01 -3.49405147e-02 -5.35371661e-01 1.65602699e-01
3.42018038e-01 -7.69495308e-01 -2.07828447e-01 4.11828279e-01
-1.40087724e+00 6.16690040e-01 6.73440546e-02 4.12969649e-01
-1.84359074e-01 5.24169803e-02 7.12792099e-01 8.54265213e-01
7.54828453e-01 5.20757914e-01 -6.80048645e-01 -5.65601587e-01
-1.38755000e+00 -1.77447140e-01 9.36404765e-01 1.17873907e+00
4.10789132e-01 2.80335605e-01 -4.54562455e-01 1.06742525e+00
4.35842536e-02 7.40865231e-01 6.86607420e-01 -7.65373528e-01
8.52376401e-01 -1.53890857e-02 -1.12065189e-01 -7.91312680e-02
-1.41293034e-01 -6.94552720e-01 -3.87980580e-01 -3.30039382e-01
2.69903261e-02 -6.98044717e-01 -1.17007971e+00 1.62201691e+00
-2.59605180e-02 -1.56758994e-01 3.39228749e-01 7.49736547e-01
6.95799589e-01 1.27323508e+00 -3.51433337e-01 -3.56163114e-01
1.06699336e+00 -1.39266205e+00 -8.87439430e-01 -2.09090456e-01
7.75592089e-01 -1.47977376e+00 1.08552206e+00 2.78475374e-01
-1.47389185e+00 -6.89999640e-01 -1.14441621e+00 -6.17968887e-02
-1.94605708e-01 3.17549318e-01 -1.95121184e-01 6.52740955e-01
-1.55625844e+00 4.30248410e-01 -7.81711161e-01 -4.56112832e-01
-2.50057906e-01 5.41252792e-01 -2.08431631e-01 3.10427278e-01
-1.36588120e+00 1.04387617e+00 1.77939132e-01 -1.70457467e-01
-8.23692918e-01 -4.79343086e-01 -5.48336923e-01 6.20184056e-02
1.16284452e-02 -6.16619706e-01 1.82361627e+00 -1.03860998e+00
-2.05203032e+00 7.15024233e-01 -2.60780126e-01 -7.89673269e-01
6.57604873e-01 -1.98146328e-02 -6.57355309e-01 8.96826908e-02
5.23058623e-02 6.70641899e-01 6.34822130e-01 -7.45674729e-01
-5.67016244e-01 1.16085149e-01 -4.31169599e-01 3.06372046e-01
-1.29087061e-01 4.02233154e-01 -7.84464121e-01 -6.30566597e-01
-1.53109998e-01 -1.10283101e+00 8.15536827e-03 -8.24978888e-01
-4.52317089e-01 6.05035573e-02 8.67926300e-01 -1.32166922e+00
1.35227430e+00 -2.02885294e+00 2.98841953e-01 9.01420861e-02
-3.23208988e-01 3.85677844e-01 -3.26442182e-01 7.39801168e-01
6.16060710e-03 1.46942049e-01 -2.26811379e-01 -7.52435684e-01
-9.16363746e-02 1.52400732e-01 -3.67972672e-01 9.99215096e-02
1.92231104e-01 9.01834846e-01 -5.97213328e-01 -2.22444460e-01
1.62059560e-01 4.48501170e-01 -3.75631899e-01 4.61793512e-01
9.62681603e-03 3.13194335e-01 6.85136840e-02 1.77621901e-01
1.93648443e-01 2.34653026e-01 1.16453461e-01 2.00151831e-01
-1.73597306e-01 9.81932282e-01 -5.64666629e-01 1.90383542e+00
-8.46299589e-01 1.09002507e+00 3.76137570e-02 -8.23522270e-01
9.51458514e-01 9.76213992e-01 1.60868838e-01 -6.66562915e-01
3.04475427e-01 6.58747613e-01 1.28645137e-01 -3.09851825e-01
4.52964455e-01 -1.01385564e-01 -2.00652257e-01 4.96661931e-01
6.89123690e-01 -6.58488870e-01 4.22561914e-02 2.51844049e-01
1.18677974e+00 -8.70018303e-02 -1.49156407e-01 -2.77345866e-01
4.28730607e-01 4.77035865e-02 2.40908191e-01 4.22234386e-01
5.20911161e-03 1.01028931e+00 2.30376303e-01 -8.28947276e-02
-1.63648057e+00 -9.96387959e-01 1.09519564e-01 9.14511085e-01
-5.61568558e-01 -5.21687388e-01 -1.20304060e+00 -5.45808077e-01
-6.20342374e-01 8.99050772e-01 -2.83370674e-01 -8.33116993e-02
-7.05320954e-01 -3.65335375e-01 9.69533086e-01 3.75880837e-01
3.14120948e-01 -8.40742469e-01 3.59209143e-02 6.36799753e-01
-7.96678603e-01 -1.28980386e+00 -9.50694263e-01 3.17061901e-01
-5.31254172e-01 -2.14864135e-01 -9.18300688e-01 -9.95885074e-01
7.15300366e-02 -2.84273247e-03 1.13768423e+00 -3.21669757e-01
9.48888734e-02 8.21830407e-02 -6.41505122e-01 -3.51294070e-01
-1.04162884e+00 5.81050098e-01 2.23161563e-01 -6.96486086e-02
2.80013382e-01 -4.18836743e-01 -1.77401289e-01 3.79038036e-01
-3.08977425e-01 2.34006360e-01 4.33448702e-01 8.80124986e-01
4.64662254e-01 -6.27255499e-01 6.05272710e-01 -5.11760473e-01
5.92922926e-01 -1.79648310e-01 -3.38541836e-01 2.78813362e-01
-7.05471933e-01 -6.43511023e-03 7.40899801e-01 -3.28122288e-01
-6.56738877e-01 -1.76384707e-03 -7.00673699e-01 -3.97330552e-01
2.43790299e-01 3.35945934e-01 -1.29424199e-01 2.28932753e-01
5.97904980e-01 4.08082992e-01 -8.97191018e-02 -4.84088719e-01
3.83350849e-01 1.45559001e+00 6.53976262e-01 -1.78078964e-01
6.30478323e-01 -4.10916418e-01 -7.59187818e-01 -6.64119303e-01
-4.49417889e-01 -5.60492933e-01 -5.28263032e-01 -1.83428541e-01
8.37867677e-01 -1.15529859e+00 3.61587442e-02 3.30847591e-01
-1.42174351e+00 -7.66730964e-01 -1.96970582e-01 6.21095181e-01
-7.91252851e-01 1.13316134e-01 -9.30117607e-01 -4.24360305e-01
-8.52138579e-01 -1.34410107e+00 1.07458615e+00 -2.90486366e-01
-5.32694042e-01 -8.47744524e-01 2.08023012e-01 5.83873391e-01
6.10764682e-01 -3.63167673e-01 5.65372050e-01 -7.97659636e-01
-6.97155222e-02 -1.84032574e-01 1.74778387e-01 7.48046875e-01
-1.96231566e-02 -1.23051265e-02 -1.16962278e+00 -4.23504442e-01
-1.77570567e-01 -2.13298216e-01 6.71128035e-01 2.16696456e-01
5.43377519e-01 -3.34017366e-01 5.13911508e-02 4.22342211e-01
1.00072193e+00 3.94860268e-01 6.95221066e-01 1.28218099e-01
3.50411981e-01 3.53661180e-01 3.24068040e-01 1.09182715e-01
1.71296284e-01 1.11303830e+00 -5.87440655e-02 -1.25414357e-01
-5.67676902e-01 -3.14292580e-01 9.09033716e-01 1.81408679e+00
5.66213802e-02 -5.35394371e-01 -9.90970433e-01 7.22925842e-01
-1.54806459e+00 -7.67542541e-01 -1.75406739e-01 2.08500385e+00
1.22560883e+00 2.11422220e-01 1.69726133e-01 2.72832274e-01
6.05525076e-01 -1.83065578e-01 -3.16446126e-02 -1.17425156e+00
-7.42005371e-03 6.24414980e-01 6.91255093e-01 1.08686268e+00
-7.59127796e-01 1.41101038e+00 6.88469505e+00 1.02651298e+00
-1.33155382e+00 6.46446109e-01 4.61547017e-01 -1.94471627e-01
-2.50875026e-01 -1.89102471e-01 -6.94323301e-01 4.07458156e-01
2.00761414e+00 -3.78470540e-01 7.01956153e-01 3.75494242e-01
4.61986363e-01 5.42436302e-01 -1.01438141e+00 7.59149253e-01
8.08354840e-02 -1.27727163e+00 -1.47146165e-01 -1.40865743e-02
7.86978960e-01 5.62098324e-01 -5.17175831e-02 6.21098578e-01
2.24693626e-01 -8.43659997e-01 1.06110322e+00 -6.49309456e-02
1.27235484e+00 -8.47191870e-01 9.90170002e-01 2.06854567e-01
-8.55664134e-01 3.89200926e-01 -2.11600721e-01 9.29260850e-02
4.25090075e-01 3.78002524e-01 -1.49811065e+00 6.75126016e-01
4.47722048e-01 3.53274137e-01 -2.81421363e-01 7.35456586e-01
-2.58134723e-01 1.37489367e+00 -3.88102829e-01 -1.02504052e-01
4.77988958e-01 1.19221918e-01 5.78541398e-01 1.69731081e+00
6.06001079e-01 -2.76742697e-01 -9.24583599e-02 3.38364661e-01
-2.81428754e-01 2.85005778e-01 -3.28471482e-01 3.98764387e-02
4.42095369e-01 9.21318233e-01 -2.85828590e-01 -6.26285493e-01
-2.32020617e-01 1.49819326e+00 2.85497177e-02 4.28641081e-01
-9.58464265e-01 -6.08141363e-01 4.83413219e-01 -7.42510632e-02
4.14670199e-01 -2.76378155e-01 -3.22791964e-01 -1.06945097e+00
2.17531640e-02 -9.72119570e-01 -2.33071685e-01 -8.45969319e-01
-6.90782487e-01 1.16952503e+00 -4.38597202e-01 -1.26325834e+00
-6.71269894e-01 -3.38970423e-01 -5.35171568e-01 1.28294277e+00
-9.72876966e-01 -1.10987282e+00 2.68926293e-01 4.06446487e-01
1.08399928e+00 -3.89747173e-01 1.10308015e+00 5.76702952e-01
-4.91301149e-01 1.12269080e+00 3.30885887e-01 2.08607018e-01
9.81316626e-01 -1.18970871e+00 9.77293015e-01 9.79768574e-01
2.99618751e-01 3.48911136e-01 9.14230049e-01 -3.98530185e-01
-1.14898276e+00 -1.29253113e+00 1.58631635e+00 -6.22458220e-01
6.59649312e-01 -6.86283231e-01 -6.03723347e-01 7.45699346e-01
8.48968685e-01 -5.15170276e-01 6.31265521e-01 -9.64995939e-03
-2.20132947e-01 -9.68594328e-02 -8.45467925e-01 5.24403334e-01
7.33904600e-01 -7.92290330e-01 -6.49945080e-01 4.01123464e-01
1.25122380e+00 -6.21730983e-01 -8.48455608e-01 3.08574252e-02
5.38198352e-01 -3.35020155e-01 5.88776350e-01 -4.77960229e-01
3.99611771e-01 -8.28931928e-02 -4.40447301e-01 -1.78782320e+00
-8.00374225e-02 -1.13775456e+00 2.08753809e-01 1.34014988e+00
1.38726759e+00 -3.75253916e-01 3.14611971e-01 1.25541240e-01
-7.01136887e-01 -5.40418327e-01 -1.28224492e+00 -9.90079761e-01
3.15467447e-01 -6.05568528e-01 3.67590696e-01 5.84245622e-01
1.86694980e-01 9.15862024e-01 -4.60389137e-01 -3.86368148e-02
3.39445695e-02 -4.09573466e-01 6.77358627e-01 -5.04899025e-01
-5.31619251e-01 -3.68963212e-01 -1.87887058e-01 -1.10770464e+00
-8.79074037e-02 -1.05308998e+00 3.27516645e-01 -1.52379704e+00
-2.09909156e-01 -5.88449314e-02 -1.12891309e-01 6.03846908e-01
1.44074753e-01 4.44257230e-01 1.77772433e-01 1.33032948e-01
-1.14928529e-01 3.44531685e-01 8.43300939e-01 -3.77766371e-01
-3.17172378e-01 1.26521707e-01 -3.48586321e-01 -3.53940018e-03
9.93315339e-01 -5.22682488e-01 -2.05342099e-01 -7.62928903e-01
-1.57833949e-01 2.41794512e-01 -1.38880342e-01 -9.38654423e-01
2.48617269e-02 3.98147702e-01 -3.11743841e-02 -5.45209706e-01
4.10860240e-01 -5.67382812e-01 1.58174574e-01 3.90080124e-01
-4.40507978e-01 2.55348653e-01 4.07049835e-01 -3.80175635e-02
-4.88010973e-01 -6.23479001e-02 7.17592239e-01 8.62312466e-02
-2.52448499e-01 -1.18928561e-02 -7.96075642e-01 9.81684029e-03
6.01138949e-01 1.25038579e-01 4.10438515e-02 -7.22886145e-01
-8.88770878e-01 1.82320535e-01 1.78203687e-01 5.16570926e-01
2.82696694e-01 -1.34719110e+00 -1.15705597e+00 2.77017325e-01
-1.03831083e-01 -5.87715089e-01 -1.35623038e-01 9.73618090e-01
-4.97883916e-01 5.97817004e-01 1.13339417e-01 -5.88737011e-01
-1.33087909e+00 9.02709365e-02 3.74978900e-01 -2.27339670e-01
-3.91111255e-01 1.07168019e+00 -4.90699500e-01 -5.99905014e-01
7.96088874e-02 -2.85267025e-01 2.74499565e-01 -3.51355202e-03
4.49882895e-01 1.35843724e-01 7.01062441e-01 -6.24047875e-01
-3.70029509e-01 4.33268309e-01 -3.73617038e-02 -1.02210367e+00
1.06119597e+00 -3.05888414e-01 2.01896504e-01 5.27271569e-01
1.58569551e+00 1.08370461e-01 -6.81130052e-01 -1.50009125e-01
-3.08616668e-01 1.68341160e-01 2.57794261e-01 -1.19423485e+00
-8.76909733e-01 1.22548175e+00 4.91311431e-01 7.72821680e-02
1.04977655e+00 -3.54563855e-02 1.17845285e+00 3.13463897e-01
2.84414500e-01 -1.26170886e+00 -3.49774867e-01 9.37396765e-01
9.92397606e-01 -9.13534641e-01 -6.81893528e-01 -9.82541516e-02
-6.69478238e-01 1.11001456e+00 9.11337212e-02 9.09321606e-02
2.49229491e-01 4.02583718e-01 5.16093254e-01 4.66479152e-01
-1.03818023e+00 -1.32499129e-01 2.39183635e-01 3.59936744e-01
5.65307438e-01 4.55963105e-01 -5.38330674e-01 6.14990354e-01
-7.65942693e-01 -2.86394954e-01 6.01654768e-01 3.73566270e-01
-4.10685420e-01 -1.34738100e+00 -5.21863252e-02 8.94774646e-02
-6.47537053e-01 -3.77471834e-01 -6.69632912e-01 4.16170716e-01
-1.87326461e-01 1.37717283e+00 -1.26286456e-02 -9.76116121e-01
5.09630501e-01 4.78351563e-01 2.77066857e-01 -6.55891240e-01
-1.03298283e+00 5.70566535e-01 6.17401958e-01 -3.80078167e-01
4.50843712e-03 -6.79734945e-01 -1.02833152e+00 -3.77182007e-01
-4.19691741e-01 5.86421192e-01 1.02631402e+00 8.87027502e-01
4.09776181e-01 6.14084840e-01 9.61246073e-01 -6.56401753e-01
-5.31834900e-01 -1.53629208e+00 -1.36571705e-01 -1.15015619e-01
3.70840013e-01 1.07299522e-01 -2.75204271e-01 3.55781585e-01] | [14.486565589904785, 7.128199100494385] |
3a2d3a04-2673-43dd-8323-b10cbdee78a7 | adaptive-deep-pyramid-matching-for-remote | 1611.03589 | null | http://arxiv.org/abs/1611.03589v1 | http://arxiv.org/pdf/1611.03589v1.pdf | Adaptive Deep Pyramid Matching for Remote Sensing Scene Classification | Convolutional neural networks (CNNs) have attracted increasing attention in
the remote sensing community. Most CNNs only take the last fully-connected
layers as features for the classification of remotely sensed images, discarding
the other convolutional layer features which may also be helpful for
classification purposes. In this paper, we propose a new adaptive deep pyramid
matching (ADPM) model that takes advantage of the features from all of the
convolutional layers for remote sensing image classification. To this end, the
optimal fusing weights for different convolutional layers are learned from the
data itself. In remotely sensed scenes, the objects of interest exhibit
different scales in distinct scenes, and even a single scene may contain
objects with different sizes. To address this issue, we select the CNN with
spatial pyramid pooling (SPP-net) as the basic deep network, and further
construct a multi-scale ADPM model to learn complementary information from
multi-scale images. Our experiments have been conducted using two widely used
remote sensing image databases, and the results show that the proposed method
significantly improves the performance when compared to other state-of-the-art
methods. | ['Fuping Zhu', 'Renlong Hang', 'Javier Plaza', 'Qingshan Liu', 'Huihui Song', 'Antonio Plaza'] | 2016-11-11 | null | null | null | null | ['remote-sensing-image-classification'] | ['miscellaneous'] | [ 2.90334672e-01 -4.81100053e-01 1.08818263e-01 -5.14390528e-01
-2.90937960e-01 -2.14122057e-01 2.67558664e-01 1.75944343e-01
-4.70407546e-01 4.03275281e-01 8.71728659e-02 -1.20408446e-01
-3.17719370e-01 -1.33473408e+00 -4.83428091e-01 -8.04207742e-01
-3.10240865e-01 -3.65004689e-01 4.67932105e-01 -1.54155359e-01
1.51413381e-01 9.28082347e-01 -1.61398077e+00 5.22141457e-01
9.22095358e-01 1.20787644e+00 5.73932111e-01 3.12473446e-01
-1.34340584e-01 6.91850424e-01 -2.74270326e-01 6.32775128e-02
6.87214196e-01 -4.11347821e-02 -5.22576272e-01 1.30002826e-01
4.78418559e-01 -6.61865234e-01 -5.29907167e-01 1.33153641e+00
5.37953615e-01 1.82874873e-01 3.29419702e-01 -7.88915336e-01
-7.80309200e-01 4.00845081e-01 -6.84189141e-01 5.10696709e-01
-5.35573661e-01 -6.90224692e-02 1.11148250e+00 -1.03051651e+00
9.99876205e-03 1.32485771e+00 7.87766755e-01 -2.00269178e-01
-8.55952382e-01 -6.95157051e-01 3.08873564e-01 2.80511111e-01
-1.79092264e+00 -9.02893618e-02 6.58319831e-01 -3.96775812e-01
6.73332989e-01 1.06089599e-01 6.96129620e-01 6.65038452e-02
2.96418935e-01 7.16399074e-01 1.09977818e+00 -8.50865245e-02
-6.51895022e-03 -2.31563568e-01 3.47117074e-02 5.87536633e-01
3.02099556e-01 -7.12241530e-02 -5.50225675e-02 -1.87923968e-01
9.73282278e-01 7.79353023e-01 -4.33338344e-01 2.11793527e-01
-1.08219600e+00 8.55688155e-01 1.30476093e+00 4.66638595e-01
-7.71812081e-01 -1.15001619e-01 -4.35649008e-02 2.02646822e-01
5.36865056e-01 1.04068212e-01 -5.52806020e-01 7.38354087e-01
-8.37057412e-01 1.80321500e-01 1.38616145e-01 4.43297982e-01
1.30750227e+00 -1.70161441e-01 -3.99099998e-02 1.00809801e+00
2.90110350e-01 4.19036180e-01 4.64998722e-01 -4.53338832e-01
4.93826300e-01 1.04942608e+00 5.81282238e-03 -1.33553064e+00
-5.08286595e-01 -5.25853038e-01 -1.37983358e+00 3.05759937e-01
-1.22695304e-01 3.35489312e-04 -1.02441704e+00 1.23439860e+00
8.45978335e-02 2.47597948e-01 1.91842392e-01 9.71171737e-01
1.06136847e+00 1.09106743e+00 1.11721724e-01 2.78385341e-01
1.25365376e+00 -8.37843180e-01 -2.99615890e-01 -4.33542579e-01
1.13272928e-01 -6.66392863e-01 9.84648347e-01 3.56700160e-02
-4.84768838e-01 -9.35574234e-01 -1.23279059e+00 1.06561303e-01
-6.22851968e-01 5.68459213e-01 6.53735042e-01 5.23089133e-02
-8.02749336e-01 8.12881172e-01 -6.93708479e-01 -2.37946659e-01
7.01200068e-01 1.81789711e-01 -3.14675272e-01 -2.84866810e-01
-1.21623194e+00 5.47527313e-01 6.60746694e-01 8.14566135e-01
-6.02141082e-01 -4.93326604e-01 -6.43605471e-01 3.79867077e-01
-5.78650348e-02 -5.96689023e-02 7.40851343e-01 -1.15108633e+00
-8.39403927e-01 5.83863735e-01 1.79516017e-01 2.29877830e-02
8.98718163e-02 -2.07926646e-01 -4.48010355e-01 1.95872024e-01
1.80687800e-01 6.85951710e-01 7.13608146e-01 -9.40410912e-01
-9.17199194e-01 -3.48554224e-01 2.78858960e-01 1.83184296e-01
-4.68812138e-01 3.91138829e-02 -1.24627240e-01 -6.90142870e-01
6.52608633e-01 -5.73952079e-01 -4.48516697e-01 3.59126866e-01
-1.09419957e-01 -6.78624734e-02 9.75058615e-01 -6.70249999e-01
8.91847372e-01 -2.39507508e+00 -1.37333050e-01 1.33723393e-01
1.46748111e-01 4.61734176e-01 -3.85716975e-01 3.61561537e-01
-7.25416616e-02 2.29698405e-01 -4.18298602e-01 1.79731563e-01
-4.47268844e-01 3.68412852e-01 -1.98297799e-01 5.44726312e-01
4.63860512e-01 8.37744892e-01 -6.23609364e-01 -5.34397185e-01
2.29097024e-01 4.51525509e-01 -1.96472749e-01 3.74910921e-01
1.38830632e-01 4.04403061e-01 -7.69495904e-01 8.18089664e-01
1.18370700e+00 -3.83367568e-01 -5.80855981e-02 -4.38354671e-01
-4.17007804e-01 -1.62746921e-01 -1.26394629e+00 1.30471420e+00
-2.89323419e-01 5.10343134e-01 1.09119251e-01 -1.01984179e+00
1.21851158e+00 5.36821596e-02 4.24133927e-01 -5.95964730e-01
-4.68851030e-02 3.23314965e-01 2.49564257e-02 -4.92567807e-01
4.96085525e-01 -1.13611653e-01 2.34769598e-01 -5.28218076e-02
-3.55657071e-01 1.47662193e-01 -1.77748218e-01 -4.76810694e-01
6.78105175e-01 -2.16325879e-01 4.03840899e-01 -3.93786788e-01
7.11283803e-01 1.94074456e-02 8.52751076e-01 8.71338964e-01
-2.50498593e-01 6.03095949e-01 1.51767479e-02 -1.06824732e+00
-8.96880567e-01 -8.10765862e-01 -3.69369686e-01 9.91876125e-01
3.34201157e-01 2.62584597e-01 -1.88716978e-01 -3.78712147e-01
1.03384823e-01 -3.01525861e-01 -6.24796271e-01 1.42614067e-01
-5.59382856e-01 -1.10725129e+00 5.20441294e-01 6.13557279e-01
1.21830904e+00 -1.24885774e+00 -5.74353635e-01 4.55974877e-01
-8.94860551e-02 -9.73336875e-01 1.13957841e-02 1.04755417e-01
-1.10896111e+00 -1.10482180e+00 -8.02459002e-01 -8.78203571e-01
6.78760111e-01 8.83518398e-01 7.04683006e-01 2.86641181e-01
-1.84159204e-01 -4.09713358e-01 -6.12343431e-01 -3.61930341e-01
1.25584185e-01 2.30401635e-01 -4.64795977e-01 4.49885249e-01
3.59912068e-01 -7.65780628e-01 -8.32402706e-01 2.30759978e-01
-1.22044194e+00 -5.34994006e-02 1.19198036e+00 7.83231974e-01
5.41608095e-01 5.83894610e-01 2.87496537e-01 -5.23655295e-01
2.83070236e-01 -5.14923871e-01 -6.45529091e-01 2.55768508e-01
-1.00272775e-01 -2.13001400e-01 8.33033562e-01 -2.19803497e-01
-9.32523787e-01 1.69460922e-01 -1.86723754e-01 -2.92496979e-01
-3.26810688e-01 1.02446997e+00 -2.71884024e-01 -3.20047319e-01
5.00762165e-01 4.67040390e-01 -2.67506540e-01 -8.67065907e-01
1.94013808e-02 1.00658262e+00 2.81132996e-01 -1.64757803e-01
8.43022108e-01 7.16274977e-01 -4.36985753e-02 -1.02505958e+00
-1.08461082e+00 -5.29195428e-01 -7.88293242e-01 4.79222536e-02
8.82838011e-01 -1.27607512e+00 -4.16210145e-01 8.62331748e-01
-1.05582440e+00 -6.98361769e-02 6.60059080e-02 6.81672394e-01
1.42496198e-01 2.94252425e-01 -6.35622919e-01 -6.09277546e-01
-4.71880227e-01 -1.06699121e+00 1.08254564e+00 6.06260359e-01
6.21672690e-01 -5.63646078e-01 -1.33153379e-01 -1.46494448e-01
6.36151612e-01 1.73675507e-01 8.74406278e-01 -4.67214286e-01
-7.21639872e-01 -1.43832356e-01 -8.68005574e-01 5.67591310e-01
4.14674670e-01 1.66433305e-02 -9.11746562e-01 -3.66057009e-01
-1.67540461e-01 -2.24208117e-01 1.25062740e+00 4.22682583e-01
1.34323192e+00 -2.99509406e-01 -8.33229199e-02 8.35713208e-01
1.97477210e+00 9.27829221e-02 7.47653186e-01 6.25668764e-01
7.13086665e-01 6.59321964e-01 3.97515386e-01 5.01483023e-01
1.70950606e-01 1.47794291e-01 7.67404914e-01 -5.20880044e-01
3.67791772e-01 -8.39658603e-02 -5.86012229e-02 5.26066780e-01
-2.98930824e-01 1.64696008e-01 -7.63084054e-01 6.49273932e-01
-1.84593320e+00 -9.69979942e-01 -1.35679260e-01 1.83698869e+00
4.35660273e-01 -2.17352971e-01 -4.46102589e-01 9.92484018e-02
8.38976145e-01 7.38449514e-01 -5.19788682e-01 2.61118442e-01
-3.69028658e-01 3.34834427e-01 8.18835258e-01 1.99608669e-01
-1.56600428e+00 8.12625647e-01 5.93062019e+00 6.94307089e-01
-1.49731863e+00 -3.22005004e-02 4.97031718e-01 4.57627952e-01
3.66505943e-02 6.45710761e-03 -6.65514171e-01 1.00963987e-01
2.36145377e-01 2.11062133e-01 5.25839590e-02 9.55149174e-01
4.00108695e-01 -7.88374916e-02 -5.51580846e-01 8.07809532e-01
-3.53404731e-01 -1.23914289e+00 2.61826307e-01 9.06956196e-02
8.57638478e-01 5.01585662e-01 -1.38436809e-01 1.78285971e-01
4.01104167e-02 -1.14697397e+00 5.82867086e-01 5.96588135e-01
3.32503647e-01 -6.38903260e-01 1.10136592e+00 4.44261909e-01
-1.91255200e+00 -3.38608682e-01 -1.12854803e+00 -3.14660847e-01
-3.41103852e-01 7.82223284e-01 -1.73185334e-01 7.58485317e-01
1.04956210e+00 1.09730971e+00 -7.28505194e-01 1.14559829e+00
-1.32695928e-01 4.44974065e-01 -3.47126245e-01 1.84666649e-01
6.45331979e-01 -2.03145728e-01 5.57618029e-02 1.04037905e+00
4.57304657e-01 3.91714424e-01 5.63161016e-01 8.06915402e-01
-1.02873035e-01 3.13328952e-01 -5.39828479e-01 2.51050889e-02
4.12513375e-01 1.45733619e+00 -5.86064398e-01 -3.26949418e-01
-6.11102402e-01 7.23240495e-01 2.54471332e-01 2.82755524e-01
-4.76306558e-01 -5.91900408e-01 9.57144797e-01 -2.31352091e-01
6.17124677e-01 -3.71496856e-01 -5.23845665e-02 -1.21934688e+00
2.04323679e-01 -5.81504345e-01 3.67366165e-01 -8.72654974e-01
-1.38344431e+00 6.64146602e-01 -8.44685584e-02 -1.54930472e+00
5.79984188e-01 -6.85909390e-01 -8.61453235e-01 1.21701896e+00
-2.43392777e+00 -1.31852567e+00 -9.08358574e-01 5.63094318e-01
3.28130871e-01 2.39625927e-02 5.98328590e-01 3.25495273e-01
-4.13869292e-01 4.92468439e-02 2.96206445e-01 5.75390935e-01
2.51805365e-01 -8.22553754e-01 2.94348031e-01 1.05740464e+00
-1.94297582e-01 4.19020444e-01 5.93723990e-02 -3.58128548e-01
-1.10502958e+00 -1.61506808e+00 7.08527565e-01 5.53274035e-01
5.02237618e-01 1.53903201e-01 -1.30696201e+00 4.93042052e-01
-8.75191540e-02 4.86154675e-01 6.33477569e-01 -2.61145681e-01
-3.29339862e-01 -6.45340085e-01 -1.04419303e+00 2.65063763e-01
7.05012500e-01 -4.73004043e-01 -5.96188188e-01 1.77489519e-01
6.08809829e-01 4.50077793e-03 -9.10270333e-01 7.02802598e-01
5.02028704e-01 -1.07145858e+00 9.72622454e-01 -3.71274173e-01
6.08652592e-01 -6.55796111e-01 -5.48006654e-01 -1.09689224e+00
-8.23993921e-01 2.19529703e-01 6.77596331e-01 7.64054298e-01
1.28485784e-01 -7.00704873e-01 6.38918161e-01 -8.36541578e-02
-1.60970435e-01 -5.46014667e-01 -7.01307237e-01 -5.96784830e-01
5.53077236e-02 -1.34746417e-01 9.31423068e-01 9.50588584e-01
-6.91684365e-01 1.39719307e-01 -4.17191237e-01 8.53463173e-01
5.82876503e-01 6.15696788e-01 5.27450144e-01 -1.59842026e+00
-9.41894129e-02 -5.36524832e-01 -5.84319115e-01 -1.29027498e+00
-1.71222791e-01 -7.53741682e-01 1.64903313e-01 -1.66071367e+00
2.35995457e-01 -5.52237868e-01 -5.27459919e-01 8.77320170e-01
-4.09293592e-01 3.65820676e-01 1.07267104e-01 7.02973247e-01
-3.21584269e-02 6.62509143e-01 1.29277635e+00 -4.66068715e-01
-1.91104814e-01 -1.11488394e-01 -6.27556920e-01 7.42745638e-01
1.00447726e+00 -3.68291706e-01 -3.97069976e-02 -7.73198605e-01
2.81115472e-02 -3.06435317e-01 5.30609906e-01 -1.23817158e+00
2.31854022e-01 -4.66763258e-01 6.32904887e-01 -8.58456075e-01
4.95744646e-02 -9.78931487e-01 1.48067132e-01 6.76623225e-01
-4.48573418e-02 -1.25901982e-01 2.81946033e-01 3.62439752e-01
-6.25004947e-01 -1.76940948e-01 9.24373090e-01 -5.36467075e-01
-1.17883277e+00 7.48558342e-01 -3.01032305e-01 -5.19405007e-01
6.71204984e-01 -1.84149668e-01 -3.05993855e-01 1.41457424e-01
-5.77520311e-01 2.06069082e-01 1.51413754e-01 2.99109221e-01
7.39448488e-01 -1.37740099e+00 -1.12510848e+00 4.06247705e-01
4.52241808e-01 3.18810701e-01 4.83789325e-01 5.07522166e-01
-8.78849864e-01 1.14122771e-01 -5.15241146e-01 -6.38134539e-01
-9.58470106e-01 1.01066090e-01 6.25563622e-01 -2.19217449e-01
-6.25958264e-01 6.21195555e-01 2.27366492e-01 -5.86867094e-01
-1.87314793e-01 -5.86676776e-01 -4.67752188e-01 3.55667993e-02
6.93851531e-01 1.34812266e-01 1.16245180e-01 -7.22448945e-01
-3.52602929e-01 7.80650437e-01 1.58037528e-01 3.01463813e-01
1.76267362e+00 -1.45476311e-02 -5.27215064e-01 2.60341257e-01
1.30531776e+00 -3.08666855e-01 -1.19791710e+00 -6.99728072e-01
-3.27181637e-01 -7.21920252e-01 2.72542477e-01 -2.37781256e-01
-1.40291178e+00 1.17212057e+00 6.84470296e-01 2.54060656e-01
1.43881428e+00 -2.40644202e-01 7.06198573e-01 6.68646872e-01
2.08354503e-01 -8.09722900e-01 -1.87594697e-01 6.33516610e-01
8.32786500e-01 -1.35565734e+00 1.74552709e-01 -2.85969168e-01
-1.61593601e-01 1.39147675e+00 5.49469113e-01 -3.66230518e-01
1.06973040e+00 -1.96091071e-01 1.70217216e-01 -1.80212393e-01
-1.59792691e-01 -5.78466356e-01 2.68432826e-01 3.10779810e-01
3.14553946e-01 2.01938197e-01 -2.66015828e-01 3.54996681e-01
3.12829316e-01 1.79843530e-02 1.78639904e-01 1.19027567e+00
-1.06201339e+00 -8.64683211e-01 -4.61956561e-01 5.97827971e-01
-4.17755127e-01 -3.07878703e-01 -1.56674474e-01 5.73096454e-01
4.39991593e-01 7.75401890e-01 2.45135561e-01 -4.96773601e-01
3.82443458e-01 -3.25138241e-01 -5.36253899e-02 -5.66029251e-01
-6.37050450e-01 -1.10788569e-01 -5.18127263e-01 -2.50615448e-01
-9.69695568e-01 -2.31571898e-01 -1.03691578e+00 -3.50876153e-01
-2.01610997e-01 4.07258645e-02 3.64866912e-01 9.47684467e-01
3.05829585e-01 4.07054991e-01 9.93384421e-01 -1.12655830e+00
-4.78949130e-01 -1.12449539e+00 -9.69855905e-01 9.83433649e-02
3.94532412e-01 -4.61729228e-01 -2.89726615e-01 -1.31392747e-01] | [9.756077766418457, -1.4288630485534668] |
ced701b2-f9c8-448e-897f-93d30b0576c8 | 2-5d-visual-relationship-detection | 2104.12727 | null | https://arxiv.org/abs/2104.12727v1 | https://arxiv.org/pdf/2104.12727v1.pdf | 2.5D Visual Relationship Detection | Visual 2.5D perception involves understanding the semantics and geometry of a scene through reasoning about object relationships with respect to the viewer in an environment. However, existing works in visual recognition primarily focus on the semantics. To bridge this gap, we study 2.5D visual relationship detection (2.5VRD), in which the goal is to jointly detect objects and predict their relative depth and occlusion relationships. Unlike general VRD, 2.5VRD is egocentric, using the camera's viewpoint as a common reference for all 2.5D relationships. Unlike depth estimation, 2.5VRD is object-centric and not only focuses on depth. To enable progress on this task, we create a new dataset consisting of 220k human-annotated 2.5D relationships among 512K objects from 11K images. We analyze this dataset and conduct extensive experiments including benchmarking multiple state-of-the-art VRD models on this task. Our results show that existing models largely rely on semantic cues and simple heuristics to solve 2.5VRD, motivating further research on models for 2.5D perception. The new dataset is available at https://github.com/google-research-datasets/2.5vrd. | ['Boqing Gong', 'Ming-Hsuan Yang', 'Matthew Brown', 'Hartwig Adam', 'Radu Soricut', 'Lior Shapira', 'Cho-Jui Hsieh', 'Sathish Thoppay', 'Xiangning Chen', 'Soravit Changpinyo', 'Yu-Chuan Su'] | 2021-04-26 | null | null | null | null | ['visual-relationship-detection'] | ['computer-vision'] | [-2.19580561e-01 -3.87120210e-02 -1.47545725e-01 -6.14952326e-01
-1.62877321e-01 -7.87993073e-01 6.20921791e-01 2.82432109e-01
-1.12231923e-02 -1.66450313e-03 2.79984236e-01 -2.90444404e-01
1.34214178e-01 -5.51073790e-01 -6.33397937e-01 -1.78573892e-01
-4.33146320e-02 5.35393715e-01 5.15657246e-01 -1.35085255e-01
6.17978096e-01 8.21597099e-01 -1.62714124e+00 4.02287543e-01
7.42573440e-02 9.67387259e-01 5.32571673e-01 8.92853975e-01
7.21998960e-02 9.41715002e-01 -5.02229929e-01 -6.91662505e-02
4.26789939e-01 -1.19129702e-01 -8.57952714e-01 4.50940788e-01
9.63612676e-01 -4.80949551e-01 -8.35440814e-01 8.51872861e-01
3.48223716e-01 3.48496139e-01 5.88753521e-01 -1.73302603e+00
-9.15325224e-01 -1.61736786e-01 -7.08882272e-01 4.99389052e-01
1.00952625e+00 1.11721374e-01 1.11093652e+00 -1.07134199e+00
8.60382795e-01 1.54681909e+00 2.73706913e-01 5.20473123e-01
-1.17961502e+00 -1.67228043e-01 5.29492021e-01 2.32096136e-01
-1.45464993e+00 -4.69532490e-01 7.95088828e-01 -7.29292750e-01
1.49620938e+00 3.14424306e-01 8.05311620e-01 8.05072725e-01
2.25641504e-02 8.20719361e-01 9.24129128e-01 -2.95226991e-01
1.58260688e-01 1.91600114e-01 1.59102291e-01 5.48766792e-01
1.44590691e-01 -8.31601396e-02 -7.17447340e-01 1.82335004e-01
1.15355968e+00 -3.73324119e-02 -2.76843905e-01 -1.02528667e+00
-1.15898120e+00 5.17975092e-01 7.33827472e-01 -1.26461104e-01
3.97491902e-02 3.35152537e-01 7.30641484e-02 1.97205409e-01
4.01366651e-01 3.59669834e-01 -3.17681372e-01 6.62893709e-03
-3.64012569e-01 4.66076732e-01 6.29821599e-01 1.26865530e+00
6.17901087e-01 -3.98357451e-01 9.53366160e-02 6.93780005e-01
5.87690055e-01 3.80794227e-01 -1.12439431e-01 -1.39378822e+00
6.16322815e-01 6.26627386e-01 2.09601671e-01 -1.32664752e+00
-4.39283609e-01 -1.18125767e-01 -3.09931725e-01 4.11122262e-01
4.31066841e-01 6.55470312e-01 -1.03607082e+00 1.44715571e+00
4.97920990e-01 1.40012518e-01 9.01707634e-02 1.41816962e+00
1.38244987e+00 4.90425289e-01 -2.49082565e-01 3.40245187e-01
1.37541199e+00 -1.05534506e+00 -2.89663225e-01 -6.68716967e-01
6.39376998e-01 -8.56285989e-01 9.53388274e-01 2.12110430e-01
-1.10551047e+00 -6.22785449e-01 -9.27643418e-01 -6.27384543e-01
-5.39610863e-01 -4.61142808e-02 7.76431322e-01 2.30069324e-01
-1.29986525e+00 -1.11336922e-02 -5.88150501e-01 -6.77588701e-01
5.68545938e-01 4.12491299e-02 -6.16295397e-01 -4.61101502e-01
-7.44276106e-01 1.21688187e+00 2.02068672e-01 8.51030573e-02
-1.06450176e+00 -5.37320673e-01 -1.18956971e+00 -3.21046770e-01
3.59310448e-01 -8.16836834e-01 1.39206016e+00 -2.00791210e-01
-7.28738964e-01 1.63641608e+00 -4.93696094e-01 -1.36429295e-01
5.18095970e-01 -2.98633456e-01 -6.35422319e-02 2.11092204e-01
2.64806122e-01 9.47592914e-01 3.73623163e-01 -1.85895658e+00
-6.49059832e-01 -5.72589457e-01 6.73909783e-01 7.57168055e-01
3.83365899e-01 -9.33891758e-02 -9.76036072e-01 -1.22465283e-01
7.77362585e-01 -8.83596718e-01 -1.10297754e-01 4.81315553e-01
-4.86685336e-01 -2.74215281e-01 9.10058379e-01 -3.52872133e-01
5.84387243e-01 -2.09402514e+00 1.73122838e-01 -6.78937957e-02
7.03574002e-01 -1.75338760e-01 -2.94310059e-02 2.83854455e-01
4.33667898e-02 -1.08700663e-01 2.46137425e-01 -8.13927770e-01
-5.81846796e-02 2.43657976e-01 -3.00202310e-01 6.91392481e-01
-1.85250454e-02 8.35903645e-01 -1.10322380e+00 -4.62934375e-01
6.57627463e-01 4.87477034e-01 -4.68968958e-01 3.62812638e-01
-2.56087899e-01 3.76597553e-01 -3.57203037e-01 8.95297527e-01
8.52778673e-01 -3.68962377e-01 -3.42159458e-02 -2.29285210e-01
-1.00540876e-01 2.83060759e-01 -1.13475049e+00 2.06271958e+00
-2.27857545e-01 1.13219416e+00 -2.82146215e-01 -7.33543456e-01
1.14867795e+00 -1.32263184e-01 3.27380776e-01 -9.21405971e-01
6.70521110e-02 -9.97688174e-02 -3.56719136e-01 -5.81115901e-01
7.41416931e-01 3.16281557e-01 3.35930996e-02 1.09794624e-01
-2.75326014e-01 -7.52106965e-01 -7.50910118e-02 5.13241053e-01
9.36613321e-01 3.31662655e-01 3.68365228e-01 1.44578427e-01
1.20364636e-01 2.03123182e-01 3.00543994e-01 7.50782311e-01
-4.87864524e-01 8.46043706e-01 5.42364240e-01 -4.10193831e-01
-9.95514810e-01 -1.45530689e+00 -5.22415787e-02 5.94909191e-01
1.05164671e+00 -6.09009802e-01 -2.97288485e-02 -4.06531364e-01
3.11882287e-01 7.71042645e-01 -7.65136242e-01 1.77366704e-01
-2.88481534e-01 -1.77286610e-01 2.36336631e-03 7.86653042e-01
3.92298341e-01 -7.04327226e-01 -9.35675204e-01 -3.53866845e-01
-2.35333979e-01 -1.55869067e+00 -2.28429377e-01 1.33792028e-01
-6.99649334e-01 -1.15502203e+00 -2.55902976e-01 -8.00271809e-01
6.54632688e-01 1.10489035e+00 1.55042434e+00 -3.43379118e-02
-4.39577013e-01 8.79491210e-01 -2.93015957e-01 -3.55980575e-01
6.94764555e-02 -3.61294389e-01 -4.26044166e-02 -5.41608512e-01
5.60132086e-01 -3.98237050e-01 -7.08516955e-01 5.96656978e-01
-2.08206058e-01 3.22000593e-01 8.92065763e-02 2.45005310e-01
8.34003031e-01 8.23847353e-02 -2.98619449e-01 -4.13687527e-01
8.86419564e-02 -5.04061699e-01 -6.56949401e-01 7.39099234e-02
-9.89428163e-02 -3.93849701e-01 -2.05740351e-02 -2.37748310e-01
-9.80485559e-01 9.28314552e-02 4.26858634e-01 -8.99884641e-01
-3.86924356e-01 -1.41517848e-01 -1.84150428e-01 2.59711761e-02
5.35447419e-01 6.10550232e-02 -3.34669530e-01 -4.54679698e-01
5.10422170e-01 4.34420198e-01 5.90599656e-01 -2.55132467e-01
6.50831163e-01 7.80839264e-01 1.01217277e-01 -9.10714447e-01
-1.16923809e+00 -8.55986536e-01 -7.90539443e-01 -4.16927427e-01
1.07540143e+00 -1.17490757e+00 -8.90945613e-01 2.95297176e-01
-1.34954405e+00 -4.84720916e-01 -7.16318563e-02 5.56435883e-01
-8.06721926e-01 9.12581682e-02 -3.05719703e-01 -8.98212492e-01
4.16512519e-01 -1.02261555e+00 1.44408989e+00 1.50201708e-01
-3.66462767e-01 -1.06211948e+00 1.67179964e-02 6.66410387e-01
-2.06975937e-01 3.97354782e-01 5.17900407e-01 -9.90754813e-02
-1.12752426e+00 7.96432123e-02 -6.75745606e-01 -1.30021879e-02
6.90357089e-02 -1.48215204e-01 -1.11303353e+00 -6.38017878e-02
-1.24781422e-01 -4.06066835e-01 8.49769652e-01 4.26996559e-01
1.16913760e+00 3.73765826e-01 -5.24288297e-01 8.10250103e-01
1.47902489e+00 2.27280378e-01 5.97484231e-01 4.96776104e-01
1.12546921e+00 7.65469074e-01 1.01267588e+00 4.53905523e-01
9.87140656e-01 9.64506328e-01 1.04237187e+00 -1.20281085e-01
-3.71625364e-01 -3.72070193e-01 -1.62815675e-01 2.05341890e-01
-1.37215257e-01 -4.71735269e-01 -1.17080092e+00 5.29875636e-01
-1.85293782e+00 -9.34029162e-01 -5.09638429e-01 1.97765923e+00
3.19081366e-01 2.10534394e-01 -4.18967605e-02 -1.01978341e-02
5.29337704e-01 3.81077439e-01 -7.74598598e-01 -3.52040559e-01
-2.43307456e-01 -5.38264632e-01 4.12422806e-01 5.93623996e-01
-1.06839526e+00 1.09225976e+00 5.77738762e+00 1.64983004e-01
-6.86225295e-01 -2.37343282e-01 4.15251404e-01 -2.17213199e-01
-3.33693504e-01 2.41060480e-01 -9.77638006e-01 -1.04935989e-01
6.49363026e-02 1.81649685e-01 3.90226096e-01 9.77041304e-01
1.55355453e-01 -5.07184565e-01 -1.57904601e+00 1.59126651e+00
3.85196030e-01 -1.19304526e+00 -9.56365094e-02 8.02500546e-02
5.82679272e-01 -2.95124687e-02 1.08473055e-01 -1.52048796e-01
1.52132213e-01 -1.18748856e+00 1.05004668e+00 5.94681561e-01
7.13847101e-01 -2.89457142e-01 3.84668916e-01 1.71621174e-01
-1.41710365e+00 6.16422333e-02 -4.12710994e-01 -4.58456278e-01
1.69892848e-01 2.85560191e-01 -1.09752035e+00 2.84615844e-01
1.16483498e+00 1.21550083e+00 -7.21050620e-01 9.49712038e-01
-2.39802912e-01 -2.44213268e-01 -2.84566045e-01 1.73922554e-01
8.52392986e-02 4.97031212e-02 5.57597697e-01 8.85174632e-01
7.29516372e-02 2.33615085e-01 1.84410468e-01 8.12593102e-01
1.29223093e-01 -3.10512394e-01 -9.34223175e-01 3.06655467e-01
6.27588272e-01 8.53067935e-01 -9.05955613e-01 -2.36037314e-01
-5.14011383e-01 1.04209101e+00 4.44280714e-01 3.75389189e-01
-9.40849841e-01 -1.66747533e-02 1.08683693e+00 1.80138037e-01
3.02871525e-01 -6.78002596e-01 -3.11549813e-01 -1.23194146e+00
1.23779088e-01 -4.67769116e-01 3.53815377e-01 -1.54915607e+00
-1.17555392e+00 4.41451341e-01 2.06097826e-01 -1.24787807e+00
2.30767488e-01 -8.08807015e-01 -7.57779926e-02 7.41313636e-01
-1.52859366e+00 -9.88843501e-01 -9.77974892e-01 6.90577030e-01
7.12541401e-01 2.95141846e-01 6.33193791e-01 -1.22782178e-01
-1.98729619e-01 3.70799601e-02 -4.12891030e-01 -3.56457196e-02
6.07563496e-01 -1.35283530e+00 7.84159005e-01 5.74104965e-01
3.03251028e-01 4.33580160e-01 7.68266261e-01 -5.87630212e-01
-1.49059010e+00 -7.63743341e-01 9.41333413e-01 -1.29690492e+00
4.66188490e-01 -7.20995843e-01 -6.71448052e-01 9.55445647e-01
-1.96912855e-01 3.91907632e-01 2.28389040e-01 4.88126464e-02
-6.30909503e-01 1.30481064e-01 -1.00858474e+00 7.97677577e-01
1.89742517e+00 -7.48461902e-01 -7.24619389e-01 3.75252903e-01
8.32614303e-01 -9.90471065e-01 -7.49835551e-01 1.44026801e-01
5.71802974e-01 -1.33219254e+00 1.55094111e+00 -4.56575632e-01
3.89095157e-01 -6.37993574e-01 -8.12816799e-01 -9.72063899e-01
-2.40022898e-01 -4.42116521e-02 -3.51912141e-01 8.16997349e-01
-5.57426959e-02 -4.06886429e-01 8.76910210e-01 6.40553415e-01
-1.03527308e-01 -6.91031516e-01 -6.40090168e-01 -7.30862677e-01
-5.60340941e-01 -7.48308480e-01 3.27084839e-01 9.43508983e-01
-4.26723957e-01 4.00703996e-01 -2.57189065e-01 5.36136270e-01
7.89502263e-01 4.22152102e-01 1.18147433e+00 -1.04910278e+00
-1.70287520e-01 -3.42154056e-01 -1.04065263e+00 -1.65030158e+00
-1.40647842e-02 -6.78822339e-01 -5.63827232e-02 -2.03580284e+00
1.42075792e-01 -5.68944693e-01 2.54272539e-02 3.74380469e-01
1.05459817e-01 4.04577821e-01 3.54167759e-01 1.98632926e-01
-8.11919332e-01 4.62194830e-01 1.48059320e+00 -5.80996610e-02
-3.38472128e-01 -1.63030848e-01 -5.77675819e-01 7.74043202e-01
9.93314266e-01 -1.20031141e-01 -9.05403197e-01 -7.10624993e-01
1.43396765e-01 1.82141140e-01 9.32508171e-01 -7.61547625e-01
1.97597161e-01 -3.66145372e-01 7.09334075e-01 -9.75230336e-01
1.09643614e+00 -7.56090343e-01 -4.61094677e-02 7.87120312e-02
-2.36524910e-01 1.49586990e-01 7.95329660e-02 6.49538517e-01
-7.51595274e-02 1.00186460e-01 4.88435209e-01 -4.24053103e-01
-1.47426772e+00 3.57979566e-01 1.58945769e-02 2.23104745e-01
1.28918672e+00 -6.88603640e-01 -5.69061756e-01 -4.70113307e-01
-7.85858989e-01 4.48724836e-01 8.11105072e-01 7.50628471e-01
1.20308411e+00 -1.17910600e+00 -5.18743753e-01 4.72280309e-02
6.08561218e-01 3.58081013e-01 4.55522805e-01 4.59455132e-01
-5.84499359e-01 5.10087907e-01 -2.04402819e-01 -1.12826300e+00
-1.56940460e+00 7.48738229e-01 4.18552250e-01 4.94696796e-01
-7.82189012e-01 1.20636499e+00 6.83865905e-01 -2.36100718e-01
4.20315176e-01 -3.96059096e-01 -2.51068324e-01 -7.84717202e-02
4.02164638e-01 1.88484699e-01 -2.27948651e-01 -8.13548684e-01
-8.41147959e-01 9.55478191e-01 7.26185068e-02 -6.65720999e-02
1.19609118e+00 -4.72253293e-01 -1.22470427e-02 7.41192162e-01
1.30622971e+00 -1.71258554e-01 -1.41959298e+00 -2.22300842e-01
-3.69848579e-01 -1.24072433e+00 7.84203261e-02 -4.61462379e-01
-8.41326892e-01 7.12301672e-01 4.03634220e-01 8.38446319e-02
9.65741456e-01 6.13795400e-01 2.00471863e-01 3.64686370e-01
5.67977607e-01 -7.47939408e-01 5.41780829e-01 5.15476704e-01
1.05041063e+00 -1.50343454e+00 2.83882707e-01 -9.09329295e-01
-7.67061949e-01 9.42737401e-01 1.19192851e+00 -9.49573964e-02
6.11305416e-01 -1.83846448e-02 1.78872660e-01 -6.24896884e-01
-8.07690084e-01 -3.62185061e-01 2.58892596e-01 7.85376787e-01
1.87144950e-01 -6.50048722e-03 5.81528842e-01 -2.55695105e-01
-3.28425616e-01 -3.67586732e-01 5.11308312e-01 1.04759479e+00
-2.43638456e-01 -6.94448888e-01 -3.57258946e-01 4.76499386e-02
3.81090164e-01 5.29665798e-02 -4.81129140e-01 8.35774958e-01
2.31912047e-01 1.07717681e+00 4.14244592e-01 -3.59212071e-01
5.53469419e-01 -5.10058045e-01 9.53460336e-01 -7.32417107e-01
1.89892501e-01 -2.86697626e-01 2.65744835e-01 -8.93450677e-01
-6.40875697e-01 -7.71684110e-01 -1.26842165e+00 -3.46344113e-01
2.36847382e-02 -4.59826171e-01 7.30196178e-01 5.05768478e-01
2.73825884e-01 3.10050339e-01 4.44962978e-01 -1.00813770e+00
1.35293797e-01 -3.66364539e-01 -5.88458121e-01 4.84221935e-01
5.85401356e-01 -1.06016672e+00 -4.29671824e-01 -1.90182030e-02] | [10.288493156433105, 1.6002289056777954] |
76a046c3-198c-417f-b333-65e07a715f70 | cross-sensor-adversarial-domain-adaptation-of | 2006.05923 | null | https://arxiv.org/abs/2006.05923v1 | https://arxiv.org/pdf/2006.05923v1.pdf | Cross-Sensor Adversarial Domain Adaptation of Landsat-8 and Proba-V images for Cloud Detection | The number of Earth observation satellites carrying optical sensors with similar characteristics is constantly growing. Despite their similarities and the potential synergies among them, derived satellite products are often developed for each sensor independently. Differences in retrieved radiances lead to significant drops in accuracy, which hampers knowledge and information sharing across sensors. This is particularly harmful for machine learning algorithms, since gathering new ground truth data to train models for each sensor is costly and requires experienced manpower. In this work, we propose a domain adaptation transformation to reduce the statistical differences between images of two satellite sensors in order to boost the performance of transfer learning models. The proposed methodology is based on the Cycle Consistent Generative Adversarial Domain Adaptation (CyCADA) framework that trains the transformation model in an unpaired manner. In particular, Landsat-8 and Proba-V satellites, which present different but compatible spatio-spectral characteristics, are used to illustrate the method. The obtained transformation significantly reduces differences between the image datasets while preserving the spatial and spectral information of adapted images, which is hence useful for any general purpose cross-sensor application. In addition, the training of the proposed adversarial domain adaptation model can be modified to improve the performance in a specific remote sensing application, such as cloud detection, by including a dedicated term in the cost function. Results show that, when the proposed transformation is applied, cloud detection models trained in Landsat-8 data increase cloud detection accuracy in Proba-V. | ['Luis Gómez-Chova', 'Dan López-Puigdollers', 'Gonzalo Mateo-García', 'Valero Laparra'] | 2020-06-10 | null | null | null | null | ['cloud-detection'] | ['computer-vision'] | [ 5.75261414e-01 -4.78047490e-01 1.72201231e-01 -1.58524737e-01
-6.72285557e-01 -8.92499566e-01 6.18037879e-01 3.54651883e-02
-4.84018117e-01 8.96298885e-01 -5.00513554e-01 -2.48740777e-01
-1.79594561e-01 -1.14665592e+00 -6.99558318e-01 -1.18035102e+00
-2.17953417e-02 5.16806655e-02 1.10674366e-01 -3.47705811e-01
-1.57394096e-01 8.56565773e-01 -1.55144668e+00 -1.55248463e-01
1.37712932e+00 1.00755894e+00 5.36491990e-01 4.39456642e-01
1.17338344e-01 2.50525147e-01 -4.32920247e-01 -1.03074506e-01
6.03474855e-01 -5.72277665e-01 -2.85076648e-01 2.42742643e-01
2.95062244e-01 -7.47485235e-02 -5.75502478e-02 1.39636612e+00
4.30560201e-01 2.10318193e-01 7.26246536e-01 -1.04931724e+00
-2.32155189e-01 5.11064008e-02 -4.33241665e-01 -9.31948870e-02
-2.65571475e-01 1.05014525e-01 5.46138823e-01 -4.91539598e-01
3.41878980e-01 6.37544513e-01 7.01444566e-01 1.93139806e-01
-1.29940999e+00 -7.10244477e-01 3.79333086e-02 2.89664026e-02
-1.77825069e+00 -6.36352375e-02 8.36675584e-01 -6.41698241e-01
3.86598468e-01 3.93961638e-01 8.13811600e-01 6.68936491e-01
2.23421708e-01 1.75926551e-01 1.55159080e+00 -3.46432656e-01
3.37430000e-01 3.71353149e-01 -5.06439447e-01 4.77021746e-02
4.53276753e-01 1.10779807e-01 -3.67509294e-03 -4.33878303e-02
4.76639062e-01 2.38191605e-01 -5.02836645e-01 -3.35959077e-01
-6.33256018e-01 9.05551791e-01 8.75469208e-01 5.21274328e-01
-6.10366523e-01 -2.94551790e-01 1.98959443e-03 3.76331836e-01
7.27234006e-01 2.09658012e-01 -1.32497892e-01 6.09062612e-01
-1.04451287e+00 2.91982949e-01 2.58337140e-01 6.83148146e-01
9.79096293e-01 3.44529659e-01 1.83817357e-01 6.49571657e-01
2.76943743e-01 1.11178684e+00 3.93886864e-01 -4.49419796e-01
3.03289264e-01 5.71901441e-01 3.21375310e-01 -1.14836562e+00
-2.47731522e-01 -8.98207307e-01 -1.14598668e+00 3.87308598e-01
1.26316380e-02 -1.22071058e-01 -8.60669792e-01 1.47941852e+00
4.03626710e-01 2.08863690e-01 4.46324617e-01 1.08342671e+00
2.69215941e-01 9.29160416e-01 2.36060277e-01 -1.62627473e-01
1.02763534e+00 -4.03818607e-01 -4.42459106e-01 -5.10906756e-01
4.45149064e-01 -6.36115134e-01 8.38105381e-01 4.24709730e-02
-3.80368412e-01 -6.05499387e-01 -1.15011382e+00 5.88829339e-01
-6.36741519e-01 1.17625087e-01 3.88628751e-01 5.17414212e-01
-8.28142703e-01 3.90328616e-01 -6.04260445e-01 -3.79370898e-01
2.04759404e-01 1.19251445e-01 -2.03016594e-01 -1.43664181e-01
-1.45618272e+00 8.16843867e-01 7.33923435e-01 2.81659424e-01
-5.63740671e-01 -5.41979313e-01 -6.57189488e-01 1.08848877e-01
-5.83515354e-02 -4.55119401e-01 6.56784713e-01 -1.46150553e+00
-1.25519323e+00 7.26208985e-01 3.25045794e-01 -6.18439853e-01
4.50159460e-01 -2.53905915e-02 -8.11494410e-01 1.04444265e-01
1.03050955e-01 3.96236032e-01 1.21554005e+00 -1.44268310e+00
-6.77694738e-01 -5.33550680e-01 -2.37436280e-01 2.34651104e-01
-4.37908232e-01 -2.65231371e-01 -2.58238129e-02 -8.38339269e-01
1.30424201e-01 -1.17326081e+00 -1.61115319e-01 -5.65313697e-02
8.13409984e-02 4.39922899e-01 9.27056074e-01 -7.98918307e-01
7.45697260e-01 -2.32470727e+00 3.58060710e-02 5.08104920e-01
-4.78876501e-01 6.12071633e-01 -2.27641150e-01 3.40166718e-01
2.75270026e-02 -9.53835249e-02 -9.24299002e-01 1.66930214e-01
-4.98742014e-01 3.86863887e-01 -2.96993643e-01 5.71966887e-01
4.66685295e-01 4.09064949e-01 -7.61721671e-01 -2.24670410e-01
3.37536782e-01 5.51016450e-01 -9.26459506e-02 3.04763287e-01
-3.26493502e-01 7.78115392e-01 -4.59270269e-01 3.73360813e-01
1.21157563e+00 1.05102353e-01 2.20010638e-01 -1.17576897e-01
-7.70373046e-02 -2.29765698e-01 -1.05983424e+00 1.30224025e+00
-6.36738658e-01 5.55165529e-01 1.43342867e-01 -1.20558119e+00
1.33236480e+00 2.21658349e-01 3.44097346e-01 -7.90994048e-01
-4.45135646e-02 3.63895625e-01 -2.74359118e-02 -3.27066898e-01
5.10370016e-01 -4.29154575e-01 7.94468746e-02 -7.56084397e-02
-3.00773740e-01 -6.36168182e-01 -2.02296063e-01 -3.41200739e-01
2.94714689e-01 1.45474404e-01 2.76736647e-01 -6.53558597e-02
8.93985271e-01 2.63007253e-01 5.09500325e-01 4.20382619e-01
2.11616904e-01 4.49195743e-01 -2.82076418e-01 -1.98879510e-01
-1.13961804e+00 -7.58968413e-01 -3.48863840e-01 6.08150125e-01
9.48068425e-02 4.92339045e-01 -6.46057963e-01 -3.94823939e-01
9.85549241e-02 7.64886260e-01 -3.41211051e-01 -1.97062299e-01
-1.02814995e-01 -9.68084037e-01 6.48554742e-01 2.29560509e-01
8.80335748e-01 -7.21594930e-01 -5.34927428e-01 2.62025028e-01
-2.38380253e-01 -1.19862413e+00 1.15397945e-01 1.50930211e-01
-1.25275660e+00 -8.31879497e-01 -8.79986405e-01 -4.92447734e-01
6.04715765e-01 5.22055268e-01 7.53314495e-01 -1.46952391e-01
1.08615339e-01 2.29166716e-01 -5.70894897e-01 -5.27306080e-01
-5.81975639e-01 1.20491877e-01 -1.42919138e-01 5.99464357e-01
2.33955771e-01 -6.16846442e-01 -3.77686024e-01 2.66718030e-01
-1.47147119e+00 -1.54660523e-01 5.99598527e-01 8.57314110e-01
5.85618913e-01 3.09560627e-01 4.19246972e-01 -6.74735248e-01
2.14224458e-01 -5.00095308e-01 -1.14086187e+00 1.85583815e-01
-6.92893445e-01 -1.11978263e-01 6.62136316e-01 -1.91582307e-01
-1.24868822e+00 4.32111710e-01 1.95194319e-01 -3.15725029e-01
-1.09836683e-01 8.83572400e-01 -1.80202335e-01 -4.75757092e-01
9.04298961e-01 6.13042414e-01 7.32053965e-02 -3.25113386e-01
1.57584029e-03 8.36367667e-01 6.93393052e-01 -5.50193787e-02
1.47620988e+00 5.78728855e-01 1.96268365e-01 -1.12510931e+00
-4.29830849e-01 -4.95760620e-01 -4.70419437e-01 -2.73503989e-01
6.74660087e-01 -1.34903657e+00 1.43953249e-01 8.46669614e-01
-9.89929438e-01 -1.30891576e-01 -6.06536455e-02 6.95998490e-01
-2.10389391e-01 2.55215466e-01 1.66616663e-01 -9.07096326e-01
-4.07895029e-01 -8.27020407e-01 6.87393248e-01 4.22200829e-01
4.27142441e-01 -1.02071249e+00 1.65829271e-01 8.24524686e-02
5.52175403e-01 5.78524888e-01 5.10406792e-01 -4.07289445e-01
-6.73703134e-01 -3.41823101e-01 -2.54546642e-01 9.33228374e-01
4.38967675e-01 -9.62731168e-02 -1.03780961e+00 -5.80975115e-01
6.84152618e-02 5.29191904e-02 7.03858495e-01 1.78735480e-01
9.08913732e-01 -2.79792726e-01 -1.20092526e-01 6.89872921e-01
1.89188075e+00 3.44595581e-01 7.80274570e-01 5.84974170e-01
4.30747002e-01 5.18387437e-01 8.13423932e-01 3.58794689e-01
-6.89915568e-02 6.66463137e-01 7.96688080e-01 -4.49418783e-01
1.85073942e-01 9.24921334e-02 3.21307033e-01 4.74237084e-01
-4.58326101e-01 -2.03664690e-01 -8.86614323e-01 7.19483316e-01
-1.63498235e+00 -9.77931082e-01 -1.92489952e-01 2.41736555e+00
5.73693752e-01 -2.67816484e-01 -2.82640755e-01 1.61532432e-01
7.99062967e-01 3.42252344e-01 -5.92952847e-01 -1.11486979e-01
-4.63587910e-01 4.34454739e-01 1.00284696e+00 3.47556740e-01
-1.18510830e+00 7.66314268e-01 4.97981119e+00 6.41984046e-01
-1.64139116e+00 1.89581990e-01 1.20403051e-01 3.92984211e-01
-4.39739525e-01 1.08188484e-02 -2.93280363e-01 4.96073723e-01
8.28051925e-01 -1.38135925e-01 4.44494307e-01 7.71384835e-01
3.61113966e-01 -2.45485649e-01 -4.08272982e-01 7.87864029e-01
-9.35810953e-02 -1.07011068e+00 1.88692566e-02 -8.50350223e-03
9.79682505e-01 1.38588220e-01 7.88145512e-02 -9.44614038e-02
-3.22265178e-02 -6.48760080e-01 3.86023402e-01 6.53656423e-01
7.11631000e-01 -8.25369179e-01 1.01118290e+00 4.01894391e-01
-1.02369082e+00 -1.00915805e-02 -4.60188031e-01 1.41915968e-02
-2.50254303e-01 6.51459575e-01 -7.76930749e-01 1.10341394e+00
7.03757048e-01 4.36337948e-01 -5.00766218e-01 1.00980794e+00
-3.36999744e-01 6.57046914e-01 -4.22573537e-01 2.29386687e-01
3.55069607e-01 -6.77796364e-01 5.98274350e-01 1.03833759e+00
8.09311807e-01 1.91785559e-01 -1.72970697e-01 7.64339924e-01
9.45756882e-02 1.78565249e-01 -9.31411147e-01 -4.40927260e-02
4.78123605e-01 1.08354986e+00 -3.55624110e-01 -1.44676372e-01
-3.91192228e-01 9.62009907e-01 -3.10050309e-01 4.43085045e-01
-8.16385567e-01 -2.62600750e-01 7.51584053e-01 2.62241900e-01
3.63575727e-01 -2.15902328e-01 -1.58941001e-01 -9.04679954e-01
8.13223049e-02 -8.56030524e-01 1.11611120e-01 -8.83907855e-01
-1.15603852e+00 6.51298523e-01 2.33286321e-02 -1.94951487e+00
-3.84582311e-01 -4.32199329e-01 -4.90428537e-01 1.28965545e+00
-1.88878179e+00 -1.25679338e+00 -8.82153988e-01 8.02633047e-01
2.50727939e-03 -2.79257983e-01 8.11756432e-01 2.46118918e-01
-1.30347475e-01 3.07490379e-01 7.28821874e-01 4.35175486e-02
5.77019334e-01 -7.28488982e-01 1.70412421e-01 1.26624918e+00
-3.50885838e-02 2.59815268e-02 7.56894350e-01 -4.43105429e-01
-1.05415559e+00 -1.65639770e+00 5.95791817e-01 4.32376236e-01
4.23381925e-01 6.70729801e-02 -1.21830714e+00 4.16834205e-01
-4.07137722e-02 2.16004208e-01 5.16734660e-01 -4.99774516e-01
-2.13101938e-01 -7.15121090e-01 -1.33042669e+00 3.31969142e-01
3.71931046e-01 -7.53555655e-01 -3.55963886e-01 4.29579884e-01
3.95825535e-01 -1.33727431e-01 -8.79450560e-01 4.90645617e-01
3.62200081e-01 -9.48076069e-01 7.63217926e-01 -2.64130414e-01
2.38973454e-01 -7.51906753e-01 -4.09248441e-01 -1.71325600e+00
-2.29442656e-01 -1.18369006e-01 5.10277987e-01 1.14774323e+00
2.44734570e-01 -9.78029251e-01 5.07064104e-01 3.33664626e-01
-1.49530647e-02 2.01494351e-01 -8.93075764e-01 -1.15779877e+00
1.78192958e-01 -1.55745000e-01 6.88177407e-01 1.11349523e+00
-7.48211324e-01 -7.76671171e-02 -3.07633609e-01 9.81100500e-01
6.00841701e-01 3.05349022e-01 9.50283110e-01 -1.47981441e+00
-2.37989336e-01 -4.18955497e-02 -4.45875019e-01 -4.14602965e-01
1.21312058e-02 -7.71719992e-01 2.79379711e-02 -1.02801788e+00
-1.89725950e-01 -7.04976201e-01 -2.66468912e-01 4.23777670e-01
-1.74593359e-01 4.99814481e-01 3.31622094e-01 3.46653193e-01
4.79557216e-01 7.43911207e-01 1.05341399e+00 -4.21050280e-01
-3.02987814e-01 3.08001339e-01 -1.74141526e-01 4.65167165e-01
1.13810718e+00 -7.09914088e-01 -2.70547599e-01 -3.40660691e-01
1.60948917e-01 -3.00075170e-02 7.26986885e-01 -1.43363953e+00
-1.77277178e-02 -2.73100078e-01 2.49013811e-01 -3.97025496e-01
1.92673981e-01 -1.40473330e+00 7.48748124e-01 5.42114437e-01
3.22939366e-01 -3.56778890e-01 4.74494755e-01 4.41093773e-01
-6.65723562e-01 -4.83818650e-01 1.03050649e+00 6.56194687e-02
-8.45299900e-01 1.95632875e-01 -3.67301077e-01 -5.01659691e-01
9.86679673e-01 -3.29312176e-01 9.23400670e-02 -3.54613215e-01
-4.96459424e-01 -1.12749748e-01 5.70545614e-01 2.08313778e-01
2.95531690e-01 -1.24621010e+00 -9.29013729e-01 2.65038520e-01
5.01722276e-01 1.44895002e-01 4.63613808e-01 5.37057996e-01
-5.54964066e-01 1.48735523e-01 -4.80326533e-01 -7.19080627e-01
-1.12018359e+00 3.96455735e-01 5.89985847e-01 -1.64640754e-01
-1.00486502e-01 3.09868842e-01 -1.04317851e-01 -4.45163697e-01
-4.00631458e-01 -2.73129135e-01 -8.67669061e-02 2.08737776e-01
2.41767287e-01 9.99197289e-02 4.25899953e-01 -6.71656966e-01
-2.61123598e-01 6.41289234e-01 3.93513948e-01 -8.89303535e-02
1.35090220e+00 -8.80099311e-02 -4.35125604e-02 2.01610550e-01
9.50869560e-01 3.91018912e-02 -1.25200856e+00 -5.51318169e-01
-2.48249799e-01 -5.96604228e-01 4.24966723e-01 -6.57764137e-01
-1.24948990e+00 7.43453205e-01 1.13123882e+00 2.39175692e-01
1.68642640e+00 -4.55591619e-01 3.56200993e-01 3.80219847e-01
4.03546154e-01 -9.28102911e-01 -5.70581377e-01 2.23608613e-01
8.95224035e-01 -1.29159498e+00 7.85075221e-03 -3.35538298e-01
-5.51747978e-01 9.21865344e-01 2.36363411e-01 -6.32217675e-02
3.48077565e-01 -9.79167819e-02 2.07481772e-01 1.05298385e-01
1.41255721e-01 -4.62827235e-01 1.85762405e-01 9.15390193e-01
2.86414493e-02 4.08561468e-01 -3.70811671e-01 1.90824911e-01
2.02198625e-01 6.77874908e-02 4.32328910e-01 7.59122550e-01
-3.75235200e-01 -1.10095203e+00 -9.40238714e-01 1.57553688e-01
-4.38288860e-02 -1.25168338e-01 -1.37816176e-01 9.14417565e-01
1.91511825e-01 8.40914726e-01 1.32773906e-01 -2.90878028e-01
2.26237923e-01 2.02419870e-02 1.34488672e-01 -2.30844826e-01
-4.13327903e-01 -8.90488997e-02 -2.63552189e-01 -8.26397240e-02
-9.58531737e-01 -7.76501119e-01 -8.69833112e-01 -3.09478760e-01
-3.09983760e-01 2.80533969e-01 8.78768384e-01 8.29933643e-01
3.23689550e-01 2.64780849e-01 1.15573931e+00 -9.21790779e-01
-6.48309290e-01 -1.05698478e+00 -9.15748179e-01 2.60942191e-01
3.92962575e-01 -6.16101205e-01 -3.25008392e-01 2.33658105e-01] | [9.923925399780273, -1.8303279876708984] |
63258a8e-9452-49a0-887e-f1901af9c5d6 | how-can-subgroup-discovery-help-aiops | 2109.04909 | null | https://arxiv.org/abs/2109.04909v1 | https://arxiv.org/pdf/2109.04909v1.pdf | How Can Subgroup Discovery Help AIOps? | The genuine supervision of modern IT systems brings new challenges as it requires higher standards of scalability, reliability and efficiency when analysing and monitoring big data streams. Rule-based inference engines are a key component of maintenance systems in detecting anomalies and automating their resolution. However, they remain confined to simple and general rules and cannot handle the huge amount of data, nor the large number of alerts raised by IT systems, a lesson learned from expert systems era. Artificial Intelligence for Operation Systems (AIOps) proposes to take advantage of advanced analytics and machine learning on big data to improve and automate every step of supervision systems and aid incident management in detecting outages, identifying root causes and applying appropriate healing actions. Nevertheless, the best AIOps techniques rely on opaque models, strongly limiting their adoption. As a part of this PhD thesis, we study how Subgroup Discovery can help AIOps. This promising data mining technique offers possibilities to extract interesting hypothesis from data and understand the underlying process behind predictive models. To ensure relevancy of our propositions, this project involves both data mining researchers and practitioners from Infologic, a French software editor. | ['Youcef Remil'] | 2021-09-10 | null | null | null | null | ['subgroup-discovery'] | ['methodology'] | [ 1.82171494e-01 2.65155017e-01 -1.61203846e-01 -1.03845306e-01
3.37120354e-01 -1.83632046e-01 3.56812418e-01 8.25322866e-01
4.60122935e-02 6.26214981e-01 -2.31928095e-01 -6.17853880e-01
-9.35518146e-01 -9.34072316e-01 -6.96722195e-02 -4.80770528e-01
-4.22900319e-01 6.41404152e-01 3.72105777e-01 -4.01309252e-01
4.53514278e-01 9.84581292e-01 -2.01991892e+00 5.81681728e-01
7.20967233e-01 1.24406660e+00 -3.01967580e-02 6.29586339e-01
-9.04502869e-02 1.10208988e+00 -9.12031114e-01 9.67746153e-02
1.50760889e-01 -4.23485965e-01 -6.80804193e-01 8.86612162e-02
-3.39582980e-01 1.59467518e-01 1.16810419e-01 5.34098446e-01
2.16114566e-01 -2.71368235e-01 2.03495070e-01 -1.68014526e+00
1.66358531e-01 3.96003723e-01 -2.73624539e-01 7.27567136e-01
3.90133321e-01 5.32305129e-02 7.00046480e-01 -3.87188017e-01
3.93155396e-01 7.95630574e-01 5.21290243e-01 -1.89668551e-01
-1.02987802e+00 -3.86878878e-01 1.53810933e-01 7.68686473e-01
-1.15046680e+00 -4.36127990e-01 4.18391109e-01 -4.35568213e-01
1.26049972e+00 6.77204788e-01 5.54096162e-01 6.35393560e-01
2.22661763e-01 2.59023547e-01 9.21004474e-01 -7.15472877e-01
4.80116397e-01 5.18728256e-01 4.82421219e-01 2.62575567e-01
5.29714644e-01 6.23448566e-02 -5.06636679e-01 -1.58847615e-01
2.42508754e-01 4.48393017e-01 -2.17545047e-01 1.50508925e-01
-8.63591671e-01 6.65117741e-01 -1.68198720e-01 8.12433660e-01
-9.44851875e-01 -5.86269557e-01 4.23303872e-01 9.03919697e-01
9.21251997e-02 5.77949226e-01 -6.59163296e-01 -4.33252901e-01
-6.54385448e-01 -1.69301368e-02 1.15936363e+00 8.05154204e-01
4.06320572e-01 1.77155390e-01 1.14405379e-01 4.87025052e-01
-1.09740170e-02 2.32308686e-01 2.93662727e-01 -8.65640342e-01
1.21816754e-01 1.32076716e+00 -2.36094192e-01 -1.20341527e+00
-7.51417756e-01 -4.78737682e-01 -8.52940679e-01 3.73169452e-01
9.62615311e-02 3.25420462e-02 -3.61104399e-01 7.45615780e-01
2.42341340e-01 -1.57713324e-01 -1.65273607e-01 4.59355861e-01
-8.83740839e-03 4.90542799e-01 -2.67270714e-01 -8.69586408e-01
1.49108398e+00 -7.79862627e-02 -1.02123570e+00 -1.03711262e-01
6.76413298e-01 -5.00730395e-01 6.64779186e-01 1.26590192e+00
-9.22837198e-01 -4.54581141e-01 -9.83975708e-01 7.90760815e-01
-6.34826064e-01 -2.30246067e-01 6.09927058e-01 5.52938163e-01
-6.62209332e-01 6.90082848e-01 -7.46509314e-01 -5.08201897e-01
3.90472084e-01 3.76268446e-01 -1.61691517e-01 -1.38321564e-01
-1.07366121e+00 1.12388086e+00 5.06855011e-01 2.28429865e-02
-2.66703874e-01 -6.00665808e-01 -3.54753375e-01 1.16351873e-01
7.86646783e-01 -2.37319842e-01 8.97628188e-01 -4.78744447e-01
-8.58194709e-01 3.42439234e-01 -3.00004613e-04 -8.45786691e-01
-4.51834165e-02 -1.30339071e-01 -9.42522943e-01 1.71939805e-01
-1.95377856e-01 -6.04977071e-01 6.31677270e-01 -9.40340221e-01
-1.06512344e+00 -6.39998496e-01 -2.27737084e-01 -7.09246039e-01
-2.50953466e-01 2.58613676e-01 3.17174941e-01 -1.93136945e-01
-3.21909264e-02 -4.84272718e-01 -2.25272551e-01 -8.20995271e-01
-1.21626452e-01 -1.86213925e-01 1.18401694e+00 -5.62894166e-01
1.82636702e+00 -1.87716579e+00 -1.86426342e-01 7.25556314e-01
2.38055557e-01 5.19096494e-01 6.55668616e-01 9.85383272e-01
-1.51597425e-01 6.81600161e-03 7.42889717e-02 5.38378000e-01
-1.19012721e-01 4.46838915e-01 -5.41346192e-01 8.88173953e-02
3.27705443e-01 2.78245807e-01 -4.49665993e-01 -3.05098802e-01
5.32936990e-01 3.67992409e-02 -2.45032802e-01 3.57840508e-01
-1.55600518e-01 2.24796265e-01 -3.84453565e-01 6.76179945e-01
2.34546572e-01 -2.67801762e-01 1.71934947e-01 -1.63116325e-02
-3.04950684e-01 1.44100547e-01 -1.30391479e+00 8.21740866e-01
-3.70390683e-01 5.02184331e-01 5.65110072e-02 -1.69404840e+00
1.21541989e+00 4.83883470e-01 7.64144480e-01 -9.57052529e-01
8.41206983e-02 1.28547266e-01 2.16685295e-01 -9.71993268e-01
6.26178309e-02 -6.54004663e-02 2.07420513e-01 5.94141304e-01
-1.51512787e-01 2.13355675e-01 6.94151044e-01 1.42024636e-01
1.58951783e+00 -5.70097089e-01 5.57996333e-01 -3.86623442e-02
6.70097113e-01 1.61121666e-01 6.07923090e-01 4.24236387e-01
1.70727432e-01 8.40331241e-02 8.71636033e-01 -8.61921608e-01
-7.27258444e-01 -4.54924405e-01 -2.68695086e-01 7.26593077e-01
-4.28373367e-01 -7.74753988e-01 -3.48777562e-01 -5.45734107e-01
2.03399554e-01 9.76895630e-01 -3.45439732e-01 -3.64068776e-01
-4.00667101e-01 -8.31687152e-01 8.14610720e-02 1.77852005e-01
1.51613638e-01 -1.16944551e+00 -1.07177758e+00 5.66858649e-01
6.99931607e-02 -1.12901068e+00 6.26012385e-01 6.08664334e-01
-9.80873346e-01 -1.52356315e+00 3.34748238e-01 -6.68143630e-02
4.48438287e-01 -8.39237273e-02 1.23468602e+00 5.86986244e-01
-6.73887551e-01 5.10893613e-02 -7.14686573e-01 -1.00446057e+00
-7.93410480e-01 -8.05378109e-02 1.89885020e-01 3.52091365e-03
7.66712487e-01 -9.39294279e-01 -1.89610988e-01 6.47138536e-01
-8.31060469e-01 -4.56832737e-01 8.94511282e-01 3.85337174e-01
1.87066883e-01 8.58229578e-01 1.03984582e+00 -9.98849809e-01
9.12465036e-01 -6.58063173e-01 -6.25406384e-01 3.54730129e-01
-1.48916781e+00 4.07109670e-02 7.31692314e-01 2.08264161e-02
-6.09868228e-01 -5.03447831e-01 1.33435726e-01 -1.72154173e-01
-7.21972346e-01 6.26945913e-01 -9.60570574e-03 3.21619332e-01
7.72913814e-01 -1.09495908e-01 2.92758137e-01 -3.92921895e-01
-3.16967100e-01 9.35597360e-01 2.67214179e-01 -2.07730323e-01
6.64494753e-01 4.26091641e-01 1.33345813e-01 -1.08209586e+00
-7.08059490e-01 -6.17510021e-01 -6.83419704e-01 -4.02560890e-01
4.60336745e-01 -2.88088799e-01 -1.08500886e+00 -1.42786987e-02
-8.41787577e-01 1.78891852e-01 -5.53774416e-01 2.87376970e-01
-1.24962151e-01 3.55601846e-03 -1.71442553e-01 -1.06274819e+00
-2.14136019e-01 -5.02091348e-01 4.84698206e-01 9.16286558e-02
-5.89091778e-01 -7.42367446e-01 9.11069587e-02 5.17850280e-01
6.99524879e-01 5.00527263e-01 9.48051989e-01 -1.00678921e+00
-5.45056164e-01 -8.99026215e-01 1.47892520e-01 3.62389445e-01
2.81823128e-01 2.54449755e-01 -8.46774757e-01 -6.93731457e-02
2.95939058e-01 8.38490203e-02 1.40643954e-01 5.21926954e-02
1.27772141e+00 -4.28477526e-01 -3.41842800e-01 -8.41319840e-03
1.02583086e+00 4.46926087e-01 7.53607750e-01 6.39506042e-01
1.21912606e-01 9.84104872e-01 9.79707181e-01 9.65231121e-01
1.31956443e-01 5.50994456e-01 6.43535316e-01 2.12162241e-01
5.15758216e-01 3.63727182e-01 1.49573222e-01 7.22638905e-01
-4.84293342e-01 3.96724008e-02 -1.15646052e+00 2.69502789e-01
-1.88227916e+00 -1.31571090e+00 -6.85592532e-01 2.27786899e+00
4.25751984e-01 7.94835925e-01 3.31077009e-01 1.17564058e+00
3.49988729e-01 -3.12145710e-01 -6.04871660e-02 -8.50489914e-01
3.54284868e-02 2.67152160e-01 2.49048695e-01 -5.73124327e-02
-5.65884531e-01 1.57175586e-01 5.54754686e+00 3.22817206e-01
-7.99358130e-01 -1.48164406e-01 8.10977146e-02 -2.07424492e-01
1.75655887e-01 1.84763875e-02 -7.34222412e-01 4.25863475e-01
1.68090248e+00 -1.92582875e-01 3.86658877e-01 8.51001918e-01
8.34562182e-01 -3.52236122e-01 -1.02818382e+00 5.80298781e-01
-1.79496914e-01 -1.39261723e+00 -2.20269144e-01 4.32264000e-01
2.25615367e-01 -2.23366082e-01 -6.35164261e-01 2.40783557e-01
-9.92277190e-02 -1.06950510e+00 1.04438037e-01 9.14250731e-01
-1.49447635e-01 -8.74965668e-01 1.16284537e+00 5.56116164e-01
-6.38667285e-01 -7.61157215e-01 -1.02906324e-01 -6.16617918e-01
3.29765290e-01 1.07864535e+00 -9.78238702e-01 7.18144178e-01
9.13820505e-01 5.06480277e-01 -6.27581298e-01 9.36693668e-01
9.67722908e-02 7.98663318e-01 -2.90214598e-01 2.97891498e-01
-1.99237034e-01 -1.64720893e-01 3.40328872e-01 8.56253684e-01
1.35865867e-01 6.62441403e-02 4.99771498e-02 5.52179694e-01
7.54613996e-01 -1.77754760e-01 -6.81925535e-01 -2.85026938e-01
5.07629514e-01 1.22267389e+00 -7.65434921e-01 -1.88009173e-01
-4.03267622e-01 1.29006624e-01 -3.17541301e-01 2.15482041e-02
-2.92610317e-01 -5.38079143e-01 6.71364903e-01 7.54639030e-01
1.22130975e-01 -6.62007928e-02 -5.97950280e-01 -5.49075842e-01
8.43427330e-02 -1.32505727e+00 6.86869562e-01 -4.57123071e-01
-9.46948230e-01 4.71584201e-01 8.57720971e-02 -1.27166581e+00
-5.97349703e-01 -4.85252142e-01 -8.86032879e-01 4.89081055e-01
-1.04926503e+00 -6.77128255e-01 -5.74776351e-01 7.31816113e-01
3.86824340e-01 -3.51127237e-01 9.56291795e-01 3.52111012e-01
-6.32599354e-01 -1.46296937e-02 -2.15873033e-01 -2.74262458e-01
4.07250464e-01 -1.23766577e+00 -1.25362918e-01 8.99063289e-01
-2.79565272e-03 2.87036419e-01 1.00198376e+00 -4.89437640e-01
-1.44260359e+00 -6.16101265e-01 9.41250205e-01 -5.44721961e-01
8.39780331e-01 -9.50490087e-02 -1.29272676e+00 2.97557682e-01
1.39810473e-01 -1.28131509e-01 8.78257096e-01 3.29222679e-01
2.72317678e-01 -5.52938461e-01 -9.03497994e-01 4.86165024e-02
5.28984368e-01 -1.52158141e-01 -7.66214073e-01 3.39888960e-01
3.42430145e-01 2.64611423e-01 -1.35757411e+00 4.32796150e-01
8.83093923e-02 -1.32326519e+00 7.03761876e-01 -7.85105228e-01
1.05866797e-01 -2.96113700e-01 1.65058702e-01 -9.34004068e-01
-8.24245214e-02 -6.99407220e-01 -4.77330953e-01 1.18486476e+00
3.31605077e-01 -7.32331157e-01 4.46227223e-01 5.35652757e-01
-2.92221874e-01 -6.97220027e-01 -6.75564349e-01 -7.07979620e-01
-7.13416517e-01 -9.27826881e-01 6.23717487e-01 1.03839195e+00
4.49615240e-01 2.49384776e-01 -2.06670091e-01 4.09518749e-01
6.70253634e-01 2.47583948e-02 8.55330288e-01 -1.90036595e+00
-1.57878652e-01 -4.17155385e-01 -8.92560601e-01 3.73002321e-01
-4.25823629e-01 -3.48875463e-01 -5.48691690e-01 -1.47238719e+00
-5.52178979e-01 -4.20291275e-01 -3.36866707e-01 5.42619586e-01
3.20602804e-01 -1.83094814e-01 -1.32976443e-01 2.47839734e-01
-4.18304145e-01 -2.85120588e-02 5.15146494e-01 1.61287963e-01
-3.98233682e-01 5.47336578e-01 -6.52550042e-01 6.43539250e-01
1.00790930e+00 -4.85516459e-01 -4.68147576e-01 2.65127063e-01
7.08067775e-01 8.36202502e-02 3.77080202e-01 -1.20634377e+00
4.13327336e-01 -3.65277380e-01 1.58410817e-01 -6.73747897e-01
-2.79048890e-01 -1.40521562e+00 4.04076666e-01 6.88318014e-01
4.63468917e-02 2.38426477e-01 1.12233691e-01 2.94347852e-01
-5.62354207e-01 -2.10563257e-01 3.75237495e-01 5.11731617e-02
-6.41859114e-01 -2.48375051e-02 -6.02327943e-01 -6.41458407e-02
1.52961838e+00 -3.74343187e-01 -4.23724562e-01 -1.49507388e-01
-8.25072289e-01 4.88674641e-01 2.02068940e-01 5.49204350e-01
5.56460977e-01 -8.33187103e-01 -5.79118252e-01 6.04531527e-01
2.45360821e-01 -3.11703533e-02 8.41276571e-02 1.47329676e+00
-4.43636149e-01 5.03186882e-01 -2.66969055e-01 -6.21129692e-01
-1.36206865e+00 8.75445783e-01 1.84184965e-02 -2.11984083e-01
-7.91702092e-01 1.14107020e-01 -7.83985436e-01 1.90671697e-01
3.45018983e-01 -2.37152785e-01 -5.15984178e-01 4.28471893e-01
8.99782419e-01 8.76667440e-01 6.89609349e-01 3.05542611e-02
-5.85477829e-01 7.54689723e-02 -1.76443920e-01 4.09990579e-01
1.73446596e+00 -1.30871475e-01 -3.74121964e-01 7.46762991e-01
4.48496312e-01 -2.16786683e-01 -5.58356822e-01 2.92033646e-02
8.92875791e-01 -5.34340441e-01 1.50467474e-02 -9.09185767e-01
-1.00979817e+00 9.62254763e-01 3.77775699e-01 1.35098064e+00
1.38289201e+00 9.65395123e-02 2.76712894e-01 5.11737943e-01
3.63433421e-01 -1.40309381e+00 -1.89418659e-01 1.47176772e-01
9.48481441e-01 -1.09868979e+00 1.63811103e-01 -1.34752601e-01
-7.11699426e-01 1.35691273e+00 2.97703624e-01 2.82680392e-01
8.82537901e-01 6.16482556e-01 -1.33693770e-01 -6.69596493e-01
-1.31295967e+00 -1.83371246e-01 -1.79321676e-01 7.89794564e-01
2.35720739e-01 -2.27242142e-01 -3.25425863e-01 6.34852707e-01
-1.87773317e-01 3.21960986e-01 5.99213600e-01 1.17256939e+00
-8.67119014e-01 -1.15286422e+00 -5.74514866e-01 9.14727509e-01
-4.39228237e-01 3.30496371e-01 -4.30160105e-01 8.64968896e-01
2.52391785e-01 1.26565588e+00 -2.18862370e-02 -6.09240472e-01
7.84949124e-01 2.70885348e-01 -1.74818665e-01 -3.93018067e-01
-6.44281805e-01 -3.46830577e-01 2.93258339e-01 -7.05025017e-01
-2.95075834e-01 -7.25398242e-01 -1.12030005e+00 -4.88028556e-01
-4.24748838e-01 6.31544828e-01 8.89865875e-01 1.16026247e+00
6.36108279e-01 9.74741220e-01 9.19610202e-01 -2.07930371e-01
-4.55327898e-01 -1.14185345e+00 -7.14540660e-01 3.29618901e-01
5.02629541e-02 -5.86485803e-01 -4.86386359e-01 9.93607864e-02] | [8.458593368530273, 5.935868263244629] |
d6c6f0a8-600e-45c2-b685-57645f2b01a5 | inside-out-visual-place-recognition | 2111.13546 | null | https://arxiv.org/abs/2111.13546v1 | https://arxiv.org/pdf/2111.13546v1.pdf | Inside Out Visual Place Recognition | Visual Place Recognition (VPR) is generally concerned with localizing outdoor images. However, localizing indoor scenes that contain part of an outdoor scene can be of large value for a wide range of applications. In this paper, we introduce Inside Out Visual Place Recognition (IOVPR), a task aiming to localize images based on outdoor scenes visible through windows. For this task we present the new large-scale dataset Amsterdam-XXXL, with images taken in Amsterdam, that consists of 6.4 million panoramic street-view images and 1000 user-generated indoor queries. Additionally, we introduce a new training protocol Inside Out Data Augmentation to adapt Visual Place Recognition methods for localizing indoor images, demonstrating the potential of Inside Out Visual Place Recognition. We empirically show the benefits of our proposed data augmentation scheme on a smaller scale, whilst demonstrating the difficulty of this large-scale dataset for existing methods. With this new task we aim to encourage development of methods for IOVPR. The dataset and code are available for research purposes at https://github.com/saibr/IOVPR | ['Marcel Worring', 'Tim Alpherts', 'Nanne van Noord', 'Sarah Ibrahimi'] | 2021-11-26 | null | null | null | null | ['visual-place-recognition'] | ['computer-vision'] | [ 1.68856621e-01 -3.62255871e-01 -1.21861277e-02 -4.12246078e-01
-7.01387823e-01 -7.77180791e-01 7.74968743e-01 2.20065013e-01
-5.92126250e-01 6.44183338e-01 2.68295854e-01 -2.98515141e-01
2.31919765e-01 -7.87337184e-01 -1.08719838e+00 -4.50499266e-01
-8.98116305e-02 -6.66038543e-02 1.42609522e-01 -1.76580533e-01
3.58937562e-01 9.12048638e-01 -1.64995325e+00 2.06127331e-01
3.87645870e-01 7.09823728e-01 6.49981439e-01 7.51572311e-01
3.07749778e-01 5.20419717e-01 -5.54754496e-01 4.35579047e-02
3.77224505e-01 -1.34022206e-01 -7.01453149e-01 2.21672341e-01
8.16959023e-01 -1.13250248e-01 -5.08987308e-01 7.74879277e-01
5.03866136e-01 7.61192977e-01 2.58225352e-01 -1.38771057e+00
-9.26534355e-01 -2.51276463e-01 -1.53764054e-01 3.51529539e-01
8.02769363e-01 1.64133996e-01 8.30514073e-01 -1.03616917e+00
8.66072953e-01 8.52519631e-01 4.67841089e-01 3.37988913e-01
-1.15408731e+00 -2.18864366e-01 3.03907186e-01 2.70057827e-01
-1.81583977e+00 -6.40908122e-01 9.55129921e-01 -2.76734442e-01
1.01702654e+00 6.69758439e-01 4.24689621e-01 1.31614304e+00
-2.57899910e-01 7.16720581e-01 1.27451384e+00 -5.73876977e-01
4.24874932e-01 2.02175885e-01 -3.20854247e-01 4.93831366e-01
-8.26197192e-02 -7.85859600e-02 -6.41628265e-01 -4.39602062e-02
9.41717982e-01 4.08664286e-01 -5.07546008e-01 -7.51362503e-01
-1.41703045e+00 6.10535443e-01 1.07916045e+00 2.99952507e-01
-1.88150853e-01 1.78140730e-01 -5.82516305e-02 4.95807566e-02
3.45852166e-01 4.78186756e-01 -1.57318085e-01 -1.03299312e-01
-8.74247253e-01 -2.11384688e-02 4.57159281e-01 1.07412338e+00
1.01977932e+00 -3.75075400e-01 -1.11262418e-01 1.08686590e+00
3.90878290e-01 6.44591153e-01 5.53493083e-01 -1.05163872e+00
5.24140000e-01 5.37970483e-01 3.21357220e-01 -1.12701917e+00
-9.51184705e-02 6.99507073e-02 -6.13473058e-01 1.02695182e-01
2.50725776e-01 3.86689663e-01 -1.08867776e+00 1.50866222e+00
3.29094589e-01 2.18025520e-01 1.48156315e-01 8.51180732e-01
8.92957985e-01 8.96807671e-01 -1.57479838e-01 5.05217433e-01
1.12299764e+00 -1.21519494e+00 -3.75510275e-01 -7.25438118e-01
4.22329009e-01 -8.36747825e-01 1.30313432e+00 -2.98618805e-02
-5.29789567e-01 -5.29098451e-01 -1.01464057e+00 -5.36279321e-01
-1.02660441e+00 3.09598178e-01 5.02039254e-01 6.32569015e-01
-1.39367890e+00 1.42471254e-01 -6.66754544e-01 -8.74833703e-01
3.16941440e-01 3.06526959e-01 -9.13908660e-01 -4.57720250e-01
-7.02072978e-01 5.96133828e-01 2.17297092e-01 1.87475547e-01
-8.31890404e-01 -3.11379135e-01 -1.41092134e+00 -1.62608996e-01
1.85475320e-01 -2.44716778e-01 8.49952221e-01 -5.32521844e-01
-9.60474014e-01 1.06700802e+00 -6.63527369e-01 -4.12132889e-01
4.14755911e-01 -1.94181487e-01 -5.15705049e-01 9.96851921e-02
4.71079767e-01 9.95989919e-01 4.65654582e-01 -1.51850832e+00
-5.60989261e-01 -4.22933251e-01 1.86207667e-01 2.86900669e-01
-4.30519581e-02 -3.09558541e-01 -8.31636727e-01 -6.00153089e-01
2.29607463e-01 -9.80413735e-01 -4.15395111e-01 -1.00878235e-02
-4.02660042e-01 1.78084269e-01 6.95867896e-01 -5.82578361e-01
8.69727671e-01 -2.40309310e+00 -3.57916832e-01 3.05972517e-01
-1.29767537e-01 3.66748497e-02 -2.73133069e-01 6.54786170e-01
-1.13159552e-01 1.56258628e-01 -3.37225914e-01 -6.02785647e-01
-1.05278626e-01 3.62589329e-01 -3.89523655e-01 6.98741078e-01
-7.83910379e-02 9.03255582e-01 -8.82564843e-01 -1.82055071e-01
7.11483657e-01 5.46651661e-01 -4.87416476e-01 1.94149002e-01
8.53028297e-02 6.49299324e-01 -1.64316863e-01 9.48107123e-01
6.53114676e-01 -1.70775682e-01 -2.46440679e-01 1.48850173e-01
-3.73025477e-01 -2.66063269e-02 -1.18715644e+00 2.01033616e+00
-6.27460957e-01 1.07853723e+00 -1.77768916e-01 -5.15934944e-01
1.02821052e+00 -5.30369841e-02 2.20232502e-01 -9.79188979e-01
-2.62765586e-01 1.52093694e-01 -6.27882004e-01 -3.00558239e-01
8.33849430e-01 3.49700391e-01 -9.94818583e-02 -5.41435406e-02
2.03381665e-02 -4.98902500e-02 1.47280812e-01 -5.49200140e-02
1.10778701e+00 6.51680902e-02 5.20216048e-01 2.40031984e-02
6.05842292e-01 3.93447876e-02 9.50291380e-02 9.57690239e-01
-3.89006078e-01 1.16129625e+00 -1.10421531e-01 -5.44183433e-01
-1.04283738e+00 -1.19006062e+00 -1.84615120e-01 1.01046193e+00
3.21322590e-01 -4.64103550e-01 -3.27703387e-01 -5.58967888e-01
-8.74632522e-02 5.59652686e-01 -7.58240402e-01 3.04457575e-01
-5.14157951e-01 -2.55323410e-01 3.82110596e-01 5.28819978e-01
7.49961495e-01 -1.14469635e+00 -6.52539372e-01 -2.78284639e-01
-4.57016528e-01 -1.44238579e+00 -3.26916963e-01 1.37206942e-01
-4.94669676e-01 -9.76951659e-01 -1.06331623e+00 -9.53687370e-01
9.43902969e-01 7.49047697e-01 9.31595683e-01 -9.45584327e-02
-3.45124215e-01 9.72711146e-01 -4.29285049e-01 -1.58327624e-01
7.53429607e-02 2.07679540e-01 -4.72482145e-02 1.37507290e-01
3.43954563e-01 -5.34530818e-01 -7.55722463e-01 5.37706017e-01
-8.78826797e-01 -1.18221186e-01 3.11032981e-01 4.78248835e-01
8.47221315e-01 -3.96971226e-01 -2.44202644e-01 -5.23335516e-01
2.65464514e-01 -4.71911967e-01 -7.12047875e-01 2.78939545e-01
-1.39925228e-02 -2.66040266e-01 4.95141715e-01 -1.61216274e-01
-6.50015414e-01 3.69146973e-01 -3.50822926e-01 -4.76890951e-01
-7.24177241e-01 2.57811788e-02 -2.80102551e-01 -3.95643383e-01
8.22198331e-01 6.00838184e-01 -4.03332889e-01 -3.87119263e-01
5.30395985e-01 7.75202751e-01 6.13430440e-01 -4.27909672e-01
9.10420835e-01 9.05173779e-01 -1.02587238e-01 -1.26365781e+00
-6.17981613e-01 -1.01438820e+00 -8.90195489e-01 -1.52766898e-01
8.96898806e-01 -1.13220775e+00 -4.44284499e-01 1.39393896e-01
-8.87148857e-01 -5.64156592e-01 -4.63509142e-01 3.44014615e-01
-6.33753717e-01 1.96329325e-01 -4.85174321e-02 -7.45193362e-01
1.98733196e-01 -1.04712617e+00 1.43536603e+00 2.97766477e-01
-4.96982932e-02 -1.11833072e+00 3.51978511e-01 5.63237071e-01
2.34992579e-01 4.17914391e-01 4.36897688e-02 -5.18256843e-01
-9.52480733e-01 -1.98029205e-01 -1.01557918e-01 3.11774015e-01
2.51227528e-01 -4.39001650e-01 -1.35017073e+00 -3.56816292e-01
-3.10771137e-01 -1.87166080e-01 7.50403464e-01 1.03477292e-01
1.21857297e+00 -2.10433349e-01 -3.30538720e-01 7.98973918e-01
1.68437290e+00 2.75930643e-01 8.34061563e-01 8.72970462e-01
9.58149135e-01 2.97674567e-01 8.08135569e-01 4.14609462e-01
4.29242939e-01 8.32765996e-01 5.42596817e-01 -4.48517412e-01
-1.87610343e-01 -5.50264657e-01 9.38041061e-02 3.04897547e-01
-1.61268547e-01 -2.88781226e-01 -1.24299014e+00 1.05601203e+00
-1.77660847e+00 -8.56384635e-01 9.54302549e-02 2.35374904e+00
3.34023684e-01 -3.65827531e-01 -2.11097524e-01 -5.56834377e-02
5.49000621e-01 6.28967345e-01 -1.90664247e-01 -1.42900839e-01
-1.57457829e-01 5.80018498e-02 7.61128128e-01 6.49103463e-01
-1.44438958e+00 1.04868114e+00 5.60859728e+00 5.89855075e-01
-1.09590280e+00 8.15240759e-03 5.99623501e-01 3.40714343e-02
2.27613118e-03 -6.96023554e-02 -7.54377782e-01 3.09302568e-01
7.05517173e-01 4.72051948e-01 7.25538850e-01 1.27943480e+00
1.00864969e-01 -3.54359239e-01 -1.06733572e+00 1.39169121e+00
3.54130208e-01 -1.32716596e+00 -1.63393259e-01 2.04412475e-01
8.31259191e-01 3.38772625e-01 2.86751896e-01 1.19582370e-01
-8.97093862e-02 -1.14433861e+00 6.63309157e-01 2.31090099e-01
6.67396605e-01 -5.65383613e-01 4.41013575e-01 2.39667609e-01
-1.33820152e+00 -1.32083833e-01 -5.03920019e-01 -6.96574599e-02
2.37411223e-02 1.49325728e-01 -1.12114155e+00 4.27449912e-01
9.75503623e-01 8.97578835e-01 -1.26615417e+00 1.33952141e+00
-3.62839222e-01 1.67967454e-01 -6.00964844e-01 9.23931412e-03
2.98097759e-01 -1.65302992e-01 2.99066126e-01 1.19392025e+00
4.47326273e-01 -2.58853525e-01 9.02394876e-02 7.84504056e-01
-1.28357559e-01 1.18769579e-01 -1.34625268e+00 2.51864761e-01
4.60149288e-01 1.13098705e+00 -8.10716569e-01 -1.27003163e-01
-2.14383125e-01 1.48415029e+00 4.07109290e-01 5.77100873e-01
-7.84796238e-01 -3.55955809e-01 7.70430684e-01 2.25486204e-01
4.60172236e-01 -6.83916390e-01 -8.01848322e-02 -1.30440462e+00
2.06811979e-01 -5.98899901e-01 6.94369599e-02 -1.08874691e+00
-8.30303550e-01 4.36559409e-01 -7.58594126e-02 -1.30815709e+00
9.65941027e-02 -6.41387582e-01 -3.69760454e-01 6.84658706e-01
-1.64341319e+00 -1.41852462e+00 -7.61093795e-01 9.48206484e-01
7.31973231e-01 1.86097220e-01 9.01287615e-01 3.88186038e-01
-3.24298590e-01 3.24766010e-01 5.10161102e-01 2.83908695e-01
7.94726253e-01 -1.20237505e+00 7.12981880e-01 1.10135114e+00
7.36757398e-01 8.32091928e-01 3.79616797e-01 -4.77002621e-01
-1.21296978e+00 -1.36120522e+00 9.80047703e-01 -9.42931473e-01
2.56878644e-01 -7.89813161e-01 -5.14568985e-01 1.01363206e+00
1.01124324e-01 4.47313845e-01 6.97935522e-01 -6.35533035e-02
-2.47430161e-01 2.02940442e-02 -1.24879825e+00 9.70324814e-01
1.21630287e+00 -8.43709946e-01 -3.76110405e-01 5.00874400e-01
4.71516669e-01 -7.05157399e-01 -6.86295092e-01 2.04515010e-01
1.03105493e-01 -1.04808915e+00 1.60201347e+00 6.41785488e-02
2.52021641e-01 -6.35767043e-01 -7.81593382e-01 -1.13839638e+00
-1.86687797e-01 -1.88398093e-01 2.29976580e-01 1.29236758e+00
1.48564622e-01 -7.78414607e-01 6.96523309e-01 4.81972218e-01
8.55398774e-02 -3.47843498e-01 -1.21570313e+00 -7.78032660e-01
-4.85473156e-01 -6.23081326e-01 4.86786366e-01 9.83628392e-01
-3.85644108e-01 -4.54026870e-02 -3.00620824e-01 6.43127561e-01
4.03680891e-01 1.16627507e-01 9.97378349e-01 -6.59133554e-01
1.78250801e-02 1.19575575e-01 -8.28362048e-01 -1.16696930e+00
-7.74763376e-02 -7.84489095e-01 1.15532756e-01 -1.81904328e+00
-1.87494278e-01 -3.84893626e-01 -2.35788181e-01 7.22144604e-01
1.78845838e-01 7.54296839e-01 3.59729141e-01 3.55551004e-01
-9.02406871e-01 6.74903691e-01 8.81394029e-01 -1.54384702e-01
-2.46253014e-01 -9.98055339e-02 -2.95314312e-01 4.65542227e-01
1.11008060e+00 -2.85303891e-01 -3.87192726e-01 -3.04873258e-01
-1.42988637e-01 -4.76387590e-01 8.47490966e-01 -1.43037546e+00
3.45915645e-01 -9.16309953e-02 7.16963887e-01 -6.03879869e-01
8.38780701e-01 -1.01203001e+00 8.18412974e-02 2.20761448e-01
-7.27969259e-02 2.45302781e-01 2.74073154e-01 6.50167525e-01
-3.37333739e-01 -3.74621563e-02 5.62570393e-01 -3.63970131e-01
-1.32188332e+00 4.31733672e-03 -2.66020119e-01 -1.16278186e-01
1.28688037e+00 -4.40337360e-01 -3.17870945e-01 -2.71274000e-01
-7.63199389e-01 3.52592692e-02 1.00377595e+00 7.31190622e-01
9.39912796e-01 -1.40969276e+00 -1.39349177e-01 5.18383861e-01
5.94120979e-01 3.23819406e-02 2.79120862e-01 3.50241244e-01
-9.67776597e-01 7.05851138e-01 -3.19348246e-01 -6.67487562e-01
-1.39626181e+00 7.13359356e-01 4.86229360e-02 1.43523380e-01
-7.15772092e-01 7.75420487e-01 3.33728611e-01 -7.79859722e-01
1.57543391e-01 -4.72445577e-01 -1.41273335e-01 -1.51866391e-01
3.79215449e-01 3.13506275e-01 6.66200370e-02 -9.54095304e-01
-6.29694462e-01 5.09410501e-01 2.96948016e-01 -2.61831194e-01
1.29208779e+00 -3.60087186e-01 -7.18145818e-02 5.43433547e-01
1.40958154e+00 3.25626761e-01 -1.21822715e+00 -2.51199324e-02
-2.73499429e-01 -9.69903767e-01 -2.19827205e-01 -6.43290102e-01
-6.30609214e-01 7.65820324e-01 1.10814071e+00 -3.54503728e-02
1.02653730e+00 8.78067017e-02 4.43567544e-01 6.37061417e-01
7.26779759e-01 -8.54477465e-01 -8.86792317e-02 4.59545463e-01
9.76091504e-01 -1.61247659e+00 -1.39956921e-01 -3.18556279e-01
-5.23660779e-01 8.34895134e-01 4.44588691e-01 -9.66407731e-02
4.21661377e-01 -2.04636171e-01 1.93072945e-01 -8.62395242e-02
5.36260791e-02 -4.08086121e-01 2.96784908e-01 9.55709577e-01
8.69251564e-02 1.34073105e-02 4.25302655e-01 -1.62254199e-01
-3.41238469e-01 -2.49581560e-01 3.96167845e-01 1.26575446e+00
-1.64812803e-01 -1.01695275e+00 -8.47151756e-01 -1.66477919e-01
-4.61464562e-02 -2.45609552e-01 -4.62903708e-01 8.80025983e-01
1.48094073e-01 9.56183851e-01 -5.72260236e-03 -1.13285542e-01
3.77588362e-01 -5.46835586e-02 2.95323759e-01 -5.82723200e-01
-3.06269735e-01 -2.90357262e-01 -1.56809077e-01 -8.54948699e-01
-3.94592553e-01 -5.22657216e-01 -7.54270971e-01 -5.96384052e-03
3.70307207e-01 -1.04408376e-01 9.54177380e-01 4.95020807e-01
5.71435630e-01 3.21766853e-01 4.73102897e-01 -1.24655938e+00
2.47767463e-01 -5.19231021e-01 -4.46972936e-01 4.85923707e-01
6.09109700e-01 -4.13836539e-01 -2.27612764e-01 1.77675933e-01] | [7.604981899261475, -1.8918046951293945] |
3907b64d-cc23-4f9d-b383-bf70367806ce | fine-grained-image-text-matching-by-cross | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Pan_Fine-Grained_Image-Text_Matching_by_Cross-Modal_Hard_Aligning_Network_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Pan_Fine-Grained_Image-Text_Matching_by_Cross-Modal_Hard_Aligning_Network_CVPR_2023_paper.pdf | Fine-Grained Image-Text Matching by Cross-Modal Hard Aligning Network | Current state-of-the-art image-text matching methods implicitly align the visual-semantic fragments, like regions in images and words in sentences, and adopt cross-attention mechanism to discover fine-grained cross-modal semantic correspondence. However, the cross-attention mechanism may bring redundant or irrelevant region-word alignments, degenerating retrieval accuracy and limiting efficiency. Although many researchers have made progress in mining meaningful alignments and thus improving accuracy, the problem of poor efficiency remains unresolved. In this work, we propose to learn fine-grained image-text matching from the perspective of information coding. Specifically, we suggest a coding framework to explain the fragments aligning process, which provides a novel view to reexamine the cross-attention mechanism and analyze the problem of redundant alignments. Based on this framework, a Cross-modal Hard Aligning Network (CHAN) is designed, which comprehensively exploits the most relevant region-word pairs and eliminates all other alignments. Extensive experiments conducted on two public datasets, MS-COCO and Flickr30K, verify that the relevance of the most associated word-region pairs is discriminative enough as an indicator of the image-text similarity, with superior accuracy and efficiency over the state-of-the-art approaches on the bidirectional image and text retrieval tasks. Our code will be available at https://github.com/ppanzx/CHAN. | ['BaiLing Zhang', 'Fangyu Wu', 'Zhengxin Pan'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['semantic-correspondence', 'text-matching'] | ['computer-vision', 'natural-language-processing'] | [ 1.41570941e-01 -4.35140431e-01 -5.28703451e-01 -4.15181905e-01
-7.91593611e-01 -3.79633874e-01 5.24140298e-01 -7.08803441e-03
-4.57310706e-01 1.81587502e-01 5.33433855e-01 -1.04461201e-01
-2.22039342e-01 -5.79063177e-01 -6.90380394e-01 -6.31528914e-01
5.71457207e-01 2.08211944e-01 8.90037343e-02 -1.04446948e-01
4.76741523e-01 2.25875393e-01 -1.60258079e+00 6.67460680e-01
9.21142161e-01 9.93890345e-01 4.64272439e-01 1.52790651e-01
-5.40509939e-01 4.40169841e-01 -1.82845682e-01 -5.07845581e-01
2.81496912e-01 -3.97812575e-01 -9.09080684e-01 -4.31216843e-02
7.52002597e-01 -3.13697100e-01 -7.38453984e-01 1.33059430e+00
4.89513278e-01 -5.55700026e-02 4.47318882e-01 -1.33460164e+00
-1.29476452e+00 4.30441171e-01 -9.09843206e-01 2.62370974e-01
2.55544722e-01 1.17937028e-02 1.37372386e+00 -1.07133687e+00
4.44116026e-01 1.23761952e+00 4.62078780e-01 3.31956148e-01
-8.38797987e-01 -8.60913396e-01 2.70803005e-01 6.04341626e-01
-1.92746699e+00 -2.75898486e-01 7.79837787e-01 -4.03503299e-01
9.54642773e-01 4.09240484e-01 6.25505328e-01 8.63907456e-01
2.13104486e-01 9.53894734e-01 8.41791511e-01 -4.57739085e-01
-4.23414409e-01 -6.07436821e-02 1.71110600e-01 7.80799985e-01
1.59719646e-01 4.62699421e-02 -4.73833054e-01 -2.22236235e-02
6.67296946e-01 3.32972467e-01 -3.63724470e-01 -3.58373493e-01
-1.34324706e+00 7.73859680e-01 7.09299445e-01 7.07023501e-01
-1.46483973e-01 1.48654968e-01 4.49233472e-01 1.40942901e-01
4.82634306e-01 1.08923636e-01 -1.75242469e-01 1.87524751e-01
-8.90809834e-01 7.72931278e-02 8.28284845e-02 1.04383063e+00
9.97609496e-01 -5.47092497e-01 -4.68020827e-01 1.10459650e+00
4.39880162e-01 6.29210591e-01 7.34633625e-01 -5.97595096e-01
6.39424384e-01 8.81455719e-01 -3.03335577e-01 -1.39893854e+00
-9.23838317e-02 -3.43361199e-01 -9.95954752e-01 -3.90635312e-01
9.09343653e-04 6.00036621e-01 -9.66733038e-01 1.49968946e+00
2.54245490e-01 5.42777516e-02 -2.97579110e-01 1.23394084e+00
9.52450156e-01 6.24133885e-01 -2.30234563e-02 1.08797669e-01
1.56694412e+00 -1.32662976e+00 -6.88323975e-01 -3.28414410e-01
4.91440803e-01 -1.08283830e+00 1.33285058e+00 -2.84578204e-01
-9.15107191e-01 -7.92881608e-01 -8.32807481e-01 -5.18858671e-01
-4.44763660e-01 4.84981611e-02 3.95526290e-01 9.05202478e-02
-8.45280230e-01 3.06873500e-01 -4.51374978e-01 -5.00309348e-01
4.11053538e-01 1.79284677e-01 -4.29972798e-01 -4.27762121e-01
-1.28267467e+00 7.44879782e-01 3.36639941e-01 1.88659906e-01
-3.31332475e-01 -6.12936199e-01 -7.18516886e-01 1.78867549e-01
2.65047103e-01 -8.52574468e-01 8.77644122e-01 -1.21555698e+00
-7.46248186e-01 1.20608723e+00 -3.48951310e-01 -9.25506745e-03
4.04006422e-01 -1.92704469e-01 -4.03802216e-01 2.01041460e-01
3.68110031e-01 9.98788357e-01 7.84326434e-01 -1.05923498e+00
-4.92610961e-01 -5.39445877e-01 -1.40662089e-01 4.11807746e-01
-5.43584704e-01 2.00084999e-01 -1.11055756e+00 -9.56481099e-01
2.34232515e-01 -7.90817559e-01 -2.22688988e-02 2.50981241e-01
-3.83885264e-01 -3.49377781e-01 6.63658917e-01 -7.03992844e-01
1.45740521e+00 -2.11018324e+00 2.15933472e-01 1.14733182e-01
1.17062718e-01 1.64835960e-01 -5.54776669e-01 5.19532681e-01
-3.10176015e-01 1.89908028e-01 -1.88046977e-01 -3.15002501e-01
-1.12265917e-02 1.31628096e-01 -4.11251873e-01 4.14018661e-01
4.13703397e-02 1.39596891e+00 -6.36195600e-01 -8.60040665e-01
2.76603371e-01 3.84307712e-01 -4.23636168e-01 2.44835198e-01
-8.01242366e-02 7.75230303e-02 -6.63832724e-01 7.92764843e-01
6.95731640e-01 -5.18747151e-01 -7.93451667e-02 -7.10162461e-01
-3.47039327e-02 4.28750739e-02 -7.67667115e-01 2.10191488e+00
-3.55386674e-01 6.87415242e-01 -2.97276169e-01 -1.16501880e+00
7.37848878e-01 -7.68411234e-02 6.35597169e-01 -1.26999640e+00
1.51902586e-01 1.04388811e-01 -2.74029464e-01 -7.97908068e-01
5.11239171e-01 1.15131080e-01 -1.42418584e-02 4.17867482e-01
-2.14233339e-01 3.49702686e-01 -4.63137329e-02 2.34117135e-01
6.69546127e-01 6.42502680e-02 3.70499700e-01 -3.09190899e-01
6.69086635e-01 8.88931379e-02 2.56068528e-01 5.45165479e-01
-1.62749052e-01 7.67989993e-01 5.08683361e-02 -4.84533906e-01
-1.18919456e+00 -8.02117527e-01 -1.32002860e-01 1.10375047e+00
6.83483064e-01 -5.43798089e-01 -5.98020852e-01 -5.86889148e-01
1.27580374e-01 1.65187404e-01 -7.75633514e-01 -2.80933589e-01
-5.18896043e-01 -5.71852922e-01 4.66052920e-01 4.28646117e-01
5.85379899e-01 -1.02884364e+00 -9.46074203e-02 -6.80557191e-02
-7.53953993e-01 -1.06105244e+00 -1.11190391e+00 -4.46153283e-01
-5.26162982e-01 -1.01299894e+00 -8.33051145e-01 -9.71219420e-01
7.93271065e-01 9.18103516e-01 1.13008773e+00 5.87300539e-01
-5.57529569e-01 2.95463413e-01 -4.85786796e-01 4.93446440e-02
-1.26908915e-02 -1.75065380e-02 -3.64841640e-01 4.34641466e-02
6.18917704e-01 -1.80303365e-01 -8.78544807e-01 4.89726245e-01
-1.02526450e+00 3.34002107e-01 7.69127548e-01 1.01157844e+00
8.71236742e-01 -2.73121834e-01 1.81855142e-01 -3.62175703e-01
4.37178552e-01 -4.00492460e-01 -5.11645317e-01 6.54159844e-01
-6.16560340e-01 1.93870232e-01 4.44285899e-01 -2.95042008e-01
-6.74761832e-01 -6.89212158e-02 -1.51082337e-01 -8.15677702e-01
-8.78539979e-02 5.63625932e-01 -2.57715672e-01 -1.26141652e-01
1.37080550e-01 5.44381559e-01 -1.75763667e-01 -4.29222912e-01
4.24025148e-01 8.39845777e-01 4.48666841e-01 -5.76709867e-01
6.63018942e-01 6.35686517e-01 -3.72284412e-01 -5.03555715e-01
-1.08008468e+00 -9.05774236e-01 -5.68606079e-01 -4.54534851e-02
9.71624494e-01 -1.04201102e+00 -3.49669814e-01 3.69814157e-01
-1.21346521e+00 6.18769489e-02 1.01804018e-01 2.45659575e-01
-4.11385536e-01 8.11244905e-01 -4.13656503e-01 -2.35586807e-01
-5.90578437e-01 -1.22826898e+00 1.44862318e+00 1.20125569e-01
2.32736617e-02 -7.28446841e-01 1.06045350e-01 7.54904866e-01
2.36068666e-01 -4.47921038e-01 9.44314063e-01 -3.78348023e-01
-7.92093575e-01 -6.14301525e-02 -9.02294338e-01 -3.03706042e-02
4.33124900e-02 -1.02670379e-01 -7.70104229e-01 -3.59365582e-01
-4.27815944e-01 -1.58290654e-01 1.02143931e+00 2.76226282e-01
1.66589606e+00 -2.72340536e-01 -4.65813190e-01 7.06694305e-01
1.51350069e+00 -1.24952309e-01 7.24156916e-01 4.33530331e-01
9.81124580e-01 6.14547014e-01 7.70674109e-01 1.61940113e-01
5.53168297e-01 1.12386131e+00 3.73726606e-01 -3.36825043e-01
-2.74472356e-01 -4.39970732e-01 6.56459779e-02 9.48638916e-01
3.76153469e-01 -3.19618970e-01 -7.94091761e-01 6.74271762e-01
-2.14955902e+00 -1.03517175e+00 2.43876334e-02 1.88447642e+00
7.69959569e-01 -3.01709533e-01 -2.07733721e-01 -3.17438394e-01
1.01715374e+00 3.62309426e-01 -5.06062746e-01 -6.21370524e-02
-2.57270575e-01 8.94182324e-02 3.93107623e-01 4.86859530e-01
-1.03294671e+00 1.02789092e+00 5.13047743e+00 1.24997735e+00
-1.09149551e+00 1.92675620e-01 6.05337381e-01 -5.49120419e-02
-5.91522753e-01 -2.79134009e-02 -7.34598875e-01 5.83181858e-01
3.60748768e-01 -5.85820638e-02 4.87750053e-01 5.57614326e-01
6.77061006e-02 2.55219042e-01 -9.12811577e-01 1.29311371e+00
5.14984846e-01 -1.51392448e+00 4.72470343e-01 6.12890497e-02
6.57182932e-01 2.38090321e-01 -3.81148458e-02 4.71170992e-02
-9.80189592e-02 -1.01472020e+00 7.52098680e-01 5.48962474e-01
9.15039301e-01 -5.16282499e-01 7.24869251e-01 5.90728521e-02
-1.52883053e+00 2.72862744e-02 -5.00873804e-01 3.98394406e-01
-2.45130882e-02 5.07543445e-01 -1.91788152e-01 8.02483201e-01
9.70304370e-01 1.01555443e+00 -6.64248407e-01 9.66112792e-01
8.67063254e-02 5.82301393e-02 -5.22991009e-02 1.11781798e-01
4.27926213e-01 -2.22314253e-01 2.52758354e-01 1.25528574e+00
3.54744613e-01 8.40060711e-02 1.67057991e-01 9.15778816e-01
-1.21308789e-01 4.15479630e-01 -4.84202236e-01 -1.90167408e-02
5.46876073e-01 1.22281325e+00 -5.57858288e-01 -3.37591261e-01
-6.73813343e-01 1.28836906e+00 4.89540786e-01 3.42367977e-01
-9.39986885e-01 -2.52353936e-01 7.98357725e-01 -4.50701974e-02
3.71350437e-01 3.97903509e-02 -2.61858761e-01 -1.36520958e+00
2.72807032e-01 -9.28088963e-01 5.68325520e-01 -1.12201989e+00
-1.59268439e+00 4.64201897e-01 -1.09177753e-01 -1.35773742e+00
2.28235930e-01 -3.69477123e-01 -4.00708944e-01 8.57189894e-01
-1.70978892e+00 -1.50125647e+00 -7.06171393e-01 8.05742443e-01
7.06365824e-01 -5.06022908e-02 5.09530663e-01 6.81464314e-01
-5.25863945e-01 9.51583803e-01 1.67092502e-01 3.00058007e-01
9.13785458e-01 -5.27257562e-01 3.49014014e-01 7.80088723e-01
4.42192405e-01 6.57271564e-01 2.73733884e-01 -4.71828669e-01
-1.41867232e+00 -1.15432918e+00 9.27955747e-01 -2.39471927e-01
6.94723427e-01 -2.93131322e-01 -1.08571362e+00 3.45432967e-01
3.62914950e-01 8.84849802e-02 4.79705423e-01 -1.63411975e-01
-7.40550995e-01 -1.66554913e-01 -6.56693637e-01 7.48881400e-01
1.31148565e+00 -9.04609144e-01 -6.53508902e-01 3.59216750e-01
8.86272550e-01 -1.64316565e-01 -7.11329818e-01 4.50657606e-01
8.42995405e-01 -9.05061424e-01 1.22370088e+00 -5.24214149e-01
6.29180610e-01 -4.15959805e-01 -3.70175153e-01 -9.07024145e-01
-5.63411295e-01 -7.04331845e-02 2.94216722e-01 1.23518860e+00
4.35585529e-02 -4.11191612e-01 6.07551038e-01 8.49658102e-02
-2.12190717e-01 -7.85934567e-01 -1.01716089e+00 -5.91180384e-01
1.26510516e-01 -3.42804253e-01 6.25356495e-01 1.14570260e+00
-1.58978492e-01 3.08032304e-01 -4.15977031e-01 7.01015145e-02
5.10442019e-01 6.58036709e-01 5.93486905e-01 -8.15590084e-01
-1.10758193e-01 -8.28956962e-01 -3.51874918e-01 -1.33912504e+00
4.65505958e-01 -1.07172346e+00 4.43703197e-02 -1.52110398e+00
8.76247525e-01 -3.92209023e-01 -4.69087809e-01 5.30694842e-01
-3.80671680e-01 5.08550406e-01 1.45763263e-01 6.23421073e-01
-7.64017522e-01 7.37479568e-01 1.28125322e+00 -4.58297521e-01
3.44885796e-01 -6.86544657e-01 -7.58218527e-01 3.62083018e-01
6.91007257e-01 -3.96902293e-01 -2.38524422e-01 -8.48516703e-01
2.25613847e-01 -2.79876173e-01 5.81524789e-01 -5.54734111e-01
4.49825048e-01 -2.67148465e-01 2.79009581e-01 -8.34646821e-01
1.37894601e-01 -9.27009702e-01 1.87088177e-01 3.54589403e-01
-5.06652713e-01 2.84208953e-01 8.91625062e-02 5.72767794e-01
-5.43378830e-01 -1.29752398e-01 6.50522888e-01 -1.14366889e-01
-9.14340258e-01 6.90032661e-01 1.62877232e-01 1.46177262e-01
8.92993629e-01 -2.98862576e-01 -6.09475911e-01 -1.74467191e-01
-1.88535765e-01 3.35273802e-01 7.12794542e-01 8.85138333e-01
7.42353797e-01 -1.63016689e+00 -7.22342014e-01 3.96584541e-01
7.24061847e-01 -3.29975128e-01 6.37034357e-01 7.73405850e-01
-2.52237320e-01 7.77173996e-01 -2.38254756e-01 -7.11934686e-01
-1.49290609e+00 7.23536074e-01 3.63918185e-01 -2.34001234e-01
-6.11917496e-01 7.76777744e-01 7.12016761e-01 -2.06124157e-01
1.20754428e-01 -8.80919695e-02 -4.65359837e-02 -2.54911417e-03
4.78078455e-01 -1.54922485e-01 7.63901621e-02 -8.97744536e-01
-4.67556804e-01 1.25490522e+00 -2.57060558e-01 2.69695908e-01
8.77971172e-01 -5.52999258e-01 -4.63303417e-01 -8.37155432e-02
1.60283363e+00 -2.58368939e-01 -7.49050021e-01 -5.33918977e-01
-1.37969077e-01 -8.76543045e-01 1.33830801e-01 -5.26203811e-01
-1.28719151e+00 1.03643715e+00 6.96370840e-01 -1.12863451e-01
1.20799756e+00 2.66852856e-01 9.99814510e-01 1.61902025e-01
6.03611954e-02 -9.25956368e-01 3.00913274e-01 4.23083276e-01
1.00857878e+00 -1.44599056e+00 8.81788507e-03 -4.11869884e-01
-4.48194861e-01 9.15652633e-01 7.86127985e-01 7.95491710e-02
5.18584669e-01 -1.43534780e-01 -2.57187039e-02 -3.20948631e-01
-5.93790770e-01 -2.91775018e-01 8.13421786e-01 2.72354126e-01
4.94734854e-01 -1.30526394e-01 -4.60363120e-01 3.45508605e-01
9.15684402e-02 -1.85774311e-01 -1.07344642e-01 6.12648249e-01
-3.76389414e-01 -1.07074296e+00 -2.85389453e-01 3.37494165e-01
-3.66251230e-01 -5.54285944e-01 -4.22189027e-01 6.20891333e-01
1.01853795e-01 6.43235743e-01 3.67654711e-01 -3.00608456e-01
2.09302142e-01 -1.79651797e-01 3.86373192e-01 -2.39297554e-01
-4.08285201e-01 1.68970108e-01 -2.48927891e-01 -8.20109844e-01
-5.88704884e-01 -4.18818980e-01 -1.03912568e+00 -2.90003538e-01
-4.72417653e-01 1.06918514e-02 4.15464193e-01 9.80125725e-01
6.34999573e-01 2.89755851e-01 5.96647620e-01 -5.20685434e-01
-2.89772868e-01 -7.33484387e-01 -1.94860339e-01 6.90074086e-01
2.01542690e-01 -5.76819420e-01 -2.10119605e-01 -5.00135049e-02] | [10.785484313964844, 1.2691649198532104] |
da57e3bc-e039-40c9-8261-a0da6f35dbba | bdanet-multiscale-convolutional-neural | 2105.07364 | null | https://arxiv.org/abs/2105.07364v1 | https://arxiv.org/pdf/2105.07364v1.pdf | BDANet: Multiscale Convolutional Neural Network with Cross-directional Attention for Building Damage Assessment from Satellite Images | Fast and effective responses are required when a natural disaster (e.g., earthquake, hurricane, etc.) strikes. Building damage assessment from satellite imagery is critical before relief effort is deployed. With a pair of pre- and post-disaster satellite images, building damage assessment aims at predicting the extent of damage to buildings. With the powerful ability of feature representation, deep neural networks have been successfully applied to building damage assessment. Most existing works simply concatenate pre- and post-disaster images as input of a deep neural network without considering their correlations. In this paper, we propose a novel two-stage convolutional neural network for Building Damage Assessment, called BDANet. In the first stage, a U-Net is used to extract the locations of buildings. Then the network weights from the first stage are shared in the second stage for building damage assessment. In the second stage, a two-branch multi-scale U-Net is employed as backbone, where pre- and post-disaster images are fed into the network separately. A cross-directional attention module is proposed to explore the correlations between pre- and post-disaster images. Moreover, CutMix data augmentation is exploited to tackle the challenge of difficult classes. The proposed method achieves state-of-the-art performance on a large-scale dataset -- xBD. The code is available at https://github.com/ShaneShen/BDANet-Building-Damage-Assessment. | ['Qian Du', 'Liang Xiao', 'Jianyu Chen', 'Delu Pan', 'Chen Chen', 'Taojiannan Yang', 'Sijie Zhu', 'Yu Shen'] | 2021-05-16 | null | null | null | null | ['2d-semantic-segmentation', 'extracting-buildings-in-remote-sensing-images'] | ['computer-vision', 'miscellaneous'] | [ 2.10876942e-01 -3.39488089e-01 2.84617066e-01 -3.57311189e-01
-8.18265498e-01 6.48079589e-02 4.05262560e-01 4.56802785e-01
-6.32803500e-01 4.66770828e-01 4.87086982e-01 -1.64096579e-01
-1.42569378e-01 -1.47857845e+00 -4.59136367e-01 -9.72901583e-01
-2.33318001e-01 1.69987008e-01 1.50275389e-02 -5.58338046e-01
4.13480178e-02 6.98532283e-01 -1.37058473e+00 1.06075823e-01
7.47009814e-01 1.01008773e+00 4.88514513e-01 5.14633298e-01
1.99908480e-01 4.97831106e-01 -3.86423379e-01 2.16519490e-01
3.63388836e-01 -7.96071216e-02 -9.06367481e-01 1.76981688e-01
-1.29100645e-03 -8.64363313e-01 -6.76242590e-01 9.07787085e-01
9.15349126e-01 2.12158024e-01 4.58433330e-01 -1.02058876e+00
-4.86413002e-01 5.62001824e-01 -8.27064753e-01 6.88271463e-01
1.32705541e-02 1.98424831e-01 8.35233986e-01 -9.92954612e-01
5.28305732e-02 1.10269856e+00 7.32335448e-01 -8.09092820e-02
-9.77284968e-01 -6.49264216e-01 1.40863165e-01 3.19992065e-01
-1.59278488e+00 -9.54554006e-02 9.86546755e-01 -5.28882325e-01
8.59689951e-01 1.20392606e-01 5.80721021e-01 6.06990337e-01
-1.11018166e-01 8.21980596e-01 7.44437516e-01 -1.02353603e-01
-5.63425431e-03 -5.39414346e-01 5.14731742e-02 3.84038329e-01
2.83445511e-02 2.78832819e-02 1.25347031e-02 7.80797228e-02
6.61863089e-01 5.04887581e-01 -3.55927229e-01 3.21903229e-01
-9.59057033e-01 9.11017656e-01 1.32094789e+00 5.79262793e-01
-9.66029167e-01 2.47064121e-02 4.59699601e-01 -5.07157110e-03
7.02873588e-01 1.18621320e-01 -1.50715530e-01 4.11399662e-01
-1.10008168e+00 1.79978818e-01 3.83483656e-02 8.07602033e-02
8.35729301e-01 -3.15550971e-03 -1.20007835e-01 8.62430036e-01
1.60124794e-01 2.83895850e-01 3.45296443e-01 -1.91402823e-01
7.65213013e-01 7.60547817e-01 -6.79700598e-02 -1.47367716e+00
-8.83896708e-01 -6.83502913e-01 -1.59085858e+00 8.11187401e-02
1.16716437e-02 -3.21588844e-01 -9.35391665e-01 1.47824669e+00
5.23017049e-01 2.37457573e-01 -1.67181835e-01 1.23491287e+00
6.50209248e-01 9.64755356e-01 1.56873956e-01 1.31578650e-02
1.14604151e+00 -7.34929621e-01 -3.11360210e-01 -4.25669611e-01
2.91670531e-01 -4.11239535e-01 9.45850849e-01 3.49433310e-02
-7.45883167e-01 -5.05854666e-01 -9.42320347e-01 -4.10427302e-02
-4.09756392e-01 1.46737754e-01 1.56007797e-01 -1.77892074e-01
-9.12894309e-01 4.93446767e-01 -7.77297616e-01 -4.01771992e-01
7.21916914e-01 1.72217458e-01 -2.88982838e-01 -3.45246762e-01
-1.37120354e+00 8.11484993e-01 5.34902155e-01 7.04103470e-01
-1.12983835e+00 -4.73155320e-01 -6.59689128e-01 2.30661005e-01
1.46599680e-01 -3.74037534e-01 8.23310554e-01 -5.29346287e-01
-4.86481667e-01 7.64795482e-01 4.53864962e-01 -2.68024176e-01
4.07324612e-01 -4.13410962e-01 -2.12732062e-01 1.39835671e-01
3.10746789e-01 5.89802980e-01 5.63621163e-01 -1.13819170e+00
-7.81565249e-01 -6.15986407e-01 1.19647138e-01 4.41557407e-01
-5.53685725e-01 4.51434851e-02 -2.21848354e-01 -9.59522605e-01
2.95422375e-01 -5.90943933e-01 -4.61305529e-01 -2.56910056e-01
-6.52561367e-01 -1.34805188e-01 9.95373964e-01 -1.05169213e+00
1.56900752e+00 -2.10242105e+00 2.05903873e-01 8.08485970e-02
3.95917520e-02 4.70346689e-01 -4.01799411e-01 6.74207985e-01
-3.55306983e-01 2.09654152e-01 -7.90789604e-01 -1.27935469e-01
-1.91825137e-01 1.82683080e-01 -3.39895070e-01 4.76500034e-01
4.95260000e-01 6.36473000e-01 -7.08400786e-01 -2.02680647e-01
1.21243492e-01 5.67562342e-01 -2.13146344e-01 3.66869807e-01
1.96578145e-01 5.20132422e-01 -5.05749941e-01 6.32524669e-01
8.83958638e-01 1.82421848e-01 -2.53733665e-01 -1.41588509e-01
-3.47186804e-01 1.75413102e-01 -1.01489377e+00 1.37456810e+00
-3.75412136e-01 3.60319465e-01 1.02533109e-01 -1.34024894e+00
1.07559061e+00 2.99534261e-01 5.37123203e-01 -7.19161034e-01
8.03530663e-02 1.70734838e-01 -1.65913656e-01 -6.67249858e-01
3.53884518e-01 -2.93528438e-01 -2.22736150e-01 6.36963129e-01
-4.20217663e-01 -9.55409408e-02 1.36350915e-01 -1.20502584e-01
1.20277691e+00 -2.74935752e-01 1.27417132e-01 -1.29115075e-01
4.97016102e-01 -1.30218044e-01 5.50354064e-01 2.72576451e-01
-1.87896162e-01 8.32993567e-01 1.45369768e-01 -1.02301109e+00
-9.98014808e-01 -6.62781656e-01 -1.45460721e-02 9.85228777e-01
-1.21927746e-01 1.91547796e-01 -6.73409700e-01 -5.53165674e-01
1.05538800e-01 4.89347279e-01 -6.50717854e-01 -2.12643966e-01
-6.25846863e-01 -1.32887185e+00 7.24500358e-01 7.21964300e-01
1.07447755e+00 -1.26049805e+00 -7.46807396e-01 5.06418347e-01
-4.03080255e-01 -5.60869396e-01 -2.11086124e-01 2.30059460e-01
-7.38083184e-01 -1.01923406e+00 -7.69568205e-01 -8.31554651e-01
7.67223716e-01 4.90924060e-01 7.83468068e-01 6.15746737e-01
-3.15159857e-01 -1.20919466e-01 -2.92240441e-01 -4.87884171e-02
3.21873754e-01 1.99670538e-01 -7.19152018e-02 1.40630141e-01
2.18985945e-01 -1.05303121e+00 -8.85405898e-01 2.72779286e-01
-1.29947162e+00 -1.02160305e-01 7.29207635e-01 8.29750359e-01
5.44057131e-01 5.07245362e-01 6.11625671e-01 -8.76849964e-02
6.79683864e-01 -8.38936567e-01 -1.86851963e-01 4.99192066e-03
-1.18640937e-01 -2.76051491e-01 4.59819347e-01 -6.38975948e-02
-8.99333477e-01 1.91540167e-01 -5.21032929e-01 -1.05423622e-01
-4.52716887e-01 1.20013380e+00 -2.50829548e-01 4.32716727e-01
5.52747726e-01 1.96104154e-01 -6.01615310e-01 -7.34657943e-01
-6.48417473e-02 7.82005012e-01 7.51518130e-01 -2.48908713e-01
1.06802928e+00 3.66369218e-01 -2.05797017e-01 -8.70501816e-01
-1.09543681e+00 -6.92815185e-01 -9.37938631e-01 -1.03066593e-01
8.40850830e-01 -1.06537032e+00 1.91448524e-03 1.11661983e+00
-1.08392227e+00 -4.79697466e-01 -1.78514048e-01 3.77663970e-01
2.03381609e-02 1.20758682e-01 -4.17987317e-01 -7.02461541e-01
-5.90636909e-01 -8.17573547e-01 9.24423754e-01 3.74723881e-01
3.37252796e-01 -6.82367325e-01 2.36739576e-01 3.03831518e-01
4.71875250e-01 6.52590632e-01 6.18509233e-01 -4.37700450e-01
-2.43496522e-01 -2.89533764e-01 -5.37850440e-01 3.46042871e-01
3.97541493e-01 -2.34291553e-01 -7.61118650e-01 -3.43340904e-01
2.24683844e-02 -4.57916617e-01 1.20303309e+00 4.56815392e-01
1.17107844e+00 -5.65235198e-01 -3.21656582e-03 5.80752075e-01
1.51468241e+00 -5.38473241e-02 8.38968992e-01 6.36587560e-01
7.87247896e-01 6.16382420e-01 6.08300090e-01 7.32384503e-01
6.39668107e-01 4.99396503e-01 8.10524762e-01 -4.95454252e-01
9.04483572e-02 3.84540879e-03 1.26470029e-01 4.98745739e-01
-1.27346247e-01 -1.69134974e-01 -1.57185674e+00 1.03571665e+00
-1.70678508e+00 -1.08419549e+00 -2.77258366e-01 1.99223793e+00
6.51015401e-01 -6.18245453e-02 1.35364020e-02 4.72107530e-01
7.68072844e-01 4.57856387e-01 -5.82558870e-01 1.27399057e-01
-2.52781063e-01 4.50051501e-02 2.03273654e-01 2.79010683e-01
-1.56581914e+00 6.63323879e-01 4.30105448e+00 7.26921201e-01
-1.24775040e+00 2.49971822e-01 9.38623846e-01 -4.82137576e-02
-3.33917290e-02 -2.97040820e-01 -4.33271378e-01 3.51098120e-01
4.71204519e-01 2.33486563e-01 1.31918609e-01 5.56687117e-01
4.74257857e-01 -2.43251801e-01 -3.21100861e-01 5.87570727e-01
-1.04734145e-01 -9.61686790e-01 -2.55683243e-01 -6.28808588e-02
7.22426236e-01 4.51328218e-01 -9.25665274e-02 1.84599832e-01
1.76360071e-01 -8.09722066e-01 4.64126378e-01 4.53509778e-01
6.36316538e-01 -1.07587135e+00 1.06320155e+00 5.55845618e-01
-1.56726992e+00 -3.83999437e-01 -2.76917070e-01 -2.50821382e-01
3.45573932e-01 8.71142805e-01 -5.27901292e-01 7.11587012e-01
9.78235662e-01 6.95051730e-01 -5.88562906e-01 1.16262019e+00
-6.74341500e-01 6.45110786e-01 -5.64971209e-01 6.50766671e-01
6.46658301e-01 -2.18797758e-01 4.01882261e-01 1.10974598e+00
6.15985692e-01 4.85065043e-01 5.68455935e-01 4.17598277e-01
-1.05012327e-01 -1.06257923e-01 -6.18903995e-01 2.38694176e-01
2.41154909e-01 1.34091222e+00 -5.81809938e-01 -2.42528379e-01
-5.56429178e-02 8.78549814e-01 4.06009823e-01 1.45925507e-01
-6.32535577e-01 -4.15292263e-01 5.61711848e-01 1.24581046e-01
3.44721228e-01 -3.22139919e-01 -3.45767379e-01 -9.55688655e-01
-1.19984716e-01 -6.19514763e-01 5.77246904e-01 -7.82725751e-01
-1.22402263e+00 8.00657332e-01 -6.60640076e-02 -1.14208567e+00
1.74480215e-01 2.50570551e-02 -1.29538524e+00 9.46519256e-01
-1.81437922e+00 -1.43947816e+00 -7.41165280e-01 5.97081900e-01
5.29523432e-01 2.52693176e-01 5.76467693e-01 6.20951951e-01
-1.07133365e+00 1.09075151e-01 -5.80892935e-02 5.20085573e-01
3.06972295e-01 -9.87269104e-01 5.09097874e-01 1.33020365e+00
-3.84356290e-01 5.24472855e-02 3.66246849e-01 -8.64525974e-01
-6.07518435e-01 -1.59089887e+00 1.01467347e+00 3.24797064e-01
5.96071064e-01 5.71291111e-02 -1.15584981e+00 3.93638313e-01
4.82950062e-02 1.81050360e-01 4.71963972e-01 -1.47464454e-01
-1.33950189e-01 -3.04744065e-01 -1.17107916e+00 2.36804456e-01
7.99831033e-01 -4.95395452e-01 -4.58387256e-01 3.53882283e-01
5.11540532e-01 -1.12507194e-01 -7.89375603e-01 3.84308964e-01
1.95178479e-01 -7.85302818e-01 1.03962457e+00 -3.84239346e-01
8.62864494e-01 -3.96350861e-01 -3.82349342e-01 -1.49478602e+00
-6.41638279e-01 -1.07402643e-02 2.89819121e-01 1.16604948e+00
8.30728337e-02 -3.54396313e-01 2.97587901e-01 3.45518053e-01
-4.00363147e-01 -8.60694766e-01 -8.80888879e-01 -4.81876105e-01
5.43761067e-04 -3.50353390e-01 7.24307716e-01 9.69725370e-01
-4.93261009e-01 3.20316225e-01 -4.01960403e-01 6.87936187e-01
4.48120981e-01 6.96693584e-02 4.57997590e-01 -1.22403300e+00
1.94742426e-01 -4.50091600e-01 -2.29383290e-01 -6.35514736e-01
-2.96237499e-01 -7.11016834e-01 1.71681434e-01 -1.88583136e+00
1.75827503e-01 -5.16041338e-01 -5.86189985e-01 1.10899711e+00
-5.25041401e-01 5.41727781e-01 1.83904007e-01 2.18302250e-01
1.83947068e-02 9.36221719e-01 8.91702175e-01 -4.27525580e-01
-2.73805588e-01 1.12084821e-01 -4.53534991e-01 7.70876527e-01
1.29392946e+00 -6.74371541e-01 -8.01599622e-02 -8.12571406e-01
8.54766443e-02 -8.14705715e-02 7.01981723e-01 -1.36915445e+00
1.86040714e-01 -1.03420228e-01 2.53067255e-01 -1.02883196e+00
1.06283017e-02 -7.22775400e-01 1.51359200e-01 7.58316219e-01
-6.50198758e-02 2.14064002e-01 2.23662421e-01 3.26604992e-01
-4.69344944e-01 -1.73022360e-01 6.90619111e-01 -6.90530688e-02
-6.51088476e-01 7.48376250e-01 -3.06083262e-01 -2.38417760e-01
8.90460610e-01 1.70471102e-01 -2.70501226e-01 -1.83978811e-01
-6.22577846e-01 6.08790874e-01 1.45740256e-01 2.69752592e-01
8.17390859e-01 -1.53414047e+00 -1.22886658e+00 8.57406855e-02
-6.60965443e-02 5.92491806e-01 8.44795704e-01 1.04842293e+00
-5.77673614e-01 -9.14700702e-02 -4.35976267e-01 -3.00705701e-01
-1.07852781e+00 3.83379817e-01 1.93253487e-01 -7.49188662e-01
-4.95350331e-01 9.19859827e-01 1.62318796e-01 -5.05190253e-01
-2.49262769e-02 -2.30881780e-01 -4.53833073e-01 6.50348842e-01
7.33359396e-01 3.22274327e-01 1.22694679e-01 -9.50445950e-01
-3.41476172e-01 4.23590690e-01 1.63503736e-01 4.51181009e-02
2.12369394e+00 -2.05017298e-01 -2.54779220e-01 9.32991579e-02
1.10598290e+00 -6.08264148e-01 -1.37841177e+00 -4.51117039e-01
-5.17100580e-02 -3.21309894e-01 6.66811645e-01 -7.50335872e-01
-1.64844334e+00 1.26920593e+00 7.44102657e-01 1.09994814e-01
1.77302325e+00 -2.91471511e-01 1.13687575e+00 3.55250716e-01
-4.39837156e-03 -9.83784974e-01 1.67545959e-01 5.03475249e-01
1.52605033e+00 -1.32789528e+00 5.52174635e-02 2.61448860e-01
-4.96839792e-01 9.16988194e-01 5.47711372e-01 -1.76374018e-01
6.68416739e-01 1.52140558e-01 -1.44052282e-01 -2.99643219e-01
-2.23514453e-01 -5.46185315e-01 1.45452365e-01 4.69194025e-01
-2.30169520e-02 1.25748694e-01 -1.55326696e-02 5.85401535e-01
1.66085869e-01 -9.19507593e-02 2.56005943e-01 1.08814347e+00
-7.14403272e-01 -6.88301265e-01 -6.96433485e-01 3.59643549e-01
-2.05220968e-01 -2.88616240e-01 -2.24099219e-01 5.60481906e-01
3.88233840e-01 9.42843616e-01 8.75749514e-02 -6.37904704e-01
4.30811971e-01 -2.47831315e-01 -2.43821889e-01 -5.06125987e-01
-8.86341333e-01 4.59886938e-02 -1.49504542e-01 -4.36967373e-01
-5.31484604e-01 -6.33527279e-01 -1.21018779e+00 -2.99436569e-01
-1.32215127e-01 -4.36470881e-02 4.52562571e-01 8.12872410e-01
2.60788232e-01 5.32141387e-01 1.00004447e+00 -1.36534202e+00
-2.70382494e-01 -1.29956925e+00 -5.51837742e-01 3.50602180e-01
5.19395649e-01 -5.46961188e-01 -1.29394576e-01 -1.95538640e-01] | [9.649314880371094, -1.3308225870132446] |
74d051e5-95bc-4d2b-9596-a26326d5b85f | single-shot-multi-person-3d-pose-estimation | 1712.03453 | null | http://arxiv.org/abs/1712.03453v3 | http://arxiv.org/pdf/1712.03453v3.pdf | Single-Shot Multi-Person 3D Pose Estimation From Monocular RGB | We propose a new single-shot method for multi-person 3D pose estimation in
general scenes from a monocular RGB camera. Our approach uses novel
occlusion-robust pose-maps (ORPM) which enable full body pose inference even
under strong partial occlusions by other people and objects in the scene. ORPM
outputs a fixed number of maps which encode the 3D joint locations of all
people in the scene. Body part associations allow us to infer 3D pose for an
arbitrary number of people without explicit bounding box prediction. To train
our approach we introduce MuCo-3DHP, the first large scale training data set
showing real images of sophisticated multi-person interactions and occlusions.
We synthesize a large corpus of multi-person images by compositing images of
individual people (with ground truth from mutli-view performance capture). We
evaluate our method on our new challenging 3D annotated multi-person test set
MuPoTs-3D where we achieve state-of-the-art performance. To further stimulate
research in multi-person 3D pose estimation, we will make our new datasets, and
associated code publicly available for research purposes. | ['Christian Theobalt', 'Gerard Pons-Moll', 'Weipeng Xu', 'Oleksandr Sotnychenko', 'Franziska Mueller', 'Dushyant Mehta', 'Srinath Sridhar'] | 2017-12-09 | null | null | null | null | ['3d-multi-person-pose-estimation-absolute', '3d-multi-person-pose-estimation-root-relative'] | ['computer-vision', 'computer-vision'] | [-2.30776116e-01 2.52312403e-02 3.21043342e-01 -3.92641246e-01
-8.34682107e-01 -5.68031549e-01 5.12665749e-01 -3.68285328e-01
-4.45851415e-01 6.05772138e-01 3.86692882e-01 6.31638169e-01
3.46790463e-01 -3.62964541e-01 -7.84053266e-01 -2.77451694e-01
-2.54295822e-02 1.33036697e+00 3.78579646e-01 -5.93991838e-02
-3.45026016e-01 4.04980123e-01 -1.59635568e+00 1.58325225e-01
1.86837718e-01 6.55551910e-01 -2.40213379e-01 9.99192357e-01
6.92726910e-01 2.17007220e-01 -6.64480329e-01 -6.43411815e-01
6.49250209e-01 -1.95322663e-01 -7.12282240e-01 6.71156704e-01
1.20260322e+00 -6.10460937e-01 -4.27571565e-01 4.68547344e-01
9.51686978e-01 1.98299333e-01 5.01060605e-01 -1.39522862e+00
-3.52197886e-02 -1.16384208e-01 -7.03158915e-01 -7.05570355e-02
1.36472070e+00 4.64093447e-01 6.85389280e-01 -8.55344594e-01
9.41780210e-01 1.74091899e+00 9.20307338e-01 6.69627726e-01
-1.10936081e+00 -2.28291824e-01 2.11259827e-01 -4.59056534e-02
-1.55996144e+00 -4.34117794e-01 5.08614659e-01 -5.07862747e-01
8.90785575e-01 4.49285179e-01 1.28648591e+00 1.51649058e+00
-5.90650588e-02 1.02144635e+00 1.19690454e+00 -3.10763121e-01
-8.11339319e-02 -3.36308554e-02 4.09241207e-02 9.21846569e-01
3.83630961e-01 -5.50285401e-03 -7.68472552e-01 -2.78007418e-01
9.16877389e-01 5.47333024e-02 -1.29522949e-01 -8.52623165e-01
-1.24391150e+00 4.05686766e-01 3.90971214e-01 -4.01891679e-01
-1.52214661e-01 2.26215392e-01 1.63038801e-02 -1.62652627e-01
3.01343143e-01 -1.38304487e-01 -5.35015643e-01 -8.98077488e-02
-6.07345700e-01 1.11282182e+00 7.38722801e-01 1.13933122e+00
3.83174449e-01 -6.26421511e-01 -3.40330511e-01 5.95004618e-01
4.17087913e-01 7.52308011e-01 -1.30094543e-01 -1.08728051e+00
6.52166843e-01 6.96688354e-01 5.65786779e-01 -6.64784193e-01
-7.45037675e-01 -1.07338376e-01 -4.01698560e-01 6.59658387e-02
8.65073800e-01 -9.98003781e-02 -9.23751175e-01 1.49882138e+00
9.99138594e-01 -2.61690557e-01 -4.49278027e-01 1.37983024e+00
1.02621949e+00 2.00007677e-01 -1.48764953e-01 2.62544781e-01
1.84002292e+00 -1.05048919e+00 -3.44832361e-01 -5.64441085e-01
2.19772667e-01 -4.65827942e-01 8.74595165e-01 3.87281239e-01
-1.19445384e+00 -6.65328503e-01 -6.04175031e-01 -2.91633219e-01
-3.15796703e-01 7.26161152e-02 6.85716033e-01 9.10397053e-01
-6.78922594e-01 3.75357181e-01 -8.54412019e-01 -7.34695375e-01
5.04968584e-01 4.34365422e-01 -7.61395693e-01 -2.48867497e-01
-7.98096359e-01 1.01868200e+00 7.08559081e-02 1.79874659e-01
-9.43050802e-01 -4.11783338e-01 -1.13881624e+00 -5.70006013e-01
6.19177401e-01 -1.28200102e+00 1.10361862e+00 -1.48452297e-01
-1.12596321e+00 1.54207718e+00 -8.46864805e-02 -8.77077952e-02
1.18384087e+00 -9.28876638e-01 -8.32727086e-03 2.28770763e-01
1.37276217e-01 9.18152094e-01 6.29805624e-01 -1.29459131e+00
-3.55634034e-01 -9.35744405e-01 9.89817381e-02 5.34537315e-01
2.17510015e-01 2.14555293e-01 -1.06524491e+00 -3.41824591e-01
7.04432651e-02 -1.39902806e+00 -3.71436328e-01 3.90291184e-01
-8.85497928e-01 -2.34121740e-01 4.60870117e-01 -7.44414866e-01
4.98554528e-01 -1.58543646e+00 6.67321324e-01 -3.37582529e-02
2.33783007e-01 -8.05747211e-02 5.69230393e-02 1.78416252e-01
1.73003435e-01 -3.51728469e-01 2.03489959e-02 -1.06184602e+00
3.22104216e-01 2.17222318e-01 3.97832453e-01 9.06857967e-01
-1.04921415e-01 1.09233606e+00 -6.42962217e-01 -8.64317000e-01
4.92257118e-01 6.94155216e-01 -6.54222429e-01 4.43590552e-01
-9.79364067e-02 8.20667982e-01 -2.29922533e-01 1.11726952e+00
5.00780404e-01 -2.87312001e-01 -1.17889509e-01 -1.35003448e-01
3.12787294e-01 -2.22089246e-01 -1.59966671e+00 2.18887019e+00
1.78946391e-01 7.26521835e-02 -1.22089386e-01 -6.75649270e-02
3.84510875e-01 3.22993517e-01 5.66586494e-01 -1.17363641e-02
2.47227058e-01 -3.08331460e-01 -4.51826692e-01 -4.24609721e-01
3.36117089e-01 1.07462026e-01 -5.71882904e-01 2.59735405e-01
1.53559923e-01 -2.03527987e-01 2.09460378e-01 1.49599761e-01
8.49079728e-01 8.09712946e-01 3.23224366e-01 3.65189761e-02
3.10456097e-01 -1.39052257e-01 3.51213127e-01 8.16485584e-01
-5.43337584e-01 9.76039529e-01 1.61181182e-01 -9.59527194e-01
-1.09581101e+00 -1.40431762e+00 -4.95971665e-02 1.01596510e+00
1.23373799e-01 -6.58610046e-01 -7.76335776e-01 -6.93670928e-01
1.48393527e-01 3.19385715e-02 -7.66772687e-01 5.03043413e-01
-6.53523207e-01 -8.62109601e-01 4.13150311e-01 5.15695691e-01
5.10576665e-01 -6.69336140e-01 -9.33107018e-01 -1.61138743e-01
-5.84780335e-01 -1.61851084e+00 -5.25123358e-01 -3.63598019e-01
-4.35350865e-01 -1.38178408e+00 -1.13557780e+00 -3.58576030e-01
6.96921468e-01 -6.03367463e-02 1.33074212e+00 -1.90230668e-01
-7.83915520e-01 7.60213971e-01 -9.86939371e-02 -2.45359525e-01
2.42364481e-01 -1.70593157e-01 5.15508831e-01 -1.87058315e-01
2.99358130e-01 -3.88159573e-01 -7.55212486e-01 5.84947288e-01
1.02678716e-01 2.54675746e-01 1.31826624e-01 2.11783916e-01
7.04050481e-01 -2.87825882e-01 -3.74016553e-01 -5.01675963e-01
-1.72038879e-02 1.40950546e-01 -4.15001273e-01 9.73536149e-02
4.54858661e-01 -4.05710697e-01 -1.37506351e-01 -3.30784351e-01
-7.86134422e-01 5.80576301e-01 -9.10265446e-02 -3.31739277e-01
-5.45918167e-01 -4.78067219e-01 -4.86151844e-01 -3.98112051e-02
7.15833545e-01 -1.44917071e-01 -3.35566431e-01 -6.16455972e-01
4.76479024e-01 3.82181644e-01 9.54725444e-01 -7.96075702e-01
7.65261173e-01 7.88707316e-01 1.54009491e-01 -7.12727785e-01
-1.12203598e+00 -5.74845731e-01 -1.39425743e+00 -6.70741975e-01
1.18746495e+00 -1.41571474e+00 -1.17471886e+00 6.70314610e-01
-1.15820682e+00 -3.38775098e-01 -2.77491678e-02 3.57150435e-01
-7.84395516e-01 2.74714261e-01 -5.95288575e-01 -9.95213747e-01
-1.07111335e-01 -9.69439507e-01 1.87316918e+00 5.00757359e-02
-6.48263812e-01 -5.85114181e-01 1.57678097e-01 8.84903610e-01
-4.92399782e-01 9.27965164e-01 -1.30114689e-01 -1.59324482e-01
-6.78019881e-01 -5.72152376e-01 1.95016161e-01 -2.72763759e-01
-2.60192096e-01 -3.80606532e-01 -1.01187098e+00 -3.67372841e-01
-5.50344050e-01 -5.84005475e-01 5.21894038e-01 4.17935640e-01
8.51891398e-01 -2.58424971e-02 -5.25490105e-01 6.34074450e-01
1.00369954e+00 -7.63774574e-01 5.09469688e-01 1.88226640e-01
1.14726007e+00 8.63279819e-01 5.61716318e-01 7.02627659e-01
8.72943521e-01 1.27308691e+00 2.65317559e-01 -6.62542740e-03
-2.76317716e-01 -4.04413670e-01 2.13538110e-01 1.49412453e-01
-7.87933588e-01 -1.68580636e-01 -8.96676004e-01 2.51924634e-01
-1.75676847e+00 -7.85965681e-01 -3.47270995e-01 2.26925039e+00
5.26441455e-01 1.91233337e-01 9.04960454e-01 8.07055924e-03
6.24712825e-01 -5.37094399e-02 -3.40334058e-01 5.47182858e-01
-6.13903366e-02 -1.56051174e-01 3.65153700e-01 5.34958839e-01
-1.43707824e+00 8.25823486e-01 6.42692137e+00 1.79254070e-01
-1.00441925e-01 9.71142352e-02 2.90571600e-01 -7.88775027e-01
3.35242152e-01 -3.41601789e-01 -1.27067423e+00 1.90533310e-01
4.96790379e-01 5.69548309e-01 1.37715325e-01 7.78693318e-01
3.65214907e-02 -4.53634799e-01 -1.41646063e+00 1.55310583e+00
3.84716839e-01 -6.78435445e-01 -3.38536531e-01 2.97765315e-01
7.07416952e-01 -2.25694582e-01 -2.70221531e-01 1.75386518e-01
3.36835921e-01 -8.84999275e-01 9.43622947e-01 7.85187602e-01
5.89240074e-01 -5.88137627e-01 4.89528269e-01 5.01979589e-01
-1.24839175e+00 1.34155542e-01 -2.87952393e-01 -2.84359545e-01
6.25086665e-01 3.09189111e-01 -5.24057925e-01 4.99583304e-01
1.05035555e+00 5.76106489e-01 -9.59284842e-01 9.28003967e-01
-2.62112260e-01 -2.66201824e-01 -7.03281760e-01 1.50087446e-01
-3.84941787e-01 1.49636760e-01 6.17358446e-01 1.03705645e+00
1.45900995e-02 2.31121853e-01 6.30644977e-01 5.26594341e-01
1.72161534e-01 -2.05163509e-01 -3.42864275e-01 5.72152853e-01
1.14495885e-02 1.24026668e+00 -8.13490391e-01 -4.05590594e-01
-1.60239324e-01 1.47525716e+00 3.24976742e-01 4.92971838e-02
-9.08516288e-01 2.87895620e-01 7.70060241e-01 4.67321426e-01
2.00824976e-01 -4.41794634e-01 1.20888427e-01 -1.44642186e+00
3.16812992e-01 -7.25626945e-01 5.08668721e-01 -1.06512308e+00
-1.19604707e+00 4.55529213e-01 5.90300441e-01 -1.14808381e+00
-3.45756918e-01 -7.35735834e-01 6.53391182e-02 7.06897676e-01
-7.84812629e-01 -1.69822085e+00 -6.87921643e-01 7.65740514e-01
4.72003549e-01 4.11273800e-02 8.58001292e-01 1.87318891e-01
-4.69399959e-01 4.02179778e-01 -8.70247364e-01 3.03473532e-01
8.38413000e-01 -1.35177386e+00 7.30382442e-01 6.63130283e-01
2.49590456e-01 5.10979891e-01 9.93892610e-01 -8.78529310e-01
-1.55039990e+00 -7.76946843e-01 8.40898514e-01 -1.70959711e+00
-4.29621302e-02 -1.06968915e+00 -1.85241893e-01 1.08075094e+00
-2.58737266e-01 3.14031392e-01 7.36126423e-01 2.70264387e-01
-2.69100845e-01 2.09516749e-01 -1.30141306e+00 6.44351542e-01
1.78534281e+00 -1.10335000e-01 -7.88436830e-01 6.84780657e-01
4.20838386e-01 -1.06540453e+00 -8.45778942e-01 2.81737745e-01
9.18439329e-01 -1.12185991e+00 1.49884820e+00 -4.73861158e-01
1.58204317e-01 -4.40431535e-01 -3.32074523e-01 -8.93611372e-01
-4.36158553e-02 -5.00658274e-01 -5.13805211e-01 7.64904916e-01
-2.84281313e-01 -1.12278245e-01 1.11491346e+00 9.68921781e-01
3.07401866e-01 -5.11909664e-01 -9.82377708e-01 -6.35320723e-01
-4.40561026e-01 -5.83357930e-01 2.96720505e-01 2.45531827e-01
-2.18500599e-01 1.70350000e-01 -8.39146435e-01 3.70131165e-01
1.23676145e+00 3.77217829e-02 1.62614667e+00 -1.27012396e+00
-5.90573132e-01 2.00748757e-01 -7.18975186e-01 -1.20975697e+00
-9.25870240e-02 -3.51689816e-01 -1.28461957e-01 -1.49876487e+00
5.96364021e-01 -3.70499119e-02 5.82883656e-01 5.90744853e-01
-3.15507621e-01 8.58095706e-01 4.51744318e-01 1.32153139e-01
-1.14600158e+00 4.26372886e-01 1.48344886e+00 1.29078910e-01
-3.50419208e-02 1.29583001e-01 -2.37069353e-01 1.05581939e+00
8.67312253e-02 -3.62336189e-01 -3.62846218e-02 -2.48392493e-01
3.48618180e-02 9.45671499e-02 1.20896840e+00 -1.39991176e+00
-3.24185775e-03 7.52381906e-02 1.27109122e+00 -1.19107068e+00
1.17159545e+00 -6.83196127e-01 5.63341796e-01 4.96541291e-01
1.00269236e-01 -8.45477581e-02 -3.70245986e-02 4.82693136e-01
5.35235405e-01 3.31303149e-01 5.47411501e-01 -8.89038563e-01
-6.20494008e-01 6.21832728e-01 2.70674944e-01 1.61757141e-01
1.03631938e+00 -3.11105579e-01 2.76344009e-02 -3.35232705e-01
-1.27954233e+00 4.27012563e-01 6.27050877e-01 6.27797067e-01
5.47402382e-01 -1.46904397e+00 -9.50331509e-01 2.19068825e-01
2.74387926e-01 2.43720382e-01 4.07183170e-01 5.35759687e-01
-5.07343411e-01 3.01472992e-01 -3.35578173e-01 -9.10529792e-01
-1.66146410e+00 2.87547141e-01 4.18199599e-01 -1.36672422e-01
-9.37671661e-01 1.01074433e+00 1.26441747e-01 -8.20173800e-01
2.22835392e-01 4.78706229e-03 2.25870639e-01 -1.77399173e-01
8.14049900e-01 4.72914994e-01 -3.21992040e-01 -1.14594889e+00
-7.10291088e-01 9.33286071e-01 2.51896948e-01 -3.49718690e-01
1.22255564e+00 -3.94043446e-01 1.82156444e-01 4.83459204e-01
9.11423385e-01 3.74006182e-02 -1.65816939e+00 -1.35137230e-01
-5.98837554e-01 -9.08937275e-01 -6.72725797e-01 -8.14496994e-01
-6.15107954e-01 6.12508833e-01 6.79534435e-01 -4.51336831e-01
5.41595519e-01 5.60823739e-01 5.90653837e-01 2.54530966e-01
9.70535457e-01 -1.07824445e+00 3.64074260e-01 1.94001362e-01
1.04922664e+00 -1.35825503e+00 5.06596506e-01 -6.74975574e-01
-5.61953902e-01 7.33208239e-01 7.87940443e-01 -1.71576530e-01
2.27801144e-01 9.19808000e-02 -2.71781497e-02 -3.83592784e-01
-2.41009593e-01 -4.88018990e-01 5.34140289e-01 9.36752439e-01
1.62705675e-01 2.45389581e-01 2.54493743e-01 5.95923960e-01
-6.58467889e-01 -1.69698969e-01 1.13894254e-01 8.68068516e-01
-9.64993834e-02 -1.03360736e+00 -1.03721726e+00 5.71053959e-02
-2.71908849e-01 5.40955663e-01 -5.37799299e-01 9.31476355e-01
4.59624171e-01 8.03389251e-01 -2.36472301e-02 -1.73845440e-01
7.37347126e-01 1.90280572e-01 1.30308652e+00 -6.69634521e-01
-3.04783881e-01 2.74038285e-01 2.81509966e-01 -8.72526884e-01
-6.71712399e-01 -1.20272708e+00 -8.10612082e-01 -4.01432484e-01
8.06988329e-02 -5.41254580e-01 3.96445602e-01 1.04549956e+00
-1.32066295e-01 2.57097065e-01 -4.20130715e-02 -1.58152008e+00
-7.40359649e-02 -9.95421171e-01 -5.25001585e-01 7.52595603e-01
1.66801974e-01 -9.83025730e-01 1.09007165e-01 2.53155768e-01] | [7.0341362953186035, -0.9356746077537537] |
074b6642-e373-45fe-8f50-43dae1ac21f2 | arraybot-reinforcement-learning-for | 2306.16857 | null | https://arxiv.org/abs/2306.16857v1 | https://arxiv.org/pdf/2306.16857v1.pdf | ArrayBot: Reinforcement Learning for Generalizable Distributed Manipulation through Touch | We present ArrayBot, a distributed manipulation system consisting of a $16 \times 16$ array of vertically sliding pillars integrated with tactile sensors, which can simultaneously support, perceive, and manipulate the tabletop objects. Towards generalizable distributed manipulation, we leverage reinforcement learning (RL) algorithms for the automatic discovery of control policies. In the face of the massively redundant actions, we propose to reshape the action space by considering the spatially local action patch and the low-frequency actions in the frequency domain. With this reshaped action space, we train RL agents that can relocate diverse objects through tactile observations only. Surprisingly, we find that the discovered policy can not only generalize to unseen object shapes in the simulator but also transfer to the physical robot without any domain randomization. Leveraging the deployed policy, we present abundant real-world manipulation tasks, illustrating the vast potential of RL on ArrayBot for distributed manipulation. | ['Huazhe Xu', 'Gu Zhang', 'Changyi Lin', 'Yuanchen Ju', 'Zhengmao He', 'Jingwen Cheng', 'Han Zhang', 'Zhengrong Xue'] | 2023-06-29 | null | null | null | null | ['reinforcement-learning-1'] | ['methodology'] | [ 2.64474005e-01 2.11505204e-01 2.95448187e-03 3.39753628e-01
-5.11662900e-01 -1.05352414e+00 1.36657104e-01 -3.30733322e-02
-3.02736044e-01 9.57223773e-01 -2.10931510e-01 1.18248269e-01
-7.31738150e-01 -7.31346428e-01 -9.92746890e-01 -8.81331980e-01
-5.38883388e-01 6.03746831e-01 4.50648844e-01 -2.78016150e-01
2.59939432e-01 6.61407709e-01 -1.91946733e+00 9.54197496e-02
5.56634665e-01 1.20178545e+00 7.22883999e-01 6.53358221e-01
3.78200769e-01 5.76437891e-01 -6.93298042e-01 4.94480789e-01
5.64523220e-01 7.51142995e-03 -1.63751617e-01 1.66559499e-02
7.24886954e-02 -5.97000301e-01 -3.07126403e-01 5.45468509e-01
5.38865209e-01 4.16493207e-01 4.35219467e-01 -1.11977351e+00
-3.74691039e-01 7.53153563e-01 -4.93627310e-01 -3.62007290e-01
4.93601263e-01 6.14937782e-01 6.00792766e-01 -3.05129588e-01
8.79259944e-01 1.28889012e+00 1.03502251e-01 4.92545754e-01
-9.12796974e-01 -6.17299199e-01 2.99406439e-01 -3.01216900e-01
-9.73563194e-01 -4.68959697e-02 5.90503275e-01 -1.55843839e-01
8.61344516e-01 1.46906704e-01 6.51372850e-01 1.38233817e+00
5.65670609e-01 5.14030457e-01 1.06013119e+00 -2.14739129e-01
7.26908267e-01 -6.36905313e-01 -7.39556968e-01 7.36533821e-01
4.18500453e-01 1.60183832e-01 -7.43755341e-01 -2.89849520e-01
1.31303072e+00 1.64137289e-01 1.08219825e-01 -9.37627435e-01
-1.32090032e+00 2.44168505e-01 5.44539809e-01 1.62105069e-01
-6.55142069e-01 7.15415120e-01 2.60759771e-01 3.49366546e-01
-2.70050496e-01 1.10058248e+00 -6.90576613e-01 -4.85545516e-01
1.23307452e-01 5.83362758e-01 7.79727995e-01 1.23897564e+00
7.43943810e-01 -1.06629953e-01 -3.99211973e-01 5.31390965e-01
-1.01557542e-02 9.87881601e-01 1.23721160e-01 -1.39950037e+00
5.53645611e-01 5.79359770e-01 7.66257107e-01 -7.61995852e-01
-3.67601305e-01 1.44079477e-01 -4.41950768e-01 5.20645559e-01
1.44735649e-01 -5.04760802e-01 -7.44218051e-01 1.57133222e+00
6.20721221e-01 -1.98989078e-01 -3.14717382e-01 8.11250091e-01
-1.19620092e-01 1.36614054e-01 -3.40536307e-03 -3.41916233e-02
7.95885801e-01 -6.95222735e-01 -7.23696947e-01 -1.23829283e-01
2.10209012e-01 -1.87387466e-01 1.01454222e+00 5.11901021e-01
-1.03788340e+00 -2.95686424e-01 -1.11137152e+00 3.32859784e-01
-2.44774610e-01 1.78746432e-01 8.79235446e-01 1.49791420e-01
-8.76835823e-01 8.24287653e-01 -1.18553376e+00 -9.88358259e-02
3.81684095e-01 7.35505104e-01 -2.21985012e-01 -1.20000251e-01
-5.11642873e-01 7.24434793e-01 2.50626981e-01 7.92214945e-02
-1.16968501e+00 -3.03462178e-01 -4.72228020e-01 -9.73507911e-02
1.03581822e+00 -5.25339365e-01 1.17732430e+00 -1.17757723e-01
-2.50060344e+00 8.48157778e-02 4.34082776e-01 -1.23501472e-01
4.49751765e-01 -4.75910842e-01 3.13371509e-01 3.65792215e-01
1.34958727e-02 6.16905510e-01 1.24386966e+00 -1.49854290e+00
-5.64749599e-01 -2.77319282e-01 7.15158358e-02 2.15005159e-01
-2.96977133e-01 -7.85399079e-01 1.00416034e-01 -3.84619117e-01
-2.01608916e-03 -1.13485014e+00 -2.51759827e-01 1.59495145e-01
-2.30741099e-01 -3.51929307e-01 9.37871158e-01 6.60992041e-02
7.69740880e-01 -2.11941028e+00 6.91202223e-01 4.13548410e-01
1.60570204e-01 -1.19921215e-01 -4.11658376e-01 9.43442881e-01
7.06370711e-01 -2.29689434e-01 8.33776444e-02 -1.14896066e-01
3.85173321e-01 4.61196542e-01 -3.61318946e-01 3.51410121e-01
-7.39079863e-02 1.11177564e+00 -1.13297033e+00 -4.97289151e-02
2.32154220e-01 -1.51708588e-01 -6.27178252e-01 3.42657328e-01
-8.13176811e-01 7.61420906e-01 -1.05182874e+00 8.31364989e-01
4.47565615e-01 -6.02646023e-02 5.83597481e-01 2.83665061e-01
-2.70530343e-01 -4.77484539e-02 -1.15994346e+00 1.94931066e+00
-4.54736918e-01 -1.49367958e-01 6.51348650e-01 -6.94953084e-01
1.07563519e+00 -5.27171381e-02 9.41719234e-01 -6.82296336e-01
1.66190609e-01 5.03819227e-01 5.56320883e-02 -7.87460566e-01
3.09886724e-01 2.92401403e-01 -4.61323559e-01 5.60415626e-01
-4.54801694e-02 -6.76377654e-01 -1.38729870e-01 -4.46698278e-01
1.76541436e+00 4.94386196e-01 -5.69318607e-02 -4.01049964e-02
-2.85247713e-01 -6.56550303e-02 8.83491188e-02 1.41849220e+00
5.26977107e-02 -5.41715557e-03 2.50994921e-01 -1.48650959e-01
-6.58980191e-01 -1.48144805e+00 3.26989919e-01 1.47345424e+00
7.34617352e-01 3.17124603e-03 -1.08598851e-01 -4.48316693e-01
7.96968818e-01 1.67856932e-01 -6.35279119e-01 -1.15062207e-01
-8.46868813e-01 3.25067431e-01 2.49560818e-01 4.30463761e-01
3.68578166e-01 -1.47075152e+00 -1.54277861e+00 3.88047308e-01
3.49098951e-01 -9.03610051e-01 -1.52392149e-01 4.25446719e-01
-5.20772219e-01 -9.39978063e-01 -2.33333662e-01 -6.47033751e-01
4.92972553e-01 3.50806296e-01 4.56125379e-01 -6.57116175e-02
-5.52171886e-01 9.69370842e-01 -5.12163877e-01 -3.54108602e-01
-3.71339247e-02 7.28859147e-03 3.11704069e-01 -3.56931150e-01
-5.22902846e-01 -8.59204650e-01 -7.22925901e-01 3.48341286e-01
-7.50280976e-01 -2.84316003e-01 9.49617743e-01 6.80689931e-01
3.03287476e-01 -1.58925783e-02 7.34608114e-01 -2.26431247e-02
8.13367903e-01 -4.39211488e-01 -6.53711915e-01 2.17560530e-01
1.98840588e-01 1.26377553e-01 4.57372069e-01 -9.89074409e-01
-7.53535628e-01 9.52700377e-02 6.65054262e-01 -5.10149956e-01
-1.21293902e-01 3.19373608e-01 8.81620347e-02 -3.07121694e-01
6.54510140e-01 -1.67782038e-01 3.03901374e-01 -3.41963351e-01
5.54557145e-01 4.52887684e-01 1.50304168e-01 -1.26267946e+00
6.15016282e-01 5.53813756e-01 2.99044400e-01 -5.77696264e-01
2.13748845e-03 -5.38695529e-02 -3.38112205e-01 -3.52212369e-01
7.13270009e-01 -4.38928723e-01 -1.65724015e+00 3.40361565e-01
-9.56969559e-01 -1.07587171e+00 -4.99266773e-01 3.08036387e-01
-1.11185086e+00 -1.60344973e-01 -5.00119984e-01 -1.00416148e+00
1.73100576e-01 -1.05209208e+00 1.67987072e+00 2.27823585e-01
-2.37864509e-01 -2.07347646e-01 7.81596452e-02 -1.72114104e-01
8.10140491e-01 4.86948699e-01 9.29659247e-01 -2.27531880e-01
-1.09719539e+00 -2.52042562e-01 1.86341465e-01 -3.20565641e-01
4.96157557e-01 -4.19316381e-01 -4.89579380e-01 -6.54672205e-01
-2.32893094e-01 -9.04894054e-01 3.15308541e-01 1.83902472e-01
1.35079873e+00 -2.46326938e-01 -7.22395718e-01 -1.06932130e-02
1.03301585e+00 4.61036563e-01 3.66291046e-01 9.97990742e-02
4.36851740e-01 4.11531210e-01 8.86906326e-01 8.67477596e-01
-7.62313139e-03 7.67955124e-01 7.86816061e-01 4.77953821e-01
2.18145087e-01 -3.95942092e-01 3.16303849e-01 3.49658877e-01
-1.73083037e-01 -3.44641834e-01 -5.00316024e-01 1.81564003e-01
-2.05157852e+00 -5.81232607e-01 9.20482457e-01 2.17371249e+00
7.80096352e-01 -1.88279077e-02 2.11975761e-02 -4.95844007e-01
5.27159333e-01 -1.19425006e-01 -1.14573503e+00 -9.01170000e-02
3.37363064e-01 5.87038636e-01 6.23027027e-01 2.57404923e-01
-7.85436928e-01 8.34755898e-01 6.23279190e+00 7.69398153e-01
-1.22671163e+00 -2.82507479e-01 -2.78932780e-01 -4.73708779e-01
-6.50234055e-03 -2.92897433e-01 -4.01965678e-01 3.63765568e-01
2.70793349e-01 1.34728923e-01 1.05178082e+00 8.00382018e-01
-7.53397942e-02 -3.98126066e-01 -1.21884501e+00 6.49491429e-01
-4.62835610e-01 -1.09176135e+00 -2.35665709e-01 1.56021103e-01
7.20346093e-01 1.04328012e-02 1.82196558e-01 5.43832839e-01
7.09502339e-01 -9.99990046e-01 6.83099627e-01 4.91661280e-01
7.12581575e-01 -2.51175702e-01 1.43552229e-01 6.77955031e-01
-1.08090329e+00 -7.58764863e-01 -2.23053545e-01 -3.28227788e-01
-6.02833293e-02 -5.92525937e-02 -8.85194898e-01 4.18402582e-01
6.68300152e-01 3.91869277e-01 9.28324163e-02 8.54719579e-01
-5.80139756e-02 -9.23914015e-02 -7.73293495e-01 -7.49456346e-01
-6.64489940e-02 -1.89405203e-01 5.24926543e-01 1.69813693e-01
5.32527089e-01 8.69070739e-02 6.49192393e-01 7.47764170e-01
6.86351731e-02 -5.13003945e-01 -1.03997052e+00 -8.04997534e-02
8.85045588e-01 8.73281896e-01 -6.56143486e-01 4.89069568e-03
2.08708823e-01 7.82776833e-01 4.91319805e-01 5.25724769e-01
-6.93155587e-01 -3.89675289e-01 7.10782468e-01 -1.89681619e-01
7.71614254e-01 -7.88190126e-01 -5.11875898e-02 -6.99646831e-01
4.05316710e-01 -9.22513604e-01 -3.98814291e-01 -5.91793478e-01
-1.36183786e+00 -1.63633019e-01 2.59892136e-01 -1.20662856e+00
-1.99210793e-01 -6.11983359e-01 -6.43358305e-02 3.27761382e-01
-8.99809420e-01 -1.01740265e+00 -2.92441308e-01 2.70745188e-01
4.26181316e-01 -5.88581599e-02 8.50962222e-01 -3.59683901e-01
4.91960049e-02 2.24956676e-01 1.46224350e-01 -4.68288839e-01
5.43100655e-01 -9.49582279e-01 -1.43988982e-01 -2.38100663e-01
-4.28202659e-01 6.77144229e-01 5.08810759e-01 -1.01606345e+00
-2.25894189e+00 -6.71237767e-01 -4.41750348e-01 -4.52695072e-01
9.17257249e-01 -8.34703803e-01 -5.29328287e-01 4.08699960e-01
-3.85759138e-02 5.00654206e-02 2.87932772e-02 -9.50002447e-02
7.63106421e-02 -1.06315836e-01 -1.25805557e+00 8.85974169e-01
1.59325647e+00 -3.06719184e-01 -2.96810478e-01 4.02144909e-01
7.51732051e-01 -8.65819216e-01 -7.71538734e-01 5.98703086e-01
9.05124426e-01 -5.04268765e-01 8.55261445e-01 -6.19093657e-01
3.71055603e-01 -2.04396427e-01 -2.01277554e-01 -1.24119961e+00
-1.86722413e-01 -9.96056795e-01 -1.92680284e-01 5.51495552e-01
-1.61175132e-02 -8.49930108e-01 7.02803254e-01 3.19463164e-01
-2.57003695e-01 -8.44625652e-01 -1.53235996e+00 -8.64397824e-01
-6.43615797e-02 1.63974464e-01 5.62176049e-01 3.57831091e-01
4.29769158e-01 -1.65875256e-01 -2.76574999e-01 2.68354833e-01
3.03269863e-01 2.12911755e-01 9.58095551e-01 -8.64894032e-01
-8.89530182e-01 -3.16953808e-01 -1.64030477e-01 -1.36763239e+00
1.06685013e-01 -3.12313735e-01 7.18920708e-01 -1.19883049e+00
-2.02186003e-01 -9.31998730e-01 -1.11849338e-01 6.87305629e-01
1.93790764e-01 -4.08494025e-01 2.31573209e-01 1.65010437e-01
-8.91195595e-01 8.17660391e-01 1.90988910e+00 -1.38540924e-01
-5.06059051e-01 -1.21654287e-01 -1.95104361e-01 2.62972534e-01
7.04063118e-01 -2.67222703e-01 -5.70721865e-01 -7.50617027e-01
2.43828282e-01 5.53836524e-01 3.37682188e-01 -1.06219375e+00
2.25375727e-01 -6.62278056e-01 4.86733243e-02 -1.66368306e-01
5.70533216e-01 -1.03606415e+00 -9.93154123e-02 6.88421369e-01
-5.13070643e-01 5.72896600e-02 2.94155955e-01 1.09520507e+00
4.44520354e-01 2.34878480e-01 1.20967016e-01 -4.86732461e-02
-2.25264728e-01 9.28153992e-02 -6.81982338e-01 -4.43011761e-01
1.49054623e+00 2.04857484e-01 -6.99753404e-01 -1.83527067e-01
-6.42565668e-01 4.66704905e-01 5.49829602e-01 5.46865761e-01
5.14856160e-01 -1.12618971e+00 5.33392467e-02 2.27724701e-01
-1.24339841e-01 2.67255634e-01 2.34934688e-01 4.55511063e-01
-2.25020185e-01 1.94779918e-01 -4.85253692e-01 -5.77768922e-01
-5.61027527e-01 4.08838868e-01 5.13282232e-02 -2.98225377e-02
-6.64764583e-01 6.26238525e-01 -1.34443954e-01 -3.81395251e-01
2.21002132e-01 -7.05293715e-01 3.89872134e-01 -5.41042030e-01
-8.03183317e-02 2.88261652e-01 -2.02060431e-01 6.06330454e-01
-2.76914686e-01 5.72250664e-01 2.16530412e-01 -2.63351232e-01
1.46082091e+00 1.48408100e-01 -9.82810929e-02 4.35791671e-01
5.34307659e-01 1.96805283e-01 -1.85649633e+00 1.78232402e-01
-2.25644335e-01 -4.96792734e-01 -7.05161631e-01 -8.28647554e-01
-4.30527270e-01 2.49573141e-01 3.32890481e-01 3.64556670e-01
6.97569311e-01 1.84037387e-01 5.02755940e-01 1.09615242e+00
1.33562756e+00 -1.29933488e+00 9.51545715e-01 6.59906209e-01
1.18724692e+00 -8.87268066e-01 -7.46471137e-02 -2.72946477e-01
-3.50384086e-01 9.03976202e-01 6.81960583e-01 -6.62568688e-01
4.25957859e-01 6.20837033e-01 -3.22071046e-01 -3.28356385e-01
-8.14770460e-01 -3.27139162e-02 -3.31653416e-01 8.66465926e-01
-4.60185677e-01 2.85937369e-01 1.46924600e-01 3.12748194e-01
5.43995723e-02 1.06521972e-01 2.20070779e-01 1.77282751e+00
-8.48043740e-01 -1.03987944e+00 -1.96280316e-01 4.53535676e-01
3.72888565e-01 6.27488375e-01 -3.15366983e-01 7.92400420e-01
1.29120588e-01 7.26114810e-01 1.18338773e-02 -3.97514105e-01
5.01753151e-01 -2.96689451e-01 1.05703557e+00 -7.67569542e-01
-3.91053140e-01 -7.41883069e-02 -2.15248346e-01 -9.82747495e-01
-1.47569636e-02 -4.63167995e-01 -1.27587473e+00 1.92519575e-01
-8.38976819e-03 -2.26757944e-01 6.53916121e-01 9.44192529e-01
8.77863646e-01 6.48286462e-01 6.64273500e-01 -1.45610929e+00
-1.16656339e+00 -9.74871516e-01 -6.04834557e-01 1.91145018e-01
4.73089069e-01 -1.27871740e+00 -1.72618434e-01 -4.49956596e-01] | [4.654242515563965, 0.6777402758598328] |
c2c40e37-85d1-4b17-a735-2fa65f36a510 | deep-multi-task-multi-label-cnn-for-effective | 2002.03683 | null | https://arxiv.org/abs/2002.03683v1 | https://arxiv.org/pdf/2002.03683v1.pdf | Deep Multi-task Multi-label CNN for Effective Facial Attribute Classification | Facial Attribute Classification (FAC) has attracted increasing attention in computer vision and pattern recognition. However, state-of-the-art FAC methods perform face detection/alignment and FAC independently. The inherent dependencies between these tasks are not fully exploited. In addition, most methods predict all facial attributes using the same CNN network architecture, which ignores the different learning complexities of facial attributes. To address the above problems, we propose a novel deep multi-task multi-label CNN, termed DMM-CNN, for effective FAC. Specifically, DMM-CNN jointly optimizes two closely-related tasks (i.e., facial landmark detection and FAC) to improve the performance of FAC by taking advantage of multi-task learning. To deal with the diverse learning complexities of facial attributes, we divide the attributes into two groups: objective attributes and subjective attributes. Two different network architectures are respectively designed to extract features for two groups of attributes, and a novel dynamic weighting scheme is proposed to automatically assign the loss weight to each facial attribute during training. Furthermore, an adaptive thresholding strategy is developed to effectively alleviate the problem of class imbalance for multi-label learning. Experimental results on the challenging CelebA and LFWA datasets show the superiority of the proposed DMM-CNN method compared with several state-of-the-art FAC methods. | ['Jing-Hao Xue', 'Longbiao Mao', 'Yan Yan', 'Hanzi Wang'] | 2020-02-10 | null | null | null | null | ['facial-attribute-classification'] | ['computer-vision'] | [ 3.36245745e-02 -4.10707921e-01 -2.34751418e-01 -9.67251301e-01
-6.88187599e-01 7.30769187e-02 2.27448896e-01 -1.16273286e-02
-4.58113253e-01 4.61786270e-01 -6.42605200e-02 2.64907837e-01
-3.22567552e-01 -5.97652674e-01 -3.94956887e-01 -9.64592516e-01
2.76267111e-01 4.95670706e-01 -2.70282447e-01 -8.09729248e-02
-4.23293673e-02 5.00360668e-01 -1.76237047e+00 1.22506477e-01
9.31172729e-01 1.53363776e+00 -1.59529537e-01 -4.68146317e-02
-1.66689098e-01 5.13847709e-01 -3.65995944e-01 -5.31435370e-01
8.75922665e-02 -1.27864197e-01 -5.10011137e-01 3.79849255e-01
7.70370364e-01 -3.52287412e-01 -3.52438837e-02 1.12178695e+00
7.59745479e-01 3.25616822e-02 6.74697399e-01 -1.59523714e+00
-5.98061442e-01 4.73737232e-02 -9.39647377e-01 -7.82578140e-02
-6.54146001e-02 -1.63355097e-01 1.04740584e+00 -1.11404538e+00
-6.55050501e-02 1.51619315e+00 6.86332881e-01 4.99260753e-01
-1.03909791e+00 -1.18051100e+00 3.30553025e-01 4.44405288e-01
-1.88869202e+00 -5.91985703e-01 9.31096852e-01 -3.28996301e-01
2.74849713e-01 -6.72040805e-02 3.97741288e-01 7.77394593e-01
5.55854365e-02 8.66555214e-01 1.08261204e+00 -1.55709714e-01
-2.09960401e-01 2.94925142e-02 -9.45962593e-02 1.02442205e+00
4.86991182e-02 -2.02692986e-01 -3.20622951e-01 -1.77780747e-01
4.96446609e-01 1.78505585e-01 -3.65137197e-02 -1.51656911e-01
-7.86264718e-01 8.69533479e-01 1.40638769e-01 -1.01887714e-02
-4.13755327e-01 1.15194872e-01 5.30959606e-01 4.27924357e-02
6.93581402e-01 -2.06040859e-01 -4.96561617e-01 3.74127865e-01
-7.77304590e-01 1.42455891e-01 2.13455707e-01 9.46677744e-01
1.04194653e+00 1.02149338e-01 -3.27930421e-01 1.26194799e+00
5.99702418e-01 4.55496132e-01 3.70244235e-01 -6.82062507e-01
3.79318893e-01 7.57593513e-01 -3.54453355e-01 -1.33381116e+00
-5.50143898e-01 -5.03516197e-01 -1.13756156e+00 1.13845922e-01
2.82057643e-01 -8.08044523e-02 -8.66666853e-01 1.98307538e+00
5.45688272e-01 2.15929732e-01 -1.69380382e-01 8.09041619e-01
9.69799280e-01 4.17931646e-01 4.71673489e-01 -1.84018552e-01
1.45708680e+00 -8.23852718e-01 -9.02292132e-01 -2.22288743e-01
3.33140045e-01 -7.52494395e-01 7.81745374e-01 1.00444250e-01
-7.88386643e-01 -7.29972959e-01 -1.04535365e+00 7.14427829e-02
-1.61082298e-01 6.07936025e-01 6.45021915e-01 4.53951895e-01
-7.64327765e-01 2.02354148e-01 -4.49133068e-01 2.64430773e-02
1.04550314e+00 8.12287211e-01 -4.04719383e-01 -1.66555271e-01
-1.18098104e+00 6.57652020e-01 3.23288769e-01 5.18834651e-01
-9.58018780e-01 -5.18509090e-01 -9.28534687e-01 6.10512458e-02
6.01585567e-01 -4.55920845e-01 1.05308604e+00 -1.30540681e+00
-1.36651421e+00 9.09677804e-01 -1.64958119e-01 2.54713476e-01
3.10423553e-01 -2.37742499e-01 -5.91869950e-01 -7.23429099e-02
1.90143153e-01 5.68383932e-01 1.03165889e+00 -1.16039765e+00
-8.76107216e-01 -6.86925173e-01 -2.96587318e-01 2.46289998e-01
-8.41382325e-01 2.39305362e-01 -5.41813672e-01 -5.90160370e-01
-2.16463208e-02 -7.54390538e-01 -2.74909139e-02 2.62639701e-01
-3.21093380e-01 -6.14163756e-01 9.79349494e-01 -4.05128777e-01
1.11093903e+00 -2.18838620e+00 1.75892934e-01 2.24913284e-01
3.02426666e-01 3.68984431e-01 -4.02590573e-01 -2.33724356e-01
-5.44720031e-02 -9.39425156e-02 -1.45641789e-01 -7.08138704e-01
-1.08494595e-01 2.41645992e-01 2.50330329e-01 6.61988676e-01
4.49208736e-01 5.78115821e-01 -5.59589148e-01 -9.18448567e-01
8.95186607e-03 6.11932158e-01 -2.97674954e-01 3.00752401e-01
-1.05910018e-01 4.49142456e-01 -8.24715614e-01 1.23648834e+00
1.02728891e+00 -1.73021600e-01 1.61738247e-02 -4.63955283e-01
6.67346641e-02 -3.86509567e-01 -1.22816348e+00 1.48216057e+00
-5.40229499e-01 1.36235818e-01 2.89134711e-01 -1.05207002e+00
1.10947990e+00 4.43465739e-01 6.42062724e-01 -7.40950286e-01
4.56052125e-01 3.51626754e-01 -3.70733738e-02 -5.70744038e-01
1.34826154e-01 -3.13397586e-01 9.58653912e-02 3.33282530e-01
2.71395713e-01 4.74947423e-01 1.95331685e-03 -4.11857545e-01
2.56650776e-01 1.15266487e-01 4.11117017e-01 -3.11638027e-01
1.02383697e+00 -4.48399186e-01 1.28088820e+00 1.57900006e-01
-4.54417586e-01 4.14757699e-01 3.34660858e-01 -8.21446478e-01
-7.74350643e-01 -5.35092175e-01 -1.83536932e-01 1.49146008e+00
1.64211571e-01 -2.39252731e-01 -7.02470124e-01 -8.19813609e-01
2.29906932e-01 2.07049251e-02 -7.74467289e-01 -2.12443218e-01
-6.43201649e-01 -1.31677067e+00 6.04543805e-01 6.00943983e-01
7.41228580e-01 -9.90149021e-01 -1.85480103e-01 2.01588914e-01
-2.41185844e-01 -1.13335681e+00 -4.88707930e-01 -1.94345415e-01
-4.82811779e-01 -1.00728536e+00 -6.32140577e-01 -8.82486582e-01
6.54470086e-01 1.66921601e-01 8.95443559e-01 3.81369293e-01
-2.46867418e-01 8.81798044e-02 -1.82799637e-01 -5.65506756e-01
-6.29416341e-03 2.52015263e-01 1.78670987e-01 9.85662937e-01
5.68800628e-01 -5.98685801e-01 -5.95825911e-01 5.08910894e-01
-7.61225104e-01 -7.01996610e-02 1.04445100e+00 9.27632213e-01
8.82960558e-01 -2.68627182e-02 7.96741545e-01 -8.39900136e-01
4.11831409e-01 -5.03515422e-01 -3.78256887e-01 4.07001078e-01
-7.55641818e-01 -8.73368755e-02 6.78928375e-01 -5.07386208e-01
-1.00516391e+00 2.78704673e-01 -2.90262878e-01 -5.74935913e-01
-2.66149461e-01 4.43667114e-01 -7.94270992e-01 -3.32379460e-01
1.08490670e-02 2.03537747e-01 2.11798161e-01 -3.49472910e-01
9.24443901e-02 7.04963863e-01 3.47033650e-01 -6.94152594e-01
5.80646932e-01 4.56678808e-01 4.00907516e-01 -4.62099254e-01
-1.28177404e+00 -3.41649443e-01 -6.01646364e-01 -3.56996119e-01
9.76505339e-01 -1.02687597e+00 -1.04421699e+00 1.03401947e+00
-1.03020835e+00 1.11405492e-01 3.15023035e-01 3.06343377e-01
-4.55123395e-01 2.32567698e-01 -4.90108401e-01 -6.78532422e-01
-4.37891036e-01 -1.38206673e+00 1.18132377e+00 4.96833622e-01
2.47404233e-01 -8.27867270e-01 -3.32370222e-01 4.23211187e-01
3.82067502e-01 3.76939923e-01 9.87253308e-01 -6.36177003e-01
-3.67932945e-01 -1.18680395e-01 -5.21241486e-01 4.31037515e-01
3.20727944e-01 1.20995594e-02 -1.19558871e+00 -3.79248470e-01
-2.34201059e-01 -5.89227736e-01 8.32344115e-01 3.39738995e-01
1.59742641e+00 -2.53395885e-01 -2.50811279e-01 8.57741058e-01
1.39949036e+00 2.34459355e-01 3.50822777e-01 4.03779328e-01
1.00474656e+00 6.90578580e-01 8.20078135e-01 6.18561924e-01
5.79104185e-01 8.39210391e-01 7.26524532e-01 -2.95043558e-01
1.73131779e-01 1.77322879e-01 -1.32919893e-01 6.67598426e-01
-1.84856012e-01 -9.39186886e-02 -7.72607386e-01 3.32015634e-01
-1.91267574e+00 -4.70579684e-01 1.16022460e-01 1.88246608e+00
7.13072658e-01 -1.86901927e-01 8.08373615e-02 1.12267137e-02
9.68380094e-01 2.84698278e-01 -5.60599029e-01 2.41687577e-02
-2.81145185e-01 5.24240956e-02 2.05755875e-01 5.24736978e-02
-1.61346877e+00 7.15473711e-01 4.95095873e+00 1.16996717e+00
-1.12678444e+00 2.46201307e-01 8.89705658e-01 1.13224186e-01
1.50702715e-01 -4.11054909e-01 -1.09272373e+00 5.13116121e-01
4.61739898e-01 1.85395584e-01 1.69900015e-01 8.19084823e-01
-3.61367241e-02 2.11364508e-01 -8.46246600e-01 1.15089142e+00
3.22405308e-01 -8.06486428e-01 3.11299145e-01 4.40764986e-02
4.68417972e-01 -4.70911175e-01 3.22714508e-01 1.91062331e-01
-1.23942241e-01 -1.09536171e+00 6.67015135e-01 6.30164504e-01
1.07736194e+00 -1.15382993e+00 9.51133192e-01 -6.78429496e-04
-1.59070969e+00 -3.81659329e-01 -3.65086198e-01 2.63832211e-01
-9.20190290e-03 6.08731925e-01 -2.80955195e-01 6.92207217e-01
6.85678899e-01 7.88356900e-01 -6.89068317e-01 8.64125729e-01
-8.35886598e-03 1.64805904e-01 -3.08978073e-02 2.34003708e-01
2.85150737e-01 -2.54277796e-01 1.72915757e-01 9.76511538e-01
2.73752362e-01 1.66226655e-01 4.90279287e-01 5.50060213e-01
-2.35058114e-01 4.29846376e-01 -5.76353781e-02 1.67453840e-01
4.61785704e-01 1.69158888e+00 -3.12863916e-01 -1.01769209e-01
-7.14289188e-01 7.09845304e-01 4.65803653e-01 5.96565306e-02
-7.60316193e-01 -3.38764280e-01 9.09622967e-01 -1.65959001e-01
1.42613634e-01 3.37846093e-02 -1.60407826e-01 -8.01986098e-01
3.47179212e-02 -8.82831097e-01 6.36736631e-01 -3.50176096e-01
-1.38784575e+00 7.31824577e-01 -2.01282024e-01 -1.35094464e+00
1.39995947e-01 -5.40734291e-01 -5.52328765e-01 8.85786474e-01
-1.98721135e+00 -1.70315135e+00 -7.01398909e-01 8.06765854e-01
4.38801527e-01 -5.08499503e-01 7.33016193e-01 8.26783061e-01
-1.00734818e+00 9.80055273e-01 -5.73586747e-02 3.49313438e-01
1.11486518e+00 -9.13112998e-01 -2.04804555e-01 5.09091258e-01
-3.99489135e-01 3.64688218e-01 1.30827531e-01 -3.19507927e-01
-1.09642684e+00 -1.56434178e+00 6.90679908e-01 8.93646181e-02
2.07930505e-01 -2.47073412e-01 -1.03285527e+00 5.25498211e-01
-1.57206386e-01 4.47880805e-01 9.48394358e-01 1.47161454e-01
-4.59537864e-01 -6.68214858e-01 -9.56155717e-01 2.52104521e-01
6.83442473e-01 -3.27388406e-01 -2.03123186e-02 2.88234502e-01
3.50664437e-01 -2.64848769e-01 -9.09727991e-01 8.46620917e-01
7.49709368e-01 -7.42314517e-01 8.71946871e-01 -5.07261217e-01
3.35227281e-01 -3.29044878e-01 -2.01682165e-01 -1.31160975e+00
-4.84873265e-01 -2.11820364e-01 -3.16898674e-02 1.56923175e+00
8.15405026e-02 -5.89454055e-01 5.93032479e-01 2.58714288e-01
-2.00914025e-01 -1.19645238e+00 -9.46801066e-01 -4.99906719e-01
-2.21726581e-01 6.79465337e-03 9.17056262e-01 1.01634741e+00
-8.23219240e-01 5.29774129e-01 -6.76544666e-01 1.71614453e-01
8.44393253e-01 2.29199916e-01 5.61799109e-01 -1.69665301e+00
1.93118244e-01 -5.29125571e-01 -4.93022501e-01 -5.94427884e-01
5.61276436e-01 -7.65719950e-01 -2.23905891e-02 -1.16799045e+00
4.93391037e-01 -6.98278785e-01 -5.41105807e-01 7.64168561e-01
-4.01038885e-01 3.21491510e-01 3.37216742e-02 2.55036891e-01
-6.91673160e-01 9.81434524e-01 1.32179761e+00 -2.12765917e-01
2.04669148e-01 1.20111652e-01 -8.55504572e-01 7.88592935e-01
6.96438551e-01 -4.91763473e-01 -1.94711879e-01 -4.87670898e-01
-4.27002199e-02 -4.09729257e-02 3.04569453e-01 -9.43212271e-01
2.84488648e-01 -3.37193191e-01 6.52891874e-01 -5.45672655e-01
4.17700738e-01 -8.59710455e-01 -1.98520198e-01 1.16430394e-01
-1.52804270e-01 8.54753777e-02 2.67990500e-01 5.29485703e-01
-4.53508198e-01 1.21211320e-01 1.12671888e+00 -2.39930432e-02
-8.84994507e-01 1.03241777e+00 -2.90177111e-02 -1.48848519e-01
1.06238019e+00 -2.41489485e-02 -1.54427707e-01 -1.67511359e-01
-6.02822244e-01 5.91858029e-01 -3.61182578e-02 4.44963485e-01
7.35910237e-01 -1.75157547e+00 -9.97449577e-01 3.27544063e-01
3.34671497e-01 1.53970242e-01 4.71298188e-01 9.40958560e-01
-1.29958674e-01 1.46950364e-01 -5.78534245e-01 -5.64762056e-01
-1.42953694e+00 4.11253959e-01 4.63295162e-01 -7.35382363e-02
-8.48054327e-03 8.02878678e-01 3.91998410e-01 -4.84354794e-01
1.86653703e-01 4.24438953e-01 -5.82593381e-01 4.86216217e-01
5.76290131e-01 2.63166606e-01 2.02646852e-01 -9.96033967e-01
-4.90234822e-01 1.00093126e+00 -4.18068767e-01 4.42300260e-01
1.39814186e+00 -2.45789722e-01 -4.05129433e-01 1.98845044e-01
1.33471859e+00 -3.96185040e-01 -1.19170380e+00 -4.97954011e-01
-1.04436591e-01 -4.09844071e-01 1.18057802e-01 -5.04054010e-01
-1.61749601e+00 1.04581404e+00 7.85952330e-01 -4.50822830e-01
1.58920562e+00 -3.34701270e-01 8.42635214e-01 1.15146391e-01
1.22802779e-01 -1.24675107e+00 1.87883139e-01 3.21649075e-01
5.75168610e-01 -1.41600835e+00 6.14877567e-02 -5.47605515e-01
-5.46276927e-01 1.19235110e+00 1.14105499e+00 2.75827557e-01
6.49557412e-01 -2.07024310e-02 1.62993968e-01 -2.41502523e-01
-3.39182466e-01 -1.66523233e-01 4.52176958e-01 2.54983932e-01
3.27033311e-01 -8.06258544e-02 -2.27969557e-01 7.39522338e-01
4.56117719e-01 -5.70232002e-03 1.61071084e-02 7.69147575e-01
-3.86819780e-01 -1.00854278e+00 -3.61563951e-01 5.55588663e-01
-7.05903828e-01 7.15153068e-02 -6.21839054e-02 7.87123382e-01
5.82613587e-01 7.19400525e-01 8.76037776e-02 -6.46975279e-01
2.30257526e-01 -1.06530013e-02 2.34857291e-01 -3.49969089e-01
-3.63313794e-01 1.12018786e-01 -2.25739837e-01 -4.85117525e-01
-7.65572131e-01 -6.76357269e-01 -9.38248515e-01 -1.70280084e-01
-3.95450741e-01 -7.71961585e-02 4.57984000e-01 1.14401090e+00
3.25246900e-01 4.87967819e-01 9.46295202e-01 -6.83894932e-01
-5.50372005e-01 -9.61072981e-01 -6.42612219e-01 4.24620777e-01
4.10015255e-01 -1.15277815e+00 -9.53466520e-02 -1.76177084e-01] | [13.531457901000977, 0.842373251914978] |
fa7f2f6b-5bac-46f4-a8bd-61632a78cd3f | multi-view-and-cross-view-brain-decoding | null | null | https://aclanthology.org/2022.coling-1.10 | https://aclanthology.org/2022.coling-1.10.pdf | Multi-view and Cross-view Brain Decoding | Can we build multi-view decoders that can decode concepts from brain recordings corresponding to any view (picture, sentence, word cloud) of stimuli? Can we build a system that can use brain recordings to automatically describe what a subject is watching using keywords or sentences? How about a system that can automatically extract important keywords from sentences that a subject is reading? Previous brain decoding efforts have focused only on single view analysis and hence cannot help us build such systems. As a first step toward building such systems, inspired by Natural Language Processing literature on multi-lingual and cross-lingual modeling, we propose two novel brain decoding setups: (1) multi-view decoding (MVD) and (2) cross-view decoding (CVD). In MVD, the goal is to build an MV decoder that can take brain recordings for any view as input and predict the concept. In CVD, the goal is to train a model which takes brain recordings for one view as input and decodes a semantic vector representation of another view. Specifically, we study practically useful CVD tasks like image captioning, image tagging, keyword extraction, and sentence formation. Our extensive experiments lead to MVD models with ~0.68 average pairwise accuracy across view pairs, and also CVD models with ~0.8 average pairwise accuracy across tasks. Analysis of the contribution of different brain networks reveals exciting cognitive insights: (1) Models trained on picture or sentence view of stimuli are better MV decoders than a model trained on word cloud view. (2) Our extensive analysis across 9 broad regions, 11 language sub-regions and 16 visual sub-regions of the brain help us localize, for the first time, the parts of the brain involved in cross-view tasks like image captioning, image tagging, sentence formation and keyword extraction. We make the code publicly available. | ['Raju S. Bapi', 'Manish Gupta', 'Jashn Arora', 'Subba Reddy Oota'] | null | null | null | null | coling-2022-10 | ['brain-decoding', 'brain-decoding', 'keyword-extraction'] | ['medical', 'miscellaneous', 'natural-language-processing'] | [ 3.57398778e-01 2.41167113e-01 2.94680834e-01 -5.86387277e-01
-9.73525703e-01 -7.87045181e-01 8.77758145e-01 -2.81302333e-02
-4.05782819e-01 2.79002637e-01 4.40101862e-01 -2.49977648e-01
4.86791641e-01 -4.29858238e-01 -9.83281732e-01 -4.96622175e-01
3.81194383e-01 3.91115606e-01 1.19042639e-02 -8.42287168e-02
2.93648958e-01 9.49503630e-02 -1.21646595e+00 7.56431043e-01
2.47623563e-01 9.77542162e-01 8.30173910e-01 6.75879478e-01
-1.20067291e-01 5.42013943e-01 -5.30604422e-01 -6.85592175e-01
-1.84343159e-02 -4.25653458e-01 -8.21708322e-01 1.26801297e-01
7.25424349e-01 -1.15611002e-01 -1.04736753e-01 1.08401871e+00
5.99784672e-01 -6.24103963e-01 7.67330229e-01 -1.06984794e+00
-9.52188075e-01 6.50565028e-01 -4.77984339e-01 2.88828284e-01
3.38837743e-01 1.85385346e-01 9.27775800e-01 -1.02923989e+00
8.52776825e-01 1.10500693e+00 1.71079174e-01 9.64343727e-01
-1.23774469e+00 -5.93936384e-01 3.50528747e-01 7.51107186e-02
-1.23341072e+00 -5.87710142e-01 4.49094653e-01 -7.51871347e-01
1.22319794e+00 1.68489918e-01 6.32038295e-01 1.56297493e+00
4.26109761e-01 8.18617940e-01 1.37464881e+00 -8.78568366e-02
-6.67269453e-02 3.76579642e-01 1.05361886e-01 6.71218157e-01
3.96065041e-02 -3.97225291e-01 -7.80595124e-01 2.43152112e-01
5.36585331e-01 -2.51848400e-01 -3.88259262e-01 2.39901356e-02
-1.73303139e+00 6.66829228e-01 1.35294765e-01 5.04243314e-01
-3.65349680e-01 9.51164141e-02 4.64198560e-01 2.82626063e-01
5.07737756e-01 4.98003393e-01 -5.26166797e-01 9.18515474e-02
-1.13871419e+00 -1.37876093e-01 7.39708722e-01 1.03374636e+00
4.86349016e-01 -1.38703540e-01 -4.16397244e-01 8.69299889e-01
2.73641378e-01 7.89458334e-01 5.58057070e-01 -5.71283638e-01
9.09860730e-01 3.69120479e-01 -4.83745039e-01 -7.27611780e-01
-4.42225367e-01 -3.03026199e-01 -6.65181100e-01 -2.71748513e-01
6.17890060e-02 -1.72469258e-01 -9.76975799e-01 1.93587816e+00
-2.67644048e-01 -2.15862677e-01 1.94363579e-01 6.25120997e-01
1.19221425e+00 4.87710178e-01 -1.62666216e-02 5.38663715e-02
2.20625401e+00 -9.28314209e-01 -3.91608983e-01 -6.13195837e-01
5.69611549e-01 -6.18227899e-01 8.87179673e-01 3.88036877e-01
-1.20253265e+00 -7.10741520e-01 -9.44238365e-01 -2.69590050e-01
-5.73124766e-01 2.64703244e-01 1.48501635e-01 2.92931527e-01
-1.54880285e+00 -1.88042253e-01 -5.23850441e-01 -5.60679734e-01
5.52700102e-01 3.25352252e-01 -8.94313276e-01 -1.54779449e-01
-9.74813879e-01 1.17373455e+00 1.56883284e-01 -1.08349673e-01
-1.34851205e+00 -5.20487607e-01 -8.72756481e-01 9.32376236e-02
1.57099739e-01 -1.13855302e+00 8.75824809e-01 -1.17468059e+00
-9.35585856e-01 1.42753136e+00 -5.56469560e-01 -4.53631759e-01
-4.44795340e-02 1.45832105e-02 -3.91995877e-01 5.27787805e-01
3.47633481e-01 1.24919236e+00 9.82586324e-01 -1.37645960e+00
-2.94924617e-01 -6.78885341e-01 -4.92645763e-02 2.13324130e-01
-2.61993170e-01 1.99987546e-01 -5.61087072e-01 -4.92396176e-01
-4.26174290e-02 -8.16115916e-01 2.63831764e-01 -3.47491533e-01
-5.76905906e-01 -8.62930641e-02 4.35781986e-01 -1.16909254e+00
6.20193243e-01 -1.97041821e+00 4.09952581e-01 -1.86853170e-01
3.72876614e-01 -1.11473784e-01 -3.96666080e-01 4.03661609e-01
-1.86088800e-01 2.58103937e-01 -2.35577628e-01 -5.27834535e-01
-1.63090885e-01 7.63800591e-02 -3.97520542e-01 2.44144946e-01
3.51242542e-01 1.10863185e+00 -6.07553601e-01 -4.61134404e-01
-1.84109658e-01 4.64225531e-01 -5.18908262e-01 2.38391265e-01
-9.93271098e-02 4.71854389e-01 -1.43366098e-01 4.54089075e-01
3.78089339e-01 -3.83350223e-01 3.01865973e-02 -5.30870080e-01
2.75480915e-02 3.46828431e-01 -2.91181356e-01 1.99052453e+00
-7.30069757e-01 8.63609493e-01 1.95964873e-01 -1.30485380e+00
6.79575264e-01 6.23967886e-01 2.96333253e-01 -8.00764024e-01
1.14654742e-01 -1.10451043e-01 -3.81932743e-02 -8.60497773e-01
7.04442561e-02 -1.66352272e-01 -1.63100988e-01 5.50428569e-01
8.76888156e-01 -5.86361289e-02 -2.67483713e-03 5.03028929e-01
1.11635923e+00 -2.51358468e-02 8.21895003e-02 -1.14127979e-01
6.05953217e-01 -2.06347018e-01 -2.71699447e-02 6.08826995e-01
1.51020393e-01 9.66783643e-01 7.97738969e-01 -1.84993565e-01
-9.49206173e-01 -1.00357533e+00 -4.85377982e-02 9.95940745e-01
-1.06489688e-01 -4.47279483e-01 -1.08642209e+00 -6.25291467e-01
-4.77461070e-01 8.33585024e-01 -6.36378407e-01 1.78914610e-02
-3.96252304e-01 -5.50297320e-01 7.32575417e-01 5.54110110e-01
5.00668764e-01 -1.00494266e+00 -6.45271242e-01 -1.49150357e-01
-5.57737947e-01 -1.74376965e+00 -5.12471557e-01 1.41928107e-01
-6.28644228e-01 -8.11022401e-01 -8.38436902e-01 -1.03251123e+00
8.09196413e-01 2.17307433e-01 1.26709855e+00 -1.73381105e-01
-1.06796078e-01 9.04552996e-01 -6.13430962e-02 -4.35564101e-01
-2.70600557e-01 5.15313484e-02 2.57842280e-02 9.35838073e-02
4.05725658e-01 -5.11877596e-01 -6.47392869e-01 5.64799495e-02
-7.84316123e-01 5.05436778e-01 8.96124840e-01 6.83594465e-01
5.51514328e-01 -8.53963017e-01 4.31946158e-01 -7.12368727e-01
5.79549968e-01 -7.01385260e-01 -3.67420226e-01 5.95758379e-01
-3.56157154e-01 8.01476464e-02 4.10921812e-01 -2.42651060e-01
-8.04855406e-01 2.49966606e-01 -4.59157437e-01 -3.34022909e-01
-4.50954616e-01 2.68978775e-01 -2.47286543e-01 2.83332080e-01
3.65671426e-01 7.76315033e-01 2.59813145e-02 -3.31600994e-01
5.77884436e-01 7.47991800e-01 5.84920824e-01 -3.34187627e-01
3.54538381e-01 5.48678935e-01 -3.97980303e-01 -8.43265474e-01
-7.12269366e-01 -4.00733471e-01 -9.00145411e-01 -3.12472314e-01
1.70004094e+00 -1.35265040e+00 -6.99482501e-01 1.01190478e-01
-1.57019186e+00 -6.36753365e-02 3.74583811e-01 2.76908785e-01
-6.64047658e-01 8.36561620e-02 -2.20463142e-01 -4.24063295e-01
-5.16714513e-01 -1.39951682e+00 1.49829853e+00 -5.38909994e-02
-3.29843909e-01 -1.07886922e+00 -1.65028080e-01 8.50567400e-01
1.37287781e-01 -1.87758178e-01 9.54134226e-01 -1.04543364e+00
-3.63489896e-01 6.51277155e-02 -3.27821821e-01 4.10257131e-01
-1.95937455e-01 -4.66564894e-01 -1.35718560e+00 -1.33937180e-01
3.45241696e-01 -2.93592572e-01 9.81904328e-01 4.21645194e-01
1.11210728e+00 -1.97931752e-01 -4.33747739e-01 6.06565893e-01
1.35584664e+00 1.60728112e-01 5.77548206e-01 1.92453191e-02
7.81440616e-01 8.65453899e-01 -1.98549837e-01 -1.73017517e-01
8.30366790e-01 6.52853012e-01 3.75854790e-01 -1.07977986e-01
-2.32712567e-01 -2.66657919e-01 8.89524102e-01 1.27648294e+00
2.06917539e-01 -3.30889821e-01 -8.87180984e-01 9.15111125e-01
-1.42392111e+00 -1.01789796e+00 -1.40210867e-01 1.75835919e+00
6.04804039e-01 -1.79834221e-03 -1.30827382e-01 -5.90415895e-01
4.25737679e-01 1.55471101e-01 -3.82169098e-01 -7.71587372e-01
-1.81952253e-01 1.15639031e-01 2.73896933e-01 2.34970793e-01
-9.58163977e-01 9.30515885e-01 5.93581820e+00 5.24486840e-01
-1.22809803e+00 5.96990228e-01 7.59063780e-01 -2.85870910e-01
-3.75422597e-01 -1.39008179e-01 -9.94399786e-01 3.91864717e-01
1.25746548e+00 3.25540662e-01 3.89908820e-01 5.15883565e-01
-3.57907303e-02 -2.08567813e-01 -1.51479363e+00 1.26084256e+00
8.42750907e-01 -1.39587212e+00 3.17829043e-01 1.91670910e-01
4.53444749e-01 5.17202795e-01 4.19655330e-02 5.67082204e-02
-1.04712829e-01 -1.21510065e+00 9.65080857e-01 7.42519498e-01
8.13749194e-01 -7.10678920e-02 5.72679222e-01 5.24589777e-01
-1.04418755e+00 1.84625760e-01 -2.23489136e-01 2.48942658e-01
2.28527188e-01 2.92544663e-01 -8.13659370e-01 3.57705504e-01
5.81691980e-01 7.55134284e-01 -8.33863914e-01 5.56392252e-01
-2.34532237e-01 5.38858593e-01 1.47135913e-01 7.63079524e-02
1.83572412e-01 -5.33027239e-02 4.17991847e-01 1.35664749e+00
4.63252187e-01 7.29352608e-02 -1.97565109e-01 1.24329638e+00
-2.32698977e-01 7.50667974e-02 -1.10269690e+00 -1.35059029e-01
7.86118358e-02 1.49410653e+00 -9.01294589e-01 -4.51956302e-01
-8.19891453e-01 1.23078442e+00 2.96386808e-01 3.16604137e-01
-6.50163531e-01 4.10203524e-02 6.30051374e-01 -8.15419108e-02
4.60433424e-01 -2.63520390e-01 -4.55193460e-01 -1.45649612e+00
6.66138008e-02 -6.69227839e-01 5.42651713e-02 -1.24261844e+00
-1.30752838e+00 8.73270452e-01 6.66011451e-03 -8.95557940e-01
-2.05896065e-01 -7.88866341e-01 -4.57279950e-01 1.00346994e+00
-1.16067874e+00 -1.36148822e+00 -1.98290050e-02 7.42938519e-01
9.87887442e-01 -4.11425084e-01 9.38685179e-01 2.34744415e-01
-2.94430852e-01 3.15340221e-01 -3.07785362e-01 3.78857464e-01
6.68646812e-01 -1.03047895e+00 5.68688810e-01 7.92638600e-01
7.98270822e-01 7.82025337e-01 2.85002798e-01 -4.78243291e-01
-1.48939681e+00 -9.13109124e-01 9.97344136e-01 -9.73087549e-01
5.28011084e-01 -1.03143811e+00 -6.95253909e-01 7.01102197e-01
7.31436908e-01 -2.85397500e-01 6.96636021e-01 -1.67233780e-01
-4.60467577e-01 3.39385904e-02 -7.28620291e-01 4.50387388e-01
9.05384600e-01 -1.03865838e+00 -6.86671793e-01 5.53957999e-01
7.69609749e-01 -6.51485547e-02 -9.45806563e-01 -6.14681914e-02
6.08277500e-01 -7.71111071e-01 1.14400649e+00 -5.36392152e-01
8.50214183e-01 4.06857654e-02 -2.14219123e-01 -1.38719296e+00
-3.52869555e-02 1.81571003e-02 2.93320924e-01 1.36419117e+00
7.93437600e-01 -5.07618129e-01 4.66308072e-02 3.70494097e-01
-1.84695959e-01 -7.53278911e-01 -8.91737521e-01 -4.58658159e-01
7.83048943e-02 -6.22898936e-01 1.71745762e-01 7.04077959e-01
-7.81696290e-02 9.20081556e-01 -2.01238260e-01 1.29518956e-01
3.17601860e-01 4.28245813e-02 4.03491378e-01 -9.37984526e-01
-9.19647440e-02 -2.30884895e-01 -2.79949218e-01 -9.75836873e-01
3.30648571e-01 -1.27866447e+00 1.55907318e-01 -1.90689540e+00
6.59713447e-01 2.73877233e-01 -7.52890557e-02 5.23056865e-01
7.41871521e-02 3.37785482e-01 3.37252647e-01 1.61030129e-01
-4.19745773e-01 7.16669187e-02 1.25404346e+00 -2.01845974e-01
5.20941615e-01 -3.22972804e-01 -1.12390566e+00 7.72699356e-01
5.35166919e-01 -4.78796393e-01 -4.48418587e-01 -9.48075652e-01
4.91357565e-01 2.81389356e-01 6.31331921e-01 -9.03306365e-01
3.04738015e-01 4.03812051e-01 4.72380191e-01 -3.49543214e-01
5.55771470e-01 -6.56568706e-01 -2.22348258e-01 2.29470178e-01
-4.37233329e-01 4.05001968e-01 9.39466506e-02 4.23519373e-01
-3.01168114e-01 -2.02750325e-01 4.65205073e-01 -6.28124654e-01
-6.24757767e-01 2.22892817e-02 -6.86219573e-01 2.23807305e-01
7.58208156e-01 -7.13046715e-02 -5.01225948e-01 -4.83867884e-01
-9.37149882e-01 1.46094441e-01 2.10855827e-01 5.30238926e-01
7.90585339e-01 -1.09425807e+00 -7.49703288e-01 2.37330303e-01
2.64326572e-01 -4.61815745e-01 -7.00740367e-02 1.23710954e+00
-9.02413577e-02 8.47925007e-01 -2.50496984e-01 -6.48310363e-01
-1.38331258e+00 5.52282095e-01 1.33238107e-01 1.39485449e-02
-5.55564344e-01 1.17910850e+00 8.91539395e-01 -2.22692057e-01
-1.08786538e-01 -4.66892391e-01 -4.47051376e-01 4.27029133e-01
5.71694493e-01 -3.93801451e-01 1.17720678e-01 -1.08911276e+00
-4.47244257e-01 6.83151007e-01 -3.49569134e-02 -5.52500129e-01
1.53972149e+00 -2.16724291e-01 -4.15298134e-01 6.52076781e-01
1.45560348e+00 -2.37611413e-01 -8.29423070e-01 6.53590634e-02
-1.42444402e-01 2.22954974e-01 4.63844351e-02 -9.49786484e-01
-1.14275575e+00 1.54958475e+00 6.03002131e-01 8.07659924e-02
1.15200841e+00 3.74494255e-01 9.02123928e-01 1.10790566e-01
3.21269035e-01 -9.66091037e-01 8.08454081e-02 5.03569603e-01
1.10513580e+00 -1.19055688e+00 -3.37467045e-01 -5.91961555e-02
-1.12030709e+00 1.07363999e+00 5.67663431e-01 7.61259720e-02
5.05329251e-01 1.92364588e-01 -1.88345879e-01 -7.13133931e-01
-1.04768622e+00 -2.84788281e-01 6.09031081e-01 4.85019088e-01
4.77139384e-01 1.01400465e-01 4.86258864e-02 8.65061462e-01
-2.43633881e-01 -3.30348492e-01 2.94349045e-01 3.76452774e-01
-3.18349749e-01 -9.03279603e-01 -2.06313431e-01 6.55191183e-01
-5.41779816e-01 -5.44201612e-01 -6.87144160e-01 3.88852686e-01
3.99320483e-01 8.35498989e-01 1.25613451e-01 -5.29558003e-01
1.09975398e-01 6.63735271e-01 6.48094356e-01 -9.69428241e-01
-7.46923268e-01 -2.30438951e-02 1.10421389e-01 -5.41502595e-01
-5.23877621e-01 -7.73852408e-01 -9.94106710e-01 3.90392929e-01
4.35909033e-02 -2.34149098e-01 1.03976071e+00 1.36455047e+00
4.62128103e-01 4.77945417e-01 1.37363598e-01 -7.04347014e-01
-2.86840647e-02 -9.61067140e-01 -1.77494273e-01 2.60268509e-01
1.57795042e-01 -4.37783301e-01 1.41395843e-02 6.41777754e-01] | [10.94041633605957, 1.688606858253479] |
f4556680-7f2c-4ebe-9b17-cad2ff90d619 | 3d-molecule-generation-by-denoising-voxel | 2306.07473 | null | https://arxiv.org/abs/2306.07473v1 | https://arxiv.org/pdf/2306.07473v1.pdf | 3D molecule generation by denoising voxel grids | We propose a new score-based approach to generate 3D molecules represented as atomic densities on regular grids. First, we train a denoising neural network that learns to map from a smooth distribution of noisy molecules to the distribution of real molecules. Then, we follow the neural empirical Bayes framework [Saremi and Hyvarinen, 2019] and generate molecules in two steps: (i) sample noisy density grids from a smooth distribution via underdamped Langevin Markov chain Monte Carlo, and (ii) recover the ``clean'' molecule by denoising the noisy grid with a single step. Our method, VoxMol, generates molecules in a fundamentally different way than the current state of the art (i.e., diffusion models applied to atom point clouds). It differs in terms of the data representation, the noise model, the network architecture and the generative modeling algorithm. VoxMol achieves comparable results to state of the art on unconditional 3D molecule generation while being simpler to train and faster to generate molecules. | ['Saeed Saremi', 'Vishnu Sresht', 'Stephen Ra', 'Andrew Martin Watkins', 'Omar Mahmood', 'Michael Maser', 'Joseph Kleinhenz', 'Joshua Rackers', 'Pedro O. Pinheiro'] | 2023-06-13 | null | null | null | null | ['3d-molecule-generation'] | ['medical'] | [ 4.79216635e-01 1.79547831e-01 2.40437463e-01 -1.23580560e-01
-8.06701660e-01 -6.63931906e-01 9.45920110e-01 3.33006121e-02
-3.76302630e-01 1.37608492e+00 -6.26443475e-02 -2.54666239e-01
3.02500695e-01 -1.30380785e+00 -1.06615901e+00 -1.06487370e+00
2.75156468e-01 1.20642483e+00 -1.73116308e-02 -5.26870377e-02
1.39230669e-01 6.83110237e-01 -1.25023758e+00 7.63794035e-02
9.46299374e-01 5.53823233e-01 6.47336319e-02 8.06680918e-01
-1.09635279e-01 7.82528460e-01 -5.71782291e-01 -4.37966377e-01
-9.17222276e-02 -1.04761744e+00 -5.86941719e-01 -1.66769713e-01
1.89418405e-01 -6.96574077e-02 -7.79680759e-02 1.18805540e+00
6.82775736e-01 2.99569607e-01 1.38028705e+00 -5.39645314e-01
-7.76069045e-01 4.90772128e-01 -2.84334868e-01 -4.76053149e-01
1.99130490e-01 4.40125227e-01 3.61480623e-01 -9.79289114e-01
1.28929543e+00 1.28746951e+00 8.34976852e-01 1.11196709e+00
-1.95789146e+00 -5.97748339e-01 -3.07156831e-01 -6.07453644e-01
-1.43835056e+00 -5.22391617e-01 5.47609448e-01 -6.80600762e-01
1.02411926e+00 -8.22036713e-02 7.91950881e-01 1.49945831e+00
3.67212862e-01 3.07084054e-01 1.20239294e+00 -1.82957560e-01
9.69529688e-01 -9.33913440e-02 -3.79560739e-01 4.36276019e-01
4.79992002e-01 1.92784965e-01 -4.53192383e-01 -5.46513855e-01
1.01917958e+00 -2.04705149e-01 1.11970529e-01 -4.89515066e-01
-1.09252965e+00 1.05413926e+00 2.72389740e-01 1.69536509e-02
-8.63661766e-01 4.71156299e-01 1.51661053e-01 -8.80340040e-02
5.37063956e-01 4.98933911e-01 6.19354621e-02 -8.14952403e-02
-1.15125668e+00 8.71727288e-01 1.19456470e+00 8.16711903e-01
8.65527272e-01 3.88335347e-01 -5.76456115e-02 3.32282424e-01
4.68156785e-01 5.94738603e-01 -3.09567750e-02 -1.23569620e+00
4.24161814e-02 -1.94895387e-01 3.56535554e-01 -3.03604901e-01
1.75617591e-01 -1.70213357e-01 -1.32636976e+00 5.45282423e-01
5.29724896e-01 -4.35699463e-01 -1.38550401e+00 1.64103234e+00
3.59163702e-01 2.01224610e-01 3.17063898e-01 6.65574670e-01
7.52697229e-01 8.90198231e-01 1.40674785e-01 -2.62063563e-01
6.73222840e-01 -4.88903284e-01 -5.82864404e-01 2.59176373e-01
3.20442617e-01 -6.70163751e-01 4.98773336e-01 7.36192584e-01
-1.59874833e+00 -4.92527425e-01 -1.00747788e+00 -1.06541894e-01
-4.61243361e-01 -3.07825714e-01 8.11208546e-01 8.20666969e-01
-1.05733359e+00 1.18481815e+00 -1.07653153e+00 8.58214870e-02
8.01421344e-01 3.30344439e-01 -3.16310585e-01 3.40661407e-02
-1.05026054e+00 5.96402287e-01 3.73317748e-01 -1.24400176e-01
-1.55809903e+00 -7.01603591e-01 -8.15792739e-01 -3.67386460e-01
-2.04029903e-01 -1.36595333e+00 1.10303283e+00 -8.44856441e-01
-2.09211659e+00 5.72819114e-01 -4.74778056e-01 -6.51058733e-01
6.42514646e-01 2.87447453e-01 3.25534940e-01 -2.01022043e-03
-1.68768719e-01 1.02070963e+00 6.49081528e-01 -1.58895230e+00
5.47742285e-02 -3.17049399e-02 -5.95664322e-01 -1.35377506e-02
6.59171164e-01 -4.85157281e-01 5.09481281e-02 -4.97351766e-01
-7.37573355e-02 -9.24142420e-01 -7.70700336e-01 -1.94650635e-01
-6.22304082e-01 -6.20413125e-02 1.32156178e-01 -3.15949708e-01
4.39251184e-01 -1.51963377e+00 6.82748914e-01 6.06035173e-01
3.04594487e-01 1.36503771e-01 -4.63086665e-02 6.11824334e-01
-1.65378049e-01 4.45755780e-01 -7.16093123e-01 -7.88224638e-01
1.15929909e-01 1.61979824e-01 -3.09111208e-01 3.15360427e-01
2.01117277e-01 9.18088794e-01 -1.14792919e+00 -7.45351613e-02
1.46412164e-01 1.08325183e+00 -7.62015164e-01 1.28811970e-01
-7.92347789e-01 8.20377827e-01 -1.76118702e-01 5.73820174e-01
8.11713994e-01 -1.65979475e-01 1.07418403e-01 -2.60348618e-02
-1.85158595e-01 3.00126940e-01 -1.06360197e+00 1.98942745e+00
-2.09717508e-02 1.39040068e-01 -2.27967836e-03 -5.00818074e-01
1.08979726e+00 4.61402029e-01 1.85756549e-01 -3.70696709e-02
-3.59644182e-02 4.36163723e-01 -1.19159378e-01 1.04103051e-01
3.17766428e-01 -8.41017783e-01 1.35019034e-01 3.24327737e-01
3.49857628e-01 -9.96573567e-01 2.88102746e-01 2.25976333e-01
8.77276242e-01 6.50076449e-01 1.69355914e-01 -8.64368230e-02
1.10104106e-01 2.07165033e-02 2.91305631e-01 1.05327916e+00
2.65086085e-01 9.45525587e-01 6.17189467e-01 -3.89224172e-01
-1.66505730e+00 -1.59116125e+00 2.73564290e-02 1.44070894e-01
-1.94967344e-01 -2.69955933e-01 -1.18759310e+00 -3.34248871e-01
1.32836848e-01 8.24949086e-01 -6.75369203e-01 -8.04202333e-02
-2.00115994e-01 -1.07978141e+00 5.92113316e-01 1.36457086e-01
1.01833940e-01 -1.28051162e+00 2.01541677e-01 4.94791299e-01
2.84368336e-01 -4.15557951e-01 -5.36552034e-02 4.19264615e-01
-8.84959817e-01 -6.34867787e-01 -7.10696340e-01 -4.81640905e-01
8.01145196e-01 -2.75636911e-01 1.44617391e+00 -3.66282403e-01
-2.85579264e-01 -2.36644000e-01 1.54724658e-01 -5.58609068e-01
-9.19790804e-01 1.63608104e-01 1.07573852e-01 -3.23293388e-01
3.30651522e-01 -8.80042553e-01 -7.56978571e-01 -3.38601202e-01
-1.12875235e+00 4.29502642e-03 4.64516670e-01 7.65974224e-01
1.35927689e+00 1.69818074e-01 5.29460967e-01 -1.16126347e+00
7.71775067e-01 -5.10359466e-01 -5.99229932e-01 -3.23587090e-01
-4.70906198e-01 1.74434856e-01 7.18024254e-01 -2.76458114e-01
-1.04593503e+00 4.30483937e-01 -7.54533648e-01 -2.95396686e-01
-6.13855302e-01 2.60556549e-01 -1.06654853e-01 -8.35374966e-02
8.60268414e-01 3.94764483e-01 1.63540080e-01 -4.78305250e-01
6.05389833e-01 2.47376800e-01 3.98514837e-01 -9.48111832e-01
8.62660110e-01 8.88499081e-01 3.56610179e-01 -7.08387852e-01
-5.35759330e-01 1.42941341e-01 -7.97969639e-01 1.48856878e-01
1.08275390e+00 -8.43513012e-01 -8.46405804e-01 7.49170721e-01
-1.39711571e+00 -6.71371102e-01 -7.78501153e-01 3.55537176e-01
-8.54237139e-01 2.21286550e-01 -9.56323802e-01 -9.48032439e-01
-4.64406788e-01 -1.10538149e+00 1.25064421e+00 1.18630730e-01
-4.38419968e-01 -1.04492545e+00 7.94125557e-01 -1.29113868e-01
4.29979146e-01 7.18665600e-01 8.81308556e-01 -2.44042531e-01
-8.04392993e-01 -9.26179960e-02 1.80323824e-01 3.35798949e-01
1.03431188e-01 3.62030596e-01 -9.47191894e-01 -2.75294125e-01
-8.14693272e-02 -5.87104142e-01 1.10084891e+00 7.01015532e-01
8.12652588e-01 -1.56739891e-01 -3.79038632e-01 7.84506559e-01
1.41748500e+00 2.76562363e-01 9.91710901e-01 -1.56744823e-01
7.08331406e-01 4.83448714e-01 -1.30613074e-01 2.13501096e-01
1.25644887e-02 8.77834186e-02 3.06443721e-01 -2.03451753e-01
-8.81839544e-02 -6.80120826e-01 2.79148281e-01 8.07973504e-01
-3.63218576e-01 -3.98252964e-01 -6.62926614e-01 2.21914142e-01
-1.68999076e+00 -1.27190089e+00 -3.20425145e-02 2.09536123e+00
1.17974210e+00 1.10034421e-01 2.23764464e-01 -1.73805237e-01
4.68436658e-01 -5.68233877e-02 -6.90873563e-01 -2.62987345e-01
-2.31244221e-01 7.74692595e-01 3.43225181e-01 8.38706732e-01
-7.38894284e-01 1.08505762e+00 7.20985270e+00 9.49329853e-01
-9.75487530e-01 9.80011374e-03 8.80248845e-01 -1.77801624e-01
-6.02251589e-01 3.36475596e-02 -9.85302031e-01 6.22589886e-01
1.17799890e+00 1.20790362e-01 6.66198552e-01 5.52644432e-01
2.20022723e-01 6.04657903e-02 -1.08940935e+00 8.47175598e-01
6.51581492e-03 -1.89665115e+00 5.47793806e-01 2.30568901e-01
1.07575214e+00 1.05194785e-01 6.68573019e-04 -3.12045719e-02
9.30449069e-01 -1.56577170e+00 7.25479722e-01 1.19415069e+00
7.33642220e-01 -1.07913172e+00 6.29955947e-01 4.78877485e-01
-6.98730409e-01 9.50027168e-01 -6.89568162e-01 -2.12020129e-02
4.53136206e-01 9.76107597e-01 -7.57175088e-01 2.53789932e-01
3.09125543e-01 6.08176351e-01 1.24447130e-01 7.28947878e-01
-2.64284223e-01 6.73303664e-01 -3.48221451e-01 -7.86882266e-02
1.27449349e-01 -1.01781774e+00 4.38999057e-01 1.19621313e+00
4.70106989e-01 -1.72847584e-01 -1.07412264e-01 1.63604379e+00
-3.82143497e-01 -2.97170818e-01 -9.93488014e-01 -1.23145774e-01
4.37996298e-01 1.07575297e+00 -7.37881482e-01 -5.52082479e-01
1.66863710e-01 9.26570237e-01 2.22039416e-01 6.81005597e-01
-5.32230616e-01 -5.40694714e-01 5.93867838e-01 1.70117319e-01
5.40829003e-01 -3.14105690e-01 -1.02732144e-01 -9.37334716e-01
-5.27039051e-01 -8.37810695e-01 -3.57422143e-01 -8.85952950e-01
-1.70059323e+00 6.89806998e-01 -4.06327188e-01 -5.34156799e-01
-3.77758861e-01 -6.18758321e-01 -5.53267479e-01 1.34536982e+00
-1.36983144e+00 -8.60986173e-01 2.46374495e-02 1.42709792e-01
1.44871429e-03 7.93144032e-02 1.08633089e+00 -8.37378129e-02
-2.68530518e-01 -2.77791284e-02 5.32299817e-01 -1.96907461e-01
5.30252516e-01 -1.70653772e+00 1.10430455e+00 3.87033045e-01
-1.85438618e-01 8.87279809e-01 7.65217662e-01 -1.08979642e+00
-1.27316916e+00 -1.32096267e+00 7.00822115e-01 -6.55169487e-01
3.38703752e-01 -8.23385060e-01 -9.55823541e-01 4.03047383e-01
3.23765188e-01 -2.29962215e-01 7.71787584e-01 -3.76130760e-01
-7.26499334e-02 3.48326921e-01 -1.47935760e+00 6.66144371e-01
9.79153514e-01 -5.87889075e-01 -1.06832951e-01 6.31123364e-01
5.86203694e-01 -5.18977165e-01 -1.16594446e+00 4.15031686e-02
5.04552186e-01 -9.60785389e-01 1.25652218e+00 -6.32365048e-01
4.46791321e-01 -5.38209438e-01 -3.51408981e-02 -1.55621505e+00
-2.70346463e-01 -1.15973771e+00 -2.42717832e-01 9.18667376e-01
5.28830111e-01 -5.85758924e-01 1.17846894e+00 2.31472462e-01
5.53499497e-02 -6.86483800e-01 -9.09968495e-01 -5.12950242e-01
7.72185683e-01 -1.11491524e-01 7.68689096e-01 5.20200074e-01
-6.17614150e-01 3.94971937e-01 -2.97297567e-01 -1.90421149e-01
1.02352560e+00 -2.17274308e-01 8.55513394e-01 -1.24574602e+00
-3.76790822e-01 -1.78694576e-01 4.20315601e-02 -1.02757990e+00
7.56677836e-02 -9.59398031e-01 3.78045589e-01 -1.59892285e+00
2.47956470e-01 -2.32503146e-01 1.21026598e-01 -1.36456564e-01
1.31069243e-01 4.26302791e-01 -2.09645361e-01 1.91472515e-01
-4.36527669e-01 7.89546728e-01 1.38084710e+00 -2.69127507e-02
-2.30699837e-01 -2.03485712e-01 -5.92172623e-01 6.85325146e-01
6.10011101e-01 -7.36763000e-01 -3.12879562e-01 5.75396372e-03
6.57556295e-01 2.33765487e-02 3.81506294e-01 -9.43554282e-01
1.06095530e-01 -1.59402058e-01 7.91959226e-01 -6.91272974e-01
5.48878133e-01 -1.53102413e-01 6.65126145e-01 4.26000178e-01
-1.43074751e-01 -1.87677503e-01 3.38401608e-02 6.81087315e-01
2.92168222e-02 -4.59145486e-01 9.70844328e-01 -6.59792006e-01
3.97909433e-01 5.51432312e-01 -8.12314212e-01 -7.57156536e-02
6.48178041e-01 -1.61478549e-01 -1.85974389e-01 -3.57770473e-01
-8.88293922e-01 -3.95488888e-01 1.21812570e+00 -5.23126185e-01
5.26674092e-01 -1.33251059e+00 -6.25268817e-01 1.63708851e-01
-5.18489003e-01 6.77630424e-01 4.08868641e-02 2.93120950e-01
-9.16985035e-01 2.67487556e-01 1.17393300e-01 -5.13631284e-01
-4.86732811e-01 3.19518805e-01 6.83564484e-01 -3.21287185e-01
-1.67558029e-01 9.39922869e-01 1.13649949e-01 -7.62154460e-01
-1.41471192e-01 -2.29752094e-01 2.98015714e-01 -1.53226838e-01
4.10796821e-01 1.08802654e-01 -8.65169801e-03 -3.74855757e-01
1.87110789e-02 5.39780557e-01 -1.12713546e-01 -2.91340858e-01
1.43642735e+00 4.26869929e-01 -2.74985492e-01 4.45590734e-01
8.86464119e-01 1.91101208e-02 -1.66280651e+00 1.57063335e-01
-4.76476669e-01 -1.24038085e-01 -1.34269238e-01 -6.55472279e-01
-5.90852022e-01 9.48313534e-01 3.36610615e-01 1.30743340e-01
2.26142019e-01 -9.31997597e-02 7.84217060e-01 5.43306470e-01
1.92111239e-01 -8.93175304e-01 3.19107510e-02 4.95043397e-01
6.27142966e-01 -9.11748648e-01 8.75182971e-02 -2.58796126e-01
-4.00908887e-01 9.60911155e-01 1.13398053e-01 -5.42921841e-01
4.96933073e-01 4.50889826e-01 -6.61321208e-02 -4.02664602e-01
-7.75590539e-01 2.71129515e-02 -7.67137706e-02 1.13366914e+00
6.03681564e-01 -7.93740526e-02 3.88466924e-01 2.92830288e-01
-2.80601978e-01 2.65911490e-01 4.71410573e-01 7.76906133e-01
-4.33176160e-01 -1.22549319e+00 -2.82684177e-01 4.09415156e-01
-2.28392720e-01 -4.58779305e-01 -5.71119130e-01 4.71618414e-01
4.21974421e-01 5.67624211e-01 5.56612946e-02 7.54110515e-02
2.58274376e-01 2.91751057e-01 7.86307693e-01 -8.41425896e-01
-3.59229773e-01 1.08844176e-01 -1.09242737e-01 -1.95839345e-01
-5.01556456e-01 -8.07088315e-01 -1.09470177e+00 -6.12420321e-01
-1.44023821e-01 5.18977761e-01 7.04915524e-01 7.18167722e-01
4.90454286e-01 5.72252154e-01 1.99713349e-01 -1.59697247e+00
-4.57051247e-01 -8.10908616e-01 -6.42623603e-01 3.61701667e-01
3.19530308e-01 -3.83962184e-01 -5.14458418e-01 1.19801074e-01] | [5.052623271942139, 5.486171722412109] |
556fe568-4993-4340-8097-d4cca9a0b782 | deepxml-scalable-accurate-deep-extreme | null | null | https://openreview.net/forum?id=SJlWyerFPS | https://openreview.net/pdf?id=SJlWyerFPS | DeepXML: Scalable & Accurate Deep Extreme Classification for Matching User Queries to Advertiser Bid Phrases | The objective in deep extreme multi-label learning is to jointly learn feature representations and classifiers to automatically tag data points with the most relevant subset of labels from an extremely large label set. Unfortunately, state-of-the-art deep extreme classifiers are either not scalable or inaccurate for short text documents. This paper develops the DeepXML algorithm which addresses both limitations by introducing a novel architecture that splits training of head and tail labels. DeepXML increases accuracy by (a) learning word embeddings on head labels and transferring them through a novel residual connection to data impoverished tail labels; (b) increasing the amount of negative training data available by extending state-of-the-art negative sub-sampling techniques; and (c) re-ranking the set of predicted labels to eliminate the hardest negatives for the original classifier. All of these contributions are implemented efficiently by extending the highly scalable Slice algorithm for pretrained embeddings to learn the proposed DeepXML architecture. As a result, DeepXML could efficiently scale to problems involving millions of labels that were beyond the pale of state-of-the-art deep extreme classifiers as it could be more than 10x faster at training than XML-CNN and AttentionXML. At the same time, DeepXML was also empirically determined to be up to 19% more accurate than leading techniques for matching search engine queries to advertiser bid phrases. | ['Manik Varma', 'Sumeet Agarwal', 'Himanshu Jain', 'Kushal Dave', 'Deepak Saini', 'Anshul Mittal', 'Kunal Dahiya'] | 2019-09-25 | null | null | null | null | ['learning-word-embeddings'] | ['methodology'] | [ 1.85074918e-02 1.84313253e-01 -7.26131201e-01 -6.95915878e-01
-1.25014126e+00 -8.81915390e-01 6.46251440e-01 3.38344127e-01
-6.65762603e-01 4.84124213e-01 -4.04877663e-02 -3.85402799e-01
-1.29069924e-01 -7.06055641e-01 -5.16238213e-01 -4.36485380e-01
5.63571639e-02 1.18999350e+00 1.38642639e-02 -6.04243129e-02
1.70464709e-01 3.61830890e-01 -1.72246063e+00 6.22451842e-01
1.82931185e-01 1.62747252e+00 -3.97522926e-01 4.25707370e-01
-6.65304959e-01 6.74489141e-01 -3.75066429e-01 -6.85947180e-01
4.31016803e-01 2.23196670e-01 -8.11444998e-01 -3.74364793e-01
9.87139702e-01 -4.08514678e-01 1.14318714e-01 8.72571588e-01
6.33980095e-01 -1.03274547e-01 6.92973435e-01 -1.29451776e+00
-9.31724966e-01 6.32419288e-01 -8.31713736e-01 1.42538086e-01
1.20004140e-01 -3.18664283e-01 1.59597933e+00 -9.54935908e-01
5.26105762e-01 1.08848619e+00 9.51416969e-01 8.74282539e-01
-1.05649126e+00 -9.84886169e-01 1.35295406e-01 -4.16309154e-03
-1.29335070e+00 -1.64314032e-01 4.87765223e-01 -1.48040086e-01
1.20655227e+00 5.04158795e-01 2.88269669e-01 1.32017934e+00
1.07040340e-02 7.96730936e-01 1.03881419e+00 -4.30969238e-01
1.33800447e-01 4.50508326e-01 4.73923594e-01 7.42515922e-01
6.79154024e-02 2.38060594e-01 -2.18829647e-01 -7.32742310e-01
-3.65349427e-02 1.80347592e-01 4.21953797e-01 -3.46856624e-01
-6.46117508e-01 1.34006214e+00 3.38344038e-01 4.08656418e-01
-1.32530183e-01 5.52816689e-01 7.66205728e-01 4.28752244e-01
8.58513534e-01 5.62698781e-01 -8.87094319e-01 3.75296269e-03
-1.18056142e+00 2.08966970e-01 7.69929051e-01 9.46008980e-01
7.90376484e-01 -3.44925851e-01 -2.03704894e-01 9.85744059e-01
3.82423669e-01 8.10943991e-02 8.76709640e-01 -9.28656459e-01
4.02043164e-01 6.42698467e-01 -3.28823961e-02 -5.33380032e-01
-5.77822685e-01 -8.67963850e-01 -2.41545647e-01 2.63871521e-01
2.25810722e-01 -4.00310569e-02 -1.06255400e+00 1.69472730e+00
3.05730164e-01 -1.62922621e-01 -2.46398166e-01 5.90251446e-01
6.31790400e-01 6.16624236e-01 4.17899281e-01 2.67093122e-01
1.51593149e+00 -1.09775996e+00 -2.82775939e-01 -5.13170958e-01
1.32543230e+00 -7.20800877e-01 1.11059749e+00 2.25207254e-01
-8.51982415e-01 -4.53339607e-01 -9.38508391e-01 -3.50017935e-01
-9.42969620e-01 -3.96051258e-02 1.09508705e+00 7.30233312e-01
-1.11147988e+00 4.07140523e-01 -2.29597092e-01 -1.17343441e-01
5.87242365e-01 8.43485713e-01 -1.57781735e-01 -1.25932395e-01
-1.50591373e+00 9.47987378e-01 2.45442152e-01 -3.14662993e-01
-5.71316719e-01 -1.04369462e+00 -7.53039002e-01 3.57058316e-01
2.49352649e-01 -4.22380894e-01 1.54777288e+00 -1.09704351e+00
-9.81886089e-01 1.24840605e+00 9.69225839e-02 -4.04224306e-01
3.38573486e-01 -1.50353193e-01 -5.38393378e-01 -9.25728753e-02
2.54138589e-01 1.14144611e+00 6.78917348e-01 -1.06888735e+00
-1.03973079e+00 -4.31391776e-01 -1.61802053e-01 1.40285045e-02
-8.59419346e-01 -2.13987418e-02 -1.96988150e-01 -4.04619813e-01
-1.04516141e-01 -8.55151474e-01 -7.89305046e-02 8.54018107e-02
-9.39745083e-02 -9.78881419e-01 9.30533707e-01 -3.34027618e-01
1.20210707e+00 -2.15745282e+00 -3.20546240e-01 2.03932136e-01
5.28483868e-01 3.49674761e-01 -2.98990101e-01 4.68069017e-01
-1.66678175e-01 1.66289195e-01 4.03544128e-01 -6.58920825e-01
5.77996194e-01 1.72671124e-01 -6.00374520e-01 4.15445507e-01
-2.34154835e-02 1.14000487e+00 -7.54431486e-01 -5.78244209e-01
-4.88026552e-02 3.12987387e-01 -5.44053555e-01 -9.39387530e-02
-6.04106367e-01 -3.54207844e-01 -3.43756914e-01 8.42985988e-01
5.00686288e-01 -3.73367041e-01 7.12389424e-02 1.89598612e-02
8.07055756e-02 2.49015182e-01 -8.50443304e-01 1.52065194e+00
-8.56098652e-01 4.94688809e-01 -1.59913957e-01 -9.86405969e-01
7.86665261e-01 2.97271580e-01 7.30022728e-01 -9.11747515e-01
2.82007873e-01 5.78411222e-01 -5.40134907e-01 -3.64189535e-01
5.39048195e-01 -3.72401297e-01 -5.13420641e-01 7.04301238e-01
2.87086815e-01 4.03091818e-01 2.48019025e-02 6.75314665e-02
9.92449284e-01 -1.60195693e-01 -7.84754101e-03 -1.02283776e-01
2.44460300e-01 -4.96081561e-02 3.94884109e-01 7.96983421e-01
-8.54578540e-02 1.90887824e-01 3.37455153e-01 -7.93635666e-01
-1.04038179e+00 -9.06011283e-01 -3.62031609e-01 2.08908510e+00
-2.89433777e-01 -4.20866430e-01 -4.54527050e-01 -1.18626225e+00
5.06430089e-01 8.95361602e-01 -8.35060418e-01 -2.56755948e-01
-5.14148593e-01 -6.08640313e-01 7.64478505e-01 7.10871696e-01
3.56350467e-03 -8.00643921e-01 -1.79994643e-01 2.01204762e-01
2.40051001e-01 -7.40119696e-01 -3.73504639e-01 9.09535229e-01
-5.00691533e-01 -8.10409188e-01 -6.16873443e-01 -1.03272188e+00
3.71362537e-01 -2.10561827e-01 1.24561656e+00 1.61743369e-02
-3.86615008e-01 2.61544771e-02 -1.63346842e-01 -3.46357256e-01
-2.81846106e-01 4.99613732e-01 7.75401248e-03 -1.92133904e-01
1.15986443e+00 -3.87640119e-01 -4.96057034e-01 4.18729931e-02
-7.19758689e-01 -4.91008967e-01 5.44177353e-01 8.37040722e-01
4.81518775e-01 -2.96390146e-01 7.89806664e-01 -1.29280913e+00
5.31848490e-01 -7.93639183e-01 -4.88693953e-01 3.28494787e-01
-1.16746676e+00 2.31346846e-01 7.30958462e-01 -6.73070312e-01
-5.68583965e-01 -7.61625767e-02 -5.36757112e-01 -1.57650918e-01
8.12636614e-02 1.03529878e-01 3.68629009e-01 -8.86546373e-02
7.00080752e-01 -2.59874195e-01 -2.20552742e-01 -6.20565236e-01
8.47215176e-01 1.16966438e+00 2.03017190e-01 -4.59406942e-01
3.58163893e-01 3.90229315e-01 -7.00292289e-02 -9.30986255e-02
-1.46780634e+00 -8.06302428e-01 -3.02033067e-01 9.05795470e-02
7.06111372e-01 -6.34435058e-01 -7.25766659e-01 4.68442142e-02
-6.07861459e-01 -1.22785620e-01 -6.37940407e-01 2.28381693e-01
-4.91700023e-01 2.31701329e-01 -9.69681084e-01 -4.44918156e-01
-6.03312969e-01 -8.68526936e-01 1.43558955e+00 -1.02251016e-01
-5.64984441e-01 -1.10699248e+00 2.48423308e-01 2.91707516e-01
5.64967275e-01 -1.56006083e-01 1.41607535e+00 -1.31687582e+00
-3.04129208e-03 -8.14802647e-01 -3.47063154e-01 3.13548923e-01
-2.71798372e-01 -3.40421259e-01 -1.19348967e+00 -4.89324242e-01
-1.41975299e-01 -9.31834757e-01 1.12970316e+00 1.44205302e-01
1.20023668e+00 -1.23721994e-01 -7.41948783e-01 6.16514564e-01
1.43217349e+00 -9.51836184e-02 1.82560548e-01 5.33142209e-01
3.80089462e-01 4.85921711e-01 5.40300429e-01 2.59458154e-01
2.37228066e-01 8.78253400e-01 5.91455758e-01 -1.87748447e-01
-2.03488246e-01 -2.76150227e-01 9.25197974e-02 4.90660638e-01
8.23048174e-01 -3.56732875e-01 -8.39026511e-01 4.86386150e-01
-1.51557732e+00 -7.74492204e-01 5.74855320e-02 1.77270222e+00
8.35589111e-01 2.59569049e-01 1.68710854e-02 6.16640784e-02
5.25163472e-01 2.10961118e-01 -5.82681417e-01 -8.22741508e-01
2.74648190e-01 4.82855439e-01 7.14345157e-01 3.20449948e-01
-1.31582701e+00 9.39810097e-01 6.18705797e+00 1.07841456e+00
-1.14535844e+00 5.98997295e-01 5.62351882e-01 -5.93795478e-01
-4.67837691e-01 -1.13723345e-01 -1.49151456e+00 4.73135114e-01
1.36595500e+00 3.13666373e-01 1.16677068e-01 1.24911916e+00
-7.03735709e-01 4.16745424e-01 -1.34063029e+00 8.58234763e-01
2.31160522e-01 -1.30014718e+00 4.48014140e-02 4.43054944e-01
5.83069503e-01 4.70584333e-01 5.02902687e-01 8.80324423e-01
5.44407845e-01 -9.45032060e-01 6.87606573e-01 1.75085634e-01
1.08998680e+00 -6.64068580e-01 7.56003737e-01 3.61237764e-01
-7.25789011e-01 -8.13514292e-01 -5.65940738e-01 2.16302902e-01
5.56940213e-02 7.56190419e-01 -6.36238992e-01 1.32247312e-02
5.17675281e-01 2.82982975e-01 -6.70291960e-01 5.94512224e-01
1.88371003e-01 3.41547787e-01 -5.61981320e-01 -3.04562151e-01
8.47301126e-01 4.18581158e-01 -2.72385418e-01 1.30783033e+00
3.61465842e-01 -5.21509528e-01 2.37422720e-01 4.66246516e-01
-4.82441097e-01 1.12089060e-01 -4.79732066e-01 -8.85611549e-02
5.08057415e-01 1.36951602e+00 -4.14782882e-01 -4.57829386e-01
-7.02289581e-01 8.20051253e-01 7.84947217e-01 3.65489908e-02
-8.48262370e-01 -1.63964078e-01 6.94545746e-01 1.81514338e-01
3.68439972e-01 2.54253119e-01 -2.09302992e-01 -7.00492442e-01
-1.16362207e-01 -5.23628116e-01 8.07326913e-01 -4.17321950e-01
-1.53649640e+00 5.82342446e-01 -3.48706603e-01 -9.15575504e-01
-6.02517962e-01 -8.97008061e-01 2.49879435e-02 7.92958200e-01
-1.65166271e+00 -1.29205036e+00 5.31069152e-02 1.04636237e-01
4.34214860e-01 -3.40447932e-01 1.36521387e+00 7.73524463e-01
-1.37503758e-01 9.87177610e-01 4.55988377e-01 -5.96395694e-02
8.56603086e-01 -1.31406832e+00 3.92539829e-01 -1.03252739e-01
3.43625039e-01 4.51111048e-01 9.74851921e-02 -3.41397434e-01
-9.12759364e-01 -1.20359170e+00 1.29791260e+00 -5.63624680e-01
7.55308747e-01 -7.24236488e-01 -4.32792336e-01 8.91196907e-01
-1.79736745e-02 3.58260989e-01 1.14821303e+00 4.12528187e-01
-8.86853218e-01 -3.00677776e-01 -1.17300439e+00 2.54415512e-01
7.87013650e-01 -7.04791248e-01 -3.52365017e-01 8.19709480e-01
5.53502083e-01 -5.69657087e-02 -7.91472971e-01 2.99630463e-01
7.21506774e-01 -6.32516026e-01 9.34835911e-01 -8.90097439e-01
3.64402652e-01 3.15029532e-01 -3.96272242e-01 -9.56950605e-01
-4.63677764e-01 -7.20537752e-02 -4.65560853e-01 1.32792926e+00
6.49901450e-01 -6.43758059e-01 1.09213805e+00 2.68331021e-01
-3.20706181e-02 -1.30619824e+00 -1.15905595e+00 -6.36867762e-01
4.31061715e-01 -5.14475644e-01 6.46643698e-01 1.04362643e+00
-1.98598787e-01 3.51197094e-01 -1.44633889e-01 -2.41083413e-01
5.52165508e-01 3.05822879e-01 1.61152795e-01 -1.46515036e+00
-3.64240348e-01 -4.51302230e-01 -4.64876384e-01 -1.10932243e+00
6.36811316e-01 -1.48973060e+00 -3.78565639e-01 -1.32156193e+00
2.76695788e-01 -9.27620709e-01 -5.98858893e-01 6.72495902e-01
2.87939072e-01 7.35153496e-01 -1.44293383e-01 1.69517204e-01
-6.80211782e-01 1.22866929e-01 5.26637673e-01 -3.32095861e-01
4.04701948e-01 1.59178600e-01 -1.03922272e+00 5.45410693e-01
7.52207279e-01 -9.23124909e-01 -4.37076092e-01 -5.44168174e-01
7.26829648e-01 -3.43093544e-01 6.43411949e-02 -7.47116148e-01
8.96352082e-02 3.14447939e-01 5.29947162e-01 -6.13507748e-01
5.12645483e-01 -1.02471101e+00 -3.30055743e-01 2.91726947e-01
-9.64978337e-01 7.71693587e-02 -7.85366222e-02 4.94694084e-01
2.80446000e-02 -7.47916222e-01 6.55006289e-01 -2.66329288e-01
-8.05688143e-01 2.46353313e-01 -9.73829553e-02 2.77501523e-01
1.22132623e+00 -3.23451236e-02 -4.46176112e-01 -1.78457260e-01
-6.23318613e-01 2.02069357e-01 2.36930445e-01 5.31006277e-01
1.51002124e-01 -1.41999662e+00 -3.52192074e-01 3.44867796e-01
2.48110965e-01 -4.08830523e-01 -6.98396116e-02 2.44469345e-01
-2.16868073e-01 8.10996592e-01 2.30011582e-01 -3.57668698e-01
-1.16300929e+00 8.21592629e-01 2.42759511e-01 -6.50363982e-01
-3.82498592e-01 1.27175140e+00 -1.50929689e-01 -8.23007941e-01
6.89903557e-01 -1.12502731e-01 1.18656033e-04 4.37558323e-01
5.32947540e-01 1.54983059e-01 4.47738975e-01 -3.81127775e-01
-3.73704135e-01 5.08981526e-01 -5.68193495e-01 -1.94641910e-02
1.31503975e+00 1.38512641e-01 3.05025820e-02 3.42649549e-01
1.97494900e+00 -7.35132992e-02 -9.33366835e-01 -1.54285789e-01
2.99105674e-01 -3.02854069e-02 2.65326113e-01 -1.11886239e+00
-1.02039552e+00 1.05308235e+00 8.99796963e-01 1.87576458e-01
7.39683986e-01 4.13832039e-01 1.26469243e+00 3.54993045e-01
2.77871609e-01 -1.34507143e+00 8.35853368e-02 3.05960357e-01
3.34595650e-01 -1.08670521e+00 -6.56798780e-02 -1.10114612e-01
-1.47373438e-01 1.10712540e+00 4.79549289e-01 1.26523152e-01
7.20104814e-01 5.11859775e-01 2.81385183e-01 -5.95881224e-01
-9.71805573e-01 -5.04557183e-03 8.61542970e-02 3.95774513e-01
5.03639638e-01 -3.70768942e-02 -3.40123653e-01 5.00193357e-01
-6.58354536e-02 -2.95373444e-02 -1.82564139e-01 8.67594421e-01
-5.12545645e-01 -1.33549738e+00 9.04919207e-02 9.88074839e-01
-8.36118996e-01 -3.39111149e-01 -3.45509052e-01 6.66344643e-01
3.19869667e-01 6.96983576e-01 3.01381350e-01 -5.01935244e-01
1.54939413e-01 5.55819511e-01 2.07314089e-01 -7.58417428e-01
-8.60460460e-01 -2.58417279e-01 1.28672406e-01 -3.90151590e-01
1.23684093e-01 -4.34195817e-01 -1.23066080e+00 -2.56833702e-01
-5.83865702e-01 2.63208598e-01 8.63580525e-01 7.38701403e-01
4.98239666e-01 2.51950413e-01 5.79016566e-01 -6.61920071e-01
-1.38686109e+00 -1.02415442e+00 -6.40882850e-01 4.28790659e-01
8.35298896e-02 -7.40862072e-01 -4.17140812e-01 -6.74534976e-01] | [9.605245590209961, 4.434901714324951] |
afdce437-093f-4624-8c1c-8777938dce24 | detect-anticipate-and-generate-semi | 1809.07075 | null | http://arxiv.org/abs/1809.07075v1 | http://arxiv.org/pdf/1809.07075v1.pdf | Detect, anticipate and generate: Semi-supervised recurrent latent variable models for human activity modeling | Successful Human-Robot collaboration requires a predictive model of human
behavior. The robot needs to be able to recognize current goals and actions and
to predict future activities in a given context. However, the spatio-temporal
sequence of human actions is difficult to model since latent factors such as
intention, task, knowledge, intuition and preference determine the action
choices of each individual. In this work we introduce semi-supervised
variational recurrent neural networks which are able to a) model temporal
distributions over latent factors and the observable feature space, b)
incorporate discrete labels such as activity type when available, and c)
generate possible future action sequences on both feature and label level. We
evaluate our model on the Cornell Activity Dataset CAD-120 dataset. Our model
outperforms state-of-the-art approaches in both activity and affordance
detection and anticipation. Additionally, we show how samples of possible
future action sequences are in line with past observations. | ['Judith Bütepage', 'Danica Kragic'] | 2018-09-19 | null | null | null | null | ['affordance-detection'] | ['computer-vision'] | [ 4.07402575e-01 2.25884631e-01 -4.93933380e-01 -4.88132119e-01
-4.14463550e-01 -2.75571764e-01 9.31443989e-01 -5.73489554e-02
-4.91913915e-01 8.34889889e-01 7.26824462e-01 1.30009830e-01
-1.27610546e-02 -3.44929367e-01 -6.30431771e-01 -3.77323210e-01
-4.95818883e-01 8.70761335e-01 7.39741325e-02 1.66959047e-01
7.92681873e-02 4.30310726e-01 -1.63987756e+00 2.84614593e-01
4.08202708e-01 6.76521122e-01 7.13585615e-01 9.59806085e-01
5.07469535e-01 1.43837297e+00 -1.41592339e-01 3.57503831e-01
2.53033619e-02 -7.07203507e-01 -8.19836199e-01 6.47840798e-01
-3.99308830e-01 -5.78184724e-01 -5.16591370e-01 6.16737425e-01
-3.97310816e-02 4.96412843e-01 1.08801758e+00 -1.56770289e+00
-3.30072284e-01 7.13384509e-01 -1.13037579e-01 -2.57193923e-01
6.57389522e-01 5.38500011e-01 1.04031658e+00 -4.81338173e-01
1.03616452e+00 1.22618330e+00 2.56248981e-01 6.47495687e-01
-1.36648130e+00 -1.47873074e-01 5.38677633e-01 5.94078004e-01
-9.29731548e-01 -3.93939674e-01 8.76285732e-01 -6.71607494e-01
1.26962686e+00 -2.42093235e-01 1.08202362e+00 1.75214756e+00
3.47647756e-01 1.31781876e+00 8.41055810e-01 -4.12677854e-01
4.95229930e-01 -3.25961828e-01 -1.23609146e-02 7.69536018e-01
-4.13467824e-01 2.04915240e-01 -8.20052326e-01 -2.08923817e-01
8.08556616e-01 3.85592699e-01 -3.92013527e-02 -8.41486514e-01
-1.84937704e+00 5.82047403e-01 3.00264657e-01 2.71262228e-01
-1.13003933e+00 5.76387644e-01 2.07008034e-01 -1.91098768e-02
2.02546213e-02 3.81492853e-01 -5.28170347e-01 -5.96820116e-01
-6.17694438e-01 5.08569121e-01 6.96907818e-01 9.73050952e-01
7.98503101e-01 -2.01951489e-01 -5.25313377e-01 4.25470322e-01
4.11501944e-01 2.74527490e-01 5.51113248e-01 -1.70473564e+00
3.70400339e-01 4.27082330e-01 7.66873538e-01 -5.97767830e-01
-5.96906304e-01 3.56732488e-01 -3.85140270e-01 1.70359924e-01
3.88879657e-01 9.28469840e-03 -1.18258595e+00 1.88793516e+00
7.05650076e-02 1.61784202e-01 1.79100409e-01 7.62076557e-01
4.67233313e-03 5.83535671e-01 3.41238856e-01 -4.52689379e-01
9.87712443e-01 -1.04150593e+00 -9.34917986e-01 -7.27940381e-01
7.65432894e-01 -1.91449448e-01 8.18223774e-01 2.50250369e-01
-9.26260173e-01 -6.72747195e-01 -7.93834686e-01 -6.12731883e-03
6.85056224e-02 4.31271881e-01 8.86397898e-01 -1.24329783e-01
-8.94092977e-01 7.06428230e-01 -1.72199655e+00 -4.68909353e-01
2.34932944e-01 2.57548600e-01 -3.97510618e-01 7.66878873e-02
-8.76133382e-01 1.07520986e+00 6.12696350e-01 1.42118737e-01
-1.48237717e+00 1.49950996e-01 -1.09088719e+00 -9.24405977e-02
6.03136480e-01 -5.32712996e-01 1.54535031e+00 -7.78445542e-01
-1.64896667e+00 5.06510854e-01 -4.81944114e-01 -8.44269276e-01
6.26322269e-01 -2.71045178e-01 1.93104949e-02 -3.85877825e-02
2.39354759e-01 1.03551030e+00 7.45991468e-01 -8.84887934e-01
-7.76664555e-01 -3.28735709e-01 -8.87370110e-02 6.62822008e-01
2.77731150e-01 -3.14087152e-01 -8.88698325e-02 -1.80631280e-01
1.19509414e-01 -1.46893406e+00 -6.01272166e-01 1.83266110e-03
-3.87198776e-01 -4.66894150e-01 5.22237062e-01 -5.20267785e-01
6.79114401e-01 -2.03115225e+00 5.17901659e-01 -1.76910281e-01
3.29032028e-03 -4.39721227e-01 1.09608971e-01 4.38001275e-01
1.81663409e-01 -3.65021586e-01 -5.56544513e-02 -6.51323020e-01
3.51938725e-01 5.92017531e-01 -2.27464721e-01 7.13226497e-01
-4.22827750e-02 9.33035433e-01 -9.77993131e-01 -5.78601420e-01
6.57816291e-01 3.15682471e-01 -5.26358306e-01 4.99254316e-01
-7.38410175e-01 7.28998125e-01 -7.23917782e-01 5.07240832e-01
-1.52056515e-01 -4.27213907e-01 6.06989563e-01 3.92923653e-01
-1.48680657e-01 5.77080965e-01 -1.05586517e+00 1.99535370e+00
-3.89595985e-01 6.45776033e-01 -2.64641523e-01 -8.15141380e-01
5.77420354e-01 5.76830089e-01 6.17993891e-01 -2.71267802e-01
2.65999101e-02 1.08723618e-01 -4.15752496e-04 -5.82394600e-01
5.35853326e-01 6.26049340e-02 -4.57616836e-01 7.36899137e-01
1.91655442e-01 2.03532353e-01 2.49928921e-01 1.13437839e-01
1.22640753e+00 9.40505981e-01 6.60804689e-01 2.33284086e-01
-1.76330917e-02 1.80541739e-01 7.02646136e-01 9.53011572e-01
-5.78976393e-01 3.42321813e-01 7.28043258e-01 -6.17250860e-01
-9.41185057e-01 -8.73499632e-01 5.05869210e-01 1.23759341e+00
3.29659879e-02 -1.32949784e-01 -3.25343072e-01 -4.62435693e-01
-3.94736022e-01 1.20801926e+00 -1.05819786e+00 -6.88025877e-02
-7.23685503e-01 -1.11301050e-01 -2.54747365e-02 8.16166461e-01
2.51422524e-01 -1.87265730e+00 -1.49245167e+00 4.74394888e-01
-4.25416708e-01 -1.15436244e+00 -3.36292297e-01 5.20290434e-01
-6.93287432e-01 -9.74137366e-01 -3.54650110e-01 -5.52345812e-01
5.27870119e-01 -1.90089911e-01 9.47708309e-01 -3.48784894e-01
-9.96741205e-02 5.36215723e-01 -3.82893801e-01 -6.99093416e-02
-5.55720747e-01 -1.26435876e-01 2.26790592e-01 -3.85609269e-02
3.23933154e-01 -6.20413125e-01 -5.33318758e-01 2.23906696e-01
-3.19873542e-01 4.92765427e-01 5.13917148e-01 8.83304298e-01
6.01221621e-01 -2.24704385e-01 -1.50113940e-01 -5.26731372e-01
3.53335261e-01 -4.58752394e-01 -2.84841806e-01 4.13418829e-01
-2.68611908e-01 2.47624263e-01 1.24776997e-01 -6.92935705e-01
-1.39643288e+00 9.01208103e-01 3.57246429e-01 -3.76129240e-01
-7.00417578e-01 4.04271036e-01 -1.22123510e-01 9.23720479e-01
5.90042293e-01 2.42880896e-01 -2.03941554e-01 -2.12955162e-01
4.12272125e-01 3.77563149e-01 4.87810344e-01 -5.56889176e-01
6.12271726e-02 4.69363987e-01 1.04636326e-02 -6.19049966e-01
-5.91979563e-01 -2.58703798e-01 -1.23421311e+00 -4.16281849e-01
1.10353267e+00 -1.01206756e+00 -8.64788890e-01 4.15428877e-01
-1.38618731e+00 -9.55311120e-01 -3.52644414e-01 7.56794989e-01
-1.38003886e+00 4.44650538e-02 -5.85197628e-01 -1.38080430e+00
4.50821161e-01 -1.41961074e+00 1.26240599e+00 -5.13343774e-02
-1.01314557e+00 -9.01639819e-01 -2.40617692e-02 1.37773812e-01
-2.18818828e-01 4.47588652e-01 4.45175260e-01 -5.24544537e-01
-8.39005947e-01 -1.90707311e-01 4.46090400e-01 -2.24530578e-01
2.01388523e-01 -2.32493162e-01 -5.20646811e-01 6.44276675e-04
-4.27736342e-02 -6.75253212e-01 7.30054319e-01 7.42268384e-01
6.05491996e-01 -2.74802029e-01 -7.57025957e-01 8.30635130e-02
6.91787779e-01 4.00807947e-01 4.68110055e-01 1.68631628e-01
5.19516885e-01 9.18413818e-01 1.01641059e+00 6.06585741e-01
6.00338519e-01 8.68905544e-01 3.81958246e-01 5.55439770e-01
3.34953636e-01 -7.13470936e-01 6.34194136e-01 4.97264713e-02
-1.44102708e-01 -3.06892723e-01 -8.93945396e-01 7.19069242e-01
-2.50537348e+00 -1.20534348e+00 6.12068288e-02 1.67054653e+00
4.42100435e-01 2.06895813e-01 2.13502899e-01 -9.59599465e-02
4.87807065e-01 2.37146497e-01 -9.76251006e-01 -1.15820818e-01
2.78584182e-01 -5.24531901e-01 3.39787036e-01 5.61681151e-01
-1.13707578e+00 1.03002346e+00 6.60197496e+00 2.12415546e-01
-4.93784010e-01 1.24383271e-01 4.88418669e-01 -1.68018341e-01
-1.63447969e-02 1.01542719e-01 -6.06689274e-01 2.22631529e-01
8.15234721e-01 1.84808865e-01 5.94635606e-01 9.95553911e-01
3.56863648e-01 -5.03724217e-01 -1.61457431e+00 7.79263616e-01
-1.23646282e-01 -9.69475925e-01 -3.98008198e-01 2.77614862e-01
5.59240222e-01 9.03022885e-02 -1.74214363e-01 5.16860366e-01
9.64422166e-01 -1.02423382e+00 1.08806384e+00 9.12533104e-01
2.49718636e-01 -4.12239790e-01 2.39682272e-01 9.88151848e-01
-9.80748415e-01 -3.45292062e-01 9.01634842e-02 -5.80218792e-01
5.96631646e-01 -2.39400744e-01 -1.08045650e+00 4.52829525e-03
4.63848770e-01 1.11697316e+00 -2.69349635e-01 2.11292118e-01
-6.65416181e-01 2.94884503e-01 -3.30907583e-01 -3.49582762e-01
3.55709255e-01 -9.17995200e-02 4.71522093e-01 6.32476985e-01
3.23702037e-01 1.23357847e-01 6.84169173e-01 1.05047452e+00
4.60888386e-01 -5.13906002e-01 -7.05643356e-01 -3.35746527e-01
3.86951596e-01 6.77944303e-01 -7.96861589e-01 -3.11505288e-01
-8.06294754e-02 1.24700940e+00 5.25656343e-01 5.20721614e-01
-6.07896507e-01 3.74526173e-01 7.90824831e-01 -2.12361589e-01
2.90708870e-01 -7.08237350e-01 -8.60099792e-02 -1.15138483e+00
2.71321721e-02 -5.15543222e-01 2.78460324e-01 -1.10094547e+00
-6.56910479e-01 4.22791183e-01 1.80085778e-01 -1.15007722e+00
-1.21572030e+00 -4.52742279e-01 -2.29221657e-01 6.18267179e-01
-7.01862454e-01 -1.24494600e+00 -3.33253928e-02 4.51776475e-01
9.48593140e-01 -6.32939264e-02 9.54573572e-01 -5.71683884e-01
-2.00371534e-01 -4.22881871e-01 -3.06646496e-01 3.43730859e-02
2.83523709e-01 -1.20453477e+00 5.79325259e-01 6.48848295e-01
4.67500955e-01 5.18248677e-01 1.19496918e+00 -1.00474477e+00
-1.10689270e+00 -6.93001390e-01 1.08929050e+00 -7.38061130e-01
5.98326981e-01 -4.18097019e-01 -6.51057422e-01 1.43302071e+00
-7.80728012e-02 -8.35891142e-02 4.28064644e-01 2.39209726e-01
1.88401565e-01 5.75181901e-01 -8.01464140e-01 9.72365916e-01
1.37442946e+00 -5.49797654e-01 -8.40513945e-01 3.94356012e-01
5.70505500e-01 -4.19792384e-01 -4.20205951e-01 2.58033752e-01
8.03367436e-01 -1.08525348e+00 7.23190367e-01 -5.96772730e-01
3.97773981e-01 -7.53642246e-02 -2.83835083e-01 -1.34352612e+00
-5.39827704e-01 -7.17482090e-01 -3.38464797e-01 6.08531773e-01
4.31755483e-01 -8.24044198e-02 9.20687735e-01 9.98944640e-01
9.16402787e-03 -4.89246011e-01 -8.53547752e-01 -5.28061986e-01
-6.38481200e-01 -7.66089797e-01 5.72236717e-01 5.49092889e-01
3.74598265e-01 2.88853914e-01 -8.79684091e-01 -1.48239776e-01
5.12190819e-01 2.28526980e-01 8.29363465e-01 -1.13542283e+00
-4.26861286e-01 -2.44422406e-01 -2.58278847e-01 -1.32505095e+00
5.46157002e-01 -4.04069573e-01 6.18784070e-01 -1.75565243e+00
1.93319142e-01 1.15118781e-02 -1.44400626e-01 6.09873533e-01
8.15060288e-02 -3.75925541e-01 1.96819559e-01 4.05478716e-01
-9.85847652e-01 9.03414607e-01 1.10477173e+00 -4.62553576e-02
-5.00972390e-01 7.67083764e-02 1.27645815e-02 9.17706609e-01
6.68278694e-01 -5.36888480e-01 -4.36022907e-01 -1.65218145e-01
1.74957007e-01 8.92810762e-01 5.21796823e-01 -9.16634500e-01
2.91674644e-01 -7.55247831e-01 4.28202659e-01 -7.23709285e-01
8.35152030e-01 -7.36098170e-01 3.17291498e-01 6.21677935e-01
-9.67988074e-01 -1.55230597e-01 -4.85058308e-01 1.05950177e+00
3.42600867e-02 -4.74938788e-02 1.86089247e-01 -5.10131240e-01
-1.04951119e+00 2.26294100e-01 -1.04975533e+00 -4.11488444e-01
1.17965567e+00 -2.78462112e-01 3.25095892e-01 -6.87648416e-01
-1.20732462e+00 3.99987519e-01 6.01812124e-01 5.92304349e-01
5.30029297e-01 -1.28918278e+00 -3.00927550e-01 1.10982239e-01
3.01341742e-01 -2.43456900e-01 2.59158701e-01 5.38122654e-01
-1.10902056e-01 4.53083068e-01 -5.99459231e-01 -6.02479100e-01
-9.57093179e-01 5.20324111e-01 1.46574363e-01 -4.80644435e-01
-6.88898325e-01 6.14923298e-01 7.08604511e-03 -4.99729931e-01
4.15788829e-01 -3.98362130e-01 -3.65952611e-01 -1.08003594e-01
2.19811022e-01 4.33761835e-01 -7.69144595e-01 -9.85495508e-01
-3.05967808e-01 -1.00777946e-01 2.84327209e-01 -6.76864684e-01
1.38991523e+00 -2.77967870e-01 1.28663689e-01 1.25005603e+00
8.49095225e-01 -8.83828223e-01 -2.11576056e+00 -2.08515123e-01
2.87671894e-01 -2.94959903e-01 -2.93015540e-01 -7.27765858e-01
-2.41891369e-01 8.03248823e-01 2.67793715e-01 6.10164963e-02
6.04560256e-01 1.51726410e-01 5.46024859e-01 5.86467028e-01
8.41668606e-01 -1.25749040e+00 4.12885487e-01 5.29600024e-01
8.40059519e-01 -1.26715577e+00 -2.88196988e-02 2.82587670e-02
-1.31610692e+00 8.08378458e-01 5.70674956e-01 -1.58460379e-01
4.99646157e-01 -4.46993299e-02 -7.02268407e-02 -1.73089981e-01
-1.27503610e+00 -3.02856237e-01 1.77957460e-01 6.12115383e-01
3.11644167e-01 5.38260877e-01 1.70886949e-01 2.72232682e-01
-1.94496900e-01 2.45991483e-01 6.19377494e-01 1.11291742e+00
-3.65095705e-01 -8.36928308e-01 -1.59730986e-01 5.06469548e-01
-1.42007032e-02 5.84934175e-01 -2.66976953e-01 4.83990431e-01
4.34637703e-02 9.75761533e-01 1.77613989e-01 -2.89237350e-01
1.72317803e-01 4.04825181e-01 6.23367131e-01 -7.79252350e-01
4.54283692e-02 1.55428797e-01 3.00631195e-01 -9.72038984e-01
-6.59362018e-01 -1.41398776e+00 -1.34629917e+00 3.10459912e-01
1.29140005e-01 -1.30202502e-01 5.03024876e-01 1.33977437e+00
1.67613998e-01 5.96787095e-01 1.15583181e-01 -1.49262178e+00
-4.77296710e-01 -1.26233256e+00 -6.69757664e-01 5.26675463e-01
4.73828584e-01 -9.72965240e-01 -1.95484951e-01 4.58865017e-01] | [7.798001766204834, 0.5155978798866272] |
4a455cc6-ab1b-4013-a30b-ad715282008e | matching-thermal-to-visible-face-images-using | 1903.00963 | null | http://arxiv.org/abs/1903.00963v1 | http://arxiv.org/pdf/1903.00963v1.pdf | Matching Thermal to Visible Face Images Using a Semantic-Guided Generative Adversarial Network | Designing face recognition systems that are capable of matching face images
obtained in the thermal spectrum with those obtained in the visible spectrum is
a challenging problem. In this work, we propose the use of semantic-guided
generative adversarial network (SG-GAN) to automatically synthesize visible
face images from their thermal counterparts. Specifically, semantic labels,
extracted by a face parsing network, are used to compute a semantic loss
function to regularize the adversarial network during training. These semantic
cues denote high-level facial component information associated with each pixel.
Further, an identity extraction network is leveraged to generate multi-scale
features to compute an identity loss function. To achieve photo-realistic
results, a perceptual loss function is introduced during network training to
ensure that the synthesized visible face is perceptually similar to the target
visible face image. We extensively evaluate the benefits of individual loss
functions, and combine them effectively to learn the mapping from thermal to
visible face images. Experiments involving two multispectral face datasets show
that the proposed method achieves promising results in both face synthesis and
cross-spectral face matching. | ['Cunjian Chen', 'Arun Ross'] | 2019-03-03 | null | null | null | null | ['face-parsing'] | ['computer-vision'] | [ 8.52988601e-01 2.88136117e-02 1.59147695e-01 -7.36125886e-01
-8.10675383e-01 -6.45378590e-01 5.80880642e-01 -9.42086041e-01
1.78476095e-01 3.55409145e-01 -1.36534616e-01 8.73037577e-02
1.29883870e-01 -8.96306574e-01 -8.30218732e-01 -9.69151437e-01
5.55342376e-01 -8.23304430e-02 -6.77288592e-01 -7.64292926e-02
-1.81163624e-01 7.70566165e-01 -1.66875136e+00 3.64617556e-01
8.91992331e-01 1.35138535e+00 1.46963634e-02 3.14007849e-01
-3.78006473e-02 6.62454486e-01 -5.64474344e-01 -7.13771880e-01
6.83698475e-01 -7.00933456e-01 -4.26706254e-01 5.01148939e-01
8.45241785e-01 -2.70924509e-01 -2.20387653e-01 1.39742088e+00
4.17654663e-01 1.21147230e-01 8.14675450e-01 -1.55112052e+00
-1.01646996e+00 3.31779346e-02 -5.47846377e-01 -4.83563930e-01
2.25226685e-01 3.66648138e-01 7.37770438e-01 -8.97285879e-01
4.22109187e-01 1.53399658e+00 6.28999770e-01 8.32282066e-01
-1.26027644e+00 -9.42157090e-01 -9.81006697e-02 8.18560179e-03
-1.29345417e+00 -8.91632020e-01 1.28834879e+00 -3.35341871e-01
2.77347237e-01 1.69518337e-01 3.84594351e-01 1.26913726e+00
-8.50730855e-03 1.98919132e-01 1.21295261e+00 -4.33744758e-01
9.18903649e-02 2.21843049e-01 -7.86638260e-01 9.73422706e-01
-1.52568147e-01 4.04761910e-01 -5.66669285e-01 1.54333720e-02
7.92955697e-01 -4.04803716e-02 -2.82680213e-01 -1.87582821e-01
-3.50472569e-01 7.68572807e-01 6.33755863e-01 -8.03570170e-03
-4.84154105e-01 2.82905370e-01 -1.88358873e-01 3.18148911e-01
6.02413714e-01 1.47230759e-01 1.57466382e-02 6.46867335e-01
-8.82539630e-01 -1.43367693e-01 4.49291110e-01 8.10534596e-01
1.09661531e+00 4.81257230e-01 -1.37857944e-01 1.29875970e+00
6.21872902e-01 9.61594701e-01 1.10406466e-01 -1.24843192e+00
2.52279729e-01 5.53077579e-01 -1.83533460e-01 -8.83810043e-01
1.98872685e-01 -3.10731262e-01 -6.00672364e-01 6.01521432e-01
1.43684417e-01 -1.58433333e-01 -1.23830318e+00 2.00036883e+00
5.17979562e-01 3.05978626e-01 3.71126860e-01 9.97408211e-01
7.26710081e-01 5.62956095e-01 2.02001870e-01 -1.87915284e-02
1.17051852e+00 -8.24997306e-01 -5.85264444e-01 -5.48501968e-01
-1.88760117e-01 -9.34624791e-01 9.12542820e-01 -2.55659401e-01
-9.87293422e-01 -7.81955063e-01 -1.11916733e+00 5.26198521e-02
-1.28100380e-01 3.08446616e-01 2.74252683e-01 1.07297850e+00
-1.06013644e+00 3.95422041e-01 -6.73931777e-01 -1.69705898e-01
6.42838240e-01 2.29923964e-01 -3.35382849e-01 -3.54665160e-01
-9.90468264e-01 5.50297320e-01 8.84870887e-02 3.45476896e-01
-1.30742037e+00 -7.72736192e-01 -1.01912045e+00 -5.18199317e-02
1.53417379e-01 -8.41912091e-01 8.35965037e-01 -1.83350337e+00
-1.72016561e+00 1.14535868e+00 -1.26488939e-01 -3.81008955e-03
3.28982741e-01 2.08874926e-01 -6.20654285e-01 4.39507306e-01
1.38878912e-01 7.07701325e-01 1.49965358e+00 -1.54949296e+00
-1.70996249e-01 -6.05350912e-01 -1.10767640e-01 2.48031601e-01
-3.89766365e-01 1.97375506e-01 -6.04268432e-01 -7.85322726e-01
-2.00142190e-02 -1.03934491e+00 1.90472633e-01 3.18067521e-01
-5.32450318e-01 3.54374915e-01 8.02948236e-01 -9.29390907e-01
2.73963273e-01 -2.16984940e+00 5.83401248e-02 4.79680985e-01
-1.92109793e-01 4.95500378e-02 -4.81570870e-01 2.60496158e-02
-1.74054682e-01 -1.89214140e-01 -4.95017618e-01 -4.39121664e-01
-4.51059081e-02 1.22499868e-01 -1.47074923e-01 4.40817088e-01
5.76697886e-01 9.07002807e-01 -4.28335071e-01 -2.06355348e-01
1.74836352e-01 9.60025251e-01 -4.26381409e-01 5.63162804e-01
-3.16157132e-01 7.67005861e-01 -4.36946005e-01 8.35633397e-01
9.40262318e-01 -6.46336824e-02 2.65003353e-01 -4.26873356e-01
4.37248558e-01 -3.29853743e-01 -6.40471995e-01 1.58047831e+00
-8.23316336e-01 2.28696927e-01 4.98420715e-01 -8.26128244e-01
1.14522016e+00 2.02945635e-01 4.26113665e-01 -8.06528389e-01
2.36281455e-01 7.78852329e-02 -3.28654647e-01 -3.60561490e-01
1.73150841e-02 -4.56436306e-01 2.11377755e-01 3.87884349e-01
8.69339705e-02 -3.17223191e-01 -4.35687631e-01 -4.02782053e-01
5.87268591e-01 2.45843947e-01 -4.71142650e-01 -3.74430139e-03
8.10338736e-01 -4.19905454e-01 7.06399918e-01 2.72853911e-01
1.92448422e-02 7.43047476e-01 1.12934753e-01 -1.17290854e-01
-9.34221864e-01 -1.16883159e+00 1.27993375e-01 1.02459764e+00
1.24120981e-01 3.05548131e-01 -1.29007757e+00 -7.09255874e-01
7.18609616e-02 4.07807082e-01 -6.46952987e-01 -5.17649055e-01
-4.10959572e-01 -4.89434212e-01 5.48975348e-01 5.00986099e-01
7.68443227e-01 -1.08761001e+00 -1.34538323e-01 -2.24810198e-01
-1.47274405e-01 -1.31003141e+00 -6.33353949e-01 -2.47813448e-01
-3.17486137e-01 -1.13294232e+00 -6.54436767e-01 -8.70009303e-01
1.03276646e+00 1.21892266e-01 7.28560686e-01 9.32211336e-03
-5.68926275e-01 4.89940673e-01 -7.42261410e-02 -1.46355242e-01
-7.09276974e-01 -3.93931389e-01 -1.06924273e-01 8.04951906e-01
1.37268886e-01 -4.75144714e-01 -7.32296228e-01 3.10509443e-01
-9.75906491e-01 -1.08344764e-01 4.64106262e-01 8.28705966e-01
4.82243478e-01 1.32914498e-01 5.70515633e-01 -8.23136330e-01
1.35591015e-01 -2.56242156e-01 -7.72651374e-01 4.90163058e-01
-4.62728113e-01 5.02374433e-02 8.29629183e-01 -2.87182748e-01
-1.62692642e+00 3.90863925e-01 -3.20727259e-01 -7.79458165e-01
-1.08274102e-01 6.82573393e-02 -6.30785048e-01 -5.82122564e-01
5.53148687e-01 2.73701876e-01 2.13938996e-01 -2.27045074e-01
6.09296203e-01 6.80438459e-01 8.45128715e-01 -6.78491414e-01
1.22249496e+00 7.04805315e-01 1.15975607e-02 -6.35227263e-01
-8.90638947e-01 5.79573289e-02 -3.00232798e-01 -3.78768355e-01
9.64657962e-01 -1.04203320e+00 -5.31914055e-01 6.68133795e-01
-7.82695711e-01 -3.01279843e-01 -1.05058290e-01 9.36990976e-02
-6.40272319e-01 -7.93908630e-03 -4.57880467e-01 -9.21701849e-01
-4.72073168e-01 -1.15743005e+00 1.29229975e+00 4.80960459e-01
4.56373185e-01 -9.92599726e-01 -3.60287666e-01 8.45120966e-01
4.41565245e-01 4.97398555e-01 8.84242415e-01 1.73623264e-02
-5.65284610e-01 1.05688404e-02 -4.17211860e-01 8.64219308e-01
5.30493736e-01 3.02852374e-02 -1.50516677e+00 -3.64017159e-01
1.25788495e-01 -4.05382127e-01 7.28079379e-01 1.79883868e-01
1.26337779e+00 -3.16644341e-01 -6.04543425e-02 1.12864518e+00
1.44771779e+00 2.92499512e-01 5.88474393e-01 -1.89489663e-01
9.48337018e-01 9.49959576e-01 3.17686737e-01 1.92510337e-01
8.79429057e-02 7.18695283e-01 3.80644828e-01 -2.17855811e-01
-6.12738490e-01 -4.28211331e-01 4.75579172e-01 6.96209371e-02
1.48496374e-01 -2.12253518e-02 -5.94581485e-01 1.40995100e-01
-1.34502828e+00 -9.30810809e-01 5.38970470e-01 2.05241442e+00
7.44401455e-01 -4.31125671e-01 -2.29785681e-01 -1.81722939e-01
9.56757367e-01 1.55392513e-01 -7.95296669e-01 -1.13489700e-03
-2.41377160e-01 5.69063246e-01 3.33771437e-01 4.69124556e-01
-9.88107502e-01 1.15981853e+00 5.73092461e+00 6.18117332e-01
-1.47436821e+00 1.71596989e-01 9.28156614e-01 2.33480111e-02
-4.24617559e-01 -2.56681621e-01 -3.77125531e-01 4.23091650e-01
8.90518665e-01 3.96205969e-02 7.86968350e-01 7.72744715e-01
8.71903598e-02 4.79246706e-01 -9.14501607e-01 1.09646094e+00
3.60818028e-01 -8.87324333e-01 2.97053512e-02 -7.09504932e-02
1.06229341e+00 -3.89334708e-01 6.73655570e-01 -1.87944651e-01
3.20866972e-01 -1.37632143e+00 7.40790844e-01 5.92570126e-01
1.27826357e+00 -8.95657897e-01 2.59386510e-01 -2.13837191e-01
-1.27938068e+00 -1.53955564e-01 -2.55428314e-01 6.49593592e-01
-1.11789241e-01 3.36667985e-01 -6.75733387e-01 6.35157764e-01
4.99361634e-01 4.73118395e-01 -4.20358896e-01 3.67890179e-01
-4.46167409e-01 3.83381844e-01 -4.47976217e-02 6.67468905e-01
-1.14235297e-01 -4.59408730e-01 1.21460095e-01 6.43499970e-01
4.40893233e-01 -1.17554925e-01 1.06027372e-01 1.38227618e+00
-5.44174552e-01 -9.48128402e-02 -6.28119767e-01 -1.53028041e-01
5.92722476e-01 1.39295280e+00 -3.95149708e-01 -5.79965767e-03
-3.55543554e-01 1.31264448e+00 1.24503486e-01 6.51789725e-01
-9.73893166e-01 -1.61032453e-01 9.59379792e-01 -1.29198223e-01
2.64551878e-01 2.66207039e-01 -1.39898196e-01 -8.94348860e-01
1.81520626e-01 -8.27778161e-01 5.75099029e-02 -9.09955740e-01
-1.53219903e+00 7.44387567e-01 -5.10168791e-01 -9.41584051e-01
-1.93993971e-01 -6.05199397e-01 -6.30109131e-01 1.30360794e+00
-1.75109088e+00 -1.65066671e+00 -6.80801690e-01 8.65509868e-01
2.28408352e-01 -4.07457292e-01 6.61786795e-01 3.43171209e-01
-6.90368176e-01 9.86296892e-01 -3.03778462e-02 2.09290907e-01
6.00745082e-01 -7.56741047e-01 3.00010204e-01 8.52752388e-01
-4.68883328e-02 5.18587649e-01 3.71754616e-01 -4.42874938e-01
-1.51543963e+00 -1.64491701e+00 1.02055587e-01 -1.78054199e-01
1.69828907e-01 -2.17530087e-01 -7.40587771e-01 4.98489201e-01
-1.05084125e-02 3.82664114e-01 6.83839202e-01 -5.59830010e-01
-8.66168559e-01 -4.70900744e-01 -1.61613929e+00 3.06770444e-01
1.11735654e+00 -1.07703829e+00 -9.75713879e-03 3.41658056e-01
4.36503649e-01 -1.36584938e-01 -9.51200128e-01 5.41538537e-01
6.63694024e-01 -8.63277674e-01 1.24630785e+00 -3.12212378e-01
4.62644845e-01 -2.91090518e-01 -2.93527275e-01 -1.36648273e+00
9.36964676e-02 -5.20997465e-01 3.71687621e-01 1.44259858e+00
2.56912440e-01 -7.08571970e-01 9.51425135e-01 6.15500093e-01
-8.13153163e-02 -2.06068248e-01 -7.24909127e-01 -8.30131590e-01
5.16145416e-02 -7.62572587e-02 8.42622995e-01 1.02115667e+00
-8.28786492e-01 1.29835114e-01 -3.48168194e-01 5.13815701e-01
1.07494903e+00 2.51237839e-01 6.08040690e-01 -1.02673697e+00
-2.97028720e-01 -2.94981956e-01 -1.35593772e-01 -5.42366803e-01
9.33424890e-01 -1.02361476e+00 2.33145028e-01 -1.07695091e+00
1.20238848e-01 -4.46850985e-01 -1.17455013e-01 5.19444287e-01
-2.47753769e-01 7.46154189e-01 1.17658347e-01 2.24822573e-02
1.62171125e-01 8.44860256e-01 1.36546993e+00 -3.53005320e-01
1.46654144e-01 -1.59834981e-01 -7.76184320e-01 5.24862051e-01
6.76362634e-01 -3.34948599e-01 -5.81965864e-01 -4.17690784e-01
-3.93028110e-01 2.01918632e-01 5.49811065e-01 -1.04912567e+00
5.92122488e-02 -3.22813809e-01 6.87056899e-01 1.74861297e-01
8.41801524e-01 -1.03096950e+00 5.50610125e-01 1.93527192e-01
-3.58321995e-01 -4.66468483e-01 2.74724178e-02 4.50999856e-01
-2.55154580e-01 1.63889374e-03 1.28520536e+00 9.88490805e-02
-4.93761688e-01 6.00855649e-01 3.65624040e-01 -1.13543555e-01
1.11743903e+00 -1.89369440e-01 -3.11251551e-01 -3.60406965e-01
-5.47416806e-01 -1.29083648e-01 7.03717172e-01 5.16280413e-01
7.16567695e-01 -1.51828790e+00 -6.80842221e-01 6.35446012e-01
3.25440407e-01 -3.77176940e-01 4.13334996e-01 2.75361538e-01
-4.74150002e-01 -5.87759987e-02 -3.95808578e-01 -4.03329641e-01
-1.33989263e+00 3.90487611e-01 7.86018848e-01 4.25662309e-01
-2.65549690e-01 9.62417662e-01 3.96045029e-01 -6.29939914e-01
8.43350664e-02 2.96248406e-01 2.04231083e-01 -2.98796833e-01
4.49152857e-01 -4.78851534e-02 7.48924166e-02 -9.99233365e-01
-2.73920387e-01 7.97779977e-01 3.29340070e-01 -2.19222367e-01
1.17609513e+00 -7.62617514e-02 -1.97413132e-01 -2.60572493e-01
1.58475399e+00 -9.91512015e-02 -1.68077874e+00 -3.87979656e-01
-4.35698211e-01 -6.17248237e-01 9.33596268e-02 -8.79237652e-01
-1.73087871e+00 7.32110441e-01 9.23502564e-01 -1.27295971e-01
1.62665224e+00 -1.25691801e-01 6.11037493e-01 -4.96763550e-02
1.28030881e-01 -9.57366943e-01 3.62958819e-01 5.50914332e-02
8.44737709e-01 -1.24510384e+00 -4.54382390e-01 -7.57191896e-01
-5.88687122e-01 1.02480912e+00 6.53988540e-01 5.81117608e-02
3.63319218e-01 1.46219105e-01 4.01947618e-01 -1.75914928e-01
-3.10000777e-01 -2.80295819e-01 4.94433314e-01 7.37718225e-01
2.55148411e-01 -4.80296500e-02 4.98293668e-01 1.37233108e-01
-2.74889052e-01 -2.20443651e-01 -1.26935700e-02 6.97143674e-01
-1.82929784e-01 -1.09148800e+00 -6.14454627e-01 7.22873807e-02
-5.37631750e-01 2.20948867e-02 -5.95587552e-01 2.36929834e-01
8.24331939e-02 1.18681324e+00 -3.12822796e-02 -4.89303529e-01
5.21677509e-02 3.45622361e-01 8.12406003e-01 -5.97641945e-01
-3.71514082e-01 -8.15554857e-02 -1.36023894e-01 -6.17766380e-01
-4.65486407e-01 -3.39489520e-01 -1.00814879e+00 -2.77468711e-01
-6.32904544e-02 -1.18174523e-01 1.04814565e+00 7.40816295e-01
3.20561379e-01 5.00223160e-01 1.30160666e+00 -7.24653184e-01
-5.45809329e-01 -6.34337902e-01 -5.66961527e-01 7.57824063e-01
3.22072834e-01 -6.43077552e-01 -3.60423803e-01 3.49324763e-01] | [12.895934104919434, 0.11870142817497253] |
e18745df-b8d0-4b2a-a066-5911701ec01d | ask2transformers-zero-shot-domain-labelling | 2101.02661 | null | https://arxiv.org/abs/2101.02661v2 | https://arxiv.org/pdf/2101.02661v2.pdf | Ask2Transformers: Zero-Shot Domain labelling with Pre-trained Language Models | In this paper we present a system that exploits different pre-trained Language Models for assigning domain labels to WordNet synsets without any kind of supervision. Furthermore, the system is not restricted to use a particular set of domain labels. We exploit the knowledge encoded within different off-the-shelf pre-trained Language Models and task formulations to infer the domain label of a particular WordNet definition. The proposed zero-shot system achieves a new state-of-the-art on the English dataset used in the evaluation. | ['German Rigau', 'Oscar Sainz'] | 2021-01-07 | null | null | null | null | ['domain-labelling'] | ['natural-language-processing'] | [ 1.05611585e-01 4.08288166e-02 -5.41941881e-01 -6.20445907e-01
-3.96326125e-01 -6.19562447e-01 7.32581437e-01 4.00188297e-01
-9.75023091e-01 8.84219229e-01 -1.02732129e-01 -1.79072097e-01
5.70961982e-02 -9.12268400e-01 4.95789805e-03 -1.12894788e-01
3.57770056e-01 8.82821798e-01 5.84559262e-01 -6.78857982e-01
1.66044146e-01 1.65720329e-01 -1.58958638e+00 1.76766783e-01
5.31078458e-01 8.31179500e-01 3.37034464e-01 4.63942170e-01
-7.57534504e-01 6.93002820e-01 -4.54172730e-01 -4.11201090e-01
-7.56779686e-02 -1.58166274e-01 -9.14381742e-01 -9.44961328e-03
2.32392371e-01 1.13337517e-01 -5.53835370e-02 1.06214631e+00
3.49290609e-01 4.60589319e-01 7.78297663e-01 -7.60356605e-01
-5.58372021e-01 4.92398322e-01 -2.42967834e-03 4.16063011e-01
2.47632012e-01 -1.14775963e-01 1.30020654e+00 -1.04577851e+00
8.74553859e-01 1.01906288e+00 3.03627282e-01 6.85629427e-01
-1.11247349e+00 -8.25420558e-01 1.92936763e-01 3.40897560e-01
-1.72463238e+00 -2.48148456e-01 6.65899515e-01 -6.56566679e-01
1.57682645e+00 -3.95839065e-01 2.17917338e-01 1.15582991e+00
-2.83827595e-02 2.65542008e-02 1.09983575e+00 -9.88290846e-01
4.84719187e-01 4.60582376e-01 1.03894520e+00 7.30263114e-01
1.74642906e-01 -2.53630638e-01 -4.41802710e-01 -2.91298211e-01
1.99861333e-01 -2.24961475e-01 2.44880512e-01 -2.12022349e-01
-6.92143857e-01 1.00856793e+00 -1.93967789e-01 6.64925575e-01
-2.63778538e-01 -6.94843829e-02 7.69103289e-01 2.77113855e-01
8.77097428e-01 5.53167820e-01 -6.64353311e-01 1.15584448e-01
-1.01073468e+00 -2.02470183e-01 9.82306182e-01 9.58684564e-01
1.11852098e+00 -8.32496807e-02 -1.36362508e-01 1.13818383e+00
3.38423491e-01 1.91701293e-01 6.68005526e-01 -5.35642326e-01
2.17121676e-01 4.42994177e-01 2.27111295e-01 -3.87549251e-01
-2.94263542e-01 -2.46965483e-01 -2.40978263e-02 -1.29766777e-01
5.87338619e-02 -1.55287459e-01 -1.22138369e+00 1.74074614e+00
6.47259206e-02 6.34417236e-01 3.74013454e-01 5.75007498e-01
1.01834738e+00 7.23695099e-01 7.91501284e-01 4.00566831e-02
1.60112882e+00 -8.22898924e-01 -7.74174154e-01 -5.88758111e-01
7.31041193e-01 -6.72370613e-01 1.27295840e+00 2.45956168e-01
-4.73975688e-01 -6.12308323e-01 -1.21220446e+00 -8.78570303e-02
-1.02447069e+00 -2.52374187e-02 2.85300136e-01 7.10801125e-01
-7.65732110e-01 3.37297499e-01 -4.38943774e-01 -9.29944694e-01
-9.40510184e-02 8.63641649e-02 -2.12745517e-01 -1.96933270e-01
-1.87125850e+00 1.34711564e+00 9.83372986e-01 -8.25720966e-01
-1.07039654e+00 -5.05022109e-01 -1.10882521e+00 -8.08578264e-03
4.88062978e-01 -3.16255808e-01 1.34285653e+00 -8.25790107e-01
-1.52969813e+00 1.53960896e+00 -2.55317390e-01 -3.29830050e-01
-1.90786868e-01 -1.49410561e-01 -9.14842427e-01 2.69868486e-02
3.09287846e-01 5.43788850e-01 6.35165036e-01 -9.94539976e-01
-6.50414765e-01 3.92606184e-02 4.23529625e-01 9.23241675e-02
-6.36949539e-01 3.73994738e-01 -2.36057505e-01 -4.61262733e-01
-6.77831948e-01 -5.84896028e-01 -1.22460939e-01 -4.00274366e-01
-7.15762097e-03 -8.45585585e-01 5.56729615e-01 -3.59217465e-01
1.34882379e+00 -2.05345106e+00 -2.06343502e-01 1.60650104e-01
-1.95459589e-01 7.47735381e-01 -3.37540388e-01 6.38603926e-01
-4.58040694e-03 1.72494829e-01 -1.82342499e-01 -4.55775023e-01
2.71359622e-01 7.14626610e-01 -5.28711617e-01 2.84418136e-01
2.05869779e-01 3.82013619e-01 -1.38704860e+00 -6.58777952e-01
5.76381922e-01 3.02053183e-01 -2.14986295e-01 3.87944162e-01
-5.00504613e-01 -1.65263548e-01 -4.17648315e-01 2.33636618e-01
4.12407547e-01 5.32920193e-03 3.63234460e-01 1.04516849e-01
-5.08997068e-02 7.85680890e-01 -1.00639153e+00 2.16820717e+00
-8.77952218e-01 3.73744160e-01 -5.65945208e-01 -1.12147522e+00
1.20782542e+00 7.45097220e-01 9.26730111e-02 -2.96780288e-01
3.09410125e-01 4.22826290e-01 -7.85994753e-02 -4.55414414e-01
4.52416986e-01 -6.77977741e-01 -3.19518179e-01 4.34863776e-01
1.00284731e+00 -8.23902190e-02 7.04248190e-01 -1.55142441e-01
9.13594425e-01 2.61469483e-01 9.41054106e-01 -5.28976679e-01
7.89536536e-01 2.33497471e-01 4.42403048e-01 4.91453320e-01
-1.45649269e-01 1.05484545e-01 4.69638743e-02 -2.97357529e-01
-1.09158015e+00 -1.02352154e+00 -4.16156292e-01 1.74326849e+00
1.22709632e-01 -4.61955577e-01 -4.68231916e-01 -7.87110925e-01
-2.75228888e-01 1.48077738e+00 -3.67469996e-01 -4.45389077e-02
-2.34754562e-01 -1.53286666e-01 8.20538342e-01 2.71250457e-01
2.17328206e-01 -1.21296418e+00 -5.42098939e-01 3.56136471e-01
4.40034382e-02 -1.36749244e+00 -3.78225237e-01 4.38451290e-01
-3.43189269e-01 -8.42606783e-01 -4.73908901e-01 -1.18045855e+00
4.01245058e-01 -2.56662499e-02 1.57414508e+00 -2.46575817e-01
-2.78964229e-02 3.31897020e-01 -6.09603286e-01 -5.42201757e-01
-3.13120097e-01 4.51647073e-01 3.70159606e-03 -1.50329858e-01
1.35478723e+00 -4.22334254e-01 5.12716211e-02 1.22283757e-01
-9.54238594e-01 -4.93109852e-01 -1.73546246e-03 5.32119155e-01
3.93331587e-01 -7.18493536e-02 7.68302023e-01 -1.13853097e+00
8.54478121e-01 -7.60213494e-01 -5.11759222e-01 5.29519320e-01
-4.94284302e-01 3.86896878e-01 7.14575052e-01 -5.73683619e-01
-1.24790347e+00 -4.53364756e-03 -2.64802158e-01 -1.68458089e-01
-5.82008779e-01 6.80034697e-01 2.62034386e-02 1.55169025e-01
7.96286821e-01 -5.88231124e-02 -7.19889879e-01 -7.33173072e-01
7.88754106e-01 9.32334244e-01 1.72907680e-01 -6.43478036e-01
3.82943422e-01 2.94093192e-01 -3.69392753e-01 -9.70023334e-01
-1.47357523e+00 -1.15354550e+00 -9.63097215e-01 -2.14139391e-02
1.02990210e+00 -9.78753924e-01 7.35056819e-03 4.95749786e-02
-1.50285983e+00 -7.40799308e-02 -3.74107689e-01 4.10841137e-01
-2.63496757e-01 3.53697926e-01 -3.61363679e-01 -7.19883502e-01
-4.95593995e-01 -5.95596671e-01 7.87208796e-01 1.03726760e-01
-5.04448175e-01 -1.46399403e+00 6.33603871e-01 -2.03138351e-01
1.99236646e-01 -2.40885302e-01 9.30888236e-01 -1.41773653e+00
3.42064112e-01 -1.65303990e-01 -1.18721537e-01 6.58652723e-01
1.84887260e-01 -4.27187890e-01 -1.12306738e+00 -6.37668669e-02
-7.94721246e-02 -6.96183980e-01 7.46252418e-01 -4.20768373e-02
6.44966662e-01 9.87364799e-02 -2.54890949e-01 2.47686282e-01
1.81312895e+00 8.82633030e-02 2.86273539e-01 2.60887355e-01
2.94402122e-01 7.22279727e-01 7.80578077e-01 3.92522186e-01
2.89749533e-01 6.76712751e-01 -1.79122314e-01 8.67671892e-02
-6.22221157e-02 -3.07886094e-01 2.73312569e-01 8.67785275e-01
3.32456321e-01 -3.92100751e-01 -1.19808924e+00 9.62328434e-01
-1.70950007e+00 -6.84630036e-01 2.72680551e-01 2.01197171e+00
1.23416221e+00 2.93889850e-01 -2.16881275e-01 -3.18468511e-01
9.54510987e-01 2.83095211e-01 -2.99270779e-01 -8.49641323e-01
1.43901929e-01 1.00582492e+00 6.38979793e-01 8.30649614e-01
-1.13706732e+00 1.49582899e+00 6.61383104e+00 1.04631531e+00
-1.04820693e+00 6.65833712e-01 -4.92489040e-02 2.52354443e-01
-1.23279318e-01 8.52041543e-02 -1.12259221e+00 1.94077119e-01
1.34774923e+00 -5.93976140e-01 2.32902139e-01 1.04236662e+00
-2.16044858e-01 8.73349383e-02 -1.11723757e+00 7.01782346e-01
3.62836480e-01 -9.52176869e-01 1.31369516e-01 -2.94803262e-01
6.35456264e-01 3.19946796e-01 -4.67439771e-01 5.92005491e-01
7.60738850e-01 -9.69340861e-01 4.25807625e-01 4.77395743e-01
1.12829006e+00 -7.14604259e-01 9.87670481e-01 4.73149836e-01
-1.17334247e+00 3.79720330e-01 -7.91791260e-01 -3.55352432e-01
3.28960031e-01 5.93369484e-01 -1.10217035e+00 3.26387525e-01
9.99801457e-02 7.37048149e-01 -2.53413498e-01 5.55743098e-01
-7.39180744e-01 7.41833925e-01 -3.68299842e-01 -2.04501599e-01
4.95150208e-01 1.07288346e-01 2.71530300e-01 1.79624140e+00
1.91446751e-01 1.37355596e-01 6.36260211e-01 6.91499233e-01
-1.91115275e-01 4.22369152e-01 -7.54838586e-01 -2.37549394e-01
6.22587204e-01 1.16580057e+00 -5.02377450e-01 -8.75432789e-01
-7.88473845e-01 7.80912995e-01 5.89652598e-01 3.44180733e-01
-6.02651119e-01 -7.58661091e-01 9.02914882e-01 8.90164152e-02
3.18913996e-01 -2.56550968e-01 -1.48256555e-01 -1.30737126e+00
-5.48609734e-01 -3.56182158e-01 6.06200218e-01 -5.57818472e-01
-1.77623177e+00 6.00255847e-01 4.12667185e-01 -8.55727732e-01
-5.34434974e-01 -8.95020604e-01 -5.43494165e-01 1.16900814e+00
-1.89842498e+00 -9.25884843e-01 -1.33600309e-01 5.54879844e-01
6.74244463e-01 -3.72630239e-01 1.42722166e+00 2.96268672e-01
-1.30991608e-01 2.57092714e-01 -6.28157854e-02 2.29567930e-01
1.02281523e+00 -1.12335646e+00 4.38318133e-01 6.63042367e-01
4.30456460e-01 5.97084701e-01 8.42949152e-01 -6.56940699e-01
-3.63245785e-01 -1.10529947e+00 1.56161380e+00 -3.48506093e-01
1.10261881e+00 -4.57978308e-01 -9.09545541e-01 6.75911069e-01
5.87962270e-01 5.99072427e-02 1.21614587e+00 2.40083188e-01
-6.79901719e-01 2.23682486e-02 -1.22749686e+00 1.16651155e-01
8.48315477e-01 -9.28350925e-01 -1.49687004e+00 3.65695119e-01
9.81548011e-01 2.00096101e-01 -7.19369233e-01 2.26327494e-01
1.33397698e-01 -1.83492407e-01 9.39734042e-01 -9.82244134e-01
1.91229060e-01 -2.93005854e-01 -3.05581897e-01 -1.38282728e+00
-3.25095266e-01 -9.84704718e-02 1.39836878e-01 1.38520825e+00
5.59271157e-01 -6.99900150e-01 3.70932221e-01 3.90068024e-01
8.79602283e-02 -2.39980802e-01 -9.73757327e-01 -9.28881824e-01
1.89824656e-01 -5.88925898e-01 1.94948792e-01 1.19477355e+00
2.73715973e-01 8.95603418e-01 -1.76271245e-01 -4.04470153e-02
4.71551359e-01 -3.26171011e-01 1.07903883e-01 -1.40426898e+00
-1.17582068e-01 -1.05568478e-02 -4.80971783e-01 -6.84719384e-01
8.52850318e-01 -1.18697155e+00 1.38691589e-01 -1.54316092e+00
-3.43551151e-02 -2.54355699e-01 -1.00361955e+00 7.26500809e-01
-4.89453301e-02 3.36288780e-01 2.49888487e-02 -1.05753191e-01
-8.67386401e-01 3.40186864e-01 4.43919957e-01 -1.43623531e-01
1.08988602e-02 -4.73992556e-01 -3.60233694e-01 8.20351720e-01
9.39906955e-01 -9.25230920e-01 -6.70629025e-01 -3.79532665e-01
1.22128114e-01 -5.64392388e-01 -1.29117593e-01 -8.00574183e-01
1.76715717e-01 -3.31003398e-01 -3.58892471e-01 -2.00813800e-01
2.53970504e-01 -7.41324961e-01 -3.52070451e-01 2.57638395e-01
-5.29904485e-01 -2.07212970e-01 3.80718589e-01 3.13361675e-01
-2.76893944e-01 -1.02562952e+00 9.59953249e-01 -4.99655962e-01
-1.50452006e+00 -1.25101358e-01 -5.00802815e-01 5.73762238e-01
1.14902055e+00 2.29493067e-01 -5.50292656e-02 4.81764264e-02
-7.83788681e-01 6.30236790e-02 3.80750030e-01 5.78469753e-01
3.10376674e-01 -1.20080841e+00 -5.75088203e-01 -8.99392813e-02
8.02542984e-01 -5.60155511e-01 -1.99187577e-01 -1.83137983e-01
-3.01270306e-01 7.07447886e-01 -2.28445143e-01 -2.08777577e-01
-1.07091486e+00 6.98267341e-01 2.13663876e-01 -6.74593151e-01
-3.33344817e-01 7.38718331e-01 -4.43317965e-02 -6.64251864e-01
6.22460581e-02 -7.36145228e-02 -5.86891413e-01 2.82468140e-01
5.26701212e-01 6.63392097e-02 1.00628305e-02 -6.61371768e-01
-5.99778473e-01 5.50221384e-01 4.75707315e-02 -4.21927124e-01
1.13293982e+00 -1.08206253e-02 2.02856399e-02 1.01165962e+00
1.08965480e+00 -3.59280527e-01 -6.02164924e-01 -8.04119408e-01
4.54557568e-01 -1.94065318e-01 1.64663628e-01 -9.22918439e-01
-3.18580568e-01 1.06567824e+00 4.37562704e-01 -1.16863720e-01
8.95204961e-01 5.25489897e-02 7.68399477e-01 5.53231537e-01
7.22105920e-01 -1.56553566e+00 -1.04727879e-01 1.17604887e+00
2.99355209e-01 -1.05146074e+00 -1.65693790e-01 -4.12279099e-01
-6.94785714e-01 1.12370431e+00 5.56926847e-01 -5.91237485e-01
9.98117208e-01 1.22609548e-01 2.28891954e-01 -2.05400899e-01
-9.58228111e-01 -8.52705300e-01 4.56552237e-01 6.82079613e-01
7.88736403e-01 1.24059692e-01 -9.60494876e-01 6.04990840e-01
5.07824905e-02 2.77226299e-01 3.26665640e-01 7.49101162e-01
-8.53657901e-01 -1.48698473e+00 1.19232975e-01 8.49900022e-02
-4.50294197e-01 -6.16325438e-01 -3.54534268e-01 5.52195013e-01
5.70141435e-01 1.06048453e+00 2.66238786e-02 -1.08135663e-01
3.91289711e-01 8.74301851e-01 2.04006389e-01 -1.50814104e+00
-4.78694558e-01 -2.86857456e-01 5.55413485e-01 -1.10895529e-01
-6.42408848e-01 -2.02218607e-01 -1.39512396e+00 1.26369409e-02
-1.96366444e-01 2.86890149e-01 6.38042629e-01 1.32747018e+00
1.06690258e-01 4.66069788e-01 1.14685856e-01 -4.91791874e-01
-4.44952965e-01 -1.34136021e+00 -6.68492198e-01 6.54233098e-01
-1.05814144e-01 -9.06410456e-01 -1.99335188e-01 1.34914443e-01] | [10.466891288757324, 8.859926223754883] |
d1f36538-07e9-4509-a639-3f08c67fb809 | boosting-multiple-sclerosis-lesion-1 | null | null | https://www.sciencedirect.com/science/article/pii/S0010482523004869 | https://www.sciencedirect.com/science/article/pii/S0010482523004869 | Boosting multiple sclerosis lesion segmentation through attention mechanism | Magnetic resonance imaging is a fundamental tool to reach a diagnosis of multiple sclerosis and monitoring its progression. Although several attempts have been made to segment multiple sclerosis lesions using artificial intelligence, fully automated analysis is not yet available. State-of-the-art methods rely on slight variations in segmentation architectures (e.g. U-Net, etc.). However, recent research has demonstrated how exploiting temporal-aware features and attention mechanisms can provide a significant boost to traditional architectures. This paper proposes a framework that exploits an augmented U-Net architecture with a convolutional long short-term memory layer and attention mechanism which is able to segment and quantify multiple sclerosis lesions detected in magnetic resonance images. Quantitative and qualitative evaluation on challenging examples demonstrated how the method outperforms previous state-of-the-art approaches, reporting an overall Dice score of 89% and also demonstrating robustness and generalization ability on never seen new test samples of a new dedicated under construction dataset. | ['S. Battiato', 'F. Pappalardo', 'D. Maimone', 'C. Di Lorenzo', 'G. Russo', 'A. Ortis', 'O. Giudice', 'F. Guarnera', 'E. Crispino', 'A. Rondinella'] | 2023-07-01 | null | null | null | computers-in-biology-and-medicine-2023-7 | ['brain-image-segmentation', 'lesion-segmentation'] | ['medical', 'medical'] | [ 4.43452060e-01 2.58780401e-02 -1.16009563e-01 -4.30761844e-01
-9.21674848e-01 -1.49514705e-01 4.93807107e-01 1.86548874e-01
-7.24278867e-01 6.64349496e-01 -1.53373331e-01 -3.89456861e-02
-4.88045841e-01 -4.99778032e-01 -4.32533026e-01 -5.17188489e-01
-6.70994699e-01 9.03222620e-01 6.15595579e-01 -1.70100741e-02
3.53252113e-01 7.82915890e-01 -1.20486808e+00 4.37030137e-01
7.93769538e-01 8.11708808e-01 4.93588120e-01 6.72392607e-01
-1.60239846e-01 5.92984498e-01 -3.57917935e-01 2.09546804e-01
3.91127020e-02 -2.89754927e-01 -1.26839244e+00 1.90146789e-01
3.01060170e-01 -3.27976108e-01 -1.38876140e-01 7.10290670e-01
6.83546364e-01 -1.76332484e-03 5.65034032e-01 -5.30411303e-01
-6.76697254e-01 4.31354910e-01 -5.94239175e-01 8.71433973e-01
1.70090750e-01 2.39240944e-01 5.66662192e-01 -3.75223994e-01
1.05547047e+00 8.57052267e-01 7.74773180e-01 5.08816004e-01
-1.41119289e+00 -3.79397482e-01 -1.79577544e-01 5.28181791e-01
-9.57145274e-01 3.58537883e-02 3.26618969e-01 -6.39177203e-01
1.06296659e+00 1.50466472e-01 6.70284390e-01 8.61094177e-01
2.69677222e-01 1.02524471e+00 1.66732049e+00 -4.39599335e-01
1.54174566e-01 -2.57230759e-01 6.52975678e-01 5.88402927e-01
-2.05940753e-01 -5.24469465e-03 -1.69513121e-01 1.94754303e-01
9.11600828e-01 -3.02380044e-02 -2.04272881e-01 -3.22275192e-01
-1.73465645e+00 6.53851271e-01 8.54985952e-01 1.12005424e+00
-7.85615683e-01 1.76281989e-01 4.33503658e-01 1.29100978e-01
5.08287966e-01 2.52117217e-01 -3.02977294e-01 2.51589231e-02
-1.51370597e+00 1.69236183e-01 1.61005720e-01 1.28849074e-01
3.51050645e-01 -3.26811641e-01 -4.72242475e-01 7.75583923e-01
-8.39812756e-02 2.71894723e-01 7.99592376e-01 -7.41275668e-01
1.19145215e-01 7.75708854e-01 -3.48067075e-01 -5.30584395e-01
-1.05324531e+00 -7.23859012e-01 -7.64328718e-01 6.95149720e-01
5.00471354e-01 -2.30873913e-01 -1.44081783e+00 1.31253338e+00
1.08554758e-01 2.96909928e-01 -4.22834843e-01 1.07784724e+00
2.82208264e-01 -7.32092336e-02 -1.89194586e-02 -2.93430630e-02
1.24798501e+00 -9.36143339e-01 -6.06461883e-01 -1.49703905e-01
5.60591936e-01 -3.79900008e-01 8.28176856e-01 3.43856156e-01
-1.15960836e+00 -2.45533735e-01 -1.00739884e+00 1.55344054e-01
-4.76264745e-01 -1.92786846e-02 6.88351274e-01 7.42585659e-01
-1.32111788e+00 1.05769134e+00 -1.16855192e+00 -6.19058371e-01
8.37117434e-01 5.27109921e-01 -6.71939373e-01 1.14803061e-01
-1.10099983e+00 1.07210541e+00 3.68387103e-01 6.58788010e-02
-5.34105062e-01 -7.72939503e-01 -3.32826495e-01 -2.47147635e-01
9.47458148e-02 -7.50982523e-01 1.10329914e+00 -8.10863853e-01
-1.31344187e+00 1.03651559e+00 5.62805198e-02 -1.04757237e+00
9.86070514e-01 -1.57839879e-01 -2.83757329e-01 7.32734859e-01
1.14945076e-01 9.13374186e-01 5.41449666e-01 -8.80636215e-01
-5.36760211e-01 -6.69737637e-01 -1.39904127e-01 -2.25873962e-01
-1.86155975e-01 2.72690624e-01 -8.02091137e-02 -7.51515865e-01
3.76858413e-01 -8.03076863e-01 -5.55448472e-01 -2.73209922e-02
-2.75616914e-01 5.72538264e-02 7.54574418e-01 -1.04167783e+00
1.07235706e+00 -1.33967078e+00 1.54091373e-01 2.68635511e-01
2.47058287e-01 5.55700481e-01 7.18463138e-02 -3.98443416e-02
-3.58145744e-01 1.75105348e-01 -6.39621556e-01 -3.34140062e-02
-2.33451039e-01 3.78752640e-03 2.73020476e-01 5.70088327e-01
2.16314331e-01 1.09676349e+00 -8.19774091e-01 -4.03270572e-01
5.15591025e-01 7.25382090e-01 -2.96931684e-01 -2.63182551e-01
-3.95219103e-02 8.18642080e-01 -3.40483040e-01 6.14682496e-01
2.72417068e-01 -3.32745969e-01 4.96144108e-02 -8.60670023e-03
-8.68022218e-02 -2.64707744e-01 -1.09986997e+00 1.93224192e+00
-3.63232255e-01 5.91341376e-01 -1.49597332e-01 -1.31950569e+00
5.55160701e-01 4.57296461e-01 8.50564599e-01 -1.03556085e+00
3.45937133e-01 5.08330941e-01 2.07677007e-01 -6.94292009e-01
-2.11322177e-02 -2.45182708e-01 4.46003318e-01 5.92293322e-01
9.77971703e-02 3.68058860e-01 3.98396194e-01 6.63584890e-03
1.47780371e+00 2.56345749e-01 -5.15574850e-02 -2.86561549e-01
6.94192171e-01 -3.16595510e-02 8.87831151e-02 7.91186929e-01
-6.24958277e-01 9.72680628e-01 3.02887976e-01 -4.37939286e-01
-1.05749500e+00 -9.44717944e-01 -4.81665224e-01 5.94333649e-01
-3.99426967e-01 2.64264047e-01 -1.15463471e+00 -5.67162931e-01
-2.47551858e-01 4.03586328e-01 -1.06945193e+00 3.56875896e-01
-7.52655685e-01 -8.70488405e-01 4.02279377e-01 7.37977684e-01
6.82967246e-01 -1.30180788e+00 -1.10890687e+00 6.31718993e-01
3.67230084e-03 -9.98805225e-01 1.17143549e-01 3.30067091e-02
-1.27680576e+00 -1.15821719e+00 -1.52976871e+00 -6.29263520e-01
4.31533813e-01 -1.25204295e-01 8.77996862e-01 -9.91972536e-02
-6.94450200e-01 2.64176816e-01 -1.93717748e-01 -4.61112894e-03
-1.90458789e-01 2.33021975e-01 -2.88752437e-01 -1.92169473e-03
1.58502012e-01 -8.14541101e-01 -8.63597393e-01 3.65046076e-02
-1.07386315e+00 -8.22618976e-02 9.38769698e-01 6.93424225e-01
8.97420883e-01 -4.36989099e-01 6.86771452e-01 -9.51891005e-01
4.94236290e-01 -3.37783158e-01 -1.16929613e-01 2.72725672e-01
-6.91573262e-01 5.42126670e-02 7.08205476e-02 -1.54698834e-01
-7.12942123e-01 -1.81326140e-02 -1.94663867e-01 -4.31141526e-01
-4.46467608e-01 4.09992695e-01 4.97441292e-01 -3.16259265e-01
5.71029186e-01 1.49296135e-01 2.18092337e-01 -5.40896237e-01
2.63897926e-01 5.76531589e-01 5.88567793e-01 -1.04222663e-01
1.40851483e-01 7.68862426e-01 7.64017850e-02 -7.68163443e-01
-6.20919466e-01 -6.84641838e-01 -1.10745490e+00 -4.36550736e-01
1.19307292e+00 -2.23066419e-01 -3.07645500e-01 6.58803940e-01
-1.06596494e+00 -4.24481422e-01 -1.48619086e-01 4.96234387e-01
-8.53194892e-01 2.29330495e-01 -5.30785441e-01 -4.30743277e-01
-6.08379960e-01 -1.28599358e+00 9.07434225e-01 7.01654926e-02
-3.25418264e-01 -1.03899729e+00 3.17702383e-01 6.37117624e-01
7.45043099e-01 5.43352008e-01 8.14821899e-01 -6.58729374e-01
-2.27531940e-01 -4.81934160e-01 -4.08746421e-01 2.29948193e-01
-6.04142807e-02 -3.92355025e-01 -7.82299876e-01 -1.14715599e-01
-1.59277827e-01 -1.09804429e-01 1.05818141e+00 8.99571717e-01
1.06904960e+00 3.53991866e-01 -4.26057994e-01 2.38054246e-01
1.42462742e+00 1.05254844e-01 8.57773066e-01 7.97683477e-01
3.03466737e-01 6.72294796e-01 1.18466474e-01 -1.82255581e-01
1.49275541e-01 6.64100349e-01 3.16030055e-01 -1.99406922e-01
-5.31779647e-01 7.61854529e-01 -1.58688053e-01 4.16900218e-01
-2.55212545e-01 3.73724967e-01 -1.45526767e+00 9.67507184e-01
-1.78331351e+00 -1.06066334e+00 -3.25876474e-01 1.94331872e+00
5.66537797e-01 3.10714126e-01 1.27123311e-01 4.15256768e-01
8.87904823e-01 -1.40276989e-02 -5.30198336e-01 -1.98606685e-01
-1.25740007e-01 6.70788348e-01 3.61668110e-01 3.09862942e-01
-1.13995600e+00 5.73428512e-01 6.77073526e+00 7.93719769e-01
-1.26993382e+00 4.53975499e-01 6.63959444e-01 -1.44371167e-01
7.02672601e-02 -2.58785069e-01 -1.99864015e-01 2.93010235e-01
9.65165555e-01 1.98988497e-01 1.70396999e-01 2.85296410e-01
3.08866084e-01 -2.93114424e-01 -7.01941013e-01 7.06584632e-01
2.99845319e-02 -1.44923615e+00 -9.60850492e-02 5.26383929e-02
6.65613830e-01 5.13920486e-01 -4.42169718e-02 9.09129381e-02
-1.75835982e-01 -1.18019664e+00 4.02958453e-01 1.03184068e+00
5.25892496e-01 -6.20187044e-01 9.25974429e-01 2.12429389e-01
-8.18361044e-01 1.15255646e-01 1.17135450e-01 1.93797201e-01
3.72729361e-01 6.31180525e-01 -7.25990891e-01 7.60748744e-01
8.25287104e-01 4.65022504e-01 -7.64749825e-01 1.29981232e+00
-5.29415382e-04 5.29110074e-01 -2.85192132e-01 2.45069191e-01
7.87728310e-01 -1.71356946e-01 8.02245796e-01 1.35706055e+00
1.91712633e-01 -1.07950889e-01 1.08794712e-01 8.31511319e-01
3.39702874e-01 2.82088578e-01 -2.82786906e-01 -6.39148504e-02
-4.00551438e-01 1.38336241e+00 -1.30770504e+00 -2.76368231e-01
-2.82565206e-01 1.04093170e+00 2.31097758e-01 2.33082920e-01
-6.74156129e-01 -3.01363021e-01 -1.24254543e-02 3.84691626e-01
2.89030284e-01 -1.61152378e-01 -5.84979892e-01 -6.93692088e-01
-7.24155009e-02 -5.96333265e-01 4.16151911e-01 -7.89574742e-01
-8.90192568e-01 8.10788453e-01 -2.93462127e-01 -8.79109859e-01
-3.29254061e-01 -6.06734931e-01 -5.26413083e-01 9.47104335e-01
-1.45875931e+00 -1.37880635e+00 -6.62662927e-03 5.82063138e-01
3.85468513e-01 -1.95115790e-01 7.82409906e-01 3.63463640e-01
-3.82732332e-01 7.24801123e-02 1.61857501e-01 6.15166500e-02
4.40454304e-01 -1.50681186e+00 1.64424971e-01 8.40530396e-01
-1.06252648e-01 5.62247038e-01 6.45921350e-01 -6.07834280e-01
-7.15912819e-01 -8.76562119e-01 5.98902047e-01 -1.21917062e-01
9.80197966e-01 4.48747456e-01 -1.08079541e+00 6.14504397e-01
3.31869066e-01 2.19403297e-01 5.25495410e-01 -2.03720257e-01
1.05681382e-01 3.71915221e-01 -1.44733000e+00 4.37489271e-01
8.35861504e-01 -3.09863746e-01 -7.43961871e-01 2.98449367e-01
2.51742005e-01 -2.66594917e-01 -1.05571735e+00 5.87507248e-01
6.54203534e-01 -1.18410528e+00 9.44723606e-01 -6.84960723e-01
4.34882015e-01 -5.54016866e-02 1.37687027e-01 -1.14026475e+00
-2.53554285e-01 -1.86035290e-01 -2.43647456e-01 7.53756106e-01
2.57122308e-01 -5.34602344e-01 9.31683958e-01 2.36108795e-01
-1.49407759e-01 -1.19371676e+00 -1.05843973e+00 -7.92274415e-01
4.62164432e-01 -5.76083362e-01 5.85121065e-02 7.58676827e-01
-3.53897661e-01 5.54481223e-02 4.68965955e-02 -1.04644150e-01
7.45350182e-01 6.35280684e-02 -2.46419348e-02 -1.35914004e+00
-5.96528426e-02 -8.55886281e-01 -8.15624416e-01 -1.73816621e-01
1.83264226e-01 -1.29437840e+00 -4.77549642e-01 -2.09537864e+00
2.27349654e-01 -2.57588089e-01 -6.62882030e-01 4.67441082e-01
7.46020824e-02 7.39779949e-01 5.99770434e-03 1.71196833e-01
-3.98678333e-01 3.31044085e-02 1.39539838e+00 -2.85688251e-01
-1.12102531e-01 -7.16051161e-02 -1.68884262e-01 6.86079860e-01
1.05642605e+00 -2.43939415e-01 -7.90306255e-02 -4.03730124e-01
-1.60399586e-01 -1.25496611e-01 7.07182527e-01 -1.64408398e+00
1.95687100e-01 3.37245733e-01 4.76472408e-01 -6.82447970e-01
-7.28241280e-02 -6.29531503e-01 -1.42926322e-02 1.00867021e+00
-3.39665562e-01 -2.74989367e-01 1.98803291e-01 4.03782934e-01
-2.59386480e-01 -3.02841157e-01 1.04382718e+00 -3.24142486e-01
-8.23239386e-01 4.64678049e-01 -3.87152910e-01 -2.02095151e-01
1.23757875e+00 -3.61881047e-01 -3.00963577e-02 1.27666742e-01
-1.32428432e+00 2.24204943e-01 1.56221613e-01 3.80106211e-01
3.96794230e-01 -1.12876177e+00 -8.08577597e-01 -1.60540402e-01
-1.45060316e-01 -5.32595813e-01 6.09455228e-01 1.54488671e+00
-8.58795166e-01 7.48391151e-01 -6.81004524e-01 -9.05264318e-01
-1.21471429e+00 3.58952552e-01 6.47933245e-01 -6.88719869e-01
-1.07210135e+00 5.91247380e-01 -4.70771581e-01 -5.48503935e-01
9.12433714e-02 -2.02870771e-01 -5.67013562e-01 3.24579298e-01
7.39914477e-01 6.65577531e-01 5.08089125e-01 -5.26019931e-01
-4.20582116e-01 6.35197341e-01 -2.87503958e-01 -4.00216997e-01
1.58902216e+00 -1.41509235e-01 -1.66094556e-01 4.53915864e-01
9.89283144e-01 -7.00826883e-01 -8.85713279e-01 -2.68637240e-01
4.63704437e-01 3.82564552e-02 5.50636232e-01 -1.16526783e+00
-1.28229940e+00 1.28301752e+00 1.47480035e+00 2.83223927e-01
9.39167976e-01 -2.42858037e-01 1.05054796e+00 2.58958369e-01
4.87691730e-01 -1.03043437e+00 -6.99135512e-02 4.27107483e-01
8.77053618e-01 -1.16401136e+00 -2.85160899e-01 6.97271107e-03
-4.23513025e-01 1.34671330e+00 1.60994619e-01 -3.07916492e-01
5.67552686e-01 2.07958803e-01 2.23438814e-01 -4.09029692e-01
-1.47910655e-01 -5.61864793e-01 3.26467574e-01 7.25765347e-01
5.14373064e-01 -1.54890949e-02 -7.00747788e-01 5.04971981e-01
2.77171254e-01 4.67041939e-01 3.30177903e-01 9.41901147e-01
-6.09895110e-01 -1.14188635e+00 -4.04037952e-01 7.13633537e-01
-9.67359841e-01 8.56933892e-02 -1.92165434e-01 8.64006400e-01
1.12376459e-01 4.46365505e-01 -9.54747871e-02 2.40656845e-02
2.25056216e-01 3.13206196e-01 8.28504920e-01 -3.46209049e-01
-8.85309756e-01 2.67805997e-03 -2.01654240e-01 -7.64547706e-01
-8.37480187e-01 -7.22357392e-01 -1.62406301e+00 5.16669974e-02
-7.38873407e-02 -2.66837955e-01 7.24647403e-01 1.18485689e+00
2.51428962e-01 1.11027348e+00 1.03141919e-01 -1.05744910e+00
-1.51119843e-01 -1.09399867e+00 -7.32530892e-01 3.76694143e-01
2.48451024e-01 -7.26542294e-01 8.84870216e-02 1.11718560e-02] | [14.239950180053711, -2.063462257385254] |
1afcce53-c5ac-49a2-bda4-f1268d0138fb | weakly-supervised-deep-learning-model-for | 2212.12844 | null | https://arxiv.org/abs/2212.12844v1 | https://arxiv.org/pdf/2212.12844v1.pdf | Weakly-Supervised Deep Learning Model for Prostate Cancer Diagnosis and Gleason Grading of Histopathology Images | Prostate cancer is the most common cancer in men worldwide and the second leading cause of cancer death in the United States. One of the prognostic features in prostate cancer is the Gleason grading of histopathology images. The Gleason grade is assigned based on tumor architecture on Hematoxylin and Eosin (H&E) stained whole slide images (WSI) by the pathologists. This process is time-consuming and has known interobserver variability. In the past few years, deep learning algorithms have been used to analyze histopathology images, delivering promising results for grading prostate cancer. However, most of the algorithms rely on the fully annotated datasets which are expensive to generate. In this work, we proposed a novel weakly-supervised algorithm to classify prostate cancer grades. The proposed algorithm consists of three steps: (1) extracting discriminative areas in a histopathology image by employing the Multiple Instance Learning (MIL) algorithm based on Transformers, (2) representing the image by constructing a graph using the discriminative patches, and (3) classifying the image into its Gleason grades by developing a Graph Convolutional Neural Network (GCN) based on the gated attention mechanism. We evaluated our algorithm using publicly available datasets, including TCGAPRAD, PANDA, and Gleason 2019 challenge datasets. We also cross validated the algorithm on an independent dataset. Results show that the proposed model achieved state-of-the-art performance in the Gleason grading task in terms of accuracy, F1 score, and cohen-kappa. The code is available at https://github.com/NabaviLab/Prostate-Cancer. | ['Sheida Nabavi', 'Ga Hie Nam', 'Harold Yamase', 'Anna Tarakanova', 'Ankit Bhardwaj', 'Jun Bai', 'Hanzhang Wang', 'Mohammad Madani', 'Mohammad Mahdi Behzadi'] | 2022-12-25 | null | null | null | null | ['multiple-instance-learning'] | ['methodology'] | [ 3.06086361e-01 2.74320275e-01 -2.21674144e-01 -6.53193951e-01
-1.17234576e+00 -5.19584298e-01 4.17196393e-01 6.71573937e-01
-4.53900665e-01 5.92283607e-01 -1.69220448e-01 -1.51896223e-01
-1.33333743e-01 -1.10277498e+00 -3.84164929e-01 -1.07357347e+00
-2.40977496e-01 7.82443166e-01 1.24896660e-01 1.93006098e-01
3.89887869e-01 5.11163175e-01 -7.88014889e-01 8.48317444e-02
9.95758712e-01 8.96325231e-01 2.17097968e-01 9.15705144e-01
-4.89817485e-02 5.63763142e-01 -1.80774420e-01 -6.48902595e-01
7.61735812e-02 -3.39735359e-01 -6.74229801e-01 -1.99532527e-02
4.43817317e-01 -1.31244525e-01 -1.65390760e-01 1.20682478e+00
3.76669049e-01 -3.65380049e-01 6.97673082e-01 -9.06728566e-01
-6.58121943e-01 2.88096488e-01 -6.19472325e-01 2.54113048e-01
-3.53592455e-01 2.54181653e-01 9.91338611e-01 -5.97420633e-01
5.48367441e-01 6.75904393e-01 6.99607134e-01 5.84038138e-01
-1.09110606e+00 -6.40733480e-01 -4.22161549e-01 3.32060784e-01
-1.18683732e+00 2.92618833e-02 7.70740569e-01 -4.09520656e-01
4.69418615e-01 3.13119560e-01 8.44157279e-01 5.95918119e-01
1.01901364e+00 5.98418355e-01 1.36790323e+00 -9.26473364e-02
3.21122408e-01 -2.47783750e-01 3.12359333e-01 9.92865562e-01
3.56913805e-01 -1.03854656e-01 1.81758150e-01 -1.07906267e-01
7.31554091e-01 4.20954615e-01 1.59524400e-02 -2.65014678e-01
-7.67680049e-01 9.98048484e-01 1.15705168e+00 3.01089257e-01
-2.66167670e-01 2.41146430e-01 4.46000367e-01 -2.41462663e-01
2.90317386e-01 2.01883957e-01 1.89519510e-01 3.15775603e-01
-8.47589254e-01 4.77163978e-02 5.09076059e-01 3.56677949e-01
5.59082389e-01 -5.24355710e-01 -7.46827871e-02 6.50368631e-01
6.14655793e-01 4.49969143e-01 7.05254316e-01 -4.97185290e-01
7.63969049e-02 1.02086377e+00 -3.54134709e-01 -1.11938322e+00
-5.64328849e-01 -6.68719709e-01 -1.24049830e+00 1.41304106e-01
4.76214111e-01 4.75557894e-01 -1.30398989e+00 1.14958942e+00
4.09531444e-02 4.34183553e-02 -1.65988475e-01 1.08225012e+00
9.55273390e-01 2.96363324e-01 4.19455945e-01 1.83223635e-01
1.60425627e+00 -9.11087096e-01 -4.41815436e-01 -3.15544426e-01
5.86002886e-01 -3.52526933e-01 7.30170846e-01 1.17285147e-01
-6.69794858e-01 -3.35143894e-01 -1.01986861e+00 -1.14618696e-01
-4.39675927e-01 4.30507571e-01 1.05322480e+00 7.32770145e-01
-1.08310843e+00 4.99039501e-01 -1.32105958e+00 -5.17252326e-01
7.56295919e-01 5.12637198e-01 -5.71462214e-01 -2.26831645e-01
-6.84230268e-01 5.17753303e-01 2.00188234e-01 5.11080325e-01
-1.15540838e+00 -5.78105211e-01 -7.84014463e-01 1.92068920e-01
-1.57075495e-01 -6.51062667e-01 8.82794082e-01 -6.22051895e-01
-1.02747643e+00 1.13368654e+00 -8.48529637e-02 -3.10659766e-01
3.13736558e-01 5.45918763e-01 -1.84100643e-01 1.52790412e-01
7.00499415e-02 5.20401061e-01 8.32401142e-02 -1.10239458e+00
-4.66761738e-01 -1.05158341e+00 -2.47873724e-01 8.31827521e-03
-1.33011132e-01 -5.13451099e-01 -4.52339143e-01 -1.22246705e-01
2.00797558e-01 -1.09155285e+00 -6.11576319e-01 -1.46708891e-01
-5.60093939e-01 -2.93010533e-01 5.20359755e-01 -9.48909938e-01
9.26710665e-01 -2.17157984e+00 5.08401468e-02 4.93055791e-01
4.75720704e-01 1.21798508e-01 1.27291754e-01 -2.79076174e-02
1.56159878e-01 3.89192611e-01 -1.65297791e-01 -2.12121814e-01
-8.04336965e-02 2.11158827e-01 3.46592993e-01 5.32860577e-01
1.69002086e-01 9.58056867e-01 -1.03541470e+00 -7.29437232e-01
9.33785588e-02 4.00367528e-01 -2.90428430e-01 1.64367631e-01
1.93274572e-01 3.81243706e-01 -4.77740616e-01 1.02416110e+00
6.05417788e-01 -5.87153614e-01 4.05952573e-01 -3.11805755e-01
3.17383349e-01 -2.38826767e-01 -5.24323046e-01 1.48343587e+00
-3.18728164e-02 2.80308247e-01 -1.35951698e-01 -7.26139963e-01
1.01339519e+00 -4.50538620e-02 3.70835930e-01 -5.90200543e-01
1.12456135e-01 4.29683983e-01 2.28651151e-01 -3.99263144e-01
9.97320041e-02 -2.18916655e-01 -2.03338832e-01 -8.75354558e-02
6.47710636e-02 -2.04003677e-02 5.73470473e-01 2.31490597e-01
1.59467673e+00 -4.26574618e-01 4.33698475e-01 -1.97640941e-01
6.00411057e-01 2.30282798e-01 6.89080715e-01 4.07555848e-01
-5.10424733e-01 5.18303633e-01 6.59711540e-01 -6.35697842e-01
-8.52386177e-01 -1.29815185e+00 -3.39135557e-01 3.98144364e-01
1.04572155e-01 -1.63351491e-01 -3.86462599e-01 -7.46219039e-01
1.49791956e-01 2.80167550e-01 -9.47646856e-01 6.50899112e-02
-3.84325534e-01 -1.11032093e+00 4.96739894e-01 6.47198439e-01
6.42530560e-01 -9.12029207e-01 -1.30543202e-01 8.07988364e-03
8.22524875e-02 -5.53965747e-01 -2.40795523e-01 1.93334356e-01
-1.11964977e+00 -1.37561119e+00 -7.86405563e-01 -8.99471760e-01
1.34295619e+00 -9.46092233e-02 9.06793773e-01 1.77751526e-01
-7.13857770e-01 8.46873596e-02 8.13853890e-02 -2.05110922e-01
-2.29978979e-01 1.76729053e-01 -6.51394367e-01 -2.99071311e-03
3.62512082e-01 -3.22943658e-01 -9.87389207e-01 -8.52869526e-02
-7.88630128e-01 1.22623695e-02 1.44755697e+00 8.57860148e-01
1.22748733e+00 1.05731804e-02 3.41708452e-01 -1.23183072e+00
4.19720322e-01 -3.66974860e-01 -4.89806652e-01 1.60788804e-01
-6.95336044e-01 -2.45178267e-01 7.03839540e-01 3.04595768e-01
-7.36371458e-01 2.28364110e-01 -6.47737533e-02 7.87878409e-02
-1.00734532e-01 7.95728743e-01 6.57484010e-02 -2.37973988e-01
2.69938886e-01 2.21317440e-01 1.07546628e-01 -1.02700233e-01
-2.98444331e-01 5.23585021e-01 5.80881774e-01 -1.36536434e-01
5.35204470e-01 4.36507940e-01 5.07885873e-01 -4.68348384e-01
-6.28737092e-01 -6.02619290e-01 -3.34358871e-01 -3.95538628e-01
1.12995446e+00 -7.09838748e-01 -7.88500965e-01 5.15997171e-01
-5.48596680e-01 -3.07066083e-01 3.06857705e-01 3.73475701e-01
-1.00250460e-01 3.61701220e-01 -1.01509619e+00 -4.11578178e-01
-6.15628660e-01 -1.18656504e+00 1.17380285e+00 7.84771740e-01
4.71357182e-02 -1.21804571e+00 3.58724922e-01 7.83727229e-01
5.08654475e-01 5.33756495e-01 1.19987249e+00 -7.50282526e-01
-6.82508051e-01 -6.90399945e-01 -4.95412558e-01 -5.06475335e-03
1.75873831e-01 -4.40596230e-02 -7.61117816e-01 -4.58436519e-01
-5.21459162e-01 -2.12282881e-01 9.81737018e-01 4.97743040e-01
1.28756273e+00 -9.79311094e-02 -5.39069831e-01 5.97606838e-01
1.99007905e+00 2.68060714e-01 7.08919644e-01 4.49547559e-01
7.17151761e-01 1.74506828e-01 2.74607450e-01 1.11110872e-02
6.17982268e-01 1.38610169e-01 7.53405631e-01 -2.84482211e-01
-4.96229120e-02 -9.61841643e-02 -1.94204792e-01 6.43883467e-01
-4.30082321e-01 1.85046680e-02 -1.30860651e+00 6.86199009e-01
-1.59398532e+00 -7.62914300e-01 -3.05227727e-01 1.98060036e+00
6.70119226e-01 2.02089310e-01 -3.03278387e-01 -1.32976770e-01
5.76303005e-01 -1.09119497e-01 -5.35451829e-01 -3.09429884e-01
3.80473703e-01 2.25472614e-01 7.36321509e-01 3.71415943e-01
-1.23878169e+00 3.13141972e-01 4.74348068e+00 3.55370462e-01
-1.14799142e+00 2.88115107e-02 1.25902510e+00 1.87757626e-01
-1.44440964e-01 -8.46526772e-02 -6.17749155e-01 4.91253942e-01
7.63549447e-01 -3.12805414e-01 7.20164180e-02 8.68583798e-01
6.98530152e-02 -1.75067201e-01 -9.00327504e-01 6.09662712e-01
1.83290944e-01 -1.31322455e+00 -1.16800897e-01 3.76759052e-01
5.55099785e-01 2.54251361e-02 -1.37566805e-01 4.25312161e-01
2.51250684e-01 -1.21665537e+00 -2.98953921e-01 6.77255511e-01
5.83848119e-01 -6.82989597e-01 1.51648569e+00 -7.18336254e-02
-1.12557662e+00 1.14907049e-01 -2.56005555e-01 5.40772319e-01
-4.25797641e-01 7.16913402e-01 -1.21949661e+00 6.31841302e-01
6.21461868e-01 6.13163769e-01 -1.15752602e+00 1.31258845e+00
-1.82131484e-01 6.65217817e-01 1.05541805e-02 -2.65897185e-01
2.67426163e-01 -1.09693795e-01 -2.25168373e-03 1.13806498e+00
4.04604465e-01 -8.97660330e-02 3.45015258e-01 6.19302452e-01
-4.59908508e-02 4.04640399e-02 -1.56712547e-01 -2.31595993e-01
5.94173037e-02 1.86799085e+00 -1.20338106e+00 1.52941048e-02
-2.50158995e-01 6.50353134e-01 3.03546548e-01 1.44840688e-01
-6.56188369e-01 -5.32707155e-01 -2.27982521e-01 6.57793581e-02
-3.47368009e-02 2.20831856e-01 -3.74309510e-01 -6.79929018e-01
-1.04929261e-01 -3.51677239e-01 6.85931027e-01 -4.49294239e-01
-1.46643698e+00 4.50369775e-01 -4.14511740e-01 -9.64249372e-01
4.83893231e-02 -6.56741917e-01 -7.90579796e-01 7.69418955e-01
-1.64360869e+00 -1.37118256e+00 -9.30770993e-01 1.09769575e-01
2.75227100e-01 3.00712883e-02 1.12677097e+00 2.88217664e-01
-6.89590693e-01 3.37217033e-01 -4.48826887e-02 5.44914722e-01
7.96564102e-01 -1.89107013e+00 -2.34458193e-01 4.79930550e-01
-6.60061002e-01 5.43732941e-01 3.74985695e-01 -6.70424998e-01
-1.56793654e+00 -1.40771484e+00 9.72275019e-01 9.72137898e-02
6.96699083e-01 -4.05083597e-02 -6.18576944e-01 8.30651581e-01
3.87149192e-02 4.19219911e-01 1.16162872e+00 -3.54415551e-02
2.46098042e-02 -3.96819293e-01 -1.39511192e+00 2.23483071e-01
3.15790474e-01 9.12381895e-03 -4.02378477e-02 3.48485917e-01
-2.47062996e-01 -4.51237798e-01 -1.25866771e+00 3.69998872e-01
5.19779265e-01 -7.79523969e-01 6.10735655e-01 -1.24973021e-01
8.22014630e-01 -4.42268342e-01 7.69656524e-02 -1.38555193e+00
-7.40996122e-01 3.24501872e-01 3.54643673e-01 1.06658340e+00
4.13441539e-01 -5.46565056e-01 1.10712838e+00 6.35264695e-01
-4.11262214e-01 -1.10526490e+00 -9.32033658e-01 -3.17033559e-01
-5.68635538e-02 4.02703494e-01 1.35278642e-01 6.70212209e-01
-1.02266997e-01 4.98002730e-02 4.05939013e-01 3.16651613e-01
1.11425018e+00 2.95564026e-01 5.25247157e-01 -1.26276374e+00
-4.37031597e-01 -5.20224512e-01 -1.11767828e+00 -1.85742714e-02
-1.90710291e-01 -1.49237275e+00 -7.02357218e-02 -1.95141757e+00
8.95675540e-01 -3.34965795e-01 -6.52737260e-01 6.05743647e-01
-3.22898090e-01 5.86943269e-01 -4.71157804e-02 2.81878620e-01
-7.54902899e-01 1.25275224e-01 1.39349985e+00 -7.01225460e-01
1.66931927e-01 -9.36644077e-02 -7.19431639e-01 4.87515688e-01
1.04022276e+00 -3.49906951e-01 -8.59145150e-02 -5.39653823e-02
5.40963113e-02 1.05762705e-01 5.65034926e-01 -1.05161595e+00
2.85900354e-01 -6.67053536e-02 9.21263039e-01 -5.37028193e-01
1.04869999e-01 -6.97785974e-01 3.14088434e-01 1.07338798e+00
-2.56319314e-01 -1.33621007e-01 7.22668767e-02 7.15164065e-01
-2.61650383e-01 -2.35077992e-01 6.32133484e-01 -2.96884626e-01
-7.29642689e-01 4.97222811e-01 -9.87946838e-02 -5.62654972e-01
1.18917668e+00 -2.41069030e-02 -7.14346647e-01 1.31718561e-01
-7.13492513e-01 4.11977798e-01 3.67747813e-01 -1.17479987e-01
5.75939894e-01 -1.43921912e+00 -8.90989423e-01 2.27039233e-02
4.17828202e-01 -1.44140236e-02 5.49642980e-01 1.07196152e+00
-9.03344870e-01 4.18907762e-01 -3.73830765e-01 -8.45754445e-01
-1.45033526e+00 4.49421406e-02 4.09455001e-01 -1.16363072e+00
-3.57699245e-01 6.97816730e-01 1.01887830e-01 -4.17255342e-01
1.06095619e-01 -2.11774334e-01 -4.28418398e-01 -3.41527194e-01
3.90221179e-01 9.02935266e-02 2.73628265e-01 -2.19569132e-01
-5.61632872e-01 3.49093795e-01 -5.46912491e-01 4.42436099e-01
1.54661870e+00 4.33671325e-01 -4.20321167e-01 2.28957087e-01
1.08476973e+00 -1.18026160e-01 -8.08525562e-01 2.43907750e-01
-3.19320932e-02 -4.84691739e-01 2.36258388e-01 -9.66442287e-01
-1.29807901e+00 5.49066722e-01 1.02047145e+00 2.11630285e-01
1.12232137e+00 -8.55338573e-02 6.70436203e-01 1.25167057e-01
3.73775810e-01 -5.89953244e-01 -9.74980891e-02 -2.48100329e-02
6.13227248e-01 -1.43727767e+00 1.23937614e-01 -3.88149500e-01
-3.92167121e-01 1.35201347e+00 5.72444975e-01 -5.62544405e-01
5.23984849e-01 1.25565380e-01 2.03320459e-01 -4.63606507e-01
-5.33944786e-01 2.50506818e-01 1.27824366e-01 3.38176012e-01
7.64822900e-01 5.31430662e-01 -6.30800605e-01 8.82026076e-01
-1.09885767e-01 2.33068883e-01 4.26266819e-01 8.52843881e-01
-2.89636761e-01 -1.06103063e+00 -1.09020896e-01 9.46121812e-01
-5.30026257e-01 2.66353488e-01 -6.66745305e-01 8.55559826e-01
-1.46156952e-01 6.17476642e-01 -4.34937030e-02 -2.15303421e-01
4.29142229e-02 -1.65331826e-01 4.67664152e-01 -4.53916401e-01
-6.95098281e-01 -4.05189954e-02 -5.06736673e-02 -3.09089839e-01
-2.32116908e-01 -4.96709287e-01 -1.39809084e+00 -2.05587566e-01
6.31364575e-03 1.31495833e-01 7.93378115e-01 6.31714761e-01
2.57279307e-01 9.19620097e-01 5.91089666e-01 -7.21856833e-01
-2.76214868e-01 -1.02706325e+00 -6.83908165e-01 5.53558528e-01
-2.31469236e-02 -3.43238384e-01 -5.13913572e-01 -4.93335128e-02] | [14.963743209838867, -2.776460647583008] |
da269988-947e-49a7-bf82-9ec5d3760aaa | deep-learning-based-inverse-method-for-layout | 1806.03182 | null | http://arxiv.org/abs/1806.03182v1 | http://arxiv.org/pdf/1806.03182v1.pdf | Deep learning based inverse method for layout design | Layout design with complex constraints is a challenging problem to solve due
to the non-uniqueness of the solution and the difficulties in incorporating the
constraints into the conventional optimization-based methods. In this paper, we
propose a design method based on the recently developed machine learning
technique, Variational Autoencoder (VAE). We utilize the learning capability of
the VAE to learn the constraints and the generative capability of the VAE to
generate design candidates that automatically satisfy all the constraints. As
such, no constraints need to be imposed during the design stage. In addition,
we show that the VAE network is also capable of learning the underlying physics
of the design problem, leading to an efficient design tool that does not need
any physical simulation once the network is constructed. We demonstrated the
performance of the method on two cases: inverse design of surface diffusion
induced morphology change and mask design for optical microlithography. | ['Yu-Jie Zhang', 'Wenjing Ye'] | 2018-06-07 | null | null | null | null | ['layout-design'] | ['computer-vision'] | [ 6.45188242e-02 1.87924922e-01 3.80765080e-01 -1.10600203e-01
1.99221715e-01 -3.97259772e-01 2.80388355e-01 -3.49406213e-01
-1.12565085e-01 8.91790509e-01 -2.34339356e-01 -2.98902780e-01
-5.12513340e-01 -8.47867906e-01 -7.59013772e-01 -8.48641992e-01
3.75957459e-01 5.14068484e-01 -2.60783315e-01 -2.26827756e-01
3.22797924e-01 9.36470389e-01 -1.72673678e+00 -2.50059307e-01
1.02472222e+00 9.39982176e-01 5.13229728e-01 4.53469336e-01
-7.70430043e-02 2.92064667e-01 -3.49415779e-01 3.22419265e-03
1.70091689e-01 -4.40943062e-01 -1.07745044e-01 3.52384388e-01
2.28195578e-01 -1.99476331e-01 -2.03832418e-01 8.65018547e-01
5.29007554e-01 4.41892743e-01 1.09111631e+00 -8.34375143e-01
-6.66558504e-01 2.49691665e-01 -2.07080156e-01 -3.21764261e-01
5.23244105e-02 7.73145854e-02 1.04650664e+00 -1.06525505e+00
9.02110636e-01 8.17715764e-01 5.95578909e-01 7.93306410e-01
-1.34125268e+00 -4.15301085e-01 -1.12273060e-01 1.41745452e-02
-1.46855927e+00 -2.47409001e-01 1.26802874e+00 -7.10373580e-01
8.77571642e-01 9.73116159e-02 8.10338140e-01 8.44854236e-01
3.72231960e-01 3.65464449e-01 8.98672819e-01 -6.48641348e-01
6.70507729e-01 3.09864968e-01 -2.33450651e-01 1.11355686e+00
3.44277740e-01 3.91771644e-01 -7.03461692e-02 2.89694462e-02
1.02042723e+00 -2.88926244e-01 -2.92319864e-01 -6.33830249e-01
-6.56462193e-01 9.86553729e-01 4.58732486e-01 4.07504052e-01
-4.02836621e-01 1.56662449e-01 -3.06562901e-01 -5.23935333e-02
1.62245318e-01 1.28865075e+00 -2.30638340e-01 4.29238588e-01
-9.86625671e-01 2.43677184e-01 8.65646780e-01 6.58675611e-01
7.78343618e-01 4.01731700e-01 2.13436589e-01 6.81740582e-01
6.07046187e-01 4.40507472e-01 1.70460239e-01 -9.68837500e-01
-7.39425048e-02 5.90494096e-01 2.01547921e-01 -1.10957849e+00
-3.16027761e-01 -5.52573502e-01 -6.55029893e-01 5.63109279e-01
7.55321980e-02 -5.70224226e-01 -1.12772238e+00 1.65260959e+00
4.47721690e-01 -9.35056284e-02 5.33632189e-03 7.79261291e-01
6.74637973e-01 7.80323982e-01 -6.72936022e-01 -3.13025147e-01
5.92598200e-01 -5.44453382e-01 -9.81911123e-01 1.08368322e-01
2.18952894e-01 -6.26717269e-01 7.34651029e-01 3.05878043e-01
-1.14728558e+00 -3.66288155e-01 -1.39986324e+00 2.49530256e-01
-3.30076754e-01 2.65949696e-01 7.53147900e-01 4.34898704e-01
-9.11384583e-01 8.94195199e-01 -7.61964083e-01 -1.34818643e-01
1.56201422e-01 7.54830182e-01 6.50964826e-02 1.95226863e-01
-7.99860239e-01 7.99493074e-01 8.21103007e-02 4.45200443e-01
-6.43060744e-01 -7.73864508e-01 -7.12924123e-01 1.76273882e-01
3.39403570e-01 -7.79098690e-01 9.80099738e-01 -7.27454126e-01
-2.15065813e+00 2.23141432e-01 -1.24310739e-01 1.66645214e-01
4.30554122e-01 2.51323014e-01 -1.19678527e-01 6.21659495e-03
-5.06195843e-01 5.03656626e-01 9.84659314e-01 -1.60069025e+00
-8.68820399e-02 1.36332780e-01 -9.24719498e-02 -1.43797770e-01
-4.01010811e-01 -6.73134625e-01 -2.20413804e-01 -4.94949639e-01
1.34922326e-01 -1.01106393e+00 -4.68844146e-01 1.55287702e-02
-4.85321224e-01 -1.04543559e-01 8.43354940e-01 -4.58381057e-01
1.12037385e+00 -1.98302150e+00 6.47151172e-01 7.85379708e-01
1.07330516e-01 1.52920589e-01 -1.32480294e-01 6.98029816e-01
1.60306796e-01 1.41322077e-03 -3.61585826e-01 -1.99507445e-01
-2.88080852e-02 4.29355234e-01 -6.27621040e-02 2.82436013e-01
3.16621184e-01 8.89623702e-01 -4.44594234e-01 -2.19588995e-01
4.69715208e-01 5.05188107e-01 -1.08590007e+00 4.01656568e-01
-6.97681606e-01 5.81795394e-01 -6.44493222e-01 2.64554650e-01
5.13886929e-01 -2.99871147e-01 3.22743833e-01 -3.89796734e-01
-4.15146649e-01 -2.88817167e-01 -1.30226457e+00 1.31412268e+00
-6.88535988e-01 5.82345963e-01 3.99128795e-01 -1.01749313e+00
1.04847121e+00 1.94437191e-01 4.89328057e-01 -4.49296474e-01
4.43597406e-01 4.72979724e-01 1.67089283e-01 -6.44562125e-01
1.92565426e-01 -1.76492080e-01 2.17227235e-01 3.46852034e-01
-3.15726921e-02 -4.90645379e-01 8.37443545e-02 -2.39279181e-01
8.59411359e-01 7.99115747e-02 -2.02119872e-01 -4.32104498e-01
6.18269324e-01 -1.62161261e-01 7.14640319e-01 6.73079193e-01
6.68951333e-01 5.06926656e-01 1.22221552e-01 -3.89697105e-01
-1.40785587e+00 -8.02020669e-01 -9.43539217e-02 3.26117501e-02
-4.30399040e-03 -7.72913545e-02 -9.63945866e-01 -2.51556963e-01
6.36134967e-02 7.09353209e-01 -5.61622679e-01 -6.01449981e-02
-7.31179297e-01 -5.16090274e-01 -4.47535999e-02 2.45709017e-01
1.56879440e-01 -9.46231663e-01 -5.66942096e-01 3.28834146e-01
3.40679586e-01 -8.51280749e-01 -2.08801374e-01 -4.47192863e-02
-7.68608630e-01 -9.83867586e-01 -4.66190666e-01 -9.77950513e-01
1.07987642e+00 -3.40714723e-01 6.35536015e-01 2.87818074e-01
-7.23726034e-01 4.53662455e-01 2.17564821e-01 -2.46557102e-01
-5.73054433e-01 1.43030484e-03 1.13730133e-01 8.79849941e-02
-3.49397182e-01 -9.07896280e-01 -5.16436219e-01 2.86398232e-01
-8.26375842e-01 3.55237722e-01 6.89929724e-01 1.03263366e+00
6.72884941e-01 6.05409086e-01 7.42352068e-01 -4.50266212e-01
5.79521537e-01 -1.46777719e-01 -1.09345603e+00 1.98824748e-01
-7.08625615e-01 4.36068714e-01 7.70052791e-01 -6.20812595e-01
-1.22327900e+00 3.67406517e-01 -2.48326808e-01 -5.72849214e-01
-3.50769870e-02 6.92898154e-01 -3.56341392e-01 -4.82485205e-01
2.16359198e-01 -9.40887034e-02 1.64027125e-01 -5.09646833e-01
2.46037126e-01 3.58937711e-01 1.65787339e-01 -5.72150290e-01
1.01242518e+00 3.11697483e-01 4.55291927e-01 -1.17717624e+00
-2.80175149e-01 2.93343425e-01 -3.29885513e-01 -4.01144236e-01
9.42174971e-01 -2.52641201e-01 -1.00597298e+00 2.81970710e-01
-1.07078385e+00 -2.89220691e-01 -2.94735730e-01 4.04713899e-01
-6.35860562e-01 1.30643353e-01 -3.95008266e-01 -9.65096951e-01
-1.96164921e-01 -1.22987878e+00 5.30296564e-01 3.45787555e-01
-1.60447404e-01 -1.13150430e+00 -5.81066012e-02 6.74146935e-02
5.15200615e-01 6.20766521e-01 1.45795369e+00 1.44400388e-01
-9.49231148e-01 -1.28728300e-01 1.15399256e-01 3.32225919e-01
2.73459673e-01 4.39699084e-01 -7.00673938e-01 -2.98851937e-01
2.50600398e-01 1.59787878e-01 6.65958762e-01 7.65076935e-01
1.10374939e+00 -2.57235557e-01 -3.61843288e-01 6.67312622e-01
1.80238509e+00 4.67853874e-01 3.92983258e-01 -2.03797743e-01
8.15623462e-01 7.63661444e-01 -6.43667951e-02 4.85923380e-01
-1.01269469e-01 7.44941056e-01 3.33543360e-01 -1.24067791e-01
-1.73872523e-02 -3.31498593e-01 3.50035131e-02 9.63423669e-01
1.56879760e-02 -3.36548895e-01 -7.67968893e-01 3.18985283e-01
-1.92931819e+00 -6.04289412e-01 -5.21521689e-03 1.92012382e+00
5.39981782e-01 -7.46061727e-02 -4.49241817e-01 -1.27317786e-01
7.70148218e-01 -2.70518333e-01 -6.66352153e-01 -6.57089472e-01
7.56670088e-02 5.09335935e-01 2.95906097e-01 7.75800049e-01
-6.50260091e-01 8.40196729e-01 6.67198277e+00 6.21502697e-01
-1.22485709e+00 -3.49673092e-01 1.04724906e-01 9.73909814e-03
-6.90351546e-01 3.26841213e-02 -6.43132329e-01 2.79400349e-01
5.50732613e-01 1.14237830e-01 7.73892701e-01 5.93822420e-01
3.61462265e-01 -9.70538035e-02 -1.11164546e+00 7.60749459e-01
-2.32791211e-02 -1.68444932e+00 -1.39240608e-01 4.03124243e-01
1.08263719e+00 -5.84617555e-01 2.45415345e-01 -1.26705378e-01
-1.60007458e-02 -1.05468357e+00 6.17736399e-01 7.41485178e-01
4.58860517e-01 -8.41790795e-01 3.97863686e-01 2.31103763e-01
-7.62654960e-01 -2.58349299e-01 -2.39663184e-01 -5.61432168e-02
3.66564572e-01 8.59368324e-01 -8.09220433e-01 3.43956798e-01
1.20460339e-01 4.84386474e-01 -8.03915411e-02 1.04627478e+00
-1.70179546e-01 2.56088793e-01 -5.67425430e-01 -4.93328363e-01
1.66511670e-01 -4.27979231e-01 7.47062445e-01 6.21827006e-01
6.29113793e-01 -7.86547922e-03 1.20342851e-01 1.54763556e+00
9.87823680e-03 -7.08644763e-02 -7.11420536e-01 -3.20280164e-01
2.50570595e-01 1.12245178e+00 -6.40413940e-01 3.32777292e-01
-6.41042814e-02 5.56402743e-01 1.51125401e-01 6.37508035e-01
-7.31294394e-01 -3.63289058e-01 5.05712152e-01 1.05697021e-01
7.98616529e-01 -5.33353508e-01 -4.24319029e-01 -9.69519973e-01
1.01196088e-01 -5.57452917e-01 -3.66352171e-01 -6.85078502e-01
-1.30451798e+00 3.48563343e-01 -1.15158513e-01 -8.47770095e-01
-2.57333249e-01 -9.13174450e-01 -6.06950045e-01 7.88279891e-01
-1.41370332e+00 -8.91350150e-01 -2.93714330e-02 2.37718880e-01
1.28748029e-01 -2.79127955e-01 7.36831009e-01 2.46269330e-01
-7.61927783e-01 2.12074921e-01 2.83527851e-01 -2.96141356e-01
1.91366881e-01 -1.24158847e+00 -2.75956262e-02 6.30540133e-01
-2.26719916e-01 6.47633255e-01 9.74724650e-01 -8.04083586e-01
-1.84084308e+00 -6.32128596e-01 6.08363748e-01 -1.79768473e-01
5.31530619e-01 -5.04977405e-01 -8.19283128e-01 1.30367696e-01
2.34064952e-01 -3.61298084e-01 6.28829300e-01 -3.34459618e-02
3.33750933e-01 6.82718828e-02 -1.14066577e+00 6.52464449e-01
8.66366386e-01 -2.50917256e-01 -4.58499581e-01 1.56973153e-01
5.77712059e-01 -3.75235885e-01 -7.92685091e-01 5.98754942e-01
6.91356003e-01 -6.43603325e-01 6.88818097e-01 -3.65563571e-01
5.18808484e-01 -3.65887493e-01 -8.36099312e-02 -1.24990487e+00
-4.32962626e-01 -6.83321714e-01 -2.28968680e-01 1.23793864e+00
8.07936370e-01 -5.61324179e-01 9.81158435e-01 8.86964262e-01
-3.11066598e-01 -9.67001319e-01 -7.87626147e-01 -7.58389473e-01
-3.16903740e-02 -5.05253524e-02 5.87259829e-01 7.27706552e-01
-5.37992537e-01 2.11444214e-01 -3.46789896e-01 3.06342006e-01
5.71020663e-01 3.78088534e-01 3.60730141e-01 -1.46090221e+00
-8.40252876e-01 -3.49128544e-01 5.86067736e-02 -8.67628396e-01
3.88340086e-01 -7.45173454e-01 2.42242977e-01 -1.54596269e+00
-3.30167264e-01 -6.93705618e-01 4.58031241e-03 6.34026825e-02
2.20922619e-01 -3.35432380e-01 -1.75934583e-02 -1.39056742e-01
1.44717664e-01 1.01469052e+00 1.57008195e+00 -1.20944232e-01
-7.00579166e-01 -2.43232753e-02 -4.13737923e-01 5.38068950e-01
7.29271948e-01 -3.25741231e-01 -5.63844979e-01 -4.46682662e-01
5.38354993e-01 -5.78116141e-02 9.27123576e-02 -1.09599447e+00
2.25668117e-01 -6.14504777e-02 4.88955855e-01 -4.55090404e-01
3.33619446e-01 -9.80124235e-01 3.96578461e-01 2.72461444e-01
-1.69011503e-01 -1.73740327e-01 3.05157125e-01 5.64703345e-01
-1.43310230e-03 -5.76226354e-01 8.44712317e-01 7.39531815e-02
-3.86463791e-01 1.01292364e-01 -6.68190181e-01 -4.11589205e-01
7.74651647e-01 -1.76978886e-01 1.43520117e-01 -1.56967327e-01
-7.50746846e-01 1.75289467e-01 5.81130147e-01 1.17293864e-01
8.59614730e-01 -1.29290617e+00 -1.21898897e-01 5.53856134e-01
-3.52816105e-01 4.20693820e-03 2.57864416e-01 3.65525812e-01
-7.02283919e-01 2.05245569e-01 -1.10270217e-01 -4.55904007e-01
-7.15537071e-01 5.44612586e-01 4.71471757e-01 -1.24987043e-01
-6.02996588e-01 6.98480189e-01 1.59087162e-02 -4.06090558e-01
4.63199653e-02 -1.79455817e-01 -7.34368414e-02 -1.28751338e-01
1.20556010e-02 1.09152988e-01 -7.59727135e-02 -2.10974380e-01
-1.30588844e-01 9.24913704e-01 2.01782495e-01 -3.10922951e-01
1.58816922e+00 8.88257772e-02 -1.96956843e-01 1.36045069e-01
1.04322028e+00 4.38446626e-02 -1.29067039e+00 9.00625885e-02
-1.20651349e-01 -1.68385312e-01 6.04457915e-01 -5.59439838e-01
-1.24226809e+00 7.52234340e-01 6.48269713e-01 -4.69160713e-02
9.57581401e-01 -3.37646514e-01 8.21796179e-01 5.52177608e-01
1.57986328e-01 -1.50069261e+00 1.66085929e-01 4.38674241e-01
8.78994167e-01 -8.52228403e-01 -7.99651667e-02 -4.72132534e-01
-3.77808288e-02 1.29302096e+00 5.54532349e-01 -2.41543069e-01
8.95725489e-01 2.98514485e-01 -2.12890670e-01 -3.64210665e-01
-4.70607460e-01 2.36981474e-02 5.78562737e-01 3.13802183e-01
1.09029815e-01 -1.66248158e-01 -2.61633307e-01 2.52749085e-01
-8.92376155e-02 -5.80128990e-02 5.30302465e-01 9.50994551e-01
-3.27732354e-01 -1.43717885e+00 -1.00753613e-01 2.40062460e-01
5.80072515e-02 1.40879139e-01 -4.42960560e-01 7.13369668e-01
2.24765867e-01 8.48071992e-01 -1.47743061e-01 -3.57305408e-01
3.64783227e-01 4.92538027e-02 7.15735614e-01 -5.94777346e-01
-2.84660131e-01 2.56259412e-01 1.51199894e-03 -3.22482795e-01
-1.34688586e-01 -4.81837451e-01 -1.31564915e+00 -5.51088527e-02
-6.61145866e-01 2.34475031e-01 9.35059965e-01 8.79883468e-01
5.35036922e-01 7.22338259e-01 7.45710611e-01 -9.89183068e-01
-1.72546640e-01 -3.48589748e-01 -5.38085341e-01 4.76300083e-02
3.99091125e-01 -1.13364851e+00 -4.06063408e-01 -3.38911086e-01] | [5.9222517013549805, 3.4329161643981934] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.