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
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
428c5b22-467d-4332-8f80-62a5a89f1cd2 | are-large-language-models-ready-for | 2304.05368 | null | https://arxiv.org/abs/2304.05368v2 | https://arxiv.org/pdf/2304.05368v2.pdf | Are Large Language Models Ready for Healthcare? A Comparative Study on Clinical Language Understanding | Large language models (LLMs) have made significant progress in various domains, including healthcare. However, the specialized nature of clinical language understanding tasks presents unique challenges and limitations that warrant further investigation. In this study, we conduct a comprehensive evaluation of state-of-the-art LLMs, namely GPT-3.5, GPT-4, and Bard, within the realm of clinical language understanding tasks. These tasks span a diverse range, including named entity recognition, relation extraction, natural language inference, semantic textual similarity, document classification, and question-answering. We also introduce a novel prompting strategy, self-questioning prompting (SQP), tailored to enhance LLMs' performance by eliciting informative questions and answers pertinent to the clinical scenarios at hand. Our evaluation underscores the significance of task-specific learning strategies and prompting techniques for improving LLMs' effectiveness in healthcare-related tasks. Additionally, our in-depth error analysis on the challenging relation extraction task offers valuable insights into error distribution and potential avenues for improvement using SQP. Our study sheds light on the practical implications of employing LLMs in the specialized domain of healthcare, serving as a foundation for future research and the development of potential applications in healthcare settings. | ['Linda Petzold', 'Yun Zhao', 'Yuqing Wang'] | 2023-04-09 | null | null | null | null | ['document-classification', 'semantic-textual-similarity'] | ['natural-language-processing', 'natural-language-processing'] | [ 6.47246480e-01 5.03604650e-01 -4.46644515e-01 -5.43689370e-01
-1.29903424e+00 -2.04548076e-01 2.05555990e-01 1.12761986e+00
-5.07543385e-01 6.29123092e-01 6.80384099e-01 -8.04895043e-01
-7.24803627e-01 -3.28302264e-01 -2.95125097e-01 -1.96639732e-01
-2.15421259e-01 7.53094494e-01 -1.10769697e-01 -1.54139802e-01
2.59002000e-01 3.50147754e-01 -8.48770797e-01 8.72881353e-01
1.14956200e+00 8.78824532e-01 7.52096772e-02 7.14441895e-01
-2.88713694e-01 1.47048831e+00 -6.34102881e-01 -4.56133246e-01
-3.74344736e-01 -4.60029125e-01 -1.41371858e+00 -1.45115882e-01
8.87542665e-02 -3.59659307e-02 -1.26782969e-01 4.75654453e-01
7.89350510e-01 7.72305951e-02 4.92594838e-01 -9.50798512e-01
-5.26576161e-01 5.71218848e-01 1.99476257e-01 6.45310163e-01
8.34944367e-01 7.13682100e-02 9.72789705e-01 -5.39273620e-01
7.13254452e-01 9.11598384e-01 9.06223118e-01 7.47616827e-01
-9.72976327e-01 -5.66072106e-01 3.60474214e-02 3.79894614e-01
-1.11254704e+00 -5.22507548e-01 2.89667815e-01 -3.75383377e-01
1.54763091e+00 4.03741509e-01 1.36235654e-01 1.10280025e+00
5.19736886e-01 1.01595962e+00 1.00331223e+00 -5.31005085e-01
1.81031480e-01 2.19087094e-01 4.07702029e-01 6.70942008e-01
2.38767073e-01 -6.13277480e-02 -6.95333064e-01 -6.79302871e-01
3.17748904e-01 -3.06551624e-02 -3.09280276e-01 9.71300900e-02
-1.22061801e+00 8.84999633e-01 1.82693943e-01 3.11200172e-01
-4.70227540e-01 -4.99552995e-01 7.01146722e-01 1.87314644e-01
4.76455867e-01 1.08319402e+00 -1.01651633e+00 -2.28435501e-01
-8.13333809e-01 1.98086858e-01 1.32390833e+00 9.89469886e-01
4.45355549e-02 -6.07105672e-01 -6.24573708e-01 8.66169453e-01
1.54149592e-01 1.68394938e-01 5.30732989e-01 -6.44948661e-01
8.01188886e-01 7.45505095e-01 -1.78261101e-01 -8.66201758e-01
-8.78711343e-01 -3.97656500e-01 -6.90909684e-01 -8.52264524e-01
1.12882949e-01 -2.36864299e-01 -6.68237090e-01 1.45872438e+00
2.77656287e-01 2.02407122e-01 4.49199528e-01 4.42709655e-01
1.41112924e+00 1.87703386e-01 7.41757035e-01 -3.50502372e-01
1.87279797e+00 -1.00128782e+00 -1.13060236e+00 -5.70596874e-01
1.35639262e+00 -9.14166391e-01 7.43114471e-01 8.40580910e-02
-8.32176805e-01 -7.58642256e-02 -4.40625668e-01 -1.41387478e-01
-2.28715599e-01 -5.68401488e-03 6.73136652e-01 2.35165223e-01
-7.28997588e-01 4.40543503e-01 -1.01600170e+00 -6.90928280e-01
5.80469728e-01 2.74230540e-01 -3.85448813e-01 -2.48016924e-01
-1.42856252e+00 1.34909093e+00 3.58216584e-01 -1.16521709e-01
-4.75965858e-01 -1.20781064e+00 -1.20131302e+00 -1.46674095e-02
5.78744531e-01 -1.09164298e+00 1.69805562e+00 -8.22196230e-02
-1.05372941e+00 1.04081261e+00 -3.51201266e-01 -6.41598165e-01
1.26567423e-01 -2.97259897e-01 -6.23491228e-01 2.99740255e-01
2.93491215e-01 3.00492853e-01 -6.95350673e-03 -7.07339764e-01
-6.14224076e-01 -3.00319850e-01 -2.27243707e-01 2.25689456e-01
-1.42273799e-01 2.84607291e-01 -6.82085082e-02 -7.09328473e-01
-1.50967836e-01 -5.93225002e-01 -6.10421538e-01 -3.51580113e-01
-6.01672590e-01 -5.38331866e-01 7.56759271e-02 -8.75899494e-01
1.81535685e+00 -1.90875721e+00 -2.43412465e-01 3.62535156e-02
4.90114391e-01 3.82236421e-01 -1.40078381e-01 1.00145233e+00
-2.31926993e-01 1.03077382e-01 -2.68822700e-01 -9.73067209e-02
-3.17919284e-01 2.93860763e-01 -8.61055627e-02 -7.54436851e-02
4.99928236e-01 1.24577034e+00 -1.21113575e+00 -9.50101435e-01
-6.77402765e-02 1.24983445e-01 -5.06934702e-01 5.47552705e-01
-1.74091011e-01 5.56189120e-01 -7.13426650e-01 6.80790424e-01
1.68555919e-02 -8.71917248e-01 3.33355367e-01 -2.06228286e-01
3.12911570e-01 8.69069040e-01 -6.15888536e-01 1.57270217e+00
-5.35763860e-01 2.05968425e-01 -1.57108083e-01 -1.15353882e+00
6.55205965e-01 5.68773270e-01 7.29384720e-01 -5.99001288e-01
-5.73558733e-03 2.26534277e-01 1.19837195e-01 -1.38300300e+00
6.96682185e-02 -3.63085538e-01 5.00823371e-02 4.15531099e-01
4.93509620e-02 1.09225633e-02 1.48228601e-01 3.17429215e-01
1.40244579e+00 -3.61091852e-01 1.09490073e+00 -8.70000571e-02
4.84415263e-01 3.15813184e-01 3.96401197e-01 9.27712977e-01
-3.54575783e-01 2.13523805e-01 3.74720603e-01 -3.22358161e-01
-4.99291986e-01 -6.13001585e-01 -5.65896332e-01 9.04352188e-01
-2.90549695e-01 -7.86601365e-01 -4.99622285e-01 -9.80159044e-01
1.13632180e-01 7.89326012e-01 -5.67283750e-01 -3.98004562e-01
-5.36336482e-01 -8.51804316e-01 6.49607062e-01 8.06396604e-01
4.80278023e-02 -1.31854272e+00 -6.53753221e-01 5.02376318e-01
-6.20431960e-01 -1.64984369e+00 -6.20473325e-01 3.15454662e-01
-1.12181246e+00 -1.47248125e+00 -4.28910524e-01 -9.64349449e-01
5.94665885e-01 -1.56553179e-01 1.44892454e+00 -1.17665939e-01
-4.26880658e-01 7.42393076e-01 -4.90357399e-01 -5.56601465e-01
-7.28384674e-01 3.83149087e-01 -2.41251618e-01 -4.75349873e-01
8.57084811e-01 -3.62859154e-03 -6.55340791e-01 8.36912766e-02
-8.81425202e-01 4.03798334e-02 8.63994896e-01 9.96612072e-01
5.15881121e-01 -2.60983378e-01 8.29400539e-01 -1.57116628e+00
1.19513178e+00 -7.28694618e-01 3.17164332e-01 5.95987678e-01
-9.17672813e-01 1.40584871e-01 4.49001223e-01 -2.49964267e-01
-1.02143526e+00 -4.72730041e-01 -5.88007092e-01 2.94533402e-01
-4.10124302e-01 9.23058510e-01 2.49576300e-01 6.80285320e-02
9.23127294e-01 2.02209771e-01 1.30769283e-01 -4.81909961e-01
4.22550924e-02 9.45669711e-01 2.59113342e-01 -4.42745447e-01
1.28220692e-01 1.82894357e-02 -1.51551813e-01 -5.44707179e-01
-1.40400338e+00 -9.59883571e-01 -2.05746800e-01 3.78454953e-01
8.62595022e-01 -7.49361098e-01 -8.45984221e-01 -6.83289096e-02
-1.05786681e+00 -2.66340792e-01 -3.43271285e-01 4.74269003e-01
-3.28036904e-01 4.86660630e-01 -9.52912509e-01 -6.33479714e-01
-9.71318722e-01 -1.04472148e+00 1.27159035e+00 1.45178139e-01
-1.06840789e+00 -1.39958298e+00 1.10282578e-01 8.76679182e-01
2.61540800e-01 1.28594192e-03 1.30923498e+00 -1.46672535e+00
-1.00862913e-01 -1.86808601e-01 -5.93257472e-02 1.60385743e-01
3.86775762e-01 -6.80219829e-01 -4.42911834e-01 -8.01469237e-02
9.97921079e-02 -4.53861564e-01 5.34460783e-01 2.05576852e-01
1.28874695e+00 -5.51963747e-01 -7.41068780e-01 3.76720846e-01
9.04391468e-01 3.66357356e-01 1.87150061e-01 4.18387622e-01
4.50138837e-01 7.27517545e-01 7.62082219e-01 4.25237566e-01
9.00157869e-01 2.72688538e-01 -1.48492262e-01 -2.74315983e-01
4.23142239e-02 -2.48907939e-01 -3.14388454e-01 8.21199179e-01
3.76023710e-01 6.62604570e-02 -1.29543829e+00 4.85261559e-01
-1.69977212e+00 -5.08433700e-01 1.12462431e-01 1.76637006e+00
1.49533403e+00 -5.04748821e-02 -2.94032961e-01 -6.87377120e-04
1.36571079e-01 -1.16360418e-01 -5.26230872e-01 -4.14476961e-01
1.83758065e-01 4.95117784e-01 1.13695830e-01 4.13639635e-01
-1.01581502e+00 6.40861452e-01 6.60164738e+00 7.20564306e-01
-6.71932042e-01 6.63814321e-02 7.09885895e-01 3.26321751e-01
-1.38896480e-01 -2.48721406e-01 -9.18640435e-01 2.09825397e-01
1.03345370e+00 -1.71022445e-01 -4.64379564e-02 7.25533903e-01
2.99343824e-01 -1.31875306e-01 -1.48203743e+00 7.71665573e-01
1.77088156e-01 -1.48910725e+00 4.18518148e-02 -3.49594623e-01
5.10440290e-01 -1.27435997e-01 -2.95358866e-01 3.34917843e-01
1.29874259e-01 -1.03615034e+00 -1.85460255e-01 5.30924022e-01
7.04873025e-01 -2.13456705e-01 1.15551519e+00 3.51558447e-01
-7.74554431e-01 -4.88910526e-02 1.74059868e-01 5.21386303e-02
2.85930872e-01 5.30538797e-01 -1.66994274e+00 8.16186666e-01
5.16921282e-01 7.28080869e-01 -4.35494155e-01 9.80603099e-01
-1.65427431e-01 7.15619564e-01 -2.00243481e-02 -8.81068259e-02
4.28826362e-02 3.40257615e-01 2.68984646e-01 1.75284779e+00
-9.28778350e-02 5.96123815e-01 4.22034830e-01 2.96170890e-01
-2.83177774e-02 4.71074432e-01 -3.94039005e-01 -4.39333260e-01
5.83104253e-01 1.08349824e+00 -4.63290840e-01 -5.79904974e-01
-3.32640648e-01 4.70800072e-01 4.13330376e-01 3.10873836e-01
-4.14092600e-01 -3.27813417e-01 4.89353687e-01 1.08771004e-01
-2.93509990e-01 2.83631861e-01 -6.56005561e-01 -1.01902163e+00
-8.23758617e-02 -1.40381384e+00 1.08089542e+00 -3.06539088e-01
-1.64427471e+00 7.67493010e-01 -4.73501952e-03 -9.87431645e-01
-3.15621853e-01 -5.42030394e-01 -2.39100829e-01 6.11332357e-01
-1.68073821e+00 -9.15919423e-01 -2.40665987e-01 4.61908221e-01
6.82166934e-01 -5.01779132e-02 1.31919086e+00 4.14736748e-01
-5.67237377e-01 7.84366846e-01 -2.76383132e-01 1.65539205e-01
9.80935276e-01 -9.57177401e-01 1.44466445e-01 2.41408765e-01
-1.22678138e-01 9.09290195e-01 5.28732777e-01 -7.71362185e-01
-1.16001105e+00 -1.07228386e+00 1.73102641e+00 -7.26793468e-01
6.78849161e-01 1.29376464e-02 -1.17084110e+00 7.24672079e-01
-3.46117616e-02 -2.60985553e-01 1.55317342e+00 3.40056360e-01
-6.69192076e-02 9.11420360e-02 -1.13689685e+00 4.62283075e-01
1.03253973e+00 -6.91097379e-01 -9.85432446e-01 8.19770873e-01
9.22739387e-01 -7.57458627e-01 -1.37960672e+00 7.93307662e-01
2.09797308e-01 -3.40451479e-01 1.04693723e+00 -1.38184023e+00
6.64732754e-01 3.09141278e-01 2.11339369e-01 -1.09676290e+00
-2.57777780e-01 -6.03541791e-01 -2.95207143e-01 8.81623447e-01
6.34057462e-01 -7.16967762e-01 5.10393441e-01 9.16580498e-01
-2.58311629e-01 -1.55523491e+00 -6.92627847e-01 -2.29189351e-01
-4.66379337e-02 -4.26847398e-01 3.20755869e-01 1.11621594e+00
6.22885287e-01 5.81993163e-01 7.53013790e-02 1.68483078e-01
2.33491495e-01 2.21599359e-02 2.11567640e-01 -9.88890350e-01
-2.89711058e-01 -2.76159048e-01 1.35440249e-02 -9.14266050e-01
1.26563281e-01 -1.00203598e+00 -1.42007722e-02 -1.98438334e+00
4.17757869e-01 -4.78204876e-01 -2.62849092e-01 6.47257209e-01
-7.55479872e-01 -3.29897016e-01 -1.49278015e-01 8.67922679e-02
-8.52902651e-01 2.14614898e-01 1.30553734e+00 -4.75234911e-02
-1.77437037e-01 2.51927018e-01 -1.04359055e+00 4.79369879e-01
6.55418217e-01 -6.74969971e-01 -3.52596581e-01 -3.01457256e-01
5.33275679e-02 4.27726775e-01 -7.57767782e-02 -3.80473107e-01
5.74453831e-01 -2.04383016e-01 -1.36515154e-02 -3.76664758e-01
-1.59726933e-01 -6.37837708e-01 -2.75458008e-01 7.58956134e-01
-8.65631402e-01 1.03922047e-01 4.26291794e-01 4.17704940e-01
-3.05939019e-01 -1.77802652e-01 3.43071938e-01 -2.02685729e-01
-4.73198920e-01 2.09825903e-01 -5.33529699e-01 7.47181535e-01
8.79950404e-01 1.49035126e-01 -2.79267222e-01 -2.00196028e-01
-9.33442354e-01 7.15973258e-01 -3.82377595e-01 4.11630720e-01
7.49990165e-01 -8.61588061e-01 -8.49601090e-01 -3.80178052e-03
6.46356463e-01 1.52486145e-01 4.42920417e-01 1.18371129e+00
-3.37651581e-01 1.01414585e+00 3.96580845e-01 -4.73411769e-01
-1.40963745e+00 5.92847824e-01 2.94141442e-01 -1.03040147e+00
-6.92771375e-01 8.74714136e-01 2.10298300e-01 -5.57709754e-01
4.42719877e-01 -4.46403503e-01 -4.14374083e-01 -1.02265805e-01
5.81598997e-01 1.57257229e-01 4.54354376e-01 -9.79612768e-03
-7.24758387e-01 1.50935784e-01 -5.45771301e-01 4.20689821e-01
1.24148536e+00 1.27926450e-02 -1.79685459e-01 9.42917764e-02
9.93228495e-01 -3.86279583e-01 -3.70269954e-01 -7.00724900e-01
7.40630686e-01 9.13759321e-02 -2.73277551e-01 -1.38774395e+00
-4.46302295e-01 6.33871913e-01 1.16320588e-01 -7.97977895e-02
1.25535500e+00 4.49870378e-01 9.79398191e-01 5.59778869e-01
9.28244963e-02 -8.31849277e-01 2.26297811e-01 6.08386576e-01
7.85322964e-01 -1.46517694e+00 -7.89861903e-02 -5.99307060e-01
-8.72334063e-01 8.89655113e-01 5.98485589e-01 5.47531068e-01
6.48529828e-01 5.31324983e-01 3.46138030e-01 -5.00342309e-01
-9.51020837e-01 -8.47382471e-02 5.81124485e-01 4.46721196e-01
9.11155641e-01 1.45876613e-02 -5.07831275e-01 9.10096705e-01
-7.76982307e-02 2.76704907e-01 -3.03933080e-02 1.20266914e+00
-6.49063438e-02 -1.20931304e+00 -2.07894444e-01 9.04037178e-01
-7.66871274e-01 -6.47586703e-01 -2.92590082e-01 5.09493709e-01
-1.74743444e-01 1.20814967e+00 -4.62978452e-01 -1.62305385e-01
7.28227794e-01 4.09946948e-01 3.58306825e-01 -1.09405696e+00
-8.54257464e-01 -1.81923479e-01 5.55163264e-01 -6.50983393e-01
-2.77833343e-01 -5.04307747e-01 -1.45427990e+00 1.77203342e-01
-2.75194138e-01 4.23328847e-01 1.50545672e-01 1.24023795e+00
8.05994391e-01 9.65239882e-01 1.51513936e-03 2.54962593e-01
-8.01495075e-01 -1.09353507e+00 3.65129188e-02 5.22267997e-01
4.40928489e-01 -3.14094603e-01 2.56212931e-02 1.69079634e-03] | [8.69277572631836, 8.642789840698242] |
41f69380-554b-4456-bf51-169810612f5a | illumination-variation-correction-using-image | 2301.09702 | null | https://arxiv.org/abs/2301.09702v1 | https://arxiv.org/pdf/2301.09702v1.pdf | Illumination Variation Correction Using Image Synthesis For Unsupervised Domain Adaptive Person Re-Identification | Unsupervised domain adaptive (UDA) person re-identification (re-ID) aims to learn identity information from labeled images in source domains and apply it to unlabeled images in a target domain. One major issue with many unsupervised re-identification methods is that they do not perform well relative to large domain variations such as illumination, viewpoint, and occlusions. In this paper, we propose a Synthesis Model Bank (SMB) to deal with illumination variation in unsupervised person re-ID. The proposed SMB consists of several convolutional neural networks (CNN) for feature extraction and Mahalanobis matrices for distance metrics. They are trained using synthetic data with different illumination conditions such that their synergistic effect makes the SMB robust against illumination variation. To better quantify the illumination intensity and improve the quality of synthetic images, we introduce a new 3D virtual-human dataset for GAN-based image synthesis. From our experiments, the proposed SMB outperforms other synthesis methods on several re-ID benchmarks. | ['Edward J. Delp', 'Amy R. Reibman', 'Jiaqi Guo'] | 2023-01-23 | null | null | null | null | ['person-re-identification'] | ['computer-vision'] | [ 2.60250717e-01 -4.85894948e-01 6.82726875e-02 -4.93345767e-01
-3.74175549e-01 -5.76884449e-01 7.78699338e-01 -3.04335654e-01
-4.46492732e-01 7.88932204e-01 3.14695001e-01 4.62509125e-01
2.04021335e-01 -6.62172854e-01 -4.32837039e-01 -6.30757391e-01
5.72636008e-01 5.61815560e-01 -2.53577292e-01 -1.41010970e-01
6.56364784e-02 6.08876050e-01 -1.54420054e+00 -1.42204538e-01
8.66079092e-01 6.09645486e-01 -2.56764144e-01 3.54059309e-01
2.35123456e-01 3.75019819e-01 -9.39679027e-01 -6.66190088e-01
7.01036632e-01 -7.31363595e-01 -5.93932748e-01 3.45568806e-01
6.72369838e-01 -4.54066575e-01 -4.33697790e-01 1.25262821e+00
8.96905899e-01 3.36566925e-01 8.99900496e-01 -1.36886501e+00
-9.31296229e-01 3.15219492e-01 -4.56445187e-01 -5.58592267e-02
3.05791885e-01 2.23428190e-01 4.24387842e-01 -8.26796710e-01
7.02480257e-01 1.43783009e+00 7.29296267e-01 9.67771113e-01
-1.45297885e+00 -8.48342836e-01 -2.73122817e-01 2.57642329e-01
-1.62703633e+00 -4.53149498e-01 9.97623563e-01 -4.73282695e-01
5.26824474e-01 6.01360053e-02 3.22997570e-01 1.43147910e+00
-4.30896580e-01 4.32677925e-01 1.33883929e+00 -5.41397393e-01
2.13064432e-01 2.83057988e-01 -2.27057636e-02 3.24062735e-01
4.82316494e-01 2.98534483e-01 -5.38393736e-01 1.91961944e-01
8.61229479e-01 -1.39209032e-01 -1.91360563e-02 -4.42297310e-01
-1.40795958e+00 6.23248696e-01 3.56796801e-01 8.87611285e-02
-2.29634997e-02 -1.35141701e-01 4.86468613e-01 3.62678558e-01
2.13033959e-01 5.35117388e-01 2.94813588e-02 1.96460187e-01
-6.80606127e-01 4.53405231e-01 3.25127333e-01 9.24073339e-01
7.61807084e-01 2.13755205e-01 -4.46361631e-01 1.26899791e+00
-7.48417899e-02 7.80919373e-01 7.62269437e-01 -8.54740798e-01
2.52845854e-01 6.04169130e-01 2.68743962e-01 -1.09671068e+00
-3.39063942e-01 -4.57544535e-01 -1.17660987e+00 1.65010974e-01
7.10093200e-01 -1.60842970e-01 -9.22669828e-01 1.89523518e+00
2.08890781e-01 2.94379383e-01 2.40832910e-01 1.12606764e+00
1.05469477e+00 1.85183480e-01 2.15932354e-02 1.02667026e-01
1.14695859e+00 -8.86650801e-01 -4.91273850e-01 -2.42073372e-01
2.96524584e-01 -5.95931828e-01 8.33198786e-01 1.56942248e-01
-7.06443071e-01 -1.12378359e+00 -1.11476445e+00 6.69780299e-02
-3.67131054e-01 4.94943291e-01 -9.37151089e-02 1.09021950e+00
-8.51871669e-01 3.25659305e-01 -1.81213230e-01 -7.95107543e-01
3.38184625e-01 3.33137095e-01 -5.72243869e-01 -1.00293465e-01
-1.24875188e+00 6.96781218e-01 4.10467833e-01 -1.28169358e-01
-9.83153760e-01 -5.01604497e-01 -8.70561600e-01 -1.64110243e-01
4.84789256e-04 -6.67097092e-01 7.67351270e-01 -1.19415736e+00
-1.62923181e+00 1.30172920e+00 -9.45512950e-02 -4.80572104e-01
6.46085858e-01 -1.38272628e-01 -6.85038865e-01 -1.77395254e-01
2.63785124e-01 6.81155264e-01 9.85675395e-01 -1.45767963e+00
-2.30100915e-01 -6.27859116e-01 -2.64642507e-01 2.42263556e-01
-3.70982170e-01 4.98271324e-02 -4.20872003e-01 -1.03098857e+00
-2.97692388e-01 -1.11472392e+00 -7.94810802e-03 -3.57077718e-01
-6.49419665e-01 -1.10077579e-02 6.40182972e-01 -7.27582753e-01
5.97636163e-01 -2.03721690e+00 2.14269653e-01 2.94248253e-01
3.55346613e-02 4.29786533e-01 -2.54285723e-01 1.21605381e-01
-1.71130404e-01 -2.52282917e-01 -1.82791010e-01 -5.10428011e-01
6.90018758e-02 -3.73684987e-02 7.65685141e-02 5.08331120e-01
-4.25239615e-02 9.38275635e-01 -6.83509469e-01 -3.60134482e-01
4.83430505e-01 4.36724812e-01 -2.18702301e-01 3.78057241e-01
1.85336471e-01 1.04271555e+00 1.04083531e-01 5.76322973e-01
8.01727772e-01 -5.62280603e-03 1.21475033e-01 -4.48281139e-01
3.15809876e-01 -2.36848295e-01 -1.37124300e+00 1.74391532e+00
-1.66238457e-01 6.87925339e-01 -4.09293383e-01 -1.00532651e+00
1.27029848e+00 3.01087834e-02 4.11210537e-01 -1.01553476e+00
3.99114043e-01 1.15536563e-01 -8.90463218e-02 -1.70813784e-01
3.39295983e-01 1.17893413e-01 1.29810888e-02 4.51746494e-01
1.98901251e-01 3.47217560e-01 1.57321915e-01 -9.53675434e-03
5.02107799e-01 -3.12570147e-02 1.27775460e-01 -2.67668992e-01
9.48830664e-01 -2.81117648e-01 7.67260373e-01 7.22031236e-01
-4.41160709e-01 8.49302053e-01 -1.26988456e-01 -5.67473710e-01
-1.53242052e+00 -1.13032091e+00 4.70553450e-02 8.06872427e-01
5.39535463e-01 5.53949326e-02 -1.03496981e+00 -6.44671619e-01
2.62117796e-02 6.09247684e-01 -6.00965500e-01 -2.09617004e-01
-4.44583982e-01 -7.94290841e-01 8.72234166e-01 5.15689552e-01
1.23207879e+00 -8.17831039e-01 -5.62247597e-02 -2.55819112e-02
-3.94510478e-01 -1.40342438e+00 -7.97542095e-01 -5.62753916e-01
-3.11689347e-01 -1.09180129e+00 -1.33136368e+00 -9.19188738e-01
8.58626783e-01 2.53993630e-01 8.88981581e-01 -3.35418165e-01
-2.94663846e-01 5.21307886e-01 -3.81162077e-01 -2.75518864e-01
-5.65424681e-01 -5.11939339e-02 6.17375135e-01 6.28008306e-01
6.46409154e-01 -4.53372180e-01 -6.85355961e-01 8.12223494e-01
-6.32661402e-01 1.38663635e-01 1.63661763e-01 1.10443389e+00
4.72803086e-01 1.24262109e-01 5.64620972e-01 -7.99442112e-01
7.39347339e-01 -5.09374449e-03 -4.88602579e-01 4.40512836e-01
-6.42403662e-01 2.33044382e-02 5.51834047e-01 -5.89681864e-01
-1.24882030e+00 1.45815268e-01 1.71700716e-01 -4.03852910e-01
-4.81552601e-01 -1.19907260e-01 -6.05432510e-01 -2.08101258e-01
9.80299234e-01 3.95306826e-01 -5.64714782e-02 -5.20966887e-01
3.64361733e-01 7.95054793e-01 1.02148080e+00 -4.98855114e-01
1.21910417e+00 5.54461002e-01 -1.28032759e-01 -7.30105340e-01
-5.16648591e-01 -3.05511326e-01 -9.52740669e-01 -2.60999531e-01
9.07144129e-01 -1.09947133e+00 -6.07490838e-01 1.11465251e+00
-1.01489985e+00 -2.97874540e-01 -2.40443602e-01 4.25694406e-01
-3.63100529e-01 4.98764873e-01 -3.09958577e-01 -6.14279032e-01
-3.97227824e-01 -1.06656432e+00 7.97643960e-01 5.76895297e-01
-2.80088693e-01 -8.30356717e-01 1.69124648e-01 5.76478124e-01
4.08538133e-01 3.88904154e-01 4.90313739e-01 -6.83925569e-01
-7.91671351e-02 -1.83198795e-01 -3.71076494e-01 6.12623811e-01
5.13433993e-01 -4.79407579e-01 -1.06170857e+00 -5.52746058e-01
-2.79258072e-01 -2.22998902e-01 6.65564179e-01 1.85715929e-01
1.03308690e+00 -2.76183248e-01 -1.44653916e-01 9.58176136e-01
1.22968388e+00 1.81747258e-01 6.95603490e-01 5.36958277e-01
1.08490467e+00 6.92775905e-01 4.72804666e-01 4.66333032e-01
3.02490115e-01 9.84013259e-01 1.02558481e-02 -3.17342043e-01
-4.79548514e-01 -3.30394506e-01 2.61915177e-01 4.05935735e-01
-3.32804412e-01 -2.20706224e-01 -6.83694839e-01 6.09721541e-01
-1.63507283e+00 -1.00081933e+00 1.64531246e-01 2.37299633e+00
5.15251696e-01 -2.42092788e-01 5.91258705e-01 2.29184434e-01
1.17512321e+00 -1.99487045e-01 -8.64433348e-01 -4.94735613e-02
-6.00200236e-01 5.88091798e-02 7.51862168e-01 1.34134516e-01
-1.17198503e+00 8.75430048e-01 5.55589104e+00 6.57555461e-01
-1.00855696e+00 5.17589264e-02 6.45778537e-01 9.97750089e-02
1.63708068e-02 -3.83967400e-01 -6.96633101e-01 6.02989912e-01
5.43875635e-01 -2.34335989e-01 6.17041171e-01 7.46398628e-01
-3.90793290e-03 2.90852547e-01 -1.06812978e+00 1.73237765e+00
4.36003029e-01 -1.05240595e+00 1.08229704e-01 -9.73129049e-02
1.12659693e+00 -2.74993509e-01 2.83033907e-01 -4.68620397e-02
5.72755337e-01 -1.18946195e+00 4.71002340e-01 3.20372105e-01
1.17996204e+00 -8.20298791e-01 8.29809785e-01 -5.15960269e-02
-9.82425272e-01 5.94574818e-03 -3.92119408e-01 3.12052697e-01
-6.69118017e-02 1.89393908e-01 -8.51073265e-01 5.15570879e-01
7.55936205e-01 7.82685459e-01 -7.98792064e-01 7.40134478e-01
-1.22251354e-01 2.49951318e-01 -6.59316487e-04 2.86691695e-01
-3.76620591e-01 -2.82321960e-01 6.07526481e-01 1.07988822e+00
3.04130733e-01 -2.33761564e-01 -2.04457968e-01 1.04295051e+00
-2.44570032e-01 1.52653113e-01 -4.54169750e-01 7.75279058e-03
5.66117942e-01 1.00356603e+00 -4.29472119e-01 -3.25803131e-01
-2.51733035e-01 1.54617715e+00 2.66978592e-02 6.12111449e-01
-7.12095618e-01 -2.46581197e-01 8.97794843e-01 1.44841820e-01
-3.12637985e-02 -1.92865983e-01 -2.45119020e-01 -1.37813807e+00
-1.81299418e-01 -1.16069973e+00 4.64045525e-01 -6.35580897e-01
-1.65203333e+00 5.47067046e-01 -8.21748897e-02 -1.38166654e+00
-2.12326437e-01 -4.92501915e-01 -2.98750371e-01 9.40955043e-01
-1.28068793e+00 -1.50511527e+00 -8.74696076e-01 9.59865391e-01
5.17594516e-01 -8.93797159e-01 7.66577899e-01 4.61240292e-01
-7.47643650e-01 1.26725519e+00 2.99310327e-01 6.41888738e-01
1.23436439e+00 -1.12085152e+00 7.12869406e-01 1.10425794e+00
1.77199155e-01 5.60657918e-01 5.19354403e-01 -5.80201030e-01
-1.06248307e+00 -1.31106603e+00 4.39005762e-01 -3.98182303e-01
-9.20158476e-02 -3.97712022e-01 -6.00745797e-01 5.77506661e-01
-5.35680316e-02 -1.30907387e-01 6.67305052e-01 -1.71599492e-01
-6.40173733e-01 -4.41680610e-01 -1.48297763e+00 5.37943184e-01
1.23357105e+00 -5.86517513e-01 -2.40546212e-01 1.65318742e-01
3.19353193e-01 -2.16948763e-01 -8.50209713e-01 3.93333822e-01
6.08972847e-01 -9.86810327e-01 1.25877559e+00 -4.27037865e-01
5.73658124e-02 -4.89446878e-01 -9.21290442e-02 -1.50532436e+00
-3.57439548e-01 -5.13454854e-01 3.45109582e-01 1.67909467e+00
-1.12753764e-01 -8.37062657e-01 8.40843260e-01 6.31562591e-01
4.19344515e-01 2.88239807e-01 -6.85876548e-01 -1.07470000e+00
-1.17481481e-02 1.26228705e-01 1.03225803e+00 1.14967906e+00
-4.59925950e-01 2.53030181e-01 -8.32275033e-01 1.46786675e-01
1.19959295e+00 -1.31176025e-01 1.39889121e+00 -1.30136728e+00
-6.78228065e-02 -1.53374165e-01 -6.07125044e-01 -7.17004359e-01
1.65634736e-01 -8.73240948e-01 -2.73566157e-01 -1.16907549e+00
3.07170987e-01 -3.78557026e-01 -3.27254832e-01 1.47908166e-01
-1.06571332e-01 6.46081448e-01 1.93746090e-01 4.56066132e-01
-3.07249546e-01 5.78397989e-01 1.15739226e+00 -5.80414057e-01
-2.54525185e-01 -1.70837909e-01 -5.81370115e-01 4.52130169e-01
8.73093367e-01 -2.40019396e-01 -3.47287923e-01 -4.90837842e-01
-3.65761459e-01 -3.27569991e-01 7.20039904e-01 -1.40692663e+00
1.69662818e-01 -5.15001733e-03 8.37962687e-01 -3.24218512e-01
1.62661508e-01 -5.72234333e-01 4.87487316e-01 3.10675859e-01
-4.84532297e-01 -1.28914937e-01 -3.49432267e-02 3.72275114e-01
-1.57046363e-01 1.07775265e-02 1.18996465e+00 -1.26811057e-01
-8.90372217e-01 3.38678926e-01 1.41852498e-01 4.06565703e-02
9.70022023e-01 -5.23895204e-01 -1.12185851e-01 -4.64203507e-01
-4.70238656e-01 -9.43409279e-02 9.04067993e-01 5.88181257e-01
5.60914993e-01 -1.67161334e+00 -1.03126216e+00 4.35680091e-01
4.57295954e-01 -1.74157500e-01 4.38944817e-01 2.88699389e-01
-4.55738038e-01 2.45776787e-01 -7.69691527e-01 -4.56380814e-01
-1.39627635e+00 6.18574083e-01 6.07590616e-01 -9.11395438e-03
-3.49479795e-01 7.78079152e-01 3.53097498e-01 -6.95225954e-01
1.41092747e-01 4.58860427e-01 -3.25394064e-01 -1.90529868e-01
5.54039717e-01 6.06655240e-01 -1.66074440e-01 -1.16478491e+00
-3.44003886e-01 7.88505912e-01 -3.75552326e-02 -1.44111574e-01
9.78796840e-01 -3.59341472e-01 1.67517588e-01 1.34821795e-02
1.10945404e+00 -1.41802996e-01 -1.13584149e+00 -6.29880607e-01
-3.16950530e-01 -6.67427897e-01 -4.97168392e-01 -7.83066332e-01
-1.08045459e+00 6.93389714e-01 1.22276950e+00 -3.48456413e-01
1.19818485e+00 -4.34711188e-01 9.29291368e-01 1.94444522e-01
3.34477365e-01 -1.47341609e+00 3.81354868e-01 1.19250596e-01
7.02687144e-01 -1.48657441e+00 -5.37319407e-02 -2.70923346e-01
-8.33426714e-01 9.23306525e-01 7.53055274e-01 -9.09464285e-02
2.20610574e-01 -1.93950102e-01 2.41818011e-01 1.75719813e-01
2.92499930e-01 -4.20284808e-01 3.34783107e-01 1.30257058e+00
-1.38043398e-02 1.31784648e-01 -2.87367497e-02 5.49043894e-01
-3.63189727e-01 -7.44128451e-02 2.09290028e-01 1.94230631e-01
1.21117570e-01 -1.31920648e+00 -6.68056309e-01 7.81044317e-03
-1.63525656e-01 4.02504057e-01 -7.07534850e-01 6.53505266e-01
3.83199096e-01 1.02898920e+00 1.88022386e-02 -7.13814735e-01
2.81888306e-01 -1.75799742e-01 6.16556048e-01 -2.94709474e-01
-3.90352428e-01 -2.29994491e-01 -3.62322778e-02 -2.73536682e-01
-6.78511620e-01 -6.48663938e-01 -7.65290022e-01 -4.15164769e-01
-6.84586763e-02 -1.78393573e-01 4.98086125e-01 6.96311951e-01
2.02149838e-01 2.71142125e-01 7.81892419e-01 -6.26628697e-01
-3.41134429e-01 -9.79523003e-01 -7.04307497e-01 1.20628774e+00
3.28825824e-02 -8.55265081e-01 -1.59043461e-01 3.88186306e-01] | [14.702515602111816, 1.0070621967315674] |
50313e54-a773-48b6-b9c2-d0fdc02d1cb2 | et5-a-novel-end-to-end-framework-for | 2209.11484 | null | https://arxiv.org/abs/2209.11484v1 | https://arxiv.org/pdf/2209.11484v1.pdf | ET5: A Novel End-to-end Framework for Conversational Machine Reading Comprehension | Conversational machine reading comprehension (CMRC) aims to assist computers to understand an natural language text and thereafter engage in a multi-turn conversation to answer questions related to the text. Existing methods typically require three steps: (1) decision making based on entailment reasoning; (2) span extraction if required by the above decision; (3) question rephrasing based on the extracted span. However, for nearly all these methods, the span extraction and question rephrasing steps cannot fully exploit the fine-grained entailment reasoning information in decision making step because of their relative independence, which will further enlarge the information gap between decision making and question phrasing. Thus, to tackle this problem, we propose a novel end-to-end framework for conversational machine reading comprehension based on shared parameter mechanism, called entailment reasoning T5 (ET5). Despite the lightweight of our proposed framework, experimental results show that the proposed ET5 achieves new state-of-the-art results on the ShARC leaderboard with the BLEU-4 score of 55.2. Our model and code are publicly available at https://github.com/Yottaxx/ET5. | ['Xian-Ling Mao', 'Zewen Chi', 'Heyan Huang', 'Xiao Zhang'] | 2022-09-23 | null | https://aclanthology.org/2022.coling-1.47 | https://aclanthology.org/2022.coling-1.47.pdf | coling-2022-10 | ['machine-reading-comprehension'] | ['natural-language-processing'] | [ 4.87467736e-01 3.58576298e-01 1.68382570e-01 -4.93027806e-01
-1.13949203e+00 -6.17380202e-01 5.53519189e-01 4.42537338e-01
-3.45084250e-01 6.43000484e-01 8.04528892e-01 -8.45833242e-01
-9.68416966e-03 -6.16251290e-01 -4.50160265e-01 -3.05568278e-01
3.82540524e-01 4.83372271e-01 2.28614002e-01 -4.38757062e-01
5.53224146e-01 -4.27008480e-01 -1.45003390e+00 7.56917953e-01
1.39604914e+00 8.73374999e-01 3.67414713e-01 1.06078160e+00
-4.62715805e-01 1.37169266e+00 -4.00722355e-01 -4.47554082e-01
-1.80665806e-01 -9.57402170e-01 -1.65484893e+00 -8.12169760e-02
5.49713336e-02 -4.65699732e-01 6.59732297e-02 6.65218174e-01
3.77238482e-01 3.90614033e-01 4.75793958e-01 -9.50644195e-01
-3.24615151e-01 8.28051746e-01 -2.95595795e-01 2.23995373e-01
9.53906834e-01 9.61416736e-02 1.32453942e+00 -7.27012157e-01
1.90660104e-01 1.27638316e+00 3.47612590e-01 5.54419994e-01
-5.41351497e-01 -2.23961487e-01 3.69255811e-01 7.57174253e-01
-8.98863852e-01 -5.11913240e-01 8.18473637e-01 -6.54678047e-02
9.31525290e-01 5.21695852e-01 3.18259865e-01 8.53626013e-01
-4.84909341e-02 1.01744258e+00 1.32006407e+00 -5.99044979e-01
2.29192495e-01 -2.62412369e-01 6.83337510e-01 5.01389682e-01
-3.75311792e-01 -5.94960511e-01 -5.26259780e-01 -1.57876741e-02
-7.16583729e-02 -2.99342811e-01 -5.70116341e-01 4.30963933e-01
-1.12342596e+00 8.47598732e-01 1.89601943e-01 4.47262451e-02
-3.88599753e-01 -1.69971257e-01 5.34755170e-01 6.92192197e-01
2.99268395e-01 2.75646091e-01 -3.31873566e-01 -6.08255327e-01
-6.04370713e-01 4.38249916e-01 1.34142673e+00 9.86958921e-01
5.87257624e-01 -8.29847336e-01 -3.76589656e-01 8.21139157e-01
2.35703081e-01 2.78278649e-01 4.36054260e-01 -9.15371537e-01
1.03549814e+00 7.36202478e-01 1.94494560e-01 -7.60926545e-01
-4.40803528e-01 -2.00511664e-01 -8.47960532e-01 -5.20816028e-01
4.25544947e-01 -2.03494996e-01 -3.58015299e-01 1.55232286e+00
5.43187976e-01 -1.59202099e-01 4.03243899e-01 9.00437057e-01
1.12012291e+00 8.06096256e-01 -4.20720540e-02 -2.66398609e-01
1.91946054e+00 -1.35391629e+00 -9.09389436e-01 -2.63397515e-01
6.92874730e-01 -1.03085554e+00 1.35698485e+00 2.38487899e-01
-1.11005175e+00 -3.51126313e-01 -9.51719820e-01 -6.40343606e-01
1.14659257e-01 -9.14050788e-02 2.54822880e-01 3.92284095e-01
-7.10246563e-01 1.08473822e-01 -4.13547635e-01 -3.07772875e-01
4.92938012e-02 -2.57453322e-01 1.23422839e-01 -4.71907139e-01
-1.44630516e+00 8.92723382e-01 2.96828091e-01 3.65090996e-01
-6.15420699e-01 -3.96312296e-01 -6.35312080e-01 1.40840024e-01
7.78275967e-01 -7.68254220e-01 1.86091876e+00 -6.14466965e-01
-1.83437598e+00 5.82843661e-01 -6.14424169e-01 -4.64375883e-01
6.55287266e-01 -6.19406998e-01 9.33147296e-02 2.83632010e-01
-3.04960664e-02 2.79826343e-01 3.90232056e-01 -8.02419603e-01
-7.30957627e-01 -2.63082564e-01 5.43673873e-01 7.81530619e-01
2.38928258e-01 1.22660268e-02 -3.88615817e-01 -2.04198152e-01
2.16749758e-01 -7.81635046e-01 1.73046559e-01 -5.43191969e-01
-5.72080970e-01 -7.97979832e-01 4.05297190e-01 -1.19850767e+00
1.53033495e+00 -1.68165040e+00 2.92539835e-01 -2.80928254e-01
2.87009656e-01 1.16914041e-01 -3.13198656e-01 1.01630008e+00
3.39677095e-01 -7.22196884e-03 -3.72515053e-01 -3.92477304e-01
1.88116714e-01 -2.27605272e-02 -3.09348494e-01 1.04057357e-01
-4.82293442e-02 8.29341412e-01 -9.15030837e-01 -5.31082749e-01
1.16300106e-01 -1.65654063e-01 -4.47772115e-01 7.52496064e-01
-6.06453478e-01 4.19542909e-01 -5.28180540e-01 3.60309750e-01
7.75920272e-01 -2.75587380e-01 1.21156074e-01 1.97544992e-01
1.11898426e-02 1.01027596e+00 -8.38860095e-01 1.82725060e+00
-6.46985590e-01 3.86863053e-01 1.23349033e-01 -8.56712878e-01
6.10427678e-01 4.27309662e-01 -2.16968268e-01 -6.52064204e-01
1.60781875e-01 5.71296923e-02 1.32552952e-01 -7.76022077e-01
4.51041579e-01 -4.52692211e-02 -1.88442841e-01 9.01137948e-01
-3.69876146e-01 -5.65170981e-02 1.71647981e-01 5.48146129e-01
1.16702020e+00 -8.28265697e-02 4.24540848e-01 -9.12426338e-02
1.00106096e+00 -8.63927081e-02 3.84595275e-01 7.26876795e-01
-3.81276876e-01 3.41391742e-01 7.83145607e-01 -1.60631076e-01
-6.65288091e-01 -7.44868577e-01 2.62860119e-01 1.28066218e+00
9.43006426e-02 -5.52447379e-01 -1.07504380e+00 -7.18960524e-01
-4.58245784e-01 1.11205399e+00 -4.12714571e-01 1.09184749e-01
-6.07976615e-01 -2.07398206e-01 6.59873843e-01 2.57838815e-01
9.92378056e-01 -1.03576231e+00 -6.27793372e-01 2.28487536e-01
-1.19279838e+00 -1.24393260e+00 -6.31057322e-01 -1.95244193e-01
-6.57495141e-01 -1.33308268e+00 -4.46778715e-01 -7.21944034e-01
3.36075455e-01 5.30710995e-01 1.19693351e+00 4.60546762e-01
3.28876406e-01 3.16020936e-01 -9.97236490e-01 -3.11525583e-01
-5.42882681e-01 4.09918219e-01 -4.93234754e-01 -1.11446649e-01
5.07854223e-01 -5.27854860e-01 -9.25617933e-01 3.40299278e-01
-6.22411728e-01 7.25856602e-01 5.51500857e-01 6.84416175e-01
1.04731597e-01 -1.68775603e-01 8.57181370e-01 -9.11600292e-01
1.11775625e+00 -5.96052289e-01 -1.45244673e-01 6.15340769e-01
-4.32743013e-01 1.25801966e-01 6.65208399e-01 8.03740881e-03
-1.56154740e+00 -5.66651642e-01 -4.43731010e-01 4.12208408e-01
-1.24641024e-01 7.73648381e-01 -2.47208610e-01 6.73063695e-01
3.76940429e-01 4.99368459e-01 2.61838026e-02 -3.90233874e-01
5.25084317e-01 1.06425428e+00 5.37083626e-01 -7.39347994e-01
4.01290148e-01 -1.07320152e-01 -4.10771072e-01 -6.49193466e-01
-1.23212171e+00 -6.35473073e-01 -3.77663374e-01 -2.92922497e-01
1.02797174e+00 -9.52800333e-01 -1.07426155e+00 6.06603503e-01
-1.45719564e+00 -2.19278991e-01 2.85376072e-01 2.29119852e-01
-4.88632500e-01 7.03261018e-01 -7.52512395e-01 -1.07427800e+00
-9.05004978e-01 -9.49709773e-01 8.02393615e-01 3.33977461e-01
-5.01332879e-01 -8.78501773e-01 -6.34777099e-02 1.31016421e+00
4.24131542e-01 3.98697294e-02 1.23450685e+00 -7.44507849e-01
-5.77932835e-01 4.14823517e-02 -2.34252319e-01 4.08177644e-01
3.93521823e-02 -5.17517567e-01 -8.84503782e-01 -8.29465613e-02
3.01995605e-01 -5.95927119e-01 6.65867388e-01 -1.69045463e-01
9.96431410e-01 -6.01825953e-01 3.20731401e-01 9.26031731e-03
1.00360286e+00 -4.85876426e-02 7.54204333e-01 1.75225794e-01
5.23903370e-01 9.71852303e-01 7.17710912e-01 3.26584309e-01
1.12167692e+00 3.40126514e-01 2.72518069e-01 4.50137943e-01
-4.60238010e-02 -5.48175693e-01 3.13069135e-01 1.44619465e+00
6.94878399e-02 -4.17936802e-01 -9.67593729e-01 5.15004575e-01
-1.99126768e+00 -8.03115904e-01 -3.58288944e-01 1.84612942e+00
1.12089431e+00 9.34132636e-02 -1.60768151e-01 1.47841185e-01
6.36279047e-01 3.25506210e-01 -5.70473075e-01 -7.26736844e-01
1.62549362e-01 1.04049677e-02 -1.65997103e-01 8.22394133e-01
-5.37161589e-01 8.05352271e-01 4.69715023e+00 8.23832572e-01
-4.66486633e-01 1.01549372e-01 4.86926407e-01 1.59119278e-01
-4.66979235e-01 2.42569864e-01 -6.08877897e-01 4.30697590e-01
8.27539384e-01 -2.59161234e-01 5.54130197e-01 3.64275277e-01
3.03193539e-01 -5.64043820e-01 -1.18488598e+00 7.03745902e-01
1.34003013e-01 -9.53843296e-01 7.88515806e-02 -5.30240774e-01
3.45280021e-01 -3.91100824e-01 -4.43349570e-01 6.73054099e-01
1.43058196e-01 -9.79669929e-01 4.82747614e-01 4.09167290e-01
3.16630840e-01 -7.74017990e-01 9.32331979e-01 9.23848748e-01
-9.93029952e-01 4.76039723e-02 -1.73426256e-01 -5.18849134e-01
2.20546171e-01 5.00604749e-01 -8.90726864e-01 7.85576463e-01
3.06298077e-01 3.00761700e-01 -3.42742115e-01 5.13833165e-01
-7.73258030e-01 9.97777700e-01 -1.07816614e-01 -4.65728819e-01
2.75979340e-01 -1.29419222e-01 5.24562657e-01 1.06275856e+00
-1.97463501e-02 5.71177959e-01 1.22375193e-03 5.85739672e-01
-3.24952096e-01 3.36849153e-01 1.40340656e-01 1.98414266e-01
7.66903520e-01 9.69788253e-01 -3.89999062e-01 -2.96036571e-01
-4.44931477e-01 1.13939977e+00 5.60397506e-01 2.52795190e-01
-7.57198632e-01 -5.56410909e-01 2.91368902e-01 -1.53296664e-01
8.54759961e-02 -4.60773297e-02 -2.91965753e-01 -1.16958177e+00
5.96189797e-01 -1.45726562e+00 4.69437152e-01 -7.22397506e-01
-1.16493130e+00 6.01742923e-01 -6.42729253e-02 -8.99771154e-01
-3.12442333e-01 -1.20005190e-01 -1.06783104e+00 1.06144118e+00
-1.58246529e+00 -9.95606065e-01 -6.33517623e-01 3.39843422e-01
1.10802424e+00 3.36739600e-01 7.72643626e-01 -1.28626496e-01
-5.61721742e-01 5.38729310e-01 -1.72774658e-01 -4.70730662e-02
7.02278972e-01 -1.34332025e+00 5.14360309e-01 8.21975052e-01
-3.44613969e-01 6.22623503e-01 7.19520390e-01 -3.94889325e-01
-1.45413613e+00 -7.20029056e-01 1.34053266e+00 -5.17661810e-01
3.46463770e-01 -2.69596398e-01 -1.05392110e+00 6.37350380e-01
7.27021754e-01 -8.89782906e-01 8.58117104e-01 5.11589944e-01
-3.77273262e-01 2.04803590e-02 -8.70898426e-01 6.22429788e-01
8.53444695e-01 -7.57239223e-01 -1.27137721e+00 3.44706506e-01
9.60148275e-01 -5.11158407e-01 -6.12389147e-01 1.36441752e-01
4.78418589e-01 -9.85728920e-01 4.37290341e-01 -4.03386891e-01
8.73209596e-01 -2.15815052e-01 -1.44553110e-01 -1.21901226e+00
9.47911814e-02 -8.01282167e-01 -1.06679648e-01 1.24973023e+00
5.00513017e-01 -6.05653644e-01 3.71371657e-01 7.48143017e-01
-1.36245996e-01 -1.00058544e+00 -9.45144176e-01 -2.48433679e-01
3.08215827e-01 -4.53984141e-01 7.72525251e-01 5.45442820e-01
4.80098128e-01 9.86932755e-01 -2.99288571e-01 3.94297242e-02
4.53690499e-01 3.84835780e-01 8.94097388e-01 -7.63288975e-01
-3.56178939e-01 -2.57524550e-01 3.84962380e-01 -1.75122046e+00
1.03575349e-01 -6.91569686e-01 3.38044524e-01 -1.94367599e+00
3.88444185e-01 -1.54408917e-01 1.26664087e-01 1.24460004e-01
-7.96635985e-01 -6.47911608e-01 2.97882587e-01 1.95856243e-01
-1.11581028e+00 7.48943865e-01 1.42144156e+00 6.37598112e-02
-9.76600647e-02 8.76783133e-02 -9.32056844e-01 6.25480115e-01
8.79971445e-01 -1.02680594e-01 -6.51504278e-01 -5.44446528e-01
3.07348728e-01 6.16357803e-01 8.09796080e-02 -8.10818613e-01
4.79152143e-01 -1.75697416e-01 -1.97884485e-01 -9.48983550e-01
-5.74715510e-02 -3.39685112e-01 -4.72311467e-01 3.47277015e-01
-8.60678434e-01 1.17000289e-01 -5.93765341e-02 5.83256960e-01
-3.32240403e-01 -4.62684035e-01 3.60249460e-01 -7.42430016e-02
-5.17362416e-01 -2.98519343e-01 -6.41167998e-01 5.34276366e-01
7.31316328e-01 2.33258620e-01 -6.19165361e-01 -9.03502107e-01
-2.97509611e-01 8.74694586e-01 3.85629159e-04 4.43768054e-01
5.37754476e-01 -9.00359333e-01 -1.22512507e+00 -3.12754542e-01
2.71257579e-01 3.68734390e-01 7.27221668e-01 8.71737778e-01
-4.96497929e-01 6.21368051e-01 1.42418995e-01 -3.59224707e-01
-1.49939692e+00 2.44527727e-01 1.91289783e-01 -8.19228649e-01
-5.20969748e-01 8.49985242e-01 1.21581644e-01 -6.18808508e-01
2.74016380e-01 -6.00836813e-01 -3.54822695e-01 3.65352444e-02
8.29636395e-01 4.48034734e-01 1.45489112e-01 -2.19859347e-01
-1.50089219e-01 2.85465360e-01 -3.36991102e-01 -1.86678857e-01
7.13958979e-01 -7.04096496e-01 -2.39013582e-01 4.16334420e-01
1.08454800e+00 -1.39057219e-01 -8.30346406e-01 -5.02603889e-01
8.61992016e-02 -9.79118198e-02 -2.67283350e-01 -1.09947562e+00
-1.77470282e-01 8.81868422e-01 -1.28214613e-01 1.22374274e-01
1.26864541e+00 -1.43911988e-01 1.36578345e+00 7.67721355e-01
1.39482662e-01 -1.02311277e+00 9.90689360e-03 9.63197768e-01
1.19165504e+00 -1.22136354e+00 -8.05386305e-02 -5.18165886e-01
-7.51748085e-01 1.05616868e+00 5.65117776e-01 2.92530119e-01
2.59507686e-01 -1.83858886e-01 2.07990259e-01 -2.53257323e-02
-1.25691962e+00 -1.69615984e-01 1.82327002e-01 1.08798482e-01
5.78251779e-01 2.46254459e-01 -7.65696645e-01 8.77201676e-01
-5.47996283e-01 -7.22115561e-02 4.82518941e-01 1.05634987e+00
-6.81769669e-01 -9.65901136e-01 -2.59308487e-01 2.23471940e-01
-3.09531629e-01 -2.73160398e-01 -5.82378805e-01 4.00180250e-01
-3.93440276e-01 1.75030220e+00 -2.37839654e-01 -3.20341587e-01
4.09934461e-01 2.90136606e-01 3.50366026e-01 -3.51618677e-01
-7.41878748e-01 -4.58411574e-01 6.67608500e-01 -4.03406620e-01
-2.94722676e-01 -4.66015667e-01 -1.36147690e+00 -5.93891382e-01
-2.67368972e-01 5.76913655e-01 3.47231984e-01 1.30547857e+00
2.80047476e-01 4.05655533e-01 7.48099685e-01 7.75666256e-03
-9.66351330e-01 -1.55199194e+00 2.54688971e-02 5.23155272e-01
3.56550694e-01 -4.09502611e-02 -4.87618029e-01 -3.94566655e-02] | [11.813239097595215, 8.064355850219727] |
1e61d32f-54b5-4732-a1ac-5e824eb2cfcd | robust-coordinated-longitudinal-control-of | 2208.05708 | null | https://arxiv.org/abs/2208.05708v1 | https://arxiv.org/pdf/2208.05708v1.pdf | Robust Coordinated Longitudinal Control of MAV Based on Energy State | Fixed-wing Miniature Air Vehicle (MAV) is not only coupled with longitudinal motion, but also more susceptible to wind disturbance due to its lighter weight, which brings more challenges to its altitude and airspeed controller design. Therefore, in this paper, an improved longitudinal control strategy based on energy state, is proposed to address the above-mentioned issues. The control strategy utilizes the Linear Extended State Observer (LESO) to observe the energy states and the disturbance of the MAV, and then designs a Multiple-Input Multiple-Output (MIMO) controller based on a more coordinated Total Energy Control (TEC) strategy to control the airspeed and altitude of the MAV. The performance of this control strategy has been successfully verified in a Model-in-the-Loop (MIL) simulation with Simulink, and a comparative test with the classical TEC algorithm is carried out. | ['Haodong Li', 'Dawei Li', 'Chenlong Zhang'] | 2022-08-11 | null | null | null | null | ['total-energy'] | ['miscellaneous'] | [-2.91552514e-01 -1.46165639e-01 -2.54480064e-01 4.59442735e-01
6.31421626e-01 -5.15320778e-01 5.40362239e-01 -2.11551651e-01
-5.94152473e-02 9.84658122e-01 -2.31451944e-01 -4.45469826e-01
-3.64150971e-01 -7.17144668e-01 -8.41057897e-02 -8.28512609e-01
1.76564351e-01 -2.85179645e-01 -1.48245785e-02 -4.80060130e-01
-1.64773032e-01 3.43808562e-01 -1.43777573e+00 -8.54256868e-01
1.13548946e+00 8.66028845e-01 2.45039389e-01 5.96730232e-01
4.42110211e-01 5.79666972e-01 -1.61814913e-01 3.57848734e-01
1.26432285e-01 -5.92670202e-01 -1.07721359e-01 1.48670405e-01
-2.18476981e-01 -2.30801389e-01 -1.78745046e-01 1.02506757e+00
4.79568064e-01 7.90070295e-01 4.71706659e-01 -1.11296630e+00
-1.20988674e-03 -2.89936185e-01 -2.03650951e-01 9.60550457e-02
-1.99087128e-01 2.40775153e-01 3.61063778e-01 -8.34041953e-01
3.47524494e-01 1.08465958e+00 4.82620925e-01 4.20754135e-01
-1.00546622e+00 -5.04418969e-01 7.74283037e-02 -4.60294262e-02
-1.47971272e+00 -1.56972706e-01 5.42152345e-01 -6.20486259e-01
6.37854993e-01 7.99898744e-01 1.09520912e+00 3.73969674e-01
8.23694289e-01 1.15922399e-01 1.04354513e+00 -1.24621697e-01
1.48529977e-01 3.07969183e-01 -1.85284242e-01 7.88931608e-01
8.96205068e-01 6.25154972e-01 3.99406701e-01 4.82923985e-02
6.41394138e-01 5.04030287e-02 -4.91429985e-01 -4.42229092e-01
-8.58895361e-01 5.55550933e-01 4.04859304e-01 2.81883478e-01
-1.60614416e-01 -1.93246812e-01 1.42957762e-01 1.12214535e-01
2.11955488e-01 5.92087746e-01 -4.30008680e-01 -5.57378940e-02
-5.96138716e-01 3.56711328e-01 5.82791805e-01 6.54077649e-01
1.63361222e-01 9.02560890e-01 2.78267473e-01 2.32226521e-01
5.63234985e-01 8.65113139e-01 3.25375795e-01 -6.89932644e-01
1.44345939e-01 8.68397534e-01 5.04598022e-01 -1.00914907e+00
-5.92508614e-01 -6.56475186e-01 -1.02709711e+00 5.51122606e-01
-3.23203355e-01 -8.18949878e-01 -4.80034918e-01 1.48870540e+00
6.57737792e-01 -2.08792821e-01 3.30808342e-01 1.22302115e+00
2.88290560e-01 1.29391873e+00 -2.63042390e-01 -9.44662809e-01
1.14024305e+00 -6.78702116e-01 -1.26283419e+00 -1.24727301e-01
4.22639698e-01 -6.05941176e-01 6.71854913e-01 1.34630382e-01
-9.54205275e-01 -7.71010578e-01 -1.47505867e+00 2.02976331e-01
-4.99124378e-01 6.11580193e-01 -2.45545059e-01 4.07562524e-01
-5.02659798e-01 4.49249268e-01 -7.51625597e-01 -1.66219637e-01
-7.82894492e-01 1.84928015e-01 6.42909715e-03 8.38687181e-01
-1.44781578e+00 1.33221757e+00 6.01540625e-01 3.52801383e-01
-3.79112452e-01 -5.72904944e-01 -7.01443315e-01 -2.31992900e-01
6.69880331e-01 -9.68179464e-01 9.92229044e-01 -5.51959813e-01
-2.08490109e+00 -2.57587284e-01 -1.41142622e-01 -1.06303848e-01
2.59368330e-01 -1.84247121e-01 -6.86051250e-01 -9.83589813e-02
-4.01056916e-01 -2.55544513e-01 6.91251397e-01 -8.61136377e-01
-7.79350996e-01 -1.09835468e-01 -1.48631081e-01 3.56420219e-01
-1.16433933e-01 -3.25957358e-01 4.60227579e-01 -5.36177218e-01
-2.78201699e-01 -1.08643055e+00 -1.13825373e-01 -3.17690372e-01
-7.35754743e-02 -1.84584543e-01 1.28982961e+00 -2.88391292e-01
1.78812921e+00 -1.89374709e+00 8.02685857e-01 7.05385357e-02
-3.25518936e-01 8.30214679e-01 6.52017772e-01 6.03885829e-01
2.02713937e-01 6.61815628e-02 9.64475647e-02 1.27838910e-01
-3.61133844e-01 1.05122380e-01 -3.00086856e-01 3.53154391e-01
1.04455613e-01 2.70865709e-01 -6.95950806e-01 -2.07694203e-01
5.04455686e-01 4.72418100e-01 -2.54700601e-01 3.17830414e-01
7.60769844e-02 4.29049462e-01 -7.23447978e-01 2.71174848e-01
6.26915276e-01 3.68355006e-01 6.87804520e-02 -3.31929535e-01
-1.15835774e+00 -5.21505535e-01 -1.47789145e+00 7.35163867e-01
-6.76938951e-01 3.24431390e-01 9.40394878e-01 -4.16753143e-01
1.08241367e+00 6.13929272e-01 2.35933721e-01 -4.22495008e-01
6.81958020e-01 3.06953162e-01 4.66664620e-02 -5.29693305e-01
2.91571259e-01 -2.85818487e-01 3.81829739e-01 -3.68855447e-01
-1.31173193e-01 -4.37853634e-01 5.27563468e-02 -9.30593461e-02
1.70171633e-01 9.58940536e-02 4.56958562e-01 -7.96995878e-01
1.44282401e+00 2.42075086e-01 9.24778461e-01 -5.19522130e-01
-4.23348546e-02 -5.60072541e-01 -1.81138441e-02 -1.80109426e-01
-1.06783557e+00 -5.29370248e-01 -2.72447914e-01 -5.12176454e-02
8.29149663e-01 -4.61787432e-01 -4.68788624e-01 2.04258077e-02
3.83820646e-02 8.75422478e-01 -7.78147951e-02 -7.30903506e-01
-3.85772973e-01 -3.51828873e-01 -2.85601944e-01 3.88355285e-01
7.57287741e-01 1.83394887e-02 -1.06200969e+00 2.66470790e-01
2.28255391e-01 -8.22487772e-01 -3.50834221e-01 -3.16244990e-01
-7.90442705e-01 -1.16398180e+00 -8.15841034e-02 -4.73302245e-01
5.17767310e-01 1.51603207e-01 9.62942839e-02 1.84721515e-01
-5.95083721e-02 9.81142968e-02 -1.23877488e-01 -6.93661213e-01
-1.47862390e-01 -2.66831785e-01 8.65646780e-01 2.79309243e-01
-3.68670017e-01 1.51548639e-01 -4.98903006e-01 7.79689491e-01
-6.40014112e-01 4.20401245e-01 4.54350173e-01 9.20227587e-01
6.10629261e-01 5.77908397e-01 6.91872478e-01 -1.29104242e-01
4.95821863e-01 -1.17437564e-01 -1.42070425e+00 -3.87216285e-02
-1.02683175e+00 -3.45303655e-01 1.41185462e+00 -1.21997327e-01
-1.22448134e+00 1.29388481e-01 2.82160006e-02 -6.07616007e-01
3.22620481e-01 3.09353828e-01 -3.66947830e-01 -2.51065046e-01
-5.96820749e-02 2.37635031e-01 5.50625741e-01 -4.79701489e-01
1.36935875e-01 6.99937999e-01 4.02721822e-01 -7.59896589e-03
1.34771585e+00 6.90750107e-02 9.21177685e-01 -1.08565867e+00
-1.91098213e-01 -2.11626217e-01 -4.08930808e-01 -3.94849181e-01
1.00895977e+00 -1.02580631e+00 -8.92432690e-01 3.71862113e-01
-7.45186150e-01 -2.13284075e-01 -1.62573941e-02 1.20347285e+00
-3.32314581e-01 2.21292977e-03 -3.16127628e-01 -1.30535996e+00
-4.56606925e-01 -1.22856927e+00 3.61154258e-01 6.96248114e-01
-1.13286391e-01 -1.05577171e+00 8.66337121e-02 -7.78517202e-02
7.76356161e-01 7.05488861e-01 6.71730101e-01 3.58044863e-01
-3.79161686e-01 -2.13587061e-01 5.17955482e-01 4.37320441e-01
-2.97774132e-02 3.28824967e-01 -1.87330931e-01 -9.51427758e-01
4.21996325e-01 1.05118141e-01 1.05164342e-01 2.10811943e-01
4.94435459e-01 -6.04349911e-01 -5.12856185e-01 7.96736836e-01
2.03640842e+00 6.31211817e-01 6.28135875e-02 3.11082274e-01
5.09942174e-01 4.55627710e-01 1.29075456e+00 3.57079208e-01
8.50329921e-02 7.54372835e-01 6.09836459e-01 -2.47993097e-01
4.49895829e-01 -1.50905803e-01 5.75912237e-01 1.00048029e+00
-3.08561236e-01 -1.97460577e-01 -1.41081735e-01 2.94018388e-01
-1.50620556e+00 -7.42465973e-01 -7.30678558e-01 2.14274025e+00
4.81357723e-01 -2.95802653e-01 -1.84073880e-01 3.37576360e-01
5.62239289e-01 7.40156472e-02 -3.38684201e-01 -9.90999103e-01
2.24482954e-01 -3.10223132e-01 8.94316435e-01 6.42518878e-01
-9.01611865e-01 6.08133636e-02 5.25221539e+00 8.51062655e-01
-1.49272215e+00 -2.41295770e-01 -2.36682408e-02 -4.98093367e-02
3.06448862e-02 3.38183641e-01 -8.32197964e-01 6.54601514e-01
1.12015069e+00 -8.95068824e-01 6.09262168e-01 9.23154593e-01
8.59831750e-01 -2.25709349e-01 -3.43000382e-01 4.42882478e-01
-2.85097480e-01 -7.55272627e-01 -2.09465623e-01 1.19084112e-01
7.49669194e-01 -6.28091455e-01 -1.20959371e-01 2.60430664e-01
-5.14927268e-01 -2.67000616e-01 6.99554026e-01 6.00393414e-01
7.48799920e-01 -1.05653489e+00 9.00082409e-01 9.45650876e-01
-1.66917658e+00 -4.83579308e-01 -4.07032549e-01 -5.16936898e-01
5.59402108e-01 2.48507306e-01 -2.10121959e-01 1.00180030e+00
1.16578698e-01 4.52781379e-01 -3.18237334e-01 6.65934622e-01
-1.19302332e-01 2.28427798e-01 -1.88303158e-01 -6.56339586e-01
1.69496521e-01 -8.33341777e-01 1.05115080e+00 4.23835933e-01
5.16995847e-01 4.81617212e-01 3.47133875e-01 7.84500182e-01
5.98582625e-01 3.97277057e-01 -7.04400897e-01 7.37618506e-02
1.53951362e-01 1.58002603e+00 1.02026761e-01 -6.17303550e-01
-1.52140871e-01 1.64707705e-01 -6.56123340e-01 1.67028323e-01
-1.16452503e+00 -1.14486706e+00 8.32641959e-01 3.60412389e-01
-1.35333627e-01 -6.90049410e-01 1.00390062e-01 -1.00159204e+00
-2.33955771e-01 -3.05603415e-01 -6.69035539e-02 -7.22610712e-01
-5.28523088e-01 3.90365034e-01 1.87007993e-01 -1.79647255e+00
-2.53361255e-01 -6.98402643e-01 -8.14816654e-01 1.14594519e+00
-1.38080657e+00 -5.65746844e-01 -2.18784854e-01 3.32973510e-01
5.08358955e-01 -1.46450270e-02 5.34422338e-01 3.42395931e-01
-1.16113901e+00 -1.69434607e-01 6.68863833e-01 -5.68677664e-01
3.42577279e-01 -9.44855928e-01 -3.98660421e-01 1.20739079e+00
-1.11982775e+00 5.88455439e-01 8.58878553e-01 -7.49694109e-01
-2.02178001e+00 -1.46049285e+00 8.13606560e-01 2.94901043e-01
6.29721761e-01 -1.03962906e-01 -8.32954705e-01 3.43576938e-01
5.27431011e-01 -2.07825333e-01 -1.64138600e-01 -8.69537771e-01
8.22783113e-01 -4.25738096e-01 -1.11482894e+00 6.89807951e-01
5.11507034e-01 -9.07731801e-02 -5.88871002e-01 2.07185955e-03
8.81542981e-01 -5.80361307e-01 -1.03626168e+00 9.61514056e-01
3.90247136e-01 -2.69170672e-01 5.85163832e-01 -3.16981345e-01
-2.38600731e-01 -1.19330752e+00 2.81786472e-01 -1.74839115e+00
-3.82687032e-01 -8.18926275e-01 -4.36639488e-01 1.31421006e+00
9.38929319e-02 -9.67129648e-01 9.90425516e-03 3.12087219e-02
-3.12876254e-01 -1.09066737e+00 -1.05438542e+00 -1.12721455e+00
-2.99667835e-01 4.61527944e-01 3.14484268e-01 5.99884450e-01
2.83354104e-01 7.26101816e-01 -5.56699574e-01 7.76109338e-01
1.82710767e-01 -8.55494142e-02 9.48790371e-01 -1.10749519e+00
6.19443692e-02 -1.99568436e-01 -1.08584929e-02 -4.68074292e-01
-1.70583352e-01 -4.40882802e-01 -7.35741556e-02 -1.70746243e+00
-5.67405343e-01 3.16861600e-01 3.19941305e-02 -2.20405713e-01
-3.17865849e-01 -2.42311463e-01 7.64437541e-02 -6.99097365e-02
4.21789400e-02 1.10399163e+00 1.37469125e+00 3.55983563e-02
-3.98012072e-01 3.90217066e-01 9.65465605e-02 5.65003157e-01
7.75980473e-01 5.70001230e-02 -8.55460823e-01 -1.34949073e-01
-1.32305741e-01 3.49798381e-01 1.64033294e-01 -1.70308673e+00
1.20765425e-01 -4.21485037e-01 9.69053134e-02 -7.63373494e-01
1.82312697e-01 -1.25397229e+00 6.08026743e-01 1.20028818e+00
1.60133645e-01 3.09944212e-01 2.22776204e-01 5.02163529e-01
-4.34712440e-01 1.23101875e-01 1.22527075e+00 4.17813838e-01
-5.00068724e-01 1.33570135e-01 -7.85971940e-01 -3.50441664e-01
1.61968791e+00 -2.24398635e-03 -4.60219741e-01 -3.26457880e-02
-3.45764250e-01 8.09361339e-01 4.68462080e-01 4.07086939e-01
3.15117866e-01 -1.28823555e+00 -2.14738041e-01 3.76762033e-01
-3.26621264e-01 -2.09277540e-01 5.14799297e-01 1.02183330e+00
-6.65421367e-01 8.33256423e-01 -2.57868141e-01 -3.73846889e-01
-1.40592349e+00 8.46801937e-01 6.99350536e-01 1.52264953e-01
-1.60595670e-01 3.83203216e-02 -1.43596709e-01 -8.61242041e-02
-5.06018460e-01 -5.15986264e-01 -3.71859461e-01 -1.58907458e-01
2.17684940e-01 8.27446699e-01 -1.96807653e-01 -9.36617494e-01
-1.22220658e-01 1.25857127e+00 9.53065574e-01 8.85332003e-02
7.08716393e-01 -4.08751249e-01 1.11105358e-02 4.81869102e-01
1.05940378e+00 2.65373051e-01 -1.01775432e+00 4.64726776e-01
-6.65618956e-01 -5.51894903e-01 3.70559514e-01 -6.05102599e-01
-7.97286749e-01 6.94517136e-01 5.83580554e-01 3.07973236e-01
1.10141373e+00 -1.09127450e+00 8.05648029e-01 7.58865923e-02
4.04013604e-01 -1.40111196e+00 -4.50972617e-01 5.69530427e-01
7.39680946e-01 -4.66179103e-01 3.21708351e-01 -4.89663929e-01
-6.19007289e-01 1.28840923e+00 1.14892673e+00 -2.31854796e-01
6.88155711e-01 5.04915528e-02 -4.70490120e-02 3.71708661e-01
-8.82412434e-01 3.22493329e-03 2.00035527e-01 -1.53738081e-01
2.12592825e-01 5.19102514e-02 -1.18536699e+00 5.01240253e-01
3.70232522e-01 1.88284189e-01 6.05140448e-01 9.32790577e-01
-8.42250526e-01 -9.47251678e-01 -7.21023381e-01 -1.70257434e-01
-4.09667283e-01 7.71925569e-01 1.42434567e-01 1.11001396e+00
6.09576643e-01 1.08337283e+00 7.05428421e-02 -6.46688879e-01
1.08822632e+00 -5.98886944e-02 -1.88030913e-01 -1.57749176e-01
-3.18789363e-01 1.72882333e-01 6.56292820e-03 -3.63377899e-01
-1.42540365e-01 -1.00086376e-01 -1.42603230e+00 -2.31306568e-01
-9.83221352e-01 8.05991471e-01 6.63023055e-01 5.88359177e-01
4.80821937e-01 8.07134151e-01 1.22020972e+00 -5.24155259e-01
-6.88825727e-01 -8.49238873e-01 -7.25686371e-01 1.43800983e-02
4.52173561e-01 -1.24078786e+00 -7.98570514e-01 -5.15581191e-01] | [5.379420280456543, 2.439556837081909] |
61c64f38-e5eb-4438-9488-d17b504a69ca | next3d-generative-neural-texture | 2211.11208 | null | https://arxiv.org/abs/2211.11208v2 | https://arxiv.org/pdf/2211.11208v2.pdf | Next3D: Generative Neural Texture Rasterization for 3D-Aware Head Avatars | 3D-aware generative adversarial networks (GANs) synthesize high-fidelity and multi-view-consistent facial images using only collections of single-view 2D imagery. Towards fine-grained control over facial attributes, recent efforts incorporate 3D Morphable Face Model (3DMM) to describe deformation in generative radiance fields either explicitly or implicitly. Explicit methods provide fine-grained expression control but cannot handle topological changes caused by hair and accessories, while implicit ones can model varied topologies but have limited generalization caused by the unconstrained deformation fields. We propose a novel 3D GAN framework for unsupervised learning of generative, high-quality and 3D-consistent facial avatars from unstructured 2D images. To achieve both deformation accuracy and topological flexibility, we propose a 3D representation called Generative Texture-Rasterized Tri-planes. The proposed representation learns Generative Neural Textures on top of parametric mesh templates and then projects them into three orthogonal-viewed feature planes through rasterization, forming a tri-plane feature representation for volume rendering. In this way, we combine both fine-grained expression control of mesh-guided explicit deformation and the flexibility of implicit volumetric representation. We further propose specific modules for modeling mouth interior which is not taken into account by 3DMM. Our method demonstrates state-of-the-art 3D-aware synthesis quality and animation ability through extensive experiments. Furthermore, serving as 3D prior, our animatable 3D representation boosts multiple applications including one-shot facial avatars and 3D-aware stylization. | ['Yebin Liu', 'Hongwen Zhang', 'Yong Zhang', 'Xiaoyu Li', 'Lizhen Wang', 'Xuan Wang', 'Jingxiang Sun'] | 2022-11-21 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Sun_Next3D_Generative_Neural_Texture_Rasterization_for_3D-Aware_Head_Avatars_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Sun_Next3D_Generative_Neural_Texture_Rasterization_for_3D-Aware_Head_Avatars_CVPR_2023_paper.pdf | cvpr-2023-1 | ['face-model'] | ['computer-vision'] | [ 1.72628477e-01 3.68360549e-01 -9.12558660e-03 -4.32769090e-01
-4.48031038e-01 -6.89630985e-01 8.69617879e-01 -8.78026307e-01
4.18066829e-01 5.03867686e-01 1.94763556e-01 1.14149615e-01
1.67574763e-01 -1.29359698e+00 -8.72738361e-01 -7.29360640e-01
2.52702951e-01 6.74600005e-01 -1.81460142e-01 -3.69061887e-01
-1.96191996e-01 1.15113235e+00 -1.53101563e+00 2.17713371e-01
8.12537670e-01 1.03617358e+00 -4.64439422e-01 4.01425838e-01
-3.65251660e-01 1.10774308e-01 -3.73048544e-01 -5.45015335e-01
4.89128441e-01 -5.18612802e-01 -2.95582652e-01 6.06537402e-01
7.28827477e-01 -5.50536454e-01 -2.39965618e-02 7.32435048e-01
4.06397223e-01 -1.25865445e-01 1.01368344e+00 -1.39322650e+00
-1.02074933e+00 -5.36073819e-02 -9.33778048e-01 -7.30996907e-01
4.87315059e-01 4.46052432e-01 5.82361579e-01 -9.76061523e-01
9.80192721e-01 1.79146075e+00 5.49651921e-01 9.61312771e-01
-1.50908566e+00 -7.21567333e-01 2.06382722e-01 -6.25124097e-01
-1.16166258e+00 -3.56433660e-01 1.35458231e+00 -5.28409719e-01
3.82730871e-01 4.65676099e-01 9.94578302e-01 1.34495163e+00
3.13922763e-01 3.61118585e-01 1.32372189e+00 4.48135696e-02
7.92024359e-02 -2.47337833e-01 -6.47734940e-01 1.07160008e+00
-1.18228242e-01 5.12292646e-02 -3.77840847e-01 -2.72824764e-01
1.77799129e+00 1.11199617e-01 -1.46964252e-01 -6.14225507e-01
-8.53481412e-01 7.92987645e-01 3.00552875e-01 -1.91718638e-01
-3.75232995e-01 4.00494814e-01 -3.54181230e-02 -4.38501202e-02
7.18908906e-01 2.76335686e-01 -2.71705717e-01 1.02424316e-01
-9.09721911e-01 5.87141335e-01 4.85662699e-01 1.02754784e+00
9.60509360e-01 7.14048266e-01 -3.61176163e-01 7.29279578e-01
4.05725747e-01 8.87626290e-01 1.15943931e-01 -1.28727674e+00
1.23939095e-02 8.30096483e-01 -1.50178418e-01 -1.04191589e+00
-9.22393948e-02 -1.14307940e-01 -1.17043781e+00 7.14099526e-01
1.05235994e-01 1.60245225e-01 -1.32907665e+00 1.88807988e+00
7.58880138e-01 3.74493822e-02 -3.29421967e-01 9.21217084e-01
9.98952389e-01 5.97286046e-01 -1.33795649e-01 -2.32271865e-01
1.39812160e+00 -5.84979355e-01 -7.00876653e-01 2.48349860e-01
3.45368707e-03 -6.00399137e-01 1.21865547e+00 1.71484768e-01
-1.54146457e+00 -3.78006846e-01 -6.95126534e-01 -1.97699353e-01
4.43019308e-02 -2.18800753e-01 6.58503234e-01 6.45080268e-01
-1.12336338e+00 4.67234910e-01 -8.03405166e-01 1.34652853e-01
7.40682065e-01 2.36878783e-01 -5.07219493e-01 1.55867934e-01
-8.49918544e-01 4.31781948e-01 -4.71137524e-01 -1.35899022e-01
-9.76747155e-01 -1.05313826e+00 -9.93010819e-01 -2.24859655e-01
1.53372869e-01 -1.26506102e+00 7.74858475e-01 -8.91406178e-01
-2.13842463e+00 1.27721000e+00 -4.64883111e-02 2.20647544e-01
6.90654397e-01 3.08103464e-03 -1.67641640e-02 6.52067065e-02
-7.49703795e-02 7.16943204e-01 1.26519859e+00 -1.64839220e+00
3.26139092e-01 -5.21144509e-01 1.27496153e-01 2.12316468e-01
-6.33722544e-02 -3.63350391e-01 -3.85362566e-01 -1.11322331e+00
2.81781584e-01 -7.76784897e-01 -1.04485944e-01 6.86718345e-01
-5.23119152e-01 1.30109727e-01 1.11385381e+00 -4.97430176e-01
6.23203635e-01 -1.83297777e+00 5.04072189e-01 3.04346591e-01
2.51777142e-01 -8.54516216e-03 -2.56224722e-01 1.44366860e-01
4.99944724e-02 4.61159796e-01 -3.00284535e-01 -5.40866137e-01
1.29787847e-01 3.80161375e-01 -2.91415036e-01 2.29418427e-01
6.22214139e-01 1.15385664e+00 -5.89093208e-01 -4.41958874e-01
1.85777128e-01 1.05656576e+00 -9.70837593e-01 4.15268272e-01
-4.88662660e-01 1.04083323e+00 -7.84916878e-01 9.46803689e-01
1.07423270e+00 -2.51998957e-02 -3.55603658e-02 -4.41600859e-01
1.27715796e-01 -1.79536015e-01 -1.00976479e+00 1.90770590e+00
-5.32661200e-01 -1.16205089e-01 5.05716026e-01 -4.19005126e-01
1.25836372e+00 2.19634131e-01 4.62995350e-01 -6.25484407e-01
1.98446363e-01 6.36668429e-02 -5.88537693e-01 -1.73520058e-01
3.18808645e-01 -4.01153922e-01 -1.07807979e-01 4.21336591e-01
6.39659092e-02 -8.47291708e-01 -5.82036436e-01 -4.91421819e-02
5.36411166e-01 7.16250122e-01 -1.25718549e-01 -2.02518493e-01
3.49421233e-01 -6.12769544e-01 6.01851463e-01 6.46285340e-03
4.25824016e-01 9.54205871e-01 5.13796747e-01 -5.60846508e-01
-1.27043021e+00 -1.31580639e+00 -1.62066206e-01 6.93653464e-01
4.69253510e-02 -2.45211229e-01 -9.90949035e-01 -5.06528437e-01
1.69450998e-01 3.48693609e-01 -9.26242113e-01 -6.23224899e-02
-7.65064895e-01 -3.93998414e-01 5.23411155e-01 3.39198709e-01
5.85830629e-01 -7.71441698e-01 -4.13835436e-01 -1.06021821e-01
2.01248303e-01 -9.27615047e-01 -6.91608369e-01 -4.91285920e-01
-9.43387628e-01 -7.34955609e-01 -8.56110394e-01 -4.19487208e-01
7.58847535e-01 -2.12039888e-01 1.13859296e+00 -9.40970320e-04
-3.19543719e-01 4.97455329e-01 4.10451367e-02 -8.88441503e-02
-5.85625231e-01 -4.28798884e-01 1.74939930e-01 3.50071877e-01
-4.78686243e-01 -1.28188682e+00 -7.61142790e-01 4.53419358e-01
-9.52744722e-01 4.10624683e-01 1.31171793e-01 7.70436347e-01
1.05312622e+00 -5.71923077e-01 3.12557101e-01 -8.70086789e-01
4.19383079e-01 -6.61882609e-02 -5.11317313e-01 2.43399329e-02
-3.06612074e-01 7.36948475e-02 5.52786827e-01 -6.94730520e-01
-1.12028801e+00 -1.74204275e-01 -2.96294928e-01 -1.05440986e+00
-2.29369536e-01 -2.41231754e-01 -7.64748752e-01 -3.90531451e-01
4.69532341e-01 2.49866158e-01 3.48090053e-01 -4.68674302e-01
7.75367916e-01 1.45084321e-01 4.15675163e-01 -1.10942256e+00
1.14477324e+00 6.52203441e-01 4.04545933e-01 -6.92320466e-01
-2.94042408e-01 6.52193367e-01 -6.24681830e-01 -2.95259416e-01
9.66371596e-01 -8.71937633e-01 -7.89214253e-01 6.78491652e-01
-1.17141175e+00 -4.91570532e-01 -6.12273097e-01 -1.61112919e-01
-9.52320158e-01 1.20353773e-01 -5.53814471e-01 -7.18524098e-01
-4.84295160e-01 -1.35673904e+00 1.68168104e+00 7.44193047e-02
-2.33946040e-01 -8.23425293e-01 -8.92428160e-02 3.12153488e-01
4.89850760e-01 1.21052837e+00 9.99938965e-01 4.20254290e-01
-7.73432016e-01 2.45667949e-01 1.56308889e-01 1.32355049e-01
3.32232356e-01 2.80109942e-01 -1.03720152e+00 -1.12178750e-01
-1.38431847e-01 -2.96995878e-01 4.34606194e-01 3.86424303e-01
1.37056792e+00 -5.81389308e-01 -1.33931115e-01 1.36721873e+00
1.16975451e+00 5.63008785e-02 7.18189597e-01 -1.74083576e-01
1.17931879e+00 6.01669192e-01 1.39246047e-01 6.38857424e-01
3.42710555e-01 9.12802875e-01 7.10283935e-01 -2.86108792e-01
-4.38871592e-01 -7.29148686e-01 8.14430863e-02 5.20223796e-01
-6.49818957e-01 -6.81508407e-02 -3.12213659e-01 6.97780252e-02
-1.21000981e+00 -7.89447188e-01 1.79486722e-01 1.85550857e+00
9.18597341e-01 -2.31493428e-01 1.85424238e-01 -1.05608977e-01
4.62147146e-01 2.90006787e-01 -7.53010929e-01 -5.76635361e-01
-1.37066752e-01 5.52952707e-01 -1.14533603e-02 4.89550263e-01
-5.89873433e-01 1.07825267e+00 5.31571102e+00 1.05927157e+00
-1.31433141e+00 1.11644827e-01 8.54643703e-01 -2.09330797e-01
-1.28607392e+00 -3.25139195e-01 -5.43423772e-01 2.91127145e-01
2.43047357e-01 1.97861165e-01 2.41335422e-01 7.55172491e-01
3.13436151e-01 5.81628084e-01 -9.47202265e-01 1.14445817e+00
-1.24841027e-01 -1.74237585e+00 8.07372451e-01 4.28734422e-01
1.01698828e+00 -7.90772676e-01 4.08933640e-01 -1.48110598e-01
2.05954432e-01 -1.36372542e+00 9.79647577e-01 7.23860502e-01
1.75638354e+00 -8.65058661e-01 -1.95327908e-01 1.20507114e-01
-1.15635943e+00 4.82518405e-01 -1.72588825e-02 3.10810298e-01
3.99668664e-01 4.17982697e-01 -2.53142446e-01 5.00411034e-01
5.91226578e-01 4.97127593e-01 -5.29999062e-02 1.27990484e-01
-2.21884415e-01 1.90718040e-01 -2.72731990e-01 3.26466411e-01
8.35631341e-02 -6.35260940e-01 6.67504489e-01 7.12096274e-01
4.60093975e-01 4.09587473e-01 -8.44917148e-02 1.45798349e+00
-3.44183803e-01 -5.39962240e-02 -8.13124955e-01 8.57181400e-02
4.34135556e-01 1.17926145e+00 -4.03892308e-01 -6.26322581e-04
7.77669176e-02 1.08511007e+00 1.89584777e-01 3.91817778e-01
-8.53541970e-01 1.65619761e-01 1.12412703e+00 6.35027289e-01
1.96381107e-01 -3.32554251e-01 -3.65905613e-01 -1.23085153e+00
-5.41860275e-02 -8.52313936e-01 -3.39793831e-01 -9.63242531e-01
-1.32069993e+00 7.89829671e-01 -4.16933037e-02 -1.09392583e+00
-3.53903443e-01 -4.07904178e-01 -6.06978357e-01 1.03621006e+00
-1.33243275e+00 -1.87538373e+00 -4.59877759e-01 7.80731022e-01
3.58866155e-01 -1.28987163e-01 1.13415289e+00 -7.31552169e-02
-1.68808296e-01 8.64418864e-01 -3.57837647e-01 -2.79580299e-02
5.91444194e-01 -9.12061036e-01 5.77239513e-01 3.78491789e-01
-1.02645338e-01 4.23960626e-01 3.63059282e-01 -6.42484128e-01
-1.81545055e+00 -1.25885260e+00 6.47245198e-02 -5.46860337e-01
1.25482604e-02 -5.45482039e-01 -8.83511961e-01 4.93887037e-01
-5.03866747e-02 2.42001086e-01 4.85173523e-01 -4.11310852e-01
-6.32335603e-01 -1.91614717e-01 -1.73036814e+00 6.46841049e-01
1.51218307e+00 -4.13629174e-01 -1.20940335e-01 2.59079728e-02
8.18310499e-01 -7.58481622e-01 -1.25125265e+00 5.50433040e-01
6.84355259e-01 -1.07977235e+00 1.05121887e+00 -4.58564818e-01
6.69063449e-01 -2.45229259e-01 -1.04502842e-01 -1.23458481e+00
-2.46561870e-01 -1.10929179e+00 -2.08445072e-01 1.36338437e+00
-1.05656087e-01 -4.34129089e-01 9.24778342e-01 5.27808428e-01
-2.95606434e-01 -1.15301728e+00 -7.76192844e-01 -4.94849145e-01
3.11615139e-01 -2.10216880e-01 1.14016175e+00 1.07813823e+00
-7.35558391e-01 -1.15778364e-01 -4.08479303e-01 -3.76693197e-02
7.96813428e-01 4.43982989e-01 1.05647290e+00 -1.13439786e+00
-2.18292549e-01 -4.30219471e-01 -4.83034283e-01 -1.12451637e+00
2.78624237e-01 -8.32189023e-01 -4.85945910e-01 -1.24531555e+00
-1.38903946e-01 -6.90739632e-01 6.87724888e-01 5.30778706e-01
3.68442208e-01 5.62619507e-01 1.52556732e-01 1.45843737e-02
3.23005199e-01 1.02369452e+00 2.25668454e+00 1.76320877e-02
-2.12058336e-01 -3.01994085e-01 -6.76762760e-01 7.77480125e-01
3.86738449e-01 3.40918787e-02 -5.40367723e-01 -6.21201038e-01
-6.97142631e-02 3.41623783e-01 6.73692882e-01 -5.42285204e-01
-3.77176702e-01 -5.34570932e-01 6.41892135e-01 -3.57123971e-01
8.71014476e-01 -6.53893828e-01 7.24422872e-01 -2.77439337e-02
-2.29491216e-05 -1.55541226e-01 2.29160145e-01 4.02260482e-01
3.28202210e-02 6.00553095e-01 1.11543512e+00 -3.36570770e-01
-1.88094795e-01 1.04786980e+00 1.14870220e-01 1.36586338e-01
9.42300737e-01 -5.05151689e-01 -1.52385153e-03 -4.73969370e-01
-7.96466768e-01 -2.61760205e-01 1.14251149e+00 3.65396142e-01
8.91201913e-01 -1.83545566e+00 -7.47088313e-01 8.06051314e-01
-1.57593757e-01 5.80196381e-01 3.72853190e-01 2.95058340e-01
-6.40679121e-01 -1.69481605e-01 -5.41722715e-01 -7.94728458e-01
-1.03131664e+00 1.91219032e-01 4.98807281e-01 -1.65356565e-02
-6.84552491e-01 9.57680702e-01 7.34724998e-01 -7.24003494e-01
-1.64427251e-01 -2.62886465e-01 1.00737482e-01 -1.93867102e-01
2.43652537e-01 -6.40780106e-03 -1.11164190e-01 -7.62385428e-01
-1.51285216e-01 1.43678737e+00 4.63856936e-01 -1.01000965e-01
1.42173982e+00 9.42335501e-02 -1.20504968e-01 1.05876766e-01
1.09150136e+00 2.56942004e-01 -1.80053091e+00 1.47742853e-01
-1.00421607e+00 -5.89478374e-01 -2.58236587e-01 -5.49730420e-01
-1.56025791e+00 9.50313866e-01 2.29537889e-01 -2.89738774e-01
1.11650252e+00 -4.90187779e-02 8.47112417e-01 -3.47406685e-01
5.81610322e-01 -4.43333536e-01 3.67853880e-01 3.09661150e-01
1.27403140e+00 -6.54823422e-01 -2.17060149e-01 -7.80276000e-01
-5.15139461e-01 1.11014354e+00 7.60620534e-01 -3.97100300e-01
6.79220974e-01 4.90215063e-01 -9.47439149e-02 -2.91922212e-01
-4.54584479e-01 4.13606584e-01 6.02042019e-01 8.94407332e-01
3.25827211e-01 5.70696928e-02 2.43729055e-02 5.97779870e-01
-4.09719825e-01 -1.99713886e-01 1.82506725e-01 5.01160920e-01
6.43576756e-02 -1.16790044e+00 -3.63838166e-01 8.27353075e-02
-1.75484940e-01 1.39435232e-01 -2.74568945e-01 8.93995762e-01
3.91758531e-01 2.63585806e-01 3.93763632e-01 -4.16931361e-01
4.30882752e-01 -8.30370773e-05 9.11208272e-01 -5.06751835e-01
-3.61534148e-01 3.98149788e-01 -2.26348922e-01 -8.57746482e-01
-3.51845890e-01 -3.29708695e-01 -1.08265150e+00 -6.09700799e-01
2.39805251e-01 -2.09333286e-01 5.81507981e-01 4.81023878e-01
7.62461782e-01 4.59227234e-01 8.29608262e-01 -1.42539227e+00
-2.27638215e-01 -6.37072384e-01 -7.30090559e-01 6.47883117e-01
2.04636738e-01 -9.25540030e-01 -2.33917475e-01 2.15618998e-01] | [12.614302635192871, -0.39362549781799316] |
42bec710-a1c7-411b-b466-8908617d5e6a | srt3d-a-sparse-region-based-3d-object | 2110.12715 | null | https://arxiv.org/abs/2110.12715v1 | https://arxiv.org/pdf/2110.12715v1.pdf | SRT3D: A Sparse Region-Based 3D Object Tracking Approach for the Real World | Region-based methods have become increasingly popular for model-based, monocular 3D tracking of texture-less objects in cluttered scenes. However, while they achieve state-of-the-art results, most methods are computationally expensive, requiring significant resources to run in real-time. In the following, we build on our previous work and develop SRT3D, a sparse region-based approach to 3D object tracking that bridges this gap in efficiency. Our method considers image information sparsely along so-called correspondence lines that model the probability of the object's contour location. We thereby improve on the current state of the art and introduce smoothed step functions that consider a defined global and local uncertainty. For the resulting probabilistic formulation, a thorough analysis is provided. Finally, we use a pre-rendered sparse viewpoint model to create a joint posterior probability for the object pose. The function is maximized using second-order Newton optimization with Tikhonov regularization. During the pose estimation, we differentiate between global and local optimization, using a novel approximation for the first-order derivative employed in the Newton method. In multiple experiments, we demonstrate that the resulting algorithm improves the current state of the art both in terms of runtime and quality, performing particularly well for noisy and cluttered images encountered in the real world. | ['Alin Albu-Schäffer', 'Rudolph Triebel', 'Klaus H. Strobl', 'Martin Pfanne', 'Manuel Stoiber'] | 2021-10-25 | null | null | null | null | ['real-time-visual-tracking', '3d-object-tracking', '3d-multi-object-tracking'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 1.13884576e-01 -2.83969641e-01 -7.05062822e-02 -1.41551524e-01
-9.44648564e-01 -5.94207764e-01 5.61709940e-01 7.59301186e-02
-3.16984922e-01 4.16296363e-01 -1.69294685e-01 9.52396020e-02
-7.61625692e-02 -3.41785580e-01 -7.07686484e-01 -6.59350991e-01
-5.18638864e-02 7.50787914e-01 6.99843347e-01 2.65315771e-01
2.43093893e-01 1.09186685e+00 -1.66297030e+00 -1.40858456e-01
4.61574554e-01 1.23943186e+00 1.22353874e-01 6.68732941e-01
9.23591107e-02 3.90343457e-01 -4.20398146e-01 -1.25056073e-01
3.11552107e-01 -4.32750098e-02 -3.72646391e-01 5.17721355e-01
9.13529396e-01 -2.97167599e-01 -1.33910000e-01 1.04827106e+00
4.44393009e-01 1.44465938e-01 6.96240842e-01 -9.50074613e-01
-5.56614399e-02 -2.54161268e-01 -6.93275273e-01 1.00536592e-01
4.60349530e-01 6.59273863e-02 5.43551683e-01 -1.26345122e+00
9.52475548e-01 1.40666521e+00 9.18503284e-01 2.79893786e-01
-1.28971374e+00 -1.29441604e-01 4.04201299e-01 -1.05487794e-01
-1.56820345e+00 -3.40596676e-01 7.74728954e-01 -6.35177076e-01
7.75549710e-01 2.71578580e-01 8.42020214e-01 5.13273060e-01
4.02974218e-01 8.62243354e-01 1.17238426e+00 -5.03479481e-01
3.07743222e-01 8.23022127e-02 -1.05219193e-01 7.36324549e-01
3.59081358e-01 4.36636716e-01 -6.04934812e-01 -4.58134919e-01
9.99417603e-01 -2.25394100e-01 -9.35596451e-02 -1.25754130e+00
-1.04798663e+00 7.21422613e-01 3.74699563e-01 1.40947074e-01
-2.67497808e-01 2.42031813e-01 -4.33260761e-02 -3.01867008e-01
9.58308160e-01 -1.27332866e-01 -1.79971024e-01 4.85862531e-02
-1.09179640e+00 6.18855476e-01 7.02898324e-01 9.00811493e-01
5.58583736e-01 -3.66104655e-02 -1.59373984e-01 5.28386414e-01
9.32261586e-01 8.30556691e-01 -3.79488885e-01 -1.14658082e+00
-9.73582268e-02 3.42964441e-01 5.39854467e-01 -1.01225579e+00
-2.67988384e-01 -7.47977853e-01 -2.53653198e-01 7.74364591e-01
6.48681879e-01 2.28579089e-01 -8.95349085e-01 1.34426343e+00
1.05580211e+00 2.09394574e-01 -2.78303295e-01 1.13434458e+00
5.39435387e-01 4.63447422e-01 -2.81529337e-01 -2.94182330e-01
1.22310054e+00 -7.59552836e-01 -6.33594334e-01 -1.51049972e-01
9.11233574e-02 -1.11965621e+00 4.71523732e-01 4.51550722e-01
-1.26380455e+00 -4.03542459e-01 -7.42606878e-01 9.48265791e-02
-1.03970930e-01 1.28823072e-01 4.96836334e-01 6.32439137e-01
-9.42664504e-01 5.29146612e-01 -1.06448066e+00 -3.41068983e-01
4.05924171e-01 2.95741767e-01 -8.12193528e-02 -1.51738152e-01
-4.99222159e-01 1.14549661e+00 3.00440416e-02 1.12676077e-01
-9.30843115e-01 -8.58623624e-01 -9.43690300e-01 -4.20377433e-01
5.85994482e-01 -7.91174233e-01 1.34988523e+00 -2.98472881e-01
-1.64846706e+00 1.11602461e+00 -3.94079626e-01 -2.70663828e-01
6.55644059e-01 -4.55452949e-01 1.57405540e-01 2.54790783e-01
-4.52753045e-02 5.22785306e-01 8.95363688e-01 -1.71450365e+00
-4.37307477e-01 -4.29382682e-01 -1.59096226e-01 2.45673612e-01
4.09631431e-01 1.61000893e-01 -8.79616082e-01 -6.32594109e-01
3.22148234e-01 -1.02496946e+00 -3.59458327e-01 7.11487770e-01
-1.89257309e-01 -4.72223796e-02 9.59196985e-01 -4.40374553e-01
7.23202527e-01 -2.02286959e+00 1.25190362e-01 2.30289117e-01
3.09912581e-02 1.51202679e-01 2.94767559e-01 1.86742216e-01
3.87900501e-01 -5.56888700e-01 -3.68274480e-01 -7.67483711e-01
-5.92803992e-02 7.35024139e-02 -1.95217244e-02 1.09003723e+00
6.85990229e-02 7.04388261e-01 -9.04364765e-01 -5.46875715e-01
7.91140497e-01 8.25800300e-01 -4.82240736e-01 1.36405885e-01
-4.54915375e-01 6.71477437e-01 -4.87542003e-01 7.98647285e-01
1.03417003e+00 -2.49368384e-01 -1.16626464e-01 -3.29498231e-01
-5.45202315e-01 1.42764179e-02 -1.55973363e+00 1.81660903e+00
-2.06193089e-01 4.48621571e-01 5.88337302e-01 -6.10660434e-01
7.67725706e-01 3.90929654e-02 8.29079628e-01 -1.06488779e-01
2.38508835e-01 2.05203176e-01 -4.70292181e-01 2.16873847e-02
6.17252469e-01 -2.01160520e-01 1.82314128e-01 1.61111712e-01
-1.86929703e-01 -5.93702018e-01 -2.87843049e-02 1.52909771e-01
6.58540905e-01 6.88643634e-01 4.11857814e-01 -4.71404791e-01
5.97931921e-01 3.94849360e-01 2.39948809e-01 6.69418812e-01
-1.84914961e-01 7.73752034e-01 -3.58295031e-02 -1.88874573e-01
-7.98164845e-01 -1.10325778e+00 -4.47068691e-01 5.44773221e-01
4.74446684e-01 -1.15797497e-01 -5.73536634e-01 -4.27961469e-01
3.84101748e-01 4.82808739e-01 -3.97996843e-01 1.95112243e-01
-6.13707840e-01 -5.22313356e-01 -3.39788832e-02 3.10201973e-01
1.51284397e-01 -5.09255290e-01 -9.81972694e-01 3.12928587e-01
1.04655519e-01 -1.28694844e+00 -3.81490648e-01 3.45751233e-02
-1.07042229e+00 -1.04094291e+00 -9.36361492e-01 -3.38886052e-01
6.89074457e-01 4.91152793e-01 1.24324811e+00 -1.12282023e-01
-4.72338408e-01 9.71910775e-01 -7.06387460e-02 -3.90676767e-01
-2.48792529e-01 -5.57640433e-01 7.29817152e-02 -1.06084019e-01
4.05709967e-02 -6.56827837e-02 -4.88880515e-01 5.32433391e-01
-5.62098503e-01 -1.67420670e-01 3.12056065e-01 5.66569328e-01
1.09434628e+00 -1.44264802e-01 -2.85257906e-01 -3.87243569e-01
-1.07565567e-01 -5.96889406e-02 -1.36670423e+00 -2.28821374e-02
-2.75507450e-01 -5.18634468e-02 -8.13095346e-02 -5.29738367e-01
-1.13561988e+00 5.10837138e-01 9.28639472e-02 -8.46128404e-01
-9.98923853e-02 2.40075693e-01 1.14319146e-01 -7.58612633e-01
3.70951116e-01 7.55180418e-02 8.88226405e-02 -6.47469997e-01
3.25740516e-01 1.53383836e-01 5.36931396e-01 -5.01760602e-01
1.01592898e+00 1.06745076e+00 2.84758449e-01 -8.71899784e-01
-9.81472373e-01 -7.87103713e-01 -6.23481333e-01 -7.16335952e-01
7.03153312e-01 -9.02118564e-01 -8.42001915e-01 4.91517782e-01
-1.25487912e+00 -2.23301873e-01 -4.57794517e-01 6.37550116e-01
-7.23966956e-01 4.72304106e-01 -2.52252758e-01 -1.30118608e+00
1.81279302e-01 -1.09265447e+00 1.61455560e+00 9.11195651e-02
1.25599340e-01 -1.06717682e+00 2.87627012e-01 1.43656254e-01
3.05335164e-01 4.26962256e-01 3.74980718e-01 1.71425901e-02
-1.08090627e+00 -2.17640474e-01 -2.21928224e-01 1.55960573e-02
-1.48024246e-01 -8.77808034e-03 -9.87787187e-01 -4.22178507e-01
2.78084666e-01 7.17178509e-02 5.99601746e-01 9.29598510e-01
6.52432740e-01 2.03128397e-01 -6.70366585e-01 5.01322091e-01
1.65768075e+00 -7.78630301e-02 2.08071873e-01 1.40058666e-01
6.69626892e-01 5.83178997e-01 1.19034374e+00 5.49587250e-01
1.44052014e-01 1.24220216e+00 7.32847393e-01 2.91880257e-02
-4.04699028e-01 -3.74380350e-02 2.89895594e-01 3.70606869e-01
-7.96178430e-02 -1.05879374e-01 -7.06592083e-01 4.12519753e-01
-1.86556041e+00 -6.62317038e-01 -3.08794141e-01 2.56383395e+00
5.22205830e-01 7.09895566e-02 1.56318590e-01 -6.60269931e-02
6.14900529e-01 4.31708805e-02 -5.28758347e-01 2.25742057e-01
2.87754759e-02 5.49308397e-02 6.51096940e-01 9.47344065e-01
-1.23819900e+00 8.88941467e-01 6.88930178e+00 7.93905318e-01
-9.78101075e-01 2.13374779e-01 1.06418066e-01 -1.03839733e-01
-1.85787231e-01 3.11475564e-02 -1.24473941e+00 1.19871430e-01
4.99654830e-01 1.91426799e-01 1.51177347e-01 8.37562442e-01
2.56167948e-01 -6.76297724e-01 -9.15829420e-01 1.00296688e+00
1.87119395e-01 -1.20169234e+00 -3.28574151e-01 2.57681251e-01
7.10067749e-01 8.77470002e-02 -3.26615846e-04 -2.93944567e-01
1.67227730e-01 -6.83565259e-01 1.16345942e+00 5.49273074e-01
5.22474110e-01 -4.78357375e-01 4.70482141e-01 4.07307535e-01
-1.44201124e+00 3.99279326e-01 -2.86900818e-01 9.87758264e-02
5.62082708e-01 1.02318823e+00 -6.26227498e-01 6.49741173e-01
5.90257883e-01 6.25921130e-01 -3.04272532e-01 1.42546678e+00
-1.90444421e-02 1.14015907e-01 -9.07478690e-01 5.33808731e-02
9.66089144e-02 -1.97104663e-01 1.14264870e+00 9.91690457e-01
2.11019471e-01 8.98149982e-02 4.28211629e-01 9.11370814e-01
4.15261060e-01 -2.23136082e-01 -5.68095922e-01 4.83140737e-01
2.07589865e-01 1.03419125e+00 -1.02293527e+00 -1.92481592e-01
-2.68944770e-01 6.95433855e-01 -2.28911247e-02 1.83416441e-01
-8.02496731e-01 2.69574553e-01 5.38546145e-01 2.48544782e-01
7.34137774e-01 -5.94751298e-01 -2.15194240e-01 -1.07294559e+00
1.31490678e-01 -5.31029582e-01 1.58389196e-01 -7.96179712e-01
-1.17930174e+00 3.83045614e-01 5.15394270e-01 -1.42045629e+00
-2.24950448e-01 -8.06143701e-01 -5.07066213e-02 7.93246865e-01
-1.65189111e+00 -1.20827687e+00 -2.26212874e-01 4.47082698e-01
5.06697416e-01 4.01230097e-01 5.90006292e-01 2.77800798e-01
7.74712339e-02 5.52940890e-02 1.43051729e-01 -3.72092575e-01
5.56856334e-01 -1.00766039e+00 2.32294500e-01 9.06723619e-01
6.85643628e-02 5.20114541e-01 9.79792535e-01 -9.64471996e-01
-1.71041644e+00 -7.79089689e-01 6.12900674e-01 -5.96092999e-01
4.33037221e-01 -4.04819965e-01 -6.10692441e-01 5.40141642e-01
-2.92124063e-01 4.81629282e-01 1.34783030e-01 -1.44511908e-01
1.40904980e-02 2.72744507e-01 -1.30415010e+00 2.90476948e-01
9.72085536e-01 -2.75118738e-01 -4.75498259e-01 3.62044871e-01
3.29851389e-01 -1.02888954e+00 -8.55336487e-01 5.54927230e-01
6.63873136e-01 -1.05403602e+00 1.33185554e+00 9.47009325e-02
-2.74539709e-01 -7.83268809e-01 -2.59448826e-01 -9.02172685e-01
-2.09922090e-01 -6.54276431e-01 -4.81719226e-01 8.14367115e-01
-1.50364995e-01 -4.83390689e-01 1.08602607e+00 3.91844273e-01
-2.09723100e-01 -7.88883865e-01 -1.27096653e+00 -1.03785801e+00
-2.80775100e-01 -6.59376204e-01 -1.09637938e-01 4.00332630e-01
-6.05925679e-01 -1.38544738e-01 -2.80065894e-01 4.08647060e-01
1.37613189e+00 3.51531297e-01 9.30373609e-01 -1.42676616e+00
-1.83896303e-01 -2.62911528e-01 -4.83872861e-01 -1.55000257e+00
-1.49144322e-01 -4.87150490e-01 3.77502471e-01 -1.32418549e+00
1.62019745e-01 -5.54525077e-01 3.46414268e-01 -2.54233889e-02
3.18854488e-02 4.60492015e-01 3.19622219e-01 2.28946343e-01
-5.47842801e-01 5.35112560e-01 1.31341028e+00 6.46847934e-02
-1.23304918e-01 2.37277940e-01 -1.64107755e-01 9.83808100e-01
2.02883452e-01 -7.43165791e-01 1.95652340e-02 -3.37496608e-01
2.68547628e-02 5.79619706e-02 7.96097577e-01 -9.29748833e-01
2.44607374e-01 -1.13008946e-01 3.83556098e-01 -1.29343438e+00
9.17731524e-01 -1.15128613e+00 2.65605390e-01 5.08833349e-01
1.08447433e-01 -2.84562558e-01 3.90164018e-01 7.49927640e-01
-6.48733750e-02 -2.90305853e-01 1.07824194e+00 -5.43544851e-02
-5.21837413e-01 3.94929796e-01 -2.90997326e-01 -1.88906848e-01
1.03966963e+00 -4.32873845e-01 1.10845916e-01 -3.10234070e-01
-8.39883864e-01 -4.65839617e-02 8.97186279e-01 3.11240219e-02
6.58443391e-01 -1.31563377e+00 -6.38357341e-01 7.04155192e-02
-7.80293718e-03 2.25713298e-01 1.49327025e-01 9.97187972e-01
-5.57424009e-01 4.48668033e-01 1.92917511e-01 -1.34167039e+00
-1.39029336e+00 3.86566758e-01 3.19304794e-01 -3.42242509e-01
-7.35147178e-01 8.13429952e-01 2.47103080e-01 -3.11054528e-01
3.01687121e-01 -4.17984098e-01 9.96254459e-02 -2.56554633e-01
3.49950731e-01 5.04676461e-01 9.05920658e-03 -8.32086146e-01
-7.09347725e-01 1.22249258e+00 2.93864042e-01 -3.72467160e-01
1.07773221e+00 -2.80429453e-01 7.15160668e-02 4.87361431e-01
8.62877667e-01 2.94455588e-01 -1.71909976e+00 -2.51905680e-01
-2.01632649e-01 -8.58482182e-01 2.78422266e-01 -5.59191465e-01
-8.03646564e-01 5.65598071e-01 7.14128017e-01 -3.19181345e-02
6.93641782e-01 1.94643840e-01 4.13703769e-01 -4.00777310e-02
6.61704063e-01 -5.98664939e-01 -1.69403255e-02 6.21037662e-01
7.10401475e-01 -1.14341354e+00 6.79357231e-01 -1.04737318e+00
-1.86663985e-01 9.29529607e-01 2.52907187e-01 -3.73088986e-01
7.89990187e-01 4.36038733e-01 9.95605253e-03 -3.14645678e-01
-2.34231487e-01 -2.74416208e-01 7.26820171e-01 7.38689482e-01
1.69615716e-01 -2.64503509e-01 -2.92380918e-02 -1.46262750e-01
3.02220374e-01 -1.22552037e-01 7.61809060e-03 1.27747035e+00
-4.84238386e-01 -9.81493652e-01 -1.04785597e+00 -4.10931371e-02
-3.82205546e-01 1.79315075e-01 -1.98038012e-01 8.94599080e-01
-1.00094773e-01 7.76147664e-01 7.44754076e-02 2.77005911e-01
3.52622271e-01 -2.74295181e-01 1.05007493e+00 -5.99601388e-01
-3.94455940e-01 6.70108974e-01 2.10001990e-02 -8.31883430e-01
-9.15614724e-01 -1.06759942e+00 -1.03577566e+00 5.37668727e-02
-6.97146833e-01 -8.88870209e-02 9.69021738e-01 9.26208138e-01
2.44754970e-01 2.41186857e-01 3.17938268e-01 -1.69806993e+00
-4.36501205e-01 -6.01248801e-01 -4.39891160e-01 -4.92507033e-02
3.37298483e-01 -1.25686693e+00 -3.56229663e-01 -1.59372278e-02] | [7.098311901092529, -2.320064067840576] |
61577b08-7c57-48dc-a208-0d3481ca74ae | sacdnet-towards-early-type-2-diabetes | 2301.04844 | null | https://arxiv.org/abs/2301.04844v2 | https://arxiv.org/pdf/2301.04844v2.pdf | SACDNet: Towards Early Type 2 Diabetes Prediction with Uncertainty for Electronic Health Records | Type 2 diabetes mellitus (T2DM) is one of the most common diseases and a leading cause of death. The problem of early diagnosis of T2DM is challenging and necessary to prevent serious complications. This study proposes a novel neural network architecture for early T2DM prediction using multi-headed self-attention and dense layers to extract features from historic diagnoses, patient vitals, and demographics. The proposed technique is called the Self-Attention for Comorbid Disease Net (SACDNet), achieving an accuracy of 89.3% and an F1-Score of 89.1%, having a 1.6% increased accuracy and 1.3% increased f1-score compared to the baseline techniques. Monte Carlo (MC) Dropout is applied to the SACDNet to get a bayesian approximation. A T2DM prediction framework based on the MC Dropout SACDNet is proposed to quantize the uncertainty associated with the predictions. A T2DM prediction dataset is also built as part of this study which is based on real-world routine Electronic Health Record (EHR) data comprising 4,124 diabetic and 181,767 non-diabetic examples, collected from 295 different EHR systems running in different parts of the United States of America. This dataset is further used to evaluate 7 different machine learning and 3 deep learning-based models. Finally, a detailed analysis of the fairness of every technique against different patient demographic groups is performed to validate the unbiased generalization of the techniques and the diversity of the data. | ['Muhammad Kamran Malik', 'Tayyab Nasir'] | 2023-01-12 | null | null | null | null | ['diabetes-prediction'] | ['medical'] | [-5.12953922e-02 1.77304000e-01 -3.85814846e-01 -9.43483770e-01
-8.32304478e-01 3.93976659e-01 2.63830543e-01 4.78248179e-01
-2.47948825e-01 8.75694335e-01 3.14735383e-01 -2.51128048e-01
-4.34569508e-01 -7.49740064e-01 -5.35673201e-01 -5.53508520e-01
-4.31389242e-01 9.37604725e-01 -5.01511991e-01 4.47482526e-01
-1.38861969e-01 1.56664863e-01 -1.18186283e+00 4.01627034e-01
1.25475800e+00 1.46920383e+00 -3.83441120e-01 3.82795453e-01
-7.12411180e-02 7.52373695e-01 -4.82749701e-01 -5.20100534e-01
4.45515841e-01 -3.20570737e-01 -3.20510775e-01 -3.05988431e-01
5.84864497e-01 -6.23001635e-01 -4.46574152e-01 8.42682064e-01
9.60515201e-01 -2.72126794e-01 8.36092174e-01 -1.18295348e+00
-6.69646084e-01 6.17938042e-01 -1.93804845e-01 1.69910088e-01
-1.53487250e-01 3.48136485e-01 5.61808944e-01 -3.68827730e-01
2.92732418e-01 1.20965886e+00 1.06203532e+00 6.04203343e-01
-1.27673113e+00 -8.28702509e-01 -9.44823623e-02 3.62699449e-01
-1.55675304e+00 -4.20628250e-01 1.18682481e-01 -5.41513801e-01
9.84930575e-01 -7.92433918e-02 6.75713658e-01 1.25594866e+00
7.56414652e-01 5.67859650e-01 8.61016691e-01 -8.43928754e-02
4.40350622e-01 2.16200769e-01 3.72955024e-01 3.84259820e-01
3.33753437e-01 5.51425219e-01 8.11863095e-02 -5.01878202e-01
7.99855232e-01 5.91027617e-01 4.22959067e-02 -5.92506602e-02
-7.89628327e-01 9.37554479e-01 3.82373244e-01 -1.98383152e-01
-7.84840882e-01 -2.33779252e-02 5.04753113e-01 1.68252304e-01
5.73132277e-01 -6.07703701e-02 -8.85524213e-01 1.84906632e-01
-6.99918568e-01 3.15909863e-01 8.85556340e-01 8.18170249e-01
-7.33694583e-02 1.25947237e-01 -5.75027764e-01 7.44237185e-01
3.63098115e-01 6.46762967e-01 3.58737677e-01 -7.10590720e-01
1.50688455e-01 7.64448583e-01 -1.69200003e-01 -4.94841635e-01
-6.72999978e-01 -7.49923944e-01 -1.42323029e+00 4.77796607e-02
2.23547250e-01 -4.98987675e-01 -1.35798466e+00 1.62140620e+00
1.50578797e-01 2.11245015e-01 1.14764743e-01 6.10979617e-01
8.80719662e-01 4.63581413e-01 4.76601839e-01 -1.01283349e-01
1.33601654e+00 -3.77742320e-01 -6.80508375e-01 3.31649274e-01
5.88952720e-01 -1.57291427e-01 2.72403657e-01 4.07767862e-01
-9.14949298e-01 -2.66070336e-01 -4.39488769e-01 2.35831533e-02
-2.14974076e-01 9.18584242e-02 6.00967586e-01 5.97705305e-01
-5.54304063e-01 7.40633845e-01 -9.11619067e-01 -5.49932301e-01
1.11216331e+00 4.32070196e-01 -2.01316122e-02 -4.64561045e-01
-1.36134017e+00 9.79359567e-01 3.09363186e-01 -1.49483606e-01
-9.71550167e-01 -1.46501839e+00 -6.20352089e-01 2.35757127e-01
-5.40709794e-02 -1.16410160e+00 1.05473661e+00 -8.21093917e-01
-1.12608588e+00 6.30615294e-01 8.11732709e-02 -9.40789223e-01
7.23605692e-01 -3.32666427e-01 -6.56184435e-01 -7.30370730e-02
-2.05893293e-02 5.83984196e-01 2.93021232e-01 -5.43835998e-01
-7.51628220e-01 -7.84724951e-01 -6.33456290e-01 4.68556993e-02
1.77595258e-01 -5.57445996e-02 -8.31441896e-04 -5.75381160e-01
-3.27026248e-01 -7.47733593e-01 -5.68631351e-01 9.68722850e-02
-4.99540836e-01 -2.33448550e-01 1.17443122e-01 -8.99919093e-01
1.15613389e+00 -1.82425988e+00 -2.56481647e-01 6.60618469e-02
5.30627429e-01 2.75936276e-01 2.02606618e-02 5.87610379e-02
-2.90896893e-01 2.13544909e-02 -2.03954265e-01 -8.38288851e-03
-1.20898724e-01 9.15481895e-02 1.94203556e-01 4.09458011e-01
4.95453089e-01 7.98736393e-01 -4.64770526e-01 -1.63197383e-01
3.05778414e-01 8.95055771e-01 -7.46747494e-01 1.42372563e-01
-1.33514300e-01 4.40043032e-01 -3.94090533e-01 7.64553070e-01
6.98550105e-01 -3.85415316e-01 1.64205477e-01 -1.82318956e-01
-6.93114614e-03 3.35330248e-01 -7.69129336e-01 1.19745564e+00
-6.07324168e-02 1.03455119e-01 -4.03450936e-01 -1.03398430e+00
8.65815103e-01 5.01787066e-01 7.97111690e-01 -9.82816875e-01
3.44167769e-01 2.00711921e-01 3.08668345e-01 -6.25185907e-01
-4.68968153e-01 -2.72760510e-01 1.38631389e-01 -1.39976023e-02
8.05748105e-02 7.22616494e-01 -6.84799105e-02 -1.78236499e-01
1.19453287e+00 -2.19023630e-01 5.81365824e-01 -2.53341228e-01
4.42099161e-02 5.15089221e-02 1.12969732e+00 8.17584574e-01
-4.08646524e-01 6.02492154e-01 4.81277406e-01 -1.04478848e+00
-1.09787405e+00 -1.07153559e+00 -8.70609522e-01 1.96872175e-01
-5.95601976e-01 -5.34131378e-02 -2.15249643e-01 -6.58194363e-01
5.73997259e-01 8.93822491e-01 -7.19903767e-01 -4.04322445e-01
-7.95134678e-02 -1.34422374e+00 6.04726374e-01 5.95759451e-01
6.47516251e-01 -7.46605575e-01 -4.06704217e-01 2.85937667e-01
-5.01177879e-03 -5.89206815e-01 1.88618317e-01 3.34574044e-01
-1.12581241e+00 -1.18861628e+00 -9.96885419e-01 -4.20349598e-01
3.14004064e-01 -7.49536574e-01 1.36761963e+00 -4.83262718e-01
-4.71618354e-01 -4.18590009e-02 -1.29515184e-02 -6.79763258e-01
-4.07820255e-01 1.33362919e-01 2.54027426e-01 -1.46021545e-01
1.20581853e+00 -5.18962741e-01 -7.93624580e-01 -2.13209495e-01
-4.45233047e-01 -2.49838829e-01 9.06075358e-01 7.78987348e-01
6.54867768e-01 1.49205312e-01 7.64318228e-01 -8.42067838e-01
1.01452261e-01 -9.92745757e-01 -5.38567483e-01 4.36858535e-02
-1.31510615e+00 -4.97777089e-02 3.76516968e-01 -3.13140959e-01
-6.52464688e-01 -7.50395283e-02 -2.40248576e-01 -5.56120157e-01
-6.80221975e-01 5.90964854e-01 -7.47953579e-02 4.99986410e-01
4.20344830e-01 7.15997964e-02 2.34320864e-01 -8.94992232e-01
-2.42823482e-01 8.54018807e-01 9.84720811e-02 -1.06403725e-02
-2.45517138e-02 -2.22880635e-02 -2.89148861e-03 -5.01266837e-01
-9.65038717e-01 -3.49318907e-02 -1.86730623e-01 3.35399777e-01
8.05760026e-01 -1.18819892e+00 -9.49599862e-01 6.37495279e-01
-6.76677704e-01 -2.77508438e-01 -2.30336815e-01 9.46438134e-01
-2.52466530e-01 -1.92790776e-01 -6.83766603e-01 -8.12201202e-01
-8.22164357e-01 -9.78332698e-01 5.33705890e-01 1.85387671e-01
-2.49412090e-01 -9.76905644e-01 1.78665861e-01 1.83009490e-01
7.51366675e-01 3.60020697e-01 1.54520011e+00 -1.18295836e+00
-2.52112865e-01 -3.10583591e-01 -5.96562386e-01 5.48183739e-01
2.01790527e-01 -8.94077122e-02 -6.79572403e-01 -1.62334725e-01
-1.61656048e-02 -6.23461902e-02 8.88559997e-01 1.37084436e+00
9.41218615e-01 -4.20918703e-01 -3.09859008e-01 6.72602415e-01
1.57908607e+00 6.13303840e-01 8.20117712e-01 1.92848429e-01
5.39365888e-01 3.27333778e-01 6.69506863e-02 8.24430227e-01
7.55561054e-01 3.84265125e-01 4.02103186e-01 -1.96938530e-01
2.95430738e-02 2.05012754e-01 9.27752182e-02 4.30450648e-01
9.52535793e-02 -4.63151112e-02 -1.25159645e+00 4.82373744e-01
-1.66862547e+00 -7.64057815e-01 -3.32788527e-01 2.46798992e+00
8.93440902e-01 -2.20282096e-03 1.94576204e-01 -3.70250285e-01
7.85371780e-01 -5.77781975e-01 -1.07236958e+00 -1.73314586e-01
8.45697597e-02 3.55710536e-01 5.84355116e-01 1.34940311e-01
-1.14595032e+00 2.61067241e-01 6.38920450e+00 3.22350591e-01
-1.14065480e+00 -9.82572362e-02 1.14518809e+00 -3.58823776e-01
3.18553567e-01 -4.27368522e-01 -7.97121465e-01 7.54037321e-01
1.48583055e+00 -1.19162522e-01 1.70696631e-01 7.48183429e-01
2.92687178e-01 3.69475447e-02 -1.23717642e+00 8.80480289e-01
-1.76610336e-01 -1.13407576e+00 1.92849442e-01 1.00206234e-01
5.56666613e-01 5.02701938e-01 1.87089950e-01 7.59556293e-01
4.84504879e-01 -1.26817846e+00 1.02505930e-01 9.54222023e-01
1.08549237e+00 -1.04181945e+00 1.25558734e+00 1.12180887e-02
-4.68104541e-01 -3.20850372e-01 -1.41679347e-01 6.09918609e-02
7.78633058e-02 1.07067311e+00 -8.83562028e-01 6.17783129e-01
9.37006712e-01 8.15345585e-01 -2.45523483e-01 1.44284630e+00
4.45360631e-01 7.29252398e-01 -4.39328700e-01 3.84272575e-01
3.70981023e-02 -8.07504877e-02 3.07605743e-01 8.50287795e-01
5.98872066e-01 1.11711822e-01 8.87921974e-02 9.90712106e-01
-3.55690643e-02 4.43535782e-02 -3.68395239e-01 -1.55514389e-01
3.31932604e-01 7.87536860e-01 1.53575584e-01 -5.64783871e-01
-4.23476458e-01 4.90322590e-01 -1.48358062e-01 3.54997605e-01
-8.73865843e-01 -3.05371791e-01 8.73789787e-01 2.56762534e-01
2.45621130e-01 5.50232649e-01 -3.55591416e-01 -1.08558655e+00
-4.06041920e-01 -1.05438650e+00 7.12220967e-01 -2.73027122e-01
-1.90469253e+00 4.32812065e-01 -3.06482136e-01 -9.84640062e-01
-2.00953439e-01 -5.27289927e-01 -1.75716341e-01 1.12720144e+00
-1.61400783e+00 -6.53960764e-01 -2.39545360e-01 4.49885994e-01
1.91168234e-01 -3.95241886e-01 1.17001605e+00 8.77771974e-01
-1.15098047e+00 7.23088205e-01 3.48730773e-01 2.34221354e-01
9.18308973e-01 -1.11972356e+00 1.35324359e-01 1.31151527e-01
-8.31650436e-01 5.38289189e-01 2.53598452e-01 -1.04899776e+00
-9.40266132e-01 -1.74375808e+00 1.15631056e+00 -3.35634202e-01
8.50354210e-02 1.32544547e-01 -8.84419978e-01 7.68511713e-01
-5.24937399e-02 -8.04701969e-02 1.01140153e+00 1.92741618e-01
-9.22226012e-02 -3.66878361e-01 -1.76282609e+00 2.51727104e-01
6.04826033e-01 1.13318175e-01 -3.68444473e-01 1.82026774e-01
5.28339326e-01 -2.60548770e-01 -1.47254503e+00 8.04012001e-01
7.43620515e-01 -7.67970145e-01 8.89892578e-01 -1.07394028e+00
7.65296876e-01 3.03724229e-01 -3.74713093e-01 -1.08268642e+00
-5.64822316e-01 -2.16805488e-01 -3.88733715e-01 9.79525626e-01
3.42821240e-01 -8.04050982e-01 5.86958587e-01 1.12854099e+00
-4.21899138e-03 -7.90959716e-01 -8.84436607e-01 -5.48015654e-01
3.45707297e-01 -3.70577842e-01 5.89455307e-01 1.23636580e+00
-5.05496264e-01 3.34623635e-01 -3.03237140e-01 2.47605756e-01
9.86089826e-01 -1.88219786e-01 2.41660893e-01 -1.82845438e+00
9.50011611e-02 -2.79759467e-01 -6.04303956e-01 -3.66203338e-01
-3.49366635e-01 -9.92494106e-01 -4.99935240e-01 -1.65999484e+00
5.34750283e-01 -6.12100065e-01 -8.39134097e-01 6.77855968e-01
-1.05083967e-02 -2.68832654e-01 -3.33065718e-01 -8.06747302e-02
-2.10312948e-01 6.88251138e-01 7.04336107e-01 -1.20892443e-01
-2.75230706e-01 3.30840468e-01 -6.79602683e-01 4.90058005e-01
8.08010817e-01 -7.41162658e-01 -1.20769024e-01 -2.12465361e-01
-3.20486605e-01 1.72391176e-01 4.37916070e-01 -9.28956509e-01
-9.96479616e-02 -7.03556463e-02 9.05139863e-01 -7.63669610e-01
5.21774217e-02 -9.24851418e-01 3.57470661e-01 9.05831039e-01
-5.71312308e-01 -1.62996069e-01 2.45193779e-01 5.27420819e-01
1.28882632e-01 2.79409736e-01 8.27646136e-01 -1.79229528e-01
-1.53746396e-01 6.90626502e-01 -3.53018969e-01 9.61396620e-02
8.78928721e-01 2.31343135e-01 -8.49553123e-02 -2.44214311e-02
-9.34096515e-01 5.13486385e-01 -3.97530422e-02 2.30144992e-01
5.44070959e-01 -1.45130360e+00 -1.23709154e+00 2.65097290e-01
1.93464473e-01 -7.30686262e-02 5.35169363e-01 1.21730852e+00
-3.09730709e-01 6.21161938e-01 -2.97249228e-01 -7.67116427e-01
-8.69822383e-01 5.68582058e-01 4.94365335e-01 -3.77242625e-01
-1.05096376e+00 4.81070548e-01 5.79741821e-02 -5.36357284e-01
6.44820094e-01 -5.02236545e-01 -8.83061066e-02 5.61246322e-03
7.20982492e-01 5.41229367e-01 1.40984446e-01 -2.78718714e-02
-4.61343795e-01 5.87036535e-02 -4.28198785e-01 4.22194302e-01
1.61468661e+00 -3.82011831e-02 7.13057220e-02 3.05816442e-01
1.00832129e+00 -8.35694194e-01 -1.07359409e+00 -3.30188245e-01
-1.11831918e-01 -9.72268879e-02 4.70186502e-01 -1.47363520e+00
-1.34699273e+00 7.24614382e-01 1.37789345e+00 -1.53781623e-01
1.00323451e+00 -3.97312611e-01 8.68718326e-01 1.75303951e-01
1.59033760e-02 -6.79743826e-01 -6.83531523e-01 3.48009229e-01
4.94263589e-01 -1.52033222e+00 -1.40777603e-01 2.88320810e-01
-7.17968047e-01 8.75212908e-01 3.57048124e-01 -2.02011317e-01
7.55378544e-01 1.91604063e-01 1.99691013e-01 -1.32935181e-01
-1.08014596e+00 3.30095645e-04 1.16124898e-01 6.49255753e-01
4.10351336e-01 1.56327084e-01 -8.38166401e-02 1.04893649e+00
4.13887560e-01 6.86206520e-01 1.01742342e-01 6.32719994e-01
9.53406245e-02 -7.04941034e-01 1.69148203e-02 1.17091763e+00
-8.21872354e-01 -4.19043422e-01 -1.85953051e-01 6.64819062e-01
2.13126153e-01 4.95454103e-01 3.02895129e-01 -1.72776401e-01
2.50024438e-01 5.74191391e-01 3.35446388e-01 -4.49599266e-01
-6.45746410e-01 7.54970461e-02 1.52822599e-01 -5.57388365e-01
-1.19376332e-01 -6.58711255e-01 -1.22804320e+00 -5.06395221e-01
8.78838822e-02 -3.24525625e-01 5.87939978e-01 7.06443906e-01
1.01656163e+00 9.09743607e-01 4.12260443e-01 -3.01502556e-01
-9.71055746e-01 -1.01627207e+00 -6.02415979e-01 2.57450253e-01
4.64020193e-01 -6.01332724e-01 -2.74741083e-01 -2.25330114e-01] | [8.005162239074707, 5.88607931137085] |
f895d73d-2ba2-415e-9532-0f63abbfba0d | learning-latent-sub-events-in-activity-videos | 1605.08140 | null | http://arxiv.org/abs/1605.08140v3 | http://arxiv.org/pdf/1605.08140v3.pdf | Learning Latent Sub-events in Activity Videos Using Temporal Attention Filters | In this paper, we newly introduce the concept of temporal attention filters,
and describe how they can be used for human activity recognition from videos.
Many high-level activities are often composed of multiple temporal parts (e.g.,
sub-events) with different duration/speed, and our objective is to make the
model explicitly learn such temporal structure using multiple attention filters
and benefit from them. Our temporal filters are designed to be fully
differentiable, allowing end-of-end training of the temporal filters together
with the underlying frame-based or segment-based convolutional neural network
architectures. This paper presents an approach of learning a set of optimal
static temporal attention filters to be shared across different videos, and
extends this approach to dynamically adjust attention filters per testing video
using recurrent long short-term memory networks (LSTMs). This allows our
temporal attention filters to learn latent sub-events specific to each
activity. We experimentally confirm that the proposed concept of temporal
attention filters benefits the activity recognition, and we visualize the
learned latent sub-events. | ['Michael S. Ryoo', 'AJ Piergiovanni', 'Chenyou Fan'] | 2016-05-26 | null | null | null | null | ['activity-recognition-in-videos'] | ['computer-vision'] | [ 3.05785686e-01 -9.63835716e-02 -3.14074785e-01 -2.63842613e-01
-2.53238738e-01 -3.30534607e-01 4.98216748e-01 -2.77959853e-01
-4.87837791e-01 5.16404390e-01 5.06695509e-01 1.89839944e-01
-1.94639608e-01 -4.81162369e-01 -1.02012837e+00 -6.63683653e-01
-8.61957252e-01 -1.61406144e-01 5.62080026e-01 2.36306846e-01
2.19474554e-01 5.96217096e-01 -1.57115257e+00 6.15946114e-01
4.04667079e-01 9.91159558e-01 3.81258309e-01 1.03134775e+00
6.15743026e-02 1.27195716e+00 -8.30904961e-01 2.08668783e-01
-6.28376305e-02 -4.72250104e-01 -9.58562136e-01 2.74386168e-01
2.91819662e-01 -3.77882361e-01 -7.60734856e-01 4.51990873e-01
1.03064135e-01 7.72829771e-01 4.12102401e-01 -1.10638988e+00
-8.50976050e-01 4.82060939e-01 -2.48645753e-01 8.65276635e-01
3.70441794e-01 3.70369732e-01 7.37964213e-01 -5.96425474e-01
4.58225220e-01 1.18503368e+00 4.89927649e-01 6.85469329e-01
-9.14204955e-01 -4.19787765e-01 8.48904252e-01 6.56429708e-01
-1.01610363e+00 -2.82900453e-01 6.99156106e-01 -6.36260152e-01
1.40008903e+00 2.65177749e-02 1.03694320e+00 1.45556724e+00
4.94049430e-01 1.04670846e+00 4.68388200e-01 -2.41420180e-01
3.83267738e-02 -4.88061219e-01 1.25754982e-01 6.08083487e-01
-3.08448136e-01 6.56635761e-02 -7.20838666e-01 2.94771641e-01
1.11082757e+00 5.96107125e-01 -5.37569940e-01 -2.78407574e-01
-1.45139670e+00 4.54826176e-01 4.88155663e-01 6.22275710e-01
-5.20974040e-01 6.72757387e-01 4.94813323e-01 2.69013882e-01
1.92544386e-01 1.55449390e-01 -6.86926603e-01 -3.06316018e-01
-7.07469940e-01 -1.83978945e-01 3.05229783e-01 7.64578462e-01
4.85560149e-01 1.71999931e-01 -8.64629149e-01 6.06252670e-01
8.88010338e-02 4.11056615e-02 9.38890040e-01 -1.02450633e+00
2.97185749e-01 3.01331252e-01 1.84234470e-01 -7.46483147e-01
-2.57594645e-01 -3.43591541e-01 -6.95162714e-01 4.80442308e-02
2.46948987e-01 -5.96410856e-02 -1.06626713e+00 1.85715485e+00
-1.08581573e-01 9.64266658e-01 -2.96098366e-02 7.55703688e-01
3.20899606e-01 9.45462286e-01 3.51752400e-01 -5.12691438e-01
1.24244571e+00 -1.38615298e+00 -9.95296419e-01 -3.82689297e-01
4.09182101e-01 -2.72651643e-01 1.03766787e+00 2.22076595e-01
-1.15467846e+00 -9.70664263e-01 -1.02638805e+00 7.18732700e-02
-4.53900546e-01 1.88192949e-01 3.60602379e-01 2.83099636e-02
-9.01488841e-01 9.29728806e-01 -1.32283616e+00 -3.86608392e-01
3.39703351e-01 3.94910306e-01 -1.94318593e-01 4.60628927e-01
-1.22763145e+00 6.75801933e-01 5.04671752e-01 3.49842221e-01
-1.38970482e+00 -2.88474649e-01 -8.59272897e-01 3.30949515e-01
4.28169757e-01 -5.46359956e-01 1.50547791e+00 -1.59863472e+00
-1.65057290e+00 4.16711509e-01 -3.60617787e-01 -7.42066503e-01
2.06725627e-01 -6.72027528e-01 -5.25793731e-01 4.50564325e-01
-1.68319046e-01 5.66608548e-01 1.07569778e+00 -5.90830326e-01
-8.26972008e-01 1.77481100e-01 3.21733534e-01 3.80337164e-02
-6.30308747e-01 1.87400803e-01 -6.62662089e-01 -1.07271624e+00
-3.90661508e-01 -6.64030790e-01 -1.67504057e-01 -2.84315720e-02
2.06316322e-01 -2.33469009e-01 1.08937097e+00 -6.78444803e-01
1.43399286e+00 -2.10630774e+00 4.08539176e-01 -2.33661681e-01
-6.09352514e-02 1.96537018e-01 -3.12398851e-01 1.43727750e-01
-2.78775781e-01 6.67285845e-02 -1.39705911e-01 -1.58771664e-01
-3.34742486e-01 4.78145391e-01 -1.89713091e-01 2.70005733e-01
5.08086801e-01 9.60335672e-01 -1.05535686e+00 -3.41260374e-01
3.86411458e-01 6.82861447e-01 -3.77722055e-01 5.21980286e-01
-3.24982792e-01 6.43414438e-01 -4.05019611e-01 3.12300742e-01
1.89247169e-02 -3.25896204e-01 1.96399763e-01 -1.15553454e-01
-1.61611930e-01 2.55009919e-01 -9.39571679e-01 1.70805597e+00
-3.80672306e-01 8.77456427e-01 -3.85240197e-01 -1.07878351e+00
4.75696295e-01 5.79328597e-01 8.00236166e-01 -6.57684207e-01
7.76231959e-02 -2.36918181e-01 -5.51589914e-02 -8.79684567e-01
2.99576640e-01 2.80224711e-01 1.29762113e-01 4.30979460e-01
5.51294386e-01 7.94040382e-01 4.43817437e-01 -2.05822021e-01
1.09426320e+00 4.70763266e-01 1.44370899e-01 7.24744722e-02
7.52002478e-01 -5.37700951e-01 7.62004733e-01 6.64472699e-01
-3.08080703e-01 4.37740266e-01 3.73227924e-01 -7.75290191e-01
-8.49269271e-01 -7.90181816e-01 3.90437186e-01 1.47280586e+00
-2.90087834e-02 -3.35578024e-01 -5.85132062e-01 -8.98207128e-01
-4.46782172e-01 3.63258868e-01 -1.02473819e+00 -3.83535832e-01
-1.00783145e+00 -3.08448434e-01 3.19200963e-01 1.07781506e+00
5.00897825e-01 -1.60656869e+00 -1.27215886e+00 5.40078402e-01
-1.45884573e-01 -9.55303133e-01 -9.91977930e-01 4.36456472e-01
-1.02168715e+00 -1.21138763e+00 -9.26078379e-01 -8.41820359e-01
4.33395863e-01 1.79869637e-01 9.91843462e-01 -7.19678327e-02
-3.60799469e-02 6.27110243e-01 -4.71373260e-01 6.66813105e-02
8.63958150e-02 -1.89563520e-02 -1.08150691e-01 5.39901376e-01
2.25505456e-01 -6.54874444e-01 -7.16203630e-01 4.92105693e-01
-1.07517028e+00 -8.92656967e-02 4.70272422e-01 7.66563714e-01
6.50973380e-01 -2.12982833e-01 3.06330860e-01 -4.40999001e-01
5.05185544e-01 -3.55040133e-01 -4.52533454e-01 5.25926292e-01
7.08670989e-02 2.81751364e-01 5.63904822e-01 -1.14515805e+00
-1.07838559e+00 5.38644120e-02 1.09367929e-01 -1.00026977e+00
-8.38525593e-02 5.18658280e-01 2.83239707e-02 1.35032147e-01
4.32430416e-01 3.00700873e-01 -4.19454902e-01 -3.59792382e-01
1.65124491e-01 1.45360023e-01 4.80516493e-01 -4.67325807e-01
3.06908011e-01 3.70510846e-01 -3.71257037e-01 -6.15591466e-01
-8.83369863e-01 -3.65850985e-01 -9.05050993e-01 -6.16337240e-01
1.19215822e+00 -7.07990229e-01 -7.76750803e-01 5.98187268e-01
-1.21535361e+00 -9.01827157e-01 -4.66016233e-01 6.35701537e-01
-8.48728478e-01 1.43715337e-01 -8.55771005e-01 -6.86754942e-01
-5.42390421e-02 -1.01478589e+00 8.99124026e-01 3.73126090e-01
-2.61443287e-01 -1.09268367e+00 1.69486552e-01 -2.69687623e-01
2.07993031e-01 3.12413752e-01 4.11903232e-01 -2.35813618e-01
-7.52731621e-01 1.31662702e-03 1.92948237e-01 3.59425515e-01
1.67601600e-01 3.17615777e-01 -8.96412194e-01 -3.90331894e-01
2.82269213e-02 -1.20479546e-01 9.92368817e-01 8.19887578e-01
1.62069833e+00 -4.85870510e-01 -4.06498969e-01 6.96187854e-01
9.94625986e-01 5.70992053e-01 8.62045228e-01 2.26377591e-01
5.64087451e-01 4.13581550e-01 6.06603503e-01 3.48534793e-01
-1.06766805e-01 7.17202783e-01 3.56874526e-01 3.97609808e-02
3.22178169e-03 -6.56956136e-02 8.41727912e-01 5.92384875e-01
-4.21160340e-01 -3.12344342e-01 -5.10160089e-01 7.02742934e-01
-2.21288681e+00 -1.48608780e+00 3.25330287e-01 2.15498972e+00
3.55731457e-01 3.29740256e-01 2.49086156e-01 -7.61530548e-02
7.97817349e-01 4.06916678e-01 -6.60835385e-01 -4.73588616e-01
4.81782220e-02 2.67064422e-01 3.44912380e-01 3.76577824e-01
-1.24225974e+00 8.11381459e-01 7.05778217e+00 4.58603114e-01
-1.28959215e+00 1.66913494e-01 4.47027564e-01 -3.70576560e-01
1.64004698e-01 -1.84972987e-01 -4.17785972e-01 5.66715658e-01
1.20855629e+00 -8.36091861e-02 3.65002275e-01 5.94132066e-01
5.29794574e-01 1.49614960e-01 -1.24837327e+00 8.93129885e-01
-4.32214476e-02 -1.30781960e+00 1.12925276e-01 -1.20315515e-01
6.72831774e-01 -1.26203716e-01 -5.48812523e-02 4.83439744e-01
1.36657134e-01 -1.01847267e+00 9.08561826e-01 1.05929899e+00
5.04448533e-01 -6.31886125e-01 4.31970656e-01 2.03282461e-01
-1.61729991e+00 -5.57158470e-01 -4.24002372e-02 -1.46308064e-01
2.34672770e-01 -1.50366127e-03 -1.89331502e-01 1.68095842e-01
1.00821042e+00 1.21722782e+00 -4.44986850e-01 1.06620145e+00
-3.64735335e-01 5.13923764e-01 9.96633526e-03 1.61869358e-02
4.10403103e-01 1.44949526e-01 3.60306084e-01 1.43777037e+00
4.13460672e-01 1.13752045e-01 2.98034012e-01 5.95018983e-01
-8.80544856e-02 -4.06289607e-01 -3.43184412e-01 -2.90900469e-01
-1.56217674e-02 1.01803207e+00 -6.28305793e-01 -6.61828756e-01
-6.13915384e-01 1.25594687e+00 4.07438934e-01 6.74615741e-01
-1.26403177e+00 -3.05928826e-01 8.52536321e-01 1.07525952e-01
6.55374825e-01 -3.73202801e-01 4.75017339e-01 -1.39728165e+00
1.19914874e-01 -5.85644364e-01 7.74579406e-01 -9.70683634e-01
-9.79649663e-01 5.78241289e-01 2.41449550e-02 -1.45074058e+00
-2.80060023e-01 -4.91260737e-01 -7.99761117e-01 6.92966580e-01
-1.29427361e+00 -1.00969815e+00 -3.31893474e-01 8.78501475e-01
1.05167389e+00 1.46755964e-01 4.87257957e-01 3.04607719e-01
-4.94418591e-01 4.00335133e-01 -2.09352314e-01 2.78479755e-01
4.44607377e-01 -1.00575995e+00 2.34096289e-01 1.14688563e+00
1.68106079e-01 5.46707571e-01 4.80368018e-01 -5.00103116e-01
-9.93165553e-01 -1.22188163e+00 6.33830488e-01 -2.98988819e-01
6.50052905e-01 -2.35663325e-01 -1.13271689e+00 1.20287776e+00
4.63060319e-01 2.23686233e-01 4.65956658e-01 -2.40640175e-02
-8.23037475e-02 -4.40197773e-02 -6.85116172e-01 5.35372615e-01
1.33330798e+00 -6.32167697e-01 -8.33826065e-01 1.97271436e-01
8.46688032e-01 -3.24615806e-01 -7.14729905e-01 4.62297499e-01
6.90814316e-01 -8.78224671e-01 9.20322537e-01 -9.10801649e-01
2.38636672e-01 -4.29905683e-01 2.68678516e-01 -1.12819815e+00
-8.02929282e-01 -5.57782412e-01 -8.23467374e-01 7.43716776e-01
-2.22947821e-02 -3.56767833e-01 7.95224726e-01 1.77974969e-01
-4.01518941e-01 -5.38165212e-01 -7.25149572e-01 -9.09805596e-01
-3.77027869e-01 -3.00954401e-01 2.67552525e-01 8.18234086e-01
-6.34144843e-02 1.38117418e-01 -7.36080825e-01 2.73994267e-01
6.87252358e-02 5.17853871e-02 3.85235637e-01 -8.04934263e-01
-5.81160903e-01 -5.43816626e-01 -3.57872844e-01 -1.48629582e+00
1.29363880e-01 -4.69577640e-01 8.74405354e-02 -1.38869858e+00
6.88581020e-02 4.31178302e-01 -7.75208592e-01 7.31132627e-01
-4.44181114e-02 -3.87783125e-02 3.01790368e-02 6.84227124e-02
-8.97319794e-01 7.16045082e-01 1.20647645e+00 -3.48158389e-01
-4.09036130e-01 5.49054965e-02 -3.76618430e-02 6.56926394e-01
6.85155630e-01 -3.41151536e-01 -4.63788986e-01 -5.58798671e-01
-1.88723058e-01 2.49309406e-01 4.60046828e-01 -1.43234265e+00
2.86416560e-01 -3.36967021e-01 7.01087058e-01 -5.76379478e-01
3.06565493e-01 -9.36931968e-01 2.93642491e-01 7.10781634e-01
-6.22086108e-01 1.01627387e-01 2.61091113e-01 7.34088182e-01
-4.54909384e-01 9.21490639e-02 6.62938356e-01 -2.79728711e-01
-1.00610864e+00 5.35911918e-01 -8.93002570e-01 -2.12722152e-01
1.16403449e+00 -5.37947357e-01 9.32034180e-02 -2.85330415e-01
-1.41576970e+00 3.81982148e-01 -1.09954411e-02 6.88189149e-01
6.08839273e-01 -1.45787513e+00 -2.50316262e-01 1.29252374e-01
8.98139104e-02 -3.91439229e-01 5.32590330e-01 7.51682341e-01
-2.94615299e-01 4.38845456e-01 -6.74643219e-01 -7.14046359e-01
-1.28587079e+00 1.02323699e+00 7.43010581e-01 -3.79409015e-01
-6.94551766e-01 7.85160422e-01 2.98528790e-01 2.03936756e-01
5.44185042e-01 -8.67131770e-01 -4.68323737e-01 9.13299248e-02
7.09456503e-01 1.53223172e-01 -3.22842717e-01 -5.12649477e-01
-2.47164384e-01 6.28419340e-01 1.25715226e-01 2.05015950e-02
1.39198625e+00 -1.00864835e-01 3.25056821e-01 7.03856528e-01
1.19717848e+00 -4.30003047e-01 -2.01549745e+00 -1.18868358e-01
1.02088027e-01 -4.31200683e-01 -2.77913094e-01 -4.40695614e-01
-1.29503739e+00 1.02167881e+00 7.57624269e-01 1.57419339e-01
1.40542114e+00 -1.15914032e-01 7.18352556e-01 2.91694760e-01
1.80510014e-01 -1.08863914e+00 7.05369592e-01 6.44982755e-01
9.29325521e-01 -6.45250320e-01 -5.15843928e-01 5.22264205e-02
-3.99354309e-01 1.34586775e+00 7.83889890e-01 -2.75031626e-01
5.64620733e-01 5.12429737e-02 -3.86677384e-01 -1.52785927e-01
-1.13274896e+00 -3.88898104e-01 4.01762366e-01 4.76060361e-01
4.92958665e-01 -2.25785002e-01 -1.84834823e-02 3.70104164e-01
6.02024794e-01 3.34794402e-01 3.66707206e-01 1.09637189e+00
-4.92272824e-01 -9.17755365e-01 -2.37286955e-01 3.29320073e-01
-4.97217953e-01 3.11730027e-01 4.93405387e-02 5.59712112e-01
2.99942195e-01 6.20796263e-01 4.32691216e-01 -3.29791129e-01
4.10883278e-01 1.93802416e-01 5.89167893e-01 -5.11446357e-01
-6.41984820e-01 2.40510643e-01 -2.71622956e-01 -8.55153024e-01
-8.09640348e-01 -6.49179161e-01 -1.20465982e+00 1.61913931e-02
6.76029921e-02 1.30986243e-01 -1.26600981e-01 8.66692126e-01
4.02101398e-01 1.22462475e+00 4.23164040e-01 -1.00551248e+00
7.26228878e-02 -1.00192451e+00 -3.18723083e-01 5.86595595e-01
6.58522248e-01 -7.94386148e-01 -1.54900923e-01 6.59054995e-01] | [8.387042045593262, 0.5654478073120117] |
fa539a32-b09d-4e14-861e-29be03a3e703 | vne-an-effective-method-for-improving-deep | 2304.01434 | null | https://arxiv.org/abs/2304.01434v1 | https://arxiv.org/pdf/2304.01434v1.pdf | VNE: An Effective Method for Improving Deep Representation by Manipulating Eigenvalue Distribution | Since the introduction of deep learning, a wide scope of representation properties, such as decorrelation, whitening, disentanglement, rank, isotropy, and mutual information, have been studied to improve the quality of representation. However, manipulating such properties can be challenging in terms of implementational effectiveness and general applicability. To address these limitations, we propose to regularize von Neumann entropy~(VNE) of representation. First, we demonstrate that the mathematical formulation of VNE is superior in effectively manipulating the eigenvalues of the representation autocorrelation matrix. Then, we demonstrate that it is widely applicable in improving state-of-the-art algorithms or popular benchmark algorithms by investigating domain-generalization, meta-learning, self-supervised learning, and generative models. In addition, we formally establish theoretical connections with rank, disentanglement, and isotropy of representation. Finally, we provide discussions on the dimension control of VNE and the relationship with Shannon entropy. Code is available at: https://github.com/jaeill/CVPR23-VNE. | ['Wonjong Rhee', 'Jungwook Shin', 'Duhun Hwang', 'Suhyun Kang', 'Jaeill Kim'] | 2023-04-04 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Kim_VNE_An_Effective_Method_for_Improving_Deep_Representation_by_Manipulating_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Kim_VNE_An_Effective_Method_for_Improving_Deep_Representation_by_Manipulating_CVPR_2023_paper.pdf | cvpr-2023-1 | ['self-supervised-image-classification', 'semi-supervised-image-classification', 'few-shot-image-classification', 'classification'] | ['computer-vision', 'computer-vision', 'computer-vision', 'methodology'] | [ 1.30914733e-01 -1.50099665e-01 -2.48301491e-01 -4.11449075e-02
-2.07844064e-01 -5.97689986e-01 7.45139956e-01 2.58836765e-02
-1.96350932e-01 7.01109052e-01 3.75425160e-01 -1.60681739e-01
-6.69020891e-01 -8.43486845e-01 -3.00506324e-01 -8.03697348e-01
-3.93200547e-01 -1.05895519e-01 -3.82066965e-01 -1.40283257e-01
2.57955492e-01 5.90009391e-01 -1.62009823e+00 7.88682774e-02
9.10136461e-01 8.95646393e-01 -7.16095641e-02 2.57542223e-01
7.22371042e-02 6.06765628e-01 -4.67793435e-01 -2.04833850e-01
2.56995469e-01 -5.47529161e-01 -5.76804519e-01 -3.16520125e-01
-1.54955974e-02 6.78430349e-02 -5.70358396e-01 1.24731302e+00
6.12211764e-01 3.72409016e-01 1.09108889e+00 -1.26809812e+00
-9.63423431e-01 8.01112711e-01 -5.31950176e-01 2.43289456e-01
1.30287886e-01 -6.71090633e-02 1.19083774e+00 -7.21077502e-01
4.42357033e-01 1.03714991e+00 5.51931441e-01 5.11344731e-01
-1.30567825e+00 -9.01594698e-01 -1.33093968e-01 1.82375252e-01
-1.62532234e+00 -3.40019137e-01 8.22096586e-01 -6.31999373e-01
6.28061235e-01 4.23287213e-01 5.99083781e-01 1.16066480e+00
2.26208463e-01 7.64229834e-01 1.22197068e+00 -5.00705600e-01
3.11927438e-01 2.66616512e-02 3.01713347e-01 6.92611873e-01
5.85458398e-01 3.48085850e-01 -4.64299351e-01 -1.75625309e-01
9.11126733e-01 -1.01305336e-01 -4.79537070e-01 -7.37788618e-01
-1.10695148e+00 9.51505363e-01 6.01268291e-01 4.40874130e-01
-2.15272382e-01 1.75218508e-01 3.90071362e-01 3.14886570e-01
3.88903111e-01 8.12746346e-01 4.06871028e-02 -1.50079235e-01
-5.23124218e-01 -1.69836178e-01 6.97219789e-01 7.42676973e-01
6.82712734e-01 2.16367587e-01 -1.67505711e-01 8.08310628e-01
1.55655041e-01 4.66507226e-01 6.80761993e-01 -1.05684519e+00
1.76066473e-01 4.37999517e-01 -3.25397849e-01 -1.02674174e+00
-4.61002916e-01 -7.87505925e-01 -1.44920039e+00 -1.04545742e-01
3.68021689e-02 -4.87367474e-02 -5.79380095e-01 2.01630187e+00
-1.22141197e-01 1.56974494e-01 1.95158020e-01 7.75735080e-01
6.88795269e-01 4.38224554e-01 -1.04996666e-01 -2.01568961e-01
1.07875812e+00 -5.63028693e-01 -7.81625926e-01 -7.58464485e-02
5.29068172e-01 -4.11796719e-01 9.01198506e-01 1.38730496e-01
-8.20008695e-01 -3.04397732e-01 -1.04540646e+00 7.60316253e-02
-2.81157047e-01 1.42389074e-01 9.12801445e-01 7.60278165e-01
-9.38299835e-01 7.11827099e-01 -9.18648720e-01 -2.19734162e-01
3.62939119e-01 2.56776243e-01 -3.92426223e-01 1.36771426e-01
-1.32097733e+00 8.20130944e-01 5.20305097e-01 -1.51802581e-02
-4.31170076e-01 -5.66624880e-01 -7.58401215e-01 2.73037314e-01
-1.05688035e-01 -8.81878436e-01 7.69567013e-01 -7.92607546e-01
-1.44679034e+00 4.92283493e-01 1.75936490e-01 -2.89188296e-01
3.00082505e-01 -1.13271065e-01 -2.64149755e-01 -6.24183118e-02
-2.16321230e-01 3.52187037e-01 4.53071892e-01 -1.04828751e+00
1.01275906e-01 -3.99886876e-01 -1.02454551e-01 2.22049117e-01
-6.86349750e-01 -4.14517850e-01 -5.59290610e-02 -7.19231188e-01
3.47353756e-01 -9.83967781e-01 -1.81005478e-01 -2.27122232e-01
-4.21126395e-01 1.25770807e-01 3.09241891e-01 -2.20773622e-01
1.41678607e+00 -2.65447664e+00 4.30412084e-01 5.18576622e-01
3.33981544e-01 1.80760682e-01 -3.86946559e-01 7.43422091e-01
-1.88945428e-01 3.80752087e-01 -1.89463273e-01 6.03763834e-02
9.82400551e-02 1.65535942e-01 -3.06553751e-01 6.64215684e-01
5.59883416e-02 7.76687682e-01 -9.16550219e-01 -2.95376536e-02
2.49719545e-01 7.66870797e-01 -6.53912365e-01 -1.53680488e-01
2.00291559e-01 4.55789119e-01 -4.97814536e-01 2.73529142e-01
5.90782583e-01 -5.88218153e-01 3.93155396e-01 -2.38607436e-01
1.37003392e-01 3.35323185e-01 -1.20308375e+00 1.33078361e+00
-4.25590158e-01 9.03981268e-01 -2.27191299e-01 -9.67853606e-01
8.11276078e-01 -1.01011686e-01 5.50659359e-01 -5.77768683e-01
2.15476289e-01 6.88860491e-02 2.97152907e-01 -7.71017075e-02
5.49956083e-01 6.00782931e-02 1.24391779e-01 6.12550795e-01
2.81137042e-02 1.23703897e-01 1.88404813e-01 1.25355631e-01
1.06691301e+00 -2.50260323e-01 6.40699863e-01 -4.68523830e-01
2.21117988e-01 -6.04838729e-01 3.67941469e-01 6.20517254e-01
-1.32119730e-01 5.26358724e-01 6.05484009e-01 7.52501283e-03
-9.82893348e-01 -1.13515019e+00 -4.45301533e-01 1.03724682e+00
2.34115258e-01 -5.91194510e-01 -4.33728099e-01 -8.94538090e-02
6.13506027e-02 8.54317844e-01 -6.37121141e-01 -6.75090432e-01
-2.64430314e-01 -9.44069028e-01 7.08327711e-01 5.44302940e-01
5.87694049e-01 -7.22769737e-01 -5.13998389e-01 -2.59105533e-01
-3.53013836e-02 -7.03330576e-01 -1.51518911e-01 2.57552177e-01
-8.73407602e-01 -8.19745719e-01 -5.09027004e-01 -2.76260376e-01
5.64143658e-01 2.20714912e-01 8.43620598e-01 -7.95478746e-02
-1.73607171e-01 4.09004211e-01 -3.17144901e-01 -2.69345883e-02
-3.28733265e-01 2.47891814e-01 1.29182845e-01 -2.81676143e-01
2.01654565e-02 -8.69300008e-01 -6.74143136e-01 2.17670664e-01
-1.13123381e+00 7.19843358e-02 6.42601550e-01 8.03391278e-01
4.10732120e-01 2.63853297e-02 4.06280130e-01 -7.04576850e-01
1.01423812e+00 -4.71938938e-01 -3.54745001e-01 1.58228859e-01
-7.44942129e-01 4.33909476e-01 3.37126851e-01 -5.33269882e-01
-6.74212873e-01 -2.63719857e-01 1.25200033e-01 -3.79096597e-01
2.08054289e-01 7.04780936e-01 2.77278526e-03 1.28788538e-02
8.50695729e-01 3.76824409e-01 2.10060090e-01 -1.95747584e-01
6.66723192e-01 5.02200067e-01 9.29486677e-02 -5.74646592e-01
6.69530928e-01 3.02219421e-01 6.11140840e-02 -8.32786679e-01
-6.30017936e-01 -2.31323659e-01 -4.91531342e-01 7.11303949e-02
3.89519930e-01 -9.02294338e-01 -5.09735644e-01 1.88366935e-01
-7.90074825e-01 -1.62678972e-01 -4.29039001e-01 6.46176815e-01
-5.42151451e-01 3.65219533e-01 -5.86726844e-01 -7.77604103e-01
-3.42983097e-01 -9.56143618e-01 6.35594010e-01 5.90071231e-02
-2.94991344e-01 -1.11528766e+00 1.15106829e-01 3.02181430e-02
5.22550821e-01 2.60636032e-01 1.08934581e+00 -6.47315443e-01
-4.31405753e-01 -1.08552771e-02 -2.35444114e-01 4.46150303e-01
2.17623398e-01 4.96987291e-02 -8.35992992e-01 -4.80238736e-01
-1.94914699e-01 -4.11007330e-02 1.11101043e+00 3.69159222e-01
1.07781589e+00 -4.48764861e-01 -1.89853370e-01 8.47515166e-01
1.22687244e+00 4.04113308e-02 6.93990171e-01 3.01646918e-01
4.73227680e-01 3.53855580e-01 1.23907521e-01 8.49104166e-01
7.56505430e-02 3.80300909e-01 2.74092287e-01 2.24820569e-01
-9.88575742e-02 -2.68241055e-02 3.00243229e-01 1.08482921e+00
-2.12735474e-01 -2.37351134e-01 -9.87801969e-01 1.07402347e-01
-1.75674164e+00 -1.00227964e+00 1.22273147e-01 2.19135952e+00
6.21189892e-01 -2.16672495e-01 -3.42297629e-02 1.15851700e-01
7.66839087e-01 4.45324004e-01 -5.64014375e-01 -9.63555425e-02
-2.81691641e-01 -4.20859717e-02 5.06667554e-01 3.81124794e-01
-9.30020809e-01 7.01340258e-01 6.43978930e+00 8.22768450e-01
-1.12102258e+00 1.01308994e-01 5.01250744e-01 -7.49659762e-02
-6.82161391e-01 -2.10710555e-01 -4.17047650e-01 3.74991298e-01
7.11031675e-01 -3.94588470e-01 7.28172898e-01 7.57558227e-01
-1.57228440e-01 2.54831046e-01 -1.18256664e+00 1.10465252e+00
4.94191758e-02 -1.25515342e+00 4.34572361e-02 2.64351577e-01
8.35788786e-01 1.40302688e-01 4.60520089e-01 3.34134191e-01
2.34763607e-01 -1.08422446e+00 3.62446249e-01 5.13715565e-01
7.21752167e-01 -7.74688005e-01 5.81492484e-01 1.68727219e-01
-1.06247699e+00 -2.18484804e-01 -5.22618055e-01 -7.69538730e-02
-3.41806382e-01 7.79180646e-01 -3.07191581e-01 5.93688846e-01
1.95204407e-01 8.16055417e-01 -6.13639116e-01 9.89452839e-01
-2.63630927e-01 2.53888935e-01 -2.35207260e-01 -1.51265681e-01
-4.86748777e-02 -4.21831816e-01 5.14853597e-01 1.01931691e+00
5.31562626e-01 1.69564575e-01 -1.33821473e-01 9.83493865e-01
-1.24019831e-01 -3.62032466e-02 -7.07038939e-01 -3.87155294e-01
7.36864984e-01 9.70111668e-01 -5.95649302e-01 -4.36781980e-02
-3.29253301e-02 7.16133833e-01 3.31936181e-01 6.25815213e-01
-7.10753858e-01 -2.50777721e-01 8.68636549e-01 -2.03021755e-03
1.03045098e-01 -3.36154997e-01 -4.13607448e-01 -1.40995193e+00
-1.29637331e-01 -9.40611184e-01 2.75264919e-01 -4.54096884e-01
-1.48321843e+00 6.01124763e-01 1.30694807e-01 -1.27758467e+00
-2.30707482e-01 -6.08629704e-01 -3.88971984e-01 6.25104427e-01
-1.09199333e+00 -6.05350018e-01 -2.43605539e-01 4.33924019e-01
-1.37470901e-01 -4.33123857e-01 9.01119173e-01 2.94761389e-01
-7.52213001e-01 5.94410956e-01 7.22589552e-01 9.39036012e-02
3.84742171e-01 -1.00623572e+00 1.44365191e-01 5.87000906e-01
2.34741971e-01 9.21936929e-01 6.02780819e-01 -3.32010776e-01
-1.55789840e+00 -9.06403482e-01 1.99002832e-01 -6.77932501e-02
9.44326162e-01 -3.42346340e-01 -7.61606753e-01 5.32355726e-01
4.23913449e-02 6.26620054e-02 1.01218939e+00 4.24880296e-01
-6.54982805e-01 -1.05994724e-01 -8.58712018e-01 7.08010197e-01
1.10528994e+00 -7.95525312e-01 -3.04090053e-01 2.90070176e-01
4.68548715e-01 -1.02914289e-01 -1.00885868e+00 5.74115336e-01
6.90850079e-01 -1.05517077e+00 1.09814334e+00 -4.66043293e-01
4.06811208e-01 1.98744368e-02 -4.16092902e-01 -1.45675886e+00
-7.22618639e-01 -4.65481907e-01 -2.14919582e-01 1.12257409e+00
3.21741492e-01 -9.28991556e-01 4.86870587e-01 5.09309649e-01
7.41667598e-02 -7.36643970e-01 -9.25789833e-01 -1.01701486e+00
2.91831553e-01 -2.34886199e-01 5.58783710e-01 1.16809440e+00
3.09424251e-01 3.36885452e-01 -3.31449330e-01 -7.44338497e-04
4.38204616e-01 1.61715463e-01 5.03222704e-01 -1.27557743e+00
-3.29423428e-01 -8.39239001e-01 -5.53618789e-01 -9.91834044e-01
3.33239645e-01 -1.31272113e+00 -4.25362438e-01 -1.39624870e+00
4.05646354e-01 -4.18920904e-01 -7.39386380e-01 3.48261029e-01
7.67211542e-02 -4.96480092e-02 4.20935541e-01 5.50790548e-01
-5.00061393e-01 7.88415968e-01 1.25781608e+00 1.68653317e-02
-2.97065258e-01 -9.02407914e-02 -8.59139204e-01 5.56353152e-01
1.12759411e+00 -3.08051020e-01 -6.19665682e-01 -3.75637621e-01
2.90394843e-01 8.38922933e-02 2.50157952e-01 -9.64191377e-01
9.11156908e-02 -3.22884656e-02 3.46582383e-01 -1.40546337e-01
4.72178310e-01 -5.02455115e-01 2.85751820e-01 5.16518056e-01
-7.19712257e-01 7.10482523e-02 8.86989683e-02 6.18302226e-01
-1.14593089e-01 -2.80401945e-01 6.75140381e-01 5.71836270e-02
-3.43666434e-01 1.96748972e-01 -3.83037716e-01 1.88849568e-01
6.79869592e-01 -2.32145637e-02 -5.86714923e-01 -4.17120010e-01
-4.13908660e-01 -8.30876976e-02 4.20846075e-01 2.24396423e-01
5.99089205e-01 -1.48541653e+00 -4.95314062e-01 3.74491811e-01
-8.73404089e-03 -5.34852386e-01 1.95195630e-01 7.17679381e-01
-3.46220046e-01 3.21739554e-01 -3.98472667e-01 -4.93014097e-01
-8.81838441e-01 4.67561066e-01 1.59258842e-01 -2.21090004e-01
-4.31687593e-01 5.22723317e-01 4.04659986e-01 -3.47497761e-01
1.97665572e-01 -1.94789335e-01 -1.26564011e-01 1.63478926e-01
2.49815106e-01 4.39634413e-01 -2.42533505e-01 -4.40675557e-01
-4.97761399e-01 3.91359925e-01 1.24308914e-01 -1.41210377e-01
1.16319811e+00 -5.92440777e-02 -2.31188804e-01 4.36288238e-01
1.29872239e+00 -1.44388899e-01 -9.65175569e-01 -1.61927998e-01
-3.12942751e-02 -4.33887720e-01 1.13329887e-01 -4.26596463e-01
-1.08023894e+00 8.90065670e-01 6.27640069e-01 3.82912815e-01
1.13576913e+00 -1.13846205e-01 1.29083082e-01 6.53743625e-01
2.38953874e-01 -7.47084677e-01 1.87334388e-01 7.20725596e-01
9.57393050e-01 -9.40981627e-01 1.49241999e-01 -2.78028160e-01
-5.74661553e-01 9.67083335e-01 4.28504944e-01 -8.48174170e-02
9.80246067e-01 1.94860280e-01 -3.02192748e-01 -2.10392866e-02
-6.98321939e-01 -1.68757141e-01 4.81545627e-01 5.37049055e-01
8.13698411e-01 2.69515693e-01 -3.63252908e-01 3.10625792e-01
-5.19315541e-01 -3.95879030e-01 4.56927568e-01 6.13801837e-01
-2.68325806e-01 -9.26363468e-01 -1.04819760e-01 3.90910119e-01
-2.86172062e-01 -3.91006827e-01 -3.55371118e-01 8.94391716e-01
-3.56973678e-01 7.43082345e-01 5.74428290e-02 -4.84842896e-01
-9.04870406e-02 -4.61783037e-02 4.91645187e-01 -5.29472172e-01
-2.43604109e-01 1.42128086e-02 7.10226297e-02 -2.46992871e-01
-3.64364624e-01 -5.08984089e-01 -9.18939471e-01 -4.79523331e-01
-3.61172080e-01 2.60118037e-01 5.85747719e-01 6.71324134e-01
7.18952119e-01 5.21204114e-01 5.38596153e-01 -5.56016147e-01
-7.58484304e-01 -9.71771955e-01 -8.07042181e-01 3.20530087e-01
2.59367466e-01 -7.60116756e-01 -5.81538737e-01 -4.41808194e-01] | [8.555490493774414, 3.9971113204956055] |
3292a283-09d4-48b7-87d1-7181cd3372e3 | explore-the-power-of-dropout-on-few-shot | 2301.11015 | null | https://arxiv.org/abs/2301.11015v1 | https://arxiv.org/pdf/2301.11015v1.pdf | Explore the Power of Dropout on Few-shot Learning | The generalization power of the pre-trained model is the key for few-shot deep learning. Dropout is a regularization technique used in traditional deep learning methods. In this paper, we explore the power of dropout on few-shot learning and provide some insights about how to use it. Extensive experiments on the few-shot object detection and few-shot image classification datasets, i.e., Pascal VOC, MS COCO, CUB, and mini-ImageNet, validate the effectiveness of our method. | ['Rui Zhao', 'Xingyu Zeng', 'Shaobo Lin'] | 2023-01-26 | null | null | null | null | ['few-shot-image-classification', 'few-shot-object-detection'] | ['computer-vision', 'computer-vision'] | [-2.74839073e-01 -3.16098928e-01 -4.99678224e-01 -5.12079537e-01
-3.74865890e-01 -2.94362940e-02 4.57498878e-01 -2.55340636e-01
-6.79402649e-01 6.53779328e-01 -6.52968213e-02 2.17546478e-01
1.86695233e-01 -8.33183050e-01 -7.87596703e-01 -5.78895450e-01
1.60899177e-01 -1.28268912e-01 7.41709232e-01 -1.26052633e-01
4.22649533e-02 2.39727274e-01 -1.55406737e+00 3.63481164e-01
3.89086246e-01 1.06168866e+00 1.58680394e-01 5.55083811e-01
-1.39798060e-01 1.37138724e+00 -6.00630760e-01 -4.14867431e-01
3.60276029e-02 -5.45342028e-01 -5.26854813e-01 4.75004725e-02
3.60475510e-01 -8.30013633e-01 -9.60191369e-01 1.24097717e+00
5.43604136e-01 7.15713918e-01 7.04383612e-01 -1.39307499e+00
-7.93894291e-01 5.77235758e-01 -4.17208403e-01 8.51362646e-01
-2.76225388e-01 5.70568144e-01 8.69608879e-01 -1.07143521e+00
7.99212039e-01 1.18757081e+00 5.98845899e-01 9.28399146e-01
-7.57845402e-01 -6.21921182e-01 1.04691843e-02 5.57741761e-01
-1.27093971e+00 -4.78660762e-01 5.62741816e-01 -3.79921913e-01
1.12381029e+00 -4.32283878e-01 6.44478202e-01 1.39577067e+00
2.89679021e-01 1.16393626e+00 7.45341778e-01 -2.55734563e-01
7.42857158e-01 7.97305256e-02 7.16239095e-01 9.14009631e-01
3.48769486e-01 1.46786585e-01 -7.03637302e-01 3.44642960e-02
5.48291028e-01 7.56172776e-01 -1.16990477e-01 -3.56748492e-01
-2.08232775e-01 1.28877521e+00 5.20468831e-01 2.78194726e-01
-6.50048405e-02 6.04672730e-01 8.37398767e-01 2.91162372e-01
4.76555169e-01 1.88010246e-01 -1.66505910e-02 -2.50233233e-01
-6.84894323e-01 -7.52267390e-02 6.92695498e-01 1.21582472e+00
9.59692478e-01 6.84695661e-01 -3.70063633e-01 1.00723648e+00
-8.12358484e-02 2.43667066e-02 8.53654742e-01 -8.48842859e-01
-5.59436828e-02 2.85109013e-01 -1.55391663e-01 -2.97393829e-01
-6.01786934e-02 -8.47100765e-02 -6.77239001e-01 1.91904128e-01
-1.62385002e-01 -3.10897082e-01 -1.53507078e+00 1.47975326e+00
-7.47123435e-02 6.38166428e-01 5.31603433e-02 8.67738843e-01
1.27672386e+00 4.59107488e-01 2.88924694e-01 -1.16661184e-01
8.35558414e-01 -1.26384413e+00 -9.30915594e-01 -3.87792379e-01
7.13318527e-01 -3.02408617e-02 1.26876569e+00 -6.38900548e-02
-6.77456141e-01 -5.12041390e-01 -1.29728401e+00 -1.93855911e-01
-5.91860890e-01 -3.76653194e-01 8.83424878e-01 5.57713866e-01
-5.16024470e-01 9.72608089e-01 -1.08841014e+00 -4.17691141e-01
1.19682384e+00 -1.28002569e-01 -9.62106064e-02 -6.50435865e-01
-1.27556360e+00 7.52738416e-01 6.30338788e-01 -4.76210266e-01
-1.58826411e+00 -7.79185295e-01 -1.19309092e+00 4.64928418e-01
4.55296159e-01 -3.22301924e-01 1.49272668e+00 -6.39710188e-01
-1.40507030e+00 7.75060534e-01 -3.71835157e-02 -8.03633034e-01
2.87556320e-01 -4.27195787e-01 -2.55078167e-01 4.26151574e-01
4.27889191e-02 7.13561118e-01 9.26613688e-01 -6.83115959e-01
-3.76103163e-01 -1.89883746e-02 1.50728434e-01 -1.69811100e-01
-5.72516918e-01 -1.74518116e-03 -4.08256859e-01 -3.48670810e-01
-5.26649833e-01 -5.43093324e-01 -3.08385730e-01 1.68134108e-01
-1.48183301e-01 -4.73717183e-01 8.91735911e-01 2.52800211e-02
8.95949006e-01 -2.48818922e+00 -3.50615650e-01 -4.60996389e-01
2.22829729e-01 5.15173614e-01 -3.67121071e-01 2.50178963e-01
1.06470302e-01 4.35889000e-03 -2.03341842e-01 -2.59400308e-01
-1.13640167e-02 5.07214904e-01 -2.84020245e-01 6.17548287e-01
3.41332823e-01 1.34321785e+00 -1.15858817e+00 -4.54757899e-01
4.57568794e-01 3.73287857e-01 -4.21748400e-01 2.30910063e-01
-1.91274807e-01 -2.64488578e-01 -2.98397481e-01 8.21032226e-01
3.17788392e-01 -2.94316202e-01 -4.33950603e-01 4.75424789e-02
1.02258109e-01 3.93714085e-02 -6.09932959e-01 1.88511360e+00
-6.01827800e-02 8.52815449e-01 -5.06008148e-01 -8.90737295e-01
6.89791143e-01 2.73575753e-01 1.44062132e-01 -9.65529203e-01
4.75972950e-01 -4.93841916e-02 -1.29141556e-02 -6.14958167e-01
1.79931059e-01 -5.07852077e-01 1.60400152e-01 3.41691643e-01
9.44641232e-01 2.15620827e-02 4.35951203e-01 3.79244804e-01
1.27349114e+00 -1.78503096e-01 3.85747880e-01 -8.52638111e-03
-3.88828248e-01 -8.05092528e-02 8.13992202e-01 1.24988711e+00
-8.28919351e-01 8.13998461e-01 5.61873198e-01 -5.57759404e-01
-9.68291342e-01 -9.59145606e-01 -1.14482678e-02 1.40274847e+00
8.33620578e-02 -4.57129359e-01 -4.95131195e-01 -6.99147046e-01
-3.83294672e-02 1.00174272e+00 -9.57934678e-01 -9.07485366e-01
-2.64411569e-01 -7.29716182e-01 5.19395709e-01 1.00601768e+00
4.70631570e-01 -1.18469262e+00 -9.76183832e-01 1.84865937e-01
4.87880379e-01 -1.18822682e+00 -1.97507486e-01 5.63544750e-01
-9.80211556e-01 -1.19878554e+00 -7.75956988e-01 -7.27312565e-01
3.57884467e-01 6.82539999e-01 9.88864064e-01 9.16096661e-03
-7.73843646e-01 3.53908211e-01 -4.66151714e-01 -5.48017740e-01
1.45579055e-01 -9.38025564e-02 8.03437531e-02 -2.87274092e-01
8.73272955e-01 -5.54619193e-01 -4.67330873e-01 -5.48075326e-02
-9.00783360e-01 -4.45780188e-01 4.41110283e-01 1.08138967e+00
4.83484000e-01 -4.79351766e-02 4.85501379e-01 -1.15166008e+00
6.39295399e-01 -7.15244770e-01 -3.38138938e-01 8.81849602e-02
-4.20511574e-01 -2.53730267e-01 5.40599346e-01 -7.24763155e-01
-1.04604173e+00 -8.23437646e-02 -1.11783020e-01 -1.24779773e+00
-3.99678051e-02 4.96012777e-01 6.15369566e-02 -1.64626360e-01
8.62233281e-01 2.60208160e-01 -3.53520244e-01 -5.88878095e-01
5.44025660e-01 3.20641011e-01 2.61422545e-01 -5.01804277e-02
5.06875336e-01 6.28051102e-01 -3.59638035e-01 -1.15844262e+00
-1.24977279e+00 -6.88454390e-01 -5.27149200e-01 -5.89690134e-02
7.84999788e-01 -1.09481490e+00 -4.66318756e-01 5.06573081e-01
-1.08064413e+00 -5.54114044e-01 -7.25107670e-01 5.58087826e-01
-6.24888837e-01 -1.03185691e-01 -1.12255883e+00 -6.73287570e-01
-2.69967109e-01 -7.51159608e-01 5.78927755e-01 6.02955997e-01
9.11212191e-02 -9.49035883e-01 2.85063088e-01 -1.79768994e-01
3.76748830e-01 -9.94098559e-03 6.34636581e-01 -8.06295753e-01
-2.87422359e-01 -3.45079839e-01 -3.70552719e-01 4.82217431e-01
-2.02707779e-02 -7.09775314e-02 -1.35068667e+00 -3.56455982e-01
2.21212998e-01 -1.14059520e+00 1.67970288e+00 4.37348217e-01
1.44521081e+00 -7.44567439e-02 -7.74871744e-03 8.37266028e-01
1.77305675e+00 4.29380462e-02 8.28333676e-01 -1.91901263e-03
7.76649535e-01 1.06880754e-01 4.84293967e-01 5.21928251e-01
-1.45784691e-01 2.33988799e-02 4.13588732e-01 2.24495694e-01
-3.01674575e-01 -1.87209934e-01 1.79277509e-01 7.14584470e-01
1.14820436e-01 9.58622843e-02 -7.88686752e-01 6.42082214e-01
-1.95488381e+00 -1.34870100e+00 4.23503160e-01 1.72536552e+00
7.03566313e-01 4.23634201e-01 -1.02073394e-01 -2.96199262e-01
5.32879472e-01 5.90821326e-01 -1.01799202e+00 -3.77310365e-01
-1.12029217e-01 3.61776859e-01 4.48160380e-01 8.51166025e-02
-1.15510690e+00 1.38843250e+00 7.20902681e+00 9.34451342e-01
-1.09621751e+00 5.68389058e-01 4.82353210e-01 -7.12639868e-01
2.34210521e-01 -1.56386748e-01 -1.09784377e+00 3.59808773e-01
1.00289571e+00 -4.03798372e-01 3.05424333e-01 1.50997746e+00
-1.80938169e-01 -1.38045132e-01 -1.15908515e+00 9.89567041e-01
3.73275012e-01 -1.64549005e+00 -2.18712687e-02 -4.26143467e-01
9.13283587e-01 5.72761953e-01 1.88754294e-02 1.13069034e+00
5.39283216e-01 -9.51668441e-01 4.21482027e-01 2.42252856e-01
8.21839690e-01 -8.70864511e-01 7.08623409e-01 3.02410096e-01
-8.89921188e-01 -2.37591833e-01 -1.07860303e+00 -2.14469612e-01
2.77241506e-02 4.39471036e-01 -4.13772881e-01 -1.15438536e-01
8.78808796e-01 1.03908741e+00 -5.01531959e-01 1.30502975e+00
-5.44479907e-01 8.91310036e-01 3.57787162e-02 -2.43574679e-01
5.64310610e-01 2.36987993e-01 2.71381766e-01 1.26991415e+00
-1.54936656e-01 3.79738748e-01 1.16373137e-01 1.24906027e+00
-4.81535077e-01 -3.89282227e-01 -9.27278340e-01 -4.32316005e-01
3.47219288e-01 1.16083968e+00 -6.70702398e-01 -7.29258001e-01
-7.94588268e-01 9.76651013e-01 6.68724000e-01 6.60828829e-01
-7.88038135e-01 -8.85842383e-01 7.13072300e-01 -1.63274676e-01
7.95548797e-01 -3.39275934e-02 -1.93133447e-02 -1.40431905e+00
-3.98731709e-01 -5.16133189e-01 4.25400734e-01 -6.88392401e-01
-1.34921122e+00 1.78206459e-01 -7.11292699e-02 -1.03269815e+00
-1.25000507e-01 -5.97642004e-01 -1.36578774e+00 4.26840395e-01
-1.68461144e+00 -7.57199287e-01 -3.14519793e-01 6.42032087e-01
9.70652997e-01 -2.07301646e-01 5.92942119e-01 2.45058641e-01
-8.38575661e-01 5.34155071e-01 2.72499710e-01 5.70002198e-01
4.46675986e-01 -7.99646497e-01 5.11559129e-01 7.87087917e-01
2.14392349e-01 6.68510318e-01 6.06868804e-01 -5.84457338e-01
-1.42612672e+00 -1.31084514e+00 1.09218299e-01 -4.58216593e-02
8.35900187e-01 -4.90377396e-01 -1.30960941e+00 8.51061344e-01
2.55349636e-01 6.50430024e-01 8.70195866e-01 8.56717825e-02
-6.07120037e-01 2.02273309e-01 -1.06975079e+00 4.41857189e-01
1.04258895e+00 -8.24092507e-01 -9.15434301e-01 2.47379452e-01
1.12662852e+00 3.22417431e-02 -3.53993326e-01 1.93432897e-01
4.39610243e-01 -9.66585636e-01 8.69027138e-01 -1.23592389e+00
5.32911360e-01 4.25561994e-01 -2.89156020e-01 -1.45140946e+00
-4.63555634e-01 -2.16495290e-01 -6.88878357e-01 9.55786765e-01
-1.61940992e-01 -2.61391997e-01 9.20948386e-01 5.54755569e-01
-2.90295273e-01 -6.73580885e-01 -8.96573246e-01 -1.30456293e+00
1.88539416e-01 -2.90232897e-01 1.29212037e-01 9.40688670e-01
-1.30147442e-01 4.07801360e-01 -5.66108108e-01 -4.29385811e-01
9.26076651e-01 -1.80226773e-01 5.26510656e-01 -1.13529742e+00
-2.41175592e-01 -3.44861656e-01 -5.32882273e-01 -5.54587781e-01
3.75236452e-01 -7.09357619e-01 1.58612370e-01 -1.37802291e+00
7.07276940e-01 3.68844628e-01 -7.50009477e-01 7.69472778e-01
-2.70936072e-01 3.26031595e-01 1.89429566e-01 7.27030933e-02
-1.20418632e+00 1.10670292e+00 8.99502277e-01 -4.16931897e-01
-1.05817743e-01 -3.31516743e-01 -3.92460376e-01 9.32316303e-01
8.76090169e-01 -9.54637825e-01 -4.54249710e-01 -1.99930266e-01
-2.55715489e-01 -4.62229460e-01 3.69216055e-01 -1.04576361e+00
3.70512098e-01 -2.98670679e-01 4.27830309e-01 -4.10032243e-01
4.88977909e-01 -2.34782591e-01 -7.39875913e-01 7.48776793e-01
-3.32280219e-01 -7.37624228e-01 1.90252438e-01 7.74590969e-01
-2.64083147e-01 -7.16676116e-01 1.35365462e+00 -5.51136136e-01
-1.46033466e+00 7.38329411e-01 -2.95648009e-01 6.48026824e-01
1.01257646e+00 -1.26912996e-01 -5.03128767e-01 -1.25782639e-01
-5.66744864e-01 2.99834162e-01 3.06279778e-01 5.26024461e-01
1.18017530e+00 -1.60392392e+00 -3.93220067e-01 7.69328373e-03
5.85300386e-01 -4.49256301e-01 3.42081100e-01 7.08620965e-01
-2.32947245e-01 1.83930829e-01 -5.13180792e-01 -3.14273089e-01
-9.56023335e-01 8.35065722e-01 3.82522643e-01 3.52542281e-01
-1.06875873e+00 1.04553747e+00 1.57467246e-01 5.53569049e-02
6.20184660e-01 4.18013055e-03 -5.03526628e-02 8.80820900e-02
1.05975175e+00 5.19345701e-01 -3.67543936e-01 2.08254001e-04
-3.85474801e-01 -8.98749009e-02 -5.04073977e-01 1.47115458e-02
1.68336904e+00 1.54299647e-01 2.50615239e-01 1.12929344e+00
1.65443647e+00 -1.02962506e+00 -1.56173956e+00 -3.84277195e-01
-5.50066717e-02 -5.90485871e-01 2.30388612e-01 -2.62705594e-01
-1.08654451e+00 1.50807500e+00 7.03334451e-01 -1.69724628e-01
5.07467091e-01 5.56144640e-02 8.95872116e-01 9.17086184e-01
3.49835068e-01 -1.49619567e+00 5.61120987e-01 7.58754551e-01
4.17651862e-01 -1.74540544e+00 -4.54396941e-02 2.21908420e-01
-6.68242395e-01 8.80162776e-01 9.23035264e-01 -5.79461396e-01
8.49708200e-01 1.55974075e-01 -2.44865879e-01 -3.94058377e-01
-1.05697858e+00 -5.74081719e-01 -4.60251458e-02 5.12348473e-01
1.90551758e-01 -2.83421993e-01 -2.87913680e-02 7.02453256e-01
5.18524110e-01 4.58717138e-01 7.56891727e-01 1.36320686e+00
-1.07620609e+00 -2.55008310e-01 3.08777511e-01 6.27478719e-01
-5.09404361e-01 -2.35055476e-01 -3.46691549e-01 5.12764633e-01
-2.48359427e-01 7.52248466e-01 -7.52290105e-03 -2.02131480e-01
1.63940609e-01 2.65171587e-01 4.56290841e-01 -1.10933185e+00
-3.35327089e-01 -2.91849077e-01 -2.30108812e-01 -8.65603566e-01
2.47831736e-02 -5.24895499e-03 -1.27092850e+00 -3.39132369e-01
-4.01210308e-01 -1.10948287e-01 3.92490953e-01 9.59109724e-01
1.65425435e-01 6.32614434e-01 5.05726874e-01 -6.49809659e-01
-9.00134683e-01 -1.02458727e+00 -1.00924623e+00 6.04783773e-01
3.57078344e-01 -8.75390351e-01 -7.37812161e-01 -2.92087823e-01] | [9.9823637008667, 2.8619766235351562] |
44c4f07b-8b2c-49dd-ae97-3cb47abeb950 | a-design-flow-for-mapping-spiking-neural | 2108.12444 | null | https://arxiv.org/abs/2108.12444v1 | https://arxiv.org/pdf/2108.12444v1.pdf | A Design Flow for Mapping Spiking Neural Networks to Many-Core Neuromorphic Hardware | The design of many-core neuromorphic hardware is getting more and more complex as these systems are expected to execute large machine learning models. To deal with the design complexity, a predictable design flow is needed to guarantee real-time performance such as latency and throughput without significantly increasing the buffer requirement of computing cores. Synchronous Data Flow Graphs (SDFGs) are used for predictable mapping of streaming applications to multiprocessor systems. We propose an SDFG-based design flow for mapping spiking neural networks (SNNs) to many-core neuromorphic hardware with the objective of exploring the tradeoff between throughput and buffer size. The proposed design flow integrates an iterative partitioning approach, based on Kernighan-Lin graph partitioning heuristic, creating SNN clusters such that each cluster can be mapped to a core of the hardware. The partitioning approach minimizes the inter-cluster spike communication, which improves latency on the shared interconnect of the hardware. Next, the design flow maps clusters to cores using an instance of the Particle Swarm Optimization (PSO), an evolutionary algorithm, exploring the design space of throughput and buffer size. Pareto optimal mappings are retained from the design flow, allowing system designers to select a Pareto mapping that satisfies throughput and buffer size requirements of the design. We evaluated the design flow using five large-scale convolutional neural network (CNN) models. Results demonstrate 63% higher maximum throughput and 10% lower buffer size requirement compared to state-of-the-art dataflow-based mapping solutions. | ['Nagarajan Kandasamy', 'Anup Das', 'M. Lakshmi Varshika', 'Shihao Song'] | 2021-08-27 | null | null | null | null | ['graph-partitioning'] | ['graphs'] | [ 2.65518725e-02 -1.11914888e-01 8.73482451e-02 -1.32238567e-01
2.24574983e-01 -3.88212979e-01 1.68252468e-01 1.79750934e-01
-4.50084001e-01 5.56494653e-01 -1.36101753e-01 -1.59931540e-01
-5.40338874e-01 -7.67177105e-01 -5.78340232e-01 -6.36230111e-01
-2.50791609e-01 4.48820204e-01 5.91259062e-01 7.79323429e-02
5.26916444e-01 8.00379038e-01 -2.01650858e+00 4.25812602e-01
3.91653687e-01 1.25397515e+00 6.27215624e-01 4.98801857e-01
-3.65428686e-01 2.94686168e-01 -7.27362156e-01 4.60869446e-02
2.55033076e-01 -4.01902378e-01 -3.61666292e-01 -3.78719777e-01
-1.01421690e-02 2.17315316e-01 -1.48592412e-01 1.03730404e+00
4.57587302e-01 -4.11100388e-01 1.58772126e-01 -1.65002096e+00
-2.69633951e-03 9.35511172e-01 -3.32632124e-01 4.10107195e-01
-3.44765753e-01 -1.27217412e-01 6.49932921e-01 -5.04853547e-01
6.17984891e-01 8.86946738e-01 3.13564956e-01 3.82720292e-01
-1.45649290e+00 -9.57446814e-01 -1.59096003e-01 1.93289250e-01
-1.22439766e+00 -3.67262900e-01 4.37643230e-01 -2.87863255e-01
1.57043433e+00 2.46792555e-01 1.32674372e+00 5.13496101e-01
1.04667842e+00 4.44835126e-02 6.83431745e-01 -1.22153431e-01
1.10874128e+00 -3.49967271e-01 2.56418675e-01 2.61400759e-01
8.55686724e-01 -1.21374428e-02 -1.02863753e+00 -3.56706344e-02
6.46745563e-01 -2.30027646e-01 -1.23063274e-01 -1.65127978e-01
-8.75461996e-01 4.13828194e-01 3.68452638e-01 3.66610467e-01
-3.77617985e-01 6.05507255e-01 5.85011184e-01 1.90353602e-01
-2.77072251e-01 6.82167113e-01 -2.58533746e-01 -3.48863333e-01
-1.25197732e+00 2.89057106e-01 1.13289690e+00 1.18263769e+00
6.00208461e-01 1.32773891e-01 5.76500334e-02 2.10965961e-01
5.10348022e-01 5.24246454e-01 4.40397322e-01 -9.61863279e-01
5.19693010e-02 9.74327087e-01 -4.50973243e-01 -6.00246012e-01
-6.82578146e-01 -6.19001806e-01 -7.18653560e-01 3.09567273e-01
7.13848844e-02 2.51415055e-02 -6.45563662e-01 1.50794411e+00
-5.02391681e-02 -2.90453807e-03 2.49701347e-02 8.03956628e-01
2.30282381e-01 7.70672023e-01 6.97251186e-02 -3.15962225e-01
1.28014731e+00 -5.57849646e-01 -5.83294570e-01 -2.95630634e-01
4.00898665e-01 -5.85843801e-01 4.79307622e-01 1.87049642e-01
-1.20659864e+00 -3.07505965e-01 -1.48420215e+00 3.56607407e-01
-2.24549457e-01 -9.87318680e-02 3.70327413e-01 7.77553737e-01
-1.29864645e+00 6.20966375e-01 -1.36292684e+00 -3.67426246e-01
4.14516568e-01 8.32425833e-01 1.75982460e-01 2.99796820e-01
-5.05605936e-01 6.88953340e-01 5.15609145e-01 -6.72773495e-02
-7.84345329e-01 -1.21318793e+00 -3.05925995e-01 6.92081571e-01
-3.85750771e-01 -5.95600247e-01 8.45942497e-01 -8.24031949e-01
-1.44569981e+00 3.59073609e-01 1.62145168e-01 -6.02324426e-01
-9.99116600e-02 3.58957052e-01 -2.16770008e-01 4.45101634e-02
-2.98996121e-01 8.33218634e-01 3.69558513e-01 -6.60548151e-01
-6.30377948e-01 -3.20852965e-01 -4.24081743e-01 -2.61989236e-01
-5.98993659e-01 8.84705335e-02 -7.75025040e-02 -4.11657505e-02
5.42797484e-02 -9.52590764e-01 -2.30666757e-01 1.83356434e-01
-6.10139631e-02 1.55946806e-01 1.03714967e+00 1.48496598e-01
1.29910398e+00 -2.08439851e+00 2.74160355e-01 4.16025192e-01
1.55851021e-01 1.93942964e-01 -9.89444181e-02 4.34994340e-01
1.85210347e-01 -3.28834116e-01 6.30460083e-02 -3.45386490e-02
-9.49577168e-02 2.25298166e-01 -1.96384341e-01 4.51685280e-01
1.48084164e-01 6.30817592e-01 -3.87290835e-01 -8.41494724e-02
-4.75302227e-02 3.36311877e-01 -7.98956096e-01 3.27742100e-02
-2.94966519e-01 -1.97596252e-02 -1.89892933e-01 4.72824991e-01
7.79074430e-01 -4.61851984e-01 3.84256810e-01 -2.08791912e-01
-6.74256384e-01 -7.16430843e-02 -1.09902263e+00 1.63356113e+00
-2.19763860e-01 7.47413635e-01 1.24726154e-01 -6.76894128e-01
1.45747018e+00 -2.17741624e-01 6.80563331e-01 -1.00907683e+00
4.87344295e-01 4.64473188e-01 5.37678242e-01 -1.92838386e-02
3.53637218e-01 2.68381268e-01 5.10125831e-02 4.93662059e-01
1.24353521e-01 3.57076705e-01 4.42597568e-01 -8.35940614e-02
1.61226857e+00 -1.73285410e-01 -1.79664791e-01 -1.28301382e+00
1.34197116e-01 2.87407219e-01 5.50424695e-01 2.85218954e-01
-1.10407144e-01 6.05429299e-02 8.67730498e-01 -3.65676671e-01
-1.42996657e+00 -1.17949307e+00 -1.32396057e-01 6.96219146e-01
2.54646599e-01 -3.92789900e-01 -1.10346985e+00 7.17564702e-01
-1.37341082e-01 3.12451065e-01 -3.30173105e-01 -3.20563972e-01
-5.57883441e-01 -7.32282639e-01 4.58455116e-01 3.15055937e-01
2.60743022e-01 -1.13978934e+00 -1.98946178e+00 6.36921227e-01
6.56874418e-01 -9.95599687e-01 -1.00822054e-01 1.00597560e+00
-9.06928420e-01 -6.62396848e-01 -6.47517964e-02 -8.00074041e-01
7.44331062e-01 -1.47635669e-01 8.23242486e-01 -9.43416823e-03
-8.32502067e-01 -1.52488351e-01 -1.84229985e-01 -3.54812771e-01
-5.88179566e-02 2.18595266e-01 3.05738747e-02 -2.06986919e-01
3.84727567e-01 -9.34801757e-01 -7.59074926e-01 1.88655809e-01
-9.17834878e-01 3.25123608e-01 5.11249006e-01 6.26253366e-01
9.11592364e-01 4.65085777e-03 7.60505736e-01 -3.02922726e-01
5.69891155e-01 -4.99277145e-01 -1.16597033e+00 1.14047967e-01
-7.01848328e-01 4.16025847e-01 1.00574207e+00 -6.09364212e-01
-4.50375408e-01 3.40873748e-01 3.05852890e-01 -5.40617287e-01
2.82870919e-01 3.33041847e-01 -2.61903793e-01 -3.34506214e-01
6.01893902e-01 -5.31221216e-04 1.18722156e-01 1.14315987e-01
-1.64793476e-01 4.75337654e-01 3.20435554e-01 -4.41260964e-01
5.60255535e-02 4.36629981e-01 5.39619148e-01 -6.86080098e-01
1.58630192e-01 -5.03623113e-03 -3.57371986e-01 -6.02448046e-01
7.06233740e-01 -5.26673317e-01 -1.03976798e+00 3.64805371e-01
-1.12138772e+00 -3.51425737e-01 -5.64310476e-02 3.74303490e-01
-7.81534195e-01 -5.44798434e-01 -4.09025431e-01 -6.08393073e-01
-6.40566766e-01 -1.40797734e+00 6.82309628e-01 7.73088038e-01
-4.18556392e-01 -4.41532671e-01 -1.53624937e-02 -4.15521562e-01
8.68970454e-01 1.31149322e-01 1.17298353e+00 -4.93164539e-01
-1.01263869e+00 2.39842981e-01 -3.87401015e-01 -4.20526206e-01
-3.26073110e-01 1.73877403e-01 -7.69076705e-01 -4.91862983e-01
-8.04645494e-02 1.58219591e-01 3.63203645e-01 6.35840058e-01
9.04494464e-01 -3.43758985e-02 -4.74221855e-01 8.38226557e-01
1.95241845e+00 6.09261215e-01 6.69760406e-01 3.88939828e-01
2.93109387e-01 5.61192214e-01 3.21150944e-02 8.72749686e-01
7.98742007e-03 4.53264236e-01 7.43191183e-01 4.64873224e-01
-8.92844424e-02 4.37746979e-02 3.38541776e-01 9.04593587e-01
7.81328857e-01 -2.83221334e-01 -1.06710243e+00 4.82878506e-01
-1.98119664e+00 -4.98221904e-01 2.79072933e-02 2.04507756e+00
3.65581065e-01 4.09126699e-01 2.01254953e-02 1.57462627e-01
7.21141934e-01 -5.16798079e-01 -9.12789583e-01 -9.00422692e-01
8.41484815e-02 4.85316277e-01 9.09744084e-01 -3.86996195e-02
-1.87181115e-01 5.90168297e-01 5.48586226e+00 4.78791535e-01
-1.54302096e+00 -4.41566482e-02 3.26765001e-01 -7.67135799e-01
-1.92522421e-01 1.05665185e-01 -1.07948935e+00 7.58295894e-01
1.70817888e+00 -5.12321711e-01 6.22357070e-01 5.57787240e-01
3.47977370e-01 -3.72619420e-01 -1.20644307e+00 1.08556259e+00
-3.11613142e-01 -1.90827739e+00 -2.11643472e-01 4.05434966e-01
5.88488460e-01 1.82294145e-01 -2.81462222e-01 -3.36882919e-01
-1.82388753e-01 -8.65970016e-01 1.07407975e+00 5.72029471e-01
5.54924488e-01 -1.12159276e+00 4.39042956e-01 1.66147992e-01
-1.21200514e+00 -4.68887746e-01 -5.69268823e-01 -1.07048497e-01
1.38238415e-01 6.12318873e-01 -3.13694119e-01 -1.29145935e-01
1.05231249e+00 1.54900238e-01 -2.31061459e-01 1.47636175e+00
6.91561520e-01 8.42627883e-02 -4.84629542e-01 -8.25885773e-01
7.65695348e-02 -5.42317331e-02 3.30579042e-01 1.05536830e+00
5.39821863e-01 -9.13610309e-02 -2.47853532e-01 1.42393923e+00
-2.44881622e-02 -9.30139571e-02 -4.57600176e-01 -9.50829834e-02
1.09280980e+00 1.37282133e+00 -1.41136253e+00 2.14662015e-01
-2.37057060e-01 4.10696357e-01 2.47314036e-01 -1.47027716e-01
-6.36270285e-01 -6.97102070e-01 9.95577097e-01 2.40295663e-01
3.69725525e-01 -3.78428429e-01 -1.09282267e+00 -2.46928230e-01
-1.72094047e-01 -3.18027705e-01 -1.94977969e-01 -5.17219901e-01
-7.07215130e-01 9.27888811e-01 -3.15583944e-01 -1.00406313e+00
2.14155391e-02 -5.50694227e-01 -6.48237288e-01 7.66096294e-01
-1.07205939e+00 -6.32848680e-01 -3.09578836e-01 2.32819438e-01
2.46485725e-01 -3.99438024e-01 7.25720525e-01 3.47255707e-01
-5.96147358e-01 5.61143279e-01 4.30949666e-02 -5.77419877e-01
1.20986186e-01 -5.40915072e-01 3.67312670e-01 8.95679772e-01
-3.10111761e-01 4.84303206e-01 7.39829838e-01 -6.77069485e-01
-2.14067483e+00 -1.08990693e+00 6.42037809e-01 3.80559593e-01
8.88992369e-01 -7.97044575e-01 -7.40604699e-01 -4.27749567e-02
2.31263131e-01 -3.81914936e-02 7.58736730e-01 -5.56389928e-01
8.71262252e-02 -4.26422596e-01 -1.12467265e+00 5.36399186e-01
9.22958970e-01 -6.60486743e-02 2.13393033e-01 -1.97476655e-01
4.88516182e-01 -1.40460536e-01 -1.02424622e+00 1.25217482e-01
5.52954018e-01 -7.08285570e-01 2.95581102e-01 -8.18827450e-02
3.31138998e-01 -3.76104504e-01 -2.48926505e-01 -1.07561326e+00
-2.67508596e-01 -6.32701993e-01 -1.03827327e-01 9.38630521e-01
4.63312179e-01 -3.58021110e-01 1.06003225e+00 6.34839773e-01
-5.57544172e-01 -9.35222268e-01 -1.24314415e+00 -9.59399819e-01
-2.82257646e-01 1.69666931e-02 5.57413340e-01 1.70152023e-01
3.75579923e-01 1.90947846e-01 3.64001185e-01 1.32686764e-01
8.12430680e-01 1.93442166e-01 2.06208110e-01 -1.31826854e+00
-1.54660240e-01 -8.59067976e-01 -8.16922784e-01 -3.88465285e-01
4.59129401e-02 -9.68977153e-01 8.53466094e-02 -1.34094441e+00
1.53288767e-01 -4.79260832e-01 -3.98353375e-02 3.76626253e-01
7.57874787e-01 -1.26794308e-01 4.06459063e-01 3.85672227e-02
-4.17638779e-01 2.27590263e-01 9.14813638e-01 3.63470376e-01
-2.31022045e-01 -5.66542208e-01 -4.88953769e-01 9.62982252e-02
8.10833454e-01 -5.75270236e-01 -6.62925065e-01 -7.01598763e-01
4.25244868e-01 2.71739423e-01 1.52583700e-02 -1.66420746e+00
1.12278903e+00 -3.57173309e-02 3.45929682e-01 -6.09622836e-01
1.35483250e-01 -1.04977322e+00 8.40007007e-01 1.04458845e+00
-2.68536627e-01 5.98547161e-01 6.02334619e-01 3.64189178e-01
5.83101511e-02 -1.53273746e-01 1.12656772e+00 4.22024250e-01
-5.66518724e-01 2.55480200e-01 -6.26345158e-01 -2.61862606e-01
1.31297648e+00 -5.13026834e-01 -7.86371589e-01 4.55149204e-01
-5.14351130e-01 2.21823424e-01 3.92468333e-01 3.98631871e-01
7.52699912e-01 -1.05729604e+00 -3.33580106e-01 5.82807720e-01
-1.28202170e-01 -3.54835987e-01 2.27528974e-01 5.53050756e-01
-9.65739727e-01 5.08696079e-01 -1.27762985e+00 -7.32099712e-01
-9.10267651e-01 2.55395770e-01 3.61765444e-01 1.38027653e-01
-4.10272837e-01 6.31803453e-01 -2.48011231e-01 3.07607889e-01
2.29447335e-01 -3.51963013e-01 1.58488259e-01 -1.55994371e-01
5.06300688e-01 4.99293447e-01 3.20082724e-01 -2.03129500e-02
-7.09257066e-01 3.55591476e-01 1.41523987e-01 -1.74513042e-01
1.52916348e+00 2.56577525e-02 -6.13317132e-01 1.54021963e-01
1.14746678e+00 -8.62615645e-01 -1.23864520e+00 4.35838193e-01
5.01546383e-01 -6.65890947e-02 2.52886742e-01 -6.34052157e-01
-1.30707419e+00 6.29658520e-01 8.59735310e-01 1.66885599e-01
1.32901430e+00 -1.66246757e-01 7.74596632e-01 8.58625993e-02
6.43209934e-01 -1.13772798e+00 7.30175972e-02 6.12768888e-01
3.86401325e-01 -1.16955258e-01 -3.17897171e-01 -9.53642353e-02
2.71781143e-02 1.38230205e+00 1.07033181e+00 -4.80423182e-01
9.48148131e-01 1.47895968e+00 -4.79004622e-01 -3.09177846e-01
-1.30179727e+00 3.37833047e-01 -3.68670583e-01 3.69254023e-01
5.16060181e-02 1.78336337e-01 -4.73476708e-01 6.40535831e-01
-3.46359074e-01 1.94278255e-01 6.19619310e-01 1.03120506e+00
-9.35985267e-01 -1.15660095e+00 -3.17932852e-02 7.96531856e-01
-2.60327701e-02 -5.14585003e-02 -7.43695125e-02 2.24957332e-01
3.46903056e-02 6.28448427e-01 8.13811839e-01 -5.72766244e-01
1.41119152e-01 -2.97559977e-01 5.24477363e-01 -5.02152622e-01
-1.00452638e+00 1.52776837e-01 -2.26402923e-01 -6.77590966e-01
1.92157686e-01 -4.83357579e-01 -1.77353561e+00 -5.71971118e-01
-7.13800490e-02 -7.97170252e-02 1.40550458e+00 6.13448620e-01
8.76569211e-01 9.93560672e-01 4.43354100e-01 -7.13932693e-01
-5.76979220e-02 -2.98673540e-01 -6.67073786e-01 -4.56646472e-01
-8.16902965e-02 -4.59434032e-01 3.15785669e-02 -1.66384652e-01] | [8.269933700561523, 2.5619332790374756] |
5d99f319-e37d-45b7-864a-e22af7d30147 | leaving-the-lines-behind-vision-based-crop | 2306.05869 | null | https://arxiv.org/abs/2306.05869v1 | https://arxiv.org/pdf/2306.05869v1.pdf | Leaving the Lines Behind: Vision-Based Crop Row Exit for Agricultural Robot Navigation | Usage of purely vision based solutions for row switching is not well explored in existing vision based crop row navigation frameworks. This method only uses RGB images for local feature matching based visual feedback to exit crop row. Depth images were used at crop row end to estimate the navigation distance within headland. The algorithm was tested on diverse headland areas with soil and vegetation. The proposed method could reach the end of the crop row and then navigate into the headland completely leaving behind the crop row with an error margin of 50 cm. | ['Junfeng Gao', 'Grzegorz Cielniak', 'Rajitha de Silva'] | 2023-06-09 | null | null | null | null | ['navigate', 'robot-navigation'] | ['reasoning', 'robots'] | [ 5.54699861e-02 -1.03383802e-01 -9.44408625e-02 8.00113752e-02
4.45112020e-01 -1.30352163e+00 -1.89968832e-02 4.33417827e-01
-1.65774614e-01 4.79179889e-01 -4.08933789e-01 -9.33118701e-01
4.03745621e-02 -1.07130361e+00 -1.50324881e-01 -4.58343863e-01
7.38244131e-02 -2.52897024e-01 3.30868751e-01 -7.30538785e-01
4.97811526e-01 7.54004240e-01 -1.49116719e+00 -1.34883851e-01
5.15419066e-01 5.50234616e-01 8.48229349e-01 1.06080890e+00
6.10833652e-02 -6.74239360e-03 1.89321786e-02 4.74337101e-01
5.67420185e-01 -2.15303659e-01 -2.23051205e-01 3.50390732e-01
8.62710252e-02 -8.00574720e-01 7.31688663e-02 8.52402329e-01
3.25648218e-01 -1.76499978e-01 5.00858366e-01 -8.06165278e-01
-2.43984297e-01 3.46340179e-01 -1.17738068e+00 -2.61144400e-01
6.99488342e-01 1.45305872e-01 5.70069671e-01 -1.12122238e+00
7.35166132e-01 5.98315239e-01 7.34626114e-01 -3.76084596e-01
-9.01301384e-01 -4.26365793e-01 1.53116301e-01 -2.53046185e-01
-1.22694337e+00 -2.54013181e-01 5.18721938e-01 -2.63927937e-01
8.26595128e-01 1.16455033e-01 9.93502796e-01 6.35441765e-02
3.70233387e-01 1.64482549e-01 1.15452135e+00 -8.25533032e-01
2.60530919e-01 -1.47616342e-01 3.79282475e-01 7.90732920e-01
8.63987803e-01 7.08741963e-01 -4.28546101e-01 2.43165985e-01
8.60733330e-01 -3.41890574e-01 -4.70066518e-01 -1.07430649e+00
-8.30762863e-01 8.35348248e-01 1.03424692e+00 2.11467892e-01
-6.32468939e-01 -1.22569799e-01 2.93489456e-01 7.66252540e-03
-2.69875377e-01 3.53516489e-01 -5.24671435e-01 2.50306457e-01
-9.44933653e-01 2.93925762e-01 7.49117613e-01 1.06132865e+00
7.21665323e-01 8.73653144e-02 3.70971173e-01 3.13099384e-01
6.12835705e-01 1.18195832e+00 -6.27686530e-02 -9.84516680e-01
2.56076872e-01 6.55487776e-01 5.13577878e-01 -1.26626503e+00
-5.65356612e-01 -2.02792451e-01 -3.73723507e-01 9.60790455e-01
5.00772178e-01 -4.81791437e-01 -1.13914931e+00 7.94693649e-01
1.64683029e-01 -6.61562443e-01 3.20801139e-01 8.05956721e-01
6.08954072e-01 6.15239918e-01 -1.76913038e-01 -1.22804143e-01
1.25073159e+00 -5.48481584e-01 -4.73910391e-01 -7.15363264e-01
7.60213375e-01 -8.46445858e-01 7.61268258e-01 1.90885052e-01
-6.59947574e-01 -3.85323375e-01 -1.56548250e+00 3.19448113e-01
-8.12372267e-01 8.50132108e-01 7.81899869e-01 7.45879829e-01
-1.21457052e+00 8.97730961e-02 -8.63353014e-01 -1.08319426e+00
-1.62346110e-01 3.07625197e-02 -6.46331668e-01 -3.38139117e-01
-8.13236475e-01 1.04341340e+00 4.83244628e-01 1.01273763e+00
-4.55958098e-01 -2.77002603e-01 -1.19550276e+00 -1.39480144e-01
7.54028112e-02 -3.66538048e-01 7.98203826e-01 -3.76665682e-01
-1.84065902e+00 1.01947367e+00 -7.33197108e-02 -2.70460010e-01
2.67771453e-01 -3.86799246e-01 1.76836729e-01 -1.00119367e-01
3.17410201e-01 6.56749785e-01 3.98008138e-01 -1.69254959e+00
-8.13499212e-01 -6.43867731e-01 -7.84225762e-02 8.74194205e-01
3.37849498e-01 -7.38111019e-01 6.98501915e-02 -5.61609343e-02
9.21139300e-01 -1.11595595e+00 -3.93757254e-01 3.14742506e-01
-3.22595507e-01 9.16813612e-01 1.17729390e+00 -3.02882373e-01
7.92699754e-01 -1.55792677e+00 -4.86270279e-01 2.86540836e-01
-2.73684740e-01 4.18726027e-01 -4.77566049e-02 7.36984968e-01
2.15620100e-01 -4.65572745e-01 -2.65261084e-01 6.52331531e-01
-7.62938559e-01 1.03485964e-01 -2.32360065e-02 3.64702255e-01
-4.17159170e-01 5.94301522e-01 -8.51944566e-01 1.15485653e-01
7.02763855e-01 2.17399031e-01 8.30512717e-02 -6.35115206e-02
4.44265097e-01 -4.81242947e-02 -2.59753704e-01 1.13805532e+00
1.52944481e+00 5.28271079e-01 3.72668505e-01 -2.07881212e-01
-1.03889775e+00 -5.84125578e-01 -1.16167212e+00 1.59704661e+00
-6.61954165e-01 7.77164400e-01 4.60770607e-01 -6.25491798e-01
1.29701710e+00 -2.80803412e-01 -2.40680370e-02 -3.63450766e-01
-1.54834732e-01 3.34624261e-01 -1.73525929e-01 -1.90016896e-01
8.65305781e-01 5.06337583e-01 1.31700978e-01 -1.33871391e-01
-5.80060780e-01 -3.40336621e-01 2.02788822e-02 -2.61276126e-01
7.69915581e-01 6.65570676e-01 5.90396881e-01 -5.29795706e-01
4.20619339e-01 1.00524426e+00 1.55152930e-02 8.60827625e-01
-3.19789857e-01 3.90617013e-01 -1.68595269e-01 -2.43353575e-01
-7.19163179e-01 -1.18486476e+00 -1.58644646e-01 9.92954195e-01
1.00090992e+00 1.06584929e-01 -5.45296729e-01 -7.98678473e-02
1.21656708e-01 5.15865684e-01 -6.37669861e-01 2.26107076e-01
-3.31420392e-01 -4.29608375e-01 -1.08041922e-02 5.44752359e-01
1.05088127e+00 -1.14117885e+00 -1.81668150e+00 3.23629320e-01
1.70094535e-01 -8.73580694e-01 4.16260988e-01 9.06982005e-01
-1.05000818e+00 -1.13719916e+00 -1.12217259e+00 -5.96388459e-01
7.07820177e-01 1.02400064e+00 7.74199009e-01 -1.19857915e-01
-6.50092304e-01 9.11442339e-02 -6.34141922e-01 -4.38423514e-01
1.04027793e-01 2.66695350e-01 -6.47737324e-01 -1.05598199e+00
3.59822869e-01 8.43492225e-02 -8.55601311e-01 2.67979473e-01
-1.21456280e-01 1.06535442e-01 4.25789744e-01 7.45057106e-01
1.92297727e-01 -2.67384917e-01 -2.00877741e-01 -5.87752402e-01
2.13845849e-01 -3.68609607e-01 -1.17228019e+00 2.57805467e-01
-5.26188731e-01 -3.43397886e-01 3.08321476e-01 2.34682649e-01
-8.74390066e-01 7.61713088e-01 6.19668245e-01 7.02943623e-01
-5.17392993e-01 7.14303076e-01 -5.54585531e-02 -8.32657441e-02
6.38140559e-01 -1.26392677e-01 -1.31282538e-01 9.96421427e-02
5.49538374e-01 2.67832667e-01 8.77441049e-01 3.51928920e-01
9.03708458e-01 4.62703437e-01 4.75461543e-01 -1.19981277e+00
-5.66388965e-02 -3.83786440e-01 -1.11803257e+00 -3.40299726e-01
6.70906007e-01 -9.15946603e-01 -6.63007319e-01 6.50218129e-01
-9.04484928e-01 -3.17368686e-01 3.02218944e-01 5.78679979e-01
8.38704128e-03 3.19654718e-02 -1.28908023e-01 -8.95238221e-01
-4.80708033e-01 -8.18282485e-01 6.79777265e-01 8.14042330e-01
-1.30964797e-02 -8.26712668e-01 -1.56875048e-02 -2.79464930e-01
4.43866253e-01 4.07538652e-01 5.93999743e-01 7.24468112e-01
-1.77160695e-01 -7.41473675e-01 -5.23230255e-01 -5.43000877e-01
3.81044120e-01 5.00119030e-01 -9.60929930e-01 -2.14662209e-01
-7.31371343e-01 -2.56127585e-02 8.57908249e-01 1.17277956e+00
-1.09977266e-02 5.72779536e-01 -8.65039051e-01 7.87252247e-01
1.99981666e+00 7.61197746e-01 4.19919878e-01 8.70305300e-01
3.66302043e-01 6.05395198e-01 1.33535922e+00 5.95888495e-01
2.18364611e-01 1.94305450e-01 9.47953701e-01 -4.24139917e-01
2.07386374e-01 -2.80576110e-01 2.92280942e-01 8.51152167e-02
2.56114334e-01 -6.54675901e-01 -1.15988970e+00 6.22299492e-01
-1.62607825e+00 -5.54210186e-01 -4.83892977e-01 2.18800473e+00
-7.14744180e-02 5.14966249e-02 -5.15882850e-01 3.17976415e-01
4.74471688e-01 -1.49006564e-02 -6.03187144e-01 -7.62115896e-01
-1.33459806e-01 -6.37372350e-03 1.74279594e+00 8.46135616e-01
-1.07918978e+00 1.39770746e+00 6.89882326e+00 -3.30809176e-01
-1.28241956e+00 -7.78983414e-01 -3.41922976e-02 8.64920497e-01
-1.00993020e-02 6.53335690e-01 -9.10299659e-01 -4.60395753e-01
1.88100919e-01 3.64782482e-01 9.12265759e-03 8.22573781e-01
4.38679338e-01 -1.40502310e+00 -3.34136248e-01 6.00643098e-01
-4.28477347e-01 -9.13563311e-01 -5.42468607e-01 -1.38803735e-01
6.26416385e-01 -4.12783563e-01 -4.86573391e-02 -1.27617195e-01
5.43353319e-01 -8.81261468e-01 2.86886007e-01 1.67638481e-01
8.40983927e-01 -5.89478195e-01 6.70863032e-01 2.48665839e-01
-1.68233240e+00 -5.80429994e-02 -5.18639147e-01 -4.79175538e-01
-5.95131665e-02 2.05112785e-01 -9.20162916e-01 7.67277002e-01
7.32002258e-01 5.43219805e-01 -6.93612039e-01 9.75795031e-01
-1.08122543e-01 2.41310298e-01 -6.85656011e-01 4.27456237e-02
8.01970124e-01 -3.94355506e-01 3.01887035e-01 7.66205132e-01
8.14718723e-01 -7.12259933e-02 1.04429394e-01 4.62488890e-01
6.31632507e-01 1.11841768e-01 -1.21061385e+00 4.40652281e-01
8.45495701e-01 1.23191178e+00 -1.15192187e+00 3.98252785e-01
3.11959684e-02 1.30984354e+00 -3.30036283e-01 4.09518719e-01
-2.20935807e-01 -1.00207591e+00 5.37682138e-02 1.38101771e-01
4.69388932e-01 -5.44450641e-01 -6.66778028e-01 -4.56653774e-01
-2.76111454e-01 -1.69484168e-01 -2.84396350e-01 -1.32230484e+00
-1.18799835e-01 1.63893849e-01 7.53191635e-02 -1.02159870e+00
-2.73875475e-01 -1.17807019e+00 -5.68405211e-01 1.16930306e+00
-1.14485347e+00 -1.47409928e+00 -1.13074851e+00 -9.18920785e-02
4.44011956e-01 -1.95492849e-01 1.43445039e+00 -5.56411862e-01
-2.11350396e-01 1.80372536e-01 4.07672912e-01 -2.56608993e-01
4.97027993e-01 -1.14597964e+00 5.85366935e-02 1.13333368e+00
-5.29036045e-01 2.52499312e-01 8.51774514e-01 -8.76587808e-01
-1.47176576e+00 -6.71922863e-01 5.98215878e-01 2.09378943e-01
8.16739947e-02 -1.57853886e-01 -4.48985517e-01 2.59653687e-01
2.90323824e-01 2.18463868e-01 2.21386760e-01 -2.30959773e-01
1.78211868e-01 -1.82525232e-01 -1.24708200e+00 5.77545941e-01
6.79405570e-01 -7.66577423e-02 3.69072318e-01 -2.92037219e-01
-7.16480166e-02 -6.25808299e-01 -2.94050813e-01 4.33632195e-01
1.17164850e+00 -1.06515992e+00 1.00269723e+00 5.17921746e-01
1.78211972e-01 -5.45845330e-01 -4.67200428e-01 -1.30769467e+00
-1.91478327e-01 -1.95206344e-01 4.69114095e-01 1.04342723e+00
4.53640372e-01 -2.84276426e-01 1.03810894e+00 1.53303787e-01
8.28971043e-02 -1.66465327e-01 -1.49435207e-01 -2.78506100e-01
-2.82419436e-02 -7.93865323e-03 1.65152192e-01 3.44348073e-01
-1.35237858e-01 -3.94099839e-02 -4.46604565e-03 8.64009738e-01
6.02314532e-01 3.48717183e-01 5.87208152e-01 -9.78460968e-01
4.88232285e-01 -1.68232873e-01 -4.18666780e-01 -7.88801312e-01
-2.50867039e-01 -5.13142169e-01 4.16404724e-01 -2.05721903e+00
-3.06603700e-01 -1.76385447e-01 3.50547493e-01 3.99175078e-01
-4.38477844e-02 -5.29216416e-02 1.60121068e-01 -7.44210407e-02
6.26223624e-01 -1.38139546e-01 9.38189924e-01 2.27063864e-01
-8.34690630e-01 4.63169739e-02 -3.52731556e-01 7.49338627e-01
8.97444963e-01 -1.46488413e-01 -7.58423805e-01 -4.69983935e-01
6.75751418e-02 6.34511113e-01 4.15931493e-01 -1.17868495e+00
1.43395066e-01 -2.73693711e-01 8.84918749e-01 -1.45991826e+00
9.43661481e-02 -1.18778110e+00 -1.61400080e-01 1.01130676e+00
-1.37279436e-01 4.08867687e-01 5.02595544e-01 5.46206057e-01
1.37466460e-01 -4.60872918e-01 6.75155044e-01 -2.30781049e-01
-1.32807207e+00 -3.38278621e-01 -8.87871981e-01 -6.80178523e-01
1.40014827e+00 -8.58585060e-01 -9.10414159e-02 -2.64160872e-01
-6.46884799e-01 3.33403230e-01 7.35205293e-01 1.51213348e-01
8.62472415e-01 -9.55433488e-01 -2.83095360e-01 5.20359516e-01
4.18136537e-01 -3.83345574e-01 -1.85573429e-01 2.20159471e-01
-1.70007908e+00 7.16324568e-01 -8.18526149e-01 -8.17175567e-01
-1.43995488e+00 2.71599621e-01 3.59786958e-01 3.62312458e-02
-4.90706444e-01 6.27665758e-01 -2.11830437e-01 -4.93973643e-01
-2.11420625e-01 -4.69424933e-01 -3.70744109e-01 -4.76161763e-02
-4.81230952e-02 3.88733029e-01 -4.91125695e-02 -5.38344800e-01
-7.88721204e-01 1.25997174e+00 2.69310594e-01 -4.93628442e-01
8.10894728e-01 -7.33228564e-01 4.38234895e-01 5.27985394e-01
6.87975287e-01 -2.03465402e-01 -1.17835188e+00 3.41982245e-01
2.36773625e-01 -6.04080737e-01 4.98783171e-01 -9.82505202e-01
-7.52649188e-01 8.97419691e-01 1.65314794e+00 -2.17686430e-01
1.18987405e+00 -1.01261091e+00 6.53876811e-02 5.87793350e-01
5.72582126e-01 -9.73379612e-01 -6.85959578e-01 6.73045814e-01
5.43014109e-01 -1.62401390e+00 4.00198340e-01 -3.56683105e-01
-7.28040636e-01 1.71297002e+00 6.98287010e-01 -3.89032245e-01
7.67033219e-01 6.66391850e-01 6.02077365e-01 -1.05186127e-01
-6.48890287e-02 -4.48412359e-01 -3.88957113e-01 1.26878273e+00
8.02262664e-01 9.38903987e-02 -3.34105909e-01 -3.84206921e-01
1.17751189e-01 1.02458172e-01 7.61867821e-01 1.57682335e+00
-9.37060714e-01 -7.64633238e-01 -6.77238226e-01 4.96097654e-02
2.97690094e-01 -1.09298579e-01 -2.27071047e-01 1.11486733e+00
1.45617142e-01 1.13388228e+00 -7.13519752e-02 -2.51743868e-02
3.39382052e-01 -2.99178600e-01 3.95011872e-01 -4.24099684e-01
-4.89064604e-01 1.85509205e-01 -1.89867735e-01 -3.58870655e-01
-1.26856372e-01 -4.88895804e-01 -1.31577325e+00 -2.33964518e-01
-6.07804000e-01 -2.97904938e-01 9.59845960e-01 3.00425470e-01
-3.58357802e-02 -3.97001356e-02 7.71103561e-01 -9.30402696e-01
-6.72719553e-02 -6.24896407e-01 -8.61689568e-01 -6.15186214e-01
4.06632006e-01 -5.52148163e-01 -1.12497151e-01 -5.87687735e-03] | [9.068540573120117, -1.6316707134246826] |
cc5942bc-72ea-4aca-836d-daa4e6fe874d | auto-completion-of-user-interface-layout-1 | 2001.05308 | null | https://arxiv.org/abs/2001.05308v1 | https://arxiv.org/pdf/2001.05308v1.pdf | Auto Completion of User Interface Layout Design Using Transformer-Based Tree Decoders | It has been of increasing interest in the field to develop automatic machineries to facilitate the design process. In this paper, we focus on assisting graphical user interface (UI) layout design, a crucial task in app development. Given a partial layout, which a designer has entered, our model learns to complete the layout by predicting the remaining UI elements with a correct position and dimension as well as the hierarchical structures. Such automation will significantly ease the effort of UI designers and developers. While we focus on interface layout prediction, our model can be generally applicable for other layout prediction problems that involve tree structures and 2-dimensional placements. Particularly, we design two versions of Transformer-based tree decoders: Pointer and Recursive Transformer, and experiment with these models on a public dataset. We also propose several metrics for measuring the accuracy of tree prediction and ground these metrics in the domain of user experience. These contribute a new task and methods to deep learning research. | ['Yang Li', 'Si Si', 'Samy Bengio', 'Xin Zhou', 'Julien Amelot'] | 2020-01-14 | null | https://openreview.net/forum?id=SylWNC4FPH | https://openreview.net/pdf?id=SylWNC4FPH | null | ['layout-design'] | ['computer-vision'] | [ 2.18800843e-01 1.32510429e-02 -1.58715233e-01 -4.20299679e-01
-4.42298293e-01 -6.74693704e-01 1.86816584e-02 1.48849860e-01
1.39052182e-01 2.95789242e-01 3.43915880e-01 -7.82381535e-01
-8.42999965e-02 -7.49617815e-01 -7.01542079e-01 -1.31540492e-01
-5.49885780e-02 4.43251252e-01 3.46884690e-02 6.32212758e-02
3.91829193e-01 3.16935807e-01 -1.38473594e+00 8.72692347e-01
6.17774606e-01 1.17363560e+00 3.22776169e-01 6.48643017e-01
-2.91217953e-01 8.24997187e-01 -5.31404078e-01 -4.76348132e-01
3.06049958e-02 -2.23641887e-01 -9.01390731e-01 -1.68141633e-01
3.49596232e-01 -3.73775512e-01 3.96221280e-02 4.99063194e-01
5.53722322e-01 -2.12519497e-01 6.58570051e-01 -1.04409635e+00
-4.41693217e-01 9.52142358e-01 -6.81519687e-01 -3.32791865e-01
3.94008666e-01 -1.87454388e-01 1.44150150e+00 -7.41546333e-01
5.62829137e-01 9.75504458e-01 8.01948607e-01 3.05020869e-01
-1.15737343e+00 -4.10104632e-01 3.91951799e-01 1.67543799e-01
-1.27482724e+00 -3.23741972e-01 8.44387054e-01 -7.21909285e-01
9.19950485e-01 3.62980872e-01 5.99710107e-01 9.61553335e-01
9.34384763e-02 1.15700364e+00 4.94791806e-01 -5.49337149e-01
5.96074462e-01 -9.83998328e-02 1.38141856e-01 7.73026526e-01
1.15811177e-01 -5.49110949e-01 -3.32715511e-01 -5.48924766e-02
8.77632678e-01 -2.24361662e-02 -1.07290857e-01 -8.45631242e-01
-8.32673490e-01 5.49538612e-01 4.37014639e-01 2.14839801e-01
-5.10325208e-02 2.12556928e-01 3.58410448e-01 2.06221063e-02
1.68135777e-01 8.66089642e-01 -8.76730859e-01 -5.67996621e-01
-8.00093114e-01 3.75965387e-01 1.12427711e+00 1.17367470e+00
3.77340972e-01 -3.23787779e-01 -2.34785050e-01 9.57917035e-01
1.84103042e-01 7.61540979e-02 1.94905314e-03 -7.33905792e-01
6.63131893e-01 8.60806286e-01 1.58188462e-01 -1.02018428e+00
-4.88051355e-01 -3.49792391e-01 -8.55368733e-01 1.63367763e-01
3.49132746e-01 -3.06888431e-01 -6.39897585e-01 1.38093781e+00
-9.52448100e-02 -2.05299661e-01 -6.10649526e-01 6.17290378e-01
6.29947186e-01 5.40539265e-01 -1.30733445e-01 1.32533625e-01
1.25195909e+00 -1.05001450e+00 -5.36318064e-01 -2.96576858e-01
9.41075563e-01 -7.46623039e-01 1.35895467e+00 7.48410821e-01
-9.24064279e-01 -5.98189116e-01 -1.16869342e+00 -2.82228887e-01
-1.49246052e-01 6.12973213e-01 8.46213758e-01 6.83641016e-01
-8.80980492e-01 8.52995157e-01 -7.18546510e-01 -2.55243391e-01
6.26892149e-01 5.43540120e-01 3.58111672e-02 -4.46264893e-02
-4.98313069e-01 5.74549437e-01 -1.47382185e-01 1.33981422e-01
-2.48528644e-01 -8.34464014e-01 -6.41984701e-01 5.20124316e-01
3.41828525e-01 -6.90308034e-01 1.57696474e+00 -5.15133739e-01
-1.16321278e+00 3.82598877e-01 -2.62701452e-01 -1.27328143e-01
2.73922652e-01 -4.38975304e-01 3.21779139e-02 -8.16414058e-01
-1.20316319e-01 5.96134663e-01 5.43619454e-01 -1.24411476e+00
-7.51006603e-01 -2.68561184e-01 1.94533259e-01 1.74741689e-02
-5.66549540e-01 -4.49085504e-01 -7.30844855e-01 -4.32974577e-01
-7.01682344e-02 -8.25800240e-01 -3.20887178e-01 8.84006098e-02
-6.99230909e-01 -1.15331329e-01 6.38508081e-01 -8.33715320e-01
1.87531281e+00 -1.99539423e+00 2.62564182e-01 2.54060984e-01
2.73368210e-01 5.69009595e-02 -4.25717590e-04 4.51812178e-01
6.96213171e-02 3.67968649e-01 2.17296332e-02 -5.00399530e-01
2.01107249e-01 -5.52362055e-02 -3.37595105e-01 -2.19495997e-01
7.73107931e-02 1.05347979e+00 -5.61576486e-01 -1.65516272e-01
1.54309213e-01 1.42902449e-01 -1.10493970e+00 3.14752460e-01
-4.56790239e-01 1.58750743e-01 -4.24091369e-01 5.36455154e-01
4.98369783e-01 -4.47044134e-01 6.50946140e-01 -4.40165043e-01
-3.18346471e-02 6.08256221e-01 -9.63210166e-01 1.85937083e+00
-8.42709661e-01 7.92867362e-01 -2.86734730e-01 -6.23489559e-01
8.58473837e-01 -4.45889644e-02 4.57373321e-01 -8.64394605e-01
2.99731433e-01 -7.89522305e-02 2.44130701e-01 -3.84493530e-01
5.22101939e-01 4.75419998e-01 -2.66287029e-01 5.28977931e-01
-4.34617043e-01 2.22569451e-01 6.85714930e-02 -1.26778454e-01
1.36361289e+00 3.74506295e-01 3.70181739e-01 -1.44485151e-03
5.12516573e-02 -3.49903792e-01 2.07030490e-01 5.47375798e-01
4.65797871e-01 6.48618519e-01 9.67502952e-01 -5.48496842e-01
-1.30798447e+00 -8.72272372e-01 9.27350894e-02 1.19359314e+00
-2.01001406e-01 -1.07719851e+00 -8.82806361e-01 -7.27335453e-01
-3.30508351e-02 5.75885296e-01 -6.78063333e-01 8.90323799e-03
-7.04055548e-01 -2.48265848e-01 2.01528579e-01 8.51073265e-01
4.69640046e-02 -9.52167928e-01 -7.48485386e-01 1.37227163e-01
-1.07878581e-01 -8.35789084e-01 -4.14108902e-01 4.45791662e-01
-8.03845763e-01 -1.03094935e+00 -4.67479050e-01 -8.55969787e-01
6.32068515e-01 -1.48267522e-01 1.29895258e+00 2.33537570e-01
-2.50871480e-01 -1.01189837e-01 -4.04879212e-01 -2.50030398e-01
-1.43149823e-01 8.17321360e-01 -3.19929421e-01 -3.00265878e-01
8.96149650e-02 -6.25164986e-01 -5.70346355e-01 3.23970705e-01
-5.12159884e-01 6.33018970e-01 6.58646047e-01 6.11672282e-01
4.04900193e-01 2.86452770e-01 1.23574264e-01 -1.07457268e+00
6.79786861e-01 -2.67810524e-01 -4.80982661e-01 3.39748085e-01
-3.73005420e-01 4.59286600e-01 7.93964148e-01 -2.37298980e-01
-8.55888546e-01 2.25836873e-01 -5.20390689e-01 -9.90283862e-02
-1.45757586e-01 6.95421040e-01 -6.47171915e-01 3.24320674e-01
4.61331785e-01 -4.29980725e-01 -5.55836439e-01 -1.08336043e+00
3.84275854e-01 6.02899313e-01 4.83998321e-02 -6.68254972e-01
5.91144621e-01 -9.03763622e-02 -7.01599941e-02 -6.16064906e-01
-7.10079193e-01 -1.02824286e-01 -9.42214727e-01 2.71917611e-01
5.98321438e-01 -4.13263977e-01 -8.07160437e-01 8.72016996e-02
-1.28798175e+00 -6.66786790e-01 1.30595835e-02 -3.01071733e-01
-3.66249114e-01 1.43330589e-01 -3.84172916e-01 -7.56399095e-01
-2.67545611e-01 -1.36508882e+00 1.26365602e+00 -1.07178830e-01
-7.61923969e-01 -8.78189981e-01 -4.38952222e-02 3.56573351e-02
2.09529698e-01 -3.99367251e-02 1.64103496e+00 -2.68282533e-01
-8.51575494e-01 -2.28419974e-02 -2.88899004e-01 7.18471482e-02
4.07430798e-01 9.38118175e-02 -9.53875721e-01 1.44737363e-02
-5.58080196e-01 7.48408446e-03 4.99640733e-01 5.17914712e-01
1.71211410e+00 -3.31505895e-01 -7.02518940e-01 7.34790325e-01
1.14352846e+00 6.17026269e-01 7.10038424e-01 2.45666549e-01
1.01275349e+00 5.94248176e-01 5.26141763e-01 6.30624354e-01
4.60660756e-01 9.94308889e-01 2.16220841e-01 -1.64625391e-01
-2.85528544e-02 -7.63857722e-01 -1.26046434e-01 7.37236440e-01
2.46548444e-01 -4.79029685e-01 -9.29591298e-01 2.42064238e-01
-1.90721750e+00 -4.32521760e-01 -6.25780523e-02 2.07825971e+00
8.39433134e-01 3.98032010e-01 1.85924351e-01 3.52088332e-01
2.53419995e-01 -1.23193510e-01 -4.73422319e-01 -5.84404528e-01
4.15314853e-01 5.06083310e-01 2.97402263e-01 2.36691639e-01
-1.18841803e+00 9.64671671e-01 6.04247665e+00 7.63118744e-01
-9.35022652e-01 -4.30438727e-01 1.03487957e+00 9.50409323e-02
-4.54482764e-01 -1.56351998e-02 -7.68819094e-01 3.50481927e-01
6.40000641e-01 2.74849623e-01 6.20839357e-01 1.11702073e+00
6.22203201e-02 -5.69509901e-02 -1.79195642e+00 1.19450259e+00
-3.41738164e-01 -1.58467233e+00 -5.07479645e-02 1.64913282e-01
5.06914616e-01 -6.42215192e-01 2.47778326e-01 2.76222020e-01
3.55620757e-02 -1.21511829e+00 6.76946998e-01 2.82661378e-01
1.02389514e+00 -7.34100163e-01 4.71580297e-01 1.32338375e-01
-1.39900267e+00 -3.02237540e-01 -1.77432910e-01 -4.77141589e-01
3.13952267e-02 3.39375287e-01 -1.13368714e+00 1.20655008e-01
7.33394861e-01 6.92467034e-01 -9.22509134e-01 1.39724600e+00
-6.23513050e-02 4.87405419e-01 -5.36631942e-02 -2.76656628e-01
-2.04123050e-01 1.22468032e-01 -2.19530761e-01 1.01548779e+00
3.95924479e-01 -2.50258058e-01 -1.01653934e-01 9.00021076e-01
-1.26224399e-01 1.94909334e-01 -5.60462117e-01 -2.17821404e-01
6.22155964e-01 1.21853173e+00 -9.00690675e-01 9.68317762e-02
-1.68459103e-01 9.52281952e-01 4.13604856e-01 1.62419975e-01
-8.56644154e-01 -4.05551940e-01 9.08267975e-01 3.41047525e-01
5.60936332e-01 -3.48834425e-01 -1.11246109e+00 -8.99140835e-01
2.08514422e-01 -8.24309349e-01 -8.81656408e-02 -8.03119123e-01
-8.64813685e-01 5.41906118e-01 -2.22929567e-01 -1.26532769e+00
-2.13606805e-01 -8.94784987e-01 -5.38202703e-01 5.15408039e-01
-7.99717903e-01 -1.07815182e+00 -2.47214735e-01 -2.48465449e-01
6.38013840e-01 9.97894630e-02 8.21430385e-01 5.24643660e-01
-5.70302010e-01 9.24486876e-01 -5.78124858e-02 2.74661869e-01
5.03568709e-01 -1.34614122e+00 1.27453578e+00 4.37590182e-01
4.16095763e-01 8.14759731e-01 5.74435115e-01 -5.63676953e-01
-1.53361166e+00 -8.73298228e-01 7.12297142e-01 -7.64539242e-01
4.73387629e-01 -9.47064996e-01 -5.69208026e-01 7.29051828e-01
-9.52694267e-02 -7.58926630e-01 7.20740497e-01 7.34559536e-01
-2.08473206e-01 -1.16536254e-02 -5.06049931e-01 9.74260271e-01
1.47770584e+00 -3.10771078e-01 -1.29957929e-01 -4.51722853e-02
7.26873994e-01 -5.61804116e-01 -7.23757148e-01 2.49344751e-01
1.02516973e+00 -9.75042701e-01 9.26068902e-01 -5.40197253e-01
6.41091287e-01 -1.25497013e-01 1.84251536e-02 -1.19054258e+00
-4.49288011e-01 -4.50668484e-01 -2.93746710e-01 1.29854143e+00
7.41145372e-01 1.98711261e-01 1.24101484e+00 7.41613448e-01
-2.09765688e-01 -1.18203056e+00 -3.28208089e-01 -1.53553814e-01
-2.90263742e-01 -7.90279031e-01 9.04915392e-01 4.67221051e-01
1.67546213e-01 8.12266946e-01 -5.37941635e-01 -1.72628745e-01
1.34951681e-01 2.05523759e-01 1.02748430e+00 -1.49782324e+00
-5.87745190e-01 -6.61508739e-01 -7.30070770e-02 -1.65926468e+00
-6.33447841e-02 -6.74865425e-01 -3.54512855e-02 -1.82388544e+00
-2.10097190e-02 -7.84477770e-01 -1.09426886e-01 6.36538029e-01
1.08523525e-01 -5.54742962e-02 3.75602543e-02 -2.28260025e-01
-7.08009839e-01 3.77482474e-01 1.11639464e+00 -1.80847794e-01
-4.74932373e-01 4.42690879e-01 -1.01761770e+00 7.19066978e-01
5.57195365e-01 -5.16794920e-02 -7.35566854e-01 -7.85045207e-01
6.74793184e-01 1.44654261e-02 1.31043727e-02 -1.04493189e+00
7.86386058e-02 7.04175904e-02 6.82227075e-01 -8.54600012e-01
2.62026370e-01 -9.84590769e-01 -1.28346151e-02 1.61355808e-01
-5.91931224e-01 2.75730938e-01 1.60784796e-01 1.67637169e-01
3.63593370e-01 -1.80446044e-01 2.11918131e-01 2.05507070e-01
-7.13171542e-01 3.55140895e-01 -3.16359907e-01 -2.79126137e-01
5.15475690e-01 -2.33771235e-01 -1.47976559e-02 -3.19959700e-01
-7.79830515e-01 5.04476689e-02 5.40441573e-01 7.64704823e-01
7.30550051e-01 -1.30174041e+00 1.14214100e-01 3.77609342e-01
1.52308986e-01 6.91000819e-02 -4.05115448e-02 4.46997762e-01
-6.73992991e-01 5.79696000e-01 -1.84141174e-01 -3.36349487e-01
-1.33549058e+00 5.71091354e-01 4.87307198e-02 -2.30841234e-01
-4.78684902e-01 8.12961280e-01 4.14816082e-01 -3.10937971e-01
7.34027386e-01 -9.99394298e-01 -2.28731081e-01 -8.39453712e-02
5.33326924e-01 1.67586848e-01 2.24768221e-01 -2.78852750e-02
-2.66172945e-01 4.61631775e-01 -1.31562084e-01 1.83064625e-01
1.19381738e+00 9.51489285e-02 -1.18524402e-01 5.00755489e-01
1.15990841e+00 9.67668444e-02 -1.21112287e+00 1.81346238e-01
5.67803264e-01 -4.61963594e-01 -1.85253277e-01 -9.30869877e-01
-7.61156976e-01 1.11013794e+00 4.84139234e-01 2.56152034e-01
9.39409673e-01 -1.06829181e-01 9.07383144e-01 5.55555642e-01
5.77653766e-01 -8.35360706e-01 9.39329565e-02 5.14909565e-01
6.43990576e-01 -8.07090402e-01 -9.10796151e-02 -4.87976819e-01
-5.15521526e-01 1.11872423e+00 8.39992166e-01 1.81233853e-01
6.67944252e-01 7.21073627e-01 -3.48392457e-01 5.03503978e-02
-8.94493043e-01 -4.20612171e-02 6.54638886e-01 6.66800320e-01
1.06512487e+00 8.94687250e-02 7.34377876e-02 9.91554201e-01
-5.55793285e-01 1.54467583e-01 2.28400260e-01 8.91570628e-01
-3.85943800e-01 -1.71606410e+00 2.17043087e-01 8.29770684e-01
-9.28913429e-02 -2.85203516e-01 -5.88441789e-01 5.10944724e-01
1.97987825e-01 6.44841671e-01 7.53339529e-02 -8.56730342e-01
4.89418060e-01 -6.54683560e-02 6.27609074e-01 -8.49336803e-01
-4.38739210e-01 1.35902286e-01 3.13740462e-01 -4.11419928e-01
4.22036588e-01 -4.50098723e-01 -1.01343858e+00 -2.23902583e-01
-5.52665442e-02 7.96874017e-02 6.14889264e-01 8.00044894e-01
5.60136795e-01 8.09992790e-01 4.28203434e-01 -9.20338094e-01
-1.42189234e-01 -7.69075751e-01 -4.41441327e-01 2.17701718e-02
6.51957095e-02 -6.67957962e-01 4.69293535e-01 5.43997586e-02] | [11.293722152709961, -0.04070484638214111] |
6b5bf6ae-97f0-4c4a-a304-22df70c81a0c | restoration-of-user-videos-shared-on-social | 2208.08597 | null | https://arxiv.org/abs/2208.08597v2 | https://arxiv.org/pdf/2208.08597v2.pdf | Restoration of User Videos Shared on Social Media | User videos shared on social media platforms usually suffer from degradations caused by unknown proprietary processing procedures, which means that their visual quality is poorer than that of the originals. This paper presents a new general video restoration framework for the restoration of user videos shared on social media platforms. In contrast to most deep learning-based video restoration methods that perform end-to-end mapping, where feature extraction is mostly treated as a black box, in the sense that what role a feature plays is often unknown, our new method, termed Video restOration through adapTive dEgradation Sensing (VOTES), introduces the concept of a degradation feature map (DFM) to explicitly guide the video restoration process. Specifically, for each video frame, we first adaptively estimate its DFM to extract features representing the difficulty of restoring its different regions. We then feed the DFM to a convolutional neural network (CNN) to compute hierarchical degradation features to modulate an end-to-end video restoration backbone network, such that more attention is paid explicitly to potentially more difficult to restore areas, which in turn leads to enhanced restoration performance. We will explain the design rationale of the VOTES framework and present extensive experimental results to show that the new VOTES method outperforms various state-of-the-art techniques both quantitatively and qualitatively. In addition, we contribute a large scale real-world database of user videos shared on different social media platforms. Codes and datasets are available at https://github.com/luohongming/VOTES.git | ['Guoping Qiu', 'Kin-Man Lam', 'Fei Zhou', 'Hongming Luo'] | 2022-08-18 | null | null | null | null | ['video-restoration'] | ['computer-vision'] | [ 1.96821526e-01 -2.74979800e-01 -9.93404984e-02 -1.14523470e-01
-5.90323627e-01 -1.94326222e-01 3.91123563e-01 -1.88874185e-01
-7.17891306e-02 6.67548835e-01 7.51452029e-01 4.30374071e-02
9.86545533e-02 -6.97409511e-01 -7.49303997e-01 -7.54057825e-01
-3.83078270e-02 -4.01566148e-01 7.84672201e-02 -3.13195318e-01
2.46455908e-01 3.79342377e-01 -1.73093021e+00 5.67069113e-01
7.71766424e-01 1.08461046e+00 4.12432432e-01 6.76675797e-01
3.22636664e-01 1.09202075e+00 -4.66933042e-01 -2.02976558e-02
3.29235435e-01 -4.29490089e-01 -7.08160400e-01 5.17688036e-01
5.52578747e-01 -7.64138639e-01 -1.00326836e+00 1.04867971e+00
4.74020571e-01 3.47280115e-01 4.53506023e-01 -1.24430025e+00
-9.70478058e-01 2.41968915e-01 -2.89250106e-01 4.85370934e-01
6.18805766e-01 2.08161965e-01 7.33864307e-01 -1.01585877e+00
7.27952719e-01 1.12817037e+00 4.83040035e-01 4.83527958e-01
-9.20727372e-01 -4.09480393e-01 2.57541388e-01 7.22196996e-01
-1.17269695e+00 -7.98392653e-01 8.96609306e-01 -5.20383418e-01
7.24433541e-01 1.76342234e-01 7.24498570e-01 9.21939909e-01
2.79520482e-01 8.86313319e-01 6.62819684e-01 -2.58159935e-01
1.97602943e-01 -4.12015229e-01 -3.30112934e-01 6.03608310e-01
-3.93785179e-01 2.95833517e-02 -7.35015094e-01 1.60767972e-01
7.36465275e-01 3.34896803e-01 -8.23420048e-01 -3.68485600e-01
-1.02988505e+00 4.63678747e-01 5.92222452e-01 7.05996007e-02
-5.96395731e-01 1.28888980e-01 5.22916973e-01 5.79337418e-01
6.72908127e-01 -2.54216403e-01 -3.40921909e-01 1.64635684e-02
-9.76912081e-01 1.29155383e-01 3.02021354e-01 5.30794144e-01
8.48987937e-01 1.78302467e-01 -2.92449474e-01 9.84701276e-01
3.53009850e-01 1.19196894e-02 4.31193054e-01 -1.41089487e+00
1.79061115e-01 3.78222227e-01 1.71784103e-01 -1.08396029e+00
-9.16673541e-02 -2.86424875e-01 -1.02784359e+00 4.69114631e-01
5.08524105e-02 3.88993435e-02 -7.47757792e-01 1.62834418e+00
2.23001197e-01 4.14975882e-01 -1.81473300e-01 1.23617935e+00
8.12180400e-01 7.22760379e-01 -1.43754542e-01 -4.43670422e-01
1.06966341e+00 -1.15163136e+00 -7.50622213e-01 -1.15695775e-01
1.90882966e-01 -8.37240160e-01 1.12798321e+00 5.12552202e-01
-1.18836749e+00 -7.03668416e-01 -1.01661634e+00 -1.72502175e-01
8.35594833e-02 -9.36351866e-02 1.64269432e-01 2.40822196e-01
-1.47240698e+00 8.67875218e-01 -6.21408582e-01 -4.17902857e-01
7.63022304e-01 1.38436839e-01 -4.23955679e-01 -4.51640755e-01
-1.03617704e+00 6.18689060e-01 -1.50742128e-01 1.86050087e-01
-1.41480374e+00 -6.21584892e-01 -8.31751168e-01 3.10459174e-02
3.85947853e-01 -7.87288249e-01 1.15828276e+00 -1.43691576e+00
-1.47422707e+00 5.51868856e-01 -1.21234402e-01 -1.80252373e-01
5.99016547e-01 -3.17340016e-01 -4.62612212e-01 4.57869321e-01
3.95819582e-02 6.32379174e-01 1.16767609e+00 -1.33620334e+00
-5.31666100e-01 -8.25074688e-02 3.22958589e-01 3.30032855e-01
-6.25295520e-01 1.12400599e-01 -6.85172737e-01 -8.88698816e-01
7.61463791e-02 -6.21555388e-01 7.36364946e-02 5.03085017e-01
-1.58563912e-01 1.55613765e-01 9.94745076e-01 -1.00221109e+00
1.14309072e+00 -2.19426823e+00 5.01401961e-01 -2.02220589e-01
4.79561746e-01 9.19556171e-02 -4.20824587e-01 4.67447937e-01
-1.50039122e-01 3.18868309e-02 -2.69739866e-01 -4.70112145e-01
-3.10884774e-01 1.31694347e-01 -2.73221046e-01 7.80014992e-01
-6.57313913e-02 5.94039023e-01 -9.31196451e-01 -2.68404901e-01
4.86015499e-01 8.34490776e-01 -7.37121046e-01 2.80424863e-01
6.04058653e-02 5.89417040e-01 -1.66692272e-01 8.47507477e-01
7.05099523e-01 -7.60870278e-02 7.08056614e-02 -4.21514690e-01
-6.42700195e-02 -1.70840528e-02 -9.11822319e-01 1.72956645e+00
-3.79522145e-01 9.22566772e-01 3.09561282e-01 -1.02287781e+00
5.41518331e-01 4.15835232e-01 6.95621014e-01 -7.14193642e-01
8.06450546e-02 -1.05883367e-01 -4.75395679e-01 -6.63438320e-01
5.91017127e-01 1.26092061e-01 2.70565122e-01 1.44956261e-01
2.00561155e-03 3.78567725e-01 2.60773569e-01 4.03426915e-01
1.35319602e+00 2.78741419e-01 5.53633496e-02 -1.56971768e-01
6.00455821e-01 -4.29111302e-01 8.35995018e-01 3.91499519e-01
-5.72773397e-01 9.69097733e-01 3.50827634e-01 -4.04878229e-01
-1.01228976e+00 -1.06985795e+00 2.09064931e-01 1.07262516e+00
3.96087348e-01 -3.80836219e-01 -7.70882368e-01 -3.80328506e-01
-2.92323172e-01 9.93099511e-02 -5.51790416e-01 -1.27555564e-01
-4.95834619e-01 -4.41127688e-01 -1.88236579e-03 2.85998255e-01
6.38147116e-01 -1.32242477e+00 -2.64785647e-01 2.21105203e-01
-6.02439523e-01 -1.01370406e+00 -8.52195144e-01 -5.37697256e-01
-8.55040491e-01 -1.27184904e+00 -7.79158056e-01 -8.75219584e-01
5.85252047e-01 8.63536119e-01 1.06574297e+00 6.26956701e-01
-1.51869074e-01 5.06460607e-01 -5.61943769e-01 5.43229520e-01
-3.36789459e-01 -4.36264992e-01 1.18892059e-01 3.04375738e-01
-1.47004873e-01 -8.80629718e-01 -1.03904998e+00 2.80026704e-01
-1.24818063e+00 2.01524019e-01 1.84476733e-01 6.92642033e-01
5.38889766e-01 1.83347359e-01 5.07936597e-01 -3.60812873e-01
6.38959646e-01 -6.70186758e-01 -1.90258086e-01 -6.53906818e-03
-3.61718923e-01 -4.86215949e-01 6.15109265e-01 -3.53642017e-01
-9.10187900e-01 -9.22503844e-02 -5.61561324e-02 -6.56341612e-01
-1.12056315e-01 4.92337495e-01 -4.58764106e-01 -1.09639660e-01
4.85523015e-01 2.37190574e-01 1.43578127e-01 -5.68732202e-01
2.13101819e-01 6.88171566e-01 6.39045477e-01 -2.49874175e-01
6.14074230e-01 5.46932459e-01 -4.49469239e-01 -7.76675403e-01
-5.27182162e-01 -2.74598747e-01 -4.59588230e-01 -8.67751896e-01
5.43805182e-01 -1.18854046e+00 -5.11825740e-01 8.07998419e-01
-9.35612500e-01 -5.93384326e-01 -2.22827360e-01 1.61463156e-01
-7.13979423e-01 7.28607178e-01 -1.06741929e+00 -3.91096383e-01
-3.20042163e-01 -1.18516243e+00 8.46789598e-01 2.33390406e-01
1.40147001e-01 -8.19597483e-01 -1.44733474e-01 6.75506055e-01
4.79729980e-01 1.98056325e-01 6.84042513e-01 2.72455454e-01
-7.82660604e-01 1.75895880e-03 -3.09831172e-01 6.99519336e-01
1.91752389e-01 5.37392236e-02 -8.08269203e-01 -5.78411043e-01
2.18269173e-02 -5.45685478e-02 1.05352819e+00 5.89384794e-01
1.31760895e+00 -3.37856531e-01 1.03730860e-03 6.59587443e-01
1.40657341e+00 -2.20777437e-01 1.26917028e+00 4.86098140e-01
7.07961977e-01 3.40056211e-01 5.82714796e-01 5.16429603e-01
6.04062796e-01 6.86660767e-01 8.51810575e-01 -8.17879885e-02
-5.92111707e-01 -8.13286528e-02 9.29106772e-01 8.74274075e-01
-1.45447031e-01 -5.60984552e-01 -4.01230633e-01 6.22486532e-01
-2.02289295e+00 -9.10858274e-01 7.13808015e-02 2.04809189e+00
5.95956981e-01 -1.26809344e-01 2.86182575e-02 3.41072351e-01
9.06993270e-01 3.75388145e-01 -4.19222414e-01 -1.07271401e-02
-2.96847224e-01 -2.72427976e-01 1.55484065e-01 6.07995510e-01
-1.12479579e+00 7.03275442e-01 5.84059095e+00 8.40510190e-01
-8.93285513e-01 3.57547045e-01 7.24061787e-01 -3.29308748e-01
-1.01740792e-01 -2.49932762e-02 -7.71405175e-02 6.28364563e-01
6.72571778e-01 7.86534101e-02 9.09244955e-01 3.91173214e-01
7.41317928e-01 -9.31265876e-02 -8.78974259e-01 9.33467805e-01
1.83098555e-01 -1.44653606e+00 5.99205159e-02 1.03035420e-01
8.07747900e-01 1.31975651e-01 -2.74393009e-03 6.17871322e-02
-5.03955930e-02 -6.14138126e-01 9.11204994e-01 8.31048369e-01
7.69618273e-01 -6.03454471e-01 5.15469074e-01 4.13249955e-02
-1.10126972e+00 -3.35726053e-01 -3.20497274e-01 -2.64052153e-01
4.52061445e-01 7.95270860e-01 -1.52288899e-01 3.80398631e-01
9.83887017e-01 1.19552135e+00 -4.55857933e-01 1.23894072e+00
-3.26949656e-01 4.26448792e-01 2.75044233e-01 7.68458366e-01
-1.72727227e-01 -2.90217977e-02 7.04379976e-01 9.40586448e-01
5.08738399e-01 -2.76462138e-02 2.43346334e-01 4.02889758e-01
-4.50924426e-01 -1.01907952e-02 -2.28556558e-01 2.29592890e-01
2.48397321e-01 1.27212083e+00 -5.19167662e-01 -3.98350924e-01
-5.30415654e-01 1.43714619e+00 1.57106608e-01 6.22133374e-01
-7.10269392e-01 7.04034939e-02 9.93863404e-01 5.13865769e-01
2.78096348e-01 -1.67510390e-01 2.58311689e-01 -1.37390554e+00
2.37463713e-01 -8.82816553e-01 2.99463779e-01 -1.20115101e+00
-1.33166075e+00 5.87211251e-01 -3.97513062e-01 -1.60076392e+00
-2.30671093e-02 -2.51512408e-01 -4.90052581e-01 5.17722785e-01
-1.68854284e+00 -9.46329832e-01 -6.75209403e-01 8.34989548e-01
9.41965282e-01 -1.77418426e-01 4.02112186e-01 7.55167484e-01
-6.02787614e-01 2.89583534e-01 2.36046121e-01 -8.84183496e-03
8.37873876e-01 -7.43709981e-01 9.16167200e-02 1.19436872e+00
-2.43360192e-01 4.44450006e-02 8.08624148e-01 -6.49265707e-01
-1.38657510e+00 -1.23341918e+00 5.05437076e-01 -2.69610249e-02
5.39978921e-01 5.85342869e-02 -9.29531991e-01 4.53494042e-01
3.84170443e-01 4.64653254e-01 4.85562146e-01 -3.23694617e-01
-2.60453194e-01 -2.00425282e-01 -1.18757558e+00 6.55509353e-01
1.33018136e+00 -6.68295622e-01 -1.31968096e-01 4.53490496e-01
7.22653687e-01 -2.18885094e-01 -8.43791544e-01 1.98751122e-01
4.22001779e-01 -1.15365744e+00 1.06455290e+00 -3.65671426e-01
7.69441903e-01 -6.61623836e-01 -4.71369654e-01 -1.34710932e+00
-4.30543959e-01 -6.06319189e-01 -5.02185524e-01 1.08504844e+00
-2.53843606e-01 -2.71610528e-01 5.55551410e-01 3.64663601e-01
-4.57468122e-01 -6.38887525e-01 -9.67602193e-01 -4.25671995e-01
-2.62842894e-01 -3.86113048e-01 4.47671384e-01 9.10032451e-01
-2.31959596e-01 -1.00721352e-01 -7.70785928e-01 1.88616097e-01
6.55461729e-01 -1.79011226e-01 4.64073420e-01 -9.23685551e-01
-1.37282670e-01 -2.76667148e-01 -6.43315434e-01 -1.08078158e+00
9.77950394e-02 -7.15530217e-01 2.55874731e-02 -1.69626606e+00
3.23059201e-01 -3.02212238e-02 -6.14161909e-01 3.90806675e-01
7.95713812e-03 8.97155583e-01 2.71260262e-01 4.57080573e-01
-8.68798554e-01 7.95196652e-01 1.34334970e+00 -1.93153352e-01
-1.52406514e-01 -3.36821079e-01 -8.85980070e-01 5.74444413e-01
6.83678150e-01 -2.66627252e-01 -1.73508525e-01 -6.95185363e-01
1.37728468e-01 2.61889338e-01 7.61480749e-01 -1.08166564e+00
2.19307676e-01 -1.32159337e-01 3.86368901e-01 -2.39766911e-01
4.08446044e-01 -5.87693930e-01 2.61038452e-01 2.76736230e-01
-6.82344511e-02 -1.95080623e-01 -3.56429294e-02 6.51411951e-01
-3.36323053e-01 8.73117521e-02 7.95748711e-01 8.52806121e-02
-9.26753938e-01 4.91971254e-01 -7.43399918e-01 -3.42171013e-01
7.38407552e-01 -3.47903371e-01 -3.26645195e-01 -8.02606165e-01
-8.87206912e-01 1.12913266e-01 8.98514748e-01 6.71082735e-01
9.07751322e-01 -1.45209312e+00 -9.44327652e-01 7.23206252e-02
-1.07875317e-01 -4.54326481e-01 9.02187407e-01 7.21459866e-01
-5.69767118e-01 -2.50194371e-01 -3.87004346e-01 -5.11182427e-01
-1.34703696e+00 5.81277966e-01 3.84977430e-01 1.96746305e-01
-9.17601943e-01 5.79714477e-01 1.54085577e-01 -6.38081804e-02
1.70314088e-01 7.93516487e-02 -2.77323246e-01 7.71654546e-02
8.82394791e-01 4.95704114e-01 1.67143717e-01 -9.92692590e-01
-1.92308366e-01 3.34944159e-01 2.82722004e-02 2.31424198e-01
1.54540098e+00 -7.21291184e-01 -2.15226948e-01 -1.29794613e-01
1.18675137e+00 -2.06496567e-01 -1.85787022e+00 -4.46663737e-01
-5.67593694e-01 -9.28707480e-01 5.10211647e-01 -5.81037581e-01
-1.80155301e+00 5.52490413e-01 9.98672664e-01 8.97803344e-03
1.70485508e+00 -1.99802697e-01 8.46903741e-01 -1.23063833e-01
1.99479163e-01 -1.07713258e+00 3.33900124e-01 1.95364594e-01
1.23745656e+00 -1.07752275e+00 1.37199447e-01 -3.32304806e-01
-4.79576439e-01 9.98429894e-01 4.40523684e-01 -2.19613418e-01
7.09177613e-01 3.87074612e-02 1.59718040e-02 4.30059806e-02
-7.04803884e-01 1.57072067e-01 5.39271086e-02 5.61714530e-01
2.25359470e-01 -8.38478953e-02 -1.25625432e-01 4.75177497e-01
1.81669578e-01 3.10016155e-01 9.35421765e-01 8.60024691e-01
-6.22526765e-01 -9.96799529e-01 -4.88848001e-01 3.46792430e-01
-3.20065051e-01 -1.36490077e-01 1.23314992e-01 1.96976289e-01
2.31203958e-01 1.26851618e+00 -2.11602580e-02 -5.52201271e-01
1.09234974e-01 -3.45558912e-01 3.38491082e-01 -3.09704185e-01
-3.60850781e-01 1.23245850e-01 -8.40593874e-02 -8.69153500e-01
-6.23690367e-01 -7.53068268e-01 -1.02747631e+00 -6.67186499e-01
3.87773290e-02 -3.08629960e-01 4.05260742e-01 8.86990845e-01
4.98691797e-01 8.42498064e-01 7.69032001e-01 -1.40011466e+00
8.62297341e-02 -8.63024116e-01 -4.68619108e-01 4.43393648e-01
6.63190365e-01 -7.27503300e-01 -4.34526712e-01 4.18620110e-01] | [11.219417572021484, -1.9180203676223755] |
5f1acfb0-07a1-4b28-bec9-e9c0c082ecdd | can-we-achieve-more-with-less-exploring-data | 2007.00875 | null | https://arxiv.org/abs/2007.00875v1 | https://arxiv.org/pdf/2007.00875v1.pdf | Can We Achieve More with Less? Exploring Data Augmentation for Toxic Comment Classification | This paper tackles one of the greatest limitations in Machine Learning: Data Scarcity. Specifically, we explore whether high accuracy classifiers can be built from small datasets, utilizing a combination of data augmentation techniques and machine learning algorithms. In this paper, we experiment with Easy Data Augmentation (EDA) and Backtranslation, as well as with three popular learning algorithms, Logistic Regression, Support Vector Machine (SVM), and Bidirectional Long Short-Term Memory Network (Bi-LSTM). For our experimentation, we utilize the Wikipedia Toxic Comments dataset so that in the process of exploring the benefits of data augmentation, we can develop a model to detect and classify toxic speech in comments to help fight back against cyberbullying and online harassment. Ultimately, we found that data augmentation techniques can be used to significantly boost the performance of classifiers and are an excellent strategy to combat lack of data in NLP problems. | ['Fang-I Hsiao', 'Chetanya Rastogi', 'Nikka Mofid'] | 2020-07-02 | null | null | null | null | ['toxic-comment-classification'] | ['natural-language-processing'] | [ 2.18623742e-01 2.44857490e-01 -2.89992899e-01 -3.27517539e-01
-6.73030853e-01 -3.68148476e-01 5.99384367e-01 5.05477190e-01
-6.70376182e-01 6.47317767e-01 5.18328130e-01 -8.24353755e-01
1.06007233e-01 -7.68898427e-01 -4.67470556e-01 -1.88180730e-01
1.28522411e-01 1.61048636e-01 -1.96216062e-01 -5.56123197e-01
5.29561877e-01 4.53409374e-01 -1.03885341e+00 4.32619631e-01
9.69062328e-01 6.26901150e-01 -3.84689957e-01 3.97564650e-01
-4.11104143e-01 8.68107319e-01 -6.41265869e-01 -7.05466986e-01
-1.21270463e-01 -2.21908450e-01 -1.10957670e+00 -3.48497868e-01
1.74649358e-01 -1.12375319e-01 -2.86268473e-01 7.19099581e-01
4.89816993e-01 1.84339628e-01 3.69941831e-01 -1.36708188e+00
-6.91606939e-01 7.09398150e-01 -4.98607457e-01 3.88344139e-01
6.15961254e-01 4.34420593e-02 6.51868045e-01 -8.50855827e-01
5.58384240e-01 1.30186629e+00 1.04827464e+00 7.15080321e-01
-1.02015960e+00 -8.21824729e-01 6.38907477e-02 2.40678564e-01
-9.74299252e-01 -3.82208019e-01 7.97173738e-01 -4.11036104e-01
9.53212500e-01 2.43286014e-01 4.62235600e-01 1.76630962e+00
-2.57343829e-01 7.44868994e-01 1.19186866e+00 -7.36684918e-01
-1.49397001e-01 5.03420293e-01 4.68679816e-01 6.74089909e-01
3.21306586e-02 -8.90829973e-03 -7.66521811e-01 -5.88721693e-01
5.95101640e-02 -2.43370995e-01 5.70956431e-03 4.58002567e-01
-9.02021945e-01 1.31444108e+00 3.59211415e-01 4.76361424e-01
-1.49476275e-01 -4.05720472e-01 6.62340522e-01 3.45795542e-01
8.81777585e-01 8.42242301e-01 -5.81836343e-01 -4.43538487e-01
-5.87756634e-01 1.95055995e-02 8.77023220e-01 2.12457255e-01
3.68698835e-01 1.16427243e-01 1.59936562e-01 1.18043125e+00
1.05817407e-01 1.99060500e-01 7.92184293e-01 -5.45443296e-01
7.47568130e-01 5.52944124e-01 -2.88065970e-01 -1.09792292e+00
-6.41580105e-01 -1.76606938e-01 -6.04819596e-01 -8.68747383e-02
4.73319829e-01 -5.06962478e-01 -6.52155876e-01 1.66693318e+00
2.80172497e-01 1.03983907e-02 7.02744573e-02 3.75587255e-01
5.95283151e-01 5.82330883e-01 4.31979954e-01 -2.24847585e-01
1.09483957e+00 -8.92940998e-01 -8.73434722e-01 -4.21923161e-01
1.40656435e+00 -6.68258607e-01 1.27298832e+00 4.70618486e-01
-6.96380854e-01 -3.10517043e-01 -8.32048535e-01 -9.13224518e-02
-9.21437263e-01 -3.03550392e-01 7.77205408e-01 1.00333583e+00
-5.93684375e-01 5.31109214e-01 -5.70376456e-01 -6.36769414e-01
3.86471421e-01 2.15916812e-01 -5.97235799e-01 -6.92639947e-02
-1.30858457e+00 1.23133516e+00 2.41375342e-01 -5.44948317e-02
-1.54426634e-01 -4.69611019e-01 -9.86982465e-01 -2.84347147e-01
2.97576547e-01 -2.81599015e-01 9.39082026e-01 -1.00408447e+00
-1.07852530e+00 7.94884145e-01 5.75120524e-02 -5.82157433e-01
1.30961239e-01 -4.11690176e-01 -4.53833193e-01 -2.70074636e-01
-3.93744186e-02 5.80165982e-01 5.67445934e-01 -9.32092667e-01
-3.83400142e-01 -7.55197525e-01 -1.22191206e-01 -1.51595414e-01
-1.30140114e+00 5.54547727e-01 4.26586866e-01 -7.57680058e-01
-1.69327557e-01 -1.05910218e+00 -1.62446484e-01 -4.76737529e-01
-5.35113513e-01 -3.21731657e-01 1.10732043e+00 -1.14352119e+00
1.29456460e+00 -2.17008328e+00 -1.74851596e-01 1.70770377e-01
-1.20941035e-01 7.03167975e-01 -2.43706539e-01 5.41782379e-01
-1.01374812e-01 8.34810197e-01 -1.19078740e-01 -3.30369771e-01
-3.31704706e-01 2.61248797e-01 -2.75779098e-01 2.13188171e-01
3.36176664e-01 8.41479301e-01 -6.84512615e-01 -3.33536953e-01
-5.61469980e-02 3.75236183e-01 -4.79367405e-01 1.20858811e-01
1.52161837e-01 3.69587868e-01 -3.56921732e-01 4.53302056e-01
3.51021975e-01 2.30342701e-01 2.25735623e-02 1.64528921e-01
-1.41376480e-01 5.85580945e-01 -6.53887689e-01 1.30822384e+00
-6.87265813e-01 7.60792017e-01 9.97319520e-02 -1.01828325e+00
9.61007774e-01 2.76258796e-01 3.14637691e-01 -8.83721054e-01
3.78037244e-01 6.08872510e-02 2.99762525e-02 -1.00804150e+00
3.85253280e-01 -7.52721950e-02 4.11868431e-02 6.39867127e-01
-1.04504243e-01 2.43528351e-01 2.61468682e-02 8.37969631e-02
9.60526347e-01 -3.45987052e-01 -2.49189101e-02 2.16686457e-01
3.92155826e-01 6.00167476e-02 4.12891209e-01 6.69943988e-01
-1.51579559e-01 2.25737765e-01 4.21830237e-01 -4.52517778e-01
-1.13400698e+00 -3.79894644e-01 -9.58467722e-02 1.53966212e+00
-6.49249554e-01 -3.93856078e-01 -7.18745410e-01 -9.04631555e-01
-1.65879384e-01 9.66444075e-01 -5.47243357e-01 -5.08453548e-01
-6.55645847e-01 -1.23930144e+00 9.98264670e-01 5.84346592e-01
3.41785729e-01 -1.04706442e+00 -1.07990064e-01 1.43226102e-01
-4.63712662e-01 -1.34815514e+00 -9.37349163e-03 2.87131935e-01
-7.83824563e-01 -1.07560039e+00 -1.86083034e-01 -7.35386431e-01
3.00901234e-01 1.43757135e-01 5.78530967e-01 3.38760704e-01
-2.23591477e-01 1.08343638e-01 -8.09316874e-01 -6.47721887e-01
-7.44737208e-01 4.65604603e-01 2.43521929e-01 -8.22872072e-02
4.65551436e-01 -5.43830514e-01 1.38863206e-01 -3.52016184e-03
-8.30031991e-01 -9.91410762e-02 4.53979731e-01 8.58152211e-01
-3.41322124e-01 -1.23838961e-01 8.26246083e-01 -1.12319458e+00
1.16014552e+00 -7.40307748e-01 1.16923869e-01 6.95841108e-03
-6.57554746e-01 -1.67491391e-01 5.51257133e-01 -6.19583786e-01
-8.70655358e-01 -1.71500191e-01 -6.99262500e-01 4.23364230e-02
-3.06802303e-01 8.67376626e-01 1.56009361e-01 -3.86094868e-01
9.17047203e-01 -6.93764985e-02 2.30014041e-01 -8.42040181e-01
3.13526273e-01 1.08565867e+00 6.81095719e-02 -3.70923132e-01
6.67077780e-01 1.10737182e-01 -2.69216508e-01 -1.12338924e+00
-1.16978121e+00 -4.05802995e-01 -8.47468138e-01 -4.32476227e-04
8.50337088e-01 -4.93371367e-01 -5.10371149e-01 5.69771767e-01
-1.10653448e+00 -2.29585871e-01 5.80217652e-02 4.33696032e-01
1.15560507e-02 3.38913053e-01 -8.79155755e-01 -1.01885748e+00
-4.48264837e-01 -5.69974720e-01 4.40711975e-01 3.95796634e-02
-4.31892365e-01 -1.00588632e+00 1.25958100e-01 1.08541620e+00
4.53977138e-01 4.13543522e-01 1.21249712e+00 -1.36173177e+00
2.72864133e-01 -3.85701120e-01 -1.15882857e-02 6.33287787e-01
-1.72305360e-01 8.77998546e-02 -1.05666983e+00 -1.88873708e-01
-3.33529860e-02 -9.07565296e-01 7.48647094e-01 -1.44266784e-01
1.21376133e+00 -8.46573234e-01 -1.16425842e-01 2.67437667e-01
8.71347725e-01 3.30154747e-02 5.17237902e-01 7.52582669e-01
6.54338062e-01 8.40181112e-01 5.26545107e-01 2.44925112e-01
3.55521917e-01 4.41082239e-01 2.24747583e-01 -2.14839876e-01
6.08579963e-02 -4.73125547e-01 2.96196669e-01 7.69261897e-01
1.35952860e-01 -1.47681683e-01 -1.25737834e+00 4.34396505e-01
-1.57604909e+00 -8.33396912e-01 -3.54642749e-01 1.84979498e+00
6.78315699e-01 1.72172919e-01 3.24316591e-01 6.22067988e-01
5.47373354e-01 -1.32833654e-02 -1.37124404e-01 -9.87409651e-01
-1.20622374e-01 1.00636527e-01 2.11682022e-01 3.05667341e-01
-1.15777886e+00 9.12682474e-01 6.85050297e+00 6.82015240e-01
-1.15914559e+00 3.14682782e-01 8.20278943e-01 1.16460904e-01
5.97781986e-02 -3.98631006e-01 -6.61855698e-01 3.70871186e-01
1.49554169e+00 2.26906657e-01 3.99833173e-01 7.54046917e-01
2.85251737e-01 8.86564180e-02 -6.15341008e-01 5.02294719e-01
4.13790137e-01 -1.04837620e+00 -8.66155103e-02 9.06537846e-02
4.37846959e-01 1.60320073e-01 2.04168275e-01 5.89798033e-01
1.69314682e-01 -1.06598341e+00 1.92142606e-01 4.45761047e-02
3.60360116e-01 -7.84125268e-01 9.47054327e-01 7.84578085e-01
-2.21544370e-01 -6.97560966e-01 -4.67972904e-02 -4.56653714e-01
-2.60754675e-02 3.13552231e-01 -1.08186972e+00 1.73162192e-01
7.28163004e-01 3.57934386e-01 -8.73452842e-01 8.43223631e-01
-1.85661018e-01 9.06453907e-01 -3.30735326e-01 -2.59479940e-01
3.52899104e-01 5.82648441e-02 3.51657897e-01 1.32507861e+00
-4.70920131e-02 1.61900759e-01 2.51986444e-01 4.43404555e-01
-1.67264372e-01 4.62171316e-01 -8.54256749e-01 -4.05996650e-01
3.11132282e-01 1.22587574e+00 -4.19963866e-01 1.66139361e-02
-4.52479422e-01 6.75260246e-01 6.66198134e-01 1.41859576e-01
-6.53877079e-01 -2.37141877e-01 3.75089049e-01 2.08665162e-01
-4.50398862e-01 -3.62584680e-01 -6.70502484e-01 -1.02973211e+00
-2.25399435e-01 -1.05962884e+00 6.42613232e-01 -7.07719803e-01
-1.42825139e+00 6.12177908e-01 -1.73628792e-01 -5.94618022e-01
-1.15930378e-01 -5.59884548e-01 -6.25354111e-01 4.13148195e-01
-1.39181793e+00 -1.52327144e+00 -2.25709602e-02 5.75597048e-01
4.78374779e-01 -2.38259897e-01 9.15542603e-01 5.61583579e-01
-9.13297892e-01 6.10915363e-01 5.60181178e-02 5.52785575e-01
6.47184372e-01 -7.79964924e-01 2.66273737e-01 6.74292922e-01
3.27866137e-01 4.69499707e-01 5.65741479e-01 -7.04144180e-01
-9.33981955e-01 -9.74034071e-01 1.16780758e+00 -3.65593463e-01
9.92180467e-01 -4.71290678e-01 -1.07980013e+00 6.54277861e-01
5.11203371e-02 -3.19023728e-01 1.09909976e+00 6.47871137e-01
-4.57183242e-01 2.13410348e-01 -1.20917416e+00 5.12593508e-01
7.18378186e-01 -6.18599713e-01 -6.07605636e-01 7.00844049e-01
8.77978802e-01 5.98059446e-02 -7.51937091e-01 4.10431862e-01
3.21847469e-01 -5.93981683e-01 8.68956447e-01 -1.41779649e+00
4.22081411e-01 5.99290311e-01 1.18323714e-01 -1.33792877e+00
-5.87346777e-02 -4.05798733e-01 1.51060954e-01 1.71199739e+00
7.45638490e-01 -7.16461837e-01 6.90059185e-01 9.19132292e-01
-3.19110714e-02 -7.47296154e-01 -9.16927636e-01 -5.06639183e-01
4.75835949e-01 -6.50504470e-01 2.52905726e-01 1.50125563e+00
3.46388668e-01 5.19004166e-01 -6.79806650e-01 -1.92043364e-01
1.53576553e-01 -5.20816684e-01 6.85836911e-01 -1.13344753e+00
2.84948707e-01 -2.64481753e-01 -1.41402066e-01 -2.71772027e-01
6.09007478e-01 -8.51774871e-01 -3.89807910e-01 -1.23139393e+00
2.57364988e-01 -5.31844676e-01 -1.43498123e-01 9.57961380e-01
-9.64802951e-02 5.15916944e-01 3.02673668e-01 -1.78081155e-01
-2.10488796e-01 5.38702726e-01 8.93667340e-01 -1.81522779e-02
-2.73926347e-01 2.74459329e-02 -1.01539075e+00 1.00984263e+00
1.09356737e+00 -6.53364658e-01 1.59370974e-02 -3.92510802e-01
2.49042675e-01 -4.02316861e-02 2.93378502e-01 -7.24143565e-01
8.72469544e-02 -9.14121792e-02 2.58896768e-01 -1.43765867e-01
4.70322073e-01 -5.19310653e-01 -6.19245410e-01 5.07575333e-01
-6.77671790e-01 1.74013287e-01 3.00081611e-01 3.45477730e-01
-1.32935256e-01 -3.28357726e-01 6.27620339e-01 -8.90761428e-03
-2.78532892e-01 5.87156750e-02 -6.95736647e-01 1.10292017e-01
7.96839774e-01 5.44719920e-02 -5.80687761e-01 -4.06326711e-01
-8.24564815e-01 1.64024130e-01 7.43285101e-03 8.99275541e-01
5.99820495e-01 -1.14375174e+00 -7.77788103e-01 2.77249545e-01
-5.46100475e-02 -7.61877060e-01 -2.78292280e-02 8.17558885e-01
-2.28800893e-01 4.49361116e-01 -2.35367581e-01 1.01289563e-02
-1.50582075e+00 6.68420076e-01 -1.12625919e-02 -1.50148720e-01
-3.36041182e-01 7.54327774e-01 -7.18363583e-01 -8.52668345e-01
4.00286168e-01 1.98616177e-01 -6.07585847e-01 2.59908408e-01
5.20410120e-01 4.58761543e-01 1.21990480e-01 -7.15027928e-01
-1.01860039e-01 -6.28276318e-02 -1.70238495e-01 -4.10659015e-02
1.65170455e+00 -4.22700159e-02 -1.71827614e-01 5.13813376e-01
1.26743162e+00 1.10539971e-02 -1.94215581e-01 -1.69336796e-01
4.53726977e-01 -5.40996492e-01 8.03114176e-02 -1.00274134e+00
-6.62960291e-01 9.94194210e-01 3.80933076e-01 5.49836814e-01
8.21866035e-01 -1.96403131e-01 1.11567771e+00 6.53365016e-01
-5.65484799e-02 -1.27919090e+00 3.14817995e-01 6.08291745e-01
7.10655689e-01 -1.49611151e+00 -2.13922977e-01 -3.86634380e-01
-6.71870291e-01 1.12568486e+00 8.59603882e-01 2.32363731e-01
6.82848275e-01 3.82713556e-01 1.75750956e-01 -9.98987034e-02
-7.78245270e-01 -1.09760344e-01 -6.47260854e-03 6.20090365e-01
6.24611676e-01 -3.08763474e-01 -5.42799234e-01 5.57788789e-01
-3.78271550e-01 -2.63257563e-01 6.73246443e-01 7.59454727e-01
-4.25188005e-01 -1.25919318e+00 -5.79464614e-01 5.88888228e-01
-8.76728594e-01 -2.00089291e-01 -9.09150541e-01 8.68463874e-01
8.58218446e-02 1.33650255e+00 -1.64193317e-01 -6.43643558e-01
2.10632667e-01 5.53754210e-01 -9.49892923e-02 -5.51134646e-01
-9.56617773e-01 -3.44426513e-01 6.27350450e-01 -3.34588408e-01
-2.85780430e-01 -6.65587008e-01 -7.78012633e-01 -5.05186677e-01
-6.70413315e-01 1.35556340e-01 1.09129524e+00 1.17523575e+00
2.62900710e-01 2.63130575e-01 5.95610201e-01 -3.90957922e-01
-5.28353333e-01 -1.36599565e+00 -2.15242237e-01 5.94534993e-01
1.79825753e-01 -4.00425285e-01 -3.59274864e-01 -2.95670629e-01] | [8.83866024017334, 10.477490425109863] |
5206d940-d690-452c-8a38-166f7cc38ff9 | korsal-key-point-detection-based-online-real | 2111.03319 | null | https://arxiv.org/abs/2111.03319v1 | https://arxiv.org/pdf/2111.03319v1.pdf | KORSAL: Key-point Detection based Online Real-Time Spatio-Temporal Action Localization | Real-time and online action localization in a video is a critical yet highly challenging problem. Accurate action localization requires the utilization of both temporal and spatial information. Recent attempts achieve this by using computationally intensive 3D CNN architectures or highly redundant two-stream architectures with optical flow, making them both unsuitable for real-time, online applications. To accomplish activity localization under highly challenging real-time constraints, we propose utilizing fast and efficient key-point based bounding box prediction to spatially localize actions. We then introduce a tube-linking algorithm that maintains the continuity of action tubes temporally in the presence of occlusions. Further, we eliminate the need for a two-stream architecture by combining temporal and spatial information into a cascaded input to a single network, allowing the network to learn from both types of information. Temporal information is efficiently extracted using a structural similarity index map as opposed to computationally intensive optical flow. Despite the simplicity of our approach, our lightweight end-to-end architecture achieves state-of-the-art frame-mAP of 74.7% on the challenging UCF101-24 dataset, demonstrating a performance gain of 6.4% over the previous best online methods. We also achieve state-of-the-art video-mAP results compared to both online and offline methods. Moreover, our model achieves a frame rate of 41.8 FPS, which is a 10.7% improvement over contemporary real-time methods. | ['Peshala Jayasekara', 'Ranga Rodrigo', 'Sachira Karunasena', 'Sakuna Jayasundara', 'Shechem Sumanthiran', 'Kalana Abeywardena'] | 2021-11-05 | null | null | null | null | ['spatio-temporal-action-localization'] | ['computer-vision'] | [ 1.72580242e-01 -4.11372334e-01 -3.39155465e-01 -3.69558744e-02
-6.01495266e-01 -4.79731858e-01 4.40446794e-01 1.58050060e-01
-7.86505997e-01 5.05543590e-01 3.23470026e-01 -4.21780460e-02
1.82180911e-01 -3.71404886e-01 -7.38898218e-01 -4.10381973e-01
-3.89745593e-01 1.46801472e-01 7.64061511e-01 6.18940331e-02
2.85368949e-01 5.57949185e-01 -1.66262639e+00 3.49526793e-01
4.39896643e-01 1.39265001e+00 1.20298289e-01 9.21118855e-01
7.68582821e-02 1.17288172e+00 -3.26381981e-01 -4.91982512e-02
5.76178730e-01 -2.68655956e-01 -6.60195291e-01 8.40214193e-02
9.43024635e-01 -9.36159968e-01 -8.54142010e-01 4.37935233e-01
5.54354787e-01 3.72743458e-01 6.94263875e-02 -1.10331297e+00
-1.72429442e-01 -2.66220644e-02 -5.97825289e-01 5.33236206e-01
5.30277193e-01 4.03989047e-01 8.48049462e-01 -7.37139523e-01
7.41292894e-01 9.97845292e-01 6.90890610e-01 4.41122949e-01
-1.14589417e+00 -4.17915970e-01 3.24621767e-01 5.04028261e-01
-1.16690409e+00 -7.31040478e-01 6.34169161e-01 -4.05449837e-01
1.39297748e+00 -1.82357967e-01 9.77270365e-01 8.78637254e-01
1.50576353e-01 7.93460727e-01 6.54868424e-01 1.37809515e-02
1.97121814e-01 -6.40575051e-01 -5.06146133e-01 1.00651205e+00
-1.70175374e-01 6.70388639e-02 -1.03328538e+00 1.63722768e-01
1.14881861e+00 2.20307752e-01 -2.50876069e-01 -4.99196500e-01
-1.49342775e+00 3.70542318e-01 5.63799202e-01 5.64219318e-02
-4.34040338e-01 7.33071566e-01 6.05827332e-01 8.81845690e-03
4.60394353e-01 1.83036774e-01 -4.82538164e-01 -8.64679873e-01
-1.06625056e+00 7.41488114e-02 5.43513715e-01 8.77376974e-01
5.72277904e-01 -8.76941979e-02 -2.44566754e-01 2.56652534e-01
7.75488913e-02 3.23728204e-01 2.42930129e-01 -1.47865391e+00
6.96823418e-01 5.63450813e-01 3.06333333e-01 -1.02944803e+00
-4.12709713e-01 -3.03574234e-01 -4.65693206e-01 2.87147790e-01
7.67666936e-01 1.30842835e-01 -7.28820086e-01 1.51172018e+00
5.29348433e-01 7.41591454e-01 -3.09514582e-01 1.03245723e+00
2.34860510e-01 4.44510162e-01 -4.76105921e-02 -1.97674602e-01
9.90030944e-01 -1.32648385e+00 -6.11326337e-01 -2.38187194e-01
8.22887957e-01 -5.58274806e-01 9.23978269e-01 2.21393570e-01
-1.26185191e+00 -5.36757290e-01 -9.05294061e-01 -4.87131417e-01
-1.48591802e-01 1.86987668e-01 6.33900404e-01 2.31508002e-01
-1.05807972e+00 8.37219656e-01 -1.40970600e+00 -2.09776685e-01
7.48558640e-01 4.95777398e-01 -5.35540462e-01 -1.33699119e-01
-6.05985761e-01 6.25446796e-01 2.03621805e-01 8.96441862e-02
-7.93259799e-01 -9.57049549e-01 -8.77448916e-01 1.33151310e-02
5.74273109e-01 -6.35382771e-01 1.28771162e+00 -7.87323952e-01
-1.65398848e+00 4.32936966e-01 -4.97952491e-01 -7.37409532e-01
8.94333661e-01 -6.32977009e-01 9.07268599e-02 8.65235627e-01
1.00841045e-01 7.49497831e-01 5.71799219e-01 -5.65033376e-01
-9.33498561e-01 -1.51515096e-01 3.52473080e-01 2.27835074e-01
-3.65038276e-01 1.22938780e-02 -7.07074642e-01 -4.44376588e-01
1.32009789e-01 -8.47109139e-01 -1.07157566e-01 9.59678292e-01
1.22244351e-01 -1.10628255e-01 1.10063791e+00 -5.91099203e-01
1.13521481e+00 -1.98956835e+00 1.07694436e-02 -2.62967110e-01
3.73030096e-01 4.10430849e-01 -7.92179555e-02 1.78594887e-01
1.96088135e-01 -1.91148266e-01 4.56237048e-02 -7.41459548e-01
-1.65757000e-01 -2.25106273e-02 -7.00180009e-02 8.18446696e-01
3.61381382e-01 1.00031066e+00 -1.22209179e+00 -5.81108689e-01
7.28314996e-01 7.52609968e-01 -7.34014332e-01 1.73291638e-01
-1.95845202e-01 5.88729620e-01 -2.81114817e-01 8.22668076e-01
1.54572949e-01 -3.67474675e-01 -2.21065320e-02 -3.06960255e-01
-2.80386001e-01 4.13370192e-01 -1.09943700e+00 2.33813763e+00
-4.53605413e-01 8.94181848e-01 -8.60428661e-02 -7.98619926e-01
5.17878950e-01 2.56721616e-01 1.16302705e+00 -8.63091707e-01
2.63217300e-01 1.33310810e-01 -2.27357417e-01 -3.68426859e-01
3.21913660e-01 3.67545426e-01 2.74640352e-01 2.95266390e-01
1.43416509e-01 4.36106652e-01 4.53129262e-01 1.64243922e-01
1.64153814e+00 7.68526018e-01 3.79630327e-02 1.37222841e-01
4.37559336e-01 -6.93011954e-02 6.36080384e-01 5.40338576e-01
-6.75941944e-01 6.45967185e-01 3.98708761e-01 -8.03705215e-01
-9.24035251e-01 -9.53532457e-01 2.95759499e-01 1.00668681e+00
2.29984909e-01 -6.44246459e-01 -4.66309011e-01 -7.21029162e-01
-2.90203430e-02 -7.28758201e-02 -5.09161532e-01 3.08205448e-02
-1.00630260e+00 -1.33467028e-02 3.55872124e-01 9.28223908e-01
8.14464390e-01 -6.79878354e-01 -1.18004835e+00 6.19763970e-01
-3.45441133e-01 -1.72286952e+00 -7.01376855e-01 -4.33889776e-02
-8.79382193e-01 -1.25419056e+00 -6.62100852e-01 -3.79210770e-01
4.85679924e-01 5.75844288e-01 9.82184768e-01 6.17212765e-02
-4.21735585e-01 4.07712370e-01 -4.43346113e-01 2.02407271e-01
1.73561603e-01 -4.07895306e-03 5.37868179e-02 1.36388034e-01
2.65269846e-01 -7.46974647e-01 -1.16205049e+00 4.35377121e-01
-5.69750130e-01 9.83663499e-02 3.76538962e-01 5.47605038e-01
7.46559501e-01 -3.16268563e-01 1.64345086e-01 -2.55676266e-02
-9.46594998e-02 -1.48371235e-01 -6.93480790e-01 3.12024262e-03
-1.97646171e-01 -1.14619479e-01 6.37297690e-01 -5.86903274e-01
-7.07083642e-01 6.52158499e-01 1.87404752e-01 -9.18530285e-01
-8.07231106e-03 5.00968769e-02 2.20727086e-01 -3.81305873e-01
5.09249508e-01 1.72204226e-01 2.33030796e-01 -2.84841478e-01
3.55018616e-01 2.91430444e-01 8.19538176e-01 -3.74704063e-01
5.59193611e-01 1.00903428e+00 2.33104378e-01 -5.86376965e-01
-9.43746030e-01 -7.89792717e-01 -1.14236844e+00 -5.06149471e-01
1.01462364e+00 -1.08079207e+00 -1.16037261e+00 4.87322450e-01
-1.14768314e+00 -5.96445024e-01 -2.60648519e-01 6.77376628e-01
-9.48537588e-01 4.79030281e-01 -6.08095169e-01 -5.31874478e-01
-1.70645073e-01 -1.04659092e+00 1.33766186e+00 -7.34426687e-03
-1.08387582e-01 -7.98001587e-01 -8.22560117e-02 4.88815427e-01
3.60633492e-01 5.16242504e-01 2.40533464e-02 -1.03667505e-01
-9.81608093e-01 -2.74359316e-01 -3.39180976e-01 9.12884921e-02
1.03286043e-01 -1.12931199e-01 -7.33037651e-01 -3.32323462e-01
-3.88303816e-01 -3.49191546e-01 9.11958575e-01 3.39152545e-01
1.17871058e+00 -3.05823777e-02 -2.43838266e-01 9.17212307e-01
1.29598737e+00 6.92476332e-02 5.73841751e-01 3.68884295e-01
8.73426616e-01 3.16240013e-01 6.93119824e-01 6.62835121e-01
4.34925079e-01 8.78919303e-01 5.37891507e-01 5.31086931e-04
-4.20796156e-01 -3.94857466e-01 4.59400445e-01 3.86200488e-01
-3.10346395e-01 -1.19556598e-01 -7.99230933e-01 5.67298114e-01
-2.25270891e+00 -1.14501047e+00 9.03976485e-02 2.24014997e+00
6.79810405e-01 2.30550155e-01 3.30147117e-01 5.02514392e-02
3.95613104e-01 3.41396540e-01 -6.79869115e-01 1.51040286e-01
9.48481262e-02 1.29068986e-01 7.79138029e-01 3.90130609e-01
-1.46872938e+00 1.03138268e+00 5.79082680e+00 5.09058535e-01
-1.19335878e+00 9.46073905e-02 3.57374400e-01 -8.61179352e-01
4.90820497e-01 -3.04916166e-02 -7.14931309e-01 4.86515760e-01
9.74784374e-01 1.42760903e-01 3.73848796e-01 6.17025316e-01
7.46537924e-01 -4.05013889e-01 -1.18265593e+00 1.15621960e+00
8.53800327e-02 -1.63954115e+00 -4.17541951e-01 1.81904599e-01
5.57191133e-01 2.23189190e-01 -3.42300892e-01 -1.37305543e-01
-4.00791131e-02 -7.72824585e-01 8.26330543e-01 4.38249379e-01
8.66979182e-01 -6.30191684e-01 3.85875791e-01 1.81182832e-01
-1.61603689e+00 -2.63212413e-01 -1.62292540e-01 -3.48097414e-01
5.09488404e-01 4.23300564e-01 -5.09701610e-01 1.99586287e-01
7.82276332e-01 1.22921908e+00 -4.11175400e-01 1.30389345e+00
-1.08346492e-01 3.59281361e-01 -6.21688545e-01 6.73115551e-02
4.20139849e-01 1.60101429e-01 3.05930257e-01 1.02509940e+00
2.08872139e-01 6.13096878e-02 4.49199706e-01 4.69539106e-01
1.96395833e-02 -1.62351012e-01 -4.93974417e-01 -4.52719033e-02
2.65996248e-01 9.43740487e-01 -7.18023956e-01 -3.16662431e-01
-6.03355169e-01 1.20087159e+00 5.04883885e-01 1.14144355e-01
-1.11279738e+00 -3.03222716e-01 8.89261305e-01 3.05028081e-01
5.47474802e-01 -8.73453438e-01 1.93830412e-02 -1.18196511e+00
5.44560969e-01 -3.64210725e-01 1.45717710e-01 -6.32356346e-01
-6.80488288e-01 2.23740637e-01 -2.02495903e-01 -1.52128899e+00
-2.28838608e-01 -6.09949291e-01 -2.58121520e-01 4.44761366e-01
-1.72312570e+00 -1.15391254e+00 -5.59910417e-01 7.71473229e-01
6.54028833e-01 1.02666877e-01 5.73385417e-01 5.23954391e-01
-5.18666446e-01 4.24006164e-01 -5.45236059e-02 2.91774362e-01
7.42618144e-01 -1.01795435e+00 4.42773014e-01 9.84485805e-01
1.99636430e-01 2.05413327e-01 2.04368487e-01 -5.76209724e-01
-1.62964869e+00 -1.15888703e+00 8.49649906e-01 -4.51271236e-01
8.01086664e-01 -5.12264371e-01 -6.19056046e-01 4.90258723e-01
-1.54752672e-01 7.40799546e-01 3.63381326e-01 -2.92893052e-01
-4.58011150e-01 -1.93411663e-01 -8.62317681e-01 5.76159775e-01
1.59320211e+00 -4.84783411e-01 -1.56781286e-01 4.94664788e-01
7.31100738e-01 -6.33673787e-01 -8.59882414e-01 1.37544662e-01
7.82079160e-01 -1.17621005e+00 1.00876296e+00 -4.96867418e-01
4.93443549e-01 -5.50349116e-01 -1.08485287e-02 -6.62828326e-01
-2.88853943e-01 -1.08944952e+00 -6.29544079e-01 6.22186840e-01
-6.35837987e-02 -3.67593795e-01 1.14104736e+00 6.78402066e-01
-1.71409696e-01 -1.06821203e+00 -1.19828892e+00 -7.87020326e-01
-4.76222247e-01 -5.78816831e-01 5.13104834e-02 4.99730974e-01
-1.56134262e-03 -7.23473281e-02 -4.04988557e-01 -1.64684542e-02
4.56978410e-01 -3.48084420e-02 8.03812087e-01 -7.98952758e-01
-1.99577630e-01 -3.76942694e-01 -9.20038819e-01 -1.62435675e+00
2.39359871e-01 -4.50962126e-01 3.30651961e-02 -1.52315927e+00
-1.54221401e-01 -7.87406936e-02 -2.29019970e-01 6.66630983e-01
1.70801163e-01 5.61512351e-01 3.35960448e-01 1.64777324e-01
-9.78917480e-01 7.08973765e-01 1.22875142e+00 8.69705752e-02
-2.90244848e-01 -2.70050287e-01 -9.20411199e-02 6.78477526e-01
6.70295298e-01 -2.56864160e-01 -5.13677537e-01 -6.83055520e-01
-1.69225469e-01 5.26463129e-02 7.53853202e-01 -1.41134083e+00
5.10821223e-01 -2.16544956e-01 4.42145497e-01 -6.76299512e-01
8.25165749e-01 -8.64764452e-01 -3.91649514e-01 4.32209402e-01
-1.77160367e-01 1.63287982e-01 1.60968110e-01 8.47874939e-01
-5.93204349e-02 3.64528596e-01 6.63146615e-01 -1.23048760e-01
-1.03395677e+00 6.53660297e-01 -3.40717733e-02 8.85330439e-02
1.31109869e+00 -6.01549804e-01 -3.01195621e-01 -3.11384469e-01
-5.65454364e-01 2.15850279e-01 4.47866261e-01 4.27347600e-01
6.79253161e-01 -1.20596468e+00 -4.60080028e-01 7.17023090e-02
-1.08631492e-01 8.47039670e-02 2.85939604e-01 1.22188306e+00
-8.44218850e-01 3.56280595e-01 -2.22283840e-01 -8.69037509e-01
-1.21933377e+00 2.08003029e-01 4.49275732e-01 -1.05774313e-01
-9.05331492e-01 7.77142465e-01 -1.85266286e-01 2.95292884e-01
4.47044551e-01 -3.02403539e-01 2.11742342e-01 -1.23222880e-01
7.05077291e-01 5.76015294e-01 -2.43838821e-02 -6.67075634e-01
-4.39180076e-01 7.25944817e-01 1.53027818e-01 -1.38018891e-01
1.26482928e+00 -1.07980520e-01 1.93663955e-01 1.56932175e-01
1.49778628e+00 -4.50574130e-01 -2.06103587e+00 -2.86223054e-01
-2.88359374e-01 -1.03609002e+00 2.00387940e-01 -5.78196824e-01
-1.12636602e+00 9.30487752e-01 5.74700236e-01 -2.34812722e-01
1.14440989e+00 -1.93611592e-01 1.12366796e+00 4.42171037e-01
3.64819884e-01 -1.20615530e+00 4.38714027e-01 6.01663291e-01
6.58852160e-01 -1.28489244e+00 1.19861111e-01 -3.99393857e-01
-1.87771350e-01 1.22217560e+00 6.76123321e-01 -2.75697887e-01
5.03029644e-01 2.76554346e-01 -1.61558643e-01 6.38572499e-02
-9.09161806e-01 -3.37238967e-01 1.53384537e-01 5.61896026e-01
2.90595710e-01 -4.05989170e-01 -8.09149072e-02 -1.77780479e-01
1.76424414e-01 2.63678968e-01 3.04120958e-01 1.33963537e+00
-4.05571103e-01 -7.40403771e-01 1.68223456e-01 1.76833451e-01
-4.53378916e-01 1.31844327e-01 -1.20682664e-01 7.05017626e-01
1.25533883e-02 9.55890298e-01 3.65921050e-01 -3.56011063e-01
3.27626020e-01 -1.04827695e-01 5.82873285e-01 -3.32022965e-01
-4.37493801e-01 1.04326837e-01 1.19128101e-01 -1.53913653e+00
-7.36095011e-01 -6.41657829e-01 -1.37864912e+00 -3.98194164e-01
-5.59994504e-02 -6.17298305e-01 6.58620298e-01 9.68425989e-01
8.52184772e-01 4.11456704e-01 5.13542056e-01 -1.30559337e+00
-2.79955089e-01 -6.03363216e-01 -4.01813835e-02 3.35978389e-01
5.00936389e-01 -8.90407622e-01 -2.48317018e-01 2.84504324e-01] | [8.398941993713379, 0.3102327883243561] |
973e7987-de43-4194-8bf7-f96dfed3a26b | flowmot-3d-multi-object-tracking-by-scene | 2012.07541 | null | https://arxiv.org/abs/2012.07541v3 | https://arxiv.org/pdf/2012.07541v3.pdf | FlowMOT: 3D Multi-Object Tracking by Scene Flow Association | Most end-to-end Multi-Object Tracking (MOT) methods face the problems of low accuracy and poor generalization ability. Although traditional filter-based methods can achieve better results, they are difficult to be endowed with optimal hyperparameters and often fail in varying scenarios. To alleviate these drawbacks, we propose a LiDAR-based 3D MOT framework named FlowMOT, which integrates point-wise motion information with the traditional matching algorithm, enhancing the robustness of the motion prediction. We firstly utilize a scene flow estimation network to obtain implicit motion information between two adjacent frames and calculate the predicted detection for each old tracklet in the previous frame. Then we use Hungarian algorithm to generate optimal matching relations with the ID propagation strategy to finish the tracking task. Experiments on KITTI MOT dataset show that our approach outperforms recent end-to-end methods and achieves competitive performance with the state-of-the-art filter-based method. In addition, ours can work steadily in the various-speed scenarios where the filter-based methods may fail. | ['Yong liu', 'Jinhao Cui', 'Zhen Yang', 'Xin Kong', 'Guangyao Zhai'] | 2020-12-14 | null | null | null | null | ['scene-flow-estimation', '3d-multi-object-tracking'] | ['computer-vision', 'computer-vision'] | [-2.14048773e-01 -7.98916280e-01 -2.44752243e-01 -5.07264398e-02
-5.55225372e-01 -2.97434419e-01 3.23953778e-01 -3.69490772e-01
-5.06078124e-01 6.06230080e-01 -1.49294466e-01 -8.80400389e-02
-1.56780273e-01 -6.99857652e-01 -5.48405409e-01 -5.91992497e-01
7.29770586e-02 5.43816507e-01 1.14711821e+00 -5.96849732e-02
1.40020505e-01 4.86131608e-01 -1.64185154e+00 -2.04026569e-02
8.63883317e-01 9.81805086e-01 4.58098263e-01 7.55288780e-01
-2.01741040e-01 8.42476189e-01 -2.07525253e-01 -2.90991038e-01
4.43114311e-01 -2.04881817e-01 -3.29512447e-01 -8.27208087e-02
1.07347810e+00 -4.21501040e-01 -8.05967033e-01 9.26103890e-01
6.96014285e-01 3.96536648e-01 3.28036249e-01 -1.42117321e+00
-1.95473522e-01 1.18427701e-01 -7.74170637e-01 4.82827932e-01
3.09684545e-01 3.87836397e-01 5.66962600e-01 -1.00881028e+00
7.71646559e-01 1.48502409e+00 9.44158435e-01 6.22024477e-01
-8.97750854e-01 -9.01548505e-01 2.47056827e-01 5.02845466e-01
-1.33631492e+00 -4.79711533e-01 5.85857153e-01 -4.29210663e-01
5.88120341e-01 -5.15225381e-02 7.68288910e-01 6.69118464e-01
2.81524599e-01 9.73701477e-01 5.60453534e-01 1.68408230e-02
-2.35659420e-01 -2.71567613e-01 -1.81067467e-01 1.11262751e+00
1.45661801e-01 5.17857254e-01 -6.01445735e-01 4.74854261e-02
8.03692758e-01 7.21345609e-03 -2.66462028e-01 -5.11584997e-01
-1.56640053e+00 6.23603761e-01 6.09544337e-01 5.06984554e-02
-3.74301285e-01 3.40403527e-01 2.18595982e-01 2.51941774e-02
1.93723351e-01 -3.07693481e-01 -2.93580383e-01 1.25761211e-01
-1.20182121e+00 6.09451652e-01 3.83360118e-01 1.10698318e+00
7.13803232e-01 1.25901639e-01 -4.45559859e-01 4.91585940e-01
6.02695405e-01 6.27114952e-01 1.36446595e-01 -1.19580233e+00
5.66599965e-01 2.85280377e-01 2.59831905e-01 -1.03578568e+00
-4.96400177e-01 -5.00623941e-01 -6.80086493e-01 4.19838697e-01
7.58581698e-01 -1.64290532e-01 -7.49969244e-01 1.40353060e+00
8.79392266e-01 9.21907425e-01 -1.72993824e-01 1.15525568e+00
8.34376633e-01 5.89279711e-01 1.58474073e-01 -2.76386470e-01
1.03307080e+00 -1.36830187e+00 -6.77078247e-01 -2.72085875e-01
5.75802386e-01 -9.78194475e-01 4.39093262e-01 2.48499706e-01
-1.06439638e+00 -9.75910485e-01 -8.05589378e-01 2.33513534e-01
5.21117076e-02 1.12047561e-01 5.15215576e-01 4.90312099e-01
-8.40801239e-01 6.38611257e-01 -1.01217604e+00 -3.08536589e-01
5.80940187e-01 2.96459913e-01 -1.45918489e-01 -2.53712475e-01
-8.30536962e-01 9.28345442e-01 5.46834111e-01 1.31158367e-01
-8.12144637e-01 -1.00021386e+00 -7.19276667e-01 -2.71309137e-01
4.85561639e-01 -1.32870460e+00 1.05483294e+00 -4.47156698e-01
-1.34107018e+00 3.30945522e-01 -3.14692765e-01 -4.71143782e-01
8.64498973e-01 -4.56314385e-01 -2.28916392e-01 1.10250600e-01
1.38437808e-01 9.66008604e-01 6.98208749e-01 -8.73367548e-01
-1.24319339e+00 -9.20630172e-02 -2.29423597e-01 1.87568367e-01
-6.25526160e-02 6.45503625e-02 -7.68904448e-01 -5.61681330e-01
1.82014123e-01 -1.01790690e+00 -4.43495303e-01 6.73728526e-01
-1.49041712e-01 -2.48930961e-01 1.42002463e+00 -2.36812890e-01
1.25804377e+00 -1.99005520e+00 4.42001894e-02 -2.98214167e-01
1.52527437e-01 4.50779706e-01 -4.65205312e-02 -1.12359323e-01
4.56632882e-01 -4.64814842e-01 1.60748556e-01 -4.19172883e-01
-2.40745962e-01 -2.00224444e-02 -1.67335682e-02 7.35903502e-01
5.94679788e-02 8.19327772e-01 -1.11324441e+00 -1.09788811e+00
8.57808053e-01 5.69465876e-01 -7.20901489e-01 1.02273181e-01
-3.32629204e-01 6.66187167e-01 -7.53079534e-01 6.01715326e-01
8.60074580e-01 -2.34092131e-01 -2.72710383e-01 -4.07383323e-01
-4.83631492e-01 -2.16923863e-01 -1.56841838e+00 1.99616098e+00
-2.01814711e-01 6.92547143e-01 -2.36323979e-02 -5.97863197e-01
6.93561256e-01 3.51130292e-02 8.41852903e-01 -3.31777364e-01
1.48397744e-01 4.35875580e-02 -4.59583253e-02 -4.33860034e-01
6.62995279e-01 5.71021531e-03 3.39351803e-01 -6.85037002e-02
-6.76986054e-02 1.93900749e-01 2.68204659e-01 1.12116337e-01
1.05800021e+00 7.41190970e-01 -2.15364974e-02 -6.37157708e-02
8.93229544e-01 2.55951554e-01 1.08971477e+00 7.64067888e-01
-6.70299530e-01 6.34230018e-01 -4.71507072e-01 -6.00402534e-01
-8.84593844e-01 -1.10512102e+00 2.14945013e-03 7.50999868e-01
7.78704882e-01 -3.35839212e-01 -1.86419889e-01 -6.42006874e-01
1.16381809e-01 2.93888837e-01 -6.20105192e-02 9.41856578e-02
-1.03306437e+00 -5.83057284e-01 3.39886963e-01 6.46776795e-01
8.73508215e-01 -6.29767478e-01 -8.69816542e-01 6.45017326e-01
-5.05174994e-01 -1.36475730e+00 -6.43940091e-01 -5.47902822e-01
-9.78660464e-01 -1.04473543e+00 -7.36023843e-01 -6.53203070e-01
2.02757806e-01 6.82392359e-01 8.58152986e-01 2.36371771e-01
-2.51158088e-01 -7.07621407e-03 -1.56826183e-01 -2.05562010e-01
-2.42157012e-01 -8.45760778e-02 1.37815684e-01 -8.59881192e-02
2.35568836e-01 -3.14648837e-01 -1.02213311e+00 7.17807472e-01
-4.98562932e-01 3.57621349e-02 3.75996560e-01 6.21926963e-01
6.67761266e-01 1.80762149e-02 2.97169834e-01 -2.16094643e-01
-2.53119051e-01 -1.88694984e-01 -8.56325448e-01 1.73525915e-01
-4.28159922e-01 -7.28986710e-02 4.29850399e-01 -6.79767251e-01
-1.09759545e+00 6.26392961e-01 -3.80178466e-02 -9.37270582e-01
4.08482589e-02 -1.16050854e-01 7.32830539e-02 -5.29730380e-01
4.84734863e-01 4.01166119e-02 -8.07114393e-02 -3.24077070e-01
3.94769818e-01 2.41757408e-01 8.37409139e-01 -2.40485042e-01
1.20823288e+00 8.81380737e-01 2.27420464e-01 -5.85647941e-01
-9.01685834e-01 -8.14620972e-01 -7.15819001e-01 -8.57393324e-01
1.00679326e+00 -1.15044153e+00 -8.18902731e-01 4.60898131e-01
-1.26777732e+00 -1.23590995e-02 1.50789604e-01 8.98370326e-01
-6.50385737e-01 6.22331619e-01 -5.73061883e-01 -8.43815744e-01
-3.48043144e-01 -1.09128106e+00 1.12906504e+00 5.18168390e-01
1.84873745e-01 -9.10343766e-01 6.51440173e-02 2.38302693e-01
3.70327473e-01 2.78771460e-01 1.63587898e-01 4.95950058e-02
-1.21823835e+00 -1.27546899e-02 -2.37937897e-01 -3.35145056e-01
-2.09376574e-01 2.23147631e-01 -7.29713500e-01 -3.46308649e-01
-3.14403325e-01 1.10746421e-01 1.01139688e+00 8.04346800e-01
8.48832071e-01 1.59588635e-01 -7.79916465e-01 8.27878356e-01
1.55647397e+00 7.02453628e-02 5.56940675e-01 5.59228480e-01
8.14439654e-01 2.69701123e-01 1.18323982e+00 4.17151302e-01
4.97506738e-01 9.94128942e-01 4.86244321e-01 2.06865996e-01
-4.39674228e-01 -2.59381175e-01 4.18606877e-01 5.18798888e-01
-1.24659009e-01 -3.28828603e-01 -6.09777510e-01 4.75257725e-01
-2.29873943e+00 -1.32059479e+00 -5.26535690e-01 2.13957834e+00
2.14868844e-01 4.57887292e-01 4.27028596e-01 -1.55813321e-01
9.73111153e-01 1.89443126e-01 -4.38609093e-01 5.63300014e-01
-9.49691981e-02 -2.35220358e-01 6.08195961e-01 4.20780838e-01
-1.33085489e+00 1.23307240e+00 5.93757725e+00 8.84662330e-01
-8.23860705e-01 3.00133497e-01 -1.32262737e-01 -1.66504055e-01
1.69711441e-01 1.50821626e-01 -1.30458200e+00 3.54372978e-01
7.49064803e-01 -2.39927899e-02 -1.62561778e-02 7.35001981e-01
4.30072188e-01 -2.10240692e-01 -8.18532646e-01 1.11178398e+00
-1.85350269e-01 -1.51749599e+00 -1.56018913e-01 -1.73640713e-01
5.55937886e-01 2.76102811e-01 -2.24028513e-01 2.48120934e-01
2.57619262e-01 -4.20002043e-01 8.85351598e-01 6.89423800e-01
4.07159537e-01 -5.57871461e-01 4.26920384e-01 5.98264456e-01
-1.86227202e+00 -6.14746846e-02 -5.79070628e-01 1.01213127e-01
7.30326593e-01 4.46418017e-01 -4.01390225e-01 8.02456796e-01
7.80644774e-01 9.58280623e-01 -4.29811120e-01 1.92305434e+00
1.42758787e-01 2.02445954e-01 -6.08041346e-01 3.39681320e-02
2.87509233e-01 -5.71054667e-02 9.59779084e-01 1.13223946e+00
5.79120338e-01 -1.19514592e-01 6.55576706e-01 5.99771261e-01
2.99278289e-01 -7.75663555e-02 -5.19555032e-01 5.62615335e-01
5.23639441e-01 1.33886147e+00 -8.48227859e-01 -4.93155628e-01
-6.09455168e-01 7.39937544e-01 4.87938859e-02 1.17889129e-01
-1.11998069e+00 -1.42068163e-01 7.49210954e-01 1.43125921e-01
6.71504319e-01 -3.42357993e-01 1.04354113e-01 -1.18987954e+00
-4.92581502e-02 -3.33713979e-01 6.18907690e-01 -8.71062338e-01
-1.04080343e+00 5.24860442e-01 -4.85848710e-02 -1.89084423e+00
-1.20437518e-01 -4.01628822e-01 -6.09020591e-01 5.22402883e-01
-1.72689176e+00 -1.16458917e+00 -5.32262683e-01 6.19633675e-01
8.59736443e-01 3.54733504e-02 1.93135813e-01 8.20133924e-01
-5.22942722e-01 3.60074818e-01 -5.45452163e-02 1.27719507e-01
7.87532210e-01 -7.17771947e-01 3.01106900e-01 1.13768625e+00
-3.07763517e-02 7.33122677e-02 7.43086278e-01 -8.83025289e-01
-1.41066360e+00 -1.47543037e+00 6.68959916e-01 -4.77717608e-01
5.86767316e-01 1.16247535e-01 -7.71972060e-01 5.13635099e-01
-5.41262627e-02 4.97632653e-01 6.11288212e-02 -3.64806384e-01
-5.54271741e-03 -1.02328770e-01 -9.08385634e-01 6.07732356e-01
1.44646430e+00 6.01159707e-02 -4.38116431e-01 2.66564965e-01
7.22315133e-01 -7.29891121e-01 -7.67699957e-01 5.71986258e-01
7.35924780e-01 -9.57108557e-01 1.31523407e+00 -3.30210000e-01
-5.23457713e-02 -9.23918068e-01 7.92489126e-02 -9.27573264e-01
-5.81323862e-01 -7.75688410e-01 -4.45357829e-01 1.10064340e+00
4.59552184e-02 -2.72855490e-01 1.13504195e+00 1.42780572e-01
-1.87889352e-01 -4.96674329e-01 -1.07691634e+00 -1.08900011e+00
-1.69530615e-01 -5.25025725e-01 3.75710368e-01 6.11011505e-01
-4.91170526e-01 1.88932493e-01 -6.01766944e-01 3.63884926e-01
1.24607706e+00 2.78569996e-01 9.62010980e-01 -1.42442000e+00
-1.63953081e-01 -4.28124160e-01 -6.71102822e-01 -1.48203957e+00
4.21708748e-02 -6.47124529e-01 2.40184128e-01 -1.32963765e+00
-5.76260686e-03 -5.53232968e-01 -1.26070708e-01 2.99123991e-02
-4.02878463e-01 2.33692318e-01 5.88548124e-01 3.12588096e-01
-9.75288510e-01 6.03664875e-01 1.51093912e+00 -9.56480429e-02
-3.05829972e-01 2.23573685e-01 2.22330764e-02 7.08847284e-01
4.57827538e-01 -7.40490258e-01 -1.32556021e-01 -4.31336701e-01
-1.83569834e-01 4.49813098e-01 6.07304573e-01 -1.49733996e+00
7.65562057e-01 -2.38765374e-01 4.93473858e-01 -1.31939721e+00
4.31575626e-01 -8.66604924e-01 4.69537139e-01 7.38066137e-01
9.69085544e-02 2.30467096e-01 1.90961987e-01 8.44805837e-01
-9.89625696e-03 -6.30629854e-03 9.30827796e-01 -1.34864692e-02
-1.04857385e+00 9.26205993e-01 -1.70502514e-01 1.94832291e-02
1.15660369e+00 -4.82566029e-01 -2.90409923e-01 -7.12565705e-02
-4.53326017e-01 6.40879035e-01 2.74295598e-01 6.10301077e-01
7.56424904e-01 -1.62263024e+00 -8.34037066e-01 -9.18809623e-02
-8.41811225e-02 6.16791621e-02 3.79999846e-01 1.20240390e+00
-5.81803977e-01 2.92778611e-01 -1.97117656e-01 -1.10476768e+00
-1.41828239e+00 6.77844703e-01 4.08027977e-01 -9.70295817e-02
-1.09161484e+00 6.12176597e-01 1.78343520e-01 -7.93924704e-02
2.53612399e-01 -6.75701946e-02 -1.38337359e-01 -1.91191033e-01
7.32825041e-01 6.56808853e-01 -2.76562631e-01 -8.36114943e-01
-5.53021133e-01 1.02953196e+00 5.72619168e-03 6.24005981e-02
9.91354704e-01 -3.51172864e-01 5.23374319e-01 3.39718647e-02
8.71736586e-01 -3.30143601e-01 -1.62919116e+00 -3.95728558e-01
-1.53375298e-01 -9.98486578e-01 2.69569248e-01 -1.57941759e-01
-1.35457182e+00 6.89430118e-01 9.15130675e-01 -2.50701368e-01
9.75421906e-01 -1.64002344e-01 1.20977390e+00 2.07056269e-01
5.54857969e-01 -8.70930135e-01 -1.74734481e-02 4.84424084e-01
1.53307810e-01 -1.24856699e+00 8.54924545e-02 -7.10505247e-01
-2.37084255e-01 1.18588948e+00 9.10676956e-01 -1.13310903e-01
5.83820224e-01 2.70720154e-01 1.49919972e-01 1.35911405e-02
-7.94192970e-01 -6.31882131e-01 2.78869420e-01 7.05175996e-01
1.11217894e-01 -5.93266726e-01 -1.52179614e-01 -3.47174197e-01
1.53114647e-01 2.58974969e-01 1.33897856e-01 9.32410777e-01
-7.76825309e-01 -8.61414075e-01 -5.94784498e-01 2.17843279e-01
-3.96298409e-01 2.27321625e-01 2.71016568e-01 9.30210710e-01
1.34033740e-01 9.61362123e-01 8.17283406e-04 -4.97485578e-01
4.67701763e-01 -2.20997840e-01 6.87046647e-01 -8.52572769e-02
-5.21833837e-01 3.30065578e-01 -1.87469292e-02 -9.02179837e-01
-8.87453198e-01 -9.74985003e-01 -1.28214908e+00 -5.88443160e-01
-6.65066779e-01 -1.88149527e-01 2.89208829e-01 9.14923012e-01
2.82316685e-01 5.08845687e-01 3.89873564e-01 -1.16388428e+00
-2.65667886e-01 -5.88414788e-01 -5.30177802e-02 2.54435420e-01
5.37753701e-01 -1.05920303e+00 -1.41675740e-01 1.91050768e-03] | [6.480399131774902, -2.105417490005493] |
c38c226d-d024-4a51-8892-a95e72f1f79b | defend-data-poisoning-attacks-on-voice | 2209.04547 | null | https://arxiv.org/abs/2209.04547v2 | https://arxiv.org/pdf/2209.04547v2.pdf | Defend Data Poisoning Attacks on Voice Authentication | With the advances in deep learning, speaker verification has achieved very high accuracy and is gaining popularity as a type of biometric authentication option in many scenes of our daily life, especially the growing market of web services. Compared to traditional passwords, "vocal passwords" are much more convenient as they relieve people from memorizing different passwords. However, new machine learning attacks are putting these voice authentication systems at risk. Without a strong security guarantee, attackers could access legitimate users' web accounts by fooling the deep neural network (DNN) based voice recognition models. In this paper, we demonstrate an easy-to-implement data poisoning attack to the voice authentication system, which can hardly be captured by existing defense mechanisms. Thus, we propose a more robust defense method, called Guardian, which is a convolutional neural network-based discriminator. The Guardian discriminator integrates a series of novel techniques including bias reduction, input augmentation, and ensemble learning. Our approach is able to distinguish about 95% of attacked accounts from normal accounts, which is much more effective than existing approaches with only 60% accuracy. | ['Dan Lin', 'Cameron Baird', 'Ke Li'] | 2022-09-09 | null | null | null | null | ['data-poisoning'] | ['adversarial'] | [-7.27163404e-02 -4.27430481e-01 -9.01457369e-02 -1.40347973e-01
-3.51047218e-01 -8.56245875e-01 5.38212538e-01 -9.69982520e-02
-5.62058389e-01 4.96805519e-01 3.55237834e-02 -8.39858413e-01
2.79968172e-01 -8.01928282e-01 5.10808975e-02 -6.62236333e-01
4.27070677e-01 4.51822951e-02 3.03063020e-02 -3.18437487e-01
2.52788574e-01 8.54231834e-01 -1.15686321e+00 1.39495462e-01
6.51884139e-01 1.02456224e+00 -5.19220591e-01 5.30525267e-01
-2.62997806e-01 2.73884118e-01 -8.33856881e-01 -7.54518211e-01
3.55281718e-02 -2.62616634e-01 -7.12215602e-01 -5.90958476e-01
1.10937983e-01 -5.03670037e-01 -5.83928823e-01 1.12195957e+00
1.04818654e+00 -8.53989273e-02 2.59400249e-01 -1.18748772e+00
-8.00032914e-01 6.60313845e-01 -5.13393283e-01 1.84313759e-01
4.43530679e-01 3.31016243e-01 7.42828906e-01 -7.64655352e-01
-2.44140834e-01 1.16821468e+00 7.82717407e-01 1.15323496e+00
-1.09950781e+00 -1.15837407e+00 -1.73030332e-01 2.45945990e-01
-1.42338765e+00 -7.78696477e-01 7.13868320e-01 2.78752595e-02
6.71958447e-01 3.56733203e-01 3.78575146e-01 1.58126485e+00
-1.39469892e-01 8.01341891e-01 9.00168002e-01 -1.73803449e-01
1.08707041e-01 3.87380004e-01 2.77025759e-01 4.19865727e-01
3.74541163e-01 1.24401987e-01 -4.47122157e-01 -5.65038860e-01
5.49985647e-01 2.43748680e-01 -5.35618186e-01 2.02019215e-01
-7.64594853e-01 8.85599792e-01 2.44646236e-01 3.70634407e-01
-1.80982098e-01 -2.07816839e-01 4.24102306e-01 7.72645324e-02
-3.00356522e-02 3.40381801e-01 -4.56115186e-01 -2.86677122e-01
-7.15815663e-01 -2.22743601e-02 1.00538075e+00 2.29359969e-01
1.47642523e-01 5.76344192e-01 1.21102579e-01 7.41473734e-01
3.60396683e-01 7.99401164e-01 8.62035275e-01 -2.09600493e-01
4.43726443e-02 6.11423731e-01 9.64037627e-02 -8.73937368e-01
-2.77944237e-01 -3.71604651e-01 -1.26305449e+00 1.25508592e-01
5.52171290e-01 -2.54335999e-01 -7.64829576e-01 1.58699608e+00
1.38832241e-01 3.17750841e-01 1.05537787e-01 5.80421269e-01
6.17907882e-01 3.26952964e-01 -8.89889058e-03 1.69775024e-01
1.64429140e+00 -4.73605096e-01 -7.45183468e-01 7.57740636e-04
1.78814158e-02 -7.06067085e-01 1.42670143e+00 5.38394392e-01
-6.74851656e-01 -3.62467051e-01 -1.19729233e+00 1.86169207e-01
-6.48214936e-01 -2.47428387e-01 7.04056084e-01 1.77067792e+00
-8.03866386e-01 5.31949699e-01 -6.44124746e-01 -2.30305329e-01
6.12848699e-01 8.50514591e-01 -3.99782836e-01 3.04919749e-01
-1.49390590e+00 7.39457190e-01 -8.04418921e-02 1.96050048e-01
-6.91714585e-01 -3.56357276e-01 -7.58516967e-01 4.17787462e-01
1.00922778e-01 -2.50678241e-01 1.20137298e+00 -5.74456811e-01
-1.76106143e+00 6.59864008e-01 -1.56545177e-01 -5.57509065e-01
2.67173529e-01 -2.32327282e-01 -8.25732768e-01 -2.63723165e-01
-4.51447219e-01 5.79427667e-02 1.11739767e+00 -8.87497783e-01
-3.72001141e-01 -3.23731333e-01 -1.07060596e-01 -3.79114449e-01
-9.81067717e-01 2.74641514e-01 1.34644985e-01 -4.82590675e-01
-2.06621632e-01 -6.72339916e-01 1.69058368e-02 -2.94466466e-01
-6.08976126e-01 -1.51037201e-01 1.35807681e+00 -7.52929151e-01
1.29622424e+00 -2.20599604e+00 -5.22397161e-01 2.91152954e-01
3.30184191e-01 1.10234153e+00 -1.20341149e-03 2.23576143e-01
-3.14058028e-02 4.99167442e-01 -1.94895178e-01 -3.32028806e-01
2.34320518e-02 -4.05159593e-02 -4.98386085e-01 3.92396569e-01
1.53715163e-02 6.13066792e-01 -6.50623977e-01 5.42867277e-03
1.17292225e-01 8.55349362e-01 -3.54507893e-01 2.66686857e-01
4.30134118e-01 3.64147484e-01 -1.97675958e-01 8.02820921e-01
8.86068344e-01 4.26205993e-02 1.86566770e-01 8.35053399e-02
2.50676692e-01 4.90078986e-01 -1.07562971e+00 9.24748302e-01
-3.40062559e-01 3.60482067e-01 1.72973916e-01 -7.03323305e-01
1.06888247e+00 5.46901822e-01 2.33085807e-02 -3.59275222e-01
5.01053691e-01 2.19227597e-01 2.36773446e-01 -1.93507969e-01
2.79812306e-01 -4.48684245e-02 -1.24346204e-01 6.03744566e-01
-1.25176713e-01 4.90605235e-01 -6.00569010e-01 -9.08450931e-02
1.02889919e+00 -5.16133666e-01 2.66130626e-01 1.28315911e-01
9.59226370e-01 -9.54963207e-01 6.14271700e-01 7.84259379e-01
-6.40259206e-01 3.89801562e-01 2.41745353e-01 -5.56308806e-01
-6.49181128e-01 -9.21372652e-01 6.97627440e-02 7.90503442e-01
-2.54755374e-02 -2.46097639e-01 -1.03972077e+00 -1.04317081e+00
6.91468120e-02 4.01791722e-01 -2.54506499e-01 -5.82504332e-01
-5.53710401e-01 -7.27379560e-01 1.50908160e+00 6.23201907e-01
9.74713862e-01 -1.09476817e+00 -8.89676884e-02 7.22446963e-02
-2.45594963e-01 -1.11717510e+00 -6.69951439e-01 -9.98244528e-03
-4.54141676e-01 -9.75925028e-01 -7.21221447e-01 -5.53910732e-01
2.19937548e-01 3.92169505e-01 4.95040804e-01 4.20608729e-01
-3.27122696e-02 1.12391680e-01 -1.04413882e-01 -5.24604857e-01
-5.17998278e-01 4.01359409e-01 7.69236088e-01 4.95350540e-01
9.64646995e-01 -6.84083223e-01 -6.15804613e-01 3.75295252e-01
-7.95321763e-01 -6.63276136e-01 5.35860300e-01 9.06817615e-01
-2.38324389e-01 1.22176379e-01 8.79371345e-01 -6.13522649e-01
9.39633906e-01 -2.62030333e-01 -3.22206348e-01 1.07147880e-01
-6.42955363e-01 1.07780941e-01 7.39944160e-01 -9.04327095e-01
-8.02826226e-01 -6.87288716e-02 -8.56871963e-01 -1.10739760e-01
-3.46927345e-01 1.68307483e-01 -8.44184816e-01 -3.88271302e-01
6.07858419e-01 4.79161978e-01 9.13462937e-02 -8.41014087e-01
6.76912367e-02 1.18855023e+00 5.58803797e-01 -3.56048435e-01
1.06796670e+00 2.37625912e-01 -5.17812133e-01 -1.06324697e+00
-3.58073741e-01 -4.17326301e-01 -2.98462361e-01 6.75666854e-02
7.22086608e-01 -3.48905712e-01 -1.57451177e+00 1.26275861e+00
-1.31390274e+00 1.63387015e-01 2.56244600e-01 2.71685600e-01
2.50097007e-01 7.30276048e-01 -7.92479575e-01 -1.24518168e+00
-7.10824966e-01 -1.07251239e+00 6.01492524e-01 6.03780448e-01
-2.41628751e-01 -8.03813338e-01 -1.89906940e-01 4.21032399e-01
1.02169764e+00 -1.26081586e-01 7.24162817e-01 -1.43486202e+00
-6.13107942e-02 -5.71792126e-01 -1.20039679e-01 7.45778322e-01
4.93253559e-01 -2.47162104e-01 -1.45773041e+00 -3.77623826e-01
3.11831415e-01 -7.77571350e-02 6.59307361e-01 -1.21844940e-01
1.31189585e+00 -5.31717598e-01 -1.16388962e-01 5.51328242e-01
1.00192142e+00 4.22182202e-01 8.39204371e-01 2.41973460e-01
7.24291384e-01 2.20604151e-01 -3.77901107e-01 3.80928844e-01
1.66047037e-01 5.18662155e-01 4.97076839e-01 -3.46492603e-02
2.10615277e-01 -3.18139464e-01 4.21587318e-01 5.17338514e-01
-1.40470013e-01 -1.76683918e-01 -9.40549314e-01 2.09207818e-01
-1.30844116e+00 -1.11853147e+00 1.93758413e-01 2.43344975e+00
8.76174808e-01 3.05067897e-01 3.37074459e-01 6.43588185e-01
9.42694068e-01 9.53825489e-02 -6.37201786e-01 -4.88866895e-01
-7.39749754e-03 6.40357435e-01 4.09844667e-01 4.46604908e-01
-1.19349182e+00 9.58607614e-01 6.00968313e+00 5.91506064e-01
-1.47076035e+00 2.83997841e-02 5.94264686e-01 2.68600792e-01
2.80898046e-02 -4.02092040e-01 -1.06555843e+00 6.91024303e-01
1.20992446e+00 1.00687601e-01 4.72793221e-01 9.80845749e-01
1.54009879e-01 5.15844882e-01 -8.26642036e-01 1.14468193e+00
6.38950095e-02 -1.11305714e+00 1.19968854e-01 4.07134861e-01
-6.90237582e-02 -4.11443114e-01 2.59612322e-01 3.98808450e-01
2.29489505e-01 -1.18131280e+00 2.10248366e-01 7.98009709e-02
7.90420413e-01 -1.01176167e+00 1.04770684e+00 2.26548553e-01
-1.05833805e+00 -2.98727244e-01 -2.02032357e-01 1.06201224e-01
8.11064765e-02 2.37956092e-01 -6.91661119e-01 2.79504806e-01
6.78160787e-01 -4.45758291e-02 -3.23632896e-01 8.84184957e-01
-2.04309955e-01 9.36589897e-01 -2.07266659e-01 -2.91166067e-01
1.16286278e-02 2.69981205e-01 3.61974150e-01 1.13450015e+00
1.88885674e-01 3.71260988e-03 -5.57541139e-02 5.44747591e-01
-4.02174652e-01 -1.07676648e-01 -6.61874950e-01 -2.69920528e-01
6.19118214e-01 1.31259775e+00 -4.47956502e-01 -8.29649717e-03
-1.49104699e-01 1.13945401e+00 -1.59454510e-01 2.18126118e-01
-8.31714511e-01 -8.29181790e-01 1.14317417e+00 1.43256010e-02
-4.14882377e-02 -3.15640092e-01 -2.22202808e-01 -1.37065327e+00
-1.87807828e-01 -1.46575546e+00 1.95898056e-01 -6.59175068e-02
-1.34355593e+00 7.02183068e-01 -7.47957647e-01 -1.01318336e+00
5.15284836e-02 -7.97018349e-01 -8.18471074e-01 1.12216949e+00
-1.36062074e+00 -1.04247844e+00 -1.24769717e-01 9.88358617e-01
2.02654496e-01 -5.93954325e-01 1.30002642e+00 5.89068592e-01
-7.51655221e-01 1.31091011e+00 -1.24597952e-01 9.78109419e-01
7.13912249e-01 -1.07942176e+00 8.19375753e-01 8.94570112e-01
1.73741058e-01 1.09952712e+00 3.06827992e-01 -2.15630859e-01
-1.28305387e+00 -5.31883955e-01 8.75087619e-01 -3.14956605e-01
4.99590486e-01 -6.26972020e-01 -1.21850526e+00 2.31406838e-01
2.29623541e-01 -2.04267994e-01 1.23243701e+00 1.30300477e-01
-8.29692721e-01 -2.90431827e-01 -1.52801037e+00 6.14910841e-01
4.56683278e-01 -9.26550269e-01 -4.55314487e-01 -1.69937298e-01
4.37261462e-01 8.04899409e-02 -3.25170994e-01 1.55084953e-01
9.47038889e-01 -9.06540573e-01 1.03343558e+00 -8.58471632e-01
-3.63885045e-01 -1.05963118e-01 -6.51770085e-02 -9.72183585e-01
-7.87365958e-02 -9.89321291e-01 -3.97974163e-01 1.63579643e+00
3.46344948e-01 -1.08730924e+00 1.05815446e+00 5.94063699e-01
5.16003370e-01 -3.97393942e-01 -9.25928116e-01 -8.02089572e-01
-1.17541440e-02 -2.99745709e-01 1.06210172e+00 1.20942593e+00
6.98326901e-02 4.53335524e-01 -7.24418044e-01 3.96192610e-01
6.40651762e-01 -5.98477125e-01 5.92038572e-01 -1.36413586e+00
-1.95730940e-01 -6.82199895e-01 -5.27898371e-01 -9.54583108e-01
1.35880202e-01 -4.97713983e-01 -2.47210726e-01 -8.60075295e-01
-5.97452819e-02 -2.06896320e-01 -5.92272282e-01 7.41843998e-01
-1.35858923e-01 2.95366317e-01 2.80073509e-02 4.04268652e-02
5.34608886e-02 4.37891692e-01 5.43787956e-01 -4.28197265e-01
-2.74544895e-01 6.32124186e-01 -1.11445522e+00 8.11963797e-01
1.04348648e+00 -3.04735392e-01 -1.86114281e-01 -9.86839458e-03
-2.27575421e-01 -1.93859324e-01 1.39380097e-01 -9.66980100e-01
2.18555689e-01 1.29790723e-01 4.13161188e-01 -1.85388416e-01
2.49035999e-01 -8.40830445e-01 -4.03259695e-01 6.89551473e-01
-1.14941075e-01 1.23149008e-01 2.88407892e-01 3.22465330e-01
1.20617729e-02 -6.23750314e-02 9.17075634e-01 9.10077244e-02
-2.93490797e-01 3.66585106e-01 -7.00554609e-01 -3.96342456e-01
4.60165590e-01 -1.88507006e-01 -4.00843024e-01 -6.31044626e-01
-6.63895786e-01 -1.38732955e-01 1.40845269e-01 5.53898752e-01
8.78411949e-01 -1.23197055e+00 -5.06998599e-01 6.30713046e-01
-3.62877667e-01 -5.10140657e-01 -1.54371532e-02 4.20850933e-01
-2.49131158e-01 2.28808865e-01 -1.92298248e-01 -2.63064355e-01
-1.53624439e+00 5.75711548e-01 5.40700436e-01 -1.65670216e-01
-3.71951520e-01 6.88502908e-01 -2.55956292e-01 -5.69192290e-01
5.93380690e-01 8.40765387e-02 -3.24775428e-01 -1.09870210e-01
1.22780025e+00 2.40508452e-01 1.41174138e-01 -4.79345381e-01
-6.36519849e-01 8.18739608e-02 -3.42555940e-01 -3.75013724e-02
1.00730598e+00 1.91384941e-01 3.34629528e-02 -3.44812274e-02
8.83141637e-01 4.28015798e-01 -6.90605462e-01 -1.52091637e-01
2.12390423e-01 -3.57743651e-01 -2.77538031e-01 -8.67021680e-01
-1.10655189e+00 1.25996816e+00 8.90490115e-01 7.81897366e-01
9.47698832e-01 -5.62210262e-01 1.37621605e+00 6.07495606e-01
2.95539439e-01 -5.87318480e-01 1.21984564e-01 6.20259404e-01
3.60434741e-01 -1.23668969e+00 -4.99833435e-01 3.45002022e-03
-5.25328994e-01 1.08770251e+00 6.16351664e-01 1.90864056e-01
7.20642149e-01 4.19676006e-01 3.23884875e-01 2.05596089e-01
-7.54629672e-02 1.01309024e-01 6.95736781e-02 9.64514554e-01
2.77603596e-01 -1.59655109e-01 -7.67612085e-02 1.10558748e+00
-7.88603425e-02 -4.22455579e-01 4.97009128e-01 6.33025944e-01
-4.47566122e-01 -1.53734779e+00 -5.21982610e-01 1.08360857e-01
-1.03211093e+00 -3.39355946e-01 -6.99721098e-01 4.39732134e-01
-2.37632349e-01 1.13518369e+00 -3.49100828e-01 -9.23852861e-01
1.97276637e-01 5.63060164e-01 -1.37700781e-01 -2.98999608e-01
-7.88128853e-01 -3.01545948e-01 -2.02072948e-01 -1.98652521e-01
-1.73251584e-01 -2.99093366e-01 -1.04281759e+00 -9.36428249e-01
-6.25500500e-01 2.85767853e-01 7.37210274e-01 1.01397026e+00
3.53927314e-01 1.20080322e-01 9.81004417e-01 -7.08895504e-01
-9.72977281e-01 -9.06606019e-01 -4.99093056e-01 9.43816006e-02
4.29110646e-01 -4.31448817e-01 -4.85681444e-01 -2.94421822e-01] | [14.017158508300781, 5.800908088684082] |
03527940-db7e-4b56-87b4-4d38d344b841 | enhancement-or-super-resolution-learning | 2201.08197 | null | https://arxiv.org/abs/2201.08197v1 | https://arxiv.org/pdf/2201.08197v1.pdf | Enhancement or Super-Resolution: Learning-based Adaptive Video Streaming with Client-Side Video Processing | The rapid development of multimedia and communication technology has resulted in an urgent need for high-quality video streaming. However, robust video streaming under fluctuating network conditions and heterogeneous client computing capabilities remains a challenge. In this paper, we consider an enhancement-enabled video streaming network under a time-varying wireless network and limited computation capacity. "Enhancement" means that the client can improve the quality of the downloaded video segments via image processing modules. We aim to design a joint bitrate adaptation and client-side enhancement algorithm toward maximizing the quality of experience (QoE). We formulate the problem as a Markov decision process (MDP) and propose a deep reinforcement learning (DRL)-based framework, named ENAVS. As video streaming quality is mainly affected by video compression, we demonstrate that the video enhancement algorithm outperforms the super-resolution algorithm in terms of signal-to-noise ratio and frames per second, suggesting a better solution for client processing in video streaming. Ultimately, we implement ENAVS and demonstrate extensive testbed results under real-world bandwidth traces and videos. The simulation shows that ENAVS is capable of delivering 5%-14% more QoE under the same bandwidth and computing power conditions as conventional ABR streaming. | ['Shuoyao Wang', 'Yang Jiang', 'Junyan Yang'] | 2022-01-20 | null | null | null | null | ['video-enhancement'] | ['computer-vision'] | [ 2.13586852e-01 -4.62610364e-01 -2.26420134e-01 -4.78689194e-01
-7.78002858e-01 -2.02905849e-01 -2.22574413e-01 -2.22866684e-01
-4.65580434e-01 7.18039334e-01 1.07202083e-01 -3.43477428e-01
-1.67476609e-01 -7.29757726e-01 -6.37511969e-01 -1.00061131e+00
-7.58678675e-01 -3.05968612e-01 5.33029675e-01 -2.70646960e-01
2.74967700e-01 3.93668115e-01 -1.48957682e+00 4.02623832e-01
5.36206067e-01 1.51921213e+00 5.42383730e-01 1.35174930e+00
1.63302466e-01 1.05511403e+00 -7.01615572e-01 -2.89431870e-01
2.47941464e-01 -1.03286751e-01 -5.71982324e-01 2.85949409e-01
-1.52283415e-01 -1.22879648e+00 -7.63852000e-01 7.30014265e-01
9.54526842e-01 1.60095692e-01 1.18236899e-01 -1.48130214e+00
-1.23568207e-01 4.92107689e-01 -6.97400928e-01 9.06697750e-01
3.73520911e-01 1.83032766e-01 8.15261424e-01 -5.46833098e-01
2.99507618e-01 1.07590282e+00 2.50265568e-01 5.69193304e-01
-9.91291642e-01 -5.67140698e-01 -1.25245070e-02 9.59430397e-01
-1.12512648e+00 -7.21434474e-01 3.96894306e-01 1.65406257e-01
9.28134561e-01 1.87322274e-01 3.44495863e-01 6.40713692e-01
4.22518015e-01 7.90744960e-01 3.95023048e-01 -1.71606854e-01
6.20612919e-01 -1.87847003e-01 -5.57205915e-01 1.60253001e-03
-3.62322599e-01 -1.28096724e-02 -6.95304573e-01 1.32090017e-01
9.97909307e-01 -2.39227816e-01 -5.15821576e-01 1.08081512e-02
-7.80317128e-01 5.05329013e-01 -1.75550599e-02 -1.48240462e-01
-9.89942372e-01 4.51715201e-01 5.92006028e-01 6.64303899e-01
2.54310012e-01 -3.11519533e-01 -2.62860149e-01 -6.87164903e-01
-1.21110272e+00 7.22762495e-02 6.71368241e-01 1.24792254e+00
2.10904121e-01 3.88231426e-01 -8.10993761e-02 8.18337023e-01
3.04578871e-01 5.70499957e-01 2.07444474e-01 -1.98601890e+00
5.34846246e-01 -4.46089685e-01 1.77842304e-01 -7.48090029e-01
-2.30561141e-02 -1.93053707e-01 -1.01461089e+00 4.67081428e-01
-2.65966326e-01 -5.41797698e-01 -2.23468333e-01 1.59832215e+00
1.26392052e-01 4.18038011e-01 4.27061766e-01 1.23272002e+00
4.05907810e-01 1.43468201e+00 1.04735985e-01 -9.59662199e-01
9.02665675e-01 -8.96434903e-01 -1.06759381e+00 2.94177115e-01
-2.23607849e-02 -6.45758510e-01 5.08449554e-01 6.15416765e-01
-1.94605112e+00 -5.72047830e-01 -9.67205107e-01 4.90632594e-01
6.59146309e-01 -7.25995064e-01 -1.90045740e-02 6.38931870e-01
-1.48633480e+00 4.99842614e-01 -7.28857517e-01 3.24769504e-02
3.17840278e-01 4.63250369e-01 1.67159691e-01 -1.92312568e-01
-1.17485738e+00 1.40020669e-01 1.02922328e-01 -3.13832819e-01
-1.24035668e+00 -7.36446023e-01 -3.23281944e-01 5.29491961e-01
5.22883475e-01 -4.85947222e-01 1.58960235e+00 -1.26900172e+00
-1.81615162e+00 1.94170140e-02 -5.31868264e-02 -4.44468826e-01
5.47908723e-01 -1.00261629e-01 -7.03411162e-01 9.94513273e-01
-4.01813537e-01 4.19373482e-01 9.47921216e-01 -1.09447718e+00
-1.19585264e+00 6.32750057e-03 2.20573336e-01 3.86619300e-01
-4.26522791e-01 5.87291420e-01 -5.31171501e-01 -4.61326510e-01
-1.65632501e-01 -2.60037154e-01 -1.97731972e-01 5.08899212e-01
5.63878059e-01 1.13622144e-01 1.03985190e+00 -7.53481686e-01
1.41182101e+00 -2.24254870e+00 -1.24055631e-01 -3.00627835e-02
2.35860378e-01 3.92423928e-01 -2.98564643e-01 2.68496066e-01
2.77809054e-01 2.80121773e-01 2.02314883e-01 -5.69018945e-02
-4.28714514e-01 4.85800803e-01 1.89025048e-02 1.34989277e-01
1.44220158e-01 2.11186647e-01 -8.40763569e-01 -7.10170627e-01
-7.01688826e-02 6.38892651e-01 -8.75814795e-01 7.93652594e-01
-1.53457113e-02 2.96253890e-01 -3.49025965e-01 5.97449362e-01
1.01655447e+00 -2.80646026e-01 8.83486494e-02 -1.28079623e-01
-2.73514837e-01 -4.29908484e-02 -1.39462376e+00 1.11597025e+00
-6.38044238e-01 8.43198419e-01 8.96437407e-01 -1.00210226e+00
5.06360352e-01 8.07619393e-01 7.54954219e-01 -1.13715661e+00
7.85648897e-02 1.04701526e-01 -1.45147920e-01 -1.13919616e+00
6.58233762e-01 -1.03649467e-01 7.59167373e-01 2.47981265e-01
-9.22520906e-02 4.06336814e-01 2.33836934e-01 2.32798681e-01
1.36444795e+00 -1.50479615e-01 -1.19337216e-01 2.13766187e-01
2.62790889e-01 -7.18412697e-01 7.90842474e-01 6.13972843e-01
-7.35589325e-01 2.46590704e-01 4.66399342e-01 6.98069558e-02
-1.39600623e+00 -1.31434166e+00 8.50607976e-02 1.50740480e+00
4.82210606e-01 1.50593132e-01 -6.69825017e-01 2.91196048e-01
-5.18276811e-01 5.90960622e-01 2.88718104e-01 4.19885367e-02
-5.75386763e-01 -3.69544297e-01 3.19830030e-01 4.48248297e-01
7.08633423e-01 -1.08090317e+00 -9.43771780e-01 8.09741199e-01
-4.35973138e-01 -1.60484254e+00 -3.60927373e-01 -2.33172998e-01
-8.80476713e-01 -4.37342167e-01 -9.08545852e-01 -7.06597090e-01
-2.22458482e-01 6.85499847e-01 1.02645981e+00 -5.25644906e-02
4.51661320e-03 4.91177410e-01 -6.56810582e-01 2.66512800e-02
-6.52906954e-01 -3.48556012e-01 -7.87081569e-02 1.83463007e-01
-2.48118758e-01 -8.77199352e-01 -9.87479866e-01 3.07874858e-01
-1.11695242e+00 -1.08423769e-01 4.74926353e-01 4.43488777e-01
5.25038362e-01 5.37460148e-01 1.03928196e+00 3.56317535e-02
6.12490237e-01 -7.38369405e-01 -5.07569671e-01 -6.69863913e-03
-3.95878971e-01 -2.93734908e-01 6.90542281e-01 -4.79904145e-01
-1.42671454e+00 -6.53845429e-01 -4.86697674e-01 -4.13412571e-01
1.66905858e-02 2.97711760e-01 -2.84298867e-01 1.00333385e-01
2.34864384e-01 2.93754339e-01 6.82942346e-02 -1.14078388e-01
-1.36986658e-01 1.22374403e+00 5.06203175e-01 -3.03900838e-01
1.95382401e-01 3.96718293e-01 -3.63014713e-02 -1.02116418e+00
-1.10200256e-01 -4.73876566e-01 8.55734199e-02 -8.74229908e-01
7.87015378e-01 -1.18003368e+00 -1.11850429e+00 4.12520140e-01
-1.12502432e+00 -4.69599217e-01 8.49293619e-02 6.76729798e-01
-9.89322960e-01 5.09210944e-01 -1.02977824e+00 -9.61207330e-01
-5.05175829e-01 -1.30722380e+00 5.98651886e-01 4.39921707e-01
4.98563260e-01 -5.70019722e-01 -2.37527192e-01 3.20062518e-01
8.76536369e-01 -1.14081331e-01 6.15721166e-01 1.85610279e-02
-8.03156853e-01 3.10068578e-01 -7.21815825e-01 7.12410867e-01
-1.21606745e-01 3.19448799e-01 -7.11354434e-01 -5.41397035e-01
1.21976547e-01 -2.38290012e-01 2.66568452e-01 6.25323713e-01
1.50199687e+00 -4.13942218e-01 3.65405917e-01 8.05115283e-01
1.77026427e+00 5.95467269e-01 7.99535692e-01 1.43631250e-01
1.41934037e-01 2.99126893e-01 7.49729276e-01 1.34126925e+00
2.09538296e-01 5.10582447e-01 1.00588882e+00 2.28240922e-01
9.58788022e-02 3.77850354e-01 7.84489453e-01 8.86804640e-01
-2.90284634e-01 -1.01049888e+00 -4.59138542e-01 4.91759449e-01
-1.72575772e+00 -1.35635531e+00 8.76759365e-02 1.93684638e+00
6.99972153e-01 2.80803323e-01 2.84251511e-01 4.64472413e-01
8.95293295e-01 1.03183717e-01 -6.70605600e-01 -7.82822967e-01
-4.43107970e-02 9.52955112e-02 5.73713660e-01 2.02355847e-01
-4.83067691e-01 4.13443923e-01 6.08050203e+00 9.00916278e-01
-1.14123523e+00 2.42428947e-02 6.21833086e-01 -3.32652479e-01
1.72425415e-02 -6.13107562e-01 -1.85069963e-01 5.93918145e-01
1.69939399e+00 -5.27864754e-01 7.48343527e-01 7.63761222e-01
1.24992764e+00 -6.69711828e-02 -8.19501638e-01 1.01579130e+00
-2.89513767e-01 -1.28978062e+00 3.87202427e-02 -1.01585627e-01
4.12873119e-01 -5.02180792e-02 -1.18096530e-01 1.60793930e-01
-1.58816934e-01 -5.61510205e-01 7.22835422e-01 3.03161860e-01
9.96115208e-01 -9.00914431e-01 7.75242746e-01 2.90962875e-01
-1.10188305e+00 -4.82276261e-01 -3.46286863e-01 2.19620727e-02
8.45776081e-01 2.28328779e-01 -2.28463635e-01 2.14205846e-01
1.08572090e+00 3.87713164e-01 2.31758699e-01 1.33960283e+00
2.88551062e-01 9.68220472e-01 -4.57922295e-02 1.46935672e-01
7.53430277e-02 1.61942914e-01 5.43874860e-01 1.35321510e+00
5.04324317e-01 7.21566498e-01 8.92979726e-02 2.93731987e-01
-3.63841444e-01 1.66855170e-03 2.54857957e-01 2.44748056e-01
4.92433280e-01 9.69354451e-01 -2.49415874e-01 -5.14392138e-01
-7.13104963e-01 1.07280087e+00 -3.57495695e-01 6.31191134e-01
-1.18668830e+00 -3.98184717e-01 1.06557214e+00 2.53013134e-01
5.59928000e-01 -2.48938814e-01 1.55183434e-01 -6.97177052e-01
5.83768748e-02 -9.49863255e-01 1.08920924e-01 -9.50846970e-01
-8.06078315e-01 4.91451055e-01 -1.92718133e-01 -1.10884738e+00
-4.13469076e-02 -1.92352816e-01 -4.83161330e-01 5.07107615e-01
-2.08022046e+00 -1.19483441e-01 -5.75516820e-01 8.80869150e-01
1.08954835e+00 -6.18629158e-02 2.24031657e-01 1.02039719e+00
-5.25531232e-01 6.24485791e-01 5.41191339e-01 -1.82363644e-01
6.62041485e-01 -6.87683463e-01 -2.45343909e-01 8.53367507e-01
-7.20672965e-01 -5.07526100e-01 9.85756695e-01 -7.58755878e-02
-1.60043883e+00 -8.89825344e-01 1.51680395e-01 8.52707088e-01
6.59356654e-01 2.28480890e-01 -1.13986766e+00 -1.14851423e-01
5.16932011e-01 2.00579628e-01 5.97277522e-01 -9.23519135e-01
9.97889414e-02 -4.05617893e-01 -1.34014118e+00 4.28739965e-01
8.68140459e-01 -1.43564031e-01 2.26322994e-01 -7.51617774e-02
1.08912241e+00 -1.06329173e-01 -1.15673411e+00 3.20487887e-01
4.86781716e-01 -1.05436945e+00 5.93084395e-01 -5.65823078e-01
6.00750387e-01 5.40429465e-02 -6.52761400e-01 -9.57963169e-01
-3.18816274e-01 -8.52125049e-01 -3.42978776e-01 1.16715968e+00
2.85335109e-02 6.81606606e-02 8.19578707e-01 5.25240362e-01
-9.73050892e-02 -6.61202192e-01 -1.13814533e+00 -7.69095600e-01
-3.66581857e-01 -6.82403088e-01 2.70627916e-01 2.09921136e-01
3.73995840e-03 -2.15326510e-02 -4.35120434e-01 5.62250793e-01
7.28037357e-01 -3.00308973e-01 3.79031092e-01 -4.91094917e-01
-5.09708464e-01 -3.16369146e-01 -3.03959280e-01 -1.49413705e+00
-1.38013542e-01 -8.50427300e-02 2.17788473e-01 -1.43631387e+00
1.90043211e-01 -1.63227975e-01 -4.56953287e-01 -1.87566727e-01
2.47402377e-02 -1.50394393e-02 2.93588310e-01 4.20745835e-02
-1.13085365e+00 6.86005235e-01 1.13104868e+00 2.62615442e-01
-2.28114158e-01 3.33866000e-01 -1.09018765e-01 3.92459363e-01
9.69648302e-01 -3.00887048e-01 -4.40436929e-01 -6.37489021e-01
2.19491169e-01 1.27726340e+00 9.34925899e-02 -9.71318066e-01
3.79891783e-01 -6.63358927e-01 -1.24961279e-01 -1.50638446e-01
3.47668141e-01 -1.09692907e+00 -1.89680859e-01 4.34488416e-01
-5.05051196e-01 1.96858451e-01 9.46205016e-03 6.83078766e-01
-3.17783922e-01 -7.20782951e-02 1.11773407e+00 3.10824692e-01
-1.01441872e+00 5.23019612e-01 -1.19625247e+00 7.81176612e-02
9.98558521e-01 -3.94944549e-01 2.40502879e-02 -1.09797871e+00
-7.56843805e-01 4.34694052e-01 1.09044559e-01 1.70834258e-01
1.06963372e+00 -9.02549446e-01 -1.09430420e+00 -1.04498342e-01
-3.55607778e-01 -4.89995420e-01 7.67868221e-01 4.84374046e-01
-8.33284378e-01 -1.81951120e-01 -5.06219923e-01 -6.13012075e-01
-1.71820033e+00 5.45452237e-01 2.57246166e-01 -1.20720319e-01
-1.89531595e-01 5.34115076e-01 -3.14950049e-01 8.45548153e-01
6.11120880e-01 1.76869407e-01 -9.08601359e-02 -3.26194137e-01
8.63414407e-01 1.04367316e+00 -1.45498291e-01 -4.83166218e-01
4.19855416e-02 -5.24537452e-02 1.30490392e-01 -2.80187905e-01
1.45608366e+00 -9.59658504e-01 2.73653716e-01 -9.55368802e-02
1.25744474e+00 -4.30086285e-01 -1.70159554e+00 -2.80765206e-01
-3.78197730e-01 -6.89584374e-01 5.82695186e-01 -4.11448300e-01
-1.42031455e+00 9.81507897e-01 1.03707862e+00 4.38357770e-01
1.79518259e+00 -3.56552005e-01 1.15295041e+00 5.51766204e-03
3.57412785e-01 -1.26005328e+00 4.59947973e-01 1.45733356e-01
5.90588629e-01 -1.24360049e+00 -2.32661262e-01 -4.36505318e-01
-7.11794674e-01 1.20015132e+00 4.98810410e-01 1.01188831e-01
6.96584702e-01 6.71547592e-01 -6.31142929e-02 4.73935157e-01
-1.28566587e+00 7.18488768e-02 -6.05413377e-01 5.80157697e-01
2.96464443e-01 -4.00029793e-02 -1.49534971e-01 1.84169069e-01
3.76480728e-01 3.82408828e-01 1.01415873e+00 8.37011516e-01
-9.28540587e-01 -7.49495208e-01 -3.04253757e-01 2.52032459e-01
-1.05192482e+00 1.66776050e-02 8.30112576e-01 1.04215167e-01
-2.19254583e-01 1.51772368e+00 3.67046773e-01 -2.74850905e-01
2.05027446e-01 -6.02267623e-01 1.24209724e-01 -4.16447148e-02
-2.11689726e-01 3.22481334e-01 1.44690927e-03 -8.66974115e-01
-6.55162752e-01 -2.89772838e-01 -1.54987097e+00 -8.48476410e-01
-7.23741874e-02 1.53085832e-02 9.42114413e-01 6.13181531e-01
4.14516300e-01 9.24547493e-01 1.22781825e+00 -7.66011119e-01
-5.41145861e-01 -3.78487021e-01 -6.63933873e-01 2.02696085e-01
6.89164281e-01 -9.79187265e-02 -3.89175087e-01 3.54276896e-01] | [11.156734466552734, -1.6554429531097412] |
f95637f0-84a7-4429-9a86-c269faec08ba | cosmetic-aware-makeup-cleanser | 2004.09147 | null | https://arxiv.org/abs/2004.09147v1 | https://arxiv.org/pdf/2004.09147v1.pdf | Cosmetic-Aware Makeup Cleanser | Face verification aims at determining whether a pair of face images belongs to the same identity. Recent studies have revealed the negative impact of facial makeup on the verification performance. With the rapid development of deep generative models, this paper proposes a semanticaware makeup cleanser (SAMC) to remove facial makeup under different poses and expressions and achieve verification via generation. The intuition lies in the fact that makeup is a combined effect of multiple cosmetics and tailored treatments should be imposed on different cosmetic regions. To this end, we present both unsupervised and supervised semantic-aware learning strategies in SAMC. At image level, an unsupervised attention module is jointly learned with the generator to locate cosmetic regions and estimate the degree. At feature level, we resort to the effort of face parsing merely in training phase and design a localized texture loss to serve complements and pursue superior synthetic quality. The experimental results on four makeuprelated datasets verify that SAMC not only produces appealing de-makeup outputs at a resolution of 256*256, but also facilitates makeup-invariant face verification through image generation. | ['Yi Li', 'Junchi Yu', 'Tieniu Tan', 'Huaibo Huang', 'Ran He'] | 2020-04-20 | null | null | null | null | ['face-parsing'] | ['computer-vision'] | [ 3.21364671e-01 3.79783571e-01 5.95728755e-02 -7.32606649e-01
-7.38765836e-01 -5.68911970e-01 4.85925376e-01 -6.19111359e-01
3.64336491e-01 4.09987777e-01 3.56403232e-01 1.70291066e-01
1.41066328e-01 -7.51924694e-01 -7.70839751e-01 -9.25582469e-01
4.45829958e-01 -6.54772902e-03 -6.36237741e-01 -1.87425137e-01
1.04911454e-01 7.61248291e-01 -1.55601156e+00 4.19649184e-01
8.79245043e-01 1.07429636e+00 -7.54870251e-02 2.89105952e-01
-1.01665528e-02 4.15798008e-01 -5.13718963e-01 -1.21111810e+00
5.69371760e-01 -7.03774691e-01 -3.28946888e-01 5.04113615e-01
8.05320859e-01 -4.15903419e-01 -1.16099015e-01 1.49501216e+00
6.95943296e-01 -4.68658507e-01 7.37335742e-01 -1.38241243e+00
-1.30974925e+00 4.82935011e-01 -7.72568703e-01 -4.68299925e-01
8.84879902e-02 3.69512647e-01 6.56813502e-01 -1.08462346e+00
5.46391189e-01 1.47515488e+00 6.79550469e-01 1.07252347e+00
-1.07238126e+00 -8.86968374e-01 -3.28537519e-03 -1.29598424e-01
-1.43700707e+00 -1.05214393e+00 1.17202902e+00 -2.98363954e-01
1.91940710e-01 2.61025399e-01 2.86056608e-01 1.08521819e+00
1.18705630e-01 3.95430595e-01 1.34064186e+00 -3.31952602e-01
1.09234437e-01 2.24756569e-01 -5.34389436e-01 1.23276556e+00
2.83902466e-01 1.79645166e-01 -5.43275177e-01 1.52805045e-01
7.43466794e-01 -1.35267645e-01 -2.44972706e-01 -1.90426707e-01
-4.48242396e-01 6.62450671e-01 4.82798487e-01 3.24198194e-02
-2.06849247e-01 2.36096069e-01 5.03927730e-02 6.63913116e-02
4.13164765e-01 3.58447909e-01 -2.02643558e-01 5.00284314e-01
-9.22995210e-01 2.21027657e-02 3.67518157e-01 1.03560138e+00
7.95733094e-01 3.71330410e-01 -4.12837327e-01 9.40707684e-01
4.49021995e-01 8.87570441e-01 2.90731311e-01 -7.58278370e-01
3.67773294e-01 6.08254969e-01 -1.21614270e-01 -1.24100399e+00
7.86312446e-02 -4.88010943e-01 -8.63148689e-01 4.28803444e-01
1.62177789e-03 -2.66918510e-01 -1.25784004e+00 1.90636706e+00
3.54906797e-01 4.99681123e-02 -3.62470672e-02 9.81097698e-01
9.57071304e-01 1.57803312e-01 2.55180717e-01 -5.69268875e-02
1.57017565e+00 -7.10826218e-01 -9.24156666e-01 -1.69369817e-01
1.18974887e-01 -1.07909572e+00 7.33396530e-01 5.23835570e-02
-9.87904310e-01 -7.24414945e-01 -1.16618979e+00 -1.28794760e-01
-2.62928158e-01 6.04537249e-01 6.83586657e-01 1.23185837e+00
-1.26143503e+00 5.15013993e-01 -3.45520437e-01 -1.50881290e-01
8.48263621e-01 2.28171736e-01 -5.60572088e-01 -1.62592173e-01
-8.21870983e-01 6.39149070e-01 -1.47859871e-01 4.52899903e-01
-1.00859165e+00 -6.66790724e-01 -1.00385618e+00 3.39600928e-02
-1.17224595e-02 -6.93148911e-01 7.03993797e-01 -1.45723176e+00
-1.55692172e+00 1.11498404e+00 -2.76428670e-01 -1.52674178e-02
5.18523932e-01 -4.95415777e-02 -4.87561285e-01 6.28834264e-03
2.50549287e-01 7.82052577e-01 1.46719897e+00 -1.58415878e+00
-1.05008632e-01 -6.24152482e-01 -1.78251877e-01 2.46348251e-02
-1.51603416e-01 1.70621604e-01 -4.61086512e-01 -9.07334507e-01
8.76308903e-02 -8.58204782e-01 1.92814440e-01 2.18277916e-01
-6.54364824e-01 9.76485536e-02 7.59231925e-01 -9.48886037e-01
7.71974742e-01 -2.14233112e+00 -8.24129060e-02 2.24635646e-01
4.70840894e-02 1.90734833e-01 -4.94817764e-01 1.26961321e-01
-4.01757538e-01 2.75389761e-01 -1.88771218e-01 -5.95704794e-01
1.85026706e-03 -1.34574816e-01 -2.93940514e-01 6.06600344e-01
7.97471166e-01 1.09879303e+00 -3.67111742e-01 -5.66830575e-01
-9.16032940e-02 7.15177953e-01 -7.24248052e-01 1.44044489e-01
-9.02872756e-02 3.19129527e-01 -4.60033596e-01 1.18178165e+00
1.34342384e+00 -5.84869320e-03 2.95970768e-01 -5.67478359e-01
2.84476966e-01 -2.87104309e-01 -8.18556964e-01 1.58617938e+00
-2.34141499e-01 2.44697303e-01 3.72780174e-01 -6.00870371e-01
9.59762037e-01 2.57517815e-01 2.22690165e-01 -7.85880804e-01
3.71290714e-01 1.31170303e-01 -2.04484850e-01 -5.02206743e-01
2.78858840e-01 -3.62109125e-01 9.94893163e-02 1.24993943e-01
1.33863404e-01 -2.34931350e-01 -2.85686314e-01 -1.47509545e-01
5.26637971e-01 1.91307381e-01 -8.93290266e-02 -3.46812606e-01
5.42172134e-01 -5.51870108e-01 7.17141688e-01 2.16262251e-01
-2.74295539e-01 8.88137162e-01 3.76614094e-01 -5.54561056e-02
-9.69179332e-01 -1.00549984e+00 1.72041301e-02 7.18723953e-01
2.83241928e-01 -1.88542902e-01 -1.21142411e+00 -7.42648363e-01
1.99776366e-02 2.67777056e-01 -9.32344794e-01 -3.71658504e-01
-3.70435447e-01 -6.81334198e-01 5.95152974e-01 4.42960113e-01
8.77916276e-01 -6.90529943e-01 4.83069494e-02 -3.67507011e-01
-1.45323023e-01 -1.04895973e+00 -8.69092166e-01 -4.44059283e-01
-5.16909897e-01 -1.08868730e+00 -7.44561553e-01 -9.21111524e-01
1.10819447e+00 1.48099706e-01 7.79871225e-01 2.04982385e-01
-5.37214041e-01 2.21893623e-01 -2.02708498e-01 -5.07826090e-01
-4.41398948e-01 -3.44565421e-01 -6.33591488e-02 7.60983944e-01
1.31473422e-01 -5.09120047e-01 -8.32628191e-01 1.52890757e-01
-6.50615990e-01 1.62224695e-01 6.68359995e-01 7.62418747e-01
4.18446541e-01 2.06361115e-02 5.76020956e-01 -8.38996947e-01
5.42380869e-01 -1.37187138e-01 -5.17872930e-01 5.13533235e-01
-6.43335581e-01 2.29777798e-01 4.11152363e-01 -1.12117209e-01
-1.42879379e+00 1.97194144e-01 -2.39733770e-01 -6.02463484e-01
-6.84555322e-02 -1.82082191e-01 -8.04669440e-01 -3.17539155e-01
4.48837310e-01 3.70910227e-01 2.39630058e-01 -2.80501127e-01
4.56024915e-01 3.82369369e-01 4.43805814e-01 -5.61840236e-01
1.23620927e+00 6.58406615e-01 5.24415821e-02 -5.04174173e-01
-6.97945297e-01 2.31809840e-01 -3.07623893e-01 -4.53265965e-01
9.99392986e-01 -1.07733643e+00 -6.67638123e-01 6.69556916e-01
-1.06092346e+00 -2.43648123e-02 -2.62149647e-02 -7.29907900e-02
-2.14455351e-01 2.13990897e-01 -4.32632029e-01 -7.98349917e-01
-5.73352456e-01 -1.14164352e+00 1.45159316e+00 5.57742894e-01
2.33881533e-01 -6.17974401e-01 -4.01930332e-01 6.81421041e-01
4.97149199e-01 3.26399475e-01 8.10187399e-01 -1.96340941e-02
-7.51888752e-01 -7.10132197e-02 -4.27415758e-01 4.46082443e-01
4.96241778e-01 1.19410053e-01 -1.49236000e+00 -2.40990892e-01
-2.82025486e-02 -1.77555889e-01 7.78190672e-01 2.36849949e-01
1.27753663e+00 -5.06031215e-01 -9.74128470e-02 1.00900340e+00
1.58825457e+00 2.05187380e-01 1.00743937e+00 -3.03602248e-01
7.40098178e-01 8.68570685e-01 3.18494558e-01 2.86028236e-01
1.29196659e-01 5.55858731e-01 5.55297315e-01 -3.57527643e-01
-7.05582500e-01 -6.77555025e-01 3.29613298e-01 3.50134969e-01
-9.48678106e-02 -8.95793140e-02 -2.85057843e-01 3.83825868e-01
-1.22062504e+00 -1.00546205e+00 2.39162654e-01 1.71697569e+00
8.55486691e-01 -4.33653533e-01 -4.31477875e-01 -2.03365326e-01
9.77687001e-01 1.33235723e-01 -5.18687844e-01 -2.92857170e-01
-3.46370518e-01 4.35254782e-01 4.60003287e-01 3.17874640e-01
-1.01815772e+00 1.11203754e+00 5.80970621e+00 1.03626561e+00
-1.44373107e+00 1.16819941e-01 1.19965959e+00 5.95251396e-02
-4.80679035e-01 -2.24639550e-01 -7.24959254e-01 6.19773746e-01
3.50057721e-01 2.60849327e-01 4.02557015e-01 8.21916997e-01
1.78600937e-01 2.44516432e-01 -8.57778430e-01 1.24133956e+00
5.14614224e-01 -1.17682219e+00 1.96099967e-01 -2.21706321e-03
9.69078064e-01 -6.08706415e-01 5.70991874e-01 -3.65115516e-02
1.73861727e-01 -1.25213063e+00 1.01535928e+00 5.28998196e-01
1.26850784e+00 -7.04356492e-01 4.60669130e-01 -3.51378918e-01
-1.13870633e+00 3.88686508e-02 -2.62741327e-01 6.02074325e-01
-1.84635445e-01 4.91549999e-01 -9.10995722e-01 5.86412668e-01
4.40193862e-01 3.56318206e-01 -6.31727993e-01 5.67405999e-01
-6.10962987e-01 2.47714996e-01 1.21857062e-01 4.40972358e-01
-3.84888053e-01 -2.87148923e-01 2.43577451e-01 8.14431489e-01
5.03172278e-01 7.06905872e-02 -3.80896300e-01 1.24233496e+00
-5.72962999e-01 3.01357806e-02 -4.78220046e-01 -1.56352580e-01
4.07958359e-01 1.35333383e+00 -6.08883858e-01 -2.32556481e-02
-7.93354064e-02 1.34760118e+00 1.31316498e-01 3.72041821e-01
-1.05207002e+00 -1.30476162e-01 9.95321870e-01 2.37172291e-01
3.30512851e-01 1.80224806e-01 -5.74149489e-01 -1.05437398e+00
1.43684894e-01 -8.46121013e-01 -1.07874334e-01 -9.05765235e-01
-1.26121843e+00 6.00093663e-01 -6.08663738e-01 -9.91109908e-01
1.31376594e-01 -5.48635423e-01 -6.93463683e-01 1.08447754e+00
-1.84218323e+00 -1.69673252e+00 -4.82613266e-01 6.58530414e-01
5.07655561e-01 -2.40241334e-01 7.88255334e-01 4.53537405e-01
-7.00263917e-01 1.11277258e+00 -3.09075147e-01 4.65281755e-01
8.72049093e-01 -7.95868397e-01 2.08002627e-01 1.05325711e+00
-1.02268830e-01 7.51107574e-01 5.82781434e-01 -6.47087932e-01
-1.66529500e+00 -1.42315972e+00 8.90739858e-01 -2.37485379e-01
3.82916071e-02 -4.37751979e-01 -4.88499343e-01 2.81732321e-01
2.44261652e-01 7.67206103e-02 5.58598399e-01 -2.12488338e-01
-7.32527733e-01 -4.66593683e-01 -1.56365895e+00 5.84321320e-01
1.26872110e+00 -8.38339567e-01 -7.14677125e-02 1.69587612e-01
3.33018869e-01 -1.86694786e-01 -6.49272561e-01 5.52965164e-01
4.67591971e-01 -1.06460440e+00 8.20001841e-01 -3.04946423e-01
7.43829906e-01 -3.40068549e-01 -2.36965157e-02 -1.06991112e+00
-2.25221738e-01 -7.02757478e-01 4.40292567e-01 1.76318383e+00
3.35802644e-01 -4.46010828e-01 9.51860428e-01 8.06872606e-01
-8.76660496e-02 -5.15910864e-01 -6.30407095e-01 -4.02165532e-01
-1.62981525e-01 1.12839483e-01 9.71551478e-01 9.19411361e-01
-5.97949326e-01 9.18406062e-03 -4.16728437e-01 4.22536582e-01
7.16572344e-01 3.72500777e-01 6.06193841e-01 -9.08872008e-01
-2.96817310e-02 -3.61907631e-01 -3.24091852e-01 -6.10131025e-01
3.31075579e-01 -8.59934032e-01 2.62661837e-02 -1.10102212e+00
3.03440630e-01 -4.31683034e-01 8.16057548e-02 6.37888193e-01
-2.51120657e-01 5.71770430e-01 1.62345935e-02 -6.18341267e-02
-2.01923717e-02 7.53757358e-01 1.51517510e+00 -3.43438536e-01
3.38277549e-01 -3.57426912e-01 -1.28530884e+00 5.71077466e-01
6.98260367e-01 -1.64482400e-01 -5.55527210e-01 -5.43640792e-01
3.04331407e-02 -1.27280787e-01 5.39053202e-01 -7.81977773e-01
-9.67874974e-02 -1.20640501e-01 7.57285237e-01 -9.92964506e-02
4.48162317e-01 -6.44366920e-01 2.32651800e-01 2.71349579e-01
-1.96776718e-01 -1.86413720e-01 7.10267574e-02 3.28391433e-01
-2.28919283e-01 1.02086231e-01 1.20349228e+00 4.68378328e-02
-6.15787268e-01 5.78968108e-01 2.72546381e-01 -2.73872942e-01
9.94018972e-01 -2.79108346e-01 -3.34615290e-01 -2.75900304e-01
-4.89983857e-01 -2.13814721e-01 5.80152810e-01 5.67909122e-01
8.38596463e-01 -1.49001658e+00 -8.90818298e-01 7.21722543e-01
1.06100991e-01 -4.35802907e-01 5.54082692e-01 3.97276700e-01
-4.45960820e-01 1.98288634e-01 -2.41123229e-01 -1.78154513e-01
-1.57806730e+00 4.85436171e-01 5.96845686e-01 3.40904295e-01
-1.70895711e-01 1.29766226e+00 5.23120105e-01 -2.05056474e-01
-6.42238483e-02 7.19123380e-03 -5.33594862e-02 -5.58745526e-02
4.56701636e-01 -9.02260914e-02 1.68829471e-01 -8.35649073e-01
-2.67693847e-01 8.37802827e-01 8.81373137e-02 2.53699273e-01
1.17648470e+00 -1.50378689e-01 -1.64740741e-01 -6.19640470e-01
1.31389117e+00 3.34882826e-01 -1.40763450e+00 1.61363974e-01
-4.59381223e-01 -7.46578515e-01 -2.58728176e-01 -9.53018129e-01
-1.76224458e+00 6.40253901e-01 7.85341680e-01 -3.50313008e-01
1.47016490e+00 4.14198004e-02 5.80987275e-01 -4.02146786e-01
2.63776749e-01 -1.02212286e+00 -1.05806440e-02 -3.66943926e-02
1.21386743e+00 -1.25425482e+00 -2.04019874e-01 -8.67210448e-01
-6.18948400e-01 8.03312898e-01 7.06529379e-01 -3.17748301e-02
6.03659689e-01 3.10203552e-01 2.07861625e-02 -2.48204783e-01
-4.12835389e-01 -1.42023578e-01 3.99854690e-01 7.33635068e-01
3.88964772e-01 1.49174213e-01 -5.32033071e-02 6.41961098e-01
-3.76779616e-01 -2.37859711e-01 1.82552692e-02 5.28341472e-01
-8.36243406e-02 -1.24387825e+00 -4.40741092e-01 1.15940757e-01
-5.80391169e-01 -1.90106064e-01 -6.19156241e-01 3.87000054e-01
7.55367041e-01 1.00079417e+00 -1.98977754e-01 -4.49442178e-01
1.47471860e-01 1.28858343e-01 7.81621814e-01 -4.31685716e-01
-4.50800627e-01 -3.47719975e-02 -2.32219826e-02 -6.35236144e-01
-3.93127471e-01 -4.12664682e-01 -9.34400737e-01 -3.81925106e-01
-3.08299065e-01 -3.44227329e-02 8.43016863e-01 6.55124903e-01
6.80492878e-01 4.88455504e-01 9.03850675e-01 -4.81858552e-01
-4.60017413e-01 -8.07547331e-01 -7.29680359e-01 7.43719280e-01
2.64399499e-01 -5.67165852e-01 -1.41059712e-01 4.11103457e-01] | [12.762114524841309, 0.10756748914718628] |
d29aca32-d5f6-4a31-a78f-04a29d352bbe | physics-informed-representation-learning-for | 2304.12586 | null | https://arxiv.org/abs/2304.12586v1 | https://arxiv.org/pdf/2304.12586v1.pdf | Physics-Informed Representation Learning for Emergent Organization in Complex Dynamical Systems | Nonlinearly interacting system components often introduce instabilities that generate phenomena with new properties and at different space-time scales than the components. This is known as spontaneous self-organization and is ubiquitous in systems far from thermodynamic equilibrium. We introduce a theoretically-grounded framework for emergent organization that, via data-driven algorithms, is constructive in practice. Its building blocks are spacetime lightcones that capture how information propagates across a system through local interactions. We show that predictive equivalence classes of lightcones, local causal states, capture organized behaviors and coherent structures in complex spatiotemporal systems. Using our unsupervised physics-informed machine learning algorithm and a high-performance computing implementation, we demonstrate the applicability of the local causal states for real-world domain science problems. We show that the local causal states capture vortices and their power-law decay behavior in two-dimensional turbulence. We then show that known (hurricanes and atmospheric rivers) and novel extreme weather events can be identified on a pixel-level basis and tracked through time in high-resolution climate data. | ['James P. Crutchfield', 'Nalini Kumar', 'Karthik Kashinath', 'Adam Rupe'] | 2023-04-25 | null | null | null | null | ['physics-informed-machine-learning'] | ['graphs'] | [-2.24632591e-01 -3.15896332e-01 3.61360669e-01 1.44558772e-01
1.01409182e-01 -9.91696715e-01 1.12276089e+00 1.63750961e-01
2.42085367e-01 9.80640650e-01 5.57794034e-01 -2.07050353e-01
-5.09886384e-01 -8.75253797e-01 -4.27996993e-01 -1.11940753e+00
-1.16206527e+00 2.35061944e-01 2.15146318e-01 -5.07613122e-01
1.78561032e-01 7.05781341e-01 -1.59466469e+00 -2.08182111e-02
1.09427333e+00 4.07561839e-01 -7.15546310e-02 1.17954147e+00
9.05656293e-02 1.03234756e+00 -3.70119750e-01 5.55433989e-01
2.93900788e-01 -8.01495433e-01 -5.76684177e-01 -1.02966474e-02
2.42418990e-01 1.50733083e-01 -2.44802102e-01 8.73880267e-01
1.24090597e-01 2.51617670e-01 8.63489032e-01 -9.69098926e-01
-5.70093811e-01 7.83646107e-02 -4.02916461e-01 5.34837782e-01
3.45415063e-02 5.23524821e-01 9.09514308e-01 -7.53513873e-01
9.27608788e-01 1.28013110e+00 5.68619013e-01 8.83887857e-02
-1.58440840e+00 -2.34663025e-01 -1.16195910e-01 -9.13574994e-02
-1.03232789e+00 -3.26453626e-01 8.08263779e-01 -1.11485958e+00
9.89179492e-01 4.34853852e-01 1.03763413e+00 7.90296257e-01
1.08481908e+00 1.63474604e-02 1.59452331e+00 -3.45829457e-01
6.10144556e-01 -3.60184491e-01 2.86714613e-01 8.58134270e-01
4.14244652e-01 7.15929151e-01 -9.47388589e-01 -5.53119481e-01
8.16663325e-01 -3.66141379e-01 -2.74482101e-01 -4.66714740e-01
-1.30020392e+00 6.39478028e-01 5.77511251e-01 4.50057536e-01
-3.78572822e-01 3.31437916e-01 -6.34019300e-02 2.54179478e-01
7.42509425e-01 7.17612803e-01 -4.64457721e-01 -1.76384628e-01
-6.42808199e-01 4.20415074e-01 8.93825233e-01 5.01137018e-01
8.58397484e-01 1.03238732e-01 2.70027757e-01 1.53373465e-01
1.51534453e-01 1.02956510e+00 1.83557779e-01 -1.15235436e+00
-3.46446157e-01 5.05346358e-01 4.12571549e-01 -1.20935082e+00
-8.00800681e-01 -5.32553256e-01 -1.34998059e+00 4.99556571e-01
3.49788725e-01 -3.89777184e-01 -4.03728545e-01 1.78132355e+00
3.38556021e-01 4.49472368e-01 2.42498845e-01 9.18696642e-01
2.51562446e-01 9.56274331e-01 -2.47282133e-01 -7.69532859e-01
1.19742501e+00 -3.84758025e-01 -7.77307272e-01 3.12978327e-01
4.40358162e-01 -3.84023458e-01 1.03305531e+00 -9.63291600e-02
-8.82799685e-01 -1.27363175e-01 -6.84037089e-01 5.94135046e-01
-4.67631072e-01 -6.28110826e-01 5.26031554e-01 1.51494995e-01
-1.15815163e+00 6.26623452e-01 -1.10030735e+00 -6.87092423e-01
-1.57746553e-01 -2.75361598e-01 -9.72060040e-02 9.24411535e-01
-9.67692137e-01 5.49825430e-01 -2.57566780e-01 -1.17860690e-01
-9.02615190e-01 -7.96747625e-01 -4.13620591e-01 -1.27394572e-01
-2.01952904e-01 -8.26120079e-01 9.32505846e-01 -6.91417575e-01
-1.32618070e+00 5.41253388e-01 -4.62541938e-01 -3.67707312e-01
1.14556380e-01 -1.31552249e-01 -4.00335610e-01 2.26136982e-01
2.60172725e-01 -1.12111554e-01 7.56295562e-01 -1.49330199e+00
-3.45176905e-01 -1.96982488e-01 -3.42069656e-01 1.63859054e-01
-3.63707513e-01 -2.34353304e-01 6.46195054e-01 -5.72523355e-01
1.34502247e-01 -1.05186319e+00 -5.31390131e-01 -3.72308433e-01
-2.48113394e-01 1.96881220e-02 9.64186966e-01 6.21636286e-02
1.02326047e+00 -1.82817960e+00 4.95034635e-01 2.15056598e-01
7.06866324e-01 -2.56921291e-01 4.63952236e-02 8.27815294e-01
1.15232736e-01 3.67677569e-01 -5.17931998e-01 4.58121568e-01
-2.92226672e-01 1.82349458e-01 -8.03210199e-01 7.05573916e-01
3.03734481e-01 6.94720030e-01 -1.34145153e+00 -1.42255917e-01
2.43559510e-01 8.82945880e-02 -6.38545394e-01 1.70890659e-01
-9.89250839e-02 1.11655998e+00 -4.31051761e-01 1.79505527e-01
6.53970778e-01 -4.47795808e-01 -1.52374834e-01 2.84043133e-01
-1.01378834e+00 -1.45074492e-02 -8.33531082e-01 1.07474160e+00
-3.25842321e-01 1.01773787e+00 2.30491191e-01 -7.06376195e-01
6.71773612e-01 2.03593075e-01 7.44384766e-01 -4.83070910e-01
-3.27563912e-01 1.45301953e-01 1.35254249e-01 -7.44143188e-01
2.50487506e-01 -3.82592410e-01 -6.54143095e-02 6.14018142e-01
-2.10532453e-02 -4.12946701e-01 5.90984374e-02 3.57191145e-01
1.59755278e+00 -2.35451475e-01 4.34474438e-01 -1.40252686e+00
2.47100174e-01 3.89871150e-01 3.56441498e-01 1.05195546e+00
-2.39075691e-01 1.22826293e-01 5.83007276e-01 -8.78735960e-01
-1.23188281e+00 -1.52720535e+00 -5.00755429e-01 8.27971339e-01
3.75761569e-01 -6.09211206e-01 -3.80512983e-01 1.56204954e-01
-4.94964700e-03 2.51826018e-01 -7.71162450e-01 -6.39938489e-02
-6.19704783e-01 -1.23035765e+00 3.49330068e-01 -1.76621586e-01
4.57919538e-01 -1.23050618e+00 -1.15263379e+00 2.87979007e-01
8.47673323e-03 -8.37274969e-01 1.39895216e-01 2.69919813e-01
-9.44620252e-01 -1.00946593e+00 -2.51961291e-01 -1.98827952e-01
5.27715623e-01 1.35008410e-01 1.40328741e+00 -1.39671296e-01
-6.79343522e-01 4.82417703e-01 -2.03158826e-01 -2.47839376e-01
-4.69812512e-01 -3.77056152e-01 8.38612199e-01 -1.37941195e-02
-1.08491078e-01 -1.01681209e+00 -6.30935073e-01 1.58120185e-01
-7.17233062e-01 3.53146233e-02 1.11129796e-02 7.58408844e-01
1.36228547e-01 2.60582000e-01 2.67089635e-01 -2.00683087e-01
8.59390557e-01 -5.76307952e-01 -9.50180352e-01 -2.11176407e-02
-2.01124847e-01 4.13584203e-01 7.79475868e-01 -1.21853739e-01
-1.01619291e+00 -1.61007404e-01 7.09498763e-01 2.68434659e-02
-4.31025505e-01 5.12967587e-01 6.65555954e-01 -7.18809217e-02
1.21385050e+00 2.60103583e-01 -2.13878810e-01 -1.65696442e-01
5.74712515e-01 3.09437424e-01 4.64689225e-01 -8.22199941e-01
1.17292762e+00 1.11710310e+00 4.18177098e-01 -1.57133949e+00
-5.76319575e-01 -3.07316482e-01 -8.48406374e-01 -5.38204849e-01
8.96399736e-01 -7.51927793e-01 -7.46052146e-01 5.04194975e-01
-1.26644683e+00 -5.05070806e-01 -7.05616593e-01 4.11801010e-01
-5.86516917e-01 -2.49586359e-01 -4.98494118e-01 -1.20896614e+00
1.20823793e-01 -4.80411649e-01 9.90603447e-01 2.38022819e-01
-1.55356750e-01 -1.29777551e+00 1.02684605e+00 -6.75418556e-01
6.57913327e-01 7.16076136e-01 9.82851982e-01 2.68704444e-01
-8.51899862e-01 5.03566861e-01 4.58933190e-02 -3.97925645e-01
4.43634927e-01 6.45194173e-01 -7.43731141e-01 -9.46904495e-02
3.00652713e-01 1.29771516e-01 7.90676653e-01 7.87274122e-01
2.94510007e-01 -5.64627528e-01 -2.76411980e-01 6.79579198e-01
1.31595099e+00 -2.88691977e-03 -6.96126595e-02 7.47821629e-02
4.70299661e-01 8.20178092e-01 -1.93773299e-01 5.61267793e-01
-3.98465581e-02 3.56170595e-01 3.24559473e-02 -2.63530225e-01
1.05805725e-01 2.31295213e-01 4.65462476e-01 1.23709691e+00
-5.45980871e-01 1.08924776e-01 -1.48282325e+00 5.91619909e-01
-2.00039196e+00 -1.48701799e+00 -7.01767981e-01 1.90567338e+00
6.40860617e-01 -2.79689252e-01 4.55669425e-02 -6.15560293e-01
3.85734349e-01 2.83152610e-01 -6.26936376e-01 -2.80858219e-01
-6.76859260e-01 4.17071171e-02 5.14057755e-01 8.54970515e-01
-1.06309831e+00 9.39448714e-01 7.75005674e+00 1.59580889e-03
-1.07145774e+00 1.97316751e-01 2.94864625e-01 -1.02747440e-01
-2.57467836e-01 2.02445522e-01 -2.49829739e-01 2.62836181e-02
1.04203582e+00 -6.69293702e-01 5.75898826e-01 2.19327301e-01
1.04419267e+00 -2.00373143e-01 -5.16984522e-01 5.78978658e-01
-5.72202146e-01 -1.72697711e+00 -2.65103191e-01 2.85914481e-01
1.24199080e+00 4.22382683e-01 9.68262553e-03 -4.37420160e-01
1.13624167e+00 -7.80043304e-01 5.59400320e-01 9.77225065e-01
3.94627422e-01 -2.56333530e-01 1.21622883e-01 4.69760090e-01
-1.47333634e+00 -2.32748881e-01 -1.85899988e-01 -7.65772820e-01
3.36286575e-01 1.17107594e+00 -1.65132731e-01 2.77249366e-01
1.01124179e+00 1.20854926e+00 -5.35636723e-01 7.30197489e-01
5.90313599e-02 8.42497587e-01 -5.83694100e-01 -1.58778161e-01
2.41618142e-01 -6.70327425e-01 1.32785821e+00 1.07114351e+00
3.14391673e-01 6.21636093e-01 -1.44609153e-01 1.09090829e+00
4.84139591e-01 -2.43761033e-01 -1.31959677e+00 1.09162383e-01
6.45833910e-02 1.33167064e+00 -9.13187742e-01 -3.89089763e-01
6.59992397e-02 4.74773109e-01 1.30254433e-01 7.87727058e-01
-5.20382285e-01 -6.06199764e-02 1.26564419e+00 2.40835965e-01
-1.41882971e-01 -9.84886169e-01 -3.60755622e-01 -1.73788917e+00
-2.46251658e-01 -4.69753295e-01 -1.53651029e-01 -7.46185184e-01
-1.71449554e+00 6.55026793e-01 2.26388082e-01 -1.40989447e+00
-1.41581520e-01 -4.17639226e-01 -9.14889216e-01 6.18188560e-01
-9.95486021e-01 -5.46370268e-01 -5.07918000e-01 6.77800596e-01
2.46554520e-02 -2.24242821e-01 1.01614928e+00 -6.51599884e-01
-3.12324524e-01 -7.28941679e-01 8.33862901e-01 -1.60031095e-01
2.16636077e-01 -1.57841468e+00 4.36991185e-01 1.03212810e+00
1.19395129e-01 8.14676166e-01 1.17483628e+00 -7.51142442e-01
-1.29909575e+00 -9.96237814e-01 6.29075408e-01 -6.84521794e-01
1.44165027e+00 -7.59488344e-01 -6.62639499e-01 1.52880952e-01
6.06249571e-01 4.35349911e-01 5.02559304e-01 1.73082024e-01
-1.50658026e-01 5.22843562e-02 -7.23692060e-01 8.26623023e-01
1.33937120e+00 -6.19477928e-01 -4.72739577e-01 7.51476943e-01
4.95744377e-01 1.39011800e-01 -5.08890152e-01 4.67336535e-01
4.82395768e-01 -1.23239112e+00 6.27887428e-01 -9.96218443e-01
4.67942536e-01 -5.65388262e-01 -1.74571648e-02 -1.64362454e+00
-6.65178716e-01 -1.01805615e+00 -1.19199470e-01 6.06447637e-01
1.67200342e-02 -8.89739871e-01 1.56127140e-01 6.76802034e-03
2.63805032e-01 -1.45669967e-01 -1.15740275e+00 -9.81561065e-01
5.55563271e-01 7.48660117e-02 2.45354071e-01 9.86302435e-01
3.52871865e-01 4.07091439e-01 -2.07130294e-02 5.88737547e-01
1.12039065e+00 4.69893277e-01 5.83954215e-01 -1.46309745e+00
6.77142143e-02 -5.53125262e-01 -5.04611135e-01 -6.13564789e-01
6.66672662e-02 -4.90616143e-01 9.69876498e-02 -1.02907813e+00
7.92180374e-02 -4.50962722e-01 -1.26293913e-01 -2.62728948e-02
1.46363527e-01 1.71674211e-02 -4.61855493e-02 8.02440643e-01
-3.02976012e-01 8.19822192e-01 1.03224063e+00 1.83591217e-01
-5.73728681e-01 -5.32191515e-01 -7.43049309e-02 9.30024862e-01
8.30487669e-01 -4.07230884e-01 -1.33951411e-01 -6.24695271e-02
3.95027310e-01 5.60269170e-02 8.05576265e-01 -1.28308630e+00
2.60481507e-01 -5.55383027e-01 -1.06453737e-02 -3.12755734e-01
-2.03989252e-01 -5.05001903e-01 1.58234864e-01 6.95032239e-01
-3.36761475e-01 3.42232198e-01 1.67196244e-01 7.21977472e-01
-6.64041340e-02 5.47145247e-01 9.33770061e-01 -2.07427129e-01
-2.08883941e-01 4.58360873e-02 -1.23996532e+00 2.38406911e-01
1.20802164e+00 4.10800457e-01 -8.10000777e-01 -5.18683374e-01
-7.62114406e-01 1.73065558e-01 6.88532948e-01 2.08039537e-01
1.60113037e-01 -1.19090796e+00 -8.75159621e-01 3.84369880e-01
-4.73527350e-02 -3.72999966e-01 -3.14916321e-03 9.52263534e-01
-8.94817233e-01 3.57395858e-01 -3.71778548e-01 -1.09469557e+00
-8.94328654e-01 2.00415760e-01 7.41964757e-01 -1.81831196e-01
-7.88807034e-01 5.52678168e-01 5.65213740e-01 -5.03270566e-01
-6.60250306e-01 -4.49566215e-01 2.16733515e-02 -7.08030388e-02
3.89819026e-01 2.35033289e-01 -3.97365898e-01 -7.94856131e-01
-3.76472980e-01 8.15058649e-01 7.82678604e-01 -5.72962165e-01
1.33884048e+00 -2.35623255e-01 -8.31007242e-01 1.11660552e+00
8.09417486e-01 2.31913015e-01 -1.34493458e+00 4.08020206e-02
5.92508949e-02 -2.10589498e-01 -7.68500790e-02 -3.10082644e-01
-5.46082854e-01 7.40447104e-01 6.36950791e-01 1.21881890e+00
1.02067995e+00 3.55795175e-01 1.94203407e-01 5.81144989e-01
4.55862880e-01 -7.40975022e-01 7.23103434e-02 8.93717229e-01
9.07991469e-01 -8.61401021e-01 -2.62998909e-01 -1.71141520e-01
-9.80815105e-03 1.09019303e+00 2.72533387e-01 -5.95067739e-01
1.24939990e+00 7.17834055e-01 8.84653255e-02 -5.27140975e-01
-1.22224116e+00 -1.23065978e-01 1.49157876e-03 4.80835587e-01
2.26599723e-01 3.77674460e-01 -1.89323872e-01 -1.45727724e-01
-4.49488491e-01 -4.39341992e-01 7.71262586e-01 8.58216584e-01
-7.40108490e-01 -6.06040835e-01 -6.29538000e-01 3.83385658e-01
1.14388950e-01 -2.22224459e-01 -5.15095115e-01 4.10337508e-01
1.16139129e-01 9.25599158e-01 4.86154050e-01 -1.82523832e-01
-1.92650240e-02 -7.73509592e-02 1.24974214e-01 -3.93119454e-01
-2.22359926e-01 -1.19859856e-02 -3.81893367e-01 -6.63165450e-01
-8.44039619e-01 -9.94948983e-01 -1.14142251e+00 -5.97478986e-01
5.16763073e-04 5.06847203e-01 1.72355115e-01 9.28149760e-01
6.74962401e-01 5.20819247e-01 7.99150467e-01 -1.14396513e+00
1.53480634e-01 -8.59176636e-01 -7.63490975e-01 4.14671689e-01
1.01250601e+00 -6.68029547e-01 -8.57158661e-01 6.13025069e-01] | [6.547235488891602, 3.873445987701416] |
6dc39e90-8d8e-457a-a1f8-8b101fd056ec | triplet-contrastive-learning-for-brain-tumor | 2108.03611 | null | https://arxiv.org/abs/2108.03611v1 | https://arxiv.org/pdf/2108.03611v1.pdf | Triplet Contrastive Learning for Brain Tumor Classification | Brain tumor is a common and fatal form of cancer which affects both adults and children. The classification of brain tumors into different types is hence a crucial task, as it greatly influences the treatment that physicians will prescribe. In light of this, medical imaging techniques, especially those applying deep convolutional networks followed by a classification layer, have been developed to make possible computer-aided classification of brain tumor types. In this paper, we present a novel approach of directly learning deep embeddings for brain tumor types, which can be used for downstream tasks such as classification. Along with using triplet loss variants, our approach applies contrastive learning to performing unsupervised pre-training, combined with a rare-case data augmentation module to effectively ameliorate the lack of data problem in the brain tumor imaging analysis domain. We evaluate our method on an extensive brain tumor dataset which consists of 27 different tumor classes, out of which 13 are defined as rare. With a common encoder during all the experiments, we compare our approach with a baseline classification-layer based model, and the results well prove the effectiveness of our approach across all measured metrics. | ['Jiashi Feng', 'Tian Yu Liu'] | 2021-08-08 | null | null | null | null | ['unsupervised-pre-training'] | ['methodology'] | [ 3.29991728e-01 2.71388054e-01 -1.55207053e-01 -4.91494805e-01
-5.25285482e-01 -1.18140355e-01 6.63890183e-01 4.69371945e-01
-8.19850802e-01 6.48857594e-01 2.36001194e-01 -3.82925212e-01
1.29627381e-02 -8.62534285e-01 -4.71352369e-01 -9.18246090e-01
2.25794744e-02 4.70378697e-01 6.99429736e-02 2.90382281e-03
-9.20760911e-03 6.56102240e-01 -1.24632025e+00 2.15933621e-01
9.72775161e-01 1.23310375e+00 1.27189741e-01 4.19639736e-01
-2.35733315e-01 8.78420651e-01 -4.02013928e-01 -4.56115812e-01
7.25610107e-02 -8.40348825e-02 -7.60129631e-01 7.82803539e-03
1.82014361e-01 -2.26717487e-01 -3.28542799e-01 1.03333139e+00
6.78622425e-01 -1.65834352e-01 9.19461012e-01 -1.01442444e+00
-1.87465653e-01 4.29752946e-01 -4.67207760e-01 1.47658393e-01
1.44088326e-03 5.76622859e-02 8.97385001e-01 -8.24271917e-01
4.35603678e-01 6.99769557e-01 6.45372748e-01 8.32287848e-01
-1.01682186e+00 -6.12686753e-01 -1.30367383e-01 3.12648028e-01
-1.30067325e+00 -2.99849451e-01 6.71534061e-01 -7.01177895e-01
5.23523808e-01 1.61704540e-01 7.09270954e-01 1.34250724e+00
4.48912412e-01 8.20585549e-01 1.06290293e+00 -3.44257385e-01
2.48057529e-01 2.05988973e-01 2.66252607e-01 7.69531369e-01
3.24501634e-01 1.26385152e-01 -1.04834087e-01 -1.29200416e-02
2.85849154e-01 3.11898291e-01 -3.80110085e-01 -3.70165139e-01
-1.10910749e+00 8.73172820e-01 5.80287874e-01 3.81011546e-01
-3.69329393e-01 1.95421964e-01 6.09985292e-01 3.67387570e-02
6.27480745e-01 1.77831382e-01 -3.61750275e-01 1.91855192e-01
-8.53875458e-01 1.43120766e-01 6.32100582e-01 6.16738975e-01
3.96347404e-01 -2.64782041e-01 -4.81154442e-01 1.00853133e+00
1.83721140e-01 -1.32171651e-02 9.26161230e-01 -3.56513783e-02
2.03217775e-01 7.67756701e-01 -5.13103068e-01 -4.02926892e-01
-8.21597695e-01 -7.90726066e-01 -1.13679695e+00 2.52425522e-01
2.98064440e-01 -9.44350883e-02 -1.14671063e+00 1.71887755e+00
3.18994105e-01 2.59714305e-01 1.18138246e-01 6.17487490e-01
1.02919292e+00 1.93393826e-02 1.32786304e-01 1.53783590e-01
1.55538881e+00 -9.21829581e-01 -6.13001466e-01 1.73938900e-01
1.10835159e+00 -5.00357509e-01 9.01825666e-01 2.71393567e-01
-6.80421174e-01 -1.83247589e-02 -8.61195922e-01 -1.40421063e-01
-7.01386034e-01 2.49605477e-01 8.74340355e-01 7.09542513e-01
-9.99049425e-01 4.65757996e-01 -8.78967106e-01 -3.33227932e-01
9.80098665e-01 4.58299458e-01 -5.83900571e-01 -2.15193048e-01
-1.00839758e+00 1.04482591e+00 4.94265765e-01 -6.37976751e-02
-1.08598864e+00 -1.07411122e+00 -6.59843802e-01 1.41583458e-01
2.47578636e-01 -7.16485202e-01 1.03448057e+00 -8.34376276e-01
-1.29526627e+00 9.80363250e-01 7.09636956e-02 -7.28111565e-01
5.67501426e-01 1.18255943e-01 -2.97911912e-01 8.23133290e-02
-1.13338128e-01 6.71882510e-01 6.58397496e-01 -8.11899960e-01
-7.57047117e-01 -3.56586099e-01 1.71374399e-02 6.36018515e-02
-6.67688370e-01 -2.50499785e-01 -9.88440588e-02 -6.53018713e-01
-2.24470705e-01 -8.44058514e-01 -3.60413492e-01 1.61457092e-01
-5.08361697e-01 -2.79995918e-01 6.59751236e-01 -6.62891507e-01
1.01743054e+00 -2.04089046e+00 1.53057665e-01 1.90168038e-01
6.76939428e-01 8.59080702e-02 5.09971417e-02 1.25697687e-01
-4.41658884e-01 4.01658379e-02 -5.33551574e-01 -4.62787151e-01
-8.25623870e-02 2.40361720e-01 1.34480044e-01 4.93704528e-01
4.88445133e-01 7.48203337e-01 -7.01679885e-01 -3.34041178e-01
1.32298812e-01 5.07724345e-01 -6.06145263e-01 3.46964359e-01
-1.63835600e-01 5.61356783e-01 -3.74504596e-01 7.22556710e-01
5.60869217e-01 -3.19015272e-02 -3.68304849e-01 -6.68157488e-02
5.43484874e-02 1.65142313e-01 -2.98835784e-01 1.82373405e+00
-6.50890231e-01 5.02121925e-01 -4.70125899e-02 -1.31307971e+00
5.00826180e-01 3.37992549e-01 7.86872685e-01 -5.79627335e-01
4.11369920e-01 2.01073557e-01 4.34065402e-01 -5.55763543e-01
-4.04429585e-02 -1.72261924e-01 1.98621303e-01 1.93137273e-01
7.80938417e-02 -4.13303711e-02 7.05848485e-02 4.72545289e-02
1.63596570e+00 -3.44508082e-01 2.90841639e-01 -3.29832315e-01
6.95761859e-01 1.98831744e-02 5.21098435e-01 2.85224110e-01
-3.22976261e-01 6.01977646e-01 5.89839876e-01 -5.26537776e-01
-6.58173144e-01 -9.27498102e-01 -5.98798096e-01 5.28837860e-01
-2.89563864e-01 -6.88732341e-02 -7.65349805e-01 -9.66302335e-01
1.07105384e-02 5.98166049e-01 -9.09999192e-01 -4.08367962e-01
-3.44482571e-01 -1.18608391e+00 5.37074149e-01 4.74550217e-01
4.46530998e-01 -7.96203494e-01 -6.03635192e-01 7.22517297e-02
3.83802801e-02 -1.12648511e+00 -1.65521964e-01 5.61672270e-01
-7.43192375e-01 -1.31497347e+00 -8.90874922e-01 -7.56701350e-01
8.48334253e-01 -1.45194173e-01 9.18166995e-01 1.95818484e-01
-6.76370084e-01 2.68776655e-01 -5.17412007e-01 -6.50290906e-01
-5.67012310e-01 3.84445757e-01 -1.36373550e-01 1.00925878e-01
4.22763735e-01 -5.99207163e-01 -6.22393370e-01 -1.35562629e-01
-1.08587492e+00 2.86924601e-01 1.02031672e+00 1.17634857e+00
3.95568281e-01 -7.17513822e-03 4.23576415e-01 -1.09643042e+00
4.81102139e-01 -6.06176019e-01 -3.72847438e-01 7.57204294e-02
-4.25779939e-01 1.39612451e-01 7.91441739e-01 -2.76460886e-01
-6.05438352e-01 2.27700546e-01 -5.47701120e-01 -3.49860758e-01
-2.42193252e-01 6.71956778e-01 -2.61471033e-01 -2.03422070e-01
5.02806544e-01 7.43083507e-02 1.17624924e-01 -3.14081937e-01
-2.70662419e-02 7.32088268e-01 3.17207277e-01 -2.33784884e-01
6.89004123e-01 4.18171525e-01 3.35630208e-01 -8.18108082e-01
-7.25622416e-01 -4.04103011e-01 -5.65809786e-01 -5.43258199e-03
1.02379310e+00 -7.68870473e-01 -3.45980525e-01 5.26418149e-01
-9.82764721e-01 -2.94851393e-01 -2.50119537e-01 6.85162961e-01
-3.91309559e-01 1.57266200e-01 -5.21602511e-01 -3.34904253e-01
-3.31449062e-01 -1.50134552e+00 8.74285102e-01 1.49514386e-02
5.84179088e-02 -1.16356707e+00 -5.79121336e-02 1.76441342e-01
5.55084646e-01 4.04514283e-01 1.22054732e+00 -1.07472634e+00
-3.32313150e-01 -2.31086314e-01 -2.75146961e-01 3.32806408e-01
2.62867957e-01 -2.98798203e-01 -1.29821277e+00 -4.47736472e-01
-1.68423817e-01 -2.44417667e-01 1.17236388e+00 2.71520495e-01
1.66983306e+00 1.28164679e-01 -6.31109595e-01 8.17483187e-01
1.32894719e+00 1.16464764e-01 6.24190688e-01 2.50089198e-01
7.87202179e-01 5.31883299e-01 1.80086523e-01 4.10823017e-01
3.44363451e-01 5.86732626e-01 8.38590205e-01 -4.02992487e-01
-2.59806871e-01 3.11569840e-01 1.10129185e-01 6.21015310e-01
1.87614813e-01 -3.59333307e-01 -1.09658074e+00 6.51142240e-01
-1.39413726e+00 -6.85890317e-01 1.23932816e-01 2.17387104e+00
9.09663558e-01 4.89137582e-02 -1.68203771e-01 2.94896185e-01
3.90961736e-01 -9.87456217e-02 -5.21018386e-01 -3.28900456e-01
2.13007480e-01 6.23839617e-01 6.22436941e-01 3.26340832e-02
-1.25685418e+00 5.80689728e-01 6.00075245e+00 8.88262331e-01
-1.56617570e+00 1.76370233e-01 7.70129561e-01 5.07192053e-02
-6.56093881e-02 -4.00930732e-01 -4.57540870e-01 5.83249807e-01
8.39440942e-01 5.55237308e-02 2.43295971e-02 5.37878096e-01
1.43469438e-01 -7.58865997e-02 -1.45829833e+00 9.83042359e-01
-3.33841443e-02 -1.13456881e+00 -6.47052824e-02 1.90710187e-01
3.27127516e-01 1.50723800e-01 1.14179343e-01 3.29662621e-01
1.30443349e-01 -1.27722549e+00 3.69611055e-01 4.66042906e-01
8.10457170e-01 -7.90720642e-01 9.63320017e-01 3.50536317e-01
-8.30264688e-01 -6.87090084e-02 -1.86057910e-01 2.81600207e-01
-2.78077900e-01 1.02925146e+00 -1.15382397e+00 7.31791914e-01
5.73795974e-01 8.51016641e-01 -5.44848502e-01 1.16897953e+00
-1.00427948e-01 7.01059580e-01 -2.74668276e-01 1.22175030e-01
1.79449767e-02 1.02301277e-01 2.28720725e-01 1.14165318e+00
4.32342976e-01 -7.93487728e-02 1.55320600e-01 8.37699115e-01
-2.24813014e-01 1.29370943e-01 -5.45772612e-01 -9.68961418e-02
1.88489556e-01 1.45905089e+00 -5.95227540e-01 -3.06819677e-01
-5.54355621e-01 8.79381955e-01 3.94549400e-01 -9.54774544e-02
-9.34539974e-01 -3.50455523e-01 7.72336781e-01 1.80226207e-01
1.40878186e-01 1.09669425e-01 -2.73514241e-01 -1.22017586e+00
-7.63231516e-02 -5.46719551e-01 2.90767252e-01 -2.77393192e-01
-1.27035487e+00 7.14423120e-01 -1.48118734e-01 -1.26083016e+00
-1.94102108e-01 -8.37005198e-01 -8.93339396e-01 6.37116730e-01
-1.99120295e+00 -1.14568686e+00 -4.63380158e-01 5.05802810e-01
3.41611385e-01 -2.86663532e-01 9.31665480e-01 6.29082739e-01
-8.72552931e-01 8.91315043e-01 -8.01684856e-02 2.72154629e-01
6.49296761e-01 -1.32690442e+00 -1.76772594e-01 6.24075174e-01
-2.39861384e-01 1.63433388e-01 3.77859086e-01 -1.80786163e-01
-1.12427294e+00 -1.49383438e+00 5.84716856e-01 -2.15996616e-02
6.78424418e-01 -3.66074979e-01 -8.26141179e-01 4.69778746e-01
2.45014548e-01 3.51469368e-01 8.72415483e-01 -1.07881971e-01
-9.18782279e-02 -7.00212792e-02 -1.32247996e+00 4.93023455e-01
9.88015175e-01 -2.33224407e-01 -3.48694801e-01 4.95281994e-01
5.32593429e-01 -4.00508583e-01 -1.04221809e+00 6.09914005e-01
2.97161818e-01 -8.39263439e-01 8.27209532e-01 -5.31214118e-01
8.57548773e-01 5.17232157e-02 1.18149079e-01 -1.75338018e+00
-1.94414228e-01 5.70005625e-02 2.14712396e-01 9.96114671e-01
4.77326542e-01 -8.42561007e-01 9.28863823e-01 4.35208082e-01
-5.07712126e-01 -1.23787045e+00 -9.41843688e-01 -5.93145728e-01
4.65176791e-01 -3.54035467e-01 5.66063464e-01 9.29409444e-01
-1.63113266e-01 -4.93735969e-02 1.21177986e-01 1.17811389e-01
4.10327017e-01 -2.10890532e-01 4.35473382e-01 -1.23717546e+00
-1.44829661e-01 -7.31647074e-01 -9.08772647e-01 -6.04336381e-01
2.78311849e-01 -1.29078484e+00 -8.09109285e-02 -1.38689244e+00
3.16451818e-01 -6.35063291e-01 -6.29249692e-01 6.55765593e-01
-1.94235835e-02 6.62362054e-02 -2.15139911e-01 -1.32372469e-01
2.35024083e-04 5.93199432e-01 1.15486920e+00 -5.80853164e-01
2.44477764e-01 1.35004167e-02 -5.76382637e-01 7.42423892e-01
6.95800483e-01 -5.12893260e-01 -3.38022649e-01 -4.57541615e-01
-2.58475095e-01 -1.61040917e-01 3.45381320e-01 -1.23980772e+00
1.37277067e-01 8.62743184e-02 4.40818816e-01 -5.22274435e-01
2.67793208e-01 -1.06701660e+00 -2.18323037e-01 8.08924735e-01
-4.29224044e-01 -9.43314806e-02 1.78992704e-01 3.01471680e-01
-4.05276954e-01 -3.71805251e-01 9.42234755e-01 2.11685151e-01
-6.68582976e-01 7.51369059e-01 -5.29630125e-01 -7.41745457e-02
1.35993266e+00 -4.06782590e-02 -8.57723579e-02 6.39642030e-02
-8.03599000e-01 2.83553809e-01 2.92141289e-01 1.23774976e-01
6.44396067e-01 -1.25042832e+00 -8.52160990e-01 2.40358651e-01
4.37992156e-01 2.73914754e-01 -3.82670090e-02 1.25398219e+00
-5.50841331e-01 2.02068657e-01 -2.24641830e-01 -6.22714281e-01
-1.28985786e+00 4.08889115e-01 5.67659080e-01 -5.86443305e-01
-7.06402063e-01 9.42755520e-01 4.70269591e-01 -4.14171755e-01
4.33437884e-01 -7.98767090e-01 -5.03220320e-01 1.16366502e-02
5.13016939e-01 -1.28457367e-01 5.41412055e-01 -3.21936011e-01
-3.70213807e-01 2.21109971e-01 -4.20588434e-01 1.89064711e-01
1.52082348e+00 3.16574574e-01 -1.25149667e-01 7.88293183e-02
1.37982333e+00 -1.43322274e-01 -7.48293161e-01 -1.50098205e-01
7.61785284e-02 -1.54513448e-01 3.22241247e-01 -8.90439212e-01
-1.52845228e+00 1.10131049e+00 9.02942240e-01 -4.85010631e-02
1.42837369e+00 -1.39851823e-01 8.00544381e-01 2.41127044e-01
2.40721375e-01 -7.46283233e-01 -2.36639991e-01 5.17786682e-01
7.22862482e-01 -1.45046830e+00 -1.41312256e-01 -5.75730205e-01
-2.70678014e-01 1.19593906e+00 4.30720985e-01 -1.35180131e-01
7.59967625e-01 4.82464761e-01 1.86705347e-02 -1.70691863e-01
-9.55526233e-01 -3.79950821e-01 -8.40235315e-03 4.36639488e-01
5.56889534e-01 1.31677762e-01 -4.59253311e-01 7.08645523e-01
-1.58910155e-01 1.88701421e-01 5.20016134e-01 8.95819962e-01
-1.36810020e-01 -1.31249034e+00 -5.68483062e-02 9.27555978e-01
-6.29135489e-01 -1.64234787e-01 -3.20740163e-01 8.04679573e-01
3.38729829e-01 4.64666724e-01 8.26975983e-03 -4.69916463e-01
2.01711595e-01 1.75598096e-02 5.07659316e-01 -7.25290775e-01
-6.48944020e-01 -4.67625737e-01 -9.94479731e-02 -3.05715352e-01
-2.75225013e-01 -4.72752869e-01 -1.30224860e+00 -9.09564197e-02
-3.73515785e-01 -1.76187709e-01 7.47220755e-01 9.75025594e-01
1.34214535e-01 1.02517200e+00 6.72817707e-01 -5.50487757e-01
-6.43852174e-01 -9.13732469e-01 -4.50379312e-01 5.39239943e-01
3.53637040e-01 -9.86136436e-01 -2.07302317e-01 -1.91691995e-01] | [14.864892959594727, -2.4847729206085205] |
12ab58f6-182b-4a54-8550-bc0efe1afca2 | sentence-centrality-revisited-for | 1906.03508 | null | https://arxiv.org/abs/1906.03508v1 | https://arxiv.org/pdf/1906.03508v1.pdf | Sentence Centrality Revisited for Unsupervised Summarization | Single document summarization has enjoyed renewed interests in recent years thanks to the popularity of neural network models and the availability of large-scale datasets. In this paper we develop an unsupervised approach arguing that it is unrealistic to expect large-scale and high-quality training data to be available or created for different types of summaries, domains, or languages. We revisit a popular graph-based ranking algorithm and modify how node (aka sentence) centrality is computed in two ways: (a)~we employ BERT, a state-of-the-art neural representation learning model to better capture sentential meaning and (b)~we build graphs with directed edges arguing that the contribution of any two nodes to their respective centrality is influenced by their relative position in a document. Experimental results on three news summarization datasets representative of different languages and writing styles show that our approach outperforms strong baselines by a wide margin. | ['Mirella Lapata', 'Hao Zheng'] | 2019-06-08 | sentence-centrality-revisited-for-1 | https://aclanthology.org/P19-1628 | https://aclanthology.org/P19-1628.pdf | acl-2019-7 | ['unsupervised-extractive-summarization'] | ['natural-language-processing'] | [ 4.70584720e-01 4.79102045e-01 -3.83318305e-01 -5.14197826e-01
-5.68512440e-01 -5.44121265e-01 9.32853401e-01 8.27340305e-01
-3.63943577e-01 8.49513054e-01 1.23918962e+00 -2.68972397e-01
-3.18455964e-01 -9.29422736e-01 -5.43383777e-01 -2.09428936e-01
-1.63927168e-01 5.76597214e-01 1.66091308e-01 -6.07795596e-01
5.95240057e-01 3.27988267e-01 -1.10602617e+00 3.36447209e-01
8.18641186e-01 3.88875097e-01 -1.22372568e-01 7.16491878e-01
-2.68905789e-01 1.19812799e+00 -9.21038985e-01 -6.85574353e-01
-1.46937966e-01 -6.05678499e-01 -1.06702995e+00 -3.24461579e-01
6.12820208e-01 -7.85861984e-02 -7.29483306e-01 9.24569428e-01
6.45948589e-01 2.55510449e-01 9.10075426e-01 -8.17306757e-01
-7.85310507e-01 1.37254941e+00 -7.16537714e-01 5.74643433e-01
4.50115710e-01 -3.23364884e-01 1.52058613e+00 -1.26097143e-01
9.69918489e-01 1.13090587e+00 8.72094750e-01 3.02236468e-01
-1.19233537e+00 -2.79495120e-01 3.85475695e-01 -2.89003942e-02
-9.02451277e-01 -4.47353393e-01 1.02120543e+00 -7.91452453e-02
1.16269243e+00 3.38406235e-01 5.92557192e-01 1.30524719e+00
2.04485297e-01 7.29497135e-01 6.83793783e-01 -4.65135723e-01
7.31130987e-02 -3.57238352e-01 4.30165678e-01 7.01044261e-01
7.61159360e-01 -7.46643007e-01 -6.92352355e-01 -2.90493637e-01
4.41747397e-01 -3.58106077e-01 -4.12105203e-01 -1.72903180e-01
-1.37920117e+00 1.00945795e+00 5.71268260e-01 6.04816079e-01
-5.25874317e-01 2.87671804e-01 7.74139583e-01 4.91233200e-01
8.06056082e-01 9.55655336e-01 -3.51086259e-01 -2.89313495e-01
-1.07878661e+00 2.40127310e-01 1.23833096e+00 6.37333870e-01
2.06858933e-01 2.67265618e-01 -1.83297589e-01 1.00726986e+00
-1.30100340e-01 5.35511319e-03 7.22819567e-01 -9.28223968e-01
8.12372148e-01 6.80178523e-01 -2.67944902e-01 -1.56563914e+00
-5.10424972e-01 -5.90980351e-01 -1.37723231e+00 -3.97604644e-01
2.80664206e-01 -2.08839312e-01 -3.58924866e-01 1.59327924e+00
-2.84728378e-01 -1.49643630e-01 5.53504005e-02 3.87021244e-01
1.20326722e+00 7.57153630e-01 -3.19345921e-01 -2.63602704e-01
9.13676023e-01 -7.92089760e-01 -5.08347571e-01 -2.72224456e-01
5.22382319e-01 -3.34125042e-01 6.49737358e-01 3.37770492e-01
-1.26380503e+00 -1.88561037e-01 -1.09416854e+00 -1.91179574e-01
-3.65239948e-01 -1.73804373e-01 7.77884007e-01 4.44363445e-01
-1.31820548e+00 1.08517265e+00 -4.41054404e-01 -6.21086359e-01
3.88882846e-01 1.78440958e-01 -3.40338588e-01 1.81562275e-01
-1.14514518e+00 9.53111470e-01 4.06414479e-01 -3.97090949e-02
-3.77296418e-01 -2.22296432e-01 -7.25236475e-01 1.98524475e-01
3.41908932e-01 -1.10611141e+00 1.03317869e+00 -8.67896676e-01
-1.32399535e+00 8.03294718e-01 -3.40641551e-02 -6.16364360e-01
3.35927814e-01 -2.07336381e-01 -3.13416839e-01 2.59755641e-01
1.00448743e-01 2.65066892e-01 5.63726664e-01 -1.12563920e+00
-4.34450805e-01 -2.70090967e-01 7.40810484e-02 4.56356227e-01
-5.30308962e-01 1.07536159e-01 -2.87479490e-01 -8.47493589e-01
-1.20540597e-01 -5.25536180e-01 -2.08655864e-01 -6.44703269e-01
-7.93466449e-01 -4.40246761e-01 2.50598341e-01 -6.94977283e-01
1.59875071e+00 -1.57391524e+00 5.19028068e-01 2.24799082e-01
6.12558901e-01 1.83977056e-02 -2.80282021e-01 8.80931973e-01
9.83852297e-02 3.81019950e-01 -2.73816049e-01 -1.70560449e-01
-7.19735306e-03 1.70170113e-01 -3.71687472e-01 3.88487458e-01
-2.98776273e-02 8.71774554e-01 -1.20949078e+00 -4.56057578e-01
-3.85452837e-01 2.87837923e-01 -3.21839631e-01 -1.35284379e-01
-2.47396808e-02 -4.72004078e-02 -4.87082452e-01 3.04408044e-01
-4.43305038e-02 -4.47553962e-01 5.58916509e-01 -4.11003642e-02
1.51249781e-01 7.47074366e-01 -7.93157518e-01 1.59922338e+00
-3.00703868e-02 9.88846898e-01 -7.31866285e-02 -1.23371208e+00
8.27617407e-01 1.25288367e-02 2.43990228e-01 -4.21808630e-01
1.72346473e-01 7.05998624e-04 1.28400996e-01 -1.11327827e-01
1.12190366e+00 2.35444888e-01 -2.97999620e-01 7.08346486e-01
-1.04525894e-01 -1.79737449e-01 6.17249608e-01 9.25091386e-01
1.49543881e+00 -2.14691743e-01 5.47503591e-01 -3.62781823e-01
2.46488914e-01 -7.10382909e-02 3.10551316e-01 1.06448150e+00
6.91168681e-02 5.66470444e-01 1.11791658e+00 -1.80907756e-01
-1.19932783e+00 -6.33571088e-01 2.91543186e-01 1.36807752e+00
-2.87484139e-01 -5.86200237e-01 -5.40292919e-01 -5.01656711e-01
-4.93196845e-02 9.59047914e-01 -7.79163778e-01 -1.89050153e-01
-8.44921887e-01 -6.29986405e-01 7.86074102e-01 5.34847975e-01
3.00338209e-01 -1.18290210e+00 -4.42896098e-01 1.67890176e-01
-3.28606904e-01 -7.50823498e-01 -2.76162505e-01 1.85314491e-01
-1.06369102e+00 -9.05232608e-01 -7.75861263e-01 -7.83883393e-01
5.95837057e-01 2.46333376e-01 1.72255552e+00 8.32728893e-02
2.34043226e-01 3.46684128e-01 -4.95524436e-01 -2.60983884e-01
-7.32263505e-01 6.96589947e-01 1.26781091e-02 -3.57361794e-01
-1.60099287e-02 -7.13067949e-01 -3.98983061e-01 -3.48887056e-01
-8.31514418e-01 -1.17019294e-02 6.99630976e-01 5.88445365e-01
-2.62964945e-02 -4.94588688e-02 8.82386744e-01 -1.23553991e+00
1.63843346e+00 -6.30684674e-01 9.71872061e-02 3.47762287e-01
-7.69405723e-01 1.36256278e-01 6.53434098e-01 -7.20488876e-02
-8.22178662e-01 -8.01046610e-01 2.15402052e-01 2.83446819e-01
9.62422565e-02 1.16755998e+00 3.05269450e-01 4.71950769e-01
8.77709627e-01 2.44231433e-01 -1.05091006e-01 -3.02662134e-01
7.39057541e-01 6.41612589e-01 6.36641324e-01 -5.82857251e-01
6.89663410e-01 2.14623883e-01 -2.91750822e-02 -1.07390404e+00
-8.28596771e-01 -3.76167595e-01 -5.19959092e-01 -4.77120839e-02
4.70013916e-01 -7.43369520e-01 -3.98907483e-01 2.71289736e-01
-1.24401510e+00 -1.26483530e-01 -3.02115440e-01 1.53736174e-01
-1.85412765e-01 6.75723791e-01 -8.16405177e-01 -4.41368550e-01
-9.16902184e-01 -4.67229307e-01 6.87696040e-01 1.99037865e-01
-5.80866873e-01 -1.16749239e+00 3.38813692e-01 2.90700912e-01
5.88578403e-01 4.05911744e-01 1.14293587e+00 -1.11399877e+00
-7.82926679e-02 -3.72036427e-01 -4.48466778e-01 1.77796364e-01
9.05309841e-02 1.74143240e-01 -4.68493283e-01 -2.34413996e-01
-5.14829397e-01 -3.75179976e-01 1.11718738e+00 5.47352910e-01
8.98802400e-01 -7.40059674e-01 -2.10257083e-01 7.33144805e-02
1.05184996e+00 -4.27700639e-01 4.56125408e-01 3.08523357e-01
7.08739221e-01 6.16061687e-01 -1.53804719e-01 3.89824361e-01
5.74274361e-01 2.02433750e-01 1.41167819e-01 1.41738445e-01
-2.00285893e-02 -4.25769657e-01 3.76320064e-01 1.44606113e+00
-3.56956601e-01 -7.06729531e-01 -8.88758242e-01 7.41918266e-01
-1.96440101e+00 -1.22387171e+00 -1.64886937e-01 2.01282024e+00
1.01702738e+00 5.04077911e-01 2.88283229e-01 -8.48224759e-02
6.65571272e-01 8.07047904e-01 -3.50865126e-01 -6.49462283e-01
-3.84961069e-01 7.20870420e-02 4.05531019e-01 2.55968571e-01
-7.39502609e-01 8.24530780e-01 6.83845949e+00 6.07361555e-01
-8.01994145e-01 -3.70762974e-01 6.69030845e-01 -7.27320090e-02
-5.87473929e-01 -1.44935936e-01 -3.95895809e-01 1.00291915e-01
1.02587485e+00 -6.40585065e-01 2.64453381e-01 5.99000573e-01
4.92165098e-03 9.55642760e-03 -1.01069283e+00 4.33012843e-01
5.54713666e-01 -1.68201244e+00 3.73462319e-01 -1.43441156e-01
7.95593441e-01 4.73952055e-01 -3.38248938e-01 3.37460130e-01
8.68704975e-01 -9.02753532e-01 3.45635474e-01 6.03988707e-01
5.15089869e-01 -7.13433206e-01 7.49271512e-01 3.74024391e-01
-9.25780416e-01 1.24950781e-01 -4.70425487e-01 -2.46063069e-01
5.74592836e-02 5.84685922e-01 -7.64251351e-01 8.23276341e-01
4.81327534e-01 9.46475208e-01 -7.95684397e-01 8.07063937e-01
-4.01211798e-01 9.92681623e-01 8.76534171e-03 -5.93585253e-01
3.30937892e-01 -2.59215146e-01 7.81721771e-01 1.48246562e+00
-8.38526264e-02 -8.85328129e-02 4.06256020e-02 4.62243766e-01
-7.53072619e-01 2.83026218e-01 -7.88074732e-01 -4.65691388e-01
4.18800235e-01 1.25686419e+00 -7.83893466e-01 -5.88349462e-01
-2.40212068e-01 8.56135368e-01 6.06529355e-01 3.76734644e-01
-3.97491693e-01 -7.24656045e-01 9.67212468e-02 1.16504915e-01
1.99690163e-01 -1.52254894e-01 -2.09383592e-01 -1.25519788e+00
-2.49179959e-01 -8.29801083e-01 4.97341752e-01 -7.25502789e-01
-1.55019009e+00 6.76507890e-01 7.24312961e-02 -6.25953078e-01
-4.41353381e-01 -1.89831764e-01 -9.15945292e-01 5.31488776e-01
-1.42137718e+00 -8.77025187e-01 -7.20244944e-02 2.21025959e-01
4.49055612e-01 -4.15968210e-01 8.33996058e-01 -1.98441252e-01
-5.28068900e-01 2.73193806e-01 3.32998067e-01 4.28680658e-01
5.73193610e-01 -1.47350442e+00 6.83253944e-01 7.29206145e-01
4.49631155e-01 8.46851110e-01 9.37752485e-01 -7.56449997e-01
-1.06348896e+00 -7.77749717e-01 1.30811679e+00 -4.13873494e-01
9.45973992e-01 -5.58425561e-02 -8.47923100e-01 7.03280091e-01
7.60683298e-01 -5.78077316e-01 7.20537603e-01 6.27766132e-01
-4.86899912e-01 -5.36510861e-03 -7.14140832e-01 6.20989442e-01
1.06063712e+00 -3.67692232e-01 -1.16970015e+00 4.80779946e-01
6.76667809e-01 1.67949703e-02 -7.70947576e-01 2.14533672e-01
4.20399159e-01 -9.14530337e-01 7.73599684e-01 -9.36223269e-01
1.01432085e+00 1.47644058e-01 1.58238746e-02 -1.67006671e+00
-6.06788099e-01 -7.78654277e-01 -2.91293472e-01 1.21802306e+00
5.62322259e-01 -6.71986401e-01 6.19272590e-01 2.44768158e-01
-1.67759761e-01 -7.76909232e-01 -6.34357214e-01 -3.68485004e-01
2.71384746e-01 2.55944226e-02 3.46045732e-01 1.20688426e+00
4.33039248e-01 1.06158137e+00 -3.00693542e-01 -3.49640727e-01
3.11448574e-01 2.57899225e-01 7.86365569e-01 -1.69062495e+00
-2.15591714e-01 -1.02166867e+00 -2.42604792e-01 -1.07483304e+00
2.88651228e-01 -1.10599947e+00 -2.17563406e-01 -2.33825064e+00
4.33236688e-01 1.51682273e-02 -3.02665412e-01 3.07908148e-01
-1.48668677e-01 -2.11166181e-02 -1.54069856e-01 2.59362876e-01
-8.91261756e-01 5.01101375e-01 8.10256541e-01 -3.45136762e-01
-1.08967841e-01 -1.57908842e-01 -1.22219658e+00 7.97990203e-01
7.77738810e-01 -3.40970725e-01 -4.31612909e-01 -4.64572549e-01
9.14399505e-01 7.85579160e-02 -6.86486289e-02 -6.86777532e-01
3.14427942e-01 1.67905018e-01 2.88697064e-01 -4.29597050e-01
-5.65760843e-02 -1.99949831e-01 -3.44663769e-01 3.33856702e-01
-9.88059878e-01 2.83516705e-01 -1.17959410e-01 9.33753073e-01
-1.52058363e-01 -2.90654600e-01 2.51708657e-01 -3.28118533e-01
-2.84086883e-01 -1.25278411e-02 -3.90682995e-01 4.29729074e-01
3.08110207e-01 -1.00630805e-01 -6.63296938e-01 -9.18843865e-01
-1.18082985e-01 6.82255328e-02 4.07744795e-01 5.31252623e-01
5.04946411e-01 -7.93100059e-01 -1.25339425e+00 -5.69871426e-01
7.99122453e-02 -1.20431364e-01 -1.53499112e-01 6.47128820e-01
-5.27125478e-01 3.22602957e-01 -7.03145340e-02 -1.43468287e-02
-1.27210939e+00 7.38787502e-02 -7.98262730e-02 -6.96826756e-01
-8.40728283e-01 8.10292184e-01 -4.89657849e-01 -3.62562835e-01
1.28855109e-01 -2.28270099e-01 -6.01676762e-01 4.02718008e-01
2.87394106e-01 5.47397256e-01 -8.22452605e-02 -6.79741800e-01
-1.39939860e-01 1.43809691e-01 -3.95954520e-01 7.56309628e-02
1.78265452e+00 1.42853692e-01 -4.60734099e-01 5.98596990e-01
9.37790096e-01 2.66938776e-01 -5.32823443e-01 -3.45922232e-01
1.46864191e-01 -7.64939040e-02 7.18565285e-02 -7.98722148e-01
-7.85736561e-01 4.06250179e-01 -4.21014518e-01 7.14511037e-01
7.04034507e-01 5.49895205e-02 6.71899259e-01 8.11971724e-01
8.10693279e-02 -1.30868304e+00 2.12546021e-01 6.12048686e-01
1.06852043e+00 -8.20721745e-01 7.62704074e-01 9.06844903e-03
-8.28896105e-01 1.12959611e+00 7.08958432e-02 -2.66764075e-01
1.28941432e-01 -2.11321279e-01 -3.35754961e-01 -4.66165572e-01
-8.94780695e-01 2.23744065e-01 5.00436962e-01 3.33561718e-01
8.03902805e-01 -3.22385207e-02 -5.18808722e-01 3.98341537e-01
-4.42399889e-01 -3.73487085e-01 1.03291690e+00 8.62739384e-01
-6.16842151e-01 -8.75133097e-01 1.46690488e-01 1.10815144e+00
-4.35041636e-01 -2.50485599e-01 -9.58132029e-01 7.15886950e-01
-6.09563947e-01 1.04244220e+00 -1.49090961e-01 -2.13224292e-01
2.06120223e-01 -3.27602811e-02 3.99396002e-01 -7.21090257e-01
-7.70995617e-01 -2.30766624e-01 7.03243911e-01 -2.26110145e-01
-5.63472748e-01 -7.03465879e-01 -1.13929594e+00 -5.63488960e-01
-2.15037495e-01 2.11030036e-01 5.33104420e-01 6.49665892e-01
5.87051332e-01 7.87882268e-01 3.19419235e-01 -6.76441014e-01
-7.57800519e-01 -1.21704280e+00 -6.18572116e-01 4.06283736e-01
6.98316097e-02 -9.64068845e-02 -3.63874853e-01 -2.58064508e-01] | [12.51528263092041, 9.52285385131836] |
c04e8898-ed8d-4db2-8ae1-88ff4f76000e | disentangling-homophily-community-structure | 2101.02510 | null | https://arxiv.org/abs/2101.02510v3 | https://arxiv.org/pdf/2101.02510v3.pdf | Disentangling homophily, community structure and triadic closure in networks | Network homophily, the tendency of similar nodes to be connected, and transitivity, the tendency of two nodes being connected if they share a common neighbor, are conflated properties in network analysis, since one mechanism can drive the other. Here we present a generative model and corresponding inference procedure that are capable of distinguishing between both mechanisms. Our approach is based on a variation of the stochastic block model (SBM) with the addition of triadic closure edges, and its inference can identify the most plausible mechanism responsible for the existence of every edge in the network, in addition to the underlying community structure itself. We show how the method can evade the detection of spurious communities caused solely by the formation of triangles in the network, and how it can improve the performance of edge prediction when compared to the pure version of the SBM without triadic closure. | ['Tiago P. Peixoto'] | 2021-01-07 | null | null | null | null | ['stochastic-block-model', 'graph-reconstruction'] | ['graphs', 'graphs'] | [ 3.45758684e-02 2.63216883e-01 3.46175954e-02 7.27527961e-02
5.34264028e-01 -6.79642200e-01 9.02308822e-01 3.03017139e-01
1.79190218e-01 7.48643219e-01 5.98972552e-02 -7.75505960e-01
-5.83231032e-01 -1.40502417e+00 -7.56168604e-01 -8.85107696e-01
-5.00728846e-01 8.15584123e-01 5.55177033e-01 -1.83040991e-01
7.78424814e-02 3.21186572e-01 -1.20791304e+00 1.17019899e-01
6.79157197e-01 4.36681092e-01 -3.79701704e-02 5.19063354e-01
-1.08770624e-01 8.62311065e-01 -2.17933252e-01 -5.01583517e-01
7.09883869e-02 -6.61280453e-01 -8.29341471e-01 -7.53416494e-02
-7.25198314e-02 2.03058958e-01 -2.82105505e-01 8.10635030e-01
4.47993651e-02 -3.30302447e-01 8.34058523e-01 -1.58993649e+00
-1.01834029e-01 9.65325296e-01 -7.42694795e-01 7.09043071e-02
4.74977881e-01 -2.40510896e-01 1.27594578e+00 -5.92707098e-01
1.01587057e+00 1.34631467e+00 9.62261796e-01 -9.87462420e-03
-1.84454238e+00 -3.58297050e-01 -3.91150080e-02 7.68662244e-02
-1.64302981e+00 -2.24281266e-01 7.95711935e-01 -6.30213559e-01
5.56611061e-01 3.62776190e-01 9.48892474e-01 9.26548541e-01
1.83905914e-01 3.89735103e-01 1.11696315e+00 -3.53147894e-01
3.28080654e-01 -4.60528471e-02 9.60078463e-02 5.55308223e-01
8.45848083e-01 3.99627201e-02 -3.93904567e-01 -6.71095490e-01
7.12623417e-01 -6.40199706e-02 -1.53650194e-01 -8.03648412e-01
-1.00769293e+00 9.64318752e-01 4.94723618e-01 4.78492379e-01
-2.38865167e-01 1.29105330e-01 1.21509962e-01 3.62193674e-01
3.84422660e-01 1.99899316e-01 -6.15671948e-02 4.43351604e-02
-1.04115379e+00 -2.65616849e-02 1.29775918e+00 6.49459481e-01
8.21154237e-01 -4.69809681e-01 3.12591344e-01 1.45716771e-01
4.52489316e-01 1.39637413e-02 -1.26003951e-01 -6.15308642e-01
-4.69544064e-03 9.45105076e-01 -8.16368088e-02 -1.42990649e+00
-2.84788430e-01 -4.67877239e-01 -1.06887138e+00 3.16930525e-02
6.17979825e-01 -1.09268121e-01 -4.20967013e-01 1.78974032e+00
3.54996711e-01 3.84739369e-01 -3.53301138e-01 2.94235140e-01
5.86159229e-01 4.28008050e-01 -3.61168683e-01 -5.01535475e-01
8.28291237e-01 -4.20886546e-01 -3.73303354e-01 2.84707278e-01
5.88446617e-01 -4.75212276e-01 4.69149739e-01 1.14141822e-01
-1.04671264e+00 -1.48938805e-01 -7.42916703e-01 5.19280910e-01
-2.91472912e-01 -6.94901109e-01 7.15833902e-01 6.81827486e-01
-1.27913594e+00 8.41700554e-01 -8.30536962e-01 -6.02796495e-01
5.52245639e-02 3.41655433e-01 -4.11529154e-01 -1.50776142e-02
-9.69262600e-01 6.04774475e-01 9.41024795e-02 1.26703933e-01
-5.60294032e-01 -4.22248304e-01 -5.20123601e-01 3.25770676e-01
3.97126079e-01 -8.06911051e-01 2.86099404e-01 -1.06129754e+00
-6.86799586e-01 8.06701839e-01 -3.84644002e-01 -4.46078956e-01
8.76612902e-01 4.98716444e-01 -1.66582912e-01 1.15400471e-01
6.48824200e-02 2.23821595e-01 6.45209312e-01 -1.33114803e+00
-1.49963126e-01 -3.05764347e-01 -2.26560131e-01 -2.18742102e-01
-4.93098199e-02 -2.53068924e-01 -1.48712561e-01 -2.59182513e-01
4.29264069e-01 -1.06081831e+00 -1.86822623e-01 -2.22463965e-01
-7.63044238e-01 -2.53327906e-01 5.68217278e-01 -1.20352000e-01
1.34716189e+00 -1.91174531e+00 2.59863466e-01 1.10270727e+00
6.50057435e-01 -2.16023028e-01 2.08712239e-02 1.08869672e+00
-2.82097548e-01 5.74273646e-01 -1.91847339e-01 5.83471432e-02
-2.58337229e-01 4.62655634e-01 -9.30535719e-02 7.06554055e-01
-6.02076314e-02 5.74014187e-01 -9.62562025e-01 -3.92719358e-01
-7.07524866e-02 2.38776520e-01 -4.60857570e-01 -2.79618293e-01
2.04342157e-01 2.24336714e-01 -7.04583824e-02 1.95266262e-01
6.92775726e-01 -6.17949724e-01 9.88164246e-01 3.02841485e-01
-1.24745756e-01 4.42068070e-01 -1.42043972e+00 6.17993176e-01
-3.09385471e-02 6.86145961e-01 1.98676467e-01 -9.06921387e-01
9.29280460e-01 3.69749457e-01 4.23347294e-01 7.53156915e-02
2.93965284e-02 1.93204567e-01 5.39258420e-01 -1.19778015e-01
-8.96355510e-02 1.05600432e-01 3.97363931e-01 8.42563689e-01
-2.86561638e-01 3.21682960e-01 5.24294257e-01 7.70940006e-01
1.45845902e+00 -3.37543786e-01 2.86184758e-01 -6.42000616e-01
2.81689137e-01 -1.56412482e-01 5.90058625e-01 1.05430937e+00
1.38221860e-01 1.75212592e-01 1.18949449e+00 -3.14121455e-01
-1.16736555e+00 -1.15078580e+00 -1.48425505e-01 4.75221246e-01
2.74130762e-01 -5.47560453e-01 -3.18213701e-01 -4.03697252e-01
3.89344335e-01 3.32668424e-01 -9.62752223e-01 -1.19501822e-01
-2.01900214e-01 -9.52824533e-01 3.07761073e-01 4.52023484e-02
1.03763200e-01 -6.38754129e-01 -1.05402127e-01 1.39124960e-01
-3.88936341e-01 -7.97580004e-01 7.49302953e-02 1.65349454e-01
-1.12428677e+00 -1.57802773e+00 -1.45439729e-01 -5.54088771e-01
8.21513057e-01 3.28651905e-01 1.36172223e+00 8.31177473e-01
-1.82832062e-01 -2.27105487e-02 -2.49572217e-01 2.51861721e-01
-8.70249569e-01 8.06952640e-02 7.24346712e-02 2.52707869e-01
2.40918264e-01 -1.23854911e+00 -4.78648871e-01 5.12344122e-01
-6.63882256e-01 1.37574688e-01 1.64186329e-01 6.23591959e-01
-1.11087728e-02 3.61061186e-01 3.07657987e-01 -1.11880016e+00
5.47573864e-01 -9.71302092e-01 -4.47111100e-01 1.66334391e-01
-9.35753047e-01 -4.72050309e-02 2.27750927e-01 -1.59734741e-01
-5.35313189e-01 -7.52570108e-02 2.44948670e-01 -8.42593089e-02
1.22950010e-01 6.66411757e-01 -5.56117669e-02 -1.85974926e-01
4.27985400e-01 2.12038264e-01 2.13767603e-01 -4.25315499e-01
2.09478244e-01 3.12368691e-01 1.10991346e-02 -3.28713268e-01
8.30913782e-01 8.76742601e-01 4.92549330e-01 -8.94679427e-01
1.63608491e-02 -5.04076242e-01 -8.22407722e-01 -3.28718930e-01
3.02408874e-01 -7.33268142e-01 -8.03725779e-01 2.44841084e-01
-1.04121614e+00 -1.63145199e-01 1.70311198e-01 2.84489334e-01
-7.60656595e-02 5.40644765e-01 -7.31698692e-01 -9.42361712e-01
1.44933879e-01 -6.09757721e-01 2.76143491e-01 -2.31542572e-01
-4.57906157e-01 -1.27638090e+00 4.98723507e-01 -4.19981331e-02
7.56186470e-02 4.15612608e-01 1.15337408e+00 -7.70787597e-01
-6.13681197e-01 -2.78992414e-01 -1.68690860e-01 -2.34222993e-01
1.23368554e-01 8.45296443e-01 -4.96033877e-01 -3.85623097e-01
-3.26312095e-01 5.09169281e-01 8.95468056e-01 5.72133005e-01
3.01449299e-01 -4.28201795e-01 -9.01940882e-01 2.62235463e-01
1.49911642e+00 -2.48709008e-01 8.52060080e-01 -2.68701068e-03
4.79726404e-01 8.20497036e-01 -2.85489440e-01 4.22187179e-01
2.03374237e-01 5.69710672e-01 5.85079789e-01 -1.49823040e-01
9.90220085e-02 -5.06760240e-01 4.04698364e-02 8.31319928e-01
-2.81681031e-01 -4.13574487e-01 -9.83174443e-01 6.98189676e-01
-1.84554982e+00 -1.18809211e+00 -1.21453702e+00 2.30478430e+00
4.93135005e-01 2.69494444e-01 4.74929720e-01 3.49783510e-01
1.19505966e+00 -1.35355577e-01 -8.94812644e-02 -2.46865734e-01
-2.43885174e-01 -1.55904755e-01 5.06859243e-01 7.23833501e-01
-4.78296816e-01 5.14737189e-01 7.87251186e+00 5.55966198e-01
-5.53538501e-01 -1.25236183e-01 5.42194605e-01 1.96080789e-01
-5.44847190e-01 5.44187844e-01 -5.33267856e-01 5.56945801e-01
7.49611318e-01 -1.22000352e-01 3.34534138e-01 5.77591360e-01
4.10275578e-01 -2.90546805e-01 -9.80809689e-01 3.43151718e-01
-3.45971227e-01 -1.34533679e+00 6.28290251e-02 6.76097631e-01
9.48684096e-01 -1.79760847e-02 -4.41094428e-01 -4.81109172e-01
8.31163287e-01 -8.47553492e-01 3.83486807e-01 3.75277191e-01
3.24990422e-01 -5.26714563e-01 6.64974570e-01 4.44710851e-01
-1.22237241e+00 1.16540045e-01 -1.37255952e-01 -5.09982824e-01
5.10757901e-02 1.11453068e+00 -9.61210132e-01 5.26961505e-01
5.30891061e-01 7.30134130e-01 -5.35658300e-01 1.16294050e+00
-3.57661784e-01 8.14824283e-01 -6.31676018e-01 -6.02847375e-02
-3.49914283e-02 -6.50445163e-01 9.89285827e-01 7.50096142e-01
9.42693055e-02 -3.69249552e-01 -7.69606158e-02 1.21882188e+00
2.30583906e-01 -9.80172083e-02 -7.53187299e-01 7.73010924e-02
7.44800985e-01 1.26748288e+00 -1.33113790e+00 -3.18951458e-01
-2.78901756e-01 6.25299513e-01 1.94227889e-01 2.26029202e-01
-4.44437504e-01 -1.01114824e-01 4.87545729e-01 6.89989746e-01
4.69932169e-01 -1.21667638e-01 -2.62584865e-01 -1.04706109e+00
-1.16202690e-01 -5.72553754e-01 3.13635886e-01 -5.42088091e-01
-1.36807895e+00 3.68806630e-01 -3.46387506e-01 -7.65839159e-01
-1.96306139e-01 -2.74953961e-01 -1.15720141e+00 7.68052757e-01
-1.00154757e+00 -9.37628388e-01 -1.17936477e-01 7.56702006e-01
-4.59789991e-01 3.47165257e-01 5.73111355e-01 5.35001755e-02
-4.80449408e-01 6.86739832e-02 4.78169918e-01 8.38933736e-02
2.91547865e-01 -1.36688709e+00 3.32935989e-01 1.02655900e+00
2.19523787e-01 8.64702821e-01 9.61884320e-01 -1.04002750e+00
-1.01067853e+00 -7.12772965e-01 1.30554879e+00 -4.79160011e-01
9.60030913e-01 -7.23447382e-01 -8.27518523e-01 5.99251091e-01
-4.70746011e-02 -2.19822019e-01 6.58733130e-01 6.98677659e-01
-3.55100185e-01 1.83433592e-01 -7.25337029e-01 6.69227600e-01
1.25597644e+00 -2.43191183e-01 -3.31828415e-01 2.99870580e-01
1.94179818e-01 5.99047482e-01 -7.47415960e-01 2.32893482e-01
6.05499327e-01 -1.50896573e+00 7.91598380e-01 -3.84869546e-01
4.83488470e-01 -3.19294840e-01 2.34403715e-01 -9.85585511e-01
-6.11847162e-01 -8.84443879e-01 -2.31005158e-03 1.10315740e+00
4.38665032e-01 -7.87356198e-01 8.60675812e-01 4.02205139e-02
5.00476480e-01 -4.60693270e-01 -1.03101921e+00 -9.22805846e-01
-3.30622382e-02 1.25841856e-01 4.60168093e-01 9.94758070e-01
3.24125916e-01 5.93602240e-01 -1.67074159e-01 -5.70004657e-02
7.05427885e-01 3.75959694e-01 8.85397971e-01 -2.04623389e+00
-4.06191230e-01 -5.90772748e-01 -4.76202011e-01 -7.14564145e-01
-1.49785593e-01 -8.52100611e-01 -2.18813211e-01 -1.39680874e+00
4.60887998e-01 -6.37190461e-01 3.12535554e-01 1.83110952e-01
7.88350031e-02 2.59780884e-01 -1.86949417e-01 5.23342848e-01
-2.63817251e-01 1.52954429e-01 7.72983491e-01 1.48031294e-01
-2.87654489e-01 2.05230936e-01 -5.40486157e-01 7.80495405e-01
4.04034555e-01 -6.04257166e-01 -1.87867522e-01 2.86964953e-01
1.16965187e+00 8.74232948e-02 5.73351085e-01 -6.12713814e-01
2.84423918e-01 2.12629944e-01 1.30190805e-01 -3.88595581e-01
-7.08905235e-02 -7.82629788e-01 9.33628976e-01 9.85162258e-01
-8.01958144e-02 1.14880718e-01 -3.64506811e-01 9.42644715e-01
-6.94659725e-02 -2.15404958e-01 5.29249251e-01 -2.02233911e-01
-6.30000560e-03 5.00037102e-03 -7.06173480e-01 -2.39892542e-01
9.42295849e-01 -4.21998948e-01 -2.80079037e-01 -6.47026002e-01
-8.74599695e-01 5.35799414e-02 7.92538822e-01 5.78826629e-02
1.33014143e-01 -1.07957387e+00 -8.55126500e-01 2.71959584e-02
-2.64940977e-01 -4.96437043e-01 -2.06038982e-01 1.31719267e+00
-5.65331280e-01 8.21431801e-02 -8.18089303e-03 -5.57162523e-01
-1.65351176e+00 6.76148772e-01 3.97120059e-01 -4.69525278e-01
-4.27924216e-01 5.08857012e-01 2.11513147e-01 -1.25162542e-01
-1.57757178e-01 2.01376647e-01 -9.05010290e-03 1.24674596e-01
8.18676949e-02 6.44841731e-01 -1.38350844e-01 -6.20912850e-01
-4.87135053e-01 2.04570189e-01 1.12601213e-01 1.08587153e-01
1.27617753e+00 -4.22534317e-01 -9.48083758e-01 7.09598303e-01
8.10858250e-01 3.70684534e-01 -6.56193137e-01 -7.16993958e-03
2.29941085e-02 -3.57745320e-01 -2.44978264e-01 -3.53215158e-01
-8.71875167e-01 5.82399666e-01 -1.59155816e-01 1.17987013e+00
5.73018610e-01 9.04692486e-02 1.78462088e-01 -2.19341870e-02
2.95556754e-01 -5.80526292e-01 -2.95227081e-01 2.44612500e-01
3.56998920e-01 -8.25240195e-01 9.81539711e-02 -1.00028312e+00
2.39918008e-02 1.05949342e+00 -1.24468515e-02 -4.27841812e-01
9.27463889e-01 2.11396087e-02 -5.39600790e-01 -5.16825080e-01
-1.04730213e+00 -1.67064205e-01 -9.01212767e-02 4.82724845e-01
3.82613868e-01 1.62056759e-01 -6.57956064e-01 -5.67717897e-03
-6.19917698e-02 -4.23133761e-01 7.22481489e-01 5.30772507e-01
-3.21077257e-01 -1.09624994e+00 -3.05751950e-01 5.44960976e-01
-2.19868898e-01 -2.79635578e-01 -1.06151199e+00 9.34525430e-01
1.31280631e-01 1.13677073e+00 4.03762907e-01 -3.58076215e-01
-3.32741410e-01 1.81754585e-02 3.62219930e-01 -3.37578565e-01
-5.28055310e-01 1.90760016e-01 2.36141339e-01 -2.56763279e-01
-5.48143685e-01 -7.05115736e-01 -5.78079164e-01 -1.14308119e+00
-8.63942504e-01 5.02822459e-01 1.54258415e-01 8.34174573e-01
4.32019621e-01 1.16381444e-01 8.13907564e-01 -3.04883718e-01
2.26988848e-02 -9.48706985e-01 -1.01180196e+00 3.96521628e-01
1.80838890e-02 -6.18710697e-01 -8.53997767e-01 -1.74942836e-01] | [6.971022605895996, 5.267307758331299] |
5e121ac8-ad8a-4fd0-97b8-e4d82320b5d4 | uncertainty-guided-mutual-consistency | 2112.02508 | null | https://arxiv.org/abs/2112.02508v2 | https://arxiv.org/pdf/2112.02508v2.pdf | Uncertainty-Guided Mutual Consistency Learning for Semi-Supervised Medical Image Segmentation | Medical image segmentation is a fundamental and critical step in many clinical approaches. Semi-supervised learning has been widely applied to medical image segmentation tasks since it alleviates the heavy burden of acquiring expert-examined annotations and takes the advantage of unlabeled data which is much easier to acquire. Although consistency learning has been proven to be an effective approach by enforcing an invariance of predictions under different distributions, existing approaches cannot make full use of region-level shape constraint and boundary-level distance information from unlabeled data. In this paper, we propose a novel uncertainty-guided mutual consistency learning framework to effectively exploit unlabeled data by integrating intra-task consistency learning from up-to-date predictions for self-ensembling and cross-task consistency learning from task-level regularization to exploit geometric shape information. The framework is guided by the estimated segmentation uncertainty of models to select out relatively certain predictions for consistency learning, so as to effectively exploit more reliable information from unlabeled data. Experiments on two publicly available benchmark datasets showed that: 1) Our proposed method can achieve significant performance improvement by leveraging unlabeled data, with up to 4.13% and 9.82% in Dice coefficient compared to supervised baseline on left atrium segmentation and brain tumor segmentation, respectively. 2) Compared with other semi-supervised segmentation methods, our proposed method achieve better segmentation performance under the same backbone network and task settings on both datasets, demonstrating the effectiveness and robustness of our method and potential transferability for other medical image segmentation tasks. | ['Dongyang Li', 'Jicong Zhang', 'Rushi Jiao', 'Qingcheng Liao', 'Yichi Zhang'] | 2021-12-05 | null | null | null | null | ['semi-supervised-medical-image-segmentation'] | ['computer-vision'] | [ 2.52581209e-01 2.85091877e-01 -5.49294591e-01 -6.24383569e-01
-1.17883575e+00 -4.24284846e-01 1.67353600e-01 2.91074514e-01
-5.65291405e-01 7.95753717e-01 -3.69435549e-02 -2.79290318e-01
-1.83561072e-01 -2.70283073e-01 -4.27677542e-01 -8.86572480e-01
1.37473390e-01 6.14505053e-01 3.25190902e-01 3.18650454e-01
4.74568754e-02 2.41249472e-01 -7.79888868e-01 -6.84328973e-02
1.46662271e+00 1.07048237e+00 4.46891457e-01 1.34551272e-01
-1.28922567e-01 5.94166875e-01 -1.39310002e-01 -1.69416219e-01
2.43250653e-01 -3.05667490e-01 -9.46774423e-01 3.81850123e-01
1.83553308e-01 -1.66342899e-01 1.00072622e-02 1.26268601e+00
4.68315661e-01 -6.75625633e-03 7.58379161e-01 -8.62322807e-01
-4.05000895e-01 5.60976624e-01 -8.35513651e-01 2.19606012e-01
-3.18537712e-01 1.01525441e-01 7.80400872e-01 -5.15052438e-01
5.34007728e-01 6.42505169e-01 6.93433642e-01 4.78747040e-01
-1.26097429e+00 -7.06434786e-01 7.00954795e-02 -1.91293031e-01
-1.31953704e+00 2.53963470e-03 7.17449844e-01 -4.90077555e-01
4.33371514e-01 1.22982100e-01 2.96925426e-01 6.59221113e-01
3.48484814e-01 9.44429755e-01 1.33697438e+00 -1.33476689e-01
2.19175741e-01 1.53743088e-01 3.69035661e-01 8.03756654e-01
2.04426512e-01 -6.05545230e-02 1.87091175e-02 -1.36771426e-01
7.88940012e-01 1.42339515e-02 -3.07325453e-01 -5.50286651e-01
-1.23068893e+00 7.09872663e-01 7.61444747e-01 3.96749467e-01
-2.35481128e-01 -1.86183944e-01 5.22876084e-01 -1.89838812e-01
7.53499150e-01 2.08825961e-01 -5.00214756e-01 1.28686562e-01
-1.22568643e+00 -3.07751268e-01 4.58862603e-01 9.02523756e-01
6.61027849e-01 -1.15202695e-01 -3.93548757e-01 9.56979215e-01
3.95483136e-01 4.57209498e-01 7.16496348e-01 -8.62733006e-01
3.64329875e-01 6.31711781e-01 -1.42946839e-01 -7.52831221e-01
-6.36193097e-01 -5.47191858e-01 -1.17508805e+00 -1.47766635e-01
5.48750460e-01 -1.31098270e-01 -1.31348300e+00 1.73079538e+00
5.04791796e-01 3.54556262e-01 -1.75025702e-01 9.24751282e-01
7.70860493e-01 1.66472390e-01 2.83317506e-01 -3.88569653e-01
1.25049913e+00 -9.47804570e-01 -7.35469341e-01 -4.24909778e-02
8.06908429e-01 -5.64031065e-01 7.91085362e-01 1.87678009e-01
-7.95539379e-01 -4.37059879e-01 -9.28647816e-01 2.55854607e-01
7.30621070e-02 1.89457729e-01 7.18569994e-01 6.16091251e-01
-7.70055354e-01 5.60647428e-01 -1.27339876e+00 3.39264870e-02
9.54068661e-01 5.73306918e-01 -3.44729632e-01 -1.92732923e-02
-9.16579247e-01 6.32979393e-01 6.50414109e-01 1.56377226e-01
-5.65375805e-01 -9.22863960e-01 -9.03838158e-01 -2.22165227e-01
5.07644176e-01 -5.17037094e-01 9.94482934e-01 -9.30525661e-01
-1.38178933e+00 9.33266878e-01 -7.48703629e-02 -5.55591941e-01
7.81518221e-01 -5.91750517e-02 2.19181813e-02 3.17798167e-01
3.26959819e-01 9.32889044e-01 6.56253695e-01 -1.08530509e+00
-3.74365956e-01 -5.88162065e-01 -4.61484760e-01 1.95852682e-01
-2.60671675e-01 -4.07770097e-01 -6.78283691e-01 -7.84550190e-01
6.02856219e-01 -1.18825328e+00 -5.56851327e-01 1.06409833e-01
-6.23126507e-01 -7.76064098e-02 6.31183684e-01 -8.27742577e-01
9.65170026e-01 -1.86116886e+00 2.50634011e-02 3.81373495e-01
3.22137088e-01 3.84228259e-01 1.73555821e-01 -3.44404221e-01
1.40351519e-01 1.60742983e-01 -8.52959216e-01 -2.96524763e-01
-3.79515469e-01 1.64063722e-01 1.70161426e-01 6.17607892e-01
1.44427419e-01 9.91892278e-01 -9.15896058e-01 -9.74623859e-01
3.52092326e-01 4.67969269e-01 -4.36182112e-01 1.80423960e-01
-1.45520777e-01 1.37787914e+00 -7.40102351e-01 5.63959181e-01
7.91481256e-01 -6.60234153e-01 2.89500147e-01 -1.92198291e-01
4.22173351e-01 -1.37206540e-01 -9.32132304e-01 2.01563215e+00
-4.58666056e-01 2.17654780e-01 5.51402150e-03 -1.17180324e+00
8.01840961e-01 3.65424156e-01 8.95997226e-01 -6.47114277e-01
1.93807676e-01 3.59355718e-01 2.84656100e-02 -4.51486260e-01
-1.70603693e-01 -2.76006550e-01 -7.33307516e-03 2.88823903e-01
-3.86157930e-02 -3.60335678e-01 -4.34128121e-02 2.39554867e-01
7.89558351e-01 1.06367402e-01 2.92759567e-01 -5.42199075e-01
6.66650832e-01 -9.15430859e-02 8.53499174e-01 5.82382143e-01
-5.30612230e-01 6.26239955e-01 2.42281720e-01 -2.29598790e-01
-7.46273458e-01 -8.99018586e-01 -6.13199949e-01 4.67052191e-01
3.42512608e-01 -8.41071247e-04 -8.46011996e-01 -1.16402924e+00
-8.35389867e-02 4.20665503e-01 -5.82612038e-01 9.97362807e-02
-4.14912909e-01 -9.06935990e-01 3.51864368e-01 7.36937106e-01
7.06345558e-01 -7.38366306e-01 -3.66292626e-01 1.48193106e-01
-3.22112113e-01 -1.37150764e+00 -6.51734233e-01 1.42420262e-01
-1.31423509e+00 -1.15008259e+00 -1.00373566e+00 -6.87522054e-01
1.05569386e+00 6.97798654e-02 8.83050203e-01 1.53687745e-01
-4.33855832e-01 1.58725470e-01 -1.23702951e-01 -2.44428053e-01
-4.06125337e-01 8.82095322e-02 -1.92982376e-01 -1.44070938e-01
1.78281479e-02 -2.83745050e-01 -6.85909986e-01 5.43033481e-01
-9.33679819e-01 2.20293730e-01 6.52960777e-01 1.21784234e+00
1.02859426e+00 5.27628735e-02 6.52338982e-01 -1.30011523e+00
2.22295672e-01 -3.53594363e-01 -6.44557953e-01 3.32458019e-01
-9.18701172e-01 1.03702687e-01 3.06700617e-01 -1.99323401e-01
-1.25613952e+00 4.15249109e-01 -1.44861221e-01 -4.20363903e-01
-6.97274432e-02 4.78220791e-01 -1.86754707e-02 -1.61483601e-01
3.23033988e-01 6.93857297e-02 3.76170725e-01 -1.57740861e-01
2.68413186e-01 6.53663218e-01 4.14173603e-01 -6.07828856e-01
4.80657071e-01 7.44970143e-01 7.74107054e-02 -3.94012660e-01
-1.15641224e+00 -7.54182100e-01 -9.69053805e-01 9.26449299e-02
1.12920082e+00 -9.03959215e-01 -4.17141825e-01 4.28197473e-01
-7.39698768e-01 -3.02222908e-01 -5.19439653e-02 6.35276258e-01
-4.27359194e-01 8.21467876e-01 -6.71770930e-01 -5.51414549e-01
-6.96720779e-01 -1.62837088e+00 1.10925651e+00 2.63631940e-01
-6.27605766e-02 -1.33522975e+00 -8.67756978e-02 6.55907691e-01
3.77577126e-01 3.48902583e-01 7.36587048e-01 -1.00011599e+00
-4.08060461e-01 -1.60260469e-01 -3.63206446e-01 5.22054195e-01
5.12860954e-01 -3.28879416e-01 -7.73481488e-01 -3.47112060e-01
8.91689211e-02 -4.67368335e-01 9.84163642e-01 8.13316107e-01
1.48050213e+00 1.01863384e-01 -4.88128096e-01 6.65243089e-01
1.32822120e+00 -6.29708311e-03 3.41243684e-01 -4.04857621e-02
9.66017902e-01 6.44999862e-01 8.03821504e-01 2.87900656e-01
3.46048951e-01 5.94993889e-01 3.22994262e-01 -4.06954139e-01
-4.56160009e-02 2.11623702e-02 -3.30397367e-01 8.65998626e-01
2.53628522e-01 2.06267551e-01 -1.04915321e+00 4.39697415e-01
-1.95974946e+00 -4.24992681e-01 -2.08428249e-01 2.21383286e+00
1.11817014e+00 2.27391407e-01 -1.70819368e-02 -1.57976776e-01
8.70601773e-01 -1.37411341e-01 -8.34964633e-01 2.76310176e-01
2.46338353e-01 1.73326269e-01 6.39225543e-01 4.29383248e-01
-1.37717485e+00 8.76767099e-01 5.56082058e+00 1.09012032e+00
-1.10532832e+00 2.56194681e-01 1.21650541e+00 2.46576548e-01
-1.51024014e-01 -1.26544416e-01 -6.39928401e-01 6.41247332e-01
5.63015103e-01 2.80734748e-01 -1.27372220e-01 7.25744188e-01
2.25511089e-01 -3.33257288e-01 -9.96003568e-01 7.86597013e-01
-1.19409643e-01 -1.27066565e+00 -2.04932004e-01 -4.40083519e-02
1.05167460e+00 -3.57891177e-03 6.14880696e-02 1.11553878e-01
2.14194864e-01 -1.06580424e+00 1.70474395e-01 2.93480575e-01
7.50873208e-01 -5.97947478e-01 1.02109802e+00 4.72654283e-01
-9.69572008e-01 2.68877327e-01 -2.84307331e-01 4.38235641e-01
2.40276575e-01 9.16324914e-01 -1.02987099e+00 7.37054646e-01
5.09722471e-01 8.77127945e-01 -4.66996640e-01 1.06319833e+00
-1.49006948e-01 7.26965427e-01 -2.15102673e-01 3.45352739e-01
2.94031471e-01 -3.78433287e-01 2.77161986e-01 1.05191028e+00
-2.60659400e-03 5.78126796e-02 4.93545681e-01 8.99389863e-01
-4.09356318e-02 2.46596366e-01 -2.01211989e-01 2.55906194e-01
3.92461330e-01 1.46279919e+00 -1.28035343e+00 -3.48939747e-01
-2.68499225e-01 8.20530236e-01 1.67402059e-01 9.39109176e-02
-1.03401744e+00 5.33937812e-02 -3.96251604e-02 -8.43602717e-02
1.55649573e-01 -3.27827223e-02 -6.61386132e-01 -1.00874567e+00
-9.74176973e-02 -6.35467231e-01 5.34665465e-01 -2.34579235e-01
-1.39971828e+00 5.75816393e-01 -1.66516304e-02 -1.26774883e+00
-4.88737002e-02 -4.32379991e-01 -5.46881914e-01 8.07425261e-01
-1.71750700e+00 -1.26692140e+00 -2.46984869e-01 4.14502740e-01
5.77843487e-01 -5.57735981e-03 5.92879176e-01 1.99863806e-01
-7.31995344e-01 5.90820491e-01 1.54210001e-01 2.43309200e-01
8.23894262e-01 -1.36817467e+00 -5.46087548e-02 6.38133109e-01
1.08074084e-01 5.35949826e-01 1.81955144e-01 -8.38478208e-01
-9.22767222e-01 -1.13186431e+00 2.66696364e-01 -3.51873398e-01
5.69110215e-01 7.05605298e-02 -1.09877264e+00 5.20595968e-01
-9.97358635e-02 5.07849097e-01 8.39247286e-01 -1.07618943e-01
-5.11463434e-02 -2.09568348e-02 -1.32385302e+00 2.83319563e-01
6.50268376e-01 -2.37164825e-01 -3.61187965e-01 5.87449491e-01
6.83125615e-01 -8.05340469e-01 -1.12605679e+00 9.20722485e-01
2.75806993e-01 -7.67551422e-01 8.15125465e-01 -2.64892876e-01
2.29467615e-01 -2.14502469e-01 1.43888935e-01 -1.00511456e+00
-5.18309372e-03 -5.17847121e-01 1.83731571e-01 1.13374567e+00
5.38026869e-01 -6.07128918e-01 8.73081625e-01 1.01861453e+00
-2.78514177e-01 -1.12530851e+00 -1.02835596e+00 -6.54627144e-01
3.35312605e-01 -4.02016163e-01 1.38416439e-01 9.62186158e-01
-1.82550326e-01 -6.04012841e-03 -2.37470046e-01 2.15809047e-01
7.45383143e-01 4.77556214e-02 4.36607063e-01 -1.17238426e+00
-3.52315307e-01 -2.69325107e-01 -3.07788640e-01 -8.44982505e-01
4.83547866e-01 -1.17922926e+00 2.29343221e-01 -1.55918455e+00
5.47145426e-01 -8.57370257e-01 -5.43367565e-01 5.11647761e-01
-6.26445353e-01 4.30462867e-01 -9.99489501e-02 5.80037355e-01
-7.19237626e-01 4.06849802e-01 1.64500797e+00 -1.84201896e-01
-2.11659029e-01 2.48633891e-01 -5.61114728e-01 8.23836923e-01
7.38586068e-01 -5.09788275e-01 -6.57458484e-01 -2.55503863e-01
-3.52015048e-01 2.31466785e-01 1.86804578e-01 -8.72826040e-01
1.88842610e-01 3.92234400e-02 3.59574914e-01 -4.38388467e-01
-1.64487168e-01 -8.65837693e-01 -2.69288663e-02 5.76977074e-01
-3.89553159e-01 -4.60363656e-01 3.15975696e-01 8.43675077e-01
-2.37889364e-01 -1.94553718e-01 1.01647389e+00 -5.70088290e-02
-5.79113722e-01 5.24324477e-01 2.00626254e-01 2.93678135e-01
1.22129941e+00 -1.57868356e-01 1.34212226e-01 -1.18792877e-01
-8.91107917e-01 4.60181385e-01 2.75649905e-01 1.00832783e-01
4.00306016e-01 -1.03975785e+00 -6.61488533e-01 -1.44161759e-02
5.72997145e-02 4.29084301e-01 4.92948413e-01 1.35313964e+00
-3.82273048e-01 4.32904780e-01 1.75783429e-02 -1.27494729e+00
-1.09125757e+00 4.33791280e-01 2.77935058e-01 -7.30135620e-01
-7.22723007e-01 9.00203407e-01 3.81196886e-01 -4.99668658e-01
1.72319740e-01 -4.27934438e-01 -4.10217009e-02 -3.26331288e-01
1.66676462e-01 5.67391329e-02 1.96828291e-01 -6.58832729e-01
-4.50058460e-01 6.70063555e-01 -3.51278633e-01 1.38950646e-01
1.01841903e+00 -1.87324315e-01 2.14343369e-02 1.44377321e-01
1.18383634e+00 -1.93870500e-01 -1.58425808e+00 -5.08215547e-01
2.39696652e-01 -2.67375618e-01 2.57748097e-01 -9.01102781e-01
-1.35121644e+00 8.56978655e-01 7.79212713e-01 -3.14103425e-01
1.03483975e+00 4.65142773e-03 9.10514474e-01 6.06041215e-02
4.07661736e-01 -1.07550526e+00 1.50172204e-01 -2.30938401e-02
3.65450829e-01 -1.87302220e+00 1.87814102e-01 -7.80516326e-01
-1.07512593e+00 8.16038370e-01 5.96964180e-01 4.06063758e-02
7.91883290e-01 2.02037960e-01 1.26809224e-01 -1.57792479e-01
-1.37288556e-01 -9.79265869e-02 7.12561846e-01 4.12895501e-01
5.52159011e-01 1.76581115e-01 -3.85846466e-01 6.18237734e-01
3.55615765e-01 -1.99266169e-02 -1.58377215e-02 7.71647751e-01
-2.69954056e-01 -1.11198890e+00 -3.16680707e-02 6.21381760e-01
-7.66655147e-01 -2.78229862e-02 7.95487240e-02 7.68874526e-01
-6.94219321e-02 7.91854739e-01 -1.74266234e-01 9.79694203e-02
-1.39328346e-01 -1.16312370e-01 3.81281406e-01 -7.92120218e-01
-4.18009460e-01 4.04197931e-01 -3.40801686e-01 -5.19328892e-01
-6.59218669e-01 -5.23285091e-01 -1.76632822e+00 3.80615711e-01
-7.20222056e-01 7.76503980e-02 5.78389823e-01 1.23516393e+00
3.70817840e-01 5.69110632e-01 6.37168765e-01 -5.74337542e-01
-6.51675701e-01 -8.46781135e-01 -5.34048557e-01 5.85129440e-01
4.55786884e-02 -6.23932064e-01 -1.89429998e-01 1.21213987e-01] | [14.708378791809082, -2.1769192218780518] |
cf049968-2a16-4dfb-8065-6621381ccb14 | nnmobile-net-rethinking-cnn-design-for-deep | 2306.01289 | null | https://arxiv.org/abs/2306.01289v1 | https://arxiv.org/pdf/2306.01289v1.pdf | NNMobile-Net: Rethinking CNN Design for Deep Learning-Based Retinopathy Research | Retinal diseases (RD) are the leading cause of severe vision loss or blindness. Deep learning-based automated tools play an indispensable role in assisting clinicians in diagnosing and monitoring RD in modern medicine. Recently, an increasing number of works in this field have taken advantage of Vision Transformer to achieve state-of-the-art performance with more parameters and higher model complexity compared to Convolutional Neural Networks (CNNs). Such sophisticated and task-specific model designs, however, are prone to be overfitting and hinder their generalizability. In this work, we argue that a channel-aware and well-calibrated CNN model may overcome these problems. To this end, we empirically studied CNN's macro and micro designs and its training strategies. Based on the investigation, we proposed a no-new-MobleNet (nn-MobileNet) developed for retinal diseases. In our experiments, our generic, simple and efficient model superseded most current state-of-the-art methods on four public datasets for multiple tasks, including diabetic retinopathy grading, fundus multi-disease detection, and diabetic macular edema classification. Our work may provide novel insights into deep learning architecture design and advance retinopathy research. | ['Yalin Wang', 'Oana M. Dumitrascu', 'Natasha Lepore', 'Peijie Qiu', 'Wenhui Zhu'] | 2023-06-02 | null | null | null | null | ['diabetic-retinopathy-grading'] | ['medical'] | [-6.29038140e-02 -3.04155499e-01 -4.83764745e-02 -3.85730624e-01
-2.35579237e-01 -9.80740786e-02 9.92862284e-02 -1.41828522e-01
-5.52569151e-01 7.22892225e-01 3.89395654e-02 -8.50216389e-01
-1.30616188e-01 -8.09626520e-01 -4.10556465e-01 -6.64168298e-01
1.46806091e-01 -3.04396115e-02 3.75680447e-01 -7.93970153e-02
2.33705387e-01 5.54450274e-01 -1.70433450e+00 3.39010298e-01
1.38554275e+00 1.00578296e+00 1.88656032e-01 7.43130147e-01
4.52526547e-02 8.06773603e-01 -4.15312648e-01 -5.28326094e-01
3.65676880e-01 -3.75330597e-01 -5.81423938e-01 -1.06603310e-01
6.77345037e-01 -6.69508278e-01 -3.35428208e-01 8.31365407e-01
1.17104089e+00 -4.76416558e-01 4.31415886e-01 -5.50203085e-01
-8.38891566e-01 1.13959774e-01 -6.19304180e-01 4.61728841e-01
-3.65880966e-01 7.26446450e-01 4.44276243e-01 -3.45397562e-01
1.49657741e-01 9.93611574e-01 7.64423728e-01 6.50691152e-01
-8.64656091e-01 -5.91737866e-01 -2.82855798e-02 4.34459031e-01
-9.37220633e-01 -5.33247948e-01 1.03276968e-01 -6.33483946e-01
7.34476805e-01 4.51035425e-02 8.89415801e-01 8.90788078e-01
2.77035594e-01 3.50229412e-01 1.36231256e+00 -4.39137787e-01
-5.41794766e-03 3.35952230e-02 6.22690022e-02 8.58731568e-01
7.94522822e-01 2.19656587e-01 3.83298956e-02 1.08744413e-01
1.06943536e+00 -1.69027671e-01 -3.43688577e-01 7.72137493e-02
-9.18702185e-01 7.51993358e-01 6.65136755e-01 2.54434459e-02
-4.00860250e-01 1.66260570e-01 2.55184263e-01 1.63190067e-01
4.98004854e-01 4.99506682e-01 -5.05067647e-01 -1.24696918e-01
-3.83184135e-01 7.17760995e-02 4.47297037e-01 5.01750767e-01
4.31267709e-01 -2.16905683e-01 -5.85023522e-01 9.33729470e-01
2.84422308e-01 2.03260675e-01 3.56365800e-01 -7.92400301e-01
2.98413038e-01 9.83022749e-01 8.16165644e-04 -5.78751445e-01
-7.34839380e-01 -9.14761424e-01 -1.22228658e+00 4.40075666e-01
5.44591427e-01 -4.09279525e-01 -1.16046333e+00 1.00342286e+00
1.17227901e-02 2.19208375e-01 -1.99970394e-01 1.16946769e+00
1.06882048e+00 -5.68088852e-02 -6.93681976e-03 3.29874717e-02
1.53468645e+00 -1.00044262e+00 -3.69739801e-01 -7.57637694e-02
5.99610507e-01 -8.93697739e-01 9.67194855e-01 3.81680936e-01
-9.98865485e-01 -5.91178596e-01 -8.46475124e-01 -2.85971284e-01
-1.19306445e-01 7.44522691e-01 9.78258729e-01 1.01808906e+00
-1.25750899e+00 3.02055359e-01 -7.26761699e-01 -6.51258349e-01
1.08300173e+00 3.36935252e-01 -3.11813019e-02 -2.42891714e-01
-6.92577004e-01 9.58359957e-01 -2.24555237e-03 4.34244692e-01
-6.13521039e-01 -5.81113338e-01 -1.70382395e-01 -3.50941420e-01
5.28733097e-02 -1.53509927e+00 1.21755385e+00 -6.46514058e-01
-1.63792562e+00 9.42849398e-01 -1.98893994e-01 -7.65302479e-01
6.18705809e-01 -3.03038210e-01 -4.96337593e-01 -1.79850291e-02
-3.59118670e-01 4.68345582e-01 5.76760352e-01 -8.81910622e-01
-8.59345078e-01 -5.15599310e-01 2.81560361e-01 -2.10460991e-01
-3.28709126e-01 2.89367288e-01 -5.41865408e-01 -2.13668823e-01
-1.43274695e-01 -6.69437289e-01 -5.23448765e-01 4.65406656e-01
-5.40995121e-01 -3.04259975e-02 3.41939986e-01 -4.61835355e-01
1.23579693e+00 -1.83285379e+00 -3.62796605e-01 -1.94889978e-02
8.52704942e-01 1.07019341e+00 -2.02695161e-01 -4.83169742e-02
3.11139166e-01 1.70483708e-01 2.26789415e-01 -1.89404756e-01
-5.97325802e-01 1.59613788e-02 3.54916811e-01 4.98320580e-01
1.52750775e-01 9.41745937e-01 -6.85977042e-01 -2.78374910e-01
3.52302253e-01 7.21446812e-01 -5.37506938e-01 1.45299599e-01
-7.24847764e-02 6.06169403e-01 -4.41634178e-01 8.39130223e-01
5.69738269e-01 -5.83503127e-01 -1.03631184e-01 -4.55640525e-01
-4.47932154e-01 1.54632047e-01 -7.09115028e-01 1.25973761e+00
-3.61458033e-01 9.78188276e-01 -3.89858305e-01 -7.16259181e-01
8.48554552e-01 -3.26835155e-03 1.71822041e-01 -8.67840230e-01
3.78621876e-01 3.68532389e-01 5.56301832e-01 -1.11288798e+00
3.37172709e-02 4.18133467e-01 1.16193914e+00 -1.06554531e-01
-3.40054423e-01 6.44465983e-01 2.27998998e-02 -4.79441106e-01
9.96115863e-01 -2.67226815e-01 3.02478999e-01 1.11028820e-01
4.01785105e-01 -3.55470359e-01 2.71876961e-01 8.89222622e-01
-3.96831751e-01 7.02372909e-01 5.11753500e-01 -7.82789528e-01
-9.32073116e-01 -6.68956280e-01 -4.27363813e-01 3.82440954e-01
-7.95424059e-02 -9.10312459e-02 -6.88203633e-01 -3.90875131e-01
1.19195946e-01 -9.03104842e-02 -5.02326131e-01 1.43279627e-01
-2.80786246e-01 -1.32635367e+00 8.56572568e-01 3.78801286e-01
9.93002117e-01 -5.11697054e-01 -6.51440144e-01 2.82483906e-01
3.70976292e-02 -9.76981580e-01 6.67754468e-03 -4.32840317e-01
-1.05965137e+00 -1.53128290e+00 -1.08199501e+00 -7.91803598e-01
6.96599722e-01 4.82613325e-01 8.09382915e-01 2.66081244e-02
-7.61741817e-01 -9.30525512e-02 -2.58854717e-01 -7.77358830e-01
-1.89990968e-01 9.37460288e-02 -2.24496439e-01 3.64083529e-01
6.29517794e-01 -4.47227478e-01 -1.49698484e+00 2.47804075e-01
-3.77951026e-01 -2.24076919e-02 1.23175275e+00 6.53878033e-01
5.62475562e-01 -1.02573983e-01 5.74894249e-01 -9.64123130e-01
8.11070263e-01 -2.97753185e-01 -7.55661607e-01 1.36492938e-01
-1.02624619e+00 -2.37464428e-01 2.93043435e-01 -3.15807879e-01
-8.71930480e-01 -3.10481489e-01 -1.09297253e-01 -1.44572183e-01
-3.36763769e-01 4.35856730e-01 1.23464368e-01 -6.32320106e-01
1.08407784e+00 -1.44924670e-02 3.11711848e-01 -8.10819387e-01
2.05658466e-01 1.13114011e+00 5.01742959e-01 -1.79156214e-01
2.48000592e-01 5.68867028e-01 1.98557019e-01 -8.79143119e-01
-7.52903700e-01 -5.95271528e-01 -1.97113737e-01 -2.46902943e-01
8.76013458e-01 -1.19900703e+00 -9.69707131e-01 1.09149575e+00
-1.21650600e+00 -1.79224491e-01 1.05436444e-01 7.03205526e-01
3.06259003e-02 3.67185980e-01 -4.22348410e-01 -6.36326432e-01
-3.79926354e-01 -1.11538327e+00 6.98614538e-01 6.40087724e-01
3.48758012e-01 -8.32652986e-01 5.50687686e-02 7.49419630e-01
9.56379116e-01 4.49026078e-02 1.04447091e+00 -4.93925698e-02
-8.45513642e-01 -6.20589731e-03 -9.84043002e-01 8.54058146e-01
2.60852903e-01 2.61940897e-01 -1.23282945e+00 -1.84433520e-01
-4.65496421e-01 -4.11949214e-03 1.22482872e+00 1.00250196e+00
1.42114592e+00 -1.38850838e-01 -3.25306594e-01 9.36563492e-01
1.52820230e+00 2.00294301e-01 1.27911353e+00 5.98179400e-01
6.29593313e-01 3.96797210e-01 1.03895115e-02 5.70640385e-01
6.36243522e-01 4.74613219e-01 6.89504206e-01 -7.58343935e-01
-5.61812639e-01 1.57493427e-01 -1.37493894e-01 2.73926497e-01
-9.42345083e-01 -2.71520972e-01 -8.89646947e-01 5.44006228e-01
-1.64959180e+00 -6.87347114e-01 -7.10716069e-01 2.06210613e+00
8.29833746e-01 -3.60531956e-02 2.46647477e-01 -2.34544128e-01
6.27082646e-01 -3.14102888e-01 -6.59487545e-01 -3.99702638e-02
-3.96998137e-01 2.99413323e-01 8.14730465e-01 -1.38687998e-01
-1.07703710e+00 6.98122561e-01 5.81973028e+00 4.45678264e-01
-1.34451950e+00 5.44775315e-02 7.32465029e-01 -8.70810822e-02
1.39120236e-01 -2.04757497e-01 -8.80184114e-01 5.36784053e-01
8.21680307e-01 3.79456878e-01 2.36774847e-01 3.97769868e-01
7.27150261e-01 -1.81640297e-01 -7.75083244e-01 1.12235379e+00
-2.21907064e-01 -1.65850937e+00 1.44240335e-01 3.87137562e-01
7.33925700e-01 4.80276585e-01 2.71440268e-01 -2.86569208e-01
1.15404418e-02 -1.23892915e+00 -1.69571653e-01 8.71700346e-01
1.04318905e+00 -2.87628144e-01 1.22755694e+00 -1.99887812e-01
-5.90409815e-01 -2.15078652e-01 -6.40242219e-01 -2.37060189e-01
-2.49194384e-01 8.96841109e-01 -7.27602422e-01 3.87905538e-01
8.15730691e-01 9.04634714e-01 -8.95280421e-01 2.17103267e+00
-8.46321657e-02 7.96170473e-01 1.45392809e-02 -8.15730393e-02
9.20649618e-02 -2.11151659e-01 3.61571610e-01 9.65404153e-01
3.82578969e-01 3.97322960e-02 -2.94233143e-01 9.41666901e-01
-5.94753698e-02 1.95152059e-01 -3.09263736e-01 1.61522865e-01
2.90040910e-01 1.07869911e+00 -2.57482052e-01 2.15951741e-01
-8.96645129e-01 3.69556397e-01 -1.21977460e-02 5.37143767e-01
-5.71875274e-01 -3.41778159e-01 9.61636722e-01 3.84730577e-01
2.40013916e-02 -2.60947067e-02 -5.27559638e-01 -9.82082009e-01
1.24163881e-01 -7.36539423e-01 1.15072176e-01 -5.93986452e-01
-1.32350504e+00 5.38779497e-01 -7.21873045e-01 -1.47326648e+00
4.97203231e-01 -9.82800722e-01 -5.59838951e-01 1.04715109e+00
-2.34732723e+00 -1.30599022e+00 -7.65881658e-01 7.35814810e-01
1.02018118e-01 -6.93809152e-01 6.74268007e-01 7.42799520e-01
-9.66560721e-01 6.01954460e-01 1.80985123e-01 1.93935946e-01
1.04702592e+00 -8.82064641e-01 2.30765939e-01 7.56759942e-01
-4.44964945e-01 5.33313811e-01 2.31624112e-01 -2.29435042e-01
-1.00134039e+00 -1.40293419e+00 7.46600151e-01 -1.70806333e-01
4.65421468e-01 2.52975792e-01 -5.75110078e-01 3.36896181e-01
-4.80505116e-02 1.56131819e-01 6.26897991e-01 1.41509846e-01
-1.68973759e-01 -5.10054410e-01 -8.43150318e-01 6.67240918e-01
1.09000540e+00 -1.88183069e-01 4.30588471e-03 3.52778167e-01
3.98724437e-01 -5.26140571e-01 -7.28360355e-01 4.10196632e-01
6.98263228e-01 -1.34249520e+00 9.03343022e-01 -7.76678622e-01
5.96828043e-01 -2.42569894e-01 2.40532175e-01 -1.17670226e+00
-2.82008648e-01 -7.38306224e-01 -1.40700758e-01 8.37059140e-01
3.90161842e-01 -1.20488989e+00 6.65091872e-01 1.64852768e-01
-3.24311018e-01 -1.09784257e+00 -7.24011302e-01 -6.28062725e-01
-1.46939866e-02 -1.77339584e-01 4.51749861e-01 5.31075537e-01
-7.11576283e-01 -1.44817343e-03 -2.84189820e-01 3.07455927e-01
5.93163550e-01 -2.43710741e-01 8.38633239e-01 -1.54096997e+00
-7.36002326e-02 -6.37676537e-01 -8.91042769e-01 -9.88045871e-01
-4.57392871e-01 -4.38926667e-01 -5.50584733e-01 -1.84838378e+00
1.68807924e-01 -4.82363224e-01 -3.77331972e-01 5.18670559e-01
-1.07215352e-01 2.97647983e-01 -2.89018810e-01 2.16163024e-01
-2.42356315e-01 1.52563348e-01 1.71992064e+00 -1.01389021e-01
-4.22390163e-01 4.57493663e-01 -1.13750732e+00 6.36622071e-01
9.87305582e-01 3.80051807e-02 -3.80924135e-01 -7.34671235e-01
2.52477050e-01 -3.98793697e-01 9.25387383e-01 -1.06613338e+00
4.47171897e-01 4.43219617e-02 3.08389872e-01 -1.21652938e-01
-1.57824438e-02 -4.15901721e-01 -2.15518519e-01 4.77575988e-01
-3.40402052e-02 -5.95978379e-01 1.80547833e-01 3.97278935e-01
-1.98879495e-01 3.84470642e-01 1.00215232e+00 -5.08969724e-02
-6.64023817e-01 5.98500907e-01 -1.61228165e-01 -5.37707992e-02
8.97749186e-01 -3.48345757e-01 -1.13178754e+00 -6.29696101e-02
-3.47219914e-01 1.71556517e-01 1.15757078e-01 2.60325849e-01
7.51281500e-01 -7.59252429e-01 -1.13096273e+00 3.15469593e-01
3.80266726e-01 5.95165864e-02 4.44636583e-01 1.24395418e+00
-8.43357980e-01 6.11772239e-01 -4.10311103e-01 -6.51689589e-01
-1.20842564e+00 -4.50480543e-02 9.59144652e-01 2.12800130e-01
-7.75642037e-01 9.55319345e-01 -3.26157957e-02 1.06314428e-01
5.38245440e-01 -7.67297208e-01 -5.44510841e-01 -1.23782478e-01
6.05149508e-01 8.64872515e-01 2.64486998e-01 1.01740435e-01
4.47997563e-02 8.01785767e-01 -4.06742305e-01 7.08904207e-01
1.22292340e+00 -3.06047320e-01 -3.16729426e-01 -4.26885411e-02
6.84517860e-01 -2.90585577e-01 -1.18711483e+00 -1.09127492e-01
-4.80650187e-01 -5.14224648e-01 3.82590503e-01 -1.20506036e+00
-1.33643723e+00 1.05128431e+00 1.23769164e+00 2.00135317e-02
1.30392206e+00 -1.73116639e-01 8.31493616e-01 3.16872090e-01
3.40686172e-01 -6.74511492e-01 -2.88400948e-01 7.97201321e-02
5.45712709e-01 -1.55135357e+00 -1.30486459e-01 -5.49467742e-01
-2.97069013e-01 1.32871532e+00 7.89963186e-01 2.30522603e-01
6.97147250e-01 -1.92538977e-01 4.54012603e-01 -1.54594719e-01
-4.99655426e-01 -7.44842291e-01 4.59283352e-01 9.96075749e-01
5.35151064e-01 9.86147150e-02 -3.09662908e-01 5.79775095e-01
2.25935504e-02 6.39553905e-01 6.56643152e-01 3.79073948e-01
-4.84061390e-01 -1.15097773e+00 4.23351415e-02 9.27455246e-01
-5.42551100e-01 -3.26297522e-01 -2.73936450e-01 7.15370119e-01
5.52332938e-01 1.12230158e+00 -1.93285584e-01 -2.41129667e-01
5.11473000e-01 -4.94868726e-01 6.05026186e-01 -5.06891310e-01
-5.07122874e-01 -5.95560856e-02 2.84852147e-01 -5.97688258e-01
-6.59021974e-01 -1.75714985e-01 -5.74283242e-01 -3.97619396e-01
-2.76100606e-01 -6.96447432e-01 4.98213321e-01 6.47544622e-01
7.94364691e-01 7.54245162e-01 4.23702419e-01 -2.61193991e-01
-4.66485828e-01 -9.84077990e-01 -6.44023418e-01 -1.74649477e-01
5.68738818e-01 -4.69023347e-01 -9.12728533e-02 2.35500839e-02] | [15.815529823303223, -3.9820003509521484] |
d39fbc49-28a8-4012-bcb2-9e3034ba7b06 | neuralizing-regular-expressions-for-slot | null | null | https://aclanthology.org/2021.emnlp-main.747 | https://aclanthology.org/2021.emnlp-main.747.pdf | Neuralizing Regular Expressions for Slot Filling | Neural models and symbolic rules such as regular expressions have their respective merits and weaknesses. In this paper, we study the integration of the two approaches for the slot filling task by converting regular expressions into neural networks. Specifically, we first convert regular expressions into a special form of finite-state transducers, then unfold its approximate inference algorithm as a bidirectional recurrent neural model that performs slot filling via sequence labeling. Experimental results show that our model has superior zero-shot and few-shot performance and stays competitive when there are sufficient training data. | ['Kewei Tu', 'Zijian Jin', 'Chengyue Jiang'] | null | null | null | null | emnlp-2021-11 | ['slot-filling'] | ['natural-language-processing'] | [ 3.19770664e-01 6.34676218e-01 -9.59311604e-01 -5.12054443e-01
-5.73269904e-01 -1.47535041e-01 5.29885769e-01 -1.90477252e-01
-3.20395619e-01 9.23111975e-01 1.66756943e-01 -7.39420116e-01
2.71432936e-01 -1.02790987e+00 -6.25686347e-01 -1.94191962e-01
2.09248066e-03 5.41434944e-01 2.48888448e-01 -2.99174547e-01
-1.78471208e-01 3.40825349e-01 -1.67843986e+00 1.88136071e-01
8.35509121e-01 1.06372392e+00 -3.25166695e-02 3.97730350e-01
-1.12884474e+00 1.38537621e+00 -3.97577077e-01 -4.51417953e-01
3.48184461e-04 -5.99243462e-01 -1.04137266e+00 -1.33240372e-01
-4.23929900e-01 -1.12575635e-01 -5.33177137e-01 1.20690072e+00
1.28148720e-01 5.49099207e-01 1.99347153e-01 -8.58508408e-01
-6.08886719e-01 9.11886454e-01 -5.33160530e-02 1.88119084e-01
2.41079643e-01 -3.52608025e-01 9.48430479e-01 -8.36746275e-01
5.12211978e-01 1.27349389e+00 7.69839883e-01 7.37517178e-01
-1.40665996e+00 -4.59355772e-01 1.88678458e-01 1.12977408e-01
-1.48276436e+00 -7.31051326e-01 4.13982898e-01 2.56849490e-02
1.79589868e+00 1.41878491e-02 4.91044253e-01 7.48021543e-01
-1.14418805e-01 1.02708042e+00 6.22691154e-01 -8.07877243e-01
5.41748703e-01 2.91848071e-02 4.28626448e-01 9.46304560e-01
-2.73845762e-01 2.54470170e-01 -4.39651340e-01 -3.82765859e-01
8.03763568e-01 7.70935118e-02 2.58508950e-01 1.74810499e-01
-4.94507402e-01 1.20473039e+00 9.00734812e-02 4.07734662e-01
-5.15277445e-01 1.38131604e-01 8.12612116e-01 4.67983037e-01
5.11325061e-01 3.06317091e-01 -1.62528232e-01 -4.14243817e-01
-1.03698719e+00 7.55918249e-02 8.85139704e-01 1.15300858e+00
8.52001429e-01 5.06183982e-01 -4.70692843e-01 1.12306190e+00
-1.17944740e-02 1.06688730e-01 7.93227732e-01 -9.32940185e-01
-5.07863751e-03 4.67043787e-01 -2.80346423e-02 -6.81270301e-01
-2.84187198e-01 -4.55751680e-02 -6.64413035e-01 -3.55965316e-01
-3.23886812e-01 -1.60796866e-01 -1.32069206e+00 1.91211629e+00
1.92288667e-01 5.91927946e-01 2.08726421e-01 4.77969944e-01
7.57058799e-01 1.22472823e+00 4.48380351e-01 -5.26435912e-01
1.34705484e+00 -1.32645428e+00 -1.35490882e+00 -6.40423059e-01
7.80145526e-01 -2.80782402e-01 8.58867288e-01 -1.58009365e-01
-1.52449441e+00 -5.49199522e-01 -9.92268384e-01 -1.71919242e-01
-5.62745929e-01 -2.08568703e-02 9.24980760e-01 4.48381901e-01
-1.01325047e+00 5.91932714e-01 -9.79642332e-01 -1.26968279e-01
1.10546529e-01 1.86212122e-01 -1.02626383e-01 1.64418846e-01
-1.86903501e+00 1.01072264e+00 7.76312470e-01 1.43505067e-01
-6.91522717e-01 -1.77496195e-01 -1.34271836e+00 1.76613703e-01
6.38835073e-01 -5.02017260e-01 1.90156555e+00 -8.66755664e-01
-1.82884717e+00 7.89646745e-01 -7.03779519e-01 -9.60740745e-01
-2.44419962e-01 -1.69812322e-01 -6.70249283e-01 -1.46037906e-01
1.73320591e-01 4.17871445e-01 4.35039669e-01 -5.29096842e-01
-7.51403213e-01 -9.90914553e-02 2.20632151e-01 -8.49458650e-02
-4.94501293e-02 1.67687044e-01 -7.54783750e-01 -5.43527305e-01
2.34380364e-01 -7.36462593e-01 -6.33479595e-01 -5.07814050e-01
-5.01249731e-01 -6.61651909e-01 3.65280300e-01 -5.05299389e-01
1.93753517e+00 -2.07130814e+00 -4.35865335e-02 3.08532149e-01
-2.82880366e-01 6.14050150e-01 -1.47868142e-01 4.18525845e-01
-1.20713808e-01 1.31725729e-01 -2.52271265e-01 -4.40998644e-01
1.98932245e-01 9.49110150e-01 -7.27050900e-01 1.56009598e-02
3.62985432e-01 1.37759769e+00 -7.73135722e-01 -4.54293221e-01
3.15618813e-02 3.37414950e-01 -2.60240406e-01 4.04480219e-01
-6.77331865e-01 -2.74118394e-01 -5.55455267e-01 6.68328643e-01
1.64952040e-01 -3.52725953e-01 4.25042093e-01 1.36281043e-01
-1.73455384e-02 8.67326796e-01 -8.46809626e-01 1.82467628e+00
-5.00611842e-01 4.54119712e-01 -1.39844626e-01 -1.56729937e+00
1.09322011e+00 4.19369578e-01 -1.98086705e-02 -9.53334332e-01
2.03839928e-01 1.59473434e-01 -7.15935409e-01 -4.04651672e-01
6.83911085e-01 -7.78033614e-01 -4.74121392e-01 3.86746436e-01
3.53310704e-01 5.72388582e-02 4.06689793e-01 1.26295492e-01
9.28431213e-01 1.41659394e-01 5.52454829e-01 1.89715296e-01
3.20174873e-01 -1.03214998e-02 8.03915918e-01 9.82667983e-01
1.20121792e-01 -1.26468847e-02 5.98016620e-01 -6.69120193e-01
-1.02953553e+00 -7.38583088e-01 2.86591407e-02 1.49497986e+00
-1.03010610e-01 -8.13960969e-01 -5.91440439e-01 -2.99620122e-01
-2.82942563e-01 1.03350055e+00 -5.84881783e-01 -4.37064230e-01
-4.40772653e-01 -3.67634386e-01 1.02647626e+00 8.16455483e-01
3.73111784e-01 -1.31331217e+00 -7.55901992e-01 4.64777917e-01
-5.86568639e-02 -1.05696857e+00 -5.96073978e-02 9.13616121e-01
-8.56033087e-01 -3.84642303e-01 -3.32246542e-01 -1.05250144e+00
3.26307654e-01 -7.30726048e-02 1.34417927e+00 5.76450676e-02
-2.98984889e-02 -3.20556492e-01 -4.54817683e-01 -1.36673935e-02
-3.84710640e-01 7.95575231e-02 -9.40176751e-03 -3.74088913e-01
6.53464437e-01 -7.70303547e-01 1.72077730e-01 2.40609217e-02
-1.04354560e+00 9.33399796e-02 6.20241284e-01 1.15922403e+00
9.99575973e-01 -2.14177761e-02 4.39807832e-01 -1.23387837e+00
8.29043388e-01 -5.15677214e-01 -5.09757221e-01 4.08936411e-01
-7.11132467e-01 4.65658486e-01 5.10060310e-01 -2.14152575e-01
-1.43435800e+00 2.03766391e-01 -5.14052987e-01 -5.22867382e-01
-3.96829322e-02 1.11378896e+00 2.94469506e-01 2.65526891e-01
7.91512966e-01 4.27715927e-01 -7.74332508e-02 -4.40905452e-01
7.08928227e-01 5.16646266e-01 7.26489067e-01 -7.19711900e-01
3.84147495e-01 1.35903388e-01 -3.27567399e-01 -5.51020265e-01
-1.09095812e+00 -1.19466946e-01 -1.75967231e-01 2.87488431e-01
6.28806770e-01 -7.68205762e-01 -3.90884280e-01 1.29787937e-01
-1.30980551e+00 -4.99551386e-01 -7.16426849e-01 7.63830543e-02
-6.92875862e-01 2.97044039e-01 -1.17265785e+00 -1.22775900e+00
-3.62175286e-01 -7.06101596e-01 7.22183943e-01 5.10811865e-01
-3.50556791e-01 -7.84952343e-01 2.71132022e-01 -3.00634533e-01
6.15752757e-01 -1.57403737e-01 9.50344920e-01 -9.14666712e-01
-2.42960870e-01 -4.61056620e-01 -1.27146408e-01 1.97149858e-01
-3.40920985e-02 -4.44246620e-01 -7.61149406e-01 2.22146228e-01
-2.27265865e-01 -6.42294407e-01 8.40372205e-01 3.84452075e-01
1.07366204e+00 -3.63714457e-01 -4.02260125e-01 6.42178953e-01
1.18893647e+00 5.56412995e-01 8.28584254e-01 9.10144020e-03
2.33004063e-01 3.58729988e-01 6.78554893e-01 3.56869578e-01
3.07631075e-01 5.96758187e-01 -8.35053250e-02 -9.18926224e-02
2.95189440e-01 -7.29971409e-01 3.64318132e-01 8.28736663e-01
1.40446484e-01 1.83047742e-01 -9.67379570e-01 4.86763924e-01
-2.26226425e+00 -9.86204445e-01 5.25359452e-01 1.85841155e+00
1.14803958e+00 2.98193038e-01 -2.19804365e-02 -1.37715623e-01
8.50666165e-01 2.90358931e-01 -6.61407411e-01 -1.06123745e+00
-1.42350942e-01 6.57494307e-01 3.00745934e-01 5.40861785e-01
-9.52966988e-01 1.49340081e+00 7.78546476e+00 1.15237558e+00
-8.57882380e-01 1.88984543e-01 7.50115275e-01 -1.81359455e-01
-3.26732159e-01 -5.45846939e-04 -8.48675430e-01 1.48540467e-01
1.65997636e+00 -3.67663115e-01 6.63946033e-01 8.58483970e-01
-2.14588419e-01 6.60732836e-02 -7.85536289e-01 9.28336382e-01
-1.81226298e-01 -1.64377105e+00 9.66616049e-02 -5.41896760e-01
4.53846008e-01 -8.37247632e-03 -3.80979717e-01 9.72322285e-01
7.30247915e-01 -1.29817450e+00 5.74471176e-01 5.05047798e-01
8.23744714e-01 -8.46761584e-01 5.95027626e-01 3.78763348e-01
-1.37918317e+00 -9.09842104e-02 -4.31945324e-01 -4.75692749e-01
7.87425876e-01 4.85396475e-01 -3.45772415e-01 3.81392270e-01
3.03798884e-01 4.70513731e-01 1.74085781e-01 4.66422528e-01
-4.47634786e-01 5.02354622e-01 -3.59412283e-01 -8.44856501e-02
5.59712529e-01 -1.69241101e-01 -3.62052396e-02 1.37319863e+00
2.35581979e-01 6.07875824e-01 3.55937302e-01 1.02227414e+00
6.77960962e-02 6.45575002e-02 -7.03743458e-01 -4.64129120e-01
8.12311411e-01 7.91546881e-01 -5.36633492e-01 -8.43822122e-01
-5.98417878e-01 1.02707696e+00 6.94833875e-01 5.08886516e-01
-8.10090423e-01 -6.58936620e-01 6.01230621e-01 -3.28506976e-01
4.13224846e-01 1.51178718e-01 -2.85415083e-01 -1.09827173e+00
-1.15149796e-01 -5.93760967e-01 6.09138727e-01 -8.03553522e-01
-7.19067693e-01 8.19799185e-01 -3.81722227e-02 -6.38841748e-01
-1.06823456e+00 -3.44956189e-01 -6.06866062e-01 9.95990694e-01
-1.38699079e+00 -8.29959154e-01 3.33947420e-01 4.22748595e-01
6.44516826e-01 -7.71827698e-02 1.32193410e+00 1.93653539e-01
-8.06092918e-01 6.36198342e-01 -1.56267270e-01 2.28341773e-01
-2.63504237e-01 -6.46464467e-01 7.77565300e-01 7.97451079e-01
2.20515370e-01 8.88115942e-01 6.77070498e-01 -4.73698318e-01
-1.13494730e+00 -1.11588824e+00 1.26504993e+00 3.37021798e-01
7.35793948e-01 -8.38026702e-02 -1.26734364e+00 1.12465155e+00
-6.62302971e-03 1.83436051e-01 8.44538152e-01 4.79872942e-01
-4.15667087e-01 4.47726160e-01 -8.51891339e-01 5.36203504e-01
1.18444610e+00 -1.24169660e+00 -9.98748362e-01 4.33700234e-02
1.01217747e+00 -4.40189421e-01 -6.68266177e-01 5.57080626e-01
4.33592856e-01 -7.27441251e-01 8.32560658e-01 -1.14251220e+00
3.89531881e-01 -1.84937287e-02 -2.12017849e-01 -1.04730260e+00
-2.21018448e-01 -6.94977105e-01 -6.22654021e-01 9.02116179e-01
7.49109507e-01 -4.48676854e-01 9.28984463e-01 7.45866001e-01
-1.96745172e-01 -1.07425821e+00 -1.07008076e+00 -7.95011997e-01
-1.42658174e-01 -7.49085307e-01 5.94005585e-01 8.62601757e-01
7.50907779e-01 6.80437863e-01 -6.40486300e-01 -4.23828065e-01
-1.11703314e-01 4.20391589e-01 1.62593916e-01 -8.34851265e-01
-4.26626056e-01 -3.37037653e-01 -2.20052078e-01 -1.50656736e+00
9.36758220e-01 -9.03121710e-01 3.36773574e-01 -1.34442472e+00
-3.03650983e-02 -2.69994766e-01 -4.64289755e-01 9.48048413e-01
7.66374841e-02 -5.98850511e-02 -1.81036279e-01 -8.36121365e-02
-8.72422397e-01 6.42747223e-01 3.99519771e-01 -1.35207579e-01
-4.50628668e-01 -1.79629132e-01 -5.12904167e-01 4.59629983e-01
6.61206365e-01 -1.38695315e-01 -3.64035279e-01 -2.33046949e-01
4.40039933e-01 4.92807060e-01 -1.43701926e-01 -5.35434008e-01
4.35660899e-01 -9.57396179e-02 -3.28275293e-01 -3.18598330e-01
3.97785574e-01 -3.21477354e-01 3.85659546e-01 3.73037159e-01
-6.20134234e-01 2.47173831e-02 5.45932889e-01 4.29397374e-01
-5.82766950e-01 -4.91307288e-01 5.37459254e-01 -3.45819801e-01
-1.09097755e+00 8.39373916e-02 -8.33616853e-01 -1.07087046e-01
6.29352927e-01 -3.11231259e-02 1.55049225e-03 -4.89948064e-01
-1.24749827e+00 1.45276770e-01 -4.90507782e-02 3.82332742e-01
6.21234417e-01 -1.66216648e+00 -7.89282694e-02 4.67842519e-01
1.33603334e-01 -6.11314848e-02 4.53625433e-02 5.26937485e-01
-2.68933252e-02 8.03280830e-01 -1.12411603e-02 -9.14271921e-02
-9.08679187e-01 7.11357832e-01 5.32945514e-01 -4.59285736e-01
-5.20247042e-01 9.08793628e-01 -2.81895310e-01 -3.93258810e-01
5.72172940e-01 -2.83278286e-01 -9.63966101e-02 9.13837925e-02
8.31285775e-01 -3.10479803e-03 -1.15097508e-01 -5.37747920e-01
-6.69635907e-02 8.02329704e-02 -8.93307105e-02 -8.62495378e-02
1.19867444e+00 7.09659830e-02 -3.52018237e-01 7.25250185e-01
1.10855889e+00 -7.39739180e-01 -7.54706323e-01 -9.55197871e-01
3.77423406e-01 -7.37998858e-02 9.89211351e-02 -4.56125826e-01
-9.19002473e-01 6.49777830e-01 2.46177062e-01 5.39633393e-01
1.09598637e+00 3.36736888e-01 1.01847649e+00 7.35464215e-01
3.00379306e-01 -1.28053164e+00 -4.14640456e-01 1.09992445e+00
2.23853886e-01 -6.95727050e-01 -5.67115188e-01 -3.52782041e-01
-3.67880613e-01 8.80949914e-01 3.39613587e-01 -4.58013080e-02
4.83063966e-01 4.78154659e-01 -2.02260055e-02 -2.85817802e-01
-1.12193990e+00 -5.44706702e-01 -1.23324968e-01 2.39765331e-01
4.33538198e-01 1.51329741e-01 -4.63275433e-01 1.09530842e+00
-2.15687249e-02 5.44565380e-01 -9.73933488e-02 1.19414461e+00
-7.27969110e-01 -1.13718081e+00 1.62006482e-01 5.51913381e-01
-5.12350678e-01 -3.56102496e-01 -3.27880085e-02 2.21239880e-01
-2.80405045e-01 5.83317757e-01 2.82824099e-01 -6.34851158e-01
5.79417884e-01 7.76102006e-01 4.66803424e-02 -8.22878420e-01
-3.22912693e-01 -1.40121788e-01 5.43963790e-01 -8.61205935e-01
-3.26622695e-01 -2.48492986e-01 -1.66440332e+00 -2.71376491e-01
-4.07975733e-01 5.75722098e-01 2.77178526e-01 1.12954032e+00
5.34391880e-01 5.98152101e-01 3.32861185e-01 -7.30273545e-01
-6.70460224e-01 -7.17569172e-01 -5.80073833e-01 2.06298724e-01
1.98771849e-01 -6.72687054e-01 -9.07670930e-02 -9.02979672e-02] | [10.780698776245117, 6.937510013580322] |
b3f1a0e2-cbfc-4b13-bf03-5eb8ea67a616 | learning-facial-representations-from-the | 2108.03427 | null | https://arxiv.org/abs/2108.03427v1 | https://arxiv.org/pdf/2108.03427v1.pdf | Learning Facial Representations from the Cycle-consistency of Face | Faces manifest large variations in many aspects, such as identity, expression, pose, and face styling. Therefore, it is a great challenge to disentangle and extract these characteristics from facial images, especially in an unsupervised manner. In this work, we introduce cycle-consistency in facial characteristics as free supervisory signal to learn facial representations from unlabeled facial images. The learning is realized by superimposing the facial motion cycle-consistency and identity cycle-consistency constraints. The main idea of the facial motion cycle-consistency is that, given a face with expression, we can perform de-expression to a neutral face via the removal of facial motion and further perform re-expression to reconstruct back to the original face. The main idea of the identity cycle-consistency is to exploit both de-identity into mean face by depriving the given neutral face of its identity via feature re-normalization and re-identity into neutral face by adding the personal attributes to the mean face. At training time, our model learns to disentangle two distinct facial representations to be useful for performing cycle-consistent face reconstruction. At test time, we use the linear protocol scheme for evaluating facial representations on various tasks, including facial expression recognition and head pose regression. We also can directly apply the learnt facial representations to person recognition, frontalization and image-to-image translation. Our experiments show that the results of our approach is competitive with those of existing methods, demonstrating the rich and unique information embedded in the disentangled representations. Code is available at https://github.com/JiaRenChang/FaceCycle . | ['Wei-Chen Chiu', 'Yong-Sheng Chen', 'Jia-Ren Chang'] | 2021-08-07 | null | http://openaccess.thecvf.com//content/ICCV2021/html/Chang_Learning_Facial_Representations_From_the_Cycle-Consistency_of_Face_ICCV_2021_paper.html | http://openaccess.thecvf.com//content/ICCV2021/papers/Chang_Learning_Facial_Representations_From_the_Cycle-Consistency_of_Face_ICCV_2021_paper.pdf | iccv-2021-1 | ['person-recognition', 'face-reconstruction'] | ['computer-vision', 'computer-vision'] | [ 8.51346999e-02 9.09735486e-02 -2.90292025e-01 -7.94666708e-01
-4.66458827e-01 -6.45988524e-01 6.50907278e-01 -8.49474072e-01
-9.46517065e-02 5.02085805e-01 2.59257048e-01 3.39365721e-01
1.42959699e-01 -3.04901212e-01 -6.67050958e-01 -1.14090276e+00
9.38375369e-02 1.84289187e-01 -7.22684860e-01 -1.31881684e-01
-2.22332552e-01 8.91909182e-01 -1.68354309e+00 1.57815993e-01
2.62801230e-01 9.62911606e-01 -3.61705810e-01 2.82660544e-01
2.39748746e-01 5.47443628e-01 -3.25143367e-01 -4.70727593e-01
3.60958040e-01 -5.40284455e-01 -5.99546075e-01 5.04991531e-01
6.23553395e-01 -5.08865356e-01 -2.45801896e-01 1.13419819e+00
4.36481029e-01 7.98824616e-03 6.94475591e-01 -1.50360501e+00
-6.38349891e-01 -3.39559801e-02 -8.73298943e-01 -2.45031208e-01
4.39001679e-01 3.62467277e-03 7.76418090e-01 -1.16535258e+00
6.29352450e-01 1.51067412e+00 3.56199950e-01 1.07346463e+00
-1.34429240e+00 -1.17049873e+00 1.65070057e-01 3.15118246e-02
-1.64490902e+00 -1.09193838e+00 9.66575861e-01 -4.12386119e-01
2.33644843e-01 3.35641831e-01 4.74308461e-01 1.37899470e+00
-1.91731140e-01 6.42011642e-01 1.18086553e+00 -3.24276537e-01
-1.29591510e-01 8.87043104e-02 -2.00660735e-01 9.32632983e-01
5.44749107e-03 2.60535598e-01 -6.64094031e-01 -3.64966318e-02
7.73955882e-01 7.81425238e-02 -4.22110170e-01 -3.35526884e-01
-7.95612156e-01 6.72225475e-01 1.96670488e-01 2.27270108e-02
-1.91202506e-01 7.53232613e-02 1.93275064e-01 2.55256772e-01
4.46721226e-01 -2.08582785e-02 -2.57120907e-01 1.50158674e-01
-7.07150996e-01 2.24561661e-01 5.45761943e-01 8.56485009e-01
9.87455964e-01 2.83676654e-01 -6.06000461e-02 8.09212804e-01
4.47884560e-01 7.69155800e-01 6.02876306e-01 -1.24015164e+00
-3.10172848e-02 3.07887584e-01 -1.84837580e-01 -1.06954455e+00
-2.32765734e-01 1.98111515e-02 -9.99783099e-01 5.41042507e-01
2.38751620e-01 -1.23764932e-01 -9.08090889e-01 2.45617008e+00
4.22462910e-01 3.34758788e-01 4.02127206e-03 9.44986999e-01
7.58197427e-01 2.83181876e-01 1.45960960e-03 -3.97327870e-01
1.47256577e+00 -6.66639805e-01 -9.25797582e-01 -8.73437524e-02
2.86418706e-01 -6.44346356e-01 7.67382801e-01 1.89747781e-01
-9.57646728e-01 -5.26075125e-01 -9.43678319e-01 -1.27700418e-02
1.08423524e-01 3.35114509e-01 5.40925205e-01 7.53980517e-01
-9.14579809e-01 4.13698047e-01 -8.01496089e-01 -1.70771569e-01
5.49256623e-01 6.12845123e-01 -1.18020880e+00 1.07482955e-01
-9.48980093e-01 7.04631627e-01 -3.03709004e-02 2.38884941e-01
-8.98613155e-01 -5.29849172e-01 -1.15324509e+00 -1.92015156e-01
1.63422793e-01 -6.15375161e-01 1.01817572e+00 -1.60915160e+00
-1.86730838e+00 1.24740255e+00 -5.41458070e-01 2.31908739e-01
4.57564682e-01 -1.17392763e-01 -4.18107450e-01 2.67666206e-02
5.43220453e-02 7.01512218e-01 1.42642462e+00 -1.26443052e+00
-9.74985212e-02 -8.62691045e-01 -2.82351911e-01 2.45078728e-01
-3.28079849e-01 1.47225589e-01 -6.80836678e-01 -6.84383273e-01
1.77857846e-01 -1.33165753e+00 2.83493251e-01 4.04413283e-01
-2.98668295e-01 1.46997973e-01 8.99389386e-01 -7.46412218e-01
7.21021891e-01 -2.42021871e+00 3.93593788e-01 2.89479345e-01
1.83444381e-01 -3.56775224e-02 -4.03993726e-01 -1.55953273e-01
-6.11589372e-01 -6.58566281e-02 -5.21605276e-02 -8.86244833e-01
-2.37974569e-01 3.09547871e-01 -1.94158956e-01 8.74008119e-01
4.52470899e-01 8.18410754e-01 -5.09725451e-01 -3.37233871e-01
-1.54896811e-01 7.35189319e-01 -6.03398383e-01 3.57318699e-01
1.28816217e-01 1.02803183e+00 -2.98459798e-01 7.37139225e-01
8.56672525e-01 1.17936850e-01 4.60640728e-01 -3.95147353e-01
3.31946254e-01 -4.35635522e-02 -1.17130148e+00 1.69689524e+00
-2.89212853e-01 4.80128139e-01 3.66515487e-01 -8.39668572e-01
9.03662503e-01 5.27243197e-01 4.84334737e-01 -6.12728119e-01
3.37622970e-01 3.92201245e-02 -9.97844115e-02 -4.48244035e-01
-2.44081132e-02 -4.92621928e-01 1.93951935e-01 5.72461843e-01
2.13594094e-01 7.94706121e-02 -2.58347750e-01 -2.79400289e-01
3.49204719e-01 3.67119193e-01 1.60394922e-01 -1.36830971e-01
7.42615402e-01 -8.72673273e-01 6.99491441e-01 -1.25071257e-01
-1.78869560e-01 7.23225832e-01 4.99534994e-01 -2.64772385e-01
-7.25732565e-01 -1.07086921e+00 2.82966010e-02 1.01422822e+00
-2.24664554e-01 -1.16723962e-01 -8.98583591e-01 -6.04862273e-01
9.36614070e-03 2.72409976e-01 -9.33350623e-01 -3.34209055e-01
-5.78621805e-01 -4.51796979e-01 7.28323042e-01 4.96103525e-01
3.03534746e-01 -8.55509937e-01 -6.02804013e-02 -3.88158828e-01
-2.28134915e-01 -1.11990130e+00 -7.97207952e-01 -3.18216681e-01
-6.24441087e-01 -8.75391841e-01 -6.53685868e-01 -6.11876845e-01
1.00599265e+00 2.11704165e-01 5.64307511e-01 1.25016317e-01
-2.46875674e-01 2.99847305e-01 5.49928993e-02 -2.66142160e-01
-2.97416419e-01 -3.61980915e-01 5.39109468e-01 7.62782395e-01
8.45869407e-02 -7.97975123e-01 -4.91240174e-01 4.47957098e-01
-8.46822083e-01 -3.68005373e-02 2.55532563e-01 8.48060250e-01
5.51255047e-01 -2.15531737e-01 4.11508083e-01 -7.92973101e-01
3.01410943e-01 -2.67675996e-01 -4.38156992e-01 7.36875534e-02
-4.68429625e-01 1.66095614e-01 2.72798270e-01 -6.32849991e-01
-1.12191212e+00 3.69056523e-01 -1.23736285e-01 -9.65672374e-01
-1.26453623e-01 6.46174476e-02 -7.56376922e-01 -2.81651262e-02
4.71169442e-01 1.28946692e-01 7.16863811e-01 -3.51302385e-01
5.63418090e-01 4.49796021e-01 7.28092492e-01 -7.07041025e-01
1.00092852e+00 8.34185123e-01 1.19390786e-01 -7.86471725e-01
-7.07213700e-01 -1.67772770e-01 -8.46432984e-01 -1.80173442e-01
8.02130640e-01 -1.11182928e+00 -9.30160880e-01 5.35261631e-01
-1.10856497e+00 4.87189442e-02 -9.43590105e-02 4.38733965e-01
-6.11562610e-01 3.98358881e-01 -2.77058572e-01 -6.80741131e-01
-2.91767567e-01 -1.06028342e+00 1.17873991e+00 1.49005905e-01
-4.00968909e-01 -7.88290799e-01 3.75193767e-02 3.84970516e-01
1.78862065e-02 4.53989983e-01 6.96366429e-01 -2.92681873e-01
-2.93268055e-01 -2.44178936e-01 5.28297238e-02 5.35363376e-01
5.82811952e-01 9.34415087e-02 -1.36635399e+00 -4.30502892e-01
1.66611135e-01 -4.10614103e-01 5.62098444e-01 1.15985230e-01
9.50601757e-01 -5.00263810e-01 -8.11986998e-02 1.10437167e+00
8.88987243e-01 -5.48161827e-02 5.78529179e-01 2.43384242e-02
7.93879390e-01 9.50577140e-01 2.09994242e-01 3.78508568e-01
2.00768128e-01 9.23134267e-01 2.41141796e-01 -1.05729073e-01
-1.41050875e-01 -3.54821682e-01 6.63629532e-01 5.27311087e-01
-4.18675721e-01 2.48819068e-01 -3.78160983e-01 1.35735378e-01
-1.60047877e+00 -1.08945858e+00 2.94304430e-01 2.20541453e+00
9.13635612e-01 -4.99567449e-01 1.13111511e-01 5.41186742e-02
6.35333061e-01 1.66705802e-01 -5.82276881e-01 -2.34526932e-01
-2.11818188e-01 2.09885642e-01 -1.55235771e-02 5.67471445e-01
-9.75642800e-01 9.29780483e-01 5.66575480e+00 3.84740055e-01
-1.42734492e+00 1.31178364e-01 6.83488905e-01 -2.62507856e-01
-2.63961494e-01 -8.35892856e-02 -7.19888031e-01 1.35369688e-01
6.28716946e-01 -2.05072761e-01 5.80790460e-01 7.55054891e-01
1.83786631e-01 3.88659716e-01 -1.39026582e+00 1.25055885e+00
5.04100919e-01 -8.69395912e-01 2.18096092e-01 1.16167851e-01
5.13150930e-01 -5.85058093e-01 4.16431576e-01 1.21350080e-01
-1.49274021e-01 -1.42196286e+00 8.78697813e-01 5.32982290e-01
1.20583797e+00 -6.20533407e-01 3.12764257e-01 1.74825549e-01
-1.06081259e+00 1.99258670e-01 -1.92879271e-02 5.11973314e-02
-4.04885337e-02 4.41598967e-02 -5.57326734e-01 4.95069146e-01
5.15721738e-01 6.70889556e-01 -2.88976461e-01 2.87169497e-02
-4.98561233e-01 1.18274465e-01 -1.06114641e-01 5.52925467e-01
-3.56943667e-01 -4.85341936e-01 4.40162301e-01 9.06170726e-01
2.89204508e-01 1.84141159e-01 -1.77668110e-01 9.95176077e-01
-3.49834561e-01 7.57678971e-02 -6.62018597e-01 9.90759805e-02
2.38426045e-01 1.33935189e+00 -1.50690332e-01 -4.45556641e-02
-3.91472816e-01 1.37479293e+00 3.18418771e-01 4.53062415e-01
-8.08147430e-01 2.01183572e-01 1.29469967e+00 4.27281745e-02
5.14685921e-02 -1.44720703e-01 -4.03498746e-02 -1.23521316e+00
1.98609009e-01 -1.13261294e+00 8.99684057e-02 -6.62562132e-01
-1.12954557e+00 7.82706320e-01 1.16169907e-01 -1.10971963e+00
-5.18394470e-01 -5.86864769e-01 -6.08729005e-01 1.07149470e+00
-1.37245524e+00 -1.52626634e+00 -4.61103112e-01 1.02833831e+00
2.56450504e-01 -2.43278012e-01 1.05949354e+00 2.79080570e-01
-7.71791697e-01 1.07878995e+00 -2.53655016e-01 4.15633649e-01
8.99871707e-01 -7.05873489e-01 4.03032498e-03 5.88763714e-01
3.43757451e-01 8.76257658e-01 4.37331587e-01 -3.52195352e-01
-1.47760785e+00 -1.01466787e+00 5.55044711e-01 -4.78588372e-01
3.52815479e-01 -5.29781461e-01 -9.23855543e-01 9.39189076e-01
-9.62221622e-02 3.89200419e-01 8.71500134e-01 1.79458018e-02
-8.39529216e-01 -3.06611568e-01 -9.88799274e-01 7.14558780e-01
1.02110136e+00 -8.40038359e-01 -2.99969405e-01 1.52172968e-01
2.86498159e-01 -3.17097396e-01 -6.39766634e-01 4.13991928e-01
1.01744223e+00 -8.79452825e-01 8.42673838e-01 -7.56512702e-01
1.49089992e-01 -2.68907994e-01 -3.07880133e-01 -1.06060338e+00
-2.42785335e-01 -8.29199851e-01 5.65589927e-02 1.42945325e+00
2.36496776e-01 -6.51912510e-01 8.62644732e-01 7.22688615e-01
4.67770308e-01 -6.16883099e-01 -1.01020038e+00 -6.90680683e-01
1.14583738e-01 -1.28623649e-01 7.64380395e-01 1.14005375e+00
-2.48519093e-01 3.24806899e-01 -6.40005469e-01 3.48121911e-01
5.33072293e-01 1.27033055e-01 1.06216681e+00 -1.04548514e+00
-2.37671793e-01 -2.76186168e-01 -5.32303214e-01 -7.21120536e-01
8.89927804e-01 -1.05198789e+00 -1.79410189e-01 -5.56807578e-01
3.85975093e-01 1.80626661e-02 -1.42140061e-01 7.38166511e-01
-2.10919548e-02 3.42473656e-01 2.32220113e-01 3.44066650e-01
1.14911079e-01 8.23134482e-01 1.37932348e+00 -8.37558806e-02
-6.25354722e-02 8.57896134e-02 -8.13028157e-01 8.25550675e-01
6.39801502e-01 -4.32467401e-01 -4.37680364e-01 -3.08206528e-01
-2.95392454e-01 1.00821704e-01 3.72684687e-01 -5.24290681e-01
-1.74829178e-02 -2.03058675e-01 6.53964937e-01 1.81289360e-01
7.24291623e-01 -7.68727422e-01 3.99685919e-01 2.95337647e-01
-1.21941276e-01 1.58142224e-01 2.97636569e-01 3.74248743e-01
-2.73352057e-01 1.18510844e-02 1.04191959e+00 8.48222598e-02
-4.38224494e-01 6.41974032e-01 2.05748510e-02 -1.67067200e-01
1.07250750e+00 -2.21161708e-01 6.94807619e-02 -7.01011896e-01
-9.02623594e-01 -2.44784772e-01 5.57435215e-01 6.32143617e-01
6.51723623e-01 -1.60188293e+00 -8.45808268e-01 9.63753819e-01
1.35603830e-01 -3.31199259e-01 1.12074457e-01 9.16281879e-01
-1.06235303e-01 3.71050462e-02 -5.34382880e-01 -6.41654789e-01
-1.71118772e+00 4.30216044e-01 5.13624847e-01 3.30774724e-01
-3.22967827e-01 7.56173551e-01 6.59553945e-01 -3.53884190e-01
7.57994056e-02 1.66127905e-01 -1.53746679e-01 1.35932475e-01
5.67301571e-01 -5.50287515e-02 -8.96538645e-02 -1.32690060e+00
-4.48608011e-01 9.92147267e-01 -4.89024930e-02 -3.61307144e-01
1.19109046e+00 -6.16763271e-02 -3.55032325e-01 1.91418573e-01
1.66649258e+00 1.46935508e-01 -1.25778365e+00 -1.72730476e-01
-3.64640146e-01 -6.16541803e-01 -1.73527300e-01 -3.63828450e-01
-1.43205261e+00 8.29089463e-01 7.79397666e-01 -7.03596532e-01
1.30583322e+00 4.65807393e-02 1.28490835e-01 1.60696000e-01
2.48734459e-01 -6.57581806e-01 2.71067619e-01 2.41670519e-01
1.24257278e+00 -1.16831517e+00 3.60444933e-02 -4.83162463e-01
-7.10996330e-01 1.00424111e+00 5.62701821e-01 3.30684781e-02
8.84581268e-01 1.93214759e-01 2.85551459e-01 -9.22230855e-02
-5.17138481e-01 -8.24254937e-03 5.43872118e-01 6.18097305e-01
4.51692104e-01 8.20280835e-02 2.01700449e-01 6.52314007e-01
-3.79656613e-01 -1.15375079e-01 1.00214817e-01 5.66590071e-01
1.53598756e-01 -1.27780282e+00 -4.15707320e-01 -1.89502254e-01
-4.06224310e-01 2.30655476e-01 -5.14176667e-01 8.74276340e-01
1.56574756e-01 6.37236297e-01 8.26593265e-02 -3.72015476e-01
3.83739322e-01 2.96140045e-01 8.04744005e-01 -5.88729501e-01
-6.97699040e-02 1.35769591e-01 -1.52232021e-01 -7.41196692e-01
-4.88661587e-01 -7.75206268e-01 -1.26211250e+00 -3.94951791e-01
-1.76407859e-01 -2.44546428e-01 6.19850636e-01 7.82716334e-01
3.29813540e-01 5.91596849e-02 1.02415538e+00 -1.07522404e+00
-6.68761909e-01 -8.09305310e-01 -6.96988344e-01 9.35066044e-01
5.48009038e-01 -8.75719309e-01 -5.14571667e-01 4.02716756e-01] | [13.024592399597168, 0.24281394481658936] |
a125cc63-18c3-4b5b-9ff8-adad6eebe7a6 | mean-reversion-and-optimization | 1408.2217 | null | http://arxiv.org/abs/1408.2217v3 | http://arxiv.org/pdf/1408.2217v3.pdf | Mean-Reversion and Optimization | The purpose of these notes is to provide a systematic quantitative framework
- in what is intended to be a "pedagogical" fashion - for discussing
mean-reversion and optimization. We start with pair trading and add complexity
by following the sequence "mean-reversion via demeaning -> regression ->
weighted regression -> (constrained) optimization -> factor models". We discuss
in detail how to do mean-reversion based on this approach, including common
pitfalls encountered in practical applications, such as the difference between
maximizing the Sharpe ratio and minimizing an objective function when trading
costs are included. We also discuss explicit algorithms for optimization with
linear costs, constraints and bounds. | [] | 2016-02-12 | null | null | null | null | ['pair-trading'] | ['time-series'] | [-9.94567201e-02 2.03224495e-01 -2.79482722e-01 -3.39603573e-01
-9.05718863e-01 -7.36723602e-01 4.83037114e-01 -2.25056648e-01
-4.52597171e-01 9.49340045e-01 1.79258913e-01 -9.77109253e-01
-5.89849055e-01 -2.73241103e-01 -5.92149079e-01 -7.36621201e-01
-2.31433973e-01 3.38794202e-01 -2.67571121e-01 -4.64420170e-01
1.10493660e+00 7.35940099e-01 -9.45910513e-01 -2.13593647e-01
9.10737991e-01 9.36373949e-01 -2.22765505e-01 8.45603645e-01
-1.46630913e-01 1.00195086e+00 -4.11874533e-01 -8.18080783e-01
7.35592365e-01 -6.34529531e-01 -9.10648346e-01 5.13738804e-02
1.19950943e-01 -3.08078140e-01 -1.50932118e-01 9.58816409e-01
1.90166458e-01 5.27337372e-01 8.28972578e-01 -1.11669767e+00
-6.89721167e-01 6.36921287e-01 -9.25525010e-01 4.53341931e-01
-3.83405178e-03 1.29554778e-01 1.20574880e+00 -5.97476065e-01
2.70622134e-01 1.11288524e+00 5.97241759e-01 4.00552899e-01
-1.23779047e+00 -5.17386734e-01 3.12178731e-01 -1.84010386e-01
-9.15682256e-01 -4.24842149e-01 6.41932547e-01 -3.27862829e-01
1.16212046e+00 7.34185755e-01 6.76743209e-01 2.91705996e-01
6.48844719e-01 5.81670761e-01 1.12377822e+00 -3.80656272e-01
-7.22264498e-02 2.66446590e-01 6.46174327e-02 6.53900683e-01
3.32256705e-01 4.57437605e-01 -3.57070029e-01 1.05898537e-01
9.97390389e-01 -2.26927593e-01 -2.10536212e-01 -3.16345572e-01
-1.15331054e+00 9.62294877e-01 1.81297451e-01 2.22406894e-01
-1.12433426e-01 5.08881629e-01 1.86784506e-01 8.25494111e-01
6.61634684e-01 9.55752611e-01 -7.23606288e-01 -3.97100061e-01
-1.15202534e+00 5.56882381e-01 1.01115370e+00 9.45138991e-01
2.77706414e-01 4.28700954e-01 2.78852396e-02 5.96883535e-01
3.47551286e-01 3.78144681e-01 2.47621194e-01 -1.60441422e+00
7.46482790e-01 -4.43523712e-02 5.78487873e-01 -5.13140023e-01
-4.97058839e-01 -5.52401245e-01 -4.35790867e-01 7.00027347e-01
6.68901443e-01 -5.01700938e-01 -3.75812948e-01 1.46878088e+00
9.32790712e-02 -1.24825321e-01 -1.12976797e-01 6.11674726e-01
-1.48822755e-01 7.76434004e-01 -4.35256124e-01 -1.18783998e+00
8.00759435e-01 -1.20390999e+00 -8.13727081e-01 2.94184566e-01
7.29625702e-01 -9.46532309e-01 1.06783807e+00 5.55582166e-01
-1.82500756e+00 1.46040231e-01 -9.53458250e-01 1.11774318e-01
-1.17373839e-01 -4.89334226e-01 5.17214060e-01 7.83958852e-01
-1.24690080e+00 1.15273213e+00 -6.34288907e-01 1.61785871e-01
-1.56932518e-01 4.94194090e-01 1.61786214e-01 8.73384655e-01
-5.49535811e-01 1.19131041e+00 -1.33646309e-01 1.36207804e-01
-3.55766475e-01 -1.23707080e+00 -4.85060483e-01 1.68807670e-01
6.00509107e-01 -3.67596745e-01 1.81788802e+00 -1.02907133e+00
-2.10721421e+00 6.19971275e-01 7.94830546e-02 -4.16909605e-01
1.06436038e+00 -4.16467935e-01 -1.97701991e-01 -2.97502518e-01
-2.23144874e-01 2.36916706e-01 6.01114988e-01 -1.04309785e+00
-6.74206018e-01 -1.07059732e-01 -2.56117042e-02 3.93101245e-01
2.14138642e-01 4.23296601e-01 9.15659070e-02 -1.01148176e+00
1.49668707e-02 -1.12709284e+00 -4.86656129e-01 -3.14120382e-01
-2.94780761e-01 7.49913082e-02 2.99917400e-01 -6.24909461e-01
1.48726165e+00 -1.85556281e+00 -1.03027105e-01 3.28044057e-01
-6.46928698e-03 -2.40898579e-01 1.26300141e-01 4.70128149e-01
-4.41878498e-01 3.88840050e-01 -2.38801658e-01 -9.48004201e-02
1.58057377e-01 -3.36701512e-01 -7.07002819e-01 6.61157787e-01
-1.20872639e-01 8.62885773e-01 -7.64337957e-01 -2.99419880e-01
2.19452865e-02 -3.56738572e-04 -8.86701047e-01 6.69103465e-05
1.03710011e-01 3.77368391e-01 -1.86732918e-01 6.39262080e-01
5.64404666e-01 -6.19735457e-02 3.89045253e-02 1.87276870e-01
-8.34429502e-01 4.71935421e-01 -1.24331558e+00 1.04961669e+00
-4.94088769e-01 6.72967732e-01 2.46078134e-01 -8.76645148e-01
5.91958880e-01 -1.20655671e-01 6.78675652e-01 -5.87249279e-01
1.02236636e-01 5.92089295e-01 -1.67839825e-01 -2.22659066e-01
7.19680130e-01 -6.22487664e-01 -2.99039613e-02 5.12422383e-01
-3.12231183e-01 -5.25503457e-01 2.72748411e-01 3.76221463e-02
6.26427889e-01 1.04482062e-01 3.26946139e-01 -8.78705919e-01
3.89527917e-01 1.12748146e-01 6.02304995e-01 6.68905973e-01
-1.34910420e-01 4.55259323e-01 1.00062740e+00 -2.90019244e-01
-1.00160384e+00 -9.76094484e-01 -1.76544935e-01 1.10413992e+00
-7.89238587e-02 -6.76473007e-02 -5.66547871e-01 -6.55864775e-01
1.16944000e-01 1.18298852e+00 -6.69934034e-01 3.19464624e-01
-7.26429999e-01 -1.13706958e+00 -3.87286348e-03 3.58698756e-01
1.95352077e-01 -7.49652982e-01 -5.05467415e-01 2.10717887e-01
3.40278000e-01 -3.62385005e-01 -9.94318843e-01 4.29542184e-01
-1.19623280e+00 -7.09410965e-01 -9.31964278e-01 -4.64722306e-01
5.63710034e-01 1.21054158e-01 1.01760864e+00 -3.25469337e-02
-3.96570517e-03 2.36883700e-01 7.22605884e-02 -2.10685819e-01
-1.38606161e-01 -3.68023664e-01 7.50733688e-02 -3.19011658e-01
-3.23743731e-01 -4.69713748e-01 -7.29175866e-01 3.24185312e-01
-5.63160002e-01 -3.07835609e-01 4.55769390e-01 8.14600587e-01
3.81315440e-01 -1.85063362e-01 4.63686138e-01 -9.28511143e-01
1.04723060e+00 -1.90590516e-01 -1.27052915e+00 4.58245605e-01
-1.39307666e+00 3.13015491e-01 4.54373449e-01 -4.02868629e-01
-1.10742271e+00 -5.59544802e-01 9.46993753e-02 -2.89997190e-01
7.90194750e-01 1.95430592e-01 2.17561424e-01 -3.10809165e-01
3.39594036e-01 -1.27295911e-01 8.43315497e-02 -5.27326047e-01
4.57511693e-01 3.84810627e-01 3.81780356e-01 -5.94747901e-01
5.89284658e-01 1.10426299e-01 4.17836159e-01 -4.11671996e-01
-6.47670150e-01 -8.65219831e-02 -2.74994016e-01 -8.08536336e-02
5.66464484e-01 -4.32318598e-01 -9.70323205e-01 1.91922918e-01
-9.10660207e-01 -6.88094854e-01 -8.92597198e-01 6.72669530e-01
-1.10909677e+00 1.60441548e-01 -9.02484298e-01 -1.00840843e+00
-2.14212120e-01 -1.27392972e+00 1.91675678e-01 5.40257931e-01
-1.45485207e-01 -1.33169806e+00 3.63372862e-01 1.32138848e-01
5.98509729e-01 -2.64789234e-03 6.31904304e-01 -7.19716907e-01
-5.08350313e-01 1.56085536e-01 -1.37423888e-01 3.16984177e-01
-8.35704580e-02 5.67366064e-01 -3.28298986e-01 -1.49842232e-01
3.07805836e-01 1.00701220e-01 4.16876495e-01 7.24982798e-01
7.71983922e-01 -8.79175782e-01 1.16262935e-01 7.89378583e-01
1.45850885e+00 5.07215500e-01 3.68166327e-01 6.90885127e-01
2.80806392e-01 9.53411698e-01 6.77571058e-01 6.06326640e-01
2.26450041e-01 3.86249959e-01 6.32748827e-02 1.44839108e-01
4.18639749e-01 -2.48834848e-01 4.51089263e-01 1.02947342e+00
-5.42445853e-02 1.87762976e-02 -4.80968058e-01 4.37557995e-01
-1.76543033e+00 -1.09902525e+00 -1.79839358e-01 2.32862449e+00
7.79277086e-01 3.45846742e-01 4.32857633e-01 -2.16598257e-01
5.54903984e-01 8.71230811e-02 -3.75947714e-01 -1.09186351e+00
2.10411072e-01 -1.18550710e-01 1.17150164e+00 1.16694713e+00
-5.14995217e-01 7.98543930e-01 8.47891331e+00 6.73033416e-01
-1.07784498e+00 -9.29167271e-02 9.54338253e-01 -4.84617919e-01
-6.88683033e-01 5.62472284e-01 -7.29535460e-01 4.18717980e-01
8.57538462e-01 -4.51302320e-01 7.53671706e-01 5.47291756e-01
6.39404058e-01 1.50487265e-02 -7.41907060e-01 6.56368136e-01
-2.23837033e-01 -1.30405307e+00 -1.14924245e-01 1.52445078e-01
1.02117777e+00 -4.10787255e-01 3.74686569e-01 3.87598634e-01
5.81234634e-01 -9.46809649e-01 8.74564707e-01 6.73825443e-01
5.69774985e-01 -9.54761386e-01 2.20250398e-01 2.04429880e-01
-9.50591922e-01 -2.43600219e-01 -1.76631227e-01 -1.86467528e-01
3.91332597e-01 4.25597459e-01 -2.44673565e-01 3.14450979e-01
5.60174823e-01 3.50693107e-01 2.62861187e-03 1.17634034e+00
-5.09317080e-03 1.81689903e-01 -2.94628173e-01 -2.48558596e-01
4.91921872e-01 -1.02383292e+00 6.61254823e-01 1.28071761e+00
4.93332565e-01 2.93935746e-01 -3.84530336e-01 7.22254395e-01
1.70133695e-01 3.35953623e-01 -3.16089571e-01 -2.82997824e-02
1.29532635e-01 1.02092004e+00 -9.26184893e-01 -2.19318554e-01
-6.10888600e-01 4.87099856e-01 -5.61245605e-02 4.67464805e-01
-9.14984703e-01 -7.69772589e-01 5.87644815e-01 3.21379423e-01
2.02606872e-01 -2.96108395e-01 -8.44085276e-01 -1.22650766e+00
-2.75480568e-01 -7.17456162e-01 3.79826844e-01 -6.88218594e-01
-8.26433778e-01 1.51069462e-01 4.62973267e-01 -9.64386225e-01
-3.89382243e-01 -5.69059372e-01 -8.83372009e-01 6.39119029e-01
-1.23575795e+00 -4.80911523e-01 5.76160192e-01 3.68380070e-01
8.19843948e-01 -4.47005257e-02 -2.13373721e-01 1.63391426e-01
-5.89409888e-01 6.86368942e-01 6.04476154e-01 -3.85630280e-01
5.40616691e-01 -1.61268246e+00 2.51569539e-01 7.24932611e-01
-3.21692191e-02 6.96027637e-01 1.21831000e+00 -4.63478684e-01
-1.17196345e+00 -3.38098675e-01 1.02772701e+00 -3.83020818e-01
1.22210288e+00 1.93887010e-01 -6.79065347e-01 9.17432070e-01
3.79484206e-01 -4.23002601e-01 3.51296127e-01 9.10457000e-02
9.97358635e-02 -4.14520890e-01 -1.20422626e+00 9.62774992e-01
9.11058664e-01 -1.67519554e-01 -5.29383540e-01 2.68270940e-01
8.06729198e-01 -3.99680883e-01 -9.06370878e-01 3.63437891e-01
7.12032259e-01 -1.00581527e+00 7.41448700e-01 -6.52709305e-01
1.06929414e-01 1.10127062e-01 2.13996172e-02 -1.29551101e+00
-2.61777878e-01 -1.80018556e+00 6.44121543e-02 1.10359740e+00
7.89105177e-01 -1.05471826e+00 6.26560271e-01 9.40649331e-01
-1.80251926e-01 -1.00771868e+00 -9.45962071e-01 -9.85443532e-01
6.90695465e-01 -2.62583315e-01 4.36247647e-01 8.49057853e-01
5.05550742e-01 2.34712643e-04 -4.68661577e-01 -3.58783484e-01
5.93686163e-01 1.89297363e-01 5.30247808e-01 -5.44848263e-01
-4.43058640e-01 -1.19464731e+00 2.64584422e-01 -1.32949626e+00
-1.92391813e-01 -5.57579994e-01 -1.78195819e-01 -8.48653853e-01
-8.01728517e-02 -2.50120908e-01 -3.45383525e-01 -9.24945846e-02
6.56502694e-02 -2.56859034e-01 5.08193314e-01 1.95882052e-01
-3.65489066e-01 4.43793327e-01 1.39297581e+00 2.50086904e-01
-6.63928509e-01 3.35707337e-01 -1.15797055e+00 6.66008770e-01
6.06947720e-01 -6.63204670e-01 -3.58171284e-01 -4.90886904e-03
5.22614539e-01 8.46205294e-01 -8.91976506e-02 -3.75134289e-01
2.19748691e-01 -9.25948918e-01 -1.29734397e-01 -4.93309915e-01
1.35027505e-02 -6.20212734e-01 2.73871541e-01 5.57529986e-01
-6.99823081e-01 8.57750356e-01 -4.76975879e-03 1.10096015e-01
3.36961597e-02 -6.35223567e-01 1.08730280e+00 -5.86249493e-02
2.97015850e-02 -1.57650188e-01 -3.48340809e-01 2.90585756e-01
1.21718860e+00 6.18492030e-02 -2.66610086e-01 -6.98600829e-01
-6.27847373e-01 4.57491964e-01 5.80206752e-01 7.11147115e-02
1.43075407e-01 -1.04170525e+00 -5.47870159e-01 -5.51384278e-02
-6.53979480e-01 -4.91967738e-01 -4.26295884e-02 1.31478620e+00
-9.94793653e-01 5.69887161e-01 5.70626743e-02 6.40035346e-02
-8.97598803e-01 7.91761279e-01 6.93323851e-01 -6.58277214e-01
-3.29429597e-01 8.13486159e-01 1.77923158e-01 -1.26223207e-01
2.34334186e-01 -5.90416431e-01 1.93563834e-01 1.11212675e-02
3.66468966e-01 1.02469552e+00 -2.13043600e-01 -1.43570483e-01
-1.28759861e-01 7.37324238e-01 -1.27487741e-02 -8.07104170e-01
1.10124063e+00 -4.48419333e-01 -3.96290310e-02 6.58837616e-01
1.12332535e+00 5.69723427e-01 -1.59370124e+00 2.36631572e-01
2.34860688e-01 -4.05896306e-01 -9.61964279e-02 -7.26388752e-01
-1.03096247e+00 6.73086762e-01 2.11366683e-01 4.27903533e-01
9.55928147e-01 -4.97954696e-01 5.12889206e-01 4.14925188e-01
-4.53946330e-02 -1.52528179e+00 -1.50308028e-01 3.47133845e-01
9.97473478e-01 -8.78276646e-01 3.76991093e-01 7.64943957e-02
-9.35969651e-01 1.17339647e+00 2.62894511e-01 -5.79929352e-01
1.04616165e+00 2.78415382e-01 -1.23553954e-01 1.25233233e-01
-9.44889486e-01 2.22029433e-01 1.83989450e-01 -1.44496754e-01
3.27968985e-01 -1.22198679e-01 -9.13862169e-01 2.57092953e-01
-4.81651932e-01 -1.47700429e-01 6.34288192e-01 9.33977365e-01
-5.79890668e-01 -1.10657048e+00 -4.85907555e-01 4.63722318e-01
-9.35245812e-01 -1.27284020e-01 -3.17410767e-01 1.09379065e+00
-5.51297188e-01 9.64247406e-01 3.76299471e-02 -2.05771253e-01
3.06735963e-01 -1.66005939e-01 4.96559858e-01 -8.20701271e-02
-9.83282387e-01 5.69785237e-01 1.47026226e-01 -4.40151870e-01
-4.53766026e-02 -8.25453401e-01 -1.00914371e+00 -1.03256738e+00
-4.54994977e-01 4.49424922e-01 7.30996013e-01 8.78485322e-01
-2.37221733e-01 3.52628380e-01 1.00467992e+00 -5.82175136e-01
-1.36499476e+00 -5.83999455e-01 -9.52522695e-01 -2.42226347e-01
5.74219108e-01 -3.15658003e-01 -9.12618041e-01 -2.52390653e-01] | [5.25218391418457, 3.9066262245178223] |
fe9f19ca-ea29-4537-8428-a88c4cf0348f | srl-scaling-distributed-reinforcement | 2306.16688 | null | https://arxiv.org/abs/2306.16688v2 | https://arxiv.org/pdf/2306.16688v2.pdf | SRL: Scaling Distributed Reinforcement Learning to Over Ten Thousand Cores | The ever-growing complexity of reinforcement learning (RL) tasks demands a distributed RL system to efficiently generate and process a massive amount of data to train intelligent agents. However, existing open-source libraries suffer from various limitations, which impede their practical use in challenging scenarios where large-scale training is necessary. While industrial systems from OpenAI and DeepMind have achieved successful large-scale RL training, their system architecture and implementation details remain undisclosed to the community. In this paper, we present a novel abstraction on the dataflows of RL training, which unifies practical RL training across diverse applications into a general framework and enables fine-grained optimizations. Following this abstraction, we develop a scalable, efficient, and extensible distributed RL system called ReaLly Scalable RL (SRL). The system architecture of SRL separates major RL computation components and allows massively parallelized training. Moreover, SRL offers user-friendly and extensible interfaces for customized algorithms. Our evaluation shows that SRL outperforms existing academic libraries in both a single machine and a medium-sized cluster. In a large-scale cluster, the novel architecture of SRL leads to up to 3.7x speedup compared to the design choices adopted by the existing libraries. We also conduct a direct benchmark comparison to OpenAI's industrial system, Rapid, in the challenging hide-and-seek environment. SRL reproduces the same solution as reported by OpenAI with up to 5x speedup in wall-clock time. Furthermore, we also examine the performance of SRL in a much harder variant of the hide-and-seek environment and achieve substantial learning speedup by scaling SRL to over 15k CPU cores and 32 A100 GPUs. Notably, SRL is the first in the academic community to perform RL experiments at such a large scale. | ['Yi Wu', 'Huanchen Zhang', 'Guangju Wang', 'Wei Fu', 'Zhiyu Mei'] | 2023-06-29 | null | null | null | null | ['reinforcement-learning-1'] | ['methodology'] | [-3.29251647e-01 -3.42382938e-01 -2.27274030e-01 -1.81892395e-01
-9.80921924e-01 -7.13712275e-01 4.40503687e-01 -3.01153034e-01
-6.93647563e-01 8.87150586e-01 -2.19695717e-01 -6.66531563e-01
-8.68299529e-02 -8.22302759e-01 -6.84081793e-01 -8.18791211e-01
-1.91205531e-01 7.04012632e-01 1.52038798e-01 -2.68249154e-01
1.37500763e-02 6.56641543e-01 -1.79328346e+00 2.39756450e-01
6.55995488e-01 7.86270320e-01 4.77489412e-01 9.24086213e-01
1.27400635e-02 1.11580694e+00 -8.59062135e-01 4.05818049e-04
5.31434238e-01 -1.64813742e-01 -8.49846601e-01 -1.59893870e-01
4.00602996e-01 -4.79459196e-01 1.31052390e-01 4.04955983e-01
9.54634547e-01 1.53759286e-01 -1.66529655e-01 -1.35532379e+00
-6.66932762e-02 4.91981298e-01 -4.14766461e-01 -4.46217507e-02
2.11801022e-01 5.83037794e-01 1.13494301e+00 -2.38506913e-01
5.49272776e-01 1.00552034e+00 7.39168882e-01 4.87026572e-01
-1.20684528e+00 -7.76039422e-01 2.53114905e-02 -2.01537088e-02
-1.12463236e+00 -4.13431853e-01 3.18534851e-01 5.85310906e-02
1.39776945e+00 2.68850833e-01 6.72002196e-01 1.20595086e+00
1.72955915e-01 1.22265947e+00 1.30458057e+00 -2.14866579e-01
5.07785439e-01 -1.40664756e-01 -2.40783811e-01 8.34251344e-01
7.81543180e-02 2.56033361e-01 -5.35133719e-01 -2.53768712e-01
9.63230908e-01 -2.78277636e-01 1.33141160e-01 -2.54688442e-01
-1.37749434e+00 9.64669645e-01 4.50253338e-01 3.14080805e-01
-2.18469843e-01 4.20007646e-01 7.07004547e-01 5.94500899e-01
2.84506053e-01 6.68162346e-01 -8.06888819e-01 -5.90356767e-01
-8.41390431e-01 5.69498479e-01 1.14288282e+00 9.30189490e-01
8.41166735e-01 2.82650799e-01 3.65071334e-02 7.46195853e-01
1.24880716e-01 3.86540294e-01 6.49576843e-01 -1.44155407e+00
4.31361586e-01 3.37981105e-01 -8.99137780e-02 -5.71362078e-01
-6.83125854e-01 -8.20709050e-01 -7.23176181e-01 7.39271283e-01
4.18417990e-01 -4.45422441e-01 -2.45066404e-01 1.53327072e+00
5.80501854e-01 2.00208515e-01 4.17994320e-01 8.62630427e-01
6.80284560e-01 6.63484991e-01 2.97111999e-02 5.96962031e-03
1.04388416e+00 -1.51299965e+00 -1.83510944e-01 -3.53296220e-01
1.03764212e+00 -8.22905838e-01 1.37019277e+00 5.93372762e-01
-1.11705458e+00 -5.43060124e-01 -1.12885046e+00 -1.09959275e-01
-1.99157298e-01 2.20306769e-01 1.36369860e+00 5.38471699e-01
-1.25164223e+00 7.50332057e-01 -9.59048867e-01 -2.77719438e-01
2.61739373e-01 6.08418107e-01 -1.45016477e-01 -7.60629326e-02
-7.67596424e-01 7.90240228e-01 4.08421397e-01 -2.10998908e-01
-8.89858186e-01 -7.87763596e-01 -6.36944652e-01 -1.84854463e-01
5.01640797e-01 -8.08744192e-01 1.77592397e+00 -1.02093494e+00
-2.17805743e+00 5.68440437e-01 2.34209150e-01 -5.14209390e-01
5.60785592e-01 -1.73376471e-01 -9.33248363e-03 -1.26797691e-01
-4.71144915e-02 6.29173160e-01 5.65990746e-01 -9.43172812e-01
-8.43924701e-01 -1.60251737e-01 3.43661278e-01 5.00227094e-01
-1.07763179e-01 -7.77319744e-02 -4.67838347e-01 -4.06429797e-01
-4.52518761e-01 -1.05945468e+00 -4.84962225e-01 -3.18707079e-01
4.08891737e-02 -3.45733315e-01 7.86451280e-01 -2.43481509e-02
8.02218199e-01 -2.03061604e+00 -4.56688441e-02 -1.21375419e-01
3.49261403e-01 3.59871864e-01 -3.61031026e-01 4.47428554e-01
4.18255031e-01 -3.83191288e-01 2.92256791e-02 -4.43568677e-01
3.47558647e-01 6.94151402e-01 -1.46485925e-01 3.29930812e-01
-1.76331893e-01 9.86086965e-01 -1.06749451e+00 -4.44416314e-01
1.56029105e-01 2.37510055e-01 -9.38717425e-01 4.40622360e-01
-4.41657066e-01 4.51700479e-01 -6.40346944e-01 4.99186665e-01
2.82127559e-01 -4.23571020e-01 4.14415240e-01 9.80035141e-02
-3.32468808e-01 3.93340498e-01 -1.15278661e+00 2.05174899e+00
-9.14140701e-01 5.13104796e-01 3.00104797e-01 -7.69179702e-01
8.94355595e-01 1.32482469e-01 6.68267965e-01 -8.06234837e-01
2.15317030e-02 2.70540178e-01 -8.98669288e-02 -7.50687048e-02
5.26127517e-01 2.29028568e-01 -1.97894186e-01 9.67261076e-01
7.86904097e-02 -2.79598594e-01 5.56103945e-01 7.17876405e-02
1.47791159e+00 5.79613149e-01 1.16809636e-01 -2.46629253e-01
2.36654833e-01 2.04301000e-01 4.04907137e-01 9.46318030e-01
-9.29874331e-02 -1.09545156e-01 3.54473799e-01 -7.41399705e-01
-9.48351383e-01 -9.62071657e-01 -5.59141226e-02 1.87813914e+00
-2.87770778e-01 -7.22086132e-01 -6.27914369e-01 -6.57711804e-01
6.19493127e-02 5.97579718e-01 -1.58041678e-02 1.91228732e-01
-5.29370308e-01 -9.76346135e-01 9.43387032e-01 3.07717413e-01
9.61183310e-01 -1.56239009e+00 -9.68667924e-01 3.48939151e-01
1.23795144e-01 -1.17167222e+00 -5.52540086e-02 3.62878472e-01
-8.31169009e-01 -9.49960351e-01 -2.12559894e-01 -8.08396041e-01
2.48077169e-01 1.02273621e-01 1.57894552e+00 1.83542117e-01
-5.06469011e-01 3.02243799e-01 -1.61576018e-01 -1.50540099e-01
-4.72108513e-01 5.04258096e-01 -7.86300972e-02 -5.71546078e-01
6.99427649e-02 -6.25019908e-01 -4.36647803e-01 1.96438640e-01
-6.35321319e-01 3.65156472e-01 8.25352192e-01 7.01115608e-01
4.25609946e-01 3.83171998e-02 6.17667794e-01 -8.72434974e-01
7.89407253e-01 -4.13108110e-01 -9.70520079e-01 -4.13540611e-03
-8.33852887e-01 3.08429092e-01 9.37362313e-01 -3.40142280e-01
-9.70978439e-01 1.58466443e-01 -7.00843751e-01 -1.34233668e-01
-3.89313638e-01 1.98824778e-01 2.16803059e-01 -1.94570661e-01
5.87702215e-01 2.24316239e-01 1.82225406e-01 -3.24201196e-01
4.01569873e-01 5.23241162e-01 1.73334032e-01 -1.13279152e+00
5.31758964e-01 1.73652023e-02 -5.39687425e-02 -6.83150411e-01
-8.58376801e-01 -1.54604778e-01 -2.86164910e-01 -3.87251489e-02
4.60351974e-01 -9.92617249e-01 -1.38795269e+00 4.55614090e-01
-7.39378870e-01 -1.32139957e+00 -4.67253119e-01 3.33289534e-01
-9.31225657e-01 -1.30698711e-01 -1.01934910e+00 -3.91478121e-01
-6.54520869e-01 -1.56175733e+00 1.22866464e+00 7.97037259e-02
-2.59029288e-02 -1.08696938e+00 3.53192478e-01 5.38973987e-01
7.16198206e-01 4.72315289e-02 5.58333755e-01 -4.41301018e-01
-7.09193230e-01 2.00546622e-01 -5.85379377e-02 2.41239026e-01
-3.29228014e-01 3.83330844e-02 -9.57986891e-01 -5.99989831e-01
-2.73419112e-01 -1.06518292e+00 3.98059070e-01 2.94028334e-02
1.12553573e+00 -7.33094513e-02 2.24849279e-03 9.46301758e-01
1.38601327e+00 -6.20884150e-02 3.74894083e-01 7.77563989e-01
6.07189596e-01 6.01062328e-02 6.18543208e-01 6.53637469e-01
4.15698081e-01 6.26700699e-01 3.33512604e-01 -3.82382691e-01
1.73577920e-01 -1.77795184e-03 6.90232456e-01 1.08490980e+00
-2.05727369e-01 1.33698180e-01 -5.91275573e-01 -7.26988688e-02
-1.91338885e+00 -7.66184866e-01 -5.13195433e-02 2.07116342e+00
1.12764847e+00 4.22023349e-02 2.44687811e-01 -5.47826886e-02
-6.76960498e-02 2.19163373e-01 -6.85341775e-01 -7.57900417e-01
1.99527726e-01 6.26854897e-01 4.91169751e-01 4.44680780e-01
-8.53231668e-01 1.39560533e+00 6.84736395e+00 1.12714314e+00
-1.23803508e+00 2.11873114e-01 5.24931550e-01 -3.90721262e-01
2.13693932e-01 -2.32381031e-01 -1.04086268e+00 7.80937225e-02
1.28806937e+00 -1.18815735e-01 1.00203526e+00 1.24022007e+00
-7.35021010e-02 -1.38116971e-01 -9.94046450e-01 8.38751376e-01
-2.60396034e-01 -1.35415149e+00 -4.15458232e-01 1.05543025e-01
7.39960730e-01 6.90948009e-01 2.95024980e-02 9.72974718e-01
1.04914558e+00 -9.44027483e-01 3.43018770e-01 -1.43245667e-01
6.00149632e-01 -1.01755607e+00 4.66714978e-01 5.85268497e-01
-1.11998510e+00 -5.59349619e-02 -5.22289157e-01 -4.59614813e-01
-4.30267364e-01 2.45693311e-01 -7.62562335e-01 3.75244707e-01
7.71163940e-01 7.89007246e-01 -5.46757936e-01 5.66689134e-01
-3.34861577e-01 5.60812294e-01 -3.34443927e-01 -1.17221132e-01
5.93320072e-01 -2.91694969e-01 1.58392210e-02 1.14978504e+00
5.01597226e-02 -2.99207032e-01 4.67241526e-01 5.78129113e-01
-1.04522996e-01 1.69118851e-01 -5.23495257e-01 1.05044991e-01
3.09535682e-01 1.79365730e+00 -5.23999453e-01 -3.00626993e-01
-4.13148463e-01 9.43632543e-01 9.12463725e-01 1.63708270e-01
-9.90325093e-01 -1.18610486e-01 9.63435173e-01 -4.93888736e-01
1.92378402e-01 -4.65745747e-01 -4.82659824e-02 -9.45377648e-01
-3.38628173e-01 -1.34002590e+00 3.29330921e-01 -5.94282210e-01
-1.09602451e+00 7.52119720e-01 -2.81584442e-01 -8.60325336e-01
-5.69080114e-01 -6.95573986e-01 -4.29699570e-01 5.27240992e-01
-1.54485130e+00 -1.01909316e+00 -2.71269709e-01 8.97802591e-01
7.12144434e-01 -4.52027678e-01 1.23418629e+00 3.03629309e-01
-8.37554753e-01 6.45806909e-01 4.28507119e-01 -1.50209129e-01
6.20907605e-01 -1.41204822e+00 4.83077168e-01 3.00957888e-01
2.48614013e-01 4.01107818e-01 5.04322171e-01 -2.64377773e-01
-2.07562709e+00 -1.12412846e+00 4.30892974e-01 -2.61056423e-01
8.61350060e-01 -4.08155710e-01 -4.96125340e-01 9.32664037e-01
3.08559746e-01 9.92381722e-02 5.63246965e-01 5.08982480e-01
-1.40407726e-01 -3.93520117e-01 -8.87683928e-01 7.14124203e-01
9.23081160e-01 -1.84162319e-01 -2.01161996e-01 7.87282765e-01
6.13175035e-01 -6.88447237e-01 -1.16058362e+00 3.57671939e-02
4.70613062e-01 -1.24774730e+00 1.01967943e+00 -3.35908681e-01
4.18370277e-01 2.42243353e-02 9.54599306e-02 -1.30409694e+00
-1.73869029e-01 -8.17858398e-01 1.30724842e-02 8.43842387e-01
1.88234955e-01 -8.62305522e-01 8.49314749e-01 1.58520296e-01
-2.57780075e-01 -1.06689215e+00 -6.24266982e-01 -7.68808901e-01
1.98085770e-01 -7.03849673e-01 6.15223050e-01 6.51313841e-01
-7.63661936e-02 5.63325882e-01 -2.34647036e-01 -1.08155906e-01
6.35957301e-01 4.95659739e-01 1.27620935e+00 -7.72570312e-01
-9.18613076e-01 -5.25290012e-01 5.04061766e-02 -1.20514834e+00
4.74393636e-01 -1.05123377e+00 -2.06078179e-02 -1.35411370e+00
-4.29705381e-02 -1.09079409e+00 -2.81399544e-02 8.24430883e-01
1.43279478e-01 3.70582283e-01 3.92247349e-01 2.47306541e-01
-1.10103285e+00 4.56225902e-01 1.35513186e+00 2.52406359e-01
-2.81357616e-01 -2.71125085e-04 -3.79199058e-01 7.50872791e-01
1.15033352e+00 -2.13963404e-01 -5.71118772e-01 -6.16338134e-01
1.95480928e-01 -4.33311015e-02 1.80047110e-01 -1.31360340e+00
1.52293578e-01 -2.38035083e-01 3.11255634e-01 -1.91984311e-01
2.69472212e-01 -6.52163386e-01 3.48541141e-02 5.27941525e-01
-2.66308397e-01 2.83925861e-01 4.21561688e-01 -5.65521084e-02
9.08541307e-02 -6.06770860e-03 6.91372156e-01 -3.06382507e-01
-7.15185404e-01 2.04881907e-01 -6.18652403e-01 2.27980986e-01
1.25262058e+00 2.45542035e-01 -6.07567251e-01 -1.12115189e-01
-2.49385357e-01 2.63212323e-01 4.14225727e-01 1.69501141e-01
2.35040948e-01 -1.04403031e+00 -5.25989950e-01 5.40479004e-01
-3.52763236e-01 1.13985889e-01 -5.03944866e-02 6.51906371e-01
-8.48086953e-01 4.01378244e-01 -4.23232198e-01 -5.17671824e-01
-1.15870607e+00 4.70303267e-01 3.12993884e-01 -9.00002301e-01
-9.03521180e-01 6.17073655e-01 -1.32873178e-01 -9.21246350e-01
2.78100222e-01 -3.34966719e-01 2.30476558e-01 -3.56332570e-01
5.20013332e-01 3.58498663e-01 1.99139029e-01 -1.01770557e-01
6.00786787e-03 2.09567860e-01 6.73037991e-02 4.46030125e-02
1.50164151e+00 3.65040094e-01 -1.18220836e-01 3.90619934e-01
1.04319501e+00 -2.47337166e-02 -1.54183662e+00 -9.12090912e-02
-1.47434682e-01 -8.78817514e-02 2.13778526e-01 -7.11554289e-01
-1.27362096e+00 5.77888310e-01 2.73898184e-01 -3.89360748e-02
1.09000134e+00 -2.08047509e-01 9.37325656e-01 8.76158834e-01
8.26082289e-01 -1.06742311e+00 7.90184662e-02 8.42800021e-01
6.69739664e-01 -1.16448152e+00 2.49765038e-01 1.98457971e-01
-7.55525589e-01 1.05637705e+00 8.41013491e-01 -3.09615165e-01
1.13826975e-01 8.49775136e-01 1.22669995e-01 -1.00552283e-01
-1.21818292e+00 -1.50581241e-01 -2.61707366e-01 3.95476639e-01
4.76583958e-01 1.28414303e-01 -1.22444797e-02 2.26346478e-01
-5.90418339e-01 1.31678581e-01 1.20395124e-01 1.26386142e+00
-3.29591900e-01 -1.33507991e+00 -3.07817847e-01 2.11130798e-01
-2.49047935e-01 -4.75238748e-02 1.45707220e-01 9.87391293e-01
6.00352790e-03 7.83042610e-01 3.36330421e-02 -2.59920865e-01
8.54104757e-03 -1.59829572e-01 6.45528376e-01 -4.39812988e-01
-1.21233284e+00 -6.79015890e-02 1.88791186e-01 -1.10155058e+00
-1.86839700e-01 -3.68400604e-01 -1.50322294e+00 -8.25870514e-01
2.27048919e-01 1.34903744e-01 9.30754900e-01 9.15987611e-01
4.96222883e-01 7.02389598e-01 3.99360150e-01 -1.16665006e+00
-7.18118846e-01 -5.96374214e-01 -5.87986171e-01 1.08422622e-01
-1.04576118e-01 -3.53915155e-01 -1.60653397e-01 -3.60833049e-01] | [3.8627443313598633, 1.660123348236084] |
4615c2b2-1388-4baa-b985-e9549efb7fce | face-image-lighting-enhancement-using-a-3d | 2207.00761 | null | https://arxiv.org/abs/2207.00761v1 | https://arxiv.org/pdf/2207.00761v1.pdf | Face Image Lighting Enhancement Using a 3D Model | Image enhancement helps to generate balanced lighting distributions over faces. Our goal is to get an illuminance-balanced enhanced face image from a single view. Traditionally, image enhancement methods ignore the 3D geometry of the face or require a complicated multi-view geometry. Other methods cause color tone shifting or over saturation. Inspired by the new research achievements in face alignment and face 3D modeling, we propose an improved face image enhancement method by leveraging 3D face models. Given a face image as input, our method will first estimate its lighting distribution. Then we build an optimization process to refine the distribution. Finally, we generate an illuminance-balanced face image from a single view. Experiments on the FiveK dataset demonstrate that our method performs well and compares favorably with other methods. | ['Jan P. Allebach', 'Qiulin Chen'] | 2022-07-02 | null | null | null | null | ['face-alignment'] | ['computer-vision'] | [ 1.91625267e-01 -3.62358481e-01 1.69101879e-01 -7.70110786e-01
-1.17582910e-01 -3.87004167e-01 2.36457676e-01 -8.97821665e-01
4.41015400e-02 3.09566915e-01 2.05552444e-01 -5.77341057e-02
2.02271298e-01 -6.46464586e-01 -4.09447521e-01 -7.68660486e-01
5.66592693e-01 -4.91338000e-02 -4.64131951e-01 -1.82492986e-01
1.48659393e-01 7.75057137e-01 -1.62738061e+00 -1.81806330e-02
5.50908744e-01 9.61227119e-01 -9.28694324e-04 5.87704062e-01
-1.42734069e-02 3.08905691e-01 -4.60729301e-01 -6.21567905e-01
5.11770666e-01 -4.34217066e-01 -4.08403546e-01 8.00319076e-01
7.37951338e-01 -7.21901357e-01 -1.85159102e-01 1.15273380e+00
7.80611992e-01 -1.19843081e-01 5.63052297e-01 -1.36910343e+00
-7.62805164e-01 -1.50828119e-02 -1.20060420e+00 -1.72735006e-01
2.26793349e-01 -9.63258222e-02 1.58105671e-01 -1.17324722e+00
4.21167135e-01 1.41778660e+00 5.46956003e-01 8.39222550e-01
-1.11909497e+00 -8.95513594e-01 2.49055579e-01 -5.36221750e-02
-1.38791406e+00 -8.71508539e-01 1.16187823e+00 -5.45055941e-02
2.59956837e-01 2.68303066e-01 5.50620139e-01 5.56635559e-01
-5.79098761e-02 3.81315976e-01 1.51034307e+00 -4.17184711e-01
-1.11193508e-01 1.47344679e-01 -4.75846916e-01 8.49234223e-01
-5.85858747e-02 2.03274250e-01 -5.60287595e-01 -2.80934107e-02
8.64745259e-01 1.58427075e-01 -3.07874441e-01 -1.22427478e-01
-6.37810349e-01 3.53621721e-01 8.28237161e-02 -1.37087014e-02
-2.38648802e-01 -1.54265419e-01 -1.25672236e-01 1.23987764e-01
7.00536668e-01 -3.11266689e-04 -1.98951572e-01 2.31208369e-01
-1.04670632e+00 7.20418990e-02 3.69490415e-01 8.84693086e-01
8.93136203e-01 8.55200663e-02 -1.17073968e-01 1.14050269e+00
4.94659334e-01 1.00835037e+00 -2.55757660e-01 -1.07552409e+00
1.50523752e-01 4.41166133e-01 6.82344288e-02 -1.06266761e+00
-1.77376643e-01 -2.14164495e-01 -8.57947409e-01 5.42619944e-01
2.30052486e-01 -2.85408169e-01 -9.16722357e-01 1.66475153e+00
6.78293645e-01 2.62817647e-02 -1.50698781e-01 9.93008614e-01
7.61188984e-01 4.99598563e-01 -3.45160425e-01 -5.20775080e-01
1.37276769e+00 -8.45281720e-01 -9.88593340e-01 -3.88347879e-02
-2.15740949e-01 -1.26973927e+00 7.98391759e-01 5.47716320e-01
-1.39193583e+00 -6.21542752e-01 -9.49428439e-01 2.82932748e-03
-6.59450772e-04 4.59345013e-01 3.15649122e-01 1.26413345e+00
-1.33322835e+00 1.86636269e-01 -3.66689980e-01 9.61838812e-02
5.01286149e-01 3.93670440e-01 -4.96983379e-01 -2.99189389e-01
-6.61030531e-01 7.87432015e-01 -2.51989216e-01 3.59193981e-01
-8.02160263e-01 -7.84040987e-01 -8.83235157e-01 -1.60206050e-01
2.15709493e-01 -5.65245748e-01 1.03663266e+00 -9.37886477e-01
-1.87128150e+00 1.04810750e+00 -6.91688061e-01 5.64930320e-01
2.45573685e-01 -4.10735048e-02 -4.40252930e-01 1.72598511e-01
-2.92842448e-01 5.78000367e-01 1.13599265e+00 -1.70894325e+00
-3.59244108e-01 -8.21195662e-01 -1.08624488e-01 4.11523342e-01
-4.60891962e-01 4.98822182e-01 -8.86073470e-01 -4.26250100e-01
2.63105839e-01 -6.64305806e-01 -2.64621321e-02 2.32451379e-01
-3.91072780e-01 3.21883500e-01 1.23622429e+00 -8.77721250e-01
9.53593194e-01 -2.02756381e+00 -4.03503366e-02 3.56441319e-01
2.63603210e-01 2.02349111e-01 -2.27915019e-01 -1.35366186e-01
-3.15023601e-01 -2.17716917e-02 7.53043070e-02 -7.48828828e-01
-1.46549642e-01 -1.83515176e-01 -1.24821195e-03 6.49451971e-01
2.28568912e-01 5.40715277e-01 -5.68210244e-01 -6.35357261e-01
3.43140334e-01 1.13406658e+00 -7.38296926e-01 3.86293560e-01
2.64449894e-01 5.01188159e-01 -7.76309073e-02 9.38585758e-01
1.63264477e+00 8.78325775e-02 2.07765162e-01 -4.74488020e-01
-9.68058184e-02 -5.38099766e-01 -1.23739517e+00 1.30751753e+00
-6.57374442e-01 4.04991329e-01 5.26639462e-01 -4.92620558e-01
1.20602798e+00 2.44082466e-01 5.67196727e-01 -7.01403916e-01
3.26224238e-01 4.85448120e-03 -4.12629426e-01 -3.93760264e-01
3.32170874e-01 -3.96169543e-01 5.36085427e-01 5.51664889e-01
-1.33636326e-01 -2.43497640e-01 -1.40934706e-01 -1.32112294e-01
1.97987169e-01 1.82858154e-01 -1.41567364e-01 -8.79118741e-02
7.52643347e-01 -8.22614551e-01 5.55114388e-01 -1.44998441e-02
-1.59294486e-01 8.89505446e-01 4.26661849e-01 -3.83611709e-01
-1.08507931e+00 -9.51288939e-01 -1.57483399e-01 9.22673523e-01
2.39254639e-01 -2.67423958e-01 -1.12887919e+00 -6.39728308e-01
-3.56530845e-01 7.93657452e-02 -5.63194335e-01 1.18940715e-02
-5.33118188e-01 -8.68838608e-01 2.43651330e-01 2.77524740e-01
8.68463278e-01 -4.89400566e-01 -2.72729993e-01 -3.17086816e-01
-3.19778860e-01 -1.07282901e+00 -8.22463691e-01 -6.48660362e-01
-6.22190952e-01 -1.00469017e+00 -1.04749191e+00 -7.98003554e-01
1.29590786e+00 5.47460020e-01 1.09594214e+00 2.42361009e-01
-4.82334733e-01 3.22698712e-01 1.74531024e-02 -5.12907267e-01
-1.03380002e-01 -5.85329711e-01 3.27412635e-01 3.57630163e-01
2.90776581e-01 -4.19292718e-01 -9.01093483e-01 7.11493611e-01
-7.23254919e-01 1.28273398e-01 2.34899387e-01 6.31843925e-01
5.95570028e-01 2.76565850e-01 3.42260122e-01 -4.67790037e-01
3.44823867e-01 7.71096423e-02 -7.93996990e-01 4.39415604e-01
-4.47337270e-01 -2.83528328e-01 3.62202346e-01 -2.03844592e-01
-1.75139475e+00 2.19902664e-01 -2.14563191e-01 -4.04067695e-01
-1.05125070e-01 -2.43201539e-01 -8.10320199e-01 -4.86710548e-01
3.56661916e-01 1.69538468e-01 2.96992421e-01 -3.63828033e-01
2.75464863e-01 8.11625421e-01 4.27841872e-01 -5.06507218e-01
8.49707425e-01 8.53728890e-01 6.68499023e-02 -9.94202733e-01
-5.85996628e-01 -1.48249507e-01 -6.26816154e-01 -6.54316187e-01
6.09433353e-01 -8.53596032e-01 -1.07458985e+00 7.87398458e-01
-1.02934229e+00 -1.29193023e-01 3.01388443e-01 2.51676053e-01
-3.47320735e-01 3.45979095e-01 -3.49268109e-01 -9.91087317e-01
-3.15416336e-01 -1.25863802e+00 1.29991758e+00 6.64931417e-01
4.56363261e-01 -7.75016904e-01 -1.54122725e-01 5.27221978e-01
5.75552762e-01 -2.15609334e-02 4.66596842e-01 6.27077520e-01
-2.90285408e-01 1.57304794e-01 -4.25432712e-01 4.95945573e-01
5.33759415e-01 3.99709582e-01 -1.39589691e+00 -2.97128916e-01
2.27582589e-01 -5.08990623e-02 4.94828701e-01 5.64562440e-01
1.57112741e+00 -1.02674134e-01 -3.04420680e-01 1.01398516e+00
1.35156548e+00 2.79359996e-01 7.93170929e-01 -1.45506948e-01
5.35080612e-01 8.97301316e-01 5.87679625e-01 5.90579748e-01
2.80224025e-01 7.91678548e-01 3.81160915e-01 -6.61345541e-01
-3.58756602e-01 -2.90241182e-01 2.06512854e-01 4.55591351e-01
-2.43529022e-01 -3.56325060e-02 -4.64255869e-01 1.48560271e-01
-9.70884621e-01 -9.49030817e-01 1.81489080e-01 2.05823994e+00
7.55077720e-01 -5.86415470e-01 5.06935120e-02 1.06741540e-01
8.77157331e-01 1.22443540e-03 -4.17930931e-01 -2.62607634e-01
-2.06903562e-01 1.03586830e-01 2.76696235e-01 7.42947638e-01
-9.37561214e-01 6.88692689e-01 7.02529478e+00 6.23705685e-01
-1.36588228e+00 -2.02286243e-02 1.27933621e+00 -3.60266924e-01
-5.17668843e-01 -3.59613955e-01 -1.14476430e+00 3.61317724e-01
3.22807044e-01 3.73568386e-02 5.92663527e-01 5.15853286e-01
4.16929513e-01 -7.11205006e-02 -6.80742681e-01 1.55992401e+00
5.71386635e-01 -9.34702098e-01 -1.39114216e-01 2.73040473e-01
9.65265810e-01 -6.46133423e-01 5.09016216e-01 -1.71831220e-01
-1.38138622e-01 -1.17244768e+00 4.62232381e-01 4.80312079e-01
1.14227319e+00 -8.83781374e-01 4.24922049e-01 -4.85181771e-02
-1.18363214e+00 2.21570686e-01 -3.96641701e-01 1.82600856e-01
2.33861580e-01 6.24111056e-01 -6.06818557e-01 3.75589907e-01
7.56967604e-01 3.52175832e-01 -4.78186160e-01 7.69242465e-01
-3.23420137e-01 1.38468549e-01 -1.51401132e-01 3.97651255e-01
-3.16222966e-01 -4.78497475e-01 2.91647196e-01 8.81905794e-01
4.15283829e-01 3.82777870e-01 -2.07863048e-01 9.18953300e-01
-1.99542478e-01 1.07131623e-01 -6.58065319e-01 2.75686681e-01
3.53958160e-01 1.67215133e+00 -7.03555763e-01 -1.43389583e-01
-5.40353835e-01 1.12562478e+00 -7.73777738e-02 5.16389370e-01
-9.59312201e-01 -1.72474608e-01 7.82774031e-01 9.82756019e-02
8.37651417e-02 2.76359320e-02 -2.17165917e-01 -9.29983675e-01
1.49212912e-01 -9.43472981e-01 -7.27246627e-02 -9.68636930e-01
-1.02175784e+00 7.19885767e-01 -1.26123443e-01 -9.23825622e-01
2.14440972e-01 -7.63502002e-01 -6.19192839e-01 1.12904322e+00
-1.85623038e+00 -1.23376679e+00 -5.68861306e-01 8.11916888e-01
3.59787405e-01 -9.17594582e-02 6.35135174e-01 6.62232816e-01
-5.29082477e-01 8.62038672e-01 -1.90837696e-01 1.06838662e-02
9.25980985e-01 -9.39712346e-01 2.08802342e-01 8.72517228e-01
-2.89520949e-01 5.87968767e-01 4.71527487e-01 -4.45853472e-01
-1.52176225e+00 -9.11196172e-01 6.71299458e-01 -2.66740322e-01
-2.42177263e-01 -2.71285295e-01 -5.28138340e-01 3.49108458e-01
4.68508273e-01 1.60594568e-01 6.85333371e-01 4.38689906e-03
-2.91706294e-01 -4.26227778e-01 -1.55893946e+00 5.52263439e-01
1.08748639e+00 -3.39527786e-01 2.92855743e-02 2.76732296e-02
6.99583963e-02 -5.88621318e-01 -6.81369781e-01 4.19957817e-01
8.45806003e-01 -1.10883617e+00 1.05926192e+00 -1.74617305e-01
4.31305319e-01 -5.49911618e-01 -1.54760838e-01 -1.41274369e+00
-6.80398494e-02 -8.40696812e-01 2.26762637e-01 1.43869245e+00
1.54314652e-01 -4.43632752e-01 8.30506921e-01 6.70889795e-01
1.95316911e-01 -5.93820691e-01 -4.10478562e-01 -4.36798483e-01
-2.77777046e-01 -8.48235339e-02 1.03610885e+00 7.26086557e-01
-2.12759390e-01 -2.58811321e-02 -6.21017218e-01 3.46230596e-01
9.61167157e-01 2.87089735e-01 9.08505201e-01 -9.67144668e-01
9.66479927e-02 -2.65338242e-01 4.58444729e-02 -8.89288723e-01
8.66195410e-02 -4.00831252e-01 -3.31696793e-02 -1.19476676e+00
6.12539351e-01 -4.03544694e-01 7.94265121e-02 3.70486289e-01
-2.76449382e-01 8.49395573e-01 8.96324888e-02 -4.47192490e-01
-1.92800358e-01 5.22759080e-01 1.62159681e+00 -1.79300196e-02
-1.10006288e-01 -4.94840927e-02 -1.08916450e+00 8.67463529e-01
8.15907717e-01 9.10923332e-02 -5.27014077e-01 -6.21786892e-01
1.94754288e-01 -4.04028520e-02 2.56964713e-01 -6.31874740e-01
-1.33717181e-02 -2.90047944e-01 9.46577072e-01 -6.91175044e-01
6.04402363e-01 -9.75221634e-01 1.65421098e-01 1.35129526e-01
2.23609537e-01 8.43139067e-02 2.26587519e-01 2.14579459e-02
-2.14711502e-01 3.04869725e-03 1.12775970e+00 -2.14020219e-02
-3.35591733e-01 6.86328113e-01 6.17861710e-02 -2.92852193e-01
9.94862378e-01 -4.32263255e-01 -3.71084005e-01 -4.27496225e-01
-5.62929928e-01 2.08297372e-02 7.61136889e-01 4.02460933e-01
9.58595216e-01 -1.52897644e+00 -8.09006870e-01 9.42091823e-01
-1.84922650e-01 -2.98917979e-01 5.62374711e-01 6.64935887e-01
-4.87017274e-01 2.73911692e-02 -4.00954276e-01 -5.61369240e-01
-1.99064565e+00 3.26562315e-01 6.28768086e-01 3.67067128e-01
-1.83419526e-01 1.08865690e+00 4.99107420e-01 -5.29481113e-01
6.62917122e-02 2.32064560e-01 -2.47109845e-01 -1.52277216e-01
1.16750479e+00 5.09269655e-01 9.05589014e-02 -8.62244427e-01
-3.45488310e-01 1.13846767e+00 4.55790609e-02 -2.72283912e-01
1.28033197e+00 -4.95413244e-01 -3.10121506e-01 -3.87856871e-01
1.28454030e+00 3.27260852e-01 -1.51299036e+00 -4.76463921e-02
-7.95649707e-01 -1.26821148e+00 1.78781733e-01 -6.62684679e-01
-1.76601100e+00 1.04129720e+00 9.19992030e-01 -2.09603101e-01
1.70412099e+00 -3.29878718e-01 4.64464188e-01 -1.85431585e-01
1.62086770e-01 -1.12072957e+00 3.82179677e-01 2.26161256e-01
9.30126548e-01 -1.22737575e+00 9.07953978e-02 -7.54529536e-01
-4.99573320e-01 1.11931515e+00 9.13338125e-01 4.02557999e-01
7.76231229e-01 6.49161875e-01 2.92928457e-01 -2.92663276e-01
-3.11424941e-01 -3.30476575e-02 2.09333673e-01 7.55409896e-01
5.84800124e-01 -1.85538724e-01 7.24826306e-02 1.41240895e-01
-1.86809033e-01 5.42096123e-02 3.35115731e-01 5.34279108e-01
-2.37975165e-01 -1.13841581e+00 -7.61454284e-01 -1.67101473e-02
-6.65277243e-01 8.18108097e-02 -1.59462601e-01 3.74577016e-01
3.21481287e-01 1.14771175e+00 -2.15108488e-02 -2.61791497e-01
3.13610971e-01 -8.99168551e-02 9.19266224e-01 -3.38395745e-01
-9.37986821e-02 3.28372657e-01 -2.75459111e-01 -5.30260503e-01
-6.54357314e-01 -3.63714486e-01 -7.91592062e-01 -7.01919377e-01
-3.86759937e-01 -1.20550662e-01 8.94040883e-01 3.99355948e-01
3.13198447e-01 3.40488404e-01 1.20404148e+00 -1.02452314e+00
-1.00592412e-01 -6.22729719e-01 -8.03871274e-01 2.68078148e-01
4.03891087e-01 -6.79679334e-01 -2.64304161e-01 2.92136490e-01] | [12.92263412475586, -0.026948630809783936] |
e43d9e89-d3cb-4040-891a-fa00d6487c57 | learning-to-learn-cropping-models-for | null | null | http://openaccess.thecvf.com/content_CVPR_2020/html/Li_Learning_to_Learn_Cropping_Models_for_Different_Aspect_Ratio_Requirements_CVPR_2020_paper.html | http://openaccess.thecvf.com/content_CVPR_2020/papers/Li_Learning_to_Learn_Cropping_Models_for_Different_Aspect_Ratio_Requirements_CVPR_2020_paper.pdf | Learning to Learn Cropping Models for Different Aspect Ratio Requirements | Image cropping aims at improving the framing of an image by removing its extraneous outer areas, which is widely used in the photography and printing industry. In some cases, the aspect ratio of cropping results is specified depending on some conditions. In this paper, we propose a meta-learning (learning to learn) based aspect ratio specified image cropping method called Mars, which can generate cropping results of different expected aspect ratios. In the proposed method, a base model and two meta-learners are obtained during the training stage. Given an aspect ratio in the test stage, a new model with new parameters can be generated from the base model. Specifically, the two meta-learners predict the parameters of the base model based on the given aspect ratio. The learning process of the proposed method is learning how to learn cropping models for different aspect ratio requirements, which is a typical meta-learning process. In the experiments, the proposed method is evaluated on three datasets and outperforms most state-of-the-art methods in terms of accuracy and speed. In addition, both the intermediate and final results show that the proposed model can predict different cropping windows for an image depending on different aspect ratio requirements.
| [' Kaiqi Huang', ' Junge Zhang', 'Debang Li'] | 2020-06-01 | null | null | null | cvpr-2020-6 | ['image-cropping'] | ['computer-vision'] | [ 3.50542665e-01 -2.77860165e-01 -1.52931362e-01 -2.36400977e-01
-3.81591678e-01 -2.45288655e-01 1.68455154e-01 3.91908325e-02
-1.98980257e-01 4.50895458e-01 -3.04747462e-01 -2.57213861e-01
-1.27063870e-01 -1.04032958e+00 -7.90907323e-01 -7.19861805e-01
3.82951766e-01 2.60784358e-01 3.67653340e-01 -1.47235334e-01
6.32891774e-01 3.35004449e-01 -1.80524445e+00 5.62902212e-01
1.22870278e+00 8.50153804e-01 8.98546278e-01 6.73734903e-01
-3.60306710e-01 3.64574850e-01 -6.34682298e-01 -8.48669782e-02
2.85254329e-01 -3.67634177e-01 -1.64786696e-01 4.84533608e-01
3.93463016e-01 -6.95256516e-02 2.71414399e-01 8.70346308e-01
1.90200523e-01 2.64465995e-03 7.97436595e-01 -1.09417570e+00
-4.48397696e-01 3.57290298e-01 -1.00112498e+00 -1.97038770e-01
-7.21973330e-02 4.20476571e-02 6.20615721e-01 -8.57375324e-01
1.60781443e-01 1.09078860e+00 4.25097167e-01 2.52200902e-01
-6.53604627e-01 -8.21283579e-01 4.03807521e-01 1.32290289e-01
-1.07506144e+00 2.43349284e-01 8.03602695e-01 -3.48447114e-01
2.75194138e-01 1.41734093e-01 7.70860136e-01 2.01106802e-01
5.50231218e-01 7.26113260e-01 1.34067345e+00 -6.91658497e-01
2.10361078e-01 4.22433436e-01 -2.22340316e-01 6.28970087e-01
2.07065076e-01 1.14164166e-01 -2.08025813e-01 1.57470480e-01
8.04863453e-01 4.96386318e-03 -9.43944156e-02 -5.65246344e-01
-1.05325365e+00 7.34805644e-01 5.33733308e-01 3.36282179e-02
-3.02713811e-01 -2.89580435e-01 1.96259856e-01 1.29729286e-01
4.12432700e-01 5.85424721e-01 -6.61462903e-01 2.63733983e-01
-7.89743900e-01 1.59842297e-01 4.27460879e-01 8.20405900e-01
9.75145638e-01 1.39990298e-03 1.11701703e-02 9.93159413e-01
2.83124149e-01 4.04897660e-01 4.72100943e-01 -4.95234251e-01
7.04573691e-01 8.29553425e-01 2.90843815e-01 -8.23795855e-01
-3.85858804e-01 -3.98015916e-01 -7.41620123e-01 6.18350625e-01
2.10129008e-01 -5.81443906e-01 -1.20979524e+00 1.21116161e+00
5.56909680e-01 3.63637686e-01 2.23953336e-01 8.41024280e-01
8.11548829e-01 1.11022627e+00 4.60206419e-02 -3.55133206e-01
1.22986650e+00 -1.05929029e+00 -4.23178941e-01 -4.52730149e-01
4.04399008e-01 -1.17802060e+00 9.57061529e-01 7.38381326e-01
-8.52895379e-01 -9.57592845e-01 -1.35755384e+00 5.53332567e-01
-2.52691627e-01 7.30424225e-01 5.09678602e-01 6.35165632e-01
-6.11720979e-01 4.77262676e-01 -3.71273339e-01 -1.94612756e-01
6.85172975e-02 1.24032564e-01 6.68792725e-02 -2.06964180e-01
-9.95596111e-01 8.53482842e-01 8.28982055e-01 1.01208910e-01
-6.58499718e-01 -7.45124519e-01 -7.81211138e-01 1.02325015e-01
4.53270644e-01 -4.01591361e-01 9.27627861e-01 -1.25666690e+00
-1.40050745e+00 5.39313972e-01 9.29291099e-02 -2.98038095e-01
4.67468888e-01 -1.60079330e-01 -4.83886778e-01 -2.62019783e-02
-3.69775504e-01 8.32345843e-01 1.27904797e+00 -1.62052965e+00
-1.13645482e+00 -1.23609200e-01 2.07246229e-01 6.03308380e-01
-3.19531292e-01 -4.97337788e-01 -5.01965821e-01 -6.67143643e-01
6.09173700e-02 -9.98716593e-01 -1.40384406e-01 -1.06980935e-01
-1.91099755e-02 5.29318191e-02 1.11631477e+00 -6.63566768e-01
1.26802289e+00 -1.97376764e+00 -7.74366781e-02 2.96596587e-01
-3.46517146e-01 6.06993496e-01 -2.25485221e-01 1.33605257e-01
-1.83454141e-01 2.64375750e-02 -2.73432046e-01 1.60681620e-01
-7.46804297e-01 -2.10591406e-01 -3.87144974e-03 -6.61365911e-02
3.90818179e-01 2.58767933e-01 -6.71425641e-01 -6.19616389e-01
7.43469000e-01 4.10071880e-01 -3.30988139e-01 4.02077615e-01
-4.90635425e-01 4.07127857e-01 -3.66299868e-01 5.07754982e-01
1.08239830e+00 1.15330197e-01 2.18206599e-01 -5.55087686e-01
-2.48699501e-01 -5.70815027e-01 -1.28791332e+00 1.28051162e+00
-1.03925049e+00 4.79879797e-01 -4.19481337e-01 -9.88611341e-01
1.58434486e+00 -4.44275141e-03 2.28041798e-01 -5.45637488e-01
1.75065950e-01 1.42542392e-01 9.17428583e-02 -5.76721311e-01
5.88536143e-01 8.50686058e-02 2.23003849e-01 3.13191265e-01
-2.70085990e-01 -3.83904696e-01 2.92135149e-01 -4.18673784e-01
1.79866195e-01 4.54526931e-01 3.56688797e-01 -1.45633658e-02
8.12138855e-01 7.34156147e-02 5.14411986e-01 4.21410143e-01
2.43416682e-01 6.69908106e-01 1.92823976e-01 -6.56715631e-01
-1.26445127e+00 -7.28735387e-01 -1.17735974e-01 8.91918600e-01
4.60797876e-01 -1.09742992e-01 -9.28030670e-01 -6.36364520e-01
-3.22091401e-01 7.77437985e-01 -5.29178560e-01 -1.32149801e-01
-6.68002903e-01 -8.29942286e-01 -3.15964460e-01 3.33412617e-01
1.03992260e+00 -1.34506583e+00 -8.65970194e-01 -3.79829369e-02
-1.26448363e-01 -8.19761336e-01 -3.44711125e-01 -2.32436627e-01
-1.06110525e+00 -1.23907220e+00 -9.41197872e-01 -1.15880632e+00
8.93455207e-01 5.96924841e-01 7.48660147e-01 3.89089048e-01
-2.36126855e-01 -1.44618183e-01 -5.38264811e-01 -7.79643238e-01
-6.48345947e-01 2.40299329e-01 -5.19395113e-01 1.33918136e-01
1.63552508e-01 -2.03741685e-01 -5.95940173e-01 5.56316316e-01
-1.12221384e+00 5.51455498e-01 1.04384136e+00 7.87879705e-01
6.86169684e-01 6.32125258e-01 6.39218807e-01 -1.01883805e+00
5.08306801e-01 -3.64722818e-01 -9.79113400e-01 6.72790170e-01
-7.56540358e-01 1.34696245e-01 7.93618858e-01 -6.95892692e-01
-1.43297970e+00 1.62850305e-01 2.97512233e-01 -3.28858018e-01
9.26224440e-02 5.80545783e-01 -2.84652948e-01 5.79684377e-02
5.36008060e-01 3.51272933e-02 1.46969691e-01 -1.17599070e-01
1.05957828e-01 7.02796161e-01 2.99003214e-01 -3.59568924e-01
8.37531745e-01 3.88649739e-02 5.26393801e-02 -7.01473176e-01
-7.70854712e-01 -1.26606748e-01 -5.74758291e-01 -5.10620892e-01
7.79641092e-01 -8.43800783e-01 -3.48732233e-01 8.82612765e-01
-8.53172660e-01 -1.98209047e-01 8.29996392e-02 5.47919452e-01
-4.10498559e-01 9.59164947e-02 6.41138256e-02 -8.20605874e-01
-4.78588164e-01 -1.18976426e+00 8.47310841e-01 9.72111285e-01
2.64839768e-01 -8.55973303e-01 -2.44689450e-01 2.14594230e-01
2.39361182e-01 2.71610379e-01 1.13227630e+00 -8.63290951e-02
-4.41509843e-01 -8.97079557e-02 -1.51428416e-01 3.79720777e-01
4.06609446e-01 3.31930876e-01 -6.79076433e-01 -2.86030978e-01
-3.06781411e-01 -3.05902562e-03 7.91548312e-01 5.29680192e-01
1.25315177e+00 -8.53587613e-02 -2.88858384e-01 5.54039776e-01
1.59852481e+00 7.65226126e-01 8.08601141e-01 6.37373269e-01
4.47152734e-01 5.91044545e-01 1.46527898e+00 4.35234874e-01
-3.46782757e-03 3.50097775e-01 5.97896338e-01 -1.50269285e-01
1.95496026e-02 -4.28921282e-01 1.76364675e-01 2.51455724e-01
1.31490171e-01 -1.58211499e-01 -5.87711513e-01 4.43764538e-01
-1.76432467e+00 -7.71263301e-01 2.43175283e-01 2.64534783e+00
5.96465707e-01 3.86141926e-01 1.05193406e-01 1.90152779e-01
1.17094517e+00 2.37369865e-01 -8.90893042e-01 -6.48025811e-01
1.80440828e-01 8.19832012e-02 4.68978256e-01 3.46263587e-01
-1.12627769e+00 9.13581073e-01 5.15767574e+00 7.80040741e-01
-1.47705543e+00 -5.39165258e-01 7.26416826e-01 2.84230947e-01
-1.63993433e-01 1.72405675e-01 -8.99284959e-01 4.91018474e-01
4.62719351e-01 -1.76208690e-01 3.90972793e-01 7.44981647e-01
2.56812185e-01 -4.77569848e-01 -5.84165633e-01 8.64383638e-01
2.42773011e-01 -1.17849374e+00 3.11893255e-01 -2.97012150e-01
8.59638512e-01 -6.54188931e-01 8.50142837e-02 3.28798503e-01
-1.40037879e-01 -7.71202147e-01 4.55500245e-01 5.47772288e-01
7.09434152e-01 -1.14702559e+00 7.53007829e-01 5.48607886e-01
-1.28317976e+00 -5.11676371e-01 -5.32608986e-01 2.96511859e-01
-1.32262632e-02 5.79454243e-01 -1.02660978e+00 6.67365134e-01
6.40790582e-01 3.56863797e-01 -6.95816994e-01 1.34761512e+00
-1.97976366e-01 4.97111708e-01 -5.55691794e-02 -1.77760407e-01
1.15389107e-02 -4.57568735e-01 2.56964147e-01 8.04902494e-01
7.59608209e-01 -1.26716375e-01 5.46364188e-02 5.28977811e-01
2.85588615e-02 3.51540864e-01 -3.06675255e-01 1.95915908e-01
6.50539517e-01 1.42891717e+00 -8.02721798e-01 -2.81548202e-01
-2.11303562e-01 5.32823086e-01 1.53632080e-02 3.50863002e-02
-7.78057218e-01 -4.23042864e-01 -6.20976230e-03 1.97323188e-01
5.04940510e-01 1.45389184e-01 -2.58932918e-01 -7.82065749e-01
-8.48973244e-02 -8.57160270e-01 3.10652465e-01 -1.13466835e+00
-7.68743038e-01 4.20490682e-01 3.19051474e-01 -1.77862191e+00
-8.95435289e-02 -6.16434574e-01 -9.66278255e-01 8.06836307e-01
-1.57318103e+00 -1.16927767e+00 -8.24997425e-01 1.36639163e-01
1.01752365e+00 -4.17621791e-01 4.43041682e-01 -1.71660572e-01
-4.08422381e-01 3.81704628e-01 -7.46393390e-03 -1.73335835e-01
8.78124774e-01 -9.76365626e-01 1.99946851e-01 7.12788224e-01
1.62857841e-03 1.42750874e-01 7.66149580e-01 -6.33741021e-01
-9.14242506e-01 -1.15656972e+00 2.83651203e-01 3.49862516e-01
-1.00514442e-01 1.12362362e-01 -8.85729909e-01 3.65925014e-01
1.66388482e-01 -1.75008833e-01 2.94518203e-01 -1.52481303e-01
-4.35265042e-02 -4.65710521e-01 -1.23221338e+00 5.83921611e-01
2.98744500e-01 4.18510169e-01 -4.18923169e-01 1.49534792e-01
5.32159209e-01 -8.30824614e-01 -6.48290217e-01 6.60715520e-01
8.79504800e-01 -1.02967477e+00 7.30156958e-01 -2.66857952e-01
6.44799411e-01 -5.17284930e-01 3.28591257e-01 -1.82294703e+00
-1.96553335e-01 -3.24292853e-02 1.19006537e-01 1.25627887e+00
5.41255414e-01 -4.12389338e-01 8.27409327e-01 2.72769481e-01
6.96141943e-02 -1.03128362e+00 -2.59470046e-01 -3.71813804e-01
-3.20166238e-02 -5.00179417e-02 8.69148433e-01 5.35019159e-01
-5.08116126e-01 1.68051898e-01 -4.95536119e-01 4.44655508e-01
3.47035348e-01 4.85458881e-01 9.75801766e-01 -1.09425533e+00
-2.84982651e-01 1.51478769e-02 -1.21464908e-01 -7.05561876e-01
-1.40236050e-01 -3.12194467e-01 -4.46757562e-02 -1.54609370e+00
1.16384335e-01 -6.33865535e-01 -1.42664865e-01 2.21874237e-01
-5.80116868e-01 -1.16598003e-01 4.75097418e-01 -1.17800271e-04
-4.87548336e-02 2.40764365e-01 1.49596679e+00 -3.40681046e-01
-7.85570264e-01 6.39567852e-01 -7.06129491e-01 7.89526761e-01
1.26567018e+00 -1.20652556e-01 -8.52088153e-01 -4.24423248e-01
1.42027333e-01 2.13261291e-01 -1.00286238e-01 -1.21079695e+00
-1.61951575e-02 -4.53452766e-01 6.52657449e-01 -9.68632281e-01
1.16856165e-01 -9.86961484e-01 2.63826102e-02 4.10708785e-01
-2.18790650e-01 1.30528510e-01 4.16035503e-01 3.37706655e-01
2.46576797e-02 -6.60118878e-01 9.79668677e-01 -1.54809266e-01
-8.90816271e-01 2.38900393e-01 -3.28993499e-02 -2.01432362e-01
1.36971641e+00 -3.94118607e-01 -1.77000642e-01 -3.96192491e-01
-4.09158915e-01 3.58357072e-01 3.35999221e-01 7.35838175e-01
8.87144625e-01 -1.15973699e+00 -7.97445238e-01 2.69706786e-01
2.58676648e-01 2.64824420e-01 5.71019053e-01 3.94142896e-01
-6.61586881e-01 -8.40772539e-02 -4.45416212e-01 -5.64379692e-01
-1.53127253e+00 6.35207713e-01 3.60798180e-01 -3.03166360e-01
-2.30805144e-01 1.94733575e-01 2.13839829e-01 -3.81681472e-01
-1.75268590e-01 -1.51420653e-01 -6.41739249e-01 -5.08795902e-02
5.29911339e-01 2.86288679e-01 1.18898660e-01 -2.80760527e-01
1.27836272e-01 9.88654613e-01 -3.66524607e-01 3.58324014e-02
1.10884333e+00 2.63513234e-02 3.33967865e-01 3.55879068e-01
8.79944205e-01 -2.14844886e-02 -1.48741341e+00 -1.09916642e-01
-3.30344886e-01 -7.14763343e-01 -1.11564547e-01 -8.82967830e-01
-1.36411071e+00 8.07214081e-01 9.91075873e-01 -2.04832166e-01
1.45944774e+00 -5.01978695e-01 6.63553536e-01 2.23110542e-01
3.57159078e-01 -1.23953915e+00 -6.25713437e-04 1.29041374e-01
8.96635771e-01 -1.30644953e+00 3.41416448e-01 -4.25205499e-01
-9.17698324e-01 1.41152382e+00 1.17925620e+00 -2.50297785e-01
5.98445654e-01 2.29407966e-01 3.57822984e-01 3.10765415e-01
-5.71511030e-01 2.41483375e-01 2.02614203e-01 8.20603192e-01
3.22659731e-01 -9.72841755e-02 -5.96380591e-01 1.35995179e-01
-2.19624847e-01 9.05321091e-02 5.96326649e-01 8.00400853e-01
-7.98784852e-01 -1.19593799e+00 -6.37642622e-01 6.15373075e-01
3.79885361e-02 1.29600152e-01 -6.78555807e-03 9.29349422e-01
3.96420836e-01 9.38046992e-01 1.94155037e-01 -5.08379102e-01
2.97102869e-01 -2.34464571e-01 5.19246459e-01 -5.13514459e-01
-4.11232263e-01 -7.52546713e-02 -1.94736660e-01 1.02760427e-01
-4.67163473e-01 -3.92041117e-01 -1.24614406e+00 -7.82432035e-03
-7.22951651e-01 2.49906361e-01 7.62129843e-01 8.25088263e-01
1.88466370e-01 5.40447414e-01 1.11368346e+00 -7.46824205e-01
-3.25974822e-01 -9.72058296e-01 -4.61202770e-01 1.64335638e-01
-6.90103546e-02 -6.92986310e-01 -1.70701653e-01 1.49598822e-01] | [11.26611328125, -1.0796282291412354] |
6003a15b-d396-4e3c-8a35-d5248d53bab3 | can-chatgpt-detect-intent-evaluating-large | 2305.13512 | null | https://arxiv.org/abs/2305.13512v1 | https://arxiv.org/pdf/2305.13512v1.pdf | Can ChatGPT Detect Intent? Evaluating Large Language Models for Spoken Language Understanding | Recently, large pretrained language models have demonstrated strong language understanding capabilities. This is particularly reflected in their zero-shot and in-context learning abilities on downstream tasks through prompting. To assess their impact on spoken language understanding (SLU), we evaluate several such models like ChatGPT and OPT of different sizes on multiple benchmarks. We verify the emergent ability unique to the largest models as they can reach intent classification accuracy close to that of supervised models with zero or few shots on various languages given oracle transcripts. By contrast, the results for smaller models fitting a single GPU fall far behind. We note that the error cases often arise from the annotation scheme of the dataset; responses from ChatGPT are still reasonable. We show, however, that the model is worse at slot filling, and its performance is sensitive to ASR errors, suggesting serious challenges for the application of those textual models on SLU. | ['Philip N. Garner', 'Mutian He'] | 2023-05-22 | null | null | null | null | ['spoken-language-understanding', 'intent-classification', 'slot-filling', 'spoken-language-understanding'] | ['natural-language-processing', 'natural-language-processing', 'natural-language-processing', 'speech'] | [ 1.67302385e-01 4.87814367e-01 -1.97600737e-01 -4.80407417e-01
-1.22267842e+00 -6.44670069e-01 6.64300859e-01 1.71109006e-01
-5.07394314e-01 7.50356436e-01 7.74684012e-01 -7.48746097e-01
4.43178676e-02 -4.09538358e-01 -6.08142197e-01 -2.82535583e-01
-5.28472699e-02 8.69960189e-01 9.30462927e-02 -3.95400465e-01
1.46016270e-01 -3.56261373e-01 -1.07294464e+00 5.14078796e-01
7.92903066e-01 6.04862154e-01 1.28216192e-01 8.58259797e-01
-5.91001451e-01 1.19378555e+00 -6.32449031e-01 -2.00371072e-01
6.20690435e-02 -3.17199737e-01 -1.27645504e+00 3.09679396e-02
1.94219440e-01 -5.84843338e-01 -2.07104057e-01 4.96595144e-01
4.19608980e-01 3.99100691e-01 5.18566191e-01 -1.04466939e+00
-4.39505607e-01 1.19569242e+00 -1.87988490e-01 1.99859336e-01
6.00035012e-01 3.07343662e-01 1.43971026e+00 -9.03147638e-01
7.09051192e-01 1.40761578e+00 8.09669852e-01 5.79038441e-01
-1.35525560e+00 -3.80067319e-01 3.20883512e-01 -9.48763713e-02
-1.17820287e+00 -8.01964700e-01 4.58144844e-02 -2.38889158e-01
1.56192935e+00 1.85055986e-01 1.52897567e-01 1.30714333e+00
-7.23233148e-02 1.03357625e+00 1.03702044e+00 -6.14657521e-01
2.49757573e-01 1.42298788e-01 7.92888820e-01 6.30037487e-01
-2.78994799e-01 -6.16371334e-02 -8.80403578e-01 -3.02831888e-01
1.62325427e-01 -2.39068419e-01 -3.43590617e-01 6.41043782e-02
-1.01227283e+00 8.92137945e-01 1.40407681e-01 3.75370026e-01
-1.31606404e-03 4.80303094e-02 6.35787606e-01 4.39790308e-01
7.81213284e-01 6.86622083e-01 -8.09222996e-01 -8.05491149e-01
-8.90992761e-01 -1.69272311e-02 1.37046015e+00 9.79863703e-01
7.44329691e-01 -9.24376026e-02 -4.98332798e-01 1.01146924e+00
-8.17013346e-03 -3.68622839e-02 5.64018250e-01 -8.65201533e-01
5.36079645e-01 2.80811757e-01 -4.34477441e-02 -2.74262220e-01
-5.89134514e-01 -3.36252749e-01 -2.49134392e-01 -3.89398992e-01
8.17594588e-01 -4.21910316e-01 -9.68201339e-01 1.79493105e+00
-2.22353205e-01 3.84925544e-01 5.40894568e-01 6.86696649e-01
9.49796498e-01 9.01192307e-01 5.59837401e-01 -6.65953085e-02
1.31673312e+00 -1.15552294e+00 -7.60945499e-01 -8.07632446e-01
1.31977665e+00 -7.09366024e-01 1.51755917e+00 1.68299794e-01
-7.10141361e-01 -3.33025098e-01 -6.49340510e-01 -3.25724572e-01
-1.67056546e-01 -1.45718142e-01 8.89795661e-01 6.57355547e-01
-1.15274727e+00 4.00391966e-01 -7.07402170e-01 -8.89155030e-01
2.31728405e-01 6.08822070e-02 -7.45519996e-02 -1.56596094e-01
-1.40776455e+00 9.06446457e-01 2.33047336e-01 -2.98897445e-01
-8.94753635e-01 -8.65098357e-01 -8.66182387e-01 4.24347550e-01
5.26169837e-01 -3.21936637e-01 1.88458884e+00 -7.67915308e-01
-1.53023458e+00 8.99103761e-01 -4.21338528e-01 -6.44628882e-01
4.02188689e-01 -4.14997488e-01 -2.02217823e-04 -1.87721625e-01
4.08026949e-02 7.04840779e-01 3.05343240e-01 -9.45735514e-01
-4.91776079e-01 -1.66603580e-01 1.22910455e-01 3.24398011e-01
-3.35203469e-01 1.50190622e-01 -2.50891030e-01 -1.66904166e-01
2.48267082e-03 -6.00215375e-01 -2.04387337e-01 -5.13544858e-01
-3.46617490e-01 -4.52059984e-01 5.60005844e-01 -7.68196285e-01
1.33565509e+00 -1.99810255e+00 -2.58724242e-01 -2.29050726e-01
-1.89086199e-02 9.31530595e-02 -3.82978946e-01 7.79100060e-01
1.32875159e-01 3.72816920e-01 -1.34334520e-01 -6.29396260e-01
6.90529030e-03 4.87573296e-01 -6.42771900e-01 -9.45514366e-02
2.37714976e-01 1.09681392e+00 -8.36204529e-01 -1.27899125e-01
-4.26241644e-02 -3.70236263e-02 -6.73781633e-01 3.06434780e-01
-4.49676812e-01 2.81008214e-01 -2.98381716e-01 5.10448694e-01
9.86980945e-02 -5.19244611e-01 3.66481334e-01 3.81151587e-01
-5.04005663e-02 1.00715923e+00 -6.02240205e-01 1.77773452e+00
-6.43275082e-01 5.63850760e-01 7.94028342e-02 -8.33693206e-01
6.31520867e-01 6.47437215e-01 -1.89180039e-02 -6.53020263e-01
-1.12362787e-01 3.21374863e-01 2.71487147e-01 -3.55188936e-01
7.26227283e-01 -3.13288122e-01 -1.17707394e-01 1.00108373e+00
2.98959821e-01 -1.31390795e-01 -1.00277737e-01 6.44026339e-01
1.24089408e+00 -1.20892398e-01 2.86375284e-01 -3.19104224e-01
-1.47947103e-01 2.14094177e-01 2.12904930e-01 1.42719901e+00
-2.36432493e-01 5.87749958e-01 6.85382843e-01 -2.15645462e-01
-9.62594330e-01 -6.71296477e-01 -1.36943117e-01 1.68661726e+00
-1.87534899e-01 -7.05105424e-01 -7.73093820e-01 -5.71769536e-01
-1.84478506e-01 1.23661172e+00 -2.80581266e-01 -1.76779956e-01
-1.72008768e-01 -6.95672810e-01 7.87274003e-01 6.68705344e-01
4.53167483e-02 -9.11164343e-01 -3.91464025e-01 4.71428663e-01
-2.87216663e-01 -1.42296052e+00 -2.70810813e-01 5.57594359e-01
-8.01233947e-01 -6.15514219e-01 -4.66672570e-01 -5.47785759e-01
1.62623316e-01 1.31274283e-01 1.43787074e+00 2.50324279e-01
7.80998692e-02 7.26426661e-01 -5.23920655e-01 -4.23778862e-01
-7.20241070e-01 6.24096394e-01 3.94703485e-02 -3.74782652e-01
5.87040126e-01 -3.58001322e-01 -4.32334617e-02 1.95299715e-01
-5.14943600e-01 1.87567025e-01 2.15515092e-01 1.18360126e+00
-1.81503356e-01 -5.40300548e-01 7.46627808e-01 -1.25394988e+00
7.75249302e-01 -6.44760549e-01 -1.58399493e-01 3.30839962e-01
-4.88243282e-01 4.89869416e-02 3.43862444e-01 -3.09558451e-01
-1.18658185e+00 -4.12383378e-01 -3.85979354e-01 4.07860838e-02
-3.58070880e-01 6.18282974e-01 2.62492627e-01 2.16887966e-01
7.37017810e-01 2.16759682e-01 1.03853114e-01 -5.47494709e-01
4.16633099e-01 8.39474201e-01 1.22351520e-01 -9.10287380e-01
2.01908484e-01 1.02970183e-01 -1.01561773e+00 -1.07535183e+00
-1.29754865e+00 -6.61239624e-01 -2.15656742e-01 1.76526189e-01
7.11456120e-01 -1.15350342e+00 -5.50169349e-01 3.79624993e-01
-1.30875015e+00 -9.51255381e-01 -2.61050820e-01 2.89635122e-01
-5.95633447e-01 2.43741035e-01 -1.27557981e+00 -1.11992836e+00
-3.17753315e-01 -1.07747412e+00 1.12100029e+00 1.04663312e-01
-8.60213161e-01 -1.33708096e+00 -1.51011989e-01 4.68836963e-01
5.29662728e-01 -7.42991626e-01 1.29596245e+00 -1.37728035e+00
-5.68152010e-01 1.19637087e-01 -1.23589225e-01 -4.84260581e-02
-8.41136947e-02 -2.47501120e-01 -1.44489598e+00 -2.82681257e-01
3.89929898e-02 -9.21521544e-01 9.18716192e-01 2.36879885e-01
8.10659409e-01 -3.50036532e-01 -2.69820422e-01 3.44754398e-01
1.09638083e+00 1.55263707e-01 2.95916319e-01 2.28968978e-01
3.80798578e-01 6.99973226e-01 5.00439703e-01 2.65773058e-01
4.19171363e-01 4.70309258e-01 -1.42960742e-01 -5.12438044e-02
1.06818512e-01 -5.55737436e-01 5.01508296e-01 8.33378673e-01
4.10973728e-01 -5.20777822e-01 -1.35370696e+00 5.13394296e-01
-1.76919854e+00 -5.86321473e-01 -5.43860346e-02 1.96158350e+00
9.55266297e-01 3.27215195e-01 -1.12779349e-01 -2.47821137e-01
1.99752212e-01 3.49718541e-01 -2.83956587e-01 -6.82176948e-01
-6.69296756e-02 2.08563328e-01 2.80937612e-01 9.32152390e-01
-6.92484260e-01 1.38326871e+00 7.58055496e+00 8.31962168e-01
-9.20984447e-01 3.28899503e-01 1.18044400e+00 5.02228588e-02
-4.77669746e-01 1.97712705e-01 -1.04231060e+00 1.82849795e-01
1.31522405e+00 -1.81917146e-01 3.08033049e-01 6.52239442e-01
1.72665849e-01 -3.97530317e-01 -1.22144246e+00 4.86059725e-01
-9.91803855e-02 -1.26036084e+00 -1.00736149e-01 -2.12476045e-01
7.24796653e-01 3.70621711e-01 -2.15246484e-01 9.47624028e-01
6.57660663e-01 -1.20250440e+00 4.39398050e-01 4.72074747e-02
6.14776909e-01 -3.59172106e-01 5.76395929e-01 8.36266220e-01
-5.54476440e-01 -1.54044271e-01 -4.62448567e-01 -7.18346000e-01
4.10242528e-01 1.85117498e-01 -1.41438985e+00 2.32375622e-01
3.40462297e-01 4.36175734e-01 -4.51160520e-01 6.18187308e-01
-1.88383475e-01 1.25071323e+00 -4.59282666e-01 -1.64369151e-01
6.55206859e-01 -1.80752929e-02 4.06546891e-01 1.30078888e+00
1.88593477e-01 3.83307666e-01 4.64410782e-01 7.11036026e-01
-3.24791372e-02 2.98140496e-01 -7.24079609e-01 -4.17147756e-01
5.02180338e-01 9.57199275e-01 -6.28178775e-01 -6.57446802e-01
-7.45126903e-01 8.03289831e-01 5.74546754e-01 6.76255822e-01
-4.00207698e-01 2.67590433e-01 8.34752977e-01 -1.71322841e-02
-5.14825024e-02 -1.84032500e-01 -4.88638908e-01 -1.12802839e+00
-3.37293416e-01 -1.09452391e+00 4.92738217e-01 -8.07839572e-01
-1.23133385e+00 4.95035827e-01 -2.29901686e-01 -5.43262303e-01
-5.66724300e-01 -6.55461311e-01 -7.49119639e-01 8.10961246e-01
-1.32713032e+00 -8.95097971e-01 5.54384701e-02 1.64081708e-01
1.01013517e+00 -9.20015424e-02 1.10344875e+00 6.22799806e-02
-5.89712381e-01 6.27659261e-01 -5.12004420e-02 1.87244132e-01
8.19594085e-01 -1.11884499e+00 8.27342331e-01 6.28544748e-01
4.74091977e-01 7.58221984e-01 8.26975882e-01 -5.17391622e-01
-1.10933375e+00 -5.66376805e-01 1.24660587e+00 -9.12297189e-01
9.61455643e-01 -6.11484408e-01 -1.26737010e+00 1.17848027e+00
3.90711665e-01 -4.20980871e-01 7.45714009e-01 8.84079456e-01
-3.23671430e-01 4.95731592e-01 -5.37989497e-01 5.23461401e-01
1.18653178e+00 -8.12826991e-01 -7.66498208e-01 4.80251849e-01
1.22214437e+00 -5.58824599e-01 -5.69255054e-01 7.16706887e-02
2.30439946e-01 -9.64765608e-01 5.70932984e-01 -1.25662613e+00
5.90081871e-01 3.80141824e-01 -2.89785452e-02 -1.39531553e+00
1.29070029e-01 -5.88856339e-01 9.83079448e-02 1.31014729e+00
8.22396994e-01 -7.49771357e-01 6.78347647e-01 9.47665691e-01
-2.00406238e-01 -6.93160117e-01 -7.54290044e-01 -6.09052896e-01
2.43062183e-01 -6.98390484e-01 2.39783764e-01 1.13543010e+00
3.66598696e-01 8.91698003e-01 -4.21420962e-01 -2.59016752e-01
1.95772707e-01 -1.05865523e-01 6.82490885e-01 -9.66303766e-01
-5.77339768e-01 -1.06475689e-01 1.92951873e-01 -1.44947386e+00
4.62145030e-01 -8.07667553e-01 3.57606918e-01 -1.31475377e+00
3.88162851e-01 -6.86000884e-01 -1.90052986e-02 7.61904001e-01
-4.52672035e-01 -1.14149906e-01 1.94675803e-01 1.37607725e-02
-7.02736139e-01 4.50149953e-01 8.64886820e-01 1.10586137e-01
-2.79998451e-01 2.09236108e-02 -8.28254044e-01 7.52687693e-01
6.03493154e-01 -2.65431851e-01 -4.20650244e-01 -7.25053012e-01
-3.83788496e-02 3.16107243e-01 7.30210990e-02 -7.70624757e-01
3.81603539e-01 1.81050465e-01 -2.30458796e-01 -1.37223646e-01
3.14636439e-01 -3.74246508e-01 -3.07949692e-01 2.60407358e-01
-8.34932268e-01 -2.80897528e-01 4.19966251e-01 3.52153540e-01
-2.86102474e-01 -3.45861048e-01 4.97015983e-01 -3.42717737e-01
-1.03540051e+00 1.86510198e-02 -7.74523675e-01 4.82126147e-01
4.15532500e-01 -1.24675028e-01 -4.80087847e-01 -9.32981908e-01
-7.00661778e-01 4.10927564e-01 3.07590812e-01 5.24259627e-01
1.86834350e-01 -6.99584484e-01 -5.92359245e-01 1.96359172e-01
2.94531435e-01 -1.20108999e-01 2.38771081e-01 9.86541748e-01
-1.68133110e-01 7.29447544e-01 3.19604307e-01 -6.85630977e-01
-1.06332326e+00 1.07440896e-01 4.20891881e-01 -5.12263000e-01
-6.06437266e-01 1.14844501e+00 5.17134547e-01 -8.55308771e-01
3.62093478e-01 -4.17533457e-01 1.82989538e-01 7.37202764e-02
6.43866181e-01 6.15819320e-02 6.13135323e-02 -1.20876066e-01
-1.48821458e-01 5.61636034e-03 -2.90542632e-01 -2.87527978e-01
1.08657277e+00 -2.24990457e-01 2.14088708e-01 7.98340082e-01
9.30311441e-01 -6.80332631e-02 -1.24878061e+00 -5.29903233e-01
4.25013721e-01 -1.99830253e-02 -1.56703785e-01 -9.21989262e-01
-3.29600960e-01 1.10285771e+00 1.37322918e-01 1.47009715e-01
5.37901819e-01 1.52648002e-01 8.24011087e-01 6.74007475e-01
5.67955375e-01 -1.02281666e+00 -1.24560595e-02 1.23070145e+00
4.16634858e-01 -1.43123639e+00 -4.10610378e-01 -2.84254819e-01
-8.93013656e-01 8.65805745e-01 7.56815732e-01 3.95737737e-01
2.04691023e-01 4.42104757e-01 5.86249053e-01 -2.39242166e-02
-1.36701703e+00 -2.50403911e-01 -1.81644082e-01 2.52743155e-01
8.68973434e-01 -8.15827101e-02 -2.16812387e-01 7.04093456e-01
-3.42956781e-01 -1.27070099e-01 4.51702684e-01 6.77782953e-01
-5.85266948e-01 -9.14233506e-01 -8.89055431e-02 4.86740023e-01
-4.02744621e-01 -6.63604319e-01 -3.87034535e-01 8.07401896e-01
-6.06475890e-01 1.03012824e+00 3.43277901e-01 -1.40041187e-01
1.11942954e-01 7.96970606e-01 1.70223843e-02 -1.20727825e+00
-7.83687413e-01 5.67262657e-02 6.47630870e-01 -6.33319259e-01
1.19451709e-01 -4.25113857e-01 -1.34780955e+00 -1.39365315e-01
-2.12507531e-01 4.02267128e-01 2.08591744e-01 1.22793210e+00
2.36996382e-01 3.76545817e-01 1.49708867e-01 -5.17336488e-01
-9.58641231e-01 -1.27202225e+00 -4.59097832e-01 3.02783668e-01
1.69907093e-01 -2.29852796e-01 -4.50599760e-01 -2.66281128e-01] | [11.268059730529785, 8.343341827392578] |
1786cc80-de6c-42a6-84dc-e29507216d4c | early-myocardial-infarction-detection-with | 2204.07253 | null | https://arxiv.org/abs/2204.07253v1 | https://arxiv.org/pdf/2204.07253v1.pdf | Early Myocardial Infarction Detection with One-Class Classification over Multi-view Echocardiography | Myocardial infarction (MI) is the leading cause of mortality and morbidity in the world. Early therapeutics of MI can ensure the prevention of further myocardial necrosis. Echocardiography is the fundamental imaging technique that can reveal the earliest sign of MI. However, the scarcity of echocardiographic datasets for the MI detection is the major issue for training data-driven classification algorithms. In this study, we propose a framework for early detection of MI over multi-view echocardiography that leverages one-class classification (OCC) techniques. The OCC techniques are used to train a model for detecting a specific target class using instances from that particular category only. We investigated the usage of uni-modal and multi-modal one-class classification techniques in the proposed framework using the HMC-QU dataset that includes apical 4-chamber (A4C) and apical 2-chamber (A2C) views in a total of 260 echocardiography recordings. Experimental results show that the multi-modal approach achieves a sensitivity level of 85.23% and F1-Score of 80.21%. | ['Moncef Gabbouj', 'Serkan Kiranyaz', 'Fahad Sohrab', 'Aysen Degerli'] | 2022-04-14 | null | null | null | null | ['myocardial-infarction-detection', 'one-class-classification'] | ['medical', 'miscellaneous'] | [ 2.51531869e-01 -4.76730645e-01 -1.63692281e-01 -4.36782129e-02
-1.03042448e+00 -7.69115865e-01 2.23233849e-01 1.46388888e-01
-9.55423713e-02 5.93242168e-01 -8.67076516e-02 -5.90531826e-01
-2.49895334e-01 -5.05024135e-01 -1.65941700e-01 -7.58222401e-01
-4.06061590e-01 4.13425833e-01 1.58620596e-01 5.20314574e-01
3.00785273e-01 4.92119730e-01 -1.24971879e+00 7.35920608e-01
5.64880192e-01 1.05017745e+00 -2.97239181e-02 1.39663899e+00
5.63025475e-01 7.34638393e-01 -5.04843533e-01 1.88312992e-01
4.19489026e-01 -8.05217564e-01 -6.17464244e-01 5.40854558e-02
4.41246599e-01 -3.97967517e-01 2.49795005e-01 3.67889732e-01
9.94721174e-01 -4.58811313e-01 9.81549442e-01 -8.80184114e-01
1.86164528e-01 2.44689912e-01 -5.50933063e-01 9.06857491e-01
-9.25502926e-02 -2.15264410e-02 6.12672329e-01 -1.03275609e+00
5.50901532e-01 5.94360650e-01 6.79998875e-01 3.03047150e-01
-9.60556746e-01 -3.71175885e-01 -5.06294847e-01 1.38151929e-01
-1.30704415e+00 -1.47987247e-01 5.66560626e-01 -8.54703546e-01
4.10375834e-01 5.21576345e-01 5.91207147e-01 3.24036300e-01
4.04701948e-01 4.63259965e-01 1.52307189e+00 -4.17187989e-01
-1.37168303e-01 1.08792722e-01 2.95926541e-01 5.56653321e-01
1.86808914e-01 4.08597142e-01 -3.41246456e-01 -4.06209081e-01
8.37203622e-01 2.45052487e-01 -1.26793116e-01 -2.84058779e-01
-1.40371239e+00 5.07024825e-01 -1.73992768e-01 3.12316775e-01
-4.33801860e-01 -2.93299228e-01 7.05599785e-01 4.21130031e-01
1.30761623e-01 2.01379150e-01 -5.27782023e-01 -2.30199099e-01
-1.17931712e+00 1.93598375e-01 4.23810482e-01 4.21533555e-01
1.15687191e-01 -1.02666110e-01 -2.63832450e-01 7.17170417e-01
1.45639345e-01 6.32243156e-01 4.05407906e-01 -9.90009069e-01
2.67904401e-01 7.35439718e-01 -5.62208220e-02 -7.74578035e-01
-3.31120878e-01 -9.33429539e-01 -1.03281069e+00 4.42853063e-01
4.50907558e-01 -3.22014421e-01 -8.44642341e-01 1.14821315e+00
2.84122795e-01 2.81745702e-01 2.27441907e-01 9.61121023e-01
8.89334679e-01 4.06427056e-01 1.22373335e-01 -5.00166357e-01
1.57760501e+00 -4.66094315e-01 -1.84408113e-01 1.05375879e-01
8.55993271e-01 -8.57893586e-01 4.88086045e-01 4.37146991e-01
-7.76669562e-01 -6.22249484e-01 -1.05435753e+00 6.36529922e-01
1.15317829e-01 7.13392377e-01 1.38332799e-01 7.58317769e-01
-6.31794572e-01 4.14032698e-01 -5.96218526e-01 -3.78779531e-01
3.47271800e-01 1.86707214e-01 -4.31817800e-01 -9.90075245e-03
-8.83113027e-01 7.42863476e-01 2.79561192e-01 5.49712665e-02
-8.48167658e-01 -8.44597220e-01 -2.00352103e-01 -1.81732923e-01
1.01096839e-01 -7.76339948e-01 5.20979881e-01 -5.52795112e-01
-8.02638590e-01 1.23753798e+00 -2.14948997e-01 -3.68306369e-01
5.33232987e-01 -2.23146006e-01 -2.61982501e-01 6.37041867e-01
2.38446921e-01 2.24640191e-01 5.54365516e-01 -1.26331806e+00
-9.41935241e-01 -4.81823713e-01 -3.09868846e-02 2.16109440e-01
1.92921653e-01 1.18323267e-01 2.04327658e-01 -7.29603350e-01
4.92784023e-01 -1.08682156e+00 -5.66021651e-02 -4.47722733e-01
-4.16846305e-01 1.50797099e-01 8.39231431e-01 -7.40276039e-01
1.50170898e+00 -2.08822513e+00 5.06761782e-02 2.92383671e-01
6.04636908e-01 4.14702952e-01 5.81031382e-01 2.66518056e-01
-1.32252604e-01 4.88255709e-01 -1.56434461e-01 1.51062503e-01
-6.77211165e-01 -1.06847659e-01 1.64940253e-01 3.39501083e-01
-1.59553707e-01 7.05216765e-01 -4.42128360e-01 -8.35993111e-01
3.58557343e-01 1.16515614e-01 -3.57060939e-01 2.90973902e-01
6.56776905e-01 8.97606373e-01 -4.39290792e-01 8.47109199e-01
5.19902766e-01 -3.33018005e-01 5.27913451e-01 -3.90953720e-01
-9.30562764e-02 -3.56442422e-01 -1.16174972e+00 1.19412374e+00
-1.08061684e-02 3.09416592e-01 -3.87613147e-01 -1.16459918e+00
7.41563320e-01 9.35336828e-01 8.35674822e-01 -2.24798039e-01
1.13446139e-01 4.71144170e-01 4.49028313e-01 -8.36113751e-01
-5.57932556e-01 -3.67747784e-01 1.22515962e-01 3.91737908e-01
-1.91010952e-01 4.57605034e-01 2.74978966e-01 2.23565428e-03
1.04044366e+00 5.51815480e-02 6.85189784e-01 -3.60451818e-01
9.04716253e-01 -1.58120357e-02 9.16372120e-01 1.04588759e+00
-5.70153296e-01 1.02760279e+00 4.85207468e-01 -8.44338834e-01
-8.37506592e-01 -1.05802941e+00 -4.08323318e-01 3.78041983e-01
-2.08366603e-01 -1.20235547e-01 -4.06217635e-01 -6.95769966e-01
-1.11782111e-01 -7.36649483e-02 -4.13733006e-01 6.40062317e-02
-8.24350595e-01 -1.08707249e+00 5.95898986e-01 6.48847163e-01
4.83601868e-01 -7.19819546e-01 -1.36728263e+00 9.10073221e-02
-4.20025557e-01 -8.13301742e-01 -4.33195606e-02 -7.40722865e-02
-1.63755965e+00 -1.48647702e+00 -1.16727102e+00 -8.76531959e-01
3.89043570e-01 -7.03270454e-03 1.24278903e+00 2.82590896e-01
-5.87372065e-01 1.01102024e-01 -3.91232640e-01 -2.95244783e-01
-6.04692817e-01 4.05437015e-02 -1.85728267e-01 1.68175325e-01
8.10137987e-02 -5.47460973e-01 -1.14270675e+00 3.30660015e-01
-3.05311292e-01 4.13112640e-02 8.70599389e-01 7.39794850e-01
6.29053712e-01 -3.36811572e-01 9.66928899e-01 -1.26586914e+00
9.28851962e-02 -4.52208400e-01 -1.50329262e-01 2.86186188e-01
-8.51499498e-01 -8.05352211e-01 2.35852078e-01 -2.64598280e-01
-5.69978178e-01 1.57929033e-01 2.31177762e-01 -3.32533926e-01
-4.36188877e-01 7.42088556e-01 2.60127991e-01 5.90040535e-02
6.85862839e-01 2.47947082e-01 -3.63377035e-02 -3.95844430e-01
-3.43121171e-01 7.05432236e-01 4.83690768e-01 -3.36396605e-01
4.12495404e-01 2.61841893e-01 5.38168848e-01 -7.60759294e-01
-6.05691135e-01 -8.87214720e-01 -7.69721925e-01 -5.30436873e-01
8.53874683e-01 -8.93085599e-01 -4.80964094e-01 5.08096337e-01
-7.95262516e-01 2.26220816e-01 1.57657772e-01 7.67897487e-01
-3.58406991e-01 4.19148326e-01 -5.46195924e-01 -8.88861954e-01
-6.76978409e-01 -1.04262841e+00 8.76695514e-01 3.53894308e-02
-1.86118364e-01 -8.61840963e-01 9.95544195e-02 5.41948974e-01
2.99590498e-01 7.43723631e-01 1.21001303e+00 -7.22168684e-01
-5.21661937e-01 -5.27003288e-01 -6.54602498e-02 4.60522622e-01
9.81093869e-02 -1.57548025e-01 -9.32600856e-01 -1.92842573e-01
2.34485250e-02 1.09294802e-01 5.32934129e-01 9.27273631e-01
7.03618824e-01 3.37061435e-01 -3.96206021e-01 4.46043462e-01
1.58026350e+00 5.89460850e-01 6.72784090e-01 3.27948123e-01
6.43093646e-01 5.45367002e-02 7.14939594e-01 3.71669471e-01
2.23021522e-01 3.93379211e-01 9.69444886e-02 -3.89264315e-01
-1.90653726e-01 3.63461852e-01 -2.00393036e-01 7.15615034e-01
-5.66173971e-01 -1.20731806e-02 -1.30519080e+00 5.76607347e-01
-1.60787213e+00 -1.01550090e+00 -4.92069274e-01 2.37948895e+00
5.48484862e-01 1.50534198e-01 2.59312838e-01 6.66040301e-01
6.95551455e-01 -3.27149779e-01 -5.13548106e-02 -6.72922134e-02
1.52378688e-02 1.66346505e-01 2.80082494e-01 2.14280248e-01
-1.50236666e+00 4.88850772e-02 6.24503565e+00 2.33415529e-01
-1.29004872e+00 4.88640778e-02 1.06672537e+00 3.86372298e-01
4.64736521e-01 1.25974134e-01 -5.58116198e-01 2.79852837e-01
9.33573902e-01 -6.01128116e-03 -3.38959247e-01 4.48717147e-01
3.59816730e-01 -3.95660698e-01 -1.07517469e+00 1.01836896e+00
1.58999655e-02 -1.48197985e+00 -5.73453419e-02 -3.21829356e-02
5.84840953e-01 -2.40478426e-01 -2.39387542e-01 1.51667759e-01
-1.06617212e+00 -7.78071344e-01 2.13074401e-01 7.40876734e-01
1.20626152e+00 -4.91915911e-01 1.20392239e+00 4.97207582e-01
-1.15777016e+00 -1.39837682e-01 2.16360271e-01 5.60873300e-02
8.19248632e-02 4.92040962e-01 -8.29312027e-01 6.62476659e-01
5.54839253e-01 8.33583653e-01 -4.66143817e-01 1.20478261e+00
5.61584234e-01 1.06837893e+00 -1.64644301e-01 5.73879898e-01
-3.41908783e-01 -6.03278242e-02 8.47468436e-01 9.45117116e-01
3.45330894e-01 2.06039965e-01 6.07525706e-01 2.21998245e-01
3.19862247e-01 2.82502025e-01 -6.14543855e-01 2.83066928e-01
3.62476647e-01 1.24933064e+00 -1.00063455e+00 -6.55587375e-01
-3.78976852e-01 3.51037025e-01 -5.26804566e-01 2.29293361e-01
-6.24377549e-01 -1.52142465e-01 -2.23402143e-01 4.95398194e-01
7.59613588e-02 1.04944654e-01 -7.45845199e-01 -8.90981853e-01
-5.75004006e-03 -1.14530361e+00 7.32751608e-01 -2.55518466e-01
-8.24429512e-01 5.16653121e-01 1.09327219e-01 -1.67895210e+00
-2.61213958e-01 -4.91015077e-01 -7.05021501e-01 1.16242790e+00
-1.39881778e+00 -1.19451988e+00 -3.08450401e-01 3.66674140e-02
5.17122030e-01 -3.81059945e-01 1.26512814e+00 6.83974624e-01
-3.31113398e-01 3.43614727e-01 -2.58822680e-01 2.59434670e-01
6.71892107e-01 -1.48094153e+00 -4.73388851e-01 8.61483157e-01
-2.98789591e-01 6.80282295e-01 4.69064385e-01 -5.66881537e-01
-9.36681390e-01 -8.79059911e-01 1.06277823e+00 -5.33270776e-01
-1.30457088e-01 3.94867152e-01 -5.11728287e-01 3.38653922e-01
-2.24551260e-01 3.53974700e-01 1.00395453e+00 -1.44977584e-01
2.77310580e-01 -2.12483600e-01 -1.06158495e+00 8.34428743e-02
4.00390983e-01 -3.77793849e-01 -2.40502551e-01 -2.29483135e-02
-4.70696360e-01 -3.88385892e-01 -1.38194001e+00 9.46030319e-01
1.09428108e+00 -1.28240740e+00 1.08934247e+00 -5.19748092e-01
4.51211959e-01 -4.40380841e-01 -1.52046725e-01 -5.56196988e-01
-1.05069235e-01 -2.41646811e-01 -1.82072699e-01 9.40598130e-01
4.52795595e-01 -4.28463757e-01 7.99735248e-01 -1.88876893e-02
-5.24924099e-02 -1.10566700e+00 -7.78478265e-01 -3.06120694e-01
2.13409916e-01 -2.65653338e-02 -2.86420226e-01 5.93642354e-01
-3.58516365e-01 1.95780069e-01 -3.88471007e-01 2.43585378e-01
7.95005500e-01 4.57879245e-01 5.99389315e-01 -1.54138267e+00
-2.35852137e-01 9.46381036e-03 -7.12416530e-01 -3.66648734e-01
-6.78564191e-01 -8.87543559e-01 -3.51039648e-01 -1.38757229e+00
7.23391533e-01 -6.29559457e-01 -6.61760628e-01 -1.12971067e-01
-5.21879137e-01 6.38351083e-01 9.91341397e-02 6.76524162e-01
-6.22865297e-02 -4.75773066e-01 1.16110039e+00 3.26904505e-01
-1.26682445e-01 4.66035873e-01 -3.80248964e-01 7.32141376e-01
1.00448382e+00 -6.94346547e-01 -6.45360410e-01 1.58673242e-01
-3.28819789e-02 7.04832852e-01 7.15735018e-01 -1.30528092e+00
-2.63542026e-01 3.79054807e-02 5.79821050e-01 -9.38544810e-01
-2.47652503e-03 -5.89627564e-01 1.83645606e-01 8.01052511e-01
-2.61622429e-01 2.23180279e-01 -9.70481113e-02 5.12147784e-01
-2.27880925e-01 1.74012370e-02 8.78912449e-01 -4.47105438e-01
-5.55025399e-01 1.46267653e-01 -6.34605289e-01 4.22760367e-01
1.11390245e+00 -4.14523661e-01 6.73758462e-02 -3.76297836e-03
-1.19255149e+00 4.49781343e-02 -1.40095606e-01 8.40720683e-02
7.14082956e-01 -9.82254863e-01 -1.13214719e+00 4.50629741e-02
1.91911399e-01 -3.51447374e-01 3.30603689e-01 1.61808765e+00
-8.39554489e-01 3.91967773e-01 -2.99791843e-01 -1.25441480e+00
-1.78721857e+00 9.67355371e-02 7.73434341e-01 -4.50550556e-01
-9.60137188e-01 4.58027929e-01 1.84827715e-01 -1.24178920e-03
-1.14183478e-01 -2.20303722e-02 -6.32208586e-01 -1.46347106e-01
3.78669530e-01 6.99828327e-01 -1.86908375e-02 -5.35098016e-01
-6.36762977e-01 7.15107143e-01 3.74851227e-02 1.95665937e-02
8.26907992e-01 -2.73797572e-01 2.09333189e-02 8.62339973e-01
9.34871495e-01 -2.06700340e-01 -4.86830801e-01 1.28726393e-01
4.43864688e-02 -3.28331172e-01 1.59856454e-01 -1.11722326e+00
-8.90954256e-01 1.05872953e+00 1.37824452e+00 1.53215239e-02
1.09543478e+00 -3.21486831e-01 5.08872986e-01 -1.86075270e-01
1.09842226e-01 -7.08397388e-01 -2.42335364e-01 -1.09429315e-01
4.05425608e-01 -1.25406456e+00 9.07415077e-02 -5.36880016e-01
-7.41383791e-01 1.35777748e+00 1.92151472e-01 -1.45366430e-01
9.93860543e-01 5.62008955e-02 6.93162084e-01 -3.22402567e-01
-5.96877754e-01 4.22864556e-02 3.85300905e-01 3.96469980e-01
9.66646969e-01 2.65600860e-01 -6.81812525e-01 4.97676253e-01
5.94789386e-01 4.59839791e-01 4.27201957e-01 1.27097070e+00
-4.63529795e-01 -9.66221035e-01 -1.77560508e-01 8.23542774e-01
-1.13583910e+00 -5.54527761e-03 8.24309662e-02 6.06128454e-01
5.01980841e-01 9.64290857e-01 -4.86713439e-01 -2.38019392e-01
2.59565562e-01 5.26478350e-01 4.04777527e-01 -5.68864346e-01
-6.51893616e-01 4.35356706e-01 3.12758610e-02 -5.80942072e-02
-7.88079262e-01 -7.72096276e-01 -1.07192814e+00 1.93429708e-01
-1.17455252e-01 -3.07301451e-02 2.31331900e-01 1.08526742e+00
4.44829792e-01 7.04785168e-01 8.04017484e-01 -3.43970567e-01
-4.85222340e-01 -9.79761183e-01 -6.89797103e-01 4.88729388e-01
4.64390278e-01 -5.72152734e-01 -3.75611931e-01 6.73707366e-01] | [14.20072078704834, -2.4072790145874023] |
5995b78b-e8b5-4131-8861-d85e14cf84d9 | boundary-aware-u-net-for-glacier-segmentation-1 | 2301.11454 | null | https://arxiv.org/abs/2301.11454v1 | https://arxiv.org/pdf/2301.11454v1.pdf | Boundary Aware U-Net for Glacier Segmentation | Large-scale study of glaciers improves our understanding of global glacier change and is imperative for monitoring the ecological environment, preventing disasters, and studying the effects of global climate change. Glaciers in the Hindu Kush Himalaya (HKH) are particularly interesting as the HKH is one of the world's most sensitive regions for climate change. In this work, we: (1) propose a modified version of the U-Net for large-scale, spatially non-overlapping, clean glacial ice, and debris-covered glacial ice segmentation; (2) introduce a novel self-learning boundary-aware loss to improve debris-covered glacial ice segmentation performance; and (3) propose a feature-wise saliency score to understand the contribution of each feature in the multispectral Landsat 7 imagery for glacier mapping. Our results show that the debris-covered glacial ice segmentation model trained using self-learning boundary-aware loss outperformed the model trained using dice loss. Furthermore, we conclude that red, shortwave infrared, and near-infrared bands have the highest contribution toward debris-covered glacial ice segmentation from Landsat 7 images. | ['Olac Fuentes', 'Sergio A. Vargas Zesati', 'Katie E. Miles', 'Bibek Aryal'] | 2023-01-26 | boundary-aware-u-net-for-glacier-segmentation | https://septentrio.uit.no/index.php/nldl/article/view/6789 | https://septentrio.uit.no/index.php/nldl/article/view/6789/7028 | proceedings-of-the-northern-lights-deep | ['self-learning'] | ['natural-language-processing'] | [ 1.38895065e-02 -1.71048015e-01 8.03173184e-02 -3.30778629e-01
-7.27357924e-01 -5.36219895e-01 2.53016174e-01 1.07177563e-01
-5.26224196e-01 9.02785838e-01 3.69363368e-01 -4.95502889e-01
1.72571048e-01 -1.04121494e+00 -8.14789593e-01 -8.30548882e-01
-8.78336489e-01 1.92960516e-01 1.73067357e-02 -5.49457312e-01
-1.03314240e-02 2.97690451e-01 -1.41931796e+00 -3.27486157e-01
1.70265007e+00 6.35454893e-01 7.95194447e-01 4.60243970e-01
1.56263903e-01 1.14446282e-01 -1.64783984e-01 5.16605318e-01
6.53134465e-01 -5.52475631e-01 -7.21743762e-01 -3.69981006e-02
7.79728532e-01 -5.04223883e-01 1.25785977e-01 1.24069726e+00
2.66195744e-01 1.16452754e-01 5.20834446e-01 -4.56257969e-01
-3.14385355e-01 5.59808612e-01 -8.14600587e-01 2.77594656e-01
-4.34300840e-01 2.72392333e-01 1.04035664e+00 -4.55912381e-01
6.71153784e-01 1.29572749e+00 6.75797760e-01 6.74127266e-02
-1.10685682e+00 -6.59407079e-01 5.57961762e-01 -1.21065438e-01
-1.02902651e+00 -5.65778054e-02 6.32082224e-01 -4.79565024e-01
5.57882547e-01 4.33947831e-01 9.55448091e-01 1.69793516e-01
4.25600737e-01 9.20483232e-01 1.32621658e+00 -1.93634093e-01
3.01216215e-01 -7.63577223e-01 1.57760918e-01 5.43803573e-01
5.72980046e-01 2.87581742e-01 -1.23536482e-01 -1.52332159e-02
7.16002226e-01 1.03521191e-01 -5.30895829e-01 1.14015276e-02
-1.09497476e+00 7.98945010e-01 1.22703254e+00 8.14850032e-02
-5.13470471e-01 1.48963273e-01 2.16077805e-01 3.06157976e-01
1.15106571e+00 5.72546363e-01 -4.88050014e-01 3.73662800e-01
-1.43193388e+00 6.16090834e-01 2.85517752e-01 3.69838566e-01
1.21776187e+00 1.95705533e-01 2.83566505e-01 8.00085783e-01
4.96431202e-01 1.18379354e+00 5.64378798e-02 -7.86362350e-01
3.31430405e-01 4.86200631e-01 1.11951843e-01 -5.92215300e-01
-5.58985412e-01 -6.72244132e-01 -1.04428184e+00 7.73456335e-01
1.55720472e-01 -3.42307061e-01 -1.22872531e+00 1.37109613e+00
2.96100080e-02 -9.88980085e-02 2.07938805e-01 1.43025851e+00
3.51707578e-01 6.72371686e-01 4.29675132e-01 -4.92403917e-02
1.34235060e+00 -5.19499660e-01 -4.78463709e-01 -7.64456153e-01
5.62588036e-01 -1.45273730e-01 1.10614276e+00 -4.58906829e-01
-5.03956139e-01 -2.49241486e-01 -1.16867995e+00 1.62043914e-01
-1.87272325e-01 9.02669579e-02 6.75625861e-01 2.49615356e-01
-1.00564849e+00 6.22501671e-01 -8.84386718e-01 -4.22458380e-01
6.97538197e-01 -1.87936425e-01 -1.99020561e-02 5.74844480e-02
-1.34200895e+00 7.40377545e-01 2.55772531e-01 6.02749586e-01
-8.30810428e-01 -6.76914155e-01 -1.07585084e+00 7.10795969e-02
1.52503047e-02 -3.34211051e-01 4.81336176e-01 -1.22992671e+00
-4.80071098e-01 9.17136729e-01 4.40390296e-02 -7.25195885e-01
4.02845085e-01 -5.51133752e-01 8.59370381e-02 -4.06990051e-02
6.22022748e-01 1.00023293e+00 4.33647454e-01 -1.57484710e+00
-9.10405576e-01 -7.57450938e-01 -3.37093681e-01 7.38076091e-01
2.37278014e-01 -2.40222096e-01 4.15979445e-01 -9.47284102e-01
5.40743232e-01 -9.01371479e-01 -4.75919813e-01 -1.60403177e-02
3.24362665e-02 2.70247757e-01 8.39179456e-01 -1.18511498e+00
8.48912001e-01 -2.21957493e+00 -1.73759893e-01 -5.94693124e-02
6.39644116e-02 3.89200300e-01 -1.23179061e-02 8.98747221e-02
3.29084456e-01 3.78215641e-01 -1.31281805e+00 2.99422801e-01
-4.39634293e-01 1.47097409e-01 -4.44488734e-01 4.21578646e-01
1.73480779e-01 7.25079417e-01 -1.03883231e+00 -2.26227239e-01
2.63793111e-01 8.10242370e-02 1.69427320e-02 -1.40337765e-01
-4.99722898e-01 5.19100785e-01 -4.64245558e-01 9.08163249e-01
1.46687734e+00 4.21404600e-01 -7.37401843e-02 1.45574704e-01
-5.78808546e-01 -1.74144469e-02 -5.86537361e-01 1.49191737e+00
-1.80233061e-01 7.67178833e-01 4.88771796e-01 -3.43570411e-01
1.27234328e+00 -4.14683461e-01 -3.09062731e-02 -9.29256439e-01
-4.59318191e-01 4.74123925e-01 -1.51704371e-01 -3.96878839e-01
7.61512756e-01 -4.62740242e-01 3.78970541e-02 2.65217602e-01
-6.18516743e-01 -1.31031111e-01 -1.53615788e-01 -1.21796820e-02
5.69980323e-01 2.79881418e-01 -1.61522999e-02 -9.11595583e-01
8.93930048e-02 3.47125351e-01 8.50810885e-01 5.64759612e-01
-5.71193695e-01 7.39283800e-01 5.03021106e-02 -1.02149856e+00
-7.96126902e-01 -1.21832585e+00 -2.97293633e-01 9.22437966e-01
4.13782954e-01 5.22462547e-01 -4.81836408e-01 -5.19432783e-01
3.57050955e-01 5.27882814e-01 -3.85202110e-01 1.63200080e-01
-6.07339144e-01 -1.35087633e+00 6.19520247e-01 4.85430568e-01
1.36081743e+00 -1.15346968e+00 -9.37946916e-01 9.31226015e-02
-6.67696893e-01 -4.47443873e-01 -3.58463317e-01 3.20190221e-01
-1.45897508e+00 -1.06683624e+00 -9.82564092e-01 -1.03892922e+00
7.03372478e-01 4.29182380e-01 1.10126436e+00 -1.84967682e-01
-1.91909477e-01 -7.03232214e-02 -4.27950919e-01 -2.57151246e-01
-9.92755890e-02 1.42502964e-01 -3.27744335e-01 -4.33505535e-01
3.31356496e-01 -4.17579174e-01 -7.83561587e-01 1.01603702e-01
-7.87277758e-01 1.28457531e-01 7.03936875e-01 6.08813643e-01
4.70447153e-01 -8.44762400e-02 6.41104102e-01 -6.78808510e-01
2.91824311e-01 -4.07982141e-01 -5.78680515e-01 2.06884041e-01
-5.70136607e-01 9.05335471e-02 9.80604813e-02 4.94471788e-01
-1.22340083e+00 2.63366103e-01 5.39740473e-02 3.09344590e-01
-1.60426170e-01 8.03808451e-01 -3.87319215e-02 -4.80486304e-02
7.34644294e-01 2.76179969e-01 2.89668664e-02 -5.02655923e-01
1.89912081e-01 8.12560260e-01 7.18712270e-01 -2.01887101e-01
7.01844037e-01 8.47469509e-01 1.14806201e-02 -1.31179392e+00
-5.11651576e-01 -5.98285973e-01 -4.23066616e-01 -3.11365157e-01
8.57116163e-01 -1.51685238e+00 -1.52345955e-01 1.00953925e+00
-6.48107350e-01 -5.93410373e-01 -2.56849766e-01 5.75653613e-01
-2.62811303e-01 4.37212586e-01 -1.70764983e-01 -9.34124470e-01
-7.98967242e-01 -5.63234627e-01 9.02340114e-01 3.63212466e-01
3.17135364e-01 -9.24629271e-01 3.91359746e-01 2.50495523e-01
5.31431913e-01 6.40003324e-01 7.61671185e-01 2.31095001e-01
-7.96580553e-01 2.84648240e-01 -5.39608896e-01 5.91622293e-01
5.11646330e-01 -3.58059704e-01 -8.39280844e-01 -4.40882355e-01
-3.92626300e-02 -6.14916012e-02 2.09129786e+00 1.13806951e+00
3.43376905e-01 -2.19262823e-01 -1.58622608e-01 9.96375382e-01
1.56460392e+00 3.54937948e-02 8.97732675e-01 8.07084560e-01
6.43595994e-01 6.63780689e-01 8.96273971e-01 2.75656283e-01
5.29753923e-01 -8.95413086e-02 6.73196614e-01 -6.46372795e-01
-1.04381360e-01 -5.29864132e-02 4.65704590e-01 6.11493923e-02
-1.47598907e-01 1.99336261e-01 -1.09465683e+00 1.07047915e+00
-1.81998181e+00 -9.72463787e-01 -3.37782651e-01 2.35708523e+00
6.01796150e-01 -3.94812822e-01 -1.75111994e-01 -3.14413458e-01
7.91725218e-01 7.27454364e-01 -7.04080999e-01 1.92085296e-01
-8.44188094e-01 4.61340509e-02 1.33117473e+00 7.85432756e-01
-1.75174034e+00 1.25460935e+00 5.70271015e+00 2.29179382e-01
-1.21582913e+00 -1.83433324e-01 9.51488972e-01 2.31874406e-01
-5.97580969e-01 3.67565930e-01 -5.55381835e-01 3.62568736e-01
4.87643301e-01 1.01187736e-01 1.50890082e-01 4.63283539e-01
7.33241558e-01 -7.76795447e-01 -6.99662417e-02 3.26068997e-01
-2.38331005e-01 -1.16973519e+00 -2.14851107e-02 -1.65463556e-02
8.15244913e-01 6.66608214e-01 -1.87674507e-01 -1.30603164e-01
4.48164791e-01 -6.88231111e-01 5.21512032e-01 5.68587661e-01
7.01307476e-01 -5.85659742e-01 6.55688822e-01 7.58672431e-02
-1.32836092e+00 -2.06476599e-02 -3.97883147e-01 -3.36501628e-01
6.30315170e-02 9.88714159e-01 -3.72781694e-01 4.25677210e-01
9.33668554e-01 1.13142264e+00 -4.94648099e-01 1.28363490e+00
-3.33624154e-01 9.80375886e-01 -3.40442270e-01 4.06482309e-01
5.83355248e-01 -5.49242198e-01 9.22512472e-01 7.99063325e-01
1.02555938e-01 -2.56865611e-03 2.87629664e-01 8.00402224e-01
-3.97560783e-02 1.44561874e-02 -4.95606333e-01 4.06432509e-01
2.05344513e-01 8.67720127e-01 -6.88820004e-01 4.82079526e-03
9.55805853e-02 7.54662454e-01 -1.44792795e-01 4.61259186e-01
-1.69961855e-01 -4.74001169e-01 1.04362428e+00 1.76669538e-01
2.08631217e-01 -4.13725227e-01 -6.21452332e-01 -9.91185248e-01
2.27937713e-01 -2.70082176e-01 2.61951655e-01 -6.37584507e-01
-7.88367987e-01 2.54820555e-01 -1.23249277e-01 -1.35489798e+00
2.31089085e-01 -9.85255018e-02 -7.55249798e-01 1.24350560e+00
-2.47208524e+00 -1.27030170e+00 -5.68906486e-01 -2.44289368e-01
5.14487684e-01 1.32071063e-01 5.61624944e-01 -1.21279731e-01
-4.32756403e-03 -3.16987842e-01 8.87183070e-01 -2.59274784e-02
7.19753385e-01 -1.37202358e+00 9.03847039e-01 1.09044361e+00
-6.23583674e-01 -8.63876045e-02 5.94545126e-01 -1.29169178e+00
-8.17949355e-01 -1.72459829e+00 8.41638207e-01 4.97136235e-01
3.57335694e-02 1.03425421e-03 -1.20700181e+00 4.74192888e-01
-1.96675256e-01 -7.52845556e-02 1.99785039e-01 2.02242285e-02
-6.39112517e-02 -5.71401656e-01 -1.37418389e+00 2.59748757e-01
7.15616882e-01 -6.18613288e-02 -4.51303333e-01 2.48331651e-01
4.15326715e-01 -4.60953526e-02 -3.31210971e-01 4.50465709e-01
6.97147548e-01 -8.98855388e-01 6.44529164e-01 1.41188771e-01
5.45295298e-01 -5.45778334e-01 -1.55842602e-01 -1.60812128e+00
-4.92796481e-01 -1.42819926e-01 8.31132650e-01 7.53691196e-01
1.50785670e-01 -6.22841597e-01 6.69506311e-01 1.05098531e-01
-5.01729190e-01 -1.27426572e-02 -9.44624841e-01 -9.33523297e-01
4.00782108e-01 1.75017342e-01 4.77402240e-01 6.72696233e-01
-4.22997981e-01 -2.82366991e-01 -3.50034952e-01 4.94830638e-01
8.81909013e-01 3.90438765e-01 2.79496312e-01 -1.67765760e+00
6.22916400e-01 -3.02377373e-01 -3.35576564e-01 -8.97015810e-01
-6.31528487e-03 -7.24895775e-01 5.66508412e-01 -1.79795873e+00
4.04787399e-02 -7.63559818e-01 -6.03955574e-02 7.45210230e-01
-5.88819742e-01 1.40580669e-01 4.06456590e-02 7.38702357e-01
-9.22122132e-03 1.01342380e+00 1.16129136e+00 -4.32500362e-01
-4.99340951e-01 -8.47981647e-02 -6.03020370e-01 5.18154263e-01
8.33408594e-01 -3.98015290e-01 1.68701857e-01 -5.73893428e-01
3.03910580e-02 -2.88807124e-01 4.40329075e-01 -1.26175129e+00
-1.51377633e-01 -4.51313794e-01 2.45353818e-01 -8.38605702e-01
-3.78840357e-01 -5.88659525e-01 -2.65892923e-01 7.89852023e-01
5.30844030e-04 -6.71382844e-01 2.51463920e-01 6.06302679e-01
-3.82456869e-01 1.19494110e-01 9.29285705e-01 -3.91490072e-01
-1.19409227e+00 1.21596240e-01 -6.07970357e-01 5.55174872e-02
7.26321161e-01 -6.72286078e-02 -5.01733959e-01 -1.30941689e-01
-2.45628953e-01 8.85655880e-01 7.68773615e-01 3.06242913e-01
4.68979985e-01 -6.47728562e-01 -1.29675090e+00 5.89112759e-01
2.14824125e-01 3.60996574e-01 6.05210721e-01 3.61029536e-01
-1.16774404e+00 2.15058830e-02 -4.80388969e-01 -5.24998486e-01
-1.05839419e+00 -4.01042283e-01 5.68157077e-01 -3.53087112e-02
-8.10246468e-01 6.40843928e-01 2.94807166e-01 -6.61840618e-01
-2.41113096e-01 -6.91594958e-01 -6.68400526e-02 2.64393866e-01
2.69904047e-01 3.21552396e-01 -8.77881646e-02 -7.58014262e-01
-3.91122133e-01 5.74378431e-01 2.24959910e-01 2.53088418e-02
1.50279677e+00 -3.44641387e-01 -2.10198745e-01 5.43266416e-01
5.43811679e-01 -6.03403747e-01 -1.81004596e+00 5.92200272e-02
-1.76448449e-01 -6.28714621e-01 6.60551965e-01 -1.09956336e+00
-1.19677401e+00 6.73936248e-01 1.05034518e+00 -3.69704366e-01
1.33947945e+00 -1.88470110e-01 7.96714127e-01 4.61877316e-01
2.52352476e-01 -9.91251409e-01 -8.84589136e-01 5.24767041e-01
8.50510478e-01 -1.52500498e+00 1.76702872e-01 -1.47820543e-02
-7.13262618e-01 1.00197494e+00 3.60221922e-01 -3.06570411e-01
6.15042210e-01 -8.13752562e-02 5.40185809e-01 -1.42533496e-01
-7.02171698e-02 -6.09951079e-01 -2.47407794e-01 6.17135763e-01
2.96248615e-01 6.82614386e-01 -4.60847944e-01 3.11695486e-02
9.23195854e-03 -6.03772625e-02 5.20749927e-01 1.06435311e+00
-1.13625181e+00 -6.69701815e-01 -7.36757159e-01 9.93200481e-01
2.71530271e-01 -4.41234559e-01 -9.94785428e-02 2.29281828e-01
-1.01628035e-01 5.33864260e-01 3.47090065e-01 -3.54912244e-02
-2.00670958e-01 -1.63321301e-01 -7.39454031e-02 -3.36308420e-01
-5.48560083e-01 3.88197065e-03 -2.31324211e-01 2.19562743e-02
-6.98611438e-01 -9.37454641e-01 -1.22106600e+00 -1.88851729e-01
-1.99448988e-01 1.51374578e-01 8.65903676e-01 7.79069006e-01
3.69910866e-01 1.82565488e-02 7.98877001e-01 -9.91084516e-01
-8.73292089e-02 -1.29137230e+00 -1.26966882e+00 9.28330645e-02
7.14427114e-01 -3.67354870e-01 -6.15413725e-01 3.60754952e-02] | [9.578673362731934, -1.4993782043457031] |
107804c0-9a4b-4cea-984f-ab89f92f9030 | learn-to-understand-negation-in-video | 2205.00132 | null | https://arxiv.org/abs/2205.00132v2 | https://arxiv.org/pdf/2205.00132v2.pdf | Learn to Understand Negation in Video Retrieval | Negation is a common linguistic skill that allows human to express what we do NOT want. Naturally, one might expect video retrieval to support natural-language queries with negation, e.g., finding shots of kids sitting on the floor and not playing with a dog. However, the state-of-the-art deep learning based video retrieval models lack such ability, as they are typically trained on video description datasets such as MSR-VTT and VATEX that lack negated descriptions. Their retrieved results basically ignore the negator in the sample query, incorrectly returning videos showing kids playing with dog. This paper presents the first study on learning to understand negation in video retrieval and make contributions as follows. By re-purposing two existing datasets (MSR-VTT and VATEX), we propose a new evaluation protocol for video retrieval with negation. We propose a learning based method for training a negation-aware video retrieval model. The key idea is to first construct a soft negative caption for a specific training video by partially negating its original caption, and then compute a bidirectionally constrained loss on the triplet. This auxiliary loss is weightedly added to a standard retrieval loss. Experiments on the re-purposed benchmarks show that re-training the CLIP (Contrastive Language-Image Pre-Training) model by the proposed method clearly improves its ability to handle queries with negation. In addition, the model performance on the original benchmarks is also improved. | ['Xirong Li', 'Fan Hu', 'Aozhu Chen', 'Ziyue Wang'] | 2022-04-30 | null | null | null | null | ['video-description'] | ['computer-vision'] | [ 6.31148368e-02 -4.35009211e-01 -5.45464277e-01 -4.95154172e-01
-7.04094470e-01 -5.99607944e-01 5.15731454e-01 -5.29082455e-02
-6.60173535e-01 4.27098989e-01 1.63918659e-01 -5.86813092e-02
1.59070596e-01 -6.58559501e-01 -1.14779341e+00 -4.77838278e-01
9.96471867e-02 3.59405994e-01 1.74807355e-01 -3.64619553e-01
-6.77903965e-02 1.57460645e-01 -1.80424774e+00 8.42087150e-01
4.04777050e-01 1.14242423e+00 1.81559786e-01 5.64892292e-01
-1.02520240e-02 1.44846880e+00 -6.04467034e-01 -6.53463066e-01
2.98193216e-01 -3.53965074e-01 -7.27862418e-01 -6.25330163e-03
1.11805189e+00 -9.19797122e-01 -9.66097236e-01 9.82412636e-01
3.50617588e-01 2.78483421e-01 6.66751564e-01 -1.54808068e+00
-1.07399058e+00 3.96585435e-01 -3.71212989e-01 2.49541983e-01
8.11399102e-01 2.61108994e-01 1.26297438e+00 -8.56150627e-01
7.30937600e-01 1.45610714e+00 2.54374534e-01 8.69194031e-01
-7.17346966e-01 -7.86497414e-01 3.25667202e-01 5.80202222e-01
-1.68184876e+00 -1.43269718e-01 6.09440684e-01 -1.10144660e-01
1.19407260e+00 4.38107163e-01 7.69738913e-01 1.22722852e+00
-3.29391688e-01 1.44822252e+00 5.69052517e-01 -2.53833622e-01
-1.11904442e-01 1.49207368e-01 -6.93559796e-02 7.77849317e-01
-5.66817522e-02 -9.49406251e-02 -8.07862461e-01 -3.03324368e-02
2.16523811e-01 1.89283684e-01 -5.64715028e-01 -5.32277405e-01
-1.17986631e+00 6.68596864e-01 3.43344748e-01 2.72428006e-01
-1.16920821e-01 4.53134775e-01 7.69729078e-01 7.78692901e-01
3.67746264e-01 1.14600368e-01 -3.05569470e-01 -7.41919056e-02
-1.12107849e+00 4.91435260e-01 7.32526004e-01 1.27267265e+00
6.27302587e-01 -1.37535051e-01 -6.13843501e-01 6.28819108e-01
1.67462096e-01 9.02747750e-01 5.08474410e-01 -9.94455159e-01
4.82186049e-01 4.55664992e-01 1.65075883e-02 -9.94494379e-01
1.77087694e-01 6.55582696e-02 -5.63404441e-01 -3.37397695e-01
9.10215527e-02 3.93579245e-01 -9.52751577e-01 2.00121117e+00
-1.67121068e-01 3.89156103e-01 3.65707666e-01 1.27289748e+00
1.20475745e+00 7.61333942e-01 9.45233405e-02 -1.11512415e-01
1.16195345e+00 -1.24715340e+00 -9.20299828e-01 -5.56858853e-02
9.40990627e-01 -5.52752018e-01 1.44781387e+00 3.56228411e-01
-1.06141031e+00 -3.71128768e-01 -9.91742969e-01 -4.33935195e-01
-5.31542242e-01 1.22955337e-01 5.09680331e-01 3.18592161e-01
-1.24586344e+00 1.67504221e-01 -4.78416204e-01 -3.37596059e-01
2.29725391e-01 2.18414694e-01 -6.32280946e-01 -5.82195163e-01
-1.61995280e+00 6.85696483e-01 2.91893274e-01 -1.93108320e-02
-1.23619294e+00 -6.40415788e-01 -1.12616026e+00 3.64580423e-01
7.05539167e-01 -7.21176922e-01 1.31225622e+00 -1.38371444e+00
-9.90439475e-01 1.29265714e+00 -1.39078826e-01 -4.83583540e-01
6.37718558e-01 -5.51069856e-01 -3.35387498e-01 7.05136061e-01
1.68827295e-01 1.31594491e+00 1.02566469e+00 -1.03989387e+00
-4.16576535e-01 2.73576961e-03 7.97225058e-01 3.12664956e-01
-7.29998946e-01 6.53898343e-02 -1.12176454e+00 -6.37207448e-01
-3.62463325e-01 -7.79210210e-01 4.75552440e-01 3.37198853e-01
-1.67237580e-01 -4.34882998e-01 1.03962541e+00 -3.19220275e-01
1.37100661e+00 -2.21741581e+00 1.43005371e-01 -6.22028522e-02
7.37968162e-02 5.12055039e-01 -5.62599480e-01 3.15689921e-01
-2.63777852e-01 2.20724940e-01 5.75615987e-02 -3.91008049e-01
1.02993742e-01 3.78322870e-01 -5.34585655e-01 3.41162175e-01
3.30675304e-01 9.81834650e-01 -1.18566489e+00 -6.49992347e-01
1.59530059e-01 6.82382643e-01 -7.96764970e-01 3.95366400e-01
-4.79195476e-01 -2.53993928e-01 -2.11143926e-01 9.16693270e-01
5.16130745e-01 -2.54016817e-01 -6.01954088e-02 -3.94151270e-01
2.99513400e-01 -8.93115252e-02 -8.93691778e-01 1.79642701e+00
-1.83490902e-01 8.44570637e-01 -2.76311915e-02 -1.04174209e+00
4.22761023e-01 4.59100068e-01 3.31912369e-01 -1.03979719e+00
-2.96977200e-02 1.29956841e-01 -4.98818755e-01 -1.00593269e+00
5.94240725e-01 -5.81409112e-02 -5.16028143e-02 3.08094054e-01
2.41900519e-01 -1.19903795e-01 4.80437100e-01 6.94072902e-01
1.15399599e+00 2.18228787e-01 -1.06148541e-01 2.61566520e-01
9.14025903e-01 -1.42697498e-01 3.77069473e-01 1.00924039e+00
-2.39671499e-01 7.09382117e-01 6.74663484e-01 -3.58348161e-01
-7.81167567e-01 -8.43560398e-01 4.19167250e-01 1.36892521e+00
3.06870669e-01 -8.13449144e-01 -6.28899038e-01 -8.72200012e-01
-3.36160734e-02 5.31882107e-01 -5.14203906e-01 -5.71221352e-01
-6.68792903e-01 -8.53863955e-02 7.67891586e-01 4.09936696e-01
5.22777855e-01 -1.14733028e+00 -4.99501497e-01 -4.01342630e-01
-5.85472882e-01 -1.39659929e+00 -7.04592466e-01 -1.67468935e-01
-4.14363086e-01 -1.10313702e+00 -9.46137190e-01 -1.05562377e+00
5.63725293e-01 5.63752174e-01 1.46159673e+00 8.75701904e-01
-1.31448001e-01 1.17841744e+00 -5.87747991e-01 -6.13220036e-02
-1.08125143e-01 -3.15071523e-01 1.28172204e-01 -1.53185636e-01
7.44270682e-01 -1.01534784e-01 -4.95638549e-01 9.08859819e-02
-1.54048669e+00 -2.94967204e-01 4.46036845e-01 1.03574467e+00
7.39717126e-01 -2.40449026e-01 2.95607626e-01 -4.72086281e-01
5.72710872e-01 -5.04568875e-01 -4.46613669e-01 6.84992611e-01
-1.32203296e-01 -5.88374026e-03 4.60301369e-01 -7.80943751e-01
-6.51605308e-01 -1.21557169e-01 4.40168232e-02 -1.32747805e+00
1.65535226e-01 4.40079510e-01 -7.45464340e-02 -8.09429437e-02
2.37843290e-01 3.73861909e-01 -2.34270424e-01 -6.60462826e-02
3.12755078e-01 3.89290452e-01 5.78575850e-01 -6.11731529e-01
6.23363256e-01 5.59627652e-01 -1.22361735e-01 -8.18120420e-01
-9.54581618e-01 -6.82110965e-01 -2.11568922e-01 -3.71085465e-01
8.24247599e-01 -1.15605128e+00 -1.00716805e+00 3.65185261e-01
-1.15140271e+00 -1.17030792e-01 -1.96545824e-01 5.51923573e-01
-7.16142774e-01 6.55313313e-01 -5.33620119e-01 -6.40788198e-01
-3.77011836e-01 -1.09233582e+00 1.27410662e+00 -1.57972008e-01
1.75091147e-01 -7.07458913e-01 -1.74751416e-01 4.70335841e-01
3.47606778e-01 -2.30203956e-01 9.66431260e-01 -8.26538503e-01
-9.26559985e-01 -1.76462114e-01 -2.15298772e-01 9.26100194e-01
-3.23333651e-01 -3.96978892e-02 -8.67060423e-01 -4.79242295e-01
-9.89330858e-02 -1.08482623e+00 1.18192196e+00 5.29638119e-03
1.33843601e+00 -1.79739922e-01 -1.38353631e-01 5.09221792e-01
1.42163432e+00 -7.17524663e-02 6.32906079e-01 3.18364292e-01
4.74354357e-01 3.59770656e-01 8.81060958e-01 4.40713391e-02
3.84400427e-01 5.88455737e-01 6.34703398e-01 8.24018046e-02
-9.17690769e-02 -3.82824630e-01 7.40375459e-01 5.47863066e-01
3.82199824e-01 -8.30301821e-01 -6.51711822e-01 7.30584204e-01
-1.97762275e+00 -1.16360402e+00 3.70528668e-01 1.96416664e+00
7.01461196e-01 3.80975343e-02 -3.14525366e-01 4.13299585e-03
3.92439991e-01 2.73338050e-01 -5.03547966e-01 -1.43827572e-01
-4.83129531e-01 7.96659589e-02 1.81154832e-01 3.40404570e-01
-1.23455906e+00 1.14458656e+00 5.82760620e+00 9.23626542e-01
-1.24171758e+00 -5.60030043e-02 3.64960700e-01 -6.19866908e-01
-3.40371758e-01 -2.12407470e-01 -5.62446594e-01 1.76249713e-01
5.85906923e-01 -1.44253016e-01 4.31025714e-01 6.99359596e-01
-2.57337582e-03 -4.19796072e-02 -1.69573164e+00 1.32748473e+00
8.17148089e-01 -1.14584219e+00 7.37920165e-01 -5.73278427e-01
4.41737950e-01 -5.29922061e-02 2.54121274e-01 8.94531429e-01
-4.30894911e-01 -1.04251850e+00 9.25935686e-01 5.77647090e-01
8.18910003e-01 -5.16286254e-01 8.81617486e-01 3.00465554e-01
-9.17723358e-01 -4.46118116e-02 -3.46433431e-01 -9.85927973e-03
2.71788165e-02 -2.06939294e-03 -5.26315331e-01 2.97032714e-01
1.07469678e+00 8.55942786e-01 -5.27042508e-01 8.84015739e-01
-2.80846477e-01 4.47816849e-01 -3.87586802e-01 -1.00588351e-01
5.60619235e-01 -9.34525020e-03 6.71255708e-01 1.34492493e+00
2.70838737e-01 2.20349446e-01 1.48475438e-01 6.52109027e-01
-6.65674567e-01 1.97861597e-01 -9.26059306e-01 -1.85415924e-01
1.52420714e-01 9.15917397e-01 -1.97643518e-01 -6.56368136e-01
-8.59271646e-01 1.14404058e+00 2.61059731e-01 6.56431317e-01
-1.03813565e+00 -1.93241864e-01 6.37668848e-01 1.82519421e-01
5.46547234e-01 3.58244032e-01 7.67164290e-01 -1.52405488e+00
5.15908718e-01 -1.06132638e+00 5.96457422e-01 -1.38168240e+00
-1.25787926e+00 4.08725291e-01 3.25724036e-01 -1.26318705e+00
-4.55514610e-01 -5.46480000e-01 -1.43382907e-01 3.51479381e-01
-1.86603129e+00 -1.13562131e+00 -1.80688202e-01 8.94336402e-01
5.17400563e-01 -1.43640742e-01 6.36045098e-01 7.42687941e-01
-1.50625229e-01 8.90362561e-01 -2.58938730e-01 3.03124487e-01
1.03890395e+00 -9.82486308e-01 -3.39773357e-01 6.90596759e-01
1.16813302e-01 6.74797773e-01 6.07906342e-01 -4.82688725e-01
-1.55497968e+00 -9.79035974e-01 8.78645062e-01 -2.74685651e-01
6.83102310e-01 -3.70309144e-01 -1.01101351e+00 8.76469791e-01
4.50061977e-01 1.85657322e-01 5.51543772e-01 -5.32938838e-01
-6.99366927e-01 -2.13328212e-01 -1.18728054e+00 5.51566482e-01
9.86671448e-01 -9.67411876e-01 -7.64181852e-01 7.66919494e-01
1.13842344e+00 -4.09410506e-01 -4.55409855e-01 5.68662703e-01
5.40215254e-01 -8.75496209e-01 1.16958630e+00 -1.00701797e+00
5.26226759e-01 -2.37668693e-01 -4.83788997e-01 -8.18487525e-01
3.30908328e-01 -3.19215149e-01 -2.89904416e-01 9.90383565e-01
1.15027517e-01 5.16058784e-03 8.02570283e-01 5.36188900e-01
-9.02927481e-03 -6.56815112e-01 -8.40306997e-01 -8.70706141e-01
-1.11202866e-01 -5.40799320e-01 3.45345557e-01 8.98213267e-01
-1.65802598e-01 1.46870717e-01 -6.53147578e-01 -1.06252208e-01
3.50148112e-01 -1.95256859e-01 5.30834019e-01 -6.33138359e-01
3.21980417e-02 -1.37324154e-01 -6.96151912e-01 -1.47654092e+00
6.80405736e-01 -9.73790646e-01 -9.96851549e-02 -1.32153296e+00
4.70321238e-01 1.36853740e-01 -4.08143520e-01 6.07716501e-01
5.13038039e-02 3.71732116e-01 3.86128128e-01 -6.54768273e-02
-1.37828314e+00 6.10677421e-01 1.20914721e+00 -7.01770604e-01
2.33707562e-01 -3.65748972e-01 -3.69814217e-01 7.58593023e-01
3.27152133e-01 -6.22471154e-01 -6.18798852e-01 -8.22613835e-01
6.52287960e-01 1.14232332e-01 6.65789366e-01 -8.48867178e-01
4.19001043e-01 9.46907848e-02 1.20209649e-01 -8.49714160e-01
7.33553410e-01 -1.06501472e+00 -3.90558273e-01 2.59817660e-01
-6.12039328e-01 3.45906794e-01 1.74542397e-01 6.04923368e-01
-8.04766119e-01 -3.64663184e-01 1.84392542e-01 -2.23287746e-01
-9.73402262e-01 3.13932031e-01 -3.55988413e-01 3.24452549e-01
9.22502637e-01 -4.69571203e-02 -3.67680162e-01 -8.78213882e-01
-5.63295484e-01 6.84979618e-01 4.56025302e-01 6.95130229e-01
1.05668521e+00 -1.43322694e+00 -6.33723915e-01 3.48120704e-02
4.69291925e-01 -2.04472288e-01 2.50583321e-01 5.87758303e-01
-5.63062966e-01 7.42442727e-01 -5.72866667e-03 -6.57485843e-01
-1.57367229e+00 8.64086032e-01 4.01478797e-01 -1.82990998e-01
-3.26099575e-01 1.01229966e+00 2.52142280e-01 -2.87306547e-01
9.31720376e-01 -5.03146172e-01 -1.96305826e-01 2.65722685e-02
6.87850773e-01 -2.78472640e-02 1.24767441e-02 -7.73265541e-01
-4.54816520e-01 4.24751252e-01 -1.81691334e-01 1.31154656e-01
1.20914674e+00 3.50176506e-02 -2.87907928e-01 1.35788321e-01
1.62780917e+00 -2.35692516e-01 -6.36827886e-01 -3.24850976e-01
-2.46750340e-01 -5.26619494e-01 3.37191625e-03 -7.32598245e-01
-1.37386870e+00 9.26586449e-01 5.55761635e-01 -1.16524741e-01
1.42107797e+00 9.89845470e-02 9.51910973e-01 1.27440834e+00
1.33634418e-01 -9.87386584e-01 5.45784652e-01 6.49222732e-01
9.54107702e-01 -1.43062901e+00 -1.01416498e-01 -1.24649070e-01
-6.01646185e-01 9.36854720e-01 8.89933050e-01 2.02157386e-02
4.54029649e-01 -5.58682382e-02 -2.06858646e-02 -4.17417973e-01
-1.18823981e+00 -1.74095586e-01 3.09065849e-01 3.63947868e-01
3.59037042e-01 -2.25156695e-01 -1.98416397e-01 4.14877623e-01
2.08363816e-01 2.81191289e-01 5.11300445e-01 1.14132941e+00
-1.47396341e-01 -8.93144071e-01 -1.38436630e-01 2.67124325e-01
-5.62555671e-01 -4.90320176e-01 -5.97308099e-01 9.09092367e-01
7.00196833e-04 7.70988584e-01 1.34099126e-01 -3.30971509e-01
5.37146866e-01 1.12142853e-01 4.78063405e-01 -4.67890382e-01
-5.92546761e-01 -2.13375360e-01 -1.54093072e-01 -8.93598258e-01
-8.24181616e-01 -3.17537189e-01 -1.11719000e+00 -1.97734922e-01
-1.76786453e-01 2.28074715e-01 1.95497990e-01 7.31516302e-01
3.15813534e-02 1.14564404e-01 1.72511011e-01 -3.51602584e-01
-4.95656341e-01 -5.98851740e-01 -2.22068757e-01 8.82750213e-01
6.45268500e-01 -7.06348777e-01 -7.75403857e-01 1.26351848e-01] | [10.379281044006348, 0.9364314675331116] |
d7f2c09d-59f4-4949-9be3-916be67e0709 | pgmpy-a-python-toolkit-for-bayesian-networks | 2304.08639 | null | https://arxiv.org/abs/2304.08639v1 | https://arxiv.org/pdf/2304.08639v1.pdf | pgmpy: A Python Toolkit for Bayesian Networks | Bayesian Networks (BNs) are used in various fields for modeling, prediction, and decision making. pgmpy is a python package that provides a collection of algorithms and tools to work with BNs and related models. It implements algorithms for structure learning, parameter estimation, approximate and exact inference, causal inference, and simulations. These implementations focus on modularity and easy extensibility to allow users to quickly modify/add to existing algorithms, or to implement new algorithms for different use cases. pgmpy is released under the MIT License; the source code is available at: https://github.com/pgmpy/pgmpy, and the documentation at: https://pgmpy.org. | ['Johannes Textor', 'Ankur Ankan'] | 2023-04-17 | null | null | null | null | ['causal-inference', 'causal-inference'] | ['knowledge-base', 'miscellaneous'] | [-4.63038743e-01 -7.76792690e-02 -7.01522470e-01 -4.84349430e-01
-2.12077111e-01 -2.51277655e-01 4.83252853e-01 1.12208210e-01
1.17282175e-01 7.86137819e-01 1.66057155e-01 -7.67285347e-01
-4.13105309e-01 -1.06874120e+00 -4.42827731e-01 -4.53696162e-01
-2.76621431e-01 5.11537373e-01 3.18005115e-01 5.72400801e-02
1.66102141e-01 4.13359076e-01 -1.10404038e+00 1.59394890e-01
4.58137363e-01 5.07526577e-01 3.33678395e-01 7.61017621e-01
1.37663245e-01 6.47386491e-01 -9.13206860e-02 -3.77689660e-01
-4.63650711e-02 -2.08563894e-01 -8.00280809e-01 -7.38351762e-01
-3.26250345e-01 -4.65597630e-01 -6.91504836e-01 9.83219147e-01
4.47051287e-01 -1.02266431e-01 6.48113966e-01 -1.81600332e+00
-3.56821209e-01 9.38087165e-01 -2.99102187e-01 4.71963018e-01
5.25791466e-01 2.39411995e-01 9.06683028e-01 -5.26957572e-01
3.42246473e-01 1.57631624e+00 6.83373749e-01 3.36924493e-01
-1.45479321e+00 -1.15785861e+00 -1.02639109e-01 5.31844914e-01
-1.56512260e+00 -4.39450234e-01 4.99803394e-01 -5.80163717e-01
8.27075660e-01 3.06217492e-01 6.16434276e-01 1.41393161e+00
3.92160952e-01 8.79158139e-01 9.62950230e-01 -3.00251931e-01
3.41825962e-01 -1.60292000e-01 7.09501028e-01 6.35273099e-01
2.62322903e-01 5.70015550e-01 -4.87815797e-01 -6.03770018e-01
9.41720307e-01 2.56398618e-01 7.15726987e-02 -3.56185138e-02
-8.47793519e-01 8.91432047e-01 1.07081763e-01 1.91075936e-01
-3.53515714e-01 4.29884523e-01 9.03014094e-03 1.74971581e-01
1.27742931e-01 -1.15117565e-01 -5.30915856e-01 -4.94889557e-01
-5.83024740e-01 6.77898407e-01 1.03677237e+00 1.08078754e+00
6.89487100e-01 -1.41030550e-01 -1.78475101e-02 9.64924395e-01
6.52674019e-01 4.04888302e-01 2.48360321e-01 -1.33178842e+00
-8.69823247e-02 3.73756796e-01 9.69142690e-02 -1.06513536e+00
-6.18465185e-01 -2.61600763e-02 -8.22118700e-01 -1.59935132e-01
2.35747159e-01 -1.66511104e-01 -6.37899578e-01 1.55909526e+00
3.48424077e-01 3.41401726e-01 -3.28803658e-01 4.83075827e-01
1.06021619e+00 6.81739509e-01 2.59027481e-01 -1.44387931e-01
1.16414440e+00 -6.10943377e-01 -6.36530697e-01 -2.17269301e-01
3.76242578e-01 -6.47474110e-01 7.82498360e-01 3.45890135e-01
-9.36545491e-01 -2.16496050e-01 -7.98565090e-01 1.19766675e-01
-3.59898090e-01 -2.29362443e-01 8.06172729e-01 6.07340932e-01
-9.15964305e-01 9.03584659e-01 -1.47667062e+00 -5.23068428e-01
3.71135265e-01 2.18182325e-01 6.98439851e-02 -3.39784175e-01
-1.36248326e+00 8.41232121e-01 6.36664271e-01 -9.09733176e-02
-9.16549861e-01 -8.17625284e-01 -6.21428072e-01 3.74267213e-02
2.33102486e-01 -8.60966563e-01 1.58730280e+00 -2.56059855e-01
-1.43672168e+00 2.22263962e-01 -2.09140524e-01 -2.84623116e-01
1.81617498e-01 -1.59357786e-01 -4.04907763e-01 -2.86315650e-01
-3.80087793e-02 3.99971664e-01 2.47312650e-01 -6.36071801e-01
-1.76538512e-01 -3.41434926e-01 -1.48833126e-01 -1.06339775e-01
7.11650178e-02 3.82889003e-01 -3.37341964e-01 -5.12460947e-01
1.08987153e-01 -7.73516476e-01 -2.17295185e-01 -2.59300411e-01
-6.87761128e-01 -3.11945349e-01 7.16486931e-01 -8.54727566e-01
1.38415754e+00 -1.98703790e+00 -1.87786877e-01 3.07188332e-01
2.30544154e-02 -1.66987535e-02 1.99312456e-02 1.11284816e+00
-1.84888482e-01 2.10956335e-01 -2.60668665e-01 7.74758533e-02
2.17364699e-01 2.66966850e-01 1.90954074e-01 2.12528333e-01
-2.21281290e-01 7.85779774e-01 -9.60132718e-01 -3.77542198e-01
5.32451391e-01 1.90323353e-01 -3.09810638e-01 1.49954617e-01
-3.02268207e-01 4.94962960e-01 -5.17185330e-01 8.07151675e-01
6.60316765e-01 -3.28490078e-01 7.11948931e-01 2.00804234e-01
-1.29986376e-01 6.34926856e-01 -1.53992033e+00 1.14162636e+00
-7.99485445e-02 3.96313578e-01 -5.15110977e-03 -1.01322687e+00
7.11442590e-01 3.99033904e-01 2.50126541e-01 -1.46918148e-01
8.90619382e-02 -8.99046436e-02 1.51843512e-02 -5.12422979e-01
6.29377663e-02 2.11955920e-01 1.00113392e-01 6.43389642e-01
1.09632134e-01 -1.26056924e-01 4.74315941e-01 2.68198073e-01
1.29686987e+00 1.63119167e-01 9.05614376e-01 -1.71045676e-01
2.13358086e-02 -1.05954997e-01 8.09164822e-01 1.03861916e+00
2.02938132e-02 -7.58913234e-02 6.69753909e-01 -3.47800553e-01
-8.61188591e-01 -1.42690754e+00 -5.26037872e-01 1.17665517e+00
-2.53044635e-01 -7.87146568e-01 -3.06416988e-01 -1.89825371e-01
2.29788214e-01 1.00183105e+00 -8.88979807e-02 8.81340504e-02
-4.82251272e-02 -9.82025445e-01 6.03423417e-01 6.80266082e-01
4.85757679e-01 -1.07619226e+00 1.33696776e-02 1.30315498e-01
-3.36259842e-01 -6.60076797e-01 9.32713449e-02 -5.58919571e-02
-1.01707113e+00 -1.18728447e+00 6.93382462e-04 -2.32799277e-01
2.84056276e-01 -1.50967926e-01 9.72573757e-01 1.30342886e-01
-3.46540928e-01 4.76167530e-01 -1.60737336e-01 -5.27736068e-01
-6.07758224e-01 -4.59233411e-02 6.60916641e-02 -7.76270509e-01
4.13609117e-01 -1.07207274e+00 -4.80582118e-01 4.54308271e-01
-5.59179962e-01 2.19822928e-01 2.87994504e-01 7.40546465e-01
1.92781255e-01 2.31816530e-01 4.55487788e-01 -8.26929867e-01
6.28141582e-01 -7.42667615e-01 -9.46026266e-01 1.05936192e-01
-8.12809289e-01 -3.32908541e-01 -4.09432910e-02 -1.27208874e-01
-8.67247581e-01 -2.10981444e-01 -5.54515183e-01 -1.79826077e-02
-5.45327067e-01 9.11901951e-01 -1.82656780e-01 3.37368459e-01
4.63439435e-01 9.61762220e-02 -7.18009695e-02 -7.35323608e-01
2.16348499e-01 8.47306848e-01 2.01873124e-01 -5.42606950e-01
3.79359007e-01 -3.34854648e-02 -8.97000059e-02 -6.14525795e-01
-4.99597639e-01 -2.53575802e-01 -6.30005002e-01 -2.05249637e-01
3.69680226e-01 -6.94190025e-01 -9.94054556e-01 4.57973123e-01
-8.27011704e-01 -5.95464528e-01 3.66211832e-01 4.69381034e-01
-2.71522224e-01 3.13808382e-01 -8.07825565e-01 -9.24883664e-01
-1.42933801e-01 -8.72752547e-01 4.12272871e-01 3.36013049e-01
-7.29276478e-01 -1.20604086e+00 8.19409639e-02 3.15210640e-01
2.56623030e-01 3.78193930e-02 9.42620337e-01 -5.98895907e-01
-5.17727017e-01 -1.67845502e-01 -8.52541029e-02 2.13407725e-01
-7.62311369e-02 5.79560816e-01 -6.19565248e-01 -2.73798615e-01
-4.50479031e-01 -1.78834438e-01 6.99815392e-01 7.40783691e-01
1.32215536e+00 -6.39924109e-01 -8.56397867e-01 3.80362898e-01
1.08608973e+00 2.95213312e-01 6.28053248e-01 2.32705504e-01
2.46182159e-01 5.01760423e-01 4.52957779e-01 6.63771331e-01
6.37123287e-01 5.75890183e-01 4.47196960e-01 4.17838842e-01
1.96866676e-01 -4.74461108e-01 1.73357815e-01 6.47592664e-01
-6.71577975e-02 -6.66943267e-02 -1.50794387e+00 2.58521736e-01
-2.23076200e+00 -1.23808634e+00 -4.67525780e-01 2.25745749e+00
9.79054511e-01 -7.15388218e-03 3.03250760e-01 -6.64659142e-02
7.97057390e-01 -9.35347751e-02 -4.64094728e-01 -3.26297909e-01
2.84674257e-01 1.60685748e-01 2.88815916e-01 6.30076289e-01
-9.65601146e-01 8.15157115e-01 7.52489901e+00 7.91385651e-01
-6.57696903e-01 4.29963976e-01 5.52469909e-01 -2.38849632e-02
8.73246696e-03 3.99343163e-01 -9.16628480e-01 6.16364837e-01
1.34293818e+00 -5.16386449e-01 6.03655338e-01 7.52074480e-01
7.99449265e-01 -4.00836855e-01 -8.84730041e-01 5.30786753e-01
-7.44790912e-01 -1.32881665e+00 -4.92710620e-01 -1.78981367e-02
4.01990980e-01 3.39513510e-01 -3.07092935e-01 2.43893445e-01
9.93963718e-01 -9.81770873e-01 2.90709406e-01 5.36630094e-01
2.56470829e-01 -5.36581278e-01 7.05361247e-01 4.88271862e-01
-8.55805397e-01 -3.60572487e-01 -4.80883926e-01 -6.03677154e-01
2.13193446e-01 9.44589972e-01 -1.01880872e+00 6.59794629e-01
1.04608488e+00 9.76373494e-01 -4.35751051e-01 1.27107000e+00
-6.81264699e-01 1.17356122e+00 -4.03818309e-01 -1.54363036e-01
-4.08565730e-01 -1.37283295e-01 5.57061911e-01 1.30797076e+00
2.14744166e-01 8.89133811e-02 2.71275997e-01 1.04140282e+00
2.77240098e-01 -2.02769294e-01 -5.13389647e-01 -1.30360916e-01
1.23622978e+00 1.12300420e+00 -5.74033141e-01 -2.41425082e-01
-2.37390563e-01 1.67511970e-01 1.21766493e-01 5.12180507e-01
-8.67829502e-01 -2.40957513e-01 7.16528177e-01 1.15577623e-01
3.42302863e-03 -3.53640467e-01 -3.26324821e-01 -7.58440375e-01
-5.24456620e-01 -7.90928721e-01 8.36326301e-01 -1.05307353e+00
-1.28899300e+00 -3.10349651e-02 7.41461754e-01 -6.27260625e-01
-5.46188056e-01 -5.45208156e-01 -8.70795608e-01 7.38250971e-01
-8.86336863e-01 -8.64265025e-01 -2.58560061e-01 4.11486536e-01
1.60544831e-02 2.70313919e-02 8.37611556e-01 2.11042315e-01
-9.23753262e-01 2.16105059e-01 2.43906379e-01 1.68031156e-01
4.52048212e-01 -1.11050522e+00 2.49299973e-01 6.33307159e-01
-2.82507211e-01 9.19670403e-01 9.22922313e-01 -8.90646338e-01
-1.15123427e+00 -1.00195611e+00 7.78076708e-01 -2.85232753e-01
9.37519193e-01 -3.86383563e-01 -6.78083539e-01 1.28292799e+00
2.13709980e-01 -4.30723727e-01 8.75582099e-01 6.71968937e-01
-2.20336989e-01 -1.28551930e-01 -9.64073598e-01 7.47594476e-01
9.51483250e-01 -2.03120857e-01 -4.27987099e-01 4.56317812e-01
2.86101520e-01 -3.37281376e-01 -1.09569705e+00 4.10155207e-01
8.14646661e-01 -1.02619088e+00 8.38179231e-01 -2.19872326e-01
2.39090949e-01 -2.96080202e-01 -1.34502262e-01 -1.07821822e+00
-6.60887957e-01 -5.00589967e-01 -4.40119505e-01 1.35519040e+00
4.32252795e-01 -1.01843250e+00 4.46296215e-01 6.18911862e-01
6.46502674e-02 -6.17964029e-01 -9.54135478e-01 -7.85234272e-01
4.50096317e-02 -1.07570612e+00 7.75109828e-01 9.40057099e-01
2.96206921e-01 1.73183784e-01 -2.72970140e-01 4.02078211e-01
6.91951573e-01 -1.11818574e-02 7.32449710e-01 -1.29366219e+00
-5.68307459e-01 -5.72659075e-01 -3.22913021e-01 -8.43955815e-01
-9.44692269e-02 -9.30866957e-01 -2.94345438e-01 -1.68872690e+00
4.78750795e-01 -4.87010866e-01 -2.03711078e-01 1.08547270e+00
8.32681134e-02 7.49059469e-02 -1.08862132e-01 2.57261723e-01
-2.04527617e-01 3.36451858e-01 7.48478055e-01 -4.00477350e-02
-1.68007478e-01 5.73036075e-01 -6.69036031e-01 8.06238770e-01
1.47854638e+00 -8.15430105e-01 -1.59309670e-01 5.73589746e-03
1.23957612e-01 8.88678804e-02 8.89956474e-01 -7.32514858e-01
1.75415754e-01 -5.43041408e-01 5.52856326e-01 -7.28766859e-01
2.95628846e-01 -2.99302250e-01 7.10791409e-01 7.81729639e-01
-1.26488544e-02 1.69945449e-01 2.21942365e-01 3.49780798e-01
1.48589730e-01 -5.89595199e-01 5.78106821e-01 -2.74530709e-01
-5.17164409e-01 2.22595990e-01 -6.61664188e-01 -3.24259311e-01
8.21784735e-01 1.83129340e-01 -5.04856825e-01 -6.83379769e-01
-9.95855808e-01 4.72658515e-01 2.05809593e-01 2.85422981e-01
4.66256529e-01 -1.23343432e+00 -5.84123671e-01 -7.29988068e-02
-1.34511173e-01 -3.94755840e-01 2.01128110e-01 1.11777020e+00
-4.25090849e-01 4.63735729e-01 -4.77745896e-03 -4.29984003e-01
-1.28369510e+00 2.61719406e-01 3.37960720e-01 -2.13253707e-01
-4.89883691e-01 4.73601520e-01 -2.16357619e-01 -7.99451232e-01
2.12586433e-01 -9.12559479e-02 -2.48616971e-02 -5.60731411e-01
6.70993507e-01 8.49456728e-01 -3.18833232e-01 2.22084790e-01
-3.37672055e-01 -4.76736240e-02 1.56666726e-01 -2.04022191e-02
1.54957235e+00 -2.22653911e-01 -5.59597075e-01 7.23232150e-01
5.35221815e-01 -4.44994271e-01 -8.78407359e-01 -1.49136811e-01
2.81293005e-01 -3.35367978e-01 8.78553241e-02 -1.16171181e+00
-6.66265070e-01 5.63107252e-01 4.18221086e-01 2.50929922e-01
9.49837148e-01 2.44139850e-01 1.84010148e-01 2.30371803e-01
3.71410072e-01 -6.99108839e-01 -4.34804052e-01 6.98550999e-01
9.20699239e-01 -9.52386379e-01 4.00903463e-01 -4.17341650e-01
-1.75894305e-01 1.04310155e+00 4.81461912e-01 2.20909968e-01
1.30975187e+00 5.08481622e-01 -1.82441235e-01 -1.08556055e-01
-1.04675031e+00 1.43138781e-01 -1.07183337e-01 8.09695959e-01
7.46179819e-01 3.64222825e-01 -5.30956984e-01 9.52104092e-01
-4.85710233e-01 3.44858795e-01 4.83744651e-01 9.14116442e-01
-1.61245137e-01 -1.47652328e+00 -5.51711679e-01 6.54730439e-01
-3.36658865e-01 -3.11173141e-01 -7.20757246e-02 6.96164072e-01
-1.13153599e-01 1.12609136e+00 -2.85241790e-02 -3.70814919e-01
-1.04729682e-01 3.07289332e-01 3.30602139e-01 -5.00595927e-01
-1.75997764e-01 6.73009306e-02 4.17114407e-01 -6.58117235e-01
-1.41674116e-01 -1.02963507e+00 -9.96511102e-01 -9.06163812e-01
-1.38028130e-01 1.69553801e-01 5.71702838e-01 6.64962292e-01
5.00876069e-01 3.83273602e-01 2.59942651e-01 -6.88302815e-01
-3.28437984e-01 -1.44913375e+00 -7.34562397e-01 -3.70869547e-01
-2.33568653e-01 -8.24813426e-01 -7.70989433e-02 -1.34797692e-01] | [7.511124134063721, 4.419738292694092] |
817d5ffc-903c-47ff-9968-3b14b2d946e0 | object-detection-by-labeling-superpixels | null | null | http://openaccess.thecvf.com/content_cvpr_2015/html/Yan_Object_Detection_by_2015_CVPR_paper.html | http://openaccess.thecvf.com/content_cvpr_2015/papers/Yan_Object_Detection_by_2015_CVPR_paper.pdf | Object Detection by Labeling Superpixels | Object detection is always conducted by object proposal generation and classification sequentially. This paper handles object detection in a superpixel oriented manner instead of the proposal oriented. Specially, this paper takes object detection as a multi-label superpixel labeling problem by minimizing an energy function. It uses the data cost term to capture the appearance, smooth cost term to encode the spatial context and label cost term to favor compact detection. The data cost is learned through a convolutional neural network and the parameters in the labeling model are learned through a structural SVM. Compared with proposal generation and classification based methods, the proposed superpixel labeling method can naturally detect objects missed by proposal generation step and capture the global image context to infer the overlapping objects. The proposed method shows its advantage in Pascal VOC and ImageNet. Notably, it performs better than the ImageNet ILSVRC2014 winner GoogLeNet (45.0% V.S. 43.9% in mAP) with much shallower and fewer CNNs. | ['Zhen Lei', 'Stan Z. Li', 'Yinan Yu', 'Xiangyu Zhu', 'Junjie Yan'] | 2015-06-01 | null | null | null | cvpr-2015-6 | ['object-proposal-generation'] | ['computer-vision'] | [ 4.31913674e-01 2.33024344e-01 -2.83089250e-01 -4.23173189e-01
-5.48079252e-01 -3.58715594e-01 4.43263590e-01 9.88791883e-02
-6.79178238e-01 6.35704875e-01 -5.52613318e-01 -5.84889799e-02
4.54312116e-01 -6.72920704e-01 -8.84006619e-01 -8.27731371e-01
3.22561979e-01 3.76121163e-01 8.67007375e-01 3.01822931e-01
3.60543340e-01 2.97848910e-01 -1.61135399e+00 3.76564860e-01
8.21871281e-01 1.21676517e+00 6.88389778e-01 5.20449221e-01
-1.80423498e-01 8.36778224e-01 -6.29616320e-01 -6.63939398e-03
4.21594054e-01 -1.00152344e-01 -7.55339205e-01 4.14887697e-01
1.18766344e+00 -3.30554813e-01 2.78758258e-01 1.35882664e+00
3.84391308e-01 -4.67395298e-02 6.51369333e-01 -1.09044051e+00
-5.88544309e-01 3.61640304e-01 -9.37004149e-01 2.18191534e-01
-1.53407037e-01 1.41779453e-01 9.15020108e-01 -1.21493089e+00
4.98044252e-01 1.26079047e+00 5.62288165e-01 3.72621775e-01
-1.35433257e+00 -6.38529897e-01 5.07033825e-01 1.24333315e-01
-1.48642349e+00 8.22658539e-02 7.58090556e-01 -5.75043857e-01
6.52377069e-01 2.37732410e-01 6.29698932e-01 7.37798393e-01
-1.24096990e-01 1.16163576e+00 1.10122275e+00 -3.34577978e-01
2.41369203e-01 5.68625212e-01 4.95044708e-01 9.70416903e-01
3.97477120e-01 9.96026322e-02 -1.19843110e-01 1.11598015e-01
8.03852975e-01 1.16056733e-01 -3.48205604e-02 -3.44986975e-01
-1.04444814e+00 6.15860224e-01 7.52483726e-01 -2.07453850e-03
-2.27129385e-01 4.34280634e-01 1.90907910e-01 -2.70880491e-01
4.74283636e-01 1.83047548e-01 -4.90064859e-01 7.63954520e-01
-1.16386163e+00 1.76785320e-01 4.31797594e-01 1.34480858e+00
9.54517722e-01 2.41397619e-01 -5.69682062e-01 8.76052141e-01
5.63545465e-01 4.25624102e-01 1.20532140e-01 -9.24907029e-01
3.58972877e-01 8.09251904e-01 3.08512419e-01 -8.51665080e-01
-3.96651864e-01 -8.08988392e-01 -5.05389631e-01 5.97612381e-01
4.12855357e-01 3.21819969e-02 -1.24246156e+00 1.23269284e+00
5.94193220e-01 2.56421149e-01 -3.25282156e-01 1.05005562e+00
1.06313097e+00 6.22230232e-01 3.45846683e-01 -3.01648281e-03
1.56065416e+00 -1.67258799e+00 -4.82953191e-01 -5.96764624e-01
5.23122489e-01 -7.48200834e-01 7.89057195e-01 3.68269682e-01
-1.02113533e+00 -9.86184776e-01 -1.30259502e+00 -2.20772102e-01
-3.66938710e-01 1.08566523e+00 5.50785244e-01 4.92149889e-01
-9.58468437e-01 2.77304053e-01 -6.27047956e-01 -2.25863665e-01
7.66390085e-01 3.51077229e-01 8.66983011e-02 -5.24765179e-02
-7.04608858e-01 8.54340851e-01 1.11295307e+00 1.15601480e-01
-1.06210995e+00 -5.64984500e-01 -6.81295097e-01 -1.82652157e-02
3.80861640e-01 -3.38161111e-01 9.46416438e-01 -1.12032723e+00
-1.24455869e+00 1.12529516e+00 1.69495210e-01 -4.10482228e-01
7.35478878e-01 -1.32889226e-01 -3.80263291e-02 9.15021971e-02
2.47309983e-01 1.35963666e+00 8.80701542e-01 -1.32677639e+00
-1.39405298e+00 -8.32792744e-02 -2.21746713e-02 5.87136298e-02
6.56602010e-02 6.67455792e-02 -6.21311665e-01 -6.04108512e-01
3.56598139e-01 -8.34027052e-01 -2.37382710e-01 4.45228457e-01
-5.30314445e-01 -5.90236306e-01 1.20735741e+00 -4.75469112e-01
8.43308270e-01 -1.99925745e+00 -2.60123104e-01 -2.07896814e-01
1.71949938e-01 4.16221380e-01 -1.33406430e-01 -3.47254306e-01
1.81649700e-01 -6.97197989e-02 -2.04026461e-01 -6.33973897e-01
-2.30565548e-01 -5.06419986e-02 2.53185928e-02 5.84961295e-01
4.44985360e-01 9.02083874e-01 -8.89206707e-01 -9.69517589e-01
4.53441530e-01 4.12929893e-01 -4.42850530e-01 1.10828936e-01
-4.57394451e-01 -8.77224933e-03 -3.54247898e-01 1.01901543e+00
1.03812945e+00 -3.87507617e-01 -1.40525535e-01 -4.81299192e-01
-4.05776352e-01 -1.30907044e-01 -1.43753541e+00 1.50407732e+00
-1.01876184e-01 7.36739516e-01 2.69688070e-01 -9.20624077e-01
1.05245638e+00 -6.34471253e-02 9.73925963e-02 -4.24517006e-01
1.72222272e-01 2.96238691e-01 -3.46633732e-01 -4.90253568e-01
6.18322790e-01 3.27650368e-01 3.41987640e-01 7.83260688e-02
1.56908497e-01 3.64275314e-02 2.77333111e-01 -1.44631527e-02
5.45917153e-01 4.81366813e-01 1.79870844e-01 -5.76365471e-01
4.78774071e-01 2.92937666e-01 6.90321386e-01 8.86547029e-01
-5.01608431e-01 6.81973100e-01 1.00298058e-02 -5.58708370e-01
-1.01372540e+00 -8.13403964e-01 -4.08484787e-01 1.22032726e+00
7.09455371e-01 3.27998847e-02 -7.28094220e-01 -9.89278138e-01
1.02168517e-02 6.81254208e-01 -5.86659849e-01 6.12294450e-02
-7.82850921e-01 -9.36170638e-01 1.80499986e-01 6.77456617e-01
9.47131276e-01 -9.43088114e-01 -7.24909782e-01 1.77721769e-01
-1.28004387e-01 -1.07692027e+00 -6.71572804e-01 4.50119942e-01
-7.87545085e-01 -1.11498380e+00 -6.07879341e-01 -1.20909452e+00
9.06046867e-01 4.07630622e-01 9.34333861e-01 -1.09580338e-01
-7.77943790e-01 -1.66614264e-01 -9.21500698e-02 -4.83972639e-01
-1.87887195e-02 -3.12239025e-02 -3.05860966e-01 1.87026009e-01
3.07105213e-01 5.16110621e-02 -9.72464025e-01 4.32499111e-01
-6.04878128e-01 4.13872391e-01 8.44943702e-01 8.91561806e-01
9.45441127e-01 -2.60256022e-01 3.43872368e-01 -7.87530601e-01
-1.28653005e-01 -1.97889462e-01 -9.67837930e-01 3.89783353e-01
-7.59286582e-01 -1.43681780e-01 2.00072587e-01 -6.22772515e-01
-1.20259631e+00 6.93697512e-01 3.08411777e-01 -2.31157586e-01
-3.60169500e-01 -2.56354004e-01 -1.41137559e-02 -4.60579455e-01
5.87160766e-01 3.27720463e-01 -4.14327532e-01 -6.25730336e-01
3.55414987e-01 5.82833827e-01 3.93514425e-01 -2.57160217e-01
4.18126494e-01 6.70669615e-01 -1.47594258e-01 -6.32442415e-01
-1.08418179e+00 -8.23257267e-01 -7.86253095e-01 -3.84545326e-01
1.18187082e+00 -9.87614155e-01 -4.77866441e-01 3.69374156e-01
-1.47324860e+00 -2.63775438e-01 -1.66722730e-01 3.26009125e-01
-2.57056087e-01 8.62783268e-02 -6.17804170e-01 -9.49394703e-01
-3.61170053e-01 -1.10053670e+00 1.43047726e+00 3.47349763e-01
4.00978059e-01 -7.00133920e-01 -4.97952849e-01 3.42357755e-01
3.08309883e-01 2.49251768e-01 3.21289212e-01 -4.35290068e-01
-1.19203115e+00 -2.97002614e-01 -9.55630779e-01 5.74331939e-01
-1.04712576e-01 7.55305067e-02 -1.20397377e+00 -3.23106289e-01
-4.83083911e-02 -3.41954768e-01 1.36780214e+00 5.32088757e-01
1.30063426e+00 -4.10686076e-01 -6.73199415e-01 6.88243985e-01
1.84736156e+00 1.91394612e-01 2.46487454e-01 2.35996559e-01
9.63038743e-01 5.73381424e-01 7.67519176e-01 3.32695097e-02
1.78173333e-01 7.84481168e-01 7.02116549e-01 -4.11509156e-01
-6.49046183e-01 -1.77250020e-02 9.66579765e-02 8.18488225e-02
8.23975354e-02 -8.53199065e-02 -8.49372506e-01 5.57414174e-01
-1.94948614e+00 -7.23848760e-01 -5.48507333e-01 1.79151380e+00
7.60946274e-01 3.20965916e-01 3.01547814e-02 -2.53480166e-01
1.07621038e+00 5.65135926e-02 -5.26220858e-01 1.42460898e-01
-2.97102869e-01 -3.05964917e-01 9.17789817e-01 4.94200438e-01
-1.66249907e+00 1.16157103e+00 6.22215939e+00 1.14578605e+00
-1.10911191e+00 3.73792142e-01 8.46656203e-01 6.05079494e-02
4.20362711e-01 -2.49258876e-01 -1.42994988e+00 5.11339664e-01
1.66056752e-01 5.29694140e-01 -2.03603894e-01 1.33942759e+00
7.70525187e-02 -4.08951879e-01 -1.18817580e+00 1.02661645e+00
1.65883273e-01 -1.35384023e+00 5.44546209e-02 -1.63724646e-01
9.99263406e-01 1.56187981e-01 7.45609179e-02 1.00189202e-01
1.59019545e-01 -8.43493998e-01 1.24437857e+00 2.74115801e-01
6.32317781e-01 -2.90777504e-01 7.31132448e-01 3.72498184e-01
-1.36478496e+00 -2.15908289e-01 -5.78083694e-01 8.89577419e-02
3.48148867e-02 5.83034039e-01 -9.81423199e-01 1.26948282e-01
5.85031271e-01 6.00691378e-01 -8.76303613e-01 1.44475091e+00
-2.81619042e-01 6.30800366e-01 -3.44370872e-01 -1.42363578e-01
5.01780748e-01 -2.39077255e-01 4.22431469e-01 1.67167592e+00
-7.77619034e-02 -1.21259250e-01 8.23451519e-01 1.23125398e+00
1.16263993e-01 6.96398094e-02 -1.01141017e-02 5.41129470e-01
3.16261321e-01 1.58072364e+00 -1.08349752e+00 -6.55887306e-01
-2.65348464e-01 1.19249368e+00 3.53881687e-01 4.07174230e-01
-9.77355182e-01 -3.38879824e-01 7.03192428e-02 6.83414415e-02
4.15071160e-01 2.60977410e-02 -5.06797612e-01 -8.47258151e-01
5.92440702e-02 -1.67654902e-01 2.84579873e-01 -7.56413221e-01
-9.12405908e-01 6.30648196e-01 -9.71372873e-02 -1.13761723e+00
3.40248495e-01 -9.76138711e-01 -5.22140324e-01 6.47962809e-01
-1.71382582e+00 -1.49732220e+00 -5.66601634e-01 8.53639767e-02
9.35578644e-01 2.33715296e-01 3.61853540e-01 4.83165652e-01
-7.77899206e-01 3.80890399e-01 -3.18872966e-02 1.49451271e-01
4.74960446e-01 -1.44986641e+00 -4.58329264e-03 8.20714951e-01
-2.34622285e-02 1.19691253e-01 4.99147922e-01 -5.54773808e-01
-5.95515847e-01 -1.71511590e+00 6.65436745e-01 -4.22268718e-01
3.40233415e-01 -4.62191492e-01 -6.97022736e-01 4.95907933e-01
6.84773251e-02 5.19525290e-01 -2.52899695e-02 -4.94323730e-01
-2.43877590e-01 -2.31759936e-01 -1.20589995e+00 1.80961102e-01
8.98440421e-01 -1.20391406e-01 -2.77294934e-01 7.64087856e-01
8.15463603e-01 -3.92776728e-01 -3.50036711e-01 4.91412580e-01
3.36430371e-01 -6.32702470e-01 1.15071929e+00 -2.24309042e-01
1.57925978e-01 -8.33232164e-01 3.78169189e-03 -6.35724843e-01
-5.00391662e-01 9.41304490e-02 -1.32151291e-01 1.06744397e+00
5.84170818e-01 -2.76058555e-01 1.03929937e+00 2.32969955e-01
-2.75536060e-01 -8.26117098e-01 -7.65673101e-01 -8.48764837e-01
-3.92223597e-01 -1.88896447e-01 1.33952692e-01 6.95674837e-01
-7.70456493e-01 5.19438721e-02 -3.22707117e-01 2.97147214e-01
1.12701428e+00 2.15057984e-01 3.93903643e-01 -1.20415485e+00
-1.20229095e-01 -5.10704219e-01 -3.88195723e-01 -1.34736586e+00
-2.10043207e-01 -8.31544042e-01 4.28131014e-01 -1.59119737e+00
4.37864751e-01 -6.29670858e-01 -4.82615620e-01 5.50571442e-01
-4.53666002e-01 7.02069938e-01 1.03078999e-01 2.56967306e-01
-1.04104996e+00 2.71156400e-01 1.30558479e+00 -6.06714666e-01
-1.80067927e-01 6.66894615e-02 -2.73169011e-01 9.05188680e-01
8.02373409e-01 -5.08321285e-01 -1.51947558e-01 -4.90002960e-01
-1.98759273e-01 -5.05720854e-01 8.46444964e-01 -1.12096643e+00
4.32626903e-01 -2.56242841e-01 4.71212566e-01 -1.03660619e+00
2.82443047e-01 -7.35827386e-01 -2.42862284e-01 6.64022148e-01
-3.39571685e-01 -4.98298824e-01 4.97562848e-02 7.76255548e-01
-1.19895928e-01 -2.88140833e-01 1.22995353e+00 -2.12851435e-01
-1.06611705e+00 3.85079980e-01 -8.70222822e-02 -1.91131398e-01
1.44725227e+00 -6.09061003e-01 -2.65173197e-01 5.55496812e-01
-6.69880092e-01 4.09283787e-01 7.33658150e-02 3.33518475e-01
6.65713251e-01 -1.25339711e+00 -7.51636386e-01 1.15932032e-01
3.04991513e-01 4.20602471e-01 1.34474151e-02 8.25537384e-01
-6.75137162e-01 4.03829157e-01 1.16726067e-02 -9.56954181e-01
-1.31204224e+00 4.66105759e-01 6.21710300e-01 2.11343899e-01
-5.96004784e-01 1.46808696e+00 5.28125882e-01 -2.24216983e-01
6.26501203e-01 -6.20972395e-01 -1.41258091e-01 3.89699414e-02
5.83022237e-01 4.87003326e-01 -1.23081148e-01 -4.85344708e-01
-4.27263677e-01 5.03833473e-01 -2.24161550e-01 3.43666494e-01
9.79770720e-01 -1.88223466e-01 -1.65944114e-01 2.79326081e-01
1.13808322e+00 -4.68933344e-01 -1.63477051e+00 -2.39637986e-01
1.80798605e-01 -3.83241832e-01 3.26846957e-01 -9.92874920e-01
-1.09245420e+00 7.92222619e-01 9.82720137e-01 5.90824010e-03
7.53618240e-01 1.52421027e-01 3.02404553e-01 1.96213931e-01
3.32892478e-01 -1.48797238e+00 3.87540251e-01 2.34471291e-01
7.51381278e-01 -1.74314821e+00 8.10455903e-02 -7.78105021e-01
-3.18605840e-01 1.03809178e+00 1.18862021e+00 -3.89721870e-01
5.82267463e-01 1.31558210e-01 -1.04416646e-01 -1.06968716e-01
-3.24170560e-01 -3.28653038e-01 5.37259161e-01 4.83556509e-01
1.98515639e-01 1.29282862e-01 -3.65895927e-01 3.46592695e-01
4.21123654e-01 -2.17964515e-01 2.69085407e-01 5.47954857e-01
-1.05327725e+00 -6.27453744e-01 -3.93490493e-01 3.98328692e-01
-3.76243353e-01 -1.95251450e-01 -2.04049930e-01 6.00810349e-01
9.59799170e-01 7.78885067e-01 2.05502197e-01 3.22519615e-02
6.19163364e-02 -5.84115200e-02 2.91907340e-01 -9.29842234e-01
-4.47550058e-01 9.74679515e-02 -9.21666622e-02 -6.69595122e-01
-5.69718361e-01 -5.11138737e-01 -1.11541867e+00 3.48760843e-01
-9.77519274e-01 -1.13528997e-01 8.56174588e-01 6.96016848e-01
8.03202167e-02 7.42924452e-01 2.81564921e-01 -1.04940259e+00
-3.70335639e-01 -9.87363100e-01 -6.27933383e-01 1.62403092e-01
3.91780078e-01 -6.09869778e-01 -2.68389821e-01 2.24951342e-01] | [9.271199226379395, 0.7917369604110718] |
951008e9-4ffb-406a-9d32-d4b7d0005265 | multi-view-representation-learning-in-multi | 2201.05829 | null | https://arxiv.org/abs/2201.05829v1 | https://arxiv.org/pdf/2201.05829v1.pdf | Multi-View representation learning in Multi-Task Scene | Over recent decades have witnessed considerable progress in whether multi-task learning or multi-view learning, but the situation that consider both learning scenes simultaneously has received not too much attention. How to utilize multiple views latent representation of each single task to improve each learning task performance is a challenge problem. Based on this, we proposed a novel semi-supervised algorithm, termed as Multi-Task Multi-View learning based on Common and Special Features (MTMVCSF). In general, multi-views are the different aspects of an object and every view includes the underlying common or special information of this object. As a consequence, we will mine multiple views jointly latent factor of each learning task which consists of each view special feature and the common feature of all views. By this way, the original multi-task multi-view data has degenerated into multi-task data, and exploring the correlations among multiple tasks enables to make an improvement on the performance of learning algorithm. Another obvious advantage of this approach is that we get latent representation of the set of unlabeled instances by the constraint of regression task with labeled instances. The performance of classification and semi-supervised clustering task in these latent representations perform obviously better than it in raw data. Furthermore, an anti-noise multi-task multi-view algorithm called AN-MTMVCSF is proposed, which has a strong adaptability to noise labels. The effectiveness of these algorithms is proved by a series of well-designed experiments on both real world and synthetic data. | ['Xin Zuo', 'Si-ming Lian', 'Jian-wei Liu', 'Run-kun Lu'] | 2022-01-15 | null | null | null | null | ['multi-view-learning'] | ['computer-vision'] | [ 1.35095611e-01 -3.42415035e-01 -1.20485350e-01 -4.05974120e-01
-7.27666140e-01 -3.74333143e-01 6.22370124e-01 -3.09079885e-01
-1.05816811e-01 4.84679252e-01 2.22067863e-01 4.23190445e-01
-3.38614941e-01 -4.73831594e-01 -3.88366461e-01 -1.09108484e+00
4.17101473e-01 4.51422900e-01 7.89049864e-02 1.50830254e-01
9.85731855e-02 -5.99174760e-02 -1.71301997e+00 6.85412288e-01
6.30196512e-01 6.41272843e-01 5.30306399e-01 2.29112774e-01
-2.03016773e-01 4.06286448e-01 -4.58137542e-01 -1.48327693e-01
5.66914231e-02 -3.01622301e-01 -5.85945964e-01 7.54589438e-01
2.70961106e-01 2.64371216e-01 3.12334865e-01 1.00924110e+00
5.70288360e-01 1.02789268e-01 8.64656925e-01 -1.38845074e+00
-4.16551590e-01 2.25727946e-01 -1.04196179e+00 8.25397074e-02
1.40893340e-01 -3.30390602e-01 8.25621128e-01 -1.06284583e+00
4.32349831e-01 1.35372508e+00 2.71126062e-01 3.01600039e-01
-1.19524896e+00 -6.24531090e-01 4.12058473e-01 2.84689248e-01
-1.15978479e+00 -2.53634602e-01 9.66158807e-01 -6.63429558e-01
4.41198587e-01 1.72292776e-02 3.87734056e-01 1.27437437e+00
3.95384550e-01 9.78108644e-01 1.63307261e+00 -3.76640707e-01
-7.38452598e-02 6.03508174e-01 1.57302946e-01 6.40083373e-01
1.84692666e-01 -3.47184747e-01 -4.91472751e-01 -1.15565546e-01
4.16181535e-01 4.97523159e-01 -2.37226129e-01 -7.66552627e-01
-1.35842884e+00 6.91653252e-01 -1.96358055e-01 4.96185780e-01
-1.89895764e-01 -4.74006414e-01 5.66943407e-01 2.66404241e-01
6.42871737e-01 -1.16927125e-01 -5.27212858e-01 2.73407787e-01
-6.51190817e-01 -1.15799300e-01 4.20486748e-01 8.79355252e-01
1.05582190e+00 9.00790393e-02 2.83474743e-01 1.08092701e+00
4.45225626e-01 6.76985383e-01 7.48743236e-01 -6.51840448e-01
5.70262253e-01 8.77616525e-01 -3.08415383e-01 -1.00452256e+00
-4.67898399e-01 -5.44796109e-01 -1.22924769e+00 2.17987299e-01
1.88692898e-01 -1.86224505e-01 -6.44530952e-01 1.61004984e+00
4.23095435e-01 6.45644665e-02 4.38098937e-01 6.19302273e-01
7.63841569e-01 7.78536379e-01 -8.82507861e-02 -8.27380419e-01
1.40327013e+00 -1.07454944e+00 -9.03076351e-01 -2.59534240e-01
4.75417614e-01 -8.74458075e-01 7.93095112e-01 7.65774906e-01
-6.11929834e-01 -1.07951653e+00 -8.88838112e-01 3.86281699e-01
-3.15955132e-01 3.40165555e-01 5.52290916e-01 5.53896308e-01
-4.72498327e-01 5.45934699e-02 -4.41792607e-01 -2.87838250e-01
9.91571993e-02 3.03816259e-01 -7.71853030e-01 -2.82288522e-01
-7.72355437e-01 6.68825626e-01 6.38442814e-01 -2.32795272e-02
-8.53391826e-01 -1.84062630e-01 -7.02082217e-01 -2.65238851e-01
8.04534614e-01 -6.24936521e-01 7.08486617e-01 -1.16715181e+00
-9.26879883e-01 7.74756551e-01 -4.82547849e-01 2.53735602e-01
2.38376141e-01 -2.83898562e-01 -7.03512669e-01 -9.98206511e-02
4.36203271e-01 4.87315953e-02 1.36705017e+00 -1.72300780e+00
-6.67323053e-01 -8.75006139e-01 -4.11619663e-01 5.49792886e-01
-5.01171291e-01 -8.62729773e-02 -4.93296713e-01 -5.33544838e-01
4.40059870e-01 -1.00160694e+00 -1.60526469e-01 -5.79417408e-01
-1.55203119e-01 -3.42168123e-01 1.31987762e+00 -2.87886411e-01
9.02962327e-01 -2.28836918e+00 4.65669483e-01 4.04797308e-02
1.93484977e-01 1.05922043e-01 5.18293716e-02 4.60458726e-01
-3.85927737e-01 -1.64176188e-02 -2.89089549e-02 -3.47232372e-01
-5.00084281e-01 3.59814614e-01 -1.56897619e-01 4.68497962e-01
-2.44623974e-01 4.92459178e-01 -6.85982049e-01 -7.75896490e-01
2.43436694e-01 1.12468876e-01 -2.93555148e-02 2.80548573e-01
1.29917890e-01 6.30082846e-01 -6.68221831e-01 5.17465770e-01
6.77633464e-01 -4.37142938e-01 2.82403141e-01 -4.07866418e-01
7.11556375e-02 -6.30543709e-01 -1.52431881e+00 1.81775761e+00
-5.09168923e-01 1.05521604e-01 6.14224188e-02 -1.28750885e+00
1.02750778e+00 7.25435615e-01 8.16754460e-01 -3.57368648e-01
-1.82136044e-01 -3.55814621e-02 -2.43637189e-01 -7.77914345e-01
-4.66729887e-02 -4.49357361e-01 -3.63414325e-02 8.18091273e-01
2.55497009e-01 1.96867108e-01 4.16636020e-02 1.12345658e-01
3.85568947e-01 2.63806492e-01 6.39357984e-01 -1.12981305e-01
8.21550667e-01 -2.19797492e-01 7.81663775e-01 4.17981654e-01
3.06836199e-02 5.08908033e-01 2.63259470e-01 -4.08682972e-01
-9.09098923e-01 -8.95641804e-01 -4.35210057e-02 1.18136299e+00
9.99859497e-02 -4.04712349e-01 -3.33497286e-01 -1.08175087e+00
-1.48509473e-01 2.23517194e-01 -6.13425732e-01 5.75801954e-02
-3.44458282e-01 -1.03868067e+00 -7.45985005e-03 3.61900985e-01
5.29271305e-01 -9.40671146e-01 -2.49765977e-01 1.86989143e-01
-3.62970173e-01 -1.22953176e+00 -1.82584256e-01 4.24886286e-01
-1.08186531e+00 -1.25560760e+00 -7.33485043e-01 -8.12630117e-01
6.38584256e-01 8.74787331e-01 8.49851489e-01 -3.84129435e-01
1.59955516e-01 5.28113186e-01 -4.39347029e-01 -4.36708122e-01
-3.20146561e-01 -7.92814195e-02 3.81386191e-01 6.30115688e-01
3.05181473e-01 -7.75412202e-01 5.42542739e-06 6.19637668e-01
-9.27552700e-01 2.99461305e-01 8.50749433e-01 8.74501348e-01
7.94804990e-01 4.10023093e-01 7.77352929e-01 -1.17833257e+00
2.37772822e-01 -5.21084487e-01 -3.88448089e-01 5.95430374e-01
-7.03545630e-01 1.19351447e-01 7.62474000e-01 -3.43379319e-01
-1.44521391e+00 3.15301120e-01 4.13351536e-01 -7.59282827e-01
-4.17011052e-01 4.85536635e-01 -5.34430385e-01 3.34152073e-01
3.04663479e-01 5.49146712e-01 7.07484186e-02 -5.50354958e-01
3.04765254e-01 7.46385574e-01 1.16998330e-01 -4.82407808e-01
7.75056779e-01 6.84083521e-01 1.99135855e-01 -8.19243729e-01
-1.24709249e+00 -8.46509278e-01 -1.06411779e+00 -4.51962560e-01
1.06365216e+00 -1.22461879e+00 -5.27679622e-01 4.88025099e-01
-8.84196281e-01 3.89006317e-01 1.63513497e-01 6.37503684e-01
-5.35314918e-01 6.96134329e-01 3.18635404e-02 -8.24093580e-01
7.43555054e-02 -1.25178754e+00 1.09857452e+00 1.32649194e-03
2.85827160e-01 -1.22588480e+00 7.14269727e-02 8.54763210e-01
-2.13854656e-01 6.80760220e-02 9.11009729e-01 -7.98224986e-01
-3.32644492e-01 1.41344011e-01 -9.71965194e-02 5.62863767e-01
5.46879649e-01 -2.38248244e-01 -1.20474851e+00 -4.73396391e-01
8.11946630e-01 -4.46992546e-01 9.99451935e-01 3.72545481e-01
1.08519828e+00 1.31326735e-01 -4.29839134e-01 4.60921973e-01
1.62522829e+00 2.61194468e-01 9.19524282e-02 2.10110784e-01
1.04371393e+00 8.06185961e-01 8.31039786e-01 4.48084295e-01
2.85103798e-01 6.73917413e-01 3.45892131e-01 -1.32045656e-01
2.97956407e-01 1.10607505e-01 5.57290673e-01 1.34130800e+00
-2.75902838e-01 -5.90890571e-02 -7.84137249e-01 2.42886886e-01
-2.06337953e+00 -1.13620663e+00 -4.01673287e-01 2.06522346e+00
2.65963674e-01 9.97524522e-03 9.76627395e-02 2.85616696e-01
8.71910453e-01 4.30218548e-01 -4.31253135e-01 -4.88551930e-02
-2.92264074e-01 -5.49280465e-01 2.10148226e-02 1.16919987e-02
-1.41513431e+00 5.51967025e-01 5.37561417e+00 9.90719676e-01
-8.03690255e-01 2.81278521e-01 5.41871548e-01 1.87775418e-01
-1.52357817e-01 -1.13941617e-01 -1.01219702e+00 1.96339056e-01
5.82934260e-01 -2.03618899e-01 1.95138618e-01 9.25656497e-01
2.72049636e-01 -3.42317939e-01 -8.89710486e-01 1.21338427e+00
6.75954878e-01 -7.24029660e-01 3.50007683e-01 1.99738190e-01
9.29308355e-01 -3.84937584e-01 1.09805301e-01 3.04942727e-01
5.07979281e-02 -6.04588032e-01 2.60355264e-01 5.22915006e-01
5.34633517e-01 -7.41442978e-01 6.98960006e-01 9.44933772e-01
-1.49736214e+00 -2.57730484e-01 -4.20361519e-01 2.00042486e-01
2.59957723e-02 6.97505474e-01 -5.01356125e-01 1.33444202e+00
6.29675210e-01 1.21470916e+00 -7.87792921e-01 5.25398552e-01
2.86056250e-02 3.79993618e-01 1.26027912e-01 5.17559946e-01
1.48923695e-01 -4.93015587e-01 4.50179279e-01 9.01770294e-01
2.11477548e-01 -1.45586386e-01 6.61288798e-01 3.14922929e-01
4.20494318e-01 3.31543028e-01 -1.07319140e+00 3.52513582e-01
-5.11361770e-02 1.48260391e+00 -6.88850641e-01 -4.97954398e-01
-8.55949402e-01 8.76300216e-01 2.04762548e-01 3.91504973e-01
-6.63811088e-01 3.11776012e-01 1.34062663e-01 -3.57038528e-01
3.03429484e-01 -1.19227499e-01 -2.26486564e-01 -1.53651285e+00
1.35330737e-01 -1.00462353e+00 6.16297424e-01 -7.07455397e-01
-1.45683682e+00 4.59445387e-01 1.26139060e-01 -1.71774817e+00
-1.22712232e-01 -4.21201259e-01 -2.89256454e-01 7.87565827e-01
-1.32175744e+00 -1.35947740e+00 -3.78532290e-01 1.03445876e+00
1.01395857e+00 -7.67191172e-01 7.98635185e-01 1.51216045e-01
-5.96170843e-01 4.69480567e-02 4.18525696e-01 -1.23752825e-01
1.06959236e+00 -1.25632608e+00 -3.36366385e-01 6.65098310e-01
5.16663432e-01 3.54647815e-01 1.92441881e-01 -6.10306978e-01
-1.32850349e+00 -1.11735618e+00 5.18760622e-01 -5.24372280e-01
3.03067356e-01 -2.96757549e-01 -9.28727508e-01 7.87372291e-01
2.22368702e-01 -1.48611099e-01 1.08087051e+00 4.08058316e-01
-3.25158328e-01 -3.24711710e-01 -5.16519666e-01 1.38262868e-01
6.20256603e-01 -5.61987162e-01 -8.27646673e-01 5.52291095e-01
3.74359429e-01 1.39762700e-01 -9.22076404e-01 5.26978970e-01
3.73260140e-01 -1.30811334e+00 8.64184678e-01 -5.33111513e-01
3.94782007e-01 -4.39997882e-01 -3.09204102e-01 -1.44535553e+00
-5.04357040e-01 -2.11170875e-02 3.11964788e-02 1.38016903e+00
1.24714181e-01 -4.13505375e-01 6.71903670e-01 -9.06046629e-02
-1.67815655e-01 -6.62702858e-01 -6.84136093e-01 -7.41093874e-01
-2.53148347e-01 -3.03220838e-01 -1.29089234e-02 1.19469059e+00
-3.40568215e-01 7.58881867e-01 -7.85820544e-01 2.51272470e-01
9.65467691e-01 6.32489622e-01 8.62654567e-01 -1.50472605e+00
-3.73144805e-01 1.59135163e-01 -1.44819751e-01 -7.10346520e-01
3.25954080e-01 -9.41346228e-01 -3.38815391e-01 -1.54762101e+00
7.82953560e-01 -2.66447067e-01 -4.12892640e-01 3.66001844e-01
-4.81383860e-01 -5.93559556e-02 1.91673368e-01 6.09934747e-01
-8.70728135e-01 6.68685973e-01 1.57356977e+00 -3.44813652e-02
-1.30823806e-01 4.17831510e-01 -5.28128684e-01 1.01411366e+00
5.24029434e-01 -6.49845958e-01 -7.63454497e-01 -3.11576098e-01
-6.15814663e-02 3.58676434e-01 7.68470019e-02 -1.09774423e+00
2.10503519e-01 -2.24977866e-01 6.15196645e-01 -8.75727475e-01
5.78181267e-01 -1.22864079e+00 3.04142863e-01 1.39616668e-01
-7.60284588e-02 9.14938971e-02 -2.10105360e-01 9.73838210e-01
-5.62228262e-01 -2.01566875e-01 7.51942992e-01 -4.99709457e-01
-9.36906099e-01 1.45151094e-01 -2.64927149e-01 4.48258258e-02
1.34667861e+00 -4.31719124e-01 -4.25473712e-02 -1.91320419e-01
-1.10536397e+00 2.22062543e-01 1.11563839e-01 5.04094005e-01
6.06391907e-01 -1.38733912e+00 -6.11240208e-01 4.26982343e-01
2.61950850e-01 -6.31612167e-02 6.70661092e-01 7.77856171e-01
3.96904022e-01 3.40409160e-01 -4.82729882e-01 -9.30293739e-01
-1.72934735e+00 9.14506376e-01 -7.35277906e-02 -4.65735853e-01
-4.23929691e-01 3.97589386e-01 7.44262159e-01 -4.44273353e-01
-2.14632750e-02 2.25597933e-01 -8.73801887e-01 6.92662299e-01
3.39941680e-01 3.24188858e-01 -9.83633026e-02 -8.68247271e-01
-1.09951623e-01 9.96603072e-01 -2.03237429e-01 -2.57434845e-02
1.44011760e+00 -3.90725344e-01 -2.09372625e-01 1.20317447e+00
1.20065415e+00 6.58588763e-03 -9.95080829e-01 -5.42201817e-01
-1.41163962e-02 -3.93964410e-01 -1.93316266e-01 -4.69065458e-01
-1.15029538e+00 1.01017022e+00 5.18643320e-01 1.90775573e-01
1.34105444e+00 -6.01637969e-03 2.17863008e-01 4.66543704e-01
3.41953903e-01 -9.98264730e-01 5.41347980e-01 4.11160231e-01
7.58506417e-01 -1.54169190e+00 3.33249182e-01 -4.96898055e-01
-1.13653183e+00 1.13365841e+00 8.49330366e-01 2.11368725e-01
7.78003037e-01 -9.72509384e-02 1.77937359e-01 -3.02395046e-01
-9.33034062e-01 -1.36009470e-01 4.24683958e-01 4.98442590e-01
1.86302796e-01 -8.59983265e-02 1.11026548e-01 4.88991678e-01
3.42056334e-01 -1.77545443e-01 3.39432210e-01 7.70664871e-01
-5.75075328e-01 -1.29998672e+00 -6.25059009e-01 4.58008468e-01
-4.60361987e-01 3.24885279e-01 -3.54984924e-02 7.12337792e-01
4.55837369e-01 9.16700780e-01 -4.34230715e-01 -5.16827345e-01
2.38172799e-01 3.41773808e-01 2.74500132e-01 -8.61063004e-01
-4.76519197e-01 6.79631352e-01 -1.59917936e-01 -3.47737879e-01
-1.06891739e+00 -8.78410161e-01 -8.31256390e-01 3.04952860e-01
-4.45856988e-01 2.49663636e-01 5.38134873e-01 1.16350663e+00
-2.82212421e-02 5.31413138e-01 1.08534992e+00 -6.61165774e-01
-5.31036854e-01 -8.56110513e-01 -9.71014857e-01 6.00973904e-01
-1.26406392e-02 -8.77973437e-01 -3.35750729e-01 4.06869709e-01] | [8.475813865661621, 4.515131950378418] |
7bf29b5e-a3bc-40f2-bd7a-54da2f261bfa | gaussian-kernel-smoothing | 2007.09539 | null | https://arxiv.org/abs/2007.09539v4 | https://arxiv.org/pdf/2007.09539v4.pdf | Gaussian kernel smoothing | Image acquisition and segmentation are likely to introduce noise. Further image processing such as image registration and parameterization can introduce additional noise. It is thus imperative to reduce noise measurements and boost signal. In order to increase the signal-to-noise ratio (SNR) and smoothness of data required for the subsequent random field theory based statistical inference, some type of smoothing is necessary. Among many image smoothing methods, Gaussian kernel smoothing has emerged as a de facto smoothing technique among brain imaging researchers due to its simplicity in numerical implementation. Gaussian kernel smoothing also increases statistical sensitivity and statistical power as well as Gausianness. Gaussian kernel smoothing can be viewed as weighted averaging of voxel values. Then from the central limit theorem, the weighted average should be more Gaussian. | ['Moo. K. Chung'] | 2020-07-19 | null | null | null | null | ['image-smoothing'] | ['computer-vision'] | [ 1.86179638e-01 -1.17370859e-01 6.04370423e-02 -5.64679503e-01
-5.60952783e-01 -1.91691443e-01 3.36185038e-01 4.78264928e-01
-8.89847219e-01 7.70535111e-01 1.31149679e-01 -1.75186232e-01
-2.42723823e-01 -5.86412311e-01 -2.14427039e-01 -1.03744161e+00
-3.30048680e-01 -1.10514641e-01 5.59889197e-01 2.35495761e-01
3.41906667e-01 7.43691623e-01 -9.34998393e-01 -4.23612148e-01
1.37992024e+00 7.78531134e-01 4.88256127e-01 3.92063290e-01
-7.47270212e-02 3.30183685e-01 -3.37114513e-01 -1.89380899e-01
3.77938412e-02 -3.43641013e-01 -6.06179476e-01 1.94167756e-02
1.25914216e-01 -2.13287458e-01 -2.33478770e-01 1.44269466e+00
5.65810084e-01 7.53508031e-01 8.03558171e-01 -6.45896256e-01
-6.84462309e-01 5.26133418e-01 -7.94623196e-01 5.90772271e-01
-3.63009185e-01 1.78666607e-01 5.14539331e-02 -5.09179652e-01
4.01686847e-01 8.19451571e-01 4.93051916e-01 3.08231533e-01
-1.53761876e+00 -4.54269916e-01 -3.51759642e-01 6.99025542e-02
-1.39057541e+00 -3.52505296e-01 8.04633319e-01 -4.03506070e-01
4.32639271e-01 2.29661182e-01 5.73878884e-01 3.86917889e-01
6.88547790e-01 3.41374159e-01 1.52595782e+00 -4.04509336e-01
4.50543821e-01 -6.59227595e-02 3.09475571e-01 3.52598310e-01
5.13984203e-01 -5.33994026e-02 1.45631269e-01 -4.49166298e-02
1.20202029e+00 -2.44418308e-01 -2.99280196e-01 -1.58015728e-01
-1.14223027e+00 7.17664719e-01 3.60015273e-01 6.89210594e-01
-7.76099741e-01 1.72343180e-01 4.31130052e-01 -2.75606692e-01
4.31554347e-01 2.64978170e-01 1.22716904e-01 -2.31303331e-02
-1.15281641e+00 7.43317381e-02 2.22560316e-01 2.27591038e-01
3.84911180e-01 2.98636675e-01 -1.24080636e-01 9.02465403e-01
2.39524379e-01 7.42271006e-01 4.11459535e-01 -1.14835167e+00
-2.07341731e-01 1.02515087e-01 -4.97234613e-02 -9.17546272e-01
-6.16989493e-01 -4.10120100e-01 -1.07733476e+00 5.09229064e-01
6.96511328e-01 -1.63916752e-01 -9.87031341e-01 1.45179546e+00
2.10949495e-01 1.04952455e-01 -3.93489897e-01 9.03471351e-01
4.29197997e-01 2.57345885e-01 5.13006449e-01 -5.45893967e-01
1.35466385e+00 -1.92652181e-01 -1.18048501e+00 -5.74391559e-02
2.72139698e-01 -9.55509543e-01 8.14376295e-01 2.05217019e-01
-1.29604304e+00 -2.84828335e-01 -8.04935634e-01 2.65881926e-01
-1.18840359e-01 -3.14032704e-01 7.03784049e-01 8.09095383e-01
-8.40554178e-01 6.01033986e-01 -1.28542542e+00 -1.29194304e-01
6.63170278e-01 3.61690253e-01 -5.13866484e-01 5.35140047e-03
-8.46424639e-01 1.27840960e+00 3.75105023e-01 2.78382272e-01
-8.74795616e-02 -7.93192327e-01 -7.92428911e-01 -5.99639714e-02
4.19660956e-02 -3.66944015e-01 9.55777586e-01 -4.29965705e-01
-1.33953297e+00 7.57346213e-01 -3.55363876e-01 -4.45167035e-01
2.04787090e-01 1.94712415e-01 -4.70435828e-01 3.36940855e-01
6.08788431e-02 4.18811738e-01 8.39839280e-01 -9.10691917e-01
-1.65362477e-01 -6.20776594e-01 -5.64448416e-01 -9.16938931e-02
2.61956960e-01 3.43149722e-01 8.00349936e-02 -6.85460806e-01
6.10806346e-01 -5.89900434e-01 -6.94855928e-01 -1.90403566e-01
3.72666754e-02 -1.72795683e-01 3.46243739e-01 -1.01900518e+00
9.78804350e-01 -2.19316649e+00 -5.30138850e-01 4.94094551e-01
5.08346021e-01 6.28571287e-02 1.94396645e-01 -2.53631651e-01
-1.39350235e-01 -1.53085098e-01 -2.33055517e-01 2.26199597e-01
-4.24924791e-01 -1.38622060e-01 3.53951514e-01 1.16819537e+00
3.14609893e-02 7.20748901e-01 -9.00069177e-01 -5.95101416e-01
5.60229599e-01 6.00142598e-01 -3.18560302e-01 -1.66643202e-01
6.13117576e-01 5.49191475e-01 -4.19440925e-01 3.06433678e-01
1.11729908e+00 4.99198139e-02 -2.18804225e-01 -4.20189321e-01
-2.68487096e-01 -8.22140351e-02 -1.14507353e+00 1.14327824e+00
-1.82350084e-01 7.63179958e-01 2.45162398e-01 -1.18410850e+00
9.36242759e-01 3.01329613e-01 7.24136770e-01 -7.33479917e-01
4.57160652e-01 2.09246904e-01 4.11266059e-01 -5.01164019e-01
3.60392094e-01 -5.63639045e-01 4.91385996e-01 2.01364115e-01
-3.08757335e-01 -2.86608756e-01 1.40521321e-02 -6.03074506e-02
7.15531886e-01 -3.14327598e-01 6.20131612e-01 -8.93680036e-01
2.99042881e-01 -7.54769593e-02 4.76117909e-01 5.68116963e-01
-3.85600686e-01 3.66631836e-01 1.53933600e-01 -3.70235890e-02
-1.06378412e+00 -1.28249931e+00 -8.50332081e-01 2.54152119e-01
-2.24277489e-02 2.70806223e-01 -9.85749900e-01 6.38079941e-02
-1.44293129e-01 8.89744699e-01 -4.36378062e-01 -1.20078556e-01
-4.59097296e-01 -9.33959544e-01 3.81176919e-01 3.54587913e-01
5.55649161e-01 -9.20829654e-01 -7.17191339e-01 5.43412268e-01
-2.11347472e-02 -9.63202834e-01 -5.49329102e-01 3.47546577e-01
-1.06920350e+00 -9.45362747e-01 -1.03111470e+00 -4.95555609e-01
9.17858481e-01 2.96393305e-01 3.73875409e-01 5.57935201e-02
-4.96934801e-01 1.95643261e-01 -1.89916492e-01 -3.26816648e-01
-3.85407090e-01 -4.43031460e-01 1.67466670e-01 -2.37040967e-01
3.22357148e-01 -5.88823318e-01 -8.31841111e-01 -2.90630497e-02
-9.50167835e-01 -4.54343349e-01 5.32487214e-01 6.51445150e-01
6.02756679e-01 6.57071948e-01 5.73701560e-01 -5.45814931e-01
1.00367355e+00 -2.90239174e-02 -6.77750051e-01 -9.95763540e-02
-4.83878583e-01 -3.49909775e-02 3.51977348e-01 -5.73186994e-01
-1.23494649e+00 -2.28373215e-01 2.68562417e-02 -1.54475579e-02
-2.87595272e-01 5.23278415e-01 1.41009808e-01 -3.95711303e-01
8.76352310e-01 2.05435008e-01 5.58184087e-01 -2.19772696e-01
1.75481409e-01 5.57465732e-01 6.78392768e-01 -2.27512509e-01
4.38870400e-01 6.65790319e-01 3.22153121e-01 -1.27212143e+00
-2.71897316e-01 -4.57504421e-01 -7.39430845e-01 -4.10721093e-01
1.26086378e+00 -3.08672041e-01 -8.55915964e-01 4.23173606e-01
-9.10416067e-01 -2.78180987e-01 -2.23752465e-02 1.25045490e+00
-4.21994299e-01 6.18553221e-01 -4.42925662e-01 -9.59447026e-01
-2.90164798e-01 -1.51794827e+00 3.69545907e-01 5.90461552e-01
-3.04736197e-01 -1.30874586e+00 -4.56806630e-01 -5.33786640e-02
7.61329114e-01 2.15086386e-01 7.55497098e-01 -1.86057746e-01
-6.30595908e-02 -5.85300982e-01 -3.35134238e-01 4.62437689e-01
4.17016357e-01 -7.64777586e-02 -6.66825056e-01 -4.28772392e-03
4.42301750e-01 2.97560722e-01 5.24302185e-01 1.33672726e+00
1.11198771e+00 2.05091804e-01 -1.35920003e-01 3.43199372e-01
1.21698201e+00 6.00186944e-01 9.82675791e-01 2.28735626e-01
4.29312378e-01 5.11840105e-01 3.46421927e-01 7.67156258e-02
-7.02068806e-02 4.04589742e-01 -1.42454728e-01 -3.07280958e-01
-2.25294754e-01 2.11226344e-01 -3.68703455e-02 6.51239634e-01
-1.55126840e-01 5.07688999e-01 -7.01768517e-01 3.86324048e-01
-1.29904485e+00 -1.22554052e+00 -6.84510112e-01 2.59511161e+00
7.47322142e-01 2.29683276e-02 1.15167022e-01 7.90898278e-02
8.49672973e-01 -1.76509261e-01 -2.89011359e-01 -2.74259031e-01
-4.66751941e-02 2.99900621e-01 1.13230872e+00 7.29829609e-01
-8.85651767e-01 6.35212839e-01 7.03708982e+00 9.88388479e-01
-1.19479680e+00 4.99759279e-02 4.88746643e-01 4.17303324e-01
-1.67203650e-01 -2.48024724e-02 -5.49567938e-01 5.69107234e-01
7.60164797e-01 -5.25750041e-01 5.20098090e-01 4.94295865e-01
9.34262276e-01 -1.03161538e+00 -3.57227921e-01 8.89745295e-01
-3.76224786e-01 -1.11565840e+00 -4.93910372e-01 2.47172907e-01
3.50250244e-01 -1.62888110e-01 -7.85753801e-02 -3.38426709e-01
6.42151311e-02 -1.03909934e+00 2.28177890e-01 7.10261643e-01
6.40350163e-01 -6.37620747e-01 8.56817961e-01 1.83566749e-01
-9.60408449e-01 5.08858562e-01 -7.24739373e-01 1.29281446e-01
5.79574704e-01 1.17772913e+00 -6.43693507e-01 1.11481212e-01
4.56868291e-01 8.91677216e-02 -4.05917019e-01 1.67210162e+00
3.18453796e-02 6.52928591e-01 -5.62321067e-01 3.66579816e-02
1.90910622e-01 -7.04122663e-01 4.89455968e-01 9.86709595e-01
9.20212194e-02 4.76944476e-01 2.42673829e-02 9.13470745e-01
5.51413774e-01 3.62614006e-01 -2.97207505e-01 5.31443991e-02
7.08873451e-01 1.29655504e+00 -1.48131800e+00 -2.55945414e-01
-5.88362217e-01 7.59779692e-01 -1.41922325e-01 3.99210155e-01
-4.98910725e-01 -5.13307810e-01 4.16823775e-01 3.23730469e-01
-1.51423723e-01 -4.57137764e-01 -7.70882189e-01 -6.41759872e-01
-1.83056056e-01 -3.46861064e-01 6.05267026e-02 -4.57406223e-01
-1.21450746e+00 3.80716860e-01 2.71321923e-01 -7.28791416e-01
1.34646118e-01 -4.12603527e-01 -7.58076906e-01 1.35939837e+00
-1.03983223e+00 -6.09783351e-01 -1.57499596e-01 4.73535985e-01
8.03126767e-02 2.13831171e-01 5.77196777e-01 2.82163411e-01
-2.17156276e-01 3.32446337e-01 4.58622798e-02 3.56146805e-02
5.79728901e-01 -1.09643006e+00 -5.13124801e-02 9.60617959e-01
-5.47733247e-01 9.87989485e-01 9.51754332e-01 -9.22642529e-01
-1.07567406e+00 -5.98061621e-01 5.97355783e-01 2.07144842e-01
1.00213420e+00 1.49916887e-01 -1.28215039e+00 3.44542861e-01
1.13047380e-02 7.98887014e-03 5.56067944e-01 -1.97876871e-01
3.68018270e-01 2.55646035e-02 -1.46448290e+00 7.97844112e-01
3.50021452e-01 -2.95991987e-01 -5.84657490e-01 9.94891971e-02
2.05486357e-01 -3.00135136e-01 -1.43342328e+00 2.35114187e-01
4.77828026e-01 -7.79598892e-01 8.34238350e-01 -1.36753067e-01
-1.89778700e-01 -3.01642448e-01 7.19661191e-02 -1.43088412e+00
-6.73277020e-01 -5.08743405e-01 8.55446756e-01 1.08793092e+00
2.33633220e-01 -8.98787141e-01 7.51625538e-01 1.13802588e+00
-1.67692944e-01 -2.65158534e-01 -1.13168335e+00 -9.85464454e-01
1.76793590e-01 -5.97082615e-01 2.04271451e-01 7.36716151e-01
3.00285876e-01 -1.84284925e-01 -1.14259860e-02 8.46048370e-02
1.18712509e+00 -5.27451277e-01 1.89731598e-01 -1.16053939e+00
3.04963291e-01 -6.54296637e-01 -5.36520243e-01 -5.26636124e-01
-2.02577021e-02 -9.11159754e-01 3.73490825e-02 -1.60439718e+00
6.55448288e-02 -2.59269267e-01 -2.55815256e-02 5.93171231e-02
-3.52315664e-01 8.92476663e-02 -1.00498810e-01 -9.46397781e-02
1.80355042e-01 1.44553021e-01 1.63633609e+00 3.32504570e-01
-4.46616620e-01 2.24722043e-01 -4.85933036e-01 8.28257680e-01
9.94120300e-01 -4.37790006e-01 -2.89384067e-01 9.07965377e-02
-3.09604704e-01 2.20856681e-01 2.76300311e-01 -8.61569643e-01
2.06347287e-01 -1.77881777e-01 5.07340789e-01 -4.69175428e-01
4.59081568e-02 -8.86437058e-01 2.55639434e-01 5.38582921e-01
-7.56755322e-02 -2.20224664e-01 1.30236000e-01 1.20414063e-01
-2.11752374e-02 -5.91425598e-01 1.35452557e+00 -2.41684597e-02
-4.99755263e-01 1.73393056e-01 -8.54245007e-01 -1.92682371e-01
9.23807085e-01 -5.05674005e-01 -1.45124301e-01 -3.06291848e-01
-8.63090396e-01 -1.21392615e-01 3.60390306e-01 -2.55813062e-01
6.42417610e-01 -1.23372924e+00 -4.92234200e-01 -5.00052311e-02
-3.46712738e-01 -2.74836987e-01 7.15964854e-01 1.76661539e+00
-5.74546814e-01 2.74325639e-01 -3.14632148e-01 -3.64094913e-01
-1.13970494e+00 4.56420124e-01 3.02726120e-01 1.52078331e-01
-9.78883862e-01 6.03214264e-01 -9.21178535e-02 2.76973993e-01
1.20039936e-02 -2.63974041e-01 -2.02700481e-01 -1.42067432e-01
9.36671674e-01 7.74430513e-01 -5.44176102e-02 -6.46703660e-01
-4.43449706e-01 3.87471884e-01 -4.92938310e-02 -1.64795309e-01
1.12491560e+00 -2.66310573e-01 -4.72438961e-01 1.64995760e-01
1.01366782e+00 1.21976346e-01 -1.15592587e+00 -6.29380569e-02
5.32811284e-02 -4.95058239e-01 8.53625894e-01 -5.47126353e-01
-8.43139470e-01 6.99867010e-01 7.96667159e-01 4.87410367e-01
1.02659714e+00 -7.19025880e-02 5.15674114e-01 -1.92544803e-01
3.30617934e-01 -1.29204834e+00 -5.92122793e-01 7.26513639e-02
5.83928764e-01 -1.11373842e+00 6.04193509e-02 -5.42857826e-01
-6.55768037e-01 1.08933342e+00 7.41324648e-02 -4.23645347e-01
9.53655422e-01 7.01652348e-01 4.35166098e-02 -1.64844185e-01
1.62625045e-01 -3.59306663e-01 5.17046213e-01 1.00788498e+00
7.71914482e-01 3.03621262e-01 -1.00254619e+00 5.30502677e-01
-1.41783670e-01 3.19272727e-01 4.65290278e-01 5.23276150e-01
-8.15956295e-01 -8.02727640e-01 -6.47882879e-01 8.83302391e-01
-6.72694802e-01 -9.32287350e-02 5.59183061e-01 7.04376876e-01
-3.02064478e-01 1.02845967e+00 3.65677439e-02 3.74948770e-01
2.13280737e-01 -1.92495048e-01 7.06708550e-01 -1.65471539e-01
-1.03229061e-01 5.36546946e-01 -2.90033400e-01 -4.64177847e-01
-4.40808058e-01 -7.95181692e-01 -1.86047947e+00 -5.67515135e-01
-4.10667509e-01 1.47548288e-01 9.88793969e-01 1.02833927e+00
-2.20558211e-01 6.10020936e-01 1.99503109e-01 -7.88173020e-01
-4.47974741e-01 -1.00033855e+00 -1.29193723e+00 1.87368035e-01
-6.99792132e-02 -7.58562803e-01 -3.00753951e-01 9.25239101e-02] | [13.997464179992676, -2.369415044784546] |
7ecb2385-5d85-4dd2-af26-256bb757b220 | deep-neural-networks-improve-radiologists | 1903.08297 | null | http://arxiv.org/abs/1903.08297v1 | http://arxiv.org/pdf/1903.08297v1.pdf | Deep Neural Networks Improve Radiologists' Performance in Breast Cancer Screening | We present a deep convolutional neural network for breast cancer screening
exam classification, trained and evaluated on over 200,000 exams (over
1,000,000 images). Our network achieves an AUC of 0.895 in predicting whether
there is a cancer in the breast, when tested on the screening population. We
attribute the high accuracy of our model to a two-stage training procedure,
which allows us to use a very high-capacity patch-level network to learn from
pixel-level labels alongside a network learning from macroscopic breast-level
labels. To validate our model, we conducted a reader study with 14 readers,
each reading 720 screening mammogram exams, and find our model to be as
accurate as experienced radiologists when presented with the same data.
Finally, we show that a hybrid model, averaging probability of malignancy
predicted by a radiologist with a prediction of our neural network, is more
accurate than either of the two separately. To better understand our results,
we conduct a thorough analysis of our network's performance on different
subpopulations of the screening population, model design, training procedure,
errors, and properties of its internal representations. | ['Naziya Samreen', 'Beatriu Reig', 'Leng Leng Young Lin', 'Thibault Févry', 'Stanisław Jastrzębski', 'Ujas Parikh', 'Nan Wu', 'Laura Heacock', 'Kyunghyun Cho', 'Krzysztof J. Geras', 'Krystal Airola', 'Jungkyu Park', 'Joshua D. Weinstein', 'Joe Katsnelson', 'Alana Lewin', 'Yiming Gao', 'Stacey Wolfson', 'Linda Moy', 'Kara Ho', 'Esther Hwang', 'Eric Kim', 'Stephanie Chung', 'S. Gene Kim', 'Masha Zorin', 'Kristine Pysarenko', 'Jiyon Lee', 'Hildegard Toth', 'Zhe Huang', 'Yiqiu Shen', 'Sushma Gaddam', 'Jason Phang', 'Eralda Mema'] | 2019-03-20 | null | null | null | null | ['breast-cancer-detection', 'breast-cancer-detection'] | ['knowledge-base', 'medical'] | [ 5.37230670e-01 6.43352449e-01 -4.53101158e-01 -5.03402114e-01
-1.05830920e+00 -5.61702430e-01 1.64296776e-02 6.70806646e-01
-4.69376862e-01 3.21799994e-01 8.98110196e-02 -1.22603095e+00
-7.97214359e-02 -1.02968299e+00 -1.15012741e+00 -3.61809522e-01
-2.48776302e-01 3.52147698e-01 3.77891362e-01 3.93903404e-01
-3.01450908e-01 2.82271445e-01 -8.51149857e-01 6.45110965e-01
7.66902626e-01 1.27046406e+00 -1.13266341e-01 1.04703498e+00
3.91454160e-01 1.25676501e+00 -4.13682759e-01 -6.45212471e-01
1.05206452e-01 -4.37736422e-01 -1.01286042e+00 2.46231407e-02
6.11720085e-01 -5.68327427e-01 -4.60333735e-01 7.91931570e-01
3.32289338e-01 -3.56199116e-01 8.23432148e-01 -2.51081526e-01
-5.22228479e-01 5.27957797e-01 -4.54293221e-01 4.57362056e-01
-1.01759113e-01 1.88883200e-01 7.23748684e-01 -1.51262835e-01
4.07441169e-01 6.64897203e-01 1.29306912e+00 4.58124340e-01
-1.06166995e+00 -4.02745157e-01 -2.68016547e-01 -3.23505133e-01
-8.70047987e-01 3.21319094e-03 -5.66559471e-02 -4.96035725e-01
4.41581547e-01 4.54474509e-01 1.04160750e+00 8.72590840e-01
7.72326052e-01 7.54316449e-01 1.03829634e+00 -4.41162616e-01
7.46421516e-02 2.78129756e-01 4.88850653e-01 1.15862012e+00
7.09315717e-01 1.74848914e-01 1.72078893e-01 -4.55303580e-01
9.60037410e-01 4.16985005e-02 -9.49422829e-03 -2.88606524e-01
-1.04064310e+00 8.12044084e-01 8.70756328e-01 1.22915126e-01
-3.03690076e-01 2.56122172e-01 3.36796671e-01 8.08295533e-02
1.98963642e-01 5.67219913e-01 -1.21277608e-01 3.39463562e-01
-9.26343441e-01 -2.69572884e-01 8.49051774e-01 3.96615684e-01
9.53845680e-03 -5.49722135e-01 -4.25754935e-01 6.99678898e-01
4.92568165e-02 3.31216455e-01 7.01848686e-01 -8.48071456e-01
-1.61196634e-01 6.51382923e-01 -1.41666204e-01 -8.05593967e-01
-9.14558589e-01 -8.24804425e-01 -9.46752548e-01 -1.03337504e-01
4.84897733e-01 -4.56255376e-01 -1.23499322e+00 1.32531369e+00
-1.45625114e-01 -1.24475226e-01 3.14595387e-03 4.68585193e-01
1.05025482e+00 1.88609898e-01 4.01046872e-01 2.57044554e-01
1.77750468e+00 -9.21666861e-01 -1.92532822e-01 -2.90192336e-01
1.10158002e+00 -3.40028822e-01 4.23405260e-01 2.30206326e-01
-1.27268922e+00 -5.80977142e-01 -1.03973913e+00 9.33181942e-02
-1.66625202e-01 6.38146102e-01 8.41251969e-01 9.15098011e-01
-1.10903788e+00 6.30934060e-01 -1.28909862e+00 -5.35071015e-01
7.50162423e-01 4.67759371e-01 -2.98603892e-01 -1.83463112e-01
-9.24268067e-01 1.01465452e+00 2.06425890e-01 -1.43979594e-01
-9.26842630e-01 -6.68856025e-01 -8.17115664e-01 2.34575748e-01
-7.29508474e-02 -8.25621963e-01 1.62088597e+00 -8.90633464e-01
-8.12153995e-01 1.16006637e+00 -1.60341337e-01 -7.73242772e-01
6.11957610e-01 3.59246105e-01 -3.39150488e-01 3.11333656e-01
1.25634551e-01 7.17148602e-01 2.36187398e-01 -1.04908371e+00
-8.74065697e-01 -2.23075092e-01 -1.00808948e-01 -1.38085067e-01
-2.42553562e-01 -2.49285460e-01 -5.05343676e-01 -3.30206990e-01
2.12328389e-01 -9.77430403e-01 -5.04325092e-01 1.20276019e-01
-3.34133506e-01 1.36424959e-01 8.74192566e-02 -8.92965615e-01
1.04154348e+00 -2.26914024e+00 -6.23531103e-01 5.22044480e-01
5.21439314e-01 1.86151341e-01 8.92525017e-02 -4.26442146e-01
-2.45255187e-01 3.91778678e-01 -1.77707061e-01 1.28662392e-01
-5.09832263e-01 -5.44642545e-02 4.01079655e-01 3.72693062e-01
5.46970427e-01 1.22404814e+00 -8.90691638e-01 -6.11271560e-01
1.42730894e-02 1.75353020e-01 -5.43285370e-01 1.35417446e-01
3.51062983e-01 1.98350385e-01 -4.64644998e-01 7.34401166e-01
4.07969356e-01 -1.16996467e+00 4.55382764e-01 -2.18035832e-01
3.44160348e-01 1.89794347e-01 -4.95855033e-01 1.15294766e+00
-3.85449767e-01 6.49200201e-01 -1.08603738e-01 -8.09511781e-01
5.69394171e-01 3.75383258e-01 4.20733392e-01 -6.59548342e-01
2.35516861e-01 5.01000881e-01 5.71224868e-01 -4.39983547e-01
4.18517627e-02 -8.81612897e-02 1.80640087e-01 3.81258905e-01
1.20670259e-01 1.44516259e-01 1.76998481e-01 6.34671450e-02
1.82975924e+00 -4.30636466e-01 4.05440092e-01 -3.76337558e-01
1.02301002e-01 3.96159559e-01 1.68054670e-01 1.39036417e+00
-3.47035795e-01 6.77546799e-01 9.34138715e-01 -7.33384967e-01
-8.24735045e-01 -1.11910367e+00 -8.76327991e-01 6.15005612e-01
-6.50469214e-02 7.77927116e-02 -4.52683687e-01 -9.44457889e-01
1.65600836e-01 3.35601449e-01 -1.27438354e+00 -3.49121392e-01
-5.59080958e-01 -1.29008162e+00 6.08196676e-01 8.17994773e-01
3.25954229e-01 -7.36464679e-01 -9.20162559e-01 1.76206142e-01
3.12019363e-02 -6.92759633e-01 3.93135771e-02 7.10214257e-01
-9.77643430e-01 -1.45922697e+00 -1.04251671e+00 -1.02764750e+00
9.01190281e-01 -7.88706914e-02 1.45116866e+00 3.65952492e-01
-5.72572589e-01 2.94645369e-01 -1.36097893e-01 -6.30345464e-01
-9.35496747e-01 2.90190130e-01 -6.53306544e-01 -4.22456890e-01
4.99039710e-01 1.64171785e-01 -8.57826293e-01 1.94776446e-01
-8.54872406e-01 6.59046471e-02 1.04087377e+00 1.08908963e+00
5.31793535e-01 -1.67569906e-01 2.05955446e-01 -1.25686026e+00
4.80858326e-01 -7.48745799e-01 -3.22438627e-01 4.74698007e-01
-3.11513662e-01 -2.36386254e-01 1.26844734e-01 -2.95202285e-01
-7.01338351e-01 1.72022343e-01 -1.75017104e-01 1.70490742e-01
-1.22234657e-01 5.76862931e-01 9.45385277e-01 -2.83671618e-01
9.67363238e-01 -6.56052157e-02 2.33770207e-01 -2.66244769e-01
-4.08709079e-01 5.82383215e-01 5.94541132e-01 -2.88876206e-01
2.17263848e-01 3.50482315e-01 2.89735794e-01 -3.48757058e-01
-1.15562248e+00 -3.95543277e-01 -3.71303886e-01 8.26443022e-04
9.75722015e-01 -9.88929808e-01 -5.88482738e-01 2.62510300e-01
-6.78552985e-01 -5.25646389e-01 -4.09227043e-01 6.90679312e-01
-6.09262548e-02 -1.34099405e-02 -1.07805002e+00 -5.02002537e-01
-2.28502810e-01 -1.38857090e+00 9.67146039e-01 3.75413567e-01
-2.45103940e-01 -1.16180074e+00 -1.03133895e-01 9.48449448e-02
5.16452909e-01 5.16600430e-01 1.17030072e+00 -9.70518053e-01
-2.01990530e-01 -6.68921053e-01 -6.44750297e-01 2.55924344e-01
3.43878150e-01 3.17514203e-02 -9.01290357e-01 -2.79563218e-01
-1.75290704e-01 -4.80571061e-01 1.27595365e+00 9.61194634e-01
1.84311044e+00 2.86133230e-01 -7.91065693e-01 6.93132520e-01
1.29321921e+00 2.19903320e-01 4.11767155e-01 3.95361871e-01
2.60255963e-01 2.66812950e-01 1.09745055e-01 -7.67938271e-02
2.90281892e-01 -7.32641816e-02 3.96848500e-01 -8.76894057e-01
-1.57535285e-01 -1.55267701e-01 -3.54625791e-01 3.84077072e-01
3.16430591e-02 -1.79160178e-01 -1.23551989e+00 5.13544440e-01
-1.29069042e+00 -3.41174573e-01 -1.34740770e-01 2.04834533e+00
6.77398622e-01 3.92722458e-01 4.95235920e-02 -1.89767510e-01
5.99713981e-01 -2.19191432e-01 -5.41655779e-01 -3.35864246e-01
7.75883570e-02 3.40274483e-01 9.80840504e-01 9.59022790e-02
-1.45717847e+00 8.25356618e-02 8.21048641e+00 3.84002298e-01
-1.07601976e+00 1.08823432e-02 1.72998703e+00 1.39117986e-01
-2.80821379e-02 -4.95840460e-01 -4.20846194e-01 3.17343801e-01
1.21399522e+00 2.60187864e-01 -2.27119684e-01 7.47224092e-01
-2.94057876e-01 -3.87997508e-01 -1.34518182e+00 2.86622524e-01
-3.58931385e-02 -1.65791118e+00 -2.15714887e-01 8.83146524e-02
8.21442842e-01 2.35982582e-01 1.49270892e-01 3.32640350e-01
5.78006387e-01 -1.37870920e+00 1.99888140e-01 4.28142279e-01
1.02355027e+00 -1.98840752e-01 1.31278169e+00 1.30102322e-01
-6.77008271e-01 -1.17104709e-01 -8.75744373e-02 2.55877972e-01
-4.42714870e-01 3.54904205e-01 -1.01788592e+00 3.25587362e-01
9.04676139e-01 4.84356314e-01 -1.12479591e+00 1.25551343e+00
3.29458117e-01 1.00259483e+00 -3.78300786e-01 -5.63448071e-02
3.55902910e-01 5.19818664e-01 -3.32436174e-01 1.54169154e+00
3.59158397e-01 1.12346582e-01 -1.57048255e-01 6.64754629e-01
-1.81591958e-01 8.88590813e-02 -2.89440900e-01 7.25413412e-02
3.48191671e-02 1.33327365e+00 -9.16739047e-01 -6.05073154e-01
-5.75830102e-01 4.73332316e-01 1.39630303e-01 7.37164542e-02
-7.71880984e-01 -1.06106631e-01 -2.43560389e-01 1.48848593e-01
2.16226578e-02 5.35319388e-01 -4.71573979e-01 -7.46738970e-01
-1.78413987e-01 -7.85580873e-01 6.51500821e-01 -6.97185457e-01
-1.37220240e+00 4.56683874e-01 -3.24004203e-01 -1.08817124e+00
-1.00143678e-01 -1.20749140e+00 -6.61171377e-01 7.59561598e-01
-1.54924846e+00 -7.29644775e-01 -4.59812641e-01 3.80819850e-02
-6.10107742e-02 -4.99323867e-02 9.22077239e-01 3.00514519e-01
-5.34063697e-01 8.04183185e-01 1.64466813e-01 7.99271464e-01
6.15030885e-01 -1.44310462e+00 2.06421614e-01 4.04117405e-01
-4.99512017e-01 5.41218340e-01 -5.74970990e-02 -5.39756835e-01
-1.03782868e+00 -1.27386189e+00 5.59200466e-01 -5.89910626e-01
5.93480289e-01 3.47561985e-01 -7.58695126e-01 1.04233873e+00
-6.03442118e-02 4.15075392e-01 9.64068949e-01 9.95705128e-02
2.57789437e-02 6.47769719e-02 -1.22789836e+00 3.14615220e-01
6.16675496e-01 4.97973673e-02 -3.66329819e-01 4.50141400e-01
3.87068659e-01 -9.12413538e-01 -1.27637327e+00 8.73782337e-01
7.22205341e-01 -9.36030507e-01 7.76646197e-01 -7.09880173e-01
9.59536493e-01 3.91767830e-01 1.74055755e-01 -9.87452626e-01
-5.42482376e-01 2.00846910e-01 1.08886883e-01 2.89099663e-01
8.70046973e-01 -6.44828439e-01 9.27836180e-01 6.15495861e-01
-3.17915739e-03 -1.27070200e+00 -6.47484183e-01 -2.89586693e-01
4.53714997e-01 -4.97114994e-02 3.41880053e-01 6.57426834e-01
-1.46813795e-01 -1.68355674e-01 1.95945978e-01 2.37912446e-01
3.19154084e-01 -1.14316687e-01 2.74942130e-01 -1.13264513e+00
-6.13937318e-01 -5.79798818e-01 -3.91237468e-01 -9.04442847e-01
-4.34524268e-01 -9.14272666e-01 -1.04685925e-01 -1.68723989e+00
8.03223908e-01 -6.42037928e-01 -6.76061749e-01 4.75533724e-01
-4.14745152e-01 5.85571826e-01 -2.16309324e-01 2.13108540e-01
-4.49491829e-01 -5.05788743e-01 1.56386769e+00 -2.92647719e-01
1.73115820e-01 2.93832809e-01 -8.94804180e-01 7.40498602e-01
5.70983112e-01 -4.19355452e-01 1.42815858e-01 -4.60278064e-01
1.04416154e-01 3.83139580e-01 5.05890489e-01 -1.31742454e+00
-1.67231122e-03 2.58388132e-01 1.26088607e+00 -3.52183282e-01
2.95536071e-02 -6.53961718e-01 -8.78019445e-03 1.20316863e+00
-6.93925083e-01 -2.71092415e-01 4.96472389e-01 4.13596869e-01
-1.60880312e-01 -4.61842418e-01 8.47116351e-01 -4.89957064e-01
-2.33952418e-01 9.39978659e-02 -5.20692229e-01 -2.23014593e-01
8.60386550e-01 -1.38874203e-01 -3.79604310e-01 -1.80802256e-01
-9.01130319e-01 2.73244917e-01 2.17674017e-01 -4.17468809e-02
8.46221894e-02 -1.04990506e+00 -7.99009621e-01 2.93701261e-01
2.35315114e-01 6.50461614e-02 1.11099072e-01 1.07502759e+00
-9.74241674e-01 5.67506850e-01 2.76252739e-02 -9.67409313e-01
-9.82524931e-01 2.18598604e-01 9.39962626e-01 -7.99300373e-01
-3.97147030e-01 8.98062706e-01 3.25095892e-01 -5.13490021e-01
2.17494339e-01 -6.25587344e-01 -9.36792195e-02 -3.87479901e-01
4.03884172e-01 1.16811298e-01 2.84112811e-01 -1.26054902e-02
-1.78140566e-01 3.17600399e-01 -2.49782860e-01 3.49549145e-01
9.25421774e-01 4.39573348e-01 9.80381146e-02 2.52102256e-01
1.07844639e+00 -3.42823178e-01 -9.01863873e-01 -1.40627936e-01
-2.51394421e-01 -2.73061488e-02 2.18817577e-01 -1.11339557e+00
-9.77284610e-01 6.75303698e-01 9.97182012e-01 5.96449018e-01
1.01210666e+00 1.24143392e-01 4.89736676e-01 5.43168247e-01
-9.84963402e-02 -5.84058762e-01 -5.45953512e-02 7.04256222e-02
2.11901143e-01 -1.88824344e+00 6.56238571e-02 -2.90921569e-01
-4.17490780e-01 1.22557020e+00 6.34571373e-01 -3.01374644e-01
6.97172999e-01 2.04817832e-01 1.51780725e-01 -2.42817968e-01
-6.46146297e-01 1.69032529e-01 3.49644393e-01 3.06233406e-01
9.05326545e-01 4.11892414e-01 -1.52611509e-01 5.99150538e-01
-2.03215685e-02 3.33498478e-01 4.98935640e-01 9.53907013e-01
-4.65566844e-01 -6.97060406e-01 -2.99227417e-01 1.42863178e+00
-9.05283988e-01 6.15490340e-02 -1.56369790e-01 1.01465142e+00
1.14626981e-01 7.24803030e-01 4.36484367e-01 -1.14075609e-01
2.34872788e-01 -1.31442640e-02 4.02282149e-01 -7.02634692e-01
-7.49183953e-01 -7.27091283e-02 1.59563571e-01 -3.99862051e-01
-2.57645935e-01 -5.61538637e-01 -9.37305033e-01 -2.35228688e-02
-2.38609403e-01 -1.49970083e-02 4.58993435e-01 6.38694882e-01
4.38579805e-02 1.12667882e+00 4.64636594e-01 -2.88010120e-01
-6.71800613e-01 -9.46559489e-01 -5.59476793e-01 3.69893521e-01
6.04632974e-01 -1.23428665e-01 -2.05649540e-01 -9.57541838e-02] | [15.210101127624512, -2.4319264888763428] |
4fc4d797-5ea5-4333-b076-440ab074cd62 | structured-label-inference-for-visual | 1802.06459 | null | http://arxiv.org/abs/1802.06459v1 | http://arxiv.org/pdf/1802.06459v1.pdf | Structured Label Inference for Visual Understanding | Visual data such as images and videos contain a rich source of structured
semantic labels as well as a wide range of interacting components. Visual
content could be assigned with fine-grained labels describing major components,
coarse-grained labels depicting high level abstractions, or a set of labels
revealing attributes. Such categorization over different, interacting layers of
labels evinces the potential for a graph-based encoding of label information.
In this paper, we exploit this rich structure for performing graph-based
inference in label space for a number of tasks: multi-label image and video
classification and action detection in untrimmed videos. We consider the use of
the Bidirectional Inference Neural Network (BINN) and Structured Inference
Neural Network (SINN) for performing graph-based inference in label space and
propose a Long Short-Term Memory (LSTM) based extension for exploiting activity
progression on untrimmed videos. The methods were evaluated on (i) the Animal
with Attributes (AwA), Scene Understanding (SUN) and NUS-WIDE datasets for
multi-label image classification, (ii) the first two releases of the YouTube-8M
large scale dataset for multi-label video classification, and (iii) the
THUMOS'14 and MultiTHUMOS video datasets for action detection. Our results
demonstrate the effectiveness of structured label inference in these
challenging tasks, achieving significant improvements against baselines. | ['Greg Mori', 'Zicheng Liao', 'Guang-Tong Zhou', 'Hexiang Hu', 'Zhiwei Deng', 'Nelson Nauata'] | 2018-02-18 | null | null | null | null | ['multi-label-image-classification'] | ['computer-vision'] | [ 6.44722760e-01 -1.72841817e-01 -6.37137115e-01 -6.73126519e-01
-4.33292955e-01 -6.57843888e-01 7.34165251e-01 1.39181197e-01
-2.56429434e-01 4.07093853e-01 5.06900072e-01 -1.79372400e-01
-1.33144200e-01 -3.79065841e-01 -7.99856246e-01 -4.62855339e-01
-4.26531374e-01 2.86084712e-01 2.05431297e-01 2.00139701e-01
1.88553721e-01 2.32968971e-01 -1.96658659e+00 1.05603123e+00
-7.74949268e-02 1.58519638e+00 3.71725187e-02 6.72715366e-01
5.21109588e-02 1.79807305e+00 -2.60015130e-01 -1.34185523e-01
1.08857006e-01 -2.51333773e-01 -1.12937164e+00 4.88024622e-01
1.12459970e+00 -3.58791739e-01 -1.84560701e-01 8.63337815e-01
-1.34099443e-02 2.02836573e-01 7.93394685e-01 -1.82487643e+00
-5.02055943e-01 3.65955919e-01 -3.88917267e-01 2.32712328e-01
5.51160991e-01 3.95721719e-02 1.17641783e+00 -4.12975818e-01
1.07786155e+00 1.53892350e+00 8.71389031e-01 4.89575356e-01
-1.22789145e+00 -6.00837111e-01 4.19772178e-01 4.24320698e-01
-1.17206359e+00 -3.97628367e-01 4.56155807e-01 -9.89885330e-01
1.00377858e+00 1.10472441e-01 6.86879754e-01 1.59956431e+00
-6.00356385e-02 8.11152875e-01 1.36322331e+00 -1.23661518e-01
1.75203681e-02 -1.44494981e-01 1.95268497e-01 1.24140382e+00
-3.35564375e-01 1.58932969e-01 -8.78592372e-01 -2.09930986e-01
5.07825613e-01 1.84778005e-01 4.37368415e-02 -3.26514810e-01
-1.69741058e+00 7.83807755e-01 4.35575753e-01 1.44554470e-02
-2.41679177e-01 8.38049710e-01 1.06495273e+00 2.65608490e-01
6.66383803e-01 3.69957536e-01 -5.01605093e-01 -2.36788794e-01
-7.71877825e-01 6.04871102e-02 6.47820711e-01 1.01930904e+00
8.50908279e-01 -1.10245302e-01 -5.30087829e-01 8.38356555e-01
3.82212073e-01 8.59257877e-02 3.69004190e-01 -1.51298630e+00
3.30508620e-01 6.50251031e-01 -1.62722692e-01 -1.02091897e+00
-7.53534853e-01 1.57131836e-01 -5.75832486e-01 2.41660938e-01
2.78117269e-01 1.77060008e-01 -1.26877546e+00 1.91337132e+00
3.01072568e-01 5.80914438e-01 -3.49834383e-01 7.10501671e-01
1.11343849e+00 6.13598168e-01 6.12136900e-01 2.51960188e-01
1.55480421e+00 -1.26735747e+00 -6.45938039e-01 -2.07080454e-01
1.09904349e+00 -8.45823884e-02 8.50528955e-01 2.74678916e-01
-6.60226941e-01 -7.29617298e-01 -7.58438289e-01 -3.34822446e-01
-8.95644784e-01 -2.05480326e-02 6.49925470e-01 2.14472666e-01
-1.24859333e+00 6.13109887e-01 -5.46114802e-01 -6.38886571e-01
7.80696452e-01 2.29647458e-01 -7.60909498e-01 -2.19102904e-01
-1.11655760e+00 7.07368672e-01 5.87807536e-01 -3.12860280e-01
-1.50146377e+00 -6.48397863e-01 -1.13873100e+00 -1.62146330e-01
5.54527879e-01 -4.99953508e-01 8.61724794e-01 -1.08874750e+00
-7.76660442e-01 1.33211684e+00 1.71686009e-01 -4.34401304e-01
1.97846219e-01 9.12049934e-02 -3.48161429e-01 4.86535460e-01
3.26358229e-01 1.38740599e+00 9.19907331e-01 -1.06254566e+00
-9.38701689e-01 -3.65714490e-01 5.11431694e-01 2.11528391e-01
-3.52963835e-01 1.64256603e-01 6.52337000e-02 -7.28313386e-01
-2.05409661e-01 -1.07312465e+00 2.76843458e-01 1.85934186e-01
-4.25062418e-01 -4.16824102e-01 1.21594429e+00 -7.33143568e-01
1.08206093e+00 -2.20298624e+00 2.93316185e-01 -2.49225438e-01
4.86311346e-01 -7.76567310e-02 -5.32599688e-01 5.11305869e-01
-1.39199212e-01 3.48074138e-01 1.85337767e-01 -4.61846948e-01
8.40585157e-02 4.31408554e-01 -1.46966372e-02 5.76405823e-01
8.55220780e-02 1.16198921e+00 -1.10273373e+00 -7.31744111e-01
2.12665841e-01 3.71548861e-01 -3.47472370e-01 2.45986402e-01
-5.91499686e-01 4.86097336e-01 -3.00207704e-01 1.15033329e+00
-2.84687430e-01 -6.21868074e-01 9.78174210e-02 -7.89600849e-01
1.17295764e-01 -3.98961119e-02 -9.02621388e-01 1.94575059e+00
-4.04994011e-01 8.07680666e-01 -8.76295790e-02 -1.04969776e+00
2.83282518e-01 4.61837173e-01 7.11710334e-01 -5.49871266e-01
6.79064393e-02 -2.57983267e-01 -5.52793145e-01 -8.19174349e-01
1.95900172e-01 5.86946979e-02 -2.93321759e-01 5.34338892e-01
4.58704621e-01 4.19346154e-01 4.09914732e-01 4.70962077e-01
1.39559209e+00 5.25321305e-01 2.60416120e-01 -2.78868258e-01
2.00759307e-01 -8.81240442e-02 3.95192176e-01 6.71105027e-01
-3.93045425e-01 2.41199866e-01 6.52764976e-01 -7.96834171e-01
-7.67596900e-01 -7.01837480e-01 -1.20587414e-02 1.95319974e+00
6.86765998e-04 -7.31584311e-01 -3.54190528e-01 -1.04169655e+00
2.52643973e-01 2.65947729e-01 -1.21898866e+00 -5.86602353e-02
-1.76883310e-01 -2.02809587e-01 8.59752357e-01 6.59124553e-01
5.24721444e-01 -1.27318835e+00 -7.04076111e-01 -1.29898667e-01
-4.56417263e-01 -1.48399472e+00 -4.35174137e-01 5.05258381e-01
-5.60116351e-01 -1.32409501e+00 -1.30691126e-01 -7.13933945e-01
3.87109995e-01 1.34382203e-01 1.36604917e+00 7.79185593e-02
-3.97606134e-01 7.60854185e-01 -4.85032499e-01 5.75034320e-02
-4.00238484e-01 -4.11726832e-01 -3.34174149e-02 3.02209198e-01
2.14575186e-01 -3.82393599e-01 -3.02698225e-01 4.42343891e-01
-8.61639917e-01 3.89012843e-01 1.63771823e-01 8.76434028e-01
5.55549324e-01 -3.05938125e-01 5.55101097e-01 -9.81086254e-01
5.16401492e-02 -7.46938586e-01 -3.07021618e-01 3.23454469e-01
-3.80236834e-01 5.66802733e-03 3.01831394e-01 -5.49117863e-01
-7.09707975e-01 1.69963583e-01 2.89034635e-01 -6.91310704e-01
-6.21750891e-01 4.56046700e-01 -1.79355647e-02 -3.22315902e-01
4.95375335e-01 -2.59390444e-01 -1.44759506e-01 -3.04707527e-01
6.94075525e-01 5.94946980e-01 4.11056072e-01 -5.51824152e-01
8.72800797e-02 6.79033518e-01 3.98025900e-01 -5.96986055e-01
-1.31007755e+00 -8.35815668e-01 -8.30729187e-01 -7.36597121e-01
1.69556367e+00 -1.18697178e+00 -8.75286937e-01 5.18961370e-01
-8.56369317e-01 -7.48464763e-01 -1.19572327e-01 8.35947841e-02
-9.16388154e-01 2.87574857e-01 -1.02719676e+00 -2.73044407e-01
1.41454682e-01 -1.21293819e+00 1.57965851e+00 -3.01515669e-01
-4.05320883e-01 -1.29614115e+00 -1.20942384e-01 8.82527530e-01
5.81134707e-02 8.41139317e-01 1.01041937e+00 -6.27325892e-01
-4.64111328e-01 -4.22298647e-02 -6.00006819e-01 3.71145934e-01
-8.61545354e-02 -2.75320888e-01 -1.10536420e+00 -3.70586991e-01
-6.05168641e-01 -1.22444296e+00 1.06213295e+00 2.93364763e-01
1.33839047e+00 -2.45847270e-01 -5.16809106e-01 6.92653596e-01
1.26629615e+00 -3.12442277e-02 2.32693985e-01 3.47965240e-01
1.42680776e+00 5.90523601e-01 5.77923417e-01 4.69710529e-01
5.74661613e-01 8.49352479e-01 9.11741078e-01 -3.62571441e-02
-5.20118594e-01 -1.65470704e-01 3.91236454e-01 3.04302812e-01
-1.81394182e-02 -3.51583779e-01 -8.22563648e-01 4.13407087e-01
-2.04815292e+00 -1.21932280e+00 -1.18326612e-01 1.75761414e+00
5.69077492e-01 -2.00980037e-01 3.14470112e-01 -2.41524473e-01
6.73351705e-01 6.30349398e-01 -6.46724522e-01 -3.48380625e-01
8.89288932e-02 -4.17415679e-01 5.86442292e-01 2.95509305e-03
-1.74468684e+00 9.19543743e-01 5.73693848e+00 7.56835520e-01
-7.43841112e-01 4.41405892e-01 6.21098995e-01 -1.43960953e-01
3.12316418e-01 -9.84189957e-02 -6.98015451e-01 5.75967252e-01
1.36352491e+00 5.95823884e-01 5.42976677e-01 6.90073788e-01
-1.11584723e-01 -1.34687990e-01 -1.45638204e+00 9.94481504e-01
2.62480617e-01 -1.45035326e+00 7.08295479e-02 1.55367211e-01
7.76565254e-01 3.40164900e-01 -1.50478154e-01 3.83811444e-01
4.12011266e-01 -1.05385840e+00 9.76230502e-01 4.75609124e-01
1.06580889e+00 -2.72114009e-01 3.24214190e-01 1.70879483e-01
-1.46425521e+00 -4.95755494e-01 4.06523198e-02 -1.24228537e-01
1.83727905e-01 6.62609637e-02 -4.76216912e-01 2.29676709e-01
8.54348063e-01 1.67663896e+00 -8.82934690e-01 5.06978512e-01
3.55241634e-02 6.01956308e-01 -2.74958909e-02 3.14075351e-01
6.45095766e-01 2.56296583e-02 1.03654593e-01 1.45390248e+00
-4.54932228e-02 -9.15262923e-02 7.31303930e-01 4.24846828e-01
-4.00028884e-01 -3.66366357e-01 -8.06333721e-01 -3.22826684e-01
1.46161571e-01 1.41195214e+00 -9.59605157e-01 -6.36324763e-01
-6.90026939e-01 1.01557529e+00 4.87550884e-01 5.04930556e-01
-1.01764154e+00 1.56370729e-01 6.68694556e-01 2.11497676e-02
1.46975800e-01 6.73482344e-02 2.31432825e-01 -1.05563402e+00
-4.90036398e-01 -8.86603355e-01 8.94561589e-01 -1.15084112e+00
-1.16211915e+00 3.85358214e-01 2.52966493e-01 -1.11347842e+00
-2.91585237e-01 -7.25087583e-01 1.33880347e-01 2.36424625e-01
-1.20300639e+00 -1.78268659e+00 -5.32699108e-01 6.67322993e-01
6.74699903e-01 -7.45778754e-02 8.46314967e-01 6.19838595e-01
-3.10764730e-01 4.07302752e-02 -1.39007732e-01 2.61832744e-01
7.87557364e-01 -1.27343142e+00 5.10436520e-02 3.69162589e-01
6.20847881e-01 -1.49484649e-01 2.38422945e-01 -6.60308480e-01
-1.15581357e+00 -1.60549867e+00 5.25322914e-01 -8.97185802e-01
8.49645197e-01 -6.30270779e-01 -5.07058799e-01 1.21286166e+00
1.66016296e-01 3.65631461e-01 7.82551348e-01 9.59492996e-02
-8.65960717e-01 2.76650846e-01 -9.30702627e-01 2.29777277e-01
1.50999212e+00 -8.99394751e-01 -2.88881123e-01 8.75462055e-01
7.48491704e-01 -1.33466229e-01 -1.30677569e+00 6.08338296e-01
6.88184738e-01 -7.96029627e-01 1.23688960e+00 -9.69564319e-01
6.22644305e-01 -1.40729606e-01 -5.37934542e-01 -1.15088975e+00
-5.47586024e-01 1.74698234e-02 -3.70887727e-01 1.08084702e+00
4.24324768e-03 -1.34295911e-01 5.36523640e-01 3.16275597e-01
-3.55533063e-01 -8.29304457e-01 -6.72128141e-01 -6.09500825e-01
-5.68116307e-01 -4.81065512e-01 1.07322805e-01 1.33874571e+00
-2.13477403e-01 4.55570847e-01 -7.61045635e-01 -1.55578673e-01
7.01331615e-01 2.60575041e-02 3.95834357e-01 -1.40736842e+00
-3.14720511e-03 -2.72424787e-01 -9.24236894e-01 -6.66663468e-01
6.98905110e-01 -1.24005926e+00 -1.54151902e-01 -1.80325675e+00
4.36127156e-01 -1.57521904e-01 -5.31541049e-01 1.09796464e+00
2.77198762e-01 7.82122970e-01 3.47839415e-01 1.59532979e-01
-1.40218508e+00 1.27475783e-01 9.12140012e-01 -4.21201587e-01
4.75620300e-01 -5.29938817e-01 -1.70702383e-01 8.43266547e-01
2.88054883e-01 -4.61398780e-01 -4.98007655e-01 -3.90649527e-01
4.72849339e-01 1.29654929e-01 8.56428921e-01 -1.12569654e+00
-9.10003707e-02 -9.63317901e-02 2.77556658e-01 -4.07819420e-01
6.34383321e-01 -8.35504353e-01 2.27265403e-01 2.57077605e-01
-8.77171636e-01 -3.00572310e-02 6.83824793e-02 9.54923034e-01
-2.49797180e-01 8.94154757e-02 6.48852706e-01 -3.40450376e-01
-1.57196677e+00 4.83361959e-01 -4.44460332e-01 2.02770010e-01
1.07952821e+00 -3.38604957e-01 -7.15617359e-01 -2.96188354e-01
-9.41875398e-01 3.47531796e-01 4.77485299e-01 7.50635505e-01
4.35940176e-01 -1.57414508e+00 -4.12609428e-01 -7.19324723e-02
6.15354061e-01 -6.90019608e-01 3.89589638e-01 8.30541611e-01
-2.19005927e-01 3.43248129e-01 -4.96642530e-01 -7.69508362e-01
-1.45857453e+00 7.83981979e-01 7.86292478e-02 -1.72527075e-01
-5.91471732e-01 9.75668907e-01 4.80779678e-01 -3.29972386e-01
3.42727095e-01 -3.84499252e-01 -3.75476778e-01 6.47873163e-01
3.69659334e-01 4.30870980e-01 -3.30311626e-01 -1.13524652e+00
-3.59081626e-01 6.36183798e-01 2.53497511e-01 3.26165110e-01
1.10979414e+00 -3.15434098e-01 -2.74921626e-01 8.64107370e-01
1.78488398e+00 -8.16074669e-01 -1.37545907e+00 1.29813245e-02
1.86524853e-01 -2.31675714e-01 1.68696210e-01 -1.02866638e+00
-9.83767390e-01 9.18347359e-01 6.28724694e-01 2.64904529e-01
1.01578641e+00 2.84523189e-01 6.17981374e-01 4.32148010e-01
5.49316645e-01 -1.03971958e+00 4.97380853e-01 3.07679057e-01
6.12124085e-01 -1.51438701e+00 -9.13258046e-02 -1.63062662e-01
-7.88232863e-01 9.91995931e-01 5.69624066e-01 2.51607478e-01
5.42370319e-01 -4.72151078e-02 1.38930053e-01 -7.40824640e-01
-9.99784112e-01 -3.47265452e-01 4.99450475e-01 4.63966072e-01
2.15018898e-01 1.15643576e-01 2.58271933e-01 -3.25636119e-02
7.17404962e-01 -3.06326002e-02 3.81889582e-01 8.48481119e-01
-3.78873467e-01 -7.30058253e-01 -1.03935348e-02 8.61609280e-01
-4.00879711e-01 -2.94394400e-02 -2.35869274e-01 6.70038700e-01
5.38420737e-01 1.04304790e+00 3.69073004e-02 -5.69835842e-01
-1.45211115e-01 3.80909115e-01 2.63231277e-01 -7.42834270e-01
-6.44134164e-01 -2.75991976e-01 5.76135933e-01 -1.29441345e+00
-1.06451690e+00 -7.01162875e-01 -9.87697423e-01 1.65690109e-01
1.39082640e-01 -3.26500297e-01 6.31535411e-01 9.71622765e-01
4.82941747e-01 7.14375019e-01 1.41591653e-01 -1.02285135e+00
1.87246986e-02 -9.46323156e-01 -7.92083859e-01 1.06807148e+00
2.94454277e-01 -1.12152314e+00 -3.54225606e-01 6.52497768e-01] | [8.595206260681152, 0.7977591753005981] |
96df2840-c256-47ca-ab50-5f544e3f04d9 | word-sense-induction-with-knowledge-1 | 2304.10642 | null | https://arxiv.org/abs/2304.10642v1 | https://arxiv.org/pdf/2304.10642v1.pdf | Word Sense Induction with Knowledge Distillation from BERT | Pre-trained contextual language models are ubiquitously employed for language understanding tasks, but are unsuitable for resource-constrained systems. Noncontextual word embeddings are an efficient alternative in these settings. Such methods typically use one vector to encode multiple different meanings of a word, and incur errors due to polysemy. This paper proposes a two-stage method to distill multiple word senses from a pre-trained language model (BERT) by using attention over the senses of a word in a context and transferring this sense information to fit multi-sense embeddings in a skip-gram-like framework. We demonstrate an effective approach to training the sense disambiguation mechanism in our model with a distribution over word senses extracted from the output layer embeddings of BERT. Experiments on the contextual word similarity and sense induction tasks show that this method is superior to or competitive with state-of-the-art multi-sense embeddings on multiple benchmark data sets, and experiments with an embedding-based topic model (ETM) demonstrates the benefits of using this multi-sense embedding in a downstream application. | ['Bulent Yener', 'Alex Gittens', 'Anik Saha'] | 2023-04-20 | word-sense-induction-with-knowledge | https://openreview.net/forum?id=-29uFS4FiDZ | https://openreview.net/pdf?id=-29uFS4FiDZ | null | ['word-sense-induction', 'word-similarity'] | ['natural-language-processing', 'natural-language-processing'] | [ 4.52241123e-01 -2.30244651e-01 -3.92251045e-01 -4.93492693e-01
-7.55240321e-01 -6.50261760e-01 7.85819948e-01 5.75630009e-01
-1.06682181e+00 3.36543083e-01 4.82427686e-01 -5.81266522e-01
1.41919434e-01 -7.95357943e-01 -1.75577074e-01 -5.78711092e-01
2.52277136e-01 4.16860193e-01 1.25132710e-01 -5.82844615e-01
4.26097721e-01 -1.14540011e-01 -1.49367750e+00 2.01735333e-01
8.79375577e-01 6.59036636e-01 7.37058103e-01 3.27232867e-01
-8.60243201e-01 -1.31215245e-01 -6.20608687e-01 -2.50927150e-01
-6.81060553e-02 -1.03432029e-01 -8.21084440e-01 -3.54842037e-01
3.18209887e-01 3.30235839e-01 1.09777553e-02 1.02177870e+00
3.92867506e-01 3.81528974e-01 5.70458412e-01 -8.81136477e-01
-1.16183233e+00 6.64617658e-01 -2.75801927e-01 4.18497533e-01
5.14526367e-01 -3.51466924e-01 1.72857535e+00 -1.24268937e+00
4.12522078e-01 1.53915286e+00 4.69098032e-01 6.45649552e-01
-1.36005831e+00 -5.57811916e-01 4.88058746e-01 2.63890654e-01
-1.39354789e+00 6.58743680e-02 6.52699411e-01 -2.44077593e-01
1.62859225e+00 7.51769245e-02 7.66961634e-01 1.38906372e+00
4.08542514e-01 7.75834918e-01 9.02639747e-01 -7.84512162e-01
3.77539635e-01 9.56609175e-02 5.63083172e-01 1.98344573e-01
4.94727790e-01 -1.79053992e-01 -5.88532150e-01 -5.00183940e-01
3.56835872e-01 2.64284909e-01 -8.50564986e-02 -3.15956086e-01
-1.19213581e+00 1.26833367e+00 3.64863217e-01 8.55113387e-01
-2.88453430e-01 -1.06510157e-02 4.42273587e-01 3.37701470e-01
7.81875014e-01 9.39906418e-01 -7.34477699e-01 -1.22782156e-01
-7.30021596e-01 2.20312878e-01 5.39806306e-01 7.53216743e-01
9.75152731e-01 -1.86752066e-01 -1.93517935e-02 1.22061789e+00
4.42823827e-01 4.14287031e-01 1.19562232e+00 -9.91650149e-02
1.47244141e-01 7.06598818e-01 -8.71478468e-02 -8.24862301e-01
-2.13666111e-01 -6.74967691e-02 -3.82659376e-01 -3.14251065e-01
-2.07435250e-01 -2.97600850e-02 -1.04692376e+00 2.00552368e+00
2.19814911e-01 6.64504051e-01 3.09174240e-01 5.94162762e-01
6.18390262e-01 6.84938669e-01 5.99964797e-01 1.67924404e-01
1.67563033e+00 -7.36796021e-01 -5.74834645e-01 -1.06909156e+00
8.68489206e-01 -7.85943508e-01 1.64428842e+00 9.70747545e-02
-4.49981272e-01 -5.31742692e-01 -1.24819386e+00 -2.62051791e-01
-1.14192939e+00 -1.20057918e-01 7.93597341e-01 6.44211650e-01
-9.87759471e-01 2.23724946e-01 -5.09907782e-01 -8.07469904e-01
-4.58137831e-03 3.84549201e-02 -3.32400322e-01 -3.25691640e-01
-1.74948418e+00 1.12814844e+00 7.71609306e-01 -3.88139606e-01
-3.94098222e-01 -8.61165702e-01 -1.43543053e+00 1.20733887e-01
2.04973131e-01 -7.32313573e-01 9.57520664e-01 -5.62144101e-01
-9.19111192e-01 9.81776237e-01 -5.33412099e-01 -2.51307875e-01
-6.63515747e-01 -5.44167161e-01 -6.91607475e-01 -1.34476662e-01
3.98759902e-01 5.95002174e-01 6.74495578e-01 -1.03210473e+00
-6.66001678e-01 -1.66110098e-01 1.76232457e-01 2.08553523e-01
-7.20249653e-01 3.22494358e-02 -8.60857889e-02 -7.99894452e-01
-2.11861115e-02 -7.73260653e-01 -5.03798246e-01 -2.55114943e-01
-7.70141184e-02 -8.26703012e-01 8.09518456e-01 -2.20739488e-02
1.53923702e+00 -2.15072489e+00 -1.26328811e-01 3.09820306e-02
7.18040904e-03 6.03257000e-01 -6.14113808e-01 6.86215401e-01
-2.74217457e-01 4.27094579e-01 -3.32085937e-01 -5.99055111e-01
2.33370394e-01 8.15373659e-01 -6.28186643e-01 -1.35824010e-01
5.19262433e-01 7.78795838e-01 -1.42946637e+00 -2.68523604e-01
3.28325570e-01 6.49736047e-01 -6.01917744e-01 3.25588614e-01
-2.80490994e-01 -2.47071788e-01 -5.75738549e-01 1.38893783e-01
3.39525968e-01 -2.21691161e-01 5.28716624e-01 6.64193509e-03
2.83599794e-01 8.95322919e-01 -1.11925280e+00 2.08442736e+00
-1.26788449e+00 4.84570950e-01 -3.25204551e-01 -9.20738459e-01
8.97807300e-01 4.01435196e-01 1.78570941e-01 -3.88711244e-01
-1.56053275e-01 2.03869194e-01 -1.12709917e-01 -4.23131526e-01
9.89049852e-01 -4.78729039e-01 -5.01915693e-01 5.72941363e-01
3.97439539e-01 -1.98828399e-01 1.79792017e-01 3.77393335e-01
1.01493359e+00 -2.72600234e-01 7.82246947e-01 -5.28437018e-01
3.16181064e-01 -2.28029698e-01 5.50373554e-01 6.60449326e-01
-2.70592105e-02 3.79346937e-01 7.63533935e-02 -3.47880214e-01
-6.64881408e-01 -1.22934234e+00 -1.95244655e-01 1.60386193e+00
2.25337416e-01 -1.06356537e+00 -2.32932433e-01 -5.86967170e-01
8.15694407e-02 1.07994425e+00 -7.01616228e-01 -2.23169416e-01
-4.56591308e-01 -6.41306520e-01 3.03003341e-01 7.88983643e-01
-4.75950651e-02 -9.93789256e-01 -4.62642431e-01 5.67791939e-01
4.21731099e-02 -1.11748672e+00 -5.28827965e-01 3.58173221e-01
-5.09469151e-01 -7.76427150e-01 -2.58016557e-01 -9.57662284e-01
3.01799804e-01 4.85617459e-01 1.48861551e+00 1.25021696e-01
-3.28452706e-01 3.91813070e-01 -5.99761307e-01 -5.03985286e-01
-1.26659065e-01 2.81048179e-01 3.67979079e-01 -1.76928446e-01
1.23659086e+00 -6.13094628e-01 -4.99580950e-01 -1.63198430e-02
-1.50659776e+00 -5.14900148e-01 3.57127100e-01 1.20225585e+00
5.15054345e-01 -4.77536410e-01 7.22077549e-01 -9.40681458e-01
1.14467740e+00 -6.70381904e-01 -6.95034042e-02 3.33370417e-01
-8.94910514e-01 3.15266937e-01 3.97673309e-01 -7.18369842e-01
-9.22379375e-01 -2.91620523e-01 -2.96831608e-01 -9.60514247e-02
-2.18787208e-01 7.01652050e-01 -1.80138424e-01 4.20577526e-01
5.71234763e-01 1.12719253e-01 -4.87518072e-01 -4.99045551e-01
1.06314695e+00 8.57084632e-01 1.27637371e-01 -8.00331116e-01
4.58323509e-01 1.65918365e-01 -5.78180313e-01 -1.04629493e+00
-1.17267013e+00 -1.12159169e+00 -6.69599652e-01 6.33201480e-01
1.08478665e+00 -8.39596450e-01 1.93303585e-01 -1.39867902e-01
-1.29779994e+00 1.16945684e-01 -1.72699526e-01 5.44516802e-01
-1.17116101e-01 3.63404036e-01 -2.50709116e-01 -5.74223578e-01
-3.09898168e-01 -9.20953214e-01 1.27622652e+00 2.96785116e-01
-6.94159031e-01 -1.73685324e+00 3.84583950e-01 -1.22298382e-01
7.31161296e-01 -2.04485670e-01 1.23095989e+00 -1.27116597e+00
9.18549523e-02 -6.27035052e-02 8.10108483e-02 2.19275326e-01
4.36087489e-01 -4.18905646e-01 -1.17307174e+00 -2.41554841e-01
-2.48904020e-01 -3.75792205e-01 1.20485306e+00 -1.02680754e-02
8.86808753e-01 -1.32425413e-01 -7.43551254e-01 1.74293712e-01
1.69951737e+00 -2.49702446e-02 2.73870409e-01 2.76996464e-01
6.18615925e-01 1.66467682e-01 6.64632499e-01 2.04239637e-01
4.08865243e-01 4.03126597e-01 9.77820382e-02 -4.64544483e-02
2.07184657e-01 -4.02857095e-01 2.05614612e-01 1.06283903e+00
6.62466049e-01 -3.56786102e-01 -9.20173585e-01 1.22532666e+00
-1.58811414e+00 -7.25904524e-01 3.08875173e-01 1.92929959e+00
1.16172397e+00 2.06366777e-01 -3.39875549e-01 -2.26964712e-01
5.50949514e-01 6.94364130e-01 -2.22586945e-01 -9.83161986e-01
-5.47368042e-02 8.45848203e-01 9.04885083e-02 6.33130789e-01
-9.87195909e-01 1.48975980e+00 6.28845882e+00 9.27764773e-01
-9.94861543e-01 2.79919833e-01 -2.36183349e-02 1.63298488e-01
-9.22333539e-01 2.48428166e-01 -7.26963818e-01 2.85471469e-01
9.58913088e-01 -2.77625710e-01 8.35232660e-02 8.83165181e-01
-2.06231281e-01 4.84537194e-03 -1.32915020e+00 1.08035100e+00
2.26774633e-01 -1.07608855e+00 4.09666568e-01 -2.12073997e-01
5.81115425e-01 3.12930197e-01 -9.48953032e-02 5.91385722e-01
4.24289316e-01 -1.02261233e+00 1.15323685e-01 1.57045349e-01
7.70410001e-01 -5.84467709e-01 7.11457372e-01 3.81732397e-02
-1.47547567e+00 1.14044048e-01 -8.33535314e-01 -2.19834074e-01
2.39292815e-01 6.76623762e-01 -9.40422416e-01 4.54113603e-01
2.88615137e-01 7.87838638e-01 -5.04008234e-01 4.97159272e-01
-6.05376065e-01 7.23898351e-01 -2.81617850e-01 -4.07067448e-01
5.12413085e-01 2.17020437e-02 5.59728801e-01 1.67006302e+00
2.43009269e-01 5.29535636e-02 4.59601492e-01 8.51134956e-01
-2.78240684e-02 1.94309399e-01 -8.43963742e-01 -3.65365595e-01
8.01440597e-01 1.26144969e+00 -2.35737309e-01 -6.83181703e-01
-6.13593400e-01 9.28676069e-01 3.32531810e-01 3.03002119e-01
-2.62606502e-01 -6.31808698e-01 1.62683678e+00 -3.69776368e-01
3.68663400e-01 -4.92466331e-01 -2.99962729e-01 -1.24291241e+00
3.46336290e-02 -4.96437728e-01 6.23215020e-01 -6.41724229e-01
-1.80777240e+00 6.97706461e-01 1.45976171e-01 -9.42893326e-01
-5.01985908e-01 -1.04486692e+00 -8.92722070e-01 1.09644389e+00
-1.81196463e+00 -1.10975420e+00 2.67533869e-01 2.19090402e-01
8.32279384e-01 -9.91810039e-02 1.49655020e+00 -9.55356192e-03
-1.36624634e-01 6.50274992e-01 -1.26482084e-01 1.38048371e-02
8.67443025e-01 -1.61201751e+00 8.42021823e-01 7.58005679e-01
7.05814600e-01 1.38938522e+00 6.22262359e-01 -4.15320694e-01
-1.20343840e+00 -1.01668286e+00 1.47542512e+00 -6.22953773e-01
1.08337092e+00 -5.88725269e-01 -1.18223035e+00 5.97712338e-01
6.71149552e-01 -5.78093007e-02 1.40476978e+00 5.82296789e-01
-8.19960237e-01 1.43041536e-01 -8.27032506e-01 7.54832745e-01
9.90993619e-01 -9.16221976e-01 -1.49953294e+00 3.12191192e-02
1.20564878e+00 1.01196609e-01 -7.13079691e-01 8.44297581e-04
3.93157423e-01 -3.52044910e-01 1.06286645e+00 -9.95874703e-01
2.37422153e-01 -2.26331815e-01 -5.37025392e-01 -1.99528778e+00
-2.51357973e-01 -5.29866099e-01 2.72846401e-01 1.24378383e+00
6.34325087e-01 -8.49200785e-01 1.29002705e-01 3.63689929e-01
-1.41449401e-03 -8.66265178e-01 -1.02300763e+00 -7.32287526e-01
6.00739896e-01 -8.01774919e-01 8.10491800e-01 1.22741473e+00
3.48403066e-01 8.66395056e-01 2.47580633e-01 1.83991641e-01
1.21556871e-01 1.92000046e-01 1.87004521e-01 -1.15212452e+00
-6.58863932e-02 -3.77950400e-01 -6.94353759e-01 -1.30967402e+00
6.95167899e-01 -1.01656139e+00 9.37903672e-02 -1.60903990e+00
8.87643844e-02 -4.93902534e-01 -9.54298913e-01 6.18142962e-01
-8.79050255e-01 3.46899778e-02 6.33822531e-02 -3.01090628e-01
-4.87502307e-01 6.17844045e-01 6.63887739e-01 -3.05444151e-01
9.31714699e-02 -5.74288547e-01 -9.45149124e-01 6.62954450e-01
7.30022967e-01 -6.86821938e-01 -7.05632448e-01 -6.69281423e-01
4.30646211e-01 -6.48664296e-01 1.49680912e-01 -5.51865578e-01
1.58635364e-03 -3.22579294e-01 1.56442914e-02 -2.31335223e-01
5.26383936e-01 -6.64246082e-01 -8.37141216e-01 -1.43311201e-02
-5.25607407e-01 4.49229658e-01 2.28046134e-01 6.40022755e-01
-3.37607056e-01 -3.43197346e-01 4.02814001e-01 -1.28884807e-01
-1.33255315e+00 -1.83192417e-02 -3.94314647e-01 5.53483605e-01
5.85045516e-01 -1.06490925e-01 -1.90595742e-02 -1.12317167e-01
-5.67363799e-01 2.14956656e-01 3.66232514e-01 1.15241110e+00
7.69340992e-01 -1.53568161e+00 -4.98200297e-01 4.34334785e-01
7.99142003e-01 -2.39098445e-01 -2.39718303e-01 -9.99945775e-03
3.61355752e-01 4.11358535e-01 2.04214081e-01 -6.64550662e-01
-1.17626321e+00 6.49096310e-01 -7.80210644e-02 -3.23709428e-01
-3.42346132e-01 1.10456216e+00 3.66137505e-01 -6.61726773e-01
-1.88372478e-01 -8.04529965e-01 -1.72317788e-01 7.29319081e-02
8.14975619e-01 -2.64544964e-01 1.07554726e-01 -5.19334197e-01
-6.49191856e-01 6.67172253e-01 -1.35437131e-01 -2.33197108e-01
1.19722033e+00 -2.41386488e-01 3.40757780e-02 8.02958250e-01
1.43526125e+00 -6.97662756e-02 -6.35814905e-01 -6.36383772e-01
3.77633482e-01 -4.21852112e-01 -1.46083841e-02 -5.91370225e-01
-4.45795029e-01 1.03297627e+00 4.22772408e-01 2.95098662e-01
8.70114744e-01 2.93206602e-01 1.11323202e+00 5.08252203e-01
3.46323222e-01 -1.12898529e+00 9.41624865e-02 9.14309502e-01
5.78931153e-01 -1.15582919e+00 -3.65697175e-01 -1.96011737e-01
-6.37511492e-01 1.04779196e+00 4.57444161e-01 -2.30203807e-01
8.43527794e-01 1.37450978e-01 5.06080329e-01 -1.38576075e-01
-1.03421724e+00 -4.99331892e-01 3.38873595e-01 6.94095850e-01
7.41922379e-01 2.43032396e-01 -6.53195798e-01 6.64874613e-01
-2.47204751e-01 -5.23046672e-01 3.40863407e-01 1.06218779e+00
-5.87087035e-01 -1.57933426e+00 -3.06527819e-02 2.40263000e-01
-4.05192852e-01 -5.82259119e-01 -3.88539106e-01 5.90818107e-01
2.57338315e-01 1.26378524e+00 1.73199728e-01 -3.56552958e-01
8.07590187e-02 5.43494523e-01 7.52502605e-02 -1.35661411e+00
-4.09845650e-01 -2.51840025e-01 7.08778650e-02 -5.99082351e-01
-4.17715579e-01 -3.36082608e-01 -1.23540103e+00 4.44355905e-01
-3.20494831e-01 3.18885446e-01 6.00416243e-01 1.30072165e+00
5.17626882e-01 3.44330132e-01 4.30100828e-01 -4.70075518e-01
-6.17000103e-01 -1.16088700e+00 -5.00095427e-01 7.93497682e-01
3.21022540e-01 -8.36345255e-01 -2.30500102e-01 -1.50606319e-01] | [10.401629447937012, 8.883500099182129] |
6ed41c8a-a4d9-4ec9-bae6-eb622795a70c | learning-online-data-association-1 | 2011.03183 | null | https://arxiv.org/abs/2011.03183v4 | https://arxiv.org/pdf/2011.03183v4.pdf | Learning Object-Based State Estimators for Household Robots | A robot operating in a household makes observations of multiple objects as it moves around over the course of days or weeks. The objects may be moved by inhabitants, but not completely at random. The robot may be called upon later to retrieve objects and will need a long-term object-based memory in order to know how to find them. Existing work in semantic slam does not attempt to capture the dynamics of object movement. In this paper, we combine some aspects of classic techniques for data-association filtering with modern attention-based neural networks to construct object-based memory systems that operate on high-dimensional observations and hypotheses. We perform end-to-end learning on labeled observation trajectories to learn both the transition and observation models. We demonstrate the system's effectiveness in maintaining memory of dynamically changing objects in both simulated environment and real images, and demonstrate improvements over classical structured approaches as well as unstructured neural approaches. Additional information available at project website: https://yilundu.github.io/obm/. | ['Tomas Lozano-Perez', 'Leslie Kaelbling', 'Yilun Du'] | 2020-11-06 | learning-online-data-association | https://openreview.net/forum?id=KjR-3lBYB3y | https://openreview.net/pdf?id=KjR-3lBYB3y | null | ['semantic-slam'] | ['computer-vision'] | [ 6.19669817e-02 9.11440849e-02 6.33975651e-05 -3.52126867e-01
-3.37044775e-01 -1.35404736e-01 4.26339686e-01 3.58689249e-01
-4.74569470e-01 6.71781421e-01 1.19451635e-01 9.39273834e-02
-2.39930704e-01 -6.93119049e-01 -1.01491153e+00 -4.75721627e-01
-6.45350993e-01 1.21178806e+00 3.92690629e-01 -9.31706131e-02
2.33846083e-01 5.28358102e-01 -1.73484254e+00 4.71213125e-02
5.44197917e-01 6.93511665e-01 1.22546291e+00 8.99797559e-01
-8.59680325e-02 9.10271943e-01 -1.44620582e-01 3.63495380e-01
9.88126174e-02 -1.24648921e-01 -1.15128410e+00 2.34585688e-01
1.85724333e-01 -1.86071649e-01 -6.18480980e-01 9.69431460e-01
2.19561279e-01 7.76747942e-01 5.02495229e-01 -1.20203876e+00
-9.28987920e-01 3.80892575e-01 -1.08533397e-01 3.41733932e-01
5.21399558e-01 1.79022864e-01 7.15328515e-01 -1.05793691e+00
6.60044432e-01 1.42080748e+00 5.61691880e-01 3.85935485e-01
-9.96439278e-01 -3.18330705e-01 5.79831123e-01 7.20842719e-01
-1.39813364e+00 -6.86520696e-01 3.15368325e-01 -3.37084353e-01
1.44803095e+00 -1.77849129e-01 8.02877486e-01 1.00325894e+00
2.20906764e-01 1.05835474e+00 4.32523429e-01 -5.45649230e-01
3.27728927e-01 -1.65038288e-01 3.28021944e-01 9.99363005e-01
3.00100893e-01 1.63527593e-01 -8.54799390e-01 -1.49329349e-01
7.02304959e-01 4.07258481e-01 -2.11105108e-01 -8.37716341e-01
-1.33031774e+00 7.00161397e-01 8.16417992e-01 2.01955944e-01
-7.02262163e-01 4.52524483e-01 -6.52323589e-02 1.81287393e-01
9.33427066e-02 3.34065497e-01 -6.37725174e-01 4.58947290e-03
-5.04186749e-01 3.02225173e-01 8.18391442e-01 1.28634465e+00
1.06811655e+00 -3.07781816e-01 2.69390404e-01 5.72469175e-01
4.21217620e-01 5.86554170e-01 6.95703387e-01 -1.06958938e+00
1.77454069e-01 3.57514173e-01 6.93504035e-01 -7.63295710e-01
-7.19671130e-01 1.54587165e-01 -5.59728682e-01 1.13581195e-01
6.31866604e-02 1.88568771e-01 -1.37077713e+00 1.61162996e+00
3.73060048e-01 4.76218998e-01 1.83902472e-01 9.42104042e-01
5.17731905e-01 5.69301307e-01 9.99721214e-02 1.61382444e-02
1.02358031e+00 -1.39105749e+00 -7.16923296e-01 -7.40144193e-01
5.63010216e-01 -2.98719287e-01 6.70533836e-01 1.44582167e-01
-1.14156079e+00 -5.96966386e-01 -8.28022540e-01 -1.30817726e-01
-6.00354612e-01 -8.79019201e-02 7.04892993e-01 -1.56952962e-01
-1.17984152e+00 7.31522024e-01 -1.68164229e+00 -1.08669150e+00
3.57261360e-01 5.51694334e-01 -3.71393561e-01 -2.19914272e-01
-8.24796319e-01 1.22127616e+00 6.65933907e-01 2.49445572e-01
-1.29225945e+00 -1.49564490e-01 -1.05695224e+00 -5.90943359e-02
3.61327708e-01 -8.80473077e-01 1.66994989e+00 -6.51820719e-01
-1.22167504e+00 5.14707446e-01 -5.14327168e-01 -6.85746133e-01
1.25928655e-01 -7.10193217e-01 -1.84930071e-01 7.50503466e-02
3.48870397e-01 9.36139822e-01 5.80397487e-01 -1.17150056e+00
-1.07896161e+00 -5.94307184e-01 -6.66275769e-02 6.65015757e-01
4.42564450e-02 -3.46169502e-01 -5.11131465e-01 -7.13595673e-02
4.86589372e-01 -1.31900489e+00 -4.27438200e-01 1.33645497e-02
-9.91525277e-02 -1.14540949e-01 7.26537764e-01 -3.40112984e-01
5.23132205e-01 -1.90855801e+00 3.03901821e-01 -1.07940167e-01
-1.92236245e-01 -8.65545794e-02 -2.17904121e-01 5.41462779e-01
2.68802345e-01 -4.37492520e-01 -1.14822917e-01 -6.59531653e-01
-3.82169709e-02 4.70689893e-01 -1.66106328e-01 6.06151879e-01
-5.28457761e-02 8.97801161e-01 -1.34653008e+00 -2.14390650e-01
4.15244848e-01 4.13918465e-01 -2.30068773e-01 1.74584240e-01
-3.76310617e-01 5.25780678e-01 -5.64470589e-01 8.16694498e-01
2.56400406e-01 -5.31353414e-01 2.49708623e-01 5.08899808e-01
-2.40480267e-02 3.31102431e-01 -1.07185185e+00 2.29980755e+00
-4.11803246e-01 6.93329871e-01 2.40039602e-01 -9.62343514e-01
5.47082245e-01 2.32837901e-01 4.17906821e-01 -5.55310786e-01
-1.28359884e-01 1.57108366e-01 -2.52730012e-01 -5.97380877e-01
8.35225165e-01 8.99343658e-03 -4.75226045e-02 1.82708234e-01
1.78199291e-01 8.20416212e-02 2.48239376e-02 1.04725376e-01
1.42443907e+00 1.47772938e-01 2.82623500e-01 -1.75874174e-01
1.11741405e-02 5.16002595e-01 3.98235679e-01 1.24008536e+00
-3.20686519e-01 4.54883128e-01 -6.39069974e-01 -7.96339631e-01
-8.60428154e-01 -1.14509714e+00 8.06301162e-02 1.45634866e+00
7.24075079e-01 1.26212444e-02 -2.49726057e-01 -2.65905350e-01
2.23537117e-01 8.70827615e-01 -6.57037020e-01 -2.14669302e-01
-5.18871367e-01 -4.79443550e-01 1.01848610e-01 6.66358531e-01
3.69267106e-01 -1.85398531e+00 -1.20836854e+00 4.27226365e-01
-2.26725981e-01 -6.47712469e-01 -4.69469428e-02 8.24487746e-01
-8.69354486e-01 -1.05177939e+00 -3.65303487e-01 -9.85536873e-01
7.07788050e-01 6.78087831e-01 1.07893705e+00 1.15368120e-01
-4.29496974e-01 7.67526031e-01 -3.67836118e-01 -5.79088926e-01
-1.38556557e-02 2.09916845e-01 4.85049248e-01 -4.95909333e-01
5.66781580e-01 -5.79328239e-01 -4.58613843e-01 1.35501862e-01
-4.57319438e-01 -1.71453729e-01 6.56029761e-01 7.01134145e-01
5.86122394e-01 1.90464020e-01 3.47700298e-01 -5.23696959e-01
1.80232495e-01 -8.42175782e-01 -5.07508576e-01 2.23297894e-01
-4.21272933e-01 7.09597692e-02 8.08448121e-02 -6.22669220e-01
-9.32791114e-01 5.10103345e-01 3.65110248e-01 -6.13091886e-01
-5.50131738e-01 4.98924464e-01 1.00003190e-01 2.69469112e-01
4.14840430e-01 2.69221485e-01 -2.07924068e-01 -5.62983692e-01
3.34112227e-01 2.67511427e-01 6.78179145e-01 -4.17654693e-01
4.33582127e-01 7.66985595e-01 -4.30475205e-01 -8.14851522e-01
-7.38479197e-01 -7.92321265e-01 -1.01643646e+00 4.61834297e-02
6.73991323e-01 -1.05382931e+00 -6.69763565e-01 4.91283000e-01
-1.14183700e+00 -9.68640685e-01 -3.65477771e-01 5.53664088e-01
-8.30488443e-01 -1.15995497e-01 -4.96570140e-01 -7.91589618e-01
1.27803400e-01 -8.00394356e-01 1.12519145e+00 2.84111619e-01
-3.67914796e-01 -9.34129119e-01 2.74875164e-01 5.83820343e-02
1.73678771e-01 -2.37039492e-01 4.99524385e-01 -5.80875039e-01
-9.70989823e-01 -9.12664607e-02 1.41461551e-01 -3.44704032e-01
1.50159538e-01 -6.02707863e-01 -5.95682681e-01 -6.51787400e-01
-1.72134787e-01 -3.65552783e-01 1.06667566e+00 5.76966941e-01
6.59943819e-01 -1.89971477e-01 -1.03452408e+00 3.19118619e-01
1.33737016e+00 2.87554950e-01 3.84861797e-01 6.84738278e-01
6.18207216e-01 5.26465356e-01 8.59671354e-01 3.95426124e-01
6.62159622e-01 5.54387927e-01 7.72060812e-01 3.41180295e-01
5.29544279e-02 -1.90992370e-01 3.03895026e-01 4.53335822e-01
1.51225939e-01 -4.91050899e-01 -1.14530671e+00 1.27154791e+00
-2.40771317e+00 -9.28393602e-01 1.07242633e-02 2.02163625e+00
9.82268378e-02 -2.25053988e-02 -2.21973017e-01 -4.84498292e-01
7.20114470e-01 2.71434104e-03 -9.05697584e-01 -5.13722636e-02
2.17343241e-01 -2.43660748e-01 7.05662668e-01 7.67154753e-01
-1.22279215e+00 1.15801442e+00 6.17247963e+00 7.25152194e-02
-7.65725195e-01 3.25187624e-01 -2.13876188e-01 -4.41159040e-01
1.80965498e-01 8.78450125e-02 -8.91025960e-01 1.39508680e-01
1.01014674e+00 1.30389869e-01 5.39114714e-01 1.05685616e+00
1.44845024e-01 -5.61089158e-01 -1.20234323e+00 7.99850762e-01
7.26017654e-02 -1.10941553e+00 -3.12546909e-01 -2.31448561e-04
7.03527331e-01 6.95587277e-01 -1.14635006e-01 3.81559432e-01
8.98805857e-01 -9.57790256e-01 1.06705832e+00 8.17575812e-01
-5.40689491e-02 -4.78933781e-01 5.02609789e-01 7.86383271e-01
-1.17603087e+00 -5.33721685e-01 -6.59025908e-01 -4.45558697e-01
3.63337964e-01 -3.09735909e-03 -1.16684651e+00 3.34550411e-01
1.15536177e+00 8.00839126e-01 -3.80096734e-01 1.17904592e+00
-1.77253470e-01 1.62619516e-01 -6.14717722e-01 -1.56274885e-01
4.51090276e-01 1.64359286e-02 6.59350157e-01 9.00665641e-01
4.87716228e-01 2.54318953e-01 6.94791794e-01 5.56953967e-01
1.82637796e-01 -3.63905251e-01 -9.07133102e-01 2.71980464e-01
8.42477679e-01 8.41985404e-01 -9.41296935e-01 -4.78466392e-01
-2.95398295e-01 1.31085837e+00 8.21368158e-01 4.18011367e-01
-5.25921583e-01 -1.05517067e-01 6.73415244e-01 2.86616683e-02
7.87908196e-01 -5.29007673e-01 2.14389905e-01 -8.86090279e-01
-5.03148837e-03 -2.95016468e-01 4.23881352e-01 -1.09142923e+00
-8.80926669e-01 3.99467617e-01 -5.91338724e-02 -8.43079209e-01
-1.44733846e-01 -3.26686382e-01 -3.88070703e-01 5.64716041e-01
-1.39578927e+00 -1.14824796e+00 -4.51592177e-01 5.02042890e-01
9.18733835e-01 6.70938194e-02 1.01796973e+00 -2.06101798e-02
-2.38170132e-01 -3.01309139e-01 4.04152602e-01 -1.81625366e-01
4.19209123e-01 -1.18366802e+00 3.87379646e-01 7.50872850e-01
3.75976622e-01 7.02836931e-01 9.03161883e-01 -9.97399449e-01
-1.42628682e+00 -1.26730859e+00 8.87050748e-01 -7.97962308e-01
4.84774888e-01 -3.43605191e-01 -1.01992857e+00 1.48478210e+00
6.46743849e-02 1.47286147e-01 1.32635921e-01 2.81448960e-01
1.53176576e-01 4.16717082e-01 -8.37349653e-01 3.47729862e-01
1.37346244e+00 -3.39488089e-01 -8.96248698e-01 5.28026879e-01
7.22818375e-01 -4.36536938e-01 -3.31524998e-01 3.33802849e-01
4.54086512e-01 -8.13232541e-01 1.06274080e+00 -7.38961339e-01
-2.04391003e-01 -4.58459646e-01 -3.93067330e-01 -1.28000581e+00
-8.49066496e-01 -3.72084111e-01 -4.38241750e-01 6.65957510e-01
3.55662704e-01 -5.37892640e-01 8.88303638e-01 5.79334319e-01
-3.11042279e-01 -3.70542258e-01 -8.32417548e-01 -9.63701963e-01
-3.80502254e-01 -3.67407441e-01 4.89075661e-01 6.30300283e-01
-5.83963133e-02 3.45222592e-01 -3.78565371e-01 9.42728162e-01
5.88872671e-01 3.52392375e-01 7.66264081e-01 -1.22827971e+00
1.64093859e-02 -3.11117042e-02 -4.76553291e-01 -1.28919971e+00
2.94131994e-01 -7.33495474e-01 7.74239600e-01 -2.01601148e+00
1.80266201e-01 -6.21780276e-01 -4.82976019e-01 5.93503296e-01
7.88003281e-02 -2.33746588e-01 5.14148660e-02 5.43770254e-01
-1.34670556e+00 6.78764939e-01 7.48604119e-01 -2.01480880e-01
-5.79766393e-01 -1.18041979e-02 -1.94735318e-01 8.69069636e-01
8.27336848e-01 -6.93887591e-01 -4.39267844e-01 -8.42958212e-01
-6.77418858e-02 9.28972438e-02 6.39013350e-01 -1.29936659e+00
7.51399100e-01 -1.77669987e-01 4.57644761e-01 -9.11580026e-01
8.46790910e-01 -1.01094913e+00 3.34726959e-01 7.31632173e-01
-3.36603522e-01 2.07701042e-01 8.73256028e-02 1.11921060e+00
7.03705475e-02 -2.41157532e-01 4.59027439e-01 -7.34160662e-01
-1.44445109e+00 3.91923845e-01 -6.62617803e-01 -4.63800907e-01
9.85751152e-01 -2.16819584e-01 -4.47657555e-02 -3.86254311e-01
-1.28606594e+00 7.22133756e-01 8.00142884e-01 7.11865783e-01
6.75484240e-01 -1.11897969e+00 -2.19179615e-01 -5.48549145e-02
1.87086046e-01 3.66765350e-01 2.79534608e-01 6.91684663e-01
-5.08410096e-01 5.50438643e-01 -1.70591325e-01 -7.05516696e-01
-1.01797593e+00 9.15511966e-01 1.83605880e-01 -1.56807881e-02
-9.19869184e-01 1.04368174e+00 1.73909590e-01 -7.61380315e-01
4.81097370e-01 -4.04545367e-01 -5.77340499e-02 -9.90234539e-02
4.81268197e-01 3.49436402e-01 -1.37676641e-01 -6.36519849e-01
-5.63213229e-01 2.76125968e-01 -1.22830411e-02 -1.09243684e-01
1.58546495e+00 -6.77128255e-01 4.17847969e-02 9.70899999e-01
9.12005186e-01 -7.26727247e-01 -1.36213565e+00 -5.12095094e-01
2.92693943e-01 -4.37787712e-01 -9.71463621e-02 -6.88877165e-01
-4.98125732e-01 5.77483237e-01 6.76731110e-01 1.70846015e-01
7.06957579e-01 4.15750206e-01 7.66966581e-01 1.18670738e+00
1.02007234e+00 -1.09599161e+00 1.80883333e-01 7.66577482e-01
7.69530356e-01 -1.39064932e+00 -1.12625800e-01 4.68808487e-02
-4.86437052e-01 7.85761118e-01 6.04476094e-01 -2.89593041e-01
6.66270196e-01 6.50178045e-02 1.83515076e-03 -5.35268188e-01
-9.28484857e-01 -5.17227530e-01 -2.93705642e-01 8.41563284e-01
-3.17452073e-01 1.83231771e-01 6.88012481e-01 -1.78548824e-02
-7.36420676e-02 -4.74568643e-02 3.98331493e-01 1.49064362e+00
-1.11763275e+00 -5.87618589e-01 -5.06640017e-01 2.89514661e-01
9.20870304e-02 9.11651403e-02 -1.03951901e-01 7.64556885e-01
5.73008657e-02 8.71716857e-01 3.85363638e-01 -2.92389467e-02
2.15372086e-01 2.75230914e-01 5.97178042e-01 -9.33171213e-01
4.04341072e-02 -2.44108170e-01 2.18330417e-02 -7.03964353e-01
-6.12770021e-01 -1.13029051e+00 -1.75898504e+00 -8.09630454e-02
-3.69048774e-01 1.46136269e-01 6.57279670e-01 9.00076628e-01
5.23581922e-01 4.54608142e-01 5.09934984e-02 -1.43354595e+00
-1.46646827e-01 -1.13658440e+00 -5.19056320e-01 2.84792751e-01
7.41995215e-01 -1.01548505e+00 -7.19428621e-03 2.78580099e-01] | [4.710126876831055, 0.6224068403244019] |
1cfc24b9-fc7e-444d-97d2-27ddd17853e9 | improving-native-language-identification-by | null | null | https://aclanthology.org/P17-2086 | https://aclanthology.org/P17-2086.pdf | Improving Native Language Identification by Using Spelling Errors | In this paper, we explore spelling errors as a source of information for detecting the native language of a writer, a previously under-explored area. We note that character n-grams from misspelled words are very indicative of the native language of the author. In combination with other lexical features, spelling error features lead to 1.2{\%} improvement in accuracy on classifying texts in the TOEFL11 corpus by the author{'}s native language, compared to systems participating in the NLI shared task. | ['Vivi Nastase', 'Lingzhen Chen', 'Carlo Strapparava'] | 2017-07-01 | null | null | null | acl-2017-7 | ['native-language-identification'] | ['natural-language-processing'] | [ 3.15727711e-01 -2.45442003e-01 -3.30487698e-01 -1.84153125e-01
-1.02421367e+00 -1.08949852e+00 9.53656852e-01 3.51404637e-01
-7.73741722e-01 7.18145490e-01 3.52699429e-01 -7.05024302e-01
1.67770520e-01 -3.93542022e-01 -2.45992377e-01 -1.90323114e-01
6.82368219e-01 5.19847333e-01 2.25508655e-03 -2.02799752e-01
1.16901433e+00 7.79657245e-01 -1.00681293e+00 5.42627692e-01
7.85122275e-01 4.01514560e-01 2.02802420e-01 1.11375713e+00
-6.87553108e-01 7.64830589e-01 -1.07217038e+00 -4.62477028e-01
-4.54450725e-03 -5.59006453e-01 -1.07682300e+00 1.67883374e-02
8.49792361e-01 -8.81722663e-03 -2.79017121e-01 1.16251397e+00
1.96324319e-01 -1.47659719e-01 8.92072201e-01 -5.40131092e-01
-5.60858369e-01 7.41646528e-01 -3.61409128e-01 9.48852658e-01
1.01354611e+00 -1.27257645e-01 1.22022426e+00 -1.12013531e+00
8.74297798e-01 1.08665264e+00 6.52513146e-01 2.89303333e-01
-8.63209903e-01 -8.60731423e-01 -1.41049042e-01 -2.50277072e-01
-1.53440320e+00 -7.08314657e-01 5.13345540e-01 -6.48773551e-01
1.13046169e+00 3.36974502e-01 4.45582755e-02 1.13988793e+00
6.95942521e-01 5.88890195e-01 1.30205774e+00 -1.09224677e+00
-1.01639122e-01 4.57306594e-01 5.72382092e-01 5.88099122e-01
3.80592585e-01 -1.56818837e-01 -8.62453282e-01 -3.48805308e-01
5.20822406e-01 -3.63557637e-01 -1.87390968e-01 5.50967813e-01
-1.24262476e+00 8.06514680e-01 -5.23047447e-01 9.73782480e-01
6.56656250e-02 -8.71856138e-02 4.44841951e-01 5.75233042e-01
4.24872369e-01 1.03587830e+00 -4.94075209e-01 -6.81463718e-01
-1.36920702e+00 4.06111151e-01 1.21396637e+00 1.10801661e+00
3.37352961e-01 -1.18190922e-01 -3.33876014e-02 6.71944022e-01
2.11700961e-01 6.49982393e-01 9.04392719e-01 -5.65696716e-01
7.61005998e-01 5.50957620e-01 4.22232747e-01 -9.25242662e-01
-1.89667076e-01 -4.86863106e-01 -1.60262883e-01 -1.47339359e-01
8.25102806e-01 -9.31670964e-02 -6.48439169e-01 1.14493477e+00
-5.32106102e-01 -5.66013515e-01 1.31034821e-01 4.02047634e-01
6.31997764e-01 6.89586282e-01 -2.58078068e-01 -3.91880393e-01
1.33748472e+00 -8.16186726e-01 -8.77476633e-01 -4.83051121e-01
8.33991289e-01 -1.43059063e+00 9.91166174e-01 6.85961008e-01
-1.02927637e+00 -4.64738190e-01 -8.37513089e-01 1.41261488e-01
-4.38976645e-01 2.86178976e-01 2.18962282e-01 1.05493665e+00
-7.70745039e-01 6.54434741e-01 -3.13380361e-01 -5.62650979e-01
-9.69499499e-02 -1.58809870e-01 -1.46033242e-01 -1.96453445e-02
-1.02028692e+00 8.80992711e-01 2.38824844e-01 -4.11301792e-01
-3.17339748e-01 -4.63567138e-01 -4.69824135e-01 -1.26635909e-01
1.06815025e-01 3.74232620e-01 1.54763806e+00 -8.29861999e-01
-1.15680695e+00 1.24678028e+00 -5.63548744e-01 -2.06488803e-01
6.25610530e-01 -1.60939038e-01 -1.19634593e+00 2.56642681e-02
4.16593552e-01 -3.18569094e-01 8.76628399e-01 -5.50736368e-01
-9.68517601e-01 -2.74200678e-01 -5.59498370e-01 -1.52976394e-01
-2.73118526e-01 8.34540486e-01 -4.23612632e-02 -1.09378541e+00
-1.06831379e-02 -7.33805597e-01 2.37315789e-01 -6.16759121e-01
-3.90037596e-01 -5.74412405e-01 3.40444088e-01 -1.12349582e+00
1.85497010e+00 -1.88218331e+00 -3.51224214e-01 5.74457526e-01
1.39734983e-01 3.91605198e-01 9.99404937e-02 8.51910651e-01
1.85947970e-01 7.04534829e-01 2.62317568e-01 9.06021893e-02
5.23697734e-02 -1.42928645e-01 -4.54104751e-01 4.93622720e-01
-2.16060817e-01 6.53367043e-01 -9.36415792e-01 -4.11010504e-01
-3.16529810e-01 -1.01905540e-01 1.54994890e-01 -7.29139149e-02
1.47693247e-01 -7.82597139e-02 -2.88400590e-01 4.83112901e-01
1.91735193e-01 -6.24758415e-02 2.03823954e-01 5.37092924e-01
-5.02650678e-01 9.89020050e-01 -1.10562396e+00 1.27678573e+00
-4.67647552e-01 1.00916636e+00 -1.59883305e-01 -1.61973804e-01
1.10975873e+00 4.60564017e-01 -4.28365529e-01 -4.98600274e-01
8.52706358e-02 7.72134840e-01 2.80312181e-01 -1.46283239e-01
7.50597060e-01 3.25748436e-02 -3.99843216e-01 6.55009687e-01
-1.63579598e-01 1.37696698e-01 5.04776955e-01 2.10681811e-01
1.15534580e+00 -2.50973880e-01 6.39127195e-01 -8.73512268e-01
6.84261620e-01 -3.16808820e-02 3.34792256e-01 1.32537866e+00
-2.21683592e-01 4.01610702e-01 3.47958654e-01 -2.50030488e-01
-1.07673502e+00 -7.44074166e-01 -2.70265132e-01 1.13995039e+00
-4.51489747e-01 -6.86152935e-01 -6.20364606e-01 -6.92308724e-01
-3.72279510e-02 1.08612370e+00 -2.63618410e-01 1.85363367e-01
-7.21097052e-01 -1.32120028e-01 1.14827621e+00 5.42501628e-01
-1.09378342e-02 -8.69009316e-01 -3.04768503e-01 3.38222593e-01
-1.08702756e-01 -1.06547701e+00 -7.93083608e-01 3.04523110e-01
-5.74759722e-01 -8.81461561e-01 -8.64039600e-01 -9.88047302e-01
4.80374426e-01 -7.70760551e-02 9.74512577e-01 1.29965737e-01
-2.18654662e-01 1.04887344e-01 -6.60875499e-01 -3.55388433e-01
-9.47488010e-01 5.39283872e-01 -1.10807270e-02 -5.09834707e-01
9.19192076e-01 1.83232978e-01 7.37237781e-02 1.99370176e-01
-4.41757619e-01 -4.67554569e-01 3.62482429e-01 5.22621155e-01
-1.86854154e-02 4.16701660e-02 3.65483642e-01 -1.19552696e+00
9.46111917e-01 -1.08143158e-01 -3.47399354e-01 2.83684731e-01
-1.01004303e+00 1.09261386e-01 8.57919872e-01 -3.46513808e-01
-1.05481434e+00 -2.61597961e-01 -3.20966065e-01 5.17920017e-01
-5.44662416e-01 4.79721904e-01 2.21336056e-02 -2.41609782e-01
7.35632956e-01 4.56085652e-01 -3.11885655e-01 -8.11448693e-01
-2.87760466e-01 1.04477537e+00 6.33543789e-01 -3.73210639e-01
8.50264132e-01 -2.81050593e-01 -4.29388463e-01 -1.07937181e+00
-7.81288028e-01 -9.68248129e-01 -7.00347364e-01 1.21461615e-01
3.87724996e-01 -5.41534305e-01 -3.07049185e-01 5.04018307e-01
-1.40265167e+00 2.14934781e-01 1.23671390e-01 4.40744489e-01
-1.84164032e-01 7.15837657e-01 -7.41537154e-01 -7.93586373e-01
-1.72942668e-01 -9.41510439e-01 6.53213918e-01 1.26080364e-01
-1.07373667e+00 -1.07784343e+00 1.25531107e-01 4.79080051e-01
-6.20904081e-02 -2.49911666e-01 7.80510664e-01 -1.45852244e+00
2.27057412e-01 -6.85514510e-01 -1.15428613e-02 2.90865332e-01
-8.95296931e-02 6.37852063e-04 -6.02948189e-01 -2.96344042e-01
-3.80873680e-02 5.46498634e-02 4.78471726e-01 -2.89565861e-01
6.85340881e-01 -5.93653202e-01 -2.71020472e-01 1.98848903e-01
1.21495032e+00 4.30251032e-01 3.97994637e-01 5.30267537e-01
4.14119780e-01 3.00216109e-01 3.93756837e-01 5.05420148e-01
-2.10231230e-01 2.99760312e-01 -4.61801022e-01 4.99264181e-01
2.34983131e-01 -3.55006725e-01 5.71596086e-01 8.42219412e-01
7.73451999e-02 -7.16941833e-01 -1.38962734e+00 5.09581625e-01
-1.16401315e+00 -9.93197918e-01 -6.22797370e-01 2.02451730e+00
8.56101990e-01 5.06273806e-01 8.46164823e-02 4.37261760e-01
8.52927327e-01 2.77329177e-01 8.03851113e-02 -1.03688836e+00
-1.45349488e-01 3.79926443e-01 9.99588966e-01 8.41818094e-01
-7.64903605e-01 1.15737951e+00 7.65060282e+00 1.01527858e+00
-8.29209626e-01 -1.97266906e-01 2.42134333e-01 1.47383675e-01
-3.05320740e-01 -7.00835064e-02 -1.50896776e+00 7.11800635e-01
1.39566398e+00 -5.68194866e-01 3.10574144e-01 6.07025504e-01
2.59581637e-02 -7.80432597e-02 -1.11492026e+00 7.86592841e-01
6.00375950e-01 -1.05880249e+00 -9.82114598e-02 2.26172805e-01
6.39637411e-01 -2.04226360e-01 -2.11084887e-01 7.56525397e-02
2.47744024e-01 -9.50319409e-01 9.40239549e-01 3.52262974e-01
7.26970315e-01 -9.09937799e-01 8.36348772e-01 7.13337302e-01
-7.46302843e-01 4.10597831e-01 -2.70263672e-01 -3.66548628e-01
-9.24283043e-02 5.11987448e-01 -1.16454542e+00 4.31844704e-02
4.16861206e-01 3.03687483e-01 -8.58618915e-01 5.99767148e-01
-3.03333968e-01 1.08185601e+00 -4.49316800e-02 -5.17918825e-01
4.28332090e-01 1.84422448e-01 7.59889305e-01 1.82228088e+00
2.19392732e-01 -4.44287397e-02 1.77638665e-01 5.70218265e-01
-1.71671614e-01 5.06443560e-01 -4.92445439e-01 -9.27371085e-01
5.48793018e-01 9.43469107e-01 -8.68859231e-01 -4.73765761e-01
-3.83318961e-01 1.45569623e+00 1.54604927e-01 1.98147655e-01
-1.34952798e-01 -1.24318635e+00 3.90962303e-01 2.32047886e-01
2.18404859e-01 -3.58062565e-01 -4.02456939e-01 -9.76542950e-01
5.77281183e-03 -1.10347652e+00 1.54049888e-01 -3.46664011e-01
-1.33274639e+00 5.64836800e-01 -5.27971864e-01 -7.42883146e-01
-5.15363097e-01 -1.21553564e+00 -5.31998038e-01 1.37003314e+00
-1.04407835e+00 -4.75660801e-01 1.52275547e-01 -6.42252564e-02
7.37490237e-01 -7.25188494e-01 7.16244459e-01 -6.53886050e-02
-4.13632751e-01 1.01155746e+00 5.37146151e-01 5.83952606e-01
1.07327056e+00 -1.40586710e+00 8.03177238e-01 1.09008467e+00
5.51963091e-01 1.04919255e+00 1.06389558e+00 -9.40029204e-01
-1.16200984e+00 -6.48889840e-01 2.24834061e+00 -7.64463544e-01
1.05277693e+00 -3.10100764e-01 -7.52043366e-01 6.65281951e-01
2.79156357e-01 -7.08883524e-01 9.64793384e-01 2.13512391e-01
-4.31852847e-01 4.94840652e-01 -1.01686859e+00 3.71800333e-01
8.81772399e-01 -9.03769016e-01 -1.20872223e+00 5.42310715e-01
2.08678871e-01 -2.23785117e-01 -9.96440709e-01 -3.80403817e-01
5.90631962e-01 -6.25220358e-01 5.80500603e-01 -5.92845857e-01
5.42160392e-01 2.00287670e-01 8.07449408e-03 -1.16708267e+00
-6.08565450e-01 -8.60451996e-01 4.00095046e-01 1.50022566e+00
6.45777225e-01 -5.37463784e-01 6.51892900e-01 5.04952908e-01
5.09495996e-02 -1.30651891e-01 -5.98111808e-01 -1.14937496e+00
4.51359272e-01 -4.96243358e-01 3.97691458e-01 8.77969503e-01
5.17243564e-01 3.17123830e-01 -2.90840447e-01 -4.03240263e-01
2.37333536e-01 -1.56851381e-01 4.09019977e-01 -1.37438965e+00
-2.00307742e-01 -7.80803502e-01 -3.38892639e-01 -1.17799902e+00
4.10344422e-01 -1.15534174e+00 2.15173125e-01 -1.08350873e+00
3.70302387e-02 2.14071982e-02 -6.35728985e-02 1.19207233e-01
-7.46171847e-02 2.21088305e-01 1.68403029e-01 3.97816986e-01
-3.25776249e-01 -3.78575683e-01 6.12763286e-01 -5.77344969e-02
4.54622060e-02 1.59257039e-01 -8.41980577e-01 8.80438030e-01
8.65028024e-01 -5.65616131e-01 2.11321205e-01 -3.72864842e-01
3.94008249e-01 -5.41829644e-03 -1.24111578e-01 -8.55672121e-01
3.13735485e-01 -3.94903272e-01 5.77041149e-01 -6.79491401e-01
-3.36259663e-01 -5.96689820e-01 -3.24749559e-01 5.41073978e-01
-7.95353413e-01 6.81200624e-01 2.52566785e-02 3.34114999e-01
-1.06481202e-01 -1.12980902e+00 6.37681305e-01 -5.00406861e-01
-6.96464658e-01 -2.65073359e-01 -1.37417293e+00 3.93096775e-01
6.18752420e-01 -5.21729350e-01 -1.24558620e-01 2.96201129e-02
-1.90756947e-01 -3.98197234e-01 7.95366943e-01 3.75251710e-01
2.11499825e-01 -8.09981287e-01 -6.44978523e-01 2.09434837e-01
2.83360839e-01 -1.00998199e+00 -4.43084896e-01 3.38042319e-01
-1.10582328e+00 9.53299642e-01 -1.64879337e-01 1.85624033e-01
-1.41502523e+00 2.73788929e-01 -1.09927110e-01 -3.73081475e-01
-4.39704835e-01 8.74220073e-01 -4.36372787e-01 -2.64909655e-01
2.11533219e-01 -1.69815034e-01 -4.10957757e-04 1.05484560e-01
8.75733793e-01 8.71170878e-01 3.65479141e-01 -8.02443981e-01
-6.21251881e-01 1.40150830e-01 -6.61905587e-01 -5.67845404e-01
6.55116677e-01 -2.21396491e-01 -4.83960398e-02 9.63074923e-01
1.32082164e+00 8.04837227e-01 -1.14171170e-01 -4.38189447e-01
6.62633121e-01 -7.37643421e-01 9.82916132e-02 -1.09042263e+00
-1.92225963e-01 6.65318727e-01 1.94646806e-01 1.36061043e-01
3.13068628e-01 -2.01934934e-01 7.91523457e-01 6.04236305e-01
3.06891769e-01 -1.63217843e+00 -2.99063593e-01 9.92321491e-01
6.41178846e-01 -1.15725851e+00 -5.96745573e-02 -1.67475849e-01
-4.57703084e-01 1.57644379e+00 4.53586787e-01 -7.69078135e-02
4.21056569e-01 1.75258011e-01 2.20676810e-01 3.86507325e-02
-3.94743621e-01 2.10088059e-01 3.55143011e-01 3.35471064e-01
1.04443216e+00 8.64526257e-02 -9.61963236e-01 6.21609926e-01
-6.24982536e-01 -2.44130597e-01 7.86249459e-01 1.13413513e+00
-9.11751986e-01 -1.22478175e+00 -6.03441119e-01 7.49122083e-01
-1.00095260e+00 -4.06258017e-01 -9.94655252e-01 6.11063182e-01
-1.41762286e-01 9.89386916e-01 1.21052600e-01 -1.40767321e-01
-6.43385947e-02 6.31055653e-01 4.13380772e-01 -9.66480255e-01
-1.06057453e+00 -3.53209972e-02 3.21357310e-01 -1.51464874e-02
1.17924668e-01 -1.16774786e+00 -9.16229129e-01 -4.96652216e-01
-2.64039308e-01 4.91143972e-01 4.45709169e-01 1.11219370e+00
-1.02988191e-01 1.18072808e-01 3.90994549e-01 -1.98192358e-01
-8.25752795e-01 -1.02297091e+00 -8.89391303e-01 2.63719112e-01
5.50366789e-02 -1.03875436e-01 -6.90935433e-01 1.43917575e-01] | [10.429397583007812, 10.515497207641602] |
44d5ba20-1f46-4aa8-9189-408c445afaa7 | pcpnet-an-efficient-and-semantic-enhanced | 2304.07773 | null | https://arxiv.org/abs/2304.07773v1 | https://arxiv.org/pdf/2304.07773v1.pdf | PCPNet: An Efficient and Semantic-Enhanced Transformer Network for Point Cloud Prediction | The ability to predict future structure features of environments based on past perception information is extremely needed by autonomous vehicles, which helps to make the following decision-making and path planning more reasonable. Recently, point cloud prediction (PCP) is utilized to predict and describe future environmental structures by the point cloud form. In this letter, we propose a novel efficient Transformer-based network to predict the future LiDAR point clouds exploiting the past point cloud sequences. We also design a semantic auxiliary training strategy to make the predicted LiDAR point cloud sequence semantically similar to the ground truth and thus improves the significance of the deployment for more tasks in real-vehicle applications. Our approach is completely self-supervised, which means it does not require any manual labeling and has a solid generalization ability toward different environments. The experimental results show that our method outperforms the state-of-the-art PCP methods on the prediction results and semantic similarity, and has a good real-time performance. Our open-source code and pre-trained models are available at https://github.com/Blurryface0814/PCPNet. | ['Guangming Xiong', 'Zijie Zhou', 'Junyi Ma', 'Zhen Luo'] | 2023-04-16 | null | null | null | null | ['semantic-textual-similarity', 'semantic-similarity'] | ['natural-language-processing', 'natural-language-processing'] | [-9.01176259e-02 -4.38910365e-01 -3.32722992e-01 -7.56735921e-01
-1.96522877e-01 -2.95549273e-01 5.44538915e-01 -3.91807109e-02
-1.09855272e-01 6.55260921e-01 -2.42066056e-01 -5.58920860e-01
-2.50970963e-02 -1.16138494e+00 -7.95415819e-01 -4.32955056e-01
-6.73005730e-02 6.40396655e-01 7.56714702e-01 -2.25356430e-01
2.77905196e-01 8.84914756e-01 -1.87790406e+00 3.78804356e-02
1.31275737e+00 1.09870136e+00 9.53902304e-01 3.06984670e-02
-4.28407937e-01 4.33134526e-01 1.61012530e-01 -2.41608635e-01
2.96883345e-01 3.29105675e-01 -4.28408593e-01 -1.46336062e-02
1.78509325e-01 -2.63306946e-01 -4.53877985e-01 1.22518981e+00
3.32377069e-02 -6.73538148e-02 3.31086159e-01 -1.53211224e+00
-5.99273264e-01 1.09526560e-01 -3.76902968e-01 -1.27172589e-01
-6.98605254e-02 3.23965907e-01 6.79278076e-01 -1.01298809e+00
4.55210000e-01 1.28392231e+00 7.71892428e-01 1.55851990e-01
-6.59483373e-01 -1.13491273e+00 4.20824617e-01 9.91584063e-01
-1.56802487e+00 -3.49505603e-01 1.01526022e+00 -3.47237080e-01
6.63323641e-01 -9.35237780e-02 8.02190006e-01 7.95770824e-01
1.92677513e-01 6.73729300e-01 9.78729248e-01 4.87086512e-02
1.23495989e-01 2.31714129e-01 -1.82600766e-01 7.92355716e-01
2.03055203e-01 3.73011112e-01 -2.00273395e-01 1.27956644e-01
4.37528193e-01 5.37373602e-01 -2.64733672e-01 -7.11192727e-01
-1.10704541e+00 5.53487837e-01 8.87876630e-01 4.32156920e-02
-4.44525540e-01 1.86896577e-01 1.96448520e-01 -1.08453035e-01
5.82073808e-01 -2.33498320e-01 -5.07174432e-01 3.55743691e-02
-6.44073784e-01 1.49692178e-01 3.32685173e-01 1.38662124e+00
1.27997017e+00 1.97590992e-01 4.68970597e-01 6.03891850e-01
4.74966228e-01 1.02444017e+00 2.20092520e-01 -1.10588694e+00
5.46864927e-01 5.07597089e-01 3.21989298e-01 -1.14518797e+00
-2.23322630e-01 -4.88388419e-01 -7.07985997e-01 1.13396950e-01
-2.53627837e-01 1.71139270e-01 -8.73939753e-01 1.33945107e+00
3.67845654e-01 7.77975380e-01 2.30706315e-02 6.91093326e-01
6.46151364e-01 9.95937645e-01 -7.40467235e-02 -4.75524105e-02
8.65133405e-01 -1.01313555e+00 -5.49876511e-01 -4.90541697e-01
4.74338561e-01 -7.28374898e-01 9.07694817e-01 2.17814341e-01
-3.02928895e-01 -1.00653076e+00 -1.19152606e+00 1.54366985e-01
-4.11069304e-01 2.04888463e-01 7.51616180e-01 1.98296636e-01
-9.48444188e-01 8.07890594e-01 -9.84835684e-01 -4.22940910e-01
5.96669614e-01 2.34272435e-01 -2.03494832e-01 -4.02723193e-01
-9.72803652e-01 1.00778818e+00 6.16466284e-01 2.56809026e-01
-7.54967749e-01 -6.55889332e-01 -8.60123873e-01 -1.19745702e-01
3.04752469e-01 -5.64977467e-01 1.10768056e+00 -4.72422361e-01
-1.17414546e+00 4.46612179e-01 -6.41885817e-01 -5.54454088e-01
3.82265568e-01 -2.48885736e-01 -6.43119454e-01 -2.00093806e-01
4.12477612e-01 9.47570264e-01 5.15897334e-01 -1.57789028e+00
-1.14192104e+00 -4.24360305e-01 -2.37610698e-01 2.49081925e-01
-9.88839641e-02 -5.01071930e-01 -5.34698665e-01 -1.01077035e-01
5.26639223e-01 -1.03283215e+00 -3.42081189e-01 4.47977930e-01
-2.45247278e-02 -3.53236020e-01 1.32323456e+00 -3.95083010e-01
5.08410096e-01 -2.08195210e+00 -5.46528280e-01 5.21456860e-02
-5.33131734e-02 3.47482353e-01 -8.52124318e-02 3.43764991e-01
6.38431162e-02 -1.22886918e-01 -1.84716821e-01 -3.00776213e-01
-1.75430641e-01 6.46217704e-01 -8.04854035e-01 3.90446186e-01
8.98342207e-02 8.62102568e-01 -9.86302674e-01 -5.13007998e-01
5.42973161e-01 4.12617952e-01 -2.93009102e-01 7.89150074e-02
-4.57067370e-01 6.34001553e-01 -7.55839586e-01 5.26837289e-01
1.16172886e+00 -9.27590057e-02 -1.45701826e-01 -1.64508075e-01
-5.32345772e-01 2.22653836e-01 -8.36924911e-01 1.87825942e+00
-6.19977236e-01 5.75671971e-01 -3.73295814e-01 -1.00525129e+00
1.50938809e+00 -1.10122643e-01 4.28385317e-01 -7.84036994e-01
-1.31862551e-01 4.89079803e-01 -2.63155729e-01 -4.78531778e-01
4.94811982e-01 -1.07028477e-01 2.48101458e-01 -2.25588948e-01
-3.37820411e-01 -3.19256306e-01 -1.91704869e-01 -1.23947293e-01
5.60316026e-01 5.62402666e-01 2.93297227e-02 -3.06688007e-02
8.30315888e-01 3.63555491e-01 1.11290538e+00 2.97718942e-01
-2.59413779e-01 2.31051698e-01 -3.02166611e-01 -5.97190976e-01
-9.72537220e-01 -1.09892225e+00 -2.38174751e-01 3.36094975e-01
8.83858442e-01 -1.83787391e-01 -4.89513502e-02 -3.82436633e-01
2.45524853e-01 1.02914560e+00 -6.28495365e-02 -2.14939684e-01
-6.41657412e-01 -1.70535594e-01 7.82086104e-02 7.05442548e-01
9.63530302e-01 -8.55638266e-01 -3.40117335e-01 2.26565301e-01
-4.35633034e-01 -1.47880590e+00 1.17319278e-01 -2.68500239e-01
-1.18091679e+00 -8.12124550e-01 -2.78050214e-01 -8.35869193e-01
4.05552804e-01 8.49207878e-01 6.44973278e-01 8.59788731e-02
2.66719967e-01 1.00933798e-01 -3.63688022e-01 -7.20455527e-01
-1.18065901e-01 -2.17350259e-01 3.36606324e-01 -1.03817537e-01
7.05086410e-01 -1.01538765e+00 -5.51070213e-01 4.18320954e-01
-2.78708279e-01 3.17700416e-01 6.14818156e-01 5.73168159e-01
9.42666531e-01 3.24539185e-01 2.93787837e-01 -4.52450514e-01
-6.95148855e-02 -4.89059806e-01 -9.05142248e-01 -1.98175889e-02
-8.27737510e-01 -1.42630368e-01 7.88377881e-01 -2.42659189e-02
-1.20081675e+00 3.08238178e-01 -2.93817818e-01 -8.45719874e-01
-3.36516887e-01 3.93338889e-01 -3.02798033e-01 -2.54620552e-01
2.81763524e-01 6.70514703e-01 -1.20702259e-01 -7.27948427e-01
3.94589841e-01 6.34557366e-01 8.20596278e-01 -4.23771918e-01
1.23880351e+00 8.45085025e-01 1.14986278e-01 -5.75571597e-01
-8.06889057e-01 -5.79652548e-01 -8.47375512e-01 -2.82912761e-01
5.85003197e-01 -1.19441116e+00 -5.47522843e-01 3.29676092e-01
-1.49959290e+00 -3.26154381e-02 -2.56519951e-02 5.01758695e-01
-6.85889721e-01 5.50258517e-01 -1.82673976e-01 -7.62085438e-01
-3.22228760e-01 -1.12586010e+00 9.64627624e-01 2.39184171e-01
4.71610963e-01 -6.31577492e-01 -2.23219588e-01 2.26824969e-01
1.72047466e-01 3.42901647e-02 6.88699067e-01 -2.17213213e-01
-1.37311256e+00 6.48894235e-02 -4.82678920e-01 2.30983913e-01
1.22372553e-01 -4.30936739e-02 -8.62863958e-01 -1.37087032e-01
5.89895695e-02 1.49956405e-01 9.33147013e-01 1.25947133e-01
1.49734128e+00 -8.57090726e-02 -8.50055277e-01 8.53635430e-01
1.43723965e+00 4.13411260e-01 5.73435664e-01 4.89398926e-01
7.53899813e-01 6.71766341e-01 1.43278348e+00 2.84797162e-01
9.16714609e-01 6.75955951e-01 9.24243331e-01 1.21465191e-01
-1.27942180e-02 -7.72858083e-01 1.03327878e-01 9.13755476e-01
9.43532884e-02 1.47983944e-02 -1.16766536e+00 7.54483342e-01
-2.05751109e+00 -1.00304627e+00 -3.20959091e-01 1.86395121e+00
2.34545991e-01 2.68379897e-01 -4.66284484e-01 -4.74148057e-02
7.94077575e-01 2.03481227e-01 -7.82347143e-01 7.01819081e-03
1.99912995e-01 -7.56360292e-02 6.15803838e-01 3.23690981e-01
-1.04894340e+00 1.31196702e+00 4.81409073e+00 1.00892711e+00
-1.35000813e+00 2.55975664e-01 1.70943245e-01 4.52395469e-01
-3.48424584e-01 2.88118243e-01 -8.81249428e-01 6.51945889e-01
7.08377659e-01 -3.52164090e-01 1.71403170e-01 1.18789995e+00
4.65438873e-01 1.10500321e-01 -7.98519790e-01 1.09055102e+00
-3.47188532e-01 -1.64730644e+00 1.05761103e-01 -1.54104149e-02
4.23305809e-01 7.28020966e-01 1.12763144e-01 2.69150943e-01
2.70782650e-01 -5.98747432e-01 8.54138136e-01 6.01748168e-01
8.98117542e-01 -7.20281959e-01 6.50543749e-01 7.96704888e-01
-1.74136090e+00 -3.38397235e-01 -8.39857280e-01 -1.02121949e-01
4.20455575e-01 6.00772500e-01 -1.05732596e+00 8.79753292e-01
8.94685328e-01 1.38475978e+00 -4.83883888e-01 1.30293429e+00
-3.29890579e-01 4.63291258e-01 -3.35714996e-01 -1.45389214e-02
2.65387654e-01 -4.15211737e-01 6.00169778e-01 7.62062788e-01
7.73031235e-01 1.22508168e-01 5.10021031e-01 8.36338997e-01
2.85780787e-01 1.84101276e-02 -8.85421932e-01 3.90545398e-01
9.19795394e-01 1.04344904e+00 -3.95679057e-01 -3.80249113e-01
-3.93891692e-01 6.30918920e-01 1.83860689e-01 2.22617164e-01
-8.99049520e-01 -2.09787533e-01 9.01305318e-01 1.60175279e-01
5.92278600e-01 -5.57337999e-01 -3.20186704e-01 -1.06385517e+00
2.90297627e-01 -2.18560338e-01 -8.92085582e-02 -1.28202105e+00
-1.14187860e+00 6.76539898e-01 2.40527661e-05 -1.99021232e+00
-1.15428805e-01 -6.03075147e-01 -7.07510531e-01 8.67538095e-01
-2.23663878e+00 -1.48044634e+00 -6.93233550e-01 4.62673604e-01
6.13320649e-01 -1.43640324e-01 5.51221073e-01 1.89195067e-01
-1.15347765e-01 -4.66840267e-02 1.02068044e-01 -1.61128297e-01
3.67927402e-01 -7.01326609e-01 7.68018544e-01 9.46185052e-01
5.95163740e-03 3.30523252e-01 7.28515089e-01 -8.71649086e-01
-1.23122990e+00 -1.71894991e+00 8.23045969e-01 -7.21869171e-02
5.97373664e-01 -7.10546151e-02 -9.91159499e-01 6.61832571e-01
-2.70151436e-01 7.40651265e-02 2.64937967e-01 -1.55247033e-01
-2.43872806e-01 -6.98036909e-01 -9.71899152e-01 5.56402206e-01
1.52354848e+00 -3.83800238e-01 -5.27794302e-01 4.75410283e-01
1.14119792e+00 -3.48791629e-01 -4.40455496e-01 1.04429948e+00
3.23726565e-01 -9.99398887e-01 9.35087979e-01 -1.04475684e-01
9.57255960e-02 -8.08595538e-01 -2.70174533e-01 -1.13003397e+00
-5.90405524e-01 8.99132807e-03 -1.47452774e-02 9.86379206e-01
1.68949276e-01 -1.00895262e+00 1.02507949e+00 1.43698022e-01
-6.81672931e-01 -7.79132247e-01 -9.68566358e-01 -1.00658083e+00
-1.50522217e-01 -7.70491302e-01 1.10980928e+00 7.66540408e-01
-5.70921302e-01 7.66179189e-02 -2.38054022e-01 6.75410330e-01
6.26563847e-01 6.79440737e-01 9.20489192e-01 -1.56588256e+00
1.92812979e-01 -5.94944842e-02 -7.70441413e-01 -1.52829003e+00
4.64681625e-01 -1.02607667e+00 7.96710774e-02 -1.66277194e+00
-1.62325487e-01 -1.03258789e+00 -1.89131275e-01 5.13202488e-01
1.43957749e-01 -6.35754615e-02 2.30515614e-01 5.59666574e-01
-3.47943068e-01 1.27185118e+00 1.28055561e+00 -3.14437866e-01
1.89008936e-02 3.51376921e-01 -3.31261754e-01 8.45116675e-01
9.27860200e-01 -4.08289909e-01 -6.72510505e-01 -7.31300890e-01
-2.06169829e-01 -2.64168512e-02 4.58118349e-01 -1.49362600e+00
4.39105362e-01 -3.82065058e-01 2.60857582e-01 -1.42290831e+00
8.01752269e-01 -1.20795286e+00 3.15455317e-01 5.84259808e-01
3.96046072e-01 1.66730523e-01 2.31912032e-01 8.73320937e-01
-1.95286050e-01 -8.49761814e-02 4.46651042e-01 -8.93636793e-02
-1.47211993e+00 9.53532934e-01 2.67657563e-02 -6.24951303e-01
1.15564072e+00 -4.06248122e-01 -3.02080601e-01 -3.37369233e-01
-2.13228539e-01 6.94358945e-01 7.06960976e-01 7.46374547e-01
1.01007366e+00 -1.48235333e+00 -4.37938005e-01 2.21004188e-01
4.60919678e-01 2.35181123e-01 2.85080612e-01 4.66406435e-01
-6.68010473e-01 6.78873181e-01 -2.28267819e-01 -1.06751800e+00
-1.11809599e+00 6.90068007e-01 3.03325683e-01 2.36651212e-01
-9.18554962e-01 5.65535486e-01 1.91132098e-01 -7.70962238e-01
-1.71628863e-01 -2.28661373e-01 -2.97081888e-01 -5.67636847e-01
3.18581015e-01 1.50696531e-01 -1.62882041e-02 -9.54023361e-01
-3.79152119e-01 7.60838568e-01 1.51476398e-01 1.22489549e-01
1.36924446e+00 -3.32485497e-01 -1.29110038e-01 4.71303940e-01
1.05280793e+00 -3.46513003e-01 -1.36659396e+00 -6.65585339e-01
-2.42926460e-02 -7.93933690e-01 8.36402457e-03 -4.77818698e-01
-1.00969207e+00 1.13525641e+00 7.48987257e-01 -3.17283601e-01
9.41825628e-01 -1.70812756e-01 1.10975051e+00 6.42894328e-01
1.16799664e+00 -6.90482855e-01 -2.65797496e-01 7.18179941e-01
7.73923278e-01 -1.14357555e+00 3.92000005e-02 -1.01357841e+00
-5.01142442e-01 1.00896382e+00 8.28314304e-01 -6.37821928e-02
8.30912054e-01 -1.80992484e-01 4.38486151e-02 1.54789798e-02
-6.80619597e-01 -3.32767963e-01 1.64417148e-01 9.53040361e-01
-2.70411730e-01 2.29194924e-01 8.19236785e-02 3.64188075e-01
-5.99796414e-01 1.65272966e-01 1.66026741e-01 7.21944928e-01
-8.44544291e-01 -1.08162892e+00 -1.52793095e-01 4.23421532e-01
3.86194557e-01 -5.53239509e-03 1.55956268e-01 5.28680682e-01
5.87098837e-01 9.24487233e-01 2.22494379e-01 -6.49005055e-01
2.49531254e-01 -1.23692356e-01 1.63969383e-01 -6.35962307e-01
5.01494467e-01 -3.65283817e-01 -3.22681218e-02 -6.50705695e-01
-4.46964443e-01 -6.54900312e-01 -1.40880597e+00 -4.00207311e-01
-3.31295103e-01 2.07010061e-01 8.29538107e-01 9.65632498e-01
6.91957235e-01 2.02748388e-01 8.64713490e-01 -1.00642824e+00
-3.41321915e-01 -6.85983002e-01 -1.83403119e-01 9.82692912e-02
1.16302274e-01 -9.73838270e-01 1.02565832e-01 -9.46356803e-02] | [8.137896537780762, -2.5844075679779053] |
3831d169-06b7-4c20-9893-8d2ac4be8d09 | convolution-channel-separation-and-frequency | 2211.01599 | null | https://arxiv.org/abs/2211.01599v1 | https://arxiv.org/pdf/2211.01599v1.pdf | Convolution channel separation and frequency sub-bands aggregation for music genre classification | In music, short-term features such as pitch and tempo constitute long-term semantic features such as melody and narrative. A music genre classification (MGC) system should be able to analyze these features. In this research, we propose a novel framework that can extract and aggregate both short- and long-term features hierarchically. Our framework is based on ECAPA-TDNN, where all the layers that extract short-term features are affected by the layers that extract long-term features because of the back-propagation training. To prevent the distortion of short-term features, we devised the convolution channel separation technique that separates short-term features from long-term feature extraction paths. To extract more diverse features from our framework, we incorporated the frequency sub-bands aggregation method, which divides the input spectrogram along frequency bandwidths and processes each segment. We evaluated our framework using the Melon Playlist dataset which is a large-scale dataset containing 600 times more data than GTZAN which is a widely used dataset in MGC studies. As the result, our framework achieved 70.4% accuracy, which was improved by 16.9% compared to a conventional framework. | ['Ha-Jin Yu', 'Chan-yeong Lim', 'Ju-ho Kim', 'Hyun-seo Shin', 'Jungwoo Heo'] | 2022-11-03 | null | null | null | null | ['genre-classification'] | ['computer-vision'] | [ 7.62404175e-03 -5.87648511e-01 1.21352956e-01 -2.30899349e-01
-6.40515327e-01 -6.82223916e-01 2.69130141e-01 6.93326211e-03
-5.52211940e-01 4.17345881e-01 3.84131074e-01 1.68448761e-01
-4.20954376e-01 -1.02576387e+00 -2.90002733e-01 -4.95674878e-01
-4.84761924e-01 -3.60320091e-01 1.60970196e-01 -3.49614352e-01
6.15823567e-01 1.82434738e-01 -1.92880404e+00 9.26719487e-01
5.33650041e-01 1.50451648e+00 5.14732674e-02 8.33508790e-01
-3.91004711e-01 8.21291983e-01 -9.50934350e-01 -3.90138701e-02
7.62962177e-02 -5.45213401e-01 -9.28694963e-01 -3.03492725e-01
1.45215588e-02 1.38885915e-01 -1.87758341e-01 9.22596812e-01
7.48226762e-01 2.69470543e-01 4.04506028e-01 -9.94694710e-01
-3.84606391e-01 1.22103214e+00 -2.55460054e-01 2.09527850e-01
3.47114176e-01 1.81423463e-02 1.21276391e+00 -7.48296320e-01
2.36174896e-01 8.96350920e-01 1.09946382e+00 2.68976957e-01
-7.05561817e-01 -1.03184247e+00 2.82604098e-02 5.53939641e-01
-1.36444783e+00 -3.08261663e-01 1.08041930e+00 -4.61532146e-01
1.02007115e+00 5.04616439e-01 8.15981865e-01 8.27829301e-01
2.08499357e-01 6.65842354e-01 8.49783897e-01 -4.58916247e-01
-1.46658227e-01 -3.81935984e-01 3.41693223e-01 1.47110805e-01
-5.01198947e-01 6.62867725e-02 -8.28351021e-01 -1.05137609e-01
7.16990352e-01 -1.63642630e-01 -1.93668738e-01 5.50038099e-01
-1.30441391e+00 7.10848749e-01 3.43057424e-01 6.87786162e-01
-3.97934973e-01 1.61348075e-01 7.32045233e-01 4.64886278e-01
2.89739788e-01 5.97693026e-01 -5.40347278e-01 -7.27546871e-01
-9.92194772e-01 1.99820042e-01 7.73232698e-01 5.52354097e-01
3.14239353e-01 1.38033435e-01 -3.95993292e-01 1.15343344e+00
-7.84279928e-02 1.02059893e-01 1.08526826e+00 -7.44274080e-01
4.48531955e-01 5.68516195e-01 -2.57219136e-01 -1.12217188e+00
-5.23199141e-01 -8.36813688e-01 -8.39742362e-01 -7.84755275e-02
3.22533429e-01 -1.14973381e-01 -5.76378345e-01 1.73015249e+00
-8.53396766e-03 3.15742195e-01 7.78502040e-03 1.06565225e+00
9.56659138e-01 7.90219843e-01 -2.34340459e-01 -2.11070880e-01
1.33653092e+00 -8.96443903e-01 -8.75561237e-01 2.77452677e-01
2.32469574e-01 -1.27234626e+00 1.35241687e+00 6.75138295e-01
-9.60792422e-01 -9.93432820e-01 -1.25765073e+00 -1.86206520e-01
-3.70742947e-01 2.78853863e-01 5.76885521e-01 3.76945555e-01
-5.06234825e-01 1.16139698e+00 -4.05908465e-01 1.58504639e-02
1.01823494e-01 2.89356261e-01 4.65958565e-02 8.47848833e-01
-1.53703201e+00 3.25090617e-01 7.04654634e-01 1.84940994e-01
-4.43533570e-01 -5.81524312e-01 -4.66365099e-01 2.48909280e-01
2.12747321e-01 -2.50207316e-02 1.26093447e+00 -1.00406861e+00
-1.75101197e+00 3.86707127e-01 1.90728530e-01 -3.66799355e-01
8.76354054e-02 -2.62535334e-01 -8.94761205e-01 1.25248963e-02
-4.21261527e-02 2.68751204e-01 7.21683562e-01 -4.02707785e-01
-9.42556739e-01 -1.98969170e-01 1.54517651e-01 3.15636754e-01
-6.15991592e-01 3.31923999e-02 -3.34256977e-01 -1.26600695e+00
1.97077140e-01 -8.19226086e-01 3.43354285e-01 -5.51335633e-01
-2.43891150e-01 -2.02720836e-01 6.03043258e-01 -6.56431079e-01
1.98574615e+00 -2.54564929e+00 -2.82128304e-02 1.43826678e-01
-2.03091484e-02 1.46182463e-01 -1.03113212e-01 4.90632772e-01
-8.59872848e-02 -4.33199033e-02 -4.96573858e-02 4.26350348e-02
9.29462388e-02 -8.73639956e-02 -5.12179673e-01 6.02926724e-02
-1.06737658e-01 6.53537631e-01 -5.47701836e-01 -3.49875033e-01
1.17949627e-01 3.18688303e-01 -4.80355352e-01 -7.61610940e-02
-1.47654071e-01 1.07850559e-01 -2.24321380e-01 5.05177259e-01
4.71017212e-01 4.46748957e-02 -1.17918678e-01 -2.31601804e-01
-4.50247884e-01 6.64513111e-01 -1.25226140e+00 2.14981818e+00
-5.65111339e-01 6.31686985e-01 -2.84487933e-01 -7.78250217e-01
1.10891306e+00 5.15247524e-01 6.46280229e-01 -5.19269407e-01
4.85021502e-01 3.56548488e-01 1.98062971e-01 -4.39798355e-01
6.25519812e-01 -3.40564102e-01 -4.51954603e-01 3.61815125e-01
2.07506657e-01 7.78026655e-02 9.43338349e-02 -3.21188092e-01
8.61219108e-01 -1.02167353e-02 7.04476088e-02 -2.16707408e-01
7.42504835e-01 -1.18322127e-01 7.28671908e-01 5.27161241e-01
-6.49011284e-02 7.14469075e-01 2.41864383e-01 -5.13821483e-01
-5.47562182e-01 -7.96409130e-01 -1.20781355e-01 1.21658206e+00
3.43266316e-02 -1.03685844e+00 -5.85700333e-01 -4.97846007e-01
-1.98128112e-02 2.72074133e-01 -4.76540625e-01 -2.67118752e-01
-4.60725129e-01 -6.51097775e-01 1.19432521e+00 4.82027411e-01
6.92735374e-01 -1.39593077e+00 -4.10724789e-01 6.92835987e-01
-5.04433990e-01 -7.13210106e-01 -8.12046587e-01 3.00072610e-01
-7.15397060e-01 -1.03742111e+00 -5.46830773e-01 -7.68940508e-01
-4.22610015e-01 -7.05800578e-02 9.15277183e-01 -1.74024373e-01
-1.81532383e-01 -3.10376346e-01 -6.03805125e-01 -4.27698731e-01
4.43056896e-02 5.12557685e-01 -2.44245455e-02 1.52614594e-01
5.69963396e-01 -9.31964457e-01 -5.95166326e-01 2.01583132e-01
-7.14670777e-01 9.41588581e-02 3.21004063e-01 7.21053898e-01
6.07668102e-01 3.15477312e-01 9.04582679e-01 -5.65711141e-01
9.29370344e-01 -2.18077198e-01 -1.26371086e-01 -6.54577836e-02
-4.75762397e-01 -2.24183202e-01 9.76268172e-01 -7.65268683e-01
-7.15817571e-01 -1.37526989e-01 -4.18452233e-01 -2.08080694e-01
4.91986796e-02 6.90171123e-01 -1.98307410e-01 1.49810612e-01
6.28011644e-01 2.78329760e-01 -3.27907294e-01 -9.32933271e-01
1.70801774e-01 1.32770288e+00 8.52456808e-01 -7.43855357e-01
3.59402120e-01 8.55133012e-02 -3.39896291e-01 -7.14517236e-01
-9.35687482e-01 -5.87815762e-01 -4.62784141e-01 -3.99476439e-01
8.35669458e-01 -7.91112900e-01 -9.67728674e-01 7.76030242e-01
-1.09553051e+00 1.28450960e-01 -5.17543435e-01 9.28455472e-01
-6.05447710e-01 -2.19813455e-02 -9.56747651e-01 -4.92794245e-01
-5.67978561e-01 -7.00240672e-01 7.91399479e-01 2.68817991e-01
-4.00143713e-01 -5.92828333e-01 1.93192810e-02 3.93472314e-02
5.08249223e-01 2.63754278e-01 9.05710161e-01 -4.34681565e-01
6.95936941e-03 1.39362827e-01 1.60519555e-01 6.02351487e-01
2.32097447e-01 -1.42735273e-01 -1.38172925e+00 -1.51962535e-02
1.81216553e-01 -2.32320055e-01 8.88979077e-01 2.58128792e-01
1.64198673e+00 -1.32265404e-01 3.05938154e-01 9.52548921e-01
1.16825533e+00 4.95399296e-01 5.88917196e-01 4.91173685e-01
6.45126045e-01 1.09205894e-01 5.40421426e-01 5.57261765e-01
1.36603788e-01 5.44891238e-01 -4.33057211e-02 1.11720189e-01
-3.11075240e-01 -3.09619218e-01 4.10525709e-01 1.50958657e+00
-3.67598444e-01 1.72370419e-01 -5.61801016e-01 3.98905516e-01
-1.85586739e+00 -1.19829392e+00 -1.89105287e-01 1.90377760e+00
1.24090827e+00 3.49828422e-01 3.21609557e-01 8.81861925e-01
6.82707667e-01 2.75234908e-01 -4.65371996e-01 -5.60010433e-01
-1.69446424e-01 6.37674510e-01 1.87785961e-02 1.44071534e-01
-1.42678452e+00 7.67181695e-01 6.07266045e+00 1.48737752e+00
-1.43581569e+00 4.43531685e-02 1.16225064e-01 -4.90621477e-01
-9.99667123e-03 -2.07584307e-01 -4.34908122e-01 5.99033296e-01
9.65044975e-01 -1.52228385e-01 5.00941038e-01 6.19670868e-01
6.69928119e-02 3.23360562e-01 -9.40544724e-01 1.28859627e+00
-8.95661488e-02 -1.26302314e+00 -7.49155134e-02 -1.36730433e-01
5.36736012e-01 -3.04140300e-01 -2.71174237e-02 6.13471627e-01
-3.30091536e-01 -1.00928223e+00 1.03021681e+00 6.59168661e-01
9.11096156e-01 -1.13916147e+00 5.93015373e-01 1.91079065e-01
-1.71078324e+00 -1.99234888e-01 -1.44985527e-01 -6.78422749e-01
-1.60485268e-01 7.41531193e-01 -1.39191121e-01 7.80729711e-01
7.32497931e-01 1.03749251e+00 -4.14195865e-01 1.03586745e+00
8.59640017e-02 7.84394622e-01 -3.20364386e-01 4.90651466e-02
2.82494962e-01 -6.69068024e-02 4.55131739e-01 1.31889236e+00
4.93348688e-01 -6.92453682e-02 1.24079809e-01 5.52716434e-01
5.88012375e-02 2.84968853e-01 -2.77419016e-02 -1.26889125e-01
5.17197490e-01 1.10189807e+00 -7.14426100e-01 -2.62713760e-01
-2.25887030e-01 9.30174768e-01 2.54051071e-02 1.16693221e-01
-1.11563849e+00 -1.17049706e+00 5.57258368e-01 -1.33397415e-01
1.95899129e-01 -5.20385727e-02 -4.25654382e-01 -1.09368587e+00
2.19801515e-01 -8.60452712e-01 5.29607058e-01 -4.93057519e-01
-1.39967144e+00 8.51161659e-01 -4.27451998e-01 -1.58570361e+00
-2.28582025e-01 -2.56163627e-01 -5.89588046e-01 9.62903380e-01
-1.11162007e+00 -1.00118291e+00 -1.67917296e-01 8.48015785e-01
5.60586691e-01 -3.78516495e-01 1.06345689e+00 6.93304360e-01
-2.65480846e-01 6.27085388e-01 3.72648053e-02 2.72349149e-01
7.62334526e-01 -1.08833110e+00 1.07345343e-01 4.75837439e-01
5.39761186e-01 5.44680059e-01 2.21099824e-01 -3.12668651e-01
-8.92336428e-01 -1.03207791e+00 8.63295078e-01 1.68569773e-01
6.98439419e-01 -3.74768376e-01 -7.64213920e-01 3.10588449e-01
-4.03215252e-02 -2.20475674e-01 9.89524007e-01 4.01521146e-01
-3.12133521e-01 -4.34161276e-01 -6.09129310e-01 2.13812858e-01
1.06521165e+00 -9.98697042e-01 -8.79391909e-01 -3.59096587e-01
6.45467460e-01 -2.35236362e-01 -1.07088578e+00 4.91365045e-01
1.10922742e+00 -1.01156688e+00 7.61958897e-01 -1.90451324e-01
4.34833676e-01 -5.45396209e-01 -2.48645857e-01 -1.27074945e+00
-4.26234424e-01 -5.52998781e-01 -1.33243278e-01 1.48291624e+00
3.54491830e-01 -2.61148155e-01 3.36324602e-01 -3.08990270e-01
-2.67145634e-01 -6.60239995e-01 -8.10279489e-01 -8.75080168e-01
-7.23911449e-03 -8.85059178e-01 1.01949012e+00 1.10100424e+00
3.26818079e-01 4.82303768e-01 -3.26012850e-01 -3.16294581e-01
1.57181263e-01 5.92773080e-01 3.79732072e-01 -1.24545944e+00
-5.54827571e-01 -7.31355667e-01 -4.28861707e-01 -7.60257483e-01
-1.19534321e-01 -1.11405540e+00 -1.72619686e-01 -1.02357507e+00
4.74121794e-02 -3.84610891e-01 -8.30564559e-01 4.37767178e-01
3.00085563e-02 5.81130028e-01 3.48500252e-01 2.40233615e-01
-4.00019586e-01 5.13936281e-01 1.27016079e+00 -1.52931213e-01
-5.14240146e-01 1.38787121e-01 -6.34962261e-01 9.09244955e-01
9.78417456e-01 -3.33439291e-01 -4.16386396e-01 -2.77989566e-01
2.11952180e-01 5.36090992e-02 1.80668905e-01 -1.53331518e+00
2.07946464e-01 -2.45391913e-02 5.01026273e-01 -8.14554513e-01
4.46897686e-01 -5.21873534e-01 3.16666603e-01 2.77655631e-01
-5.02559543e-01 -3.29543173e-01 2.73150474e-01 8.59636068e-02
-7.91200936e-01 -4.74744737e-02 3.91636223e-01 2.65263114e-02
-6.69177532e-01 3.92290158e-03 -3.51390958e-01 -1.36920601e-01
5.95678568e-01 -6.67828321e-02 -2.47195318e-01 -2.28449106e-01
-1.04064274e+00 -1.09464273e-01 -1.67560428e-01 6.58612490e-01
4.17727411e-01 -1.83194637e+00 -5.99222004e-01 3.24864149e-01
1.01866819e-01 -3.36195052e-01 3.89352709e-01 5.91041684e-01
-4.08485383e-01 2.09335163e-01 -2.65707880e-01 -2.87913740e-01
-1.29188514e+00 3.43892783e-01 2.81604469e-01 -1.22789346e-01
-7.83918440e-01 7.22233236e-01 -2.17701584e-01 -2.59960771e-01
4.50810075e-01 -6.60468280e-01 -6.24678552e-01 4.27697659e-01
6.21094108e-01 3.80416095e-01 1.68625321e-02 -6.41365886e-01
-4.00621504e-01 8.04494321e-01 2.31596053e-01 -3.22443813e-01
1.33230841e+00 -7.04716817e-02 -1.55633196e-01 1.10412490e+00
1.14569247e+00 1.73375428e-01 -8.60390186e-01 6.90026069e-03
4.11032401e-02 -3.11212003e-01 1.93762898e-01 -1.00308788e+00
-1.17631555e+00 9.57381785e-01 5.83261430e-01 4.77000207e-01
1.60105741e+00 -4.09597039e-01 1.18388498e+00 6.44747317e-02
1.81306854e-01 -1.43795788e+00 -1.25221655e-01 9.42704439e-01
8.72127473e-01 -4.59031314e-01 -3.75399739e-01 -1.90080836e-01
-3.96708399e-01 1.43521464e+00 3.22845399e-01 -4.59644347e-01
7.99939513e-01 3.27403516e-01 1.00705169e-01 -1.97489843e-01
-6.59958541e-01 -3.14291358e-01 5.37806571e-01 -2.14137640e-02
6.84583724e-01 2.36129850e-01 -6.74789429e-01 1.64836729e+00
-1.04230595e+00 2.22841337e-01 2.36767203e-01 7.35222459e-01
-3.51200700e-01 -1.09717000e+00 -4.10875708e-01 3.11722428e-01
-7.93282747e-01 -1.37380660e-01 -2.85985947e-01 5.60026050e-01
7.45544672e-01 1.19223773e+00 2.38417372e-01 -1.06982136e+00
3.79417211e-01 2.91100562e-01 4.94589686e-01 -3.77175748e-01
-1.13532293e+00 6.22311890e-01 1.08558573e-01 -5.93586683e-01
-5.76984346e-01 -2.90637612e-01 -1.45253789e+00 -2.82593846e-01
-3.45387429e-01 4.01488990e-01 6.24189675e-01 7.94606686e-01
2.63249159e-01 1.23234403e+00 8.67232621e-01 -6.48661971e-01
-3.51599783e-01 -1.34056568e+00 -1.01383245e+00 5.78153849e-01
2.28375748e-01 -4.80089217e-01 -2.79330611e-01 2.01321423e-01] | [15.775984764099121, 5.255076885223389] |
b7e68e2c-be71-4c74-a05b-cb5ae2f8038e | duluth-at-semeval-2017-task-6-language-models | 1704.08390 | null | http://arxiv.org/abs/1704.08390v1 | http://arxiv.org/pdf/1704.08390v1.pdf | Duluth at SemEval-2017 Task 6: Language Models in Humor Detection | This paper describes the Duluth system that participated in SemEval-2017 Task
6 #HashtagWars: Learning a Sense of Humor. The system participated in Subtasks
A and B using N-gram language models, ranking highly in the task evaluation.
This paper discusses the results of our system in the development and
evaluation stages and from two post-evaluation runs. | ['Ted Pedersen', 'Xinru Yan'] | 2017-04-27 | duluth-at-semeval-2017-task-6-language-models-1 | https://aclanthology.org/S17-2064 | https://aclanthology.org/S17-2064.pdf | semeval-2017-8 | ['humor-detection'] | ['natural-language-processing'] | [-5.90452015e-01 2.00824440e-01 7.20285699e-02 -4.06182945e-01
-8.32138777e-01 -5.62856972e-01 8.59983563e-01 3.67494643e-01
-8.14584494e-01 4.34591919e-01 1.11963773e+00 -3.24874371e-01
4.16578919e-01 -3.75375569e-01 -2.17841849e-01 -8.41157734e-02
-3.01035464e-01 6.29771173e-01 3.68064940e-01 -9.56141710e-01
6.59954011e-01 -1.62883714e-01 -8.17105353e-01 1.11826265e+00
1.53210819e-01 2.65108854e-01 -3.90566885e-01 1.06385601e+00
2.70698786e-01 1.94033611e+00 -7.34145045e-01 -1.02718163e+00
-8.58517811e-02 -3.46940458e-01 -1.34905982e+00 -4.43034679e-01
4.57006574e-01 -1.77039176e-01 -3.72557640e-01 6.07117772e-01
6.46310210e-01 1.96958363e-01 2.71865726e-01 -9.87571001e-01
-6.19089484e-01 1.38641620e+00 -1.92571670e-01 2.84862280e-01
4.90063399e-01 2.18390778e-01 1.57884836e+00 -1.52080500e+00
4.48488504e-01 1.16930187e+00 9.00570571e-01 6.67495072e-01
-8.73981237e-01 -6.10372365e-01 -8.36064219e-01 3.55605662e-01
-1.12467110e+00 -4.46549147e-01 6.96198523e-01 -7.69028723e-01
1.45754445e+00 2.43967727e-01 3.41425210e-01 1.15922105e+00
3.91192250e-02 1.40129840e+00 1.27210009e+00 -2.88971752e-01
3.30504104e-02 3.92928839e-01 7.59849727e-01 5.25792897e-01
6.21885993e-02 -1.55084610e-01 -9.24211383e-01 -4.14236516e-01
-1.73702091e-01 -7.46124625e-01 -2.51943413e-02 5.06731033e-01
-9.02895033e-01 1.44411623e+00 4.04653609e-01 6.90158010e-01
-3.87848586e-01 -8.55468810e-02 9.52003181e-01 3.83622944e-01
3.57044131e-01 9.83365119e-01 -2.10769773e-01 -5.00560343e-01
-1.08704150e+00 7.14010477e-01 1.56796861e+00 6.27389371e-01
1.48905218e-01 4.19838075e-03 -2.67738044e-01 7.80492544e-01
1.06549613e-01 3.59949529e-01 6.57860637e-01 -7.00119436e-01
3.47433060e-01 2.80607283e-01 2.55642384e-01 -9.28271592e-01
-7.83631563e-01 -4.84718561e-01 -2.84401178e-01 -1.78405270e-01
2.06479967e-01 -2.02503502e-01 -3.44977647e-01 1.34011745e+00
-3.81033629e-01 -1.98523581e-01 1.78806782e-01 7.86324799e-01
1.23467004e+00 5.65721691e-01 4.51589525e-01 -1.38941854e-01
1.13254559e+00 -1.31498456e+00 -7.91630507e-01 -3.54922205e-01
1.14287984e+00 -1.18235564e+00 1.31597817e+00 6.35854721e-01
-1.32037818e+00 -2.91326433e-01 -1.22159100e+00 -4.50429648e-01
-4.07113910e-01 -1.44990608e-01 1.34671032e-01 5.08112669e-01
-1.04122925e+00 2.81056851e-01 -2.42532954e-01 -2.44982138e-01
-3.99864689e-02 -1.75156116e-01 -1.34262830e-01 3.12861502e-01
-1.82522941e+00 1.49241960e+00 4.83557075e-01 -3.04059863e-01
-8.78929615e-01 -4.22936320e-01 -8.71095359e-01 -2.37879738e-01
-2.91169465e-01 -1.74927235e-01 1.92244458e+00 -5.17281532e-01
-1.05910575e+00 1.33078957e+00 1.26123518e-01 -8.80242407e-01
3.31948459e-01 -6.82916939e-01 -4.29341108e-01 -3.10582906e-01
1.98815733e-01 -3.64951380e-02 2.96598643e-01 -9.70781684e-01
-4.79465097e-01 -7.91844446e-03 1.45783365e-01 1.28267303e-01
-4.30973411e-01 6.38694763e-01 2.45571882e-01 -4.73622620e-01
-5.42371154e-01 -8.20406556e-01 -1.80078321e-03 -1.49927711e+00
-2.28627339e-01 -5.63137650e-01 5.29921830e-01 -9.65224147e-01
1.88177288e+00 -2.03715420e+00 -1.22146159e-01 -8.03496540e-02
3.22063297e-01 3.02399933e-01 1.21187316e-02 8.83544624e-01
8.99038166e-02 2.04040915e-01 1.44444108e-01 -5.33946157e-01
2.83593655e-01 -3.17302905e-02 -6.47552431e-01 2.18279004e-01
-4.51658815e-01 9.95439291e-01 -1.15159404e+00 -3.74554455e-01
-2.21553698e-01 8.07880461e-02 -2.79180050e-01 4.97298956e-01
-2.20590048e-02 -1.99744627e-01 -9.90199391e-03 2.37373903e-01
2.15505198e-01 -1.74660757e-01 3.07990666e-02 1.52562693e-01
-2.67823845e-01 1.09434843e+00 -1.93084389e-01 1.28956282e+00
-5.58928072e-01 8.39074552e-01 -3.46005827e-01 -9.62138474e-02
6.58059061e-01 4.60429788e-01 1.66818112e-01 -8.02708030e-01
1.93545967e-01 4.57199454e-01 9.20905769e-02 -6.60686910e-01
1.04997981e+00 -7.39307046e-01 -8.03832114e-01 7.60013640e-01
1.86552197e-01 -3.20593536e-01 3.78765017e-01 7.25878239e-01
1.12911713e+00 -4.19932127e-01 6.66746795e-01 -5.33965468e-01
7.50301898e-01 1.74054399e-01 -3.10025457e-02 8.26115370e-01
-2.63799250e-01 3.78577679e-01 5.02904296e-01 -9.05996919e-01
-1.07065463e+00 -5.99503458e-01 1.05068937e-01 1.58066547e+00
-4.52265739e-01 -1.10769176e+00 -4.88265187e-01 -7.52511442e-01
-2.39892021e-01 1.55025470e+00 -8.50998044e-01 7.97454268e-02
-7.28940785e-01 -5.75351059e-01 1.22435236e+00 4.18708891e-01
3.24944586e-01 -1.39220393e+00 -7.32694626e-01 3.28942239e-01
-4.27625597e-01 -1.18204129e+00 -2.68213362e-01 1.33563221e-01
-3.88972044e-01 -9.48398173e-01 -8.48979205e-02 -6.67702973e-01
-2.60586828e-01 1.49712160e-01 1.63642812e+00 4.23487484e-01
2.13720694e-01 -1.98461786e-02 -6.67497098e-01 -4.74555582e-01
-5.13504207e-01 2.48930305e-01 -1.28591135e-01 -6.99887276e-01
1.11639857e+00 -4.49522197e-01 -3.18447471e-01 5.14425896e-02
-6.59283638e-01 1.93980321e-01 3.03047985e-01 9.13199067e-01
-4.78539318e-01 -3.32086921e-01 3.95310372e-01 -1.28468180e+00
1.04491925e+00 -8.01459670e-01 5.20579843e-03 -8.26942921e-02
-4.88041580e-01 -2.80080795e-01 5.82571745e-01 2.29491964e-01
-7.70121992e-01 -4.22558337e-01 -5.15601695e-01 2.55985200e-01
3.46666545e-01 8.59291553e-01 3.55815500e-01 3.32680196e-01
1.31716168e+00 -7.43905380e-02 -4.74869907e-01 -3.37613195e-01
4.02893096e-01 9.37215447e-01 5.46409428e-01 -5.26769340e-01
8.08517218e-01 -1.45129219e-01 -5.44502020e-01 -9.52945650e-01
-1.47146809e+00 -1.10075951e+00 -4.56484109e-01 -3.39635342e-01
6.70214057e-01 -1.15394437e+00 -5.56316435e-01 4.81696546e-01
-1.46180940e+00 -4.98526692e-01 -6.00006655e-02 2.56332278e-01
-3.68840575e-01 -7.80463368e-02 -1.20243514e+00 -8.13492358e-01
-8.62746179e-01 -4.43064690e-01 5.00094175e-01 5.90118347e-03
-9.67670381e-01 -1.20158529e+00 8.32683444e-01 6.93262994e-01
5.52834928e-01 4.82384153e-02 6.71061158e-01 -1.36827993e+00
3.05883199e-01 -5.42893767e-01 1.39678851e-01 3.21743429e-01
-8.10808301e-01 -4.41029936e-01 -1.17336619e+00 -1.72018901e-01
-5.58060557e-02 -1.27363610e+00 9.24364746e-01 -2.70207256e-01
5.74613571e-01 -8.03391099e-01 3.89789730e-01 -1.11230418e-01
1.09940958e+00 -4.45287436e-01 6.54587567e-01 6.25626087e-01
5.19332170e-01 7.49794900e-01 4.31108952e-01 7.09268034e-01
7.35883236e-01 4.68783200e-01 1.06222671e-03 3.40862125e-01
-5.37480190e-02 -6.88598335e-01 9.43068862e-01 1.27347124e+00
2.34330118e-01 -2.40657136e-01 -1.34314597e+00 7.26045370e-01
-1.94934893e+00 -1.32092643e+00 -7.61258423e-01 1.81938791e+00
1.05584478e+00 3.40959102e-01 6.13025904e-01 -1.84872195e-01
4.50606178e-03 5.29754460e-01 3.84588808e-01 -9.03211713e-01
-3.27525139e-01 4.21505958e-01 9.59233642e-02 1.28776324e+00
-7.67536283e-01 1.25677395e+00 7.13417053e+00 6.47252381e-01
-7.32523680e-01 6.48206651e-01 2.06135422e-01 -1.90052912e-01
-5.61788380e-01 1.75457850e-01 -6.60891652e-01 1.35719955e-01
1.40618527e+00 -5.01230240e-01 1.90181449e-01 8.39045644e-01
-2.74362788e-02 -1.05378285e-01 -9.96507347e-01 5.39631307e-01
7.36930728e-01 -9.64678705e-01 -7.67023191e-02 -4.21241105e-01
6.21599078e-01 7.14195251e-01 -1.80479601e-01 1.08226311e+00
6.30776942e-01 -1.26669812e+00 9.66616750e-01 3.22531983e-02
3.33978117e-01 -6.76414073e-01 1.25098526e+00 8.38656843e-01
-3.54551107e-01 -6.74825208e-03 -2.08967507e-01 -5.94561815e-01
2.52502441e-01 3.36380541e-01 -1.34423518e+00 -1.87493026e-01
3.61177981e-01 5.67412615e-01 -9.78068352e-01 6.22099817e-01
-8.28445017e-01 1.35210097e+00 1.52815267e-01 -5.51970899e-01
3.44425201e-01 5.41461170e-01 7.41544664e-01 2.21586800e+00
-2.61861682e-01 1.98865786e-01 3.86848152e-01 4.91680145e-01
-3.92708212e-01 3.19204271e-01 -4.00602162e-01 -1.49916887e-01
2.87247151e-01 1.46241784e+00 -6.04972728e-02 -4.26711828e-01
-2.78648853e-01 6.92326069e-01 4.43341136e-01 -5.97193576e-02
-5.65524578e-01 -3.60722393e-01 -9.86970291e-02 4.40919936e-01
-3.56920451e-01 -2.67869592e-01 -6.36797428e-01 -1.09540582e+00
-4.47729915e-01 -1.08749568e+00 8.41609418e-01 -7.80449808e-01
-1.40120852e+00 8.78356755e-01 -5.14819212e-02 -4.81445611e-01
-4.21577066e-01 -5.64642549e-01 -8.27500999e-01 7.69392312e-01
-9.93651569e-01 -1.50689638e+00 -4.74912999e-03 2.57808477e-01
5.95752478e-01 -3.03610355e-01 8.85088444e-01 9.71212685e-02
8.83035436e-02 7.25373089e-01 -4.09385026e-01 3.95026654e-01
9.97982383e-01 -1.62836838e+00 5.14413834e-01 5.52833855e-01
1.14563040e-01 9.10568237e-01 1.34151745e+00 -5.87559760e-01
-7.48725474e-01 -5.44701993e-01 2.24748254e+00 -1.00062561e+00
1.47578335e+00 -2.97118902e-01 -9.20644760e-01 7.95942605e-01
9.35732067e-01 -6.41506135e-01 1.09955013e+00 6.44847572e-01
-9.44412231e-01 5.32978892e-01 -6.85317099e-01 2.22465843e-01
3.91782939e-01 -9.69328463e-01 -1.21095109e+00 4.80976552e-01
5.98922074e-01 -4.10531938e-01 -5.02178073e-01 3.87883574e-01
6.20906293e-01 -8.83472383e-01 3.66676271e-01 -1.06491590e+00
1.29263902e+00 -1.08274065e-01 -2.44718730e-01 -1.17744422e+00
-4.77623969e-01 -8.70912731e-01 -1.70106724e-01 8.29300702e-01
6.54416203e-01 1.54313684e-01 3.55040908e-01 3.76865596e-01
-5.73745966e-02 -4.93634254e-01 -6.94490194e-01 -2.80503660e-01
7.75219381e-01 -4.95849311e-01 -1.30411670e-01 1.18965805e+00
8.70586812e-01 1.23374164e+00 -5.86299181e-01 -4.37125474e-01
4.47204739e-01 -2.38604948e-01 9.28685665e-01 -8.19848001e-01
-5.23260713e-01 -6.48114800e-01 -2.01725870e-01 -7.45899916e-01
3.73828471e-01 -1.07761610e+00 1.52567118e-01 -9.77932632e-01
8.97942781e-01 4.79077965e-01 -3.19722414e-01 6.35388672e-01
-2.70217210e-01 5.14947355e-01 3.05454999e-01 2.68185824e-01
-1.14158058e+00 2.20384181e-01 4.33146089e-01 -4.92121046e-03
8.66078362e-02 -1.45130754e-01 -8.13138187e-01 7.66346931e-01
9.60891724e-01 -4.78246033e-01 4.94731367e-02 -3.20798516e-01
8.04219306e-01 -5.55231795e-02 4.07275349e-01 -6.05188072e-01
2.95760274e-01 2.00985521e-01 -1.44842237e-01 -6.42304718e-01
-8.79326537e-02 -2.15379789e-01 -3.22352827e-01 4.74607676e-01
-8.64368320e-01 3.62501323e-01 2.27729976e-01 -2.65946239e-01
-4.38187033e-01 -5.18341362e-01 9.09302473e-01 -2.99577355e-01
-5.36778867e-01 -3.23984891e-01 -6.02286279e-01 4.63995695e-01
4.31968480e-01 4.13625389e-01 -5.51778197e-01 -7.38365471e-01
-5.49768209e-01 3.85515571e-01 2.60291845e-01 5.29594719e-01
4.27903503e-01 -1.33636534e+00 -1.45367432e+00 -4.60187942e-01
4.56429332e-01 -8.97951603e-01 9.03285667e-02 1.03062522e+00
-5.88425100e-01 4.16219532e-01 4.77384701e-02 1.02196457e-02
-1.39695084e+00 8.76208469e-02 2.08751753e-01 -7.95769036e-01
-2.97304571e-01 9.41081345e-01 -1.97161019e-01 -4.68284160e-01
1.50781486e-03 9.02255595e-01 -5.92817426e-01 1.52865767e-01
1.09179831e+00 5.00272512e-01 1.13525704e-01 -1.02884102e+00
-1.28044605e-01 -3.54788780e-01 -4.19692039e-01 -7.52225101e-01
1.25252235e+00 3.73182036e-02 -5.30808628e-01 1.02599251e+00
1.39627588e+00 4.58207011e-01 9.34252236e-03 -4.15659010e-01
7.82510340e-01 -2.81357020e-01 1.14716060e-01 -1.33720410e+00
-1.02548137e-01 1.03380692e+00 -1.04009829e-01 3.17286462e-01
4.76381093e-01 -8.16603601e-02 1.16609454e+00 6.83984280e-01
3.98871064e-01 -1.40918469e+00 2.96931624e-01 1.69437325e+00
1.18412519e+00 -9.89057362e-01 -3.51845915e-03 3.69148761e-01
-1.10460067e+00 1.27101815e+00 8.50360811e-01 -1.91001669e-01
4.71239746e-01 2.45706961e-01 3.77887249e-01 -7.04204023e-01
-1.07172096e+00 7.17064515e-02 4.78039384e-01 -2.29520380e-01
1.40219438e+00 3.65335494e-01 -1.12457657e+00 1.04662752e+00
-8.50646496e-01 -2.23863527e-01 7.33897924e-01 7.73951590e-01
-8.71369183e-01 -5.94208777e-01 5.15098758e-02 -2.26261169e-01
-8.30436468e-01 -4.34201121e-01 -1.26261401e+00 8.85200977e-01
-1.80019423e-01 1.06565571e+00 -5.71312726e-01 -1.12840819e+00
3.63920867e-01 2.53503740e-01 1.53378323e-01 -6.43800378e-01
-1.35568619e+00 -2.01819405e-01 7.96705484e-01 -4.03800666e-01
-1.60255224e-01 -6.82885885e-01 -1.33034050e+00 -8.11478257e-01
-2.24961564e-01 8.83351684e-01 1.86916828e-01 9.53211427e-01
-4.54496026e-01 -1.64039567e-01 8.27607691e-01 -1.61598846e-01
-8.71489465e-01 -1.56508768e+00 -4.61029947e-01 7.86627650e-01
1.69324666e-01 1.64074108e-01 -6.76744282e-01 -3.68058844e-03] | [8.86762523651123, 11.057747840881348] |
8ebb399e-601b-404f-b065-b13139b62115 | ts-net-combining-modality-specific-and-common | 1806.01550 | null | http://arxiv.org/abs/1806.01550v1 | http://arxiv.org/pdf/1806.01550v1.pdf | TS-Net: Combining modality specific and common features for multimodal patch matching | Multimodal patch matching addresses the problem of finding the
correspondences between image patches from two different modalities, e.g. RGB
vs sketch or RGB vs near-infrared. The comparison of patches of different
modalities can be done by discovering the information common to both modalities
(Siamese like approaches) or the modality-specific information (Pseudo-Siamese
like approaches). We observed that none of these two scenarios is optimal. This
motivates us to propose a three-stream architecture, dubbed as TS-Net,
combining the benefits of the two. In addition, we show that adding extra
constraints in the intermediate layers of such networks further boosts the
performance. Experimentations on three multimodal datasets show significant
performance gains in comparison with Siamese and Pseudo-Siamese networks. | ['Frédéric Jurie', 'Sovann En', 'Alexis Lechervy'] | 2018-06-05 | null | null | null | null | ['multimodal-patch-matching', 'patch-matching'] | ['computer-vision', 'computer-vision'] | [ 2.98397928e-01 -1.74252614e-01 -1.42325372e-01 -3.78636688e-01
-8.90141189e-01 -7.05286086e-01 7.88025320e-01 8.42013657e-02
-4.71942306e-01 4.04313564e-01 1.17252372e-01 7.91289359e-02
-3.03709507e-01 -6.24205887e-01 -8.31296265e-01 -5.12853026e-01
1.09141685e-01 2.20259681e-01 4.80430007e-01 -2.37242118e-01
2.58397102e-01 6.54831588e-01 -1.66121686e+00 5.32680571e-01
2.96209395e-01 1.23906553e+00 -1.04449891e-01 4.57129836e-01
-3.50260347e-01 3.29418033e-01 2.22118720e-01 -5.00689924e-01
6.04021728e-01 -4.14992332e-01 -7.10065186e-01 -2.71948650e-02
8.01679552e-01 3.82088423e-02 -3.03745627e-01 9.69598234e-01
3.29167187e-01 1.59760341e-01 1.97696745e-01 -1.56438553e+00
-3.01992983e-01 3.19517195e-01 -7.17627525e-01 -1.86724320e-01
7.53159642e-01 -1.42294914e-01 1.08284092e+00 -9.99687314e-01
8.06415677e-01 1.19184399e+00 7.49386370e-01 4.46485847e-01
-1.23944414e+00 -3.16396654e-01 1.03631943e-01 1.20370321e-01
-1.45158648e+00 -5.80172420e-01 1.04159200e+00 2.49395594e-02
6.22077405e-01 3.17413539e-01 4.96709347e-01 1.01991928e+00
-4.69429910e-01 1.00776172e+00 1.31232166e+00 -3.18732679e-01
1.70451641e-01 1.53295562e-01 -2.28114292e-01 6.23857558e-01
5.32660410e-02 1.24988429e-01 -1.13563466e+00 -3.41374308e-01
6.92873776e-01 3.29816967e-01 -3.15416962e-01 -8.48298490e-01
-1.49356723e+00 2.93096572e-01 6.59579217e-01 4.13702220e-01
-4.02608931e-01 1.68847963e-01 2.38438204e-01 5.62446356e-01
7.79188052e-02 3.31495345e-01 -3.23581487e-01 -3.37366685e-02
-1.03188574e+00 -9.68399923e-03 5.56666791e-01 9.94504035e-01
1.06949687e+00 -5.48756599e-01 2.83446431e-01 7.06931651e-01
4.62302774e-01 3.29368502e-01 3.62270959e-02 -9.62047935e-01
7.03265607e-01 7.95644879e-01 9.28342268e-02 -9.38832164e-01
-2.34596819e-01 -1.16021167e-02 -6.58201396e-01 -2.72419825e-02
7.32080102e-01 1.47167556e-02 -9.75329936e-01 1.67422545e+00
3.50448936e-01 2.81413794e-01 1.77309528e-01 1.20298016e+00
8.58664989e-01 3.81058276e-01 -6.17578812e-02 3.13433349e-01
1.25048792e+00 -8.31453383e-01 -3.03682178e-01 -1.66149810e-01
2.04714075e-01 -8.48206580e-01 1.00602746e+00 1.23983443e-01
-1.09226990e+00 -3.54577988e-01 -1.05742145e+00 -2.13444922e-02
-7.42354453e-01 2.29292409e-03 4.10102338e-01 4.77003872e-01
-1.03434336e+00 5.85624874e-01 -8.14049661e-01 -8.53029549e-01
3.44431132e-01 4.69771445e-01 -7.54559398e-01 -4.45249200e-01
-8.42560887e-01 5.84228575e-01 2.54141867e-01 8.60679001e-02
-3.68072778e-01 -5.71104646e-01 -7.81273067e-01 1.84893645e-02
4.86747533e-01 -6.78120375e-01 7.56502688e-01 -1.32371724e+00
-1.32199454e+00 9.14355993e-01 -3.12404484e-01 -9.41961482e-02
5.77757180e-01 1.72730852e-02 -2.99584687e-01 5.31770527e-01
-2.62972653e-01 1.00194871e+00 7.38486528e-01 -1.38497794e+00
-7.79817045e-01 -4.98431027e-01 4.02030855e-01 2.88520604e-01
-3.06130290e-01 7.01735681e-03 -7.73110449e-01 -3.11483800e-01
6.29833877e-01 -1.13399839e+00 1.10121630e-01 4.81504261e-01
-4.34530377e-01 -1.88070565e-01 8.00590336e-01 -3.76554221e-01
4.93529916e-01 -2.22694349e+00 2.92363942e-01 7.35435665e-01
4.14296761e-02 -6.57008588e-02 -4.86825615e-01 7.55124450e-01
-1.19747050e-01 -1.17776588e-01 -1.80216119e-01 -4.63463038e-01
8.42678845e-02 2.82703757e-01 4.15576473e-02 6.17115438e-01
1.98202446e-01 8.57784927e-01 -7.96211779e-01 -5.15536606e-01
7.97628239e-03 5.12940228e-01 -8.50589499e-02 4.22608890e-02
-3.42844352e-02 4.28016275e-01 -8.72108191e-02 1.00290763e+00
8.10501575e-01 -3.06860059e-01 2.60604501e-01 -5.59075415e-01
1.46547835e-02 -4.64936867e-02 -1.48367250e+00 2.21129727e+00
-4.53345358e-01 5.25710642e-01 1.61845371e-01 -7.43311226e-01
6.77736342e-01 3.67568761e-01 4.56496596e-01 -9.21972990e-01
-5.85869998e-02 4.24494296e-01 -2.15659499e-01 -3.70768219e-01
5.68201363e-01 -1.83064789e-01 1.26473784e-01 5.48667133e-01
1.65234521e-01 3.46181810e-01 2.28357688e-01 2.15887249e-01
7.96155930e-01 3.86391550e-01 -6.22372068e-02 3.97073254e-02
4.86464918e-01 2.75972821e-02 3.08692634e-01 7.91737616e-01
-6.22700974e-02 8.91964972e-01 4.00558770e-01 -3.12777191e-01
-7.03454435e-01 -1.12571275e+00 2.59458125e-01 1.02816081e+00
6.46285415e-01 -4.28372025e-01 -2.82192588e-01 -7.69852638e-01
7.83700719e-02 -6.69645565e-03 -7.07276165e-01 1.51583046e-01
-5.45422733e-01 -2.59055585e-01 5.61535597e-01 6.73387408e-01
5.88157594e-01 -5.06301224e-01 -8.29415381e-01 -3.81426722e-01
-2.48075366e-01 -1.19693625e+00 -3.00469875e-01 1.00060001e-01
-9.14593756e-01 -1.20187569e+00 -8.20116460e-01 -5.43946624e-01
7.38315940e-01 5.16863823e-01 8.27285349e-01 9.10242647e-02
1.23287551e-01 8.04196477e-01 -4.90059108e-01 1.92273527e-01
-7.11431773e-03 8.82431045e-02 -1.68843746e-01 4.77984637e-01
3.14712167e-01 -7.17848361e-01 -6.72721863e-01 4.07240182e-01
-1.19802380e+00 -9.81406271e-02 7.42948651e-01 5.76893091e-01
6.44022346e-01 -3.03059340e-01 8.14440846e-02 -5.40225148e-01
3.35275978e-01 -3.47595900e-01 -4.52845275e-01 7.36788273e-01
-2.95764714e-01 2.21311942e-01 5.58900118e-01 -5.49324155e-01
-8.34762394e-01 5.98127425e-01 2.32127517e-01 -6.59791052e-01
-4.42331493e-01 5.90050817e-01 -2.13234738e-01 -5.89066327e-01
2.89324522e-01 1.72561988e-01 5.46583198e-02 -5.54463565e-01
4.53725606e-01 3.60112876e-01 6.83142424e-01 -6.78437769e-01
9.62514520e-01 8.21286976e-01 2.27869883e-01 -6.68842435e-01
-3.73563409e-01 -6.98515594e-01 -6.08873546e-01 -3.13297480e-01
6.74911499e-01 -6.33745253e-01 -7.02090383e-01 2.29362637e-01
-9.78225291e-01 5.87193631e-02 -2.16044828e-01 3.87610346e-01
-4.66652572e-01 4.12517458e-01 -2.87811637e-01 -6.00456357e-01
2.16888580e-02 -1.05397415e+00 1.25668919e+00 4.57573444e-01
-6.92920107e-03 -9.04538631e-01 -7.29795266e-03 2.73525864e-01
5.11633456e-01 3.72766256e-01 6.45500004e-01 -5.83965957e-01
-8.04846883e-01 -3.94263595e-01 -5.41601419e-01 -1.49748191e-01
2.13077962e-02 -4.13663611e-02 -1.03578973e+00 -1.86427876e-01
-5.84530294e-01 -3.35764647e-01 8.37324083e-01 -1.30589589e-01
7.19036818e-01 8.99557322e-02 -2.33638868e-01 6.05093598e-01
1.73816633e+00 -1.03300489e-01 6.59355760e-01 4.00757819e-01
6.69073820e-01 7.33314753e-01 4.30625707e-01 1.16334639e-01
4.96808529e-01 8.29652727e-01 5.17329633e-01 -3.42335343e-01
-4.34266552e-02 -3.99482816e-01 1.91208735e-01 3.94626230e-01
-1.25991702e-01 -6.59016818e-02 -1.01116967e+00 6.16565228e-01
-2.04121208e+00 -8.47703815e-01 2.01697629e-02 2.36928892e+00
3.40619236e-01 -2.68653065e-01 3.01294655e-01 5.59638441e-02
6.29862368e-01 6.40786737e-02 -4.53333467e-01 -1.49930090e-01
-3.73240083e-01 9.25604850e-02 5.08682609e-01 1.75767794e-01
-1.12631559e+00 5.10625184e-01 5.60084772e+00 6.34131014e-01
-1.10069919e+00 1.24489538e-01 2.02996209e-01 -1.58614695e-01
-4.38726753e-01 3.12541157e-01 -2.43701160e-01 1.40617222e-01
4.87579852e-01 2.94741184e-01 5.41315377e-01 2.96391129e-01
-3.34836841e-01 -4.46629703e-01 -1.39183843e+00 1.16513872e+00
2.05701783e-01 -1.48462820e+00 -7.91502930e-03 -9.68358442e-02
6.16611421e-01 2.41926715e-01 -7.81228691e-02 -2.80389458e-01
1.48890698e-02 -6.73606098e-01 7.54422188e-01 7.69651651e-01
6.93750381e-01 -5.75598478e-01 7.67515063e-01 1.26965582e-01
-1.37810063e+00 1.97213948e-01 3.51008698e-02 4.43152994e-01
1.57112181e-01 3.64341103e-02 -3.53722453e-01 9.93113458e-01
8.18235099e-01 7.63389587e-01 -8.02182436e-01 1.27929652e+00
1.03126511e-01 7.81132355e-02 -7.78868020e-01 2.53314614e-01
3.32901537e-01 -1.21844649e-01 5.00915110e-01 1.00667918e+00
3.21082920e-01 -2.30113685e-01 4.68994444e-03 8.93231273e-01
-6.82581812e-02 -1.55890107e-01 -6.43497586e-01 4.52108644e-02
3.30832481e-01 1.33390152e+00 -8.67828190e-01 -1.95891246e-01
-8.13530803e-01 1.10923052e+00 5.27032204e-02 5.14695287e-01
-5.76326191e-01 -4.48190272e-01 4.34823185e-01 -1.27760749e-02
4.25164372e-01 -2.09130868e-01 -1.38501182e-01 -1.18947411e+00
1.55915514e-01 -7.06141770e-01 5.65928161e-01 -8.80656540e-01
-1.50299633e+00 4.26738501e-01 1.27260059e-01 -1.44225943e+00
2.08307341e-01 -5.97099423e-01 -4.43133205e-01 8.64523470e-01
-1.82571256e+00 -1.38572788e+00 -4.86486495e-01 1.08426857e+00
5.78361228e-02 1.68460220e-01 4.32970256e-01 3.95776927e-01
-3.01203668e-01 5.87252378e-01 -1.11557476e-01 3.92643586e-02
7.70625949e-01 -1.14656007e+00 7.82154500e-02 7.76158035e-01
4.64026421e-01 7.06685126e-01 3.77925813e-01 -2.55788058e-01
-1.79620683e+00 -6.27356112e-01 8.72029066e-01 -4.49964777e-02
5.62849283e-01 -1.51134208e-01 -7.95023859e-01 2.39237010e-01
3.67340684e-01 2.86947973e-02 7.18009830e-01 1.12901218e-01
-8.33842695e-01 -3.12458485e-01 -1.21383190e+00 5.80665588e-01
1.00270641e+00 -9.70938563e-01 -2.27525979e-01 1.36156186e-01
3.18453789e-01 -3.42459649e-01 -9.40966725e-01 3.51831406e-01
8.94902527e-01 -1.22042584e+00 1.06849253e+00 -6.02969766e-01
4.09770608e-01 -4.76423919e-01 -6.19751215e-01 -9.77016330e-01
2.55435109e-01 -4.20775086e-01 2.78644748e-02 1.10700035e+00
4.73222613e-01 -6.07304454e-01 9.44513500e-01 7.32899487e-01
2.85037100e-01 -5.73065937e-01 -1.04741991e+00 -8.36087525e-01
-2.37776428e-01 -4.09104884e-01 8.30895066e-01 1.01438725e+00
6.79015592e-02 4.85436134e-02 -3.08674991e-01 3.27189922e-01
4.05506074e-01 5.96696854e-01 8.35013330e-01 -1.05745959e+00
-1.26504496e-01 -4.52622265e-01 -5.55133164e-01 -1.01790810e+00
-7.27659166e-02 -9.15082872e-01 -2.83260755e-02 -1.43398619e+00
2.18008965e-01 -4.83872741e-01 -4.46399927e-01 7.46943116e-01
1.82894394e-01 6.01588368e-01 4.75032151e-01 2.79892772e-01
-7.69985497e-01 2.98478037e-01 9.18732285e-01 -8.93229395e-02
-9.20288861e-02 -2.02868983e-01 -5.12742460e-01 4.78989244e-01
9.16977644e-01 -3.86747062e-01 -2.96975732e-01 -5.89065075e-01
5.07674992e-01 2.41636306e-01 7.05463350e-01 -1.04160202e+00
6.92186654e-01 -1.65124521e-01 1.13661200e-01 -6.29011273e-01
6.18247390e-01 -1.31515110e+00 2.26638094e-01 -7.53676593e-02
-4.34389591e-01 2.70022899e-01 3.35579664e-02 7.84900248e-01
-6.10656500e-01 4.12808359e-02 5.57999074e-01 -1.21562280e-01
-7.88390756e-01 1.21911779e-01 1.09573528e-01 -1.14395849e-01
7.37907171e-01 -6.89041674e-01 -4.71832871e-01 -3.78574282e-01
-6.78664386e-01 1.31247804e-01 7.80302346e-01 4.92638141e-01
6.90358400e-01 -1.44814622e+00 -2.08412543e-01 1.68804958e-01
4.79219079e-01 -4.05887514e-01 1.02171116e-01 1.50372052e+00
-3.73662531e-01 2.64369905e-01 -3.06864947e-01 -7.43976474e-01
-1.33781397e+00 2.58962601e-01 4.47611034e-01 7.12937787e-02
-2.63497800e-01 7.01766968e-01 -2.33191624e-01 -5.83798468e-01
4.40020680e-01 -5.26911616e-02 7.39371181e-02 1.84426233e-01
1.09809130e-01 4.22475547e-01 6.62470832e-02 -8.91035974e-01
-7.67282844e-01 7.41780818e-01 2.33182132e-01 -4.06733513e-01
1.31727672e+00 -1.77332461e-01 -3.17360520e-01 4.36040461e-01
1.37980032e+00 -5.56574240e-02 -1.04379416e+00 -5.73947012e-01
1.41475648e-01 -6.08081996e-01 -1.30189002e-01 -9.14350808e-01
-1.39497042e+00 8.49574685e-01 7.35904336e-01 1.33331165e-01
1.43354023e+00 1.95181981e-01 7.86053419e-01 4.59355056e-01
3.71180117e-01 -9.86587942e-01 -1.78064778e-02 1.58467397e-01
6.38585031e-01 -1.34933496e+00 -7.04762414e-02 -3.22645187e-01
-5.26950121e-01 1.27950299e+00 4.47084159e-01 -2.82779392e-02
4.82834071e-01 -5.84815629e-02 1.33298665e-01 -4.32154626e-01
-5.51581025e-01 -5.89019477e-01 5.87087274e-01 4.64673132e-01
2.70730615e-01 -1.94102630e-01 -9.49273780e-02 -1.13821812e-01
4.21039522e-01 -2.51534749e-02 2.33345386e-02 1.32392585e+00
1.04065225e-01 -1.38349712e+00 -4.98677343e-01 3.13639849e-01
-2.20788121e-02 1.63813066e-02 -8.12241614e-01 8.95914853e-01
1.83956787e-01 9.27978337e-01 -1.67703684e-02 -4.12550837e-01
4.32914138e-01 1.87291384e-01 6.71618462e-01 7.51576340e-03
-7.44434476e-01 1.08268276e-01 4.76481952e-02 -8.96226943e-01
-1.30241573e+00 -6.82361305e-01 -8.59519064e-01 -2.33677760e-01
-1.83196768e-01 -8.67202058e-02 7.21335649e-01 8.14312994e-01
4.57678646e-01 -1.20843023e-01 4.74230677e-01 -1.00959396e+00
-6.16973080e-02 -4.06861246e-01 -3.97071540e-01 4.93999124e-01
4.33406770e-01 -5.54603755e-01 -2.30561942e-01 -5.51969856e-02] | [7.877450942993164, -2.072140693664551] |
71faa12a-8f5d-45df-8abf-73d8d051b2c7 | is-chatgpt-a-general-purpose-natural-language | 2302.06476 | null | https://arxiv.org/abs/2302.06476v2 | https://arxiv.org/pdf/2302.06476v2.pdf | Is ChatGPT a General-Purpose Natural Language Processing Task Solver? | Spurred by advancements in scale, large language models (LLMs) have demonstrated the ability to perform a variety of natural language processing (NLP) tasks zero-shot -- i.e., without adaptation on downstream data. Recently, the debut of ChatGPT has drawn a great deal of attention from the natural language processing (NLP) community due to the fact that it can generate high-quality responses to human input and self-correct previous mistakes based on subsequent conversations. However, it is not yet known whether ChatGPT can serve as a generalist model that can perform many NLP tasks zero-shot. In this work, we empirically analyze the zero-shot learning ability of ChatGPT by evaluating it on 20 popular NLP datasets covering 7 representative task categories. With extensive empirical studies, we demonstrate both the effectiveness and limitations of the current version of ChatGPT. We find that ChatGPT performs well on many tasks favoring reasoning capabilities (e.g., arithmetic reasoning) while it still faces challenges when solving specific tasks such as sequence tagging. We additionally provide in-depth analysis through qualitative case studies. | ['Diyi Yang', 'Michihiro Yasunaga', 'Jiaao Chen', 'Zhuosheng Zhang', 'Aston Zhang', 'Chengwei Qin'] | 2023-02-08 | null | null | null | null | ['arithmetic-reasoning'] | ['reasoning'] | [ 2.30594397e-01 4.61791128e-01 -9.65432227e-02 -4.17219549e-01
-1.16872871e+00 -5.01870096e-01 4.37007695e-01 4.32188243e-01
-4.62049186e-01 6.98840857e-01 6.89999580e-01 -6.49242103e-01
9.29882377e-03 -5.86269796e-01 -4.35716480e-01 -1.57994688e-01
2.52388805e-01 7.18466222e-01 1.34970054e-01 -5.34543455e-01
4.03741181e-01 -1.55737340e-01 -1.24819410e+00 8.00916433e-01
9.66989517e-01 5.43172121e-01 1.37987852e-01 6.95719779e-01
-4.20396537e-01 1.47644520e+00 -5.52824080e-01 -7.94058919e-01
-9.36961994e-02 -2.73393750e-01 -1.22211039e+00 -5.42684853e-01
1.69987768e-01 -1.56884193e-01 -1.31247908e-01 7.45595634e-01
6.81959927e-01 4.92105454e-01 3.97051245e-01 -1.31371307e+00
-8.92397225e-01 1.02272022e+00 -1.41753852e-01 2.25678131e-01
5.61463356e-01 5.70130348e-01 1.42290950e+00 -7.79171109e-01
7.59346366e-01 1.50042486e+00 8.03550601e-01 7.44507253e-01
-1.26663172e+00 -6.18923306e-01 -2.39666309e-02 1.90568358e-01
-8.52060318e-01 -5.52269936e-01 3.64000976e-01 -3.87874871e-01
1.56885827e+00 -1.90537367e-02 1.04152493e-01 1.36195731e+00
4.29200590e-01 1.08989668e+00 1.13959789e+00 -4.93595570e-01
3.57765794e-01 -7.18796030e-02 4.16026533e-01 5.77300906e-01
-2.54336923e-01 -3.86966765e-01 -7.68792033e-01 -2.33230561e-01
2.09933370e-01 -2.91126460e-01 -8.92246813e-02 1.97570175e-01
-1.18598378e+00 8.83819699e-01 6.26709610e-02 4.55952525e-01
-2.58844465e-01 -8.27330165e-03 5.13225436e-01 5.89510560e-01
6.51678801e-01 1.07276583e+00 -6.61129534e-01 -8.98605645e-01
-5.89947164e-01 3.08506399e-01 1.27494740e+00 8.97574604e-01
5.29618800e-01 -1.05345272e-01 -5.69172084e-01 9.94682729e-01
-1.30679384e-01 1.98780537e-01 7.38029897e-01 -1.18975413e+00
8.35576653e-01 6.70533419e-01 1.27156168e-01 -8.82698298e-01
-5.55763602e-01 3.39810774e-02 -5.55009544e-01 -2.77119845e-01
6.42246604e-01 -4.39083844e-01 -3.87431920e-01 1.93204117e+00
-4.40851972e-02 7.31415600e-02 3.52378488e-01 7.09007323e-01
7.60063231e-01 9.98521149e-01 5.11161983e-01 -1.15460992e-01
1.30811214e+00 -9.74373996e-01 -5.94842732e-01 -7.16601193e-01
1.20480967e+00 -7.19828069e-01 1.52774215e+00 3.02333325e-01
-1.28293526e+00 -4.06450599e-01 -4.77294594e-01 -3.68079692e-01
-1.53079316e-01 -3.51733387e-01 7.15465784e-01 3.87221426e-01
-1.03553748e+00 4.86901194e-01 -6.13220930e-01 -5.71633995e-01
1.58131078e-01 -1.06393740e-01 -2.62163013e-01 -3.75560999e-01
-1.52732778e+00 1.19704640e+00 1.65820837e-01 -3.37235928e-01
-5.88692486e-01 -1.01808393e+00 -8.42956066e-01 3.96980703e-01
5.84136069e-01 -6.17415905e-01 1.99829972e+00 -6.57099605e-01
-1.55284286e+00 7.36549973e-01 -3.39297712e-01 -5.66604197e-01
3.73899937e-01 -2.16006875e-01 -1.02756798e-01 5.77847138e-02
2.02899098e-01 6.59446776e-01 1.88165531e-01 -5.40748954e-01
-4.77997094e-01 -7.65760988e-02 2.53789604e-01 1.98913917e-01
-4.01865363e-01 4.13503855e-01 5.29483110e-02 -3.81635159e-01
-2.30535582e-01 -8.06311071e-01 -2.13694066e-01 -1.74291953e-01
-3.50393690e-02 -7.63873339e-01 2.98626453e-01 -7.77166545e-01
1.25542736e+00 -1.96892548e+00 -8.46130401e-02 -3.57955217e-01
1.73603252e-01 4.36906099e-01 -4.59198385e-01 9.00980175e-01
1.31232589e-01 3.58592182e-01 -1.52948558e-01 -4.06839758e-01
1.41206726e-01 2.09659308e-01 -4.51523215e-01 -1.61034781e-02
3.23414743e-01 1.43403816e+00 -1.24256337e+00 -4.73437190e-01
-7.99399316e-02 1.73772406e-02 -6.64119959e-01 2.89939016e-01
-5.33375740e-01 7.88842887e-02 -3.43078941e-01 4.35438454e-01
7.35671222e-02 -6.18698955e-01 2.17973784e-01 5.91302037e-01
-6.46601394e-02 7.21484065e-01 -5.22883952e-01 1.77481961e+00
-6.09484613e-01 6.78796232e-01 -9.64267403e-02 -9.08673704e-01
8.28256905e-01 5.89471400e-01 1.19861409e-01 -7.93188393e-01
7.30990171e-02 9.21217799e-02 4.35388029e-01 -8.32095563e-01
4.67248410e-01 -3.42167139e-01 -4.63942558e-01 1.06807268e+00
1.08992763e-01 -4.93394071e-03 2.34695464e-01 6.06479526e-01
1.45461893e+00 -1.64542928e-01 4.94141400e-01 -1.67326897e-01
1.48006305e-01 4.83881861e-01 4.49814916e-01 1.07112670e+00
-3.67894709e-01 3.69132608e-01 6.54965878e-01 -2.12553099e-01
-1.07263660e+00 -7.12186992e-01 3.59124184e-01 1.61883628e+00
-3.24523807e-01 -4.32743609e-01 -6.92939699e-01 -3.57355624e-01
-4.09216359e-02 1.12030315e+00 -2.91635156e-01 -3.21216017e-01
-4.77358878e-01 -5.20442724e-01 9.11702156e-01 6.70637310e-01
5.16937256e-01 -1.41344476e+00 -5.23229361e-01 4.32191372e-01
-7.20478654e-01 -1.31280148e+00 -1.89274833e-01 -4.92075123e-02
-6.46247447e-01 -8.70515883e-01 -4.11792994e-01 -8.58050346e-01
2.03106880e-01 4.70082581e-01 1.35008872e+00 6.24601245e-02
-7.00520352e-02 4.43664700e-01 -6.65808916e-01 -4.14919972e-01
-8.43068779e-01 2.56616890e-01 1.76187139e-02 -4.60026562e-01
7.62082875e-01 -4.66664732e-01 -8.82901438e-03 1.66765273e-01
-4.87988263e-01 2.64519095e-01 6.32213235e-01 6.88040018e-01
-2.22668901e-01 -1.98635176e-01 9.20536160e-01 -1.01613367e+00
1.33964598e+00 -7.68087566e-01 -7.19937906e-02 5.51020801e-01
-4.12107140e-01 -1.60821509e-02 5.83853424e-01 -5.06117165e-01
-1.42781627e+00 -5.97589731e-01 -1.90589443e-01 -1.34189904e-01
-1.40861407e-01 7.83327997e-01 1.90215245e-01 2.23461658e-01
9.22225893e-01 1.33669630e-01 -5.79675892e-04 -3.19717288e-01
3.38764369e-01 8.14824581e-01 4.24520195e-01 -9.35404956e-01
5.18548787e-01 -1.34024352e-01 -5.43904305e-01 -8.83184910e-01
-1.26672256e+00 -5.30724823e-01 -1.65284917e-01 -1.59523450e-02
7.96836257e-01 -9.80003536e-01 -9.02843297e-01 3.03224891e-01
-1.33608949e+00 -7.87795305e-01 5.18938992e-03 2.92577922e-01
-5.14330745e-01 2.73626000e-01 -1.21231568e+00 -1.01737332e+00
-4.89940792e-01 -7.68798232e-01 7.58780837e-01 1.45617545e-01
-1.02176404e+00 -1.06855166e+00 4.72081713e-02 6.73860550e-01
4.68267620e-01 -3.18888038e-01 1.34008491e+00 -1.08438516e+00
-2.85698235e-01 -1.03811212e-01 -1.78477570e-01 1.09255314e-01
-4.04023796e-01 -2.89798796e-01 -9.73208487e-01 6.97539523e-02
1.31585106e-01 -8.85410368e-01 6.08640194e-01 1.02856308e-01
7.53465593e-01 -4.43642139e-01 -9.62956622e-02 4.63012531e-02
1.04822016e+00 9.46112648e-02 5.04360974e-01 1.44494057e-01
2.59632200e-01 9.09242332e-01 7.38838673e-01 3.50773275e-01
4.99261349e-01 2.48668283e-01 -2.24349976e-01 6.06323779e-01
1.87168151e-01 -7.01188207e-01 3.42669517e-01 7.72943139e-01
1.82061002e-01 -5.40676236e-01 -1.45777786e+00 5.43832600e-01
-2.12166476e+00 -1.02211344e+00 1.17554486e-01 1.55209899e+00
1.01630998e+00 2.11314350e-01 -2.52232909e-01 -1.12542540e-01
4.92255360e-01 1.27765626e-01 -3.88225824e-01 -7.88735032e-01
1.05586544e-01 2.77370691e-01 -2.48705938e-01 4.20112282e-01
-5.91231585e-01 1.26128566e+00 6.16766548e+00 6.14028513e-01
-7.75556326e-01 3.03575933e-01 7.02819169e-01 6.98395297e-02
-2.75710464e-01 5.09493910e-02 -7.36925244e-01 3.14380378e-01
1.26495469e+00 -5.06064355e-01 3.23864907e-01 8.80418897e-01
2.48630449e-01 -1.77549973e-01 -1.12152660e+00 7.34911859e-01
7.85200149e-02 -1.31964612e+00 -6.66099489e-02 -3.75169069e-01
7.73596525e-01 4.99781482e-02 -2.41104946e-01 1.05078650e+00
7.39048004e-01 -9.60019112e-01 4.82009947e-01 1.98617682e-01
4.68833119e-01 -3.56765658e-01 7.03203082e-01 1.06800222e+00
-7.26595402e-01 -4.36638445e-01 -4.58689153e-01 -8.37626994e-01
4.20736939e-01 2.16399357e-01 -1.12607360e+00 4.90401201e-02
4.48633134e-01 3.76755297e-01 -2.14469180e-01 8.45753014e-01
-6.32464349e-01 8.01711440e-01 9.95305628e-02 -4.37326461e-01
4.32280302e-01 1.29651144e-01 2.55394548e-01 1.24686408e+00
2.00412676e-01 6.83810592e-01 2.90809274e-01 8.38821113e-01
-2.09093958e-01 1.64448708e-01 -5.67638218e-01 -6.45978332e-01
7.29947209e-01 1.13410866e+00 -4.38446045e-01 -5.48727751e-01
-5.59055984e-01 5.45591235e-01 7.27582097e-01 3.45984280e-01
-4.99966532e-01 -2.06788614e-01 7.43242383e-01 4.34306487e-02
-1.95516825e-01 -2.69472480e-01 -3.75636429e-01 -1.20783699e+00
-1.05992883e-01 -1.08423507e+00 4.31023300e-01 -1.25944328e+00
-1.57252467e+00 3.53241414e-01 -1.08310841e-01 -6.20793819e-01
-5.08965969e-01 -6.28491104e-01 -8.34897399e-01 9.49444890e-01
-1.11958778e+00 -7.63804555e-01 -2.06871573e-02 3.53985965e-01
8.75223994e-01 -7.89514109e-02 9.19452667e-01 -9.78327692e-02
-3.94601941e-01 3.61555785e-01 -3.15106869e-01 3.73090148e-01
9.03840721e-01 -9.13632870e-01 7.74823248e-01 6.59135997e-01
-3.21527608e-02 8.86417270e-01 7.24280357e-01 -7.75927842e-01
-1.36219990e+00 -8.51116538e-01 1.75200367e+00 -6.91655219e-01
1.06665397e+00 -4.12717551e-01 -1.08254015e+00 1.07026279e+00
2.38444149e-01 -4.52223212e-01 7.27103174e-01 4.54598933e-01
-4.53270644e-01 2.71066338e-01 -9.56270039e-01 7.46049166e-01
1.05023599e+00 -8.18505168e-01 -1.19166589e+00 3.64379555e-01
1.07097864e+00 -3.53353202e-01 -6.23375535e-01 1.81035753e-02
3.17153960e-01 -7.85523236e-01 7.22749710e-01 -1.04302859e+00
9.47342515e-01 4.45801049e-01 5.40403388e-02 -1.43619061e+00
-4.21957046e-01 -9.26448822e-01 8.33267570e-02 1.24046910e+00
5.91351807e-01 -5.86650848e-01 5.32932937e-01 1.24536574e+00
-2.91329354e-01 -6.84802115e-01 -6.01502895e-01 -6.07532084e-01
1.94166362e-01 -6.36787832e-01 2.64452398e-01 1.06614113e+00
8.22759926e-01 7.79680312e-01 -2.97910839e-01 -2.41957605e-01
1.29578114e-01 7.12534934e-02 7.53812611e-01 -1.36601412e+00
-4.37180519e-01 -3.73917997e-01 9.76403877e-02 -9.60113943e-01
5.52376091e-01 -1.02549899e+00 3.13566208e-01 -1.69204462e+00
3.05060714e-01 -1.36708617e-01 2.29233243e-02 7.84391761e-01
-3.42706710e-01 -2.11469263e-01 2.81637818e-01 1.70891896e-01
-8.33395779e-01 2.31461167e-01 1.22116208e+00 4.94064614e-02
-1.74496964e-01 -9.99345109e-02 -1.02170038e+00 6.48833454e-01
9.08569753e-01 -4.44186866e-01 -5.01410544e-01 -6.99824095e-01
5.89584529e-01 4.20348257e-01 6.19861074e-02 -9.00083542e-01
4.56133008e-01 -2.83746660e-01 -1.14483699e-01 6.52593225e-02
3.18586081e-01 -1.23821735e-01 -4.36779499e-01 3.22188705e-01
-8.57663453e-01 2.80127555e-01 1.77488625e-01 3.65604818e-01
-1.19600616e-01 -3.27188909e-01 6.03740394e-01 -5.60379326e-01
-1.01655281e+00 -6.93340451e-02 -6.64597154e-01 6.48374915e-01
8.93625438e-01 1.47120908e-01 -7.58966804e-01 -6.68120801e-01
-3.61145228e-01 4.17241842e-01 1.84611306e-01 6.10578656e-01
4.24088240e-01 -8.44159961e-01 -7.34079599e-01 -4.30214256e-02
2.46656686e-01 -1.78914323e-01 3.66693109e-01 1.02090394e+00
-1.75398216e-01 8.09584260e-01 -1.43418074e-01 -1.65356159e-01
-1.07859468e+00 6.08323276e-01 -8.94421414e-02 -4.48036313e-01
-8.15778852e-01 1.11868191e+00 7.52440691e-02 -5.49563468e-01
1.22886956e-01 -4.10866827e-01 6.67087734e-02 5.80103137e-02
7.95362592e-01 2.17912912e-01 -1.20070830e-01 -5.43305539e-02
-8.46264213e-02 -1.72287151e-01 -9.35995653e-02 -2.52928823e-01
1.21618176e+00 -4.71150726e-02 -9.04313475e-02 7.50332832e-01
8.90998781e-01 -3.53943020e-01 -9.78251398e-01 -5.00258982e-01
3.83046627e-01 -2.23226845e-01 -4.20283228e-01 -9.84676123e-01
-2.85561860e-01 1.11058354e+00 -4.85465378e-01 9.14797261e-02
5.13668418e-01 3.22114639e-02 1.06900644e+00 9.20571744e-01
6.25385821e-01 -1.24815834e+00 3.71148556e-01 1.11062431e+00
5.07955015e-01 -1.20831716e+00 -5.65462768e-01 -9.47545096e-02
-1.07620764e+00 9.16998684e-01 8.46348763e-01 1.65375188e-01
1.90429628e-01 2.64812022e-01 1.10677719e-01 3.53591181e-02
-1.55008197e+00 3.20291370e-02 -1.93594366e-01 3.81096303e-01
7.89398193e-01 2.92092445e-03 -3.31913173e-01 8.61012459e-01
-5.10926545e-01 2.05334008e-01 5.82820117e-01 9.70192432e-01
-7.60024548e-01 -7.98031092e-01 -1.22450531e-01 3.04639757e-01
-3.19645554e-01 -4.20665681e-01 -4.88701016e-01 6.37925923e-01
-4.11648721e-01 1.31583226e+00 1.07453324e-01 -2.44910166e-01
1.47413164e-01 8.07419896e-01 1.81789890e-01 -1.01130402e+00
-7.72113323e-01 -5.39972007e-01 4.80599910e-01 -6.19891047e-01
-2.98404451e-02 -6.18585527e-01 -1.14989078e+00 -7.21234083e-01
1.42408445e-01 3.66867036e-01 1.85027152e-01 1.20684028e+00
4.01453048e-01 2.99840391e-01 -6.13572784e-02 -3.68998110e-01
-1.03300273e+00 -1.22662663e+00 -2.03315765e-01 3.92411709e-01
-1.48379073e-01 -2.82211214e-01 -2.45266467e-01 -2.38148838e-01] | [11.195412635803223, 8.415201187133789] |
f187ea8b-4dbe-46f6-935c-1338b7720de4 | deep-network-embedding-for-graph | 1901.01718 | null | http://arxiv.org/abs/1901.01718v1 | http://arxiv.org/pdf/1901.01718v1.pdf | Deep Network Embedding for Graph Representation Learning in Signed Networks | Network embedding has attracted an increasing attention over the past few
years. As an effective approach to solve graph mining problems, network
embedding aims to learn a low-dimensional feature vector representation for
each node of a given network. The vast majority of existing network embedding
algorithms, however, are only designed for unsigned networks, and the signed
networks containing both positive and negative links, have pretty distinct
properties from the unsigned counterpart. In this paper, we propose a deep
network embedding model to learn the low-dimensional node vector
representations with structural balance preservation for the signed networks.
The model employs a semi-supervised stacked auto-encoder to reconstruct the
adjacency connections of a given signed network. As the adjacency connections
are overwhelmingly positive in the real-world signed networks, we impose a
larger penalty to make the auto-encoder focus more on reconstructing the scarce
negative links than the abundant positive links. In addition, to preserve the
structural balance property of signed networks, we design the pairwise
constraints to make the positively connected nodes much closer than the
negatively connected nodes in the embedding space. Based on the network
representations learned by the proposed model, we conduct link sign prediction
and community detection in signed networks. Extensive experimental results in
real-world datasets demonstrate the superiority of the proposed model over the
state-of-the-art network embedding algorithms for graph representation learning
in signed networks. | ['Fu-Lai Chung', 'Xiao Shen'] | 2019-01-07 | null | null | null | null | ['link-sign-prediction'] | ['graphs'] | [-1.76975261e-02 4.54278171e-01 -5.06193578e-01 -4.44661945e-01
5.20529747e-01 -3.78684640e-01 3.64537328e-01 7.87640437e-02
1.34851635e-01 4.89710987e-01 2.73227036e-01 -2.99176753e-01
-6.77608132e-01 -1.17238700e+00 -4.68045920e-01 -6.19973779e-01
-5.27529120e-01 4.73675460e-01 8.84652212e-02 -3.68029296e-01
6.31309068e-03 2.71983027e-01 -9.55329895e-01 -1.05697922e-01
7.70931005e-01 7.72305191e-01 -2.44460776e-01 2.59544462e-01
-2.88921982e-01 6.55336499e-01 -1.54281601e-01 -6.07891262e-01
2.55434453e-01 -4.20951575e-01 -3.67119461e-01 -2.58667190e-02
3.74429166e-01 -2.00463519e-01 -1.30245662e+00 1.30881488e+00
2.87590116e-01 -3.90468001e-01 6.25048518e-01 -1.75262451e+00
-1.04101884e+00 9.17038381e-01 -6.39526010e-01 3.59561406e-02
-3.90922790e-03 -8.72791708e-02 1.63967454e+00 -6.60360634e-01
6.97192729e-01 1.15501547e+00 5.71174622e-01 2.45879367e-01
-1.14439726e+00 -8.77196789e-01 2.47348160e-01 2.67225504e-01
-1.21155190e+00 -5.71592338e-02 1.49785185e+00 -4.40354794e-01
5.22148550e-01 1.00253105e-01 1.06064880e+00 6.92706764e-01
4.46344167e-03 5.19255579e-01 5.94543934e-01 -6.97008073e-02
-1.89989626e-01 -9.21665970e-03 1.87383652e-01 1.09488189e+00
8.76948178e-01 -3.52064222e-02 -3.46170336e-01 -1.44080132e-01
7.95906365e-01 4.32902277e-01 -4.85730737e-01 -1.15909398e+00
-1.38720453e+00 1.17894650e+00 1.30677700e+00 3.50184143e-01
-1.95981726e-01 2.85825133e-01 6.34516716e-01 4.04926956e-01
3.96096915e-01 5.16869910e-02 -2.17482343e-01 4.17357385e-01
-5.67877650e-01 -2.79223591e-01 7.92804778e-01 9.13236618e-01
7.49585211e-01 -1.80627666e-02 2.34985635e-01 6.50968611e-01
8.15881312e-01 2.85314143e-01 2.76481837e-01 -5.03432453e-01
7.10544407e-01 1.34184742e+00 -5.84648371e-01 -1.87654066e+00
-2.01928273e-01 -4.84623283e-01 -1.37561083e+00 -1.40289977e-01
-9.28304144e-06 8.74772891e-02 -5.43417513e-01 1.59322917e+00
1.32451028e-01 2.12546691e-01 -9.72732753e-02 8.09331417e-01
8.97875309e-01 6.60834372e-01 -4.86835182e-01 -1.55548215e-01
9.87367392e-01 -7.93759644e-01 -8.18103433e-01 -1.85140058e-01
4.65788513e-01 -2.61424303e-01 6.48278892e-01 -2.56857216e-01
-6.88976228e-01 -9.31134447e-02 -1.43297911e+00 1.25385538e-01
-3.11600417e-01 -2.88693942e-02 1.09297264e+00 2.41209760e-01
-7.71102965e-01 6.09212995e-01 -5.20910263e-01 -2.74836212e-01
6.03171408e-01 5.26767075e-01 -7.48976290e-01 -2.99740523e-01
-1.35645163e+00 2.05765799e-01 3.66325676e-01 5.16219318e-01
-1.75424680e-01 -2.74794579e-01 -1.12356150e+00 3.79604191e-01
2.76855737e-01 -4.07637030e-01 2.44357944e-01 -8.87697458e-01
-9.05789793e-01 7.43561924e-01 8.53098631e-02 -7.96562359e-02
3.77480447e-01 4.88520026e-01 -6.83097303e-01 2.87683576e-01
-4.31557782e-02 4.38604444e-01 6.72155499e-01 -1.29337132e+00
6.78782463e-02 -2.74501354e-01 1.00813605e-01 -1.20122373e-01
-1.05390692e+00 -4.32062447e-01 -2.70604074e-01 -7.05170214e-01
6.47884488e-01 -7.67250299e-01 6.87879790e-03 7.77859092e-01
-5.22137642e-01 -9.41537246e-02 1.04085493e+00 -3.49281311e-01
1.37327492e+00 -2.10098267e+00 5.09038568e-01 7.27778375e-01
8.02874386e-01 3.58365417e-01 -5.41118383e-01 6.53374791e-01
-4.91577148e-01 2.57662326e-01 -3.45459074e-01 3.65419611e-02
3.86823490e-02 5.19711792e-01 -1.29397616e-01 5.34190357e-01
2.94833750e-01 9.22650278e-01 -1.27939796e+00 -6.74108267e-01
-1.58776008e-02 6.98675573e-01 -6.03734970e-01 5.42511158e-02
2.72867769e-01 -1.42465144e-01 -4.15455937e-01 4.93195474e-01
6.96340024e-01 -5.41579604e-01 6.88724637e-01 -6.55377984e-01
3.75845671e-01 1.98609963e-01 -1.14580357e+00 1.35953617e+00
3.34325172e-02 6.85787201e-01 9.51476768e-02 -1.24345315e+00
1.17817903e+00 2.09884927e-01 6.47117317e-01 -4.69227493e-01
2.15848461e-01 3.26885313e-01 5.46338081e-01 -2.94638902e-01
6.80602193e-02 -7.67818466e-02 3.07179332e-01 4.72435981e-01
1.44748360e-01 2.48362452e-01 3.99802804e-01 7.85714388e-01
1.07736850e+00 -5.68488598e-01 8.37210193e-02 -3.14189047e-02
6.80519640e-01 -5.46242952e-01 8.54519784e-01 -1.72976598e-01
-3.78782898e-01 2.71555930e-01 1.15471947e+00 -3.11406791e-01
-8.65037203e-01 -1.12312150e+00 5.51233953e-03 3.99454623e-01
2.14007139e-01 -5.33665299e-01 -1.82063386e-01 -9.78504419e-01
3.96867573e-01 -2.76144564e-01 -7.07233667e-01 -7.22198009e-01
-6.80103302e-01 -5.11283576e-01 3.15643579e-01 4.86539841e-01
4.66246277e-01 -9.82896507e-01 5.30297339e-01 1.07460581e-01
2.01161038e-02 -7.30179310e-01 -7.13943958e-01 -4.84057069e-02
-9.72496390e-01 -1.58492947e+00 -5.52951157e-01 -1.45192146e+00
1.28612638e+00 4.59861666e-01 7.31112421e-01 7.03666151e-01
-1.82577729e-01 5.83081022e-02 -3.50993663e-01 2.20739484e-01
-8.01889226e-02 -1.26708537e-01 1.39385372e-01 3.94304693e-01
1.64986074e-01 -8.05781841e-01 -6.74093127e-01 2.71594495e-01
-1.03087306e+00 -1.28270462e-01 7.43323803e-01 1.31669140e+00
3.05118710e-01 3.04195315e-01 8.26095641e-01 -9.03679669e-01
6.66096210e-01 -5.52763760e-01 -3.14509630e-01 2.84164429e-01
-7.73139060e-01 1.97147489e-01 5.71904540e-01 -4.61005926e-01
-1.67537808e-01 -2.58822262e-01 2.64631003e-01 -5.66637754e-01
8.36693525e-01 8.80321324e-01 -4.59965736e-01 -3.22201669e-01
-4.22461294e-02 1.97782204e-01 3.66025329e-01 -2.78037041e-01
4.11527187e-01 3.97631139e-01 8.70925561e-02 -9.50002149e-02
1.15296650e+00 4.66125906e-01 3.24436188e-01 -4.60878968e-01
-4.01794195e-01 -2.80801535e-01 -8.53949189e-01 -9.36580524e-02
3.78928930e-01 -5.86903691e-01 -8.33320379e-01 2.25231543e-01
-9.87030387e-01 2.14277089e-01 -1.65396497e-01 3.62276345e-01
-1.16634525e-01 7.60616660e-01 -7.79908478e-01 -4.32731420e-01
-3.99623871e-01 -7.37380624e-01 6.13770902e-01 6.30629659e-02
6.69587636e-04 -1.34250140e+00 2.27575049e-01 1.21192485e-01
1.83988586e-01 2.38778919e-01 1.34931469e+00 -4.56259191e-01
-6.09556556e-01 -6.19205475e-01 -7.74050117e-01 3.08604658e-01
4.66342390e-01 2.18532935e-01 -2.93237835e-01 -4.63930994e-01
-6.33469701e-01 -7.48147592e-02 1.10859811e+00 -1.16650924e-01
8.95953715e-01 -5.80649555e-01 -5.12843311e-01 6.90171719e-01
1.26246202e+00 -3.93884957e-01 3.81914467e-01 2.15574484e-02
1.08278203e+00 6.69256866e-01 2.96322137e-01 2.36174226e-01
4.85794783e-01 1.53579757e-01 8.37609768e-01 -5.67664765e-02
-4.74520586e-02 -7.78668582e-01 1.99054480e-01 1.45924699e+00
1.18558735e-01 -1.49204999e-01 -6.19570911e-01 5.78149974e-01
-1.90870130e+00 -1.03689456e+00 -2.58020371e-01 2.02756572e+00
7.06825376e-01 4.05371517e-01 -8.17640945e-02 5.06539583e-01
8.91061723e-01 7.01456070e-01 -5.18350720e-01 1.60363354e-02
-2.77376175e-01 -1.32769242e-01 3.56388986e-01 4.14945483e-01
-8.55180085e-01 5.00545204e-01 5.60447598e+00 2.65980393e-01
-1.09814179e+00 -2.16295689e-01 1.81019515e-01 3.92311104e-02
-7.51207530e-01 2.66032219e-01 -2.11125180e-01 7.20391154e-01
3.05962920e-01 -2.82924116e-01 3.57374966e-01 6.46298170e-01
-3.97669710e-02 7.32151031e-01 -1.01440263e+00 9.74115968e-01
1.36126995e-01 -1.41539383e+00 3.66107821e-01 3.95527095e-01
7.12618351e-01 -3.59356962e-02 -6.30121632e-03 2.07704931e-01
1.12403758e-01 -1.01520157e+00 2.88604498e-01 5.32844067e-01
8.71974170e-01 -7.54252911e-01 8.77480686e-01 1.06850147e-01
-1.63630068e+00 -2.12878272e-01 -5.78749955e-01 -6.30118698e-02
2.25265414e-01 9.02582049e-01 -4.51383740e-01 5.02797067e-01
3.47469181e-01 1.56185758e+00 -5.46367705e-01 1.10613859e+00
-4.89404023e-01 4.48901415e-01 -1.34356946e-01 -1.14999384e-01
1.71297297e-01 -7.54541516e-01 4.84654397e-01 6.94584668e-01
1.02625094e-01 -1.93335980e-01 1.29796535e-01 8.47855210e-01
-6.02720499e-01 2.33869329e-02 -9.70781267e-01 -6.35527372e-01
5.75730920e-01 1.42848265e+00 -5.40587723e-01 -1.74511477e-01
-3.93296957e-01 9.92571592e-01 5.03800809e-01 3.62619042e-01
-6.29600704e-01 -8.06903005e-01 8.05882692e-01 2.72474825e-01
2.92878926e-01 -2.60502577e-01 5.76926842e-02 -1.26942348e+00
3.16657424e-01 -5.84900439e-01 3.86659235e-01 -4.82233167e-01
-1.66858113e+00 4.30468798e-01 -5.19547105e-01 -1.41223478e+00
2.28011340e-01 -8.28684807e-01 -1.09136677e+00 5.55303991e-01
-1.75422037e+00 -1.34153390e+00 -3.46711278e-01 4.59742665e-01
-3.64635974e-01 -2.20903397e-01 6.85050130e-01 7.04041243e-01
-6.86447322e-01 5.56820273e-01 2.16609135e-01 7.87612796e-01
5.00859499e-01 -1.14570367e+00 6.07799478e-02 4.09080595e-01
1.21081591e-01 9.05873060e-01 1.02163821e-01 -7.72346914e-01
-1.68090546e+00 -1.02477431e+00 1.01324308e+00 2.11856380e-01
1.14173675e+00 -5.48402607e-01 -1.01521456e+00 8.36995661e-01
-5.20056598e-02 7.25616813e-01 7.91058183e-01 1.48213089e-01
-7.78194427e-01 -4.31578189e-01 -9.80094910e-01 4.89787668e-01
1.38390708e+00 -7.98961937e-01 -4.28235441e-01 4.45025146e-01
7.13907778e-01 2.42823780e-01 -9.78838921e-01 5.65449595e-01
6.38549864e-01 -6.76202655e-01 1.13374257e+00 -7.55395889e-01
4.35025215e-01 -3.76469731e-01 9.16344076e-02 -1.42726111e+00
-5.09542823e-01 -3.05291176e-01 -5.18415868e-01 1.29206574e+00
1.60794258e-01 -9.71657336e-01 1.04246497e+00 -6.11311123e-02
2.64297038e-01 -9.76511180e-01 -8.16656888e-01 -6.81720078e-01
-1.48304552e-01 3.50565046e-01 6.08074963e-01 1.42279792e+00
3.03303510e-01 5.17099500e-01 -3.49741042e-01 9.44104418e-03
8.59156668e-01 2.74130851e-01 4.49985147e-01 -1.66374707e+00
-1.41368404e-01 -6.84054971e-01 -1.07103503e+00 -1.00268626e+00
3.74265999e-01 -1.37688899e+00 -4.45711017e-01 -1.77763128e+00
3.04890692e-01 -6.92307651e-01 -5.60311615e-01 4.58546847e-01
2.83128805e-02 2.26577923e-01 3.65816429e-02 1.14256024e-01
-4.17110413e-01 1.07986975e+00 1.43142462e+00 -6.64769888e-01
1.78880185e-01 -3.05764675e-01 -7.08360851e-01 6.43893003e-01
4.52534497e-01 -3.99283141e-01 -5.99911511e-01 -3.41512203e-01
7.65539527e-01 -1.60041004e-01 3.14035386e-01 -5.05209923e-01
3.37435484e-01 1.20823450e-01 2.49335662e-01 -5.53074121e-01
3.20960999e-01 -1.15186083e+00 -1.54781178e-01 7.74113178e-01
-1.25294968e-01 -6.95554540e-02 -3.66525173e-01 9.68010426e-01
-4.78957236e-01 1.88399889e-02 4.86237019e-01 3.27416331e-01
-4.38889056e-01 8.36779833e-01 3.59136611e-01 -1.65634736e-01
9.74354744e-01 -2.99691558e-01 -5.64868033e-01 -4.29372191e-01
-4.95902777e-01 6.23574972e-01 4.02437419e-01 5.54047525e-01
1.14310753e+00 -1.82237089e+00 -5.26864111e-01 4.84255791e-01
3.52833718e-01 -7.67839253e-02 4.87204455e-02 8.62471700e-01
-5.60761988e-01 2.28468791e-01 -3.24765503e-01 -2.13824660e-01
-1.36075819e+00 5.23091555e-01 3.58016461e-01 -3.11859041e-01
-5.61415315e-01 7.48114228e-01 -4.35559265e-02 -8.63415718e-01
2.08093166e-01 -8.53017047e-02 -2.97808856e-01 2.08242595e-01
1.57103047e-01 2.64764041e-01 -4.11895186e-01 -7.64343739e-01
-4.80543524e-01 6.41177297e-01 -5.60819581e-02 4.92030561e-01
1.65207231e+00 1.57178968e-01 -7.03159451e-01 3.53606343e-01
1.63346207e+00 8.07350948e-02 -9.36750293e-01 -4.24122512e-01
-1.87161282e-01 -7.28104711e-01 1.02284566e-01 -1.71855301e-01
-1.67759442e+00 1.03325117e+00 1.07631601e-01 2.68877119e-01
6.96285605e-01 -4.80387136e-02 8.09187710e-01 4.51796055e-01
2.44688705e-01 -5.91751277e-01 4.13156658e-01 4.61818367e-01
8.32687855e-01 -1.35770249e+00 2.78853416e-01 -8.16688597e-01
-3.72799456e-01 1.14974797e+00 6.43503129e-01 -4.06424969e-01
1.13941431e+00 -3.41165274e-01 -4.47252482e-01 -6.46098256e-01
-4.86313432e-01 1.16621949e-01 3.86977017e-01 4.59088564e-01
4.65499490e-01 1.38100117e-01 -2.88177550e-01 3.66201282e-01
3.53623703e-02 -3.43197882e-01 4.97440785e-01 7.61413276e-01
-3.71524751e-01 -1.23296523e+00 1.35325387e-01 7.84729064e-01
2.14281633e-01 -7.95372203e-02 -8.44470322e-01 6.67771399e-01
-1.52400523e-01 6.24256968e-01 1.56653345e-01 -5.55002034e-01
2.11029723e-01 -2.71056294e-01 2.88662076e-01 -6.22491419e-01
-1.04619458e-01 -3.75763148e-01 -7.46681839e-02 -2.99421549e-01
-2.62898386e-01 -4.04727608e-01 -1.51073861e+00 -6.27673686e-01
-5.87632060e-01 1.99024737e-01 3.73859942e-01 4.35779721e-01
4.24650818e-01 6.36101365e-01 8.89293551e-01 -6.74488008e-01
-4.52782780e-01 -7.95956314e-01 -1.08506489e+00 5.26761830e-01
3.33368719e-01 -7.78362870e-01 -6.59453809e-01 -4.88453567e-01] | [7.211602210998535, 6.206033229827881] |
744cee47-2391-4cef-ba38-e8fd86c54415 | proto-value-networks-scaling-representation | 2304.12567 | null | https://arxiv.org/abs/2304.12567v1 | https://arxiv.org/pdf/2304.12567v1.pdf | Proto-Value Networks: Scaling Representation Learning with Auxiliary Tasks | Auxiliary tasks improve the representations learned by deep reinforcement learning agents. Analytically, their effect is reasonably well understood; in practice, however, their primary use remains in support of a main learning objective, rather than as a method for learning representations. This is perhaps surprising given that many auxiliary tasks are defined procedurally, and hence can be treated as an essentially infinite source of information about the environment. Based on this observation, we study the effectiveness of auxiliary tasks for learning rich representations, focusing on the setting where the number of tasks and the size of the agent's network are simultaneously increased. For this purpose, we derive a new family of auxiliary tasks based on the successor measure. These tasks are easy to implement and have appealing theoretical properties. Combined with a suitable off-policy learning rule, the result is a representation learning algorithm that can be understood as extending Mahadevan & Maggioni (2007)'s proto-value functions to deep reinforcement learning -- accordingly, we call the resulting object proto-value networks. Through a series of experiments on the Arcade Learning Environment, we demonstrate that proto-value networks produce rich features that may be used to obtain performance comparable to established algorithms, using only linear approximation and a small number (~4M) of interactions with the environment's reward function. | ['Marc G. Bellemare', 'Pablo Samuel Castro', 'Ross Goroshin', 'Charline Le Lan', 'Rishabh Agarwal', 'Joshua Greaves', 'Jesse Farebrother'] | 2023-04-25 | null | null | null | null | ['atari-games'] | ['playing-games'] | [ 2.48466894e-01 4.43898350e-01 -1.80249304e-01 5.62714525e-02
-3.95103514e-01 -5.94171047e-01 9.98120010e-01 3.82355064e-01
-8.36530864e-01 9.39219832e-01 1.75372094e-01 -2.60400921e-01
-5.17383695e-01 -8.76691759e-01 -8.34298611e-01 -9.98690784e-01
-3.92756104e-01 6.72095239e-01 1.68817192e-01 -5.44967294e-01
3.97041947e-01 5.70869625e-01 -1.65130985e+00 -1.36388555e-01
7.56137669e-01 7.54168451e-01 4.90081370e-01 5.98240376e-01
2.15841219e-01 1.15804541e+00 -6.28305018e-01 -6.85229525e-02
3.41835827e-01 -3.97010863e-01 -9.97465312e-01 2.61517875e-02
-2.17281833e-01 -2.96782702e-01 -2.00167209e-01 8.27919185e-01
3.47140074e-01 6.97212398e-01 7.99114883e-01 -1.12566721e+00
-5.16208112e-01 7.61905909e-01 -2.83651739e-01 2.89494783e-01
8.64866748e-02 2.19924867e-01 9.73268986e-01 -3.77076387e-01
4.82692748e-01 1.15512788e+00 4.76588696e-01 6.69545054e-01
-1.51560569e+00 -3.56066287e-01 3.21101993e-01 3.28266174e-02
-8.08068871e-01 -4.41802353e-01 6.37179673e-01 -2.92744339e-01
1.00848770e+00 7.81210884e-03 8.40413451e-01 1.13462710e+00
4.13537724e-03 8.05389225e-01 1.23657465e+00 -6.37623072e-01
4.77593571e-01 2.54141957e-01 1.08324058e-01 6.74331903e-01
3.30085218e-01 6.92161739e-01 -3.33059907e-01 -2.04433650e-01
1.00405216e+00 9.78488009e-04 -1.64501786e-01 -8.23299050e-01
-1.02401125e+00 1.13412046e+00 5.34842134e-01 2.80550361e-01
-5.17181993e-01 4.34658349e-01 4.38784599e-01 7.72836626e-01
2.66663522e-01 7.64551342e-01 -3.32834005e-01 -2.27107510e-01
-1.85626179e-01 5.22911847e-01 7.45182276e-01 7.22440481e-01
9.22381163e-01 3.13704282e-01 1.07761681e-01 7.83395886e-01
1.67462483e-01 2.22137183e-01 5.49147964e-01 -1.32597780e+00
1.78574353e-01 2.21519962e-01 2.49571592e-01 -5.85105360e-01
-5.55781066e-01 -6.35802627e-01 -6.03085339e-01 6.37795150e-01
5.98712385e-01 -1.83249176e-01 -6.86888635e-01 2.17920232e+00
8.97053331e-02 -8.62516388e-02 2.77706772e-01 6.08638942e-01
5.99951185e-02 5.83484769e-01 1.27720088e-01 -2.73972362e-01
9.27776456e-01 -8.93412590e-01 -3.97626519e-01 -3.31782103e-01
7.99480200e-01 -1.45120949e-01 1.12645805e+00 4.02299643e-01
-1.33896315e+00 -4.37635690e-01 -7.40961790e-01 1.82702556e-01
-4.78497088e-01 -3.13037038e-01 1.09723294e+00 4.10917908e-01
-1.23108685e+00 8.55750561e-01 -7.98287153e-01 -3.16962093e-01
5.21214902e-01 5.43445230e-01 -1.94240406e-01 2.30338857e-01
-9.33736384e-01 1.34889674e+00 5.74576795e-01 -1.13618106e-01
-1.35830045e+00 -3.60379577e-01 -8.02394152e-01 2.80958176e-01
5.64745367e-01 -8.15207779e-01 1.58587110e+00 -1.08355713e+00
-1.65981126e+00 6.21341467e-01 1.92757472e-01 -6.87454343e-01
2.67334580e-01 -7.81193301e-02 2.73216188e-01 2.36657143e-01
-5.83485775e-02 6.38140500e-01 1.02471411e+00 -1.45191884e+00
-7.51483619e-01 -3.33743453e-01 7.25340843e-01 4.47563231e-01
-3.11554432e-01 -2.63769597e-01 2.47524694e-01 -7.84962893e-01
-4.50581431e-01 -1.00387967e+00 -7.27604032e-01 -9.30748321e-03
1.12905346e-01 -3.66570354e-01 3.71004134e-01 -1.00271486e-01
7.77376354e-01 -2.19368505e+00 4.57951039e-01 3.21622044e-01
3.51688415e-01 1.74177393e-01 -4.12732035e-01 4.73542958e-01
-1.54806137e-01 -3.47504951e-02 -4.26298171e-01 -2.01644972e-01
1.37101799e-01 5.45497298e-01 -3.22847098e-01 3.59622180e-01
1.41501734e-02 1.07406938e+00 -1.13922083e+00 -1.60125911e-01
3.04400027e-01 9.39771757e-02 -7.52752781e-01 3.78555298e-01
-2.62783915e-01 2.86004156e-01 -6.37299657e-01 3.88696529e-02
1.34842217e-01 -3.59413028e-02 3.76592577e-02 5.42348027e-01
3.44551913e-02 4.13583487e-01 -1.03820193e+00 1.63181353e+00
-6.56495988e-01 4.38835055e-01 2.69932479e-01 -1.55798292e+00
7.22476184e-01 1.57903686e-01 4.23890084e-01 -6.76045537e-01
1.21856764e-01 8.54335800e-02 4.33105916e-01 -2.76613504e-01
3.19722354e-01 -3.18541259e-01 2.52685491e-02 8.77637923e-01
1.42267302e-01 -1.92045748e-01 4.65179265e-01 2.16586515e-01
1.11425889e+00 3.26546341e-01 4.60437149e-01 -4.97668445e-01
2.99316823e-01 1.13011152e-02 1.54425636e-01 1.08107114e+00
3.00229937e-02 -9.13453773e-02 7.92411447e-01 -3.40424299e-01
-1.04613864e+00 -1.04514921e+00 -1.40468823e-02 1.58998716e+00
-1.09278247e-01 -2.42179587e-01 -6.27441525e-01 -5.72719574e-01
2.17989832e-02 6.71964824e-01 -1.08817625e+00 -4.81478393e-01
-6.23584926e-01 -6.08285725e-01 3.08887869e-01 7.02653348e-01
2.56678075e-01 -1.59806657e+00 -1.20961010e+00 3.48585606e-01
4.49558467e-01 -4.96042252e-01 -8.68576989e-02 9.66970563e-01
-1.05922115e+00 -1.14835382e+00 -8.13471675e-01 -8.23960841e-01
6.42511010e-01 3.30729425e-01 1.15291405e+00 2.39134386e-01
6.51399195e-02 6.85436547e-01 -3.25443596e-01 -4.09028500e-01
-5.52336931e-01 1.71423808e-01 3.69401723e-02 -5.09937644e-01
1.42914588e-02 -1.06602108e+00 -3.70471418e-01 -1.17291264e-01
-1.06147206e+00 -2.84540690e-02 8.03105116e-01 1.15648210e+00
1.64445639e-01 -1.89019050e-02 9.02743578e-01 -1.05401576e+00
9.50675726e-01 -4.57428485e-01 -7.13855743e-01 2.72743180e-02
-6.27443671e-01 5.44166744e-01 8.05427730e-01 -6.54463112e-01
-1.05685306e+00 -1.42714055e-02 -6.85722977e-02 -1.25362635e-01
-1.61625087e-01 5.72363198e-01 1.92572683e-01 -5.43361902e-02
7.59453952e-01 3.93494040e-01 1.79126665e-01 -2.94676214e-01
4.77834642e-01 2.36307979e-01 3.39745551e-01 -9.63245749e-01
9.11077142e-01 1.81062385e-01 9.09960866e-02 -7.00464725e-01
-8.18153799e-01 -7.84632564e-02 -3.51007819e-01 6.42669853e-03
4.12672073e-01 -5.84968507e-01 -1.06572914e+00 2.01275468e-01
-8.64861786e-01 -1.04375339e+00 -8.00370932e-01 3.68255436e-01
-1.13274670e+00 3.34378742e-02 -5.87424517e-01 -9.21701670e-01
2.47668087e-01 -1.15351629e+00 6.67239726e-01 -6.62055537e-02
-5.34307323e-02 -1.26753962e+00 1.20775841e-01 -1.76032782e-01
5.21069646e-01 7.32054235e-04 1.31290114e+00 -7.09249377e-01
-4.23608929e-01 3.91657412e-01 1.38945533e-02 2.87694097e-01
1.12731289e-02 -5.52717507e-01 -8.94389331e-01 -3.60984802e-01
1.34598762e-01 -7.45882869e-01 1.09419000e+00 3.88838202e-01
1.30961525e+00 -3.82319719e-01 -3.88367139e-02 3.50983143e-01
1.23988068e+00 2.19370008e-01 5.21212876e-01 7.20982671e-01
3.64269018e-01 7.68210948e-01 4.32242185e-01 6.12735748e-01
7.60055482e-02 5.79904497e-01 6.97892368e-01 1.98976591e-01
1.48705348e-01 -2.40592986e-01 3.79034430e-01 4.38918024e-01
-4.38282818e-01 1.43869311e-01 -5.47452390e-01 4.28519666e-01
-2.05078173e+00 -1.19218385e+00 3.30389172e-01 2.26815820e+00
8.73015821e-01 2.42397055e-01 4.64249909e-01 6.69612288e-02
3.98381352e-01 1.92531273e-01 -7.40121961e-01 -6.08339131e-01
1.41587362e-01 3.62193257e-01 2.72339493e-01 4.79752213e-01
-7.91777909e-01 7.91690767e-01 6.64037704e+00 6.56004906e-01
-8.28041196e-01 4.90048714e-02 3.63019407e-01 -1.99932177e-02
-4.33307648e-01 -1.25766233e-01 -4.60285574e-01 2.28636488e-01
1.00301778e+00 -3.92791182e-01 9.34247077e-01 1.00691867e+00
5.17915711e-02 -2.09076449e-01 -1.44038773e+00 7.25031495e-01
-3.07375699e-01 -1.26674640e+00 -5.91035075e-02 3.84945750e-01
6.53322816e-01 -1.25205353e-01 2.64839232e-01 8.65503550e-01
8.45938087e-01 -1.13690174e+00 7.47120142e-01 1.08111665e-01
3.14199954e-01 -9.67648804e-01 4.34500128e-01 6.93694413e-01
-8.65209341e-01 -5.22536874e-01 -4.90464002e-01 -3.50288004e-01
-2.55737513e-01 -3.50615717e-02 -4.66850728e-01 1.16829135e-01
3.62426400e-01 6.19979024e-01 -4.17775393e-01 8.12292159e-01
-2.45476559e-01 3.74325037e-01 -2.83477098e-01 -1.16111726e-01
6.89162254e-01 -1.85219750e-01 2.86087155e-01 9.52545285e-01
1.61682427e-01 1.71068519e-01 2.69467294e-01 6.54244900e-01
-1.53611124e-01 -1.14280768e-01 -1.10485590e+00 3.46725695e-02
4.08826292e-01 1.05305099e+00 -7.08331168e-01 -1.46280304e-01
-2.53034860e-01 5.54302633e-01 9.68456328e-01 4.79223907e-01
-5.78681588e-01 -1.04274288e-01 5.44117332e-01 2.06434410e-02
2.70622373e-01 -1.20964766e-01 -5.51852249e-02 -8.94846916e-01
-2.48533502e-01 -1.09084845e+00 2.56043762e-01 -4.90691602e-01
-1.00040305e+00 4.54555184e-01 2.03686059e-01 -9.34331000e-01
-7.79358804e-01 -7.36881256e-01 -4.58333999e-01 5.81552863e-01
-1.54891932e+00 -8.61250401e-01 1.02493100e-01 7.79114723e-01
6.01054430e-01 -3.23593855e-01 1.12625194e+00 -2.99761832e-01
-2.91663647e-01 3.36222619e-01 4.13719982e-01 -1.54818207e-01
6.26699775e-02 -1.55800068e+00 1.59139670e-02 5.29103458e-01
2.91656137e-01 6.73332214e-01 6.91592634e-01 -1.14453554e-01
-1.42943144e+00 -8.68294060e-01 1.31519496e-01 -4.44878638e-01
7.79253542e-01 -3.33701015e-01 -8.88500392e-01 8.64920080e-01
3.65750521e-01 -9.44366157e-02 3.92285466e-01 3.86430621e-01
-1.52603447e-01 3.33500355e-02 -8.80819261e-01 6.87046826e-01
1.19059467e+00 -4.44396675e-01 -9.60081518e-01 1.76936701e-01
6.16920173e-01 -1.57132417e-01 -6.32777095e-01 1.53072894e-01
2.91193843e-01 -8.37549806e-01 9.27230120e-01 -9.42086160e-01
5.45384526e-01 1.67501032e-01 -7.30724484e-02 -1.88321841e+00
-5.54146826e-01 -6.60327494e-01 -3.87773216e-01 7.80702531e-01
3.18733901e-01 -8.49206269e-01 6.15222335e-01 2.36535981e-01
-2.40479618e-01 -8.85993898e-01 -8.28632176e-01 -9.34717417e-01
3.80852669e-01 -2.35964179e-01 2.19282851e-01 7.69695938e-01
9.30045173e-02 5.94996810e-01 -2.02136144e-01 -2.92703301e-01
6.33171916e-01 5.47654033e-02 6.49145603e-01 -1.41292989e+00
-6.83430135e-01 -7.58673847e-01 1.69816345e-01 -1.11922824e+00
6.21203065e-01 -9.00872052e-01 1.24425523e-01 -1.27246070e+00
2.51218468e-01 -7.16091514e-01 -5.20632982e-01 6.38424098e-01
1.08910441e-01 -1.40768275e-01 3.94660592e-01 1.61022127e-01
-4.98198122e-01 7.53559947e-01 1.37035573e+00 -6.27337024e-03
-2.77701795e-01 2.88563401e-01 -1.08532238e+00 9.34699237e-01
8.90809536e-01 -5.21746814e-01 -7.81233251e-01 -3.82307738e-01
4.82594609e-01 1.09560199e-01 3.47216308e-01 -6.35845304e-01
6.83480524e-04 -3.52891117e-01 2.52700716e-01 1.32172415e-02
3.93486261e-01 -9.06039655e-01 -3.47939581e-01 6.91458523e-01
-1.06250250e+00 -9.62648448e-03 1.80980250e-01 6.18911624e-01
1.38451532e-01 -8.63773823e-01 7.43349969e-01 -4.17568803e-01
-6.09140635e-01 2.04637840e-01 -8.08799267e-01 1.61702141e-01
1.02782881e+00 -2.19176039e-01 -3.85354608e-01 -5.06333351e-01
-8.35205555e-01 2.26663187e-01 3.36806655e-01 1.27169698e-01
5.89521468e-01 -1.20259154e+00 -6.45558536e-01 1.23958131e-02
-2.71862186e-02 1.24364927e-01 -1.86628759e-01 6.70241714e-01
-1.10250115e-01 2.60127008e-01 -5.32082260e-01 -1.96788415e-01
-7.73156285e-01 8.49678457e-01 2.84822792e-01 -5.69227755e-01
-6.60418451e-01 4.62941527e-01 6.31591678e-01 -3.16583693e-01
4.15806890e-01 -2.31787413e-01 -3.91899973e-01 1.37161016e-01
4.68523264e-01 2.74703920e-01 -2.76943803e-01 -7.30600283e-02
1.18846141e-01 3.61910164e-02 -3.18184763e-01 -2.56942898e-01
1.82032669e+00 5.10461582e-03 1.16179794e-01 4.18553829e-01
8.85059297e-01 -3.28790545e-01 -1.49859548e+00 -2.60903865e-01
5.84409088e-02 -3.50219309e-01 -1.11435682e-01 -5.29928565e-01
-8.66955936e-01 9.39690173e-01 3.72155458e-01 5.72473705e-01
1.11269820e+00 1.38475239e-01 8.60018283e-02 1.02084923e+00
5.59058607e-01 -1.08848882e+00 4.88303900e-01 6.82467520e-01
1.05585670e+00 -9.99545872e-01 -1.07798368e-01 8.95451084e-02
-5.80731094e-01 9.54229593e-01 6.36491716e-01 -4.71838325e-01
3.43565971e-01 2.17184573e-01 -3.18468750e-01 -5.70006184e-02
-1.08283925e+00 -5.50590038e-01 -3.12616885e-01 1.17932665e+00
2.08943814e-01 -2.06941873e-01 -5.43665886e-02 3.20274830e-01
-2.26523191e-01 -2.29840815e-01 6.97354734e-01 9.69217658e-01
-8.30542386e-01 -1.12593520e+00 -1.22981995e-01 6.08133078e-01
-3.10039431e-01 -1.95593182e-02 -1.04740061e-01 1.02926314e+00
-1.62838742e-01 6.20232224e-01 -3.35648879e-02 1.96469724e-01
1.54148683e-01 -1.94400847e-01 7.93019950e-01 -9.00782347e-01
-7.16706991e-01 -9.73666906e-02 8.54305476e-02 -4.79966432e-01
-5.20270169e-01 -5.35472333e-01 -1.09908116e+00 -3.70201290e-01
-7.90927708e-02 3.81290704e-01 2.25574538e-01 1.09531331e+00
-1.07502803e-01 6.13952518e-01 7.63863504e-01 -1.05947244e+00
-1.14886844e+00 -9.55142975e-01 -7.32499659e-01 4.08392012e-01
4.59358454e-01 -9.41288352e-01 -2.62126058e-01 -1.35523245e-01] | [4.074431896209717, 1.6780434846878052] |
f043fba3-da36-4a4d-96f6-ff4739f9400e | artificial-bandwidth-extension-using-deep | 2108.13326 | null | https://arxiv.org/abs/2108.13326v1 | https://arxiv.org/pdf/2108.13326v1.pdf | Artificial bandwidth extension using deep neural network and $H^\infty$ sampled-data control theory | Artificial bandwidth extension is applied to speech signals to improve their quality in narrowband telephonic communication. For accomplishing this, the missing high-frequency (high-band) components of speech signals are recovered by utilizing a new extrapolation process based on sampled-data control theory and deep neural network (DNN). The $H^\infty$ sampled-data control theory helps in designing of a high-band filter to recover the high-frequency signals by optimally utilizing the inter-sample signals. Non-stationary (time-varying) characteristics of speech signals forces to use numerous high-band filters. Hence, we use a deep neural network for estimating the high-band filter information and a gain factor for a specified narrowband information of the unseen signal. The objective analysis is done on the TIMIT dataset and RSR15 dataset. Additionally, the objective analysis is performed separately for the voiced speech as well as for the unvoiced speech as generally needed in speech processing. Subjective analysis is done on the RSR15 dataset. | ['Hanumant Singh Shekhawat', 'Deepika Gupta'] | 2021-08-30 | null | null | null | null | ['bandwidth-extension', 'bandwidth-extension'] | ['audio', 'speech'] | [ 7.13601410e-02 4.23678905e-02 4.79400195e-02 -2.79965878e-01
-6.12158775e-01 -1.75671771e-01 1.82240512e-02 -3.38284373e-01
-1.31249517e-01 8.20487380e-01 3.02747101e-01 -4.79543716e-01
-5.10094523e-01 -5.14099419e-01 -2.90570468e-01 -9.88002837e-01
-4.76242788e-03 -1.21964868e-02 -2.56088763e-01 -2.07917050e-01
-1.93238929e-01 5.15263081e-01 -1.75083101e+00 2.59846419e-01
6.56747043e-01 1.22143555e+00 3.95452529e-01 9.47887003e-01
-7.78745040e-02 2.78364152e-01 -8.98901880e-01 1.51398867e-01
1.80984572e-01 -5.59908450e-01 -3.96723032e-01 4.11454231e-01
-1.72525570e-01 -4.49507445e-01 -4.82429504e-01 1.08833015e+00
9.21176672e-01 5.31563044e-01 5.08596480e-01 -6.44678056e-01
-2.76773483e-01 4.90187228e-01 -3.17798257e-02 4.37307030e-01
1.94127336e-01 6.92512244e-02 8.59400034e-01 -8.68876755e-01
2.23364189e-01 1.00376427e+00 5.98778188e-01 2.49669671e-01
-9.27614689e-01 -5.58081448e-01 -3.29335004e-01 1.73699588e-01
-1.04900205e+00 -8.46519172e-01 1.21680498e+00 -2.97883570e-01
8.42574358e-01 3.10988098e-01 7.78359592e-01 1.03602946e+00
-8.43119994e-02 5.55112004e-01 6.24341428e-01 -7.17660129e-01
1.26450181e-01 4.75068912e-02 2.27276340e-01 7.93899447e-02
-4.85119611e-01 6.58784270e-01 -3.31344992e-01 3.50214634e-03
6.51275873e-01 -3.11808914e-01 -8.63825798e-01 4.55386698e-01
-5.85247099e-01 7.67588198e-01 5.48572373e-03 6.37568355e-01
-8.02707255e-01 -7.50954673e-02 5.79222202e-01 6.75201774e-01
6.89240158e-01 2.42962435e-01 -7.71054566e-01 -2.65110433e-01
-9.84863400e-01 1.27304763e-01 7.57122993e-01 6.64934278e-01
3.93860281e-01 9.32768941e-01 4.49319184e-02 1.26808429e+00
2.42901564e-01 4.62691516e-01 5.79563618e-01 -9.01802540e-01
3.53190720e-01 -2.72605151e-01 3.02390337e-01 -6.45491123e-01
-5.71691632e-01 -9.68226910e-01 -9.76358354e-01 -5.97577617e-02
3.94376725e-01 -6.51549876e-01 -8.54369342e-01 1.53482342e+00
3.76388043e-01 1.00054584e-01 1.24127157e-01 9.80906367e-01
4.94209826e-01 1.19100869e+00 -3.67324471e-01 -8.22617531e-01
1.14457273e+00 -6.66904807e-01 -1.44148517e+00 2.75052153e-02
-8.48919526e-02 -1.05026102e+00 1.02773547e+00 6.61108792e-01
-1.09668732e+00 -1.03613412e+00 -9.04153645e-01 2.06693649e-01
-1.72718972e-01 4.01507437e-01 9.96832252e-02 7.79710174e-01
-7.80510068e-01 5.55073261e-01 -2.48240456e-01 3.80191416e-01
-1.40011460e-01 3.23547870e-01 -4.49873246e-02 3.66241246e-01
-1.67055583e+00 5.43337166e-01 3.22298586e-01 3.78528595e-01
-6.65168047e-01 -9.23938334e-01 -5.20471036e-01 2.21355408e-01
1.21523425e-01 -1.65851340e-01 1.32384646e+00 -1.18503356e+00
-1.96396267e+00 2.77775049e-01 4.72617475e-03 -6.29218459e-01
3.32378894e-01 -4.73929979e-02 -1.09135091e+00 4.86079365e-01
-4.55451071e-01 -1.08828560e-01 1.59583771e+00 -7.33358204e-01
-6.24340296e-01 -1.13597669e-01 -4.71151918e-01 8.06770697e-02
-2.64340371e-01 -3.56796980e-01 1.69359416e-01 -1.02018285e+00
4.21952516e-01 -5.68569064e-01 1.74921870e-01 -3.99864763e-01
-1.67324319e-01 -9.63010266e-02 1.03487468e+00 -1.30113506e+00
1.23956144e+00 -2.59169888e+00 -9.10221189e-02 5.58888018e-02
-3.11621815e-01 7.33442962e-01 -7.83018023e-03 4.22737181e-01
-3.97748828e-01 -2.64213800e-01 6.86344132e-02 -1.57457620e-01
-1.59842521e-01 -1.44924834e-01 -5.05464792e-01 5.43227792e-01
1.46108255e-01 -1.32439300e-01 -5.26678503e-01 4.81120013e-02
4.90712970e-01 7.74639845e-01 -5.27415216e-01 4.41122174e-01
-1.57874376e-01 6.98355019e-01 -1.67257249e-01 2.55753994e-01
8.97430301e-01 4.87107933e-01 -1.81786180e-01 -6.30964756e-01
-1.43595457e-01 4.57147062e-01 -1.01403606e+00 1.30459917e+00
-8.32861125e-01 8.32152009e-01 8.52486253e-01 -1.08961368e+00
1.08975434e+00 1.02455938e+00 4.50181842e-01 -6.80064201e-01
3.04176390e-01 4.39231157e-01 1.90225065e-01 -9.50688481e-01
2.78530508e-01 -3.91723931e-01 5.90674043e-01 3.35251689e-02
2.31663093e-01 -3.51113886e-01 -1.61751658e-01 -5.00273705e-01
4.77224678e-01 -2.87672073e-01 1.76467061e-01 -2.85662264e-01
7.58312285e-01 -5.66889226e-01 4.59871113e-01 4.01734442e-01
-3.11740667e-01 4.31320250e-01 8.22109655e-02 -3.37719113e-01
-1.25366616e+00 -6.66527092e-01 -4.78563458e-01 1.12808669e+00
-5.16412854e-01 2.80444682e-01 -8.35521758e-01 9.99914482e-02
-1.62054211e-01 9.08819497e-01 -5.44607602e-02 -2.37490207e-01
-6.06513143e-01 -3.98210555e-01 3.45588624e-01 1.50808319e-01
3.19173396e-01 -1.00768328e+00 -1.56077936e-01 7.95244515e-01
-3.41755480e-01 -9.34412420e-01 -5.30963540e-01 6.53811038e-01
-6.84145451e-01 -6.09171748e-01 -8.87823164e-01 -7.95249581e-01
1.87841684e-01 -3.33901122e-02 5.41233659e-01 -1.81542858e-01
-8.99664238e-02 3.02043390e-02 -4.29575354e-01 -3.56160045e-01
-7.19607949e-01 -1.65355489e-01 2.60362446e-01 2.98389047e-01
1.30156398e-01 -7.50539124e-01 -5.45791388e-01 3.96828353e-01
-6.48158908e-01 -3.14071655e-01 3.19862515e-01 1.07218194e+00
1.91817716e-01 7.54172504e-01 1.27375937e+00 -1.58690810e-01
1.11854219e+00 -2.11871997e-01 -8.75142217e-01 -3.04064333e-01
9.86722950e-03 -3.35266620e-01 1.07558835e+00 -8.45074475e-01
-1.26493609e+00 -2.83787757e-01 -7.99152136e-01 -4.90183681e-01
-1.12850502e-01 4.11386847e-01 -2.56924033e-01 3.23116660e-01
5.79006135e-01 3.33020985e-01 5.13951518e-02 -7.01862931e-01
1.06667042e-01 1.44671750e+00 4.86890554e-01 -3.08218122e-01
3.10866654e-01 1.15697287e-01 -4.22691375e-01 -1.54704452e+00
-6.51874244e-01 -6.13705456e-01 -2.05479607e-01 -1.77158505e-01
5.59310496e-01 -6.90346420e-01 -8.84513259e-01 6.53258026e-01
-1.26455915e+00 -2.23720763e-02 -3.96112919e-01 1.17648315e+00
-6.72590554e-01 1.72050878e-01 -8.88163745e-01 -1.57288492e+00
-4.23279047e-01 -1.27851236e+00 6.57862008e-01 2.03292772e-01
-1.64853290e-01 -7.89621472e-01 -2.74730206e-01 3.69556487e-01
5.26744008e-01 -2.80485868e-01 9.43182886e-01 -5.45850992e-01
-2.93371491e-02 -3.90599489e-01 2.78168112e-01 9.29907024e-01
1.68450698e-01 -1.58589289e-01 -1.28254783e+00 -2.15561807e-01
9.18238580e-01 3.36101651e-02 5.34677029e-01 1.05105925e+00
1.05438673e+00 -4.16486621e-01 2.95080125e-01 5.41649818e-01
1.04389107e+00 7.94889629e-01 5.83573341e-01 -3.10409755e-01
6.67275563e-02 7.27874160e-01 5.15792191e-01 4.74387914e-01
-4.67771411e-01 3.04344118e-01 1.21063195e-01 -2.01727990e-02
-2.90754646e-01 2.81302966e-02 9.98563394e-02 1.30257666e+00
1.38393566e-01 -4.43646044e-01 -2.51866728e-01 6.97925508e-01
-1.18109226e+00 -8.37024331e-01 -1.72007009e-01 2.19122767e+00
9.55075920e-01 1.97754979e-01 6.22824125e-04 8.71149004e-01
1.15769494e+00 2.86644936e-01 -6.13605022e-01 -5.54690897e-01
3.13222595e-02 3.97569776e-01 2.93615878e-01 7.53310740e-01
-8.86160731e-01 4.52493280e-01 6.19595718e+00 1.32719052e+00
-1.26482236e+00 -1.21208355e-01 5.58170259e-01 3.11775841e-02
-3.46664935e-02 -3.89317483e-01 -6.63883090e-01 4.25715238e-01
1.33509195e+00 9.57284216e-03 8.61514270e-01 6.91066265e-01
9.13245857e-01 5.67909107e-02 -6.07102156e-01 9.94101882e-01
-4.59409088e-01 -1.03992450e+00 -4.35919911e-01 -1.10010855e-01
4.52855229e-01 -3.59762162e-01 2.27941588e-01 2.58825958e-01
-3.66138339e-01 -6.74709797e-01 6.78740323e-01 4.20138508e-01
8.27856719e-01 -9.56112981e-01 6.49323106e-01 5.92478752e-01
-9.61239576e-01 -3.86738449e-01 -4.54065681e-01 -5.79053573e-02
3.32652926e-01 1.27553630e+00 -9.09010708e-01 3.46603036e-01
3.21952462e-01 2.05302238e-01 8.29970479e-01 8.75804663e-01
-1.99021101e-01 9.15157080e-01 -1.81246966e-01 4.91808094e-02
1.79089174e-01 -2.17754677e-01 8.77048254e-01 9.37761724e-01
4.44793642e-01 2.16674164e-01 -2.71072686e-01 6.96181774e-01
-8.36187303e-02 -3.96124609e-02 -1.59810618e-01 -3.83727461e-01
5.06589234e-01 8.71260166e-01 -2.42599860e-01 -1.89390957e-01
-4.79796417e-02 4.72767115e-01 -3.81558448e-01 7.94366539e-01
-6.60231233e-01 -7.14644551e-01 5.33043623e-01 1.26672655e-01
3.61105293e-01 -2.49388456e-01 4.85568214e-03 -6.21323764e-01
-1.63102850e-01 -1.06245720e+00 -5.65319322e-02 -8.42098176e-01
-1.07856882e+00 7.21637487e-01 -3.69852126e-01 -1.29943299e+00
-5.40011823e-01 -4.67178285e-01 -3.98113489e-01 1.31291401e+00
-1.51830566e+00 -5.48599660e-01 2.89344221e-01 6.37931943e-01
1.10243070e+00 -3.96932751e-01 8.30488145e-01 6.20785415e-01
-2.78001755e-01 7.27709383e-02 4.10514027e-01 -9.71368551e-02
3.26055288e-01 -8.34445894e-01 9.26217288e-02 5.14133513e-01
-3.95105720e-01 5.00575066e-01 1.05593574e+00 -3.81525159e-01
-1.25279927e+00 -7.84304500e-01 7.46655524e-01 5.29218376e-01
5.61541498e-01 -2.78425783e-01 -9.73742783e-01 1.28076449e-01
1.49409875e-01 -7.48525336e-02 5.35759330e-01 -3.24172407e-01
3.43450725e-01 -3.65468651e-01 -1.28559387e+00 2.37361118e-01
3.74649256e-01 -8.11129510e-01 -7.77938962e-01 2.60617018e-01
1.08006477e+00 -4.15729851e-01 -7.63100982e-01 7.35759363e-02
3.38181943e-01 -9.65634704e-01 8.94142747e-01 -5.16223311e-01
1.07073426e-01 -1.16659485e-01 -3.26456934e-01 -1.74824131e+00
-7.72768930e-02 -1.13247955e+00 -6.29351437e-02 9.87481773e-01
2.31133223e-01 -8.12564194e-01 6.00542367e-01 1.85326234e-01
-3.24259013e-01 -2.62585938e-01 -9.58829939e-01 -6.37070060e-01
-3.16781998e-01 -6.00562692e-01 4.88175094e-01 6.75367653e-01
1.99130923e-02 1.22950524e-01 -6.95239365e-01 5.03799736e-01
5.54625511e-01 -1.97534293e-01 2.64916986e-01 -9.57417965e-01
-3.96471202e-01 5.54826260e-02 8.62517282e-02 -1.18319547e+00
1.16734520e-01 -8.57348740e-02 2.02792630e-01 -1.06606233e+00
-9.48856115e-01 -3.20363455e-02 -1.14810131e-01 -2.84548789e-01
3.14088196e-01 -2.36001372e-01 -1.96027234e-01 -2.40684077e-01
6.46524131e-01 8.19310367e-01 1.52791524e+00 -1.47422910e-01
-4.47236389e-01 7.91129351e-01 -6.17476515e-02 8.11772168e-01
6.99554145e-01 -2.68981218e-01 -6.77430630e-01 -1.69288278e-01
-2.03280330e-01 9.38959777e-01 2.07659565e-02 -9.47037995e-01
1.21077530e-01 3.88485156e-02 1.88106984e-01 -1.07886052e+00
6.93884194e-01 -1.20046449e+00 1.60168961e-01 3.74323964e-01
-3.86272430e-01 -6.27285838e-01 2.27368265e-01 4.96839285e-01
-5.87092578e-01 -4.30609494e-01 1.11166012e+00 -9.01419744e-02
-2.51178920e-01 -1.16536625e-01 -6.43511653e-01 -2.12189600e-01
3.36777806e-01 -1.85596436e-01 -6.87925320e-04 -8.30372632e-01
-1.18205798e+00 -3.76000732e-01 -6.67084396e-01 3.72732170e-02
5.48195541e-01 -1.08076465e+00 -5.99339008e-01 7.24272072e-01
-6.83197737e-01 -1.50649175e-01 7.46616602e-01 6.10397696e-01
-3.95650685e-01 4.93639976e-01 5.97065501e-02 -4.39873904e-01
-9.30499256e-01 4.05157655e-01 7.05671906e-01 1.62161916e-01
-5.42262316e-01 7.35842288e-01 -2.09619239e-01 -9.43983570e-02
4.43149775e-01 -6.12242937e-01 -3.02020162e-01 3.14007282e-01
7.11291492e-01 5.25447667e-01 3.49586248e-01 -4.69413906e-01
2.23770496e-02 3.81622165e-01 1.79388374e-01 -3.21147740e-01
1.45490825e+00 -4.80094045e-01 -1.72966602e-03 4.21289682e-01
1.76212263e+00 2.15379536e-01 -1.28876591e+00 -8.82526711e-02
-4.52793628e-01 -3.60811353e-01 6.84614420e-01 -5.88458955e-01
-1.11307430e+00 9.41684008e-01 6.32868469e-01 9.19124186e-01
1.50144327e+00 -7.06301332e-01 1.04975879e+00 1.59685016e-01
4.77006026e-02 -1.59695947e+00 7.75569351e-03 5.75725079e-01
1.04880404e+00 -7.69618332e-01 -3.09363633e-01 -2.39792541e-01
-1.28655925e-01 1.49869978e+00 -5.54948859e-02 -1.21985741e-01
9.79586840e-01 2.71770269e-01 1.92317784e-01 2.56911725e-01
-3.39797825e-01 1.70394517e-02 1.72333345e-01 6.54054880e-01
3.77446026e-01 -1.08990679e-02 -4.01024550e-01 6.53071284e-01
-4.76216227e-01 8.52914453e-02 5.01924992e-01 4.26530957e-01
-7.43319094e-01 -7.23156333e-01 -8.88271809e-01 2.43556812e-01
-6.21405840e-01 -6.76948801e-02 2.92434096e-01 4.56490099e-01
-1.68476686e-01 1.49709654e+00 5.76230772e-02 -2.19427466e-01
4.27556664e-01 2.34552726e-01 1.60076711e-02 -7.79291382e-03
-5.88147342e-01 8.93127561e-01 3.15405071e-01 1.52888950e-02
-1.09894261e-01 -1.48636729e-01 -1.10644472e+00 -2.19231963e-01
-7.39006758e-01 3.46100062e-01 1.09932220e+00 9.11314964e-01
-6.28321916e-02 9.68279719e-01 1.09391010e+00 -8.01454782e-01
-8.53735507e-01 -1.18500578e+00 -1.05562305e+00 2.56028920e-01
1.07688129e+00 -2.54196376e-01 -9.54955280e-01 -3.76979820e-02] | [15.026348114013672, 5.897350788116455] |
8239fdf0-8aba-427a-b0e6-86504f2be455 | a-physics-informed-ai-method-for-calculating | 2306.13345 | null | https://arxiv.org/abs/2306.13345v1 | https://arxiv.org/pdf/2306.13345v1.pdf | A physics-informed AI method for calculating melting points with uncertainty control and optimal sampling | We present an artificial intelligence (AI) method for automatically computing the melting point based on coexistence simulations in the NPT ensemble. Given the interatomic interaction model, the method makes decisions regarding the number of atoms and temperature at which to conduct simulations, and based on the collected data predicts the melting point along with the uncertainty, which can be systematically improved with more data. We demonstrate how incorporating physical models of the solid-liquid coexistence evolution enhances the AI method's accuracy and enables optimal decision-making to effectively reduce predictive uncertainty. To validate our approach, we compare our results with approximately 20 melting point calculations from the literature. Remarkably, we observe significant deviations in about one-third of the cases, underscoring the need for accurate and reliable AI-based algorithms for materials property calculations. | ['Alexander Shapeev', 'Timofei Miryashkin', 'Olga Klimanova'] | 2023-06-23 | null | null | null | null | ['decision-making'] | ['reasoning'] | [ 2.16418043e-01 -1.37358412e-01 -3.06190878e-01 -1.62594751e-01
-5.07991552e-01 -2.45263249e-01 4.90434080e-01 4.09856260e-01
-3.58458191e-01 9.70335960e-01 -1.56813949e-01 -4.92352575e-01
-2.54647970e-01 -9.10281956e-01 -6.14415407e-01 -1.06503201e+00
4.34195548e-02 1.18029320e+00 -1.09214783e-01 -2.97893226e-01
5.34986377e-01 4.23386395e-01 -1.94396186e+00 -5.94201647e-02
1.75237501e+00 1.08353961e+00 -1.15958832e-01 4.17389601e-01
-1.46273062e-01 2.43405759e-01 -2.93127477e-01 -1.48737609e-01
2.66500175e-01 -3.10823888e-01 -7.37894952e-01 -3.92688453e-01
9.61659662e-03 -6.68520704e-02 6.48226812e-02 1.00105000e+00
-4.36368063e-02 2.36699909e-01 1.25585186e+00 -1.05348432e+00
-3.62929165e-01 8.13392282e-01 -3.38972330e-01 -1.16024621e-01
3.46919686e-01 6.65431321e-01 1.07756186e+00 -3.26731354e-01
4.91031677e-01 8.01573455e-01 5.10793805e-01 3.33629787e-01
-1.29009223e+00 -4.18301553e-01 -3.39215040e-01 9.38207284e-02
-1.45523739e+00 -1.94262341e-01 7.50701129e-01 -4.30061489e-01
1.22743165e+00 5.09922028e-01 1.00321198e+00 4.07048345e-01
6.81735039e-01 1.78917170e-01 1.16021669e+00 -9.74186540e-01
7.66451836e-01 -2.00480479e-03 2.35499695e-01 2.05777198e-01
6.68022990e-01 3.70889306e-01 -1.10084891e-01 -3.97384435e-01
2.20055208e-01 -3.76333028e-01 1.46023542e-01 -2.28914320e-01
-8.41496527e-01 4.76933688e-01 1.25654623e-01 4.62074101e-01
-4.89938229e-01 1.67612597e-01 2.29694054e-01 1.53955743e-01
5.99170625e-01 1.04332387e+00 -4.75542575e-01 -1.40255854e-01
-1.04638994e+00 7.03674316e-01 8.64109159e-01 4.56920892e-01
6.94524109e-01 -1.31638154e-01 3.36600274e-01 3.99269044e-01
5.19035280e-01 7.67110765e-01 1.45994127e-01 -1.05844080e+00
1.45062611e-01 7.34149456e-01 4.89841342e-01 -4.27850485e-01
-3.42494458e-01 1.10444866e-01 -5.16178608e-01 3.50520402e-01
5.28198361e-01 3.21358331e-02 -1.05589938e+00 1.10262036e+00
1.96025163e-01 -2.00152233e-01 9.80131999e-02 6.50205135e-01
2.88603574e-01 6.57653689e-01 2.49364406e-01 -7.17267096e-01
9.76918161e-01 -4.48273748e-01 -6.37689173e-01 1.74008608e-01
6.74564123e-01 -4.43891168e-01 7.78288782e-01 6.67360485e-01
-1.07758594e+00 -2.57420391e-01 -1.45492291e+00 2.28574932e-01
-4.01868165e-01 -6.31859362e-01 9.68044281e-01 9.18294847e-01
-7.20086753e-01 1.37327361e+00 -8.68772209e-01 -1.54817596e-01
-2.64410794e-01 8.57579291e-01 1.26101807e-01 2.42801502e-01
-1.34241223e+00 1.14052749e+00 6.73597872e-01 -1.12223245e-01
-2.68868774e-01 -8.33226979e-01 -6.80614769e-01 -2.16756642e-01
5.98531477e-02 -5.52014887e-01 1.22219372e+00 -7.41213500e-01
-1.58424938e+00 6.99296653e-01 -2.79717714e-01 -3.95064920e-01
3.16725641e-01 2.66715854e-01 -5.80657661e-01 7.56465867e-02
-3.48352075e-01 5.56887150e-01 5.13849735e-01 -1.35626674e+00
-2.01869622e-01 -3.42302084e-01 -3.74721348e-01 -6.05820604e-02
1.73763976e-01 -2.06156313e-01 1.75820947e-01 -3.68410982e-02
2.58519232e-01 -9.87694919e-01 -5.16551971e-01 -6.87431216e-01
-2.55160391e-01 -2.84034163e-01 4.77655649e-01 -6.47938788e-01
1.20177710e+00 -1.55069935e+00 2.86895782e-01 8.17032516e-01
2.02537924e-01 -1.53946385e-01 2.71326810e-01 7.23824739e-01
-1.17610060e-02 3.01855683e-01 -5.79859674e-01 -4.42794990e-03
9.56462026e-02 2.73660887e-02 -8.47014636e-02 3.29330891e-01
5.09459227e-02 8.84621978e-01 -5.60270965e-01 -3.81102890e-01
4.99438792e-01 -1.73730180e-02 -5.22670269e-01 1.07044950e-01
-9.33520794e-01 6.30856931e-01 -4.13932651e-01 7.87557423e-01
8.29434752e-01 2.78230011e-02 4.40414816e-01 9.67560615e-03
-4.70152557e-01 3.17558587e-01 -7.28373706e-01 1.24048567e+00
-2.76550978e-01 7.46262670e-02 -2.57742018e-01 -7.69807339e-01
1.15151680e+00 1.84380591e-01 8.34046245e-01 -7.56720066e-01
3.65410984e-01 4.90081489e-01 6.40113950e-01 -3.67135882e-01
6.29136980e-01 -5.68185270e-01 -1.69457227e-01 7.76466727e-01
-1.49672955e-01 -9.21545088e-01 2.38210425e-01 -2.04132989e-01
6.68694615e-01 1.76453963e-01 2.30124965e-01 -8.97553623e-01
5.05515993e-01 1.86289459e-01 3.13972533e-01 7.19974220e-01
-8.44972432e-02 2.79341757e-01 3.96778226e-01 -7.36120105e-01
-1.70139408e+00 -6.67630076e-01 -5.40202379e-01 6.14800394e-01
1.98987409e-01 -4.42185998e-01 -1.20527911e+00 5.49222119e-02
4.02655691e-01 1.18400931e+00 -6.24103487e-01 -2.62604058e-01
-3.72945040e-01 -1.14010549e+00 -8.27210397e-02 3.16812575e-01
2.31750399e-01 -1.13830662e+00 -3.22073460e-01 3.15869451e-01
1.15432382e-01 -6.93447113e-01 1.73246875e-01 3.98192734e-01
-9.38583851e-01 -7.00821757e-01 -9.08804387e-02 8.72326940e-02
4.33307827e-01 -4.56012160e-01 1.11402225e+00 4.62355018e-01
-3.61523122e-01 1.41293174e-02 -8.54047462e-02 -5.12563288e-01
-1.15474033e+00 1.25880882e-01 4.40183342e-01 -6.47527993e-01
6.89248621e-01 -8.80930960e-01 -5.12085795e-01 9.42682922e-02
-7.25269854e-01 -1.06947301e-02 1.61171794e-01 3.51363719e-01
4.07239228e-01 3.73826653e-01 2.81694293e-01 -6.14953220e-01
5.98935962e-01 -3.07283789e-01 -7.96524882e-01 3.85811716e-01
-1.28968108e+00 6.75018668e-01 6.85742676e-01 2.53169041e-04
-9.50839639e-01 -3.82787064e-02 -5.21432720e-02 -8.25878307e-02
-1.45767152e-01 2.95250803e-01 -1.75535172e-01 -1.91094726e-01
4.69850302e-01 5.92946485e-02 -8.03594515e-02 -3.05812269e-01
1.74055934e-01 8.16281855e-01 3.82937133e-01 -1.26394820e+00
4.45248067e-01 1.95864245e-01 2.35588908e-01 -6.50119305e-01
-2.93412298e-01 2.12004855e-02 -1.19133174e+00 -3.61022770e-01
7.92191327e-01 -2.44394317e-01 -1.25795329e+00 3.91924769e-01
-7.62695432e-01 -1.94291905e-01 -2.16688141e-01 3.51485789e-01
-7.36257792e-01 3.70136142e-01 -4.43528175e-01 -1.34268701e+00
-6.99608564e-01 -1.31167924e+00 8.91396165e-01 3.84374738e-01
-7.21111178e-01 -1.00384998e+00 2.71346956e-01 5.58004439e-01
2.31692329e-01 3.36634904e-01 1.42740846e+00 -5.16431153e-01
-5.46748102e-01 -1.77240148e-01 2.70065099e-01 -9.10314247e-02
9.95949507e-02 7.33006775e-01 -9.31501389e-01 -2.83447113e-02
-1.44398317e-01 -3.36990729e-02 6.90316379e-01 5.73932350e-01
1.08034611e+00 -2.63776351e-02 -5.42911470e-01 1.32155403e-01
1.44543552e+00 5.90292037e-01 7.65051723e-01 4.29452181e-01
4.12507385e-01 6.23040497e-01 6.15897000e-01 5.31562209e-01
-4.82632667e-02 6.02414846e-01 2.76391745e-01 5.15965939e-01
6.00028694e-01 -1.20609574e-01 -4.42067832e-02 8.63687217e-01
-7.53085673e-01 -4.84237298e-02 -1.34586036e+00 2.18417891e-03
-1.78543842e+00 -7.77509749e-01 -4.84790429e-02 2.48610163e+00
9.32628036e-01 3.91306102e-01 2.74972409e-01 2.83352017e-01
4.70744491e-01 -1.30283102e-01 -9.03127730e-01 -9.34677005e-01
2.62414664e-01 4.05186266e-01 6.34530723e-01 6.85807407e-01
-7.31590509e-01 9.56566095e-01 8.25284767e+00 7.34394312e-01
-1.11113465e+00 -3.89072239e-01 9.29696143e-01 7.13422894e-02
-7.05083787e-01 1.56281292e-01 -4.53785449e-01 6.40661776e-01
1.53230894e+00 -4.85825270e-01 7.73767889e-01 4.28517938e-01
2.63888508e-01 -5.70467770e-01 -1.16429317e+00 4.22346383e-01
-5.42669356e-01 -1.30998027e+00 -5.96271493e-02 1.45513594e-01
7.11413085e-01 -3.13585877e-01 -3.29054631e-02 -1.12410322e-01
4.26267087e-01 -1.11209679e+00 6.15870297e-01 6.29349589e-01
7.20606089e-01 -1.19660950e+00 6.68782055e-01 3.32766324e-01
-8.56549323e-01 1.33058857e-02 -3.84607136e-01 -4.71160024e-01
-2.48869509e-02 8.47382724e-01 -8.96140695e-01 6.72603607e-01
4.92005408e-01 1.84605524e-01 -9.25360247e-02 6.52823269e-01
1.46245390e-01 5.93056917e-01 -6.65239215e-01 -3.91661674e-01
1.12235114e-01 -1.01766467e+00 2.60486513e-01 6.60438180e-01
2.70457089e-01 3.66230398e-01 -2.43417293e-01 1.28075385e+00
2.25843713e-01 -4.31843624e-02 -4.03362244e-01 -3.11975449e-01
6.50494099e-01 8.07538331e-01 -8.69659603e-01 -2.14741886e-01
4.08988297e-02 4.02447820e-01 1.21001944e-01 1.60969511e-01
-5.01347601e-01 -9.86786038e-02 6.59781396e-01 1.36049554e-01
-6.38606027e-02 -4.31778640e-01 -8.61747503e-01 -8.93681824e-01
-8.69370326e-02 -7.27669120e-01 2.98113562e-02 -6.37935519e-01
-1.08583450e+00 1.71325222e-01 2.66186178e-01 -6.54342115e-01
-5.21036744e-01 -7.77762473e-01 -9.16770875e-01 6.48371160e-01
-8.47392738e-01 -6.65533662e-01 3.57911825e-01 -2.56031096e-01
-2.20923513e-01 1.96355611e-01 7.95541465e-01 -2.83369184e-01
-5.05567312e-01 2.54853487e-01 6.80434346e-01 -5.06534994e-01
2.59608895e-01 -1.23099625e+00 5.93273342e-01 1.56334624e-01
-3.87292594e-01 5.06865203e-01 1.43127167e+00 -9.21821058e-01
-1.65637302e+00 -3.36250871e-01 5.55922329e-01 -5.26460350e-01
7.57649004e-01 -3.58440131e-01 -9.31272686e-01 2.46539205e-01
1.22802563e-01 -4.27103490e-01 6.27039790e-01 3.00866514e-01
5.15948720e-02 3.92874815e-02 -1.54050672e+00 6.18640721e-01
8.96193385e-01 -2.02806443e-01 -5.36264598e-01 3.14214140e-01
8.19657326e-01 -1.56573668e-01 -1.24683619e+00 4.82035756e-01
6.65504098e-01 -8.30335915e-01 7.66593397e-01 -5.45666099e-01
5.42611063e-01 3.47796269e-02 -1.14403009e-01 -1.10063326e+00
-1.76014960e-01 -4.48218524e-01 5.48195615e-02 1.02104568e+00
6.27307892e-01 -8.45610619e-01 7.17066228e-01 1.85605419e+00
2.16278791e-01 -7.03254163e-01 -1.18692863e+00 -7.68627048e-01
7.36355245e-01 -4.57629412e-01 1.08768988e+00 5.79152167e-01
5.57969809e-01 -2.97988951e-01 -1.58899292e-01 -1.63250323e-02
8.55946302e-01 5.19195855e-01 2.23102257e-01 -1.54588997e+00
-7.46204928e-02 -2.73103774e-01 -1.95599839e-01 -2.33595550e-01
3.97875786e-01 -6.09369636e-01 -2.15644725e-02 -7.83103168e-01
4.27008569e-01 -7.32374251e-01 -2.42670551e-01 2.16165051e-01
9.76395905e-02 -1.59941882e-01 2.09106365e-03 2.68903822e-01
-3.54434967e-01 5.63178003e-01 1.13769066e+00 -1.20976716e-01
-2.81968832e-01 -2.86083907e-01 -4.50069398e-01 6.28567636e-01
8.38694751e-01 -3.18756282e-01 1.21777445e-01 4.40305889e-01
3.48917514e-01 1.12150675e-02 -2.05304459e-01 -1.21047747e+00
-1.26404867e-01 -5.85186779e-01 3.98652554e-01 -5.36481082e-01
1.84913814e-01 -8.63717616e-01 4.70085293e-01 5.56987822e-01
-2.34839141e-01 -2.56423771e-01 3.42442751e-01 1.69136241e-01
1.91519395e-01 -6.24416173e-01 6.56764090e-01 -4.95035648e-02
-2.59746522e-01 1.19506814e-01 -6.19807959e-01 -5.09000123e-01
1.08535624e+00 -2.18475819e-01 -3.39581594e-02 -2.42679045e-01
-5.57281613e-01 8.77649933e-02 1.14936459e+00 -4.37858924e-02
1.67802110e-01 -1.13496590e+00 -3.84139121e-01 1.65581942e-01
-1.13948865e-03 -1.60334989e-01 1.68517679e-01 5.23142636e-01
-9.64757264e-01 6.53332591e-01 -1.08495392e-01 -4.90675330e-01
-9.32839751e-01 8.53611887e-01 5.07486045e-01 -3.67327124e-01
-3.54346037e-01 8.17223638e-02 -3.23660135e-01 -3.94766599e-01
-5.31475663e-01 -6.32548034e-01 2.17519179e-01 -3.83958459e-01
3.01490128e-01 4.05130595e-01 2.63577789e-01 -4.00420815e-01
-3.17312777e-01 7.29525030e-01 -2.08195999e-01 -2.58439898e-01
1.29233265e+00 1.08419791e-01 -4.61422622e-01 6.15074575e-01
6.58495188e-01 -1.73324391e-01 -1.15903437e+00 2.23746628e-01
2.06841618e-01 -2.57312596e-01 1.16253354e-01 -7.09685624e-01
-6.15749359e-01 6.16832733e-01 2.95295656e-01 2.98390329e-01
8.22639525e-01 -3.65896039e-02 7.57171631e-01 4.48745906e-01
7.46689200e-01 -1.46160722e+00 -6.73562348e-01 3.30917269e-01
5.78551412e-01 -1.22066855e+00 5.16734421e-01 -4.61457640e-01
-4.70431864e-01 1.22584140e+00 6.91431284e-01 6.99496567e-02
5.74185491e-01 5.92674851e-01 -2.43545443e-01 -1.27070755e-01
-8.27833056e-01 1.81998730e-01 2.02264294e-01 3.55530828e-01
4.06500846e-01 4.69221205e-01 -3.82231891e-01 4.30841178e-01
-4.05885696e-01 3.90946902e-02 2.40050986e-01 9.25764918e-01
-6.99214637e-01 -1.48327315e+00 -6.94076836e-01 6.24922097e-01
-1.74238794e-02 -1.21733218e-01 -5.62727451e-01 7.00082004e-01
1.55538693e-01 8.11631024e-01 3.04606348e-01 -4.27632362e-01
-2.70109683e-01 5.07452428e-01 7.87454784e-01 -6.52710423e-02
-4.23830777e-01 -3.25061977e-01 1.27685383e-01 -1.80502027e-01
-4.06681091e-01 -9.02784050e-01 -1.79342782e+00 -9.34806228e-01
-5.18720150e-01 6.40703142e-01 6.50095642e-01 1.22364545e+00
2.72572368e-01 9.92139895e-03 4.77756053e-01 -1.09288037e+00
-2.95215100e-01 -8.21165621e-01 -9.35442269e-01 3.26758772e-01
-8.42803270e-02 -7.60297120e-01 -5.12261629e-01 -4.33313429e-01] | [5.5447773933410645, 4.693122386932373] |
2efddcd0-3c7f-47f6-831e-6676f81c0c19 | scene-text-magnifier | 1907.00693 | null | https://arxiv.org/abs/1907.00693v2 | https://arxiv.org/pdf/1907.00693v2.pdf | Scene Text Magnifier | Scene text magnifier aims to magnify text in natural scene images without recognition. It could help the special groups, who have myopia or dyslexia to better understand the scene. In this paper, we design the scene text magnifier through interacted four CNN-based networks: character erasing, character extraction, character magnify, and image synthesis. The architecture of the networks are extended based on the hourglass encoder-decoders. It inputs the original scene text image and outputs the text magnified image while keeps the background unchange. Intermediately, we can get the side-output results of text erasing and text extraction. The four sub-networks are first trained independently and fine-tuned in end-to-end mode. The training samples for each stage are processed through a flow with original image and text annotation in ICDAR2013 and Flickr dataset as input, and corresponding text erased image, magnified text annotation, and text magnified scene image as output. To evaluate the performance of text magnifier, the Structural Similarity is used to measure the regional changes in each character region. The experimental results demonstrate our method can magnify scene text effectively without effecting the background. | ['Toshiki Nakamura', 'Seiichi Uchida', 'Anna Zhu'] | 2019-06-17 | null | null | null | null | ['text-annotation'] | ['natural-language-processing'] | [ 8.28425825e-01 4.32542339e-02 1.17001779e-01 -4.13570970e-01
1.43866181e-01 -8.52356702e-02 5.73905945e-01 -6.27201200e-01
-7.14178503e-01 5.35023987e-01 5.29205561e-01 -1.91652298e-01
4.41677660e-01 -8.66217256e-01 -7.34018922e-01 -5.79702795e-01
8.20061982e-01 1.05815187e-01 4.88936752e-01 -6.25274703e-02
4.29793864e-01 2.15744480e-01 -1.59854245e+00 8.63446593e-01
1.20260370e+00 6.93544626e-01 7.29394615e-01 1.16449332e+00
-3.55889231e-01 1.25502610e+00 -8.77568901e-01 -4.72741991e-01
2.40499839e-01 -4.11852598e-01 -3.30076307e-01 3.77146274e-01
9.50789988e-01 -8.92671824e-01 -1.09268475e+00 1.39191258e+00
7.01240540e-01 3.49588580e-02 5.28672755e-01 -7.53803134e-01
-1.15615594e+00 9.98999536e-01 -8.27688634e-01 2.75666535e-01
-1.47089839e-01 4.14138615e-01 1.15204290e-01 -1.03029394e+00
5.83456695e-01 1.37180507e+00 3.22695613e-01 7.48801053e-01
-5.69463134e-01 -6.18308425e-01 7.56271109e-02 1.37927070e-01
-1.10309005e+00 -4.77846861e-01 4.58495647e-01 -4.54760730e-01
7.63068914e-01 3.63938630e-01 8.37692618e-01 6.69260740e-01
3.64769131e-01 1.36199427e+00 8.97414029e-01 -3.41453761e-01
-3.13016981e-01 1.28532156e-01 -1.23596992e-02 7.47355461e-01
1.88826278e-01 -1.40194362e-02 -3.44726443e-01 9.19618368e-01
8.86567533e-01 2.41311446e-01 -5.00585318e-01 3.22635084e-01
-1.44964957e+00 2.83595413e-01 2.75815547e-01 5.38291708e-02
-3.57280113e-02 2.79892892e-01 4.59317297e-01 5.70762977e-02
3.05882871e-01 1.50471136e-01 -2.43818313e-01 -2.55815983e-01
-1.14272118e+00 8.94882604e-02 1.98561281e-01 1.28137434e+00
5.45931578e-01 4.75118786e-01 -5.42607367e-01 9.09085751e-01
7.77033865e-02 1.03894997e+00 8.18252265e-01 -2.53915846e-01
8.71960878e-01 7.49837279e-01 -3.23833168e-01 -8.49206686e-01
-3.55699390e-01 -1.14032634e-01 -9.88877058e-01 3.22770804e-01
3.47448550e-02 -5.35249889e-01 -1.61891186e+00 8.47637713e-01
6.63446262e-02 4.89519536e-02 1.23416889e-03 9.84639168e-01
1.33364093e+00 9.35951471e-01 -1.89773649e-01 6.18512742e-02
1.32280743e+00 -1.47741914e+00 -1.10869384e+00 -3.21121782e-01
4.28239822e-01 -1.09113503e+00 1.54023910e+00 5.14000535e-01
-1.27294600e+00 -6.93332136e-01 -1.13650107e+00 -6.19946063e-01
-5.25926948e-01 9.53779519e-01 1.10540077e-01 5.90912104e-01
-8.65264535e-01 2.71883845e-01 -5.89517713e-01 -1.31314561e-01
7.07488120e-01 2.92639256e-01 -4.54821363e-02 6.42764047e-02
-9.95143116e-01 7.63773024e-01 7.91098535e-01 2.26960361e-01
-6.00215971e-01 -6.88490093e-01 -7.54205108e-01 2.37561420e-01
2.26913452e-01 -5.77748060e-01 1.17358446e+00 -1.08921838e+00
-1.78718770e+00 9.38724995e-01 1.11440323e-01 -2.78189629e-01
8.59203458e-01 -2.84340709e-01 -6.23346627e-01 1.28825381e-01
-1.02447376e-01 7.67588258e-01 1.22826612e+00 -7.90036201e-01
-1.21449947e+00 -4.90348488e-02 -3.18112552e-01 6.24174893e-01
-3.88902515e-01 1.30609199e-01 -9.78526652e-01 -9.41933572e-01
-2.75960565e-01 -4.69677716e-01 6.21525897e-03 -1.02114700e-01
-9.04107273e-01 3.94531935e-01 1.50138760e+00 -9.80271399e-01
1.40580463e+00 -2.23508286e+00 2.79461052e-02 -2.30243742e-01
5.22416174e-01 6.76708996e-01 -9.03235003e-02 -1.49634294e-02
-5.82102239e-02 7.09146261e-02 -1.88050032e-01 -3.07518780e-01
-2.24601641e-01 -1.44252807e-01 -4.42347288e-01 2.88039744e-01
9.44145992e-02 1.17724586e+00 -5.73775411e-01 -6.78957701e-01
8.93100560e-01 3.57780159e-01 -4.48674709e-01 -8.34104344e-02
-4.08534944e-01 1.36905938e-01 -1.10979453e-01 4.32672143e-01
1.03566897e+00 8.19523446e-03 -4.45998132e-01 -3.68035913e-01
-4.31802899e-01 -2.68579274e-01 -1.08455920e+00 1.24143386e+00
-1.28156602e-01 1.40050197e+00 -2.96087980e-01 -5.92466950e-01
6.67626619e-01 -2.64155477e-01 -1.31191656e-01 -1.08594000e+00
4.14472550e-01 -7.41118789e-02 -3.89358662e-02 -1.07671034e+00
1.05573237e+00 3.99840772e-01 1.52550370e-01 3.07732314e-01
-3.33269775e-01 -3.52667004e-01 3.31901461e-01 1.41200796e-01
5.13939857e-01 2.60775886e-03 -1.97129235e-01 5.66657074e-02
5.70325553e-01 1.02459602e-02 8.33333936e-03 6.26843691e-01
2.15836883e-01 8.54915917e-01 4.73672211e-01 -5.47590315e-01
-1.45895100e+00 -7.36385047e-01 -2.16314107e-01 1.05042255e+00
4.43363756e-01 -2.82897562e-01 -9.69669521e-01 -4.98627514e-01
-3.49338949e-01 7.62245297e-01 -8.29166353e-01 -1.75468862e-01
-8.56496453e-01 -5.05580723e-01 5.85494220e-01 4.30362791e-01
1.40657914e+00 -1.23663116e+00 -6.37176156e-01 -2.17266157e-01
-1.99184209e-01 -9.79407489e-01 -9.80958402e-01 -3.23631167e-01
-4.47045743e-01 -7.15364754e-01 -1.15148807e+00 -1.10266316e+00
8.43986452e-01 3.52080107e-01 4.59571451e-01 1.20790556e-01
-5.40158272e-01 -8.78806189e-02 -2.14706346e-01 -7.54052877e-01
-3.65044534e-01 -3.86827104e-02 -2.94577360e-01 1.54217109e-01
1.83027327e-01 -2.53064185e-01 -7.23149538e-01 3.15790661e-02
-1.25364423e+00 9.70752835e-01 8.85203600e-01 5.73228180e-01
5.35614550e-01 2.99026281e-01 -7.14396164e-02 -1.00553048e+00
6.46551549e-01 2.16250941e-01 -6.80772007e-01 1.72483429e-01
-3.26623827e-01 -1.97337136e-01 8.15699220e-01 -6.87051356e-01
-1.33023715e+00 8.69435593e-02 3.75114381e-02 -4.83816892e-01
-1.27195224e-01 3.55133504e-01 -2.69538075e-01 7.59830773e-02
6.51266694e-01 8.82403910e-01 -3.21810007e-01 -2.59167731e-01
6.76866472e-01 1.15879369e+00 1.02194405e+00 2.73446292e-01
7.63327003e-01 5.32225788e-01 -4.90723282e-01 -1.19460380e+00
-6.96962774e-01 1.24678034e-02 -5.67434371e-01 -4.94500637e-01
1.22630060e+00 -7.83988357e-01 -5.48885167e-01 9.75459039e-01
-1.18282783e+00 -6.09821737e-01 -3.44510436e-01 4.54216450e-01
-4.78688270e-01 5.02778590e-01 -3.93146008e-01 -2.89534509e-01
-5.99501669e-01 -1.10052514e+00 1.21975958e+00 7.98711240e-01
5.48782229e-01 -6.15353882e-01 -3.99354279e-01 2.56821454e-01
3.64628345e-01 -5.36905415e-02 9.05139506e-01 -2.33482689e-01
-6.42531991e-01 -2.23139122e-01 -7.23923266e-01 3.90576601e-01
-3.85026969e-02 3.39151293e-01 -1.00308418e+00 1.95547134e-01
-4.41230625e-01 1.18000664e-01 1.22035468e+00 6.70818925e-01
1.53874278e+00 -5.22057533e-01 -1.44480944e-01 1.07346189e+00
1.16281104e+00 4.64358926e-01 1.32234430e+00 2.07457468e-01
1.17488861e+00 2.36470208e-01 3.99083823e-01 2.75618047e-01
2.39403769e-01 2.40243107e-01 2.41729766e-01 -5.72383285e-01
-7.77659535e-01 -3.86697203e-01 2.74054915e-01 6.83425605e-01
9.37033668e-02 -6.22252941e-01 -7.99125314e-01 3.68983388e-01
-1.75008428e+00 -1.28344440e+00 -4.86037821e-01 1.84011865e+00
8.91242683e-01 1.18068121e-01 -2.31811181e-01 -1.46261886e-01
1.09968865e+00 1.37450740e-01 -8.11949313e-01 -2.50232488e-01
-5.09566188e-01 -1.78776085e-02 7.73712516e-01 4.63676661e-01
-1.17629588e+00 1.68189049e+00 5.76953030e+00 9.96521294e-01
-1.59999490e+00 -3.59549403e-01 6.72124445e-01 -3.40865016e-01
-3.08439173e-02 -4.03237313e-01 -8.36570203e-01 8.01022947e-01
2.26133332e-01 -2.24836886e-01 7.03704536e-01 7.08710313e-01
3.04466993e-01 -2.34101668e-01 -6.32901430e-01 1.33957028e+00
4.09785986e-01 -1.61207926e+00 4.58172232e-01 -2.85277426e-01
1.11233640e+00 9.86471325e-02 3.66828799e-01 2.28162095e-01
7.08305612e-02 -1.19880986e+00 7.68298268e-01 9.20652151e-01
1.49499762e+00 -6.24540627e-01 3.66419971e-01 3.24974358e-01
-9.35757220e-01 -1.71784371e-01 -6.69654787e-01 4.04233299e-02
-8.97616148e-03 4.96857435e-01 -8.33695710e-01 6.13960065e-03
5.77061176e-01 8.00349414e-01 -7.54736304e-01 1.18436277e+00
-3.38187814e-01 4.93110657e-01 -4.25957479e-02 -4.15070266e-01
2.33877823e-02 -1.73065960e-01 4.75663781e-01 1.43481839e+00
5.32599650e-02 1.18819565e-01 -2.27049693e-01 7.69480884e-01
-2.67564684e-01 2.26562440e-01 -4.04149652e-01 -3.69434416e-01
1.49614766e-01 1.26675141e+00 -7.64282525e-01 -9.65708017e-01
-3.75340492e-01 1.37364674e+00 -1.77290425e-01 5.84315538e-01
-9.26023304e-01 -1.32580268e+00 -1.36489779e-01 5.01297005e-02
3.63447040e-01 1.22410774e-01 -6.41404152e-01 -1.37884188e+00
-8.50194767e-02 -9.65929627e-01 -1.00375034e-01 -1.46833050e+00
-5.36818504e-01 4.22040403e-01 -2.52876461e-01 -1.11638856e+00
4.24429357e-01 -8.25041950e-01 -7.50507772e-01 8.82940829e-01
-1.17152512e+00 -1.22520101e+00 -7.96235144e-01 6.65432215e-01
1.07902563e+00 -2.91929215e-01 -3.98684815e-02 4.79220599e-01
-8.51870894e-01 7.37121284e-01 3.00621420e-01 5.12672186e-01
7.22841322e-01 -1.05770254e+00 5.98169029e-01 1.16769981e+00
-3.00911874e-01 1.47388652e-01 5.23793638e-01 -1.17641199e+00
-1.06133461e+00 -1.40973806e+00 4.14335668e-01 -2.55595088e-01
2.73681074e-01 -4.84639853e-01 -5.45236826e-01 7.42094934e-01
4.26954836e-01 -3.38864446e-01 -1.59437642e-01 -7.64382899e-01
2.76387155e-01 2.46533304e-02 -8.14696670e-01 1.18040586e+00
9.97346401e-01 -1.04701273e-01 -4.83352095e-01 6.32355094e-01
1.05773258e+00 -1.07595360e+00 -2.83923090e-01 1.82448998e-01
5.64271808e-01 -7.70367205e-01 6.92274213e-01 -4.62121606e-01
1.18059278e+00 -4.62749302e-01 1.58680335e-01 -1.05003285e+00
-9.15915519e-02 -5.01015067e-01 1.50881156e-01 8.80231023e-01
4.57044244e-01 -4.03476149e-01 7.88583398e-01 3.71751696e-01
-4.89082336e-01 -5.74853122e-01 -2.98135579e-01 1.44997463e-02
-1.07729383e-01 -2.54626542e-01 7.82718658e-01 8.56247723e-01
-1.89783588e-01 3.28519255e-01 -7.72058845e-01 3.63286957e-02
1.86006308e-01 -1.30637884e-01 1.12416101e+00 -6.22731686e-01
5.73904477e-02 -7.69218862e-01 -3.69839847e-01 -1.38615716e+00
-3.96812141e-01 -6.73024058e-01 6.49390593e-02 -1.68588805e+00
3.06549639e-01 1.79352224e-01 5.94983339e-01 2.86297083e-01
-4.31454599e-01 2.88568586e-01 3.08281451e-01 7.85736144e-02
-5.08090019e-01 6.42488301e-01 2.07858968e+00 -4.91830885e-01
-2.73757875e-01 -1.65396258e-01 -5.52794635e-01 9.32092547e-01
7.53922760e-01 8.26024860e-02 -5.59695125e-01 -9.55269992e-01
2.31654972e-01 -6.09317981e-02 1.36631474e-01 -1.05495644e+00
4.25263315e-01 -3.25169772e-01 1.06626618e+00 -1.22413909e+00
3.67957540e-02 -4.42775220e-01 -4.38509852e-01 4.27653819e-01
-5.94379902e-01 -2.75201440e-01 3.60106051e-01 2.78200150e-01
1.86182275e-01 -3.62781465e-01 1.06249774e+00 -5.00105843e-02
-7.84711599e-01 5.19794047e-01 -5.39081812e-01 9.39365476e-02
9.40366924e-01 -6.93639398e-01 -7.98406780e-01 -4.17891234e-01
-5.56630909e-01 2.47158960e-01 3.28402907e-01 4.52493131e-01
1.05968356e+00 -1.17999542e+00 -7.66296089e-01 4.11869109e-01
-1.93427667e-01 2.28200763e-01 7.52359986e-01 7.23565161e-01
-1.27587938e+00 2.78269738e-01 -3.13027620e-01 -4.77319509e-01
-1.46068585e+00 6.68298662e-01 6.78512812e-01 3.05798680e-01
-1.03222227e+00 8.97024691e-01 6.80209041e-01 -1.83887601e-01
4.41922247e-01 -7.30544448e-01 -4.32200134e-01 -3.63403082e-01
1.11648381e+00 4.13422197e-01 -3.32957238e-01 -3.32833260e-01
4.15702313e-01 7.23845840e-01 -5.17303348e-01 -1.88384384e-01
9.90735412e-01 -3.81895661e-01 -1.88852251e-01 5.60379103e-02
8.78484905e-01 3.78276378e-01 -1.27958584e+00 -1.95659190e-01
-7.46934235e-01 -6.99765623e-01 1.72277521e-02 -1.03371859e+00
-1.23996770e+00 1.22057402e+00 7.52635598e-01 -8.54660273e-02
1.33644176e+00 -4.10179526e-01 8.91352117e-01 4.52178389e-01
-5.99965096e-01 -1.50572371e+00 2.84031540e-01 5.79276800e-01
9.50764656e-01 -9.19799805e-01 8.04954171e-02 -3.12865913e-01
-8.68516624e-01 1.28088427e+00 1.11012757e+00 -6.35289997e-02
2.04754934e-01 4.94965613e-01 1.80116683e-01 -1.51156008e-01
-3.03155541e-01 -2.66034245e-01 3.22650045e-01 6.56531155e-01
-6.17714180e-03 -1.26475558e-01 -1.10910818e-01 3.87288272e-01
-7.50226498e-01 -8.28459412e-02 1.04497015e+00 6.02607965e-01
-8.58179033e-01 -3.16730469e-01 -3.20445210e-01 8.95600080e-01
-3.83637398e-01 -5.75219214e-01 -6.39030457e-01 7.39973366e-01
4.24923956e-01 4.36457038e-01 3.70709687e-01 -6.36623979e-01
3.10321808e-01 -3.41521353e-01 3.37433040e-01 -4.91481781e-01
-3.16577971e-01 1.41260013e-01 -3.96171138e-02 -5.89340851e-02
1.09828472e-01 -2.55722046e-01 -1.47483099e+00 -4.62283641e-01
-4.74425882e-01 -4.61831897e-01 7.03234673e-01 7.83024549e-01
1.83118880e-01 9.03848588e-01 4.11398113e-01 -6.71445787e-01
4.16750200e-02 -1.20735621e+00 -3.74645859e-01 2.03306496e-01
5.04344583e-01 1.85376015e-02 8.07594284e-02 7.27423251e-01] | [11.862859725952148, 2.0419490337371826] |
fcafbb2d-4dd4-48e8-bd7d-74dcd6c858d5 | synthesizing-training-data-for-object | 1702.07836 | null | http://arxiv.org/abs/1702.07836v2 | http://arxiv.org/pdf/1702.07836v2.pdf | Synthesizing Training Data for Object Detection in Indoor Scenes | Detection of objects in cluttered indoor environments is one of the key
enabling functionalities for service robots. The best performing object
detection approaches in computer vision exploit deep Convolutional Neural
Networks (CNN) to simultaneously detect and categorize the objects of interest
in cluttered scenes. Training of such models typically requires large amounts
of annotated training data which is time consuming and costly to obtain. In
this work we explore the ability of using synthetically generated composite
images for training state-of-the-art object detectors, especially for object
instance detection. We superimpose 2D images of textured object models into
images of real environments at variety of locations and scales. Our experiments
evaluate different superimposition strategies ranging from purely image-based
blending all the way to depth and semantics informed positioning of the object
models into real scenes. We demonstrate the effectiveness of these object
detector training strategies on two publicly available datasets, the
GMU-Kitchens and the Washington RGB-D Scenes v2. As one observation, augmenting
some hand-labeled training data with synthetic examples carefully composed onto
scenes yields object detectors with comparable performance to using much more
hand-labeled data. Broadly, this work charts new opportunities for training
detectors for new objects by exploiting existing object model repositories in
either a purely automatic fashion or with only a very small number of
human-annotated examples. | ['Arsalan Mousavian', 'Jana Kosecka', 'Georgios Georgakis', 'Alexander C. Berg'] | 2017-02-25 | null | null | null | null | ['object-detection-in-indoor-scenes'] | ['computer-vision'] | [ 1.99293375e-01 2.35338822e-01 4.63410497e-01 -3.68361920e-01
-4.60912704e-01 -6.78899050e-01 8.59915078e-01 9.90281701e-02
-6.36318147e-01 2.57086396e-01 -3.87654871e-01 -1.29092857e-01
2.12299153e-01 -5.50931156e-01 -1.07755697e+00 -4.47407573e-01
-1.92009106e-01 8.63884687e-01 9.11665738e-01 -3.15481395e-01
-1.13982223e-02 1.01783741e+00 -1.95186710e+00 3.50233078e-01
1.67170301e-01 1.00230372e+00 7.39906728e-01 7.27495670e-01
-6.75782040e-02 4.55237091e-01 -6.11586988e-01 -3.87644507e-02
8.56764615e-01 3.13042819e-01 -4.29526210e-01 6.84070170e-01
5.74311197e-01 -5.59568644e-01 -3.77313554e-01 9.37862217e-01
3.01027685e-01 -5.17765619e-02 4.72068965e-01 -1.37226284e+00
-4.44310695e-01 1.67177960e-01 -3.66964698e-01 2.71806031e-01
3.35751623e-01 6.90057397e-01 5.68951130e-01 -1.18868792e+00
6.66514456e-01 1.34941053e+00 6.76380873e-01 3.97984713e-01
-1.43798113e+00 -4.80912119e-01 2.20223829e-01 4.33600508e-02
-1.33754551e+00 -6.15888238e-01 7.04625726e-01 -5.85964680e-01
1.08397520e+00 2.34567411e-02 7.06973672e-01 1.18092179e+00
-4.32261407e-01 9.93540347e-01 9.69541967e-01 -4.96674329e-01
3.55852276e-01 5.01658320e-01 -1.48828924e-01 6.99897349e-01
5.93182743e-01 3.31565470e-01 -1.51664793e-01 1.31441072e-01
9.63674664e-01 1.87786549e-01 1.46087095e-01 -9.71476376e-01
-1.47236788e+00 4.85095710e-01 7.80257702e-01 1.32252052e-01
-4.22920436e-01 2.78871417e-01 2.54124135e-01 -1.20875482e-02
1.96602106e-01 5.02445877e-01 -6.80435956e-01 4.73139882e-01
-5.52268982e-01 4.55578268e-01 4.63909060e-01 1.38498724e+00
8.08553934e-01 -3.99594903e-02 1.17368668e-01 4.92278874e-01
3.69148850e-01 5.29567838e-01 1.74595535e-01 -1.01094854e+00
2.44776875e-01 6.80951059e-01 5.14659584e-01 -7.10755348e-01
-5.01830041e-01 -4.70768392e-01 -2.14317724e-01 6.28014505e-01
7.24445105e-01 2.36742258e-01 -1.20182693e+00 1.30237091e+00
5.59914947e-01 -1.54756263e-01 1.54083118e-01 9.51020777e-01
8.42036426e-01 4.17764068e-01 3.95853855e-02 5.76893389e-01
1.44443595e+00 -1.02365112e+00 -8.93874019e-02 -7.14541554e-01
4.09387916e-01 -6.78551137e-01 1.09441376e+00 1.58870265e-01
-8.36619437e-01 -7.76923656e-01 -9.65918720e-01 3.66057344e-02
-6.77689612e-01 5.22802949e-01 7.16847718e-01 4.50679630e-01
-8.68626773e-01 1.70706928e-01 -8.60288262e-01 -8.42428088e-01
7.21590340e-01 3.60663563e-01 -6.26089990e-01 -2.64532954e-01
-4.96606588e-01 1.02479982e+00 7.62216985e-01 7.19225965e-04
-1.50835466e+00 -4.19917345e-01 -8.76665294e-01 -1.56657770e-01
5.71225524e-01 -3.90460581e-01 1.51804280e+00 -9.67576802e-01
-8.81110609e-01 1.26613986e+00 3.86073917e-01 -5.35569966e-01
7.47229457e-01 -1.49876744e-01 -2.02024981e-01 5.50262034e-02
3.32239240e-01 1.12182522e+00 7.13868320e-01 -1.73113990e+00
-8.75839591e-01 -3.51673096e-01 2.52627879e-01 -6.39285101e-03
8.24316442e-02 1.32611275e-01 -5.60311019e-01 -2.97362089e-01
4.64756936e-01 -8.24769318e-01 -4.25719827e-01 5.34311295e-01
-5.20397961e-01 -3.94587079e-03 9.53393996e-01 -2.17448190e-01
1.03687547e-01 -1.91840219e+00 -2.56332457e-01 -1.21802479e-01
6.94902688e-02 2.94285148e-01 -1.97072655e-01 3.30930144e-01
1.23493254e-01 -2.92762548e-01 2.61558518e-02 -3.80432993e-01
1.41336143e-01 2.98932076e-01 -1.21828899e-01 5.91194987e-01
4.25023884e-01 9.97141242e-01 -1.00108051e+00 -2.66616076e-01
6.20642662e-01 2.66011685e-01 -5.97649992e-01 2.93262005e-01
-5.49701869e-01 5.17304718e-01 -2.67237604e-01 9.43994462e-01
5.96489489e-01 -8.78371596e-02 7.92381763e-02 -2.87194967e-01
-1.21523134e-01 1.40085295e-01 -1.32129109e+00 1.69835055e+00
-2.75015622e-01 7.61227131e-01 5.50260320e-02 -8.87413442e-01
9.19759929e-01 -9.02151018e-02 2.54711986e-01 -5.52117169e-01
1.05736911e-01 3.31647426e-01 -7.36404508e-02 -7.36055851e-01
5.36793530e-01 -2.87642237e-02 -9.73393843e-02 2.15766802e-01
3.01483512e-01 -3.50483716e-01 3.09979469e-01 -4.88559995e-03
1.23411000e+00 3.27197731e-01 2.17307732e-01 -1.50109425e-01
7.00241551e-02 4.47335064e-01 7.88239762e-02 1.08049428e+00
-3.58790129e-01 6.92478180e-01 -6.85602007e-03 -8.17662656e-01
-1.54784834e+00 -1.14682162e+00 -2.19513059e-01 1.12552571e+00
2.71021724e-01 1.29764929e-01 -5.42931020e-01 -6.23091638e-01
3.40478897e-01 5.97741306e-01 -7.76048958e-01 4.58403416e-02
-5.09375095e-01 -5.08298695e-01 4.68438238e-01 7.13412404e-01
5.50461650e-01 -1.27846396e+00 -1.30688143e+00 1.74405307e-01
2.21865430e-01 -1.48800373e+00 2.99649805e-01 7.06334472e-01
-4.98599738e-01 -9.69771087e-01 -6.19105399e-01 -8.83544028e-01
8.52877438e-01 6.81318760e-01 1.26630878e+00 1.14333034e-02
-7.47526109e-01 4.64873135e-01 -2.50654370e-01 -8.76858056e-01
-3.90667975e-01 -2.11847320e-01 2.77790546e-01 -1.33413076e-01
4.31044489e-01 -2.68425375e-01 -6.05097473e-01 4.72697735e-01
-8.73453021e-01 2.39101589e-01 7.77986467e-01 4.64542776e-01
4.11405802e-01 -2.93328255e-01 1.47901237e-01 -4.70179111e-01
3.84785607e-02 -3.35570157e-01 -9.26188231e-01 9.65278149e-02
1.94089394e-02 -9.44266617e-02 1.81198388e-01 -6.51635170e-01
-8.03510487e-01 5.59991241e-01 1.73141107e-01 -6.03596151e-01
-7.52791584e-01 -1.87293515e-01 -1.20564900e-01 -5.46786226e-02
1.04233801e+00 1.33983940e-01 -2.32794613e-01 -5.12821853e-01
4.97340560e-01 4.69374567e-01 7.75257349e-01 -6.77438617e-01
9.79750812e-01 8.03731084e-01 -3.05165708e-01 -7.07680821e-01
-7.97681332e-01 -7.34440327e-01 -9.31972504e-01 -3.56204927e-01
7.64276326e-01 -9.56958771e-01 -3.79188925e-01 3.18929851e-01
-1.23197424e+00 -6.28839612e-01 -5.38219273e-01 2.72074759e-01
-6.63410485e-01 -1.82049662e-01 -2.19978854e-01 -8.84019554e-01
2.17801467e-01 -1.14060378e+00 1.67225111e+00 6.62363041e-03
1.87246457e-01 -4.35787797e-01 -2.56595075e-01 1.01747185e-01
3.46210390e-01 3.55856329e-01 4.03602898e-01 -7.70208418e-01
-9.34138179e-01 -6.95306420e-01 -5.35135388e-01 2.22369313e-01
3.49968858e-02 -1.14054456e-01 -1.25511014e+00 -1.51413813e-01
-2.44740859e-01 -4.09954816e-01 6.89876020e-01 -5.03027625e-02
9.82976913e-01 2.28659555e-04 -5.79195380e-01 3.14737856e-01
1.48712313e+00 1.75729364e-01 2.90954560e-01 7.88388848e-01
5.75043678e-01 6.57256067e-01 6.18984520e-01 3.89170647e-01
1.69510841e-01 8.45148087e-01 8.34230423e-01 -3.24916869e-01
-1.31641105e-01 -1.14216082e-01 -3.34332511e-02 -1.68464690e-01
1.24593690e-01 -1.75678611e-01 -1.24205756e+00 7.06567705e-01
-1.69666088e+00 -7.92558610e-01 -3.90102901e-02 1.97147226e+00
3.79407793e-01 4.54644084e-01 1.44350499e-01 1.00106858e-01
5.79190791e-01 -3.32612157e-01 -5.72910726e-01 3.59291762e-01
-1.78766266e-01 -9.02522355e-02 8.09749424e-01 -9.38274339e-03
-1.34601104e+00 9.04250443e-01 6.03560925e+00 1.39876068e-01
-1.01145947e+00 1.40004441e-01 4.29777205e-01 -8.64743963e-02
3.08221489e-01 -1.81178913e-01 -8.01176846e-01 1.18570454e-01
5.35096526e-01 2.79787779e-01 2.63433248e-01 1.46066475e+00
2.08669081e-01 -2.68274069e-01 -1.45296478e+00 1.00040567e+00
-9.31703672e-02 -1.28705525e+00 -2.01433539e-01 -3.43153588e-02
6.36725962e-01 4.85100418e-01 -4.98700328e-02 3.73329550e-01
6.08201504e-01 -7.73501277e-01 1.44894087e+00 3.17407489e-01
4.22408640e-01 -8.84719342e-02 5.90756774e-01 4.64593977e-01
-9.62204278e-01 -3.12368006e-01 -5.84543467e-01 -2.79042386e-02
9.77343395e-02 1.10048153e-01 -1.37677920e+00 -4.13251743e-02
1.02258158e+00 2.48153687e-01 -9.19192314e-01 1.31442463e+00
-1.46014750e-01 8.88483003e-02 -6.99416339e-01 -4.23128158e-03
3.94394964e-01 2.96653241e-01 4.13698852e-01 1.17884421e+00
1.26625419e-01 -1.06053784e-01 3.71093392e-01 1.05730331e+00
1.07367493e-01 -3.57260495e-01 -8.75489235e-01 2.74181604e-01
4.15807813e-01 1.35153615e+00 -1.06110466e+00 -5.68529069e-01
-3.58094931e-01 7.65625119e-01 4.07897294e-01 2.90311128e-01
-6.81149781e-01 5.38341515e-02 6.36556387e-01 3.82264912e-01
5.90482116e-01 -5.91664374e-01 -8.75085369e-02 -8.98965180e-01
3.57234888e-02 -5.95274150e-01 -1.25838146e-01 -1.16374326e+00
-1.06956911e+00 6.52875900e-01 2.30682135e-01 -1.31946576e+00
-4.41021752e-03 -1.05883276e+00 -2.09901676e-01 5.46295762e-01
-1.31380558e+00 -1.32489479e+00 -8.36442530e-01 3.45950544e-01
7.71882355e-01 -1.48451269e-01 7.43924618e-01 2.06217200e-01
-3.48644704e-01 4.27503362e-02 7.86887389e-03 2.07127243e-01
2.35432208e-01 -1.10612404e+00 7.66233504e-01 7.97576189e-01
3.68889779e-01 3.57854217e-01 9.22119141e-01 -2.90561289e-01
-1.32857502e+00 -1.22972834e+00 1.50514498e-01 -8.52172732e-01
4.57935214e-01 -9.38116550e-01 -7.99980998e-01 8.02331269e-01
-1.82062358e-01 5.24428070e-01 6.03124015e-02 -4.96819586e-01
-3.64095092e-01 -5.77152111e-02 -1.27485895e+00 5.33137381e-01
1.19474530e+00 -3.10887516e-01 -5.29075325e-01 8.44468653e-01
7.59178162e-01 -4.59861219e-01 -3.24906141e-01 5.41219592e-01
5.22826552e-01 -1.17583072e+00 1.34969521e+00 -8.20567131e-01
1.08451173e-01 -7.32033253e-01 -6.01609886e-01 -9.13211465e-01
-1.16230354e-01 -6.61673471e-02 1.72103986e-01 7.89171755e-01
3.05341989e-01 -2.95700610e-01 9.14491951e-01 6.23636007e-01
-3.87814462e-01 -3.45972747e-01 -6.64165854e-01 -9.15934980e-01
-4.95959014e-01 -8.52027595e-01 4.59045917e-01 5.95391929e-01
-7.05198586e-01 2.36958656e-02 2.27733210e-01 3.81249219e-01
7.16962636e-01 -1.25114247e-01 1.44915915e+00 -1.24745774e+00
-2.62574166e-01 -4.25614119e-01 -8.53720725e-01 -8.23005617e-01
-2.06906691e-01 -6.57461822e-01 4.39285845e-01 -1.74262190e+00
-6.30305633e-02 -8.54475558e-01 4.26969454e-02 5.23883343e-01
2.72739977e-01 4.69383687e-01 1.71431646e-01 2.27460667e-01
-7.90577829e-01 3.03536147e-01 9.68412042e-01 -2.37384230e-01
5.25895022e-02 1.82390958e-02 -1.62000820e-01 9.64609027e-01
5.91390073e-01 -3.91108513e-01 -1.28824905e-01 -6.50432587e-01
-2.22472206e-01 -4.22780782e-01 8.64580989e-01 -1.42791593e+00
3.45199928e-02 -8.18060860e-02 6.61418319e-01 -6.26461506e-01
6.16855919e-01 -1.13265526e+00 1.24521613e-01 4.51816887e-01
-2.03457162e-01 3.43193412e-02 5.06135881e-01 6.76996231e-01
2.30538219e-01 -2.47868180e-01 8.17015886e-01 -8.39189112e-01
-1.30706632e+00 5.97078577e-02 -2.59015799e-01 -3.56988102e-01
1.35684049e+00 -4.73337620e-01 -1.67137012e-01 2.12159362e-02
-7.72436917e-01 -1.28831983e-01 6.87039137e-01 6.29353464e-01
6.63243711e-01 -1.06234956e+00 -5.10771811e-01 5.32127321e-01
6.45339191e-01 4.22437161e-01 -2.07872599e-01 5.34033954e-01
-8.97691071e-01 4.77271289e-01 -3.11002880e-01 -1.14770877e+00
-9.29069757e-01 8.70754719e-01 5.13401091e-01 3.30219001e-01
-8.31947029e-01 1.01979434e+00 5.27348101e-01 -7.83121228e-01
4.81908023e-01 -5.87564647e-01 2.70385623e-01 -3.25375080e-01
4.23697323e-01 -2.63230111e-02 3.62217665e-01 -6.77565634e-01
-3.92513275e-01 6.09512478e-02 4.82601821e-02 -1.06951289e-01
1.37113392e+00 -1.66702978e-02 2.34169155e-01 2.76774079e-01
8.79799128e-01 -4.66482311e-01 -1.67825413e+00 -4.84425128e-01
2.72470802e-01 -5.64874411e-01 -2.13054851e-01 -7.75542378e-01
-8.09268713e-01 6.78893387e-01 9.04391170e-01 1.61766171e-01
7.05735087e-01 5.10973215e-01 1.04552679e-01 8.35219800e-01
8.74880433e-01 -7.89646327e-01 5.67176402e-01 2.82686085e-01
8.67482901e-01 -1.60501552e+00 -1.68215632e-01 -2.12611437e-01
-3.57590169e-01 1.10522926e+00 8.46566021e-01 -2.82532126e-01
4.22311962e-01 3.61219376e-01 1.90516964e-01 -4.62790191e-01
-4.53590930e-01 -6.12656116e-01 1.53805837e-01 8.99538517e-01
-1.96018010e-01 6.77945167e-02 6.21111393e-01 2.04548761e-01
5.62658068e-03 -4.23260033e-01 4.51358765e-01 1.22068405e+00
-7.38713562e-01 -6.86123013e-01 -6.89200103e-01 3.70638222e-01
1.56558752e-02 5.68004437e-02 -3.22774291e-01 1.18125403e+00
3.25093269e-01 6.67116404e-01 2.27736190e-01 -2.49619558e-01
5.87582886e-01 -7.61522204e-02 5.74389696e-01 -9.90343809e-01
-3.15083086e-01 -8.46072882e-02 7.77458847e-02 -4.39528406e-01
-5.63863218e-01 -8.60159159e-01 -1.08375609e+00 3.32151830e-01
-4.29007709e-01 -5.52455366e-01 1.20365357e+00 1.01591706e+00
1.74568236e-01 6.39733136e-01 1.02643207e-01 -1.80293334e+00
-3.53703797e-01 -1.12564635e+00 -3.04953396e-01 3.94650400e-01
4.02037740e-01 -9.68068063e-01 -1.31239176e-01 1.03250347e-01] | [7.937332630157471, -1.1587871313095093] |
e95b188d-1fab-4118-8d5a-cc3d0bcb54a7 | a-novel-image-descriptor-with-aggregated | 2202.03677 | null | https://arxiv.org/abs/2202.03677v1 | https://arxiv.org/pdf/2202.03677v1.pdf | A Novel Image Descriptor with Aggregated Semantic Skeleton Representation for Long-term Visual Place Recognition | In a Simultaneous Localization and Mapping (SLAM) system, a loop-closure can eliminate accumulated errors, which is accomplished by Visual Place Recognition (VPR), a task that retrieves the current scene from a set of pre-stored sequential images through matching specific scene-descriptors. In urban scenes, the appearance variation caused by seasons and illumination has brought great challenges to the robustness of scene descriptors. Semantic segmentation images can not only deliver the shape information of objects but also their categories and spatial relations that will not be affected by the appearance variation of the scene. Innovated by the Vector of Locally Aggregated Descriptor (VLAD), in this paper, we propose a novel image descriptor with aggregated semantic skeleton representation (SSR), dubbed SSR-VLAD, for the VPR under drastic appearance-variation of environments. The SSR-VLAD of one image aggregates the semantic skeleton features of each category and encodes the spatial-temporal distribution information of the image semantic information. We conduct a series of experiments on three public datasets of challenging urban scenes. Compared with four state-of-the-art VPR methods- CoHOG, NetVLAD, LOST-X, and Region-VLAD, VPR by matching SSR-VLAD outperforms those methods and maintains competitive real-time performance at the same time. | ['Cheng Shuai', 'Hu Jun', 'Liu Wei', 'Pan Feng', 'Xue Dingyu', 'Feng Joe-Mei', 'Nie Jiwei'] | 2022-02-08 | null | null | null | null | ['visual-place-recognition'] | ['computer-vision'] | [-3.99891138e-02 -7.41271853e-01 -1.01018801e-01 -3.99282038e-01
-6.38888061e-01 -5.39518416e-01 6.57466173e-01 1.96711391e-01
-4.54654843e-01 5.03378928e-01 2.87237260e-02 2.24657163e-01
-3.27088147e-01 -8.46259475e-01 -6.54689431e-01 -6.11910582e-01
4.30992134e-02 2.62294203e-01 8.04866850e-01 -4.37454283e-01
5.01234829e-01 9.07418668e-01 -1.85405970e+00 -7.40196407e-02
7.91627944e-01 1.06807244e+00 7.48992085e-01 1.79961041e-01
-3.02360147e-01 5.16120374e-01 -3.47862124e-01 1.75248638e-01
3.24382991e-01 -5.11101075e-02 -5.19489229e-01 1.80337965e-01
4.86300021e-01 6.25742227e-02 -5.65748811e-01 1.24842227e+00
2.61550784e-01 5.69815814e-01 7.04582453e-01 -1.33978760e+00
-7.87116885e-01 -2.70530283e-01 -7.30713248e-01 1.73688322e-01
4.13488925e-01 8.98111612e-02 6.75863922e-01 -8.25617790e-01
9.09516156e-01 1.41379094e+00 5.71362078e-01 1.70886114e-01
-1.08197773e+00 -5.52279472e-01 4.06950325e-01 4.69158888e-01
-2.08590698e+00 -2.93053061e-01 8.49933267e-01 -4.52938408e-01
7.86374390e-01 3.63666266e-01 4.84972000e-01 6.40829206e-01
3.39772403e-01 6.01683795e-01 1.00230598e+00 3.96301621e-04
2.13554040e-01 -7.69131184e-02 2.63870452e-02 7.58452117e-01
2.55638778e-01 -9.84224081e-02 -6.80850446e-01 -1.32540846e-03
6.23615682e-01 5.04081666e-01 -2.89571047e-01 -8.24790239e-01
-1.28939450e+00 5.51191926e-01 8.79231453e-01 2.58497775e-01
-2.69995213e-01 9.11018476e-02 3.14746201e-01 -4.32673059e-02
2.90439576e-01 -1.23627946e-01 -2.65761673e-01 2.89634854e-01
-9.39565301e-01 1.39141157e-01 1.55932665e-01 1.06464958e+00
1.35694647e+00 -8.08994099e-02 -5.75423688e-02 9.78762627e-01
6.31149948e-01 1.08213592e+00 6.08673573e-01 -6.28399134e-01
3.83188605e-01 6.81010246e-01 -3.99956070e-02 -1.64392316e+00
-3.39161754e-01 -5.43053038e-02 -7.61514902e-01 -2.46751886e-02
-6.39131516e-02 5.88470697e-01 -1.37991476e+00 1.58827734e+00
5.05020976e-01 2.07596868e-01 2.01595873e-01 1.04506767e+00
1.17234159e+00 8.89461219e-01 6.51973784e-02 1.75989252e-02
1.11844277e+00 -9.22588944e-01 -5.23738027e-01 -5.31060338e-01
1.70545012e-01 -8.39256108e-01 6.63332045e-01 -8.33126679e-02
-3.15561473e-01 -7.35626757e-01 -1.01665115e+00 -8.54459703e-02
-7.05280602e-01 4.59709652e-02 5.91792285e-01 5.57529703e-02
-1.02497554e+00 2.31686965e-01 -4.75165367e-01 -8.81289721e-01
3.16453934e-01 9.77368876e-02 -7.99280822e-01 -3.12161773e-01
-8.58525574e-01 7.38761008e-01 5.02168119e-01 2.17298284e-01
-1.06865072e+00 -4.02594239e-01 -1.17905056e+00 -3.46382201e-01
2.30196461e-01 -2.07544327e-01 5.36138773e-01 -6.80694938e-01
-9.95421469e-01 1.09064114e+00 -4.28524584e-01 -1.01853408e-01
3.74058634e-01 9.22938958e-02 -5.76967299e-01 5.26519045e-02
6.81298494e-01 7.50428975e-01 5.21309495e-01 -1.50295174e+00
-8.51695895e-01 -6.53131843e-01 -4.77078289e-01 4.45968598e-01
4.24729764e-01 -1.48505360e-01 -8.80062640e-01 -3.33166897e-01
1.01139104e+00 -8.91154647e-01 -2.52848774e-01 1.78306252e-01
-2.14529604e-01 -8.84989202e-02 1.00736451e+00 -5.34799874e-01
8.25423658e-01 -2.56012487e+00 -8.37272927e-02 3.00296336e-01
-1.56106770e-01 9.95512232e-02 -2.31453925e-01 3.46668124e-01
2.56212801e-01 -1.52912989e-01 -2.69285887e-01 -2.41975144e-01
-1.21195264e-01 6.29856646e-01 -4.44734544e-01 9.74408746e-01
-2.49612227e-01 7.49128997e-01 -1.05696285e+00 -9.33460891e-01
6.04660273e-01 3.45666021e-01 -7.21101612e-02 1.19352132e-01
-3.40773091e-02 4.71845239e-01 -7.07130611e-01 9.78477061e-01
9.68289018e-01 2.44057745e-01 -4.06370252e-01 -1.52039587e-01
-2.90943950e-01 -2.77036518e-01 -1.35906017e+00 1.97891593e+00
-2.49129385e-01 8.33237529e-01 -1.55106306e-01 -7.97339141e-01
1.48741066e+00 -3.45324337e-01 4.64198738e-01 -1.22907853e+00
-1.70220435e-01 3.82823020e-01 -6.41824901e-01 -4.41798329e-01
7.57345617e-01 3.73460650e-01 -2.97580868e-01 -2.21652374e-01
1.63632520e-02 -2.48455808e-01 -5.49594276e-02 1.69497266e-01
7.99124897e-01 1.73874184e-01 2.72097141e-01 -2.97440320e-01
6.85914516e-01 3.66191179e-01 8.59638155e-01 8.83089364e-01
-5.11323512e-01 8.28764260e-01 -1.36645466e-01 -5.12602508e-01
-9.58234072e-01 -1.45784962e+00 -3.73412162e-01 7.49076962e-01
1.16001761e+00 -1.12123743e-01 -6.92234039e-02 -3.89883578e-01
2.86017120e-01 4.04569924e-01 -4.70386475e-01 -1.51320219e-01
-2.98412830e-01 -5.45315504e-01 3.35767120e-01 2.01838136e-01
1.13232076e+00 -8.76973987e-01 -5.03799856e-01 4.92873080e-02
-3.77457827e-01 -1.17256689e+00 -4.26017314e-01 -4.97486778e-02
-3.90218496e-01 -1.01737082e+00 -4.14577246e-01 -9.16472554e-01
6.33518100e-01 8.47446740e-01 6.30838633e-01 -1.32214027e-02
-3.96625042e-01 3.72555614e-01 -5.31698465e-01 -6.42718822e-02
2.28006959e-01 -2.82195151e-01 1.02640480e-01 4.18027401e-01
2.00263306e-01 -4.75713998e-01 -5.92364848e-01 5.52576065e-01
-7.25200593e-01 -7.69874975e-02 4.37728286e-01 6.12915158e-01
1.15934658e+00 3.77791375e-02 -1.36910945e-01 -4.09544975e-01
-3.21634263e-02 -5.26722372e-01 -6.86400473e-01 4.17538643e-01
-4.45762575e-01 -1.19108871e-01 2.78271139e-01 -2.25169331e-01
-9.48231161e-01 3.78498197e-01 1.61187485e-01 -4.65825826e-01
-3.17150623e-01 2.18698502e-01 -4.62308049e-01 -4.27847236e-01
4.97214556e-01 8.76502395e-01 -3.08435529e-01 -4.38673556e-01
4.19879258e-01 7.27522373e-01 1.03052306e+00 -3.93013597e-01
9.58137512e-01 7.76905477e-01 1.86473221e-01 -9.63136315e-01
-6.64453387e-01 -1.16614985e+00 -7.59240866e-01 -3.51358324e-01
9.33894515e-01 -1.17360044e+00 -2.03030363e-01 5.41266799e-01
-1.18275452e+00 5.56827569e-03 -3.41278464e-02 3.98193300e-01
-5.57290792e-01 3.54568988e-01 4.58568707e-03 -7.51724899e-01
-1.83952916e-02 -1.17175221e+00 1.43292999e+00 3.35784405e-01
3.36700499e-01 -7.39036500e-01 1.85295328e-01 2.76301891e-01
2.42253318e-01 5.67957342e-01 4.11497653e-01 -4.65299785e-01
-9.37376797e-01 -9.26888958e-02 -4.37284917e-01 1.73773855e-01
1.10277824e-01 2.35200743e-03 -8.54589939e-01 -1.58922434e-01
-2.98109293e-01 1.16697550e-01 9.43544745e-01 2.55330950e-01
9.08557236e-01 -4.34664376e-02 -6.60372734e-01 1.13810074e+00
1.83146131e+00 3.65989208e-01 7.86510408e-01 6.81148529e-01
1.00072718e+00 2.70439118e-01 1.16913581e+00 3.06656688e-01
5.98465502e-01 8.57271254e-01 5.38061619e-01 -1.22673266e-01
-2.32761234e-01 -5.68899632e-01 1.78917676e-01 4.88361567e-01
1.00342199e-01 -1.56660378e-01 -1.04639089e+00 7.76309967e-01
-2.20271754e+00 -9.35540557e-01 -3.45707327e-01 2.11700058e+00
2.45889246e-01 -1.94828704e-01 -4.67422664e-01 -3.30434263e-01
7.77287304e-01 6.46948397e-01 -4.85355377e-01 -6.32540807e-02
-5.29282093e-01 -2.37775832e-01 1.00485063e+00 4.06011909e-01
-1.23022127e+00 1.26238227e+00 5.29170752e+00 1.11254299e+00
-1.05811179e+00 1.93642303e-01 1.58188269e-01 5.38396299e-01
-1.32168755e-01 2.26537019e-01 -7.18115568e-01 5.41470766e-01
3.02748322e-01 -8.98183435e-02 4.13696349e-01 1.15006983e+00
6.28910288e-02 -4.75788325e-01 -6.72442853e-01 1.47643673e+00
3.12936217e-01 -1.09661865e+00 1.45991489e-01 -8.08025748e-02
8.48492146e-01 4.24844265e-01 4.91875708e-02 1.16426133e-01
1.40951142e-01 -7.76107848e-01 1.00175297e+00 8.78584862e-01
6.93267107e-01 -6.31922305e-01 6.84928596e-01 2.15170816e-01
-1.81722248e+00 -7.95306489e-02 -6.85318828e-01 2.81439602e-01
-4.31813393e-03 3.24795753e-01 -6.12039268e-01 8.44715834e-01
9.46553826e-01 1.03981650e+00 -9.32356477e-01 1.28230369e+00
-2.30314836e-01 -7.80410618e-02 -3.10905069e-01 1.17490634e-01
4.00351644e-01 -3.08637559e-01 6.35830879e-01 1.11518478e+00
3.07034999e-01 8.75895023e-02 4.66259539e-01 7.93489218e-01
3.74967873e-01 1.59588382e-01 -8.68430972e-01 3.77533078e-01
8.38772416e-01 1.16640806e+00 -8.58279824e-01 -2.33441278e-01
6.11024350e-03 1.08647823e+00 1.05765097e-01 3.83738846e-01
-6.57064319e-01 -3.83367509e-01 7.77745664e-01 -6.94248527e-02
2.76388317e-01 -5.66250801e-01 -6.62023947e-02 -1.04827070e+00
4.93222848e-02 -3.70827377e-01 2.79998481e-01 -1.17291737e+00
-9.70615804e-01 4.62316275e-01 2.85792351e-03 -1.31448412e+00
2.82630622e-01 -3.00268441e-01 -3.84794712e-01 7.36340106e-01
-1.68699145e+00 -1.38517678e+00 -8.35782707e-01 9.23220813e-01
6.38025820e-01 -2.16621444e-01 5.03858209e-01 2.53431946e-01
-3.73747826e-01 8.08429644e-02 4.68484730e-01 1.74617574e-01
5.71206391e-01 -9.01615262e-01 2.23000869e-01 1.07944989e+00
2.50272810e-01 2.52441585e-01 6.43704832e-01 -7.94384181e-01
-1.33525729e+00 -1.50860035e+00 8.43294084e-01 -2.66799271e-01
4.29165900e-01 -3.03722829e-01 -7.56590366e-01 5.69214463e-01
-4.30901647e-01 4.57242936e-01 7.45701492e-02 -3.94833833e-01
-3.40447128e-01 -5.47897696e-01 -1.25219285e+00 3.99868250e-01
1.29415405e+00 -7.89394617e-01 -7.01448500e-01 4.05255377e-01
8.62448633e-01 -5.48185647e-01 -4.78258878e-01 4.52155679e-01
2.29830444e-01 -9.23401773e-01 1.17860484e+00 -4.58605774e-03
-2.14550495e-01 -9.13704216e-01 -8.66500795e-01 -1.12319219e+00
-4.55039859e-01 6.28467351e-02 6.71296000e-01 1.47272396e+00
-2.16248810e-01 -6.64551258e-01 2.61600643e-01 2.01961473e-01
-2.62763292e-01 -2.34745562e-01 -1.15236235e+00 -9.97967899e-01
-4.97713268e-01 -3.89566094e-01 8.19097579e-01 8.03649426e-01
-6.14593387e-01 -1.93117708e-01 -1.89545915e-01 6.55735433e-01
8.77130508e-01 2.78044224e-01 8.43489349e-01 -1.22280753e+00
4.79478300e-01 -1.58314288e-01 -1.24065912e+00 -7.86707819e-01
2.98356086e-01 -9.27654088e-01 3.84079158e-01 -1.84726501e+00
2.72058427e-01 -5.80668628e-01 -3.69482845e-01 4.66229439e-01
2.03739226e-01 3.55375409e-01 2.95550495e-01 6.25677645e-01
-1.11340463e+00 8.03704858e-01 9.59946811e-01 -3.75674039e-01
-2.28778541e-01 -3.72207612e-01 -4.80927117e-02 6.07318640e-01
5.05158126e-01 -5.64169824e-01 -3.23287517e-01 -4.64913785e-01
-2.08879545e-01 -1.10624284e-01 7.18129277e-01 -1.34984243e+00
5.33556759e-01 -5.73696733e-01 3.65691841e-01 -1.11679888e+00
4.02802795e-01 -9.06812429e-01 6.03287876e-01 4.37566221e-01
6.12793975e-02 1.09556466e-01 2.15162244e-03 7.95083940e-01
-4.08646375e-01 3.68725732e-02 7.95092940e-01 -2.16472641e-01
-1.78781998e+00 5.31753480e-01 -6.34422004e-02 -2.68564895e-02
1.22286355e+00 -4.17201787e-01 -4.21029538e-01 5.41078188e-02
-3.53290021e-01 3.32431912e-01 7.62074590e-01 6.75689220e-01
1.08463860e+00 -1.57404006e+00 -4.11207438e-01 4.21113640e-01
6.71081424e-01 3.24522883e-01 5.42563260e-01 5.55435538e-01
-9.57590163e-01 3.24499637e-01 -4.14060652e-01 -1.04153574e+00
-1.26241386e+00 6.79142475e-01 4.61044937e-01 3.01964849e-01
-6.62134707e-01 5.99265993e-01 3.79132777e-01 -5.44109881e-01
6.89576790e-02 5.84978901e-04 -1.73420951e-01 2.99627595e-02
2.65457511e-01 2.32879788e-01 -8.29467643e-03 -1.50612009e+00
-9.17296767e-01 1.22082770e+00 3.32094073e-01 9.93546993e-02
1.26551199e+00 -6.43087268e-01 -3.43878090e-01 4.85368729e-01
1.36325359e+00 -1.16360903e-01 -1.08227146e+00 -5.12743771e-01
-7.56023750e-02 -9.29628074e-01 1.20683648e-01 -5.13497770e-01
-1.14379311e+00 4.77543354e-01 9.84251440e-01 -3.86372596e-01
9.96570528e-01 1.86545759e-01 5.56020796e-01 3.72039676e-01
1.01726782e+00 -1.21652150e+00 -4.42589112e-02 5.97345352e-01
7.51413107e-01 -1.36164391e+00 1.68170765e-01 -2.75790125e-01
-7.44570553e-01 8.96846116e-01 5.51591337e-01 -1.57561883e-01
6.52294338e-01 -4.00924265e-01 9.88073349e-02 -3.12240303e-01
-6.42995462e-02 -6.95292771e-01 4.20747131e-01 9.34142172e-01
-5.29726684e-01 1.93340257e-01 -3.80122289e-02 1.23640671e-01
-9.04398933e-02 -4.45685804e-01 1.04682855e-01 7.87120640e-01
-8.08080494e-01 -5.01783490e-01 -5.89989543e-01 -8.03372338e-02
2.37219542e-01 2.49819487e-01 -1.68306842e-01 8.73028338e-01
5.75502574e-01 9.08316016e-01 2.29251266e-01 -5.54527998e-01
4.42512929e-01 -4.26481158e-01 1.58996001e-01 -4.36570406e-01
8.61298740e-02 -2.45890602e-01 -3.13648164e-01 -9.57736492e-01
-4.22951192e-01 -7.78428674e-01 -1.48544443e+00 -2.66934037e-01
-7.90847093e-02 -2.17872951e-02 9.04516935e-01 8.10325742e-01
2.61127084e-01 2.08362907e-01 8.83521736e-01 -8.86682451e-01
-1.37572140e-01 -3.28356355e-01 -1.01452470e+00 6.28306925e-01
4.61146027e-01 -1.02768457e+00 -3.79219264e-01 -1.53068736e-01] | [7.623830318450928, -1.9522496461868286] |
8fb94c3f-06d7-4693-8108-8941a6bd7df7 | joint-representation-classification-for | 1505.04617 | null | http://arxiv.org/abs/1505.04617v1 | http://arxiv.org/pdf/1505.04617v1.pdf | Joint Representation Classification for Collective Face Recognition | Sparse representation based classification (SRC) is popularly used in many
applications such as face recognition, and implemented in two steps:
representation coding and classification. For a given set of testing images,
SRC codes every image over the base images as a sparse representation then
classifies it to the class with the least representation error. This scheme
utilizes an individual representation rather than the collective one to
classify such a set of images, doing so obviously ignores the correlation among
the given images. In this paper, a joint representation classification (JRC)
for collective face recognition is proposed. JRC takes the correlation of
multiple images as well as a single representation into account. Under the
assumption that the given face images are generally related to each other, JRC
codes all the testing images over the base images simultaneously to facilitate
recognition. To this end, the testing inputs are aligned into a matrix and the
joint representation coding is formulated to a generalized
$l_{2,q}-l_{2,p}$-minimization problem. To uniformly solve the induced
optimization problems for any $q\in[1,2]$ and $p\in (0,2]$, an iterative
quadratic method (IQM) is developed. IQM is proved to be a strict descent
algorithm with convergence to the optimal solution. Moreover, a more practical
IQM is proposed for large-scale case. Experimental results on three public
databases show that the JRC with practical IQM no only saves much computational
cost but also achieves better performance in collective face recognition than
the state-of-the-arts. | ['Songcan Chen', 'Liping Wang'] | 2015-05-18 | null | null | null | null | ['sparse-representation-based-classification'] | ['computer-vision'] | [ 5.15110314e-01 -1.78464338e-01 -2.25806117e-01 -3.91459733e-01
-8.57856691e-01 -6.43638819e-02 -8.35518241e-02 -2.49814808e-01
-3.57322693e-02 5.23891628e-01 -8.57986733e-02 -4.95971702e-02
-3.98055732e-01 -7.80232906e-01 -4.16957587e-01 -8.56284380e-01
3.71705629e-02 1.98273703e-01 -2.77825117e-01 4.30010259e-02
2.13238448e-01 4.89765704e-01 -1.65484047e+00 1.83882535e-01
7.69119740e-01 1.21058166e+00 2.76228070e-01 2.50685424e-01
-1.42106697e-01 8.16054046e-01 -7.12197065e-01 -2.81467289e-01
4.63353425e-01 -7.29934335e-01 -5.32324195e-01 5.46886206e-01
4.12372082e-01 1.59400389e-01 -3.70055288e-01 1.48902464e+00
3.34539801e-01 3.00319493e-01 6.49125338e-01 -1.24126482e+00
-7.66515493e-01 1.21998340e-01 -1.08806229e+00 7.78739676e-02
2.01880693e-01 -3.05982590e-01 7.33182073e-01 -9.73497450e-01
3.42936665e-01 1.31796646e+00 3.53294224e-01 3.90294671e-01
-1.21591389e+00 -9.79859233e-01 2.14531839e-01 1.64322734e-01
-1.94512820e+00 -3.53994906e-01 8.76262188e-01 -2.87294716e-01
5.95122695e-01 4.71107095e-01 1.94080755e-01 2.88901001e-01
-7.34746456e-02 4.97529507e-01 1.12330472e+00 -5.08746564e-01
1.97998345e-01 1.50148556e-01 2.23533720e-01 9.10417914e-01
2.83583075e-01 -2.96861976e-01 -5.01682103e-01 -2.45348930e-01
6.46849930e-01 3.53683174e-01 -4.27921265e-01 -3.41130085e-02
-8.37012827e-01 8.96283805e-01 3.11465204e-01 3.03191751e-01
-4.16735381e-01 -4.87538502e-02 -1.49387568e-02 4.36067134e-01
2.96674371e-01 -5.00464477e-02 -4.72198464e-02 2.18667969e-01
-9.36219275e-01 -1.79532260e-01 5.90937793e-01 8.82157207e-01
1.05813694e+00 3.34225535e-01 2.01964602e-02 1.32447422e+00
4.15954411e-01 6.06361926e-01 4.99058753e-01 -8.60650003e-01
7.39390492e-01 8.41431797e-01 -2.11443961e-01 -1.68640745e+00
7.23823265e-04 -5.53464353e-01 -1.26260793e+00 4.56416756e-02
-4.77397293e-02 -4.10812199e-02 -7.03337610e-01 1.68982613e+00
1.58766299e-01 4.86784190e-01 2.53005803e-01 9.20179486e-01
6.08984828e-01 8.55810344e-01 -2.34262884e-01 -7.58180141e-01
1.21632552e+00 -6.31041229e-01 -5.68664670e-01 -2.20256999e-01
3.73613358e-01 -9.12064433e-01 3.45276326e-01 3.00237864e-01
-8.63797963e-01 -6.78951681e-01 -1.08700633e+00 4.17628318e-01
4.29349281e-02 6.31395757e-01 4.72918540e-01 7.17431426e-01
-9.13027108e-01 6.90057427e-02 -4.29873347e-01 -1.22382410e-01
4.52515304e-01 7.17694163e-01 -6.06132627e-01 -4.99140471e-01
-7.22788870e-01 4.48207885e-01 1.42900333e-01 3.61107737e-01
-5.56228518e-01 -3.46278191e-01 -9.52527225e-01 -6.45898059e-02
3.62614363e-01 -1.21754095e-01 5.94648242e-01 -1.20272446e+00
-1.22877836e+00 8.04285586e-01 -5.22041917e-01 -9.12754312e-02
-8.40740185e-03 1.76271021e-01 -6.40869260e-01 1.00274213e-01
3.17543328e-01 2.59124130e-01 1.04330063e+00 -1.23244596e+00
-5.30270576e-01 -6.98016584e-01 -1.77988470e-01 2.13511452e-01
-3.73638600e-01 1.12339683e-01 -7.12532163e-01 -7.50881255e-01
7.15014100e-01 -8.59601021e-01 -1.97558358e-01 1.17853303e-02
-3.79400820e-01 -7.07466993e-03 9.90020156e-01 -4.95508432e-01
1.12102199e+00 -2.52922487e+00 4.06162292e-01 6.65032625e-01
4.75266725e-02 3.09335083e-01 -3.66050065e-01 2.82232702e-01
-2.49789387e-01 -2.45107964e-01 -4.96974111e-01 -4.49588418e-01
-5.02108812e-01 3.49976093e-01 -2.22639948e-01 7.73066640e-01
3.51040587e-02 3.64259273e-01 -5.46467364e-01 -4.58181143e-01
6.72494024e-02 5.85644901e-01 -5.50574660e-01 2.90707707e-01
3.76618803e-01 3.30734998e-01 -5.56191981e-01 8.16931844e-01
1.12151086e+00 -2.20134288e-01 3.82716835e-01 -3.19409817e-01
-2.11008894e-03 -4.12048280e-01 -1.71709526e+00 1.31407809e+00
-3.78542304e-01 3.04520220e-01 3.04516524e-01 -1.70120788e+00
1.33551741e+00 3.74089003e-01 5.92744052e-01 -6.84034765e-01
-3.70913371e-02 2.58464813e-01 -8.77133012e-03 -2.32661486e-01
-7.55665172e-03 -6.76477179e-02 2.71708041e-01 3.95503700e-01
1.19888924e-01 2.46031389e-01 1.58383995e-01 1.14758387e-01
7.15441048e-01 -6.13485217e-01 3.38669449e-01 -2.27715105e-01
1.00511420e+00 -4.07546699e-01 9.65961933e-01 4.77896512e-01
6.05588630e-02 7.25508153e-01 4.03530657e-01 -2.41069302e-01
-4.53837246e-01 -7.07315087e-01 -7.28618875e-02 7.39855587e-01
3.56330007e-01 -3.41083467e-01 -4.40909714e-01 -6.06503010e-01
7.65600353e-02 2.58673102e-01 -6.12914562e-01 -1.50517225e-01
-4.05639172e-01 -8.56727660e-01 1.98863432e-01 3.41711164e-01
8.22446048e-01 -8.15299869e-01 -2.16945678e-01 4.28531179e-03
9.12291650e-03 -8.33051383e-01 -6.38904929e-01 -6.37926683e-02
-7.25018799e-01 -1.20553088e+00 -8.03672254e-01 -1.07679117e+00
1.27324700e+00 8.13212872e-01 4.34867918e-01 2.27252647e-01
-4.47080344e-01 3.74188334e-01 -3.55341196e-01 1.13974959e-01
1.00767240e-01 -5.25724769e-01 9.50890407e-02 7.58655310e-01
2.27566257e-01 -5.30248880e-01 -5.17077506e-01 5.71656823e-01
-7.79040456e-01 -1.83730513e-01 6.69629216e-01 1.04958713e+00
7.54409015e-01 4.36933041e-01 5.99833488e-01 -8.91612470e-01
3.33301663e-01 -6.08897507e-01 -5.59994340e-01 3.30137730e-01
-5.34998238e-01 -2.80072987e-01 6.41442716e-01 -3.30014229e-01
-8.27124357e-01 2.30177075e-01 1.36930242e-01 -7.60957301e-01
2.80268222e-01 7.29396641e-01 -3.46401513e-01 -2.63912141e-01
3.00222218e-01 8.48857045e-01 3.90641809e-01 -5.53088188e-01
-7.06422701e-03 7.82578468e-01 4.18130159e-01 -4.83166188e-01
9.15444136e-01 3.02626938e-01 1.36236683e-01 -9.92331088e-01
-5.52700639e-01 -5.85637510e-01 -2.41110221e-01 -6.32796660e-02
4.38776821e-01 -1.05712330e+00 -7.07970500e-01 2.13280886e-01
-8.19170654e-01 4.12768781e-01 -4.18241927e-03 5.81797838e-01
-2.14926109e-01 4.91059482e-01 -2.44511455e-01 -1.07144928e+00
-2.34320104e-01 -1.33837414e+00 8.08493376e-01 3.69835764e-01
1.38836488e-01 -4.84383732e-01 -2.82796979e-01 5.45386553e-01
2.21770108e-01 -8.97636695e-04 8.64086390e-01 -4.27362084e-01
-5.77313423e-01 -3.62694025e-01 -4.60607886e-01 6.47213936e-01
4.63389248e-01 -1.80541962e-01 -6.37952983e-01 -6.43856466e-01
1.72684282e-01 -2.75046647e-01 6.31349504e-01 1.38329595e-01
1.33928251e+00 -5.70347488e-01 -1.80536002e-01 7.41276145e-01
1.58092058e+00 6.87659144e-01 7.33005047e-01 -1.04938649e-01
5.48474491e-01 5.89840531e-01 4.93424118e-01 5.30550659e-01
5.53186797e-02 6.45760655e-01 2.96839178e-01 4.90795523e-02
-4.12958674e-02 2.28543263e-02 3.37170750e-01 9.71006691e-01
1.12854410e-02 -4.62881066e-02 -5.45099080e-01 3.34473222e-01
-1.80001485e+00 -1.01127613e+00 3.83104123e-02 2.41560555e+00
5.95793128e-01 -5.07646322e-01 -2.85803676e-01 3.39920789e-01
1.05916977e+00 1.85747430e-01 -4.50073540e-01 -1.37691006e-01
-2.75578380e-01 4.86953348e-01 1.07770614e-01 2.47210220e-01
-7.97260463e-01 4.70187634e-01 5.46994686e+00 1.07121944e+00
-1.27004838e+00 1.13716990e-01 9.15649056e-01 1.37841225e-01
-3.57806981e-02 7.42453486e-02 -7.75082767e-01 7.08356440e-01
4.24225241e-01 -1.93376794e-01 5.92510045e-01 8.73819292e-01
-1.06122933e-01 -1.33256659e-01 -7.97647417e-01 1.53330266e+00
4.77166116e-01 -1.11749065e+00 2.39297152e-01 9.86393988e-02
9.25707996e-01 -4.75562096e-01 3.45803589e-01 2.42819890e-01
-1.06531627e-01 -1.20686936e+00 4.74840254e-01 3.23025614e-01
1.06209719e+00 -8.75416040e-01 6.81442618e-01 3.18281651e-01
-1.56322360e+00 -2.33664900e-01 -6.16914570e-01 1.67609841e-01
-2.87995070e-01 6.05873466e-01 -1.99528754e-01 7.38436103e-01
4.50194567e-01 8.64484787e-01 -4.08758521e-01 7.91985750e-01
1.53722256e-01 4.50345814e-01 -1.72760189e-01 3.69286388e-01
9.15194079e-02 -5.55905819e-01 3.74717414e-01 9.18011010e-01
6.18489981e-01 5.57985187e-01 4.97580290e-01 4.90676701e-01
-2.10179582e-01 4.86350209e-01 -5.93758345e-01 3.04349959e-02
5.82936645e-01 1.01555550e+00 -5.73394418e-01 -2.33517125e-01
-4.81337845e-01 9.51450109e-01 3.20000082e-01 4.10575300e-01
-4.72336948e-01 -4.58222479e-01 6.27801061e-01 -3.06726933e-01
4.46572214e-01 -5.00759929e-02 -4.16179784e-02 -1.13883114e+00
1.70563176e-01 -1.03125477e+00 5.64798713e-01 -3.10020030e-01
-1.07842660e+00 6.69985831e-01 -2.88865805e-01 -1.55500698e+00
-1.59645546e-02 -3.74121577e-01 -5.48177779e-01 1.09365129e+00
-1.39348137e+00 -8.27698231e-01 -2.86887825e-01 9.02932882e-01
5.39072990e-01 -6.91032290e-01 7.95827627e-01 4.10679609e-01
-9.32923675e-01 8.81981611e-01 2.42817283e-01 2.29037642e-01
3.57336521e-01 -5.84635198e-01 -5.90868711e-01 8.25125396e-01
2.15760797e-01 9.00224864e-01 1.99014276e-01 -4.09425616e-01
-1.76841426e+00 -1.24808204e+00 6.87461972e-01 3.01724136e-01
1.72379225e-01 1.29337862e-01 -9.22325015e-01 4.14950222e-01
-1.13229327e-01 4.35611308e-01 9.62211072e-01 -4.44694944e-02
-6.07094586e-01 -6.42406225e-01 -1.27836668e+00 2.34318331e-01
8.46114993e-01 -7.87414193e-01 -1.04280025e-01 4.36054438e-01
1.02338068e-01 -1.57767802e-01 -8.02613795e-01 3.57939124e-01
3.94693226e-01 -9.12607312e-01 9.61353958e-01 -2.22724676e-01
2.11660773e-01 -6.77238882e-01 -8.54600966e-01 -1.05966580e+00
-3.22158039e-01 -3.49473834e-01 1.88224744e-02 1.29279089e+00
1.32087752e-01 -8.66183996e-01 7.40250885e-01 2.58066952e-01
1.81058615e-01 -8.31329226e-01 -1.33178592e+00 -7.93952286e-01
-4.50345635e-01 -1.35349825e-01 4.00201261e-01 9.26756203e-01
-1.41665235e-01 1.24892823e-01 -4.38898414e-01 4.53163475e-01
8.50844979e-01 4.84133244e-01 4.77627099e-01 -1.09477854e+00
-3.04381132e-01 -1.31380215e-01 -8.93981934e-01 -9.52909172e-01
3.48309815e-01 -9.62194204e-01 -1.34382427e-01 -1.06784332e+00
5.02535105e-01 -7.53024936e-01 -5.39994121e-01 5.50303221e-01
-6.76262081e-02 5.46448588e-01 3.32872093e-01 5.54813445e-01
-4.33719426e-01 5.77002645e-01 1.00125110e+00 -4.85758513e-01
-3.47752906e-02 4.89033498e-02 -8.91092062e-01 2.68679142e-01
4.72612947e-01 -3.27961534e-01 -6.15474641e-01 -3.96892607e-01
-3.13604861e-01 4.99512047e-01 -3.28673236e-02 -1.08226883e+00
3.94265145e-01 -2.08554089e-01 3.16084027e-01 -2.62518704e-01
5.97703040e-01 -8.22664917e-01 4.45047975e-01 4.48145241e-01
-1.55226365e-01 -9.12708938e-02 -8.81503597e-02 7.18465090e-01
-6.69338107e-01 -3.24202418e-01 1.00972736e+00 4.53099273e-02
-5.84552228e-01 4.57791120e-01 -3.05434428e-02 -3.04466307e-01
1.12126017e+00 -4.47331399e-01 -3.90699171e-02 -3.71616483e-01
-6.00961924e-01 7.68117011e-02 2.71530505e-02 8.89339074e-02
9.93000865e-01 -1.40321445e+00 -8.14628839e-01 6.09883070e-01
9.82356593e-02 -1.67522907e-01 5.10317624e-01 7.32583344e-01
-2.62410551e-01 3.67653459e-01 -1.39687443e-02 -6.37687743e-01
-1.44814861e+00 3.36495757e-01 1.91001549e-01 -4.06303108e-02
-2.53711909e-01 9.22810972e-01 3.27748418e-01 -1.29502460e-01
1.55688390e-01 2.84827381e-01 -3.79011959e-01 8.52146819e-02
7.16580033e-01 2.32368067e-01 1.06106270e-02 -1.21971714e+00
-4.29022372e-01 9.92952824e-01 -2.15230003e-01 1.14697285e-01
1.21318722e+00 3.67970876e-02 -5.20571947e-01 1.25051990e-01
1.80369878e+00 5.01960143e-02 -8.73625755e-01 -4.22722608e-01
-4.33974802e-01 -9.50141668e-01 4.61554378e-02 -3.51539701e-01
-1.75607502e+00 5.39587557e-01 7.16302633e-01 -8.73361975e-02
1.41054428e+00 -3.76345336e-01 3.39947820e-01 2.66937256e-01
5.89719951e-01 -7.04050601e-01 1.82481900e-01 2.85536170e-01
9.56134856e-01 -1.20288956e+00 3.38178538e-02 -7.17097759e-01
-6.34563744e-01 1.12493575e+00 6.63041055e-01 -1.94906369e-01
8.17762673e-01 -1.35653913e-01 -1.70586407e-01 -5.25048226e-02
-5.26919007e-01 3.11173834e-02 2.50075996e-01 5.01500189e-01
2.74618745e-01 7.96447098e-02 -3.38499874e-01 4.10129666e-01
1.67261869e-01 -2.53523663e-02 8.99561420e-02 9.51196015e-01
-3.48073989e-01 -1.06496763e+00 -7.36044943e-01 6.42276704e-01
-2.39113048e-01 1.64718181e-01 8.22781995e-02 2.01666787e-01
3.35793734e-01 1.17109561e+00 6.62339404e-02 -5.30392647e-01
1.69630602e-01 -2.11620167e-01 2.83584356e-01 -7.63207197e-01
-2.36849606e-01 -1.33853080e-02 -4.06605601e-01 -4.34901565e-01
-3.59804690e-01 -6.33392334e-01 -1.10198820e+00 -1.88394457e-01
-3.14126253e-01 5.34285545e-01 5.32548964e-01 7.18535781e-01
4.90761191e-01 2.02740073e-01 1.26161194e+00 -4.80993986e-01
-5.17020345e-01 -7.16419339e-01 -9.80440974e-01 2.87584901e-01
1.71900988e-01 -7.57092476e-01 -3.54529619e-01 7.05979615e-02] | [12.49701976776123, 0.4112582802772522] |
d46aeeb0-6eab-4618-bf10-4c4e9584f7cb | symbiotic-adversarial-learning-for-attribute | 2007.09609 | null | https://arxiv.org/abs/2007.09609v2 | https://arxiv.org/pdf/2007.09609v2.pdf | Symbiotic Adversarial Learning for Attribute-based Person Search | Attribute-based person search is in significant demand for applications where no detected query images are available, such as identifying a criminal from witness. However, the task itself is quite challenging because there is a huge modality gap between images and physical descriptions of attributes. Often, there may also be a large number of unseen categories (attribute combinations). The current state-of-the-art methods either focus on learning better cross-modal embeddings by mining only seen data, or they explicitly use generative adversarial networks (GANs) to synthesize unseen features. The former tends to produce poor embeddings due to insufficient data, while the latter does not preserve intra-class compactness during generation. In this paper, we present a symbiotic adversarial learning framework, called SAL.Two GANs sit at the base of the framework in a symbiotic learning scheme: one synthesizes features of unseen classes/categories, while the other optimizes the embedding and performs the cross-modal alignment on the common embedding space .Specifically, two different types of generative adversarial networks learn collaboratively throughout the training process and the interactions between the two mutually benefit each other. Extensive evaluations show SAL's superiority over nine state-of-the-art methods with two challenging pedestrian benchmarks, PETA and Market-1501. The code is publicly available at: https://github.com/ycao5602/SAL . | ['DaCheng Tao', 'Yu-Tong Cao', 'Jingya Wang'] | 2020-07-19 | null | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/2116_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123590222.pdf | eccv-2020-8 | ['person-search'] | ['computer-vision'] | [ 5.79457358e-03 -3.02227102e-02 -4.19536754e-02 -4.25587565e-01
-8.54723752e-01 -5.48931897e-01 7.38458037e-01 -1.63339227e-01
-3.22574675e-01 7.10126281e-01 1.95373625e-01 2.96236277e-02
1.21927954e-01 -1.03507066e+00 -6.73405766e-01 -8.97411406e-01
3.52569044e-01 6.19425714e-01 -6.05442412e-02 -2.09302738e-01
-2.36964405e-01 2.74719387e-01 -1.37026715e+00 3.00285906e-01
8.89127553e-01 1.15560830e+00 -2.37535328e-01 3.79964352e-01
-1.33418217e-01 5.54441392e-01 -5.76562524e-01 -1.13862324e+00
3.81314725e-01 -3.13731104e-01 -6.20003521e-01 -6.06584623e-02
3.81851435e-01 -2.89578229e-01 -7.89105594e-01 1.12686825e+00
7.78189063e-01 3.15483063e-02 6.54968560e-01 -1.69867778e+00
-1.09985375e+00 2.75402158e-01 -5.64822614e-01 -6.00258410e-02
3.18551511e-01 4.55644310e-01 9.42658544e-01 -9.09531474e-01
3.25209618e-01 1.37221217e+00 6.16670072e-01 8.25760424e-01
-1.20365763e+00 -9.37783182e-01 6.06433898e-02 5.58276772e-01
-1.65359092e+00 -4.62994337e-01 9.21237111e-01 -3.32658261e-01
4.40225571e-01 4.04785722e-01 6.53004169e-01 1.68134594e+00
-9.46271569e-02 9.37809110e-01 9.96687651e-01 -1.30688980e-01
1.23434216e-01 3.33759278e-01 -2.62805313e-01 4.35350239e-01
1.54148713e-01 3.16712767e-01 -4.90280420e-01 -3.25971007e-01
5.75351179e-01 2.42790833e-01 -1.11461245e-01 -5.60646057e-01
-1.07402289e+00 1.02187002e+00 6.21457338e-01 -8.19951203e-03
-3.64926487e-01 -1.09038940e-02 3.39603454e-01 1.95418045e-01
3.13338131e-01 2.12691650e-01 -1.00900428e-02 6.62263110e-02
-6.03193700e-01 4.20393616e-01 5.18382668e-01 9.34006989e-01
7.03850389e-01 -1.27875879e-01 -2.49272138e-01 9.36892629e-01
2.91085720e-01 5.10904193e-01 3.03442240e-01 -6.18187726e-01
5.74704230e-01 7.15355277e-01 -5.83883608e-03 -9.60481882e-01
9.10412669e-02 -3.31091970e-01 -1.09256506e+00 7.20672756e-02
5.27431488e-01 -1.75217614e-01 -1.05026996e+00 1.86152101e+00
5.25104642e-01 1.47748232e-01 9.76623967e-02 1.00730133e+00
1.01043510e+00 5.91746807e-01 3.33246022e-01 3.88651341e-01
1.39758503e+00 -9.25854146e-01 -6.38198793e-01 -5.14445186e-01
-4.68244590e-02 -6.99514151e-01 8.84596825e-01 8.04864988e-03
-9.91036713e-01 -7.02854455e-01 -9.82157230e-01 -9.79048759e-02
-7.37866044e-01 4.38112691e-02 5.45001268e-01 7.18630373e-01
-6.44656062e-01 1.90737337e-01 -5.01211405e-01 -1.11025952e-01
8.14042866e-01 2.71010607e-01 -4.30605859e-01 -3.12298954e-01
-1.43409419e+00 6.28847480e-01 3.15651208e-01 1.53013989e-01
-8.19635153e-01 -5.83962798e-01 -8.47952664e-01 -6.60380125e-02
3.25051963e-01 -9.81081903e-01 7.53596783e-01 -9.30823028e-01
-1.00723243e+00 1.04458022e+00 1.23771027e-01 -1.76721811e-01
9.28448260e-01 -3.16110462e-01 -5.74153900e-01 -1.34506421e-02
2.47596234e-01 8.70472431e-01 1.01049173e+00 -1.42051196e+00
-5.15880108e-01 -5.65717041e-01 1.20545849e-01 1.36255488e-01
-5.28118014e-01 -9.88196731e-02 -6.03929758e-01 -9.17217314e-01
-1.56742215e-01 -9.18480694e-01 -1.41641915e-01 2.39710137e-01
-6.38929009e-01 -2.37454981e-01 8.10244918e-01 -4.97724533e-01
9.59534645e-01 -2.16738653e+00 2.33474672e-01 3.97092625e-02
2.38425300e-01 4.06788915e-01 -1.13575593e-01 3.76792222e-01
-1.31218731e-01 4.86182459e-02 -1.78043023e-01 -4.60450351e-01
1.71642020e-01 2.01330364e-01 -3.46985281e-01 3.96235526e-01
2.81626701e-01 1.20944667e+00 -9.66707110e-01 -6.54543102e-01
1.58426180e-01 7.25633264e-01 -3.40167612e-01 4.47385609e-01
3.56461369e-02 5.76409757e-01 -5.50942242e-01 9.70754862e-01
8.23514998e-01 -1.00612760e-01 -9.10167396e-02 -3.79525304e-01
3.67648929e-01 -1.38890594e-01 -1.19199336e+00 1.68658090e+00
-2.14131743e-01 2.84490973e-01 -2.35616639e-01 -1.05982041e+00
7.69255042e-01 1.90050185e-01 4.59003061e-01 -6.72165334e-01
1.42651945e-01 5.20189777e-02 -1.58576831e-01 -4.60485220e-01
3.42859715e-01 -7.99520165e-02 -2.91574597e-01 2.85820961e-01
1.01365000e-01 1.33735880e-01 2.19884086e-02 9.64564234e-02
9.36127961e-01 6.33120388e-02 1.42807752e-01 1.81174010e-01
5.98761141e-01 -2.22101256e-01 7.97804713e-01 6.66967809e-01
-2.76354074e-01 7.43214846e-01 3.19863170e-01 -6.48564816e-01
-1.16893780e+00 -1.43142521e+00 -1.33908093e-01 9.16838646e-01
4.17849123e-01 -2.47697040e-01 -7.49363959e-01 -9.13931131e-01
1.29114747e-01 5.60437381e-01 -8.64186645e-01 -3.81256700e-01
-3.78991604e-01 -8.80011201e-01 7.67888129e-01 6.04400396e-01
7.46147871e-01 -1.19117463e+00 -3.44234169e-01 1.00249566e-01
-3.86783689e-01 -1.22444487e+00 -5.50140798e-01 -2.70044714e-01
-2.64443517e-01 -1.09739995e+00 -7.32513428e-01 -5.43742597e-01
6.83065116e-01 1.87688842e-02 1.23335743e+00 4.19394746e-02
-4.47764903e-01 3.43956798e-01 -4.46957350e-01 -5.29226899e-01
-2.88778991e-01 4.14182879e-02 1.05648175e-01 4.63004678e-01
7.29983568e-01 -5.69811583e-01 -8.59978080e-01 4.89845425e-01
-7.81623781e-01 -5.95141388e-02 6.63944185e-01 1.09742379e+00
6.70353591e-01 -6.18323758e-02 5.77884436e-01 -8.29778612e-01
4.28751379e-01 -7.67273307e-01 -2.93431580e-01 3.31479788e-01
-3.85452211e-01 -1.98013246e-01 6.15248442e-01 -7.09999263e-01
-9.06130910e-01 5.63557632e-02 -2.82530218e-01 -6.92653477e-01
-3.20971876e-01 -4.93131429e-02 -7.54505217e-01 1.82398483e-01
5.47607958e-01 3.30194712e-01 -3.40832733e-02 -2.92355001e-01
3.63532007e-01 6.52090132e-01 7.66786397e-01 -6.40597463e-01
1.16594648e+00 5.38652539e-01 -3.05666476e-01 -4.54706699e-01
-7.57324159e-01 -3.06879014e-01 -5.17212808e-01 -1.47096500e-01
9.52292502e-01 -8.66576910e-01 -4.93993789e-01 7.34949052e-01
-1.04669106e+00 -7.98617527e-02 -4.23807889e-01 2.47066021e-01
-5.24879158e-01 1.76890984e-01 -3.24994177e-01 -5.79927266e-01
-4.16314423e-01 -1.22830737e+00 1.06239522e+00 5.11579812e-01
-2.27437578e-02 -7.38019764e-01 -1.22481875e-01 6.18491888e-01
3.42119128e-01 5.54422677e-01 7.45322943e-01 -7.92899191e-01
-5.32768309e-01 -3.75008672e-01 -2.69029021e-01 2.65735000e-01
3.29984963e-01 -3.25526953e-01 -1.13599443e+00 -3.24873447e-01
-3.64863604e-01 -4.98236924e-01 7.37405598e-01 1.78603863e-04
1.43161142e+00 -3.73741269e-01 -3.70129347e-01 8.15229774e-01
1.28033042e+00 1.71797916e-01 9.04235840e-01 3.10412824e-01
9.19173658e-01 5.56298316e-01 4.88004565e-01 3.24769825e-01
5.14767826e-01 8.15872312e-01 5.75897098e-01 -1.93632439e-01
-1.95874408e-01 -4.49851871e-01 1.31078064e-01 3.43181998e-01
-4.89468388e-02 -3.42151523e-01 -9.11475420e-01 5.94687879e-01
-1.81954801e+00 -1.21665370e+00 2.25368440e-01 2.20347619e+00
8.73052835e-01 -2.21558027e-02 2.47199073e-01 1.12366766e-01
7.90482938e-01 1.08204290e-01 -8.16398919e-01 -4.36531566e-02
-1.74860090e-01 1.08735122e-01 2.67650574e-01 1.32725626e-01
-1.40950811e+00 7.52686620e-01 4.85740471e+00 9.28796530e-01
-8.19062054e-01 3.27286005e-01 8.05064201e-01 -1.07010618e-01
-2.02338904e-01 -2.02524304e-01 -8.02373230e-01 8.75059843e-01
4.99185950e-01 -2.33099777e-02 3.45027238e-01 8.24508727e-01
-3.39334458e-01 2.33906463e-01 -1.19642484e+00 1.21075594e+00
9.79144350e-02 -1.24160731e+00 1.48172930e-01 2.34907344e-01
5.45139015e-01 -2.39379182e-01 3.26025456e-01 3.81701499e-01
2.90698528e-01 -1.20590889e+00 6.68211877e-01 7.00186849e-01
8.68340313e-01 -9.38498318e-01 7.69389749e-01 3.10029089e-01
-1.16985106e+00 -1.33869141e-01 -3.11405241e-01 5.22443831e-01
1.27043396e-01 3.00330848e-01 -3.38018894e-01 5.74771285e-01
1.00947261e+00 4.72258657e-01 -7.09849417e-01 9.95612621e-01
-2.38243952e-01 2.64237523e-01 -1.00305416e-01 2.32329234e-01
1.43613502e-01 -2.32665271e-01 6.29806221e-01 1.10709858e+00
8.43028203e-02 1.52330965e-01 1.02546483e-01 1.05533218e+00
-3.63376588e-01 -1.12435400e-01 -6.33968115e-01 1.83274940e-01
7.55960345e-01 1.20043349e+00 -3.19653630e-01 -3.00221115e-01
-5.43901145e-01 1.16108000e+00 1.80673540e-01 2.47179657e-01
-1.12202585e+00 -2.68526465e-01 1.00192153e+00 1.14314660e-01
1.85620666e-01 2.09633023e-01 -2.49800324e-01 -1.13124096e+00
1.72283545e-01 -1.12450063e+00 5.27134955e-01 -5.78019261e-01
-1.91165137e+00 6.66142583e-01 -6.62504211e-02 -1.26386464e+00
-2.93757260e-01 -3.71942759e-01 -6.30414546e-01 1.00681138e+00
-1.39912295e+00 -1.58511007e+00 -5.78056276e-01 8.35347772e-01
5.60454607e-01 -4.17834729e-01 8.97783697e-01 6.13725424e-01
-6.29399598e-01 1.13003922e+00 -2.04515196e-02 5.34186602e-01
8.29919338e-01 -1.27797174e+00 4.51456726e-01 7.59360015e-01
1.05594099e-01 5.26368499e-01 5.03289163e-01 -4.85587656e-01
-1.27028346e+00 -1.22555268e+00 6.46993876e-01 -5.88919163e-01
4.56251889e-01 -6.11974776e-01 -9.99307513e-01 5.71288824e-01
2.91970998e-01 4.37459052e-01 8.99243355e-01 -1.38415590e-01
-4.87844497e-01 -2.98215091e-01 -1.27623510e+00 4.82773364e-01
1.08190513e+00 -4.76225197e-01 -4.78525490e-01 3.79362673e-01
3.99286330e-01 -4.56158191e-01 -8.90120685e-01 3.49748731e-01
6.09169543e-01 -9.78223681e-01 1.55664337e+00 -7.16764331e-01
4.94534194e-01 -3.12586606e-01 -2.96857953e-01 -1.34417939e+00
-2.64675736e-01 -2.87771970e-01 -2.18543932e-01 1.59875441e+00
1.35757759e-01 -6.82211161e-01 7.96345890e-01 7.13772357e-01
1.87857777e-01 -1.03081214e+00 -9.03950512e-01 -8.37040961e-01
1.49222761e-01 -7.95567781e-02 9.96235013e-01 9.52443600e-01
-6.88070714e-01 3.50358456e-01 -4.64004517e-01 2.50576675e-01
1.03631449e+00 3.16950046e-02 9.41924632e-01 -1.03610873e+00
-2.43580401e-01 -3.28487575e-01 -5.76824367e-01 -7.25781143e-01
1.33529514e-01 -6.39947414e-01 -2.55135983e-01 -1.23915577e+00
3.39414060e-01 -8.26829791e-01 -3.07746589e-01 6.68705106e-01
-4.25715417e-01 4.09293830e-01 2.71216720e-01 2.81390011e-01
-3.91918212e-01 7.79694557e-01 1.07235396e+00 -5.20771682e-01
5.50948158e-02 1.48907542e-01 -9.26744163e-01 6.81305110e-01
8.42144310e-01 -4.35413748e-01 -4.45803910e-01 -5.31724870e-01
-5.07378094e-02 -2.80302644e-01 8.32819283e-01 -9.51651812e-01
3.23029190e-01 -1.49364278e-01 8.41434062e-01 -4.73933727e-01
6.75131977e-01 -7.78785944e-01 2.90704966e-01 1.92193866e-01
-9.29006860e-02 9.77636650e-02 -1.77526865e-02 6.28385127e-01
-3.16892624e-01 -1.75574347e-02 9.56090987e-01 -6.32074103e-02
-6.81819379e-01 7.08146989e-01 4.13120419e-01 2.61754721e-01
1.19733536e+00 -3.19446862e-01 -3.03121865e-01 -3.41858000e-01
-6.75382793e-01 2.85913169e-01 4.87919509e-01 7.37286687e-01
7.55530715e-01 -1.76209712e+00 -9.26817238e-01 1.71839073e-01
2.36416265e-01 4.16354463e-02 3.90544087e-01 5.12704492e-01
-9.90936458e-02 5.66870160e-02 -3.78566980e-01 -4.90149587e-01
-1.09904802e+00 7.62837827e-01 4.08343554e-01 -2.68157154e-01
-4.17251468e-01 9.40602362e-01 3.99233520e-01 -4.44826990e-01
1.66909784e-01 6.27401650e-01 -1.74802691e-01 2.75685519e-01
5.27769923e-01 2.64706314e-01 -1.92656264e-01 -9.14223492e-01
-4.60049033e-01 4.32021886e-01 -2.41200641e-01 1.90119788e-01
1.34205782e+00 2.67738830e-02 1.22502349e-01 1.17178801e-02
1.19982255e+00 -8.74454528e-02 -1.31801975e+00 -4.45375681e-01
-4.23004985e-01 -7.62393534e-01 -3.13611984e-01 -7.72758245e-01
-1.35324669e+00 8.30973506e-01 8.21421146e-01 1.18626148e-01
1.29787230e+00 1.39206618e-01 1.00683033e+00 -1.48603797e-03
2.07470611e-01 -1.08167100e+00 3.01175237e-01 -3.35061625e-02
9.42452490e-01 -1.42841327e+00 -6.98955059e-02 -4.50644225e-01
-8.38878930e-01 6.92188442e-01 8.79640877e-01 -3.48974429e-02
3.14613193e-01 -1.78716574e-02 -7.62298256e-02 -7.76417479e-02
-4.37727004e-01 -2.33091921e-01 3.10518652e-01 8.23242426e-01
-2.43405588e-02 1.44074574e-01 2.72941329e-02 8.05907130e-01
-2.66172498e-01 -3.63330871e-01 1.90629326e-02 6.88615322e-01
2.34549671e-01 -1.40984786e+00 -5.81021667e-01 4.52053905e-01
-5.01972556e-01 4.87777777e-03 -4.39098477e-01 8.17691505e-01
6.39895916e-01 8.84247720e-01 2.19689235e-02 -5.14991224e-01
3.81880760e-01 -4.07952443e-02 3.46236378e-01 -2.49121130e-01
-4.72622424e-01 -2.92826653e-01 -1.95651695e-01 -5.02967775e-01
-2.95584619e-01 -8.74096096e-01 -7.72444844e-01 -4.79985476e-01
-9.46681947e-03 3.29180285e-02 4.14943695e-01 6.61481440e-01
4.10663992e-01 3.58009219e-01 7.49919951e-01 -5.14601707e-01
-5.07290840e-01 -7.56252646e-01 -1.71799913e-01 7.28605568e-01
2.18661487e-01 -9.33006525e-01 -1.35949939e-01 1.19959787e-01] | [14.63484001159668, 0.9980476498603821] |
236b857b-7507-408d-816b-979102301296 | federated-learning-with-noisy-labels | 2208.09378 | null | https://arxiv.org/abs/2208.09378v3 | https://arxiv.org/pdf/2208.09378v3.pdf | Labeling Chaos to Learning Harmony: Federated Learning with Noisy Labels | Federated Learning (FL) is a distributed machine learning paradigm that enables learning models from decentralized private datasets, where the labeling effort is entrusted to the clients. While most existing FL approaches assume high-quality labels are readily available on users' devices; in reality, label noise can naturally occur in FL and is closely related to clients' characteristics. Due to scarcity of available data and significant label noise variations among clients in FL, existing state-of-the-art centralized approaches exhibit unsatisfactory performance, while prior FL studies rely on excessive on-device computational schemes or additional clean data available on server. Here, we propose FedLN, a framework to deal with label noise across different FL training stages; namely, FL initialization, on-device model training, and server model aggregation, able to accommodate the diverse computational capabilities of devices in a FL system. Specifically, FedLN computes per-client noise-level estimation in a single federated round and improves the models' performance by either correcting or mitigating the effect of noisy samples. Our evaluation on various publicly available vision and audio datasets demonstrate a 22% improvement on average compared to other existing methods for a label noise level of 60%. We further validate the efficiency of FedLN in human-annotated real-world noisy datasets and report a 4.8% increase on average in models' recognition performance, highlighting that~\method~can be useful for improving FL services provided to everyday users. | ['Nirvana Meratnia', 'Tanir Ozcelebi', 'Aaqib Saeed', 'Vasileios Tsouvalas'] | 2022-08-19 | null | null | null | null | ['learning-with-noisy-labels', 'learning-with-noisy-labels'] | ['computer-vision', 'natural-language-processing'] | [ 6.88683316e-02 -2.84404129e-01 -4.28827927e-02 -6.01718426e-01
-1.41239417e+00 -7.81281590e-01 5.47474846e-02 -1.59985796e-01
-8.95769745e-02 4.05964762e-01 1.06290758e-01 5.01514934e-02
4.06130590e-02 -2.82864183e-01 -5.62565327e-01 -1.03268564e+00
2.61848897e-01 3.77638310e-01 6.10836968e-02 3.80873680e-01
-5.73958516e-01 4.94889557e-01 -1.69602668e+00 6.29658580e-01
3.49833667e-01 1.77868402e+00 3.27283703e-02 6.94739640e-01
-2.07757413e-01 1.17226815e+00 -9.13307965e-01 -5.75057447e-01
4.52857077e-01 4.22036648e-02 -4.84851331e-01 5.13094440e-02
3.84040564e-01 -5.14455736e-01 -2.07139373e-01 1.13266253e+00
1.13469374e+00 -2.76498854e-01 1.73773706e-01 -1.40515947e+00
-4.10736620e-01 8.98925543e-01 -1.96352571e-01 -2.75943000e-02
1.65927514e-01 3.98789614e-01 8.56856644e-01 -5.75555444e-01
3.29492241e-01 1.10081685e+00 1.03398609e+00 7.49607682e-01
-1.50181341e+00 -9.28458929e-01 3.52879949e-02 2.23465592e-01
-1.40245569e+00 -8.50897133e-01 8.27788889e-01 -2.94849306e-01
6.16826534e-01 4.95027542e-01 -2.66540665e-02 1.30992258e+00
-3.82161707e-01 8.86803806e-01 1.11151290e+00 -3.15953672e-01
6.34261608e-01 1.67341456e-01 -7.86537006e-02 2.70122975e-01
-1.09975606e-01 -2.73862153e-01 -9.44590509e-01 -6.64121687e-01
1.88532546e-01 7.98183605e-02 -7.59709030e-02 -1.41466841e-01
-9.17055547e-01 4.30405438e-01 1.31234691e-01 2.40675043e-02
-5.91459095e-01 3.80173773e-01 8.64192069e-01 2.55583704e-01
7.04869866e-01 -1.79854512e-01 -8.38131309e-01 -3.70782286e-01
-8.94811988e-01 -1.41881719e-01 9.23442960e-01 1.01645863e+00
6.92144156e-01 -8.34498927e-02 -3.14448088e-01 9.10136580e-01
1.16180748e-01 8.42163324e-01 4.12288338e-01 -1.33510113e+00
4.27747756e-01 4.17467952e-01 1.90237001e-01 -7.76098132e-01
-1.42437056e-01 -4.26962048e-01 -9.27334726e-01 -2.29056045e-01
5.34120314e-02 -4.53646630e-01 -4.27476287e-01 1.78299391e+00
5.59500575e-01 4.53860253e-01 -1.01318285e-02 7.39981472e-01
6.45479262e-01 2.18831748e-01 2.15307772e-01 -3.71794403e-01
1.04270232e+00 -1.06711614e+00 -1.02725983e+00 1.84317857e-01
6.24796331e-01 -8.15275848e-01 1.12181926e+00 9.46570575e-01
-7.99631655e-01 -2.80290961e-01 -4.48272496e-01 1.51257485e-01
-9.22980905e-02 2.39907369e-01 7.42888033e-01 1.03100502e+00
-1.13674450e+00 3.69921774e-01 -8.72852445e-01 -1.32765546e-01
7.68419564e-01 6.06330812e-01 -2.22833619e-01 -2.22171575e-01
-5.74918866e-01 1.53594971e-01 -3.40756595e-01 1.62709221e-01
-1.07039440e+00 -5.79232574e-01 -1.90203443e-01 7.96156824e-02
4.09767747e-01 -5.71021795e-01 1.93127692e+00 -8.53256404e-01
-1.55416834e+00 5.92776299e-01 -1.80535018e-01 -4.69089478e-01
6.65654600e-01 -1.89057812e-01 -8.19055855e-01 1.18843026e-01
-4.63045314e-02 7.99664482e-02 9.67157006e-01 -1.37915897e+00
-7.39332557e-01 -3.77047986e-01 -5.72536141e-02 -1.93909779e-01
-7.19122469e-01 2.15216041e-01 -4.73439962e-01 -4.84985083e-01
-3.64961103e-02 -7.95802593e-01 -1.18816115e-01 2.40244001e-01
-4.18003470e-01 -2.83712000e-01 1.24973643e+00 -4.37085062e-01
1.26225495e+00 -2.43323565e+00 -5.94663382e-01 2.04151273e-01
3.88023645e-01 3.25428486e-01 -9.00357813e-02 3.80020142e-01
2.80780613e-01 5.60128689e-02 2.81371057e-01 -1.01557517e+00
3.08603197e-01 3.61895531e-01 -2.32173607e-01 5.70352733e-01
-3.84032667e-01 5.34502506e-01 -7.40434468e-01 -4.70445991e-01
1.12644061e-01 5.17671168e-01 -5.85136771e-01 3.67803395e-01
-4.24939066e-01 6.33649409e-01 -3.79721254e-01 9.32139635e-01
6.86529219e-01 -4.88694906e-01 4.15722787e-01 -4.82656747e-01
3.49353462e-01 2.07268164e-01 -1.54409516e+00 1.78563190e+00
-7.39038944e-01 2.66971141e-01 7.30416238e-01 -6.42783880e-01
6.49953663e-01 8.48522842e-01 8.71754110e-01 -4.70832556e-01
3.66345674e-01 3.49720776e-01 -7.29959369e-01 -6.42442346e-01
2.03102473e-02 2.79864043e-01 4.03846148e-03 6.62035286e-01
2.44989675e-02 2.43172660e-01 -3.98976684e-01 3.77361253e-02
1.60258830e+00 -5.00968754e-01 -3.28760147e-01 1.02643430e-01
2.22121850e-01 -5.10129631e-01 7.56422460e-01 1.05942631e+00
-5.27238429e-01 5.02174973e-01 4.20244075e-02 -5.06599069e-01
-6.48982525e-01 -9.55606401e-01 2.35431120e-02 1.46863496e+00
-4.34769802e-02 -4.28176761e-01 -9.49397206e-01 -7.70014107e-01
-5.45835756e-02 2.05601200e-01 -2.29335546e-01 9.51225758e-02
-2.93919116e-01 -5.50942361e-01 8.51103604e-01 3.99705440e-01
6.07314527e-01 -9.16348279e-01 -4.93057132e-01 3.43851388e-01
-4.76092517e-01 -1.25066674e+00 -4.66261595e-01 4.64648217e-01
-4.49123412e-01 -9.25018549e-01 -2.77013518e-02 -5.27238011e-01
3.93928856e-01 3.00601631e-01 1.15895319e+00 -1.20123886e-01
-3.22931588e-01 6.61798716e-01 -3.26121897e-01 -5.40872395e-01
-2.44359091e-01 1.14880845e-01 3.14868152e-01 5.36818922e-01
3.48512650e-01 -7.72749305e-01 -7.90064037e-01 4.31439638e-01
-8.65654349e-01 -4.87099916e-01 2.14306623e-01 7.77330935e-01
7.18048394e-01 1.92417353e-01 7.02784061e-01 -9.19517279e-01
4.38310176e-01 -6.08347416e-01 -4.83520985e-01 2.72263318e-01
-6.13122761e-01 -1.99920580e-01 9.02649164e-01 -6.60149634e-01
-1.09027040e+00 3.58986467e-01 -2.22008228e-01 -6.75538778e-01
-4.08196956e-01 -1.01996519e-01 -4.96122360e-01 -1.14215314e-01
7.74898946e-01 -1.51063621e-01 -1.69077173e-01 -1.04462850e+00
4.85756665e-01 1.38182545e+00 6.67343676e-01 -8.27360868e-01
4.00841326e-01 7.51264930e-01 -3.91098768e-01 -2.13134676e-01
-8.23712230e-01 -7.14744329e-01 -1.38919085e-01 -1.72226012e-01
4.09097582e-01 -1.25750542e+00 -1.30373287e+00 6.60148025e-01
-9.23505664e-01 -4.20695692e-01 -6.83848441e-01 1.38888001e-01
-4.90962684e-01 1.26321062e-01 -7.57487059e-01 -1.02313817e+00
-8.69162977e-01 -1.26300275e+00 1.38629889e+00 -2.95773149e-02
7.44037926e-02 -6.11876547e-01 -1.94233611e-01 6.69696212e-01
7.70512283e-01 -5.59037644e-03 6.51930630e-01 -7.30564415e-01
-5.89745760e-01 -3.68518919e-01 -2.61239754e-03 6.75041258e-01
1.97102308e-01 -2.15170205e-01 -1.75542092e+00 -3.71498734e-01
1.54521048e-01 -6.40043437e-01 2.30700761e-01 6.87428489e-02
1.59703207e+00 -5.36320031e-01 -1.51205048e-01 4.73033130e-01
1.38818169e+00 -1.58514813e-01 6.52808771e-02 -1.17996529e-01
5.62587321e-01 5.69408424e-02 1.89958081e-01 9.64935184e-01
3.12920183e-01 7.52689362e-01 5.95210075e-01 -5.07834069e-02
-3.80577028e-01 -1.10591784e-01 2.52737314e-01 9.66642380e-01
5.65268040e-01 -1.97928205e-01 -6.55373752e-01 6.17560565e-01
-1.94613242e+00 -6.99495196e-01 -1.61347743e-02 2.01680994e+00
9.42492068e-01 -2.66583949e-01 1.78162128e-01 3.66454512e-01
7.53883362e-01 -8.91554654e-02 -9.16052878e-01 -1.03432022e-01
-1.85579807e-01 9.32382941e-02 6.93577230e-01 -1.01942867e-01
-9.80104029e-01 6.49643600e-01 6.50737715e+00 1.16025329e+00
-1.27635813e+00 8.02103519e-01 7.45049775e-01 -5.57814002e-01
-4.13853005e-02 -5.13970733e-01 -6.31907701e-01 7.12098539e-01
1.21569514e+00 1.91488117e-01 7.76783049e-01 1.38680303e+00
3.18203628e-01 1.23845473e-01 -1.10826147e+00 1.63595951e+00
-2.43073881e-01 -1.22345793e+00 -4.02916282e-01 -8.07363018e-02
8.67385447e-01 5.86915612e-01 1.13892391e-01 1.15156978e-01
6.64950132e-01 -5.78029454e-01 9.09986913e-01 4.82596576e-01
9.72883642e-01 -6.36917651e-01 8.64202797e-01 6.17269158e-01
-1.10062051e+00 -5.14946878e-01 -1.53985754e-01 7.92607144e-02
7.85480216e-02 1.12139845e+00 -7.04233825e-01 3.19240540e-01
1.25275028e+00 1.96958706e-01 -5.59566379e-01 8.91988635e-01
1.56336874e-01 1.01859725e+00 -6.91018641e-01 2.77637005e-01
-2.95593113e-01 2.96724528e-01 -3.99299078e-02 1.08657646e+00
1.14716575e-01 -7.24988505e-02 2.70882159e-01 3.42871308e-01
-5.30047417e-01 1.27743080e-01 -2.75312036e-01 4.35244739e-01
8.82149696e-01 1.35246646e+00 -3.37509274e-01 -3.27531964e-01
-4.65419859e-01 1.00629246e+00 1.20348386e-01 2.48989135e-01
-8.09319139e-01 1.21119611e-01 9.68260765e-01 -1.50335357e-01
2.98178434e-01 1.80976316e-01 -4.37922716e-01 -1.16506565e+00
3.48475426e-01 -1.15409088e+00 2.91402489e-01 -5.03123760e-01
-1.75128090e+00 7.30364382e-01 -5.81850111e-01 -1.06794083e+00
-1.34738848e-01 -1.58276074e-02 -1.12631492e-01 5.93038976e-01
-1.20062768e+00 -1.35494173e+00 -3.98523331e-01 1.09787202e+00
3.80316734e-01 -4.76145297e-02 1.24769640e+00 9.54222143e-01
-6.10885620e-01 9.89293575e-01 6.54320300e-01 -1.01881459e-01
7.60591984e-01 -9.48238850e-01 1.80963099e-01 6.11117244e-01
2.95976192e-01 2.34346166e-01 4.55775708e-01 -3.71488959e-01
-1.62191975e+00 -1.23548269e+00 6.72086895e-01 -4.72314358e-01
6.34279907e-01 -7.37912595e-01 -4.60111469e-01 4.88367498e-01
-1.06724970e-01 8.69786739e-01 8.95051956e-01 -2.58688945e-02
-6.36230588e-01 -7.55393088e-01 -1.66446590e+00 2.52717525e-01
1.15457976e+00 -9.64540899e-01 4.31922019e-01 7.55737543e-01
7.83403575e-01 -1.99992299e-01 -9.75616336e-01 1.85820520e-01
3.82549822e-01 -1.02899075e+00 6.41517580e-01 -2.29696393e-01
-4.92771327e-01 -3.01555544e-01 -5.95153630e-01 -8.26927006e-01
5.91439269e-02 -1.16337025e+00 -4.31023389e-01 1.73864269e+00
1.19164303e-01 -4.29356843e-01 1.00317669e+00 1.10061955e+00
1.07457243e-01 -5.61468422e-01 -1.24121761e+00 -6.98532939e-01
-6.35972977e-01 -1.09488845e+00 9.87587988e-01 7.67462671e-01
-3.46108168e-01 -5.98699190e-02 -5.54454327e-01 3.38289797e-01
7.84899056e-01 -2.54462421e-01 8.85255992e-01 -1.20265687e+00
-6.14539385e-01 2.49307305e-01 -1.20270647e-01 -9.35486495e-01
2.01818213e-01 -5.59958994e-01 1.00576587e-01 -1.07201648e+00
-2.30571777e-02 -1.01384234e+00 -3.42520446e-01 9.19619977e-01
3.46264958e-01 5.32943666e-01 3.09927195e-01 6.23299301e-01
-1.26386976e+00 3.25139165e-01 4.70044583e-01 -2.65090227e-01
-5.89417554e-02 2.51116425e-01 -5.89749277e-01 6.74767375e-01
6.11921191e-01 -6.31643891e-01 -3.91404420e-01 -8.75921130e-01
1.74412562e-03 -2.27389842e-01 2.36342862e-01 -1.19353187e+00
4.60810602e-01 1.50634259e-01 2.86583733e-02 -2.92758495e-01
3.69156659e-01 -1.64057052e+00 6.32037461e-01 -1.50807202e-02
-1.64986446e-01 -1.96189493e-01 -1.04249530e-01 6.59911811e-01
-3.00087761e-02 1.56671613e-01 5.67965388e-01 1.36236802e-01
-2.29731068e-01 2.37761453e-01 -1.89566568e-01 -2.33888373e-01
8.45278084e-01 1.19020410e-01 -4.49714512e-01 -5.04287064e-01
-8.31757724e-01 -7.64066726e-02 2.60888785e-01 1.19145781e-01
2.03696955e-02 -1.22203815e+00 -1.66121438e-01 3.09012562e-01
-4.51560020e-02 1.06515914e-01 4.68286961e-01 6.84103906e-01
-1.68740392e-01 5.95240220e-02 5.44578612e-01 -6.97854102e-01
-1.23876977e+00 5.25376439e-01 3.14711750e-01 -1.64004460e-01
-5.70288002e-01 9.32355285e-01 -3.18850875e-01 -4.50703323e-01
9.88753140e-01 -2.84426600e-01 5.98402858e-01 -6.68033734e-02
6.45810962e-01 5.83835542e-01 7.83630788e-01 -5.24705112e-01
-3.15923482e-01 1.22894645e-01 1.40712738e-01 2.04159737e-01
1.37243283e+00 -4.49660718e-01 1.05933407e-02 4.20958310e-01
1.32147849e+00 -2.80895960e-02 -1.50943851e+00 -7.79311001e-01
-1.15252420e-01 -3.16726476e-01 7.60178566e-02 -1.05130255e+00
-1.46862662e+00 4.73001838e-01 1.27767122e+00 2.40565121e-01
1.50375938e+00 1.16444990e-01 1.10398006e+00 3.79300922e-01
9.81451392e-01 -1.15949214e+00 1.94066241e-02 -3.60938162e-02
3.02591532e-01 -1.26763558e+00 -4.33471680e-01 -3.97626787e-01
-4.06305820e-01 5.85551381e-01 2.00562716e-01 4.27976668e-01
8.98410857e-01 7.63552070e-01 5.72025061e-01 -3.72170992e-02
-9.28750813e-01 2.35145122e-01 -3.19124460e-01 6.84482574e-01
-1.58794016e-01 1.27912641e-01 2.81743914e-01 1.07643425e+00
-8.10469443e-04 1.82043120e-01 3.11796982e-02 9.75337982e-01
1.88755356e-02 -1.14941132e+00 -5.76154172e-01 5.36368847e-01
-8.30292404e-01 2.72153574e-03 -1.13033347e-01 -2.26156458e-01
6.60643756e-01 1.63660765e+00 -1.92328677e-01 -4.42593426e-01
4.41025883e-01 2.17329144e-01 -9.20715481e-02 -5.29470801e-01
-1.10543609e+00 3.90976518e-01 -8.03720579e-02 -9.73314464e-01
-5.57636738e-01 -5.59196413e-01 -1.04680920e+00 -5.38089097e-01
-4.94643152e-01 4.02958728e-02 1.06826115e+00 7.16207623e-01
8.76888275e-01 2.83296943e-01 9.65046644e-01 -9.06199515e-01
-9.64602888e-01 -7.80834973e-01 -9.16161716e-01 5.90708315e-01
3.62149537e-01 -3.50654542e-01 -5.72038829e-01 4.33507979e-01] | [5.881345272064209, 6.264899730682373] |
07864df5-a38a-43ae-9b72-4725c4d2cb70 | shadow-optimization-from-structured-deep-edge | 1505.01589 | null | http://arxiv.org/abs/1505.01589v2 | http://arxiv.org/pdf/1505.01589v2.pdf | Shadow Optimization from Structured Deep Edge Detection | Local structures of shadow boundaries as well as complex interactions of
image regions remain largely unexploited by previous shadow detection
approaches. In this paper, we present a novel learning-based framework for
shadow region recovery from a single image. We exploit the local structures of
shadow edges by using a structured CNN learning framework. We show that using
the structured label information in the classification can improve the local
consistency of the results and avoid spurious labelling. We further propose and
formulate a shadow/bright measure to model the complex interactions among image
regions. The shadow and bright measures of each patch are computed from the
shadow edges detected in the image. Using the global interaction constraints on
patches, we formulate a least-square optimization problem for shadow recovery
that can be solved efficiently. Our shadow recovery method achieves
state-of-the-art results on the major shadow benchmark databases collected
under various conditions. | ['Karianto Leman', 'Li Shen', 'Teck Wee Chua'] | 2015-05-07 | shadow-optimization-from-structured-deep-edge-1 | http://openaccess.thecvf.com/content_cvpr_2015/html/Shen_Shadow_Optimization_From_2015_CVPR_paper.html | http://openaccess.thecvf.com/content_cvpr_2015/papers/Shen_Shadow_Optimization_From_2015_CVPR_paper.pdf | cvpr-2015-6 | ['shadow-detection'] | ['computer-vision'] | [ 7.19992518e-01 1.78963169e-01 3.51361409e-02 -5.41207850e-01
-6.13252282e-01 -3.97862583e-01 4.28520560e-01 -2.83324182e-01
-4.13497388e-02 6.65991724e-01 -6.05863221e-02 -1.68171436e-01
2.22478181e-01 -3.95638913e-01 -9.74254370e-01 -1.15437961e+00
4.02673632e-02 3.11736345e-01 9.29039598e-01 -4.05301712e-03
2.32767612e-01 6.15528405e-01 -1.51794052e+00 6.13973618e-01
8.15674245e-01 1.29456699e+00 4.75715727e-01 5.52806616e-01
4.90634590e-02 8.70635629e-01 -6.19466305e-01 2.84862071e-01
2.52853006e-01 -3.66617560e-01 -4.63425428e-01 3.42565507e-01
9.09241617e-01 -4.09722775e-01 -1.73635677e-01 6.23167694e-01
3.64056259e-01 1.59759559e-02 5.52403688e-01 -1.26010013e+00
-1.91546217e-01 -1.04130842e-01 -5.83489537e-01 1.06972069e-01
8.39628875e-02 -2.82836199e-01 1.08737457e+00 -1.19206786e+00
7.87614644e-01 1.07800746e+00 8.69541824e-01 -2.15459719e-01
-1.21398830e+00 -5.61202407e-01 2.74900973e-01 4.74303663e-01
-1.32317340e+00 -2.51854181e-01 1.02592456e+00 -9.05457288e-02
6.42583430e-01 2.97424346e-01 6.05679750e-01 7.78915465e-01
7.72367477e-01 1.04137623e+00 1.67466855e+00 -5.50765276e-01
1.41617924e-01 9.73078832e-02 2.60053247e-01 1.22236919e+00
4.61301021e-02 1.72256693e-01 -9.12255645e-01 -1.01861961e-01
4.56386656e-01 -4.82194982e-02 -4.29145545e-01 -6.68092251e-01
-8.31924975e-01 7.35604405e-01 6.66225016e-01 -1.90651909e-01
1.96936782e-02 1.60685956e-01 -9.24216285e-02 -2.22754255e-01
8.10403228e-01 9.26289149e-03 -5.40206611e-01 7.08803654e-01
-1.13160777e+00 1.45685598e-01 9.43992972e-01 7.24094033e-01
1.11630762e+00 -2.04902425e-01 -1.91854194e-01 5.80378771e-01
1.35348335e-01 9.20175850e-01 -5.39662957e-01 -1.10171306e+00
-5.44078425e-02 4.26882982e-01 1.87872946e-01 -1.20601475e+00
-5.16444147e-01 -5.20876825e-01 -5.02196670e-01 5.66107273e-01
4.75646257e-01 2.80337930e-01 -1.24977064e+00 1.14200306e+00
5.01850367e-01 4.34302241e-01 -1.88387379e-01 7.10071743e-01
1.00981617e+00 5.45240641e-01 -6.26122355e-01 -1.90553755e-01
1.18579316e+00 -1.39100957e+00 -7.23240674e-01 -5.38084865e-01
7.42397010e-02 -9.21168804e-01 9.15626466e-01 3.55062038e-01
-6.42682314e-01 -3.00204277e-01 -1.17281723e+00 -8.29266608e-02
-3.87508094e-01 2.43894160e-01 6.11290872e-01 2.55002648e-01
-8.81894886e-01 3.49685192e-01 -5.79753578e-01 1.44545045e-02
6.85892820e-01 3.16863179e-01 -3.94839933e-03 -3.55939746e-01
-6.34489536e-01 7.22986817e-01 -3.54253836e-02 4.20227975e-01
-1.14945507e+00 -7.50408769e-01 -8.30649734e-01 -1.15120746e-01
8.62638175e-01 -2.76140925e-02 1.02264524e+00 -1.02700424e+00
-1.05947101e+00 7.35120714e-01 -5.02704561e-01 -2.17471689e-01
3.03337216e-01 -3.01333189e-01 1.22751914e-01 1.73194155e-01
1.61233321e-01 3.24018061e-01 9.61479723e-01 -2.19608831e+00
-6.91805720e-01 -6.50224909e-02 2.41036899e-02 2.83965707e-01
1.58025146e-01 -5.24885245e-02 -6.66070938e-01 -4.67037380e-01
3.18802074e-02 -1.24988973e+00 -5.22943586e-02 4.30077463e-01
-5.60398996e-01 3.35278422e-01 1.47340763e+00 -7.18983889e-01
6.73439085e-01 -1.89658678e+00 -1.03029251e-01 3.68813753e-01
3.34617883e-01 3.85588929e-02 5.60539439e-02 7.00295046e-02
4.37730372e-01 -5.96589148e-01 -7.57861078e-01 -7.04221368e-01
-1.75171554e-01 6.85328364e-01 -5.57470381e-01 7.34951556e-01
5.98837296e-03 1.07849038e+00 -6.55135512e-01 -8.03117275e-01
2.49150202e-01 7.49843180e-01 -1.11625455e-01 4.69700158e-01
-2.44586706e-01 3.64387721e-01 -3.42949748e-01 1.05571651e+00
1.14053953e+00 -2.19932914e-01 8.39255899e-02 -3.64885986e-01
-9.63335484e-02 1.33480616e-02 -8.66669595e-01 1.24679792e+00
-5.30114830e-01 1.21942878e+00 5.04965127e-01 -6.92727566e-01
6.24572158e-01 -2.06938997e-01 4.80673462e-01 -8.11434150e-01
2.86660735e-02 7.06501901e-02 -2.48877779e-01 -1.79400921e-01
2.75019079e-01 1.99215934e-02 1.36471391e-01 6.19189441e-01
-1.89116821e-01 -2.86128581e-01 -4.58417594e-01 6.43538013e-02
1.10883093e+00 2.40149230e-01 2.87368596e-02 -5.13723671e-01
1.98414311e-01 -7.42869452e-04 5.89588106e-01 9.36166704e-01
6.38466477e-02 9.39755678e-01 2.98281312e-01 -3.56143475e-01
-5.32818079e-01 -1.16451263e+00 -4.13399965e-01 1.38785958e+00
6.67059720e-01 -6.43438995e-02 -7.69683123e-01 -9.08049464e-01
1.63700670e-01 3.98695141e-01 -9.33762550e-01 1.50289744e-01
-8.04966271e-01 -8.62407804e-01 2.38170192e-01 5.48089325e-01
5.73301077e-01 -1.29541862e+00 -1.01261783e+00 -1.67808548e-01
-3.20679903e-01 -1.57158053e+00 -4.45696592e-01 8.99268568e-01
-4.45783943e-01 -1.20300019e+00 -4.96771544e-01 -8.50383162e-01
8.42013836e-01 8.78472626e-01 1.24108887e+00 4.46161389e-01
-8.47393632e-01 2.26965889e-01 -3.34008306e-01 -6.51089609e-01
1.35883033e-01 -3.62212420e-01 -5.39049685e-01 2.42051572e-01
-4.36955124e-01 -3.24965656e-01 -7.01392353e-01 6.48555696e-01
-7.77764678e-01 2.67646551e-01 7.55359948e-01 1.11441147e+00
7.26485133e-01 2.37737283e-01 -1.67118311e-01 -1.17957199e+00
3.80188674e-02 -1.17069960e-01 -6.82994246e-01 4.47092950e-01
-8.31616640e-01 3.35008353e-02 2.73756623e-01 -2.17046961e-01
-1.56849682e+00 4.70379651e-01 6.28930509e-01 -2.42167503e-01
-9.88516510e-02 -9.29189697e-02 -2.50339687e-01 -8.58420074e-01
3.75816882e-01 1.94026530e-01 -4.08256710e-01 -1.73718750e-01
3.50174487e-01 2.59415954e-01 5.96767545e-01 -5.01199067e-01
1.01089430e+00 1.31635153e+00 1.41854882e-01 -9.09951031e-01
-1.46209669e+00 -8.27211142e-01 -9.85430777e-01 -5.96240759e-01
7.08392143e-01 -6.48485184e-01 -4.59745526e-01 6.84853613e-01
-1.23941612e+00 -1.00921917e+00 5.15665077e-02 -3.18170100e-01
-3.93071979e-01 4.33086187e-01 -2.98509747e-01 -7.80721545e-01
-2.23562047e-01 -1.08869517e+00 1.70658326e+00 1.46171734e-01
2.82349229e-01 -1.14703083e+00 -2.04668101e-02 5.55170894e-01
3.08887303e-01 4.90446806e-01 7.53020167e-01 1.34190932e-01
-1.27985728e+00 1.64795950e-01 -6.84457421e-01 3.49948436e-01
4.51285876e-02 -3.48687410e-01 -1.36343646e+00 -6.91149607e-02
6.25951812e-02 -1.93314821e-01 1.36408889e+00 5.65486968e-01
1.09516275e+00 -8.03184956e-02 -6.32559955e-01 8.87833357e-01
1.43566680e+00 -2.95898621e-03 5.22619963e-01 1.92993924e-01
1.08850849e+00 8.02107513e-01 1.05406380e+00 3.25616539e-01
1.92690313e-01 7.79142201e-01 4.52613562e-01 -8.64252210e-01
-5.15055418e-01 1.65797487e-01 2.28278741e-01 1.50511637e-01
-3.18034664e-02 -3.76510620e-01 -8.48690927e-01 3.98575485e-01
-2.02534986e+00 -5.12644053e-01 -3.64820272e-01 1.71221256e+00
6.22392297e-01 1.77486748e-01 -5.31379759e-01 -1.50072828e-01
3.16216320e-01 5.90285361e-01 -5.18113732e-01 -1.51125202e-02
-4.58230913e-01 4.89430398e-01 9.87199426e-01 9.34688509e-01
-1.13171005e+00 1.29083252e+00 7.02540207e+00 7.30389714e-01
-8.54584873e-01 1.59524381e-01 5.69292545e-01 2.98961759e-01
-2.72496790e-01 3.51589918e-01 -8.15584362e-01 -9.93528636e-04
2.28125915e-01 5.42280376e-01 4.61695701e-01 6.46883190e-01
2.62658056e-02 -9.65727687e-01 -7.76488900e-01 7.10179925e-01
6.19508326e-01 -1.31338906e+00 -5.87775767e-01 1.59039721e-01
1.16639674e+00 1.36344731e-01 1.70798898e-01 -2.96914279e-01
2.56801456e-01 -1.15856850e+00 8.27344537e-01 5.51095188e-01
6.08094692e-01 -5.36769986e-01 9.21828508e-01 8.17227587e-02
-1.53552878e+00 -1.22643337e-01 -1.72416791e-01 2.09188476e-01
1.15901090e-01 1.04176259e+00 -1.13943291e+00 3.88006628e-01
1.01239109e+00 5.97386658e-01 -7.39047706e-01 7.30422199e-01
-3.77822578e-01 5.28146386e-01 -2.15563297e-01 2.86477089e-01
1.61758326e-02 -2.30333000e-01 3.22040945e-01 1.53511834e+00
-2.48103872e-01 6.02147775e-03 3.97288442e-01 8.27941298e-01
9.38509107e-02 -1.76102459e-01 -5.33994377e-01 4.48525518e-01
-3.21515240e-02 1.52539992e+00 -1.50567913e+00 -1.43668875e-01
-3.34560424e-01 1.29170072e+00 2.29621276e-01 6.57927573e-01
-1.05861652e+00 -9.48548913e-02 2.21257657e-01 1.91474736e-01
4.25626159e-01 -2.80579567e-01 -4.93303508e-01 -7.86003172e-01
2.19949961e-01 -7.34112680e-01 -6.48040250e-02 -9.95231032e-01
-8.37590873e-01 3.52512121e-01 1.47843421e-01 -8.27927351e-01
4.67668474e-01 -6.71821058e-01 -8.62346590e-01 4.00749922e-01
-1.96881783e+00 -1.45462680e+00 -1.06361544e+00 5.68136036e-01
5.95602095e-01 1.52358904e-01 6.03411436e-01 -7.60453567e-02
-2.11919308e-01 1.56238571e-01 1.94752887e-01 -3.65343541e-02
9.07703638e-01 -1.44174612e+00 8.03333372e-02 7.24254847e-01
1.19175330e-01 6.90590963e-02 7.30638862e-01 -7.55282760e-01
-1.22768307e+00 -1.02263403e+00 5.53543389e-01 -4.84205216e-01
4.13715273e-01 -8.19340527e-01 -8.65569711e-01 3.30086887e-01
2.72128016e-01 4.19424653e-01 3.31117451e-01 -2.67819669e-02
-4.90200788e-01 -1.97410539e-01 -7.82147467e-01 2.89162755e-01
1.17524874e+00 -8.63458633e-01 -1.32795066e-01 6.81758165e-01
5.01580179e-01 -6.09012306e-01 -2.03400135e-01 6.30052388e-01
6.80410206e-01 -1.20399702e+00 1.22675908e+00 1.32957548e-01
2.39946693e-01 -2.70270377e-01 -3.20943862e-01 -9.77096915e-01
1.21519482e-02 -3.41200024e-01 -1.33703545e-01 8.57280076e-01
2.42798179e-01 -4.49620128e-01 1.00976396e+00 1.51663676e-01
-4.01904941e-01 -1.16564226e+00 -8.56518567e-01 -6.50118887e-01
-5.40044308e-01 -2.98680276e-01 -4.31204066e-02 3.47256690e-01
-6.88699365e-01 1.08412273e-01 -5.13991535e-01 4.61928755e-01
1.08813477e+00 8.19037855e-01 7.94152260e-01 -1.25584102e+00
-4.38069314e-01 8.57776254e-02 5.33452481e-02 -8.98231566e-01
6.55611336e-01 -5.84029257e-01 1.02826321e+00 -1.69798589e+00
6.92458510e-01 -5.39461255e-01 -1.80169418e-01 6.31164610e-01
-1.31409958e-01 6.86016977e-01 1.13440976e-02 2.43419647e-01
-8.68065238e-01 6.56648397e-01 1.23035896e+00 -2.55768090e-01
-1.18481643e-01 2.97247022e-02 -1.81005925e-01 8.83196354e-01
5.47865808e-01 -6.55210853e-01 -3.02762866e-01 -4.08343449e-02
-1.63131729e-01 -2.03801498e-01 8.17414463e-01 -8.66598487e-01
2.00458899e-01 -3.33923638e-01 5.73004365e-01 -9.14367497e-01
8.77391219e-01 -7.96376705e-01 -2.81760842e-01 2.55245835e-01
-3.54920030e-02 -5.10024071e-01 1.40199745e-02 8.71917367e-01
4.65426706e-02 2.40746200e-01 9.24438834e-01 8.16517323e-02
-6.06980503e-01 7.09363725e-03 -2.70543218e-01 -7.89451450e-02
9.50974226e-01 -1.23328574e-01 -4.54477251e-01 -4.86208349e-01
-3.11054587e-01 1.12904362e-01 5.06859720e-01 1.64338663e-01
9.51018035e-01 -8.74687195e-01 -2.64763355e-01 9.75661129e-02
4.83470932e-02 1.92975029e-01 -3.25559676e-02 1.00551665e+00
-5.83526969e-01 2.79122889e-01 -1.60137072e-01 -7.96515524e-01
-1.93086612e+00 1.55499011e-01 5.17971456e-01 -9.14622024e-02
-9.06735361e-01 8.95621181e-01 1.02615654e+00 -2.39051342e-01
5.26526093e-01 -3.81100804e-01 1.97251990e-01 -1.52817667e-01
1.62882268e-01 1.60438210e-01 1.31549001e-01 -8.62214208e-01
-5.02983868e-01 8.21512043e-01 1.74406290e-01 -1.43608883e-01
1.23710966e+00 2.65241954e-02 -2.82295972e-01 4.41503078e-01
1.22975433e+00 2.42385074e-01 -1.73289490e+00 -3.89470935e-01
-9.10702869e-02 -5.88992417e-01 3.92231941e-01 -8.03231239e-01
-1.17490065e+00 7.05308437e-01 7.27118731e-01 -2.90253550e-01
1.16214705e+00 3.88938367e-01 9.07307088e-01 5.19145489e-01
3.17193866e-01 -1.31328785e+00 5.06863713e-01 5.26561320e-01
9.28121924e-01 -1.36820412e+00 4.11096543e-01 -1.02642500e+00
-4.81175244e-01 1.02821374e+00 3.31561863e-01 -1.95267692e-01
9.94347215e-01 8.34723175e-01 2.00453490e-01 -4.43133295e-01
-4.13465261e-01 -4.03069943e-01 6.52116299e-01 7.04150379e-01
-1.27869546e-01 7.66910017e-02 1.18922241e-01 1.69665486e-01
2.63027072e-01 -5.64136922e-01 3.93093586e-01 1.10605538e+00
-7.90416837e-01 -8.99439692e-01 -7.49560297e-01 1.72361523e-01
-6.70291558e-02 -2.19746560e-01 -1.02816260e+00 6.56070530e-01
4.55215424e-01 9.58060086e-01 -1.61800489e-01 -2.49000371e-01
9.93819535e-03 -2.04044163e-01 5.88410079e-01 -5.55571914e-01
-7.39681721e-02 3.45555842e-01 9.49776843e-02 -8.66748571e-01
-5.93537807e-01 -5.90630949e-01 -1.54430664e+00 1.93139091e-01
-6.13416791e-01 -2.21953750e-01 6.16853595e-01 1.19761837e+00
5.81477210e-03 6.30390584e-01 6.42650425e-01 -1.40330720e+00
1.16381787e-01 -4.95691001e-01 -6.52389705e-01 1.72592085e-02
5.83958149e-01 -1.10351348e+00 -5.71570277e-01 1.47369325e-01] | [10.843816757202148, -4.1017231941223145] |
c118f282-7b43-4241-b96e-239b544743a0 | sleep-stage-classification-using | 1909.11141 | null | http://arxiv.org/abs/1909.11141v1 | http://arxiv.org/pdf/1909.11141v1.pdf | Sleep Stage Classification Using Bidirectional LSTM in Wearable Multi-sensor Systems | Understanding the sleep quality and architecture is essential to human
being's health, which is usually represented using multiple sleep stages. A
standard sleep stage determination requires Electroencephalography (EEG)
signals during the expensive and labor-intensive Polysomnography (PSG) test. To
overcome this inconvenience, cardiorespiratory signals are proposed for the
same purpose because of the easy and comfortable acquisition by simplified
devices. In this paper, we leverage our low-cost wearable multi-sensor system
to acquire the cardiorespiratory signals from subjects. Three novel features
are designed during the feature extraction. We then apply a Bi-directional
Recurrent Neural Network architecture with Long Short-term Memory (BLSTM) to
predict the four-class sleep stages. Our prediction accuracy is 80.25% on a
large public dataset (417 subjects), and 80.75% on our 32 enrolled subjects,
respectively. Our results outperform the previous works which either used small
data sets and had the potential over-fitting issues, or used the conventional
machine learning methods on large data sets. | [] | 2019-09-24 | null | null | null | null | ['sleep-quality-prediction'] | ['medical'] | [ 1.11165307e-01 -3.03558409e-01 4.39442992e-02 -5.83069324e-01
-4.20515597e-01 5.73539287e-02 -1.93733469e-01 -2.75619626e-01
-5.53305566e-01 9.02944028e-01 1.11831315e-01 -1.21305615e-01
-9.85477939e-02 -4.54689056e-01 -1.49245024e-01 -7.00158298e-01
2.50268411e-02 -4.91330512e-02 -2.01439206e-02 8.20029825e-02
7.53847733e-02 1.65970642e-02 -1.45629716e+00 -2.75229290e-02
1.08161223e+00 1.28028643e+00 3.69220644e-01 4.81627703e-01
3.24136108e-01 3.18312526e-01 -6.77506447e-01 -9.60241929e-02
-1.01641826e-01 -5.96004307e-01 -2.45006397e-01 -1.33699834e-01
-2.71472275e-01 -5.30278198e-02 -2.15910643e-01 9.64980841e-01
7.97727466e-01 -1.76017396e-02 6.21096529e-02 -1.10552418e+00
-5.50428927e-01 2.19533399e-01 -3.59456658e-01 8.20952773e-01
4.46219802e-01 -3.88292708e-02 6.21728897e-01 -5.67362249e-01
-4.01765227e-01 3.53318125e-01 8.67557704e-01 8.47341597e-01
-7.98095703e-01 -1.15883470e+00 -2.08010420e-01 6.48097873e-01
-1.49898243e+00 -7.64033854e-01 8.49552035e-01 -1.97470561e-01
1.11022830e+00 2.67507881e-01 1.22275233e+00 1.23533690e+00
7.70359397e-01 2.44333431e-01 1.45004857e+00 1.34192675e-01
2.60356098e-01 2.33428150e-01 3.51589292e-01 4.30370480e-01
2.51111984e-01 -1.34937719e-01 -6.55005634e-01 -3.48267742e-02
4.55766231e-01 6.98902965e-01 -3.73771846e-01 4.69849557e-01
-9.75531161e-01 3.21129858e-01 3.57652366e-01 5.39106965e-01
-5.23595214e-01 -1.53709561e-01 2.96413749e-01 2.94360846e-01
4.66325641e-01 3.60099226e-01 -6.10457003e-01 -6.06026351e-01
-1.22570860e+00 -5.89137912e-01 6.95116639e-01 6.70039475e-01
4.04809594e-01 1.06394805e-01 1.99876074e-02 7.03412652e-01
4.43781674e-01 4.75175679e-01 1.20620573e+00 -5.53292036e-01
2.52377182e-01 8.41234207e-01 2.11658282e-03 -8.69517028e-01
-9.81350243e-01 -6.11680269e-01 -1.17108202e+00 -5.40613234e-01
-1.93536952e-01 -1.82158455e-01 -5.03286541e-01 1.41600788e+00
-2.72545785e-01 5.58893740e-01 -9.40710083e-02 8.78692687e-01
1.00292587e+00 4.54494208e-01 1.67879853e-02 -5.70417464e-01
1.63032663e+00 -7.12542176e-01 -9.58997726e-01 -5.97779572e-01
7.48139918e-02 -2.86098778e-01 1.24667490e+00 6.63775206e-01
-9.55051422e-01 -5.40637434e-01 -1.32945693e+00 1.15255989e-01
-1.72971517e-01 3.67460221e-01 6.44709170e-01 9.64551032e-01
-1.00814128e+00 6.49314761e-01 -1.36068594e+00 -2.91883320e-01
3.67332309e-01 8.46040666e-01 -2.10242674e-01 5.41272283e-01
-1.16166234e+00 7.16985285e-01 -1.58830091e-01 5.56642652e-01
-5.57474077e-01 -3.68230253e-01 -6.88998938e-01 9.17298421e-02
2.20391788e-02 -7.10452735e-01 8.65015507e-01 -6.30749047e-01
-1.78127062e+00 7.66321182e-01 -5.70033729e-01 -4.76840317e-01
-2.80463696e-01 -3.86472702e-01 -8.32361162e-01 1.55722976e-01
-1.19423106e-01 5.37413824e-03 9.43171263e-01 -2.29823887e-01
-7.59885162e-02 -7.96710074e-01 -4.02135789e-01 -2.34963093e-02
-6.38727903e-01 3.24781179e-01 -8.94502103e-02 -3.86178911e-01
2.70102948e-01 -9.12070394e-01 -1.41500965e-01 -5.55484474e-01
-2.03816682e-01 -1.74353048e-01 2.41931006e-01 -7.30814278e-01
1.53603625e+00 -2.10102439e+00 -7.40392059e-02 -2.38974646e-01
5.12148559e-01 2.04759583e-01 4.00191844e-01 -6.51543160e-05
2.86772437e-02 -8.59301910e-02 -1.19253593e-02 -6.79731429e-01
-2.04554543e-01 1.47695735e-01 -4.69920151e-02 6.43700957e-01
-1.88687101e-01 9.62829113e-01 -5.84988177e-01 -3.44626129e-01
1.58417985e-01 4.88353968e-01 -2.76085645e-01 5.32500088e-01
6.18330359e-01 4.96245474e-01 -4.51559842e-01 6.45009637e-01
2.10706115e-01 -4.93873626e-01 -1.50502026e-01 -1.98058158e-01
-5.06982394e-02 7.61432707e-01 -5.16602755e-01 2.12993431e+00
-5.24868846e-01 4.37893927e-01 -1.47639707e-01 -8.35464895e-01
1.09512532e+00 5.36181390e-01 3.88002992e-01 -8.27116013e-01
4.07228798e-01 1.57137424e-01 1.06743380e-01 -7.57033706e-01
-4.90779581e-04 -4.88016486e-01 -4.20212038e-02 3.93036038e-01
-9.97313038e-02 3.39782000e-01 -2.41349384e-01 -2.14912504e-01
1.13243139e+00 -4.18763235e-02 5.68483114e-01 -3.66509080e-01
4.38071012e-01 -8.15522909e-01 1.05845010e+00 3.17577302e-01
-4.95238543e-01 5.42799711e-01 7.28013963e-02 -4.22057390e-01
-2.77699083e-01 -7.62076199e-01 -1.93795219e-01 6.47198617e-01
1.36578500e-01 -8.25271726e-01 -5.12426257e-01 -2.70651489e-01
-4.38450903e-01 3.68557483e-01 -5.73286235e-01 -6.39641047e-01
-2.58927941e-01 -1.20192790e+00 4.97407287e-01 7.57454813e-01
5.63162565e-01 -1.15047240e+00 -1.00059652e+00 1.90127894e-01
-1.13927491e-01 -9.00603950e-01 -3.48371118e-01 2.64445335e-01
-9.93383110e-01 -8.54213893e-01 -5.33196270e-01 -4.93683934e-01
3.13574135e-01 2.19365582e-01 9.11658108e-01 3.99654359e-02
-2.38771543e-01 -1.51399150e-01 -1.76205814e-01 -3.61471534e-01
5.13762951e-01 1.84420303e-01 7.69647300e-01 1.43416584e-01
1.06921077e+00 -1.37609804e+00 -1.08236003e+00 4.64932352e-01
-2.74212003e-01 2.99922079e-02 8.05066168e-01 5.91749907e-01
5.35079896e-01 -2.59030730e-01 1.05677652e+00 -2.48324737e-01
6.20077312e-01 -6.11287713e-01 -2.75548130e-01 1.50885716e-01
-9.70903873e-01 -3.02587509e-01 8.40610266e-01 -2.94678122e-01
-6.27786577e-01 -2.33221427e-01 -3.69233668e-01 -2.91179836e-01
-8.31079856e-02 1.96754530e-01 -2.20564961e-01 5.22172451e-02
3.13452184e-01 5.22381842e-01 -2.48686954e-01 -6.70366168e-01
-4.94768053e-01 1.18528700e+00 4.26906258e-01 -7.95546696e-02
2.79135972e-01 1.24292701e-01 -2.61350632e-01 -9.36947703e-01
-9.55959558e-01 -4.80692953e-01 -5.30942261e-01 4.25850935e-02
9.73713279e-01 -1.11559761e+00 -1.12115526e+00 3.64594936e-01
-7.94452667e-01 -8.89509171e-02 1.48119614e-01 9.40967917e-01
-1.97231635e-01 2.43053600e-01 -6.36110187e-01 -9.02804255e-01
-1.02774906e+00 -1.01345396e+00 9.05013382e-01 5.37093282e-01
-3.24765295e-01 -6.58294082e-01 1.26945943e-01 4.41676229e-01
5.18602490e-01 -8.94722641e-02 3.11701387e-01 -4.73457932e-01
-1.82511359e-02 -2.15623707e-01 -7.48362616e-02 3.60416472e-01
5.42497993e-01 -5.95494330e-01 -1.16840994e+00 -4.30573702e-01
8.21233749e-01 -1.90157682e-01 4.72748727e-01 4.59115177e-01
1.33525920e+00 -1.35185853e-01 -2.11156532e-01 1.06712139e+00
9.95453238e-01 4.60221708e-01 6.73813999e-01 3.49535882e-01
5.40328979e-01 1.07027374e-01 8.94352496e-02 4.49136466e-01
7.36484289e-01 3.79624665e-01 5.97311184e-03 1.11623548e-01
3.30577582e-01 -9.72996932e-03 6.34925783e-01 1.49294972e+00
-1.59217551e-01 2.00026244e-01 -6.25097513e-01 2.91361868e-01
-1.46042371e+00 -7.05131114e-01 4.99031320e-02 2.07982779e+00
9.04296815e-01 2.14480579e-01 2.57117927e-01 2.27446154e-01
3.80492479e-01 1.11347875e-02 -7.53028095e-01 -2.58514613e-01
1.77976906e-01 3.93117547e-01 1.85506359e-01 -1.66996300e-01
-7.35426188e-01 3.38927120e-01 6.38934517e+00 1.90667093e-01
-1.22418153e+00 3.91390234e-01 3.75495076e-01 -7.09859192e-01
1.27395421e-01 -3.05909425e-01 -7.47934520e-01 1.07688284e+00
1.75199258e+00 -1.13421552e-01 6.39327407e-01 6.79014087e-01
5.66145897e-01 -7.31505379e-02 -8.73655677e-01 1.71565211e+00
3.37368131e-01 -7.20866323e-01 -7.73259461e-01 -5.38699366e-02
1.04069687e-01 2.55714118e-01 -1.71955839e-01 3.23982984e-01
-7.13820636e-01 -1.15129006e+00 8.97918344e-02 9.04374361e-01
8.83265316e-01 -5.85454941e-01 6.47403479e-01 4.80503649e-01
-1.12142086e+00 -2.55868912e-01 -4.60466117e-01 -6.59707785e-01
1.61262527e-01 8.66599381e-01 -2.98369974e-01 2.12500229e-01
1.09332895e+00 1.24772084e+00 -8.36794257e-01 1.02255797e+00
-3.48266810e-01 9.42168355e-01 -4.32886153e-01 -3.15553039e-01
-3.90034407e-01 -4.81984407e-01 2.37869427e-01 7.39502251e-01
4.15068209e-01 2.98108101e-01 -3.16081077e-01 9.73789394e-01
-4.96038534e-02 -1.99841350e-01 -4.05028015e-01 3.00470944e-02
2.01054603e-01 1.53976917e+00 -5.76364994e-01 -4.32009958e-02
-5.39438546e-01 1.06272912e+00 -2.87500713e-02 7.68177211e-02
-7.70634890e-01 -7.04539716e-01 4.60245848e-01 -8.69036764e-02
-2.72724420e-01 -1.70853555e-01 -5.78934073e-01 -1.70802712e+00
3.43628436e-01 -5.26770353e-01 2.28300318e-01 -8.21250975e-01
-1.17082632e+00 9.87555683e-01 -3.35399866e-01 -1.21147954e+00
-7.92961940e-02 -1.86403483e-01 -9.11993563e-01 8.74204457e-01
-1.44498265e+00 -7.54246235e-01 -5.67712724e-01 6.74293756e-01
4.85742986e-01 -1.39061613e-02 9.50812995e-01 7.48024404e-01
-1.31942594e+00 6.66724563e-01 -2.58228064e-01 -2.12522343e-01
5.93685508e-01 -1.05434358e+00 2.26918474e-01 8.25794637e-01
-1.65266886e-01 1.18933094e+00 3.14967901e-01 -4.55878913e-01
-1.52088022e+00 -8.28480959e-01 1.02237153e+00 -5.06115913e-01
6.51450455e-01 -6.14914954e-01 -7.61606872e-01 7.27491915e-01
1.24374822e-01 -1.19260617e-01 1.26823688e+00 5.93033917e-02
3.99020284e-01 -5.23842692e-01 -1.04937661e+00 1.78509846e-01
1.00984144e+00 -6.40338957e-01 -1.07580280e+00 -1.17696844e-01
5.08363843e-01 -2.12215930e-02 -1.11431789e+00 3.23370934e-01
7.98713565e-01 -9.69996333e-01 6.37289405e-01 -3.02646935e-01
3.97534184e-02 -1.98836952e-01 3.08667812e-02 -1.12225473e+00
-2.38817140e-01 -1.00879800e+00 -3.77751172e-01 1.06765163e+00
1.38794988e-01 -9.28884447e-01 8.08356285e-01 9.06779766e-01
-3.10484350e-01 -1.07469845e+00 -9.12331045e-01 -4.88318205e-01
-6.12911224e-01 -4.13697422e-01 7.95671046e-01 3.94091457e-01
3.85447711e-01 7.22924531e-01 -6.79059148e-01 -2.03106534e-02
2.80024111e-01 2.41116121e-01 2.17001721e-01 -1.30133200e+00
-3.26869994e-01 8.11440125e-02 -4.62054342e-01 -1.00248551e+00
6.43883049e-02 -7.43024945e-01 -1.37101486e-01 -1.37056494e+00
3.99956226e-01 -1.73073426e-01 -9.53094423e-01 4.49152410e-01
-2.07442030e-01 5.76594472e-01 -3.30316693e-01 2.04165936e-01
-8.52525890e-01 8.79850507e-01 7.52053440e-01 2.44050026e-01
-4.92536336e-01 5.19057870e-01 -1.04167426e+00 9.66053843e-01
1.10457039e+00 -5.67443013e-01 -6.07106984e-01 -3.89761418e-01
5.12586713e-01 1.84319705e-01 1.48030922e-01 -1.33699453e+00
3.71612132e-01 2.81102955e-01 7.42716491e-01 -3.96500409e-01
7.62418509e-01 -7.57465899e-01 7.99337849e-02 5.17074168e-01
3.42113785e-02 4.47468191e-01 6.85526878e-02 3.89922589e-01
6.90196231e-02 1.59462050e-01 4.23473448e-01 -1.46211786e-02
-2.40028843e-01 3.26953590e-01 -4.13924783e-01 -2.72321105e-01
8.45252037e-01 -3.76978636e-01 -1.09062657e-01 -1.34823501e-01
-7.49581635e-01 1.35629416e-01 2.57303089e-01 2.79091239e-01
9.56654310e-01 -1.22314477e+00 -1.87536821e-01 7.94505060e-01
-6.41697496e-02 -3.30477029e-01 2.52933890e-01 1.51628840e+00
4.39928770e-02 4.25092548e-01 -4.33640808e-01 -4.68609512e-01
-1.07910216e+00 2.72414297e-01 4.39240724e-01 1.67469144e-01
-7.97523558e-01 6.55156970e-01 -3.32146436e-01 8.87014344e-02
1.44405186e-01 -5.95412493e-01 -5.24959683e-01 5.72395697e-02
8.21963251e-01 5.18821478e-01 3.96449924e-01 -4.23266083e-01
-7.67961502e-01 6.08228922e-01 2.02513292e-01 3.35144252e-01
1.56950140e+00 -3.94372135e-01 -3.15928727e-01 9.06381547e-01
1.20198035e+00 -1.42809093e-01 -9.00563776e-01 1.39614522e-01
-6.99792355e-02 -2.36260280e-01 8.86066183e-02 -5.59817612e-01
-1.11052263e+00 1.04473329e+00 9.10411835e-01 1.58063725e-01
1.50520146e+00 -2.67075002e-01 1.37393689e+00 5.41873693e-01
6.63411617e-01 -7.34663904e-01 -2.68104196e-01 -3.19030434e-02
4.34696883e-01 -1.08996439e+00 1.27738088e-01 1.86891362e-01
-6.21466577e-01 9.32019591e-01 5.61881900e-01 -2.95411915e-01
7.92734802e-01 1.59319788e-02 -6.25869483e-02 -4.31426823e-01
-8.33557248e-01 3.24446738e-01 3.66011262e-01 2.39018396e-01
3.61296326e-01 5.78654669e-02 -4.17845488e-01 1.57220829e+00
-6.15646601e-01 3.97428423e-01 3.62869322e-01 6.53517604e-01
-3.18920076e-01 -8.59432161e-01 -2.06441386e-03 1.05190301e+00
-9.36800122e-01 -2.17343912e-01 -1.12274736e-01 1.49731353e-01
8.11113715e-02 1.38978374e+00 -4.59895842e-02 -7.36507773e-01
2.31811598e-01 1.42410651e-01 3.66910309e-01 -7.59527981e-01
-6.54003978e-01 -1.23480774e-01 -2.54059196e-01 -8.08515191e-01
-6.85078919e-01 -5.63570738e-01 -1.14641845e+00 3.34826671e-02
-3.02022725e-01 3.70890290e-01 6.23038113e-01 1.04048288e+00
7.77065635e-01 6.58877254e-01 7.32526541e-01 -7.87915170e-01
-1.67874187e-01 -1.38219559e+00 -8.19987118e-01 1.38529509e-01
5.56677818e-01 -6.62206173e-01 -4.58189785e-01 -1.55992001e-01] | [13.506937980651855, 3.4969656467437744] |
61f1b4bb-4c1f-4341-9bdb-29873c137c02 | learning-hierarchical-cross-modal-association | 2203.13161 | null | https://arxiv.org/abs/2203.13161v1 | https://arxiv.org/pdf/2203.13161v1.pdf | Learning Hierarchical Cross-Modal Association for Co-Speech Gesture Generation | Generating speech-consistent body and gesture movements is a long-standing problem in virtual avatar creation. Previous studies often synthesize pose movement in a holistic manner, where poses of all joints are generated simultaneously. Such a straightforward pipeline fails to generate fine-grained co-speech gestures. One observation is that the hierarchical semantics in speech and the hierarchical structures of human gestures can be naturally described into multiple granularities and associated together. To fully utilize the rich connections between speech audio and human gestures, we propose a novel framework named Hierarchical Audio-to-Gesture (HA2G) for co-speech gesture generation. In HA2G, a Hierarchical Audio Learner extracts audio representations across semantic granularities. A Hierarchical Pose Inferer subsequently renders the entire human pose gradually in a hierarchical manner. To enhance the quality of synthesized gestures, we develop a contrastive learning strategy based on audio-text alignment for better audio representations. Extensive experiments and human evaluation demonstrate that the proposed method renders realistic co-speech gestures and outperforms previous methods in a clear margin. Project page: https://alvinliu0.github.io/projects/HA2G | ['Bolei Zhou', 'Bo Dai', 'Wayne Wu', 'Xiaowei Zhou', 'Xinyi Lin', 'Rui Qian', 'Yinghao Xu', 'Hang Zhou', 'Qianyi Wu', 'Xian Liu'] | 2022-03-24 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Liu_Learning_Hierarchical_Cross-Modal_Association_for_Co-Speech_Gesture_Generation_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Liu_Learning_Hierarchical_Cross-Modal_Association_for_Co-Speech_Gesture_Generation_CVPR_2022_paper.pdf | cvpr-2022-1 | ['gesture-generation'] | ['robots'] | [ 2.78657436e-01 1.08130917e-01 -1.02214843e-01 -3.12199503e-01
-1.13100016e+00 -5.04325569e-01 7.36793101e-01 -5.78960419e-01
2.77397484e-01 2.27570295e-01 9.64282632e-01 2.23439604e-01
2.14302853e-01 -3.90088886e-01 -6.73528373e-01 -4.92424011e-01
1.65430889e-01 6.64094925e-01 2.32131988e-01 -4.10829484e-01
-4.49107215e-02 2.85023242e-01 -1.77111816e+00 6.01638615e-01
3.63330781e-01 6.73806727e-01 1.53360575e-01 1.06551015e+00
-3.56019080e-01 5.97771347e-01 -9.00457144e-01 -2.51643181e-01
1.90131485e-01 -8.18939149e-01 -8.26538861e-01 1.90341860e-01
4.97200966e-01 -5.31678677e-01 -2.65231699e-01 6.41033351e-01
8.04540455e-01 2.99278557e-01 6.55617476e-01 -1.35763454e+00
-1.43502444e-01 9.69420075e-01 -3.99654239e-01 -4.69363302e-01
8.83246124e-01 1.56011388e-01 1.36668718e+00 -8.14501882e-01
6.92615211e-01 1.70884156e+00 4.97622609e-01 9.01166201e-01
-1.00435698e+00 -9.60958123e-01 4.07486320e-01 -6.16998374e-02
-1.46388936e+00 -4.98072922e-01 7.97189713e-01 -5.01447618e-01
6.20182574e-01 6.38369918e-01 1.01689494e+00 1.46908295e+00
-4.57278103e-01 1.22977984e+00 6.07482195e-01 -3.38165760e-01
-8.62975121e-02 -6.91908479e-01 -3.15909952e-01 5.56949377e-01
-5.03270507e-01 1.36811793e-01 -1.18686688e+00 -1.29120216e-01
1.16286576e+00 -6.88194409e-02 8.36804733e-02 -1.99117213e-01
-1.64945424e+00 5.36116362e-01 2.46504351e-01 1.96661010e-01
-2.58083254e-01 6.24703288e-01 5.25637746e-01 7.11844340e-02
-6.71193674e-02 2.94645578e-01 -1.28084257e-01 -4.37954158e-01
-9.84029889e-01 7.86603808e-01 6.31682873e-01 1.15806520e+00
1.55179545e-01 3.02925736e-01 -3.39274824e-01 7.75582731e-01
4.74083304e-01 5.40323079e-01 6.16482139e-01 -1.13115251e+00
5.41630387e-01 2.19356984e-01 -7.04503730e-02 -7.60201216e-01
-2.67652690e-01 -7.98264891e-02 -6.49420381e-01 7.79597238e-02
3.91210914e-01 -6.91057183e-03 -8.32970321e-01 1.67332160e+00
6.93995476e-01 3.34746003e-01 -2.39582941e-01 1.08599472e+00
1.24630475e+00 6.66811466e-01 3.14387321e-01 -1.42604345e-02
1.39383590e+00 -1.21842897e+00 -8.57965767e-01 5.30551486e-02
2.02801526e-01 -9.66895998e-01 1.38178861e+00 3.22627455e-01
-1.07723475e+00 -7.40640342e-01 -9.53917265e-01 -2.47784600e-01
3.38433444e-01 1.83984205e-01 5.72097003e-01 4.03867841e-01
-6.98474288e-01 4.62035865e-01 -1.09198129e+00 -2.35566705e-01
1.01289362e-01 1.81087345e-01 -1.19384795e-01 5.27688503e-01
-9.54795122e-01 2.17638433e-01 2.75212616e-01 -2.33249683e-02
-9.59012687e-01 -3.97798747e-01 -1.00610614e+00 -2.54108071e-01
4.00223970e-01 -7.92585373e-01 1.75046968e+00 -8.24546516e-01
-2.03022528e+00 4.86413598e-01 -2.72126675e-01 -4.90543917e-02
7.97659338e-01 -6.11891568e-01 -2.39205007e-02 2.75460213e-01
-1.82788875e-02 1.15359759e+00 1.00673091e+00 -1.39348173e+00
-7.65214741e-01 -3.87745649e-01 -2.17937946e-01 6.62921965e-01
-2.44216565e-02 1.87182635e-01 -7.16932833e-01 -1.20104694e+00
4.84095871e-01 -1.21311450e+00 -8.55999216e-02 -8.53888616e-02
-6.67065799e-01 -3.80242020e-01 6.38102055e-01 -8.07443261e-01
1.36930692e+00 -1.80381739e+00 5.81898868e-01 1.35065332e-01
1.11308500e-01 -2.05002680e-01 -2.03552127e-01 4.58780587e-01
1.60469770e-01 -1.76093966e-01 4.03111354e-02 -6.98296309e-01
3.05957347e-01 2.19204754e-01 -4.97143686e-01 1.31911919e-01
-9.99786705e-02 1.04527020e+00 -9.22674596e-01 -7.67507792e-01
3.25210035e-01 7.50072122e-01 -6.73650324e-01 3.58082861e-01
-4.19360459e-01 9.08261716e-01 -5.32075405e-01 6.91692770e-01
5.51569052e-02 -3.22328396e-02 3.58837187e-01 -2.90950418e-01
1.58700004e-01 4.93558139e-01 -1.35821402e+00 2.24906206e+00
-3.41534376e-01 3.45580220e-01 -4.37549315e-02 -4.65855330e-01
8.22962880e-01 6.67472899e-01 5.40939808e-01 -3.33551526e-01
1.36407122e-01 7.08190873e-02 -1.28318310e-01 -5.20525873e-01
5.21269679e-01 -1.98570251e-01 -3.98309678e-01 4.15406734e-01
6.26273528e-02 -5.75440466e-01 -1.15290284e-01 8.42081904e-02
7.00213015e-01 7.74331093e-01 4.08829272e-01 4.72269326e-01
1.16572939e-01 -1.61997810e-01 3.91509026e-01 4.76667821e-01
2.70623956e-02 1.05895174e+00 2.47374371e-01 -8.85854438e-02
-9.67770278e-01 -1.20151746e+00 5.02305210e-01 1.73426080e+00
1.07594632e-01 -8.24248433e-01 -8.52896631e-01 -4.04970407e-01
-3.91059488e-01 3.67690235e-01 -4.07200724e-01 3.81553620e-01
-8.93670022e-01 -1.16221510e-01 9.90886867e-01 7.72006094e-01
6.44287318e-02 -1.46929777e+00 -6.31707728e-01 2.02050477e-01
-5.83352029e-01 -1.16497540e+00 -8.07071090e-01 -3.81141096e-01
-7.19904304e-01 -7.63745070e-01 -1.03605115e+00 -8.78352761e-01
1.64264619e-01 -4.65985350e-02 9.16032791e-01 3.38518098e-02
-3.65819484e-01 2.22001210e-01 -6.73267305e-01 -2.18838662e-01
-6.42921507e-01 5.01682237e-02 7.41852075e-02 -8.87155384e-02
-1.31641135e-01 -6.81804419e-01 -4.29603487e-01 3.16311330e-01
-5.38387656e-01 5.68094075e-01 4.60031152e-01 7.79990315e-01
6.37649953e-01 -5.81887960e-01 4.00596678e-01 -3.51698488e-01
5.79700232e-01 4.28655930e-02 -8.93285051e-02 -5.76246195e-02
1.30404919e-01 1.13878869e-01 2.03739733e-01 -8.02025020e-01
-1.05269217e+00 6.13448739e-01 -5.40115595e-01 -5.96139550e-01
-4.01518852e-01 -1.09411497e-02 -5.64871967e-01 5.01914501e-01
5.02014995e-01 3.61580014e-01 -2.70298943e-02 -5.93737602e-01
8.01521778e-01 6.20767057e-01 8.97906661e-01 -9.75126565e-01
8.76213729e-01 2.30197802e-01 -1.76900819e-01 -1.05025530e+00
-4.43108290e-01 -4.91001815e-01 -7.63282657e-01 -4.81040686e-01
1.01848960e+00 -9.51273441e-01 -8.53205562e-01 5.21311402e-01
-1.28263330e+00 -5.68992376e-01 -2.40782842e-01 5.09541392e-01
-9.41365957e-01 3.00841659e-01 -6.92280471e-01 -8.31765294e-01
-2.38178864e-01 -1.37792253e+00 1.90838766e+00 6.91207460e-05
-9.52615798e-01 -4.28292662e-01 -1.01040560e-03 6.38800919e-01
-1.88104272e-01 4.95260388e-01 4.95038122e-01 -3.66570503e-01
-4.05897468e-01 7.40943104e-02 1.99351221e-01 -1.32044181e-01
1.91076905e-01 1.03441983e-01 -7.66452134e-01 5.08952290e-02
-6.22263730e-01 -4.94871467e-01 4.11153287e-01 2.37978622e-01
9.77747262e-01 -4.79473203e-01 -1.02956727e-01 5.22820294e-01
5.80652952e-01 8.94981846e-02 4.35743898e-01 -6.17738850e-02
1.19415820e+00 8.25050116e-01 8.40676367e-01 6.64359987e-01
4.22281206e-01 1.24793434e+00 1.74792916e-01 2.44753554e-01
-7.50710368e-01 -8.24907064e-01 5.63284218e-01 1.17235744e+00
-3.95620853e-01 1.24005228e-01 -7.63816953e-01 4.51844811e-01
-1.78559089e+00 -1.11636758e+00 -1.01100085e-02 1.78825188e+00
1.15691507e+00 -1.18962653e-01 7.40751565e-01 2.55854189e-01
7.88097978e-01 2.53279895e-01 -3.04298073e-01 -5.14874840e-03
1.61284506e-01 2.98636794e-01 -1.61485180e-01 6.01924241e-01
-1.02224386e+00 1.29243946e+00 5.55041361e+00 1.04141915e+00
-1.19570553e+00 3.55099253e-02 3.94470152e-03 -4.44862634e-01
-4.57089275e-01 -2.67763495e-01 -7.27567613e-01 1.42925784e-01
4.80562121e-01 7.55073875e-02 2.44249299e-01 7.37496793e-01
3.17034692e-01 6.58194840e-01 -1.19722521e+00 1.29014063e+00
3.67746651e-02 -1.26881516e+00 5.04742086e-01 -1.27137721e-01
6.15469813e-01 -3.53727669e-01 -8.96630362e-02 3.58374357e-01
3.74415070e-01 -1.09690380e+00 1.37526155e+00 2.93269902e-01
1.02964211e+00 -5.64797461e-01 -2.59294976e-02 3.13091457e-01
-1.79000151e+00 1.70535952e-01 3.87724906e-01 2.81883460e-02
5.81207156e-01 -3.71328473e-01 -1.06014907e+00 4.87098277e-01
6.07388496e-01 3.76615375e-01 -2.33626261e-01 6.63720608e-01
-6.10339940e-01 6.46668434e-01 -2.55188733e-01 -1.52239054e-01
6.96751177e-02 2.17627749e-01 8.06382358e-01 1.33183515e+00
3.91914845e-01 2.64588237e-01 4.37544048e-01 4.99009341e-01
2.17186376e-01 3.11886847e-01 -2.77988553e-01 -2.57565290e-01
8.04559946e-01 9.28101063e-01 -7.31678545e-01 -4.33604956e-01
-4.85603996e-02 1.18567455e+00 -1.45827726e-01 2.96437979e-01
-8.88922215e-01 -1.58536077e-01 7.76905298e-01 2.56586820e-01
8.46533701e-02 -3.29546094e-01 -2.89928347e-01 -9.26680267e-01
4.06312980e-02 -1.19425642e+00 2.40016177e-01 -7.84731269e-01
-8.82761419e-01 6.04707718e-01 1.01040170e-01 -1.51609528e+00
-9.21693087e-01 -1.88449547e-01 -3.97031456e-01 4.19004232e-01
-5.67343533e-01 -1.66001964e+00 -4.93041873e-01 6.93450391e-01
1.20754921e+00 -6.15223162e-02 9.66706634e-01 2.17295125e-01
-1.20724134e-01 8.15049291e-01 -6.81171894e-01 3.74809504e-01
5.80929339e-01 -1.10043788e+00 7.08124042e-01 4.95525688e-01
6.29063070e-01 5.83906829e-01 8.56637537e-01 -8.12083960e-01
-1.18183029e+00 -9.48815107e-01 7.01261699e-01 -4.61060286e-01
5.56932867e-01 -4.65865165e-01 -5.99277258e-01 6.35050714e-01
3.38518061e-02 -4.03734744e-01 6.27388060e-01 5.68714784e-03
-5.11603355e-01 3.01185250e-01 -5.65786660e-01 9.71431196e-01
1.54656291e+00 -6.16038918e-01 -9.22935724e-01 1.19870424e-01
1.12255621e+00 -7.86836922e-01 -7.19626546e-01 3.90462905e-01
1.09532940e+00 -6.28994346e-01 1.12799227e+00 -5.75811505e-01
5.10220945e-01 -4.23775077e-01 -2.76438206e-01 -1.03413129e+00
1.80660158e-01 -1.10896754e+00 -2.55322814e-01 1.23404062e+00
1.11138098e-01 3.56409922e-02 8.74445677e-01 4.67253439e-02
-2.33076423e-01 -5.20527124e-01 -7.35239685e-01 -6.65266871e-01
-1.85667053e-01 -6.77460492e-01 8.83009255e-01 7.80956626e-01
1.64605588e-01 4.92512912e-01 -5.88450134e-01 1.88510776e-01
6.58484519e-01 1.84326932e-01 1.26837373e+00 -1.13179350e+00
-6.03194714e-01 -6.03296697e-01 -3.92163992e-01 -1.44571626e+00
2.40419105e-01 -7.88265347e-01 4.65419888e-01 -1.54875851e+00
-4.54972573e-02 -3.22612673e-01 2.80070096e-01 5.44939101e-01
-1.53484255e-01 3.86701465e-01 5.77348411e-01 4.47474808e-01
-6.24137044e-01 7.31685340e-01 1.57532847e+00 -1.01994991e-01
-6.09884799e-01 2.36120090e-01 -2.76809275e-01 8.88809025e-01
5.57985008e-01 -2.84642935e-01 -5.06012797e-01 -2.89714187e-01
-1.34585708e-01 6.28599942e-01 4.20411110e-01 -1.06371772e+00
2.19687596e-01 -1.98709324e-01 1.44081756e-01 -8.50738347e-01
7.64141619e-01 -3.38578403e-01 4.28354800e-01 3.35116953e-01
-6.07124388e-01 -8.10898200e-04 -1.30111262e-01 3.19499880e-01
-4.32117581e-01 2.62771696e-01 2.38275900e-01 -1.65832862e-01
-6.40475571e-01 2.50539005e-01 -2.52488673e-01 4.98167649e-02
6.90298200e-01 -2.90452510e-01 2.20808744e-01 -6.98444426e-01
-1.25268996e+00 6.03827424e-02 1.97406132e-02 9.27946687e-01
6.64697409e-01 -1.69269764e+00 -7.20332026e-01 -6.67771772e-02
2.52271324e-01 2.28338733e-01 3.09029460e-01 5.97834706e-01
-5.33196926e-01 2.15947941e-01 -1.79723665e-01 -7.78279006e-01
-1.69460571e+00 -1.11707762e-01 3.01996619e-02 7.79323280e-02
-8.39882731e-01 1.13270283e+00 3.58966589e-01 -6.26790643e-01
6.69886947e-01 -3.92310202e-01 -1.46473065e-01 2.00590834e-01
6.14339352e-01 3.25669974e-01 -2.90456921e-01 -1.14707887e+00
-1.51651040e-01 6.50613487e-01 4.04679298e-01 -8.41658354e-01
1.13920856e+00 -4.01819460e-02 3.27353925e-01 6.16039813e-01
9.29556072e-01 2.97455966e-01 -1.44126225e+00 6.98475726e-03
-1.65477246e-01 -3.21745425e-01 -5.35107493e-01 -5.40814161e-01
-8.00776184e-01 9.27327275e-01 1.90737724e-01 -2.67167091e-01
8.97130370e-01 2.64831185e-01 1.04288530e+00 2.28337198e-01
3.56536001e-01 -9.53497946e-01 6.44534826e-01 4.88063902e-01
1.38479805e+00 -6.56216204e-01 -2.79808998e-01 -6.59564674e-01
-1.04243934e+00 1.03298187e+00 7.20472872e-01 6.89375624e-02
2.07477555e-01 4.53950435e-01 3.68365407e-01 -1.07582077e-01
-4.42293048e-01 -3.71466100e-01 4.95068043e-01 7.27139354e-01
6.30756259e-01 4.09204841e-01 -2.05067649e-01 8.47620845e-01
-1.04709101e+00 -2.34096527e-01 2.17081055e-01 8.08076382e-01
-2.82814950e-01 -1.27962196e+00 -6.66680932e-01 -2.47645631e-01
-2.16545314e-01 -3.69955190e-02 -6.20098829e-01 7.11454213e-01
-4.22999635e-02 7.90164053e-01 -8.47364292e-02 -7.11292982e-01
4.14443880e-01 2.78230548e-01 7.64624774e-01 -7.74151266e-01
-6.05626881e-01 7.32687831e-01 8.58363584e-02 -9.00437534e-01
-3.32881451e-01 -9.13684070e-01 -1.83554840e+00 7.86891654e-02
7.89433792e-02 -1.25253782e-01 5.53075016e-01 9.11332548e-01
7.21087400e-03 9.05321777e-01 2.16555625e-01 -1.61283123e+00
-5.37919283e-01 -1.04647422e+00 -3.60426128e-01 6.05327666e-01
2.38500372e-01 -6.85477078e-01 -5.92235141e-02 4.42888975e-01] | [5.659525394439697, -0.14955255389213562] |
e5ed403d-8c4e-4efb-bbb6-001cb9ad2c5a | graph-neural-networks-for-breast-cancer-data | 2211.15561 | null | https://arxiv.org/abs/2211.15561v1 | https://arxiv.org/pdf/2211.15561v1.pdf | Graph Neural Networks for Breast Cancer Data Integration | International initiatives such as METABRIC (Molecular Taxonomy of Breast Cancer International Consortium) have collected several multigenomic and clinical data sets to identify the undergoing molecular processes taking place throughout the evolution of various cancers. Numerous Machine Learning and statistical models have been designed and trained to analyze these types of data independently, however, the integration of such differently shaped and sourced information streams has not been extensively studied. To better integrate these data sets and generate meaningful representations that can ultimately be leveraged for cancer detection tasks could lead to giving well-suited treatments to patients. Hence, we propose a novel learning pipeline comprising three steps - the integration of cancer data modalities as graphs, followed by the application of Graph Neural Networks in an unsupervised setting to generate lower-dimensional embeddings from the combined data, and finally feeding the new representations on a cancer sub-type classification model for evaluation. The graph construction algorithms are described in-depth as METABRIC does not store relationships between the patient modalities, with a discussion of their influence over the quality of the generated embeddings. We also present the models used to generate the lower-latent space representations: Graph Neural Networks, Variational Graph Autoencoders and Deep Graph Infomax. In parallel, the pipeline is tested on a synthetic dataset to demonstrate that the characteristics of the underlying data, such as homophily levels, greatly influence the performance of the pipeline, which ranges between 51\% to 98\% accuracy on artificial data, and 13\% and 80\% on METABRIC. This project has the potential to improve cancer data understanding and encourages the transition of regular data sets to graph-shaped data. | ['Teodora Reu'] | 2022-11-28 | null | null | null | null | ['data-integration'] | ['knowledge-base'] | [ 4.31760103e-01 3.42897445e-01 -2.42499486e-01 -1.86203331e-01
-3.64848316e-01 -2.90870905e-01 7.03860521e-01 8.86251748e-01
-3.00815970e-01 3.38664532e-01 6.14438653e-01 -5.28286695e-01
-2.83910573e-01 -1.04056442e+00 -4.67879206e-01 -8.93144608e-01
-2.24186465e-01 6.01486742e-01 -1.73881233e-01 -5.91131039e-02
-1.41138107e-01 4.52131659e-01 -1.24638259e+00 2.60376066e-01
4.55308884e-01 6.42688036e-01 -1.07306868e-01 9.32443380e-01
-1.69707611e-01 2.79496759e-01 -3.39059472e-01 -2.27435857e-01
4.02652025e-02 -4.60578918e-01 -6.11388981e-01 7.60146976e-02
-6.14057109e-02 1.48397073e-01 -3.75641048e-01 7.46788442e-01
5.23692429e-01 -1.98512495e-01 8.26141059e-01 -9.65183616e-01
-5.69953740e-01 5.89673817e-01 -1.52329132e-01 -8.69600773e-02
4.60428558e-02 2.78855175e-01 9.73142445e-01 -4.40621614e-01
1.09377170e+00 1.11047101e+00 6.62555754e-01 5.01883209e-01
-1.39338660e+00 -1.27927080e-01 -2.83058405e-01 -2.33874857e-01
-1.11180592e+00 -2.36630946e-01 5.89312732e-01 -5.56325555e-01
9.20159817e-01 7.84268677e-02 8.42145383e-01 1.28599060e+00
4.76066768e-01 3.29539508e-01 7.56652176e-01 -4.00646925e-01
3.56535882e-01 1.67609751e-01 2.50904471e-01 9.14595723e-01
4.14540321e-01 -1.28373340e-01 -3.59405279e-01 -4.01592493e-01
4.19359803e-01 3.04749906e-01 -1.74064770e-01 -3.95063937e-01
-1.10817790e+00 1.13326740e+00 6.54650033e-01 5.59099197e-01
-4.05898780e-01 6.90487400e-02 5.69737375e-01 1.52346998e-01
4.79987681e-01 4.93884861e-01 -2.53096074e-01 1.63066417e-01
-6.56243742e-01 -1.60100654e-01 7.88387835e-01 4.18990135e-01
5.02832711e-01 -1.42581418e-01 -1.48563029e-03 7.08052337e-01
4.93891388e-01 -9.24538746e-02 7.97773838e-01 -3.99614125e-01
3.84686068e-02 1.15560234e+00 -5.59088230e-01 -9.81656492e-01
-8.03226829e-01 -5.75183272e-01 -8.93867016e-01 -4.82904576e-02
4.54223305e-01 -6.36085719e-02 -9.88611698e-01 1.66747940e+00
4.31344151e-01 8.59108195e-02 2.46822476e-01 5.35986066e-01
9.66295302e-01 2.98724502e-01 2.21061215e-01 1.33914500e-01
1.42278862e+00 -4.47427362e-01 -5.29654443e-01 1.83344111e-01
1.18184888e+00 -3.51023555e-01 6.96623504e-01 6.74314946e-02
-7.60088742e-01 -2.75554925e-01 -1.04862356e+00 -1.51581094e-01
-8.97872925e-01 -2.28458211e-01 7.75107861e-01 6.30678236e-01
-1.18146563e+00 7.72712588e-01 -1.12451541e+00 -7.15990543e-01
6.84511900e-01 3.77604455e-01 -6.10585868e-01 -2.72751749e-01
-1.05989873e+00 7.16638386e-01 4.69223678e-01 -1.58440843e-01
-7.54143357e-01 -7.97705948e-01 -1.04317844e+00 1.48985330e-02
-8.33878219e-02 -9.58153427e-01 3.65933776e-01 -6.93092823e-01
-1.10950875e+00 9.60869133e-01 9.31804813e-03 -4.01527256e-01
1.69501275e-01 3.72145027e-01 -2.77435035e-01 -4.98272032e-02
-2.10096806e-01 6.16605103e-01 2.27384537e-01 -7.69701600e-01
-2.36416906e-01 -6.95145011e-01 -4.13145185e-01 9.04087052e-02
-6.46175444e-01 -3.38383883e-01 -5.13638973e-01 -3.19761544e-01
-6.78414777e-02 -9.84470487e-01 -3.26733410e-01 1.83659762e-01
-3.67495924e-01 -6.31541759e-02 5.46602309e-01 -5.47847509e-01
8.98212969e-01 -2.10375237e+00 5.25485218e-01 2.88551480e-01
4.45764035e-01 8.64652917e-02 -2.87896752e-01 7.80230939e-01
-2.00896457e-01 4.03245151e-01 -3.64242733e-01 -4.17262942e-01
-1.13683559e-01 2.38557503e-01 3.81465971e-01 6.48145318e-01
3.50328833e-01 9.80416179e-01 -9.21936393e-01 -1.21459089e-01
1.21952564e-01 8.80101919e-01 -5.81051886e-01 2.07989216e-01
-3.29694420e-01 1.20055743e-01 -2.15030417e-01 5.98480284e-01
2.59504765e-01 -4.35356319e-01 6.05995536e-01 -1.97396412e-01
3.15311372e-01 -4.59110774e-02 -5.44267893e-01 1.63020515e+00
-1.62856653e-01 6.80513918e-01 5.38137630e-02 -9.59393561e-01
8.84179831e-01 1.43479437e-01 7.74484873e-01 -3.02890122e-01
3.26590866e-01 -5.65219112e-02 2.55381912e-01 -4.21078563e-01
1.97322190e-01 -1.82857111e-01 1.02890275e-01 3.03517014e-01
2.36222818e-01 4.77205925e-02 1.68378711e-01 4.04763877e-01
1.48548949e+00 -2.03363284e-01 5.24141304e-02 -1.86327577e-01
2.00656414e-01 1.79694816e-01 4.46898103e-01 2.20872566e-01
-1.60963297e-01 5.79690039e-01 9.08557475e-01 -2.88984179e-01
-9.00151968e-01 -8.78452718e-01 -2.70706236e-01 7.13478744e-01
-1.98329881e-01 -5.96069217e-01 -4.23820883e-01 -6.72771096e-01
2.24219516e-01 6.26395643e-01 -1.11884844e+00 -3.91800970e-01
6.61671907e-02 -1.36093152e+00 6.51467144e-01 2.86248207e-01
-3.19092780e-01 -8.28121543e-01 -2.47144341e-01 1.73195913e-01
3.91621470e-01 -9.37723815e-01 7.93142170e-02 4.64496881e-01
-8.76578510e-01 -1.50151491e+00 -3.39465708e-01 -5.77948153e-01
8.48061442e-01 -1.09813951e-01 9.70046222e-01 2.18380302e-01
-7.69915760e-01 2.28788584e-01 -2.82480210e-01 -3.61949503e-01
-7.85022318e-01 -2.71588806e-02 -1.28454819e-01 7.19676539e-02
3.44467342e-01 -3.96779180e-01 -5.37187696e-01 -7.98706859e-02
-1.36328745e+00 7.07524195e-02 5.90829074e-01 1.07399809e+00
6.72455728e-01 5.45252524e-02 3.79238725e-01 -1.22125995e+00
8.02592456e-01 -8.75826776e-01 -3.24664116e-01 1.58016071e-01
-6.40062988e-01 3.01068753e-01 6.70694590e-01 -1.18831940e-01
-4.77154046e-01 -1.26750730e-02 -2.86327124e-01 -3.83009106e-01
-1.10910714e-01 8.98546457e-01 -1.33520074e-03 2.10091159e-01
6.91169083e-01 2.59116106e-02 5.96108139e-01 -2.49887496e-01
5.30210614e-01 6.25971198e-01 1.30560800e-01 -3.36997360e-02
4.12877411e-01 5.45791030e-01 3.37989479e-01 -9.04448509e-01
-2.62367755e-01 -3.77830446e-01 -6.02410972e-01 1.14388891e-01
1.17306554e+00 -7.09447443e-01 -3.53930533e-01 3.05126131e-01
-4.84135181e-01 -5.09162545e-01 -2.41942227e-01 5.11439741e-01
-2.23385558e-01 6.08717240e-02 -7.73109078e-01 -3.43183875e-01
-4.10552919e-01 -1.40615737e+00 1.00673115e+00 7.79789612e-02
-2.80717820e-01 -1.56412840e+00 5.15126288e-01 2.92984068e-01
3.48242611e-01 7.90031910e-01 1.39331794e+00 -1.02841926e+00
-2.09587947e-01 -3.38127851e-01 -1.43771812e-01 -7.15424716e-02
3.64956170e-01 2.94739723e-01 -9.13995922e-01 -4.18696880e-01
-5.07727265e-01 -2.42994174e-01 9.26961422e-01 3.82302612e-01
9.05380547e-01 -3.09008430e-03 -8.09819043e-01 8.58456194e-01
1.45612574e+00 -8.51584673e-02 4.29453969e-01 -9.61000938e-03
7.64138103e-01 6.90657020e-01 -1.26612008e-01 5.56043461e-02
3.23884308e-01 4.16731894e-01 6.42674446e-01 -2.62966484e-01
-1.95870563e-01 -2.73426771e-01 2.34996423e-01 7.72185504e-01
3.63620698e-01 -4.36716139e-01 -1.03759027e+00 5.28436899e-01
-1.49249208e+00 -7.13284671e-01 -1.05109341e-01 1.98403573e+00
6.05157077e-01 1.61647022e-01 3.19793485e-02 -8.24484378e-02
4.46712345e-01 2.84503177e-02 -5.71597517e-01 -5.78600645e-01
-1.32474974e-01 1.93593532e-01 3.92037630e-01 3.96619737e-01
-6.86582983e-01 7.17275083e-01 6.35139847e+00 2.35609666e-01
-1.32038045e+00 -1.32619217e-01 8.29021394e-01 -4.99507301e-02
-4.87500519e-01 -9.52868387e-02 -4.62469786e-01 3.02135736e-01
1.40567338e+00 -4.58159819e-02 2.69315571e-01 5.10349989e-01
2.57599682e-01 1.05668679e-01 -1.13193357e+00 6.92592978e-01
-1.38467327e-02 -1.59579897e+00 -4.99721207e-02 4.12767470e-01
4.45839763e-01 5.81642687e-01 -2.10470513e-01 1.69680342e-01
6.93062782e-01 -1.22558916e+00 -2.01913670e-01 5.63745975e-01
6.47474706e-01 -6.03838205e-01 7.93466091e-01 2.23997712e-01
-8.07590425e-01 4.91909683e-02 -2.83746600e-01 3.22929680e-01
-1.36035472e-01 8.08670878e-01 -1.57238805e+00 8.22860539e-01
2.87522107e-01 8.36573005e-01 -8.58442247e-01 6.70227289e-01
1.51946813e-01 5.79360843e-01 -2.31254265e-01 -1.66305944e-01
2.75709834e-02 -1.40761063e-01 2.81811386e-01 1.17104447e+00
2.28201956e-01 -3.14950496e-01 7.67930821e-02 8.92325103e-01
-1.50528342e-01 1.34096757e-01 -6.34958506e-01 -6.44441605e-01
1.97910473e-01 1.47775042e+00 -8.48886013e-01 -9.51454118e-02
-4.38652992e-01 6.79141104e-01 4.54756051e-01 1.92118675e-01
-4.80280370e-01 -1.62370875e-01 7.35209167e-01 2.75929481e-01
-4.87059392e-02 -4.81534563e-02 -1.15052938e-01 -1.03899729e+00
-6.04239225e-01 -9.45434451e-01 7.62707651e-01 -4.30855602e-01
-1.35068130e+00 5.38433850e-01 -2.93714136e-01 -6.05033159e-01
-1.73874751e-01 -8.37446690e-01 -6.68226838e-01 9.22243357e-01
-1.21044123e+00 -1.13220978e+00 -4.23786968e-01 2.59936303e-01
1.40474781e-01 -1.87294275e-01 1.15438199e+00 -7.10080639e-02
-9.46875691e-01 4.28069025e-01 3.32068354e-01 3.10357928e-01
5.08118629e-01 -1.30054951e+00 2.20427766e-01 4.50784296e-01
9.04445201e-02 5.72782099e-01 4.31468546e-01 -7.16223061e-01
-1.75209522e+00 -1.34643757e+00 4.97913301e-01 -4.51652139e-01
6.97053313e-01 -4.58637297e-01 -8.86064708e-01 5.18857419e-01
2.01658413e-01 1.04655817e-01 1.26646364e+00 2.73420066e-01
-1.86627358e-01 5.97662851e-02 -1.08736467e+00 5.53645015e-01
7.95708239e-01 -4.82521802e-01 -1.50189638e-01 4.87720728e-01
5.14881372e-01 -5.61913438e-02 -1.35704291e+00 2.70122707e-01
4.13339108e-01 -8.69261086e-01 6.97683752e-01 -8.10489058e-01
5.03538370e-01 -1.00914851e-01 -5.87452315e-02 -1.51389039e+00
-4.24069196e-01 -2.36310840e-01 1.78472757e-01 9.11041081e-01
6.41314030e-01 -7.78825045e-01 9.25257623e-01 4.93644238e-01
-8.90046805e-02 -1.14678347e+00 -6.59510851e-01 -1.35519549e-01
1.83592990e-01 -3.06725830e-01 4.66270119e-01 9.84921038e-01
2.14882985e-01 2.13309467e-01 2.39904523e-01 1.02964211e-02
3.90607566e-01 -1.49272099e-01 7.80511022e-01 -1.20144820e+00
-4.10913020e-01 -6.42629385e-01 -8.24245214e-01 -5.70877939e-02
2.21089255e-02 -1.56726074e+00 -4.20517236e-01 -1.62324202e+00
1.89444438e-01 -3.79958332e-01 -4.43994582e-01 5.12773216e-01
-2.62391210e-01 -2.03940496e-02 -2.56566733e-01 -1.52091918e-04
-1.21603668e-01 4.88389373e-01 8.88778508e-01 -2.41216630e-01
-3.22348624e-01 -4.85567063e-01 -8.20429862e-01 3.43308121e-01
6.62803531e-01 -3.32140297e-01 -4.03218716e-01 -2.02071369e-01
1.91671208e-01 7.99502209e-02 1.41529977e-01 -8.96364212e-01
1.79013815e-02 5.77265434e-02 6.31359637e-01 -1.95775017e-01
2.63407260e-01 -6.32228494e-01 6.19605541e-01 7.27093816e-01
-5.36328256e-01 -3.52882035e-02 1.84333473e-01 7.05245316e-01
-3.77921164e-02 1.85227599e-02 5.86686730e-01 -8.66244063e-02
-1.02591924e-01 5.18096924e-01 -2.70298809e-01 -2.22722441e-01
1.11412108e+00 -1.33527517e-01 -3.23062807e-01 -2.36725748e-01
-8.55348110e-01 2.02261791e-01 6.81089640e-01 4.62093979e-01
3.36791486e-01 -1.06787801e+00 -8.84364486e-01 4.46435750e-01
2.82369912e-01 -1.09148107e-01 1.65068209e-01 6.90998197e-01
-6.59065127e-01 2.48375699e-01 -1.73376903e-01 -6.60621941e-01
-1.33609700e+00 6.29240096e-01 3.88179451e-01 -4.74248171e-01
-4.57856447e-01 6.25670135e-01 -4.25684750e-02 -5.42918503e-01
1.07604195e-03 -2.06017107e-01 -3.52681488e-01 3.62158120e-01
1.68984875e-01 8.83615613e-02 1.80595994e-01 -4.93221521e-01
-3.30306888e-01 1.49919182e-01 -1.02300927e-01 9.32850018e-02
1.67539847e+00 3.26224148e-01 -1.40587270e-01 3.14762264e-01
1.52633643e+00 -1.35626897e-01 -8.81973445e-01 6.94468096e-02
-4.90340777e-03 7.97628984e-03 1.96079940e-01 -5.31680524e-01
-1.09898102e+00 8.76600862e-01 5.86572409e-01 2.91320026e-01
9.58505511e-01 3.56482565e-02 3.55687022e-01 8.18452388e-02
-1.91911057e-01 -6.69628263e-01 -1.10931829e-01 1.07793905e-01
5.42775273e-01 -1.01309276e+00 8.17888230e-02 -2.85588413e-01
-2.56686985e-01 1.24232578e+00 2.09581912e-01 3.37688625e-02
7.04885125e-01 2.10452139e-01 1.31637707e-01 -5.32949209e-01
-1.03854597e+00 -8.20643082e-02 1.25298753e-01 5.69233954e-01
6.84800625e-01 1.45679474e-01 -1.58553824e-01 3.88061553e-01
-3.93705666e-02 5.01094647e-02 5.27564824e-01 6.81062102e-01
-2.53246892e-02 -1.33934855e+00 -1.16038732e-01 7.54711509e-01
-4.44024026e-01 1.58813000e-02 -6.74261272e-01 7.28537858e-01
4.40503359e-02 6.40872180e-01 8.36902037e-02 -4.75018442e-01
-2.02345964e-03 2.49070480e-01 1.69874996e-01 -7.04179287e-01
-5.79255402e-01 3.27238929e-03 1.17101200e-01 -3.35610926e-01
-1.44633651e-01 -5.61637163e-01 -1.20874715e+00 -1.42234936e-01
-3.11718941e-01 3.43196541e-02 8.89387906e-01 6.88409328e-01
7.62374759e-01 9.10982370e-01 3.33446831e-01 -6.14354610e-01
-1.99676275e-01 -8.91400278e-01 -3.79531115e-01 7.12186456e-01
-4.07985831e-03 -3.51819158e-01 -3.84923697e-01 -1.30065352e-01] | [5.980751991271973, 5.696719169616699] |
28da45ad-0eb4-4296-935e-35c21dbaf9e2 | source-code-summarization-with-structural | 2202.06521 | null | https://arxiv.org/abs/2202.06521v1 | https://arxiv.org/pdf/2202.06521v1.pdf | Source Code Summarization with Structural Relative Position Guided Transformer | Source code summarization aims at generating concise and clear natural language descriptions for programming languages. Well-written code summaries are beneficial for programmers to participate in the software development and maintenance process. To learn the semantic representations of source code, recent efforts focus on incorporating the syntax structure of code into neural networks such as Transformer. Such Transformer-based approaches can better capture the long-range dependencies than other neural networks including Recurrent Neural Networks (RNNs), however, most of them do not consider the structural relative correlations between tokens, e.g., relative positions in Abstract Syntax Trees (ASTs), which is beneficial for code semantics learning. To model the structural dependency, we propose a Structural Relative Position guided Transformer, named SCRIPT. SCRIPT first obtains the structural relative positions between tokens via parsing the ASTs of source code, and then passes them into two types of Transformer encoders. One Transformer directly adjusts the input according to the structural relative distance; and the other Transformer encodes the structural relative positions during computing the self-attention scores. Finally, we stack these two types of Transformer encoders to learn representations of source code. Experimental results show that the proposed SCRIPT outperforms the state-of-the-art methods by at least 1.6%, 1.4% and 2.8% with respect to BLEU, ROUGE-L and METEOR on benchmark datasets, respectively. We further show that how the proposed SCRIPT captures the structural relative dependencies. | ['Zenglin Xu', 'Yun Peng', 'Wenchao Gu', 'Yasheng Wang', 'Cuiyun Gao', 'Zi Gong'] | 2022-02-14 | null | null | null | null | ['code-summarization'] | ['computer-code'] | [ 1.84822261e-01 2.09806971e-02 -3.98855746e-01 -5.97720861e-01
-5.67265868e-01 -3.45576048e-01 2.12097794e-01 4.18592721e-01
-1.19725995e-01 8.19451585e-02 6.33953154e-01 -4.15459305e-01
1.46102220e-01 -8.50840330e-01 -8.37493658e-01 -4.36426401e-01
1.05718002e-01 6.99473768e-02 2.60285467e-01 -1.52593344e-01
5.38429260e-01 -4.75558862e-02 -1.51000619e+00 4.78844255e-01
1.10350001e+00 7.48436272e-01 5.58265686e-01 4.20541197e-01
-9.78839397e-01 1.43055367e+00 -5.46093702e-01 -3.98981541e-01
-4.45936382e-01 -5.33353806e-01 -7.34489381e-01 -3.37804198e-01
-1.13742173e-01 -1.11548617e-01 -2.50254124e-01 1.30645227e+00
8.36931095e-02 1.49769913e-02 4.07864153e-01 -8.55083406e-01
-9.36228514e-01 1.35953891e+00 -6.75952256e-01 2.08534390e-01
1.83433160e-01 -2.38750979e-01 1.45264828e+00 -8.73558998e-01
2.52869010e-01 1.18528044e+00 7.07594872e-01 4.68862444e-01
-7.75957584e-01 -6.42571032e-01 3.51515353e-01 1.35676324e-01
-1.05353630e+00 -2.68680185e-01 1.00692618e+00 -5.95834613e-01
1.27602780e+00 1.84366145e-05 2.05977648e-01 7.95714796e-01
2.14240134e-01 9.55223620e-01 3.71528953e-01 -3.32532793e-01
7.18511641e-03 -1.16857029e-02 5.94947159e-01 1.04981852e+00
1.69817477e-01 -4.32697833e-01 -2.25966573e-01 -1.08143844e-01
5.71051717e-01 3.85792673e-01 -2.13219941e-01 -2.04217196e-01
-1.20179617e+00 8.84574056e-01 6.85325682e-01 4.13874209e-01
-2.62034684e-01 3.88990998e-01 8.87147665e-01 2.34418828e-02
3.03024918e-01 2.87060887e-01 -4.69379574e-01 -3.24716032e-01
-6.91628575e-01 -6.63315207e-02 4.91663277e-01 1.19749033e+00
9.11273718e-01 1.60045892e-01 -2.90470600e-01 1.15536964e+00
5.25867999e-01 2.83182263e-01 1.01089168e+00 -5.44910371e-01
1.00376427e+00 1.18407226e+00 -3.21268290e-01 -1.14301193e+00
-1.99991897e-01 -6.29300594e-01 -9.13135052e-01 -3.85479599e-01
-2.19327271e-01 1.07536718e-01 -8.00160885e-01 1.69815397e+00
-1.89729631e-01 -2.02099290e-02 2.28722721e-01 5.15457809e-01
1.20493948e+00 1.06502354e+00 3.63692567e-02 -3.82679962e-02
1.23864210e+00 -1.31596756e+00 -5.95726371e-01 -5.32392740e-01
8.89453948e-01 -6.97065592e-01 1.17187357e+00 -1.33261830e-01
-1.00387967e+00 -5.31621516e-01 -1.03301501e+00 -1.04499653e-01
-1.36381522e-01 5.26897371e-01 4.80559051e-01 3.98727089e-01
-8.01030695e-01 6.56861126e-01 -1.01140499e+00 -1.17102653e-01
4.00089443e-01 1.00724921e-01 4.07711528e-02 5.38334176e-02
-1.02103996e+00 4.81394410e-01 5.71155667e-01 1.39338195e-01
-8.55862141e-01 -6.70737267e-01 -1.28200400e+00 5.51700532e-01
2.60567904e-01 -3.45096976e-01 1.47592604e+00 -1.05630851e+00
-1.40338445e+00 6.22300565e-01 -4.91951197e-01 -3.61122817e-01
-6.68437406e-02 -1.35112718e-01 -2.02687338e-01 -3.45563143e-01
1.95017785e-01 2.03309879e-01 4.19508517e-01 -9.99650538e-01
-6.20082378e-01 -1.33915633e-01 2.10874602e-01 1.27151925e-02
-4.88992870e-01 3.37553799e-01 -5.67287683e-01 -7.11994350e-01
5.27418591e-02 -6.90446436e-01 -2.54632324e-01 -4.20990616e-01
-6.26586080e-01 -4.48184490e-01 5.77349842e-01 -7.35100925e-01
1.63355076e+00 -2.20995998e+00 3.44242781e-01 -7.21269548e-02
2.64430463e-01 4.49815363e-01 -2.48321906e-01 3.56021076e-01
-1.63593546e-01 2.60599762e-01 -5.23371518e-01 -2.79908508e-01
5.49199097e-02 1.98822156e-01 -5.43797553e-01 2.70680282e-02
1.77521572e-01 8.49299610e-01 -9.63165045e-01 -3.47110629e-01
-4.34065796e-02 3.15900832e-01 -7.26687074e-01 5.82433641e-01
-3.88630748e-01 -1.74162373e-01 -6.97156489e-01 2.97277838e-01
3.94999057e-01 -4.27200705e-01 9.20306966e-02 -4.38274145e-02
-1.96298420e-01 1.01947248e+00 -6.04504287e-01 1.91621780e+00
-8.55783045e-01 4.76182818e-01 -4.73308682e-01 -1.17208552e+00
1.30882227e+00 1.60710350e-01 -3.75407026e-03 -6.28111184e-01
-5.14115542e-02 2.26704955e-01 2.16333736e-02 -5.76369584e-01
4.32531327e-01 1.60956621e-01 -3.38987321e-01 5.42131484e-01
-9.14227888e-02 3.72306883e-01 1.79856390e-01 3.90483022e-01
1.08283770e+00 2.29629889e-01 3.42670083e-01 -1.72180831e-01
8.09964657e-01 -3.11475247e-01 7.95939922e-01 4.37059850e-01
2.37253785e-01 4.80009466e-01 9.44201112e-01 -5.77819467e-01
-9.20859218e-01 -8.31243455e-01 3.80952954e-01 1.29898167e+00
6.09944612e-02 -8.59364212e-01 -8.78470004e-01 -8.02405477e-01
-3.28873664e-01 7.92861402e-01 -7.51734614e-01 -4.67885464e-01
-8.10343921e-01 -6.61976576e-01 6.92424834e-01 8.62044215e-01
6.90454543e-01 -1.23180079e+00 -7.38308072e-01 3.23067814e-01
-3.61392856e-01 -7.33737350e-01 -6.55151904e-01 2.56040543e-01
-9.51462090e-01 -8.87384295e-01 -5.15945435e-01 -8.14879715e-01
7.38348484e-01 1.59211203e-01 1.14906430e+00 4.19035077e-01
2.29000568e-01 -2.65781671e-01 -5.76448739e-01 -7.30431154e-02
-6.28050566e-01 4.70676273e-01 -5.13499618e-01 -1.58087701e-01
2.93180048e-01 -7.30401874e-01 -2.83040076e-01 5.73657565e-02
-8.74466181e-01 3.90872896e-01 9.16346192e-01 8.73239219e-01
4.63611543e-01 3.21934931e-02 2.64305323e-01 -1.22644067e+00
5.56833923e-01 -6.95183039e-01 -5.23973942e-01 3.75026196e-01
-5.39510429e-01 6.61652863e-01 1.14395916e+00 -7.44920000e-02
-1.37242651e+00 -3.98928039e-02 -4.01737809e-01 -2.64685631e-01
1.12959102e-01 1.12548101e+00 -1.96399599e-01 3.97005320e-01
4.75443006e-01 6.73739195e-01 -3.46758127e-01 -6.01939142e-01
1.56916469e-01 7.72776067e-01 4.53565627e-01 -7.47430742e-01
6.05986118e-01 -6.99240118e-02 -5.05054593e-01 -2.62281090e-01
-9.68837976e-01 -2.66007423e-01 -4.76547033e-01 3.45172167e-01
7.20196605e-01 -7.17588603e-01 -2.10436389e-01 4.62207794e-01
-1.57037067e+00 -9.68883559e-02 1.44280136e-01 7.92033151e-02
-4.05260980e-01 4.28163677e-01 -7.93350220e-01 -4.60479319e-01
-4.92967993e-01 -1.46131504e+00 9.23046887e-01 4.06779051e-01
-9.47811827e-02 -9.12876725e-01 4.77499254e-02 1.23406008e-01
5.35683632e-01 1.31462678e-01 1.52513373e+00 -6.92738235e-01
-5.75257778e-01 1.11234471e-01 -3.89107585e-01 3.06951493e-01
3.66308957e-01 1.73510358e-01 -7.79031575e-01 -6.52236165e-03
-7.32333139e-02 -9.64341983e-02 9.47324574e-01 4.46180329e-02
1.46286738e+00 -6.22016728e-01 -3.15543175e-01 7.43866503e-01
1.39718962e+00 4.43483442e-01 6.55096591e-01 2.34473243e-01
1.06159377e+00 5.65939367e-01 2.88566232e-01 4.09679830e-01
5.85885882e-01 5.04217625e-01 5.22902369e-01 2.22182527e-01
-2.08348129e-03 -6.13369226e-01 6.85969114e-01 1.46727550e+00
1.26369059e-01 5.15355729e-02 -1.14581084e+00 6.51920974e-01
-1.95968211e+00 -8.31743121e-01 -1.69081628e-01 2.03038478e+00
1.03492188e+00 2.80593812e-01 -2.36234605e-01 -8.00623521e-02
8.68335783e-01 4.01179731e-01 -5.41041315e-01 -6.03674293e-01
4.18877870e-01 -6.40152693e-02 3.07372898e-01 1.95067152e-01
-7.30170488e-01 8.59026849e-01 4.76294041e+00 7.56042182e-01
-9.76914108e-01 3.43578272e-02 4.27785099e-01 2.62386322e-01
-6.85546696e-01 1.85741678e-01 -7.11346030e-01 7.21473694e-01
1.14183331e+00 -4.95912373e-01 4.47796613e-01 1.22348034e+00
-3.68526913e-02 3.56005043e-01 -1.17042053e+00 7.21832037e-01
5.40617518e-02 -1.45941758e+00 2.73649901e-01 -4.88279879e-01
5.81355870e-01 2.32820958e-01 -2.26724699e-01 5.68403184e-01
5.12724817e-01 -9.08368886e-01 8.38069320e-01 2.60723114e-01
6.14408851e-01 -9.63512778e-01 9.15924788e-01 3.71582001e-01
-1.64703584e+00 -1.72809318e-01 -5.98830342e-01 2.31556036e-02
-2.36093521e-01 6.42136514e-01 -5.15700996e-01 5.82136571e-01
7.39630163e-01 1.27071083e+00 -7.76088893e-01 7.13262737e-01
-4.97023702e-01 7.41347134e-01 2.19128549e-01 -3.84373665e-01
4.38533425e-01 -1.36442054e-02 2.17719868e-01 1.46727884e+00
3.73715848e-01 -1.92301750e-01 -6.84831068e-02 1.36450326e+00
-3.67395997e-01 8.10009018e-02 -4.40466523e-01 -2.83883393e-01
6.45682573e-01 1.03363752e+00 -5.92759669e-01 -4.74990577e-01
-6.46841645e-01 7.84774423e-01 5.42303383e-01 2.39935473e-01
-8.40307832e-01 -1.06564963e+00 6.57412529e-01 -2.87774414e-01
4.13283467e-01 4.85291667e-02 -1.98751181e-01 -1.28759933e+00
4.05456305e-01 -8.62991452e-01 3.45198005e-01 -7.69054532e-01
-9.25732315e-01 9.14377689e-01 -1.31780684e-01 -1.06164753e+00
-3.18663925e-01 -3.38258862e-01 -9.26316142e-01 9.12789762e-01
-1.60140908e+00 -9.03003335e-01 -3.72851551e-01 1.77680686e-01
9.15095687e-01 -2.27430761e-01 7.80447125e-01 2.10195705e-01
-9.28419948e-01 6.00051165e-01 1.81625143e-01 5.55848241e-01
1.89158008e-01 -1.30180311e+00 9.40729856e-01 1.05142426e+00
1.28678754e-01 1.20498610e+00 4.26429540e-01 -4.53894526e-01
-1.36253047e+00 -1.44378269e+00 1.04651248e+00 -1.19734414e-01
5.76602340e-01 -4.02678967e-01 -1.26352143e+00 9.69850302e-01
2.00167626e-01 5.67848198e-02 5.46047270e-01 1.17870212e-01
-8.54876220e-01 -1.64195344e-01 -5.86000919e-01 3.85530233e-01
1.03676641e+00 -6.84527338e-01 -9.82276797e-01 1.00002013e-01
1.10618198e+00 -5.38529694e-01 -5.50342917e-01 1.57458022e-01
2.74499476e-01 -1.19315970e+00 6.15254343e-01 -4.99079674e-01
1.11517859e+00 -3.95776957e-01 -5.05918860e-02 -1.41379869e+00
-4.49375212e-01 -3.36945236e-01 1.67556927e-01 1.71083462e+00
4.69292462e-01 -4.77553695e-01 5.40640295e-01 2.09122226e-01
-5.09740472e-01 -7.72666276e-01 -4.56729263e-01 -5.22862673e-01
7.38497600e-02 -3.26726407e-01 8.26325178e-01 8.89417648e-01
1.09013334e-01 4.60503042e-01 -9.41934884e-02 7.46772662e-02
3.40100080e-01 4.78240550e-01 4.49983060e-01 -1.14628410e+00
-3.91291618e-01 -7.36847579e-01 -2.88720548e-01 -1.16610789e+00
6.38370752e-01 -1.17575133e+00 2.46019915e-01 -1.67138267e+00
5.45596302e-01 -4.86517280e-01 -5.16624331e-01 7.04457939e-01
-3.55745822e-01 -5.06237209e-01 -2.03094035e-01 3.18889260e-01
-6.09618604e-01 6.96424842e-01 5.81247509e-01 -3.78199011e-01
3.38259782e-03 -9.85688046e-02 -8.87967765e-01 8.22816253e-01
7.13258207e-01 -8.55899811e-01 -5.96541584e-01 -9.69565630e-01
4.79559064e-01 9.49189439e-02 6.76287934e-02 -8.18011701e-01
3.11254591e-01 -2.34162793e-01 -7.24643767e-02 -4.51355815e-01
-4.42905068e-01 -5.93487442e-01 -3.12364567e-02 4.78831768e-01
-7.89392889e-01 3.61409545e-01 2.09017798e-01 3.76687199e-01
-4.68864113e-01 -6.24370933e-01 6.75067246e-01 -3.71230781e-01
-6.83532417e-01 2.14747041e-01 -2.85996228e-01 2.51418829e-01
5.91744959e-01 1.49762720e-01 -4.05789882e-01 2.34468915e-02
9.86424312e-02 1.62229851e-01 4.06231135e-01 7.74454772e-01
6.90051436e-01 -1.44083226e+00 -5.81562936e-01 2.19072178e-01
4.19451714e-01 2.92260408e-01 2.15732545e-01 4.35941815e-01
-4.68744248e-01 4.72493887e-01 -1.99676231e-01 -4.07574117e-01
-1.15870118e+00 4.55110162e-01 5.11461675e-01 -4.88401234e-01
-6.39355481e-01 1.04862320e+00 4.66632098e-01 -5.68693280e-01
2.40158245e-01 -6.97387576e-01 -4.03484106e-01 -3.02869618e-01
7.18356311e-01 -2.01371256e-02 6.86659738e-02 -5.59856892e-01
-5.09813368e-01 7.65601039e-01 -5.22579253e-01 4.88705248e-01
1.52614629e+00 1.69969171e-01 -6.30681872e-01 5.98322272e-01
1.47454202e+00 4.63998243e-02 -1.06023729e+00 -5.51625371e-01
4.37217236e-01 -3.02974492e-01 -1.06035315e-01 -5.45081794e-01
-1.39398348e+00 1.23045933e+00 1.50352961e-03 5.47520518e-02
1.13381243e+00 2.51025381e-03 8.84132922e-01 5.53649664e-01
1.87975317e-01 -4.99394774e-01 1.02112122e-01 7.86650717e-01
6.02473080e-01 -7.85054982e-01 -4.05699104e-01 -3.09189796e-01
-5.18910348e-01 1.27022731e+00 7.42388487e-01 -1.20514050e-01
1.26400009e-01 3.58731657e-01 -1.12106428e-01 -1.53587416e-01
-9.19271469e-01 1.68100715e-01 2.09269494e-01 3.10079098e-01
8.91871810e-01 -1.64650306e-01 -6.06083423e-02 8.87511253e-01
-8.30346569e-02 -1.67007178e-01 6.71760440e-01 8.05591524e-01
-5.69513202e-01 -1.18714857e+00 9.19636860e-02 5.11479080e-01
-5.05746663e-01 -5.16463220e-01 -1.86972842e-01 2.75657475e-01
-7.67839774e-02 6.49230838e-01 4.17497978e-02 -4.53435510e-01
4.65071440e-01 3.31225456e-03 -6.56363592e-02 -1.04331052e+00
-7.41696894e-01 -3.93008351e-01 -1.62446693e-01 -5.23034692e-01
-2.08899900e-01 -4.98157859e-01 -1.77727461e+00 -4.86279055e-02
-1.95127726e-01 5.61362147e-01 5.12823284e-01 8.56393039e-01
4.80383098e-01 9.62517202e-01 7.07450330e-01 -4.07768011e-01
-5.64867020e-01 -8.98067296e-01 -1.09487481e-01 2.92447656e-01
5.47349930e-01 -3.98197204e-01 -3.50055218e-01 1.70686826e-01] | [7.558337688446045, 7.965846538543701] |
9bc69c89-a232-46e5-950e-9cef1900c6c8 | improving-passage-re-ranking-with-word-n-gram | null | null | https://aclanthology.org/2020.icon-main.21 | https://aclanthology.org/2020.icon-main.21.pdf | Improving Passage Re-Ranking with Word N-Gram Aware Coattention Encoder | In text matching applications, coattentions have proved to be highly effective attention mechanisms. Coattention enables the learning to attend based on computing word level affinity scores between two texts. In this paper, we propose two improvements to coattention mechanism in the context of passage ranking (re-ranking). First, we extend the coattention mechanism by applying it across all word n-grams of query and passage. We show that these word n-gram coattentions can capture local context in query and passage to better judge the relevance between them. Second, we further improve the model performance by proposing a query based attention pooling on passage encodings. We evaluate these two methods on MSMARCO passage re-ranking task. The experiment results shows that these two methods resulted in a relative increase of 8.04% in Mean Reciprocal Rank @10 (MRR@10) compared to the naive coattention mechanism. At the time of writing this paper, our methods are the best non transformer model on MS MARCO passage re-ranking task and are competitive to BERT base while only having less than 10% of the parameters. | ['Manish Shrivastava', 'Chaitanya Alaparthi'] | null | null | null | null | icon-2020-12 | ['passage-ranking', 'passage-re-ranking'] | ['natural-language-processing', 'natural-language-processing'] | [ 4.46392298e-02 -2.79459387e-01 -9.54870805e-02 -2.63985902e-01
-1.23517907e+00 -4.50608313e-01 9.54026163e-01 5.57108700e-01
-9.73458052e-01 5.38182437e-01 6.48377657e-01 -2.93302268e-01
-2.92467594e-01 -6.86549842e-01 -7.60290921e-01 -1.77281454e-01
6.58983365e-02 6.94048166e-01 7.29184687e-01 -7.91537404e-01
7.68347859e-01 2.29084283e-01 -1.32017338e+00 8.32735956e-01
9.18124855e-01 5.99865735e-01 2.58317262e-01 9.25178230e-01
-2.20882639e-01 7.77639687e-01 -6.21752322e-01 -6.85353220e-01
3.65037844e-02 -4.67288882e-01 -1.36909235e+00 -8.53867173e-01
7.46167421e-01 -2.89601296e-01 -6.41831219e-01 7.82858074e-01
8.56095850e-01 5.50123334e-01 6.51279926e-01 -5.62932253e-01
-8.59169364e-01 1.00995338e+00 -4.71022397e-01 1.20195138e+00
6.10860705e-01 -5.08189440e-01 1.63001406e+00 -1.14766216e+00
2.55138576e-01 1.09283364e+00 6.94878042e-01 3.95516634e-01
-7.13256836e-01 -2.64015913e-01 -1.10106133e-01 6.47221208e-01
-1.17284679e+00 -4.59325075e-01 4.34276789e-01 -1.14096612e-01
1.60731339e+00 7.22818077e-01 3.16392213e-01 5.48529506e-01
1.06993005e-01 9.90310073e-01 6.36875927e-01 -8.00451338e-01
-2.48906255e-01 -2.79690940e-02 6.08495414e-01 2.86990494e-01
-1.91377044e-01 -2.91558027e-01 -5.80639541e-01 -3.42071503e-01
2.46103108e-01 -1.96057945e-01 -3.90558153e-01 3.49194884e-01
-9.69861627e-01 6.74909353e-01 7.13885307e-01 6.36193454e-01
-2.35895023e-01 1.80967972e-01 3.91972780e-01 5.66524804e-01
5.50120175e-01 8.06147754e-01 -3.23401570e-01 -1.31249458e-01
-8.22337925e-01 3.85323495e-01 5.69480538e-01 7.05472827e-01
3.11681956e-01 -7.94403613e-01 -9.74946499e-01 1.25185573e+00
3.26673388e-01 2.94278264e-01 9.67998862e-01 -5.22651792e-01
7.69566000e-01 4.34683174e-01 9.73497853e-02 -8.97875309e-01
-2.84819007e-01 -6.00218952e-01 -3.08098227e-01 -5.88068008e-01
2.09758505e-01 2.86822110e-01 -6.19676709e-01 1.54384279e+00
-9.86272544e-02 3.99068892e-02 -4.60797474e-02 9.19068038e-01
9.81755018e-01 7.17445910e-01 7.89219290e-02 1.28749371e-01
1.35609019e+00 -1.42857969e+00 -4.71788317e-01 -1.53481126e-01
8.45746398e-01 -1.09586596e+00 1.30101168e+00 -2.85810828e-01
-1.30415881e+00 -5.92031300e-01 -1.04797316e+00 -4.27643210e-01
-3.80048186e-01 -4.92518097e-02 3.71016860e-01 4.84479874e-01
-1.25135493e+00 8.09137285e-01 -3.16966474e-01 -5.50404310e-01
-4.64782231e-02 4.07124013e-01 8.19380507e-02 -5.36278449e-02
-1.80600965e+00 1.02693939e+00 6.05900399e-02 -3.01726665e-02
-1.75562069e-01 -8.72042239e-01 -2.85490900e-01 5.66248178e-01
-3.48604798e-01 -8.48201215e-01 1.44122434e+00 -6.65500879e-01
-1.26897836e+00 1.00191534e+00 -3.34850460e-01 -5.09863138e-01
2.89378673e-01 -3.23764861e-01 -4.44492251e-01 2.03832075e-01
1.28469974e-01 4.78173286e-01 1.91996634e-01 -4.86884296e-01
-6.72690451e-01 -2.45330594e-02 3.91975492e-01 6.28920138e-01
-5.04752278e-01 3.62048894e-01 -6.62008047e-01 -5.37611365e-01
2.70280056e-02 -6.63064241e-01 1.01314439e-02 -6.00003779e-01
1.19565176e-02 -7.32779026e-01 2.81051397e-01 -7.13960111e-01
1.54354191e+00 -1.74541581e+00 -1.42963409e-01 1.25970826e-01
7.76836053e-02 5.05130231e-01 -5.09218693e-01 9.00516331e-01
1.20060742e-01 2.81172186e-01 1.45347700e-01 -1.92017317e-01
4.98385355e-02 -3.48497152e-01 -3.58049542e-01 2.10251063e-01
-9.41501334e-02 1.34948337e+00 -1.04760420e+00 -4.86169785e-01
-3.31622183e-01 2.48169050e-01 -6.60591483e-01 1.65786982e-01
3.01664346e-03 -1.18103661e-01 -4.12776768e-01 1.60786748e-01
3.26948076e-01 -2.79161394e-01 -1.83236059e-02 5.77213615e-02
1.96327999e-01 1.09836316e+00 -2.24196509e-01 1.55179036e+00
-4.80874926e-01 8.73822153e-01 -6.99636638e-01 -7.37967849e-01
6.14525437e-01 3.11426282e-01 2.23222286e-01 -1.19431531e+00
3.64470370e-02 2.87444800e-01 1.46339312e-01 -4.82757717e-01
9.27041411e-01 1.63950637e-01 4.99476865e-02 6.34680390e-01
-1.72743931e-01 2.27266356e-01 4.63123590e-01 5.96268177e-01
1.21581972e+00 -2.03282923e-01 8.09383765e-02 -4.72066402e-01
8.24820280e-01 -1.57448456e-01 -6.40866905e-02 1.17421544e+00
-4.12634872e-02 5.87145507e-01 1.68467641e-01 -1.16653614e-01
-1.09711778e+00 -7.78129816e-01 -1.74895898e-01 1.63190544e+00
1.87825590e-01 -3.18433851e-01 -6.49419844e-01 -5.76820910e-01
-1.21891089e-01 5.54522753e-01 -5.89391828e-01 -2.79034495e-01
-8.39815557e-01 -9.47481811e-01 7.50877500e-01 7.81173408e-01
4.15802151e-01 -1.11460936e+00 -1.34791201e-02 2.70338774e-01
-7.63623297e-01 -8.71641994e-01 -1.03762591e+00 -2.13483170e-01
-8.67626846e-01 -8.46640408e-01 -1.08179784e+00 -9.92762566e-01
3.25613201e-01 5.22365272e-01 1.43895483e+00 6.01229250e-01
-1.18498690e-01 3.64552408e-01 -7.88692057e-01 -1.01848230e-01
-8.21650326e-02 5.35348296e-01 -1.50280476e-01 -2.93795198e-01
7.99038708e-01 -4.91165429e-01 -7.63562620e-01 1.74065992e-01
-7.30977237e-01 -2.89822310e-01 4.48475569e-01 9.94874954e-01
2.25115225e-01 -6.88911974e-01 8.09649527e-01 -7.79854357e-01
1.19172335e+00 -5.45847952e-01 -1.62481859e-01 5.74770927e-01
-7.23198295e-01 1.28212884e-01 3.08349073e-01 -3.90402496e-01
-7.78340638e-01 -6.55309737e-01 -4.43347722e-01 2.04040065e-01
4.56441998e-01 7.45473921e-01 4.61711556e-01 -1.51306642e-02
8.11334789e-01 9.65430811e-02 -5.42002380e-01 -7.28210628e-01
2.30437279e-01 1.01208389e+00 2.70438701e-01 -4.41577017e-01
2.96666592e-01 -1.15669012e-01 -3.79212946e-01 -5.76838195e-01
-9.11170542e-01 -1.08779597e+00 -3.94616753e-01 -1.44263380e-04
5.45483708e-01 -6.86823130e-01 -7.15370834e-01 2.32075676e-01
-1.08089685e+00 -2.44029880e-01 1.39311716e-01 5.88014483e-01
-2.13715911e-01 6.95061803e-01 -1.16935885e+00 -4.50023323e-01
-7.05978334e-01 -6.42041564e-01 7.54430771e-01 3.08247089e-01
-1.77761704e-01 -9.46108401e-01 4.05759603e-01 6.53071821e-01
7.14761674e-01 -6.98290467e-01 1.10355985e+00 -1.08477664e+00
-2.31729731e-01 -4.19519871e-01 -4.29622948e-01 -2.59532303e-01
-1.60523355e-01 -3.86461169e-01 -9.61615145e-01 -4.06271994e-01
-2.72100002e-01 -1.74172431e-01 1.19351947e+00 1.49915129e-01
8.88907135e-01 -1.28754750e-01 -2.15570852e-01 1.12188347e-01
1.11558807e+00 1.00154012e-01 7.89714217e-01 7.15313017e-01
2.60848224e-01 3.94255191e-01 5.09606242e-01 1.04978010e-01
3.66337627e-01 1.05565000e+00 1.06907770e-01 8.64143446e-02
-3.37692469e-01 -2.31127590e-01 3.10655922e-01 1.30989659e+00
-7.24860728e-02 -6.41921818e-01 -7.85328805e-01 7.70067155e-01
-1.90278506e+00 -1.28755772e+00 -3.70330215e-01 2.25328541e+00
8.42225313e-01 -1.83866881e-02 8.43526125e-02 -1.31023362e-01
7.43945599e-01 -7.29952306e-02 -1.15220286e-01 -5.75579822e-01
-8.72386843e-02 2.72973478e-01 2.79896945e-01 9.97751355e-01
-8.47416341e-01 8.38850617e-01 6.57608080e+00 8.04223359e-01
-6.84536695e-01 1.55020654e-01 3.27867240e-01 -1.32440999e-01
-4.24236745e-01 -2.02177629e-01 -7.96927094e-01 4.08689290e-01
1.07800269e+00 -4.04303253e-01 4.23014432e-01 2.94875503e-01
-2.19971925e-01 1.48162514e-01 -1.08024096e+00 5.57998478e-01
3.59484524e-01 -1.20307851e+00 1.78299055e-01 -2.99314350e-01
6.71172798e-01 4.67426717e-01 -1.14215963e-01 6.89147651e-01
5.97517751e-02 -8.50254118e-01 3.57560754e-01 5.79890251e-01
2.78466254e-01 -6.28196776e-01 1.15811574e+00 2.18911469e-01
-1.04853380e+00 2.67469790e-02 -7.19310164e-01 -1.07306063e-01
2.44523045e-02 2.16818750e-01 -7.85618365e-01 6.23363853e-01
7.19772458e-01 3.87274653e-01 -7.66200244e-01 1.50228775e+00
-5.84786646e-02 6.58405483e-01 -1.98819980e-01 -6.29760325e-01
3.05407584e-01 2.01550469e-01 5.35375834e-01 1.48996067e+00
1.37615398e-01 1.22403279e-01 -1.93722844e-01 3.05675715e-01
-3.44911695e-01 5.83380938e-01 1.85711961e-02 2.36180611e-02
3.45299751e-01 8.32526922e-01 -2.68031418e-01 -5.24774492e-01
-4.04981464e-01 1.34263742e+00 7.92553723e-01 3.66726816e-01
-7.64623463e-01 -8.06136608e-01 2.33119950e-01 3.42457257e-02
2.57973611e-01 2.59867907e-01 7.01548234e-02 -1.08193564e+00
1.97086036e-01 -5.79243720e-01 5.13813972e-01 -7.55433917e-01
-1.57845581e+00 9.08898950e-01 -1.16214499e-01 -9.13795054e-01
-1.04750946e-01 -4.11825210e-01 -5.97947478e-01 1.29250526e+00
-1.70139217e+00 -8.16552997e-01 1.91924758e-02 3.47242355e-01
6.67637348e-01 -2.06486076e-01 7.41373181e-01 7.88704455e-01
-1.79087445e-01 1.03364539e+00 5.34671247e-01 2.76072145e-01
7.68451214e-01 -1.24163604e+00 7.48566210e-01 6.15103066e-01
5.27982950e-01 9.80971754e-01 4.32429999e-01 -4.04719502e-01
-9.05592382e-01 -6.49940848e-01 1.94489086e+00 -5.69795668e-01
5.37360966e-01 -3.64336297e-02 -1.01404440e+00 4.39825535e-01
3.93550843e-01 -4.03212339e-01 7.87962556e-01 5.52781224e-01
-5.16267776e-01 -4.65042666e-02 -8.09983075e-01 7.37832844e-01
1.01586056e+00 -8.71195734e-01 -1.17371559e+00 5.13483047e-01
8.71746778e-01 -2.65579611e-01 -7.32640862e-01 3.68794769e-01
5.77461362e-01 -7.80674696e-01 1.13157594e+00 -9.91344154e-01
4.01176751e-01 -1.18143432e-01 -2.12506428e-01 -1.05610681e+00
-6.97347283e-01 -4.99136627e-01 -7.71709830e-02 1.27024066e+00
8.23303759e-01 -6.59961224e-01 5.49700856e-01 3.81383181e-01
-3.08432311e-01 -9.18275058e-01 -8.49931121e-01 -5.81986904e-01
4.14508611e-01 -9.90642831e-02 6.42177880e-01 9.11197662e-01
6.06381774e-01 7.34403074e-01 -1.79293498e-01 -1.46642208e-01
-4.66431491e-03 1.49636403e-01 2.44751066e-01 -9.03121591e-01
-5.88728845e-01 -8.50725651e-01 -2.53568351e-01 -1.60504508e+00
-3.74243557e-02 -1.31859183e+00 9.50185657e-02 -1.70260346e+00
7.03346372e-01 -3.59678686e-01 -8.73696089e-01 2.96133995e-01
-7.12335050e-01 5.27797699e-01 9.77885127e-02 4.09411132e-01
-7.95010090e-01 5.39182305e-01 9.46363449e-01 -3.59424561e-01
-1.80493761e-02 1.46873444e-01 -6.97493732e-01 9.80148837e-02
8.25978458e-01 -5.21685660e-01 -3.76488239e-01 -8.72947395e-01
4.90882695e-01 -2.86515895e-02 9.07231122e-02 -7.23973632e-01
4.64062393e-01 2.81608284e-01 1.13484807e-01 -5.32492340e-01
1.67370304e-01 -2.06048310e-01 -4.76485252e-01 4.27283764e-01
-9.43835378e-01 5.57926416e-01 1.61175132e-01 4.69937027e-01
-2.67267168e-01 -6.09744966e-01 4.36064512e-01 3.25291529e-02
-5.65559506e-01 8.06076676e-02 -5.17509997e-01 2.73908556e-01
2.24140584e-01 4.04304676e-02 -5.78882456e-01 -5.14123261e-01
-3.80882293e-01 3.70949179e-01 5.01558185e-02 6.90232337e-01
3.34243506e-01 -1.42059600e+00 -1.04534256e+00 -1.76480532e-01
2.20595464e-01 -8.14512670e-01 2.60371208e-01 9.04280782e-01
-4.46338236e-01 1.02613401e+00 1.58000305e-01 -1.13928191e-01
-1.39610863e+00 4.40763950e-01 4.17936057e-01 -6.22266173e-01
-4.49384391e-01 1.18541288e+00 -1.36496738e-01 -4.69999015e-01
1.89747244e-01 -7.89208114e-02 -6.54388964e-01 -7.44676739e-02
7.91344523e-01 4.29650187e-01 4.41987991e-01 -5.89549601e-01
-4.06248719e-01 6.19173050e-01 -5.21539748e-01 -3.27544987e-01
9.70103204e-01 -2.21144676e-01 -2.68835485e-01 1.28700763e-01
1.42103755e+00 2.73473978e-01 -3.36015671e-01 -3.70088845e-01
3.99988472e-01 -5.04587710e-01 1.83832049e-02 -1.08033395e+00
-5.48296630e-01 8.68284762e-01 3.97049934e-01 2.37776950e-01
9.10364866e-01 4.65646796e-02 1.00702369e+00 6.85117066e-01
8.06662217e-02 -8.93348575e-01 5.30083366e-02 8.30541670e-01
8.59949112e-01 -9.20395374e-01 -6.50050715e-02 -4.65973504e-02
-3.10506463e-01 8.82536769e-01 5.00273347e-01 -8.56415778e-02
4.04202372e-01 -4.46160793e-01 9.97063667e-02 -1.39260799e-01
-8.49069953e-01 -3.63191515e-01 9.83158648e-01 1.20523922e-01
1.03253198e+00 -3.33731472e-01 -1.05887723e+00 4.15640652e-01
-1.14798345e-01 -3.38640988e-01 1.73051998e-01 6.69379473e-01
-6.42706156e-01 -1.34353518e+00 6.74606068e-03 5.86292505e-01
-9.67254519e-01 -7.64127851e-01 -5.16726494e-01 3.42336416e-01
-5.15721858e-01 1.18145692e+00 1.21444225e-01 -3.46867800e-01
3.23763341e-01 1.07915714e-01 5.42231262e-01 -4.62484330e-01
-1.07044041e+00 -2.27043018e-01 3.31404567e-01 -2.51265913e-01
-1.68387890e-01 -4.93381053e-01 -1.02679932e+00 -2.75061905e-01
-6.31625354e-01 6.88462853e-01 3.38474870e-01 9.65713918e-01
3.93960565e-01 4.55608189e-01 6.39815807e-01 -4.49886858e-01
-6.88039839e-01 -1.29604840e+00 -1.57878146e-01 5.54179609e-01
1.44935191e-01 -2.64148355e-01 -2.51967579e-01 -5.04789293e-01] | [11.513226509094238, 7.7163519859313965] |
0c78524c-e170-4d77-af10-9885b0f0db7c | gen-ir-sigir-2023-the-first-workshop-on | 2306.02887 | null | https://arxiv.org/abs/2306.02887v2 | https://arxiv.org/pdf/2306.02887v2.pdf | Gen-IR @ SIGIR 2023: The First Workshop on Generative Information Retrieval | Generative information retrieval (IR) has experienced substantial growth across multiple research communities (e.g., information retrieval, computer vision, natural language processing, and machine learning), and has been highly visible in the popular press. Theoretical, empirical, and actual user-facing products have been released that retrieve documents (via generation) or directly generate answers given an input request. We would like to investigate whether end-to-end generative models are just another trend or, as some claim, a paradigm change for IR. This necessitates new metrics, theoretical grounding, evaluation methods, task definitions, models, user interfaces, etc. The goal of this workshop (https://coda.io/@sigir/gen-ir) is to focus on previously explored Generative IR techniques like document retrieval and direct Grounded Answer Generation, while also offering a venue for the discussion and exploration of how Generative IR can be applied to new domains like recommendation systems, summarization, etc. The format of the workshop is interactive, including roundtable and keynote sessions and tends to avoid the one-sided dialogue of a mini-conference. | ['Donald Metzler', 'Ruqing Zhang', 'Gabriel Bénédict'] | 2023-06-05 | null | null | null | null | ['information-retrieval', 'answer-generation'] | ['natural-language-processing', 'natural-language-processing'] | [ 4.07891601e-01 2.68622547e-01 -1.74700066e-01 -2.48217031e-01
-1.33177328e+00 -9.31655765e-01 1.09941638e+00 3.18901986e-01
-1.79090604e-01 4.52852517e-01 8.42029750e-01 -2.86622614e-01
-3.11884075e-01 -4.55758482e-01 -9.63104442e-02 -4.31928366e-01
3.55591893e-01 7.50864983e-01 -1.39188811e-01 -4.56856668e-01
9.27888215e-01 2.48568673e-02 -1.70639884e+00 7.79276490e-01
6.70803547e-01 4.98858035e-01 1.36533305e-01 8.23725760e-01
-4.40340787e-01 7.74631560e-01 -7.32789397e-01 -3.87503237e-01
-2.15852395e-01 -7.91276574e-01 -1.21485436e+00 -1.81428641e-01
3.08462918e-01 1.37227833e-01 -2.11252552e-02 6.81977868e-01
8.32280278e-01 3.47076535e-01 6.70857847e-01 -9.08856213e-01
-1.33185124e+00 9.12549496e-01 -3.72564495e-01 1.48445651e-01
8.57115865e-01 -1.83512047e-01 1.08702457e+00 -9.35591638e-01
8.75606656e-01 1.47072184e+00 2.55729228e-01 6.37397945e-01
-1.06127667e+00 -1.80246517e-01 1.19668104e-01 1.27095863e-01
-1.14866149e+00 -4.85332698e-01 5.15634835e-01 -4.17648852e-01
8.67333591e-01 7.43431687e-01 1.81312367e-01 1.10176206e+00
-5.72526678e-02 1.27346957e+00 7.30641663e-01 -8.48782063e-01
2.10755840e-01 5.91169655e-01 5.69695175e-01 9.85634327e-02
2.45204046e-01 -2.74131447e-01 -3.61575246e-01 -2.73404300e-01
4.06621844e-01 -7.80150369e-02 -4.87447828e-01 4.46384363e-02
-1.22934175e+00 1.07648563e+00 3.46345045e-02 7.13490903e-01
-3.90650749e-01 -5.95795624e-02 4.94676530e-01 5.11275113e-01
7.20587909e-01 8.41172218e-01 7.48228207e-02 -3.08372617e-01
-8.34967554e-01 7.04152703e-01 1.06706107e+00 9.06837404e-01
4.68858927e-01 -1.89516217e-01 -6.36702955e-01 1.08867490e+00
3.63079667e-01 3.44253540e-01 7.62977481e-01 -9.78668213e-01
2.62308151e-01 6.16265953e-01 1.80124685e-01 -1.00483036e+00
-2.86291093e-01 -2.12123513e-01 -5.68849564e-01 -3.56820226e-01
2.82845832e-02 -2.45621502e-01 -7.15578079e-01 1.28357017e+00
8.29231068e-02 -7.00947881e-01 -2.72393730e-02 1.04143238e+00
1.27145493e+00 9.06497002e-01 -2.47761654e-03 -4.96169358e-01
1.30812860e+00 -7.60457814e-01 -7.55786538e-01 -1.97697729e-01
8.32306325e-01 -1.36936927e+00 1.03503478e+00 4.14477527e-01
-1.33718288e+00 -4.80123699e-01 -6.30525827e-01 -2.81352609e-01
-2.33717784e-01 6.67727068e-02 6.01917326e-01 4.95297432e-01
-1.31028044e+00 1.11922845e-01 -4.32296216e-01 -8.18445385e-01
-3.56850535e-01 -1.05061404e-01 6.39934838e-02 -4.11625266e-01
-1.09707201e+00 9.19747651e-01 1.08949646e-01 -1.26399949e-01
-1.81503743e-01 -5.62867820e-01 -5.55203259e-01 -1.94734484e-01
3.91890049e-01 -9.59890902e-01 1.48420465e+00 -9.09337163e-01
-1.36975014e+00 8.49101603e-01 -1.18714243e-01 -1.54330164e-01
1.45556167e-01 -4.96469915e-01 -3.48282367e-01 -6.00686483e-02
-2.71574222e-02 6.14033997e-01 5.35440683e-01 -1.24112666e+00
-4.44115758e-01 -5.78084648e-01 -8.69924948e-03 5.36011279e-01
-2.80963391e-01 5.81147313e-01 -2.89095104e-01 -6.02675974e-01
6.21882081e-02 -1.01697934e+00 -1.21503547e-01 -8.49003792e-01
-3.45195234e-01 -9.91840541e-01 5.14534175e-01 -5.04434109e-01
1.48421097e+00 -1.90714514e+00 1.45850375e-01 -1.15148187e-01
-8.18006247e-02 2.77511805e-01 -1.80191949e-01 1.29849458e+00
9.49084088e-02 3.96028012e-01 1.61884978e-01 -1.32690936e-01
3.07698697e-01 -2.98510075e-01 -7.57566512e-01 4.06753942e-02
-2.17627689e-01 9.29692149e-01 -9.03036892e-01 -2.39823982e-01
-3.94415632e-02 7.14278996e-01 -3.65928084e-01 -3.61478887e-02
-3.72850031e-01 5.33271171e-02 -6.27707064e-01 2.36479282e-01
1.55641660e-01 -3.19744885e-01 1.17873982e-01 9.23295468e-02
-3.03498566e-01 7.00125515e-01 -1.15679193e+00 1.73504496e+00
-3.24431300e-01 7.83772826e-01 -3.19314718e-01 -6.13088429e-01
1.00703633e+00 5.16045392e-01 3.25654030e-01 -9.08141494e-01
8.90280157e-02 1.21292613e-01 -1.71506137e-01 -4.65530068e-01
1.26252723e+00 1.45458326e-01 -1.10972084e-01 1.02421844e+00
-2.28694156e-01 -1.64040700e-01 4.54308957e-01 5.87358952e-01
1.01309645e+00 1.40876010e-01 2.29710545e-02 -3.90922464e-02
2.26214573e-01 3.21808904e-01 -1.59899443e-01 9.97015119e-01
3.85073721e-01 7.91139841e-01 7.97459632e-02 -3.85018915e-01
-8.43157709e-01 -6.40884638e-01 6.72378540e-02 1.34526789e+00
1.57997720e-02 -6.95477962e-01 -7.25494325e-01 -1.91663727e-01
-3.65562797e-01 1.04646194e+00 -3.85804564e-01 -3.29816401e-01
-5.59191704e-01 -6.81372464e-01 3.11769426e-01 1.48679942e-01
-3.25112306e-02 -1.50708544e+00 -4.14320916e-01 3.26722920e-01
-4.84072447e-01 -5.65888524e-01 -4.47849065e-01 -1.99069083e-01
-9.09838498e-01 -5.96676350e-01 -9.86850619e-01 -7.05320418e-01
3.87183011e-01 6.82141244e-01 1.39352226e+00 1.66618064e-01
-1.95404768e-01 9.14335012e-01 -8.51952314e-01 -5.24955392e-01
-5.52019119e-01 2.17889592e-01 -1.71670869e-01 -2.34003887e-01
5.71547151e-01 -2.13108823e-01 -9.03743029e-01 2.95459837e-01
-1.24993098e+00 -1.22227348e-01 6.42794132e-01 4.64445531e-01
2.95599192e-01 -5.53567350e-01 8.20552707e-01 -9.96563196e-01
1.50210524e+00 -3.83845806e-01 -1.36378616e-01 4.95331347e-01
-1.16005719e+00 1.22144289e-01 -8.66897255e-02 -2.02232644e-01
-1.20852160e+00 -6.56311452e-01 -8.29772651e-02 1.17474243e-01
-8.93298239e-02 9.44259405e-01 2.27178007e-01 6.03247762e-01
1.21350873e+00 2.57773638e-01 -1.01083465e-01 -5.90238631e-01
6.69100225e-01 9.75800216e-01 2.11515948e-01 -5.53983569e-01
5.95352292e-01 1.73224762e-01 -7.09010780e-01 -8.10839474e-01
-7.21627951e-01 -8.52298856e-01 -4.65349928e-02 -2.05867156e-01
5.43637693e-01 -6.22980297e-01 -2.66667336e-01 -1.99889988e-01
-1.34724808e+00 -2.01110944e-01 -4.84833330e-01 3.72336090e-01
-4.04852033e-01 2.15969190e-01 -5.62122643e-01 -8.46270084e-01
-8.93263817e-01 -7.92790532e-01 1.06307447e+00 4.29523796e-01
-7.84225464e-01 -1.08804536e+00 5.91145217e-01 7.26397812e-01
7.52424419e-01 -2.84983635e-01 7.78127670e-01 -8.69030774e-01
-3.84953111e-01 -5.69054902e-01 -3.87047641e-02 1.04100935e-01
4.41054702e-02 2.47674063e-01 -8.47264826e-01 -1.51191890e-01
-2.54167646e-01 -4.41162348e-01 9.10767376e-01 4.14282501e-01
6.99435472e-01 -6.93368137e-01 -3.19570780e-01 -2.74377733e-01
1.15812552e+00 3.22146952e-01 8.80248487e-01 2.91500717e-01
3.41657221e-01 9.33297634e-01 4.68375444e-01 3.98835897e-01
3.68292779e-01 8.06548476e-01 -1.13834940e-01 4.14773762e-01
-2.24914819e-01 -2.16668069e-01 4.33367997e-01 9.92023230e-01
-2.34671250e-01 -5.77186108e-01 -7.28521943e-01 4.88291204e-01
-1.95001066e+00 -1.31172299e+00 -3.88296008e-01 2.50475764e+00
6.68300748e-01 -2.46243864e-01 1.02809168e-01 -7.60688260e-02
6.31363571e-01 4.47529368e-02 -1.08507760e-01 -7.13271141e-01
8.95338655e-02 1.49331808e-01 -2.02946037e-01 4.47246015e-01
-5.90125859e-01 6.48515582e-01 6.33052588e+00 7.24116504e-01
-1.16107678e+00 -1.17965482e-01 4.89800155e-01 -1.21986091e-01
-8.06666195e-01 5.11370040e-02 -8.61251712e-01 1.04490660e-01
9.87682581e-01 -6.93990290e-01 2.94281214e-01 7.21491456e-01
2.10820138e-01 8.50772951e-03 -9.92025733e-01 8.21570754e-01
4.94643837e-01 -1.41473722e+00 2.88666904e-01 5.88336699e-02
8.33809495e-01 8.75980482e-02 1.97776467e-01 6.40109777e-01
4.63921279e-01 -8.60862792e-01 5.54297209e-01 4.31871384e-01
4.69714940e-01 -4.93777364e-01 7.14803755e-01 2.61171013e-01
-5.84086299e-01 1.46060169e-01 -3.19390416e-01 -2.03682091e-02
2.54806131e-01 4.19669122e-01 -7.45934784e-01 6.66134655e-01
5.05665302e-01 2.66982913e-01 -5.13837516e-01 1.15904701e+00
-9.12302453e-03 5.33955336e-01 9.95123312e-02 -4.06253725e-01
1.58709168e-01 -6.06897771e-02 6.43288851e-01 1.41747117e+00
3.41107845e-01 1.01332471e-01 1.55014545e-02 4.99717891e-01
1.26469042e-02 5.38638592e-01 -5.60112476e-01 -3.91787797e-01
3.91732663e-01 1.49774301e+00 -7.67221451e-01 -1.76749304e-01
-1.67190824e-02 8.00628543e-01 -1.36283010e-01 4.91339296e-01
-2.68664688e-01 -3.59421134e-01 1.63278535e-01 2.33072519e-01
-1.35979071e-01 7.21977800e-02 -3.13071400e-01 -9.15264606e-01
-6.26244769e-02 -1.14693618e+00 5.14534473e-01 -9.95908797e-01
-1.17333055e+00 7.17101574e-01 2.46250555e-01 -1.06503761e+00
-9.14953232e-01 -1.89197600e-01 -5.64663291e-01 8.54611576e-01
-9.76521790e-01 -6.10252559e-01 -2.96545476e-01 2.63678849e-01
7.75411069e-01 3.80927362e-02 9.91111815e-01 2.76159972e-01
-2.41846338e-01 4.68819916e-01 1.26115799e-01 -1.39108449e-01
1.08022225e+00 -1.03272939e+00 2.81000227e-01 4.45680946e-01
6.28730059e-01 9.14172947e-01 9.59622860e-01 -2.69265741e-01
-1.77448821e+00 -6.82596743e-01 1.35430670e+00 -8.59229863e-01
4.60863531e-01 1.79738943e-02 -8.20925891e-01 6.41112447e-01
6.76070809e-01 -7.30779707e-01 7.52384067e-01 4.00843471e-01
-2.84732133e-01 6.53745756e-02 -7.77807176e-01 7.01631606e-01
7.44746625e-01 -3.33894521e-01 -5.98556578e-01 6.91842139e-01
8.20052981e-01 -2.61659592e-01 -5.33056736e-01 1.71750844e-01
5.37267029e-01 -7.33141541e-01 8.42308044e-01 -5.26423395e-01
5.36512017e-01 -1.31782163e-02 -2.46363860e-02 -1.27658451e+00
-5.97985685e-01 -9.42793369e-01 1.27339914e-01 1.45331848e+00
6.06647551e-01 -4.20683563e-01 4.23371911e-01 8.95883799e-01
-5.03217816e-01 -4.42199320e-01 -4.63129759e-01 -5.25802612e-01
8.41827914e-02 -4.06658083e-01 1.25848562e-01 7.51239300e-01
4.21075433e-01 1.09677088e+00 -5.80625050e-02 -5.71538031e-01
2.24114910e-01 2.65288830e-01 9.47567403e-01 -1.26190376e+00
-2.53928721e-01 -8.74548256e-01 -8.16316530e-02 -1.07034993e+00
-6.59941435e-01 -9.94752049e-01 -4.12275717e-02 -2.29418468e+00
3.96908581e-01 -1.56464413e-01 -1.92218885e-01 2.59315729e-01
-1.39241904e-01 3.06149751e-01 1.38047934e-01 5.87503672e-01
-9.33094323e-01 3.07177633e-01 1.30554807e+00 -5.94154559e-02
-5.22320986e-01 1.20393455e-01 -1.41680777e+00 2.13954508e-01
6.79868340e-01 -3.77140552e-01 -6.70511782e-01 -5.22160947e-01
9.52506959e-01 -3.91447097e-02 1.55357212e-01 -5.26062727e-01
4.01835054e-01 -4.92752856e-03 -3.99697833e-02 -3.70034337e-01
9.67048779e-02 -1.81735799e-01 1.63353562e-01 1.72277968e-02
-8.86015654e-01 1.83206007e-01 7.68084303e-02 2.65233785e-01
-3.99297893e-01 -4.94879872e-01 1.74200293e-02 -2.46870592e-01
-3.12093645e-01 8.30430910e-02 -4.98814315e-01 3.10949415e-01
5.28566122e-01 -1.76381916e-01 -6.01601601e-01 -7.87398696e-01
-3.38526726e-01 1.11811720e-01 1.15073361e-01 9.07570064e-01
6.02463305e-01 -1.11549652e+00 -1.24885607e+00 -4.29258376e-01
4.00921166e-01 -1.55046985e-01 3.77045810e-01 7.93881893e-01
-1.41769052e-01 9.52720761e-01 3.75810564e-01 -4.02534544e-01
-1.32144761e+00 4.09551233e-01 -1.90583840e-01 -4.65989023e-01
-5.93354225e-01 7.11775422e-01 1.03059553e-01 -3.32935154e-01
2.00114578e-01 9.00350213e-02 -4.67158705e-01 3.27164561e-01
9.87479985e-01 1.95730448e-01 1.49594516e-01 -1.05104208e-01
-9.48176309e-02 1.43747434e-01 -4.87084091e-01 -4.57690924e-01
1.38034928e+00 -1.95403516e-01 -1.98400944e-01 6.15842462e-01
1.03488719e+00 -1.28019273e-01 -3.40711325e-01 -3.50664496e-01
2.81802893e-01 -6.13535494e-02 5.74463513e-03 -1.08193648e+00
-5.06502688e-01 6.40540004e-01 3.53703976e-01 7.52185762e-01
9.11715567e-01 3.65029514e-01 5.21287084e-01 4.97393787e-01
2.90826112e-01 -1.10434365e+00 1.57061204e-01 5.49602985e-01
1.47486925e+00 -1.02161241e+00 7.53842145e-02 1.06764048e-01
-6.11581624e-01 1.07936525e+00 1.67900085e-01 1.24580696e-01
3.94540220e-01 -1.56968683e-01 1.02084376e-01 -3.87885064e-01
-1.09561980e+00 -1.74112946e-01 6.80074334e-01 3.80619258e-01
1.23882771e+00 -1.03167675e-01 -8.92661929e-01 3.46381336e-01
-3.28918576e-01 1.06138244e-01 3.99450988e-01 8.82595539e-01
-6.69644475e-01 -1.29637289e+00 -3.13104898e-01 3.32988232e-01
-5.30380726e-01 -2.41195217e-01 -9.07187939e-01 6.92947328e-01
-6.44944668e-01 1.30088437e+00 -6.98039755e-02 -2.17177674e-01
2.77231485e-01 2.79526234e-01 5.46701193e-01 -9.21268880e-01
-8.96683276e-01 1.36279374e-01 3.23077053e-01 -1.31547183e-01
-3.52194548e-01 -7.20792174e-01 -9.41501260e-01 -3.52468282e-01
-3.25557381e-01 5.69960892e-01 1.06823182e+00 8.55067492e-01
6.75301850e-01 3.11913133e-01 5.12027621e-01 -6.51895285e-01
-4.88040835e-01 -1.47800934e+00 -9.40086544e-02 2.59085894e-01
-1.80998370e-01 -1.22579604e-01 -2.82559484e-01 6.37429729e-02] | [12.231797218322754, 9.2449951171875] |
754ba408-19c7-49c6-bb0e-ee669f6cbe0b | hero-hierarchical-encoder-for-video-language | 2005.00200 | null | https://arxiv.org/abs/2005.00200v2 | https://arxiv.org/pdf/2005.00200v2.pdf | HERO: Hierarchical Encoder for Video+Language Omni-representation Pre-training | We present HERO, a novel framework for large-scale video+language omni-representation learning. HERO encodes multimodal inputs in a hierarchical structure, where local context of a video frame is captured by a Cross-modal Transformer via multimodal fusion, and global video context is captured by a Temporal Transformer. In addition to standard Masked Language Modeling (MLM) and Masked Frame Modeling (MFM) objectives, we design two new pre-training tasks: (i) Video-Subtitle Matching (VSM), where the model predicts both global and local temporal alignment; and (ii) Frame Order Modeling (FOM), where the model predicts the right order of shuffled video frames. HERO is jointly trained on HowTo100M and large-scale TV datasets to gain deep understanding of complex social dynamics with multi-character interactions. Comprehensive experiments demonstrate that HERO achieves new state of the art on multiple benchmarks over Text-based Video/Video-moment Retrieval, Video Question Answering (QA), Video-and-language Inference and Video Captioning tasks across different domains. We also introduce two new challenging benchmarks How2QA and How2R for Video QA and Retrieval, collected from diverse video content over multimodalities. | ['Jingjing Liu', 'Zhe Gan', 'Yen-Chun Chen', 'Yu Cheng', 'Linjie Li', 'Licheng Yu'] | 2020-05-01 | null | https://aclanthology.org/2020.emnlp-main.161 | https://aclanthology.org/2020.emnlp-main.161.pdf | emnlp-2020-11 | ['moment-retrieval'] | ['computer-vision'] | [ 3.67130429e-01 -2.76680142e-01 -2.95156091e-01 -4.45591420e-01
-1.32335389e+00 -3.88040185e-01 8.55013072e-01 -1.76899672e-01
-3.46112251e-01 3.21414292e-01 6.99752569e-01 -1.30256683e-01
2.39553615e-01 -3.69493186e-01 -1.40415108e+00 -4.05487537e-01
-1.86448887e-01 4.55672413e-01 2.83367008e-01 -1.76792324e-01
1.75354108e-02 -1.48432717e-01 -1.62262213e+00 1.24305582e+00
1.33999690e-01 1.31002319e+00 3.25980663e-01 1.21557689e+00
-1.86479479e-01 1.67832994e+00 -3.20065349e-01 -6.50540113e-01
6.44368902e-02 -4.51052278e-01 -9.75708365e-01 2.62812793e-01
1.07179964e+00 -7.80541718e-01 -1.14891922e+00 6.68245375e-01
3.92858952e-01 2.50674605e-01 4.29176986e-01 -1.44515204e+00
-7.14482367e-01 4.85651553e-01 -4.02368516e-01 3.64033431e-01
9.85259891e-01 4.24812526e-01 9.38036919e-01 -9.85151410e-01
8.55166018e-01 1.90925896e+00 2.38347843e-01 4.99889255e-01
-7.14771211e-01 -3.99896979e-01 3.15532506e-01 6.85844660e-01
-1.29661965e+00 -7.66754210e-01 4.10410464e-01 -4.90395635e-01
9.47441757e-01 2.25792021e-01 3.89016181e-01 1.49601448e+00
4.13991809e-02 1.28621900e+00 4.91810381e-01 -5.03591299e-02
-1.80637613e-01 -4.94014651e-01 -1.29038706e-01 1.05547273e+00
-7.04548717e-01 -1.68360725e-01 -1.11403000e+00 -1.02279373e-01
7.11343348e-01 1.81611534e-02 -2.04617858e-01 1.41215518e-01
-1.74169719e+00 5.68223417e-01 -4.98049818e-02 2.59827338e-02
-2.70036012e-01 6.53577805e-01 7.39498198e-01 6.47891819e-01
2.31981501e-01 -1.29328892e-01 -1.71061903e-01 -4.91118848e-01
-1.14604354e+00 3.11595470e-01 5.23171544e-01 1.25492322e+00
6.01314664e-01 1.36926576e-01 -7.09112108e-01 8.15827310e-01
4.73028123e-01 8.03819954e-01 5.05013824e-01 -1.52095807e+00
8.91490936e-01 2.62646109e-01 2.79295798e-02 -1.19675684e+00
3.58697362e-02 4.18771297e-01 -7.57092535e-01 -7.10611641e-01
2.61719048e-01 1.54115647e-01 -1.04141641e+00 1.75265610e+00
-2.26719072e-03 8.30909848e-01 -7.85007849e-02 1.06471109e+00
1.36009669e+00 1.26069069e+00 8.00834820e-02 -2.34399021e-01
1.56208336e+00 -1.40441585e+00 -8.14178944e-01 -2.79775150e-02
4.02611613e-01 -7.03418493e-01 9.95852649e-01 1.83032095e-01
-1.63111746e+00 -7.42943168e-01 -5.35315514e-01 -6.31948352e-01
-7.46398941e-02 -7.81260654e-02 2.05584943e-01 -6.32213876e-02
-1.27730274e+00 1.30752012e-01 -6.94566429e-01 -3.57050836e-01
2.70342886e-01 1.42306492e-01 -6.04712605e-01 -4.97331172e-01
-1.33698547e+00 2.96530515e-01 6.37441427e-02 1.99810162e-01
-1.53357410e+00 -6.11383915e-01 -1.14866960e+00 -3.70177440e-02
4.13877040e-01 -8.65853965e-01 1.23021519e+00 -1.26596987e+00
-1.33310747e+00 1.09368825e+00 -8.32340479e-01 -6.75164104e-01
4.61752266e-01 -4.14566875e-01 -4.94184524e-01 9.28868592e-01
5.09325936e-02 1.12642574e+00 1.26503778e+00 -1.03138232e+00
-5.98884642e-01 -1.62836015e-01 3.21423143e-01 3.50114107e-01
-1.61973268e-01 4.85961378e-01 -1.28734851e+00 -6.82867646e-01
-3.52905482e-01 -7.25823998e-01 2.57168680e-01 1.37845427e-01
-1.81226566e-01 -1.54399186e-01 1.06455803e+00 -1.15132892e+00
1.37920225e+00 -2.10516906e+00 6.74144447e-01 -2.71517247e-01
3.91249239e-01 6.80804923e-02 -6.67891920e-01 5.10534644e-01
4.72948328e-03 -9.38710645e-02 9.34625417e-02 -6.98447764e-01
5.48233613e-02 4.41067338e-01 -6.39656246e-01 2.58560032e-01
2.71134079e-01 1.35645473e+00 -9.50339139e-01 -8.81695926e-01
2.19920799e-01 4.94313180e-01 -6.78135097e-01 6.17705822e-01
-5.79419374e-01 1.80838272e-01 -1.69976979e-01 1.04715991e+00
2.92152017e-01 -6.43150806e-01 1.18022384e-02 -6.08470619e-01
3.00512850e-01 -2.12574080e-02 -8.51549923e-01 2.02831793e+00
-1.89278468e-01 1.07260478e+00 2.09315181e-01 -9.67726290e-01
3.58091742e-01 6.93383753e-01 6.43963754e-01 -9.57815647e-01
-5.02673686e-02 -1.65635407e-01 -6.46991193e-01 -1.07270885e+00
7.97296464e-01 4.42759782e-01 -1.49269819e-01 1.75699294e-01
4.11825687e-01 3.26697707e-01 2.98917174e-01 8.12886357e-01
1.11230505e+00 3.18892926e-01 -2.08895847e-01 2.27333501e-01
5.49364328e-01 -2.36461982e-01 2.06625640e-01 7.58728623e-01
-3.11068892e-01 9.16121244e-01 7.20528841e-01 -4.86855268e-01
-1.06458807e+00 -9.76574302e-01 4.34670657e-01 1.65123475e+00
3.13837707e-01 -7.81035006e-01 -5.91135442e-01 -3.28103900e-01
-1.83872163e-01 1.67819515e-01 -5.69194496e-01 -3.52687836e-02
-8.33943248e-01 -2.42936581e-01 7.35423088e-01 3.68766069e-01
5.21655381e-01 -9.16719139e-01 -1.90802291e-01 7.27487579e-02
-9.66831028e-01 -1.77549386e+00 -1.10488534e+00 -6.79287195e-01
-5.22402823e-01 -1.11689222e+00 -7.74834931e-01 -9.69474733e-01
2.39370689e-01 5.48520625e-01 1.51794720e+00 2.10985467e-01
-4.03078273e-02 1.09830189e+00 -5.22661567e-01 4.85570431e-01
-3.22698534e-01 -4.53527778e-01 1.63597427e-03 4.71049607e-01
1.23590156e-01 -2.68617958e-01 -5.93388379e-01 6.57785654e-01
-1.15113664e+00 2.82442212e-01 2.41428480e-01 9.54707086e-01
7.05570877e-01 -6.02266490e-01 2.11931705e-01 -3.11617941e-01
1.24899410e-01 -8.69651556e-01 -2.62968421e-01 6.50562763e-01
4.38828230e-01 -2.74072319e-01 2.89555967e-01 -6.18979752e-01
-8.45385194e-01 -2.19283074e-01 -6.05672672e-02 -1.19543064e+00
-5.04557043e-03 4.34319079e-01 -1.81447297e-01 1.98299304e-01
1.46697402e-01 5.90736806e-01 3.71498838e-02 -1.14959173e-01
4.40012962e-01 4.19834167e-01 9.45641339e-01 -8.40704739e-01
4.78145570e-01 5.44191599e-01 -2.43231401e-01 -1.04701900e+00
-5.72263539e-01 -5.46778262e-01 -4.21293676e-01 -5.86616695e-01
1.33303535e+00 -1.48314965e+00 -1.04726398e+00 4.81570184e-01
-1.37851739e+00 -4.85546052e-01 8.67285952e-02 2.16800347e-01
-8.18493545e-01 6.65331125e-01 -1.02664506e+00 -4.83255982e-01
-2.45902702e-01 -1.39151216e+00 1.75086284e+00 -1.86705947e-01
4.00909558e-02 -8.05747926e-01 -2.61580497e-01 1.03925550e+00
6.88630939e-02 1.19872287e-01 5.57601154e-01 -2.63843894e-01
-1.08963692e+00 1.77566931e-01 -3.57770056e-01 2.29508057e-01
-3.56678158e-01 2.34335303e-01 -6.83857799e-01 -3.90731186e-01
-2.76365489e-01 -8.21207464e-01 9.89464164e-01 3.79214495e-01
1.33599925e+00 -4.74422693e-01 -9.10065845e-02 7.40409195e-01
1.03256023e+00 1.00229204e-01 9.59306061e-01 -2.70966087e-02
9.28594410e-01 3.55481327e-01 6.48675203e-01 3.69246334e-01
9.03578281e-01 7.89982617e-01 5.81672847e-01 2.14217007e-01
-2.28832260e-01 -4.87554997e-01 8.91500771e-01 1.02784681e+00
5.65780606e-03 -6.19816899e-01 -7.27436304e-01 5.98356903e-01
-2.16968513e+00 -1.52801633e+00 1.23337343e-01 1.70261443e+00
4.42485660e-01 -2.96998948e-01 2.56069392e-01 -4.36597884e-01
7.29692876e-01 5.44288695e-01 -5.05418420e-01 1.89397596e-02
-6.00663304e-01 -4.10727501e-01 5.55275902e-02 6.02720678e-01
-1.10494232e+00 9.25706029e-01 6.22992897e+00 1.01272964e+00
-8.60473096e-01 3.06746721e-01 8.16820860e-01 -3.76147687e-01
-4.68695194e-01 -1.98439002e-01 -4.96543229e-01 6.88607454e-01
1.14923739e+00 1.76486790e-01 6.32633388e-01 3.67801398e-01
3.03936362e-01 -4.27808985e-02 -1.22656822e+00 1.47180474e+00
5.15714049e-01 -1.79349840e+00 5.88537931e-01 -3.88233006e-01
6.55903101e-01 3.03790927e-01 1.70293227e-02 4.90370750e-01
-1.52399316e-01 -1.07338619e+00 1.06901622e+00 8.02957356e-01
9.57197249e-01 -2.05739498e-01 3.57691318e-01 1.02376699e-01
-1.53953993e+00 -1.22115396e-01 -1.15583405e-01 2.73431033e-01
6.23970091e-01 -4.93013449e-02 -2.08971485e-01 5.49114764e-01
9.23861086e-01 1.19458807e+00 -4.77817476e-01 4.65342104e-01
4.94586527e-01 5.73099852e-01 -1.97878018e-01 3.52577686e-01
3.50175649e-01 -2.42267549e-02 6.06495500e-01 1.47713506e+00
1.67003289e-01 1.48048311e-01 1.90298572e-01 4.08678561e-01
-3.91236454e-01 -2.27670923e-01 -4.35308754e-01 -2.74274647e-01
3.77141416e-01 9.18314397e-01 -2.39146829e-01 -6.76010311e-01
-7.91562855e-01 1.26479995e+00 -4.78525274e-02 6.07119977e-01
-1.08956504e+00 3.68591666e-01 8.23339880e-01 1.65437207e-01
3.38714004e-01 -2.86737531e-01 7.14653730e-01 -1.78551018e+00
-1.26321362e-02 -1.25580513e+00 7.45503902e-01 -1.31746411e+00
-1.18454075e+00 5.73430121e-01 1.31553471e-01 -1.20611823e+00
-4.21772093e-01 -4.93393809e-01 -2.97252148e-01 2.77761668e-01
-1.49448621e+00 -1.43313694e+00 -4.18311298e-01 1.27224946e+00
1.14679933e+00 -4.01008129e-01 3.86257708e-01 8.60643983e-01
-3.79767716e-01 5.70140898e-01 -1.74915977e-02 2.61003405e-01
9.77716029e-01 -8.10389102e-01 3.15568566e-01 6.98667347e-01
1.90946117e-01 3.26229006e-01 4.62474555e-01 -4.57238406e-01
-2.19001842e+00 -1.21655023e+00 7.08853364e-01 -6.65814996e-01
8.29057097e-01 -3.76325846e-01 -6.92053139e-01 9.46030796e-01
3.94711792e-01 1.60928652e-01 3.76762986e-01 -6.17211640e-01
-4.74467427e-01 6.95509762e-02 -5.90376794e-01 6.03225112e-01
1.03665030e+00 -1.21951830e+00 -4.05265719e-01 6.71583533e-01
1.29672873e+00 -6.42287314e-01 -1.03390157e+00 2.66611814e-01
7.33365595e-01 -8.28690529e-01 1.28129518e+00 -8.86043727e-01
9.69108164e-01 -3.27569664e-01 -8.83199453e-01 -4.61112767e-01
4.57304604e-02 -1.02142239e+00 -6.90203905e-01 1.18745744e+00
4.27054539e-02 2.14753032e-01 6.14419937e-01 4.09794599e-01
-1.54463187e-01 -6.14158928e-01 -9.50919747e-01 -3.52069199e-01
-3.82862300e-01 -6.93122327e-01 2.98184395e-01 8.05398107e-01
-3.00398350e-01 3.30873758e-01 -1.20140135e+00 1.29022852e-01
4.06936944e-01 -1.09137379e-01 8.59060287e-01 -3.55674237e-01
-5.12070715e-01 -1.89988419e-01 -5.34009457e-01 -1.77590084e+00
4.08994824e-01 -5.89663625e-01 -1.58092126e-01 -1.30267143e+00
4.39786434e-01 6.93260491e-01 6.79855328e-03 1.62985653e-01
-1.20024160e-02 3.84750515e-01 4.83569920e-01 3.74078363e-01
-1.56522703e+00 6.32133603e-01 1.21388459e+00 -4.01231587e-01
2.52855778e-01 -5.33939302e-01 -1.74525186e-01 4.71598804e-01
-3.52764130e-02 1.11419261e-02 -3.65987122e-01 -1.10050583e+00
4.59380925e-01 9.51694012e-01 8.04423988e-01 -6.76304400e-01
3.76242816e-01 -1.08292043e-01 1.39879882e-01 -8.63616824e-01
8.60947013e-01 -6.88765824e-01 3.48130576e-02 4.67138328e-02
-6.54318154e-01 5.48117280e-01 1.20166473e-01 8.48563671e-01
-6.82671487e-01 4.53712702e-01 3.08535010e-01 -6.86956197e-02
-1.16919088e+00 7.91665494e-01 -4.00901556e-01 3.70657176e-01
6.81609511e-01 -7.40389451e-02 -6.72771692e-01 -1.05677927e+00
-7.83743322e-01 7.82282770e-01 2.17413813e-01 9.29657578e-01
9.82633412e-01 -1.58124542e+00 -8.01525533e-01 -1.80497676e-01
1.99423909e-01 -2.50563622e-01 7.63589442e-01 9.26444173e-01
-5.00283480e-01 4.18136537e-01 5.42840622e-02 -1.04195559e+00
-1.40132844e+00 5.77944696e-01 2.52392262e-01 -5.85271940e-02
-4.03307915e-01 9.74327624e-01 4.42124099e-01 -1.19247168e-01
4.78459388e-01 -2.14231417e-01 -5.11945970e-02 1.26047418e-01
9.12358224e-01 2.56425709e-01 -3.65152955e-01 -1.17840886e+00
-1.77220792e-01 6.41538501e-01 5.38615286e-02 -1.11316770e-01
1.04812038e+00 -6.84790671e-01 -2.95882195e-01 4.65827793e-01
1.64393544e+00 -5.69376230e-01 -1.45104873e+00 -5.18293560e-01
-3.42498034e-01 -4.31670398e-01 -2.22138494e-01 -4.34344530e-01
-9.46488559e-01 9.89561260e-01 2.92264909e-01 -2.12852299e-01
1.15271533e+00 1.78633168e-01 1.21786284e+00 5.71468830e-01
1.18146524e-01 -8.91582906e-01 6.81003034e-01 7.09986389e-01
9.98151600e-01 -1.41773403e+00 -5.05898476e-01 5.64431213e-02
-9.19153631e-01 1.03053713e+00 4.49582249e-01 3.64059471e-02
4.89538133e-01 1.87245905e-02 -2.06300154e-01 -1.15591817e-01
-1.42482698e+00 4.81624678e-02 5.63179612e-01 2.32299253e-01
1.43924803e-01 -2.39118353e-01 4.21395540e-01 5.01420498e-01
2.14513332e-01 8.34698156e-02 3.11577320e-01 7.30604947e-01
-2.05815822e-01 -6.92159832e-01 -4.34987932e-01 2.69224852e-01
-4.83675927e-01 -3.24520499e-01 -1.41639978e-01 5.58002412e-01
-2.00494528e-01 1.03467047e+00 2.97058761e-01 -6.68097377e-01
-6.64637461e-02 3.14598083e-02 4.82447028e-01 -1.29415631e-01
-3.60037506e-01 2.75072157e-01 1.88783720e-01 -1.24522471e+00
-7.05893397e-01 -6.23206735e-01 -9.81198847e-01 -5.09990633e-01
3.56616735e-01 4.70134839e-02 3.14309567e-01 1.13814449e+00
4.76981729e-01 3.85899156e-01 4.02087450e-01 -1.06095111e+00
-7.05757886e-02 -7.14966893e-01 -3.67655270e-02 7.33190179e-01
7.21841395e-01 -2.85585761e-01 -1.07080400e-01 8.13076913e-01] | [10.32602596282959, 0.8692918419837952] |
e3528c02-9a0c-4e58-8d7f-3a54337928fa | alp-action-aware-embodied-learning-for | 2306.10190 | null | https://arxiv.org/abs/2306.10190v1 | https://arxiv.org/pdf/2306.10190v1.pdf | ALP: Action-Aware Embodied Learning for Perception | Current methods in training and benchmarking vision models exhibit an over-reliance on passive, curated datasets. Although models trained on these datasets have shown strong performance in a wide variety of tasks such as classification, detection, and segmentation, they fundamentally are unable to generalize to an ever-evolving world due to constant out-of-distribution shifts of input data. Therefore, instead of training on fixed datasets, can we approach learning in a more human-centric and adaptive manner? In this paper, we introduce \textbf{A}ction-aware Embodied \textbf{L}earning for \textbf{P}erception (ALP), an embodied learning framework that incorporates action information into representation learning through a combination of optimizing policy gradients through reinforcement learning and inverse dynamics prediction objectives. Our method actively explores complex 3D environments to both learn generalizable task-agnostic representations as well as collect downstream training data. We show that ALP outperforms existing baselines in object detection and semantic segmentation. In addition, we show that by training on actively collected data more relevant to the environment and task, our method generalizes more robustly to downstream tasks compared to models pre-trained on fixed datasets such as ImageNet. | ['Pieter Abbeel', 'aditi raghunathan', 'Wilson Yan', 'Anthony Han', 'Xinran Liang'] | 2023-06-16 | null | null | null | null | ['benchmarking', 'benchmarking'] | ['miscellaneous', 'robots'] | [ 3.73007357e-01 1.48501083e-01 -5.94799686e-03 -5.41821241e-01
-6.92615807e-01 -6.79382205e-01 7.70180047e-01 -4.46682535e-02
-8.11140478e-01 4.13150460e-01 1.55896664e-01 -2.46006712e-01
-1.04089051e-01 -5.45361578e-01 -1.29360271e+00 -5.44861972e-01
-5.89901358e-02 6.03293836e-01 2.42129907e-01 -3.30051869e-01
2.18162000e-01 4.35383409e-01 -1.58321142e+00 7.13310242e-02
7.13919640e-01 7.97495484e-01 6.73850954e-01 8.04290473e-01
2.86582261e-01 7.48818636e-01 -4.15254742e-01 7.78963119e-02
4.96652693e-01 -2.67548293e-01 -8.73877168e-01 1.50462463e-01
5.83336711e-01 -4.71823722e-01 -2.89126277e-01 8.08641672e-01
5.07865012e-01 4.52489763e-01 6.47922337e-01 -1.05419362e+00
-1.09399021e+00 4.52389300e-01 -4.74387825e-01 4.50399876e-01
-1.18209748e-02 7.84727871e-01 8.20110679e-01 -6.57039106e-01
8.14873159e-01 1.36591673e+00 6.21465921e-01 1.06556034e+00
-1.54917347e+00 -5.13727427e-01 8.60982120e-01 -1.57863110e-01
-6.75780773e-01 -4.22998309e-01 5.30077040e-01 -4.99825686e-01
1.11502767e+00 -1.12684863e-02 7.71036685e-01 1.66278255e+00
-4.83424962e-02 1.16386652e+00 1.22081101e+00 -9.79324281e-02
3.22699726e-01 -8.13965723e-02 -3.62203680e-02 6.96206510e-01
1.93092879e-02 5.29142380e-01 -5.02866983e-01 2.89621025e-01
7.64173210e-01 -1.47887215e-01 -1.34159401e-01 -6.53909683e-01
-1.33949387e+00 6.81439400e-01 7.95639277e-01 -1.71556443e-01
-3.91225070e-01 8.54773819e-01 2.41549447e-01 -6.19848445e-02
3.88153911e-01 7.92635381e-01 -8.61201108e-01 -1.24184318e-01
-6.09035134e-01 5.36482513e-01 5.86062372e-01 9.08510268e-01
7.26823509e-01 1.94304138e-01 -3.30994159e-01 5.61424375e-01
3.71221632e-01 6.30299747e-01 5.36002934e-01 -1.41308308e+00
3.39625299e-01 3.12044561e-01 1.14442602e-01 -4.30902243e-01
-4.91181076e-01 -4.82350200e-01 -1.01874538e-01 5.85629046e-01
5.44168711e-01 -3.99856120e-01 -1.54222524e+00 2.17301345e+00
2.30118170e-01 1.68105334e-01 1.14189878e-01 1.04200685e+00
6.57349348e-01 4.67529505e-01 5.06429732e-01 4.43892360e-01
9.23224866e-01 -1.19237649e+00 4.44584489e-02 -6.14245415e-01
3.78617704e-01 -1.69727236e-01 1.31058896e+00 2.32126772e-01
-1.01646256e+00 -5.99793315e-01 -7.98033237e-01 -2.73566186e-01
-4.29505259e-01 -2.46344164e-01 7.88901448e-01 4.21209812e-01
-1.43017435e+00 5.71454406e-01 -1.04801023e+00 -6.48190856e-01
9.63516831e-01 4.38172191e-01 -1.51643068e-01 1.29154205e-01
-6.31766856e-01 9.58738089e-01 3.19282800e-01 7.66969426e-03
-1.62456310e+00 -8.49815190e-01 -7.63508618e-01 -2.14321539e-01
3.82283807e-01 -9.83667433e-01 1.51909494e+00 -1.25396454e+00
-1.49529636e+00 9.67095494e-01 1.88764200e-01 -7.05738783e-01
6.45860732e-01 -4.73698080e-01 2.14764535e-01 2.55645588e-02
1.96655288e-01 1.33829284e+00 9.15010989e-01 -1.55996418e+00
-5.90596437e-01 -5.08445621e-01 3.88348460e-01 4.84873652e-01
-1.45677611e-01 -2.95776904e-01 -5.16877949e-01 -5.13067782e-01
-1.45477578e-01 -9.93761957e-01 -5.30166447e-01 3.38097841e-01
-1.44735605e-01 -5.28185777e-02 1.02159619e+00 -4.90867555e-01
4.08033222e-01 -1.89605665e+00 3.14030915e-01 -1.98604211e-01
1.72414944e-01 3.31804216e-01 -3.89975727e-01 1.17292017e-01
3.04715276e-01 1.54738262e-01 -2.59208232e-01 -4.01271731e-01
5.89884445e-02 3.65587950e-01 -3.78993928e-01 3.16185236e-01
4.45837080e-01 1.31718433e+00 -1.34485924e+00 -2.32399508e-01
2.54254788e-01 2.45496616e-01 -6.76540732e-01 5.71837462e-02
-9.11878347e-01 8.63923132e-01 -7.05192208e-01 6.64429724e-01
1.36828169e-01 -2.47324035e-01 -2.20310509e-01 1.81589425e-01
-2.28483081e-02 -1.31160662e-01 -4.74650174e-01 2.04913974e+00
-6.26056075e-01 6.41990840e-01 1.76710367e-01 -1.23929060e+00
6.50689125e-01 -1.60201371e-01 4.79168087e-01 -1.01001632e+00
1.37979552e-01 -1.54722631e-01 1.82041805e-02 -7.65142977e-01
3.18019181e-01 1.41287878e-01 4.33739237e-02 4.88389462e-01
2.83671170e-01 -4.31834579e-01 -5.64346164e-02 1.75970659e-01
1.15401185e+00 7.92142630e-01 -2.93101966e-01 -2.51017690e-01
-6.62228442e-04 3.31430376e-01 4.72176403e-01 1.03678858e+00
-5.11255324e-01 5.01953244e-01 2.22525299e-01 -2.81282276e-01
-1.05507183e+00 -1.15327930e+00 -1.72118038e-01 1.52219284e+00
4.10022974e-01 1.51708812e-01 -8.64470363e-01 -8.18361342e-01
3.16882074e-01 8.77391636e-01 -9.76311266e-01 -5.22176743e-01
-4.79197562e-01 -8.00815225e-01 5.79784572e-01 8.73572350e-01
7.60791361e-01 -1.40477896e+00 -1.09958255e+00 2.65390098e-01
3.48917693e-01 -9.65007424e-01 -2.67231643e-01 4.66316342e-01
-9.15770173e-01 -1.00244272e+00 -6.52958155e-01 -6.50923371e-01
6.23661160e-01 2.76680499e-01 1.21415877e+00 -1.11664079e-01
-4.38055754e-01 1.09969485e+00 -2.12468103e-01 -8.72804105e-01
-2.35084534e-01 4.12385389e-02 -1.72825560e-01 -1.84516162e-01
3.07984024e-01 -3.59741956e-01 -9.42063272e-01 2.18145117e-01
-6.50021672e-01 -1.26157384e-02 7.08918333e-01 9.03247118e-01
6.22288942e-01 -6.26647234e-01 6.24873579e-01 -8.23797464e-01
6.39055967e-01 -5.17199099e-01 -5.09496152e-01 1.70125097e-01
-7.39783108e-01 2.22632840e-01 3.17081779e-01 -6.02537096e-01
-1.33172607e+00 1.85722634e-01 -2.05867775e-02 -5.93816876e-01
-3.45545441e-01 1.00728022e-02 8.09193254e-02 -6.48737177e-02
1.01986504e+00 1.31677672e-01 2.25022789e-02 -3.10695767e-01
6.96954727e-01 2.32863769e-01 5.59943736e-01 -1.07708049e+00
7.27237105e-01 5.56764901e-01 -2.33039498e-01 -5.78574717e-01
-1.00675595e+00 -2.12019399e-01 -4.90840435e-01 -2.59874880e-01
1.24952519e+00 -9.59316373e-01 -7.70813882e-01 4.93596613e-01
-7.64255047e-01 -1.19673049e+00 -6.66923285e-01 3.34604174e-01
-1.01245594e+00 -2.28429615e-01 -2.66197473e-01 -7.59426057e-01
-6.38098419e-02 -1.23916352e+00 1.49487507e+00 3.38913649e-01
-1.08118542e-01 -1.14475954e+00 -7.32271373e-03 4.97484893e-01
4.47893023e-01 4.17124152e-01 7.28218019e-01 -4.87917483e-01
-8.41183424e-01 1.57021567e-01 -2.49612480e-01 4.73154664e-01
-1.11833140e-01 -8.11325833e-02 -1.22236669e+00 -3.19157809e-01
-2.80284435e-01 -9.56633151e-01 1.25735641e+00 4.11489040e-01
1.63235259e+00 -8.56041759e-02 -4.60830450e-01 9.59321320e-01
1.11479449e+00 1.18018284e-01 1.82632580e-01 5.48099637e-01
5.77778280e-01 6.21459484e-01 5.12415588e-01 7.81310946e-02
5.48205495e-01 3.72525990e-01 9.38782036e-01 -1.19610980e-01
-2.29636997e-01 -5.80273032e-01 2.62023807e-01 -8.80273879e-02
-3.21408153e-01 -2.58027732e-01 -9.81782019e-01 6.83865905e-01
-2.02247286e+00 -9.35113549e-01 3.63892794e-01 1.91544223e+00
7.32116699e-01 1.67913273e-01 1.72355369e-01 -6.02660537e-01
4.50773209e-01 1.09808415e-01 -1.48142791e+00 -2.48066559e-01
-6.20327033e-02 2.95244426e-01 6.62519157e-01 1.30366564e-01
-1.10359120e+00 1.20803690e+00 6.62769938e+00 2.49168664e-01
-1.28671360e+00 2.10665777e-01 8.58236074e-01 -3.00200909e-01
-3.60818774e-01 -7.74486288e-02 -5.72287023e-01 3.09571683e-01
7.32727110e-01 1.67349000e-02 5.34701586e-01 9.38803673e-01
2.04213545e-01 -8.13813731e-02 -1.31051338e+00 7.32347310e-01
-8.58161319e-03 -1.35634387e+00 -4.49262410e-02 1.63610503e-01
8.31685483e-01 5.36649406e-01 6.01909816e-01 6.63521171e-01
9.12325203e-01 -1.21072829e+00 1.10536754e+00 6.13436699e-01
4.52426553e-01 -1.97315931e-01 1.20750099e-01 4.97105718e-01
-6.18622422e-01 -4.54199463e-01 -1.67056724e-01 2.72532310e-02
-1.24135137e-01 -3.01422309e-02 -7.29547620e-01 6.07239222e-03
1.05623615e+00 8.76751304e-01 -6.79233849e-01 8.39585185e-01
-2.22745135e-01 5.30520976e-01 -1.79521665e-01 -1.32854119e-01
5.13122082e-01 -2.49359757e-02 4.60648447e-01 1.02917302e+00
1.81724932e-02 1.45111457e-01 3.52016866e-01 1.11469698e+00
-1.94617391e-01 -3.32894266e-01 -7.16385722e-01 -1.18023746e-01
2.38131359e-01 1.00301421e+00 -6.59682095e-01 -1.38455376e-01
-1.64095566e-01 8.87100101e-01 5.65739810e-01 8.36381733e-01
-7.69971669e-01 2.19417408e-01 7.95878470e-01 -3.09029240e-02
3.46965194e-01 -3.70407403e-01 -1.88768551e-01 -1.11851013e+00
-2.58451760e-01 -7.35267878e-01 1.46935508e-01 -9.05013382e-01
-1.33025014e+00 3.32141608e-01 1.21979691e-01 -6.69492126e-01
-2.05300108e-01 -9.30129528e-01 -5.55095613e-01 7.07436562e-01
-1.56828213e+00 -1.39415491e+00 -3.42182755e-01 7.64460981e-01
8.18398356e-01 -7.96481520e-02 4.78569776e-01 -2.89573759e-01
-5.31059623e-01 3.00610125e-01 1.26740197e-02 8.69269967e-02
4.91675168e-01 -1.55333626e+00 5.84456265e-01 6.66665912e-01
2.94796765e-01 4.62488621e-01 6.63736999e-01 -4.97446865e-01
-1.45853674e+00 -1.22469819e+00 -2.61104375e-01 -8.68681490e-01
6.27559781e-01 -5.16753495e-01 -7.06338108e-01 1.01386786e+00
7.74930343e-02 3.93388957e-01 1.68886006e-01 1.33867055e-01
-4.43419486e-01 -8.66738185e-02 -1.16659427e+00 6.25655234e-01
1.67936802e+00 -4.54375982e-01 -5.74912488e-01 5.41701257e-01
9.13373888e-01 -6.09629214e-01 -5.60406923e-01 4.27251875e-01
3.86698604e-01 -8.51911068e-01 1.18951344e+00 -1.01328588e+00
3.39155495e-01 -3.04489899e-02 -1.72962010e-01 -1.36019874e+00
-2.35282898e-01 -4.47386950e-01 -2.77217448e-01 9.37494814e-01
4.37750369e-01 -6.07761919e-01 8.78905654e-01 7.12000728e-01
-2.71690100e-01 -8.05725753e-01 -6.32531583e-01 -7.04526305e-01
4.00880396e-01 -4.85073417e-01 4.07926470e-01 6.23393714e-01
-5.54004192e-01 1.80143714e-01 2.96996683e-02 1.07653551e-01
7.49770164e-01 3.69201638e-02 8.61706972e-01 -1.06711543e+00
-4.58860189e-01 -6.21505499e-01 -2.83304900e-01 -1.18869317e+00
3.46611470e-01 -8.97983432e-01 5.39099514e-01 -1.59207475e+00
8.56868401e-02 -8.14204156e-01 -3.82330567e-01 5.46789169e-01
-1.18382670e-01 5.16892225e-02 3.01067799e-01 8.79571140e-02
-7.79706955e-01 7.03373075e-01 1.50638032e+00 -2.32026011e-01
-1.37572721e-01 -2.86353640e-02 -8.35194647e-01 7.53521562e-01
6.68625236e-01 -2.24242270e-01 -7.45406210e-01 -9.66137886e-01
-1.13860480e-02 -3.11540812e-01 8.51645768e-01 -9.73559618e-01
1.23095177e-01 -4.06310439e-01 7.28468597e-01 -1.14853121e-01
4.80611533e-01 -5.10718465e-01 -4.32201922e-01 3.20137858e-01
-5.98790526e-01 -9.49143916e-02 4.04601008e-01 9.93087411e-01
2.54621416e-01 -6.42726570e-02 6.16836309e-01 -5.94053566e-01
-1.29702234e+00 5.88060439e-01 -1.98986456e-01 4.24405903e-01
1.25331330e+00 -4.25374150e-01 -5.40376544e-01 -2.07068413e-01
-8.22807193e-01 6.09190047e-01 5.90408683e-01 7.09625781e-01
4.99876529e-01 -7.36921847e-01 -4.95826453e-01 -2.05059163e-03
2.52473384e-01 4.28334653e-01 6.89162537e-02 4.05416459e-01
-3.70027184e-01 7.57612213e-02 -2.76454836e-01 -9.16857600e-01
-5.95423639e-01 5.21886766e-01 6.80120051e-01 1.43788859e-01
-7.60174274e-01 1.15084267e+00 5.58065653e-01 -7.57103682e-01
3.50321323e-01 -2.80757487e-01 6.88368306e-02 -2.55919904e-01
1.04280144e-01 5.15233055e-02 -2.05037460e-01 -4.07063186e-01
-1.49893537e-01 4.73034352e-01 -1.15825035e-01 -2.66826540e-01
1.51634145e+00 -1.14455432e-01 4.68024522e-01 3.52541268e-01
8.32497954e-01 -6.61444008e-01 -2.39641547e+00 1.46672120e-02
1.09008513e-01 -2.82315135e-01 1.45543113e-01 -1.13271010e+00
-1.09066749e+00 9.15960431e-01 6.98273063e-01 -1.78408995e-01
9.24878180e-01 3.45734954e-01 5.32870591e-01 6.80655003e-01
5.12301266e-01 -1.42058206e+00 7.17543006e-01 4.07350093e-01
9.20502782e-01 -1.51180232e+00 -2.97593474e-01 2.31303141e-01
-8.19440663e-01 7.69467533e-01 1.14309490e+00 -3.43517393e-01
5.39827645e-01 1.09169051e-01 1.41253695e-01 -1.50430098e-01
-8.78183603e-01 -5.00417531e-01 5.20214625e-02 1.15032852e+00
3.55528146e-02 -2.38950387e-01 3.98168743e-01 1.67676255e-01
-7.46877268e-02 -6.75652251e-02 3.20115983e-01 1.04088199e+00
-4.83461469e-01 -8.68083179e-01 -1.58913031e-01 4.50927705e-01
-7.53499866e-02 2.36656010e-01 -3.59185249e-01 1.05379701e+00
3.94927233e-01 6.33496046e-01 1.02023862e-01 -1.34071112e-01
3.44299346e-01 -1.35646194e-01 7.75324285e-01 -9.27850306e-01
-5.42986214e-01 -3.16748828e-01 -8.44435021e-02 -7.70940244e-01
-5.64102590e-01 -1.01201975e+00 -1.40379846e+00 1.45545125e-01
3.00223827e-02 -4.75366861e-01 8.55373204e-01 9.78153884e-01
3.98218781e-01 7.97099471e-01 2.18623087e-01 -1.29412568e+00
-7.92461932e-01 -6.44643009e-01 -2.42163718e-01 6.12672627e-01
5.21287739e-01 -7.35871315e-01 -6.59623817e-02 1.65894672e-01] | [4.430741310119629, 0.7804831862449646] |
0f0d4fd4-cff5-41b8-8433-7d46c0fc8346 | attributable-and-scalable-opinion | 2305.11603 | null | https://arxiv.org/abs/2305.11603v1 | https://arxiv.org/pdf/2305.11603v1.pdf | Attributable and Scalable Opinion Summarization | We propose a method for unsupervised opinion summarization that encodes sentences from customer reviews into a hierarchical discrete latent space, then identifies common opinions based on the frequency of their encodings. We are able to generate both abstractive summaries by decoding these frequent encodings, and extractive summaries by selecting the sentences assigned to the same frequent encodings. Our method is attributable, because the model identifies sentences used to generate the summary as part of the summarization process. It scales easily to many hundreds of input reviews, because aggregation is performed in the latent space rather than over long sequences of tokens. We also demonstrate that our appraoch enables a degree of control, generating aspect-specific summaries by restricting the model to parts of the encoding space that correspond to desired aspects (e.g., location or food). Automatic and human evaluation on two datasets from different domains demonstrates that our method generates summaries that are more informative than prior work and better grounded in the input reviews. | ['Mirella Lapata', 'Hao Tang', 'Tom Hosking'] | 2023-05-19 | null | null | null | null | ['unsupervised-opinion-summarization'] | ['natural-language-processing'] | [ 5.98597765e-01 7.29449868e-01 -2.91783929e-01 -5.85214317e-01
-1.14932120e+00 -1.03662968e+00 8.43433142e-01 8.88518333e-01
-2.08752882e-02 8.70811641e-01 1.44525719e+00 4.52919193e-02
1.78784743e-01 -9.20949996e-01 -4.78010356e-01 -4.77112651e-01
2.45115012e-01 8.21385503e-01 -1.31878093e-01 -1.82424068e-01
6.68950140e-01 5.69373220e-02 -1.70077562e+00 7.70415545e-01
1.22528481e+00 5.18110812e-01 -8.20978805e-02 8.44162226e-01
-5.52262306e-01 5.91570735e-01 -1.25302279e+00 -5.05889118e-01
-2.18555152e-01 -7.85657763e-01 -6.62559509e-01 4.68857825e-01
3.71551275e-01 -2.70972848e-01 6.31542921e-01 8.45569193e-01
2.44557634e-01 1.45978972e-01 1.19659793e+00 -6.11298144e-01
-7.34057188e-01 1.22401047e+00 -2.84323961e-01 -2.13472277e-01
8.13484788e-01 -2.61315197e-01 1.69749427e+00 -6.48982406e-01
7.75123656e-01 1.23084044e+00 4.01598155e-01 2.87806958e-01
-1.33293521e+00 -7.27897435e-02 4.86211985e-01 -5.25718927e-01
-5.04532039e-01 -5.99585235e-01 6.07249916e-01 -3.34499031e-01
1.20887327e+00 6.89414084e-01 8.33012104e-01 7.95890510e-01
6.09645009e-01 9.98704851e-01 5.49326837e-01 -4.68586147e-01
7.22868979e-01 2.93707997e-01 5.06636679e-01 2.11749807e-01
8.44270408e-01 -1.03397131e+00 -8.72444928e-01 -5.90940654e-01
1.06093138e-01 4.87487614e-02 -4.14508581e-02 -1.66061416e-01
-1.17691481e+00 1.06015742e+00 -2.12610692e-01 1.94938242e-01
-9.84878778e-01 -6.36818856e-02 5.37826836e-01 -3.48948501e-02
1.04789162e+00 1.02346432e+00 -4.20444697e-01 -2.64943182e-01
-1.29571521e+00 4.10191983e-01 1.36536872e+00 1.11313820e+00
7.13344276e-01 -1.14642300e-01 -5.26192009e-01 5.92296243e-01
1.17557727e-01 4.73848015e-01 5.79548478e-01 -9.76302862e-01
5.14112234e-01 8.51058066e-01 4.51424003e-01 -1.00128937e+00
-1.21094016e-02 -3.70256782e-01 -5.84136188e-01 -4.65265304e-01
-4.74046379e-01 -3.64626646e-01 -8.75110686e-01 1.28510594e+00
-9.98315886e-02 -7.16827929e-01 5.30038297e-01 1.58367023e-01
8.69991541e-01 1.02206194e+00 9.13625732e-02 -6.28138185e-01
1.47319245e+00 -8.50814044e-01 -1.03954363e+00 -3.39535058e-01
6.26174450e-01 -5.59418380e-01 9.01652694e-01 4.34749305e-01
-1.46716464e+00 -3.46370280e-01 -1.13372564e+00 -1.54252812e-01
-2.98338383e-01 2.25387812e-01 6.66948378e-01 3.96149755e-01
-1.30632055e+00 5.51223934e-01 -5.85571229e-01 -4.83397722e-01
9.47192591e-03 1.38899729e-01 -9.18912739e-02 3.75009835e-01
-1.00795972e+00 5.54352343e-01 3.40818763e-01 -3.55615914e-01
-3.42245311e-01 -2.28954241e-01 -1.11666775e+00 3.89408529e-01
5.36099561e-02 -1.03907216e+00 1.38019729e+00 -9.90674615e-01
-1.42479169e+00 5.67274988e-01 -7.14220047e-01 -5.25431454e-01
-4.25876258e-03 -3.74280065e-01 -4.43428792e-02 4.27846789e-01
7.58514583e-01 7.09811509e-01 7.54442751e-01 -1.32849443e+00
-8.52135539e-01 -1.02060705e-01 9.09506530e-02 4.34237242e-01
-5.30409157e-01 -2.80368142e-02 9.94200166e-03 -5.79508305e-01
1.16130672e-01 -6.12736464e-01 -3.74896199e-01 -9.66908038e-01
-7.62442648e-01 -4.64549392e-01 2.84590304e-01 -6.93985522e-01
1.49516940e+00 -1.72979975e+00 5.15572309e-01 -5.46473404e-03
2.79372960e-01 -2.86503285e-01 -1.28091518e-02 1.00012577e+00
2.02818543e-01 5.45951784e-01 -3.65328133e-01 -7.07410991e-01
3.37828785e-01 7.17560574e-02 -8.08674812e-01 -1.67282559e-02
7.58593976e-02 8.63281071e-01 -1.17903650e+00 -5.40850043e-01
-2.53757417e-01 -7.27035850e-03 -5.81011057e-01 1.08036801e-01
-5.00595391e-01 -2.25332901e-01 -6.28565252e-01 3.15723538e-01
1.35964483e-01 -1.82474211e-01 2.01296240e-01 -2.60212064e-01
-3.25574219e-01 8.43231380e-01 -6.54361486e-01 1.38664472e+00
-3.11130852e-01 4.73795235e-01 -2.72559792e-01 -5.37008345e-01
1.03798783e+00 4.78283018e-01 3.47212195e-01 -1.52673833e-02
-2.14937106e-01 3.38910259e-02 -6.39286518e-01 -3.54165494e-01
1.35519397e+00 -1.91331461e-01 -8.16147566e-01 1.03468072e+00
-4.62750755e-02 -8.46319735e-01 7.63553739e-01 7.52410591e-01
1.01289570e+00 -1.52283221e-01 7.44072855e-01 -2.70740360e-01
8.38265866e-02 3.72597814e-01 3.37884128e-01 9.97948349e-01
4.43642676e-01 7.07993269e-01 9.36553955e-01 -2.13158458e-01
-1.15360510e+00 -9.71015751e-01 2.36146495e-01 1.00854325e+00
-1.74020991e-01 -9.47881222e-01 -6.72029316e-01 -7.13246882e-01
-1.23432375e-01 1.44304252e+00 -7.71233618e-01 -9.25135314e-02
-2.53756285e-01 -5.81765592e-01 1.34474590e-01 3.95131767e-01
6.72240853e-02 -1.20958269e+00 -8.11341107e-01 4.76813436e-01
-5.09744108e-01 -7.02504992e-01 -5.28565705e-01 2.01140165e-01
-8.12773943e-01 -5.41532516e-01 -5.34212112e-01 -6.08811796e-01
1.01084089e+00 8.82365406e-02 1.44669354e+00 -4.04232353e-01
3.07338804e-01 6.56994343e-01 -8.42709482e-01 -5.91557264e-01
-6.28119648e-01 4.76638794e-01 -6.02709875e-02 -3.19457836e-02
4.76868391e-01 -4.60624486e-01 -2.70708859e-01 -4.72044706e-01
-1.04021120e+00 2.32377633e-01 5.63921332e-01 4.96848345e-01
4.29470569e-01 9.53879058e-02 7.68687189e-01 -1.32046080e+00
1.47427142e+00 -5.18326104e-01 8.78395587e-02 3.15213859e-01
-3.13007742e-01 3.91340256e-01 5.30166328e-01 -1.18954070e-01
-1.07883036e+00 -1.24436885e-01 1.29380137e-01 6.17366314e-01
-1.72584251e-01 9.12289858e-01 9.73120555e-02 1.12328351e+00
6.22313678e-01 4.14357990e-01 -2.25706786e-01 -2.18216375e-01
6.27736926e-01 7.26393640e-01 2.44911373e-01 -4.50175554e-01
3.55602235e-01 3.55673164e-01 -6.38101518e-01 -9.57957029e-01
-1.11721504e+00 -3.80920410e-01 -6.05607748e-01 3.14750243e-03
1.00424910e+00 -8.66556406e-01 -4.05001342e-02 -4.86006662e-02
-1.37584901e+00 2.25448847e-01 -1.02257407e+00 1.28970534e-01
-5.25356889e-01 2.92567462e-01 -6.19214714e-01 -9.36168313e-01
-8.52482796e-01 -8.63442361e-01 1.48535538e+00 2.21963286e-01
-1.21763289e+00 -8.83735001e-01 3.02496493e-01 1.05063029e-01
1.88710541e-01 3.85926276e-01 1.00987756e+00 -9.31968749e-01
-2.03821138e-01 -4.57653970e-01 3.43292236e-01 2.53579676e-01
6.54867768e-01 3.37457448e-01 -4.83674943e-01 -1.51341200e-01
1.36508480e-01 -2.32049569e-01 1.04208720e+00 5.73925793e-01
3.33045483e-01 -1.03792953e+00 -3.40668201e-01 -8.42906162e-02
8.84565771e-01 6.85653985e-02 4.44131494e-01 1.06047221e-01
4.20351624e-01 1.11669183e+00 4.97748882e-01 5.21258533e-01
6.78434789e-01 2.41278931e-01 -3.51504162e-02 6.85173422e-02
2.71248490e-01 -3.14453214e-01 7.10102677e-01 1.27167535e+00
2.46245384e-01 -5.02113581e-01 -5.12751460e-01 9.68810618e-01
-1.89456809e+00 -1.09030020e+00 5.05226664e-02 1.71953571e+00
1.04485536e+00 2.32803062e-01 6.29974157e-02 -7.53292814e-02
5.11630356e-01 4.47221756e-01 -3.66903305e-01 -9.98366177e-01
-7.04443082e-02 1.36352275e-02 1.40196159e-01 7.25035191e-01
-7.07219958e-01 9.21093404e-01 7.23757553e+00 2.92005479e-01
-5.69205463e-01 -3.03743631e-01 7.54264235e-01 -3.30303520e-01
-1.30531323e+00 1.02461800e-01 -9.29648757e-01 2.73618668e-01
9.59814668e-01 -6.23762190e-01 -1.23376481e-01 7.39229798e-01
4.33231562e-01 -2.93087184e-01 -1.16588223e+00 2.43997246e-01
5.18366277e-01 -1.46870601e+00 8.42640460e-01 5.55171482e-02
1.10465264e+00 -5.43936729e-01 -1.00202017e-01 1.09827127e-02
6.82590246e-01 -7.25960433e-01 9.37415898e-01 5.71788490e-01
5.90076029e-01 -8.46984565e-01 8.30283701e-01 5.01200616e-01
-1.01269758e+00 1.61588445e-01 -4.81119990e-01 -2.06445858e-01
4.96662199e-01 8.65336359e-01 -8.82449269e-01 6.34010851e-01
3.17447394e-01 1.11613679e+00 -6.63590848e-01 3.84202391e-01
-5.29487431e-01 7.25385606e-01 -6.81541637e-02 -5.07566869e-01
2.57595927e-01 -4.05448556e-01 6.78424120e-01 1.44546402e+00
5.28219104e-01 1.41856223e-01 1.32689849e-01 6.51061654e-01
4.81919497e-02 3.31349909e-01 -8.17728221e-01 -3.44842076e-01
4.98300731e-01 1.12561917e+00 -8.65953207e-01 -9.52213287e-01
2.75208086e-01 9.55416799e-01 -1.20278157e-01 4.78954822e-01
-3.23314369e-01 -6.76869690e-01 3.25740635e-01 1.53616362e-03
4.21945691e-01 -6.88949525e-02 -5.94045818e-01 -1.35032439e+00
3.22843529e-02 -9.78420138e-01 9.81020033e-02 -8.40096593e-01
-9.63336468e-01 8.07283878e-01 2.95053303e-01 -1.05415821e+00
-8.06121826e-01 1.95420384e-01 -8.30315173e-01 6.83401287e-01
-8.80984545e-01 -8.22694480e-01 1.78687647e-01 -3.11669469e-01
9.17750239e-01 2.16304157e-02 1.17969537e+00 -5.19265115e-01
-2.50041157e-01 1.15256712e-01 -5.77292666e-02 -3.35131317e-01
5.72621226e-01 -1.69944835e+00 7.79650092e-01 8.80506456e-01
1.38220906e-01 1.10910630e+00 1.18564200e+00 -9.22010660e-01
-7.38960326e-01 -8.19788396e-01 1.71261322e+00 -6.69392645e-01
4.54173535e-01 -4.27366942e-01 -5.37763119e-01 7.44275570e-01
8.22470725e-01 -1.26276565e+00 1.06689656e+00 2.34614715e-01
3.22236978e-02 2.32959501e-02 -9.98043418e-01 6.38707578e-01
5.76210320e-01 -3.53829026e-01 -1.18598521e+00 3.73707891e-01
1.17053747e+00 -2.67995941e-03 -5.02192140e-01 -6.59448877e-02
4.65480953e-01 -9.25652325e-01 4.71974224e-01 -4.66878653e-01
9.20036912e-01 -2.18231559e-01 -3.70631181e-02 -1.72050405e+00
-2.89888024e-01 -7.61465251e-01 -1.56429231e-01 1.59389043e+00
6.93007708e-01 -6.27173901e-01 5.48252463e-01 5.23843527e-01
-3.00065458e-01 -6.81160629e-01 -3.70959431e-01 8.45784098e-02
-2.79696196e-01 4.82711792e-02 7.15570450e-01 4.26330656e-01
3.59044015e-01 8.17180336e-01 -1.86619774e-01 -9.82314423e-02
5.79193056e-01 4.68802661e-01 6.53789043e-01 -1.21904206e+00
-1.22609012e-01 -4.62026298e-01 9.52978879e-02 -1.22058988e+00
2.02754401e-02 -5.43644905e-01 3.28431100e-01 -2.58225036e+00
2.62493312e-01 1.07805230e-01 1.52032018e-01 4.54911590e-01
-2.25681648e-01 -2.01445408e-02 -1.49307176e-02 2.66724020e-01
-9.83207345e-01 5.34484148e-01 8.99506092e-01 -3.71998638e-01
-5.69527686e-01 3.88540551e-02 -1.55153501e+00 7.64513075e-01
7.44491875e-01 -5.27162731e-01 -5.91165304e-01 -2.51333863e-01
7.22966313e-01 -4.47658673e-02 -3.45189929e-01 -6.19500220e-01
2.24779338e-01 -5.51140904e-02 3.40168089e-01 -1.15003181e+00
1.18931033e-01 -3.22837830e-01 1.16802767e-01 2.34852567e-01
-1.03675508e+00 3.35671276e-01 -7.09897503e-02 3.89727354e-01
-3.81793469e-01 -3.94171536e-01 1.58349592e-02 -3.07777733e-01
-1.83826499e-02 -3.88713330e-01 -8.59629631e-01 -4.53961827e-02
6.25575185e-01 -3.24846566e-01 -2.51584649e-01 -9.16581869e-01
-6.29807532e-01 3.24942619e-01 6.23435557e-01 1.46505654e-01
5.79586685e-01 -1.08204758e+00 -9.95844722e-01 2.03038659e-02
1.78674564e-01 2.41732046e-01 -1.54939666e-01 1.41892344e-01
-4.09786016e-01 5.84092200e-01 1.37994468e-01 -3.74721378e-01
-1.06546450e+00 5.25508039e-02 -3.24016988e-01 -7.13310301e-01
-6.12913609e-01 6.38597548e-01 1.87647995e-02 -2.28354320e-01
3.07615120e-02 -6.93931639e-01 -7.20396280e-01 8.80497515e-01
6.74614608e-01 1.14400752e-01 -1.52470872e-01 -4.36020732e-01
-1.98765412e-01 4.54222620e-01 -3.38074327e-01 -4.25476313e-01
1.63200009e+00 -3.85096967e-01 -6.76248193e-01 8.27262938e-01
8.54007065e-01 5.94777703e-01 -1.08998990e+00 2.42026880e-01
1.32581562e-01 1.14568323e-01 -4.16037053e-01 -6.29182637e-01
-1.85457945e-01 3.20779204e-01 -5.47441721e-01 1.02709866e+00
1.00382817e+00 2.10659280e-01 7.68566847e-01 4.93129075e-01
1.85728520e-01 -1.01187396e+00 2.66798347e-01 5.45070529e-01
9.91467893e-01 -6.07122123e-01 4.81322974e-01 -1.40708283e-01
-1.04525816e+00 1.01332617e+00 -2.15961207e-02 -3.22025061e-01
1.02634124e-01 1.52381167e-01 8.20524991e-02 -3.86821121e-01
-1.11616158e+00 2.08361134e-01 3.58980000e-01 2.60849029e-01
7.09686577e-01 2.67407209e-01 -6.65062606e-01 5.64307988e-01
-7.87863851e-01 -4.93054092e-01 1.01790881e+00 9.62751925e-01
-9.49877739e-01 -1.08237648e+00 -1.05375007e-01 9.77482319e-01
-3.65986437e-01 -3.07263494e-01 -8.95470321e-01 1.61210239e-01
-2.11127564e-01 1.19769764e+00 1.96815848e-01 -1.23510703e-01
3.54301810e-01 2.43529871e-01 3.43913175e-02 -1.54486966e+00
-8.18860531e-01 2.15954900e-01 5.16075015e-01 -1.35576651e-01
-4.48078781e-01 -1.05645275e+00 -1.13268566e+00 4.55804355e-02
-2.15219170e-01 9.43562269e-01 4.48293805e-01 9.35052454e-01
5.91162682e-01 7.81567335e-01 5.14791548e-01 -1.04210079e+00
-4.94860739e-01 -1.06400764e+00 -4.90314305e-01 3.41306210e-01
5.47806203e-01 3.13229747e-02 -5.39398789e-01 3.96280229e-01] | [12.465489387512207, 9.36670970916748] |
49700779-00d9-46d5-9a44-0a48e125b445 | conversational-automated-program-repair | 2301.13246 | null | https://arxiv.org/abs/2301.13246v1 | https://arxiv.org/pdf/2301.13246v1.pdf | Conversational Automated Program Repair | Automated Program Repair (APR) can help developers automatically generate patches for bugs. Due to the impressive performance obtained using Large Pre-Trained Language Models (LLMs) on many code related tasks, researchers have started to directly use LLMs for APR. However, prior approaches simply repeatedly sample the LLM given the same constructed input/prompt created from the original buggy code, which not only leads to generating the same incorrect patches repeatedly but also miss the critical information in testcases. To address these limitations, we propose conversational APR, a new paradigm for program repair that alternates between patch generation and validation in a conversational manner. In conversational APR, we iteratively build the input to the model by combining previously generated patches with validation feedback. As such, we leverage the long-term context window of LLMs to not only avoid generating previously incorrect patches but also incorporate validation feedback to help the model understand the semantic meaning of the program under test. We evaluate 10 different LLM including the newly developed ChatGPT model to demonstrate the improvement of conversational APR over the prior LLM for APR approach. | ['Lingming Zhang', 'Chunqiu Steven Xia'] | 2023-01-30 | null | null | null | null | ['program-repair', 'program-repair'] | ['computer-code', 'reasoning'] | [ 1.92747161e-01 5.36439657e-01 -1.56963378e-01 -2.78096944e-01
-1.14285791e+00 -6.01299107e-01 8.67252201e-02 3.44220340e-01
5.74359596e-01 4.45737123e-01 9.31793004e-02 -6.94728673e-01
5.94002366e-01 -6.26931846e-01 -9.51748550e-01 6.52869493e-02
-2.26429254e-02 -9.54625010e-02 2.92703122e-01 -8.03724900e-02
6.12150311e-01 -4.50052828e-01 -1.50921690e+00 8.90585005e-01
1.34377396e+00 1.27368718e-01 5.29435158e-01 8.81964922e-01
-4.22206074e-01 1.19807827e+00 -1.09907520e+00 -4.44317937e-01
-2.84758240e-01 -7.62785196e-01 -1.02072251e+00 -9.89941582e-02
3.37439805e-01 -7.14601725e-02 3.98569524e-01 1.03807497e+00
1.92368835e-01 -2.39209384e-01 5.88934273e-02 -1.29956877e+00
-8.67476583e-01 1.07011616e+00 -5.82553267e-01 -1.18645608e-01
8.69035542e-01 3.78866762e-01 1.12004614e+00 -8.74340773e-01
5.87818563e-01 1.13072646e+00 9.46442604e-01 9.33306694e-01
-1.30995440e+00 -3.35394174e-01 3.42541456e-01 -4.86728959e-02
-9.87188756e-01 -1.62816644e-01 6.89735591e-01 -6.85876131e-01
1.59760606e+00 3.37686062e-01 5.34184754e-01 1.14318883e+00
4.98366117e-01 6.16180897e-01 7.00864017e-01 -7.68793166e-01
1.75943598e-01 3.09407562e-01 4.13362235e-01 9.05136526e-01
2.84991004e-02 -2.79466838e-01 -2.71980971e-01 -8.91720057e-01
8.96138549e-02 -3.78631130e-02 -3.01470220e-01 2.11796492e-01
-8.64448547e-01 6.41969502e-01 1.73612852e-02 2.37255722e-01
-1.13335149e-02 3.04546267e-01 3.71224850e-01 5.53914309e-01
4.06335354e-01 8.33625793e-01 -6.01339817e-01 -4.07871842e-01
-7.41429448e-01 2.83614218e-01 1.12432981e+00 1.19474065e+00
1.08877695e+00 -8.93090144e-02 -4.74944234e-01 8.93940151e-01
4.68216777e-01 2.63341129e-01 5.29985130e-01 -6.87283754e-01
8.88713002e-01 1.34979713e+00 1.10942736e-01 -8.84625018e-01
1.54471144e-01 -3.63639116e-01 -5.45004755e-02 1.17015466e-01
-1.12147234e-01 -8.34752321e-02 -5.32472134e-01 1.71490586e+00
2.15408191e-01 1.21556878e-01 7.69938305e-02 3.64797622e-01
5.91462493e-01 6.25425458e-01 -1.63089648e-01 8.11815858e-02
9.54024732e-01 -1.38132107e+00 -2.74309248e-01 -6.00817621e-01
1.25508237e+00 -8.56353164e-01 1.66360497e+00 4.19323921e-01
-7.53574073e-01 -4.68254805e-01 -1.14705539e+00 1.95067674e-01
2.42167979e-01 4.40126628e-01 4.99736339e-01 5.37000000e-01
-1.23754096e+00 7.07376957e-01 -9.02407765e-01 -2.11144209e-01
2.78521031e-01 -1.37825534e-01 -3.17198448e-02 -2.81314284e-01
-4.90836263e-01 4.57375556e-01 -3.12947154e-01 1.91934016e-02
-1.10061502e+00 -8.24634731e-01 -9.93890643e-01 7.90945590e-02
3.44466954e-01 -2.71817863e-01 1.75560594e+00 -1.19323409e+00
-1.32336640e+00 3.83715242e-01 -5.18322229e-01 -1.11201182e-01
-7.69354776e-02 -4.22291100e-01 -8.66937190e-02 -3.06259096e-01
4.54104394e-01 3.92098725e-01 7.48707592e-01 -1.22012186e+00
-3.91057670e-01 2.43292555e-01 4.02805507e-01 -4.33098316e-01
-2.11273551e-01 1.74859688e-01 -1.67280018e-01 -4.76257682e-01
-3.87502164e-01 -9.44645941e-01 -2.47757271e-01 -4.97237623e-01
-4.64988351e-01 -3.74611944e-01 6.56890571e-01 -1.04144681e+00
1.77922666e+00 -2.09657693e+00 1.91074416e-01 -8.22139233e-02
1.98873743e-01 1.48467764e-01 -7.00426400e-01 7.59778500e-01
-2.36792162e-01 4.35500443e-01 -5.03033876e-01 -4.56604958e-01
-6.53043613e-02 3.21377963e-02 -6.62386894e-01 -1.80967152e-01
7.06800044e-01 9.61085498e-01 -1.13532746e+00 -2.29454845e-01
-4.68589038e-01 -5.86049594e-02 -1.27434754e+00 7.00053215e-01
-1.00761962e+00 1.59661472e-01 -3.34403008e-01 6.78203285e-01
5.33743799e-01 -3.36576760e-01 1.34350672e-01 4.78224963e-01
-1.58048458e-02 7.02175438e-01 -6.48402750e-01 1.70633447e+00
-8.87960732e-01 6.25683129e-01 -2.25297049e-01 -6.44301295e-01
9.68516886e-01 4.31802839e-01 -2.92314947e-01 -3.74458253e-01
-5.93228579e-01 2.38275722e-01 8.32681265e-03 -1.09748518e+00
4.02402073e-01 3.33130866e-01 -1.53642163e-01 1.01023352e+00
-8.02478418e-02 -9.89980027e-02 5.55693777e-03 3.34459573e-01
1.80201840e+00 5.03805220e-01 1.48574278e-01 1.52648970e-01
4.60851610e-01 2.78231531e-01 8.11579704e-01 8.55472386e-01
3.10362488e-01 6.66786373e-01 1.04581034e+00 -1.73840255e-01
-7.40629017e-01 -5.52216291e-01 4.94627386e-01 9.80414093e-01
-2.05323145e-01 -1.15979016e+00 -1.05232131e+00 -1.42144585e+00
-3.33666623e-01 1.03953874e+00 -5.68764091e-01 -3.59116197e-01
-8.18253040e-01 -4.84733641e-01 6.77386165e-01 4.44667161e-01
-4.21888083e-02 -1.35244906e+00 -5.31826735e-01 2.69173175e-01
-4.30041969e-01 -4.04955298e-01 -7.15294003e-01 -5.04049100e-02
-6.33741379e-01 -1.34742260e+00 -2.08062306e-02 -8.84354711e-01
1.00315309e+00 3.06299925e-01 1.43481386e+00 1.07808697e+00
-3.55599612e-01 3.37311417e-01 -8.64093244e-01 5.24244122e-02
-1.23736405e+00 1.31359756e-01 -7.01730132e-01 -3.56800556e-01
2.84462482e-01 -6.66581213e-01 -1.86446905e-01 2.92007089e-01
-7.89905667e-01 1.19707286e-01 6.33835137e-01 9.08072650e-01
1.39952302e-01 -3.15735549e-01 8.39798272e-01 -1.19190109e+00
8.09943140e-01 -7.72413671e-01 -6.44447923e-01 6.45237029e-01
-4.99205410e-01 1.98014230e-01 7.33137667e-01 -6.86238766e-01
-1.18261003e+00 -3.43347758e-01 -3.42102289e-01 -3.76247354e-02
8.51442385e-03 1.00468826e+00 3.86325382e-02 -3.82603817e-02
9.72899556e-01 1.42093420e-01 -2.50439078e-01 -5.68690002e-01
-4.98823486e-02 7.46703029e-01 1.14302851e-01 -1.02683449e+00
6.43999398e-01 -3.67846251e-01 -7.81873822e-01 -3.97743434e-01
-5.83447695e-01 -1.53089212e-02 3.08960546e-02 8.54710024e-03
4.59165931e-01 -5.82342803e-01 -3.00016731e-01 1.86067104e-01
-1.71198082e+00 -7.70108461e-01 -1.88852996e-02 -1.84955552e-01
-1.48928970e-01 5.68521976e-01 -6.71582401e-01 -8.09066474e-01
-3.96589309e-01 -1.41502666e+00 1.11037052e+00 -1.50899030e-02
-7.45948792e-01 -7.14123309e-01 4.23303306e-01 1.82389721e-01
5.36625743e-01 1.09763123e-01 1.65427101e+00 -3.68517220e-01
-8.46671462e-01 -3.45504463e-01 8.49654749e-02 4.86686796e-01
1.89063236e-01 2.97670096e-01 -8.31200123e-01 -1.58934891e-01
1.85363993e-01 -2.99137414e-01 2.69317448e-01 -2.81845957e-01
8.35893810e-01 -6.45154536e-01 -4.38760787e-01 1.04817010e-01
1.24435937e+00 6.50857985e-02 6.86473668e-01 3.78020331e-02
6.89276457e-01 5.43842375e-01 6.73249722e-01 3.12087774e-01
4.18887019e-01 4.82722729e-01 3.41583610e-01 3.80898267e-01
-2.43948236e-01 -6.26790583e-01 8.70293736e-01 1.20043921e+00
5.65490842e-01 -1.38490096e-01 -1.12953031e+00 8.07025790e-01
-1.89759374e+00 -6.75481975e-01 -2.92855471e-01 2.19320774e+00
1.26454699e+00 1.91670842e-02 -3.42417598e-01 -2.06645206e-01
5.18331766e-01 -2.61836141e-01 -2.56785512e-01 -5.85216224e-01
4.48504359e-01 3.31866771e-01 -4.34856653e-01 6.66172683e-01
-5.15752316e-01 8.15749586e-01 5.73977566e+00 5.57142079e-01
-1.14731348e+00 2.90127099e-01 2.25829422e-01 2.67542839e-01
-8.51469755e-01 7.48533368e-01 -8.09865594e-01 3.07961881e-01
9.37328696e-01 -2.19236255e-01 5.17540395e-01 1.26191568e+00
2.38735937e-02 -2.80825615e-01 -1.54150963e+00 3.49459976e-01
1.73987970e-01 -1.26138759e+00 -6.56091049e-02 -2.19714493e-01
9.41058397e-01 -1.12317741e-01 -3.06066632e-01 7.85362303e-01
2.48030961e-01 -8.72502863e-01 7.83216894e-01 4.25915211e-01
3.32228631e-01 -3.66461754e-01 7.46468484e-01 5.72147906e-01
-1.07606936e+00 -1.80274174e-01 -3.85951400e-01 -2.55963504e-01
-3.15263897e-01 6.37474000e-01 -1.35879374e+00 3.72052401e-01
5.01708150e-01 8.20191741e-01 -1.13048995e+00 1.21019232e+00
-5.66849470e-01 1.06754315e+00 2.24092677e-01 -1.31948262e-01
-2.29712382e-01 1.60217464e-01 5.08089781e-01 1.31910610e+00
5.44946611e-01 -4.49035585e-01 3.25146705e-01 1.52503729e+00
4.46911715e-02 -3.18921693e-02 -5.67322850e-01 -2.63729930e-01
5.74672222e-01 1.12479544e+00 -2.00792283e-01 -2.04507157e-01
-6.66945338e-01 7.06440568e-01 6.24186635e-01 4.16502953e-01
-8.95095229e-01 -6.03923500e-01 6.02667093e-01 1.41689539e-01
1.52375087e-01 -8.78866762e-02 -2.59577453e-01 -1.15020311e+00
6.59193695e-01 -1.34652853e+00 -7.98025057e-02 -8.11875820e-01
-1.05159712e+00 8.44155550e-01 -1.85686126e-01 -1.33119535e+00
-2.99591988e-01 1.09557599e-01 -1.23942649e+00 9.39414918e-01
-1.21894717e+00 -8.75692070e-01 -2.86248386e-01 -2.98806608e-01
9.81234848e-01 2.44961735e-02 9.48723614e-01 1.27575025e-01
-6.57972932e-01 8.14570963e-01 -6.57779992e-01 -1.48920730e-01
7.64711738e-01 -1.22463560e+00 9.19634819e-01 1.21201849e+00
-1.24307185e-01 1.21714497e+00 6.01650596e-01 -1.09707320e+00
-1.41017401e+00 -1.49896443e+00 1.01166248e+00 -8.00773442e-01
6.37254715e-01 -5.86828768e-01 -1.38422406e+00 7.58683622e-01
3.83859515e-01 -5.07948339e-01 5.20209908e-01 1.41868174e-01
-6.89614832e-01 1.32344633e-01 -8.29421937e-01 6.39587045e-01
7.42170572e-01 -6.39579594e-01 -6.39828265e-01 3.61104459e-01
1.27755952e+00 -2.95073748e-01 -4.02396232e-01 1.26909405e-01
4.35760170e-02 -1.00147510e+00 2.47344896e-01 -3.04620802e-01
1.02981484e+00 -5.90919316e-01 3.05188242e-02 -1.32010901e+00
1.26483023e-01 -9.64140058e-01 -8.68764669e-02 1.49813128e+00
9.38181162e-01 -6.54317677e-01 5.21814466e-01 6.31953895e-01
-6.49908662e-01 -8.89031053e-01 -4.81894761e-01 -4.88854140e-01
-2.86769688e-01 -6.70793891e-01 7.18277276e-01 7.83272147e-01
6.66942000e-01 2.88063794e-01 -1.93552434e-01 3.11941206e-01
1.74652353e-01 3.91282201e-01 1.05887139e+00 -8.45335126e-01
-1.10322464e+00 -1.20653644e-01 8.34050104e-02 -1.05912995e+00
4.09478068e-01 -1.07072365e+00 6.11447573e-01 -1.22809792e+00
5.32241762e-01 -6.49615824e-01 3.52777600e-01 9.16325331e-01
-6.37686431e-01 -2.00771898e-01 -2.92896658e-01 8.45938697e-02
-5.49821138e-01 1.96688503e-01 7.81968355e-01 -2.37538487e-01
-4.78237659e-01 1.70016900e-01 -9.13191497e-01 4.73467797e-01
4.98975813e-01 -9.34024870e-01 -5.12785316e-01 -7.29483783e-01
8.35125804e-01 3.92867088e-01 4.09096062e-01 -9.26045060e-01
1.00838706e-01 -1.76977739e-01 -3.81149292e-01 -1.38943881e-01
-3.74568760e-01 -2.01928943e-01 1.94384247e-01 3.43307912e-01
-5.55467606e-01 3.49171519e-01 4.17867035e-01 5.31729937e-01
-3.62724513e-01 -8.11953425e-01 2.77054578e-01 -1.86873287e-01
-3.41404587e-01 -2.23982975e-01 -5.76703668e-01 1.42016470e-01
6.68716788e-01 -1.66354328e-02 -8.04935873e-01 -6.06229268e-02
-4.10156071e-01 2.34913528e-01 7.71794379e-01 7.71193743e-01
6.01539254e-01 -9.31958437e-01 -4.32477742e-01 3.43386889e-01
5.29349089e-01 -1.16254995e-03 3.92895341e-02 9.35132146e-01
-4.02830869e-01 1.92958321e-02 4.10631716e-01 -6.54692531e-01
-1.35343659e+00 5.22218049e-01 1.23410545e-01 -2.12963283e-01
-6.69297338e-01 8.48457932e-01 2.22142458e-01 -7.27089405e-01
5.08727767e-02 -6.89649701e-01 2.13595793e-01 -6.89373195e-01
6.35154665e-01 -8.16003457e-02 3.82963449e-01 2.86870629e-01
-2.86307126e-01 2.52865881e-01 -2.41579652e-01 1.61833614e-01
1.39641690e+00 3.56968313e-01 -7.34706700e-01 4.02367800e-01
9.89936888e-01 4.12655592e-01 -1.03543723e+00 4.13698331e-02
4.46301848e-01 -4.42160159e-01 -5.61496377e-01 -9.77301955e-01
-6.59478128e-01 8.17261159e-01 -5.22749089e-02 1.96835414e-01
7.56085157e-01 9.45013389e-02 6.51924014e-01 5.02015352e-01
6.23081088e-01 -5.93331158e-01 5.19211113e-01 5.56390822e-01
1.05754304e+00 -9.22072470e-01 -4.63305086e-01 -5.93940258e-01
-4.26611871e-01 1.15060222e+00 1.13556218e+00 -1.11519463e-01
6.40890896e-02 6.16246700e-01 -1.84435621e-02 -3.17305475e-02
-1.43636358e+00 4.92757589e-01 -1.32369800e-02 7.47226536e-01
8.42922568e-01 -2.79050291e-01 -1.43551528e-01 1.02356410e+00
-7.19326688e-03 -8.73424858e-03 9.44721580e-01 1.40714574e+00
-4.45459247e-01 -1.64516914e+00 -2.59968549e-01 4.45806563e-01
-1.39073730e-01 -4.87079650e-01 -6.30599141e-01 2.62920320e-01
6.92979768e-02 1.05972898e+00 -4.38001275e-01 -6.31925285e-01
2.82085508e-01 3.58085424e-01 4.65129822e-01 -1.46254241e+00
-1.09676445e+00 -1.51017576e-01 3.73311669e-01 -6.27800345e-01
2.29389384e-01 -5.33138394e-01 -1.14950442e+00 1.90042585e-01
-5.40135980e-01 5.39496601e-01 2.73337275e-01 9.19880211e-01
8.24331462e-01 5.96866667e-01 6.17369771e-01 -6.63454771e-01
-5.73940694e-01 -1.16500223e+00 3.55772316e-01 1.50849268e-01
4.50899899e-01 -2.80943394e-01 -4.67401326e-01 2.88386881e-01] | [7.658444881439209, 7.730049133300781] |
79b4b199-c3ec-498d-9267-f58fde1b87ab | grammatical-error-detection-using-tagger | null | null | https://aclanthology.org/W14-1706 | https://aclanthology.org/W14-1706.pdf | Grammatical Error Detection Using Tagger Disagreement | null | ['Anubhav Gupta'] | 2014-06-01 | null | null | null | ws-2014-6 | ['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] | [-7.385798454284668, 3.7128329277038574] |
764fe5fc-ab81-4462-b052-3a466d36aa10 | a-cascade-sequence-to-sequence-model-for | 1908.04917 | null | https://arxiv.org/abs/1908.04917v2 | https://arxiv.org/pdf/1908.04917v2.pdf | A Cascade Sequence-to-Sequence Model for Chinese Mandarin Lip Reading | Lip reading aims at decoding texts from the movement of a speaker's mouth. In recent years, lip reading methods have made great progress for English, at both word-level and sentence-level. Unlike English, however, Chinese Mandarin is a tone-based language and relies on pitches to distinguish lexical or grammatical meaning, which significantly increases the ambiguity for the lip reading task. In this paper, we propose a Cascade Sequence-to-Sequence Model for Chinese Mandarin (CSSMCM) lip reading, which explicitly models tones when predicting sentence. Tones are modeled based on visual information and syntactic structure, and are used to predict sentence along with visual information and syntactic structure. In order to evaluate CSSMCM, a dataset called CMLR (Chinese Mandarin Lip Reading) is collected and released, consisting of over 100,000 natural sentences from China Network Television website. When trained on CMLR dataset, the proposed CSSMCM surpasses the performance of state-of-the-art lip reading frameworks, which confirms the effectiveness of explicit modeling of tones for Chinese Mandarin lip reading. | ['Rui Xu', 'Mingli Song', 'Ya Zhao'] | 2019-08-14 | null | null | null | null | ['lipreading'] | ['computer-vision'] | [ 3.87454599e-01 -1.09493136e-01 -7.35355794e-01 -1.59768358e-01
-1.04964614e+00 -8.87469053e-02 3.86547178e-01 -2.65620172e-01
-2.09682778e-01 4.14114386e-01 7.93137014e-01 -4.57324892e-01
6.63707972e-01 -3.51873964e-01 -4.81981575e-01 -4.83800828e-01
5.68707943e-01 -4.59984243e-01 7.93446749e-02 6.44520521e-02
6.05013549e-01 -2.29661062e-01 -1.84545350e+00 4.24024224e-01
9.40765262e-01 1.16173494e+00 8.18719447e-01 6.35562658e-01
-6.32150471e-01 7.89214253e-01 -3.98734957e-01 -2.12795854e-01
-2.94614464e-01 -7.62099683e-01 -5.86586416e-01 -1.11332163e-01
4.21606243e-01 -2.54094005e-01 -6.73134923e-02 1.13365638e+00
9.69617128e-01 -4.43721950e-01 6.24006271e-01 -9.47864711e-01
-1.09871805e+00 8.16412687e-01 -6.18162632e-01 -2.27274090e-01
3.61688286e-01 2.76086330e-01 9.98413980e-01 -1.31137550e+00
2.73212254e-01 1.30316663e+00 5.04917741e-01 9.86140788e-01
-8.50350440e-01 -6.82480812e-01 1.49932936e-01 3.52240235e-01
-1.47997141e+00 -1.08283293e+00 1.11486411e+00 -2.91053087e-01
7.86982000e-01 2.00723395e-01 4.51911926e-01 1.06856656e+00
1.10373437e-01 1.31987035e+00 1.20458841e+00 -8.13033342e-01
7.15117529e-03 7.88705647e-02 -1.34010678e-02 2.13313192e-01
-4.38814968e-01 -5.84991788e-03 -9.55869377e-01 4.27397579e-01
2.19414175e-01 -7.09983766e-01 -5.74518621e-01 2.51678824e-01
-1.11660528e+00 6.68035448e-01 -1.03088714e-01 1.68532163e-01
-2.65054405e-01 -2.42603749e-01 4.63414490e-01 -1.55982420e-01
5.19970477e-01 -3.55459064e-01 -3.08495820e-01 -3.55759978e-01
-1.00747395e+00 -3.31071675e-01 3.49550724e-01 1.13896775e+00
3.82913083e-01 8.15679133e-02 -1.73183411e-01 1.39783049e+00
7.97742188e-01 9.65148091e-01 8.11036825e-01 -5.71639657e-01
5.17738461e-01 7.28024244e-02 -3.19444686e-01 -5.40904284e-01
1.07809184e-02 2.30546758e-01 -7.73101747e-01 1.30832136e-01
2.79078007e-01 -1.81146532e-01 -7.03254521e-01 1.86881697e+00
-9.52631701e-03 -6.42882362e-02 4.05232072e-01 7.18955100e-01
1.22918785e+00 7.38206625e-01 3.37022305e-01 -4.90756392e-01
1.68594933e+00 -9.46870744e-01 -1.16710746e+00 -2.06627861e-01
2.22181529e-01 -1.16177046e+00 1.66305614e+00 3.04300249e-01
-1.37573850e+00 -9.96302903e-01 -9.54127967e-01 -4.15470988e-01
-1.99022487e-01 4.95785683e-01 -1.70212407e-02 9.13941801e-01
-1.12255847e+00 -1.77591667e-01 -1.90725967e-01 -1.95030257e-01
3.28343689e-01 -1.09653249e-01 1.94212385e-02 1.25757128e-01
-1.37763631e+00 6.79169655e-01 3.01070940e-02 1.68402374e-01
-4.92025316e-01 -5.31192720e-01 -1.15736854e+00 -2.72797532e-02
9.70621258e-02 -2.75420807e-02 1.35323906e+00 -7.96878099e-01
-2.13067698e+00 1.00582564e+00 -8.61277461e-01 -1.32186934e-01
4.43418086e-01 -2.75974311e-02 -5.92884839e-01 2.79863179e-01
-4.67172787e-02 1.21080577e+00 1.10180497e+00 -9.77588415e-01
-6.25837445e-01 3.61446589e-02 -4.01451856e-01 3.07752520e-01
7.61303380e-02 4.10885990e-01 -4.55799192e-01 -7.14224339e-01
-1.36782646e-01 -3.58177692e-01 6.29814506e-01 -1.82007607e-02
-5.20224094e-01 -7.24826455e-01 9.02059734e-01 -1.09757471e+00
1.27686977e+00 -2.37861872e+00 -3.53908062e-01 -4.40705389e-01
-1.42511323e-01 1.92462876e-01 -1.82256714e-01 2.86419660e-01
3.33319865e-02 4.01712507e-01 -2.07071453e-01 -6.19041383e-01
2.96579838e-01 -3.41495425e-01 -3.57750416e-01 2.25937456e-01
3.81372124e-01 1.16938567e+00 -4.18816090e-01 -1.06957650e+00
2.16131017e-01 5.91022193e-01 -3.14923316e-01 1.37668476e-01
-4.20283824e-01 1.32304892e-01 -9.52111334e-02 1.01512575e+00
9.84360158e-01 1.92074001e-01 -1.27597570e-01 -5.05105108e-02
-5.42203665e-01 7.08738029e-01 -4.97886866e-01 1.74063313e+00
-6.45435572e-01 9.54687834e-01 6.43857569e-02 -4.13082033e-01
9.95296597e-01 5.94754696e-01 6.95088366e-03 -1.01022649e+00
8.22659023e-03 1.11705497e-01 -1.70682445e-01 -7.53401279e-01
4.20492649e-01 -1.27791464e-01 1.87610656e-01 3.08278084e-01
-3.35202813e-01 -3.13226849e-01 -1.45214975e-01 -3.57054472e-01
8.99318755e-02 2.90011227e-01 9.89351869e-02 -4.42819327e-01
1.05420005e+00 -4.50121015e-01 2.54843116e-01 2.56600022e-01
-5.10199964e-01 8.40561748e-01 4.73691016e-01 3.16387594e-01
-7.49493301e-01 -1.12042153e+00 -5.01043916e-01 1.12522697e+00
2.46172935e-01 -2.54774421e-01 -1.42067659e+00 -3.55210185e-01
-4.15619999e-01 9.24996138e-01 -2.83733994e-01 8.57048407e-02
-3.33182693e-01 -1.16927333e-01 7.79956102e-01 3.10596317e-01
8.74895334e-01 -1.59621227e+00 -3.26076239e-01 8.24717134e-02
-6.67935908e-01 -1.23036349e+00 -1.30473971e+00 -5.90302765e-01
-1.33571364e-02 -9.01110291e-01 -1.24429798e+00 -1.37947834e+00
2.37378985e-01 9.11426172e-02 8.44865978e-01 -6.28132448e-02
-1.56425461e-01 1.67039379e-01 -2.83143342e-01 -8.29470813e-01
-7.52208114e-01 7.92000890e-02 2.23562326e-02 1.32590756e-01
7.85192668e-01 2.18058173e-02 -4.43735063e-01 3.44792992e-01
-6.30779028e-01 4.43933159e-01 6.75030053e-01 5.07790446e-01
5.06121159e-01 -2.90408403e-01 1.09080088e+00 -4.14231718e-02
8.62665653e-01 -5.51870130e-02 -5.21565020e-01 5.09408414e-01
-4.42169130e-01 -3.10148597e-01 3.01877230e-01 -7.30779707e-01
-1.13642812e+00 -3.31752151e-01 -7.06400275e-01 -7.18862936e-03
-3.71645600e-01 3.96223634e-01 -6.69828653e-01 3.14109772e-01
7.55281094e-03 8.85573506e-01 4.10167605e-01 -6.74875915e-01
1.44976303e-01 1.71070921e+00 7.78354824e-01 -3.64034623e-01
6.05657846e-02 -1.13335907e-01 -4.69089895e-01 -1.47673869e+00
-5.03830373e-01 -1.55959666e-01 -5.46845436e-01 -4.65464294e-01
1.32255459e+00 -9.63011742e-01 -1.10494876e+00 1.30223775e+00
-1.27754104e+00 -4.76283848e-01 4.96540517e-02 5.22023320e-01
-7.62240827e-01 4.89560574e-01 -7.34153926e-01 -1.07625234e+00
-3.74118984e-01 -1.33074510e+00 1.17248499e+00 3.93613160e-01
-3.66435423e-02 -8.02530944e-01 -3.51008445e-01 4.60538715e-01
6.75304592e-01 -5.65841854e-01 1.05846560e+00 -1.10469855e-01
-2.97362804e-01 2.27243617e-01 -3.06641698e-01 6.57712221e-01
4.63631034e-01 -2.77470704e-02 -1.27423072e+00 1.41969904e-01
1.64855868e-01 -4.46390450e-01 6.52419090e-01 7.54078209e-01
1.10036087e+00 -1.24450922e-01 3.72696966e-02 1.48802400e-01
1.03352904e+00 3.26312661e-01 1.06453335e+00 -2.03045994e-01
3.20036650e-01 7.21249342e-01 4.59907621e-01 1.19717158e-01
7.63234794e-01 5.34065902e-01 8.78729001e-02 -1.80786714e-01
-7.64631689e-01 -8.98722887e-01 6.81665897e-01 1.18712556e+00
4.20417100e-01 -2.33745769e-01 -8.51376712e-01 4.91558075e-01
-1.15056992e+00 -8.78925860e-01 1.28926203e-01 2.11135221e+00
1.37991440e+00 2.40450352e-01 1.65970922e-02 1.02300920e-01
1.14248013e+00 3.73502910e-01 -5.61429501e-01 -3.82688880e-01
-4.59822297e-01 4.51834910e-02 -2.41608471e-02 8.46425951e-01
-8.72374117e-01 1.29400516e+00 6.43177366e+00 1.21294332e+00
-1.54865861e+00 -2.10711975e-02 7.36303091e-01 5.11808217e-01
-2.49083206e-01 -4.42577899e-01 -1.02401042e+00 7.42280841e-01
7.93062806e-01 -1.86347857e-01 3.72860163e-01 5.84757209e-01
6.11531138e-01 -2.53127754e-01 -8.19171786e-01 1.08478844e+00
4.23541963e-01 -1.25936234e+00 -8.48792642e-02 -9.08774957e-02
3.63347858e-01 -6.97963014e-02 3.58005524e-01 2.97684669e-01
-5.19339263e-01 -1.23310399e+00 8.94554198e-01 6.20531917e-01
1.36812401e+00 -5.04033685e-01 5.11150479e-01 5.44267833e-01
-1.44806552e+00 4.54786122e-01 -2.28160217e-01 1.31018803e-01
2.05418974e-01 -2.30609342e-01 -4.59110022e-01 1.01400226e-01
4.43889320e-01 6.66454434e-01 -3.48490119e-01 7.25354850e-01
-3.70922118e-01 9.06092942e-01 -5.70460036e-02 -4.21536922e-01
-5.32450080e-02 1.20676845e-01 2.52368778e-01 1.19548273e+00
1.50905520e-01 -1.48922652e-01 -5.75620532e-02 1.05179358e+00
-1.32487833e-01 6.19937778e-01 -2.65086770e-01 -1.41177714e-01
4.99462426e-01 7.63511777e-01 -2.30407223e-01 -1.87658191e-01
-7.87566602e-01 6.31574750e-01 -3.45963478e-01 4.09241974e-01
-7.55844474e-01 -4.09584880e-01 6.87990904e-01 1.54858567e-02
2.53536016e-01 -4.86868508e-02 -2.59915620e-01 -9.26578581e-01
9.62190982e-03 -1.12341070e+00 -2.84489393e-01 -1.15526414e+00
-1.23718691e+00 4.15208697e-01 -3.90417218e-01 -1.22198033e+00
-1.96728691e-01 -6.00629747e-01 -6.62869990e-01 1.36478269e+00
-2.14967227e+00 -1.21386170e+00 -8.84234831e-02 5.86809993e-01
1.40805960e+00 -1.04113750e-01 7.15620458e-01 6.30044285e-03
-3.94581378e-01 8.62873137e-01 -2.57984698e-01 4.34609711e-01
8.05623293e-01 -8.61522555e-01 5.29365182e-01 6.54655516e-01
-1.33635893e-01 3.70417953e-01 2.68217951e-01 -5.11152089e-01
-1.15906453e+00 -7.14424312e-01 1.46555364e+00 -1.34230135e-02
4.25630897e-01 -5.45636058e-01 -7.84562945e-01 1.67870700e-01
7.53363609e-01 -4.30031300e-01 7.28645504e-01 -4.40522403e-01
-7.11051449e-02 4.73065302e-02 -1.04556084e+00 8.14083219e-01
7.67154872e-01 -1.08867073e+00 -6.86161220e-01 -7.46374577e-02
1.01435852e+00 -2.08493665e-01 -5.07017255e-01 3.43422085e-01
6.37452662e-01 -8.55875194e-01 6.20203495e-01 -2.79543459e-01
4.49556738e-01 -3.01319420e-01 -3.40896845e-01 -8.67637813e-01
3.08527857e-01 -8.79876912e-01 2.28622571e-01 1.68450809e+00
5.19964278e-01 -5.43467402e-01 5.50523698e-01 2.75896132e-01
-1.14396207e-01 -6.48252726e-01 -8.65075588e-01 -6.02924049e-01
5.40005803e-01 -8.08173239e-01 6.70406640e-01 4.62940723e-01
2.71109432e-01 4.67351407e-01 -3.98571968e-01 -4.28138860e-02
5.37507534e-01 -3.89157124e-02 4.73053187e-01 -8.04264486e-01
8.58113542e-02 -7.31405377e-01 9.15733948e-02 -1.75701880e+00
6.26558185e-01 -5.91764987e-01 3.96226734e-01 -1.41078997e+00
1.73434153e-01 -1.14498377e-01 -5.46948500e-02 4.70294029e-01
-3.18208426e-01 1.48711905e-01 3.51966053e-01 1.05977356e-01
-5.66441230e-02 6.60151243e-01 1.34421349e+00 -4.00668442e-01
-8.80091488e-02 2.59417772e-01 -5.23355186e-01 7.18785763e-01
7.38554001e-01 1.76292554e-01 -3.03376377e-01 -2.73418218e-01
-3.64269942e-01 3.71451020e-01 -4.36256081e-02 -6.92736387e-01
3.51393074e-01 -6.93599358e-02 1.70884073e-01 -1.02091718e+00
4.38526452e-01 -3.48622918e-01 -5.80428600e-01 2.65554100e-01
-7.49144971e-01 -1.59981206e-01 2.05412850e-01 2.68339980e-02
-4.28424358e-01 -4.25397664e-01 9.14039254e-01 1.80033430e-01
-7.04313099e-01 1.67513609e-01 -6.06164336e-01 3.61572206e-01
7.43200123e-01 -3.30454379e-01 -4.85217422e-01 -6.58715963e-01
-3.82064551e-01 2.81458721e-02 3.05132955e-01 7.30232596e-01
8.23962629e-01 -1.17147231e+00 -9.14922655e-01 5.99108696e-01
1.67605057e-02 -3.27695668e-01 4.47566718e-01 8.72144699e-01
-2.47652963e-01 4.87176985e-01 8.77176523e-02 -7.15963304e-01
-1.36203825e+00 6.41484797e-01 3.02673578e-01 5.27811170e-01
-5.02673507e-01 6.43829167e-01 3.21630955e-01 -1.68446042e-02
6.20195270e-01 -5.30775249e-01 -4.11864817e-01 5.03537208e-02
7.88212001e-01 -4.74608652e-02 -3.50812316e-01 -9.50040221e-01
-2.72784024e-01 9.73002672e-01 1.64200425e-01 -3.39422137e-01
6.46812379e-01 -6.32739484e-01 1.26250654e-01 5.61025143e-01
1.15661621e+00 7.47918606e-01 -1.34589541e+00 -2.71208230e-02
-4.78780530e-02 -1.35716796e-01 9.28601623e-02 -8.84544194e-01
-6.58419609e-01 1.50568640e+00 4.26605970e-01 1.05810538e-01
1.18607771e+00 1.22601122e-01 9.87303913e-01 -2.32040733e-01
4.33753841e-02 -1.25461626e+00 6.47842363e-02 4.65541601e-01
9.53178704e-01 -1.32250261e+00 -7.72082031e-01 -4.98542577e-01
-9.99689400e-01 1.13559437e+00 4.10301119e-01 5.61168194e-01
6.76361859e-01 3.86893153e-01 7.09877372e-01 6.72231793e-01
-6.52800262e-01 -4.33391839e-01 3.06570172e-01 8.08172166e-01
5.80440879e-01 -4.84830402e-02 -2.97882169e-01 6.10984981e-01
-4.65889454e-01 4.36265804e-02 2.37081498e-01 4.80222732e-01
-6.23154521e-01 -1.11304867e+00 -2.97312349e-01 -1.74878538e-01
-6.16213143e-01 -7.56739795e-01 -2.89718449e-01 6.40560925e-01
3.54642347e-02 1.38009644e+00 1.21089183e-01 -1.71875030e-01
-4.17015143e-03 4.12243783e-01 1.02765679e-01 -1.71866268e-01
9.13386121e-02 4.93621081e-01 -1.41618133e-01 -7.61982948e-02
-3.31015438e-01 -7.61956632e-01 -1.31961477e+00 5.45254424e-02
-3.95853817e-01 -2.16468740e-02 1.00177884e+00 9.09907520e-01
1.66168660e-01 8.02893117e-02 7.62863219e-01 -6.20246947e-01
-5.71972013e-01 -1.31057024e+00 -5.29895604e-01 -2.49503497e-02
5.97812295e-01 -2.59912580e-01 -4.86425042e-01 2.43511677e-01] | [14.305388450622559, 4.96316385269165] |
87d25195-c2ef-4ca2-99d1-e7e85fd85afe | limoseg-real-time-bird-s-eye-view-based-lidar | 2111.04875 | null | https://arxiv.org/abs/2111.04875v3 | https://arxiv.org/pdf/2111.04875v3.pdf | LiMoSeg: Real-time Bird's Eye View based LiDAR Motion Segmentation | Moving object detection and segmentation is an essential task in the Autonomous Driving pipeline. Detecting and isolating static and moving components of a vehicle's surroundings are particularly crucial in path planning and localization tasks. This paper proposes a novel real-time architecture for motion segmentation of Light Detection and Ranging (LiDAR) data. We use three successive scans of LiDAR data in 2D Bird's Eye View (BEV) representation to perform pixel-wise classification as static or moving. Furthermore, we propose a novel data augmentation technique to reduce the significant class imbalance between static and moving objects. We achieve this by artificially synthesizing moving objects by cutting and pasting static vehicles. We demonstrate a low latency of 8 ms on a commonly used automotive embedded platform, namely Nvidia Jetson Xavier. To the best of our knowledge, this is the first work directly performing motion segmentation in LiDAR BEV space. We provide quantitative results on the challenging SemanticKITTI dataset, and qualitative results are provided in https://youtu.be/2aJ-cL8b0LI. | ['Martin Simon', 'Heinrich Gotzig', 'Hazem Rashed', 'Patrick Maeder', 'Stefan Milz', 'Senthil Yogamani', 'Mona Hodaei', 'Sambit Mohapatra'] | 2021-11-08 | null | null | null | null | ['moving-object-detection', 'motion-segmentation'] | ['computer-vision', 'computer-vision'] | [ 2.03931913e-01 -3.42894942e-01 -1.85099229e-01 -6.70348704e-01
-4.01235402e-01 -6.84375942e-01 3.71757448e-01 -6.20835973e-03
-6.90001726e-01 3.77761871e-01 -6.21158481e-01 -8.17941487e-01
2.55233437e-01 -6.79318368e-01 -7.77549148e-01 -3.35983396e-01
9.07518938e-02 5.54904163e-01 1.04700661e+00 -1.01924717e-01
3.96174580e-01 8.72885644e-01 -2.01675677e+00 1.66840628e-01
7.00818062e-01 1.01530468e+00 4.54910845e-01 9.81425345e-01
-1.89545989e-01 5.62419772e-01 -4.16975379e-01 1.06288999e-01
4.92463589e-01 -8.40447173e-02 -5.19377291e-01 -2.28788126e-02
9.49719548e-01 -3.01996261e-01 -1.25299722e-01 1.00496852e+00
2.48741925e-01 2.52059907e-01 2.75586694e-01 -1.80352330e+00
2.55763650e-01 4.78032790e-02 -7.67655313e-01 7.36043572e-01
1.46852108e-02 3.94133389e-01 4.12300676e-01 -1.01508021e+00
6.29794538e-01 1.21632171e+00 4.76560086e-01 6.30983293e-01
-8.57536733e-01 -9.51863110e-01 1.66605353e-01 6.72513187e-01
-1.39937031e+00 -8.17585528e-01 7.55388558e-01 -5.28551102e-01
8.96728218e-01 3.68180811e-01 4.79742497e-01 5.45289278e-01
3.22160661e-01 8.21791589e-01 7.90544510e-01 -7.93433636e-02
4.13144052e-01 5.32727763e-02 2.82906145e-01 7.62614608e-01
1.85947433e-01 9.77793261e-02 -4.92958188e-01 1.85490280e-01
2.05912262e-01 -1.75129697e-01 2.21057907e-02 -5.55705965e-01
-1.04834259e+00 5.90029240e-01 2.87103117e-01 -2.57752895e-01
2.76967883e-02 4.83323485e-01 2.76412129e-01 -5.23637906e-02
3.45808029e-01 -3.17069679e-01 -4.21015054e-01 -1.67989343e-01
-1.04011416e+00 1.81233823e-01 2.84218907e-01 1.14335299e+00
9.39946473e-01 1.10581264e-01 2.57016420e-01 3.35758805e-01
3.95072043e-01 8.10121000e-01 1.87608272e-01 -1.23419809e+00
3.34070742e-01 4.31839615e-01 5.98646887e-02 -6.39646173e-01
-4.45667773e-01 1.00780398e-01 -3.01849563e-02 8.67047250e-01
3.68611991e-01 -7.94300810e-02 -1.18180168e+00 1.07641613e+00
7.87606180e-01 4.54207391e-01 1.41758826e-02 1.04117846e+00
1.08214235e+00 5.34153163e-01 4.16344218e-02 1.71050370e-01
1.30810666e+00 -9.27079380e-01 -5.86526573e-01 -6.59197450e-01
5.75062096e-01 -7.38155723e-01 7.34742761e-01 2.81718314e-01
-9.62176919e-01 -8.09795916e-01 -1.25038719e+00 -4.19537365e-01
-4.40994769e-01 1.38628691e-01 4.95921582e-01 8.01138103e-01
-8.42123628e-01 1.96815699e-01 -1.07976401e+00 -1.27695888e-01
6.82805955e-01 3.10882062e-01 -2.12211505e-01 -1.34101436e-01
-7.67415226e-01 7.07037210e-01 1.81838676e-01 1.58751100e-01
-9.14591253e-01 -7.04008281e-01 -9.07336295e-01 -5.23305416e-01
6.07748210e-01 -3.57006401e-01 1.35431278e+00 -5.60740530e-01
-1.06109071e+00 8.27031612e-01 -6.37670457e-01 -7.25336850e-01
4.97196734e-01 -2.39617988e-01 -4.60365474e-01 1.78848475e-01
1.25705659e-01 1.19681394e+00 9.48190451e-01 -1.29223490e+00
-1.32061708e+00 -6.15997553e-01 -2.28560716e-01 -3.74356983e-03
2.51785666e-01 -7.06184804e-02 -5.52714586e-01 3.57639394e-03
1.96316525e-01 -1.05482304e+00 -3.30705851e-01 1.47806421e-01
-2.77877390e-01 -2.35953167e-01 1.61622715e+00 -3.13942760e-01
8.47764671e-01 -2.12299824e+00 -4.31252390e-01 -5.98208793e-03
2.13642672e-01 4.19275165e-01 5.83491586e-02 -2.36182302e-01
2.42559090e-01 -1.67760357e-01 -2.61555910e-01 -4.73624498e-01
-2.01677561e-01 2.42675528e-01 -3.54896367e-01 6.24130189e-01
1.95150480e-01 9.81261611e-01 -8.28233004e-01 -6.12236857e-01
5.38532197e-01 4.40623552e-01 -2.08215088e-01 -3.71304035e-01
-3.34415674e-01 4.42754358e-01 -2.92865902e-01 8.01495731e-01
1.08372951e+00 5.63897491e-01 -3.55259001e-01 -1.56866923e-01
-6.37502789e-01 1.27229586e-01 -1.17859876e+00 1.59572983e+00
-2.60057837e-01 1.32739329e+00 2.67940760e-01 -5.66219032e-01
8.49767148e-01 -2.38334000e-01 5.01575291e-01 -8.90955448e-01
3.13760668e-01 1.05537891e-01 8.06988552e-02 -3.38174969e-01
9.76816714e-01 1.96655244e-01 4.59187524e-03 2.35580742e-01
-5.51500320e-01 -4.30955529e-01 2.86251187e-01 1.60890728e-01
1.04346585e+00 2.28065208e-01 -2.14345038e-01 -1.68828487e-01
4.29674298e-01 5.62785029e-01 7.01205432e-01 4.29541737e-01
-6.86784923e-01 5.11049330e-01 5.74529953e-02 -2.91962981e-01
-8.20699096e-01 -1.17599022e+00 -3.69307220e-01 7.58454800e-01
5.99366486e-01 -1.57598764e-01 -7.33995020e-01 -5.83691597e-01
1.95982784e-01 9.86309886e-01 -2.94283926e-01 1.16619855e-01
-7.85575807e-01 -2.28861257e-01 2.70565569e-01 8.04300964e-01
3.43232274e-01 -8.21708858e-01 -1.57390440e+00 3.00613558e-03
2.08947748e-01 -1.53686237e+00 -3.18931788e-01 1.76079899e-01
-7.35003531e-01 -1.13504899e+00 2.83109006e-02 -6.39330924e-01
4.14535940e-01 1.00605977e+00 9.52062845e-01 -2.63914950e-02
-8.45874429e-01 3.76829624e-01 -6.01164363e-02 -7.97938645e-01
-1.72559485e-01 -1.88131914e-01 1.36093693e-02 -2.03225821e-01
7.21341252e-01 2.45500617e-02 -8.86262476e-01 5.42458773e-01
-5.33546507e-01 7.87013471e-02 2.89945453e-01 4.43795882e-02
8.54504704e-01 1.72057763e-01 5.42931408e-02 -6.86594188e-01
-7.67261833e-02 -4.46045816e-01 -1.11205149e+00 -3.85978520e-01
-3.45546633e-01 -3.82637054e-01 2.07143560e-01 -7.84482211e-02
-8.58198166e-01 7.21766353e-01 1.01231865e-03 -6.67593718e-01
-5.92656732e-01 -3.00074518e-01 -9.81718302e-02 -1.74797550e-01
4.77097660e-01 -1.33851379e-01 -4.04350534e-02 -1.84153363e-01
5.69869578e-01 7.56233513e-01 8.23225319e-01 -1.00655153e-01
7.76159406e-01 1.19792819e+00 1.88971162e-01 -1.17287683e+00
-2.86261976e-01 -9.67055440e-01 -9.42884445e-01 -5.23104191e-01
1.08927763e+00 -7.88062394e-01 -7.72555232e-01 2.18223855e-01
-1.12024128e+00 -5.68791926e-01 -2.47946739e-01 4.22862977e-01
-4.91635412e-01 2.33770758e-01 -7.00119734e-02 -8.27427089e-01
-1.63975134e-01 -1.36928904e+00 1.31956899e+00 2.95796990e-01
-9.33428667e-03 -6.42610192e-01 -1.35904804e-01 4.78950292e-01
2.46486347e-02 2.17059791e-01 1.61813542e-01 -2.59846717e-01
-9.64611471e-01 -6.64984509e-02 -2.72532791e-01 1.53744653e-01
-1.41509071e-01 3.11243832e-01 -1.23517013e+00 -2.05112789e-02
-2.09961608e-01 2.47482091e-01 1.18055987e+00 4.19923663e-01
9.33416665e-01 5.49185932e-01 -7.48116016e-01 5.26326060e-01
1.24453318e+00 5.73401928e-01 5.56686819e-01 2.69812286e-01
9.62915182e-01 8.25815618e-01 1.29308963e+00 9.82373506e-02
5.09679615e-01 7.46781945e-01 6.67540908e-01 -1.86353289e-02
-4.71643418e-01 1.47729740e-01 3.65006983e-01 1.79340243e-01
3.20179880e-01 -1.28168270e-01 -1.28102589e+00 6.99897885e-01
-1.65892601e+00 -9.14563715e-01 -9.98608053e-01 2.07528210e+00
2.48944193e-01 4.09507811e-01 1.25593871e-01 1.52082875e-01
4.65755999e-01 -4.85570282e-02 -6.54072881e-01 -4.87234086e-01
2.54381537e-01 2.78971344e-01 1.07818222e+00 8.33440781e-01
-1.33767831e+00 1.16920686e+00 4.88757277e+00 6.60913467e-01
-1.16659129e+00 2.45121315e-01 3.35467517e-01 -3.85962993e-01
-8.20331275e-02 6.08385578e-02 -1.31698322e+00 4.97578263e-01
1.23155379e+00 6.36655539e-02 -3.27348262e-02 8.81677449e-01
5.01832426e-01 -6.66993976e-01 -9.75480258e-01 9.11106527e-01
-2.99788117e-02 -1.10110152e+00 -5.10468423e-01 5.02509698e-02
4.70411658e-01 4.79487658e-01 1.06359906e-01 -4.30196058e-03
1.81331374e-02 -8.81437480e-01 1.07569861e+00 2.84479707e-01
5.93400300e-01 -8.86408746e-01 3.67017776e-01 4.93028641e-01
-1.51240158e+00 -8.60377178e-02 -4.07807529e-01 1.57794550e-01
4.18923736e-01 4.79782313e-01 -1.03408420e+00 2.70495534e-01
8.14331710e-01 5.78690231e-01 -7.08283603e-01 1.01225245e+00
5.62873669e-02 2.94834793e-01 -3.68645698e-01 1.64346382e-01
2.25344241e-01 -1.86784670e-01 6.78294599e-01 1.16317987e+00
3.22248816e-01 3.46544497e-02 2.63141721e-01 7.44790852e-01
2.26636529e-01 -3.24089944e-01 -7.12438822e-01 2.76381105e-01
5.51684737e-01 1.40699971e+00 -1.20626378e+00 -3.00624818e-01
-4.29239243e-01 7.76547909e-01 -2.21217632e-01 8.27657804e-02
-1.14662957e+00 -4.80177283e-01 1.02062643e+00 4.80005383e-01
5.45670807e-01 -7.28512943e-01 -4.00851965e-01 -4.75419432e-01
9.82317701e-02 -1.23839132e-01 6.85375230e-03 -6.60922170e-01
-4.52072918e-01 4.76037681e-01 1.26836225e-01 -1.30353689e+00
-5.48626482e-02 -7.81068146e-01 -4.44647640e-01 5.57328343e-01
-1.80905497e+00 -8.95704687e-01 -7.28109598e-01 5.35574496e-01
9.91758406e-01 3.94073576e-02 1.53449163e-01 5.61739743e-01
-4.59344000e-01 2.02663258e-01 -1.36198848e-01 -2.12718517e-01
5.13000429e-01 -1.00535119e+00 9.23429847e-01 1.02208364e+00
6.41241670e-02 1.23330548e-01 7.03420103e-01 -8.10079098e-01
-1.52752376e+00 -1.45524609e+00 4.43964481e-01 -7.93991029e-01
4.29591417e-01 -4.40216959e-01 -8.91829967e-01 6.16814196e-01
-6.15830906e-03 2.80873090e-01 2.70549476e-01 -7.50785232e-01
1.29197225e-01 -1.85762987e-01 -1.16680264e+00 5.40763617e-01
1.24465358e+00 -1.59435093e-01 -8.77686068e-02 2.33202279e-01
6.22211993e-01 -7.26068854e-01 -1.81166828e-01 5.36278248e-01
4.82894391e-01 -1.02729809e+00 1.23117685e+00 -3.66379768e-01
-4.94172834e-02 -9.24528599e-01 -1.54652268e-01 -7.11181939e-01
2.68189073e-01 -3.19656968e-01 -7.39719272e-02 9.69568431e-01
8.95663276e-02 -4.83884931e-01 1.19491565e+00 3.45822275e-01
-5.74450135e-01 -4.60438728e-01 -1.15120566e+00 -7.61994541e-01
-1.77152410e-01 -1.06870759e+00 3.30783427e-01 4.82038200e-01
-6.63488925e-01 7.64650181e-02 1.69079095e-01 4.57421958e-01
7.27043331e-01 1.16832308e-01 1.10049689e+00 -1.20977473e+00
3.75966340e-01 -3.70472401e-01 -8.67222369e-01 -1.03574324e+00
2.56863534e-01 -7.49141932e-01 5.16950965e-01 -1.43627763e+00
-2.74491638e-01 -4.89212245e-01 1.63385689e-01 1.64893419e-01
6.40887991e-02 6.22103870e-01 2.08963174e-02 1.35716591e-02
-5.62466800e-01 1.53024688e-01 9.18573081e-01 -6.90707117e-02
-1.63867012e-01 1.02710508e-01 -5.39855063e-02 8.79197180e-01
1.22449076e+00 -4.98013079e-01 -6.54351532e-01 -5.13068974e-01
-1.77431375e-01 -2.85874873e-01 5.80168784e-01 -1.24599969e+00
6.20236158e-01 -2.92829424e-01 2.95291066e-01 -1.51401734e+00
7.21009910e-01 -8.48791242e-01 2.54992507e-02 6.06534600e-01
3.36317182e-01 2.19966546e-01 7.36289144e-01 5.41613042e-01
6.68703318e-02 -2.63233393e-01 8.46902966e-01 2.47157831e-02
-1.58003223e+00 3.01548153e-01 -5.90754747e-01 -1.16527960e-01
1.65961385e+00 -6.62074745e-01 -4.51347589e-01 1.08628638e-01
-4.22281891e-01 6.89999759e-01 5.49169183e-01 6.25501335e-01
1.07295966e+00 -1.12253976e+00 -5.32222688e-01 4.48553920e-01
8.54453072e-02 3.19199532e-01 3.72989178e-01 8.56044352e-01
-9.67012227e-01 5.99584699e-01 -1.95508197e-01 -1.09283900e+00
-1.82417977e+00 4.64186937e-01 2.27717906e-01 7.48175919e-01
-7.01988459e-01 9.68938231e-01 6.92689642e-02 -1.09924242e-01
8.08158796e-03 -4.72140372e-01 -1.14870667e-01 7.74228722e-02
5.39172411e-01 7.55915463e-01 3.14032197e-01 -1.01733780e+00
-7.80835390e-01 6.44786179e-01 -3.22306454e-02 -2.75268555e-01
7.15999186e-01 -3.37152034e-01 1.88320458e-01 4.50350523e-01
1.09176135e+00 2.13824689e-01 -1.37802815e+00 1.78907245e-01
8.22497457e-02 -7.36199081e-01 3.28077555e-01 -1.33366972e-01
-1.01127732e+00 1.15810132e+00 9.83076811e-01 -4.13282961e-02
7.73710310e-01 1.54225409e-01 1.09095657e+00 1.56229630e-01
4.52135116e-01 -1.18120193e+00 -3.19677442e-01 4.32735652e-01
3.39847028e-01 -1.37149894e+00 8.48040730e-02 -6.88988328e-01
-4.52586412e-01 1.04095972e+00 7.96846271e-01 7.14031830e-02
4.98822123e-01 5.55233061e-01 3.59895676e-01 -3.16663176e-01
-6.02917790e-01 -6.05732501e-01 2.54195809e-01 6.81564987e-01
1.57811958e-02 7.63526112e-02 -6.65720478e-02 -7.34366253e-02
-3.49871039e-01 -3.18281263e-01 5.39741695e-01 1.34961247e+00
-9.04020488e-01 -8.44412982e-01 -4.78270173e-01 1.90563560e-01
-1.41248032e-01 1.99102297e-01 -3.08156341e-01 8.03801894e-01
4.23029274e-01 1.24657667e+00 6.64225280e-01 -2.84713000e-01
3.72678220e-01 -5.78745306e-02 4.42875952e-01 -6.88764215e-01
-3.78094949e-02 -1.08823046e-01 -8.52997601e-03 -8.66781712e-01
-2.59163886e-01 -8.70810688e-01 -1.79870999e+00 -2.53706694e-01
-2.74819247e-02 -1.46921888e-01 1.30158293e+00 4.79774415e-01
4.67482030e-01 6.99915290e-01 5.87209105e-01 -1.27000904e+00
5.23031205e-02 -2.92526037e-01 -2.09842518e-01 -3.28056216e-02
5.18558323e-01 -7.85275400e-01 -1.77666396e-01 8.13251957e-02] | [8.047029495239258, -1.5997437238693237] |
ace1360e-779b-4358-9461-3289cea7ca54 | emgse-acoustic-emg-fusion-for-multimodal | 2202.06507 | null | https://arxiv.org/abs/2202.06507v1 | https://arxiv.org/pdf/2202.06507v1.pdf | EMGSE: Acoustic/EMG Fusion for Multimodal Speech Enhancement | Multimodal learning has been proven to be an effective method to improve speech enhancement (SE) performance, especially in challenging situations such as low signal-to-noise ratios, speech noise, or unseen noise types. In previous studies, several types of auxiliary data have been used to construct multimodal SE systems, such as lip images, electropalatography, or electromagnetic midsagittal articulography. In this paper, we propose a novel EMGSE framework for multimodal SE, which integrates audio and facial electromyography (EMG) signals. Facial EMG is a biological signal containing articulatory movement information, which can be measured in a non-invasive way. Experimental results show that the proposed EMGSE system can achieve better performance than the audio-only SE system. The benefits of fusing EMG signals with acoustic signals for SE are notable under challenging circumstances. Furthermore, this study reveals that cheek EMG is sufficient for SE. | ['Yu Tsao', 'Hsin-Min Wang', 'Kai-Chun Liu', 'Kuan-Chen Wang'] | 2022-02-14 | null | null | null | null | ['electromyography-emg'] | ['medical'] | [ 3.36829364e-01 -1.91915825e-01 -1.74155518e-01 1.25609249e-01
-1.26330924e+00 -8.27351213e-02 2.05373377e-01 -6.52675092e-01
-3.62214863e-01 6.54430628e-01 4.57085282e-01 2.00418204e-01
-1.66052774e-01 -1.51864424e-01 -3.75894070e-01 -1.03471935e+00
2.97119379e-01 -5.41596293e-01 -1.37175143e-01 -2.24258319e-01
8.72913301e-02 2.65470415e-01 -2.04435849e+00 1.51530027e-01
8.89197886e-01 1.16663170e+00 6.50967181e-01 6.46127999e-01
-1.78230032e-01 -8.66097808e-02 -8.21090877e-01 -1.47189483e-01
-2.95125991e-01 -4.85158682e-01 -1.14780135e-01 1.44810989e-01
1.81692958e-01 -1.32265240e-01 -2.76650757e-01 1.04304028e+00
1.10265052e+00 6.31102622e-02 4.40911561e-01 -1.27040112e+00
-1.84484988e-01 3.91464561e-01 -4.87537205e-01 -1.17364548e-01
5.14862120e-01 3.03852484e-02 5.69251299e-01 -1.00768507e+00
3.94659191e-01 1.17181003e+00 5.61499178e-01 9.12839115e-01
-9.07369852e-01 -8.33434701e-01 -1.77042112e-01 4.53435004e-01
-1.17845750e+00 -8.12813640e-01 1.15073502e+00 -1.05205894e-01
5.72854102e-01 3.22750539e-01 4.67359483e-01 1.45500040e+00
9.80565026e-02 1.04706335e+00 1.25944638e+00 -3.91121000e-01
-1.33346245e-01 8.27380121e-02 -3.20745349e-01 1.84336141e-01
-3.22138429e-01 3.66444588e-01 -8.93566668e-01 -8.66096243e-02
3.28941047e-01 -2.23350301e-01 -6.65925086e-01 2.11351395e-01
-1.01809502e+00 4.10132259e-01 -1.81254148e-01 5.42417407e-01
-5.66633880e-01 -6.24747723e-02 4.55784351e-01 2.96417981e-01
2.42259428e-01 2.70898417e-02 -2.33435020e-01 -6.77654207e-01
-7.81935632e-01 -2.75837362e-01 5.17570078e-01 4.72101778e-01
1.88969865e-01 4.79745120e-01 1.81406125e-01 1.43603218e+00
5.01209319e-01 9.12870169e-01 8.27452302e-01 -8.59652638e-01
4.05835360e-01 -6.79153502e-02 -1.58016548e-01 -9.04373527e-01
-2.99402326e-01 -2.25266248e-01 -7.37700105e-01 8.71673375e-02
-2.74820179e-02 -3.39213103e-01 -6.24876380e-01 1.99856687e+00
2.88118541e-01 3.38160008e-01 2.16868117e-01 1.03509927e+00
1.11871219e+00 5.39277673e-01 2.29931623e-01 -7.15752780e-01
1.19956565e+00 -5.62322497e-01 -1.50859416e+00 -6.85728416e-02
1.50396898e-01 -1.01067984e+00 1.11527610e+00 7.27091372e-01
-1.02907181e+00 -6.35962784e-01 -8.63323569e-01 2.83610076e-01
-8.93860161e-02 5.32296956e-01 3.08802992e-01 9.24512327e-01
-7.93478549e-01 3.28568846e-01 -8.54794502e-01 -1.18118919e-01
1.38091043e-01 3.68689090e-01 -7.90532768e-01 3.82050633e-01
-1.28127658e+00 6.29614115e-01 4.31610644e-02 3.46824288e-01
-5.51444530e-01 -2.89341301e-01 -1.00323236e+00 -6.76373020e-03
5.13708651e-01 -2.29584187e-01 1.02392280e+00 -6.44829392e-01
-2.04881024e+00 5.66098273e-01 -3.84672433e-01 2.51872569e-01
5.65206297e-02 -3.64698708e-01 -9.14434314e-01 4.99600887e-01
-4.06722933e-01 4.11280304e-01 1.12240350e+00 -1.22007608e+00
-4.11180317e-01 -5.60951471e-01 -5.86042225e-01 2.12330207e-01
-5.91734111e-01 2.84866840e-01 -2.50998914e-01 -7.84720242e-01
3.37623447e-01 -7.63325334e-01 3.25311840e-01 -1.54155985e-01
-1.89804986e-01 -2.57865548e-01 1.01989818e+00 -1.05484343e+00
1.30572724e+00 -2.48473549e+00 3.07246059e-01 1.61515027e-01
-1.49410337e-01 6.23281002e-01 -2.79601842e-01 3.59298706e-01
-4.98690978e-02 1.04457382e-02 -1.12375177e-01 -2.90997565e-01
1.46943882e-01 2.62814611e-02 1.84426755e-01 3.16925049e-01
1.64523497e-01 8.07651341e-01 -4.36198980e-01 -6.01468682e-01
3.08135539e-01 7.25849032e-01 -6.25132769e-02 1.68844089e-01
3.01889360e-01 3.65579516e-01 -1.84486002e-01 1.03551936e+00
6.44773364e-01 4.74094898e-01 2.14220257e-03 -4.62453783e-01
1.85108885e-01 1.13488548e-02 -1.33845615e+00 1.69951725e+00
-4.87494469e-01 6.48486137e-01 9.07963097e-01 -9.29884613e-01
9.76488829e-01 6.61858737e-01 4.89645243e-01 -5.95690489e-01
4.35413063e-01 2.78011888e-01 -3.39538604e-02 -1.09475136e+00
1.64444238e-01 -4.49155420e-01 2.33135715e-01 2.33377472e-01
2.88201541e-01 -1.98220938e-01 -1.15967415e-01 -3.68070871e-01
5.64325809e-01 1.00319199e-01 6.75444528e-02 3.54080498e-01
9.27420676e-01 -6.79721236e-01 5.47815919e-01 2.10800320e-01
-3.29616487e-01 4.47772950e-01 5.01274765e-02 7.22738206e-01
-4.46424514e-01 -1.13247812e+00 -1.73167601e-01 9.09416199e-01
1.33304060e-01 -3.63158852e-01 -8.40460837e-01 -2.59667426e-01
-2.64937699e-01 2.58785605e-01 -8.70576967e-03 -1.77246898e-01
-3.26016694e-01 -5.04480243e-01 7.37631202e-01 4.98216242e-01
4.79654789e-01 -1.15297472e+00 7.62020051e-02 2.89624661e-01
-5.45916975e-01 -1.16731274e+00 -6.30844176e-01 -1.91426143e-01
-7.38388538e-01 -8.18377912e-01 -9.95301366e-01 -1.00045681e+00
1.94033414e-01 2.11944923e-01 2.25222602e-01 -3.09427828e-01
-2.73182988e-01 7.09137619e-01 -3.41779083e-01 -4.57253486e-01
-5.22580981e-01 -1.22959182e-01 4.12170529e-01 1.03343740e-01
4.70626175e-01 -7.28377640e-01 -4.30913121e-01 6.80854857e-01
-8.74678433e-01 -2.80283570e-01 7.66200423e-01 1.16253448e+00
4.88380849e-01 5.82349189e-02 1.17302346e+00 -1.27548818e-04
1.18532526e+00 -2.41890624e-01 -1.42621160e-01 2.29214802e-01
-2.32835948e-01 -3.96520883e-01 2.89691418e-01 -9.25555408e-01
-1.37068093e+00 -2.49383867e-01 -7.93765008e-01 -4.06528115e-01
-4.85119164e-01 6.77176833e-01 -7.17158079e-01 -1.40134960e-01
2.97586203e-01 5.63773334e-01 7.69066095e-01 -6.34503901e-01
-1.13311499e-01 1.67971289e+00 9.39745843e-01 -5.10636389e-01
4.34815288e-01 7.83799887e-02 -1.19932339e-01 -1.27246082e+00
-1.07837901e-01 -4.51770991e-01 -1.27664387e-01 -6.20013714e-01
5.25733948e-01 -6.66445971e-01 -1.04611135e+00 8.32466960e-01
-7.89655447e-01 1.09398507e-01 2.98494220e-01 1.21318400e+00
-6.05251908e-01 7.14918375e-01 -6.63727403e-01 -1.26514053e+00
-5.09670436e-01 -1.28951108e+00 1.21209586e+00 3.75150532e-01
-1.03376970e-01 -6.32394910e-01 -4.18542117e-01 7.00930357e-01
3.84197861e-01 -9.02263448e-02 4.58718687e-01 -2.04574138e-01
-1.07348161e-02 -2.24815533e-01 2.00303525e-01 7.78676271e-01
4.12124544e-01 -1.14497192e-01 -1.12087691e+00 -1.47818804e-01
2.16216773e-01 -3.16399187e-01 4.65195328e-01 3.96573782e-01
1.04634809e+00 1.96775664e-02 -1.55485779e-01 4.34475392e-01
8.81902039e-01 4.83760059e-01 8.89128625e-01 -1.81265354e-01
1.82028145e-01 7.50219345e-01 9.06469762e-01 2.69583344e-01
-3.64769995e-02 7.70652175e-01 2.13282958e-01 -6.79317713e-02
-3.71738225e-01 -2.05278054e-01 7.36431003e-01 1.26332045e+00
-7.01254085e-02 -1.90507054e-01 -3.76027584e-01 4.28457350e-01
-1.48434150e+00 -9.29535925e-01 1.02486119e-01 2.01678157e+00
9.05554831e-01 -2.23530635e-01 1.02681302e-01 5.42392910e-01
8.76558721e-01 -8.55809674e-02 -4.24582452e-01 -1.41603485e-01
-3.46751392e-01 3.88983935e-01 -1.27711490e-01 2.63638526e-01
-8.59384179e-01 6.29809201e-01 6.49612522e+00 1.27900136e+00
-1.59720123e+00 1.60325348e-01 -2.39794761e-01 -7.55467787e-02
-2.93364733e-01 -7.63343215e-01 -5.32887220e-01 7.65302002e-01
9.53967035e-01 5.60885668e-02 2.63251901e-01 5.62742829e-01
5.18714368e-01 -2.21786112e-01 -5.93130469e-01 1.43790686e+00
2.93598622e-01 -8.67986202e-01 -4.90533412e-01 1.53509341e-02
1.62493199e-01 -3.45434487e-01 2.33974293e-01 1.84975162e-01
-7.20330656e-01 -9.24229443e-01 4.95340452e-02 5.80318630e-01
1.05399597e+00 -5.56047201e-01 8.14997792e-01 4.18675154e-01
-1.06054521e+00 -2.50432058e-03 1.39426187e-01 4.58386779e-01
3.68618757e-01 2.03744337e-01 -4.61259574e-01 5.72917819e-01
5.29885590e-01 4.67101544e-01 1.84324607e-01 1.15868390e+00
-3.58178377e-01 6.54986978e-01 -5.04346550e-01 -2.08108395e-01
-2.12922335e-01 -1.21854164e-01 9.54276443e-01 1.00646305e+00
7.36895680e-01 1.12143382e-01 -3.31237853e-01 5.95056772e-01
9.58435908e-02 3.80403489e-01 -6.50037110e-01 -3.24112236e-01
3.84714365e-01 1.22245193e+00 -5.67114726e-02 1.58442155e-01
-4.14010406e-01 7.89395571e-01 -5.94021320e-01 4.77583855e-01
-6.28147304e-01 -6.94334149e-01 6.51554286e-01 -1.82877615e-01
8.87766778e-02 -1.42000765e-01 1.57271668e-01 -9.88777399e-01
2.34225646e-01 -9.69464362e-01 3.37830409e-02 -1.06163037e+00
-1.06829762e+00 3.76950175e-01 -1.80687696e-01 -1.30471897e+00
-4.18226093e-01 -6.90414488e-01 -5.53199887e-01 8.66024852e-01
-1.46208549e+00 -1.05420327e+00 -2.41023481e-01 6.05158031e-01
4.90541041e-01 -3.91233802e-01 7.33402073e-01 6.57874346e-01
-4.07830715e-01 8.92966330e-01 1.20926715e-01 -2.28065372e-01
7.41827190e-01 -6.83418393e-01 -6.23646617e-01 4.82559860e-01
-1.37119800e-01 5.98123968e-01 6.54397190e-01 -5.07787883e-01
-1.59850347e+00 -5.04658997e-01 5.82538188e-01 3.09449881e-01
4.32448566e-01 -1.83049396e-01 -9.42920327e-01 -4.57933284e-02
1.02285527e-01 -3.69818687e-01 9.43454027e-01 -7.07804337e-02
1.86290756e-01 -3.34053546e-01 -1.13437307e+00 5.48217177e-01
1.02604008e+00 -8.21404576e-01 -7.82824516e-01 -2.06062809e-01
3.48353356e-01 -3.37083846e-01 -1.30178452e+00 5.85695982e-01
9.51263368e-01 -4.25052285e-01 7.85548031e-01 -2.96287268e-01
1.36090010e-01 -2.48290360e-01 -1.76945835e-01 -1.59662390e+00
3.51515234e-01 -9.95129824e-01 -9.86029133e-02 1.56778872e+00
2.43461236e-01 -6.96695447e-01 6.24401271e-01 9.46491435e-02
-3.32056999e-01 -4.87896472e-01 -1.12375927e+00 -8.47658575e-01
-2.87869602e-01 -6.84670329e-01 4.22197521e-01 5.12260199e-01
4.48882401e-01 7.38905445e-02 -6.28594816e-01 1.35304078e-01
4.59626973e-01 -3.05559039e-01 7.67094433e-01 -9.14548814e-01
-8.31901729e-02 -4.97246206e-01 -4.83262450e-01 -9.83081698e-01
5.29207945e-01 -5.78615189e-01 2.29085207e-01 -1.28101480e+00
-1.27855092e-01 1.27872184e-01 -4.57169086e-01 2.68256783e-01
-2.34607443e-01 1.17926359e-01 6.62859976e-02 -2.04053774e-01
-4.55901586e-03 9.14142907e-01 1.53601336e+00 -1.81057572e-01
-3.20436388e-01 2.47009560e-01 -5.08171022e-01 6.93374455e-01
7.01000810e-01 -1.40877843e-01 -3.47526968e-01 1.32716998e-01
-4.37870234e-01 7.02562392e-01 5.33042997e-02 -7.14817524e-01
3.63378823e-01 5.70539758e-03 -6.05297536e-02 -7.38059402e-01
8.77916217e-01 -8.51474047e-01 1.66422307e-01 1.19659796e-01
-1.86914623e-01 -3.88850838e-01 3.74998450e-01 4.32556063e-01
-7.83884346e-01 -2.54787028e-01 4.94153291e-01 2.91694999e-01
-6.11851215e-01 -5.12590446e-03 -6.13586068e-01 -3.39695394e-01
8.28483999e-01 -5.40689230e-01 -2.83690393e-01 -7.19061136e-01
-1.09842849e+00 1.81877632e-02 -2.92489976e-01 5.37814319e-01
9.85079527e-01 -1.41185462e+00 -6.00439310e-01 5.27088046e-01
3.52254361e-02 -7.17968166e-01 6.86440170e-01 1.46194184e+00
2.63245642e-01 2.27473408e-01 -3.39803666e-01 -7.01582849e-01
-1.90601707e+00 1.07794171e-02 2.91839242e-01 5.11147976e-01
-3.59316766e-01 6.02903903e-01 -2.23853245e-01 -3.03231627e-01
6.10331118e-01 1.56496949e-02 -4.01355714e-01 3.26294512e-01
5.07937908e-01 5.24777412e-01 9.72693637e-02 -7.42123365e-01
-1.80734158e-01 8.32206488e-01 3.65350336e-01 -3.82770032e-01
1.26752758e+00 -5.11552393e-01 7.05333874e-02 3.31248075e-01
1.17890179e+00 3.11838061e-01 -7.39803851e-01 -1.54201180e-01
-3.26750547e-01 -5.50157428e-01 3.37926835e-01 -8.83322358e-01
-9.95910704e-01 1.14656377e+00 7.51600683e-01 -3.21020558e-02
1.51347101e+00 -8.97676945e-02 1.08853412e+00 4.29917350e-02
3.53350729e-01 -1.49493527e+00 1.54561415e-01 -1.14932559e-01
1.11126852e+00 -1.20702732e+00 -5.48458695e-01 -5.54387152e-01
-6.67110264e-01 1.15095532e+00 4.73825693e-01 6.17148876e-01
7.13930845e-01 5.11348248e-01 1.70088708e-01 2.72886127e-01
-5.95000744e-01 -4.46943581e-01 5.26867926e-01 9.37462747e-01
3.04587126e-01 -5.24670631e-02 -6.40760541e-01 1.04134166e+00
-1.52032882e-01 2.04249367e-01 1.83674604e-01 6.76466644e-01
-3.72003376e-01 -1.08344221e+00 -8.00809622e-01 2.29116544e-01
-6.67273641e-01 8.20501521e-02 -2.06223831e-01 7.77547538e-01
-1.20607167e-01 1.39356291e+00 -3.02046895e-01 -6.41507566e-01
4.08780336e-01 3.77526611e-01 5.16759574e-01 -8.84925872e-02
-3.46741557e-01 9.57868338e-01 3.00581813e-01 -5.58558106e-01
-6.25538170e-01 -6.14063382e-01 -1.26691687e+00 -1.01264501e-02
-6.93192840e-01 6.73387870e-02 1.19720745e+00 9.63352799e-01
3.37569743e-01 5.33956230e-01 7.08923817e-01 -7.58813679e-01
-4.81397659e-01 -1.27189636e+00 -9.12771225e-01 4.40346003e-01
4.06393856e-01 -9.41791832e-01 -7.25050092e-01 -1.23701513e-01] | [14.324756622314453, 5.014816761016846] |
01231e25-4345-4b86-a639-d74a6dccfc30 | toward-robust-diagnosis-a-contour-attention | 2211.16806 | null | https://arxiv.org/abs/2211.16806v1 | https://arxiv.org/pdf/2211.16806v1.pdf | Toward Robust Diagnosis: A Contour Attention Preserving Adversarial Defense for COVID-19 Detection | As the COVID-19 pandemic puts pressure on healthcare systems worldwide, the computed tomography image based AI diagnostic system has become a sustainable solution for early diagnosis. However, the model-wise vulnerability under adversarial perturbation hinders its deployment in practical situation. The existing adversarial training strategies are difficult to generalized into medical imaging field challenged by complex medical texture features. To overcome this challenge, we propose a Contour Attention Preserving (CAP) method based on lung cavity edge extraction. The contour prior features are injected to attention layer via a parameter regularization and we optimize the robust empirical risk with hybrid distance metric. We then introduce a new cross-nation CT scan dataset to evaluate the generalization capability of the adversarial robustness under distribution shift. Experimental results indicate that the proposed method achieves state-of-the-art performance in multiple adversarial defense and generalization tasks. The code and dataset are available at https://github.com/Quinn777/CAP. | ['Shancheng Jiang', 'Shiqi Deng', 'Haohan Wang', 'Jinpeng Liu', 'Jinwen She', 'Xing Zhang', 'Kun Xiang'] | 2022-11-30 | null | null | null | null | ['adversarial-defense'] | ['adversarial'] | [ 1.39019072e-01 6.10118248e-02 7.40116015e-02 -8.60279873e-02
-1.12111652e+00 -5.36832154e-01 3.30799311e-01 -3.64177339e-02
-3.72524798e-01 6.19135380e-01 2.00425223e-01 -4.92746264e-01
-1.44686565e-01 -5.39634943e-01 -4.87989843e-01 -9.47604358e-01
-2.62041897e-01 5.24720311e-01 7.91054443e-02 -8.95276442e-02
-6.24661185e-02 6.32306755e-01 -4.38443571e-01 2.65306801e-01
1.11817408e+00 1.12358391e+00 -2.33799085e-01 6.82051718e-01
3.86654794e-01 6.66138768e-01 -5.04955232e-01 -4.84755099e-01
6.18773162e-01 -4.00856465e-01 -7.01787412e-01 -3.37151825e-01
5.72581850e-02 -4.48158979e-01 -6.43892348e-01 1.23939312e+00
1.00548565e+00 -1.07814215e-01 9.60963547e-01 -1.13174450e+00
-8.41361940e-01 4.06193227e-01 -5.57516098e-01 7.34254837e-01
2.68998463e-02 2.73606598e-01 2.80722201e-01 -6.94161952e-01
4.81394589e-01 9.01336491e-01 6.83438778e-01 8.71702552e-01
-7.01117992e-01 -8.05605233e-01 -7.82282650e-02 1.08064495e-01
-1.23113334e+00 9.18329582e-02 8.51400495e-01 -5.04966080e-01
4.90059018e-01 3.69584501e-01 2.90809840e-01 1.45369756e+00
7.87519157e-01 6.22444928e-01 1.02260816e+00 1.16265178e-01
-2.97931675e-03 3.60817350e-02 -2.37179726e-01 7.71055102e-01
1.80697247e-01 3.12072724e-01 2.37029523e-01 -4.47612911e-01
7.71434009e-01 1.53387457e-01 -5.87105095e-01 -9.20028761e-02
-1.01045442e+00 9.31433618e-01 9.02755082e-01 2.84575403e-01
-4.08139110e-01 4.52342294e-02 5.53944051e-01 3.15372124e-02
4.58028883e-01 3.95892352e-01 2.66766883e-02 2.71831870e-01
-6.18404567e-01 1.86109647e-01 3.60978544e-01 6.10677719e-01
-3.23629797e-01 2.62020797e-01 -3.94291908e-01 5.31320572e-01
2.61296839e-01 7.06475616e-01 5.80207229e-01 -7.54415929e-01
4.37512934e-01 1.93497375e-01 -1.83684081e-01 -9.93325174e-01
-4.97400671e-01 -6.76883817e-01 -1.40750027e+00 1.98458016e-01
3.01984251e-01 -3.44789088e-01 -1.12082171e+00 1.47552836e+00
3.82209003e-01 4.87933338e-01 3.11714895e-02 9.91567850e-01
6.76239192e-01 3.86074573e-01 2.92466223e-01 -1.67715564e-01
1.33986044e+00 -8.53046536e-01 -6.36306882e-01 3.59257869e-02
2.03416705e-01 -6.15784109e-01 7.57853270e-01 3.37736547e-01
-9.95281398e-01 -1.96855560e-01 -9.58066821e-01 3.39322984e-01
-2.29615226e-01 -5.29764771e-01 3.86849165e-01 7.63182819e-01
-5.54842889e-01 4.18115467e-01 -1.15290141e+00 -4.30338643e-02
1.00228620e+00 2.43084684e-01 -1.45991176e-01 -1.90965086e-01
-1.43500757e+00 8.57529759e-01 3.61974835e-02 3.36990990e-02
-1.16991782e+00 -9.62122023e-01 -5.96384108e-01 -1.94268391e-01
2.29322374e-01 -8.66200387e-01 9.95813012e-01 -6.03978813e-01
-1.27253020e+00 7.19465971e-01 5.10321438e-01 -5.69570482e-01
1.09202564e+00 -1.74285397e-01 -3.71766388e-01 5.33375978e-01
-1.76409796e-01 3.30898017e-01 8.74743819e-01 -1.25794792e+00
-1.09788954e-01 -2.89018035e-01 -2.29034230e-01 9.75810066e-02
-2.51748890e-01 1.72302470e-01 -6.08398505e-02 -1.24420273e+00
-1.10835791e-01 -1.02500117e+00 -5.20034432e-01 1.89964369e-01
-5.36855638e-01 2.65476763e-01 7.76528180e-01 -7.70401180e-01
1.16822517e+00 -2.03796601e+00 8.22073892e-02 3.18864107e-01
2.97514558e-01 3.45858067e-01 2.19764665e-01 2.28163246e-02
-1.54454157e-01 4.03110862e-01 -8.30742836e-01 -4.76357862e-02
-2.46946692e-01 -7.17212493e-03 -4.35322404e-01 7.86720812e-01
2.58768559e-01 1.04563320e+00 -7.74630368e-01 -6.97934747e-01
3.63832712e-02 6.11038446e-01 -6.37085915e-01 4.01764929e-01
1.56710923e-01 1.04279053e+00 -9.26298380e-01 9.05952513e-01
8.89624476e-01 -2.07075194e-01 -2.32140958e-01 -9.24935006e-03
5.10705590e-01 -6.11366808e-01 -5.34302413e-01 1.57091880e+00
-9.62170735e-02 4.22702804e-02 9.11181048e-02 -9.65307474e-01
4.34255868e-01 5.68776250e-01 8.23343813e-01 -3.62105012e-01
5.24957478e-01 2.86895961e-01 3.04615349e-01 -7.55172968e-01
-9.63635817e-02 -4.75828081e-01 -3.37769657e-01 2.18634620e-01
-3.73812735e-01 -2.99926877e-01 -5.75848162e-01 1.79256633e-01
1.16103983e+00 -2.23487228e-01 1.87643364e-01 -4.51975018e-01
5.91941535e-01 -6.01198338e-02 6.37495339e-01 6.33365393e-01
-7.23834813e-01 9.71014619e-01 2.41913080e-01 -4.14358407e-01
-9.03329968e-01 -1.45525789e+00 -5.89572966e-01 4.67207968e-01
1.09629713e-01 4.70221758e-01 -8.04835677e-01 -1.10463679e+00
8.13781470e-02 5.06323934e-01 -8.04989457e-01 -4.51126933e-01
-8.06272686e-01 -1.01395440e+00 9.43749666e-01 6.86739087e-01
5.76792479e-01 -1.06250608e+00 -6.32015884e-01 1.42650530e-02
-1.58081993e-01 -8.68697703e-01 -7.16701090e-01 -2.84569621e-01
-6.65163338e-01 -1.04230523e+00 -1.11870313e+00 -4.98864233e-01
7.81363130e-01 -2.42508426e-01 7.90192604e-01 1.98890150e-01
-7.73365617e-01 4.86934930e-01 -2.96119094e-01 -6.18884206e-01
-6.39255643e-01 -1.65735289e-01 1.73740581e-01 -1.24866731e-01
-2.01367766e-01 -4.68833685e-01 -1.02665949e+00 1.27300769e-01
-9.96642590e-01 -3.34172755e-01 5.71628153e-01 1.01376939e+00
6.66725576e-01 -2.04919249e-01 7.79635489e-01 -9.70793605e-01
6.74910784e-01 -9.74228203e-01 -2.36271635e-01 1.40606895e-01
-5.63143611e-01 -2.43201539e-01 7.44184852e-01 -4.92572397e-01
-1.02357006e+00 -3.19185734e-01 -1.26561686e-01 -7.95528471e-01
1.21380426e-02 4.38937008e-01 7.44266883e-02 -1.77582577e-01
7.41885304e-01 4.72734198e-02 -2.08714455e-02 -1.63933560e-01
-7.14656105e-03 5.43876350e-01 6.62719429e-01 -5.99318922e-01
9.90098476e-01 6.40098333e-01 1.27655789e-01 -3.27143133e-01
-6.75668240e-01 -8.37738160e-03 -2.98393905e-01 -2.58654088e-01
1.26044142e+00 -5.93955517e-01 -4.22195882e-01 5.44842660e-01
-9.53625381e-01 -1.19992875e-01 -1.97900236e-01 5.18446565e-01
-4.91161346e-01 4.62842882e-01 -7.59874105e-01 -5.46624482e-01
-8.99482191e-01 -1.38294435e+00 6.95052028e-01 1.47190049e-01
1.03106610e-01 -1.04409719e+00 1.00509129e-01 3.87833834e-01
6.40169203e-01 1.01679385e+00 9.40163672e-01 -8.83852780e-01
-4.30038184e-01 -3.64951074e-01 -1.48344085e-01 3.98903906e-01
7.90657923e-02 -3.00386578e-01 -9.59745884e-01 -5.90587020e-01
4.43771601e-01 -3.07931691e-01 7.48396277e-01 4.93871897e-01
1.63945889e+00 -2.85979569e-01 -4.11134690e-01 1.11797988e+00
1.21729088e+00 4.09861445e-01 5.39554477e-01 1.29788727e-01
8.68525863e-01 1.96383491e-01 3.78732264e-01 4.01763201e-01
1.81557890e-02 1.92466542e-01 6.33012652e-01 -2.57209420e-01
-5.41422442e-02 1.08941302e-01 3.68452109e-02 6.13214612e-01
-1.57454789e-01 -5.05019248e-01 -1.42139935e+00 3.80361706e-01
-1.44176209e+00 -8.41058016e-01 3.24705124e-01 1.87741685e+00
7.19221056e-01 1.71992004e-01 -2.71812409e-01 -2.15804860e-01
5.84774792e-01 1.59133635e-02 -8.72149229e-01 -2.54800320e-01
9.82971862e-02 2.16786027e-01 5.42039454e-01 4.48915213e-01
-1.25984061e+00 4.36390817e-01 5.99105167e+00 8.16483200e-01
-1.27929866e+00 3.72932523e-01 1.03334653e+00 -4.37948518e-02
-3.08704585e-01 -6.92903280e-01 -5.95136508e-02 5.72023869e-01
6.53769791e-01 -2.67578602e-01 1.97531313e-01 6.78168237e-01
-7.28295892e-02 4.14949387e-01 -7.37295628e-01 7.39847422e-01
1.26789168e-01 -1.18266940e+00 4.55042012e-02 1.19503895e-02
6.97544634e-01 1.81811899e-01 6.28698230e-01 1.28618255e-01
8.71166363e-02 -1.43184114e+00 2.90052533e-01 6.66202962e-01
1.12688828e+00 -8.09055746e-01 8.45490932e-01 1.62490308e-01
-8.62196922e-01 -1.48091078e-01 -1.21619500e-01 6.36767089e-01
1.84319913e-01 9.13234800e-02 -8.61521244e-01 8.46336305e-01
7.51076937e-01 3.47712874e-01 -3.65159512e-01 8.84709537e-01
6.65192977e-02 7.03236699e-01 -2.02466547e-01 4.03673768e-01
1.48788810e-01 6.80290759e-02 1.01796341e+00 1.18056822e+00
4.68874305e-01 4.13257957e-01 2.64117092e-01 7.44670272e-01
-1.09025791e-01 8.16959590e-02 -7.69849002e-01 2.13726372e-01
4.15210992e-01 1.11135924e+00 -8.19024146e-01 -1.36183903e-01
-1.33953765e-01 1.00816190e+00 -1.12229154e-01 3.10456812e-01
-1.34650826e+00 -2.40358174e-01 5.68303227e-01 1.63483143e-01
2.08332363e-04 2.05111206e-01 -4.22707587e-01 -1.12649798e+00
-1.93022832e-01 -9.38931048e-01 7.55857885e-01 -4.91285414e-01
-1.73562038e+00 1.07283044e+00 1.54043958e-01 -1.48760378e+00
-8.96835625e-02 -6.92339718e-01 -9.91635501e-01 6.88084066e-01
-1.36917269e+00 -1.28520525e+00 -3.67207080e-01 7.96254277e-01
2.67284334e-01 -4.29972380e-01 9.33510840e-01 4.42021847e-01
-6.56851709e-01 1.15098190e+00 1.06791891e-01 3.11539084e-01
6.62661433e-01 -1.08034015e+00 1.54665127e-01 9.43023860e-01
-6.15663290e-01 4.37124342e-01 5.19277692e-01 -7.27793455e-01
-1.07253575e+00 -1.26090086e+00 -1.81203093e-02 -5.61634958e-01
6.44067109e-01 -8.88961554e-02 -1.10498929e+00 6.96775436e-01
3.16650987e-01 4.97041017e-01 6.57106936e-01 -9.76264954e-01
-2.64124066e-01 1.04291856e-01 -1.78674674e+00 5.48685551e-01
8.32166672e-01 -2.45732605e-01 -4.90679324e-01 4.74555343e-01
9.09873486e-01 -7.75451958e-01 -1.21555138e+00 7.53429592e-01
3.19163680e-01 -5.83460271e-01 1.18814111e+00 -8.34494829e-01
4.74295795e-01 -1.11273825e-01 -7.65809864e-02 -1.15353918e+00
-3.01450759e-01 -6.92486763e-01 4.66484465e-02 7.64490187e-01
3.67786914e-01 -9.94013667e-01 5.62675238e-01 5.90406656e-01
-3.08652282e-01 -1.24542689e+00 -1.30176187e+00 -6.26109123e-01
8.41109812e-01 -3.03940892e-01 4.24791783e-01 1.18876731e+00
-1.92486674e-01 -3.65825385e-01 -2.78385639e-01 4.74610478e-01
7.80466378e-01 -2.46370926e-01 1.28761396e-01 -8.40245962e-01
-3.00016075e-01 -4.10418153e-01 -4.09271330e-01 -2.30920628e-01
-9.88973528e-02 -1.09948909e+00 -7.82180056e-02 -1.16373062e+00
3.26944321e-01 -5.36403358e-01 -7.62761116e-01 3.60479146e-01
-5.05400419e-01 3.87277216e-01 1.99392915e-01 1.42963737e-01
-1.56082258e-01 4.64534968e-01 1.61429775e+00 -2.85558164e-01
2.53666818e-01 -1.74182318e-02 -5.62828124e-01 7.33591318e-01
1.10232306e+00 -7.61352301e-01 -4.29948956e-01 -2.46176317e-01
-2.43593141e-01 2.93656468e-01 3.95562947e-01 -1.09051502e+00
1.43853396e-01 -2.16916502e-01 4.55536038e-01 -1.37075037e-01
7.34286234e-02 -9.93745327e-01 1.00637496e-01 1.05597997e+00
-1.11614719e-01 2.25467488e-01 4.26078022e-01 5.63954890e-01
-3.25996019e-02 2.77727749e-02 1.19815278e+00 -1.16224680e-02
-1.45246610e-02 8.52359414e-01 -2.34434515e-01 5.88591516e-01
1.32468569e+00 1.64498106e-01 -4.60538834e-01 -2.22239390e-01
-6.25461757e-01 3.28516394e-01 2.76266187e-01 3.08740497e-01
8.36919487e-01 -1.34376001e+00 -1.12568176e+00 2.55306393e-01
-6.06846064e-02 2.24902555e-01 6.12577915e-01 1.05425954e+00
-8.47103298e-01 9.41530615e-03 -2.92635709e-01 -4.40140247e-01
-1.06317544e+00 8.73584688e-01 8.19239736e-01 -4.96406078e-01
-7.37284780e-01 1.06912220e+00 5.37148237e-01 -1.67345181e-01
1.04824364e-01 -2.48048052e-01 1.75347272e-02 -4.93574232e-01
3.74453127e-01 2.48386890e-01 -2.93839514e-01 -5.55217803e-01
-5.65373302e-01 6.29748285e-01 -2.85546452e-01 2.25084443e-02
1.18890727e+00 1.50569960e-01 1.93471476e-01 -1.42706439e-01
1.05678153e+00 7.75310099e-02 -1.19085944e+00 3.67171168e-02
-4.18864995e-01 -3.26084793e-01 -6.43787012e-02 -9.29804206e-01
-1.30093658e+00 9.38980997e-01 1.00536668e+00 4.15291525e-02
1.28721476e+00 -1.73771322e-01 1.03369343e+00 -1.62056480e-02
4.08492684e-02 -4.90650207e-01 2.26538315e-01 1.84983835e-01
1.28133583e+00 -1.39052558e+00 1.29272789e-01 -3.13721955e-01
-9.85296190e-01 7.30816245e-01 6.81807697e-01 -4.42127585e-01
1.04242480e+00 5.04106224e-01 5.34156501e-01 -1.71758652e-01
-2.38561645e-01 4.90469426e-01 4.41083223e-01 6.08256638e-01
1.03781849e-01 1.53337866e-01 -7.90919140e-02 9.12015140e-01
-1.80557335e-03 -3.40321243e-01 2.82806128e-01 8.26650739e-01
-8.18697214e-02 -8.47109616e-01 -3.92274052e-01 3.38828266e-01
-1.07519054e+00 -1.68188393e-01 -6.61758259e-02 7.84911811e-01
9.91112217e-02 5.25347352e-01 -2.66672105e-01 -4.00126517e-01
2.83387482e-01 3.14516276e-02 3.99420202e-01 -3.27066094e-01
-7.54434824e-01 -7.78194666e-02 -4.61826771e-01 -4.74331170e-01
1.70634780e-02 -5.94466686e-01 -1.47332311e+00 -1.37248516e-01
2.00902577e-02 6.72909385e-03 2.62753874e-01 5.64293325e-01
4.01918471e-01 9.10263479e-01 7.59408712e-01 -5.76405168e-01
-9.56632972e-01 -9.19415832e-01 -1.95827365e-01 6.78072453e-01
4.64236230e-01 -5.60551226e-01 -4.18027937e-01 -1.03113107e-01] | [14.234519004821777, -2.2603821754455566] |
9971cb6b-6e67-4036-8ab8-5e38187829cf | a-generalized-doubly-robust-learning | 2211.06684 | null | https://arxiv.org/abs/2211.06684v1 | https://arxiv.org/pdf/2211.06684v1.pdf | A Generalized Doubly Robust Learning Framework for Debiasing Post-Click Conversion Rate Prediction | Post-click conversion rate (CVR) prediction is an essential task for discovering user interests and increasing platform revenues in a range of industrial applications. One of the most challenging problems of this task is the existence of severe selection bias caused by the inherent self-selection behavior of users and the item selection process of systems. Currently, doubly robust (DR) learning approaches achieve the state-of-the-art performance for debiasing CVR prediction. However, in this paper, by theoretically analyzing the bias, variance and generalization bounds of DR methods, we find that existing DR approaches may have poor generalization caused by inaccurate estimation of propensity scores and imputation errors, which often occur in practice. Motivated by such analysis, we propose a generalized learning framework that not only unifies existing DR methods, but also provides a valuable opportunity to develop a series of new debiasing techniques to accommodate different application scenarios. Based on the framework, we propose two new DR methods, namely DR-BIAS and DR-MSE. DR-BIAS directly controls the bias of DR loss, while DR-MSE balances the bias and variance flexibly, which achieves better generalization performance. In addition, we propose a novel tri-level joint learning optimization method for DR-MSE in CVR prediction, and an efficient training algorithm correspondingly. We conduct extensive experiments on both real-world and semi-synthetic datasets, which validate the effectiveness of our proposed methods. | ['Jie Sun', 'Rui Zhang', 'Xiao-Hua Zhou', 'Zhenhua Dong', 'Peng Wu', 'Haoxuan Li', 'Quanyu Dai'] | 2022-11-12 | null | null | null | null | ['selection-bias'] | ['natural-language-processing'] | [ 1.98658602e-03 -6.58771515e-01 -8.33122849e-01 -3.94627035e-01
-8.73024702e-01 -1.35653406e-01 2.20198646e-01 -1.22628674e-01
-7.09829107e-02 8.66039634e-01 -1.73215181e-01 -4.32002634e-01
-4.00972575e-01 -6.62503183e-01 -5.54225624e-01 -7.05173850e-01
2.35137358e-01 1.01782948e-01 1.73644885e-01 -1.09608307e-01
5.25471210e-01 1.70927688e-01 -1.44697905e+00 -1.58254877e-02
1.29879963e+00 1.50265372e+00 2.03944325e-01 -8.28407854e-02
1.26031414e-01 6.44430816e-01 -4.95432556e-01 -5.09556711e-01
3.46572012e-01 -2.15302259e-01 -1.62213057e-01 -1.81856647e-01
-1.46080861e-02 -2.19866142e-01 -3.55920583e-01 1.08447945e+00
4.57270443e-01 1.72772914e-01 7.21770942e-01 -1.41752529e+00
-8.49911571e-01 6.24309480e-01 -1.02678442e+00 1.92634389e-01
-5.70721412e-03 -7.76679115e-03 1.08523285e+00 -7.05371678e-01
-2.47308463e-02 1.08753562e+00 6.52784586e-01 2.57102489e-01
-1.15371954e+00 -1.12986505e+00 3.01098406e-01 2.51553625e-01
-1.44331324e+00 -1.74469337e-01 8.37464690e-01 -3.95816624e-01
2.54704356e-01 2.09202379e-01 2.23757308e-02 1.04288590e+00
4.02356923e-01 1.13301802e+00 1.26921797e+00 -1.24503419e-01
3.46010000e-01 5.82954109e-01 3.87394994e-01 1.73012525e-01
5.81556678e-01 2.70740002e-01 -2.30432764e-01 -3.48811656e-01
8.46585512e-01 3.25885326e-01 -9.27578285e-02 -6.90827370e-01
-9.16141331e-01 9.29762721e-01 1.28151938e-01 -1.37579978e-01
-2.40863085e-01 -2.06347406e-01 3.21492612e-01 3.84628952e-01
5.19047678e-01 2.21542612e-01 -5.28339744e-01 -1.35216907e-01
-9.36035395e-01 2.23323479e-01 5.30189157e-01 9.91257787e-01
3.68742555e-01 4.96254563e-02 -5.17650485e-01 1.07489288e+00
1.36133209e-01 4.28930551e-01 7.63138592e-01 -6.03571117e-01
7.77921736e-01 4.26316649e-01 3.74465942e-01 -1.05376983e+00
-7.36815780e-02 -9.15557623e-01 -1.11753726e+00 -8.21567047e-03
3.71941477e-01 -9.72302258e-02 -4.21587557e-01 1.67894185e+00
1.68322340e-01 7.23062828e-02 -1.58758059e-01 6.56834722e-01
9.71799716e-02 4.53118056e-01 3.72462235e-02 -7.50167429e-01
1.00197399e+00 -7.80322969e-01 -7.69409716e-01 -4.01554778e-02
3.66039306e-01 -6.30876541e-01 1.29841220e+00 6.24154449e-01
-1.05381870e+00 -6.32012844e-01 -1.06186032e+00 4.62534666e-01
1.31811795e-03 4.09179032e-01 6.25119209e-01 7.55507827e-01
-5.19286580e-02 7.58578897e-01 -4.13764060e-01 2.01325521e-01
4.95392233e-01 3.39258462e-01 3.03842008e-01 6.34089112e-02
-1.34246731e+00 5.71544707e-01 8.78285617e-02 -6.67588860e-02
-6.52795076e-01 -8.51725698e-01 -3.08259606e-01 1.02371871e-01
6.35582745e-01 -3.28078836e-01 1.24880576e+00 -7.57239103e-01
-1.49450278e+00 2.83899993e-01 1.49797443e-02 -4.32679027e-01
8.51069629e-01 -4.14802223e-01 -7.21924126e-01 -5.88297367e-01
-1.01216160e-01 -2.82650828e-01 9.61679518e-01 -1.08198059e+00
-7.61343360e-01 -6.39992952e-01 -2.22083434e-01 -5.64798191e-02
-5.50833523e-01 -2.49374762e-01 -2.12964699e-01 -9.62537408e-01
-5.21607772e-02 -5.88115752e-01 -2.41223529e-01 -3.96558493e-01
-4.29084271e-01 -3.25901359e-01 6.28644943e-01 -6.01340353e-01
1.92773008e+00 -2.21217847e+00 -1.57638311e-01 3.43790829e-01
1.41969090e-03 3.30367506e-01 1.01220161e-01 1.37255535e-01
-1.07168369e-01 1.46803424e-01 1.09238038e-02 2.22638529e-02
1.75776958e-01 -2.27493733e-01 -5.67480206e-01 4.18221921e-01
-1.83185935e-01 5.27177632e-01 -6.54652536e-01 -3.36734474e-01
5.40613420e-02 8.26400444e-02 -3.55289966e-01 2.62189567e-01
-7.47714043e-02 2.14029893e-01 -7.91521490e-01 7.45414138e-01
8.58835816e-01 -4.79781836e-01 -7.42634833e-02 -3.50912780e-01
-5.43974750e-02 8.03031027e-02 -1.47273517e+00 9.85459447e-01
-6.35729611e-01 -4.71285507e-02 -1.29463643e-01 -1.22141254e+00
1.16791749e+00 -4.79162373e-02 4.76619810e-01 -8.04415822e-01
2.20396087e-01 3.71278673e-01 -2.85475045e-01 -3.19883347e-01
3.74402881e-01 -9.92320478e-02 -8.06715935e-02 3.95590991e-01
-5.11097431e-01 4.56149668e-01 -1.88794121e-01 -1.62418023e-01
8.02425504e-01 -1.73556581e-01 4.25688118e-01 -1.19637683e-01
7.82093227e-01 -5.04438400e-01 8.78261566e-01 7.72284746e-01
-3.63337606e-01 2.69813687e-01 5.24127960e-01 -1.33180574e-01
-9.84185040e-01 -1.04969680e+00 -4.79870349e-01 9.33279634e-01
4.04421777e-01 1.06646931e-02 -2.79651642e-01 -8.70561659e-01
3.99493337e-01 9.99787867e-01 -3.27025980e-01 -5.04657328e-01
-1.93404868e-01 -1.14168608e+00 1.54125705e-01 6.21319532e-01
5.96641481e-01 -5.28388202e-01 1.72369212e-01 1.29938200e-01
-2.93644052e-02 -7.78132379e-01 -5.93782842e-01 -1.56047624e-02
-9.26917970e-01 -8.37162495e-01 -7.54623652e-01 -5.80278277e-01
3.07638377e-01 5.52025437e-01 9.07959223e-01 -3.50969106e-01
-1.18595444e-01 -5.01573347e-02 -2.38218665e-01 -2.54207611e-01
-2.10397065e-01 2.74710596e-01 2.57052243e-01 2.11347416e-01
4.58767563e-01 -5.04408300e-01 -6.92719042e-01 9.08515990e-01
-6.61549807e-01 -4.52718675e-01 9.36410308e-01 9.97726321e-01
8.31499338e-01 4.87463504e-01 1.26232553e+00 -9.91836786e-01
9.65516210e-01 -8.67678523e-01 -8.74011934e-01 4.86874998e-01
-1.33503520e+00 4.71188575e-02 6.90718770e-01 -7.73793936e-01
-1.24167919e+00 -2.16948345e-01 5.48046231e-02 -4.52317894e-01
2.23662868e-01 3.88112336e-01 -3.63827378e-01 1.28576636e-01
3.63621265e-01 3.58581543e-01 3.44183184e-02 -6.94126546e-01
2.06904352e-01 1.26854515e+00 2.83183157e-01 -6.10917330e-01
7.60111451e-01 1.27969027e-01 -9.91427004e-02 -3.74621660e-01
-1.07947457e+00 -5.55946171e-01 -1.58996761e-01 1.14349805e-01
2.28938818e-01 -1.05249846e+00 -8.12416971e-01 4.31868732e-01
-4.47326958e-01 -7.52487732e-03 -8.25856328e-02 5.78841925e-01
-6.90101862e-01 5.06276250e-01 -3.36574674e-01 -1.16783452e+00
-4.04427439e-01 -1.07483077e+00 6.23145044e-01 2.06193000e-01
1.86357945e-01 -8.13440263e-01 -7.63643384e-02 5.18128097e-01
4.39074636e-01 -1.44456565e-01 1.09193480e+00 -7.22447634e-01
-4.40590441e-01 -5.40647507e-01 -4.21721131e-01 6.71371639e-01
2.72790819e-01 -3.04569691e-01 -6.03584290e-01 -4.89295900e-01
2.22429261e-02 -3.23188186e-01 6.87233329e-01 5.14453292e-01
1.75800312e+00 -1.90628380e-01 -3.63218486e-01 3.76041383e-01
1.48405147e+00 2.82598972e-01 6.50473893e-01 3.98182094e-01
4.62327003e-01 2.98507422e-01 1.06249344e+00 8.10880542e-01
1.82594731e-01 8.71341884e-01 3.04336160e-01 2.05468088e-01
3.22944015e-01 -4.51772630e-01 4.35469985e-01 6.51550293e-01
1.47848979e-01 -1.12431839e-01 -2.40354851e-01 1.96872622e-01
-2.03017545e+00 -9.20937598e-01 -7.82232061e-02 3.04671764e+00
9.42537129e-01 4.00889218e-01 4.07000810e-01 1.77838400e-01
9.99058127e-01 -1.26250073e-01 -1.03925538e+00 -6.92703128e-02
8.15967992e-02 -2.01367468e-01 7.61460662e-01 -4.10312191e-02
-1.01127863e+00 4.95779067e-01 6.21835756e+00 1.46439016e+00
-8.01284015e-01 8.14456400e-03 8.40206683e-01 -7.07686841e-02
-2.23022655e-01 -2.24653468e-01 -1.17479444e+00 9.59349513e-01
6.61735713e-01 -4.69681650e-01 5.37839532e-01 1.16548431e+00
4.33421522e-01 2.57042736e-01 -9.81046855e-01 1.16924858e+00
-6.62055686e-02 -1.02194130e+00 -6.84476495e-02 9.47219878e-02
7.87771046e-01 -3.29973608e-01 5.70622921e-01 7.05910087e-01
2.31054723e-01 -7.60033429e-01 3.78372580e-01 5.93672991e-01
6.82829678e-01 -9.30215478e-01 8.24728370e-01 4.12618965e-01
-9.18392718e-01 -6.12177193e-01 -5.65395057e-01 1.65330723e-01
9.33602639e-03 1.16594374e+00 -1.61426529e-01 5.82660913e-01
5.36460400e-01 7.82752097e-01 -5.07573187e-01 1.30935442e+00
3.46892439e-02 6.62990451e-01 -5.73151570e-04 -2.03155965e-01
-3.31178367e-01 -2.87407368e-01 4.59986538e-01 7.15394557e-01
3.73255908e-01 -2.31704921e-01 2.93040909e-02 9.36985135e-01
-1.85546756e-01 3.78661841e-01 -1.86734661e-01 1.01992518e-01
8.53424668e-01 1.08452535e+00 -1.10287860e-01 -4.32887375e-02
-4.75621581e-01 5.06545484e-01 1.48236394e-01 4.03262258e-01
-1.08276689e+00 -6.46893501e-01 6.20449185e-01 2.11261183e-01
4.00058329e-01 4.47981693e-02 -4.34621632e-01 -1.26109576e+00
1.95024684e-01 -9.87167239e-01 3.90515000e-01 -3.09045792e-01
-1.86966860e+00 2.89504472e-02 -1.08365148e-01 -1.61990261e+00
1.62808418e-01 -3.57298553e-01 -6.63505018e-01 5.83602011e-01
-1.68126547e+00 -7.85341680e-01 -2.12311164e-01 5.02416134e-01
5.64282596e-01 -4.40236956e-01 1.65190175e-01 5.98722160e-01
-1.01789486e+00 1.15687466e+00 8.43214035e-01 -3.07420701e-01
9.82347190e-01 -9.89745557e-01 -1.03511132e-01 3.52396488e-01
-3.69183928e-01 7.09020019e-01 6.01638019e-01 -6.11436665e-01
-1.20137703e+00 -1.20826578e+00 3.53056192e-01 -1.81200057e-01
8.10867071e-01 -1.23702765e-01 -9.00625587e-01 3.61525595e-01
-4.87683117e-01 -2.15426162e-01 8.24833572e-01 3.80699843e-01
-3.16388100e-01 -7.50287533e-01 -1.13356006e+00 4.83805418e-01
8.14192653e-01 -1.89931244e-01 -3.79763901e-01 4.09593821e-01
7.82336652e-01 -6.20846897e-02 -1.06964755e+00 5.74594617e-01
6.82592988e-01 -8.00402701e-01 1.11901128e+00 -6.51896179e-01
3.56685340e-01 -7.07900971e-02 -3.07819575e-01 -1.11529529e+00
-5.06447017e-01 -5.25825441e-01 -3.84701401e-01 1.60047460e+00
4.13348168e-01 -7.59373963e-01 8.55722189e-01 5.83947003e-01
2.20704988e-01 -9.26279426e-01 -7.53902376e-01 -1.31879497e+00
2.64086813e-01 -2.87361562e-01 6.53976500e-01 7.24058986e-01
-3.35748270e-02 2.67486900e-01 -8.00630689e-01 -3.87695469e-02
9.55166399e-01 3.78458023e-01 6.91364825e-01 -1.22728443e+00
-5.75393796e-01 -4.96721655e-01 -9.35638696e-02 -1.23011041e+00
4.78008427e-02 -5.29488087e-01 -1.58964992e-01 -9.58919287e-01
4.43284422e-01 -8.21841061e-01 -8.09624851e-01 5.60720414e-02
-3.63610148e-01 -2.61292368e-01 -2.11378664e-01 5.13562500e-01
-7.04126298e-01 8.11818182e-01 1.08192587e+00 1.79869160e-02
-4.98626918e-01 8.66650581e-01 -9.19968426e-01 5.30774832e-01
8.62889886e-01 -3.01617593e-01 -5.31034112e-01 1.21784568e-01
1.84578374e-01 8.89127254e-02 1.49091780e-01 -7.88873434e-01
-3.85304680e-03 -4.05418068e-01 2.84924209e-01 -6.50460720e-01
-3.79384831e-02 -7.38097727e-01 -9.90388244e-02 2.34385192e-01
-5.27414382e-01 -1.31475836e-01 -2.43836790e-01 1.03622818e+00
-6.43819198e-02 -1.96721733e-01 8.17493260e-01 1.71331555e-01
-2.75286585e-01 3.86522084e-01 -4.15773876e-02 1.17378421e-01
1.09327340e+00 -3.24593931e-02 -3.44244212e-01 -4.18480992e-01
-2.64195412e-01 4.68914121e-01 1.49197713e-01 5.05694807e-01
3.64885300e-01 -1.62309194e+00 -4.54135329e-01 1.06137536e-01
1.71287760e-01 -4.18182999e-01 3.30356359e-01 9.44087744e-01
3.04421455e-01 5.32457590e-01 4.91351895e-02 -2.31188312e-01
-9.43357050e-01 9.21781301e-01 3.42721194e-02 -5.05614221e-01
-1.67077094e-01 4.24002469e-01 1.46594122e-01 -1.27075762e-01
4.98408496e-01 -6.80777654e-02 -2.15608239e-01 -5.67448027e-02
5.34557879e-01 8.66796970e-01 3.52688916e-02 -1.64626107e-01
-7.89596960e-02 4.12596732e-01 -4.36396420e-01 3.16708803e-01
1.09018731e+00 -4.49921608e-01 2.71120667e-01 5.14746606e-01
1.03354204e+00 2.48487517e-02 -1.23374152e+00 -5.08104742e-01
-8.53736401e-02 -7.24159479e-01 3.17808419e-01 -8.19075584e-01
-1.07241666e+00 7.52475202e-01 7.71261811e-01 3.21659446e-01
1.11735332e+00 -3.02341878e-01 1.03338814e+00 1.73580885e-01
5.29034913e-01 -1.35792291e+00 2.69908875e-01 -4.91965860e-02
4.66142565e-01 -1.34406352e+00 2.45213225e-01 -3.70903254e-01
-7.94394135e-01 7.52820849e-01 6.86543167e-01 1.04476186e-03
7.73633659e-01 -2.11955518e-01 -3.46990079e-01 3.73911679e-01
-5.89466393e-01 2.03067034e-01 1.83565468e-01 3.92695069e-01
3.02473575e-01 8.19737837e-02 -7.70260692e-01 1.39294720e+00
2.00636476e-01 1.94171071e-01 2.20094115e-01 5.94047725e-01
-3.90186936e-01 -1.31438446e+00 -1.82637438e-01 8.27467501e-01
-5.84140778e-01 8.74058083e-02 7.30193630e-02 6.33476853e-01
-1.00731872e-01 9.19362962e-01 -2.91277647e-01 -7.05995798e-01
5.41180491e-01 -2.58184642e-01 1.11655004e-01 -1.40004247e-01
-7.56547973e-02 3.88856535e-03 -2.59229690e-01 -3.49161863e-01
-3.04435343e-02 -8.31661820e-01 -7.43993282e-01 -2.96755284e-01
-9.11309958e-01 1.70817941e-01 5.02153039e-01 8.63952816e-01
3.94599885e-01 5.35710931e-01 1.31241012e+00 -2.29156047e-01
-1.66098654e+00 -7.72111893e-01 -1.06201231e+00 4.22061831e-01
1.07960731e-01 -1.05326593e+00 -6.37403846e-01 -4.16551977e-01] | [9.698671340942383, 5.320231914520264] |
b5b4a879-dbb3-4948-b6bb-1c3022715dd8 | gpt-self-supervision-for-a-better-data | 2306.04349 | null | https://arxiv.org/abs/2306.04349v2 | https://arxiv.org/pdf/2306.04349v2.pdf | GPT Self-Supervision for a Better Data Annotator | The task of annotating data into concise summaries poses a significant challenge across various domains, frequently requiring the allocation of significant time and specialized knowledge by human experts. Despite existing efforts to use large language models for annotation tasks, significant problems such as limited applicability to unlabeled data, the absence of self-supervised methods, and the lack of focus on complex structured data still persist. In this work, we propose a GPT self-supervision annotation method, which embodies a generating-recovering paradigm that leverages the one-shot learning capabilities of the Generative Pretrained Transformer (GPT). The proposed approach comprises a one-shot tuning phase followed by a generation phase. In the one-shot tuning phase, we sample a data from the support set as part of the prompt for GPT to generate a textual summary, which is then used to recover the original data. The alignment score between the recovered and original data serves as a self-supervision navigator to refine the process. In the generation stage, the optimally selected one-shot sample serves as a template in the prompt and is applied to generating summaries from challenging datasets. The annotation performance is evaluated by tuning several human feedback reward networks and by calculating alignment scores between original and recovered data at both sentence and structure levels. Our self-supervised annotation method consistently achieves competitive scores, convincingly demonstrating its robust strength in various data-to-summary annotation tasks. | ['Chang Xu', 'Yanxi Li', 'Xiaohuan Pei'] | 2023-06-07 | null | null | null | null | ['one-shot-learning'] | ['methodology'] | [ 6.84778929e-01 3.94346923e-01 -3.21193486e-01 -4.56950337e-01
-1.33275187e+00 -6.04498804e-01 7.30879247e-01 4.63516384e-01
-2.36241221e-01 8.46333265e-01 4.70100582e-01 1.82786193e-02
1.12988897e-01 -4.88257349e-01 -4.96648490e-01 -5.97292662e-01
4.77200598e-01 8.10172975e-01 5.22192195e-02 -2.25673303e-01
3.45156938e-01 -9.12659243e-02 -1.22487068e+00 2.73628265e-01
1.25203884e+00 9.05487597e-01 4.97363865e-01 4.50600743e-01
-4.11333591e-01 7.31722236e-01 -7.75602520e-01 -4.86389697e-01
-3.89294736e-02 -7.04122007e-01 -7.62100935e-01 2.47068301e-01
1.48717269e-01 -6.95687979e-02 4.04699109e-02 8.51017654e-01
7.15871513e-01 3.64988029e-01 4.99301195e-01 -1.07592750e+00
-6.29622161e-01 9.83723164e-01 -3.68695468e-01 1.30407482e-01
4.96635348e-01 1.63132891e-01 1.25660157e+00 -1.11715388e+00
7.38870442e-01 8.99789929e-01 4.68987018e-01 6.33063257e-01
-1.15734255e+00 -4.44467723e-01 -1.92977972e-02 -1.50353357e-01
-1.00258756e+00 -6.52113080e-01 8.15727592e-01 -3.83005410e-01
9.26853061e-01 3.82450148e-02 4.25553173e-01 1.14130569e+00
-1.64004833e-01 8.05862665e-01 5.61046362e-01 -3.79345268e-01
3.91597599e-01 2.45012760e-01 2.07740396e-01 4.89109516e-01
2.87200212e-02 -3.53503466e-01 -7.08321333e-01 -1.86956525e-01
3.02831531e-01 -1.81748524e-01 -1.31288961e-01 -9.17490795e-02
-1.06941783e+00 7.97189295e-01 2.42980927e-01 2.79702067e-01
-6.39007151e-01 -2.03945622e-01 6.11300647e-01 4.04926240e-02
6.89785302e-01 6.22836709e-01 -1.08052008e-01 -2.11703613e-01
-1.19281065e+00 3.26763213e-01 8.15035582e-01 1.08355999e+00
6.30990207e-01 1.20128833e-01 -8.97611976e-01 8.25200856e-01
2.48067994e-02 1.51355535e-01 6.73811615e-01 -6.82085276e-01
9.67486262e-01 9.59257424e-01 1.51465371e-01 -6.82915390e-01
-7.60844797e-02 -5.48688114e-01 -8.48715723e-01 -3.34436268e-01
-9.71499085e-03 -2.29782298e-01 -6.38204873e-01 1.73215663e+00
4.09436435e-01 -1.30263820e-01 3.56256694e-01 8.09962690e-01
9.61000264e-01 7.42070198e-01 1.26534462e-01 -4.09146607e-01
1.29317331e+00 -1.20395231e+00 -6.95001364e-01 -4.71418649e-01
6.48854852e-01 -6.57093644e-01 1.13976109e+00 1.30329847e-01
-1.12589157e+00 -6.18915379e-01 -1.01934659e+00 -2.18986586e-01
6.55190572e-02 5.23092687e-01 1.63982026e-02 9.89969596e-02
-8.30914915e-01 5.90074182e-01 -5.90673268e-01 -3.88036221e-01
5.28953373e-01 3.62492017e-02 -3.31775367e-01 -1.01848401e-01
-1.05884755e+00 6.90568566e-01 7.10860431e-01 -1.28468558e-01
-9.45577919e-01 -6.50957108e-01 -9.92133856e-01 3.46589863e-01
6.55780673e-01 -6.60233438e-01 1.49589813e+00 -8.41466963e-01
-1.38400555e+00 6.33272231e-01 -2.45313138e-01 -5.72771311e-01
5.34408391e-01 -2.58678377e-01 -3.30577865e-02 5.13393208e-02
5.20321786e-01 5.14634311e-01 7.55065262e-01 -1.04322469e+00
-6.01949990e-01 -2.48874947e-01 -1.32176414e-01 4.28335547e-01
-3.35331649e-01 1.46824978e-02 -4.31556374e-01 -7.46274054e-01
-1.81089401e-01 -8.29013348e-01 -1.90695822e-01 -4.76290494e-01
-7.06738532e-01 -5.59391618e-01 7.22051024e-01 -7.36733437e-01
1.45876026e+00 -2.11052608e+00 2.78117120e-01 -1.48950562e-01
2.81355102e-02 3.92376274e-01 -2.01888174e-01 6.94984078e-01
-1.33506373e-01 -3.57373990e-02 -4.46139127e-01 -7.65129924e-01
-7.48289675e-02 -8.31558648e-03 -6.10420167e-01 -8.94058719e-02
6.99365795e-01 9.16575611e-01 -1.22449386e+00 -5.97952187e-01
-2.69769460e-01 1.29801035e-01 -3.34747076e-01 7.80142546e-01
-5.73713422e-01 5.77196598e-01 -5.10942578e-01 2.92543352e-01
8.12966079e-02 -5.79201579e-01 9.31175873e-02 -2.23168135e-01
-1.81517258e-01 6.04334414e-01 -8.84926379e-01 1.88672435e+00
-3.26873004e-01 3.62237751e-01 -3.45961750e-01 -1.00098372e+00
1.30613303e+00 5.91763377e-01 1.59175962e-01 -5.47717571e-01
1.01700239e-01 2.32920662e-01 -2.15349540e-01 -6.04126573e-01
9.16604936e-01 -1.43607691e-01 -3.43897611e-01 8.59583855e-01
3.15510482e-01 -1.85376614e-01 6.29319966e-01 5.60791314e-01
1.11712623e+00 3.31077874e-01 4.49469715e-01 1.45502225e-01
5.15691280e-01 2.90732652e-01 6.14350319e-01 5.83306015e-01
2.42596015e-01 6.95187688e-01 6.43676639e-01 -1.63377464e-01
-1.33474660e+00 -4.56933886e-01 4.73108411e-01 1.13158214e+00
-1.94866419e-01 -5.54841280e-01 -7.89227486e-01 -7.65216649e-01
-3.11572939e-01 1.19502890e+00 -5.48812509e-01 -3.97597551e-01
-3.51579070e-01 -4.57029074e-01 5.28714538e-01 6.73018038e-01
3.55836540e-01 -1.35612559e+00 -7.57186830e-01 3.57857108e-01
-4.71082270e-01 -1.04799128e+00 -7.05395699e-01 4.49856520e-01
-7.83766448e-01 -7.93563306e-01 -5.50490677e-01 -8.03265691e-01
7.15908051e-01 1.81520864e-01 9.65655029e-01 -7.65866637e-02
8.24036747e-02 -1.25211533e-02 -4.56462443e-01 -1.96008384e-01
-8.32182944e-01 4.21979964e-01 -1.80856913e-01 1.43964946e-01
5.18968031e-02 -3.46437395e-01 -3.62465352e-01 4.07627001e-02
-9.30171013e-01 3.35214049e-01 6.40688121e-01 9.51346397e-01
7.41536617e-01 -2.70374000e-01 9.39872742e-01 -1.12825871e+00
1.01398385e+00 -6.93124413e-01 -4.15590018e-01 3.39156479e-01
-5.96182823e-01 3.25427681e-01 8.68329048e-01 -4.13581669e-01
-1.26397657e+00 1.16402172e-01 7.81500190e-02 -3.89665008e-01
1.10589825e-01 7.92142510e-01 -1.97951674e-01 7.75462091e-01
7.59009659e-01 3.73268038e-01 1.83814783e-02 -3.70764583e-01
3.09494704e-01 7.64677405e-01 7.71905363e-01 -4.70352232e-01
7.95040667e-01 -7.50971735e-02 -5.54535806e-01 -4.11661834e-01
-1.39562142e+00 -3.29990685e-01 -5.70858002e-01 8.04081708e-02
7.20510304e-01 -8.68138969e-01 -3.03390380e-02 4.90954034e-02
-1.19562995e+00 -3.04843307e-01 -6.61347628e-01 7.73292929e-02
-5.52284539e-01 2.37285197e-01 -3.15781474e-01 -8.02382052e-01
-9.59322691e-01 -1.00079834e+00 1.18229842e+00 4.53862160e-01
-5.03520191e-01 -7.39782691e-01 2.18704551e-01 6.29037797e-01
3.09505314e-01 1.14848070e-01 9.48101342e-01 -1.31166565e+00
-4.28372949e-01 -5.30907452e-01 -9.46539193e-02 2.93505192e-01
1.28989592e-01 -1.04368515e-01 -8.64150643e-01 -1.60308868e-01
-6.02885559e-02 -6.92086160e-01 5.19380629e-01 -2.76291296e-02
9.66361463e-01 -5.05818725e-01 -1.86098769e-01 2.27009475e-01
1.03729153e+00 9.29799601e-02 2.56525576e-01 1.87475495e-02
5.56500375e-01 6.88931167e-01 8.88286710e-01 5.78467011e-01
3.73011023e-01 5.36991239e-01 2.29671001e-01 1.21912569e-01
-9.84053835e-02 -6.78332448e-01 2.00134382e-01 1.02036250e+00
1.45290822e-01 -4.88307923e-01 -7.49229372e-01 5.77014208e-01
-1.95302093e+00 -1.03666079e+00 2.42821559e-01 2.20753336e+00
1.16190422e+00 3.40618759e-01 2.41351891e-02 1.26911372e-01
7.98330307e-01 2.22802088e-01 -6.79702699e-01 -1.65613264e-01
-4.47989628e-02 -2.04617679e-02 -2.05492020e-01 1.76219448e-01
-8.36993575e-01 9.27552819e-01 4.95941830e+00 7.58640587e-01
-9.86809671e-01 5.32991812e-02 6.13475859e-01 -9.55719687e-03
-4.12154615e-01 2.64441490e-01 -8.29233527e-01 4.93463606e-01
1.06359684e+00 -6.45794094e-01 1.81032076e-01 8.88227999e-01
4.55521017e-01 -7.62888342e-02 -1.18371892e+00 6.49496734e-01
2.53919005e-01 -1.30255890e+00 1.56413153e-01 -2.90132135e-01
6.50412261e-01 -1.55666053e-01 -2.66578436e-01 4.65903074e-01
3.37453872e-01 -7.71127522e-01 7.75699615e-01 4.21878487e-01
9.49107766e-01 -5.60007691e-01 7.69581437e-01 6.32967889e-01
-1.05814445e+00 -8.48704278e-02 -1.73800766e-01 1.65135160e-01
4.12437946e-01 6.21141136e-01 -1.40894508e+00 6.58316076e-01
2.35947087e-01 7.90122390e-01 -4.23584610e-01 8.91469061e-01
-4.57843781e-01 6.15373075e-01 1.14050880e-01 -5.04882112e-02
2.08789364e-01 -1.80048719e-01 5.68453431e-01 1.12711692e+00
3.11905414e-01 4.95046042e-02 4.62897420e-01 1.00065064e+00
-3.91543359e-01 2.82265842e-01 -5.01457989e-01 -4.39269543e-01
7.03261971e-01 1.50395191e+00 -6.32216513e-01 -6.63109660e-01
-1.54079214e-01 9.10722733e-01 5.19472361e-01 1.90113366e-01
-5.93877435e-01 -3.95404905e-01 -5.04390933e-02 -5.28213568e-02
2.70786941e-01 1.07168689e-01 -3.66055578e-01 -1.05667019e+00
1.25789970e-01 -9.30119097e-01 3.90633702e-01 -9.32875574e-01
-1.09320199e+00 8.85606408e-01 -9.02859271e-02 -1.35522771e+00
-5.49816012e-01 3.55400771e-01 -9.35455203e-01 9.00738180e-01
-1.20378864e+00 -1.00950539e+00 -6.11925960e-01 1.51690885e-01
1.07957923e+00 -2.93603539e-01 8.53115797e-01 1.03576034e-01
-8.30639362e-01 5.34168601e-01 -2.23692313e-01 1.20861635e-01
7.08786845e-01 -1.33071208e+00 6.21203125e-01 9.34329510e-01
-1.18873641e-02 5.91064453e-01 7.85614908e-01 -1.00245869e+00
-1.12523544e+00 -1.22358942e+00 1.10760021e+00 -2.54449904e-01
6.51235402e-01 -2.53663510e-01 -1.25171816e+00 4.84411150e-01
2.49353945e-01 -9.36753228e-02 7.83167779e-01 -3.00530255e-01
-2.13559002e-01 3.21781859e-02 -8.58554065e-01 3.51931095e-01
7.57520258e-01 -3.86446357e-01 -8.12348008e-01 2.52149969e-01
9.03003037e-01 -5.33870518e-01 -6.31482303e-01 7.10601434e-02
2.60112137e-01 -3.72102827e-01 3.98107439e-01 -6.28995061e-01
8.21266353e-01 -2.02677369e-01 1.47199064e-01 -1.37819147e+00
-6.67290613e-02 -8.55907917e-01 -8.99731070e-02 1.69716823e+00
5.91239870e-01 -1.80230170e-01 5.02739131e-01 5.74536443e-01
-4.70512748e-01 -7.40028560e-01 -7.53456414e-01 -4.58073735e-01
-4.64707553e-01 -5.06723719e-03 5.09831607e-01 7.25384057e-01
9.12046507e-02 1.08610475e+00 -3.67018670e-01 -2.07072824e-01
4.57457632e-01 2.78287798e-01 8.59482944e-01 -1.26604915e+00
-2.80569434e-01 -1.85546085e-01 2.92590320e-01 -7.98209608e-01
1.09786682e-01 -1.13612950e+00 4.59691286e-01 -1.76541102e+00
3.90059888e-01 -4.80499268e-01 9.43723917e-02 7.06097245e-01
-4.87974226e-01 -1.07761674e-01 2.87195053e-02 4.41016525e-01
-7.84452260e-01 7.99325287e-01 9.07830715e-01 -1.32732183e-01
-4.81497347e-01 1.16718709e-01 -8.87253642e-01 3.93121094e-01
6.93565190e-01 -6.02359116e-01 -6.29855037e-01 -3.80979747e-01
1.51094154e-01 3.91639680e-01 4.86733206e-02 -8.58367503e-01
3.03016037e-01 -1.58662926e-02 1.80758357e-01 -6.73974872e-01
1.25905305e-01 -4.16369617e-01 7.23831952e-02 1.98405683e-01
-7.96257079e-01 1.49718180e-01 -6.25898391e-02 5.82782030e-01
-1.92337096e-01 -5.75278938e-01 5.54091752e-01 -1.54795229e-01
-2.73807406e-01 1.98937625e-01 1.09584577e-01 3.60902011e-01
8.03757012e-01 -1.36940911e-01 -3.96228880e-01 -3.11210543e-01
-5.14062047e-01 3.97392780e-01 4.45161194e-01 4.82971072e-01
3.78865212e-01 -1.17716634e+00 -8.65135610e-01 6.84912279e-02
3.21958154e-01 4.94811147e-01 1.14013158e-01 7.88459778e-01
-4.53933999e-02 3.44909549e-01 6.17980286e-02 -4.69694138e-01
-9.99502599e-01 3.89475048e-01 -4.92259227e-02 -6.62309766e-01
-7.24957585e-01 4.81619507e-01 -5.96993007e-02 -1.72367498e-01
2.47753486e-01 -7.98171461e-02 -3.71507734e-01 4.16292578e-01
6.33700788e-01 2.10128456e-01 2.82598406e-01 -4.61979270e-01
4.43634614e-02 -1.00203194e-02 -2.70127922e-01 -2.14820370e-01
1.58845818e+00 -2.62533929e-02 7.17868358e-02 5.61145008e-01
8.05238426e-01 -1.48507550e-01 -1.38864243e+00 -5.01891017e-01
3.47632945e-01 -4.04747576e-02 -3.24576765e-01 -8.06997538e-01
-7.43811131e-01 6.94614410e-01 1.38649447e-02 2.17739493e-01
9.10272419e-01 1.56363085e-01 8.41532111e-01 3.35665226e-01
3.15156952e-03 -9.87287760e-01 5.44392169e-01 5.26368797e-01
1.03579640e+00 -1.24692070e+00 -1.66892737e-01 -2.16968685e-01
-1.12513030e+00 9.41394508e-01 7.32693613e-01 4.76436988e-02
-2.10399374e-01 1.65120468e-01 -1.44425854e-01 -1.19834356e-01
-1.09898233e+00 5.77537715e-02 3.40168923e-01 3.73773515e-01
5.18640518e-01 -2.99269706e-01 -2.83695161e-01 9.92128015e-01
-3.17316055e-01 9.66100991e-02 4.87134248e-01 1.01404119e+00
-5.21702766e-01 -9.72204030e-01 -6.22722432e-02 5.69380820e-01
-2.23387390e-01 -6.36926591e-02 -5.11815488e-01 2.02568591e-01
-2.49144360e-01 7.70810723e-01 5.13756415e-03 -1.95578605e-01
4.23034966e-01 3.78838003e-01 -9.63971019e-02 -1.19677842e+00
-8.73473704e-01 1.33615628e-01 2.60450721e-01 -1.58259153e-01
-1.88296065e-01 -6.55911922e-01 -1.39346147e+00 3.25242072e-01
-4.53994125e-01 5.37034392e-01 5.94075501e-01 1.07721126e+00
6.14463866e-01 5.54610610e-01 6.61034226e-01 -7.50582099e-01
-8.42370868e-01 -1.35870552e+00 3.01664672e-03 5.44448495e-01
1.72057107e-01 -4.44713056e-01 -1.14086926e-01 2.29706138e-01] | [11.74892807006836, 8.874403953552246] |
24c6c9cd-0b05-4361-89c0-1f8677de22f8 | arts-eliminating-inconsistency-between-text | 2110.10405 | null | https://arxiv.org/abs/2110.10405v1 | https://arxiv.org/pdf/2110.10405v1.pdf | ARTS: Eliminating Inconsistency between Text Detection and Recognition with Auto-Rectification Text Spotter | Recent approaches for end-to-end text spotting have achieved promising results. However, most of the current spotters were plagued by the inconsistency problem between text detection and recognition. In this work, we introduce and prove the existence of the inconsistency problem and analyze it from two aspects: (1) inconsistency of text recognition features between training and testing, and (2) inconsistency of optimization targets between text detection and recognition. To solve the aforementioned issues, we propose a differentiable Auto-Rectification Module (ARM) together with a new training strategy to enable propagating recognition loss back into detection branch, so that our detection branch can be jointly optimized by detection and recognition targets, which largely alleviates the inconsistency problem between text detection and recognition. Based on these designs, we present a simple yet robust end-to-end text spotting framework, termed Auto-Rectification Text Spotter (ARTS), to detect and recognize arbitrarily-shaped text in natural scenes. Extensive experiments demonstrate the superiority of our method. In particular, our ARTS-S achieves 77.1% end-to-end text spotting F-measure on Total-Text at a competitive speed of 10.5 FPS, which significantly outperforms previous methods in both accuracy and inference speed. | ['Tong Lu', 'Cong Yao', 'Zhibo Yang', 'Wenhai Wang', 'Jun Tang', 'Humen Zhong'] | 2021-10-20 | null | null | null | null | ['text-spotting'] | ['computer-vision'] | [ 5.12558281e-01 -5.92596471e-01 7.83351138e-02 -2.85953015e-01
-7.79924870e-01 -4.09534812e-01 3.66054237e-01 -4.46245253e-01
-2.80804843e-01 2.42134362e-01 -7.16115832e-02 -3.58004302e-01
1.11740299e-01 -4.04572785e-01 -5.82495630e-01 -5.82041323e-01
8.71273100e-01 3.95208329e-01 2.32196376e-01 -4.07361500e-02
5.01135170e-01 2.99462885e-01 -1.43475556e+00 2.29937375e-01
1.21382213e+00 9.39728022e-01 4.42024946e-01 9.38579917e-01
-3.34507823e-01 5.15487373e-01 -6.34696901e-01 -3.49675298e-01
1.36187732e-01 -4.10838991e-01 -2.79351711e-01 4.29167926e-01
6.81381643e-01 -7.19471455e-01 -5.87221980e-01 9.11188304e-01
7.02609956e-01 -8.69811177e-02 4.36337501e-01 -1.06463838e+00
-7.29510307e-01 3.26187372e-01 -9.69775558e-01 -9.39135626e-02
3.22756022e-01 2.47294232e-01 1.00841415e+00 -1.42891145e+00
1.88223794e-01 1.03125429e+00 7.43110478e-01 5.52084446e-01
-8.45096409e-01 -5.51960707e-01 1.87187076e-01 -1.57543018e-01
-1.44341052e+00 -6.43343866e-01 4.97425407e-01 -3.60501021e-01
7.94123709e-01 4.53606457e-01 3.18950862e-01 8.59756529e-01
1.29870540e-02 1.49167514e+00 5.74760675e-01 -5.81283987e-01
-1.96788713e-01 1.35052940e-02 3.00071668e-02 9.59700763e-01
1.92194656e-01 -1.38671562e-01 -7.20107019e-01 2.96743572e-01
9.23513114e-01 1.48218378e-01 -3.50720733e-01 5.82131557e-02
-1.42113650e+00 4.98682380e-01 1.11392826e-01 1.09849311e-01
5.61638437e-02 1.81258470e-02 4.46865022e-01 1.52229279e-01
4.28827286e-01 1.64946925e-03 -1.82013810e-01 -2.55561560e-01
-1.12789178e+00 -1.58949420e-01 4.63690788e-01 1.03062081e+00
3.19112778e-01 2.64912426e-01 -4.83518869e-01 1.09012485e+00
4.63469774e-01 1.07900000e+00 6.58976436e-01 2.97248214e-02
9.01270449e-01 7.45752871e-01 -2.26553679e-02 -1.03248215e+00
-2.22174287e-01 -4.24855620e-01 -8.69367719e-01 -7.78430840e-03
3.40767145e-01 -1.39391840e-01 -9.52772915e-01 1.09625149e+00
1.91545129e-01 1.34185910e-01 -1.07302368e-01 1.31194305e+00
7.76199818e-01 6.63706243e-01 -4.22540843e-01 8.19561034e-02
1.36344886e+00 -1.20757842e+00 -7.48726189e-01 -3.74107093e-01
8.27566862e-01 -1.05827868e+00 1.39891303e+00 3.49544466e-01
-8.67715418e-01 -5.15210688e-01 -1.01147115e+00 -2.64122576e-01
8.87302961e-03 9.85573590e-01 1.08015701e-01 7.88680673e-01
-7.95412660e-01 1.92075446e-01 -7.17420876e-01 -3.18238080e-01
3.42565686e-01 5.13480783e-01 2.10717432e-02 1.04491793e-01
-7.04362035e-01 4.11807984e-01 6.01704605e-02 3.06345731e-01
-3.63923788e-01 -3.20535958e-01 -5.51002324e-01 1.21535070e-01
5.18745661e-01 -4.86768126e-01 1.15372455e+00 -1.07494998e+00
-1.80099976e+00 7.83817470e-01 -4.28546280e-01 -1.20314837e-01
8.53154898e-01 -5.78028738e-01 -4.60928112e-01 -8.03347304e-02
6.74839616e-02 3.83393139e-01 1.24386144e+00 -8.24155033e-01
-9.60935593e-01 -5.03136218e-01 -6.84888542e-01 4.45038319e-01
-6.27869129e-01 6.36177510e-02 -9.22406554e-01 -9.80424404e-01
1.88782588e-01 -6.04543209e-01 3.29612225e-01 3.93687725e-01
-7.13511467e-01 -1.35229990e-01 1.28554130e+00 -6.48971915e-01
1.32492805e+00 -2.25472260e+00 -3.19321640e-02 -1.71150029e-01
2.48274386e-01 5.45398474e-01 -2.00821713e-01 7.66355097e-02
2.79659092e-01 2.24343501e-02 -6.71262667e-02 -7.62095571e-01
3.24842930e-02 -1.49770811e-01 -9.33497250e-01 6.78717017e-01
2.01512128e-01 1.00027227e+00 -5.97185373e-01 -5.51072538e-01
5.57599604e-01 5.41983128e-01 -2.80855298e-01 2.44510040e-01
-2.54725009e-01 1.15226619e-01 -5.86974621e-01 1.02134645e+00
7.45474100e-01 -2.17444167e-01 -2.29346931e-01 2.12774388e-02
-2.67706335e-01 8.15331042e-02 -1.05952275e+00 1.45415854e+00
-1.53563306e-01 1.06043816e+00 7.85836056e-02 -6.64350688e-01
1.12206817e+00 -4.98008728e-02 2.01983243e-01 -7.83388197e-01
2.53180772e-01 2.89670676e-01 -6.50334597e-01 -5.35198748e-01
8.77605557e-01 -9.05850343e-03 1.91360161e-01 5.54325163e-01
-3.96875471e-01 1.06720857e-01 -1.73441231e-01 -5.29892854e-02
8.35718334e-01 2.66265064e-01 -1.36642426e-01 1.35185897e-01
7.03647792e-01 -1.11650638e-01 3.45324010e-01 8.55101883e-01
-1.90040275e-01 9.31046188e-01 1.71843752e-01 -3.80059898e-01
-9.89455760e-01 -8.10683191e-01 -5.63107915e-02 1.28166902e+00
5.02575994e-01 -4.12819058e-01 -6.34460509e-01 -8.42878401e-01
-2.24642660e-02 4.73169088e-01 -2.70592362e-01 2.75919847e-02
-6.53760493e-01 -6.67918086e-01 9.90812063e-01 5.95276535e-01
8.06844652e-01 -7.13662684e-01 -5.82353532e-01 -7.44498670e-02
-3.22758496e-01 -1.11396563e+00 -1.11027277e+00 -1.72795922e-01
-8.69206369e-01 -8.32425177e-01 -1.02093697e+00 -8.78362656e-01
8.62630904e-01 7.91660130e-01 5.16864181e-01 3.68808210e-01
-5.34547925e-01 2.30002031e-01 -4.41845983e-01 -1.62329957e-01
-2.25442439e-01 6.76173717e-02 -9.18261781e-02 3.26128840e-01
2.09167555e-01 1.93437487e-01 -5.47654033e-01 8.70659232e-01
-9.21676338e-01 3.47243935e-01 8.28891933e-01 1.05936456e+00
5.10838926e-01 4.49961498e-02 2.68463165e-01 -4.33265537e-01
5.03437996e-01 3.27740912e-03 -8.64244640e-01 5.19814134e-01
-6.35432422e-01 2.05698907e-02 8.05708945e-01 -5.22813380e-01
-9.04149711e-01 3.51160139e-01 -7.07837194e-02 -6.17968798e-01
-4.92183398e-03 1.84272438e-01 -1.22725673e-01 -2.10903764e-01
4.58305985e-01 8.35500717e-01 1.19725212e-01 -3.45823675e-01
1.68057352e-01 1.14338672e+00 6.01039708e-01 -3.17006409e-01
8.83914709e-01 5.55406094e-01 -3.31365645e-01 -1.04807115e+00
-6.74611092e-01 -6.52344823e-01 -3.98205251e-01 -2.40862951e-01
4.96686369e-01 -9.51178133e-01 -8.75897646e-01 8.84685636e-01
-1.01745391e+00 -3.91863167e-01 1.65703416e-01 3.13266188e-01
-4.32216287e-01 8.68399739e-01 -5.01434505e-01 -1.07799566e+00
-7.92209089e-01 -1.08842790e+00 1.79476917e+00 2.39271089e-01
2.56203681e-01 -7.13089228e-01 -2.54997641e-01 3.70441258e-01
4.97224957e-01 -3.93950224e-01 3.23001564e-01 -3.92606497e-01
-5.55495620e-01 -4.02604312e-01 -8.03891897e-01 2.49695443e-02
1.14452414e-01 6.93587065e-02 -8.44963372e-01 -4.41822827e-01
-1.20717578e-01 -1.16446428e-01 1.01021874e+00 2.06220880e-01
9.47166681e-01 -1.71706438e-01 -4.33802336e-01 6.95890367e-01
1.30324435e+00 -2.04844065e-02 7.44059741e-01 2.77168363e-01
8.72772753e-01 2.62088448e-01 7.09533811e-01 6.55929387e-01
5.46319820e-02 8.41338992e-01 1.60825521e-01 -1.89447582e-01
-2.98756123e-01 -3.81157398e-01 6.07561290e-01 6.72857881e-01
5.04146993e-01 -7.04760790e-01 -7.84291267e-01 1.21412799e-01
-2.00903106e+00 -6.94276690e-01 -2.32693002e-01 2.31524992e+00
5.12616038e-01 1.40291110e-01 2.44434342e-01 1.82998508e-01
1.24484670e+00 7.61782378e-02 -8.30358326e-01 1.16241708e-01
-2.77975887e-01 -3.96239996e-01 4.65978980e-01 5.04920363e-01
-1.14595711e+00 1.16008723e+00 6.11462069e+00 1.05611718e+00
-1.39510155e+00 -3.30191195e-01 6.06579185e-01 1.28902337e-02
7.42711052e-02 -1.81542665e-01 -1.15367806e+00 5.92603862e-01
3.18615168e-01 5.49323931e-02 5.48135638e-01 7.28262305e-01
1.10062182e-01 1.64334640e-01 -8.42989683e-01 1.33994770e+00
5.40062726e-01 -1.05101252e+00 2.39110067e-01 -2.53842860e-01
3.73208314e-01 -6.18979335e-02 2.10767388e-01 2.59965986e-01
-1.95897669e-01 -8.98040295e-01 9.03454065e-01 3.84020030e-01
1.31576574e+00 -4.00985181e-01 4.43959892e-01 4.98350024e-01
-1.36138761e+00 -4.76116464e-02 -2.56848425e-01 4.19512093e-02
8.98700878e-02 7.61797905e-01 -9.64889169e-01 3.87078524e-01
1.84213385e-01 8.89880538e-01 -4.34169710e-01 1.03502011e+00
-1.98163271e-01 5.68611324e-01 -4.55163717e-01 -4.54657197e-01
6.52722046e-02 -6.46397471e-02 6.73486471e-01 1.38463020e+00
4.31193233e-01 -1.77179113e-01 9.92091000e-02 1.10794723e+00
-2.26963818e-01 1.21538214e-01 -2.74084628e-01 -1.34278119e-01
3.29657108e-01 1.09285736e+00 -9.17205334e-01 -2.53673881e-01
-3.35716128e-01 1.42057180e+00 1.33324459e-01 3.00608248e-01
-1.10815287e+00 -8.77229512e-01 2.38964632e-01 -1.06796093e-01
4.97141540e-01 -2.03916997e-01 -7.62112141e-01 -1.53420866e+00
5.43030441e-01 -8.12218010e-01 1.84444264e-01 -8.50121737e-01
-1.08625686e+00 3.15811187e-01 -8.65208745e-01 -1.36766195e+00
3.90332788e-01 -7.57262111e-01 -5.78283310e-01 7.09422767e-01
-1.47861791e+00 -1.16620445e+00 -5.76007426e-01 4.71423566e-01
1.05125511e+00 -1.79217190e-01 3.86767238e-01 3.42537612e-01
-1.08285868e+00 1.22476375e+00 3.60841244e-01 3.66976768e-01
8.30466509e-01 -9.11003768e-01 7.93226540e-01 1.01362967e+00
1.87275708e-01 2.00932205e-01 4.27034616e-01 -7.96042919e-01
-2.00760055e+00 -1.10452890e+00 6.65400445e-01 -3.64665926e-01
4.80340362e-01 -5.22063375e-01 -8.55720103e-01 4.92566019e-01
-3.83334368e-01 -2.55500525e-01 1.32007495e-01 -1.51408300e-01
-3.32940340e-01 -8.86506811e-02 -9.41380918e-01 8.20923686e-01
1.00145710e+00 -4.17496711e-01 -4.39754933e-01 4.62532699e-01
5.26422560e-01 -6.85267508e-01 -8.61327946e-02 2.36874849e-01
7.40248740e-01 -8.19414198e-01 7.19275653e-01 -2.97228806e-02
3.76342982e-01 -3.83184463e-01 -8.04871023e-02 -6.71538413e-01
-3.63717303e-02 -8.46208155e-01 -1.78157732e-01 1.16984034e+00
2.75563776e-01 -6.99248791e-01 9.75649118e-01 2.21254215e-01
-1.78265542e-01 -7.56765604e-01 -1.02231979e+00 -7.78491020e-01
-1.98885530e-01 -3.63295615e-01 3.30763280e-01 6.25632405e-01
1.76467076e-01 4.73916024e-01 -6.72620475e-01 1.70503512e-01
4.60747421e-01 3.40060085e-01 8.87539208e-01 -9.38449681e-01
-2.77784467e-01 -7.74936140e-01 -2.97602355e-01 -2.10741591e+00
-9.55550894e-02 -6.20732069e-01 5.05572319e-01 -1.33513832e+00
2.79577345e-01 -4.09286737e-01 1.92989156e-01 4.74153221e-01
-4.17402208e-01 8.84485915e-02 2.23361775e-01 5.77357233e-01
-8.45714867e-01 7.36173451e-01 1.19374299e+00 -1.68847680e-01
-1.90491870e-01 1.08706998e-02 -5.03741145e-01 4.44736451e-01
5.83074212e-01 -1.88372061e-01 5.67586757e-02 -9.24520612e-01
2.37114862e-01 9.12950113e-02 2.55407274e-01 -8.42252135e-01
6.45973504e-01 4.13714945e-02 4.21503156e-01 -8.91450226e-01
9.34620723e-02 -7.15994239e-01 -5.77933609e-01 4.07609224e-01
-3.37289095e-01 -2.43091524e-01 2.50975877e-01 7.06772804e-01
4.53111120e-02 -1.19790070e-01 6.22705638e-01 5.53310156e-01
-5.68045914e-01 1.64440259e-01 -2.86356986e-01 -8.48513544e-02
8.46431017e-01 -4.97180641e-01 -6.49679959e-01 -2.39293531e-01
-1.64480314e-01 4.17235374e-01 4.96212453e-01 5.15722334e-01
9.59596753e-01 -1.05693293e+00 -8.93057585e-01 7.08172321e-01
8.29081051e-03 -1.08156176e-02 1.70498714e-01 1.06003094e+00
-5.37048101e-01 5.47380805e-01 3.26808959e-01 -8.83897841e-01
-1.34364116e+00 3.33951294e-01 3.58610243e-01 -1.62347347e-01
-1.00802362e+00 7.12332249e-01 1.24994688e-01 -2.49924645e-01
5.93952656e-01 -2.23606721e-01 3.39567035e-01 -4.15204883e-01
7.35683620e-01 4.48670059e-01 -2.63036638e-02 -5.03915191e-01
-2.43501082e-01 9.68614995e-01 -3.28366667e-01 2.10679341e-02
8.09713840e-01 -3.70424658e-01 2.72720098e-01 2.48116806e-01
9.00365770e-01 6.59581050e-02 -1.40209460e+00 -3.18625242e-01
-1.94008112e-01 -7.80393124e-01 9.47354212e-02 -9.09668326e-01
-9.87737834e-01 9.05514359e-01 5.38430810e-01 2.68504396e-02
1.12800670e+00 -3.60868514e-01 1.21598089e+00 6.00886881e-01
-1.83907654e-02 -1.13755155e+00 3.32757264e-01 7.05628216e-01
7.41264522e-01 -1.30160308e+00 -4.17526923e-02 -4.54933733e-01
-6.51417732e-01 1.40248442e+00 5.55662394e-01 -1.35899959e-02
9.93333943e-03 4.34055179e-01 -2.69459821e-02 -3.83838564e-02
-4.91330653e-01 1.40604619e-02 3.12748909e-01 1.27803773e-01
3.01022947e-01 -1.06176950e-01 1.14314310e-01 4.12236214e-01
2.18501955e-01 1.02859465e-02 2.38061383e-01 9.12731647e-01
-7.06845641e-01 -5.87773561e-01 -6.50578320e-01 5.09504020e-01
-2.93172807e-01 -2.08161637e-01 -6.71307683e-01 4.44374710e-01
-3.98954093e-01 9.94600415e-01 5.88755272e-02 -6.19335651e-01
4.52824831e-01 -1.23528823e-01 1.06175445e-01 -3.42643946e-01
-2.91915178e-01 3.56816620e-01 -2.72596270e-01 -2.42745101e-01
1.56875983e-01 -3.75909746e-01 -1.36635184e+00 -4.36451852e-01
-1.12106979e+00 -2.57524401e-01 7.38445938e-01 1.00082517e+00
6.24531269e-01 3.86843681e-01 7.98013926e-01 -7.61633575e-01
-8.51793826e-01 -7.95599520e-01 -5.26919782e-01 1.47853300e-01
4.98143852e-01 -3.72291028e-01 -5.95939696e-01 1.16251878e-01] | [12.006939888000488, 2.216689348220825] |
9d66fb06-47f5-4cdf-8ac2-42d4cb3ae4fc | lidar-snowfall-simulation-for-robust-3d | 2203.15118 | null | https://arxiv.org/abs/2203.15118v2 | https://arxiv.org/pdf/2203.15118v2.pdf | LiDAR Snowfall Simulation for Robust 3D Object Detection | 3D object detection is a central task for applications such as autonomous driving, in which the system needs to localize and classify surrounding traffic agents, even in the presence of adverse weather. In this paper, we address the problem of LiDAR-based 3D object detection under snowfall. Due to the difficulty of collecting and annotating training data in this setting, we propose a physically based method to simulate the effect of snowfall on real clear-weather LiDAR point clouds. Our method samples snow particles in 2D space for each LiDAR line and uses the induced geometry to modify the measurement for each LiDAR beam accordingly. Moreover, as snowfall often causes wetness on the ground, we also simulate ground wetness on LiDAR point clouds. We use our simulation to generate partially synthetic snowy LiDAR data and leverage these data for training 3D object detection models that are robust to snowfall. We conduct an extensive evaluation using several state-of-the-art 3D object detection methods and show that our simulation consistently yields significant performance gains on the real snowy STF dataset compared to clear-weather baselines and competing simulation approaches, while not sacrificing performance in clear weather. Our code is available at www.github.com/SysCV/LiDAR_snow_sim. | ['Luc van Gool', 'Dengxin Dai', 'Fisher Yu', 'Felix Heide', 'Mario Bijelic', 'Christos Sakaridis', 'Martin Hahner'] | 2022-03-28 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Hahner_LiDAR_Snowfall_Simulation_for_Robust_3D_Object_Detection_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Hahner_LiDAR_Snowfall_Simulation_for_Robust_3D_Object_Detection_CVPR_2022_paper.pdf | cvpr-2022-1 | ['robust-3d-object-detection', 'physical-simulations'] | ['computer-vision', 'miscellaneous'] | [ 3.60282324e-02 -4.34290916e-01 3.09706688e-01 -2.37670958e-01
-4.28152949e-01 -5.97994685e-01 4.59122449e-01 3.82087938e-02
-4.28152353e-01 7.32202053e-01 -5.53992808e-01 -7.79625595e-01
4.85771835e-01 -1.04552555e+00 -7.50132143e-01 -5.08079231e-01
-1.53234273e-01 8.72035325e-01 7.70963073e-01 -3.43564332e-01
5.87437861e-02 1.00698566e+00 -2.09090638e+00 -4.05673742e-01
1.18281853e+00 5.52254021e-01 4.18529242e-01 9.11140621e-01
-2.66902775e-01 9.42354184e-03 -5.95795035e-01 2.11218953e-01
7.51606345e-01 1.16593845e-01 1.99114308e-01 1.34209946e-01
1.03276396e+00 -3.75656396e-01 -2.68859774e-01 8.08765709e-01
5.03205061e-01 1.75639257e-01 6.95748150e-01 -1.43604791e+00
8.96258503e-02 -2.17367530e-01 -9.16038513e-01 6.99179471e-01
1.25240773e-01 7.27734745e-01 4.89571244e-01 -1.01549101e+00
2.45683402e-01 1.23030674e+00 7.33555615e-01 2.50028402e-01
-1.12396801e+00 -1.09938478e+00 2.77805418e-01 3.28930728e-02
-1.50694883e+00 -3.59900087e-01 2.25448683e-01 -5.80989242e-01
9.09644127e-01 2.46060118e-01 9.54378068e-01 5.79444289e-01
3.77024412e-01 5.79506338e-01 1.32488418e+00 -1.43929809e-01
3.29655677e-01 1.08868688e-01 3.95873070e-01 5.52136242e-01
9.21464384e-01 8.00594270e-01 -5.53156912e-01 -1.79678112e-01
1.96924850e-01 -9.37409773e-02 2.34123841e-02 -3.94818068e-01
-7.76951492e-01 8.10055315e-01 5.73673010e-01 -7.03172505e-01
-2.39947006e-01 2.15409026e-01 -4.47155014e-02 7.50459731e-02
7.31158555e-01 -7.02873692e-02 -1.21823162e-01 6.17264956e-02
-1.03326046e+00 9.80148017e-01 5.31170607e-01 1.17992783e+00
8.53022337e-01 2.62652159e-01 1.14299908e-01 3.14028800e-01
4.13295627e-01 1.69799614e+00 -4.32164341e-01 -8.05689692e-01
6.09174788e-01 4.14457560e-01 5.86607575e-01 -4.29303765e-01
-4.29596096e-01 -3.46499652e-01 -2.48289317e-01 1.05162454e+00
2.13224247e-01 -1.56583533e-01 -1.39511561e+00 1.11513531e+00
8.01105142e-01 6.88621104e-01 1.23507574e-01 1.02089202e+00
8.78649235e-01 5.88617206e-01 1.15402939e-03 4.25538868e-02
1.19310069e+00 -4.17963237e-01 -3.84110928e-01 -7.34846294e-01
5.27155042e-01 -7.68831909e-01 9.91515100e-01 -2.04585001e-01
-6.54908597e-01 -4.26798463e-01 -1.04253972e+00 2.44586602e-01
-4.83206302e-01 -1.33411989e-01 4.45838332e-01 7.27233827e-01
-8.32844615e-01 3.46631743e-02 -1.03337908e+00 -5.85155785e-01
4.47704673e-01 6.02464043e-02 3.41717243e-01 -1.17781609e-01
-1.00787389e+00 1.32518673e+00 -1.45662159e-01 2.22742990e-01
-9.77942526e-01 -9.98720646e-01 -9.01618958e-01 -4.18073297e-01
3.05339307e-01 -7.66924500e-01 1.38183522e+00 1.41834438e-01
-8.60362291e-01 8.32998872e-01 -7.10252285e-01 -8.50887120e-01
7.04703212e-01 -3.86767328e-01 -2.44387761e-01 -2.10567832e-01
5.58424056e-01 7.55874515e-01 6.38513446e-01 -1.59538019e+00
-1.01616156e+00 -5.02708256e-01 -6.79366887e-02 4.40302253e-01
6.91295922e-01 -2.75190055e-01 -1.54994288e-02 -1.52761936e-02
-2.92059616e-03 -1.16994202e+00 -5.85804582e-01 3.07613164e-01
-3.10434252e-01 2.83603936e-01 1.23052549e+00 9.58145559e-02
5.25637388e-01 -1.93006086e+00 -7.72718728e-01 2.04029471e-01
1.95554540e-01 2.21055090e-01 6.26605004e-02 1.80547833e-01
5.11997938e-01 -5.92217445e-02 -5.59833109e-01 -4.65412110e-01
-7.26750270e-02 4.37996358e-01 -6.96361005e-01 5.06800771e-01
2.35725671e-01 8.07983756e-01 -8.45376492e-01 -5.46619117e-01
6.52671814e-01 4.26859796e-01 -2.53911186e-02 -6.94360733e-02
-1.58876240e-01 1.68339208e-01 -5.05093575e-01 7.77321100e-01
1.49581027e+00 3.95354480e-01 -4.74685013e-01 3.92654389e-01
-5.92870593e-01 4.87687737e-01 -1.23057973e+00 9.07211483e-01
-5.64064562e-01 7.88713634e-01 2.70932019e-01 -1.15658335e-01
9.12500501e-01 -3.65358531e-01 5.95867075e-02 -6.71530485e-01
7.69861564e-02 7.71584362e-02 -1.02945574e-01 -3.16156119e-01
8.08094203e-01 -2.50429928e-01 2.43834127e-02 4.18117017e-01
-9.96912479e-01 -1.06531978e+00 1.24891005e-01 2.46167943e-01
1.00189221e+00 6.04918052e-04 -2.43916195e-02 -1.60693958e-01
-1.59944631e-02 7.22533107e-01 3.69193256e-01 1.09846020e+00
-3.24901193e-01 4.24395770e-01 -3.15484941e-01 -4.16128129e-01
-8.96687806e-01 -1.67277348e+00 -6.54097557e-01 5.34366608e-01
7.76865363e-01 6.80577978e-02 -3.59048516e-01 -3.70903999e-01
8.07528913e-01 1.01472795e+00 -4.61944938e-01 3.42727304e-02
-3.25868249e-01 -1.09019077e+00 4.78935391e-01 5.22178650e-01
4.58869845e-01 -6.43490016e-01 -1.19568706e+00 1.54144660e-01
1.73639670e-01 -1.39887738e+00 -3.04323304e-02 1.66129619e-01
-8.87291729e-01 -1.04528606e+00 -3.25063281e-02 -1.10601641e-01
5.35327852e-01 1.25688112e+00 1.13539076e+00 3.20549250e-01
-5.92255771e-01 1.56885937e-01 -2.40433037e-01 -1.23362482e+00
-4.62617636e-01 -1.11725852e-01 2.85648614e-01 -5.84863544e-01
7.21335053e-01 -3.43222916e-01 -4.21521574e-01 5.34451127e-01
-4.46233690e-01 -4.05726545e-02 2.78450519e-01 1.21676020e-01
5.68952322e-01 -3.51532884e-02 7.05096647e-02 -6.34617090e-01
2.51436859e-01 -3.91401947e-01 -1.32000899e+00 -4.46419477e-01
-2.84272581e-01 -3.76138449e-01 -1.73455998e-01 -4.86303382e-02
-7.11436749e-01 3.92964810e-01 1.97767720e-01 -4.27259803e-01
-3.59787345e-01 6.38108552e-02 1.36435777e-01 -3.56500477e-01
8.73132885e-01 1.52233228e-01 1.04479957e-02 -1.55904323e-01
2.11563647e-01 7.29021072e-01 3.20283055e-01 -4.37219709e-01
1.62213421e+00 1.09120655e+00 3.58252883e-01 -1.25886905e+00
-9.07811999e-01 -6.58278644e-01 -6.46010399e-01 -3.39427620e-01
5.91388404e-01 -1.28532588e+00 -4.96658117e-01 4.03169960e-01
-1.05448031e+00 -7.01676786e-01 -2.99294740e-01 3.34467590e-01
-2.68065929e-01 2.04012975e-01 8.44570249e-02 -1.32355118e+00
-1.22141294e-01 -1.00766218e+00 1.27240217e+00 1.73868001e-01
9.30065364e-02 -4.87336397e-01 9.22180936e-02 1.90476134e-01
1.86479464e-01 3.05575252e-01 3.00174534e-01 1.56130627e-01
-1.15039444e+00 -2.54597843e-01 -3.41441959e-01 -2.16191247e-01
1.23507760e-01 1.67286471e-01 -1.17813265e+00 -1.82995990e-01
-3.41626555e-01 1.72448725e-01 1.19081628e+00 4.96990442e-01
4.96039897e-01 3.71286243e-01 -8.03991020e-01 4.27779377e-01
9.97049391e-01 -1.40577465e-01 3.17520171e-01 3.74346793e-01
4.98938590e-01 4.14766073e-01 1.21220171e+00 2.11479336e-01
7.12159753e-01 7.04976976e-01 6.39992535e-01 -2.59418488e-01
-2.66008317e-01 -1.20663811e-02 3.45432103e-01 -1.69670507e-01
1.07614093e-01 -2.91467816e-01 -9.95116055e-01 5.90781689e-01
-1.62738216e+00 -9.64795649e-01 -8.71317923e-01 2.49417377e+00
1.37413979e-01 5.69977045e-01 1.11074552e-01 -1.46685109e-01
6.93753541e-01 3.82540748e-02 -7.92066753e-01 -1.48250446e-01
-6.49723262e-02 1.90053448e-01 1.27341259e+00 9.73862588e-01
-1.11542296e+00 1.23469186e+00 6.18719149e+00 1.07092090e-01
-1.11086023e+00 1.95575014e-01 -2.29904801e-01 -3.89490217e-01
-1.09590471e-01 -1.80218332e-02 -1.53116357e+00 4.40059036e-01
9.81105685e-01 -2.14080557e-01 -1.84666831e-02 5.94628632e-01
1.09714806e+00 -7.61297464e-01 -4.78831291e-01 5.33616602e-01
-3.10879529e-01 -1.20677257e+00 -5.39254695e-02 3.59956563e-01
5.42323947e-01 8.54982615e-01 -4.44566458e-02 1.38473570e-01
8.33179176e-01 -7.17887819e-01 8.24352026e-01 1.52541444e-01
6.20009482e-01 -4.59344685e-01 4.54880118e-01 4.54916328e-01
-1.51499927e+00 2.56939739e-01 -5.81547618e-01 -4.08108115e-01
5.15315831e-01 9.48940516e-01 -1.38103676e+00 1.55944809e-01
9.02013659e-01 5.64710438e-01 -5.57108581e-01 1.60100424e+00
-5.07588387e-01 7.24956930e-01 -1.12242150e+00 -1.03113957e-01
2.66404510e-01 -1.73459053e-01 1.11829567e+00 1.08345771e+00
4.19363469e-01 1.06304757e-01 4.96283174e-01 9.62936282e-01
2.64312625e-01 -3.44277114e-01 -1.04140377e+00 6.03216290e-01
8.95256281e-01 9.86805379e-01 -6.57941282e-01 -3.28905135e-01
-1.57855824e-01 2.05271140e-01 -2.60746896e-01 3.34109962e-01
-9.98880327e-01 -1.13143325e-01 1.40166926e+00 7.80508459e-01
4.43607330e-01 -7.22698569e-01 -5.67569435e-01 -8.29016864e-01
2.14167386e-01 -1.79712519e-01 -5.31823970e-02 -1.00475168e+00
-9.42149818e-01 3.72600883e-01 3.40790093e-01 -1.59498525e+00
1.65348172e-01 -4.94859219e-01 -8.33109677e-01 1.08192945e+00
-2.18362379e+00 -1.07454300e+00 -9.61831272e-01 6.54490590e-02
4.63790685e-01 3.23696077e-01 3.00323129e-01 1.33484513e-01
-1.59605473e-01 -5.46487235e-02 -1.78010464e-02 -4.11495268e-01
5.70591807e-01 -1.19411767e+00 1.33559489e+00 9.44396973e-01
-1.14164995e-02 3.50527801e-02 1.15753329e+00 -1.20763564e+00
-1.24804091e+00 -1.59989250e+00 5.84082603e-01 -8.92490506e-01
6.52341127e-01 -7.33946979e-01 -9.01684940e-01 5.27386010e-01
-3.94507676e-01 3.40396702e-01 1.75788343e-01 -1.79207012e-01
-4.97821718e-02 -2.07944065e-01 -1.15630949e+00 5.90837419e-01
1.17444229e+00 -6.13381863e-02 -5.57836235e-01 5.63597620e-01
6.25278354e-01 -8.30873311e-01 8.47179368e-02 5.68672955e-01
3.71085554e-01 -8.16857100e-01 9.96834576e-01 -2.46176839e-01
-3.14451873e-01 -1.06563556e+00 -1.34363636e-01 -1.41237152e+00
5.30406125e-02 -2.60541677e-01 1.14788353e-01 5.46744108e-01
6.61126375e-01 -8.50154996e-01 1.13984954e+00 2.29871437e-01
-4.55491543e-01 -1.33429527e-01 -1.25803983e+00 -1.08415747e+00
-1.58261135e-03 -8.52739096e-01 6.46950126e-01 2.66930908e-01
-8.02594364e-01 1.76647082e-01 1.35984197e-01 1.12117481e+00
1.12293422e+00 5.01882851e-01 1.40543425e+00 -1.48738730e+00
4.53539521e-01 -8.77979845e-02 -5.44094026e-01 -1.00838888e+00
2.06563503e-01 -6.78626537e-01 5.79686522e-01 -1.64396036e+00
-1.23553127e-01 -1.02590978e+00 3.41443777e-01 2.84866720e-01
-2.22670674e-01 4.76354808e-01 1.65998802e-01 1.69650570e-01
-1.40835926e-01 6.08799040e-01 1.11786580e+00 -1.89722434e-01
-3.86791527e-01 5.05590439e-01 -1.60881639e-01 6.85336649e-01
7.67944574e-01 -7.65403509e-01 -1.66134015e-01 -5.64383388e-01
-1.02773704e-01 -4.10341084e-01 8.51164341e-01 -1.45931005e+00
4.06032242e-02 -5.02881169e-01 3.05855840e-01 -1.39398265e+00
8.10832679e-01 -8.34874868e-01 -3.11872602e-01 4.95872408e-01
5.28296232e-01 -6.30157366e-02 7.42693126e-01 4.88976449e-01
2.95650452e-01 -1.09543085e-01 8.87560546e-01 -4.24821563e-02
-6.86775088e-01 5.41502953e-01 -8.03671002e-01 -1.49966339e-02
1.16922390e+00 -5.59375584e-01 -4.22276348e-01 -2.87262470e-01
-3.63601089e-01 7.30927467e-01 7.92626500e-01 3.32081169e-01
6.79547608e-01 -7.85900235e-01 -9.60299969e-01 3.37947100e-01
3.71072084e-01 3.88068825e-01 -4.59311046e-02 5.82656085e-01
-6.06226206e-01 1.89805388e-01 2.15926439e-01 -1.11622858e+00
-1.41096699e+00 1.56326264e-01 6.24662399e-01 3.94440651e-01
-8.18996906e-01 4.35340732e-01 3.06172311e-01 -5.59488833e-01
-2.04356655e-01 -7.16579616e-01 3.66716921e-01 -1.92778468e-01
4.86316502e-01 2.43318915e-01 2.78255522e-01 -7.07961261e-01
-4.12321717e-01 7.73254991e-01 1.45539761e-01 6.04021875e-03
9.50950325e-01 -2.37806067e-01 5.29879570e-01 4.41594750e-01
2.82456517e-01 1.52985543e-01 -1.42949009e+00 -4.84477952e-02
-4.06523019e-01 -7.93577433e-01 3.09067369e-01 -4.40186769e-01
-7.00550079e-01 9.10746813e-01 8.99341702e-01 -1.24621904e-02
3.24999481e-01 2.44008396e-02 8.13211858e-01 6.42878294e-01
5.18633187e-01 -6.35850072e-01 -6.12620056e-01 8.00234437e-01
5.49152374e-01 -1.38232541e+00 3.12309951e-01 -8.25374424e-01
-3.65834057e-01 6.05532467e-01 7.81585574e-01 -2.03342259e-01
6.43509984e-01 8.41884851e-01 5.80420732e-01 -4.40975010e-01
-4.54683781e-01 -8.33786309e-01 -3.18830431e-01 8.53480458e-01
-5.01515806e-01 3.43658268e-01 2.45204434e-01 -2.79962212e-01
-6.34111345e-01 -1.59856603e-02 8.76123965e-01 1.26134527e+00
-1.08080554e+00 -6.61284208e-01 -9.36700344e-01 6.98897958e-01
4.13805962e-01 -1.57096848e-01 -2.78277308e-01 9.33130026e-01
1.24430276e-01 9.81682181e-01 5.21179914e-01 -3.66919562e-02
7.27699280e-01 -1.97647378e-01 3.62397432e-01 -9.25748944e-01
-1.01608299e-01 -3.14040422e-01 2.05077559e-01 -2.00149626e-01
-2.43188128e-01 -1.04594588e+00 -1.46754396e+00 -3.79853040e-01
-3.93079460e-01 6.96429312e-02 8.56989145e-01 8.68157387e-01
4.53345984e-01 1.04795165e-01 5.78198135e-01 -1.33274448e+00
-2.99053103e-01 -7.89444089e-01 -5.78043163e-01 -1.40859589e-01
6.45049632e-01 -1.41656971e+00 -6.62743211e-01 -3.11202437e-01] | [7.818630695343018, -2.4845192432403564] |
2b7d59b4-1ebc-43fa-861f-a30c31ef6c1c | delving-deeper-into-the-decoder-for-video | 2001.05614 | null | https://arxiv.org/abs/2001.05614v3 | https://arxiv.org/pdf/2001.05614v3.pdf | Delving Deeper into the Decoder for Video Captioning | Video captioning is an advanced multi-modal task which aims to describe a video clip using a natural language sentence. The encoder-decoder framework is the most popular paradigm for this task in recent years. However, there exist some problems in the decoder of a video captioning model. We make a thorough investigation into the decoder and adopt three techniques to improve the performance of the model. First of all, a combination of variational dropout and layer normalization is embedded into a recurrent unit to alleviate the problem of overfitting. Secondly, a new online method is proposed to evaluate the performance of a model on a validation set so as to select the best checkpoint for testing. Finally, a new training strategy called professional learning is proposed which uses the strengths of a captioning model and bypasses its weaknesses. It is demonstrated in the experiments on Microsoft Research Video Description Corpus (MSVD) and MSR-Video to Text (MSR-VTT) datasets that our model has achieved the best results evaluated by BLEU, CIDEr, METEOR and ROUGE-L metrics with significant gains of up to 18% on MSVD and 3.5% on MSR-VTT compared with the previous state-of-the-art models. | ['Jianmin Li', 'Xiaolin Hu', 'Haoran Chen'] | 2020-01-16 | null | null | null | null | ['video-description'] | ['computer-vision'] | [ 2.42343843e-01 -7.39964768e-02 -2.53088951e-01 -3.74213636e-01
-1.22304940e+00 -2.21201316e-01 5.93913019e-01 -1.04122967e-01
-4.48659569e-01 8.39952648e-01 4.57596481e-01 -1.20276980e-01
3.89608264e-01 -6.79303482e-02 -9.54528987e-01 -4.46162999e-01
2.55712420e-01 3.53393316e-01 3.28454137e-01 -1.68146983e-01
2.17198581e-01 -1.43807307e-01 -1.37686539e+00 5.75972259e-01
7.32103765e-01 9.12228703e-01 6.25529647e-01 6.99128389e-01
-1.83897734e-01 1.13699555e+00 -5.76027393e-01 -7.25951076e-01
-1.92204118e-01 -5.90639532e-01 -6.87455714e-01 9.80370864e-02
1.68873459e-01 -2.93044448e-01 -5.13280451e-01 8.62925112e-01
5.97250879e-01 6.17406517e-02 6.68678701e-01 -1.25784600e+00
-7.87436843e-01 6.53713644e-01 -5.91545165e-01 1.87463433e-01
4.84923661e-01 -5.51658757e-02 9.56158161e-01 -7.65809417e-01
5.27118444e-01 1.19874215e+00 4.01841670e-01 8.39404106e-01
-7.83478320e-01 -6.01334929e-01 1.44518733e-01 3.28972697e-01
-1.50182045e+00 -4.27474082e-01 5.20308197e-01 -3.57619077e-01
1.03427279e+00 1.29873842e-01 2.11675331e-01 1.36274695e+00
2.05734655e-01 1.02218378e+00 5.74021697e-01 -2.71109968e-01
2.86472659e-03 3.34155679e-01 -1.10101467e-02 5.78188002e-01
-2.72997897e-02 -3.55678171e-01 -4.56357628e-01 1.11925818e-01
5.51913917e-01 -2.76009113e-01 -3.77265960e-01 -8.68870616e-02
-1.09510159e+00 8.92557681e-01 1.15783982e-01 2.95972258e-01
-3.81743729e-01 2.41216063e-01 7.91096091e-01 -2.05536862e-03
5.22578359e-01 1.55054063e-01 -2.21587017e-01 -4.64333475e-01
-1.09302557e+00 1.43403877e-02 6.47104084e-01 1.11103249e+00
4.46688712e-01 1.29203051e-02 -4.51242626e-01 1.00298524e+00
5.47634900e-01 5.16781330e-01 6.79467022e-01 -6.62281036e-01
9.10109162e-01 1.44553423e-01 5.04758991e-02 -7.01807618e-01
5.90940304e-02 -3.20154786e-01 -7.51420319e-01 -2.41645545e-01
-1.36471897e-01 -2.94555068e-01 -1.14975357e+00 1.66068935e+00
-3.08383465e-01 5.02449632e-01 3.47654492e-01 9.94089901e-01
1.13424671e+00 1.23013806e+00 3.24956268e-01 -2.98735857e-01
1.20436609e+00 -1.25508201e+00 -1.08448720e+00 -3.53442729e-01
5.23134649e-01 -6.99365616e-01 8.08041632e-01 1.03682816e-01
-1.10379267e+00 -6.07432961e-01 -1.06728017e+00 -3.41660716e-02
-7.03281686e-02 2.33962268e-01 3.97044450e-01 3.10027510e-01
-1.02115929e+00 3.14283520e-01 -7.09455788e-01 -3.02047938e-01
1.96162283e-01 2.50923634e-01 -2.48002231e-01 -7.66737163e-02
-1.42033064e+00 9.81120408e-01 6.64723933e-01 -1.57325901e-02
-1.17091203e+00 -2.74537623e-01 -1.01794171e+00 1.26167148e-01
3.14199984e-01 -5.19653678e-01 1.41942585e+00 -1.19617724e+00
-1.56212533e+00 6.92566752e-01 -3.19059908e-01 -6.30472779e-01
5.24750590e-01 -2.96294570e-01 -5.81663072e-01 3.43880296e-01
2.19110660e-02 7.91030109e-01 7.61311114e-01 -1.15394473e+00
-7.06973255e-01 1.97931170e-01 1.00641429e-01 4.77013648e-01
-2.56931484e-01 1.45330399e-01 -9.26493466e-01 -6.13586962e-01
-3.18316251e-01 -8.62677217e-01 -4.36049439e-02 -4.97682482e-01
-2.68584728e-01 -2.81420022e-01 7.65919626e-01 -9.36735868e-01
1.46635425e+00 -2.23125553e+00 3.01555693e-01 -3.12642694e-01
-8.55509564e-02 5.83114088e-01 -1.57867596e-01 4.46298629e-01
-4.68554273e-02 1.35064423e-01 -1.48726016e-01 -6.20858192e-01
-1.15639083e-01 1.24894626e-01 -1.86412871e-01 2.63799787e-01
1.21576712e-01 7.22816348e-01 -7.28322566e-01 -6.16733491e-01
1.32875115e-01 7.92108715e-01 -5.10144949e-01 5.42346001e-01
-4.32745486e-01 2.44838387e-01 -4.14868265e-01 3.98006409e-01
4.31491345e-01 -3.73105526e-01 -9.02390480e-02 -8.95681456e-02
2.45214775e-02 1.62519187e-01 -9.04722393e-01 1.92353880e+00
-4.29008394e-01 7.24571705e-01 -7.91991502e-02 -8.46611977e-01
8.57133746e-01 8.90169322e-01 2.85175234e-01 -5.89022398e-01
2.43603691e-01 2.73340523e-01 -3.58225793e-01 -9.23354626e-01
6.41128838e-01 -7.51899183e-02 -1.18357100e-01 2.20763043e-01
2.88450480e-01 4.21895683e-01 3.23160291e-01 2.13762030e-01
8.41279507e-01 3.51584911e-01 1.04063161e-01 1.44219443e-01
8.01314414e-01 1.16914753e-02 6.10152006e-01 5.71866095e-01
-1.15813285e-01 1.01635301e+00 4.60644066e-01 8.97087250e-03
-1.15664828e+00 -6.40224874e-01 2.65625715e-01 9.88877714e-01
1.91281319e-01 -3.88674140e-01 -8.41045499e-01 -5.19006014e-01
-3.78028840e-01 9.61441934e-01 -3.46825629e-01 -2.06688270e-01
-5.11786640e-01 -8.56621802e-01 5.98199010e-01 5.08273661e-01
8.25440288e-01 -1.10943568e+00 -2.00896934e-01 1.33048236e-01
-5.74316382e-01 -1.54253983e+00 -6.70556605e-01 -8.59877244e-02
-7.36459196e-01 -6.40330791e-01 -1.14174116e+00 -1.12157381e+00
4.33559090e-01 1.47586480e-01 8.86903465e-01 -1.92350164e-01
3.23987395e-01 1.57367915e-01 -7.09083676e-01 -1.56831443e-01
-6.92260027e-01 1.77458286e-01 -1.94810778e-01 2.03629151e-01
4.69255477e-01 -1.16036542e-01 -2.81100094e-01 -2.52148136e-02
-9.72918391e-01 4.13814992e-01 8.62756014e-01 7.43661344e-01
4.76519197e-01 -2.89373130e-01 6.51330829e-01 -6.66287780e-01
7.14971662e-01 -7.00712085e-01 -4.30613667e-01 4.43086505e-01
-4.72287029e-01 9.88385603e-02 5.29313505e-01 -4.97316331e-01
-1.05623055e+00 9.98566449e-02 -3.36787194e-01 -7.54120886e-01
9.45020840e-03 6.74705625e-01 -1.61070853e-01 3.41089457e-01
2.13182822e-01 4.03423071e-01 -9.50810034e-03 -4.93871510e-01
1.74658284e-01 1.17114067e+00 5.57909608e-01 -1.19711980e-01
5.53145230e-01 2.72937887e-03 -4.98551577e-01 -7.37889409e-01
-8.29975009e-01 -5.47827482e-01 -3.66301566e-01 -3.28558177e-01
1.26605737e+00 -1.31836104e+00 -6.31091595e-01 3.72254997e-01
-1.42540264e+00 -5.35375141e-02 4.46750432e-01 7.28120327e-01
-6.09379649e-01 3.23949188e-01 -6.58394516e-01 -7.17449009e-01
-5.00686884e-01 -1.39990163e+00 8.50030303e-01 2.83501238e-01
4.60200422e-02 -1.03117895e+00 -8.18457305e-02 5.93634367e-01
4.27380174e-01 1.70348719e-01 6.82663262e-01 -8.12973380e-01
-4.93776888e-01 -2.14289933e-01 -1.44713029e-01 6.25209451e-01
-1.44665807e-01 -1.68459415e-01 -8.70485485e-01 -4.03350890e-01
-4.74992767e-03 -3.46522123e-01 8.38638127e-01 3.12740475e-01
1.04698634e+00 -2.61323959e-01 -3.07186246e-01 5.46304047e-01
1.78121817e+00 4.51749653e-01 7.78492093e-01 6.15276575e-01
7.53300071e-01 1.26568884e-01 5.79517663e-01 2.13877752e-01
3.73377711e-01 6.89211905e-01 4.67312396e-01 1.53185725e-01
-8.92446414e-02 -4.72201884e-01 6.50521696e-01 1.12166965e+00
1.43481076e-01 -7.06500649e-01 -7.75075138e-01 5.09890020e-01
-2.00172162e+00 -8.93900990e-01 -3.16486023e-02 2.21558022e+00
6.50672495e-01 3.58792424e-01 -4.51619290e-02 -1.68247655e-01
1.18754649e+00 3.24901313e-01 -3.60591978e-01 -6.20189071e-01
-7.89766014e-02 -4.14961785e-01 5.43575704e-01 4.30876464e-01
-1.15687096e+00 9.70288038e-01 6.06213808e+00 7.17058778e-01
-1.11419654e+00 2.58882552e-01 5.16609669e-01 6.80515394e-02
-6.81985319e-02 -1.90026045e-01 -9.34574962e-01 7.39924192e-01
1.45683062e+00 -2.99247742e-01 3.81996423e-01 7.37124503e-01
4.24102902e-01 7.25925863e-02 -1.09929943e+00 1.32826281e+00
6.51548207e-01 -1.20953989e+00 3.32848489e-01 -3.38342965e-01
4.72613305e-01 1.54722869e-01 -5.44722341e-02 6.38641238e-01
-2.29534924e-01 -9.98087227e-01 8.04745257e-01 4.13033277e-01
8.95473659e-01 -7.86754906e-01 1.08627856e+00 3.40134472e-01
-9.48583484e-01 4.61245477e-02 -4.14087564e-01 2.00179070e-01
5.96128941e-01 1.34160846e-01 -7.95062661e-01 5.76851368e-01
4.34010953e-01 8.68722141e-01 -5.04334569e-01 1.27612114e+00
-9.32097286e-02 5.79824269e-01 1.39317796e-01 -2.19370872e-01
5.78514695e-01 1.93179790e-02 5.64237058e-01 1.43674397e+00
3.41776401e-01 -6.76929653e-02 -9.98683944e-02 6.44942582e-01
-3.34703922e-01 1.09681122e-01 -4.64536905e-01 -2.90929854e-01
2.32345849e-01 8.76859128e-01 -2.46746987e-01 -4.12470311e-01
-6.99609041e-01 1.13060784e+00 7.14414492e-02 4.57743615e-01
-1.25880456e+00 -3.86346906e-01 3.88408214e-01 1.38423806e-02
5.91439545e-01 -7.15823323e-02 1.51315331e-01 -1.23990059e+00
1.25806138e-01 -9.44975853e-01 2.61294276e-01 -1.10081398e+00
-9.14840519e-01 1.00932360e+00 1.17844857e-01 -1.51850951e+00
-5.29429674e-01 -3.13573390e-01 -3.93345922e-01 8.27554166e-01
-1.68387008e+00 -8.91880333e-01 -2.26098865e-01 4.56141412e-01
9.16171014e-01 -4.71429467e-01 7.07179487e-01 6.40712082e-01
-7.51242995e-01 5.59996605e-01 4.09567893e-01 3.44368517e-01
6.47829115e-01 -9.58668470e-01 2.31705114e-01 8.68696630e-01
-1.50906533e-01 1.41648844e-01 9.44726765e-01 -6.07666552e-01
-1.26450050e+00 -1.19060624e+00 1.13501048e+00 -9.83438864e-02
5.05464911e-01 -3.55011284e-01 -1.01686430e+00 7.91878045e-01
5.78480303e-01 -3.16408902e-01 5.42067945e-01 -4.63250101e-01
-6.03330694e-02 2.31990125e-02 -8.92604530e-01 4.81613785e-01
4.67276514e-01 -4.75698918e-01 -8.36453497e-01 4.69560951e-01
1.06721771e+00 -5.40010393e-01 -7.59239912e-01 2.24939302e-01
2.48280525e-01 -6.97558880e-01 5.46647191e-01 -4.59978133e-01
6.46579146e-01 -2.54275709e-01 -1.79482803e-01 -1.15811098e+00
-1.09127760e-01 -8.02085698e-01 -2.21655801e-01 1.51847386e+00
5.98857224e-01 -6.55806363e-02 5.42483330e-01 4.34951544e-01
-1.83652297e-01 -5.14148533e-01 -8.59762073e-01 -7.70042360e-01
-1.96883425e-01 -2.43487939e-01 1.69004217e-01 6.03635490e-01
-1.97716191e-01 7.60950327e-01 -7.61594296e-01 7.17248544e-02
5.04620314e-01 -3.03123832e-01 3.80859077e-01 -8.56264651e-01
-4.62760217e-02 -1.16185859e-01 -3.49563926e-01 -1.19636691e+00
3.53235990e-01 -9.18295085e-01 2.04020679e-01 -1.89311278e+00
5.12757659e-01 1.91383198e-01 -3.33686858e-01 2.22360805e-01
-9.83464494e-02 1.39684707e-01 4.16300178e-01 3.42230141e-01
-1.03234649e+00 6.56544447e-01 1.09418607e+00 -1.43735722e-01
-1.52374029e-01 -5.93123324e-02 -6.13068223e-01 3.64593327e-01
6.48838580e-01 -6.40487850e-01 -3.46962124e-01 -8.89602005e-01
-9.94197950e-02 4.90780383e-01 1.85672846e-02 -1.09783900e+00
3.23183149e-01 1.89648286e-01 3.84208113e-02 -6.21574044e-01
4.29811388e-01 -6.89352453e-01 1.26518421e-02 3.24468434e-01
-6.59396708e-01 3.18309516e-01 2.36066997e-01 5.96064091e-01
-5.17331541e-01 -5.11233151e-01 7.84601450e-01 -1.24085143e-01
-9.57887411e-01 2.61949211e-01 -4.53912318e-01 1.33885264e-01
1.02238655e+00 1.49360215e-02 -2.13782907e-01 -7.30062783e-01
-5.21237314e-01 5.04332125e-01 1.87116876e-01 9.17908728e-01
8.46466184e-01 -1.37420118e+00 -1.12496161e+00 -1.06315158e-01
1.80282786e-01 -3.11251819e-01 1.27436921e-01 6.81871712e-01
-7.11584508e-01 8.18102360e-01 -1.63676411e-01 -5.73393703e-01
-1.27734637e+00 5.32654166e-01 3.35739642e-01 -1.32771179e-01
-6.00327313e-01 6.09171212e-01 -1.26249447e-01 2.23516628e-01
6.35373712e-01 5.86614721e-02 -6.68852031e-01 -2.07512844e-02
6.57661915e-01 4.99878526e-02 -1.06675111e-01 -1.04760706e+00
-3.24039191e-01 3.13063443e-01 -3.12837511e-01 -2.13555470e-01
1.43417227e+00 -4.26044405e-01 1.06370881e-01 5.32350302e-01
1.62102795e+00 -5.79015732e-01 -1.24975216e+00 -1.65558755e-01
4.80778590e-02 -1.58544868e-01 1.65480018e-01 -8.20201457e-01
-9.62438643e-01 8.40520263e-01 5.77923715e-01 -9.40869004e-03
9.62919354e-01 -1.53705463e-01 1.03925359e+00 2.44389117e-01
-1.80237014e-02 -1.13319540e+00 2.65974440e-02 6.23637199e-01
8.69274020e-01 -1.38103676e+00 -3.31353545e-01 2.30844729e-02
-1.18217170e+00 1.04640388e+00 5.20528555e-01 -1.98382959e-02
3.59408915e-01 4.24056388e-02 1.19468868e-01 2.03532606e-01
-9.32798386e-01 -8.64912271e-02 2.64024615e-01 3.01381230e-01
5.48704863e-01 -2.95648426e-01 -3.69164526e-01 5.41371584e-01
2.69268364e-01 1.32695615e-01 7.25867510e-01 6.67327583e-01
-5.07521689e-01 -8.27231348e-01 -1.48564562e-01 2.17824563e-01
-9.47013140e-01 -1.45699561e-01 -5.83185162e-03 7.15640604e-01
-2.28770867e-01 1.13642335e+00 5.08198049e-03 -5.10608137e-01
2.06337422e-01 1.81727663e-01 -1.38971778e-02 -6.83967352e-01
-5.26098609e-01 2.83366472e-01 8.21332410e-02 -3.81801367e-01
-7.07439780e-01 -4.73294199e-01 -1.16037190e+00 -8.30982253e-02
-2.38682717e-01 5.76111794e-01 9.02895749e-01 1.02408445e+00
1.81625620e-01 7.26690531e-01 5.61721325e-01 -6.20277166e-01
-4.51814622e-01 -1.11453164e+00 -3.93298477e-01 3.90718967e-01
4.41053361e-01 -4.08378482e-01 -5.64660251e-01 1.96030155e-01] | [10.58139705657959, 0.6758863925933838] |
bf77cfe7-16be-4c01-9c89-0d68bc810856 | casino-a-corpus-of-campsite-negotiation | 2103.15721 | null | https://arxiv.org/abs/2103.15721v2 | https://arxiv.org/pdf/2103.15721v2.pdf | CaSiNo: A Corpus of Campsite Negotiation Dialogues for Automatic Negotiation Systems | Automated systems that negotiate with humans have broad applications in pedagogy and conversational AI. To advance the development of practical negotiation systems, we present CaSiNo: a novel corpus of over a thousand negotiation dialogues in English. Participants take the role of campsite neighbors and negotiate for food, water, and firewood packages for their upcoming trip. Our design results in diverse and linguistically rich negotiations while maintaining a tractable, closed-domain environment. Inspired by the literature in human-human negotiations, we annotate persuasion strategies and perform correlation analysis to understand how the dialogue behaviors are associated with the negotiation performance. We further propose and evaluate a multi-task framework to recognize these strategies in a given utterance. We find that multi-task learning substantially improves the performance for all strategy labels, especially for the ones that are the most skewed. We release the dataset, annotations, and the code to propel future work in human-machine negotiations: https://github.com/kushalchawla/CaSiNo | ['Jonathan Gratch', 'Jonathan May', 'Gale Lucas', 'Rene Clever', 'Jaysa Ramirez', 'Kushal Chawla'] | 2021-03-29 | null | https://aclanthology.org/2021.naacl-main.254 | https://aclanthology.org/2021.naacl-main.254.pdf | naacl-2021-4 | ['persuasion-strategies'] | ['computer-vision'] | [-0.07303953 0.44543058 -0.4877368 -0.767216 -0.8288813 -0.88230485
0.9407642 0.20234789 -0.43137273 0.9608061 0.70967776 -0.31585053
-0.16783153 -0.4253675 -0.12780364 -0.44807833 0.1931431 0.99512243
-0.12699191 -0.8787966 0.3432819 -0.19813578 -0.9836948 0.47390231
0.7338399 0.382597 -0.13451499 0.62113 -0.24821408 1.0508721
-0.76286197 -0.6417889 0.05338665 -0.6161853 -1.4014957 -0.27161482
0.14934118 -0.2881597 0.04693243 0.5669178 0.5239842 0.17337966
0.80141294 -1.6701779 -0.09769598 1.4459654 -0.35543355 0.03008769
0.6887761 -0.04262744 1.0562912 -0.36555088 0.8543067 1.514468
0.6474867 0.71598524 -0.96121615 -0.95126694 -0.18155667 0.22249338
-0.9650702 -0.68847656 0.7828655 -0.27055067 0.5925125 0.2199507
0.51985323 1.495977 0.15251088 1.0106453 1.3581339 -0.4484555
-0.10982248 0.01478665 0.12795478 0.5413774 -0.4608 -0.2998344
-0.8968596 -0.46511513 0.4673724 -0.7132356 0.14582579 -0.2971736
-1.0759428 1.161584 -0.02300933 0.5138043 -0.20136134 0.10892379
0.68572223 0.81031364 0.32779846 0.6824416 -0.500016 -1.0073817
-0.2824313 0.7256115 1.5377319 0.7148411 0.5319411 -0.45523614
0.06043173 1.1830163 -0.09580471 -0.07520492 0.21924119 -1.5807586
0.4414917 0.42360383 0.30127314 -0.7932152 -0.893399 0.19848771
-0.42434147 -0.12894969 0.9505627 -0.5385554 -0.05366978 1.6769942
0.6197595 -0.2597545 0.30529425 0.8874372 1.0964835 0.38294762
0.25265723 -0.24790832 1.5356537 -1.0988923 -0.67003506 -0.1546996
0.65558225 -1.0900052 1.0469278 0.3413028 -1.0172191 -0.12500103
-0.49945953 -0.19815812 -0.02456699 -0.27383575 1.1339246 0.45745483
-0.5727656 0.34055847 -0.6316691 -0.40430963 -0.11772103 0.04834453
-0.2478662 0.39354646 -1.4520521 1.236068 0.11348423 -0.21170406
-0.48685253 -0.36143836 -0.93840855 -0.17252377 0.6296662 -0.14056799
1.9432106 -0.98713356 -1.9886879 1.0478123 0.09660666 -0.23758087
0.66477346 -0.06155841 0.20030862 -0.26496172 0.3771334 0.73959976
0.09715944 -1.0197605 -0.80872047 -0.02820852 0.35665053 0.66299284
-0.02170669 0.3798236 -0.07580736 -0.2972344 -0.12929562 -1.058339
-0.07886623 -0.69635206 -0.16486937 -0.9317717 0.26387343 -0.4991252
0.89911294 -1.8595421 0.39980745 0.09868494 0.25646424 -0.11288144
-0.22030674 0.8704871 0.4759457 0.08935601 0.29838413 -0.20064323
0.38907048 0.20406829 0.01618502 0.2969898 -0.26582107 0.8180064
-0.9509748 -0.62527424 -0.0533598 -0.07166618 -0.21026517 0.33616439
-0.21343122 0.75114316 -0.79414123 0.40996188 0.16415238 0.09266244
0.6857226 0.33631876 -0.28716835 0.92537 -0.8479217 1.8293225
-0.4763759 0.5585995 0.74722224 -0.8222352 0.82252425 0.47022146
0.49667403 -0.55847716 0.37473848 0.13099001 0.5479229 -0.3923749
0.59631306 -0.03979731 -0.91086584 0.98931164 -0.06188444 -0.6699422
0.3149406 0.20719802 0.81503654 0.09164184 0.7215647 -0.16391802
0.16279726 0.320502 0.7485717 0.70911795 -0.54807013 -0.31144992
0.8816669 -0.6644286 -0.8771892 -0.646316 -0.06306451 1.9819984
0.17270477 -0.40188897 -0.42862156 -0.5486665 -0.01454276 0.55840605
-0.38242275 0.38710865 -0.7141583 -0.19250615 0.9456756 0.08243688
0.5639228 -1.1204039 -0.56950986 0.42830673 -0.7266946 -1.0759634
-0.430609 0.01979823 0.0916447 -1.1478006 -0.03976389 -0.8603476
-0.11414993 -0.12971373 1.1358956 0.16627695 -0.02534839 0.23253526
-0.48841158 -0.7098352 -1.0003393 0.4396893 0.09832928 -0.652711
0.46826866 -0.5568704 -0.07123596 0.5447299 -0.1004665 0.48271206
0.11509213 0.97121173 -0.43596506 -0.3406715 0.7288429 -0.9770686
1.2388556 -0.43119535 -0.38091156 0.13170776 -0.38473347 -0.15864162
0.14055422 -0.39406848 -1.0425414 -0.1883117 -0.0706732 0.31307945
-0.21482246 0.5497087 0.30883253 -0.03746031 0.67991996 -0.40015048
0.19694127 -0.14359681 0.3875023 1.0401454 0.40915608 -1.2165735
0.46798968 -0.19980814 -0.4157634 -0.6657926 -0.8301125 -0.5385539
-0.4564679 -0.517586 0.6866847 -0.8846625 -1.5565795 0.28044423
-1.0824583 -0.8696295 0.32450238 0.33327815 -0.81229573 0.13018739
-1.0301778 -0.9094218 -0.15738326 -1.0353287 0.72886336 0.5434319
-1.074062 -0.93408376 0.1836706 1.0704691 0.35203546 0.03509087
0.92564064 -1.0900929 0.1279698 0.3796085 0.1018116 -0.2814601
0.12815897 -0.01786441 -0.5702922 0.06713763 -0.32820955 -1.1768106
0.2448048 0.13051423 0.49028343 -0.30118832 -0.22107424 -0.04347738
0.2880078 0.22632042 0.04864458 0.51103115 0.11339183 1.0252004
0.9806776 0.4378843 1.1652793 0.8286018 -0.3244209 -0.02184582
0.35736963 -0.2459016 0.3717759 0.7083142 -0.2617847 -0.15719505
-1.3497515 0.68530667 -2.1509438 -1.1486897 0.04710701 1.3723142
1.5207744 0.10206849 0.42207414 -0.26287624 0.4833872 0.30050963
-0.1381686 -0.98567146 0.11962726 0.24978958 0.173239 0.92105967
-1.1976051 1.4782145 6.360125 0.6213995 -0.8211353 0.2571657
0.7098483 0.02625745 -0.1628846 0.06358948 -0.51896954 -0.03081534
0.8946125 -0.484527 0.6015742 0.67791396 0.39624837 -0.21665603
-1.1081096 0.5292233 -0.09228863 -1.3541049 -0.6400109 -0.1148867
0.4969254 -0.02290322 -0.4103201 0.68195385 0.95420533 -1.0369532
0.44728255 -0.08102427 0.40014052 -0.5664188 0.7613609 0.77579623
-0.71446294 -0.10572853 0.16462907 -0.658432 0.10575138 -0.3569119
-1.4023683 0.26295108 0.44186547 0.31544632 0.2359565 0.21349561
-0.4430217 0.6280177 -0.2702872 -0.41519624 0.3342711 -0.44263703
0.537959 1.0778936 -0.4922869 0.61652744 0.91641134 0.47960496
-0.2949446 0.34874776 -0.41332033 -0.17245062 0.9952027 1.3772703
-0.5230335 -0.14212272 -0.24511795 0.5870986 0.4757054 0.05135895
-0.6880094 -0.08752976 0.91034025 -0.26162192 -0.37566513 -0.3342557
-0.24518396 -0.83565235 -0.30335993 -1.486858 0.329439 -0.25117335
-1.3034136 0.2457718 0.22417325 -0.5562056 -0.8261549 -0.14365865
-0.74982923 0.5195078 -1.0432127 -1.2876625 -0.2007071 0.22319703
0.87118137 -0.32206392 1.1035991 -0.02558774 -0.28536126 0.5456834
-0.40657344 0.56494576 1.1337774 -1.076462 0.27097344 -0.05900886
-0.00998918 0.4170462 0.8944928 -0.45141813 -1.161829 -0.11298679
0.9718046 -0.4636969 0.9914471 -0.52059513 -0.5907865 0.8337117
0.6906625 -0.9755265 0.9847892 0.83752775 -0.20768705 0.3802525
-1.0311991 0.59966356 0.88788974 -0.41353303 -1.0355042 0.46073428
0.30629697 -1.0643967 -1.0542195 0.18041228 0.913267 -0.68616253
0.39951718 -0.7522761 0.49600938 0.30404672 0.17970036 -1.3973556
0.03629817 -1.0958407 0.62639904 1.1851176 0.6538224 -0.5054175
0.8386413 1.0060942 0.10766801 -0.5209709 -1.0724858 -0.13765818
0.4891085 -0.2694156 0.5144841 1.4063553 0.88443965 0.71589565
-0.46893272 -0.20686632 0.3291362 0.41640392 1.268991 -1.1737205
-0.18070577 -0.58882385 0.06614244 -0.9456106 0.7734544 -0.5986658
0.2761417 -1.4673098 0.1377033 -0.8322565 0.361671 0.9747589
-0.05516886 -0.08723616 0.29300246 0.06583741 -0.6424588 0.42013386
1.1044387 -0.06519541 -0.44988662 0.02897698 -0.7195596 0.8264381
0.8131927 -0.30712962 -0.13277102 -0.2784921 -0.05977749 0.5045196
-0.25271022 -0.23454016 0.4668037 -0.8261073 -0.18831347 0.0684527
0.48640108 -0.17230989 -0.14677908 0.24686083 -0.8183983 -0.02845737
0.07111629 -0.06575836 -0.19644013 -0.07274067 0.43310308 -0.05953316
-0.5806543 -0.31492302 -0.8483151 0.07094912 1.1417382 0.15357342
-0.608917 -1.2011997 -0.58487505 0.95417917 0.28909644 0.79264295
0.06320525 -0.893972 -1.0946342 -0.25630882 -0.07280031 -0.18499865
-0.04913993 0.59460473 -0.38900274 0.23828308 -0.5500127 -0.24529211
-1.6588475 -0.4358716 0.3488822 -0.19380845 -0.2818551 0.83047444
-0.30335903 -0.9778948 0.34404877 -0.07766315 -0.42391276 0.2514397
0.1867572 0.1623496 -0.4839748 -0.5272325 -0.283602 0.0506857
-0.08462067 -0.43007672 1.3870138 -0.07986946 -0.2883306 0.6911352
0.47358283 0.38030985 -0.85785544 -0.2328565 0.3324864 -0.23989071
-0.32353482 -0.99916065 -0.24631527 0.14232343 -0.22651468 0.36697373
0.48921704 0.22169465 0.568386 0.47677234 0.73898816 -1.3042505
-0.03250087 1.0434005 0.9516382 -1.3246539 0.09810726 -0.5205599
-1.1916392 1.042766 0.83624727 0.37110686 0.17985618 0.46111417
0.74482703 -0.3269001 -1.2598951 0.00999484 -0.25614023 0.2830099
0.89497346 0.49027818 -0.8626165 0.6925344 -0.8043353 -0.24228527
0.81391436 0.870839 -0.24411424 -1.589293 -0.15601148 0.05446602
-0.27960962 0.28135553 -1.123722 1.0517645 -0.14488791 1.32109
0.00730411 -0.23935172 0.27238068 0.3598312 0.4523017 -0.71658033
-1.1263534 0.03279 1.1493723 -0.52742463 -0.6564031 -0.8670672
-1.4396334 -0.81844634 -0.18948954 0.77520293 0.7411117 1.105779
-0.05191565 -0.11946562 0.7204478 -0.6802932 -0.70560277 -1.4385607
-0.13980058 0.40099797 -0.02411075 -0.72176564 -0.02606387 -0.39990023] | [12.78298282623291, 7.988338947296143] |
ff4304c8-b3ac-46b3-b890-fe471c357c65 | evaluating-off-the-shelf-machine-listening | 2110.07410 | null | https://arxiv.org/abs/2110.07410v1 | https://arxiv.org/pdf/2110.07410v1.pdf | Evaluating Off-the-Shelf Machine Listening and Natural Language Models for Automated Audio Captioning | Automated audio captioning (AAC) is the task of automatically generating textual descriptions for general audio signals. A captioning system has to identify various information from the input signal and express it with natural language. Existing works mainly focus on investigating new methods and try to improve their performance measured on existing datasets. Having attracted attention only recently, very few works on AAC study the performance of existing pre-trained audio and natural language processing resources. In this paper, we evaluate the performance of off-the-shelf models with a Transformer-based captioning approach. We utilize the freely available Clotho dataset to compare four different pre-trained machine listening models, four word embedding models, and their combinations in many different settings. Our evaluation suggests that YAMNet combined with BERT embeddings produces the best captions. Moreover, in general, fine-tuning pre-trained word embeddings can lead to better performance. Finally, we show that sequences of audio embeddings can be processed using a Transformer encoder to produce higher-quality captions. | ['Xavier Serra', 'Konstantinos Drossos', 'Xavier Favory', 'Benno Weck'] | 2021-10-14 | null | null | null | null | ['audio-captioning'] | ['audio'] | [ 4.18490976e-01 1.21681318e-01 8.49216729e-02 -2.85111666e-01
-1.36482382e+00 -4.80769098e-01 4.71288830e-01 6.66786209e-02
-2.50758111e-01 6.62257373e-01 9.00993228e-01 -1.25195488e-01
2.04896659e-01 -3.92247349e-01 -5.86541533e-01 -2.80631095e-01
-3.72722792e-03 6.24712765e-01 2.05115199e-01 -3.79949868e-01
-1.14652187e-01 -6.51673228e-03 -1.57995605e+00 5.78574955e-01
4.24763530e-01 1.03433359e+00 3.33103508e-01 1.12616289e+00
-3.17458600e-01 6.34995639e-01 -7.59349048e-01 -5.42082727e-01
-7.25427195e-02 -5.19125462e-01 -8.47010314e-01 -6.23225458e-02
2.29386270e-01 -1.19794026e-01 -1.99251905e-01 6.25660062e-01
9.83862162e-01 -6.43688589e-02 5.32171905e-01 -1.26931989e+00
-8.51814210e-01 1.10650384e+00 2.89192975e-01 1.25507787e-01
8.93097520e-01 1.41556770e-01 1.41020060e+00 -9.21740115e-01
3.82654995e-01 1.03043675e+00 7.18887568e-01 7.75236845e-01
-1.22134376e+00 -4.87387449e-01 -1.55269086e-01 4.74728584e-01
-1.24459136e+00 -6.13703251e-01 1.06370437e+00 -2.16412425e-01
9.88722146e-01 3.82592291e-01 4.72787410e-01 1.58322132e+00
-1.38570026e-01 6.98076963e-01 7.73976326e-01 -7.17522323e-01
1.62079588e-01 2.77324259e-01 -2.63708308e-02 1.97678372e-01
-1.96234599e-01 -2.81437039e-01 -7.99683332e-01 -2.09529668e-01
2.82478482e-01 -7.35473752e-01 -6.05189025e-01 -1.51120424e-01
-1.36789978e+00 7.04505265e-01 4.51876298e-02 6.05404079e-01
-4.83193189e-01 3.85822803e-01 6.70864999e-01 2.22601131e-01
4.08891112e-01 1.04218304e+00 -4.52008963e-01 -6.51368439e-01
-9.08906460e-01 1.52942821e-01 7.48387575e-01 9.61438239e-01
3.77201349e-01 1.29397735e-01 -5.19764960e-01 1.15568328e+00
1.83891267e-01 1.80249527e-01 8.81268263e-01 -6.78854048e-01
6.38885796e-01 -1.58060081e-02 1.79269448e-01 -6.48707688e-01
-1.52209103e-01 -4.12496239e-01 -2.82853663e-01 -4.16226268e-01
3.75396550e-01 -4.12384242e-01 -7.96561837e-01 1.64934182e+00
-2.83978015e-01 4.66129929e-01 2.75136441e-01 1.06250989e+00
7.63068616e-01 1.18511736e+00 1.98916480e-01 1.29528508e-01
1.50037110e+00 -1.21931350e+00 -1.05035639e+00 -3.64388138e-01
4.63734776e-01 -1.24677539e+00 1.47177613e+00 3.63278121e-01
-9.81548190e-01 -6.93411827e-01 -1.06967938e+00 -1.45567641e-01
-3.59207332e-01 2.38056645e-01 3.83884609e-01 7.20105648e-01
-1.08815801e+00 3.31309259e-01 -5.34491539e-01 -3.84430259e-01
-2.00149529e-02 1.92335322e-01 -2.14377403e-01 1.47744408e-02
-1.50234306e+00 6.86478019e-01 2.65810490e-01 -7.57043809e-02
-1.04500079e+00 -8.84777069e-01 -9.35362935e-01 3.10773313e-01
8.22785273e-02 -6.74146414e-01 1.69320416e+00 -9.15983856e-01
-1.75302494e+00 3.56708586e-01 -1.72950774e-01 -7.30667472e-01
1.91100538e-01 -3.69002879e-01 -7.62187481e-01 3.37441206e-01
-1.04524598e-01 1.04093027e+00 8.66488874e-01 -1.11939013e+00
-3.61601710e-01 2.35117942e-01 1.54436633e-01 1.34894669e-01
-7.83745766e-01 1.71373352e-01 -2.79773563e-01 -8.39653790e-01
-5.35063982e-01 -8.69057953e-01 -2.56867185e-02 -9.64467302e-02
-3.17446887e-01 -2.63308585e-01 6.07671022e-01 -7.02087820e-01
1.31773186e+00 -2.42457151e+00 7.36671612e-02 -2.86792070e-01
-2.51508355e-01 3.28034908e-01 -5.32858670e-01 8.38806689e-01
-2.44841918e-01 2.26835936e-01 -8.71323887e-03 -5.87825775e-01
4.47425216e-01 1.15273751e-01 -7.46755064e-01 -1.42010510e-01
5.53844154e-01 7.29132056e-01 -9.33173060e-01 -5.16291797e-01
1.15513362e-01 8.77837420e-01 -6.71062350e-01 5.61646402e-01
-3.68461788e-01 2.02363402e-01 -1.79441854e-01 2.28767499e-01
1.70298129e-01 8.41148570e-02 -4.47995067e-02 -2.13593557e-01
9.95552614e-02 7.56775856e-01 -9.05948699e-01 2.01465416e+00
-1.02123666e+00 1.11203182e+00 -2.44988844e-01 -8.03619802e-01
9.47144747e-01 9.30390596e-01 3.38318825e-01 -3.92072022e-01
2.23237753e-01 3.42900306e-01 -9.77934599e-02 -6.97774768e-01
7.01078176e-01 -2.20792055e-01 -1.69620752e-01 3.91358733e-01
4.21556413e-01 -3.66013050e-01 9.18834060e-02 3.80500359e-03
1.11672056e+00 -7.02696294e-02 -6.48568347e-02 1.04910523e-01
5.21442473e-01 1.41029563e-02 -1.29119843e-01 4.74217623e-01
-1.34013548e-01 1.21859753e+00 3.43380809e-01 -8.40014368e-02
-1.19063318e+00 -1.05050743e+00 5.67686222e-02 1.31090891e+00
-2.62060106e-01 -7.34534562e-01 -7.77838886e-01 -2.45706975e-01
-4.86329466e-01 1.01052260e+00 -4.18235362e-01 -1.23290010e-01
-4.49595422e-01 -2.92305976e-01 7.72296607e-01 5.80563366e-01
5.69773726e-02 -1.21653366e+00 -2.64231354e-01 6.08794570e-01
-5.23781240e-01 -1.37187636e+00 -7.29944170e-01 2.08868384e-01
-4.90970731e-01 -4.93750483e-01 -1.13947153e+00 -1.12041521e+00
1.16553299e-01 -9.83511880e-02 1.12518227e+00 -3.71999979e-01
8.08017552e-02 5.97600996e-01 -9.23872173e-01 -6.93814158e-01
-6.29713237e-01 4.49686706e-01 1.34832188e-01 2.29943365e-01
4.21137601e-01 -6.52931750e-01 -1.91179231e-01 1.20484702e-01
-8.86368334e-01 -7.67189870e-03 5.12935996e-01 6.39366925e-01
2.95997471e-01 -4.01374072e-01 8.15791428e-01 -2.88119763e-01
1.21511292e+00 -2.23393917e-01 3.70149463e-02 1.18973449e-01
-1.89115465e-01 3.16648722e-01 8.87236834e-01 -7.88562179e-01
-8.41175079e-01 1.58277899e-01 -6.35526836e-01 -4.28857505e-01
-2.30530143e-01 4.16068763e-01 -1.08354092e-01 4.11939472e-01
6.43197119e-01 3.37524861e-01 -1.60296127e-01 -6.78579807e-01
5.25319636e-01 1.22397304e+00 5.03013730e-01 -6.23197138e-01
7.05358982e-01 -4.58658719e-03 -6.56555831e-01 -8.90232503e-01
-7.79585779e-01 -4.43407834e-01 -2.42945284e-01 -2.96316594e-01
9.29155171e-01 -9.65748310e-01 -1.10557295e-01 -6.42374083e-02
-1.48654783e+00 -2.53050357e-01 -5.58911443e-01 7.19105482e-01
-8.46392810e-01 1.97721511e-01 -4.05165315e-01 -7.73380697e-01
-3.41534585e-01 -1.17321575e+00 1.43336511e+00 -1.52611360e-01
-6.68977320e-01 -7.70182848e-01 4.66254413e-01 3.20322841e-01
6.26958728e-01 -2.22654656e-01 7.31258333e-01 -8.56007516e-01
-1.74716398e-01 -1.18848689e-01 3.38771641e-02 5.40818870e-01
1.67355716e-01 -1.75126523e-01 -1.37937665e+00 6.46179765e-02
-3.85739774e-01 -4.34953481e-01 6.06913626e-01 1.39765188e-01
1.22647536e+00 -2.55659908e-01 8.67238864e-02 1.65599853e-01
1.14663446e+00 1.97094470e-01 7.11881936e-01 2.24635541e-01
2.69580871e-01 4.88006234e-01 4.67008710e-01 3.64165008e-01
1.08763576e-01 9.14211094e-01 3.09582949e-01 1.29577413e-01
-5.89829624e-01 -6.83274567e-01 6.52807176e-01 1.31413162e+00
2.10107610e-01 -4.51654941e-01 -7.61050582e-01 9.27049816e-01
-1.41345406e+00 -8.39960992e-01 8.96206945e-02 1.70777571e+00
1.11037803e+00 1.18973270e-01 1.02999188e-01 5.56969404e-01
7.27310419e-01 2.48173386e-01 3.39657962e-02 -8.63670409e-01
1.40779689e-01 5.46716928e-01 -5.31414859e-02 5.15992820e-01
-8.71456683e-01 6.29196703e-01 6.80033827e+00 8.24902117e-01
-1.23397028e+00 3.30728859e-01 2.00012594e-01 -1.71617568e-02
-4.73501652e-01 -2.88135111e-01 -4.96130675e-01 4.50451612e-01
1.58615279e+00 -3.04882824e-01 3.41158748e-01 8.47435117e-01
3.63602132e-01 5.47278106e-01 -1.38522863e+00 1.12840652e+00
3.31309766e-01 -1.28885972e+00 3.06899339e-01 -2.90727228e-01
3.98459375e-01 -1.63906708e-01 1.12990126e-01 5.42092562e-01
-1.98652074e-01 -9.21853423e-01 9.29648876e-01 3.33010852e-01
7.53878653e-01 -5.02614141e-01 9.60364878e-01 -2.48074412e-01
-1.26263452e+00 2.55875718e-02 -1.82969704e-01 -2.31796011e-01
5.71213365e-01 3.53795707e-01 -1.29788268e+00 6.35912359e-01
5.75680792e-01 4.70544964e-01 -6.11638606e-01 1.20556784e+00
-2.74060011e-01 1.06238496e+00 -2.72282958e-01 -4.94601220e-01
3.44099015e-01 3.26249033e-01 5.55223942e-01 1.41155446e+00
6.02262855e-01 -3.68587017e-01 -2.77467817e-01 5.85065365e-01
-1.37845069e-01 4.76177692e-01 -5.08645356e-01 -5.29482782e-01
3.07489783e-01 9.10667300e-01 -1.40971228e-01 -3.22696418e-01
-3.69720668e-01 9.23854649e-01 -4.74287905e-02 2.29908004e-01
-1.10898507e+00 -7.02287316e-01 7.23319471e-01 2.85227507e-01
4.29305375e-01 -3.03996623e-01 1.05237812e-01 -1.03088760e+00
7.01731741e-02 -8.75433207e-01 1.51442200e-01 -1.44033825e+00
-1.32173491e+00 1.15281403e+00 -1.07954144e-02 -1.48409927e+00
-4.77466792e-01 -6.22492075e-01 -6.27370834e-01 6.18274271e-01
-1.57890475e+00 -9.30708289e-01 -2.04491332e-01 2.43308619e-01
1.01743424e+00 -1.14479594e-01 1.14812255e+00 6.80313230e-01
-2.43024513e-01 5.68924963e-01 -1.82374313e-01 1.00231119e-01
9.55062330e-01 -1.14859414e+00 2.91497678e-01 5.80597460e-01
6.63454652e-01 1.99728251e-01 1.21359110e+00 2.47138627e-02
-1.01350236e+00 -1.18482327e+00 1.26786542e+00 -4.02627289e-01
9.92180824e-01 -6.41976535e-01 -8.73499215e-01 5.70310771e-01
8.19892704e-01 -2.95136392e-01 9.07390773e-01 1.04740250e-03
-2.55941540e-01 -2.74487704e-01 -5.71742058e-01 5.48725903e-01
8.49939883e-01 -7.59274542e-01 -1.06811142e+00 3.06887001e-01
1.18179810e+00 -1.12823211e-01 -7.55745471e-01 -8.65315460e-03
5.09106934e-01 -5.03207386e-01 8.16107631e-01 -6.85340703e-01
5.68451524e-01 -2.60405034e-01 -2.22549617e-01 -1.62475228e+00
-5.13879769e-02 -8.83105516e-01 2.41185263e-01 1.61034381e+00
8.23688805e-01 -3.30461115e-01 5.65821588e-01 2.35627711e-01
-5.26461959e-01 -4.99069989e-01 -1.04446924e+00 -8.94306004e-01
-2.04905197e-01 -9.78026748e-01 8.40923846e-01 5.76511860e-01
3.16838712e-01 6.65921271e-01 -5.80362678e-01 1.10959746e-01
-1.54136131e-02 -1.68656215e-01 7.76161134e-01 -9.42757487e-01
-2.68796295e-01 -1.12478696e-01 -5.68657458e-01 -1.12279916e+00
2.19608009e-01 -8.81093860e-01 4.21949446e-01 -1.56251955e+00
-3.16456765e-01 -1.61663413e-01 -2.61512488e-01 5.19512236e-01
1.22831501e-01 4.68648106e-01 3.19927990e-01 -2.70205438e-01
-4.61335868e-01 8.51107657e-01 9.84601676e-01 -5.00818014e-01
-2.28012860e-01 -2.32886165e-01 -6.05325878e-01 2.45749429e-01
1.00812495e+00 -6.31965995e-01 -4.55268592e-01 -7.69987822e-01
1.23137124e-01 -9.58809108e-02 1.90230370e-01 -1.43755662e+00
1.02797963e-01 3.16331863e-01 -1.39719844e-01 -1.76721111e-01
8.58507991e-01 -9.33069944e-01 6.49479851e-02 7.85948187e-02
-7.28889585e-01 1.08571030e-01 3.65400881e-01 3.59371513e-01
-6.01979017e-01 -6.48655295e-01 4.38029855e-01 2.29984120e-01
-5.04379928e-01 -1.11521967e-01 -7.45809257e-01 3.85861322e-02
7.55348444e-01 2.80682035e-02 2.94962581e-02 -8.51383865e-01
-8.39134693e-01 -1.01421483e-01 -8.92267749e-02 8.61260533e-01
6.08227253e-01 -1.56684446e+00 -8.62161398e-01 -4.88794502e-03
3.72278363e-01 -5.32218575e-01 -2.75655054e-02 4.56656784e-01
-3.99134040e-01 7.74398148e-01 -2.54306585e-01 -4.28449839e-01
-1.11018503e+00 4.26369041e-01 1.42089605e-01 -2.02341840e-01
-2.47427866e-01 7.40445793e-01 -4.09358948e-01 -1.25944406e-01
4.12553668e-01 -7.05176830e-01 -3.36061418e-01 1.42417252e-01
6.00969434e-01 1.54485650e-04 6.49010763e-02 -6.34173453e-01
-3.44178587e-01 4.78878707e-01 2.84595817e-01 -5.08517087e-01
1.29488993e+00 9.08650905e-02 3.74116153e-01 6.10055864e-01
1.38096941e+00 1.89747840e-01 -7.01014221e-01 6.46780655e-02
-1.33705005e-01 -2.56378859e-01 1.35999262e-01 -5.46481252e-01
-7.13080347e-01 1.23648548e+00 6.25497878e-01 4.99595076e-01
1.17612410e+00 1.18053630e-01 1.00852406e+00 2.23688781e-01
2.99365729e-01 -9.28200185e-01 3.93040866e-01 3.11669022e-01
1.13954675e+00 -8.49979162e-01 -4.63518023e-01 -3.71615529e-01
-8.08703482e-01 1.28082991e+00 3.86855036e-01 2.46549100e-02
4.25257832e-01 1.48292154e-01 2.24189684e-01 2.48830914e-01
-8.34370852e-01 -4.16372061e-01 3.85519147e-01 8.19391549e-01
5.53607762e-01 -1.21311501e-01 -2.15504006e-01 9.34706926e-01
-6.73925638e-01 2.66685039e-02 8.07047009e-01 4.50738758e-01
-2.63357103e-01 -1.50390804e+00 -4.31957573e-01 2.27418482e-01
-5.48116863e-01 -2.29767799e-01 -5.19338012e-01 4.61948067e-01
-6.37389868e-02 1.22823036e+00 1.45115405e-01 -5.43614745e-01
5.07667899e-01 4.83276010e-01 2.03786612e-01 -8.32700491e-01
-5.24713278e-01 -3.09097189e-02 5.53184688e-01 -2.74009109e-01
-4.83024925e-01 -4.51252371e-01 -8.85775208e-01 3.64593327e-01
-3.47641855e-01 6.93123877e-01 7.90433109e-01 7.25582123e-01
4.50926036e-01 8.89285624e-01 4.70885068e-01 -9.29729939e-01
-4.87521887e-01 -1.28020024e+00 -2.42222130e-01 3.22321236e-01
2.90635765e-01 -3.64700615e-01 -4.74229693e-01 3.59991372e-01] | [15.2935152053833, 4.884670734405518] |
840af2ce-6e9a-4629-ad1e-4e6d132f3971 | data-augmentation-with-sentence-recombination | null | null | https://openreview.net/forum?id=WSxxdJdByLS | https://openreview.net/pdf?id=WSxxdJdByLS | Data Augmentation with Sentence Recombination Method for Semi-supervised Text Classification | As the need of large amount of time and expertise to obtain enough labeled data, semi-supervised learning has received much attention to utilize both labeled and unlabeled data. In this paper, we present SeRe: a Sentence Recombination method to augment training data for semi-supervised text classification. SeRe makes full use of the similarities between sentences in different samples through the grouping and recombining process to form rich and varied training data. SeRe generates data from three combinations, including labeled, unlabeled, and mixed data. Meanwhile, SeRe combines the self-training framework to improve the quality of augmented training data iteratively. We apply SeRe to text classification tasks and conduct extensive experiments on four publicly available benchmarks. Experimental results show that SeRe achieves new state-of-the-art performances on all of them. | ['Anonymous'] | 2021-11-16 | null | null | null | acl-arr-november-2021-11 | ['semi-supervised-text-classification-1'] | ['natural-language-processing'] | [ 2.75451928e-01 -7.14644343e-02 -4.14376706e-01 -7.49253988e-01
-8.50870311e-01 -4.52968836e-01 5.89903653e-01 2.26626948e-01
-4.64323103e-01 9.79716718e-01 1.53769180e-01 -2.07514122e-01
4.02608931e-01 -4.57234651e-01 -4.07349527e-01 -3.87820363e-01
4.32393909e-01 5.05251408e-01 5.03713340e-02 -1.78121880e-01
-4.21489263e-03 -2.58556515e-01 -1.30131590e+00 7.07431734e-01
1.37935436e+00 8.04754376e-01 2.23879904e-01 3.45064878e-01
-6.69646740e-01 9.46985483e-01 -6.76741600e-01 -5.01088858e-01
4.54880577e-03 -7.17555463e-01 -9.58882630e-01 6.26518071e-01
1.72993332e-01 -5.64605854e-02 -1.76492661e-01 6.98403358e-01
4.91169870e-01 3.06890532e-02 4.67513382e-01 -1.31289995e+00
-7.47697234e-01 1.13021588e+00 -5.67203581e-01 5.42045869e-02
3.23444486e-01 -3.61972660e-01 8.91486466e-01 -1.44344306e+00
3.71053427e-01 9.65438843e-01 4.62821484e-01 5.74249625e-01
-1.09866667e+00 -7.21534193e-01 2.62824208e-01 -4.97636274e-02
-1.20698440e+00 -4.52450514e-01 9.49682713e-01 -1.01656042e-01
6.48184776e-01 9.15790945e-02 3.03676158e-01 1.20870852e+00
-3.07116181e-01 1.52369177e+00 1.21938396e+00 -8.74010444e-01
1.50179923e-01 5.54180980e-01 7.11902678e-01 6.59310579e-01
1.03273712e-01 -4.13772315e-01 -7.31055856e-01 -2.19210401e-01
1.29072964e-01 1.04904972e-01 -3.53107989e-01 -8.97381976e-02
-1.21685457e+00 6.54020607e-01 1.69135243e-01 2.83984303e-01
8.87632649e-03 -6.23602808e-01 6.67379677e-01 4.28667426e-01
9.51815963e-01 4.01667416e-01 -6.67154014e-01 -8.80909190e-02
-8.85470986e-01 -2.69036084e-01 7.01032102e-01 1.22755539e+00
6.73535764e-01 -5.00009805e-02 -3.86536926e-01 1.45858431e+00
1.28000736e-01 2.56860465e-01 8.73449922e-01 -4.40553933e-01
1.12917733e+00 1.01952028e+00 -1.56230286e-01 -4.02440399e-01
-2.26714000e-01 -5.64950287e-01 -1.16412425e+00 -5.63690364e-01
1.29219159e-01 -3.84907842e-01 -8.31623852e-01 1.44585335e+00
2.62237877e-01 6.40036315e-02 2.32182771e-01 4.03779864e-01
1.16177368e+00 7.88598895e-01 -8.97081345e-02 -3.71744424e-01
9.22691822e-01 -1.69919240e+00 -9.01406348e-01 -3.79221976e-01
1.02358234e+00 -5.70789874e-01 1.17538214e+00 2.76746154e-01
-8.94887447e-01 -7.08508313e-01 -1.01944268e+00 -5.12869433e-02
-2.98687369e-01 6.65782034e-01 4.74738419e-01 5.50233841e-01
-5.43678820e-01 2.90810436e-01 -6.92150652e-01 -1.16150424e-01
5.83183825e-01 1.36664361e-01 -3.52529794e-01 -2.95489937e-01
-1.23801267e+00 3.32263798e-01 6.94814920e-01 1.71155289e-01
-2.79531300e-01 -5.52822351e-01 -9.62398648e-01 1.65000279e-02
4.80014861e-01 -4.44999218e-01 1.29196429e+00 -1.02317119e+00
-1.40835714e+00 9.05386209e-01 -4.41686928e-01 -2.38995001e-01
5.55577278e-01 -2.02180833e-01 -3.99595290e-01 -7.21139163e-02
6.94096684e-02 6.93403661e-01 6.02600098e-01 -1.33345127e+00
-5.88031173e-01 -4.02455568e-01 -5.30132413e-01 3.10215265e-01
-9.60046828e-01 1.66020021e-02 -5.15560925e-01 -9.40939844e-01
9.72200036e-02 -8.07380021e-01 -5.50594665e-02 -2.97278970e-01
-5.95726013e-01 -6.19695961e-01 1.01706004e+00 -4.90411937e-01
1.27941215e+00 -2.02523541e+00 -3.42967361e-02 -2.14143276e-01
2.94482857e-01 5.80954850e-01 -2.29117259e-01 5.23319364e-01
-1.38911054e-01 2.04227313e-01 -4.01023388e-01 -9.31568563e-01
-9.12786275e-02 2.04882085e-01 -4.72922653e-01 -2.10710932e-02
2.60719150e-01 9.74150002e-01 -1.00731218e+00 -9.25487757e-01
-4.17162031e-02 3.03965844e-02 -1.21799648e-01 3.56677055e-01
-1.90033019e-01 4.86062825e-01 -6.70380235e-01 5.62773526e-01
6.76804245e-01 -5.38974702e-01 8.04605260e-02 1.48788318e-01
3.54849696e-01 3.84997249e-01 -8.25988233e-01 1.82016182e+00
-3.91075015e-01 5.01145899e-01 -2.80886978e-01 -1.01427376e+00
1.10108209e+00 2.90012658e-01 3.16554844e-01 -4.07868564e-01
2.38909379e-01 2.27118924e-01 -4.40872192e-01 -3.72621000e-01
6.32808089e-01 7.47278035e-02 -1.17144883e-01 8.03127646e-01
2.28369296e-01 -6.10493086e-02 5.50192177e-01 6.55510604e-01
7.48425603e-01 -5.21297939e-02 2.90977359e-01 3.65783200e-02
7.17892110e-01 -4.81427424e-02 5.89651883e-01 5.78170657e-01
-1.17129505e-01 4.80916142e-01 2.97277719e-01 -8.42216313e-02
-8.85005891e-01 -6.55693591e-01 -2.21508339e-01 1.21184719e+00
2.53748838e-02 -4.99962598e-01 -6.05176270e-01 -1.38809931e+00
-1.05201863e-01 6.43479824e-01 -6.59639359e-01 -2.59638250e-01
-2.79165000e-01 -7.98489511e-01 4.31461245e-01 6.88622952e-01
8.54965091e-01 -1.04623580e+00 3.01031947e-01 6.98588192e-02
-4.39432114e-01 -1.10309684e+00 -6.95293128e-01 2.90962040e-01
-1.02491987e+00 -7.37370729e-01 -6.89356923e-01 -1.17817235e+00
1.11693561e+00 7.35414386e-01 1.24887466e+00 2.44661257e-01
4.54475246e-02 1.58258285e-02 -1.07195115e+00 -2.92942554e-01
-7.31288135e-01 3.16461533e-01 -1.77020609e-01 4.24739160e-02
2.75038689e-01 -1.81042612e-01 -1.69353969e-02 6.36284053e-01
-9.26030278e-01 4.62345749e-01 3.94186527e-01 1.37738383e+00
4.27216202e-01 1.35967791e-01 1.00389349e+00 -1.27953970e+00
7.03036904e-01 -4.74950433e-01 -2.13795066e-01 6.50881469e-01
-7.13710070e-01 2.94696204e-02 9.37820435e-01 -5.13151646e-01
-1.27810538e+00 -3.44669037e-02 2.95411926e-02 -1.32378355e-01
-4.16416563e-02 8.32694709e-01 -2.38495648e-01 3.51051122e-01
5.67486823e-01 4.40370768e-01 -1.98370799e-01 -5.41617990e-01
3.43625724e-01 1.41314268e+00 4.15675282e-01 -6.06108785e-01
5.63617945e-01 2.15648443e-01 -6.50705636e-01 -5.83542764e-01
-1.41147709e+00 -5.22331655e-01 -8.04150641e-01 -4.27718386e-02
2.16478541e-01 -8.96194160e-01 7.45842084e-02 7.08423734e-01
-9.86203611e-01 -3.10898155e-01 -2.96684176e-01 4.28499162e-01
2.08424237e-02 5.66099465e-01 -7.56122768e-01 -6.29075110e-01
-6.04923069e-01 -8.88771892e-01 1.05980408e+00 1.91024393e-01
9.68680680e-02 -1.04586172e+00 4.36564982e-02 8.16363275e-01
-5.32826483e-02 -3.01780671e-01 5.62828898e-01 -1.14193428e+00
-3.77163216e-02 -4.08357263e-01 -1.81175321e-01 6.88732088e-01
5.55828691e-01 -8.47298503e-02 -8.60469878e-01 -4.53475535e-01
-4.63175168e-03 -1.03289139e+00 1.10348308e+00 -1.58080444e-01
1.44372201e+00 -1.57084599e-01 -3.76268059e-01 2.49209493e-01
8.31745505e-01 9.67104286e-02 3.00739527e-01 5.88852018e-02
8.79866719e-01 7.46418893e-01 8.25834274e-01 5.13040245e-01
4.80116963e-01 5.06850481e-01 -2.70757526e-01 -2.63973862e-01
5.66737428e-02 -3.60778838e-01 2.35815570e-01 1.43075955e+00
6.37725472e-01 -5.83144903e-01 -8.64817977e-01 3.47756863e-01
-1.99372470e+00 -7.98011124e-01 -2.32788280e-01 1.99755704e+00
1.45152152e+00 2.10959703e-01 2.17412170e-02 4.27921712e-01
1.00448799e+00 2.01628841e-02 -6.98960841e-01 9.44980606e-02
-3.89611334e-01 4.38088030e-02 9.32931900e-02 1.33295313e-01
-1.25470078e+00 1.00560474e+00 6.36675358e+00 1.08912337e+00
-9.27572727e-01 -3.33643099e-03 9.44252670e-01 5.71456971e-04
-1.74673721e-01 -2.88081467e-01 -9.08127964e-01 7.60486662e-01
6.44677639e-01 -2.44222075e-01 7.43060783e-02 7.74883628e-01
5.37814982e-02 -1.38817746e-02 -1.00048888e+00 9.09901202e-01
3.02006066e-01 -1.14175034e+00 8.47477391e-02 -4.11943138e-01
1.18986809e+00 7.32137635e-03 -1.80738986e-01 5.50833404e-01
5.98115683e-01 -7.16506720e-01 4.96007085e-01 2.38926578e-02
6.12662137e-01 -5.40886521e-01 9.35269773e-01 7.04901516e-01
-1.17233706e+00 1.53483180e-02 -2.09671497e-01 8.37711021e-02
3.21102999e-02 7.90721536e-01 -9.30425882e-01 8.28588963e-01
3.01586568e-01 1.14766121e+00 -1.00818563e+00 7.99466550e-01
-3.72784615e-01 1.08291495e+00 -1.44083813e-01 -2.82676786e-01
-7.64102163e-03 -1.83542117e-01 6.43135682e-02 1.20519280e+00
-1.87053502e-01 -3.52949984e-02 6.28330648e-01 5.50107896e-01
-4.11938548e-01 3.67263556e-01 -2.07100123e-01 -1.49575293e-01
7.77302146e-01 1.26259863e+00 -6.89289629e-01 -6.41745508e-01
-5.68781137e-01 1.02816296e+00 6.11613810e-01 2.88471878e-01
-5.01535952e-01 -4.53015357e-01 -3.07119906e-01 -4.22066778e-01
1.64816797e-01 -2.57524452e-03 -4.22109872e-01 -1.49506032e+00
4.38229233e-01 -8.65637124e-01 4.82413590e-01 -8.54684174e-01
-1.61774647e+00 8.21350455e-01 -8.25711861e-02 -1.38611841e+00
-3.17378223e-01 -2.84977138e-01 -5.42906523e-01 6.22466564e-01
-1.35964870e+00 -1.17978954e+00 -4.74157274e-01 3.71674657e-01
1.03058934e+00 -4.79247779e-01 5.15821099e-01 2.05635682e-01
-1.01265669e+00 8.82809460e-01 3.15239400e-01 5.18309295e-01
1.01269817e+00 -1.29897237e+00 5.78139246e-01 6.95032835e-01
3.15331936e-01 4.50402051e-01 1.62795365e-01 -6.73994064e-01
-1.07753241e+00 -1.33322084e+00 9.20079648e-01 -3.20768625e-01
5.26027799e-01 -5.47827303e-01 -1.17999566e+00 7.23069012e-01
3.17766368e-01 7.07509518e-02 1.05035543e+00 4.78125066e-02
-3.73652339e-01 -1.30639464e-01 -8.77713978e-01 3.79382849e-01
8.79544675e-01 -5.91540754e-01 -8.86878014e-01 5.28977692e-01
8.86632919e-01 -3.26267958e-01 -5.86860001e-01 3.23283136e-01
6.14770502e-02 -6.75371945e-01 4.51927096e-01 -6.30408466e-01
6.92793548e-01 -8.45327824e-02 1.81500524e-01 -1.51430082e+00
2.22039208e-01 -4.92455572e-01 5.83799146e-02 1.79782820e+00
7.03642964e-01 -6.87850773e-01 9.18526947e-01 4.85106200e-01
-3.20877433e-01 -1.01055086e+00 -5.55709064e-01 -8.03422689e-01
8.53635445e-02 -1.58480391e-01 4.30699468e-01 1.45451832e+00
4.74961728e-01 9.05057549e-01 -2.87536055e-01 -4.33777809e-01
4.96582121e-01 3.07148367e-01 8.16085458e-01 -1.24521077e+00
-1.73071608e-01 -1.21775381e-01 1.37159303e-01 -1.45685041e+00
6.02635205e-01 -1.23289561e+00 1.39179543e-01 -1.54623771e+00
6.52150393e-01 -5.23289442e-01 -1.45695105e-01 8.28202963e-01
-8.37175727e-01 7.74920061e-02 3.18985619e-02 4.26121831e-01
-9.92023230e-01 1.11219859e+00 1.36684585e+00 -2.70278543e-01
-4.19291168e-01 3.67577933e-02 -7.36546159e-01 4.85239804e-01
9.20262575e-01 -4.71066028e-01 -5.57833731e-01 -4.93467540e-01
-2.93158870e-02 -3.10777798e-02 -2.28350952e-01 -7.06983030e-01
2.26566687e-01 9.45209414e-02 2.88793951e-01 -8.87192547e-01
5.24632148e-02 -6.17423236e-01 -5.18158317e-01 5.60051017e-02
-8.64294350e-01 -1.50734961e-01 8.96033272e-02 6.05349422e-01
-5.81016123e-01 -4.11832541e-01 5.89963078e-01 1.95059106e-01
-4.16790158e-01 2.45262906e-01 -6.03843965e-02 3.47387940e-01
7.45607376e-01 4.42714728e-02 -5.67409217e-01 -2.79417813e-01
-4.78057683e-01 8.67071629e-01 2.40307704e-01 5.73849380e-01
5.79042315e-01 -1.40596175e+00 -8.67465615e-01 2.12300122e-01
3.42863262e-01 2.98912317e-01 1.39055163e-01 4.28296119e-01
1.72419380e-03 3.59504789e-01 2.17775181e-01 -5.69023967e-01
-1.49691582e+00 6.55659914e-01 3.49829644e-02 -6.11712992e-01
-2.85980731e-01 7.68877506e-01 -1.50863394e-01 -9.39912319e-01
3.58583093e-01 -1.36297390e-01 -2.21763566e-01 -1.83769260e-02
5.08190811e-01 2.14136332e-01 1.30306929e-01 -3.12228411e-01
-6.40241802e-02 1.13431074e-01 -4.93016660e-01 5.64332418e-02
1.05349755e+00 -2.83406675e-01 -2.22238258e-01 6.33230686e-01
1.30425549e+00 1.41258696e-02 -1.03921843e+00 -8.02152872e-01
7.73411989e-02 -3.14874589e-01 6.73934668e-02 -8.91937137e-01
-9.53902841e-01 6.31657064e-01 -5.84761873e-02 2.10556358e-01
1.08695233e+00 4.07973379e-02 9.01708126e-01 8.39627504e-01
7.69022629e-02 -1.17225456e+00 4.29630935e-01 6.23392284e-01
6.20892048e-01 -1.55080748e+00 1.07166588e-01 -7.97458112e-01
-1.00575328e+00 7.95254648e-01 8.11237276e-01 3.25259268e-01
5.84559798e-01 1.87242389e-01 2.01772526e-01 2.78388351e-01
-9.91959274e-01 3.02502550e-02 3.65309238e-01 2.66145021e-01
6.95428193e-01 -2.19389558e-01 -5.06877184e-01 8.97451818e-01
-7.39384741e-02 8.75471681e-02 3.08796674e-01 1.38558221e+00
-3.35385293e-01 -1.60887563e+00 -2.18225271e-01 7.01334715e-01
-6.60258010e-02 -2.30454952e-01 -7.69667029e-01 4.17200148e-01
-2.48654261e-01 1.29609489e+00 -1.38236776e-01 -4.80508745e-01
1.00606777e-01 2.18952343e-01 2.24240243e-01 -1.00175083e+00
-6.02030933e-01 3.67009416e-02 1.83770716e-01 2.61506379e-01
-6.00600541e-01 -5.13746977e-01 -1.31413198e+00 1.68522015e-01
-8.42827857e-01 6.21016979e-01 5.04535139e-01 1.16125679e+00
4.58518177e-01 5.26608229e-01 1.35896993e+00 -5.47104716e-01
-7.53184378e-01 -1.36570156e+00 -5.24199426e-01 4.44432497e-01
1.52103230e-01 -4.10226464e-01 -3.36836040e-01 1.94889382e-01] | [10.635576248168945, 7.406373977661133] |
1d3fc220-168f-4ac6-a445-d140e7049ecb | algorithm-and-hardware-co-design-of-energy | 2212.02046 | null | https://arxiv.org/abs/2212.02046v1 | https://arxiv.org/pdf/2212.02046v1.pdf | Algorithm and Hardware Co-Design of Energy-Efficient LSTM Networks for Video Recognition with Hierarchical Tucker Tensor Decomposition | Long short-term memory (LSTM) is a type of powerful deep neural network that has been widely used in many sequence analysis and modeling applications. However, the large model size problem of LSTM networks make their practical deployment still very challenging, especially for the video recognition tasks that require high-dimensional input data. Aiming to overcome this limitation and fully unlock the potentials of LSTM models, in this paper we propose to perform algorithm and hardware co-design towards high-performance energy-efficient LSTM networks. At algorithm level, we propose to develop fully decomposed hierarchical Tucker (FDHT) structure-based LSTM, namely FDHT-LSTM, which enjoys ultra-low model complexity while still achieving high accuracy. In order to fully reap such attractive algorithmic benefit, we further develop the corresponding customized hardware architecture to support the efficient execution of the proposed FDHT-LSTM model. With the delicate design of memory access scheme, the complicated matrix transformation can be efficiently supported by the underlying hardware without any access conflict in an on-the-fly way. Our evaluation results show that both the proposed ultra-compact FDHT-LSTM models and the corresponding hardware accelerator achieve very high performance. Compared with the state-of-the-art compressed LSTM models, FDHT-LSTM enjoys both order-of-magnitude reduction in model size and significant accuracy improvement across different video recognition datasets. Meanwhile, compared with the state-of-the-art tensor decomposed model-oriented hardware TIE, our proposed FDHT-LSTM architecture achieves better performance in throughput, area efficiency and energy efficiency, respectively on LSTM-Youtube workload. For LSTM-UCF workload, our proposed design also outperforms TIE with higher throughput, higher energy efficiency and comparable area efficiency. | ['Bo Yuan', 'Yang Sui', 'Chunhua Deng', 'Lingyi Huang', 'Miao Yin', 'Yu Gong'] | 2022-12-05 | null | null | null | null | ['video-recognition'] | ['computer-vision'] | [ 1.45540059e-01 -4.43426043e-01 -4.31878895e-01 1.28185311e-02
-4.23588365e-01 7.19592422e-02 1.35574758e-01 -3.99081632e-02
-6.39581740e-01 2.60501057e-01 9.25946981e-03 -8.19133818e-01
8.81444216e-02 -7.24052072e-01 -8.58184218e-01 -8.03843915e-01
-6.92896545e-02 7.35184327e-02 1.94434360e-01 7.06973150e-02
6.00852743e-02 2.37217933e-01 -1.49465215e+00 4.15125579e-01
8.42087865e-01 1.55044663e+00 6.34127319e-01 3.76783729e-01
-2.43669137e-01 1.07014263e+00 -4.06322330e-02 -3.35687906e-01
3.11153512e-02 4.21329029e-02 -4.68905628e-01 -1.73717096e-01
4.37104076e-01 -7.31030941e-01 -7.34250069e-01 8.25819850e-01
3.39339614e-01 -6.80274470e-03 1.32611528e-01 -9.45149302e-01
-3.06959003e-01 6.65175736e-01 -7.58790433e-01 3.98381352e-01
-1.77774221e-01 5.65220155e-02 7.50166953e-01 -9.89308774e-01
1.21647954e-01 8.27087879e-01 7.46202052e-01 3.58611047e-01
-7.59012580e-01 -8.65169704e-01 8.17605108e-02 6.89855516e-01
-1.27499259e+00 -4.84130979e-01 5.98177731e-01 -1.01014022e-02
1.64073360e+00 1.77482590e-01 8.52601111e-01 1.04382885e+00
5.33378839e-01 1.00806940e+00 6.67919219e-01 -3.83131623e-01
1.41072497e-01 -3.41796517e-01 3.65259945e-01 1.09537327e+00
3.75592113e-01 -9.18151215e-02 -7.11818218e-01 -2.44452693e-02
6.62342370e-01 3.61282647e-01 -2.38323078e-01 -1.09318057e-02
-1.30305719e+00 3.49580318e-01 5.36193848e-01 6.24711752e-01
-5.17789066e-01 7.14579403e-01 1.06786752e+00 5.68312667e-02
1.99469239e-01 -2.86501974e-01 -3.89198959e-01 -5.66189706e-01
-1.31636786e+00 -2.04316303e-01 2.76360065e-01 9.27201986e-01
2.88612723e-01 7.43209124e-01 -6.80313781e-02 6.35748267e-01
6.37191683e-02 3.92243683e-01 1.00333345e+00 -4.54501122e-01
6.91173196e-01 5.19149959e-01 -4.42307472e-01 -1.07581604e+00
-4.48437512e-01 -8.09347868e-01 -1.45499301e+00 -5.17468572e-01
-2.27223352e-01 1.82889581e-01 -9.72908795e-01 1.47004247e+00
1.62434205e-01 4.32757795e-01 1.10957749e-01 7.09462464e-01
6.98362350e-01 1.16583586e+00 1.72861278e-01 -5.08575022e-01
1.67359710e+00 -1.29611647e+00 -8.02702069e-01 -2.23225266e-01
1.29126263e+00 -4.92421269e-01 1.00538671e+00 3.59611839e-01
-9.12822187e-01 -7.18257487e-01 -1.38278759e+00 -2.54635066e-01
3.09620090e-02 5.77530921e-01 9.04579163e-01 7.35871732e-01
-9.47252333e-01 6.36206269e-01 -1.46573675e+00 -3.66704464e-01
3.56429040e-01 6.38708055e-01 -4.09179777e-02 8.68044645e-02
-9.37751174e-01 6.22488797e-01 6.66366160e-01 4.16376621e-01
-8.04396272e-01 -6.89645410e-01 -5.06135523e-01 4.69966173e-01
3.85407746e-01 -6.35620832e-01 1.18986607e+00 -6.36519492e-01
-1.53518009e+00 2.68466502e-01 -2.18302622e-01 -9.36491013e-01
9.09109414e-02 -1.31218567e-01 -4.33670759e-01 6.59264997e-02
-6.27780020e-01 4.81256604e-01 6.53331518e-01 -4.70661581e-01
-5.88119268e-01 -4.45845127e-01 -2.32584238e-01 1.85728505e-01
-1.36285210e+00 -3.17123570e-02 -4.05482113e-01 -7.60381043e-01
2.06367850e-01 -1.04604352e+00 -1.15415402e-01 -3.22995409e-02
-1.55471653e-01 -2.23537892e-01 1.33023334e+00 -4.94299948e-01
1.58643818e+00 -1.97262096e+00 -3.90773490e-02 -2.81892180e-01
3.35501730e-01 7.07320273e-01 1.70819148e-01 1.74973696e-01
1.89714253e-01 -1.55510291e-01 1.85587272e-01 -3.93789083e-01
-1.14707582e-01 2.41095766e-01 -5.45304477e-01 3.32458258e-01
-3.43146116e-01 9.18444395e-01 -4.22863662e-01 -5.70309877e-01
2.52243370e-01 5.60471773e-01 -4.97574180e-01 5.29270386e-03
-4.26967405e-02 -2.21449256e-01 -6.05316341e-01 6.46021903e-01
5.67417264e-01 -3.21945012e-01 4.37099695e-01 -7.69401431e-01
-1.40461683e-01 5.29570766e-02 -6.14243329e-01 1.90056479e+00
-6.03327990e-01 7.68521488e-01 -7.64703155e-02 -1.13628709e+00
7.97911763e-01 4.18367088e-01 4.83789235e-01 -9.57997203e-01
3.89328718e-01 5.48045933e-01 4.47198749e-02 -3.45965087e-01
8.14941049e-01 -2.16557160e-02 8.98048654e-02 4.59792048e-01
9.22106281e-02 8.48037064e-01 -4.31739679e-03 1.07691616e-01
8.77370119e-01 1.77325249e-01 -4.80328389e-02 -4.71337110e-01
4.63451028e-01 -1.03231698e-01 5.97921073e-01 2.73115605e-01
-2.82323230e-02 -2.56062686e-01 -2.54982878e-02 -8.18655610e-01
-1.14941859e+00 -4.13792640e-01 -2.59256046e-02 1.06516492e+00
-9.27947089e-02 -6.52698934e-01 -9.34062719e-01 -3.78730074e-02
-4.73854512e-01 4.36617941e-01 -2.64093459e-01 -2.41367936e-01
-1.02030802e+00 -8.80612433e-01 8.76984656e-01 6.86705709e-01
1.24019611e+00 -5.84696770e-01 -1.35262382e+00 4.04881507e-01
-1.79557279e-01 -1.47003460e+00 -4.94804651e-01 1.35468587e-01
-1.41473830e+00 -3.80179286e-01 -5.02973437e-01 -9.24046457e-01
3.01060319e-01 4.96525198e-01 5.95971942e-01 2.28932306e-01
-6.00829814e-03 -3.17770094e-01 -2.57905394e-01 9.35955495e-02
-1.02786897e-02 2.84990162e-01 3.85277063e-01 -7.88802374e-03
2.40968347e-01 -7.88223803e-01 -7.87757397e-01 1.16428018e-01
-8.87210906e-01 6.16342187e-01 6.94687188e-01 9.45124686e-01
5.84674954e-01 1.98391750e-01 1.63601816e-01 -4.17194784e-01
1.10325754e-01 -1.36268824e-01 -5.58769464e-01 3.69534582e-01
-7.58724809e-01 2.61680752e-01 1.17685974e+00 -5.13622880e-01
-9.01708424e-01 1.51946088e-02 -3.18850964e-01 -9.22153592e-01
6.40530646e-01 8.01191330e-01 3.63373868e-02 -2.32136562e-01
1.22299448e-01 8.05010974e-01 -3.12568277e-01 -4.66924250e-01
-1.53993070e-01 5.51571369e-01 5.08777201e-01 -5.61704338e-01
2.39158615e-01 3.00300419e-01 2.14711547e-01 -8.10178339e-01
-4.48241651e-01 1.15122244e-01 -2.42039204e-01 -2.11608380e-01
7.09713757e-01 -1.01355302e+00 -9.71170545e-01 6.51587129e-01
-1.09236681e+00 -3.48803133e-01 3.17662120e-01 5.78403771e-01
-2.51702309e-01 4.71075475e-01 -1.03954756e+00 -7.27733076e-01
-1.19458604e+00 -1.44873691e+00 8.04588318e-01 -1.19207107e-01
2.75444776e-01 -8.41196954e-01 -5.03895104e-01 2.76122123e-01
6.53054655e-01 1.49562612e-01 1.15388548e+00 -2.57485628e-01
-8.65145147e-01 -7.84610435e-02 -2.62127370e-01 5.12048192e-02
-2.20953509e-01 8.67176056e-03 -8.23380649e-01 -7.19949007e-01
3.50987613e-01 -3.32190692e-01 1.12049949e+00 3.65721911e-01
1.45640695e+00 -5.47221482e-01 -5.69451213e-01 9.73004162e-01
1.61443138e+00 3.46681446e-01 5.77996373e-01 2.28622139e-01
1.10116363e+00 8.06046836e-03 5.18729210e-01 3.91050130e-01
3.28196079e-01 8.35810542e-01 3.23347539e-01 -5.97037487e-02
2.13493079e-01 -3.41944426e-01 5.87696433e-01 1.82260323e+00
-5.23707345e-02 -3.02127898e-01 -8.43105435e-01 1.62010074e-01
-1.79069412e+00 -7.82690942e-01 -3.16734344e-01 2.12984705e+00
4.46156025e-01 4.34061021e-01 -1.00073695e-01 3.94268602e-01
3.81742239e-01 4.41551238e-01 -6.57851219e-01 -7.09488153e-01
1.70592189e-01 5.80554642e-02 7.30564952e-01 -8.86181742e-02
-8.48773360e-01 7.95378268e-01 4.75486517e+00 1.48327148e+00
-1.52317178e+00 4.61258888e-01 7.12652683e-01 -3.89802903e-01
9.44341123e-02 -1.19828135e-01 -7.51055777e-01 3.54561090e-01
1.65183008e+00 -1.57104298e-01 1.64666265e-01 8.98561478e-01
1.95205793e-01 1.08204238e-01 -1.13786376e+00 1.31009150e+00
-9.05807763e-02 -1.67080462e+00 2.62518287e-01 3.84393692e-01
2.98190504e-01 1.89785913e-01 2.27847874e-01 4.65744823e-01
-4.69530791e-01 -7.87492216e-01 8.07302237e-01 -2.00058576e-02
1.07939112e+00 -1.03722203e+00 6.19822919e-01 5.60818255e-01
-1.77358711e+00 -3.57153267e-01 -7.35723317e-01 -1.09180883e-01
1.45303816e-01 6.58636153e-01 -2.70665139e-01 5.31394362e-01
9.10821915e-01 5.61336100e-01 -1.71890557e-01 6.15004182e-01
7.39370942e-01 6.05190814e-01 -2.91026980e-01 -1.68252677e-01
6.25032604e-01 -7.38224164e-02 1.26317307e-01 1.24867249e+00
8.29322398e-01 3.57683331e-01 1.79052263e-01 4.57422018e-01
-2.62777269e-01 1.72036126e-01 -4.52484637e-01 -2.33757675e-01
5.39489090e-01 1.32003319e+00 -8.19358110e-01 -7.43009865e-01
-3.05774450e-01 9.59851623e-01 2.33192846e-01 -1.09012172e-01
-1.21710813e+00 -1.48178831e-01 5.34742475e-01 -9.29701403e-02
3.75060946e-01 -5.23380578e-01 -1.75120547e-01 -1.33546829e+00
2.55679846e-01 -7.59321809e-01 2.13981643e-01 -4.88802075e-01
-3.38709146e-01 1.17552304e+00 -3.98011416e-01 -1.41546178e+00
-4.95242178e-02 -6.47009909e-01 -3.84194165e-01 3.01264405e-01
-1.26680911e+00 -1.32720780e+00 -4.35906112e-01 5.02008796e-01
8.04568112e-01 -1.55854777e-01 8.05356681e-01 6.67911828e-01
-1.15600538e+00 1.01177907e+00 3.78577501e-01 -5.46920411e-02
4.59948443e-02 -3.74682307e-01 5.95350504e-01 8.93938720e-01
-1.11444853e-03 7.27376223e-01 5.51312804e-01 -4.77118373e-01
-2.26186252e+00 -1.30614257e+00 3.61605823e-01 3.73975396e-01
6.61250651e-01 -4.28346246e-01 -9.65196490e-01 6.46197379e-01
2.37430230e-01 7.09693655e-02 5.43438554e-01 -3.50223720e-01
-2.45810911e-01 -3.80444944e-01 -8.67711425e-01 6.22725964e-01
1.11839569e+00 -4.87871647e-01 -7.06273317e-02 4.12219286e-01
8.77210557e-01 -5.68662703e-01 -9.96974289e-01 5.16471982e-01
8.43293190e-01 -1.00169182e+00 7.46180177e-01 -1.43150687e-01
3.75789583e-01 -5.26624806e-02 -4.11110997e-01 -7.52096474e-01
-2.70414799e-01 -4.74975199e-01 -6.23545468e-01 9.08728719e-01
3.39738131e-02 -4.50528890e-01 9.97979760e-01 4.41447854e-01
-4.74930346e-01 -1.35461628e+00 -1.21854782e+00 -8.89527857e-01
-2.37940177e-01 -6.21492386e-01 5.30707657e-01 6.84334576e-01
2.02369168e-01 2.50393212e-01 -7.21955776e-01 -1.12588234e-01
6.49477124e-01 9.67959166e-02 4.93515015e-01 -7.09111512e-01
-3.67444068e-01 -3.58344495e-01 -4.08660710e-01 -1.42063475e+00
1.46122752e-02 -8.34587812e-01 -3.96152556e-01 -1.34324503e+00
4.22121495e-01 -4.01693285e-01 -3.74649376e-01 5.14135122e-01
3.74206960e-01 3.01487744e-01 2.64497638e-01 3.27894002e-01
-5.15844107e-01 8.88650537e-01 9.53748703e-01 -1.17475107e-01
3.10796835e-02 -6.26959741e-01 -1.26540512e-01 5.73478520e-01
4.87584740e-01 -3.46415997e-01 -5.32758415e-01 -1.01479053e+00
-1.36398718e-01 4.53922063e-01 9.92074311e-02 -1.45285165e+00
6.75901353e-01 1.40646011e-01 2.83873558e-01 -7.94823110e-01
6.31504536e-01 -8.35173368e-01 3.51304919e-01 1.05603111e+00
-1.33455517e-02 5.31372070e-01 4.48584020e-01 2.73428410e-01
-1.19409479e-01 7.40444288e-02 5.62211215e-01 1.77313492e-01
-7.31446803e-01 5.73373616e-01 -5.59226155e-01 -6.82265580e-01
9.57406104e-01 -5.30105948e-01 -5.03308952e-01 3.15591693e-02
-1.09117828e-01 -2.98295938e-03 1.79763421e-01 2.63950914e-01
8.67020011e-01 -1.41500676e+00 -3.66000772e-01 2.38598451e-01
-2.46504754e-01 -3.97202343e-01 8.55566204e-01 1.16540456e+00
-6.85153186e-01 1.08338153e+00 -4.05760467e-01 -7.22503424e-01
-1.38540077e+00 7.50824571e-01 2.46034518e-01 -4.88164812e-01
-7.48878360e-01 6.94114268e-01 2.05735788e-01 2.01848030e-01
1.64672658e-01 -7.14947760e-01 2.44072303e-01 -2.92360336e-01
6.23426616e-01 5.45937181e-01 4.41914707e-01 -5.62793136e-01
-3.85859460e-01 5.70169210e-01 -1.46631747e-01 4.43860471e-01
1.26347375e+00 4.54497337e-02 -2.72345215e-01 2.18250364e-01
1.44764364e+00 -5.52584350e-01 -8.85638356e-01 -2.66085029e-01
-1.14296310e-01 -2.89929897e-01 7.44476557e-01 -7.90743157e-02
-1.59367299e+00 1.18716955e+00 9.57980096e-01 -2.53551692e-01
1.45708096e+00 -7.32034504e-01 1.74075294e+00 6.03188097e-01
8.22989881e-01 -9.83036458e-01 6.59295842e-02 3.92824709e-01
3.74608070e-01 -7.62893558e-01 2.07053944e-01 -1.63759485e-01
-3.10085505e-01 1.34680426e+00 8.36120248e-01 1.29759982e-01
4.92737204e-01 4.34326679e-01 -5.72468638e-01 -6.79549202e-02
-1.21456230e+00 4.06277478e-01 1.49879843e-01 -9.30551961e-02
3.99640799e-01 3.15465964e-02 -2.81034708e-01 4.36500281e-01
-1.15923963e-01 3.43944192e-01 3.89412880e-01 9.36694443e-01
-4.31256920e-01 -8.77974927e-01 -1.00192942e-01 7.90488660e-01
-4.54017907e-01 -3.11746925e-01 4.71837670e-01 4.76246476e-01
-1.30152538e-01 5.69372118e-01 2.65387625e-01 -1.10842395e+00
-5.51115200e-02 -2.08163217e-01 4.33596134e-01 2.63170544e-02
-6.20766819e-01 2.12332264e-01 7.25109801e-02 -6.18845642e-01
-2.03231901e-01 -2.57863939e-01 -1.37094128e+00 -9.00433242e-01
-4.41024065e-01 -8.07961300e-02 7.16705263e-01 1.05052733e+00
6.51133060e-01 6.54178321e-01 3.83836687e-01 -9.91910219e-01
-4.77323055e-01 -9.19457376e-01 -3.58233333e-01 -3.68797749e-01
1.52821183e-01 -5.19213796e-01 2.30119482e-01 -2.60990798e-01] | [8.484189987182617, 2.914473295211792] |
f2f879cf-42e1-400f-b5a5-b2093381ea2b | generalizable-pose-estimation-using-implicit | 2305.17252 | null | https://arxiv.org/abs/2305.17252v1 | https://arxiv.org/pdf/2305.17252v1.pdf | Generalizable Pose Estimation Using Implicit Scene Representations | 6-DoF pose estimation is an essential component of robotic manipulation pipelines. However, it usually suffers from a lack of generalization to new instances and object types. Most widely used methods learn to infer the object pose in a discriminative setup where the model filters useful information to infer the exact pose of the object. While such methods offer accurate poses, the model does not store enough information to generalize to new objects. In this work, we address the generalization capability of pose estimation using models that contain enough information about the object to render it in different poses. We follow the line of work that inverts neural renderers to infer the pose. We propose i-$\sigma$SRN to maximize the information flowing from the input pose to the rendered scene and invert them to infer the pose given an input image. Specifically, we extend Scene Representation Networks (SRNs) by incorporating a separate network for density estimation and introduce a new way of obtaining a weighted scene representation. We investigate several ways of initial pose estimates and losses for the neural renderer. Our final evaluation shows a significant improvement in inference performance and speed compared to existing approaches. | ['Yotto Koga', 'Linh Tran', 'Kamal Rahimi Malekshan', 'Vaibhav Saxena'] | 2023-05-26 | null | null | null | null | ['pose-estimation', 'density-estimation'] | ['computer-vision', 'methodology'] | [ 2.46379882e-01 1.35062933e-01 7.03710094e-02 -6.12787664e-01
-5.61196804e-01 -6.00998700e-01 4.98692900e-01 -3.83450426e-02
-3.55568945e-01 4.67174321e-01 -1.45612717e-01 1.32547155e-01
-2.18360871e-01 -9.02039051e-01 -1.23198795e+00 -4.81547564e-01
1.13744803e-01 1.03678143e+00 4.25278485e-01 -5.37003316e-02
3.26272249e-01 1.01260328e+00 -1.61695862e+00 6.92607760e-02
6.24794424e-01 1.08759296e+00 7.08798051e-01 7.52109170e-01
-4.45162207e-02 6.66832507e-01 -6.92731917e-01 -4.15429249e-02
3.86013836e-01 -1.48251891e-01 -8.55560780e-01 1.68052450e-01
7.67504930e-01 -8.12893808e-01 -6.10294521e-01 9.37115669e-01
3.56629163e-01 3.82798791e-01 8.71702075e-01 -1.06029606e+00
-2.24943221e-01 3.95638376e-01 -5.17119706e-01 -6.50839359e-02
4.52985466e-01 1.55286007e-02 6.08768523e-01 -8.36749256e-01
8.85895729e-01 1.59657383e+00 4.90003675e-01 4.25176054e-01
-1.20581186e+00 -3.90724063e-01 5.04529834e-01 -8.18755757e-03
-1.23672080e+00 -2.77677715e-01 6.80477083e-01 -4.81265187e-01
8.09199870e-01 3.15313563e-02 5.45435190e-01 9.45556223e-01
1.68337807e-01 8.91000032e-01 7.50370085e-01 -1.46280676e-01
3.63937289e-01 -7.75287598e-02 -8.87117088e-02 7.71875739e-01
1.03640102e-01 -1.97945654e-01 -7.15798914e-01 -1.83234792e-02
1.28411019e+00 -2.85864584e-02 -1.12199366e-01 -9.71203327e-01
-1.15517688e+00 4.48723674e-01 7.69343793e-01 -2.45806694e-01
-3.75538707e-01 6.19503438e-01 2.63254344e-01 -1.48952827e-02
4.37084019e-01 6.11851215e-01 -3.92004132e-01 -1.70775950e-01
-7.43494451e-01 5.06885588e-01 9.52674508e-01 1.18889391e+00
8.67323279e-01 -9.77425724e-02 -2.17885777e-01 5.56528270e-01
2.52033204e-01 3.42708319e-01 -1.76270857e-01 -1.28924274e+00
4.71007019e-01 3.70345861e-01 1.87656134e-01 -1.11846685e+00
-2.84344882e-01 -4.88258541e-01 -3.76730859e-01 3.54116201e-01
3.35821241e-01 1.03973925e-01 -1.07948804e+00 1.58991599e+00
5.19959331e-01 2.25375757e-01 -3.44187349e-01 9.48466241e-01
5.07620454e-01 5.63036442e-01 -7.94313177e-02 4.19644147e-01
8.78110647e-01 -9.00290787e-01 -4.66400027e-01 -4.55302656e-01
1.19985633e-01 -6.22062027e-01 8.18261981e-01 6.73897147e-01
-1.11029494e+00 -6.28070891e-01 -1.09327674e+00 -4.99970496e-01
-2.61007220e-01 3.51395249e-01 8.91187191e-01 2.56877989e-01
-9.93253112e-01 9.05868769e-01 -1.22517073e+00 -3.25862765e-01
4.80414689e-01 5.38124979e-01 -3.29074234e-01 -3.25571686e-01
-4.88180757e-01 1.14161837e+00 6.84444666e-01 3.22122276e-01
-1.32145894e+00 -6.04972601e-01 -1.10121918e+00 -1.31103396e-01
5.24417400e-01 -8.82988930e-01 1.30266392e+00 -4.65635926e-01
-1.55200851e+00 5.17485321e-01 -2.17943236e-01 -2.03958318e-01
5.82902730e-01 -6.70182049e-01 4.75654393e-01 1.37170151e-01
-2.33184323e-02 9.21643972e-01 9.88311410e-01 -1.66246665e+00
-4.23092753e-01 -5.22831559e-01 5.16362131e-01 3.18728685e-01
5.76030575e-02 -4.76676613e-01 -5.91973960e-01 -4.54118401e-01
6.90938592e-01 -8.30077887e-01 -2.84329891e-01 6.47683918e-01
-4.21950430e-01 -1.32473454e-01 1.04355013e+00 -7.49850333e-01
5.70192575e-01 -1.94673848e+00 6.01918340e-01 1.25587195e-01
1.77989528e-01 -6.31670654e-02 -1.58780009e-01 3.47049892e-01
2.97343284e-01 -3.89591277e-01 -1.04075074e-01 -7.30661690e-01
1.40241101e-01 6.18316412e-01 -4.79073167e-01 5.96178710e-01
4.38091666e-01 7.97274947e-01 -1.00735283e+00 -2.75974065e-01
5.25819123e-01 8.22189867e-01 -9.60726559e-01 3.62057537e-01
-6.46736920e-01 6.95427716e-01 -5.89414656e-01 3.79523963e-01
8.61216784e-01 -2.16916084e-01 7.31258746e-03 -4.87855673e-01
6.91061318e-02 4.34391856e-01 -1.35302699e+00 2.44987345e+00
-4.98901010e-01 4.20057267e-01 1.95493817e-01 -9.67458010e-01
1.02542293e+00 -1.63224101e-01 4.18417603e-01 1.42658958e-02
2.56005973e-01 -3.48299965e-02 -3.35352480e-01 -4.67302650e-01
7.63456762e-01 -1.79708637e-02 -6.31045103e-02 1.18306130e-01
3.74078751e-01 -8.16909909e-01 9.94027704e-02 1.21633753e-01
8.87995541e-01 9.76668596e-01 3.85113508e-02 -4.51220572e-02
1.04588941e-01 -7.21231243e-03 2.58821249e-01 7.00078607e-01
2.65304595e-01 7.46209800e-01 3.60502601e-01 -3.13333124e-01
-9.75479782e-01 -1.39630687e+00 -1.67657450e-01 9.81228590e-01
4.62968379e-01 -3.06726158e-01 -5.55794239e-01 -4.04648572e-01
2.02696621e-01 7.58934140e-01 -4.32307750e-01 -2.28256777e-01
-6.75134778e-01 -2.53133535e-01 1.88871056e-01 7.63640165e-01
4.22671646e-01 -7.05143154e-01 -8.47579718e-01 3.59890014e-02
-2.02097982e-01 -1.14306927e+00 -1.26612410e-01 2.13611096e-01
-1.01830220e+00 -8.82372916e-01 -4.77074057e-01 -5.52924335e-01
1.03138828e+00 4.25778031e-02 9.74015892e-01 -3.14494878e-01
-6.00315034e-01 4.89252895e-01 -8.63761306e-02 -4.97139722e-01
-2.67652810e-01 1.61753744e-01 5.97193837e-02 -3.40800107e-01
-5.76817393e-02 -6.25009358e-01 -5.36706865e-01 5.76981008e-02
-8.00926089e-01 6.02557696e-02 6.09307826e-01 5.08725107e-01
7.10110068e-01 -2.92721204e-02 9.99078751e-02 -6.05984211e-01
2.54754096e-01 -2.54834265e-01 -6.33318424e-01 5.92529848e-02
1.31961899e-02 4.72734839e-01 3.33225816e-01 -4.51205850e-01
-1.20410740e+00 5.20376980e-01 -1.98589653e-01 -6.92991376e-01
-5.98619059e-02 3.05559725e-01 -1.36536047e-01 -1.19810872e-01
6.11455977e-01 -1.14495285e-01 1.39153287e-01 -7.98820555e-01
5.29133201e-01 2.08804816e-01 6.77041292e-01 -1.08959436e+00
8.49313080e-01 6.14764988e-01 2.18791679e-01 -5.36660850e-01
-9.42912519e-01 -4.30391967e-01 -9.70025182e-01 -3.12551796e-01
7.21195936e-01 -8.34078312e-01 -9.88092542e-01 3.50714445e-01
-1.42480254e+00 -2.92064011e-01 -4.59649980e-01 5.40863514e-01
-8.26347709e-01 2.83204377e-01 -5.07153273e-01 -7.90022612e-01
1.41548157e-01 -1.42383933e+00 1.72632670e+00 3.45841469e-03
-8.05518180e-02 -6.46682858e-01 -1.72173589e-01 1.13977212e-02
2.65466988e-01 4.40659136e-01 6.68891251e-01 -9.03474391e-02
-1.05094922e+00 -2.26102144e-01 -3.36947262e-01 2.71995902e-01
1.04565553e-01 -1.60771772e-01 -1.12622547e+00 -3.79194975e-01
-3.82654667e-02 -3.81660789e-01 8.47754776e-01 3.78191918e-01
1.72531533e+00 -7.99281895e-02 -4.38467801e-01 7.49820590e-01
1.40627968e+00 -1.52176693e-01 6.33429587e-01 1.44098297e-01
8.25132549e-01 5.85648298e-01 6.01887286e-01 4.82359499e-01
2.76449680e-01 6.68623388e-01 9.54129398e-01 3.09796661e-01
-8.86422023e-02 -5.96592128e-01 2.07918733e-01 6.52696371e-01
-1.70459747e-01 -1.51112124e-01 -6.93070531e-01 2.39600822e-01
-1.76295257e+00 -4.20982957e-01 2.68005997e-01 2.24825263e+00
6.40918136e-01 9.70553011e-02 -2.61719167e-01 -2.02891901e-01
3.24984640e-01 8.64890590e-02 -8.80473554e-01 -1.04162470e-01
2.74346262e-01 2.85778224e-01 5.14567077e-01 6.45061433e-01
-9.51240122e-01 9.24389064e-01 6.63034344e+00 7.08193421e-01
-9.27234113e-01 -2.21771300e-01 2.25376830e-01 -1.13586627e-01
-2.45581046e-01 1.21200815e-01 -8.79755139e-01 -9.06735659e-02
4.07743812e-01 1.30373135e-01 5.33206284e-01 1.11532688e+00
-1.41750306e-01 -4.14627969e-01 -1.58686042e+00 9.64988470e-01
2.72094071e-01 -1.10030305e+00 2.94177383e-01 -1.18843026e-01
3.67374688e-01 -1.45309195e-02 -1.18407957e-01 2.82810956e-01
3.17425251e-01 -9.67988014e-01 1.05076349e+00 8.10895562e-01
6.71643198e-01 -5.42002678e-01 3.40267450e-01 6.52889073e-01
-1.07081234e+00 5.34645393e-02 -6.25149488e-01 -2.58529782e-01
3.19066763e-01 4.68277574e-01 -9.89940822e-01 6.15751147e-01
6.63576603e-01 6.81990802e-01 -4.58072484e-01 1.08278835e+00
-3.11305672e-01 -4.34954204e-02 -6.61472380e-01 1.34035289e-01
-4.63888533e-02 -1.52578324e-01 6.54152334e-01 6.97097719e-01
2.87901551e-01 -2.41083041e-01 3.74482483e-01 1.24220681e+00
4.70635761e-03 -3.94777000e-01 -6.38411939e-01 1.92038283e-01
4.28528100e-01 1.01310217e+00 -8.25621068e-01 -2.80989468e-01
7.32806772e-02 1.23640954e+00 6.92536950e-01 3.20440799e-01
-7.60562956e-01 -1.38007060e-01 5.61040461e-01 1.45327196e-01
2.64028847e-01 -7.14904130e-01 -1.50103152e-01 -1.02055645e+00
2.98104346e-01 -4.54015762e-01 -7.81845823e-02 -1.04646981e+00
-1.12723517e+00 3.01160365e-01 5.54945111e-01 -9.75727439e-01
-3.50483924e-01 -9.92818475e-01 5.02836667e-02 9.10659015e-01
-1.33507526e+00 -1.05525506e+00 -5.44725358e-01 1.67282745e-01
6.40286505e-01 3.81515950e-01 7.00836837e-01 1.19708620e-01
-1.15370303e-01 2.23787859e-01 -2.75855124e-01 -6.52096700e-03
4.84913677e-01 -1.21107829e+00 3.24015319e-01 5.41363478e-01
-6.78077266e-02 8.31739306e-01 7.76845217e-01 -7.00831711e-01
-1.73828685e+00 -9.16561604e-01 2.17568442e-01 -7.32352793e-01
2.95844436e-01 -6.21612668e-01 -6.38493478e-01 1.06392920e+00
-3.04746449e-01 6.50651827e-02 -1.81833804e-01 6.51535690e-02
-2.02569500e-01 -1.23499252e-03 -1.16325903e+00 5.62773764e-01
1.26495945e+00 -5.30346870e-01 -5.21154463e-01 3.55136603e-01
8.59581172e-01 -1.10185969e+00 -9.04893160e-01 5.17661095e-01
6.63831770e-01 -7.75450408e-01 1.30404675e+00 -4.30278897e-01
5.40318131e-01 -4.34195727e-01 -2.26406366e-01 -1.16958106e+00
-8.14352632e-02 -3.34796131e-01 -3.39548558e-01 8.10143113e-01
-7.73595506e-03 -5.23818612e-01 8.86876583e-01 7.69079685e-01
-3.37644815e-01 -7.78732300e-01 -8.67082953e-01 -6.58678710e-01
-6.37099817e-02 -5.64756691e-01 5.83465695e-01 4.13700819e-01
-3.71526688e-01 1.34978235e-01 -2.20217690e-01 4.35680062e-01
6.54984772e-01 7.60731101e-02 1.09792447e+00 -1.14582872e+00
-3.88441116e-01 -8.51054415e-02 -6.92545712e-01 -1.76743996e+00
2.20715493e-01 -7.18157530e-01 4.67918605e-01 -1.83845258e+00
2.07232028e-01 -5.37625790e-01 1.80128768e-01 2.87693620e-01
7.96590820e-02 1.71922259e-02 2.11002663e-01 1.68468114e-02
-4.52400029e-01 5.63790679e-01 1.58222091e+00 -1.34420097e-01
-7.92035535e-02 8.04142952e-02 -2.31801972e-01 7.95389891e-01
5.80773413e-01 -2.99531817e-01 -5.71574390e-01 -8.82752180e-01
4.00785625e-01 4.04113438e-03 7.93719232e-01 -1.08636832e+00
2.78708071e-01 -7.68363327e-02 5.77291012e-01 -9.68707025e-01
9.33492541e-01 -9.11570072e-01 2.33844951e-01 2.91321129e-01
-3.50821137e-01 -1.54845282e-01 2.40419745e-01 7.96293616e-01
7.32000619e-02 -4.25749004e-01 4.78949577e-01 -4.23119932e-01
-8.06492209e-01 4.86829758e-01 2.82072667e-02 -3.38135839e-01
8.10789287e-01 -2.92761087e-01 2.14804988e-03 -3.14412683e-01
-7.82470167e-01 1.23851709e-01 6.41505539e-01 4.83023494e-01
6.83845520e-01 -1.16669178e+00 -4.29880977e-01 9.42801610e-02
3.31517532e-02 9.10079062e-01 2.16663498e-02 4.50455368e-01
-8.79811823e-01 1.05287187e-01 -1.39968574e-01 -9.66766596e-01
-7.86663055e-01 4.47636962e-01 3.63525629e-01 9.14090797e-02
-7.18787611e-01 1.06139767e+00 3.48488629e-01 -7.49437571e-01
5.14882326e-01 -6.95454001e-01 2.07283288e-01 -2.41904989e-01
2.74087012e-01 4.46662396e-01 -2.42735799e-02 -5.30337155e-01
-1.35358036e-01 5.85416853e-01 3.89523245e-02 -5.77259138e-02
1.37668204e+00 -5.08999452e-02 -3.55160952e-01 6.86321974e-01
1.22055209e+00 -3.60047489e-01 -1.76952040e+00 -1.61867812e-01
-2.80108780e-01 -8.09447587e-01 -3.58079523e-02 -5.34584761e-01
-9.97010529e-01 9.88270164e-01 2.92920172e-01 -3.68336886e-01
6.82382643e-01 2.05865085e-01 4.28650111e-01 6.73795342e-01
8.27676892e-01 -1.04429173e+00 2.30629832e-01 6.90163732e-01
1.23920763e+00 -1.03264618e+00 2.88276285e-01 -7.34886825e-01
-1.33573294e-01 1.25397205e+00 7.99672127e-01 -4.47189391e-01
6.83808863e-01 4.75110948e-01 -3.67258966e-01 -4.01855171e-01
-1.78259730e-01 2.89789662e-02 4.15868878e-01 7.12939620e-01
1.58135086e-01 -1.14882037e-01 3.33623976e-01 1.28511339e-02
-3.99319947e-01 -5.35821496e-03 2.11088747e-01 1.19649851e+00
-4.51807320e-01 -8.60581398e-01 -3.02848339e-01 4.65760052e-01
6.04056893e-03 2.31875688e-01 -2.91081756e-01 6.63802624e-01
-9.92182121e-02 3.69694531e-01 2.04805270e-01 -9.42879394e-02
4.05562967e-01 -9.44339335e-02 1.06420147e+00 -8.35324287e-01
-9.39060450e-02 -1.27357960e-01 -7.41049647e-02 -8.75869215e-01
-4.66171503e-01 -6.20383441e-01 -1.20562243e+00 -2.09955834e-02
-3.97036582e-01 -3.53445083e-01 9.25697446e-01 8.83411348e-01
2.31044874e-01 8.82628858e-01 1.63291872e-01 -1.75831306e+00
-6.43059731e-01 -9.03556406e-01 -4.61827457e-01 3.12742054e-01
3.55113775e-01 -1.17832303e+00 -1.08536437e-01 1.22131914e-01] | [6.279313087463379, -1.2832573652267456] |
88d87f3d-bd20-4e66-ad98-424e0e20fc35 | danes-deep-neural-network-ensemble | 2302.01756 | null | https://arxiv.org/abs/2302.01756v1 | https://arxiv.org/pdf/2302.01756v1.pdf | DANES: Deep Neural Network Ensemble Architecture for Social and Textual Context-aware Fake News Detection | The growing popularity of social media platforms has simplified the creation and distribution of news articles but also creates a conduit for spreading fake news. In consequence, the need arises for effective context-aware fake news detection mechanisms, where the contextual information can be built either from the textual content of posts or from available social data (e.g., information about the users, reactions to posts, or the social network). In this paper, we propose DANES, a Deep Neural Network Ensemble Architecture for Social and Textual Context-aware Fake News Detection. DANES comprises a Text Branch for a textual content-based context and a Social Branch for the social context. These two branches are used to create a novel Network Embedding. Preliminary ablation results on 3 real-world datasets, i.e., BuzzFace, Twitter15, and Twitter16, are promising, with an accuracy that outperforms state-of-the-art solutions when employing both social and textual content features. | ['Panagiotis Karras', 'Elena-Simona Apostol', 'Ciprian-Octavian Truică'] | 2023-02-01 | null | null | null | null | ['network-embedding'] | ['methodology'] | [-1.32663280e-01 -1.19171605e-01 -4.24704969e-01 -3.72162044e-01
-2.11308509e-01 -4.77369964e-01 1.17180204e+00 5.20379782e-01
-4.22939926e-01 6.41295910e-01 6.58307433e-01 -2.57957369e-01
3.08159709e-01 -1.02428031e+00 -4.07836825e-01 -2.21731439e-01
1.82050347e-01 7.32480735e-02 2.84877747e-01 -8.29183698e-01
4.16744202e-01 4.91660945e-02 -1.25403941e+00 6.95843339e-01
6.39367819e-01 9.54682946e-01 -4.35188979e-01 2.34231338e-01
-3.29832077e-01 1.17389071e+00 -9.67446625e-01 -5.72229743e-01
-1.16008431e-01 -6.35890841e-01 -4.49987590e-01 -2.56464899e-01
3.15103203e-01 -4.34098482e-01 -8.10500085e-01 9.48578775e-01
2.56774455e-01 -1.94390342e-01 3.42149049e-01 -1.20960939e+00
-7.46256828e-01 9.44484413e-01 -3.47996742e-01 2.41034269e-01
2.64926761e-01 -1.73331812e-01 1.05519760e+00 -8.89966965e-01
9.56809342e-01 1.06896448e+00 7.80828655e-01 3.94982040e-01
-7.72816241e-01 -5.87892115e-01 1.30448908e-01 2.22740486e-01
-1.01300764e+00 -2.70626038e-01 1.04625070e+00 -5.58460176e-01
6.68258369e-01 3.85714978e-01 1.07554090e+00 2.01185012e+00
3.77327353e-01 7.78312325e-01 1.18694270e+00 -2.17645437e-01
1.07562304e-01 3.12121660e-01 1.02668971e-01 4.14279640e-01
3.68273795e-01 1.40069202e-01 -6.83330119e-01 -6.89386427e-01
1.99344084e-01 2.65054703e-01 -5.29336594e-02 1.15987338e-01
-1.17114425e+00 1.26384997e+00 7.59907305e-01 5.90299189e-01
-2.98354477e-01 1.36796743e-01 7.62034655e-01 5.25821269e-01
1.18807983e+00 7.53167093e-01 -1.16480343e-01 2.02073567e-02
-1.04976201e+00 4.95296836e-01 1.17896426e+00 3.40874642e-01
5.76437533e-01 9.64187756e-02 -1.39471471e-01 7.80982137e-01
2.96114504e-01 4.54771876e-01 6.29382789e-01 -2.45700553e-01
5.15653014e-01 7.36719728e-01 2.17456371e-01 -2.02346325e+00
-3.90412003e-01 -5.47328949e-01 -6.27811849e-01 -5.02220452e-01
2.88361967e-01 -3.66759866e-01 -6.16813719e-01 1.47165394e+00
4.24058229e-01 4.75974381e-01 -2.70731866e-01 8.51751804e-01
1.01627457e+00 7.57846534e-01 -2.65658557e-01 2.11451165e-02
1.21036839e+00 -1.04040492e+00 -9.73785698e-01 -5.34535587e-01
6.50021374e-01 -7.95351624e-01 7.05861926e-01 5.90016395e-02
-5.70305765e-01 1.87046409e-01 -1.23430395e+00 -1.28957406e-01
-1.01460540e+00 -6.04934990e-02 5.14427364e-01 3.97154957e-01
-6.41359508e-01 7.33719349e-01 -4.76069540e-01 -1.38253346e-01
4.58992898e-01 -2.44792655e-01 -2.02996954e-01 3.41662318e-02
-1.78983843e+00 8.46692443e-01 1.99547350e-01 1.56486288e-01
-6.57237291e-01 -6.91707060e-02 -6.26391709e-01 -7.36258477e-02
6.02461934e-01 -2.95214623e-01 9.70337808e-01 -1.22974098e+00
-1.36066747e+00 7.26936817e-01 2.39579856e-01 -5.34919262e-01
8.21735680e-01 -1.73245281e-01 -8.03191304e-01 2.05168411e-01
2.07777172e-01 9.58930515e-03 1.27979410e+00 -1.08905470e+00
-3.70275378e-01 -1.34634152e-01 7.93637931e-02 -2.72802621e-01
-7.16567159e-01 5.66704385e-02 1.07824979e-02 -1.04011595e+00
-1.15486659e-01 -1.06716561e+00 7.62232840e-02 -2.38750741e-01
-7.62362897e-01 -8.50646794e-02 1.44928658e+00 -8.06502521e-01
1.40496385e+00 -2.07564974e+00 -4.99200188e-02 2.21744761e-01
6.28167212e-01 3.99246454e-01 -1.00830846e-01 8.35998237e-01
1.07595451e-01 3.41100961e-01 4.64966483e-02 -5.92171013e-01
-9.50128883e-02 3.86868715e-02 -5.13037741e-01 6.42113388e-01
1.97039038e-01 8.87520194e-01 -1.27284169e+00 -2.25094706e-01
-1.45179674e-01 5.28262675e-01 -5.67244112e-01 -2.41164133e-01
-5.27111590e-01 3.44110698e-01 -6.01330221e-01 4.68573481e-01
2.76720017e-01 -5.14322460e-01 1.29450366e-01 1.94678120e-02
-2.56471913e-02 7.99312294e-01 -4.11826968e-01 9.40838218e-01
-4.08335537e-01 1.13345098e+00 9.41238739e-03 -9.37373042e-01
8.32132101e-01 3.25874150e-01 3.98017913e-01 -5.02018988e-01
6.76640928e-01 2.43150875e-01 -1.49002060e-01 -2.85203069e-01
8.64680946e-01 3.72721329e-02 -3.81515652e-01 8.67307365e-01
-1.30165830e-01 1.17412031e-01 -5.84951825e-02 5.22816002e-01
1.00004613e+00 -3.70382786e-01 1.98231220e-01 -9.14989319e-03
2.14767069e-01 8.82728994e-02 1.99145675e-01 8.20644617e-01
-2.84622014e-01 3.78727138e-01 8.27900052e-01 -5.71662307e-01
-9.70661879e-01 -4.29587960e-01 1.53620392e-01 1.11315024e+00
3.26508254e-01 -7.31932044e-01 -5.77825367e-01 -1.06123734e+00
1.73125073e-01 6.47436917e-01 -1.00515687e+00 -4.53179121e-01
-5.49695015e-01 -7.63993144e-01 5.03213167e-01 -1.46983787e-01
5.28449714e-01 -9.38677907e-01 -8.88134539e-02 4.66990024e-01
-6.06995523e-01 -1.29867399e+00 -4.71435189e-01 -3.35341752e-01
-5.48807025e-01 -9.77405190e-01 -2.08350658e-01 -3.41468543e-01
2.01017901e-01 4.02618587e-01 9.84745741e-01 3.60097110e-01
8.60344991e-02 -1.22620715e-02 -6.32429898e-01 -3.85882825e-01
-6.44496799e-01 2.25315347e-01 -6.36042804e-02 3.98245275e-01
3.41705501e-01 -5.46308398e-01 -5.51353514e-01 3.95739496e-01
-9.98427510e-01 2.35048942e-02 2.18114957e-01 1.06455660e+00
-6.77614100e-03 -1.53478265e-01 9.97450650e-01 -1.15567911e+00
1.00253034e+00 -1.11149502e+00 -2.65208155e-01 -1.86193779e-01
-6.08399034e-01 -4.89756703e-01 6.89031243e-01 -7.18182683e-01
-6.52848005e-01 -6.57893181e-01 2.92463582e-02 -3.70756499e-02
3.29821110e-01 1.06829417e+00 5.15149236e-01 7.93309510e-02
1.04121780e+00 1.31375149e-01 2.88443118e-02 -3.10839802e-01
3.45629901e-01 1.15304351e+00 1.12423986e-01 -6.14087954e-02
6.57396555e-01 7.41432548e-01 -4.04870033e-01 -1.04779387e+00
-1.19941795e+00 -4.52665001e-01 1.22422827e-02 -1.63425088e-01
3.66229504e-01 -7.87434995e-01 -1.57817796e-01 6.78760648e-01
-1.26959884e+00 -5.88742830e-02 -8.00248515e-03 2.25057021e-01
5.96585162e-02 2.74041563e-01 -9.41703737e-01 -4.90686536e-01
-2.67006308e-01 -7.40615129e-01 6.64101243e-01 -2.94672698e-01
-3.16186607e-01 -1.10706818e+00 -7.05225021e-02 5.45086324e-01
8.04138005e-01 7.82971561e-01 3.57448608e-01 -1.19444168e+00
-3.42118770e-01 -9.22897100e-01 -4.81095970e-01 3.83964300e-01
2.19536304e-01 7.64191002e-02 -9.10865963e-01 -2.35832155e-01
4.34421189e-02 -2.74997860e-01 9.79632556e-01 -1.05189778e-01
8.12319219e-01 -9.84304309e-01 -5.12411952e-01 8.19612592e-02
9.85996127e-01 -3.28448415e-01 4.39939708e-01 3.91499311e-01
8.62749398e-01 5.04264116e-01 2.53858149e-01 6.79118574e-01
3.04746419e-01 5.33204794e-01 5.74601233e-01 1.56287000e-01
2.54041553e-02 -5.36801338e-01 4.05426949e-01 1.04899848e+00
3.05995554e-01 -6.87117517e-01 -8.62537861e-01 5.58631241e-01
-1.83816218e+00 -1.23850977e+00 -1.65299743e-01 1.68200195e+00
7.66427279e-01 2.81403422e-01 1.53162315e-01 3.41555886e-02
7.90766716e-01 7.92593002e-01 -2.29607776e-01 -2.44371936e-01
-2.86308646e-01 -3.40109289e-01 3.87266070e-01 5.04449427e-01
-1.31157005e+00 9.45219100e-01 5.76628399e+00 7.55776107e-01
-1.64572072e+00 4.83571351e-01 4.04276401e-01 -2.00396433e-01
-3.17402631e-01 -1.50708944e-01 -5.18406689e-01 9.33898628e-01
8.96148980e-01 -8.32103193e-02 4.49477881e-01 8.10046911e-01
2.49766946e-01 7.63934702e-02 -7.71298587e-01 7.08207309e-01
3.32634091e-01 -1.87939060e+00 -9.92665887e-02 1.74448490e-01
9.09094036e-01 6.34985209e-01 6.98290169e-02 2.77260661e-01
1.16916969e-01 -6.18409097e-01 8.68907213e-01 1.90068543e-01
5.50981522e-01 -4.22620982e-01 1.00085664e+00 5.00327885e-01
-4.17069823e-01 -1.73200550e-03 1.27177626e-01 -2.60103315e-01
1.15417622e-01 7.75702298e-01 -9.71281052e-01 2.81968087e-01
4.07685637e-01 1.27390420e+00 -5.20027459e-01 6.59848571e-01
-4.46122050e-01 9.18783069e-01 -3.11984390e-01 -6.43209159e-01
7.03828871e-01 2.41061766e-02 8.89608085e-01 1.38337362e+00
-5.68657815e-02 -1.45804644e-01 6.20442927e-02 7.92929351e-01
-4.50752437e-01 3.96185368e-02 -7.70620346e-01 -7.10669816e-01
4.33731109e-01 1.07303858e+00 -5.97027481e-01 -4.84558105e-01
-4.28707600e-01 8.68115067e-01 2.94607580e-01 3.35818797e-01
-7.93849945e-01 -4.27881449e-01 5.50425589e-01 3.90122086e-01
7.18936473e-02 -2.27621391e-01 -2.51749940e-02 -1.59352827e+00
-8.93761218e-02 -9.44354713e-01 1.50400713e-01 -5.08671463e-01
-1.69128668e+00 6.70254946e-01 -1.40929878e-01 -1.04746914e+00
-2.08259732e-01 -4.10623670e-01 -6.38502598e-01 4.24225271e-01
-1.38711452e+00 -1.06341946e+00 -3.51691931e-01 4.27686185e-01
7.97126964e-02 -2.03595906e-01 5.75834274e-01 3.77889574e-01
-5.03329933e-01 5.81637442e-01 4.52266723e-01 6.04877949e-01
7.14876771e-01 -7.77327478e-01 5.70750296e-01 5.91220498e-01
1.36286646e-01 5.53881407e-01 8.05364430e-01 -8.13780963e-01
-1.08784556e+00 -1.03461707e+00 1.17580104e+00 -5.25007010e-01
1.30491149e+00 -5.87207258e-01 -7.78414071e-01 6.57260478e-01
7.89678171e-02 3.84126157e-02 7.90363193e-01 4.45309356e-02
-7.52234519e-01 2.09913358e-01 -1.35816777e+00 8.79658937e-01
8.21641445e-01 -8.95551383e-01 -5.82673788e-01 6.71655416e-01
7.65391290e-01 -3.80465716e-01 -4.08798367e-01 -1.00340262e-01
6.52136803e-01 -9.05991375e-01 6.92074120e-01 -6.67644560e-01
9.58141088e-01 6.85020611e-02 4.35864069e-02 -1.63512635e+00
1.72757600e-02 -6.34900093e-01 -3.76091421e-01 7.78869152e-01
4.72385049e-01 -1.05663657e+00 6.58105195e-01 6.27082288e-02
6.73154071e-02 -6.77756786e-01 -8.86855483e-01 -5.49047410e-01
-1.34117156e-01 -3.28025043e-01 4.45948243e-01 1.74798095e+00
2.59661466e-01 6.26704693e-01 -7.51175284e-01 -1.57856449e-01
2.40822643e-01 1.65141165e-01 5.75523078e-01 -1.21197307e+00
-1.30668819e-01 -6.04070485e-01 -4.14614141e-01 -9.28235054e-01
1.96884960e-01 -8.00245821e-01 -3.88887495e-01 -1.30636871e+00
-1.65228590e-01 -3.65325391e-01 -3.82742174e-02 3.13157797e-01
1.03235215e-01 4.25643027e-01 1.07165046e-01 3.35142583e-01
-5.46257973e-01 6.20221913e-01 1.36471999e+00 -3.96099150e-01
-1.30954385e-01 -1.21532954e-01 -6.39411449e-01 7.81133354e-01
9.34052050e-01 -7.75148332e-01 -6.39577433e-02 -2.57275760e-01
8.42377782e-01 -6.15852922e-02 5.28355360e-01 -5.10099530e-01
1.03404298e-01 -3.00240964e-02 -6.79380819e-02 -4.34459209e-01
5.07869482e-01 -5.88877857e-01 -4.18504417e-01 3.51201594e-01
-4.22171444e-01 -1.79782823e-01 -2.24755593e-02 1.07253420e+00
-4.98611361e-01 1.95518322e-02 6.25793040e-01 -1.17672063e-01
-1.79710537e-01 2.11162239e-01 -5.70028126e-01 2.18617946e-01
5.51016450e-01 -5.50949872e-02 -9.26962614e-01 -9.36360419e-01
-7.56653607e-01 -4.10059467e-03 3.29147756e-01 7.60915995e-01
6.24935389e-01 -1.44637096e+00 -9.35725868e-01 5.03328666e-02
2.63520807e-01 -5.10997474e-01 -4.48559336e-02 8.22594047e-01
-4.16622758e-01 1.29249036e-01 1.44281074e-01 -2.61415899e-01
-9.71232653e-01 4.22667533e-01 2.33325914e-01 -1.46110609e-01
-4.18978512e-01 6.99647546e-01 -3.94929767e-01 -3.96735042e-01
-1.37977421e-01 -2.83258289e-01 -2.02623114e-01 4.86543655e-01
7.01317728e-01 3.49878788e-01 1.19718745e-01 -9.13121819e-01
-2.03975424e-01 -2.05236465e-01 -1.71260551e-01 -6.52849628e-03
1.40632725e+00 1.88468024e-02 -4.17087466e-01 5.90086162e-01
1.43463528e+00 3.03293347e-01 -6.57678962e-01 -6.03840172e-01
1.67060092e-01 -7.41830170e-01 3.81428927e-01 -9.25161839e-01
-9.95732009e-01 4.83118951e-01 -1.22489735e-01 1.06064367e+00
4.92997050e-01 -6.82435334e-02 1.12181175e+00 4.54965115e-01
2.74449915e-01 -1.18594456e+00 4.62438792e-01 7.58703411e-01
1.13818753e+00 -1.42530406e+00 5.12126572e-02 -4.11696702e-01
-4.33552563e-01 8.67033124e-01 2.59091944e-01 -3.60487700e-01
9.69800651e-01 -1.57790154e-01 6.65159598e-02 -4.93935138e-01
-5.91992497e-01 3.35984439e-01 2.48540267e-01 6.90528611e-03
3.74675423e-01 1.14310205e-01 -5.73310673e-01 6.13728762e-01
-2.53957421e-01 -2.07175016e-01 8.91710758e-01 8.65887046e-01
-3.40731412e-01 -7.67172515e-01 -1.99109912e-01 6.67366982e-01
-7.40854084e-01 -1.34652376e-01 -9.76246715e-01 7.09172845e-01
1.35262579e-01 1.11386371e+00 -6.74485639e-02 -7.03122616e-01
-4.74909954e-02 -8.16912130e-02 -1.96368456e-01 -6.59032643e-01
-9.05848801e-01 -2.54641920e-01 6.36483610e-01 -5.99389255e-01
-2.61286318e-01 -4.70194519e-01 -5.67726851e-01 -7.26933002e-01
-6.97872281e-01 1.21831000e-01 9.11573410e-01 1.06308734e+00
5.30410409e-01 1.61485583e-01 8.42913687e-01 -7.95942783e-01
-4.52426612e-01 -1.22081137e+00 -2.64603972e-01 5.50050378e-01
7.79253781e-01 -6.41407430e-01 -7.92866766e-01 -4.89726245e-01] | [8.179155349731445, 10.252546310424805] |
25529604-71af-48b3-9145-7e2df69ecc82 | robust-3d-object-detection-in-cold-weather | 2205.11925 | null | https://arxiv.org/abs/2205.11925v2 | https://arxiv.org/pdf/2205.11925v2.pdf | Robust 3D Object Detection in Cold Weather Conditions | Adverse weather conditions can negatively affect LiDAR-based object detectors. In this work, we focus on the phenomenon of vehicle gas exhaust condensation in cold weather conditions. This everyday effect can influence the estimation of object sizes, orientations and introduce ghost object detections, compromising the reliability of the state of the art object detectors. We propose to solve this problem by using data augmentation and a novel training loss term. To effectively train deep neural networks, a large set of labeled data is needed. In case of adverse weather conditions, this process can be extremely laborious and expensive. We address this issue in two steps: First, we present a gas exhaust data generation method based on 3D surface reconstruction and sampling which allows us to generate large sets of gas exhaust clouds from a small pool of labeled data. Second, we introduce a point cloud augmentation process that can be used to add gas exhaust to datasets recorded in good weather conditions. Finally, we formulate a new training loss term that leverages the augmented point cloud to increase object detection robustness by penalizing predictions that include noise. In contrast to other works, our method can be used with both grid-based and point-based detectors. Moreover, since our approach does not require any network architecture changes, inference times remain unchanged. Experimental results on real data show that our proposed method greatly increases robustness to gas exhaust and noisy data. | ['Klaus Dietmayer', 'Johannes Kopp', 'Daniel Meissner', 'Marc Walessa', 'Vinzenz Dallabetta', 'Aldi Piroli'] | 2022-05-24 | null | null | null | null | ['robust-3d-object-detection'] | ['computer-vision'] | [ 1.23072430e-01 -1.40753806e-01 1.80493772e-01 -3.48636389e-01
-5.23696601e-01 -6.77196503e-01 6.92728698e-01 8.51893798e-02
-4.54322755e-01 5.68898082e-01 -5.30466616e-01 -4.20709968e-01
3.40432823e-01 -1.20404232e+00 -9.84296560e-01 -6.97467029e-01
3.57187033e-01 3.96052927e-01 6.88349009e-01 -1.04397707e-01
1.68988615e-01 9.43935871e-01 -1.64217508e+00 -2.07422540e-01
9.42183793e-01 1.02627110e+00 3.83826345e-01 3.87264937e-01
-8.35795030e-02 2.29248703e-01 -6.78518534e-01 -1.98847517e-01
7.42622375e-01 2.28186443e-01 -6.15229532e-02 1.05171017e-02
8.04301679e-01 -4.69163716e-01 -1.88240722e-01 9.49847698e-01
4.98904109e-01 4.44875747e-01 7.41908431e-01 -1.34893548e+00
-1.24287255e-01 -1.39525875e-01 -5.25234520e-01 1.24903060e-01
-3.55631948e-01 5.06743848e-01 5.74833333e-01 -1.11599982e+00
1.67161003e-01 1.10609066e+00 9.66796398e-01 4.65284348e-01
-1.03337002e+00 -1.01155794e+00 3.10991347e-01 1.35041833e-01
-1.41914976e+00 -3.46242130e-01 8.87395024e-01 -5.61285794e-01
7.38471985e-01 2.97161281e-01 6.44799292e-01 9.13364649e-01
-2.40824640e-01 3.82161558e-01 1.01418924e+00 -1.00637935e-01
4.95274037e-01 3.68930101e-01 -2.06606299e-01 3.91199321e-01
5.49805582e-01 4.09685403e-01 -6.28010705e-02 3.79413366e-03
4.58506703e-01 1.47607461e-01 -1.20994233e-01 -2.97931194e-01
-6.54031217e-01 8.88939440e-01 7.62727201e-01 -2.99042404e-01
-3.53012651e-01 9.82298106e-02 7.66144395e-02 -4.43607755e-02
7.36595154e-01 4.47598130e-01 -2.80661017e-01 3.35923940e-01
-1.02704477e+00 5.68734944e-01 4.60265696e-01 8.08635712e-01
1.05259931e+00 1.41888782e-01 -4.03484628e-02 6.39762223e-01
5.06099284e-01 1.00087833e+00 -1.15729392e-01 -6.75561428e-01
5.32553852e-01 5.46455324e-01 3.53915364e-01 -9.62471902e-01
-3.41729283e-01 -3.83505225e-01 -5.51327527e-01 7.37299562e-01
1.72794878e-01 -6.15148246e-02 -1.11827111e+00 1.44327641e+00
6.57256603e-01 5.69809794e-01 -9.70777720e-02 1.11131620e+00
8.03245187e-01 5.10407627e-01 7.61232972e-02 2.17457309e-01
1.05248868e+00 -7.20484257e-01 -5.47926188e-01 -4.43922371e-01
3.87539834e-01 -5.43107152e-01 9.75792587e-01 1.15228213e-01
-8.24017048e-01 -6.24087334e-01 -1.27278686e+00 1.68915465e-01
-7.45804071e-01 1.50163308e-01 4.45838749e-01 7.10281730e-01
-6.96108878e-01 5.27487755e-01 -8.54202509e-01 -6.86165169e-02
4.50405359e-01 3.01071405e-01 2.90005095e-02 -1.05690565e-02
-9.80533957e-01 1.16232657e+00 3.74595195e-01 3.26574206e-01
-7.97266662e-01 -9.08424973e-01 -9.04163003e-01 -1.63851544e-01
6.36051416e-01 -4.79192853e-01 1.23372281e+00 -4.17780101e-01
-1.25052571e+00 4.31393415e-01 -2.78573543e-01 -5.57619452e-01
6.87837481e-01 -4.05542701e-01 -1.97107226e-01 -1.71702392e-02
1.04202546e-01 6.86627865e-01 8.55045795e-01 -1.55553961e+00
-6.40814364e-01 -2.70770311e-01 -1.00566633e-01 1.09434195e-01
-7.33376518e-02 -4.04985994e-01 -4.52679574e-01 -3.81349802e-01
5.20978756e-02 -1.05860317e+00 -3.35260630e-01 2.55602181e-01
-2.55437732e-01 -2.00154498e-01 1.16596019e+00 -3.17736059e-01
6.23034775e-01 -2.11107087e+00 -5.99969327e-01 3.52529705e-01
2.00686380e-01 5.76646984e-01 -8.95731617e-03 5.90517521e-02
1.75413445e-01 3.30656022e-01 -4.09211844e-01 -6.04310095e-01
-2.43122708e-02 2.57598847e-01 -4.57819700e-01 5.54251075e-01
7.99173832e-01 8.23011756e-01 -7.73373604e-01 -3.30223411e-01
7.28943527e-01 5.92505038e-01 -5.85081697e-01 2.36481756e-01
-4.14188564e-01 5.00645041e-01 -3.98365080e-01 5.96723318e-01
1.24983239e+00 2.01473042e-01 -3.75801384e-01 -2.20732763e-01
-4.08749789e-01 4.11837667e-01 -1.29293537e+00 1.10770547e+00
-7.03866601e-01 6.52233362e-01 1.34855106e-01 -8.03894818e-01
1.04200435e+00 2.24843677e-02 1.79096505e-01 -6.71179175e-01
2.01741725e-01 1.70524910e-01 -3.02799553e-01 -4.88347322e-01
7.11096585e-01 -3.67678225e-01 2.87084758e-01 1.27561569e-01
-3.84122968e-01 -6.54575288e-01 -8.16916600e-02 -9.87728238e-02
7.84245551e-01 6.36105686e-02 -2.12715536e-01 1.49495542e-01
3.27921212e-01 7.80803040e-02 7.40264833e-01 7.31891990e-01
-7.48353526e-02 7.18106210e-01 -1.69282109e-02 -2.21342817e-01
-1.27368271e+00 -1.06809378e+00 -2.26168320e-01 6.52038217e-01
2.65732467e-01 -5.63782752e-02 -4.46819782e-01 -6.27142251e-01
2.37225890e-01 8.62844706e-01 -4.00869101e-01 -2.88888216e-01
-6.92727387e-01 -8.38692963e-01 4.62497830e-01 8.30636203e-01
5.06045401e-01 -7.49438465e-01 -6.82139635e-01 5.96041679e-02
7.31974989e-02 -1.35855520e+00 -3.17585357e-02 3.72270755e-02
-8.43644202e-01 -1.08119178e+00 -2.60693669e-01 -2.52118677e-01
6.53621614e-01 5.82352519e-01 1.07586563e+00 2.81487465e-01
-3.15828711e-01 7.13993907e-02 -3.30451965e-01 -1.13738751e+00
-4.05246735e-01 2.89257281e-02 2.09002629e-01 -4.96840551e-02
3.57486010e-01 -4.52505350e-01 -7.24518538e-01 2.59667367e-01
-9.60226417e-01 -9.77370143e-02 5.10786712e-01 4.51153219e-01
4.64310735e-01 2.67279953e-01 6.06495559e-01 -6.68956935e-01
2.20537946e-01 -4.99502003e-01 -1.12331772e+00 -3.75776857e-01
-5.94229639e-01 -1.24059647e-01 5.03344953e-01 -3.53641182e-01
-1.01214290e+00 3.87274563e-01 -2.23716512e-01 -8.16279888e-01
-3.38323593e-01 1.09310329e-01 -5.33105396e-02 -2.08301291e-01
7.33468413e-01 -3.02393008e-02 -2.38684565e-02 -5.46021461e-01
3.04846853e-01 7.58295715e-01 5.81383646e-01 -3.54844421e-01
1.49404907e+00 9.42333460e-01 2.86215603e-01 -1.05414367e+00
-6.85693383e-01 -6.69835448e-01 -5.33356309e-01 -2.30606750e-01
7.08309829e-01 -1.03200293e+00 -6.16138816e-01 2.99846530e-01
-1.19966912e+00 -3.60192984e-01 -2.05899492e-01 4.34433430e-01
-1.55015603e-01 2.94942379e-01 -4.50694077e-02 -1.34983873e+00
-2.20120445e-01 -1.00889683e+00 1.17894590e+00 2.48976141e-01
2.39321932e-01 -6.76851153e-01 -6.66821897e-02 2.54601628e-01
4.31331426e-01 5.18847167e-01 4.69591916e-01 -4.61428881e-01
-8.72600675e-01 -4.05260324e-01 -4.04410094e-01 4.67625588e-01
7.53940456e-03 5.31363897e-02 -1.32655418e+00 -2.81832784e-01
1.53526627e-02 -1.07033536e-01 1.08754683e+00 1.00457512e-01
1.09940970e+00 8.81278515e-03 -5.25640845e-01 5.28058112e-01
1.32923090e+00 1.33732349e-01 4.19446915e-01 3.31416398e-01
8.26718748e-01 6.68436110e-01 8.17953110e-01 1.90024719e-01
3.00646037e-01 7.27765203e-01 9.62567031e-01 -3.96262079e-01
-1.15296915e-01 -1.12137318e-01 3.42636257e-02 8.99066627e-02
-7.68169761e-02 -2.51792133e-01 -8.38546753e-01 6.64687634e-01
-1.65529442e+00 -7.99739838e-01 -3.61843199e-01 2.54282928e+00
3.26540381e-01 2.20485628e-01 3.68408263e-02 1.50543496e-01
5.90023577e-01 8.77707452e-03 -6.92553818e-01 7.94821605e-02
1.64026961e-01 2.74224669e-01 9.13089693e-01 4.20094669e-01
-1.15721178e+00 9.06449199e-01 5.76358032e+00 3.72877508e-01
-1.40272915e+00 2.31998891e-01 1.90060034e-01 -3.34293187e-01
-1.87362075e-01 -1.96295291e-01 -9.59392071e-01 5.03135800e-01
8.13598096e-01 2.99564302e-01 1.34696037e-01 9.23289478e-01
6.29698157e-01 -5.20030558e-02 -8.36236477e-01 8.27058077e-01
-1.18543454e-01 -1.18199599e+00 -3.02628815e-01 4.36457470e-02
5.16006291e-01 5.21873713e-01 -5.42307682e-02 3.88415158e-01
2.09735453e-01 -8.73430908e-01 7.69607961e-01 3.65191728e-01
6.79545999e-01 -7.01279461e-01 6.59417629e-01 5.01701534e-01
-1.25029039e+00 -3.24922539e-02 -5.22380412e-01 -9.65057537e-02
2.28248686e-01 7.72984862e-01 -1.30946040e+00 3.80810678e-01
7.57338107e-01 3.10451329e-01 -6.09011233e-01 1.31388628e+00
-3.23343724e-01 6.31648362e-01 -8.56011391e-01 -3.02638020e-02
1.77128166e-01 -1.25691637e-01 6.99724674e-01 1.05814183e+00
2.78919786e-01 -4.66702692e-02 2.51946777e-01 1.27183557e+00
4.01558541e-02 -2.20206544e-01 -9.06712711e-01 4.35417384e-01
7.52268910e-01 1.35829520e+00 -5.97816229e-01 -3.39238763e-01
-3.46910745e-01 5.15756190e-01 2.02801540e-01 3.62663031e-01
-1.02093101e+00 -1.24844104e-01 7.17690945e-01 5.79136372e-01
4.33041453e-01 -3.64200681e-01 -4.03916895e-01 -9.50235903e-01
2.51281679e-01 -3.45437735e-01 -1.51303932e-01 -5.70408165e-01
-1.29092085e+00 3.08387041e-01 1.17820457e-01 -1.29447460e+00
-6.50862083e-02 -6.32386684e-01 -8.91509533e-01 9.02723312e-01
-2.10317850e+00 -1.01155877e+00 -6.98749304e-01 2.06617698e-01
5.16503096e-01 2.07543433e-01 3.64520907e-01 5.34112036e-01
-5.04182816e-01 3.86092454e-01 -2.24213526e-02 -1.93585698e-02
5.20315945e-01 -1.24319160e+00 8.03334892e-01 9.86131668e-01
-9.13309082e-02 3.97643328e-01 7.05992699e-01 -7.78293252e-01
-1.13200140e+00 -1.72700715e+00 3.89101624e-01 -5.65583289e-01
2.81619430e-01 -6.33033097e-01 -1.15064287e+00 4.20283556e-01
-3.72401506e-01 2.95017868e-01 2.82311887e-01 -2.07493857e-01
-2.86608517e-01 -1.81572914e-01 -1.29349613e+00 5.41502297e-01
8.03420544e-01 -3.10965359e-01 -4.56747800e-01 4.53310519e-01
6.72377765e-01 -5.49736917e-01 -3.22886139e-01 8.17224383e-01
3.20669830e-01 -7.40979910e-01 1.01069987e+00 -3.83035898e-01
-7.37540126e-02 -6.94537222e-01 3.79197560e-02 -1.31427562e+00
-1.07757397e-01 -2.99805701e-01 -6.07090332e-02 1.04206848e+00
4.20309216e-01 -8.24643373e-01 1.00372112e+00 7.32569396e-01
-5.13477623e-01 -4.68545228e-01 -1.02743864e+00 -1.00453925e+00
1.66720226e-01 -7.42225885e-01 7.52079666e-01 6.93594098e-01
-7.28893518e-01 1.32376105e-01 -1.75970912e-01 7.26328909e-01
5.75820267e-01 -1.52315674e-02 1.12423217e+00 -1.47717702e+00
1.73904225e-02 -2.14327395e-01 -1.82165280e-01 -8.29892337e-01
4.43237424e-02 -6.35769010e-01 5.24280012e-01 -1.40897501e+00
-1.83752760e-01 -1.00606918e+00 -4.00236733e-02 3.19686890e-01
-2.86214173e-01 5.78748226e-01 3.28441054e-01 1.74656317e-01
-1.03738047e-01 6.26135170e-01 8.62300873e-01 -8.49860981e-02
-2.10762113e-01 2.92586893e-01 -2.57598162e-01 8.74148428e-01
1.00417221e+00 -5.96563160e-01 -2.34806418e-01 -3.50697070e-01
2.61768252e-01 -5.82266867e-01 7.72223175e-01 -1.25268066e+00
-1.04519026e-02 -1.83508992e-01 4.18957293e-01 -9.44521725e-01
6.34552777e-01 -1.07288337e+00 -9.12564099e-02 3.72967452e-01
2.90822357e-01 -1.58219621e-01 4.99825984e-01 6.06919110e-01
1.29068822e-01 -3.32298040e-01 8.92571926e-01 -1.01448484e-01
-4.05758172e-01 3.35211635e-01 -3.74352634e-01 -3.48153621e-01
1.09519219e+00 -2.32003674e-01 -1.50745004e-01 -1.46108180e-01
-2.74009764e-01 2.72393674e-01 5.68227351e-01 5.52963793e-01
5.02254486e-01 -1.21032584e+00 -6.50585532e-01 3.02993268e-01
1.83826834e-01 6.84046984e-01 2.25042049e-02 6.01415157e-01
-4.07508105e-01 2.79021531e-01 3.69775504e-01 -7.46590376e-01
-1.12673318e+00 5.62654495e-01 3.77787858e-01 4.27588187e-02
-6.54491901e-01 5.96301138e-01 1.90683424e-01 -6.53927088e-01
-8.07742216e-03 -6.19250357e-01 -1.80307791e-01 3.82066891e-02
4.15045410e-01 4.11254346e-01 4.19289410e-01 -5.62025249e-01
-2.36833408e-01 3.93486381e-01 5.43682016e-02 5.04824258e-02
1.16946745e+00 2.24660467e-02 3.40168625e-01 2.74136394e-01
9.39740658e-01 9.31475312e-02 -1.44211376e+00 -7.15423673e-02
-3.29621732e-01 -5.53372085e-01 3.84192884e-01 -5.31387031e-01
-1.18833649e+00 9.49667454e-01 8.72083902e-01 1.80993617e-01
8.85121405e-01 -1.20616756e-01 8.79327118e-01 4.83087540e-01
2.50821207e-02 -1.03884482e+00 -2.53772940e-02 4.35013801e-01
6.85694933e-01 -1.69911110e+00 2.08977517e-02 -5.77349007e-01
-2.50952184e-01 7.65697122e-01 8.47933948e-01 -1.70581535e-01
4.69028205e-01 3.69869351e-01 2.03262195e-01 -3.39373827e-01
-3.78036380e-01 -3.95114928e-01 3.08581859e-01 6.64624095e-01
-1.06482483e-01 -6.69157356e-02 -7.25588873e-02 2.15836883e-01
-2.35355631e-01 -3.80432829e-02 3.41531157e-01 9.75269794e-01
-6.44346058e-01 -7.59981513e-01 -7.09936380e-01 5.57950497e-01
-2.41868243e-01 -1.04409635e-01 -2.42114350e-01 7.19091535e-01
4.16149646e-01 1.02461469e+00 4.27388668e-01 -3.32620978e-01
4.76158917e-01 1.71770714e-02 7.11613670e-02 -7.22926259e-01
-3.07372749e-01 -1.42259985e-01 -8.09018612e-02 -3.52674067e-01
-2.25505158e-01 -5.50109863e-01 -1.24669611e+00 -7.19935894e-02
-7.62028515e-01 -1.30760953e-01 1.14946532e+00 8.85653138e-01
2.58661211e-01 4.88874435e-01 6.28058612e-01 -1.30941224e+00
-5.17101109e-01 -1.01774120e+00 -2.91830510e-01 3.35186988e-01
5.74247003e-01 -1.10622549e+00 -6.18174672e-01 -2.67462969e-01] | [7.930769920349121, -2.5252139568328857] |
b0992e13-4d96-4cee-adb9-a6f45dd68408 | efficient-explainable-face-verification-based | 2304.13409 | null | https://arxiv.org/abs/2304.13409v1 | https://arxiv.org/pdf/2304.13409v1.pdf | Efficient Explainable Face Verification based on Similarity Score Argument Backpropagation | Explainable Face Recognition is gaining growing attention as the use of the technology is gaining ground in security-critical applications. Understanding why two faces images are matched or not matched by a given face recognition system is important to operators, users, anddevelopers to increase trust, accountability, develop better systems, and highlight unfair behavior. In this work, we propose xSSAB, an approach to back-propagate similarity score-based arguments that support or oppose the face matching decision to visualize spatial maps that indicate similar and dissimilar areas as interpreted by the underlying FR model. Furthermore, we present Patch-LFW, a new explainable face verification benchmark that enables along with a novel evaluation protocol, the first quantitative evaluation of the validity of similarity and dissimilarity maps in explainable face recognition approaches. We compare our efficient approach to state-of-the-art approaches demonstrating a superior trade-off between efficiency and performance. The code as well as the proposed Patch-LFW is publicly available at: https://github.com/marcohuber/xSSAB. | ['Naser Damer', 'Philipp Terhörst', 'Anh Thi Luu', 'Marco Huber'] | 2023-04-26 | null | null | null | null | ['face-recognition', 'face-verification'] | ['computer-vision', 'computer-vision'] | [ 2.88485944e-01 2.60571212e-01 -2.03810647e-01 -1.06536722e+00
-3.76637429e-01 -5.76990366e-01 6.37710631e-01 8.17886740e-02
4.02812541e-01 4.29503351e-01 1.23297505e-01 -5.62326133e-01
-4.18739706e-01 -5.02057612e-01 -5.78017712e-01 -4.16940123e-01
4.40829992e-02 1.96940035e-01 -2.77039200e-01 -7.51359239e-02
5.04477918e-01 9.31033909e-01 -1.86353481e+00 6.56079590e-01
4.88799036e-01 1.24425793e+00 -5.43551207e-01 3.78515929e-01
1.95036426e-01 5.10236323e-01 -4.83189493e-01 -9.21501219e-01
4.06998664e-01 -5.40112913e-01 -6.17015481e-01 -2.80739844e-01
1.06713033e+00 -2.62630910e-01 2.96855494e-02 1.16760671e+00
2.44161069e-01 -3.93487841e-01 4.00553375e-01 -1.90663493e+00
-9.85424936e-01 3.23396653e-01 -4.79053944e-01 1.32855028e-01
5.82406998e-01 2.01547116e-01 8.31952155e-01 -9.18883204e-01
5.87655842e-01 1.33408499e+00 5.33791482e-01 8.33680153e-01
-1.25293899e+00 -9.38851893e-01 -1.24845169e-01 4.78839070e-01
-1.47971213e+00 -9.22311842e-01 5.82407773e-01 -2.86701173e-01
7.20651686e-01 9.20583487e-01 2.97855407e-01 1.07570708e+00
1.99450720e-02 2.01168373e-01 1.28801787e+00 -4.01238829e-01
1.72777861e-01 3.06204140e-01 2.48360291e-01 7.87887275e-01
5.26820123e-01 3.12696278e-01 -8.28815997e-01 -4.16018456e-01
3.27802837e-01 1.18892118e-01 -2.60076463e-01 -4.52917367e-01
-5.97908497e-01 6.51427746e-01 4.12511975e-01 1.32377803e-01
-2.77372807e-01 2.13508263e-01 -8.29213858e-02 3.06108445e-01
4.60380971e-01 2.85961032e-01 -1.13150179e-02 -9.59316418e-02
-1.09978604e+00 2.53386289e-01 5.52895725e-01 5.16690731e-01
6.01812780e-01 -2.33278185e-01 -1.77753225e-01 3.27675998e-01
5.64189196e-01 5.85693359e-01 -6.95612058e-02 -1.08318949e+00
-7.71718053e-03 7.27555633e-01 3.94520536e-03 -1.38386488e+00
1.87145025e-02 -1.23625912e-01 -5.31396985e-01 6.72992229e-01
3.36267322e-01 4.21617419e-01 -4.88800198e-01 1.61067617e+00
2.92437881e-01 4.24911708e-01 -1.65736288e-01 9.03499722e-01
7.56648362e-01 1.30055308e-01 -5.29523231e-02 1.09727092e-01
1.46595252e+00 -4.76428807e-01 -6.13185346e-01 1.91397354e-01
3.02134305e-01 -9.42136824e-01 9.79922295e-01 1.97183698e-01
-9.76384938e-01 -3.47253084e-01 -1.11200416e+00 1.11519404e-01
-4.68282163e-01 2.32980736e-02 5.23514628e-01 1.19609523e+00
-1.22220719e+00 8.04509103e-01 -5.29357612e-01 -5.39435923e-01
9.87229705e-01 4.81680095e-01 -7.51479149e-01 9.17810947e-03
-7.02584982e-01 8.42261910e-01 -2.87982583e-01 1.91597626e-01
-6.11979842e-01 -8.12029898e-01 -6.07739508e-01 2.76311189e-01
5.74233383e-02 -2.08727986e-01 8.88377070e-01 -1.00130522e+00
-9.23740327e-01 1.13268828e+00 -3.38705719e-01 -3.68762523e-01
4.75822598e-01 -3.12752686e-02 -6.59450471e-01 -8.79237130e-02
-8.81477520e-02 7.28733718e-01 8.71167183e-01 -1.26923561e+00
-2.40093902e-01 -5.93340993e-01 5.24560846e-02 -4.78256077e-01
-3.10852408e-01 3.95037413e-01 2.90457923e-02 -4.97108519e-01
-5.82783744e-02 -8.59485149e-01 3.46600324e-01 5.64985454e-01
-5.99538207e-01 -1.00742601e-01 1.16456866e+00 -5.98647773e-01
1.11996162e+00 -2.24329686e+00 -2.85794228e-01 6.87331855e-01
2.12349877e-01 3.57470393e-01 -2.02997148e-01 1.63610265e-01
-4.91543651e-01 5.98607779e-01 -1.81683868e-01 -3.40314955e-01
2.58959681e-01 -9.99579653e-02 -4.95788574e-01 7.11626828e-01
4.92386222e-01 6.47148788e-01 -5.09401143e-01 -2.02359125e-01
1.68350711e-01 9.20764208e-01 -4.65268046e-01 4.66593876e-02
2.26093173e-01 2.36359730e-01 -1.46731809e-01 8.50467563e-01
1.11520386e+00 -2.02104762e-01 3.19860131e-01 -3.55605304e-01
4.94736172e-02 1.25352919e-01 -1.10950053e+00 1.23982477e+00
-3.30567406e-03 9.41558421e-01 -2.29370505e-01 -5.91180384e-01
1.07069445e+00 2.21751124e-01 7.66996518e-02 -7.25161433e-01
7.28959143e-02 1.59677938e-01 -1.50297835e-01 -3.38475615e-01
2.89814442e-01 2.00198427e-01 5.02010286e-01 7.00049579e-01
-3.25528830e-01 1.98873341e-01 -2.42892236e-01 1.02381699e-01
9.89369929e-01 1.00666143e-01 3.85466158e-01 -4.58577126e-01
7.74733603e-01 -4.00346309e-01 2.34693155e-01 4.63676482e-01
-3.63659084e-01 6.86869264e-01 7.35072255e-01 -6.31152511e-01
-7.81524122e-01 -9.31057930e-01 -2.44773373e-01 6.37756884e-01
2.11298913e-01 -4.74877894e-01 -1.01118767e+00 -8.88506711e-01
3.40767652e-01 7.47952521e-01 -9.89444852e-01 -1.65926844e-01
-2.59715497e-01 -9.95731503e-02 7.52026916e-01 3.18762809e-01
3.60282153e-01 -8.75812650e-01 -8.10535252e-01 -6.80417538e-01
1.15915067e-01 -8.24018657e-01 -3.38022292e-01 -4.80829656e-01
-4.08241630e-01 -1.34078646e+00 -3.06258291e-01 -1.97005615e-01
1.06801105e+00 2.79393256e-01 1.05070066e+00 8.60421062e-01
-5.40303767e-01 3.63794953e-01 -2.57882595e-01 -2.90890306e-01
-4.55402911e-01 -2.43663788e-01 1.43067822e-01 2.49013111e-01
5.60570002e-01 -3.41685206e-01 -7.31711984e-01 8.36999655e-01
-8.22180808e-01 -1.26919001e-01 4.29575890e-02 4.85380948e-01
2.19283774e-01 -4.10472333e-01 3.08572948e-01 -7.35549629e-01
6.14952028e-01 -3.52743715e-01 -6.92712605e-01 7.26590753e-01
-8.61937165e-01 9.23590884e-02 2.98509467e-03 -3.64503339e-02
-7.25053728e-01 -1.37224272e-01 7.89428353e-02 -3.54165584e-01
-2.50225306e-01 -9.79723260e-02 -3.56695324e-01 -4.36090052e-01
8.03288579e-01 -3.15218717e-01 2.64746487e-01 -3.47079545e-01
3.98452133e-01 8.34069908e-01 3.92795742e-01 -3.20283055e-01
9.68124092e-01 6.83585227e-01 2.82114577e-02 -3.35820168e-01
-2.48572692e-01 -6.87479079e-02 -3.61003995e-01 -4.25306141e-01
3.82817268e-01 -2.79577702e-01 -1.12104166e+00 -4.37988229e-02
-1.33238780e+00 1.38411820e-01 -1.44616663e-01 -3.92159112e-02
-2.29307905e-01 1.97876573e-01 9.54901502e-02 -9.25417542e-01
-2.24253133e-01 -1.23609459e+00 1.09519684e+00 3.42345327e-01
-6.34423375e-01 -5.36839545e-01 -2.44017780e-01 5.94633877e-01
7.64355600e-01 3.53406996e-01 7.70441473e-01 -7.29349732e-01
-6.06163442e-01 -2.79746860e-01 -5.27557135e-01 1.28553614e-01
2.60772467e-01 4.66299564e-01 -1.38906145e+00 -2.25935891e-01
-3.94129425e-01 -5.44048697e-02 3.76418829e-01 1.52527094e-01
1.16157901e+00 -6.13598764e-01 -2.69847840e-01 4.62991059e-01
1.03672540e+00 4.08292562e-02 8.99742782e-01 2.12358460e-01
2.33291045e-01 1.03558135e+00 5.24055004e-01 4.14596945e-01
2.77965665e-02 1.02030516e+00 7.15337515e-01 -1.88147902e-01
-1.91793919e-01 -2.74461478e-01 3.35431784e-01 -2.29909316e-01
-1.71510711e-01 -1.09101184e-01 -8.66968989e-01 2.87567936e-02
-1.80557299e+00 -1.30803466e+00 -2.57203821e-02 2.33001685e+00
2.29770228e-01 -2.97292292e-01 6.71068281e-02 3.16734314e-01
8.39444220e-01 2.55030999e-03 -4.17089701e-01 -6.60986483e-01
-6.44891933e-02 2.92162567e-01 8.78000706e-02 5.91094911e-01
-7.23995507e-01 6.93477094e-01 5.92440557e+00 6.50081873e-01
-1.40470982e+00 1.09671675e-01 1.10279047e+00 -2.95117367e-02
-5.01706004e-01 4.07150388e-02 -5.64139485e-01 3.31247211e-01
8.74178171e-01 -2.96065956e-01 6.19367003e-01 9.12839115e-01
2.70063460e-01 1.67377546e-01 -1.36161089e+00 1.06282902e+00
3.03480268e-01 -1.66921771e+00 5.59854731e-02 2.41774246e-01
3.35606277e-01 -4.56847370e-01 6.55007243e-01 -4.60523903e-01
-1.19991638e-01 -1.47892547e+00 9.62785184e-01 4.77919906e-01
9.00872588e-01 -5.25682330e-01 6.43685281e-01 -4.66934562e-01
-1.05826604e+00 9.31228548e-02 -2.64583856e-01 1.08126141e-01
-2.66219109e-01 3.22540402e-01 -8.79449844e-01 4.16488290e-01
8.27965200e-01 4.24029440e-01 -8.28656375e-01 7.76243985e-01
-3.20210725e-01 4.12850410e-01 -1.93049833e-02 1.80829108e-01
-3.35034281e-01 1.77653745e-01 3.68197560e-01 9.52435017e-01
4.87093300e-01 -6.34910688e-02 -6.68642938e-01 1.29467213e+00
-1.28150508e-01 5.97988181e-02 -6.92819834e-01 -5.74429967e-02
6.65655732e-01 1.27851689e+00 -7.17770159e-01 -7.48423338e-02
-1.55513585e-01 8.82973194e-01 -3.90484035e-02 1.39832377e-01
-9.67225730e-01 -1.03446767e-01 1.09440053e+00 4.91047233e-01
1.02308288e-01 1.71499535e-01 -3.39334577e-01 -7.54181623e-01
2.22030431e-01 -1.13318157e+00 3.04016232e-01 -9.05665278e-01
-1.07750666e+00 1.01801038e+00 -5.73271550e-02 -1.17126215e+00
-1.02518514e-01 -6.39863372e-01 -5.56308925e-01 8.80387306e-01
-1.53792882e+00 -1.20688140e+00 -6.11255527e-01 4.35825527e-01
4.84728031e-02 -3.20316464e-01 9.53685820e-01 2.37003490e-01
-4.38490838e-01 9.19450223e-01 -3.91264737e-01 -7.26336986e-02
7.07136929e-01 -7.33852983e-01 6.01193607e-01 8.77028465e-01
3.94631416e-01 1.01315820e+00 7.77975380e-01 -3.86191398e-01
-1.24259460e+00 -7.64589429e-01 8.89403462e-01 -5.57963550e-01
3.64253819e-01 -3.42268765e-01 -9.22787964e-01 4.33869839e-01
2.84593612e-01 4.04240042e-01 9.67295647e-01 1.21660993e-01
-9.83311415e-01 -5.14606535e-01 -1.79542804e+00 4.27959800e-01
1.01106000e+00 -6.93639398e-01 -1.90731987e-01 1.31037325e-01
1.50660947e-01 -8.06344673e-02 -6.29141510e-01 3.44255090e-01
9.95004773e-01 -1.55589163e+00 7.84493566e-01 -6.32804155e-01
3.01767766e-01 -5.53732991e-01 -3.66215587e-01 -8.39091897e-01
-1.22168690e-01 -7.22151041e-01 6.40299842e-02 1.27402270e+00
6.51136816e-01 -9.15945351e-01 9.92385209e-01 1.04022241e+00
3.98591667e-01 -6.36785686e-01 -1.14863443e+00 -7.64648020e-01
-4.67382073e-01 -2.37671211e-01 1.32215559e+00 9.94915068e-01
1.40607968e-01 -5.33372879e-01 -3.03694695e-01 5.18730402e-01
7.12634921e-01 4.51378971e-02 7.75886178e-01 -1.17458057e+00
1.16053440e-01 -4.76691008e-01 -1.06851113e+00 -2.82885492e-01
2.11499497e-01 -8.71630073e-01 -4.18513834e-01 -1.04244101e+00
3.69759440e-01 -4.13700372e-01 -3.41147989e-01 9.50521648e-01
1.92492649e-01 6.64471567e-01 4.27514255e-01 2.69861549e-01
-3.16451192e-01 1.30657658e-01 5.93101263e-01 -1.47032663e-01
3.63862067e-01 -3.05767179e-01 -9.62220073e-01 4.17415947e-01
8.56472909e-01 -5.31363070e-01 -3.03259879e-01 -2.94630229e-01
1.11799203e-01 -2.66288757e-01 8.74250948e-01 -1.00732255e+00
2.17900306e-01 -1.11691616e-01 3.39002132e-01 -7.54316673e-02
3.36679310e-01 -1.11731648e+00 7.09231794e-01 6.02291644e-01
-3.99014115e-01 4.16023016e-01 3.44196558e-01 1.69123232e-01
-1.08009145e-01 6.15006536e-02 7.12990463e-01 4.26399231e-01
-4.12604302e-01 2.77851582e-01 2.37426862e-01 -5.04251599e-01
1.20334029e+00 -4.80518013e-01 -9.36571598e-01 -4.56274688e-01
-4.14625734e-01 -2.52225548e-01 7.42017925e-01 7.24000096e-01
1.00751662e+00 -1.40400290e+00 -7.42987096e-01 6.37674034e-01
2.20939144e-01 -1.07910419e+00 2.24217966e-01 6.82798564e-01
-5.46465695e-01 3.03039223e-01 -4.46978092e-01 -5.63324690e-01
-2.03964114e+00 1.91238433e-01 4.58476067e-01 3.93389106e-01
-2.18974933e-01 8.78903389e-01 -2.34163217e-02 -2.55582243e-01
6.42244443e-02 -1.17498890e-01 1.32849753e-01 -3.70510757e-01
8.89864743e-01 3.72594357e-01 1.81023553e-01 -8.96715522e-01
-8.98399472e-01 4.36509013e-01 1.46219302e-02 2.77597364e-02
1.19269180e+00 2.11252898e-01 -2.73811877e-01 -1.78109586e-01
1.00558770e+00 1.07946955e-01 -9.58457470e-01 3.54868352e-01
8.72612372e-02 -1.23780835e+00 -2.88872242e-01 -1.01185596e+00
-1.28292382e+00 7.63245106e-01 1.19590342e+00 4.90973204e-01
1.02115917e+00 9.29518603e-04 1.93689212e-01 -4.19281498e-02
2.46487185e-01 -3.91371220e-01 -1.06858015e-01 -3.85616630e-01
1.14709818e+00 -1.27279496e+00 2.38407657e-01 -6.50573611e-01
-5.53564668e-01 1.11818326e+00 3.97680223e-01 3.02816063e-01
6.90325797e-01 3.43447059e-01 3.54487211e-01 -4.00752693e-01
-6.06719255e-01 3.42163712e-01 4.19157058e-01 6.52837932e-01
6.03965998e-01 1.83698341e-01 -8.71792715e-03 3.72648209e-01
-2.82436192e-01 9.44822505e-02 1.71130955e-01 6.80990100e-01
-1.01816326e-01 -1.33797073e+00 -5.48016846e-01 2.09210262e-01
-3.02223235e-01 -2.58919019e-02 -8.47770214e-01 6.75971031e-01
2.24003538e-01 1.08496451e+00 1.06463715e-01 -5.48140585e-01
2.67295808e-01 -9.05823112e-02 4.03736234e-01 -7.56690949e-02
-6.23942256e-01 -7.13581979e-01 1.23461941e-02 -9.55585182e-01
-5.49536169e-01 -5.98385572e-01 -9.95619357e-01 -9.15784419e-01
-3.79789591e-01 2.25694641e-01 1.05371618e+00 6.43555820e-01
9.53943253e-01 -6.26938567e-02 7.15055287e-01 -5.42019546e-01
-2.15065971e-01 -5.10835648e-01 -2.13963300e-01 5.69603026e-01
2.24546328e-01 -6.15541697e-01 -4.36980575e-01 5.92608750e-02] | [12.93997859954834, 0.9519627094268799] |
83185a46-5706-468e-854f-0b008a1eb955 | statistical-tomography-of-microscopic-life | null | null | http://openaccess.thecvf.com/content_cvpr_2018/html/Levis_Statistical_Tomography_of_CVPR_2018_paper.html | http://openaccess.thecvf.com/content_cvpr_2018/papers/Levis_Statistical_Tomography_of_CVPR_2018_paper.pdf | Statistical Tomography of Microscopic Life | We achieve tomography of 3D volumetric natural objects, where each projected 2D image corresponds to a different specimen. Each specimen has unknown random 3D orientation, location, and scale. This imaging scenario is relevant to microscopic and mesoscopic organisms, aerosols and hydrosols viewed naturally by a microscope. In-class scale variation inhibits prior single-particle reconstruction methods. We thus generalize tomographic recovery to account for all degrees of freedom of a similarity transformation. This enables geometric self-calibration in imaging of transparent objects. We make the computational load manageable and reach good quality reconstruction in a short time. This enables extraction of statistics that are important for a scientific study of specimen populations, specifically size distribution parameters. We apply the method to study of plankton. | ['Aviad Levis', 'Ronen Talmon', 'Yoav Y. Schechner'] | 2018-06-01 | null | null | null | cvpr-2018-6 | ['transparent-objects'] | ['computer-vision'] | [ 2.34285742e-01 -1.99387565e-01 7.44448721e-01 3.39362212e-02
-2.54499555e-01 -6.08707309e-01 5.11870921e-01 1.17417216e-01
-6.89494252e-01 7.82928586e-01 -4.57974702e-01 -5.62976720e-03
9.47307795e-02 -1.02036047e+00 -6.71579719e-01 -1.01518726e+00
-6.48593530e-02 1.21701574e+00 6.33598685e-01 3.15455467e-01
2.44030416e-01 8.44694972e-01 -1.28006244e+00 -3.79149586e-01
3.89340878e-01 5.29930532e-01 7.67744601e-01 8.78868937e-01
1.31478891e-01 1.65859640e-01 -3.10536414e-01 -9.00283605e-02
3.81312788e-01 -2.25597873e-01 -5.26025832e-01 5.36517024e-01
5.14548123e-01 -4.29628611e-01 2.09965825e-01 1.11871243e+00
2.55271584e-01 2.41790369e-01 1.27474296e+00 -3.11560631e-01
-4.79322731e-01 7.22931623e-02 -8.05502236e-01 2.10526705e-01
1.10596297e-02 8.67139623e-02 6.91085041e-01 -8.48158896e-01
5.15367091e-01 1.20093894e+00 6.74091756e-01 4.50244486e-01
-1.38464999e+00 3.82302795e-03 -3.27660680e-01 -2.56810993e-01
-1.40303850e+00 -2.19794109e-01 3.19438308e-01 -7.62891173e-01
6.65769994e-01 2.34555587e-01 7.89222062e-01 4.54361647e-01
5.42227924e-01 -1.43967226e-01 1.40728748e+00 -4.28179711e-01
4.62957799e-01 2.17871666e-01 -3.83052342e-02 5.74775159e-01
9.66896892e-01 -2.37001613e-01 -1.53877318e-01 -2.75765330e-01
1.24769819e+00 2.44800642e-01 -4.59883690e-01 -2.34504104e-01
-1.11888170e+00 5.83600700e-01 -5.96045144e-02 -2.48180591e-02
-2.35254973e-01 -2.26199627e-02 -2.78480738e-01 7.07275942e-02
5.96035898e-01 4.86646265e-01 -3.97001982e-01 1.62993416e-01
-6.79848194e-01 2.34097257e-01 1.08307660e+00 8.60746324e-01
1.04753327e+00 5.00670895e-02 4.84178573e-01 4.31065291e-01
5.07558465e-01 1.56864738e+00 2.43502453e-01 -1.13464975e+00
-9.85936150e-02 2.20065191e-01 4.79068488e-01 -5.42984784e-01
-4.28630710e-01 -1.33324787e-01 -7.76130319e-01 5.90304971e-01
4.72329825e-01 1.78043887e-01 -6.46316230e-01 1.14502168e+00
7.55477011e-01 1.16016187e-01 8.38635266e-02 6.81800723e-01
6.73120558e-01 4.08101737e-01 -6.66710317e-01 -5.74015975e-01
1.57030833e+00 -2.86184102e-01 -6.39626905e-02 -1.34308606e-01
1.31669091e-02 -7.03815222e-01 7.55466759e-01 2.18227655e-01
-1.27947581e+00 -1.18539862e-01 -6.47033811e-01 1.77129209e-01
2.77841017e-02 -1.91538692e-01 1.72439680e-01 5.54345191e-01
-8.07683706e-01 9.43942904e-01 -1.30928314e+00 -4.98237967e-01
8.19690153e-02 3.33614081e-01 -3.43862265e-01 3.07264894e-01
-2.22235039e-01 7.49826670e-01 1.02486918e-02 -1.40734151e-01
-7.23596632e-01 -7.71576583e-01 -4.14489955e-01 -2.78354406e-01
-1.49150938e-01 -1.03266573e+00 1.15278113e+00 -3.02807689e-01
-1.51894879e+00 1.21616328e+00 -3.78831714e-01 -2.20555425e-01
3.64771128e-01 1.15123615e-01 3.78684923e-02 6.04348898e-01
1.42336905e-01 1.67259738e-01 8.93726408e-01 -1.54211068e+00
-1.79525867e-01 -6.62910998e-01 -2.57333785e-01 4.19955581e-01
1.27234116e-01 -1.89171806e-01 6.43614382e-02 6.66748956e-02
3.20970953e-01 -1.09142470e+00 -1.30329698e-01 1.50388733e-01
-2.13561282e-01 5.05442083e-01 8.64892483e-01 -3.69223237e-01
2.08866164e-01 -1.71667838e+00 3.36093277e-01 -1.33537248e-01
3.86568427e-01 -2.00019509e-01 2.31972560e-01 4.45467979e-01
5.45853555e-01 -3.21090966e-02 -5.07050216e-01 -6.78749144e-01
-3.39188576e-01 4.24866617e-01 -8.95483345e-02 1.11092937e+00
3.79862785e-02 3.01467657e-01 -6.22307897e-01 -5.93172669e-01
4.21784252e-01 5.17930984e-01 -6.17814124e-01 3.90062809e-01
-6.66757077e-02 8.82254422e-01 -6.57145858e-01 4.36342627e-01
1.27812779e+00 -5.06421566e-01 -1.03231668e-01 8.10351744e-02
-6.16970956e-01 9.76483375e-02 -1.01587582e+00 9.93791580e-01
-6.07632220e-01 5.16061127e-01 5.10115683e-01 -4.49138612e-01
5.90706110e-01 1.76809117e-01 5.21454155e-01 -2.29835752e-02
2.66632378e-01 1.65383726e-01 4.87694256e-02 -5.12566268e-01
5.34207582e-01 -8.66752088e-01 4.79365885e-01 4.64400560e-01
-1.80692807e-01 -1.03337312e+00 -1.02739871e-01 1.23136632e-01
8.73158753e-01 -8.94513428e-02 5.94043136e-01 -8.77436936e-01
2.01620013e-01 -2.65884876e-01 1.71972513e-01 6.77436471e-01
3.54652017e-01 9.65508759e-01 -8.76310561e-03 -3.13500911e-01
-1.53530121e+00 -1.19136477e+00 -8.70845318e-01 2.16323107e-01
2.67290890e-01 2.21633092e-01 -6.00124061e-01 2.09448621e-01
1.65890262e-01 6.40913993e-02 -6.52216017e-01 2.92587340e-01
-5.95344365e-01 -1.20324981e+00 -6.77748919e-02 -4.20178138e-02
3.11339200e-01 -8.52397859e-01 -1.08689988e+00 -4.62129787e-02
1.01617888e-01 -1.32335806e+00 1.39056090e-02 7.97259882e-02
-1.22719657e+00 -1.16270578e+00 -7.13221192e-01 -5.22517979e-01
8.80218685e-01 6.20441139e-01 1.25478446e+00 6.33081868e-02
-2.50994176e-01 6.19697332e-01 -2.35794365e-01 -3.37259769e-01
-5.20674884e-01 -2.42033318e-01 5.13576806e-01 -1.42893448e-01
2.12366618e-02 -9.40660775e-01 -6.06554449e-01 4.66476649e-01
-6.67614400e-01 -1.49873823e-01 7.09679797e-02 5.17753184e-01
7.97873497e-01 -2.88537936e-03 -4.14544821e-01 -7.45851815e-01
-4.98352796e-02 -4.06250894e-01 -1.13948715e+00 -4.11372036e-02
-1.73968241e-01 -2.09328622e-01 7.81245828e-01 -4.32684600e-01
-9.09003377e-01 -2.81632513e-01 2.25277722e-01 -2.88062155e-01
-3.14814806e-01 2.24226154e-02 2.44788602e-01 -4.72916096e-01
6.08821213e-01 3.54962617e-01 8.96203145e-02 -6.40882194e-01
-3.43397498e-01 4.15571958e-01 3.82259160e-01 -6.74451232e-01
1.06361425e+00 1.34812593e+00 7.90174365e-01 -1.86555076e+00
-6.14871919e-01 -6.05375886e-01 -8.84448111e-01 -1.30678788e-01
7.47027040e-01 -9.30180371e-01 -1.12554061e+00 4.23521012e-01
-1.02297413e+00 -5.85808873e-01 -6.63303494e-01 9.23158944e-01
-3.59249532e-01 6.42947257e-01 -4.96740192e-01 -9.54249144e-01
-1.32619843e-01 -1.15796280e+00 1.46248019e+00 3.03750277e-01
3.30554724e-01 -1.55046177e+00 4.97904152e-01 1.35103175e-02
2.54837275e-01 1.23695225e-01 4.22130615e-01 7.41828233e-02
-9.08539593e-01 3.83445829e-01 -1.54027581e-01 8.57810453e-02
1.30340636e-01 4.06900436e-01 -9.55739439e-01 -5.48267424e-01
7.84885705e-01 -1.36469260e-01 7.85975456e-01 7.81581104e-01
7.88326621e-01 -2.34428674e-01 -2.34363660e-01 1.02231359e+00
1.58235729e+00 -2.38975778e-01 2.20859334e-01 7.43326545e-02
7.71648824e-01 6.12913072e-01 1.22791015e-01 7.85637379e-01
2.83649832e-01 6.66530311e-01 6.80903792e-01 1.33293182e-01
-1.18493550e-01 4.43857640e-01 1.40640423e-01 1.03378117e+00
-6.92491412e-01 -4.82033581e-01 -8.49540770e-01 3.96241993e-01
-1.12565899e+00 -8.68026257e-01 -7.95394719e-01 2.55760312e+00
5.81749320e-01 -4.77967411e-01 -1.12288095e-01 -3.37694943e-01
4.14054841e-01 -2.15074554e-01 -4.57926989e-01 1.84403405e-01
-2.87641793e-01 3.44030708e-01 8.91287625e-01 1.08536947e+00
-6.85753047e-01 6.22199237e-01 6.76280308e+00 2.97163308e-01
-9.45359588e-01 2.11910591e-01 1.31307334e-01 1.28115907e-01
-5.95175147e-01 2.50762939e-01 -1.18628514e+00 3.24228078e-01
7.75978267e-01 -1.98277142e-02 3.64422888e-01 2.50051737e-01
4.78899181e-01 -4.90073800e-01 -8.14870477e-01 6.91399097e-01
-8.94364715e-02 -9.73500371e-01 1.63102597e-01 6.66229665e-01
9.42801952e-01 5.84212482e-01 -6.66889101e-02 -6.28997386e-01
2.72321105e-01 -5.89717567e-01 4.86908048e-01 6.25403464e-01
9.42968249e-01 -2.12636560e-01 4.76156294e-01 4.45106715e-01
-1.15257859e+00 4.33063090e-01 -1.05758631e+00 -3.26792836e-01
2.82553524e-01 9.26595271e-01 -8.70627522e-01 1.78284734e-01
6.51528955e-01 5.44818103e-01 -2.95034081e-01 9.53324735e-01
1.09146908e-02 6.65925086e-01 -7.61777520e-01 -6.97269943e-03
-1.20631725e-01 -1.06410575e+00 9.82271492e-01 1.02561128e+00
7.04646289e-01 4.47657406e-01 5.09577282e-02 9.54072714e-01
8.11337829e-02 1.49178207e-01 -6.67810798e-01 3.65469515e-01
2.56544113e-01 1.25894761e+00 -1.19778192e+00 -1.74136370e-01
-2.63166636e-01 7.97286749e-01 6.93035945e-02 1.81037933e-01
-3.60319644e-01 6.34594560e-02 5.71330369e-01 5.76768160e-01
5.05917788e-01 -5.48957527e-01 -1.15776844e-01 -1.60823452e+00
-7.15954974e-02 -1.55681878e-01 -2.21354544e-01 -7.72865415e-01
-1.37504673e+00 3.25912327e-01 3.51790547e-01 -1.18106401e+00
1.01592086e-01 -1.05089879e+00 -7.05554724e-01 8.00509274e-01
-1.40448427e+00 -8.15571427e-01 -5.43475866e-01 3.52121264e-01
8.95116702e-02 2.09075376e-01 8.32147121e-01 -2.15875506e-01
-1.63344070e-01 -3.99845093e-01 8.55246127e-01 -3.93055350e-01
3.63922775e-01 -1.55109584e+00 3.08872283e-01 6.25577211e-01
-3.27671468e-02 5.47294855e-01 1.00453281e+00 -7.28363276e-01
-1.48160410e+00 -9.06315029e-01 4.74436134e-01 -7.26731956e-01
8.62908244e-01 -2.81401336e-01 -1.02327013e+00 5.88898122e-01
8.92423391e-02 3.33142757e-01 6.00116432e-01 -3.80405664e-01
3.33024934e-02 8.26449543e-02 -1.30557954e+00 2.55111933e-01
7.62059152e-01 -4.58001882e-01 -3.46076012e-01 6.69762790e-01
2.43104368e-01 -3.36183935e-01 -9.18802023e-01 1.31027400e-01
5.86021841e-01 -1.04572415e+00 1.03863204e+00 -8.40213224e-02
1.96090311e-01 -4.26357806e-01 -3.23819816e-01 -1.16442597e+00
-1.82927221e-01 -5.51438570e-01 2.81584978e-01 7.03844607e-01
1.16932064e-01 -8.35016012e-01 6.67557180e-01 2.20665678e-01
-1.35897934e-01 -6.75136819e-02 -9.42851365e-01 -9.28348124e-01
2.59428382e-01 1.20795816e-01 1.91171840e-01 6.78323209e-01
-4.15108204e-01 2.85077125e-01 -2.64607698e-01 7.47501373e-01
1.36475492e+00 4.97503728e-01 7.08282709e-01 -1.65730858e+00
-7.41274178e-01 -1.62450835e-01 -3.97310257e-01 -1.19716275e+00
-2.72466186e-02 -5.82895398e-01 1.63566470e-01 -1.16390944e+00
6.95349395e-01 -3.71694833e-01 5.90831816e-01 -3.50348473e-01
7.45404586e-02 6.70098424e-01 -1.28982067e-01 6.71086550e-01
-2.54034102e-01 4.97153789e-01 1.48712492e+00 2.72757351e-01
-6.87435549e-03 2.48568863e-01 4.64900816e-03 1.00831795e+00
8.54571402e-01 -7.34809101e-01 -2.02646032e-01 -7.48256803e-01
2.11472228e-01 3.43834385e-02 5.31127632e-01 -9.85028863e-01
-6.92483485e-02 -4.73303407e-01 1.67608038e-01 -3.27077389e-01
7.42820263e-01 -9.69187319e-01 2.76459545e-01 4.91591901e-01
4.26495314e-01 -1.53128088e-01 -1.05734885e-01 5.64127922e-01
2.31379509e-01 -7.45464683e-01 1.36026943e+00 -7.21532285e-01
3.57054681e-01 4.32908595e-01 -5.17372847e-01 2.23604336e-01
6.84791088e-01 -9.02077928e-02 -3.64576191e-01 -1.64233461e-01
-4.74268287e-01 -3.96253735e-01 1.31031370e+00 -6.97843015e-01
6.76503122e-01 -7.02769935e-01 -7.91434109e-01 7.23131001e-02
-3.37257743e-01 5.25358319e-01 1.43384323e-01 8.34667444e-01
-1.32105303e+00 -1.08915791e-01 -1.61668867e-01 -9.82371807e-01
-1.37891698e+00 1.74842939e-01 4.50988263e-01 1.16882948e-02
-8.30325603e-01 9.49885368e-01 6.51931286e-01 -4.47233558e-01
-5.52552760e-01 -4.64164823e-01 -1.92076992e-02 -3.61573696e-01
6.07974887e-01 4.76837754e-01 -7.16464520e-02 -8.08618486e-01
-2.94315368e-01 1.20117068e+00 2.56153226e-01 -1.44507542e-01
1.46187043e+00 -3.79553199e-01 -4.49338824e-01 9.51226354e-01
9.75140452e-01 4.65915143e-01 -1.55751085e+00 -1.48419768e-01
-7.41535664e-01 -7.24295497e-01 -1.26880601e-01 1.64200872e-01
-8.96999180e-01 1.10754716e+00 2.22379062e-02 5.58045208e-01
5.49338758e-01 2.23984614e-01 1.66822091e-01 4.99209225e-01
5.64657688e-01 -2.83596545e-01 -1.10718325e-01 5.66899896e-01
6.08486772e-01 -1.30083323e+00 6.07788801e-01 -3.03052455e-01
1.43794855e-02 1.19575346e+00 2.68230319e-01 -3.36046100e-01
8.69600356e-01 4.83066022e-01 -2.07169518e-01 -3.12420756e-01
-5.92308581e-01 -7.10973144e-02 2.98407134e-02 6.50348723e-01
1.71030164e-01 1.79189190e-01 -2.33161505e-02 -3.98121744e-01
-1.43464386e-01 -7.13566959e-01 1.01276827e+00 6.76485300e-01
-8.02606702e-01 -9.15509164e-01 -5.55834293e-01 3.47454339e-01
-3.82526159e-01 -1.17435902e-01 -2.08646641e-03 6.74170077e-01
-7.70612583e-02 4.57023323e-01 4.50329304e-01 2.25240946e-01
2.90149171e-03 -4.19425666e-01 9.40852761e-01 -9.08164203e-01
2.53651943e-02 2.56979793e-01 -2.91691154e-01 1.43685073e-01
-7.69698203e-01 -1.25689375e+00 -1.32271457e+00 -3.89455795e-01
-4.05225605e-01 3.75007719e-01 6.56218708e-01 5.86342037e-01
-2.47391477e-01 6.16500080e-02 7.07391977e-01 -1.32297111e+00
-4.18570012e-01 -1.03437793e+00 -1.32772052e+00 1.35683715e-01
4.37729031e-01 -6.41202450e-01 -9.15473878e-01 3.19936603e-01] | [13.04355525970459, -2.979276657104492] |
66d80137-8a9d-4324-b184-88855d9b9bb8 | end-to-end-sleep-staging-with-raw-single | 1904.10255 | null | http://arxiv.org/abs/1904.10255v1 | http://arxiv.org/pdf/1904.10255v1.pdf | End-to-end Sleep Staging with Raw Single Channel EEG using Deep Residual ConvNets | Humans approximately spend a third of their life sleeping, which makes
monitoring sleep an integral part of well-being. In this paper, a 34-layer deep
residual ConvNet architecture for end-to-end sleep staging is proposed. The
network takes raw single channel electroencephalogram (Fpz-Cz) signal as input
and yields hypnogram annotations for each 30s segments as output. Experiments
are carried out for two different scoring standards (5 and 6 stage
classification) on the expanded PhysioNet Sleep-EDF dataset, which contains
multi-source data from hospital and household polysomnography setups. The
performance of the proposed network is compared with that of the
state-of-the-art algorithms in patient independent validation tasks. The
experimental results demonstrate the superiority of the proposed network
compared to the best existing method, providing a relative improvement in
epoch-wise average accuracy of 6.8% and 6.3% on the household data and
multi-source data, respectively. Codes are made publicly available on Github. | ['Asif Shahriyar Sushmit', 'Taufiq Hasan', 'Ahmed Imtiaz Humayun', 'Mohammed Imamul Hassan Bhuiyan'] | 2019-04-23 | null | null | null | null | ['sleep-staging'] | ['medical'] | [-1.25458762e-02 -6.27553044e-03 -5.54593876e-02 -6.89658344e-01
-4.13977921e-01 3.30782123e-02 -2.30644703e-01 -3.84386592e-02
-8.80497992e-01 1.02662349e+00 3.19739819e-01 -1.08403787e-01
-2.90065426e-02 -7.59032518e-02 7.93824270e-02 -5.20691335e-01
-4.40239936e-01 1.91619724e-01 -1.78051502e-01 -6.56190887e-02
-5.72406268e-03 1.21104315e-01 -1.00658965e+00 1.09103628e-01
1.01850235e+00 1.38022065e+00 8.91528428e-02 5.66737592e-01
6.09603763e-01 6.22569978e-01 -8.98492396e-01 -2.65188783e-01
-9.65092182e-02 -4.59624380e-01 -6.89039469e-01 -1.04651660e-01
-1.00674786e-01 -2.44898617e-01 -4.10550833e-01 8.37157190e-01
1.09589636e+00 9.73984003e-02 1.44785903e-02 -1.11588264e+00
-4.00533617e-01 3.65624189e-01 -2.40712687e-02 9.92941380e-01
1.39658228e-01 7.50744641e-02 6.08044922e-01 -5.49420536e-01
-7.02387094e-02 3.06967437e-01 7.87780702e-01 7.85369158e-01
-9.91408885e-01 -1.13972127e+00 -3.77716035e-01 5.05514860e-01
-1.47959399e+00 -7.11570859e-01 5.60372591e-01 -2.84832809e-02
1.40948308e+00 6.71529621e-02 1.26146042e+00 1.24893057e+00
8.57897222e-01 4.64413285e-01 1.05943108e+00 8.75098854e-02
5.38200557e-01 -9.80953798e-02 5.13509154e-01 5.82596064e-01
3.07240307e-01 -2.79830813e-01 -9.32774305e-01 1.32157490e-01
3.34718436e-01 4.45924282e-01 -5.03043115e-01 2.73019850e-01
-1.26295030e+00 5.15457213e-01 6.74548864e-01 4.11676079e-01
-7.70917535e-01 -1.71765331e-02 4.41316873e-01 1.03152826e-01
7.26551712e-01 2.78274834e-01 -4.99613285e-01 -5.46481788e-01
-1.54512393e+00 -2.37416178e-01 7.85347402e-01 5.90416193e-01
3.33284549e-02 -2.67701061e-03 -1.88859746e-01 7.17283428e-01
1.08920336e-01 3.53262275e-01 1.13276649e+00 -7.27621198e-01
4.21657801e-01 6.94125175e-01 -2.17656177e-02 -3.56515586e-01
-1.21716678e+00 -7.61831641e-01 -1.07983446e+00 -1.53005749e-01
-2.34365270e-01 -2.73630857e-01 -8.17736924e-01 1.60553265e+00
-3.98091465e-01 3.40229273e-01 1.37604073e-01 8.35423470e-01
1.27354264e+00 -3.44905257e-02 -6.28139675e-02 -3.40774685e-01
1.34416175e+00 -1.12632704e+00 -9.25699174e-01 -4.98767018e-01
1.99325591e-01 3.67629007e-02 9.74166095e-01 5.82731605e-01
-1.28256500e+00 -4.92509544e-01 -1.26721692e+00 -5.59509769e-02
-1.28899843e-01 3.42306316e-01 4.42853719e-01 6.79944336e-01
-1.54589903e+00 8.57198775e-01 -1.53799629e+00 -5.79305291e-01
8.92929077e-01 9.80343044e-01 -4.61000353e-01 4.01263744e-01
-1.01077175e+00 8.30832601e-01 1.38401017e-01 2.17236966e-01
-1.00617695e+00 -5.87125719e-01 -6.20512962e-01 3.38305235e-01
-2.65148431e-01 -9.48221087e-01 1.23544240e+00 -6.58251524e-01
-1.40291786e+00 8.91777813e-01 -3.25267881e-01 -7.16537058e-01
2.21801430e-01 -2.82366276e-01 -7.34917819e-01 3.31184894e-01
2.40295917e-01 4.79149193e-01 5.54826975e-01 -1.69123203e-01
-2.72821456e-01 -7.65023172e-01 -8.13797861e-02 2.52615362e-01
-3.95700604e-01 1.26650453e-01 -2.37473994e-01 -2.80142784e-01
-2.79600155e-02 -8.62753391e-01 -1.85231686e-01 -3.41265589e-01
-5.62677622e-01 -4.58010621e-02 7.79558048e-02 -8.40442419e-01
1.38932145e+00 -2.09772992e+00 5.58223762e-02 -3.47404659e-01
7.00230658e-01 -5.98917380e-02 3.31524730e-01 1.14418313e-01
-4.15415436e-01 -1.32649451e-01 -3.82910550e-01 -1.13574708e+00
-2.74623454e-01 9.36948881e-02 4.09711629e-01 9.52983975e-01
-2.60083199e-01 9.00221765e-01 -7.87546575e-01 -7.63450265e-02
1.86218292e-01 5.70563436e-01 -2.89921403e-01 4.08622921e-01
8.78993452e-01 7.02027261e-01 4.31447104e-02 7.02432394e-01
2.54667342e-01 -5.75400174e-01 -1.88616946e-01 1.29503563e-01
7.65088573e-03 8.04783881e-01 -2.53961623e-01 2.42381430e+00
-3.65532458e-01 6.08439624e-01 -2.64317274e-01 -6.65026963e-01
5.74783862e-01 5.46251297e-01 4.25956130e-01 -7.09607601e-01
5.02926886e-01 1.12040065e-01 4.02628601e-01 -3.56997997e-01
1.36434466e-01 -3.24532092e-01 -3.82755622e-02 2.43294641e-01
3.04276466e-01 3.37635010e-01 1.38080984e-01 1.69367775e-01
1.70465028e+00 -4.97376889e-01 6.20307565e-01 -6.25180602e-01
3.96582216e-01 -5.82705855e-01 7.83373177e-01 4.23902899e-01
-5.99739611e-01 7.33165622e-01 3.81402850e-01 -4.62886900e-01
-5.22329152e-01 -1.14291859e+00 -2.35295042e-01 5.61724007e-01
-3.68122980e-02 -7.23977566e-01 -8.81588995e-01 -5.76139212e-01
-4.29394037e-01 6.82427526e-01 -8.78053784e-01 -4.71008301e-01
-1.32169664e-01 -1.09014452e+00 6.02391303e-01 9.69421268e-01
7.95843959e-01 -1.56021237e+00 -1.35279608e+00 1.64781928e-01
-2.16852099e-01 -1.05053627e+00 -2.38370135e-01 7.01853395e-01
-1.00892580e+00 -1.11085784e+00 -8.07883799e-01 -4.74783003e-01
4.70728755e-01 3.14937253e-03 1.18266141e+00 1.38110325e-01
-1.92522302e-01 -1.64224301e-02 -1.43397391e-01 -3.68433058e-01
3.10574681e-01 4.09806758e-01 5.21093607e-01 -5.65978847e-02
6.11302078e-01 -1.23427188e+00 -1.29153824e+00 1.86460502e-02
-4.44483548e-01 3.99813391e-02 5.47105491e-01 8.56127441e-01
4.47697848e-01 -2.97575742e-01 8.30231667e-01 -5.43081701e-01
8.62974048e-01 -6.80016160e-01 -7.92776942e-02 -1.65730089e-01
-1.05123329e+00 -4.82912630e-01 7.50532985e-01 7.43633881e-02
-4.96957213e-01 -1.64611310e-01 -3.25434268e-01 -5.01579821e-01
-2.77746290e-01 2.25978583e-01 1.16555877e-01 2.88653731e-01
4.89286661e-01 3.74881953e-01 -2.03701481e-01 -3.46123815e-01
-4.41777289e-01 7.23771691e-01 8.56095910e-01 3.06885928e-01
1.04780653e-02 4.19498861e-01 -3.63842100e-01 -6.97760165e-01
-9.35631633e-01 -8.03467512e-01 -6.99116230e-01 8.60731155e-02
1.26819754e+00 -1.21814203e+00 -7.00664699e-01 5.58757782e-01
-7.57991791e-01 -5.70368230e-01 -1.41255021e-01 5.98290503e-01
-5.49699068e-01 6.01724796e-02 -4.79616970e-01 -6.46712720e-01
-1.32810354e+00 -1.13566387e+00 8.82616758e-01 4.25091475e-01
-4.58002180e-01 -8.90356779e-01 1.99878842e-01 3.31265539e-01
5.29075086e-01 2.20590785e-01 3.23694080e-01 -8.90404522e-01
3.15320939e-01 -2.56946176e-01 1.79028839e-01 5.61315238e-01
2.69821167e-01 -8.65934074e-01 -1.17199433e+00 -6.18974149e-01
5.34599900e-01 -3.94458830e-01 7.81802237e-01 6.11144662e-01
1.43818879e+00 2.45912243e-02 -1.62299126e-01 1.10736108e+00
1.20533073e+00 2.18300194e-01 8.26751769e-01 3.50543737e-01
3.92393768e-01 -1.00511581e-01 -1.11463964e-01 6.01008892e-01
6.76657856e-01 8.83917809e-02 5.23658454e-01 -1.99144751e-01
6.21795729e-02 4.70499575e-01 2.93444186e-01 1.07118237e+00
-1.55577257e-01 -1.53633267e-01 -9.53614473e-01 5.82442999e-01
-1.62561417e+00 -7.20623374e-01 2.96725649e-02 2.18144298e+00
3.98069531e-01 2.63567537e-01 2.38259733e-01 3.70185494e-01
4.40934032e-01 1.61489859e-01 -1.01168907e+00 -1.69524118e-01
2.20021263e-01 8.37856174e-01 4.27849144e-01 -2.22142965e-01
-9.00744200e-01 3.48050654e-01 5.91522694e+00 5.80884889e-02
-1.14623952e+00 7.51206994e-01 6.86716616e-01 -9.46639240e-01
6.46482170e-01 -5.34880996e-01 -5.10307491e-01 1.01149523e+00
1.83601677e+00 -1.40975952e-01 7.94826388e-01 6.82383657e-01
6.02408648e-01 -3.16366166e-01 -1.21167946e+00 1.52112222e+00
2.87737459e-01 -1.03309655e+00 -1.04492176e+00 -1.04227871e-01
4.66048598e-01 8.16552103e-01 -1.60580575e-01 3.25984836e-01
-4.56700861e-01 -1.24714684e+00 7.17910409e-01 5.67909479e-01
1.22670877e+00 -8.63252282e-01 1.04160774e+00 3.41560841e-01
-9.93610859e-01 -1.84968024e-01 -1.13585316e-01 -2.91974634e-01
1.65533438e-01 3.59193712e-01 -6.01657331e-01 2.71070063e-01
1.27339375e+00 1.17140281e+00 -9.23573911e-01 1.20590651e+00
-5.30001462e-01 8.32072437e-01 -1.37846619e-01 -1.15282228e-02
-4.19375226e-02 1.12004489e-01 1.09447680e-01 1.03817034e+00
2.87985384e-01 2.25079983e-01 -3.27347666e-01 9.03277218e-01
-4.52270061e-01 -3.33184570e-01 -3.07827532e-01 1.12116061e-01
2.04464704e-01 1.55476165e+00 -8.84765089e-01 -2.75030345e-01
-2.86388606e-01 1.32435954e+00 3.79677892e-01 8.81298110e-02
-8.40024948e-01 -5.91900289e-01 4.34446305e-01 -1.50745034e-01
-9.21026245e-02 1.90028086e-01 -6.73063576e-01 -1.36794770e+00
1.56603470e-01 -6.61961675e-01 2.90135384e-01 -8.03025484e-01
-1.01916337e+00 1.06814170e+00 -2.19496191e-01 -9.91351843e-01
-1.68915894e-02 -1.89709008e-01 -1.05203176e+00 8.71049523e-01
-1.44113350e+00 -5.86890459e-01 -8.41297209e-01 8.94496441e-01
6.98587656e-01 -1.50951788e-01 1.11818004e+00 5.30094504e-01
-1.04230750e+00 7.78855443e-01 -7.58242011e-02 -4.37544435e-02
4.21190977e-01 -1.29810989e+00 6.21917963e-01 9.48552549e-01
-2.72396684e-01 9.12437856e-01 1.80409789e-01 -3.04537624e-01
-9.76159215e-01 -1.06288433e+00 9.63124990e-01 -4.54147846e-01
4.67213184e-01 -4.52426165e-01 -6.02782011e-01 6.02767169e-01
5.16157031e-01 3.76022458e-02 1.01111853e+00 5.31520285e-02
5.45707941e-01 -2.89559722e-01 -1.39153814e+00 9.27183181e-02
1.02679884e+00 -4.28148925e-01 -8.10347974e-01 2.09920734e-01
4.36346024e-01 -5.13493657e-01 -9.23734069e-01 3.42897117e-01
5.38725138e-01 -1.44183958e+00 4.66567367e-01 -2.17557773e-01
2.78288484e-01 1.13002740e-01 -1.04735466e-02 -1.53127086e+00
-2.88182020e-01 -7.99074173e-01 -2.45291173e-01 5.65618217e-01
3.20126534e-01 -7.41891801e-01 7.79091597e-01 6.04911864e-01
-8.49979520e-01 -1.17073596e+00 -1.11316741e+00 -4.81387496e-01
-3.81714076e-01 -4.68368471e-01 4.91720319e-01 4.24068481e-01
3.23497862e-01 8.33190143e-01 -2.68494487e-01 -7.80921280e-02
3.62053066e-01 -3.17996293e-01 2.00409904e-01 -1.25379145e+00
3.26713307e-05 -1.29800290e-01 -5.67960858e-01 -4.02425677e-01
2.13078171e-01 -1.03053534e+00 3.78975794e-02 -1.82128561e+00
4.68064338e-01 8.49919301e-03 -1.17222369e+00 8.19898963e-01
-3.56662758e-02 6.07272565e-01 -1.65984526e-01 7.99623504e-03
-9.47416604e-01 6.61225796e-01 5.78101277e-01 9.85191986e-02
-3.59984547e-01 2.00378001e-01 -7.93977141e-01 7.47731566e-01
1.17774880e+00 -5.85389256e-01 -6.14202499e-01 -3.85771394e-01
-1.01382881e-01 2.07692787e-01 2.64875859e-01 -1.77853405e+00
2.14093775e-01 6.79091454e-01 8.90020072e-01 -5.87515771e-01
6.86494589e-01 -5.17302215e-01 3.45817581e-02 6.86725676e-01
-2.23328069e-01 5.34661114e-01 1.56252816e-01 4.00296658e-01
4.52348553e-02 1.44651502e-01 9.09917295e-01 -2.13804051e-01
-3.48010242e-01 5.25735438e-01 -2.55090386e-01 1.52981937e-01
7.19525397e-01 -1.69487670e-01 -3.70630354e-01 -1.91464916e-01
-9.25624430e-01 1.19619481e-01 3.58904541e-01 2.62665302e-01
9.37847972e-01 -1.09970140e+00 -3.43081564e-01 4.64784831e-01
1.15298204e-01 -2.40653753e-01 3.21811616e-01 1.66142023e+00
-3.67372364e-01 4.95284408e-01 -5.31611145e-01 -4.58473891e-01
-1.02853680e+00 1.54832378e-01 5.30982375e-01 -2.14287207e-01
-1.02675521e+00 1.07127023e+00 1.41675619e-03 1.74190760e-01
4.47550088e-01 -7.79867530e-01 -2.73757666e-01 -2.53742784e-01
7.78695822e-01 4.56673801e-01 6.84312284e-01 -3.71349782e-01
-7.65114844e-01 -2.45539859e-01 7.45099410e-02 1.69366315e-01
1.80236757e+00 -1.79273437e-03 -2.07204506e-01 6.71229482e-01
1.15868866e+00 -4.68966693e-01 -9.63906229e-01 4.30289835e-01
-2.63017327e-01 9.29432735e-02 3.50225598e-01 -1.09658027e+00
-1.30853248e+00 1.02387726e+00 1.29746985e+00 3.48150469e-02
1.54325557e+00 -1.95244715e-01 1.12376547e+00 2.82338679e-01
4.93450582e-01 -7.39967823e-01 -1.20961897e-01 1.19162120e-01
6.31105423e-01 -1.16563749e+00 -3.51755805e-02 4.81198549e-01
-7.30173767e-01 9.40907776e-01 5.89667022e-01 -5.22954047e-01
5.76516390e-01 -5.14936596e-02 -1.34063408e-01 -6.14548504e-01
-7.09598422e-01 9.09170285e-02 2.77521759e-01 2.98965454e-01
3.99762630e-01 2.25294366e-01 -4.77601111e-01 1.34454358e+00
-5.66361845e-01 2.70646095e-01 4.18619633e-01 7.64902949e-01
-1.31409332e-01 -5.43163896e-01 2.86431789e-01 1.07091773e+00
-1.12972999e+00 -4.70725089e-01 -1.56094655e-01 5.21406651e-01
1.37351170e-01 1.33074832e+00 6.77471608e-02 -6.03840828e-01
3.84907126e-01 2.06429735e-01 2.40051106e-01 -7.97447562e-01
-9.80490506e-01 -1.83880419e-01 6.97069094e-02 -9.26675797e-01
-4.74517494e-01 -5.06032705e-01 -1.51891637e+00 -5.44223189e-02
3.45552899e-02 1.56849578e-01 6.49767339e-01 9.45686579e-01
6.17078841e-01 1.04519176e+00 3.91957521e-01 -9.99722004e-01
-1.92694142e-01 -1.42062008e+00 -7.52418935e-01 1.05539985e-01
6.42949104e-01 -7.19732046e-01 -3.95845979e-01 -1.84576705e-01] | [13.504759788513184, 3.522120714187622] |
2ec69317-bd84-40ad-8ef8-3a59c833b63c | automatic-method-of-domain-ontology | 1405.1346 | null | http://arxiv.org/abs/1405.1346v1 | http://arxiv.org/pdf/1405.1346v1.pdf | Automatic Method Of Domain Ontology Construction based on Characteristics of Corpora POS-Analysis | It is now widely recognized that ontologies, are one of the fundamental
cornerstones of knowledge-based systems. What is lacking, however, is a
currently accepted strategy of how to build ontology; what kinds of the
resources and techniques are indispensables to optimize the expenses and the
time on the one hand and the amplitude, the completeness, the robustness of en
ontology on the other hand. The paper offers a semi-automatic ontology
construction method from text corpora in the domain of radiological protection.
This method is composed from next steps: 1) text annotation with part-of-speech
tags; 2) revelation of the significant linguistic structures and forming the
templates; 3) search of text fragments corresponding to these templates; 4)
basic ontology instantiation process | ['Olena Orobinska'] | 2014-05-06 | null | null | null | null | ['text-annotation'] | ['natural-language-processing'] | [ 3.75864178e-01 3.76321286e-01 1.59116954e-01 2.52276734e-02
-2.64130175e-01 -3.28855187e-01 7.09963083e-01 5.86504996e-01
-4.46802467e-01 7.53263772e-01 3.18628103e-01 -3.91460508e-01
-9.63287532e-01 -9.72228467e-01 -2.22937297e-02 -5.49584150e-01
1.25447541e-01 7.88340330e-01 4.17127192e-01 -3.83737415e-01
2.72107512e-01 6.27239287e-01 -1.93300581e+00 1.03925772e-01
1.00964987e+00 8.93978655e-01 3.38180453e-01 2.95942992e-01
-8.26341629e-01 7.38099873e-01 -3.23876083e-01 -1.63648248e-01
-1.32422736e-02 -3.75101060e-01 -1.17079198e+00 1.70571938e-01
-6.60175681e-01 1.03092879e-01 1.20605975e-01 1.16346431e+00
3.04653615e-01 -4.08542249e-03 7.80304372e-01 -6.38755977e-01
-4.18950200e-01 6.09439552e-01 1.35178432e-01 8.50155577e-02
4.99110729e-01 -1.90695092e-01 6.25339746e-01 -4.25289571e-01
8.61366808e-01 9.88314390e-01 2.09400207e-01 1.87963665e-01
-4.46567178e-01 -9.43189487e-02 -3.04762572e-01 2.32911199e-01
-1.38985491e+00 -2.28643879e-01 4.81948763e-01 -7.33178079e-01
6.82154119e-01 2.16520995e-01 6.14638388e-01 6.45683408e-01
2.81494737e-01 2.43621469e-02 1.08754134e+00 -9.98775303e-01
2.28796184e-01 4.73848224e-01 3.25476825e-01 6.79063976e-01
2.93181121e-01 -1.33392870e-01 -3.21190894e-01 -3.03681437e-02
2.99423963e-01 -2.23334968e-01 -2.82768816e-01 -1.79078490e-01
-7.57879436e-01 6.85013831e-01 -4.24644411e-01 1.28504884e+00
-6.29066050e-01 -5.96949875e-01 5.72370887e-01 7.04434365e-02
9.45180133e-02 2.31971025e-01 -3.39320540e-01 -1.26353070e-01
-5.39430320e-01 9.18578207e-02 8.41767311e-01 1.12884486e+00
5.48753500e-01 -2.20467836e-01 3.73037070e-01 4.97201830e-01
3.03051740e-01 2.59004056e-01 7.35541105e-01 -3.35174173e-01
1.68131590e-01 1.15632868e+00 1.03604369e-01 -7.67197430e-01
-3.39698881e-01 -2.83416659e-01 -1.45420864e-01 -4.46504587e-03
4.76160794e-01 -8.75063613e-02 -5.63326120e-01 1.30734360e+00
4.73375052e-01 -4.78975624e-01 3.35919827e-01 2.61980891e-01
9.87610877e-01 4.96948749e-01 3.20786953e-01 -4.24514502e-01
2.02935958e+00 -2.89240479e-01 -1.11077082e+00 1.50402293e-01
5.94622910e-01 -1.08633530e+00 6.76087558e-01 3.63535196e-01
-8.10706913e-01 -2.87100673e-01 -1.03184688e+00 -1.99985020e-02
-9.88848865e-01 1.60742164e-01 4.80216026e-01 8.23729455e-01
-5.88566720e-01 4.27839905e-01 -4.46446687e-01 -8.24732482e-01
-1.28452346e-01 2.20117390e-01 -5.09186327e-01 2.12406456e-01
-1.22193253e+00 1.25224876e+00 1.10275269e+00 1.21887028e-01
-4.60222691e-01 -1.83736578e-01 -5.78692436e-01 2.61156112e-01
7.01185405e-01 -3.96050721e-01 7.16487765e-01 -7.96020687e-01
-1.17639303e+00 1.00863767e+00 3.64122465e-02 -2.14029640e-01
4.38325405e-01 -4.20504138e-02 -8.81727934e-01 2.04402104e-01
1.69854254e-01 -1.68995280e-02 1.45258918e-01 -1.07405818e+00
-1.00514746e+00 -5.82880199e-01 -1.17060639e-01 1.63866207e-01
-4.03749913e-01 2.43456542e-01 -2.32010499e-01 -1.24417432e-01
2.78081745e-01 -5.37355065e-01 5.74916974e-02 -2.24230200e-01
-1.17973238e-01 -5.19411206e-01 6.17325127e-01 -1.07770991e+00
1.32233810e+00 -1.96923864e+00 9.99445692e-02 4.66067493e-01
-1.55839995e-01 3.55395377e-01 3.40884238e-01 8.76483738e-01
-1.09224841e-01 -9.65467691e-02 -2.90619701e-01 5.53768396e-01
2.19103083e-01 3.55210483e-01 -2.37847164e-01 1.47692472e-01
-2.53912747e-01 4.55994867e-02 -8.08700442e-01 -8.63654554e-01
4.42225844e-01 3.60686988e-01 3.40069719e-02 -1.04894966e-01
-2.70343721e-01 3.11877280e-01 -1.02294588e+00 5.55585027e-01
1.57705858e-01 1.79913506e-01 5.48953116e-01 -2.51240224e-01
-6.28677785e-01 2.49380007e-01 -1.36201954e+00 1.55907142e+00
2.10777596e-02 9.60145742e-02 -1.81907952e-01 -1.03169334e+00
1.15414548e+00 8.89918387e-01 8.53320599e-01 -6.86896384e-01
5.59637189e-01 6.32534504e-01 -2.19976142e-01 -1.12939072e+00
4.35904652e-01 -2.52136141e-01 5.94901480e-02 1.35457858e-01
9.33116749e-02 4.71406952e-02 5.54476619e-01 -2.05482885e-01
6.67549193e-01 4.01894659e-01 1.03114760e+00 -4.38696295e-01
1.04970670e+00 4.40105200e-02 3.84874225e-01 5.79075189e-03
-3.84394638e-03 -2.62668610e-01 4.60142404e-01 -5.53040445e-01
-1.03845859e+00 -6.43451035e-01 -5.76479197e-01 4.82939541e-01
-1.95251614e-01 -5.96164316e-02 -9.73306894e-01 -3.52755994e-01
-1.63559586e-01 7.05113649e-01 -3.55110466e-01 4.79246229e-01
-3.20205569e-01 -6.64447367e-01 4.98099834e-01 -2.34249219e-01
3.94632310e-01 -1.29486144e+00 -7.87007570e-01 3.86111915e-01
-3.69810015e-01 -1.14329839e+00 3.34822387e-01 1.78399712e-01
-9.20414209e-01 -1.44793332e+00 -2.23984838e-01 -6.83389187e-01
4.94703382e-01 -2.50747174e-01 5.90588331e-01 4.57271725e-01
-3.76350760e-01 2.70984713e-02 -5.56861699e-01 -8.82269621e-01
-8.08771849e-01 6.73983097e-02 -2.38471627e-01 -2.25325361e-01
5.40055096e-01 -6.65618539e-01 -4.14762311e-02 1.82876006e-01
-1.30599916e+00 -3.52200478e-01 7.67277718e-01 2.33406320e-01
5.29170096e-01 5.20652533e-01 5.69455981e-01 -8.60440195e-01
4.82144177e-01 -3.60017508e-01 -7.50906348e-01 5.49256384e-01
-6.61074221e-01 -3.95353399e-02 4.38915819e-01 2.75527030e-01
-1.27055442e+00 -8.22091773e-02 -5.41059256e-01 6.92742109e-01
-4.88862514e-01 6.35314405e-01 -5.36724269e-01 8.92680138e-02
7.26447165e-01 6.47807121e-01 3.77035886e-02 -6.48846745e-01
2.04607561e-01 1.01898348e+00 5.87325692e-01 -7.43361712e-01
5.71364522e-01 5.30571222e-01 1.01276256e-01 -1.19378924e+00
-4.93801326e-01 -7.19903350e-01 -7.97991335e-01 -2.55314529e-01
1.19718850e+00 -3.79295081e-01 -7.11526811e-01 7.59100690e-02
-1.02686346e+00 2.53306508e-01 -1.87556505e-01 4.57479864e-01
-4.59387064e-01 7.81064749e-01 -1.28017783e-01 -1.02217686e+00
-5.80498219e-01 -9.33099031e-01 5.58471501e-01 1.28971741e-01
-1.96404502e-01 -8.07672620e-01 -1.52069166e-01 5.83567917e-01
-3.87672633e-02 3.76561075e-01 1.35709333e+00 -8.02934170e-01
-4.09717292e-01 -4.00976628e-01 6.99138492e-02 3.47550124e-01
2.51651675e-01 1.23758782e-02 -7.83122480e-01 2.80722588e-01
2.76011705e-01 1.84367776e-01 2.94868788e-03 3.46684866e-02
6.95779324e-01 -9.27019864e-02 -4.06789660e-01 6.05596863e-02
1.63746369e+00 9.28877175e-01 1.14157104e+00 6.25340164e-01
-3.05366330e-02 9.83109534e-01 7.35588133e-01 3.07392210e-01
1.97647195e-02 7.62728691e-01 3.08955759e-01 4.44119215e-01
-1.46426586e-02 -1.61539689e-01 4.42625619e-02 9.49465513e-01
-3.98745090e-01 -1.77320436e-01 -1.00723958e+00 6.45454824e-01
-1.68173039e+00 -1.25859702e+00 -1.61889523e-01 2.10371327e+00
4.39887136e-01 3.57325673e-01 -7.61973187e-02 7.02451169e-01
6.24592483e-01 -1.97435290e-01 2.93936044e-01 -4.78187576e-02
-1.70697242e-01 2.38725483e-01 2.35540748e-01 5.75992882e-01
-8.83705020e-01 5.60860574e-01 5.49989557e+00 6.94225252e-01
-8.17026138e-01 1.35904700e-01 -3.00093830e-01 5.32115400e-01
-3.36897224e-01 2.48995617e-01 -7.40822673e-01 5.39765835e-01
6.72868550e-01 -2.30394751e-01 1.39086768e-01 6.88597798e-01
2.58631706e-01 -3.25938612e-01 -6.05742574e-01 3.44073087e-01
1.90626010e-01 -1.14114821e+00 1.76444575e-01 3.07937890e-01
2.23538533e-01 -3.90092820e-01 -7.50540853e-01 -6.07665675e-03
1.67270944e-01 -5.85717142e-01 9.83139038e-01 7.51234114e-01
2.52826691e-01 -5.78429520e-01 1.09434545e+00 4.13757950e-01
-1.11709857e+00 -2.28208080e-01 -3.71985793e-01 2.82296300e-01
4.54583347e-01 6.28925979e-01 -6.82143629e-01 1.42144597e+00
4.17593390e-01 1.08922936e-01 -2.00915486e-01 1.02525139e+00
-4.06833529e-01 2.18074694e-01 -2.30920747e-01 -2.54635066e-01
1.94193214e-01 -5.86556792e-01 6.20091200e-01 9.79596555e-01
4.89206940e-01 2.03681946e-01 -4.99836206e-02 5.84270477e-01
3.93222958e-01 8.05548012e-01 -6.01630807e-01 -2.29784116e-01
5.72644234e-01 1.03618908e+00 -9.67376888e-01 -3.49731266e-01
-3.01778972e-01 3.70275319e-01 -2.47692615e-01 -5.80694713e-02
-5.20195544e-01 -6.06413782e-01 6.82690814e-02 2.21641496e-01
2.01724261e-01 3.16731855e-02 -2.51302183e-01 -6.19516015e-01
1.90185368e-01 -9.25122082e-01 7.72065639e-01 -5.99979043e-01
-7.15382516e-01 5.15216053e-01 8.79389122e-02 -1.08686638e+00
-5.36444224e-02 -7.27720618e-01 4.36902046e-03 1.00785422e+00
-1.11746001e+00 -1.27990544e+00 -2.59860784e-01 3.98880780e-01
4.10826176e-01 -1.44093275e-01 1.32771480e+00 6.82401776e-01
-2.78033763e-01 -2.62340248e-01 8.03892165e-02 1.03978306e-01
1.65255815e-01 -7.96273887e-01 -5.11636972e-01 8.54508877e-01
-1.54677793e-01 5.90380788e-01 9.43818569e-01 -8.27241182e-01
-8.89402449e-01 -1.82754666e-01 1.62528467e+00 -7.91160315e-02
7.86712050e-01 8.32050368e-02 -7.71608770e-01 4.73653018e-01
4.04876880e-02 -7.94248462e-01 7.56994486e-01 -4.33377735e-03
9.76408571e-02 -1.73948988e-01 -1.18722999e+00 5.36794662e-01
7.77778089e-01 -5.58975041e-01 -1.08458662e+00 5.68325937e-01
2.66233712e-01 -2.31259406e-01 -1.26345599e+00 -4.64508832e-02
6.95909977e-01 -8.32108915e-01 8.36000085e-01 -4.70112979e-01
1.45763114e-01 -7.26821005e-01 -9.41714272e-02 -5.24065375e-01
3.91871482e-02 -4.44408685e-01 7.28776395e-01 1.46279120e+00
4.27304387e-01 -6.85574889e-01 4.37910467e-01 1.85791984e-01
-5.57294905e-01 -3.55574280e-01 -1.30775297e+00 -6.66802824e-01
-3.10440272e-01 -2.77866960e-01 7.42712200e-01 9.73000467e-01
4.25331622e-01 1.02881938e-01 1.31593913e-01 1.65197715e-01
4.99653518e-01 -2.65062869e-01 4.37072843e-01 -1.64606953e+00
-5.86304665e-02 -4.66034442e-01 -6.03200197e-01 -1.21254995e-01
-1.46669239e-01 -6.04055166e-01 -2.12280318e-01 -1.90071559e+00
-3.60566564e-02 -3.33025008e-01 6.06513843e-02 3.93169671e-01
3.11890870e-01 -6.40348256e-01 -9.33167785e-02 3.07418197e-01
-4.46379520e-02 1.13397881e-01 9.78144944e-01 1.75409824e-01
5.01003005e-02 -2.54772276e-01 -5.64815998e-01 9.33418989e-01
3.72713566e-01 -6.75793886e-01 -5.42101026e-01 -1.05677910e-01
3.19153160e-01 2.16300666e-01 -8.57554972e-02 -1.05919993e+00
1.44785911e-01 -3.21272641e-01 -9.98602360e-02 -6.20781660e-01
1.82936750e-02 -1.22395730e+00 7.63544321e-01 7.70425975e-01
3.96589339e-02 7.92320166e-03 -2.56348252e-02 3.60842079e-01
-1.04437999e-01 -1.09986985e+00 7.47962892e-01 -3.90861362e-01
-8.79626334e-01 -9.42690894e-02 -5.00256479e-01 -1.28050759e-01
1.15518630e+00 -4.47871357e-01 -2.52230823e-01 2.06914499e-01
-8.41000319e-01 -2.61167735e-01 2.65591562e-01 1.60033271e-01
-8.41867924e-02 -7.83510208e-01 -2.44327396e-01 -1.69981048e-01
1.30970195e-01 -2.47510701e-01 4.60991174e-01 5.90021968e-01
-9.75849211e-01 7.85100222e-01 -6.01131797e-01 -2.25165803e-02
-1.37859416e+00 7.79413462e-01 4.28856701e-01 -6.25436366e-01
-6.33377552e-01 9.68206581e-03 -1.31992295e-01 -7.04233423e-02
1.26929700e-01 -5.68077341e-02 -1.02894425e+00 1.91301703e-01
4.51990157e-01 5.13296843e-01 3.48847568e-01 -9.16162014e-01
-2.87393838e-01 5.69717407e-01 3.18620026e-01 -3.07111919e-01
1.29845071e+00 -1.65136293e-01 -4.26168323e-01 1.56698391e-01
4.33023781e-01 1.52935550e-01 -2.59332806e-01 -3.12896855e-02
5.73848069e-01 -1.73497424e-01 -1.77893326e-01 -9.00865257e-01
-4.99099821e-01 4.75967854e-01 5.79199553e-01 4.90743816e-01
1.02944076e+00 -2.31026530e-01 4.87756550e-01 2.96007305e-01
5.93326032e-01 -1.42190790e+00 -4.14318055e-01 3.00275683e-01
7.02216268e-01 -3.74458224e-01 8.03024769e-02 -1.03914201e+00
-2.09381059e-01 1.42624104e+00 -5.31936325e-02 6.09308362e-01
6.47441804e-01 2.45380834e-01 -5.49141690e-02 -6.43096626e-01
-2.29291946e-01 -4.64978725e-01 2.10953906e-01 6.31523371e-01
5.75123668e-01 -9.10365656e-02 -1.30366969e+00 6.03759170e-01
-1.99585095e-01 4.62121129e-01 3.45293194e-01 1.10770762e+00
-1.09337437e+00 -1.47823870e+00 -6.12516403e-01 6.25708997e-02
-7.44391203e-01 1.19482130e-01 -2.30202109e-01 1.12001312e+00
8.73121738e-01 1.03276122e+00 -4.11615461e-01 1.93680599e-01
6.79766297e-01 5.24742842e-01 3.41944844e-01 -5.06696463e-01
-4.98670459e-01 -4.34057601e-02 6.86032116e-01 2.83983350e-02
-5.23565173e-01 -8.67439151e-01 -1.32884634e+00 -6.33638650e-02
-3.31847370e-01 6.40690506e-01 1.01490629e+00 1.46020043e+00
-1.69160381e-01 4.90212202e-01 9.43398569e-03 -1.56845094e-03
-6.05475605e-01 -9.49608386e-01 -7.46796787e-01 6.34509802e-01
-5.20050466e-01 -7.52978623e-01 -1.88069537e-01 1.52702004e-01] | [9.339373588562012, 8.544754981994629] |
4590235a-ee7f-4b4c-b5a0-f72b40f103d7 | visualhints-a-visual-lingual-environment-for | 2010.13839 | null | https://arxiv.org/abs/2010.13839v1 | https://arxiv.org/pdf/2010.13839v1.pdf | VisualHints: A Visual-Lingual Environment for Multimodal Reinforcement Learning | We present VisualHints, a novel environment for multimodal reinforcement learning (RL) involving text-based interactions along with visual hints (obtained from the environment). Real-life problems often demand that agents interact with the environment using both natural language information and visual perception towards solving a goal. However, most traditional RL environments either solve pure vision-based tasks like Atari games or video-based robotic manipulation; or entirely use natural language as a mode of interaction, like Text-based games and dialog systems. In this work, we aim to bridge this gap and unify these two approaches in a single environment for multimodal RL. We introduce an extension of the TextWorld cooking environment with the addition of visual clues interspersed throughout the environment. The goal is to force an RL agent to use both text and visual features to predict natural language action commands for solving the final task of cooking a meal. We enable variations and difficulties in our environment to emulate various interactive real-world scenarios. We present a baseline multimodal agent for solving such problems using CNN-based feature extraction from visual hints and LSTMs for textual feature extraction. We believe that our proposed visual-lingual environment will facilitate novel problem settings for the RL community. | ['Michiaki Tatsubori', 'Kartik Talamadupula', 'Subhajit Chaudhury', 'Thomas Carta'] | 2020-10-26 | null | null | null | null | ['text-based-games'] | ['playing-games'] | [-1.08767003e-01 3.00172940e-02 2.07631420e-02 -3.24617207e-01
-5.35130858e-01 -9.85851824e-01 8.17142248e-01 8.77062380e-02
-7.30173230e-01 5.01700163e-01 2.74499178e-01 -4.18318331e-01
3.15835536e-01 -6.16643012e-01 -5.74985862e-01 -3.86563450e-01
-9.64493230e-02 4.49930191e-01 1.49800092e-01 -8.50553930e-01
3.39744955e-01 3.14626545e-01 -1.47269797e+00 9.12748337e-01
3.80826145e-01 5.39949596e-01 7.66530097e-01 1.20425057e+00
-3.98533911e-01 1.66268170e+00 -4.91190314e-01 5.23500890e-03
4.04915392e-01 -4.27037299e-01 -1.01311314e+00 2.03277588e-01
1.24350086e-01 -6.50199652e-01 -3.88443172e-01 6.75465286e-01
3.41739863e-01 5.81053674e-01 4.77755576e-01 -1.50147080e+00
-4.97485757e-01 6.75979495e-01 -1.79237112e-01 -1.84050739e-01
1.07148826e+00 8.61279428e-01 8.20179701e-01 -5.36573350e-01
7.29103208e-01 1.45084834e+00 9.93807614e-02 8.57461512e-01
-9.20347631e-01 -1.47487402e-01 2.74849057e-01 4.67103004e-01
-8.36854756e-01 -5.11731863e-01 7.77681649e-01 -1.64543837e-01
1.49774694e+00 2.64733374e-01 6.52438700e-01 1.41358602e+00
-8.89821127e-02 1.26543021e+00 9.92835283e-01 -8.10405672e-01
1.48697019e-01 2.18836948e-01 -3.21403891e-01 1.17545176e+00
-6.37142718e-01 3.61124337e-01 -7.98656523e-01 1.12156488e-01
8.56046855e-01 -6.73486665e-02 -2.61646211e-01 -7.87171423e-01
-1.53122902e+00 9.80704665e-01 4.74297106e-01 3.05248797e-01
-3.26394439e-01 1.45470321e-01 6.21543825e-01 6.93836749e-01
-2.61392981e-01 6.79089487e-01 -3.56637776e-01 -3.81343096e-01
-3.67087662e-01 4.29436386e-01 8.67023766e-01 9.58231926e-01
6.90813363e-01 1.85988516e-01 -3.47674102e-01 8.54671478e-01
6.85333252e-01 4.96986777e-01 4.15216178e-01 -1.32379150e+00
5.76698184e-01 7.08219290e-01 3.30289572e-01 -7.54491031e-01
-6.88370943e-01 5.45300722e-01 -9.97369289e-02 8.28627169e-01
8.28401268e-01 -2.60486186e-01 -8.38786960e-01 1.48172748e+00
3.70093465e-01 -4.15736407e-01 4.34973806e-01 1.13405168e+00
1.30358684e+00 7.40156591e-01 3.08574289e-01 1.12275526e-01
1.36059082e+00 -1.28429806e+00 -6.99836016e-01 -3.65939975e-01
9.19624925e-01 -5.91468990e-01 1.65133703e+00 5.33551693e-01
-1.07283580e+00 -4.77508992e-01 -9.91527081e-01 -2.66194612e-01
-5.95494926e-01 2.60287505e-02 6.10304475e-01 2.48827979e-01
-1.27250469e+00 2.12096795e-01 -7.54042149e-01 -7.85105109e-01
-2.13065609e-01 3.29579473e-01 -4.57072526e-01 -9.27468464e-02
-8.78597975e-01 1.21054530e+00 4.50655848e-01 1.11529253e-01
-1.09104133e+00 3.06177028e-02 -1.56251049e+00 -3.00337613e-01
7.72254586e-01 -5.37486255e-01 1.75506222e+00 -1.09509146e+00
-2.02927589e+00 9.75175261e-01 5.01094125e-02 -2.17100441e-01
5.34037232e-01 -1.32594168e-01 1.10316470e-01 2.34066173e-01
-9.50328335e-02 1.07723725e+00 5.02896965e-01 -1.56847453e+00
-6.87537134e-01 -2.97372222e-01 7.31659889e-01 7.34211326e-01
2.39920333e-01 2.25444753e-02 -4.10936624e-01 -1.58810973e-01
-4.42668289e-01 -7.77688384e-01 -3.53406131e-01 -5.93750551e-02
-4.02595222e-01 -3.73373181e-01 8.90091419e-01 -3.36759001e-01
4.16190565e-01 -1.96184754e+00 3.00512493e-01 -6.65800869e-02
2.67652065e-01 1.70677677e-01 -5.88543713e-01 9.05144930e-01
3.78581643e-01 -3.53119701e-01 3.18652689e-01 -2.77032137e-01
4.70276743e-01 5.03410161e-01 -3.13287191e-02 1.43408209e-01
9.12933052e-02 1.22446179e+00 -1.08286381e+00 -6.33549631e-01
9.08148944e-01 2.42720410e-01 -3.73934776e-01 7.32319832e-01
-7.70332158e-01 5.23720443e-01 -4.50592399e-01 4.91614491e-01
1.03488959e-01 4.23354609e-03 4.10881102e-01 -1.83821842e-02
-3.87759924e-01 1.48498848e-01 -9.63135362e-01 2.19508123e+00
-5.60887873e-01 6.78117275e-01 3.44546765e-01 -7.84904957e-01
6.83572233e-01 2.77794063e-01 1.59050286e-01 -1.01241231e+00
2.25240111e-01 -2.55803525e-01 -8.42959210e-02 -1.03071475e+00
6.55919611e-01 1.19414896e-01 -2.18979537e-01 6.14757717e-01
1.64397016e-01 -4.55275923e-01 3.35284591e-01 4.00563955e-01
9.06633377e-01 8.44112992e-01 4.26151425e-01 3.09992403e-01
2.62187690e-01 5.33027530e-01 -1.91028610e-01 1.10356903e+00
-4.71274555e-01 7.65051469e-02 2.74768025e-01 -7.35081911e-01
-9.54876125e-01 -9.80891645e-01 7.93086112e-01 1.82613039e+00
3.85666154e-02 -4.25433636e-01 -4.52085614e-01 -8.95708501e-01
-4.88527238e-01 8.76271248e-01 -4.43018258e-01 2.02954799e-01
-6.50697529e-01 -3.41365151e-02 4.29993153e-01 3.44560534e-01
4.97797728e-01 -1.98634565e+00 -1.35821116e+00 5.83578497e-02
-2.40130678e-01 -1.23085129e+00 -2.24221975e-01 5.52178383e-01
-3.29954118e-01 -9.74461257e-01 -4.43861693e-01 -1.03833818e+00
2.33558744e-01 2.84942806e-01 1.28814411e+00 1.52495027e-01
-2.29500711e-01 1.06795466e+00 -8.55612576e-01 -2.02305168e-01
-6.74588323e-01 -3.61162007e-01 -2.60181189e-01 -5.91579199e-01
4.42941695e-01 -8.75931680e-02 -3.15671176e-01 -5.50111569e-02
-7.19793677e-01 5.25506496e-01 3.53503019e-01 7.24658191e-01
-1.28066614e-01 -5.21579802e-01 -7.80849308e-02 -3.19028944e-01
1.02381408e+00 -1.20533682e-01 -4.50692415e-01 3.71496707e-01
1.90522566e-01 1.94679841e-01 5.60389757e-01 -5.87303579e-01
-9.13019001e-01 3.82552028e-01 -2.68011913e-02 -3.30430061e-01
-7.87333369e-01 4.53538656e-01 1.02590071e-02 -2.98361722e-02
7.76669919e-01 3.43722075e-01 -4.87766303e-02 1.49870187e-01
7.92085767e-01 5.80505490e-01 5.90436816e-01 -8.70803058e-01
5.47897518e-01 1.39775142e-01 -3.26901942e-01 -9.03008401e-01
-1.73050731e-01 -3.48948151e-01 -6.25739694e-01 -5.87782264e-01
1.03962791e+00 -6.06184006e-01 -1.57601774e+00 9.86087844e-02
-1.29444599e+00 -1.24903286e+00 -1.27453238e-01 4.08854067e-01
-9.84570503e-01 4.18062001e-01 -7.58931994e-01 -1.07675529e+00
-5.41604497e-02 -1.40176690e+00 1.09796345e+00 2.45969400e-01
-3.95654172e-01 -9.43365693e-01 1.77863821e-01 3.55251461e-01
2.15071216e-01 2.67986208e-01 8.18342686e-01 -5.60945570e-01
-3.25397402e-01 4.28584248e-01 -2.78328210e-01 -3.02821398e-01
5.39581887e-02 -2.02411383e-01 -9.02740717e-01 -7.74871260e-02
-2.32497826e-01 -1.21930075e+00 4.26744521e-01 1.53925925e-01
5.75185657e-01 -2.00530827e-01 -5.02931550e-02 2.61138231e-01
1.11270845e+00 6.73763871e-01 3.76773983e-01 7.25944340e-01
6.80312037e-01 7.67954588e-01 7.10971594e-01 4.83592838e-01
8.03492725e-01 8.94577205e-01 6.86365783e-01 -3.94473493e-01
1.94426745e-01 -2.25020319e-01 7.28334725e-01 2.71031111e-01
-1.59555942e-01 -3.10527027e-01 -9.73605037e-01 2.16052607e-01
-2.08786345e+00 -1.08338952e+00 8.51974934e-02 1.79388118e+00
7.85094142e-01 -2.86733776e-01 3.96578103e-01 -2.32740924e-01
1.15547724e-01 1.21834643e-01 -4.37702924e-01 -9.04818177e-01
2.59874195e-01 -7.66239315e-02 6.75326362e-02 6.42283797e-01
-1.14487076e+00 1.40239906e+00 6.15244102e+00 2.44611770e-01
-1.16129959e+00 -2.74040550e-01 2.04171389e-01 -5.17169237e-02
1.83902457e-01 -4.66951460e-01 -2.95898646e-01 -3.09416980e-01
5.76472700e-01 5.67361951e-01 1.00784016e+00 6.91302598e-01
4.27792281e-01 -3.99973273e-01 -1.36102176e+00 1.15758026e+00
2.49669984e-01 -1.01896083e+00 -1.00705788e-01 -2.69752741e-01
-8.01967159e-02 2.36257851e-01 6.65451214e-02 8.05626869e-01
1.06179595e+00 -1.27886295e+00 7.32131302e-01 1.19229026e-01
5.22570074e-01 -6.20723486e-01 3.23542476e-01 5.74807525e-01
-1.12880075e+00 -1.66904584e-01 -2.82403044e-02 -4.07855004e-01
1.39819026e-01 -8.01503897e-01 -1.33603561e+00 2.98193872e-01
8.11589420e-01 4.27664757e-01 -3.31831396e-01 6.11992180e-01
-2.78453261e-01 -1.54849380e-01 -7.38105848e-02 -5.84687829e-01
6.97544575e-01 -1.75188944e-01 3.82441968e-01 1.30349278e+00
-2.40656380e-02 1.29372895e-01 8.33099544e-01 7.67719924e-01
4.14607286e-01 2.50096858e-01 -1.09877944e+00 -2.31094971e-01
-1.21757127e-02 1.22202933e+00 -7.51985550e-01 -2.33301878e-01
-7.45447934e-01 1.26433682e+00 6.80082858e-01 6.82247818e-01
-7.24906802e-01 -2.92015910e-01 4.95088279e-01 -4.27725166e-01
2.02095315e-01 -6.18015051e-01 2.90819705e-01 -1.21076465e+00
-4.20384198e-01 -1.50938165e+00 3.52389276e-01 -1.28197110e+00
-1.00315130e+00 7.11717665e-01 4.18960974e-02 -9.41074014e-01
-7.34870672e-01 -1.05846846e+00 -4.62157756e-01 4.64621156e-01
-1.34080470e+00 -1.38583362e+00 -3.52000505e-01 1.01707315e+00
1.20205784e+00 -1.88748538e-01 1.08554041e+00 -4.03235942e-01
-1.17848568e-01 5.26062325e-02 -4.80154186e-01 3.17201465e-01
7.30453908e-01 -1.52097213e+00 4.06780243e-02 2.68085629e-01
5.05020201e-01 4.26268607e-01 9.44204390e-01 -4.04184431e-01
-1.86562228e+00 -3.75777483e-01 3.67438972e-01 -5.21450996e-01
4.67190474e-01 -6.18486047e-01 -3.97675127e-01 8.64091992e-01
1.09708595e+00 -3.97720277e-01 5.23034453e-01 2.50349361e-02
-3.18890750e-01 3.57464790e-01 -1.13085628e+00 1.18865788e+00
7.87715256e-01 -5.15214622e-01 -9.18623805e-01 2.95056045e-01
6.05992615e-01 -5.61754227e-01 -2.44666770e-01 1.43502861e-01
5.98341763e-01 -1.03776181e+00 1.08794415e+00 -8.31426442e-01
4.32065159e-01 -3.95781666e-01 -3.69070619e-01 -1.59694314e+00
-2.52804486e-03 -6.05893373e-01 2.91282862e-01 7.22242117e-01
3.55944961e-01 -1.06949106e-01 6.24226451e-01 6.12924159e-01
-5.31865358e-02 -1.33338794e-01 -3.55658799e-01 -9.65598300e-02
-1.84355825e-01 -7.12772071e-01 7.98714831e-02 8.97911608e-01
8.29820275e-01 4.31804359e-01 -5.33009648e-01 -6.67404244e-03
9.69049707e-02 7.98410252e-02 1.30626130e+00 -7.14487910e-01
-4.13009286e-01 -3.95221412e-01 1.44821899e-02 -1.15471268e+00
2.57952124e-01 -6.99388623e-01 5.89209080e-01 -1.73588908e+00
1.36331543e-01 5.35814799e-02 1.43211529e-01 8.64132404e-01
3.43360126e-01 4.56030890e-02 6.57632291e-01 -2.36758888e-01
-1.10013270e+00 4.29218382e-01 1.42928922e+00 -3.23978007e-01
-6.49498522e-01 -4.86376971e-01 -1.66108355e-01 8.14001679e-01
8.18918347e-01 6.04225360e-02 -5.69138110e-01 -5.32733500e-01
3.27197492e-01 3.84148717e-01 4.46529955e-01 -7.40816176e-01
3.26293021e-01 -6.57404602e-01 4.87051785e-01 -3.56626004e-01
7.15634525e-01 -9.39178586e-01 -4.94367421e-01 3.37380052e-01
-6.49324596e-01 3.34836662e-01 4.12736446e-01 9.70514119e-02
2.94041578e-02 -1.31798372e-01 2.06801221e-01 -7.79008210e-01
-1.19729269e+00 -3.25595140e-01 -1.16159165e+00 -1.90481618e-01
1.08886325e+00 -2.27816716e-01 -5.25176525e-01 -9.88720000e-01
-9.44323778e-01 4.79667455e-01 4.70264614e-01 5.58585227e-01
1.03102100e+00 -1.06441975e+00 -5.04652858e-01 1.36990950e-01
3.17302018e-01 -4.34663713e-01 3.17645371e-02 4.96506244e-01
-7.18989313e-01 4.76723313e-01 -5.61545551e-01 -5.13353467e-01
-1.65882349e+00 7.67978668e-01 3.62453878e-01 -1.41524658e-01
-6.37972057e-01 6.14826500e-01 4.53005403e-01 -9.46491778e-01
5.50078630e-01 -4.01477247e-01 -5.84452868e-01 -2.71758169e-01
6.74219310e-01 -2.26871595e-01 -3.51202458e-01 -5.96296728e-01
-8.55629370e-02 2.06095561e-01 1.17975302e-01 -5.62543333e-01
1.25418591e+00 -2.96855241e-01 1.90020278e-01 7.12903976e-01
6.16541862e-01 -1.47267580e-01 -1.28962207e+00 -2.00120538e-01
-7.71784708e-02 -1.77657560e-01 -2.10263893e-01 -1.11204433e+00
-6.54378414e-01 8.69881749e-01 5.80342174e-01 3.14449638e-01
8.85162354e-01 9.18236673e-02 3.35791618e-01 1.09961438e+00
5.53568006e-01 -1.05531073e+00 6.36084676e-01 8.02008450e-01
1.07026875e+00 -1.53990114e+00 -4.02049989e-01 1.20320775e-01
-1.33530736e+00 1.45755887e+00 9.25574183e-01 8.25746059e-02
-1.05941266e-01 4.51902807e-01 8.09162199e-01 -4.59644973e-01
-9.82064784e-01 -6.29894435e-01 1.54961467e-01 1.10671139e+00
5.94394743e-01 1.72121916e-02 3.35156083e-01 1.20428205e-01
-1.36295885e-01 -1.85754746e-01 4.33109730e-01 1.21677518e+00
-5.45687437e-01 -1.13757789e+00 -4.96400774e-01 -2.19668016e-01
-8.78886506e-03 -1.07537039e-01 -6.95342898e-01 9.85490441e-01
-1.27344579e-01 1.44257176e+00 -2.04668432e-01 -3.48251849e-01
2.95000017e-01 2.46996954e-01 8.07164788e-01 -6.24644160e-01
-9.13569748e-01 7.98821300e-02 3.11567813e-01 -1.00003445e+00
-5.41200459e-01 -4.72699374e-01 -1.38445067e+00 -1.22592740e-01
2.18177199e-01 -1.26276329e-01 5.94640553e-01 1.10901940e+00
-2.06553996e-01 3.39790493e-01 9.73508433e-02 -1.41054201e+00
-1.70605391e-01 -9.25356388e-01 -1.69428781e-01 3.95579785e-01
5.02928555e-01 -5.11712670e-01 1.42762199e-01 7.74885491e-02] | [4.416485786437988, 0.7930276989936829] |
60453adb-3022-47ac-a0c3-60e126cc9548 | rethinking-data-free-quantization-as-a-zero | 2302.09572 | null | https://arxiv.org/abs/2302.09572v1 | https://arxiv.org/pdf/2302.09572v1.pdf | Rethinking Data-Free Quantization as a Zero-Sum Game | Data-free quantization (DFQ) recovers the performance of quantized network (Q) without accessing the real data, but generates the fake sample via a generator (G) by learning from full-precision network (P) instead. However, such sample generation process is totally independent of Q, specialized as failing to consider the adaptability of the generated samples, i.e., beneficial or adversarial, over the learning process of Q, resulting into non-ignorable performance loss. Building on this, several crucial questions -- how to measure and exploit the sample adaptability to Q under varied bit-width scenarios? how to generate the samples with desirable adaptability to benefit the quantized network? -- impel us to revisit DFQ. In this paper, we answer the above questions from a game-theory perspective to specialize DFQ as a zero-sum game between two players -- a generator and a quantized network, and further propose an Adaptability-aware Sample Generation (AdaSG) method. Technically, AdaSG reformulates DFQ as a dynamic maximization-vs-minimization game process anchored on the sample adaptability. The maximization process aims to generate the sample with desirable adaptability, such sample adaptability is further reduced by the minimization process after calibrating Q for performance recovery. The Balance Gap is defined to guide the stationarity of the game process to maximally benefit Q. The theoretical analysis and empirical studies verify the superiority of AdaSG over the state-of-the-arts. Our code is available at https://github.com/hfutqian/AdaSG. | ['Meng Wang', 'Richang Hong', 'Yang Wang', 'Biao Qian'] | 2023-02-19 | null | null | null | null | ['data-free-quantization', 'data-free-quantization'] | ['computer-vision', 'methodology'] | [ 9.92787182e-02 4.01040226e-01 -3.69375974e-01 1.42059013e-01
-8.26007128e-01 -7.50328302e-01 1.57855093e-01 -4.02292490e-01
-3.49455148e-01 9.62317586e-01 -1.83615431e-01 -5.10826886e-01
-3.53113651e-01 -1.18488097e+00 -6.15101278e-01 -1.00012600e+00
-3.07061344e-01 4.97947671e-02 6.08448349e-02 -3.96908164e-01
3.83875966e-01 2.11883783e-01 -1.03025746e+00 -1.82522327e-01
9.39247847e-01 1.03604639e+00 -1.89488068e-01 7.87845373e-01
3.29923838e-01 7.42056549e-01 -9.75030184e-01 -7.13395774e-01
7.84551799e-01 -7.04478860e-01 -4.73381549e-01 -2.25904375e-01
-2.92206287e-01 -7.70142555e-01 -5.80974340e-01 1.52947509e+00
8.33508909e-01 -2.59124488e-01 3.07705432e-01 -1.61456788e+00
-4.76504028e-01 8.52493465e-01 -3.73937488e-01 1.76857248e-01
5.08565903e-02 5.49171567e-01 1.06458247e+00 -2.18032688e-01
4.47392523e-01 1.11248136e+00 2.86466181e-01 5.03158689e-01
-1.22569716e+00 -1.27427697e+00 -3.24372858e-01 -5.32853603e-02
-1.70487607e+00 -6.75883651e-01 8.13894689e-01 -4.64794599e-02
3.41729552e-01 2.05631211e-01 7.50508189e-01 7.72264183e-01
2.39488780e-01 6.09678745e-01 8.85798037e-01 -2.37007856e-01
6.43765450e-01 2.10617613e-02 -4.12439853e-01 5.42469561e-01
3.27454120e-01 6.63994551e-01 -5.83472073e-01 -3.07478994e-01
1.12908208e+00 -4.94434714e-01 -4.91200268e-01 -2.98673838e-01
-9.59887564e-01 8.29973400e-01 4.73382384e-01 -2.81437159e-01
-4.57766950e-01 6.42783105e-01 4.31966931e-01 7.97114968e-01
1.25002801e-01 3.35417479e-01 -2.31149644e-01 -4.36244547e-01
-1.02671325e+00 3.87468457e-01 7.91302204e-01 1.12986791e+00
8.92846882e-01 3.02237391e-01 -3.40398669e-01 1.39042497e-01
2.02322304e-01 6.43016517e-01 1.75301671e-01 -1.38216388e+00
6.63579345e-01 6.67893514e-02 2.48290777e-01 -1.06038880e+00
-2.22803857e-02 -6.09165728e-01 -1.01770115e+00 2.76069432e-01
4.74582225e-01 -6.02672160e-01 -2.75733501e-01 2.02871704e+00
1.39758378e-01 2.23245993e-01 1.16046764e-01 1.05112195e+00
2.23786071e-01 6.29653275e-01 -4.02313054e-01 -4.46048796e-01
1.00666952e+00 -3.98893803e-01 -6.45908535e-01 1.33453056e-01
4.96504396e-01 -2.94720531e-01 1.09892142e+00 7.58715510e-01
-1.26978207e+00 -3.14726651e-01 -1.43085551e+00 5.38271248e-01
1.23995051e-01 -1.70021906e-01 2.05488950e-01 1.36990654e+00
-1.24316502e+00 6.27934098e-01 -4.72184569e-01 1.22351378e-01
6.20938063e-01 6.75832748e-01 8.32845569e-02 1.36273161e-01
-1.63710034e+00 2.65135199e-01 6.05263174e-01 -3.71426605e-02
-1.01687181e+00 -5.54346681e-01 -2.82873720e-01 6.61000535e-02
7.62671828e-01 -7.96836436e-01 1.24772787e+00 -1.05623055e+00
-2.05288100e+00 3.41072142e-01 2.35740572e-01 -7.16009319e-01
8.76431584e-01 3.43876332e-01 -1.66355744e-01 3.53995055e-01
4.38676327e-02 3.86650264e-01 9.92479622e-01 -1.09868062e+00
-6.71857595e-01 -1.40299708e-01 4.30677533e-01 4.27248403e-02
-1.84672713e-01 -3.87521625e-01 -3.71175289e-01 -4.94019121e-01
-8.19439292e-02 -8.22883010e-01 -2.55424201e-01 1.16375357e-01
-6.92102790e-01 2.27466226e-01 4.84002262e-01 -3.81005585e-01
1.25516391e+00 -2.24778557e+00 -3.15155655e-01 5.32896638e-01
4.74806517e-01 2.50900328e-01 -2.11378992e-01 4.93647307e-01
2.84335405e-01 4.38911438e-01 4.05830666e-02 1.81648619e-02
7.70538822e-02 1.53976515e-01 -3.35570961e-01 6.04392707e-01
1.78957701e-01 1.06825340e+00 -1.37914562e+00 -3.45314264e-01
1.22784503e-01 9.98685881e-02 -8.05675924e-01 5.22782877e-02
7.91689008e-03 3.30443352e-01 -7.10599482e-01 5.92676282e-01
8.91971588e-01 -5.41089289e-02 3.31278175e-01 -2.49052033e-01
1.02850623e-01 1.19517058e-01 -1.30774367e+00 1.42070758e+00
-8.44791234e-02 1.80361852e-01 4.38470960e-01 -1.05544376e+00
9.89282191e-01 3.10310662e-01 2.66039461e-01 -9.52908933e-01
3.02564770e-01 4.58889604e-01 2.93055344e-02 -2.62682617e-01
6.63474202e-01 -2.56171644e-01 -2.63990849e-01 4.85408515e-01
-1.75829157e-02 -2.43428007e-01 -1.28699280e-02 3.59898031e-01
1.09779799e+00 -2.35291526e-01 3.56364340e-01 -2.12022766e-01
2.94818670e-01 -4.94762927e-01 6.83948457e-01 1.07076395e+00
-7.03295767e-01 2.94157386e-01 1.07507372e+00 1.77210838e-01
-1.06076014e+00 -1.33148980e+00 2.78787553e-01 5.80663383e-01
4.85380173e-01 -3.64891946e-01 -6.68124259e-01 -5.75234950e-01
-1.21019140e-01 5.80009043e-01 -3.03587496e-01 -4.51215565e-01
-8.23461115e-02 -6.39570415e-01 9.60997403e-01 -1.37640506e-01
9.12063777e-01 -7.86163688e-01 -3.86681318e-01 2.04336300e-01
-1.81045309e-01 -7.00968087e-01 -4.39509243e-01 5.92739135e-02
-5.15328050e-01 -9.12337959e-01 -3.40277165e-01 -1.50052384e-01
3.61764610e-01 1.79289281e-01 9.09883082e-01 7.12494552e-02
3.32419127e-01 2.92018279e-02 -4.28364635e-01 -1.55790418e-01
-7.04763651e-01 1.51434526e-01 1.03559159e-01 -5.72620183e-02
-1.48761198e-01 -9.67452049e-01 -9.28057730e-01 2.42190301e-01
-9.74380374e-01 -2.22849354e-01 7.61270821e-01 6.94036543e-01
4.95131880e-01 6.16289198e-01 1.02213562e+00 -7.28718460e-01
9.33618844e-01 -6.38980269e-01 -7.59246945e-01 -4.01011854e-02
-5.75727582e-01 1.55364677e-01 1.03094542e+00 -3.27457249e-01
-3.46281141e-01 -2.66309172e-01 -1.24138743e-01 -4.40558285e-01
4.29181576e-01 2.27963269e-01 -6.23015404e-01 -3.37031811e-01
6.74144387e-01 3.68261307e-01 2.96967655e-01 3.62623602e-01
7.99572706e-01 7.34949708e-01 5.20347357e-01 -7.10362494e-01
1.04477990e+00 4.44476247e-01 1.26770392e-01 -5.32916784e-01
-3.37516963e-01 8.83351117e-02 -4.78424393e-02 -4.28004384e-01
2.29458347e-01 -8.74196947e-01 -1.14739275e+00 3.64254177e-01
-8.23882759e-01 -5.68677723e-01 -6.80324554e-01 1.57596275e-01
-8.13724577e-01 4.22799259e-01 -4.51194912e-01 -1.03489983e+00
-4.14971530e-01 -1.16027057e+00 6.28185034e-01 1.58962756e-01
2.59719908e-01 -5.92365265e-01 -2.47900635e-01 1.06910922e-01
3.98456365e-01 2.46105507e-01 5.31255186e-01 -2.87961721e-01
-1.05873084e+00 -2.17030093e-01 -3.75660360e-01 4.23659474e-01
2.84288060e-02 1.90729629e-02 -8.71777833e-01 -6.46543324e-01
3.17659117e-02 -3.51283193e-01 4.01336610e-01 3.95703465e-01
1.19639289e+00 -6.68263853e-01 1.39780194e-01 8.09972703e-01
1.57053125e+00 2.67253574e-02 9.03422117e-01 1.84816763e-01
2.32542709e-01 1.96320117e-01 5.72432041e-01 9.17806208e-01
2.47157797e-01 4.86317903e-01 9.17887270e-01 3.83537829e-01
9.20025706e-02 -5.29772878e-01 5.43702900e-01 5.98605931e-01
1.58740595e-01 -5.79265296e-01 -3.33688617e-01 3.20619583e-01
-1.67150116e+00 -9.00617421e-01 1.52227089e-01 2.55795312e+00
9.48724687e-01 5.35501719e-01 3.55686575e-01 4.97269541e-01
7.46975362e-01 2.83765852e-01 -7.02765107e-01 -2.31752530e-01
1.46789346e-02 2.56149113e-01 1.01583707e+00 3.58053774e-01
-5.73489428e-01 7.13961601e-01 5.93157673e+00 1.60970747e+00
-1.23295903e+00 1.30203843e-01 7.79570401e-01 -2.06710368e-01
-4.19625193e-01 1.83814362e-01 -4.41559732e-01 8.36899579e-01
1.10722268e+00 -6.61558867e-01 7.72852898e-01 4.65028077e-01
6.45194530e-01 -3.02466601e-02 -7.22477019e-01 8.19136322e-01
-4.49716210e-01 -1.25245643e+00 9.62977763e-03 4.15404230e-01
5.34804463e-01 -2.34849900e-01 2.58636504e-01 4.10948664e-01
4.70291197e-01 -7.84682572e-01 9.71873224e-01 5.20301461e-01
1.16631091e+00 -1.09616983e+00 7.46313274e-01 5.74430704e-01
-9.49051499e-01 -4.00910467e-01 -4.40497786e-01 -1.94531858e-01
2.21718252e-02 9.80742097e-01 -6.60589278e-01 7.99858332e-01
3.01608175e-01 3.30497116e-01 -2.25163862e-01 8.86175096e-01
-4.68066394e-01 8.46668005e-01 -1.93872556e-01 -9.84278992e-02
1.01486787e-01 -3.63976866e-01 6.91558242e-01 6.23273849e-01
4.32222545e-01 2.47238781e-02 1.21086523e-01 1.28379059e+00
-2.86642224e-01 -1.63845867e-01 -2.29305029e-01 -2.36890372e-02
1.23844826e+00 1.06575167e+00 -4.14641291e-01 -1.86632410e-01
-7.60502275e-03 7.29411364e-01 -1.13083087e-02 4.40841287e-01
-9.15828705e-01 -6.50052011e-01 5.50908923e-01 2.16499299e-01
1.04821675e-01 -2.49469653e-02 -2.94526428e-01 -8.55976641e-01
4.81375121e-02 -9.54232931e-01 8.05892888e-03 -5.23348749e-01
-1.30360842e+00 5.77072948e-02 -2.40854010e-01 -1.61885440e+00
-2.40030661e-01 3.26810703e-02 -5.14745116e-01 9.70894217e-01
-1.67328811e+00 -7.03263164e-01 -1.47451684e-01 6.39858723e-01
-6.44411072e-02 -8.53066444e-02 3.51269633e-01 2.36937240e-01
-4.66568381e-01 1.16909897e+00 3.38140637e-01 2.22033009e-01
5.46965301e-01 -1.01506746e+00 2.58575350e-01 1.00962543e+00
-3.73262405e-01 3.15180510e-01 5.95758736e-01 -4.22932774e-01
-1.39317524e+00 -1.19005597e+00 2.49917358e-01 -5.30517697e-02
7.98083186e-01 -3.68841141e-01 -4.67944175e-01 2.85355728e-02
-1.86830625e-01 2.25688860e-01 3.96507561e-01 -6.58994675e-01
-1.64998233e-01 -3.67968202e-01 -1.55118704e+00 6.16604149e-01
8.76908839e-01 -6.12869382e-01 1.66903883e-01 -1.14855632e-01
8.96104872e-01 -2.70429760e-01 -8.02700818e-01 5.85248368e-03
4.56593394e-01 -1.35611165e+00 9.34346080e-01 6.95761070e-02
3.68689924e-01 -4.08331692e-01 -2.15834215e-01 -1.21698904e+00
-1.52897328e-01 -1.23956096e+00 -2.79342860e-01 1.13840342e+00
7.84961954e-02 -9.83402729e-01 1.21384370e+00 1.04716614e-01
2.05392689e-01 -6.40590906e-01 -1.23413205e+00 -9.92401242e-01
3.48503858e-01 -4.02921677e-01 8.59746873e-01 6.54146492e-01
5.72572425e-02 -2.30245572e-02 -4.63508189e-01 3.36000443e-01
8.53176594e-01 -2.90362954e-01 8.99860263e-01 -8.11522782e-01
-5.60885906e-01 -4.68460649e-01 -5.31860709e-01 -1.33731079e+00
-1.93628460e-01 -7.64627397e-01 -3.87868173e-02 -7.84598887e-01
-3.70219424e-02 -6.67966008e-01 -1.23408213e-01 1.92543790e-01
-1.95350647e-01 2.88812518e-01 4.69315708e-01 4.46618974e-01
-6.44468248e-01 7.36318767e-01 1.56423104e+00 1.03617392e-01
-3.74377638e-01 3.54805171e-01 -1.00625348e+00 -1.21674379e-02
1.10170650e+00 -3.88742685e-01 -8.07827234e-01 1.97899193e-01
6.32956743e-01 5.84170759e-01 4.55249667e-01 -1.08554208e+00
6.29549325e-02 -1.74294159e-01 -2.25102067e-01 -2.48395264e-01
1.04030162e-01 -6.14635885e-01 2.33406536e-02 8.31653953e-01
-1.87465698e-01 -4.53804731e-01 -2.97342300e-01 8.27080131e-01
-2.42004488e-02 -2.01432228e-01 6.66089356e-01 -7.22981095e-02
-3.38433534e-01 5.62079847e-01 -4.23971742e-01 3.90745640e-01
8.52461636e-01 -4.10000145e-01 -5.25236964e-01 -8.83463204e-01
-4.88781333e-01 3.78482521e-01 5.21713734e-01 -1.81192398e-01
4.09637868e-01 -1.34737122e+00 -6.95889771e-01 3.09853435e-01
-1.46194369e-01 -4.77926210e-02 1.44745752e-01 6.98078394e-01
-4.10119474e-01 8.79426003e-02 -7.50130564e-02 -2.88075000e-01
-5.37606537e-01 4.47232723e-01 6.38084888e-01 -4.49471265e-01
-5.41737378e-02 6.75093114e-01 -1.91488370e-01 -5.23241997e-01
1.92568712e-02 -2.33984023e-01 3.84477705e-01 -1.90118283e-01
2.90773302e-01 3.77745032e-01 -6.12929165e-02 -1.87161565e-01
-5.10608188e-05 1.33378357e-01 4.35243219e-01 -3.44235778e-01
8.04231226e-01 -4.77120340e-01 -5.90825081e-03 5.37843779e-02
1.06088233e+00 7.41694942e-02 -1.41139078e+00 -2.26276040e-01
-5.37368894e-01 -6.60594404e-01 1.12097368e-01 -4.99106646e-01
-1.33247781e+00 7.15413451e-01 4.45351839e-01 6.31486237e-01
1.29580426e+00 -4.94654387e-01 7.68166602e-01 4.98921536e-02
9.09083605e-01 -9.52018142e-01 2.47152984e-01 2.07377031e-01
3.91221404e-01 -7.92426646e-01 -6.56420439e-02 -3.72184336e-01
-3.39552432e-01 1.04306841e+00 3.44144136e-01 -2.32080117e-01
6.98129058e-01 8.69842619e-02 -3.16454440e-01 2.03367285e-02
-6.71085358e-01 1.99873634e-02 -4.42648619e-01 9.37423587e-01
-1.05212741e-01 2.69752175e-01 -2.79394537e-01 8.30410302e-01
-6.70453072e-01 3.52615088e-01 9.50306177e-01 8.09463143e-01
-4.56643105e-01 -1.10191917e+00 -4.01119828e-01 6.67609572e-01
-3.02253008e-01 -6.16851635e-02 -1.90036252e-01 5.54449975e-01
3.82099807e-01 1.10037160e+00 -5.48285320e-02 -6.97116613e-01
2.40925431e-01 -4.83723789e-01 1.77259743e-01 -2.83484280e-01
-4.53661233e-01 -1.30803555e-01 -2.90363654e-02 -7.62905419e-01
-2.01216578e-01 -2.95687288e-01 -1.03814030e+00 -9.51865077e-01
-6.31587923e-01 2.49485224e-01 2.53447622e-01 6.20850205e-01
3.67178321e-01 4.36697066e-01 1.17093337e+00 -6.09145343e-01
-1.12913215e+00 -4.16536361e-01 -1.13005793e+00 -5.13839945e-02
4.50642377e-01 -2.24826336e-01 -7.13299394e-01 -4.25449252e-01] | [8.688275337219238, 2.900325059890747] |
a0187258-bbf0-44c4-97cf-0df33f455b6f | wsense-a-robust-feature-learning-module-for | 2303.17845 | null | https://arxiv.org/abs/2303.17845v1 | https://arxiv.org/pdf/2303.17845v1.pdf | WSense: A Robust Feature Learning Module for Lightweight Human Activity Recognition | In recent times, various modules such as squeeze-and-excitation, and others have been proposed to improve the quality of features learned from wearable sensor signals. However, these modules often cause the number of parameters to be large, which is not suitable for building lightweight human activity recognition models which can be easily deployed on end devices. In this research, we propose a feature learning module, termed WSense, which uses two 1D CNN and global max pooling layers to extract similar quality features from wearable sensor data while ignoring the difference in activity recognition models caused by the size of the sliding window. Experiments were carried out using CNN and ConvLSTM feature learning pipelines on a dataset obtained with a single accelerometer (WISDM) and another obtained using the fusion of accelerometers, gyroscopes, and magnetometers (PAMAP2) under various sliding window sizes. A total of nine hundred sixty (960) experiments were conducted to validate the WSense module against baselines and existing methods on the two datasets. The results showed that the WSense module aided pipelines in learning similar quality features and outperformed the baselines and existing models with a minimal and uniform model size across all sliding window segmentations. The code is available at https://github.com/AOige/WSense. | ['Mohd Halim Mohd Noor', 'Ayokunle Olalekan Ige'] | 2023-03-31 | null | null | null | null | ['human-activity-recognition', 'human-activity-recognition'] | ['computer-vision', 'time-series'] | [ 3.05368323e-02 -3.52244586e-01 -2.00308621e-01 -5.36515415e-01
-3.83064508e-01 -7.31655955e-02 3.02343190e-01 8.43346342e-02
-6.62013769e-01 5.26943743e-01 2.81103671e-01 1.61278710e-01
-1.42207155e-02 -8.40478063e-01 -6.33955479e-01 -4.28021282e-01
-2.98639983e-01 -4.67362076e-01 1.85228720e-01 1.34436265e-01
-5.88200577e-02 9.71052423e-02 -1.35340488e+00 1.40158683e-01
6.25045955e-01 1.40533745e+00 -1.77377731e-01 5.66107213e-01
2.39162132e-01 3.69287580e-01 -9.16352272e-01 5.22006117e-02
2.32948765e-01 -1.51379272e-01 -3.73133063e-01 -1.09793141e-01
2.43587956e-01 -3.35272819e-01 -3.39839667e-01 5.58349967e-01
8.11805606e-01 2.46953681e-01 -3.79585847e-02 -1.23240066e+00
-4.46134090e-01 4.41436470e-01 -3.94568920e-01 4.61741656e-01
6.28705859e-01 1.70420736e-01 5.27833402e-01 -7.39459395e-01
-4.78674509e-02 9.08192635e-01 1.14929247e+00 3.36437672e-01
-8.93698692e-01 -8.42762113e-01 -3.58521082e-02 8.87460709e-02
-1.42103601e+00 -3.05301189e-01 6.65798306e-01 -3.72314155e-01
1.43846107e+00 2.20937937e-01 1.11537206e+00 1.50595582e+00
4.72292364e-01 7.76569724e-01 8.61182034e-01 8.22434202e-02
3.47098500e-01 -1.38638437e-01 4.43836451e-01 6.13821805e-01
5.26856422e-01 -1.35654360e-01 -9.08231676e-01 -3.05891186e-01
8.05946827e-01 5.63436985e-01 -1.38469115e-02 6.08207546e-02
-1.40379238e+00 5.27931750e-01 4.94223386e-01 4.34167981e-01
-4.33710635e-01 8.11688676e-02 4.79933590e-01 7.93398395e-02
3.99973601e-01 3.65327597e-01 -7.18096852e-01 -6.26524746e-01
-1.03102970e+00 8.94872472e-02 6.91656053e-01 6.24350965e-01
4.31983709e-01 1.78547949e-01 -2.50440612e-02 6.03565395e-01
2.53972501e-01 4.60341960e-01 1.02142763e+00 -5.35522461e-01
5.16861975e-01 9.72746432e-01 1.04610935e-01 -1.05387545e+00
-9.62811112e-01 -3.67244363e-01 -1.06257856e+00 -3.06590468e-01
1.14095211e-01 -5.93915343e-01 -7.62340546e-01 1.40172327e+00
2.65801668e-01 5.38319707e-01 -3.00276935e-01 8.25409770e-01
9.44342017e-01 5.54216564e-01 1.54413477e-01 1.91601723e-01
1.33002865e+00 -8.42254639e-01 -7.61102438e-01 -4.54358488e-01
6.24325752e-01 -2.71989763e-01 1.20563304e+00 4.39133078e-01
-7.78474271e-01 -1.01951241e+00 -1.70641720e+00 -6.14179336e-02
-5.48748970e-01 2.94355452e-01 7.24715471e-01 9.75671589e-01
-8.00432622e-01 9.63235259e-01 -1.73095846e+00 -4.58074003e-01
6.60471618e-01 6.18253708e-01 -3.67237836e-01 3.94190878e-01
-1.25131047e+00 7.10072041e-01 4.29942101e-01 4.38780695e-01
-6.07631683e-01 -3.17075580e-01 -9.41057265e-01 -3.75117026e-02
-8.53222087e-02 -4.80960816e-01 9.87014294e-01 -6.60366774e-01
-1.51768672e+00 2.45346814e-01 -6.75876513e-02 -6.10394418e-01
2.82696277e-01 -8.67688954e-01 -6.87901318e-01 -2.87165344e-01
-9.12514925e-02 3.44344944e-01 6.00530326e-01 -6.43469989e-02
-4.98220503e-01 -4.29950118e-01 -2.30103672e-01 -5.70787005e-02
-6.43311381e-01 -2.69325767e-02 -6.65640533e-02 -5.40087223e-01
-1.87810361e-02 -7.92368174e-01 -1.32196873e-01 -2.42219895e-01
-3.64090532e-01 -1.83609903e-01 7.26792097e-01 -7.26179600e-01
1.38959539e+00 -2.23030996e+00 -3.04147601e-01 6.63123652e-02
1.57269001e-01 4.33088601e-01 1.64632797e-01 2.31796786e-01
3.48777045e-04 1.97278429e-02 -1.16987832e-01 -3.76215398e-01
-1.16452858e-01 2.56668568e-01 1.91504747e-01 6.27913535e-01
3.81393671e-01 8.91235828e-01 -8.01748395e-01 -2.46771559e-01
4.16907579e-01 7.48949826e-01 -2.91610897e-01 2.20139742e-01
4.34934884e-01 4.85773832e-01 -5.48597038e-01 7.09867954e-01
4.42002773e-01 -2.97591358e-01 -1.00962035e-01 -2.61271626e-01
-3.66704836e-02 5.22271574e-01 -1.39907837e+00 1.86928511e+00
-2.89317757e-01 5.51585138e-01 -5.61910212e-01 -9.37536061e-01
1.03767776e+00 4.09954458e-01 7.60984719e-01 -6.08978510e-01
4.37507331e-01 1.58750601e-02 -1.13028578e-01 -6.61295474e-01
3.29455405e-01 3.69609714e-01 -4.07067955e-01 2.79025018e-01
2.73861378e-01 3.60945463e-01 1.13267794e-01 -4.16546732e-01
1.32277620e+00 2.95689106e-01 4.06513035e-01 -2.18967628e-02
1.49009496e-01 -3.55057776e-01 9.59667921e-01 6.39380634e-01
-4.85376537e-01 5.11907518e-01 -8.35345462e-02 -8.38279247e-01
-5.88445604e-01 -1.13616717e+00 1.48077514e-02 9.78095293e-01
-4.31221537e-02 -7.42672861e-01 -7.92352617e-01 -3.82733792e-01
-9.14326757e-02 3.63564231e-02 -6.92013383e-01 -3.16571623e-01
-6.34394646e-01 -1.25286996e+00 9.80005443e-01 1.02002752e+00
1.15114510e+00 -1.04559338e+00 -1.25197256e+00 4.10362810e-01
-2.48864204e-01 -1.08028734e+00 -4.09602255e-01 3.50923657e-01
-1.21646786e+00 -1.08587217e+00 -2.96142012e-01 -3.92841578e-01
1.75855145e-01 3.79327126e-02 6.17311776e-01 -1.75060466e-01
-1.73550576e-01 7.34220743e-02 -3.67988735e-01 -6.95415258e-01
6.05196416e-01 4.87403303e-01 3.89127731e-01 4.30422537e-02
9.23894703e-01 -7.55396783e-01 -7.92520285e-01 1.56063512e-01
-7.52701938e-01 -6.01837263e-02 5.31603098e-01 5.16744137e-01
4.13706571e-01 -1.71547905e-01 5.55016279e-01 -8.59187245e-02
8.28384995e-01 -5.94473064e-01 -1.16511904e-01 -1.95101708e-01
-5.48413694e-01 -2.32640617e-02 4.65133667e-01 -7.63568938e-01
-4.90041673e-01 1.42670244e-01 -3.55844140e-01 1.06231486e-02
-1.88818172e-01 4.16759908e-01 -3.09221268e-01 1.64242983e-01
7.59087801e-01 1.05284303e-01 -2.09503397e-01 -6.88069582e-01
-2.64519483e-01 8.79999638e-01 4.47063863e-01 -2.10284948e-01
3.15590739e-01 3.81408691e-01 -4.34772968e-01 -8.08746696e-01
-7.04660118e-01 -3.64400297e-01 -6.50255144e-01 3.89307812e-02
9.06491220e-01 -1.13154685e+00 -4.90698665e-01 9.81602848e-01
-8.69213998e-01 -3.19839031e-01 -2.59190649e-01 7.69357920e-01
-1.59755692e-01 4.00483832e-02 -7.24263251e-01 -6.09261096e-01
-7.72055864e-01 -8.80944967e-01 1.08650112e+00 6.78857505e-01
-7.00385094e-01 -6.49651945e-01 6.13500774e-02 2.04835922e-01
6.54701769e-01 7.37133265e-01 -2.00612610e-03 -7.71787047e-01
1.35131508e-01 -5.18607140e-01 2.52422482e-01 5.18076241e-01
4.96685088e-01 -1.43742353e-01 -1.21345866e+00 -3.56656611e-01
2.35949308e-01 -1.04843616e-01 5.45149624e-01 4.14976269e-01
1.09409058e+00 -2.78343886e-01 -3.13212126e-01 7.14809537e-01
9.73904252e-01 2.53527373e-01 7.07334697e-01 5.83111286e-01
7.07407176e-01 -1.75953314e-01 2.68410593e-01 6.29141688e-01
4.47972387e-01 5.47006667e-01 2.31887504e-01 -2.42674977e-01
3.12144548e-01 -1.15410537e-01 6.09150469e-01 9.53122139e-01
-3.12961638e-01 1.92532882e-01 -7.12396622e-01 5.74923456e-01
-1.74880672e+00 -7.97905862e-01 7.94379339e-02 2.07946444e+00
6.51172757e-01 3.49208653e-01 4.86638546e-01 5.38618386e-01
5.13858438e-01 3.57235998e-01 -6.69581532e-01 -3.22591722e-01
9.22732875e-02 4.88130629e-01 4.95762348e-01 -5.95029444e-02
-1.38560772e+00 3.26699406e-01 6.35462093e+00 2.33556360e-01
-1.30993927e+00 3.92320037e-01 3.57852995e-01 -6.53618813e-01
6.36301219e-01 -4.45343584e-01 -8.25890183e-01 9.11251068e-01
1.42430174e+00 2.22357631e-01 9.22007859e-02 1.05648839e+00
4.58922237e-01 -2.15463340e-01 -9.46348548e-01 1.19466555e+00
-1.79702088e-01 -8.86016667e-01 -4.56372648e-01 -3.95430066e-02
3.88209939e-01 4.97091591e-01 -2.91917354e-01 3.51204038e-01
-8.96839947e-02 -9.12510335e-01 4.69953775e-01 4.76454347e-01
3.23165715e-01 -6.04196489e-01 9.18280602e-01 2.32538536e-01
-1.41498959e+00 -3.25747460e-01 -1.16742253e-01 -7.27587759e-01
1.12713076e-01 8.21517825e-01 -6.02880597e-01 4.01342154e-01
1.18132710e+00 9.78835344e-01 -8.09574664e-01 1.13069510e+00
-4.79479283e-02 7.49395192e-01 -6.44284546e-01 -2.36293674e-01
7.25878477e-02 1.11257404e-01 1.49243101e-01 1.41012919e+00
5.15129209e-01 -2.32391149e-01 1.47725075e-01 4.40288961e-01
-6.71065506e-03 -2.18124554e-01 -3.66958946e-01 -2.84611457e-03
3.85461837e-01 1.19120336e+00 -6.04501188e-01 -4.32419568e-01
-5.62995374e-01 7.68583953e-01 -1.96782295e-02 1.12000868e-01
-1.08956289e+00 -6.99576735e-01 8.28063011e-01 2.63751000e-01
7.29107112e-02 -5.46664536e-01 -4.56973225e-01 -1.32306421e+00
4.34235692e-01 -6.73963606e-01 3.70997071e-01 -4.29647565e-01
-1.01703763e+00 5.06770432e-01 -2.32096743e-02 -1.31806707e+00
-8.74835774e-02 -6.22131050e-01 -8.38460267e-01 5.73106587e-01
-9.50356662e-01 -8.30534816e-01 -7.88768888e-01 7.10530221e-01
5.32472610e-01 1.94324389e-01 7.54807174e-01 6.50972307e-01
-1.01745856e+00 4.87100661e-01 -2.28219345e-01 5.75869739e-01
5.70197940e-01 -9.23842788e-01 8.76765370e-01 9.43568885e-01
-9.86714661e-03 9.59034443e-01 3.72571677e-01 -5.48761904e-01
-1.31557083e+00 -1.23996544e+00 6.89265966e-01 -3.96028399e-01
3.64977121e-01 -5.10815024e-01 -8.19913983e-01 8.29903781e-01
4.98329960e-02 4.16633099e-01 1.00349474e+00 8.79451632e-03
6.64265230e-02 -2.84286261e-01 -1.13397932e+00 1.72810838e-01
1.06762135e+00 -2.75213927e-01 -7.63950825e-01 -1.36867240e-01
4.14424896e-01 -5.35732031e-01 -1.25047112e+00 5.73214591e-01
9.81242180e-01 -6.57922566e-01 8.36435258e-01 -4.34566915e-01
-1.07404768e-01 -3.52523685e-01 -9.00650322e-02 -1.24179494e+00
-2.86378562e-01 -4.07730520e-01 -5.77951729e-01 9.33532953e-01
1.73549220e-01 -8.56875598e-01 5.85956693e-01 6.51365578e-01
-1.38458595e-01 -8.72272372e-01 -9.29697037e-01 -8.15005004e-01
-5.32473385e-01 -5.94700813e-01 9.78766024e-01 8.88631582e-01
1.75127238e-01 3.62530559e-01 -4.46095854e-01 -2.88120210e-02
8.80247355e-02 -3.75443697e-01 6.96251512e-01 -1.31703794e+00
-2.30304480e-01 -7.85578489e-02 -7.69385040e-01 -8.65660012e-01
-5.84577441e-01 -4.22485977e-01 -1.57478884e-01 -1.44048178e+00
-1.12943277e-01 3.15545797e-02 -7.11585939e-01 9.32753682e-01
-1.42519429e-01 4.12422448e-01 -1.69217333e-01 -5.88054806e-02
-6.81838274e-01 4.98520762e-01 7.53491580e-01 -1.29129887e-01
-6.36759698e-01 2.10101873e-01 -6.78833127e-01 8.55316758e-01
1.24885368e+00 -4.60330427e-01 -2.64381230e-01 -4.69913691e-01
2.27585480e-01 -4.79495496e-01 5.06088495e-01 -1.69070017e+00
1.58664718e-01 1.99210480e-01 1.13120615e+00 -4.04472172e-01
4.14525241e-01 -5.16568184e-01 2.89244860e-01 5.21822631e-01
-4.85454919e-03 3.74629676e-01 4.35043514e-01 1.81478158e-01
-4.79404479e-02 3.13445538e-01 3.73728603e-01 4.90038842e-02
-5.70635915e-01 2.89925069e-01 -4.05212522e-01 -1.78531215e-01
7.39366472e-01 -5.12211680e-01 -1.64260283e-01 4.51885834e-02
-8.40594471e-01 1.15247510e-01 5.82125373e-02 8.45000982e-01
6.29879475e-01 -1.59774482e+00 -2.46554464e-01 5.14581323e-01
2.02666279e-02 1.20736055e-01 1.25426680e-01 1.02804863e+00
-2.01560065e-01 3.21417361e-01 -5.56809187e-01 -5.62819183e-01
-9.57711041e-01 8.39812979e-02 5.23121476e-01 -3.06408733e-01
-4.94142592e-01 5.21819890e-01 -6.20621562e-01 -4.49682415e-01
3.15047890e-01 -1.05447137e+00 -1.06296226e-01 5.19152824e-03
8.35780621e-01 6.73545778e-01 3.98178041e-01 -3.62176657e-01
-7.08866715e-01 4.54994917e-01 2.18722627e-01 2.30853453e-01
1.39544392e+00 4.48562503e-02 4.42673296e-01 5.72290361e-01
1.16615462e+00 -4.36248630e-01 -1.36596513e+00 -3.70680355e-02
1.41242743e-01 -2.27515146e-01 -1.13463290e-01 -3.52576345e-01
-1.07345045e+00 8.69898021e-01 1.21349704e+00 1.51173770e-01
1.15202057e+00 -4.32551503e-01 1.16460931e+00 3.24544132e-01
4.07838702e-01 -1.34059429e+00 3.22011387e-04 2.95121402e-01
4.06198740e-01 -1.15762293e+00 -1.73499193e-02 2.07856908e-01
-3.83074135e-01 9.20299053e-01 6.85777128e-01 -4.43892926e-01
7.78532445e-01 2.91458964e-01 -1.37200758e-01 -1.42677769e-01
-3.03395987e-01 4.26576398e-02 2.59912908e-01 6.22475386e-01
6.27848566e-01 1.10759243e-01 -2.60481775e-01 1.05406487e+00
-3.48096788e-01 4.97650385e-01 1.77681133e-01 1.41518795e+00
-3.69089812e-01 -8.21532249e-01 -2.65620232e-01 7.09442377e-01
-7.58585751e-01 1.70260996e-01 -2.11400777e-01 6.78973258e-01
5.30810118e-01 1.27548635e+00 1.23873189e-01 -9.72976923e-01
5.57116449e-01 2.50890940e-01 2.04122931e-01 -4.47339356e-01
-9.10756111e-01 -1.96257249e-01 -1.51463943e-02 -1.04533517e+00
-5.48761606e-01 -6.68707073e-01 -1.22917449e+00 -1.02784537e-01
-1.39122620e-01 -1.29718497e-01 6.36113584e-01 1.05963159e+00
7.61598647e-01 7.42495596e-01 4.00056362e-01 -1.10057759e+00
-3.21234941e-01 -1.51589429e+00 -5.26029050e-01 1.72736183e-01
3.57267559e-01 -7.77788699e-01 3.13345692e-03 1.29826263e-01] | [7.563937664031982, 0.8184187412261963] |
8fb83dc9-0c14-47f2-8323-5f7e5ea62f91 | shape-complementarity-optimization-of | 2107.07295 | null | https://arxiv.org/abs/2107.07295v4 | https://arxiv.org/pdf/2107.07295v4.pdf | Shape Complementarity Optimization of Antibody-Antigen Interfaces: the Application to SARS-CoV-2 Spike Protein | Many factors influence biomolecules binding, and its assessment constitutes an elusive challenge in computational structural biology. In this respect, the evaluation of shape complementarity at molecular interfaces is one of the main factors to be considered. We focus on the particular case of antibody-antigen complexes to quantify the complementarities occurring at molecular interfaces. We relied on a method we recently developed, which employs the 2D Zernike descriptors, to characterize investigated regions with an ordered set of numbers summarizing the local shape properties. Collected a structural dataset of antibody-antigen complexes, we applied this method and we statistically distinguished, in terms of shape complementarity, pairs of interacting regions from non-interacting ones. Thus, we set up a novel computational strategy based on \textit{in-silico} mutagenesis of antibody binding site residues. We developed a Monte Carlo procedure to increase the shape complementarity between the antibody paratope and a given epitope on a target protein surface. We applied our protocol against several molecular targets in SARS-CoV-2 spike protein, known to be indispensable for viral cell invasion. We, therefore, optimized the shape of template antibodies for the interaction with such regions. As the last step of our procedure, we performed an independent molecular docking validation of the results of our Monte Carlo simulations. | ['Giancarlo Ruocco', 'Edoardo Milanetti', 'Pier Paolo Olimpieri', 'Mattia Miotto', 'Lorenzo Di Rienzo', 'Alfredo De Lauro'] | 2021-07-15 | null | null | null | null | ['molecular-docking'] | ['medical'] | [ 3.90556186e-01 -2.74399996e-01 2.83697218e-01 -3.24547142e-01
-4.91279632e-01 -8.41214836e-01 3.89264315e-01 7.05803752e-01
-6.95059001e-01 1.14852250e+00 -1.55644849e-01 -4.02407497e-01
-1.36955529e-01 -6.34466290e-01 -8.57850015e-01 -8.34992945e-01
-2.75159836e-01 1.02622724e+00 1.55375913e-01 -4.45835471e-01
4.27738011e-01 1.09758282e+00 -1.27784646e+00 3.88125062e-01
1.14846349e+00 3.19155991e-01 2.25156650e-01 3.51659715e-01
-2.38986894e-01 -4.21978533e-02 -2.75653601e-01 -1.70177102e-01
-7.96701536e-02 -4.57208425e-01 -3.27052087e-01 -3.65713447e-01
-2.96813566e-02 4.14196610e-01 7.27955282e-01 7.84037769e-01
4.38601553e-01 5.98712713e-02 1.26096582e+00 -5.13697982e-01
-4.81871590e-02 -1.42490312e-01 -9.21562314e-02 2.38345459e-01
5.56088924e-01 1.09947339e-01 9.69453096e-01 -1.15620029e+00
1.04286611e+00 8.90167892e-01 4.71927166e-01 3.06046009e-01
-1.64045334e+00 -4.21619117e-01 -2.77140647e-01 3.05172920e-01
-1.30865121e+00 -3.02563190e-01 5.77306628e-01 -9.76511300e-01
8.72911394e-01 3.36375564e-01 8.74232173e-01 6.02625608e-01
5.82803011e-01 -4.82515357e-02 1.42451918e+00 -3.32163930e-01
4.32129145e-01 3.83142680e-02 3.37884456e-01 4.31171030e-01
4.33224678e-01 1.14280753e-01 -1.29366353e-01 -8.47275794e-01
3.11712205e-01 1.46578997e-01 -2.38523990e-01 -9.11922812e-01
-5.75103223e-01 1.00377953e+00 1.85237646e-01 6.13716006e-01
-5.99344194e-01 -4.35840040e-01 -6.39009196e-03 -2.46550292e-01
9.57900435e-02 4.23814565e-01 -5.30974746e-01 1.71139002e-01
-3.34568888e-01 3.18954825e-01 8.29589367e-01 -1.51162326e-01
9.48807716e-01 -6.85732245e-01 7.45456899e-03 2.63199389e-01
1.56654790e-01 4.18868780e-01 -1.72603607e-01 6.39824197e-02
4.52864356e-03 7.30650127e-01 4.13808018e-01 -9.47485924e-01
-6.68845952e-01 -2.75114775e-01 -7.36554682e-01 3.14190924e-01
5.70588112e-01 -6.06503412e-02 -5.57451189e-01 1.66110098e+00
5.67803085e-01 -2.11634740e-01 1.08424082e-01 7.92366982e-01
3.46241742e-01 3.86408836e-01 4.05448049e-01 -7.33221173e-01
1.57591963e+00 -1.64944842e-01 -3.72222871e-01 5.51521361e-01
3.84714723e-01 -8.08896184e-01 5.47456741e-01 2.51138538e-01
-7.52231121e-01 -3.23230177e-01 -8.73653352e-01 6.65027916e-01
-4.12482649e-01 2.00109817e-02 2.93730855e-01 4.97573793e-01
-7.22182095e-01 8.24330091e-01 -6.18371367e-01 -2.67462581e-01
-1.34503275e-01 6.74826086e-01 -5.80709934e-01 3.02663714e-01
-1.29149437e+00 1.03897917e+00 3.63234043e-01 1.02168597e-01
-4.30915207e-01 -3.20242077e-01 -4.16336358e-01 5.43553606e-02
2.81628463e-02 -7.30407774e-01 4.39068884e-01 -9.08765972e-01
-1.30562699e+00 1.02793372e+00 -3.60357791e-01 -3.19799334e-01
2.55910248e-01 3.47645789e-01 -2.62481272e-01 4.29686084e-02
-2.85554767e-01 1.66650206e-01 5.46107769e-01 -1.32547760e+00
7.76998699e-02 -8.04170847e-01 -2.73876160e-01 -5.27832322e-02
4.22520101e-01 2.32793197e-01 1.27432525e-01 -3.78410488e-01
-5.87357953e-02 -1.00620329e+00 -5.21672964e-01 -7.76041925e-01
3.73999104e-02 2.02494487e-02 6.30554110e-02 -3.21960866e-01
8.10114980e-01 -1.87922645e+00 5.06570995e-01 9.91672039e-01
2.42431462e-01 6.26627505e-01 7.23554194e-02 9.31140840e-01
-3.02047849e-01 -2.57833805e-02 -2.40424752e-01 3.85276258e-01
-3.82539779e-01 -3.01039368e-01 -1.94365889e-01 6.74846113e-01
1.34932399e-01 7.26846218e-01 -5.88337302e-01 -2.77242035e-01
1.41424596e-01 7.55436718e-01 -8.52796435e-01 3.02004397e-01
-6.38170302e-01 9.89205539e-01 -7.57401049e-01 1.65047631e-01
1.02405834e+00 -7.40788430e-02 7.21560538e-01 -2.33552784e-01
-3.36617440e-01 -2.22038273e-02 -7.76846111e-01 1.03811026e+00
5.02663516e-02 -9.43494290e-02 -1.49392724e-01 -5.29933453e-01
1.11668992e+00 3.16811413e-01 6.41000330e-01 -7.10472763e-01
3.08885187e-01 4.26903576e-01 4.54865485e-01 -1.62300140e-01
-8.36058110e-02 -3.35014224e-01 3.34854931e-01 8.77460241e-02
-3.03271919e-01 3.31344604e-01 2.26243585e-01 -2.71074951e-01
7.33751476e-01 -3.87422144e-02 6.10450923e-01 -5.34304082e-01
1.13112295e+00 1.86578155e-01 1.83301538e-01 3.27286184e-01
1.13508947e-01 3.39574784e-01 7.72446454e-01 -7.54518211e-01
-1.14381897e+00 -8.29400420e-01 -4.30680215e-01 6.90454543e-01
-7.77573809e-02 -4.89302464e-02 -1.04610801e+00 -2.09499344e-01
5.12946323e-02 2.66411543e-01 -5.96758306e-01 4.27289121e-02
-7.07206488e-01 -1.03231490e+00 1.40329391e-01 -3.62176597e-01
-2.11030811e-01 -1.02499723e+00 -5.70623636e-01 4.14673924e-01
2.33720765e-01 -5.58821261e-01 -2.18412966e-01 3.60945672e-01
-7.26552725e-01 -1.51684535e+00 -6.56807363e-01 -3.49123836e-01
6.16441131e-01 -1.43750012e-01 8.03583801e-01 1.72392244e-03
-3.06809008e-01 -1.84846476e-01 -3.09797525e-01 -3.48903984e-01
-5.97847819e-01 -1.67992935e-01 8.53647292e-02 1.70976326e-01
5.08815229e-01 -5.24466872e-01 -8.48432064e-01 4.37900782e-01
-8.42343152e-01 -2.83381164e-01 6.58789873e-01 7.10802019e-01
7.66544700e-01 -4.89619821e-01 4.41724896e-01 -8.80010426e-01
6.74364865e-01 -2.89734125e-01 -8.74392271e-01 2.00090110e-01
-2.08226204e-01 5.29078007e-01 6.85264826e-01 -1.81571573e-01
-7.74109721e-01 4.20723289e-01 -6.06488526e-01 1.16705552e-01
-1.98527589e-01 4.13721234e-01 -4.81198817e-01 -4.47087586e-01
5.21215022e-01 3.33341092e-01 3.21687162e-01 -3.99019659e-01
-3.40254456e-02 2.14330137e-01 -1.43975139e-01 -6.46565020e-01
5.66293895e-01 4.49113846e-01 3.10102910e-01 -8.34024012e-01
6.70718625e-02 -4.94961321e-01 -3.73482436e-01 -6.41551316e-02
1.07022822e+00 -3.55558783e-01 -1.39904356e+00 2.10172862e-01
-1.54994905e+00 1.37681603e-01 2.87243515e-01 6.13338888e-01
-4.56203073e-01 5.95995784e-01 -1.89147472e-01 -6.32568657e-01
-3.37625116e-01 -1.42621815e+00 8.09711099e-01 -1.28445789e-01
-1.60670310e-01 -7.46372104e-01 1.09470308e+00 1.38523355e-01
2.76911020e-01 3.25358331e-01 1.32299149e+00 -1.07031322e+00
-5.71394622e-01 -4.50980403e-02 1.28910718e-02 -1.45313546e-01
-3.78384627e-02 8.78618062e-02 -5.59643209e-01 -2.45203570e-01
-9.54001769e-02 1.88966513e-01 7.82888770e-01 4.50075030e-01
3.68058920e-01 8.41500163e-02 -4.35422957e-01 3.02881271e-01
1.46673918e+00 6.69275224e-01 7.23577201e-01 1.92892253e-01
2.11503476e-01 7.19414413e-01 7.97006845e-01 5.44337153e-01
-2.67043471e-01 1.21667111e+00 3.72139841e-01 -2.20149264e-01
7.53855586e-01 2.19447061e-01 2.31918648e-01 1.97251022e-01
-5.97352743e-01 -1.84807718e-01 -8.80077958e-01 -5.79373166e-02
-1.64049876e+00 -8.73161554e-01 -3.91119927e-01 2.45446014e+00
8.00572276e-01 1.64829254e-01 2.99183458e-01 -4.73330826e-01
7.09456623e-01 -2.69173115e-01 -2.21142933e-01 -6.09470725e-01
-1.96572930e-01 5.52054405e-01 2.73856074e-01 9.76223648e-01
-7.42937863e-01 6.20069683e-01 5.97241783e+00 7.24346757e-01
-1.03477430e+00 -1.38374358e-01 2.21279830e-01 3.09756935e-01
-5.63727558e-01 2.68814832e-01 -8.30860257e-01 3.68519813e-01
7.70208478e-01 2.86223888e-01 1.37394205e-01 3.88322890e-01
4.15636510e-01 -1.80462465e-01 -8.46694112e-01 5.66652477e-01
-2.23950237e-01 -1.53352547e+00 2.66262263e-01 2.78816074e-01
4.68455583e-01 -3.35728884e-01 -2.15007126e-01 -2.91984648e-01
-2.11895153e-01 -1.00135219e+00 4.87351343e-02 7.70323694e-01
4.32252407e-01 -7.98739851e-01 8.37980866e-01 2.19743565e-01
-1.13299060e+00 4.36019331e-01 -1.31421477e-01 8.33633021e-02
2.19010234e-01 5.96994698e-01 -1.00252688e+00 2.61298805e-01
6.00798912e-02 -2.59708554e-01 -3.27620029e-01 8.19079995e-01
2.16695905e-01 -5.21853417e-02 -1.70379341e-01 -4.01401341e-01
2.56988443e-02 -8.82701635e-01 6.80828869e-01 1.03633893e+00
-9.12859291e-02 2.91651815e-01 -1.06841370e-01 8.90174210e-01
2.70036638e-01 7.10288882e-01 -6.50831580e-01 5.74168190e-02
9.56191719e-02 9.07402277e-01 -7.19336212e-01 -1.59240738e-01
-8.59591141e-02 6.71107531e-01 3.03725660e-01 2.14783311e-01
-7.16282964e-01 -1.08346589e-01 1.01608586e+00 2.51038194e-01
3.63934875e-01 -1.77331299e-01 4.17522609e-01 -8.88170183e-01
-2.49702349e-01 -1.00327003e+00 2.53234267e-01 -2.75215715e-01
-8.61752093e-01 6.60022199e-01 -5.60899414e-02 -7.78512359e-01
-2.21928880e-01 -7.33782351e-01 -3.81826580e-01 1.21035182e+00
-1.08124197e+00 -7.33346403e-01 1.51181668e-01 4.37372416e-01
-1.48592265e-02 6.64913356e-02 1.04912221e+00 2.13723108e-01
-3.98906082e-01 2.46891845e-03 5.61387420e-01 -5.35700798e-01
6.38731658e-01 -8.62404883e-01 1.52959034e-01 3.75115782e-01
-2.88995564e-01 1.01449060e+00 1.18931913e+00 -9.42380309e-01
-1.18773127e+00 -4.64427173e-01 1.00877142e+00 -4.41483945e-01
5.35609603e-01 -4.95828390e-01 -8.71554613e-01 3.48126918e-01
-5.79282753e-02 -5.85595429e-01 9.67154682e-01 -1.59136966e-01
-5.65003455e-02 2.71307021e-01 -9.94383812e-01 5.76368272e-01
5.55113375e-01 -3.39840204e-01 -5.93496621e-01 2.31795534e-01
3.11032593e-01 -7.00944960e-02 -8.20136547e-01 6.51629090e-01
6.14509642e-01 -1.23846948e+00 1.07599831e+00 -9.35571134e-01
-3.73036623e-01 -5.23684382e-01 3.20607424e-02 -9.90592718e-01
-3.00130576e-01 -3.98903161e-01 4.93517846e-01 6.39636993e-01
3.25562090e-01 -8.37469101e-01 7.96292245e-01 2.04839230e-01
1.68686211e-01 -6.50031567e-01 -9.65127110e-01 -3.75550687e-01
-9.94853023e-03 2.58620054e-01 4.21116680e-01 3.04653466e-01
-6.47389442e-02 3.79770786e-01 -8.91930535e-02 1.27267569e-01
4.22446609e-01 4.16264594e-01 7.99476981e-01 -1.40469539e+00
-3.48705947e-01 -3.68247807e-01 -4.26352233e-01 -2.81430960e-01
1.97144777e-01 -6.31754458e-01 -3.98768514e-01 -1.00842094e+00
3.87661755e-01 -3.47213268e-01 -2.23512843e-01 -2.06592903e-01
-3.45677622e-02 8.09064656e-02 4.61863801e-02 3.20523798e-01
-2.93771505e-01 4.26816821e-01 9.53005195e-01 1.11707650e-01
-3.84452105e-01 1.57609984e-01 -1.66643322e-01 5.80596387e-01
6.75098002e-01 -5.33152580e-01 5.93711697e-02 6.13865316e-01
5.64180136e-01 5.49736358e-02 2.12800935e-01 -6.49670064e-01
-1.77605793e-01 -3.70002866e-01 1.90646648e-01 -7.62550712e-01
4.36396807e-01 -1.14397740e+00 9.30062115e-01 1.08283544e+00
1.88014835e-01 -1.01288110e-01 1.85337245e-01 3.36744487e-01
-7.26521313e-02 -2.87836641e-01 9.27530885e-01 1.70363814e-01
-3.74710739e-01 4.88439985e-02 -7.90315211e-01 -2.76905805e-01
1.15692711e+00 -3.14653158e-01 -1.19368453e-02 2.76840806e-01
-1.05190229e+00 -4.73012716e-01 8.78248990e-01 -1.88395277e-01
5.43477893e-01 -7.97238350e-01 -5.97419202e-01 2.99479753e-01
3.71755034e-01 -8.61576676e-01 4.69023138e-01 1.10486186e+00
-9.92735147e-01 9.14774477e-01 -5.76149821e-01 -6.33099139e-01
-1.62102151e+00 1.04901874e+00 6.21016204e-01 -4.59420711e-01
2.36878693e-01 1.41741574e-01 5.72845399e-01 -3.07404310e-01
-2.88455039e-01 -7.10994825e-02 -7.25013256e-01 1.79974332e-01
2.20477298e-01 1.52500331e-01 3.48952949e-01 -9.05573905e-01
-8.47324610e-01 1.14582098e+00 1.38067761e-02 3.11206847e-01
1.18467808e+00 2.02167794e-01 -3.73839498e-01 -6.44999789e-03
1.15811527e+00 4.71786052e-01 -7.71685302e-01 6.44200593e-02
6.58942014e-02 -2.07314387e-01 -7.22448826e-01 -4.71802831e-01
-2.21708819e-01 4.40778553e-01 8.37057650e-01 -1.14701986e-01
8.08692396e-01 6.96797445e-02 1.88208014e-01 4.21362430e-01
6.74449742e-01 -4.47895616e-01 -2.43737414e-01 4.37588036e-01
8.71362448e-01 -1.01290154e+00 -1.48455188e-01 -4.43117559e-01
-3.66903841e-01 9.60791111e-01 -1.42785832e-02 -1.99805275e-02
3.84421378e-01 -2.39991769e-02 -3.19942459e-02 -5.45463741e-01
-6.52812779e-01 -4.60633159e-01 3.66373688e-01 2.81855732e-01
6.55294597e-01 1.56046674e-01 -1.29064584e+00 2.46548742e-01
3.43297869e-01 3.67629081e-02 2.83647962e-02 6.53474569e-01
-6.13686442e-01 -1.78970242e+00 -7.20415175e-01 -1.77789479e-01
-3.76891255e-01 -2.16104701e-01 -8.63752246e-01 8.12997460e-01
1.04997352e-01 5.55717111e-01 -2.38352552e-01 -2.15344936e-01
4.07804787e-01 1.29280627e-01 5.67268789e-01 -2.61383951e-01
-7.61876702e-01 2.21983835e-01 -9.73256025e-03 -1.64878070e-01
-3.20936263e-01 -5.35212278e-01 -1.34597838e+00 -2.36788139e-01
-2.76207238e-01 8.56367409e-01 7.81565487e-01 9.04041350e-01
5.22240043e-01 5.73189743e-02 7.55745828e-01 -9.17698979e-01
-2.12133825e-01 -6.08136177e-01 -5.50522506e-01 5.86240888e-01
1.23115905e-01 -6.83878362e-01 -2.09370628e-01 -3.70284230e-01] | [4.769105911254883, 5.343353748321533] |
2bf98a97-b570-406b-8be3-8761bea12241 | addressing-bias-in-face-detectors-using | 2210.16024 | null | https://arxiv.org/abs/2210.16024v1 | https://arxiv.org/pdf/2210.16024v1.pdf | Addressing Bias in Face Detectors using Decentralised Data collection with incentives | Recent developments in machine learning have shown that successful models do not rely only on huge amounts of data but the right kind of data. We show in this paper how this data-centric approach can be facilitated in a decentralized manner to enable efficient data collection for algorithms. Face detectors are a class of models that suffer heavily from bias issues as they have to work on a large variety of different data. We also propose a face detection and anonymization approach using a hybrid MultiTask Cascaded CNN with FaceNet Embeddings to benchmark multiple datasets to describe and evaluate the bias in the models towards different ethnicities, gender, and age groups along with ways to enrich fairness in a decentralized system of data labeling, correction, and verification by users to create a robust pipeline for model retraining. | ['Richard Blythman', 'Robin Lehmann', 'M. R. Ahan'] | 2022-10-28 | null | null | null | null | ['face-detection'] | ['computer-vision'] | [-3.68347466e-01 1.03404447e-01 -1.21666320e-01 -1.05414701e+00
-3.03093642e-01 -5.77787280e-01 7.69131660e-01 1.46121398e-01
-8.41172159e-01 5.29183030e-01 2.41913959e-01 -2.82166954e-02
1.78628415e-01 -6.86164975e-01 -5.17428935e-01 -3.55714649e-01
2.87977934e-01 7.62257993e-01 -2.02316284e-01 -1.67002290e-01
-2.65364926e-02 8.80105019e-01 -1.40387368e+00 3.48216712e-01
3.80792618e-01 7.26692855e-01 -7.05470264e-01 3.71007800e-01
4.58347127e-02 3.15199286e-01 -5.37330925e-01 -1.10724247e+00
8.17667842e-01 6.25519156e-02 -6.37289822e-01 -2.87447929e-01
1.11998057e+00 -8.20801973e-01 -1.47624314e-01 9.05956089e-01
9.35743570e-01 -2.93939263e-01 4.42095965e-01 -1.63516879e+00
-7.32007325e-01 5.92904031e-01 -6.40613437e-01 -6.05572611e-02
-6.32088706e-02 2.24249035e-01 6.41305804e-01 -9.30047452e-01
7.58725703e-01 1.63827050e+00 9.81268287e-01 1.02293980e+00
-1.32531273e+00 -1.09186280e+00 -2.43915439e-01 -1.09677598e-01
-1.30794883e+00 -8.20238769e-01 4.93581146e-01 -5.19460499e-01
5.51949143e-01 1.08137965e-01 3.74927819e-01 1.31481075e+00
-3.47838223e-01 4.08481061e-01 1.00491571e+00 -3.04499149e-01
2.13998575e-02 7.58325994e-01 1.13313660e-01 8.23153377e-01
5.34663796e-01 2.21105620e-01 -5.38138866e-01 -4.86743450e-01
5.53036392e-01 1.08981282e-02 2.36160159e-01 -5.75693965e-01
-6.17756546e-01 1.17012072e+00 2.85252452e-01 1.16692290e-01
-6.43478781e-02 6.85313046e-02 7.55327225e-01 3.56459230e-01
6.05427265e-01 4.27693099e-01 -6.50453687e-01 5.53903997e-01
-1.13870633e+00 4.51033324e-01 7.71089733e-01 6.24172330e-01
9.32461679e-01 -1.00080259e-01 -3.95599365e-01 6.69065237e-01
3.38458806e-01 4.74071115e-01 3.38367075e-01 -1.05344915e+00
3.94593984e-01 6.99458420e-01 4.43558656e-02 -9.73064721e-01
-4.34439003e-01 -2.64596611e-01 -8.58463585e-01 4.94804770e-01
7.32738376e-01 -4.00403351e-01 -7.38071322e-01 1.94974673e+00
6.15202367e-01 -5.17812632e-02 -3.77480835e-01 6.04206920e-01
5.50459445e-01 -1.07799418e-01 4.53238130e-01 4.00538355e-01
1.18808722e+00 -6.99459732e-01 -4.50030267e-01 3.61197367e-02
7.37048209e-01 -6.75316513e-01 8.10095906e-01 1.18776642e-01
-9.73979056e-01 -4.46770608e-01 -6.67510688e-01 -3.91037881e-01
-7.22101152e-01 2.29691386e-01 7.52541184e-01 1.34719741e+00
-1.42942476e+00 6.72357678e-01 -4.27499563e-01 -5.34555137e-01
1.12186408e+00 7.99952388e-01 -8.91616940e-01 -1.29304886e-01
-9.17991638e-01 1.18128908e+00 1.48594752e-01 3.98200797e-03
-8.96949768e-01 -8.87738168e-01 -7.54762590e-01 1.15313001e-01
-2.40768772e-02 -7.48995245e-01 9.92736816e-01 -1.33289075e+00
-8.90080273e-01 1.34678948e+00 6.27947925e-03 -5.50119460e-01
8.45207393e-01 2.31270376e-03 -2.15420455e-01 -2.31503516e-01
-6.32360252e-03 9.51168656e-01 1.08446026e+00 -9.86774325e-01
-5.79530776e-01 -7.62590051e-01 -2.40881726e-01 -2.08573923e-01
-8.38258803e-01 5.67408025e-01 -4.52313125e-02 -3.50045145e-01
-6.25954807e-01 -7.73899198e-01 -1.43126994e-02 3.45538884e-01
-1.96421117e-01 -2.31478006e-01 1.05185890e+00 -8.16463411e-01
8.71101201e-01 -1.92118168e+00 -2.21275970e-01 3.22341442e-01
3.42632085e-01 7.15017617e-01 -2.92405456e-01 1.47050500e-01
-2.07259208e-01 4.76631910e-01 1.73255980e-01 -8.76004934e-01
2.07957000e-01 -8.69115144e-02 -7.43275285e-02 6.48456395e-01
4.13220912e-01 7.01046586e-01 -5.01293540e-01 -4.15628374e-01
1.02669187e-01 6.75547600e-01 -8.85612369e-01 7.24269524e-02
2.16611639e-01 4.19099569e-01 -1.05849043e-01 6.54520094e-01
1.05429721e+00 2.47534066e-01 1.83455825e-01 7.24927038e-02
-4.00871551e-03 -1.56453907e-01 -1.18000221e+00 1.29506218e+00
-1.69346213e-01 3.36132348e-01 7.19380558e-01 -7.67110169e-01
7.34551847e-01 3.11408341e-01 4.99972403e-01 -3.29408377e-01
4.48457956e-01 1.94122553e-01 -8.16275552e-02 -2.10537300e-01
4.18618321e-01 -7.02244863e-02 1.78524982e-02 6.99088395e-01
4.36642021e-01 2.13713810e-01 -2.47109737e-02 1.83560684e-01
8.39880705e-01 -2.71078646e-01 -2.70564621e-03 -4.25677866e-01
4.93115634e-01 -2.75354713e-01 8.26644897e-01 6.83898091e-01
-6.69896722e-01 5.15160441e-01 3.76424313e-01 -7.69813240e-01
-1.49054146e+00 -6.16572499e-01 -1.96449056e-01 1.41669142e+00
-6.96078002e-01 -2.29069263e-01 -1.01548731e+00 -9.29187775e-01
5.15412629e-01 2.25766405e-01 -8.79643440e-01 -2.06987888e-01
-4.16068405e-01 -8.74189317e-01 8.86211693e-01 4.06237185e-01
3.46371055e-01 -8.25262785e-01 -1.85571998e-01 -1.57392904e-01
3.47470105e-01 -8.45084369e-01 -4.30468619e-01 -2.64589787e-01
-3.90442967e-01 -1.25088131e+00 -6.25746608e-01 -5.99539340e-01
7.35659182e-01 -7.30217099e-02 1.12656260e+00 3.50152940e-01
-3.85065526e-01 3.39768142e-01 7.22525269e-02 -9.98751223e-01
-5.14859796e-01 1.96851805e-01 2.96750814e-01 3.66662413e-01
7.89357364e-01 -3.39371741e-01 -5.04440486e-01 2.78771669e-01
-8.75358164e-01 -2.66555130e-01 2.01600119e-01 7.25134313e-01
-1.30698651e-01 -3.99205208e-01 5.87243378e-01 -1.23094010e+00
5.81546366e-01 -5.03763616e-01 -7.08697021e-01 3.40025276e-01
-7.91432381e-01 -1.26310587e-01 3.69140565e-01 -2.40950167e-01
-9.20439303e-01 3.56906116e-01 -4.62984182e-02 -4.12065595e-01
-2.54460812e-01 -2.62093246e-01 -2.51430720e-01 -4.25944000e-01
7.67183959e-01 -5.11689484e-01 3.21110547e-01 -6.09398067e-01
4.96973366e-01 8.66465628e-01 1.72752410e-01 -4.68363047e-01
9.22756493e-01 7.28315175e-01 -3.14690992e-02 -5.32342553e-01
-5.12472272e-01 -3.18045855e-01 -8.82526994e-01 5.17717376e-03
6.95738375e-01 -1.13788843e+00 -7.45631993e-01 7.21136034e-01
-1.12615502e+00 -1.84602767e-01 -3.49752218e-01 3.01620483e-01
-5.58299050e-02 1.09133020e-01 -5.09810865e-01 -6.52422309e-01
-5.12274802e-01 -8.03372681e-01 9.02014911e-01 3.25356960e-01
-2.07588851e-01 -9.21950877e-01 1.04371130e-01 5.03574967e-01
7.07743227e-01 1.10790983e-01 6.70102894e-01 -1.06897557e+00
-3.20473343e-01 -4.01846707e-01 -5.22259235e-01 4.35022891e-01
-9.77413822e-03 3.45691770e-01 -1.36908603e+00 -5.12935519e-01
-2.84405231e-01 -5.85139453e-01 7.47489333e-01 2.19740987e-01
1.34752190e+00 -3.43394548e-01 -2.24971354e-01 8.46320212e-01
1.31962800e+00 -6.40817404e-01 3.46292138e-01 -1.98898278e-02
8.59493732e-01 9.61558640e-01 -5.03969193e-02 3.82854134e-01
5.30829310e-01 4.68217462e-01 4.22548085e-01 -4.26247925e-01
-1.28061190e-01 -3.06364328e-01 -6.45568520e-02 4.11985666e-02
1.38353735e-01 2.21513554e-01 -9.74362910e-01 6.65251195e-01
-1.69201970e+00 -1.02188385e+00 -4.16816026e-02 2.19469881e+00
8.62819791e-01 -4.88459945e-01 5.34405172e-01 -3.46878737e-01
9.75533307e-01 -2.36088991e-01 -4.27034378e-01 -7.27847934e-01
-1.56946346e-01 4.45246369e-01 7.96479225e-01 3.67607564e-01
-1.18699193e+00 1.07022750e+00 7.24561262e+00 4.64926779e-01
-1.20289183e+00 3.50952774e-01 8.58609200e-01 -4.15873110e-01
-1.25436127e-01 -1.68910295e-01 -1.04923165e+00 4.62021470e-01
1.05816567e+00 -1.49868771e-01 4.55961287e-01 8.52688730e-01
6.81638494e-02 1.16204105e-01 -1.09948933e+00 9.57318127e-01
1.72467008e-01 -1.31987488e+00 3.79238538e-02 2.34436795e-01
7.91853607e-01 1.85704082e-01 1.51332006e-01 1.71909273e-01
8.33396494e-01 -1.28036904e+00 4.38624501e-01 2.96167761e-01
7.65786648e-01 -6.35295212e-01 6.54976487e-01 1.39876559e-01
-6.26969457e-01 -5.42636752e-01 -5.61721683e-01 -1.21720582e-02
-2.40201026e-01 4.14239973e-01 -8.92193496e-01 2.21623257e-01
8.94954801e-01 2.62316227e-01 -9.43675101e-01 1.00049722e+00
1.25727504e-01 2.39870608e-01 -4.25584167e-01 3.53138566e-01
-2.18101695e-01 1.33296728e-01 -2.58520376e-02 1.02344990e+00
2.11871311e-01 -3.37648451e-01 -9.68148708e-02 6.81782067e-01
-7.08933115e-01 2.95259327e-01 -7.65822589e-01 1.06189780e-01
5.18853188e-01 1.63016939e+00 -3.45818698e-01 -1.25241056e-01
-6.39236629e-01 5.56570470e-01 6.65690541e-01 -1.30443741e-02
-5.08328438e-01 1.10400774e-01 9.33543861e-01 2.80906051e-01
1.33749992e-01 4.42220317e-03 -2.72393584e-01 -1.22263408e+00
-1.96798697e-01 -9.95728731e-01 6.93962276e-01 -3.25940400e-01
-1.64004970e+00 2.02835470e-01 -2.16056064e-01 -4.16678220e-01
-4.97422889e-02 -7.17799842e-01 -5.53573430e-01 1.04397488e+00
-1.70777702e+00 -1.49458992e+00 -2.12693483e-01 1.00341725e+00
-2.97071069e-01 -5.99266648e-01 9.88036513e-01 7.35729337e-01
-8.04044008e-01 1.08763993e+00 3.98166776e-02 6.86537623e-01
1.33254254e+00 -1.01586294e+00 3.55793059e-01 6.32808089e-01
4.77362750e-03 7.58261025e-01 4.44478750e-01 -4.98316109e-01
-9.80001986e-01 -1.28515112e+00 1.07607245e+00 -8.83557260e-01
3.41399997e-01 -6.21790349e-01 -6.87412143e-01 9.62713003e-01
1.85994029e-01 2.95236498e-01 9.86688256e-01 4.23419267e-01
-8.03747237e-01 -4.79456186e-01 -1.73043060e+00 3.08566153e-01
8.43996108e-01 -6.57663763e-01 -2.18096897e-01 3.95759284e-01
1.10074826e-01 3.91591638e-02 -5.53438067e-01 1.57684818e-01
5.80211222e-01 -1.21892905e+00 7.83363521e-01 -1.14768004e+00
6.83980584e-02 5.93595095e-02 7.68003687e-02 -1.06204712e+00
-2.18311459e-01 -6.76996112e-01 2.22433880e-01 1.73879218e+00
2.60035545e-01 -8.38124692e-01 1.01235712e+00 1.38873541e+00
5.75040817e-01 -2.14225650e-01 -9.12276268e-01 -3.77765983e-01
5.49764276e-01 3.06894304e-03 1.08764899e+00 1.27768719e+00
-5.53646982e-01 8.51483122e-02 -5.65268099e-01 -2.32317075e-02
7.62699246e-01 -2.63627529e-01 1.11356103e+00 -1.65628386e+00
2.56517202e-01 -3.67112786e-01 -4.89593506e-01 2.56403118e-01
5.85498452e-01 -1.01557779e+00 -6.07728958e-01 -9.80879664e-01
4.10017103e-01 -6.64473653e-01 -2.46264458e-01 7.49073684e-01
5.96272014e-02 5.11686683e-01 1.54793203e-01 3.83136198e-02
-2.01593354e-01 2.64813006e-01 5.96881270e-01 3.41060795e-02
1.31322101e-01 -1.61375791e-01 -1.18997979e+00 6.86854839e-01
7.96895027e-01 -7.45017827e-01 -1.01256840e-01 -7.72147715e-01
2.02112660e-01 -7.29273915e-01 6.42377555e-01 -8.49419832e-01
1.99035749e-01 6.97581246e-02 9.53308523e-01 1.05473898e-01
2.39248857e-01 -1.05380785e+00 -6.18857369e-02 2.99312651e-01
-2.74377733e-01 1.61190242e-01 2.04547927e-01 2.21795052e-01
1.61665961e-01 -4.89207096e-02 1.03201771e+00 -2.86503524e-01
-2.68420964e-01 8.10623288e-01 2.15752751e-01 -6.01834804e-03
1.14481843e+00 1.30816773e-01 -4.18390661e-01 -2.39099398e-01
-6.48267686e-01 4.04062063e-01 8.63120079e-01 4.75909173e-01
-7.49194771e-02 -1.33610165e+00 -1.06061614e+00 6.44411087e-01
-1.14212990e-01 -5.18317461e-01 1.20377921e-01 5.06070495e-01
-3.12017441e-01 2.25030869e-01 -7.28748977e-01 -1.83135375e-01
-1.49595320e+00 4.43559855e-01 4.98928457e-01 -9.51500162e-02
-7.22495094e-03 9.90298271e-01 -1.65132254e-01 -9.53253508e-01
2.53528357e-01 3.93499821e-01 -9.92753543e-03 4.55619127e-01
6.57258153e-01 3.05290699e-01 1.93311110e-01 -8.42813253e-01
-4.57970440e-01 2.20619246e-01 -1.85990721e-01 -3.26674879e-02
1.53977978e+00 3.48950662e-02 -3.35664749e-01 -2.70766616e-01
1.18966186e+00 1.82476550e-01 -1.23629785e+00 -1.20990999e-01
-3.53013985e-02 -7.55679965e-01 -1.20985396e-01 -7.76414931e-01
-1.34868705e+00 1.00613225e+00 7.76440620e-01 -2.25068182e-02
7.74938703e-01 -2.77843356e-01 4.40494925e-01 3.63761634e-02
1.70008078e-01 -1.38328433e+00 -4.03872490e-01 2.20260277e-01
5.68178594e-01 -1.57941902e+00 1.88265949e-01 -6.04075529e-02
-4.14476544e-01 9.30189252e-01 7.26410091e-01 4.78902686e-04
9.02472794e-01 1.72725484e-01 3.88595521e-01 -1.01181678e-01
-4.67307270e-01 -3.42692481e-03 -1.38588715e-02 8.51474106e-01
2.78631538e-01 -9.17718187e-02 -8.65421519e-02 5.20242572e-01
-1.22090667e-01 8.13011453e-02 3.69342029e-01 6.31081283e-01
6.12617135e-02 -1.64415157e+00 -3.77642900e-01 4.80801374e-01
-8.01649094e-01 1.60731822e-01 -8.03220391e-01 7.53705025e-01
4.30780262e-01 7.96853304e-01 1.04219362e-01 -1.61571965e-01
1.33049890e-01 4.42209572e-01 4.45236683e-01 -7.48573422e-01
-1.15852964e+00 -5.30126154e-01 -2.28855945e-02 -6.92551970e-01
-2.63391495e-01 -7.53036618e-01 -6.03878975e-01 -8.58617663e-01
-7.90002868e-02 -1.33892700e-01 8.89017165e-01 7.20201910e-01
7.62061477e-01 -2.55773753e-01 7.53172338e-01 -8.41200590e-01
-7.88218141e-01 -8.67610455e-01 -6.18907034e-01 7.51561940e-01
3.18348944e-01 -3.80234003e-01 -2.87459046e-01 -2.05795631e-01] | [13.057777404785156, 0.9427109360694885] |
468a9d9b-f788-4883-8f57-a753cb69316b | empty-cities-a-dynamic-object-invariant-space | 2010.07646 | null | https://arxiv.org/abs/2010.07646v1 | https://arxiv.org/pdf/2010.07646v1.pdf | Empty Cities: a Dynamic-Object-Invariant Space for Visual SLAM | In this paper we present a data-driven approach to obtain the static image of a scene, eliminating dynamic objects that might have been present at the time of traversing the scene with a camera. The general objective is to improve vision-based localization and mapping tasks in dynamic environments, where the presence (or absence) of different dynamic objects in different moments makes these tasks less robust. We introduce an end-to-end deep learning framework to turn images of an urban environment that include dynamic content, such as vehicles or pedestrians, into realistic static frames suitable for localization and mapping. This objective faces two main challenges: detecting the dynamic objects, and inpainting the static occluded back-ground. The first challenge is addressed by the use of a convolutional network that learns a multi-class semantic segmentation of the image. The second challenge is approached with a generative adversarial model that, taking as input the original dynamic image and the computed dynamic/static binary mask, is capable of generating the final static image. This framework makes use of two new losses, one based on image steganalysis techniques, useful to improve the inpainting quality, and another one based on ORB features, designed to enhance feature matching between real and hallucinated image regions. To validate our approach, we perform an extensive evaluation on different tasks that are affected by dynamic entities, i.e., visual odometry, place recognition and multi-view stereo, with the hallucinated images. Code has been made available on https://github.com/bertabescos/EmptyCities_SLAM. | ['Jose Neira', 'Cesar Cadena', 'Berta Bescos'] | 2020-10-15 | null | null | null | null | ['steganalysis'] | ['computer-vision'] | [ 3.71477723e-01 1.50403261e-01 3.80132198e-01 -1.36437654e-01
-5.02429008e-01 -6.44752979e-01 6.72330081e-01 -4.04259712e-01
-6.06114566e-01 6.54548705e-01 -1.84160039e-01 2.43676156e-02
2.96047717e-01 -9.51167583e-01 -1.30054390e+00 -8.45748425e-01
1.38515443e-01 5.85679412e-01 5.35261631e-01 -3.59454036e-01
1.27021536e-01 7.90029705e-01 -1.80696416e+00 2.56467342e-01
6.89854681e-01 7.97161043e-01 6.10172629e-01 7.21331298e-01
2.81206876e-01 7.01156497e-01 -5.25308609e-01 -3.11402261e-01
6.48767769e-01 -2.62261391e-01 -4.86434519e-01 2.96705455e-01
4.91088927e-01 -4.26254660e-01 -6.74863040e-01 1.21016848e+00
1.78054452e-01 3.48093450e-01 3.01933855e-01 -1.36369789e+00
-3.61475587e-01 9.87493340e-03 -3.41371238e-01 1.35419965e-01
6.58431053e-01 4.10927385e-01 3.03993940e-01 -7.66126394e-01
9.89570975e-01 1.32057202e+00 5.35617054e-01 3.99605185e-01
-1.10913622e+00 -4.69999760e-01 -4.92618494e-02 5.37684083e-01
-1.52649188e+00 -6.49046183e-01 9.59943593e-01 -5.10911644e-01
4.59947526e-01 2.04235196e-01 6.24774456e-01 1.33285940e+00
2.84163952e-01 4.70294684e-01 1.17946446e+00 -5.49349129e-01
2.18176380e-01 3.38297367e-01 -5.53343296e-01 5.95744610e-01
2.05502838e-01 5.29714763e-01 -2.25326329e-01 3.60721834e-02
5.48420131e-01 2.34726280e-01 -4.35339779e-01 -7.22042322e-01
-1.03967798e+00 7.76311934e-01 5.53784192e-01 2.80593574e-01
-3.72424632e-01 1.44014105e-01 -9.72800478e-02 2.40884230e-01
2.73761451e-01 1.79318547e-01 -9.67030898e-02 2.56925911e-01
-9.70292926e-01 3.11912179e-01 6.62411690e-01 7.49270618e-01
1.15822029e+00 1.64125994e-01 2.39852756e-01 4.39701617e-01
2.29239747e-01 8.97312343e-01 4.13948566e-01 -1.05252302e+00
3.56722683e-01 1.63197801e-01 2.84952849e-01 -1.25866032e+00
-2.02518493e-01 -2.70406663e-01 -4.92479891e-01 6.58822000e-01
4.86125737e-01 -9.89845470e-02 -1.09466624e+00 1.72428083e+00
5.24530888e-01 3.39805692e-01 2.81265825e-01 1.05218256e+00
5.67537725e-01 6.78040385e-01 -3.75401706e-01 5.99584989e-02
1.01905167e+00 -9.42305088e-01 -5.27411997e-01 -7.11545169e-01
4.89984490e-02 -8.39485526e-01 5.51670730e-01 2.06231013e-01
-9.71737623e-01 -5.94830930e-01 -9.49687958e-01 -8.71183127e-02
-6.27054214e-01 -1.34883210e-01 1.16105370e-01 4.72698241e-01
-1.08249152e+00 3.74402344e-01 -6.81243062e-01 -4.81934816e-01
1.60937250e-01 1.06378578e-01 -6.23509586e-01 -3.61493170e-01
-1.26658976e+00 1.13761044e+00 4.87238497e-01 9.78789851e-02
-1.38166976e+00 -3.10368270e-01 -1.05621958e+00 -1.04054518e-01
4.09356445e-01 -6.43861592e-01 7.32404828e-01 -1.33953786e+00
-1.25811183e+00 9.64282751e-01 -1.58835873e-01 -5.47835708e-01
1.05825806e+00 -1.50811216e-02 -3.20240647e-01 3.27629507e-01
4.34437245e-01 7.15744317e-01 1.10022843e+00 -1.73438859e+00
-5.24275243e-01 -3.95808160e-01 5.09382226e-02 4.19591814e-01
3.33281249e-01 -3.10438156e-01 -5.77259123e-01 -4.91039008e-01
1.22177370e-01 -1.17563498e+00 -1.97498128e-01 1.30959511e-01
-3.98536921e-01 8.22340488e-01 1.00211048e+00 -8.79187942e-01
2.27002203e-01 -2.13004875e+00 2.59164602e-01 2.12597013e-01
-1.59037024e-01 2.35096946e-01 -1.87386394e-01 3.85368913e-01
-5.28515242e-02 -3.59377652e-01 -3.84089172e-01 -6.15868211e-01
-3.76362145e-01 4.10258889e-01 -4.06663835e-01 8.22788000e-01
-8.55063200e-02 8.64848077e-01 -8.73426139e-01 -2.34646678e-01
8.23883176e-01 7.01715648e-01 -3.85186225e-01 2.26614341e-01
-1.55948818e-01 1.00907362e+00 -1.41957805e-01 5.41497707e-01
1.03446984e+00 3.63110512e-01 -1.97884277e-01 7.51227885e-02
-1.85963809e-01 -1.85956284e-01 -1.40897119e+00 1.76647055e+00
-5.60291409e-01 7.25178123e-01 1.95991307e-01 -7.70524859e-01
8.05814445e-01 -5.10912351e-02 5.39173007e-01 -8.93484652e-01
1.66983545e-01 2.73532331e-01 -2.83423066e-01 -5.90027392e-01
6.69839680e-01 4.28033620e-02 -9.91706550e-03 -1.95811819e-02
-3.06014400e-02 -4.19702172e-01 5.45495152e-02 -7.22665340e-02
1.01543570e+00 1.12609044e-01 -6.62303269e-02 1.66803181e-01
6.52020216e-01 6.70873672e-02 4.89281893e-01 6.27922654e-01
-5.99411428e-02 1.04981303e+00 1.52385175e-01 -4.27678138e-01
-1.23337412e+00 -1.10912466e+00 9.52855200e-02 4.30959910e-01
6.66906536e-01 2.82766908e-01 -6.61179364e-01 -3.78667414e-01
-5.38259819e-02 8.37810993e-01 -5.78097343e-01 -4.13046122e-01
-5.11279643e-01 -3.40812176e-01 4.54411626e-01 -1.67067423e-01
7.00783372e-01 -1.04975712e+00 -1.03397059e+00 1.19575746e-01
-6.86906815e-01 -1.44728994e+00 -1.53244466e-01 9.39860940e-02
-4.21822578e-01 -1.13451242e+00 -7.02756703e-01 -5.21684825e-01
5.13215542e-01 5.54992199e-01 7.70665586e-01 -2.18418688e-01
-3.32615733e-01 3.70358944e-01 -4.60893065e-01 -2.51564890e-01
-6.82530284e-01 -3.26338828e-01 -7.96234980e-02 4.76585537e-01
-1.57258496e-01 -5.89769244e-01 -5.86387873e-01 3.28089297e-01
-1.30507076e+00 7.43034482e-02 5.98304212e-01 7.59681046e-01
5.30405819e-01 7.61832222e-02 -3.40230584e-01 -5.82319498e-01
-1.31956652e-01 -5.15666723e-01 -6.38031363e-01 -3.11706378e-03
-1.50146663e-01 -1.45859510e-01 6.11930847e-01 -4.52754617e-01
-1.00929928e+00 4.05951351e-01 -2.57691383e-01 -8.58539820e-01
-3.04235816e-01 -6.00788034e-02 -4.01804388e-01 -4.18875307e-01
6.62757993e-01 5.81799448e-01 1.45488366e-01 -1.64483398e-01
3.43722790e-01 3.98934126e-01 7.16946781e-01 -8.06727707e-02
1.20671701e+00 1.17316699e+00 6.01724302e-03 -8.97152841e-01
-4.89983410e-01 -3.99996549e-01 -6.37928903e-01 -4.46359247e-01
9.12318528e-01 -9.84116197e-01 -1.49967864e-01 6.55873179e-01
-1.40686119e+00 -4.62072939e-01 -4.52907681e-01 4.16534990e-01
-8.74375224e-01 4.04008508e-01 -2.41063491e-01 -4.77852374e-01
6.06595874e-02 -1.30950618e+00 1.17462766e+00 2.17304260e-01
3.40620697e-01 -9.20630336e-01 1.25753582e-01 5.01431584e-01
1.81380138e-01 6.63227499e-01 3.68551642e-01 -2.87886918e-01
-9.75380540e-01 -4.39521782e-02 1.30723238e-01 3.92431557e-01
-2.13393360e-01 -2.51794368e-01 -1.14430487e+00 -4.66664791e-01
1.32194266e-01 -9.90556255e-02 9.61489499e-01 3.04273397e-01
6.44220769e-01 -2.41474703e-01 -3.75546932e-01 1.00060689e+00
1.59948385e+00 2.65556455e-01 9.94712472e-01 4.97756302e-01
6.96675599e-01 6.32034898e-01 7.90089309e-01 2.62881637e-01
3.49484414e-01 1.03312075e+00 1.02808774e+00 -2.44113952e-01
-3.19365650e-01 -2.80832171e-01 5.50834715e-01 -7.57121146e-02
-3.33087556e-02 -4.16500598e-01 -8.06781054e-01 6.46443546e-01
-1.73846889e+00 -1.10839748e+00 -6.77453279e-02 2.21542478e+00
1.60516322e-01 -1.42277434e-01 -3.12439412e-01 -9.66074094e-02
7.72740602e-01 3.75005633e-01 -5.10886967e-01 -1.93394408e-01
-2.78271347e-01 -5.54506592e-02 7.88223743e-01 8.23606908e-01
-1.09306157e+00 1.06059790e+00 4.51427555e+00 5.69697142e-01
-1.38247156e+00 4.47935224e-01 3.12768579e-01 2.08810195e-01
-1.55427754e-01 1.59297273e-01 -5.03012836e-01 6.26774013e-01
6.93992376e-01 1.84796512e-01 5.75504065e-01 8.92830789e-01
3.05469602e-01 -4.54833984e-01 -7.68543363e-01 9.70184505e-01
5.61799228e-01 -1.20910418e+00 -1.38850048e-01 9.40677971e-02
8.65620315e-01 3.18608016e-01 1.90880910e-01 -5.78767285e-02
9.94442850e-02 -7.37432182e-01 1.27248824e+00 7.28761017e-01
7.52673090e-01 -6.28192306e-01 7.21286654e-01 4.78219062e-01
-8.66674840e-01 -7.88052753e-02 -3.26472014e-01 1.01896070e-01
4.07263517e-01 4.81831133e-01 -7.25483179e-01 7.89456189e-01
7.21729517e-01 4.94931668e-01 -5.63547373e-01 1.05702126e+00
-3.48965049e-01 -2.30458081e-01 -3.50954533e-01 5.59773624e-01
2.04165488e-01 -2.68614769e-01 1.00887895e+00 9.67191875e-01
4.72593039e-01 -2.47623891e-01 1.23462014e-01 1.11058486e+00
2.50692815e-01 -3.66803765e-01 -1.07025349e+00 6.75205946e-01
2.60685116e-01 1.10221148e+00 -7.09341824e-01 -8.77898782e-02
-2.37781599e-01 1.46986735e+00 -8.83818325e-03 5.28990567e-01
-1.02320564e+00 -1.90427706e-01 5.60088456e-01 3.10472548e-01
4.63765532e-01 -2.70131469e-01 1.25924259e-01 -1.47759390e+00
1.11010820e-01 -6.85945094e-01 -6.60943706e-03 -1.10395551e+00
-5.92219174e-01 7.71944582e-01 4.80794087e-02 -1.30335379e+00
-5.26257277e-01 -2.38937423e-01 -4.76837456e-01 8.33732486e-01
-1.62266433e+00 -1.30845094e+00 -6.98692679e-01 8.93907249e-01
5.14152408e-01 -2.05916911e-02 4.97442603e-01 4.61249083e-01
-5.36416210e-02 8.05650353e-02 3.12946737e-01 3.51258107e-02
6.18761122e-01 -8.20908725e-01 2.29152367e-01 1.44660330e+00
1.43501729e-01 -6.27740771e-02 1.03050387e+00 -5.24941325e-01
-1.20525897e+00 -1.37262917e+00 7.22089946e-01 -4.23327625e-01
2.00692028e-01 -4.88906950e-01 -6.93831682e-01 8.21648180e-01
-1.85433388e-01 2.27757305e-01 -2.27776572e-01 -8.81409287e-01
-6.62432984e-02 -8.98689628e-02 -1.41313732e+00 5.09703100e-01
8.73431444e-01 -5.00756800e-01 -4.13258761e-01 3.26384068e-01
5.14336050e-01 -7.74825573e-01 -2.26551384e-01 3.36308390e-01
3.45982999e-01 -1.38768017e+00 1.19037092e+00 1.05015934e-02
2.77201772e-01 -5.54096222e-01 -3.09080929e-01 -1.38526130e+00
2.04906225e-01 -4.39071476e-01 1.57867655e-01 9.06968594e-01
-4.30281833e-02 -7.71608055e-01 7.51891911e-01 2.71624833e-01
-1.71276286e-01 -1.58610880e-01 -1.18173432e+00 -7.91441143e-01
-1.52556226e-01 -3.97884011e-01 3.04025799e-01 7.44741559e-01
-8.91376019e-01 -9.47568044e-02 -6.62749469e-01 5.01695454e-01
6.48661733e-01 1.05262682e-01 1.10720372e+00 -8.66198063e-01
-3.51508528e-01 7.28528351e-02 -8.24335039e-01 -7.60560274e-01
2.07648471e-01 -6.83932483e-01 1.72767878e-01 -1.23842049e+00
-1.70467585e-01 -3.05383742e-01 2.75537282e-01 1.57613546e-01
2.53517836e-01 5.73322237e-01 4.38853681e-01 3.01203281e-01
-3.95003438e-01 6.85336828e-01 1.03970230e+00 -1.86985090e-01
-1.27370492e-01 1.33133262e-01 -6.45599365e-02 7.79497981e-01
6.03824973e-01 -6.96852386e-01 -2.75923699e-01 -4.61692899e-01
-1.32565871e-01 1.82745919e-01 9.38286901e-01 -1.37964165e+00
2.08483249e-01 -1.10319875e-01 4.34457332e-01 -3.78117561e-01
9.38988566e-01 -1.17523730e+00 6.76722825e-01 8.41638327e-01
1.55616641e-01 -2.46767756e-02 9.83783677e-02 6.27541959e-01
-3.31424117e-01 -3.39735001e-01 1.09637249e+00 -3.53437722e-01
-1.06579065e+00 2.20514387e-01 -2.60695189e-01 -1.10847451e-01
1.34081674e+00 -4.39297229e-01 -1.17905185e-01 -5.74175715e-01
-8.52014780e-01 -4.80871834e-02 9.50069189e-01 4.63385016e-01
8.16655517e-01 -9.99952018e-01 -5.92146218e-01 5.72314918e-01
1.54464349e-01 1.11122377e-01 4.02417094e-01 7.96043634e-01
-9.60012317e-01 1.20690353e-01 -5.23095429e-01 -5.66475451e-01
-1.20129550e+00 6.98475242e-01 6.38116419e-01 -1.23074174e-01
-6.73232853e-01 3.90774727e-01 2.56181210e-01 -4.37492192e-01
-8.67813826e-03 -7.94813707e-02 1.17699429e-02 -1.57202676e-01
4.11300838e-01 2.96795398e-01 1.36488080e-01 -1.09856653e+00
-2.41288751e-01 6.91889644e-01 3.29328984e-01 -4.00747746e-01
1.26970088e+00 -4.86965925e-01 1.55593678e-02 2.08132863e-01
1.26156831e+00 1.39885128e-01 -1.52175510e+00 -2.30465308e-01
-3.85281950e-01 -8.57158124e-01 -1.19231679e-01 -5.96988618e-01
-1.07742000e+00 7.35558093e-01 8.53373647e-01 -8.15228522e-02
9.94911730e-01 -2.92674452e-02 7.28331208e-01 1.44486949e-01
6.89038873e-01 -8.40515256e-01 9.75371972e-02 6.49772406e-01
9.51276124e-01 -1.24667335e+00 -3.44167233e-01 -4.24554944e-01
-6.22523248e-01 1.11028922e+00 4.12988752e-01 -2.86555320e-01
3.05665821e-01 9.21287537e-02 1.27505630e-01 4.41619158e-02
-1.50100306e-01 -4.79909718e-01 -7.79279843e-02 8.83348525e-01
-6.46606088e-01 -1.13935187e-01 7.68741518e-02 -2.44104564e-01
-2.18308970e-01 -2.59087503e-01 7.90440440e-01 7.79186606e-01
-2.12154478e-01 -8.24572444e-01 -8.47664297e-01 -2.79000431e-01
-1.49201885e-01 1.60588682e-01 -2.88083941e-01 9.82090950e-01
7.29734838e-01 9.19267952e-01 5.98927476e-02 -3.18326384e-01
3.94547552e-01 -1.21040627e-01 4.08971548e-01 -4.79816258e-01
-1.95799455e-01 -2.19101474e-01 -1.45919070e-01 -8.39395821e-01
-3.48987788e-01 -8.79110575e-01 -1.00335467e+00 -3.73956233e-01
7.58116916e-02 -1.51673719e-01 9.60159957e-01 8.34854007e-01
1.91869035e-01 4.37028974e-01 7.23067403e-01 -1.46945870e+00
-1.58493638e-01 -5.10064542e-01 -5.06607831e-01 6.39341652e-01
7.72873938e-01 -8.43223035e-01 -4.61213857e-01 1.19470656e-01] | [8.471084594726562, -2.225710391998291] |
7344f4d6-30d6-45d0-bbf1-b9a84698ed25 | dynamic-memory-induction-networks-for-few | 2005.05727 | null | https://arxiv.org/abs/2005.05727v1 | https://arxiv.org/pdf/2005.05727v1.pdf | Dynamic Memory Induction Networks for Few-Shot Text Classification | This paper proposes Dynamic Memory Induction Networks (DMIN) for few-shot text classification. The model utilizes dynamic routing to provide more flexibility to memory-based few-shot learning in order to better adapt the support sets, which is a critical capacity of few-shot classification models. Based on that, we further develop induction models with query information, aiming to enhance the generalization ability of meta-learning. The proposed model achieves new state-of-the-art results on the miniRCV1 and ODIC dataset, improving the best performance (accuracy) by 2~4%. Detailed analysis is further performed to show the effectiveness of each component. | ['Xiaodan Zhu', 'Yongbin Li', 'Ruiying Geng', 'Jian Sun', 'Binhua Li'] | 2020-05-12 | dynamic-memory-induction-networks-for-few-1 | https://aclanthology.org/2020.acl-main.102 | https://aclanthology.org/2020.acl-main.102.pdf | acl-2020-6 | ['few-shot-text-classification'] | ['natural-language-processing'] | [-2.97610968e-01 -2.77821928e-01 -7.27437556e-01 -3.02720219e-01
-1.46098718e-01 1.42837003e-01 3.96772712e-01 2.35605597e-01
-6.37240231e-01 5.67441642e-01 5.48529997e-02 -1.84101075e-01
-3.71282518e-01 -1.24515522e+00 -1.94902852e-01 -3.62905681e-01
1.51392147e-02 5.23938179e-01 8.07753086e-01 -4.10746843e-01
4.74194646e-01 2.52131611e-01 -1.83665931e+00 6.09888852e-01
7.52758980e-01 1.01373804e+00 2.79766589e-01 4.97961909e-01
-7.91997731e-01 9.55623806e-01 -4.07904953e-01 -2.10873470e-01
-3.83012086e-01 -2.42363140e-01 -9.77525413e-01 -3.45443249e-01
-1.74456149e-01 -2.79628009e-01 -6.21808827e-01 8.48386467e-01
4.37081665e-01 8.16541910e-01 6.80345476e-01 -9.11220551e-01
-5.12958765e-01 1.03187287e+00 -1.76674634e-01 7.41202891e-01
-2.00936198e-02 -7.01961592e-02 8.71149898e-01 -9.83306289e-01
6.22580767e-01 9.66923177e-01 2.63057023e-01 6.97225571e-01
-5.31795323e-01 -8.09826851e-01 1.35519728e-01 9.54921365e-01
-1.26772094e+00 -5.05401492e-01 7.79652059e-01 -2.60195732e-02
1.17031562e+00 -5.02776578e-02 4.49321717e-01 1.02972507e+00
1.85751170e-01 9.15241957e-01 4.87798512e-01 -8.23805571e-01
6.33004189e-01 2.69030094e-01 1.12490404e+00 6.82135582e-01
2.08751738e-01 -2.53673673e-01 -5.94948351e-01 -2.00930866e-04
1.79774076e-01 5.89993238e-01 1.29937172e-01 -3.05565950e-02
-3.59660238e-01 1.10808754e+00 4.85615820e-01 8.04464877e-01
-1.02674335e-01 -1.41382277e-01 7.64271617e-01 3.21901739e-01
5.49757540e-01 2.59116888e-01 -2.26235852e-01 -2.90832669e-01
-7.48180866e-01 -3.20630968e-01 6.87091231e-01 1.04519475e+00
6.78432465e-01 2.58530051e-01 -5.57106555e-01 1.12379146e+00
4.61940244e-02 -8.42736382e-03 1.18836010e+00 -2.90607393e-01
4.09666687e-01 9.66776848e-01 -4.08450305e-01 -7.55872548e-01
-6.25888109e-01 -5.04790068e-01 -9.60667074e-01 -3.88720870e-01
-5.25963545e-01 -5.58535196e-02 -1.11425090e+00 1.09821153e+00
-2.29909923e-03 4.25112993e-01 4.94163334e-02 3.69911551e-01
1.20174122e+00 7.81533778e-01 4.08296049e-01 -5.60139239e-01
1.25714505e+00 -1.08754182e+00 -8.18034172e-01 -3.25326502e-01
1.07191002e+00 -6.62216023e-02 1.01105738e+00 -4.10934575e-02
-6.36506736e-01 -6.81971610e-01 -1.31480384e+00 3.34503531e-01
-9.04518485e-01 -4.03307825e-01 8.15981865e-01 7.96901941e-01
-5.13583124e-01 5.24529099e-01 -6.32597864e-01 -4.52595472e-01
5.29968023e-01 1.39074996e-01 1.72545806e-01 -2.82564282e-01
-1.81048226e+00 7.48239636e-01 9.81689572e-01 -5.59879184e-01
-5.48908591e-01 -7.18344152e-01 -6.70438886e-01 3.86767775e-01
4.79586840e-01 -3.32690626e-01 1.10284376e+00 -2.82733083e-01
-1.47503865e+00 4.16521668e-01 -6.22312501e-02 -7.35586584e-01
1.10279664e-01 8.93320739e-02 -8.00780416e-01 3.22046280e-01
-2.89033413e-01 2.09497392e-01 5.08087397e-01 -6.30754173e-01
-5.85604489e-01 -4.95984405e-01 6.73591206e-03 9.16819349e-02
-1.14508128e+00 -6.65505007e-02 -5.87364316e-01 -4.83540356e-01
-1.29478633e-01 -5.81651628e-01 -4.38976176e-02 -6.66052163e-01
2.17823498e-02 -5.74262559e-01 1.04710054e+00 -2.76337694e-02
1.94545674e+00 -1.94947445e+00 -1.89825311e-01 -3.63356136e-02
-2.64123902e-02 7.05915868e-01 -4.48933057e-02 3.54829401e-01
2.38893017e-01 9.31595787e-02 3.24922472e-01 -2.06232011e-01
-2.17773899e-01 2.58206218e-01 -3.32630068e-01 -7.41525590e-02
-3.96820724e-01 1.17102194e+00 -6.48733437e-01 -7.06019163e-01
4.47407484e-01 -7.59404600e-02 -1.85254544e-01 9.67413634e-02
-3.48583698e-01 -4.38890338e-01 -4.84048158e-01 6.47075415e-01
4.29624289e-01 -3.89720827e-01 3.69650386e-02 4.20236811e-02
-1.24465153e-02 -8.50901306e-02 -1.05658948e+00 1.56293261e+00
-6.30127192e-01 3.75348628e-01 -7.50529408e-01 -1.11120760e+00
1.06379998e+00 1.44370794e-01 3.18396896e-01 -8.39390814e-01
4.60697442e-01 -2.54407167e-01 -1.24404788e-01 -3.25940460e-01
5.31327009e-01 -1.34451106e-01 -7.09849969e-02 5.28439999e-01
4.69472706e-01 4.53671157e-01 3.99136990e-01 2.51384795e-01
9.73223805e-01 -8.01083207e-01 4.02764589e-01 -1.12305924e-01
4.38338935e-01 -6.29296601e-02 2.98272401e-01 1.12237346e+00
-3.08148265e-01 1.33471832e-01 1.84767712e-02 -6.38051033e-01
-6.33343935e-01 -5.51123023e-01 -2.48530239e-01 1.81511593e+00
2.21217051e-01 -4.73761648e-01 -4.59523857e-01 -6.52621508e-01
-1.17079213e-01 1.48283494e+00 -8.29827607e-01 -7.89064407e-01
-1.16175570e-01 -1.04666984e+00 4.63344514e-01 7.88665175e-01
8.30756128e-01 -1.08776653e+00 -7.35639334e-01 4.09890831e-01
-7.51682278e-03 -8.23128462e-01 -1.08886182e-01 6.04300320e-01
-1.13857543e+00 -8.35433125e-01 -4.10760790e-01 -6.62032127e-01
9.15829465e-02 5.22366166e-01 7.12949753e-01 1.91598803e-01
-3.89271379e-01 2.29353502e-01 -9.30377841e-01 -5.00290930e-01
-1.58960134e-01 5.78960776e-01 2.09556550e-01 -2.01206625e-01
8.78687143e-01 -5.03136277e-01 -2.95541018e-01 3.84284496e-01
-1.05873394e+00 -1.39932737e-01 3.50097537e-01 9.78445947e-01
2.84731209e-01 3.41910660e-01 7.95529664e-01 -1.03019834e+00
7.05803394e-01 -7.20619917e-01 -2.60486305e-01 4.87929672e-01
-1.10012400e+00 5.64094968e-02 8.77909362e-01 -6.66150510e-01
-1.25747240e+00 -5.49993992e-01 -7.38392025e-02 -4.99219149e-01
-7.33150989e-02 6.60280049e-01 1.59880787e-01 -3.83194350e-02
7.78800488e-01 5.02080202e-01 -3.78075600e-01 -4.94208246e-01
3.52483392e-01 1.07696021e+00 6.57254159e-02 -1.91211700e-01
2.54511207e-01 2.90880322e-01 -1.41772136e-01 -1.03277910e+00
-1.09331095e+00 -8.64442289e-01 -9.05454099e-01 -8.77380073e-02
4.89902139e-01 -5.75019002e-01 -5.12167931e-01 2.92948425e-01
-8.91408861e-01 -1.23161688e-01 -6.90621063e-02 2.97788918e-01
-3.67500544e-01 1.47718653e-01 -8.42933536e-01 -8.65662456e-01
-8.99829030e-01 -5.79932034e-01 1.95885271e-01 4.51720685e-01
4.41880301e-02 -9.93016303e-01 1.58345431e-01 2.64669582e-02
5.85492194e-01 -6.53293610e-01 1.22882652e+00 -1.23465061e+00
-6.09998778e-02 -4.14444238e-01 -3.78749259e-02 -8.41067880e-02
-1.02679431e-01 -2.42965758e-01 -9.02569056e-01 -2.33492315e-01
-1.29423514e-01 -4.62334186e-01 1.29670572e+00 3.50734532e-01
1.40518963e+00 -7.11414847e-04 -5.52755177e-01 4.99643147e-01
1.47369087e+00 6.61276162e-01 8.57218683e-01 5.90612113e-01
3.12343001e-01 6.91180974e-02 7.03622878e-01 8.75586689e-01
8.64486792e-04 5.44617295e-01 2.27727249e-01 5.32437325e-01
2.89458558e-02 -5.79213500e-02 -7.17179701e-02 1.04869306e+00
7.61829540e-02 -2.26043671e-01 -9.85321283e-01 2.97000557e-01
-1.97923255e+00 -1.27778852e+00 4.18352753e-01 1.87692618e+00
5.85232794e-01 4.80729759e-01 6.97739189e-03 1.44544691e-01
8.02706718e-01 5.39645493e-01 -6.62901759e-01 -5.07226825e-01
1.42152533e-01 3.02850217e-01 1.62635773e-01 2.09714938e-02
-1.16831923e+00 1.23278689e+00 6.58236361e+00 1.23628664e+00
-9.86933053e-01 2.39370137e-01 6.25055313e-01 -2.57350564e-01
1.39837146e-01 -3.41131628e-01 -1.33465493e+00 5.73848128e-01
1.35701406e+00 -5.85340321e-01 2.76919812e-01 1.30241525e+00
-2.67181396e-01 1.18456908e-01 -5.12388408e-01 7.71693289e-01
3.18380713e-01 -1.77703094e+00 5.25525391e-01 -5.24256170e-01
6.70905948e-01 8.29522982e-02 -5.08459285e-02 9.99820232e-01
1.87329262e-01 -7.10055292e-01 -5.80559671e-02 4.81466860e-01
7.14189589e-01 -1.25603366e+00 9.77321029e-01 7.26479530e-01
-1.21391666e+00 -6.43196225e-01 -9.65341032e-01 -6.63946494e-02
-1.37527362e-01 4.03168082e-01 -7.15772927e-01 4.69383150e-01
6.51758611e-01 4.86874670e-01 -7.61003494e-01 8.66181493e-01
7.57288113e-02 6.19936705e-01 -3.62925604e-02 -7.80231178e-01
2.33692959e-01 2.54343361e-01 2.04826355e-01 1.33714664e+00
7.89014921e-02 4.93522286e-01 1.87609911e-01 3.60869139e-01
-3.17508519e-01 4.46467400e-01 -5.88919461e-01 -2.52737366e-02
7.84916103e-01 1.11778069e+00 -7.66446888e-01 -7.18587458e-01
-3.63523632e-01 7.65085518e-01 5.24848163e-01 6.35937154e-02
-6.00524068e-01 -9.28288400e-01 2.93193787e-01 -9.00478140e-02
4.43532377e-01 1.63782313e-01 -2.93174386e-01 -1.15557671e+00
-5.90720296e-01 -3.79188776e-01 1.03767288e+00 -5.56723475e-01
-1.05836642e+00 5.36936522e-01 1.20843731e-01 -1.05005074e+00
-3.49212885e-01 -4.62159634e-01 -9.26825523e-01 4.00074929e-01
-1.15569282e+00 -1.02120531e+00 -4.38095003e-01 4.74366844e-01
1.03007698e+00 -6.27328575e-01 1.07547641e+00 9.21500102e-02
-8.09543908e-01 9.22918558e-01 5.06761789e-01 1.77515954e-01
4.82103348e-01 -6.73559070e-01 2.43557781e-01 6.88519537e-01
1.83233142e-01 6.83629692e-01 3.10376048e-01 -6.30223274e-01
-1.43613231e+00 -1.20646048e+00 4.02570784e-01 -1.13151036e-01
6.59159005e-01 1.25329688e-01 -1.16488481e+00 5.63949943e-01
-1.37795627e-01 8.86123776e-02 9.63453054e-01 4.75766122e-01
-3.56636375e-01 -2.31050804e-01 -8.78505111e-01 6.19985998e-01
9.53807712e-01 -3.69137526e-01 -9.19938087e-01 3.61781046e-02
1.11020958e+00 1.22461826e-01 -6.30561888e-01 5.47197998e-01
3.83826107e-01 -7.79673696e-01 8.77473831e-01 -1.05123854e+00
3.18711430e-01 2.47289374e-01 -1.84119448e-01 -1.33173466e+00
-7.96926081e-01 9.87112746e-02 -8.60282421e-01 1.15195930e+00
1.80752680e-01 -4.33538228e-01 8.71972680e-01 3.36986899e-01
-3.74322534e-02 -7.70831823e-01 -9.73676264e-01 -1.01457858e+00
9.82761662e-03 -7.08177090e-01 5.62661886e-01 9.86346483e-01
6.49708152e-01 5.28524816e-01 -3.31575781e-01 -3.89267951e-01
4.42086875e-01 2.36394420e-01 3.25627983e-01 -1.46719909e+00
-2.90486887e-02 -4.64016616e-01 -4.45775211e-01 -8.26724291e-01
3.73247623e-01 -1.04953003e+00 -2.54381597e-01 -1.29319739e+00
5.64934194e-01 -6.67748153e-02 -1.03076422e+00 5.13369799e-01
-2.48770252e-01 -7.94629902e-02 1.26981094e-01 1.39278874e-01
-1.20348930e+00 8.11391652e-01 8.32903504e-01 -4.09520477e-01
-4.96673405e-01 2.79832691e-01 -6.40073597e-01 6.24205172e-01
9.24817026e-01 -5.07960141e-01 -6.34472072e-01 1.75289493e-02
-2.35668033e-01 2.00126674e-02 -2.70086467e-01 -1.20097780e+00
7.65137017e-01 -3.50770414e-01 4.93698597e-01 -7.50443757e-01
1.09598607e-01 -5.42404115e-01 -3.83594692e-01 7.94001520e-01
-6.38413072e-01 -3.32379252e-01 3.44614655e-01 7.60094285e-01
2.16846392e-02 -8.85778844e-01 9.60747778e-01 -1.86072290e-01
-1.39670968e+00 6.03133678e-01 -4.37883317e-01 9.52469185e-02
1.01417887e+00 6.36335611e-02 -6.00486279e-01 -2.19069302e-01
-4.86394942e-01 1.29762322e-01 5.35895005e-02 5.72712243e-01
8.68388653e-01 -1.30838871e+00 -1.66630402e-01 1.07801579e-01
6.83662653e-01 -5.51319718e-01 6.76447690e-01 3.41458052e-01
-1.62330866e-01 6.99090362e-01 -2.62597859e-01 -1.66827530e-01
-1.14758873e+00 9.99435186e-01 7.73735791e-02 -4.02728528e-01
-7.22359896e-01 6.32476032e-01 -5.03169179e-01 -2.01330096e-01
3.26199800e-01 2.64812112e-01 -8.06103587e-01 3.27706069e-01
1.19643939e+00 8.18135560e-01 1.45063862e-01 2.67556794e-02
-3.27393621e-01 9.22425911e-02 -5.45067072e-01 1.59099624e-01
1.18299925e+00 -5.88577501e-02 1.61921412e-01 8.54889691e-01
1.08508694e+00 -7.16471016e-01 -4.52329516e-01 -4.98268813e-01
1.13156259e-01 -2.65878260e-01 3.74089956e-01 -6.45613909e-01
-6.07811093e-01 9.68383789e-01 7.21169770e-01 4.79948595e-02
8.54123056e-01 -1.21190116e-01 8.18051279e-01 9.75259364e-01
4.82840329e-01 -1.59734070e+00 3.85780454e-01 1.07144678e+00
1.52519509e-01 -1.19479799e+00 -2.40458697e-02 3.44018936e-02
-4.88745779e-01 1.39135396e+00 8.31099331e-01 2.70533025e-01
1.00657392e+00 2.56519057e-02 -2.71508634e-01 -6.61332309e-02
-1.06117475e+00 -2.97702968e-01 3.73155594e-01 3.43815237e-01
1.31561533e-01 -4.41382006e-02 -3.04844290e-01 1.02840567e+00
3.96783084e-01 2.06781849e-01 1.77286386e-01 1.00273788e+00
-1.21258843e+00 -6.93102181e-01 1.07959069e-01 9.06996667e-01
-8.18322971e-02 -2.43424594e-01 6.10390026e-03 4.96162117e-01
-3.01239222e-01 1.08936536e+00 1.13269068e-01 -8.22355330e-01
2.37744257e-01 6.57804310e-01 9.62044820e-02 -8.56675327e-01
-1.44330695e-01 -5.00210047e-01 -1.36662245e-01 -8.54676962e-02
3.79710487e-04 1.75247371e-01 -1.30534065e+00 -4.95315731e-01
-6.40714049e-01 4.03913856e-01 4.42964077e-01 1.18032146e+00
4.59781826e-01 7.05302536e-01 7.55686462e-01 -3.50224644e-01
-6.31715059e-01 -1.26873434e+00 -7.64511585e-01 2.10561112e-01
-4.55431491e-01 -8.70617509e-01 -4.14465725e-01 -4.03747648e-01] | [10.190155029296875, 3.540890693664551] |
2266c139-1f38-49bb-8709-23861c5a2b04 | unified-fully-and-timestamp-supervised | 2209.00638 | null | https://arxiv.org/abs/2209.00638v2 | https://arxiv.org/pdf/2209.00638v2.pdf | Unified Fully and Timestamp Supervised Temporal Action Segmentation via Sequence to Sequence Translation | This paper introduces a unified framework for video action segmentation via sequence to sequence (seq2seq) translation in a fully and timestamp supervised setup. In contrast to current state-of-the-art frame-level prediction methods, we view action segmentation as a seq2seq translation task, i.e., mapping a sequence of video frames to a sequence of action segments. Our proposed method involves a series of modifications and auxiliary loss functions on the standard Transformer seq2seq translation model to cope with long input sequences opposed to short output sequences and relatively few videos. We incorporate an auxiliary supervision signal for the encoder via a frame-wise loss and propose a separate alignment decoder for an implicit duration prediction. Finally, we extend our framework to the timestamp supervised setting via our proposed constrained k-medoids algorithm to generate pseudo-segmentations. Our proposed framework performs consistently on both fully and timestamp supervised settings, outperforming or competing state-of-the-art on several datasets. Our code is publicly available at https://github.com/boschresearch/UVAST. | ['Mehdi Noroozi', 'Juergen Gall', 'Zico Kolter', 'S. Alireza Golestaneh', 'Nadine Behrmann'] | 2022-09-01 | null | null | null | null | ['action-segmentation'] | ['computer-vision'] | [ 9.65102434e-01 2.41653472e-01 -3.98423642e-01 -5.68412304e-01
-1.18967032e+00 -6.13866210e-01 7.12433279e-01 -2.54519731e-01
-5.74559808e-01 6.59270227e-01 3.35122675e-01 -7.79277906e-02
4.46967155e-01 -2.69154400e-01 -1.05883479e+00 -6.12001300e-01
2.10067675e-01 4.52821642e-01 4.34199512e-01 1.52996033e-01
2.45316610e-01 -1.14348568e-01 -1.21645999e+00 4.87566710e-01
6.40300155e-01 1.00102854e+00 1.12338468e-01 8.74689400e-01
1.54369250e-01 1.08250499e+00 -2.64199644e-01 -4.78467703e-01
5.48834026e-01 -8.83045018e-01 -9.51895952e-01 4.37098563e-01
5.11353970e-01 -3.59127671e-01 -3.85793030e-01 6.79499626e-01
3.46773446e-01 1.80955008e-01 3.41401219e-01 -1.27756488e+00
-2.62388170e-01 6.38591826e-01 -5.56441784e-01 1.24989837e-01
5.37969112e-01 2.17608020e-01 9.81978357e-01 -7.13593125e-01
7.20323503e-01 1.09418857e+00 5.48211694e-01 5.85487902e-01
-1.24823964e+00 -3.94663244e-01 2.96451926e-01 2.42986768e-01
-1.03593206e+00 -6.73449695e-01 5.53463519e-01 -5.62736571e-01
9.46907580e-01 5.65909557e-02 5.75800419e-01 1.49405599e+00
1.25332326e-01 1.11473155e+00 8.97916317e-01 -1.86559156e-01
1.82485893e-01 -3.13743532e-01 -3.14711452e-01 5.82685709e-01
-5.38769901e-01 -1.74055308e-01 -6.98953032e-01 6.10944070e-02
7.25808442e-01 5.78634106e-02 -6.64553046e-02 -2.58085459e-01
-1.50114167e+00 5.75932086e-01 -1.28582537e-01 -1.60594597e-01
-3.56616765e-01 5.00659525e-01 7.98925579e-01 8.61846730e-02
6.07779622e-01 -9.29623842e-02 -5.72956145e-01 -6.92006171e-01
-1.25611734e+00 4.10536468e-01 6.25398099e-01 1.22224820e+00
6.35057926e-01 -4.55558412e-02 -5.79838991e-01 4.87710595e-01
9.52641368e-02 1.70383900e-01 4.55331296e-01 -1.44202352e+00
6.64454401e-01 5.60688749e-02 1.82368919e-01 -2.93318301e-01
2.26685740e-02 -1.26369506e-01 -3.83436203e-01 -2.02436313e-01
3.54409248e-01 -2.34697655e-01 -8.32695723e-01 1.79883337e+00
3.82162571e-01 1.08320963e+00 -7.58393249e-03 9.69153821e-01
9.43497941e-02 6.37162447e-01 3.43958437e-02 -3.69220525e-01
1.15632129e+00 -1.54913998e+00 -6.29733205e-01 -1.63637832e-01
7.42460549e-01 -7.10872591e-01 1.00136805e+00 3.47470731e-01
-1.34959018e+00 -5.83736122e-01 -6.96599483e-01 -2.75125712e-01
1.64439768e-01 2.52000451e-01 2.72225767e-01 3.04773241e-01
-9.67572570e-01 6.85649872e-01 -1.40927827e+00 -3.49454314e-01
5.47314107e-01 1.93932444e-01 -2.15083212e-01 3.96600552e-02
-8.75174344e-01 4.12658542e-01 5.13343751e-01 -5.64037673e-02
-1.09254289e+00 -7.31647015e-01 -1.04684770e+00 -2.45332971e-01
7.56568611e-01 -8.76575887e-01 1.83384538e+00 -1.43124664e+00
-1.84114254e+00 6.89591289e-01 -4.14815545e-01 -8.81431580e-01
9.33862448e-01 -7.12104380e-01 8.99820179e-02 4.13135380e-01
2.89008170e-01 8.28721046e-01 8.98902893e-01 -8.09659064e-01
-8.23980093e-01 -2.45825164e-02 3.28669101e-02 2.98395574e-01
6.51743039e-02 2.06245080e-01 -7.55543470e-01 -9.31153238e-01
-3.37183654e-01 -1.16299534e+00 -2.80936152e-01 -1.28247002e-02
-3.51976097e-01 6.01647906e-02 7.78657317e-01 -8.32031250e-01
1.24503672e+00 -1.97937810e+00 6.01016939e-01 -3.89104635e-01
-1.28646225e-01 2.32890427e-01 -1.71285763e-01 6.49404287e-01
-1.44026563e-01 -2.26602629e-01 -7.47282565e-01 -9.04185891e-01
7.37674087e-02 2.52308398e-01 -2.04493731e-01 5.08950174e-01
3.40500236e-01 8.81358325e-01 -1.14194787e+00 -6.11372113e-01
2.09124014e-01 4.35330540e-01 -7.51829326e-01 3.89032573e-01
-5.50948977e-01 9.42786634e-01 -3.61071974e-01 5.14572501e-01
2.82218099e-01 -1.34770960e-01 2.04130590e-01 7.75342435e-02
-3.00919414e-01 4.76788342e-01 -8.04688931e-01 2.38698053e+00
-4.08882499e-01 4.64918941e-01 -2.10852608e-01 -1.25712943e+00
6.37790859e-01 5.36684513e-01 8.01653087e-01 -3.80037308e-01
4.76678275e-03 6.52720481e-02 -4.23671991e-01 -5.57677269e-01
5.40657640e-01 -7.29949772e-02 -2.02836290e-01 4.47645426e-01
2.55925208e-01 2.01461166e-01 3.63829553e-01 1.70090586e-01
1.21259654e+00 1.19608712e+00 2.76786864e-01 1.66813076e-01
4.66294467e-01 -3.28306295e-02 8.74554873e-01 4.43451613e-01
-2.10639343e-01 7.78208673e-01 6.48528337e-01 -1.22808918e-01
-1.22972226e+00 -8.25920999e-01 3.48218828e-01 1.09671545e+00
-8.14790204e-02 -6.33130550e-01 -1.07692802e+00 -7.99334288e-01
-3.58440191e-01 6.33397579e-01 -3.24838847e-01 1.04253471e-01
-8.26104343e-01 -1.72140777e-01 7.98381448e-01 7.38546968e-01
3.83544445e-01 -1.00801265e+00 -7.23112643e-01 3.38634729e-01
-4.33450699e-01 -1.62740338e+00 -9.78683472e-01 -1.02824576e-01
-9.91310000e-01 -9.62081075e-01 -8.22187662e-01 -6.97385073e-01
3.87026489e-01 1.43962130e-02 9.49201286e-01 -3.78320545e-01
-2.17003394e-02 3.60606939e-01 -6.45824492e-01 -2.04564650e-02
-4.37938720e-01 1.10212848e-01 -4.03290689e-02 3.52564305e-01
1.92481399e-01 -6.10392749e-01 -8.71844530e-01 3.05857122e-01
-1.05869710e+00 5.28377473e-01 5.05184948e-01 6.35341465e-01
9.86341178e-01 -4.35493112e-01 4.63748246e-01 -9.69036639e-01
1.02760792e-01 -5.22668123e-01 -5.32018602e-01 -1.40268803e-02
-2.70654529e-01 9.11596604e-03 8.80096436e-01 -2.71969587e-01
-1.08564436e+00 4.92549449e-01 -3.50867286e-02 -7.89741099e-01
-3.06614429e-01 1.74602270e-01 -7.70631060e-02 4.03773278e-01
1.17945708e-01 4.66722310e-01 -8.97193793e-03 -4.93937761e-01
4.17205334e-01 6.10480547e-01 7.25265861e-01 -5.28724611e-01
6.20398104e-01 5.58520317e-01 -1.08700231e-01 -4.44775671e-01
-8.62677276e-01 -6.04846895e-01 -1.04321754e+00 -2.44474232e-01
1.13388419e+00 -1.10199761e+00 -5.02686262e-01 7.19973207e-01
-1.17965627e+00 -8.54916871e-01 -3.12799692e-01 3.38469476e-01
-1.24454236e+00 6.71152771e-01 -6.59489393e-01 -6.69045329e-01
-1.92982137e-01 -1.23464990e+00 1.60393262e+00 -8.24405476e-02
-1.46805972e-01 -8.52712035e-01 1.97597057e-01 7.25349844e-01
-7.93893710e-02 5.52953362e-01 1.78796977e-01 -6.18654788e-01
-7.00709939e-01 2.07981709e-02 4.76960950e-02 5.95828414e-01
9.89033058e-02 -1.32068992e-01 -7.06356227e-01 -1.59237012e-01
-1.10598266e-01 -4.46979612e-01 9.88725543e-01 3.06116790e-01
1.18836141e+00 -5.19916296e-01 -1.95905775e-01 7.16874719e-01
1.24744213e+00 1.61551178e-01 7.54901171e-01 2.57260144e-01
8.02396297e-01 4.45844829e-01 1.13394356e+00 7.44111121e-01
6.44014955e-01 1.00953329e+00 2.64633328e-01 1.69953227e-01
3.16209607e-02 -5.32294035e-01 8.16696584e-01 7.22295642e-01
-1.28135651e-01 -4.29233789e-01 -7.35011876e-01 5.80308557e-01
-2.33378363e+00 -1.14800406e+00 3.88316326e-02 2.19162321e+00
8.39270651e-01 7.24992156e-02 5.29336035e-01 -2.29722895e-02
6.59136415e-01 4.52198237e-01 -6.74660265e-01 -2.92356730e-01
2.41461948e-01 3.29923034e-01 7.56609023e-01 4.78952110e-01
-1.26368439e+00 1.19179702e+00 5.23317146e+00 7.68110037e-01
-1.00576723e+00 3.07741463e-01 5.31959414e-01 -3.44667643e-01
-1.87918264e-02 2.84756988e-01 -5.93374908e-01 7.32724428e-01
1.32721281e+00 1.92784593e-02 3.65028381e-01 5.23230910e-01
7.74847388e-01 -3.60384360e-02 -1.35345101e+00 8.71937990e-01
1.57835290e-01 -1.27590454e+00 -1.49323329e-01 -8.52963030e-02
7.61752188e-01 5.00898175e-02 -2.14825884e-01 6.62913993e-02
1.93803441e-02 -8.19782078e-01 1.14216268e+00 6.05212331e-01
9.30493593e-01 -3.77500504e-01 4.18275118e-01 2.02812269e-01
-1.37382996e+00 -5.86574599e-02 4.59197648e-02 -1.17285080e-01
8.44690204e-01 2.03061223e-01 -6.29203618e-01 7.53775358e-01
3.58516484e-01 1.32566583e+00 -2.23614365e-01 8.24889600e-01
-3.56909305e-01 9.33048844e-01 -5.61489984e-02 6.45878196e-01
4.86491144e-01 -3.49235535e-01 4.89355028e-01 1.43123364e+00
3.70188653e-01 4.53833155e-02 4.90471035e-01 4.05864209e-01
-5.94854914e-02 -1.01320378e-01 -3.12797129e-01 -1.67965144e-01
3.41712475e-01 1.02760136e+00 -6.60154700e-01 -6.20145917e-01
-6.44516289e-01 1.63840246e+00 4.49183062e-02 2.13040501e-01
-1.30102682e+00 -8.76653343e-02 7.50059724e-01 1.97955042e-01
7.64696360e-01 -2.61403114e-01 4.41807136e-02 -1.35063922e+00
1.97110206e-01 -9.46537495e-01 3.77792507e-01 -6.22341156e-01
-7.68554270e-01 3.75556320e-01 1.26771480e-01 -1.63535845e+00
-5.42015791e-01 -1.74321324e-01 -5.57501316e-01 4.27136332e-01
-1.59932029e+00 -1.28175330e+00 -6.98001459e-02 5.05188823e-01
1.22465324e+00 1.78906143e-01 4.20271635e-01 3.51809829e-01
-5.65160632e-01 5.27329862e-01 -7.56518077e-03 9.41146165e-02
9.04250383e-01 -1.20924985e+00 8.24748516e-01 9.79891539e-01
-4.92825322e-02 4.09944318e-02 6.93341494e-01 -5.88805616e-01
-1.38660526e+00 -1.44247627e+00 8.36983442e-01 -4.03140515e-01
7.13417828e-01 -5.81570268e-01 -5.95184207e-01 1.18035221e+00
3.76731664e-01 1.48588866e-02 5.91033876e-01 -6.63189054e-01
-1.64660633e-01 6.17423058e-02 -8.30101311e-01 5.04506469e-01
1.40111983e+00 -4.02954489e-01 -3.85985434e-01 3.83958131e-01
9.51695502e-01 -7.37353504e-01 -9.50876415e-01 2.30860934e-01
6.77781582e-01 -9.05584395e-01 7.33568668e-01 -5.43031812e-01
7.35042751e-01 -3.98785472e-01 -1.48269266e-01 -9.62452531e-01
1.13980444e-02 -1.22896051e+00 -2.53559381e-01 1.11192429e+00
1.89633116e-01 -3.27903122e-01 8.84662688e-01 1.80093899e-01
-5.13486147e-01 -8.02561760e-01 -9.53049660e-01 -8.32196712e-01
-2.45660141e-01 -5.53353608e-01 2.08007410e-01 5.48231483e-01
-9.15207118e-02 1.97238013e-01 -7.88729727e-01 1.17212407e-01
6.25551403e-01 1.30972579e-01 8.57764363e-01 -5.62279880e-01
-6.66534841e-01 -5.78117184e-02 -4.19778824e-01 -1.51761591e+00
4.50054526e-01 -7.50737548e-01 2.44989753e-01 -1.38492155e+00
1.86144933e-01 8.17737952e-02 -8.78011882e-02 5.21233559e-01
-2.31740065e-02 4.56943452e-01 3.58716488e-01 1.87743440e-01
-9.10838664e-01 7.37484038e-01 1.06557918e+00 1.34319827e-01
-1.48340046e-01 4.54860739e-02 -3.16769630e-01 5.98654568e-01
8.40781271e-01 -5.40233970e-01 -5.29217958e-01 -5.83931088e-01
-1.64564416e-01 5.35056114e-01 4.11227584e-01 -1.00584614e+00
2.82708853e-02 -2.86763489e-01 -1.50230274e-01 -4.00616527e-01
4.76631105e-01 -5.48050940e-01 2.46597260e-01 2.97918022e-01
-6.00470662e-01 9.70694125e-02 1.65155157e-02 7.32145011e-01
-2.49049067e-01 -3.07491585e-03 5.84430873e-01 -1.38197720e-01
-7.03635752e-01 5.17381430e-01 -3.70093018e-01 2.18753368e-01
1.25689197e+00 -4.00278538e-01 -4.60853651e-02 -3.49889457e-01
-7.62938976e-01 2.53487110e-01 7.28741884e-01 5.01419306e-01
4.84793931e-01 -1.22535133e+00 -7.85654187e-01 -1.05020851e-01
1.77805461e-02 7.02406913e-02 1.75730407e-01 1.20052397e+00
-3.76695395e-01 4.47172433e-01 -1.54541880e-01 -7.01348126e-01
-1.28413081e+00 3.49468261e-01 1.21112972e-01 -2.85207242e-01
-7.20255494e-01 6.58411622e-01 2.19006494e-01 -2.41810769e-01
1.80148751e-01 -4.12337989e-01 2.05823019e-01 -6.61287680e-02
3.36187065e-01 4.94906336e-01 -2.60025322e-01 -9.14121270e-01
-2.80165195e-01 4.20098901e-01 -3.07308435e-02 -3.82491201e-01
1.30596733e+00 -1.85227752e-01 1.76680669e-01 6.10146105e-01
1.26446891e+00 -3.44331920e-01 -1.93056822e+00 -1.24566041e-01
2.66069081e-02 -4.71168518e-01 -3.38420361e-01 -4.88083452e-01
-9.78745222e-01 5.98555148e-01 2.12829545e-01 -3.88672292e-01
1.27294564e+00 -6.96829855e-02 1.30098927e+00 2.43990514e-02
2.11095303e-01 -1.11560988e+00 1.67587280e-01 4.53316718e-01
4.45743084e-01 -1.10367417e+00 -2.24333167e-01 -3.43115419e-01
-9.90292609e-01 9.05700564e-01 4.09064025e-01 -2.58129627e-01
1.80424452e-01 3.06639761e-01 -2.61064991e-03 1.67939886e-01
-1.07804930e+00 -8.12079981e-02 9.85464826e-03 3.99200588e-01
6.60099804e-01 -6.90351948e-02 -5.29119611e-01 3.20475966e-01
4.49948870e-02 6.14263237e-01 6.96152747e-01 1.09481549e+00
-1.16210662e-01 -1.40322399e+00 -6.10652305e-02 2.43645161e-01
-6.47643209e-01 -1.65409505e-01 -1.19264670e-01 4.51458752e-01
2.12134942e-02 7.50205040e-01 1.66254640e-01 -2.93592960e-01
2.08005264e-01 3.42694372e-01 4.37756687e-01 -8.16908538e-01
-5.33303082e-01 3.36854458e-01 1.57922089e-01 -1.08799338e+00
-8.53657544e-01 -1.23795915e+00 -1.28586173e+00 1.16613224e-01
3.47572640e-02 -9.69675705e-02 3.86719316e-01 1.01861894e+00
4.87370342e-01 5.20609796e-01 6.27503872e-01 -1.11357594e+00
-4.39203233e-01 -9.23553586e-01 -1.35944575e-01 4.05871570e-01
3.03081751e-01 -3.61463755e-01 5.81767000e-02 7.81383038e-01] | [8.478521347045898, 0.5468856692314148] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.